Skip to main content
Download PDF

Signatures of functional bacteriome structure in a tropical direct-developing amphibian species

Animal Microbiome volume 4, Article number: 40 (2022) Cite this article

Abstract

Background

Host microbiomes may differ under the same environmental conditions and these differences may influence susceptibility to infection. Amphibians are ideal for comparing microbiomes in the context of disease defense because hundreds of species face infection with the skin-invading microbe Batrachochytrium dendrobatidis (Bd), and species richness of host communities, including their skin bacteria (bacteriome), may be exceptionally high. We conducted a landscape-scale Bd survey of six co-occurring amphibian species in Brazil’s Atlantic Forest. To test the bacteriome as a driver of differential Bd prevalence, we compared bacteriome composition and co-occurrence network structure among the six focal host species.

Results

Intensive sampling yielded divergent Bd prevalence in two ecologically similar terrestrial-breeding species, a group with historically low Bd resistance. Specifically, we detected the highest Bd prevalence in Ischnocnema henselii but no Bd detections in Haddadus binotatus. Haddadus binotatus carried the highest bacteriome alpha and common core diversity, and a modular network partitioned by negative co-occurrences, characteristics associated with community stability and competitive interactions that could inhibit Bd colonization.

Conclusions

Our findings suggest that community structure of the bacteriome might drive Bd resistance in H. binotatus, which could guide microbiome manipulation as a conservation strategy to protect diverse radiations of direct-developing species from Bd-induced population collapses.

Background

One of the most fundamental questions in disease ecology is: in a given host-microbe encounter, why do some individuals get sick while others do not? When germ theory gained acceptance in the late nineteenth century, research centered on highly transmissible, toxin-producing bacteria that consistently caused disease (e.g., diphtheria and cholera), leading to the microbe-centric view that virulence is a fixed microbial trait [1]. The role of the host gained traction in the twentieth century when large-scale disease emergences in immune-impaired human patients were attributed to microbes considered to be commensal. For example, widespread use of antibiotics increased oral candidiasis by the common fungal gut microbe Candida albicans [2] and patients immunocompromised from HIV were more susceptible to pneumococcal pneumonia from the common respiratory tract bacteria Streptococcus pneumoniae [3]. However, even this more host-centric view failed to account for a role of the host in disease outcomes in apparently healthy patients [4].

To better integrate microbe and host contributions to disease, the ‘damage-response framework’ of microbial pathogenicity rejects defining microbes as pathogens, commensals, or opportunists, instead defining only microbes and hosts which interact [5]. Microbial pathogenicity is then characterized by placing these context-dependent interactions on a continuum of host damage arising from the microbe, the host response, or both, and ranging from beneficial or neutral interactions to a level of cell, tissue, or organ damage that disrupts homeostasis, causing disease [6]. Eleven interrelated factors governing the outcome of these host-microbe interactions have been proposed, with the first letter of each factor assembling into the acronym MISTEACHING to underscore the rapidly evolving and complex nature of the field: microbiome, inoculum, sex, temperature, environment, age, chance, history, immunity, nutrition, and genetics [7]. While this list was originally proposed for human hosts, it is also applicable to other vertebrates.

As new approaches for studying microbes have exposed the host microbiome as a central player in host physiology, immunity, and health [8], the damage-response framework has been revised to consider the host component as a new entity encompassing the host and its microbiota [7]. Assembly of the host microbiome is a combination of deterministic and stochastic processes and can be viewed through a metacommunity lens: at the scale of the host individual, communities are influenced by host behavior and physiology, including immunity and chemical conditions, with individual communities linked through dispersal among conspecifics, heterospecifics, and free-living microbial communities, all against the backdrop of abiotic environmental conditions [9, 10]. In turn, host-associated microbes may mediate host interactions with invading microbes, and subsequent host damage, by stimulating the host immune system, competing with microbial intruders, or producing bioactive compounds [11].

Amphibians are an ideal system to study the spectrum of host responses to a given microbial species. Not only are they the most threatened vertebrate taxon [12], but they are also hosts to the skin-infecting chytrid fungus Batrachochytrium dendrobatidis (Bd), a microbe with a deadly combination of the greatest known vertebrate host breadth and, in hundreds of species, a high capacity for host damage [13]. The amphibian skin tissue that Bd infects is highly biologically active. In addition to the skin’s role in breathing and homeostasis, it also produces mucus secretions and chemical compounds that create distinct microhabitats for diverse host-associated bacterial communities (bacteriome) [14] known to influence host-Bd interactions [15,16,17,18,19]. Extreme variation in Bd pathogenicity has been reported across hundreds of tested host species, with many examples of host bacteriome composition correlating with Bd pathogenicity [20, 21]. For example, inoculations with Bd-inhibitory bacteria or bacterial metabolites increased survival in mountain yellow-legged frogs Rana muscosa [22], red-backed salamanders Plethodon cinereus [23] and boreal toads Anaxyrus boreas [24], and higher skin bacterial diversity was associated with lower Bd loads in the coqui frog Eleutherodactylus coqui [25].

Most other documented host contributors to Bd pathogenicity match the attributes of MISTEACHING: temperature [26,27,28], environmental contamination [29, 30], life stage [31, 32], behavior [33,34,35], acquired immunity [36], and innate immunity including MHC class II expression and genetic diversity [37, 38] and antimicrobial skin peptides [18, 39], with many of these factors interrelated with the bacteriome. An additional factor especially relevant to amphibians is host community structure. Amphibian communities can reach high levels of species richness relative to other terrestrial vertebrates, especially in the tropics, and often exhibit extreme within-community variation in life history strategies and other traits that may result in widely divergent host-microbe interactions, both from the perspective of the host bacteriome and invading microbes [40,41,42].

Brazil’s Atlantic Forest supports a megadiverse fauna of amphibians [43] and Bd is generally widespread in the region [44]. As an aquatic pathogen, Bd is easily transmitted among aquatic-breeding frogs during seasonal breeding aggregations in water bodies [45,46,47]. In contrast, terrestrial-breeding species exhibiting direct development (i.e., without a free-living, aquatic larval stage; individuals hatch as froglets from terrestrial eggs) are rarely exposed to water bodies, in many cases translating to low immunity against Bd, but may encounter the pathogen sporadically through spillover from aquatic-breeding species or in response to climate anomalies [40, 48]. These terrestrial host-Bd interactions typically score high on the host damage continuum including high infection loads and mortality rates [48, 49]. We investigated Bd infection patterns in six Atlantic Forest anuran species. Four species, Boana faber (Hylidae), Bokermannohyla circumdata (Hylidae), Bokermannohyla hylax (Hylidae), and Rhinella ornata (Bufonidae) are aquatic-breeding, and two species, Haddadus binotatus (Craugastoridae) and Ischnocnema henselii (Brachycephalidae), are terrestrial-breeding. While we expected similar Bd infection patterns in the two ecologically similar terrestrial-breeding species, intensive sampling of 40 natural forest sites across eight landscapes yielded no detections of Bd-positive individuals of H. binotatus but substantial levels of infection in the other species, including relatively high Bd prevalence in Ischnocnema henselii (Fig. 1). Previous community-wide Bd surveys suggest that in habitats where sympatric species are infected, consistent non-detection of Bd in a single host species after exhaustive sampling is rare [40]. In addition, previous studies on Atlantic Forest amphibian assemblages indicate that the two terrestrial-breeding species experience similar probability of Bd exposure through spillover from aquatic-breeding species [48]. Together, these findings suggest that host factors such as the bacteriome may be the primary drivers of differential host damage between H. binotatus and I. henselii.

Fig. 1
figure 1

Occurrence of six focal amphibian species (shown at top left) in eight focal landscapes (circles of 15-km diameter) and 40 sample sites (five red points per landscape) in the Atlantic Forest in the State of São Paulo, southeastern Brazil. Colored bars within circles indicate species sampled within each landscape. Barplots show relative Batrachochytrium dendrobatidis prevalence among species observed at each landscape. Landscapes are (1) Ribeirão Branco, (2) Intervales State Park, (3) Pilar do Sul, (4) Cotia, (5) Serra do Japi, (6) São Luíz do Paraitinga, (7) Serra do Mar State Park—Núcleo Santa Virgínia, and (8) Bananal

Full size image

To test the hypothesis that the observed variation in Bd prevalence was associated with host bacterial communities, we compared bacteriome composition and structure among the six focal amphibian species, including alpha and beta diversity, a metric of dysbiosis known as dispersion, and co-occurrence network topology. If bacteriomes drive variation in host damage in this system, we would expect distinct bacteriome metrics in H. binotatus compared to the other species, especially with regard to characteristics previously associated with bacteriome stability and pathogen inhibition such as high microbial diversity [50], low dispersion [51], and strongly nonrandom network topology [52, 53]. Combined, our results shed new light on the potential role of bacteriome structure in microbial defense in historically understudied tropical amphibians.

