Abstract
Stable isotopes of carbon and nitrogen characterize trophic relationships in predator-prey relationships, with clear differences between consumer and diet (discrimination factor, Δ13C, Δ15N). However, parasite-host isotopic relationships remain unclear, with Δ13C and Δ15N remaining incompletely characterized, especially for helminths. In this study, we used stable isotopes to determine discrimination factors for 13 parasite-host pairings of helminths in coral reef fish. Δ15N differences grouped according to phylogeny and attachment site on the hosts: Δ15N was positive for trematodes and nematodes from the digestive tract and varied for cestodes and nematodes from the general cavity. Δ13C showed more complex patterns with no effect of phylogeny or attachment site. A negative relationship was observed between Δ15N and host δ15N value among different host-parasite pairings as well as within 7 out of the 13 parings, indicating that host metabolic processing affects host-parasite discrimination values. In contrast, no relationships were observed for Δ13C. Our results indicate that host phylogeny, attachment site and host stable isotope value drive Δ15N of helminths in coral reef fish while Δ13C is more idiosyncratic. These results call for use of taxon- or species-specific and scaled framework for bulk stable isotopes in the trophic ecology of parasites.
Introduction
Parasitism is the most common life strategy for consumers and is ubiquitous amongst food webs1. However, parasites remain a neglected component during the evaluation of biodiversity2 and trophic relationships for parasites3 remain poorly characterized within food webs as small size, multidisciplinary requirements for identification, and cryptic lifestyles (e.g. multiple hosts associated to multiple larval stages) make identification and characterization of these relationships difficult. The role of parasites in aquatic food chains has been shown to be fundamental4,5 and the inclusion of parasitic relationships within food webs dramatically increases the number of trophic links within ecosystems6. Despite this, host-parasite relationships remain poorly described, especially for systems with high biodiversity such as coral reefs7.
Stable isotope techniques are routinely utilized to study trophic relationships within food webs8,9 by using trophic discrimination factors for carbon and nitrogen (Δ13C and Δ15N) to account for the stepwise increase in δ13C and δ15N (‰) that occurs between diet and consumer during metabolism10–12. However, parasites do not generally follow this relationship with parasite-host discrimination factors (Δ13C or Δ15N) observed ranging from considerably higher than typical trophic discrimination between predator and prey to negative values for Δ13C or Δ15N across a variety of taxa1,13,14.
Amongst helminths in fish hosts, cestodes and nematodes are usually depleted in δ15N values versus their hosts13,15–17 and vary in Δ13C, while trematodes have been found to vary in both Δ13C and Δ15N18,19. However, there can also be considerable variation within these helminth groups14,20. Distinct differences for trophic discrimination in parasitic relationships are potentially caused by the combined effects of unique feeding ecologies, often reduced metabolic capabilities of the parasitic taxa being investigated21,22, and host metabolic effects due to parasitism23. Feeding ecology varies depending on whether the parasite feeds upon host tissue exclusively (on host or within; tissue type dependent18), or is able to supplement with material from within the dietary tract as the host feeds or from the environment (e.g. prey items, detritus, mucus, or blood24). In addition, it has been suggested that trophic discrimination factors of parasites may not be fixed but scale with the isotopic signature of their hosts, both within25 and among parasite species14.
Despite multiple investigations, clear discrimination patterns between taxa have not emerged, making simple incorporation of parasites into food web studies using a single universal trophic discrimination factor impossible. This knowledge gap warrants further investigation into the drivers of discrimination factors in helminths. In this study, we examine both δ13C and δ15N values from whole tissue of 136 helminth parasite-host pairings from coral reef fish to determine the isotopic relationship between taxonomically distinct groups of helminth parasites and to investigate the effect of the habitat in the host and host isotopic signature on helminth isotopic discrimination. We expect that parasite isotopic enrichments and variability versus their host will be larger in parasites located in the dietary tract while parasites solely utilizing host tissues will have less variability in their isotopic discrimination. In addition, we expect a negative scaling of parasite discrimination factors with the δ13C and δ15N values of their hosts.
