Introduction

The epilithic biofilm is a complex assemblage of organisms (bacteria, fungi, algae, heterotrophic protozoans and metazoans) embedded in a mucous matrix of exopolymeric substances (EPS) together with entrapped organic and inorganic particles [1,2]. This mat coats any hard submerged substrate, and when enough light is available, microphytes (and their EPS exudates) contribute copiously to its organic content, then fuelling biofilm-dwelling consumers [1,3]. In the Garonne River, this biofilm is deeply impacted by hydrological scenarios, and consequently the development of its fauna and flora depends on its stability [4-6]. Even if diatoms usually dominate the biofilm phototrophic community in terms of biomass, fluctuating contributions of green algae and cyanobacteria can occur [7,8], constraining resource availability for biofilm-dwelling grazers.

Meiobenthic organisms (meiofauna) are, by definition, benthic invertebrates passing through 500 µm meshes but retained on 42 µm meshes [9]. Through its activity (e.g. bioturbation, grazing and excretion), the meiofauna can affect key processes such as primary production, bacterial development and organic matter fluxes [10-15]. A few recent studies address biofilm-dwelling meiofauna feeding habits, highlighting the importance of pelagic to benthic vertical imports through rotifer and ciliate filtration activity [16], or exploring nematode feeding selectivity to budget their implication in phototrophic carbon transfer [17,18].

Hence, understanding meiofauna trophic role is pivotal to disentangle energy flows in many freshwater food webs [19-21], given that their high production to biomass ratios places meiofauna as relevant trophic intermediaries between micro- and macro-organisms [20-22]. However, this issue, and especially the quantitative and qualitative in situ uptake of primary producers by meiofauna still receives little consideration in freshwater systems [10,17,18,22-25].

Rotifers are important contributors to freshwater meiofauna and constitute significant prey for macrobenthic and other meiobenthic invertebrates [26,27]. In the Garonne River, bdelloid rotifers frequently dwell epilithic biofilms [6]. Bdelloid rotifers can consume a wide variety of preys (algae, bacteria and yeasts) either by suspension feeding, scraping or browsing [28]. While some rotifer species are known to feed selectively [26], the feeding behaviour of Bdelloids is poorly documented and in situ data are particularly lacking. For biofilm-dwelling rotifers, this is mostly due to the difficulty of measuring their feeding behaviour in such complex habitats: not only are epilithic biofilms composed of a complex matrix containing a variety of potential food sources, but the mucous nature of the biofilm itself poses practical experimental problems.

High-performance liquid chromatography (HPLC) quantification of biomarker pigments contained in guts is useful to obtain quantitative and qualitative in situ data on the diet of post-mortem isolated taxa of animals on characterised preys (e.g. microphytic groups). To date, this technique is routinely used to detail the feeding behaviour of aquatic invertebrates in planktonic and benthic habitats (e.g. [29,30]). Especially, HPLC gut pigment quantification allows to examine selective microphyte feeding within complex-clustered benthic habitats such as muds or biofilms [18,31].

As stated above, the epilithic biofilm can offer a wide range of potential substrates for the growth of its inhabiting rotifer community. The aim of this study was: (1) to quantify in situ meiobenthic rotifer feeding activity and possible selectivity; (2) to calculate the pressure exerted on the various potential microphytic prey groups as identified by marker pigments, to evaluate the importance of these for the biofilm phototrophic biomass dynamic.

Material and Methods

Results

Biomarker pigments that were detected in biofilm samples were (a) chlorophyll c, fucoxanthin and diadinoxanthin, (b) lutein and chlorophyll b and (c) zeaxanthin and myxoxanthophyll, indicating the presence of diatoms (a), green microalgae (b) and cyanobacteria (c) respectively (Table 1, Figure 1 a). Myxoxanthophyll is a characteristic carotenoid of some filamentous cyanobacteria whereas zeaxanthin can also originate from some green algae [41,42]. In terms of biomass, diatoms strongly dominated biofilm phototrophic community (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Examples of HPLC absorbance chromatograms obtained at 440 nm from field samples.

(A) biofilm, (B): rotifer (946 individuals). 1: chlorophylls c; 2: fucoxanthin-like; 3: pheophorbide a; 4: fucoxanthin; 5: diadinoxanthin; 6: myxoxanthophyll; 7: cis-fucoxanthin; 8: zeaxanthin-like; 9: zeaxanthin; 10: lutein; 11: chlorophyll b; 12: echinenone; 13: chlorophyll a-like; 14: chlorophyll a; 15: pheophytin a; 16: α-carotene; 17: β-carotene 18: echinenone-like1; 19: echinenone-like 2.

https://doi.org/10.1371/journal.pone.0075352.g001

Echinenone and Myxoxanthophyll, both characteristic of cyanobacteria, were present in relative high concentration in bdelloid rotifers (Table 1, Figure 1b). Myxoxanthophyll has not been reported as an animal carotenoid [43], and is characteristic of filamentous cyanobacterial taxa [41,42]. Hence, its presence among rotifer pigments reveals that they fed on biofilm filamentous cyanobacteria. The presence of chlorophylls c, fucoxanthin and diadinoxanthin among rotifer pigments indicates that they also fed on diatoms (Table 1, Figure 1b). Lutein and chlorophyll b were not detected in rotifers suggesting that they did not feed on green microalgae.