Results

Bd prevalence and infection loads

Bd prevalence varied significantly by species (X2 = 103.476, n = 903, p < 0.0001; Fig. 2A). We did not detect Bd in any of the 126 sampled H. binotatus individuals. Bd prevalence was 34.8% (n = 204) in I. henselii, 5.4% (n = 112) in R. ornata, 13.8% (n = 356) in Bo. faber, 7.0% (n = 43) in Bk. circumdata, and 12.7% (n = 63) in Bk. hylax (Fig. 2A). Prevalence patterns among the full dataset (n = 903) were similar to patterns for the subset of samples (n = 666) analyzed after rarefying bacteriome data to account for unequal sequencing depths among samples (X2 = 101.599, n = 666, p < 0.0001; Additional file 1: Table S1). I. henselii had an average (± standard error) Bd load of 246.9 ± 91.3 ITS copies, R. ornata 24.8 ± 15.5 ITS copies, Bo. faber 145.0 ± 46.0 ITS copies, Bk. circumdata 738.0 ± 612.9 ITS copies, and Bk. hylax 2,441.3 ± 2,379.7 ITS copies, but these averages did not differ significantly (F = 1.309(4,132), R2 = 0.038; p = 0.270).

Fig. 2
figure 2

For six anuran species sampled in Brazil’s Atlantic Forest, A average prevalence of Batrachochytrium dendrobatidis (± standard error) across all samples, and violin plots showing likelihood of different B numbers of bacterial sub-operational taxonomic units (sOTU) detected on the skin (after rarefying sequence reads) and C levels of skin bacterial dispersion (distance to group centroid) based on unweighted UniFrac distances. Bars on violin plots indicate mean ± standard error

Full size image

Patterns of skin bacterial alpha diversity, composition and dispersion

Bacteriome alpha diversity metrics differed among host species: Shannon index (F(5,548.7) = 36.12, R2 = 0.24, p < 0.0001), Faith’s phylogenetic diversity (F(5,598.5) = 48.66, R2 = 0.30, p < 0.0001), and sOTU richness (F(5,501.7) = 55.10, R2 = 0.32, p < 0.0001; Fig. 2B). Haddadus binotatus had higher bacteriome alpha diversity in all pairwise comparisons (Additional file 1: Table S2).

Bacteriome beta diversity metrics also differed among species, both in terms of sOTU presence/absence (unweighted UniFrac PermANOVA: F(4,660) = 16.82, R2 = 0.11, p = 0.001; Fig. 3) and relative abundance (weighted UniFrac PermANOVA: F(4,660) = 49.89, R2 = 0.27, p = 0.001; Additional file 1: Fig. S1; Additional file 1: Table S3). The two terrestrial-breeding species and the toad species carried distinct skin communities while the three treefrog species hosted more similar communities (Fig. 3; Additional file 1: Table S3). Skin bacterial community composition did not differ between Bd-positive and Bd-negative I. henselii (unweighted UniFrac permANOVA: F(1, 180) = 0.996, R2 = 0.006, p = 0.452; weighted UniFrac permANOVA: F(1, 180) = 0.779, R2 = 0.004, p = 0.502).

Fig. 3
figure 3

Community composition of host skin bacteria calculated using principal coordinates analysis based on unweighted UniFrac distances. Small circles indicate data points. Large circles indicate group centroids. Lines indicate distance to group centroids

Full size image

Differential abundance analysis (Linear discriminant analysis Effect Size; LEfSe) revealed 87 differentially abundant sOTUs across the six landscapes where H. binotatus were sampled (Additional file 1: Fig. S2). Of these sOTUs, 65 (75%) were differentially abundant in H. binotatus in at least one landscape (Additional file 1: Fig. S2). Haddadus binotatus carried the highest number of differentially abundant sOTUs at all but one landscape where it was sampled (Additional file 1: Table S4). Four soil-associated sOTUs were enriched in H. binotatus at three landscapes, including Devosia sp. (Hyphomicrobiaceae), Aminobacter sp. (Phyllobacteriaceae), Agrobacterium sp. (Rhizobiaceae), and Variovorax sp. (Comamonadaceae; Additional file 1: Fig. S2; Additional file 1: Table S4).

Skin bacterial dispersion differed among host species (unweighted UniFrac: F(5,122.1) = 33.02, R2 = 0.19, p < 0.0001; weighted UniFrac: F(5,609.3) = 38.47, R2 = 0.24, p < 0.0001; Jaccard: F(5, 553.8) = 56.42; R2 = 0.32; p < 0.0001; Bray–Curtis: F(5, 652.6) = 40.76, R2 = 0.27, p < 0.0001; Additional file 1: Tables S5, S6). Haddadus binotatus had the lowest values of unweighted Unifrac (Fig. 2C) and Jaccard (Additional file 1: Fig. S3A) dispersion. Patterns of weighted UniFrac (Additional file 1: Fig. S3B) and Bray–Curtis (Additional file 1: Fig. S3C) dispersion were less consistent (Additional file 1: Tables S5, S6).

Skin bacterial communities across all six host species were dominated by taxa in the phyla Proteobacteria, Actinobacteria, and Bacteroidetes (Additional file 1: Fig. S4). More than half of detected sOTUs (58.9%) were shared by a proportion of individuals of all 6 host species and only 3.7% of sOTUs were unique to one host. Haddadus binotatus carried only two unique sOTUs (Additional file 1: Table S7) but hosted the highest diversity of common core (detected in at least 90% of individuals) sOTUs (12), compared to one common core sOTU for I. henselii, three common core sOTUs for R. ornata, two common core sOTUs for Bk. hylax, one common core sOTU for Bk. circumdata, and no common core sOTUs identified for Bo. faber (Additional file 1: Table S8). Most of the common core sOTUs for H. binotatus were also differentially abundant in this species (Additional file 1: Table S8). None of the common core sOTUs detected in H. binotatus were detected in the common core bacteriome of the other host species, except for an Agrobacterium sp. which was also detected in the common core bacteriome of R. ornata (Additional file 1: Table S8).

Co-occurrence networks

To assess community interactions within host bacteriomes, we used co-occurrence networks based on significant pairwise correlations (both positive and negative) among sOTUs. Overall, bacteriome co-occurrence network topology in Serra do Japi, the best-sampled landscape, varied among H. binotatus, I. henselii, R. ornata, and Bo. faber (Table 1). To compare network topologies, we used a set of network metrics to characterize the interconnectedness (density, average path length, diameter, and average degree), clustering (average clustering coefficient and modularity), and centrality (betweenness centrality) of sOTUs within co-occurrence networks (Additional file 1: Methods). These comparisons revealed that observed networks were more interconnected (2- to fourfold higher diameters and average path lengths) and clustered (20- to 80-fold higher average clustering coefficients) than expected based on random networks. The exception to this pattern was the modularity metric (i.e., presence of correlated subsets [modules] of sOTUs), which was strikingly lower than expected for I. henselii, but four to six times higher than expected for the other host species. The proportion of positive correlations in the networks was high across all host species, ranging from 98% (H. binotatus) to 100% (Bo. faber). Average path length (a metric of network interconnectedness) was similar across all four host networks.

Table 1 Topology metrics of similarity-based bacterial networks using Spearman correlations (coefficient ρ > 0.6 and < − 0.6; P ≤ 0.01)
Full size table

Haddadus binotatus, the species harboring no detectable Bd, had the smallest bacterial network in number of nodes (sOTUs) and edges (correlations). This network was relatively sparsely connected (based on average degree), and highly clustered (based on average clustering coefficient and modularity; Table 1). This network was partitioned into two main modules connected by a subset of nodes and primarily separated by negative edges, a distinct clustering pattern not observed in the other host species (Fig. 4A).