Results
We examined the isotopic discrimination of 136 helminth parasite-host pairings including trematodes (n=27), cestodes (n=19), and nematodes (n=90) from 4 host reef lagoon-associated fish species (Table 1, Fig. 1). Trematodes and five of the nematode species were sampled from the dietary tracts (DT) of host species while the remainder were samples from the general cavity (GC). Six of the Δ15N values for parasite-host pairs were negative (Supplementary Table 1), with positive relationships (1.05 to 1.58‰) predominately occurring in dietary tract associated parasites (Fig. 1). In 9 cases δ15N values were statistically significantly different between parasites and hosts (Supplemental Table 1 & Fig. 2), excluding two pairs from the general cavity: A. novacaledonica-L. genivitattus and Philometra sp.-S. undosquamis (0.58 and 1.18‰, respectively). For the three parasite pairs found in both L. genivittatus and N. furcosus (i.e. A. novacaledonica, Callamanus sp., and Pseudophyllidae), Δ13C and Δ15N were consistently the same between the parasite-host pairings within the same host, either positive or negative, except for carbon in A. novacaledonica (0.47‰ L. genivitattus versus −1.10‰ N. furcosus) (Supplemental Table 1, Fig. 2). Comparison of Δ15N between taxa and tissue types for nematodes found significant differences (One-way ANOVA: F3, 132: 26.9 p< 0.001; Fig. 3). Host δ15N was examined versus Δ15N and found that a linear regression for all of the pairings combined minus the herbivore pairing had a negative slope of −1.2 (R2 = 0.071; Fig. 4) with negative slopes observed within individual pairings that were statistically different than 0 for 7 of the parasite-host pairings at α=0.05 (Fig. 4; Supplementary Table 2).
Δ13C values of the pairings were generally negative with lower δ13C in the parasite than those of the host fish, except for S. undosquamis / Philometra sp. (Δ13C −0.06 to 1.19‰), S. lineatus / white nematode (Δ13C −0.38 to 3.81‰) and L. genivittatus / A. novacaledonica (Δ13C 0.47 to 0.98‰, Figs. 2 & 3). The first two of these relationships occur for parasites in the digestive tract and the last one in the general cavity of the host fish. Among the 13 fish-parasite pairings tested, statistically significant differences between host and parasite δ13C were observed for 10 cases with a Δ13C range from −0.73 to −1.97 ‰. There was no clear Δ13C distinction observed between parasites found in the digestive tract versus those found in the general cavity. Comparison of Δ13C between taxa and tissue types for nematodes found no significant differences (One-way ANOVA: F3, 132:1.4 p= 0.2; Fig. 3). Host δ13C was examined versus Δ13C and found that a linear regression for all of the pairings combined minus the herbivore pairing had a negative slope of −0.56 (R2 =0.196, p<0.001; Fig. 4), but that linear regressions for individual pairings indicated no statistically significant difference from 0 for those relationships (Supplementary Table 2).
Mean δ13C and δ15N values for the host species L. genivittus were −14.04 ± 0.51‰ and 8.97 ± 0.52‰, N. furcosus were −14.16 ± 0.49‰ and 9.31 ± 0.50‰, S. undosquamis were - 16.25 ± 0.49‰ and 8.07 ± 1.09‰, and S. lineatus were −15.75 ± 0.40‰ and 8.97 ± 0.59‰, respectively (Fig. 1; Supplementary Table 1), with significant differences observed between species for both δ13C and δ15N values (One-way ANOVA: δ13C, F3,177= 52.6, p< 0.001; δ15N, F3,177=22.8, p< 0.001). Body size did not influence δ13C and δ15N values for S. undosquamis and S. lineatus for the size range and number of individuals considered here (Pearson correlation, p> 0.05 in all cases). By contrast, fish size significantly influenced δ15N values for L. genivittus and N. furcosus (p< 0.001 for both species; Fig. 5) but not for δ13C values (p> 0.05). Calculated trophic levels were similar for the whole populations of L. genivittus and S. undosquamis (2.86 and 2.79, respectively) with S. lineatus having the lowest and N. furcosus having the highest trophic levels (2.52 and 2.99, respectively; One-way ANOVA, F3,137=42.7, p< 0.001). Trophic levels were also higher for larger fish for both L. genivattus (2.6 and 2.86 for 11-15 cm and 18-21.5 cm individuals, respectively; one-way ANOVA: F1,39, p< 0.001) and N. furcosus (2.78 and 2.99 for 12.1-15 cm and 20-25.3 cm individuals, respectively; one-way ANOVA: F1,61=27.6, p< 0.001).
