Community composition and interactions of biofilm bacteria on submerged freshwater macrophytes

Dissertation

zur Erlangung des akademischen Grades des Doktors der Naturwissenschaften (Dr. rer. nat.)

an der Universität Konstanz Mathematische-Naturwissenschaftliche Sektion Fachbereich Biologie

vorgelegt von Melanie Hempel

Konstanz, April 2008

Tag der mündlichen Prüfung: 06. Oktober 2008 Referent: PD Dr. Elisabeth Groß Referent: PD Dr. Hans-Peter Grossart Referent: Prof. Dr. B. Schink

Success consists of going from failure to failure without loss of enthusiasm.

Winston Churchill (1874-1965)

Table of Contents

Abbreviations i Summary in German – deutsche Zusammenfassung iii Summary in English – englische Zusammenfassung vii

I General Introduction 1 II Epiphytic bacterial community composition on two common submerged macrophytes in brackish water and freshwater Abstract 21 Introduction 22 Material & Methods 25 Results 28 Discussion 34 III Bacterial community composition of biofilms on two submerged macrophytes and an artificial surface in Lake Constance Abstract 43 Introduction 44 Material & Methods 47 Results 51 Discussion 60 Supplementary 64 IV Spatio-temporal dynamics of the bacterial biofilm on two freshwater macrophytes and an artificial substrate in Lower Lake Constance Abstract 69 Introduction 70 Material & Methods 72 Results 75 Discussion 85

V Impact of the polyphenol degrading bacterium Matsuebacter sp. FB25 on the growth of Acentria ephemrella larvae Abstract 95 Introduction 96 Material & Methods 98 Results 100 Discussion 107 VI Single- and multispecies biofilm formation of tannin degrading bacteria on an aquatic macrophyte Abstract 115 Introduction 116 Material & Methods 118 Results 121 Discussion 125 VII General Discussion and Perspectives 129 VIII References 139 Record of Achievement 159 List of Publications 161 Acknowledgement – Danksagung 165 Curriculum vitae 167 ABBREVIATIONS

AFDM Ash free dry mass ANOVA Analysis of variance BCC Bacterial community composition BLAST Basic local alignment search tool CFB Cytophaga-Flavobacteria-Bacteroidetes CLSM Confocal laser scanning microscopy DAPI 4ʹ,6–Diamidino–2–phenylindol DGGE Denaturing gradient gel electrophoresis dm Dry mass DNA Deoxyribonucleic acid DOC Dissolved organic carbon DOM Dissolved organic matter EPS Exopolysaccharide FA Formamide FISH Fluorescence in situ hybridization GIS Geographic information system HPLC High performance liquid chromatography IGKB Internationale Gewässerschutzkommission für den Bodensee MICRO-FISH Microautoradiography combined with FISH NaPPi Sodium pyrophosphate NMDS Non metric dimensional scaling OD600nm Optical density at 600 nm PCR Polymerase chain reaction PVPP Polyvinylpolypyrrolidon RFLP Restriction fragment length polymorphism rRNA Ribosomal ribonucleic acid SB Schaproder Bodden SD Standard deviation SDS-PAGE Sodium dodecylsulfate polyacrylamide gel electrophoresis SEM Scanning electron microscope TEP Transparent exopolymer particles UPGMA Unweighted pair group method with arithmetic mean

i

ii ZUSAMMENFASSUNG

Ziel dieser Arbeit war es, die bakterielle Biofilmgemeinschaft auf aquatischen

Oberflächen, insbesondere von submersen Makrophyten, zu untersuchen. Im

Mittelpunkt standen dabei die Zusammensetzung und Sukzession des bakteriellen

Biofilms, der mögliche Einfluss von Umweltfaktoren, Habitat und der Pflanzen

(Substrat) auf den Biofilm und die Interaktionen einzelner Isolate untereinander und mit aquatischen Herbivoren. In der Litoralzone von Seen bieten Makrophyten eine große Oberfläche zur Besiedlung von Bakterien und Algen. Zwischen Pflanzen und

Epiphyten finden häufig Interaktionen statt, die für beide Seiten positiv oder auch negativ sein können. Interaktionen zwischen Pflanzen und epiphytischen Biofilm können z.B. durch Oberflächenveränderungen oder ausgeschiedene organische

Substanzen stattfinden. Besonders Sekundärstoffe von Pflanzen (z.B. Phenole) sind bekannt dafür, dass sie andere Phototrophe oder Mikroorgansimen beeinflussen können.

Ich erwartete, dass das phenolreiche Tausendblatt Myriophyllum spicatum L. eine andere Biofilmzusammensetzung hat als das Laichkraut Potamogeton perfoliatus, die

Armleuchteralge Chara aspera oder künstliche Substrate (Plastikstreifen).

Myriophyllum spicatum scheidet algizide und bakterizide Polyphenole aus, während einige Chara-Arten algizid wirkende zyklische Thiane produzieren. Für Potamogeton perfoliatus ist nicht bekannt, dass es Polyphenole synthetisiert und Bakterien oder

Algen im Wachstum hemmt. Da M. spicatum zudem einen deutlichen Gradient von

Makronährstoffen und Polyphenolen von den jungen Apikalmeristemen zu den

älteren Blättern hinweg aufweist, war ein weiterer Aspekt dieser Arbeit, den Einfluss des Blattalters der Pflanzen auf die Biofilmzusammensetzung zu untersuchen. Chara aspera und M. spicatum kommen sowohl im Bodensee (Süßwasser) als auch im

Schaproder Bodden (Brackwasser) vor. Daher haben wir die bakterielle

Biofilmzusammensetzung auf beiden Pflanzen in beiden Habitaten verglichen. Alle

Biofilmanalysen dieser Arbeit sind mit molekularen Methoden wie FISH

iii (Fluoreszenz in situ Hybridisierung) und im Bodensee ergänzend mit DGGE

(Denaturierende Gradienten Gel Elektrophorese) und der Erstellung einer

Klonbibliothek durchgeführt worden. Durch die Kombination mehrerer Methoden wurden die erhaltenen Ergebnisse bestätigt und eventuelle Schwachstellen einer

Methode ausgeglichen.

Alle Untersuchungen des bakteriellen Biofilms lassen darauf schließen, dass die

Biofilme auf den jeweiligen Substraten von Bakterien der CFB-Gruppe dominiert waren. Alpha- und Betaproteobakterien waren am zweithäufigsten, während

Planktomyceten fast nur auf C. aspera im Schaproder Bodden gefunden wurden. Wir konnten einen deutlichen Einfluss des Habitats und des Substrates auf

Planktomyceten nachweisen, während CFB-Bakterien eher durch die Pflanze und das Blattalter beeinflusst wurden.

Beim Vergleich des Biofilms auf M. spicatum mit dem auf P. perfoliatus und

Plastikstreifen wurde deutlich, dass die Jahreszeit keinen Einfluss auf die

Zusammensetzung des bakteriellen Biofilms hat. Umweltfaktoren wie Wasserstand und –temperatur, Leitfähigkeit, pH und der Kohlenstoff- und

Gesamtpolyphenolgehalt der Pflanze haben hingegen einen Einfluss. Die bakterielle

Biofilmgemeinschaft auf P. perfoliatus und den Plastikstreifen ähnelte einander mehr als diejenige auf M. spicatum. Alle Ergebnisse aus den Analysen der

Biofilmgemeinschaften deuten darauf hin, dass gerade Apikalmeristeme von M. spicatum einen sehr heterogenen und speziellen Biofilm haben. Wir vermuten, dass dies mit dem erhöhten Polyphenolgehalt in diesen Abschnitten zusammenhängt. Die

Ergebnisse der Klonbibliothek bestätigen dies, und laut den Sequenzvergleichen mit

GenBank liegen viele dieser Biofilmbakterien auch größtenteils noch nicht in Kultur vor.

Wir konnten drei Isolate aus dem Biofilm (Pantoea agglomerans & Agrobacterium vitis) und dem Umgebungswasser (Matsuebacter sp.) von M. spicatum isolieren, die

iv Polyphenole abbauen können. Mit einem eigens entwickelten Versuchsaufbau haben wir untersucht, ob diese Isolate axenische M. spicatum besiedeln können.

Da Epiphyten zwangsläufig von Herbivoren aufgenommen werden, können diese so deren Verdauung und Darmflora beeinflussen. In Fütterungsexperimenten testete ich, ob das polyphenolabbauende Bakterium Matsuebacter sp. einen Einfluss auf das

Larvenwachstum des Wasserzünslers Acentria ephemerella (DENIS & SCHIFFERMÜLLER) hat.

Das Wachstum der Larven wurde im Vergleich zu axenischem M. spicatum weder negativ noch positiv beeinflusst, wenn sie mit Matsuebacter sp. besiedeltem M. spicatum gefüttert wurden. Daraus schließen wir, dass Matsuebacter sp. weder als zusätzliche Nährstoffquelle dient, noch die vorhandene Darmflora der Larve beeinflusst oder nennenswert die Polyphenole verändert hat.

Während der Einfluss von Matsuebacter sp. auf das Larvenwachstum vernachlässigbar war, ist es uns gelungen, eine dichte und schnelle Biofilmbildung dieses Bakteriums auf axenischem M. spicatum zu zeigen. Dadurch wird die

Biofilmbildung des landwirtschaftlich genutzten „Biocontrollers“ P. agglomerans gehemmt und diejenige des Pflanzenpathogens A. vitis verstärkt.

Mit dieser Arbeit ist es mir gelungen, einen Beitrag zu dem bisher recht spärlichen

Wissen über bakterielle Biofilm auf aquatischen Pflanzen zu leisten, weiterhin die

Interaktionen einzelner Bakterien miteinander zu beleuchten und auch den Bogen zu höheren trophischen Ebenen zu spannen. Einzelne Bakteriengruppen werden offensichtlich von dem jeweiligen Substrat und Habitat beeinflusst, während für ganze Bakteriengemeinschaften auch Umweltparameter wie Wasserstand und– temperatur und Leitfähigkeit, aber auch der Kohlenstoffgehalt und

Gesamtphenolgehalt der Pflanzen von Bedeutung sind.

v vi SUMMARY

The aim of my PhD thesis was to investigate the bacterial biofilm community composition (BCC), especially on submerged macrophytes. The special interest was the composition and succession of the heterotrophic biofilm and possible influences such as environmental factors, habitat and plants (substrate) on the biofilm and the interaction of isolates with each other and with aquatic herbivores. On the littoral zones of lakes, macrophytes offer a large area for colonization of bacteria and algae.

Interactions between plant and epiphytes are frequent and can be positive and negative for both sides. Interactions between macrophytes and epiphytic biofilm can be mediated by structural changes of the surface or by exuded organic compounds.

Especially secondary metabolites of Plants (e.g., phenols) are known to have an impact on other phototrophs or microorganisms.

I expected that the phenol-rich milfoil Myriophyllum spicatum L. would have a different BCC than the pondweed Potamogeton perfoliatus, the stonewort Chara aspera or artificial substrates (polypropylene sheets). Myriophyllum spicatum exudes algicidal and bactericidal polyphenols, while some Chara species produce algicidal cyclic sulphur compounds. It is not known if P. perfoliatus synthesizes polyphenols and if it may inhibit bacterial and algal growth. Another aspect of this work was to investigate the influence of leaf age on the BCC, since M. spicatum displays a distinct gradient of macronutrients and polyphenols from young apical meristems to older leaves. Both Char aspera and M. spicatum occur in Lake Constance (freshwater) and in the Schaproder Bodden (brackish water). Thus, we compared the BCC on both macrophytes in both habitats. All analyses of the BCC in this study have been done with FISH (fluorescence in situ hybridization) and in Lake Constance additionally with DGGE (denaturing gradient gel electrophoresis) and the construction of a clone library. With the combination of several methods, I could verify the obtained results and possible disadvantages of one method could be evened out by the other.

vii All investigations of the biofilm community lead to the conclusion that biofilms on the respective substrates were dominated by bacteria of the CFB-group. Alpha- and betaproteobacteria were the second most abundant groups, while planctomycetes were only found on brackish water C. aspera. Planctomycetes were largely influenced by the habitat and the substrate type (plant species) while bacteria of the CFB-group were rather influenced through plant species and leaf age.

The BCC comparison on M. spicatum, P. perfoliatus and the artificial substrates was not influenced by season. However, environmental factors such as water level and temperature, conductivity, pH and the carbon and total phenolic content of the plant tissue influenced the bacterial biofilm community composition. The BCC on the artificial substrates was more similar to that on P. perfoliatus than to that on

M spicatum. The results obtained in all those community studies revealed a rather distinct and heterogeneous BCC on M. spicatum apices. We assume that this is a consequence of the high polyphenol content in these plant parts. The data obtained in the clone library support this finding. According to GenBank, most of the sequences obtained in the clone library do belong to bacteria not yet cultured.

We were able to isolate three bacterial strains from the biofilm (Pantoea agglomerans

& Agrobacterium vitis) and the surrounding water (Matsuebacter sp.) of M. spicatum.

All three are able to degrade polyphenols. With an especially designed experimental set-up, we tested if the three isolates were capable to colonize axenic M. spicatum.

Since epiphytes are taken up inevitably during feeding of herbivores they may have an impact on digestion and gut microbiota. In no choice feeding experiments I investigated, if the polyphenol degrading Matsuebacter sp. has an impact on the larval growth of the aquatic moth Acentria ephemerella (DENIS & SCHIFFERMÜLLER).

In comparison to axenic M. spicatum, plants colonized with Matsuebacter sp. had no negative or positive impact on larval growth. Thus we conclude that Matsuebacter sp. neither serves as an additional nutrient source, nor influences the gut microbiota or alters the exuded plant polyphenols. viii While the influence of Matsuebacter sp. on larval growth was negligible, we could prove that this bacterium forms dense biofilms on M. spicatum rather quickly. The presence of Matsuebacter sp. reduces the biofilm formation of the agriculturally used bio control agent P. agglomerans, and thus that of the plant pathogen A. vitis is enhanced.

With this work, I contributed to the scarce knowledge on bacterial biofilms on aquatic plants. Further I elucidated the biofilm formation and interactions of single strains and investigated their impact on higher tropic levels. Single bacterial groups are obviously influenced by the substrate and habitat type. Bacterial communities in their whole are rather determined by environmental factors like water level and temperature, and conductivity, and also the plant carbon and total phenolic content.

ix

x I General Introduction

Chapter I

General Introduction

1 I General Introduction

2 I General Introduction

FUNCTIONS OF MACROPHYTES IN LITTORAL ZONES OF LAKES In the littoral zones of lakes, macrophytes are the major primary producers and a vital structural element. Especially in shallow eutrophic lakes, they often are responsible for clear water states, caused by light and nutrient competition advantages over algae or allelopathy (Scheffer et al. 1993, Hilt & Gross 2008).

Submerged macrophytes reduce sediment resuspension. They transport selected mineral nutrients taken up by the root through the lacunar system to the leaf surface. For example, methane and manganese are released in substantial amounts for epiphytic bacteria and algae (Jackson et al. 1994, Schuette 1996, Heilman &

Carlton 2001). Besides providing shelter for young fish and zooplankton, macrophytes are also frequently consumed by waterfowl. Macrophytes provide a large surface area for egg deposition of snails and fish and for the colonization of microorganisms such as algae, fungi and bacteria. Thus, they fulfil many functions in this heterogeneous and active habitat.

With a total volume of 48.4 km³, the prealpine Lake Constance contributes

0.04% of all freshwater lakes on earth (Figure 1.1A). Its shore line is 273 km long and comprises the larger and deeper oligotrophic Upper Lake (186 km shoreline;

101 mean depth; 8 mg total phosphorus m-3) and the smaller mesotrophic Lower

Lake (87 km shoreline; 13 m mean depth; 17 mg total phosphorus m-3; (IGKB 2006,

2007)). Especially in the Upper Lake, the phosphorus concentration has been declining constantly in the past years.

In lakes, the littoral zone is usually defined as the near shore area where sunlight penetrates all the way to the sediment and allows macrophytes or other benthic primary producers to grow. Light levels of about 1% or less of surface values usually define this depth. In 1993, the littoral zone of Lower Lake was

18 km² as determined by GIS measurements (geographic information system;

(Schmieder 1997)).

3 I General Introduction

Figure 1.1. Lake Constance and study site A) Lake Constance (Circle indicates Island of Reichenau) B) Island of Reichenau C) Study site located at Niederzell.

Pictures are taken from: A) www.igkb.de B) ww.insel-reichenau.de C) Dr. T. Heege, DLR Oberpfaffenhofen.

According to this method, the Upper Lake had littoral zone of 57 km² contrasting

18 km² in the Lower Lake with the littoral zone defined as the area above 10 m water depth. Due to the ongoing reoligotrophication, light is no longer a limiting factor for macrophyte growth and with an increased maximum depth of macrophytes, the littoral zone of Lower Lake has most likely increased.

The sampling site of this study was located in lower Lake Constance near a gravel ridge close to the Island of Reichenau in 2 – 3 m water depth (N47°42,247,

E9°02,289; Figure 1.1Band C). This site is at the wind exposed north-western part of the island and water currents are higher than in the northern part. While

Myriophyllum spicatum (Figure 1.2A) is found on the ridge or on the slope of it, alongside the ridge dense mats of Chara spp. cover the sandy sediment

(Figure 1.2B). In this bay two larger Potamogeton perfoliatus stands (Figure 1.2C) are located close to the shore on the right and left hand side of the ridge.

4 I General Introduction

Figure 1.2. Aquatic substrates investigated in this study. A) Myriophyllum spicatum stand in a dense Chara sp. mat B) Chara sp. C) Potamogeton perfoliatus patch C) Exposed polypropylene sheets used as artificial substrates. All pictures by the courtesy of Dr. M. Mörtl.

IMPORTANCE OF BACTERIAL BIOFILMS IN AQUATIC HABITATS

Bacteria occur ubiquitous in nature. They are found in soil, marine and freshwater, sewage sludge and even in extreme environments such as hydrothermal vents (Hugenholtz et al. 1998). In the last decades, it has been commonly acknowledged that bacteria prefer an attached life style if nutrient conditions are favourable and thus are mainly found on surfaces (Costerton et al.

1995, Stanley & Lazazzera 2004). Since then, much research has been done on biofilms, their role and function in wastewater treatment, health care, industries and ecology (Paerl & Pinckney 1996, Morris & Monier 2003, Pasmore & Costerton

2003, Parsek & Fuqua 2004, Stanley & Lazazzera 2004). Bacteria, if in biofilms or not, contribute to the overall nutrient cycling in pelagic and littoral zones, and in streams (Ardon & Pringle 2007). They are involved in the degradation of sinking dead organisms (algae, zooplankton), are associated with lake snow, produce

5 I General Introduction metabolites that can be used further and are the essential part of the microbial loop (Riemann et al. 2000, Cotner & Biddanda 2002). In industrial systems and in water supply biofilms often cause problems and immense costs since they are hard to remove, induce corrosion and are a source of contamination (Pasmore &

Costerton 2003).

The formation of biofilms has many ecological implications for bacteria and their surfaces. Biofilms found in nature are usually multispecies biofilms, embedded in a matrix of extracellular polymeric substances (EPS), in which bacteria with different metabolically characteristics coexist and may act as symbionts (Eberl 1999, Burmolle et al. 2006). The EPS matrix surrounding the biofilm prevents the cells from desiccation, and channels that form in the matrix are used to transport substrates or proteins, which would otherwise be lost in the surrounding water (Czaczyk & Myszka 2007). It is well known from clinical studies that bacterial biofilms are more resistant to antibiotics or detergents than single cells due to the surrounding matrix (Stewart & Costerton 2001, Burmolle et al. 2006, Harrison et al. 2007). This aspect can certainly be transferred to nature since bacteria are often exposed to bactericidal substances in the environment

(Burmolle et al. 2006). Bacteria in biofilms are also more resistant to grazing by flagellates. Assumingly, this is caused by the thickness of the biofilm and the EPS– matrix, which makes them less accessible (Jürgens & Matz 2002). Further, biofilms provide a very structured and heterogeneous habitat on a very small scale. Cells at the bottom of the biofilm often experience anaerobic conditions while cells in the outer layers are more exposed to grazing, toxins or UV–radiation.

One mechanism in biofilm formation that has received much attention in the past years is quorum sensing. It is defined as cell density dependent gene regulation and is mediated via signal molecules, e.g., N–acylhomoserinlactones in gram– negative or peptides or butyrolactone in gram–positive bacteria (Bibb 1996,

Kleerebezem et al. 1997, Chhabra et al. 2005). Several functions have been shown to be regulated by quorum sensing, among them siderophore production, motility,

6 I General Introduction pathogenicity, bioluminescence, biofilm formation and EPS production (Guan et al. 2000, Hammer & Bassler 2003, Lupp et al. 2003, Marketon et al. 2003). In the past, quorum sensing–studies focussed mainly on medical implications. Studies on the ecological relevance of quorum sensing are scarce (Manefield & Turner 2002) but gave hints that it might not be as important as assumed (Styp von Rekowski et al. 2008). In preliminary experiments, I tested whether bacteria capable to use tannic acid as a sole carbon source produce N–acylhomoserinlactones. In bioassays, I found several strains that produce those signal molecules but found no evidence that those strains grew faster than others that did not produce these compounds. Other studies found that plant derived furanones and polyphenols may interfere with quorum sensing regulated biofilm formation (see below;

(Hentzer et al. 2002, Huber et al. 2003)). Thus, plants and algae have developed strategies against this mechanism, indicating that the ecological role of quorum sensing is controversial and needs further investigation.

Studies on biofilms (initial colonization, succession, interactions) are often conducted in artificial systems with only two or three strains. The advantage of those set ups is that all environmental parameters can be kept constant. Further, the selected strains can be modified genetically, which is attractive to investigate the importance of certain genes in biofilm formation or interactions between different bacteria in the biofilm. Biofilms are usually investigated by a combination of different dyes and microscopy. With the introduction of fluorescence in situ hybridization, green fluorescent protein labelled strains and confocal laser microscopy it has become possible to observe biofilm formation from the beginning on in situ. Researchers have shown that biofilms do not consist of several flat layers but form mushroom like structures that are connected with transport channels (De Beer et al. 1994, Picioreanu et al. 2000). Further, biofilms are no static communities but are shaped by settling and sloughing of organisms.

Thus, biofilms are characterized by a constantly changing community composition in which participants have to react and act constantly to new metabolic partners or

7 I General Introduction even competitors from the beginning on (Grossart et al. 2003, Kiorboe et al. 2003,

Pasmore & Costerton 2003).

On the surface of M. spicatum, bacteria will also live as a biofilm ‘entity’ that interacts in multiple ways To investigate biofilm formation of selected strains and their interactions on M. spicatum leaves, we developed a set up, which allows to do exactly this. We chose three strains originating from the surrounding water

(Matsuebacter sp.) and the biofilm of M. spicatum (Agrobacterium vitis and Pantoea agglomerans) all with different characteristics and abilities to degrade polyphenols.

Matsuebacter sp. (Mitsuaria sp.; betaproteobacteria) is able to constitutively degrade polyphenols (Müller et al. 2007), while A. vitis, a grapevine pathogen, can only degrade polyphenols if the degradation pathways are induced (Ophel & Kerr

1990, Müller et al. 2007). The gammaproteobacterium P. agglomerans is frequently used in agriculture as a biological control agent and possesses the required enzymes for phenol degradation (Zeida et al. 1998). It is further found throughout the world in the gut of ruminants and herbivores (Nelson et al. 1998, Pidiyar et al.

2004).

8 I General Introduction

INTERACTIONS OF BACTERIA WITH PHOTOTROPHS AND HERBIVORES

In aquatic systems it is well known that epiphytes and plants do interact in various ways (Beattie & Lindow 1999, Kubanek et al. 2003, Mathesius et al. 2003).

In general it can be assumed that epiphytes have quite an impact on host organisms and vice versa.

As mostly sessile organisms, plants have to defend themselves against potential enemies, e.g., herbivores, epiphytic algae and pathogenic bacteria or fungi. Thus, they developed several methods to hold off enemies. Most prominent examples are thorns or spikes, thick leaves and also chemical defence. In aquatic systems, chemical defence is most commonly used against other phototrophs or microorganisms. Chemical defence results in compounds that taste bad to herbivores, or inhibit epiphytic phototrophs or heterotrophic microorganisms. In

1937, Hans Molisch coined the term allelopathy for plant-plant and plant-bacteria interactions. Allelopathy covers biochemical interactions, both stimulatory and inhibitory, among different primary producers or between primary producers and microorganisms (Molisch 1937). Here, we focus on and elucidate the impact of allelochemicals produced by different aquatic plants on the bacterial community composition in the epiphytic biofilm, especially those of Myriophyllum spicatum.

Since organic metabolites always leak from the macrophyte tissue (Godmaire &

Planas 1986), epiphytic bacteria should be directly affected by these compounds.

Biofilms on plants may have beneficial or detrimental effects to their host. First of all, they reduce the light availability of the plant, especially if algae or phototrophic bacteria are embedded. Further, the biofilm may contain pathogens that invade plant cells or promote biofouling. On planktonic algae, attached bacteria increase the sinking speed, and the algae reach light limitation earlier than without the biofilm (Grossart et al. 2005). On the other hand, the biofilm may lead to reduced grazing, and the plant can benefit from degraded compounds and CO2 produced by bacteria. Phototrophs might even influence the bacterioplankton community composition. The macrophyte Vallisneria americana influences the

9 I General Introduction bacterioplankton community indirectly over DOC (dissolved organic carbon) and phosphorus release (Huss & Wehr 2004). If the ammonia level was high, the bacterioplankton numbers were positively influenced by DOC and phosphorus release of the plant. If ammonia was limiting, non–rooted plants had a negative effect on bacterioplankton abundance, while rooted macrophytes had no effect.

Further, the mode of interaction between plants and phototrophs can be influenced by the trophic state (Danger et al. 2007a, Danger et al. 2007b). Here, nitrogen limitation of algae resulted in commensalism, phosphorus limitation in competition and nutrient rich situations in mutualism.

In the past, many studies described allelochemical interactions between bacteria and phototrophic organisms. The marine red algae Bonnemaisonia hamifera and

Delisea pulchra produce and release furanones that inhibit the settling of bacteria by interfering with quorum sensing regulated biofilm formation (Maximilien et al.

1998, Huber et al. 2003, Nylund et al. 2005). These furanones also affect the swarming motility of Serratia liquefaciens and indirectly larval attachment

(Rasmussen et al. 2000, Dobretsov et al. 2007). Positive effects of bacteria were found for Roseobacter gallaciencis and Pseudoalteromonas tunicata on the seaweed

Ulva australis where they produce antifouling compounds (Rao et al. 2006). Axenic

Ulva linza need bacteria to promote growth and restore their growth form (Matsuo et al. 2005, Marshall et al. 2006).

For all plant species investigated in this study, the existence of allelopathic compounds has been reported. In Eurasian watermilfoil Myriophyllum spicatum, tellimagrandin II, a hydrolysable polyphenol (Figure 1.3A), is present in concentration of 5% of the dry mass (up to 25% of all tannins on apices and leaves), and is responsible for inhibition of the photosystem II in cyanobacteria

(Leu et al. 2002). Further, plant polyphenols inhibit the gut microbiota of the herbivorous larvae Acentria ephemerella (DENIS & SCHIFFERMÜLLER, Figure 1.4A and

B), and are supposed to retard the growth of this larvae (Choi et al. 2002,

Walenciak et al. 2002). To compare the impact of plant chemistry on the bacterial

10 I General Introduction community composition of M. spicatum, we chose the pondweed Potamogeton perfoliatus that contains hardly any phenolic compounds (Choi et al. 2002). It grows in the vicinity of M. spicatum and although algae-inhibiting compounds, mostly di-terpenes, have been described for several Potamogeton species, none of these effects or compounds have been described and isolated from P. perfoliatus

(DellaGreca et al. 2001). We also investigated the bacterial community composition of a third macrophyte, the macroalgae Chara aspera. Some Chara species contain cyclic sulphur compounds that inhibit bacteria and algae (Figure 1.3C; (Anthoni et al. 1987)).

A) C) S H C S 3 S

B) S

H3C S S S

Figure 1.3. Chemical compounds in freshwater macrophytes. A) Tellimagrandin II B) Tannic acid (http://www.pharmainfo.net) C) Trithiane and dithiane from Chara globularis.

The herbivore Acentria ephemerella causes substantial damage to aquatic plants such as P. perfoliatus and M. spicatum (Gross et al. 2002). Although it fully develops on the latter, its growth is retarded compared to polyphenol-free P. perfoliatus

11 I General Introduction assumingly due to the high polyphenol content of M. spicatum (Choi et al. 2002).

Many studies showed the importance of gut bacteria for insects, esp. herbivores (Ji et al. 2000, Ji & Brune 2001, Dillon & Dillon 2004) and it has been assumed that the negative impacts of plant tannins on herbivores are often a consequence of plant- bacteria interactions (Schultz et al. 1992). This could especially be the case for bacteriostatic hydrolysable polyphenols. In the case of A. ephemerella, isolated gut bacteria were inhibited in the presence of M. spicatum-derived tannins. To test if tannin-degrading bacteria will have an influence on the growth of A. ephemerella larvae, we conducted no choice experiments in which larvae were fed

Matsuebacter sp.-colonized M. spicatum leaves in comparison to axenic plants or those colonized by a natural bacterial biofilm. Matsuebacter sp. grows constitutively with tannic and gallic acid (see above).

Figure 1.4. A) Aquatic moth Acentria ephemerella B) A. ephemerella larvae of different instar ages (courtesy of Dr. E.M. Gross)

BACTERIAL GROUPS ON DIFFERENT SURFACES AND HABITATS

Surprisingly, if the ISI web of science database is searched for the terms

‘bacteria’, ‘biofilm’, ‘freshwater’, ‘community’ and ‘composition’ in the years 1956 to 2008 (week 16), only nine hits are displayed. If the term ‘freshwater’ is not included, the number of hits rises to 116. This is surprising, since biofilms have received much attention in the last decade but obviously, freshwater biofilms did not. Many studies that deal with freshwater biofilms often focus on certain bacterial groups rather than on whole communities (Brümmer et al. 2004,

12 I General Introduction

Tadonléké 2007). Thus, studies describing whole biofilm communities are scarce, especially on macrophytes.

The invention and introduction of molecular methods to microbial ecology has increased our knowledge of bacterial communities immensely over the last two decades (Head et al. 1998). Especially methods like denaturing gradient gel electrophoresis (DGGE) and fluorescence in situ hybridisation (FISH) contributed to this gain in knowledge. They allow to process large sample numbers and provide data not only of community structures (DGGE), but also on species composition and spatial distribution within a given community (FISH).

In the past, many aquatic bacterial communities have been described on the basis of these new methods in various habitats, giving evidence that most aquatic bacteria occur throughout the world, though in various abundances.

Betaproteobacteria, for example, are commonly found in freshwater habitats, while they do not occur in the marine bacterioplankton (Glöckner et al. 1999).

Aquatic biofilm communities have been described on and in various substrates such as sediments, chlorophytes, diatoms, sponges, lake and marine snow, stones, steel foil, ceramic tiles or propylene sheets in marine and freshwater by various molecular methods for the 16S rRNA gene or functional genes.

On green algae and diatoms as well as on lake snow and copepod carcasses, researchers found high abundance of the different proteobacteria groups and of

Cytophaga-Flavobacteria-Bacteroidetes (CFB). The abundance of different groups was somewhat dependent on the substrates investigated. Lake and marine snow are often composed of dead zooplankton or algae, but also transparent exopolymer particles (TEP) are important hot spots for the microbial degradation of organic matter (Simon et al. 2002). On Lake Constance lake snow, mainly alpha- and betaproteobacteria were found in the epilimnion, while in the hypolimnion bacteria of the CFB group were dominant (Schweitzer et al. 2001).

Gammaproteobacteria were of minor importance on particles from all depth investigated. The researchers assume that a limited number of alpha- and

13 I General Introduction betaproteobacteria first consume and then release amino acids, while in deeper water layers CFB-bacteria consume the refractory components. On diatom microaggregates, initially alphaproteobacteria dominate, while betaproteobacteria and CFB increase towards the end. Brachvogel and colleagues defined four types of microaggregates with different bacterial communities. Here, betaproteobacteria dominated particles originating from diatoms, while on “DAPI yellow particles” betaproteobacteria and CFB-bacteria were equally abundant. Interestingly, gammaproteobacteria, which were not numerous in other studies, were most abundant on particles originating from zooplankton and phytoplankton (62% of

DAPI counts; (Brachvogel et al. 2001)). Although in all these ‘Lake Constance’ studies only probes for alpha-, beta- and gammaproteobacteria and CFB were used, these groups accounted for more than 50% of all cells detected and thus, the community on lake snow aggregates in Lake Constance seems to be made out of a few bacterial groups.