Fig. 4
figure 4

Correlation-based bacterial networks for the Serra do Japi landscape using Spearman correlations (coefficient ρ > 0.6 and < − 0.6; p ≤ 0.01) and the Fruchterman-Reingold layout. Colored points indicate nodes (sOTUs) and lines indicate edges (positive or negative correlations). Nodes (sOTUs) are color-coded based on bacterial phylum or class (in the case of Proteobacteria). Square-shaped nodes represent common core sOTUs (90%). In the Haddadus binotatus network, the common core sOTU classified as Alphaproteobacteria and shown at the top left of the network (marked with a black star) ranked highest in betweenness centrality. Edge lengths are a function of layout and are not biologically meaningful

Full size image

In I. henselii, the species that harbored the highest Bd prevalence, network topology ranked low in measures of both overall interconnectedness and clustering. This network showed the lowest diameter, average clustering coefficient and modularity (Table 1), visualized as one indistinct cluster of nodes (Fig. 4B). R. ornata had the largest and most interconnected network with the highest average degree and density, but intermediate clustering (Fig. 4C; Table 1). The Bo. faber network ranked highest in modularity, though it differed from H. binotatus in the lack of negative edges and high number of small modules at the network periphery (Fig. 4D; Table 1).

Of the 12 common core sOTUs detected in H. binotatus, 9 sOTUs were included in its network, one of which (Paracoccus marcusli from the class Alphaproteobacteria) presented the highest betweenness centrality value (i.e., degree to which the positioning of an sOTU bridges different parts of the network) in the entire network (and 22°, Fig. 4A). Most common core sOTUs for I. henselii and R. ornata were included in their networks but did not present high centrality values.

Discussion

We conducted the first comparative bacteriome study of terrestrial and aquatic-breeding frogs in Brazil’s Atlantic Forest, with high spatial replication (40 sampling sites across eight landscapes) and robust landscape-level spatial coverage (400 km). Strikingly, Bd was not detected in more than 100 sampled individuals of the terrestrial-breeding species H. binotatus. In contrast, we detected 35% Bd prevalence among 204 individuals of the ecologically similar, terrestrial-breeding species I. henselii, 7–14% Bd prevalence among 462 individuals of three aquatic-breeding treefrog species, and 5% Bd prevalence among 112 individuals of the aquatic-breeding toad Rhinella ornata (Fig. 2A). Previous work on Bd dynamics in Brazilian terrestrial-breeding frogs suggests low Bd resistance in these species [49] and substantial exposure to Bd in the wild [44, 54]. For example, Brachycephalus ephippium, another terrestrial-breeding frog species endemic to Brazil’s Atlantic Forest with similar habitat use as our focal terrestrial species, was exposed to substantial levels of Bd in terrestrial habitats through spillover from aquatic-breeding species during the breeding season [48]. Thus, the contrasting Bd occurrence patterns we observed between H. binotatus and I. henselii may be more strongly driven by host factors than environmental or pathogen factors, prompting a comparative skin bacteriome analysis of an ecologically diverse assemblage of amphibian species, some for the first time.

In contrast to consistent non-detections of Bd in H. binotatus, the highest Bd prevalence in our study was found in I. henselii, the only other species in our study with a terrestrial-breeding life history strategy, and which can be found side by side with H. binotatus in the field. While the skin bacteriomes of Atlantic Forest treefrog species tracked bioclimatic variables at large spatial scales [55], the finer spatial resolution of our study revealed significant species-level variation in host bacteriome attributes under constant biogeographic and climatic conditions, consistent with previous work showing differences in the skin bacterial communities of co-occurring amphibian species [56,57,58,59]. Compared to I. henselii, H. binotatus carried significantly higher overall bacterial alpha diversity (Fig. 2B). High alpha diversity could enhance Bd inhibition if diverse microbial communities have a greater chance of containing Bd-suppressing taxa and these taxa are able to become dominant [50, 60]. However, I. henselii carried the second highest bacteriome alpha diversity of the six species evaluated (Fig. 2B), suggesting that overall bacterial diversity on its own does not explain differential Bd prevalence among the focal host assemblage. Similarly, H. binotatus scored the lowest in relative bacterial dispersion but I. henselii had the second lowest dispersion (Fig. 2C), similarly suggesting that the dispersion metric alone does not fully explain host-Bd interactions in this system.

The most contrasting bacteriome characteristics between H. binotatus and I. henselii were common core diversity and co-occurrence network structure. While we detected only one common core sOTU in I. henselii, the common core bacteriome of H. binotatus contained 12 sOTUs. Species richness of common core microbial communities, defined as the taxa occurring above an occupancy frequency threshold, is not necessarily linked to bacteriome function or evolutionary history with the host [61]. Certain taxa can have high colonization frequency within host populations for many reasons, including high competitive ability or widespread distribution in the environment, and certain taxa may play important functional roles in host populations even when rare [61]. As a result of these findings, recent work has differentiated the common core from other types of core microbiomes including the temporal core, ecological core, functional core, and host-adapted core [61]. Co-occurrence network analysis uses centrality metrics to identify sOTUs that may play an important role in structuring the microbiome, paralleling the concept of the ecological core [61,62,63]. Aside from high relative diversity, the common core bacteriome of H. binotatus included the sOTU with the highest betweenness centrality score in the co-occurrence network for this species (black star in Fig. 4A), suggesting a highly connected position that could stabilize and regulate community structure and occurrence of taxa with beneficial attributes [62, 63]. Our results support recent perspectives on core microbial communities as good candidates from which to build on our understanding of host-microbe interactions and bacteriome function [61].

A conspicuous attribute of the I. henselii network was low modularity, both compared to other observed host networks and randomly generated networks (Fig. 4B). In contrast, the H. binotatus network had a highly clustered structure with two main modules joined by a subset of sOTUs and distinct partitioning between positive and negative sOTU frequency correlations (Fig. 4A). This partitioning is unlikely to be a geographical artifact because our bacteriome networks represent frogs from a single landscape. Network modularity reveals clusters of associated species that cohabitate, potentially indicating niche partitioning mediated by competition, mutualism, or host microhabitat characteristics [53, 64]. In contrast, the central nodes joining the two modules could be more generalist taxa. Across different types of ecological networks, modularity has been associated with network function and robustness [64,65,66,67,68]. For instance, high modularity could increase efficiency in sharing of metabolites across the network, reduce ripple effects of network perturbations, or promote functional redundancy to guard against failure of certain metabolic reactions [69, 70]. Our findings suggest a high level of network organization in H. binotatus, which could indicate functional communities with patterns of assembly and ecological interactions that happen to be effective against Bd invasion, even under the probable scenario that Bd is not a prominent selective force on bacteriome structure in terrestrially associated frogs [9, 21, 53].

Specifically, the highly structured negative correlations among sOTUs in the H. binotatus network (blue lines in Fig. 4A) could indicate a bacteriome primed for antagonistic interactions and thus more resilient to Bd invasion [52, 71], consistent with the finding that modules of marine sponge bacteriomes were associated with production of bioactive secondary metabolites involved in chemical defense [70]. Similar to H. binotatus, Bd prevalence is typically low or absent in North American eastern red-backed salamanders Plethodon cinereus, a species that carries high concentrations of violacein, a strongly antifungal secondary metabolite of the dominant skin bacterium Janthinobacterium lividum [72,73,74,75,76,77]; thus, a comparison of modularity in the bacteriomes of these two species would be useful. The H. binotatus network also stood apart in its relative sparseness with the fewest nodes, edges, and edges per node, which contrasts with the high overall taxonomic diversity of the bacteriome. Together, these results suggest that high overall bacteriome alpha diversity could function as a rich bacterial reservoir facilitating the formation of a highly structured and synergistic subset of member taxa, consistent with the finding that synergistic interactions among diverse microbial assemblages could boost the Bd-inhibitory functions of the community [50].

Compared to the focal terrestrial-breeding species, the focal aquatic-breeding treefrog species carried intermediate Bd prevalence, low overall bacteriome alpha diversity, high bacterial dispersion, and scant to absent common cores (0–2 sOTUs). The most distinctive feature of the Bo. faber co-occurrence network was a high number of small node clusters loosely connected to the overall network. This finding suggests a substantial level of bacterial community variability among individuals, consistent with variable microhabitat use and potentially helping to explain the high dispersion and few common core microbes characteristic of the generalist treefrog species. The Bo. faber network also lacked negative connections, suggesting limited antagonistic activity, such as production of anti-microbial metabolites that could protect against microbial invasion [76].

