Institut de Ciències del Mar...

224

Transcript of Institut de Ciències del Mar...

Institut de Ciències del Mar (CSIC)

Universitat Politècnica de Catalunya

Effects of different allochthonous carbon sources on marine bacterioplankton diversity and function

Memòria presentada per optar al grau de Doctor en Ciències del Mar per la

Universitat Politècnica de Catalunya per

Itziar Lekunberri García

Vist-i-plau del director de tesi

Dr. Josep M. GasolInvestigador–CSIC

“Marine food web” by Joaquim Buil

“Lo esencial es invisible a los ojos”

“El Principito” (Antoine de Saint-Exupéry)

A mis padres y a mi hermanica, Maitane

7

11

25

61

95

119

149

171

193199

207

217

Contents

SUMMARY/ RESUM/ RESUMEN/LABURPENA

GENERAL INTRODUCTION

Chapter I Increased marine bacterioplankton diversity detected by combining culturing and culture-independent techniques

Chapter II Changes in bacterial activity and community composition caused by realistic exposure to oil in mesocosms experiment

Chapter III Effects of different types of phytoplankton blooms caused by nutrient additions on bacterial heterotrophic activity and assemblage composition in sweater mesocososms

Chapter IV Effects of Sahara Dust deposition on marine microbial abundance, bacterial community structure and ecosystem function

Chapter V Transplant experiments show dominance of bottom-up control of prokaryote abundance, heterotrophic production and bacterial community structure in the NE Atlantic Ocean

SYNTHESIS AND DISCUSSION

ANNEXES

I: Traditional methods for bacterial culturing

II: Protocols for Scanning Electronic Microscopy (SEM) of bacteria

REFERENCES (Introduction, discussion and annexes sections)

AGRADECIMIENTOS

7

SUMMARY

Bacteria are the most abundant organisms on Earth and mediate many critical ecosystem processes. This indicates the great importance of the knowledge of the identity and population dynamics of the indigenous microorganisms. Initial studies in aquatic microbiology were carried out by traditional cultivation techniques. Later on, cell counts by epi uorescence techniques demonstrated how abundant bacteria were, and molecular techniques were developed and used to detect the most representatives bacteria in the environment. This is mainly because the culture media have nutrient contents (both organic and inorganic) which are several times more concentrated than those in the natural environment. This makes retrieving dominant organisms from an oligotrophic environment not an easy task, and generates selectivity for particular bacterial types, which are not very representative. A similar selection of bacteria also occurs in the natural environment, at another scale, when different allochthonous inputs take place. These can be due to natural processes, such as river run-offs, aerosol deposition, etc., or may have an anthropogenic origin, such as oil spills. In this thesis we evaluate the effect of several of these additions on bacterial community structure and function.

There is an increasing interest in the isolation of naturally dominating organisms since it provides a better information about the metabolic capabilities of the organisms that are growing in the environment. Thus, we decided to evaluate how representative were the isolates obtained in the Mediterrranean Sea as compared to the sequences retrieved by molecular techniques such as cloning and DGGE ngerprinting. Our analysis highlighted the selectivity created by the growth media, but simultaneously allowed to increase our knowledge about the diversity in our study area, and provided a means for looking into the “rare microbial biosphere”, which cannot be analyzed with other techniques.

To address the question of how nutrient addition modi es bacterial community structure and affects microbial function, we performed different mesocosm experiments, where we added different types of nutrients to water collected from Blanes Bay (NW Mediterranean Sea). Another mesocosm experiment was carried out to evaluate the effect of the oil spill that took place after the Prestige tanker accident in front of the Galician coast. Mesocosms experiments have been widely used to evaluate addition effects as a means to reproduce of the natural conditions. During these experiments different environmental parameters were measured and also bacterial production and abundance were monitored. Bacterial community structure was determined by the DGGE ngerprinting technique, which allows the comparison of different samples and further analysis of the dominant richness by the Shannon-Weaver index, based on the presence and absence of bands. These analyses showed the in uence of grazers on the bacterial selection after the different additions and, for example, explained the effects of Saharan Dust deposition on microbial function and community metabolism, as an effect of the inorganic phosphorus and organic carbon contributed by the dust. In the oil spill mesocosm experiments, effects on bacterial structure, production and abundance were detected only with additions four times higher than the concentration found after the Prestige tanker accident.

We also include an experimental study carried out in the North Atlantic during two different cruises, in which using dyalisis bags and transplant experiments we tested whether bottom-up conditions were stronger than top-down factors in determining microbial function and community structure. The results showed differences between experiments, but a general higher effect of bottom-up conditions than the effect of predators.

The different experimental work included in this thesis allows for some generalizations, but also show some discrepancies, which we discuss in the last chapter.

8

RESUM

Els bacteris són els organismes més abundants a la Terra i els encarregats de pariticipar multitud de processos que són crítics per a l´entorn. Això fa que sigui de gran importància el coneixement de la identitat i dinàmica de les poblacions de microorganismes indígenes. Els primers estudis en microbiologia es van realitzar mitjançant tècniques de cultiu tradicional. Més endavant, tècniques de comptatge per epi uorescència van mostrar com n’eren d’abundants els bacteris, i es van emprar també tècniques moleculars per a detectar els bacteris més representatius del medi. Però les concentracions de bacteris en els medis de cultiu (tant orgànics com inorgànics) són diverses vegades superiors a les que es troben en el medi natural. En conseqüència, l’obtenció dels cultius dels organismes dominants en un medi oligotrò c no és una tasca fàcil i acaba seleccionant bacteris que no són gaire representatius. Una selecció similar de bacteris també es dóna en el medi natural, a diferent escala, quan hi ha aports de diferents compostos al·lòctons. Aquests es poden deure a processos naturals com ara la descàrregues de rius, deposicions mitjançant aerosols, etc., o bé poden ser d’origen antropogènic, com els vessaments de petroli. En aquesta tesi avaluem l’efecte de diverses d’aquestes addicions en la diversitat, l´estructura de les comunitats i la funció bacterianes.

Hi ha un interès creixent en l’aïllament d’organismes dominants en el medi, ja que això permet tenir més informació sobre les capacitats metabòliques dels organismes que hi estan creixent. Per tant, vam decidir mesurar ns a quin punt són representatius els aïllats obtinguts en el mar Mediterrani comparant-los amb les seqüències obtingudes a través de tècniques moleculars com ara les biblioteques de clons i tècniques coma ara la DGGE ngerprinting. La nostra anàlisi destaca la selectivitat creada pel medi marí, però alhora permet tenir un coneixement més ampli sobre la diversitat en la nostra àrea d’estudi, i proporciona mitjans per a explorar la “biosfera microbiana rara”, que no pot ésser analitzada mitjançant altres tècniques.

Per a abordar la pregunta de com l’addició de nutrients selecciona la diversitat bacteriana i afecta la funció microbiana, vam realitzar una sèrie d’experiments de mesocosmos, on es van afegir diferents tipus de nutrients a aigua recollida a la Badia de Blanes (Nord-oest del mar Mediterrani). Un altre experiment de mesocosmos va ser realitzat amb l’objectiu d’avaluar l’efecte del vessament de petroli que va donar-se després de l’enfonsament del petrolier Prestige davant les costes gallegues. Els experiments de mesocosmos han estat utilitzats àmpliament per a determinar els efectes de les addicions amb l’intenció de simular les condicions naturals. Durant aquests experiments es van avaluar diferents paràmetres ambientals i també es van monitoritzar la producció i l’abundància bacterianes. L’estructura de la comunitat bacteriana va ser determinada amb la tècnica de DGGE

ngerprinting, que permet la comparació de diferents mostres i una anàlisi posterior de la diversitat dominant mitjançant l’índex de Shannon-Weaver, basat en la presència i absència de bandes. Aquestes anàlisis mostren la in uència dels depredadors en la selecció bacteriana després de les diferents addicions i, per exemple, explica els efectes de la deposició de pols del Sàhara en la funció microbiana i en el metabolisme de la comunitat, especialment per causa del fòsfor inorgànic i el carboni aportats amb la pols. En els experiments de vessament de petroli, es van poder detectar canvis en l’estructura, producció i abundància bacterianes només en el cas d’addicions quatre vegades més grans que la concentració trobada després de l’accident del Prestige.

També incloem un experiment realitzat a l’Atlàntic Nord durant dues campanyes diferents, en les quals es van utilitzar bosses de diàlisi i experiments de trasplantaments per a demostrar com les condicions de bottom-up van ser més grans que la regulació top-down en la determinació de la funció i estructura de la comunitat microbiana. Els resultats mostren diferències entre experiments, però en general es van detectar efectes més grans de bottom-up que no pas deguts a depredadors.

Finalment, els diferents treballs experimentals inclosos en aquesta tesi permeten fer algunes generalitzacions però també mostren algunes discrepàncies, fet que es discuteix en l’últim apartat.

9

RESUMEN

Las bacterias son los organismos más abundantes de la Tierra y los encargados de mediar en multitud de procesos críticos para el medio ambiente. Esto hace que sea de gran importancia el conocimiento de la identidad y dinámica de poblaciones de los microorganismos indígenas. Los primeros estudios en microbiología se realizaron mediante técnicas de cultivo tradicionales. Más tarde, técnicas de contaje por epi uorescencia mostraron lo abundantes que eran las bacterias, y técnicas moleculares se usaron también para detectar las bacterias más representativas del medio. Pero la concentración de nutrientes en los medios de cultivo (tanto orgánicos como inorgánicos), son varias veces superiores a los que se encuentran en el medio natural. En consecuencia, la obtención de los organismos dominantes en un medio oligotró co no es una tarea fácil y acaba seleccionando organismos bacterianos, que no son muy representativos. Una selección similar de bacterias ocurre también en el medio natural, a diferente escala, cuando hay aportes de diferentes compuestos alóctonos. Esto puede ser debido a procesos naturales tales como descargas de ríos, deposiciones de aerosoles, etc., o bien pueden ser de origen antropogénico, como los derrames de petróleo. En esta tesis evaluamos el efecto de varias de estas adiciones en la diversidad y función bacterianas.

Hay un creciente interés en el aislamiento de organismos dominantes en el medio, ya que esto permite tener mayor información sobre las capacidades metabólicas de los organismos que están creciendo. Por tanto, decidimos evaluar hasta qué punto son representativos los aislados obtenidos en el mar Mediterráneo comparándolos con las secuencias obtenidas mediante técnicas moleculares como bibliotecas de clones y DGGE ngerprinting. Nuestro análisis destaca la selectividad creada por el medio de cultivo, pero a la vez permite ampliar el conocimiento sobre la diversidad en nuestra área de estudio, y proporciona medios para explorar la “biosfera microbiana rara”, que no puede ser analizada mediante otras técnicas.

Para abordar la cuestión de cómo la adición de nutrientes selecciona la diversidad bacteriana y afecta a la función microbiana, realizamos diferentes experimentos de mesocosmos, donde añadimos diferentes tipos de nutrientes a agua recogida en la Bahía de Blanes (Noroeste del mar Mediterráneo). Otro experimento de mesocosmos fue llevado a cabo para evaluar el efecto del derrame de petróleo que tuvo lugar después del hundimiento del petrolero Prestige frente a las costas gallegas. Los experimentos de mesocosmos han sido ampliamente utilizados para determinar los efectos de adiciones con la intención de simular las condiciones naturales. Durante estos experimentos se evaluaron diferentes parámetros ambientales y también se monitorizó la producción y la abundancia bacterianas. La estructura de la comunidad bacteriana fue determinada por la técnica de DGGE ngerprinting, que permite la comparación de diferentes muestras y un análisis posterior de la diversidad dominante mediante el índice de Shannon-Weaver, basado en la presencia y ausencia de bandas. Estos análisis muestran la in uencia de los depredadores en la selección bacteriana después de las diferentes adiciones y, por ejemplo, explica los efectos de la deposición de polvo del Sáhara en la función microbiana y en el metabolismo de la comunidad, especialmente por el fósforo inorgánico y el carbono aportado por el polvo. En los experimentos de derrame de petróleo, sólo se detectaron efectos en la estructura, producción y abundancia bacterianas en el caso de adiciones cuatro veces mayores que la concentración encontrada después del accidente del petrolero Prestige.

También incluimos un experimento llevado acabo en el Atlántico Norte durante dos campañas diferentes, en las que se usaron bolsas de diálisis y experimentos de trasplantes para demostrar cómo las condiciones de bottom-up fueron mayores a los factores de top-down en la determinación de la función y estructura de la comunidad microbiana. Los resultados mostraron diferencias entre experimentos, pero en general se detectaron mayores efectos de bottom-up que los debidos a predadores.

Los diferentes trabajos experimentales incluidos en esta tesis permiten hacer algunas generalizaciones, pero también muestran algunas discrepancias, lo que se discute en el último apartado.

10

LABURPENA

Bakterioak Lurreko izakiriki oparoenak dirak eta ingurumeneko prozesu kritiko askotan bitartekari. Guzti honek izaugarrizko garrantzia ematen dio lekuko mikroorganismoen nortasuna eta taldeko dinamikaren berri izateari. Lehen mikrobiologiako ikerketak ohiko kultibo tekniken bidez egin ziren. Ondoren, epi uoreszentzia bidez egindako zenbatze teknikek bakterioen ugaritasuna azaldu zuten, eta teknika molekularrak bakterio adierazgarrienak aurkitzeko erabili izan ziren. Baino elikagai bilketa kultibo baliabideetan (nahiz organiko nahiz inorganikoak) hainbat aldiz haundiagoa da ingurune naturalean aurkitu daitekeena baino. Ondorioz, organismo gainartzaileen lorpena ingurune oligotro koetan ez da lan erreza eta adierazgarriak ez diren bakterioak aukeratzen bukatzen da. Antzeko aukeraketa bat gertazten da ingurune naturalean, maila ezberdin batean, bertakoak ez diren konposatuen ekarpenak daudenean. Hau prozesu naturalen ondorio izan daiteke hala nola erreken isuria, aerosolen uzteen bidez, e.a. edo gizakiaren jardueraren ondorioz, petrolio isuriak bezela. Tesi honetan hauetako gehitze batzuk bakterioen dibertsitate eta funtzioetan duten eragina aztertzen dugu.

Geroz eta jakin min handiagoa dago bere ingurunean organismo gainartzaileen isolamenduan, honek hazten ari diren organismoen ahalmen metabolikoari buruzko informazioa edukitzen lagunduko baitu. Beraz, aztertzea erabaki duguna zera da, Mediterraneo itsasoan isolatutakoen adierazgarritasuna, klonen liburutegi eta DGGE ngerprinting teknika molekularren bidez lortutako sekuentziekin alderatuz. Gure analisiak kultibo baliabideen

bidez sortutako hautaketa nabarmentzen du, baina era berean gure azterketa eremuko dibertsitateari buruzko ezaguera haunditzea laguntzen digu eta beste tekniken bidez aztertu ezin daitekeen “mikrobiar biosfera arraroa” aztertzeko baliabideak ahalbideratzen dizkigu.

Elikagaien gehitzeak nola bakterioen dibertsitatea aukeratzen eta mikrobiar funtzioan eragiten duen aztertzeko, mesokosmosean saiakuntza ezberdinak egin ditugu, non Blaneseko badian (Mediterraneo itsasoko ipar-mendebaldea) hartutako ur laginei elikagai ezberdinak gehitu dizkiogun. Mesokosmoseko beste saiakuntza bat galizar kostaren parean egin zen, Prestige petroliontzia ondoratu ondoren, petrolio isuriak duen eragina aztertzeko. Mesokosmoseko saiakuntzak ugariak izan dira gehitzeen ondorioak zehazteko, egoera naturalak simulatzeko intentzioarekin. Saiakuntza hauetan ingurune parametro ezberdinak aztertu ziren, era berean bakterio produkzioaren eta ugaritasunaren datuak bildu ziren. Bakteriar taldearen estruktura DGGE ngerprintingteknikaren bidez nkatu zen, teknika honek lagin ezberdinak alderatzea eta dibertsitate gainartzailearen geroagoko analisi bat ahalbideratzen ditu Shannon-Weaver indizearen bidez, banden presentzia eta ausentzian oinarrituz. Analisi hauek, gehiketa ezberdinen ondoren, harrapariek bakterioen aukeraketan duten eragina erakusten dute eta, adibidez, Saharako hautsen pilatzeak mikrobiar funtzioetan eta taldeko metabolismoan duen eragina argitzen du, bereziki hautsak ekartzen duen fosforo inorganiko eta karbonoaren ondorioz. Petrolio isurien saiakuntzetan, bakterioen estruktura, produkzio eta ugaritasunean ondorioak antzeman ziren, isuriak Prestige petroliontziaren ezbeharraren ondoren bilatutako kontzentrazioak laukoizten ziren gehitzeetan soilik.

Bi kanpaina ezberdinetan, Atlantiko Iparraldean egindako saiakuntza bat ere gehitzen dugu, non dialisi boltsak eta trasplante saiakuntzak erabili ziren, mikrobiar komunitatearen funtzioa eta estruktura zehazteko garaian, bottom-up egoerak top-down faktoreak baino gehiago izan zirela egiaztatzeko. Emaitzek saiakuntzen arteko ezberdintasunak erakutsi zituzten, baina orokorrean bottom-up faktoreen eragin handiagoak antzeman ziren harrapariei dagokiena baino.

Tesi hontan azaltzen diren saiakuntza ezberdinak jeneralizazio batzuk aurkeztea zilegitzen digu, nahiz eta desadostasun batuk agertu, hauek azken atalean eztabaidatzen dira.

INTRODUCTION

Los grandes conocimientos engendran las grandes dudas

Aristóteles

13

Introduction

GENERAL INTRODUCTION

BACTERIA IN THE ENVIRONMENT

The ocean is probably the oldest and largest ecosystem on the planet. Microorganisms living there

are the closest living descendants of the original forms of life on Earth (Hunter-Cervera et al. 2005).

By increasing the oxygen concentration in the atmosphere they shaped a new aerobic environment

where plants, animals and all other life forms evolved.

Bacteria are the most abundant and species-rich groups of microorganisms. They mediate many

critical ecosystem processes that sustain life on Earth, and other organisms are unable to achieve

(Horne-Devine et al. 2003). Some examples of processes carried out exclusively by microorganisms

are the key processes that set the pace of the nitrogen cycle. Nitri cation, denitri cation and half of

the microbially-mediated nitrogen xation occur in the oceans. Some bacteria play a role in cloud

formation by cycling compounds such as dimethylsul de into the atmosphere. Other prokaryotes

adapted to extreme environments are known to participate in the constant ef ux of methane from

the surface of the oceans, although the process is not yet entirely understood. These are only some

of the physiological characteristics that have been discovered so far, but the huge microbial diversity

indicate that there may be many more metabolic capabilities yet to be discovered.

APPROACHES TO PROKARYOTE DIVERSITY: CULTURING vs MOLECULAR

Early marine microbiologists were convinced that a high percentage of abundant pelagic could

be recovered on substrate-amended agar plates (Zobell 1943). Lately, epi uorescense microscopy

revealed that the bacteria readily cultivable on standard rich growth media were only about 0.1 percent

of the observable bacteria. (Ferguson et al 1984). Jannash & Jones (1959) reported such observations

in seawater, although this problem was universal, and was later named “the great plate anomaly”

(Staley & Konopka 1985).

14

Introduction

In the mid 1980s molecular methods were introduced in Microbial Ecology. Somehow surprisingly,

the results generated by these new techniques did not match the results obtained by culturing. They

revealed that some dominant bacteria are not cultivable using standard methodologies, and others

that are easily cultivable occur at such low abundance in the natural environment that molecular

techniques cannot detect them. Molecular tools made possible to obtain the genetic sequence data of

the dominant organisms in spite of not being able to culture them.

While by ngerprinting techniques not more than 100 sequences are commonly detected, it has

been shown that with a higher effort in cloning, ca1600 sequences can be obtained (Acinas et al. 2004).

Recently appeared different molecular approaches allow massive sequencing of the bacterioplankton

community. Some examples are the study by Venter et al. (2004) in the Sargasso Sea that used a

brute-force shotgun sequencing approach that permit them to estimate that the sample analyzed

contained ca.1800 different bacterial taxa. Sogin et al. (2006), using a more powerful and cheaper

sequencing approach to analyze deep-water samples from the North Atlantic, estimating richness to

be of thousand taxa. Recent molecular studies thus show that the taxonomic composition of natural

bacterial communities comprises a tremendous diversity, although relatively few species appear to

form the bulk of the biomass (Pedrós-Alió 2006).

Fig.1 Plots of number of individuals versus taxon rank. The total curve represents biodiversity and is postulate by Pedros-Alió (2006) to be composed of two sections: dominant diversity (red) and to rare taxa (blue), which survive in the ecosystem at low abundances. Molecular techniques can presumably access most taxa in the diversity part of the curve and the biodiversity tail is usually not accessible, however isolation can retrieve some taxa from this part of the curve.

15

Introduction

Margalef (1994) distinguished between ecosystem diversity (a property of the ecosystem) and

biodiversity, de ned as the total genetic information on Earth (or any part of it) while diversity, in

turn, would be the components that are active and abundant at one particular time and place. Based

on this statement, Pedrós-Alió (2006) suggested to consider the taxa retrieved by PCR as a proxy for

the diversity of an ecosystem. However, rare taxa (the seed bank) that form the long tail of a rank-

abundance curve are necessary to complete the biodiversity. Also, it is important to note that the

number of representatives of a given group is not necessarily linked to the importance of that group

in the functioning to the community (Fig.1).

In relation with this idea, Chapter 1 of this thesis focuses in trying to increase the diversity

previously estimated by molecular approaches (DGGE and clone libraries) at the Blanes Bay Microbial

Observatory (BBMO) with the diversity that would be retrieved by traditional culturing techniques.

Questions at play are not only a (bacterial) group by group comparison of the differences between

what is retrieved by culturing and what is retrieved by molecular techniques, but also whether there

are new bacterial groups appearing, whether some cultured bacteria are good representatives of groups

of sequences appearing in situ, etc. A special attention is given to the microdiversity patterns present

in the culture collection vs. in the molecular techniques. Microdiversity can be de ned as the diversity

measured above the division of species: sequences that are more than 97% similar in their 16S rRNA,

but that they are not exactly identical

RELATING BACTERIAL DIVERSITY AND FUNCTION

There is an increasing interest in relating bacterial diversity to their function in the environment.

However, bacterial activity has mostly been studied in bulk. There are numerous approaches useful

for establishing a functional role for bacterioplankton species and some are used for determining the

activity of microorganisms that are uncultured. Some examples are the use of bromodeoxyuridiene

uptake to identify growing cells (Urbach et al. 1999), and 13C incorporation into lipids to identify

organisms that assimilate acetate. The MARFISH technique combines autoradiography and in situ

hybridation to identify the cells that are able to assimilate speci c organic compounds (Cottrell &

Kirchman 2000). Another approach that would allow us to investigate the organisms’ physiology is to

isolate it and sequence the genome. Indeed, sequencing of complete microbial genomes, now routinely

reveals organisms capacities to perform unexpected metabolic functions, that were impossible to

identify before.

16

Introduction

Traditional methods (e.g. culturing on ZobBell’s marine growth media) have limitations, because

this media has ca. 170-fold more DOC than natural seawater and thus, selects for “eutrophic” bacteria,

those that can develop in such a large amount of carbon. Consequently, the isolation of microbes

found in oligotrophic environments was a great challenge. Many culture techniques have been used

in order to obtain isolates of the most representative members of the bacterial communities. One

approach is the “dilution to extinction” culturing method for the isolation of oligotrophic bacteria

in seawater (Button et al. 1993) and later modi ed by Rappé et al. (2002). Other approaches are

microencapsulation techniques, and the use of low-nutrient agar plates (Eilers et al. 2001). Relatively

new strains that represent many of the most abundant marine clades have been isolated. This is the

case of members of the SAR11 clade, which is the most abundant group of bacteria detected in

seawater DNA by gene cloning and by FISH (Morris et al. 2002, Rappé et al. 2002).

Cultures are important because they provide complete genomes and the means to test the hypotheses

that emerge from genomic data. Cultures supported by predictions from genomic data, together with

metagenomics are a powerful combination. The major success of metagenomics was the discovery

of the light-driven, proton pump proteorhodopsin in the fragmentary genome data of SAR86 (Bejá

et al. 2000). This was the rst report about the presence of such retinaldehyde protein in the domain

bacteria, and this report sparked considerable interest in this potentially important mechanism for

capturing light energy in marine ecosystems.

The Moore Foundation had the initiative to sequence the whole genome of isolates from different

environments, and 9 of our isolates were selected, giving the possibility to learn more about their

function in the environment. For example the manual annotation of M152 (Polaribacter sp.) revealed

the presence of proteorhodopsine (Gonzalez et al. 2008). Furthermore inM134 (Dokdonia sp.) the

bacterial light-dependent proton pump was shown to provide M134 with phototrophic potential

(Gómez-Consarnau et al. 2007).

MARINE MICROBIAL ECOSYSTEMS

Marine microorganisms are known to live in every corner of the ocean, as they are metabolically

versatile to adapt to any environment. The microbial ecologists’ trade is to examine whether bacterial

communities differ according to the type of environment. Although there are likely millions of

species of bacteria, we are only starting to investigate patterns in their diversity and which are the

forces that governs these patterns (Ward et al. 1998). Decades ago it was unclear whether marine

17

Introduction

microbial species are either cosmopolitan, or are more provincial and limited to certain geographical

areas. Fenchel & Finlay (2004), after Baas-Becking (1934), proposed that the majority of marine

planktonic microorganism were cosmopolitan. The arguments for the “everything is everywhere” idea

is that spatial barriers do not occur due to the great capability for dispersal into new locations (Finlay

2002, Fenchel & Finlay 2004, Collins et al. 2001). The Baas-Becking (1934) saying continued as

“everything is everywhere, but the environment selects”. This theory argued that spatial differentiation

would depend on the capability of each microorganism to succeed in a new environment where it

is transported. An example of clear differences in actual community composition are presented by

Pommier et al. (2007), showing 16S rRNA gene libraries collected at nine distant localities spread

around the world were quite different, and some organisms could not be considered cosmopolitan. Of

course, cosmpolitanism could be present in the “rare species” pool (sensu Pedrós-Alió 2006) and the

Pommier study would have failed to detect that part of the bacterial biodiversity, because it had used

a technique that sampled only the “diversity” (Fig. 2).

A priori, each physical, chemical and biological parameter in the oceans could have an impact on

the diversity of microbial communities which would express in a relationship between the observed

diversity pattern as a function of the contextual environmental parameter (Ramette et al. 2007).

Fig.2 Global sampling of marine bacterioplankton communities. (Pommier et al. 2007)

18

Introduction

Bacterioplankton composition has been shown to change along temporal (seasonal cycles, Pinhassi

& Hagstrom 2000, Morris et al. 2005, Schauer et al. 2003, Eilers et al. 2001) and spatial scales

(Pinhassi & Berman et al. 2003, Selje et al. 2004). Since bacteria respond quickly to biotic and abiotic

changes of their environment, the vertical distribution of physical and chemical parameters in the

marine water column should be the basis for a pronounced strati cation of the prominent microbial

populations (Rheinheimer & Gocke 1994, Ramsig et al.1996. Morris et al. 2004, Giovannoni et

al. 1996, Wright et al. 1997). Diversity (or bacterial community structure) has also been found to

vary following latitudinal gradients (Selje et al. 2004, Pommier et al. 2005). Thus, bacterial species

composition in the ocean appears to be determined by factors known to affect the species in culture,

such as temperature, light, nutrients, and the quantity and quality of DOM. Those are all parameters

that we can consider “bottom-up” forces.

To evaluate the magnitude of that “bottom-up” forcing compared to the effect caused by trophic

interactions (predators, “top-down”), we designed a transplant experiment, that appears in Chapter 5,

using dialysis bags to incubate water from different depths in the original conditions and at different

depths with their associated ecological conditions. We aimed to address the possibility of bacterial

community structure to signi cantly vary depending on the environmental conditions and the presence/

absence of bacterial predators.

EFFECTS OF NUTRIENT ADDITIONS ON BACTERIOPLANKTON COMMUNITY STRUCTURE

Changes in inorganic nutrient availability are of decisive importance for de ning ocean primary

productivity and for regulating phytoplankton community composition, succession and distribution.

In response to the increase or decay of different phytoplankton species, bacterioplankton community

structure changes relatively fast (Pinhassi et al. 2004, Sapp et al. 2007). Since heterotrophic bacteria

are the major consumers of dissolved organic matter in the aquatic environment, limitation of bacteria

by organic or inorganic nutrients can have important consequences in biogeochemical Carbon cycling.

The speci c limiting nutrients for bacterial growth may differ depending on the site. Blanes Bay,

and the whole of the Mediterranean show P-limitation (Thingstad et al. 1998, Pinhassi et al. 2006)

although not always (Sala et al. 2002).

19

Introduction

The function and diversity of the marine microbial communities is especially affected in coastal

areas, which exhibit variability in factors such as salinity, and organic and inorganic nutrients. This

variability is consequence of external forcing factors due to natural causes (river-inputs, rainfalls) or

anthropogenic effects. Human activities are responsible for most of the allochthonous additions, and

industrialization is a component of global change and most of the allochthonous sources of carbon

and inorganic nutrients to the ocean are predicted to vary in response to global change. One of the

consequences of climate change is a decrease in rainfalls that might lead to deserti cation of many

areas and to an increase in the deposition of dust originated in these areas.

Besides these processes, organic contamination may also cause bacterioplankton community

successions (Revilla et al. 2000, Schendel et al. 2004). Coastal eutrophication caused by human

activities provides allochthonous sources of nutrients for bacterioplankton communities. Some of

the most widespread contaminants in the environments are the petroleum hydrocarbons (Santas et

al. 1999) and oil spills are a constant problem in many coastal waters. Bacteria can use the PAHs

(Policyclic Aromatic Hidrocarbons, a component of oil) as an allochthonous source of carbon and

energy (Castle et al. 2006), but these compounds can also negatively affect bacterial composition and

function, as quite a number of studies have shown (Siron et al. 1993, Macnaughton et al. 1999, Kasai

et al. 2001, Ogino et al. 2001, Castle et al. 2006, Ohwada et al. 2003, Maruyama et al. 2003, Capello

et al. 2006 and 2007, Teira et al. 2007)

The aim of Chapters 2, 3 and 4 of this Thesis is to assess the effects caused by different additions on

bacterial function and bacterioplankton community structure. In Chapter 2 we simulated the Prestige

oil spill in waters of the Ría de Vigo kept in mesocosm tanks. Chapters 2 and 3, in contrast, were

performed in the NW Mediterranean. In Chapter 3 we wanted to determine the effect of different

inorganic nutrients on phytoplankton species development and the subsequent effects that they could

cause on bacterioplankton. Chapter 4 evaluated the effect of a Sahara dust deposition event on the

microbial communities of the coastal NW Mediterranean.

20

Introduction

BACTERIAL DIVERSITY DETERMINED BY ENVIRONMENTAL CONDITIONS

Different environmental conditions (temperature, salinity, nutrient concentrations, light…) are

responsible for structuring microbial communities in the ocean. Many past studies have focused on

de ning the selective role of organic and inorganic nutrient availability on the growth of speci c

bacteria (Pinhassi et al. 2004, Allers et al. 2007). In some areas the dominant phylogenetic groups

differ depending on the season. For example, Eilers et al. (2001) found that Bacteroidetes dominate

during spring time and Alphaproteobacteria dominate during summer in the Baltic Sea. However,

in the North Sea, the dominant group during spring was Bacteroidetes while Gammaproteobacteria

took over during summer time. Therefore, bacterial successions present diverse patterns in different

environments.

The Alphaproteobacteria may have special adaptation which allow them to use compounds in

low concentrations and may explain why the SAR11 clade is that abundant in open oceans (Morris

et al. 2002, Malmstrom et al. 2007). Concretely in the NW Mediterranean sea Alphaproteobacteria

was found to account for 40% of total cell counts during summer and spring (Alonso-Sáez et al.

2007). Cytophaga-like bacteria are known to be abundant on particles in estuarines and other marine

environments (DeLong et al. 1993, Crump et al. 1999). Indeed, Pinhassi et al. (2004) found that

Flavobacteriaceae could be particularly important in the processing of organic matter during algal

blooms. Gammaproteobacteria have been shown to have the capacity for fast growth in response to

elevated nutrient concentrations (Pinhassi & Berman 2003).

Changes in bacterial composition have also been reported from the Mediterranean Sea (Schauer et

al. 2003, Alonso-Sáez et al. 2007) where different studies have shown shifts in bacterial community

structure due to inorganic nutrient additions (Schäfer et al. 2001, Lebaron et al. 1999, Pinhassi et al.

2006, Allers et al. 2007). Microbial communities from ecosystems receiving nutrient additions tend

to be dominated by those organisms capable of competitively dominate the uptake of the nutrients

added. Microbial communities from contaminated ecosystems also tend to be dominated by those

organisms capable of utilizing the toxic compounds. As a result, these communities are less diverse

in comparison to pristine environments (Torsvik et al. 2002). Examples of this specialization of the

marine communities due to the addition of oil spills are reported by Macnaughton et al. (1999), Ogino

et al. (2001), Kasai et al. (2002), Teira et al. (2007). After an exposition to petroleum contamination

there is a strong selection for hydrocarbon degrading bacteria, such as genus Alcanivorax and

Cycloclasticus.

21

Introduction

PREDATION PRESSURE ON BACTERIAL ACTIVITY AND COMMUNITY STRUCTURE

Depending on the type of marine system, the main factors responsible for the regulation of bacterial

activity and growth can be temperature (Shiah & Ducklow 1994), grazing (Thingstad et al. 1997) or

competition for limiting nutrients (Rivkin & Anderson 1997, Thingstad et al. 1997). Indeed, grazing,

especially due to heterotrophic nano agellates (HNF), has been identi ed as the main factor leading

to bacterial loss (Fenchel et al. 1982, Sherr & Sher 1984, Wright & Cof n 1984, Pace 1988), and

can often balance bacterial production. In oligotrophic environments the substrate supply is more

relevant than predators (Sanders et al. 1992). Several studies indicate that agellates tend to graze

preferentially upon the active component of the bacterioplankton assemblage (del Giorgio et al. 1996),

thus selectively exerting grazing pressure upon some speci c groups of bacteria, a phenomenon that

could modulate assemblage composition (Pernthaler 2005, Pernthaler & Amann 2005).

Although top-down and bottom-up controls are widely known concepts in aquatic microbial ecology

(Pace & Cole 1994), there is still disagreement to whether the abundance and productivity of bacteria

are determined mainly by predators or nutrients (Pernthaler et al. 1997). Furthermore, from a structural

point of view, both predators and nutrients may be major forces shaping the genotypic composition

of bacterial communities (Jürgens & Güde 1994). In addition, succession patterns might appear as

shaped by selective mortality, since heterotrophic nano agelates or viral lysis can be phylotype-

selective and affect the relative proportions of different microbial populations (Thingstad et al. 2000,

Pernthaler 2005). Allers et al. (2007) reported selective agellate grazing on the bacterial population

in response to glucose and phosphorus manipulations in a mesocosm. Grazing resulted in a change in

bacterial community composition, varing from Alteromonadaceae dominance to Rhodobacteriaceae

dominance. Viruses are also important in the control of bacterial abundance (Bratbak et al. 1990)

In all the experiments performed in this thesis we wanted to address the question of which were the

bottom-up and top-down factors controlling the bacterial communities. In order to do so we took into

consideration the grazing effect, either by measuring the HNF abundance or by controlling predator

presence (by ltration).

22

Introduction

EXPERIMENTAL APPROACHES

Biological oceanographers try to understand how marine organisms and biological processes

affect and are affected by physical, chemical and geochemical ocean processes. The relative marine

microbial activities and their in uences in global oceanic processes are dif cult to assess. There

are two approaches to address these questions. The rst, and most realistic one, is to perform in

situ sampling of different environments. Negative aspects of this approach include that it is time

consuming, that it requires economical effort, and that depends on favourable natural conditions.

Because of these problems with carrying out in situ studies, laboratory experiments are designed to

determine the factors affecting metabolism and diversity of bacteria in the environment. However,

this second approach is handicapped by the dif culty of simulating natural conditions. As reported by

Pinhassi & Berman (2003), growth of bacteria from particular phylogenetic lineages are commonly

stimulated by the input of certain labile substrates which might be released during experimental

manipulations.

Hence, it is important to restrict unnecessary manipulations. To avoid these laboratory limitations

mesocosms experiments are commonly used in aquatic ecology to simulate natural behaviour of

communities. These are easy to manipulate and minimize disturbances in the experimental setup by

increasing the incubation volume (Odum et al. 1984). Many studies have performed these approaches

in order to evaluate the response of marine microorganisms to nutrient additions concretely in the

Mediterranean sea: Schäfer et al. (2001), Pinhassi et al. (2006) and Allers et al. (2007). In Chapter

3 and 4 we also used mesocosms experiments to determine the effect of different resources in Blanes

Bay (NW Mediterranean) communities. However, this kind of experiments can also mimic other

situations in the environment such as the release of contaminants to the environment e.g. oil-spills

(Siron et al. 1993, Ohwada et al. 2003, Kasai et al. 2002, Teira et al. 2007). In Chapter 2 we used a

mesocosm experiment with a similar set-up as reported by Olsen et al. 2006). It allows generalization

of some conclusions to situations other than the one assayed, but also allow to see differences between

experiments. A more realistic approach to study the in uences of the environmental conditions to a

speci c bacterial community is the used of the dyalisis bags, a kind of “cage “ that allows the ux

of nutrients. In Chapter 5 we combine mesocosm with transplants using dyalisis bags to assess the

bottom-up or top-down factors controlling the microbial communities at different depths.

23

Introduction

RESOURCES

Fig. 3 Conceptual outline of the thesis. Bacterial community composition and functioning are controlled by two different groups of factors (top-down and bottom-up). The aim of the thesis is to assess the effects of substrates from different origins. In Chapter 1, by determining the bacteria thast grow on traditional culturing, media. In Chapter 2, by evaluating inputs of anthropogenic origin (oil-spills). In Chapter 3, by atmospheric depositions (Saharan dust). In Chapter 4, by stimulation of phytoplankton blooms and the subsequent effect on bacteria, and in Chapter 5, by the effect of mixing of waters of different origin along the vertical strati cationin the ocean.

Top-down

Bottom-up

24

Introduction

IN SUMMARY, AIMS OF THIS THESIS:

The main goal of this thesis was to determine how the different allochthonous resources

affected bacterial community structure, abundance and activity. In order to assess this question

we used different approaches, such as traditional culture techniques, mesocosm experiments

and transplants: which include a combination of mesocosms and dialysis bags.

Schauer et al. (2003) and Alonso-Sáez et al. (2007) assessed the seasonality of bacterioplankton

diversity by clone libraries and DGGE (a ngerprinting technique) in Blanes Bay. As a

consequence of the increasing interest for isolation of microorganisms, we put some effort at

isolating in plates bacteria from this environment. Chapter 1 was focused on the comparison

of the sequences retrieved by culturing with the sequences obtained by culture independent

techniques previously reported in other studies.

Mesocosms experiments was an approach used in this thesis to determine the effect of different

allochthonous additions, of natural or anthropogenic origin. In Chapter 2 we described the

simulation of the Prestige oil spill in front of the Galician coast, in Chapter 3 we analyze bacterial

function and community structure after different phytoplankton bloom stimulation experiments

using different nutrient sources in the NW Mediterranean sea. In Chapter 4 we mimicked a

Saharan dust deposition event in the NW Mediterranean and describe their consequences on the

bacterial community structure and on microbial community function, including the metabolic

balance.

In Chapter 5 we considered “allochthonous” the water from depths different from those where

the resident bacteria developed. We used a transplant experiment to determine the controlling

factors of bacterioplankton communities of the epipelagic, the DCM and the mesopelagic in the

North Atlantic Ocean.

Increased marine bacterioplankton diversity detected by combining culturing and culture-independent techniques

Itziar Lekunberri, Josep M Gasol, Silvia G. Acinas, Laura Gómez-Consarnau, Emilio O. Casamayor, Ramon Massana, Carlos Pedrós-Alió and Jarone Pinhassi

FEMS Microbial Ecology, submitted

CHAPTER 1

Per descriure la complexitat d´un arbre és necessari un poeta

Ramon Margalef

27

Chapter I

ABSTRACT

Despite the recognition that microbial biodiversity is an important characteristic of marine

ecosystems, knowledge about the diversity of marine bacteria that can be cultured on traditional

media is very limited. We studied the phylogenetic diversity of 136 bacterial isolates obtained from

coastal Mediterranean seawater at the Blanes Bay Microbial Observatory (www.icm.csic/bio/projects/

icmicrobis/bbmo). Nearly all isolates belonged to the Gammaproteobacteria, Alphaproteobacteria, or

the Bacteroidetes (70, 40 and 20 isolates respectively). The 16S rRNA gene sequences of the isolates

were compared to environmental sequences previously obtained by molecular methods, including

DGGE (60 sequences) and clone libraries (303 sequences), from samples collected at different seasons.

Nine cases of complete or nearly complete identity between isolates and in situ sequences were found:

ve in the Gammaproteobacteria, three in the Alphaproteobacteria and one in the Bacteroidetes. This

suggests that culturing to some extent provides representative organisms for physiological response

experiments for various taxa detected by molecular methods. The limited overlap between sequences

obtained by culturing and molecular techniques resulted in a substantial increase in the bacterial

diversity (richness) detected at this sampling site by combining the two approaches. Rarefaction curve

analysis showed similar patterns for sequences from isolates and molecular approaches, except for

Alphaproteobacteria where culturing retrieved a higher diversity per unit effort. Around 30% of the

environmental clones and isolates formed microdiversity clusters constrained at 99% 16S rRNA gene

sequence similarity. The nding of microdiversity clusters among isolate and clone sequences alike,

suggests that this type of microdiversity is not an artifact and that similar evolutionary processes act on

cultured and uncultured bacteria. Thus, marine bacterial isolates could serve as models to investigate

how microdiversity clusters originate and provide indications of the ecological signi cance of such

genetic variability.

28

Isolate bacterioplankton diversity in the NW Mediterranean

INTRODUCTION

The application of molecular methods to retrieve small subunit rRNA sequences from natural

environments revolutionized microbial ecology (Pace et al. 1985). About 20 years later, environmental

sequences are increasingly reported to data bases and have signi cantly expanded the known diversity

of bacteria (Giovannoni et al. 1989), archaea (DeLong 1994, Fuhrman et al. 1994) and eukarya (Díez

et al. 2001, López-García et al. 2001, Moon van der Stay et al. 2001). For most of this period, culturing

has been looked upon almost with suspicion, as a technique that provided a biased picture of the

“true” diversity of natural communities. In fact, the number of studies of cultured diversity has been

extremely low compared to those reporting clone libraries or analyses using different ngerprinting

techniques (Donachie et al. 2007).

Ribosomal RNA sequences, however, are a poor substitute for the organisms themselves. The

cultured organisms can be used as model organisms for physiological studies (Giovannoni &Stingl

2007, Gómez-Consarnau et al. 2007) and their genomes can be sequenced (Moran et al. 2004, Bauer

et al. 2006, González et al. 2008) and used as scaffolds to reassemble genomes from metagenomic

analyses (Venter et al. 2004, Rusch et al. 2007). Recently, novel protocols for isolating bacteria have

been developed (Zengler et al. 2002, Simu et al. 2005, Giovanonni et al. 2007, Stingl et al. 2007,

Nichols et al. 2008) generating new interest in culturing. However, even with the traditional culturing

methods, new microorganisms can be easily isolated from any environment (Donachie et al. 2007).

Obviously, these microorganisms are present in concentrations that are too low to be retrieved by

molecular methods that essentially detect only the most abundant sequences (Pedrós-Alió 2007).

These easily cultured organisms may form part of the rare biosphere (Pedrós-Alió 2007): the very

large collection of microorganisms present in ecosystems at very low abundance (Sogin et al. 2006).

This low abundance protects them from viral lysis and, partially, from predation, and they can persist

for very long periods of time with little or no activity. It has been suggested that if conditions change,

members of the rare biosphere can quickly multiply and become part of the abundant taxa and

participate actively in carbon and energy ow (Pedrós-Alió 2006). Thus, they constitute a “seed-bank”

that may become extremely useful for the community to respond to changing conditions, including

the appearance of pollutants or other human-made perturbations. The biodiversity of any ecosystem

is not completely known until this rare biosphere is characterized (Pedrós-Alió 2006, Donachie et al.

2007). Since molecular methods are generally incapable of retrieving these rare organisms, isolation

in pure culture remains one of the best windows into the rare biosphere.

29

Chapter I

Comparisons between molecular and culturing approaches have been attempted in a very limited

extent, comparing only one or a few samples and a low number of clones and/or cultures (Suzuki et al.

1997, Eilers et al. 2000, 2001, Hagström et al. 2002, Kisand & Wikner 2003, Stevens et al. 2005). Here,

we analyze a collection of bacteria isolated in pure culture from the Blanes Bay Microbial Observatory

(BBMO). This is an oligotrophic coastal site in the Northwestern Mediterranean Sea, where extensive

molecular data exist (Schauer et al. 2003, Alonso-Sáez et al. 2007) allowing for comparisons between

both approaches. Thus, the aim of the present study was to compare the diversity of cultured isolates

to the diversity of environmental sequences obtained by cultivation independent techniques. Further,

we aimed at comparing the patterns of microdiversity among cultures and clones. In addition, nine of

the isolates were selected by the Gordon and Betty Moore Foundation Marine Microbiology initiative

(www.moore.org/marine-micro.aspx) for whole genome sequencing. The present study intends to

provide the molecular environmental context for these bacteria.

MATERIALS & METHODS

Study area and sampling. The site studied was Blanes Bay, in the NW Mediterranean (The Blanes

Bay Microbial Observatory), at 70 km North of Barcelona. Water was sampled at about 1 km offshore

(41º40’N, 2º 48’E) from a boat, and immediately ltered through a 200 µm mesh net. Sampling was

done in parallel to studies of carbon and sulfur cycling (see details in Alonso-Sáez et al. 2008 and

Vila-Costa et al. 2008) through the period 1998-2004 (see details in Table 4). Seawater was kept in

25-l polycarbonate carboys and transported under dim light to the laboratory (less than 2 h).

Origin of bacterial DNA and isolates.

1) Natural seawater. Samples were collected as explained above and inoculated directly on agar

plates as detailed below.

2) Whole water enrichments. The effect of nutrient additions on the growth of heterotrophic

bacteria was examined in un ltered, whole-water samples (Pinhassi et al. 2006a). Brie y, nutrients

were added to 250 ml samples in nal concentrations of 40 µM C (as glucose), 2 µM N (NH4Cl),

and 0.6 µM P (NaH2PO

4), singly and in all different combinations. After incubation for 24 h at in situ

temperature in the dark, samples for bacterial production, total bacterial numbers and colony forming

units were collected.

30

Isolate bacterioplankton diversity in the NW Mediterranean

3) Dilution cultures. Several times during the period 2001-2004, 1.9 l of seawater were used as

growth media after ltration through 0.2 µm pore-size sterivex units (Durapore-Millipore) using a

peristaltic pump at pressures <200 mm Hg (for details see Pinhassi et al. 2006a). Inocula consisted in

100 ml seawater prepared by gravity ltration through 0.8 µm pore-size lters (Nuclepore). Bacteria

developed in 2 L Nalgene bottles maintained at in situ temperatures (15-22ºC) and in the dark.

Cultures were unamended controls or enriched with C and P, alone or in combination. Substrates

added were e.g. glucose, dimethylsulfoniopropionate (DMSP), inorganic P, ATP, or DNA, all at a nal

concentrations of 20 µM C (Glucose, DMSP) or 0.6 µM P (phosphate, ATP, or DNA).

4) Mesocosm experiments. Two mesocosm (20-100 L) experiments reported in detail in Pinhassi

et al. (2004) and Allers et al. (2007) provided additional sequences. For these experiments surface

seawater was collected from Blanes Bay and was transferred to the laboratory. Bacterial growth and

diversity were monitored in mesocosms enriched with N and P and/or C for approximately one week

(see details in the references mentioned above).

Collection of community DNA. To collect in situ microbial biomass, between 5-15 L of seawater

were ltered using a peristaltic pump through a 5 µm pore size Durapore lter (Milipore) and a 0.2 µm

Sterivex lter (Durapore, Milipore) in succession. The 0.2 µm Sterivex unit was lled with 1.8 ml of

lysis buffer (40 mM EDTA, 50 mM Tris-HCl, 0.75 M sucrose) and stored at –70ºC. Microbial biomass

from the experiments was collected by ltering 0.5-1 L water onto 0.2 µm-pore-size polycarbonate

lters (Milipore) at <200 mm Hg, and lters were stored in lysis buffer at –70ºC. Nucleic acids were

extracted by a standard protocol using phenol/chloroform (details in Schauer et al. 2003)

Origin of the sequences used in the phylogenetic comparison.

1) Isolates. A total of 625 colonies were collected from the above-mentioned sources, from which we

selected 136 isolates for partial 16S rDNA sequencing based on differences in colony morphology and

abundance on agar plates at different sampling occasions or experiments. The isolates were obtained

by plating 100 µl of undiluted, or 10x, 100x and 1000x diluted seawater of untreated or experimental

samples, in triplicates, onto Zobell agar plates (Zobell 1943). Agar plates were incubated at in situ

temperature in the dark for 10-15 days. The 16S rDNA sequences of the 136 bacterial isolates were

ampli ed by means of PCR using Taq polymerase (Boehringer-Mannheim) from DNA preparations

using bacterial 16S rDNA primers 27f and 1492r. Isolates were partially sequenced with internal

primer 358f, which yielded sequences of approximately 700 - 850 bp length. Accession numbers of

the isolates are given in the Table 4.

31

Chapter I

2) DGGE bands. DGGE bands from both environmental samples (27) and enrichment cultures (33)

(whole seawater enrichments and dilution cultures) were retrieved. DGGE ngerprints were performed

as previously described (Schauer et al. 2003, Sánchez et al. 2007). The 16S fragments (around 550

bp in length) were ampli ed with bacterial speci c primer 358f complementary to position 341 to

358, with GC-clamp and universal primer 907rm complimentary to position 927 to 907 (Muyzer et al.

1997). The PCR products were loaded in a 6% polyacrylamide gel with a DNA-denaturant gradient

ranging from 40 to 80%. The gel was run at 100 V for 16 h at 60°C in 1x TAE running buffer. The

DGGE bands were excised, reampli ed, and veri ed by a second DGGE. Bands were sequenced

using primer 358f without the GC-clamp; with the Big Dye terminator cycle-sequencing kit (Perkin

Elmer Corporation) and an ABI PRISM model 377 (v3.3) automated sequencer. These sequences

have been reported in Schauer et al. (2003), Alonso-Sáez et al. (2007), Pinhassi et al. (2004, 2006a),

and Allers et al. (2007).

3) Clone libraries. Clone libraries were constructed from in situ samples collected on 3 March,

14 May, 14 July, 4 August and 21 October 2003 (Alonso-Sáez et al. 2007). Brie y, 16S rDNA

sequences were ampli ed from total community DNA by PCR with universal primers 27f and 1492r,

and resulting amplicons were cloned using the TOPO-TA cloning kit. PCR amplicons were digested

with the restriction enzyme HaeIII (Invitrogen), and the RFLP patterns were compared to identify

unique clones. Clones were partially sequenced with internal primer 358f, which yielded sequences

of approximately 700 - 850 bp.

4) Additional sequences included in the analysis. For comparison, eight DGGE sequences from

the neighboring Banyuls Bay (ca. 150 km away) were included (Schäfer et al. 2001).

Phylogenetic analysis. The 16S rRNA gene partial sequences of our isolates were compared to

existing prokaryotic sequences in GenBank (NCBI) using BLAST (basic local alignment search

tool). Comparison of BLAST results from database searches with different regions of the sequences

showed that none of our sequences were chimeras. Phylogenetic analyses were performed using the

parsimony methods implemented in the ARB sequence analysis package by which sequences were

grouped into major phylogenetic groups. The trees were calculated from sequences approximately

corresponding to nucleotide positions 420 to 770 (E. coli numbering). Sequences of close relatives

included in GenBank after the BLAST process were included in the trees as references in order to

delimitate and name different clusters of sequences.

32

Isolate bacterioplankton diversity in the NW Mediterranean

Comparison of sequence diversity. Rarefaction analysis and diversity estimates by the non-

parametric Chao-1 richness estimator were carried out using the online rarefaction calculator

software (http:// www2.biology.ualberta.ca/jbrzusto/rarefact.php). For clustering sequence analysis,

the software Clusterer was used (Klepac-Ceraj et al. 2006). This allows grouping of sequences into

percent similarity clusters (100, 99% etc.) by the nearest neighbor approach (i.e., sequences are added

to clusters if there is at least one sequence which is within a set similarity threshold). These clusters

based on 99% similarity cut-off served as the basic units for statistical extrapolation of total sequence

diversity using the Chao-1 non-parametric richness estimator and for comparison of sequence

composition among the major phylogenetic groups and between molecular and isolation approaches.

RESULTS

General patterns of richness. The 136 bacterial cultures (84 unique) isolated from Blanes Bay

and analyzed here are detailed in the Table 4, where their origin, closest sequence, and closest cultured

relative are presented. The 16S rRNA sequences of these isolates were compared with 363 sequences

(194 unique) obtained from 16S rDNA clone libraries and denaturing gradient gel electrophoresis

(DGGE) from the same site, retrieved in previous studies (Schauer et al. 2003, Alonso-Sáez et al.

2007). Fig. 1 shows the main groups of bacteria retrieved from the BBMO indicating whether they

were found by culturing, molecular methods or both.

We explored how the percent similarity level chosen would affect the number of OTUs for isolates

and clones (Fig. 2A). When the % similarity was increased from 99 to 100% the number of clone

OTUs increased substantially while the number of isolate OTUs remained nearly the same. This

increase would appear to indicate a larger microdiversity among the clones compared to the cultures.

However, when the same exercise was carried out within each of the three main groups of bacteria,

signi cant differences appeared. In the Alphaproteobacteria there was a large increase in OTUs

among the clones, while only minor increases in OTUs among isolates was observed (Fig. 2B). The

Gammaproteobacteria showed precisely the opposite trend, with a marked increase in the microdiversity

of the cultures and a very small increase in that of clones (Fig. 2C). The Bacteroidetes showed almost

no changes either in isolates or in clones (Fig. 2D). Other than this, there was a clear discontinuity

above 97%, suggesting that this percentage is a reasonable cutoff point for this data set. Both the

Alpha- and Gammaproteobacteria showed the in ection point around 97% similarity. Interestingly,

33

Chapter I

the Bacteroidetes showed a much more uniform increase of OTUs with increases in percent similarity.