While community composition of aggregated bacteria in Lake Constance seemed rather constant, the community composition of aggregated bacteria in

Lake Baikal was different. Here, gammaproteobacteria were quite abundant in all water layers, except for 100 m depth. Bacteria of the CFB-group were not as abundant and betaproteobacteria increased with depth (Ahn & Burne 2007).

Several studies dealing with freshwater biofilms in streams found high abundances of betaproteobacteria and CFB. Alphaproteobacteria were frequently found while gammaproteobacteria were of minor importance (Brummer et al.

2000, Olapade & Leff 2004, 2005, 2006). Olapade and co-workers proved very nicely that the BCC depends on the type and quality of dissolved organic matter

(DOM) provided. Labile DOM (e.g., glucose) evoked more response in all bacterial groups investigated than leaf leachate, algal exudates or inorganic nutrients

(Olapade & Leff 2005).

Descriptions of the heterotrophic biofilm community composition (BCC) on macrophytes are scarce but do exist for Myriophyllum spicatum and some

14 I General Introduction

Potamogeton species. The BCC on M. spicatum has been described by cultivation based methods at the beginning of 1990 (Chand et al. 1992) and again with DGGE in 2000 (Walenciak 2004). The latter study described differences between attached and free living communities.

AIMS OF THIS STUDY

The main objective of my Ph.D. thesis was to increase the scarce knowledge on bacterial biofilms on freshwater macrophytes and to investigate patterns and possible processes determining their community structure. Whole community compositions on different macrophytes in different habitats have not been described before. Although environmental factors as DOM, chlorophyll content and conductivity have been identified to determine differences of bacterial communities from different habitats, so far no one attempted to investigate the influence of plant age, plant species and plant chemistry on the bacterial community composition (BCC). It is important to understand internal and external influences on bacterial dynamics, since they might further influence higher trophic levels.

Based on my initial studies in Lake Constance, we asked how BCC is influenced by different habitats and plants (Chapter 2). We chose a higher plant

(Myriophyllum spicatum) and a macroalga (Chara aspera), both occurring in the same habitats (Lake Constance and Schaproder Bodden). In cooperation with Maja

Blume from the University of Greifswald, we investigated differences and similarities of the BCC on young and old plant parts of both macrophytes in both habitats with fluorescence in situ hybridization (FISH).

In Chapters 3 and 4, I focus on spatial and temporal differences of the BCC on

M. spicatum in comparison to other substrates. First, I investigated the seasonal and within-plant differences of the BCC on this macrophyte with denaturing gradient gel electrophoresis (DGGE) during summer 2005. I extended these investigations in summer 2006 to explore whether the BCC depends on the type of

15 I General Introduction substrate, the plant age (apices and leaves) or the season. Environmental parameters and plant chemistry were measured as potential determining factors. I investigated the BCC on Myriophyllum spicatum, Potamogeton perfoliatus and polypropylene sheets used as artificial substrate (Figure 1.2D) with DGGE

(Chapter 3) and FISH (Chapter 4).

Herbivorous insects may take up biofilm bacteria with their host plants. We knew that M. spicatum derived tannins decrease the growth of larvae of the aquatic moth Acentria ephemerella and inhibit selected strains of the gut microbiota. We further had bacteria isolated from M. spicatum stands that are able to degrade the plants’ tannins. Therefore, I compared the impact of axenic (= bacteria-free) plants to those with a biofilm of Matsuebacter sp. or plants with a natural bacterial biofilm on the growth of A. ephemerella larvae in no choice feeding experiments

(Chapter 5).

Finally, we were interested in the recolonization of M. spicatum with previously isolated bacterial strains and how these strains would interact (Chapter 6). In her diploma thesis, Sonja Wicks developed, based on my findings and with my guidance, an experimental set up to follow the recolonization of M. spicatum and the interactions of the different proteobacterial strains Matsuebacter sp.,

Agrobacterium vitis and Pantoea agglomerans by DAPI staining and FISH.

16 I General Introduction

17 I General Introduction

18

Chapter II

Epiphytic bacterial community composition on two common submerged macrophytes in brackish water and freshwater

Melanie Hempel, Maja Blume, Irmgard Blindow & Elisabeth M. Gross

BMC Microbiology 8(1):58

II Bacterial Biofilms on submerged macrophytes

20 II Bacterial biofilms on submerged macrophytes

ABSTRACT: Plants and their heterotrophic bacterial biofilm communities possibly strongly interact, especially in aquatic systems. We aimed to ascertain whether different macrophytes or their habitats determine bacterial community composition. We compared the composition of epiphytic bacteria on two common aquatic macrophytes, the macroalga Chara aspera Willd. and the angiosperm Myriophyllum spicatum L., in two habitats, freshwater (Lake Constance) and brackish water (Schaproder Bodden), using fluorescence in situ hybridization. The bacterial community composition was analysed based on habitat, plant species, and plant part. The bacterial abundance was higher on plants from brackish water [5.3×107 cells (g dry mass)–1] than on plants from freshwater [1.3×107 cells (g dry mass)-1], with older shoots having a higher abundance. The organic content of freshwater plants was lower than that of brackish water plants (35 vs. 58%), and lower in C. aspera than in M. spicatum (41 vs. 52%). The content of nutrients, chlorophyll, total phenolic compounds, and anthocyanin differed in the plants and habitats. Especially the content of total phenolic compounds and anthocyanin was higher in M. spicatum, and in general higher in the freshwater than in the brackish water habitat. Members of the Cytophaga– Flavobacteria–Bacteroidetes group were abundant in all samples (5–35% of the total cell counts) and were especially dominant in M. spicatum samples. Alphaproteobacteria were the second major group (3–17% of the total cell counts). Betaproteobacteria, gammaproteobacteria, and actinomycetes were present in all samples (5 or 10% of the total cell counts). Planctomycetes were almost absent on M. spicatum in freshwater, but present on C. aspera in freshwater and on both plants in brackish water. Bacterial biofilm communities on the surface of aquatic plants might be influenced by the host plant and environmental factors. Distinct plant species, plant part and habitat specific differences in total cell counts and two bacterial groups (CFB, planctomycetes) support the combined impact of substrate (plant) and habitat on epiphytic bacterial community composition. The presence of polyphenols might explain the distinct bacterial community on freshwater M. spicatum compared to that of M. spicatum in brackish water and of C. aspera in both habitats.

21 II Bacterial Biofilms on submerged macrophytes

INTRODUCTION In aquatic systems, bacteria occur often associated with surfaces, e.g., in biofilms or on lake or marine snow (Costerton et al. 1995). Biofilm associated bacteria are most abundant at intermediate nutrient availability while either low or high nutrient conditions favour planktonic growth of bacteria (Stanley

& Lazazzera 2004). Biofilms are not only formed on abiotic surfaces but also on living organisms such as aquatic plants and algae.

In freshwater and marine habitats, bacteria associated with cyanobacterial blooms, diatom blooms, phytoplankton (Jasti et al. 2005), lake snow (Weiss et al. 1996), and bacterioplankton (Glöckner et al. 1999, Shade et al. 2007) have been investigated. Betaproteobacteria occur almost exclusively in freshwater but not in saline habitats, while alphaproteobacteria are more abundant in marine than in freshwater samples (Glöckner et al. 1999). Alphaproteobacteria dominate the planktonic bacteria in the North Sea, followed by the

Cytophaga–Flavobacteria–Bacteroidetes (CFB) group, and all groups of bacteria display a seasonal succession (Sapp et al. 2007). Diverse bacterial communities dominate in cyanobacterial blooms, including members of the

CFB group and betaproteobacteria (Eiler & Bertilsson 2004). Mainly members of the CFB group and alphaproteobacteria, especially Roseobacter, are attached to marine diatoms (Riemann et al. 2000, Grossart et al. 2005). Members of the

CFB group and alpha-, beta-, and gammaproteobacteria have been identified by molecular methods on the chlorophytes Desmidium devillii, Hyalothexca dissliens, and Spondylosium pulchrum (Fisher et al. 1998). In general, the bacteria associated with diatoms and some chlorophytes that have been studied are mostly heterotrophic. In contrast, information about bacterial biofilms on aquatic macrophytes is scarce. A general overview and comparisons of attached and planktonic bacterial communities in freshwater and marine

22 II Bacterial biofilms on submerged macrophytes habitats is given in (Simon et al. 2002, Pernthaler & Amann 2005) and references therein.

Submerged macrophytes are, in addition to algae, the main primary producers in lakes; they structure the littoral zone and prevent resuspension of sediments, thus enabling clear water states (Scheffer et al. 1993). The freshwater macrophytes Myriophyllum spicatum and Chara globularis, and possibly also other Chara species, produce secondary compounds such as polyphenols and cyclic sulfur compounds, which exert allelopathic activity against algae and cyanobacteria (Anthoni et al. 1987, Nakai et al. 2000).

Antibacterial cyclic quaternary amines have been isolated from C. globularis

(Anthoni et al. 1987). Hydrolysable polyphenols of M. spicatum, especially tellimagrandin II, inhibit photosystem II of cyanobacteria (Leu et al. 2002).

Plant polyphenols may have antimicrobial activity, but some bacteria may also overcome polyphenol-based plant defences (Scalbert 1991).

Not only secondary metabolites but also nutrients possibly affect biofilm density and composition. Depending on their life cycle stage, macrophytes may release low to substantial amounts of macronutrients (Carignan & Kalff

1982), and at times high concentrations of micronutrients (Jackson et al. 1994).

Especially older plant parts may leak both organic compounds and inorganic nutrients (Sondergaard 1981). Nutrient conditions affect the impact of submerged macrophytes on bacterioplankton: Vallisneria americana has a positive impact on bacterioplankton density under high NH4+ conditions, but a neutral or negative impact when NH4+ is limiting (Huss & Wehr 2004).

Biofilms can be both beneficial and detrimental for submerged macrophytes.

On the positive side, epiphytic biofilms provide organic compounds and carbon dioxide to the macrophytes and enhance nutrient recycling (Wetzel

1993). Further, the biofilm bacteria Roseobacter gallaciencis and

Pseudoalteromonas tunicata that colonize the marine alga Ulva australis produce compounds against fouling organisms (Rao et al. 2006), and axenic Ulva linza

23 II Bacterial Biofilms on submerged macrophytes require bacteria to restore the typical growth form, and some bacteria even enhance the algal growth rate (Matsuo et al. 2005, Marshall et al. 2006).

Negative impacts on submerged macrophytes could arise from increased shading by thick biofilms and possibly also from pathogenic bacteria present in the biofilm. Macroalgae can also have negative effects on epiphytic bacteria.

For instance, bacterial colonization of the marine red algae Bonnemaisonia hamifera and Delisea pulchra is inhibited by algal-released secondary metabolites (Maximilien et al. 1998, Huber et al. 2003). These furanones also affect the swarming motility of Serratia liquefaciens (Rasmussen et al. 2000) and indirectly affect larval attachment (Dobretsov et al. 2007). Whether or not such chemical interactions between plants and bacteria are important for biofilm density and community composition on aquatic macrophytes is unknown. The only study addressing microbial diversity on M. spicatum showed that the biofilm was dominated by gammaproteobacteria and members of the CFB group (Chand et al. 1992). Bacterial epiphytes of C. aspera have not been described before.

Given that a strong interaction might exist between plants and their associated heterotrophic biofilm, especially in aquatic systems, we questioned whether different macrophytes (substrate, plant age) or the respective habitat determines bacterial community composition. We selected two common, allelochemically active, submerged macrophytes, Chara aspera and

Myriophyllum spicatum, sampled in freshwater (Lake Constance) and brackish water (Schaproder Bodden). We identified plant species, plant age, and habitat-specific differences and similarities of the bacterial density and community composition.

24 II Bacterial biofilms on submerged macrophytes

MATERIAL & METHODS

Plants. Brackish water samples of Chara aspera and Myriophyllum spicatum were collected on 24 October 2006 in the Schaproder Bodden, east of the Isle of

Hiddensee (N 54°27.4627’; E 13°07.5664’). Three plants each were collected by snorkelling in 0.7–1 m depth, stored in artificial brackish water (8‰, the same salinity as in the Bodden; (Blindow et al. 2003)) with 3.5% formaldehyde (final concentration). Freshwater plants were sampled at the southwest shore of the

Isle of Reichenau, Lake Constance, near a gravel ridge (N 47°42.247,

E 9°02.289). Three replicates each were collected on 6 November 2006 at a depth of 0.7–1.2 m for M. spicatum and 2.5–3 m for C. aspera. The plants were transported in separate sterile tubes to the laboratory, where they were fixed with 3.5% formaldehyde (final concentration). All samples were stored at 4 °C until processing started on 7 November 2006. The plant samples were divided into an upper section of the plants apices, approx. 5 cm long, and a lower section, approximately 5–10 cm of stem length above the sediment.

Biomass and chemical analyses. Myriophyllum spicatum was processed as part of our routine sampling campaign, in which plants are dissected into apices, upper and lower leaves, and stems. For C. aspera chemical analyses, we did not differentiate between upper and lower plant parts. Sub-samples of each plant part were incinerated for 6 h at 550 °C to determine the ash-free dry mass. We measured the carbon, nitrogen, and phosphorus content of all plant samples using standard methods (Choi et al. 2002). The total phenolic content of M. spicatum and C. aspera was determined using a modified Folin–Ciocalteau method (Box 1983). The concentration of non-phenolic compounds interfering with the Folin reagent are <5% in M. spicatum (Choi et al. 2002). Those in

C. aspera were determined using a modified polyvinylpyrrolidon method

(Gross et al. 1996). The major allelochemical of M. spicatum, tellimagrandin II, was quantified by reverse-phase HPLC (Müller et al. 2007). All measurements were based on dry mass since the inorganic incrustations of C. aspera also

25 II Bacterial Biofilms on submerged macrophytes provide settlement surfaces for bacteria. The antibacterial and allelopathically active compounds in Chara spp. (Scheffer et al. 1993, Fisher et al. 1998) are difficult to isolate and were not determined here.

Detachment of biofilm. Plant parts were transferred into sterile 50 ml polypropylene tubes containing 50 ml of formaldehyde (3.7% final concentration) and sodium pyrophosphate (0.1 M Na4P2O2×10 H2O, NaPPi).

The biofilm was detached by ultrasonication for 60 s (Laboson 200 ultrasonic bath, Bender & Hobein), followed by 15 min of vigorous shaking (18.3 Hz,

Thermomixer Eppendorf) and again 60 s of ultrasonication. Two millilitres of the detached biofilm were filtered onto white polycarbonate filters (0.2 µm, Ø

25 mm Nucleopore) and stored at –20 °C.

We optimized the detachment procedure prior to this experiment. NaPPi was a suitable detergent to detach bacteria from macrophyte leaves as shown by a previous study in our group (Müller et al. 2007). We further varied the sonication time and shaking duration to obtain the best results for a gentle but effective detachment of the biofilm (Buesing & Gessner 2002, Bockelmann et al.

2003). Detachment with an ultrasonic probe (Bandelin electronic GM 70 HD, 20 kHz, 57W) resulted in 0.13 ± 0.03×106 cells cm-2 but the plant tissue was severely damaged and numerous bacterial cells were still attached to the leaf surface as observed by microscopic examination. We then tried are more gentle detachment with shorter sonication times in an ultrasonic bath and constant, gentle shaking afterwards, rather than permanent ultrasonication. This method yielded 1.9 ± 0.6×106 cells cm-2 and the plant tissue was not visibly damaged except at the cut surface on the petiole. A thorough microscopy of the leaves proved hardly any attached bacterial cells left.

Fluorescence in situ hybridization (FISH). FISH was conducted following a protocol by Pernthaler et al. (Pernthaler et al. 2001) consisting of a hybridization step at 46 °C for 3 h and a washing step for 15 min at 48 °C.

Filters were counterstained with DAPI (4ʹ,6-diamidino-2-phenylindol,

26 II Bacterial biofilms on submerged macrophytes

1 µg ml-1, 5 min). Stained cells were counted under an epifluorescence microscope (Labophot 2, Nikon) at an excitation wavelength of 549 nm. Probes used are listed in Table 2.1 and further details are available at probeBase (Loy et al. 2003).

Table 2.1. Oligonucleotide probes used in this study. % Probe a) Sequence Target group Reference FA EUB338 GCTGCCTCCCGTAGGAGT 35 Most bacteria (Amann et al. 1990)

NON338 ACTCCTACGGGAGGCAGC 35 Competitor of EUB (Wallner et al. 1993)

ALF968 GGTAAGGTTCTGCGCGT 20 Alphaproteobacteria (Neef 1997)

BET42a b) GCCTTCCCACTTCGTTT 35 Betaproteobacteria (Manz et al. 1992)

GAM42ab) GCCTTCCCACATCGTTT 35 Gammaproteobacteria (Manz et al. 1992) b) PLA886 b) GCCTTGCGACCATACTCCC 35 Planctomycetes (Neef et al. 1998)

HGC96a TATAGTTACCACCGCCGT 25 Actinomycetes (Roller et al. 1994)

CF319a TGGTCCGTGTCTCAGTAC 35 Bacteroidetes (Manz et al. 1996) a) Probes were labelled with Cy 3 b) For these probes, a competitor probe was used; FA: Formamide

Statistical analyses. Data of FISH analysis were arcsin transformed. For planctomycetes, data were additionally x¼ transformed to ensure equal variances. Plant species-, plant part-, and habitat-specific differences were analysed by 3-way ANOVAs (Sigma STAT 3.0). Non-metric dimensional scaling plots were generated with square-root transformation of data and Bray-Curtis similarity (Primer 5.0). For correlations, the Pearson correlation was used (Sigma STAT 3.0).

27 II Bacterial Biofilms on submerged macrophytes

RESULTS

The two plant species, each from two different habitats, exhibited distinct morphological and chemical characteristics. The organic content of the plants of each species and from each habitat differed, but the upper or lower parts of each plant sampled did not differ in organic content (Figure 2.1; 3-way

ANOVA, Table 2.2). Chara aspera had a lower organic content (40.9% ± 14.4, mean ± SD, n = 12) than Myriophyllum spicatum (51.7% ± 11.6, mean ± SD, n = 12). 80 Figure 2.1 Proportion of organic dry mass in plant samples collected at all sites. SB: Schaproder Bodden, LC: Lake 60 Constance, C: Chara aspera, M: Myriophyllum spicatum; upper and lower indicate plant parts analysed; n = 3; error bars indicate 40 SE

% organic dry mass % organic 20

0

r r

B M uppe LC C lower C LC SB C lower C SB LC M lower SB M lower SB LC C upper SB C upper LC M uppe S

Freshwater plants had a much lower organic content (35.1% ± 9.4, mean ± SD, n = 12) than brackish water plants (57.5% ± 6.6, mean ± SD, n = 12).

Only a marginal interaction of plant × habitat was found (p = 0.079), owing to a larger difference in organic content of C. aspera from the two sites than that of

M. spicatum. The significant interaction term between habitat and plant part is due to the observed differences of plant parts in Lake Constance; the organic content of the plant parts did not differ in plants from Schaproder Bodden.

28 II Bacterial biofilms on submerged macrophytes

Myriophyllum spicatum contained more phenolic compounds than C. aspera

[97–173 mg (g dry mass)–1 vs. <1 mg (g dry mass)–1; Table 2.3] and M. spicatum from Lake Constance had a slightly higher polyphenol content than

M. spicatum from Schaproder Bodden [apices: 173 ± 21 mg (g dry mass)–1 and

120 ± 33 mg (g dry mass)–1, respectively; Student’s t-test: P = 0.02]. Also the anthocyanin content was much higher in M. spicatum than in C. aspera. In both habitats, the anthocyanin content of C. aspera was <0.1 mg (g dry mass)–1; the anthocyanin content of M. spicatum from Schaproder Bodden was slightly lower than that of M. spicatum from Lake Constance [approx. 0.5 mg (g dry mass)–1 vs. 1.0 mg (g dry mass)–1; Student’s t–test: P = 0.005, Table 2.2]. The chlorophyll a and b contents were highest in the apical shoots and upper leaves of M. spicatum from Lake Constance (Table 2.3).

Table 2.2. Statistical analysis. 3-way ANOVA for selected parameters. Data for the CFB were arcsin transformed; for planctomycetes, x1/4 transformation was used.

% Plant Total Ash-free organic bacterial cell Planctomycetes CFB dry mass matter counts Source of varia- DF F P F P F P F P F P tion Habitat 1 101.63 <0.001 13.721 0.002 25.963 <0.001 30.970 <0.001 0.467 0.504 Plant 1 24.481 <0.001 3.746 0.071 1.944 0.182 26.623 <0.001 45.454 <0.001 Plant part 1 0.0563 0.815 0.183 0.674 21.229 <0.001 3.705 0.072 21.018 <0.001 (PP) Habitat 1 3.510 0.079 0.0307 0.863 2.606 0.126 10.618 0.005 0.538 0.474 × Plant Habitat 1 5.249 0.036 2.253 0.153 10.499 0.005 0.484 0.497 4.901 0.042 × PP Plant × 1 1.087 0.313 0.0479 0.830 0.0246 0.877 0.0998 0.756 7.105 0.017 PP Habitat × Plant 1 0.505 0.488 0.121 0.732 < .001 0.995 1.179 0.294 14.113 0.002 × PP

29 II Bacterial Biofilms on submerged macrophytes

The carbon content (Table 2.3) of C. aspera was about half of that of

M. spicatum, possibly in part owing to the overall lower organic dry mass of the former. Chara aspera also contained less nitrogen and phosphorus per g dry mass than M. spicatum when whole plants were considered. The C/N molar ratio ranged from about 15 in apices of M. spicatum from Lake Constance to 31 in stems of M. spicatum from Schaproder Bodden. The C/P molar ratio ranged from 436 in apices of M. spicatum to more than 1373 in C. aspera from Lake

Constance.

Table 2.3. Chemical parameters measured in plants. LC: Lake Constance, SB: Schaproder Bodden, n = 3, mean ± SD Total phenolic content Anthocyanin Chlorophyll a and b [mg (g dry mass)–1] [mg (g dry mass)–1] [mg (g dry mass)-–1] LC SB LC SB LC SB C. aspera 0.9 ± 0.08 0.7 ± 0.12 0.05 ± 0 0.06 ± 0 1.9 ± 0.17 1.2 ± 0.02 M. spicatum Apex 173 ± 21 120 ± 33 0.9 ± 0.02 0.52 ± 0.08 6.8 ± 1.3 2.3 ± 0.26 Upper leaves 120 ± 29 97 ± 5 0.8 ± 0.2 0.50 ± 0.01 8.4 ± 2.3 2.3 ± 0.28 Upper stem 133 ± 13 100 ± 9 1.4 ± 0.1 0.81 ± 0.15 1.8 ± 0.6 1.0 ± 0.04

C N P

[mg (g dry mass)–1] [mg (g dry mass)–1] [mg (g dry mass)–1] LC SB LC SB LC SB C. aspera 206 ± 4 179 ± 15 11 ± 1.1 14 ± 0.12 0.4 ± 0.1 0.68 ± 0.06 M. spicatum Apex 425 ± 20 357 ± 35 36 ± 9 24 ± 8 2.6 ± 1.2 2.4 ± 1.5 Upper leaves 384 ± 37 363 ± 5 27 ± 6 16 ± 3 1.2 1.3± 0.33 Upper stem 400 ± 7 379 ± 25 15 ± 3 14 ± 3 0.9 ± 0.2 ––

We determined the bacterial abundance based on plant dry mass since there are no reliable surface area-to-biomass ratios for M. spicatum and C. aspera from the two habitats. The bacterial abundance in the two habitats and on the different plant parts differed significantly, but did not differ significantly between the two plant species (Figure 2.2, Table 2.2). In general, we found a higher bacterial abundance on plants from Schaproder Bodden

[5.1×107 ± 3.9×107 cells (g dry mass)–1; mean ± 1 SD] than on plants from Lake

Constance [1.3×107 ± 0.7×107 cells (g dry mass)–1]. The lower plants parts from

30 II Bacterial biofilms on submerged macrophytes

Schaproder Bodden had higher bacterial cell counts than the upper plant parts, while cell counts on lower plant parts from Lake Constance were only marginally higher than the counts on upper plant parts (Figure 2.2), resulting in a significant habitat × plant part interaction (Table 2.2, P = 0.005). The ash- free dry mass differed significantly between habitats, and the organic content of the plant samples differed significantly between habitats and plant species but not between plant parts (Table 2.2). The general pattern of bacterial abundance remained when calculated on an organic dry matter basis

(Figure 2.2). 25 7 x 10

-1 20

15

10

5

Total bacterial cell counts (g dm) cell bacterial Total 0 LC C lower SB C lower LC M lower SB M lower LC C upper SB C upper LC M upper SB M upper Figure 2.2.Total bacterial cell counts determined by DAPI staining. Black bars: counts (g dry mass)–1; grey bars: counts (g ash-free dry mass)–1. SB: Schaproder Bodden; LC: Lake Constance; M: M. spicatum; C: C. aspera. n = 3; error bars indicate SE.

31 II Bacterial Biofilms on submerged macrophytes

The composition of the bacterial biofilm on the two plant species was

similar except for the abundance of members of the CFB group and

planctomycetes (Figure 2.3). On both plant species in both habitats, bacteria of

the CFB group were the most abundant bacterial group and reached up to 35%

A) B) 30 30 Figure 2.3. Biofilm composition in Lake Constance (left) and Schaproder Bodden 20 20 (right). A and B, Myriophyllum spicatum upper 10 10 section; C and D, Chara aspera upper section; 0 0 E and F, M. spicatum lower section; C) D) 30 30 G and H, Chara aspera lower section. n = 3; errors bars indicate SD. 20 20 ALF: alphaproteobacteria; BET: betaproteobacteria; 10 10 GAM: gammaproteobacteria; PLA: planctomycetes; 0 0 E) F) HGC: actinomycetes; 30 30 CFB: Cytophaga–Flavobacteria–

% of DAPI counts Bacteroidetes 20 20

10 10

0 0 G) H) 30 30

20 20

10 10

0 0 ALF ALF PLA PLA BET BET CFB CFB HGC HGC GAM GAM

of the total cell counts. The CFB counts correlated positively with all

measured chemical parameters (Pearson correlation: carbon: r = 0.637,

P = 0.0008; nitrogen: r = 0.666, P = 0.0003; phosphorus: r = 0.755, P < 0.0001;

chlorophyll: r = 0.433, P = 0.0344; total phenolic compounds: r = 0.685,

P = 0.0002). The number of CFB cells was generally higher on M. spicatum than

on C. aspera and higher on upper parts of both plant species. The differences

were not uniform and resulted in significant interaction terms (Figure 2.3;

Table 2.2), which indicated specific habitat, plant, and plant part patterns. The

32 II Bacterial biofilms on submerged macrophytes second major group of bacteria in the biofilms were alphaproteobacteria, which accounted for 3–17% of the DAPI counts. The abundance of alphaproteobacteria did not differ between plant species and habitats (3-way

ANOVA, df = 1, F = 4.1, P = 0.05). Beta- and gammaproteobacteria abundance was similar on both plant species and in both habitats (3-way ANOVA, df = 1,

F = 1.257, P = 0.279; df = 1, F = 1.982, P = 0.178). Actinomycetes were the least- abundant group, and their abundance did not differ between plant species

(0.7–2.0% of DAPI counts¸ 3-way ANOVA, df = 1, F = 1.179, P = 0.294).

Stress: 0.12

Figure 2.4. Non-metric dimensional scaling plot of the bacterial community composition on all plant samples. Grey triangles: Myriophyllum spicatum, white triangles: Chara aspera. Striped triangles: samples from Schaproder Bodden; non-striped triangles: samples from Lake Constance. Upper and lower plant parts are denoted by triangles pointing upwards and downwards, respectively. Data are x1/4 transformed.

Interestingly, the proportion of planctomycetes differed between habitat and plant species. In Lake Constance, almost no planctomycetes were detected on M. spicatum, but they made up 2–3% of all cell counts on C. aspera. In

Schaproder Bodden, planctomycetes were found on both plant species, with slightly higher numbers on the upper plant parts (2–6% of DAPI counts) than

33 II Bacterial Biofilms on submerged macrophytes on the lower plant parts, but there were no differences between the plant species (Figure 2.3, Table 2.2).

We found negative correlations of this group with carbon (Pearson correlation; r = –0.507, P = 0.0114), nitrogen (r = –0.433, P = 0.0343), chlorophyll

(r = –0.648, P = 0.0006), and total phenolic content (r = –0.449, P = 0.0278).

Overall, the bacterial community composition on M. spicatum in Lake

Constance differed from that on C. aspera in both habitats and even from M. spicatum in Schaproder Bodden (Figure 2.4).

DISCUSSION

To our knowledge, this is the first study comparing bacterial biofilms on two macrophytes in brackish and freshwater habitats. Our data support the findings of other studies of biofilms on aquatic organisms, especially diatoms and cyanobacteria, where CFB and alphaproteobacteria make up major parts of the biofilm (Eiler & Bertilsson 2004, Grossart et al. 2005).

The total bacterial cell counts on the two plant species revealed that habitat and plant part seem to be more important for epiphyte bacterial abundance than the plant species. Although surface area-to-dry mass ratios have been determined in other studies, e.g., Myriophyllum spicatum 1205 cm2 (g dry mass)–1 and Nitellopsis obtusa (starry stonewort) 560 cm2 (g dry mass)–1 (Sher-Kaul et al.

1995), we decided not to calculate bacterial density based on plant surface area because our computer-based image analysis of M. spicatum leaf area showed that the calculation of surface area based on dry mass cannot be averaged over the whole plant. The surface area-to-dry mass ratio was 3500 cm2 (g dry mass)–1 for freshwater M. spicatum apices and 1600 cm2 (g dry mass)–1 for the lower parts of the same plant. Such a difference would amplify our findings that lower shoots harbour a higher abundance of bacteria. Freshwater Chara spp. had a surface area-to-biomass ratio of only 122 cm2 (g dry mass)–1, which would yield even higher bacterial densities on this plant. In general, bacterial

34 II Bacterial biofilms on submerged macrophytes counts were highest on lower leaves close to the sediment (Figure 2). This seems reasonable since biofilm on older leaves should be thicker, thus containing more cells. Older leaves also contain less allelopathic compounds and are leakier than younger leaves, which might influence the nutrient availability. The nutrient content of the water column could be higher close to the sediment; this could also have an impact, but was not assessed in this study. Differences between the total bacterial cell counts on plant species from the different habitats might also be a consequence of pH, temperature, salinity and water retention time, which have been also found to influence community composition (Lindström et al. 2005, Lozupone & Knight 2007).

Alpha-, beta-, and gammaproteobacteria were present on both macrophytes in similar abundance, with gammaproteobacteria having the lowest counts of the proteobacteria. The least-abundant group was the actinomycetes (0.7–2.0% of DAPI counts). Not all members of this group might have been detected with the FISH probes because of the generally thicker cell walls of gram-positive bacteria. Our EUB probe, for example, detected only 50–80% of all DAPI cell counts (data not shown). The coverage of all bacteria together could probably have been higher if a combination of three different EUB probes were used

(Daims et al. 1999), but since the planctomycetes, which are often missed by the single EUB probe used, did not make up a major amount of the biofilm, our results would not have changed dramatically. The total counts of all group-specific probes did not account for all eubacterial counts. We therefore assume that we did not detect all bacterial groups present in the biofilm of the two plant species, but we did use probes for the most common groups in biofilms and aquatic systems.

Alpha- and betaproteobacteria are the most abundant bacteria in lake snow aggregates in Lake Constance, and CFB are only found in hypolimnic particles, where they are considered to degrade refractory compounds such as chitin and cellulose (Schweitzer et al. 2001). Betaproteobacteria in the polluted river

35 II Bacterial Biofilms on submerged macrophytes

Spittelwasser dominated biofilms formed on glass slides throughout the year, followed by alphaproteobacteria, with seasonal maxima of CFB and planctomycetes, but gammaproteobacteria were never abundant (Brümmer et al. 2004). Comparably, alphaproteobacteria, followed by CFB, were dominant in biofilms on stainless steel and polycarbonate exposed in Delaware Bay, and betaproteobacteria were almost absent (Jones et al. 2007). In our study, we saw a comparable picture, with CFB mostly dominating the biofilm on macrophytes, followed by alpha- and betaproteobacteria. Distinct differences for habitat and plant species were found for the members of the CFB group and planctomycetes. Especially the abundance of planctomycetes differed between plant species and between the plant parts. Planctomycetes are found in a wide variety of habitats and are known to colonize surfaces (Brümmer et al. 2000). Earlier studies on lake snow in Lake Constance did not look for this bacterial group, and a comparison of lakes and oceans found only low numbers in freshwater and hardly any in the marine bacterioplankton

(Glöckner et al. 1999, Schweitzer et al. 2001). The authors of the latter study argue that this might be due to low abundance and that FISH was at the range of its detection limit. This could also be the case in our study since we only found low abundances. Based on our data and the meagre knowledge about planctomycetes ecology, we propose that either nutrient content or plant age

(senescence) might account for differences in the abundance of planctomycetes because of the strong negative correlations with carbon, nitrogen, chlorophyll, and total phenolic content. This is supported by a study of marine planctomycetes, which were affected by organic compounds (Tadonléké 2007), and the observation that M. spicatum excretes substantial amounts of organic compounds (Godmaire & Planas 1986). Planctomycetes occur in many different habitats, yet their ecology is unexplored since only a few species have been cultivated (Wang et al. 2002).