Bdelloid fucoxanthin/Chl a-eq and myxoxanthophyll/Chl a-eq ratios (i.e. respective biomass proportions of diatoms and cyanobacteria to total microphytes in the rotifer gut) were compared to fucoxanthin/Chl a-eq and myxoxanthophyll/Chl a-eq ratios in the biofilm to detect any potential selective uptake of diatoms and/or cyanobacteria. Results are displayed in Figure 2: Rotifer fucoxanthin/Chl a-eq was lower than biofilm fucoxanthin/Chl a, showing that diatoms were ingested in lower proportions than their availability in the biofilm. In contrast, rotifer myxoxanthophyll/Chl a-eq was much higher than biofilm myxoxanthophyll/Chl a. This shows that bdelloids consumed cyanobacteria in much higher proportions than their availability within the biofilm matrix. Ingestion rates and the maximum feeding pressure exerted by bdelloid rotifers on various potential sources (total phototrophic biomass, cyanobacteria, diatoms) were estimated assuming that the rotifers fed only within the biofilm (Table 3; see discussion). Ingestion rates and grazing pressure of the biofilm-dwelling rotifer community on cyanobacteria (particularly on filamentous species, Table 3) were high compared with those exerted on diatoms and on the total phototrophic biomass.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Comparison of pigment proportions between rotifer gut contents (rotifers) and their habitat (biofilm).

(A) Fucoxanthin: Chla-eq; (B): Myxoxanthophyll:Chla-eq; Error bars are SD (n = 3). Fucoxanthin (Fuco) and myxoxanthophyll (myxo) are biomarkers for diatoms and filamentous cyanobacteria respectively.

https://doi.org/10.1371/journal.pone.0075352.g002

Between 67–317 fluorescent objects (FO) were detected by CLSM in centremost layers of five bdelloid rotifer individuals (Figure 3A–E). Mean FO area was 14 µm² spanning 1.8–416 µm². Mean and median FO circularity were high: 0.8 and 0.9, respectively—a perfect circle has a circularity of 1. FO distribution from rotifer boundaries was normal (N = 692; χ² = 89, Df = 16, P < 0.001). The area and fluorescence (CTCF) of FO correlated positively with increasing distance from rotifer boundaries (N = 692; area: R = 0.11, P < 0.01; CTCF: R = 0.1, P < 0.01). The circularity of FO correlated negatively with increasing distance from rotifer boundaries (N = 692; R = –0.14, P < 0.001). These trends are echoed by the observation of large body-centred fluorescent rods (i.e. ingested diatom cells: d on Figure 3A). Small round FO were more evenly distributed in rotifer bodies.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Chlorophyll fluorescence (red objects) in five bdelloid rotifer individuals under a confocal laser scanning microscope.

Scale bar is 100 µm. d: ingested diatoms. (A–E) Five randomly selected bdelloid individuals.

https://doi.org/10.1371/journal.pone.0075352.g003

Discussion

It must be envisaged that the presence of myxoxanthophyll in rotifer bodies could result from symbiotic associations between cyanobacteria and the invertebrates. To the best of our knowledge, observations of endosymbiotic cyanobacteria have not been reported yet for rotifers. However, it is known that bdelloids have foreign genes deriving from prokaryotes, parasitic fungi, plants, algae, and protists, by horizontal gene transfer (HGT) [44]. HGT processes in rotifers are still not well understood. Among the hypotheses, it is proposed that the source organisms of HGT could be either ingested prey or endosymbionts [45]. Nevertheless, endosymbiotic cyanobacteria were observed in neotropical snails forming rounded corpuscles in the midgut gland [39] and in temperate sponges [40]. These symbiotic cyanobacteria belong to Chroococcale, Pleurocapsale and Synechococcale orders, including non-filamentous species. However, myxoxantophyll pigment is the signature of filamentous cyanobacteria [42] and the fluorescent objects (FO) observed in bdelloids in our study were all circular (Figure 3), suggesting that myxoxanthophyll in rotifers, did not originate from endosymbiotic cyanobacteria. We also observed a large variation of FO area per individual (up to 231-fold) weakening the hypothesis that FO could be symbiotic cells. A more plausible rationale for FO is that bdelloids possess digestive circular vesicles which can include partly digested material [46]. Within bdelloids, we observed that fringing FO were small and showed the lowest fluorescence. This is consistent with the hypothesis that bdelloids contained digestive vesicles—including partly digested algal material—originating from ingested cyanobacteria and migrating towards rotifer body fringes during digestion.

