Direct and Indirect Horizontal Transmission of the Antifungal Probiotic Bacterium

Transcription

1 AEM Accepted Manuscript Posted Online 12 February 2016 Appl. Environ. Microbiol. doi: /aem Copyright 2016, American Society for Microbiology. All Rights Reserved. 1 2 Direct and Indirect Horizontal Transmission of the Antifungal Probiotic Bacterium Janthinobacterium lividum on Green Frog (Lithobates clamitans) Tadpoles 3 4 Eria A. Rebollar 1 #, Stephen J. Simonetti 1, William R. Shoemaker 1 *, Reid N. Harris Department of Biology, James Madison University, Harrisonburg, VA, USA # Address correspondence to Eria A. Rebollar, ea.rebollar@gmail.com. * Present address: Department of Biology, Indiana University, Bloomington, IN, USA Running title: Transmission of J. lividum on green frog tadpoles

2 Abstract Amphibian populations worldwide are being threatened by the disease chitridiomycosis, which is caused by Batrachochytrium dendrobatidis, Bd. To mitigate the effects of Bd, bioaugmentation of antifungal bacteria has been shown as a promising strategy. One way to implement bioaugmentation is through indirect horizontal transmission, defined as the transfer of bacteria from a host to the environment and to another host. In addition, direct horizontal transmission among individuals can facilitate spread of a probiotic in a population. In this study, we tested whether the antifungal bacterium Janthinobacterium lividum could be horizontally transferred, directly or indirectly, in a laboratory experiment using Lithobates clamitans tadpoles. We evaluated the ability of J. lividum to colonize the tadpoles' skin and to persist through time using culture-dependent and cultureindependent techniques. We also tested whether the addition of J. lividum affected the skin community in L. clamitans tadpoles. We found that transmission occurred rapidly by direct and indirect horizontal transmission, but was more effective with indirect transmission that included a potential substrate. Even though J. lividum colonized the skin, its relative abundance on the tadpole skin decreased over time. The inoculation of J. lividum did not significantly alter the skin bacterial diversity of L. clamitans tadpoles, which was dominated by Pseudomonas. Our results show that indirect horizontal transmission can be an effective bioaugmentation method. Future research is needed to determine the best conditions, including the presence of substrates, in which a probiotic can persist on the skin so that bioaugmentation becomes a successful strategy to mitigate chytridiomycosis. 2

3 Introduction Animal and plant populations in nature are threatened by several emerging infectious diseases (1-6). In the case of amphibians, many populations have been devastated by chytridomycosis, a disease caused by Batrachochytrium dendrobatidis (Bd) (Berger et al., 1998) and the recently discovered B. salamandrivorans (Bsal) (7,8). Bd has caused extinctions of amphibian species in the tropics including tropical Australia (9) and the neotropics in Central America (10,11). However, some species are considered resistant or tolerant since they are persisting in the wild despite the presence of the pathogen (12, 13). In order to reduce the risk of population declines or extinctions in susceptible amphibians, effective mitigation protocols must be developed Amphibian defenses against Bd include the adaptive immune system, the innate immune system (14), which includes the production of antimicrobial peptides (AMPs) (15) and alkaloids (16), and skin symbiotic bacteria that produce antifungal metabolites (17-21). The adaptive immune system s effectiveness appears to be limited by Bd s secretions that disrupt lymphocyte proliferation (22). This result may explain why vaccination strategies are either ineffective or require multiple pathogen exposures to achieve immunity (23). While some amphibian species produce AMPs that are effective against pathogens like Bd (24), many species do not produce AMPs (25, 26). Moreover, these innate immune defensive secretions may be difficult to manipulate as part of a disease mitigation management strategy. On the other hand, certain cutaneous symbiotic bacterial species 3