Discussion
Δ15N varied inconsistently within and between taxa, with the most consistent result being elevated Δ15N (>0‰) for dietary tract associated nematodes likely associated with feeding on host dietary items in addition to tissue. Δ13C was consistently negative between parasite taxa and likely indicates increased reliance on fatty acids from the host to support tissue growth in reef fish-associated helminths. The varied relationships amongst and between taxa provide further evidence that parasite-host pairings are distinctly different than typical trophic relationships and warrant further investigation to adequately characterize parasite contributions to food webs.
Δ15N values showed no difference or were positive for the dietary tract associated trematode and nematodes (Allardia, Callamanus, Rhaphidascaris, and unidentified white) while the cestodes had negative values for both pairings examined. Δ15N values for the general cavity-associated nematodes varied, with a strong positive value for the gonad-associated Philometra sp.- S. undosquamis pairing, no difference observed for unidentified white nematode - N. furcosus pairing, and strong negative relationships for both the nematode cyst-type – N. furcosus and the unidentified white nematode – L. genivittatus pairings (Fig. 2).
Trematodes have been found to have positive to neutral Δ15N values regardless of infection site (e.g. dietary tract or general cavity18,19) indicating utilization of host-metabolized nitrogen derived from tissue in addition to nitrogen compounds derived from the host diet. The low Δ15N values (0.18 to 0.6‰) observed in this study demonstrate close association of the trematode with the host diet. Sole reliance on host tissue and therefore host-metabolized N would be expected to yield a “classical” trophic enrichment of ~3.4‰, which is considerably larger than the observed Δ15N. Small Δ15N values likely reflect combined utilization of more 15N enriched sources of nitrogen derived from host tissue as well as more 15N depleted compounds either metabolized or directly assimilated. Negative Δ15N values have been previously observed for cestodes within fish hosts15,16,18,19,24,26 and may be caused by direct utilization of relatively 15N-depleted compounds from the host diet16 or metabolically recycled N-depleted AAs produced by the gut microbial community27 that both cestodes and trematodes are well-positioned to utilize while residing in the dietary tract.
Similarly, three of the general-cavity associated nematodes displayed negative or neutral Δ15N values, indicating that direct uptake from the host without further metabolic processing by the parasite of nitrogen compounds as a likely pathway for N in this taxa. This uptake was not consistent across the taxa, with different species that target similar infection sites (e.g. dietary tract versus general cavity) displaying considerably different Δ15N values. The single gonad-associated nematode (Philometra sp. - S. undosquamis) examined had a positive Δ15N indicating at least partial reliance on direct utilization of metabolized N from host tissues. The dietary tract-associated nematodes were generally 15N-enriched in comparison to their hosts indicating at least partial utilization of host-metabolized compounds from tissue or the taxa-dependent ability for nematodes to biosynthesize AAs from nitrogenous compounds22.
Δ13C values were predominately neutral or negative for ten of the pairings examined, with positive Δ13C only observed for the L. genivittatus - A. novacaledonica, unidentified white – S. lineatus, and Philometra sp. - S. undosquamis pairings. No change or depletion in δ13C does not agree with the expected 0.5-1‰ increase that is usually expected for trophic interactions11, but likely reflects reliance on lipids and fatty acids directly derived from the host or host diet to support helminth tissue growth. Plathelminthes and some nematodes have been found to be incapable of de novo fatty acid synthesis and to have to rely on fatty acids derived from the host22,28 due to incomplete metabolic pathways for lipid biosynthesis. Direct uptake of fatty acids and other lipids from the host would be expected to coincide with a minimum carbon fractionation as no further metabolic processing is required, thereby maintaining the relatively low δ13C values associated with lipids as they are incorporated into the parasite. The relatively uniform neutral or negative relationships for Δ13C values across the pairings located from both the dietary tract and the general cavity indicate that lipid carbon is likely utilized to support tissue growth beyond species closely associated with fatty tissues (e.g. blood and liver)13,14. This relationship should be examined further with methods that incorporate metabolic pathway techniques for target species beyond model organisms targeted for pathogenicity21,22. Pairings that have elevated Δ13C values likely reflect decreased utilization of host lipids and increased reliance on either host sugars or proteins processed through more complete metabolic pathways within the helminths to provide tissue carbon or potentially different δ13C compositions from different host tissues18.