We also combined all sequences (isolates, clones, and DGGE bands) for each of the three major

groups (Fig. 2E), showing the clear difference in microdiversity between the Proteobacteria, on the

one hand, and the Bacteroidetes on the other. The next analysis consisted of calculating rarefaction

curves for all the sequences together (Fig. 3A) or for the three groups separating isolates and clones

Fig. 1.- Schematic phylogenetic tree displaying the major lineages in the marine prokaryotic plankton. The color of each branch indicates whether only clones (green), isolates (red) or both, clones and isolates (yellow), were retrieved from Blanes Bay. Lack of color indicates that we could not detect this group with any of these techniques. Cyanobacteria and algae are drawn in dark green since they appeared in the clone libraries in this study but no effort was made to isolate them.

Bacteroidetes

Archaea

Planctomyces

Alteromonas

Marinomonas

Halomonas

Epsilonproteobacteria

Pelagibacter

Roseobacter

Sphingomonas

Rhodobacter

Betaproteobacteria

Chloroflexi

Fibrobacter

Sphingobacteria

BacteroidetesFlavobacteria

Cyanobacteria Algae

Marine Actinobacteria

SAR86

Deltaproteobacteria

Gammaproteobacteria

Alphaproteobacteria

Verrumicrobia

Firmicutes

34

Isolate bacterioplankton diversity in the NW Mediterranean

Fig. 2.- Number of OTUs as a function of the percentage of similarity cutoff chosen to group sequences. A) The three major phylogenetic groups of bacteria combined as retrieved by culturing, cloning or both together B) Number of Alphaproteobacteria OTUs obtained with each technique. C) Same for Gammaproteobacteria; D) Same for Bacteroidetes. E) Total number of OTUs obtained using clone libraries and isolation techniques combined for each of the three main groups.

0

10

20

30

40

50

60

80 85 90 100

Gammaproteobacteria

0

5

10

15

20

25

30

35

40

80 85 90 95 100

Bacteroidetes

Nu

mb

er o

f OTU

sN

um

ber

of O

TUs

Percentage of similarity (%)

(A) (B)

(C) (D)

Percentage of similarity (%)

Nu

mb

er o

f OTU

s

Percentage of similarity (%)

All Alphaproteobacteria

0

10

20

30

40

50

60

70

80

80 85 90 95 100

Group comparison AlphaGammaCFB

95

0

10

20

30

40

50

60

70

80

80 85 90 95 100

totClonesIsolates

(E)

0

20

40

60

80

100

120

140

160

70 75 80 85 90 95 100

35

Chapter I

Table 1.- Number of characterized isolates, clones, and DGGE bands belonging

to different taxonomic groups retrieved from Blanes Bay.

__________________________________________________________________________

Isolates Clones DGGE bands

Total Unique Total Unique In situ Experiments__________________________________________________________________________

Alphaproteobacteria 40 20 219* 80 16 17

Betaproteobacteria 2 2 1 1 0 0

Gammaproteobacteria 70 39 50 26 4 2

Deltaproteobacteria 0 0 0 0 1 0

Epsilonproteobacteria 0 0 1 1 0 0

Bacteroidetes 20 19 18 14 7 13

Firmicutes 3 3 0 0 0 0

Actinobacteria 1 1 4 3 0 0

Verrucomicrobia 0 0 10 9 0 0

All 136 84 303 134 28 32__________________________________________________________________________

* 95 belonged to the SAR11 group.

(Figs. 3B-D). The pooled data set was used to estimate the total number of OTUs present in Blanes

Bay by using the Chao 1 index. When the percent similarity was increased from 97 to 100%, the total

richness increased from 125 to 420 OTUs.

Next, we chose a 99% similarity level to compare the estimates of richness provided by isolates,

clones or the two combined, for the three main groups (Table 2, Figs. 3B-D). Unexpectedly, the

Chao estimates from isolates were similar to (Alphaproteobacteria and Bacteroidetes), or larger than

(Gammaproteobacteria), the estimates from clones alone (Table 2). Thus, there was at least as much

richness in the isolate collection as in the clone libraries. It must be recalled that all cultures were

isolated in Zobell medium under similar conditions, with no particular efforts to provide a variety of

culture conditions. Combination of isolates and clones provided Chao estimates similar to, or larger

than, those obtained with either isolates or clones separately (Table 2). Thus, the total richness estimate

obtained by combining the molecular and the isolation approaches was similar to that obtained for the

isolates alone. The rarefaction curves con rmed the differences among the three groups: more OTUs

from cultures for the Gammaproteobacteria, and similar estimates for the Alphaproteobacteria and

the Bacteroidetes (Fig. 3B-D).

36

Isolate bacterioplankton diversity in the NW Mediterranean

Fig. 3.- Rarefaction analysis for sequences obtained by isolation and by molecular approaches. A) All sequences together. B) Alphaproteobacteria. C) Gammaproteobacteria. D) Bacteroidetes.

0

50

100

150

200

0 50 100 150 200 250 300 350 400

100% 166 420.3 49.1 99% 116 231.8 28.3 98% 94 159 19.2 97% 74 125.6 17.6

Number of sequences screened

Num

ber o

f OTU

s obs

erve

d

OTUs Chao1 SD

0

5

10

15

20

25

0 10 20 30 40 50 60 70 80

Gammaproteobacteria

isolatesclones

0

5

10

15

20

0 5 10 15 20 25

Bacteroidetesisolatesclones

Number of sequences screened Number of sequences screened

Number of sequences screened

All sequences

Num

ber o

f OTU

s obs

erve

d Nu

mbe

r of O

TUs o

bser

ved

Num

ber o

f OTU

s obs

erve

d

0

5

10

15

20

25

30

35

0 50 100 150 200

Alphaproteobacteria

isolatesclones

(A)

(B)

(C)

(D)

37

Chapter I

Tabl

e 2.

-Cha

o1 r

ichn

ess

estim

ates

for

the

maj

or b

acte

rial

gro

ups,

bas

ed o

n 99

% s

imila

rity

cut

-off

usin

g th

e se

quen

ces

from

eith

er

isol

ates

and

clo

ne li

brar

ies

sepa

rate

ly, o

r al

l com

bine

d.

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

Is

olat

es

Clo

nes

Isol

ates

and

clo

nes

Taxo

n A

ll O

TU

s C

hao

SD

All

OT

Us

Cha

o SD

A

ll O

TU

s C

hao

SD

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

Alp

hapr

oteo

bact

eria

40

19

47

16

18

3 33

58

12

22

3 52

14

0 31

Gam

map

rote

obac

teri

a 70

27

10

8 39

46

16

36

14

11

6 39

92

23

Bac

tero

idet

es

20

19

34

11

19

16

44

16

39

32

77

19

Tota

l 13

0 67

24

6 52

24

8 62

12

5 21

37

8 11

6 23

2 28

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

____

38

Isolate bacterioplankton diversity in the NW Mediterranean

Identity of the cultures.

Alphaproteobacteria. Phylogenetic analysis of the Alphaproteobacteria revealed that a substantial

fraction of the isolates belonged to different clusters within the Roseobacter group (clusters Alpha

1 to 4, Fig. 4). In addition, we found isolates belonging to Erythrobacteraceae (cluster Alpha 5),

Hyphomonadaceae and Aurantimonadaceae.

Clusters Alpha 1 and 4 included only sequences obtained by culture independent techniques (Fig.

4, Table 3). Sequences within cluster Alpha 1 (related to clone NAC11-7) and within cluster Alpha 4

(related to clone CHAB-1-5) showed within group similarities >95% and >96% respectively (Table

3). On the other hand, cluster Alpha 2 included only sequences from isolates and P-enrichment

experiments (Mes9 and B-260) but none retrieved directly from the in situ samples. Isolate M193

within this cluster has been whole-genome sequenced. In addition, six isolates represented bacterial

groups that were not detected by molecular techniques (M588, M464, M441, M624, M623 and

M463).

Two clusters containing both isolates and environmental sequences were found. In cluster Alpha 3

the sequences showed within group similarities larger than 97% and belonged to the genus Nereida.

The isolate M128 was 99.8% similar to the clone sequence SPR23 and 99.6% to DGGE bands

B90 and B94, obtained in P, ATP and DNA addition experiments. Cluster Alpha 5 consisted of the

genus Erythrobacter, with four isolates, two clone library sequences, and one in situ DGGE band.

Erythrobacter isolate M442 was 99.8% similar to the clone sequence AUT4 and was identical to

DGGE band 27. Apart from these clusters, isolate M483 was 99.8% similar to clone SPR41. Thus,

isolates M128, M442, and M483 seem good candidates for investigating pure cultures representative

of marine bacterioplankton.

All together, 64% of the 16S rRNA sequences grouped into >97% similarity clusters. The

Alphaproteobacteria had substantial microdiversity with only 33% of the OTUs grouped at 99%

similarity level (Fig. 2B). The rarefaction curves for the isolates were steeper than for the environmental

clones (Fig 3B), implying that, with the same effort, there is a potential to recover larger species

diversity through isolation than through clone analysis.

39

Chapter I

Fig. 4.- Maximum Parsimony tree illustrating the relationships among the partial 16S rRNA sequences from Blanes Bay for the Alphaproteobacteria. Colors indicate the origin of the sequences: bacteria isolated on solid media (red), DGGE bands from experiments (orange), DGGE bands from in situ samples (green), and clones retrieved from in situ samples (blue). Clusters are indicated by vertical red lines. Identities between isolates and sequences are marked by white arrows. The sequences marked with asterisks have been genome sequenced. The SAR11 sequences are not displayed in this tree.

SUM32, Clone

M128, Isolate M149, Isolate

Alpha 1

Alpha 2

Alpha 3

Alpha 4

Alpha 5

0.10

NAC11-7, AF245635 SPR9, Clone WIN4, Clone BL98-28, DGGE band in situ BL03-35, DGGE band in situ

Roseobacter sp. DQ660384.1 B115, DGGE band experiments B140, DGGE band experiments

Roseovarius crassostreae, AF114484Mes23, DGGE band Mesocosmos

Roseoisalinus antarticus, AJ605747.2 SPR10, Clone

BL98-31, DGGE band in situ Mes38, DGGE band Mesocosmos

M612, Isolate Mes11, DGGE band Mesocosmos

Aquamarinicola italica, AM904563M605, Isolate

Mes26, DGGE band Mesocosmos Mes47, DGGE band Mesocosmos Ruegeria gelatinovorans, AF247972M606, Isolate

Mes28, DGGE band Mesocosmos Silicibacter pomeroyi, CF000031

B-230, DGGE band experiments BL98-33, DGGE band in situ

M193, Isolate * Mes9, DGGE band Mesocosmos B-260, DGGE band experiments

M165, Isolate M618, Isolate

M483, Isolate SPR41, Clone Roseobacter litoralis, X78312

BL03-37, DGGE band in situ B147, DGGE band experiments

SPR2, Clone Octadecabacter orientus, DQ167247

Jannashia cystaugens, AB121783M588, Isolate Nereida ignava, AJ748748

SPR23, Clone B90, DGGE band experiments B94, DGGE band experiments

M613, Isolate Mes8, DGGE band Mesocosmos

Rhodobacteraceae bacterium, AY962292.1 Mes16, DGGE band Mesocosmos

AUT20, Clone CHAB-1-5,AJ240910

AY-57, DGGE band Banyuls B79, DGGE band experiments BL98-34, DGGE band in situ BL03-43, DGGE band in situ

BL98-36, DGGE band in situ AY-51, DGGE band Banyuls

Mes15, DGGE band Mesocosmos Loktanella hongkongensis,AY600300.1

M464, Isolate Paracoccus sp.DQ108402.1

M441, Isolate AY-58, DGGE band Banyuls

BL03-55, DGGE band in situ Uncultured organism, DQ396058.1

M623, Isolate Hyphomonas oceanitis, F261046Maricaulis maris, AJ227803.1

M624, Isolate Aurantimonas coralicida, AY065627

M463, Isolate Unculture alphaproteobacteria, AF245619.1

AUT80, Clone Citromicrobium sp., DQ985045.1

M458, Isolate M443, Isolate

Erythrobacter citreus, AF118020 AUT4, Clone

M442, Isolate AUT12, Clone

BL03-48, DGGE in situ Erythrobacter aquimaris, AY461443

M155, Isolate Nisaea denitrificans, DQ665839.1

BL98-50, DGGE band in situ BL03-51, DGGE band in situ

Candidatus Pelagibacte ubique, CP000084.1

40

Isolate bacterioplankton diversity in the NW Mediterranean

Gammaproteobacteria. The Gammaproteobacteria was the most diverse group in terms of richness

(Table 1, Fig. 5) and could be divided into 6 major clusters. In cluster Gamma 1 (Alteromonas) there

were several coincidences of isolates with sequences from in situ clone libraries. This cluster consisted

of 9 isolates from different experiments with combined enrichments of inorganic P and C sources like

glycerol, glucose and DMSP. In this cluster we found that three isolates (M111, M185 and M292) were

identical or nearly identical (>99%) to DGGE bands (labeled as AY and BY) obtained in a mesocosm

experiment carried out in 1996 in Banyuls, ca. 150 km northeast of Blanes Bay (Schäffer et al. 2001).

Furthermore, isolate M169 was identical to DGGE band B145 from an experiment (isolate M604 was

99.8% similar to this sequence). All the sequences from this Alteromonas group were distantly related

to sequence JUL4, which represented 30% of the clones from the July clone library).

Three clusters were formed only by isolates: Gamma 2 (Pseudoalteromonas), Gamma 3 (Vibrio),

and Gamma 4 (Marinomonas). Among the Vibrio sequences there was a substantial phylogenetic

diversity (within-cluster similarity 94%, Table 3). The Pseudoalteromonas cluster included isolates

obtained from unenriched seawater samples from different seasons (March, June and November). In

contrast, most Vibrio sequences were isolated from C+P enrichment experiments. Sequences related

to the Marinomonas genus were primarily obtained from enrichments with ATP

Isolate MED266, from unenriched seawater cultures, was 99.3% similar to clone JUL18 that

was found in three copies in the July clone library. Isolates af liated with the genus Marinobacter

were retrieved from enrichments with three different carbon sources (glucose, pyruvate or DMSP) in

combination with phosphate. Cluster Gamma 5 consisted of distantly related sequences containing

only clone sequences.

The Gammaproteobacteria, like the Alphaproteobacteria, showed high levels of microdiversity:

only 32% of the OTUs were grouped at 99% similarity level. In contrast to the Alphaproteobacteria,

this was due to the large microdiversity of isolates but not of clones (Fig. 2C). Moreover, based on

the Chao-1 estimator, Gammaprotebacteria had the largest number of OTUs (108). This number of

taxa was slightly higher when only isolates were taken into account than with the combination of both

molecular and isolation approaches (Table 2).

Six genera were detected by isolation that had not been detected by molecular techniques: Pseudoalteromanoas, Vibrio, Halomonas, Marinomonas, Oleispira and Marinobacter. Isolates M222,

M121 and M297 were whole-genome sequenced. MED297 has been formally described as the new

species Reinekea blandensis (Pinhassi et al. 2007).

41

Chapter I

Fig. 5.- As previous gure, for the Gammaproteobacteria.

BY-92, DGGE Banyuls Gamma 1

Gamma 2

Gamma 3

Gamma 4

Gamma 5

0.10

M113, Isolate SUM81, Clone

M609, Isolate Alteromonas sp., EF061412.1

M275, Isolate Alteromonas alvinellae, AF288360

B145, DGGE experiments M169, Isolate M604, Isolate

M111, Isolate SUM78 , Clone

AY-86, DGGE Banyuls M185, Isolate

AY-84, DGGE Banyuls M292, Isolate

Alteromonas macleodii, Y18231 AY-85, DGGE Banyuls

Alteromonas addita, AY682202 M517, Isolate

B-227, DGGE experiments JUL4_14, Clone

Glaciecola mesoplila, AY771709 M620, Isolate

M107, Isolate M290, Isolate

Pseudoalteromonas atlantica, X82134 Alteromonas sp., AB016267.1

M188, Isolate Pseudoalteromonas citrea, X82137

M246, Isolate Idiomarina sp., DQ985070.1

M206, Isolate M181, Isolate

M428, Isolate M140, Isolate

M222, Isolate* Vibrio lentus, AJ278880.1

M511, Isolate M227, Isolate

M241, Isolate Vibrio splendidus, AY227706

Vibrio parahaemolyticus, DQ026024.1M418, Isolate

Vibrio pectenicida, Y13830.1 M535, Isolate

M608, Isolate Enterovibrio norvegicus, AJ437193

M126, Isolate Aeromonas sobria, X74683

JUL3, Clone M297, Isolate*

Reinekea marinisedimentorum, AJ561121M121, Isolate*

M238, Isolate M254, Isolate

M123, Isolate M119, Isolate

M269, Isolate M302, Isolate

Marinomonas pontica, AY539835 JUL18_3, Clone

M266, Isolate Oleispira antarticam, AJ426421

M498, Isolate Marinobacter aquaeolei, AF173969

M503, Isolate M203, Isolate

Halomonas marisflava, AF251143 M543, Isolate

AUT70, Clone JUL12, Clone

Noriaki katanozaka, AB022713.1 AUT94, Clone

WIN38, Clone AUT48_2, Clone

Marine gammaproteobacterium, AY3863391 SPR78, Clone

WIN52, Clone WIN63, Clone

42

Isolate bacterioplankton diversity in the NW Mediterranean

Bacteroidetes. Most of the Bacteroidetes sequences belonged to the Flavobacteriaceae family

(Fig. 6, Table 3). Most isolates belonged to clusters Bacteroidetes 1 and 3. Cluster Bacteroidetes

1 contained ve different genera: Gramella, Suf avibacter, Dokdonia (including Krokinobacter),

Salegentibacter and Leeuwenhoekiella (Fig. 6). The last two genera had not been found in Blanes by

molecular techniques. There were three subclusters in cluster Bacteroidetes 3: sequences related to

Polaribacter including isolates and DGGE sequences from P-enrichment experiments and mesocosms

algal blooms; sequences related to Tenacibaculum with a similar origin; and isolate M341 related to

two sequences (BL98-5 and BL03-7) obtained by DGGE analysis of winter samples.

Cluster Bacteroidetes 2 included a large number of sequences af liated to the Delaware Bay

sequences ZD0403 and CF6 (Kirchman et al. 2003). These sequences were obtained exclusively by

molecular techniques. Cluster Bacteroidetes 4 was also composed of uncultivated phylotypes and

largely corresponded to the AGG58 clade (O’Sullivan et al. 2004). Isolates M466 and the M571,

belonging to the genus Microscilla, had not been found before in Blanes Bay by cultivation-independent

techniques.

Only one coincidence between isolates and environmental sequences was found: isolate M134

was very similar to a DGGE band from Banyuls and to the reference strain Dokdonia donghaensis.

This isolate has been whole-genome sequenced and is a promising candidate for environmentally

relevant physiological studies. In fact, MED134 is the only marine bacterium that has been shown to

grow better in the light than in the dark thanks to the presence of proteorhodopsin (Gómez-Consarnau

et al. 2007). Isolates M217 (Leeuwenhoekiella blandensis) and M152 (Polaribacter sp.) have also

been whole-genome sequenced. The genome of MED152 has been manually annotated and published

(González et al. 2008) and MED217 has been formally described as a new species (Pinhassi et al.

2006b).

43

Chapter I

Fig. 6.- As previous gures, for the Bacteroidetes.

Bacteroidetes 1

Bacteroidetes 2

Bacteroidetes 3

Bacteroidetes 4

0.10

Gramella echinicolla, AY608409 M454, Isolate M367, Isolate

WIN75, Clone Sufflavibacter litoralis, DQ868538.2

M220, Isolate Krokinobacter genikus, AB198086 BY-64, DGGE band Banyuls

Dokdonia donghaensis, DQ003277.1 M134, Isolate*

Dokdonia sp., EU195942.1M329, Isolate

Salegentibacter salinus, EF486353 M532, Isolate M445, Isolate Leeuwenhoekiella aequorea, AJ278780

M217, Isolate* M495, Isolate

Croceibacter atlanticus, AY163576.1 B-203.2, DGGE band experiments

Maribacter forsetii, AM712900 M381, Isolate

Auerimarina marisflavi, EF108215.1 B122, DGGE band experiments

SUM25, Clone WIN97, Clone

Unculture marine bacterium ZD0403, AJ400347 BL98-14, DGGE band in situ

B76-2, DGGE band experiments BL03-28, DGGE band in situ

SUM18_2, Clone WIN28, Clone

AUT25, Clone Unculture Bacteroidetes CF6, AY274860

AUT33, Clone Polaribacter dokdonensis, DQ004686

M152, Isolate* B-240, DGGE band experiments

Polaribacter irgensii, AY771712 B98, DGGE band experiments

M568, Isolate Tenacibaculum maritimum, M64629

M595, Isolate M621, Isolate

M622, Isolate B170, DGGE band experiments

M341, Isolate BL98-5, DGGE band in situ B100, DGGE band experiments BL03-7, DGGE band in situ

Flavobacterium succinicans, AM230494.1 AUT87, Clone

Fluviicola taffensis, AF493694.2B108, DGGE band experiments

AUT100, Clone WIN14, Clone

B180, DGGE band experiments B131, DGGE band experiments

B88, DGGE band experiments B-270, DGGE band experiments

Marine Eubacterial AGG58, L10946 B175, DGGE band experiments

SUM2, Clone AUT19_2, Clone

BL03-11, DGGE band in situ SPR33, Clone

Flexibacter tractuosus, M58789.1 M599, Isolate

AUT46, Clone Cyclobacterium marium, M62788

Microscilla arenaria, AB078078 M466, Isolate

M571, Isolate AUT77, Clone

Chitinophaga pinensis, AF078775 BL98-40, DGGE band in situ

44

Isolate bacterioplankton diversity in the NW Mediterranean

DISCUSSION

Since the introduction of molecular biology techniques in marine microbial ecology, a great effort

to describe the diversity of marine bacterioplankton communities has resulted in a comprehensive

inventory of primarily uncultured bacteria (e.g. Giovannoni & Stingl, 2007). As a consequence,

a great majority (86%) of clones of our different clone libraries were more than 97% similar to

sequences already deposited in GenBank. At the same time, a number of microbial ecology studies

have investigated the diversity of cultured isolates (Pinhassi et al. 1997; Eilers et al. 2000, 2001;

Hagström et al. 2000; Long & Azam 2001; Pinhassi & Berman, 2003; Stevens et al. 2005; Suzuki et

al. 1997; Selje et al. 2005). The increase in information about culturable marine bacteria is evident

when one considers that at the time of the pioneering study on the Roseobacter group by González

et al. (1997), the group consisted of only 6 described species while in 2005 more than 35 species

had been described forming up to 13 different major clusters (Buchan et al. 2005). The potential of

culturing to increase the known richness of any environment was also manifest in the present study.

The three main groups of bacteria increased the number of OTUs signi cantly. Moreover, at least for

the Gammaproteobacteria, the shape of the rarefaction curve suggested that potentially more diversity

could be uncovered by culturing than by cloning.

The most abundant groups with different techniques. Both culture dependent and independent

methods indicated that Alpha- and Gammaproteobacteria and Bacteroidetes were the predominant

groups. Together, these three groups accounted for 96% of the isolates, 95% of the clones, and 96% of

the DGGE bands (Table 1). A seasonal study carried out at the same sampling site by FISH indicated

that these three groups accounted for 75% of the EUB+ count (Alonso-Sáez et al. 2007). On average,

Alpha-, Gammaproteobacteria and Bacteroidetes accounted for 30, 4-8, and 11% of the DAPI count

respectively. If we take FISH to be the most quantitative technique to identify bacteria in marine

waters, Gammaproteobacteria were overrepresented in the three sets of sequences considered here,

particularly in the culture collection, where they represented 51% of the isolates, but also in clones

(17%) and DGGE bands (14%, Table 1). Alphaproteobacteria, on the other hand, showed the same

importance in the culture collection (30%) as in FISH, and were overrepresented by the two molecular

approaches (72% of clones and 57% of DGGE bands). Finally, the Bacteroidetes were represented

in percentages not too different form FISH by the three approaches (15% in cultures, 6% in clone

libraries, and 25% in DGGE bands). These comparisons should be kept in mind to interpret our

results.

45

Chapter I

Among the least represented groups, Betaproteobacteria and Actinobacteria were recovered by

both culturing and molecular approaches, while Deltaproteobacteria, Epsilonproteobacteria, and

Verrucomicrobia were only found by molecular techniques (Fig. 1). Obviously, our sequencing and

isolating efforts were not suf cient to retrieve these other groups in signi cant numbers, and we can

say very little about their diversity. The culturing and sequencing effort would have to be increased

at least by one order of magnitude to retrieve signi cant numbers of these groups. Recently, a sample

from the BBMO collected on September 2007 was analyzed by 454 pyrosequencing within the

International Census of Marine Microbes program (http://icomm.mbl.edu). Of the 8605 tag sequences

that could be assigned a taxonomic identity and were not cyanobacteria or chloroplasts, sequence

tags were distributes as follows: Proteobacteria 84%, Bacteroidetes 9%, Verrucomicrobia 2.5%,

Actinobacteria 1.7%, Firmicutes 1.9%, and ten other groups making less than 0.5% each (unpublished

data). Therefore, it is clear that Proteobacteria and Bacteroidetes were the major components of the

bacterioplankton, while the remaining groups were quantitatively minor components.

Estimates of richness in Blanes Bay. It is well known that bacteria isolated in pure culture frequently

do not coincide with sequences retrieved by molecular methods from the same ecosystem. Thus, it is

not surprising that the culturing effort contributed signi cantly to increase the richness of bacterial taxa

known from the BBMO. At least two new genera of Bacteroidetes and six of Gammaproteobacteria

were added to the taxa list of Blanes Bay. In terms of OTUs, known richness increased from 62 (clones

alone) to 116 (clones plus isolates). And considering the estimates from the Chao 1 indicator, richness

increased from 125 to 232 based on clones plus isolates. The degree of novelty provided by cultures

was highest for the Gammaproteobacteria and lowest for the Alphaproteobacteria. However, this is

still a minor fraction of the total richness in Blanes Bay, as a comparison to the 454 pyrosequencing

data mentioned above makes obvious: from a total of 14775 tags, there were 2719 unique ones

(representing different OTUs). This is one order of magnitude higher than the highest estimates from

the culturing and clone library approaches combined, and increases our knowledge about microbial

diversity on the one hand, and exemplify the large efforts needed to completely characterize it. With

the used methods we have barely scratched the surface of microbial biodiversity in this site.

46

Isolate bacterioplankton diversity in the NW Mediterranean

Isolates of potential environmental relevance. Both molecular and culturing methods showed

the major groups to be Alphaproteobacteria, Gammaproteobacteria and Bacteroidetes, as in most other

marine waters (Hagström et al. 2002). However, at the level of speci c 16S rRNA gene sequences

we found only minimal overlap between cultivated and uncultivated bacteria. This suggested that

most of the isolates were of low signi cance in the studied environment at the time of sampling.

This is a fairly general situation (e.g. Donachie et al. 2007). Thus, Suzuki et al. (1997) compared

127 isolates (37 different) and 58 clones (25 different) from one coastal Paci c seawater sample.

They found little overlap between isolate and clone sequences, although a few coincidences occurred

in the Gammaproteobacteria (1 Pseudomonas) and in the Alphaproteobacteria (2 Roseobacter, 1

Sphingomonas). In a similar comparison in North Sea tidal ats, Stevens et al. (2005) found four

identical clone and isolate sequences among a total of 188 isolates and 77 clones. In Blanes we

found 9 cases of identical or nearly identical 16S rRNA gene sequences shared by isolates and in situ

sequences (from clones or DGGE bands): 5 in the gammaproteobacteria, 3 in the alphaproteobacteria

and 1 in the Bacteroidetes (marked with arrows in Figs. 2-4). Obviously these cultures are good

candidates for laboratory studies of environmentally relevant microorganisms.

Expecting the results mentioned above, we reasoned that enrichment cultures would be a good

source of isolates of potential importance in the environment. By incubating sea water samples with

different carbon sources and/or nutrients we hoped to increase the abundance of bacteria that could

respond to such enrichments in nature. It is a frequent observation that novel bands appear in DGGE

gels in response to experimental manipulations (e.g. Massana et al. 2001, Allers et al. 2007). We had

speculated that these bands could correspond to bacteria that could be relatively easy to isolate in

pure culture. Indeed, we found some coincidences between isolates and DGGE bands that appeared

in response to enrichment experiments (Figs. 4-6). We envision these enrichment experiments as an

opportunity for bacteria in the seed-bank to grow and become members of the diversity of the system,

hopefully mimicking episodic enrichments in nature. One particular example is shown in Fig. 7.

The DGGE gel shown corresponds to an enrichment experiment with different P-rich compounds.

A speci c band (band B90) was seen to increase in intensity in the three treatments with P-rich

compounds, but it disappeared in the unenriched control. This band was identical to isolate M128, a

member of the genus Nereida. The band can be appreciated in the lane corresponding to the natural

samples before the beginning of the experiment. This band was too faint for successful sequencing

and, thus, would have remained unidenti ed by conventional molecular techniques. However, under

the appropriate conditions, it became one of the dominant members of the assemblage. This suggests

that M128 is normally a member of the rare biosphere, but may become important when episodes of

47

Chapter I

P enrichment occur. This also indicates that the composition of the bacterial assemblage likely has a

strong dependence on nutrient supply (or bottom-up control), something that gives an explanation for

the recurrence of the same bacterial populations year after year at the same season (Fuhrman et al.

2006).

Fig. 7.- DGGE gel showing the band patterns from different treatments in a nutrient addition experiment.

The in situ sample (labeled T0) showed a larger number of bands than the enrichments. A band increased in

intensity in P additions (B90 and B94), and was identical to isolate M128 (see Supplementary Table). This

band was also present at time 0, but in a much lower relative abundance.

To PO4 ATP DNA K

specific for P =M128 in situ as “seed bank”

48

Isolate bacterioplankton diversity in the NW Mediterranean

Microdiversity. We observed that a high fraction of the total diversity for both environmental

clones and isolates formed microdiversity clusters, constrained at 99% 16S rRNA gene sequence

similarity according to Acinas et al. (2004). While microdiversity clusters are often observed in

environmental clone libraries (Field et al. 1997, Acinas et al. 2004, Morris et al. 2005, Pommier et

al. 2006), we are not aware of any report on microdiversity for larger collections of marine bacterial

isolates. Thirty percent of the total observed OTUs (by both approaches) and 45% of the total species

estimated by Chao-1 were grouped as microdiversity clusters (Figs. 2A and 3A). However, this

microdiversity fraction was not equally distributed among all taxonomic groups, it was mostly due

to the Gammaproteobacteria in the case of isolates and to the Alphaproteobacteria in the case of

clones, while Bacteroidetes exhibited the lowest microdiversity. This low microdiversity could be

the effect a lower number of Bacteroidetes sequences (45) compared to 223 and 116 sequences from

Alpha- and Gammaproteobacteria, respectively. However, the low microdiversity of Bacteroidetes

found in our study is a characteristic also observed in other studies (Acinas et al. 2004; Pommier et

al. 2006). The ecological signi cance of the relatively high abundance of very divergent lineages for

Bacteroidetes remains unresolved. Previous studies show that microdiversity clusters could represent

important units of evolutionary differentiation as ecotypes in natural populations of bacteria (Cohan

2006, Thompson et al. 2005, Polz et al. 2007, Cohan & Perry 2007, Zo et al. 2008). The nding of

microdiversity clusters among isolate and clone sequences alike, suggests that similar evolutionary

processes act on cultured and uncultured bacteria. Thus, marine bacterial isolates could serve as

models to investigate how microdiversity clusters arise and provide indications of the ecological

signi cance of such genetic variability.

ACKNOWLEDGEMENTS

We thank Vanessa Balagué for her work with the clone libraries, and Laura Alonso-Sáez for sharing

data. Thanks also to Isabel Ferrera who helped with Figure 1, and to Thomas Pommier for the

preliminary 454 data. This work was supported by EU-project BASICS (CTM2005-04795/MAR),

Spanish MEC projects GenµMAR (CTM2004-02586/MAR), MODIVUS (CTM2005-04795/MAR)

and GEMMA (CTM2007-63753-C02-01/MAR), and the Swedish Research Council (grant 621-203-

2692 to JP).

49

Chapter I

Tabl

e 4.

- Id

entit

y of

the

bac

teri

a is

olat

ed f

rom

Bla

nes

Bay

. The

seq

uenc

es h

ave

diffe

rent

ori

gins

and

com

e fr

om d

iffer

ent

year

s. d

iffer

ent

seas

ons.

and

al

so f

rom

diff

eren

t ty

pes

of e

nric

hmen

t ex

peri

men

ts. I

ndic

ated

is

the

% s

eque

nce

sim

ilari

ty t

o ot

her

isol

ates

or

sequ

ence

s in

Gen

Ban

k. S

imila

rity

val

ues

are

base

d on

app

roxi

mat

ely

500

base

pai

rs. A

cces

sion

num

bers

are

sho

wn

for

the

sequ

ence

s in

clud

ed in

Gen

Ban

k. D

ate

of is

olat

ion

in d

d/m

m/y

y. O

rigi

ns:

in s

itu. f

rom

unt

reat

ed s

eaw

ater

sam

ples

; A

TP. f

rom

ATP

-am

ende

d se

awat

er c

ultu

res;

P. f

rom

inor

gani

c P

-am

ende

d se

awat

er c

ultu

res;

Con

trol

. sea

wat

er

cultu

res

with

out

enri

chm

ents

; et

c.

Whe

n th

ere

is m

ore

than

one

Gen

Ban

k ac

cess

ion

num

ber

for

a gi

ven

isol

ate

type

, onl

y th

e on

e w

ithou

t pa

rent

hesi

s is

us

ed in

the

text

.

Acc

#

Nam

e D

ate

isol

. O

rigi

n M

ost s

imila

r or

gani

sm(s

) si

mila

rity

lin

eage

fa

mily

• D

Q68

1131

M

107

16/0

3/01

C

ontr

ol

Pseu

doal

tero

mon

as a

gari

vora

ns A

J417

594

99.8

G

amm

a Ps

eudo

alte

rom

onad

acea

e (

DQ

6811

65)

M27

1 13

/11/

01

Con

trol

M

107

DQ

6811

31

100

(E

U25

3571

) M

306

13/1

1/01

P

M10

7 D

Q68

1131

10

0

• D

Q68

1132

M

111

16/0

3/01

C

ontr

ol

Alte

rom

onas

sp.

AY

5766

89

100

Gam

ma

Alte

rom

onad

acea

e

• D

Q68

1133

M

113

16

/03/

01

Con

trol

A

ltero

mon

as a

lvin

ella

e A

F288

360

98.7

G

amm

a A

ltero

mon

adac

eae

Alte

rom

onas

sp.

AJ3

9119

1 99

.8

• D

Q68

1134

M

114

16/0

3/01

C

ontr

ol

Tha

lass

obac

illus

dev

oran

s A

J717

299

99.8

Fi

rmic

utes

B

acill

acea

ea

• D

Q68

1136

M

119

20/0

3/01

A

TP

Mar

inom

onas

dok

done

nsis

DQ

0115

27.1

95.7

G

amm

a M

arin

omon

as (

DQ

6811

35)

M11

7 16

/03/

01

Con

trol

M

119

DQ

6811

36

99

• D

Q40

3809

M

121

20/0

3/01

A

TP

Mar

inom

onas

dok

done

nsis

DQ

0115

26

95

Gam

ma

Mar

inom

onas

Mar

ine

bact

eriu

m A

Y02

8196

98

.9

M23

4 25

/06/

01

DM

SP/P

M

252

25/0

6/01

A

TP

• D

Q68

1137

M

123

20/0

3/01

A

TP

Mar

inom

onas

pon

tica

AY

5398

35

95.5

G

amm

a

Hal

omon

adac

eae

M

221

23/0

5/01

A

TP

M23

6 25

/06/

01

DM

SP/P

• D

Q68

1138

M

126

20/0

3/01

A

TP

Ent

erov

ibri

o no

rveg

icus

AJ4

3719

3 99

.8

Gam

ma

Vib

rion

acea

e

V

ibri

o sp

. AJ3

1620

7 10

0

50

Isolate bacterioplankton diversity in the NW Mediterranean•

DQ

6811

39

M12

8-M

130

20/0

3/01

C

ontr

ol

Ner

eida

igna

va A

J748

748

99.4

A

lpha

R

hodo

bact

erac

eae

Alp

ha p

rote

obac

teri

um A

Y61

2764

99

.8

M15

6-M

157

20/0

3/01

In

situ

M

479

22/0

3/04

C

ontr

ol

M58

4-M

585

29/0

6/04

In

situ

M

592

29/0

6/04

In

situ

M

625

11/1

0/04

C

ontr

ol

(

DQ

6811

40)

M13

2 20

/03/

01

Con

trol

M

128

DQ

6811

39

100

• D

Q48

1462

M

134

20/0

3/01

C

ontr

ol

Dok

doni

a do

ngha

ensi

s D

Q00

3277

99

.4

CFB

Fl

avob

acte

riac

eae

Kro

kino

bact

er g

enik

us A

B19

8086

10

0

• D

Q68

1142

M

140

20/0

3/01

G

lu/P

V

ibri

o sp

. AF2

4227

4 99

.6

Gam

ma

Vib

rion

acea

e

V

ibri

o sp

. DQ

2193

66

100

• D

Q68

1143

M

146

20/0

3/01

P

Tha

lass

obac

illus

dev

oran

s A

J114

484

99.8

Fi

rmic

utes

B

acill

acea

e

H

alob

acill

us s

p. A

Y55

3123

10

0

• D

Q68

1144

M

149

20/0

3/01

P

Ros

eova

rius

cra

ssos

trea

e A

F114

484

98.1

A

lpha

R

hodo

bact

erac

eae

Unc

ultu

red

alph

apro

teob

acte

rium

DQ

3761

56

100

• D

Q48

1463

M

152

20/0

3/01

P

Pola

riba

cter

dok

done

nsis

DQ

0046

86

99.5

C

FB

Flav

obac

teri

acea

e

M

arin

e ba

cter

ium

AY

7458

62

99.9

• D

Q68

1145

M

155

20/0

3/01

In

situ

E

ryth

roba

cter

lito

ralis

AF4

6583

6 99

.0

Alp

ha

Ery

thro

bact

erac

eae

• D

Q68

1146

M

165

15/0

5/01

P

Phae

obac

ter

inhi

bens

AY

1777

12

98.4

A

lpha

R

hodo

bact

erac

eae

Ros

eoba

cter

sp.

AY

5766

90

99.8

• D

Q68

1147

M

169

15/0

5/01

P

Alte

rom

onas

alv

inel

lae

AF2

8836

0 99

.1

Gam

ma

Alte

rom

onad

aeca

e

U

ncul

ture

d m

arin

e ba

cter

ium

DQ

0711

62

100

M20

9 23

/05/

01

Gly

cero

l

M

480

22/0

3/04

In

situ

M

496

22/0

3/04

D

MSP

• D

Q68

1148

M

181

15/0

5/01

P

Vib

rio

tasm

anie

nsis

AJ5

1491

2 99

.1

Gam

ma

Vib

rion

acea

e

V

ibri

o sp

. AF0

2240

9

51

Chapter I

• D

Q68

1149

M

185

15/0

5/01

P

Alte

rom

onas

add

ita A

Y68

2202

99

.6

Gam

ma

Alte

rom

onad

acea

e

A

ltero

mon

as s

p. A

SP29

4361

10

0

M

188

Pseu

doal

tero

mon

as s

p. E

U44

0056

• D

Q68

1150

M

193

15/0

5/01

P

Phae

obac

ter

inhi

bens

AY

1777

12

97

Alp

ha

Rho

doba

cter

acea

eM

177-

M16

3-M

192-

M19

4 15

/05/

01

P

• D

Q68

1151

M

203

23/0

5/01

Py

ruva

te

Mar

inob

acte

r aq

uaeo

lei A

J000

726

100

Gam

ma

Alte

rom

onad

acea

e

M

. hyd

roca

rbon

ocla

stic

us A

Y66

9169

10

0

• D

Q68

1152

M

206

23/0

5/01

In

situ

Id

iom

arin

a se

osin

ensi

s A

Y63

5468

99

.2

Gam

ma

Idio

mar

inac

eae

Mar

inob

acte

r sp

. DQ

2707

61

99.9

• D

Q68

1154

M

216

23/0

5/01

P

Plan

ococ

cus

citr

eus

AF5

0000

8 99

.6

Firm

icut

es

Plan

ococ

cace

ae

Pl

anoc

occu

s sp

. DQ

1773

34

99.9

• D

Q29

4291

M

217

23/0

5/01

P

Lee

uwen

hoek

iella

acc

omm

odim

aris

AJ7

8098

0 97

C

FB

Flav

obac

teri

acea

e

C

ytop

haga

sp.

AY

7458

17

99.6

• D

Q68

1155

M

219

23/0

5/01

P

Koc

uria

pol

aris

AJ2

7886

8 99

.7

Aci

noba

ct.

Mic

roco

ccac

eae

Koc

uria

ros

ea D

Q17

6452

10

0

• D

Q68

1156

M

220

23/0

5/01

P

Sale

gent

ibac

ter

mis

hust

inae

AY

5766

53

94.1

C

FB

Flav

obac

teri

acea

e

U

ncul

ture

d or

gani

sm D

Q39

6362

99

.9

• D

Q68

1157

M

222

25/0

6/01

G

lu/P

V

ibri

o sp

. AF2

4227

4 10

0 G

amm

a V

ibri

onac

eae

Vib

rio

sp. D

Q49

2722

10

0

M

226

25/0

6/01

G

lu/P

• D

Q68

1158

M

227

25/0

6/01

G

lu/P

V

ibri

o gi

gant

is A

J582

807

97.8

G

amm

a V

ibri

onac

eae

Mar

ine

bact

eriu

m A

Y02

8201

98

.5

M23

0 25

/06/

01

Glu

/P

• D

Q68

1159

M

238

25/0

6/01

D

MSP

M

arin

omon

as b

land

ensi

s D

Q40

3809

99

.8

Gam

ma

O

cean

ospi

rilla

les

52

Isolate bacterioplankton diversity in the NW Mediterranean•

DQ

6811

60

M24

1 25

/06/

01

Glu

/P

Vib

rio

giga

ntis

AJ5

8280

7 10

0 G

amm

a V

ibri

onac

eae

M

285

13/1

1/01

C

ontr

ol

M30

0 13

/11/

01

P

(D

Q68

1141

) M

137

20/0

3/01

(E

U25

3572

) M

310

28/0

1/03

P

M24

1 D

Q68

1160

99

(E

U25

3575

) M

344

04/0

3/03

P

M

241

DQ

6811

60

99 (

EU

2535

91)

M50

9 22

/03/

04

Am

ino

acid

s M

241

DQ

6811

60

98

• D

Q68

1161

M

246

25/0

6/01

D

NA

Ps

eudo

alte

rom

onas

av

ipul

chra

AF2

9795

8 99

G

amm

a Ps

eudo

alte

rom

onad

acea

e

• D

Q68

1162

M

254

25/0

6/01

A

TP

Mar

ine

bact

eriu

m A

Y02

8196

99.2

G

amm

a M

arin

omon

as

M

arin

omon

as p

ontic

a A

Y53

9835

95.2

• D

Q68

1163

M

266

13/1

1/01

C

ontr

ol

Sacc

haro

phag

us d

egra

dans

AF0

5526

9 93

.6

Gam

ma

Oce

anos

piri

llale

s

O

leis

pira

ant

arct

icam

AJ4

2642

1 92

M

299-

M29

4-M

279

13/1

1/01

C

ontr

ol

M27

8-M

268

1

3/11

/01

C

ontr

ol

M31

9-M

324-

M32

6-M

327

28/0

1/03

C

ontr

ol

• D

Q68

1164

M

269

13/1

1/01

C

ontr

ol

Mar

inom

onas

dok

done

nsis

DQ

0115

27

97.6

G

amm

a M

arin

omon

as

M

arin

e ba

cter

ium

AY

0282

04

• D

Q68

1166

M

275

13/1

1/01

C

ontr

ol

Alte

rom

onas

alv

inel

lae

AF2

8836

0 99

.6

Gam

ma

Alte

rom

onad

acea

e

G

amm

a pr

oteo

bact

eriu

m A

B01

0855

(

DQ

6811

53)

M21

2 23

/05/

01

Gly

cero

l M

275

DQ

6811

66

99

M61

4 11

/10/

04

P

• D

Q68

1167

M

290

13/1

1/01

In

situ

Ps

eudo

alte

rom

onas

aga

rivo

rans

AJ4

1759

4 99

.6

Gam

ma

Pseu

doal

tero

mon

acea

e

Ps

eudo

alte

rom

onas

sp.

AY

7811

55

99.8

• D

Q68

1168

M

292

13/1

1/01

In

situ

A

ltero

mon

as a

ddita

AY

6822

02

99.6

G

amm

a A

ltero

mon

adac

eae

Unc

ultu

red

Alte

rom

onas

sp.

AY

6642

13

99.6

(E

U25

3590

) M

506

22/0

3/04

D

MSP

M

292

DQ

6811

68

99

• D

Q40

3810

M

297

13/1

1/01

In

situ

R

eine

kea

mar

inis

edim

ento

rum

AJ5

6112

1 94

.8

Gam

ma

Rei

neke

a

U

ncul

ture

d ba

cter

ium

AY

1723

02

96.1

53

Chapter I

• D

Q68

1169

M

302

13/1

1/01

P

Mar

inom

onas

dok

done

nsis

DQ

0115

27

98.3

G

amm

a M

arin

omon

as

U

ncul

ture

d M

arin

omon

as s

p. D

Q42

1663

98

.4

• E

U25

3573

M

329

04/0

3/03

In

situ

K

roki

noba

cter

dia

feri

tkos

AB

1980

89

96

CFB

Fl

avob

acte

race

ae

• E

U25

3574

.1

M34

1

Te

naci

bacu

lum

sp.

EU

0212

93

95

CFB

Fl

avob

acte

race

ae

• E

U25

3577

M

367

13/0

5/03

In

situ

G

ram

ella

ech

inic

ola

AY

6084

09

97

CFB

Fl

avob

acte

race

ae

• E

U25

3579

M

381

13/0

5/03

C

ontr

ol

Mar

ine

bact

eriu

m A

F359

541

99

CFB

Fl

avob

acte

race

ae

M

arib

acte

r fo

rset

ii su

bsp.

For

setii

AM

7190

0 97

• E

U25

3580

M

418

16/0

9/03

P

Gam

ma

prot

eoba

cter

ium

AF3

8414

2 99

G

amm

a V

ibri

onac

eae

• E

U25

3581

M

428

16/0

9/03

P

Gam

ma

prot

eoba

cter

ium

AF3

8414

2 99

G

amm

a V

ibri

onac

eae

• E

U25

3582

M

441

16/0

9/03

C

ontr

ol

Para

cocc

us s

p. A

B26

4129

99

A

lpha

R

hodo

bact

erac

eae

• D

Q68

1170

M

442

16/0

9/03

C

ontr

ol

Ery

thro

bact

er c

itreu

s A

F118

020

100

Alp

ha

Ery

thro

bact

erac

eae

M

537

25/0

5/04

In

situ

(E

U25

3576

) M

365

13/0

5/03

In

situ

M

442

DQ

6811

70

100

(E

U25

3578

) M

380

13/0

5/03

In

situ

M

442

DQ

6811

70

99 (

DQ

6811

72)

M45

6 22

/03/

04

In s

itu

M44

2 D

Q68

1170

99

.4

(E

U25

3594

) M

525

25/0

5/04

In

situ

M

442

DQ

6811

70

100

(D

Q68

1181

) M

539

25/0

5/04

In

situ

M

442

DQ

6811

70

96

• E

U25

3583

M

443

16/0

9/03

C

ontr

ol

Uni

dent

ied

bac

teri

um D

Q98

5889

98

A

lpha

E

ryth

roba

cter

acea

e

E

ryth

roba

cter

sp.

AY

6461

57

• E

U25

3584

M

445

16/0

9/03

In

situ

Sa

lege

ntib

acte

r sp

. EF5

2000

7 97

C

FB

Flav

obac

tera

ceae

• E

U25

3585

M

447

16/0

9/03

In

situ

Pa

llero

nia

mar

ism

inor

is A

Y92

6462

96

A

lpha

R

hodo

bact

erac

eae

• D

Q68

1171

M

454

22/0

3/04

In

situ

G

ram

ella

ech

inic

ola

AY

6084

09

99.9

C

FB

Flav

obac

teri

acea

e

54

Isolate bacterioplankton diversity in the NW Mediterranean•

DQ

6811

73

M45

8 22

/03/

04

In s

itu

Ery

thro

bact

er a

quim

aris

AY

4614

43

98.1

A

lpha

E

ryth

roba

cter

acea

e (

EU

2535

95)

M53

1 25

/05/

04

In s

itu

M45

8 D

Q68

1173

10

0

• D

Q68

1174

M

463

22/0

3/04

In

situ

A

uran

timon

as c

oral

icid

a A

Y06

5627

10

0 A

lpha

A

uran

timon

adac

eae

• E

U25

3587

M

464

22/0

3/04

In

situ

U

nide

nti

ed b

acte

rium

DQ

9858

96

99

Alp

ha

Rho

doba

cter

acea

e

M

ethy

larc

ula

sp. A

J534

207

98

• E

U25

3588

M

466

22/0

3/04

In

situ

M

icro

scill

a ar

enar

ia A

B07

8078

98

C

FB

Flam

meo

virg

acea

e

• D

Q68

1175

M

483-

M48

5 22

/03/

04

Con

trol

A

ntar

ctob

acte

r he

lioth

erm

us R

S219

16

97.3

A

lpha

R

hodo

bact

erac

eae

Unc

ultu

red

Ros

eoba

cter

AF2

4563

4 97

.9

• D

Q68

1176

M

495

22/0

3/04

D

MSP

L

eeuw

enho

ekie

lla a

equo

rea

AJ7

8098

0 99

.6

CFB

Fl

avob

acte

riac

eae

• D

Q68

1177

M

498

22/0

3/04

G

luco

se

Mar

inob

acte

r av

imar

is A

Y51

7632

99

.5

Gam

ma

Alte

rom

onad

acea

e

U

ncul

ture

d A

rctic

sea

ice

bt. A

Y16

5592

99

.9

• D

Q68

1178

M

503

22/0

3/04

D

MSP

M

arin

obac

ter

avim

aris

AY

5176

32

97.0

G

amm

a A

ltero

mon

adac

eae

Unc

ultu

red

Mar

inob

acte

r sp

. AY

6875

37

100

• E

U25

3589

M

504

22/0

3/04

D

MSP

L

imno

bact

er th

ioox

idan

s A

J289

885

99.9

B

eta

Bur

khol

deri

acea

e

• E

U25

3592

M

511

22/0

3/04

A

min

o ac

ids

Vib

rio

sp. D

Q92

3444

10

0 G

amm

a V

ibri

onac

eae

Vib

rio

sple

ndid

us A

J874

364

100

• E

U25

3593

M

513

22/0

3/03

A

ceta

te

Lim

noba

cter

thio

oxid

ans

AJ2

8988

5 99

.9

Bet

a B

urkh

olde

riac

eae

• D

Q68

1179

M

517

22/0

3/03

D

MSP

A

ltero

mon

as a

ddita

AY

6822

02

99.9

G

amm

a A

ltero

mon

adac

eae

(E

U25

3598

) M

555

25/0

5/04

In

situ

M

517

DQ

6811

79

97

• E

U25

3596

M

532

25/0

5/04

In

situ

Sa

lege

ntib

acte

r sl

inus

EF4

8635

3 99

C

FB

Flav

obac

teri

acea

e

• D

Q68

1180

M

535

25/0

5/04

In

situ

V

ibri

o pe

cten

icid

a Y

1383

0

99.7

G

amm

a V

ibri

onac

eae

55

Chapter I

(E

U25

3597

) M

536

25/0

5/04

In

situ

M

535

DQ

6811

80

100

• D

Q68

1182

M

543

25/0

5/04

In

situ

H

alom

onas

mar

isav

a A

F251

143

99.3

G

amm

a H

alom

onad

acea

e

• E

U25

3599

M

568

25/0

5/04

In

situ

Po

lari

bact

er d

okdo

nens

is D

Q48

1463

96

C

FB

Flav

obac

teri

acea

e

M

arin

e ba

cter

ium

AF3

5953

9 97

• E

U25

3600

M

571

25/0

5/04

In

situ

M

icro

scill

a ar

enar

ia A

B07

8078

99

C

FB

Flam

meo

virg

acea

e

• D

Q68

1183

M

588

29/0

6/04

In

situ

T

hala

ssob

acte

r st

enot

roph

icus

AB

1217

82

100

Alp

ha

Rho

doba

cter

acea

e

• D

Q68

1184

M

595

29/0

6/04

In

situ

Te

naci

bacu

lum

ska

gerr

aken

se A

F469

612

97.2

C

FB

Flav

obac

teri

acea

e

C

ytop

haga

sp.

AB

0735

87

100

• E

U25

3601

M

599

19/0

7/04

In

situ

M

icro

scill

a se

rice

a A

B07

8081

99

C

FB

Flex

ibac

teri

acea

e

M60

2 11

/10/

04

Glu

/P

• D

Q68

1185

M

604

11/1

0/04

P

Unc

ultu

red

Alte

rom

onas

sp.