4 isolated from amphibians secrete antifungal metabolites that are effective against Bd (18, 19, 27), while other bacterial species may deter Bd from colonizing an amphibian host (28). In addition, populations with a higher proportion of individuals with anti-bd bacteria tend to persist with the pathogen, while populations with lower proportions tend to dramatically decline (29,30). These anti-fungal bacteria can be used to prevent or mitigate Bd or Bsal infections using a bioagumentation strategy (31) Bioaugmentation is the process of increasing the population density of locally occurring anti-fungal bacteria on amphibian hosts (31). Bioaugmentation methods include environmental and individual probiotic inoculation (31, 32). Probiotic bacterial inoculations have been effective in certain laboratory and mesocosm trials (33, 34), although not all of them have been successful (35-36). A key example is that of Janthinobacterium lividum, a gram-negative bacterium, which produces an antifungal metabolite called violacein (18). Bioaugmentation of J. lividum was effective in laboratory and field trials in preventing mortality due to Bd in the frog species Rana muscosa (33, 34), but has not been effective in the Panamanian golden frog, Atelopus zeteki (35) The success of probiotic bioaugmentation in an at-risk amphibian population would increase if probiotic bacteria could be horizontally transmitted from an inoculated individual to un-inoculated individuals by direct contact in the population or by transmission from the environment to amphibians (38). Environmental transmission has been shown in a laboratory study in which J. lividum was successfully introduced in soil 4

5 and transmitted from soil to salamanders, resulting in a reduced risk of Bd infection (32). A third option for establishing a probiotic in an amphibian population or community is through indirect horizontal transmission, which is defined as the transfer of bacteria from a host to the environment and then to another host. The use of indirect horizontal transmission may allow for the colonization of a probiotic bacterium in an at-risk amphibian population through the inoculation of relatively few individuals Moreover, the success of a probiotic will depend on whether it can become established in the existing cutaneous bacterial community and persist for an extended period of time while inhibiting the pathogen or by modulating the host's immune response to inhibit the pathogen (39-41). Successful probiotic establishment is likely a function of the cutaneous microbial community structure, the ecological interactions occurring in the microbial community, and the competitive abilities of probiotic bacteria. Ideally, the probiotic would be added and persist with relatively little perturbation of the existing community structure, as seen in other systems (42, 43). Alternatively, the probiotic might alter the cutaneous community structure and eventually reach an alternative stable state (44). For the long-term health of the host, it is important for the probiotic to not displace microbes that have relevant symbiotic interactions with the host In this study, we tested whether the skin bacterium J. lividum could be horizontally transferred directly or indirectly, by way of the environment, in a laboratory experiment using green frog tadpoles (Lithobates clamitans) as a model system. The colonization of a 5

6 probiotic bacterium in the larval stage may be more effective than colonization in the adult stage, since the larval skin community has a lower diversity than the adult stage (45, 46), and the host immune system is not yet fully developed (14). We hypothesized that transmission of the probiotic would be greatest when both direct and indirect horizontal transmission was possible. We evaluated the ability of J. lividum to colonize the tadpoles' skin and to persist throughout the experiment. We tracked the fate of J. lividum using both culture-dependent and culture-independent techniques and compared the congruence of methods. We also described the bacterial community structure of the tadpoles skins and tested whether the addition of J. lividum affected the skin community structure in L. clamitans tadpoles Materials and Methods Experimental Design The experiment consisted of two treatments with their appropriate controls (Figure 1a). In the first treatment we tested for both direct plus indirect horizontal transmission (DIH), and included two free-swimming tadpoles per container, with one tadpole inoculated with the probiotic bacterium J. lividum. In this treatment, tadpoles could come into direct contact, which allows for the possibility of direct horizontal transmission. In addition, J. lividum from an inoculated tadpole could migrate to the water and subsequently colonize the inoculated tadpole partner, which allows for the possibility of indirect horizontal transmission. In the second treatment we tested for indirect horizontal transmission only (IH), and included two tadpoles in a container separated by two plastic mesh barriers with 6

7 only one tadpole inoculated with J. lividum. The mesh was composed of bleached fiberglass window screening and served to prevent direct contact between tadpoles, allowing only indirect horizontal transmission to occur. Controls for both treatments (CIH and CDIH) were identical to their respective treatments with the exception of neither tadpole being inoculated with J. lividum (Figure 1a). Instead tadpoles from controls received a mock bath with sterile artificial pond water, which is known as provasoli medium (47). For this experiment, there were a total of 20 4L-tanks. Each tank was filled with 3L of provasoli medium: seven tanks per experimental treatment (14 tanks total for both treatments) and three tanks for each control (6 tanks total for controls). The experiment was done in an environmental chamber with a constant temperature of 17ºC and light cycle of 12h/12h. The animal care IACUC protocol A03-12 was approved for conducting this experiment Development of a J. lividum rifampicin resistant strain used for probiotic treatments A rifampicin resistant strain of J. lividum was used to track the bacterial abundance throughout the experiment (32). The purpose of this was to have a bacterial strain that could be distinguished from all other bacterial strains that are naturally present on the tadpole skin and that are not resistant to rifampicin. To obtain a rifampicin resistant J. lividum strain, a J. lividum culture was plated onto a 1% tryptone plate with a rifampicin gradient that ranged from 0 to 0.01% rifampicin (from 0 g/l to 0.1 g/l). The strain used in this experiment was isolated from the skin of the salamander Hemidactylium scutatum collected in the George Washington National Forest, Rockingham Co., VA, USA, and it has 7