36 II Bacterial biofilms on submerged macrophytes

The dominance of the CFB group in all our samples is not unusual, but is nevertheless interesting because we found distinct differences between plant species and plant parts. The CFB counts correlated positively with plant carbon and nutrient content as well as with chlorophyll and phenolic compounds. Members of the CFB group have often been described as major components of biofilms and are known to degrade rather complex molecules that occur in the high molecular mass fraction of dissolved organic matter

(Kirchman 2002). This is important for other bacterial groups that are not capable of degrading such molecules but thrive on the degradation products

(Wetzel 1993). Considering that alphaproteobacteria are more likely to degrade labile organic matter (Cottrell & Kirchman 2000, Schweitzer et al. 2001), this group could depend on degradation products of CFB or betaproteobacteria.

The correlation of alphaproteobacteria with CFB (r = 0.444, P = 0.03) might be evidence for this.

We found that habitat had a distinct influence on both planctomycetes and members of the CFB group. The Schaproder Bodden is a shallow coastal area of the Baltic Sea and has a higher salinity than Lake Constance. Salinity is in fact a major environmental determinant of microbial community composition

(Lozupone & Knight 2007). Whether or not the trophic state of the habitat influences biofilm density and composition remains open. Epiphytic algae are influenced by the trophic state, especially nitrogen availability, but only indirectly or not at all by host species (Phillips et al. 1978). Under eutrophic conditions, epiphytes should receive more organic and inorganic resources from the surrounding water and should be less dependent on plant-exuded compounds. It is unlikely that plant nutrient content influences the algal biofilm since plants relocate only small amounts of macronutrients (Carignan

& Kalff 1982). This, however, does not exclude the possibility that the epiphytic algal composition might influence the composition of heterotrophic bacteria.

37 II Bacterial Biofilms on submerged macrophytes

Only M. spicatum from Lake Constance exhibited a distinct bacterial biofilm community compared to M. spicatum from Schaproder Bodden and C. aspera from both habitats (Figure 4). Perhaps the phenolic content of the plant species in the different habitats is responsible for this effect. While C. aspera contains almost no phenolic compounds, M. spicatum has high concentrations of total phenolic compounds and anthocyanin, especially in the samples from Lake

Constance.

We propose that the bacterial community composition is rather determined by the presence or (near) absence of phenolic compounds and not by their concentrations, since the concentrations between the upper and lower shoots in both habitats are rather similar. The association of polyphenol-degrading bacteria with M. spicatum (Müller et al. 2007) might be evidence for this. A direct proof of the impact of phenolic compounds on biofilm composition is difficult to achieve, and complex molecules, especially tannins, can be a difficult substrate for some bacteria (Scalbert 1991) and may have caused this pattern. We also cannot rule out other factors such as salinity, pH, temperature, and dissolved organic carbon to explain these differences

(Allgaier & Grossart 2006). Dissolved organic matter produced by plants and epiphytic algae is usually subjected to photolysis (Wetzel et al. 1995) but can also be degraded by bacteria capable of degrading high molecular weight compounds (Kirchman 2002). The resulting degradation products as well as carbon dioxide and oxygen recycling can be of mutual benefit for both primary producers and heterotrophic bacteria. Biofilm bacteria are also known to produce compounds that can influence phototrophs beneficially or detrimentally (Rao et al. 2006). Both plant species contain allelochemicals known to inhibit bacterial or algal growth (Wium-Andersen et al. 1982, Gross et al. 1996).

We present one of the first studies investigating heterotrophic bacterial communities on aquatic plants. To elucidate further which bacterial groups are

38 II Bacterial biofilms on submerged macrophytes active and contribute metabolically to processes within the biofilm, further approaches such as MICRO–FISH could be applied. To gain insight into the bacterial groups involved more specific probes for proteobacteria and CFB should be used. Also archaea could be of interest since they play a major role in the root region of various aquatic plants. Our data suggest an apparent impact of plant species, plant age and habitat on epiphytic bacterial communities.

Acknowledgements This work was supported by the German Science Foundation with grant CRC454, project A2 to EMG. Claudia Feldbaum is acknowledged for technical assistance. We thank M. Mörtl for providing the statistical program PRIMER 5. Karen Brune improved the English, and two anonymous reviewers made valuable comments on the manuscript.

39 II Bacterial Biofilms on submerged macrophytes

40

Chapter III

Bacterial community composition of biofilms on two submerged macrophytes and an artificial surface in Lake Constance

Melanie Hempel, Hans–Peter Grossart & Elisabeth M. Gross

III Bacterial biofilms on aquatic surfaces

42 III Bacterial biofilms on aquatic surfaces

ABSTRACT: Submerged macrophytes provide vast surfaces for the settling of microorganisms. Those biofilms contribute substantially to the overall nutrient cycling in the littoral zones and interact in multiple ways with macrophytes and other epiphytes. Although studies on aquatic biofilms are numerous, aquatic macrophytes and their heterotrophic biofilms have rather been neglected. During two seasons, we investigated the bacterial community composition (BCC) on two submerged macrophytes with DGGE and with a clone library. We followed the spatial BCC on Myriophyllum spicatum in Lower Lake Constance in 2005 and found differences between the apices and the leaves. These differences lessened towards autumn. In 2006, we investigated the substrate–specific BCC on M. spicatum, Potamogeton perfoliatus and polypropylene sheets. The comparison between all substrates also exhibited a distinct BCC on M. spicatum apices. On the leaves of both plants, the BCC was rather similar, also compared to the artificial substrate. A comparison between the BCC on mesocosm M. spicatum and artificial substrates gave similar results as in the field. Bacterial sequences from excised DGGE bands and the clone library were mainly affiliated with yet uncultured clones originating from various freshwater habitats. We also found bacteria capable of degrading phenolic and aromatic compounds. Our results indicate that the BCC on M. spicatum apices is rather unique and may point to specific bacterial functions in this microenvironment. To better elucidate these functions, further studies are needed, preferentially with cultivation–based approaches, which allow for more detailed physiological studies of biofilm bacteria.

Keywords: Myriophyllum spicatum, Potamogeton perfoliatus, DGGE, phenolic compounds, clone library.

43 III Bacterial biofilms on aquatic surfaces

INTRODUCTION

Several studies have shown that plants are no neutral substrates for epiphytic algae and cyanobacteria, and that some plants might influence the density and composition of their autotrophic biofilm, while a dense biofilm formation may severely hamper the growth of submerged macrophytes (Phillips et al. 1978,

Eminson & Moss 1980, Blindow 1987). Comparable studies for the heterotrophic biofilm on freshwater macrophytes are scarce. We were interested whether substrate–specific differences exist in the heterotrophic bacterial community composition (BCC) between artificial and natural substrates. We chose the submerged macrophytes Myriophyllum spicatum L. (Haloragaceae) and Potamogeton perfoliatus L. (Potamogetonaceae) as the two natural substrates. Both are common in lakes of different trophic state, also in the oligo– to mesotrophic Lower Lake

Constance, Germany. Myriophyllum spicatum is known for its allelopathic activity against algae and cyanobacteria caused by hydrolysable polyphenols (Gross et al.

1996, Leu et al. 2002). Many species of the Potamogetonaceae, among them

P. perfoliatus, contain antibacterial compounds, although none has been identified so far for P. perfoliatus (Bushmann & Ailstock 2006).

Both plant species differ in their leaf structure, growth form, life–cycle and chemical composition. Myriophyllum spicatum has pinnate leaves with individual filaments of about 1 mm diameter. In Lake Constance, it forms small and distinct stands of 10 – 15 shoots with a length of 30 – 100 cm. Potamogeton perfoliatus has laminar, oval leaves and grows in stands of several square meters up to the surface with shoot length up to 4 m, resembling ‘underwater forests’. In Lake Constance,

M. spicatum starts growing in late June and declines not before November.

Potamogeton perfoliatus emerges already in May and is senescing in September to

October. Our long–term analysis shows that the chemical composition of M. spicatum in Lake Constance exhibits seasonal patterns and a gradient of macro– and micronutrients and phenolic compounds from the apices to the older leaves, while

P. perfoliatus contains very little phenolic compounds (Choi et al. 2002), and in

44 III Bacterial biofilms on aquatic surfaces

general exhibits no pronounced gradients in macronutrients (Gross et al., unpublished; this study).

Submerged macrophytes do not only provide shelter and nutrition for other organisms, they also structure the littoral zone, prevent sediments from re– suspension and may change the nutrient composition of the water column (Stoner

1980, Jeppesen et al. 1998). They provide a vast surface for the settlement of microorganisms, algae and meio– and macrofauna. Bacteria in biofilms may have beneficial effects for the plants: they can attract zoospores, enhance and restore growth or growth form of plantlets and produce metabolites beneficial for the plant

(Joint et al. 2000, Marshall et al. 2006). Secondary metabolites of bacteria associated with terrestrial plants may reduce herbivory (Hoagland 2001, Sturz & Kimpinski

2004). Yet, dense layers of autotrophic epiphytes decrease the light and nutrient availability of submerged macrophytes (Sand-Jensen 1990). Further, sedimentation and particle retention are a source of nutrients both for macrophytes and epiphytes.

Thus an optimum between nutrient supply and light availability for photosynthesis has to be kept (Schulz et al. 2003).

However, also the plant may influence its epiphytes. Secondary metabolites of red algae interfering with quorum sensing may inhibit bacterial attachment (Maximilien et al. 1998, Nylund et al. 2005). Plant polyphenols similar to those present in

M. spicatum apparently hamper the quorum sensing regulated biofilm formation

(Huber et al. 2003). The antimicrobial properties of polyphenols are due to their ability to chelate proteins, nutrients and iron (Scalbert 1991). These allelochemicals also inhibit gut microbiota of herbivores feeding on M. spicatum (Walenciak et al.

2002). If polyphenols would interfere with epiphytic bacteria, this could lead to a reduced biofilm formation and in turn to an increased light and nutrient availability for the plant.

So far, only a few studies dealt with the heterotrophic biofilm composition on freshwater macrophytes. They often focus on marine and brackish species or are restricted to certain bacterial functions such as nitrification and denitrification

45 III Bacterial biofilms on aquatic surfaces

(Eriksson & Weisner 1999). On the chlorophytes Desmidium grevillii, Hyalotheca dissiliens and Spondylosium pulchrum bacteria belonging to the CFB phylum and alpha–, beta– and gammaproteobacteria were found (Fisher et al. 1998), and the obtained sequences were not closely related to known taxa. Bacteria belonging to the

Flavobacteria–Sphingobacteria group were most abundant on marine diatoms (Grossart et al. 2005). Bacterial biofilm community composition and the reaction of individual taxa were selectively influenced by the type of dissolved organic matter provided

(Olapade & Leff 2006). Submerged plants as producers of dissolved organic matter might thus influence bacterial community composition. Cultivation–based assays detected gammaproteobacteria and Cytophaga–Flavobacteria in the biofilm on

M. spicatum (Chand et al. 1992). The biofilm community composition on Potamogeton crispus in comparison to cellulose filters yielded variable numbers of alpha–, beta– and gammaproteobacteria (Hong et al. 1999). Molecular studies investigating aquatic surfaces have mostly been on artificial substrates (Brummer et al. 2000, Olapade &

Leff 2006), algae (Rao et al. 2006), among them diatoms (Grossart et al. 2005). Others investigated the heterotrophic biofilm on macrophytes with cultivation dependent techniques (Chand et al. 1992), which can be rather selective depending on the growth conditions applied. In the last decade, our knowledge on microbial communities increased continuously, mainly due to the establishment of new molecular methods such as FISH (fluorescence in situ hybridisation) and DGGE

(denaturing gradient gel electrophoresis). These methods allow the investigation of bacterial communities without cultivation (Head et al. 1998).

By using DGGE and clone library, we investigated whether the biofilm BCC differed between different parts of the same plant, different macrophytes and/or substrates. We expected differences between the apices and the lower leaves on

M. spicatum reflecting the unique chemical gradient of this plant (Hempel & Gross, submitted, Gross et al., unpublished). Further, we propose that the biofilm BCC on the polyphenol–rich M. spicatum is distinct from the polyphenol–free P. perfoliatus and an artificial substrate. Plant quality differences between plants and upper and

46 III Bacterial biofilms on aquatic surfaces

lower leaves was analysed by measuring carbon, nitrogen, phosphorus and chlorophyll content as well as total phenolic content, anthocyanins and, only in

M. spicatum, the hydrolysable polyphenol tellimagrandin II. In 2005 we investigated the spatial differences in biofilm BCC between younger and older plant parts of

M. spicatum. In summer 2006, we extended our study to the comparison of the BCC on three different substrates: M. spicatum, P. perfoliatus and polypropylene sheets, with extensive plant chemistry. To get a more detailed community composition we constructed a clone library of the biofilm on the apices of M. spicatum.

MATERIALS AND METHODS

Sampling site. We sampled submerged macrophytes during the growing season in 2005 and 2006 near the Isle of Reichenau in Lower Lake Constance, Germany

(N47°42.247, E9°02.289). In July, August and October 2005, we investigated whether the biofilm community composition on Myriophyllum spicatum differed between younger and older plant parts. Triplicates of three different plant stands were analyzed. In 2006, we investigated the BCC on the two submerged macrophytes

M. spicatum and Potamogeton perfoliatus and polypropylene sheets as an artificial substrate. The artificial substrates were deployed in 2.6 m depth two weeks before the sampling campaign started on 17 July. They were made out of 0.3 mm thick polypropylene sheets cut to the size of 9.7×1.2 cm and punctuated at each end with a hole–puncher. A floater was tied to one end for an upright position; the other was fixed with a lace to a plastic bar fixed to the ground by tent pegs. Myriophyllum spicatum, Potamogeton perfoliatus and artificial substrates were sampled in triplicates by snorkelling in a depth of 1.5 – 2 m every two weeks between 17 July and

09 October 2006. Plants and artificial substrate samples were stored individually in sterile 50 ml polyethylene tubes at 4 °C until processing started at latest 24 hours later. Additionally, we deployed artificial substrates in an outdoor mesocosm densely stocked with M. spicatum. Samples of the mesocosm plants (apex and leaf)

47 III Bacterial biofilms on aquatic surfaces and artificial substrates were taken on the same dates as in the field. Only samples from September and October were analysed.

For chemical analysis, plants were stored in plastic bags, three replicates consisting of at least five plants from one stand. Samples were stored at 4 °C until analysis the next day. At each sampling date, temperature, oxygen, conductivity and pH were measured in the water column 20 cm below water surface.

Detachment of epiphytic biofilm. In the laboratory, the plant length was measured and the overall state of the plant was recorded. Artificial substrates were documented by photography and approximately 2 cm of the middle part were used for biofilm analysis. The apex and 13 leaves of the 11 – 25 cm shoot section (lower leaf) of both plant species were taken and transferred to 15 ml sodium pyrophosphate (0.1 M Na4P2O7×10 H2O). Since the leaf surface of Potamogeton perfoliatus was much larger, only five leaves were sampled. The biofilm of all samples was detached by 1 min ultrasonication (Laboson 200 ultrasonic bath, Bender &

Hobein, Germany), 15 min of shaking (18.3 Hz, horizontal shaker Eppendorf,

Germany) and subsequent ultrasonication for 1 min. The detachment of epiphytic bacteria had been optimized before as described in (Hempel et al. 2008). The suspension containing the detached biofilm was filtered onto ME 24 membrane filters (0.2 µm; Ø 45 mm, Schleicher & Schuell) for DNA extraction, and stored at –20

°C until use.

DNA extraction. Filters were cut into small pieces and DNA was extracted following a standard phenol/chloroform protocol with an additionally lysozyme step

(8 mg ml–1; 260 µl sample–1; 30 min at 65 °C; (Walenciak 2004)). Extracted DNA was dried, taken up in 40 µl of filter sterilized Millipore water and quantified.

Polymerase chain reaction (PCR). PCR was performed in a Thermocycler T–

Gradient (Biometra, Germany). We used the primers 341f [5`–CCT ACG GGA GGC

AGC AG–3`(Muyzer et al. 1993)] and 907r [5`–CCG TCA ATT CMT TTG AGT TT–

3`(Lane et al. 1985)]. To perform DGGE, primer 341f was supplemented with a GC– clamp [5`–CGC CCG CCG CGC CCC GCG CCC GTC CCG CCG CCC CCG CCC–3´

48 III Bacterial biofilms on aquatic surfaces

(Muyzer et al. 1995)]. One 50 µl PCR reaction contained 5 µl PCR–buffer (10×Taq– buffer, Eppendorf, Germany), 5 µl dNTP–Mix (500 mM, Eppendorf, Germany), 0.5 µl of the forward primer at 25 pmol µl–1, 0.5 µl of the reverse primer at 25 pmol µl–1, 3 µl

MgCl2 (25 mM, Eppendorf, Germany), 10 µl 6 mg ml–1 BSA (Sigma) and 1 U Taq polymerase (0.2 µl, Eppendorf, Germany). The following protocol was used for amplification: (1) 5 min 95 °C; (2) 1 min 95 °C; (3) 1 min 55 °C; (4) 2 min 72 °C; repeat

(2) – (4) 29×; (5) 15 min 72 °C. We did not retrieve PCR products from all replicates, probably due to the high polyphenol content in M. spicatum plants resulting in variable replicate numbers.

Denaturing gradient gel electrophoresis (DGGE). DGGE was performed in an

INGENY PhorU system (Ingeny). For better comparison of DGGE banding patterns, equal amounts of PCR products (ca 50 ng) were loaded to the gel and an external standard was used. DGGE was performed in a 7% (v/v) polyacrylamide gel with a denaturing gradient of 40 to 70% urea and formamide, and run at 60 °C for 20 hours.

Gels were stained with 1× SybrGold (Invitrogen), washed in deionised water and documented with an AlphaImager 2200 Transilluminator (Biozym) under UV light.

Bands were excised from the gel with a sterile scalpel and immediately transferred to a sterile PCR cup, in which DNA was eluted with sterile water. DNA was amplified using the primer pair 341f – 907r (without GC–clamp) and conditions as described above. Sequencing was performed by 4base lab, Reutlingen, Germany. DGGE gels were analyzed with the software GelCompar II version 3.5 (Applied Maths,

Belgium). Cluster analysis was performed with Pearson correlation using the unweighted pair group method with arithmetic mean (UPGMA).

Clone library. In May 2006, apices of M. spicatum grown in an outdoor mesocosm

(a concrete basin of 2×2×1 m) were used to analyze the BCC with a clone library. The mesocosm had a flow–through with Lake Constance water of approx. 20 l h–1, and the plants were subjected to the same climatic conditions as in the field except that the water level was kept constant. Biofilm detachment, DNA extraction and PCR conditions were the same as described above, except that PCR was performed with

49 III Bacterial biofilms on aquatic surfaces primers 27f (5`–AGA GTT TGA TCC TGG CTC AG–3`) and 1492r (5`–TAC GGY TAC

CTT GTT ACG ACT T–3`). We used the TA Cloning Kit by Invitrogen. Instead of sequencing each single clone, we separated distinct clones by performing a restriction fragment length polymorphism (RFLP) with MspI (0.08 U sample–1, 37 °C over night). Only clones exhibiting different banding patterns were sequenced (4base lab, Reutlingen, Germany).

Statistics. We assumed that plant chemistry and environmental conditions might influence the biofilm community composition. Thus, biofilms experiencing comparable conditions should display a similar community composition. We related the DGGE banding patterns to plant chemistry (Hempel & Gross, submitted) and environmental conditions with a BEST–ENV analysis to see which factors explain the differences between both plants best. First of all, the DGGE banding patterns were transformed into a presence-absence matrix. With these data a dissimilarity matrix was calculated based on Bray Curtis dissimilarity. A second dissimilarity matrix was calculated for standardised environmental data with Euclidean distance. For the plant chemistry, we chose tissue nitrogen, carbon, phosphorus, chlorophyll and total phenolic content and as environmental factors water level, temperature, conductivity and pH. The data were normalised to allow a comparison between different units.

This means, that all data are placed on a common scale by subtracting the mean of each variable from each value and divide the product by the standard deviation. This yields values in the range of –2 to +2. The ranks of both matrices were compared by

Spearman rank coefficient (ρ) to find the best match between them. To provide statistic validation, 999 permutations were carried out. These analyses were performed with Primer 6 (Version 6.1.6, Primer E Ltd.).

50 III Bacterial biofilms on aquatic surfaces

RESULTS

Environmental variables and plant condition

Environmental conditions changed during the sampling period from July to

October (Table 3.1). The temperature dropped from the beginning to the end by about 10 °C. The water level at the sampling dates was more or less constant at

319 cm, but was at maximum 25 cm higher or 27 cm lower. Conductivity and pH were also relatively constant (267 ± 14 µS cm–1 and 8.3 ± 0.2, respectively).

Throughout the sampling period in 2006, Myriophyllum spicatum shoots were 30 –

45 cm long and exhibited dark green leaves, the typical red coloured stems, and were not densely colonized with epiphytic algae. Potamogeton perfoliatus shoots were

20 – 50 cm long and during summer had intact, brightly green coloured leaves. Both plant species did not show severe signs of grazing. While M. spicatum did not show any sign of senescence throughout the sampling period, P. perfoliatus had brownish leaves at the last two sampling dates. At the beginning of the sampling period, the artificial substrates were covered with a thin layer of bacteria and algae and with increasing exposition time by several layers of Dreissena polypmorpha (zebramussel;

Mollusca).

Table 3.1. Environmental variables measured at the sampling dates. Water level data are of the water gauge at the harbour in Konstanz. Sampling Temperature Water level Conductivity No. sampling date pH Date [°C] [cm] [µS cm–1] 17. July 1 25.5 331 251 8.30 31. July 2 25.6 310 248 8.61 15. August 3 18.7 324 265 8.53 29. August 4 17.6 321 263 8.27 12. September 5 20.0 324 263 8.06 22. September 6 18.9 329 274 8.41 09. October 7 16.2 322 286 8.01 23. October 8 15.6 292 282 8.23

51 III Bacterial biofilms on aquatic surfaces

Denaturing gradient gel electrophoresis (DGGE)

Biofilm bacterial community composition on apices and leaves of Myriophyllum spicatum

(summer 2005). In July and August, the DGGE banding pattern differed between apices and leaves. Most of the apices samples clustered together separately from the leaf samples (Figure 3.1A and B). In both month, however, some replicates diverged and clustered with the respective other plant part. In October 2005, the differences between apex and leaves were less apparent (Figure 3.1C) indicating changes in the plants’ physiology. In July, the bacterial diversity based on the banding pattern seemed to be highest on M. spicatum leaves. In contrast, the BCC of the apices was less diverse, and exhibited a higher variability of the banding pattern than that of the leaves (Figure S3.1A, see supplementary). These differences seemed to fade when macrophytes aged (Figure S3.1B, see supplementary).

Figure 3.1. Cluster analysis of DGGE banding patterns of Myriophyllum spicatum samples in 2005. A) July B) August C) October as determined by unweighted pair group method with arithmetic mean (UPGMA).

Biofilm community composition on different substrates. In summer 2006, DGGE banding patterns of plant apices revealed that the bacterial biofilm on M. spicatum differed from that on P. perfoliatus and the artificial substrate (Figure 3.2A). Most of the P. perfoliatus apex and artificial substrate samples clustered together, whereas 52 III Bacterial biofilms on aquatic surfaces

apex samples of M. spicatum formed a distinct cluster. Only both plant samples from the end of August differed from all other samples. The BCC on the leaves of both macrophytes and the artificial substrates was rather similar (Figure 3.2B) and a succession in BCC was not observed neither in apices nor leaves or artificial substrates.

Figure 3.2. Cluster analysis of DGGE banding patterns of the heterotrophic biofilm community on Myriophyllum spicatum, Potamogeton perfoliatus and the artificial substrates in 2006. A: Apices of both plant species in comparison to the artificial substrates. B: Lower leaves of both plant species in comparison to the artificial substrate. A & B: Samples from Lake Constance. C: Comparison between mesocosm M. spicatum and mesocosm artificial substrates MS: M. spicatum, PP: P. perfoliatus, Art: artificial substrate.

In the mesocosm, we observed a distinct cluster of the early fall samples of

M. spicatum apices, another cluster of the two leaf samples from October and a third cluster, which contains all artificial substrates and each one apex and leaf sample

(Figure 3.2C).

Sequencing of single DGGE bands. Bands cut from the gels were re–amplified and sequenced. We analysed 14 bands of the apices (Figure S3.2A, see supplementary)

53 III Bacterial biofilms on aquatic surfaces with BLAST (Altschul et al. 1990). Most of the retrieved sequences belonged to the betaproteobacteria (50%), gammaproteobacteria (21%) and the rest (29%) could be only assigned to the domain bacteria (Table 3.2). The closest relatives based on

BLAST search are from soil or freshwater habitats. We analysed 16 bands of the leaves, of which 4% belonged to the gammaproteobacteria, 6% each to the actinobacteria, betaproteobacteria, cyanobacteria and chloroplasts. The rest (50%) could only be assigned to the domain bacteria (Table 3.3, Figure S3.2B, see supplementary). Here, the sequences were mostly similar to those from other freshwater studies.

54 III Bacterial biofilms on aquatic surfaces

Table 3.2. BLAST analysis of 16S rRNA gene sequences from biofilm on the apices of Myriophyllum spicatum, Potamogeton perfoliatus and the artificial substrate.

Identity Substrate Sampling Datea) Most similar to Bacterial group Accession no. Sourceb) (%)

M. spicatum 31. July (18) Uncultured bacterium clone 164ds20 100 Bacteria; environmental samples. AY212616 Equine faecal contamination

P. perfoliatus 31. July (19) Uncultured Burkholderiales bacterium clone Hv(lab)_2.15 99 Betaproteobacteria EF667915 Basal metazoan Hydra

P. perfoliatus 31. July (16) Methylophilus sp. U33 98 Betaproteobacteria EU375653 Organic pollutants degradation

P. perfoliatus 15. August (14) Acinetobacter sp. HTYC28 98 Gammaproteobacteria EU372908 China sea

Artificial 15. August (15) Ralstonia sp. JB1B3 100 Betaproteobacteria EU375662 Organic pollutants degradation

M. spicatum 12. September (20) Uncultured Ideonel la sp. clone GASP–MA2S1_A04 98 Betaproteobacteria EF662829 Bacterial soil communities in Michigan

P. perfoliatus 12. September (11) Uncultured bacterium clone MA34_2003DFa_B05 90 uncultured bacterium EF378328 Agricultural soil community

M. spicatum 22. September (8) Uncultured bacterium clone 164ds20 93 Bacteria; environmental samples AY212616 Equine faecal contamination

P. perfoliatus 22. September (10) Clonothrix fusca strain AW–b 93 Gammaproteobacteria DQ984190 Clonothrix fusca Roze 1896

M. spicatum 09. October (4) Uncultured betaproteobacterium clone CH_02 97 Betaproteobacteria EF562573 Complex organic matter degradation

M. spicatum 09. October (12) Uncultured Ideonel la sp. clone GASP–MA2S1_A04 97 Betaproteobacteria EF662829 Bacterial soil communities in Michigan

M. spicatum 09. October (5) Uncultured Rubrivivax sp. clone GASP–WDOW1_D03 97 Betaproteobacteria EF075729 Soil in pasture and cropping systems

P. perfoliatus 23. October (1) Methylomonas methanica clone VAS23 72 Gammaproteobacteria AM489704 Baltic Sea sediments

Artificial 23. October (3) Uncultured bacterium isolate DGGE gel band D2/3_1 98 Bacteria; environmental samples. EF208596 Daggyai Tso geothermal field of Tibet

a) Numbers in brackets indicate the bands excised from the gels in figure S2 b) Identical source names represent identical studies

55 III Bacterial biofilms on aquatic surfaces

Table 3.3. BLAST analysis of 16S rRNA gene sequences from biofilm on the lower leaves of Myriophyllum spicatum, Potamogeton perfoliatus and the artificial substrate. Identity Accession Substrate Sampling Datea) Most similar to Bacterial group Source b) (%) no.

M. spicatum 31. July (6) Uncultured bacterium clone YCC126 95 Bacteria; environmental samples EF205477 Geothermal regions in central Tibet

M. spicatum 31. July (7) Uncultured actinobacterium clone IRD18A09 96 Bacteria; environmental samples AY947900 River bacterioplankton

P. perfoliatus 31. July (2) Calycanthus floridus chloroplast 94 DQ629462 Calycanthus

P. perfoliatus 31. July (3) Clonothrix fusca strain AW–b 93 Gammaproteobacteria DQ984190 Clonothrix fusca Roze 1896

P. perfoliatus 31. July (4) Clonothrix fusca strain AW–b 92 Gammaproteobacteria DQ984190 Clonothrix fusca Roze 1896

P. perfoliatus 15. August (9) Acinetobacter sp. Hg4–05 16S 99 Gammaproteobacteria EU372903 China sea

Artificial 15. August (8) Ralstonia sp. JB1B 99 Actinobacteria EU375662 Organic pollutants degradation

M. spicatum 29. August (12) Uncultured bacterium clone M1–53 96 Bacteria; environmental samples EU015116 Estrogen–degrading membrane bioreactors

M. spicatum 29. August (13) Hydrogenophaga taeniospiralis clone SE57 94 Betaproteobacteria AY771764 Arctic

P. perfoliatus 12. September (15) Uncultured bacterium clone cams48–2 95 Bacteria; environmental samples AY544224 Lake Constance M. spicatum

P. perfoliatus 12. September (17) Clonothrix fusca strain AW–b 93 Gammaproteobacteria DQ984190 Clonothrix fusca Roze 1896

M. spicatum 09. October (21) Uncultured cyanobacterium clone RD107 96 Cyanobacteria DQ181677 East Antarctic lakes

M. spicatum 09. October (23) Uncultured bacterium clone YCC126 95 Bacteria; environmental samples EF205477 Geothermal regions in central Tibet

P. perfoliatus 09. October (20) Uncultured bacterium clone cams48–2 95 Bacteria; environmental samples AY544224 Lake Constance M. spicatum

P. perfoliatus 23. October (26) Uncultured bacterium clone M1–53 88 Bacteria; environmental samples EU015116 Estrogen–degrading membrane bioreactors

Uncultured bacterium isolate DGGE gel band P. perfoliatus 23. October (27) 84 Bacteria; environmental samples EF396239 Stream biofilm OTU_1 a ) Numbers in brackets indicate the bands excised from the gels in figure S2 b) Identical source names represent identical studies

56 III Bacterial biofilms on aquatic surfaces

Clone library

We obtained 64 clones of the M. spicatum apex biofilm with 31 different restriction patterns of which 18 were sequenced (Table 3.4). The majority of the clones could not be assigned to different bacterial groups, although related strains have been found in various other studies in aquatic and soil environments throughout the world. Nearly all of the sequences referred to uncultured clones except the Escherichia coli strain SA–

Y1–39, which was found four times. Eight clones were closely related to cyanobacteria found in agricultural soil in Michigan, and seven were affiliated with

M. spicatum chloroplasts.

57 III Bacterial biofilms on aquatic surfaces

Table 4.3. BLAST analysis of 16S rRNA gene sequences of clones obtained from Myriophyllum spicatum apices

RFLP Identity Most similar to Bacterial group Accession no. Source group (%) 1 Uncultured bacterium clone Pia–s–9 98 Bacteria; environmental samples EF632943 Aquatic environments of the Andean Altiplano 2 Uncultured bacterium clone 1H_31 97 Bacteria; environmental samples AY546500 Transbaikal soda lake sediments 3 Uncultured bacterium clone 57–ORF18 94 Bacteria; environmental samples DQ376571 Phosphate removal from wastewater 4 Myriophyllum spicatum chloroplast 100 EF207450 M. spicatum, Saxifragales 5 Uncultured freshwater bacterium clone 965006H10.x1 98 Bacteria; environmental samples DQ065573 primary producer 6 Uncultured proteobacterium clone MS098A1_A01 98 Proteobacteria; environmental samples EF698192 Human inflammatory bowel diseases 7 Uncultured Methylophilus sp. clone BR03BD02 98 Betaproteobacteria DQ857216 Rat gastrointestinal tract 8 Uncultured cyanobacterium clone GASP–MA1S2_G12 99 Cyanobacteria; environmental samples EF662391 Bacterial soil communities in Michigan 9 Uncultured bacterium clone Pia–s–9 98 Bacteria; environmental samples EF632943 Aquatic environments of the Andean Altiplano 10 Uncultured bacterium clone 2\SC\48 96 Bacteria; environmental samples EU340184 Aquatic Macrophytes 11 Uncultured bacterium clone Pia–s–9 98 Bacteria; environmental sample EF632943 Aquatic environments of the Andean Altiplano 12 Escherichia coli strain SA–Y1–39 100 Gammaproteobacteria EU420955 Jiaozhou Bay, China 13 Uncultured Rhizobium sp. clone AUVE_04B08 98 Alphaproteobacteria EF651117 Australian Vertisol 14 Uncultured bacterium clone 231ds5 98 Bacteria; environmental samples AY212678 Equine fecal contamination 15 Uncultured bacterium clone ANTLV2_G07 98 Bacteria; environmental samples DQ521527 Lake Vida, Antarctica 16 Uncultured soil bacterium clone L1A.3A11 98 Bacteria; environmental samples AY988750 Soil 17 Bacterium RSB–3–5 94 Bacteria AY822525 Sand dune plant species 18 Uncultured betaproteobacterium clone CH_02 97 Betaproteobacteria EF562573 Complex organic matter degradation a) Identical source names represent identical studies

58 III Bacterial biofilms on aquatic surfaces

Impact of plant chemistry and environmental factors on community composition

We performed a separate BEST–ENV analysis for leaves and apices of both plants to elucidate the major factors influencing bacterial community composition as determined with DGGE. The highest correlation coefficient was achieved for carbon and phosphorus content of the apices, which was not improved when adding water conductivity yet due to the low number of replicates did not result in significant effects (ρ = 0.249, P > 0.16, n = 13). The same was true for the environmental factors water level and conductivity (ρ = 0.140, P = 0.27, n = 13).

