Iola G. Boëchat and Rita Adrian
(FEMS – Microbiology Ecology, under revision)
Keywords: Heterotrophic protists; Biochemical composition; Fatty acids; Amino acids; Dietary composition
The focus of our study was to determine whether the biochemical composition of heterotrophic protists resembles that of their diet. Carbon- and cell-specific concentrations of fatty acids and essential amino acids were investigated for two ciliates (Balanionplanctonicum, Urotrichafarcta) grown on algal diet (the cryptomonad Cryptomonas phaseolus), and a ciliate and a flagellate (Cyclidium sp. and Chilomonasparamecium) grown on a mixed diet consisting of bacteria and small rice particles. Stepwise discriminant analyses (SDA) indicated differences in the fatty acid and amino acid composition between heterotrophic protists and their diet, as well as among protist species. Carbon-specific fatty acid and amino acid concentrations were usually higher in the heterotrophic protists than in their diet. B. planctonicum and U. farcta showed higher concentrations of monounsaturated and some polyunsaturated fatty acids than their algal diet. Moreover, except for tryptophan, valine, and lysine, higher carbon-specific amino acid concentrations were observed in both B. planctonicum and U. farcta than in Cryptomonas. Cyclidium sp. and C.paramecium had higher carbon-specific concentrations of polyunsaturated fatty acids and amino acids than their diet, except for histidine, methionine, and leucine. Cellular-specific fatty acid concentrations were generally higher in the protists fed the algae than in the protists fed bacteria and rice particles, while cellular-specific amino acid concentrations were similar among protists. The higher fatty acid and amino acid concentrations in the heterotrophic protists compared to their diet may suggest that these species are capable of efficiently assimilating or even synthesising biochemical compounds.We conclude that dietary fatty acid and amino acid composition influences the composition of the four freshwater protist species to a minor extent, and that species-specific differences in fatty acid and amino acid metabolism are more important determinants of the biochemical composition in the analysed heterotrophic protists.
Aquatic food webs rely on autotrophs as these are the most important resource of essential chemical components for the mesozooplankton. Heterotrophic protists on the other hand are an important trophic link in aquatic food webs, as they prey on primary producers and bacteria, and are themselves preyed upon by the mesozooplankton (De Biase et al., 1990; Wickham et al., 1993; Sanders et al., 1996; Adrian and Schneider-Olt, 1999; Adrian et al., 2001; Burns and Schallenberg, 2001; Mohr and Adrian 2002a). Depending on their ability to assimilate and incorporate chemical compounds obtained from their diet, heterotrophic protists have the potential to modify the chemical composition of organic matter at an early stage in the food chain.
In zooplankton species, the fatty acid composition of a predator usually resembles that of its prey (Ederington et al., 1995; Fernández-Reiriz and Labarta, 1996; Von Elert and Stampfl, 2000). Daphnia galeata exhibited higher concentrations of the polyunsaturated fatty acids EPA (eicosapenenoic acid) and DHA (docosahexaenoic acid) when fed a cryptomonad than D. galeata fed green algae (Weers et al., 1997). Similarly, Acartia tonsa exhibited higher concentrations of saturated and monounsaturated fatty acids when fed a bacterivorous ciliate and higher concentrations of polyunsaturated fatty acids when fed diatoms (Ederington et al., 1995). A study involving two marine ciliates (Harvey et al., 1997) showed that the lipid composition (fatty acids, neutral lipids, and sterols) of the ciliates resembled that of their prey (either bacteria or algae). Assuming that these findings are also valid for freshwater heterotrophic protists, one would expect differences in the chemical make-up of heterotrophic protists fed different diets. Indeed, bacterivorous protists contained large amounts of saturated (SAFA) and monounsaturated (MUFA) fatty acids (together comprising more than 85% of the total fatty acids), and very low quantities of polyunsaturated fatty acids (PUFA), and highly unsaturated fatty acids (HUFA) (Ederington et al., 1995). In contrast, algivorous ciliates are expected to contain high quantities of PUFA, HUFA, and some essential amino acids (Desvilettes et al., 1997). In algivorous marine ciliates, SAFA and PUFA comprised 32% and 57% of the total fatty acids (Klein Breteler et al., 1999). The amino acid composition of mesozooplankton predators seems to be rather constant, and more or less independent of the amino acid composition of their diet (Frolov et al., 1991; Cowie and Hedges, 1994; Guisande et al., 1999; Guisande et al., 2000; Helland et al., 2003a,b). Except for [page 23↓]studies on Tetrahymena (Holz, 1973), the amino acid composition of freshwater heterotrophic protists is virtually unknown.