8 been used in previous probiotic trials due to its capacity to produce antifungal metabolites (18, 33, 34). Bacterial colonies that grew closest to the highest concentration of rifampicin were isolated and re-plated onto the same kind of plate. This process was repeated twice, which was enough to isolate a completely rifampicin resistant strain. Plating the bacteria on a 0.01% rifampicin plate (0.1 g/l) and observing its growth confirmed the capacity of this J. lividum strain to grow in the presence of this antibiotic Data collection (swabbing) We collected 40 tadpoles from George Washington National Forest, Rockingham Co., VA, USA (10 April 2014) that were between Gosner stages 26-36, making sure tadpoles were no smaller than 1.9 cm. Once in the lab, tadpoles were placed into two separate 33L-tanks for 12 days. Each tank contained 15L of sterile provasoli medium. During these 12 days tadpoles were fed every three days with commercially available tadpole food (Zoo Med Aquatic Frog & Tadpole Food, USA). The tanks were constantly aerated with separate air pumps. Provasoli medium was replaced every two days with fresh sterile provasoli medium in order to reduce the risk of altering water chemistry. The 12-day acclimation period allowed the tadpoles to become habituated to their new environment before the experiment began. After the acclimation period (Day -1, Figure 1b) each tadpole was swabbed back and forth five times on both the lateral sides and on the mouth. Swabbing did not have any noticeably negative effect on the tadpoles. Swabs were then plated onto rifampicin plates to test for the occurrence of naturally rifampicin-resistant strains on the tadpoles. After 8

9 pre-swabbing, each tadpole was placed individually into a plastic sterile container for the purpose of randomization, each of which contained 50mL of sterile provasoli medium. A random number generator was used to determine the assignment of individual tadpoles to treatments. The location of treatment tanks was also randomized on shelves in the environmental chamber. J. lividum baths were prepared for tadpoles assigned to the experimental treatment tanks in the following way: 700 ml of J. lividum in 1% tryptone liquid culture were incubated until the culture reached a concentration of 1.7X10 7 cells of J. lividum/ml, as determined by optical density units (ODs) using a spectrophotometer Multiscan GO (Thermo Scientific USA). The cell concentration estimated by ODs was calculated based on a growth curve of the same bacterial strain. The culture was divided into ml falcon tubes, which were centrifuged at 5,000RPM (G-force = 3,214) for 10 minutes. Remaining liquid was decanted, and cells were resuspended in an equivalent volume of sterile provasoli medium. A second centrifugation was performed, the liquid was decanted, and cells were finally resuspended in 50 ml of sterile provasoli medium. Resuspended cells in each 50mL tube were added to 250mL-containers that contained 200mL of sterile provasoli medium for a total volume of 250mL. These centrifugation steps were necessary to remove waste products from the culture medium. The final number of J. lividum cells in each bath container was 8.5 X10 8 cells. Controls were identically constructed with the exception of sterile provasoli medium being used in place of the J. lividum liquid culture. Tadpoles from each treatment (20 tadpoles including controls) were individually rinsed with 25mL of sterile provasoli medium to remove any transient bacteria before being placed into individual J. lividum (or 9