The focal aquatic-breeding toad R. ornata carried relatively low Bd prevalence consistent with previous surveillance [54], intermediate bacteriome alpha diversity and dispersion, and three common core sOTUs. The R. ornata network was highly interconnected, scoring twice as high as other species in overall network density and average number of connections per sOTU, suggesting an organized structure with complex paths of communication. Thus, acknowledging the limitations of our correlational results and the vast life history differences among our focal host species, one consistency across life history guilds was lower Bd prevalence in species with higher bacteriome network structure (H. binotatus and R. ornata) and higher Bd prevalence in species with less defined network structure (I. henselii and Bo. faber). This pattern is consistent with the previous finding, in a comparison of three Panamanian amphibian species, that the more Bd-resistant species had more bacteriome network connections and greater clustering than the more Bd-vulnerable species [78]. Similarly, the bacteriome of Bd-vulnerable species were more similar to environmental bacterial pools compared to Bd-resistant species [58], suggesting that skin communities formed mainly by neutral processes may be less protective than more structured communities formed by deterministic processes.

While the distinctive skin bacteriome of H. binotatus is compelling as a mechanism explaining the lack of Bd detected in this species, not all bacteriome metrics that we measured were correlated with Bd infection (e.g., sOTU richness, dispersion) and, thus, it is critical to consider other possible non-mutually exclusive factors that could contribute to this pattern. Studies on the physiology of H. binotatus are extremely scarce and thus we have limited ability to assess skin structure and other physiological factors that could contribute to Bd resistance, but distinctive characteristics of the skin such as sloughing rate have been associated with Bd resistance in other species [79]. Behavioral studies are also rare. Both I. henselii and H. binotatus were observed in forested areas near and away from streams. During dryer months, I. henselii may be observed very close to streams and occasionally submerged in water [80], whereas H. binotatus has been observed in holes and under debris during dryer periods in other studies, possibly because they are larger-bodied and more tolerant of dry conditions [81]. These divergent strategies for overcoming hydric stress may lead to different encounter rates with environmental Bd reservoirs (streams) and environmental bacterial reservoirs. However, our study took place in the rainy season when frogs were unlikely to be experiencing hydric stress and both terrestrial-breeding species were likely exposed to similar Bd and bacterial reservoirs within the same terrestrial microhabitats.

The bacteriome is only a portion of host-associated microbial communities, with protists, multicellular fungi, viruses, archaea, and microscopic metazoans also often in great abundances [82,83,84,85]. Microbial assemblages also exist in a complex mucosal substrate that includes host-produced antimicrobial peptides [86, 87]. These additional components of host skin communities interact with host-associated bacteria [84, 85] and may also independently interact with invading microbes [86, 87], suggesting that analysis at the level of the entire chemical and biological mucosal matrix may be necessary to fully evaluate Bd resistance mechanisms in H. binotatus. Immunogenetic variation is another factor that could not only help to explain host variation in microbiome assembly [85, 88, 89], but could also contribute to Bd resistance independently of the microbiome [37, 38, 90, 91]. Thus, immunogenetic analysis of our focal species could reveal a more complete mechanistic picture of host variation in Bd resistance.

We cannot rule out the possibility that H. binotatus possesses low resistance and tolerance to Bd and thus succumbs to chytridiomycosis extremely quickly, to an extent that lowers the chances of detecting infected individuals. This scenario seems unlikely, however, because we would expect to find a small number of individuals with very high infection loads over the course of our intensive sampling effort. Variation in landscape characteristics is also unlikely to explain the absence of detectable Bd in our H. binotatus samples because the pattern was consistent across all six landscapes where H. binotatus were found. Similarly, variation in host reproductive strategy and phylogenetic classification are unlikely explanations because H. binotatus and I. henselii are phylogenetically similar, both belonging to the Brachycephaloidea superfamily, and both exhibit direct development. I. henselii typically exhibit higher population densities than H. binotatus, which could increase disease transfer between conspecifics. However, H. binotatus were also generally abundant at our sample sites, so this probably does not explain the lack of Bd found in H. binotatus. Haddadus binotatus and I. henselii generally share the same forested habitats and activity patterns. However, we observed I. henselii vocalizing on the ground as well as within bromeliads and on leaves, branches, and trunks up to 1.5 m above the ground, consistent with the few previous natural history observations for the species [80]. In contrast, we never observed H. binotatus in bromeliads or vocalizing at heights greater than 20 cm from the ground, although they have occasionally been observed up to 80 cm above the ground [92]. Additional research is needed to characterize the habitat use of these poorly studied species more thoroughly, but if H. binotatus and I. henselii come in contact with certain substrates at different rates, this could lead to different patterns of bacterial recruitment.

Conclusion

Direct-developing frogs have been found to be at high risk of population declines caused by climate change and chytridiomycosis [28, 40, 49, 93]; however, our study has shown evidence of Bd resistance in the direct-developing species H. binotatus. The bacteriomes of H. binotatus were highly diverse and structured, most notably with high common core taxonomic diversity and a tightly clustered network topology partitioned by negative correlations, characteristics that may be associated with community stability and competitive interactions that could inhibit Bd colonization. Together, our results indicate that bacteriome community structure could be contributing to Bd resistance in H. binotatus, although further research is needed to rule out other resistance factors, especially because our study represents the first analysis of the H. binotatus microbiome and natural history information on this species is scarce. Continued research on the mechanisms that prevent infection in H. binotatus may prove to be beneficial in conservation efforts to protect diverse radiations of cryptic direct-developing species from Bd-induced population declines and extinctions. This includes culturing and challenging H. binotatus skin bacteria with Bd to determine the influence of key taxa, or networks of taxa, on Bd resistance; studies on behavioral, genetic, and mucosomal drivers of host-associated bacterial composition; genome-resolved metagenomics and metatranscriptomics to elucidate functional roles of the bacteriome in Bd resistance [94]; and manipulation experiments to detect keystone bacteriome members and probiotic assemblages.

Methods

Study location and focal species

We sampled eight landscapes (circular areas 15 km in diameter) in the Atlantic Forest in the state of São Paulo, Brazil (Fig. 1). Within each landscape, we sampled five natural forest sites (Fig. 1). We conducted field sampling during the breeding season from September 2018 to January 2019, visiting each of the 40 sites twice within that time.

We sampled six common species (Fig. 1) comprising two modes of reproduction. Four species are aquatic-breeding, including three arboreal treefrog species, Boana faber, Bokermannohyla circumdata, and Bk. hylax, and the toad Rhinella ornata. Two species, Haddadus binotatus and Ischnocnema henselii, are terrestrial-breeding. All of these species are endemic to the Brazilian Atlantic Forest [43]. Frogs were captured primarily in terrestrial leaf litter and vegetation, rarely within water bodies.

Across all sites, we captured 903 individual frogs: 356 Bo. faber, 42 Bk. circumdata, 63 Bk. hylax, 126 H. binotatus, 204 I. henselii, and 112 R. ornata (Additional file 1: Table S1). We captured each individual frog with fresh gloves, rinsed the skin surface with sterile distilled water to remove transient bacteria, and collected skin swabs following Hyatt et al. [95]. We stored swabs in sterile tubes on ice in the field and transferred samples to -20 ºC until DNA extraction (Qiagen DNeasy; Additional file 1: Methods). After swabbing, we released all individuals at the point of capture.

Bd and bacteriome characterization

To quantify Bd loads, we followed qPCR protocols by Boyle et al. [96]. For qPCRs, we diluted an aliquot of each extracted DNA sample 1:10 and used primers to amplify the ITS and 5.8S rRNA regions of Bd, including synthetic standards ranging from 102 to 106 gene copies [96].

To characterize frog skin bacterial communities, we followed the Earth Microbiome Project 16S Illumina Amplicon Protocol, targeting the V4 region of the bacterial 16S rRNA gene and including negative controls without template DNA (Additional file 1: Methods) [97, 98]. We used Quantitative Insights into Microbial Ecology (QIIME 2) v. 2019.1 for initial processing of bacterial sequences (Additional file 1: Methods). We clustered amplicon sequences into sub-operational taxonomic units (sOTUs) using Deblur [99], discarded sOTUs with fewer than 0.005% of the total sequence reads in the dataset [100] and rarefied samples to 2,000 reads to standardize read counts across samples (Additional file 1: Methods).