AY

6642

13

100

Gam

ma

Alte

rom

onad

acea

e

• D

Q68

1186

M

605

11/1

0/04

G

lu/P

T

hala

ssob

ius

med

iterr

aneu

s A

J878

874

99.3

A

lpha

R

hodo

bact

erac

eae

• D

Q68

1187

M

606

11/1

0/04

G

lu/P

R

uege

ria

atla

ntic

a D

8852

7 97

.9

Alp

ha

Rho

doba

cter

acea

e

R

oseo

bact

er s

p. A

Y74

5859

10

0

• D

Q68

1188

M

608

11/1

0/04

G

lu/P

V

ibri

o pr

obio

ticus

AJ3

4506

3 99

.1

Gam

ma

Vib

rion

acea

e

V

ibri

o sp

. AJ3

1616

8 99

.6

• D

Q68

1189

M

609

11/1

0/04

P

Alte

rom

onas

alv

inel

lae

AF2

8836

0 99

.6

Gam

ma

A

ltero

mon

adac

eae

• D

Q68

1190

M

612

11/1

0/04

P

Tha

lass

obiu

s ge

latin

ovor

us D

8852

3 97

.5

Alp

ha

Rho

doba

cter

acea

e

U

ncul

ture

d ba

cter

ium

DQ

1174

34

99.9

• D

Q68

1191

M

613

11/1

0/04

Tha

lass

obiu

s ae

stua

rii A

Y44

2178

97

.9

Alp

ha

Rho

doba

cter

acea

e

U

ncul

ture

d al

pha

prot

eoba

cter

ium

AY

6639

68

100

56

Isolate bacterioplankton diversity in the NW Mediterranean•

DQ

6811

92

M61

8 11

/10/

04

Glu

/P

Phae

obac

ter

inhi

bens

AY

1777

12

97.6

A

lpha

R

hodo

bact

erac

eae

Unc

ultu

red

Ros

eoba

cter

sp.

AY

6639

66

98

• D

Q68

1193

M

620

11/1

0/04

G

luco

se

Gla

ciec

ola

mes

ophi

la A

J488

501

92.8

G

amm

a A

ltero

mon

adac

eae

Unc

ultu

red

mar

ine

euba

cter

ium

AF1

5967

2 97

.3

• D

Q68

1194

M

621

11/1

0/04

G

luco

se

Tena

ciba

culu

m s

kage

rrak

ense

AF4

6961

2 98

.2

CFB

Fl

avob

acte

riac

eae

• D

Q68

1195

M

622

11/1

0/04

G

luco

se

Tena

ciba

culu

m s

kage

rrak

ense

AF4

6961

2 98

.8

CFB

Fl

avob

acte

riac

eae

• D

Q68

1196

M

623

11/1

0/04

G

luco

se

Hyp

hom

onas

oce

aniti

s A

F082

797

99.1

A

lpha

H

ypho

mon

adac

eae

Mar

ine

bact

eriu

m A

F359

546

100

• D

Q68

1197

M

624

11/1

0/04

G

luco

se

Mar

icau

lis s

p. A

J301

665

99.7

A

lpha

H

ypho

mon

adac

eae

M

aric

aulis

mar

is A

J301

665

99

.9

57

Chapter I

REFERENCESAcinas SG, Klepac-Ceraj V, Hunt DE, Pharino Ch, Ceraj I, Distel DL & Polz M (2004) Fine-scale

phylogenetic architecture of a complex bacterial community. Nature. 430: 551-554.

Allers E, Gómez-Consarnau L, Pinhassi J, Gasol JM, Simek K & Pernthaler J (2007) Response of Alteromonadaceae and Rhodobacteriaceae to glucose and phosphorus manipulation in marine mesocosms. Environ Microbiol 9: 2417-2429.

Alonso-Sáez L, Balagué V, Sà EL, Sánchez O, González JM, Pinhassi J, Massana R, Pernthaler A, Pedrós-Alió C & Gasol JM (2007) Seasonality in bacterial diversity in north-west Mediterranean coastal waters: assessment through clone libraries, ngerprinting and FISH. FEMS Microbiol Ecol 60: 98-112.

Alonso-Sáez L, Vázquez-Domínguez E, Cardelús C, Pinhassi J, Sala M, Lekunberri I, Balagué V, Vila-Costa M, Unrein F, Massana R, Simó R & Gasol JM (2008) Factors controlling the year-round variability in carbon ux through bacteria in a coastal marine System. Ecosystems 11:397-409.

Bauer M, Kube M, Teeling H, Richter M, Lombardot T, Allers E et al. (2006) Whole genome analysis of the marine Bacteroidetes ‘Gramella forsetii’ reveals adaptations to degradation of polymeric organic matter. Environ Microbiol 8: 2201-2213.

Buchan A, González JM & Moran MA (2005) Overview of the marine Roseobacter lineage. Appl Environ Microbiol 71: 5665-5677.

Cohan FM (2006) Towards a conceptual and operational union of bacterial systematics, ecology, and evolution. Philos Trans Roy Soc B 361:1985-1996.

Cohan FM & Perry EB (2007) A systematics for discovering the fundamental units of bacterial diversity. Curr Biol 17:R373-R386.

DeLong EF (1992) Archaea in coastal marine environments. Proc Natl Acad Sci USA 89:5685-5689.

DeLong EF, Franks DG & Alldredge AL (1993) Phylogenetic diversity of aggregate-attached vs. free-living marine bacterial assemblages. Limnol Oceanogr 38: 924-934.

Díez B, Pedrós-Alió C, Massana R (2001) Study of genetic diversity of eukaryotic picoplankton in different oceanic regions by small-subunit rRNA gene cloning and sequencing. Appl Environ Microbiol 67:2932-2941.

Donachie, SP, Foster JS & Brown MV (2007). Culture clash: Challenging the dogma of microbial diversity. ISME J. 1: 97-99.

Eilers H, Pernthaler J, Glockner FO & Amann R (2000) Culturability and in situ abundance of pelagic bacteria from the North Sea. Appl Environ Microbiol 66: 3044-3051.

Eilers H, Pernthaler J, Peplies J, Glockner FO, Gerdts G & Amann R (2001) Isolation of novel pelagic bacteria from the German bight and their seasonal contributions to surface picoplankton. Appl Environ Microbiol 67: 5134-5142.

Field KG, Gordon D, Wright T, Rappe M, Urbach E, Vergin K & Giovannoni SJ (1997) Diversity and depth-speci c distribution of SAR11 cluster rRNA genes from marine planktonic bacteria. Appl Environ Microbiol 63:63-70.

Fuhrman JA, McCallum K & Davis AA (1992) Novel major archaebacterial group from marine plankton. Nature 356: 148-149

Fuhrman JA, Hewson I, Schwalbach MS, Steele JA, Brown MV & Naeem S (2006) Annually reoccurring

58

Isolate bacterioplankton diversity in the NW Mediterranean

bacterial communities are predictable from ocean conditions. Proc Natl Acad Sci USA 103: 13104-13109

Giovannoni SJ & Stingl U (2007) The importance of culturing bacterioplankton in the “omics” age. Nature 5: 820-826.

Giovannoni SJ, Britschgi TB, Moyer CL, Field KG (1990) Genetic diversity in Sargasso Sea bacterioplankton. Nature 345: 60-63.

Giovannoni SJ, Foster RA, Rappé MS & Epstein S (2007) New cultivation strategies bring more microbial plankton species into the laboratory. Oceanography 20: 62-69.

Gómez-Consarnau L, González JM, Coll-Lladó M, Gourdon P, Pascher T, Neutze R, Pedrós-Alió C & Pinhassi J (2007) Light stimulates growth of proteorhodopsin-containing marine Flavobacteria. Nature445: 210-213.

González JM & Moran MA (1997) Numerical dominance of a group of marine bacteria in the alpha-subclass of the class Proteobacteria in coastal seawater. Appl Environ Microbiol 63: 4237-4242.

González JM, Fernández-Gómez B, Fernàndez-Guerra A, Gómez-Consarnau L, Sánchez O, Coll-Lladó M, del Campo J, Escudero L, Rodríguez-Martínez R, Alonso-Sáez L, Latasa M, Paulsen I, Nedashkovskaya O, Lekunberri I, Pinhassi J & Pedrós-Alió C (2008) Genome analysis of the proteorhodopsin-containing marine bacterium Polaribacter sp. MED152 (Flavobacteria). Proc Nat Acad Sci USA 105: 8724-8729.

Hagström Å, Pommier T, Rohwer F, Simu K, Stolte W, Svensson D & Zweifel UL (2002) Use of 16S ribosomal DNA for delineation of marine bacterioplankton species. Appl Environ Microbiol 68: 3628-3633.

Klepac-Ceraj V, Ceraj I, Polz MF (2006) Clusterer: extendable java application for sequence grouping and cluster analyses, Online J Bioinformat, 7: 15-21

Kirchman DL, Yu LY & Cottrell MT (2003) Diversity and abundance of uncultured Cytophaga-like bacteria in the delaware estuary. Appl Environ Microbiol 69: 6587-6596.

Kisand V & Wikner J (2003) Combining culture-dependent and –independent methodologies for estimation of richness of estuarine bacterioplankton consuming riverine dissolved organic matter. Appl Environ Microbiol 69: 3607-3616

Long RA & Azam F (2001) Antagonistic interactions among marine pelagic bacteria. Appl Environ Microbiol67: 4975-4983.

López-García P, Rodríguez-Valera F, Pedrós-Alió C, Moreira D (2001) Unexpected diversity of small eukaryotes in deep-sea Antarctic plankton. Nature 409:603-607.

Massana R, Pedrós-Alió C, Casamayor EO, Gasol JM (2001) Changes in marine bacterioplankton phylogenetic composition during incubations designed to measure biogeochemically signi cant parameters. Limnol Oceanogr 46: 1181-1188.

Moran MA, Buchan A, González JM, Heidelberg JF, Whitman WB et al. (2004) Genome sequence of Silicibacter pomeroyi reveals adaptations to the marine environment. Nature 432: 910-913.

Morris RM, Vergin KL, Cho JC, Rappe MS, Carlson CA& Giovannoni SJ (2005) Temporal and spatial response of bacterioplankton lineages to annual convective overturn at the Bermuda Atlantic Time-series Study site. Limnol Oceanogr 50:1687-1696.

59

Chapter I

Muyzer AE, Brinkhoff T, Nübel U, Santegoeds C, Schäfer H & Wawer C (1997) Denaturing gradient gel electrophoresis (DGGE) in microbial ecology. In: Molecular Microbial Ecology Manual (Addermans ADL, van Elsas JD & Brujin FJ Eds.) Vol. 3.4.4, pp. 1-27.

Nichols D, Lewis K, Orjala J, Mo S, Ortenberg R, O’Connor P, Zhao C, Vouros P, Kaeberlein T & Epstein SS (2008) Short peptide induces an “uncultivable” microorganism to grow in vitro. Appl Environ Microbiol74: 4889-4897

O’Sullivan LA, Fuller KE, Thomas EM, Turley CM, Fry JC & Weightman AJ (2004) Distribution and culturability of the uncultivated “AGG58” cluster of the Bacteroidetes phylum in aquatic environments. FEMS Microbiol Ecol 47: 359-370.

Pace NR, Stahl DA, Lane DJ, et al. (1986) The analysis of natural microbial populations by ribosomal RNA sequences. Adv Microb Ecol 9: 1-55

Pedrós-Alió C (2006) Marine microbial diversity: can it be determined? Trends Microbiol 14: 257-263.

Pedrós-Alió C (2007) Dipping into the rare biosphere. Science 315: 192-193.

Pinhassi J, Zweifel UL & Hagström Å (1997) Dominant marine bacterioplankton species found among colony-forming bacteria. Appl Environ Microb 63: 3359-3366.

Pinhassi J & Berman T (2003) Differential growth response of colony-forming alpha- and gamma-proteobacteria in dilution culture and nutrient addition experiments from lake Kinneret (Israel), the eastern Mediterranean Sea, and the Gulf of Eilat. Appl Environ Microbiol 69: 199-211.

Pinhassi J, Sala MM, Havskum H, Peters F, Guadayol Ò, Malits A & Marrasé C (2004). Changes in bacterioplankton composition under different phytoplankton regimens. Appl Environ Microbiol 70: 6753-6766

Pinhassi J, Gómez-Consarnau L, Alonso-Sáez L, Sala MM, Vidal M, Pedrós-Alió C & Gasol JM (2006) Seasonal changes in bacterioplankton nutriente limitation and their effects on bacterial community composition in the NW Mediterranean Sea. Aquat Microb Ecol 44: 241-252

Pinhassi J, Pujalte MJ, Macián MC, Lekunberri I, González JM, Pedrós-Alió C & Arahal DR (2007) Reinekea blandensis sp. nov., a marine genome-sequenced gammaproteobacterium. Internat J. Syst Evol Microbiol 57: 2370-375.

Polz MF, Hunt DE, Preheim SP & Weinreich DM (2006) Patterns and mechanisms of genetic and phenotypic differentiation in marine microbes. Philos Trans Roy Soc B 361:2009-2021.

Pommier T, Canbäck L, Riemann L, Boström KH, Simu K, Lundberg P, Tunlid A & Hagström Å (2006) Global patterns of diversity and community structure in marine bacterioplankton. Mol Ecol 16: 867-880.

Rusch DB, Halpern A, Sutton G, Heidelberg KB et al. (2007) The Sorcerer II: Global ocean sampling expedition: northwest Atlantic through eastern tropical Paci c. PLOS biology 5: 0398-0431.

Sánchez O, Gasol JM, Massana R, Mas J & Pedrós-Alió C (2007) Comparison of different denaturing gradient gel electrophoresis primer sets for the study of marine bacterioplankton communities. Appl Environ Microbiol 73: 5962-5962.

Schäfer H, Bernard L, Courties C, Lebaron P, Servais P, Pukall R, Stackebrandt E, Troussellier M, Guindulain T, Vives-Rego J & Muyzer G (2001) Microbial community dynamics in Mediterranean nutrient-enriched seawater mesocosms: changes in the genetic diversity of bacterial populations. FEMS Microbiol Ecol 34: 243-253.

60

Isolate bacterioplankton diversity in the NW Mediterranean

Schauer M, Balagué V, Pedrós-Alió C & Massana R (2003) Seasonal changes in the taxonomic composition of bacterioplankton in a coastal oligotrophic system. Aquat Microb Ecol 31: 163-174.

Selje N, Brinkhoff T & Simon M (2005) Detection of abundant bacteria in the Weser estuary using culture-dependent and culture-independent approaches. Aquat Microb Ecol 39: 17-34.

Simu K, Holmfeldt K, Zweifel UL & Hagström Å (2005) Culturability and coexistence of colony-forming and single-cell marine bacterioplankton. Appl Environ Microbiol 71: 4793-4800

Sogin ML, Morrison HG, Huber JA, Welch DBM, Huse SM, Neal PR, Arrieta JM & Herndl GJ (2006) Microbial diversity in the deep sea and the underexplored “rare biosphere”. Proc Natl Acad Sci USA 103: 12115-12120.

Stevens H, Stubner M, Simon M & Brinkhoff T (2005) Phylogeny of Proteobacteria and Bacteroidetes from oxic habitats of a tidal at ecosystem. FEMS Microbiol Ecol 54: 351-365.

Stingl U, Tripp HJ & Giovannoni (2007) Improvements of high-throughput culturing yielded novel SAR11 strains and other abundant marine bacteria from the Oregon coast and the Bermuda atlantic time series study site. ISME J 1: 361-371.

Suzuki MT, Rappe MS, Haimberger ZW, Win eld H, Adair N, Strobel J & Giovannoni SJ (1997) Bacterial diversity among small-subunit rRNA gene clones and cellular isolates from the same seawater sample. Appl Environ Microb 63: 983-989.

Thompson JR, Pacocha S, Pharino C, Klepac-Ceraj V, Hunt DE, Benoit J, Sarma-Rupavtarm R, Distel DL & Polz MF (2005) Genotypic diversity within a natural coastal bacterioplankton population. Science 307: 1311-1313.

Venter JC, Remington K, Heidelberg JF, et al. (2004) Environmental genome shotgun sequencing of the Sargasso Sea. Science 304: 66-74.

Vila-Costa M, Kiene RP & Simó R (2008) Seasonal variability of the dynamics of dimethulated sulfur compounds in a coastal northwest Mediterranean site. Limnol Oceanogr 53: 198-211

Zengler K, Toledo G, Rappe M, Elkins J, Mathur EJ, Short JM & Keller M (2002) Cultivating the uncultured. P Natl Acad Sci USA 99: 15681-15686.

Zo YG, Chokesajjawatee N, Arakawa E, Watanabe J, Huq A & Colwell RR (2008) Appl Environ Microb 74: 2915-2920

ZoBell CE (1943) The effect of solid surfaces upon bacterial activity. J Bacteriol 46: 39-56

Changes in bacterial activity and community composition caused by realistic exposure to oil in mesocosm experiments

Itziar Lekunberri, Alejandra Calvo-Díaz, Eva Teira, Xelu A.G. Morán, Francesc Peters, Mar Nieto-Cid, Oscar Espinoza-González, Isabel G. Teixeira and Josep M Gasol

CHAPTER 2

La Tierra no es herencia de nuestros padres sino prestamo de nuestros hijos

Proverbio indio

63

Chapter II

ABSTRACT

We studied the effects of the Prestige oil spill on the Ría de Vigo bacterial abundance, production

and community structure, by using mesocosms experiments (ca. 3500 L) lled with water from the

center of the Ría, to which we added a realistic concentration of polyciclic aromatic hydrocarbons

(approximately 20-30 μg L-1 chrysene eq). We followed the changes in bacterial activity by leucine

and thymidine incorporation, and changes in bacterial assemblage structure by 16S-rDNA DGGE as

well as CARD-FISH, during the four relevant periods of the seasonal cycle: spring bloom, summer

strati cation, autumn downwelling and winter. In addition, simultaneous to one of the mesocosms

experiments, we run microcosms with fuel additions equivalent to 0.5x, 1x, 2x and 4x the treatment

imposed to the mesocosms. We detected effects on bacterial diversity (increased richness) and

function (more bacteria and more production) only in the summer experiment. In the microcosm

experiments we found similar effects at PAH concentrations of ca. 40 µg L-1 (2x 20 µg L-1) and

clear detrimental effects on phytoplankton at concentrations of ca. 80 µg L-1, with large development

of gammaproteobacteria. We also show that the development of heterotrophic nano agellate in the

mesocosms is what causes the most important changes in bacterial assemblage structure.

64

Effects of oil on bacterial activity and diversity

INTRODUCTION

Bacterioplankton communities inhabiting the plankton of coastal regions use multiple sources of

carbon and nutrients (e.g. Moran et al. 1999, Mou et al. 2008). These sources have different origins:

autochthonous phytoplankton production, carbon and nutrients coming from the watersheds, or

allochthonous nutrient and C sources that have originated because of human activities. Petroleum

hydrocarbons are some of these organic carbon sources, and are some of the most widespread

contaminants in the environment (Santas et al. 1999). Polyciclic aromatic hydrocarbons (PAHs) are

particularly in uential because of their continuous release into the water, their persistence, and their

well-known detrimental effects on marine invertebrates (Preston, 2002), which make them subject of

public concern.

But bacteria are able to decompose these carbon molecules (e.g. Harayama et al. 1999, Kasai

et al. 2002) and may consider chronic PAHs inputs to seawater just one allochthonous source of

carbon and energy (Castle et al. 2006). However, immediately after an oil spill, the soluble fraction

of polycyclic aromatic hydrocarbons (PAHs) is released into the water column and may represent

point concentrations of PAHs much higher than the values measured in chronic exposures and, thus,

affect bacterioplankton communities in ways different than just as a regular C source. The effect

of oil spills on the natural bacterioplankton communities of coastal ecosystems has been studied

using in situ measurements (e.g. Bode et al. 2006), laboratory experiments (Maruyama et al. 2003,

Castle et al. 2006, Cappello et al. 2006), and experimental systems such as mesocosms (Siron et al.

1993, Ohwada et al. 2003, Cappello et al. 2007). This last type of approach avoids the effects caused

by advection, diffusion and mixing under natural conditions, and it also offers the possibility of

comparison with natural populations to which PAHs were not added, thus allowing statistical testing

of hypotheses. These types of analyses have reported contradictory results about the effects of the oil

spills on plankton communities. Often, no effect on primary producers has been found (e.g. Varela

et al. 2006) while simultaneous signi cant effects on bacterial variables have been reported (Bode

et al. 2006). In several experiments, bacteria have been shown to be stimulated (e.g. Delille & Siron

1993, Ohwada et al. 2003, Bode et al. 2006, Castle et al. 2006, Cappello et al. 2007, Dalby et al.

2008), neutrally affected (Maruyama et al. 2003) or repressed (Garcia et al. 1998, Sargian et al. 2005)

by high concentrations of PAHs. In most cases, a shift in bacterial community structure has been

observed (Gerdts et al. 2004, Castle et al. 2006, McKew et al. 2007) resulting in decreasing richness

65

Chapter II

(Röling et al. 2002) and elevated contributions of gammaproteobacteria (Yakimov et al. 2004, Gerdts

et al. 2004, Cappello et al. 2007), particularly groups such as Alcalinovorax (Cappelloet al. 2007) or

Cycloclasticus (Kasai et al. 2002, Teira et al. 2007).

The concentration of PAHs used to experimentally simulate oil spills have ranged from 900 mg crude

oil L-1 (Cappello et al. 2007) to 4.5 µg L-1 chrysene equivalents (Ohwada et al. 2003), with all possible

intermediate values. It might be possible that some of the reported effects of oil are concentration-

dependent, so that they are signi cant at concentrations which are never experienced by the natural

bacteria assemblages. In fact, the lack of effect of oil spills on phytoplankton communities has been

attributed to the reduction in exposure concentration related to the dispersion of the oil caused of the

hydrographic water movements. Accordingly, it can be hypothesized that if PAHs do not cause effects

on microbial diversity and function, it is because the communities are habituated to a certain level of

these substances in the water, as parts of their normal carbon pools used for growth.

To explore this hypothesis, we added realistic oil concentrations to mesocosms in order to test the

effect of PAHs on plankton communities of the Ría de Vigo. To evaluate the response of bacterioplankton

we measured bacterial activity, diversity and community structure in a series of experiments. Bacterial

activity and diversity measurements can represent useful tools for characterizing the way in which

natural and human stressors drive changes in natural ecosystems. We used relatively low PAHs

concentrations (20-30 μg L-1chrysene eq.) that were, however, sixfold higher than the 90% percentile

of the concentrations found along the Galician coast after the accident (González et al. 2006). To assess

the changes in bacterial assemblage structure we used the DGGE ngerprinting technique, which

serves well to compare microbial assemblages and assess temporal and spatial changes in structure

(Casamayor et al. 2002), also providing estimates of diversity. In general, we found little effects with

the chosen concentrations and, thus, we also performed a concentration-dependent experiment using

a gradient from 0 to 80 µg L-1 chrysene eq. to explore the threshold concentration at which bacterial

assemblage composition and metabolism were affected.

The experiments were part of the project IMPRESION, designed to test the effect of the Prestige

oil spill on plankton communities of the Ría de Vigo. This coastal embayment is an area affected by

marked seasonal cycles of coastal winds (Nogueira et. al. 1997). The annual cycle is divided in an

upwelling season (March to September), with short relaxation intervals that enhance productivity

(Álvarez-Salgado et al. 2002), and a downwelling season (from October to March), characterized

by low phytoplankton biomass and primary production. Microbial plankton composition is different

66

Effects of oil on bacterial activity and diversity

in the different situations: upwelling, downwelling or transition (Figueiras et al. 2002, Nogueira et

al. 1997, Alonso-Gutiérrez et al. submited, Teira et al. 2008). Thus, we explored the effects of oil

on bacterial community structure and activity at the four relevant parts of the seasonal cycle: spring

bloom, summer strati cation, autumn upwelling and winter.

MATERIALS & METHODS

Experimental setup and sampling. The experimental setup was established in a small harbor of

a protected bay, four times during a year. As detailed in Teira et al. (2007), six mesocosms of 1.5 m in

diameter and 2 m deep were lled with seawater at the center of the Ría and transported to the harbor

to which they were tight. The bags were lled from their bottom through a 200 µm mesh in order to

exclude mesozooplankton and facilitate good replication within bags. Two of the mesocosms were

used as controls, received no additions and were not treated in any way except for the 200 µm pre-

ltration. Two were added a low concentration of soluble PAHs (5-10 µg L-1 chrysene equivalents)

and two a high concentration of soluble PAHs (approx. 20-30 µg L-1 chrysene eq.). The experiments

lasted 9 days after the addition. Samples were taken every day during the rst 5 days, and thereafter

every 2 days. Since during the rst two experiments there was no response in the low-concentration

mesocosms, in the subsequent experiments this treatment was eliminated and instead we triplicated

the control and the high-PAH concentration treatments.

The PAHs were added as a soluble fraction which was prepared from 15 kg of Prestige-like heavy

oil provided by the “O cina Técnica de Coordinación del Programa de Intervención Cientí ca en la

Catástrofe del Prestige” mixed with 300 L of 0.2 µm- ltered seawater. The mixture was vigorously

stirred during 4 hours and the resulting extract, of approximately 700 µg L-1 soluble PAHs, was separated

from the insoluble fuel oil by decantation, collected on 25 litres polyethylene barrels, from which it

was added to the mesocosms. PAH concentrations were measured following the MARPOLMON

protocol (UNESCO, 1984), with modi ed volumes, and referred to a chrysene standard (see details

in González et al. 2006).

The Prestige oil spill was found to consist of a complex mixture of hydrocarbons, where the aromatic

fraction (mainly naphthalene, phenanthrene, and alkyl derivatives) comprised ca. 53% (Alzaga et al.

2004). PAHs represented 99.7% of the water soluble fraction of the Prestige oil and alkanes were

almost undetectable in that fraction (J. Albaigés, pers. comm.). Although it was not possible to use

exactly the Prestige oil, we used oil with a very similar composition.

67

Chapter II

Addition of oil was done after having taken the time 0 sample, and sampling of the mesocosms

was then done daily at sunset with an integrated tube that collected water down to 1 meter above

the bottom of the tanks. The experiments were carried out at the 4 periods of the seasonal cycle in

the coastal NE Iberian Atlantic waters (Nogueira et al. 1997): spring bloom (March 2005), summer

strati cation (July 2005), autumn upwelling (September 2005), and winter (January 2006).

Additional microcosm experiments. As described in Teira et al. (2007), a microcosm experiment

was run in parallel to the winter mesocosms using the same initial seawater, in order to determine the

response of the community to a larger gradient of PAH concentration. For the microcosm experiment

ten 5 L-PET bottles were kept opened and refrigerated by circulating surface seawater. A gradient of

four concentrations (10, 20, 40 and 80 µg L-1 equivalent chrysene) plus the control were prepared in

duplicates. This is equivalent to 0.5x, 1x, 2x, and 4x what was added in the simultaneous mesocosms.

The microcosms were kept for 8 days, sampling every 24 h for bacterial abundances, bulk heterotrophic

production and bacterial community structure analyses.

Organism abundances. Chlorophyll a concentration was used as an estimator of phytoplankton

abundance. Two hundred and fty millilitres of seawater were sequentially ltered through 20, 2

and 0.2 µm polycarbonate lters. The lters were frozen and chlorophyll-a extracted in 90% acetone

for 24 h at 4°C. The uorescence of the extracted chlorophyll-a was measured with a Turner-TD-

700 uorometer previously calibrated with pure chlorophyll-a. Total chlorophyll-a concentration was

obtained by addition of the concentration in each of the three size classes.

Bacterial abundances were measured by ow cytometry as described in Gasol & del Giorgio (2000).

The samples were run in a Becton-Dickinson FACSCalibur cytometer after staining with SybrGreen

2.5 µM nal concentration, Molecular Probes) and bacteria were detected by their signature in a plot

of side scatter (SCC) vs FL1 (green uorescence). Calibration of the cytometer, and of the SSC-size

relationship was done as described in Calvo-Díaz & Morán (2006).

HNF abundances- Samples for nano agellate abundance were collected four or ve times in every

experiment. Subsamples of 60-80 mL were xed with glutaraldehyde (1%, nal concentration).

The samples were ltered through 0.6 µm black polycarbonate lters (Milipore), and stained with

DAPI (4, 6 - diamidino-2- phenylindole, Porter and Feig, 1980) at a nal concentration of 5 µg mL-1

(Sieracki et al. 1995). Abundance of these microorganisms was determined at 1000X magni cation

with an Olympus-BX40-102/E epi uorescence microscope. The heterotrophic and phototrophic

nano agellates were counted under both UV excitation (blue uorescence) and blue excitation (B2

68

Effects of oil on bacterial activity and diversity

lter). Under blue light we discriminated phototrophic nano agellates (PNF, showing red-orange

auto uorescence, and/or plastidic structures) from colorless nano agellates classi ed as heterotrophic

(HNF). With this method we could not distinguish mixotrophic nano agellates. Random 10 mm

transects (100 µm width) were examined and between 20 and 100 HNF, and between 10 and 50 PNF

per lter were counted.

Bacterial activity and production. Bacterial heterotrophic production was estimated using both

the 3H-leucine (Kirchman et al. 1985) and 3H-thymidine (Fuhrman & Azam 1982) incorporation

methods. Triplicate or cuatriplicate aliquots of 1.2 ml were taken for each sample and one or two TCA

killed controls. The Leu tracer was used at 40 nM and the TdR tracer at 20 nM nal concentrations

in incubations lasting approximately 2 h. The incorporation was stopped with the addition of 120 µl

of cold 50% TCA to the samples and, after mixing, they were kept frozen at –20ºC until processing,

which was carry out by the centrifugation method of Smith & Azam (1992). The samples were counted

on a Beckman scintillation counter, 24 h after addition of 1 ml of scintillation cocktail. We used the

standard conversion factors of 3.1 kg C mol Leu-1 and 40 kg C mol TdR-1, since we had previously

determined CFs in the Ría to be on average close to these theoretical ones (Morán et al. 2002). The

nal value of heterotrophic bacterial production was computed as the average of the Leucine and

Thymidine-based determinations.

Collection of community DNA. Microbial biomass was collected by sequentially ltering around

5 L of seawater through a 3 μm pore size polycarbonate lter (Milipore, 46 mm) and 0.2 μm sterivex

lter (Milipore). The Sterivex units were lled with 1.8 ml of lysis buffer (50 mM Tris-HCl pH 8.3, 40

mM EDTA pH 8.0, 0.75 M sucrose) and stored at -80ºC. We used the sterivex units for the analyses.

Microbial biomass was treated with lysozime, proteinase K and sodium dodecyl sulphate, and the

nucleic acids were extracted with phenol and concentrated in a Centricon-100 (Milipore). Nucleic

acids were extracted by a standard protocol using phenol/chloroform (see details in Schauer et al.

2003).

Fingerprinting analysis. DGGE and gel analysis were performed essentially as described before

(Schauer et al. 2000; Sánchez et al. 2007). Brie y, 16S rRNA gene fragments (around 550 bp in length)

were ampli ed by PCR using the universal primer 907rm and the bacterial speci c primer 358f, with

a GC-clamp. PCR products were loaded on a 6% polyacrylamide gel with a DNA-denaturant gradient

ranging from 40-80%. The gel was run at 100V for 16h at 60ºC in 1x TAE running buffer. DGGE gel

images were analyzed using the Diversity Database software (BIO-RAD). A matrix was constructed

69

Chapter II

for all lanes taking into account the relative contribution of each band (in percentage) to the total

intensity of the lane. Based on this matrix, we obtained a dendogram by UPGMA clustering method

(Euclidean distances, software Statistica 6.0).

Bacterial community composition. We used CARDFISH to study the composition of the bacterial

communities in the microcosm experiment. Five mililitres were xed with paraformaldehyde (2% nal

concentration) and after 12–18 h storage at 4°C in the dark were ltered through a 0.2 µm polycarbonate

lter (Millipore, GTTP, 25 mm lter diameter) supported by a cellulose nitrate lter (Millipore, HAWP,

0.45 mm), washed twice with Milli-Q water, dried and stored in a microfuge vial at -20°C until further

processing. We used oligonucleotide probes speci c for the domain Eubacteria (EUB338) (Amann et

al. 1990), the Alpha- (ALF968) (Glöckner et al. 1999) and Gammaproteobacteria (GAM42a) (Manz

et al. 1992) subclasses and the class Flavobacteria of phylum Bacteroidetes (CF319a) (Manz et al.

1996). We also used probe CYU829 for Cycloclasticus (Maruyama et al. 2003, Teira et al. 2007).

The Eub antisense probe Non 338 probe was used as negative control. Further details in Teira et al.

(2008).

Statistical analysis. We used ANOVA to test for the differences between treatments. A repeated

measures ANOVA (RMANOVA) with one within-subjects factor (treatment) was conducted to

discriminate the treatment from the time effects and all possible interactions. Time is a within-subject

factor because the same mesocosms is sampled at sequential time periods (every-24-48 h). One way

ANOVA was done to determine differences in the microcosm experiment. We used software Statistica

6.0 and JMP 7 for all analyses.

RESULTS

The initial environmental conditions were very different for each season (Table 1). Inorganic

nutrients were very abundant in winter and spring, while chlorophyll was very high particularly in

autumn, when dissolved inorganic nitrogen (DIN) -but not silicate- was also high, as our sampling

coincided with the decline of an algal bloom (Fig. 1C). Bacterial abundance and heterotrophic

production were high in summer and fall.

70

Effects of oil on bacterial activity and diversity

Tabl

e 1.

- Ini

tial c

ondi

tions

of t

he m

esoc

osm

s ex

peri

men

ts. T

he m

icro

cosm

exp

erim

ent w

as s

imul

tane

ous

to th

e Ja

nuar

y’06

mes

ocos

ms.

Mea

n ±

stan

dard

err

ors.

Tem

pera

ture

(T)

in ºC

, sal

inity

(Sa

l), d

isso

lved

inor

gani

c ni

trog

en (

DIN

) in

μM

, dis

solv

ed in

orga

nic

phos

phor

us (

DIP

) in

μM

, si

licat

e (S

iO4)

in μ

M, p

artic

ulat

e or

gani

c ca

rbon

(P

OC

) in

μM

, par

ticul

ate

orga

nic

nitr

ogen

(P

ON

) in

μM

, chl

orop

hyll-

a co

ncen

trat

ion

(Chl

a) in

m

g m

-3, t

otal

pro

kary

otic

abu

ndan

ce (P

A) i

n x1

05 cel

l ml-1

and

BH

P in

μg

C L

-1 d

-1 (m

ean

of L

eu a

nd T

dR w

ith s

tand

ard

conv

ersi

on fa

ctor

(CF

) 3.1

kg

C m

ol L

eu-1 a

nd 4

0 kg

C m

ol T

dR).

N=

6.

Exp

erim

ent

Dat

e T

Sa

l D

IN

DIP

Si

O4

Chl

a PA

B

HP

Spri

ng

02/0

3/05

10

.5 ±

0.0

35

.48 ±

0.02

4.

40 ±

0.0

8 0.

52 ±

0.0

0 3.

17 ±

0.0

4 3.

01 ±

0.0

1 6.

13 ±

0.1

8 3.

05 ±

0.4

Sum

mer

02

/07/

05

20.8

± 0

.0

35.0

2 ±

0.01

0.

58 ±

0.1

1 0.

15 ±

0.0

1 0.

59 ±

0.0

5 1.

50 ±

0.0

1 9.

92 ±

0.7

2 69

.94 ±

3.4

Aut

um

22/0

9/05

15

.4 ±

0.0

35

.73 ±

0.01

5.

66 ±

0.7

3 0.

51 ±

0.0

8 0.

41 ±

0.0

2 9.

22 ±

0.6

3 13

.82 ±

0.40

26

.25 ±

8.11

Win

ter

24/0

1/06

12

.4 ±

0.0

35

.60 ±

0.00

7.

70 ±

0.4

4 0.

48 ±

0.0

2 3.

72 ±

0.1

5 0.

54 ±

0.0

1 5.

29 ±

0.0

6 16

.88 ±

0.31

71

Chapter II

The added PAHs disappeared from the water in a similar exponential way in the 4 experiments

(Fig. 1). On day 2 of the experiments, only ca. 35% of the added PAHs were found in solution except

in summer, where the high ambient temperature caused faster volatilization of the added compounds

and only 10% remained in the water (Fig. 1). No PAHs were detected in summer after day 4, while

in the other months between 5 and 15% of the added substrates remained in the mesocosms at day 4.

The PAHs exponential decay rate was pretty similar in all the experiments (around -0.35 d-1) except

for the summer experiment in which it was higher -0.7 d-1 (Table 2).

Fig 1.- Changes in dissolved PAH concentration and in chlorophyll a concentration in the control mesocosms (white circles) and the mesocosms to which PAHs had been added (black circles). Spring experiment (A), Summer experiment (B), Autumn experiment (C) and Winter experiment (D). Average plus standard deviation of 2 (spring) or 3 (the other months) replicated mesocosms.

Chla

(mg

m-3)

Spring Summer

% o

f PAH

s rem

ainin

g

Autumn

0

5

10

15

0

20

40

60

80

100

120

0 2 4 6 8

Winter

Chla

(mg m

-3)

% of

PAHs

rem

ainin

g

0

5

10

15

0

20

40

60

80

100

120

0 2 4 6 8

0

5

10

15

0

20

40

60

80

100

120

0 2 4 6 80

5

10

15

0

20

40

60

80

100

120

0 2 4 6 8

Control

PAHs

PAHs concentration

A

C

B

D

Time (days)Time (days)

72

Effects of oil on bacterial activity and diversity

Chlorophyll a. The evolution of chlorophyll in the mesocosms was different in the different

experiments. In spring (Fig 1A, C) water con nement caused a bloom (up to 10 mg m-3), probably

because we had sampled growing healthy algae. In contrast, in autumn, the starting values of Chla

were 9 mg m-3 and decreased drastically during the four subsequent days. In summer and winter there

were few changes in chlorophyll during the experiments, just small slow increases. In general, no

differences were found because of the oil additions. They were only signi cant (<0.01) in the summer

experiment (Fig 1B, Table 3) when the increase after day four to the end of the experiment, was higher

(ca. 2x) in the oil-added mesocosms (Fig. 1B).

Bacterial abundance. The changes in bacterial abundances in the experiments also showed different

trends in the different seasons (Fig. 2). Except in the autumn experiment, there was a general initial

increase in abundance (day 1) that soon was followed by a decrease in bacterial abundance (days 3-4)

and by a second peak, higher in magnitude and at about days 6-8. The autumn experiment showed just

one peak. The effect of oil was obvious in the summer experiment (Fig 2b) when the two peaks were

higher in magnitude when PAHs had been added. When separating the rst (days 0-4) from the second

(days 5-8) part of the experiments, the PAH treatments affected the second part of the experiment in

all experiments (Table 3). This suggests that the signi cance observed using all the data together is

due to the effect in the second part of the experiment.

Table 2.- Conditions of the treatments with added PAHs in the different experiments. Average of2 or 3 replicates ±Stdev.

Experiment Date Initial PAHs Conc Exponential decay rate

(µg l-1) (d-1)

Spring 02/03/05 14.95 ± 5.59 -0.370 ± 0.024

Summer 02/07/05 29.03 ± 6.08 -0.714 ± 0.015

Autum 22/09/05 23.63 ± 6.11 -0.415 ± 0.045

Winter 24/01/06 18.51 ± 4.68 -0.326 ± 0.014

Microcosms (Winter):

0.5x 9.60 ± 0.99 -0.231 ± 0.009

1x 19.81 ± 0.42 -0.241 ± 0.001

2x 38.85 ± 0.64 -0.234 ± 0.003

4x 73.21 ± 2.97 -0.212 ± 0.009

73

Chapter II

Fig 2.- Changes in dissolved PAHs concentration and in bacterial abundances in the control mesocosms (white circles) and the mesocosms to which PAHs had been added (black circles). Spring experiment (A), Summer experiment (B), Autumn experiment (C) and Winter experiment (D). Average plus standard deviation of 2 (spring) or 3 (the other months) replicated mesocosms.

0

1 10 6

2 10 6

3 10 6

4 10 6

5 10 6

0

20

40

60

80

100

120

0 2 4 6 8

ControlPAHsPAHs concentration

A

0

1 10 6

2 10 6

3 10 6

4 10 6

5 10 6

0

20

40

60

80

100

120

0 2 4 6 8

C

Time (days)

0

1 10 6

2 10 6

3 10 6

4 10 6

5 10 6

0

20

40

60

80

100

120

0 2 4 6 8

BSpring Summer

0

1 10 6

2 10 6

3 10 6

4 10 6

5 10 6

0

20

40

60

80

100

120

0 2 4 6 8

D

Time (days)

Autum Winter

% of

PAHs

rem

ainin

g%

of PA

Hs re

main

ing

Bacte

rial a

bund

ance

(cell

ml-1

)Ba

cteria

l abu

ndan

ce (c

ell m

l-1)

74

Effects of oil on bacterial activity and diversity

Table 3.- Results of the repeated measures ANOVA with one within-subject factor (time) and treatment as a between-subjects factor. N = 2-3 for the mesocosms experiments, and N = 4 for the microcosm experiment. Signi cant (P< 0.05) effects are in bold.

Treatment Time (treatment) F-value Prob. F-value Prob.

Chlorophyll a

Spring 0.43 0.525 68.04 <0.01Summer 9.81 <0.01 4.09 <0.01Autumn 0.11 0.733 45.55 <0.01Winter 3.21 0.084 15.63 <0.01Microcosms 130.85 <0.01 79.49 <0.01

Total prokaryotic abundance

Spring 16.10 <0.01 85.07 <0.01 rst half 1.18 0.31 second half 32.91 <0.01Summer 35.08 <0.01 7.59 <0.01 rst half 1.54 0.24 second half 37.29 <0.01Autumn 5.95 0.021 4.67 <0.01 rst half 1.34 0.29 second half 60.89 <0.01Winter 20.31 <0.01 32.60 <0.01 rst half 2.25 0.15 second half 8.88 0.02Microcosms 727.79 <0.01 94.81 <0.01

Prokaryotic heterotrophic production

Spring 4.71 0.47 55.65 <0.01Summer 43.54 <0.01 5.44 <0.01Autumn 2.35 0.136 9.43 <0.01Winter 0.215 0.646 2.73 0.014Microcosms 325.44 <0.01 87.15 <0.01

Bacterial growth rate

Spring 0.56 0.47 103.09 <0.01Summer 0.92 0.36 6.55 <0.01Autumn 3.98 0.06 5.21 <0.01Winter 5.46 0.03 12.82 <0.01Microcosms 4.39 <0.01 18.53 <0.01

75

Chapter II

Fig. 3.- Changes in dissolved PAHs concentration and in bacterial production (leucine incorporation) in the control mesocosms (white circles) and the mesocosms to which PAHs had been added (black circles). Spring experiment (A), Summer experiment (B), Autumn experiment (C) and Winter experiment (D). Average plus standard deviation of 2 (spring) or 3 (the other months) replicated mesocosms.

0

20

40

60

80

100

120

140

160

0

20

40

60

80

100

120

4 6 80

20

40

60

80

100

120

140

160

0

20

40

60

80

100

120

0 2 4 6 8

AControl

PAHs

PAHs concentration

0

20

40

60

80

100

120

140

160

0

20

40

60

80

100

120

0 2 4 6 8

Bacte

rial p

rodu

ction

(mg

C L-1

d-1)

Bacte

rial p

rodu

ction

(mg

C L-

1 d-

1)

0

20

40

60

80

100

120

140

160

0

20

40

60

80

100

120

0 2 4 6 8

0 2

% o

f PAH

s rem

ainin

g%

of P

AHs r

emain

ing

B

C D

Spring Summer

Autum Winter

Times (days)Times (days)

76

Effects of oil on bacterial activity and diversity

Fig. 4.- Evolution of heterotrophic nano agellate abundance in the control (empty symbols) and the oil-added treatment (solid symbols) in the Spring experiment. Averages of 2 replicated mesocosms ± Stdev. Bacterial abundances in the same control (empty squares) and treatment (solid squares) mesocosms are shown for reference (same data of Fig. 2A).

Table 4.- Heterotrophic nano agellate evolution in the rst three mesocosms with experiments and microcosm. Changes in abundance (growth rates) before and after the abundance peak reached in each experiment. Data are not available for the Winter experiment and for 8x treatment of microcosm.

INITIAL FINAL

Time interval (d) GR (d-1) Time interval (d) GR (d-1)

Spring 0-4 0.48 >4 -0.23Summer 0-2 0.71 >2 -0.59Autumn 0-2 0.40 >2 -0.76Winter - - - -

Microcosms 0x 0-3 0.3 >3 0.02 2x 0-3 -0.04 >3 0.06 4x 0-3 -0.07 >3 0.18

0

5 105

1 106

1.5 106

2 106

2.5 106

0 2 4 6 8

Time (days)

0

1000

2000

3000

4000

5000

Hete

rotro

phic

nano

flage

llate

s (c

ells

ml-1

)Bacteria (cells m

l -1)

77

Chapter II

Bacterial production. Thymidine and leucine incorporations were well correlated throughout the

experiments (TdR = 51.9*Leu^0.69, r2= 0.56, N=165), with an average ratio Leu:TdR of 30.4 ± 2.8.

The ratio was lower (average 12.8 ± 1.2) in summer, and larger in spring. It was relatively stable

throughout the experiment in summer and winter, but increased throughout the spring experiment

(46.0 ± 8.9), and decreased in the autumn experiment. Using repeated measures ANOVA the only

signi cant differences between the ratio Leu:TdR in PAHs treated vs. control mesocosms was observed

in summer (details not shown), when it was signi cantly lower in the mesocosms that had received

PAHs (9.7 ± 0.7) than in those that had not (13.7 ± 0.9).

Bacterial production was much higher in the autumn experiment than in the rest of the experiments

(Fig. 3), in concordance with the fact that this experiment coincided with the end of a phytoplankton

bloom that slowly decomposed during our mesocosm experiment (Fig. 1C). In general, bacterial

production tended to reproduce the same pattern as bacterial abundances (Fig. 2), with an initial peak

and a secondary peak after day 4. This was evident in the spring and winter experiments (Fig. 3A

and 3D), less obvious in the summer experiment, and different in the autumn experiment (Fig. 3C).

Differences between treatments were only signi cant in the summer experiment (Fig. 3B and Table 3),

when the mesocosms that had received PAHs had productions about 1.5 fold higher than the controls.

The autumn experiment showed a small tendency for production being higher in the mesocosms that

had received PAHs (Fig 3).

Bacterial growth rates showed a pattern similar to that of production, with initial values ranging

from 0.3 d-1 (spring) to 3 d-1 (summer) (details not shown). They were not signi cantly affected by

the oil addition, except in the microcosm experiment (see below), where bacterial growth rates at the

end of the experiment in the control treatment were twice those of the treatment that had received the

maximum amount of oil. There was also a slight effect of oil on bacterial growth rates in the winter

experiment, where bacterial growth rates were slightly faster in the oil-treatments than in the controls

(ca. 25% decrease in growth rate) between day 3 and day 8.

Heterotrophic nano agellates. Flagellates were enumerated in the rst three experiments. The

evolution was rather similar: they rst increased up to day 3-4, and then decreased until the end of

the experiment (see example in Fig. 4, growth rates for the two different phases in Table 4). The

peak of HNF abundance was reached at different time in the different seasons, at day 4 in the spring

experiment (Fig. 4), at day 2 in the summer and autumn experiments (Table 4). At the peak of agellate

development, the abundance was higher in the mesocosms that had received PAHs than in the controls

(see example for the spring experiment in Fig. 4).

78

Effects of oil on bacterial activity and diversity

Fig. 5.- Time course evolution in the microcosm experiments: PAH concentration (A), Chlorophyl a(B), Bacterial abundance (C) and Prokaryotic production (D). A gradient of four concentrations represented: control (white circle), 20 μg L-1 (black circle), 40 μgL-1 (white square) and 80 μg L-1 (black square). For simplicity, we omit the treatment with ca. 10 µg L-1.

Bacte

rail

prod

uctio

n (m

g C L

-1 h-1

)

0

5 10 5

1 10 6

1,5 10 6

2 10 6

0 1 2 3 4 5 6 7

Bacte

rial a

bund

ance

(cell

s ml-1

)

0

20

40

60

80

100

120

0 1 2 3 4 5 6 7

Control20 µg L-1

40 µg L-1

80 µg L-1

Time (days) Time (days)

Chla

(mg m

-3)

% PA

Hs re

main

ing

A B

C D

0

50

100

150

200

250

0 1 2 3 4 5 6 7

0

0.5

1

1.5

2

0 1 2 3 4 5 6 7

79

Tabl

e 5.

-C

ompo

sitio

n of

the

bact

eria

l ass

embl

age

in th

e m

icro

cosm

exp

erim

ent d

eter

min

ed b

y C

AR

DF

ISH

: C

ells

pos

itive

with

pro

be A

LF96

8 fo

r A

lpha

prot

eoba

cter

ia (

Alp

ha),

pro

be G

AM

42a

for

Gam

map

rote

obac

teri

a (G

amm

a) a

nd p

robe

CF

319a

for

Bac

tero

idet

es (

CF

B).

Mea

sure

s at

th

e be

ginn

ing

(t0)

and

at t

he e

nd (

tf) o

f the

mic

roco

sm e

xper

imen

t in

the

four

diff

eren

t con

ditio

ns:

cont

rol (

0x),

20

mg

L-1 e

q. c

hrys

ene

(1x)

, 40

mg

L-1 (

2x)

and

80 m

gL-1(4

x). T

he “

Cha

nge

in a

bund

ance

” is

the

num

ber

of c

ells

pro

duce

d (o

r lo

st)

duri

ng th

e ex

peri

men

t.

A

lpha

prot

eoba

cter

ia

Gam

map

rote

obac

teri

a B

acte

roid

etes

t0

tf

C

hang

e in

abu

nd.

t0

tf

Cha

nge

in a

bund

. t0

tf

C

hang

e in

abu

nd.

0x

1.48

105

3.31

104

-115

,000

± 5

,300

9.

25 1

04 0.

76 1

05 -1

6,00

0 ±

800

2.50

105

0.18

105

-232

,000

± 1

4,00

01x

1.

32 1

05 0.

68 1

04 -1

26,0

00 ±

6,0

00

5.09

104

1.97

105

146,

000 ±

13,0

00

1.83

105

0.51

105

-132

,000

± 5

,700

2x

0.6

105

1.70

104

-59,

000 ±

1,70

0 4.

07 1

04 1.

29 1

05 88

,000

± 6

,000

1.

68 1

05 0.

64 1

05 -1

04,0

00 ±

2,8

004x

1.

63 1

05 6.

36 1

04 -9

9,00

0 ±

5,10

0 8.

14 1

04 6.

87 1

05 60

6,00

0 ±

44,0

00

1.83

105

1.

11 1

05 -7

19,0

00 ±

2,7

00

80

Effects of oil on bacterial activity and diversity

Fig 6.- Examples of DGGE gels of bacterial 16S rRNA gene fragments from the Spring (A), Summer (B) and microcosm (C) experiments and dendogram classi cation (Ward’s Method, Euclidean distances according to the band pattern). DGGE lanes represent the different treatments: control mesocosms (C1 and C2) and PAHs-amended mesocosms (A1 and A2) and the samplings times in days (T0, T2, T4, T6 and T8). In the microcosm experiment the different PAHs concentrations (control, 20 μg L-1, 40 μg L-1 and 80 μg L-1) and the samplings days (T0, T3 and T7). Different grey bands reunite together samples that cluster together.

0A1C1C1C2C1A1C1A1A1C1 T2

T6ToToT8T8T6T6T4T2

C1 A1 C1 A1 C1 A1 C1 C2 A1 A2 C1 C2 A1 A2

To T2 T4 T6 T8

0 5 10 15 20 25 30T

AC1A1

2

C2 T8T8T8T6T6T6T6T6T4T4T2T2

A1 A2 C1 A1 C2 A1 A2 C1 C2 A2 C1 C2

A T2T2

C1A1 T4

T4

2 T6T6T6T6T6

C2 T8T8T8

A

B

0 5 10 15 20 25 30

20 �g/L

80 �g/L

40 �g/L

40 �g/L20 �g/L

control

Linkage Distance

ToToToT3T3T3ToT3T7T7T7T7

control

20 �g/L

80 �g/L

control

40 �g/L80 �g/L

20 �g/L T7T7T7

40 �g/L80 �g/Llortnoc 02

�Lg

1-

04�

Lg1-

08�

Lg1- lo rtno c 0 2

�Lg

1-

04�

Lg1-

08�

Lg1- lor tnoc 0 2

�Lg

1-

04�

Lg1-

08�

Lg1-

To T3 T7

A

2C

CCA

A

2

1

1

1A2C

21

C

Summer

Minicosmos

Spring

Linkage Distance

Time (days)

Time (days)

0 50 100 150 200 250

C1 T0A1 T0A1 T2C1 T2C2 T6C1 T8C1 T4C2 T8C1 T6A1 T4A2 T6A1 T8A1 T6A2 T8

T2 T4 T6 T8

A1 T4A2 T6A1 T8A1 T6A2 T8

C1 T0A1 T0

Linkage Distance

81

Chapter II

Microcosm experiments. As we have shown before (Teira et al. 2008), the evolution of bacterial

communities was similar in the mesocosms and the microcosms that had received the same oil additions.

PAHs, however, disappeared more slowly from the microcosms (Fig. 5A) than from the mesocosms

(Fig 1D), probably because the water surface in contact with the atmosphere was smaller in the case of

the microcosms. PAHs had an effect on chlorophyll a (Fig. 5B, Table 3), which was much stronger at

80 µg L-1. Bacterial abundance was also affected by the added PAHs, so that the decrease in the peak

occurred later the more PAHs had been added, and did not occur when the maximal concentration of

80 µg L-1 was added (Fig. 5C). Bacterial production did not show any signi cant differences between

treatments until time 4, when the 80 µg L-1 sample increase tremendously (Fig. 5D). HNF developed

much more in the control of the microcosms (up to 3500 cells mL-1) than in the PAH treatments,

which were all rather similar and had few changes in HNF concentration (ca. 1500 mL-1 with a slight

increase towards the end of the experiment, details no shown).

Bacterial community structure and estimates of diversity. Figure 6 shows two examples

of the ngerprints of bacterial community structure. In three cases (spring, autumn and winter) the

band patterns showed no differences between controls and PAHs-added mesoscosms, as exempli ed

by the March DGGE gel and corresponding dendrogram (Fig. 6A). Only in the case of the summer

experiment there was a detectable oil effect (Fig. 6B), the dendrogram showing a cluster of samples

labeled “A” (PAH addition) after day 4 onwards. The presence and absence of bands in the DGGE

analysis was used to estimate bacterial richness of each sample. Band number varied between 7 and

17 with a global average of 13. Differences in bacterial richness were not signi cant for any of the

experiments, except for the summer one, where we detected higher diversity in the PAHs-ammended

mesocosms, particularly in the rst part of the experiment.

