A controlled laboratory study has demonstrated that symbiotic skin bacteria provide an essential defensive barrier against the lethal fungal pathogen Batrachochytrium dendrobatidis in the neotropical leaf-litter frog Haddadus binotatus, while the closely related frog Ischnocnema henselii suffers high mortality regardless of microbiome disruption. The research, published in Animal Microbiome by an international team led by Laura K. Schuck, Julia R. Ernetti, and C. Guilherme Becker, reveals fundamentally divergent defense strategies between two ecologically similar amphibian species from Brazil’s Atlantic Forest. Using a two-by-two factorial experimental design crossing fungal exposure with antibiotic-mediated bacteriome suppression, the authors monitored frog survival, pathogen burden, and bacterial community structure over forty days. Their results provide direct in vivo evidence that microbially mediated disease defense is driven by community-level functional attributes rather than simple taxonomic diversity, elucidating why sympatric amphibian species display starkly different disease susceptibilities under identical pathogen challenges.
The investigation addresses an ecological contrast in Brazil’s Atlantic Forest, a global biodiversity hotspot where amphibian declines during the late 1970s were primarily driven by chytridiomycosis. While many contemporary amphibian populations persist with enzootic fungal infections, Haddadus binotatus represents an extraordinary exception. In an earlier survey spanning forty Atlantic Forest sites across 450 kilometers, none of the 126 sampled H. binotatus individuals tested positive for Batrachochytrium dendrobatidis. Conversely, Ischnocnema henselii—a co-occurring species in the superfamily Brachycephaloidea sharing direct-developing reproduction and terrestrial leaf-litter microhabitats—displayed an infection prevalence of 34.8 percent, the highest among all surveyed species. Previous field data also revealed that H. binotatus harbored distinctive bacterial network structures and elevated skin bacterial diversity. These contrasting field patterns led researchers to hypothesize that the natural resistance of H. binotatus is actively mediated by functional cutaneous bacteria rather than behavioral pathogen avoidance or microhabitat differentiation.
To experimentally evaluate this protective hypothesis, the researchers conducted a field sampling campaign during the 2023 rainy breeding season, collecting forty Haddadus binotatus and forty Ischnocnema henselii individuals across three Atlantic Forest fragments in São Luís do Paraitinga, São Paulo, Brazil. Handled with fresh nitrile gloves, each frog was rinsed with autoclaved distilled water to eliminate transient microbes before receiving standardized diagnostic swabs across the ventral abdomen and inner thighs. Two sequential swabs were collected per frog: one cryopreserved in twenty percent glycerol solution at minus eighty degrees Celsius for bacterial isolation and culturing, and one stored dry for molecular assays. Before laboratory trials commenced, animals were stratified across experimental cohorts. Haddadus binotatus individuals were balanced across treatments by sex and body size, with mean snout-vent lengths of 39.4 to 41.4 millimeters and body masses of 5.1 to 6.0 grams, while I. henselii individuals were stratified strictly by body size.
The forty-day experimental protocol employed a balanced two-by-two factorial design consisting of four treatment groups per species, each containing ten frogs: antibiotic-treated with fungal exposure, antibiotic-treated unexposed control, untreated with fungal exposure, and untreated unexposed control. To suppress cutaneous bacteria, twenty frogs of each species were immersed overnight for fourteen hours in shallow antibiotic baths covering their legs, following an established protocol from Holden and colleagues. Control frogs spent fourteen hours submerged in sterile artificial pond water. Following treatments, animals were rested undisturbed for thirty-six hours before undergoing pre-exposure swabbing. Pathogen challenge was administered using Batrachochytrium dendrobatidis isolate CLFT 278, representing the local Bd-GPL/Bd-Brazil lineage. Exposed individuals were maintained for forty-five minutes in ten milliliters of solution containing 4.2 million zoospores per milliliter—delivering ten million zoospores total—while unexposed frogs received sterile control baths before all subjects returned to individual terraria for daily monitoring.
Survival curves monitored across forty days revealed stark differences in treatment responses between the two species. Global Cox proportional hazards models showed that Ischnocnema henselii experienced a significantly higher overall mortality rate than Haddadus binotatus, with a hazard ratio of 2.848 (p = 0.012). A significant interaction between antibiotic treatment and pathogen exposure also emerged across the study cohort (hazard ratio = 11.290, p = 0.048), indicating that bacteriome disruption drastically increased mortality under infectious challenge. In species-specific planned log-rank tests, bacteriome suppression significantly increased mortality among fungus-exposed H. binotatus (chi-square = 5.500, df = 1, p = 0.019), where survival remained high under fungal exposure alone but declined sharply following antibiotic treatment. In contrast, I. henselii experienced high mortality under fungal exposure regardless of antibiotic administration (chi-square = 0.500, df = 1, p = 0.474). Crucially, antibiotic treatment alone did not reduce survival in either species, demonstrating that mortality was contingent on pathogen exposure.
Quantitative tracking of fungal infection intensities using TaqMan quantitative PCR assays targeting internal transcribed spacer gene copy numbers corroborated the divergent survival patterns. At field capture, only two of forty H. binotatus individuals tested positive for the fungus, confirming low baseline wild infection. In laboratory mixed-effects models, antibiotic application per se did not dictate infection intensity in H. binotatus (beta = -0.082, t = -0.18, p = 0.861). Instead, fungal intensity in H. binotatus was significantly governed by bacteriome function, displaying a strong negative association with the proportion of putative pathogen-inhibitory sequence reads (beta = -0.97, t = -2.32, p = 0.022). Conversely, I. henselii exhibited no significant moderation of infection intensity by antibiotic treatment (beta = 0.20, t = 0.32, p = 0.751), and the association between inhibitory sequence reads and fungal load was non-significant (beta = -0.62, t = -1.76, p = 0.082), indicating weaker or less consistent bacterial defense in this host species.
