Nutritional Aspects in the Invasive Freshwater Bivalve Corbicula Fluminea: the Role of Essential Lipids

Nutritional aspects in the invasive freshwater bivalve Corbicula fluminea: The role of essential lipids

Dissertation

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

an der

Mathematisch-Naturwissenschaftliche Sektion Fachbereich Biologie

vorgelegt von

Basen, Timo

Tag der mündlichen Prüfung: 27.07.2012 1. Referent: Prof. Dr. Karl-Otto Rothhaupt 2. Referent: Prof. Dr. Alexander Wacker

ich was suchen ich nicht wissen was suchen ich nicht wissen wie wissen was suchen ich suchen wie wissen was suchen

ich wissen was suchen ich suchen wie wissen was suchen ich wissen ich suchen wie wissen was suchen ich was wissen

Ernst Jandl

Table of contents Chapter 1 General introduction 1 Aims and objectives 8 Chapter 2 Role of essential lipids in determining food quality for the invasive freshwater clam Corbicula fluminea 11 Abstract 11 Introduction 12 Methods 13 Results 17 Discussion 21 Acknowledgements 25 Chapter 3 Absence of sterols constrains food quality of cyanobacteria for an invasive freshwater bivalve 27 Abstract 27 Introduction 28 Materials and methods 29 Results 32 Discussion 36 Acknowledgements 37 Chapter 4 Impact of temperature and seston dynamics on growth and survival of Corbicula fluminea: A field study in Lake Constance 39 Abstract 39 Introduction 40 Materials and methods 42 Results 48 Discussion 65 Acknowledgements 72 Supplementary data 72 Chapter 5 Phytoplankton food quality effects on gammarids: benthic-pelagic coupling mediated by an invasive freshwater clam 73 Abstract 73 Introduction 74

Materials and methods 75 Results 79 Discussion 85 Acknowledgements 88 Supplementary data 88 Chapter 6 Concluding remarks and perspectives 89 Abstract 95 Zusammenfassung 97 Bibliography 101 Appendix 119 Record of achievement / Abgrenzung der Eigenleistung 124 Danksagung 125 Curriculum Vitae 126 List of publications 127

1 General introduction

Chapter 1

General introduction

Autecology of Corbicula fluminea The Asian clam Corbicula fluminea (OF Müller 1774) is a hermaphroditic self fertilizing clam with a life span of 1 - 5 years (McMahon 2002; Sousa et al. 2008). The clams start to reproduce at a shell size of about 10 mm (i.e. at the age of 5 - 9 months); between one and three reproduction cycles are possible depending to ecosystem conditions. Adult Corbicula have a high fecundity and conduct brood hatchery (Kraemer and Galloway 1986; Araujo et al. 1993), whereat larvae are kept in inner demibranches until they reach a size of approximately 250 µm. The larvae (Fig. 1c) are then released into the water column and settle to the ground within 24 h, i.e. the time span for drift distribution is limited. A temperature tolerance of > 2 and < 36 °C has been reported (Britton and Morton 1979; Karatayev et al. 2005), but longer periods with temperatures about 3 °C (Werner and Rothhaupt 2008) and times with low food supply and high temperatures during summer months have been shown to cause high mortality rates in C. fluminea (Weitere et al. 2009; Vohmann et al. 2010). At a temperature of 10 - 11 °C growth and reproduction are possible (Karatayev et al. 2005). Clams prefer habitats with soft sediments (small gravel, sand) which are well oxygenated and contain high proportions of organic matter (Hakenkamp and Palmer 1999; Vaughn and Hakenkamp 2001). Juvenile and adult clams are capable of deposition feeding, the so called ‘pedal feeding’ (Reid et al. 1992), which is considered as the primary mechanism of larval nutrition until gill structures and in- and exhalant siphons are developed properly. However, although adult Corbicula are still able to take up food via its muscular foot, filter feeding clearly is the main process of food uptake in the Asian clam.

a b c

Fig. 1: Different stadia of Corbicula fluminea. Adult clams with closed shells (a), with expelled siphons and muscular foot (b) and D-shaped larvae (size approximately 300 µm; c).

1 1 General introduction

The clams circulate water for respiration and feeding, and remove particles from the water, which are either consumed or bound as pseudofaeces and expelled. Corbicula fluminea can effectively remove detritus, bacteria and algae from the water column but is regarded as a non-selective suspension feeder (Lauritsen 1986; Way et al. 1990; Boltovskoy et al. 1995). Preferred size range for ingestion is known to be between 1 μm and 20 - 25 μm (Way et al. 1990; McMahon and Bogan 2001), with reported maximal values of up to 170 μm (Boltovskoy et al. 1995).

Ecological impacts of C. fluminea Bivalves can dramatically alter the structure of benthic communities. This was shown for the zebra mussel Dreissena polymorpha in Lake Constance in the 1960s (Mörtl and Rothhaupt 2003) and in other freshwater ecosystems in Europe and North and South America (Stanczykowska 1977; MacIsaac et al. 2002). One contributing factor is the physical alteration of habitat structure, the so-called ecosystem engineering (Jones et al. 1994). Bivalve shells can provide shelter and substrate for other benthic species (Strayer et al. 1999; Crooks 2002; Sousa et al. 2008). Burrowing bivalves, like C. fluminea, can impact benthic processes such as nutrient and organic matter cycling in sediments (Hakenkamp and Palmer 1999; Vaughn and Hakenkamp 2001). Additionally, filter feeding bivalves are often considered as important benthic-pelagic couplers, which remove large amounts of seston (algae, bacteria, particulate organic carbon) from the water column and transfer these resources to the benthos as biodeposition material (faeces and pseudofaeces), thereby stimulating benthic productivity (Hakenkamp et al. 2001; Gergs et al. 2009). Filtration processes mediated by bivalves can severely disturb the recruitment of other bivalves species (Hakenkamp and Palmer 1999), when larval swimming stadia are ingested (Strayer et al. 1999). Additionally, a high filtration rate can regulate pelagic nutrient cycling (Cohen et al. 1984; Cahoon and Owen 1996; Hwang et al. 2011), reduce eutrophication processes (Phelps 1994; McMahon 2002) and increase water clarity and therefore promote submerged vegetation (Phelps 1994). Moreover, the excretion of inorganic nutrients to the water column by filter feeding may play an important role in stimulating primary production (Yamamuro and Koike 1993; Vaughn et al. 2008). Filtration, nutrient excretion and benthic- pelagic coupling are regarded as the main water column processes completed by C. fluminea (Lauritsen and Mozley 1989; Vaughn and Hakenkamp 2001). A high filtration rate favours Corbicula in competition with other bivalves (Sousa et al. 2005), which has been shown for German water ways, where Corbicula massively spread and quickly replaced D. polymorpha as dominant mollusc (Tittizer et al. 2000; Bachmann et al. 2001)

2 1 General introduction

Fig. 2 Map of Lake Constance (after IGKB 2010). Numbers indicate the site of first occurrence of Corbicula fluminea in the lake in 2003 (1) and the maximal distribution on the northern shore up to Immenstaad (2) and on the southern shore up to Altnau (4) in 2010. An isolated discovery was detected in the “Überlinger See” in 2008 (3). In 2011 first observations of C. fluminea were recorded in the “Seerhein” near the city of Konstanz (5).

Invasion history of C. fluminea Since its introduction into many aquatic freshwater systems all over the world during the last decades, C. fluminea has undergone a remarkable range expansion to become a ubiquitous benthic invertebrate in freshwater ecosystems. Originating from Southeast Asia, C. fluminea was introduced to North America (McMahon 1982) and South America (Darrigran 2002) in the early 20th century. In the 1980s Corbicula spp. invaded Europe, presumably via ballast water used in ships originating from North America (Mouthon 1981), and this event was followed by a rapid expansion in European inland waters (Den Hartog et al. 1992; Araujo et al. 1993). In 1988 it was also first detected in the River Rhine delta (Bij de Vaate and Greijdanus-Klaas 1990). Within years it inhabited the whole navigable River Rhine up to the Swiss border (Turner et al. 1998; Tittizer et al. 2000). The first individuals of C. fluminea in Lake Constance were discovered at the Rohrspitz (Austria) in 2003 (Fig. 2; position1; Werner and Mörtl 2004). In the following years the clams spread within the eastern part of the lake. Further occurrence of C. fluminea was discovered at the southern shore (Switzerland) and at the northern shore (Austria, Germany) of Lake Constance. In 2008, C. fluminea spread along the northern shore up to Friedrichshafen (Fig. 2, position2), an isolated appearance in the western part of the lake close to Konstanz-Egg (3) was also discovered. The maximum

3 1 General introduction western distribution in 2010 reached the axis Altnau (4) – Friedrichshafen (2). In 2011, first observations of C. fluminea were reported near the city of Konstanz (“Seerhein”, 5). A further establishment of C. fluminea in the lower Lake Constance (Untersee) is merely a matter of time.

Lake Constance is an oligotrophic pre-alpine lake, situated in Middle Europe, with Germany, Switzerland and Austria as riparian states. With a maximum depth of 253 m, a surface of 473 km², and ~50 km³ water volume, Lake Constance is one of the largest and deepest lakes in Middle Europe. The lake consists of two basins, the larger and deeper Upper Lake and the smaller and shallower Lower Lake Constance. The major contributing river is the Alpine Rhine, located in the eastern part, in Austria. First specimens of C. fluminea in Lake Constance were found in a shallow, sandy bay between the old and new Alpine Rhine river bed (Fig. 2, position1; Fig. 3; Werner and Mörtl 2004).

The C. fluminea population in Lake Constance is characterized by slow growth, a reduced maximum shell size, and only one reproductive period per year (Werner and Rothhaupt 2008). At the Rohrspitz (Austria), the site of maximum density, up to 90% of total biomass of littoral community was represented by C. fluminea (Werner and Rothhaupt 2007).

a b

c d e

Fig. 3: Impressions from the sampling site of Corbicula fluminea at the Rohrspitz (Austria; a), collected clams with gravel and debris (b) and underwater pictures of typical clam habitats (c-e; © 2012 John Hesselschwerdt - RHEOS).

4 1 General introduction

Many invasions in Lake Constance took place in the last decades, with successful invasions of gastropods (Viviparus ater), crayfish (Orconectes limosus), amphipods (Dikerogammarus villosus), mysids (Limnomysis benedeni ) and bivalves like D. polymorpha and C. fluminea (for details see Hanselmann 2011). Most non-native species do not successfully establish or have only little impact on communities. However, the introduction of new species to ecosystems is potentially associated with a loss of diversity and ecosystem stability (Sala et al. 2000; Chandra and Gerhardt 2008; Strayer 2010). In the last decades, the spread of invasive species in new freshwater ecosystems rapidly increased (Richardson and Pysek 2008). Especially in lakes, invaders were identified as the main cause of extinction of native species (Lodge 2001; Strayer 2010). Invasive species can disturb food web processes, introduce diseases and parasites and alter the natural species composition (Tittizer et al. 2000; Ellis et al. 2011; Poulin et al. 2011). Hence, it is important to understand the factors which are responsible for the successful geographic spread of these invasive species in order to predict possible expansions and to assess ecosystem consequences.

How to be an invasive clam? One main factor limiting expansion of C. fluminea is water temperature. It has been suggested that anthropogenic increase of minimum temperature in winter, caused by power plants (cooling water outflow), can favour the successful establishment in German water ways (Schöll 2000). Additionally, climate change is known to favour species adapted to warmer temperatures or prone to extreme cold temperatures (Mooij et al. 2005). For the pelagic food chain severe effects are postulated to show up with global warming, e.g. increase in frequency of cyanobacterial bloom formation (Paerl and Huisman 2008; Wagner and Adrian 2009). Climate scenarios with rising temperatures and increased atmospheric

CO2 supplies are expected to favour cyanobacterial dominance, because cyanobacteria potentially have a competitive advantage over other phytoplankton groups in coping with the expected increase in abiotic challenges (Jöhnk et al. 2008; Paerl and Huisman 2009; Wagner and Adrian 2009; Paerl et al. 2011). Increased periods of thermal stratification have been suggested to shift phytoplankton assemblages towards a higher proportion of cyanobacteria capable of N2-fixation, such as Aphanizomenon or Anabaena, which may also affect predator-prey interactions in aquatic food webs (Wagner and Adrian 2011). Cyanobacterial blooms can be associated with hazards to human health and livestock and reduced quality of water bodies (Carmichael 1992; Codd 1995; Paerl et al. 2011). Therefore, it is important to investigate the consequences of cyanobacterial mass development for the aquatic food web.

The elementary process in food webs is the transfer of energy from one trophic level to the next. Most important for the carbon flow in aquatic food webs is the plant-herbivore interface, at which the transfer efficiency is highly variable and affected by many factors. Structure and morphology of planktonic cells can affect food quality (DeMott et al. 2001; Van Donk et al. 2011). Digestion efficiency is also a factor determining food quality, e.g. when cells are imbedded in gelatinous layers (Van Donk and Hessen 1993; Van Donk et al. 2011). However, not only morphology but also the compounds within the cells have to be

5 1 General introduction considered in the context of food quality. Secondary metabolites, such as toxins, affect primary consumers when feeding on toxic strains of phytoplankton (Lampert 1987; Carmichael 1992). Various taxa of eukaryotic algae and especially cyanobacteria produce toxins that greatly reduce their nutritional value for herbivores. Selective feeding bivalves are known to promote blooms of toxin producing cyanobacteria and additionally increase the benthic pelagic coupling, due to high rates of pseudofaeces production and low digestion rates (Vanderploeg et al. 2001; Pires et al. 2005; Bontes et al. 2007). Besides toxins, protease inhibitors may also reduce the digestibility of ingested food sources and thus may reduce food quality for filter feeding cladocerans (Schwarzenberger et al. 2010).

Furthermore, the lack or low availability of essential nutrients can constrain the quality of food. Among mineral nutrients, nitrogen and phosphorus are of particular importance in aquatic ecosystems (Elser et al. 2000; Sterner et al. 2008). In marine ecosystems, nitrogen is known as the main limiting parameters for energy transfer between trophic levels (Tyrell 1999), whereas in freshwater systems phosphorus is considered to be the most important limiting nutrient (Elser et al. 2000). Somatic growth especially depends on the supply with phosphorus, because growth is linked to the protein synthesis rate and thus to the concentration of phosphorus-rich ribosomal RNA (Elser et al. 2003a). In autotrophs the C:N:P ratio is known to be highly variable, whereas consumers tend to maintain homeostasis, i.e. relatively constant body C:N:P ratios (Hessen and Lyche 1991; Frost et al. 2002; Elser et al. 2003b). Thus, insufficient availability of N and P for herbivores, i.e. high C:N and C:P ratios in primary producers, can lead to limitation of herbivore growth (Urabe and Watanabe 1992; Sterner 1993). However, most of the studies in freshwater ecosystems focused on interactions between planktonic organisms. Within the last years, stoichiometric demands in the benthic food web has come into focus (Frost et al. 2002; Frost et al. 2003; Cross et al. 2005; Fink et al. 2006; Gergs and Rothhaupt 2008). However, the importance of N and P limitation in benthic food webs is still barely understood, especially for bivalves. In addition, new challenges in food webs mediated by climate change or introduction of non-native species disturb established patterns and therefore might influence micronutrient effects.

Reports on food quality effects on freshwater bivalves, apart from those mediated by mineral nutrient stoichiometry, are scarce. In a previous attempt, Foe and Knight (1986) examined the growth of C. fluminea on six genera of green algae provided in various combinations, and recorded positive tissue growth on different mixtures of Ankistrodesmus, Chlamydomonas, Chlorella, and Scenedesmus, but not for unialgal food sources or mixtures containing Pedinomonas or Selenastrum. However, requirements of Corbicula for single nutrients have not been further characterized. Corbicula, like filter feeding daphnids, are known to be non-selective filter feeders and thus are unable to select food particles with regard to their quality. Therefore, they might respond particularly sensitively to the predominance of nutritionally inadequate food sources (e.g. lipid or micronutrient depleted algae). Research on marine bivalves used for aquaculture revealed that growth and reproduction of bivalves is significantly affected by dietary lipids (e.g. Delaunay et al. 1993; Soudant et al. 1996a; Knauer et al. 1999; Ben Kheder et al. 2010). For the freshwater bivalve

6 1 General introduction

D. polymorpha, Wacker and Von Elert (2002; 2003; 2004) reported evidence that growth, survival and juvenile recruitment are affected by the dietary essential fatty acid supply, especially the supply with long chain polyunsaturated fatty acids (PUFAs) like eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA). PUFAs play an important role in the physiology of animals (Cook 1996). They are components of cell membranes and precursors of bioactive molecules, such as eicosanoids (Stanley-Samuelson 1994), and cannot be synthesized by animals de novo, and thus are essential for their nutrition.

Sterols are another class of lipids that are required for a multitude of physiological processes, e.g. as indispensable components of cell membranes, as precursors for vitamin D and steroid hormones and for regulation of signal transduction (for more detail see Martin- Creuzburg and Von Elert 2009). The ability to synthesize sterols de novo is related to eukaryotic cells, whereas in prokaryotes like cyanobacteria, sterols are usually absent (Volkman 2003; Summons et al. 2006). Cholesterol is the predominant sterol found in animals (Goad 1981). In aquatic ecosystems the role of sterols for food quality was first investigated in filter feeding zooplankton. Arthropods are not capable of synthesising sterols de novo and thus require a dietary source of sterols for growth and reproduction (Martin- Creuzburg and Von Elert 2009). The absence of dietary sterols impairs the fitness of filter feeding zooplankton, as has been shown by sterol supplementation of sterol deficient food sources (Hassett 2004; Martin-Creuzburg and Von Elert 2004). So far, sterol demands in benthic invertebrates have not been investigated. Experiments with marine bivalves, mostly with species used in aquaculture, suggest that the ability to synthesize sterols de novo is generally low or absent among bivalve species, which implies that a dietary source of sterols is necessary for growth (Voogt 1975; Soudant et al. 2000; Park et al. 2002). Data on sterol requirements of freshwater bivalves have not been collected yet.

7 1 General introduction

a b

d e f

Fig. 4 Impressions of experimental setups in field and laboratory experiments: (a) plastic boxes used for the determination of clam growth in Lake Constance (b); different algae (d) and clam cultivation vessels (c, e, f) used in laboratory experiments.

Aims and objectives To elucidate the energy flow in freshwater ecosystems it is important to understand the factors affecting the efficiency of energy transfer across trophic levels. My work was focused on food quality effects on the pelagic-benthic coupling, with key aspect on the nutrition of the benthic freshwater clam C. fluminea.

The thesis starts with two laboratory studies about the role of essential biochemicals in clam nutrition. At first, nutritional requirements of C. fluminea were investigated in a standardized experimental setup, where food and clam tissue parameters (fatty acids, sterols, elemental stoichiometry) were estimated and related to somatic growth rates of individual clams (chapter 2). Phytoplankton species were used which differed in their morphology and in their biochemical composition to assess the significance of different food quality constraints for the growth of clams.

8 1 General introduction

Based on these results, the role of sterols for growth of C. fluminea became a focal point of this study. In a second experimental approach clams were fed with different sterol containing eukaryotic algae and sterol-free cyanobacteria diets (chapter 3). A supplementation method was developed to enrich cyanobacteria with sterols to be able to investigate whether growth conditions for C. fluminea are constrained by a deficiency in dietary sterols.

To assess the significance of environmental factors on growth and survival of C. fluminea in Lake Constance, an experimental field study was performed in Lake Constance during the year 2010 (chapter 4). I hypothesized that seasonal variations in water temperature, phytoplankton succession and seston nutrient composition influence clam growth. Additionally, concomitant laboratory experiments were conducted to study the influence of temperature on clam growth rates and to compare the physiological states of C. fluminea collected from the field in different phases during the season.

Corbicula fluminea is an important pelagic-benthic coupler and thus might be able to transfer not only energy but also essential nutrients to detritivorous benthic invertebrates when they feed on clam-generated biodeposition material. Therefore, I explored the mechanisms responsible for growth and survival of a benthic invertebrate, i.e. Gammarus roeselii, which I fed with different sestonic phytoplankton species and clam generated biodeposition material. The benthic-pelagic coupling mediated by C. fluminea was investigated in an experimental food chain in laboratory experiments (chapter 5), where the role of bioavailability, micronutrients and biochemicals in pelagic autotrophs and clam generated biodeposition materials for gammarid growth and survival was examined.

9 1 General introduction

10 2 Effect of food quality on clam growth

Chapter 2

Role of essential lipids in determining food quality for the invasive freshwater clam Corbicula fluminea

Timo Basen, Dominik Martin-Creuzburg and Karl-Otto Rothhaupt

Freshwater Science (formerly Journal of the North American Benthological Society) doi: 10.1899/10-087.1

Received: 23 June 2010

Accepted: 16 March 2011

Abstract The invasive clam Corbicula fluminea has become a widespread benthic invertebrate in many freshwater ecosystems throughout Europe and North and South America. Invasive bivalves can dramatically alter the structure of native benthic communities, so understanding the factors responsible for successful invasion is important. We investigated C. fluminea nutritional requirements for essential lipids in a standardized growth experiment. Juvenile clams were fed different cyanobacteria (Aphanizomenon flos-aquae, Anabaena variabilis, Synechoccocus elongatus) or eukaryotic algae (Scenedesmus obliquus, Cryptomonas sp.). Somatic growth rates were then correlated with elemental (C:N and C:P) and biochemical (sterol and fatty acid content) components of the food sources and clam tissue. Somatic growth rates were significantly higher when juveniles were fed eukaryotic algae than when fed cyanobacteria. Linear regression analyses revealed significant positive relationships between somatic growth rates and dietary sterol and polyunsaturated fatty acid content. Somatic growth rates also were highly correlated with the total sterol and α- linolenic acid content of clam tissues. This result suggests that the growth of C. fluminea is partially dependent on the availability of these essential lipids in the diet. Algal nutritional value may influence the successful geographic spread of this highly invasive species because food quality and quantity are changing as a result of global warming.

Key words: Corbicula fluminea, sterol, food quality, cyanobacteria, PUFA, invasive species, essential lipids.

11 2 Effect of food quality on clam growth

Introduction The clam Corbicula fluminea (Bivalvia, Corbiculidae), indigenous to Australia, Asia, and Africa, has been introduced to Europe and North and South America in the last century (Britton and Morton 1979; Araujo et al. 1993). Since its introduction, C. fluminea has undergone a remarkable range expansion to become an ubiquitous benthic invertebrate in freshwater ecosystems. In Germany, C. fluminea has successfully invaded many rivers, streams, and lakes where it often maintains large populations. Corbicula fluminea was discovered in the lower river Rhine in 1988, spread rapidly upstream, and was first recorded in Lake Constance in 2003 (Werner and Mörtl 2004). Invasive bivalves can dramatically alter the structure of native benthic communities as evidenced by the results of invasions of C. fluminea (Aldridge and McMahon 1978; Belanger et al. 1990; Williams et al. 1993) and the zebra mussel Dreissena polymorpha in many freshwater ecosystems in Europe and North and South America (Stanczykowska 1977; MacIsaac et al. 2002; Mörtl and Rothhaupt 2003).

Bivalves like C. fluminea and D. polymorpha are important benthic–pelagic couplers that remove large numbers of particles from the water column and transfer these resources to the substrate as biodeposits (faeces and pseudofaeces), thereby stimulating benthic productivity (Hakenkamp et al. 2001; Gergs et al. 2009). Corbicula fluminea and D. polymorpha also may play an important role in stimulating primary production by excreting inorganic nutrients to the water column (Yamamuro and Koike 1993; Vaughn et al. 2008). Hence, it is important to understand the factors responsible for the successful geographic spread of these invasive species.

Invasion patterns are affected by abiotic challenges, such as temperature, turbidity, dissolved O2, or NH3 (McMahon 1979; Cooper et al. 2005; Walther et al. 2009). They also might be affected by food and substrate availability (Foe and Knight 1985; Mouthon 2001; Schmidlin and Baur 2007) or by availability of essential nutritional components.

Long-chain polyunsaturated fatty acids (PUFAs) are essential dietary compounds that have important physiological functions (Cook 1996). For example, arachidonic acid (ARA) and eicosapentaenoic acid (EPA) are precursors of eicosanoids, which are thought to be relevant to bivalve reproduction and osmoregulation (Stanley-Samuelson 1994). Research on marine bivalves (almost exclusively on species relevant for aquaculture) has shown that a dietary source of certain PUFAs is required for proper growth and development (Delaunay et al. 1993; Soudant et al. 1996b). However, lipid requirements of freshwater bivalves are poorly understood. Recent research showed that certain long-chain PUFAs (e.g., EPA and docosahexaenoic acid [DHA]) are important to D. polymorpha (Vanderploeg et al. 1996; Wright et al. 1996; Wacker et al. 2002; Wacker and Von Elert 2003; 2004).

Sterols are another class of lipids required for many physiological processes. They are indispensable components of cell membranes and precursors for steroid hormones (Goad 1981; Martin-Creuzburg and Von Elert 2009). Experiments with marine bivalves suggest that the ability to synthesize sterols de novo is generally low or absent among bivalves. Thus, a

12 2 Effect of food quality on clam growth dietary source of sterols is necessary for growth (Soudant et al. 1998; Park et al. 2002). However, data on sterol requirements of freshwater bivalves are not yet available.

We investigated the nutritional requirements of the invasive filter-feeding freshwater clam C. fluminea for essential lipids in standardized growth experiments. Corbicula fluminea were fed common cyanobacterial and eukaryotic pelagic food sources that differed in their biochemical composition, and their growth rates were related to the availability of essential lipids in their diets and to lipids incorporated in clam tissues.

Methods

Clam collection Corbicula fluminea individuals were collected in the upper basin of Lake Constance at a sampling site described by Werner and Rothhaupt (2008). The clams were collected by SCUBA divers at water depths of 1.5 to 3 m, depending on lake-level fluctuations. Living individuals were separated from debris, sand, and gravel. They were transferred to flow- through systems with filtered (<30 µm), aerated lake water and precombusted sediment at an ambient temperature of 20°C until the start of growth experiments.

Cultivation of cyanobacteria and algae Food sources for C. fluminea were cultivated semi continuously in aerated 5-L vessels at a dilution rate of 0.25/d at 20°C with illumination at 100 to120 μmol quanta m–2 s–1 and were harvested in the late-exponential-growth phase. The coccoid cyanobacterium Synechococcus elongatus (SAG 89.70; Sammlung für Algenkulturen, Göttingen, Germany), the filamentous cyanobacteria Anabaena variabilis (ATCC 29413; American Type Culture Collection, Manassas, Virginia), and Aphanizomenon flos-aquae (CCAP1401-1; Culture collection of algae and protozoa, Oban, Scotland), and the green alga Scenedesmus obliquus (SAG 276-3a) were grown in Cyano medium (Jüttner et al. 1983). The flagellate Cryptomonas sp. (SAG 26.80) was grown in modified Woods Hole (WC) medium (Guillard 1975) enriched with vitamins and P. These foods were used because they differ in PUFA and sterol content. Food suspensions were prepared by concentrating the cells by centrifugation and resuspension in fresh media. Concentrations of the food suspensions were estimated from photometric light extinction (800 nm) and from C-extinction equations determined prior to the experiments.

Growth experiments Adult bivalves expend most of their energy in reproduction (gametogenesis) and little in somatic growth (Soudant et al. 1999). Therefore, we used juveniles that were sexually immature to maximize somatic growth rates and to enlarge possible differences in response among food sources (initial fresh mass 70.0 ± 55.1 mg, range between 10.5 - 222.9 mg). The 13 2 Effect of food quality on clam growth

28-d experiments were carried out at 20 °C in glass beakers filled with 200 mL of filtered lake water (0.45-μm pore-size membrane filter) and precombusted sediment (550 °C for 5 h) to allow the clams to burrow. Each of the 5 food treatments was replicated in 15 beakers, and 1 randomly chosen clam was transferred to each beaker. Clams were fed daily with 3 mg C/L of the food suspensions during the experiment. Water was exchanged daily to remove fecal pellets, and sediment was exchanged once a week to reduce biofilm formation. No clams died during the experiment.

Somatic growth rates (g) were determined as the increase in dry mass from the beginning of the experiment (W0) to day 28 (Wt) over time (t) with the equation used by Martin– Creuzburg et al. (2005b):

(lnW  lnW ) g  t 0 t .

A subsample of clams (n = 48) was taken at the beginning of the experiment to estimate the individual fresh (FM) and dry mass (DM) after 24 h freeze-drying to establish a fresh-dry- mass regression (DM = 0.599FM). Samples were weighed on an electronic balance (±0.1 μg; XP2U, Mettler Toledo GmbH, Gießen, Germany). The start-dry mass of clams used in the growth experiment was estimated from their actual fresh mass and the previously determined fresh-dry-mass regression. Growth rates of all specimens were calculated as means (n = 15) for each treatment.

Analyses of food organisms and clam tissues Aliquots of the food suspensions were filtered onto precombusted glass-fiber filters (Whatman GF/F, 25-mm diameter) and analyzed for particulate organic C (POC) and N with an NCS-2500 analyzer (ThermoQuest, Biberach, Germany). Particulate P was measured in aliquots collected on acid-rinsed polysulfon filters (HT-200; Pall, Dreieich, Germany) and digested with a solution of 10% potassium peroxodisulfate and 1.5% sodium hydroxide for 60 min at 121°C. Soluble reactive P was determined with the molybdate-ascorbic acid method (APHA 1985).

Soft tissues of freeze-dried clams were isolated and weighed before determination of elemental composition. The C and N content of soft tissues (n = 3 per treatment) was measured as described above with an NCS-2500 analyzer. C and N contents were expressed per unit tissue dry mass and were converted to molar C:N ratios. The P in soft tissues (n = 3 for each treatment) was solubilized by mechanical shearing with a mortar and by sonication in ultrapure water. P content was measured as described above. P content was expressed per unit tissue dry mass and was converted to C:P molar ratios.

Lipids in food sources were extracted twice in a mixture of dichloromethane/methanol (2:1, volume/volume [v/v]) from precombusted GF/F filters (Whatman, 25-mm diameter) loaded with ~0.5 mg (for fatty acid analysis) or ~1 mg (for sterol analysis) POC of the algal food 14 2 Effect of food quality on clam growth sources. Soft tissues of freeze-dried clams (n = 2 for each treatment) were separated from their shell, weighed, and placed in dichloromethane/methanol (2:1, v/v). Lipids in soft tissues were solubilized by mechanical shearing with a mortar and by sonication. Half of the resulting tissue solution was used for fatty acid analysis and the other half was used for sterol analysis.

For analysis of sterols, the pooled cell-free extracts were dried under a stream of N2 and saponified with 0.2 mol/L methanolic KOH (70°C, 1 h). Subsequently, sterols were partitioned into iso-hexane:diethyl ether (9:1, v/v), dried under a stream of N2, and resuspended in a volume of 10 to 30 mL iso-hexane. For analysis of fatty acids, the cell-free extracts were dried under a stream of N2 and esterified with 3 mol/L methanolic HCl (60°C, 20 min). Subsequently, fatty acid methyl esters (FAMEs) were partitioned into iso-hexane, dried under a stream of N2, and resuspended in a volume of 25 to 100 mL iso-hexane. Lipids were analyzed by gas chromatography on an HP 6890 gas chromatograph (GC; Agilent Technologies, Böblingen, Germany) equipped with a flame ionization detector and either a DB-225 (J&W Scientific, Cologne, Germany) capillary column to analyze FAMEs or an HP-5 (Agilent) capillary column to analyze sterols. Details of GC configurations are given elsewhere (Martin-Creuzburg et al. 2009; 2010).

Lipids were quantified by comparison to internal standards (C17:0 and C23:0 methyl esters, 5α-cholestane). The detection limit was ~20 ng of sterol/fatty acid. Lipids were identified by their retention times and their mass spectra, which were recorded with a GC/mass spectrometer (GCQ, Thermo Finnigan MAT, Bremen, Germany) equipped with a fused Si capillary column (DB-225MS, J&W Scientific for FAMEs; DB-5MS, Agilent for sterols). Sterols were analyzed in their free form and as their trimethylsilyl derivatives. Mass spectra were recorded between 50 and 600 amu in the electron impact ionization mode, and lipids were identified by comparison with mass spectra of reference substances purchased from Sigma- Aldrich (Munich, Germany) or Steraloids (Newport, RI, USA), mass spectra found in a self- generated spectra library, or in the literature (e.g. Toyama et al. 1952; Belanger et al. 1973; Goad and Akihisa 1997). The absolute amount of each lipid was quantified per mg POC or soft-tissue C content of clams. Fatty acids are reported using shorthand nomenclature as follows: a:bn-x, where a represents the number of C atoms, b is the number of double bonds, and x is the position of the first double bond counted from the methyl end.

Statistical analyses All statistical analyses were done with the statistical software package R (R Development Core Team 2006). Differences among growth rates were analyzed with analysis of covariance (ANCOVA) with food treatment as the categorial variable and

15 2 Effect of food quality on clam growth

Tab. 1 Mean (±1 SD; n = 3) elemental and biochemical composition of cyanobacteria (Synechococcus elongatus, Anabaena variabilis, Aphanizomenon flos-aquae) and eukaryotic algae (Scenedesmus obliquus, Cryptomonas sp.) used as food sources for Corbicula fluminea. Elemental ratios for C:N and C:P are molar, fatty acids (FA) are reported as µg FA/mg C. Lipid classes are saturated (SAFA), mono- (MUFA), and polyunsaturated fatty acids (PUFA) and fatty acids with double bond at n-3 position, n-6 position, and ratio between those lipid classes (n- 3:n-6). Fatty acids are described as a:b(n-c) where a = number of C atoms, b = number of double bonds, c = position of the first double bond counted from the methyl end. n.d. = below detection limits.

Variable S. eolongatus A. variabilis A. flos-aquae S. obliquus Cryptomonas sp.

C:N 4.4 ± 0.1 4.4 ± 0.1 4.9 ± 0.1 5.8 ± 0.1 6.0 ± 0.1

C:P 53.0 ± 0.3 50.7 ± 1.0 80.2 ± 0.4 40.7 ± 3.0 124.5 ± 3.5

18:2(n-6) n.d. 38.5 ± 1.2 11.1 ± 1.3 46.9 ± 3.1 21.8 ± 1.4

18:3(n-3) n.d. 66.4 ± 2.3 66.2 ± 9.8 134.9 ± 9.5 76.0 ± 5.0

18:4(n-3) n.d. n.d. n.d. 11.8 ± 0.6 36.50 ± 1.7

20:2(n-6) n.d. n.d. n.d. n.d. n.d.

20:4(n-6) n.d. n.d. n.d. n.d. n.d.

20:5(n-3) n.d. n.d. n.d. n.d. 46.9 ± 2.0

22:5(n-3) n.d. n.d. n.d. n.d. n.d.

