Surviving Chytridiomycosis: Differential Anti-Batrachochytrium dendrobatidis Activity in Bacterial Isolates from Three Lowland Species of Atelopus

In the Neotropics, almost every species of the stream-dwelling harlequin toads (genus Atelopus) have experienced catastrophic declines. The persistence of lowland species of Atelopus could be explained by the lower growth rate of Batrachochytrium dendrobatidis (Bd) at temperatures above 25°C. We tested the complementary hypothesis that the toads' skin bacterial microbiota acts as a protective barrier against the pathogen, perhaps delaying or impeding the symptomatic phase of chytridiomycosis. We isolated 148 cultivable bacterial strains from three lowland Atelopus species and quantified the anti-Bd activity through antagonism assays. Twenty-six percent (38 strains representing 12 species) of the bacteria inhibited Bd growth and just two of them were shared among the toad species sampled in different localities. Interestingly, the strongest anti-Bd activity was measured in bacteria isolated from A. elegans, the only species that tested positive for the pathogen. The cutaneous bacterial microbiota is thus likely a fitness-enhancing trait that may (adaptation) or not (exaptation) have appeared because of natural selection mediated by chytridiomycosis. Our findings reveal bacterial strains for development of local probiotic treatments against chytridiomycosis and also shed light on the mechanisms behind the frog-bacteria-pathogen interaction.


Introduction
The world's amphibians are facing a conservation crisis due to a variety of factors [1], with the newest and perhaps most insidious threat coming from an emerging global pathogen, the microscopic fungus Batrachochytrium dendrobatidis, or Bd [2]. This fungus is the etiological agent of chytridiomycosis, and has been implicated in many enigmatic declines worldwide [3,4,5]. In the Western Hemisphere, the disease has mainly affected populations in highland environments [6,7,8,9]. Species occurring in tropical regions at middle and high elevations are thought to face a higher risk presumably because (1) Bd grows faster at the low temperatures characteristic of montane areas [10,11], and (2) the amphibian immune capacity may decrease at lower temperatures [7,12]. Despite the current focus on montane declines, a growing body of evidence suggests that Bd may also be widespread and abundant in lowland amphibian populations, for example, in Australia and Central America [11,13,14], yet amphibian declines are rarely observed in the lowlands [14,15].
Individuals, populations and species of amphibians are known to vary in their susceptibility to chytridiomycosis, but the causes of this variation are not well understood [16]. Some species consist of both resistant and susceptible populations [6]. Some species maintain remnant populations that persist in the presence of Bd [17,18], whereas others are known to be infected but show no signs of chytridiomycosis [5,19]. The allegedly well developed immune system of amphibians, which shows both adaptive and innate responses [7], is probably involved on the resistance to infection [20,21]. Amphibians may also benefit from symbiotic microbiota present on their skin that function as a barrier against pathogens, including Bd, which specifically targets amphibian skin [22,23,24]. Previous work has demonstrated that bacterial isolates from the skin of certain species of amphibians exhibit strong antifungal activity against Mariannaea sp. [25] and Bd [22,26]. For example, the bacteria Janthinobacterium lividum and Lysobacter gummosus, produce the antimicrobial peptides violacein and 2,4-diacetylphloroglucinol respectively, which inhibit Bd growth in vitro [27,28]. Thus, variation in the cutaneous microbial community of amphibian skin may be a key factor in resistance to chytridiomycosis [26,28,29,30].
Conservation strategies focused on amphibians threatened by chytridiomycosis have mainly sought to prevent the spread of Bd to naïve populations and establish ex situ assurance colonies, yet solving the problem may require prophylactic treatments and enabling populations to persist with the pathogen [31]. Since the discovery of cutaneous bacteria with strong antifungal properties, increasing effort has focused on probiotic therapies [25,30]. Perhaps the most promising approach to date is to use beneficial symbiotic microorganisms, or their metabolic products, to increase the resistance to infection or disease through environment bioaugmentation or host therapy [16,30,31,32]. However, the majority of the studies had not considered the role of the environmental conditions on this host-pathogen interaction, since small variations could strongly affect anti-Bd activity of protective bacteria [33].
While most research on the immune-like properties of the cutaneous microbiota of amphibian skin has been conducted on temperate zone systems [32,34], the vast majority of amphibian diversity lies in the tropical realm, especially in South America [35]. If microbial communities on tropical amphibians differ from those on temperate hosts, then probiotic therapies optimized for these highly endangered tropical species are needed [36]. Furthermore, since bioaugmentation and other probiotic approaches would eventually require the anthropogenic introduction of bacteria into an environment, using locally obtained bacteria might minimize the risks associated with this procedure. Therefore, surveying the diversity and evaluating the anti-Bd action of microbial communities in the amphibian skin of Neotropical species could potentially expand the array of tools to help mitigate the impact of Bd.
One of the more severely impacted groups of amphibians are the montane harlequin toads of the genus, Atelopus (Anura: Bufonidae) with 80% of species listed as critically endangered [37]. In Colombia nearly all the 33 species of Atelopus have declined sharply, yet four lowland species are persisting at 0-600 masl [38]. For at least one species, A. elegans from Gorgona Island, we know that the pathogen has been present for at least five years without causing obvious disease or declines [39]. Among the various possible factors that may account for the persistence of A. elegans despite infection with Bd we hypothesize that cutaneous symbiotic bacteria may be a contributing factor to disease resistance in this population.
The aim of our study was to test whether the frog A. elegans harbors cutaneous bacteria capable of inhibiting Bd growth. We also compared the antifungal bacterial communities found in A. elegans with two other Atelopus species that persist in the lowlands without evidence of Bd infection. Thus, our null hypothesis is that bacteria isolated from the infected species (A. elegans) should exhibit stronger anti-Bd activity compared with bacteria from the other two toad species. Our long-term goal is to know whether these potentially beneficial strains could be used in bioaugmentation experiments or the metabolites they produce employed in host therapy to protect threatened Neotropical species.