We performed the same analyses for the leaf sections. Plant chemistry did not explain the variability of our samples. Of all environmental factors, conductivity explained most of the variability (ρ = 0.383, P = 0.026). This correlation is not improved, if all environmental factors together with plant chlorophyll are taken into account (ρ = 0.378 (P = 0.059)).

The NMDS analysis of the BCC based on DGGE banding pattern does not show clear groups based on plant species or season, neither for plant apices nor leaves

(Figure 3.3).

Figure 3.3. Non metric Dimensional scaling analysis of DGGE banding patterns of Myriophyllum spicatum (white triangles) and Potamogeton perfoliatus (grey triangles). A) Apices B) Leaves. The numbers given indicate the sampling date details see Table 3.1).

59 III Bacterial biofilms on aquatic surfaces

DISCUSSION

We assessed the bacterial community composition on two submerged macrophytes, and an artificial substrate in Lake Constance with DGGE and a clone library. To elucidate the impact of allelochemicals, e.g., polyphenols, in the water column of a dense Myriophyllum spicatum stand, we investigated the bacterial community composition on M. spicatum and artificial substrates in a densely planted mesocosm. The biofilm community composition differed between M. spicatum apices and leaves in 2005 and between plants (M. spicatum and Potamogeton perfoliatus) and the artificial substrate in 2006. In the mesocosm, M. spicatum apices and leaves had a different biofilm community composition than the artificial substrates exposed directly in the plant patch.

The 2005 DGGE data indicate that spatial differences exist between apices and leaves of M. spicatum. These differences lessen towards autumn, and might be related to differences in plant chemistry. Conductivity, total phosphorus and phytoplankton biomass influenced the pelagic and particle–associated attached BCC in four lakes in northern Germany (Allgaier & Grossart 2006). In our study, phosphorus and carbon content did not explain changes in BCC on the apices, and was also not affected by conductivity. The BCC on the leaves, however, was influenced by conductivity. It could be that the influence of conductivity on BCC is more important closer to the ground due to sediment resuspension or local water currents and seepage. Other factors that might affect community composition are leaf structure, especially surface to biomass ratio, grazing and nutrient availability.

We additionally assessed the bacterial diversity on M. spicatum apices with a clone library to get a more detailed picture of diversity in the biofilm. With DGGE, we can assess the overall diversity but the information on single strains is often lost, especially if they occur in low abundance. In addition, sequencing of DGGE bands usually yields <500 bp, which is sufficient for phyla assignment, but not for single strain affiliation. The results of the clone library suggest a high diversity with almost every second clone representing a different 16S rRNA gene but hardly any affiliation

60 III Bacterial biofilms on aquatic surfaces

to bacterial groups, although we used only high quality sequences to assure correctness of our data. Some of the sequenced DGGE bands could be assigned to single strains or at least to the specific phylogenetic group. Not surprisingly, we found high amounts of M. spicatum chloroplasts in the clone library, most likely due to DNA extraction of damaged plant tissue. The leaves get damaged when they are plucked and the biofilm detachment procedure may have enhanced tissue damage

(Hempel et al. 2008).

Betaproteobacteria affiliated with Ralstonia sp. are known to degrade aromatic compounds (Steinle et al. 1998, Ryan et al. 2007). In contrast, Acinetobacter sp. and

Clonothrix fusca are gammaproteobacteria. While Acinetobacter sp. is a known as a common saprophyte using a wide range of organic compounds, C. fusca is a sheathed methanotroph living in biofilms of running waters (Vigliotta et al. 2007).

Interestingly, several of our sequences were affiliated with methanotrophic bacteria, indicating the availability of methane as a bacterial substrate. This is reasonable, since methane is transported through the lacunar system from the roots to the leaves in aquatic plants. Previously, it has been shown that this mechanism supports methane oxidation by epiphytic bacteria (Schuette 1996, Heilman & Carlton 2001).

These results indicate that the biofilm community composition on M. spicatum is well adapted to the various organic compounds, such as polyphenols, and other substrates released by the plant.

The majority of closely related sequences found in the BLAST database, however, refer to yet uncultured strains and, thus, do not allow conclusions on their ecosystem functions. On the other hand, these findings indicate that the biofilm BCC on the macrophytes is a rather unique one esp. on M. spicatum, which is mostly likely well adapted to the specific environmental conditions in this microenvironment. To better elucidate functions of biofilm BCC, further studies are needed, preferentially with cultivation–based approaches allowing for in-depth physiological studies.

Both DGGE and clone library sequences were mainly affiliated to strains originating in various limnetic habitats suggesting that - like many other freshwater

61 III Bacterial biofilms on aquatic surfaces bacteria (Lindstrom et al. 2005) - they are widely distributed among habitats (lakes, rivers, sewage). The relative high amount of bacteria associated with soil or agriculture indicate a constant terrestrial source. This might be explained by the vicinity of our sampling site to the Island of Reichenau, which is intensively used for agriculture and lake water is used for irrigation causing run-off from the fields into the lake. Two sequences of our study closely related to sequences found in a year

2000 study on bacterial biofilms on M. spicatum in Lake Constance (Acc. no.

AY544224, Table 3.2, (Walenciak 2004)) indicate the presence of specific and constant bacterial members of these biofilms.

The presence of a specific biofilm BCC on M. spicatum is also demonstrated by the fact that its BCC in 2006 with one exception was quite distinct from that on

P. perfoliatus apices and the artificial substrate. Interestingly, at the end of August

2006, both plant apices do not cluster with the related samples, when the concentration of phenolic compounds (total phenolics, anthocyanin, tellimagrandin

II) of M. spicatum were approx. 25 – 50% lower than on sampling dates before and afterwards (Hempel & Gross, submitted). In previous studies we found that light strongly affects the concentration of phenolic compounds in M. spicatum (Gross

2003). August 2006 was very rainy and cold. The temperature dropped from 26 °C in

July to 16 °C in August (Table 1). Thus light and temperature might have caused the observed changes.

On the leaves of both plant species, however, we did not find such distinct differences as on the apices. Although the bacterial biofilm composition on all substrates formed some subclusters, the different substrates did not form distinct clusters (Figure 3A and B). In parallel to our findings of the FISH analyses, we suggest that the BCC of the older biofilm on the leaves is more similar between different substrates than the younger one on the apices.

One general aspect we did not investigate due to time constrains based on the large sample numbers is the presence of fungi and grazers. Fungi, like bacteria, occur in a wide range of habitats and serve different functions (Nikolcheva et al. 2003).

62 III Bacterial biofilms on aquatic surfaces

Often they are responsible for degrading complex matter, but are also known to produce antibiotics. Thus, they could influence the BCC by providing degradation products or inhibiting bacteria. A second factor not determined was bacterivorous flagellates. As part of the bacterial loop, they impose grazing pressure upon the bacterial community, which will not only change the size of bacteria (more filamentous bacteria) but may select to certain strains (Jürgens & Matz 2002).

In summary, our results reveal a distinct BCC on the apices vs. leaves of

M spicatum in summer, but less pronounced differences in autumn. Compared to

P. perfoliatus and artificial substrates, M. spicatum apices exhibit a particular biofilm

BCC. This is in good agreement with our findings of M. spicatum and Chara aspera from Lake Constance in comparison to those from a brackish habitat (Hempel et al.

2008). This is one of the first studies investigating the BCC on living and neutral substrates in a meso- to oligotrophic lake over a complete growing season with different molecular tools. While other studies described the BCC on riverine biofilms, terrestrial leaves or artificial substrates, we sought to fill the knowledge gap that exits on bacterial biofilms in oligotrophic lakes on natural substrates.

Acknowledgment: This work was supported by the German Science Foundation with grant CRC454, project A2 to EMG. We would like to thank S. Nadj for technical assistance and J. Hesselschwerdt and S. Werner for help with sampling. Special thanks go to the Microbial Ecology group of HPG at the IGB Neuglobsow for help with DGGE.

63 III Bacterial biofilms on aquatic surfaces

SUPPLEMENTARY

Figure S3.1: DGGE profiles from amplified 16S rRNA gene fragments of the biofilm on Myriophyllum spicatum apices and leaves in 2005. A) July B) August and October. STD: standard

64 III Bacterial biofilms on aquatic surfaces

A) MAP MPPPA MAMM M A SSPPPAAM A 87 6543 2 11 4 1 16 15 19 12 14 3 20 8 5 18

10

B) S AMMM P AAP MPAMMSS P AAAPPMP 123 45 67

9 8 21

23 26 6 12

13 20 27 2 7 3 15 17 4

Figure S3.2: DGGE banding patterns and excised bands of Myriophyllum spicatum, Potamogeton perfoliatus and the artificial substrates in 2006. A) Apices B) Leaves. M: M. spicatum, P: P. perfoliatus A: Artificial substrates, S: Standard. Number 1 – 7 indicate the sampling dates as listed in Table 3.1. The closest relatives based on BLAST search are listed in Table 3.2 (apices) and Table 3.3 (leaves).

65 III Bacterial biofilms on aquatic surfaces

66

Chapter IV

Spatio–temporal dynamics of the bacterial biofilm on two freshwater macrophytes and an artificial substrate in Lower Lake Constance

Melanie Hempel & Elisabeth M. Gross

IV Spatio-temporal biofilm dynamics

68 IV Spatio-temporal biofilm dynamics

ABSTRACT: The heterotrophic biofilm community composition (BCC) on the aquatic macrophytes Myriophyllum spicatum (MS), Potamogeton perfoliatus (PP) and artificial substrates (Art) was compared on a spatial (plant age) and temporal (season) scale in 2006. If plants were neutral substrates for microorganisms, no differences in BCC compared with the propylene sheets should be observed. Myriophyllum spicatum contains polyphenolic allelochemicals, which inhibit algae, cyanobacteria and bacteria, and might influence the BCC. Major bacterial groups were identified by FISH (fluorescence in situ hybridization). We also measured plant stoichiometry and content of phenolic compounds. To investigate the initial bacterial colonization, we exposed axenic M. spicatum in an outdoor mesocosm for 72 hrs. Total bacterial cell counts were highest on the artificial substrate (1.3 ± 0.3×106 cm–2), followed by M. spicatum (0.8 ± 0.3×106 cm–2) and P. perfoliatus (0.3 ± 0.2×106 cm–2), and increased towards autumn. Members of the CFB group (mean on MS 22%, PP 21%, Art 16%), alphaproteobacteria (mean on all substrates 19%) and betaproteobacteria (mean on MS 17%, PP 31%, Art 7%) were always abundant, while actinobacteria and planctomycetes occurred less frequently. Potamogeton perfoliatus had the highest proportion of gammaproteobacteria (average 19%), while on M. spicatum and the artificial substrate the shares were 9% and 4%, respectively. Alpha– and betaproteobacteria displayed a slight seasonality. The initial colonization of axenic M. spicatum was dominated by the CFB group at 5 hrs, followed by betaproteobacteria. Plant chemistry differed substantially between plants, and within M. spicatum. Differences in BCC were mainly explained by environmental factors (water level, conductivity, temperature, pH) and by plant chemistry, especially carbon and total phenolic content.

Keywords: Myriophyllum spicatum, Potamogeton perfoliatus, allelochemical interactions, fluorescence in situ hybridization, phenolic compounds, water level, conductivity

69 IV Spatio-temporal biofilm dynamics

INTRODUCTION

Submerged macrophytes are not only the major primary producers in the littoral zones of lakes they also structure this zone by reducing sediment resuspension, providing spawning area and shelter for young fishes and zooplankton. They further offer a vast surface area for the attachment of various organisms, from bacteria and algae to invertebrates (Jeppesen et al. 1998).

Heterotrophic bacteria largely contribute to the overall nutrient cycling and interact in various ways with other organisms by relocating nutrients, converting degradation products, restoring growth forms of macroalgae, facilitating spore attachment and preventing grazing (Joint et al. 2000, Buesing & Gessner 2006,

Marshall et al. 2006). Further, they can invade and damage tissue and promote biofouling (Underwood 1991). Excessive biofilm formation also has potential negative consequences for the host due to decreased exchange of nutrients and reduced photosynthesis (Phillips et al. 1978, Sand-Jensen & Søndergaard 1981). In the root sections of macrophytes, bacteria are generally recognized as important mediators for macrophytes` nutrient uptake, esp. nitrogen (Eriksson & Weisner

1999). In return, macrophytes provide resources for bacteria, e.g., exuded organic compounds or gases like methane from the root zone, which is transported through the lacunar system to the above–ground plant parts and released into the water column (Gross et al. 1996, Schuette 1996, Heilman & Carlton 2001).

Taking these aspects into account, the littoral zone is not solely characterized by the macrophyte community but also by their autotrophic and heterotrophic biofilms.

From terrestrial plants it is known that they often invest in chemical defence against competitors, pathogens and herbivores. A macrophyte with a high allelochemical potential in Lake Constance is Myriophyllum spicatum L. Besides its canopy forming growth, it produces high amounts of hydrolysable polyphenols that retard larval growth and inhibit bacteria and cyanobacteria (Choi et al. 2002, Leu et al. 2002, Walenciak et al. 2002). These polyphenols are not only in the plant tissue but may gradually leak from leaves into the surrounding water. Thus, biofilms

70 IV Spatio-temporal biofilm dynamics establishing on the surface of these plants are exposed to these compounds in high concentrations, which may result specific adaptations, e.g., to polyphenols, which can be utilized as substrate (Müller et al. 2007). Another macrophyte growing in the vicinity of M. spicatum in Lake Constance is the pondweed Potamogeton perfoliatus. It forms large stands up to the water surface, but does not contain polyphenols, only few other phenolic compounds. Several pondweeds, however, contain diterpenes which inhibit microalgae (DellaGreca et al. 2001), but none of them has been found in

P. perfoliatus.

Little attention has been paid to the heterotrophic biofilm on submerged macrophytes, especially to spatial differences between older and younger leaves.

Hence, the knowledge on the bacterial community composition (BCC) of these biofilms is extremely scarce. Molecular studies investigating aquatic surfaces have mostly been on artificial substrates (Brummer et al. 2000, Olapade & Leff 2006), algae

(Rao et al. 2006), among them diatoms (Grossart et al. 2005). Others investigated the heterotrophic biofilm on macrophytes with cultivation dependent techniques (Chand et al. 1992), which can be rather selective depending on growth conditions applied.

In this study, we focus on the spatial and temporal community composition of the heterotrophic biofilm on aquatic surfaces. We investigated Myriophyllum spicatum,

Potamogeton perfoliatus and an artificial substrate throughout the vegetation period in summer 2006 with fluorescence in situ hybridization (FISH). We distinguished along a vertical axis between plant apices, middle and lower leaves. Plant quality along this gradient was analysed by measuring carbon, nitrogen, phosphorus, chlorophyll and total phenolic content (anthocyanins and tellimagrandin II).

71 IV Spatio-temporal biofilm dynamics

MATERIAL & METHODS

Sampling site. The sampling site (N47°42.247, E9°02.289) is located at the Isle of

Reichenau in Lower Lake Constance, Germany, near a gravel ridge. Myriophyllum spicatum, Potamogeton perfoliatus and artificial substrates were taken biweekly in triplicates by snorkelling in a depth of 1.5 – 2 m in the time from 17 July to 09 October

2006. Plants and artificial substrate samples were stored individually in polyethylene tubes at 4 °C until processing. For chemical analysis, plants were stored in plastic bags at 4 °C until analysis the next day. Temperature, oxygen, conductivity and pH were measured in the water column at each sampling date. For a detailed description of the experimental set–up see Hempel et al. (submitted).

Detachment of epiphytic biofilm. In the laboratory, plant length was measured and the overall state of the plant was recorded. Artificial substrates were documented through photography. Each one leaf of the apex, one leaf of the 1 – 10 cm (middle leaf) and 10 – 25 cm (lower leaf) shoot section were taken from three plants originating from three different stands and these nine leaves were transferred to 1 ml sodium pyrophosphate (0.1 M Na4P2O7×10 H2O) containing

3.7% formaldehyde. Of the artificial substrates, one section of approx. 1 cm2 was cut off and also transferred into 1 ml sodium pyrophosphate. The detachment of the biofilm had been optimized and was conducted as described in (Hempel et al. 2008).

After detachment, leaves were transferred into one millilitre of tap water and stored at 4 °C until leaf surface was measured. The detached biofilm was filtered onto white polycarbonate filters (0.2 µm; Ø 25 mm, Schleicher & Schuell) and stored at –20 °C.

Fluorescence in situ hybridization (FISH). FISH was performed following a protocol consisting of a hybridization step at 46 °C for three hours and a washing step for 15 min at 48 °C (Pernthaler et al. 2001). Filters were counterstained with

DAPI (4’,6–Diamidino–2–phenylindol, 1 µg ml–1, 5 min). Stained cells were counted under an epifluorescence microscope (Labophot 2, Nikon) at an excitation of 549 nm and 1000× magnification. The probes used are listed in Table 4.1, and further details are available at probeBase (Loy et al. 2003).

72 IV Spatio-temporal biofilm dynamics

Table 1. Oligonucleotide probes used in this study. % Form– Probe a) Sequence Target group Reference amide EUB338 GCTGCCTCCCGTAGGAGT 35 Most Bacteria (Amann et al. 1990) NON338 ACTCCTACGGGAGGCAGC 35 Competitor to EUB (Wallner et al. 1993) ALF968 GGTAAGGTTCTGCGCGT 20 Alphaproteobacteria (Neef 1997) BET42a b) GCCTTCCCACTTCGTTT 35 Betaproteobacteria (Manz et al. 1992) GAM42a b) GCCTTCCCACATCGTTT 35 Gammaproteobacteria (Manz et al. 1992) PLA886 b) GCCTTGCGACCATACTCCC 35 Planctomycetes (Neef et al. 1998) HGC96a TATAGTTACCACCGCCGT 25 Actinomycetes (Roller et al. 1994) CF319a TGGTCCGTGTCTCAGTAC 35 Bacteroidetes (Manz et al. 1996) a) Probes were labelled with cy 3. b) For these probes a competitor probe was used

Measurement of leaf surface. To relate total cell counts to the surface area of the plants, leaves were photographed with a Nikon D70S and analysed with

“Makrophyt” a computational program designed by the scientific workshops of the

University Constance. Each leaf was photographed with three different exposure times and the mean leaf size calculated. The calculated area of M. spicatum was multiplied by π (3.14) to account for the circular shape of the leaves. To calculate the leaf surface of P. perfoliatus, the area was multiplied by two since the oval leaves are laminar.

Chemical analysis. Different plant parts were analysed by spectrophotometry for total phenolic content (Folin–Ciocalteau assay; (Box 1983)), anthocyanin (Murray &

Hackett 1991), carbon, nitrogen, phosphorus (Choi et al. 2002), chlorophyll a and b

(Porra 1990) content and for M. spicatum only, by HPLC for tellimagrandin II (Müller et al. 2007). Folin–sensitive compounds in P. perfoliatus are only to 50% phenolic compounds (Choi et al. 2002), thus the Folin–results were halved to reflect the true total phenolic content. Myriophyllum spicatum was part of our routine sampling campaign, and three replicates resulting from three different stands were measured.

We measured P. perfoliatus plants originating from one stand, and thus only one measurement for each sampling date is available. We experienced that plants originating from one location usually do not differ substantially in chemical composition.

73 IV Spatio-temporal biofilm dynamics

Initial colonization of axenic M. spicatum. Axenic M. spicatum rooted plants of ca.

8 cm length were planted into an outdoor concrete basin (2×2×1 m) in the yard of the

Limnological Institute in May 2007. The plants were anchored into the sediment by tying them to stainless steel screw–nuts. After 5, 24, 36, 48 and 72 hours, three leaves from three different plants were taken and analyzed for bacterial community composition with FISH. Detachment of bacteria and FISH were performed as described above.

Statistics. To analyse significant differences and potential interactions between the biofilm community composition on the substrates at different times, we used one– way ANOVAs to compare differences between all three substrates or between individual sampling dates, Mann–Whitney rank sum tests to distinguish differences between parts of both plants and Pearson correlations to investigate continuous seasonal changes (Sigma Stat 3.11, Systat Software, Inc.). The proportional FISH data were arcsin transformed, and data for gammaproteobacteria, planctomycetes, actinomycetes and CFB were additionally x1/4 transformed to yield equal variance. To account for the multiple comparisons, we set our level of significance at α = 0.01.

We related the FISH abundance data to plant chemistry and environmental conditions with a BEST–ENV analysis to see which factors explain the differences between both plant species best. A dissimilarity matrix was calculated based on Bray

Curtis dissimilarity for square root transformed FISH data, and a dissimilarity matrix was calculated for standardised environmental data with Euclidean distance as described before (Hempel et al., submitted). For the plant chemistry, we chose tissue nitrogen, carbon, phosphorus, chlorophyll and total phenolic content and as environmental factors water level, temperature, conductivity and pH. The analyses were performed with Primer 6 (Version 6.1.6, Primer E Ltd.).

74 IV Spatio-temporal biofilm dynamics

RESULTS

Initial colonization of axenic Myriophyllum spicatum

Total bacterial cell counts on freshly colonized M. spicatum were in the same range as data from the lake sampling site or even higher (0.75 – 2.3×106 cells cm–2), thus plants are colonized very rapidly within 5 hours (Figure 4.1A). A distinct succession of bacterial groups was observed (Figure 4.1B). In the first five hours, members of the

CFB–group made up the major portions of the biofilm (46% of DAPI counts), followed by betaproteobacteria (24% of DAPI counts) and actinomycetes (6% of

DAPI counts). After 24 h, the biofilm was dominated by betaproteobacteria (56 – 62% of DAPI counts), followed by members of the CFB–group. Alpha– and gammaproteobacteria increased steadily at low numbers to a maximum of 4% and

5% of total cell counts, respectively.

Figure 4.1: Initial colonization of axenic Myriophyllum spicatum in a mesocosm over 72 hours. A) Total bacterial cell counts over the sampling duration B) Bacterial biofilm community composition as determined by FISH.

Alphaproteobacteria Betaproteobacteria Gammaproteobacteria Actinobacteria Planctomycetes CFB-Phlyum

75 IV Spatio-temporal biofilm dynamics

Total bacterial cell counts

On M. spicatum, we did not observe any significant influence of plant age or season on the total bacterial cell counts. On the apices, cell counts

(0.63 ± 0.24×106 cells cm–2) were in the range to that on middle leaves

(0.66 ± 0.1×106 cells cm–2), and both were slightly lower than on lower leaves

(1 ± 0.31×106 cells cm–2). Towards autumn, a slight increase in total bacterial cell counts on the lower leaves was monitored (Figure 4.2A). On Potamogeton perfoliatus, total bacterial cell counts were rather similar on the different plant parts throughout the season (apex: 0.43 ± 0.31×106, middle leaves: 0.2 ± 0.1×106; lower leaves:

0.28 ± 0.13×106 cells cm–2) with slightly higher cell counts for the apices (Figure 4.2B).

We observed a marked rise on 12 September (0.6 ± 0.3×106 cells cm–2) and a subsequent decline afterwards. The artificial substrates had about 13–fold higher cell counts at the end of the sampling campaign than at the beginning (from

0.3± 0.03×106 to 5 ± 3×106 cells cm–2). With values of 1.78 ± 1.5×106 cells cm–2 over the whole season, the bacterial numbers were higher on the artificial substrate than on both macrophytes (Figure 4.2C). The bacterial cell counts on P. perfoliatus were lower than those on M. spicatum, especially between the middle and lower leaves (Mann-

Whitney rank sum test, middle and lower leaves both P < 0.001).

1.5 A B C

5 -2 -2

) cm 4 ) cm 6 1.0 6 3

0.5 2

1 Total cell counts (x10 Total cell counts (x10

0.0 0 Aug Sep Oct Aug Sep Oct Aug Sep Oct

Figure 4.2. Total cell counts on all substrates in the vegetated period. A) Myriophyllum spicatum B) Potamogeton perfoliatus C) Artificial substrate. White circles: Apex; triangle up: Middle leaves; triangle down: lower leaves. Note that the y–axis in 2C is differently scaled.

76 IV Spatio-temporal biofilm dynamics

Bacterial community composition

The bacterial community composition was assessed by FISH (fluorescence in situ hybridization). Usually, eubacterial counts were above 50% of DAPI counts

(65% ± 16, mean ± SD, for all dates and substrates). On all substrates, bacteria of the

CFB and beta– and alphaproteobacteria were abundant, followed by gammaproteobacteria, while actinomycetes and planctomycetes were of minor importance (Figure 4.3).

Spatial and temporal variability on different substrates. On M. spicatum, the biofilm community composition did not differ strongly between sampling dates or plant parts (Figure 4.3A, C and E). In general, proportions of CFB and betaproteobacteria were high throughout the vegetation period and ranged between 3 – 75% and 3 –

58% of DAPI counts, respectively, and no seasonal trends were observed. In a few cases, the hybridization efficiency was low, and no CFB bacteria could be detected, which might be caused by the low hybridization efficiency of less than 50% of this probe (see Figure 4.3).

The apices of M. spicatum had the highest CFB shares (32 ± 17% of DAPI counts) followed by the middle (16 ± 12% of DAPI counts) and lower leaves (15 ± 10% of

DAPI counts). Alphaproteobacteria increased on the apices during the season (2 –

43% of DAPI counts, Pearson correlation P < 0.01), stayed more or less constant on middle leaves (18 ± 7% of DAPI counts, Pearson correlation P = 0.761) and decreased on the lower leaves towards fall (from 46% to 10% of DAPI counts, Pearson correlation P = 0.0155). Planctomycetes and actinomycetes accounted together for

13% of the DAPI counts.

On P. perfoliatus, differences in biofilm community composition between different plant parts were even less pronounced (Figure 3B, D and F). Betaproteobacteria on the leaves doubled from July (13 ± 9% of DAPI counts) to September (52 ± 5% of

DAPI counts: One–way–ANOVA, DF = 6, F = 6.25, P < 0.001, Holm Sidak post hoc test

P < 0.005 for comparisons between July and September).

77 IV Spatio-temporal biofilm dynamics

A B 80 80

60 60

40 40

20 20

0 0 80 C D 80

60 60

40 40

20 20

0 0 80 E F 80 % of DAPI counts DAPI of % 60 60

40 40

20 20

0 0

80 G Aug Sep Oct

Alphaproteobacteria 60 Betaproteobacteria Gammaproteobacteria 40 Actinobacteria Planctomycetes CFB group 20

0

Aug Sep Oct

Figure 4.3: Biofilm community composition on Myriophyllum spicatum, Potamogeton perfoliatus and artificial substrates. A) M. spicatum apex B) P. perfoliatus apex C) M. spicatum middle leaf D) P. perfoliatus middle leaf E) M. spicatum lower leaf F) P. perfoliatus lower leaf G) Artificial substrate. n = 3. The standard deviation was between 7 – 135% and is not displayed for more visibility

78 IV Spatio-temporal biofilm dynamics

Bacteria belonging to the CFB–group made up the largest portion of all detected bacteria on all plant parts (range 10 – 50% of DAPI counts, Figure 3B, D and F). In general, CFB counts on all plant parts declined towards fall with an intermediate peak in mid August (54 ± 21% of DAPI counts). Alphaproteobacteria ranged from 8 –

27% of DAPI counts on all plant parts, and exhibited no seasonal trend with the exception of the apices, where their numbers increased (Pearson correlation

P = 0.0186), comparable to M. spicatum.

The biofilm on artificial substrates was dominated by alphaproteobacteria

(23 ± 10% of DAPI counts), CFB (16 ± 10% of DAPI counts) and betaproteobacteria

(8% of DAPI counts, Figure 4.3G). Gammaproteobacteria, planctomycetes and actinomycetes accounted together for up to 10% of the biofilm community. We did not find any seasonal trend for all bacterial groups on this substrate.

Comparison between substrates. Potamogeton perfoliatus had the highest proportions of betaproteobacteria, especially on the middle and lower leaves compared to the respective M. spicatum parts (t– test, P = 0.08 and P < 0.001, respectively), and both

P. perfoliatus parts had much higher counts than M. spicatum and the artificial substrates (17 ± 8% of DAPI counts for M. spicatum, 31 ± 12% of DAPI counts for

P. perfoliatus and 7 ± 8% of DAPI counts for artificial substrates, one–way ANOVA,

Holm Sidak post hoc test P < 0.0001 for both comparisons). CFB and alphaproteobacteria did not differ between plants, irrespective of plant part, and were found in the same range (16 – 20% of DAPI counts) on the artificial substrates.

Gammaproteobacteria were higher on P. perfoliatus than on M. spicatum on every plant part (19 ± 10% and 9 ± 4% of DAPI counts, respectively; Mann–Whitney rank sum test P < 0.001). Gammaproteobacteria on the artificial substrate were present on all sampling dates but in rather variable numbers (1 – 11% of DAPI counts).

Actinomycetes were present in low but consistent numbers on all substrates, ranging between 1 – 22% of DAPI counts but we did not find any effect of the substrate nor the plant age. Planctomycetes occurred more often on M. spicatum, and there at times in rather high abundance of up to 29% of DAPI counts on the apices, but they were

79 IV Spatio-temporal biofilm dynamics

also lacking frequently. On P. perfoliatus, planctomycetes were rather absent throughout the vegetation period, while we found constant but low numbers on the artificial substrates.

Chemical analysis

Plant stoichiometry. In M. spicatum, the molar C/N ratio ranged between 15 and 21 and was highly variable during the season and between plant parts. We observed both a gradient from apices (lowest) to lower leaves (highest), and additionally a decline with season (Figure 4.4A). The seasonal change was due to a rise in nitrogen in the plants [apices 19 – 44 mg (g dm)–1, middle leaf 12 – 36 mg (g dm)–1and lower leaf 7 – 25 mg (g dm)–1] and differences in carbon content [apices 403 – 455 mg

(g dm)–1, middle leaf 289 – 437 mg (g dm)–1and lower leaf 201 – 350 mg (g dm)–1]. The molar C/N ratio in P. perfoliatus ranged from 10 to 27, and was more constant throughout the season in leaves compared to the apices (Figure 4.4B). Phosphorus concentration was highest in M. spicatum apices [1.8 to 3.5 mg (g dm)–1] with a rise in autumn. Also in the middle and lower leaves phosphorus increased with season, and the content was slightly higher in lower leaves compared to middle leaves [0.8 ± 0.3 and 0.5 ± 0.2 mg (g dm)–1, respectively, mean ± SD; Figure 4.4C]. In P. perfoliatus, the phosphorus content was very similar in apices and leaves [0.5 to 1.2 mg (g dm)–1], with higher values in mid September and end of October in all plant parts

(Figure 4.4D). The chlorophyll content in P. perfoliatus was slightly higher than in

M. spicatum (Figure 4E and F), with a strong decrease from the beginning of

September until the end of the sampling campaign. The apices [6 ± 2 mg (g dm–1)] always contained less chlorophyll than the middle and lower leaves [7 ± 2 and

9 ± 3 mg (g dm–1), respectively]. In M. spicatum, chlorophyll content increased in all plant parts with season, and was higher in apices and middle leaves than in lower leaves [5 ± 1, 6 ± 2 and 4 ± 2 mg (g dm–1), respectively; Figure 4.4E].