δ15N values for both L. genivattus and N. furcosus increased with body size leading to higher trophic levels in larger fish, a pattern commonly observed for coral reef-associated fish, while there was no corresponding increase or shift in δ13C throughout ontogeny. No change in δ13C indicates that both smaller and larger individuals are relying on similar sources of underlying carbon production29, and that the larger individuals are feeding on larger prey with an elevated trophic level, i.e. elevated δ15N values30.
Additionally, negative scaling of both Δ13C and Δ15N values versus δ13C and δ15N of fish hosts were observed with statistically significant negative slopes observed within pairings for nitrogen in 7 of the 13 parasite-host pairs (Fig. 4), while no statistically significant within-pairings relationships were found for carbon (Supplementary Table 2). Negative scaling of trophic discrimination factors with increased host carbon and nitrogen values has previously been observed for parasite-host, predator-prey and herbivore-plant relationships12,14,31 and may be associated with dietary quality. In this study, there appears to be an increased spread in the fractionation (Δ15N values) observed within the herbivore S. lineatus, with the lowest δ15N values occurring with the highest Δ15N values and a distinct grouping of individuals with lower Δ15N values corresponding with the highest δ15N values (Fig. 4). This wide range of values may represent the relative richness in diet, with strictly herbivorous individuals causing a shift in their parasites towards exclusive utilization of host tissues, while individuals that supplement with animal protein (more omnivorous) provide additional material within their diet for their parasites to supplement from. Increased protein quality in a predator’s diet results in a smaller ‰ difference between the diet and consumer, i.e. a smaller trophic fractionation32. This relationship coincides with the trend of decreased Δ15N values being observed for increased trophic level predation within predators in this study (Fig. 4). In herbivores, larger trophic discrimination factors for nitrogen are often observed33, and supplementation with protein (omnivory) would be expected to generate a negative offset in Δ15N if the parasites are supplementing from dietary protein in addition to host tissues.
In conclusion, this study characterized discrimination factors for carbon and nitrogen within helminths living in coral reef fish and highlights the uncertainties that remain in adequately describing parasitic relationships within food webs. These uncertainties call for the development of a taxon- or species-specific and scaled framework for using bulk stable isotope analysis to study the trophic ecology of parasites. In addition, further work using metabolomics and compound specific stable isotope techniques is warranted in order to better characterize the underlying metabolic differences that are driving the differences observed for trophic discrimination factors between parasite and hosts.
Methods
Sampled areas and studied species
Individual fish were captured in New Caledonia, southwestern Pacific Ocean. The three species Lethrinus genivittatus Cuvier & Valenciennes, 1830, Nemipterus furcosus (Cuvier & Valenciennes, 1830) and Saurida undosquamis (Richardson, 1848) were caught using hand lines in the lagoon off the city of Nouméa (22°18’S and 166°25’E) at approximately 10-12 m depth, in August 2011, 2013 and 2014. Three years of data from catches were pooled as a preliminary two-way ANOVA (year x size) revealed that year was not a significant factor (p > 0.05). L. genivittatus feeds on crabs and worms, N. furcosus feeds on crabs and shrimp and S. undosquamis is predominantly piscivorous34. The species Siganus lineatus (Cuvier & Valenciennes, 1835) was caught in coastal mangroves in the southeast coast at Yaté (22°16’S and 167°0l’l5E) using gillnets in June-August 2014. This species is usually considered an herbivore35, but has been observed to predominately feed on algae and supplement with minor consumption of invertebrates in Yaté36. Parasites were present in all fish that were examined and appear to be ubiquitous within the species examined in this study.