22:6(n-3) n.d. n.d. n.d. n.d. 5.0 ± 0.1

Total FA content 125.0 ± 11.0 242.5 ± 8.3 157.2 ± 13.9 272.3 ± 22.0 251.2 ± 14.0

SAFA 56.4 ± 8.9 70.6 ± 2.4 55.01 ± 2.50 52.05 ± 3.5 42.4 ± 4.6

MUFA 68.6 ± 2.1 67.1 ± 1.8 24.9 ± 1.6 23.5 ± 3.6 27.5 ± 1.3

PUFA n.d. 104.9 ± 4.4 77.3 ± 11.1 196.7 ± 15.8 181.2 ± 9.8

n-3 n.d. 66.4 ± 3.2 66.2 ± 1.3 146.8 ± 5.7 159.4 ± 1.4

n-6 n.d. 38.5 ± 2.3 11.1 ± 9.8 50.0 ± 10.1 21.8 ± 8.7

n-3:n-6 n.d. 1.7 6.0 2.9 7.3

individual dry mass at the beginning of each experiment as the covariates. Differences among treatments were analyzed with Tukey’s Honestly Significant Difference (HSD) post hoc test. The dependence of growth rates of C. fluminea on both food components and clam tissue components was assessed by linear regression analyses.

16 2 Effect of food quality on clam growth

Results

Elemental and biochemical composition of food sources Molar C:N and C:P ratios of the cyanobacteria S. elongatus, A. variabilis, and A. flos-aquae and the eukaryotic algae S. obliquus and Cryptomonas sp. were low, results indicating high N and P content (Tab. 1). The lipid composition of cyanobacteria was characterized by high amounts of short-chain saturated (SAFA) and monounsaturated (MUFA) fatty acids and by the absence of sterols. In contrast to A. variabilis and A. flos-aquae, which contained comparatively high amounts of the 2 PUFAs linoleic acid (18:2n-6, LIN) and α-linolenic acid (18:3n-3, ALA), S. elongatus contained no PUFAs. Scenedesmus obliquus contained considerable amounts of ALA and 18:4n-3, and Cryptomonas sp. contained high amounts of the long-chain fatty acid eicosapentaenoic acid (20:5n-3, EPA) and small amounts of docosahexaenoic acid (22:6n-3, DHA) (Tab. 1). Both eukaryotic species were characterized by high total fatty acid, PUFA, and n-3 levels.

Sterols were detected only in the eukaryotic algae. Scenedesmus obliquus contained a total of 9.9 ± 4.2 µg sterol/mg C. Dominant sterols were fungisterol (5α-ergost-7-en-3β-ol; 2.1 ± 0.7 µg/mg C); chondrillasterol ((22E)-5α-poriferasta-7,22-dien-3-ol; 7.4 ± 3.4 µg/mg C); and 22-dihydrochondrillasterol (5α-poriferast-7-en-3β-ol; 0.5 ± 0.1 µg/mg C). In Cryptomonas sp., total sterols averaged 13.4 ± 4.5 µg/mg C and consisted of 2 principal sterols: brassicasterol ((22E)-ergosta-5,22-dien-3β-ol; 4.2 ± 1.4 µg/mg C) and stigmasterol ((22E)-stigmasta-5,22- dien-3β-ol; 9.1 ± 3.0 µg/mg C).

Elemental and biochemical composition of clam tissues The elemental composition of soft-tissues was characterized by low C:N (4.6 - 5.3) and C:P (~100 - ~140) (Tab. 2), indicating a high N and P content. Nine sterols were detected in clam tissues: cholesterol (cholest-5-en-3β-ol), stigmasterol, brassicasterol, campesterol (campest- 5-en-3β-ol), corbisterol ((22E)-stigmasta-5,7,22E-trien-3β-ol), ergosterol ((22E)-ergosta- 5,7,22-trien-3β-ol), fungisterol, chondrillasterol, and 22-dihydrochondrillasterol (Tab. 2). The total sterol level of field-collected clams at the start of the experiment was 13.8 ± 0.6 µg sterol/mg C. High amounts of the sterols detected in Cryptomonas sp. were found in tissues of clams fed Cryptomonas sp. (Tab. 2), and small amounts of the sterols detected in S. obliquus were found in tissues of clams fed S. obliquus, results indicating incorporation of dietary phytosterols.

The total FA content of C. fluminea was between 115 and 150 µg/mg C. Highest amounts were measured in field-collected clams at the beginning of the experiment and in individuals fed either of the 2 eukaryotic algae (Tab. 2). Fatty-acid profiles were dominated by saturated fatty acids (SAFAs, 16:0, 18:0, 20:0), monounsaturated fatty acids (MUFAs, 18:1), and PUFAs with ≥18 C atoms, i.e., LIN, ALA, ARA, EPA, docosapentaenoic acid (22:5n-3, DPA), and DHA, but the relative proportion of these fatty acids differed with food sources (Tab. 2).

17 2 Effect of food quality on clam growth

Tab. 2 Mean (±1 SD; n = 3 for C:N and C:P, n = 2 for lipids and sterols) elemental and biochemical composition of soft-tissues of Corbicula fluminea fed different cyanobacteria (Synechococcus elongatus, Anabaena variabilis, Aphanizomenon flos-aquae) and eukaryotic algae (Scenedesmus obliquus, Cryptomonas sp.) after 28 d in a laboratory growth experiment. Abbreviations, units, and fatty acid structure are as in Tab. 1. n.d. = below detection limits.

Variable Start S. eolongatus A. variabilis A. flos-aquae S. obliquus Cryptomonas sp.

C:N 4.9 ± 0.2 5.3 ± 0.2 5.0 ± 0.2 4.6 ± 0.0 4.9 ± 0.1 5.3 ± 0.1

C:P 139.2 ± 3.9 122.9 ± 7.5 125.0 ± 0.5 103.3 ± 0.0 131.9 ± 4.5 121.2 ± 3.5

18:2(n-6) 3.9 ± 1.8 n.d. 7.5 ± 1.1 2.9 ± 0.4 8.1 ± 0.2 5.4 ± 0.6

18:3(n-3) 6.8 ± 2.6 3.4 ± 1.2 15.8 ± 2.2 12.7 ± 2.4 17.9 ± 1.3 18.4 ± 1.9

18:4(n-3) 1.8 ± 2.6 n.d. n.d. n.d. 1.1 ± 1.5 4.5 ± 0.6

20:2(n-6) n.d. n.d. 3.2 ± 0.5 n.d. 2.9 ± 0.4 3.9 ± 0.1

20:4(n-6) 10.1 ± 1.3 5.1 ± 0.1 5.7 ± 0.8 7.3 ± 0.3 3.8 ± 1.7 5.6 ± 0.3

20:5(n-3) 9.5 ± 2.00 6.3 ± 1.4 4.6 ± 0.6 5.2 ± 1.1 3.3 ± 0.7 15.7 ± 0.8

22:5(n-3) 8.2 ± 0.2 5.2 ± 0.8 5.0 ± 0.7 6.1 ± 0.4 n.d. 6.3 ± 0.1

22:6(n-3) 17.5 ± 1.3 10.2 ± 1.9 9.1 ± 1.3 11.3 ± 0.3 5.0 ± 1.5 7.0 ± 0.1

Total FA content 149.6 ± 14.7 116.1 ± 25.4 128.8 ± 18.2 115.3 ± 11.7 134.3 ± 20.3 133.8 ± 3.4

SAFA 31.0 ± 4.3 28.8 ± 4.8 24.4 ± 3.5 25.8 ± 3.6 29.9 ± 2.5 29.2 ± 0.2

MUFA 16.6 ± 4.4 19.4 ± 2.2 23.1 ± 3.3 15.4 ± 2.0 27.7 ± 1.6 12.4 ± 0.7

PUFA 57.9 ± 8.8 42.8 ± 5.4 51.0 ± 7.2 45.6 ± 5.0 42.1 ± 1.2 66.7 ± 3.6

n-3 43.9 ± 8.3 25.0 ± 5.3 34.5 ± 4.9 35.4 ± 4.3 27.2 ± 0.7 51.9 ± 3.5

n-6 14.0 ± 0.5 5.1 ± 0.1 16.4 ± 2.3 10.2 ± 0.7 14.9 ± 1.8 14.8 ± 0.2

n-3:n-6 3.1 4.9 2.1 3.5 1.8 3.5

Unidentified sterol 0.7 ± 0.2 0.1 ± 0.2 0.2 ± 0.0 0.3 ± 0.0 0.2 ± 0.0 0.2 ± 0.2

Cholesterol 8.3 ± 1.0 3.4 ± 2.8 4.0 ± 0.4 6.8 ± 0.8 5.7 ± 0.1 7.3 ± 1.1

Brassicasterol 1.8 ± 0.4 0.8 ± 0.6 0.9 ± 0.1 1.6± 0.1 1.4 ± 0.3 9.7 ± 7.1

Ergosterol n.d. n.d. n.d. n.d. 0.2 ± 0.2 1.8 ± 2.5

Campesterol 0.9 ± 0.2 0.3 ± 0.2 0.3 ± 0.0 0.5 ± 0.1 1.7 ± 0.4 0.8 ± 0.6

Stigmasterol 1.2 ± 0.1 0.7 ± 0.5 0.5 ± 0.0 0.9 ± 0.2 3.1 ± 0.8 9.2 ± 7.2

Corbisterol 0.8 ± 0.1 n.d. n.d. n.d. 0.8 ± 0.1 1.3 ± 1.3

Fungisterol 0.1 ± 0.1 n.d. n.d. n.d. 0.5 ± 0.0 n.d.

Chondrillasterol n.d. 0.3 ± 0.4 n.d. n.d. 0.3 ± 0.4 n.d.

22-Dihydrochondrillasterol 0.1 ± 0.2 n.d. n.d. n.d. 0.1 ± 0.1 n.d.

Total sterol content 13.8 ± 0.6 5.7 ± 4.7 5.9 ± 0.6 10.1 ± 1.2 14.0 ± 1.9 30.4 ± 19.8

18 2 Effect of food quality on clam growth

Fig. 5 Mean (±1 SD; n = 15) somatic growth rates of Corbicula fluminea fed 5 different food sources for 28 d in a laboratory growth experiment. Bars labeled with different letters differ significantly (Tukey’s Honestly Significant Difference, p < 0.05).

Growth of C. fluminea In all treatments, the dry mass of C. fluminea increased during the 28 d experiment (positive somatic growth rates; Fig. 5). However, neither the initial mass (ANCOVA, F = 2.96, df = 1, p = 0.09) nor the initial mass × food source interaction (ANCOVA, F =1.30, df = 4, p = 0.28) significantly explained the growth of C. fluminea. Instead, growth rates were significantly affected only by food treatments (ANCOVA, F = 32.9, df = 4, p < 0.001; Tukey’s HSD, p < 0.05). Clams fed the 3 cyanobacteria had lower growth rates than clams fed the 2 eukaryotic algae. Growth rates did not differ significantly among clams fed filamentous (A. variabilis, A. flos-aquae) or single-celled (S. elongates) cyanobacteria (Fig. 5). Clams fed S. obliquus (0.023 ± 0.004/d) and Cryptomonas sp. (0.027 ± 0.004/d) achieved the highest growth rates, but these growth rates were not significantly different.

19 2 Effect of food quality on clam growth

Fig. 6 Linear regression models between molar C:N (A) and C:P (B); total fatty acid (FA) (C), total sterol (D), n-3 (E), n-6 (F), saturated (G), monounsaturated (H), and polyunsaturated (I) FA levels; n-3:n-6 ratio (K); and α- linolenic acid (ALA (L) and linoleic acid (LIN) (M) levels in food sources (x-axis) and growth rates of Corbicula fluminea (y-axis). Regression equations are plotted for significant models.

Growth rates of clams were positively correlated with many biochemical components of their food (Fig. 6A–M). For example, C:N (Fig. 6A), total fatty acids (Fig. 6C), total sterols (Fig. 6D), n-3 PUFAs (Fig. 6E), total PUFAs (Fig. 6I), ALA (Fig. 6L), and LIN (Fig. 6M) content of food sources were all significantly positively correlated with clam growth. SAFA (Fig. 6G) and MUFA (Fig. 6H) content of food sources were significantly negatively correlated with growth. Growth rates of clams (Fig. 7A–M) were significantly positively correlated with total sterols (Fig. 7D), ALA (Fig. 7L), and EPA (Fig. 7M) concentrations in clam soft-tissues.

20 2 Effect of food quality on clam growth

Fig. 7 Linear regression models between molar C:N (A) and C:P (B); total fatty acid (FA) (C), total sterol (D), n-3 (E), n-6 (F), saturated (G), monounsaturated (H), and polyunsaturated (I) FA levels; n-3:n-6 ratio (K); and α- linolenic acid (ALA) (L), and eicosapentaenoic acid (EPA) (M) levels in Corbicula fluminea tissue (x-axis) and growth rate of Corbicula fluminea individuals (y-axis). Regression equations are plotted for significant models.

Discussion Few studies have been done to determine the effects of the food quality of phytoplankton species on freshwater bivalves. Foe and Knight (1986) examined the growth of C. fluminea fed 6 genera of green algae in various combinations. They observed strongly positive tissue growth when clams were fed various mixtures of Ankistrodesmus, Chlamydomonas, Chlorella, and Scenedesmus, but not when they were fed unialgal food sources or mixtures containing Pedinomonas or Selenastrum. However, the nutritional requirements of Corbicula that make some algal species higher-quality food than others have not been further characterized.

In our study, the growth of C. fluminea was significantly affected by the lipid composition of its food. Corbicula fluminea growth rates were significantly lower when juveniles were fed 21 2 Effect of food quality on clam growth cyanobacterial diets than when they were fed eukaryotic algae. This result suggests that cyanobacteria are of poorer food quality than eukaryotic algae for C. fluminea and adds to previous findings that cyanobacteria are a nutritionally inadequate food source for a number of aquatic invertebrates, including D. polymorpha (Wacker et al. 2002; Vanderploeg et al. 2009) or cladocerans of the genus Daphnia (Wilson et al. 2006; Martin-Creuzburg et al. 2008). Cyanobacteria are a poor-quality food for Daphnia because they cannot supply essential lipids, especially sterols (Martin-Creuzburg et al. 2008).

Few studies have been done on the sterol requirements of bivalves, and none have been done on freshwater bivalves. Experiments with marine bivalves suggest that sterols are essential for gametogenesis and for biosynthesis of membranes in the early embryo (Soudant et al. 1996a). Duncan et al. (1987) reported that cholesterol is the main sterol in C. fluminea with lesser amounts of campesterol, sitosterol, stigmasterol, and brassicasterol. Recent studies confirm the presence of these sterols in C. fluminea (Chijimatsu et al. 2011). However, whether these sterols are synthesized de novo by Corbicula or are of dietary origin has not been investigated. The 5 sterols (cholesterol, brassicasterol, campesterol, stigmasterol and corbisterol) we found in C. fluminea agree with those reported by Duncan et al. (1987). In addition, sterols present in the eukaryotic algae S. obliquus and Cryptomonas sp., which were in accordance with former lipid analysis in algae (Martin-Creuzburg et al. 2005b; Martin-Creuzburg et al. 2006), also occurred in Corbicula tissues. Thus, the biochemical composition of clam tissues reflected the biochemical composition of their food sources. However, we cannot determine whether these sterols were assimilated or simply present in the gut of the animals because gut tissues were not dissected prior to analyses.

Total sterol levels in clam tissues decreased relative to initial levels within 4 wk of growth on either of the cyanobacterial diets. In contrast, total sterol levels in clams fed the sterol- containing green alga S. obliquus were constant over the experiment, a result that suggests dietary uptake and possibly homeostatic regulation of sterol levels. Total sterol levels in clams fed Cryptomonas sp. increased during the experiment primarily because of the incorporation of large quantities of dietary brassicasterol and stigmasterol. We cannot show definitively that the correlations we obtained represent causal relationships. Nevertheless, the finding that total sterol levels in the food organisms and in clam tissues were highly correlated with somatic growth rates of the clams suggests that C. fluminea relies on a dietary source of sterols to obtain high growth rates. Hence, growth of clams fed cyanobacterial diets presumably was limited by the absence of sterols, as has been shown for the crustacean Daphnia (Martin-Creuzburg et al. 2008).

PUFAs are another important and well documented class of essential lipids. In particular, molecules with >18 C atoms, such as EPA, play a significant role in animal physiology, and their absence can constrain growth and reproduction of various invertebrate species, such as Daphnia (Müller-Navarra et al. 2000; Wacker and Von Elert 2001) or D. polymorpha (Vanderploeg et al. 1996; Wacker et al. 2002; Wacker and Von Elert 2004). We found strong positive correlations between somatic growth rates and the levels of total PUFAs and n-3

22 2 Effect of food quality on clam growth

PUFAs in the diet, which suggests that the growth of C. fluminea is strongly affected by the availability of dietary PUFAs. Moreover, the 2 PUFAs LIN and ALA, which are important precursors for long-chain PUFAs of the n-6 and n-3 series, respectively, also were positively correlated with somatic growth rates of C. fluminea (Fig. 6L, M). Somatic growth rates and the availability of SAFAs and MUFAs in the diet were negatively correlated, results suggesting that growth was not affected by these FAs. Somatic growth rates and tissue levels of ALA and EPA were strongly positively correlated, results suggesting that long-chain PUFAs are of particular importance for growth of C. fluminea. This conclusion is in accordance with studies on nutritional requirements of marine bivalves, which suggest that egg numbers, hatching success, and growth and survival of larvae are closely linked to dietary lipids, particularly to ARA, EPA, and DHA (Soudant et al. 1996a; 1996b; 1996c). Dietary PUFAs also affect egg mass, egg quality, and larval growth of D. polymorpha (Wacker et al. 2002; Wacker and Von Elert 2003; 2004). Hence, in our study, growth of C. fluminea on cyanobacterial diets was possibly constrained by the absence of sterols and, simultaneously, by the absence of long-chain PUFAs. Martin-Creuzburg et al. (2009) used Daphnia as a model organism to show that feeding on a cyanobacterial diet leads to colimitation by sterols and PUFAs. Our data strongly suggest that the growth of juvenile C. fluminea depends on the availability of these essential lipids in the diet. However, we cannot separate the effects mediated by dietary sterols from those mediated by dietary PUFAs (or other potentially colimiting nutrients not considered here) to assess the relative importance of these essential nutrients.

Corbicula fluminea is regarded as a nonselective suspension feeder (Lauritsen 1986; Way et al. 1990; Boltovskoy et al. 1995; Vaughn and Hakenkamp 2001). However, the finding that Anodonta anatina is able to reject filamentous cyanobacteria suggests that size and shape of suspended particles affect particle sorting in bivalves (Ward et al. 1998; Vanderploeg et al. 2001; Bontes et al. 2007). Furthermore, variation in gill surface structures might lead to differences in particle sorting in bivalve species (Payne et al. 1995; Ward et al. 1998; Baker et al. 2000). Chemoreception via recognition of surface structures has been proposed as another mechanism by which bivalves avoid uptake of unsuitable food sources (Espinosa et al. 2010). Vanderploeg et al. (2001; 2009) suggested that D. polymorpha is able to discriminate against large toxic colonies of Microcystis aeruginosa and, thereby, promotes toxic Microcystis blooms. We provided C. fluminea with different phytoplankton species as sole food sources to assess differences in food quality rather than to assess capability for selective feeding. To assess the significance of a lipid limitation in the field, research is needed to test whether C. fluminea can select food particles on the basis of biochemical quality, e.g., sterol or PUFA content.

Preliminary experiments suggested that all food sources used in our study, including the filamentous cyanobacteria, were readily removed from the water column and ingested by C. fluminea. Moreover, in contrast to starving clams, which did not grow at all, even the clams fed cyanobacterial diets had positive somatic growth rates, which indicated that all offered food sources were of some nutritional use. After 24 h of feeding, significant amounts of food particles were present in the water column, an observation suggesting that clams did not

23 2 Effect of food quality on clam growth reduce food quantity to limiting levels and that a daily supply of 3 mg C/L was sufficient. Moreover, significant amounts of biodeposited material (feces and pseudofeces) were observed on sediment surfaces in all food treatments. Corbicula fluminea can feed on sedimented or biodeposited organic matter (Boltovskoy et al. 1995; Cahoon and Owen 1996; Hakenkamp and Palmer 1999). Thus, C limitation in our experiment is rather unlikely.

The biomass-related somatic growth rates obtained in our experiment correspond to an increase in shell length of 0.6 to 1.8 mm over the 28-d experiment (0.02–0.06 mm/d), which is comparable with published data on shell growth of natural C. fluminea populations (Belanger et al. 1990; French and Schloesser 1991; Vohmann et al. 2010).

Molar C:N and C:P of clam tissues were low and comparable to elemental tissue ratios of Corbicula described in the literature (Evans-White et al. 2005). Negative correlations of somatic growth rates with C:N or C:P of the different food sources (i.e., increasing growth rates with decreasing C:N or C:P) would be expected if growth were limited by low availability of N or P. However, in our experiment, somatic growth rates of C. fluminea were positively correlated with dietary C:N, i.e., growth rates increased with a decreasing availability of dietary N. This result together with the finding that the highest growth rates were obtained with algae containing high amounts of lipids but low levels of N suggest that somatic growth of C. fluminea was constrained by a low availability of dietary lipids and not by the availability of dietary N. Corbicula fluminea has a low demand for N (Atkinson et al. 2010) and, therefore, limitation by N in our experimental setup was rather unlikely. Bivalves might bypass elemental nutrient limitations by accumulating N and P within body tissues, a mechanism that could contribute to the success of invasive species (Naddafi et al. 2009). Moreover, bivalves are involved in nutrient recycling. For instance, bivalves have the potential to affect phytoplankton communities directly by grazing or indirectly via resuspension of limiting nutrients (Arnott and Vanni 1996; Vanni 2002), thereby supporting growth of phytoplankton (Lauritsen 1986; Yamamuro and Koike 1993), which is largely driven by the availability of N and P (Tilman et al. 1982; Elser et al. 1990). Thus, on one hand, bivalve invasion might be influenced by the composition of the phytoplankton community, and on the other hand, bivalve invasion might affect nutrient flows in lake ecosystems.

Annual succession of suspended food organisms (mainly phytoplankton) is associated with changes in food quantity and quality, e.g., changes in the availability of biochemical nutrients for filter-feeding organisms (Wacker and Von Elert 2001), which also may affect the fitness of bivalves. In eutrophic systems, summer months are often characterized by a high proportion of cyanobacteria in the phytoplankton, which may limit availability of essential substances like sterols and PUFAs (Müller-Navarra et al. 2004). The frequency of cyanobacterial bloom formation is expected to increase with global warming (Paerl and Huisman 2008; Wagner and Adrian 2009), and this increased frequency could influence the geographic spread of C. fluminea and other invasive bivalves. However, clams like C. fluminea might change their feeding mode from seston filtration to deposition feeding. Increased reliance on benthic food sources could allow clams to outlast possible limitations

24 2 Effect of food quality on clam growth caused by shifts in pelagic food sources (Hakenkamp and Palmer 1999; Nichols and Garling 2000; Raikow and Hamilton 2001; Nichols et al. 2005).

Our finding that the growth of C. fluminea is significantly affected by the availability of dietary sterols and PUFAs contributes to our understanding of how benthic food web processes are affected by biochemical food-quality constraints, a topic that has been studied almost exclusively for pelagic components of the food web. Investigating and comparing the physiological demands of native and invasive species may help us understand invasion patterns and improve risk assessment of upcoming invasions.

Acknowledgements We thank R. Basen, N. Schlotz and 2 anonymous referees for helpful comments on the manuscript that improved its quality. This work was supported by the DFG (German Research Foundation) within the collaborative research centre SFB 454 “Littoral of Lake Constance”.

25 2 Effect of food quality on clam growth

26 3 Sterol limitation of Corbicula

Chapter 3

Absence of sterols constrains food quality of cyanobacteria for an invasive freshwater bivalve

Timo Basen, Karl-Otto Rothhaupt and Dominik Martin-Creuzburg

Oecologia doi: 10.1007/s00442-012-2294-z received: 28. September 2011 accepted: 22. February 2012

Abstract The accumulation of cyanobacterial biomass may severely affect the performance of aquatic consumers. Here, we investigated the role of sterols in determining the food quality of cyanobacteria for the invasive clam Corbicula fluminea, which has become a common benthic invertebrate in many freshwater ecosystems throughout the world. In standardized growth experiments, juvenile clams were fed mixtures of different cyanobacteria (Anabaena variabilis, Aphanothece clathrata, Synechococcus elongatus) or sterol-containing eukaryotic algae (Cryptomonas sp., Nannochloropsis limnetica, Scenedesmus obliquus). In addition, the cyanobacterial food was supplemented with different sterols. We provide evidence that somatic growth of C. fluminea on cyanobacterial diets is constrained by the absence of sterols, as indicated by a growth-enhancing effect of sterol supplementation. Thus, our findings contribute to our understanding of the consequences of cyanobacterial mass developments for benthic consumers and highlight the importance of considering sterols as potentially limiting nutrients in aquatic food webs.

Key words: benthic-pelagic coupling, Corbicula fluminea, cyanobacterial blooms, essential lipids, food quality

27 3 Sterol limitation of Corbicula

Introduction The accumulation of cyanobacterial biomass, frequently observed in aquatic ecosystems around the globe, severely affects food web processes and is often associated with hazards to human health and livestock and reduced recreational quality of water bodies (Carmichael 1992; Codd 1995). As the frequency of cyanobacterial bloom formation is expected to increase with global warming (Paerl and Huisman 2008), it is important to investigate the consequences of cyanobacterial mass developments for ecosystem processes, e.g., the role of cyanobacterial carbon within the food web and its food quality for aquatic consumers.

In general, cyanobacteria represent a nutritionally inadequate food source for aquatic consumers, which is due to (1) morphological properties, i.e. to the formation of filaments or colonies that hamper ingestion (De Bernardi and Giussani 1990; Van Donk et al. 2011), (2) the production of toxins (Carmichael 1992; Wilson et al. 2006) or other secondary metabolites that reduce the efficiency of digestion (protease inhibitors; Schwarzenberger et al. 2010), and/or (3) a deficiency in essential biochemical nutrients, in particular sterols (Von Elert et al. 2003; Martin-Creuzburg et al. 2008). It is generally assumed that cyanobacteria, in contrast to eukaryotic algae, are unable to synthesize sterols de novo (Volkman 2003; Summons et al. 2006). Sterols are required for a multitude of physiological processes in eukaryotic consumers, e.g., they are indispensable structural components of cell membranes and serve as precursors for steroid hormones (Goad 1981; Martin-Creuzburg and Von Elert 2009). The lack of sterols has been suggested to constrain the carbon transfer efficiency from cyanobacteria to herbivorous zooplankton (Von Elert et al. 2003; Martin-Creuzburg et al. 2008; Martin-Creuzburg et al. 2010).

When blooms settle down, a huge pelagic carbon pool is transferred to the benthic food web (Nascimento et al. 2008). It has been shown that cyanobacterial carbon can be ingested and assimilated by benthic invertebrates, but the nutritional value seems to be rather low (Karlson et al. 2008; Nascimento et al. 2009). In laboratory growth experiments, bivalves showed reduced growth and reproduction rates when fed cyanobacterial food sources (Wacker and Von Elert 2003; Basen et al. 2011). Despite the fact that they are capable of removing colonial and filamentous cyanobacteria from the water column, feeding on cyanobacteria usually results in low ingestion rates and in an increased production of pseudofaeces, irrespective of cyanobacterial toxin production (Pires et al. 2005; Bontes et al. 2007). In contrast to morphological properties and toxicity, the role of essential dietary compounds in determining food quality for bivalve species has been poorly investigated. Experiments with marine bivalves, primarily species relevant for aquaculture, suggest that the ability to synthesize sterols de novo is generally low or absent (Walton and Pennock 1972; Goad 1981). Thus, a dietary source of sterols is presumably required for growth and reproduction (Soudant et al. 2000; Park et al. 2002), which may explain why cyanobacteria are of low nutritional quality for bivalve species.

One of the most spreading invaders in freshwater ecosystems is the Asian clam Corbicula fluminea. Originating in Southeast Asia, C. fluminea has been introduced to North and South 28 3 Sterol limitation of Corbicula

America and Europe in recent decades and has undergone a remarkable range expansion to become an ubiquitous benthic invertebrate in lentic and lotic freshwater ecosystems (McMahon 1982; Araujo et al. 1993; Darrigran 2002). Invasive bivalves, like C. fluminea and Dreissena polymorpha, can strongly affect invaded native ecosystems. As highly efficient filter feeders they can alter phytoplankton dynamics, influence the pelagic nutrient cycling (Cahoon and Owen 1996; Hwang et al. 2011), and affect native benthic communities. By transferring pelagic carbon into the benthic food web, bivalves improve the benthic-pelagic coupling thereby stimulating benthic productivity (Strayer et al. 1999; Sousa et al. 2008).

In the present study, we investigated the consequences of a dietary sterol deficiency for the growth of C. fluminea in standardized growth experiments. Juvenile clams were fed different sterol-free (cyanobacteria) and sterol-containing food sources (algae). In addition, clams were fed a mixture of three different cyanobacteria supplemented with sterols to assess whether the growth of C. fluminea on a cyanobacterial diet is constrained by a deficiency in dietary sterols. In addition, the sterol composition of food sources and clam tissues was recorded.

Materials and methods

Cultivation of cyanobacteria and algae Food sources for C. fluminea were cultivated semi-continuously in aerated 5 l vessels at a dilution rate of 0.25 d–1 at 20 °C with illumination at 100 - 120 μmol quanta m–2 s–1, and harvested in the late-exponential growth phase. The coccoid cyanobacterium Synechococcus elongatus (SAG 89.70, Sammlung für Algenkulturen Göttingen, Germany), the filamentous cyanobacterium Anabaena variabilis (ATCC 29413, American Type Culture Collection, Manassas, USA), the gelatinous cyanobacterium Aphanotece clathrata (SAG 23.99), the green alga Scenedesmus obliquus (SAG 276-3a) and the eustigmatophyte Nannochloropsis limnetica (SAG 18.99) were grown in Cyano medium (Jüttner et al. 1983). The flagellate Cryptomonas sp. (SAG 26.80) was grown in modified Woods Hole (WC) medium enriched with vitamins (Guillard 1975), nitrogen (2 mM, final concentration) and phosphorus (100 µM, final concentration). These food organisms were used because they differ in their sterol content and composition. Food suspensions were prepared by concentrating the cells by centrifugation and resuspension in fresh media. Carbon concentrations of the food suspensions were estimated from photometric light extinctions (800 nm) and from carbon- extinction equations determined prior to the experiment.

Sterol supplementation of cyanobacteria Cyanobacteria were enriched with a mixture of cholesterol (Sigma, C8667, purity 99%), stigmasterol (Sigma, S2424, 95%), and ergosterol (Sigma, E2000, 95%) using a modified protocol of a method described by Von Elert (2002). Sterols were dissolved in ethanol (2.5

29 3 Sterol limitation of Corbicula mg ml-1) to prepare ethanolic stock solutions. For supplementation, 40 mg of bovine serum albumin (BSA, Sigma A7906, 98%) was dissolved in 10 ml of ultrapure water and 266.7 µl of each sterol stock solution were added during gentle stirring. Subsequently, 10 ml of Cyano medium and 2.67 mg particulate organic carbon (POC) of the three cyanobacterial stock solutions (8 mg POC in total) were added and after 5 min of incubation the volume was brought to 80 ml with Cyano medium. The resulting suspension was incubated on a rotary shaker (100 rpm) for 4 h with illumination at 100 µmol m-2 s-1. To remove excess BSA and free sterols, cyanobacterial cells were then concentrated by centrifugation and resuspended in fresh medium; this process was repeated twice. The obtained cyanobacterial food suspension (“Mix + BSA + Sterols”) was then used as food for C. fluminea in the growth experiment. Control food suspensions (“Mix + BSA”) were prepared similarly but without adding sterols.

Clam sampling Corbicula fluminea were collected in February 2009 in the upper basin of Lake Constance at a sampling site described by (Werner and Rothhaupt 2008). The clams were collected at a water depth of 2 - 3 m by scuba-diving. After separation of living individuals from debris, sand and gravel, they were placed in flow-through systems with filtered (< 30 µm), aerated lake water and pre-combusted sediment at an ambient temperature of 20°C. Clams were kept under these conditions for two weeks until the start of the growth experiment.

Growth experiments Adult bivalves invest most of their energy in reproduction (gametogenesis) and little in somatic growth (Soudant et al. 1999), therefore we used juveniles (initial dry mass 13.3 - 201.0 mg, including shells; size range 5 - 10 mm) which were not sexually mature to maximize somatic growth rates. The 28 d lasting experiment (24.02. - 24.03.2009) was carried out at 20°C. Glass beakers were filled with 200 ml of filtered lake water (0.45 μm pore-sized membrane filter) and about one centimetre of precombusted sediment (550°C for 5 h) to allow the clams to burrow. Clams were randomly transferred to each beaker. Each of the 10 food treatments consisted of 10 replicates, i.e. individual clams. Clams were fed daily with 3 mg C l-1 of the food suspensions or starved without adding food. The experiment comprised the following food treatments: the three cyanobacteria S. elongatus, A. variabilis, and A. clathrata, the three eukaryotic algae S. obliquus, N. limnetica, and Cryptomonas sp., a mixture of the three cyanobacteria (1/3 POC provided by each species, “Mix”) either unsupplemented or supplemented with BSA (control, “Mix + BSA”) or BSA and sterols (“Mix + BSA + Sterols”), and a mixture of the three eukaryotic algae. Water was exchanged daily to remove faecal pellets; sediment was exchanged once a week to reduce biofilm formation. Somatic growth rates (g) were determined as the increase in total dry mass from the beginning of the experiment (M0) to day 28 (Mt) over time (t) using the equation:

30 3 Sterol limitation of Corbicula

(ln M  ln M ) g  t 0 t

A subsample of clams (n = 47) was taken at the beginning of the experiment to estimate the individual fresh and dry mass after 24 h of freeze drying. Samples were weighed on an electronic balance (Mettler Toledo XP2U; ± 0.1 μg). The dry mass (DM, including shells) of clams at the start of each experiment was estimated from their actual fresh mass (FM, including shells) and previously established fresh-dry-mass regressions (DM = 0.625FM, R² = 0.994). Growth rates of clams were calculated as means (n = 10) ± standard deviation for each treatment (n = 10).

Analyses of food organisms and clam tissues Aliquots of the food suspensions were filtered onto precombusted glass-fibre filters (Whatman GF/F, 25 mm diameter) and analysed for particulate organic carbon (POC, n = 3 per treatment) using an NCS-2500 analyser (ThermoQuest). The carbon content of clams was determined by analysing freeze-dried soft-body tissues dissected from subsampled individuals at the end of the experiment (n = 3 per treatment).

