Feeding ecology of Limnoperna fortunei in southern China: insights from stable isotopes and fatty-acid biomarkers

ABSTRACT Limnoperna fortunei (L. fortunei) is one of the most abundant freshwater bivalves in southeast Asia, with wide-ranging direct and indirect impacts on ecosystems. To estimate material flows in the habitats of L. fortunei, a combination of stable-isotope and fatty-acid analyses were applied to assess the feeding spectrum of L. fortunei in southern China. Using the isotope-mixing model, the contribution proportions to the diet of L. fortunei were estimated as 19.8%–28.2% for plankton, 57.6%–65.2% for particulate organic matter (POM) and 10.2%–21.1% for sediment organic matter. We conclude that POM is the principal food source of L. fortunei. The δ13C enrichment of fixed carbon from POM to L. fortunei was 0.67%–2.41%. Based on the fatty acid data, it was estimated that L. fortunei consumed or selectively accumulated Chlorophyceae, Cryptophyceae, Dinophyceae, bacteria and terrestrial organic matter. The feeding spectrum of L. fortunei is similar to that of Dreissena polymorpha. We suggest that L. fortunei is able to differentiate suitable food items using chemical cues and the surface properties of particles.


Introduction
The bivalve mussel, Limnoperna fortune (L.fortunei), is a benthic suspension feeder with a widespread distribution in fresh waters of Southeast Asia and South America (Paolucci et al. 2010;Zhang et al. 2014;Zhang et al. 2015). Adults of L.fortunei are able to firmly attach to hard substrata using the byssus (Nishino 2012). The density of L.fortunei is extremely high, reaching 10,000 individuals m ¡2 near the bank of the Xizhijiang River, southern China (Xu et al. 2009). With individual filtration rates up to 350 mL h ¡1 , L.fortunei has a marked effect on suspended and sediment organic matter (SSOM), which modified the nutrient supply (Boltovskoy et al. 2009;Di Fiori et al. 2012). Therefore, L.fortunei provides an important link between SSOM and consumers in fresh water. Consequently, to understand the energy and material flows in the habitats of L.fortunei, it is necessary to clarify the composition of diet of this species.
Physical methods to determine the diet of bivalves include analysis of stomach contents and faecal pellets (Lehane and Davenport 2004). Stomach-contents analysis of L.fortunei living in the Middle Paran a River indicated that the main food ingested comprised plankton, particularly Euglenophyta, Rotifera, Chydoridae and Bosminidae (Molina et al. 2010). However, this method did not identify which dietary items were assimilated and can provide only rough estimates of recent feeding activity (seconds to hours) (Kang et al. 1999). More recently, biochemical methods, such as the analysis of stable isotopes or fatty acids, have been used with considerable success in determining the diet of bivalves (K€ urten et al. 2013;Najdek et al. 2013;Braeckman et al. 2015). Stable-isotope analysis is a useful approach for identifying the food sources assimilated by bivalve mussels (Vaughn and Hakenkamp 2001). Consumers typically contain more heavy isotopes than is present in their diets, with stepwise heavy-isotope enrichment occurring at each trophic level (Peterson and Fry 1987). The ratio of 13 C/ 12 C (d 13 C) is preferentially used as a dietary tracer because there is little fractionation (ca. 1.0%) between the consumer and its food sources (Asante et al. 2010;Sun et al. 2012). In contrast, the method of using fatty acids as dietary tracers in aquatic ecosystems relies on the observation that different primary producers (e.g. diatoms, dinoflagellates, bacteria) generally have specific fatty-acid profiles in their biomass. Furthermore, these fatty acid profiles tend to be transferred conservatively and thus are present in consumers' tissues in an unchanged form (Shin et al. 2008). Therefore, it is possible to distinguish dietary and non-dietary compounds in the body of a consumer via fatty-acid analysis. These new approaches are therefore able to improve the accuracy of information on the diet of L.fortunei. Such information is essential for understanding energy and material flows in the habitats in which it occurs.
Bivalves are considered to be herbivores and it is assumed that plankton, particulate organic matter (POM) and sediment organic matter (SOM) are the main components of their diet (Molina et al. 2010;Zhao et al. 2013). In this work, we used a combination of stable-isotope and fatty-acid analyses to evaluate the contribution of plankton, POM and SOM to the diet of L.fortunei. Using these data, we assessed the feeding spectrum of L.fortunei in the Xijiang River, southern China.

