Our results showed differences in dorsal skin bacterial assemblage structure and diversity between two subspecies of hellbenders from populations not previously sampled. Furthermore, we showed that variation in these communities could be explained by subspecies, severity of chronic lesions in the feet, and partially by life stage. Bd infection status and adult size class (mass) were not correlated with dorsal skin bacterial assemblages. Several factors including both environment and host health may be contributing to these differences. Host populations are separate subspecies from different ecoregions. Furthermore, the ETN eastern hellbender population we sampled is considered healthy and stable and surrounded by forested land (USFWS 2018; Da Silva Neto et al. 2019) whereas the AR Ozark population is in danger of extirpation with little surrounding forest buffer (USFWS 2018).
Distinction between ETN and AR dorsal skin microbial assemblage was greatest via Sorenson Index for average assemblage structure and via lower Hill numbers for alpha diversity. This highlights that subspecies differences were driven by presence of relatively rarer phylotypes. Similar to other studies, we determined that skin microbial assemblage structure of hellbenders differed by subspecies (Hernández-Gómez et al. 2017b, 2018). However, in contrast to those previous studies, we showed eastern hellbenders to have greater richness. We sampled Ozark hellbenders from AR and eastern hellbenders from ETN, whereas previous work focused on populations of both subspecies within Missouri (MO). This is an important distinction for eastern hellbender sampling. While Ozark hellbenders have a small contiguous range within the Ozark highlands of AR and MO only, eastern hellbenders span a large geographic range with eastern MO populations disjunct from all other eastern hellbenders and phylogenetically distinct from ETN (Unger et al. 2013; Hime 2017). Disparate results are further confounded by differential population health within the eastern hellbender subspecies. Eastern hellbenders of MO have been classified as an endangered population alongside the Ozark subspecies whereas the rest of the eastern hellbenders remain unlisted (USFWS 2018).
Interestingly, we found a trend of decreased alpha diversity with increased lesion severity with the apparently unhealthy Ozark hellbender population of AR. This, along with subspecies differences, may suggest that skin bacterial diversity is negatively associated with hellbender skin health. This is in contrast to Hernández-Gómez et al. (2017b) who found toe lesions to have richer assemblages than apparently healthy dorsal skin within a given individual. However, our study differed in that we compared dorsum between individuals of varying toe lesion severity, as opposed to direct comparison between lesioned and lesion-free skin within an individual. Furthermore, along with decreased diversity, we showed skin microbiomes of AR hellbenders, particularly those with severe lesions, had increased beta dispersion. This suggests severe lesions may be associated with uneven, random, and potentially dysbiotic, skin bacterial communities. In contrast to subspecies comparisons, dorsal skin bacterial assemblage differences between AR individuals of varying toe lesion severity were driven by relative abundance of more common phylotypes. A similar relationship was observed between dorsum and toe lesions in Ozark hellbenders of MO in that differences were driven by changes in relative abundance of shared OTUs (Hernández-Gómez et al. 2017b).
These distal limb lesions are progressive and can take several years to increase in severity (Hardman et al. 2020a), and the same factors affecting toe lesions may also be affecting dorsal skin communities. Changes we observed between AR hellbenders of varying toe lesion severity could represent a decreased ability to maintain an optimal microbial community as disease progresses in an individual. Another explanation could be that local microbial changes due to chronic lesions can eventually generate changes over the entire skin surface as has been seen around chronic ulceration due to leishmaniasis in humans (Gimblet et al. 2017). Alternatively, shifts in microbial communities could be the main mechanism by which lesions manifest, and lesion severity represent a sequela of this dysbiosis that manifests specifically in an area of increased use (e.g. toes). The mechanism by which these patterns may arise is still speculative and it is possible these trends are not causal or are potentially spurious from limited sample size within AR (n = 15). More intensive within-river surveys alongside more detailed measurements of health (i.e., blood and immunological parameters, body condition score, and detailed lesion scoring) will be needed to better understand within watershed variation and drivers from host health and environment on changes in skin bacterial community parameters.
Unexpectedly, infection status of Bd had no effect on beta or alpha diversity. This was an unexpected result based on other studies of Bd infection being associated with skin microbial community shifts in amphibians (Jani and Briggs 2014; Muletz-Wolz et al. 2019). Wild hellbenders appear to be tolerant to Bd infection and there are no confirmed reports of chytridiomycosis in a wild hellbender despite the impressive amount of Bd surveillance studies with often 20–30% infection prevalence (Gonynor et al. 2011; Souza et al. 2012; Bales et al. 2015; Seeley et al. 2016; Hardman et al. 2020b). However, Bd may be an important secondary invader during instances of decreased skin immunity as chytridiomycosis is reported in captive and stressful conditions (Dean et al. 2016; Dusick et al. 2017), and Bd infection rates are positively associated with higher lesion scores in wild AR populations (Hardman et al. 2020a). Still, data from this study suggest that even if Bd affects AR hellbender skin health, it is not in association with changes in the skin bacterial community.
Life stage also had minimal effect on the bacterial skin community, at least within the ETN population for which we were able to sample juveniles. Similar to subspecies differences, community membership but not relative abundance, was predictive of hellbender life stage. Our study was limited in the fact that only two juveniles were still gilled larvae (pre-metamorphosis). Evaluation of larval and post-metamorphic juvenile substages within a single site may reveal significant microbiome shifts after metamorphosis we were unable to evaluate.
We found commonalities in indicator taxa that overlapped with those found in previous hellbender studies. The top indicator phylotypes for AR (genera Limnohabitans and Acinetobacter) belonged to families Comamonadaceae and Moralleaceae, respectively. These taxa were previously identified with increased abundance in toe lesions compared to dorsal skin in Ozark hellbenders of MO (Hernández-Gómez et al. 2017b), and further characterized as significantly more abundant in Ozark vs eastern hellbender skin of MO (Hernández-Gómez et al. 2018). The genus Acinetobacter is of particular interest as an indicator taxon because this genus has been previously cultured from Ozark hellbender toe lesions (Nickerson et al. 2011) and is also highly correlated with differential amphibian host responses to Bd infection in both laboratory and field experiments (Bates et al. 2018). Similarly, the only indicator taxon for moderate lesions was an unclassified Burkholderiales, and an OTU in the Burkholderiales was recently shown to increase in abundance on skin of Sierra Nevada yellow-legged frogs, Rana sierrae, experimentally infected with Bd (Ellison et al. 2019).
Some phylotypes may associate with healthy hellbender skin. Several of the indicator taxa that distinguished dorsal skin assemblages of AR animals with mild versus moderate or severe toe lesions were the same that distinguished ETN from AR. One particular (genus Ferruginibacter) has also been identified as a core taxon on alpine newt (Ichthyosaura alpestris) skin (Bletz et al. 2017b).
No indicator phylotype for ETN eastern hellbenders from this study overlapped with taxa previously identified as significantly associated with eastern hellbender skin. As mentioned above, our study evaluated eastern hellbenders from ETN whereas the previous studies comparing hellbender subspecies evaluated those from MO. Populations from ETN and MO vary greatly in phylogeny, ecoregion, and population health. Conversely, AR Ozark hellbenders are phylogenetically very similar to MO Ozark hellbenders and individuals sampled from our study were from one of the same watersheds sampled in these previous studies. It is no surprise then that our indicator species analysis revealed similar results for Ozark hellbenders yet disparate results for eastern hellbenders. The importance of specific taxa of bacteria and their function within a community context remains an understudied area in amphibian microbiome research. This study contributes to our understanding of the microbial assemblages of hellbenders and their correlation with skin health.