10 mock) baths for 24 hours. This was considered day 0 of the experiment. Following the 24 h probiotic baths, tadpoles were placed in their respective tanks. This was defined as day 1 of the experiment (Figure 1). Tadpoles were swabbed on days 3, 10, and 17 of the experiment (Figure 1b). During swabbing, each tadpole was removed from its tank, rinsed with 25mL of sterile provasoli medium to remove transient bacteria, and then swabbed 5 times back and forth on both the lateral side and the mouth. Swabs were then placed into 1.5mL eppendorf tubes containing 1mL of PBS solution Data collection (culturing and colony counting) The bacteria from each swab were extracted by placing each swab in a separate tube in a shaking incubator at 25 C and 500 RPM for 30 minutes. 1:10 and 1:100 dilutions were obtained from the swab extracts using PBS buffer. This resulted in three dilutions per swab/sample (D10 0, D10-1, D10-2 ). After the first swabbing day, it was determined that the D10-2 dilution was no longer necessary for the remaining days due to low colony counts. After making the dilutions, two replicates of each extract sample were plated onto 1% tryptone-0.01% rifampicin plates using sterile 3mm glass beads. Colony forming units (CFUs) were counted after 48 hours of incubation at 21 C Statistical analyses: transmission experiment We addressed the following hypotheses: (1) If probiotic bioaugmentation was successful, tadpoles in experimental treatments would differ from their respective controls in terms of CFU counts of J. lividum, (2) CFU counts for DIH treatment will differ from IH treatment, 10

11 and (3) if maximal transmission occurs between tadpoles within each treatment, then uninoculated tadpoles will end up with equivalent CFU counts as their co-housed inoculated partners. To test these hypotheses we performed the following comparisons: (1) Each treatment (DIH or IH) was compared to its controls on each swabbing day (days 3, 10 and 17), (2) Treatments (DIH or IH) were compared to each other on each swabbing day (days 3, 10 and 17), (3) inoculated and un-inoculated tadpoles were compared within each treatment (DIH or IH) on each swabbing day (days 3, 10 and 17). All comparisons based on CFU counts were assessed with a Kruskal-Wallis test using R (48). A nonparametric analysis was used because the data were not normally distributed according to the Shapiro-Wilk normality test (W = , p-value < 2.2e-16) performed in R (48). All controls had no presence of J. lividum (CFUs=0), therefore we concluded that there was no contamination in the experiment Molecular methods and sequencing Whole genomic DNA was extracted from 78 swab samples from days -1, 3 and 17 of the experiment which were a randomly-selected subset of the total number of samples: IH (N=24), DIH (N=24) and their respective controls CIH (N=15), CDIH (N=15) using the DNeasy Blood and Tissue kit (Qiagen, Valencia CA, USA) according to the manufacturer s instructions including a pretreatment with lysozyme. DNA extracted from swabs was used to amplify the V4 region of the 16S rrna gene using barcoded primers (F515/R806) and PCR conditions adapted from Caporaso et al. (49). Amplicons were quantified using Quantifluor TM (Promega Madison, WI). Composite samples for sequencing were created 11

12 by combining equimolar ratios of amplicons from the individual samples, followed by cleaning with the QIAquick PCR clean up kit (Qiagen, Valencia CA, USA). Barcoded composite PCR products were sent to the Dana Farber Cancer Institute s Molecular Biology Core Facilities (Boston, MA) for MiSeq Illumina sequencing using a 250 bp single read strategy S amplicon data processing The 250bp single reads were filtered and processed with the Quantitative Insights Into Microbial Ecology (QIIME) pipeline (50). Sequences were de-multiplexed and filtered to retain high quality reads using the following filtering parameters: no N characters were allowed in retained sequences, no errors in barcode sequence were allowed, a minimum of 5 consecutive base pairs were needed to include a read, and a maximum of 5 consecutive low quality base pairs were allowed before truncating a read. After filtering, 10,978,923 sequences were retained for the 78 samples. De-multiplexed and filtered sequences were clustered into operational taxonomic units (OTUs) at a sequence similarity threshold of 97% with the UCLUST method (51). Sequences were matched against the Greengenes database (52), and those that did not match were clustered as de novo OTUs at 97% sequence similarity. Taxonomy was assigned using the RDP classifier (53) and the Greengenes database. Representative sequences were aligned to the Greengenes database with PyNAST (54), and a ML phylogenetic tree was constructed with FastTree 2 (55). The OTU table for each data set was filtered using a minimum cluster size of 0.001% of the total reads (56). The final 12