For analyses of alpha diversity, we calculated sOTU richness, Shannon index, and Faith’s phylogenetic diversity for each sample. We also identified common core sOTUs for each host species, defined as the set of sOTUs detected in at least 90% of sampled individuals, independent of relative abundance.

For analyses of beta diversity, we calculated Jaccard, Bray–Curtis, and unweighted and weighted UniFrac distances between samples [101, 102]. To quantify bacterial dysbiosis, referring to imbalance that may disrupt healthy functioning, we used the metric known as dispersion. In many cases, dysbiosis occurs through stochastic changes in the microbiome, leading to increased microbial variability among hosts [103]. Thus, dispersion, a measure of microbial variability, is routinely used as a measure of dysbiosis [103,104,105]. We used Jaccard, Bray–Curtis, and unweighted and weighted UniFrac distances between samples to calculate bacterial dispersion for each sample as an index of dysbiosis (betadisper function in vegan package in Program R). Dispersion is estimated by reducing original distances to principal coordinates and calculating the non-Euclidean distance between each object and the group centroid. We used amphibian species as the grouping variable.

Statistical and network analyses

We compared infection prevalence among amphibian species using a Generalized Linear Model (GLM) with binomial distribution and logit link. We used General Linear Mixed Models (GLMMs; standard least squares) to compare bacterial alpha diversity indices and bacteriome dispersion among amphibian species. Study landscape (1–8) and site (1–5) within landscape were included as random effects. We performed Tukey a posteriori tests for each pairwise comparison. We compared average Bd infection loads (log transformed; only Bd-positive individuals) among amphibian species using a similar GLM approach. We used permutational analysis of variance (permANOVA) to compare skin bacterial composition among species (Additional file 1: Methods). To test for effects of Bd infection status on microbiome composition, we used permANOVAs to compare skin bacterial composition (weighted and unweighted UniFrac distance) between Bd-positive (n = 66) and Bd-negative (n = 116) I. henselii, the focal species with the most robust sample sizes of Bd-positive and Bd-negative individuals. To identify differentially abundant sOTUs among host species for each study landscape, we used the linear discriminant analysis (LDA) effect size (LEfSe) method [106](Additional file 1: Methods).

To compare patterns of sOTU co-occurrence among host skin bacterial communities, we computed correlation-based networks (Additional file 1: Methods). In these networks, each node represents an sOTU and edges between nodes represent highly significant pairwise correlations, which can be inferred as co-occurrences (when ρ is positive) or antagonistic interactions (when ρ is negative). We characterized network structure from the following metrics: density, average shortest path length, diameter, average degree, average clustering coefficient, modularity, and node betweenness centrality (Additional file 1: Methods). We limited this analysis to a single landscape to reduce effects of confounding factors associated with variation between landscapes. Thus, we selected Serra do Japi, the landscape with the largest sample size (n = 120 individuals) of the eight focal landscapes. Sample sizes at this landscape were well-balanced between terrestrial-breeding (n = 21 H. binotatus, n = 43 I. henselii) and aquatic-breeding (n = 24 Bo. faber, n = 26 R. ornata) host species. Bk. hylax (n = 5) and Bk. circumdata (n = 1) were rarely sampled at this landscape and were excluded. Bacteriome diversity and Bd infection prevalence at this landscape followed trends for the eight landscapes combined (Table 1; Additional file 1: Table S1).

Availability of data and materials

The datasets generated and analysed during the current study are available from the corresponding authors on reasonable request.

Abbreviations

Bd:

Batrachochytrium dendrobatidis

GLM:

Generalized Linear Model

GLMM:

General Linear Mixed Models

LEfSe:

Linear discriminant analysis (LDA) effect size

permANOVA:

Permutation analysis of variance

QIIME:

Quantitative Insights into Microbial Ecology

sOTU:

Sub-operational taxonomic unit

References

  1. Johnson DI. Bacterial virulence factors. Bacterial pathogens and their virulence factors. Berlin: Springer; 2018. p. 1–38.

    Book  Google Scholar 

  2. Homei A, Worboys M. Candida: a disease of antibiotics. Fungal disease in Britain and the United States 1850–2000 mycoses and modernity. Berlin: Springer; 2013.

    Google Scholar 

  3. Megarbane B, Bruneel F, Chevret S, Thuong M, Wolff M, Regnier B, et al. Severe community-acquired bacterial pneumonia from Streptococcus pneumonia in HIV-infected patients: epidemiology and prognostic features of mortality. Pathol Biol (Paris). 1999;47:422–9.

    CAS  Google Scholar 

  4. Casadevall A, Pirofski LA. Host-pathogen interactions: the attributes of virulence. J Infect Dis. 2001;184:337–44.

    Article  CAS  PubMed  Google Scholar 

  5. Casadevall A, Pirofski LA. The damage-response framework of microbial pathogenesis. Nat Rev Microbiol. 2003;1:17–24.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Casadevall A, Pirofski LA. Host-pathogen interactions: basic concepts of microbial commensalism, colonization, infection, and disease. Infect Immun. 2000;68:6511–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Casadevall A, Pirofski LA. What is a host? Attributes of individual susceptibility. Infect Immun. 2018;86:1–12.

    Article  Google Scholar 

  8. Clemente JC, Ursell LK, Parfrey LW, Knight R. The impact of the gut microbiota on human health: an integrative view. Cell. 2012;148:1258–70.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Adair KL, Douglas AE. Making a microbiome: the many determinants of host-associated microbial community composition. Curr Opin Microbiol. 2017;35:23–9.

    Article  PubMed  Google Scholar 

  10. Costello EK, Stagaman K, Dethlefsen L, Bohannan BJM, Relman DA. The application of ecological theory toward an understanding of the human microbiome. Science (80- ). 2012;336:1255–62.

    Article  CAS  Google Scholar 

  11. Flórez LV, Biedermann PHW, Engl T, Kaltenpoth M. Defensive symbioses of animals with prokaryotic and eukaryotic microorganisms. Nat Prod Rep. 2015;32:904–36.

    Article  PubMed  Google Scholar 

  12. IUCN. IUCN Red List of Threatened Species. Version 2021-1. https://www.iucnredlist.org. [Internet]. 2021 [cited 2021 May 11]. www.iucnredlist.org

  13. Scheele BC, Pasmans F, Skerratt LF, Berger L, Martel A, Beukema W, et al. Amphibian fungal panzootic causes catastrophic and ongoing loss of biodiversity. Science (80- ). 2019;1463:1459–63.

    Article  CAS  Google Scholar 

  14. Donachie SP, Fraser CJ, Hill EC, Butler MA. The problem with ‘microbiome.’ Diversity. 2021;13:138.

    Article  CAS  Google Scholar 

  15. Rollins-Smith LA, Ramsey JP, Pask JD, Reinert LK, Woodhams DC. Amphibian immune defenses against chytridiomycosis: Impacts of changing environments. Integr Comp Biol. 2011;51:552–62.

    Article  CAS  PubMed  Google Scholar 

  16. Jani AJ, Briggs CJ. The pathogen Batrachochytrium dendrobatidis disturbs the frog skin microbiome during a natural epidemic and experimental infection. Proc Natl Acad Sci. 2014;111:E5049–58.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Walke JB, Becker MH, Loftus SC, House LL, Cormier G, Jensen RV, et al. Amphibian skin may select for rare environmental microbes. ISME J. 2014;8:2207–17.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Varga JFA, Bui-Marinos MP, Katzenback BA. Frog skin innate immune defences: sensing and surviving pathogens. Front Immunol. 2019;10:1–21.

    Article  CAS  Google Scholar 

  19. Bataille A, Lee-Cruz L, Tripathi B, Kim H, Waldman B. Microbiome variation across amphibian skin regions: implications for chytridiomycosis mitigation efforts. Microb Ecol. 2016;71:221–32.

    Article  PubMed  Google Scholar 

  20. Bernardo-Cravo AP, Schmeller DS, Chatzinotas A, Vredenburg VT, Loyau A. Environmental factors and host microbiomes shape host–pathogen dynamics. Trends Parasitol. 2020;36:616–33.

    Article  CAS  PubMed  Google Scholar 

  21. Rebollar EA, Martínez-Ugalde E, Orta AH. The amphibian skin microbiome and its protective role against chytridiomycosis. Herpetologica. 2020;76:167–77.