The maximum feeding pressure exerted by bdelloid rotifers on the total MPB biomass in the biofilm was low, as it was previously observed for nematodes in the same site of the Garonne river [18]. Values for rotifer ingestion rates were however on average 70 fold higher than those of nematodes, resulting in a 14 fold higher grazing pressure exerted on the MPB biomass by rotifers than by nematodes.

It is generally known that meiofauna feed on diatoms [11,18,29,31,47], and the fact that biofilm-dwelling rotifers live in diatom-rich biotopes makes it logic that they use these algae as food [6] as observed in this study. However, as our results showed that the fucoxanthin/Chl a-eq ratio was lower in the animals than in the feeding medium, diatoms were clearly not selected by the bdelloids, on the contrary, they were probably avoided. Previous studies have highlighted that many species of diatoms can produce secondary metabolites (oxylipins) with toxic effects on marine invertebrates such as planktonic copepods (e.g. Calanus helgolandicus) and cladocerans for example 48 but to the best of our knowledge, no data on diatom toxicity on bdelloid rotifers is available. Thus, further investigation is needed to examine the cause of the observed limited ingestion of diatoms: toxic effect or a morphological effect (e.g. size of food particles [26]). Green microalgal biomarker pigments (lutein and chlorophyll b) were not detected in rotifers indicating that they were not used as food source by the rotifers. It can be alternatively envisaged that both diatom and cyanobacterial pigments could have been either strongly degraded during their passage through rotifer guts. However, although chlorophyll b is a labile pigment, lutein and fucoxanthin in contrast, has been often found in pigment gut contents of meiofauna indicating that they are particularly resistant to degradation [29,31,47].

Unexpectedly, our results showed that the relative cyanobacterial biomass (represented by the myxoxantophyll /Chla-eq ratio) in rotifer body was much higher than the proportion of filamentous cyanobacterial biomass in the biofilm [42]. These results show that the rotifers were able to selectively ingest benthic filamentous cyanobacteria from the biofilm, and additionally, that rotifers can exert a high feeding pressure on these filamentous cyanobacteria in comparison with the pressure exerted on diatoms (Table 3). Assuming a relatively low microphytobenthic production of the biofilm at a daily scale (a turnover of 8 and 16 days for grazed and ungrazed stream epilithic biofilms respectively, according to Lamberti & Resh [49]), this grazing activity suggests that rotifer community could regulate microphytobenthic cyanobacteria stocks by selective feeding. Previous studies hypothesized that direct top-down control of MPB biomass by meiofaunal grazing activity is not a primary regulating mechanism. Meiofauna play a secondary role by modifying the potential bottom up controls of MPB through, e.g. bioturbation [6,11]. While observations previously reported for nematodes support this hypothesis [18], our results however, suggest that for rotifers, in contrast, this question merits further investigations. The possibility that biofilm rotifers also fed partially on phytoplankton in the water column should also be kept in mind [16], especially as the Garonne’s high flow rates provide a constant replenishment of phytoplankton passing over the biofilms [50]. However, considering the low Chla concentration measured in the Garonne river at the study site [50], it seems unlikely that the biofilm rotifers obtained a considerable fraction of their food from the water column.

Cyanobacteria as a preferred prey for rotifers is surprising since: 1) diatoms dominated both biofilm MPB, while filamentous cyanobacteria were minor contributors to biofilm MPB biomass and 2) cyanobacteria are generally considered as non-attractive preys because they show poor nutritional quality (low EPA and P contents [51,52]) together with chemical and structural defence against grazing [51-53]. Some marine harpacticoid copepods have been reported to avoid ingesting cyanobacteria in marine superficial sediments [29,31,47]. However, it is known that Brachionus planktonic rotifers are able to ingest some cyanobacteria but their growth and their toxin sensitivity are variable according to the cyanobacterial species fed upon [53,54]. Battaglia & Valencia [55] mention that in an Antarctic lake, HPLC analyses of pigments in the bdelloid rotifer Philodina gregaria Murray 1910, revealed accumulations of myxoxanthophyll and other carotenoids that are characteristic of the cyanobacterial mat on which they feed. However, no rotifer pigment content data were provided by these authors. Gaudes et al. [56] report that the biomass of Diplogaster rivalis (Schultze, 1857), the most common meiobenthic nematode species in cyanobacterial biofilms of the Llobregat river (Spain) was positively correlated to the availability of cyanobacterial biomass, suggesting that these nematodes could consume cyanobacteria. These authors hypothesised that cyanobacterial high protein content could compensate the lack of polyunsaturated fatty acids in grazer diet.