The DGGE analysis and the associated dendrogram of the microcosm experiment showed a clear

clustering for the highest concentration treatments, particularly at the start of the experiment (times 0

and 3, Fig. 6C). The samples from time 7 all clustered together, indicating that the time evolution was

probably more important than the PAHs additions in determining community structure. The lowest

richness was detected with the highest PAHs concentration.

In the microcosm experiments we also followed the changes in bacterial subgroup abundance by

CARDFISH. Alphaproteobacteria abundances had similar decreases (1x105 cells mL-1) in all treatments

except in the 2x, when it was lower (5x104 cells mL-1). Bacteroidetes showed a clearly higher decrease

in the tank that was amended with higher concentration of PAHs (Table 5). Somehow different results

were observed when comparing the relative contribution of each subgroup to total DAPI-stained

82

Effects of oil on bacterial activity and diversity

cells abundance instead of the absolute abundance (Teira et al. 2007). The most relevant changes

occurred, however, also in the Gammaproteobacteria, which slightly decreased in the treatment that

received no PAHs, but increased a lot in the treatment that had received most PAHs. Community

structure in the microcosms that had received PAHs shifted from ca. 50% Bacteroidetes, 35%

Alphaproteobacteria and 15% Gammaproteobacteria to 20% Bacteroidetes, 10% Alphaproteobacteria

and 73% Gammaproteobacteria. Community structure also shifted in the microcosms that had not

received PAHs, but the dominance by Gammaproteobacteria was smaller (55%) and driven by the

large decrease in Bacteroidetes (from 50 to 14%). Gammaproteobacteria in the control microcosm

changed little, and Alphaproteobacteria lost a large number of cells, but continued to be a relatively

large part of the assemblage, ca. 26%.

DISCUSSION

We used a mesocosms approach to test for the effects of realistic additions of oil on bacterial

function and assemblage structure. We repeated the experiment four times at different phases of the

seasonal cycle of Ría de Vigo to discern whether the assayed levels of PAHs affected microbial

communities at all times or only depending on the initial microbial communities or environmental

settings. Furthermore we also run a variable-concentration experiment to nd the threshold generating

signi cant effects. The four experiments started in four different plankton conditions. Spring with a

starting phytoplankton bloom, summer with little development of algae, autumn corresponded to a

declining algal bloom, and winter had low chlorophyll values. The fact that the algal communities

maintained constant populations in the mesocosms in summer and winter indicates that our sampling

and bag- lling protocols were accurate and did not disturb too much the communities.

Following the accident of the Prestige tanker in front of the Galician coast in November 2002,

several research projects were carried out to determine the effects of the spill on the ecosystem. Our

objective was to test to what extend the bacterial assemblages of Ría de Vigo were pre-adapted to the

oil concentrations likely to be found in an oil spill accident. We hypothesized that if the assemblages

did not change in diversity nor in function after point additions of PAHs, it would indicate that the

assemblages were accustomed to the presence of similar PAHs concentrations in the waters, and that

they could use and metabolize this type of allochthonous carbon, which would then not dramatically

modify the structure and the function of the bacterial community.

83

Chapter II

The concentrations of PAHs measured after similar oil spill accidents are very variable (as are in fact

the methodologies used to measure the concentration). After the accident in Paraíso Bay, Antarctica,

values of 50-100 µg L-1 were measured, while the values measured after the Exxon Valdez spill were

much lower, of ca. 10-30 ng L-1 (González et al. 2006). The values measured in the area of the Prestige

oil spill a few weeks after the accident were on average 0.3 µg L-1 (González et al. 2006).

It is somehow dif cult to compare different units of PAHs concentrations reported before because

not all the units are interchangeable. In any case, the values used in our experiment tend to be much

lower than those used in the past in experimental studies. Siron et al. (1993), for example, used a

range of oil concentrations of 0.7-4.4 mg L-1 in mesocosms.Some authors used very large crude oil

concentrations 900 mg L-1, Cappello et al. 2006; 100 mg L-1, Cappello et al. 2007) because they

targeted the isolation of hydrocarbon-degrading bacteria, or wanted to detect the speci c use of one

PAH molecule (naphtalene, 6.4 g L-1, Castle et al. 2006). It is certain that when crude oil is added,

the real concentration of PAHs that the microorganisms experience will be much smaller. In our

experiments we chose values (Table 2) intermediate between what was detected in the sea, and what

had been used in previous similar studies, concentrations that we believed are more realistically close

to what in situ bacteria could experience. The fast disappearance of PAHs from the water (Table 1)

further assured that PAH concentrations in the water in our experiments were realistic.

Our results indicate that the concentrations of PAHs tested in the mesocosms appeared to affect

bacterial abundance in all experiments while, however, chlorophyll, bacterial production and growth

rates were only affected signi cantly during the summer experiment (Table 2). This contrasting result

(effects on abundance and not on the rest of the variables) can be explained by the fact that the

effects on bacterial abundance can be seen in the second part of the experiment, after agellates

have grazed on the community and have, presumably, affected its composition (see below). That the

effects are signi cative in summer can be assigned to the fact that in the period with less available

inorganic nutrients (summer, Table 1), the increased mineralization activity by bacteria stimulated

by the introduced PAHs probably generated more inorganic nutrients, which in turn stimulated algal

growth, while in winter the excess of nutrients in the water would mask the increase caused by

increased bacterioplankton mineralization. It is interesting to note that we found the highest effect in

the season in which PAHs disappeared faster from the water (0.70 d-1 in summer and 0.35 d-1 in the

other seasons) probably due to the higher temperatures (Table 1). Temperature also plays a signi cant

role in controlling the nature and extent of microbial hydrocarbon metabolism (Nedwell et al. 1999),

and directly affects both the rate of biodegradation (Brakstad et al. 2006), and the physico-chemical

behaviour of oil hydrocarbons, such as viscosity, diffusion and volatilization, which changes oil

84

Effects of oil on bacterial activity and diversity

composition, and bioavailability of the water-soluble components (Northcott & Jones, 2000, Rowland

et al. 2000). Coulon et al. (2007) showed that the change in temperature had a much more pronounced

effect on the oil-degrading microbial communities than nutrient additions, even though biodegradation

was possible at all temperatures.

The summer season in the NW Iberian zone is characterized by the upwelling due to the dominance

of North winds over the adjacent shelf. Surface water leaves the Ría moving offshore and this water

is replaced by subsurface oceanic Eastern North Atlantic Central Water, which enters the Ría and has

a profound effect on its hydrography (e.g. Álvarez-Salgado et al. 1993). The bacterial assemblages

at this time of the year, previously surviving in deep oceanic waters, are probably less adapted to the

amounts of allochthonous organic matter present in the Ría. In fact, we detected using CARD-FISH the

presence of some Betaproteobacteria (typically considered “freshwater” groups, Methé et al. 1998) at

all seasons except Summer, and the gammaproteobacteria PAH-degrading genus Cycloclasticus was

detected in September and January, but not in summer (Alonso-Gutiérrez et al. submitted). SAR11

was also not detected in summer, something which is consistent with a deep-water origin of the

bacterial assemblage (SAR11 decreases at mesopelagic depths, e.g. Baltar et al. 2007).

The microcosm experiments run in parallell to the January mesocosms demonstrate that the lack

of effect of the added PAH concentrations in the spring, autumn and winter experiments was most

probably due to a too low PAH concentration being assayed. In that experiment, ca. 20 µg L-1 of PAHs

had no effect, while twice this amount started to be detectable, and 4x this amount had a very relevant

effect on the microbial communities, decreasing algal development, increasing bacterial abundance

and production, and decreasing bacterial richness. The PAHs addition affected positively bacterial

abundance and production both in the summer experiment and in the microcosm experiments (Fig. 2

and Fig. 5), but had contrasting effects on chlorophyll development, which was stimulated in summer,

and depressed in the microcosms. Previous studies reported a stimulatory effect of oil addition to

autotrophs but a toxic effect at higher concentrations, while bacterial response was however positive

at all concentrations (Wang & Koshikawa 2007).

Inspection of the DGGE gels let us conclude that there were signi cant changes in bacterial

community structure caused by the PAH additions only in the summer experiment where an increase

in the number of bands was observed (an estimate of higher richness). This result differs from the

previous idea that hydrocarbon polluted areas should have low diversity values and a high dominance

of a few bacterial groups (Harayama et al. 1999, MacNaughton et al. 1999, Kasai et al. 2001, Röling

et al. 2002), and also differs from Schäfer et al. (2000) that reported that experimental manipulations

85

Chapter II

should reduce the diversity of microbial communities. Since the experimentally added PAHs can be

considered as a C source additional to other C sources present in the water, it could be speculated that

increased richness would indicate the appearance of more niches for bacterial development when the

PAHs were present. Interestingly, in the microcosm experiments we observed an increase in richness

with the additions of 20 and 40 µg L-1 as compared to the control, but then a decrease with the highest

concentration (80 µg L-1). The global structure of the community changed little, however, as detected

by DGGE (Fig. 6C), and it seemed to do so more at the end of the experiment, when the 80 µg L-1

treatment was more distinct from the control treatment. This is in contrast with the CARD FISH results

(Table 5) that indicate a clear shift towards gamma-proteobacteria in the 4x (80 µg L-1) treatment,

also apparent, but much lower in magnitude, in the 1x and 2x treatments. This discrepancy is maybe

due to the fact that the high addition treatment presents much higher abundances respect to the other

treatments. Bacteroidetes and alphaproteobacteria decreased in all treatments, but with contrasting

patterns: alphaproteobacteria decreased more at low than at high oil additions, while Bacteroidetes

were replaced by gammaproteobacteria particularly at the highest PAH addition. These results at

times seem to contradict a previous analysis (Teira et al. 2008) which was based on %contribution to

each group and not in strict abundance. The differences can be assigned to a different variable being

observed.

In previous research, gammaproteobacteria have generally been seen to dominate after oil

additions (e.g. Grossman et al. 1990, Kasai et al. 2001, Syutsubo et al. 2001, Röling et al. 2002,

McKew et al. 2007, Cappello et al. 2007, etc.). For example, Alcanivorax were shown to dominate

in oil-contaminated seawater when nutrients were adequately supplied (Kasai et al. 2002) and to

have broad substrate speci city for alkane degradation (Hara et al. 2003). One of the predominant

oil degraders after the Nakhodka oil spill was the aromatic hydrocarbon decomposer Cycloclasticus

pugetii (23-25%), followed by the aliphatic hydrocarbon decomposer Alcanivorax borkumensis (4-

7%) (Maruyama et al. 2003). Since the type of oil we used was rich in PAHs and poor in aliphatics,

we expected to see the development of Cycloclasticus. Indeed, an increase of Cycloclasticus bacteria

in the four mesocosms experiments, reaching different levels depending on the season, was reported

by Teira et al. (2007) and in the microcosm experiments this organism showed a proportional increase

depending on the amount of added PAHs. The Cycloclasticus genus is well known to be implicated in

PAH degradation (Kasai et al. 2002, Yakimov et al. 2004). In the microcosm experiments this group

was a large fraction of the total gammaproteobacteria specially in the middle of the experiment. As

an example, 3, 12, 28, 43 and 89% of the cells positive with the gamma- probe (GAM42a) were also

postive with probe CYPU829 for Cycloclasticus in the 0x, 0.5x, 1x, 2x and 4x treatments respectively

at day 5. Later on in the experiment, Cycloclasticus disappeared from the microcosms (Teira et al.

2007) and were replaced by other gammaprotebacteria. It is important to note, however, that probe

86

Effects of oil on bacterial activity and diversity

GAM42a has a broad gammaproteobacteria group coverage of only 76% (Amann & Fuchs 2008),

and does not target Alcanivorax, nor many other typically hydrocarbon-degrading bacteria, such

Cycloclasticus, Oleispira, Thalassolituus, etc. It targets, however, other gammaproteobacteria with

capability to degrade hydrocarbons, such as Pseudomonas or Marinobacter hydrocarbonoclasticus

(unpublished results).

For the analyses of bacterial community structure in this experiments we used the DGGE approach,

as other studies have done in oil polluted marine environments. (Macnaughton et al. 1999, Ogino et al.

2001, Kasai et al. 2001, Castle et al. 2006) Bacterial richness (as estimated by DGGE) or community

structure was not affected in the experiments other than in the summer one. Observation of the DGGE

band patterns (Fig. 6) and the statistical analyses reported in Table 3 indicates that “Time” is one of the

most important factors determining the changes in bacterial community structure in the experiments,

and also is one of the main factors determining the changes in chlorophyll, prokaryote abundance and

prokaryotic heterotrophic production (Table 3). Since we showed that the oil addition had little effect

on bacterial diversity and function, we could speculate that besides the environmental factors (nutrient

levels, DOC content, water, temperature, etc.) protist development could be one of the reasons why

bacterial function and community structure varied in the different experiments. We expected that

changes in bacterial community structure (as detected by DGGE) should be correlated to the changes

in bacterial activity.

Indeed, the similarity between bacterial community structure in two consecutive time points was

correlated to the total bacterial activity between these two points (Fig. 7A) so that the lower bacterial

production, the closer are the two communities, and vice-versa. The relationship seemed to be different

in the summer experiment than in the others, again supporting the idea that the summer community

developed differently from the bacterial communities of the other experiments. The correlations were

signi cant but not great. To improve them, we analyzed separately the rst part of the experiments

(Fig. 7B), when HNF developed (see example in Fig. 5) with positive growth rates (Table 4) from

the second part of the experiment, when HNF declined (Fig. 4) and presented negative growth rates

(Table 4). The changes in bacterial community structure were well linked to bacterial production in

the rst part of the experiment, when HNF probably induced relatively small changes in the bacterial

assemblage. Later on, once HNF had developed and depleted the bacterial population (Fig. 4), bacterial

community structure changes were not linked any more to bacterial production.

As in many other published mesocosms experiments (e.g. Lebaron et al. 1999, Pinhassi et al.

2006, Allers et al. 2007) bacterial abundance in our study had a small initial positive response to the

87

Chapter II

mesocosms startup (days 1-2), then they decreased in abundances (days 3-4), and then they increased

again towards the end of the experiments. It was this second bacterial peak the one that was most

affected by oil.

There is still disagreement as to whether the abundance, productivity and diversity of pelagic

bacteria are determined mainly by predators or nutrients (Pernthaler 2005). Grazing has been shown

to in uence bacterial community composition in laboratory and eld experiments (van Hannen et

al. 1999, Gasol et al. 2002, Simek et al. 2002). Allers et al. (2007), in a mesocosm experiment with

Mediterranean waters found a succession of bacterial communities. The initial abundance peak consisted

of predominantly Alteromonadaceae, very similar phylotypes in the different treatments. These

organisms were depleted by an increase in HNF numbers and afterwards different Rhodobacteraceae

phylotypes developed in the different treatments, particularly those that had received P additions. If

Fig. 7.- A) Relationships between integrated bacterial production between samples (X axis) and the similarity in bacterial community structure (Y, Pearson correlation coef cients between two consecutive lanes in a DGGE analysis, considering presence/absence of bands as well as intensity), with separation of experiments run in Spring, Autumn and Winter (solid circles) from the Summer experiment (empty circles). A similitude of 1 indicates that the community did not change. For the Summer experiment, N=8, r=-0.71, P<0.05, for the other experiments, N=22, r=-0.58, P<0.005. B) Same plot for the three experiments of Spring, Autumn and Winter, separating the points before the time= 4 (solid squares) and after that time point (empty squares). This division corresponding to changes before the peak of agellate development, or after that peak (see Fig. 5). See text for further explanations. The relationship is only signi cant for the period before agellate development, N=10, r=-0.94, P<0.005. Regression lines are drawn for orientation only

Bacterial production between two time points (µgC l-1 d-1)

Spring, Autumn and Winter experimentsSummer experiment

Sim

ilarit

y of

bac

teria

l com

mun

ity s

truct

ure

A-0.2

0

0.2

0.4

0.6

0.8

1

0 50 100 150 200 250-0.2

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100 120 140

B

Bacterial production between two time points (µgC l-1 d-1)

From time=0 to time=4From time=4 to time=8

88

Effects of oil on bacterial activity and diversity

the bacterial assemblages of coastal waters are composed of rather stable populations with peaks of

fast-growing organisms showing rapid temporal uctuations, manipulation of the ecosystem might

destabilize this system and promote the growth of these fast-growing populations, that are the ones

that will be responsible for the changes in bacterial production and community structure (Pernthaler

2005).

Conclusion

Policyclic aromatic hydrocarbons (PAHs), which are well-known to have toxic effects on

organisms also are, however, a signi cant allochthonous C source for bacteria. Our microcosm study

showed that the addition of PAHSs resulted in an immediate change in bacterial community structure

and rapid rate of oil degradation, suggesting the presence of a pre-adapted oil-degrading microbial

community assemblage, in agreement with the data presented by Coulon et al. (2007). Our mesocosms

experiments showed that, at least in three of the seasons, the natural assemblage seldom reacted to the

PAHs additions, did not show any toxic effect, and did not produce any bacterial stimulation, maybe

because the concentration was negligible compared to that of the total biovailable DOC. In the Ría de

Vigo, we had observed that PAH addition triggered short-lived peaks of Cycloclasticus (Teira et al.

2007) that used the PAHs, and then were virally-lysed or predated by protists. The organisms with the

potential to use the allochthonous added substances were present, albeit at very low concentrations,

and had the potential to develop immediately after PAHs additions. Our results also indicate that when

realistic (i.e. not exagerately large) oil or PAHs concentrations are added in experimental containers,

the response of the communities will be more likely to shed light into the realistic behaviour of natural

microbial communities.

ACKNOWLEDGEMENTS

This work was supported by project IMPRESION (VEM2003-20021). We thank all the people

involved, particularly those who helped with the preparation and sampling of the mesocosms during

the four experiments, and F.G. Figueiras and X.A. Álvarez-Salgado who were the driving forces

behind the whole project. We also thank J. González (Universidad de Vigo) for access to data and

D. Vaqué for help with the HNF counts. We would also like to thank the captain and crew on board

RV Mytilus. Financial support to IL and ACD was provided by PhD fellowships from the Spanish

government (MEC).

89

Chapter II

REFERENCESAllers E, Gómez-Consarnau L, Pinhassi J, Gasol JM, Simek K & Pernthaler J (2007) Response of

Alteromonadaceae and Rodobacteriaceae to glucose and phosphorus manipulation in marine mesocosms. Environ Microbiol 9: 2417-2429.

Álvarez-Salgado XA, Rosón G, Pérez FF, & Pazos Y (1993). Hydrographic variability ofF the Rias Baixas (NW Spain) during the upwelling season. J. Geophy. Res., 98: 14447-14455

Alzaga R, Montuori P, Ortiz L, Bayona JM, & Albaigés J (2004) Fast solid-phase extraction gas chromatography mass spectrometry procedure for oil ngerprinting. Application to Prestige oil spill. J Chromat A 1025: 133–138

Amann R, Binder BJ, Olson RJ, Chisholm SW, Devereux R & Stahl DA (1990) Combination of 16S rRNA-targeted oligonucleotide probes with ow cytometry for analyzing mixed microbial populations. Appl Environ Microbiol56: 1919-1925.

Alonso-Gutiérrez J, Lekunberri I, Teira E, Gasol JM, Figueras A & Novoa B. Strong dominance of Roseobacter and detection of novel SAR86 members in a Coastal upwelling ecosystem (Ría de Vigo, NW Spain). Appl Environ Microbiol (submited.)

Blanton JO, Tenore KR, Castillejo F, Atkinson LP, Schwing FB & Lavin A (1987) The relationship of upwelling to mussel production in the rias of the Western Coast of Spain. J Mar Res45: 497-511

Brakstad OG & Bonaumet K. (2006) Biodegradation of petroleum hydrocaarbons in seawater at low temperatures (0-5 ºC) and bacterial communities associated with degradation. Biodegradation17: 71-82.

Bode A, González N, Lorenzo J, Valencia J, Varela M, & Varela M (2006) Enhanced bacterioplankton activity after the ´Prestige´oil spill off Galicia, NW Spain. Aquat Microb Ecol43: 33-41.

Bjørnsen PK & Kuparinen J (1991) Determination of bacterioplankton biomass, net production and growth ef ciency in the Southern Ocean. Mar Ecol Prog Ser 71: 185-194

Calvo-Díaz A & Morán XAG (2006) Seasonal dynamics of picoplankton in shelf waters of the southern Bay of Biscay. Aquat Microb Ecology 42: 159-174.

Cappello S, Denaro, R, Genoves M, Giuliano L & Yakimov MM (2007) Predominant growth of Alcanivorax during experiments on “oil spill bioremediation” in mesocosms. Microbiol Res 162: 185-190

Cappello S, Caruso G, Zampino D, Monticelli LS, Maimone G, Denaro R, Tripodo, Troussellier m, Yakimov M & Giuliano L (2006) Microbial community dynamics during assays of harbour oil spill bioremediation: a microscale simulation study. J Appl Microbiol 102: 184-194

Casamayor EO, Massana R, Benlloch S, Overas L, Diez B, Goddard VJ, Gasol JM, Joint I, Rodríguez-Valera F & Pedrós-Alió C (2002) Changes in archaeal, bacterial and eukaryal assemblages along a salinity gradient by comparison of genetic ngerprinting methods in a multipond solar saltern. Environ Microbiol 4: 338-348

90

Effects of oil on bacterial activity and diversity

Castle DM, Montgomery MT, & Kirchman DL (2006) Effects of naphthalene on microbial community composition in the Delaware stuary. FEMS Microb Ecol 56: 55-63

Coulon F, McKew BA, Osborn MA, McGenity TJ & Timmis KN (2007) Effects of temperature and biostimulation on oil-degrading mmicrobial communities in temperatre estuarine waters. Environ Microbiol 9: 177-186.

Dalby AP, Konstantinos AK, Christaki U & Karayanni H (2008) Cosmopolitan heterotrophic microeukaryotes are active bacterial grazers in experimental oil-polluted systems. Environ Microbiol 10: 7-56

Delille B & Siron R (1993) Effect of dispersed oil on heterotrophic bacterial communities in cold marine waters. Microb Ecol 25: 263-273.

Figueiras FG, Labarta U & Reiriz MJF (2002) Coastal upwelling, primary production and mussel growth in the Rias Baixas of Galicia. Hydrobiologia 484: 121-131

Fuhrman J A & F Azam (1982) Thymidine incorporation as a measure of heterotrophic bacterioplankton production in marine surface waters: Evaluation and eld results. Mar Biol 66: 109-12

Garcia EM, Siegert IG & Suarez P (1998) Toxicity assays and naphtahalene utlization by natural bacteria selected in marine environments. Bull Environ Contam Toxicol 61: 0370-0377

Gasol JM, Pedrós-Alió C, & Vaqué D (2002) Regulation of bacterial assemblages in oligotrophic plankton systems: results from experimental and empirical approaches. Ant. van Leeuwenhoek81: 435-452.

Gasol JM & del Giorgio PA (2000) Using ow cytometry for counting natural planktonic bacteria and understanding the structure of planktonic bacterial communities. Scientia Marina. 64: 197-224

González JJ, Viñas L, Franco MA, Fumega J, Soriano JA, Grueiro G, Muniategui S, López-Mahía P, Prada D, Bayona JM, Alzaga R & Albaigés J (2006) Spatial and temporal distribution of dissolved/dispersed aromatic hydrocarbons in seawater in the area affected by the Prestige oil spill. Mar Pollut Bull 53: 250-259.

Glöckner FO, Fuchs BM & Amann R (1999) Bacterioplankton composition of lakes and oceans: a rst comparison based on uorescence in situ hybridization. Appl Environ Microbiol 65: 3721-

3726.

Gerdts G, Wichels A, Döpke H, Kings K, Gunkel W & Schütt C (2004) 40-year long-term study of microbial parameters near Helgoland(Germand Bight, North Sea): historical view and future perspectives. Helgol Mar Res 58: 230-242.

Hara A, Syutxubo K & Harayama S (2003) Alcanivorax which prevails in oil-contaminated seawater exhibits broad substrate speci city for alkane degradation. Environ Microbiol 5: 746-753

Harayama S, Kishira H, Kasai Y & Shutsubo K (1999) Petroleum biodegradation inmarine environments. J Mol Microbiol Biotechnol 1: 63-70

91

Chapter II

Kasai Y, Kishira, H & Harayama, S (2002) Bacteria Belonging to the Genus Cycloclasticus Play a Primary Role in the Degradation of Aromatic Hydrocarbons Released in a Marine Environment. Appl Environ Microbiol 68: 5625-5633.

Kasai Y, Kishira H, Syutsubo K & Harayama S (2001) Molecular detection of marine bacterial populations on beaches contaminated by the Nakhodka tanker oil-spill accident. Environ Microbiol 3: 246-255.

Kirchman DL, K´nees E, & Hodson R (1985) Leucine incorporation and its potential as a measure of protein synthesis by bacteria innatural aquatic ecosystems. Appl Enviro Microbiol 49: 1681-1688

Koshikawa H, Xu DQ, Liu AL, Kohata K, Kawachi M, Maki H, Zhu MY & Watanabe M (2007) Effect of the water-soluble fraction of diesel oil on bacterial and primary production an dthe trophic transfer to mesozooplankton through a microbial food web in Yangtze estuary, China. Est, Coast Shelf Sci 71: 68-80

Lebaron P, Servais P, Troussellier M, Courties C, Vives-Rego J & Muyzer G (1999) Changes in bacterial community structure in seawater mesocosms differing in their nutrient satatus. Aquatic Microb Ecol 19: 255-267.

Maruyama A, Ishiwata H, Kitamura K, Sunamura M, Fujita T, Matsuo M & Higashihara T (2003) Dynamics of Microbial Populations and Strong Selction for Cycloclasticus pugetii following the Nakhodka Oil Spill. Microb Ecol 46: 442-453

Macnaughton SJ, Stephen JR, Venosa AD, Davis GA, Chang Y & White DC (1999) Microbial Population Changes during Bioremediation of an Experimental Oil Spill. Appl Environ Microbiol65:3 566-3574.

Manz W, Amann R, Ludwig, W, Wagner M & Schleifer KH (1992) Phylogenetic oligonucleotide probes for the major subclasses or Proteobacteria: problems and solutions. Syst Appl Microbiol15: 593-600.

Manz W, Amann R, Ludwig W, Vancanneyt M & Schleifer KH (1996) Application of a suite of 16S r RNA-speci c oligonucleotide probes designed to investigat bacteria of the phylum Cytophaga-Flavobacter-Bacteroides in the natural environment. Microbiology-UK 142: 1097-1106.

McKew BA, Coulon F, Yakimov M, Denaro R, Genovese M, Smith CJ, Osborn MA, Timmis KN & McGenity TJ. (2007) Ef cacy of intervention strategies for bioremediation of crude oil in marine systems and effects on indigenous dydrocarbonoclstic bacteria. Environ Microbiol. 9: 165-17

Methé BA, Hiorns WD & Zehr JP (1998) Contrasts between marine and freshwater bacterial communities in Lake George and six other Adirondack lakes. Limn Oceanogr 43: 368-374

Morán XAG, Estrada M, Gasol JM & Pedrós-Alió C (2002) Dissolved primary production and the strength of phytoplankton-bacterioplankton coupling in contrasting marine regions Microb Ecol44: 217-223

Moran M A, Sheldon WM & Sheldon JE (1999) Biodegradation of riverine dissolved organic carbon in ve estuaries of the Southeastern United States. Estuaries 22: 55–64.

92

Effects of oil on bacterial activity and diversity

Nayar S, GohB PL & Chou LM (2005) Environmental impacts of diessel fuel on bacteria and phytoplankton in a tropical estuary assessed using in situ mesocsms. Ecotoxicology 14: 397-412

Nedwell DB (1999) Effect of low temperature on microbial growth: lowered af nity for substrates limits growth at low temperature, FEMS Microbiol. Ecol. 30: 101–111

Nogueira E, Ibáñez F, & Figueras FG (1997) Seasonal patterns and long-term trends in an estuarine upwelling ecosystem (Ría de Vigo, NW Spain). Est Coast Shelf Sci 44: 285-300.

Northcott & Jones GL (2000) Experimtntal approaches andanalytical techniques for determining organic compound bound residues in soil and sediment. Environ Poll 108: 19-43.

Ogino A, Koshikawa H, Njahara T & Uchiyama H (2001) Succession of microbial communities during a biostimulation process as evaluated by DGGE and clone library analyses. J Appl Microbiol 91: 625-635

Ohwada K, Nishimura M, Wada M, Nomura H, Shibata A, Okamoto K, Toyoda K, Yoshida A, Takada H & Yamada M (2003) Study of the effect of water-soluble fractions of heavy-oil on coastal marine organism using enclosed ecosystems, mesocosms. Mar Pollut Bull 47: 78-84

Pernthaler A, Pernthaler J, Eilers H, & Amann R (2001) Growth patterns of two marine isolates:adaptations to substrate patchiness. Appl Environ Microbiol 67: 4077-4083

Pinhassi J, Gómez-Consarnau L, Alonso-Sáez L, Sala M, Vidal M, Pedros-Alió C, Gasol JM (2006) Seasonal changes in bacterioplankton nutrient limitaion and their effects on bacterial community composition in the NW Mediterranean Sea Aquat Microb Ecol 44: 241-252

Porter KG & Feig YS (1980) The use of DAPI for identi cation and enumartion of bacteria and blu-green algae. Limn Oceanogr 25: 943-948

Preston BL. (2002) Spatial patterns in benthic biodiversity of Chesapeake Bay,USA (1984-99): association with water quality and sediment toxicity. Environ Tox Chem 21: 151-162.

Rowland AP, Lindley DK, Hall MJ, Rossall MJ, Wilson DR & Bernham DG (2000) Effects of beach sand properties, temperature and rainfall aon the degradation rates of oil in buried oil/beach sand mixtures. Environ Poll 109: 109-118.

Röling WFM, Milner MG, Jones DM, Lee K, Daniel F, Swannell RJP & Head IM (2002) Robust hydrocarbon degradation and dynamics of bacterial communities during nutrient-enhanced oil spill bioremediation. Appl Environ Microbiol 68: 5537-5548.

Sánchez O, Gasol JM, Massana R, Mas J, & Pedrós-Alió C (2007) Comparison of different Denaturing Gradient Gel Electrophoresis primer sets for the study of marine bacterioplankton communities. Applied Environmental Microbiology 73: 5962-5962.

Santas R, Korda A, Tenente A, Buchholz K & Santas PH. (1999) Mesocosm assays of oil spill bioremediation with oleophilic fertilizers: Inipol, F1 or Both? Mar Pollut Bull 38: 44-48.

Sargian P, Mostajir B, Chatila K, Ferreyra GA, Pelletier E & Demers S (2005) Non-synergistic effects of water-soluble crude oil and enhanced ultraviolet-B radiation on a naturual plankton assemblage. Mar Ecol Prog Ser 294: 63-77

93

Chapter II

Schauer M, Massana R, & Pedrós-Alió C (2000) Spatial differences in bacterioplankton composition along the Catalan coast (NW Mediterranean) assessed by molecular ngerprinting. FEMS Microbiol Ecol. 33: 51-59

Schauer M, Balagué V, Pedrós-Alió C, & Massana R (2003) Seasonal changes in the taxonomic composition of bacterioplankton in a coastal oligotrophic system. Aquat Microb Ecol. 31: 163-174

Sieracki ME, Haugen EM & Cucci TL (1995) Overestimation of heterotrophic bacteria in the Sargasso Se: direct evidence by ow and imaging cytometry. Deep-Sea Res 42: 1399-1409.

Simek K, Nedoma J, Pernthaler A, Posch T & Dolan JR (2002) Altering the balance between bacterial production and protistan bacterivory triggers shifts in freshwater bacterial community composition. Antonie van Leewenhoek Int J Gen Mol Microb 81: 453-463.

Siron R, Pelletier E, Delille D & Roy S (1993) Fate and effects of dispersed crude oil under icy conditions simulated in mesocosms. Mar Environ Res 35: 273-302

Smith DC & Azam F (1992) A simple, economical method for measuring bacteria protein synthesis rates in seawter using 3H-leucine. Mar Microb Food Webs 6: 107-114

Teira E, Lekunberri I, Gasol JM, Nieto-Cid M, Álvarez-Salgado XA & Figueiras F (2007) Dynamics of the hydrocarbon-degrading Cycloclasticus bacteria during mesocosm-simulated oil spills. Environ Microb 9: 165-176

Teira E, Gasol JM, Aranguren-Gassis M, Fernández A, González J, Lekunberri I & Álvarez-Salgado XA (2008) Linkages between bacterioplankton community composition, heterotrophic carbnon cycling and envrionmental conditions in a highly dynamic coastal ecosystem. Environ Microb10: 906-917.

UNESCO (IOC) 1984. Manual for monitoring oil and disslved/dispersed hydrocarbons in marine waters and on beaches. Intergovernamental Oceanographic Commission, Manuals and GuidesNo13, 35 p.

van Hannen EJ, Zwart G, van Agterveld MP, Gons HJ, Ebert J & Lananbroek HJ (1999) Changes in bacterial and eukaryotic community structure after mass lysis of lamentous Cyanobacteria associated with viruses. Appl Environl Microbiol 65: 795-801.

Varela M, Bode A, Lorenzo J, Álvarez-Ossorio MT, Miranda A, Patrocinio T, Anadón R, Biseca L, Rodríguez N, Valdés L, Cabal J, Urrutia Á, García-Soto C, Rodríguez M, Álvarez-Salgado XA,& Groom S ( 2006). The effect of the “Prestige” oil spill on the plankton of the N-NW Spanish coast. Mar. Pol. Bull. 53: 272-286g

Yakimov M, Gentile G, Bruni V, Cappello S, D´Auria G, Golyshin P & Giuliano L (2004) Crude oil-induced structural shift of coastal bacterial communities of rod bay (Terra Nova Bay, Ross Sea, Antarctica) and characterization of cultured cold-adapted hydrocarbonoclastic bacteria. FEMS Microb Ecol 49: 419-432.

Different types of phytoplankton blooms caused by different additions in seawater mesocosms, and their effect on bacterial heterotrophic activity and assemblage composition

Itziar Lekunberri, Thomas Lefort,Clara Ruíz-González, Montse Coll, Clara Cardelús, Cèlia Marrasé and Josep M. Gasol

CHAPTER 3

Il est d’un petit esprit, de vouloir ne s’être jamais trompé

Louis XIV

97

Chapter III

ABSTRACT

Bacterial community structure is thought to be in part related to the type of phytoplankton species

dominating during algal blooms. The proliferation of speci c organisms in such blooms is due to solar

radiation and the presence of speci c nutrients. We performed a mesocosms experiment to generate

different phytoplankton blooms through the additions of: silica (SiO4), urea and/or phosphorous to

oligotrophic water from the Blanes Bay Microbial Observatory (NW Mediterranean). Over 10 days

incubations, bacterial activity, abundance, nutrient composition and bacterial community structure

were analyzed. The aims of this experiment were: i) to relate different additions to the phytoplankton

and bacterioplankton community composition. ii) to investigate whether the changes in bacterial

community structure occur by a bottom-up or top-down control. Our results reveal that the change in

the bacterial composition is due to a combination of changes in the phytoplankton communities and

grazing pressure. And the differences in abundances and activity of bacteria are due to a shift in the

bacterial composition.

98

Bacterial response to different types of phytoplankton

INTRODUCTION

Phytoplankton blooms consist in a marked increase in algal populations, reaching concentrations of

up to millions of cells per milliliter, involving typically only one or a few species. These phenomena

occur when light and nutrients are available in excess. Cullen et al. (2002) de ned 3 types of blooms:

upwelling blooms which are induced by deep (nutrient-rich) waters reaching the lit layers of the

ocean, entrainment blooms those that appear when the mixed layer deepens due to wind erosion

or solar heating reduction, and detraiment blooms, which occur when near-surface strati cation is

induced in a deep mixed layer of nutrient-rich water. In coastal waters algal blooms are frequently the

result of nutrients inputs (Rosenberg et al. 1990, Smayda 1990) or of the start of strati cation. And

these inputs, even if seasonally recurrent, can be equated with sporadic inputs. Variation in nutrient

regimes affects phytoplankton production, diversity and succession, and also in uences the overall

structure of the microbial food web (Smayda & Reynolds 2001, 2003, Cullen et al. 2002).

Previous studies have correlated increases in biomass of individual phytoplankton species with

nutrient inputs to coastal regions (Smayda 1990, Hallegraeff 1993). Diatom growth has been reported

to be dependent on silica availability (e.g. Allen et al. 2005), and urea was found to be preferentially

be used as a nitrogen source by some cyanobacteria and dino agellates (Glibert et al. 2008), thus

potentially selecting for these species. Other algae are adapted to low nutrient concentrations, and

these species bloom after opportunistic species have done it and exhausted nutrients. For example,

blooms of the eukaryotic picoplankter Aureococcus anophagefferens occur mainly after nitrate and

ammonium have decreased below detection limits (Keller & Rice 1989, Smayda & Villareal 1989).

Primary producers release DOC that is consumed by bacteria (Hagström et al. 1979, Fuhrman &

Azam 1980) and comparative studies show a dependence of bacterial abundances on Chl a suggesting

a control on bacterial abundance derived from the availability of resources derived from primary

production (Bird & Kalff 1984, Cole et al. 1988, Gasol & Duarte 2000). Thingstad & Lignell (1997)

enumerated the 5 factors which in uence growth of heterotrophic bacteria: i) DOC, ii) inorganic

phosphate, iii) organic/inorganic nitrogen, iv) protozoan predation and v) lysis by viruses. Control

by resources of bacterial abundance and activity is often named “bottom-up control” and control by

predation or lysis is called “top-down”.

99

Chapter III

While many studies have shown that the speci city of a phytoplankton blooms is a consequence of

the speci c nutrient added, few have related this selection to shifts in bacterial community structure

(but see Schäfer et al. 2001, Pinhassi et al. 2006). Laboratory enrichment experiments show a

phytoplankton response followed by changes in bacterioplankton community composition (Pinhassi

et al. 2006, Allers et al. 2007) for example diatom blooms tended to be followed by Flavobacteria

developments. Besides Sapp et al. (2007a) found strong shifts in the associated bacterial communities

(members of the Gammaproteobacteria) to diatom species. However, Sapp et al (2007b) reported that

it could not be generalized whether there were bacterial communities strictly microalgal species-

speci c. Similarly, Garcés et al. (2007) found little correspondance between the type of dino agellate

species developing in blooms and their associated bacterial communities.

Different mesocosms experiments have been used to study the effect of nutrient additions: globally

on the phytoplankton communities (Duarte et al. 2000), on the activity and diversity of bacterial

assemblages (Schäfer et al. 2001), and on microbial succession after nutrient additions (Allers

et al. 2007). In this study we conducted a mesocosms experiment in which we tried to generate

three different bloom situations, in order to determine the effect of the different additions on the

phytoplankton, and of different phytoplankton on bacterioplankton community structure and function

during the summer season in Blanes Bay (NW Mediterranean). Experimental studies have always

shown a general phosphorous limitations in the NW Mediterranean ( Thinsgstad et al. 1998, Pinhassi

et al. 2006), particularly during the summer (e.g. Pinhassi et al. 2006). To avoid any nutrient limitation

phosphorous was added to all the tanks except for the controls. The three enriched treatments included

addition of silica, urea, and P alone. We monitored phytoplankton diversity daily and at the same

time sampled for bacterial abundance, physiological status, activity and diversity were collected. In

addition we also monitored the development of the HNF community to explore whether bottom-up or

top-down control led to bacterial population changes.

MATERIALS & METHODS

Experimental setting. The mesocoms experiment was carried out with water collected from Blanes

Bay (The Blanes Bay Microbial Observatory, NW Mediterranean). Eight 100 L tanks (2 replicates, 4

treatments) were lled with seawater ltered by 150 µm, in duplicates with different additions, and

incubated at in situ temperature (19ºC) with a 12 h photoperiod to simulate natural conditions. The

rst sample was taken 1 h after nutrient addition. The experiment was conducted during 10 days from

the 5 th to 15 th June 2007.

100

Bacterial response to different types of phytoplankton

Additions. Within the 8 tanks, 2 were left as controls with no nutrient additions (Ka, Kb) and 6

received phosphate in the same proportion (1 µM nal concentration). Of these 6, 4 were enriched

with 16 µM N nal concentration, 2 with inorganic N (NO3) and 2 with organic N as urea (CO(NH

2)

2).

The NO3 amended tanks also received SiO

4. To keep the silicate in excess we added silicate twice

(in day 1 and day 2), and obtained a concentration (23-22 µM) similar to a target of 25 µM (thus

generating an addition with the ratio P:N:Si of 1:16:25, higher than the standard Red eld’s (1:16:16

to assure that the Si was in excess).To all tanks we added a metal solution in the same proportion to

phosphate as in the f/2 medium (Guillard et al. 1975).

Microplankton enumeration. For the identi cation of microphytoplankton (mainly diatoms,

dino agelates and coccolithophorids), samples were xed with formalin-hexamine solution (0.4%

nal concentration). Counts were made with the Utermöhl et al. (1958) methodology using 50 cm3

settling chambers. One transect of the chamber was observed at 400x magni cation to count the

smaller (<20 µm) and more frequent organisms. Additionally, 1 transect of half of the chamber was

inspected at 200x to enumerate cells of intermediate size (roughly 20 and 50 µm), and the whole

chamber was scanned at 200x to count the larger forms.

Pico- and nanoplankton abundances. Determination of algal and bacterial abundance was

performed with ow cytometry (Marie et al. 1999, Gasol & del Giorgio 2000). Samples were run in

a Becton-Dickinson FACSCalibur ow cytometer with a laser emiting at 488 nm. For phytoplankton,

the samples were not xed and were run at high speed (ca. 100 µl min-1), Four populations

(Prochlorococcus, Synechococcus, picoeukaryotes and nanoeukaryotes) were differentiated according

to size and red and orange uorescence. For non-phototrophic bacteria, 1.2 ml samples were xed with

a 1% paraformaldehyde + 0.05% glutaraldehyde solution and deep-frozen in liquid N2. Afterwards,

the samples were unfrozen, stained with SybrGreen at 1:10,000 dilution and run at low speed (ca 10

µl min-1). Cells were identi ed in a plot of side scatter vs green uorescence (FL1).

Heterotrophic nano agellate abundances were measured following the Rose et al. (2004) protocol.

From a stock solution of 1 mM Lysotracker Green (Molecular Probes) 1 µl was added to 99 µl of <

0.2 µm MiliQ, and 3.8 µl of the diluted Lysotracker were added to 0.5 ml of the sample, generating

a 75 nM Lysotracker nal concentration. We analyzed the samples as in Rose et al. (2004), using

a combination of side scatter and green and red uorescence. The samples were run alive, at high

(ca. 100 µl min-1) speed. Concentrations were obtained from weighting measurement of the volume

analyzed.

101

Chapter III

Bacterial single-cell activity. Measurements of the physiological status of bacteria were done in

two ways: i) Highly respiring prokaryotes, as those able to reduce 5-cyano-2,3-ditiolyl tetrazolium

chloride (CTC; Polysciences), CTC turns into a red uorescent formazan that is detectable by

epi uorescence and ow cytometry (Sherr et al. 1999, Sieracki et al. 1999). Sample aliquots (0.4

ml) were amended with 5 mM CTC (from a fresh stock solution at 50 mM) immediately following

collection and were incubated for 90 min in the dark at room temperature. CTC-positive (CTC+) cells

were enumerated by ow cytometry using the FL2-versus-FL3 dot plot (see Gasol & Arístegui 2007).

For these analyses, we used a high speed (ca. 100 µl ml-1) and a threshold set in red uorescence.

ii) Cells with intact membranes were enumerated using the NADS viability protocol, based on the

combination of the cell-permanent nucleic acid strain SybrGreen I (Molecular Probes, Eugene, OR)

and the cell-impermeant propidium iodine (PI, Sigma Chemical Co.) uorescent probe. We used a

1:10,000 SG1 and 10 µg ml-1 PI concentrations. After simultaneous addition of each stain, the samples

were incubated for 20 min in the dark at room temperature and then analyzed by ow cytometry.

SG1 and PI uorescence were detected in the green (FL1) and orange-red (FL3) cytometric channels,

respectively. A dot plot of red versus green uorescence allowed distinction of the “live” cell cluster

(i.e., cells with intact membranes and DNA present) from the “dead” cell one (i.e., with compromised

membranes) (Gregori et al. 2003, Falcioni et al. 2008).

Bacterial production. Bacterial heterotrophic production was estimated using the 3H-leucine

incorporation method (Kirchman et al. 1985). Quadruplicate aliquots of 1.2 ml and 2 TCA-killed

controls were taken daily. The samples were incubated with 40 nM nal concentration for about 2 h in

the dark and at in situ temperature. The incorporation was stopped with the addition of 120 μl of cold

TCA 50% to each replicate and the samples were kept frozen at –20ºC until processing, which was

carried out by the centrifugation method described by of Smith & Azam (1992).

Collection of community DNA. Microbial biomass was collected by sequential ltration of about

5 l of seawater through a 3 µm pore size polycarbonate lter (Millipore, 46 mm) and a 0.2 µm

sterivex lter unit (Millipore). Microbial biomass was treated with lysozyme, proteinase K and sodium

dodecyl sulphate, and the nucleic acids were extracted with phenol and concentrated in a Centricon-

100 (Milipore). The nucleic acids were extracted by a standard protocol using phenol/chloroform

(Schauer et al. 2003).

102

Bacterial response to different types of phytoplankton

Fingerprinting analysis. DGGE and gel analysis were performed essentially as previously described

(Schauer et al. 2000). Brie y, 16S rRNA gene fragments (around 550 bp in length) were ampli ed by

PCR, using the universal primer 907rm and the bacterial speci c primer 358f, with a GC-clamp. The

PCR products were loaded on a 6% polyacrylamide gel with a DNA-denaturant gradient ranging from

40-80%. The gel was run at 100V for 16 h at 60ºC in 1x TAE running buffer. DGGE gel images were

analyzed using the Diversity Database software (BIO-RAD).

Statistical analysis. A matrix was constructed for all DGGE lanes taking into account the relative

contribution of each band (in percentage) to the total intensity of the lane. Based on this matrix, we

obtained a dendrogram by the Ward’s clustering method (Euclidean distances, Statistica 6.0), and

ordinations of nonmetric multidimensional scaling based on the Bray-Curtis similarity index (MDS,

Kruskal & Wish 1979, Clarke & Green 1988). A one-way ANOSIM was computed to test the observe

differences between treatments and the control samples, and another to contrast differences among

time within treatment samples (software Primer v5). We also used one-way ANOVAs to test for the

differences in Chla, bacterial abundance and bacterial activity between treatments at the different

times when samples were taken for bacterial diversity.

RESULTS

Evolution of inorganic nutrient concentrations. Variations of the inorganic nutrient

concentrations for the different treatments during the experiment are shown in Fig.1. Except from

the control, all tanks had received 1 µM phosphate, which was rapidly consumed in the N amended

treatments (Si and U) but remained quite high in the P tank. The N source in the Si tank was NO3,

which started to decrease after day 6 (Fig 1C). In the representation of the evolution of silica (Fig

1B) the maximal value in treatment Si was only reached on day 2 (25 µM, after the second nutrient

addition), and this concentration was only reduced to ca. 10 µM at the end of the experiment, thus

indicating that there was much leftover unused Si in this treatment. Si remained at the initial value

of ca 4 µM in the control treatment, but it was used in the Urea treatment after day 3, and in the P

treatment after day 5, down to values < 0.1 µM.

103

Table 1.- Concentrations of nutrients (µM) in the 4 different experimental conditions 1 h after the addition. aand b are the replicates.

Experimental Nomenclature PO4 NO

3 SiO

4 DON

conditions

Initial water 0.02 0.13 0.21 -

Control Ka, Kb 0.02 ± 0.00 0.34 ± 0.03 0.38 ± 0.01 4.48 ± 0.41Phosphorous Pa, Pb 0.67 ± 0.03 0.40 ± 0.05 0.34 ± 0.03 6.92 ± 1.05Urea+P Ua, Ub 0.79 ± 0.01 0.031 ± 0.01 0.31 ± 0.00 16.70 ± 1.09N+P+SiO

3 Sia, Sib 0.75 ± 0.01 12.5 ± 1.03 22.7 ± 0.79* 17.08 ± 0.70

*The Silicate values correspond to day 2 (after the second addtion of the SiO3, see text)

Fig.1.- Nutrient evolution throughout the mesocosms experiment. Phosphate (A), Silicate (B) and nitrate+nitrite (C). Treatments: K tank ( ), P tank ( ), Si tank ( ) and U tank ( ). The arrow indicates the second addition of silicate.

0

0.2

0.4

0.6

0.8

1

0 2 4 6 8 10Time (days)

KaKbPaPbSiaSibUaUb

PO43+

(μM

)

SiO 4-2

(μM

)

NO3- +

NO2(μ

M)

Time (days)

0.01

0.1

1

10

100

0 2 4 6 8 10

C

Time (days)

0.001

0.01

0.1

1

10

100

0 2 4 6 8 10

BA

104

Bacterial response to different types of phytoplankton

Time (days)

Leuc

ine i

ncor

pora

tion

(pm

ol L

eu L

-1h-1

)

HNF

(cell

s ml-1

)BA

(cell

s ml-1

)

Chlo

roph

yll (

mg

m-3)

Time (days)

100

1000

104

0 2 4 6 8 100

1000

2000

3000

4000

0 2 4 6 8 10

0.1

1

10

100

0 2 4 6 8 10

AKaKbPaPbSiaSibUaUb

6 105

8 105

106

3 106

5 106

7 106

0 2 4 6 8 10

C D

B

Dynamics of chlorophyll a and phytoplankton composition. The initial value of chlorophyll a

(Chl a) was around 0.2 µg L-1 in all tanks (Fig. 2A). Signi cant differences (<0.05) between treatments

were found after time 3. The Si tank reached concentrations about 16 fold higher than the control and

the P tank while the U tanks reached values ca 5-fold those of the K and P tanks. Chl a in all treatments

remained high until the end of the experiment. Maximum value of Chl a in the P and K tanks was

reached at day 2, in the U tank at day 4, and in the Si tank at day 6.

Fig. 2.- Chlorophyll a concentration (A), bacterial abundance (B), leucine incorporation rates (C) and heterotrophic nano agelate abundance (D) throughout the mesocosms experiment. Treatments: K tank ( ), P tank ( ), Si tank ( ) and U tank ( ).Points used for the statistical analyses are indicated with arrows.

105

Chapter III

0

20

40

60

80

100

2 4 6 8 10

A%Dinoflagellates%Diatoms%Cocolitoforales%NanoflagellatesOthers

0

20

40

60

80

100

2 4 6 8 10

B

0

20

40

60

80

100

2 4 6 8 10

C

0

20

40

60

80

100

2 4 6 8 10

D

0 0

Time (days) Time (days)00

% V

olum

e (�m

3 )%

Vol

ume (

�m3 )

CONTROL P

Si U

The evolution of phytoplankton composition along the experiment in terms of contribution to

biovolume is shown in Fig. 3. The K and P tanks present similar patterns, with alternation between

diatoms and dino agellates. The Si tank showed a large increase in the diatoms comprising ca. 90%

of the community. In contrast the U tank had a decrease in the diatom composition compensated by

an increased in the dino agellates from time 4 to the end.

Fig. 3.- Phytoplankton diversity as % of volume. K tank (A), P tank (B), Si tank (C) and U tank (D). Dino agellates ( ), diatoms ( ), cocolitoforids ( ), nano agellates (X) and others ( ).

106

Bacterial response to different types of phytoplankton

On the basis of size and pigment content, we identi ed two different populations of picophytoplankton

(Synechococcus and picoeukaryotes). Prochlorococcus, are not present in Blanes Bay at this time of

the year (details not shown). Synechococcus in the control and P treatments had a peak at day 3 and

then decreased to remain with populations similar to the initial ones. The U and Si treatments had

highest values with a peak at day 5 and then a decrease to the disappearance at the end (Fig. 4A).

Picoeukaryotes had a clear peak at day 4 with maximal values in the Si and U tanks (Fig. 4B).

0

5 104

1 105

1.5 105

2 105

0 2 4 6 8 10

KaKbPaPbSiaSibUaUb

Syne

choc

occu

s (ce

lls m

l-1)

Pico

euka

ryot

es (c

ells m

l-1)

Time (days)

0

5000

1 104

1.5 104

2 104

2.5 104

3 104

3.5 104

0 2 4 6 8 10

B

A

Fig. 4.- Evolution of Synechococcus (A) and Picoeukaryotes (B). Treatments: K tank ( ), P tank ( ), Si tank ( ) and U tank ( ).

107

Chapter III

Bacterial abundance. Initial bacterial cell numbers were 1.2 x 106 cells ml-1 (Fig. 2B). There was

an initial peak at time 2-3 days in all treatments, which was much larger in treatments P, U and Si than

in the control. A second peak at times 6-8 did not occur in the control. This peak was higher in the U

treatment than in the Si and P treatments. There were very signi cant differences (one-way ANOVA

P<0.01, Table 2) between controls and amended tanks at the beginning (time 3) of the experiment

Bacterial production. All the treatments presented an increased at days 2-3 but no signi cant

differences within the amended tanks were found (Fig 2C), at the beginning they were all very different

from the control tank. While the Si treatment maintained a high value al through the experiment, the

P treatment decreased at time 3, and the U treatment at time 7.

Heterotrophic nano agellates. They were determined by Lysotracker staining. The numbers

generally followed the changes in bacterial abundances with a time lag, and HNF reached highest

densities always at least 1 day after the peak of bacterial abundances (Fig. 2D). The rst maxima

occurred at day 3-4, and a second peak at day 9. In the rst peak there were not strong differences

between treatments, with a slightly higher value for treatment Si. The second peak was higher for

treatments Si and U.

Bacterial single-cell activity. All the amended treatments presented a signi cantly (<0.05) higher

number of respiring cells compared to the controls (Fig. 5A, Table 2). The U tanks show the highest

number of respiring cells, followed by the Si and then the P tanks. The %NADS, which is another

indicator of the physiological stage of the bacteria, did not show differences between treatments except

for the controls which decreased from time 4 towards the end of the experiments (Fig. 5B). Differences

between treatments were also observed by measuring the % of high nucleic acid containing bacteria

(Fig. 5C), which showed almost exactly the same pattern as the amount of CTC+ cells.