    Article  Google Scholar 

  22. Harris RN, Brucker RM, Walke JB, Becker MH, Schwantes CR, Flaherty DC, et al. Skin microbes on frogs prevent morbidity and mortality caused by a lethal skin fungus. ISME J. 2009;3:818–24.

    Article  CAS  PubMed  Google Scholar 

  23. Harris RN, Lauer A, Simon MA, Banning JL, Alford RA. Addition of antifungal skin bacteria to salamanders ameliorates the effects of chytridiomycosis. Dis Aquat Org. 2009;83:11–6.

    Article  Google Scholar 

  24. Kueneman JG, Woodhams DC, Harris R, Archer HM, Knight R, McKenzie VJ, et al. Probiotic treatment restores protection against lethal fungal infection lost during amphibian captivity. Proc R Soc B Biol Sci. 2016;283:20161553.

    Article  CAS  Google Scholar 

  25. Longo AV, Savage AE, Hewson I, Zamudio KR. Seasonal and ontogenetic variation of skin microbial communities and relationships to natural disease dynamics in declining amphibians. R Soc Open Sci. 2015;2:140377.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Cohen JM, Venesky MD, Sauer EL, Civitello DJ, McMahon TA, Roznik EA, et al. The thermal mismatch hypothesis explains host susceptibility to an emerging infectious disease. Ecol Lett. 2017;20:184–93.

    Article  PubMed  Google Scholar 

  27. Greenspan SE, Bower DS, Webb RJ, Roznik EA, Stevenson LA, Berger L, et al. Realistic heat pulses protect frogs from disease under simulated rainforest frog thermal regimes. Funct Ecol. 2017;31:2274–86.

    Article  Google Scholar 

  28. Neely WJ, Greenspan SE, Ribeiro LP, Carvalho T, Martins RA, Rodriguez D, et al. Synergistic effects of warming and disease linked to high mortality in cool-adapted terrestrial frogs. Biol Conserv. 2020;245:108521.

    Article  Google Scholar 

  29. Preuss JF, Greenspan SE, Rossi EM, Lucas Gonsales EM, Neely WJ, Valiati VH, et al. Widespread pig farming practice linked to shifts in skin microbiomes and disease in pond-breeding amphibians. Environ Sci Technol. 2020;54:11301–12.

    Article  CAS  PubMed  Google Scholar 

  30. McMahon TA, Romansic JM, Rohr JR. Nonmonotonic and monotonic effects of pesticides on the pathogenic fungus Batrachochytrium dendrobatidis in culture and on tadpoles. Environ Sci Technol. 2013;47:7958–64.

    Article  CAS  PubMed  Google Scholar 

  31. Gervasi S, Gondhalekar C, Olson DH, Blaustein AR. Host identity matters in the amphibian-Batrachochytrium dendrobatidis system: fine-scale patterns of variation in responses to a multi-host pathogen. PLoS ONE. 2013;8:e54490.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Bresciano JC, Salvador CA, Paz-y-Miño C, Parody-Merino AM, Bosch J, Woodhams DC. Variation in the presence of anti-Batrachochytrium dendrobatidis bacteria of amphibians across life stages and elevations in Ecuador. EcoHealth. 2015;12:310–9.

    Article  CAS  PubMed  Google Scholar 

  33. Rowley JJL, Alford RA. Behaviour of Australian rainforest stream frogs may affect the transmission of chytridiomycosis. Dis Aquat Organ. 2007;77:1–9.

    Article  PubMed  Google Scholar 

  34. Brem FMR, Lips KR. Batrachochytrium dendrobatidis infection patterns among Panamanian amphibian species, habitats and elevations during epizootic and enzootic stages. Dis Aquat Organ. 2008;81:189–202.

    Article  PubMed  Google Scholar 

  35. Richards-Zawacki CL. Thermoregulatory behaviour affects prevalence of chytrid fungal infection in a wild population of Panamanian golden frogs. Proc Biol Sci. 2010;277:519–28.

    PubMed  Google Scholar 

  36. McMahon TA, Sears BF, Venesky MD, Bessler SM, Brown JM, Deutsch K, et al. Amphibians acquire resistance to live and dead fungus overcoming fungal immunosuppression. Nature. 2014;511:224–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Bataille A, Cashins SD, Grogan L, Skerratt LF, Hunter D, Mcfadden M, et al. Susceptibility of amphibians to chytridiomycosis is associated with MHC class II conformation. Proc B. 2015;282:20143127.

    Google Scholar 

  38. Savage AE, Gratwicke B, Hope K, Bronikowski E, Fleischer RC. Sustained immune activation is associated with susceptibility to the amphibian chytrid fungus. Mol Ecol. 2020;29:2889–903.

    Article  CAS  PubMed  Google Scholar 

  39. Demori I, El Rashed Z, Corradino V, Catalano A, Rovegno L, Queirolo L, et al. Peptides for skin protection and healing in amphibians. Molecules. 2019;24:347.

    Article  CAS  PubMed Central  Google Scholar 

  40. Becker CG, Bletz MC, Greenspan SE, Rodriguez D, Lambertini C, Jenkinson TS, et al. Low-load pathogen spillover predicts shifts in skin microbiome and survival of a terrestrial-breeding amphibian. Proc R Soc B Biol Sci. 2019;286:20191114.

    Article  Google Scholar 

  41. Becker CG, Rodriguez D, Longo AV, Toledo LF, Lambertini C, Leite DS, et al. Deforestation, host community structure, and amphibian disease risk. Basic Appl Ecol. 2016;17:72–80.

    Article  Google Scholar 

  42. Becker CG, Rodriguez D, Toledo LF, Longo AV, Leite DS, Lambertini C, et al. Partitioning the net effect of host diversity on an emerging amphibian pathogen. Proc R Soc B Biol Sci. 2014;281:20141796.

    Article  Google Scholar 

  43. Haddad CFB, Toledo LF, Prado CPA, Loebmann D, Gasparini JL, Sazima I. Guia dos anfíbios da Mata Atlântica: diversidade e biologia. São Paulo: Anolis Books; 2013.

    Google Scholar 

  44. Lambertini C, Becker CG, Belasen AM, Valencia-Aguilar A, Nunes-de-Almeida CHL, Betancourt-Román CM, et al. Biotic and abiotic determinants of Batrachochytrium dendrobatidis infections in amphibians of the Brazilian Atlantic Forest. Fungal Ecol. 2021;49:100995.

    Article  Google Scholar 

  45. Becker CG, Fonseca CR, Haddad CFB, Batista RF, Prado PI. Habitat split and the global decline of amphibians. Science(80- ). 2007;318:1775–7.

    CAS  Google Scholar 

  46. Lips KR, Reeve JD, Witters LR. Ecological traits predicting amphibian population declines in Central America. Conserv Biol. 2003;17:1078–88.

    Article  Google Scholar 

  47. Longo AV, Burrowes PA, Joglar RL. Seasonality of Batrachochytrium dendrobatidis infection in direct-developing frogs suggests a mechanism for persistence. Dis Aquat Organ. 2010;92:253–60.

    Article  PubMed  Google Scholar 

  48. Moura-Campos D, Greenspan SE, DiRenzo GV, Neely WJ, Toledo LF, Becker CG. Fungal disease cluster in tropical terrestrial frogs predicted by low rainfall. Biol Conserv. 2021;261:109246.

    Article  Google Scholar 

  49. Mesquita AFC, Lambertini C, Lyra M, Malagoli LR, James TY, Toledo LF, et al. Low resistance to chytridiomycosis in direct-developing amphibians. Sci Rep. 2017;7:16605.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  50. Piovia-Scott J, Rejmanek D, Woodhams DC, Worth SJ, Kenny H, McKenzie V, et al. Greater species richness of bacterial skin symbionts better suppresses the amphibian fungal pathogen Batrachochytrium dendrobatidis. Microb Ecol. 2017;74:217–26.

    Article  PubMed  Google Scholar 

  51. Greenspan SE, Migliorini GH, Lyra ML, Pontes MR, Carvalho T, Ribeiro LP, et al. Warming drives ecological community changes linked to host-associated microbiome dysbiosis. Nat Clim Change. 2020;10:1057–61.

    Article  Google Scholar 

  52. Poudel R, Jumpponen A, Schlatter DC, Paulitz TC, Gardener BBM, Kinkel LL, et al. Microbiome networks: a systems framework for identifying candidate microbial assemblages for disease management. Anal Theor Plant Pathol. 2016;106:1083–96.