In rivers, benthic rotifers can be highly selective during food collection - size of preys being the most important discriminating factor for selection [26]. The main cyanobacteria present in the studied biofilm are the filamentous taxa Phormidium uncinatum (Gomont, 1890), Lyngbia sp., and Phormidium sp. [57]. They are mat forming cyanobacteria with long unbranched filaments (1–5 × 10–50 µm) organized in fine and smooth layered strata [58]. Unfortunately, knowledge about bdelloid rotifer food-size preferenda is lacking. Nevertheless, even if the biomass contribution of cyanobacteria is minor, filamentous cyanobacterial cells may provide a higher “encounter rate” to biofilm-grazers than diatom cells.

Cyanobacteria produce and release a number of volatile organic compounds (VOC) which may be infochemicals and are responsible for musty odours of water [59,60]. This is the case for the cyanobacterial species present in the studied biofilm: Phormidium uncinatum, Phormidium sp. and Lyngbia sp. produce geosmin, dihydroβ-ionone, β-ionone and 2-methylisoborneol [58,59]. Dihydroβ-ionone is an attractant for the nematode Caenorhabditis elegans (Maupas, 1900), and this cyanobacterial chemotaxis effect differs among nematode species [59]. Chemotaxis of nematodes may affect both their capacity to locate nearby biofilms and to position themselves within a biofilm [61,62]. Some planktonic rotifers are able of chemical recognition of substances (e.g. genus Brachionus) [61,62]. It is plausible that chemotaxis aids biofilm-dwelling rotifers to detect cyanobacteria within the biofilm. Some studies highlight the ingestion of chemically defended plant tissues by invertebrates as a strategy to deter predators. In laboratory, Hyalella azteca (Saussure, 1858), a freshwater amphipod was less preyed upon when its diet was based on a chemically defended macrophyte, Berula erecta (Coville, 1893) [63]. Some laboratory experiments showed that small crustaceans can develop a toxin resistance, mostly after a period of pre-exposition [64]. As some cyanobacteria are well known for their toxic properties (e.g. [65] for Phormidium species), their hypothetical ingestion by benthic rotifers as a strategy to reduce predation deserves further exploration.

Echinenone was detected in the rotifers, but not in the biofilm nor in the phytoplankton samples. Thus, the echinenone detected within bdelloids was not gained through cyanobacteria ingestion. Echinenone is a β-carotene derivative pigment that has been found in many invertebrates (e.g. in molluscs, crustaceans, echinoderms and tunicates [43,66,67]). Nevertheless, bdelloids possess carotenoid biosynthetic genes, originating from other organisms by horizontal gene transfer [44]. It seems that these genes allow production of carotene pigments and derivatives by rotifers [68]. Thus, since β-carotene was also a major pigment found in rotifers and in microphytes (Table 1, Figure 1 a, b), it can be hypothesised that rotifers were able to convert β-carotene to echinenone. Gilchrist & Green have shown that the two major groups of rotifers (Bdelloidea and Ploima) can accumulate carotenoids from their algal food and convert these into other carotenoid pigments (e.g. astaxanthin and astacene) [69]. However, authors did not observe accumulation of echinenone but of astaxanthin and astacene in the Bdelloidea rotifer Philodina roseola (Ehrenberg, 1832), likely because they used a chromatography method which did not allow separating the carotenoids contained in rotifer samples prior to their identification. Considering that astaxanthin and echinenone have a similar spectral shape but different retention times, it is possible that these authors were unable to distinguish the presence of echinenone in their rotifer samples. It has been previously hypothesized that the photoprotection of microalgae, assured by the accumulation of photoprotector carotenoids in their cells, could be transferred to grazers in the form of pigments. In addition, the survival of certain species of meiobenthic copepods is directly related to their carotenoid content when exposed to copper [70]. In this context, our results suggest that the potential link between the accumulation of echinenone and a possible improvement of resistance to metal toxicity and UV light exposures deserves to be tested for bdelloid rotifers.

Conclusion

To conclude, our results showed that biofilm-dwelling rotifers fed on cyanobacteria and diatoms. It is strongly suggested that this feeding was highly selective on cyanobacteria while diatoms were avoided. Our results support the hypothesis that the role of cyanobacteria as food source for the meiobenthos has been underestimated as previously suggested [59]. We feel that both the potential of biofilm-dwelling rotifers to control cyanobacterial development and the idea that cyanobacteria are ingested by the rotifers to obtain predation protecting toxins merit further investigation. Another aspect soliciting further interest is the apparent potential of bdelloid rotifers to convert microphyte-ingested β-carotene to the photoprotective pigment echinenone.

Supporting Information

Acknowledgments

We thank F. Julien and R. Fernandez for field assistance; A. Jauneau and A. Le Ru for their help in CLSM analyses.