For the analysis of sterols, glass-fibre filters loaded with ~1 mg POC of the food suspensions (n = 3) were sonicated and stored at -20 °C in a mixture of dichloromethane/methanol (2:1, v/v). Soft-tissues of freeze-dried clams (n = 3 for each treatment) were separated from their shell, weighed, crushed by mechanical shearing using a mortar, sonicated and subsequently stored at -20 °C in dichloromethane/methanol (2:1, v/v). Total lipids of clam tissue or algae suspensions were extracted three times from each sample using dichloromethane/methanol (2:1, v/v) and the pooled cell-free extracts were dried under a stream of nitrogen and saponified with 0.2 M methanolic KOH (70 °C, 1 h). Subsequently, sterols were partitioned into iso-hexane:diethyl ether (9:1, v/v), dried under a stream of nitrogen, and resuspended in iso-hexane. Sterols were analyzed by gas chromatography (GC) on a HP 6890 GC (Agilent Technologies, Waldbronn, Germany) equipped with a flame ionization detector and a HP-5 (Agilent, 30 m, 0.25 mm I.D., 0.25 µm film) capillary column. Details of GC configurations are given elsewhere (Martin-Creuzburg et al. 2009). Sterols were quantified by comparison to an internal standard (5α-cholestan). The detection limit was 20 ng of sterol. Sterols were identified by their retention times and their mass spectra, which were recorded with a gas chromatograph-mass spectrometer (Finnigan MAT GCQ) equipped with a fused silica capillary column (DB-5MS, Agilent, 30 m, 0.25 mm I.D., 0.25 µm). Sterols were analyzed in their free form and as their trimethylsilyl derivatives which were prepared by incubating 20 ml of iso-hexane sterol extract with 10 ml of N,O-bis(trimethylsilyl)trifluoroacetamide including 1% trimethylchlorosilane for 1 h at room temperature. Mass spectra were recorded between 50 and 600 amu in the EI ionization mode. Sterols were identified by comparison with mass spectra of reference substances purchased from Sigma or Steraloids and/or mass spectra found in a self-generated spectra library or in the literature (e.g. Toyama et al. 1952; Belanger et al. 1973; Goad and Akihisa 1997). The C-24 stereochemistry 31 3 Sterol limitation of Corbicula and the cis-trans isomery of sterols could not be identified with certainty and thus, if procurable, was adopted from the literature. The absolute amount of each sterol was related to the POC of the food sources or to the carbon content of clam soft-tissues and given as mean ± standard deviation.

Statistical analyses All statistical analyses were carried out using the statistical software package Statistica 6.0 (StatSoft). Differences among growth rates and among sterol levels in clam tissue were analyzed using one-way analyses of variance (ANOVA) and Tukey’s HSD post-hoc tests. The correlation between somatic growth rates of C. fluminea and sterol levels in clam tissue was assessed by linear regression analyses.

Results

Sterol composition of food sources Sterols were not detected in cyanobacterial food suspensions, i.e. neither in A. variabilis, A. clathrata, S. elongatus nor in the unsupplemented or BSA-treated cyanobacterial food mixtures. When enriched with sterols, the cyanobacterial food mixtures contained on average 38.3 ± 15.3 µg mg C-1 of sterols in total, consisting of cholesterol (cholest-5-en-3β-ol, 13.5 ± 3.9 µg mg C-1), ergosterol ((22E)-ergosta-5,7,22-trien-3β-ol, 7.8 ± 4.3 µg mg C-1) and stigmasterol ((22E)-stigmasta-5,22-dien-3β-ol, 17.1 ± 7.4 µg mg C-1). The green alga S. obliquus contained on average 13.6 ± 5.4 µg mg C-1 of sterols in total, the principal sterols were fungisterol (5α-ergost-7-en-3β-ol, 4.6 ± 1.8 µg mg C-1), chondrillasterol ((22E)-5α- poriferasta-7,22-dien-3-ol, 7.8 ± 3.4 µg mg C-1), and 22-dihydrochondrillasterol (5α- poriferast-7-en-3β-ol, 1.1 ± 0.2 µg mg C-1). In Cryptomonas sp., total sterols averaged 8.4 ± 4.2 µg mg C-1 and consisted of two principal sterols: brassicasterol ((22E)-ergosta-5,22-dien- 3β-ol, 2.7 ± 1.3 µg mg C-1) and stigmasterol (5.7 ± 2.9 µg mg C-1). N. limnetica contained 7.6 ± 8.9 µg mg C-1 of sterols in total, the principal sterols were cholesterol (5.4 ± 6.3 µg mg C-1), sitosterol (stigmast-5-en-3-ol, 1.2 ± 1.4 µg mg C-1) and fucosterol ((24E)-stigmasta-5,24(28)- dien-3-ol, 1.1 ± 1.2 µg mg C-1).

32 3 Sterol limitation of Corbicula

Fig. 8 a) Sterol content of invasive clam Corbicula fluminea after 28 days of growth on a cyanobacterial diet (data obtained for the three cyanobacterial diets, Synechococcus elongatus, Anabaena variabilis, and Aphanotece clathrata combined, n = 9) and on eukaryotic algae (data obtained for the three algal diets, Scenedesmus obliquus, Cryptomonas sp., and Nannochloropsis limnetica combined, n = 9). b) The sterol content of clams fed a mixture of all three cyanobacteria (Mix), a cyanobacterial mixture treated merely with bovine- serum-albumin (Mix + BSA), or a cyanobacterial mixture treated with BSA and sterols (coarse bar, Mix + BSA + Sterols) is shown for comparison (n=3). The shaded background represents tissue sterol levels of clams at the beginning of the experiment. Data represent means and standard deviations (SD). Bars labeled with the same letters are not significantly different based on Tukey’s HSD, P <0.05 following ANOVA; each panel represents a separate statistical analysis

Sterol composition of clam tissues Six different sterols were identified in clam tissues: cholesterol (40-72% of total sterols), brassicasterol (8-26%), corbisterol ((22E)-stigmasta-5,7,22-trien-3β-ol, 7-13%), campesterol (campest-5-en-3β-ol, 2-12%), stigmasterol (5-12%), and ergosterol (0-5%) (Tab. 3). The total sterol content of field-collected clams at the start of the experiment was on average 9.8 ± 2.1 µg mg C-1. Lowest sterol concentrations were found in clams fed A. variabilis (4.8 ± 1.6 µg mg C-1) and highest levels were measured in clams fed N. limnetica (12.9 ± 5.0 µg mg C-1) and in clams fed sterol supplemented cyanobacteria (12.6 ± 0.6 µg mg C-1). In general, total sterol levels of clams fed eukaryotic algae (10.5 ± 4.0 µg mg C-1, n = 9) were significantly higher -1 than those of clams fed cyanobacterial diets (5.6 ± 1.8 µg mg C , n = 9, ANOVA, F1,16 = 11.22, P = 0.004, Fig. 8a) and the total sterol content of clams fed sterol supplemented cyanobacteria was significantly higher than those of clams fed unsupplemented or merely

BSA-treated cyanobacteria (ANOVA, F2,6 = 17.95, P = 0.003; Tukey’s HSD, P < 0.05, Fig. 8b).

33 3 Sterol limitation of Corbicula

Tab. 3 Sterol composition of soft-body tissues of Corbicula fluminea fed either one of the three cyanobacteria (Synechococcus elongatus, Anabaena variabilis, Aphanotece clathrata) or one of the three eukaryotic algae (Scenedesmus obliquus, Cryptomonas sp., Nannochloropsis limnetica). In addition, the sterol composition of clams fed a mixture of all three cyanobacteria (Mix), a cyanobacterial mixture treated merely with bovine serum albumin (Mix + BSA), or a cyanobacterial mixture treated with BSA and sterols (Mix + BSA + Sterols) is presented. Values are given as means ± standard deviations (n = 3). Sterols are reported as µg sterol mg C-1. n.d. = not detectable.

Cryptomonas Cryptomonas

S. elongatus S.

N. limnetica

A. clathrata

A. variabilis

S. obliquus S.

+ Sterols

Starving

+ BSA

Start

Mix Mix Mix

sp.

5.21 5.67 5.22 2.98 3.85 3.34 3.60 9.28 4.31 2.30 7.33 Cholesterol ± 1.01 ± 0.56 ± 1.29 ± 1.14 ± 0.53 ± 1.49 ± 1.36 ± 3.64 ± 1.09 ± 2.10 ± 1.48

2.35 2.15 0.82 0.49 0.63 0.68 2.20 1.07 0.69 1.47 1.93 Brassicasterol ± 0.54 ± 0.91 ± 0.16 ± 0.16 ± 0.11 ± 0.46 ± 1.37 ± 0.78 ± 0.22 ± 1.56 ± 0.56

0.52 0.54 0.24 0.08 0.16 0.55 Ergosterol n.d. n.d. n.d. n.d. n.d. ± 0.46 ± 0.49 ± 0.42 ± 0.14 ± 0.28 ± 0.50

0.43 0.44 1.20 0.23 0.27 0.27 0.31 0.32 0.34 0.29 0.31 Campesterol ± 0.09 ± 0.11 ± 0.21 ± 0.07 ± 0.01 ± 0.13 ± 0.17 ± 0.14 ± 0.10 ± 0.08 ± 0.08

0.86 0.73 1.73 0.37 0.76 0.74 1.84 0.63 0.88 0.65 1.16 Stigmasterol ± 0.19 ± 0.29 ± 0.63 ± 0.09 ± 0.24 ± 0.43 ± 1.29 ± 0.24 ± 0.32 ± 0.14 ± 0.37

0.42 0.71 0.56 0.80 0.65 0.60 1.09 0.75 0.71 0.88 Corbisterol n.d. ± 0.32 ± 0.17 ± 0.10 ± 0.09 ± 0.24 ± 0.14 ± 0.58 ± 0.28 ± 0.25 ± 0.41

0.41 0.31 0.13 0.14 0.10 0.46 0.05 0.42 unknown n.d. n.d. n.d. ± 0.12 ± 0.30 ± 0.11 ± 0.13 ± 0.18 ± 0.34 ± 0.08 ± 0.39

9.78 10.25 9.80 4.78 6.40 5.67 8.78 12.85 7.06 5.63 12.58 Total sterol content ± 2.07 ± 3.34 ± 1.64 ± 1.62 ± 0.76 ± 2.74 ± 4.74 ± 4.95 ± 2.00 ± 1.55 ± 0.59

34 3 Sterol limitation of Corbicula

Fig. 9 Somatic growth rates of Corbicula fluminea (a) either starved or fed different cyanobacterial (Synechococcus elongatus, Anabaena variabilis, Aphanotece clathrata; white bars) or eukaryotic food sources (Scenedesmus obliquus, Cryptomonas sp., Nannochloropsis limnetica; gray bars). In addition, (b) growth rates of clams fed a mixture of all three cyanobacteria (Mix), a cyanobacterial mixture treated merely with bovine serum albumin (Mix + BSA), or a cyanobacterial mixture treated with BSA and sterols (coarse bar, Mix + BSA + Sterols) are presented. Data are means + SD, n = 10. Bars labeled with the same letters are not significantly different based on Tukey’s HSD, P > 0.05 following ANOVA

Growth of C. fluminea The dry mass of C. fluminea increased during the 28 d lasting experiment in all food treatments, leading to positive somatic growth rates. In contrast, the dry mass of starving individuals slightly decreased during the experiment (Fig. 9a). No clams died during the experiment. In general, clams fed one of the three cyanobacterial diets had lower growth rates than clams fed one of the three algal diets (ANOVA, F1,9 = 19.44 , P < 0.001; Tukey’s HSD, P < 0.05). Growth rates of clams fed the filamentous A. variabilis, the single-celled picocyanobacterium S. elongatus, the gelatinous A. clathrata or the unsupplemented or merely BSA-treated cyanobacterial diets (Mix, Mix + BSA) did not differ significantly (Tukey’s HSD, P > 0.05; Fig. 9b). Likewise, the significantly higher growth rates obtained with the eukaryotic algae S. obliquus, N. limnetica and Cryptomonas sp. did not differ significantly from each other (Tukey’s HSD, P > 0.05). Growth rates of C. fluminea fed sterol supplemented cyanobacterial food (Mix + BSA + Sterols) were significantly higher than those of clams fed unsupplemented cyanobacterial food (Mix, Mix + BSA) but did not differ from growth rates obtained with the eukaryotic food.

35 3 Sterol limitation of Corbicula

Discussion Nutritional requirements of benthic invertebrates have been poorly studied, in particular with regard to essential biochemicals. We show here that somatic growth of the invasive freshwater clam C. fluminea on cyanobacterial diets is constrained by the absence of sterols, as indicated by a growth-enhancing effect of sterol supplementation. This adds to previous findings showing that the growth of zooplankton (i.e. Daphnia) on cyanobacterial diets is constrained by the absence of sterols (Martin-Creuzburg et al. 2005b; 2008) and thus highlights the importance of considering sterols as potentially limiting nutrients in aquatic food webs.

It has been suggested that the capability of synthesizing sterols de novo is low or even absent in bivalve species, which suggests that they rely on sufficient supply with dietary sterols to cover their physiological demands (Goad 1981; Napolitano et al. 1993; Soudant et al. 1996a). Total sterol levels in the soft-body of clams fed the sterol-containing eukaryotic algae (Cryptomonas sp., N. limnetica or S. obliquus) were significantly higher than those of clams fed cyanobacterial diets, but did not differ from sterol levels determined in clams at the beginning of the experiment. In contrast, when grown on cyanobacterial diets, total sterol levels in clam tissues decreased compared to initial sterol levels. The supplementation of a cyanobacterial diet with sterols led to significantly increased sterol levels in clam soft- bodies, indicating an incorporation of supplemented sterols. In accordance with previous studies, the sterol composition of C. fluminea was dominated by cholesterol, with lesser amounts of brassicasterol, campesterol, corbisterol, ergosterol, sitosterol and stigmasterol (Duncan et al. 1987; Basen et al. 2011; Chijimatsu et al. 2011). Taking into account that bivalve species are presumably incapable of synthesising sterols de novo, sterols detected in soft-bodies of clams fed sterol-free cyanobacterial food were presumably incorporated and stored from dietary sources they had received prior to the experiment. The finding that sterols detected in clam tissues did not differ qualitatively, irrespective of the sterol composition of the food, suggests that C. fluminea is capable of converting dietary phytosterols to clam-specific sterols and that these sterols are functionally important in clam physiology. Besides sterols, long-chain polyunsaturated fatty acids (PUFAs) have been suggested to play an important role in bivalve nutrition (Soudant et al. 1996a; Wacker et al. 2002; Basen et al. 2011). As both, sterols and long-chain PUFAs, are either absent or hardly represented in cyanobacteria, the growth of bivalves feeding on cyanobacteria dominated diets might be simultaneously constrained by the availability of dietary sterols and certain PUFAs, as has been shown in laboratory experiments with Daphnia (Martin-Creuzburg et al. 2009). Further detailed investigations on bivalve nutrition are needed to separate effects mediated by dietary sterols from those mediated by dietary PUFAs to assess the relative importance of these essential nutrients.

In many aquatic ecosystems throughout the world, the phytoplankton is, at least seasonally, dominated by cyanobacteria (Reynolds and Walsby 1975; Oliver and Ganf 2000). Climate scenarios with rising temperatures, increased atmospheric CO2 supplies and increased

36 3 Sterol limitation of Corbicula periods of thermal stratification are expected to favour cyanobacterial dominance (Jöhnk et al. 2008; Paerl and Huisman 2008), which may also affect trophic interactions in aquatic food webs as cyanobacteria are a nutritionally inadequate food source for most aquatic consumers (De Bernardi and Giussani 1990; Martin-Creuzburg et al. 2008). Recently, it has been stated that cyanobacterial carbon, deposited during bloom conditions in a marine system, did not support the benthic detritus-based food web (Karlson et al. 2008; Nascimento et al. 2009). Bivalves are significantly involved in transferring organic matter from pelagic sources to the sediment and thus provide a crucial link between pelagic and benthic food web processes (Newell 2004; Vaughn et al. 2008). Hence, it is important to understand the impact of cyanobacterial mass developments on bivalve species to more accurately assess consequences for benthic food web processes. Our data suggests that the benthic-pelagic coupling between cyanobacteria and filter-feeding bivalves is at least partially constrained by a dietary sterol deficiency.

C. fluminea is considered to be one of the most important invaders in aquatic ecosystems in the last decades (Araujo et al. 1993; McMahon 2000). In contrast to some other bivalves, which are able to sort food particles according to their size, shape or surface structure (Bontes et al. 2007; Espinosa et al. 2010), C. fluminea is regarded as a non-selective suspension feeder (Way et al. 1990; Vaughn and Hakenkamp 2001) and thus is presumably not able to discriminate against nutritionally inadequate food particles. It remains to be tested whether C. fluminea is able to adjust its feeding or assimilation rate in order to gain more of a limiting nutrient (i.e. compensatory feeding). Moreover, Corbicula might be able to change its feeding mode from seston filtration to deposition feeding via its muscular foot (Vaughn and Hakenkamp 2001; Nichols et al. 2005) and in this way may potentially avoid the uptake of nutritionally inadequate food sources present in the water column. Thus, the predominance of cyanobacterial carbon in the water column may result in an increased utilization of benthic food sources to overcome a possible sterol limitation. However, the availability of adequate benthic food sources is potentially scarce, in particular during cyanobacterial bloom condition, as significant amounts of cyanobacterial carbon are deposited to the sediment. Considering our data, this suggests that somatic growth of C. fluminea and potentially other filter-feeding bivalves is constrained by a deficiency in dietary sterols when cyanobacteria dominate the phytoplankton. Consequently, the expected increase in the frequency of cyanobacterial bloom formation in response to global warming may severely impair the growth of filter-feeding bivalves.

Acknowledgements We thank S. Oexle and M. Bauer for the support with the experiments, R. Basen and 3 anonymous referees for helpful comments on the manuscript that improved its quality. This work was funded by the DFG (German Research Foundation) within the collaborative research centre CRC 454 “Littoral of Lake Constance”.

37 3 Sterol limitation of Corbicula

38 4 Corbicula in Lake Constance

Chapter 4

Impact of temperature and seston dynamics on growth and survival of Corbicula fluminea: A field study in Lake Constance

Timo Basen, Katja Maren Fleckenstein, Karsten Rinke, Karl-Otto Rothhaupt and Dominik Martin-Creuzburg

Abstract The invasive clam Corbicula fluminea was first recorded in Lake Constance (Germany) in 2003; since then its distribution spread in the lake. Clams being filter feeders largely depend on food supply in the water column. Thus, seasonal variation in water temperature and seston composition may determine the individual growth performance of benthic clams. To understand factors affecting growth and survival of C. fluminea in Lake Constance, field experiments were performed throughout the year 2010. Therefore, the temperature, algal composition and quality of seston (lipid, macronutrient composition) in different water depths were documented and growth rates of clams were estimated. Additionally, elemental and biochemical composition of clam tissue was analysed. Accompanying the field studies, standardized laboratory growth experiments were performed at temperatures between 4 and 25 °C. Using principal component analysis and linear models, correlations between seston- or clam tissue-parameters and clam growth rates were statistically investigated. In winter, clams were exposed to low temperatures (< 5 °C) combined with low food supply resulting in high mortality and no shell or tissue growth. Throughout the year, when temperatures exceeded 10 °C, growth of clams was recorded. Analyses showed that the growth of C. fluminea in Lake Constance is mainly determined by water temperature. Additionally, a strong linear correlation between growth and temperature was confirmed in laboratory tests. Since growth and development of C. fluminea occurs only above a temperature 10 °C, only 7 months are available for reproduction and growth of the clam population in Lake Constance. Corbicula fluminea may benefit from climate change which leads to milder winter temperatures, causing reduced mortality rates, and to an earlier onset of thermal stratification in spring, extending the growing season of C. fluminea. This may support further spread in Lake Constance.

Key words: essential lipids, food quality, invasive species, seasonal succession, temperature

39 4 Corbicula in Lake Constance

Introduction The introduction of new species in ecosystems is often highly correlated with a decrease of biodiversity and ecosystem stability, disturbance of food web processes, introduction of new diseases and parasites and displacement of native species (Sala et al. 2000; Chandra and Gerhardt 2008; Ellis et al. 2011; Poulin et al. 2011). Especially in freshwater systems, the introduction of new species is often caused by human activities (Pollux et al. 2003; Briski et al. 2011). A challenge of invasion biology is to identify the mechanisms enabling a successful establishment of non-native species in new habitats. In this context, global warming is discussed as a factor that can favour the successful migration of non-native species (Stachowicz et al. 2002). In addition to increasing temperatures, survival and growth of non- native species and thus the probability of a successful establishment in an invaded habitat may be affected by food availability and food quality. Food quality of phytoplankton for freshwater invertebrates depends on a variety of different factors like morphology or toxicity (Carmichael 1992; Van Donk et al. 2011), but also on the elemental (Elser et al. 2000; Sterner and Elser 2002) and biochemical composition (Brett and Müller-Navarra 1997; Martin-Creuzburg and Von Elert 2009). For instance, the biochemical composition of seston can change over the growth period (Tilman et al. 1982; Gächter and Bloesch 1985), due to the succession of species and to seasonal changes in growth conditions, e.g. light and nutrient availability (Harrison et al. 1990; Hessen et al. 2002). Seasonal changes in food quality for filter feeding cladocerans have been investigated in Lake Constance before (Wacker and Von Elert 2001; Hartwich et al. 2012).

Fig. 10 Map of Lake Constance. The circle shows the sampling site of Corbicula fluminea; the arrow indicates the experimental area.

40 4 Corbicula in Lake Constance

In the last decades, the occurrence of invasive species in freshwater systems has increased considerably (Richardson and Pysek 2008). In central Europe the River Rhine system is a region with ongoing invasions (Tittizer et al. 2000; Bij de Vaate et al. 2002) with potential threads for Lake Constance ecosystem (Hanselmann 2011). In particular, among the molluscs there are many invasive freshwater species that spread worldwide. Quite often they can form massive stocks in rivers and affect biomass and composition of primary producers, which in turn can have significant consequences for the ecosystem (Strayer 2010). Invasive bivalves can significantly affect the structure of native benthic communities, as has been shown for the successful invasion of Dreissena polymorpha in Lake Constance in the 1960s (Mörtl and Rothhaupt 2003). Another non-native bivalve in Lake Constance is the Asian clam Corbicula fluminea, which has been first recorded in the lake in 2003 (Werner and Mörtl 2004; see Fig. 10). Corbicula fluminea originates from East Asia and has been introduced in many freshwater ecosystems in North America and Europe in the past (Mouthon 1981; McMahon 1982; Den Hartog et al. 1992). A rapid expansion in European inland waters followed (Kinzelbach 1991; Turner et al. 1998). Although C. fluminea has successfully invaded Lake Constance, it has a reduced maximum size and only one reproductive period per year which may explain why the population size increased rather slowly since 2003 (Werner 2008). Nevertheless, when habitat conditions are favourable (i.e. sand, absence of D. polymorpha), it can build up massive stocks and dominate the local benthos community comprising up to 90% of the total benthic biomass (Werner and Rothhaupt 2007).

Lake Constance is a large, monomictic, oligotrophic, prealpine lake situated on the northern edge of the central European Alps. After winter circulation, thermal stratification induces phytoplankton growth (Rinke et al. 2010). Phytoplankton community in spring is dominated by centric diatoms and cryptophytes summer species are dominated by grazing resistant diatoms (Sommer 1985; Gaedke 1992).

Aim of this study was to investigate the relative importance of abiotic (i.e. temperature) and biotic (i.e. food quantity and quality) factors potentially influencing growth and survival of C. fluminea in Lake Constance. Therefore, field experiments were conducted during winter, spring, summer and autumn in 2010 (one in each season) in which field-collected clams were exposed in different depths in the lake for seven weeks. Clam growth rates were determined and related to prevailing temperatures and to different quantitative and qualitative seston characteristics, i.e. chlorophyll a, species composition, elemental (carbon, nitrogen, phosphorus) and biochemical composition (fatty acids, sterols). Each field experiment was accompanied by a standardized laboratory growth experiment conducted at different temperatures and defined food supply in order to assess conditional changes among field- collected clams during the season.

41 4 Corbicula in Lake Constance

Materials and methods

Clam sampling Juvenile C. fluminea were collected in the upper basin of Lake Constance at a sampling site close to the Austrian border (E 9°37′/N 47°30′) as described by Werner and Rothhaupt (2008). The clams were collected at a water depth of 2 - 3 m by scuba-diving 1 - 2 weeks prior to the experiments (see Tab. 5). After separation of living individuals from debris, sand and gravel, they were placed in flow-through systems with filtered (< 30 µm), aerated lake water and pre-combusted sediment at an ambient water temperature of 9 - 12 °C until the start of the respective growth experiments.

Growth experiments For growth experiments we used juveniles which were not sexually mature to maximize somatic growth rates. Prior to the experiments, each clam was twofold measured (length, width, height) with an electrical calliper (Digi-Met, IP 65), weighed (Mettler UMT2, ± 0.1 µg) and marked individually with nail polish. Field and laboratory experiments were conducted simultaneously during winter, spring, summer and autumn 2010, the former lasted seven weeks and the latter four weeks (see Tab. 5). Slightly varying durations of the field experiments (47 - 51 days) were due to bad weather conditions during the intended days of completion.

Fig. 11 Setup of field experiments. Clams were placed in boxes, installed on a chain, adjustable in depth via deflector rolls and floating buoys.

42 4 Corbicula in Lake Constance

Field experiment For field experiments, 400 clams were randomly grouped into subsamples of 20 clams which were required for each treatment and stored in a flow through system with lake water until the experiments started (< 12 h). At the beginning of the experiments, these individuals were placed in plastic boxes (0.01 m²) containing sediment (1.5 l volume, grain size < 2 mm) collected at the clam sampling site. Boxes were attached to a deflector roll construction with floating buoys to enable a flexible adjustment to water level variations during the experiments (Fig. 11). Four chains each with five boxes were exposed in the littoral of Upper Lake Constance (E9°12.163'/N47°41.505'; Fig. 10, see Mörtl and Rothhaupt 2003 for detailed description of the study site); the boxes were exposed at 2, 4, 6, 8 and 10 m below surface. After 7 weeks of exposure, boxes were recovered and clams were separated from sediment.

Water temperature was recorded hourly with data loggers attached to each box at one of the four chains (HOBO UA-002-64, Onset). Seston samples were taken weekly at five different water depths (2, 4, 6, 8 and 10 m below surface) close to the boxes using a standard water sampler. The water was filtered on site (< 140 µm) and transported into the laboratory for chemical analyses.

During the field experiments, chlorophyll a (chlA) measurements were conducted twice a week between 0 and 12 m water depth using a multi-channel fluorescence probe (FluoroProbe, bbe Moldaenke). With this probe four different algae classes (Cryptophyte, Chlorophyte, Cyanobacteria, diatoms) could be classified via their specific fluorescence profiles. Probe parameters and calibration procedures are given elsewhere (Rinke et al. 2009). Data samples for time spans between experimental periods were measured weekly, only datapoints from 17.8. - 28.9.2010 were taken from routine sampling from “Überlinger See” (E9°7.743’/ N47°45.453’). A mean of data points for depth step ± 1 m were taken for specific depths.

Laboratory experiments For laboratory growth experiments the eustigmatophyte Nannochloropsis limnetica (SAG 18.99) was cultivated semi-continuously in Cyano medium (Jüttner et al. 1983) in aerated 5 l vessels at a dilution rate of 0.2 d–1 at 22 °C with illumination at 160 μmol quanta m–2 s–1. Food suspensions were prepared every other day by concentrating the cells via centrifugation (3000 g for 10 min) and resuspension in fresh medium. Carbon concentrations of the food suspensions were estimated from photometric light extinctions (800 nm) and from carbon-extinction equations determined prior to the experiment.

The 30 days lasting laboratory experiments were carried out in climate chambers at temperatures between 4 and 25°C (see Tab. 5) in glass beakers filled with 1 l of filtered lake water (0.45 μm pore-sized membrane filter) and about one centimetre of precombusted

43 4 Corbicula in Lake Constance sediment (grain size < 2 mm, 550°C for 5 h) to allow the clams to burrow. Each of the five temperature treatments consisted of ten replicates, i.e. individual clams. Clams were taken from same population used for field experiments and were randomly transferred to each beaker and fed daily with saturating amounts (4 mg C l-1) of the N. limnetica food suspension. Water was exchanged every other day to remove faecal pellets; sediment was exchanged once a week to reduce biofilm formation. Beakers were slightly aerated, to reduce sedimentation of added algae.

Water temperatures in the experimental beakers were documented using data loggers (HOBO). Aliquots of the N. limnetica food suspensions were taken three times during each experiment for chemical analyses.

Determination of clam parameters

Somatic growth rates (gm) of C. fluminea in field and laboratory experiments were determined as the increase in total dry mass of surviving clams from the beginning (M0) to the end of the experiments (Mt) using the equation:

(ln M t  ln M 0 ) g  m t

To estimate the initial dry mass of clams at the beginning of each experiment (M0) a fresh- dry-mass regression was established prior to the experiments using the fresh and dry masses (after 48 h of freeze-drying) of 50 individuals. Growth rates of clams were calculated as means ± standard deviations (SD) for each temperature treatment (laboratory experiments; n = 10) and for each water depth (field experiment; n = 4), respectively.

-1 Shell growth (gL, in mm d ) of C. fluminea was estimated from differences in shell lengths at the start (L0) and at the end of each experiment (Lt). The determination of growth rates via the increase in shell heights and widths revealed similar results as the determination of growth rates via the increase in shell lengths and, thus, they are not presented here.

(Lt  L0 ) g L  t Clams used for chemical analyses (carbon, nitrogen, phosphorus, fatty acids, sterols; n = 9) were dissected and the soft body was separated from the shells. The percentage of soft body dry mass (m) on total dry mass including shells (Mt) was estimated as tissue condition index (TCI).

m TCI  100 M t

44 4 Corbicula in Lake Constance

Chemical analyses Aliquots of the N. limnetica food suspensions or of lake water containing seston (< 140 µm) were filtered onto precombusted glass-fibre filters (Whatman GF/F, 25 mm diameter) and analysed for particulate organic carbon (POC) and nitrogen (PON) using an EuroEA3000 elemental analyzer (HEKAtech GmbH). The water volume required for the analyses of lake seston was estimated from concomitantly conducted measurements of the Secchi depths, which is roughly correlated with seston concentrations.

To determine the elemental composition of clams, soft body tissues of freeze-dried clams were separated from shells and weighed. The C and N content of soft body tissues was expressed as molar C:N ratios.

For the determination of particulate phosphorus, aliquots of the N. limnetica food suspension or of lake water containing seston (< 140 µm) were collected on acid-rinsed polysulfone filters (HT-200; Pall). Clam soft-tissues were solubilized by mechanical shearing using a mortar and by ultrasound treatment in ultrapure water. Subsequently, clam, seston and algae samples were digested using a solution of 10 % potassium peroxodisulfate and 1.5 % sodium hydroxide for 60 min at 121 °C. Soluble reactive phosphorus was determined using the molybdate-ascorbic acid method (Greenberg et al. 1985).

For lipid analysis of N. limnetica or lake seston, glass-fibre filters loaded with either ~0.5 mg C for fatty acids or ~1.0 mg C for sterols were sonicated and stored at -20 °C in a mixture of dichloromethane/methanol (2:1, v/v). Soft-tissues of freeze-dried clams were separated from their shell, weighed, crushed by mechanical shearing using a mortar, sonicated and subsequently stored at -20 °C in dichloromethane/methanol (2:1, v/v). Total lipids of clam tissue or algae suspensions were extracted three times from each sample using dichloromethane/methanol (2:1, v/v) and the pooled cell-free extracts were dried under a stream of nitrogen and saponified with methanolic KOH (0.2 M, 70 °C, 1 h) for sterols or were transesterified with methanolic HCl (3 M, 60 °C, 15 min) for the analysis of fatty acids.

Subsequently, fatty acid methylesters (FAME) were extracted three times with iso-hexane (2 ml); the neutral lipids were partitioned into iso-hexane:diethylether (9:1, v:v). The lipid- containing fraction was evaporated to dryness under N2 and resuspended in iso-hexane (10 – 20 µl). Lipids were analysed by gas chromatography-flame ionization detection (GC-FID; Hewlett-Packard 6890, Agilent Technologies) equipped with a DB-225 (J&W Scientific, 30 m, 0.25 mm inner diameter, 0.25 µm film) capillary column for FAME analysis and a HP-5 (Agilent, 30 m, 0.25 mm inner diameter, 0.25 µm film) capillary column for sterol analysis. Details of GC configurations are given elsewhere (Martin-Creuzburg et al. 2009; 2010). Lipids were quantified by comparison to internal standards (17:0 ME and 23:0 ME; 5a-cholestan) of known concentrations, considering response factors determined previously with lipid standards (Sigma or Steraloids). Lipids were identified by their retention times and their mass spectra, which were recorded with a GC-mass spectrometer (7890A GC system, 5975C inert MSD, Agilent Technologies) equipped with a fused-silica capillary column (DB-225MS, J&W for FAMEs; DB-5MS, Agilent for sterols; GC configurations as described for FID). Sterols 45 4 Corbicula in Lake Constance were analysed in their free form. Spectra were recorded between 50 and 600 amu in the electron impact (EI) ionization mode. The limit for quantitation of fatty acids and sterols was 20 ng. The absolute amount of each sterol was related to the POC of the food sources or to the carbon content of clam soft-tissues and expressed as means ± SD.

Statistical analyses Mortality rates of clams in the field trials were transformed (square root, arcsin, see Underwood 1997) and analysed using analysis of variance (ANOVA, Statistica, Sigmastat 6.0) followed by Tukey’s HSD test (p <0.05). Differences between TCIs at the beginning of the consecutive field experiments were also analysed using ANOVA. An analysis of covariance (ANCOVA, STATISTICA) was used to analyse the effects of water temperature and date of the experiment on clam growth rates in laboratory experiments.

Principal component analyses (PCA) and linear models were carried out using the statistical software package R (R Development Core Team 2006). Analysed parameters used in the four PCAs are presented in Tab. 4. First PCA was calculated on measured seston parameters of Lake Constance (PCA1). For each of the 18 variables, 140 data points (5 * 7 * 4; depth, weeks, experiments) were incorporated into the data set. A second PCA was performed on clam growth rates, length increase, TCI, water temperature and seston parameters (PCA2). Each of the 21 variables were represented by 20 data points (5 depth * 4 experiments). To estimate possible correlations between clam growth rates and clam tissue parameters, a third PCA was performed (PCA3). Each of the 15 variables were represented by 20 data points (5 depth * 4 experiments). For laboratory experiments, a fourth PCA was performed on clam growth rates and tissue parameters with 16 variables represented by 20 data points (5 temperatures * 4 experiments, PCA4).

In addition to PCAs, general linear models were calculated to test for significant differences among variables (for PCA2,3,4). To explain the residual variance in the dataset, the residuals of the linear model of the growth rates in dependence on the temperature were correlated in linear models with various seston variables (for PCA2) and two-factor linear models were calculated subsequently. Akaikes information criterion (AIC) was used to evaluate the goodness of fit of the different models (Akaike 1974).