Bacterial community composition and estimates of diversity. DNA samples were collected

at day 0, 3, 6, 8 and 10. The analysis of the bacterial DGGE ngerprints (Fig. 6A) by the Ward’s

clustering method, (Fig. 6B) and that of the MDS analyses (Fig. 6C). show similar results: the samples

were organized by time rather than by treatment, thus indicating that the internal evolution in the tanks

superseded the effects of the additions. The one-way ANOSIM analysis of similarity showed that the

108

Bacterial response to different types of phytoplankton

105

106

0 2 4 6 8 10

CTC

(cell

s ml-1

)

% L

ive c

ells (

NAD

S)

% H

DNA

Time (days)

Time (days)

Time (days)

KaKbPaPbSiaSibUaUb

20

30

40

50

60

70

80

90

0 2 4 6 8 10

C

A

70

75

80

85

90

95

0 2 4 6 8 10

B

Fig. 5.- Measures of single-cell bacterial activity: CTC-positive cell abundance (A), % of NADS-determined live cells (B) and %HNA (C). Treatments: K tank ( ), P tank ( ), Si tank ( ) and U tank ( ). Points used for the statistical analyses are indicated with arrows.

109

Chapter III

samples from treatments clustered together by timing (R= 0.517, P: 0.003) and were different from

the control samples (R= 0.549, P: 0.002). The Shannon-Weaver diversity index for phytoplankton did

not change along the experiment in any of the treatments except for the Si tank where it decreased

drastically from time 2 towards the end (Fig. 7A). Bacterial diversity (estimated from the DGGE) had

a similar magnitude than phytoplankton (ca. 2 bits), and showed non-signi cant differences between

treatments (Fig. 7B).

Table 2.- Results of one-way ANOVAs done at every time point when DNA samples for the DGGE analysis were taken. N = 2 for all the mesocosm experiments. Signi cant (P< 0.05) effects are in bold. Treatments with the same letters are not signi cantly different from each other.

T3 T6 T8 T10

Chlorophyll a P 0.011 0.002 <0.0001 0.057

U AB B B ASi A A A AP B A C AK B A C A

Bacterial abundance P <0.001 0.032 0.003 NS

U A A A ASi A AB AB AP A AB BC AK B B C A

Bacterial production P 0.004 0.005 0.008 0.012

U A AB B ABSi A A A AP A BC B BK B C B B

CTC + cell abundance P 0.009 0.011 0.001 NS

U A A A ASi A A A AP A AB B AK B B B A

110

Bacterial response to different types of phytoplankton

Tabl

e.3.

-Phy

topl

ankt

onpo

pula

tions

at T

7 (a

t the

pea

k of

the

bloo

m)

In

pare

ntes

is a

re th

e ab

unda

nce

perc

enta

ge o

f eac

h gr

oup.

The

“O

ther

s” g

roup

incl

ude

the

rest

of p

hyto

plan

kton

pop

ulat

ions

whi

ch a

re n

ot r

epre

sent

ed in

the

prev

ious

gro

ups.

In

situ

K

(ce

lls L

-1)

P (c

ells

L-1)

Si (

cells

L-1)

U (

cells

L-1)

Din

oag

lela

tes

2290

(5.

7%)

4350

(2.

8%)

5854

(3.

76%

) 81

41 (

0.04

%)

9861

(3.

64%

)

Dia

tom

s 16

340

(41.

3%)

6670

8 (4

3.8%

) 11

4783

(74

%)

20,0

00,0

00 (

99.5

%)

2017

22(7

4.5%

)

Coc

olith

ofor

ids

3060

(7.

7%)

1031

8 (6

.8%

) 48

52 (

3.1%

) 59

70 (

0.03

%)

5990

(2.

2%)

Nan

oag

ella

tes

1782

0 (4

5%)

7082

1 (4

6.5%

) 29

646

(19%

) 82

380

(0.4

1%)

5215

5(19

.3%

)

Oth

ers

40(0

.1%

) 18

0 (0

.1%

) 36

0 (0

.2%

) 60

(0.

0%)

829

(0.3

%)

111

Chapter III

K P Si U K P Si U U2º K P Si U K P Si U T3 T6 T8 T10T0

Transform: Square rootResemblance: S17 Bray Curtis similarity

Treatment

K

P

Si

U

0

3

3

3

3

6

666

6

88

8

8

10

10

10

10

2D Stress: 0.16

0 50 100 150 200 250 300

Linkage Distance

U T10

Si T10

P T10

U2º T6

U T6

Si T6

P T6

Si T8

U T8

P T8

K T10

U T3

Si T3

K T8

P T3

K T6

K T3

T0

T3

A

B C

Fig. 6.- A) Denaturing Gradient Gel Electrophoresis (DGGE) of bacterial 16S rRNA gene fragments from samples of different treatments and at different times of incubation. Two different gels are presented, and one lane (U t6) is presented in both gels. B) Dendogram classi cation (Ward’s Method, Euclidean distances according to the band pattern) of DNA samples. (C), Nonmetric multidimensional scaling (NMDS) of the bacterial assemblage composition data.

112

Bacterial response to different types of phytoplankton

DISCUSSION

By adding different nutrients to a NW Mediterranean plankton community, we expected to

generate differences in phytoplankton biomass levels and in phytoplankton community structure and

diversity. And, as a consequence, we expected to generate changes in bacterial biomass, in bacterial

function, and in bacterial community structure and diversity. While we did observe the changes in

phytoplankton biomass (Fig. 2) and community structure (Fig. 3), we only observed strong changes

in bacterial function (Fig 2C), while we observed little changes in bacterial abundances (Fig. 2B) and

bacterial diversity (Fig 7). And bacterial community structure was seen to vary mainly following time,

and much less following the different nutrient additions (Fig. 6).

Our study area (NW Mediterranean) is known to be phosphorous-limited mainly during the

summer (Thingstad et al. 1998, Pinhassi et al. 2006). As the experiment was carried out in June, and

to avoid P limitation, all treatments except for the control (K tank) were amended with phosphate

(PO4) at 1 µM nal concentration. In one tank we added only this phosphorous (the P tank) where we

expected the development of picoeukaryotes and small algae. In another treatment urea was added as

an organic nitrogen source (U tank). Our interest in urea addition was because it has been suggested

that urea enrichment should preferentially lead to dominance by Cyanobacteria, picoeukaryotes and

dino agellates rather than diatoms (Gilbert et al. 2008 and references therein). The last treatment

(Si tank) had NO3 additions besides the phosphate addition to ensure nutrient suf ciency, and also

received SiO4 to promote diatoms dominance, as is well known that they need silica for growth (Allen

et al. 2005).

Indeed, diatoms developed in the Si tank, but they also did so in the other treatments specially in

the rst part of the experiments. This is clearly seen from the decreases in Si concentration in the tanks

(Fig 1B). Phosphate was rapidly consumed in the tanks, which had some N sources (urea or NO3) but

remained mostly unused in the P tank probably because of the lack of a N source. This indicates that

a simple nutrient addition was not enough to promote the selection of a speci c algal group as there

was co-limitation and the added nutrient was in most cases not fully used (Fig. 1).

The Chla plot shows the dramatic differences between treatments caused mainly by the Si

treatment, which reached around 14 fold higher concentrations than the rest of treatments. In this case

the sample phytoplankton diversity decreased strongly (Fig. 8). In the U treatment dino agellates

increased towards the end (Fig. 5), and Synechococcus reached the higher values from time 5 to

time 9 compared to the other treatments. In contrast, phytoplankton diversity in the K, U and P tanks

113

Chapter III

did not show much differences throughout the experiment. In summary, we did induce changes on

phytoplankton species by the different additions, but those were particularly clear in the Si treatment,

while the P and U treatments evolved rather similarly, with little clear-cut differences.

In terms of bacterial abundance and bacterial activity measured as bacterial production and CTC+

cells, at time 3 there were signi cant differences (One way ANOVA P<0.005) between the control

tanks respect to the others (Table 2), a further demonstration of bacterioplankton P limitation in the

NW Mediterranean. We believe this strong limitation, well known from other studies (Sala et al. 2002,

Pinhassi et al. 2006, Thingstad et al. 1998 and references therein) is what drove mainly the initial

evolution of the bacterial community, which reached a rather similar peak at times 2-3 (Fig. 2B),

and presented changes in community structure that were very similar between treatments from time

0 to time 3 (Fig. 6). Thus, this suggests that the bacterioplankton rst of all responded to the nutrient

addition they were mostly dependent, i.e. P. Afterwards, HNF developed (Fig 2C) and depleted

bacterial communities which, only after that situation became differentiated, with differences between

treatments in abundance and activity. For example, the differences in CTC + cells were caused by the

U and Si tanks, but in bacterial production only the Si tank differs from the rest.

The DGGE technique has been widely used in comparative microbial ecology to assess the diversity

of microbial communities and their response to changing environments (Labbé et al. 2007, Castle et

al. 2006, Macnaughton et al. 1999). The ordination of the samples by non-metric multidimensional

scaling (NMDS) based on the DGGE gel band pattern agreed with the clustering results by Ward’s

Methods using Euclidean distances. These analyses show that the samples from amended tanks cluster

together due to different incubation times. The ANOSIM analyses determined a quite good clustering

of samples by the factor of time (R= 0.517, P:0.003). However the control samples had an independent

evolution (R=0.549, P:0.002). Thus, in spite of the observed changes in phytoplankton biomass and

diversity, bacteria showed much less changes. We believe, mostly driven by the initial P limitation and

posterior selection by agellate predation.

In fact, the small changes detected in the community structure analysis at time 3 (Fig. 6) corresponded

with some changes in the Shannon diversity index, which were lower in the K and P treatments at

time = 3 (the rst bacterial abundance peak), and then also slightly lower in the U treatment during

the second bacterial peak (times 6-8). Unfortunately, we had no replicates of these measurements, and

no proper statistical tests can be run. But the lower diversity values in the U plot during the second

peak suggest some specialization of bacterial composition, in agreement with Schäfer et al. (2001)

that found that the DGGE patterns indicated that the addition of nutrients not only had a quantitative

114

Bacterial response to different types of phytoplankton

effect on bacterial community dynamics (higher bacterial cell counts and production, Lebaron et al

1999), but also had a qualitative effect on the development of the bacterial composition.

It is interesting to mention here that the maximum Shannon diversity values were seen in the Si

treatment, simultaneous to the minimum phytoplankton diversity values (Fig. 7). Although the DGGE-

based estimates of richness are subject to polemics (Bent et al. 2007, Danovaro et al. 2007) some

authors think they are as good (or as bad) as any other estimate, since accuate richness estimates are

impossible (Pedrós-Alió 2006). There are two implications to this observation: one is that in ecology

it is commonly believed that diversity is a property that applies to all parts of the ecosystem, and that

by measuring diversity of a group of organisms we are estimating diversity of all other groups. Our

data boot work againsts that hypothesis. Secondly, it is perfectly possible that an evolving diatom

bloom, in which the blooming species age, has different associated bacterial communities, depending

on whether the bloom is young and growing, or old and decomposing, each with different bacterial

communities of similar richness.

0

0,5

1

1,5

2

2,5

0 2 4 6 8 10

Bacterial richness

0

0,5

1

1,5

2

2,5

2 4 6 8 10

Phytoplankton richness

0

Shan

non-

Wea

ver I

ndex

Shan

non-

Wea

ver I

ndex

SiKUP

Time (days) Time (days)

A B

Fig.7.- Shanon-Weaver diversity index of the (micro)phytoplankton (A) and bacteria (B) in the different treatments: K tank ( ), P tank ( ), Si tank ( ) and U tank ( )

115

Chapter III

HNF are assumed to be the primary grazers of bacteria. Thus, an important effect of HNF on

bacterial community structure is expected. Lebaron et al. (1999) suggested that nutrients and grazing

play a key role in shaping the morphologic, genotypic and phenotypic composition of bacterial

communities. The pattern of the HNF abundance in our experiment was the same as that of bacterial

abundance pattern but with one day lag (Fig. 2D), and also the bacterial biomass increase during

the second peak was follow by increases of HNF biomass. Since bacterial abundance and activity

parameters showed strong differences between treatments in the second bacterial peak, and even the

Shannon index showed differentiation of the samples in the second peak, grazing pressure seemed to

be important in the shift of the bacterial assemblage.

Many mesocosms experiments have been performed to investigate the effect of nutrient addition

on marine planktonic communities in the Mediterranean Sea (Lebaron et al. 1999, Schäfer et al.

2001, Olsen et al. 2006, Allers et al. 2007) The input of either phosphorus or nitrogen has often been

reported to be signi cant in the Mediterranean, particularly during strati cation periods (Eker-Develi

et al. 2006). Recent studies in the NW Mediterranean have found differentiation in the bacterial

composition due to different nutrient addition (Schäfer et al. 2001 and Pinhassi et al. 2004) and the

effect of grazing on bacterial biomass (Lebaron et al. 1999). Our results concur with Allers et al. (2007)

that a combination of bottom-up and top-down factors controls bacterial succession and growth in

these type of experiments. We observed that the initial P de cit of the bacterial communities strongly

determined the evolution of the communities after nutrient additions. After an initial very similar

bacterial bloom, depleted by HNF grazing activity, however, we observed that the different nutrients

added did modify slightly bacterial community structure and function. The changes observed were

more clear in the Silica treatment, as the addition of all potentially limiting nutrients (Si, P, N) allowed

the development of a clear phytoplankton bloom, with high chlorophyll and low diversity values.

But even during this bloom we observed an evolution of bacterial diversity probably shaped by the

evolution of HNF predators. The bacterial community then, seems to be controlled by the top-down

factors in contrast to the bottom-up factors that control the initial bacterial bloom as the differences

between treatments in bacterial abundances and activity were not detected until after theHNF grazing

pressure forced changes.

ACKNOWLEDGMENTS

This work was supported by the NoE Marine Genomics Europe and MARBEF, and by Spanish

project MODIVUS (CTM2005-04795/MAR). We thank Vanessa Balagué, Irene Forn, and all the

people participating from the Blanes Bay sampling program, particularly Montse Sala, Evaristo

Vázquez-Domínguez and Cristina Rodríguez for various help.

116

Bacterial response to different types of phytoplankton

REFERENCESAllen JT, Brown L, Sander r, Moore CM, Mustard A, Fielding S, Lucas M, Rixen M, Savidge G, Henson S &

Mayor (2005) Diatom carbon export enhanced by silicate upwelling in the northeast Atlantic. Nature 437:

Allers E, Gómez-Consarnau L, Pinhassi J, Gasol JM, Simek K & Pernthaler J (2007) Response of Alteromonadaceae and Rodobacteriaceae to glucose and phosphorus manipulation in marine mesocosms. Environ Microbiol 9: 2417-2429.

Bent SJ, Pierson JD, Forney LJ, Danovaro R (2007) Measuring species richness based on microbial community ngerprints: the emperor has no clothes. Appl and Environ Microb 73:2399-2401

Bird DF & Kalff J (1984) Empirical relationships between bacterial abundance and chlorophyll concentraion in fresh and marine waters. Can. J. Fish Aquat. Sci 41: 1015-1023.

Castle DM, Montgomery MT & Kirchman DL (2006) Effects of napthalene on microbial community composition in the Delaware estuary. FEMS Microb Ecol 56: 55-63.

Clarke KR & Green RH (1988) Statistical design and analysis for a biological effects study. Marine Ecology Progress Series 46: 213-226.

Cole JJ, Finlay S & Pace ML (1988) Bacterial production in fresh and saltwater ecosystems: a cross-systme overview. Mar Ecol Progr Ser 43: 1-10.

Cullen JJ, Franks PJ, Karl DM, Longhurst A (2002) Physical in uences on marine ecosystem dynamics. The Sea 12, eds Allan R. Robinson, James J, McCarthy and Brian J Rothschild. John Wiley & Sons, Inc., New York

Duarte C, Agustí S, Gasol JM, Vaqué D, Vazquez-Dominguez E (2000) Effect of nutrient supply on the biomass structure of planktonic communities: an experimental test on a Mediterranean coastal community. Mar Ecol Prog Ser 206: 87-96

Eker-Develi E, Kideys AE, Tugrul S (2006) Role of Saharan dust on phytoplankton dynamics in thenortheeastern Mediterranean. Mar Ecol Prog Ser 314: 61-75

Falcioni T, Papa S & Gasol JM (2008) Evaluating the ow cytometric Nucleic Acid Double Staining Protocol (NADS) in realistic planktonic bacterial death situations Appl Environ Microb74: 1767-1779.

Fuhrman, JA & Azam F (1980) Bacterioplankton secondary production estimates for coastal waters of British Columbia, Antarctica, and California. Appl Environ Microbiol 39: 1085–1095

Garcés E, Vila M, Reñé, Alonso-Sáez L, Anglès S, Lugliè M, Masó M & Gasol JM (2007) Natural bacterioplankton assemablge composition during blooms of Alexandirum spp.(Dinophyceae) in NW Mediterranean coastal waters. Aquat Microb Ecol 46: 55-70

Gasol, J.M., del Giorgio, P.A. (2000) Using ow cytometry for counting natural planktonic bacteria and understanding the structure of planktonic bacterial communities. Scientia Marina. 64: 197-224

Gasol, J.M., Duarte, C.M. (2000) Comparative analyses in aquatic microbial ecology: How far do they go ? FEMS Microb Ecol 31: 99-106.

Gasol, J.M. & J. Arístegui (2007) Cytometric evidence reconciling the toxicity and usefulness of CTC as a marker of bacterial activity Aquat Microb Ecol 46: 71-83.

Glibert PM et al. (2008) Ocean ure fertilization for carbon credits poses high ecological risks Mar Pollution Bulletin 56: 1049-1056

117

Chapter III

Guillard RRL (1975) Culture of phytoplankton for feeding marine invertebrates. In: Smith W, Chanley MH (eds) Culture of marine invertebrates. Plenum Publishing, New York 26-60.

Gregori G, Denis M, Lefevre D & Romano JC (2003) Viability of heterotrophic bacteria in the Bay of Marseille.C.R.Biol. 326: 739-750

Hagström Å, Larsson U, Hörstedt P, & Normark S (1979) Frequency of dividing cells, a new approach to the determination of bacterial growth rates in aquatic environments. Appl Environ Microbiol 37:805-812.

Hallegraeff (1993) Review of harmful algal blooms and their apparent global increase. Phycologia 32: 79-99

Keller & Rice 1989 Effects of nutrient enrichment on natural populations of the beteewn tide phytoplankton Aureocccus anophagefferens (Chrysophyceae). Phycologia 25: 636-646

Kirchman DL, Knees E & Hodson R (1985) Leucine incorporation and its potential as a measure of protein synthesis by bacteria in natural aquatic systems. Appl Environ Microb 49: 599-607.

Kruskal JB & Wish M (1979) Multidimensional scaling. Wiley, Beverly Hills, CA.

Labbé D, Margesin R, Schinner F, Whyte LG & Greer CW (2007) Comparative phylogenetic analysis of microbial communities in pristine and hydrocarbon-contaminated Alpine soils. FEMS Microb Ecol 59: 466-475.

Lebaron P, Servais P, Troussellier M, Courties C, Vives-Rego J & Muyzer G (1999) Changes in bacterial community structure in seawater mesocosms differing in their nutrient status. Aquat Microb Ecol 19: 255-267.

Macnaughton SJ, Stephen JR, Venosa AD, Davis GA, Chang YJ & White DC (1999) Microbial population changes during bioremediation of an experimental oil spill. Appl Environ Microb65: 3566-3574.

Marie D, Brussard C, Partensky F, & Vaulot D. (1999) Flow cytometric analysis of phytoplankton, bacteria and viruses. In Current Protocols in Cytometry. Robinson, J.P., Darzynkiewicz, Z., Dean, P.N., Orfao, A., Rabinovitch, P.S., Stewart, C.C., . New York, USA: John Wiley and Sons, pp. 11.11.1–11.11.15.

Olsen Y, Agustí S, Andersen T, Duarte CM, Gasol JM, Gismervik I, et al. (2006) A comparative study of responses in planktonic food web structure and funstion in contrasting European coastal waters exposed to expereimental nutrient addition. Limn Oceanogr 51: 488-503.

Pedrós-Alió C (2006) Marine microbial diversity: can it be determined? Trends Microbiol 14: 257-263.

Pinhassi J, Sala M, Havskum H, Peters F, Guadayol O, Malits A & Marrasé C. (2004) Changes in bacterioplankton composition under different phytoplankton regismens. Appl EnvironMicrob 70: 6753-6766

Pinhassi J, Gómez-Consarnau L, Alonso-Sáez L, Sala M, Vidal M, Pedros-Alió C, Gasol JM (2006) Seasonal changes in bacterioplankton nutrient limitaion and their effects on bacterial community composition in the NW Mediterranean Sea. Aquat Microb Ecol 44: 241-252

Red eld AC (1934) in James Johnstone Memorial Volume (ed. Daniel, R.J.) Liverpool University Press 176-192.

Rose JM, Caron DA, Sieracki ME & Poulton N (2004) Counting heterotrophic nanoplanktonic protists in cultures and aquatic communities by ow cytometry. Aquat Microb Ecol 34: 263-277.

Rosenberg R, Elmgren R, Fleischer S, Jonsson P, Persson G & Dahlin H (1990) Marine eutrophication case studies in Sweden. Ambio 19: 102-108.

118

Bacterial response to different types of phytoplankton

Sala M, Peters F, Gasol JM, Pedrós-Alió C, Marrasé C & Vaqué D (2002) Seasonal and spatial variations in the nutrient limitation of bacterioplankton growth in the northwestern Mediterranean. Aquat Microb Ecol27: 47-56.

Sapp M, Wichels A, Gerdts G. (2007a). Impacts of cultivation of marine diatoms on the associated bacterial community, Appl Environ Microb 73: 3117-3120.

Sapp M, Schwaderer AS, Wiltshire KH, Hoppe HG, Gredsts G & Wichels A (2007b). Species-speci c bacterial communities in the phycosphere of miroalgae? Microb Ecol 53:683-699

Schäfer H, Bernard L, Courties C, Lebaron P servais P, Pukall T, Stackebrandt E, Troussellier , Guindulain T, Vives-Rego J, Muyzer (2001) Microbial community dynamics in Mediterranean nutrient-enriched seawater mesocosms: changes in the genetic diversity of bacterial populations. FEMS Microb Ecol 34: 243-253.

Schauer M, Massana R & Pedrós-Alió C (2000) Spatial differences in bacterioplankton composition along the Catalan coast (NW Mediterranean) assessed by molecular ngerprinting. FEMS Microb Ecol 33: 51-59.

Schauer M, Balague V, Pedrós-Alió C & Massana R (2003) Seasonal changes in the taxonomic composition of bacterioplankton in a coastal oligotrophic system. Aquat Microb Ecol 31: 163-174.

Smayda TJ & Villareal TA (1989) The 1985 “brown-tide” and the open phytoplankton niche in Narragansett Bay during summer. In: Cosper, E.M., Bricelj V.M, Carpenter, E.J. (Eds.), Novel Phytoplankton Blooms: Causes and Impacts of Recurring Brown Tides and Other Unusual Blooms. Springer Verlag, Berlin, pp. 159–187.

Smayda TJ (1990) Novel and nisance phytoplankton blooms in the sea: Evidence for a global epidemic. In Granéli, E. et al. (eds), Tox Mar Phyto. Elsevier 29-40

Smayda TJ & Reynolds CS (2001) Community assembly in marine phytoplankton: application of recent models to harmful dino agellate blooms. J. Plank Res 23: 477-461.

Smayda TJ & Reynolds CS (2003) Strategies of marine dino agellated survival and some rules of assembly. Journal of Sea Research 49: 95-106.

Smith DC & Azam F (1992) A simple, economical method for measuring bacterial protein synthesis rates in seawater using 3H-leucine. Mar Microb Food Webs 6: 107-114.

Thigstad TF & Lignell R (1997) Theoretical models for the control of bacterial growth rate, abundance, diversity and carbon demand. Aquat Microb Ecolo 13: 19-27

Thingstad F, Zweifel UL & Rassoulzadegan F (1998) P limitation of heterotrophic bacteria and phytoplankton inb the northwest Mediterranean. Limn Oceanogr 43: 88-94.

Utermöhl H ( 1958) Zur Vervolkommung der quantitativen Phytoplankton-Methodik. Mitt Int Verein Theor Angew. Limnol 9: 1-38

Effects of Saharan Dust deposition on marine microbial abundance, activity and bacterial community structure and ecosystem function

Itziar Lekunberri, Thomas Lefort, Estela Romero, Evaristo Vázquez-Domínguez, Cèlia Marrasé, Francesc Peters, Markus Weinbauer and Josep M Gasol

CHAPTER 4

Nothing in life is to be feared. It is only to be understood

Marie Curie

121

Chapter IV

ABSTRACT

The Mediterranean coast receives large inputs of Saharan dust. Due to its nutrient content, dust has a

potential fertilization effect on this oligotrophic environment. The aim of this paper was to evaluate

the effect of dust depositions on bacterial abundance, activity and community structure, and on the

metabolic balance on the microbial community. In order to assess this goal we performed a 10 day

mesocosms experiment with additions of realistic dust concentration (0.05 g L-1). Due to the known

P limitation of our study area, we included a P amended tank to distinguish the effect of dust alone

from that of the phosphorus contained in the dust. To determine the saturating concentration of dust

we included a higher dust concentration tank (0.5 g L-1). We found that dust increased the initial water

phosphorus concentration in 0.3 μM and the DOC concentration in 14 μM. Dust addition increased

bacterial abundance (1.8 fold) and bacterial production (5 fold) thus creating higher growth rates at

the initial times of the incubation. Primary production and community respiration were stimulated by

dust and by P, but the net result of the addition of low amounts of dust was a heterotrophic system,

while the net result of the high dust and P additions were net autotrophic communities. Bacterial

community structure changed little between the P-addition and the 0.5 g L-1 dust addition, but the

resulting communities were different from the control and from the high-dust addition communities.

122

Effects of Saharan dust on bacterial activity and community structure

INTRODUCTION

Coastal microbial communities are subject to external forcing events, such as the inputs of

allochthonous nutrient from rivers, coastal runoff, submarine groundwater discharge, resuspension

of benthic particles, etc. These allochthonous inputs contain organic and inorganic nutrients, which

affect microbial diversity and function (e.g. Wassmann & Olli 2004, Markaki et al. 2003, Bonnet et

al. 2005, Pulido-Villena et al. 2008). Among these sources of allochthonous nutrients to the coastal

ocean, atmospheric transport is a particularly important type (Herut et al. 1999). Wind-transported

particles can settle by wet deposition associated with rain, and this has been shown to be an effective

mechanism of dust deposition in the western Mediterranean Sea (Prodi & Fea 1979, Guerzoni et al.

1997, Loyë-Pilot & Martin 1996). The Sahara and the Sahel are, by far, the largest source of dust

particles (Prospero et al. 2003): they are responsible for more than half of the world’s mineral dust

emission (Perez et al. 2007) and the Mediterranean Sea, because of its closeness to this area, receives

large inputs of Saharan Dust (Löye-Pilot & Marin 1996, Guerzoni et al. 1999). Most of the SD

(Saharan dust) deposition events in the coastal NW Mediterranean occur during late spring and early

summer (Guerzoni et al. 1999, Ridame & Guieu 2002), when “dust rains” are common.

The Mediterranean Sea, and particularly the NW Mediterranean, are known to have phytoplankton

and bacterial communities which are P-limited, particularly during the summer season (Thingstad

et al. 1998, Pinhassi et al. 2006), and since SD has been seen to contain bioavailable phosphorus

(Herut et al. 2002), SD deposition events can stimulate the activity and the production of P-limited

microorganisms. In particular, primary production has been seen to respond to SD inputs (Baroli et al.

2005, Bonnet et al. 2005, Lenes et al. 2001, Herut et al. 1999, 2002, 2005, Ridame & Guieu 2002).

However also bacterial abundance and activity have been seen to respond to SD enrichments (Herut et

al. 2005, Pulido-Villena et al. 2008). A priori, one could speculate that bacteria should be stimulated

after algae would be stimulated rst by the nutrients contributed in the deposition.

But SD contains nutrients other than P, particularly N and Fe (Jickells et al.1995), and is also

known to be also a source of organic carbon to the aquatic ecosystems (Pulido-Villena et al. 2008).

Heterotrophic bacteria are often the best competitors for phosphorus uptake (Currie 1984), and the

simultaneous addition of DOC and P might shift the competition for P to be won by bacteria (Thingstad

et al. 1999). Thus, it could be hypothesized that the processes driven by heterotrophic bacteria would

be more stimulated by a SD event than the autotrophic processes. SD events would, thus, have the

capacity to shift a system towards net heterotrophy by stimulating bacterial respiration more than

primary production. Bacterial respiration has been seen to be stimulated by P additions (Obernosterer

123

Chapter IV

et al. 2003) and the addition of DOC, presumably of low quality (one can pressume the DOC travelling

in the SD is not very labile), in the presence of enough inorganic P but no other sources of labile DOC,

could force bacteria to respire most of the low-quality C, thus greatly increasing community respiration.

In fact, Pulido-Villena et al. (2008) observed that a natural dust deposition event of 2.6 g m-2 induced a

1.5-fold increase in bacterial abundance and a 2-fold increase in respiration, while experimental dust

additions (between 2 and 20 g m-2) also stimulated abundance and respiration (1.5-3 fold increases).

These authors, however, did not speci cally compare the heterotrophy stimulation caused by the dust

with the simultaneous expected stimulation of primary production (as found by others, e.g. Herut et

al. 2005). Thus, while SD inputs are expected to stimulate both bacterial and phytoplanton function,

it is uncertain whether the input will shift the metabolic balance of the community.

Addition of P and DOC from the SD, but also changes in phytoplankton community structure caused

by the fertilization (e.g. Pinhassi et al. 2004), can affect the structure of the bacterial assemblages.

It is particularly interesting to understand whether it is the P or the DOC component of the SD what

in uences in a larger degree community structure, and whether there are phylotypes particularly

stimulated by the SD pulses. In addition to the inputs of nutrients associated with dust, viable bacteria

can be transported long-distances (Kellogg & Grif n 2006), with unknown effects on the resident

bacterial community structure.

We designed an experiment to i) con rm the stimulation of primary production (PP) and bacterial

function caused by additions of SD to a P-limited oligotrophic plankton community, ii) tell apart the

effects of the fertilization by the P contained in the SD, from the effects of the other components of

the SD, particularly the bioavailable DOC, iii) test whether SD additions would shift the microbial

community to autotrophic or heterotrophic net metabolism, and iv) see whether SD fertilization would

alter bacterial community structure in a way fundamentally different from the changes caused by P

fertilization alone. We added dust to a series of microcosms at an “experimental” concentration of 50

mg l-1 and used two “controls” on top of a no-addition treatment: one with inorganic P alone at the same

concentration that in the SD addition, and one with a much higher SD addition to observe the effects

of SD in a saturation pulse of the same magnitude as found sporadically in the Mediterranean (Loyë-

Pilot & Martin 1996). Furthermore, as the dust depositions in this area are due to storms (Ridame

& Guieu 2002), we added a turbulence treatment to discriminate the effects of the associated water

movement from the effects of SD alone.

124

Effects of Saharan dust on bacterial activity and community structure

MATERIALS & METHODS

Experiment setting. The experiment was carried out with water collected in Blanes Bay, NW

Mediterranean (The Blanes Bay Microbial Observatory, 41°40’0“ N, 2°48’0“ E) the 16 May 2006.

Eight tanks of 7.5 L were used for the experiment lled with seawater ltered by 150 μm. Of the

8 tanks, four were established as controls, four received a low dust concentration (0.05 g L-1), two

received a high dust concentration (0.5 g L-1) and two received an inorganic phosphorus addition. Half

of the tanks were incubated with turbulence, and the other half without. While tank replicability was

ensured by using two replicates per condition in half of the treatments, space constraints limited the

possibility of establishing 2 replicates of all treatments. The incubation was done at in situ temperature

(17ºC), and the experiment lasted for 8 days. At the end of the experiment 66% of the volume still

remained in the tanks.

The Saharan dust used in the experiment was collected during an intense dust storm over

Villefranche-sur-Mer, France (43°42’18“ N, 7°18’45“ E) on February 21, 2004. A hard plastic tray

was exposed 1 m above the ground to the rain (during ca. 24 h). The tray had 4 cm high side walls

and the dust accumulated at the bottom of the tray. Several hours after the rain had stopped, the

supernatant was poured off and the dust was transferred with a plastic spatula into an acid-washed

Nalgene high-density polyethylene (HDPE) bottle. The dust was then dried at 60ºC. The dry dust was

grinned before a single initial addition to the experimental tanks. Saharan dust deposition events carry

between 5 and 8000 mg of dust per litre (Ridame & Guieu 2002). A 0.05 g L-1 dust addition released

0.38 ± 0.08 µmol PO4

3- L-1 (Romero et al. in prep.), and this value was used as our low addition level

(DL) and a 10x higher value (0.5 g L-1) as our high addition level (DH). P treatments were enriched

with inorganic PO4

3- at a concentration of 0.5 µM, again as a single initial addition. This was estimated

to be the amount of P that we would add with the Saharan Dust, and would act as a control of the effect

of a P addition without the rest of components of the Saharan Dust.

Small-scale turbulence was generated with vertically-oscillating grids with the mechanical device

described before (Peters et al. 2002). We used a turbulent kinetic energy dissipation rate of 10-2 cm2

s-3. This value is within the range of turbulence intensities in coastal areas (Kiørboe & Saiz, 1995). We

applied turbulence only during the rst three days of the experiment, a typical duration of turbulence

events of the applied mean intensity in the Blanes area (Guadayol & Peters, 2006). Thus, containers

corresponding to turbulence (T) treatments underwent turbulent conditions for three days and remained

still until the end of the experiment, whereas still (S) treatments were kept still all along. Temperature

was adjusted to the in situ water temperature (17ºC) and light conditions were set to 225 µmol photons

m-2 s-1 inside the containers. The light:dark cycle was also adjusted to that of the time of the year. The

experiment started within 3 to 4 hours of water collection.

125

Chapter IV

Chlorophyll a and nutrients. Samples (100 ml) for estimating chlorophyll a concentration (Chla)

were ltered through Whatman GF/F lters of 25 mm diameter, which were immediately frozen

until extraction. Filters were then placed in 90% acetone at 4ºC for 24 h and the uorescence of the

extract measured using a Turner Designs uorometer (Yentsch & Menzel 1963). Total phosphorus

was determined by wet oxidation and colorimetry following standard protocols (further details in

Romero et al. in prep). For analysis of DOC concentration, ltered samples (20 ml) were acidi ed

with 16 mM HCl ( nal conc.) in acid-clean polypropylene tubes and stored at 4°C until analysis. DOC

was measured with a high temperature carbon analyzer (Shimadzu TOC 5000).

Organism abundance. Bacterial abundance and the percentage of high nucleic acid cells were

measured by ow cytometry (Gasol & del Giorgio 2000). Samples were run in a Becton-Dickinson

FACSCalibur cytometer after staining with SybrGreen (1:10,000 nal concentration, Molecular

Probes), and bacteria were detected by their signature in a plot of side scatter (SCC) vs FL1 (green

uorescence). Regions were established on the SSC versus green uorescence plot in order to

discriminate cells with HNA (high nucleic acid) content from cells with LNA (low nucleic acid)

content. In the same SybrGreen stained samples, we were able to enumerate the concentration of

dust particles (see below). Cell abundance was determined for each subgroup. Bacterial biomass was

calculated from abundance assuming a carbon content of 12 fg C cell-1 (Fukuda et al. 1998).

Live samples were analyzed in a FACSCalibur ow cytometer (Becton Dickinson) in unstained

aliquots for the enumeration of picophytoplankton populations. Milli-Q water was used as sheath

uid and 1-µm yellow-green latex Polysciences beads were used as an internal uorescence standard.

The samples were run at the highest possible speed (around 60 µl min-1) for 5 min, and the data were

acquired in log mode. Abundances were calculated by the ratiometric method from the known amount

of added beads, calibrated daily against TrueCount (Becton & Dickinson) beads. Synechococcus

were detected by their signature in a plot of orange uorescence (FL2) vs. red uorescence (FL3),

picoeukaryotes (Peuk) had higher FL3 signals and no FL2 signals. No Prochlorococcus were present

at the time of sampling.

Single-cell bacterial physiology. We used the activity probe CTC (5 cyano-2,3 ditolyl tetrazolium

chloride, Polysciences) to enumerate highly actively-respiring prokaryotes, as an indication of the

number of very active cells. Two 0.5 ml sample aliquots were spiked with 5 μM nal concentration

of CTC from a freshly prepared stock solution (50 mM in Milli-Q water) and incubated for 90 min

in the dark, at in situ temperature. The samples were immediately run through the FACSCalibur ow

cytometer. An additional 0.5 ml aliquot was xed with paraformaldehyde at time 0, and spiked with 5

126

Effects of Saharan dust on bacterial activity and community structure

mM nal concentration of CTC for a background control of CTC uorescence on dead samples. CTC-

formazan is excited by wavelengths between 460 and 530 nm and has bright red uorescence. CTC+

particles were those that showed red uorescence (above 630 nm, FL3 in our instrument), higher than

the background uorescence level (Gasol & Arístegui 2007). A dual plot of 90° light scatter and red

uorescence was used to separate CTC+ cells from background noise. The FL2 (orange uorescence)

vs FL3 (red uorescence) plot was used to differentiate the populations of photosynthetic microbes

from the CTC+ particles.

Bacterial production. Bacterial heterotrophic production was estimated using the 3H-leucine

incorporation method (Kirchman et al. 1985). Triplicate or cuatriplicate aliquots of 1.2 ml were

taken for each sample and one or two TCA killed controls. The Leu tracer was used at 40 nM nal

concentration in incubations of approximately 2h. The incorporation was stopped with the addition

of 120 µl of cold TCA 50% to the samples and, after mixing, they were kept frozen at –20ºC until

processing, which was carry out by the centrifugation method of Smith & Azam (1992).

Community metabolism.For primary production, two light and one dark acid-cleaned polycarbonate

bottles were lled with 160 ml water samples from each sample, inoculated with ~3.7 105 Bq (10 µCi)

of 14C-bicarbonate (VKI, Denmark). Incubations were done in the same chamber where the main

samples were incubating, and with the same light levels, and lasted for ~2 h at in situ temperature.

Samples were then ltered onto Whatman GF/F lters (25 mm diameter). Filters were placed in 6 ml

vials and fumed overnight with HCl 35%. Finally, 4.5 ml of ReadySafe liquid scintillation cocktail

were added before determination of disintegrations per minute (dpm) in the laboratory by means of a

Beckman LS6500 liquid scintillation counter. Dark bottle dpm were subtracted for correction of non-

photosynthetic 14C xation. No isotope discrimination factor was used. Conversion to carbon units

was performed assuming an ambient inorganic C concentration of 25000 mg C m-3.

For community respiration, we followed the changes in dissolved oxygen during dark incubations

of un ltered water. Eight BOD bottles were carefully lled, and three replicate bottles were

immediately xed with Winkler reagents to determine the initial oxygen concentration. Four replicate

bottles were incubated in the darkness at in situ temperature and xed with Winkler reagents after

approximately 24 hours. Dissolved oxygen measurements were made with an automatic titrator

(DL50 Graphix, Mettler Toledo) based on potentiometric endpoint detection (Outdot et al. 1988).

The rate of respiration was determined by regressing the oxygen concentration against the time in

which the samples were withdrawn. This estimation assumes that the disappearance of oxygen was

linear (Model I regression) and the slope of the regression is equal to the respiration rate. We assumed

127

Chapter IV

a respiratory quotient of 0.88 (see further details in Alonso-Sáez et al, 2008). Since, on an annual

basis, bacterial respiration at the site is 71±12% of community respiration (Alonso-Sáez et al. 2008),

we estimated bacterial respiration from the value of community respiration in order to obtain some

estimates of bacterial growth ef ciency, which is the fraction of bacterial production over bacterial

production+respiration.

Collection of community DNA. Microbial biomass was collected by sequentially ltering around

5 L of seawater through a 3 µm pore size polycarbonate lter (Millipore, 46 mm) and a 0.2 µm

polycarbonate lter (Millipore). Microbial biomass was treated with lysozime, proteinase K and sodium

dodecyl sulphate, and the nucleic acids were extracted with phenol and concentrated in a Centricon-

100 (Millipore). Nucleic acids were extracted by a standard protocol using phenol/chloroform (see

details in Schauer et al. 2003).

Fingerprinting analysis. DGGE and gel analysis were performed essentially as previously described

(Schauer et al. 2003, Sánchez et al. 2007). Brie y, 16S rRNA gene fragments (around 550 bp in length)

were ampli ed by PCR, using the universal primer 907rm and the bacterial speci c primer 358f, with

a GC-clamp. the PCR products were loaded on a 6% polyacrylamide gel with a DNA-denaturant

gradient ranging from 40-80%. The gel was run at 100V for 16 h at 60ºC in 1x TAE running buffer.

DGGE gel images were analyzed using the Diversity Database software (BIO-RAD). A matrix was

constructed for all lanes taking into account the relative contribution of each band (in percentage) to

the total intensity of the lane. Based on this matrix, we obtained a dendogram by UPGMA clustering

method (Euclidean distances, Statistica 6.0). DGGE bands were excised, reampli ed, and veri ed by

a second DGGE. Bands were sequenced using primer 358f without the GC clamp, with the BigDye

terminator cycle-sequencing kit (Perkin Elmer Corporation) and an ABI PRISM model 377 (v3.3)

automated sequencer.

Non-metric multidimensional scaling plots (NMDS) were also constructed from the binary matrix.

NMDS plots present data in an Euclidean plane (with dimensions of no special signi cance) in

which similar measurements appear close together. Each band pattern is a point in the plane and, by

connecting consecutive points, changes in community structure can be visualized and interpreted

(see van Hannen et al. 1999). Dendrograms and NMDS plots are a robust and widely used way of

comparison of microbial assemblage structure.

128

Effects of Saharan dust on bacterial activity and community structure

RESULTS

General experimental considerations. Our low SD additions resulted in an increase in P

concentration of ca. 0.3 µM above the 0.1 µM present in the water at the time of sampling (Table

1). In the DH (High dust) treatment the increase was ten times higher, up to 3.3 µM. Similarly, the

addition of low dust (DL) concentrations enriched the background DOC concentration in ca. 14 µM

(from 59 to 73 µM), while the DH treatment supposed an enrichment of slightly more than 10-fold, up

to ca. 260 µM (Table 1). Maximum DOC recorded at Blanes Bay is ca. 170 µM (Pinhassi et al. 2006,

Alonso-Sáez et al. 2008) and, thus, the DH treatment was well above the common values at the Bay.

Total P at the Bay oscillates between 0.01 and 0.2 µM (Guadayol et al. unpublished)

It was possible to detect the Saharan dust particles by ow cytometry. The particles appeared in the

cytograms that were used to enumerate bacteria by SybrGreen I (Fig. 1). They appeared as particles

with high SSC (light scatter) and FL3 (red uorescence), but low FL1 (green uorescence) as is shown

in the plots of FL3 vs FL1. The measured particles were present at ca. 3 x 106 particles ml-1 in the DH

treatments, and at about 6 x 105 ml-1 in the DL treatments. Dust particles slowly settled or disappeared

during the rst three days of the experiment and were absent at day 4. Perhaps counterintuitively, dust

disappeared faster in the T than in the S treatments (Fig. 1).

Fig.1.- Representation of Dust particles measurements by ow cytometry: concentration (left side) and green uorescence (FL1) vs red uorescence (FL3) plot of a dust sample (upper panel) and control (lower).

129

Chapter IV

Microorganism abundance. The experiment started with an initial chlorophyll a (Chla) value of

3 mg m-3, which is one of the largest values that can be detected in the Bay (see Alonso-Sáez et al.

2008), and that indicates that we had sampled a bloom. The bloom decayed throughout the experiment

(Fig. 2A), but did less so in the DL treatment. The P treatment reached values intermediate between

those of the C and DL treatments, with a small bloom developing towards the end of the experiment.

In the DH treatment, a bloom developed reaching values of up to 10 mg m-3 (Fig. 2B). Treatment

was a signi cant variable explaining the data (Table 2), particularly between treatments DH and

C. Treatment DL maintained on average 1.5x the Chla in the control. Synechococcus, which were

initially at 3 x 104 cells ml-1 disappeared from all treatments at similar rates (details not shown), while

picoeukaryotes responded different to the different treatments: they maintained abundances and even

increased slightly at the end of the experiment in the DL treatment (Fig 2C), while they decreased in

the control. They reached a peak of 4 x 104 cells ml-1 in the DH treatment in days 3-4, and a peak of

2 x 105 ml-1 in the P treatments towards the end of the experiment (Fig. 2D). Average Picoeukaryote

(Peuk) abundance was 2.7x in the DL treatment as compared to the C treatment (Table 2). Comparison

of the evolution of Chla and Peuk abundance suggests that while the Chla peak at the end of the

experiment can, in part, be assigned to the Peuk development, the one occurring in the DH treatment

needs be assigned to cells larger than picoeukaryotes. Indeed, the responsible for this Chla peak in the

DH treatment were nano agellates and diatoms, mostly of genus Nitzschia (Romero et al. in prep).

Table 1.- Concentrations of total phosphorus (µM) and of DOC (µM) at the start of the experiment (t0), but after addition of dust (SD) and inorganic phosphorus (IP), and after three days of incubation (t3). C stands for Control, DL for low dust concentration, DH for High dust concentration, P for Phosphorus addition, T for turbulence treatment, and S for Still treatment. When possible, average ± st.dev. of two replicated tanks.

Target concentrations Measured concentrations SD IP Total phosphorus (µM) DOC(µM) (g L-1) (µM) t0 t3 t(18h)

source seawater 0.095 ± 0.007 58.98 ± 0.65CT 0 0 0.115 ± 0.021 CS 0 0 0.120 ± 0.014 DHT 0.5 0 3.34 264.84 ± 2.55DHS 0.5 0 3.37 DLT 0.05 0 0.425 ± 0.021 73.17 ± 1.96DLS 0.05 0 0.445 ± 0.007 PT 0 0.5 0.42 PS 0 0.5 0.47

130

Effects of Saharan dust on bacterial activity and community structure

1

10

0 1 2 3 4 5 6 7

1

10

0 1 2 3 4 5 6 7

104

105

0 1 2 3 4 5 6 7

104

105

0 1 2 3 4 5 6 7

Control- StillControl - TurbulenceDust (low) - StillDust (low) - Turbulence

Dust (high) - StillDust (high) - TurbuencePhosphorus - StillPhosphorus - TurbulenceA B

C D

Time (days) Time (days)

Fig.2.- Upper panels: Changes in chlorophyll a concentrations (mg m-3): In the control tank (white circles) and dust low concentration tank (black circles) in still and turbulence conditions (A). In phosphorus addition tank (white squares) and dust high concentration tank (black squares)in still and turbulence conditions (B). Lower panels: Picoeukaryotes abundances (cell ml-1) in the same treatments.

131

Chapter IV

While heterotrophic bacteria decreased in the control tanks until day 3, and then increased towards

the end of the experiment (Fig. 3A), they presented an initial peak in treatment DL at day 1, which was

suppressed in day 2. The response in the DH treatment was much higher (> 10x in just one day, Fig.

3B), and the high bacterial concentrations lasted for three days after coming back to concentrations

similar than those in the control treatment at day 4. In both, P and DH treatments, bacteria increased

steadely from day 4 to the end of the experiment. The nal bacterial yield in the P and DH treatments

was slightly higher (4-8 106 cells ml-1) than in the C or DL treatments (3-5 106 cells ml-1). Addition of

dust (DL) increased bacterial abundance by a factor of 1.8 on average (Table 2).

The differences between treatments were higher in the physiological structure of the bacterial

assemblage. The %HNA cells increased immediately in all treatments, particularly in the DL, except

in the C (Fig 3C-D). Some effects of turbulence were obvious in this parameter, with %HNA being

higher in the T replicates of the C, P and DL treatments. The number of actively-respiring bacteria

Table 2.- Results of a One-way ANOVA with the time-integrated (for production) or time-averaged (for stocks) values for each treatment. Treatments with the same letters are not signi cantly different from each other. The Post-hoc tests give the probability of DL treatments being different from the C treatments, and of the P treatmens from the DL treatments. The “factor” is the ratio between the time-integrated value of treatment DL to that of treatment C.

Chla total BA BP Peuk %HNA P/B

Global ANOVA P <0.001 <0.001 0.005 0.002 0.002 NS

TreatmentCS D D D C D C ACT C D D C D D B C ADLS B C C A B A ADLT B C C A B B C A ADHS A B A B B A ADHT A A A B C B A APS C D C D B C D A A B APT B C C D A B C A A A

Post-Hoc tests

P Difference DL / C 0.001 0.004 0.002 0.007 <0.001 NSFactor (x) 1.5 1.8 2.3 2.7 1.4 -

P Difference DL / P NS NS NS 0.01 NS NS

132

Effects of Saharan dust on bacterial activity and community structure

� �� ��

�� � ����

Fig.3.- Upper panels: Bacterial abundance (cell ml-1): In the control tank (white circles) and dust low concentration tank (black circles) in still and turbulence conditions (A). In the phosphorus addition tank (white squares) and dust high concentration tank (black squares) in still and turbulence conditions (B). Lower panels: Physiological state measured by %HNA in the same treatments.

133

Chapter IV

(CTC+) followed a pattern quite similar to that of the %HNA cells, with a clear difference between the

DL and the C treatments (Table 3), a much larger response in the DH, and a relatively higher response

in the turbulent replicates of the C, P and DL treatments. While the differences in %HNA between

the C and DL treatments (Fig. 3) lasted for the whole incubation, those in the number or proportion

of CTC+ cells were mostly apparent during the rst 4 days of incubation. Table 3 was split into two

parts (times 1-3 and 4-7 days) to re ect better these differences. In any case, the difference in the

number of CTC+ cells was signi cant between the C and DL treatments for both periods, and it was

also signi cant for the %CTC+ cells in the rst period (Table 3).

Bacterial activity. Dust stimulated bacterial production by a factor of ca. 5x, from the control

treatment to that of the DL treatment (Fig. 4A, Table 2). The effect of the addition of P was more

pronounced when turbulence was applied in the incubation (Fig. 4B), but a positive enhancement

in the turbulent treatments was evident in almost all additions. Bacterial growth rates were higher

Table 3.- Average concentrations of CTC+ cells, %CTC (over total bacteria), for days 1 to 3, and for days 4 to 7. Average values ± SE of all replicates and times. In the lower part, results of the Student’s t comparisons between treatments C (control) and LD (Low Dust), or between treatments P (Phosphorus) and LD (Low Dust). In this last case, separating Still from Turbulence treatments.

CTC+ abund %CTC Days 1-3 Days 4-7 Days 1-3 Days 4-7

CS 0.82 105 ± 0.20 1.91 105 ± 0.28 10.4 ± 2.9 7.7 ± 3.0

CT 0.94 105 ± 0.05 0.74 105 ± 0.12 10.9 ± 2.3 5.2 ± 0.7

DHS 49.30 105 ± 14.7 32.46 105 ± 0.05 45.7 ± 22.4 6.3 ± 0.9

DHT 51.22 105 ± 14.60 3.55 105 ± 2.29 29.3 ± 8.2 26.1 ± 17.1

DLS 3.85 105 ± 0.81 3.52 105 ± 0.85 26.8 ± 8.5 6.9 ± 1.1

DLT 6.80 105 ± 1.82 2.17 105 ± 0.25 30.6 ± 6.2 8.5 ± 2.0

PS 0.91 105 ± 0.19 1.49 105 ± 0.48 9.7 ± 2.1 4.0 ± 0.4PT 2.24 105 ± 0.21 2.76 105 ± 1.07 16.2 ± 1.4 8.4 ± 2.4

Prob. diff DL / C P < 0.001 P < 0.001 P < 0.001 NSFactor (X) 6.02 2.15 2.69 1.20Prob. diff DL / P Still NS NS NS NS Turb P < 0.001 NS NS NS

134

Effects of Saharan dust on bacterial activity and community structure

� �� ��

�� � ����

in the DL treatment than in the C at days 2 and 3, but were similar later on (Fig. 4C). Since days 2-

3 are when bacterial abundances decreased, this could indicate the effects of grazing by agellates

developing and feeding on bacteria at those times. The addition of a high amount of Dust (DH) did

not promote bacterial production above the values created by the addition of lower amounts of dust

(DL). It is possible that incorporation at the added 40 nM leucine was actually saturated in these

samples that had, at the beginning ca. 107 cells ml-1. The relatively low growth rates measured for the

DH treatments at the start of the experiment (Fig. 4D), are not compatible with the large increase in

abundances measured (Fig. 3B), supports this possibility, and suggests that the measured values of

production and growth rates in the DH treatments might be unreliable.

Fig.4.- Upper panels: Bacterial production (μg C L-1): In the control tank (white circles) and dust low concentration tank (black circles) in still and turbulence conditions (A). In the phosphorus addition tank (white squares) and dust high concentration tank (black squares) in still and turbulence conditions (B). Lower panels: Bacterial growth rate (d-1) in the same treatments.

135

Chapter IV

Tree Diagram for 13 CasesUnweighted pair-group average

Euclidean distances

0 10 20 30 40 50 60Linkage Distance

ToCS T3CT T3

DHS T3DHT T3DHS T7DHT T7DLS T3DLT T3PS T3PT T3PS T7PT T7

Resemblance: S17 Bray Curtis similarity

PT T7

PS T7

DHT T7 DHS T7

PT T3

PS T3

DHT T3DHS T3

DLT T3DLS T3

CT T3CS T3

To

2D Stress: 0.07

T7T3T0

CS CT DLS DLT DHS DHT PS PT DHS DHT PS PT

SD1SD2SD3SD4 SD6

SD7SD8SD10

SD11SD12SD13SD14SD15

Community metabolism. We measured primary production twice during the experiment, and

community respiration once. Dust addition increased both, PP and CR (Table 4), the low dust addition

by factors ranging from 1.6 to 3.6. The DH addition produced much larger increases, with factors of

6-15. In all cases the increases in metabolism caused by the dust were stronger for PP than for CR,

thus turning the communities towards net autotrophy from a near-balanced initial community. There

was a clear difference between the DL and the P treatments: P stimulated PP much more than DL,

while CR was similarly stimulated by both treatments. The net result of the addition of DL was a

heterotrophic system, while the net result of the DH and P additions were autotrophic communities.

It is somehow dif cult to compare to the controls since CR in the control-Turbulence treatment was

highly stimulated.