Different protist species may show different metabolic features. To date little is known about species-specific differences in fatty acid and amino acid metabolism of freshwater heterotrophic protist species. Such differences are expected to occur and may contribute to the transfer of essential compounds between primary producers and higher level consumers in aquatic food webs. In addition to the assimilation and incorporation of dietary compounds, de novo synthesis of some essential compounds has been observed in ciliates and flagellates. The bacterivorous ciliate Pleuronema sp. was found to contain high concentrations of a triterpenoid alcohol (tetrahymanol), as a major neutral lipid, which was not observed in its bacterial prey (Ederington et al., 1995). Also the ability of the heterotrophic dinoflagellate Oxyrrhis marina to synthesize EPA, DHA, and some sterols has already been described (Klein Breteler et al., 1999).
We now address the question of how the biochemical profile (fatty acids and essential amino acids) of freshwater heterotrophic protists reflects that of their diet. Furthermore, we looked for differences in the biochemical composition between protist species fed the same diet. To elucidate the relationship between the biochemical composition of heterotrophic protists and their diet, we compare the fatty acid and essential amino acid composition of four protist species with the fatty acid and essential amino acid composition of their diet. The ciliates Balanion planctonicum and Urotricha farcta were cultured on the cryptomonad Cryptomonas phaseolus and the ciliate Cyclidium sp. and the flagellate Chilomonas paramecium were cultured on a mixed diet consisting of bacteria and small rice particles. Our study examines the fate of the biochemical compounds, which confer food quality to heterotrophic protists and emphasizes species-specific differences in protist ability to transform energy and organic matter at an early stage in the aquatic food web, when the disparity in the biochemical composition may be large.
The ciliates Balanion planctonicum WULLF, 1922 (average biovolume 3256 ± 1331 μm3) and Urotricha farcta CLAPARÈDE and LACHMANN, 1858 (2778 ± 1707 μm3) were cultured in WC medium (Guillard and Lorenzen, 1972) in frequently diluted batch cultures incubated at 17°C under a 12:12 h light : dark regime. The ciliates were fed the cryptomonad Cryptomonas phaseolus EHRENBERG, 1832 (392 ± 125 μm3), obtained from the Algal Collection of the University of Göttingen, Germany, and cultured in WC medium at 17 ± 1°C under a 16:8 h light : dark regime. Our cultures were non axenic, but bacteria accounted for less than 2 % of the total organic carbon contents in the cultures of the protists and C. phaseolus. Although these ciliates are also known to ingest bacteria, especially U. farcta (Foissner et al., 1999), we consider these ciliates as algivores, as they were mainly fed an algal prey. According to microscopic observations, they indeed preyed upon the algae, which were reduced to less than 10% of the total carbon content of the cultures, usually within a five days growth period. The flagellate Chilomonas paramecium EHRENBERG, 1832 (403 ± 288 μm3) and the ciliate Cyclidium sp. (1315 ± 617 μm3) were cultured in Volvic water (a spring water, poor in minerals, sold worldwide by Société des Eaux de Volvic, Puy-de-Dôme, France) at 18 ± 1°C under a 16:8 h light : dark regime. Both were cultured on a mix of bacteria grown on previously autoclaved polished rice corns (referred to as “bacteria + rice” diet, since the protists might have been able to ingest rice particles along with bacteria in the medium). Although the contribution of small rice particles to the total carbon biomass in the cultures was small compared to bacterial biomass (based on microscopic observations), the influence of the biochemical composition of rice should not be ignored, and it has to be considered when discussing the results. Henceforth, species are referred to by their genus names only.
Samples of heterotrophic protists and the algal diet were taken from the cultures, fixed with Lugol’s solution and counted under a stereomicroscope at x 20 magnification. For bacteria enumeration, 1 mL samples were fixed in 2% formaldehyde, filtered onto 0.2 μm pore size black polycarbonate membranes, stained with DAPI (4,6–diamidino–2–phenylindol) (Porter and Feig, 1980) and counted under an epifluorescence microscope [page 25↓](excitation wavelength 450–490 nm combined with an FT 510 beam splitter and a LP 520 suppression filter). Carbon concentrations in heterotrophic protists and their diet were derived from cell volume estimates using carbon : biovolume conversion factors of 0.10 pgC μm-3 for the alga Cryptomonas(Montagnes et al., 1994), 0.19 pgC μm-3 for heterotrophic protists (Putt and Stoecker, 1989), and 0.125 pgC μm-3 for the bacteria (Pelegrí et al., 1999). Cell biovolumes of the protists, algae, and bacteria were estimated based on commonly used geometric forms (sphere, cylinder and ellipsoid). Unfortunately, we were not able to separate the algivorous ciliates from their algal diet via filtration or centrifugation due to their similar cell dimensions – an inherent problem in working with micrograzers. To minimize this problem we collected samples for the biochemical analyses at the end of the exponential growth phase, when the algivorous ciliates had diminished Cryptomonas to 4–16% of the total organic carbon content of the cultures. To a lesser extent, this was also the case for the heterotrophic protists growing on bacteria and rice corns, although they were previously filtered successively over a 10 μm and a 5 μm mesh. Due to this previous filtration, bacterial biomass was reduced to less than 20% of the total bacterial carbon content in the cultures. The filtration