46 4 Corbicula in Lake Constance

Tab. 4: Variables and abbreviations used in principal component analyses.

variables value PCA 1 2 3 4 ALA α-linolenic acid content seston µg mg C-1 ARA arachidonic acid content seston µg mg C-1 chlA total chlorophyll a concentration seston µg l-1 CN molar carbon to nitrogen ratio seston - CP molar carbon to phosphorus ratio seston - clamALA α-linolenic acid content clam tissue µg mg DW-1 clamARA arachidonic acid content clam tissue µg mg DW-1 clamC carbon content clam tissue mg mg TG-1 clamCN molar carbon to nitrogen ratio clam tissue - clamCP molar carbon to phosphorus ratio clam tissue - clamDHA docosahexaenoic acid content clam tissue µg mg DW-1 clamEPA eicosapentaenoic acid content clam tissue µg mg DW-1 clamFA total fatty acid content clam tissue µg mg DW-1 clamN nitrogen content clam tissue mg mg TG-1 clamP phosphorus content clam tissue mg mg TG-1 clamPUFA PUFA content clam tissue µg mg DW-1 clamST total sterol content clam tissue µg mg DW-1 crypto cryptophyta seston µg l-1 cyan cyanobacteria seston µg l-1 depth depth seston m diat diatoms seston µg l-1 DHA docosahexaenoic acid content seston µg mg C EPA eicosapentaenoic acid content seston µg mg C green green algae seston µg l-1 growth.rate growth rate via dry weight increase clam tissue d-1 lengthincrement length increase clam tissue mm d-1 mean.temp mean water temperature seston °C POC particulate organic carbon content seston µg l-1 PON particulate organic nitrogen content seston µg l-1 Ppart particulate organic phosphorus content seston µg l-1 PUFA PUFA content seston µg mg C-1 TCI tissue condition index clam tissue - temp water temperature seston °C tot.FA total fatty acid content seston µg mg C tot.ST total sterol content seston µg mg C

47 4 Corbicula in Lake Constance

Results

a b

c d

e f

Fig. 12 Conture plots of water temperature (a), total chlorophyll a (b) and algal composition (c-f) at the study site in Lake Constance in 2010. The values between sampling dates and between sampled water depths were interpolated linearly. The four experimental periods are accentuated.

Field experiments During winter 2010 water of Lake Constance was mixed well and temperatures of 4 - 5 °C were measured in the water column at the study site until early April (Fig. 12a). Spring warming increased the temperature in the upper water layers during the second experimental period up to 13 °C at 2 m water depth respectively 10 °C at 10 m depth. The temperature gradient from top to bottom increased further during summer and was highest in July, during the third experimental period, with water temperatures of up to 25 °C in the upper water layers and up to 19 °C in 10 m depth. Temperature decrease started from 48 4 Corbicula in Lake Constance

August onwards. In late autumn, i.e. during experimental period four, water temperatures were between 8 and 12 °C (mean temperatures during the experiments are presented in Tab. 5).

The total chlA concentrations (Fig. 12b) roughly followed the seasonal temperature changes. From January to April, chlA was barely detectable (<1 µg l-1). In April, chlA concentrations increased up to 4 µg l-1, with maximum concentrations of 5 - 7 µg l-1 measured in June and July. This chlorophyll maximum shifted to deeper water layers with time (Fig. 12b). A second period of higher chlA concentrations in the upper water layer was recorded between mid August and mid October. Afterwards, the chlA concentrations dropped again to 1 - 2 µg l-1.

The abundance of the four different algae groups (Chlorophyceae, Cryptophyceae, cyanobacteria and diatoms) showed a time- and depth-dependent pattern (Fig. 12c - f). Green algae were abundant from June to October with the highest concentrations occurring in 0 - 6 m water depth (maximum 3.4 µg l-1). Diatoms were abundant already at the end of April and then again in June and July with a maximum of 4.5 µg l-1 in 6 m depth. In late summer, a massive diatom assemblage was present in water layers between 6 and 8 m (Fig. 12f). Diatoms constituted the highest proportion of total chlA during the whole growth period. Measurable concentrations of Cryptophyceae (2 µg l-1) were located only in 4 - 6 m depth in May. In the other months, the concentration was ≤ 1 µg l-1. Over the whole growth period, the abundance of cyanobacteria was low and detectable concentrations were found only in the upper 2 - 4 m. In early April, the highest concentration of cyanobacteria with 0.45 µg l-1 was detected; otherwise the concentration was between 0 and 0.2 µg l-1.

49 4 Corbicula in Lake Constance

a During the first field experiments in winter, POC concentrations in the water column were low (0.1 mg -1 C l ). In the course of the second experiment (spring), the POC concentration increased from initially 0.2 mg C l-1 up to 0.5 mg l-1 C in early May. In the summer experiment, a POC maximum b was present between 4 and 8 m (see Fig. 13a). The highest POC concentration occurred in July with 0.9 mg C l-1 in 6 m depth. Subsequently, the POC concentration decreased again until the end of the experiment to 0.4

mg C l-1. Carbon concentrations c from October to December (experiment four) were similar to those measured during the winter experiment (i.e. < 0.2 mg C l-1). Seston carbon to nitrogen ratios (C: N) were higher during winter (9-12) and summer (4-11) than during

spring (6-7) and autumn (6-9). Fig. 13 Conture plots of particulate organic carbon (a) and Seston carbon to phosphorus ratios molar carbon to nitrogen (b) and carbon to phosphorus (C: P) showed a strong seasonal ratios (c) at the study site in Lake Constance in 2010. The variation (Fig. 13c). During the first values between sampling dates and between sampled water depths were interpolated linearly. The four two experiments (i.e. in winter and experimental periods are accentuated spring), the C: P ratios were most of the time and in all depths below

180. During summer, the C: P ratios of seston increased to a maximum of 380 in the upper water layers and decreased again until the end of July. In November and December, C:P ratios dropped again to values similar to the early growth period (see Fig. 13c).

The concentrations of fatty acids (total FA) and polyunsaturated fatty acids (PUFAs) in lake seston showed similar patterns with low values during winter and autumn and high values during spring and summer (Fig. 14a, c; b, d). For α-linolenic acid (ALA), arachidonic acid (ARA), eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) highest values were measured also during spring and summer (i.e. in experiment two and three). ALA was

50 4 Corbicula in Lake Constance detectable in spring and autumn (i.e. in experiment two and four). Seston ARA concentrations peaked during the summer (experiment three) with concentrations of up to 4 µg mg C-1. EPA and DHA were both present during spring and summer; the highest concentrations were detected in summer (experiment three) in 6 m water depth. Total sterol concentrations of lake seston were low during most times of the year (< 4 µg l-1; Fig. 14f); somewhat higher values were detected only during short times in spring and summer in the lower and in the upper water layers. A similar pattern was observed when total sterol concentrations were related to the available POC (i.e. µg mg C-1; Fig. 14e). The carbon- related concentrations of total fatty acids, total PUFAs, ALA and DHA were highest during spring (experiment two). During summer, these values were reduced, contrary to ARA and EPA which showed highest concentrations during experiment three (Fig. 15b, d), however lipid content for ARA, DHA and EPA in lake seston (µg l-1) during summer experiment reflected highest values found during the year (Appendix Fig. S1).

51 4 Corbicula in Lake Constance a b

c d

e f

Fig. 14 Seasonal changes in fatty acid concentrations (all in µg mg C-1 and µg l-1) in seston of Lake Constance sampled at the study site in 2010. Shown are the total fatty acid concentrations (a, b) and the concentrations of polyunsaturated fatty acids (PUFA; c, d) and of total sterols (e, f). The values between sampling dates and between sampled water depths were interpolated linearly. The four experimental periods are accentuated.

52 4 Corbicula in Lake Constance a b

c d

Fig. 15 Seasonal changes in fatty acid concentrations (all in µg mg C-1) in seston of Lake Constance sampled at the study site in 2010. Shown are the α-linolenic acid (ALA; a), arachidonic acid (ARA; b), docosahexaenoic acid (DHA; c) and eicosapentaenoic acid (EPA; d). The values between sampling dates and between sampled water depths were interpolated linearly. The four experimental periods are accentuated.

The environmental conditions during the first field experiment in winter were characterized by cold water temperatures and low chlA and POC concentrations in all depths at the study site. During the second experiment in spring, gradual warming initialized the growth of phytoplankton. The spring bloom was dominated by diatoms and Cryptophyceae. In addition, the second experiment was characterized by low C:P and C:N ratios of seston and a high concentration of PUFA (especially ALA, DHA and EPA). While ALA was represented mainly in the spring seston, ARA concentrations peaked in July. The third experiment comprised the summer maximum of phytoplankton, which was also dominated by diatoms and Chlorophyceae. Water temperatures were highest in the third experiment, wherein a temperature and chlorophyll gradient was recorded with depth. The C:P and C:N ratios of seston and POC concentrations were was particularly high in the summer experiment. After cooling of Lake Constance during the last experiment in autumn, water temperatures were comparable to those measured during spring and, consequently, the concentrations of chlA, algal classes, and of lipid compounds dropped also to spring levels.

53 4 Corbicula in Lake Constance

Mortality of C. fluminea The mortality of C. fluminea in the four field experiments differed significantly (ANVOA, F =1112.6; p < 0.001, Tukey´s HSD p < 0.05). During the first field trial in winter 93.0 ± 1.4% of the clams died across all depths. In the second field experiment in spring, the mortality of clams was significantly lower (28.0 ± 1.4 %). During the summer

experiment, altogether only four individuals died, the average mortality was 1.0 ± 0.0 %. Mortality during the fourth experiment in autumn (1.0 ± 0.1 %) was not significantly different from the mortality observed during the summer experiment.

Growth of C. fluminea During the winter experiment, growth was recorded neither in dry

weight nor in shell growth rate

(Tab. 5, Fig. 16). In the second field experiment in spring, growth rates of clams increased marginally. Highest growth rates of C. fluminea, irrespective of whether they were calculated via the dry weight increase of soft-body tissues or via the increase in shell length, were recorded in the third experiment during summer (Fig. 16a, b; red circles). During the fourth

experiment in autumn, tissue and Fig. 16 Somatic growth rates (calculated via the increase in A dry weight or B shell length) and tissue condition indices (TCI; C) of shell growth rates decreased again. Corbicula fluminea in the four field experiments (winter, spring, summer, autumn) plotted against the water temperature. Dots represent mean values ± SD for each box (i.e. means of living individuals at the end of an experiment, max. n = 20).

54 4 Corbicula in Lake Constance

The TCI of clams at the beginning of the experiments did not differ from each other (4.64% -

5.24%, ANOVA, F = 0.4384, p = 0.7265). However, there were significant differences at the end of the field trials (Tab. 16c). The lowest proportion of clam soft body was found at the end of the first experiment in winter with values between 2.98% and 4.17% when TCI values decreased during the experiment. Mean values of the second experiment in spring were between 4.64% and 5.38% and thus changed little during the experiment. During the summer experiment the proportion of soft body increased up to 6.53% - 7.99% and during the autumn experiment the clam TCI increased up to 5.75% -6.33% (Tab. 5).

Tab. 5 (part1) Time table of experiments conducted in Lake Constance in different depths (F1-5, “Field”) and in climate chambers at different temperatures (L1-5, “Laboratory”). Water temperatures recorded at the study site in Lake Constance and in the climate chambers are given as means ± standard deviations. Mortality of clams is given as overall percentages.

Experiment 1 Experiment 2 Experiment 3 Period 4

„Winter“ „Spring“ „Summer“ „Autumn“

Field Laboratory Field Laboratory Field Laboratory Field Laboratory

25.01.10 25.01.10 30.03.10 31.03.10 17.06.10 16.06.10 14.10.10 13.10.10 duration ------(days) 17.03.10 22.02.10 20.05.10 28.04.10 03.08.10 14.07.10 02.12.10 10.11.10 (51) (28) (51) (28) (47) (28) (49) (28)

4.64 3.85 9.37 4.08 20.18 3.75 10.58 3.57 F1 L1 ± 0.19 ± 0.35 ± 2.35 ± 0.32 ± 2.85 ± 0.39 ± 1.71 ± 0.26

4.59 10.17 8.91 10.46 18.30 9.71 10.52 9.39 F2 L2 ± 0.18 ± 1.06 ± 2.22 ± 0.68 ± 2.45 ± 0.52 ± 1.71 ± 0.43

water temp. 4.61 15.51 8.56 16.04 16.45 15.90 10.52 14.90 F3 L3 (° C) ± 0.17 ± 0.18 ± 2.13 ± 0.26 ± 2.10 ± 0.62 ± 1.70 ± 0.26

4.52 17.99 8.11 18.85 14.78 20.20 10.41 19.69 F4 L4 ± 0.17 ± 0.54 ± 2.03 ± 0.39 ± 1.84 ± 0.47 ± 1.69 ± 0.24

4.52 24.50 7.68 25.04 13.54 24.48 10.36 24.18 F5 L5 ± 0.17 ± 0.50 ± 1.83 ± 0.22 ± 1.61 ± 0.22 ± 1.68 ± 0.51

F1 L1 93 0 29 90 1 0 4 0

F2 L2 91 0 26 0 0 0 0 0 mortality F3 L3 91 0 30 0 1 0 3 10 (%) F4 L4 94 0 28 10 1 0 6 0

F5 L5 95 0 28 0 1 10 0 0

55 4 Corbicula in Lake Constance

Tab. 5 (part2) Time table of experiments conducted in Lake Constance in different depths (F1-5, “Field”) and in climate chambers at different temperatures (L1-5, “Laboratory”). Somatic growth rates (calculated via the increase in dry weight or shell length) and tissue condition indices (TCI; number in brackets are represented, if differing from n = 9) are presented.

Experiment Experiment Experiment Experiment Experiment Experiment Period 4 Period 4 1 2 3 1 2 3 „Autumn“ „Autumn“ „Winter“ „Spring“ „Summer“ „Winter“ „Spring“ „Summer“

Field Laboratory Field Laboratory Field Laboratory Field Laboratory

0.000 -0.001 0.002 0.024 0.001 0.007 -0.001 F1 L1 -0.001 ± 0.000 ± 0.002 ± 0.000 ± 0.001 ± 0.001 ± 0.001 ± 0.002

0.000 0.006 0.002 0.007 0.021 0.004 0.007 0.006 F2 L2 ± 0.000 ± 0.003 ± 0.000 ± 0.002 ± 0.000 ± 0.002 ± 0.001 ± 0.002 somatic 0.000 0.024 0.001 0.024 0.017 0.022 0.007 0.017 growth F3 L3 rate ± 0.000 ± 0.004 ± 0.000 ± 0.005 ± 0.001 ± 0.009 ± 0.001 ± 0.005 (d-1) 0.000 0.030 0.001 0.029 0.015 0.032 0.007 0.031 F4 L4 ± 0.000 ± 0.007 ± 0.000 ± 0.007 ± 0.000 ± 0.009 ± 0.001 ± 0.011

0.000 0.051 0.001 0.053 0.013 0.037 0.007 0.041 F5 L5 ± 0.000 ± 0.008 ± 0.000 ± 0.017 ± 0.000 ± 0.012 ± 0.000 ±0.017

0.000 -0.001 0.003 0.083 0.001 0.011 -0.001 F1 L1 -0.002 ± 0.001 ± 0.003 ± 0.001 ± 0.003 ± 0.004 ± 0.001 ± 0.002

-0.001 0.003 0.002 0.006 0.069 0.003 0.009 0.003 F2 L2 ± 0.002 ± 0.002 ± 0.001 ± 0.003 ± 0.002 ± 0.005 ± 0.003 ± 0.003 shell growth -0.004 0.040 0.002 0.042 0.051 0.034 0.009 0.022 F3 L3 rate ± 0.006 ± 0.006 ± 0.000 ± 0.007 ± 0.001 ± 0.013 ± 0.003 ± 0.010 (mm d-1) -0.001 0.051 0.002 0.054 0.043 0.070 0.011 0.064 F4 L4 ± 0.000 ± 0.012 ± 0.000 ± 0.011 ± 0.001 ± 0.013 ± 0.002 ± 0.013

-0.001 0.118 0.001 0.113 0.034 0.101 0.011 0.096 F5 L5 ± 0.000 ± 0.015 ± 0.000 ± 0.027 ± 0.000 ± 0.018 ± 0.001 ± 0.022

3.82 4.95 4.94 6.53 4.05 6.20 4.01 F1 L1 3.91 (1) ± 0.09 (6) ± 0.99 ± 0.64 ± 0.46 ± 0.77 ± 0.97 ± 0.51

4.17 10.12 5.38 10.92 7.13 7.82 6.33 6.31 F2 L2 ± 0.94 (7) ± 1.73 ± 0.71 ± 2.88 ± 0.40 ± 2.48 ± 0.86 ± 1.61

TCI 2.98 13.54 4.64 15.84 6.92 14.22 6.60 10.07 F3 L3 (%) ± 1.38 (7) ± 1.80 ± 0.58 ± 2.75 ± 0.56 ± 2.46 ± 0.95 ± 1.90

4.00 14.35 4.80 16.47 8.00 12.35 6.04 9.11 F4 L4 ± 0.77 (5) ± 2.07 ± 0.98 ± 3.83 ± 0.40 ± 3.00 ± 1.80 ± 1.16

3.99 11.07 5.04 11.69 7.99 10.48 5.75 7.50 F5 L5 ± 0.56 (3) ± 0.56 ± 0.74 ± 1.92 ± 0.83 ± 1.56 ± 0.92 ± 1.73

56 4 Corbicula in Lake Constance

Fig. 17 Biplot of the first principal component analysis (PCA) of weekly measured seston parameter of Lake Constance in 2010 (for detail see Tab. 4). Left and lower axis represent data values (numbers), right and upper axis represents variables (red arrows). Arrows represents loading of variables on first and second principal component PCA1 and PCA2. Numbers are grouped as follows: 1-35 experiment one (winter), 36-70 experiment two (spring), 71-105 experiment three (summer), 106-140 experiment four (autumn).

Principal Component Analysis and linear models The high temporal resolution analysis of seston parameters (Fig. 17) revealed positive correlations among lipid compounds, i.e. total FA, total PUFA, ALA, EPA, DHA and total sterols, with the exception of ARA, which was strongly correlated with water temperature and POC. The C:N ratio of seston was negatively correlated with the concentrations of lipid compounds (total FA, total PUFA, ALA, EPA, DHA, with exception of ARA) and total sterols (Fig. 17, Tab. 6). The water temperature was not only correlated positively with the concentration of ARA, but also with the C:P ratio of seston, the abundance of green algae and diatoms, and the POC, PON and total chlA concentration in the lake.

57 4 Corbicula in Lake Constance

Fig. 18 Biplot of 2. principal component analysis (PCA), with mean values of seston parameters and growth rate, length increment and tissue condition index of clams per experimental period (see detail Tab. 4). Left and lower axis represent data values (numbers), right and upper axis represents variables (red arrows). Arrows represents loading of variables on first and second principal component PCA1 and PCA2. Numbers are grouped as follows: 1 - 5 experiment 1 (winter), 6 - 10 experiment 2 (spring), 11 - 15 experiment 3 (summer), 16 - 20 experiment 4 (autumn).

Analysis of mean values of growth experiments with seston parameters and clam fitness factors (Fig. 18) revealed a positive correlation among growth parameters, i.e. dry weight, shell length, and TCI. Temperature, the occurrence of green algae (green), and the concentration of ARA in lake seston were also positively correlated with growth parameters, as indicated by a similar direction of the respective arrows in the PCA plot (Fig. 18). The first component (PCA1) explains 45% of the total variance of the data and is negatively correlated with almost all variables (arrows), with the exception of C:N and depth. PCA2 explains 31% of the total variance and is negatively correlated with all growth parameters (growth rate, length, TCI), mean temperature, elemental seston stoichiometry (C:P and C:N ratio), seston quantity (POC, chlA), abundance of green algae and cyanobacteria, and with seston concentrations of ARA.

58 4 Corbicula in Lake Constance

Tab. 6 Linear models of somatic growth rates of Corbicula fluminea correlated with different seston parameters, clam parameters for field and laboratory experiments. Significant models in bold.

Field experiments Laboratory experiments factor R² p slope factor R² p slope

seston temp 0.937 <0.001 0.002 mean.temp 0.937 <0.001 0.002 ARA 0.789 <0.001 0.006 CP 0.640 <0.001 0.000 green 0.570 <0.001 0.015 chlA 0.522 <0.001 0.005 POC 0.486 0.001 0.033 diatom 0.354 0.006 0.006 PON 0.195 0.051 Ppart 0.133 0.115 tot.FA 0.115 0.143 CN 0.113 0.147 cyano 0.038 0.411 PUFA 0.031 0.457 crytpo 0.009 0.699 tot.ST 0.008 0.713

EPA 0.000 0.955 ALA 0.000 0.972 DHA 0.000 0.977

clam length 0.960 <0.001 0.289 length 0.957 <0.001 0.413 TCI 0.645 <0.001 0.004 TCI 0.337 0.007 0.003 clamCN 0.298 0.013 0.004 clamCP 0.484 <0.001 0.000 clamN 0.292 0.014 -0.223 clamDHA 0.453 0.001 -0.002 clamCP 0.217 0.051 clamP 0.400 0.004 -6.793 clamEPA 0.169 0.072 clamCN 0.348 0.008 0.008

clamST 0.154 0.120 clamN 0.337 0.009 -0.544 clamFA 0.135 0.112 clamC 0.189 0.063 clamARA 0.131 0.116 clamST 0.148 0.273 clamP 0.122 0.156 clamFA 0.134 0.113 clamPUFA 0.120 0.135 clamPUFA 0.120 0.134 clamALA 0.118 0.139 clamEPA 0.091 0.195 clamDHA 0.069 0.265 clamARA 0.015 0.611 clamC 0.001 0.881 clamALA 0.004 0.785

59 4 Corbicula in Lake Constance

The relationships between seston variables in the first PCA (Fig. 17) and in the second PCA (Fig. 18) were rather similar. However, the positive correlation between the concentrations of DHA and ALA in seston and the abundance of Cryptophyceae was more pronounced in the second PCA, as well as the POC concentration and the total chlA concentration correlation. In contrast, the positive correlation between the abundance of green algae and C:P ratios of seston, between DHA and sterol concentrations and between diatom abundance and seston PON concentrations is more pronounced in the first PCA (Fig. 18).

The linear model analysis revealed a significant effect of water temperature on clam growth rates in the field. This relationship explains 93.7%, i.e. the largest proportion, of the variance of the data (Tab. 6). Additionally, length increase (R² = 0.8804, p < 0.001) and TCI (R² = 0.7186, p < 0.001) were significantly positively correlated with water temperature. For TCI a maximum appears to be at around 14 - 15 ° C (Fig. 16c). Due to strong positive correlations among these three growth parameters (Fig. 18, Tab. 6), the following linear models were calculated only for seston parameters and somatic growth rates.

Linear models showed that ARA concentration, C:P ratio of seston, the abundance of green algae and diatoms, POC concentration and total chlA content also have a significant positive impact on somatic growth rates of C. fluminea (Tab. 6); however, they also correlate positively with temperature (Fig. 18).

Residual variance of the linear model (growth rate versus water temperature) was significantly negatively correlated with seston PUFA concentrations and explained 89.01% of the variance of the residuals. The concentrations of DHA, ALA and total FA in lake seston had also high correlation coefficients resulting in significant negative linear models with residual variance of linear model (Tab. 7). The C:N ratio of seston was positively correlated with the residuals. The abundance of Cryptophyceae and the concentration of Ppart, PON, EPA and total sterols show a smaller but still significant negative correlation with the residuals.

The calculation of two factorial linear models with growth rates and temperature and a further factor resulted in a maximum correlation coefficient with the PUFA concentration as a second independent variable. This 2-factor model explains 99.6% of the variance of the data (Appendix, Tab. S1). A two-factorial analysis of variance (ANOVA) revealed both a significant effect of water temperature and the PUFA concentration and a significant interaction between the two factors (Appendix, Tab. S2). Other highly significant models with residuals of temperature and growth rates are POC content of the seston, total chlA concentration and total FA (Tab. 7).

60 4 Corbicula in Lake Constance

Tab. 7 Linear models of residuals of clam growth vs. mean water temperature and other seston parameters.

R² p slope

PUFA 0.8901 <0.001 -0.0001

DHA 0.8784 <0.001 -0.0002

ALA 0.8603 <0.001 -0.0003

tot.FA 0.7998 <0.001 -0.0001

CN 0.7859 <0.001 0.0013

crypto 0.7632 <0.001 -0.0059

EPA 0.4632 <0.001 -0.0003

tot.ST 0.4424 0.0014 -0.0003

Ppart 0.4123 0.0023 -0.0006

PON 0.3905 0.0032 -0.0517

POC 0.1080 0.1571

green 0.1071 0.1589

diat 0.0979 0.1792

CP 0.0964 0.1828

ARA 0.0074 0.7176

TCI 0.0046 0.7765

cyan 0.0001 0.9633

mean.temp <0.001 1

61 4 Corbicula in Lake Constance

Fig. 19 Biplot of 3. principal component analysis (PCA), with mean values of tissue parameters and growth rate, length increment and tissue condition index of clams per experimental period (see detail Tab. 4). Left and lower axis represent data values (numbers), right and upper axis represents variables (red arrows). Arrows represents loading of variables on first and second principal component PCA1 and PCA2. Numbers are grouped as follows: 1 - 5 experiment one (winter), 6 - 10 experiment two (spring), 11 - 15 experiment three (summer), 16 - 20 experiment four (autumn).

A separate PCA, performed on clam growth rates and clam tissue components (Fig. 19), revealed that 76 % of the total variance of the data can be explained by the first two principal components. The various fatty acid components recorded in clam tissues were positively correlated with each other, except for ARA. Growth rates and the increase in shell length coincided almost completely and the TCI also acts in the same direction. The growth parameters were positively correlated with the elemental stoichiometry (C:N, C:P) of clam tissues and negatively with the N, P, and ARA content of the soft body (Fig. 19).

The calculation of linear models showed a significant positive correlation between growth rates and C:N ratios of clam tissues (R² = 0.298, P = 0.013) and a significantly negative correlation with the N content of the soft body (R² = 0.292, P = 0.014).

62 4 Corbicula in Lake Constance

Laboratory growth experiments

Overall, the mortality of C. fluminea was low in laboratory experiments, except for the second experiment in which 9 out of 10 clams died in the 5 °C treatment. The alga N. limnetica, which was used as food source for C. fluminea, was characterized by a high total fatty acid content, with the

main fraction being represented by

PUFAs, in particular by EPA (134 –

172 µg mg C-1; see Appendix, Tab. S3).

Dry mass growth rates and shell growth rates in laboratory experiments were also positively correlated with the mean water temperature (Fig. 20, Tab. 6), similar as in the field experiments. The TCI was highest between 15 and 18 °C; at higher temperatures the proportion of soft body tissue was reduced

(Fig. 20c), resulting in a reduced R² of the linear model (Tab. 6).

Analysis of covariance (ANCOVA) revealed a significant influence of water temperature on somatic growth rates of clams (F = 612.31, p <0.001), on the increase in shell- length (F = 830.14 , p <0.001) and on clam TCI (F = 69.03, p <0.001) , while the time of the experiment did not influence growth rates (F = 1.70, p =0.168) or the increase in length (F = Fig. 20 Somatic growth rates (calculated via the increase in A 0.753, p =0.522 ). The clam TCI in the dry weight or B shell length) and tissue condition indices (TCI; second experiment in spring was C) of Corbicula fluminea in the four laboratory experiments increased, whereas during the fourth plotted against the water temperature. Circles represent mean experiment in autumn lowest values value of climate chamber replicates (n = 10).

63 4 Corbicula in Lake Constance were detected (F = 21.79, p < 0.001).

A PCA (Fig. 21) and the calculation of linear models (Tab. 6) revealed a significant positive correlation among dry mass growth rates, shell growth, and TCI of clam tissue. Also, the C:P and C:N ratio of clam tissues correlated with clam growth rates. A significant negative correlation was found between growth rates and the phosphorus and nitrogen content of clam tissue as well as DHA concentrations in the soft body of clams. The total FA and ST content, total PUFAs, ARA and EPA had no significant effect on the growth rate of the clams in laboratory experiments (Tab. 6).

Fig. 21 Biplot of 4. principal component analysis (PCA), with mean values of tissue parameters and growth rate, length increment and tissue condition index of clams in the laboratory experiment (see detail Tab. 4). Left and lower axis represents data values (numbers), right and upper axis represents variables (arrows). Arrows represents loading of variables on first and second principal component PCA1 and PCA2. Numbers are grouped as follows: 1 - 5 experiment 1, 6 - 10 experiment 2, 11 - 15 experiment 3, 16 - 20 experiment 4.

64 4 Corbicula in Lake Constance

Discussion

Lake Constance seston Lake Constance, a large warm-monomictic lake, circulates during the winter months. A stable thermal stratification is usually reached between April and October, primarily depending on wind events (Bäuerle et al. 1998). With beginning of thermal stratification water temperatures exceeded 10 °C in late April 2010 in upper water layers and at the beginning of May at 10 m depth. Largest vertical temperature gradients were present during summer months. Autumn temperatures were reduced and upper water layers in Lake Constance were mixed, due to storm wind events typical for Western Europe.

The seasonal variation in abundance and taxonomic composition of phytoplankton can be affected by many factors, such as the availability of nutrients, water temperature, formation of thermal layers, light intensity and grazing intensity of zooplankton (Sommer 1985; Tirok and Gaedke 2007; Rinke et al. 2010). In general, the phytoplankton concentration in Lake Constance during winter is very low (see here, Rinke et al. 2010). The proliferation of algae in spring starts with establishment of a stable thermal layer. Time of the occurrence and height of the spring bloom is largely dependent on the weather in spring (Gaedke 1992). Grazing by filter feeding zooplankton, especially Daphnia, leads to a clear-water phase in June (Lampert 1978). The summer plankton is usually dominated by large grazing resistant phytoplankton species (Sommer 1985). Compared to spring species the summer phytoplankton may be of lower food quality regarding to ingestibility or nutrient content, due to nutrient depleted growth over summer month when P is reduced. The temporal and spatial variation of the composition and distribution of phytoplankton in the field experiments in late 2010 can be explained according to these mechanisms, however, in the study year 2010 neither a strong spring bloom nor a clear-water phase were distinctive. Cryptophyceae and diatoms provide the majority of total biomass of phytoplankton in Lake Constance (Hartwich et al. 2012), whereas the proportion of Chlorophyceae and cyanobacteria decreased within the last decades, due to a decrease of phosphorus concentrations (IGKB 2009). In our field experiments, cyanobacteria represented minor proportion in seston, in contrast to the quantitatively dominant diatoms, which represented the main fraction of the total chlorophyll content. In general, algae growth is located in surface layers in spring, where growing conditions are well represented by the rising temperature and high light availability. Over time, the concentration of available phosphorus, limiting nutrient for algae in most freshwater ecosystems (Elser et al. 2000), gradually decreases and the maximum of the algae concentration is shifted to deeper layers, where phosphorus is still available (Eder et al. 2008; Rinke and Rothhaupt 2008).

Long-term measurements of seston in Lake Constance showed a decrease of the phosphorus as a result of reoligotrophication (Gaedke and Schweizer 1993) and an increase in the seasonal variability and maximum values of seston C:P (Hochstädter 2000). Former studies on Lake Constance showed C:P variations between 100 - 300 even up to 500 in late summer months (Hochstädter 2000; Hartwich et al. 2012), which can be affirmed in the present 65 4 Corbicula in Lake Constance study. In oligotrophic lakes usually high seston C:P ratios are observed (Elser et al. 2000; Sterner et al. 2008), which can lead to insufficient phosphorus supply for consumers (Brett et al. 2000).

The strong correlation of the Chlorophyceae with the C:P of seston (first PCA) confirms the picture of the temporal and vertical distribution of green algae. Chlorophyceae were present only during summer months, particularly in the upper water layers with highest C:P ratios, while the highest abundance of diatoms (higher phosphorus requirement) can be found in deeper water layers with lower C:P. Further evidence of the high nutrient requirements of diatoms and Cryptophyceae can be found in strong correlation of diatoms with Ppart concentration respectively of Cryptophyceae with PON concentration of seston. POC was also correlated with total chlorophyll concentration, total values reached throughout the year are comparable with other studies on Lake Constance (Wacker and Von Elert 2001; Hartwich et al. 2012), although the study by Wacker and Von Elert was conducted in times of higher productivity of Lake Constance.

The seasonal variation of environmental conditions such as light, nutrient availability and temperature can affect levels of PUFAs and sterols in algae significantly (Wright et al. 1980; Müller-Navarra 1995). It has been shown that the relative proportion of PUFA in phytoplankton decreases under phosphorus and nitrogen limitation, while the proportion of monounsaturated fatty acids and lipid-storage substances increases (Harrison et al. 1990; Reitan et al. 1994; Lynn et al. 2000). This could be an explanation for the reduced PUFA content (per mg C) of summer seston, compared to PUFA-rich spring seston. Another indication is the strong negative correlation between the concentrations of PUFAs (ALA, EPA, DHA and total PUFA) and the C:N ratio of the seston and the somewhat weaker correlation with the C:P ratio of seston. Additionally, the amount of bacteria may increase during the year (Hartwich et al. 2012) and these phylum is generally known for low PUFA and sterol levels, therefore, their occurrence might reduce the lipid:carbon ratio, observed during summer months.

Not only the quantity but also the fatty acid composition of seston varied significantly throughout the experimental period. One reason is the different fatty acid composition of each dominant species of algae. For green algae fatty acids with chain lengths of 16 to 18 carbon atoms are typical (Ahlgren et al. 1992). Cryptophyceae and diatoms have a particularly high concentration of long-chain PUFAs such as EPA and DHA (Dunstan et al. 1993; Ahlgren et al. 1997). This was also reflected in the present study, in which a high PUFA content of the seston, especially EPA and DHA, coincided with the dominance of these two groups of algae in spring. In addition, the concentration of ALA, a potential precursor for the n-3 fatty acids EPA and DHA, is enhanced in spring. These three PUFAs correlated in the statistical analyses. The n-6 arachidonic acid (ARA) showed a completely different pattern, with high concentrations during summer months and strong correlation with water temperature. Throughout the year, ARA levels were relatively low and hard to detect. ARA is

66 4 Corbicula in Lake Constance often associated to diatoms (Dunstan et al. 1993), therefore, during times of diatoms dominance in summer months, this fatty acid was measureable.

Clam growth and survival Our study revealed a strong impact of water temperature on growth and survival of C. fluminea in Lake Constance in 2010. For C. fluminea, lethal temperatures of ≤ 2 °C and > 36 - 38 °C have been reported (Britton and Morton 1979; Karatayev et al. 2005). The mortality of C. fluminea in the first field experiment in winter was remarkably high, most likely because the clams were exposed to very low temperatures down to 4 °C throughout the experiment. In Lake Constance, water temperatures of ≤ 3 °C persisting over several weeks have already been found to cause mass mortality of C. fluminea (Werner and Rothhaupt 2008). Similar observations were reported from the USA (French and Schloesser 1991; Morgan et al. 2003). However, in the first concomitant laboratory experiment none of the clams died at temperatures comparable to the prevailing field temperatures. Hence, low temperature alone cannot explain the high mortality observed in the first field experiment in winter. During this winter experiment, the lowest POC and PUFA concentrations of the year were detected in the water column (0.1 mg POC l-1, <50 µg PUFA mg C-1), whereas the clams in the concomitant laboratory tests were always provided with saturating amounts of N. limnetica, which is rich in PUFAs, particularly in EPA and can result in high growth rates of C. fluminea (Basen et al. 2012). As dietary PUFA requirements of aquatic invertebrates may increase with decreasing temperatures (Martin-Creuzburg et al. 2012), we suggest that the high mortality observed in the first field experiment in winter was caused by low temperatures combined with an insufficient supply with PUFA-containing food sources.