Fig.5.- Denaturing Gradient Gel Electrophoresis (DGGE) of bacterial 16S rRNA gene fragments from samples of different treatments and at different times of incubation (A). Dendogram classi cation (Ward’s Method, Euclidean distances according to the band pattern) of DNA samples. (B). Nonmetric multidimensional scaling (MDS) of the bacterial assemblage composition data (C).

136

Effects of Saharan dust on bacterial activity and community structure

Bacterial community structure. We ngerprinted bacterial assemblage composition by

denaturing gradient gel electrophoresis (DGGE). The Nonmetric multidimensional scaling (MDS)

and Ward’s clustering method were used in order to explore the similarities between the samples

based on bacterial assemblage structure, and to compare the effect of the dust addition to those of

phosphorus and control treatments. The P and DL communities after the addition clustered together

and were distant from the C or DH communites. (Fig. 4). The control tanks were similar to the initial

sample. There were some effects of Turbulence, more obvious at the end of the experiment (Time 7)

than at time 3. We retrieved 13 bands that were dominant in different treatments (Table 5). All the

sequences that could be retrieved from the DGGE gel were found to have similarities in GenBank to

previously described organisms, with values > 97% (Table 5). Just one phylotype was dominant in all

the treatments (SD2), four appeared only in phosphorus treatments, 3 only where Saharan dust had

been added, and 3 in both P and Dust treatments.

DISCUSSION

The fertilizing nature of Saharan dust. The amount of SD that we added in our main experimental

treatment, 50 mg L-1 was slightly higher than the doses used experimentally in a couple of recent

studies (Herut et al. 2005, Pulido-Villena et al. 2008), but were within the limits of what can be

expected to occur in the ocean. According to Ridame & Guieu (2002), 60% of ‘Saharan rains’ carry

between 5 and 8000 mg of dust per liter. A 0.05 g L-1 dust concentration released, abiotically, 0.38 ±

0.08 µmol PO43- L-1 (Romero et al. in prep) a value which is somewhat lower than the range reported

in Ridame & Guieu (2002), where 0.05 g L-1 of dust released anywhere between 0.97 and 3.2 µmol

PO4

3- L-1. In the experimental tanks, however, we obtained ca. 3 µmol PO43- per gram of dust (close

to the yields reported by Pan et al. 2002, and Herut et al. 2005), which led to nal concentrations of

1.5 µmol PO4

3- L-1 and 0.15 µmol PO4

3- L-1 in the DH and DL containers respectively (Romero et al.

in prep). This is an amount of P that increased the initial P concentration by a modest factor of 1.3

(Table 1) and, thus, addition of a smaller dust amount would most probably not have affected the

microbial communities of the coastal waters we used for the experiment. This might contrast with

the experiments of Herut et al. (2005) and Pulido-Villena et al. (2008), which used oceanic waters,

initially more oligotrophic than the waters we used.

In spite of the fact that the P addition was relatively modest, it triggered clear effects in Chla

development (an average factor of 1.5x, Fig 2), in picoeukaryotes development (an average factor of

2.7x), and in primary production (an average factor of 1.5x at day 3 and 2.7x at day 7, Table 4). That

SD enhances primary production had been previously demonstrated in several experimental studies

(Blain et al. 2004, Mills et al. 2004, Herut et al. 2005), and the yield we observed (on average 0.5 µg

137

Chapter IV

Chla per 0.05 g. dust, i.e. 50 µg g-1) is exactly the same yield observed by Herut et al. (2005) in the

Eastern Mediterranean, but smaller than the observations of Blain et al. (2004) and Mills et al.(2004)

in the North Atlantic. Enhancement of chlorophyll in the DH treatment was on the order of 10x the

enhancement generated in the DL treatment (Fig. 2), which is consistent with the 10-fold difference in

the addition. Since SD contains nutrients other than P (Jickells et al. 2005), the hypothesis that other

limiting nutrients were the responsible for the primary producers enhancement cannot be disregarded

and, in fact, the addition of P alone resulted in a smaller enhancement than the DL addition (Fig 2),

although on average the difference was not statistically signi cant (Table 2).

Bacteria were also stimulated by the dust addition: abundance (an average factor of 1.8x),

production (factor of 2.3x), number of actively respiring cells (factors of 2-6x), and also the single-

cell indications of cellular activity, such as the %HNA (factor of 1.4x) and the %CTC+ cells (a

factor of 2x). However, bacterial growth rates were not statistically enhanced by the DL addition. The

stimulation of bacterial heterotrophic activity resulted also in a stimulation of community respiration

(at this site, BR is aprox. 70% of CR, Alonso-Sáez et al. 2008), by a factor of 1.6-2.4 (Table 4).

Bacterial abundance and production were, thus, enhanced by a factor higher than the enhancement of

chlorophyll a. P enrichment in the DL treatment was of a factor of 1.3x and DOC enrichment was also

of a similar factor (1.24x). However, the background DOC concentration at the start of the experiment

(59 µM, Table 1) is typically formed by mostly nonreactive C (Hansell & Carlson 1998). The 14 µM

added with the DL treatment could indicate a large enrichment factor, although since we do not know

the nature of this DOC, we can not con rm this observation except through its effects on bacterial

abundance, single-cell activity and heterotrophic production.

A response of the heterotrophic community to the SD addition had already been observed by

Bonnet et al. (2005) and Herut et al. (2005), even though they did not measure the DOC added. In

these studies, as in our DL treatment, the stimulation of bacterial production was not followed by a

proportional stimulation of bacterial abundance, something that they assigned to HNF development

cropping the stimulated production. HNF reached a peak at day 3 in our experiment (Romero et al. in

prep) and likely cropped most of the new production in the rst days of the experiment, but not after

day 4 (Fig. 3), maybe because bacteria then developed grazing-defence strategies. Bacteria seemed to

escape HNF checking in the DH treatments (Fig. 3B): the burst of growth caused by the large nutrient

dose escaped grazer control.

The bacterial activity “yield” that we obtained was similar to that measured by Herut et al. (2005):

they obtained ca. 75 pmol leucine h-1 per mg dust, and we obtained 70 pmol h-1 per mg dust. The

bacterial abundance yield we measured (2 x 109 cells per 50 mg dust, in the DL treatment and 2 x

138

Effects of Saharan dust on bacterial activity and community structure

1010 cells per 500 mg in the DH treatment) was consistent between treatments, but lower than the 1

x 109 cells per 2 mg measured by Pulido-Villena et al. (2008). In the DH treatment, however, leucine

incorporation rates should had been 10x those we measured in the DL treatment and were similar,

something that we think was caused because we added too low amount of leucine tracer for the

enrichment we had generated.

Biogeochemical effects of SD: autotrophy vs. heterotrophy. Saharan Dust enhanced

community respiration in our experiment, as it did in the experiments and natural observations of

Pulido-Villena et al. (2008). These authors measured a 2x enhancement of bacterial respiration, while

we measured a factor of 1.6-2.4x (Table 4) with a higher SD dose. Our results suggest that this

enhancement had to do with the inorganic nutrient fertilization caused by the SD, but also with the

DOC that was introduced with the dust. SD contains carbon in amounts that might easily reach 1% dw

(Eglinton et al. 2002). This carbon originated in the past in vegetation res accumulated and stored

in soil (Eglinton et al. 2002), and is thus, likely not to be very labile. Our results suggest some use of

this carbon by the resident (see below) bacterial community.

We only measured respiration at the end of the experiment (Table 4), when respiration in the Still

treatment was not signi cantly different between the P and DL treatments, but was much higher in the

DL treatment under turbulence. A similar result could be observed in the number of actively respiring

cells (CTC+, Table 2), which were, under turbulence, much higher in the DL treatment than in the P

treatment. In some studies the number of CTC+ bacterial cells have been seen to vary coherently with

bacterial respiration (Smith, 1998) and increases in the number of CTC+ cells are usually associated

to bursts in growth of bacteria (del Giorgio & Gasol, 2008).

The metabolic balance of the microbial community was altered by the dust addition in a way

different from the alteration caused by P addition. P shifted the community to net autotrophic ( nal

CR/PP values of 0.4-0.5). In the DL treatment the values were net heterotrophic (1.47 ± 0.73 and

1.63 ± 0.52). The control treatment stayed close to 1 (0.95 ± 0.35) in the still treatment but became

very heterotrophic under turbulence (see discussion in next section). Interestingly, the high addition

of dust (DH, 0.5 g L-1) shifted the community towards net autotrophy while the DL treatment shifted

it towards heterotrophy. In the DH treatment a large diatom bloom developed with Chla values of ca.

10 mg m-3 (see details in Romero et al. in prep). In contrast, the 1.5x increase in Chla measured in

the DL treatment consisted in pico- and nanoeukaryotic cells. Apparently, at least from what can be

extrapolated from our few data, a low addition of SD can shift the system to heterotrophy, but a larger

input might trigger the development of diatom blooms that would turn the system net autotrophic.

139

Chapter IV

Table 4.- Values of community metabolism in the different treatments. When errors are present, they correspond to the standard error of two replicated microcosms.

Primary Production Community Respiration CR/PP

(mgC m-3 h-1)

Initial 5.02 4.66 0.93Control – Still day 3 2.32 ± 0.23 day 7 1.24 ± 0.38 1.13 ± 0.10 0.95 ± 0.35

Control – Turbulence day 3 2.40 ± 0.28 day 7 1.04 ± 0.06 3.40 ± 0.11 3.27 ± 0.08

Dust (low) – Still day 3 3.64 ± 0.16 day 7 2.26 ± 0.78 2.74 ± 0.50 1.21 ± 0.73

Dust (low) – Turbulence day 3 3.25 ± 0.12 day 7 3.71 ± 1.16 5.42 ± 0.03 1.46 ± 0.52

FACTOR DL / C day 7-Still 1.82 2.42 day 7-Turbulence 3.56 1.59

Dust (high) – Still day 3 15.76 day 7 39.29 18.27 0.46

Dust (high) – Turbulence day 3 25.20 day 7 21.70 19.70 0.91

Phosphorus – Still day 3 1.55 day 7 6.21 3.08 0.50

Phosphorus – Turbulence day 3 2.13 day 7 8.98 3.62 0.40

FACTOR P/DL day 7-Still 2.75 1.12 day 7-Turbulence 2.42 0.67

140

Effects of Saharan dust on bacterial activity and community structure

The type of algae using the added P might explain this difference, as well as the possibility that

some nutrient thresholds exist that shift the “winners” from picoeukaryotes to diatoms. Silica, which

was known to be present in SD but with low availability (Bonnet et al. 2005) might be used by the

autotrophic plankton when added at some concentrations. Indeed, we observed that the SD that we

used contained silica, that this was released into the water, and that it was presumably used up by the

microbial community in the DH treatment (Romero et al. in prep).

In this sense it is interesting to refer to the models of the microbial food web developed by Thingstad

(in e.g. Thingstad et al. 2007). The models idealize competition for P between bacteria, agellates

and diatoms. Addition of labile-DOC shifts the uptake towards the bacteria (and, thus, the system

would turn net heterotrophic), while addition of silicate shifts it towards the diatoms (and the system

would turn net autotrophic). We hypothesize that our DL treatment would be a situation of the rst

type, while the DH treatment would be of the second type. The metabolic response of the plankton

community to Saharan Dust additions would be concentration-dependent.

Pulido-Villena et al. (2008) presented evidence that, on top of the fertilizing nature of SD events

that could increase primary production and C drawback to the ocean, the stimulation of heterotrophic

bacteria could actually reduce the amount of C sequestration caused by these events. These authors

calculated that the dust-induced bacterial growth could mineralize up to 70% of the annual bioavailable

DOC annually exported to the deep Mediterranean. Our results concur with those of these authors

in assigning a very relevant role to bacteria in the use of the nutrients accompanying the dust pulse,

but suggest that the real metabolic effects are concentration-dependent, and might depend on the

bioavailability of the P, Si and organic C attached to the dust, something that is currently very poorly

constrained.

Interactions with turbulence. Our experiment was not designed to be a proper test of the effect of

turbulence as an external factor modifying the response of he microbial community. We just wanted to

create a situation likely to be close to that encountered by the plankton community at the time when a

dust pulse is more probable. Perhaps counterintuitively, the applied turbulence levels facilitated dust

sedimentation or dissolution (Fig. 1), instead of retarding it. This result is consistent with experimental

work that shows that algae sediment faster under turbulence than under still conditions (Ruiz et

al. 2004) and with work showing that at low levels of the shear caused by turbulence, the balance

between aggregation and coagulation of particles of sizes 2-10 µm was dominated by coagulation

(Colomer et al. 2005). Cytometric detection of dust was likely not complete since very small particles

probably escaped detection by the ow cytometer (Fig. 1). In the SSC (a surrogate of particle size)

141

Chapter IV

Tabl

e 5.

-Phy

loge

netic

af

liatio

n of

16S

rR

NA

gen

e se

quen

ces

from

exc

ised

DG

GE

ban

ds o

btai

ned

in th

e di

ffere

nt tr

eatm

ents

.

For

each

phy

lotyp

e, th

e clo

sest

rela

tive

in G

enBa

nk a

re s

hown

, tog

ether

with

their

acc

essio

n nu

mber

s an

d se

quen

ce s

imila

rity.

We a

lso

indi

cate

wheth

er th

e phy

lotyp

e was

dete

cted

in th

e P tr

eatm

ents,

in th

e D tr

eatm

ents,

or w

heth

er th

e ban

d ten

ded

to b

e mor

e com

mon

in th

e Tu

rbul

ence

trea

tmen

ts.

Ban

d ID

# P

D

T

C

lose

st r

elat

ive

Acc

ess.

#

% s

imila

rity

Ph

ylog

enet

ic

grou

p

SD1

X

- -

U

ncul

ture

d Fl

avob

acte

ria

bact

eriu

m c

lone

MS0

24-1

F E

F202

335

99%

B

acte

roid

etes

Te

naci

bacu

lum

sp.

E

F416

572

98%

SD2

X

X

-

Unc

ultu

red

Tena

ciba

culu

m s

p.

AY

5735

21

100%

B

acte

roid

etes

Po

lari

bact

er d

okdo

nens

is s

trai

n M

ED

152

AY

5735

21

98%

SD3

- X

X

Unc

ultu

red

Tena

ciba

culu

m s

p.

AY

5735

21

99%

B

acte

roid

etes

Te

naci

bacu

lum

sp.

A

B27

4770

98

%

SD4

- X

-

U

ncul

ture

d Te

naci

bacu

lum

sp.

A

Y57

3521

99

%

Bac

tero

idet

es

Pola

riba

cter

dok

done

nsis

str

ain

ME

D15

2 D

Q48

1463

98

%

SD6

- X

X

Unc

ultu

red

cyto

phag

a sp

. A

J487

533

98%

B

acte

roid

etes

SD7

X

- X

Unc

ultu

red

Bac

tero

idet

es

AB

2660

10

96%

B

acte

roid

etes

L

ewin

ella

nig

rica

ns

AB

3016

15

92%

SD8

X

- -

U

ncul

ture

d B

acte

roid

etes

A

B26

6010

99

%

Bac

tero

idet

es

Lew

inel

la n

igri

cans

A

B30

1615

90

%

SD10

X

-

-

Unc

ultu

red

alph

apro

teob

acte

rium

E

F506

911

99%

A

lpha

prot

eoba

cter

ia

Mar

inos

ulfo

nom

onas

met

hylo

trop

ha

AY

7717

69

97%

SD11

X

X

-

U

ncul

ture

d al

phap

rote

obac

teri

um

AY

9222

43

97%

A

lpha

prot

eoba

cter

ia

Ros

eoba

cter

gal

laec

iens

is

AJ8

6725

5 96

%

SD12

X

-

-

Unc

ultu

red

alph

apro

teob

acte

rium

D

Q44

6171

98

%

Alp

hapr

oteo

bact

eria

T

hala

ssob

ius

sp.

EF5

8795

1 97

%

SD13

-

X

-

Ros

eoba

cter

sp.

A

Y25

8088

97

%

Alp

hapr

oteo

bact

eria

A

ntar

ctob

acte

r he

lioth

erm

us

Y11

552

97%

SD14

-

X

-

Rho

doba

cter

acea

e ba

cter

ium

A

J810

845

99%

A

lpha

prot

eoba

cter

ia

Rue

geri

a at

lánt

ica

AF1

2452

1 99

%

SD15

X

X

-

R

uege

ria

algo

colu

s X

7831

5 98

%

Alp

hapr

oteo

bact

eria

142

Effects of Saharan dust on bacterial activity and community structure

vs. uorescence plots (cytogram not shown), dust particles appeared to range from sizes smaller than

bacteria to >20 µm and, thus, we can expect the dust particles detected by the cytometer to be subject,

most likely, to the same aggregation/coagulation regulation shown by Colomer et al. (2005).

We also found a systematic, but weak, positive effect of turbulence on Chla development (Fig. 2A-

B). Experimental turbulence has been seen to enhance chlorophyll development in several laboratory

studies (e.g. Arin et al. 2002, Pinhassi et al. 2004), and in non-enriched and slightly enriched waters

of Blanes Bay particularly in spring (Guadayol et al. unpublish). Turbulence had more effect on

the sample receiving P (Fig. 2B) and less effect in the control sample, something that was also seen

in Arin et al. (2002). While Chla showed a positive enhancement by turbulence, picoeukaryote

development (Fig 2C-D and Table 2) and bacteria (Table 2) were less affected. Shifts in food web

structure and agellate feeding rates caused by turbulence could explain these differences (Peters et

al. 1998, 2002). The most relevant effects on bacterial activity were again seen in the treatment that

had received inorganic nutrients alone (P), which showed a large difference in leucine incorporation

(Fig. 4) and in the number of CTC+ cells (Table 3). These effects resulted in relatively similar levels

of autotrophic and heterotrophic metabolism at day 7 in the P treatment (Table 4), but in very strong

differences in the control treatment, which respired more under turbulence than under still conditions

(t-test, P<0.001). The DL treatment also respired more under turbulence than under still conditions

(t-test, P<0.001), while the balance CR/PP was signi cantly higher for the control under turbulence

than for the still control (t-test, P<0.001). The metabolic balance difference was not signi cant for

treatment DL.

It has to be taken into account that turbulence was applied only from day 1 to day 3 and, thus,

the difference reported in Table 4 occurred 4 days after turbulence had stopped. In the only study to

date in which microbial community respiration has been studied under turbulence or still conditions,

turbulence always generated a higher level of respiration, and a higher heterotrophic metabolic

balance (Alcaraz et al. 2002). Other previous studies had not found effects of turbulence on respiration

(Peters et al. 1998). Turbulence affected little bacterial community structure (Fg. 5), something that

is compatible with the explanation in Pinhassi et al. (2004) that claimed that changes in community

structure were linked to the changes in phytoplankton community structure caused by the turbulence

and not to the physical stress by itself.

The potential for affecting bacterial community structure. We tested whether bacterial

community structure was affected by the different treatments. In particular, we were interested to

see whether the DL and P treatments were signi cantly different. It has been speculated that airborne

143

Chapter IV

bacteria can travel associated to Saharan dust particles (Grif n et al. 2001, Kellogg & Grif n 2006,

Grif n 2007). If that were to be a relevant source of allochthonous bacteria able to develop in the

ocean, we should see it in our analyses. Of course, if that were only a tiny contribution of inactive

bacteria, we would not see it with DGGE, as this technique has a low resolution (picks up only <30

dominant phylotypes, Muyzer et al. 1997).

The main differences observed in community structure were those between the DH treatments and

the DL and P treatments (Fig. 5), indicating that the evolution with time of the bacterial community

caused by the low addition of SD was very similar to that caused by P alone. Even the high addition

of dust didn’t seem to affect the number of bands that were detected (about 12-15 per sample).

Mainly alphaproteobacteria and bacteroidetes could be retrieved from the DGGE and identi ed. We

did not retrieve any gammaproteobacteria, but this might be related to the fact that gammaproteobacteria,

particularly Alteromonas, tend to develop at the beginning of these mesocosm experiments (Schäfer

et al. 2001, Allers et al. 2007) and tend to be replaced by alphaproteobacteria at the end of these same

experiments, particularly by Rhodobacteraceae of types similar to the SD11 and SD13 bands seen in

our experiment.

Several of the Bacteroidetes that we retrieved were similar to Tenacibaculum and Polaribacter

(Table 5). This group contains the genome-sequenced Polaribacter dokdonensis strain M152 (very

similar to band SD2) which has been seen to contain a substantial number of genes for attachment

to surfaces or particles, gliding motility, and polymer degradation (in agreement with the currently

assumed life strategy of marine Bacteroidetes), but also the proteorhodopsin gene, which together

with a remarkable suite of genes to sense and respond to light, may provide a survival advantage

in the nutrient-poor sun-lit ocean surface when in search of fresh particles to colonize (González et

al. 2008). Organisms of this group have been retrieved in culture, in DGGE bands from enrichment

experiments, but are also appearing in situ in Blanes Bay (Lekunberri et al. submitted Chapter 1),

something that make them part of the most interesting marine Bacteroidetes, as they grow in culture

but at times dominate in situ. Some Tenacibaculum-like organisms (SD3 and SD4) appeared in SD

additions, but the sequence closer to M152 (SD2) appeared in almost all experimental treatments.

Bacteroidetes of the Lewinella type (bands SD7 and SD8), relatives of Chitinophaga, are not

common in the area (Lekunberri et al. submitted Chapter 1). They appeared only in the P treatment.

Within the alphaproteobacteria identi ed, bacteria of the Ruegeria/Silicibacter cluster appeared in

the D and P treatments at the end of the experiment (SD14 and SD15). These are common organisms

144

Effects of Saharan dust on bacterial activity and community structure

in Blanes Bay, easy to isolate in plates, and common in mesocosmss experiments (Lekunberri et al.

submitted Chapter 1). The other alphaproteobacteria identi ed (SD10 to SD13) were also commonly

developing in mesocosmss experiments (Allers et al. 2007). In that study different types of sequences

related to Rhodobacteraceae dominated the different treatments in nutrient-enriched mesocosmsss.

Most of the treatments elicited small phytoplankton growth (the mesocosmsss were light-limited),

and the development of picoeukaryotes. This might be the reason for the similitude of these sequences

to those of Allers et al. (2007) and for the difference between the DL and DH treatments (Fig 5). As

pointed out by Pinhassi et al. (2004), the type of algae developing in the system determined bacterial

community structure.

Bands 3, 4, 6, 13, 14 and 15 were associated in some way with the dust addition, but most of them

appeared also as faint bands in some P treatments (e.g. band SD15). Band SD2, which was very

similar to band SD4, appeared in all treatments. It is, thus, unlikely that any of the bands identi ed

can be considered as speci c of the SD addition treatments, and the af liation of the bands also do

not support this speci city. Bacterial community structure in the DL treatment was likely to be mainly

generated by the P pulse, and the associated development of algae. The situation was probably slightly

different in the DH treatment.

Conclusion

A small Saharan dust addition affected bacterial abundance, activity and diversity in our mesocosm

experiments. It also shifted the metabolic balance of the community to heterotrophy, something that we

assign to the DOC addition simultaneous to the P addition with the dust. However, a higher addition of

dust caused a shift towards autotrophy. Bacterial community structure was affected simialrly by dust

than by P alone. The results reinforce the potential for Sahara Dust inputs to affect bacterial diversity

and function, but clearly indicate the need of knowing better the composition of the inputs, as the

results encountered were, in part, dependent on the concentration of added nutrients.

ACKNOWLEDGEMENTS

We thank Raffaela Cattaneo and Òscar Guadayol for help during the experiment. We also thank V.

Balagué, C. Cardelús and I. Forn for careful organization of sampling at the Microbial Observatory.

This work was supported by Spanish projects MODIVUS (CTM2005-04795/MAR) and VARITEC

(CTM2004-0442-C02), and is a contribution to the european Networks of Excellence Eur-Oceans and

MARBEF.

145

Chapter IV

REFERENCESAllers E, Gómez-Consarnau L, Pinhassi J, Gasol JM, Simek K & Pernthaler J (2007) Response of

Alteromonadaceae and Rodobacteriaceae to glucose and phosphorus manipulation in marine mesocosms. Environ Microb 9: 2417-2429.

Alonso-Sáez L., E. Vázquez-Domínguez, C. Cardelús, J. Pinhassi, M.M. Sala, I. Lekunberri, V. Balagué, M. Vila-Costa, F. Unrein, R. Massana, R. Simó & J.M. Gasol (2008) Factors controlling the year-round variability in carbon ux through bacteria in a coastal marine system Ecosystems. 11: 397-409.

Alcaraz M, Marrasé C, Peters F, Arin L, Malits A (2002) Effects of turbulence conditions on the balance between production and respitation in marine planktonic communities. Mar Ecol Prog Ser 242: 63-71.

Arin, L., Morán, X.A.G., Estrada, M. (2002) Phytoplankton size distribution and growth rates in the Alboran Sea (SW Mediterranean): short term variability related to mesoscale hydrodinamics. J. Plankton Res 24: 1019-1033.

Blain S et al. (2004) Availability of iron for phytoplankton in the nort-east Atlantic Ocean. Limn Oceanogr. 49: 2095-2104.

Bartoli G, Migon C & Losno R (2005). Atmospheric Input of dissolved inorganic phophorus and silicon to the coastal northwestern mediterranean sea: uxes, variability and possible impact on phytoplabkton dyanamics. Deep-Sea Res I 52: 2005-2016.

Bonnet, S., C. Guieu, et al. (2005). Effect of atmospheric nutrients on the autotrophic communities in a low nutrient, low chlorophyll system. Limn Oceanogr 50: 1810-1819.

Colomer J, Peters F, Marrasé C. (2005) Experimental analysis of coagulation of particles under low-shear ow. Water Res 39: 2994-3000.

Currie DJ (1984) Microscale nutrient patches: do they matter to the plankton? Limn Oceanogr 29: 211-214

del Giorgio PA & Gasol JM (2008) Physiological structure and single-cell activity in marine bacterioplankton. pp. 243-298 In: Kirchman, DL (ed). Microbial Ecology of the Ocean, 2nd Ed. John Wiley & Sons Inc., Hoboken, NJ.

Eglinton TI, Eglinton G, Dupont L, Sholkovitz ER, Montlucon D & Reddy CM (2002) Composition, age, and provenanace of organic matter in NW African dust over the Atlantic Ocean. Geochem. Geophys, Geosyst, 3:10.1029/2001GC000269

Fukuda R, Ogawa T, Nagata T & Koike I (1998) Direct determination of carbon and nitrogen contents of natural bacterial assemblages in marine environments. Appl Environ Microb 64: 3352-3358.

Gasol JM., del Giorgio PA. (2000) Using ow cytometry for counting natural planktonic bacteria and understanding the structure of planktonic bacterial communities. Scientia Marina 64: 197-224.

Gasol JM & Arístegui J. (2007) Cytometric evidence reconciling the toxicity and usefulness of CTC as a marker of bacterial activity Aquat Microb Ecol 46: 71-83.

González JM et al. (2008) Genome análisis of the proteorhodopsin-contining marine bacterium Polaribacter sp. MED152 (Flavobacteria). PNAS 105:8724-8720.

Grif n DW, Garrison VH, Kellogg C, Shinn EA. (2001) African desert dust in the Caribbean atmosphere: Microbiology and public health. Aerobiologia 17: 203-213

Grif n DW (2007). Atmospheric movement of microorganisms in clouds of desert dust and implications for human health. Clin Microb Rev 20: 459–477.

146

Effects of Saharan dust on bacterial activity and community structure

Guerzoni S et al. (1997). Saharan dust inputs to the western Mediterranean Sea: Depositional patterns, geochemistry and sedimentological implications. Deep-Sea Res II 44(3-4): 631-654.

Guadayol Ò & Peters F (2006) Analysis of wind events in a coastal area: a tool for assessing turbulence variability for studies on plankton. Scientia Marina 70: 9-20.

Guerzoni S, et al. (1999). The role of atmospheric deposition in the biogeochemistry of the Mediterranean Sea. Progr Oceanogr 44: 147-190.

Hansell DA & Carlson CA (1998) Deep ocean gradients in dissolved organic carbon concentrations. Nature395: 263-266.

Herut B, Krom M D, Pan G & Mortimer R (1999). Atmospheric input of nitrogen and phosphorus to the Southeast Medieterranean: Sources, uxes and possible impact. Limn Oceanogr 44: 1683-1692.

Herut B, Collier R &Krom MD (2002). The role of dust in supplying nitrogen and phosphorus to the Southeast Mediterranean. Limn Oceanogr 47: 870-878.

Herut B et al. (2005). Response of East Mediterranean surface water to Saharan dust: On-board microcosm experiment and eld observations. Deep-Sea Res II 52: 3024-3040.

Jickells T. (2005). Atmospheric Inputs of Metals and Nutrients to the Oceans - Their Magnitude and Effects. Mar Chem 48: 199-214.

Kellogg CA & Grif n, DW. [2006] Aerobiology and the global transport of desert dust. Trends Ecol Evol 21: 638-644

Kiørboe T & Saiz E (1995) Planktivorous feeding in clam and turbulent enriornments, with amphasis on copepods. Mar Ecol Pro Ser 60:35-44

Kirchman DL, Knees E & Hodson R (1985) Leucine incorporation and its potential as a measure of protein synthesis by bacteria in natural aquatic systems. Appl Environ Microb 49: 599-607.

Klein C, Dolan JR, et al. (1997). Experimental examination of the effects of rainwater on microbial communities in the surface layer of the NW Mediterranean Sea. Mar Ecol-Progr Ser 158: 41-50.

Lenes JM et al (2001) Iron fertilization and the Trichodesmium response on the West Florida shelf. 46::1261-1277.

Loyë-Pilot MD & Martin JM (1996). Saharan dust input to the western Mediterranean: An eleven years record in Corsica. The Impact of Desert Dust across the Mediterranean:, edited by S. Guerzoni and Chester, pp. 191-199 Springer, New York.

Markaki 2003 Markaki Z, Oikonomou K, Kocak m, Kouvarakis G, Chaniotaki A, Kubilay N, Mihalopoulos N (2003). Atmospheric depsition of inorganic phophorus in the Levantine Basin, eastern Mediterranean: spatial and temporal variability and its role in sweater productivity. Limn Oceoanogr 48:1557-1568.

Mills MM, Ridame C, Davey M, La Roche J & Geider RJ (2004) Iron and phosphorus co-limit nitrogen xation in the eastern tropical North Atlantic. Nature 429: 292-232

Muyzer AE, Brinkhoff T, Nübel U, Santegoeds C, Schäfer H & Wawer C (1997) Denaturing gradient gel electrophoresis (DGGE) in microbial ecology. In: Molecular Microbial Ecology Manual (ADDERMANDS ADL, van Elsas JD & Brujin FJ Eds.) Vol. 3.4.4, pp.

Obernosterer I, Kawasaki N & Benner R (2003) P-limitation of respiration in the Sargasso Sea and uncoupling of bacteria from P-regeneration in size-fractionation experiments. Aquat Microb. Ecol 32: 229-237.

147

Chapter IV

Outdot CR, Gerard R, Morin P, Gningue I (1988) Precise shipboard determination of dissolved oxygen (Winkler procedure) for productivity studies with a commuercial system. Limn Oceanogr 33: 146-150

Pan G, Krom MD, et al. (2002). Adsorption-desorption of phosphate on airborne dust and riverborne particulates in East Mediterranean seawater. Environmental Science & Technology 36: 3519-3524.

Pérez C, Jiménez P, et al. (2007). Long-term trends (1987-2006) of Saharan dust over the Mediterranean and the Canary Islands with the DREAM regional dust model. Geophys Res Abstr 9: 08525.

Peters F, Marrasé C, Gasol, JM, Sala, MM, Arin L (1998) Effects of turbulence on bacterial growth mediated through food web interactions. Mar Ecol Progr Ser 172: 293-303.

Peters F, Marrasé C, Havskum H, Rassoulzadegan F, Dolan J, Alcaraz M, Gasol JM (2002) Turbulence and the microbial food web: effects on bacterial losses to predation and on community structure. J. Plank Res24: 321-331.

Pinhassi J, Sala M, Havskum H, Peters F, Guadayol O, Malits A & Marrasé C. (2004) Changes in bacterioplankton composition under different phytoplankton regismens. Appl Envir Microb 70: 6753-6766

Pinhassi J, Gómez-Consarnau L, Alonso-Sáez L, Sala M, Vidal M, Pedros-Alió C, Gasol JM (2006) Seasonal changes in bacterioplankton nutrient limitaion and their effects on bacterial community composition in the NW Mediterranean Sea. Aquat Microb Ecol 44: 241-252

Prodi F & Fea G (1979). A case of transport and depsition of Sarna dust over the Italian peininsula and southern Europe. J.Gesophys. Res. 84: 6951-6960

Prospero JM, Barrett K, et al. (1996). Atmospheric deposition of nutrients to the North Atlantic Basin. Biogeochemistry 35: 27-73.

Prospero JM & Lamb PJ (2003). African droughts and dust transport to the Caribbean: climate change implications. Science 302: 1024-1027

Pulido-Villena, E., T. Wagener, et al. (2008). Bacterial response to dust pulses in the western Mediterranean: Implications for carbon cycling in the oligotrophic ocean. Global Biogeochemical Cycles 22, doi:10.1029/2007GB003091.

Ridame C, & Guieu C (2002). Saharan input of phosphate to the oligotrophic water of the open western Mediterranean Sea. Limn Oceanogr 47: 856-869.

Ruiz, J, Macías, D., Peters, F. (2004) Turbulence increases the average settling velocity of phytoplankton cells. Proceedings of the National Academy of Science USA 101: 17720-17724.

Sánchez O, Gasol JM, Massana R, Mas J & Pedrós-Alió C (2007) Comparison of different denaturing gradient gel electrophoresis primer sets for the study of marine bacterioplankton communities. Appl Environ Microb 73: 5962-5962.

Schäfer H, et al. (2001) Microbial community dynamics in Mediterranean nutrient-enriched seawater mesocosms: changes in the genetic diversity of bacterial populations. FEMS Microbiology Ecology 34: 243-253.

Schauer M, Balague V, Pedros-Alio C & Massana R (2003) Seasonal changes in the taxonomic composition of bacterioplankton in a coastal oligotrophic system. Aquatic Microbial Ecology 31: 163-174.

Smith DC & Azam F (1992) A simple, economical method for measuring bacterial protein synthesis rates in seawater using 3H-leucine. Mar Microb Food Webs 6: 107-114.

148

Effects of Saharan dust on bacterial activity and community structure

Smith EM (1998) Coherence of microbial respiration rate and cell-speici c bacterial activity in a coastal planktonic community. Aquat Microb Ecol 16:27-35

Thingstad, T.F., Havskum, H., Zweifel, U.L., Berdalet, E., Sala, M.M. Peters, F., Alcaraz, M., Scharek, R., Perez, M., Jacquet, S., Fonnes Flaten, G.A., Dolan, J.R., Marrasé, C., Rassoulzadegan, F., Hagstrøm, Å., Vaulot, D. (2007) Ability of a “minimum” microbial food web model to reproduce response patterns observed in mesocosms manipulated with N and P, glucose, and Si. J. Mar Syst 64: 15-34.

Thingstad TF, Zweifel UL, et al. (1998). P limitation of heterotrophic bacteria and phytoplankton in the northwest Mediterranean. Limn Oceanogr 43: 88-94.

Thingstad TF, Pérez M, Pelegrí S, Dolan J, Rassoulzadegan F (1999) Trophic control of bacterial growth in microcosms containing a natural community from northwest Mediterranean surface water. Aquat Microb Ecol 18: 145-156

Thingstad, T. F., M. D. Krom, et al. (2005). Nature of Phosphorus Limitation in the Ultraoligotrophic Eastern Mediterranean. Science 309: 1068-1071.

van Hannen EJ, Zwart G, van Agterveld MP, Gons HJ, Ebert J & Laanbroek HJ (1999) Changes in bacterial and eukaryotic community structure after mass lysis of lamentous cyanobacteria associated with viruses. Appl Environ Microb 65: 795–801.

Wassmann P & Olli K 2004 (eds). Drainage basin nutrient inputs and eutrophication: an integrated approach. http://www.ut.ee/˜olli/eutr.

Yentsch CS & Menzel DW (1963) A method for the determination of phytoplankton chlorophyll and phaephytin by uorescene. Deep-Sea Res 10: 221-231

Transplant experiments demonstrate strong bottom-up control of prokaryote abundance, heterotrophic production and bacterial community structure in the NE Atlantic Ocean

Itziar Lekunberri, Federico Baltar, Alejandra Calvo-Díaz, Eva Teira, Javier Arístegui and Josep M. Gasol

CHAPTER 5

The purpose of science is to create opportunities for unpredictable things to happen.

Freeman Dyson

151

Chapter V

ABSTRACT

Two transplant experiments between different depths were conducted in the eastern North Atlantic

Ocean to test whether predators or nutrient supply were more relevant in determining bacterial biomass,

production and community structure. To attain this goal, un ltered and 0.8 µm- ltered water from

surface or from deeper layers (either mesopelagic or DCM) were placed in dialysis bags. These bags

were incubated in surface and depth conditions tanks (in tanks containing surface or deep water and

the corresponding temperature). The rst experiment, took place during the RODA cruise in August

2006 south of the Canary Islands (26º50´N, 16º18´W), and the second during the CARPOS cruise in

October 2006 near Madeira (31º50´N, 16º40´W). Dialysis bags of 12000-14000 Da pore size were used

to maintain the microbial communities and let them be in uenced by the surrounding environmental

conditions (allowing the free exchange of nutrient salts and dissolved organic matter). We did not

observe any strong effect of predators in any of the different conditions. Bacterial activity increased

faster when incubated at surface than at depth conditions. Signi cant differences were found between

incubation conditions in the bacterial community structure analyzed in the rst experiment. Together,

our results indicate that bottom-up forces are much more important than top-down regulation in order

to determine bacterial community structure and function, at least in the NE Atlantic Ocean.

152

Regulation of bacterial activity and community structure

INTRODUCTION

The microbial loop hypothesis postulates a major role for bacteria in the oceanic carbon cycle

(Pomeroy 1974, Williams 1981, Azam et al. 1983), actively growing through the consumption

of DOC released by primary producers, C of allochthonous origin, and C originated from trophic

interactions between plankton organisms (e.g. Nagata 2000). Bacterioplankton can consume 30-60%

of the primary production via dissolved organic matter (Williams 1981, Azam et al. 1983, Cole et al.

1988), reaching close to 100% in very oligotrophic sites (Gasol et al. submitted). Bacterial biomass

often comprises a major fraction of particulate organic carbon, in oligotrophic systems at times even

surpassing phytoplankton biomass (Cho & Azam 1988, Fuhrman et al. 1989, Simon et al. 1992, Gasol

et al. 1997).

Depending on the speci c characteristics of each environment, factors that may control prokaryote

communities include temperature (Shiah & Ducklow 1994), grazing pressure (Thingstad et al. 1997),

and availability of limiting nutrients (Rivkin & Anderson 1997, Thingstad et al. 1997). Bacterial

production is considered to be controlled by the rate of nutrient supply (bottom-up), whereas nal

standing stocks and the rates of speci c growth are considered to be determined by predation pressure

(top-down), by substrate supply (bottom-up), or by both mechanisms (Wright & Cof n 1984, Pace &

Cole 1994, Thingstad & Lignell 1997). Thus, more nutrients would mean more biomass and production,

and more predators could mean less biomass or activity. Grazing is also known to be a selective factor

determining bacterial community structure. Predators prefer to predate the largest bacteria (Jürgens &

Güde 1994), the most active cells (del Giorgio et al 1996), and prefer some speci c bacterial groups

over others (Simek et al 1997, Jürgens et al 1999, Suzuki 1999).

It is generally accepted that bacterial communities seem to be rather similar at relatively large

spatial scales on the horizontal axis (Acinas et al. 1997, Baldwin et al. 2005), except where different

water masses meet (Suzuki et al. 2001, Pinhassi et al. 2003). Since the largest distinction in microbial

communities occurs at the boundary of the euphotic and aphotic zones, light and the vertical distribution

of autotrophs seem to be an important factor determining bacterial distributions, through their effects on

the DOC pool, although vertical clines in the composition of DOC are as yet poorly resolved (Carlson

et al. 2002). Furthermore, temperature, nitrogen and phosphorus availability, vary dramatically on the

vertical axis, so these factors must be considered together to understand the vertical strati cation of

microbial communities.

153

Chapter V

The assessment and comprehension of the factors that control bacterial abundance and production

will facilitate the understanding of how carbon and nutrients circulate in planktonic food webs.

Because of the dif culty of naturally replicating nutrient supply in mesocosm, transplant experiments

are a research approach of choice for the study of these factors (e.g. Gasol et al. 2002).

Here we describe two transplant experiments designed to assess the factors that control prokaryote

abundance, activity, and bacterial community composition at two different depths in the eastern North

Atlantic Ocean. To assess these questions, un ltered and 0.8 µm- ltered water from surface or from

depth (deeper layers, either mesopelagic or DCM) were placed in dialysis bags.. We used dialysis

bags that are sort of “microbial cages” that allow exposure of the microbes to ambient nutrient

concentrations (Herndl et al. 1993). Samples from the two different depths were incubated in dialysis

bags placed in times of in situ conditions and transplanted to another tank with the conditions of the

other depth. The experiments lasted 6-8 days and the bags were sampled at 24 h intervals to estimate

bacterial abundance and whole community activity.

MATERIALS & METHODS

Study site and water collection. We present data from two experiments conducted in the North

Atlantic Ocean in August and October 2006. In the August experiment the water was collected during

the RODA cruise near the Canary Islands (26º50´N, 16º18´W), and in the October experiment the

water was collected during the CARPOS cruise near the Island of Madeira (31º50´ N, 16º40´ W). In

the August experiment, the water used in the experiments came from the surface (ca. 5 m) and the

mesopelagic zone (500 m). In October the water for the experiments was collected from the surface

(ca. 5 m) and from the depth of the DCM (deep chlorophyll maximum), which was located at 100 m.

In both cruises the water was collected with Niskin bottles and the parameters of temperature, salinity

and uorescence in each station were recorded by a SeaBird 911 CTD system.

Transplant experiments. The two experiments started by collecting surface and deep water (from

the mesopelagic in the August experiment, and from the DCM in the October experiment). Two acid

clean 50 L tanks were lled up with water one from each different depth and placed to incubate in the

dark at in situ temperatures. The temperature conditions wer simulated with running super cial water

for the super cial water conditions tank and with an incubator for the depth water-conditions tanks

(DCM or mesopelagic). In both tanks 4 dialysis bags (of 1 L each) were placed with un ltered water

from the two different depths and with water that was gently pre- ltered through 0.8 µm polycarbonate

154

Regulation of bacterial activity and community structure

lters, to simultaneously address the effects of grazers and of changes in the natural conditions on the

bacterial communities. The dialysis bags (Spectrum laboratories), with a cut-off size of 12000-14000

Da and a maximal width of ~ 8 cm, were cut in lengths of 50 cm to hold ca, 1000 ml. The bags were

thoroughly washed in HCl 4% overnight, and then soaked for >3h in Milli-Q water before use. After

they were lled, they were closed and placed in the two tanks, the one holding surface water and the

other one having deep water. Samples for bacterial abundance and bacterial heterotrophic production

were taken daily from the dialysis bags, making sure not to contaminate the inside of the bags with

surrounding water, and DNA to analyze bacterial community structure was sampled at the beginning

and at the end of the experiments. Unfortunately, the DNA samples of the CARPOS cruise experiment

were lost during transport.

Environmental variables. Measurements of chlorophyll a (Chla) and nutrients were done in both

cruises (Table 1). Particulate organic carbon (POC) and particulate organic nitrogen (PON) were

measured only in the RODA cruise. Organic carbon was determined with a CHN elemental analyzer

after ltration through Whatman GF/F lters (POC) or catalytic combustion at high temperature of

un ltered and acidi ed samples (TOC). Dissolved organic carbon (DOC) was computed as DOC =

TOC-POC. The procedures employed are detailed in previous studies describing the variability of

these parameters in the study area (Alonso-Sáez et al. 2007, Arístegui et al. 1997).

Prokaryote abundance and biomass. Heterotrophic prokaryotes were measured by ow

cytometry as described in Gasol & del Giorgio (2000). The samples were run in a Becton-Dickinson

FACSCalibur cytometer after staining with SybrGreen (1:10,000 nal concentration, Molecular

Probes), and bacteria were detected by their signature in a plot of side scatter (SCC) vs FL1 (green

uorescence). Calibration of the cytometer, and of the uorescence-size relationship was done as

described in Calvo-Díaz & Morán (2006).

Prokaryotic heterotrophic production. Prokaryotic heterotrophic production was estimated

using the 3H-leucine incorporation method (Kirchman et al., 1985). Quadruplicate aliquots of 1.2 ml

were taken for each sample plus two TCA killed controls. The Leu tracer was used at 40 nM nal

concentrations in incubations lasting 2-3 h. The incorporation was stopped with the addition of 120 µl

of cold 50% TCA to the samples and, after mixing, were kept frozen at -20ºC until processing, which

was carried out by the centrifugation method of Smith & Azam (1992). The samples were counted on

a Beckman scintillation counter, 24 h after addition of 1 ml of scintillation cocktail.

155

Chapter V

Bacterial community structure. Microbial biomass was collected by ltering around 250-500 ml

of seawater through 0.2 µm polycarbonate lter (Millipore, 25 mm). The lters were stored at -80ºC.

Microbial biomass was treated with lysis buffer (50 mM Tris-HCl pH 8.3, 40 mM EDTA pH 8.0, 0.75

M sucrose), lysozime, proteinase K and sodium dodecyl sulphate. Nucleic acids were extracted by a

standard protocol using phenol/chloroform (see details in Schauer et al. 2003).

Fingerprinting analysis. DGGE and gel analysis were performed essentially as described (Schauer

et al. 2000; Sánchez et al., 2007). Brie y, 16S rRNA gene fragments (around 550 bp in length) were

ampli ed by PCR using the universal primer 907rm and the bacterial speci c primer 358f, with a

GC-clamp. PCR products were loaded on a 6% polyacrylamide gel with a DNA-denaturant gradient

ranging from 40-80%. The gel was run at 100V for 16 h at 60ºC in 1x TAE running buffer. DGGE gel

images were analyzed using the Diversity Database software (BIO-RAD).

Statistical analysis. A matrix was constructed for all DGGE lanes taking into account the relative

contribution of each band (in percentage) to the total intensity of the lane. Based on this matrix,

we obtained a dendrogram by the Ward’s clustering method (Euclidean distances, Statistica 6.0),

and ordinations of nonmetric multidimensional scaling based on Bray-Curtis similarity index (MDS,

Kruskal & Wish 1979, Clarke & Green 1988). An statistic test (one way ANOSIM) was used to test

for differences between treatments and the control samples, and another to contrast differences among

time within treatments (Primer v5).

156

Regulation of bacterial activity and community structure

0

50

100

150

200

250

30014 16 18 20 22 24 26

193.7 200 206.2 212.5 218.7 225 231.2

Oxigen (�mol Kg-1)

Salinity (PSU)

Temperature (ºC)

0 0.1 0.2 0.3 0.4 0.5

Fluorescence (vol)

36 36.2 36.4 36.6 36.8 37

0

200

400

600

800

10005 10 15 20 25

Temperature (ºC)

120 140 160 180 200 220 240 260

Oxigen (�mol Kg-1)34 34.5 35 35.5 36 36.5 37

Salinity (PSU)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7Fluorescence(vol)

A B

RESULTS

As expected, we used for our experiments very

different source waters (Fig. 2 and Table 1). In cruise

RODA, we compared surface with mesopelagic water.

This last one was colder (10 degrees lower), nutrient-

replete, and with half the concentration of bacteria than

to the surface. It was also 20x less active than the surface

community (Table 1). Stn E01B of cruise CARPOS was

relatively richer (surface Chla was 0.18 as compared

to < 0.05 mg m-3 in Stn 64 of cruise RODA), and we

chose to perform the transplant experiment with water

of the deep chlorophyll maximum (DCM), which was

very well formed and which reached 0.46 mg m-3 of chlorophyll (Fig. 2). For the interpretation of the

results it has to be taken into account that the incubations of the experiments were carried out at dark

conditions and at in situ temperature. Since the photosynthetic processes were suppressed, this could

affect the original DOC quality during the incubation.

Fig.1.- Map to the sampling area showing the two station locations.

Fig.2.- CTD pro les of water temperature (red), oxygen (blue), uorescence (green) and salinity (black) of the stations in RODA (A) and CARPOS (B) from where the experiments were stablished. Arrows indicate the selected depths.

S

O

T

F

S

TO

F

157

Chapter V

Bacterial production was 2 orders of magnitude higher at the DCM than at the surface, while

bacterial abundance was higher at the surface and, thus, bacterial growth rates were even higher at

depth. That the surface community of RODA was more active than that of the mesopelagic, and that

the DCM community of CARPOS was more active than that of the surface was also evidenced by the

%HNA cells (Table 1).

Bacte

rial a

bund

ance

(cell

ml-1

)Ba

cteria

l abu

ndan

ce (c

ell m

l-1)

Bacterial abundance

Time (days)

Leuc

ine i

ncor

pora

tion (

pmol

Leu

L-1 h-1

)

Leucine incorporation

Leuc

ine i

ncor

pora

tion (

pmol

Leu

L-1 h

-1)

Time (days)

MESOPELAGIC MESOPELAGIC

SURFACE SURFACE

A B

DC104

105

106

107

0 1 2 3 4 5 6

Super CTLS s 0.8 S s SFS m 0.8S m SF

104

105

106

107

0 1 2 3 4 5 6

Meso CTLM 0.8 mM SF mM 0.8 sM SF s

1

10

100

1000

104

0 1 2 3 4 5 6

0,1

1

10

100

1000

104

0 1 2 3 4 5 6

Fig.3.- Estimates of bacterial abundance (left panels, BA) and leucine incorporation rates (right panels, LIR) in the RODA mesopelagic-epipelagic transplant experiment. The upper panels (A, B) correspond to the surface water. The lower panels to the mesopelagic water. In both panels, the control of the tank (outside the bags) is referred with the black round symbol and a continuous line. Water incubated in bags placed in the same type of water are labelled with white symbols (this is Surface water incubated in surface water, or S s, and Mesopelagic water incubated in mesopelagic water, or M m). Transplanted water (S m and M s), stands with dark symbols. Water ltered through 0.8 µm (w/o predators) is plotted with round symbols (e.g. S s 0.8), while untouched water is referred as UF and is plotted with round symbols.

158

Regulation of bacterial activity and community structure

Both, average (and maximal) bacterial abundances were strongly determined by the characteristics

of the water where the samples were incubated in the transplant experiment done with meso- and

epipelagic waters (cruise RODA experiment, Table 2). For the surface water (Fig. 3A) there was

slightly more growth in the incubations at the surface (S s) than in the mesopelagic water (S m).

105

106

107

0 1 2 3 4 5 6

Bacterial abundance S controlS s 0.8S s UFS dcm 0.8S dcm UF

0.1

1

10

100

1000

104

1 2 3 4 5 6

Leucine incorporation

105

106

107

0 1 2 3 4 5 6

DCM controlDCM dcm 0.8DCM dcm UFDCM s 0.8DCM s UF

1

10

100

1000

104

0 1 2 3 4 5 6

0

Time (days) Time (days)

DCM DCM

SURFACEBacte

rial a

bund

ance

(cell

ml-1

)Ba

cteria

l abu

ndan

ce (c

ell m

l-1)

Leuc

ine i

ncor

pora

tion

(pm

ol L

eu L

-1 h-1

)Le

ucin

e inc

orpo

ratio

n (pm

ol L

eu L

-1 h-1

)

SURFACE

A

C D

B

Fig.4.- Estimates of bacterial abundance (left panels, BA) and leucine incorporation rates (right panels, LIR) in the CARPOS DCM-epipelagic transplant experiment. The upper panels (A, B) correspond to the surface water. The lower panels to the DCM water. In both panels, the control of the tank (outside the bags) is referred with the black round symbol and a continuous line. Water incubated in bags placed in the same type of water are labelled with white symbols (this is Surface water incubated in surface water, or S s, and DCM water incubated in mesopelagic water, or DCM dcm). Transplanted water (S dcm and DCM s), stands with dark symbols. Water ltered through 0.8 µm (w/o predators) is plotted with round symbols (e.g. S s 0.8), while untouched water is

referred as UF and is plotted with round symbols.

159

Chapter V

Treatment S-s-UF (un ltered) was rather similar to what happened outside the bags, and there

was little difference between the un ltered and the ltered samples in bacterial abundance, but the

statistical tests show signi cant effects of ltration on bacterial activity (Table 2, Fig. 5B). In terms

of leucine incorporation (Fig 3B), all samples started incorporation at about the same day (day 2), but

samples incubated in the mesopelagic had an initial lowering, an then they recovered and increased

leucine incorporation at a rate very similar to that of the water incubated at the surface, in spite of the

difference in incubation temperature. Bacterial activity was slightly higher in the un ltered than in

ltered samples. Some more differences were visible when we incubated water originated from the

mesopelagic. Bacteria developed strongly when incubated in surface water (Fig. 3C), reaching 4 106

cells ml-1 from a start of 1.5 105 cells ml-1 in 6 days (growth rate of 4.3 d-1). The un ltered sample

developed a higher yield (Fig. 5) than the ltered sample, suggesting that ltration through 0.8 µm

retained some growing bacteria, probably those attached to particles. The rate of growth of the ltered

and un ltered bacteria was roughly the same. When incubated in the mesopelagic, however, there was

a signi cant difference between the samples that had been ltered through 0.8 µm, which did not

develop until day 5, and those which were un ltered, that developed similarly as the control external

water (Fig. 3C). Leucine incorporation rates started to be higher earlier in the samples incubated at

a higher temperature (i.e. samples M s, Fig 3D). Paired Student’s t-tests indicate that the maximum

abundance was reached when incubated at the surface, while the maximum production was achieved

in the presence of predators (Table 2).

In the second experiment, with transplanted water from the surface and the depth of the deep

chlorophyll maximum (DCM), the starting situation was different: the DCM sample was richer and

more active than the surface sample. Surface water incubated at the DCM reached higher abundances

than incubated at the surface (Fig. 4A, Fig 5C). The control water was similar to the S-s-UF sample.

In terms of activity, however, values were rather similar between treatments with a systematic trend of

less production in the UF sample than in the F sample (Table 2, Fig. 5). DCM waters presented a very

similar pattern: they developed more when incubated in DCM water, while activity was rather similar

between treatments but systematically lower in the UF sample than in the F treatment.