Ethics statement
Procedures for capture and handling of live animals in the field were approved by the Colombian National Parks authority and the Ministry of the Environment, under permits DTSO 019-09, DTSO 001-09 and Nu 10-07032012.

Study species
Cutaneous bacterial microbiota was sampled from three latitudinally separated species occurring in coastal forests of Colombia. Atelopus aff. limosus, probably an undescribed species, occurs near the municipality of Capurganá (8u 379N, 77u 229W, 150 masl), close to the border between Panama and Colombia; Atelopus spurrelli, considered as Vulnerable (VU) by the IUCN [40], was sampled near the municipality of Arusí (5u309N, 77u319W, 90 masl) on the Pacific coast; and A. elegans, Critically Endangered (CR), was sampled in the insular Gorgona National Park located 56 Km off the Pacific coast of Colombia (2u479-3u69 N, 78u69-78u189 W, between 6-115 masl). A. elegans had tested positive (17%, 15 out of 78 individuals) for Bd at the time we conducted this study [39]. We sampled 80 individuals of A. spurrelli and 82 of A. aff. limosus and detected no infected individuals. We thus concluded that Bd might be either absent or rare in those populations.

Bacterial isolation
To obtain bacterial isolates from toads' skin, samples were collected from five A. spurrelli, eight Atelopus aff. limosus and seven A. elegans adults. Toads were manipulated using fresh disposable nitrile gloves and rinsed twice in sterile dechlorinated water to remove transient bacteria [25]. Individuals were swabbed on their left, right and ventral surfaces, hindlimbs and interdigital membranes, using a sterile cotton swab. Swabs were preserved in 2 mL cryovials containing 1 mL DS solution, a weak salt solution resembling pond water [41]. All swab samples were refrigerated within 24 h and processed within 48 hours after sampling. To isolate pure colonies, serial dilutions were performed until 1610 25 . To recover the highest possible number of bacterial morphotypes, each dilution was plated in R2A media in duplicate and incubated at 23uC for two days. Bacterial morphotypes were defined according to the macroscopic characteristics of the obtained colonies (i.e. color, form, elevation and margin). Single colonies of each bacterial morphotype were streaked on fresh nutritive agar plates until pure cultures were obtained. Each isolate was cryopreserved in nutritive broth with 30% glycerol at 280uC.