In the second laboratory experiment in spring 90% of the clams died within the 4 - 5 °C treatment, most likely because the field-collected animals were exposed to cold temperatures in the laboratory for additional four weeks. In contrast, 70% of the clams survived in the simultaneously conducted field experiment with mean water temperatures of 8 - 9 °C. Mortality was negligible in all other laboratory and field experiments conducted later in the year.

The clams, which survived the winter experiment, were the largest and heaviest at the beginning of the experiment, indicating a size-dependent winter mortality as already described by Werner (2008). A depletion of reserve materials stored in clam soft bodies, as has been shown to occur at low and high water temperatures, probably impairs smaller clams much earlier than larger clams, i.e. smaller clams are more sensitive to low temperatures than larger animals (French and Schloesser 1991; Werner and Rothhaupt 2008; Vohmann et al. 2010).

Temperature did not only affect survival but also the growth of C. fluminea. At a temperature of 10 - 11 °C, growth and reproduction is possible (Kraemer and Galloway 1986; Karatayev et al. 2005), but favourable conditions are between 14 - 22 °C (Foe and Knight 1986). Werner (2008) observed stagnation in growth of C. fluminea in Lake Constance during winter months and a decrease in growth rate with increasing water depth. Here, a positive 67 4 Corbicula in Lake Constance correlation between temperature and growth rates was confirmed in the principal component analysis and by high correlation coefficients of linear models. Below 10 °C barely any increase in length or weight was detected at field and laboratory conditions. Temperatures above 10 °C usually occur in Lake Constance to a maximum depth of 25 m and only over a period of seven months during the year (Rinke et al. 2010), i.e. favourable conditions for C. fluminea in Lake Constance are constricted both spatially and temporally.

Comparing results of field and laboratory experiments showed that above 15 °C average somatic growth rates were higher in the laboratory than in the field. Superior growth conditions for C. fluminea in the laboratory are also indicated by higher TCI values in laboratory clams. Shell and soft body growth in bivalves does not necessarily occur simultaneously (Hilbish 1986). Changes of the index give meaningful information about the physiological status of individuals (Lucas and Beninger 1985; Cain and Luoma 1990; Cataldo and Boltovskoy 1998). Vohmann et al. (2010) found a high plasticity of the body mass of C. fluminea at different food availabilities and concluded an adaptation strategy to longer starvation periods. A decrease of the soft body mass with simultaneous positive shell growth suggests an independent control of shell and body mass growth. In the study by Vohmann et al. (2010), the observed decrease of the soft body mass of C. fluminea at high temperatures was also attributed to the simultaneous low food availability. Foe and Knight (1985) speculate about a decoupling of shell and soft body growth caused by an investment in reproductive organs. Reproduction of C. fluminea starts with a shell length of approximately 10 mm (Aldridge and McMahon 1978). To minimize the effect of reproduction on somatic growth, juvenile clams were used in our study. However, we cannot completely exclude that the determination of somatic growth rates and TCI values were to some extent affected by the investment in reproduction, in particular during the summer and autumn experiment.

As the experimental date did not significantly affect growth rates (tissue and shell) in the laboratory tests and as the TCIs of the clams at the beginning of the experiments were not statistically different, it can be assumed that all animals were in a similar physiological state at the start of the experiments. This is somehow surprising, as one would expect different preconditions for clams collected in the field at different times of the year. However, under optimal conditions with high quality food supply these effects merely affect growth rates in clams.

Under laboratory growth conditions clams were fed with saturating amounts of algae. In the field experiments, food supply was defined by plankton succession. For C. fluminea, food quantity limitation was observed at chlorophyll a concentrations of < 20 µg l-1 at 15 °C water temperature (Foe and Knight 1985; Mouthon 2001). Additionally, with increasing temperatures a higher carbon supply is needed. During our field experiments, chlorophyll a concentrations did not exceed 10 µg l-1 (> 10 °C). However, potential food limitation at high temperatures in the lake were not lethal as shown for the River Rhine, where a negative energy balance, due to increased basal metabolic rate at warmer temperatures, resulted in reduced growth rates and concurrent increased mortality (Weitere et al. 2009; Vohmann et

68 4 Corbicula in Lake Constance al. 2010). An adaptation mechanism to summer heat stress was found in shell closure and therefore metabolism regulation for C. fluminea (Ortmann and Grieshaber 2003). However, in recent laboratory growth experiments the reduced activity resulting in shell closure was only observed at low temperatures. It was also recorded during winter times when temperature falls below 5 °C over a longer period (Ortmann and Grieshaber 2003).

Bivalves like C. fluminea are highly reliant on seston food availability. Thus, seasonal changes in seston carbon concentrations and in the phytoplankton species composition may impair the performance of filter feeders. C. fluminea favours particles of 1 - 20 µm but is able to ingest particles up to 170 µm (Way et al. 1990; Boltovskoy et al. 1995; McMahon and Bogan 2001) and can effectively remove detritus, bacteria and algae from the water column (Boltovskoy et al. 1995). It is assumed that C. fluminea is a non-selective filter feeder (Way et al. 1990; Boltovskoy et al. 1995), and therefore can probably not select food particles with regard to their quality. So far, only few studies have been conducted to determine the effects of food quality of phytoplankton species on freshwater bivalves (Foe and Knight 1986; Wacker and Von Elert 2003; Basen et al. 2011). Especially the absence of essential lipids, i.e. PUFAs and sterols, has been suggested to impair the growth of filter feeding bivalves (Soudant et al. 1996a; Wacker et al. 2002; Wacker and Von Elert 2002; 2003; 2004; Basen et al. 2012). The PUFA- and sterol-rich alga N. limnetica, which was used as food source in our laboratory experiments, was previously found to be a good quality food for C. fluminea (Basen et al. 2012). Compared to clams exposed in the field, clams in the laboratory experiments were provided with non-limiting concentrations of a high quality food, which explains why the clams in the laboratory experiments had higher somatic growth rates and higher proportions of soft body tissue than the clams in the field experiments.

Food quality effects become more important when food quantity is high (Sterner 1997). However, it has been shown that filter feeding cladocerans of the genus Daphnia can be limited by phosphorus even at low food quantity (Boersma and Kreutzer 2002). For Lake Constance recent studies revealed that food quantity and quality with regard to PUFAs affect filter feeding cladocerens (Wacker and Von Elert 2001; Hartwich et al. 2012). Simultaneous limitation by multiple nutrients (i.e. co-limitation) is a phenomenon which only recently has been considered in food quality research with two limiting essential biochemicals determining fitness in filter feeding cladocerans (Martin-Creuzburg et al. 2009; Martin- Creuzburg et al. 2010; Sperfeld et al. 2012). With fast growing organisms like filter feeding cladocerans and shorter experimental time spans indications for co-limitation for feeding on seston of Lake Constance were found with altering POC, ALA and Ppart limitations throughout the year (Hartwich et al. 2012).

In our study we investigated potential food quality effects of lake seston on C. fluminea growth performance throughout the year. Seston ARA concentration correlates with clam growth rates, however also strongly correlated to lake temperature. Temperature is the dominant driver in plankton succession in Lake Constance and many factors depend on water temperature. The high correlation between water temperature and POC, PON,

69 4 Corbicula in Lake Constance diatoms, green algae and chlA content in the seston limits the model comparison as well. Significant linear models for other factors like POC might be biased by co-linearity, which we cannot exclude. Apart from temperature alone, explaining 93% of the variance, only minor improvement of statistical models was reached when adding a second factor (PUFA, ARA content, C:P ratio, green algae, chlA, diatoms, total sterol or total fatty acid) to a two- factorial model on clam growth rates correlated to temperature. Adding temperature and seston PUFA concentrations as factors in the linear model resulted in a higher R² value, i.e. growth rates of clams in the field were affected by both factors simultaneously. However, the linear model of growth rate with temperature underestimates the actual growth rates in periods of high seston PUFA concentrations (negative residuals, spring), but overestimates growth rates at lower seston PUFA concentrations (positive residuals, winter; Appendix Fig. S2). Probably, a temperature optimum curve would describe the entire data set even better; however, the simplified linear model used confirms the strong influence of temperature on growth rates of C. fluminea in the field.

The elemental nutrient composition (phosphorus, nitrogen) of phytoplankton is highly variable, while the elemental stoichiometry of consumers is less variable (Sterner and Elser 2002; Liess and Hillebrand 2005). This can lead to a decoupling of carbon transport in the food chain from primary producers to herbivorous consumers. The elemental composition of macroinvertebrates changes seasonally because of changes in the nutrient composition of their diet (Liess and Hillebrand 2005; Naddafi et al. 2009). The invasive D. polymorpha seems to be adapted to avoid short term elemental nutrient limitation and therefore fitness is probably less affected by stoichiometric constraints (Naddafi et al. 2009), which could be important for the establishment in new ecosystems. Our results suggest that the elemental composition of C. fluminea is subject to similar mechanisms. The growth rates of clams correlate positively with the C:P and C:N ratio of the tissue (i.e. negative with P and N values). This suggests that C. fluminea has no strict homeostasis (Evans-White et al. 2005; Atkinson et al. 2010). Naddafi & Petterson (2008) showed a significant seasonal variation of tissue C:P and C:N ratios of D. polymorpha in meso-eutrophic Lake Erken (Sweden), at which the C:P ratio was greatest during summer months. Similar observations were also made in the present work. Besides, laboratory experiments suggest that changing nutrient composition of C. fluminea may be independent of food stoichiometry. Possible limitation in P supply for clams in Lake Constance might occur during short periods in summer. P-rich sources in seston are bacteria, which may also serve as food source for clams. Those particles were not detectable in our setup, but achieve up to 50% in Lake Constance (Hartwich et al. 2012) and might help clams to overpass times of P depression. Thus, we conclude that C. fluminea is less susceptible to P-limitation. This could be another factor influencing successful spread of this species in freshwater ecosystems.

In Lake Constance, the cyanobacterial biomass steadily decreased in the last decades as a consequence of the re-oligotrophication process (Gaedke and Schweizer 1993; Sommer et al. 1993). The role of sterols as limiting nutrient in freshwater ecosystems dominated by eukaryotic phytoplankton is of minor importance. However, when the seston is dominated

70 4 Corbicula in Lake Constance by prokaryotic food sources, e.g. during cyanobacterial blooms, the importance of sterols for pelagic and benthic food webs might increase (Martin-Creuzburg and Von Elert 2009). Cyanobacterial blooms may negatively affect consumers due to a variety of factors such as toxicity, grazing or digestion resistance, and a deficiency in essential biochemical nutrients, i.e. sterols and long-chain PUFAs (DeMott et al. 1991; Martin-Creuzburg and Von Elert 2009; Van Donk et al. 2011). Experiments with C. fluminea showed that somatic growth of clams feeding on pure cyanobacterial diets is constrained by the absence of dietary sterols (Basen et al. 2011; Basen et al. 2012). For filter feeding zooplankton it has been proposed that at least 20% of eukaryotic carbon are required to satisfy sterol requirements (Martin-Creuzburg et al. 2005b). Thus, sterol limitation might occur in the field only in times of massive cyanobacterial bloom formation, which occurs more often in eutrophic ecosystems. In addition, cyanobacteria and possible mediated food quality effects on consumer level gain importance when bloom formation will be favoured by climate change (Jöhnk et al. 2008; Wagner and Adrian 2009).

The expected increase in global temperature may also alter the geographical distributions of invasive species, which are recognised as a main cause of the omnipresent decline of freshwater biodiversity (Sala et al. 2000; Walther et al. 2009). Stachowicz et al. (2002) postulated that the greatest effects of climate change on biotic communities are mediated by changes in maximum and minimum temperatures. Changes in plankton succession and species composition, lake water exchange and vertical mixing, ice cover in winter and nutrient recycling are factors predicted to be effected by changing temperature regimes. There are indications that thermal stratification in Lake Constance in spring begins earlier and ends later in the year (Straile et al. 2003), which may extend the growing season for C. fluminea in Lake Constance. Moreover, milder winter temperatures (Straile et al. 2010) may reduce winter mortality (Werner and Rothhaupt 2008) and may improve growth conditions for clams and increase the reproductive success (Mouthon 2001; Sousa et al. 2008) of C. fluminea in Lake Constance. Similar observations were made at thermally impaired aquatic systems, when clams inhabited areas of power plant cooling water outflows (Cairns and Cherry 1983; French and Schloesser 1991).

Perspectives A challenge of invasion biology is to understand mechanisms that determine the success of invasive species. We conclude from our data that the growth of C. fluminea in Lake Constance is mainly controlled by water temperature. Since growth and development of C. fluminea occur above a temperature of about 10 °C, only seven months of the year are available for growth and reproduction, and this is limited to a maximum water depth of 25 m. The relationship between growth rates and water temperature could be confirmed in laboratory experiments. However, seasonality in seston characteristics, e.g. short term depletion of food quantity and quality, might also impair the growth of clams. For C. fluminea clear indications for food quality could not be distinguished apart from

71 4 Corbicula in Lake Constance temperature effects, which impair seston development throughout the year. Seston parameters vary over 7 weeks, therefore clams integrate conditions over experimental duration, and only severe changes in seston succession might affect data processing. A higher temporal resolution of the clam growth in Lake Constance would certainly be helpful to understand the relationship between the changing environmental conditions and performance of C. fluminea to point out times of possible food quantity and quality limitation for clams.

Acknowledgements We thank S. Oexle and H. Recknagel for the support with the experiments and R. Basen for helpful comments on how to improve the quality of this manuscript. This work was funded by the DFG (German Research Foundation) within the collaborative research centre CRC 454 “Littoral of Lake Constance”.

Supplementary data

Appendix Tab. S1 Comparison of linear models with one independent factor or with 2 factors including temperature correlated with clam tissue growth rates

Appendix Tab. S2 Two-factorial analyses of variances (ANOVA) with different seston parameters on growth rate of Corbicula fluminea

Appendix Tab. S3 Composition of N. limnetica used as sole food source for C. fluminea in the laboratory experiments

Appendix Fig. S1 Seasonal changes in fatty acid concentrations in seston of Lake Constance sampled at the study site in 2010. Shown are the α-linolenic acid (ALA), arachidonic acid (ARA), docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) (all in µg mg C-1 and in µg l-1). The values between sampling dates and between sampled water depths were interpolated linearly. The four experimental periods are accentuated.

Appendix Fig. S2 Linear model of growth rate of Corbicula fluminea in correlation to lake water temperature and residuals of linear model plotted against PUFA content of Lake Constance seston in 2010

72 5 Clam mediated benthic pelagic coupling

Chapter 5

Phytoplankton food quality effects on gammarids: benthic-pelagic coupling mediated by an invasive freshwater clam

Timo Basen, Rene Gergs, Karl-Otto Rothhaupt and Dominik Martin-Creuzburg

submitted

Abstract Benthic-pelagic coupling mediated by bivalves has been shown to increase the flow of energy towards the benthos. Here, to assess the capability of clams to process and therewith modify the quality of pelagic food sources for subsequent use by benthic invertebrates, a growth experiment was conducted in which juvenile Gammarus roeselii were raised either directly on sedimented pelagic autotrophs (algae, cyanobacteria) or on the same autotrophs biodeposited by the invasive freshwater clam Corbicula fluminea either as feces or pseudofeces. We show that growth and survival of G. roeselii are significantly improved when autotrophs are offered as biodeposition material, i.e. after processing by the clams, and conclude that this clam-mediated upgrading of food quality is due to both an increased bioavailability of pelagic food particles, which are packed in mucus during clam processing, and an increased dietary provisioning with essential lipids originating from the clams. Hence, filter-feeding bivalves provide a crucial link between the pelagic and benthic food web not only by deflecting energy fluxes, but also by processing and upgrading pelagic food for benthic invertebrates.

73 5 Clam mediated benthic pelagic coupling

Introduction Geographical spread of invasive species is recognized as a main cause of the omnipresent decline of freshwater biodiversity (Sala et al. 2000). In the last century the Asian clam Corbicula fluminea has become an ubiquitous benthic invertebrate in freshwater ecosystems worldwide (Araujo et al. 1993; Darrigran 2002; Lee et al. 2005). Filter-feeding bivalves can considerably increase the pelagic–benthic coupling, i.e. the flow of pelagic organic matter to the benthos, thereby stimulating benthic productivity (Strayer et al. 1999; Sousa et al. 2008; Gergs et al. 2009). It has been shown that the occurrence of Corbicula populations lead to an increase in sediment organic matter concentrations (Hakenkamp and Palmer 1999). The biodeposited material mainly consists of digested (feces) and undigested, rejected (pseudofeces ´PSF´) seston particles. At low particle concentrations, the deposited material consists mainly of feces (MacIsaac and Rocha 1995; Roditi et al. 1997), whereas with rising food concentrations (above 0.2 mg C l-1 for Dreissena polymorpha) the proportion of PSF is increasing (Walz 1978; Gergs et al. 2009). For many benthic invertebrates these bivalve- generated food packages represent a suitable food source (Karatayev et al. 1997; Roditi et al. 1997). It has been shown that biodeposition material of bivalves supports gammarid nutrition both in the laboratory and in the field (Gergs and Rothhaupt 2008; Gergs et al. 2011).

Besides bivalve driven deposition of organic matter, sedimentation of phytoplankton per se, especially under bloom conditions, provides a huge pelagic carbon pool for the benthic food web (Nascimento et al. 2008; Suikkanen et al. 2010). It has been shown that benthic invertebrates can ingest and assimilate sedimented cyanobacteria, but the nutritional value seems to be rather low (Karlson et al. 2008; Nascimento et al. 2009). As the frequency of cyanobacterial bloom formation is expected to increase with global warming (Paerl and Huisman 2009), it is important to investigate the consequences of cyanobacterial mass developments for ecosystem processes, e.g. the role of cyanobacterial carbon within the food web and its food quality for pelagic and benthic consumers. In general, cyanobacteria represent a nutritionally inadequate food source for aquatic consumers, which can be due to morphological properties (Van Donk et al. 2011), toxin production (DeMott et al. 1991), and/or to a deficiency in essential biochemical nutrients (Martin-Creuzburg et al. 2008; Martin-Creuzburg et al. 2009). In particular, the lack of sterols has been suggested to constrain the carbon transfer efficiency from cyanobacteria to herbivorous zooplankton and benthic filter feeders (Martin-Creuzburg et al. 2008; Basen et al. 2011; Basen et al. 2012).

Although several studies have shown that bivalves are significantly involved in transferring pelagic organic matter (i.e. mainly phytoplankton) to the benthic food web (e.g. Gergs et al. 2009), the quality of this biodeposited material as food for benthic invertebrates has been poorly studied. In the present study, we investigated whether biodeposition materials produced by C. fluminea feeding on different pelagic food sources differ in their food quality for Gammarus roeselii. We hypothesized that the biodeposition activity does not only provide increased access to pelagic food sources, but also affects the nutritional quality of

74 5 Clam mediated benthic pelagic coupling phytoplankton as food for benthic invertebrates by modifying the dietary elemental and biochemical composition. In laboratory experiments, two different concentrations of algae (Nannochloropsis limnetica, Scenedesmus obliquus) and cyanobacteria (Anabaena variabilis, Synechococcus elongatus) were fed to adult C. fluminea to obtain biodeposition material consisting either of mostly digested (feces) or undigested (PSF) material. We investigated the survival and growth of G. roeselii feeding on collected biodeposition materials and on the autotrophic food sources without clam conditioning, and related the results to the elemental (C, N, P) and biochemical (fatty acid, sterol) composition of the different food sources to assess the role of clam filtration and digestion in determining the nutritional quality of phytoplankton-derived food for benthic invertebrates.

Materials and methods

Sampling and cultivation of animals Gammarids (G. roeselii) and clams (C. fluminea) were obtained from the littoral zone of the oligotrophic prealpine Lake Constance. Adult G. roeselii were collected via kick sampling at the shoreline and clams were collected at a water depth of 2 – 3 m by scuba-diving. Until the start of the experiments, both species were kept separately in climate chambers with a diurnal dark-light cycle of 12 h :12 h G. roeselii were kept at 15°C in aquaria containing lake water, gravel of different grain sizes for shelter, and dried alder leaves as a food source; C. fluminea were kept at 20°C in a ﬂow-through system with seston-containing lake water (<30 µm) and washed sediment.

Food preparation Autotrophic food sources were cultivated semi-continuously in aerated 5 l vessels at a dilution rate of 0.20 d–1 at 20°C with illumination at 100 - 120 μmol quanta m–2 s–1 and harvested in the late-exponential growth phase. The coccoid cyanobacterium Synechococcus elongatus (SAG 89.70, Sammlung für Algenkulturen Göttingen, Germany), the filamentous cyanobacterium Anabaena variabilis (ATCC 29413, American Type Culture Collection, Manassas, USA), the green alga Scenedesmus obliquus (SAG 276-3a), and the eustigmatophyte Nannochloropsis limnetica (SAG 18.99) were each grown in Cyano medium (Jüttner et al. 1983). These food organisms were used because they differ considerably in size, shape, lipid content and composition. Food suspensions were prepared by concentrating the cells via centrifugation (3000 g, 10 min) and resuspension in fresh medium. Carbon concentrations of the food suspensions were estimated from photometric light extinctions (800 nm) and from carbon-extinction equations determined prior to the experiment.

To obtain biodeposition material produced by clams, C. fluminea were kept in ﬂow-through systems with filtered, aerated lake water (<1 µm, 200 mL min-1, 20° C) enriched with well-

75 5 Clam mediated benthic pelagic coupling defined amounts of food, i.e. cyanobacteria or algae. The flow-through systems consisted of experimental basins with a size of 34 × 40 × 7.5 cm (width × depth × height); the water level was adjusted to 6 cm, resulting in a water volume of approximately 8 l. To minimize sedimentation of the algae, the water in the basins was gently aerated. All flow-through systems were equipped with small plastic boxes (8 × 8 × 5 cm; n = 5 per basin) containing approximately 12 g of clam biomass (shell length 10 - 20 mm). The autotroph food suspensions were added continuously using a peristaltic pump. Two different food concentrations were adjusted to produce two different kinds of biodeposition material: A low food concentration (0.2 mg C l-1) was used to gain mostly digested autotrophic carbon (henceforth referred to as ‘feces’) and a high food concentration (1 mg C l-1) was used to increase the fraction of undigested autotrophic carbon (henceforth referred to as ‘pseudofeces’). Clam conditioned ‘feces’ and ‘pseudofeces’ differed in color and thus could be easily distinguished. The organic matter biodeposited by C. fluminea in the plastic boxes was collected three times a week, i.e. the plastic boxes were replaced by clean boxes and the biodeposited material was rinsed from the boxes using a pipette, centrifuged, resuspended in filtered lake water and adjusted to a certain optical density to ensure constant feeding conditions for G. roeselii. Preliminary experiments revealed that food suspensions prepared according to this procedure are highly suitable to provide G. roeselii with well-defined food concentrations.

Analyses of food sources The elemental (C, N, P) and biochemical (fatty acids, sterols) composition of the food sources were determined weekly from aliquots of the food suspensions. Aliquots were filtered onto precombusted glass-fiber filters (Whatman GF/F, 25 mm diameter) and analyzed for particulate organic carbon (POC) and nitrogen using an NCS-2500 analyzer (ThermoQuest). For the determination of particulate phosphorus, aliquots were collected on acid-rinsed polysulfone filters (HT-200; Pall) and digested with a solution of 10 % potassium peroxodisulfate and 1.5 % sodium hydroxide for 60 min at 121°C, and soluble reactive phosphorus was determined using the molybdate-ascorbic acid method (Greenberg et al. 1985). Values are expressed as molar carbon to nitrogen (C:N) and molar carbon to phosphorus ratios (C:P).

Lipids were extracted two times from pre-combusted GF/F filters (Whatman, 25 mm diameter) loaded with approximately 0.5 mg (for fatty acid analysis) or 1 mg (for sterol analysis) POC of the food suspensions using a mixture of dichloromethane/methanol (2:1, v/v). For the analysis of sterols, the pooled cell-free extracts were dried under a stream of pure gaseous nitrogen and saponified with 0.2 mol l-1 methanolic KOH (70°C, 1 h). Subsequently, sterols were partitioned into iso-hexane:diethyl ether (9:1, v/v), again dried under a stream of nitrogen, and resuspended in a volume of 20 µl iso-hexane. For the analysis of fatty acids, the cell-free extracts were dried under a stream of nitrogen and esterified with 3 mol l-1 methanolic HCl (60°C, 20 min). Subsequently, fatty acid methyl esters

76 5 Clam mediated benthic pelagic coupling

(FAMEs) were partitioned into iso-hexane, dried under a stream of nitrogen, and resuspended in a volume of 50 µl iso-hexane. Lipids were analyzed by gas chromatography on an HP 6890 GC (Agilent Technologies) equipped with a flame ionization detector and either a DB-225 (J&W Scientific) capillary column to analyze FAMEs or an HP-5 (Agilent Technologies) capillary column to analyze sterols. Details of GC configurations are given elsewhere (Martin-Creuzburg et al. 2009; Martin-Creuzburg et al. 2010). Lipids were quantified by comparison to internal standards (C17:0 and C23:0 methyl esters; 5α- cholestan). The detection limit was approximately 20 ng of sterol/fatty acid. Lipids were identified by their retention times and their mass spectra, which were recorded with a gas chromatograph/mass spectrometer (Finnigan MAT GCQ) equipped with a fused silica capillary column (DB-225MS, J&W Scientific for FAMEs; DB-5MS, Agilent for sterols). Sterols were analyzed in their free form and as their trimethylsilyl derivatives. Mass spectra were recorded between 50 and 600 amu in the EI ionization mode, and lipids were identified by comparison with mass spectra of reference substances purchased from Sigma or Steraloids and/or mass spectra found in a self-generated spectra library or in the literature (e.g. Goad and Akihisa 1997). The absolute amount of each lipid was related to POC.

Growth experiments with G. roeselii Gammarids were maintained in tanks filled with water from Lake Constance, containing gravels of different grain sizes for shelter. Gammarids were fed on dried alder leaves. For the growth experiments, juvenile gammarids (2 – 3 mm body length) hatched in these tanks were used. Gammarid growth experiments were conducted from June to October 2009 in glass beakers filled with 100 ml of filtrated lake water (<1 µm); a small stone (organic matter removed using a muffle furnace) was provided as shelter in each beaker. Juvenile G. roeselii were randomly transferred to the experimental beakers. The experiment comprised the following food treatments: the two cyanobacteria S. elongatus and A. variabilis and the eukaryotic algae S. obliquus and N. limnetica without passage through the clam and the same food organisms as biodeposited material produced by feeding the clam with either high or low concentrations of these food organisms (i.e. ‘feces’ or ‘pseudofeces’). The food suspensions were prepared and renewed daily. Gammarids were fed ad libitum (Gergs and Rothhaupt 2008) or starved without adding food (starvation as a control treatment). Each food treatment was replicated 20 times; each replicate consisted of one individual. Three times a week gammarids were transferred into new beakers to avoid the accumulation of food, fecal pellets and the formation of biofilm. Body lengths of the gammarids were measured once a week as described in Gergs and Rothhaupt (2008), survival was recorded three times a week.

Statistical analyses All results were statistically analyzed using the statistical software package R (R Development Core Team 2006). The survival of gammarids in the growth experiments was 77 5 Clam mediated benthic pelagic coupling analyzed in dependence of the offered food species (A. variabilis, N limnetica, S. elongatus and S. obliquus) and the type of the food treatment (i.e. autotrophs, feces and pseudofeces ‘PSF’). We used the parametric survival model ‘survreg’ fitted to an exponential data distribution (α = 0.05). Comparisons between treatments were done using General linear hypotheses and multiple comparisons with Tukey’s post-hoc test for parametric models. For the two-way analyses, model simplification was performed stepwise. Model comparison was done using ANOVA with Chi² test; significant differences from the full model indicate a loss of information through model simplification. Differences in body length at the end of the growth experiment were analyzed using ANOVAs followed by Tukey’s post-hoc tests which were performed separately for the different food species and food treatments. A t-test was performed in those treatments where only two gammarids survived. Homogeneity of variances was tested using Bartlett’s test. Differences between food species and food treatments in total sterol and total PUFA concentration (in µg mg C-1), and molar C:N and C:P ratios, respectively, were analyzed using ANOVAs followed by Tukey’s post-hoc tests separately for the different food species and food treatments. Data were square root (sterols, C:N, C:P) and ln(x + 2) (PUFAs) transformed to obtain homogeneity of variances (Levene’s test).

Tab. 8 Statistical analysis of molar carbon to nitrogen (C:N) and carbon to phosporus (C:P) ratios and of total PUFA and sterol contents of the different food sources used to raise Gammarus roeselii (part a: Two-way analysis; part b and c: One-way analyses of variance (ANOVA) and Tukey’s HSD post hoc test) . species treatment effect df C:N C:P PUFAs sterols

a all all species 3 <0.001* <0.001* <0.001* <0.001*

treatment 2 <0.001* 0.46 0.03* 0.03*

species*treatment 6 0.02* 0.03* 0.26 0.03*

b S. elongatus all treatment 2 0.004* 0.08 - <0.001*

A. variabilis 2 <0.001* 0.41 0.07 0.10

S. obliquus 2 0.06 0.09 0.19 0.21

N. limnetica 2 <0.001* 0.22 0.004* 0.97

c all autotrophs species 3 <0.001* 0.14 <0.001* <0.001*

feces 3 <0.001* 0.002* 0.004* 0.005*

PSF 3 0.01* 0.003* <0.001* 0.27

78 5 Clam mediated benthic pelagic coupling

Results

Analyses of food sources Food stoichiometry - In all 3 food treatments the cyanobacteria had lower molar C:N ratios compared to eukaryotic algae (Tab. 8). Nitrogen levels were reduced in feces produced on cyanobacterial diets, leading to a significant increase in C:N ratios (S. elongatus: P = 0.004, A. variabilis: P < 0.001). C:N ratios of the alga N. limnetica were significantly lower than C:N ratios of biodeposition materials produced by C. fluminea feeding on N. limnetica (Tab. 8, Fig. 22). In contrast, C:N ratios of the green alga S. obliquus did not differ significantly from C:N ratios of S. obliquus–based biodeposition materials (P = 0.06). The molar C:P ratios showed no significant differences between food treatments in any of the used cyanobacteria or algae. However, C:P ratios of PSF and feces produced on a S. elongatus diet were lower than C:P ratios of PSF and feces produced on the other autotrophs (Tab. 8).

Fig. 22 Molar carbon to nitrogen (C:N) and carbon to phosphorus (C:P) ratios of the different food sources used to raise Gammarus roeselii. The two cyanobacteria (S. elongatus and A. variabilis) and the two eukaryotic algae (S. obliquus and N. limnetica) were either fed directly to G. roeselii (white bars) or after passage through the clam C. fluminea, i.e. either as ‘feces’ (coarse bar) or as ‘pseudofeces’ (PSF, gray bars). The different biodeposition materials were produced by feeding C. fluminea with low (‘feces’) and high (‘PSF’) autotroph concentrations. Data are shown as means and standard deviations; numbers at the bottom of the bars indicate sample sizes. Statistical analyses were performed separate for C:N and C:P ratios in all four food species. Bars labeled with the same letters are not significantly different.

79 5 Clam mediated benthic pelagic coupling

Fatty acid composition - In the cyanobacterium S. elongatus only saturated (14:0, 16:0) and monounsaturated fatty acids (16:1, 18:1) were detected. In contrast, small amounts of the polyunsaturated fatty acids (PUFAs) 20:4n-6 (arachidonic acid, ARA) and 20:5n-3 (eicosapentaenoic acid, EPA) were detected in feces of C. fluminea while feeding on S. elongatus (Appendix, Tab. S4). The filamentous cyanobacterium A. variabilis contained two PUFAs, 18:2n-6 and 18:3n-3 (α-linolenic acid, ALA), which were detected also in feces and PSF produced by C. fluminea while feeding on this cyanobacterium. In the green alga S. obliquus, four C-18 PUFAs were detected (18:2n-6, 18:3n-6, ALA, 18:4n-3) but no PUFA with more than 18 carbon atoms. N. limnetica contained the highest amounts of PUFAs; the principal fatty acid was EPA. Besides EPA, also ALA, ARA, and 18:4n-3 were present in all N. limnetica-based food treatments; traces of 22:6n-3 (docosahexaenoic acid, DHA) were additionally detected in PSF.

Fig. 23 Total polyunsaturated fatty acids (PUFA) and total sterol levels (µg mg C-1) of the different food sources used to raise Gammarus roeselii. The two cyanobacteria (S. elongatus and A. variabilis) and the two eukaryotic algae (S. obliquus and N. limnetica) were either fed directly to G. roeselii (white bars) or after passage through the clam C. fluminea, i.e. either as ‘feces’ (coarse bar) or as ‘pseudofeces’ (PSF, gray bars). The different biodeposition materials were produced by feeding C. fluminea with low (‘feces’) and high (‘PSF’) autotroph concentrations. Data are shown as means and standard deviations; numbers at the bottom of the bars indicate sample sizes. Statistical analyses were performed separate for total PUFAs and total sterol levels in all four food species. Bars labeled with the same letters are not significantly different.

80 5 Clam mediated benthic pelagic coupling

In all biodeposition materials (feces and PSF), the composition of fatty acids did not differ from that of the respective autotrophs without passage through the clam. However, total PUFA levels were reduced in biodeposited material compared to cyanobacteria and algae fed directly to gammarids; significant only for N. limnetica-based diets (N. limnetica: P = 0.003; Tab. 8, Fig. 23). Among all autotroph species, the highest PUFA concentration was found in N. limnetica. The total fatty acid composition and the concentrations of single fatty acids in the different food sources are presented in table S4. Without food addition, water samples taken in the flow through systems contained on average 0.2 ± 0.0 mg C l-1, providing small amounts of fatty acids (total fatty acids: 22.8 ± 6.9 µg l-1, respectively 52.1 ± 15.9 µg mg C-1 for the low food concentration (i.e. feces) and 18.4 ± 5.6 µg mg C-1 for the high food concentration (i.e. PSF)). The small amounts of fatty acids detected in the water samples divide into 16:0, 16:1, ALA and EPA.