160

Regulation of bacterial activity and community structure

Fig. 5 presents two integrative variables of what happened in the experiment: the average abundance

of bacteria in each treatment, and the total integrated bacterial production. The analysis in Table 2

includes two extra variables, the maximal abundance of bacteria and the maximal leucine incorporation

rates: these variables show similar trends than the average abundance and the integrated production.

Fig. 5 clearly shows the effect on the microbial community due to the incubation water conditions

(more abundance and more production in the samples incubated in the surface in RODA ), and a

slight, but signi cant effect of predators, mainly on activity.

Fig.5.- Left-side panels, maximum bacterial yield in each treatment of experiment RODA (A) and CARPOS (C), and integrated leucine incorporation (right-side panels) through the whole of the experiment RODA (B) and CARPOS (D). Letters indicate the source water, the incubation water, and the treatment (e.g. M s 0.8 indicates mesopelagic water incubated at surface and after ltration through 0.8 µm).

0

5 105

1 106

1,5 106

2 106

2,5 106

3 106

SUPE

R

S s

0.8

S s

UF

S m

0.8

S m

UF

MES

O

M m

0.8

M m

UF

M s

0.8

M s

UF

0

5 105

1 106

1,5 106

2 106

SUPE

R

S s

0.8

S s

UF

S m

0.8

S m

UF

MES

O

M m

0.8

M m

UF

M s

0.8

M s

UF

RODA

CARPOS CARPOS

0

1000

2000

3000

4000

SUPE

R

S s

0.8

S s

UF

S m

0.8

S m

UF

MES

O

M m

0.8

M m

UF

M s

0.8

M s

UF

0

2000

4000

6000

8000

1 104

SUPE

R

S s

0.8

S s

UF

S m

0.8

S m

UF

MES

O

M m

0.8

M m

UF

M s

0.8

M s

UF

RODA

Aver

age b

acter

ial ab

unda

nce (

cells

ml-1

)Av

erag

e bac

terial

abun

danc

e (ce

lls m

l-1)

Integ

rated

leuc

ine i

ncor

pora

tion (

pM)

Integ

rated

leuc

ine i

ncor

pora

tion (

pM)

A

C

B

D

161

Chapter V

In the RODA experiment, we analyzed community structure by PCR-16S rRNA DGGE

ngerprinting. The Ward’s clustering method showed a separation between the tanks, all samples

incubated at surface conditions clustered together and all samples incubated in Mesopelagic conditions

grouped together (Fig. 6a). The ordination of the samples by NMDS agrees well with the clustering

results, and shows separation of the samples incubated at the different depths (Fig. 6b). The ANOSIM

analysis determined signi cant differences between initial and nal time of incubations (R=0.915,

P:0.015) and incubation conditions (R=1, P: 0.008).

Ward`s methodCity-block (Manhattan) distances

0 50 100 150 200 250 300 350 400

Linkage Distance

MESO Tf

MmUF

SmUF

Mm0.8

Sm0.8

SUPER Tf

MsUF

Ms0.8

SsUF

Ss0.8

MESO To

SUPER To

DGGE gelDendogram

Resemblance: S17 Bray Curtis similarity

SUPER To

Ss0.8

SsUF

Ms0.8MsUF

SUPER Tf

MESOTo

Sm0.8

SmUF

Mm0.8

MmUFMESO Tf

2D Stress: 0.08

A B

C

SUPER Ss0.8 SsUF Ms0.8 MsUF SUPER MESO Sm0.8 SmUF Mm0.8 MmUF MESO To Tf To Tf

Fig.6.- DGGE gel of the RODA experiment (A). Dendogram classi cation (B)(Ward’s Method, Euclidean distances according to the band pattern). S (water from surface origin), M (water form mesopelagic origin), s (incubation conditions: surface tank), m (incubation conditions: mesopelagic tank). 0.8 (sample ltrated by 0.8 µm) and UF (sample without ltration). Nonmetric multidimensional scaling (NMDS) (C) of the bacterial assemblage composition data (DGGE band pattern). The samples have been clustered in agreement with the Ward’s clustering methods results.

162

Regulation of bacterial activity and community structure

DISCUSSION

Nutrient, light and temperature conditions vary dramatically on the vertical axis in the ocean.

Bacterial function, particularly bacterial production, varies also dramatically with depth (with

exponential slopes of -0.90 km-1, Arístegui et al., submitted). Bacterial community structure is also

known to vary dramatically on the vertical axis, more so than in the horizontal plane (Hewson et al.

2006). It is thus interesting to evaluate the effect that an input of surface waters could have in the DCM

or mesopelagic zone, as it would occur in an anticyclonic eddy. In this situation DOC accumulated

in the surface water masses could be available for the upper mesopelagic microorganisms. Carlson

et al. (2004) showed that DOC that could not be degraded by the surface bacterial community could

be easily remineralized at depth. Similarly, we were interested in predicting the effects caused by a

cyclonic eddy, when deep water masses would move upwards to the surface layer carrying inorganic

nutrients.

The objective of this study was to know to which extent bacterioplankton communities were

determined by its original conditions (e.g. nutrient concentration, abiotic factors), depending on

the depth at which the community had been formed. Hence, we hypothesized that a change in the

conditions should reveal the factor that determines a speci c level of bacterial metabolism and a

speci c community structure. As laboratory experiments have the dif culty of mimicking natural

nutrient conditions, to address this question we performed a transplant experiment using dialysis bags.

We used the concept of a “transplant” experiment, because in this approach organisms are allowed to

grow at different environmental conditions from their origin, sort of in a “microbial cage” (Herndl et

al. 1993).

In addition, we wanted to test whether the growth was nutrient controlled (“bottom-up”) and/

or biomass predator-controlled (“top-down”). A complex interaction exists between the top-down

and bottom-up controls of bacteria in plankton environments. Bacterial production is thought to be

regulated by the supply of nutrients and nal abundances are either determined both by predation-

pressure and substrate supply (Wright & Cof n 1984) or just by predators (Thingstad & Lignell 1997).

However in compiling the results of several experimental and empirical studies, Pace & Cole (1994)

found little evidence of systems where predators were limiting bacterial abundances. Similarly, we

found little evidence of abundances being determined by the presence or absence of predators (Table

2).

163

Chapter V

Table 2.- Results of paired Student’s t-tests (tests whether the average difference is signi cantly different from 0) between each paired treatment in the experiments: source water (e.g. surface vs. mesopelagic water), Incubation water, or presence/absence of predators. The variables analyzed are Maximum and experiment-average bacterial abundance, and integrated and maximum leucine incorporation rates. The trend explains whether the Surface (S) value is higher than the mesopelagic (M) or DCM value, or whether the 0.8 µm ltered (F) is higher or lower than the ltered (F) value. The ratio is the average increase in the variable within the experiments (Surface/Mesopelagic, UnFiltered/Filtered or Surface/DCM).

Exp. Variable Effect Prob. Trend Ratio

RODA Max BA Source 0.010 S > M 1.8 ± 0.5 Incubation 0.009 S > M 1.7 ± 0.3 Predators 0.009 UF > F 1.6 ± 0.3 Average BA Source 0.014 S > M 2.7 ± 1.1 Incubation 0.032 S > M 2.2 ± 1.1 Predators 0.041 UF > F 1.6 ± 0.4 Integrated LIR Source 0.055 S > M 1.7 ± 0.5 Incubation NS 1.0 ± 0.3 Predators 0.051 UF > F 0.7 ± 0.2 Max LIR Source NS 1.5 ± 0.7 Incubation 0.058 M > S 0.7 ± 0.3 Predators 0.086 UF > F 0.6 ± 0.3

CARPOS Max BA Source NS 1.2 ± 0.6 Incubation NS 1.3 ± 0.4 Predators NS 1.2 ± 0.4 Average BA Source NS 0.9 ± 0.4 Incubation NS 1.1 ± 0.4 Predators NS 1.0 ± 0.3 Integrated LIR Source NS 1.1 ± 0.1 Incubation NS 1.0 ± 0.1 Predators 0.019 UF > F 1.2 ± 0.1 Max LIR Source NS 1.0 ± 0.2 Incubation 0.002 DCM > S 0.7 ± 0.1 Predators 0.048 UF > F 1.1 ± 0.7

164

Regulation of bacterial activity and community structure

In both of our experiments, the bags which should be similar to the control –the incubation

tank itself, (e.g. the S-s-UF sample)- showed a similar pattern in bacterial production and bacterial

abundance, thus suggesting that the evolution of bacteria in the bags was not affected further than the

global effect caused by the con nement in the tank. As a general pattern, there was a clear tendency

for the bags with the same origin (e.g. surface water) to diverge depending on whether they were

transplanted or incubated in the same water. Furthermore, when e.g. surface water was transplanted

into more oligotrophic deep (mesopelagic), growth was lower than in the control, while when the

surface water was transplanted into more active (DCM) water, growth was enhanced. In general, the

effect of predators was small, and mainly affected leucine incorporation (Fig. 3-5) The abundance was

higher in the treatments without ltration that is contrary to what we should expect, an explanation

would be that the hot-spots of bacterial activity were excluded in the ltration process.

Grazing, particularly by heterotrophic nano agellates (HNF), has been identi ed as the main

bacterial loss factor (Fenchel 1982, Sherr & Sherr 1984, Wright & Cof n 1984, Pace et al. 1988) and

has been considered to balance more or less bacterial production. In a nutrient addition experiment

with waters of Blanes Bay, Allers et al. (2007) detected a shift in bacterial diversity after an increase

in agellate abundance in nutrient-enriched mescosms. Thus, this suggests that nutrient supply may

favour a bacterial increase which would be followed by a agellate increase, which will act as a

selective force on bacterial community structure (Pernthaler 2005). However, as we did not nd any

major differences between bags with or without predators, the bottom-up factors seemed to explain

(P< 0.01) differences in community structure (Fig. 6). Previous studies have reported that succession

within the bacterioplankton in coastal marine systems may be lead by autochthonous processes, or by

the input of certain labile organic substrates (nutrient-rich runoff from rivers or organic contamination,

Revilla et al. 2000, Schendel et al 2004). This may stimulate the growth of bacteria from particular

phylogenetic lineages (e.g. fast-growing-gammaproteobacteria, Pinhassi & Berman 2003). Our data

seem to suggest the existence of a shift in bacterioplankton community structure due to the waters in

which the samples are incubated.

Gasol et al. (2002) run a transplant experiment in a canyon-shaped reservoir, submitting bacterial

communities from the river side to the conditions in the reservoir side. Even with very different growth

rates for different bacterial groups, the nal community structure was quite similar to the initial one. In

contrast, in our experiments we found signi cant differences in the community composition between

incubation conditions revealing bottom-up effect on the community structure. This agrees with the

results of Carlson et al. (2004), that suggests controls on the bacterial community in space and time

during transplant experiments.Fuhrman et al. (2007) showed that the presence of annually reoccurring

165

Chapter V

bacterial communities in the Southern California coast, and that the occurrence of each population

could be predicted from the ocean environmental conditions. This would imply a large degree of

bottom-up control of bacterial community structure, as predators tend to vary more stochastically

than the nutrient sources (Tanaka & Taniguchi 1999) that are most likely driven by temperature,

light availability, rain and its direct effects on bacteria, or through effects mediated by phytoplankton

species succession, all parameters that we consider “bottom-up”. Our results show experimentally

this to be the case for surface, mesopelagic and DCM bacterial communities of the North Atlantic

Ocean.

ACKNOWLEDGEMENTS

We thank the chief scientist of cruise CARPOS, Dr. Ramiro Varela as well as the rest of coworkers

on both cruises for various help. We particularly thank A. Bode for ancillary data and X.A.G. Morán

for the opportunity to join cruise CARPOS. The eld work of this study was supported by projects

RODA (CTM 2004-06842-C03/MAR) and CARPOS (REN 2003-09532-C03-01), and the lab work

was supported by project MODIVUS (CTM2005-04795/MAR). This is a contribution to the EUR-

OCEANS and MARBEF European Networks of Excellence.

166

Regulation of bacterial activity and community structure

REFERENCESAcinas SG, Rodríguez-Valera F, Pedrós-Alió C (1997) Spatial and temporal variation in marine

bacterioplankton diversity as shown by RFLP ngerprinting of PCR ampli ed 16S rDNA. FEMS Microb Ecol 24: 27-40

Allers E, Gómez-Consarnau L, Pinhassi J, Gasol JM, Simek K & Pernthaler J (2007) Response of Alteromonadaceae and Rodobacteriaceae to glucose and phosphorus manipulation in marine mesocosms. Environ Microbiol 9: 2417-2429.

Alonso-Sáez L, Gasol JM, Arístegui J, Vilas JC, Vaqué D, Duarte CM & Agustí S (2007) Larg-scale variability in surface bacterial carbon demand and growth ef ciency in the subtropical North East Atlantic Ocean. Limn Oceanogr 52:533-546.

Arístegui J, Tett, A. Hernández-Guerra, G. Basterretxea, M. F. Montero, K. Wild, P. Sangrá, S. Hernández-Leon, M. Canton, J. A. García-Braun, M. Pacheco, E. D. Barton. (1997) The in uence of island-generated eddies on chlorophyll distribution: a study of mesoscale variation around Gran Canaria. Deep-Sea Research Part I: Oceanogr Res Papers 44: 71-96.

Azam F, Fenchel T, Field JG, Gray JS, Meyerreil LA & Thingstad F (1983) The Ecological Role of Water-Column Microbes in the Sea. Marine Ecology-Progress Series 10: 257-263.

Azam F & Worden AZ (2004) Oceanography: microbes, molecules, and marine ecosystems. Science 303: 1622-1624.

Baldwin AJ, Moss JA, Pakulski JD, Catala P, Joux F, Jeffrey WH (2005) Microbial diversity in a Paci c Ocean transect from the Arctic to Antarctic circles. Aquat Microb Ecol 41:91-102

Barton ED & Arístegui J (1998) The transition Zone of the Canary Current. Progr Oceanogr 41: 455-504.

Bode A, Barquero S, González N, Álvarez-Ossorio MT & Varela M (2004a) Cotnribution of heterotrophic plankton to nitrogen regeneration in the upwelling ecosystem of A Coruña (NW Spain). J. Plank Res 26: 1-18.

Bode A, Varela M, Teira E, Fernández E, González N & Varela M (2004) Planktonic carbon and nitrogen cycling off NW Spin: variations in productio of particulate and dissolved organic pools. Aquat Microb Ecol 37: 95-107

Bode A, González N, Rodríguez C, Varela M & Varela MM (2005) Seasonal variability of plankton blooms in the Ria de Ferrol (NW Spain): I. Nutrient concentrations and nitrogen uptake rates. Estuarine, Coastal Shelf Science 63: 269-284.

Carlson CA (2002) Production and removal proceses, p. 91-151. In D.A. Hansell and C.A. Carlson [eds.], Biogeochemistry of marine disolved organic matter. Academic Press.

Carlson CA, Giovannoni SJ, Hansell DA & Goldberg SJ (2004) Interactions among dissolved organic carbon, microbial processes, and community structure in the mesopelagic zone of the northwestern Sargasso Sea.Limn Oceanogr 49: 1073-1083.

Casas B, Varela M, Canle M, González N & Bode A (1997) Seasonal variations of nutrients, seston and phytoplankton, and upwelling intensity off La Coruña (NW Spain). Est, Coast Shelf Science 44: 767-778.

Calvo-Díaz A & Morán XAG (2006) Seasonal dynamics of picoplankton in shelf waters of the southern Bay of Biscay. Aquat Microb Ecol 42: 159-174.

167

Chapter V

Cho BC & Azam F (1988) Major role of bacteia in biogeochemical uxes in the ocean´s interior. Nature 332: 441-443.

Clarke KR& Green RH (1988) Statisitcal design and analysis for a biological effects study. Mar Ecol Progr Ser 46: 213-226.

Cole JJ, Finlay S & Pace ML (1988) Bacterial production in fresh and saltwater ecosystems: a cross-systme overview. Mar Ecol Progr Ser 43: 1-10.

Del Giorgio, Gasol JM, Vaqué D, Mura P, Agustí S, Duarte CM (1996) Bacterioplankton community structure: Protist control net production and the proportion of active bacteria in a coastal marine community. Limn Oceanogr 41: 1169-1179

DeLong EF (2005). Microbial community genomics in the ocean. Nat rev Microbiol 3: 459-468

Fenchel T (1982) Ecology of heterotrophic micro agellates.Quantitative occurrence and importance as consumers of bacteria. Marine Ecology Progress Series 9: 35-42.

Fuchs BM, Zubkov MV, Sahm K, Burkill PH & Amann R (2000) Changes in community composition during dilution cultures of marine bacterioplankton as assessed by ow cytometric and molecular bio.ogical techniques. Environ Microb 2: 191-201.

Fuhrman JA, Sleeter TD, Carlson CA & Proctor LM (1989) Dominance of bacterial biomass in the Sargasso Sea and its ecological implications. Mar Ecol Progr Ser 57: 207-217.

Fuhrman JA, Community Structure: Bacteria and Archaea. 2007. In: Christon J. Hurst, Ronald L. Crawford, Jay L. Garland, David A. Lipson, Aaron L. Mills, Linda D. Stetzenbach (eds.). Manual of Environmental Microbiology, 3rd edition. ASM Press, Washington, D.C. Section IV, Chap 35.

Fukuda R, Ogawa T, Nagata T & Koike I (1998) Direct determination of carbon and nitrogen contents of natural bacterial assemblages in marine environments. Appl Environ Microb 64: 3352-3358.

Gasol, J. M., del Giorgio, P.A., Duarte, C.M. (1997) Biomass distribution in marine planktonic communities. Limn Oceanog 42: 1353-1363.

Gasol, J.M., del Giorgio, P.A. (2000) Using ow cytometry for counting natural planktonic bacteria and understanding the structure of planktonic bacterial communities. Scientia Marina. 64: 197-224

Gasol, J.M., Comerma, M., García, J.C., Armengol, J., Casamayor, E.O., Kojecká, P., Simek, K. (2002) A transplant experiment to identify the factors controlling bacterial abundance, activity, production, and community composition in a eutrophic canyon-shaped reservoir Limn Oceanogr 47: 62-77.

Giovannoni SJ & Rappe M (2000) Evolution, diversity, and molecular ecology of marine prokaryotes. Microbial Ecology of the Oceans, Edited by Dvid I., Kirchman.

Hansell DA & Carlson CA (1998) Deep ocean gradients in dissolved organic carbon concentrations. Nature395: 263-266.

Hansell DA (2002) DOC in the global ocean carbon cycle. pp. 685-716 In: Hansell, D.A. and Carlson C.A.(eds.), Biogeochemistry of marine dissolved organic matter. Academic Press.

Herndl GJ, Kaltenböck E & Müller-Niklas G (1993) Dialysis bag incubation as a nonradiolabeling technique to estimate bacteriplankton production in situ. In. P.F. Kemp, B.F. Sherr, E.B. Sherr, and J.J. Cole (eds), Handbook of methods in aquatic microbial ecology. Lewis 553-556.

168

Regulation of bacterial activity and community structure

Hewson I & Fuhrman JA (2006) Spatial and vertical biogeography of coral ref. sediment bacterial and diazotroph communities. Mar Ecol Prog Ser 306: 79-8

Jürgens K, Pernthaler A, Schalla S & Amann R (1999) Morphological and compsitional changes in a planktonic bacterial community in response to enhanced protozoan grazing. Appl Environ Microb 65: 1241-1250.

Jürgens & Güde (1994) The potential importance of grazing resistant bacteria in planktonic systems. Mar. Ecol. Prog. Ser 112:169-188

Kepkay PE (1994) Particle aggregation and the biological reactivity of colloids. Mar Ecol Progr Ser 109: 293-304.

Kirchman, D.L., K´nees, E., and Hodson, R (1985) Leucine incorporation and its potential as a measure of protein synthesis by bacteria innatural aquatic ecosystems. Appl Environ Microb 49: 1681-1688

Kruskal JB& Wish M (1979) Multidimensional scaling. Wiley, Beverly Hills, CA.

Lebaron P, Servais P, Troussellier M, Courties C, Vives-Rego J & Muyzer G (1999) Changes in bacterial community structure in seawater mesocosms differing in their nutrient satatus. Aquat Microb Ecolo 19: 255-267

Massana R, Pedrós-Alió C, Casamayor EO & Gasol JM (2001) Changes in marine bacterioplankton phylogenetic composition during incubations designed to measure biogeochemically signi cant parameters. Limn Oceanogr 46: 1181-1188.

Morris RM, Rappé MS, Urbach E, Connon SA, Giovanonni SJ (2004) Prevalence of the Chloro exi-related SAR202 bacterioplankton cluster throughout the mesopelagic zone and deep ocean. Appl Environ Microb70: 2836-2842

Nagata T. (2000) Production mechanisms of dissolved organic matter. Microbial Ecology of the Oceans, Edited by Dvid I., Kirchman.

Pace ML (1988) Bacterial mortality and the fate of bacterial produc tion. Hydrobiologia 9: 41-49.

Pace MF & Cole JJ (1994) Comparative and experimental approaches to top-down an bottom-up regulation of bacteria. Microbial Ecology 28: 181-193.

Pedrós-Alió C (2006) Marine microbial diversity: can it be determined? Trends in Microbiology 14: 257-263.

Pernthaler J (2005) Predation on prokaryotes in the water column and its ecological implications. Nat Rev Microbial 3: 537-546

Pinhassi J & Berman T (2003) Differential Growth Response of Colony-Forming alpha- and gamma-proteobacteria in dilution culture and nutrient addition experiments from lake Kinneret (Israel), the eastern Mediterranean Sea, and the Gulf of Eilat. Appl Environ Microb 69: 199-211.

Pinhassi J, Winding A, Binnerup S, Zweifel UL, Riemann B, Hagström Å (2003) Spatial variability in bacterioplankton community composition at the Skagerrak-Kattegat front. Mar Ecol Prog Ser 255:1-13

Pomeroy LR (1974) The ocean’s food web, a changing paradigm. Bioscience 24: 499-504.

Pommier T, Canbäck L, Riemann L, Boström KH, Simu K, Lundberg P, Tunlid A & Hagström A (2007) Global patterns of diversity and community structure in marine bacterioplankton. Molecular Ecology.

Revilla M, Iriarte A, Madariaga I & Orive E (2000) Bacterial and phytoplankton dynamics along a trophic gradient in a shallow temperature estuary. Est, Coast Shelf Sci 50: 297-313.

169

Chapter V

Rush DB, et al. (2007) The Sorcerer II Global Ocean Sampling Expedition: Northwest Atlantic through Eastern Tropical Paci c. PLOS biology 5: 0398-0431.

Rivkin RB & Anderson MR (1997) Inorganic mutrient limitation of oceanic bacterioplankton. Limn Oceanogr 42: 730-740.

Sánchez O, Gasol JM, Massana R, Mas J & Pedrós-Alió C (2007) Comparison of different denaturing gradient gel electrophoresis primer sets for the study of marine bacterioplankton communities. Appl Environ Microb 73: 5962-5962.

Schauer M, Massana R & Pedrós-Alió C (2000) Spatial differences in bacterioplankton composition along the Catalan coast (NW Mediterranean) assessed by molecular ngerprinting. FEMS Microbiol Ecol 33: 51-59.

Schauer M, Balague V, Pedrós-Alió C & Massana R (2003) Seasonal changes in the taxonomic composition of bacterioplankton in a coastal oligotrophic system. Aquat Microb Ecol 31: 163-174.

Schendel EK, Nordström SE & Lavkulich LM (2004) Floc and sediment properties and their environmental distribution form a marine sh farm. Aquac Res 35: 483-493.

Selje N, Simon M, Brinkhoff T (2004) A newly discovered Roseobacter cluster in temperate and polar oceans. Nature 427:445-448

Sherr BF & Sherr EB (1984) Role of heterotrophic protozoa in carbon and energy glow in aquatic ecosystem. In: Klug. M.J., Reddy, C.A. (eds) Current perspectives in microbial ecology. American Society for Microbiology, Washington, DC. 412-423.

Shiah FK & Ducklow HW (1994) Temperature regulation speci c growht rate in Chesapeake Bay. Limn Oceanogr 39: 1243-1258.

Simek K, Bodkova J, Macek M, Nedoma J & Psenner R (1995) Ciliate grazing on picoplankton in a eutrophic reservoir during the summer phytoplankton maximum: a study at the species and community level. Limn Oceanogr 40.

Simek K, Vrba J, Pernthaler A, Posch T, Hartman P, Nedoma J & Psenner R (1997) Morphological and genotypic shifts in an experimental bacterial community in uenced by protists of contrasting feeding modes. Appl EnvironMicrob 63: 587-595.

Simon M, Cho BC & Azam F (1992) Signi cance of bacterial biomass inlakes and the ocean: comparison to phytoplankton biomass and biogeochemical implications. Mar Ecol Prog Ser 86: 103-110.

Smith DC & Azam F (1992) A simple, economical method for measuring bacterial protein synthesis rates in seawater using 3H-leucine. Mar Microb Food Webs 6: 107-114.

Suzuki MT, Preston CM, Chavez FP, DeLong EF (2001) Quantitative mapping of bacterioplankton populations in seawater: eld tests across an upwelling plume in Monterey Bay. Aquat Microb Ecol24:117-127

Suzuki MT (1999) Effect of protistan bacteriovory on coastal bacterioplankton diversity. Aquat Microb Ecol20: 261-272.

Tanaka T & Taniguchi A (1999) Predator-prey eddy in heterotrophic nano agellate-bacteria relationships in a bay on the northeastern Paci c coast of Japan. Mar Ecol Prog Ser 179: 123-134

Taylor GT, Iabichella M, Ho T-Y, Scranton MI, Thunell RC, Muller-Karger F, Varela R (2001) Chemoautotrophy in the redox transition zone of the Cariaco Basin: a signi cant mid-water source of organic carbon production. Limn Oceanogr 46: 148-16

170

Regulation of bacterial activity and community structure

Thingstad F & Lignell R (1997) Theoretical models for the control of bacterial growth rate, abundance, diversity and carbon demand. Aquat Microb Ecol 13: 19-27.

Thingstad F, Hagström A & Rassoulzadegan F (1997) Accumulation of degradable DOC in surface waters: Is it caused by a malfunctioning microbial loop? Limn Oceanogr 42: 398-404.

Varela MM, van Aken HM, Sintes E, Herndl GJ (2008) Latitudinal trends of Crenarchaeota and Bacteriain the meso- and bathypelagic water masses of the Eastern North Atlantic. Environ Microb doi:10.1111/j.1462-2920.2007.01437.x

Venter JC, et al. (2004) Environmental genome shotgun sequencing of the Sargasso Sea. Science 304: 66-74.

Williams PJ (1981) Incorporation of microheterotrophic processes into the classical paradigm of the planktonic food web. Kiel. Meeresforsch. 5: 1-28.

Wright RT & Cof n RB (1984) Factors affecting bacterioplankton density and productivity an salt marsh estuaries. In M.J. Klug and C.A. Reddy (eds). CUrrent perspectives in microbial ecology. Ame Soc Micr485-494.

SYNTHESIS AND DISCUSSION

La armonía total de este mundo está formada por una natural aglomeración de discordancias

Séneca

173

Synthesis and dicussion

SYNTHESIS AND DISCUSSION

The aim of this thesis was to describe the effects of different allochthonous carbon sources on the

diversity and function of marine bacterioplankton. In some of the chapters (#1 particularly), the focus

was on diversity, and in some others it was placed on the link between the changes in diversity and

those in bacterial function, in some cases extending beyond the purely bacterial function to that of

the whole microbial ecosystem (e.g. Chapter 4). In Chapter 1 we compared the diversity obtained

by traditional culture techniques and by molecular approaches in the NW Mediterranean. The culture

media can be considered a special type of allochthonous carbon source, and we encountered low

similitude between the bacteria present in the environment, and those selected by the media, suggesting

a strong effect of the resource types and levels on diversity The diversity detected by cultivation was

largely invisible by the molecular PCR-based techniques, and using both methodologies we increased

substantially the diversity reported from the area in previous papers. Coastal planktonic communities

are subject to the introduction of C and nutrients of anthropogenic origin. We analyzed the effect

of one of these sources of C and nutrients of basically human origin in Chapter 2, in which, to

determine the effects of the Prestige tanker oil spill tanker in front of the Galician coast, we used

realistic additions of oil to mesocosms. We detected little effects of oil on the diversity and function

of the authochtonous bacterial microbiota. To evaluate the effect of different nutrient inputs from

different origins (terrestrial or athmospheric), two enrichment experiments were carried out in the NW

Mediterranean coast. We added different sources to stimulate phytoplankton blooms and determined

changes in bacterial community structure, but little differences were found even when a diatom bloom

developed (Chapter 3), In the experiment that tried to mimick an event of Saharan dust deposition

(Chapter 4) we saw clear effects on bacterial function. In Chapter 5, however, and we investigated

the effects of different types of “nutrient environments” on diversity and function of epipelagic, and

mesopelagic bacterioplankton. We did so with transplant experiments, and we found that, in contrast

to what we had seen in the NW Mediterranean and partly in the Atlantic coast, the bottom-up factors

(nutrient environment) strongly determined the nal results of the experiments.

Several elements were common between the different experiments/chapters. For example, i) we

used a mesocosms approach in 3 of the chapters (2, 3 and 4). In these experiments we detected a

common pattern of development of bacterial abundance that was also seen in previously published

experiments. Below, we analyze the meaning and implications of these patterns, also as they have

to do with agellate development. ii) Two of the experiments/chapters were done in Blanes Bay

174

Synthesis and dicussion

(NW Mediterranean) waters, but started at different times of the year (chapters 3 and 4). Similarly,

in Chapter 2 we considered the variation in effects depending on the time of the year in which the

experiment was done. Also, iii) we used a genetic ngerprinting technique (DGGE) in all the chapters.

In Chapter 1 we compared the bands appearing in DGGEs done in Blanes Bay at different years with

the diversity seen by other techniques (clone libraries and isolation). In other chapters we analyzed

changes in bacterial community structure in the different experiments. This allowed us to compare the

“amount of change” in structure between different manipulations and in different seasonal studies.

In the following sections we explore some of these general issues that appear different times

throughout this thesis.

TRANSPLANTS OIL-SPILLSIMULATION SAHARAN DUST

DEPOSITIONPHYTOPLANKTON BLOOMS INDUCTION

+ REALISM -

- CONTROL +

Fig.1.- Summary of different approaches used in this thesis and the associated levels of control, inversely related to the realism levels. From the more realistic and less controlled, to the less realistic and more controlled: transplant experiments (Chapter 5), oil addition (Chapter 2), Saharan dust addition (Chapter 3) and inorganic nutrient addition and phytoplankton bloom induction (Chapter 4). Chapter 1 could somehow be included in the right side of the graph, as the culture media are well de ned, and we know well that provides unrealistic views of bacterial community structure.

175

Synthesis and dicussion

i) EXPERIMENTAL APPROCHES

To address the different questions posed in this thesis, we mainly used mesocosms and transplant

experiments using dialysis bags. These two kinds of approaches are very useful to experimentally

mimic natural situations. Mesocosms are a common tool used in aquatic ecology to simulate natural

behaviour of communities, at a scale that can be handled and providing a degree of realism not possible

in the laboratory (Odum 1984). They have been widely used to study the effect of nutrient additions

on phytoplankton communities (e.g. Duarte et al. 2000), on the activity and diversity of bacterial

assemblages (e.g. Schäfer et al. 2001), and on microbial succession (e.g. Allers et al. 2007). This type

of experimental approach avoids the effects caused by advection, diffusion and mixing under natural

conditions and also offers the possibility of comparison with natural populations where there were no

additions, thus allowing statistical testing of hypotheses. We have performed mesocosms experiments

to evaluate the effects of oil spills in Chapter 2, to stimulate different phytoplankton blooms Chapter

3 and to mimic a Saharan dust deposition in Chapter 4. And we even used in Chapter 5 mesocoms

approaches as a way to maintain the transplant experiment with the dialysis bags. These bags are

sort of microbial cages that allow nutrients to ow, and are thus, very useful to exposure a determine

bacterial community to a particular ambient nutrient concentration (Herndl et al. 1993).

To the enclosures we added different types of substrates, ranging from simple inorganic nutrients,

somehow uncharacterized dust originated from the Sahara, or the soluble fraction of Prestige oil.

Furthermore, in one case we applied a “black-box” approach to the substrates by incubating samples

from some depths into other depths that we expected should have different nutrient elds (DOC and

inorganic nutrients, but also temperature). Our target variables (bacterial diversity and function) and

our questions (do the target variables change with these additions?) were analyzed in all these different

types of approaches. They can be ordered in two types of opposite gradients (Fig. 1) depending on

the degree of control we can have over the nutrients added, or on the degree of realism that we can

assume each approach has. The dialysis bag experiments are probably more realistic than the other

approaches, but we control little the nutrient eld. In contrast, in the nutrient-addition experiments of

Chapter 3 we controlled well the nutrients added, but the situation was probably the less realistic one

and the one that shed less light into what really occurs in nature. Plate growth media as those used in

Chapter 1 are an extreme example of unrealistic, but well controlled, approach.

It is well known that enclosures can either enhance the growth of certain microzooplankton by

excluding their predators, or induce their mortality, and this generates the question of whether the

mesocosms can also affect the microbial communities. The Mediterranean sea is known to be an

oligotrophic ecosystem and, thus, it is likely that the so-called “bottle effect” (Marshall et al. 1971) is

more intense than in a relatively more eutrophic system such as Ría de Vigo (Chapter 2). These bottle-

effects, maybe related to the dominant P-limitation of Mediterranean Sea bacterioplankton (Thingstad

176

Synthesis and dicussion

et al. 1998, Pinhassi et al. 2006) seemed larger in the initial times of the experiments performed in

the Mediterranean. Figures 2-4 show a comparison between the changes recorded in the experiments

of Chapters 3 and 4 with two other experiments done with waters from the NW Mediterranean, and

published in Pinhassi et al. (2004) and Allers et al. (2007). These plots show that the initial start of

the these experiments presented shift-ups of 2-3 orders of magnitude in bacterial production, which

are much higher than the shift-ups seen in the Ría de Vigo mesocosms experiments (Fig. 3 in Chapter

2).

� ���

� �

� �

Fig.2.- Chlorophyll a values (mg m-3) along the incubation of four different mesocosms experiments with NW Mediterranean waters: (A) nutrient addition and turbulence experient (Pinhassi et al. 2004); (B) inorganic and organic addition experiment (Allers et al. 2007); (C) phytoplankton bloom stimulation experiment (Chapter 3) and (D) Saharan dust deposition experiment (Chapter 4). Treatments are K: Control, C: Carbon addition, P, Phosphorus addition, CP: both additions. Si: Silica and P addition, U: Urea and P addition. Low and High Dust are different ammounts of Saharan Dust.

177

Synthesis and dicussion

Some mesocosms experiments had been performed before to evaluate the changes of bacterial

community structure due to changes in nutrient availability and predators (e.g. Lebaron et al. 1999

and Schäfer et al. 2001). Previous to the studies in this thesis two mesocosms experiments were

performed with coastal waters of the NW Mediterranean (Pinhassi et al. 2004, Allers et al. 2007).

All of these experiments show differences between treatments and control in Chla concentration,

bacterial abundance and bacterial production (Fig 2, 3 and 4) that can be inspected further.

Hence, the utilization of mesocosms as an approach to study the effect of different nutrients additions

can be very useful. However, we also show, in the oil-spill simulation experiments in the Galician

coast (Chapter 2) that sometimes there may not be any bacterial response because, as shown by the

Fig.3.- Bacterial abundance (cells ml-1) values along the incubation of four different mesocosms experiments with NW Mediterranean waters. Panels as in Figure 1.

106

107

0 2 4 6 8 10

106

107

106

107

KLow DustHigh DustP

KPSiU

Allers et al. 2007Pinhassi et al. 2004

ANuts + TurbNuts - Turb

KCPCP B

0 2 4 6 8 10

DChapter 3 Chapter 4

Time (days) Time (days)

106

107

178

Synthesis and dicussion

concentration dependent experiment, the concentrations of the compound added in the experiment

were not high enough to generate a response. The complexity of mesocosms experiments implies that

not many experimental situations can be analyzed at once. However, the comparison (meta-analysis)

of similar mesocosms experiments performed in different environments or settings allows for an

increase in the signi cance of the conclusions over what would be possible to know from just an

experiment alone. It allows generalizing some conclusions to situations other than the one assayed,

but also allows seeing differences between experiments (Olsen et al. 2006).

These results indicate a bacterial community uesd to the oil concentration tested. From the transplant

experiments done in North Atlantic Ocean, the bags with the same origin as the incubation tank water

had the same pattern in bacterial abundance, bacterial production and community structure, suggesting

Fig.4.- Bacterial production (µgC l-1 d-1) along the incubation of four different mesocosms experiments with NW Mediterranean waters. Panels as in Figure 1.

1

10

100

0 2 4 6 8 101

10

100

0 2 4 6 8 10

1

10

100

1

10

100

0 2 4 6 8 10

KLow DustHigh DustP

KPSiU

Allers et al. 2007Pinhassi et al. 2004Nuts + TurbNuts - Turb

KCPCP

0 2 4 6 8 10

Chapter 3 Chapter 4

Time (days) Time (days)

Bacte

rial P

rodu

ction

(µg

C L-1 d

-1 )Ba

cteria

l Pro

ducti

on (µ

g C

L-1 d-1 )

179

Synthesis and dicussion

no effect of enclosure in the evolution of the bags. All together con rms that enclosure often did not

mask the effects on bacterial communities due to changes in the environment.

ii) THE STATUS OF THE INITIAL COMMUNITIES

Inspection of the effect of oil on plankton community structure, as we did in Chapter 2, running

similarly designed mesocosms experiments at different times of the seasonal cycle in the planktonic

ecosystem of Ría de Vigo allowed us to detect differences in the response of the microbial community

depending on the initial situation. In those experiments, we started a mesocosms with a decaying

phytoplankton community and another one with a blooming community. However, the community

that was more sensitive to oil seemed to be the summer one, something that we assigned to the fact

that the water in the Ría at htat time of the year comes mainly from the upwelling of oceanic deep

water that enters the Ría and fertilizes the planktonic community. The bacteria in these waters are

likely not to be naturally exposed to oil components nor to other allochthonous C sources and might

thus be more sensitive to stressors such as the polycyclic aromatic hydrocarbons added.

The response of the NW Mediterranean communities to nutrient additions was also varied

depending in part on the time of the year when the experiment was done, but also dependoing on

the experimental conditions used. For example, chlorophyll decreased in the control treatment of the

experiment reported in Fig. 2B (Allers et al. 2007), but this experiment was run with very low light.

Chlorophyll also decreased in the Saharan Dust addition experiment (Chapter 4), but this was due to

the fact that the starting community was the one with the largest Chla value (ca. 3 mg m-3), during a

natural bloom situation, and the experiment evolved during the decline of that bloom.

A few other things are interesting and can be seen in Figs. 2-5. For example, bacteria seem to

develop a small initial bloom (except in the Saharan Dust experiment), before there is any noticeable

development in Chla. This can be related to the very important P-limitation of bacteria in the area

(Pinhassi et al. 2006). The effect of P on bacteria (note that all treatments in Chapters 3 and 4 had

received P except the controls) is much stronger than this effect on phytoplankton, as can also be seen

comparing Fig. 2 (Chla) and Fig. 2 (bacterial abundance). Finally, the chlorophyll maximum yield in

these experiments is ca. 2 orders of magnitude (Chapter 3, addition of Si + P), while the max. bacterial

yield did not go over one order of magnitude (in the experiment reported by Pinhassi et al. 2004, and

in the addition of a high concentration of Saharan Dust). This can probably be traced back again to

the well known lack of proportionality between primary producers and bacteria when their natural

changes are compared through regression lines (Gasol & Duarte 2000), but can also be masked by the

fact that bacterial numbers might hide increases in bacterial cell size and, thus, in bacterial biomass.

180

Synthesis and dicussion

The role of agellates as grazers of bacteria is very apparent in these experiments. In almost all

of them bacteria present an initial peak before algal development, then there is a through, and a

second peak of bacteria. In one of these studies (Allers et al. 2007) we could identify most bacteria

developing in the rst peak as Alteromonadaceae, while most of the bacteria in the second peak were

Rhodobacteraceae. The rst group of sequences were much more homogeneous between treatments

than the second group which differed according to the treatment. In fact, it appears as if the initial

bacterial peak, particularly when P is added, is rather independent of the treatment, and only the second

bacterial peak has to do with the treatment applied (see Chapter 3). The rst bacteria developing

would be oportunistic phylotypes that would then be grazed down by agellates.

iii) MEASURING BACTERIAL COMMUNITY STRUCTURE WITH DGGE

For the comparison of bacterial community structure in the different treatments, we used the

ngerprinting technique DGGE (denaturing gradient gel electrophoresis, Muyzer et al. 1997). The

level of diversity detection of this method is quite low (<30 sequences) compared to other molecular

techniques such as clone libraries (which were used in chapter one with a target of 100 sequence

types per sample) and much lower compared with the newer shotgun techniques, which are able to

detect thousand of sequences. However, this is a very useful methodology for comparative analyses of

microbial communities and their response to changing environments. (Kasai et al. 2001, Labbé et al.

2007, Brakstad et al. 2006, Castle et al. 2006, Ogino et al. 2001, Macnaughton et al. 1999, Casamayor

et al. 2002, Sánchez et al. 2007).

In Chapter 1 we presented a comparative study of the potential of sequence detection by culturing

and culture-independent techniques. It was observed that the molecular techniques differed substantially

from the cultivation techniques and, therefore, it was the combination of all the techniques what

allowed to enlarge our view of the bacterial biodiversity present at the investigated location (Donachie

et al. 2007). In Figure 5 we show a review of the different studies presented in Donachie et al. (2007)

with the addition of the results of our study. It is a plot of the percentage of sequences retrieved by

clone libraries in comparison with the sequences from isolates and to the sequences detected by

both techniques. We found similar overlap as in the other marine study (Eilers et al. 2000), and non-

overlapping detection by the two sets of techniques, suggests that the isolate sequences do not represent

the diversity of the environment. The concept of biodiversity has been explained by Pedrós-Alió

(2006) as a pool of very abundant species and a tail (of unknown length) of rare/occasional species.

This author proposes a framework in which each type can be studied with different strategies. PCR

primers will predominantly hybridize and represent the common taxa, and the rare ones will only be

181

Synthesis and dicussion

retrieved by PCR-independent techniques such as cultivation or metagenomics techniques. Certainly,

there is an increasing interest in cultivation (for example, the Moore Foundation initiative http://www.

moore.org/marine-micro.aspx) because it provides the complete genome of the isolate and is a way to

test the hypotheses that emerge from metagenomic data, since procedures for reconstructing genomes

from whole-genome shotgun sequences are not yet very reliable.

Apart from differences from culture and culture independent techniques, differences in the

diversity detected between molecular procedures have been described (e.g. Alonso-Sáez et al. 2007).

In particular, this study showed that the dominant group present in Blanes Bay (SAR11) was not well

detected by the DGGE technique. Castle & Kirchman (2004) carried out a comparative study between

DGGE and FISH, and also showed that DGGE failed to detect the most abundant phylogenetic group

Fracti

on of

dive

rsity

(%)

0

20

40

60

80

100

1 2 3 4 5 6 7 8 9

OverlapClonesIsolates

Study

9 2

120

19

78 3519

16

25

5147

3515

26

44

8

8159

1917

028103020

Fig. 5.- Comparative analysis of diversity detected by culture-dependent and culture-independent approaches in the same habitat (taken from Donachie et al. 2007 with the addition of our data). Percentages of unique sequence groups reported in libraries done using molecular analyses (PCR-based 16S rRNA clones), in cultivation, or by both techniques (if a phylotype is detected by both methods it is listed separately: ´overlap´). The habitats shown are (1) Blanes Bay Microbial Observatory (Lekunberri et al. chapter 1); (3) endodontic pathogenic oral bio lms (Munson et al., 2002); (4) exodontic oral bio lms (Munson et al. 2004); (5) Hawaiian lakes (Donaiche et al. 2004); (6) rhizoplane (Kaiser et al. 2001); (7) bird feathers (Shawkey et al. 2005); (8) North Sea (Eilers et al., 2000); (9) hypersaline salterns (Maturrano et al. 2006).

182

Synthesis and dicussion

detected by FISH in some samples (Betaproteobacteria in their samples). Besides primer speci city,

these authors argued that high richness within groups could lead to an underestimation as compared

with FISH, because different sequences would appear as different faint bands, which could be dif cult

to excise from the gels for sequencing. Indeed, Sánchez et al. (submitted) have shown that the likely

reason for the lack of detection of SAR11 in DGGEs of Blanes Bay is the high evenness of multiple

microdiverse phyloypes, individually non-dominant in the area.

iv) DIFFERENT FERTILIZING EFFECTS OF THE DIFFERENT ADDITIONS

Mesocosms experiments have the dif culty of nding the right balance between realistic conditions

and controlled changes. The different studies presented in the thesis are in a scale from more control

to more realistic approaches (Fig. 1). The transplant experiment (Chapter 5) represent the most

realistic and less controlled approach, where the microbial community was transplanted to another

conditions that we did not modify. In a second step of complexity, we tried to simulate realistic events

occurring to natural communities. The real concentrations of PAHs are very dif cult to accurately

determine and in addressing the effects of an oil spill, in doing the experiments reported in Chapter

2, we concluded that we should better create a concentration-dependent experiment to determine the

threshold concentration that affected the communities. In the experiment mimicking a Saharan dust

deposition event, the problem was in the dif culty of determining the dust composition (Chapter

4). We decided to use two different concentrations and we observed very different whole microbial

metabolism effects with each concentration. The most controlled approach is the addition of simple

compounds, adding well known concentrations as we did in Chapter 3 where we tried to induce

different phytoplankton blooms.

In the experiments done in the Ría de Vigo, and using concentrations similar to those measured after

the Prestige accident (20-30 μg PAH L-1) we did not get an effect on bacteria. Not seeing a signi cant

response to the realistic addition of oil suggests that the natural communities are used to a certain

level of these substances in the water and maybe even the addition remains below the concentration of

available DOC. Our results contrasted to the measurements of bacterial production after the accident

in the A Coruña area that were higher as compared to previous years (Bode et al. 2006) indicating an

effect of oil. In contrast, in the concentration dependent experiment the concentration of 40 μg PAH

L-1 indicated a detectable effect and the 80 μg PAH L-1 a clear detrimental effect on phytoplankton,

and a large development of Gammaproteobacteria. These results coincided with Nayar et al. (2004)

that reported toxicity to the autotrophs and stimulatory effect to the bacteria with the addition of high

oil concentrations.

183

Synthesis and dicussion

The addition of Saharan dust only generated signi cant increases in bacterial abundance at the

highest concentration, suggesting a high contribution of the DOC content of the dust on the bacterial

community, and a limitation in phytoplankton growth by other nutrients such as silicate or nitrogen.

Previous studies have shown that addition of glucose alone as a C source does not necessarily stimulate

bacterial development (Pinhassi et al. 2006, Allers et al. 2007).

Finally, our results indicate a very strong effect of phosphorus on bacterial activity and abundance

that seems to supersede the effects of Carbon. Figures 3 and 4 indicate that all the treatments with

added P generated similar levels of bacterial biomass and production: treatments P and CP in Allers

et al. (2007), treatments P, Si, and U in Chapter 3, and treatments DH, DL and P in the Chapter 4

experiment.

v) FACTORS CONTROLLING BACTERIAL COMMUNITY STRUCTURE

In all ecosystems there is a ux of energy and material from autotrophic to heterotrophic organisms.

Phytoplankton organisms are the major producers of new production in the open ocean, while

bacteria utilise DOM as an energy source, and this is mainly of phytoplankton origin (5-50%, Nagata

2000). Bacteria also compete for nutrients with phytoplankton. At the same time the heterotrophic

nano agellates are effective bacterivores in the sea (e.g. Sorokin et al. 1979, Sierburth et al. 1982,

Unrein et al. 2007). Bacteria have been regarded as remineralizers, responsible for converting organic

matter to inorganic and recycling nutrients to primary produces and predation is known to provide a

feedback of some material ow to the food webs. Thus, the nutrient ow within the microbial loop is

tightly coupled. And the microbial loop concept originated from the idea that bacteria and agellates

recirculated the carbon that was previously though to be lost from the system. Grazers are important

in regulating bacterial abundance and production, and also seem to be a key factor controlling the

recycling of nutrients accumulated as bacteria (Caron & Goldman 1990).

It is also well known that nutrient additions cause changes on phytoplankton and these changes

are followed by shifts in bacterial community composition (Pinhassi et al. 2006, Allers et al. 2007),

in biomass and in bacterial function (production) at different stages of phytoplankton blooms (Smith

et al. 1995). Pernthaler et al. (1997) suggested that there was disagreement to whether the abundance

and productivity of bacteria are determined mainly by predators or nutrients. In a recent study Allers

184

Synthesis and dicussion

et al. (2007) considered a combination of both factors (bottom-up and top-down) controlling bacterial

succession and growth. This was in agreement with Lebaron et al. (1999) that suggested that nutrients

and grazing play a key role in shaping the morphologic, genotypic and phenotypic composition of

bacterial communities.

In our experiments we focused on the effect of the different additions on the bacterial community.

Besides we measure the HNF abundance in all the experiments except in the transplant experiment

(Chapter 5), where we controlled grazing pressure by ltration. From our results we can conclude the

importance of the predators on determining bacterial community structure. In Chapter 2, where we

correlated the changes in bacterial community structure with the changes in bacterial production (as no

change in structure is possible without production), we observed that this correlation was only signi cant

before the peak of HNF but not after that peak, suggesting a change in bacterial community structure,

maybe towards grazing-resistant forms (Pernthaler 2005). In the phytoplankton bloom stimulation

experiment (Chapter 3), the differences in bacterial activity and abundance between treatments were

found in the second bacterial peak, after the appearance of the HNF. Hence, there seems to be a shift

in bacterial community structure caused by the grazing pressure. In the experiments of transplant

of bacterial communities to different depths in the water column (Chapter 5) we observed that the

factors that determined a speci c community structure were mainly the environmental conditions,

although grazing pressure appeared to also play a role in determining bacterial abundance.

When analyzing the relative effects of nutrients or predators on community structure, the problem

of comparison between experiments arises. Did bacteria change more in the Sahara dust addition

experiment? or when we added nutrients? When we detect that community structure is different in the

ltered (no predators) than in the un ltered samples because the communities are not placed exactly

in the same position in the NMDS plot (Fig 6 in Chapter 5), how can we evaluate the signi cance of

this difference? To approach answering these questions, we decide to evaluate each of the experiments

presented in this thesis by averaging taking each DGGE analysis, and computing the average similitude

in each experiment. If, in an experiment, all treatments were to evolve equally, the average Bray-

Curtis similarity index would be 100. If the samples were to immediately change and no single band

would be repeated from treatment to treatment, then average similarity would be 0. Figure 6 compares

the values that we have measured during this thesis, with the typical seasonal evolution recorded in

Blanes Bay.

In the experiments where we detected little change, such as those in Ría de Vigo in spring, fall

and winter, the value of the average Bray-Curtis was about 65. In the summer experiment when we

did see differences, the value was 45. The seasonal changes in Blanes Bay provided values of 35-

45. And the experiments reported values ranging from 35 (transplant experiments, chapter 5) to 65

185

Synthesis and dicussion

(Saharan dust addition, Chapter 4), with intermediate values in the nutrient addition experiment. This

comparison allows us to quantitatively say that the changes produced by the transplant were of the

same magnitude than those occurring on average along the seasonal cycle in the NW Mediterranean,

but that the changes produced by Saharan dust were as modest as those produced by oil additions in

Ría de Vigo communities.

��

��

��

��

��

��

���

Fig. 6.- Comparison of the average (± SE) similarity indices (Bray-Curtis) in each of the studies for which we had analyzed the variability of bacterial community structure using the DGGE technique. We included the seasonal variability at the Blanes Bay Microbial Observatory during years 1997-1998 (in Schauer et al. 2003), and years 2003-2004 (Alonso-Sáez et al. 2007).

186

Synthesis and dicussion

vi) WHICH BACTERIA ARE SELECTED IN THESE TYPE OF ADDITION EXPERIMENTS?

Almost every ecophysiological parameter in the oceans is thought to have an impact on the diversity

of microbial communities. Recent studies show that some bacterial taxa exhibit biogeographical

patterns, which are traced to the changes in the environmental characteristics (Pace 2003). Previous

studies have shown that the different conditions of the different times of the year affect the bacterial

assemblage in our study site (Schauer et al. 2003, Alonso-Sáez et al. 2007), i.e. there is seasonality. The

bacterial community structure of our study area was also investigated previously by two mesocoms

approaches. In the rst one, Pinhassi et al. (2004) found the presence of Roseobacter genera in all

treatments and a remarkable predominance of Bacteroidetes (in particular Flavobacteria) following

the phytoplankton blooms. Later on Allers et al. (2007) showed a clear contribution of the Roseobacter

genera (Alfaproteobacteria subclass) dependent on the Chla concentrations and an until increase in

Alteromodaceae (Gammaproteobacteria) at the start of the blooms. Other studies carried out in the

NW Mediterranean also found similar patterns, Schäfer et al. (2001), e.g. in an enrichment experiment

found particular dominance of genera Ruegeria belonging to the Alphaproteobacteria subclass, and

members of the Bacteroidetes. Lebaron et al. (1999) also found that phylotype Alteromonas macleodii

was dominant at the end of the incubation.

Langenhender (2005) suggests that bacterial community composition and bacterial functioning do

not need to be very tightly linked being however both mostly in uenced by the environment. There

are some bacteria that have speci c functionalities and only become dominant in speci c situations.

For example, the Gammaproteobacteria are generally the dominant bacterial group after oil additions

(Grossman et al. 1999, Kasai et al. 2001). There are two key organisms identi ed with the major roles in

the degradation in the petroleum hydrocarbons. Alcanivorax is responsible for alkane biodegradation,

whereas Cycloclasticus degrades various aromatic hydrocarbons (Harayama et al. 2004). In our oil

spill stimulation experiment we found a clear shift towards Gammaproteobacteria in the 80 μg L-1

treatment, where Cycloclasticus ended up being a large faction of the total Gammaproteobacteria.