Batrachochytrium dendrobatidis growth inhibition assays
To test for anti-Bd activity in bacteria isolated from frogs' skin, we used growth inhibition assays [22]. At the time of the assays, no Colombian Bd strains were available, so we decided to use Bd strain JEL 423 (University of Maine, Orono, USA) isolated from Phyllomedusa lemur in lowland forests of neighboring Panama. Bd was grown in TGh media (10 g tryptone, 10 g agar, 4 g gelatine hydrolisate, 1000 mL distilled water) for three days at 23uC until maximal zoospore production was observed. Bd was then harvested from three Petri dishes and transferred to a sterile tube containing 16 mL of sterile water in order to obtain a solution with a high concentration of zoospores. Plates with 25 mL of TGh media were coated with 1 mL of the zoospore suspension, and plates were allowed to dry until most of the solution had diffused into the agar. Unknown or ''query'' bacteria from active isolated cultures were streaked in a line across one side of a Petri dish. Escherichia coli (strain DH5a) was streaked as a negative control in a parallel line on the opposite side of the Petri dish; this control streak was included to observe the Bd growth around a bacterial species that shows no inhibition activity on Bd growth. Each bacterial isolate was tested in triplicate and the plates were incubated at 23uC. On the third day of incubation, query bacteria were checked for signs of inhibition of Bd growth. For those isolates showing inhibition, photographs were taken and anti-Bd activity was estimated by using photometric techniques.
We found two kinds of evidence of anti-Bd activity in the tested cultures. First, we detected an inhibition zone around the query bacteria line. Second, we found strong variation in the density of Bd colonies growing throughout the culture medium outside the inhibition zone. To summarize both effects in a single measurement we took photographs of the cultures under similar lighting conditions with a dark background. Then, we estimated Bd growth by measuring the light (i.e. grey value) reflected at variable distances of the query bacteria line (Fig. 1). Data from the three plates were averaged in order to obtain the statistical unit of analysis. To facilitate comparisons we re-expressed grey values as percentages of the maximum light intensity reflected by a single Bd colony within the same Petri dish. All photometric measurements were conducted on the software ImageJ [42] after distance and light calibration.
To quantify Bd growth inhibition we modeled Grey values (hereafter Bd growth) as a function of distance to the query bacteria by using spline models on JMP statistics software (JMP, Version 8. SAS Institute Inc., Cary, NC, 1989-2007). Spline models combine several polynomial functions of relatively low degree (often cubic) to fit piecewise several subsets of X-Y values. A Lambda parameter allows modifying the shape of the curve between a pure cubic function (Lambda = 0) and eventually a straight line, as Lambda diverges to infinity [43]. The functions are blend smoothly allowing both interpolation of Y-values as well as estimation of model residuals.
The anti-Bd activity was compared between bacterial strains and frogs in two ways. First, we fitted a general spline model for all readings of Bd growth at every tested distance from all bacteria. The residuals of this model were then calculated by frog species and plotted to show between-frogs differences in the anti-Bd activity of their corresponding bacteria. Second, we fitted Bd growth against Distance to bacteria by using a single spline model for each bacterial strain. Because the Lambda value (l = 1) and the range of Distance values (0.17-1.77 cm) were identical for every model, we could interpolate Bd growth values at the first, second and third quartile of the growth function and save them as new variables. Then, we tested the effect of frog species on the Bd growth quartiles by using multivariate analysis of variance. The first approach allowed easy visualization of the general pattern; the second one permitted statistical testing.

Identification of bacterial isolates
Bacterial strains were identified by sequencing the 16S ribosomal gene. One colony of each morphotype was resuspended in 10 mL of distilled water in a 0.2 mL PCR tube and boiled for 7 min at 95uC. The solution was used directly in a PCR reaction with the universal eubacterial primers, 27F  and 1492R (59-GGTTACCTTGTTACGACTT-39) [44]. Thermocycling parameters were: an initial denaturation of 3 min at 95uC, followed by 35 cycles of 45 s at 95uC, 45 s at 52uC and 90 s at 72uC. A final Figure 1. Photometric method. Estimation of bacterial anti-Bd effect as a combined measurement of the inhibition halo around the query bacteria (thick width band, above) and the decreasing density of Bd colonies. A spline regression model is fitted to grey values (i.e. density of Bd colonies) against distance to the query bacteria (middle). Bd growth is then interpolated at the first (25), second (50), and third (75) quartiles (middle and below). The bacterial isolate in the right column (red) have a stronger anti-Bd effect than the bacteria in the left column (blue). doi:10.1371/journal.pone.0044832.g001 extension of 7 min at 72uC was performed to complete polymerization. PCR results were checked by electrophoresis in 1% agarose gels. Products were sent to Macrogen (Korea) for sequencing. DNA sequences were cleaned and assembled using Geneious [45]. 16S sequences were then identified using BLASTn [46] against the complete GenBank nucleotide database (http:// www.ncbi.nlm.hih.gov) and the Greengenes database (http:// greengenes.lbl.gov), using default parameter settings in both cases.