Sterol composition - Sterols were not detected in the two cyanobacteria A. variabilis and S. elongatus). However, sterols were detected in biodeposition material produced by C. fluminea while feeding on the cyanobacteria, consisting of cholesterol (cholest-5-en-3β-ol), sitosterol (stigmast-5-en-3-ol) and stigmasterol ((22E)-stigmasta-5,22-dien-3β-ol) (Fig. 22; Appendix Tab. S1). The green alga S. obliquus contained fungisterol (5α-ergost-7-en-3β-ol), chondrillasterol ((22E)-5α-poriferasta-7,22-dien-3-ol) and schottenol (5α-stigmast-7-en-3β- ol). Principal sterols in N. limnetica were cholesterol, sitosterol and isofucosterol ((24Z)- stigmasta-5,24(28)-dien-3-ol). The sterols detected in biodeposition materials did not differ from those detected in the corresponding algal food sources, neither in composition nor in quantity (S. obliquus P = 0.21, N. limnetica P = 0.97; Fig. 23; Tab. 9; Appendix Tab. S4). In water samples taken in the flow through systems small amounts of sterols were also detected (total sterols: 0.6 µg l-1, respectively 2.5 ± 0.6 µg mg C-1 for feces, 0.9 ± 0.2 µg mg C- 1 for PSF), with cholesterol (1.4 ± 0.3 µg mg C-1 for feces, 0.5 ± 0.1 µg mg C-1 for PSF) as major component.

Gammarid growth and survival experiment Without food addition, all individuals died within one week. Thus, all survival and growth effects can be attributed to the offered food sources. Survival of gammarids on the different food sources was affected by both the food treatment (i.e. feces, PSF, and without clam conditioning) and the autotrophic species as well as by the interactions of these factors (P < 0.01, Tab. 9). Model simplification resulted in a significant loss of information (P < 0.01) (Tab. 9) and thus the full model without simplification was used. With regard to the S. elongatus- based diets, survival of G. roeselii was higher when fed digested S. elongatus (i.e. feces) than when fed S. elongatus without passage through the clam or when fed S. elongatus biodeposited as PSF (Tab. 9, Fig. 24; P < 0.001). Survival rates of G. roeselii did not differ significantly among the A. variabilis-based diets (P = 0.52). With both the S. obliquus- and the N. limnetica-based diets, survival of G. roeselii was lowest when algae were provided 81 5 Clam mediated benthic pelagic coupling without passage through the clam, intermediate when packed as PSF and highest when packed as feces (Tab. 9, Fig. 24). Interspecific comparison among food sources revealed no significant differences between survival rates of gammarids, neither for autotrophs nor for feces (Tab. 9). PSF produced by C. fluminea while feeding on S. elongatus resulted in lower survival of G. roeselii than PSF based on eukaryotic algal diets. PSF based on A. variabilis did not differ from all other PSF diets.

Within the 7 weeks lasting experiment, the body length of G. roeselii increased in all food treatments. Differences in body length among food treatments were noticeable not before 3 weeks of growth (Fig. 24). The starved individuals in the control treatment died within the first week of the experiment; body growth of these animals was not evaluated. Overall, the increase in body length was less pronounced on S. elongatus-based diets than on diets based on the eukaryotic algae (P < 0.001; Tab. 10, Fig. 24).

Fig. 24 Growth and survival of Gammarus roeselii fed different cyanobacteria (S. elongatus, A. variabilis) and eukaryotic algae (S. obliquus, N. limnetica) either directly (white circles) or after passage through the clam C. fluminea, i.e. either as ‘feces’ (triangles) or as ‘pseudofeces’ (PSF, squares). The different biodeposition materials were produced by feeding C. fluminea with low (‘feces’) and high (‘PSF’) autotroph concentrations. Differences between food treatments in final length or surviving gammarids were analyzed separately for each algae species. Different letters indicate significant differences between treatments (P < 0.05).

82 5 Clam mediated benthic pelagic coupling

Tab. 9 Survival of Gammarus roeselii analyzed using a parametric survival regression model (survreg) and multiple comparisons followed by Tukey’s test for parametric models. Tested models were compared with the full model by ANOVA using the Chi² test. The most simple model not significantly different from the full model, indicating no loss of information through model simplification, was selected. Different letters indicate significant differences between treatments (Tukey’s HSD, P < 0.05).

Chi2 df p-value residual diff.

species*treatment 75.16 11 <0.001 1065.7

species+treatment 59.04 5 <0.001 1081.8*

species 5.37 3 0.15 1135.5*

treatment 52.95 2 <0.01 1087.9*

Chi2 df p-value autotrophs feces PSF

S. elongatus 34.62 2 <0.001* A B A

A. variabilis 1.29 2 0.52 A A A

S. obliquus 23.10 2 <0.001* A B B

N. limnetica 10.79 2 0.005* A B AB

Chi2 df p-value S. elongatus A. variabilis S. obliquus N. limnetica

autotrophs 4.55 3 0.21 A A A A

feces 7.83 3 0.05 A A A A

PSF 9.83 3 0.02* A AB B B

Furthermore, feces and PSF, irrespective of whether they were produced on cyanobacterial or on algal diets, resulted in higher growth rates of G. roeselii than pure cyanobacteria or algae (P < 0.001). When fed S. elongatus-based diets, the final body length of G. roeselii could be determined only on feces, as all individuals died in the two other food treatments during the experiment. When fed A. variabilis-based diets, the increase in body length was significantly less pronounced on the pure cyanobacterium than on A. variabilis-containing feces or PSF (P < 0.001). Final lengths of gammarids raised on the different N. limnetica- or S. obliquus-based diets differed significantly with higher values obtained on PSF than on feces or pure algae. No gammarid survived on a pure S. obliquus diet and thus the final body length could not be determined (Tab. 10, Fig. 24). The final body lengths of gammarids raised on the cyanobacterium A. variabilis were significantly lower than the body lengths of gammarids raised on the eukaryotic alga N. limnetica (P = 0.036). Final body lengths of G. roeselii were significantly lower when fed S. elongatus-based feces than when fed the other autotroph-based feces.

83 5 Clam mediated benthic pelagic coupling

Tab. 10 Two-way analyses of final lengths of surviving Gammarus roeselii with food species (i.e. S. elongatus, A. variabilis, S. obliquus, N. limnetica) and food treatments (i.e. autotroph, feces, PSF) as independent factors. Differences between food treatments were analysed separately for each autotroph species and differences between autotroph species were analysed separately for each food treatment using Tukey’s HSD tests. In cases where only two gammarids survived, a t-test was performed. Different letters indicate significant differences between treatments (P < 0.05).

F value df p-value Tukey’s HSD post-hoc test

S. elongatus (A), A. variabilis (AB), species 7.68 2 <0.001* S. obliquus (B), N. limnetica (B);

treatment 35.60 3 <0.001* autotrophs (A), feces (B), PSF (B)

species*treatment 7.74 4 <0.001*

F value df p-value autotrophs feces PSF

S. elongatus - - - - A -

A. variabilis 14.79 2 <0.001* A B B

S. obliquus t = 2,13 - <.049 - A B

N. limnetica 22.62 2 <0.001* A B C

F value df p-value S. elongatus A. variabilis S. obliquus N. limnetica

autotrophs t = -2.98 - 0.036* - A - B

feces 7.28 3 <0.001* A B B AB

PSF 10.66 3 <0.001* - A A B

Apart from that, body lengths did not differ significantly among the different kinds of feces offered to G. roeselii. The final body lengths of G. roeselii fed N. limnetica-based PSF was higher than the that of gammarids fed A. variabilis- or S. obliquus-based PSF; gammarids fed S. elongatus-based PSF did not survive the experimental period (Tab. 10, Fig. 24).

84 5 Clam mediated benthic pelagic coupling

Discussion Benthic-pelagic coupling mediated by bivalves has been shown to quantitatively affect the flux of energy from pelagic sources to the benthic food web (Newell 2004). It has been poorly studied, however, whether this bivalve-mediated coupling also qualitatively improves phytoplankton-derived carbon for benthic invertebrates in comparison to direct sedimentation of phytoplankton. Our results indicate that biodeposition of organic matter by bivalves not only shifts pelagic resources to the benthos (e.g. shown by Gergs et al. 2009), but also increases the nutritional value of pelagic autotrophs as food for a benthic invertebrate. We show that survival and somatic growth of the gammarid G. roeselii is enhanced by feeding on biodeposition material produced by the clam C. fluminea as compared to feeding directly on sedimented pelagic algae or cyanobacteria.

In our laboratory study, growth and survival of gammarids were lowest when raised directly on sedimented cyanobacteria or algae, i.e. without passage through the clam. Remarkable, however, is the finding that G. roeselii was able to grow and survive the experimental period of 7 weeks with sedimented autotrophs as the sole food source; this was the case for the filamentous cyanobacterium A. variabilis and the eukaryotic alga N. limnetica. Compared to starved animals, which all died within the first days of the experiment, survival of gammarids was extended also when provided with S. elongatus or S. obliquus, indicating that the animals were able to use these sedimented food sources to some extent. This suggests that freshwater benthic invertebrates benefit from sedimented pelagic carbon sources and that the intensity of this beneficial effect depends on the phytoplankton species composition. Survival of G. roeselii was highest on clam feces in all four autotroph treatments, followed by survival on pseudofeces; survival was lowest on autotrophs without clam conditioning. The observed clam mediated food quality improvement might be due to packing of food items into larger particles during passage through the clam (Atkinson et al. 2011). Filtered autotroph cells are concentrated in clam born mucus prior to their release as feces or PSF and these ‘packets’ of food might be more easily accessible for G. roeselii than the unconditioned autotroph cells. Moreover, besides trapping of particles from the water column, clam born mucus itself might be also of nutritional interest for gammarids, as it harbors and potentially supports growth of microorganisms (Connor and Quinn 1984; Davies et al. 1992; Guo et al. 2009). Differences in food quality between feces and PSF might be due to a different constitution of deposited organic matter. When food particles are ingested and at least partly digested by the clams, changes in their morphology or stability may improve the efficiency of ingestion or digestion by deposition feeders, potentially explaining differences in food quality between feces and PSF as found in our study.

The qualitative improvement of pelagic carbon by the clams was most evident when G. roeselii was fed S. elongatus-based diets, as growth and survival of G. roeselii was possible only on feces produced by clams feeding on S. elongatus. In contrast, diets based on the filamentous cyanobacterium A. variabilis did not differ in their effects on survival rates of G. roeselii. We propose that these differences in food quality between the two cyanobacteria

85 5 Clam mediated benthic pelagic coupling are based on morphological differences between single celled S. elongatus and colony forming A. variabilis. However, the final body length of G. roeselii was higher in animals fed A. variabilis-based biodeposition materials (i.e. feces and PSF) than in animals raised directly on A. variabilis. The structural benefit of clam biodeposited material for G. roeselii is not evident for colony forming autotrophs, as gammarids were able to ingest A. variabilis even without clam conditioning. Our data imply that the observed nutritional upgrading of cyanobacterial food mediated by C. fluminea is at least partially due to an enrichment of the biodeposited material with biochemical nutrients during passage through the clam. Thus, the traces of sterols and PUFAs detected in S. elongatus-based biodeposits, in particular in feces, may have increased the probability of survival and may have allowed for at least moderate growth of G. roeselii. This adds to previous findings showing that the growth of aquatic invertebrates (among them C. fluminea) on cyanobacterial diets is constrained by a deficiency in essential biochemicals (Martin-Creuzburg et al. 2008; Basen et al. 2012).

The role of cyanobacteria in aquatic food webs is of scientific interest both in freshwater and marine ecosystems. As a consequence of cyanobacterial mass developments, high amounts of pelagic carbon are transferred to the benthic food web (Nascimento et al. 2008; Suikkanen et al. 2010). When bivalves feed on cyanobacteria, this usually results in low ingestion rates and in an increased production of pseudofeces, irrespective of cyanobacterial shape or toxin production (Bastviken et al. 1998; Pires et al. 2005; Bontes et al. 2007). It has been shown that cyanobacterial carbon can be ingested and assimilated by benthic invertebrates, but the additional carbon supply did not improve growth conditions (Karlson et al. 2008; Nascimento et al. 2008).

It has been reported that the formation of biodeposition material is associated with changes in the elemental stoichiometry, i.e. the elemental composition of ingested seston differed from that of biodeposited material in particular with regard to the phosphorus content, which was significantly reduced by bivalve conditioning (Gergs et al. 2009). Our results indicate similar mechanisms by showing that total fatty acid and PUFA levels are reduced in biodeposited material as compared to autotrophs which were not processed by the clams and thus support previous findings obtained with marine bivalves (Bradshaw et al. 1991). Therefore, the improved food quality of biodeposited material for G. roeselii was presumably not related to changes in the total dietary fatty acid content. However, food quality is defined not only by the total fatty acid content but also by the availability of essential lipids. The finding that cyanobacteria did not contain any sterols and long chain PUFAs and that sterols and EPA were detected in cyanobacterial biodeposits suggests that the observed food quality improvement for G. roeselii was due to a clam-mediated dietary enrichment with these essential lipids. Although we cannot totally exclude that the small amounts of sterols and PUFAs detected in feces and PSF produced with cyanobacterial diets partially originated from water microbes passing through the 1 µm filter in our flow-through system, we propose that these lipids were released by clams during processing of ingested particles in the mantle cavity, e.g. as part of mucus or of excretion products, and then were packed in feces and PSF. This hypothesis is supported by Bradshaw et al. (1991) who already reported

86 5 Clam mediated benthic pelagic coupling an enrichment of individual sterols and fatty acids in biodeposits of a marine bivalve. These sterols and PUFAs were detectable only in feces and PSF produced on cyanobacterial diets, i.e. in the absence of phytoplankton derived sterols and PUFA. In feces and PSF produced on the eukaryotic algae, these sterols and PUFAs could not be detected presumably because such small amounts were masked by other sterols and PUFAs originating from the algae in the respective analyses.

In marine amphipods, cholesterol was found to be the dominating sterol (Nelson et al. 2001). Also high levels of PUFAs, especially EPA and DHA, have been reported to occur in gammarid tissues (Nelson et al. 2001; Kolanowski et al. 2007). It was suggested that benthic invertebrates (e.g. gammarids) may selectively consume PUFA-rich particles from sediments to cover their PUFA demands (Goedkoop et al. 2000; Makhutova et al. 2003). In our study, high levels of PUFAs and sterols were provided by the eukaryotic food sources, especially by N. limnetica, which might explain why this alga is of superior food quality for G. roeselii. Additionally, enrichment with PUFAs in S. elongatus feces (ARA, EPA) and N. limnetica PSF (DHA) may have led to increased growth in juvenile gammarids.

Another important factor determining food quality for aquatic invertebrates is the stoichiometry of carbon, nitrogen and phosphorus (Elser et al. 2000; Frost et al. 2002). A stoichiometric mismatch between diet and consumer can lead to lower growth rates of the consumer even at saturating food quantity (Frost and Elser 2002). For example, allochthonous leaves are an important energy source for benthic shredders, but are a low- quality food because of low phosphorus and nitrogen contents (Cross et al. 2005; Gergs and Rothhaupt 2008). However, in our study elemental stoichiometric constraints can be excluded, because food organisms and their biodeposition material had overall high N and P loads and little to none changes were found in stoichiometry after clam treatment.

Conclusions Biodeposition constitutes a mechanism that potentially supports benthic food webs. As hypothesized, we found that biodeposition by C. fluminea provides a nutritional link between phytoplankton and benthic invertebrates and that this process is associated with an upgrading of phytoplankton food quality. We show that growth and survival of juvenile gammarids on sedimented autotrophs differs among autotroph species but in general are lower than on the same autotrophs biodeposited by C. fluminea. This upgrading of food quality is partially due to structural changes of food particles during food processing by the clams (i.e. packing in mucus), as direct consumption of sedimented phytoplankton by gammarids is presumably hampered by small cell sizes, but also to a clam-mediated supplementation with lipids essential for gammarid growth. In particular, we show that clam conditioning increases the food quality of cyanobacteria for G. roeselii, presumably by providing sterols. Similarly, in experimental pelagic food chains it has been shown that the food quality of cyanobacterial food sources can be upgraded by heterotrophic protists for subsequent use by Daphnia, by dietary provisioning with essential lipids, i.e. sterols and 87 5 Clam mediated benthic pelagic coupling

PUFAs (Martin-Creuzburg et al. 2005a; Bec et al. 2006). Our results will help to elucidate the importance of bivalve-mediated benthic-pelagic coupling for benthic invertebrates, in particular with regard to a potential channeling of essential nutrients to the detritus based food web.

Acknowledgements We thank J. Kim for establishing the setup and help conducting the experiments. This work was funded by the DFG (German Research Foundation) within the collaborative research centre CRC 454 ‘Littoral of Lake Constance’ and the RISE (Research Internships in Science and Engineering) program funded by the German Academic Exchange Service (DAAD).

Supplementary data Appendix Table S4 Table of lipid composition of food sources fed to Gammarus roeselii

88 6 Concluding remarks and perspectives

Chapter 6

Concluding remarks and perspectives

The Asian clam Corbicula fluminea was one of many non-native species, which invaded Lake Constance in the last decades. Since its first observation in 2003, the clam established a huge population, mostly in the eastern part of the lake in the vicinity of its first observation (ANEBO ; Werner 2008). From there, C. fluminea spread along the southern and northern shore of Lake Constance and now, nine years after its first observation, has invaded almost the complete basin of upper Lake Constance. However, in comparison to the invasion of D. polymorpha in the late 60s, the invasion and dispersal of C. fluminea occurred rather slowly (ANEBO ; Siessegger 1969). Dispersal of bivalves can be attributed to crawling and to larval drift via mucus draglines, although C. fluminea has no pelagic stage (Prezant and Chalermwat 1984). Werner (2008) speculated that apart from natural drift of larvae the attachment on boats and bird legs has contributed to the establishment in remote areas of the lake.

The successful spread of C. fluminea is presumably favored by many autecological traits such as early sexual maturity, high fertility, brood care and a wide tolerance to various environmental factors (Rajagopal et al. 2000; McMahon 2002; Sousa et al. 2008). Nowadays, numbers of non-native species in freshwater ecosystems increase rapidly, quite often with impacts on the stability of ecosystems and global species diversity (Sala et al. 2000). Introduction of alien species can have severe effects, e.g. on the species composition, ecosystem processes and modified nutrient flows and trophic interactions within a community. To predict and assess such impacts, a thorough knowledge of the autecology and life cycle of the alien species is required. We are only beginning to understand effects resulting from invasions of non-native species (Sala et al. 2000; Walther et al. 2009). In this context, nutritional requirements of species have to be considered. In general, nutritional requirements of freshwater bivalves are poorly investigated, in particular with regard to essential biochemicals.

Due to their high filtration rate bivalve populations are able to influence phytoplankton community composition (Lauritsen 1986; Cahoon and Owen 1996; Bastviken et al. 1998; Welker and Walz 1998). Likewise, annual succession of phytoplankton is associated with changes in the availability of elemental and biochemical nutrients for filter-feeding organisms (Wacker and Von Elert 2001; Hartwich et al. 2012; see chapter 4). This is often reflected in the lipid content of benthic organisms and can provide clues to the ecology of those organisms (Cavaletto and Gardner 1999). Seasonality in seston development and short term depletion of food quantity and quality might influence clams variously, and at least temporarily they might be affected by the supply of food and by the supply of nutrients. Co- limitation and interactions between potentially limiting nutrients, as discussed for filter- 89 6 Concluding remarks and perspectives feeding cladocerans (Martin-Creuzburg et al. 2009; Hartwich et al. 2012), are likely to happen in the field and also to affect also the performance of benthic organisms. However, with their internal storages adult clams might temporarily overcome unfavourable conditions.

For growth season in oligotrophic Lake Constance my study revealed a strong impact of water temperature on growth and survival of C. fluminea (chapter 4), thereby confirming previous assumptions (Werner 2008; Werner and Rothhaupt 2008). Plankton succession and seston nutrient parameters were also strongly controlled by temperature regime. Seston was dominated by high proportion of PUFA and sterol containing algae resulting in favourable growth conditions for C. fluminea during spring and summer months. In winter months clams were faced low carbon and low essential lipid supply resulting in growth stagnation and massive die-off events (chapter 4). Clam growth was affected by other seston parameters (i.e. C:P ratio, seston green algae- diatom-, carbon- or ARA- content), however, many of these factors were also influenced by water temperature (co-linearity) and hardly separable from temperature effects in our experimental dataset.

Invasion of C. fluminea in lower Lake Constance probably will take place in the next years (see Fig. 1, Werner pers. communication), with unpredictable consequences for benthic (and pelagic) communities. Lower Lake Constance provides the warmer habitat, has a higher phosphorus load in comparison to the upper basin of Lake Constance and the sediment is dominated by soft substrate (IGKB 2010), factors which all favour C. fluminea. Laboratory findings (chapter 2, 3), showing that C. fluminea growth is constrained by limitation in essential lipids, might impair in habitats, where nutrient load is higher (eutrophic systems) or in times of massive bloom formation, when temperature is sufficient for clam growth and plankton succession and composition might regulate clam growth performance, possibly present in lower Lake Constance.

However, in upper Lake Constance temperature was the main driver for seston dynamics as well as clam growth (chapter 4). For lotic ecosystems, temperature regimes have been shown to support clam establishment and to favour the spread of C. fluminea (Schöll 2000). Beneficial effects of increased temperatures on clam growth were observed in European water ways at cooling water outflows of power plants, which are often inhabited by invasive species (Cairns and Cherry 1983; French and Schloesser 1991). With ongoing dynamics of global warming multiple factors will bias ecosystems processes. In aquatic ecosystems, changing temperature regimes will affect plankton succession and species composition, lake water exchange and vertical mixing, ice cover in winter, nutrient recycling and eutrophication symptoms (Mooij et al. 2005; Jöhnk et al. 2008; Adrian et al. 2009). Additionally, the invasion of non-native species is favoured by global warming (Stachowicz et al. 2002; Walther et al. 2009).

For Lake Constance, recent monitoring indicates that the duration of thermal stratification and therefore the phytoplankton growing season is prolonged by climate change (Straile et al. 2003). In addition, expected milder winter temperatures will reduce thermal stress 90 6 Concluding remarks and perspectives organisms have to cope with (Straile et al. 2010). For C. fluminea, an increase in water temperature will increase the reproductive success and potentially may facilitate a second reproduction event per year (Mouthon 2001; Sousa et al. 2008; Werner 2008). If clams are supported by increasing temperature, this may result in an enhanced top-down control of phytoplankton by filter feeding clams. Additionally, C. fluminea itself may possibly be shifted from temperature- to nutrient-regulated performance regarding growth, reproduction and mortality. Mortality of C. fluminea may change from low temperature determined winter die-off to heat stress regulated summer mortality. In combination with low or even limiting carbon supply during summer months, thermal stress may result in higher summer mortality (Weitere et al. 2009; Vohmann et al. 2010).

Mortality of clams can tremendously change habitat structure as shown by Werner et al. (2008) for Lake Constance during winter in 2005/2006, when the majority of the C. fluminea population suffered from low water level and extremely cold water temperatures. Massive clam die-off events result in an accumulation of bare shells on the ground, whereas shells are typically covered with sand. The clam-mediated structural modification of the habitat at the Rohrspitz bay (Fig. 1) increased substrate diversity and settlement surface for benthic macroinvertebrates, like D. polymorpha or G. roeselii (Werner and Rothhaupt 2007; Werner 2008; Gergs et al. 2011; personal observations).

Due to its filtration capacity bivalves can shift a substantial part of pelagic primary production as biodeposition material, i.e. either as digested faeces or undigested pseudofaeces, to the benthic food web (Gergs et al. 2009; Atkinson et al. 2011). Bivalves act as a food ‘donor’ and food quality ‘converter’ for benthic invertebrates by altering the bioavailability of food and its elemental and biochemical (fatty acids, sterols) composition in biodeposition material (Gergs et al. 2009; chapter 5). The impact of this transformation process itself and the role of essential biochemical nutrients have not been considered as a mechanism determining trophic interactions between benthic herbivores and its phytoplankton food source. My results show, for the first time, that filter-feeding bivalves not only provide a crucial link between the pelagic and benthic food web by increasing the bioavailability of pelagic resources in comparison to sedimented algae (chapter 5), but that clam conditioning also is associated with changes in the biochemical composition of pelagic food sources. Especially the modification in essential lipids (sterols, PUFAs) improves growth conditions for gammarids, that are constrained by a low availability of dietary lipids when fed cyanobacterial carbon. In a tritrophic food chain ‘trophic upgrading’ by intermediary protists has shown to improve cyanobacterial carbon for filter-feeding cladocerans (Klein Breteler et al. 1999; Martin-Creuzburg et al. 2005a). In my setup, sterol-depleted food sources might have also been biochemically ‘upgraded’ during clam conditioning for subsequent use by benthic consumers via the incorporation of clam-derived lipids in biodeposition materials. Hence, my results will help to elucidate the importance of bivalves in mediating a benthic-pelagic coupling and to assess effects of food quality on the benthic community.

91 6 Concluding remarks and perspectives

However, apart from carbon transfer via pelagic-benthic coupling also during bloom conditions huge biomass is introduced into benthic food web. Diatom blooms settling to the sediment are considered to be the main sources of high quality food for benthos (Fitzgerald and Gardner 1993), whereas cyanobacterial carbon has been reported to be of minor nutritional quality for benthic invertebrates (Karlson et al. 2008; Nascimento et al. 2009; Suikkanen et al. 2010). Aquatic cyanobacteria often form extensive blooms, both in freshwater and marine systems, and tend to increase with global warming. The accumulation of cyanobacterial biomass may severely affect herbivorous consumers. C. fluminea growth rates were significantly lower when juveniles were fed cyanobacterial diets instead of eukaryotic algae (chapter 2). My results present first insight into food quality effects on C. fluminea nutrition which can be related to essential dietary substances missing in former laboratory studies (Foe and Knight 1986). However, correlations of nutritional parameters with somatic growth rates are a first step towards the identification of mechanisms affecting clam fitness. Based on results from chapter 2, the diet of clams was experimentally manipulated in chapter 3 to assess the significance of lipid substances for clam nutrition. The absence of essential lipids in cyanobacteria (sterols, PUFAs) has been identified as a major food quality constraint for filter-feeding Daphnia and herein for Corbicula as well (Müller-Navarra et al. 2000; Von Elert et al. 2003; chapter 2, chapter 3). Whereas C. fluminea could not benefit from PUFA supplementation when fed with cyanobacteria (data not presented here), sterols enrichment improved growth conditions for clams (chapter 3), indicating strong dependency on dietary sterols. For the first time, I attributed growth conditions of C. fluminea to nutritional lipids (chapter 2) and identified sterols as determining factor in clam nutrition on cyanobacterial diets (chapter 3).

In times of insufficient nutritional supply filter feeding bivalves might suffer from seston composition, especially in times of cyanobacterial blooms. In contrast to other filter feeding bivalves, C. fluminea is regarded as a non-selective suspension feeder (Way et al. 1990; Vaughn and Hakenkamp 2001) and thus might not be able to discriminate against nutritionally inadequate food particles. However, other mechanisms might help C. fluminea to overcome unfavourable nutritional conditions. It remains to be tested if C. fluminea is able to adjust filtration or assimilation rate in order to increase utilization of food sources and gain more specific limiting nutrients (compensatory feeding) or switch from filtration to deposition feeding to obtain its food from the sediment (pedal feeding) and thus avoid nutritionally inadequate food sources present in the water column. However, as organic carbon concentrations in sediments of Lake Constance are very low (1 - 5%, not presented; Deutzmann pers. communication; Van Der Velden and Schwartz 1976), I assume that additional carbon will not completely satisfy clam maintenance. Nevertheless, pedal feeding might contribute to the dietary provisioning with essential nutrients and therefore might improve the performance of clams and possibly provide advantage over exclusively filter- feeding bivalves.

Certainly, it would be very interesting to see whether differences between C. fluminea and native/established bivalves in their ability to persist and survive cyanobacterial blooms affect

92 6 Concluding remarks and perspectives the invasion success. Investigating and comparing the physiological demands of native and invasive species may help to understand invasion patterns and to improve risk assessment of upcoming invasions. Additionally, mechanisms to avoid unfavourable nutritional conditions are known in bivalve filtration (selective feeding), it remains to be tested if pedal feeding might allow C. fluminea to decouple their performance from (unfavourable) seston composition.

Aquatic herbivores face a multitude of nutritional challenges in natural environments. Sterol limitation might be one factor with the potential to affect the structure of aquatic food webs, at least seasonally and in certain habitats. Apart from sterols, biochemical food quality depends on a multitude of other factors such as PUFAs, amino acids, vitamins and presumably their elemental stoichiometric composition. Over the last years, considerable advance has been made in identifying food quality aspects and their role in invertebrate performance. Food quality research has mainly focused on elemental nutrients and polyunsaturated fatty acids as potentially limiting resources. This thesis provided data showing that the dietary sterols should also be taken into account when analyzing nutritional constraints and their effects on trophic interactions in freshwater food webs. Following studies with C. fluminea might investigate co-limitation aspects of factors mediating food quality in cyanobacteria and other eukaryotic algae under field and laboratory conditions. Apart from growth performance of juvenile clams, reproduction in adult specimen, larval development and juvenile recruitment might face different nutritional demands, necessary for understanding the life history of C. fluminea. My data might help to improve our knowledge on the consequences of cyanobacterial mass developments for benthic filter-feeders and highlights the importance of sterols as potentially limiting nutrients in aquatic food webs.

93 6 Concluding remarks and perspectives

94 Abstract

Abstract

The Asian clam Corbicula fluminea, originating from East Asia, has become a widespread benthic invertebrate in many freshwater ecosystems throughout Europe and North and South America. With the introduction of invasive bivalves drastic changes in established benthic communities often occur. Knowing the nutritional demands in ecosystem processes is crucial for understanding the flow of energy in food webs, and for investigating species interactions within the food web. In this context, nutritional requirements of species have to be considered, which are poorly investigated for freshwater bivalves, in particular with regard to essential biochemicals. Therefore, the aim of my study was to gain knowledge on nutritional demands of C. fluminea. In particular, I examined the role of essential lipids (sterols, polyunsaturated fatty acids) for juvenile clams both in the laboratory and in the field, i.e. in Lake Constance. Further on, I focused on the benthic-pelagic coupling mediated by C. fluminea, and the possible nutrient transfer and modulation of food quality effects onto detritivorous benthic invertebrates.

Laboratory growth experiments with C. fluminea raised on eukaryotic and prokaryotic food sources revealed strong dependency of clam growth rates on lipid composition of food sources. Somatic growth rates were significantly higher when juveniles were fed eukaryotic algae instead of cyanobacteria. Linear regression analyses revealed significant positive relationships between clam somatic growth rates and dietary sterol and polyunsaturated fatty acid content. Based on these results, a supplementation approach was conducted, in which C. fluminea was fed again with pro- and eukaryotic food sources. Additionally, sterol depleted cyanobacterial food was enriched with sterols. The growth rate improvement of clams fed with enriched food sources revealed sterol limited growth on cyanobacteria. I have shown for the first time that a benthic freshwater invertebrate is constrained by an absence of dietary sterols and I pointed out growth enhancing effects by sterols addition when feeding on cyanobacterial food sources. Cyanobacteria, as prokaryotic phototrophs, lack essential lipids (sterols, long chain PUFAs) and are already known as nutritional inadequate food source for pelagic filter-feeding organisms. The accumulation of cyanobacterial biomass may severely affect the performance of aquatic consumers, not only pelagic zooplankton, but also benthic filter feeding bivalves might be affected by cyanobacterial carbon and be limited by low sterol and PUFA supply.

Filter feeding clams are largely affected by seston dynamics in the pelagial. Thus, seasonal variation in water temperature and seston composition may determine the growth performance of benthic clams. To elucidate C. fluminea somatic growth under natural conditions, a field study of clam growth and survival was conducted in Lake Constance throughout 2010. Temperature and food quantity and quality effects were investigated and correlated with clam fitness parameters. Analyses showed that the growth and survival of C. fluminea in Lake Constance is basically determined by water temperature, with high

95 Abstract mortality rates during winter and increased growth rates during summer when temperature and carbon supply favour growth of C. fluminea. Concomitant laboratory experiments confirmed the dominant influence of temperature on growth of C. fluminea. The field study also revealed that the conditions favouring clam growth are restricted to only seven months of the year, when temperature exceeded 10°C. Hence, the C. fluminea population in Lake Constance may benefit from the expected climate change resulting in milder winter temperatures and an earlier onset of thermal stratification in spring because of reduced winter mortality and an extended growing season, i.e. the period in which water temperatures exceed 10°C.

Apart from nutritional aspects, I investigated the role of C. fluminea as mediator between phytoplankton dynamics and benthic detritus feeders, via the establishment of a laboratory experimental food chain consisting of different pelagic primary producers (algae, cyanobacteria), the freshwater clam C. fluminea as primary consumer, and Gammarus roeselii as omnivorous consumer feeding on biodeposition material produced by C. fluminea. The clam was used as benthic-pelagic coupler and the produced biodeposition material was collected and fed to G. roeselii. Data showed that the survival and growth of gammarids significantly increased when autotrophs were offered to G. roeselii as biodeposition material compared to sedimented autotrophs. This upgrading of food quality was hypothesised to derive from an increased accessibility of pelagic food for G. roeselii and to a clam-mediated dietary enrichment with essential elemental and biochemical (fatty acids, sterols) nutrients. My results show that filter-feeding clams not only provide a crucial link between the pelagic and benthic food web by changing energy fluxes, but also transform pelagic resources and increase the bioavailability of pelagic resources in comparison to sedimented algae. Furthermore, data suggest that this pelagic-benthic coupling is constrained by a low availability of essential biochemical nutrients; in particular during cyanobacterial blooms dietary lipids are scarce for gammarids.

Thus, my findings contribute significantly to our understanding of the consequences associated with cyanobacterial mass developments, which are expected to be favoured by global warming, for benthic filter-feeding clams. They also highlight the importance of considering sterols as potentially limiting nutrients in aquatic food webs. My results suggest that the growth of C. fluminea is partially dependent on the availability of essential lipids in the diet. Additionally, C. fluminea has the potential to mediate nutritional effects from pelagic to benthic food webs and may therefore play an important trophic role in the benthic food web.

96 Zusammenfassung

Zusammenfassung

Die aus Ostasien stammende asiatische Körbchenmuschel Corbicula fluminea hat sich in den letzten Jahrzenten zu einem weitverbreiteten benthischen Wirbellosen in vielen Süßwasser- Ökosystemen entwickelt. Mit der Einführung der invasiven Muscheln in neue Lebensräume sind oftmals drastische Veränderungen in etablierten Lebensgemeinschaften verbunden. Um Wechselwirkungen und Energieflüsse zwischen Organismen im Nahrungsnetz zu verstehen, ist im Besonderen die Kenntnis des Nährstoffbedarfs der verschiedenen Arten wesentlich. Dieser wurde bisher für Süßwassermuscheln kaum untersucht, insbesondere im Hinblick auf essentielle biochemische Makromoleküle. Ziel meiner Arbeit war es daher, Erkenntnisse über den Nährstoffbedarf von C. fluminea zu gewinnen. Im Besonderen untersuchte ich die Rolle von essentiellen Lipiden (Sterole, mehrfachungesättigte Fettsäuren) für juvenile Muscheln im Labor und unter natürlichen Bedingungen im Jahresverlauf des Bodensees. Außerdem beschäftigte ich mich mit der bentisch-pelagischen Kopplung durch C. fluminea, sowie mit einem möglichen Nährstofftransfer und einer Modulation von Futterqualitätseffekten auf detrivore, bentische Invertebraten.