procedure also reduced the occurrence of large rice particles in the samples. However, rice particles and bacterial cells smaller than 5 μm were still present in the samples for biochemical analyses. Since a complete separation of heterotrophic protists and their diet was not possible, we adopted a subtraction method in order to estimate the fatty acid and amino acid composition and concentration in heterotrophic protists separately. By knowing the carbon-specific fatty acid and amino acid contents of the diets, obtained from analysis of either the algal or “bacteria + rice” sole-cultures, we determined the biochemical concentration of the heterotrophic protists indirectly through subtraction. For the biochemical analyses, 200 to 300 mL of each culture were filtered in 2–3 replicates on pre-incinerated (550°C, 4 h) GF/C Whatman glass fiber filters. Samples of bacterial cultures were collected on GF/F Whatman glass fiber filters. Samples had a minimum concentration of 1–3 mgC. The analyses were repeated for 4–8 different batches of each consumer-diet combination and the sole dietary cultures. We collected protist cells at the end of the exponential phase because algal and bacterial prey densities were minimal in the cultures by this time (see sample preparation above). Nevertheless, protist cells were still within the exponential growth phase, during which food quality has been found to be the highest (Sterner, 1993). Filters containing the material for the fatty acid analyses were immediately extracted with chloroform-methanol 2:1, v/v (Folch et al., 1957) and homogenised by sonification for 5 min at 5000 cycles min-1(Ultrashall-Desintegrator USD 20, VDE Wiss. Gerätebau, Berlin). [page 26↓]Cryptomonas samples, however, were submitted to a 10 min sonification time period, which assured cell destruction. After homogenising, an internal standard was added to the samples (tricosanoic acid, 0.2 mg mL-1). The samples were allowed to extract for 3 h, at 20°C. After extraction, samples were dried under nitrogen flux and promptly stored at –20°C until analysis. Samples for amino acid analyses were lyophilized and promptly stored at –20°C pending further analysis.
Fatty acid methyl-esters (FAME) were formed by addition of 5 mL sulphuric acid (5% v/v) and heating the samples for 4 h at 80°C (Weiler, 2001). An aliquot of 0.2 μL of the samples was finally injected into a Varian Star 3600 CX series gas chromatograph, equipped with a fused silica capillary column (Omegawax 320, SUPELCO, 30 m x 0.32 mm). The following heating program was applied: initial temperature of 180°C (2 min), subsequent heating at 2°C min-1 to 200°C, which was held isothermally for 33 min Injector and FID detector temperatures were 250°C and 260°C, respectively. Helium was used as a carrier gas. FAMEs identification were made by comparing the retention times with retention times of a calibration standard solution (Supelco FAME Mix 47885–4, PUFA Nr. 3–47085–4 and PUFA Nr. 1–47033) and quantified by comparing the peak areas with the peak area of the internal standard. For presentation, we selected some fatty acids from the total measured pool. It should be kept in mind that the sums of saturated, monounsaturated, and polyunsaturated fatty acids presented in the results refer to the whole pool of measured fatty acids.
Lyophilized filters were hydrolysed by the addition of 6N HCL and incubation of the samples at 110°C for 24 h. After hydrolysis, samples were neutralized with 6N NaOH. A 200 μL aliquot of the neutralized amino acid solution was diluted with 1 mL of a methanol : water (80:20, v/v) solution (Ogunji and Wirth, 2001). A 50 μL internal standard (4 μg mL-1 homoserine) was added to an aliquot of 500 μl of the diluted samples and the vials were transferred to an Auto Sampler. A 50 μL aliquot of the diluted mixture was injected in triplicate into a Merck/Hitachi HPLC system (Ogunji and Wirth, 2001) using a Nova-Pak C18 column, 4 μm, 3.9 x 300 mm (Waters GmbH, Germany).
Differences in the fatty acid and amino acid composition between heterotrophic protists and their diet were tested using one-way ANOVA, followed by Dunnett’s [page 27↓]test. In this case, the biochemical composition of the protists was tested against dietary biochemical composition, used as experimental control for this purpose. Pairwise comparisons of the biochemical composition among protist species were performed with a Tukey HSD – test following one-way ANOVA. To test the hypothesis that fatty acid and amino acid composition in the protists was dependent on dietary composition, we performed stepwise discriminant analyses (SDA). Discriminant analysis is used to determine which variables better discriminate between two or more defined groups. In our case, the analyses were performed to determine which fatty acids and amino acids (variables) better separated the organisms we analysed (groups). The basic idea underlying the discriminant analysis is to determine whether groups differ with regard to the mean of a variable, and then to use that variable to predict group membership. In the case of a multiple group stepwise discriminant analysis (6 different groups – 4 protist species and two diets in our analysis), the ultimate calculations correspond to a canonical correlation analysis, which provides the successive functions and canonical roots containing the variables mostly responsible for separating the groups. For the fatty acid data, SDA were performed separately for saturated, monounsaturated, and polyunsaturated fatty acids. Although we do not present the entire pool of fatty acids we measured, all fatty acids were considered for the statistical analyses. We performed one SDA for the whole pool of amino acid data. All statistical procedures were run in Statistica for Windows (version 5.01, Stat Soft).