80 IV Spatio-temporal biofilm dynamics

A) B) 35 35

30 30

25 25

20 20

C/N [molar ratio] [molar C/N 15 15

10 10 C) D) ] -1 4 4 [mg (g dm)

- 2 2 4- P-PO 0 0

] 15 15

-1 E) F)

12 12

9 9

6 6

3 3

Chlorophyll a & b [mg (g dm) 0 0 Aug Sep Oct Aug Sep Oct

Figure 4.4: Chemical parameters of Myriophyllum spicatum (left) and Potamogeton perfoliatus (right). Circles: Apex; triangle up: Middle leaf; triangle down: Lower leaf. n = 3, mean ± SD.

81 IV Spatio-temporal biofilm dynamics

Phenolic compounds. The total phenolic content was highest in M. spicatum apices

[200 – 250 mg (g dm)–1], followed by middle and lower leaves [67 – 138 and 50 –

70 mg (g dm)–1, respectively; Figure 4.5A]. Potamogeton perfoliatus contained only

21 ± 9 mg (g dm)–1 of total phenolics, and no difference occurred between apices and leaves (Figure 4.5B).

The anthocyanin content was higher in M. spicatum apices [1.5 ± 0.3 –

2 mg (g dm)-1] than in both leaf sections [0.6 ± 0.3 mg (g dm)–1; Figure 4.5C]. In

P. perfoliatus, the anthocyanin content was rather similar between all plant parts

[average 0.3 ± 0.07 mg (g dm)–1], and no seasonal variation was observed

(Figure 4.5D). Tellimagrandin II is the major hydrolysable polyphenol in M. spicatum and does not occur in P. perfoliatus. In the apices of M. spicatum, the tellimagrandin II content was higher [30 – 70 mg (g dm)–1] than in both leaf sections [2 –20 mg (g dm)–1;

Figure 4.5E].

Table 4.2. Environmental variables measured at the sampling dates. Water level data are of the water gauge at the harbour in Konstanz. Sampling Temperature Water level Conductivity No. sampling date pH Date [°C] [cm] [µS cm–1]

17. July 1 25.5 331 251 8.30 31. July 2 25.6 310 248 8.61 15. August 3 18.7 324 265 8.53 29. August 4 17.6 321 263 8.27 12. September 5 20.0 324 263 8.06 22. September 6 18.9 329 274 8.41 09. October 7 16.2 322 286 8.01 23. October 8 15.6 292 282 8.23

Impact of plant chemistry and environmental factors on community composition

We performed a BEST–ENV analysis to elucidate the major factors influencing bacterial community composition. The analysis indicated that of all plant chemistry parameters measured, only carbon and total phenolic content marginally explained the variation in bacterial community composition between the two plant species

(ρ = 0.175, P = 0.1, n = 42).

82 IV Spatio-temporal biofilm dynamics

A) B) 300 300 ] -1

200 200

100 100 TPC [mg (g dm)

0 0

] C) D) -1

2 2

1 1 Anthocyanin [mg (g dm) 0 0 ] E) -1 Aug Sep Oct 60

40

20

Tellimagrandin II [mg (g dm) 0 Aug Sep Oct

Figure 4.5: Concentration of different phenolic compounds in Myriophyllum spicatum (left) and Potamogeton perfoliatus (right). Circles: Apex; triangle up: Middle leaf; triangle down: Lower leaf. n = 3 mean ± SD. Tellimagrandin II or hydrolysable polyphenols are not present in P. perfoliatus.

83 IV Spatio-temporal biofilm dynamics

The environmental factors water level and conductivity explained most of the variability of the biofilm community composition (ρ = 0.33, P = 0.002, n = 42). If both plant chemistry and environmental variables were considered, carbon, total phenolic content, water temperature, water level, conductivity and pH were the major predictors, and the correlation coefficient increased (ρ = 0.354, P = 0.009). To compare the biofilm on the artificial substrates with that on the plants, we carried out the

BEST–ENV analysis only with environmental variables. Here, conductivity explained most of the variability in the biofilm community composition (ρ = 0.217, P = 0.002, n = 49).

Despite these differences, the overall bacterial community composition did not differ much depending on substrate, plant part or season as the NMDS analysis indicated (Figure 4.6). Slight changes occurred with season, especially at the beginning of the growing period.

Stress: 0.17 Apex 6 Leaf 10 cm Leaf 25 cm Artificial substrate 2

2 4 4 5 2 4 6 2 7 7 4 7 3 3 4 7 6 6 5 2 2 5 1 6 2 65 5 5 7 6 1 1 4 3 4 1 5 3 1 3 1 3 3

7

Figure 4.6. NMDS plot if FISH abundance data based on a Bray Curtis dissimilarity matrix. White: Myriophyllum spicatum Grey: Potamogeton perfoliatus. Sampling date numbers are listed in Table 4.2.

84 IV Spatio-temporal biofilm dynamics

DISCUSSION

Our study shows that the bacterial biofilm community composition on an artificial substrate and two common freshwater macrophytes was rather similar. All major bacterial groups determined by FISH were found on all substrates with varying abundance depending on time, plant species and plant age. Differences were especially pronounced between the apices of the plant species while older plant parts were more alike. The environmental conditions were rather stable (Hempel et al., submitted), but plant chemistry differed between both macrophytes. Still, based on a

BEST–ENV analysis, environmental factors were stronger predictors of the bacterial community composition than plant chemistry.

Initial colonization of axenic plants

We investigated the initial colonization of axenic Myriophyllum spicatum under more controlled conditions in a mesocosm. This approach enabled us to follow which bacterial groups would first colonize the plant surface and to see if these groups would be present at later stages. We assume that the initial colonization of axenic plants is comparable to that of apical shoots in the lake. The bacterial density increased very fast during the first five hours. The total bacterial cell counts were within the range found on plants in the lake and reached a plateau after 36 hours

(Figure 4.1A). Agar beads, used in other studies, are rapidly colonized by single strains within 30 - 50 min after exposure. Settling of bacteria was influenced by their motility and by the nutrient content of the agar beads, especially for tumbling bacteria (Kiorboe et al. 2002). Total cell counts on the agar beads were within the range of our findings.

The bacterial community composition developing on the axenic M. spicatum exhibited a distinct succession. Bacteria of the CFB group dominated the first 10 hours, later betaproteobacteria were most abundant. This succession pattern was also found on ceramic tiles (Olapade & Leff 2006), while on stainless steel alphaproteobacteria were the most abundant group after two hours of incubation

(Jones et al. 2007). Both groups are known to contain strains capable of degrading

85 IV Spatio-temporal biofilm dynamics

polymeric compounds, humic matter and polyphenols (Grossart et al. 2008). In the field, bacteria of the CFB group were more abundant than betaproteobacteria on the apices of M. spicatum during the summer, but both groups contributed always the highest proportions of the community composition. Maybe three days of incubation for the axenic shoots was too short to see a shift back towards members of the CFB group and a rise of alphaproteobacteria. The latter are known to thrive on labile organic matter compounds (Cottrell & Kirchman 2000, Olapade & Leff 2006) and may be more abundant in an established biofilm regulated by defined recycling processes. Similar to the field data, we observed only minor proportions of gammaproteobacteria, actinomycetes and planctomycetes.

Total cell counts

Submerged macrophytes greatly increase the attachment area for organisms in littoral habitats (Jeppesen et al. 1998). Due to distinct leaf morphologies, the relationship between plant surface and biomass varies substantially intra– and interspecifically (Sher-Kaul et al. 1995). We therefore decided to precisely measure each leaf analysed for bacterial biofilm composition by image analysis. Younger and older leaves have a distinct surface to biomass ratio that differed by a factor of two, especially in M. spicatum (Hempel et al. 2008).

Cell counts were highest on the artificial substrates, and rose on all substrates towards autumn. The higher cell counts on the plants might be a consequence of nutrient leaching from senescing plants. The higher cell counts on the propylene sheets are probably caused by the high settlement of Dreissena polymorpha and snail egg clutches after one month of exposure, which increased the settling area substantially. Further, zebramussels produce pseudofaeces and probably provide a high nutrient availability on the polypropylene sheets (Stewart et al. 1998). We can also not exclude that the surface itself exudes compounds that facilitate the attachment of organisms. Both macrophytes were much less colonized by

D. polymorpha and other organisms, and if, only at the stem sections close to the sediment.

86 IV Spatio-temporal biofilm dynamics

Bacterial densities on the macrophytes in 2006 were in the range we found in 2005

(1.3 – 1.7×106 cells cm–2; MH, unpubl. data). Compared to other studies, cell numbers on Lake Constance macrophytes and artificial substrates were rather low. On

Rorippa sp., young submerged leaflets had a bacterial density of about 17×106 cells cm–2 and old submerged plants up to 35×106 cells cm–2 (Hossell & Baker 1979). In

Lake Monchoon, Korea, Potamogeton crispus and cellulose films had cell numbers as high as 20×106 cells cm–2 (Hong et al. 1999). Artificial substrates in rivers had bacterial densities to about 2×106 cells cm–2 (Olapade & Leff 2006). The total bacterial cell counts were rather constant on both plants throughout the sampling period. Thus, the loss of bacterial cells by sloughing or grazing is compensated by the settling of bacteria or cell proliferation within the biofilm. In general, a high variation both in total cell counts and different bacterial groups, as we observed in our study, is not uncommon in heterogeneous habitats such as the littoral zone or plant surfaces

(Peters et al. 2007).

Higher bacterial numbers on M. spicatum in comparison to Potamogeton perfoliatus may have been caused by several factors. A higher surface to volume ratio of

M. spicatum leaves and the whorl–like leaf structure promote the settling of epiphytes, as was shown for epiphytic algae (Lalonde & Downing 1991). Whether this higher surface to volume ratio also enhances the release of nutrients and soluble organic carbon is unclear. The production and release of allelochemicals by these macrophytes might either negatively or positively affect bacterial epiphytes. The hydrolysable polyphenols produced by M. spicatum have antibacterial activity, although only at rather high concentrations (Scalbert 1991, Walenciak et al. 2002).

These tannins undergo fast photolysis and derivatisation when released (Bertilsson &

Tranvik 1998, Maie et al. 2008) and are also easily degradable by specialized bacteria, since both, the sugar moieties and the gallic acid from tellimagrandin II, are good substrates for a variety of microorganisms (Müller et al. 2007). Potamogeton natans produces various antialgal diterpenes (Cangiano et al. 2001, DellaGreca et al. 2001), but nothing is known about their release nor microbial degradation. Also other

87 IV Spatio-temporal biofilm dynamics

Potamogeton species contain algicidal and cyanobactericidal compounds (Wu et al.

2007a, Wu et al. 2007b). Whether antibacterial compounds present in P. perfoliatus

(Bushmann & Ailstock 2006) have an impact on total cell counts is questionable, since the extract concentrations were rather high.

Bacterial community composition

We selected the probes for different bacterial groups based on related studies in the field (Brummer et al. 2000, Schweitzer et al. 2001). In most of our samples, the sum of bacteria detected by all probes accounted for more than 100% of those detected by the EUB probe. To detect all eubacteria, others used a combination of different EUB probes, which are especially sensitive for planctomycetes (Daims et al.

1999). Since the numbers of planctomycetes in our study are very low, the use of these probes would probably not have yielded higher EUB counts. We applied a probe for Archaea (Arch915) on M. spicatum leaves but did only find a few scattered signals. Thus we concluded that Archaea probably do not contribute much to the total bacterial cell counts on M. spicatum. Also the variable numbers of CFB might be explained by the probe quality. Less than 50% of this group is detected by this probe, but other probes available detect even less (Loy et al. 2003).

Differences in the biofilm community composition were mainly caused by beta–, alpha– and gammaproteobacteria. The most dominant group on all substrates were bacteria belonging to the CFB group. This group is frequently found in biofilms in high abundance. Overall, the biofilm community composition on all substrates is similar to that found on lake snow particles in Lake Constance (Weiss et al. 1996,

Schweitzer et al. 2001). Alpha– and betaproteobacteria were most abundant on these particles, while gammaproteobacteria were less abundant. Only on hypolimnetic particles, CFB bacteria had the highest proportion of the bacterial community composition, apparently because of their capacity to degrade refractory material as chitin or cellulose. The high abundance of CFB on our substrates might also be due to complex organic compounds such as allelochemicals released by the plants or compounds recycled within the biofilm. CFB and betaproteobacteria are considered

88 IV Spatio-temporal biofilm dynamics to degrade high molecular weight dissolved organic matter (DOM, (Kirchman 2002)).

Surprisingly, we found high numbers of gammaproteobacteria on P. perfoliatus in autumn. Maybe the high nutrient availability at the end of the vegetation period, when plants are senescing and nutrient leakage is higher, is favorable for this bacterial group. Myriophyllum spicatum does not decline so early in the year and nutrient leakage in autumn might be less than in P. perfoliatus. Actually, nitrogen and phosphorus content increased in M. spicatum in autumn (Figure 4.4).

The biofilm developing on glass plates in two more or less polluted rivers had similar abundance of all bacterial groups analyzed except betaproteobacteria, which occurred more often in the polluted Spittelwasser (Brummer et al. 2000). Numbers of

CFB were rather low compared to our study, but the abundance of planctomycetes is in accordance with our data, although we did not find a seasonal trend for this group. The random occurrence of planctomycetes especially on M. spicatum remains enigmatic. In November 2006, we found almost no planctomycetes on this macrophyte in Lower Lake Constance, while they were present on the macroalga

Chara aspera (Hempel et al. 2008). Since planctomycetes were more variable on

M. spicatum apices, certain plant related factors might be responsible for this.

In general, M. spicatum displayed a higher spatio–temporal variability of the biofilm community composition than P. perfoliatus, especially of alphaproteobacteria.

Myriophyllum spicatum also exhibited pronounced chemical gradients, especially of phenolic compounds, nitrogen and phosphorus, while P. perfoliatus did not display such a spatial or temporal heterogeneity (Hempel et al, submitted). Terrestrial plant apices often contain higher amounts of anthocyanins than older leaves (Gould 2004).

This is also the case in M. spicatum. The spatial differences of the bacterial community composition are also confirmed by DGGE analysis of the same dataset (Hempel et al. submitted), where especially M. spicatum apices differed to those of P. perfoliatus and the artificial substrate. Betaproteobacteria were more abundant in the biofilm on

P. perfoliatus, especially towards autumn. One reason for this might be the earlier senescence of this macrophyte. Betaproteobacteria have been shown to degrade

89 IV Spatio-temporal biofilm dynamics

proteins and amino acids, compounds that should be readily available on senescent plants (Cottrell & Kirchman 2000). But nevertheless, this phylogenetic group is very diverse concerning ecology and degradation capabilities. Thus, it is not adequate to speculate on ecological functions. A related DGGE study on these biofilms proved, that there are indeed betaproteobacteria, e.g., Ralstonia sp., which are capable of polyphenol degradation (Hempel et al., submitted).

The NMDS analysis did not show distinct differences of the biofilm community composition on the three substrates. Nevertheless, we saw slight seasonal community changes that we also observed for alphaproteobacteria and betaproteobacteria. Sampling dates around autumn cluster more closely than sampling dates in summer, which might reflect seasonal differences in plant chemistry and senescence of both macrophytes.

Our analysis of the three data sets environmental variables, plant chemistry and bacterial community composition by BEST–ENV showed that water level and conductivity were the strongest predictors of the biofilm community composition.

Tissue carbon and, in contrast to our predictions, total phenolic content in plants did not explain much of the variation, but together with all environmental variables had the highest correlation coefficient. These findings are in accordance with other studies dealing with the impact of environmental factors on bacterial communities.

In lakes from northern Europe, pH, conductivity and temperature had an impact on the bacterial community composition (Lindstrom et al. 2005, Allgaier & Grossart

2006). Of course, the environmental conditions at the microscale within the biofilm still may vary between the substrates, but such measurements were beyond the scope of this study.

Unfortunately, only few studies have dealt so far with the heterotrophic biofilm on freshwater submerged macrophytes. This is a pity since epiphytic biofilm community composition might be important for the overall nutrient cycling in the littoral zone and thus might affect the plants and associated organisms (Costerton et al. 1995, Ardon & Pringle 2007).

90 IV Spatio-temporal biofilm dynamics

In comparison to our DGGE analysis of this data set, differences in bacterial community composition analysed by FISH between the three substrates were not as pronounced. With DGGE, community differences were better detected due to a higher resolution of this method. Using only probes for major bacterial groups, we might have overlooked changes in subgroups with distinct physiological capacities

(Schweitzer et al. 2001). Further studies should involve more specific probes and the detection of active cells by e.g., MICRO–FISH.

The small differences in bacterial community composition on all parts of both plants might indicate established biofilms that depend less on the substrate than on internal recycling processes and external nutrient supply from the water column. Of course, the probes selected do not give a detailed insight into the biofilms, and thus we might have missed changes within the groups. Apparently, a constant bacterial density on both plants is reached within few hours and no distinct changes in major bacterial groups occur past this time, as seen with axenic M. spicatum. Axenic shoots of P. perfoliatus have not been available for initial colonization experiments to see plant specific differences and especially to test whether the developing biofilm depends more on plant chemistry.

Acknowledgements. This work was supported by the German Science Foundation with grant CRC454, project A2 to EMG. We would like to thank C. Feldbaum and S. Nadj for technical assistance and J. Hesselschwerdt and S. Werner for help with sampling. We appreciate S. Wicks for conduction the initial colonization experiment. S. Hilt indicated suitable artificial substrate material. G. Heine adapted the digital imaging system ‘Makrophyt’. H.–P. Grossart gave helpful comments on the manuscript.

91 IV Spatio-temporal biofilm dynamics

92

Chapter V

Impact of the polyphenol degrading bacterium Matsuebacter sp. FB25 on the growth of Acentria ephemerella larvae

Melanie Hempel & Elisabeth M. Gross

102 V Acentria ephemerella and Matsuebacter sp.

94 V Acentria ephemerella and Matsuebacter sp.

ABSTRACT: The larvae of the aquatic pyralid moth Acentria ephemerella (DENIS & SCHIFFERMÜLLER) cause severe feeding damage on macrophytes in Lake Constance, esp. on Potamogeton spp. and Myriophyllum spicatum L. Earlier studies showed a better growth performance on tannin poor pondweeds than on tannin rich M. spicatum. In two independent experiments, we tested if the tannin degrading betaproteobacterium Matsuebacter sp. would enhance the growth of A. ephemerella larvae by detoxifying the plant secondary metabolites and by providing further carbon and nitrogen sources. First instar larvae were offered three differently treated M. spicatum plants: axenic M. spicatum (AX), axenic M. spicatum colonized with Matsuebacter sp. (MATS) and mesocosm M. spicatum (MESO). Fresh food was supplied every three to four days ad libidum when headcapsule width of the larvae was measured. The daily growth rates of larvae were 9 – 13 µm day-1 on axenic, 7 – 15 µm day-1 on Matsuebacter sp.-colonized plants and 5 – 12 µm day-1 on mesocosm plants. The mortality of larvae was highest in the MESO set-up (80%), while the other two set-ups had mortalities from 35 - 52% in both experiments. Treatments with AX and MATS had similar carbon and nitrogen contents [388 ± 6 mg C (g dm)-1 and 43 ± 2 mg N (g dm)-1] while the MESO treatment contained 228 ± 17 mg carbon (g dm)-1 and 12 ± 3 mg nitrogen (g dm)-1. Since no growth difference was found between axenic and Matsuebacter sp.-colonized plants, the bacteria were neither an additional nutrient source for the larvae, nor did they affect the larval growth by possibly modifying polyphenols. Thus, one single bacterial strain will most likely not effectively influence larval growth.

Keywords: Myriophyllum spicatum, herbivores, Lepidoptera, betaproteobacteria, biofilm.

95 V Acentria ephemerella and Matsuebacter sp.

INTRODUCTION

Bacteria are found in almost every environment where they cover the range from commensals to pathogens. In many organisms, the majority of bacteria is not pathogenic but rather commensal. Especially important are bacteria in the intestines.

In t has been shown for a variety of habitats that gut microbiota are essential for the successful processing of forage or to synthesize specific amino acids or sterols (Dillon

& Dillon 2004). Usually, the gut bacteria are adapted to certain environments and food sources. But not all forage can be equally well digested. Some diets contain plant secondary compounds that hamper the digestion by inhibiting the gut microbiota. Well known compounds in nature with this effect are tannins. They chelate iron, complex proteins and nutrients (Scalbert 1991), and decrease the nitrogen availability in ruminants (McSweeney, 1999).

Tannins are well recognized as plant allelochemicals in terrestrial and aquatic ecosystems. The hydrolysable tannins of the macrophyte Myriophyllum spicatum L.

(Haloragaceae) inhibit the photosystem II of cyanobacteria (Leu et al. 2002), and larvae of the aquatic moth Acentria ephemerella larvae (Pyralidae, DENIS &

SCHIFFERMÜLLER) grow slower on M. spicatum than on the pondweed Potamogeton perfoliatus, which does not contain hydrolysable tannins (Choi et al. 2002). Further, gut bacteria of A. ephemerella were inhibited by tannins from M. spicatum (Walenciak et al. 2002). Although in the past, herbivory on aquatic macrophytes was considered rare (Shelford 1918, Gregory 1983), high abundance of herbivores may have a substantial impact on macrophytes (Newman 1991). In Lake Constance,

A. ephemerella larvae cause substantial damage to apical meristems of M. spicatum and

P. perfoliatus. Larvae are usually found in densities higher than 0.8 individuals per shoot, which is considered to cause a decline of M. spicatum (Painter & McCabe 1988,

Gross et al. 2002).

While tannin rich plants are considered to be well protected against herbivory

(Feeny 1970), a tannin rich forage does not have to be detrimental. Rats fed with a diet rich in condensed tannins, select towards gut microbiota capable to degrade

96 V Acentria ephemerella and Matsuebacter sp. tannins (Smith & Mackie 2004). Further, tannins in green tea, for example, act as prooxidants in yeasts because they reduce reactive oxygen species (Maeta et al. 2007).

Besides plant allelochemicals, plant stoichiometry is important for insect growth and food choice. Aquatic herbivores consuming elementally imbalanced food have a diminished conversion efficiency of ingested carbon into new biomass, and the gross growth efficiency is reduced (Elser et al. 2000). Food plants containing low amounts of phosphorus also impair growth and reproduction (Urabe & Watanabe 1992).

Lepidoptera prefer plants with the highest nitrogen content, which is also one of the major determinants of food quality (Newman 1991, White 1993).

Schultz and others (Schultz et al. 1992) assume that negative impacts of plant polyphenols on herbivores may be nothing more than a by-product of plant-microbe interactions. Thus, we asked whether A. ephemerella growth on tannin rich

M. spicatum is promoted if the plant is colonized with a bacterial strain that degrades tannins. In 2005, Matsuebacter sp. was isolated in enrichment experiments from the surrounding water of M. spicatum. This betaproteobacterium is able to degrade tannic and gallic acid constitutively (Müller et al. 2007). To investigate the effect of this bacterium on the growth of freshly hatched A. ephemerella larvae, we also offered axenic (= bacteria-free) M. spicatum and M. spicatum grown in an outdoor mesocosm, which were colonized with a natural biofilm.

97 V Acentria ephemerella and Matsuebacter sp.

MATERIALS & METHODS

Larvae. Two egg clutches with 90 – 200 eggs were collected in July 2007 on

Potamogeton perfoliatus. These egg clutches were reared in sterile filtered lake water

(0.2 µm) at 20 ± 2 °C in with permanent aeration. After hatching, larvae were fed for seven days with axenic M. spicatum to avoid further bacterial contamination and to accustom the larvae to the food supplied in the assay. We performed one feeding experiment with each of the egg clutches.

Food sources/plant material. Three differently treated kinds of M. spicatum were offered during the experiment resulting in the three set-ups: axenic plants (AX), axenic plants colonized with the tannin degrading bacterium Matsuebacter sp.

(MATS) and mesocosm M. spicatum (MESO). Axenic plants were cultured as described in (Gross 2003) and supplied with fresh medium every two weeks to ensure the plants fed were of the same nutritious state. Myriophyllum spicatum was readily colonized within 48 h by an actively growing liquid culture of Matsuebacter sp. that was diluted to an optical density (OD600nm) of 0.2 in 100 ml sterile artificial tap water (DIN EN ISO 7346-3:1998-03). Previous studies proved that 48 h are sufficient to ensure colonization of the plant (Wicks, Hempel and Gross, submitted for publication). We used mesocosm M. spicatum to monitor the larval growth on plants with a natural biofilm. The mesocosm is a 2×2×1 m concrete basin in the yard of the

Limnological Institute filled with Lake Constance sediment and water (flow rate ca.

20 l h-1), and densely planted with M. spicatum. The plants were exposed to the same environmental conditions as in the field and thus should be comparable to field plants. Mesocosm plants were collected 48 h before they were fed to the larvae. To provide comparable conditions for all plants compared to Matsuebacter sp.–colonized

M. spicatum, axenic and mesocosm plants were also incubated for 48 h in sterile artificial tap water prior to feeding.

Bacterial cultures. Matsuebacter sp. cultures were grown at 20 ± 2 C° in medium B

(Müller et al, 2007) without yeast and tryptone but with 3 mM succinate for permanent cultures, and 5 mM succinate for experiments. The bacteria were

98 V Acentria ephemerella and Matsuebacter sp. harvested after 24 h hours in the exponential growth phase. The cells were washed twice with artificial tap water to remove Medium B and the cell suspension was adjusted to an OD600 of 0.2.

Feeding assays. Feeding assays were performed in sterile cell culture dishes. Each dish contained 4 ml artificial tap water, 3 – 4 leaves of the respective plant material and one larva. Larvae were distributed randomly to the three set-ups. Both experiments started when the larvae were in average seven days old and ended after

21 days. Every three to four days, the headcapsule width of the larvae was measured under the stereomicroscope with digital imaging, and fresh food and water were supplied in a new, sterile dish. Food was always supplied ad libidum. The experiments were conducted in a Sanyo MLR 350 environmental test chamber

(SANYO Electric Biomedical Co., Ltd., Japan) at 20 °C and a light regime of 16 hours light (level 3; equal to 100 µmol photons m-2 s-1) and 8 hours darkness.

Chemistry. At each feeding day, a subsample of the plants of each treatment were analysed for total phenolic content, carbon and nitrogen (Choi et al. 2002). At the end of the experiments, all larvae were frozen and the intestines removed. The gut was transferred into sodium pyrophosphate (see below). The larval bodies and heads were dried at 60 °C and each 6 - 9 larvae were pooled and subjected to carbon and nitrogen measurements to obtain the nutrients accumulated during the experiments.

We removed the intestines, because we only wanted to analyze the accumulated nutrients in the body and no those present in the gut, which would reflect recent feeding rather than real nutrient uptake. To remove the intestines, the larval head was fixed with tweezers and the body was cut off behind the headcapsule with a scalpel without cutting off the intestines from the head. The head with the intestines was pulled out of the body, and the intestines cut off. Due to the small size of the larvae, all dissections were performed under a stereomicroscope.

Bacterial counts. At the end of the experiment, three guts of each treatment were resuspended in 0.9 ml 0.1 M sodium pyrophosphate (Na4P2O7×10 H2O) and fixed with 0.1 ml 37% formaldehyde for bacterial enumeration. The guts were transferred

99 V Acentria ephemerella and Matsuebacter sp. to an ultrasonic bath (Laboson 200 ultrasonic bath, Bender & Hobein) for 60 s, shaken for 15 min (1100 rpm, Eppendorf Theromixer, 20 °C) and again exposed to 60 s ultrasonication (Walenciak 2004). The suspension was filtered onto polycarbonate filters (0.2 µm) and bacteria were stained with DAPI (4ʹ,6–Diamidino–2–phenylindol,

1 µg ml-1, 5 min). Stained cells were counted under an epifluorescence microscope

(Labophot 2, Nikon) at an excitation wavelength of 549 nm. The bacterial cell counts were related to gut volume. We estimated the gut volume with a correlation for headcapsule width to gut length (100 µm headcapsule width equals 1.6 mm gut length). A mean gut diameter of one millimetre was assumed (Walenciak 2004).

Statistical analysis. We compared the different treatments with one–way

ANOVAs. If the normality test failed despite transformation of the data, we used

ANOVA on ranks with the Tukey post hoc test or Dunn´s method. The comparison of the food quality between the different experimental days was also performed with one–way ANOVA.

RESULTS

Experiment I. At the beginning of experiment 1 in July 2007 the headcapsule width of the 78 larvae was 173 ± 5 µm. The larvae were distributed to the three different treatments (Ax n = 27, MATS n = 26, MESO n = 25). In the set-up with Matsuebacter sp.–colonized plants and axenic plants, the larvae grew better than with mesocosm plants (Figure 1A). The average headcapsule growth rates were 15 ± 3 µm day-1

(mean ± 1SD for all data given), 13 ± 3 µm day-1 and 12 ± 2 µm day-1, respectively and did not differ significantly (P = 0.67). At the end of experiment I, the larvae in treatments MATS and AX had a headcapsule width of 496 ± 44 µm (n = 17) and

455 ± 67 µm (n = 16), respectively, while larvae fed with mesocosm plants had a headcapsule width of 448 ± 46 µm (n = 5; Figure 5.1A).

During the 21 days, several larvae died. The highest mortality was observed in the mesocosm treatment. Here, 80% of the larvae died during the experiment, while in the AX and MATS treatments 40% and 35% of the larvae died, respectively.

100 V Acentria ephemerella and Matsuebacter sp.

600 A) 500

400

300

200

100

600 B)

500

400 Headcapsuslse width [µm]

300

200 AX MATS 100 MESO

5 1015202530

Figure 5.1. Growth of Acentria ephemerella larvae on differently treated Myriophyllum spicatum leaves. A) Feeding experiment I B) Feeding experiment II with three differently treated M. spicatum. AX axenic M. spicatum, MATS M. spicatum colonized with Matsuebacter sp., MESO mesocosm M. spicatum. Mean ± 1SD.

The total bacterial cell counts in the larval gut as determined by DAPI ranged from

4 ± 3×106 cells (mm gut volume)-3 to 8 ± 3×106 cells (mm gut volume)-3 with no distinct differences for the set–ups (Figure 5.2A, one–way ANOVA, F = 1.105, P = 0.39).

The carbon/nitrogen molar ratio (C/N) was 11 ± 1 and 10 ± 0.3 for axenic and

Matsuebacter sp.–colonized plants, respectively (Table 5.2). Mesocosm plants had a

C/N ratio of 21 ± 3 (Table 5.2). The total phenolic content was 60 ± 5 mg (g dm)-1 in axenic, 64 ± 5 mg (g dm)-1 in Matsuebacter sp.–colonized and 35 ± 11 mg (g dm)-1 in

101 mesocosm plants 12 A) throughout the experiment

10 (Table 5.2). The

8 phosphorus content was 9 ± 1 mg (g dm)-1 and 9 ± 2 gut volume gut 6 -1 -3 mg (g dm) and in axenic

4 and colonized plants, ) mm 6 2 respectively, while mesocosm plants had only 0 B) 1.5 ± 0.2 mg (g dm)-1 tissue 12 phosphorus content. The

10 food sources within each 8 treatment did not differ

6 between the different feeding days (ANOVA on total bacterial cell counts (x10 4 ranks, P values were 2 always > 0.6 for all

0 treatments). In general, AX MATS MESO axenic and Matsuebacter Figure 5.2. Total bacterial cell counts in the gut of differently fed Acentria ephemerella larvae at the end sp.-colonized plants had a of the experiment. A) Feeding experiment I B) Feeding experiment II. AX axenic Myriophyllum spicatum, similar plant chemistry MATS Matsuebacter sp. colonized-M. spicatum, MESO mesocosm M. spicatum. n = 3, mean ± 1SD compared to mesocosm plants (Table 5.2).

The larval bodies accumulated 462 – 511 mg C (g dm)-1 and 85 – 96 mg N (g dm)-1 during the experiment with axenic and Matsuebacter sp.-colonized plants, respectively. For each treatment only two measurements were possible since at least five larvae had to be pooled for one assay. Larvae fed mesocosm plants accumulated

102 V Acentria ephemerella and Matsuebacter sp.

242 mg C (g dm)-1 and 85 mg N (g dm)-1 in 21 days (Table 5.4). Due to the high mortality in some of the treatments, a higher replication was not possible.

Experiment II. In the second experiment, 132 larvae were distributed to the different treatments (AX n = 45, MATS n = 43, MESO n = 44) and had an averageheadcapsule width of 325 ± 27 mm. The daily growth rate was lower than in the first experiment.

Larvae fed axenic M. spicatum grew 9 ± 3 mm day-1, larvae fed Matsuebacter sp.– colonized plants 7 ± 3 mm day-1 and larvae on mesocosm plants grew 5 ± 3 mm day-1.

The daily growth rate of larvae fed axenic M. spicatum was significantly higher, than those fed mesocosm plants (ANOVA on ranks, P = 0.006). In this experiment, larvae fed axenic M. spicatum and Matsuebacter sp.–colonized M. spicatum grew 488 ± 58 µm and 466 ± 57 µm, respectively. The larvae grown on mesocosm plants had a final headcapsule width of 433 ± 52 µm (Figure 5.1B).