Zu diesem Zweck etablierte ich standardisierte Laborwachstumsexperimente mit C. fluminea. Dort konnte ich mit eukaryotischen und prokaryotischen Nahrungsquellen für die Muscheln einen Zusammenhang zwischen Muschelwachstumsraten und der Lipidzusammensetzung der Nahrungsquellen aufzeigen. Somatische Wachstumsraten waren signifikant höher, wenn Jungtiere mit eukaryotischen Algen gefüttert wurden, im Vergleich zu cyanobakteriellem Futter. Lineare Regressionsanalysen konnten signifikante positive Zusammenhänge zwischen somatischen Muschelwachstumsraten und dem Gehalt an Sterol sowie mehrfach-ungesättigten Fettsäuren der Nahrungsquellen hervorheben. Basierend auf diesen Ergebnissen wurde ein weiterer Laborversuch durchgeführt, in dem C. fluminea erneut mit pro- und eukaryotischen Nahrungsquellen gefüttert wurden. Zusätzlich wurden sterolfreie Cyanobakterien mit Sterolen angereichert und als Nahrung für die Muscheln angeboten. Durch die Steigerung der Muschelwachstumsrate auf sterolangereicherten Cyanobakterien konnte ich nachweisen, dass das Wachstum von C. fluminea sterollimitiert ist, wenn Cyanobakterien als einzige Futterquelle zur Verfügung stehen. Zum ersten Mal wurde gezeigt, dass ein benthischer Süßwasserwirbelloser durch die Ernährung mit Cyanobakterien sterollimitiert ist und eine Anreicherung der Kost mit Sterolen das Wachstum fördern kann. Cyanobakterien als prokaryotische Phototrophe enthalten keine essentiellen Lipide (Sterole, langkettige PUFA) und sind bereits als schlechte Nahrungsquellen für pelagische Filtrierer bekannt. Das Entstehen von Cyanobakterienblüten kann somit ernsthafte Auswirkungen auf filtrierende Konsumenten wie pelagisches Zooplankton oder auch benthische Muscheln haben. Diese Filtrierer könnten durch cyanobakteriellen Kohlenstoff beeinflusst und durch eine geringe Sterol- und PUFA- versorgung limitiert werden.

97 Zusammenfassung

Filtrierende Muscheln sind weitgehend von der Sestondynamik im Pelagial abhängig. So können jahreszeitliche Schwankungen der Wassertemperatur und der Sestonzusammensetzung einen erheblichen Einfluss auf das Wachstum benthischer Muscheln haben. Um die Entwicklung von C. fluminea unter natürlichen Bedingungen zu untersuchen, wurde eine Feldstudie im Bodensee im Jahr 2010 durchgeführt und Wachstum und Überleben der Muscheln dokumentiert. Temperatur und Futterqualitäts- und Futterquantitätsparameter wurden untersucht und mit diversen Muschelfitnessparametern korreliert. Die Ergebnisse zeigen deutlich, dass Wachstum und Überleben von C. fluminea im Wesentlichen von der Wassertemperatur bestimmt wird. Im Winter, bei geringen Temperaturen und niedriger Futterverfügbarkeit, konnten hohe Sterblichkeit und stagnierendes Wachstum beobachtet werden, wohingegen im Sommer, bei erhöhter Wassertemperatur und Kohlenstoffversorgung, deutlich verbesserte Wachstumsbedingungen für Corbicula vorgefunden wurden. In zeitgleichen Laborversuchen konnte der dominante Einfluss der Temperatur ebenfalls aufgezeigt werden. Für die Population der Muschel im Bodensee bedeutet dies, dass nur etwa sieben Monate im Jahr für Wachstum und Reproduktion zur Verfügung stehen, in denen die Wassertemperatur 10 °C übersteigt. Daher wird die Population von C. fluminea im Bodensee durch den zu erwartenden Klimawandel begünstigt werden, da vermutlich die Temperaturbelastung im Winter und somit die Sterblichkeit verringert werden und sich die Vegetationsperiode durch eine längere stabile Schichtung des Epilimnions zusätzlich ausdehnen wird.

Neben den ernährungsphysiologischen Aspekten der Muschel untersuchte ich die Rolle von C. fluminea als Mittler zwischen Phytoplankton und einem benthischen Detritusfresser. Dazu erzeugte ich im Labor eine experimentelle Nahrungskette aus Primärproduzenten (Algen, Cyanobakterien), über einen Primärkonsumenten (C. fluminea) hin zu einem detritivoren Konsumenten (Gammarus roeselii). Die Muschel wurde als pelagisch-benthischer Koppler genutzt und produziertes Biodepositionsmaterial von C. fluminea wurde gesammelt und an G. roeselii verfüttert. Überleben und Wachstum der Gammariden wurden signifikant erhöht, wenn die Gammariden statt sedimentierten Autotrophen das Biodepositionsmaterial der Muscheln fressen konnten. Neben der Bioverfügbarkeit von Nahrungspartikeln für die Gammariden veränderte die Muschel durch Einlagerung von essentiellen Lipiden auch die biochemische Zusammensetzung in den Futterpartikeln, was sich ebenfalls positiv auf Wachstum und Überleben der Gammariden auswirkte. Meine Ergebnisse zeigen, dass filtrierende Muscheln nicht nur ein wichtiges Bindeglied zwischen dem pelagischen und benthischen Nahrungsnetz sind, hervorgerufen durch den quantitativen Eintrag von Ressourcen, sondern auch eine qualitative Modulation im Stofffluss hervorrufen können. Weiter belegen die Daten, dass diese pelagisch-benthische Kopplung durch eine geringe Verfügbarkeit von essentiellen biochemischen Inhaltsstoffen eingeschränkt ist, insbesondere wenn bei Cyanobakteriendominanz im Pelagial Lipidinhaltsstoffe limitierend für Gammariden werden können.

Somit können meine Erkenntnisse dazu beitragen, die Auswirkungen von Cyanobakterienblüten auf Ökosysteme zu verstehen, deren Erscheinen in Folge der

98 Zusammenfassung

Klimaerwärmung begünstigt wird. Ich konnte zeigen, dass das Wachstum von C. fluminea teilweise von der Verfügbarkeit von essentiellen Lipiden in der Nahrung abhängt. Des Weiteren kann die Muschel Dynamiken des pelagischen ins benthische Nahrungsnetz transferieren und somit trophische Interaktionen im Benthos beeinflussen. Meine Arbeit unterstreicht die Bedeutung von Sterolen als potenziell limitierendem Nährstoff im aquatischen Nahrungsnetz.

99 Zusammenfassung

100 Bibliography

Bibliography

Adrian R, O'Reilly CM, Zagarese H et al. (2009) Lakes as sentinels of climate change. Limnology and Oceanography 54:2283-2297

Ahlgren G, Goedkoop W, Markensten H, Sonesten and Boberg M (1997) Seasonal variations in food quality for pelagic and benthic invertebrates in Lake Erken – the role of fatty acids. Freshwater Biology 38:555-570

Ahlgren G, Gustafsson IB and Boberg M (1992) Fatty acid content and chemical composition of freshwater microalgae. Journal of Phycology 28:37-50

Akaike H (1974) A new look at the statistical model identification. IEEE Transactions on Automatic Control 19:716-723

Aldridge DW and McMahon RF (1978) Growth, fecundity, and bioenergetics in a natural population of Asiatic freshwater clam, Corbicula manilensis (Philippi), from North Central Texas. Journal of Molluscan Studies 44:49-70

ANEBO Aquatische Neozoen im Bodensee, http://www.neozoen-bodensee.de/. In:

APHA (1985) Standard methods for the examination of water and wastewater. In, 16th edn. American Public Health Association, American Water Works Association, and Water Environment Federation, Washington, DC

Araujo R, Moreno D and Ramos MA (1993) The Asiatic clam Corbicula fluminea (Müller, 1774) (Bivalvia, Corbiculidae) in Europe. American Malacological Bulletin 10:39-49

Arnott DL and Vanni MJ (1996) Nitrogen and phosphorus recycling by the zebra mussel (Dreissena polymorpha) in the western basin of Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 53:646-659

Atkinson CL, First MR, Covich AP, Opsahl SP and Golladay SW (2011) Suspended material availability and filtration-biodeposition processes performed by a native and invasive bivalve species in streams. Hydrobiologia 667:191-204

Atkinson CL, Opsahl SP, Covich AP, Golladay SW and Conner LM (2010) Stable isotopic signatures, tissue stoichiometry, and nutrient cycling (C and N) of native and invasive freshwater bivalves. Journal of the North American Benthological Society 29:496-505

Bachmann V, Beisel JN, Usseglio-Polatera P and Moreteau JC (2001) Decline of Dreissena polymorpha in the River Moselle: biotic and abiotic key factors involved in dynamics of invasive species. Archiv für Hydrobiologie 151:263-281

Baker S, Levinton J and Ward J (2000) Particle transport in the zebra mussel, Dreissena polymorpha (Pallas). The Biological Bulletin 199:116-125

101 Bibliography

Basen T, Martin-Creuzburg D and Rothhaupt K-O (2011) Role of essential lipids in determining food quality for the invasive freshwater clam Corbicula fluminea. Journal of the North American Benthological Society 30:653-664

Basen T, Rothhaupt K-O and Martin-Creuzburg D (2012) Absence of sterols constrains food quality of cyanobacteria for an invasive freshwater bivalve. Oecologia -online early- 1- 8

Bastviken DTE, Caraco NF and Cole JJ (1998) Experimental measurements of zebra mussel (Dreissena polymorpha) impacts on phytoplankton community composition. Freshwater Biology 39:375-386

Bäuerle E, Ollinger D and Imberger J (1998) Some meteorological, hydrological, and hydrodynamical aspects of Upper Lake Constance. Arch. Hydrobiol. Spec. Issues Advanc. Limnol. 53

Bec A, Martin-Creuzburg D and Von Elert E (2006) Trophic upgrading of autotrophic picoplankton by the heterotrophic nanoflagellate Paraphysomonas sp. Limnology and Oceanography 51:1699-1707

Belanger P, Zintel JA, Vandenhe.Wj and Smith JL (1973) Identification by combined gas-liquid chromatography-mass spectrometry of sterols isolated from Scenedesmus obliquus grown in light and heavy water. Canadian Journal of Chemistry - Revue Canadienne de Chimie 51:3294-3298

Belanger TV, Annis CG and Vanepps DD (1990) Growth rates of the Asiatic clam, Corbicula fluminea, in the upper and middle St. Johns River, Florida. Nautilus 104:4-9

Ben Kheder R, Quéré C, Moal J and Robert R (2010) Effect of nutrition on Crassostrea gigas larval development and the evolution of physiological indices: Part B: Effects of temporary food deprivation. Aquaculture 308:174-182

Bij de Vaate A and Greijdanus-Klaas M (1990) The Asiatic Clam, Corbicula Fluminea (Müller, 1774) (Pelecypoda, Corbiculidae), a new immigrant in the Netherlands. Bull Zool Mus Amsterdam 12:173-178

Bij de Vaate A, Jazdzewski K, Ketelaars HAM, Gollasch S and Velde GVd (2002) Geographical patterns in range extension of Ponto-Caspian macroinvertebrate species in Europe. Canadian Journal of Fisheries and Aquatic Sciences 59:1159-1174

Boersma M and Kreutzer C (2002) Life at the edge: Is food quality really of minor importance at low quantities? Ecology 83:2552-2561

Boltovskoy D, Izaguirre I and Correa N (1995) Feeding selectivity of Corbicula fluminea (Bivalvia) on natural phytoplankton. Hydrobiologia 312:171-182

Bontes BM, Verschoor AM, Pires LMD, van Donk E and Ibelings BW (2007) Functional response of Anodonta anatina feeding on a green alga and four strains of cyanobacteria, differing in shape, size and toxicity. Hydrobiologia 584:191-204

102 Bibliography

Bradshaw SA, Ohara SCM, Corner EDS and Eglinton G (1991) Effects on dietary lipids of the marine bivlave Scrobicularia plana feeding in different modes. Journal of the Marine Biological Association of the United Kingdom 71:635-653

Brett MT and Müller-Navarra DC (1997) The role of highly unsaturated fatty acids in aquatic food web processes. Freshwater Biology 38:483-499

Brett MT, Müller-Navarra DC and Park SK (2000) Empirical analysis of the effect of phosphorus limitation on algal food quality for freshwater zooplankton. Limnology and Oceanography 45:1564-1575

Briski E, Bailey SA and MacIsaac HJ (2011) Invertebrates and their dormant eggs transported in ballast sediments of ships arriving to the Canadian coasts and the Laurentian Great Lakes. Limnology and Oceanography 56:1929-1939

Britton JC and Morton B (1979) Corbicula in North America: the evidence reviewed and evaluated. In: Britton JC, Mattice, J.S., Murphy, C.E., Newland, L.W. (ed) Proceedings, First International Corbicula Symposium. Christian University Research Foundation Publication, Fort Worth, Texas, pp pp. 249-287

Cahoon LB and Owen DA (1996) Can suspension feeding by bivalves regulate phytoplankton biomass in Lake Waccamaw, North Carolina? Hydrobiologia 325:193-200

Cain DJ and Luoma SN (1990) Influence of seasonal growth, age, and environmental exposure on Cu and Ag in a bivalve indicator, Macoma balthica, in San Francisco Bay. Marine Ecology-Progress Series 60:45-55

Cairns J and Cherry DS (1983) A site specific field and laboratory evaluation of fish and Asiatic clam Corbicula fluminea population responses to coal fired power plant discharge. Water Science and Technology 15:31-38

Carmichael WW (1992) Cyanobacteria secondary metabolites - the cyanotoxins. Journal of Applied Bacteriology 72:445-459

Cataldo D and Boltovskoy D (1998) Population dynamics of Corbicula fluminea (Bivalvia) in the Parana River Delta (Argentina). Hydrobiologia 380:153-163

Cavaletto JF and Gardner WS (1999) Seasonal dynamics of lipids in freshwater benthic invertebrates. In: Arts MT, Wainmann BC (eds) Lipids in aquatic ecosystems. Springer, New York, pp 109-131

Chandra S and Gerhardt A (2008) Invasive species in aquatic ecosystems: issue of global concern. Aquatic Invasions 3:1-2

Chijimatsu T, Umeki M, Okuda Y, Yamada K, Oda H and Mochizuki S (2011) The fat and protein fractions of freshwater clam (Corbicula fluminea) extract reduce serum cholesterol and enhance bile acid biosynthesis and sterol excretion in hypercholesterolaemic rats fed a high-cholesterol diet. The British Journal of Nutrition:1-9

103 Bibliography

Codd GA (1995) Cyanobacterial toxins: Occurrence, properties and biological significance. Water Science and Technology 32:149-156

Cohen RRH, Dresler PV, Phillips EJP and Cory RL (1984) The effect of the Asiatic clam, Corbicula fluminea, on phytoplankton of the Potomac River, Maryland. Limnology and Oceanography 29:170-180

Connor VM and Quinn JF (1984) Stimulation of food species growth by Limpet mucus. Science 225:843-844

Cook HW (1996) Fatty acid desaturation and chain elongation in eukaryotes. In: Vance DE, Vance JE (eds) Biochemistry of lipids, lipoproteins and membranes. Elsevier Science, Amsterdam, pp 129-152

Cooper NL, Bidwell JR and Cherry DS (2005) Potential effects of Asian clam (Corbicula fluminea) die-offs on native freshwater mussels (Unionidae) II: porewater ammonia. Journal of the North American Benthological Society 24:381-394

Crooks JA (2002) Characterizing ecosystem-level consequences of biological invasions: the role of ecosystem engineers. Oikos 97:153-166

Cross WF, Benstead JP, Frost PC and Thomas SA (2005) Ecological stoichiometry in freshwater benthic systems: recent progress and perspectives. Freshwater Biology 50:1895-1912

Darrigran G (2002) Potential impact of filter-feeding invaders on temperate inland freshwater environments. Biological Invasions 4:145-156

Davies MS, Hawkins SJ and Jones HD (1992) Pedal mucus and its influence on the microbial food-supply of 2 intertidal Gastropods, Patella vulgata and Littorina littorea (L). Journal of Experimental Marine Biology and Ecology 161:57-77

De Bernardi R and Giussani G (1990) Are blue-green algae a suitable food for zooplankton? An overview. Hydrobiologia 200:29-41

Delaunay F, Marty Y, Moal J and Samain JF (1993) The effect of monospecific algal diets on growth and fatty acid composition of Pecten maximus (L) larvae. Journal of Experimental Marine Biology and Ecology 173:163-179

DeMott WR, Gulati RD and Van Donk E (2001) Daphnia food limitation in three hypereutrophic Dutch lakes: Evidence for exclusion of large-bodied species by interfering filaments of cyanobacteria. Limnology and Oceanography 46:2054-2060

DeMott WR, Zhang Q-X and Carmichael WW (1991) Effects of toxic cyanobacteria and purified toxins on the survival and feeding of a copepod and three species of Daphnia. Limnology and Oceanography 36:1346-1357

Den Hartog C, Vandenbrink FWB and Vandervelde G (1992) Why was the invasion of the river Rhine by Corophium curvispinum and Corbicula species so successful? Journal of Natural History 26:1121-1129

104 Bibliography

Duncan M, Fried B, Sherma J and Hoskin GP (1987) Lipids and sterols in Corbicula fluminea (Bivalvia). Comparative Biochemistry and Physiology B-Biochemistry & Molecular Biology 87:881-883

Dunstan GA, Volkman JK, Barrett SM, Leroi J-M and Jeffrey SW (1993) Essential polyunsaturated fatty acids from 14 species of diatom (Bacillariophyceae). Phytochemistry 35:155-161

Eder M, Rinke K, Kempke S, Huber A and Wolf T (2008) Lake-wide field campaign in Lake Constance 2007. Wasserwirtschaft 98:34-38

Ellis BK, Stanford JA, Goodman D et al. (2011) Long-term effects of a trophic cascade in a large lake ecosystem. Proceedings of the National Academy of Sciences of the United States of America 108:1070-1075

Elser JJ, Acharya K, Kyle M et al. (2003a) Growth rate-stoichiometry couplings in diverse biota. Ecology Letters 6:936-943

Elser JJ, Fagan WF, Denno RF et al. (2000) Nutritional constraints in terrestrial and freshwater food webs. Nature 408:578-580

Elser JJ, Kyle M, Makino W, Yoshida T and Urabe J (2003b) Ecological stoichiometry in the microbial food web: a test of the light : nutrient hypothesis. Aquatic Microbial Ecology 31:49-65

Elser JJ, Marzolf ER and Goldman CR (1990) Phosphours and nitrogen limitation of phytoplankton growth in the fresh-waters of North America - A review and critique of experimental enrichments. Canadian Journal of Fisheries and Aquatic Sciences 47:1468-1477

Espinosa EP, Perrigault M, Ward JE, Shumway SE and Allam B (2010) Microalgal cell surface carbohydrates as recognition sites for particle sorting in suspension-feeding bivalves. Biological Bulletin 218:75-86

Evans-White MA, Stelzer RS and Lamberti GA (2005) Taxonomic and regional patterns in benthic macroinvertebrate elemental composition in streams. Freshwater Biology 50:1786-1799

Fink P, Peters L and Von Elert E (2006) Stoichiometric mismatch between littoral invertebrates and their periphyton food. Archiv für Hydrobiologie 165:145-165

Fitzgerald SA and Gardner WS (1993) An algal carbon budget for pelagic-benthic coupling in Lake Michigan. Limnology and Oceanography 38:547-560

Foe C and Knight A (1985) The effect of phytoplankton and suspended sediment on the growth of Corbicula fluminea (Bivalvia). Hydrobiologia 127:105-115

Foe C and Knight A (1986) Growth of Corbicula fluminea (Bivalvia) fed artificial and algal diets. Hydrobiologia 133:155-164

105 Bibliography

French JRP and Schloesser DW (1991) Growth and overwinter survival of the asiatic clam Corbicula fluminea, in the St Clair River, Michigan. Hydrobiologia 219:165-170

Frost PC and Elser JJ (2002) Growth responses of littoral mayflies to the phosphorus content of their food. Ecology Letters 5:232-240

Frost PC, Stelzer RS, Lamberti GA and Elser JJ (2002) Ecological stoichiometry of trophic interactions in the benthos: understanding the role of C : N : P ratios in lentic and lotic habitats. Journal of the North American Benthological Society 21:515-528

Frost PC, Tank SE, Turner MA and Elser JJ (2003) Elemental composition of littoral invertebrates from oligotrophic and eutrophic Canadian lakes. Journal of the North American Benthological Society 22:51-62

Gächter R and Bloesch J (1985) Seasonal and vertical variation in the C : P ratio of suspended and settling seston of lakes. Hydrobiologia 128:193-200

Gaedke U (1992) The size distribution of plankton biomass in a large lake and its seasonal variability. Limnology and Oceanography 37:1202-1220

Gaedke U and Schweizer A (1993) The 1st decade of oligotrophication in Lake Constance .1. The response of phytoplankton biomass and cell-size. Oecologia 93:268-275

Gergs R, Grey J and Rothhaupt KO (2011) Temporal variation in zebra mussel (Dreissena polymorpha) density structure the benthic food web and community composition on hard substrates in Lake Constance, Germany. Biological Invasions 13:2727-2738

Gergs R, Rinke K and Rothhaupt KO (2009) Zebra mussels mediate benthic-pelagic coupling by biodeposition and changing detrital stoichiometry. Freshwater Biology 54:1379- 1391

Gergs R and Rothhaupt KO (2008) Feeding rates, assimilation efficiencies and growth of two amphipod species on biodeposited material from zebra mussels. Freshwater Biology 53:2494-2503

Goad LJ (1981) Sterol biosynthesis and metabolism in marine invertebrates. Pure and Applied Chemistry 53:837-852

Goad LJ and Akihisa T (1997) Analysis of sterol. Blackie Academic & Professional.

Goedkoop W, Sonesten L, Ahlgren G and Boberg M (2000) Fatty acids in profundal benthic invertebrates and their major food resources in Lake Erken, Sweden: seasonal variation and trophic indications. Canadian Journal of Fisheries and Aquatic Sciences 57:2267-2279

Greenberg AE, Trussell RR and Clesceri LS (1985) Standard methods for the examination of water and wastewater. American Public Health Association

Guillard RRL (1975) Cultures of phytoplankton for feeding of marine invertebrates. Smith, W.L. Chanley, M.H. (eds) In Culture of marine invertebrate animals.:29-60

106 Bibliography

Guo F, Huang ZB, Huang MQ, Zhao J and Ke CH (2009) Effects of small abalone, Haliotis diversicolor, pedal mucus on bacterial growth, attachment, biofilm formation and community structure. Aquaculture 293:35-41

Hakenkamp CC and Palmer MA (1999) Introduced bivalves in freshwater ecosystems: the impact of Corbicula on organic matter dynamics in a sandy stream. Oecologia 119:445-451

Hakenkamp CC, Ribblett SG, Palmer MA, Swan CM, Reid JW and Goodison MR (2001) The impact of an introduced bivalve (Corbicula fluminea) on the benthos of a sandy stream. Freshwater Biology 46:491-501

Hanselmann AJ (2011) A review of spatio-temporal patterns of the colonisation of Lake Constance with alien macrozoobenthos. Lauterbornia 72:131-148

Harrison P, Thompson P and Calderwood G (1990) Effects of nutrient and light limitation on the biochemical composition of phytoplankton. Journal of Applied Phycology 2:45-56

Hartwich M, Martin-Creuzburg D, Rothhaupt K-O and Wacker A (2012) Oligotrophication of a large, deep lake alters food quantity and quality constraints at the primary producer- consumer interface. Oikos -online early-

Hassett RP (2004) Supplementation of a diatom diet with cholesterol can enhance copepod egg-production rates. Limnology and Oceanography 49:488-494

Hessen DO, Faerovig PJ and Andersen T (2002) Light, nutrients, and P:C ratios in algae: Grazer performance related to food quality and quantity. Ecology 83:1886-1898

Hessen DO and Lyche A (1991) Interspecific and intraspecific variations in zooplankton element composition. Archiv für Hydrobiologie 121:343-353

Hilbish TJ (1986) Growth trajectories of shell and soft tissue in bivalves: Seasonal variation in Mytilus edulis L. Journal Of Experimental Marine Biology And Ecology 96:103-113

Hochstädter S (2000) Seasonal changes of C:P ratios of seston, bacteria, phytoplankton and zooplankton in a deep, mesotrophic lake. Freshwater Biology 44:453-463

Hwang SJ, Kim HS, Park JH and Kim BH (2011) Shift in nutrient and plankton community in eutrophic lake following introduction of a freshwater bivalve. Journal of Environmental Biology 32:227-234

IGKB (2009) Jahresbericht der Internationalen Gewässerschutzkommission für den Bodensee: Limnologischer Zustand des Bodensees, Bericht Nr. 37.

IGKB (2010) Jahresbericht der Internationalen Gewässerschutzkommission für den Bodensee: Limnologischer Zustand des Bodensees, Bericht Nr. 38.

Jöhnk KD, Huisman J, Sharples J, Sommeijer B, Visser PM and Stroom JM (2008) Summer heatwaves promote blooms of harmful cyanobacteria. Global Change Biology 14:495- 512

107 Bibliography

Jones CG, Lawton JH and Shachak M (1994) Organisms as ecosystem engineers. Oikos 69:373-386

Jüttner F, Leonhardt J and Mohren S (1983) Environmental factors affecting the formation of mesityloxide, dimethylallylic alcohol and other volatile compounds excreted by Anabaena cylindrica. Journal of General Microbiology 129:407-412

Karatayev AY, Burlakova LE and Padilla DK (1997) The effects of Dreissena polymorpha (Pallas) invasion on aquatic communities in eastern Europe. Journal of Shellfish Research 16:187-203

Karatayev AY, Burlakova LE and Padilla DK (2005) Contrasting disrtibution and impacts of two freshwater exotic suspension-feeders, Dreissena polymorpha and Corbicula fluminea. In: Dame R, Olenin S (eds) The comparative roles of suspension-feeders in ecosystems, vol 47. Nato Science Series: IV: Earth and Environmental Sciences, pp 239-262

Karlson AML, Nascimento FJA and Elmgren R (2008) Incorporation and burial of carbon from settling cyanobacterial blooms by deposit-feeding macrofauna. Limnology and Oceanography 53:2754-2758

Kinzelbach R (1991) Die Körbchenmuscheln Corbicula fluminalis, Corbicula fluminea und Corbicula fluviatilis in Europa (Bivalvia:Corbiculidae). Mainzer Naturw. Archiv 29:215- 228

Klein Breteler WCM, Schogt N, Baas M, Schouten S and Kraay GW (1999) Trophic upgrading of food quality by protozoans enhancing copepod growth: role of essential lipids. Marine Biology 135:191-198

Knauer J, Barrett SM, Volkman JK and Southgate PC (1999) Assimilation of dietary phytosterols by Pacific oyster Crassostrea gigas spat. Aquaculture Nutrition 5:257- 266

Kolanowski W, Stolyhwo A and Grabowski M (2007) Fatty acid composition of selected fresh water gammarids (Amphipoda, Crustacea): A potentially innovative source of omega- 3 LC PUFA. Journal of the American Oil Chemists Society 84:827-833

Kraemer LR and Galloway ML (1986) Larval development of Corbicula fluminea (Müller) (Bivalvia, Corbiculacea) - An appraisal of its heterochrony. American Malacological Bulletin 4:61-79

Lampert W (1978) A field study on the dependence of the fecundity of Daphnia spec. on food concentration. Oecologia 36:363-369

Lampert W (1987) Laboratory studies on zooplankton-cyanobacteria interactions. New Zealand Journal of Marine and Freshwater Research 21:483 - 490

Lauritsen DD (1986) Filter-feeding in Corbicula fluminea and its effect on seston removal. Journal of North American Benthological Society 5:165-172

108 Bibliography

Lauritsen DD and Mozley SC (1989) Nutrient excretion by the Asiatic clam Corbicula fluminea. Journal of the North American Benthological Society 8:134-139

Lee T, Siripattrawan S, Ituarte CF and O Foighil D (2005) Invasion of the clonal clams: Corbicula lineages in the New World. American Malacological Bulletin 20:113-122

Liess A and Hillebrand H (2005) Stoichiometric variation in C : N, C : P, and N : P ratios of littoral benthic invertebrates. Journal of the North American Benthological Society 24:256-269

Lodge DM (2001) Lakes. In: Chapin FS, Sala OE, Huber-Sannwald E (eds) Global biodiversity in a changing environment: scenarios for the 21st century. Springer, New York, pp 277- 313

Lucas A and Beninger PG (1985) The use of physiological condition indicies in marine bivalve aquaculture. Aquaculture 44:187-200

Lynn SG, Kilham SS, Kreeger DA and Interlandi SJ (2000) Effect of nutrient availability on the biochemical and elemental stoichiometry in the freshwater diatom Stephanodiscus minutulus (Bacillariophyceae) Journal of Phycology 36:510-522

MacIsaac HJ, Robbins TC and Lewis MA (2002) Biological invasions of aquatic habitats in Europe and the Great Lakes. Modeling ships' ballast water as invasion threats to the Great Lakes. Canadian Journal of Fisheries and Aquatic Sciences 59:1245-1256

MacIsaac HJ and Rocha R (1995) Effects of suspended clay on zebra mussel (Dreissena polymorpha) feces and pseudofeces production. Archiv für Hydrobiologie 135:53-64

Makhutova ON, Kalachova GS and Gladyshev MI (2003) A comparison of the fatty acid composition of Gammarus lacustris and its food sources from a freshwater reservoir, Bugach, and the saline Lake Shira in Siberia, Russia. Aquatic Ecology 37:159-167

Martin-Creuzburg D, Bec A and Von Elert E (2005a) Trophic upgrading of picocyanobacterial carbon by ciliates for nutrition of Daphnia magna. Aquatic Microbial Ecology 41:271- 280

Martin-Creuzburg D, Bec A and von Elert E (2006) Supplementation with sterols improves food quality of a ciliate for Daphnia magna. Protist 157:477-486

Martin-Creuzburg D, Sperfeld E and Wacker A (2009) Colimitation of a freshwater herbivore by sterols and polyunsaturated fatty acids. Proceedings of the Royal Society B - Biological Sciences 276:1805-1814

Martin-Creuzburg D and Von Elert E (2004) Impact of 10 dietary sterols on growth and reproduction of Daphnia galeata. Journal of Chemical Ecology 30:483-500

Martin-Creuzburg D and Von Elert E (2009) Ecological significance of sterols in aquatic food webs. In: Arts MT, Brett MT, Kainz MJ (eds) Lipids in Aquatic Ecosystems. Springer, New York, pp pp. 43-64

109 Bibliography

Martin-Creuzburg D, Von Elert E and Hoffmann KH (2008) Nutritional constraints at the cyanobacteria-Daphnia magna interface: The role of sterols. Limnology and Oceanography 53:456-468

Martin-Creuzburg D, Wacker A and Basen T (2010) Interactions between limiting nutrients: Consequences for somatic and population growth of Daphnia magna. Limnology and Oceanography 55:2597-2607

Martin-Creuzburg D, Wacker A and von Elert E (2005b) Life history consequences of sterol availability in the aquatic keystone species Daphnia. Oecologia 144:362-372

Martin-Creuzburg D, Wacker A, Ziese C and Kainz M (2012) Dietary lipid quality affects temperature-mediated reaction norms of a freshwater key herbivore. Oecologia 168:901–912

McMahon RF (1979) Response to temperature and hypoxia in the oxygen consumption of the introduced Asiatic freshwater clam Corbicula fluminea (Müller). Comparative Biochemistry and Physiology A - Physiology 63:383-388

McMahon RF (1982) The occurence and spread of the introducwed Asiatic freshwater clam, Corbicula fluminea (Müller), in North-America 1924-1982. Nautilus 96:134-141

McMahon RF (2000) Invasive charakteristics of freshwater bivalve Corbicula fluminea. In: Leach JH, Claudi R (eds) Nonidigenous freshwater organisms: Vectors, Biology and Impacts. Lewis Publishers

McMahon RF (2002) Evolutionary and physiological adaptations of aquatic invasive animals: r selection versus resistance. Canadian Journal of Fisheries and Aquatic Sciences 59:1235-1244

McMahon RF and Bogan AE (2001) Mollusca: Bivalvia. In: Ecology and Classification of North American Freshwater Invertebrates, 2nd Edition, Thorp, J. H. and A. P. Covich (Eds.), Academic Press, Inc., pp. 331-430

Mooij WM, Hulsmann S, Domis LND et al. (2005) The impact of climate change on lakes in the Netherlands: a review. Aquatic Ecology 39:381-400

Morgan DE, Keser M, Swenarton JT and Foertch JF (2003) Population dynamics of the Asiatic clam, Corbicula fluminea (Müller) in the Lower Connecticut River: Establishing a foothold in New England. Journal of Shellfish Research 22:193-203

Mörtl M and Rothhaupt KO (2003) Effects of adult Dreissena polymorpha on settling juveniles and associated macroinvertebrates. International Review of Hydrobiology 88:561-569

Mouthon J (1981) Sur la présence en France et au Portugal de Corbicula (Bivalvia, Corbiculidae) originaire d’Asie. Basteria 45:109-116

Mouthon J (2001) Life cycle and population dynamics of the Asian clam Corbicula fluminea (Bivalvia : Corbiculidae) in the Saone River at Lyon (France). Hydrobiologia 452:109- 119 110 Bibliography

Müller-Navarra D (1995) Evidence that a highly unsaturated fatty-acid limits Daphnia growth in nature. Archiv für Hydrobiologie 132:297-307

Müller-Navarra DC, Brett MT, Liston AM and Goldman CR (2000) A highly unsaturated fatty acid predicts carbon transfer between primary producers and consumers. Nature 403:74-77

Müller-Navarra DC, Brett MT, Park S et al. (2004) Unsaturated fatty acid content in seston and tropho-dynamic coupling in lakes. Nature 427:69-72

Naddafi R, Eklov P and Pettersson K (2009) Stoichiometric constraints do not limit successful invaders: Zebra Mussels in Swedish lakes. Plos One 4

Naddafi R and Pettersson K (2008) Effects of the zebra mussel, an exotic freshwater species, on seston stoichiometry. Limnology and Oceanography 53:1973-1987

Napolitano GE, Ackman RG and Silvaserra MA (1993) Incorporation of dietary sterols by the sea scallop Placopecten magellanicus (Gmelin) fed on microalgae. Marine Biology 117:647-654

Nascimento FJA, Karlson AML and Elmgren R (2008) Settling blooms of filamentous cyanobacteria as food for meiofauna assemblages. Limnology and Oceanography 53:2636-2643

Nascimento FJA, Karlson AML, Naslund J and Gorokhova E (2009) Settling cyanobacterial blooms do not improve growth conditions for soft bottom meiofauna. Journal of Experimental Marine Biology and Ecology 368:138-146