We found higher amounts of saturated and polyunsaturated than of monounsaturated fatty acids in Cryptomonas (Table 1). High concentrations of the saturated acids 16:0 and the polyunsaturated acids 18:3ω3 and 18:4ω3, as well as high concentrations of eicosapentaenoic acid (EPA) for Cryptomonas sp. have already been described (Ahlgren et al., 1990; Ahlgren et al., 1992; Von Elert and Stampfl, 2000). Our Cryptomonas species contained also high carbon-specific concentrations of linoleic acid (18:2ω6), 20:(2–3)ω6, arachidonic acid (20:4ω6), docosapentaenoic acid (DPA, 22:5ω3), and docosahexaenoic acid DHA (22:6ω3).
The composition of all major fatty acid classes (SAFA, MUFA, PUFA) were similar in the two ciliates fed Cryptomonas and basically resembled the composition of the algal diet (Table 1). The only exceptions were DPA and DHA, which were not found in the ciliates, yet present in Cryptomonas, and the polyunsaturated fatty acid 20:3ω6, which was not found in Balanion (Table 1). However, carbon-specific concentrations of almost all monounsaturated and some polyunsaturated fatty acids were higher in the ciliates than in their algal diet (Dunetts’ test, P<0.05; Table 1). Cryptomonas contained more ω6 than ω3 fatty acids, while both Balanion and Urotricha showed the inverse pattern, as shown by the ω6:ω3 ratios (Table 1). Total SAFA, MUFA, and PUFA concentrations in Balanion and Urotricha were similar, although concentrations differed for a few fatty acids (e.g. DHA, Tukey HSD – test, P<0.05).
A fatty acid distribution typical of bacteria was observed in the “bacteria + rice” diet (Table1). High carbon-specific concentrations of saturated fatty acids such as 16:0, 18:0, 19:0, and 24:0 were found, in agreement with other studies (Fredrickson et al., 1986; Kaneda, 1991; Harvey et al., 1997). We also found high carbon-specific concentrations of the monounsaturated fatty acids 16:1ω9 and 16:1ω5 both considered typical for bacteria (Gillan et al., 1981). However, we observed high carbon-specific concentrations of polyunsaturated fatty acids, in contrast with previous studies (Kaneda, 1991; Harvey et al., 1997).
The fatty acid composition of Cyclidium and Chilomonas deviated more considerably from that of their “bacteria + rice” diet (Table 1). Although not present in their “bacteria + rice” diet, the fatty acids 14:0, 18:3ω6, 20:2ω6, 20:3ω6, and 21:5ω3 were found in both Cyclidium and Chilomonas (Table 1). Higher SAFA, MUFA, and PUFA carbon-specific concentrations were observed in the heterotrophic protists than in their “bacteria + rice” diet, except for the saturated acid 24:0, which was found in very low concentrations in the protists, despite its elevated concentration in the “bacteria + rice” diet (Dunnetts’ test, P<0.05). On the other hand, the monounsaturated acid 20:1ω9 was found in high carbon-specific concentrations in the “bacteria + rice” diet and in Cyclidium, but could not be detected in Chilomonas (Table 1). We detected DHA in Chilomonas, although not present in the “bacteria + rice” diet and in Cyclidium (Table 1). Carbon-specific concentrations of monounsaturated fatty acids in Cyclidium were higher than those found in Chilomonas, largely attributable to discrepancies in the concentrations of 18:1ω7 and 20:1ω9 (Table 1). Cyclidium contained higher EPA concentrations than Chilomonas and the algal diet Cryptomonas (ANOVA, and subsequent Tukey HSD – test, P<0.01; Table 1).
Balanion and Urotricha had significantly higher carbon-specific concentrations of SAFA (P<0.001), MUFA (P<0.001), and PUFA (P<0.01) than either Cyclidium or Chilomonas, or both (ANOVA, and subsequent Tukey HSD – test). Carbon-specific EPA concentrations were similar among Balanion, Urotricha, and Cyclidium, and were higher than in Chilomonas (ANOVA, and subsequent Tukey HSD – test, P<0.01; Table 1). These results are in agreement with the significantly higher total carbon-specific fatty acid concentrations observed in the algal diet Cryptomonas than in the “bacteria + rice” diet, except for monounsaturated fatty acids (Table 1).