Sample collection and preparation
Triplicate samples of the plankton, POM and SOM were collected monthly between March and November 2013. The sampling site (23 08 0 12 00 N, 112 48 0 7 00 E) ( Figure 1) is locate at the Xijiang River Figure 1. Sampling site (23 08 0 12 00 N, 112 48 0 7 00 E) locating at the Xijiang River is about 140 kilometres from seaport. The Xijiang River cover a full distance of over 2214 kilometres.
(south China), and its riverbed is rock based. The water quality of sampling site belong to Class I of the Surface water Environmental Quality Standard (2002). Samples of plankton were collected with a net of 60 mm mesh by repeated multiple horizontal short duration tows (5-10 min) carried out between 2-8 m depth. About 30 L of fresh water were collected and stored in acid-cleaned polyethylene bottles to obtain samples of POM. SOM samples were collected at depths of 8-10 m via the sediment corer (300 mm in diameter, PC-300, Mooring Systems Inc., Cataumet, MA). Adult L.fortunei (22 mm) were collected from the riverbed at same site using SCUBA at depths of 2-10 m. All samples were immediately stored at 4 C before being transported to the laboratory.
In the laboratory, samples of plankton were freeze-dried, ground into powder using a pestle and mortar, and stored in acid-cleaned polyethylene bags at ¡80 C. For samples of POM, water samples were pre-sieved through a 200 mm mesh net to remove large particles and then filtered using Whatman GF/F glass fibre (pre-combusted at 550 C for 5 h). The filter papers were rinsed with ultrapure water, freeze-dried and stored in acid-cleaned polyethylene bags at ¡80 C. Samples of SOM were freeze-dried, sieved through a 300 mm stainless steel screen, homogenised and stored in acid-cleaned polyethylene bags at ¡80 C. After depuration in filtered water for 24 h, samples of L.fortunei tissues were dissected with a plastic knife and rinsed with ultrapure water.
For each analysis, 10 individuals of L.fortunei were pooled, freeze-dried, homogenised, and stored in acid-cleaned polyethylene bags at ¡80 C.

Stable-isotope analyses
For stable-isotope measurements, plankton, POM and SOM samples were acidified with 10% HCl, rinsed with distilled water and oven-dried at 40 C for 24 h, to remove carbonates (Deniro and Epstein 1978). When production of CO 2 bubbles ceased, the samples were dried and stored in acidcleaned polyethylene bags. For analysis, about 1 mg of the powdered sample was packed into a 4 £ 6-mm tin capsule.
Samples were combusted in an elemental analyser (Vario MICRO cube, Elementar Analysensysteme GmbH, Lagensebold, Germany) attached to an isotope-ratio mass spectrometer (MAT 253, Thermo Fisher Scientific, Waltham, MA) to determine 13 C / 12 C. The value of d 13 C was expressed as the deviation from a standard in parts per thousand (%) according to the following equation: d 13 C = [(R sample /R standard )¡1] £ 1000 where R is the corresponding ratio of 13 C / 12 C. Carbon values were referenced to standard Pee Dee Belemnite (PDB). Measurements were made with a precision of approximately 0.2%.

Fatty acid analyses
Lipids were extracted from plankton, POM, SOM and L.fortunei samples following the method of Folch et al. (1957) and Zhao et al. (2013). Lipids were extracted ultrasonically for 10 min using a solvent mixture (two parts chloroform to one part methanol). The lower chloroform phase containing lipids was collected and separated further by centrifugation. The lipid extracts were saponified, transmethylated, separated and purified to transform fatty acids to fatty-acid methyl esters (FAMEs). FAMEs were analysed using a gas chromatograph (GC-9A; Shimadzu, Tokyo, Japan) on a DB-FFAP capillary column (30 m £ 0.32 mm internal diameter, 0.25 mm film). Hydrogen was used as the carrier gas. The injector temperature was 250 C. FAMEs were identified by comparing their retention times with those of standards.

Data analyses
Statistical analyses were performed using SPSS software (Version 19.0; SPSS, Chicago, IL). Significant differences (P < 0.05) in d 13 C were tested using Student's t-test and one-way analysis of variance (ANOVA). For fatty acid data, means, standard errors (SE) and Fisher's least significant difference (LSD) post-hoc tests were calculated conventionally.
To evaluate the relative contributions of plankton, POM and SOM to the diet of L.fortunei, the isotope-mixing model (Philips 2001) was used with slight modifications. The fractionation of d 13 C values for bivalves was set at 0.8 in the model (Fukumori et al. 2008), defined as follows: d 13 C Lf = f pl (d 13 C pl +0.8)+f p (d 13 C p +0.8)+f s (d 13 C s +0.8) where the subscripts Lf, pl, P and S refer to L.fortunei, plankton, POM and SOM, respectively, and f pl, f p and f s are the fractional contributions of plankton, POM and SOM, respectively.
Fatty acids that are commonly used as biochemical markers for particular taxonomic groups that occurred in our study are shown in Table 1.