All individuals caught were immediately placed in ice until further processing in the laboratory. Each fish was measured to the nearest 0.1 cm (total length) and a small piece of dorsal muscle of each fish was sampled and immediately frozen at −20°C. To extract the parasites, the general cavity was first examined to collect parasites found outside the digestive tracts embedded in or attached to fish tissues. In a second step, the method presented in Justine, et al. was applied to extract the parasites found alive within the digestive tracts. All parasites having a sufficient biomass were collected and immediately frozen, i.e. nematodes, cestodes and trematodes. A total of 54 L. genivittatus were caught with 36 exploitable fish-parasite pairings, 99 N. furcosus with 75 exploitable fish-parasite pairings, and respectively 7 S. undosquamis and 18 S. lineatus were exploitable as fish-parasite pairings (Table 1).
Stable isotope preparation and analyses
Carbon and nitrogen stable-isotope ratios (δ13C and δ15N) were determined for dorsal muscle tissue of all fishes collected. Fish muscle tissue is routinely utilized for stable isotope values for fish as it does not require lipid extraction prior to analysis38. Samples were freeze-dried and ground into a fine powder using a mortar and pestle. One milligram of powdered material was loaded into tin capsules and analysed for each sample without prior treatment. This same procedure was used for parasites (whole animal) for samples that had sufficient dry mass (≥ 0.3 mg).
13C/12C and 15N/14N ratios were determined with a continuous-flow isotope-ratio mass spectrometer (Thermo Scientific Delta V Advantage, Bremen, Germany) coupled to an elemental analyser (Thermo Scientific Flash EA1112, Bremen, Germany). The analytical precision was 0.1‰ for 13C and 0.15‰ for 15N, estimated using the internal standards leucine calibrated against ‘Europa flour’ and IAEA standards N1 and N2. Isotope ratios were expressed as δ notation (‰) differences from a standard reference material:
Where R is the corresponding ratio (13C/12C or 15N/14N) for both sample and reference standard and δX is the measured isotopic value in per mil (‰) relative to the international standard references are Vienna Pee Dee Belemnite (vPDB) for carbon and atmospheric N2 for nitrogen.
Parasite-host discrimination factors were calculated using:
Where δX represents the isotopic value of carbon or nitrogen for each parasite-host tissue pairing examined.
Data analysis
The significance of differences in δ13C and δ15N between a fish and its parasite was tested with the Wilcoxon signed rank test when homogeneity of variances was not verified or paired samples t-tests when homogeneity of variances was verified, dependant on fish species. The relationships between fish size and isotopic values (δ15N or δ13C) were investigated with Pearson correlation coefficients. One-way analysis of variance (ANOVA) was used to determine significant differences between host and parasite δ13C and δ15N values and to explore the relationship between host size and trophic level. The relationship between host δ13C and δ15N values and Δ13C and Δ15N values were determined through linear regression followed with subsequent application of an F test of the modelled slope against a slope of 0. This was done among and within the host-parasite parings. For the analysis among pairings we excluded samples from the herbivorous fish host S. lineatus (19 samples) due to their very different isotope values to avoid skewing the relationship due to explained outliers. The trophic level (TL) of fish individuals was calculated following the formulae of10: where λ is the trophic level of the source of organic matter, i.e. 1,: δ15Nfish is the isotopic value of nitrogen for the considered fish, δ15Nsource is the isotopic value of the source of organic matter at the base of the food web, i.e. 3.59 for sedimentary organic matter29 that concerns L. genivittatus, N. furcosus and S. undosquamis, all caught of sandy unvegetated bottom; and 2.12 for the most eaten algae by Siganus lineatus and Δ15N that is the trophic enrichment factor (TEF) between a food item and its consumer. Here, we adopted a value of 3.9 ‰ for S. lineatus36,39 reflecting usually higher TEF for herbivores compared to the conventional 3.4 ‰ value33. For the three other species, we adopted a TEF of 3.0 ‰ because TEF are usually lower than the conventional value for carnivores33,40.
Additional Information
Declarations of interest: none. MB, TM, YL and PS contributed to experimental design, sample collection and processing, and data collation. PMR wrote the manuscript with considerable consultation from DT, MvdM and YL for data analysis. All authors reviewed the manuscript prior to submission.
Acknowledgments
We are grateful to the students in “Licence Sciences de la Vie, de la Terre et de l’Environnement” de l’Université de la Nouvelle-Calédonie for their valuable help for the parasites sampling during dissections, to C. Pigot for his help during fish catches and samples preparation, to G. Guillou for stable isotopes analyses.