The mortality was 52% for larvae on axenic plants (n = 22), 35% for larvae on

Matsuebacter sp.–colonized plants (n = 28) and 75% for larvae reared on mesocosm plants (n = 11).

Table 5.1. Statistical analysis for the comparison between the headcapsule widths during the experiment in the three set-ups by one–way ANOVA with Holm-Sidak post hoc test. On all other sampling days, but those displayed, larval headcapsule width was not significantly different. Ax: axenic Myriophyllum spicatum, MATS: Matsuebacter sp.–colonized M. spicatum, MESO: mesocosm M. spicatum Experiment 1 MESO AX Sampling day 4 a) MATS <0.05 <0.05 AX >0.05 Sampling day 5 MATS 0.001 0.137 AX 0.021 Sampling day 6 MATS 0.001 0.2 AX 0.011 Experiment 2 MESO AX Sampling day 4 MATS <0.0001 0.189 AX <0.001 a) ANOVA on ranks with Dunn´s method was used

103 V Acentria ephemerella and Matsuebacter sp.

Table 5.2. Chemical parameters of the differently treated Myriophyllum spicatum in Experiment I. Statistical analysis of data was performed with one-way ANOVA with α < 0.05, except for phosphorus and C/N, which were analyzed by ANOVA on ranks with a Tukey post hoc test. Statistically significant different groups after post-hoc tests are indicated by different groups. AX axenic Myriophyllum spicatum MATS Matsuebacter sp.-colonized M. spicatum MESO mesocosm M. spicatum

Sampling Time since Tissue carbon content Tissue nitrogen content C/N

Day hatching [d] [mg (g dm-1)] [mg (g dm-1)] [molar ratio]

AX MATS MESO AX MATS MESO AX MATS MESO 1 7 382.60 382.48 220.81 43.75 44.53 12.87 10.20 10.02 20.02 2 11 384.91 383.00 257.84 45.65 43.69 18.09 9.84 10.23 16.63 3 14 371.57 388.74 219.39 43.36 42.91 12.39 10.00 10.57 20.65 4 18 385.18 393.22 219.39 42.01 41.85 12.55 10.70 10.96 20.39 5 21 395.14 397.19 216.15 38.58 43.82 9.85 11.95 10.57 25.59 6 25 397.41 388.55 251.57 40.96 44.67 13.78 11.32 10.15 21.29 group a a b a a b a a b

Phosphorus content Total phenolic content Tellimagrandin II

[mg (g dm-1)] [mg (g dm-1)] [mg (g dm-1)] AX MATS MESO AX MATS MESO AX MATS MESO 1 7 9.86 10.02 1.47 55.00 65.00 36.50 4.96 6.14 2.35 2 11 10.27 11.12 1.84 58.00 58.25 51.25 5.06 4.89 3.31 3 14 9.86 9.87 1.31 55.50 72.00 33.00 4.35 6.12 2.11 4 18 9.86 8.20 1.58 59.50 65.00 32.00 5.77 5.08 1.79 5 21 6.63 6.32 1.36 66.50 62.00 32.00 5.54 4.42 1.96 6 25 8.71 10.23 1.50 63.00 63.50 42.50 6.41 5.13 3.31 group a a b a a b a a b

104 V Acentria ephemerella and Matsuebacter sp.

Table 3 Chemical parameters of the differently treated Myriophyllum spicatum in Experiment II. Statistical analysis of data was performed with one-way ANOVA and α < 0.05, except for carbon and C/N, which were analyzed by ANOVA on ranks with a Tukey post hoc test. Statistically significant different groups after post-hoc tests are indicated by different groups. AX axenic Myriophyllum spicatum MATS Matsuebacter sp.-colonized M. spicatum MESO mesocosm M. spicatum Sampling Time since Tissue carbon content Tissue nitrogen content C/N

Day hatching [d] [mg (g dm-1)] [mg (g dm-1)] [molar ratio] AX MATS MESO AX MATS MESO AX MATS MESO 1 7 379.64 391.08 218.74 42.77 41.87 12.49 10.36 10.90 20.43 2 11 386.88 391.32 220.61 44.30 41.23 11.58 10.19 11.07 22.23 3 14 390.77 392.47 245.73 42.53 42.97 12.58 10.72 10.66 22.79 4 18 391.78 389.81 211.60 43.62 46.75 9.35 10.48 9.73 26.40 5 21 387.51 393.80 227.74 47.29 44.99 9.49 9.56 10.21 28.01 group a ab a a a b a ab b

Phosphorus content Total phenolic content Tellimagrandin II

[mg (g dm-1)] [mg (g dm-1)] [mg (g dm-1)] AX MATS MESO AX MATS MESO AX MATS MESO 1 7 10.73 9.15 0.96 58.50 64.50 32.50 5.90 5.01 1.63 2 11 8.65 8.91 0.82 53.50 59.50 31.00 3.82 5.01 1.17 3 14 9.73 8.68 1.47 62.50 63.00 44.50 5.56 4.40 3.53 4 18 9.20 8.86 1.04 71.00 58.50 28.00 5.59 5.07 2.19 5 21 6.82 11.29 0.88 58.50 61.50 32.50 4.11 5.38 1.70 group a a b a a b a a b

105 V Acentria ephemerella and Matsuebacter sp.

The total bacterial cell counts were slightly lower than in the first experiment.

Larvae fed with mesocosm plants had a slightly higher bacterial cell counts in the gut

[7 ± 1×106cells (mm gut volume)-3] than the larval guts of the other two treatments, which were rather identical [Figure 5.2B, AX 5 ± 1 ×106cells (mm gut volume)-3, MATS

5 ± 0.8 ×106cells (mm gut volume)-3, one–way ANOVA, F = 2.309, P = 0.095].

The plant chemistry between both experiments was quite similar. In the second experiment the C/N ratio of axenic and Matsuebacter sp.–colonized plants was 10 ± 0.4 and 11 ± 0.5 respectively, and 24 ± 3 in mesocosm plants (Table 5.3). The total phenolic content during the experiment did not vary much (Table 5.3) with

61 ± 7 mg (g dm)-1 for axenic, 62 ± 3 mg (g dm)-1 for Matsuebacter sp.–colonized and

34 ± 6 mg (g dm)-1 for mesocosm plants. The phosphorus content was similar in axenic and Matsuebacter sp.–colonized plants [9 ± 2 and 9 ± 1 mg (g dm)-1, respectively] and 1 ± 0.3 mg (g dm)-1 in mesocosm plants. The food sources did not differ between the different feeding time points (ANOVA on ranks, P values were always > 0.6). In this experiment, too, the set-ups AX and MATS had rather identical plant chemistry (Table 5.3).

Carbon and nitrogen content in the larval bodies were assessed after the experiment ended at day 21. In this experiment, the larvae accumulated 486 – 511 mg

C (g dm)-1 and 88 – 97 mg N (g dm)-1 in the set-ups AX and MATS. Larvae fed mesocosm plants had a nitrogen content of 110 mg (g dm)-1 and a carbon content of

467 mg (g dm)-1 (Table 5.4).

106 V Acentria ephemerella and Matsuebacter sp.

Table 5.4. Carbon and nitrogen content in larval bodies after 21 days on different diets. Six to nine larvae were pooled for each replicate. AX axenic Myriophyllum spicatum, MATS Matsuebacter sp.- colonized M. spicatum, MESO mesocosm M. spicatum, na no larvae were available for measurements AX MATS MESO a) Replicate 1 2 3 1 2 3 1 Experiment I C [mg (g dm)-1] 462.44 491.25 na 510.97 501.51 na 241.79 N [mg (g dm)-1] 84.81 85.58 na 95.84 92.29 na 85.15 C/N [molar ratio] 6.36 6.7 na 6.22 6.34 na 3.32 Experiment II C [mg (g dm)-1] 485.52 517.43 502.42 511.03 501.22 507.07 466.77 N [mg (g dm)-1] 90.66 87.70 95.30 97.19 88.61 89.67 110.86 C/N [molar ratio] 6.25 6.88 6.15 6.13 6.60 6.60 4.91 a) Due to the high mortality in this treatment, no further replicates were possible.

DISCUSSION

We compared the growth performance of Acentria ephemerella larvae on differently treated Myriophyllum spicatum plants and were interested, if the tannin–degrading bacterium Matsuebacter sp. would enhance the growth of the larvae, which usually grow slower on tannin rich M. spicatum than on tannin poor Potamogeton perfoliatus

(Choi et al. 2002).

The high mortality we observed in the treatment with mesocosm plants is probably due to the leaf toughness. The mesocosm plants were very stiff and covered with lime. Usually, larvae in this treatment did not show signs of food consumption

(faecal pellets, coloured gut) in contrast to larvae fed axenic or Matsuebacter sp.- colonized plants. Those leaves were light green and soft, the larval guts were green coloured and we found faecal pellets frequently. Thus, we assume that larvae could not mechanically handle mesocosm plants due the high ash content. In another growth experiment, we offered lake-M. spicatum and it was well used by the larvae

(data not shown). We further assessed the ash free dry mass (AFDM) of field and mesocosm M. spicatum and found lower AFDM values in the leaves of mesocosm plants (79% ± 10% and 53% ± 15%, respectively, Student’s t-test, P < 0.001), indicating a higher ash content. We assume that the higher anorganic content in mesocosm plants was caused by high densities of Lymnea stagnalis that dwell in the mesocosm

107 V Acentria ephemerella and Matsuebacter sp. without natural enemies. Probably the leaf toughness is an adaption to increased grazing pressure.

That the larval growth was rather identical on axenic and Matsuebacter sp.- colonized M. spicatum may be related to the overall better nutrient content of the plants and the lower ash content in comparison to mesocosms plants. The M. spicatum plants for those set ups originated from a non limiting plant medium and the bacteria were obviously not an additional carbon or nitrogen source (Table 5.2 and 5.3). Since Lepidoptera are known to prefer food with a high nitrogen content

(Newman 1991, White 1993), the higher larval growth rate on both axenic and

Matsuebacter sp.-colonized plants could also be explained by the twice as high C/N ratio of mesocosm plants (Table 5.2 and 5.3). Further, mesocosm plants had a lower phosphorus content. The growth of mayflies and Manduca sexta larvae was also negatively affected at high C:P ratios (Frost & Elser 2002, Perkins et al. 2004). Thus, larvae that were able to feed on mesocosm M. spicatum had less nutritious food than in the other two assays and might even be P-limited. These data are also confirmed by the carbon and nitrogen content that accumulated in the larval bodies. In the AX and MATS treatments of experiment I, the carbon content of larval bodies was twice as high as in larvae fed the mesocosm plants. The nitrogen content in the larval bodies was similar between the three treatments with slightly higher values in larvae fed Matsuebacter sp.-colonized plants (Table 5.4). This can be explained by homeostasis of the larvae. Despite abundant food availability, invertebrates might grow slower because of nutrient deficiency of their food that has to be compensated

(Sterner & Hessen 1994). Unfortunately due to the high mortality and necessity to pool many larvae for one C/N measurement, we do not have enough replicates to perform statistical analysis and confirm these data. We also had not enough larvae to perform analyses of larval P content. Thus the C/N values can only be used as indicators for possible limitations and storage mechanisms. In the second experiment the carbon content was rather similar between all treatments while the accumulated nitrogen in the larval bodies was rather similar in all food sources.

108 V Acentria ephemerella and Matsuebacter sp.

Besides measuring the headcapsule width and nutrient content of plants and larvae, we also assessed the bacterial numbers in the gut. Interestingly, the bacterial numbers in the gut were equal between different set ups. This is further evidence that bacteria are inherited in early larval stages and also that bacteria pass rather quickly through the gut (Dillon & Dillon 2004). We expected that larvae fed mesocosm and Matsuebacter sp.-colonized plants would inherit more bacteria than those fed axenic plants. One reason for equal bacterial numbers could arise from the rearing of the larvae. As long as larvae were hatching, the Potamogeton perfoliatus leaf, which had a natural bacterial biofilm, and on which the egg clutch was found, was left in the Erlenmeyer flask and only removed if all larvae were hatched. Usually, hatching took three days in which also axenic M. spicatum was offered, but newly hatched larvae often preferred P. perfoliatus. It is unclear yet, how the biofilm

Matsuebacter strains would survive or establish in the larval gut together with other bacteria. Earlier studies showed that Lepidoptera have a transient gut microbiota

(Appel 1994). Thus, it is possible that Matsuebacter sp. was excreted rather fast, and it seems not surprising to find equal bacterial numbers in the guts of differently fed larvae.

We conclude that the growth of herbivores on tannin rich forage was not influenced by a single tannin-degrading bacterium alone. In contrast, the anorganic content of the food plant had a larger influence on the larval growth and accessibility of the food. In these experiments, A. ephemerella larval growth was not necessarily slowed down by a tannin–rich diet but because of the high ash fraction of the diet, and the necessity to store nutrients at the cost of a reduced growth. To elucidate the impact of tannins and bacteria on the growth performance of A. ephemerella larvae, the use of artificial diets supplemented with tannins and selected bacterial strains would be a possible solution. In marine systems, agar based diets have been successfully fed to sea urchins or amphipods (Hay et al. 1994), but so far, all attempts to feed artificial diets to A. ephemerella failed (Choi et al. 2002). This leaves at present only experimental approaches as we used them in this set-up.

109 V Acentria ephemerella and Matsuebacter sp.

Acknowledgement: We thank O. Miler for help with larval egg clutches, C. Feldbaum for technical assistance with plant chemistry and B. Mothes for help with measuring and feeding of the larvae.

110 V Acentria ephemerella and Matsuebacter sp.

111 V Acentria ephemerella and Matsuebacter sp.

112

Chapter VI

Single- and multispecies biofilm formation of tannin degrading bacteria on an aquatic macrophyte

Sonja Wicks, Melanie Hempel & Elisabeth M. Gross

VI Colonization of M. spicatum by tannin-degrading bacteria

114 VI Colonization of M. spicatum by tannin-degrading bacteria

ABSTRACT: We developed a microcosm set-up to study the distribution of single and mixed cultures between the planktonic or attached life-style of three proteobacterial strains. The three isolates were capable to metabolise hydrolysable polyphenols produced by the submerged macrophyte Myriophyllum spicatum. In single culture, Matsuebacter sp., capable to constitutively degrade polyphenols, reached the highest biofilm density with up to 80×106 cells cm-2 and declined in the surrounding water. The abundance of Agrobacterium vitis increased both in the water and on the leaves (18×106 cells cm-2). The density of Pantoea agglomerans stayed constant in the water column and only marginally increased on the leaves to a maximum of 6×106 cells cm-2. In mixed cultures, we determined the relative proportion of each strain by fluorescence in situ hybridisation. We show strong interactions and different colonization patterns depending on the different combination of isolates. Matsuebacter sp. has a high colonization potential in single species biofilms but not in combination with other species. If P. agglomerans was cultivated with A. vitis, it dominated the biofilm, but biofilm formation was less if cultivated with Matsuebacter sp. When all three strains were co-cultivated, A. vitis was dominant over both Matsuebacter sp. and P. agglomerans. We show that the terrestrial bio-control agent P. agglomerans may prevent a high density of the plant pathogen A. vitis also in aquatic systems. When Matsuebacter sp. is added, A. vitis regains dominance over P. agglomerans. Our system allows a close investigation of the reciprocal impact of plants and attached microorganisms and among biofilm bacteria.

Keywords: Myriophyllum spicatum, Matsuebacter sp., Pantoea agglomerans, Agrobacterium vitis, FISH

115 VI Colonization of M. spicatum by tannin-degrading bacteria

INTRODUCTION Especially under low light conditions, bacteria are the first to colonize any surface submerged in water, and may facilitate the further attachment of algae and other organisms (Roeselers et al., 2006). Within the biofilm, heterotrophic and autotrophic organisms depend on each other for the production, degradation or recycling of mineral nutrients and organic compounds (Romani et al., 2004). Information on the bacterial community composition, structure and succession on different abiotic and biotic substrates in aquatic systems is increasing and mainly based on culture- independent methods (DGGE, FISH, clone libraries) and to some extend on culture- based approaches, e.g., the isolation of strains with certain metabolic capacities.

These ‘microbial landscapes’ now require explicit ecological theories to explain their spatial structure and function (Battin et al., 2007).

Bacteria colonizing living organisms, e.g., plants, may be subjected to both nutritional and detrimental compounds produced and released by the host

(Karamanoli et al., 2005). Some bacteria might be pathogenic to submerged living aquatic plants (macrophytes) (Chand et al., 1992; Shabana et al., 2003), while other bacterial strains are known to protect plants from hazardous pathogens, such as the specific bacterial community associated with Sphagnum mosses (Opelt et al., 2007).

Within the bacterial community on plants, complex interactions will develop that depend on external environmental factors and the interaction between plant and microorganisms and among the bacteria (Beattie and Lindow, 1999; Burmolle et al.,

2006; Hansen et al., 2007).

In the course of our investigation of the epiphytic bacterial community composition on the submerged macrophyte Myriophyllum spicatum, we isolated several strains capable to degrade the major secondary metabolites produced by this plant, hydrolysable polyphenols (Müller et al., 2007). From the water surrounding the plant, we retrieved the betaproteobacterium Matsuebacter (Mitsuaria) sp. FB25 that constitutively degrades tannic and gallic acid, while the alphaproteobacterium

Agrobacterium vitis EB26 was derived from the biofilm on leaves and needs prior

116 VI Colonization of M. spicatum by tannin-degrading bacteria

induction to degrade these polyphenols and is a well known pathogen for grapevines

(Ophel and Kerr, 1990; Müller et al., 2007). The Matsuebacter sp. strain is related to

Mitsuaria (Matsuebacter) chitosanotabidus that produces the exoenzyme chitosanase

(Park et al., 1999; Amakata et al., 2005). Not much is known about these bacteria regarding their physiology and ecology so far. We also isolated the gammaproteobacterium Pantoea agglomerans from the biofilm on M. spicatum leaves.

This bacterium occurs ubiquitous in nature and can act as a biological control agent against pathogenic bacteria through its fast and dense colonization of plant surfaces and the vast variety of antibiotics produced (Sabaratnam and Beattie, 2003). It has also been isolated from the faeces of koala bears that feed on tannin-rich diets and is frequently found in the gut of ruminants and herbivores (Osawa, 1992; Nelson et al.,

1998; Pidiyar et al., 2004). Pantoea agglomerans T71 exhibited a high tannase and gallic acid decarboxylase activity (Zeida et al., 1998), enzymes important to initiate the anaerobic degradation of hydrolysable polyphenols.

Thus, we had three strains capable to metabolise the secondary metabolites in

M. spicatum as a sole carbon source with different metabolic and ecological characteristics. We tested whether all strains are capable to recolonize axenic shoots of M. spicatum. Therefore, we developed a specific microcosm where bacteria suspended in sterile filtered Lake Constance water could attach to plant shoots rooted in a mineral medium (Figure 6.1). We were able to monitor the possible interactions of two or all three strains on the plant surface by FISH (fluorescence in situ hybridisation) due to the affiliation of all strains to different proteobacterial groups. We expected Matsuebacter sp. FB25 to cope best with exuded polyphenols but had no information on its capability to form biofilms. We hypothesized that A. vitis and P. agglomerans would colonize the plant quickly and show distinct signs of interactions due to their ecological function as a pathogen and a biological control agent. The combined results of the single and mixed species colonization experiments should reveal insights into the role of these bacteria in the biofilm of the plant.

117 VI Colonization of M. spicatum by tannin-degrading bacteria

MATERIAL & METHODS

Plants. For this biofilm formation assay we used axenic Myriophyllum spicatum shoots. Axenic shoots were grown in mineral medium A (Min A; (Gross, 2003)) at

22 ± 2 °C with a 16:8 hours light [80 µE m-2 s-1]:dark regime.

Bacteria. The strains Agrobacterium vitis and Pantoea agglomerans were isolated from the biofilm of M .spicatum, while Matsuebacter sp. FB25 was isolated from the surrounding lake water (Müller et al., 2007). Agrobacterium vitis and P. agglomerans were grown in liquid LB medium, while Matsuebacter sp. was cultured in liquid

Medium B (Müller et al., 2007). All cultures were kept at 22 °C and permanently shaken at 120 rpm. Permanent cultures were transferred to fresh medium once a week. For experiments, cultures were inoculated 72 h before the experiment started, and harvested in the exponential growth phase. The centrifuged cultures were washed two times with sterile lake water (15 min, 10,500 g at 16 °C) to remove growth medium. Cell densities of single species set-ups were adjusted to a final

OD600nm of 0.18 in 900 ml of sterile lake water for A. vitis and P. agglomerans. Since

Matsuebacter sp. has just one third of the size of the other two isolates, we inoculated this bacterium at the final OD600nm of 0.07 to ensure comparable cell densities. In the mixed species colonization assay, the bacteria were adjusted to an entire OD600nm of

0.2 with adequate parts of each bacterium (P. agglomerans and A. vitis OD600nm = 0.08,

Matsuebacter sp. OD600nm = 0.04).

Colonization assays. To investigate the bacterial colonization on M. spicatum, we designed a microcosm set-up with the characteristic that the plant growth medium was separated from the bacteria (Figure 6.1). Thus, the plant was the only nutrient source supplied, and effects would not be distorted by bacteria growing on the medium leaking into the water column. A 100 ml Erlenmeyer flask was glued to the lid of a 1.5 l preserving jar (Weck, Germany) with heat-stable silicone (Akfix, Gasket

Maker). The Erlenmeyer flask was filled with 70 ml of a sterile solution of Min A containing Gelrite (6 g l-1, Roth, Germany). Two axenic M. spicatum shoots of 5 – 8 cm length were planted into the medium before it solidified. When the medium was

118 VI Colonization of M. spicatum by tannin-degrading bacteria

solidified and the plants were fixed the lid was placed on the jar so that plants hang into the water columns upside down (Figure 6.1).

Figure 6.1. Experimental set-up. Autoclavable microcosm for the examination of biofilm formation on plants. The water column A with bacteria does not come into contact with the plant-growth media. A) Erlenmeyer flask (100 ml) filled with 70 ml of solidified Min A glued to the lid of a Weck-jar with heat-stable silicone B) Axenic plant shoot embedded in the solidified medium C) Sterile filtered Lake Constance water (900 ml). B

The jar had been filled with 900 ml of sterile filtered

lake water (nitrocellulose filters, 0.2 µm, Millipore).

C The plants were acclimatised for three to five days at experimental conditions before bacteria were added.

The experiments were conducted in a climate chamber (Sanyo MLR 350, Japan) at

22 ± 0.5 °C with 16 hours light [80 µE (m-2 s)-1] and 8 hours darkness and the light coming from the sides instead from the top so that all plants got equal light irradiance, which was not attenuated by the lid. After acclimatisation, the plants were inoculated with bacteria in mono- or mixed cultures (see above). We conducted three replicates of each treatment and one sterile control.

Sampling. Leaf samples were taken after 5, 10, 24, 48 and 72 h under sterile conditions at a laminar flow. From each jar, one leaf was removed (n = 3) with sterile tweezers and transferred into 1.5 ml reaction tubes with 0.9 ml 0.1 M sodium pyrophosphate (Na4P2O7×10 H2O). The bacterial density in the surrounding water was determined by measuring the optical density (OD600nm). The bacterial biofilm was detached as described before and the resulting suspension fixed with 0.1 ml

37% formaldehyde (Hempel et al., 2008). The leaves were transferred to 1 ml of tap water and stored at 4 °C until the leaf surface was measured.

Samples of single species treatments were filtered onto black filters (0.2 µm;

Whatman, Maidstone UK) and stained with DAPI (4,6-diamidino-2-phenylindol,

10 µg ml-1) for 5 min in the dark. To describe the colonization in the mixed species

119 VI Colonization of M. spicatum by tannin-degrading bacteria treatments we applied FISH (fluorescence in situ hybridization ) with specific probes for alpha-, beta- and gammaproteobacteria using probes ALF968, BET42a and

GAM42a with adequate competitor probes (Hempel et al., 2008).

Measurements of leaf surface. We estimated the cell density based on leaf surface area to properly compare the different experiments. Therefore, the leaves were photographed with a single lens reflex camera (Nikon D70s) fixed on a tripod and a piece of millimetre paper next to the leaf for scaling. Each leaf was photographed four times at different exposure times. The pinnate leaves of M. spicatum have round filaments of about one millimetre in diameter. Hence the surface of one leave was calculated as the total area multiplied with π (3.142) as adequate approximation. The leaf surface was determined with the programme “Macrophyte” (G. Heine, scientific workshop, University of Konstanz) and the variance of measurements of one leaf was always below 5%.

Statistical analysis. To elucidate statistically significant differences in the experiments, we performed one-way ANOVA for the single strains assays and two- way ANOVA for multi species assays. Prior to analysis, the proportional FISH data were arcsin transformed. When differences in growth or proportion were clearly visible, we waived statistical analysis. All calculations were carried out with

SigmaStat 3.11 (Systat Software, Inc).

120 VI Colonization of M. spicatum by tannin-degrading bacteria

RESULTS

Single species assays. Every bacterial isolate showed a different distribution between planktonic and attached growth. 0.025 30 A) Agrobacterium vitis grew

0.020 25 both on Myriophyllum 20 spicatum leaves and in the 0.015 15 water column. The OD600 nm 0.010 10 in the lake water steadily

0.005 5 -2 increased from 0.012 ± 0.003

) cm to 0.020 ± 0.003 in 72 hours, 0.000 0 6 0.020 B) 25 indicating an increase in cell

20 density of approx. 70%. 0.015 6 15 ll counts (x10 After 5 h 9.7 ± 2.9×10 cells 0.010 cm-2 [mean ± 1SD] were

of surrounding medium 10 0.005 attached, and 17.9 ± 2.2×106

600nm 5 -2

OD cells cm at the end (Figure 0.000 0 total bacterial ce C) 6.2A). 0.020 120

100 0.015 80 Figure 6.2. Single–species colonization 60 0.010 assays of the three isolates on 40 Myriophyllum spicatum. 0.005 A) Agrobacterium vitis B) Pantoea 20 agglomerans C) Matsuebacter sp. Grey bars: Colonization of the leaf surface [bacterial 0.000 0 cells cm-2 (×106)]; black circles: OD600 nm of 0 5 10 24 36 48 72 the surrounding lake water. Note: the scale 6 -2 Time [h] is stretched up to 140×10 cells cm for 1C. n=3, mean ± 1SD.

Of all isolates, Pantoea agglomerans established the lowest cell density on the leaf surface (Figure 6.2B). The cell densities on the plant surface increased significantly

(One-way ANOVA, F = 5.629; P = 0.004) during the first 48 h of the experiment from

1.5 ± 0.5×106 at 10 h to 6.0 ± 1.5×106 cells cm-2 at 48 h and then declined slightly to

121 VI Colonization of M. spicatum by tannin-degrading bacteria

3.5 ± 2.3×106 cells cm-2 until 72 h. The OD600 nm in the lake water ranged from 0.014 to

0.017 and did not change significantly (one-way ANOVA, F = 1.457; P = 0.26).

Matsuebacter sp. had the strongest potential to colonize the plant surface of all isolates (Figure 6.2C). Bacterial cell numbers on the leaf surface increased from

8.2 ± 0.7×106 to 84.0 ± 43.2×106 cells cm-2, thus the increase was about tenfold. The

OD600 nm in the lake water decreased significantly from 0.01 ± 0.002 to 0.005 ± 0.001

(one-way ANOVA, F = 5.417; P = 0.004).

Mixed species assays. In the mixed species assays, we found different interactions and settling patterns. Agrobacterium vitis was very abundant in the biofilm if co- cultivated with Matsuebacter sp. (Figure 6.3A). If co-cultivated with P. agglomerans the abundance of A. vitis remained low (Figure 6.3B). When P. agglomerans and

Matsuebacter sp. were co-cultivated, we observed fluctuations in the abundance of both strains, though P. agglomerans prevailed towards the end (Figure 6.3C). If all three strains were present, A. vitis dominated the biofilm and Matsuebacter sp. and

P. agglomerans showed a similar pattern as before (Figure 6.3D).

Agrobacterium vitis and Matsuebacter sp. Agrobacterium vitis always dominated the surface colonization in the presence of Matsuebacter sp (Figure 6.3A).

The former established constant cell densities on the leaves (68 ± 14% of DAPI counts), while abundance of the latter remained low (11 ± 6% of DAPI counts). In this assay, the total bacterial cell counts increased from 5 to 36 hours (3 ± 0.7×106 to

12 ± 5×106 cells cm-2) and stayed more or less constant until the end. Agrobacterium vitis and P. agglomerans. In the co-cultivation of A. vitis with

P. agglomerans, the biofilm on the leaves was dominated by P. agglomerans (Figure

6.3B). Here, A. vitis made up 14 ± 9% of DAPI counts, while P. agglomerans dominated the biofilm from the beginning on with 61 ± 10% of DAPI counts. The portions of both strains stayed constant, while total cell counts on the leave increased from

1 ± 0.3×106 cells cm-2 at 5 h to 9 ± 2 ×106 cells cm-2 at 72 h.

122 VI Colonization of M. spicatum by tannin-degrading bacteria

A) 14 80 12 60 10 8 40 6 ]

4 -2 20

2 cm 6 0 0 B) 14 80 12 10

60 ll counts [x 10 8 40 6

% of total cells total of % 4 20 2

0 0 total bacterial ce C) 14 80 Agrobacterium vitis Matsuebacter sp. 12 Pantoea agglomerans 60 10 8 40 6 4 20 2 0 0 D) 14 Figure 6.3. Mixed–species colonization 80 assays of the three isolates on Myriophyllum 12 spicatum. 60 10 A) Agrobacterium vitis and Matsuebacter sp. 8 B) A. vitis and Pantoea agglomerans C) Matsuebacter sp. and P. agglomerans 40 6 D) A. vitis, Matsuebacter sp. and P. agglomerans. 4 Black triangles: A. vitis, white triangles: 20 Matsuebacter sp., squares: P. agglomerans, dotted 2 line w/ circles: total bacterial counts cm-2. n = 3, 0 0 mean ± 1SD. 5 1024364872 Time [h]

123 VI Colonization of M. spicatum by tannin-degrading bacteria

Matsuebacter sp. and P. agglomerans. The co-cultivation experiment of

Matsuebacter sp. and P. agglomerans showed two stages (Figure 6.3C). In the first five hours, Matsuebacter sp. dominated the biofilm (46 ± 6% of DAPI counts; two-way

ANOVA, Holm Sidak post hoc, test P < 0.001), and at 24 h still dominated over

P. agglomerans. Then a shift occurred, and at 48 h and 72 h, P. agglomerans dominated the biofilm with 53 ± 20% of DAPI counts (Two-way ANOVA, Holm Sidak post hoc test, P48h 0.006 and P72h < 0.001) and Matsuebacter sp. decreased to 16 ± 5% of DAPI counts at 72 h. In this combination, biofilm formation was lowest. Total cell counts did not rise above 2 ± 1×106 cells cm-2 during the 72 h.

Agrobacterium vitis, Matsuebacter sp. and P. agglomerans. When all strains were co-cultivated, a combination of the two species treatments occurred: A. vitis made up the major proportion of the biofilm (56 ± 10% of all cell counts; Figure 6.3D; two-way

ANOVA, Holm Sidak post hoc test; P < 0.0001) as it did in the assay with

Matsuebacter sp., while the abundance of P. agglomerans and Matsuebacter sp. in the biofilm were equally low until 48 h. At the end of the experiment, Matsuebacter sp. is the second most abundant strain in the biofilm (26 ± 6% of DAPI counts). The cell density of the biofilm increased constantly during the first 48 h of this assay from

3 ± 0.7 to 8 ± 3×106 cells cm-2 and then stayed more or less constant until the end.

124 VI Colonization of M. spicatum by tannin-degrading bacteria

DISCUSSION

The aim of this study was to investigate the recolonization pattern of previously isolated strains and their interactions on the polyphenol containing macrophyte

Myriophyllum spicatum. Therefore, we developed an experimental set-up to study these effects in situ. The three bacterial strains chosen were all capable to degrade polyphenols (Zeida et al., 1998; Müller et al., 2007) and originated from the biofilm or the surrounding water of the plant.

Colonization patterns of the strains gave evidence that all three of them were able to form biofilms, although with distinct patterns. While the grapevine pathogen

Agrobacterium vitis displayed the expected high cell densities on the plant and in the water column, the agriculturally used bio control agent Pantoea agglomerans had the lowest cell densities of all three strains and hardly grew in the surrounding water. If co-cultivated with A. vitis, however, P. agglomerans dominated the biofilm throughout the experiment. Thus, it exhibits comparable colonization patterns as in terrestrial habitats. Pantoea agglomerans efficiently protects grapevine berries against the fungal pathogen Botrytis cinerea (Magnin-Robert et al., 2007). Since A. vitis is also a grapevine pathogen, similar interactions between both strains are possible.

In all mixed species assays the biofilm formation was less than in the single species assays. This could be related to competition not only for resources (nutrients and space) but also to the production of signal molecules or antibiotics.