Results
A total of 148 cultivable bacterial morphotypes were isolated, 40 from Atelopus elegans, 85 from Atelopus aff. limosus, and 23 from A. spurrelli. In antagonism assays, we observed anti-Bd activity in 16 of 40 (40%) bacterial morphotypes from A. elegans, 16 of 85 from Atelopus aff. limosus (19%), and six of 23 (26%) from A. spurrelli. We detected an inhibition zone around the query bacteria line, and also found strong variation in the density of Bd growing in the Petri dish outside the inhibition halo (Fig. 1).
Among all three hosts we identified 12 bacterial species with anti-Bd activity belonging to six genera: Pseudomonas, Acinetobacter, Stenotrophomonas, Comamonas, Chryseobacterium and Elizabethkingia (Table 1). Toad species differed more in the composition of bacteria with anti-Bd activity (Anosim R = 0.42, P,0.0001, 9999 permutations) than in the whole bacterial communities (R = 0.51, P,0.0001) as inferred as well from Bray-Curtiss indices of dissimilarity: 0.71 (0.59 for all cultivable bacteria) between Atelopus aff. limosus and A. spurrelli, 0.84 (0.69) between Atelopus aff. limosus and A. elegans, and 0.75 (0.71) between A. spurrelli and A. elegans, ranging between 0 = identical and 1 = totally dissimilar. Two out of the three strains that exhibited the highest anti-Bd action, both tentatively assigned to P. tolaasii, were exclusive to A. elegans. The third one, tentatively assigned to P. putida, was isolated from A. spurrelli skins. A generalized spline model was fitted to Bd growth as measured within the Petri dish at variable distances to query bacteria (Fig. 1). The regression residuals were re-analyzed by toad species and showed clearly that bacterial strains isolated from A. elegans caused greater Bd growth inhibition in vitro as compared to the other toad species (Fig. 2). To formally compare the results of antagonism assays, we fitted a spline model (Bd growth as a function of distance to the query bacteria) for each bacterial strain (Fig. 3) and interpolated Bd growth values at the first, second and third quartile of the growth function. Strains isolated from the skin of A. elegans showed higher anti-Bd activity (i.e. lower Bd growth; Manova, Frog, F = 3.292, DF = 2, P = 0.0490) especially at the first and second quartiles of the growth curves (Manova, Frog x Quartile interaction, Wilks' lambda value = 0.619, DF = 4, P = 0.0024, N = 38 tests; Fig. 4).