Stepwise discriminant analyses for fatty acids were run separately for saturated, monounsaturated, and polyunsaturated fatty acids. Although the analyses often provided more than two significant canonical roots (discriminant functions) only the first two roots are presented in each case (Fig. 1), because together they explained more than 85% of the total variation among groups. The results of the first 3 roots are presented in Table 2.
|Fig. 1 – Canonical roots provided by stepwise discriminant analyses (SDA) run separately for saturated (SAFA), monounsaturated (MUFA), and polyunsaturated fatty acids (PUFA). See text and Table 2 for the fatty acids discriminating within each root.|
The SDA on saturated fatty acids provided 4 significant roots, each root representing one or more fatty acids responsible for discriminating between protists and dietary assemblages (Wilks’ Lambda for SDA<0.0001, approximated F=25.1, P<0.001). The “bacteria + rice” diet was separated from all other organisms at the first root, due to their significant lower 20:0 concentrations and their higher 24:0 concentrations (Fig. 1, Table 2). Three groups were discriminated at the second root, basically due to the significant differences in the concentrations of the saturated fatty acids 12:0, 19:0, and 21:0 (Fig. 1, Table 2). Cryptomonas was separated from Balanion and Urotricha, which formed a second group together with Chilomonas and the “bacteria + rice” diet. The third group discriminated at this root was composed only by Cyclidium. The third root discriminated between Urotricha and Balanion, due to the significant differences in the saturated acids 8:0 and 15:0 (Table 2). Cyclidium and Chilomonas were better separated at the fourth root of the SDA, basically due to the differences in the saturated fatty acid 10:0.
The second SDA, based on differences in monounsaturated fatty acids between protists and diets, provided 4 significant canonical roots (Wilks’ Lambda for SDA<0.0001, approximated F=16.5, P<0.001). The first root discriminated among three groups as follows: the first group was formed by Chilomonas and Cyclidium, the second by both the algal and “bacteria + rice” diets, and the third one was formed by the ciliates fed the algal diet (Fig. 1). Especially the significant differences involving the monounsaturated fatty acids 18:1ω7, 22:1ω11, and 16:1ω5 were responsible for discriminating between these groups (Table 2). The second root more efficiently separated diets from protists (Fig. 1) basically due to the significant differences involving the monounsaturated fatty acids 12:1, 14:1ω5, and 16:1ω7 (Table 2). Cyclidium and Chilomonas were separated at the third root, due to the significant differences involving the 20:1ω9 (Table 2). The two ciliates fed the algae were separated at the fourth root, due to the differences in the concentrations of 18:1ω9. Given their similar concentrations of monounsaturated fatty acids, the algal and the “bacteria + rice” diets were not discriminated by any root provided by the SDA on monounsaturated fatty acids.
The SDA based on polyunsaturated fatty acids provided the best discrimination among groups, resulting in three significant roots (Wilks’ Lambda for SDA<0.0001, approximated F=35.9, P<0.001). The first root discriminated between two groups. The first group was composed by Cyclidium and Chilomonas, the second one by Urotricha, Balanion, the algal diet Cryptomonas, and the “bacteria + rice” diet (Fig. 1). Moreover, a slight separation was observed between Cyclidium and Chilomonas at this root. Overall, differences in the [page 31↓]polyunsaturated fatty acids 22:3ω6, 18:2ω6, 22:5ω3, 18:3ω6, 20:2ω6, 20:4ω3, and 18:3ω3 contributed most to the separation observed at this root (Table 2). The second root separated Balanion and Urotricha from their algal diet Cryptomonas (Fig. 1). The algal diet was also separated from the “bacteria + rice” diet at this root. The polyunsaturated fatty acids separating groups within this root were 22:2ω6, DPA, DHA, 16:3ω4, 16:2ω4, and 20:3ω6. The third root efficientlyseparated Balanion from Urotricha, and Cyclidium from Chilomonas, due to the significant differences observed for 20:3ω3, EPA, and 18:3ω4 (Table 2).
All essential amino acids found in the diet were also found in the heterotrophic protists (Table 3). However, the carbon-specific amino acid concentrations differed between protists and diet, as revealed by the Dunnett’s tests following one-way ANOVA (Table 3, P<0.05). Balanion and Urotricha showed higher carbon-specific concentrations than their algal diet Cryptomonas for 6 of 10 amino acids (threonine, arginine, methionine, phenylalanine, isoleucine, and leucine; Table 3). Moreover, Balanion had higher carbopn-specific histidine concentrations than Cryptomonas, whereas Urotricha had higher tryptophan and lysine concentrations than the algal diet (Table 3). Only valine concentrations did not differ significantly between both ciliates and their algal diet Cryptomonas. Except for leucine and tryptophan, carbon-specific amino acid concentrations were similar in Balanion and Urotricha (Tukey HSD – tests, P<0.05 for leucine and tryptophan).
Except for histidine, methionine, and leucine concentrations, the carbon-specific amino acid concentrations were significantly higher in both Cyclidium and Chilomonas than in their “bacteria + rice” diet (Table 3). Moreover, higher leucine concentrationswere found in Cyclidium, and higher histidine and methionine concentrations were found in Chilomonas than in the “bacteria + rice” diet (Dunnetts’ test, P<0.05). Chilomonas had significantly higher carbon specific histidine, arginine, tryptophan, valine, isoleucine, and lysine concentrations than Cyclidium (Tukey HSD – test, P<0.01), although both species were fed the same diet (Table 3).