Not much is known about the ecology of the Mitsuaria. First records go back to

1999, when the chitosanase of Matsuebacter chitosanotabidus 3001 (Mitsuaria chitosanitabida) originating from Japanese soil was described (Park et al., 1999;

Amakata et al., 2005). Here, we show that Matsuebacter sp. is able to form dense biofilms on polyphenol containing macrophytes even in the presence of a supposedly strong competitor like P. agglomerans.

The dominance of Matsuebacter sp. over P. agglomerans during the first 36 h of the experiments might be explained by several reasons. Firstly, P. agglomerans was grown on a nutrient rich medium (LB) and the transfer to the rather nutrient poor

125 VI Colonization of M. spicatum by tannin-degrading bacteria lake water might have resulted in a growth retardation. This is also in accordance with the single species set-up. Agrobacterium vitis also had reduced doubling times when transferred to lake water. Matsuebacter sp., however, is grown on nutrient poor medium B and thus the change to lake water might not be as severe as for

P. agglomerans. In all our preliminary tests, Matsuebacter sp. did not grow on nutrient rich media (LB), while P. agglomerans did not grow on Medium B. Further, in terrestrial systems, P. agglomerans is applied directly to the respective host plants

(Remus et al., 2000). Thus, it might be possible that the attachment of P. agglomerans is impeded if not directly sprayed on plants but the bacteria have to actively attach to a surface.

Also a synergistic effect could be responsible for the observed colonization pattern. Matsuebacter sp. grows constitutively with polyphenols, while P. agglomerans possesses the required enzymes for anaerobic milieus (Zeida et al., 1998; Müller et al.,

2007). In this scenario, Matsuebacter sp. would be the initial colonizer that degrades the exuded plant polyphenols, which would facilitate the colonization of

P. agglomerans and it could thrive on the degradation products. The mechanisms and importance of facilitation are described in detail by (Stachowicz, 2001). However, in growth experiments P. agglomerans grew readily with tannic acid in aerobic environments (data not shown). Thus, P. agglomerans does not necessarily need to grow anaerobic to degrade tannic acid. The success of P. agglomerans in the second half of the assay with Matsuebacter sp. could be caused by the production of antibiotics and bactericides that are produced by this strain (Wright et al., 2001;

Montesinos, 2007). If Matsuebacter sp. is inhibited by P. agglomerans we could not investigate, but the three species assay indicates that Matsuebacter sp. influences

P. agglomerans in a way that the latter cannot compete with A. vitis. One reason for this could be that Matsuebacter sp. attenuates the formerly strong inhibition of A. vitis by P. agglomerans by metabolizing the EPS (extracellular polymeric substances) matrix produced by A. vitis. This would reduce the bacterial attachment, since EPS is used for attachment. Matsuebacter sp. could also modify tannins exuded by

126 VI Colonization of M. spicatum by tannin-degrading bacteria

M. spicatum, which could modify the competitive ability of the other strains. Or

Matsuebacter sp. could change signal molecules produced by P. agglomerans. But these arguments are rather speculative and need further evidence.

With this microcosm approach, we were able to show that the polyphenol degrading soil bacterium Matsuebacter sp. is able to form dense biofilms in aquatic habitats and can inhibit the biological control agent P. agglomerans in such way, that it will not displace A. vitis. We assume that the interactions of these model strains on

M. spicatum are regulated both by the different growth rates and production of EPS and antibiotics, but also by the different capability to degrade plant polyphenols.

Acknowledgement: This work was supported by the German Science Foundation with grant CRC454, project A2 to EMG.

127 VI Colonization of M. spicatum by tannin-degrading bacteria

128

Chapter VII

General Discussion and Perspectives

129 VII General Discussion and Perspectives

130 VII General Discussion and Perspectives

GENERAL DISCUSSION AND PERSPECTIVES

With this work, I present one of the first comprehensive studies dealing with the bacterial community composition (BCC) on submerged macrophytes in freshwater and brackish water. My aim was to expand the scarce knowledge on heterotrophic biofilms on natural aquatic surfaces, especially submerged macrophytes, and the factors determining the BCC. With cultivation-dependent and -independent techniques (DGGE, FISH, clone library), I described the BCC on different macrophytes in a freshwater (Lake Constance) and a brackish water habitat

(Schaproder Bodden). We investigated the influence of habitat, plant species and plant part (age), plant chemistry and environmental factors (conductivity, water temperature, water level, pH) on the BCC.

In almost all samples analyzed, we found a BCC dominated by bacteria of the CFB group with fluctuating abundance, followed by alpha- and betaproteobacteria. Only on Potamogeton perfoliatus, the abundance of gammaproteobacteria was slightly higher compared to the other substrates. Planctomycetes were most abundant on brackish water plants and only seldom occurred on freshwater plants. Actinomycetes were of minor abundance on all substrates analyzed. I found that differences in BCC on submerged macrophytes depend on habitat, substrate (plant species) and in part also on plant age (plant part). These findings are in accordance with other studies on freshwater biofilms. There, also alpha- and betaproteobacteria, and bacteria of the

CFB-group were most abundant (Brümmer et al. 2000, Schweitzer et al. 2001,

Grossart et al. 2008). In my samples, I seldom found gammaproteobacteria, only associated with senescing P. perfoliatus. This is probably related to the higher release of organic compounds and nutrients during senescence and in accordance with other studies (Wagner et al. 1993, Brachvogel et al. 2001). The BCC found in the field was confirmed by the initial colonization of axenic Myriophyllum spicatum in an outdoor mesocosm. After an initial dominance of CFB bacteria, betaproteobacteria dominated the biofilm. This indicates that macrophytes are colonized very quickly and the BCC is very constant with regard to major bacterial groups.

131

VII General Discussion and Perspectives

In all our community analyses, we found a distinct heterogeneous BCC for freshwater M. spicatum apices compared to all other samples. We hypothesized that the distinct BCC of M. spicatum apices might be related to the high total phenolic content in these samples. I used detailed statistical analyses (BIO-ENV) relating plant chemistry and environmental factors to the relative abundance of bacterial groups identified by FISH and the DGGE banding pattern of BCC on different substrates.

With DGGE data, we did not get significant correlations except for the leaf section, where conductivity explained some of the BCC variability. This was surprising, since the community analysis with DGGE gives a finer resolution since ideally individual strains are detected compared to FISH, where we used six probes to detect major bacterial groups. Probably the lower replicate number of DGGE (n = 13) may have lead to insignificant results. In contrast, the BIO-ENV analysis performed with the

FISH data set resulted in significant effects. Both environmental factors, especially water level and temperature, conductivity and pH as well as plant chemistry, i.e., plant carbon and total phenolic content, explained most of the BCC variability.

Similar factors have also been found to influence the BCC of attached and free-living bacteria in other freshwater habitats (Lindström et al. 2005, Allgaier & Grossart 2006).

The different outcome of and conclusions drawn from DGGE and FISH analyses show that it is very important how determinants for BCC are calculated. It also shows that community analyses should always be done with two or more complementary methods. This is also supported by our finding that M. spicatum apices had a distinct biofilm based on cluster analysis of the banding pattern in

DGGE. If the BCC was analysed with FISH, the difference between the substrates were not as obvious. This can also be explained by the probes we chose for FISH.

They are designed for major bacterial groups, which comprise many ecologically and physiologically different strains. Thus, the information of single strain shifts during the season or between different substrates gets lost, if only general probes are applied. The high variance of CFB bacteria in our samples is probably also related to the probe chosen. It detects only 38% of this group, but presently no better probes are

132 VII General Discussion and Perspectives

available (Loy et al. 2003). To avoid this bias, one could use a combined set of probes for CFB-bacteria to minimize detections errors.

Seasonal dynamics of the bacterial groups were observed for alpha- and betaproteobacteria as determined by FISH, while cluster analysis of the DGGE banding pattern did not reveal distinct seasonal changes. On lake snow, a distinct succession of different subgroups within the betaproteobacterial community was detected with only three probes for close relatives of this group (Schweitzer et al.

2001), although the total abundance of the betaproteobacteria did not change. Thus, I propose that further community analysis should be performed with more specific probes for the most abundant groups to get are detailed community analysis and to follow distinct successions more precisely.

The BCC data combined with the total cell counts of the substrates show that the plant chemistry might have an impact on bacteria, but not necessarily a negative one.

Total cell counts were higher on M. spicatum than on P. perfoliatus. This could be a consequence of higher nutrient leakage or leaf structure. Feathery, finely dissected leaves tend to support higher densities of epiphytes compared to laminar ones

(Lalonde & Downing 1991). Flagellates also might influence the BCC and bacterial numbers (Jürgens & Matz 2002). Maybe the high polyphenolic content of M. spicatum negatively affected bacterivorous flagellates, which could increase total cell counts.

Lethal and sublethal effects of M. spicatum exudates on invertebrates have been shown recently (Linden & Lehtiniemi 2005). Unfortunately, due to time constraints, we did not evaluate flagellate abundances.

If the total phenolic content really influences the BCC can only be determined in controlled experiments with substrates releasing defined concentrations and types of phenolic plant secondary metabolites (see below).

Exuded plant compounds or the quality of the DOC in a water body may influence the BCC, as has been shown previously (Eiler et al. 2003, Huss & Wehr

2004, Tadonléké 2007). Sphagnum mosses, for example, harbour a unique epiphytic

133

VII General Discussion and Perspectives community (Opelt et al. 2007), most likely caused by the extreme habitats they live in.

Polyphenols have been shown to be detrimental to bacteria by chelating proteins, iron and nutrients (Scalbert 1991). Defined synthetic polyphenols may inhibit biofilm formation (Huber et al. 2003). Thus, exuded polyphenols from M. spicatum might even inhibit quorum sensing biofilm formation. We successfully isolated three strains that are capable of degrading the exuded polyphenols of M. spicatum (this study,

(Müller et al. 2007)), providing further evidence that this plant selects towards a defined bacterial community. Very special among those isolates is Matsuebacter sp.

FB 25. This betaproteobacterium is able to degrade polyphenols constitutively and almost nothing is known about its ecology so far.

We tested if this bacterium would influence the impaired growth of Acentria ephemerella larvae when fed M. spicatum. Epiphytic bacteria are inevitably taken up during feeding and thus may interact with the herbivore (Dillon & Dillon 2004). We designed two no-choice feeding assays to compare larval growth on M. spicatum with different bacterial biofilms. Unfortunately, larvae fed with mesocosm plants exhibited a high mortality. Most likely this was caused by the higher leaf toughness due to a higher ash content of these plants. Thus, we can presently only conclude that

Matsuebacter sp.-colonized M. spicatum did not enhance the growth of A. ephemerella compared to axenic M. spicatum. We expected an influence of this bacterium either by attenuating the impact of polyphenols by degrading them, influencing the gut microbiota or as additional carbon or nitrogen source, thus raising the food quality of the plants. That the effect of the bacteria was rather negligible might also be caused by the short gut passage of the larvae that accounts for the transient nature of most gut bacteria. Alternately, Matsuebacter sp. might not be able to gain dominance or even co-exist in a gut already colonized by other gut bacteria. The knowledge on the gut microbiota of adults and larvae is still scarce. To further investigate the impact of bacteria and food quality on larval growth, we could use Pantoea agglomerans, isolated from the biofilm of M. spicatum and frequently found in herbivore guts

134 VII General Discussion and Perspectives

(Dillon et al. 2002). To elucidate the impact of M. spicatum derived tannins and associated bacteria on the growth it would be necessary to raise A. ephemerella larvae on an artificial diet. Thus, it would be possible to exclude other factors such as food quality. But all attempts to feed those larvae with an artificial diet failed (Choi et al.

2002, Erhard et al. 2007).

We also studied the initial colonisation of axenic M. spicatum with selected bacterial isolates. Matsuebacter sp. was a strong colonizer in single species set-ups but not in combination with Agrobacterium vitis. Interestingly, Matsuebacter sp. was able to inhibit P. agglomerans in such a way that it could not dominate the biofilm if A. vitis was present. The mechanism of this interaction remains unclear so far. We assume that Matsuebacter sp. may be able to degrade bacteriostatic compounds P. agglomerans releases against A. vitis. This would facilitate the biofilm formation of A. vitis. Or competition for resources (space, nutrients) affects the interaction, e.g., the differences in the constitutive capacity to degrade plant polyphenols. In preliminary investigation, we also tested if those strains produce N–acylhomoserinlactones, but all tests were negative even though quorum sensing activities have been described before for A. vitis and P. agglomerans (Holden et al. 1999, Wang et al. 2008). We think that the interactions of those strains are an ideal model for further investigations.

With our new experimental set-up, we are able to analyze the complex interactions between the three strains. This assay allows also investigating the reciprocal impact of the bacteria on the plant and vice versa.

135

VII General Discussion and Perspectives

Perspectives

For a better understanding of the processes shaping biofilm formation, succession and interactions of bacteria within the biofilm, but also with the host or herbivores, further investigations are needed.

I suggest that the biofilm on the investigated macrophytes should be analyzed with more specific probes and to use a combination of methods such as MICRO-FISH or stable isotope probing to obtain detailed data on bacterial abundance and activity.

The inclusion of more microscopic methods such as scanning electron microscopy

(SEM) or confocal laser scanning microscopy (CLSM) should allow answers to the spatial distribution of bacteria in the biofilm. Green fluorescent protein labelling of certain bacterial strains would facilitate the biofilm research, since FISH is very laborious and the complete detachment of the bacterial biofilm always has a certain uncertainty. Strains that do not need to be stained would facilitate the analysis.

Cyanobacteria, eukaryotic algae, fungi and flagellates should also be included into further analysis, since they are important components of the biofilms on aquatic surfaces.

The final evidence that the polyphenols of M. spicatum have a crucial influence on the BCC is still lacking. The proof for this can only be given experimentally, since field work includes too many unknown or uncontrolled factors that would also influence the BCC. Field work would also require an immense sampling number.

Thus, we need an experimental set-up with natural lake water flow-through and a substrate that can be manipulated with respect to the provided carbon source. The first trials showed that due to the high chemical reactivity of tannic acid and other polyphenols, experiments are difficult and susceptible to artefacts.

To investigate the interactions of bacterial isolates and their impact on the plant, not only the standard plant chemistry (total phenolic content, chlorophyll) should be analysed but also the reactions of the plant on the genomic (real time PCR) or protein level (SDS-PAGE). If the strains induce plant responses, they could further be manipulated with respect to degradation pathways, biofilm formation or single

136 VII General Discussion and Perspectives

pathways to elucidate, which physiological traits are responsible for the interactions with the plants or other bacteria.

Thus, aquatic biofilms on aquatic macrophytes and the allelochemical interactions of aquatic plants and biofilm bacteria remains an interesting and exciting field of research. With my PhD thesis, I elucidated some of the open questions but still, many more studies have to be done to understand biofilms and their ecological impact on macrophytes.

137

VII General Discussion and Perspectives

138 VIII References

REFERENCES

Ahn SJ, Burne RA (2007) Effects of oxygen on biofilm formation and the AtlA autolysin of Streptococcus mutans. J Bacteriol 189:6293-6302

Allgaier M, Grossart HP (2006) Seasonal dynamics and phylogenetic diversity of free-living and particle-associated bacterial communities in four lakes in northeastern Germany. Aquat Microb Ecol 45:115-128

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignment search tool. J Mol Biol 215:403-410

Amakata D, Matsuo Y, Shimono K, Park JK, Yun CS, Matsuda H, Yokota A, Kawamukai M (2005) Mitsuaria chitosanitabida gen. nov., sp nov., an aerobic, chitosanase-producing member of the ʹBetaproteobacteriaʹ. Int J Syst Evol Microbiol 55:1927-1932

Amann RI, Binder BJ, Olson RJ, Chisholm SW, Devereux R, Stahl DA (1990) Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analyzing mixed microbial populations. Appl Environ Microbiol 56:1919-1925

Anthoni U, Nielsen PH, Smith-Hansen L, Wium-Andersen S, Christophersen C (1987) Charamin, a quaternary ammonium ion antibiotic from the green alga Chara globularis. J Org Chem 52:694-695

Appel HM (1994) The chewing herbivore gut lumen: physicochemical conditions and their impact on plant nutrients, allelochemicals, and insect pathogens. In: Bernays EA (ed) Insect-plant interactions. CRC Press, Boca Raton, p 240

Ardon M, Pringle CM (2007) The quality of organic matter mediates the response of heterotrophic biofilms to phosphorus enrichment of the water column and substratum. Freshw Biol 52:1762-1772

Azam F, Fenchel T, Field JG, Gray JS, Meyer-Reil LA, Thingstad F (1983) The ecological role of water-column microbes in the sea. Mar Ecol Prog Ser 10:257- 263

Battin TJ, Sloan WT, Kjelleberg S, Daims H, Head IM, Curtis TP, Eberl L (2007) Microbial landscapes: new paths to biofilm research. Nature Reviews Microbiology 5:76-81

139 VIII References

Beattie GA, Lindow SE (1999) Bacterial colonization of leaves: A spectrum of strategies. Phytopathology 89:353-359

Bertilsson S, Tranvik LJ (1998) Photochemically produced carboxylic acids as substrates for freshwater bacterioplankton. Limnol Oceanogr 43:885-895

Bibb M (1996) The regulation of antibiotic production in Streptomyces coelicolor A3(2). Microbiology 142:1335-1344

Blindow I (1987) The composition and density of epiphyton on several species of submerged macrophytes - the neutral substrate hypothesis tested. Aquat Bot 29:157-168

Blindow I, Dietrich J, Mollmann N, Schubert H (2003) Growth, photosynthesis and fertility of Chara aspera under different light and salinity conditions. Aquat Bot 76:213-234

Bockelmann U, Szewzyk U, Grohmann E (2003) A new enzymatic method for the detachment of particle associated soil bacteria. J Microbiol Methods 55:201-211

Box JD (1983) Investigation of the Folin-Ciocalteau phenol reagent for the determination of polyphenolic substances in natural waters. Water Res 17:511- 525

Brachvogel T, Schweitzer B, Simon M (2001) Dynamics and bacterial colonization of microaggregates in a large mesotrophic lake. Aquat Microb Ecol 26:23-35

Brümmer IH, Fehr W, Wagner-Dobler I (2000) Biofilm community structure in polluted rivers: abundance of dominant phylogenetic groups over a complete annual cycle. Appl Environ Microbiol 66:3078-3082

Brümmer IH, Felske AD, Wagner-Dobler I (2004) Diversity and seasonal changes of uncultured Planctomycetales in river biofilms. Appl Environ Microbiol 70:5094-5101

Buesing N, Gessner MO (2002) Comparison of detachment procedures for direct counts of bacteria associated with sediment particles, plant litter and epiphytic biofilms. Aquat Microb Ecol 27:29-36

Buesing N, Gessner MO (2006) Benthic bacterial and fungal productivity and carbon turnover in a freshwater marsh. Appl Environ Microbiol 72:596-605

140 VIII References

Burmolle M, Webb JS, Rao D, Hansen LH, Sorensen SJ, Kjelleberg S (2006) Enhanced biofilm formation and increased resistance to antimicrobial agents and bacterial invasion are caused by synergistic interactions in multispecies biofilms. Appl Environ Microbiol 72:3916-3923

Bushmann PJ, Ailstock MS (2006) Antibacterial compounds in estuarine submersed aquatic plants. J Exp Mar Biol Ecol 331:41-50

Cangiano T, DellaGreca M, Fiorentino A, Isidori M, Monaco P, Zarrelli A (2001) Lactone diterpenes from the aquatic plant Potamogeton natans. Phytochemistry 56:469-473

Carignan R, Kalff J (1982) Phosphorus release by submerged macrophytes: Significance to epiphyton and phytoplankton. Limnol Oceanogr 27:419-427

Chand T, Harris RF, Andrews JH (1992) Enumeration and characterization of bacterial colonists of a submersed aquatic plant, Eurasian watermilfoil (Myriophyllum spicatum L.). Appl Environ Microbiol 58:3374-3379

Chhabra SR, Philipp B, Eberl L, Givskov M, Williams P, Camara M (2005) Extracellular communication in bacteria. In: Chemistry of pheromones and other semiochemicals, Vol 240, p 279-315

Choi C, Bareiss C, Walenciak O, Gross EM (2002) Impact of polyphenols on growth of the aquatic herbivore Acentria ephemerella. J Chem Ecol 28:2245-2256

Costerton JW, Lewandowski Z, Caldwell DE, Korber DR, Lappinscott HM (1995) Microbial Biofilms. Annu Rev Microbiol 49:711-745

Cotner JB, Biddanda BA (2002) Small players, large role: Microbial influence on biogeochemical processes in pelagic aquatic ecosystems. Ecosystems 5:105-121

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 Microbiol 66:1692- 1697

Czaczyk K, Myszka K (2007) Biosynthesis of extracellular polymeric substances (EPS) and its role in microbial biofilm formation. Pol J Environ Stud 16:799-806

Daims H, Bruhl A, Amann R, Schleifer KH, Wagner M (1999) The domain-specific probe EUB338 is insufficient for the detection of all Bacteria: development and evaluation of a more comprehensive probe set. Syst Appl Microbiol 22:434-444

141 VIII References

Danger M, Leflaive J, Oumarou C, Ten-Hage L, Lacroix G (2007a) Control of phytoplankton-bacteria interactions by stoichiometric constraints. Oikos 116:1079-1086

Danger M, Oumarou C, Benest D, Lacroix G (2007b) Bacteria can control stoichiometry and nutrient limitation of phytoplankton. Funct Ecol 21:202-210

De Beer D, Stoodley P, Lewandowski Z (1994) Liquid flow in heterogeneous biofilms. Biotechnol. Bioeng. 44:636-641

DellaGreca M, Fiorentino A, Isidori M, Monaco P, Temussi F, Zarrelli A (2001) Antialgal furano-diterpenes from Potamogeton natans L. Phytochemistry 58:299-304

Dillon RJ, Dillon VM (2004) The gut bacteria of insects: nonpathogenic interactions. Annu Rev Entomol 49:71-92

Dillon RJ, Vennard CT, Charnley AK (2002) A Note: Gut bacteria produce components of a locust cohesion pheromone. J Appl Microbiol 92:759-763

Dobretsov S, Dahms HU, Huang YL, Wahl M, Qian PY (2007) The effect of quorum- sensing blockers on the formation of marine microbial communities and larval attachment. FEMS Microbiol Ecol 60:177-188

Eberl L (1999) N-acyl homoserinelactone-mediated gene regulation in gram-negative bacteria. Syst Appl Microbiol 22:493-506

Eiler A, Bertilsson S (2004) Composition of freshwater bacterial communities associated with cyanobacterial blooms in four Swedish lakes. Environ Microbiol 6:1228-1243

Eiler A, Langenheder S, Bertilsson S, Tranvik LJ (2003) Heterotrophic bacterial growth efficiency and community structure at different natural organic carbon concentrations. Appl Environ Microbiol 69:3701-3709

Elser JJ, Fagan WF, Denn RF, Dobberfuhl DR, Folarin A, Huberty A, Interlandi S, Kilham SS, McCauley E, Schulz KL, Siemann EH, Sterner RW (2000) Nutritional constraints in terrestrial and freshwater food webs. Nature 408: 578-580

Eminson D, Moss B (1980) The composition and ecology of periphyton communities in freshwaters. 1. The influence of host type and external environment on community composition. Br Phycol J 15:429-446

142 VIII References

Erhard D, Pohnert G, Gross EM (2007) Chemical defense in Elodea nuttallii reduces feeding and growth of aquatic herbivorous Lepidoptera. J Chem Ecol 33:1646- 1661

Eriksson PG, Weisner SEB (1999) An experimental study on effects of submersed macrophytes on nitrification and denitrification in ammonium-rich aquatic systems. Limnol Oceanogr 44:1993-1999

Feeny P (1970) Seasonal changes in oak leaf tannins and nutrients as a cause of spring feeding by winter moth caterpillars. Ecology 51:565-581

Fisher MM, Wilcox LW, Gram LE (1998) Molecular characterization of epiphytic bacterial communities on charophycean green algae. Appl Environ Microbiol 64:4384-4389

Frost PC, Cross WF, Benstead JP (2005) Ecological stoichiometry in freshwater benthic ecosystems: an introduction. Freshw Biol 50:1781-1785

Frost PC, Elser JJ (2002) Growth response of littoral mayflies to the phosphorus content of their food. Ecol Lett 5:232-240

Glöckner FO, Fuchs BM, Amann R (1999) Bacterioplankton compositions of lakes and oceans: a first comparison based on fluorescence in situ hybridization. Appl Environ Microbiol 65:3721-3726

Godmaire H, Planas D (1986) Influence of Myriophyllum spicatum on the species composition biomass and primary productivity of phytoplankton. Aquat Bot 23:299-308

Gould KS (2004) Natureʹs Swiss army knife: The diverse protective roles of anthocyanins in leaves. J Biomed Biotechnol:314-320

Gregory SV (1983) Plant -herbivore interactions in stream ecosystems. In: Barnes JR (ed) Sream Ecology. Plenum Press, p 157-190

Gross EM (2003) Differential response of tellimagrandin II and total bioactive hydrolysable tannins in an aquatic angiosperm to changes in light and nitrogen. Oikos 103:497-504

Gross EM, Feldbaum C, Choi C (2002a) High abundance of herbivorous Lepidoptera larvae (Acentria ephemerella DENIS & SCHIFFERMULLER) on submersed macrophytes in Lake Constance (Germany). Arch Hydrobiol 155:1-21

143 VIII References

Gross EM, Meyer H, Schilling G (1996) Release and ecological impact of algicidal hydrolysable polyphenols in Myriophyllum spicatum. Phytochemistry 41:133- 138

Grossart HP, Jezbera J, Hornak K, Hutalle KML, Buck U, Simek K (2008) Top-down and bottom-up induced shifts in bacterial abundance, production and community composition in an experimentally divided humic lake. Environ Microbiol 10:635-652

Grossart HP, Kiorboe T, Tang K, Ploug H (2003) Bacterial colonization of particles: Growth and interactions. Appl Environ Microbiol 69:3500-3509

Grossart HP, Levold F, Allgaier M, Simon M, Brinkhoff T (2005) Marine diatom species harbour distinct bacterial communities. Environ Microbiol 7:860-873

Guan LL, Onuki H, Kamino K (2000) Bacterial growth stimulation with exogenous siderophore and synthetic N-acyl homoserine lactone autoinducers under iron-limited and low-nutrient conditions. Appl Environ Microbiol 66:2797- 2803

Hammer BK, Bassler BL (2003) Quorum sensing controls biofilm formation in Vibrio cholerae. Mol Microbiol 50:101-114

Hansen SK, Rainey PB, Haagensen JAJ, Molin S (2007) Evolution of species interactions in a biofilm community. Nature 445:533-536

Harrison JJ, Ceri H, Turner RJ (2007) Multimetal resistance and tolerance in microbial biofilms. Nat Rev Microbiol 5:928-938

Hay ME, Kappel QE, Fenical W (1994) Synergisms in plant defenses against herbivores: interactions of chemistry, calcification, and plant quality. Ecology 75:1714-1726

Head IM, Saunders JR, Pickup RW (1998) Microbial evolution, diversity, and ecology: A decade of ribosomal RNA analysis of uncultivated microorganisms. Microb Ecol 35:1-21

Heilman MA, Carlton RG (2001) Methane oxidation associated with submersed vascular macrophytes and its impact on plant diffusive methane flux. Biogeochemistry 52:207-224

144 VIII References

Hempel M, Blume M, Blindow I, Gross EM (2008) Epiphytic bacterial community composition on two common submerged macrophytes in brackish water and freshwater. BMC Microbiol 8:58

Hempel M, Gross EM (submitted) Spatio-temporal dynamics of the bacterial biofilm on two submerged macrophytes and an artificial substrate in Lower Lake Constance.

Hempel M, Grossart HP, Gross EM (submitted) Bacterial community composition of biofilms on two submerged macrophytes and an artificial substrate in Lake Constance.

Hentzer M, Riedel K, Rasmussen TB, Heydorn A, Andersen JB, Parsek MR, Rice SA, Eberl L, Molin S, Hoiby N, Kjelleberg S, Givskov M (2002) Inhibition of quorum sensing in Pseudomonas aeruginosa biofilm bacteria by a halogenated furanone compound. Microbiology 148:87-102

Hilt S, Gross EM (2008) Can allelopathically active submerged macrophytes stabilise clear-water states in shallow lakes? Basic Appl Ecol 9:422-432

Hoagland RE (2001) Microbial allelochemicals and pathogens as bioherbicidal agents. Weed Technol 15:835-857

Holden MTG, Chhabra SR, de Nys R, Stead P, Bainton NJ, Hill PJ, Manefield M, Kumar N, Labatte M, England D, Rice S, Givskov M, Salmond GPC, Stewart G, Bycroft BW, Kjelleberg SA, Williams P (1999) Quorum-sensing cross talk: isolation and chemical characterization of cyclic dipeptides from Pseudomonas aeruginosa and other Gram-negative bacteria. Mol Microbiol 33:1254-1266

Hong SH, Lee YO, Kim HW, Ahn TS (1999) Succession and diversity of attached bacteria on cellulose film and leaves of Potamogeton crispus in Lake Moonchon, Korea. Arch Hydrobiol Spec Issues Advanc Limnol 54:273-282

Hossell JC, Baker JH (1979) Note on the enumeration of epiphytic bacteria by microscopic methods with particular reference to 2 freshwater plants. J Appl Bact 46:87-92

Huber B, Eberl L, Feucht W, Polster J (2003) Influence of polyphenols on bacterial biofilm formation and quorum-sensing. Z Naturforsch Sect C J Biosci 58:879- 884

145 VIII References

Hugenholtz P, Goebel BM, Pace NR (1998) Impact of culture-independent studies on the emerging phylogenetic view of bacterial diversity. J Bacteriol 180:4765- 4774

Huss AA, Wehr JD (2004) Strong indirect effects of a submersed aquatic macrophyte, Vallisneria americana, on bacterioplankton densities in a mesotrophic lake. Microb Ecol 47:305-315

IGKB (2006) Der Bodensee Zustand-Fakten-Perspektiven. Report No. 57, Internationale Gewässerschutzkomission für den Bodensee, Stuttgart

IGKB (2007) Limnologischer Zustand des Bodensees. Report No. 34, Internationale Gewässerschutzkomission für den Bodensee

Jackson LJ, Rowan DJ, Cornett RJ, Kalff J (1994) Myriophyllum spicatum pumps essential and nonessential trace elements from sediments to epiphytes. Can J Fish Aquat Sci 51:1769-1773

Jasti S, Sieracki ME, Poulton NJ, Giewat MW, Rooney-Varga JN (2005) Phylogenetic diversity and specificity of bacteria closely associated with Alexandrium spp. and other phytoplankton. Appl Environ Microbiol 71:3483-3494

Jeppesen E, Sondergaard M, Sondergaard M, Christoffersen K (1998) The structuring role of submerged macrophytes in lakes, Vol 131. Springer, New York

Ji R, Brune A (2001) Transformation and mineralization of C-14-labeled cellulose, peptidoglycan, and protein by the soil-feeding termite Cubitermes orthognathus. Biol Fertil Soils 33:166-174

Ji R, Kappler A, Brune A (2000) Transformation and mineralization of synthetic C-14- labeled humic model compounds by soil-feeding termites. Soil Biol Biochem 32:1281-1291

Joint I, Callow ME, Callow JA, Clarke KR (2000) The attachment of Enteromorpha zoospores to a bacterial biofilm assemblage. Biofouling 16:151-158

Jones PR, Cottrell MT, Kirchman DL, Dexter SC (2007) Bacterial community structure of biofilms on artificial surfaces in an estuary. Microb Ecol 53:153-162

Jürgens K, Matz C (2002) Predation as a shaping force for the phenotypic and genotypic composition of planktonic bacteria. Antonie Van Leeuwenhoek 81:413-434

146 VIII References

Karamanoli K, Menkissoglu-Spiroudi U, Bosabalidis AM, Vokou D, Constantinidou HIA (2005) Bacterial colonization of the phyllosphere of nineteen plant species and antimicrobial activity of their leaf secondary metabolites against leaf associated bacteria. Chemoecology 15:59-67

Kiorboe T, Grossart HP, Ploug H, Tang K (2002) Mechanisms and rates of bacterial colonization of sinking aggregates. Appl Environ Microbiol 68:3996-4006

Kiorboe T, Tang K, Grossart HP, Ploug H (2003) Dynamics of microbial communities on marine snow aggregates: Colonization, growth, detachment, and grazing mortality of attached bacteria. Appl Environ Microbiol 69:3036-3047

Kirchman DL (2002) The ecology of Cytophaga-Flavobacteria in aquatic environments. FEMS Microbiol Ecol 39:91-100

Kleerebezem M, Quadri LEN, Kuipers OP, deVos WM (1997) Quorum sensing by peptide pheromones and two-component signal-transduction systems in Gram-positive bacteria. Mol Microbiol 24:895-904