Nelson MM, Mooney BD, Nichols PD and Phleger CF (2001) Lipids of Antarctic Ocean amphipods: food chain interactions and the occurrence of novel biomarkers. Marine Chemistry 73:53-64

Newell RIE (2004) Ecosystem influences of natural and cultivated populations of suspension- feeding bivalve molluscs: A review. Journal of Shellfish Research 23:51-61

Nichols SJ and Garling D (2000) Food-web dynamics and trophic-level interactions in a multispecies community of freshwater unionids. Canadian Journal of Zoology - Revue Canadienne de Zoologie 78:871-882

Nichols SJ, Silverman H, Dietz TH, Lynn JW and Garling DL (2005) Pathways of food uptake in native (Unionidae) and introduced (Corbiculidae and Dreissenidae) freshwater bivalves. Journal of Great Lakes Research 31:87-96

Oliver R and Ganf G (2000) Freshwater blooms. In: Whitton B, Potts M (eds) The ecology of cyanobacteria. Kluwer Academic Publishers, Netherlands, pp 149-194

Ortmann C and Grieshaber MK (2003) Energy metabolism and valve closure behaviour in the Asian clam Corbicula fluminea. Journal of Experimental Biology 206:4167-4178

111 Bibliography

Paerl HW, Hall NS and Calandrino ES (2011) Controlling harmful cyanobacterial blooms in a world experiencing anthropogenic and climatic-induced change. Science of the Total Environment 409:1739-1745

Paerl HW and Huisman J (2008) Climate - Blooms like it hot. Science 320:57-58

Paerl HW and Huisman J (2009) Climate change: a catalyst for global expansion of harmful cyanobacterial blooms. Environmental Microbiology Reports 1:27-37

Park DW, Jo QT, Lim HJ and Veron B (2002) Sterol composition of dark-grown Isochrysis galbana and its implication in the seed production of Pacific oyster, Crassostrea gigas. Journal of Applied Phycology 14:351-355

Payne BS, Lei J, Miller AC and Hubertz ED (1995) Adaptive variation in palp and gill size of the zebra mussel (Dreissena polymorpha) and Asian clam (Corbicula fluminea). Canadian Journal of Fisheries and Aquatic Sciences 52:1130-1134

Phelps HL (1994) The Asiatic clam (Corbicula fluminea) invasion and system-level ecological change in the Potomac River estuary near Washington, Dc. Estuaries 17:614-621

Pires LMD, Bontes BM, Van Donk E and Ibelings BW (2005) Grazing on colonial and filamentous, toxic and non-toxic cyanobacteria by the zebra mussel Dreissena polymorpha. Journal of Plankton Research 27:331-339

Pollux B, Minchin D, van der Velde G, van Alen T, Moon-van der Staay SY and Hackstein J (2003) Zebra mussels (Dreissena polymorpha) in Ireland, AFLP-fingerprinting and boat traffic both indicate an origin from Britain. Freshwater Biology 48:1127–1139

Poulin R, Paterson RA, Townsend CR, Tompkins DM and Kelly DW (2011) Biological invasions and the dynamics of endemic diseases in freshwater ecosystems. Freshwater Biology 56:676-688

Prezant RS and Chalermwat K (1984) Flotation of the bivalve Corbicula fluminea as a means of dispersal. Science 225:1491-1493

R Development Core Team (2006) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria

Raikow DF and Hamilton SK (2001) Bivalve diets in a midwestern US stream: A stable isotope enrichment study. Limnology and Oceanography 46:514-522

Rajagopal S, van der Velde G and de Vaate AB (2000) Reproductive biology of the Asiatic clams Corbicula fluminalis and Corbicula fluminea in the river Rhine. Archiv für Hydrobiologie 149:403-420

Reid RGB, McMahon RF, Foighil DO and Finnigan R (1992) Anterior inhalant currents and pedal feeding in bivalves. Veliger 35:93-104

Reitan KI, Rainuzzo JR and Olsen Y (1994) Effect of the nutrient limitation on fatty acid and lipid content of marine microalgae Journal of Phycology 30:972-979

112 Bibliography

Reynolds CS and Walsby AE (1975) Water-blooms. Biological Reviews of the Cambridge Philosophical Society 50:437-481

Richardson DM and Pysek P (2008) Fifty years of invasion ecology - the legacy of Charles Elton. Diversity and Distributions 14:161-168

Rinke K, Huber AMR, Kempke S et al. (2009) Lake-wide distributions of temperature, phytoplankton, zooplankton, and fish in the pelagic zone of a large lake. Limnology and Oceanography 54:1306-1322

Rinke K and Rothhaupt KO (2008) An ecological model of Lake Constance: Model concept, validation and long-term simulations. Wasserwirtschaft 98:26-30

Rinke K, Yeates P and Rothhaupt KO (2010) A simulation study of the feedback of phytoplankton on thermal structure via light extinction. Freshwater Biology 55:1674- 1693

Roditi HA, Strayer DL and Findlay SEG (1997) Characteristics of zebra mussel (Dreissena polymorpha) biodeposits in a tidal freshwater estuary. Archiv für Hydrobiologie 140:207-219

Sala OE, Chapin FS, Armesto JJ et al. (2000) Biodiversity - Global biodiversity scenarios for the year 2100. Science 287:1770-1774

Schmidlin S and Baur B (2007) Distribution and substrate preference of the invasive clam Corbicula fluminea in the river Rhine in the region of Basel (Switzerland, Germany, France). Aquatic Sciences 69:153-161

Schöll F (2000) Die Wassertemperatur als verbreitungsregulierender Faktor von Corbicula fluminea (O.F. MÜLLER 1774). Hydrologie und Wasserbewirtschaftung 44:318–321

Schwarzenberger A, Zitt A, Kroth P, Mueller S and Von Elert E (2010) Gene expression and activity of digestive proteases in Daphnia: effects of cyanobacterial protease inhibitors. BMC Physiology 10:6

Siessegger B (1969) Vorkommen und Verbreitung von "Dreissena polymorpha PALLAS" im Bodensee. gwf-Wasser/Abwasser 110:414 – 415

Sommer U (1985) Seasonal succession of phytoplankton in Lake Constance. Bioscience 35:351-357

Sommer U, Gaedke U and Schweizer A (1993) The 1st decade of oligotrophication of Lake Constance. 2. The response of phytoplankton taxonomic composition. Oecologia 93:276-284

Soudant P, Le Coz JR, Marty Y, Moal J, Robert R and Samain JF (1998) Incorporation of microalgae sterols by scallop Pecten maximus (L.) larvae. Comparative Biochemistry and Physiology A - Molecular and Integrative Physiology 119:451-457

Soudant P, Marty Y, Moal J et al. (1996a) Effect of food fatty acid and sterol quality on Pecten maximus gonad composition and reproduction process. Aquaculture 143:361-378 113 Bibliography

Soudant P, Marty Y, Moal J and Samain JF (1996b) Fatty acids and egg quality in great scallop. Aquaculture International 4:191-200

Soudant P, Moal J, Marty Y and Samain JF (1996c) Impact of the quality of dietary fatty acids on metabolism and the composition of polar lipid classes in female gonads of Pecten maximus (L). Journal of Experimental Marine Biology and Ecology 205:149-163

Soudant P, Sanles MV, Quere C et al. (2000) The use of lipid emulsions for sterol supplementation of spat of the Pacific oyster, Crassostrea gigas. Aquaculture 184:315-326

Soudant P, Van Ryckeghem K, Marty Y, Moal J, Samain JF and Sorgeloos P (1999) Comparison of the lipid class and fatty acid composition between a reproductive cycle in nature and a standard hatchery conditioning of the Pacific Oyster Crassostrea gigas. Comparative Biochemistry and Physiology B - Biochemistry & Molecular Biology 123:209-222

Sousa R, Antunes C and Guilhermino L (2008) Ecology of the invasive Asian clam Corbicula fluminea (Müller, 1774) in aquatic ecosystems: an overview. Annales de Limnologie - International Journal of Limnology 44:85-94

Sousa R, Guilhermino L and Antunes C (2005) Molluscan fauna in the freshwater tidal area of the River Minho estuary, NW of Iberian Peninsula. Annales de Limnologie - International Journal of Limnology 41:141-147

Sperfeld E, Martin-Creuzburg D and Wacker A (2012) Multiple resource limitation theory applied to herbivorous consumers: Liebig’s minimum rule vs. interactive co- limitation. Ecology Letters 15:142–150

Stachowicz JJ, Terwin JR, Whitlatch RB and Osman RW (2002) Linking climate change and biological invasions: Ocean warming facilitates nonindigenous species invasions. Proceedings of the National Academy of Sciences of the United States of America 99:15497-15500

Stanczykowska A (1977) Ecology of Dreissena polymorpha (Pall.) (Bivalvia) in lakes. Polish Archives of Hydrobiology 24:481-530

Stanley-Samuelson DW (1994) The biological significance of prostaglandins and related eicosanoids in invertebrates. American Zoologist 34:589-598

Sterner RW (1993) Daphnia growth on varying quality of Scenedesmus - mineral limitation of zooplankton. Ecology 74:2351-2360

Sterner RW (1997) Modelling interactions of food quality and quantity in homeostatic consumers. Freshwater Biology 38:473-481

Sterner RW, Andersen T, Elser JJ et al. (2008) Scale-dependent carbon : nitrogen : phosphorus seston stoichiometry in marine and freshwaters. Limnology and Oceanography 53:1169-1180

114 Bibliography

Sterner RW and Elser JJ (2002) Ecological stoichiometry: the biology of elements from molecules to the biosphere Princeton University Press, New Jersey

Straile D, Johnk K and Rossknecht H (2003) Complex effects of winter warming on the physicochemical characteristics of a deep lake. Limnology and Oceanography 48:1432-1438

Straile D, Kerimoglu O, Peeters F et al. (2010) Effects of a half a millennium winter on a deep lake - a shape of things to come? Global Change Biology 16:2844-2856

Strayer DL (2010) Alien species in fresh waters: ecological effects, interactions with other stressors, and prospects for the future. Freshwater Biology 55:152-174

Strayer DL, Caraco NF, Cole JJ, Findlay S and Pace ML (1999) Transformation of freshwater ecosystems by bivalves - A case study of zebra mussels in the Hudson River. Bioscience 49:19-27

Suikkanen S, Kaartokallio H, Hallfors S, Huttunen M and Laamanen M (2010) Life cycle strategies of bloom-forming, filamentous cyanobacteria in the Baltic Sea. Deep-Sea Research Part II-Topical Studies in Oceanography 57:199-209

Summons RE, Bradley AS, Jahnke LL and Waldbauer JR (2006) Steroids, triterpenoids and molecular oxygen. Philosophical Transactions of the Royal Society B - Biological Sciences 361:951-968

Tilman D, Kilham SS and Kilham P (1982) Phytoplankton community ecology - The role of limiting nutrients. Annual Review of Ecology and Systematics 13:349-372

Tirok K and Gaedke U (2007) The effect of irradiance, vertical mixing and temperature on spring phytoplankton dynamics under climate change: long-term observations and model analysis. Oecologia 150:625-642

Tittizer T, Schöll F, Banning M, Haybach A and Schleuter M (2000) Aquatische Neozoen im Makrozoobenthos der Binnenwasserstraßen Deutschlands. Lauterbornia 39:1–72

Toyama Y, Kita M and Tanaka T (1952) Sterols and other unsaponifiable substances in the fats of shell fishes, Crustacea and Echinodermata.10. Corbisterol. Bulletin of the Chemical Society of Japan 25:355-357

Turner H, Kuiper J, Thew N et al. (1998) Fauna Helvetica2: Atlas der Mollusken der Schweiz und Liechtensteins.

Tyrell T (1999) The relative influences of nitrogen and phosphorus on oceanic primary production. Nature 400:525-531

Underwood AJ (1997) Experiments in ecology: Their logical design and interpretation using analysis of variance. Cambridge University Press

Urabe J and Watanabe Y (1992) Possibility of N-limitation or P-limitation for planktonic cladocerans - an experimental test. Limnology and Oceanography 37:244-251

115 Bibliography

Van Der Velden W and Schwartz AW (1976) Purines, pyrimidines and organic carbon in lake sediments — a comparison of sediments from lakes of varying degrees of eutrophication. Chemical Geology 18:273-284

Van Donk E and Hessen DO (1993) Grazing resistance in nutrient-stressed phytoplankton. Oecologia 93:508-511

Van Donk E, Ianora A and Vos M (2011) Induced defences in marine and freshwater phytoplankton: a review. Hydrobiologia 668:3-19

Vanderploeg HA, Johengen TH and Liebig JR (2009) Feedback between zebra mussel selective feeding and algal composition affects mussel condition: did the regime changer pay a price for its success? Freshwater Biology 54:47-63

Vanderploeg HA, Liebig JR, Carmichael WW et al. (2001) Zebra mussel (Dreissena polymorpha) selective filtration promoted toxic Microcystis blooms in Saginaw Bay (Lake Huron) and Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 58:1208-1221

Vanderploeg HA, Liebig JR and Gluck AA (1996) Evaluation of different phytoplankton for supporting development of zebra mussel larvae (Dreissena polymorpha): The importance of size and polyunsaturated fatty acid content. Journal of Great Lakes Research 22:36-45

Vanni MJ (2002) Nutrient cycling by animals in freshwater ecosystems. Annual Review of Ecology and Systematics 33:341-370

Vaughn CC and Hakenkamp CC (2001) The functional role of burrowing bivalves in freshwater ecosystems. Freshwater Biology 46:1431-1446

Vaughn CC, Nichols SJ and Spooner DE (2008) Community and foodweb ecology of freshwater mussels. Journal of the North American Benthological Society 27:409-423

Vohmann A, Borcherding J, Kureck A, bij de Vaate A, Arndt H and Weitere M (2010) Strong body mass decrease of the invasive clam Corbicula fluminea during summer. Biological Invasions 12:53-64

Volkman JK (2003) Sterols in microorganisms. Applied Microbiology and Biotechnology 60:495-506

Von Elert E (2002) Determination of limiting polyunsaturated fatty acids in Daphnia galeata using a new method to enrich food algae with single fatty acids. Limnology and Oceanography 47:1764-1773

Von Elert E, Martin-Creuzburg D and Le Coz JR (2003) Absence of sterols constrains carbon transfer between cyanobacteria and a freshwater herbivore (Daphnia galeata). Proceedings of the Royal Society of London Series B - Biological Sciences 270:1209- 1214

116 Bibliography

Voogt PA (1975) Investigations of capacity of synthesizing 3beta-sterols in mollusca.13. Biosynthesis and composition of sterols in some bivalves (Anisomyaria). Comparative Biochemistry and Physiology B - Biochemistry & Molecular Biology 50:499-504

Wacker A, Becher P and Von Elert E (2002) Food quality effects of unsaturated fatty acids on larvae of the zebra mussel Dreissena polymorpha. Limnology and Oceanography 47:1242-1248

Wacker A and Von Elert E (2001) Polyunsaturated fatty acids: Evidence for non-substitutable biochemical resources in Daphnia galeata. Ecology 82:2507-2520

Wacker A and Von Elert E (2002) Strong influences of larval diet history on subsequent post- settlement growth in the freshwater mollusc Dreissena polymorpha. Proceedings of the Royal Society of London Series B - Biological Sciences 269:2113-2119

Wacker A and Von Elert E (2003) Food quality controls reproduction of the zebra mussel (Dreissena polymorpha). Oecologia 135:332-338

Wacker A and Von Elert E (2004) Food quality controls egg quality of the zebra mussel Dreissena polymorpha: The role of fatty acids. Limnology and Oceanography 49:1794- 1801

Wagner C and Adrian R (2009) Cyanobacteria dominance: Quantifying the effects of climate change. Limnology and Oceanography 54:2460-2468

Wagner C and Adrian R (2011) Consequences of changes in thermal regime for plankton diversity and trait composition in a polymictic lake: a matter of temporal scale. Freshwater Biology 56:1949–1961

Walther G-R, Roques A, Hulme PE et al. (2009) Alien species in a warmer world: risks and opportunities. Trends in Ecology & Evolution 24:686-693

Walton MJ and Pennock JF (1972) Some studies on biosynthesis of ubiqiunone, isoprenoid alcohols, squalene and sterols by marine invertebrates. Biochemical Journal 127:471- 479

Walz N (1978) The energy balance of the freshwater mussel Dreissena ploymorpha PALLAS in labratory experiments and in Lake Constance. I. Pattern of activity, feeding, and assimilation efficiency. Archiv für Hydrobiologie Supplement 55:83-105

Ward JE, Levinton JS, Shumway SE and Cucci T (1998) Particle sorting in bivalves: in vivo determination of the pallial organs of selection. Marine Biology 131:283-292

Way CM, Hornbach DJ, Millerway CA, Payne BS and Miller AC (1990) Dynamics of filter feeding in Corbicula fluminea (Bivalvia, Corbiculidae). Canadian Journal of Zoology - Revue Canadienne de Zoologie 68:115-120

Weitere M, Vohmann A, Schulz N, Linn C, Dietrich D and Arndt H (2009) Linking environmental warming to the fitness of the invasive clam Corbicula fluminea. Global Change Biology 15:2838-2851

117 Bibliography

Welker M and Walz N (1998) Can mussels control the plankton in rivers? - A planktological approach applying a Lagrangian sampling strategy. Limnology and Oceanography 43:753-762

Werner S (2008) Effects of the invasive Asian clam Corbicula fluminea on the littoral communities of Lake Constance In: Dissertation, Mathematischen- Naturwissenschaftlichen Sektion, Biology University of Constance, Konstanz, p 130

Werner S and Mörtl M (2004) Erstnachweis der Fluss-Körbchenmuschel Corbicula fluminea im Bodensee. Lauterbornia 49:93-97

Werner S and Rothhaupt K-O (2008) Mass mortality of the invasive bivalve Corbicula fluminea induced by a severe low-water event and associated low water temperatures. Hydrobiologia 613:143-150

Werner S and Rothhaupt KO (2007) Effects of the invasive bivalve Corbicula fluminea on settling juveniles and other benthic taxa. Journal of the North American Benthological Society 26:673-680

Williams JD, Warren ML, Cummings KS, Harris JL and Neves RJ (1993) Conservation status of freshwater mussels of the United States and Canada. Fisheries 18:6-22

Wilson AE, Sarnelle O and Tillmanns AR (2006) Effects of cyanobacterial toxicity and morphology on the population growth of freshwater zooplankton: Meta-analyses of laboratory experiments. Limnology and Oceanography 51:1915-1924

Wright DA, SetzlerHamilton EM, Magee JA and Harvey HR (1996) Laboratory culture of zebra (Dreissena polymorpha) and quagga (D. bugensis) mussel larvae using estuarine algae. Journal of Great Lakes Research 22:46-54

Wright DC, Berg LR and Patterson GW (1980) Effect of cultural conditions on the sterols and fatty acids of green algae. Phytochemistry 19:783-785

Yamamuro M and Koike I (1993) Nitrogen metabolism of the filter-feeding bivalve Corbicula japonica and its significance in primary production of a brackish lake in Japan. Limnology and Oceanography 38:997-1007

118 Appendix

Appendix

Chapter 4 Tab. S1 Comparison of linear models with one independent factor or with 2 factors including temperature correlated with clam tissue growth rates.

AIC R² p AIC R² p mean.temp -247.96 0.9365 <0.001 ARA -223.95 0.7889 <0.001 mean.temp*ARA -254.30 0.9621 <0.001 CP -213.27 0.6400 <0.001 mean.temp*CP -248.90 0.9504 <0.001 green -209.72 0.5700 0.0001 mean.temp*green -248.62 0.9497 <0.001 chlA -207.59 0.5218 0.0003 mean.temp*chla -285.07 0.9919 <0.001 POC -206.13 0.4855 0.0006 mean.temp*POC -285.03 0.9918 <0.001 diatom -201.57 0.3538 0.0057 mean.temp*diatom -275.94 0.9872 <0.001 PUFA -193.47 n.s. mean.temp*PUFA -300.11 0.9962 <0.001 tot.ST -192.99 n.s. mean.temp*tot.ST -263.17 0.9757 <0.001 tot.FA -195.28 n.s. mean.temp*tot.FA -285.01 0.9918 <0.001

Chapter 4 Tab. S2 Two-factorial analyses of variances (ANOVA) with different seston parameters on growth rate of Corbicula fluminea

F p ARA 333.13 <0.001 mean.temp 63.31 <0.001 ARA*mean.temp 9.83 0.006

CP 206.31 <0.001 mean.temp 100. 01 <0.001 CP*mean.temp 0.04 0.853 n.s.

green 181.19 <0.001 mean.temp 120.65 <0.001 green*mean.temp 0.01 0.907 n.s.

chlA 1026.53 <0.001 mean.temp 861.27 <0.001 chlA*mean.temp 63.44 <0.001

POC 952.89 <0.001 mean.temp 923.55 <0.001 POC*mean.temp 70.34 <0.001

diatom 440.75 <0.001 mean.temp 740.81 <0.001 diatom*mean.temp 48.12 <0.001

PUFA 129.88 <0.001 mean.temp 4013.23 <0.001 PUFA*mean.temp 12.17 0.003

119 Appendix

Chapter 4 Tab. S3 Composition of Nannochloropsis limnetica used as sole food source for Corbicula fluminea in the laboratory experiments.

Experiment 1 Experiment 2 Experiment 3 Experiment 4 C:N 6.23 ±0.04 6.39 ± 0.36 6.57 ± 0.33 5.98 ± 0.19 C:P 63.38 ± 21.75 52.82 ± 18.51 50.09 ± 7.59 50.99 ± 1.54 total FA (µg mg C-1) 338.95 ± 20.89 344.94 ±27.55 313.86 ± 26.20 323.68 ± 28.16 PUFA (µg mg C-1) 207.60 ± 10.98 205.26 ± 13.51 171.30 ± 5.03 208.37 ± 7.42 EPA (µg mg C-1) 170.72 ± 10.76 169.37 ± 10.08 134.06 ± 0.70 172.26 ± 7.42

120 Appendix

Chapter 4 Fig. S1 Seasonal changes in fatty acid concentrations in seston of Lake Constance sampled at the study site in 2010. Shown are the α-linolenic acid (ALA), arachidonic acid (ARA), docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) (all in µg mg C-1 and µg l-1). The values between sampling dates and between sampled water depths were interpolated linearly. The four experimental periods are accentuated.

121 Appendix

Chapter 4 Fig. S2 Linear model of growth rate of Corbicula fluminea in correlation to lake water temperature and residuals of linear model plotted against PUFA content of Lake Constance seston in 2010

122 Appendix

Chapter 5 Tab. S4 Lipid composition of food sources fed to Gammarus roeselii. The two cyanobacteria (S. elongatus, A. variabilis) and the two eukaryotic algae (S. obliquus, N. limnetica) were either fed directly to G. roeselii or as biodeposition material produced by C. fluminea, i.e. either as ‘feces’ or as ‘pseudofeces’ (PSF). Data are given as means (µg -1 sterol/fatty acid mg C ) ± standard deviations; numbers of samples are provided in the last two lines (n.d. = not detectable). S. elongatus A. variabilis S. obliquus N. limnetica autotrophs feces PSF autotrophs feces PSF autotrophs feces PSF autotrophs feces PSF 18:2 n-6 n.d. n.d. n.d. 25.3 ±7.3 11.7 ±1.8 12.7 ±7.1 27.8 ±16.8 9.2 ±9.7 18.3 ±11.6 11.9 ±2.6 7.2 ±5.9 9.2 ±2.5 18:3 n-6 n.d. n.d. n.d. n.d. n.d. n.d. 1.9 ±2.7 0.8 ±1.9 0.7 ±1.5 n.d. n.d. n.d. 18:3 n-3 n.d. n.d. n.d. 37.4 ±7.1 19.5 ±2.5 21.8 ±12.4 58.3 ±35.5 27.1 ±21.3 39.9 ±24.6 2.7 ±4.1 4.9 ±9.0 2.2 ±4.9 18:4 n-3 n.d. n.d. n.d. n.d. n.d. n.d. 7.6 ±5.7 8.8 ±13.4 5.8 ±4.6 2.3 ±5.7 2.7 ±6.5 0.0 ±0.0 20:4 n-6 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 17.4 ±2.8 7.6 ±6.6 9.6 ±5.6 20:4n-3 n.d. 1.4 ±2.5 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 20:5 n-3 n.d. 1.7 ±2.9 n.d. n.d. n.d. n.d. n.d. n.d. n.d. 134.0 ±15.1 65.2 ±24.1 76.8 ±28.5 22:6n-3 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 4.6 ±5.2 SAFA 52.8 ±6.1 29.9 ±10.6 32.7 ±11.2 39.1 ±6.7 22.2 ±4.0 21.8 ±11.3 34.6 ±15.0 35.1 ±20.2 31.5 ±10.0 69.9 ±22.9 31.9 ±9.5 34.4 ±9.5 MUFA 59.2 ±4.7 33.2 ±12.1 40.2 ±10.2 55.6 ±11.6 28.1 ±7.7 31.1 ±17.2 19.2 ±16.2 9.9 ±6.5 12.5 ±6.7 60.0 ±11.5 22.0 ±9.8 28.6 ±10.7 PUFA n.d. 3.1 ±5.3 n.d. 62.7 ±12.1 31.2 ±4.2 34.4 ±19.4 95.7 ±58.6 45.9 ±35.2 64.7 ±39.9 168.2 ±21.8 87.5 ±35.7 102.4 ±33.6 Cholesterol n.d. 2.5 ±0.6 1.1 ±0.2 n.d. 1.8 ±2.5 1.2 ±1.4 n.d. n.d. n.d. 7.2 ±3.6 8.7 ±6.7 7.6 ±4.7 Fungisterol n.d. n.d. n.d. n.d. n.d. n.d. 2.3 ±1.0 1.3 ±1.3 2.3 ±0.6 n.d. n.d. n.d. Stigmastanol n.d. 1.8 ±0.3 1.2 ±0.3 n.d. 1.1 ±1.5 1.2 ±1.7 n.d. n.d. n.d. n.d. n.d. n.d. Sitosterol n.d. 0.6 ±0.5 0.4 ±0.2 n.d. 0.7 ±1.0 0.8 ±0.8 n.d. n.d. n.d. 1.8 ±0.6 2.1 ±2.2 1.3 ±1.6 Fucosterol n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 1.6 ±1.4 1.0 ±1.7 0.9 ±1.0 Chondrilla-sterol n.d. n.d. n.d. n.d. n.d. n.d. 4.8 ±2.2 4.3 ±4.3 6.8 ±1.7 n.d. n.d. n.d. Schottenol n.d. n.d. n.d. n.d. n.d. n.d. 0.5 ±0.6 0.3 ±0.5 1.0 ±0.5 n.d. n.d. n.d. n (fatty acid) 4 3 4 4 3 4 6 6 5 6 6 5 n (sterols) 4 3 4 4 4 4 6 5 4 4 5 7

123

Record of achievement / Abgrenzung der Eigenleistung

All chapters were exclusively written by me and contribution of the co-authors is mainly based on improvements and modification statements with regard of content.

Chapter 2, 3, 5: Experimental design and results described in these chapters were exclusively performed by me or under my direct supervision.

Chapter 4: I designed the study. Katja Fleckenstein, whom I was supervising during her Diploma thesis, partly conducted growth experiments and data processing.

124

Danksagung

Ich möchte mich bei all denjenigen ganz herzlich danken, die zum Gelingen dieser Arbeit beigetragen haben.

Mein ganz besonderer Dank gilt meinem Betreuer Dr. Dominik Martin-Creuzburg, der durch seinen unermüdlichen Einsatz das Bestmögliche aus mir, meinen Muscheln und meiner Arbeit herausgeholt hat. Danke für die vielen Stunden, die du in mich investiert hast, sei es für hilfreiche Gespräche, umfangreiche Korrekturen oder zahlreiche Kaffees.

Prof. Dr. Karl-Otto Rothhaupt möchte ich für seine wissenschaftlichen Ratschläge, sein Engagement als Gutachter dieser Arbeit und für die Bereitstellung der Arbeit im Rahmen des SFB 454 Bodenseelitoral danken.

Mein Dank gilt auch Petra Merkel für die exzellente Unterstützung bei der Lipidanalytik und für das Messen und vor allem Einpacken der Kohlenstoffproben. Danke auch an Christine Gebauer und Christian Fiek für die vielen Phosphatanalysen, an Martin Wolf für den unermüdlichen Einsatz bei der Konstruktion neuer Aufbauten und Anlagen, an Alfred, Dank dem ich nicht immer mit der Kröte zum Rohrspitz fahren musste und besonders an die Forschungstaucher, die zu jeder Jahreszeit meine Muscheln holen „durften“.

Im Rahmen meiner Dissertation haben viele Studenten zum Gelingen dieses Projektes beigetragen. Besonderer Dank gilt Hans, Katja, Miriam, Nicole, Piet und Sarah.

Meinen Bürokollegen Karsten und Nina möchte ich ebenfalls „Danke“ sagen für unzählige Stunden des Spaßes und der Ablenkung, wenn ich mal eine Auszeit vom Laboralltag brauchte. Ich werde die Zeit und den exzessiven Kaffee- und Kuchengenuss in U219 nie vergessen.

Weiterhin möchte ich mich bei all den Leuten im Limnologischen Institut und im Speziellen bei Almut, Birgit, John, Marieke, Melanie, Rene, Sonja, Stefan und Tine bedanken, die während der ganzen Jahre gute Freunde und Kollegen waren. Danke auch an Jörg, der mich vom ersten bis zum letzten Semester in Konstanz ertragen hat. Vielleicht klappt das ja noch mit unserer gemeinsamen Karriere in Plön .

Zu guter Letzt möchte ich mich ganz herzlich bei meiner Familie bedanken. Meine Eltern haben mich seit jeher in all meinen Vorhaben bestärkt und mir überhaupt erst die Möglichkeit gegeben, meine Interessen durch das Studium der Biologie zu vertiefen. Meinem Bruder Mirko danke ich dafür, dass er mich mit seiner Begeisterung für die Natur schon in frühen Jahren angesteckt hat und auch heute noch in vielen Bereichen ein Vorbild ist. Meiner Frau Rita danke ich für ihre Kraft, ihre aufbauenden Worte zur rechten Zeit und ihre Liebe. Ohne sie wäre ich vermutlich wahnsinnig geworden .

125

Curriculum Vitae

NAME Timo Basen

DATE OF BIRTH 01.12.1981

PLACE OF BIRTH Düren

HIGHER EDUCATION 1992 – 2001 Gymnasium Zitadelle, Jülich

UNIVERSITY EDUCATION 2002 – 2006 Studies of biology at the University of Konstanz, Konstanz

DIPLOMA THESIS 2007 “Der Einfluss von Licht und Nährstoffen auf den Lipidgehalt von Algen und mögliche Konsequenzen für das Wachstum von Daphnia magna“ DOCTORAL THESIS 2007 – 2012 “Nutritional aspects in the invasive freshwater bivalve Corbicula fluminea: The role of essential lipids”

TEACHING 2009 Bachelor thesis Miriam Bauer: “Der Einfluss von Sterolen auf das Wachstum von Süßwassermuscheln“ 2010 Diploma thesis Katja Fleckenstein: „Wachstum der invasiven Süßwassermuschel Corbicula fluminea im Bodensee: Einfluss von Futterqualität und Temperatur“

126

List of publications

Reviewed publications

Martin-Creuzburg, D., Wacker, A., and Basen, T. (2010). Interactions between limiting nutrients: Consequences for somatic and population growth of Daphnia magna. Limnology and Oceanography 55: 2597–2607

Basen T., Martin-Creuzburg D., and Rothhaupt K.-O. (2011). Role of essential lipids in determining food quality for the invasive freshwater clam Corbicula fluminea. Journal of the North American Benthological Society, 30(3):653–664

Basen T., Rothhaupt K.-O. and Martin-Creuzburg D. (2012). Absence of sterols constrains pelagic-benthic coupling between cyanobacteria and an invasive freshwater clam. Oecologia, online early (doi: 10.1007/s00442-012-2294-z)

Manuscripts in preparation

Basen T., Gergs R., Rothhaupt K.-O. and Martin-Creuzburg D. (submitted). Phytoplankton food quality effects on gammarids: benthic-pelagic coupling mediated by an invasive freshwater clam. Canadian Journal of Fisheries and Aquatic Sciences

Basen T., Fleckenstein K. M., Rinke K., Rothhaupt K.-O. and Martin-Creuzburg D. (in prep.) Impact of temperature and seston dynamics on growth and survival of Corbicula fluminea: A field study in Lake Constance

127

Non-per-reviewed publications and scientific presentations

Basen T., Martin-Creuzburg D. and Rothhaupt K.-O. (2007). Einfluss von Licht und Nährstoffen auf den Lipidgehalt von Scenedesmus obliquus und mögliche Konsequenzen für das Wachstum von Daphnia magna. (Jahrestagung der Deutschen Gesellschaft für Limnologie, DGL Münster, poster presentation).

Basen T., Martin-Creuzburg D. and Rothhaupt K.-O. (2008). Wachstum der invasiven Süßwassermuschel Corbicula fluminea: Einfluss essentieller Lipide. (DGL Konstanz, oral presentation).

Basen T., Martin-Creuzburg D., and Rothhaupt K.-O. (2009). Nutrition of the invasive freshwater bivalve Corbicula fluminea: Role of essential lipids. American Society of Limnology and Oceanography, ASLO Nizza, Frankreich (oral presentation).

Basen T., Martin-Creuzburg D., and Rothhaupt K.-O. (2009). Nutrition of the invasive freshwater bivalve Corbicula fluminea: role of essential lipids. Jahrestagung der Deutschen Zoologischen Gesellschaft (Tagung der Deutschen Zooglogischen Gesellschaft, DZG Regensburg, poster presentation).

Basen T., Gergs R. and Martin-Creuzburg D. (2010). Benthisch-pelagische Kopplung durch die invasive Süßwassermuschel Corbicula fluminea. Jahrestagung der Deutschen Gesellschaft für Limnologie (DGL Bayreuth, oral presentation).

Fleckenstein K., Basen T. and Martin-Creuzburg D (2010). Wachstum der invasiven Süßwassermuschel Corbicula fluminea im Bodensee: Einfluss von Temperatur und Futterqualität. Jahrestagung der Deutschen Gesellschaft für Limnologie (DGL Bayreuth, oral presentation).

Basen T., Martin-Creuzburg D., and Rothhaupt K.-O. (2010). Food quality mediated benthic- pelagic coupling by an invasive freshwater bivalve. International Symposium: The Role of Littoral Processes in Lake Ecology, Hegne, Deutschland (oral presentation).

Martin-Creuzburg D., Basen T., and Rothhaupt K.-O. (2010). Absence of sterols constrains pelagic-benthic coupling between cyanobacteria and an invasive freshwater bivalve. International Symposium: The Role of Littoral Processes in Lake Ecology, Hegne, Deutschland (poster presentation).

Basen T., Fleckenstein K. and Martin-Creuzburg D. (2011). Corbicula fluminea im Bodensee. Jahrestagung der Deutschen Gesellschaft für Limnologie (DGL Freising, oral presentation).

128