Cross-Domain co-occurrence in the amphibian skin microbiome
network
Our microbiome network analyses revealed a number of notable patterns.
The bacterial network consisted of a major and minor group, and the
majority of microeukaryote groups did not form significant connections
in the eukaryote-only microbial network. However, in the overall
microbial network, a number of microbial groups exhibited cross-Domain
co-occurrence: a majority of previously unconnected microeukaryote
groups (10/16) and both previously unconnected bacterial groups became
connected in the overall microbiome network. To our knowledge, ours is
the first study to demonstrate these cross-Domain network connections in
the amphibian skin microbiome.
One important implication of this result is that ecological interactions
may exist between microbiome bacteria and eukaryotes that may
significantly impact microbiome assembly. It is currently unclear how
widespread these associations are, as previous studies that have
examined both bacteria and microeukaryotes on amphibian skin have
focused on taxon-specific associations, namely betweenBd -inhibitory bacteria and fungi (14), and between Bd and
either bacteria or microeukaryotes (16). While potential antagonistic
interactions with Bd have been the focus in cross-Domain research
on the amphibian skin microbiome, microbial interactions can occur
across the spectrum of biological symbioses (reviewed in 64).
Mutualistic interactions between bacteria and microeukaryotes have been
documented in other systems, for example mycorrhizae-helper bacteria are
known to indirectly facilitate plant-fungal interactions in the
multitrophic mycorrhizal complex (65). An alternative explanation for
our network analysis results is that bacteria and eukaryotes positively
co-occur due to co-filtering via specific host, environmental, or other
exogenous factors unrelated to microbial interactions. Further research
is needed on cross-Domain microbial co-occurrence patterns, microbial
interactions and their implications for amphibian host health.
Bd inhibitory and enhancing bacteria have variable effects
on microbiome fungi and protists
Our dataset included a number of bacteria previously shown to inhibitBd , which have been generally termed “antifungal” in the
literature (38) although empirical support for this broad designation
comes from only a single study (14). Bacteria with previously
demonstrated effects on Bd growth did not show general patterns
with T. taophora skin microbiome eukaryotes. Bacteria previously
found to enhance Bd growth were positively associated with the
Chytridiomycota more broadly, although Bd was not present in our
18S dataset. However, these bacteria were negatively associated with
Ascomycota and Basidiomycota fungi. Perhaps more critical are the
relationships with Bd inhibitory bacteria, as these bacteria have
been proposed for use in probiotic treatments for the management ofBd infections (11,66). Bd inhibitory bacteria showed weak
positive associations with Cryptomycota fungi and significant negative
associations with Basidiomycota fungi and other unidentified fungi in
the T. taophora skin microbiome. Bd -inhibitory bacteria
were also positively associated with Choanoflagellates, and showed
strong though non-significant positive associations with the
Zoopagomycota and Ichthyosporea.
Our results suggest that probiotic treatments in wild populations may
have unintended consequences for microbiome stability. According to our
analyses, specific attempts to increase Bd inhibiting bacteria
and/or reduce Bd enhancing bacteria in wild frog populations
could have unwanted effects, such as potentially reducing fungi in the
Dikarya (Ascomycota and Basidiomycota) that are known to benefit
amphibian health (Kearns et al. 2017), and/or augmenting poorly studied
parasites such as Ichthyosporea protists (67) and fungi including
Ascomycota and Zoopagomycota (68,69). These hypothetical effects warrant
further study, for example through culture-based or in vivochallenges between proposed probiotic bacteria and these potentially
impacted microeukaryotes.