Introduction
The host-associated microbiome has recently captured the attention of
researchers seeking to understand and predict disease-associated
wildlife population declines. In particular, research on the skin
microbiome is burgeoning in the field of amphibian disease, in which a
majority of studies focus on the skin disease chytridiomycosis caused by
the pathogenic chytrid fungus Batrachochytrium dendrobatidis(Bd ). Bd and other pathogens have been linked to severe
amphibian declines around the world since at least the 1970s (1–4). In
some regions, however, declines have not been observed. Plethodontid
salamander populations from the Eastern United States showed no evidence
of disease-associated declines despite the presence of Bd in the
environment (5). In a series of foundational studies, many of which were
performed in vitro , bacteria cultured from salamander skin were
correlated with reduced disease risk (6–8). Further studies pointed to
antifungal bacterial metabolite production as the main mechanism behind
this correlation between bacteria and reduced disease risk (7,9,10).
These findings among others gave rise to interest in characterizing
amphibian microbiome bacteria as a means to determine Bdsusceptibility, and in using “probiotic strategies” (manipulating
amphibian skin bacteria) to mitigate disease-associated amphibian
declines (10–13).
However, despite a growing body of research on specificBd -inhibitory bacteria, much remains to be understood about the
diversity and assembly of the overall amphibian skin microbiome,
including the ecological roles of non-bacterial taxa (but see Kueneman
et al. 2016; Kearns et al. 2017) and interactions between microbiome
bacteria and microeukaryotes other than Bd . A diversity of
microeukaryotes including fungi, microscopic metazoans, and protists
have been identified on amphibian skin using high-throughput sequencing
(16,17). In previous studies, fungi comprised the dominant eukaryotic
taxon on adult amphibians (14), and explained more variation inBd susceptibility than bacteria (15). Although little is
known about the ecological roles of these fungi in the amphibian skin
microbiome, symbiotic fungi are known to serve important roles in
protection against fungal pathogens in other host-microbial systems (Gao
et al. 2010; Newsham et al. 1995). Fungi also serve as hyperparasites,i.e. , parasites of pathogens/parasites. For example, the
cryptomycete fungus Rozella parasitizes chytrid fungi (20). Less
is known about the symbiotic roles of host-associated protists, although
microbiome eukaryotes on the whole have been shown to impact health
(21,22) and immune function (23) in mammals. Thus, microbiome eukaryotes
could be equally important as bacteria in determining disease
susceptibility in vertebrates including amphibians. Without an
understanding of the interactions between microbiome eukaryotes and
bacteria, it is impossible to predict the potential microbiome-wide
effects of proposed measures to manipulate bacteria to control Bdoutbreaks.
In addition, few studies to date have examined the genetic mechanisms
that determine animal microbiome assembly and diversity. From research
on mammals, it is known that microbiome assembly and diversity can
co-vary with overall host genetic diversity (24) as well as host
immunogenetics (25,26), with the latter relationship hypothetically
resulting from interactions between immune cells and microbes including
commensals and pathogens. In amphibians, previous studies have
demonstrated that geography, host identity, and developmental stage can
influence microbiome diversity (27–29). Yet only a single study to date
has linked amphibian skin microbiome diversity with overall host genetic
variability (29). Although no studies have directly examined the
relationship between immunogenetics and microbiome diversity or
structure in amphibians, results from an experimental study on the
laboratory model frog Xenopus laevis suggest that MHC (major
histocompatibility complex) immunogenes may determine the ability of
hosts to tolerate different microbes (30). The relationship between
immunogenes and the amphibian host-associated microbiome remains to be
explored, and is increasingly relevant for wild amphibian populations
threatened by emerging disease.
In a number of amphibian species, genetic diversity has been compromised
due to anthropogenic habitat fragmentation (31). Although it is unknown
to what extent habitat fragmentation impacts the amphibian skin
microbiome, genetic erosion in fragmented amphibian populations has been
observed at neutral loci as well as immunogenetic regions (17) which may
have implications for microbiome structure (25). In addition,
fragmentation may cause a decline in microbial transmission, which in
turn may alter microbial interactions and networks in host-associated
microbiomes. However, the effects of habitat fragmentation on wildlife
are subject to time lags (32); genetic erosion resulting from inbreeding
may not be detectable for several generations following habitat
fragmentation, making the impacts on genetics and related factors
difficult to detect in recently fragmented populations. Historically
fragmented populations offer an opportunity to examine the effects of
genetic erosion on the microbiome and broader animal health.
We evaluated the effects of long-term habitat fragmentation on the
amphibian skin microbiome using a historically fragmented model system
in the Brazilian Atlantic Forest. This system consists of dozens of
land-bridge islands, which were naturally separated from the mainland
12,000-20,000 years ago via sea level rise (33) and thus represent
ancient forest fragments. Contemporary insular frog populations were
once part of contiguous coastal populations, and are now isolated to the
islands (34,35). Using this geographic setting, we examined the impacts
of geography and host genetics on skin microbiome diversity and
structure. We used amplicon-based high-throughput DNA sequencing to
analyze bacterial and eukaryotic microbes found in skin swab samples
collected from a single frog species (Thoropa taophora[Cycloramphidae]) found across coastal mainland and island sites.
The island populations of T. taophora have experienced
fragmentation-induced genetic erosion at both neutral and immunogenetic
loci (17,35), offering an opportunity to examine the relationship
between host genetic diversity and microbiome diversity and structure.
We compared the bacteria we recovered from T. taophora skin swabs
to a database of amphibian microbiome bacterial isolates that have been
previously categorized as Bd inhibitory, Bd enhancing, and
having no effect on Bd . This allowed us to test for corresponding
ecological relationships between these bacteria and other
microeukaryotes found in the T. taophora microbiome. Our study
was designed to address the following research questions: (1) Does
geography and/or host genetic diversity structure the microbiome
community? (2) How is bacterial diversity and community assembly related
to microeukaryotic diversity and community assembly in the skin
microbiome? (3) Do bacteria that affect Bd growth have
predictable associations with other microbiome eukaryotes?