Characteristics of stable carbon isotopes in plankton, POM, SOM and L.fortunei
Mean d 13 C values for plankton, POM, SOM and L.fortunei tissue sample are shown in Table 2. Significant differences in d 13 C values were observed among plankton, POM, SOM and L.fortunei samples (one-way ANOVA, P < 0.05). The d 13 C values varied between ¡23.38% and ¡20.68% for plankton, between ¡23.52% and ¡20.99% for POM, and between ¡26.72% and ¡19.30% for SOM. The d 13 C values for L.fortunei ranged from ¡21.62% to ¡19.91%.

Contribution of plankton, POM and SOM to the dietary regime of L.fortunei
The relative contributions of plankton, POM and SOM in the diet of L.fortunei, calculated using the isotope-mixing model, are presented in Figure 2. The contribution of POM to the carbon content of L.fortunei ranged between 57.6% and 65.2%, which was significantly higher than that of plankton (19.8%-28.2%) and SOM (10.2%-21.1%) (t-tests, P < 0.05). Table 1. Fatty acids as biochemical markers of certain taxonomic groups that occur in the Xijiang River, southern China. SFAs, MUFAs, PUFAs and BrFAs refer to saturated fatty acids, monounsaturated fatty acids, polyunsaturated fatty acids and branched fatty acids, respectively. DHA refers tor C22:6(n-3) and EPA refers to C20:5(n-3).