Not in all experiments we decided to explore phylogenetically the bacteria generating relevant

bands in the DGGE. We did so only in the Saharan Dust experiment (Chapter 4), where we obtained

bands belonging to Bacteroidetes (known to have a life-style of attachment to particles), Within

the Alphaproteobacteria identi ed, bacteria of the Ruegeria/Silicibacter cluster appeared in several

treatments, as well as several Rhodobacteraceae. These are all common organisms in Blanes Bay,

easy to isolate in plates, and common in mesocosms experiments (Chapter 1). In fact, addition of

sequences retrieved (in the past) from enrichment experiments allowed to nd quite some phylotypes

appearing both in culture, and in the natural environment, but when the natural environment had been

modi ed by the different allochthonous enrichments studied here.

187

Conclusions

1.- Combination of culturing and culture independent approaches has increased substantially the diversity detected in Blanes Bay. We analyzed 136 isolates which belonged mainly to the Gammaproteobacteria, Alphaproteobacteria, or the Bacteroidetes. Addition of these organisms increased the estimates of richness in ca. two times the OTUs richness previously described with molecular tools alone.

2.- Nine cases of complete or nearly complete identity between isolates and in situ sequences were found: ve in the Gammaproteobacteria, three in the Alphaproteobacteria and one in the Bacteroidetes. This suggests that culturing to some extent provides representative organisms for physiological response experiments for various taxa detected by molecular methods.

3.- Around 30% of the environmental clones and isolates formed microdiversity clusters constrained at 99% 16S rRNA gene sequence similarity. The nding of microdiversity clusters among isolate and clone sequences alike, suggests that this type of microdiversity is not an artifact and that similar evolutionary processes act on cultured and uncultured bacteria.

4.- The bacterioplankton communities of coastal areas where there is an intense impact of human activities, such as Ría de Vigo, are used to certain concentrations of PAHs dissolved in the water. However, during the summer when in that area there is an entrance of oceanic water oil additions affected much more the function and community structure of the resident microbial communities.

5.- Development of agellate populations in the experiments performed in diverse marine areas was seen to explain some of the patterns in bacterial function and community structure seen. In the mesocosm experiments performed in Ría de Vigo, for example, we detected a signi cant effect of

agellates on bacterial community structure.

6.- Addition of different types of inorganic nutrient to NW Mediterranean waters generated responses in bacterial function and community structure mainly because of the P-limited nature of bacterial growth in the area. An initial bacterial peak appeared in all samples to which P was added, independently of the phytoplankton species selected by the addition. A second bacterial peak later on during the experiment was more diverse, showing effects of the different phytoplankton that developed.

7.- Besides the fertilization produced by the phosphorus content of Saharan dust we found that an addition of dust also stimulated directly bacterial activity and community respiration.Dust addition increased bacterial abundance by ca. 2-fold and bacterial production by 5-fold. Primary production and community respiration were stimulated by dust and by P, but the net result of the addition of low amounts of dust was a heterotrophic system, while the net result of the high dust and P additions were net autotrophic communities.

CONCLUSIONS

188

Conclusions

8.- Bacterial community structure changed little between the P-addition and a low addition of dust (0.05 g L-1), but the resulting communities were very different from the control with no additions, and from the communities that developed following a large dust addition (0.5 g L-1),

9.- The incubation of bacterial communities in different environmental conditions (water from surface transplanted to the mesopelagic or the DCM, and viceversa) led to bacterial function and growth to mainly follow the conditions of the incubation tank, with little effects of the initial community and of the presence or absence of predators. Predator presence had a small effect on bacterial production but not on nal bacterial yield.

10.- The incubation of bacterial communities in different environmental conditions (water from surface transplanted to the mesopelagic and viceversa) lead to changes in bacterial community structure, indicating bottom-up forces to be much more important than top-down regulation.

11.- Comparison of different experimental approaches ranging from extremely modi ed nutrient elds (in Chapter 1) and little touched environmental transplants (chapter 5) show how each approach illustrates different aspects of the dynamic interplay between bacterial community structure, diversity and function in the ocean. The comparison also allows detection of general patterns in bacterial development in experimental mesocosms, which vary depending on the initial conditions of the samples.

12.- Comparison of the changes in bacterial community structure induced by the different experimental treatments presented in this work allows the discrimination of large induced changes (e.g. the transplant experiments of Chapter 5) from low induced changes (Saharan dust addition). These can, in turn, be compared to the seasonal variability do discriminate “large” from “small” effects on bacterial community struture.

189

Future perspectives

FUTURE PERSPECTIVES

In the different enrichment experiments presented, we used the ngerprinting technique DGGE that permits the comparison between samples but, however, has a limitation and seldom ever presents more than 20 bands. We suggest that should promote a technique such as ARISA (Ampli ed Ribosomal Intergenic Spacer Anlysis, Brown et al. 2005), which is a technique from which a number of around 100 sequences can be retrieved. The RNA section utilized by this technique is from 16S to 23S, including the ITS fragment that has large variability in length and sequence. These products are run on automated sequencer generating an electropherogram where the sequences are differentiated. Standards (e.g. from clones) can be run and used to identify the peaks (Brown et al. 2005, Hewson & Fuhrman 2006).

Besides the culture and molecular techniques used in this thesis to study the diversity of a speci c

environment, we suggest the use of PCR-independent approaches such as uorescence in situ

hybridization (FISH) to complement the analyses of biodiversity of the ecosystem and to follow the

temporal dynamics of bacterial populations in enrichment experiments. We used this approach in

Teira et al. (2007) to detect the dynamics of small contributors to community structure (such as the

Gammaproteobacteria Cycloclasticus) in the oil-spill simulation experiment. In Chapter 2 we show

how, this technique complements well what we can see in DGGE analyses.

There is an increasing interest in relating the ecological function of microbes and diversity. We, thus,

suggest that more effort in this issue would provide a better understanding of ecosystem functioning.

Different approaches could be used to address this question:

i) One approach would be the isolation of bacteria that allow the sequencing of the complete

genome and the subsequent physiological analysis of the capacity of the isolates (e.g. González et al.

2008) specially during enrichment experiment where the concentration of the addition is more or less

controlled.

ii) The combination of FISH with microautoradiography provides the information of the single-

cell activity of speci c bacterial groups on carbon processing. This technique utilizes radio labelled

compounds, and thus labelling of the allochthonous compounds to determine the bacteria taking up

these substrates isn’t an easy task. We tried hard with 14C-labelled hydrocarbons, without success.

190

Future perspectives

iii) Another approach to determine which are the organisms growing because of a speci c substrate

addition, is a recently developed technique that consists in the incubation with compounds labelled

with Bromodioxidouridine (BrdU), that is an halogenated nucleoside that works as an analoge to

thymidine and as an alternative to [H3] TdR. It is an alternative to radioactive substrates and can be

applied to determine bacterial production if one follows a subsequent molecular analysis. It is used

as a speci c ngerprinting technique (BUMP-DGGE) that reveals the phylogenetic af liations of

the presumably actively growing bacteria that have incorporated this compound (Urbach et al. 1999;

Hamasaki et al. 2007). The method has been used simultaneously to FISH (Pernthaler et al. 2002).

ANNEXES

Non gogoa, han zangoa

Proverbio vasco

193

Annexes

ANNEX I:

Traditional methods for bacterial isolation and cultivation

Bacteria play signi cant roles in marine food webs and in most marine biogeochemical cycles

(Fenchel 1988). The large biomass of bacteria is believed to be a repository of relatively unknown

diversity. Despite the overwhelming bacterial diversity present in the world’s oceans, the majority

of marine bacteria can not be easily cultivated in the laboratory (Giovannoni & Rappé 2000).

The discrepancy between direct microscopic enumeration and plate counts of bacteria was rst

pointed by Jannasch & Jones (1959). They attributed it to the presence of bacteria in aggregates, to

selective effects of the used media, and to the presence of inactive cells. Since values estimated by

epi uorescence direct counts are orders of magnitude higher than those estimated by plate counts (the

so-called CFU or colony forming units), this has been referred to as “the great plate count anomaly”

(Staley & Konopka 1985). Comparison of direct cell counts and classic cultivation approaches, have

shown that only a minor fraction, <1% of total bacterial numbers, can be cultivated in common media

(Staley & Konopka 1985)

In contrast, based on DNA-DNA hybridation of the genomic DNAs of isolates obtained with the

traditional ZoBell medium run against community DNA, it has been suggested that readily culturable

bacteria can be abundant in the marine water column (Rehnstam et al. 1993, González et al. 1996,

González & Moran 1997, Pinhassi et al. 1997).

The great challenge in microbial ecology is to relate diversity to the ecological function and

biogeochemical activities of speci c bacteria (Hamasaki et al. 2007). To correlate the activity of

speci c microorganisms with de ned environmental or physiological parameters

Among the different ways for associating diversity and functioning, one possibility is the isolation

of relevant members of the assemblage and their study in the laboratory. If we know that a given

organism does a given function, we might be able to search for its presence to better know about

that linkage. But currently, the general opinion is that the isolates tend not to be representatives of

the indigenous bacteria. Lack of knowledge about the dominant uncultured microorganisms can be

attributed directly to their low cultivability by standard microbiological techniques.

194

Annexes

Progress has been made in recent years at cultivating members of the numerically abundant

clades; for example, the SAR11 clade using sterile marine water supplemented with either phosphate,

ammonium and de ned mixture of organic carbon compounds (Rappé et al. 2002), the marine group

I Crenarchaeota was grown aerobically at 28ºC in Synthetic Crenarchaeota Media amended with

autoclaved vitamin solution, KH2PO

4 solution, selenite-tungstate solution, bicarbonate solution

and ammonium chloride per litre of media (Konneke et al. 2005) and the OM43 clade (Connon &

Giovannoni, 2002) were obtained in pure cultures. Cultivation allows a comprehensive physiological

characterization of relevant members of bacterial communities. Therefore, is important to complement

the culture-independent ndings of the microbial communities in aquatic and other environments

(Stevens et al. 2005).

The interest on the characterization of new species is increasing, particularly since it is now possible

to know their genomes through initiatives such as the Gordon & Betty Moore Foundation Marine

Microbial Initiative. Knowing genomic composition allows us to determine the function of microbial

organisms in speci c environments. Rush et al. (2007) published part of the Sorcereer II Global

Ocean Sampling Expedition and suggested the utility of having isolates from a well characterized

biogeochemical site to be used as scafolds for reconstruction of the genomes.

Using traditional cultivation techniques we obtained isolates collected offshore at the Blanes Bay

Microbial Observatory (32º 80’N 34º 93’E) Some of these isolates were described as novel. Of some,

whole-genome sequencing was carried out by the J.Craig Venter Institute (USA) through the Gordon

and Betty Moore Foundation initiative in Marine Microbiology (https://www.moore.org/marine-

micro.aspx).

195

Annexes

A.- PROTOCOL OF ISOLATION

1.- Petri dish preparation

- 4 g of Peptone (Bacto TM Peptone DIFCO 211677)

- 0.8 g of yeast extract (Bacto Yeast Extract DIFCO 212750)

- 12.5 g of Agar (Bacto TM Agar DIFCO 214010)

- Add 600 ml seawater ltrated by 0.2 μm (sterivex Millipore SVBV010RS

or Membrane lters GVWP02500 or GVWP04700)

- Add 200 ml of sterile Milli Q water

- Mix until is dissolved (with a magnetic stirrer)

- Autoclave 20 minutes at 121ºC (1 kg/cm2)

- Mix the bottle and autoclave again 20 minutes

- Let it stand until is warm and can be touched

1.1-Prepare Petri dishes

- Leave the Petri dishes in the laminar ux chamber for 30 minutes with ultraviolet light on

- Dispense the liquid into the Petri dishes (40 ml)

- Close the plates

- Leave for 24 h to solidify

- Turn upside-down the Petri dishes (To prevent agar from drying too much)

- Leave standing for 2 days (To completely dry out)

- Close each Petri dish with para lm. (To avoid any contamination)

- Label them with the preparation date.

196

Annexes

1.2- Spread out sample onto the Petri dishes

- Prepare different dilutions of the sample (with sterile seawater): 1x, 10x and 100x

- Add 100 μl with a pipette and spread with a digralsky spreader

- Leave the Petri dishes growing for one week at in situ temperature.

1.3- Isolation of the colonies

- Pick the colonies (with an inoculating loop) and spread them in a new Petri dish.

- Divide the Petri dish in four slices

- Spread each colony in one square of the Petri dish.

- Label each square with the number or name of colony and date of the inoculation.

Fig.1.- Colony forming bacteria from Blanes Bay growing on agar plates.

197

Annexes

1.4- Growth in liquid media

Preparation of 1 L of the Zobell liquid media:

- 4 g of Peptone

- 0.8 g of yeast extract

- Add 600 ml seawater ltrated by 0.2 μm

- Add 200 ml of sterile MilliQ water

- Mix until everything is dissolved

- Dispense 20 ml in 100 ml sterile polycarbonate Nalgene bottles.

- Autoclave for 20 minutes at 121ºC

- Pick the chosen colony from the Petri dish

- Place it in the liquid media

- Leave one or two days until the liquid is opaque.

1.5- DNA collection

- Place all the liquid media in a 15 ml centrifuge tube (DELTALAB) and centrifugate for 10

minutes at 4500 rpm

- Remove the supernatant with a pipette.

- Maintain the pellet at -80ºC

1.6- Check for contamination

- Take 15 μl of the liquid media and spread them on a Petri dish.

- When the drop is dry, spread in the Petri dish with an inoculation loop.

- If the growing colonies are only of one type we can presume that there is no contamination.

198

Annexes

1.7- Glicerination

- Mix 900 μl of sterile glycerol (SIGMA G5516) at 50% with 900 μl of the liquid media

- Keep at –80ºC.

Fig.2.- Two different strains growing on agar plates.

199

Annexes

ANNEX II:

Scanning Electronic Microscopy (SEM) images of bacterial

cultures

An electron microscope is a type of microscope that uses an electron beam to produce and enlarged

an image. Electron microscopes have much greater resolving power and can obtain much higher

magni cations than light microscopes. The two important contributions to electronic microscopy

were the Broglie (1924) theory about the wave proprieties of electron movement and the Bush (1926)

discovery of the magnetic or electrostatic elds.

There are two kinds of electron microscopes: the transmission (TEM) and scanning (SEM), or the

combination of both (METB). TEM has to be used in ultra-thin cuts (from 10 to 200 nm) depending

on the sample. The electrons pass through the sample to form the image. The electron beam scan

over the surface, and the signals produced include secondary electrons, backscattered electrons,

characteristic X rays, Auger electrons, and photons of various energies. The interactions that the

beam electrons undergo in a specimen reveal information on the specimen’s composition, topography,

crystallography and other properties.

Samples preparation is a critical process. The main steps are, the selection of the sample, cleaning,

xation, deshydratation, placement on the supports and coverage with a conductive product or a

metal. All these have to be gently done to avoid the cells to shrink or to break in their surface.

200

Annexes

2- PROTOCOL SEM

2.1- Sample preparation

- Counts of bacteria should not exceed 10 6 cells ml-1

- Prepare different dilutions of the liquid media: 1x, 10x and 100x in 10 ml

2.2- Sample xation

- Add Glutaraldehyde at 2-3% nal concentration to 10 ml of the different dilutions.

- Leave xing during 2-4 h.

2.3- Poli-L-Lysine treatment: to adhere the sample to the lter

- Cover the lter with a drop of Poli-L-Lysine 0.1% w/v solution (SIGMA)

- Wait 1 minute and submerge the lters in Milli Q water for 30 seconds.

2.4- Filtration

Filtrate the samples with 0.2 μm polycarbonate lter previously prepared with Poly-L-Lysine

(Millipore).

After the ltration, place the lters in a metal net capsule to avoid coiling of the lter.

2.5- Dehydratation

Deshydratate in an alcoholic series with Ethanol and esterile MilliQ water:

Leave for 15 minutes at 30%, 15 min. at 50 %, 15 min at 70 %, 15 min at 80 %, 15 minutes at

90 % (two times) and 15 minutes at 100 % (two times)

2.6- Dry by critical point

Critical point dryer (BAL-TEC, CPD 0309). is used to end the dry process without any surface tension on the samples.

1º- Place the sample in a cage of the dryer.2º- Add ethanol just to cover the sample 3º- Lower temperature until 4-7ºC4º- Fill the rest of the cage with liquid CO

2

5º-Remove half of the total liquid every 15 minutes around 7-8 times until is all CO2

6º- Increase temperature until 37ºC to avoid liquifying of the CO2.( The critical point of CO

2 is

reached at a temperature of 31ºC and a pressure of 73,8 bar)7º- Decrease pressure 6 bar per minute to avoid liquifying of the CO

2

201

Annexes

Images of Blanes Bay Isolates: (check Chapter 1 - Table 4 for details of the strains).

Fig.1 M134 : Dokdonia donghensis

Fig.2 M193: Phaeobacter sp.

Fig3 M222: Vibrio sp.

202

Annexes

Fig.4 M121: Marinomonas blandensis

Fig.5 M217: Leeuwenhoekiella blandensis

Pinhassi P, Bowman JP, Nedashkovskaya OI, Lekunberri I, Gomez-Consarnau L & Pedrós-Alió C (2006). Leeuwenhoekiella blandensis, sp.nov., a novel marine member of the family Flavobacteriaceae. Inter J Syst Evol Microb 56: 1489-1493.

Pedrós-Alió, C. 2007. Dipping into the rare biosphere. Science 315:192-193

203

Annexes

Fig.6 M297: Reinekea blandensis

Pinhassi J, Pujalte MJ, Macián MC, Lekunberri I, González JM, Pedros-Alió C & Arahal DR (2007). Reinekea blandensis sp. nov., a marine genome-sequenced gammaproteobacterium. Inter J Syst and Evol Microb 57: 2370-2375

Fig.7 M152: Polaribacter dokdonensis

González JM, Fernández-Gómez B, Fernàndez-Guerra A, Sánchez O, Gómez-Consarnau L, Coll-Lladó M, del Campo J, Escudero L, Rodríguez-Martínez R, Alonso-Sáez L, Latasa M, Paulsen I, Nedashkovskaya O, Lekunberri I, Pinhassi J, & Pedrós-Alió C (2008) Genome analysis of the proteorhodopsin-containing marine bacterium Polaribacter sp. MED152 (Flavobacteria): a tale of two environments. PNAS doi: 10.1073/pnas.0712027105

204

Annexes

Fig.8 RED 65:Bermanella marisrubri

Pinhassi J, Pujalte MJ, Pascual J, González JM, Lekunberri I, Pedrós-Alió C & Arahal DR (2008) Bermanella marisrubri gen. nov., sp. nov., a novel genome-sequenced Gammaproteobacterium from the Red Sea. Inter J Syst Evol Microb (in press)

Fig.9 SCB36 Winogradskyella sp.

205

Annexes

Fig.10 MED92: Neuptuniibacter caesariensi

Arahal DA, Lekunberri I, Gonzalez JM, Pascual J, Pujalte MJ, Pedrós-Alió C, & Pinhassi J (2007) Neuptuniibacter caesariensis gen. nov., sp. nov., a novel marine genome-sequenced gammaproteobacterium.Inter J Syst Evol Microb 57: 1000-1006

REFERENCES(introduction discussion and annexes)

Imagination is more important than knowledge

Albert Einstein

209

References

REFERENCES

Acinas SG, Klepac-Ceraj V, Hunt DE, Pharino Ch, Ceraj I, Distel DL & Polz M (2004) Fine-scale phylogenetic architecture of a complex bacterial community. Nature 430: 551-554.

Allers E, Gómez-Consarnau L, Pinhassi J, Gasol JM, Simek K & Pernthaler J (2007) Response of Alteromonadaceae and Rodobacteriaceae to glucose and phosphorus manipulation in marine mesocosms. Environ Microb 9: 2417-2429.

Alonso-Sáez L, Balague V, Sá EL, Sánchez O, Gonzalez JM, Pinhassi J, Massana R, Pernthaler A, Pedrós-Alió C & Gasol JM (2007) Seasonality in bacterial diversity in north-west Mediterranean coastal waters: assessment through clone libraries, ngerprinting and FISH. FEMS Microbiology Ecology.10:2444-2454

Baas-Becking LGM (1934) Geologie of INleiding Tot de Milieukunde. The Hague, The Netherlands: W.P. Van Stokum.

Bratbak G, Hendal M, Norland S, Thingstad TF (1990) Viruses as partners in spring bloom microbial trophodynamis. Appl Envir Microb 56: 1400-1405

Broglie L (1926)Ondes et mouvements (Waves and Motions). Paris: Gauthier-Villars,

Béjà O, et al. (2000) Bacterial Rhodopsin: Evidence for a nex type of phototrophy in the sea. Science 289: 1902-1905.

Bode A, González N, Lorenzo J, Valencia J, Varela M, & Varela M. (2006) Enhanced bacterioplankton activity after the ´Prestige´oil spill off Galicia, NW Spain. Aquat Microb Ecol 43: 33-41.

Brown, M.V., M.S. Schwalbach, I. Hewson and J.A. Fuhrman. 2005. Coupling 16S-ITS rDNA clone libraries and ARISA to show marine microbial diversity; development and application to a time series. Envir Microb 7:1466-1479

Button DK, Schut F, Quang P, Martín R, Robertshon BR (1993) Viability and isolation of marine bacteria in dilution culture: theory, procedures and initial results. Appl Env Microbiol 59: 881-891

Cappello S, Denaro R, Genovese M, Giuliano L & Yakimov M (2007) Predominant growth of Alcanivorax during expereiments on “oil spill bioremediation “ in mesocosms. Microb Res 162: 185-190.

Cappello S, Denaro R, Genovese M, Giuliano L, & Yakimov MM (2006) Predominant growth of Alcanivoraxduring experiments on “oil spill bioremediation” in mesocosms. Microb Res 162: 185-190

Caron & Goldman (1990) Protozoan nutrient regeneration. Ecology of marine protozoa. Oxford

Castle D, Kirchman DL (2004) Composition of estuarine bacterial communities assessed by denaturing gradien gel electrophoresis and uorescence in situ hybridization. Limn aOceanogr Meth 2:303-314

Castle DM, Montgomery MT & Kirchman DL (2006) Effects of napthalene on microbial community composition in the Delaware estuary. FEMS Microbiology Ecology 56: 55-63

Collins A (2001) Ocean circulation. Ocean Circulation (ed. Heineman, B.) Boston, MA 208-209.

Connon SA & Giovannoni SJ (2002) High-throughput methods for culturing microorganisms in very-low-nutrient media yield diverse new marine isolates. Appl Environ Microb 68: 3878-3885.

Cottrell MT & Kirchman DL (2000) Natural assemblages of marine proteobacteria and members of the Cytophaga-Flavobacter cluster consuming Low- and High-molecular weight dissolved organic matter. Appl Environ Microb 66: 1692-1697.

210

References

Crump BC, Kling GW, Bahr M & Hobbie JE (2003) Bacterioplankton community shifts in an artic lake correlate with seasonal Changes in organic matter source. Appl Environ Microb 69:2253-2268

del Giorgio PA, Gasol JM, Vaqué D, Mura P, Agusti S & Duarte CM (1996) Bacterioplankton community structure: protists control net produ tion and the proportion of active bactteria in a coastal marine community. Limn Oceanogr 41:1169-1179.

DeLong, E.F., Franks, D.G. and Alldredge, A.L. (1993) Phylogenetic diversity of aggregate-attached vs. free-living marine bacterial assemblages. Limn Oceanogr 38, 924-934.Donachie, SP, Foster JS & Brown MV (2007). Culture clash: Challenging the dogma of microbial diversity. ISME J. 1: 97-99.

Duarte CM, Agustí S, Gasol JM, Vaqué D, Vazquez-Dominguez E (2000) Effect ofnutrient suplí on the biomass structure of planktonic ocmmunities: an expereimental test on a Mediterranean coastal community. Mar Ecol Prog Ser 206: 87-95

Eilers H, Pernthaler J, Glockner FO & Amann R (2000) Culturability and in situ abundance of pelagic bacteria from the North Sea. Appl Environ Microb 66: 3044-3051.

Eilers H, Pernthaler J, Peplies J, Glockner FO, Gerdts G & Amann R (2001) Isolation of novel pelagic bacteria from the German bight and their seasonal contributions to surface picoplankton. Appl Environ Microb 67: 5134-5142.

Fenchel T (1982) Ecology of heterotrophic micro agellates. Quantitative occurrence and importance as consumers of bacteria. Mar Ecol Progr Ser 9: 35-42.

Fenchel T & Finlay BJ (2004) The ubiquity of small species: pattersn of local and global diversity. Bioscience 54: 777-784.

Ferguson RL, Buckley EN & Palumbo AV (1984) Response of marine bacterioplankton to differential ltration and con nement. Appl Environ Microb 47: 49-55.

Finlay BJ (2002) Global Dispersal of free-living microbial eukaryote species. Science 296: 1061-1063.

Gasol JM, del Giorgio PA & Duarte CM (1997) Biomass distributions in marine planktonic communities. Limn Oceanogr 42: 1353-1363.

Gasol, J.M., Duarte, C.M. (2000) Comparative analyses in aquatic microbial ecology: How far do they go ? FEMS Microbiology Ecology 31: 99-106.

Giovannoni SJ & Rappe M (2000) Evolution, diversity, and molecular ecology of marine prokaryotes. Microbial Ecology of the Oceans, Edited by Dvid I., Kirchman.

Giovannoni SJ, Rappe M, Vergin KL & Adair N (1996) 16S rRNA genes reveal strati ed open ocean bacterioplankton populations related to the Green Non-Sulfur bacteria. Proceedings of the National Academy of Sciences of the United States of America 93: 7979-7984.

Gómez-Consarnau L, González JM, Coll-Lladó M, Gourdon P, Pascher T, Neutze R, Pedrós-Alió C & Pinhassi J (2007) Light stimulates growth of proteorhodopsin-containing marine Flavobacteria. Nature445: 210-213.

González JM, Whitman WB, Hodson RE, Moran MA. Identifying numerically abundant culturable bacteria from complex communities: an example from a lignin enrichment culture. Appl Environ Microbiol. 1996 Dec;62(12):4433–4440.

González J M& Moran MA (1997) Numerical dominance of a group of marine bacteria in the alpha-subclass of the class proteobacteria in coastal seawater. Appl Environ Microbiol 63: 4237-4242.

211

References

González JM, Fernández-Gómez B, Fernàndez-Guerra A, Gómez-Consarnau L, Sánchez O, Coll-Lladó M, del Campo J, Escudero L, Rodríguez-Martínez R, Alonso-Sáez L, Latasa M, Paulsen I, Nedashkovskaya O, Lekunberri I, Pinhassi J & Pedrós-Alió C (2008) Genome analysis of the proteorhodopsin-containing marine bacterium Polaribacter sp. MED152 (Flavobacteria). Proc Nat Acad Sci USA 105: 8724-8729.

Grossman M, Prince R, Garrett R, Garrett K, Bare R, Lee K, Sergy G, Owens E & Guénette C (1999) Microbial diversity in oiled and un-oiled shoreline sediments in the Norwegian Arctic. Proc. 8th Int. Symp. Microb. Ecol., Halifax, Canada

Harayama S, Kishira H, Kasai Y & Shutsubo K. (1999) Petroleum biodegradation in marine environments. J Mol Microbiol Biotechnol 1: 63-70

Hamasaki K, Taniguchi A, Tada Y, Long RA & Azam F (2007) Actively growing bacteria in the inland sea of Japan, identi ed by combined bromodeoxyuridine immunocapture and denatruing gradient gel electrophoresis. Appl Environ Microb 73: 2787-2798.

Herndl GJ, Kaltenböck E & Müller-Niklas G (1993) Dialysis bag incubation as a nonradiolabeling technique to estimate bacteriplankton production in situ. In. P.F. Kemp, B.F. Sherr, E.B. Sherr, and J.J. Cole (eds), Handbook of methods in aquatic microbial ecology. Lewis 553-556.

Hewson I & Fuhrman JA (2006) Spatial and vertical biogeography of coral ref. sediment bacterial and diazotroph communities. Mar Ecol Prog Ser 306: 79-86

Horner-Devine MC, Carney KM & Bohannan BJM (2003) An ecological perspective on bacterial biodiversity. The Royal Society 271: 113-122.

Hunter-Cervera J, Karl D & Buckley M (2005) Marine microbial diversity: The key to Earth’s habitability. American Academy of Microbiology, Washington.

Jannash HW & Jones GE (1959) Bacterial populations in sea water as determined by different methods of enumeration. Limn Oceanogr 4: 521-527.

Jürgens K & Güde H (1994) The potential importance of grazing-resistant bacteria in planktonic systems. Mar Ecol Progr Ser 112: 169-188.

Kasai Y, Kishira H, Syutsubo K & Harayama S (2001) Molecular detection of marine bacterial populations on beaches contaminated by the Nakhodka tanker oil-spill accident. Environ Microb 3: 246-255.

Kasai Y, Kishira H & Harayama S (2002) Bacteria belonging to the genus Cycloclasticus play a primary role in the degradation of aromatic hydrocarbons released in a marine environment. Appl Environ Microb 68: 5625-5633.

Konneke M, Bernard AE, de la Torre JR, Walker CB & Waterbury JB (2005) Isolation of an autotrophic ammoni-oxidizing marine arhceon. Nature 437: 542-546.

Langenheder S, Lindstrom ES & Tranvik LJ (2005) Weak coupling between community composition and funtioning of auqatic bacteria. Limn Oceanogr 50: 957-967.

Labbé D, Margesin R, Schinner F, Whyte LG & Greer CW (2007) Comparative phylogenetic analysis of microbial communities in pristine and hydrocarbon-contaminated Alpine soils. FEMS Microbiology Ecology 59: 466-475.

Lebaron P, Servais P, Troussellier M, Courties C, Vives-Rego J & Muyzer G (1999) Changes in bacterial community structure in seawater mesocosms differing in their nutrient satatus. Aquat Microb Ecol 19: 255-267.

212

References

Malmstrom RR, Kiene RP, Cottrell MT, Kirchman DL (2004) Contribution of SAR11 bacteria to disolved dimethylsulfoniopropionate and amino acid uptake in the North Atlantic Ocean. Appl Envión Microbl 70: 4129-4135

Macnaughton SJ, Stephen JR, Venosa AD, Davis GA, Chang YJ & White DC (1999) Microbial population changes during bioremediation of an experimental oil spill. Appl Envión Microb 65: 3566-3574.

Margalef, R. (1994) Diversity and biodiveristy – their posible meaning in relation with the wish for sustainable develpment. An Acad. Bras. Ciênc. 66: 3-14

Marshall, K. C., R. Stout, and R. Mitchell. 1971. Mechanisms of the initial events in the sorption of marine bacteria to solid surfaces. J Gen Microbiol 68: 337-348.

Maruyama A, Ishiwata H, Kitamura K, Sunamura M, Fujita T, Matsuo M & Higashihara T (2003) Dynamics of microbial populations and strong selection for Cycloclasticus pugetii following the Nakhodka oil spill. Microb Ecol 46: 442-453.

Morris RM, Rappe M, Connon SA, Vergin KL, Slebold WA, Carlson CA & Giovannoni SJ (2002) SAR11 clade dominates ocean surface bacterioplankton communities. Nature 420: 806-810.

Morris RM, Rappe M, Urbach E, Connon SA & Giovannoni SJ (2004) Prevalence of the Chloro exi-Related SAR202 Bacterioplankton Cluster thoughout the Mesopelagic Zone and Deep Ocean. Appl Envión Microb70: 2836-2842.

Morris RM, Vergin KL, Cho JC, Rappe M, Carlson CA & Giovannoni SJ (2005) Temporal and spatial response of bacterioplankton lineages to annual convective overturn at the Bermuda Atlantic Time-series study site. Limn Oceanogr 50: 1687-1696.

Muyzer AE, Brinkhoff T, Nübel U, Santegoeds C, Schäfer H & Wawer C (1997) Denaturing gradient gel electrophoresis (DGGE) in microbial ecology. In: Molecular Microbial Ecology Manual (ADDERMANDS ADL, van Elsas JD & Brujin FJ Eds.) Vol. 3.4.4, pp. 1-27

Nayar S, GohB PL & Chou LM. (2005) Environmental impacts of diessel fuel on bacteria and phytoplankton in a tropical estuary assessed using in situ mesocsms. Ecotoxicology 14: 397-412

Odum, EP. 1984. The mesocosm. BioScience 34: 558-562

Ogino A, Koshikawa H, Nakahara T & Uchiyama H (2001) Succession of microbial communities during a biostimulation process as evaluated by DGGE and clone libray analyses. J.Appl Microb 91: 625-635.

Ohwada, K., Nishimura, M., Wada, M., Nomura, H., Shibata, A., Okamoto, K., Toyoda, K., Yoshida, A., Takada, H., and Yamada, M. (2003) Study of the effect of water-soluble fractions of heavy-oil on coastal marine organism using enclosed ecosystems, mesocosms. Mar Pollut Bull 47: 78-84

Olsen Y, et al. (2006) A comparative study of responses in planktonic food web structure and funstion in contrasting European coastal waters exposed to expereimental nutrient addition. Limn Oceanogr 51: 488-503.

Pace MF & Cole JJ (1994) Comparative and experimental approaches to top-down an dbottom-up regulation of bacteria. Microb Ecol 28: 181-193.

Pace ML (1988) Bacterial mortality and the fate of bacterial produc tion. Hydrobiologia 9: 41-49.

Pace A, Vidiri A, Galie E, Carosi M, Telera S, Cianciulli AM, Canalini P, Giannarelli D, Jandolo B, & Carapella CM (2003) Temozolomide chemotherapy for progressive low-grade glioma: Clinical bene ts and radiological response. Ann. Oncol. 14: 1722-1726

213

References

Pedrós-Alió C (2006) Marine microbial diversity: can it be determined? Trends in Microbiology 14: 257-263.

Pedrós-Alió C (1993) Diversity of bacterioplankton. Trends in Microbiology 8.

Pernthaler J, Simek K, Sattler B, Schwaderer A, Bobkova J & Psenner R (1997) Short-term changes of protozoan control on autotrophic picoplankton in an oligomesotrophic lake. J. Plank Res 18: 443-462.

Pernthaler A, Pernthaler J, Schattenhofer M, & Amann R (2002) Identi action of DNA-synthesizing bacterial cells in coastal North Sea plankton. Appl Environ Microb 68: 5728-57-36

Pernthaler J (2005) Predation on prokaryotes in the water column and its ecological implications. Nature reviews Microbiology 3: 537-546.

Pernthaler J & Amann R (2005) Fate of heterotrophic microbes in pelagic habitats: Focus on populations. Microb Mol Bio Rev 69: 440.

Pinhassi J, Zweifel UL & Hagstrom A (1997) Dominant marine bacterioplankton species found among colony-forming bacteria. Appl Envión Microb 63: 3359-3366.

Pinhassi J & Berman T (2003) Differential growth Response of colony-forming alpha- and gamma-proteobacteria in dilution culture and nutrient addition experiments from lake Kinneret (Israel), the Eastern Mediterranean Sea, and the Gulf of Eilat. Appl Envión Microb 69: 199-211.

Pinhassi J & Hagstrom A (2000) Seasonal succesion in marine bacterioplankton. Aquat Microb Ecol 21: 245-256.

Pinhassi J, Sala M, Havskum H, Peters F, Guadayol O, Malits A & Marrasé C. (2004) Changes in bacterioplankton composition under different phytoplankton regismens. Appl Envir Microb 70: 6753-6766

Pinhassi J, Gómez-Consarnau L, Alonso-Sáez L, Sala M, Vidal M, Pedros-Alió C, Gasol JM (2006) Seasonal changes in bacterioplankton nutrient limitaion and their effects on bacterial community composition in the NW Mediterranean Sea. Aquat Microb Ecol 44: 241-252

Pommier T, Pinhassi J & Hagstrom A (2005) Biogeographic analysis of ribosomal RNA clusters from marine bacterioplankton. Aquat Microb Ecol 41: 79-89.

Pommier T, Canbäck L, Riemann L, Boström KH, Simu K, Lundberg P, Tunlid A & Hagström A (2007) Global patterns of diversity and community structure in marine bacterioplankton. Mol Ecol.

Ramette A (2007) Multivariate analyses in microbial ecology. FEMS Microbiology Ecology 62: 142-160.

Ramsing NB, Fossing H, Ferdelman TG, Andersen F & Thamdrup B (1996) The distribution of bacterial populations in a strati ed Fjord (Mariager Fjord, Denmark) quanti ed by in situ hybridization and related to chemical gradients in the water column. Appl Envir Microb 62: 1391-1404.

Rappé MS, Connon SA, Vergin KL & Giovannoni SJ (2002) Cultivation of the ubiquitous SAR11 marine bacterioplankton clade. Nature 418: 630-633.

Revilla M, Iriarte A, Madariaga I & Orive E (2000) Bacterial and phytoplankton dynamics along a trophic gradient in a shallow temperature estuary. Est, Coast Shelf Sci 50: 297-313.

Rehnstam AS Backman S, Smith DC, Azam F, Hagström A (1993) Blooms of séquense-speci c of culture bacteria in the sea FEMS Microbiology Ecology 102: 161-166

Rheinheimer, G. & Gocke, K. (1994), The Use of Bacteriological Variables for the Characterization of Different Water Bodies. Int. Revue ges. Hydrobiol. 79: 605-619

214

References

Rivkin RB & Anderson MR (1997) Inorganic mutrient limitation of oceanic bacterioplankton. Limnology Oceanog 42: 730-740.

Rush DB, et al. (2007) The Sorcerer II Global Ocean Sampling Expedition: Northwest Atlantic through Eastern Tropical Paci c. PLOS biology 5: 0398-0431.

Sala M, Peters F, Gasol JM, Pedrós-Alió C, Marrasé C & Vaqué D (2002) Seasonal and spatial variations in the nutrient limitation of bacterioplankton growth in the northwestern Mediterranean. Aquat Microb Ecol27: 47-56.

Sanders RW, Porter KG, Bennett SJ, DeBrase AE (1989) Seasonal patterns of bacteriovory by agellates, ciliates, rotifers and cladocerans Limn Oceanogr 34:673-687

Santas, R., Korda, A., Tenente, A., Buchholz, K. and Santas P.H. (1999) Mesocosm assays of oil spill bioremediation with oleophilic fertilizers: Inipol, F1 or Both? Mar Pollut Bull 38: 44-48.

Sapp M, Wichels A, Wiltshire KH, Gerdts G (2007) Bacterail community dynamics during the winter-spring transition in the North Sea. FEMS Microbiol Ecol 59: 622-637

Schauer M, Balague V, Pedros-Alio C & Massana R (2003) Seasonal changes in the taxonomic composition of bacterioplankton in a coastal oligotrophic system. Aquat Microb Ecol 31: 163-174.

Schäfer H, et al. (2001) Microbial community dynamics in Mediterranean nutrient-enriched seawater mesocosms: changes in the genetic diversity of bacterial populations. FEMS Microbiology Ecology 34: 243-253.

Schendel EK, Nordström SE & Lavkulich LM (2004) Floc and sediment properties and their environmental distribution form a marine sh farm. Aqua Res 35: 483-493.

Sherr BF & Sherr EB (1984) Role of heterotrophic protozoa in carbon and energy glow in aquatic ecosystem. In: Klug. M.J., Reddy, C.A. (eds) Current perspectives in microbial ecology. American Society for Microbiology, Washington, DC. 412-423.

Selje N, Brinkhoff T & Simon M (2005) Detection of abundant bacteria in the Weser estuary using culture-dependent and culture-independent approaches. Aqua Microb Ecol 39: 17-34.

Shiah FK & Ducklow HW (1994) Temperature regulation speci c growht rate in Chesapeake Bay. Limn Oceanogr 39: 1243-1258.

Sieburth JM & Davis PG (1982) The role of heterotrophic nanoplankton in the grazing and nurturing of planktonic bacteria in the Sargasso and Caribbean Seas. Ann. Inst Oceangr 58: 285-296.

Siron, R., Pelletier, E., Delille, D., and Roy, S. (1993) Fate and effects of dispersed crude oil under icy conditions simulated in mesocosms. Mar Environ Res. 35: 273-302

Sogin ML, Morrison HG, Huber JA, Welch DBM, Huse SM, Neal PR, Arrieta JM & Herndl GJ (2006) Microbial diversity in the deep sea and the underexplored “rare biosphere”. PNAS 103: 12115-12120.

Sorokin YI 1977 The heterotrophic phase of plankton sucesión in the Japan Sea. Mar Biol. 41: 107-117

Staley JT & Konopka A (1985) Measurement of in situ activities of nonphotosynthetic microorganisms in aquatic and terrestrial habitats. Ann Rev Microb 39: 321-346.

Stevens H, Stubner M, Simon M & Brinkhoff T (2005) Phylogeny of Proteobacteria and Bacteroidetes from oxic habitats of a tidal at ecosystem. FEMS Microbiology Ecology 54: 351-365.

215

References

Teira E, Lekunberri I, Gasol JM, Nieto-Cid M, Álvarez-Salgado XA & Figueras FG (2007) Dynamics of the hydrocarbon-degrading Cycloclasticus bacteria during mesocosm-simulated oil spills. Environ Microb 9: 165-176.

Thingstad F, Hagstrom A & Rassoulzadegan F (1997) Accumulation of degradable DOC in surface waters: Is it caused by a malfunctioning microbial loop? Limn Oceanogr 42: 398-404.

Thingstad F, Zweifel UL & Rassoulzadegan F (1998) P limitation of heterotrophic bacteria and phytoplankton inb the northwest Mediterranean. Limn Oceanogr 43: 88-94.

Thingstad F (2000) Elements of a theory for the mechanisms controlling abundance, diversity, and biogeochemical role of lytic bacterial viruses in aquatic systems. Limn Oceanogr 45: 1320-1328.

Torsvik V, Ovreas L, Thingstad TF (2002) Prokaryotic diversity-magnitude, dynamics and controlling factors. Science 296:1064-1066

Unrein F, Alonso-Sáez L, Massana R & Gasol JM 2007. Signi cant year-round impact of small mixotrophic agellates on bacterioplankton in an oligotrophic coastal ocean. Limn Oceanogr. 52: 456-469

Urbach E, Vergin KL & Giovannoni SJ (1999) Immunochemical Detection and Isolation of DNA from Metabolically Active Bacteria. Appl Envir Microb 65: 1207-1213.

Venter JC, et al. (2004) Environmental genome shotgun sequencing of the Sargasso Sea. Science 304: 66-74.

Ward DM, Ferris MJ, Nold SC & Bateson MM (1998) A natural view of microbial biodiversity within hot spring cyanobacterial mat communities. Microb Mol Biol Rev 62: 1353-1370.

Wright TD, Vergin KL, Boyd PW & Giovannoni SJ (1997) A novel delta-subdivision proteobacterial lineage from the lower ocean surface layer. Appl Envir Microb 63: 1441-1448.

Wright RT & Cof n RB (1984) Factors affecting bacterioplankton density and productivity an salt marsh estuaries. In M.J. Klug and C.A. Reddy (eds). Current perspectives in microbial ecology. American Society of Micrbiologists 485-494.

ZoBell CE (1943) The effect of solid surfaces upon bacterial activity. J Bacteriol 46: 39-56

AGRADECIMIENTOS

Just can´t live that negative way…. Make way for the positive day!

Bob Marley

219

Aunque aún no me lo crea parece que ya llego el nal de esta tesis, donde queda re ejado el trabajo que he conseguido realizar gracias al apoyo y ayuda de tantísima gente que día a día han estado ahí para compartir cada momento y ofrecerme la mejor de sus sonrisas, pero también gracias a esa otra parte de mi que a pesar de la distancia siempre han estado tan cerca.

Primero de todo agradecer a mi director de tesis sin quién no habría sido posible este trabajo. Gracias por haberme apoyado desde el primer momento en el que me presenté en tu despacho (con un día de plazo para pedir la beca) y por todo lo que me has enseñado. Sobretodo gracias por tener esa actitud positiva para afrontar cada problema que tanto me ha motivado para realizar mi trabajo. Muchísimas gracias por darme la oportunidad de conocer el mar (tan lejano para mí) desde su punto de vista más pequeñito hasta el más grande en las campañas oceanográ cas. Mil gracias en este último periodo por todo el esfuerzo dedicado.

Mis primeros pasitos en la biología molecular fueron dados de la mano de Jarone, a quien agradezco enormemente que me transmitiera su gran pasión por la ciencia, pero también agradezco haber sido siempre tan cercano para mí. Muchas gracias por valorar siempre mi trabajo y seguir mis pasos aún desde Suecia.

En esta tesis han sido cruciales los consejos y correcciones de Carles, muchas gracias por toda la ayuda. A Silvia Acinas agradezco muy especialmente la ayuda y el toque tan especial que dio a mi trabajo. A Emilio por transmitirme sus conocimientos de ARB. A Jose González por contar con mi colaboración y por los varios favores de última hora. A Dolors por los días tan divertidos en Vigo y las aventurillas varias que lo rodearon (con parada en Lerma) y por irradiar siempre alegría. Mila esker horrela izateagatik! A Cèlia por los animos y ayuda especialmente esta última temporada. A Ramon, Cesc y Rafel por sus ánimos y comentarios. Y a todos los que habéis conseguido transmitir tan buen ambiente a todas las generaciones, mil gracias por hacer del ICM un lugar tan especial, de donde cuesta tanto marchar!

Muy especialmente quiero dar las gracias a Andrés que ha sido mi compañero en este viaje desde su inicio, sobre todo en los últimos años, con él que he compartido el estrés, la alegría, las penas, el hambre, el sueño... cada día. Como darte las gracias por haber sido siempre tan compresivo, animarme incluso en tus peores momentos y intentar hacerme ver los problemas de otra manera. Mil gracias por ser el mejor compañero y un gran amigo. Askenian lortu dugu! Orain ospatzeko.. Argentinara!

A Laura Gómez quiero agradecer su gran ayuda en todo momento y disponibilidad. Muchísimas gracias por darme tan buenos consejos y por estar ahí en momentos difíciles y transmitirme tranquilidad. Aunque te marcharas a Suecia (no te lo tendré en cuenta) ha sido un lujo que siempre consiguiéramos el encuentro. Quedan muchos otros!

Gracias a mis chicas por haber compartido conmigo las mejores sobremesas y hacer que volviese mi sonrisa incluso en la peor de las situaciones. A Vane por dar un toque tan especial a cada momento y conseguir sacarme de la monotonía, a Irene por transmitirme siempre mucha energía y optimismo, a Clarita por tantos y tantos favores y por haberme entendido siempre, a Dacha por haber cuidado también de sus niños, a la María que aunque ya marchó siempre alegró nuestras comidas, a Pilar

AGRADECIMIENTOS

220

por su agradable compañía y por su puesto no se me olvida el chico Javier que siempre tiene los comentarios más oportunos y me hace reír tanto. A todas y cada uno gracias por tan buenos consejos y por haber compartido conmigo los mejores momentos en el Instituto.

Gracias a Laurita por todo lo que me has enseñado y todo lo que hemos compartido juntas(desde jefe, viajes a piso) y espero que algún día nos volvamos a encontrar, quien sabe quizá Plentzian?

A Imelda que (aunque de Burgos) siempre consiguió hacerme reír y que tanto la eché de menos cuando se marcho, pero a pesar de todo siempre estuvo muy cerquita. Aún nos queda mucho por celebrar!!

A Fer que cuando estaba en Barcelona tanto me ayudo incluso acogiéndome en su casa, que también consiguió hacer más amena mi estancia en Holanda. Preparate boludo que en breve nos volveremos a ver!

A Pati por todos esos momentos compartidos y las interminables charlas que durante este último año tanto me ayudaron.

A Georgios por descubrirme lo fascinante de las profundidades del mar y estar siempre dispuesto a ayudarme. Gracias por dejarme el legado de enseñar el magni co programa de enmaquetación. A Cris por transmitirme siempre alegría y cuidar de mÍ esta última temporada y Lorenzo por sacarme siempre una sonrisa. A mis compis de piso Meri e Inma quienes siempre me hicieron reír. A Ale y Vero por la celebración tan especial del txupinazo en las calles de la Barceloneta. A mis venios italianos Ida y Danielle por las estupendas cenas compartidas.

A Jose Manuel sin quien no habrían dado lugar esas maravillosas fotos del SEM (a pesar de encontrar tan aburridas a las bacterias) Olga por su ayuda en molecular, a Enric por sus consejos de estadística, a Mickhail por su ayuda con las conversión de unidades, a Silvia también por su ayuda con la estadística, a Hugo por resolverme tantas dudas y a muchos otros que en momentos cruciales me echaron un cable. A Susana y Vicky las estudiantes que estuvieron de prácticas y tanto me ayudaron .

El trabajo que se realizó en Vigo fue gracias a la colaboración de mucha gente, gracias a todos por los esfuerzos que hicieron posible que saliera a delante, muy especialmente a Paco y Pepe por todas las sugerencias y ayuda. A todo el resto de gente involucrada Luisa, Belen, Mar, Isabel, Oscar, Pili, Eva, Jose, Maria y muchos otros. Por supuesto fue indispensable la ayuda de los asturianos expertos en citometría Xelu y Ale, muchas gracias a los dos por lo bien que lo pasamos juntos y por haber estado ahí en cada momento para lo que fuera.

A toda la gente que me ha ayudado en mi estancia en Holanda. Eva y Vero que tan bien me cuidaron y hicieron mis estancias más divertidas y emocionantes con la colaboración el de Bilbo y Goa. A mi compañero de bungalow Pedro que consiguió que cada día fuera un aventura en la pequeña isla. A Txetxu que justo marchaba cuando llegue y con quien pasé ratos tan divertidos, gracias lo compartido en Texel y en el Hesperides. Special thanks to Gerhard for all the help in my stays, for looking after me so well in any situation, promise next time I will not drive in Texel!

221

En mi estancia en Kalmar agradecer muy especialmente la hospitalidad de Neus y Jarone y a Nuria y la perrita: Twiggy que tanta compañía me hizo. A Anna por llevarme a conocer un Julefrokost en un granja donde conocí la esencia de la sociedad escandinava.

De las campañas realizadas quisiera agradecer a todos mis compañeros con los que compartí esta aventura tan especial para mí. Especialmente a Javier, Fede y Ale que los volví un poco loquitos con mis experimentos y a Nahiara, Eva, Nacho, Marc, Xavi, Ramón, Kim...mil gracias por tantas risas compartidas. También quiero agradecer al O cial Alfonso que tan bien me cuido.

A toda la gente que siempre ha tenido una sonrisa que ofrecer y tan bien me han hecho sentir, algunos que ya marcharon, algunas nuevas incorporaciones y a los de siempre MIL GRACIAS: Violeta, Gemma Oscar, Mariona, Ana Sabata, Estela, Hugo, Montse Sala, Albert René, Silvia Anglés, Nagore, Laura Arin, Elisa, Nuria Gómez, Sofía, Sonia, Ester, Montse, Raquel, Lorena, Bea, Elisabet, Arantza, , Clara, Juancho, Thomas, Ero, Mireia, Bego, Lidia, Andrea, Jordi, Xavi Leal, Rodrigo, Julia, Evarist, Albert Calbet, Eva Calvo, Carles, Eva Flo, Enrique Isla, Josep Maria, Cristina Linares, Cova, Sergi Rossi, Elisabeta...

Agradecer a Sara por toda la ayuda que me prestó para que viniese a Barcelona y a Jordi Dach por sus consejos de venir al ICM.

Toda la gente que estuvieron ahí pero más alejados de la ciencia:

A Antonia por buscarme el mejor hogar (“Parallel”) donde encontré una gran familia, y tuve un inversión total en la cultura catalana, gracias a todos y muy especialmente a Miquel, Pau, Jordi, Cunci y Serrano, por haber cuidado también de vuestra “chiqueta” en sus primeras andanzas por la gran ciudad, y por esas riquísimas paellas, arroz al forn y por todos los rincones que descubrimos juntos de Barcelona. Muy especialmente a Alfred por todo su cariño y apoyo en los inicios de esta aventura.

A Blanca y a Mateo por contar conmigo y entenderme, y a todos los del Bitacora, especialmente al Arnau y Juantxo.

A toda la cuadrrrrrrilla de Iruña: Maite, Izaskun, Sandra, Eva, Isa y Nerea, que han seguido toda la vida mis pasos día a día, sin perder detalle, para las que nunca la distancia fue un problema. Gracias por todas la visitas y por la amistad incondicional. Por ese sentimiento de cuadrilla que siempre nos ha mantenido juntas. Mila esker por todo lo vivido por esperar siempre son ansia mi llegada. A Maitetxu especialmente por su gran imaginación, sus buenos consejos de diseño y por las largas conversaciones de Skype que tanto me han ayudado. A todas gracias por escucharme siempre y hacerme sentir tan especial. Mil gracias por pedir que Sanfermin me echara un capote y conseguirlo! A pesar de que no veáis llegar el día de mi vuelta: Benetan laster itzuliko naiz!

A mis primicos: Mikel, Edurne, Amaia, Mirentxu, Lander, Iosu por venirme a visitar con tanta ilusión, por animarme siempre, por tantas conversaciones con las que hemos intentado arreglar el mundo, por

222

hacer que el hecho de estar juntos fuera tan especial. Por incluso alguno veniros a vivir a Barcelona y estar ahí siempre que ha hecho falta.

A mi tío Joaquim que rápidamente captó mi idea, muchas gracias por el diseño.

A Maria que durante todos estos años (y muchos más) estuviera donde estuviera nunca se separó de mí, que siempre me entendió. Por esos ratos compartidos en Leoz, Barcelona, Etxaleku... y por supuesto a Nuria muduaren onen amatxoa! Por comprenderme y mostrarme la vida de otra manera.

A Arantza que siempre estuvo ahí me fuera donde me fuera mil gracias por la amistad y con anza de saber que puedo contar contigo. Gracias a Dani quien me animó a salir de la vieja Iruña, a Felix y Amaia que fueron mi familia danesa y me animaron a venir. A Unai: dudarik gabe lagundu zenidan urte asko pasa izan arren, mila esker denagatik!

A toda la gente de Iruña que cada vez que voy me reciben con tanta ilusión Jon, Maite Aginaga, David, Monica, Mikel Goñi, Mikel de Goñi, Iosu, Maider, Eloisa...y muy especialmente a Aritz por todo el cariño.

Finalmente agradecer a mis padres y Maitanica que siempre con aron en mí y me dieron las energías, la tranquilidad y el cariño incondicional que me ha motivado cada día.

MUCHAS GRACIAS A TODOS!!!MOLTES GRACIES A TOTS!!!MILA ESKER DENORI!!!THANKS TO EVERYONE!!!