Discussion
Our results demonstrated that 12 bacterial species isolated from three Atelopus species from the Colombian lowlands inhibit Bd growth. The composition of anti-Bd bacterial communities significantly varied between toad species or perhaps localities. The strongest anti-Bd activity was measured in bacteria isolated from A. elegans, the only species that tested positive for the pathogen. Our data suggest two evolutionary mechanisms behind amphibians' resistance to chytridiomycosis, an exaptation or an adaptation against the selective pressure represented by Bd.
Several studies have demonstrated that the presence of bacteria with antifungal activity in amphibian skin inhibits Bd and prevents morbidity and mortality. The addition of bacteria with antifungal properties seems to reduce the effects of the disease on experimentally infected individuals [25,26,29,30]. Our results revealed that two Pseudomonas (P. tolaasii and P. aeruginosa) species exhibited stronger (63 and 64% of maximum Bd growth as measured at the 25% quartile) anti-Bd activity than bacteria of any other isolated genera (69%). These results are supporting the findings of Lam et al. [47], and Walke et al. [48], which recognized that in Rana muscosa and R. sierrae, and in Hyalinobactrachium collymbiphyllum respectively, bacterial strains of the genus Pseudomonas were common residents and showed strong inhibition of Bd growth. Our results allow us to suggest that those bacterial species producing a larger inhibition zone may be good candidates to be used in probiotic therapies. Nevertheless, since the laboratory conditions probably differ from the host's skin, it is important to run anti-Bd assays using bacteria with proved efficiency for delaying or inhibiting Bd growth in vitro.
McKenzie et al. [49] suggest that skin bacteria are strongly species-specific across sites or even among species occurring in the same pond. The low number of shared bacteria found in this and other research [49], suggest that studies at the local level are important for understanding the mechanisms behind the frogpathogen-bacteria interaction. Of course, the proportion of bacteria shared among toad species is probably underestimated since we worked only with cultivable bacteria. In any case, since bacterial communities may covary with distance or geographic conditions, it is important to look for candidate local strains with strong anti-Bd activity before considering bioaugmentation assays and eventual probiotic treatment of chytridiomycosis.
Atelopus elegans, the only species in our study that tested positive for Bd, holds the skin bacteria with the strongest anti-Bd action. Although admittedly non-replicated, the pattern suggests that the species' current bacterial community may have resulted from natural selection represented by Bd infection. If so, we would be witnessing a post-infection or post-decline event in A. elegans, where frogs and pathogen are now coexisting after a critical period of strong natural selection. Alternatively, the anti-Bd bacterial microbiota of the three toad species may represent an exaptation, a pre-existing (extended) phenotype that eventually subserves protection against Bd pathogenic infection. To discriminate between both the adaptation and exaptation scenarios, we need longitudinal (i.e. historical within the same population) data or geographic (i.e. between populations) comparisons on bacterial composition of infected and uninfected frogs. In the latter case, Woodhams et al. [26], have already showed that Rana muscosa from Conness in Yosemite National Park hosts a significant proportion of anti-Bd bacteria that inhibited Bd growth and persisted for six years in the presence of the chytrid fungus, whereas the Sixty Lake population was devastated by chytridiomycosis. For evolution to occur on skin microbiota, the ability to acquire or maintain certain bacteria should be heritable. Frog skins might vary in their habitability to different bacterial taxa due to their skin secretions or skin humidity; also, between species differences in habitat use may affect the probability of acquiring certain bacterial taxa. Both represent just speculative hypotheses that deserve rigorous testing.
We cannot rule out the possibility that high temperatures delay or reduce Bd growth in the studied Atelopus species, since they inhabit lowland forests of Colombian coasts with average annual temperatures around 27uC. High environmental temperatures can be directly involved with the growth control of the pathogen. Laboratory tests have shown that at temperatures higher than 25uC the pathogen either decreases zoospore production or dies [10]. Moreover, other studies showed that Bd-infected frogs exposed to warmer temperatures lived longer [50] and that the prevalence of infection was considerably lower [51]. In our case, high temperatures may have interacted with species-specific microbiota in allowing the survival of A. elegans despite the infection by Bd. The effect of environmental conditions in shaping the interaction between frogs and Bd has been already suggested [33]. Also, strain-level differences in bacteria and overall community structure may prohibit probiotics discovered on one frog species from persisting on another one. For example, Janthinobacterium lividum, a bacterium isolated from North American amphibians, failed to prevent or delay mortality in Bd exposed individuals of Atelopus zeteki [36]. Likewise, we cannot rule out a synergistic or detrimental anti-Bd effect of the whole consortium of skin microbiota, because we only evaluated the effects of individual bacteria against Bd. It is well known that bacterial consortia are frequently established to carry out several processes that are otherwise harder to accomplish when grown alone [52], for example the oxidation of anaerobic methane [53,54], the metabolism of explosive compounds [55] and the enhancement of bioremediation strategies [56], among others. It has been shown that disruption in antifungal microbial communities is likely to lead to a breakdown of the protective effects of beneficial microorganisms and may lead to disease emergence [57]. Further studies are thus needed to better understand how the bacterial symbionts of frogs' skin interact among themselves and with their amphibian hosts, and how the activities performed by the skin bacteria could benefit diseased frogs.
Colombia has one of the most diverse amphibian faunas of the world with 751 described species [58], but the current conserva-tion status of many populations and species remains unknown. This study is at the forefront of attempts to seek local bacteria involved in resistance to amphibian chytridiomycosis [59]. We found that amphibian skin microbiota is an important component of disease resistance, and moreover, we describe Neotropical bacteria that provide promising avenues for disease mitigation. In particular, probiotic therapy may be applied as a management tool to reduce the vulnerability of Neotropical amphibians to the devastating effects of Bd as has been demonstrated with frogs from temperate zones, where the treatment with beneficial bacteria has been shown to be highly successful to protect susceptible amphibian species from Bd infections [59].  . Residuals of Bd growth after the toad species from which the bacteria were recovered. Residual analysis of the aforementioned model to summarize differences between bacterial isolates grouped after the frog species from which they were recovered. Atelopus elegans is the only species we found infected with Bd in its natural habitat. doi:10.1371/journal.pone.0044832.g004