Rare inventory of trematode diversity in a protected natural reserve

In the face of ongoing habitat degradation and the biodiversity crisis, natural reserves are important refuges for wildlife. Since most free-living organisms serve as hosts to parasites, the diverse communities in protected areas can be expected to provide suitable habitats for a species-rich parasite fauna. However, to date, assessments of parasite diversity in protected nature reserves are rare. To expand our knowledge of parasite communities in natural habitats, we examined 1994 molluscs belonging to 15 species for trematode infections in a central European natural reserve. The parasitological examination revealed an overall prevalence of 17.3% and a total species richness of 40 trematode species. However, the parasite diversity and prevalence did not differ markedly from trematode communities in non-protected environments, which might be partly explained by a dilution effect caused by a high number of non-host organisms in our study system. The proportion of complex and long life cycles of parasites in the present study is high, indicating complex biotic interactions. We conclude that life cycle complexity, in addition to parasite diversity and trematode species richness, can provide valuable information on ecosystem health and should therefore be considered in future studies.


Results
Occurrence of snails and trematodes. A total of 1994 snails was collected at three different sites during monthly samplings from May to October 2017 and from May to September in 2019 in the natural reserve "Bienener Altrhein" in North Rhine-Westphalia, Germany (Fig. 1). Fifteen snail species belonging to seven families were sampled (Bithynidae, Hydrobiidae, Lymnaeida, Physidae, Planorbidae, Valvatidae, Viviparidae). Nine of these species harboured larval trematodes. The lymnaeid Ampullaceana balthica (former name: Radix balthica), which was collected in 2017 and 2019, was the most frequently collected snail species (n = 574), followed by the planorbids Planorbis planorbis (n = 309) and Planorbarius corneus (n = 228) ( Table 1).
With 13 trematode species, A. balthica showed the highest species richness among the snail hosts studied, followed by Bithynia tentaculata, P. corneus and Lymnaea stagnalis with 11, 10 and 9 species, respectively. The remaining snail hosts harboured only few species (Table 1). The relative species richness is on average four trematode species per snail species (Table 2).
Trematode life cycle complexity. Based on a literature survey, the life cycles of the trematode species identified in this study were reconstructed. Our investigation revealed that the proportional occurrence of detected digenean trematodes with a shorter life cycle (i.e. one intermediate host) accounts for 15%, compared to 85% with longer life cycles (i.e. two or three intermediate hosts) (Fig. 4). Species of the genera Asymphylodora, Notocotylus, Sanguinicola and Trichobilharzia are the main representatives that require only one intermediate host. Species belonging to the genera Asymphylodora and Sanguinicola utilise fish as final hosts, whereas species of the genera Notocotylus and Trichobilharzia parasitize waterfowl (Table 3). No significant association between sampling location and life cycle length were detected (Fisher's exact test, p = 0.648).
Host taxa. An extensive literature search was conducted to compile an overview of all recorded trematode species in their second intermediate and final host groups (Table 3). We reviewed final host specificity to detect trematode species that use rare and/or protected host species. Additionally, we looked at host specificity regarding the first intermediate host, for which trematodes usually exhibit high specificity 55 . With a few exceptions, individual trematode species were mainly found in one snail host species or in different species of the same family. The exceptions include, for instance, P. radiatus or E. recurvatum, which were detected in five and four snail species from three families, respectively. A large proportion of digeneans parasitize waterfowl (37.5%), in www.nature.com/scientificreports/ most cases anatid birds. Both, fish-eating birds and a very unspecific group consisting of mammals, birds and amphibians each account for almost 18% of hosts (Fig. 5).