    CAS  Google Scholar 

  53. Faust K, Raes J. Microbial interactions: from networks to models. Nat Rev Microbiol. 2012;10:538–50.

    Article  CAS  PubMed  Google Scholar 

  54. Ribeiro JW, Siqueira T, DiRenzo GV, Lambertini C, Lyra ML, Toledo LF, et al. Assessing amphibian disease risk across tropical streams while accounting for imperfect pathogen detection. Oecologia. 2020;193:237–48.

    Article  PubMed  Google Scholar 

  55. Ruthsatz K, Lyra ML, Lambertini C, Belasen AM, Jenkinson TS, da Silva LD, et al. Skin microbiome correlates with bioclimate and Batrachochytrium dendrobatidis infection intensity in Brazil’s Atlantic Forest treefrogs. Sci Rep. 2020;10:22311.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  56. McKenzie VJ, Bowers RM, Fierer N, Knight R, Lauber CL. Co-habiting amphibian species harbor unique skin bacterial communities in wild populations. ISME J. 2012;6:588–96.

    Article  CAS  PubMed  Google Scholar 

  57. Kueneman JG, Parfrey LW, Woodhams DC, Archer HM, Knight R, McKenzie VJ. The amphibian skin-associated microbiome across species, space and life history stages. Mol Ecol. 2014;23:1238–50.

    Article  PubMed  Google Scholar 

  58. Rebollar EA, Hughey MC, Medina D, Harris RN, Ibáñez R, Belden LK. Skin bacterial diversity of Panamanian frogs is associated with host susceptibility and presence of Batrachochytrium dendrobatidis. ISME J. 2016;10:1682–95.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  59. Abarca JG, Vargas G, Zuniga I, Whitfield SM, Woodhams DC, Kerby J, et al. Assessment of bacterial communities associated with the skin of Costa Rican amphibians at La Selva Biological Station. Front Microbiol. 2018;9:1–12.

    Article  Google Scholar 

  60. Walke JB, Becker MH, Hughey MC, Swartwout MC, Jensen RV, Belden LK. Dominance-function relationships in the amphibian skin microbiome. Environ Microbiol. 2017;19:3387–97.

    Article  CAS  PubMed  Google Scholar 

  61. Risely A. Applying the core microbiome to understand host–microbe systems. J Anim Ecol. 2020;89:1549–58.

    Article  PubMed  Google Scholar 

  62. Martín González AM, Dalsgaard B, Olesen JM. Centrality measures and the importance of generalist species in pollination networks. Ecol Complex. 2010;7:36–43.

    Article  Google Scholar 

  63. Trosvik P, de Muinck EJ. Ecology of bacteria in the human gastrointestinal tract—identification of keystone and foundation taxa. Microbiome. 2015;3:44.

    Article  PubMed  PubMed Central  Google Scholar 

  64. Baldassano SN, Bassett DS. Topological distortion and reorganized modular structure of gut microbial co-occurrence networks in inflammatory bowel disease. Sci Rep. 2016;6:1–14.

    Article  CAS  Google Scholar 

  65. Guimerà R, Nunes Amaral LA. Functional cartography of complex metabolic networks. Nature. 2005;433:895–900.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  66. Papin JA, Reed JL, Palsson BO. Hierarchical thinking in network biology: the unbiased modularization of biochemical networks. Trends Biochem Sci. 2004;29:641–7.

    Article  CAS  PubMed  Google Scholar 

  67. Ravasz E, Somera AL, Mongru DA, Oltvai ZN, Barabási AL. Hierarchical organization of modularity in metabolic networks. Science (80- ). 2002;297:1551–5.

    Article  CAS  Google Scholar 

  68. Hartwell LH, Hopfield JJ, Leibler S, Murray AW. From molecular to modular cell biology. Nature. 1999;402:C47-52.

    Article  CAS  PubMed  Google Scholar 

  69. Rezende EL, Albert EM, Fortuna MA, Bascompte J. Compartments in a marine food web associated with phylogeny, body mass, and habitat structure. Ecol Lett. 2009;12:779–88.

    Article  PubMed  Google Scholar 

  70. Lurgi M, Thomas T, Wemheuer B, Webster NS, Montoya JM. Modularity and predicted functions of the global sponge-microbiome network. Nat Commun. 2019;10:992.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  71. Shade A, Peter H, Allison SD, Baho DL, Berga M, Bürgmann H, et al. Fundamentals of microbial community resistance and resilience. Front Microbiol. 2012;3:1–19.

    Article  Google Scholar 

  72. Muletz CR, Myers JM, Domangue RJ, Herrick JB, Harris RN. Soil bioaugmentation with amphibian cutaneous bacteria protects amphibian hosts from infection by Batrachochytrium dendrobatidis. Biol Conserv. 2012;152:119–26.

    Article  Google Scholar 

  73. Brucker RM, Baylor CM, Walters RL, Lauer A, Harris RN, Minbiole KPC. The identification of 2,4-diacetylphloroglucinol as an antifungal metabolite produced by cutaneous bacteria of the salamander Plethodon cinereus. J Chem Ecol. 2008;34:39–43.

    Article  CAS  PubMed  Google Scholar 

  74. Lauer A, Simon MA, Banning JL, Andre E, Duncan K, Harris RN. Common cutaneous bacteria from the eastern red-backed salamander can inhibit pathogenic fungi. Copeia. 2007;2007:630–40.

    Article  Google Scholar 

  75. Brucker RM, Harris RN, Schwantes CR, Gallaher TN, Flaherty DC, Lam BA, et al. Amphibian chemical defense: antifungal metabolites of the microsymbiont Janthinobacterium lividum on the salamander Plethodon cinereus. J Chem Ecol. 2008;34:1422–9.

    Article  CAS  PubMed  Google Scholar 

  76. Loudon AH, Holland JA, Umile TP, Burzynski EA, Minbiole KPC, Harris RN. Interactions between amphibians’ symbiotic bacteria cause the production of emergent anti-fungal metabolites. Front Microbiol. 2014;5:441.

    Article  PubMed  PubMed Central  Google Scholar 

  77. Becker MH, Brucker RM, Schwantes CR, Harris RN, Minbiole KPC. The bacterially produced metabolite violacein is associated with survival of amphibians infected with a lethal fungus. Appl Environ Microbiol. 2009;75:6635–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  78. Rebollar EA, Bridges T, Hughey MC, Medina D, Belden LK, Harris RN. Integrating the role of antifungal bacteria into skin symbiotic communities of three Neotropical frog species. ISME J. 2019;13:1763–75.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  79. Ohmer MEB, Cramp RL, Russo CJM, White CR, Franklin CE. Skin sloughing in susceptible and resistant amphibians regulates infection with a fungal pathogen. Sci Rep. 2017;7:3529.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  80. Lucas EM, Molinari De Bastiani VI, Lingnau R. Geographic distribution, habitat use and vocalizations of the leaf-litter frog Ischnocnema henselii (Anura: Brachycephalidae) in the subtropical Atlantic Forest. Rev Bras Zoociências. 2018;19:151–62.

    Article  Google Scholar 

  81. Levy DL, Heald R. Biological scaling problems and solutions in amphibians. Cold Spring Harb Perspect Biol. 2016;8:1–16.

    Article  Google Scholar 

  82. Kueneman JG, Woodhams DC, Van Treuren W, Archer HM, Knight R, McKenzie VJ. Inhibitory bacteria reduce fungi on early life stages of endangered Colorado boreal toads (Anaxyrus boreas). ISME J. 2016;10:934–44.

    Article  PubMed  Google Scholar 

  83. Kearns PJ, Fischer S, Fernández-Beaskoetxea S, Gabor CR, Bosch J, Bowen JL, et al. Fight fungi with fungi: antifungal properties of the amphibian mycobiome. Front Microbiol. 2017;8:2494.

    Article  PubMed  PubMed Central  Google Scholar 

  84. Medina D, Hughey MC, Walke JB, Becker MH, Pontarelli K, Sun S, et al. Amphibian skin fungal communities vary across host species and do not correlate with infection by a pathogenic fungus. Environ Microbiol. 2019;21:2905–20.

    Article  CAS  PubMed  Google Scholar 

  85. Belasen AM, Riolo MA, Bletz MC, Lyra ML, Toledo LF, James TY. Geography, host genetics, and cross-domain microbial networks structure the skin microbiota of fragmented Brazilian Atlantic Forest frog populations. Ecol Evol. 2021;11:9293–307.