The SDA provided 5 discriminant functions (given by canonical roots), but only the first two were significant (Wilks’ Lambda for SDA=0.00007, approximated F=7.96, P<0.001) (Table 4). The first canonical root separated a group composed by Balanion, Urotricha, and Chilomonas, from a group formed by Cyclidium and both diets (Fig. 2). Phenylalanine, lysine, tryptophan, methionine, isoleucine, and [page 32↓]valine were the amino acids separating the groups at this root. The second root separated Cyclidium from the dietary assemblages (Fig. 2), mainly due to differences in histidine and leucine concentrations (Table 4). According to the SDA no significant differences were detected for the amino acid concentrations between the algal and the “bacteria + rice” diet.
|Fig. 2 – Canonical roots provided by stepwise discriminant analysis (SDA) for amino acid data. The first root separated Chilomonas, Balanion, and Urotricha. Second root separated Cyclidium from both diets. See text and Table 4 for the amino acids discriminating within each root.|
We found a positive correlation between cell-specific fatty acid concentration and cell size. The algivores Balanion (3256 µm3) and Urotricha (2778 µm3) had significantly higher concentrations of saturated, monounsaturated, and polyunsaturated fatty acids, including EPA (Tukey HSD – Test, all P values <0.001) than Cyclidium (1315 µm3) and Chilomonas (403 µm3) (Fig. 3). Cell-specific amino acid concentrations were similar in all protist species, except for threonine (P<0.01) and leucine (P<0.001), which were significantly higher in Balanion (Fig. 3).
|Fig. 3 – Cell-specific concentrations of fatty acids (upper panels) and amino acids (lower panels) in the ciliates B. planctonicum and U. farcta fed the algae C. phaseolus,and the heterotrophic protists Cyclidium sp. and C. paramecium fed the “bacteria + rice” diet. Abbreviations are as follows: SAFA=saturated fatty acids, MUFA=monounsaturated fatty acids, PUFA=polyunsaturated fatty acids, EPA=Eicosapentaenioc acid, DHA=Docosahexaenoic acid; His=histidine, Thr=threonine, Arg=argenine, Trp=tryptophan, Met=metheonine, Val=valine, Phe=phenylalanine, Ile=isoleucine, Leu=leucine, Lys=lysine.|
The majority of the dietary biochemical components were also observed in the heterotrophic protists. However, the protists ensured higher concentrations of several biochemical compounds than their algal diet, a pattern already observed for other herbivores (Raubenheimer, 1992). Moreover, differences in fatty acid and amino acid composition and concentrations among protist species suggest species-specific features in the metabolism of those biochemical compounds in the studied heterotrophic protists.
One could expect to find high fatty acid concentrations in heterotrophic protists fed a diet rich in fatty acids. If so, Balanion and Urotricha should contain higher fatty acid concentrations than Cyclidium and Chilomonas, as the algal diet had [page 34↓]higher concentrations of saturated and polyunsaturated fatty acids than the “bacteria + rice” diet. In fact we observed significant higher total SAFA, MUFA, and PUFA concentrations in the ciliates fed the algae than in the protists fed the “bacteria + rice” diet.
However, concentrations of monounsaturated fatty acids did not differ significantly between the algal and “bacteria + rice” diets. Since we found higher concentrations of monounsaturated fatty acids in both Balanion and Urotricha than in Cyclidium and Chilomonas, some metabolic differences in accumulation of monounsaturated fatty acids are supposed to occur among those protist species. These mechanisms could also be responsible for the high concentrations of some polyunsaturated fatty acids in the protists, despite low dietary concentrations (e.g. 16:2ω4, 18:3ω3, and the 22(ω6) fatty acids). Metabolic mechanisms underlying the accumulation of mono- and polyunsaturated fatty acids in the protists may include high fatty acid assimilation efficiencies and protist ability to preferentially use self-synthesized carbohydrates or other organic compounds as a source of energy. In both cases, carbon-specific concentrations of essential compounds obtained in the diet are expected to increase in the heterotrophic protists. In addition, the protist species we investigated may have compensated for occasionally low concentrations of mono and polyunsaturated fatty acids in their diet simply by exhibiting high ingestion rates. In this case, the cellular amount of fatty acids in the protists would be originated only from the diet, given that the studied protists are able to retain fatty acids, but not able to synthesise them. An estimate of the ingestion rates needed to reach the EPA concentrations we measured in the protists, versus known published ingestion rates for Balanion planctonicum (190 Cryptomonas cells ciliate-1 day-1, Müller and Schlegel, 1999) and Cyclidium sp. (28000 bacteria ciliate-1 day-1, Šimek et al., 1994) makes this a reasonable assumption. As we took samples for biochemical analysis around the sixth day of the growth phase, all protist species we investigated may have been able to obtain their entire cellular fatty acids from the diet. It muss be kept in mind, however, that estimates based on published ingestion rates have a very speculative character, although we have worked with the same species.