Discussion
Our present knowledge of parasite diversity is still very limited and reveals major gaps in our understanding of the distribution of parasites in different ecosystems, especially in natural freshwaters. Due to this incomplete knowledge of parasite diversity, predictions about how anthropogenic environmental changes will affect parasite communities in the future remain difficult. The present study therefore presents a valuable and rare inventory of trematode diversity in a protected and natural aquatic system and is, to the best of our knowledge, the first evaluation of trematode communities in a natural freshwater reserve. While we can confirm our hypothesis that the studied freshwater nature reserve supports diverse and species-rich trematode communities, no significant differences in parasite life cycle complexity could be shown in comparison to other habitats, and the patchy data basis on host specificity of digenean trematodes hinders accurate conclusions on the role of rare host taxa in our study system. With 40 trematode taxa in 1994 snails, the trematode species richness in our system ranks at the top compared to similar faunistic surveys from freshwater systems 51,52,54,56,57 . However, due to the rather high number of snail host species analyzed in the present study (15 species), of which some hosts harboured only few or no parasites, our study area shows a rather medium to low average species richness per host species with four trematode species per snail species. Moreover, with a prevalence of 17.3%, the natural reserve investigated in the present study is among the lower range, while the overall trematode prevalence in comparable studies ranged from 13.2 56 to 35.1% 52 . In our system, the snail hosts belonging to the two families Lymnaeidae and Planorbidae harboured the vast majority of trematodes we have identified (92%). This is in line with other broad faunistic surveys (e.g. [57][58][59][60][61][62], which indicate that these two snail host families show a particular high prevalence and species richness and can therefore be considered key host groups in aquatic systems 54,63 . Since these trematode-rich key host taxa seem to be prevalent in heavily modified waterbodies, e.g. man-made reservoirs 54 or fish ponds 52 , we do not find trematode communities with an extraordinary high species richness in the natural reserve compared to other systems, as initially expected. Overall, with higher free-living diversity in restored habitats and natural reserves, not only does the diversity of suitable trematode hosts increase but also that of non-host organisms which can act as potential diluters of parasites. High free-living diversity can result in various dilution mechanisms, e.g. an encounter reduction of parasites and suitable hosts, a transmission reduction, in which fewer parasites successfully infect a host, or a reduction of reproduction, in which infected hosts produce fewer infectious stages 64 . Experiments under laboratory as well as under natural conditions have shown that dilution effects particularly affect parasites with complex life cycles and with vulnerable larval stages, such as trematodes and other helminths [64][65][66] . Due to their short-lived and fragile transmission stages (miracidia and cercariae), trematodes are particularly susceptible to reduction by predation or physical disruptions [67][68][69] . Furthermore, a high density and diversity of non-host species, decoy hosts or alternative hosts, especially benthic organisms, may decrease the probability of transmission stages finding suitable hosts 65 . Miracidia in particular show a high host specificity for their mollusc intermediate hosts 55 , so that high mollusc diversity is likely to reduce trematode transmission to target hosts, while parasite species richness might be less impacted by the diversity of vertebrates, for which trematodes show lower host specificity 64 . The  (14). (e) Australapatemon burti (16). (f) Hypoderaeum conoideum (19). (g) Echinostoma revolutum (38 www.nature.com/scientificreports/ species rich and abundant mollusc fauna detected in our study system (15 species) could thus represent one of the main diluters, causing a lower parasite species richness than initially expected. However, the dilution effect of free-living species diversity on parasitism is not universal and may depend on various factors, such as the individual species composition in a particular habitat, the characteristics of certain host-parasite assemblages, or the scale of the study 65 . To further explore such essential concepts in host-parasite ecology, comparable studies from various aquatic systems and habitats with different species compositions will be required. However, comparing parasite diversity and functional ecology across different ecosystem often remains hindered and limited by a patchy knowledge of parasite taxonomy, host-specificity, and by dissimilar sampling and study approaches in different regions. For instance, some of the available freshwater studies assessing parasite diversity are broad faunistic surveys, in which many snail hosts, including those with low parasite diversity (e.g. Potamopyrgus antipodarum, Physa acuta, Aplexa hypnorum), were sampled 51 (present study). Other studies, in contrast, specifically targeted selected species-rich key trematode hosts in the habitat 54,57 . In addition, we find different degrees of resolution regarding parasite identification across studies. Some assessments have focused intensively on selected taxonomic parasite groups using morphological and molecular approaches to discover cryptic species diversity within problematic trematode lineages. Selbach et al. 54 , for example, focused intensively on the genera Diplostomum and Neopetasiger, whereas Soldánová et al. 57 examined the species complex in the genus Plagiorchis in detail, resulting in the discovery of a high number of species in these groups. Altogether, such differences need to be considered when comparing parasite diversity in different systems, and when using parasites as bioindicators to determine environmental conditions and changes over time and between habitats [35][36][37] .
Moreover, for many taxa, data on host specificity of trematodes in their second intermediate and final hosts are still too scarce, and many life cycles of parasites are not fully known 70 . The decreasing parasitological taxonomic expertise and the limited number of large-scale studies of vertebrate final hosts are major obstacles in elucidating many parasite life cycles 19 . For instance, parasites considered to be host-specific might be able to infect a wider host range, as host specificity data are usually based on a limited number of samples 71 . On the other hand, taxa considered generalists might be much more host specific. For example, Diplostomum mergi was long considered a generalist species until it was recently discovered to constitute a species complex whose individual members are highly host-specific in their final host 72 . Accordingly, the available literature information on final hosts is www.nature.com/scientificreports/ often designated in rather broad categories, so it may appear that many parasites are not particularly specific with respect to their final host taxa. This is also reflected in the results of our study, which show that most of the parasites detected seem to parasitize very common and widespread host taxa, e.g., waterfowl or fish-eating birds.
Keeping the above-mentioned limitations in mind, the results from our study allow some valuable comparisons with host-parasite communities from other habitats and ecosystems. Even the rather broad final host categories of parasite communities can reveal information on distinctive free-living communities utilized by the parasites in different aquatic ecosystems that represent a characteristic final host composition. For example, despite the close geographic proximity, similar study approaches as well as similar reported trematode species richness and prevalence, the results of Selbach et al. 54 and from the present study reveal fundamentally different final host structures in these waterbodies (see Fig. 5a vs b). Moreover, the parasites' life cycle complexity, i.e. the number of hosts required to complete the full life cycle, can serve as an indicator to assess biotic interactions in ecosystems. While parasite species with longer life cycles that rely on trophic transmission can indicate longer food chains and thus more complex biotic interactions in a habitat, a higher share of less complex parasite life cycles can indicate ecosystems that are more severely impacted 73 . Assessments from marine systems and estuarine marshes using parasite communities as bioindicators could show contaminated or only recently restored habitats to contain mostly parasites with short life cycles with only one intermediate host, or direct transmission pathways 74,75 . The natural reserve of the present study is predominantly characterized by trematode species with long life cycles that require two or three intermediate hosts (85% of all species, see Fig. 4). Although the statistical analysis did not reveal any significant differences between the studies compared, the complex life cycles and diverse final host groups of the trematode communities in the present study suggest a diverse and trophically well-connected free-living fauna in the protected nature reserve adjacent to the River Rhine. These long chains of host-parasite connections can indicate stable community structures that provide the basis for ecosystem resilience and persistence 21,76 . Overall, in addition to the diversity of parasites that can indicate ecosystem health 21 , their functional ecology and life histories can provide valuable information on interspecific interactions in these habitats.
In summary, we investigated a protected freshwater ecosystem that provides a valuable model system for freeliving and parasite diversity. However, our study also highlights the patchy and heterogeneous data availability, especially with regard to final hosts specificity, which makes it difficult to draw conclusions on the functional diversity and compare parasite communities across various habitats. We therefore encourage future studies on host-parasite communities in natural and protected aquatic systems as well as taxonomic research in this field that will allow a better resolution of cryptic trematode diversity and host specificity for future comparisons. Ultimately, understanding and predicting how global climate change and other anthropogenic pressures will affect complex and fragile freshwater ecosystems and their biotic communities will require a thorough awareness of the parasite communities and their ecological roles that too often remain overlooked in environmental assessments and conservation approaches.