Kubanek J, Jensen PR, Keifer PA, Sullards MC, Collins DO, Fenical W (2003) Seaweed resistance to microbial attack: A targeted chemical defense against marine fungi. Proc Natl Acad Sci USA 100:6916-6921

Lalonde S, Downing JA (1991) Epiphyton biomass is related to lake trophic status, depth, and macrophyte architecture. Can. J. Fish. Aquat. Sci. 48:2285-2291

Lane DJ, Pace B, Olsen GJ, Stahl DA, Sogin ML, Pace NR (1985) Rapid determination of 16s ribosomal RNA sequences for phylogenetic analyses. Proc Natl Acad Sci 82:6955-6959

Leu E, Krieger-Liszkay A, Goussias C, Gross EM (2002) Polyphenolic allelochemicals from the aquatic angiosperm Myriophyllum spicatum inhibit photosystem II. Plant Physiol 130:2011-2018

Linden E, Lehtiniemi M (2005) The lethal and sublethal effects of the aquatic macrophyte Myriophyllum spicatum on Baltic littoral planktivores. Limnol Oceanogr 50:405-411

Lindström ES, Kamst-Van Agterveld MP, Zwart G (2005) Distribution of typical freshwater bacterial groups is associated with pH, temperature, and lake water retention time. Appl Environ Microbiol 71:8201-8206

147 VIII References

Loy A, Horn M, Wagner M (2003) probeBase: an online resource for rRNA-targeted oligonucleotide probes. Nucleic Acids Res 31:514-516

Lozupone CA, Knight R (2007) Global patterns in bacterial diversity. Proc Natl Acad Sci 104:11436-11440

Lupp C, Urbanowski M, Greenberg EP, Ruby EG (2003) The Vibrio fischeri quorum sensing systems ain and lux sequentially induce luminescence gene expression and are important for persistence in the squid host. Mol Microbiol 50:319-331

Maeta K, Nomura W, Takatsume Y, Izawa S, Inoue Y (2007) Green tea polyphenols function as prooxidants to activate oxidative-stress-responsive transcription factors in yeasts. Appl Environ Microbiol 73:572-580

Magnin-Robert M, Trotel-Aziz P, Quantinet D, Biagianti S, Aziz A (2007) Biological control of Botrytis cinerea by selected grapevine-associated bacteria and stimulation of chitinase and beta-1,3 glucanase activities under field conditions. Eur J Plant Pathol 118:43-57

Maie N, Pisani O, Jaffé R (2008) Mangrove tannins in aquatic ecosystems: Their fate and possible influence on dissolved organic carbon and nitrogen cycling. Limnol Oceanogr 53:160–171

Manefield M, Turner SL (2002) Quorum sensing in context: out of molecular biology and into microbial ecology. Microbiology 148:3762-3764

Manz W, Amann R, Ludwig W, Vancanneyt M, Schleifer KH (1996) Application of a suite of 16S rRNA-specific oligonucleotide probes designed to investigate bacteria of the phylum cytophaga-flavobacter-bacteroides in the natural environment. Microbiology 142 1097-1106

Manz W, Amann R, Ludwig W, Wagner M, Schleifer KH (1992) Phylogenetic oligodeoxynucleotide probes for the major subclasses of proteobacteria - Problems and solutions. Syst Appl Microbiol 15:593-600

Marketon MM, Glenn SA, Eberhard A, Gonzalez JE (2003) Quorum sensing controls exopolysaccharide production in Sinorhizobium meliloti. J Bacteriol 185:325-331

Marshall K, Joint I, Callow ME, Callow JA (2006) Effect of marine bacterial isolates on the growth and morphology of axenic plantlets of the green alga Ulva linza. Microb Ecol 52:302-310

148 VIII References

Mathesius U, Mulders S, Gao M, Teplitski M, Caetano-Anolles G, Rolfe BG, Bauer WD (2003) Extensive and specific responses of a eukaryote to bacterial quorum-sensing signals. Proc Natl Acad Sci USA 100:1444-1449

Matsuo Y, Imagawa H, Nishizawa M, Shizuri Y (2005) Isolation of an algal morphogenesis inducer from a marine bacterium. Science 307:1598

Maximilien R, de Nys R, Holmstrom C, Gram L, Givskov M, Crass K, Kjelleberg S, Steinberg PD (1998) Chemical mediation of bacterial surface colonisation by secondary metabolites from the red alga Delisea pulchra. Aquat Microb Ecol 15:233-246

McSweeny CS, Palmer B, Bunch R, Krause DO (1999) Isolation and characterization of proteolytic ruminal bacteria from sheep and goats fed the tannin-containing shrub legume Calliandra calothyrsus. Appl Environ Microbiol 65:3075-83

Molisch H (1937) Der Einfluss einer Pflanze auf die andere - Allelopathie, Vol. Fischer, Jena, Germany

Montesinos E (2007) Antimicrobial peptides and plant disease control. Fems Microbiology Letters 270:1-11

Morris CE, Monier JM (2003) The ecological significance of biofilm formation by plant-associated bacteria. Annu Rev Phytopathol 41:429-453

Müller N, Hempel M, Philipp B, Gross EM (2007) Degradation of gallic acid and hydrolysable polyphenols is constitutively activated in the freshwater plant- associated bacterium Matsuebacter sp FB25. Aquat Microb Ecol 47:83-90

Murray JR, Hackett WP (1991) Dihydroflavonol reductase activity in relation to differential anthocyanin accumulation in juvenile and mature phase Hedera helix L. Plant Physiology 97:343-351

Muyzer G, Dewaal EC, Uitterlinden AG (1993) Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16s ribosomal RNA. Appl Environ Microbiol 59:695-700

Muyzer G, Teske A, Wirsen CO, Jannasch HW (1995) Phylogenetic relationships of Thiomicrospira species and their identification in deep-sea hydrothermal vent samples by denaturing gradient gel electrophoresis of 16S rDNA fragments. Arch Microbiol 164:165-172

149 VIII References

Nakai S, Inoue Y, Hosomi M, Murakami A (2000) Myriophyllum spicatum-released allelopathic polyphenols inhibiting growth of blue-green algae Microcystis aeruginosa. Water Res 34:3026-3032

Neef A (1997) Anwendung der In-situ-Einzelzell-Identifizierung von Bakterien zur Populationsanalyse in komplexen mikrobiellen Biozönosen. PhD thesis, Technical University Munich, Biology Department

Neef A, Amann R, Schlesner H, Schleifer KH (1998) Monitoring a widespread bacterial group: in situ detection of planctomycetes with 16S rRNA-targeted probes. Microbiology 144 3257-3266

Nelson KE, Thonney ML, Woolston TK, Zinder SH, Pell AN (1998) Phenotypic and phylogenetic characterization of ruminal tannin-tolerant bacteria. Appl Environ Microbiol 64:3824-3830

Newman RM (1991) Herbivory and detritivory on freshwater macrophytes by invertebrates: a review. J N Am Benthol Soc 10:89-114

Nikolcheva LG, Cockshutt AM, Barlocher F (2003) Determining diversity of freshwater fungi on decaying leaves: comparison of traditional and molecular approaches. Appl Environ Microbiol 69:2548-2554

Nylund GM, Cervin G, Hermansson M, Pavia H (2005) Chemical inhibition of bacterial colonization by the red alga Bonnemaisonia hamifera. Mar Ecol Prog Ser 302:27-36

Olapade OA, Leff LG (2004) Seasonal dynamics of bacterial assemblages in epilithic biofilms in a northeastern Ohio stream. J N Am Benthol Soc 23:686-700

Olapade OA, Leff LG (2005) Seasonal response of stream biofilm communities to dissolved organic matter and nutrient enrichments. Appl Environ Microbiol 71:2278-2287

Olapade OA, Leff LG (2006) Influence of dissolved organic matter and inorganic nutrients on the biofilm bacterial community on artificial substrates in a northeastern Ohio, USA, stream. Can J Microbiol 52:540-549

Opelt K, Berg C, Berg G (2007) The bryophyte genus Sphagnum is a reservoir for powerful and extraordinary antagonists and potentially facultative human pathogens. FEMS Microbiol Ecol 61:38-53

150 VIII References

Ophel K, Kerr A (1990) Agrobacterium vitis sp. nov. for strains of Agrobacterium biovar 3 from grapevines. Int J Syst Bacteriol 40:236-241

Osawa R (1992) Tannin-protein complex-degrading enterobacteria isolated from the alimentary tracts of koalas and a selective medium for their enumeration. Appl Environ Microbiol 58:1754-1759

Paerl HW, Pinckney JL (1996) A mini-review of microbial consortia: Their roles in aquatic production and biogeochemical cycling. Microb Ecol 31:225-247

Painter DS, McCabe KJ (1988) Investigation into the disappearance of Eurasian watermilfoil from the Kawartha Lakes, Canada. J Aquat Plant Management 26:3-12

Park JK, Shimono K, Ochiai N, Shigeru K, Kurita M, Ohta Y, Tanaka K, Matsuda H, Kawamukai M (1999) Purification, characterization, and gene analysis of a chitosanase (ChoA) from Matsuebacter chitosanotabidus 3001. J Bacteriol 181:6642-6649

Parsek MR, Fuqua C (2004) Biofilms 2003: Emerging themes and challenges in studies of surface-associated microbial life. J Bact 186:4427-4440

Pasmore M, Costerton JW (2003) Biofilms, bacterial signaling, and their ties to marine biology. J Ind Microbiol Biotechnol 30:407-413

Perkins MC, Woods HA, Harrison JF, Elser JJ (2004) Dietary phosphorus affects the growth of larval Manduca sexta. Arch Insect Biochem Physiol 55:153-168

Pernthaler J, Amann R (2005) Fate of heterotrophic microbes in pelagic habitats: Focus on populations. Microbiol Mol Biol Rev 69:440-+

Pernthaler J, Glöckner FO, Schonhuber W, Amann R (2001a) Fluorescence in situ hybridization (FISH) with rRNA-targeted oligonucleotide probes. In: J. P (ed) Methods in Microbiology, Vol 30. Academic Press, San Diego, p 207-226

Peters L, Hillebrand H, Traunspurger W (2007) Spatial variation of grazer effects on epilithic meiofauna and algae. J N Am Benthol Soc 26:78-91

Phillips GL, Eminson D, Moss B (1978) A mechanism to account for macrophyte decline in progressively eutrophicated freshwaters. Aquat Bot 4:103-126

Picioreanu C, Van Loosdrecht MC, Heijnen JJ (2000) Effect of diffusive and convective substrate transport on biofilm structure formation: a two- dimensional modeling study. Biotechnol Bioeng 69:504-515

151 VIII References

Pidiyar VJ, Jangid K, Patole MS, Shouche YS (2004) Studies on cultured and uncultured microbiota of wild Culex quinquefasciatus mosquito midgut based on 16s ribosomal RNA gene analysis. Am J Trop Med Hyg 70:597-603

Porra RJ (1990) The assay of chlorophyll-a and chlorophyll-b converted to their respective magnesium-rhodochlorin derivatives by extraction from recalcitrant algal cells with aqueous alkaline methanol: prevention of allomerization with reductants. Biochi Biophys Acta 1015:493-502

Rao D, Webb JS, Kjelleberg S (2006) Microbial colonization and competition on the marine alga Ulva australis. Appl Environ Microbiol 72:5547-5555

Rasmussen TB, Manefield M, Andersen JB, Eberl L, Anthoni U, Christophersen C, Steinberg P, Kjelleberg S, Givskov M (2000) How Delisea pulchra furanones affect quorum sensing and swarming motility in Serratia liquefaciens MG1. Microbiology 146:3237-3244

Remus R, Ruppel S, Jacob HJ, Hecht-Buchholz C, Merbach W (2000) Colonization behaviour of two enterobacterial strains on cereals. Biol Fertil Soils 30:550-557

Riemann L, Steward GF, Azam F (2000) Dynamics of bacterial community composition and activity during a mesocosm diatom bloom. Appl Environ Microbiol 66:578-587

Roeselers G, Zippel B, Staal M, van Loosdrecht M, Muyzer G (2006) On the reproducibility of microcosm experiments - different community composition in parallel phototrophic biofilm microcosms. FEMS Microbiol Ecol 58:169-178

Roller C, Wagner M, Amann R, Ludwig W, Schleifer KH (1994) In situ probing of gram-positive bacteria with high DNA G + C content using 23S rRNA-targeted oligonucleotides. Microbiology 140 2849-2858

Romani AM, Guasch H, Munoz I, Ruana J, Vilalta E, Schwartz T, Emtiazi F, Sabater S (2004) Biofilm structure and function and possible implications for riverine DOC dynamics. Microbial Ecology 47:316-328

Ryan MP, Pembroke JT, Adley CC (2007) Ralstonia pickettii in environmental biotechnology: potential and applications. J Appl Microbiol 103:754-764

Sabaratnam S, Beattie GA (2003) Differences between Pseudomonas syringae pv. syringae B728a and Pantoea agglomerans BRT98 in epiphytic and endophytic colonization of leaves. Appl Environ Microbiol 69:1220-1228

152 VIII References

Sand-Jensen K (1990) Epiphyte shading - its role in resulting depth distribution of submerged aquatic macrophytes. Fol Geobot Phytotax 25:315-320

Sand-Jensen K, Søndergaard M (1981) Phytoplankton and epiphyte development and their shading effect on submerged macrophytes in lakes of different nutrient status. Int Rev ges Hydrobiol 66:529-552

Sapp M, Wichels A, Wiltshire KH, Gerdts G (2007) Bacterial community dynamics during the winter-spring transition in the North Sea. FEMS Microbiol Ecol 59:622-637

Scalbert A (1991) Antimicrobial properties of tannins. Phytochemistry 30:3875-3883

Scheffer M, Hosper SH, Meijer ML, Moss B, Jeppesen E (1993) Alternative equilibria in shallow lakes. Trends Ecol Evol 8:275-279

Schmieder K (1997) Littoral zone - GIS of Lake Constance: a useful tool in lake monitoring and autecological studies with submersed macrophytes. Aquat Bot 58:333-346

Schuette JL (1996) Lacunar pressures in Myriophyllum heterophyllum: Manometric measurement and diurnal gas dynamics in field populations. Aquat Bot 54:321-336

Schultz JC, Hunter MD, Appel HM (1992) Antimicrobial activity of polyphenols mediates plant-herbivore interactions. In: Hemingway RW, Laks PE (eds) Plant Polyphenols, Vol 59. Plenum Press, New York, p 621-638

Schulz M, Rinke K, Kohler J (2003) A combined approach of photogrammetrical methods and field studies to determine nutrient retention by submersed macrophytes in running waters. Aquat Bot 76:17-29

Schweitzer B, Huber I, Amann R, Ludwig W, Simon M (2001a) Alpha- and beta- proteobacteria control the consumption and release of amino acids on lake snow aggregates. Appl Environ Microbiol 67:632-645

Shabana YM, Cuda JP, Charudattan R (2003) Evaluation of pathogens as potential biocontrol agents of hydrilla. J Phytopathol 151:607-613

Shade A, Kent AD, Jones SE, Newton RJ, Triplett EW, McMahon KD (2007) Interannual dynamics and phenology of bacterial communities in a eutrophic lake. Limnol Oceanogr 52:487-494

153 VIII References

Shelford VW (1918) Conditions of existence. In: Ward HB, Whipple GC (eds) Freshwater Biology. John Wiley & Sons, New York, p 21-60

Sher-Kaul S, Oertli B, Castella E, Lachavanne JB (1995) Relationship between biomass and surface area of six submerged aquatic plant species. Aquat Bot 51:147-154

Simon M, Grossart HP, Schweitzer B, Ploug H (2002) Microbial ecology of organic aggregates in aquatic ecosystems. Aquat Microb Ecol 28:175-211

Smith AH, Mackie RI (2004) Effect of condensed tannins on bacterial diversity and metabolic activity in the rat gastrointestinal tract. Appl Environ Microbiol 70:1104-1115

Sondergaard M (1981) Kinetics of extracellular release of 14C-labelled organic carbon by submersed macrophytes. Oikos 36:331-347

Stachowicz JJ (2001) Mutualism, facilitation, and the structure of ecological communities. Bioscience 51:235-246

Stanley NR, Lazazzera BA (2004) Environmental signals and regulatory pathways that influence biofilm formation. Mol Microbiol 52:917-924

Steinle P, Stucki G, Stettler R, Hanselmann KW (1998) Aerobic mineralization of 2,6- dichlorophenol by Ralstonia sp. strain RK1. Appl Environ Microbiol 64:2566- 2571

Sterner RW, Hessen DO (1994) Algal nutrient limitation and the nutrition of aquatic herbivores. Ann Rev Ecol Syst 25:1-29

Stewart PS, Costerton JW (2001) Antibiotic resistance of bacteria in biofilms. Lancet 358:135-138

Stewart TW, Miner JG, Lowe RL (1998) Quantifying mechanisms for zebra mussel effects on benthic macroinvertebrates: organic matter production and shell- generated habitat. J North Am Benthol Soc 17:81-94

Stoner AW (1980) The role of seagrass biomass in the organization of benthic macrofaunal assemblages. Bull Mar Sci 30:537-551

Sturz AV, Kimpinski J (2004) Endoroot bacteria derived from marigolds (Tagetes spp.) can decrease soil population densities of root-lesion nematodes in the potato root zone. Plant and Soil 262:241-249

154 VIII References

Styp von Rekowski K, Hempel M, Philipp B (2008) Quorum sensing by N- acylhomoserine lactones is not required for Aeromonas hydrophila during growth with organic particles in lake water microcosms. Arch Microbiol 189:475-482

Tadonléké RD (2007) Strong coupling between natural Planctomycetes and changes in the quality of dissolved organic matter in freshwater samples. FEMS Microbiol Ecol 59:543-555

Underwood GJC (1991) Colonization and invasion of leaves of the aquatic macrophyte Ceratophyllum demersum L by epiphytic bacteria. Microb Ecol 21:267-275

Urabe J, Watanabe Y (1992) Possibility of N-limitation or P-limitation for planktonic Cladocerans - an experimental test. Limnol Oceanogr 37:244-251

Vigliotta G, Nutricati E, Carata E, Tredici SM, De Stefano M, Pontieri P, Massardo DR, Prati MV, De Bellis L, Alifano P (2007) Clonothrix fusca Roze 1896, a filamentous, sheathed, methanotrophic gamma-proteobacterium. Appl Environ Microbiol 73:3556-3565

Wagner M, Amann R, Lemmer H, Schleifer KH (1993) Probing activated sludge with oligonucleotides specific for proteobacteria: inadequacy of culture-dependent methods for describing microbial community structure. Appl Environ Microbiol 59:1520-1525

Walenciak O (2004) Impact of polyphenols from Myriophyllum spicatum on its herbivore Acentria ephemerella and associated microorganisms. Doctoral thesis, University of Constance

Walenciak O, Zwisler W, Gross EM (2002) Influence of Myriophyllum spicatum- derived tannins on gut microbiota of its herbivore Acentria ephemerella. J Chem Ecol 28:2045-2056

Wallner G, Amann R, Beisker W (1993) Optimizing fluorescent in situ hybridization with rRNA-targeted oligonucleotide probes for flow cytometric identification of microorganisms. Cytometry 14:136-143

Wang J, Jenkins C, Webb RI, Fuerst JA (2002) Isolation of Gemmata-like and Isosphaera-like planctomycete bacteria from soil and freshwater. Appl Environ Microbiol 68:417-422

155 VIII References

Wang YH, Zhang LQ, Li JY, Wang JH, Wang HM (2008) The quorum-sensing system AvsR-AvsI regulates both long-chain and short-chain acyl-homoserine lactones in Agrobacterium vitis E26. Antonie Van Leeuwenhoek 93:267-273

Weiss P, Schweitzer B, Amann R, Simon M (1996) Identification in situ and dynamics of bacteria on limnetic organic aggregates (Lake Snow). Appl Environ Microbiol 62:1998-2005

Wetzel RG (1993) Microcommunities and microgradients: Linking nutrient regeneration, microbial mutualism and high sustained aquatic primary production. Netherlands Journal of Aquatic Ecology 27:3-9

Wetzel RG, Hatcher PG, Bianchi TS (1995) Natural photolysis by ultraviolet irradiance of recalcitrant dissolved organic matter to simple substrates for rapid bacterial metabolism. Limnol Oceanogr 40:1369-1380

White TCR (1993) The inadequate environment. Nitrogen and the abundance of animals, Vol. Springer Verlag, Berlin

Wicks S, Hempel M, Gross EM (submitted) Single- and multispecies biofilm formation of tannin degrading bacteria on an aquatic macrophyte.

Wium-Andersen S, Anthoni U, Christophersen C (1982) Allelopathic effects on phytoplankton by substances ieolated from aquatic macrophytes (Charales). Oikos 39:187-90

Wright SA, Zumoff CH, Schneider L, Beer SV (2001) Pantoea agglomerans strain EH318 produces two antibiotics that inhibit Erwinia amylovora in vitro. Appl Environ Microbiol 67:284-292

Wu ZB, Deng P, Wu XH, Luo S, Gao YN (2007a) Allelopathic effects of the submerged macrophyte Potamogeton malaianus on Scenedesmus obliquus. Hydrobiologia 592:465-474

Wu ZB, Zhang SH, Wu XH, Cheng SP, He F (2007b) Allelopathic interactions between Potamogeton maackianus and Microcystis aeruginosa. Allelopathy J 20:327-338

Zeida M, Wieser M, Yoshida T, Sugio T, Nagasawa T (1998) Purification and characterization of gallic acid decarboxylase from Pantoea agglomerans T71. Appl Environ Microbiol 64:4743-4747

156 VIII References

WEB PAGES Probe Base http://www.microbial-ecology.net/probebase BLAST http://www.ncbi.nlm.nih.gov/BLAST http://www.pharmainfo.net

157 VIII References

158 RECORD OF ACHIEVEMENT - ABGRENZUNG DER EIGENLEISTUNG

Data presented in this study are based on my experimental design, performance and supervision. I performed sampling, evaluated and analyzed data with the following exceptions:

Chapter 2: Sampling and data collection was performed by Maja Blume. The final data analysis was performed by me.

Chapter 4: The plant chemistry was measured as part of the routine sampling campaign and data were kindly provided by E. Gross. Data on initial colonization of Myriophyllum spicatum were provided by S. Wicks

Chapter 6: Data from this Chapter were derived out of the diploma thesis by S. Wicks.

159 160 LIST OF PUBLICATIONS

Wicks S, Hempel M, Gross EM (submitted) Single- and multispecies biofilm formation of tannin-degrading bacteria on an aquatic macrophyte

Hempel M and Gross EM (submitted) Spatio–temporal dynamics of the bacterial biofilm on two freshwater macrophytes and an artificial substrate in Lower Lake Constance

Hempel M, Grossart HP, Gross EM (submitted) Bacterial community composition of biofilms on two submerged macrophytes and an artificial surface in Lake Constance

Premke K, Hempel M, Rothhaupt KO, Fischer P (submitted) Ecological studies on the decomposition rate of fish carcasses by benthic organisms in the littoral zone of Lake Constance, Germany

Hempel M, Blume M, Blindow I, Gross EM (2008) Epiphytic bacterial community composition on two common submerged macrophytes in brackish water and freshwater BMC Microbiology 8:58

Styp von Rekowski K, Hempel M, Philipp B (2008) Quorum sensing by N -acylhomoserine lactones is not required for Aeromonas hydrophila during growth with organic particles in lake water microcosms. Archives of Microbiology 189:475-482

Müller N, Hempel M, Philipp B, Gross EM (2007) Degradation of gallic acid and hydrolysable polyphenols is constitutively activated in the freshwater plant-associated bacterium Matsuebacter sp. FB25. Aquatic Microbial Ecology 47:83-90

Salzburger W, Brandstätter A, Gilles A, Parson W, Hempel M, Sturmbauer C, Meyer A (2003) Phylogeography of the vairone (Leuciscus souffia, Risso 1826) in Central Europe. Molecular Ecology 12(9):2371-86

161 162 CONFERENCE CONTRIBUTIONS

Biofilms on submerged macrophytes in Lake Constance, Germany: Which factors structure abundance and community composition? Symposium for Aquatic Microbial Ecology (SAME), Faro, Portugal 2007 (Poster).

Räumliche Zusammensetzung des heterotrophen Biofilms auf Myriophyllum spic- atum L. Deutsche Gesellschaft für Limnologie (DGL), Dresden 2006 (Vortrag).

Effect of secondary metabolites from Myriophyllum spicatum on biofilm formation and composition on a spatial and temporal scale. Vereinigung allgemeiner und angewandter Mikrobiologie (VAAM), Jena 2006 (Poster).

Wirkung sekundärer Metabolite aus Myriophyllum spicatum auf Bildung und Zu- sammensetzung des heterotrophen epiphytischen Biofilms. Deutsche Gesellschaft für Limnologie (DGL), Karlsruhe 2005 (Poster).

FURTHER COOPERATION WITHIN THE CRC 454

With J. Hesselschwerdt (Project A6) on microbial aspects of leaf litter degradation in the littoral zone of Lake Constance.

163 164 DANKE Diese Arbeit wurde in der Arbeitsgruppe von Frau PD Dr. E. Groß an der Universität Konstanz in der Zeit von Januar 2005 bis April 2008 durchgeführt.

Mein besonderer Dank gilt:

Frau PD Dr. E.M. Groß für die Vergabe des Themas, die Begutachtung der Arbeit, die angenehme Betreuung, gute Diskussionen und die unbezahlbare Hilfe in den letzten Wochen. Ich habe nicht nur fachlich viel von ihr lernen können.

Herrn PD Dr. H.-P. Grossart für die fachliche Betreuung während meiner Neuglobsow-Aufenthalte, Ratschläge und Tipps zur Auswertung, sowie die Begutachtung dieser Arbeit.

Claudia Feldbaum für das unermüdliche Messen der Pflanzenchemie und aufmunternde Worte.

Sonja Wicks, die so mutig war und sich bereit erklärte, unter meiner Regie ihre Diplomarbeit anzufertigen. Für die gute Zusammenarbeit und das in mich gesetzte Vertrauen bedanke ich mich besonders.

Stefanie Nadj und Benedikt Mothes für die Geduld und Zuverlässigkeit bei Routineaufgaben.

Allen Instituts- und Universitätsmitarbeitern hier und in Neuglobsow, die darüber hinaus zum Gelingen dieser Arbeit beigetragen haben. Stellvertretend den Kollegen von Z8 und U, S. Berger, K. Huppertz, M. Wolf, M. Allgaier, S. Bruckner, C. Koppe und vielen anderen aus Beschaffungsabteilung und Werkstatt.

Stefan und John: Es waren drei schöne Jahre mit Euch und Ihr werdet mir sehr fehlen. Die Arbeit mit Euch, egal ob Büro oder Freiland, war ein großes Erlebnis und immer wieder ein Riesenspaß.

Sarah, Jutta, Steffi und René für die gute Freundschaft, den Austausch, die Kollegialität und gute Nachbarschaft.

Meinen Eltern, Kai und Ann-Kristin und der restlichen Familie, die meine „Gewässerrandforschung“ aufmerksam verfolgt haben, für nimmer endendes Interesse und ihre Unterstützung. Um die brennendste Frage zu beantworten: „Ja, ich habe Ergebnisse!“

Thomas. Für all die Liebe, Hilfe und Unterstützung, ohne die ich mich so manches Mal verrannt hätte. Und für F&P.

165 166 Curriculum vitae

Persönliche Daten

Name: Hempel Vorname: Melanie Geburtsdatum: 17. Juni 1980 Geburtsort: Detmold, Deutschland

Aktuelle Tätigkeit Seit Januar 2005 Wissenschaftliche Angestellte der Universität Konstanz in der Gruppe von Frau PD Dr. E. Groß (Chemische Ökologie) mit dem Ziel der Promotion zu dem Thema: „Community composition and interactions of biofilm bacteria on submerged freshwater macrophytes“

Studium 01/2004 – 10/2004 Diplomarbeit zu dem Thema: „Untersuchungen zur Bedeutung AHL – vermittelter Zell-Zell-Kommunikation heterotropher Bakterien im Bodenseelitoral“ bei Dr. B. Philipp am Lehrstuhl „Mikrobielle Ökologie“ (Prof. Dr. B. Schink) der Universität Konstanz 10/2001 – 10/2004 Hauptstudium Biologie an der Universität Konstanz; Abschluss: Diplom (Note: 1,9; Fächer: Gewässermikrobiologie, Limnologie, Phytopathologie, Evolutionsbiologie) 10/1999 – 09/2001 Grundstudium Biologie an der Universität Bielefeld; Abschluss: Vordiplom (Note: 1,8; Fächer: Botanik und Limnologie)

Schulischer Werdegang 1997 – 1999 Besuch des Stadtgymnasiums Detmold; Abschluss: Abitur (Note: 1,9) 1996 – 1997 Schulaufenthalt in den USA (Ohio & Oklahoma) 1990 – 1996 Besuch des Stadtgymnasiums Detmold 1986 – 1990 Besuch der Grundschule Heiligenkirchen, Detmold

Berufsbezogene Tätigkeiten 01/2005 – heute Betreuung einer Diplomarbeit an der Universität Konstanz Leitung und Organisation eines S1 – Labors Vorlesung zu dem Thema „Molekularbiologische Methoden in der mikrobiellen Ökologie“ Betreuung und Einweisung studentischer Hilfskräfte Betreuung dreier Hauptstudiumskurse in der Limnologie 10/2004 – 12/2004 Praktikum bei der GATC Biotech AG, Konstanz, im Bereich „Custom Sequencing“ 10/2004 – 12/2004 Studentische Hilfskraft am Lehrstuhl für „Mikrobielle Ökologie“ von Prof. Dr. B. Schink an der Universität Konstanz; abschließ- ende Arbeiten, Betreuung eines DAAD – Studenten (Deutscher Akademischer Austausch Dienst) 07/2004 – 08/2004 Studentische Hilfskraft am Lehrstuhl für „Mikrobielle Ökologie“ von Prof. Dr. B. Schink zur Betreuung des Hauptstudiumkurses „Limnische Mikrobiologie“ an der Universität Konstanz 167 06/2004 – 10/2004 Werksstudent bei der GATC Biotech AG, Konstanz Sommersemester Besuch der Weiterbildung „Einführung in die Sicherheitsproble- 2004 matik der Gentechnik“ an der Universität Konstanz 03/2004 – 10/2004 Studentischer Vertreter des IAESTE – Komitees (International Association for the Exchange of Students for Technical Experience) an der Universität Konstanz; Betreuung auslän- discher Studierender und Vermittlung deutscher Studenten ins Ausland 10/2003 – 12/2003 Studentische Hilfskraft am Lehrstuhl für „Evolutionsbiologie“ bei Prof. Dr. A. Meyer an der Universität Konstanz; DNA – Extraktion, PCR – Aufreinigung und allgemeine Laborarbeiten 07/2003 – 08/2003 Studentische Hilfskraft am Lehrstuhl für „Mikrobielle Ökologie“ von Prof. Dr. B. Schink an der Universität Konstanz zur Betreuung des Hauptstudiumkurses „Limnische Mikrobiologie“ 01/2003 – 03/2003 Studentische Hilfskraft in der Arbeitsgruppe von Dr. E. v. Elert am Limnologischen Institut der Universität Konstanz; Präparation von Schnecken und Messungen zur enzymatischen Darm- aktivität 06/2002 – 09/2002 Studentische Hilfskraft am Lehrstuhl für „Mikrobielle Ökologie“ bei Prof. Dr. B. Schink an der Universität Konstanz unter der Betreuung von Dr. B. Philipp. Etablierung eines Systems zur Quantifizierung von Biofilmen (Kolter – Assays)

Sprachkenntnisse Deutsch: Muttersprache Englisch: fließend in Wort und Schrift

Labortätigkeit Molekularbiologische Methoden: DNA – Extraktion, PCR, DGGE, Klonierungen Chemische Analysen: Photometrische– und HPLC – Verfahren Mikroskopie: Färbungen mit DAPI & SYBR Green, Fluoreszenz in situ Hybridisierung Allgemeines: steriles Arbeiten; Pflege von Stammkulturen (Pflanzen und Bakterien); Ansetzen von Medien und Puffern

Computerkenntnisse MS Office; Internet Recherche; EndNote/Reference Manager; SigmaStat; SigmaPlot; Primer; JMP; ARB; BioEdit; GelCompar; CorelDraw; LINUX; LIMS

Weiteres Sommer 2005 Bodenseeschifferpatent 2000 – 2001 Sitz im Jugendausschuss der ev. ref. Kirchengemeinde Heiligen- kirchen 1998 Ausbildung zum Ehrenamt in der Kinder- und Jugendarbeit. Kurs der Jugendzentrale der Lippischen Landeskirche 1997 Fahrausweis der Klasse 3 für Personenwagen 1994 – 2001 Betreuung eine Kindergruppe des 1. & 2. Schuljahres in der ev. ref. Kirchengemeinde Heiligenkirchen

168 169

170