However, efficient ingestion rates alone cannot explain the higher carbon-specific concentrations of mono- and polyunsaturated fatty acids in the heterotrophic protists. For instance, despite the lower EPA cellular concentrations observed in Cyclidium, which obviously reflect its small size, this ciliate had carbon-specific EPA concentrations, which were similar to the EPA concentrations found in Balanion and Urotricha. Even if Cyclidium was able to get its entire cellular EPA amount from the diet, EPA incorporation efficiency in this species would need to [page 35↓]be considerably higher than in both Balanion and Urotricha, in order to result in similar carbon-specific concentrations. Moreover, since linolenic acid (18:3ω3), a precursor of EPA and DHA (see Appendix 1) was present in Cyclidium (Table 1) its conversion into EPA may be likely to occur in this ciliate.
The experiments of Erwin and Bloch (1963) and Lees and Korn (1966) illustrated the ability of Tetrahymena pyriformis to take up and incorporate, and/or convert and incorporate a considerable variety of short-chain and long-chain fatty acids, and by doing so, to modify its fatty acid composition. Moreover, this ciliate is known to synthesize unsaturated fatty acids by two distinct pathways from palmitic acid (Koroly and Connor, 1976). First evidences for the synthesis of highly unsaturated fatty acids, like EPA and DHA, and also sterols by the heterotrophic dinoflagellate Oxyrrhis marina (Klein Breteler et al., 1999) suggest reconsidering the assumption that heterotrophic protists are unable to synthesize complex lipid molecules. However, synthesis of unsaturated fatty acids may be inefficient in phagotrophic protists. Inefficient conversion ability was claimed for zooplankton consumers, which showed enhanced growth and reproduction in media supplemented with EPA or DHA (Brett and Müller-Navarra, 1997; DeMott and Müller-Navarra, 1997). Our study suggests synthesis of DHA by Chilomonas, possibly using 18:3ω3 as a precursor (Table 1). This could indicate a good prey qualityof Chilomonas, as indeed observed for Daphnia (Sanders and Porter, 1990; Lair and Picard, 2000). One could argue that DHA in Chilomonas might have originated from ingested rice particles, assuming that this flagellate is able to ingest and to digest plant matter. However, the polished rice we used did not contain EPA and DHA (Boëchat unpublished data). Moreover, Cyclidium was grown in the same medium and did not contain any DHA, the same being true for the “bacteria + rice” diet. On the other hand, although present in Cryptomonas, DHA was not found in Balanion and Urotricha. Possibly, Balanion and Urotricha have low metabolic demands for DHA, or are unable to convert DHA from the polyunsaturated fatty acid 18:3ω3.
The SDA based on fatty acid data suggested a relative independence of our heterotrophic protists from dietary fatty acids. The polyunsaturated fatty acids were the best parameter in separating Balanion and Urotricha from their algal diet, the same being valid for Cyclidium and Chilomonas. All protists as well as their respective diets were discriminated already at the first and second roots, with clear differences originating from a broader palette of polyunsaturated fatty acids than observed for the SDA based on saturated or monounsaturated fatty acids. Although these fatty acid classes were also able to [page 36↓]significantly discriminate among protist species and between diets, the composition of polyunsaturated fatty acids seems to be more diverse and maybe strongly affected by species-specific metabolism. This is not surprisingly, since energetic costs of building polyunsaturated fatty acids are much higher than those for saturated or monounsaturated fatty acids. Also the presence of a broader number of accessory enzymatic systems is necessary to insert double bounds into the fatty acid carbon chain (Stryer, 1995). Heterotrophic protists fed the same diet were only separated at the third or fourth root of the SDA based on saturated or monounsaturated fatty acids, which indicates a rather similar and conservative composition among those species, which could arises partially from the diet or perhaps from similar features in metabolism and synthesis of those fatty acid classes. The SDA also suggested a more efficient accumulation of dietary saturated and monounsaturated fatty acids by Cyclidium and Chilomonas, as those species were clearly separated from their “bacteria + rice” diet at the first canonical root of the analyses. Balanion and Urotricha were still placed together with their algal diet at the first root of the SDA based on saturated fatty acids, which indicates a rather similar composition in those ciliates and in Cryptomonas.