    Article  PubMed  PubMed Central  Google Scholar 

  86. Ramsey JP, Reinert LK, Harper LK, Woodhams DC, Rollins-Smith LA. Immune defenses against Batrachochytrium dendrobatidis, a fungus linked to global amphibian declines, in the South African clawed frog, Xenopus laevis. Infect Immun. 2010;78:3981–92.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  87. Rollins-Smith LA, Fites JS, Reinert LK, Shiakolas AR, Umile TP, Minbiole KPC. Immunomodulatory metabolites released by the frog-killing fungus Batrachochytrium dendrobatidis. Infect Immun. 2015;83:4565–70.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  88. Hernández-Gómez O, Briggler JT, Williams RN. Influence of immunogenetics, sex and body condition on the cutaneous microbial communities of two giant salamanders. Mol Ecol. 2018;27:1915–29.

    Article  PubMed  Google Scholar 

  89. Griffiths SM, Harrison XA, Weldon C, Wood MD, Pretorius A, Hopkins K, et al. Genetic variability and ontogeny predict microbiome structure in a disease-challenged montane amphibian. ISME J. 2018;12:2506–17.

    Article  PubMed  PubMed Central  Google Scholar 

  90. Savage AE, Zamudio KR. MHC genotypes associate with resistance to a frog-killing fungus. Proc Natl Acad Sci. 2011;108:16705–10.

    Article  PubMed  PubMed Central  Google Scholar 

  91. Savage AE, Mulder KP, Torres T, Wells S. Lost but not forgotten: MHC genotypes predict overwinter survival despite depauperate MHC diversity in a declining frog. Conserv Genet. 2018;19:309–22.

    Article  CAS  Google Scholar 

  92. Dias IR, Lourenço-De-Moraes R, Solé M. Description of the advertisement call and morphometry of Haddadus binotatus (Spix, 1824) from a population from southern Bahia, Brazil. North West J Zool. 2012;8:107–11.

    Google Scholar 

  93. Carvalho T, Becker CG, Toledo LF. Historical amphibian declines and extinctions in Brazil linked to chytridiomycosis. Proc R Soc B Biol Sci. 2017;284:20162254.

    Article  Google Scholar 

  94. Rebollar EA, Antwis RE, Becker MH, Belden LK, Bletz MC, Brucker RM, et al. Using “omics” and integrated multi-omics approaches to guide probiotic selection to mitigate chytridiomycosis and other emerging infectious diseases. Front Microbiol. 2016;7:68.

    Article  PubMed  PubMed Central  Google Scholar 

  95. Hyatt AD, Boyle DG, Olsen V, Boyle DB, Berger L, Obendorf D, et al. Diagnostic assays and sampling protocols for the detection of Batrachochytrium dendrobatidis. Dis Aquat Organ. 2007;73:175–92.

    Article  CAS  PubMed  Google Scholar 

  96. Boyle DG, Olsen V, Morgan JAT, Hyatt AD. Rapid quantitative detection of chytridiomycosis (Batrachochytrium dendrobatidis) in amphibian samples using real-time Taqman PCR assay. Dis Aquat Organ. 2004;60:141–8.

    Article  CAS  PubMed  Google Scholar 

  97. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010;7:335–6.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  98. Kozich JJ, Westcott SL, Baxter NT, Highlander SK, Schloss PD. Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing platform. Appl Environ Microbiol. 2013;79:5112–20.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  99. Amir A, McDonald D, Navas-Molina JA, Kopylova E, Morton JT, Xu ZZ, et al. Deblur rapidly resolves single-nucleotide community sequence patterns. mSystems. 2017;2:e0019116.

    Article  Google Scholar 

  100. Bokulich NA, Subramanian S, Faith JJ, Gevers D, Gordon JI, Knight R, et al. Quality-filtering vastly improves diversity estimates from Illumina amplicon sequencing. Nat Methods. 2013;10:57–9.

    Article  CAS  PubMed  Google Scholar 

  101. Lozupone C, Knight R. UniFrac: a new phylogenetic method for comparing microbial communities. Appl Environ Microbiol. 2005;71:8228–35.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  102. Lozupone C, Lladser ME, Knights D, Stombaugh J, Knight R. UniFrac: an effective distance metric for microbial community comparison. ISME J. 2011;5:169–72.

    Article  PubMed  Google Scholar 

  103. Zaneveld JR, McMinds R, Vega TR. Stress and stability: applying the Anna Karenina principle to animal microbiomes. Nat Microbiol. 2017;2:17121.

    Article  CAS  PubMed  Google Scholar 

  104. Jin Song S, Woodhams DC, Martino C, Allaband C, Mu A, Javorschi-Miller-Montgomery S, et al. Engineering the microbiome for animal health and conservation. Exp Biol Med. 2019;244:494–504.

    Article  CAS  Google Scholar 

  105. Kriss M, Hazleton KZ, Nusbacher NM, Martin CG, Lozupone CA. Low diversity gut microbiota dysbiosis: drivers, functional implications and recovery. Curr Opin Microbiol. 2018;44:34–40.

    Article  PubMed  PubMed Central  Google Scholar 

  106. Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS, et al. Metagenomic biomarker discovery and explanation. Genome Biol. 2011;12:R60.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank Ingrid Scheeffer, Cinthia Brasileiro, Diego Moura, João Mendes, Emily Guimarães, Roberta Mariáh, and Fundação Parque Zoológico de São Paulo.

Funding

The National Science Foundation provided funding to C.G.B. (IOS-1947681).

Author information

Author notes

  1. Renato A. Martins and Sasha E. Greenspan have contributed equally to this work

Authors and Affiliations

  1. Programa de Pós-Graduação em Conservação da Fauna, Universidade Federal de São Carlos, São Carlos, SP, 13565-905, Brazil

    Renato A. Martins

  2. Department of Biological Sciences, The University of Alabama, Tuscaloosa, AL, 35487, USA

    Sasha E. Greenspan, Shannon Buttimer, Vanessa M. Marshall, Wesley J. Neely, Samantha Siomko & C. Guilherme Becker

  3. Sistema Nacional de Investigación, SENACYT, Building 205, City of Knowledge, Clayton, Panama, Republic of Panama

    Daniel Medina

  4. Department of Biology, The Pennsylvania State University, University Park, PA, 16803, USA

    Daniel Medina & C. Guilherme Becker

  5. Department of Biodiversity and Aquaculture Center (CAUNESP), Universidade Estadual Paulista, Rio Claro, SP, 13506-900, Brazil

    Mariana L. Lyra & Célio F. B. Haddad

  6. Centro de Ciências da Natureza, Universidade Federal de São Carlos, Campus Lagoa do Sino, Buri, SP, 18290-000, Brazil

    Vinícius São-Pedro

Authors
  1. Renato A. Martins
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Sasha E. Greenspan
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Daniel Medina
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Shannon Buttimer
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Vanessa M. Marshall
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Wesley J. Neely
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Samantha Siomko
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Mariana L. Lyra
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Célio F. B. Haddad
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Vinícius São-Pedro
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. C. Guilherme Becker
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

CGB designed the study. RAM and CGB carried out fieldwork. RAM, SEG, WJN, and MLL carried out laboratory work. CFBH and VSP facilitated field and laboratory work. SEG, DM, WJN, SB, and CGB analyzed the data. All authors interpreted the data and contributed to drafting the work. SEG and DM revised the manuscript. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Sasha E. Greenspan or C. Guilherme Becker.

Ethics declarations

Ethics approval and consent to participate

All procedures were approved by the Universidade Federal de São Carlos Animal Ethics Committee (protocol 2464230518). Our research was conducted under the following permits: SISBIO: 63917-3; COTEC: 260108-005.880/2018; SISGEN: A663126.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1:

 Supplementary methods, tables and figures.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Martins, R.A., Greenspan, S.E., Medina, D. et al. Signatures of functional bacteriome structure in a tropical direct-developing amphibian species. anim microbiome 4, 40 (2022). https://doi.org/10.1186/s42523-022-00188-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s42523-022-00188-7

Keywords

  • Bacterial co-occurrence network analysis
  • Batrachochytrium dendrobatidis
  • Brazil’s Atlantic Forest
  • Ecological core
  • Haddadus binotatus
  • Ischnocnema henselii
  • Microbiome

Animal Microbiome

ISSN: 2524-4671