To quantify antifungal function directly, the research team isolated and cultured 786 bacterial isolates from the glycerol-preserved skin swabs taken at initial field capture. They recovered 374 isolates from H. binotatus, representing 139 species, 62 genera, and 27 families, alongside 412 isolates from I. henselii, representing 244 species, 87 genera, and 35 families. Cell-free metabolite extracts were tested against Batrachochytrium dendrobatidis reference strain CLFT 041 in 96-well microplate challenge assays, monitoring optical density at 492 nanometers across eight days. Isolates inhibiting fungal growth by fifty percent or more relative to controls were classified as inhibitory. At field capture, 43.2 percent of isolates from H. binotatus inhibited fungal growth compared to 36.3 percent from I. henselii (t = 0.421, p = 0.673). Logistic regression revealed that genus identity significantly explained inhibition probability (likelihood ratio chi-square = 44.509, df = 13, p < 0.001), with Stenotrophomonas exhibiting an odds ratio of approximately 13 (beta = 2.56, p = 0.004).
Molecular profiling of baseline skin bacteriomes using PacBio HiFi long-read sequencing of full-length 16S ribosomal RNA genes revealed subtle but critical distinctions between host species prior to experimental manipulation. Total bacterial amplicon sequence variant richness did not differ significantly between H. binotatus and I. henselii at capture (S = 1026, Z = -0.772, p = 0.440), nor did unweighted UniFrac community composition diverge (PERMANOVA F = 1.093, R-squared = 0.016, p = 0.242). The baseline proportion of inhibitory sequence reads was also statistically indistinguishable between species (S = 992, Z = -1.207, p = 0.227). However, when sequences were matched against the cultured isolate database, H. binotatus harbored a significantly higher richness of pathogen-inhibitory amplicon sequence variants (S = 952.5, Z = -1.708, p = 0.043). Furthermore, ANCOM-BC2 analysis detected field enrichment of Cutibacterium, Flavobacterium, and Sphingobacterium in I. henselii, whereas H. binotatus showed selective enrichment of Xiphinematobacter, Phenylobacterium, Actinoallomurus, and Moraxella.
The investigation also evaluated community stability by measuring multivariate dispersion based on unweighted UniFrac distances. In unperturbed animals of both species, dispersion declined with increasing fungal load, reflecting community homogenization under aseptic captive conditions (beta = -0.018, p < 0.001 for H. binotatus; beta = -0.015, p = 0.006 for I. henselii). However, in H. binotatus exposed to both antibiotics and the fungus, this relationship reversed through a significant positive interaction (beta = 0.026, t = 3.39, p = 0.001), showing that microbial suppression triggered community destabilization under infection, consistent with the Anna Karenina principle. In I. henselii, antibiotic administration reduced dispersion overall (beta = -0.045, t = -2.07, p = 0.046) without interacting with fungal load (beta = 0.007, p = 0.299). Concurrently, experimental treatments reshaped community composition: antibiotics enriched Chitinophaga while depleting Herbiconiux, Pseudescherichia, Aeromicrobium, and Luteolibacter, whereas combined antibiotic and fungal exposure specifically enriched Achromobacter and Brucella while depleting Dyella, Herbiconiux, and Luteolibacter.
While these empirical findings substantiate bacteriome-mediated defense in Haddadus binotatus, the authors acknowledged important physiological and methodological considerations. Broad-spectrum antibiotic administration may generate non-target host effects, such as elevated oxidative stress or impacts on host-secreted antimicrobial peptides, which were not directly quantified in this investigation. In time-varying Cox models, higher fungal loads independently predicted reduced survival (hazard ratio = 2.212, p < 0.001), whereas unsuppressed bacteriomes conferred higher survival (hazard ratio = 0.372, p = 0.006), supporting the protective influence of intact commensal communities. Although separate fungal strains were used for in vivo challenges (CLFT 278) and in vitro assays (CLFT 041), both belong to the global pandemic lineage Bd-GPL, maintaining experimental relevance. Ultimately, the contrast between Haddadus binotatus and Ischnocnema henselii demonstrates that amphibian disease resistance depends on specific functional richness and inhibitory metabolic capacities rather than broad microbial diversity alone, highlighting microbiome function as a vital component of wildlife disease defense.
Subject of Research: Bacteriome-mediated defense against Batrachochytrium dendrobatidis in neotropical frogs
Article Title: Differential skin-bacteriome-mediated defense against chytridiomycosis in two neotropical frog species
Article References: Schuck, L. K., Ernetti, J. R., Buttimer, S., de Assis, A. B., Martins, R. A., Pontes, M. R., Zanatta, A. C., Tosta, M. B., Toledo, L. F., & Becker, C. G. (2026). Differential skin-bacteriome-mediated defense against chytridiomycosis in two neotropical frog species. Animal Microbiome, 8(1), Article 89. https://doi.org/10.1186/s42523-026-00593-2
Image Credits: AI Generated
DOI: 10.1186/s42523-026-00593-2
Keywords: Amphibians, Batrachochytrium dendrobatidis, Skin Microbiome, Chytridiomycosis, Haddadus binotatus, Ischnocnema henselii, Disease Ecology, Atlantic Forest, Differential, skin-bacteriome-mediated, defense, against