Discussion
In the present study, the d 13 C signatures of plankton, POM and SOM were significantly different. These distinct signatures made it possible to calculate the relative contributions of plankton, POM and SOM to the diet of L.fortunei using a mixing model. These calculations indicated that L.fortunei primarily feeds on POM, followed by plankton and SOM (Figure 2). Previous studies on the feeding ecology of bivalves indicated that Pinctada fucata martensii obtained 78% (Kanaya et al. 2005) and Ruditapes philippinarum 61.0% (Fukumori et al. 2008) of their carbon from POM in their natural habitats. Our results are consistent with their conclusion that POM is the principal food source of bivalves. The stable-isotope approach assumes a fixed isotopic enrichment between the bivalve and its food items. Ruditapes philippinarum, Mactra veneriformis and Nihonotrypaea japonica were reported to be enriched by 0.6%-2.0% for d 13 C relative to POM (Yokoyama et al. 2005). Therefore, our observation of the d 13 C enrichment of L.fortunei relative to POM (0.67%-2.41%) ( Table 1) is consistent with those findings. However, plankton, POM and SOM are heterogeneous mixtures of phytoplankton, bacteria, benthic microalgae and other OM (Dalsgaard et al. 2003). The isotopic signatures of these sources often overlap in natural conditions, making it difficult to separate specific components (Phillips and Gregg 2003). Primary producers, such as diatoms, dinoflagellates and bacteria, are characterised by distinct fatty-acid profiles (Kharlamenko et al. 2001). Therefore, these profiles can be used to identify the relative contribution of each component to the plankton, POM and SOM mixtures.
According to the fatty acid profiles of L.fortunei food items (Figure 3), plankton, POM and SOM differ in the percentages of several fatty acid markers, indicating high levels Chlorophyceae, Cryptophyceae, Bacillariophyceae and cyanobacteria in plankton, high levels of plant detritus in POM, and high levels of heterotrophic bacteria and Copepoda in SOM. Although the dominant taxa in the phytoplankton were Chlorophyceae, Cryptophyceae, Dinophyceae and Bacillariophyceae, fatty acid markers of these algae were not abundant in plankton, and were particularly scarce in POM. The relatively low proportion of algal biomass in the water column is a possible explanation of this result. The higher percentages of SFAs (C18:0, C20:0 and C22:0) and BrFAs in POM and SOM suggest that they contain high proportions of detrital particles derived from plant debris, pseudofaeces, faeces and other molluscan excreta .
The fatty acid compositions of various bivalve taxa are highly variable. As in other freshwater bivalve, e.g. D. polymorpha and D. bugensis , L.fortunei possesses high levels of C20:5(n-3) and C22:6(n-3), which are considered physiologically crucial and probably are conservatively retained in tissues relative to other compounds Kelly and Scheibling 2012). In addition, many researchers emphasised the importance of the ratio C22:6(n-3)/C20:4(n-6) for the growth and reproduction of the zoobenthos (Ahlfren et al. 2009). Data collected from the literature ) indicated values for the C22:6(n-3)/C20:4(n-6) ratio in D. polymorpha and D. bugensis of 1.44 and 1.49, respectively, and in Potamocorbula amurensis the ratio was ca. 2 (Canuel et al. 1995). In the present study, the ratio in L. fortunei was somewhat higher (2.02; Table 3).
Algal fatty acids are used as energetic resources and are catabolised in animal tissues (Brett and Goldman 2006;Gladyshev et al. 2011). The percentage of C18:2(n-6) (Figure 3a), an essential fatty acid synthesised by algae (Maria-Jos e et al. 2010), was higher in L.fortunei than in its food items. Likewise, percentages of the essential fatty acid, C18:4(n-3), which was abundant in Cryptophyceae and Dinophyceae (Gugger et al. 2002;Maria-Jos e et al. 2010), was also higher in L.fortunei (Table 3). Apparently, L.fortunei consumed or selectively accumulated C18:2(n-6) and C18:4(n-3) from certain algae species, i.e. Chlorophyceae, Cryptophyceae and Dinophyceae. In contrast, Bacillariophyceae are abundant in plankton, POM and SOM, but are deficient in L.fortunei (Figure 3(d)). Nevertheless, we could not estimate the consumption of Bacillariophyceae by L.fortunei by using fatty-acid markers because bivalves are reported to preferentially store fatty acids from Bacillariophyceae as triacylglycerols and to use them for catabolism ). Heterotrophic bacterial fatty acids are reported to be a significant source of organic carbon and nitrogen for some bivalve molluscs (Nichols and Garling 2000). Bacterial taxa differ markedly in fatty markers (Napolitano 1999), e.g. bacterial i17:0 and 17:0 for sulfate-reducing bacteria, and C18:1(n-7) for cyanobacteria, sulphur-oxidising and aerobic bacteria. Thus, L.fortunei in the present paper apparently consumed the latter bacteria, which usually dwell in the water column. Decomposed material, particles of plant detritus and copepods can also be potential food sources for L.fortunei but the high percentage of their fatty acid markers in plankton, POM and SOM, and low levels in L.fortunei indicated that L.fortunei did not prefer these food items. The high level of the biomarkers of terrestrial OM in L.fortunei (Table 3) suggest the indirect assimilation of terrestrial OM via heterotrophic bacteria, but could also result from the direct assimilation of terrestrial OM using cellulase and hemicellulase in this species (Mcleod and Wing 2009).
According to the fatty-acid marker analyses, L.fortunei preferred planktonic algae and bacteria. This feeding spectrum is similar to that of D. polymorpha (Cole and Solomon 2012;Makhutova et al. 2013) and could be explained in terms of active selection, more efficient assimilation of the selected diet from POM and/or preferential ingestion. Although some bivalve species appear to capture their diet indiscriminately (Ward et al. 1997), L.fortunei is able to differentiate between suitable and unsuitable particles by gill sorting mechanisms. This selection mechanism appears not to be based on particle size but on chemical cues and their surface properties (Wong and Cheung 1999).

Conclusions
The feeding spectrum of L.fortunei in the Xijiang River in summer was described. Using the isotopemixing model, the relative contributions to the diet of L.fortunei were estimated as 19.8%-28.2% for plankton, 57.6%-65.2% for POM and 10.2%-21.1% for SOM. Fatty acid biomarkers specific to Chlorophyceae, Cryptophyceae, Dinophyceae, heterotrophic bacteria and terrestrial OM were identified in the tissues of L.fortunei, indicating that there were substantial algal, bacterial and terrestrial inputs into the diet of L.fortunei. However, the present study did not include the winter months. The possibility of seasonal variation in the levels of plankton, POM and SOM in the water column should be addressed in future research.
Bin Cui is an associate professor of the school of civil engineering, Guangzhou University. She received doctoral degree from South China University of Technology. Her research focuses on L.fortunei control and water pollution control.
Zhimin Sun is a professor of the Guangzhou Municipal Engineering Design & Research Institute. He received doctoral degree from the Harbin Institute of Technolog. His research focuses on water pollution control.
Wuyang Zhou received his MSc degree from the school of civil engineering, Guangzhou University. His research focuses on water pollution control.
Pengfei Ren received doctoral degree from the Harbin Institute of Technolog. His research focuses on water pollution control.
Zhili Du received his MSc degree from the Illinois Institute of Technology (iit). His research focuses on water pollution control.