Materials and methods
In total, 1648 snails of 15 species belonging to seven families were collected and examined for trematode infections during monthly collections in spring (May), summer (June-August) and autumn (September, October) in 2017. Additionally, 346 snails of the species A. balthica were sampled monthly from May to September in 2019. To allow comparability, the same sampling effort (time spent sampling) was applied at each sampling site, during each sampling trip, and for all snail host species. Snails were identified to the species level using the identification keys of Glöer 77 and Welter-Schultes 78 . Snails were collected with hand-nets from the submersed vegetation or picked by hand from sediments, stones, deadwood and macrophytes along the shoreline of the oxbow or the littoral zone of the pond. All snails were taken to the laboratory, placed in individual beakers filled with filtered  www.nature.com/scientificreports/ river water at 20 °C room temperature and exposed to artificial light to induce the emission of cercariae. After the day of sampling, every beaker was screened daily for three consecutive days under a stereomicroscope for the presence of cercariae. Snails that did not show an infection with trematodes during this time were dissected and examined for prepatent infections (rediae/sporocsysts).

Data analysis.
Trematode prevalence (p) was calculated separately for parasite assemblages in all infected snail species as the proportion of infected host individuals in relation to the total number of host individuals in a population (p = n inf /N × 100, where n inf is the number of infected snails of one species and N are all snails of one species in a sampled population) following Bush et al. 84 . The overall prevalence was calculated for all parasites in all snail species as the proportion of all infected host individuals in relation to the total number of all host individuals (Op = a inf /S × 100, where a inf is the number of infected snails of all species and S are all snails collected). Species richness was calculated as the total number of trematode species in all host individuals. The relative species richness (S r ) was calculated for parasite assemblages in all infected snail species as a quotient of the number of trematode species in relation to the total number of infected snail host species studied (S r = t/M, where t is the number of detected trematode species and M are all infected snail species sampled). (www.r-proje ct. org)). A χ 2 test followed by a post-hoc pairwise test of independence with Bonferroni correction was applied to compare overall trematode prevalence between sampling months. While a Fisher's exact test was employed to assess differences in the proportions of parasite species with one and two or three intermediate hosts between individual studies.

Data availability
The raw data that support the findings of this study are available from the corresponding author, [J.S.], upon reasonable request.