We found high concentrations of polyunsaturated fatty acids in the “bacteria + rice” diet, in contrast to previous studies (Kaneda, 1991; Harvey et al., 1997). This might have resulted from an indirect influence of small rice particles from the culture medium, still present after filtering the cultures subsequently through 10 μm and 5 μm meshes (see sample preparation). Unfortunately, a quantification of that influence was not possible. Microscopic examinations, however, gave no evidence for high numbers of rice corn particles in our samples. Except for linoleic and linolenic acid, polished rice corns do not contain measurable amounts of long chain fatty acids (Souci et al., 1994). Fatty acid analyses of the rice corns used for our study confirmed this assumption (Boëchat, unpublished data), although we found high amounts of the saturated fatty acid 24:0 in our rice samples. Nevertheless, the results on dietary polyunsaturated fatty acids in the “bacteria + rice” samples should be interpreted with caution and should be considered when comparing our results with natural bacterial assemblages.
Although the amino acid composition in the heterotrophic protists generally reflected that of their diet, carbon-specific concentrations were higher in the protists than in their diet. Moreover, carbon-specific amino acid concentrations differed species-specifically among the four investigated heterotrophic protists (Table 3). Again, metabolic capabilities of the protists involving ingestion and assimilation efficiencies may be the underlying mechanisms explaining the [page 37↓]discrepancies between protists and diet. Amino acid requirement and assimilation in Tetrahymena has been shown to be dependent on the concentrations of other nutrients in the culture medium such as sodium (Holz, 1973). This ciliate is also able to recover changes in cell volume resulting from changes in extracellular osmolarity, partially by adjusting the intracellular concentration of free amino acids (Dunham, 1973). Moreover, similar toplants and bacteria, which are able to synthesize their entire set of essential amino acids (Stryer, 1995) amino acid synthesis in heterotrophic protists may be expected to occur. Since different biosynthetic pathways for a single amino acid can differ among species (Umbarger, 1981), this may have contributed to the observed differences among protist species (e.g. the higher carbon-specific concentrations in Chilomonas than in Cyclidium).
The tendency of higher amino acid concentrations in the protists than in their diet, also observed for fatty acids, strengthens our hypothesis that heterotrophic protists can compensate for low dietary biochemical contents, simply by exhibiting high ingestion rates and high assimilationefficiencies. Here, dietary biochemical composition should have only a limited influence on the biochemical composition of the heterotrophic protists. The actually determining factors are possibly the metabolic features of the protists. This possibility is fascinating and deserves further investigation. Experiments with radioactively labeled amino acids are the next necessary step to clarify species-specific features in amino acid metabolism of heterotrophic protists.
When considering overall energy uptake by predators, data on carbon-specific biochemical concentrations are important, because carbon provides a good biomass and energy estimate. However, in light of the entire feeding process, which includes capture of prey, handling time and ingestion of prey, the prey size (Mohr and Adrian, 2001) and consequently the cell-specific biochemical content is crucial. There is a trade-off between the effort of prey handling and the nutrition of the single prey. Cell-specific fatty acid concentrations in both Balanion and Urotricha significantly exceeded those found in Cyclidium and Chilomonas (Fig. 3). With respect to the fatty acid input per single predator/prey encounter, higher fatty acid concentrations in Balanion and Urotricha could guarantee a better food quality for rotifers fed algivores as compared to bacterivores, a fact already observed (Mohr and Adrian, 2002b). In contrast, amino acid concentrations seemed to be not so dependent of diet composition and protist cell size. Thus, amino acid [page 38↓]concentration cannot be regarded as a good indicator of protist nutritional quality when comparing algivorous and bacterivorous protist prey. This is consistent with findings of rather constant amino acid concentrations in freshwater microalgae (Ahlgren et al., 1992) and rotifers (Frolov et al., 1991), and is supported by previous studies, which demonstrate the absence of relationship between prey amino acid composition and nutritional quality (Watanabe et al., 1978; Frolov et al., 1991).
In aquatic ecosystems, the second level of the Eltonian biomass pyramid can be larger than the producers’ level (Brett and Müller-Navarra, 1997). In our case, the protist matter was more concentrated in various chemical compounds than the algal or “bacteria + rice” matter. Therefore, heterotrophic protists might be able to ingest and/or accumulate these compounds efficiently. Although we did not measure ingestion and assimilation rates, we assume that both Balanion and Urotricha, and especially Cyclidium were probably better in accumulating monounsaturated and some polyunsaturated fatty acids (Fig. 1). All protist species, but especially Chilomonas,seemed to be efficient in amino acid accumulation. Overall, with respect to various essential chemical compounds, heterotrophic protists may be viewed as prey of upgraded quality at an early stage in aquatic food webs. The fact that the biochemical (the present study) and elemental compositions (Caron et al., 1990) of heterotrophic protists can indeed differ from dietary composition suggest species-specific differences in metabolic pathways of heterotrophic protists. This possibility may stimulate further research concerning assimilation and metabolism of organic matter in heterotrophic protists. In light of trophic interactions and trophic transfer efficiency, our results may be very useful to substantiate the view of heterotrophic protists as an important link in transferring biochemical matter and energy between the microbial and the classical food web.
|© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.|
|DiML DTD Version 4.0||Zertifizierter Dokumentenserver|
der Humboldt-Universität zu Berlin