13 266 rarefied OTU table had OTUs and reads per sample S amplicon data analysis To determine whether OTUs from the genus Janthinobacterium (determined by using the RDP classifier and greengenes database) changed in relative abundance across time in each treatment as compared to its respective control, we calculated differences in relative abundance between days -1 and 3 (dif[-1/3]) and between days 3 and 17 (dif[3/17]). Oneway ANOVAs were calculated to determine the effect of each treatment for dif[-1/3] and dif[3/17]. To determine if J. lividum bioaugmentation altered the tadpoles skin community structure, alpha and beta diversity indices were calculated. For alpha diversity we obtained Faith s Phylogenetic Diversity (PD) and Shannon index. To determine the effects of treatment and swabbing day we performed a two-factor ANOVA for PD and Shannon index. For beta diversity, a Weighted Unifrac distance matrix was calculated. Adonis non-parametric analysis was performed to test for differences in beta diversity among experimental treatments within each swabbing day (Days -1, 3 and 17). Weighted Unifrac matrices and Adonis tests were calculated using QIIME (50). Finally, OTU relative abundances of all individuals were grouped by day and treatment and these were represented in stacked barchart graphs. This analysis was done by grouping OTUs at the genus level based on the classification obtained through the RDP classifier using QIIME. The 16S rrna amplicon sequences obtained in this study through Illumina sequencing were deposited in the NCBI Sequence Read Archive (SRA study accession number: SRP067767). 13

14 Results J. lividum inoculation was successful in both indirect and direct plus indirect horizontal treatments By day 3 of the experiment, J. lividum inoculation was successful in both indirect (IH) and direct plus indirect (DIH) horizontal transmission treatments: 92.85% of the tadpoles (13 out of 14 individuals) in each treatment had rifampicin resistant colonies that corresponded to the incoulated rifampicin restistant J. lividum strain. Tadpoles from both DIH and IH treatments at day 3 had significantly higher numbers of J. lividum on their skins than did their respective controls, which all had no J. lividum colonies (DIH: H=10.8, 1 df, p=1 X 10-3 ; IH: H=10.8, 1 df, p=1 X 10-3 ), indicating that J. lividum inoculation was successful in both treatments and that there was no contamination throughout the experiment. In addition, all swabs collected before the inoculation with rifampicin resistant J. lividum (day -1) showed no colony forming units (CFUs) on rifampicin plates, indicating that no natural rifampicin resistant bacteria was present on the tadpoles' skin before starting the experiment Indirect horizontal treatment had a higher concentration of J. lividum over time than direct plus indirect treatment

15 CFUs of J. lividum from DIH and IH treatments were not significantly different on day 3 (H=2.81, 1 df, p=0.09; Figure 2), however IH treatment had a significantly higher number of CFUs in comparison to DIH on days 10 and 17 (Day10: H=21.33, df = 1, p-value = 3.8 x10-6 ; Day17: H=12.08, df = 1, p-value = 5 x10-4 ; Figure 2). In addition, CFUs from the IH treatment remained significantly higher when compared to controls throughout the rest of the experiment (Day10: H=12.32, df = 1, p-value = 4.4 x10-4 ; Day17: H=7.04, df = 1, p- value = 7.9 x10-3 ). In contrast, CFUs from the DIH treatment were no longer significantly different from controls on days 10 and 17 (Day10: H=0.90, df = 1, p-value = 0.34; Day17: H=0.42, df = 1, p-value = 0.51) When comparing the number of CFUs between originally inoculated tadpoles and uninoculated tadpoles within each treatment, we found that CFUs on tadpoles from IH treatment were not significantly different throughout the experiment (Day3: H= 0.49, df = 1, p-value = 0.48; Day10: H=1.8, df = 1, p-value = 0.17; Day17: H= 0.51, df = 1, p-value = 0.47; Figure 3b). In contrast, in the DIH treatment, inoculated tadpoles had a significantly higher number of CFUs than un-inoculated tadpoles on day 3 (H= 3.92, df = 1, p-value = 0.04; Figure 3a). We did not perform this comparison for days 10 and 17 since by then, CFUs from DIH were not significantly different to their controls (CFUs=0 see above) Overall, these results indicate that transmission was readily achieved in both treatments, and the number of CFUs differed between experimental treatments, with IH treatment showing a higher amount of J. lividum CFUs over time than DIH (Figure 2 and 3). 15

16 Presence of J. lividum on the tadpoles decreased over time in both treatments In both treatments, the number of J. lividum CFUs decreased throughout the experiment (Figure 2). This trend was also obtained in the 16S amplicon sequencing data (Figure 4a). Since the rifampicin resistant strain of J. lividum can not be distinguished from the natural J. lividum present on the tadpoles' skin using 16S amplicon sequencing, we analyzed the additive relative abundance of the 25 OTUs that were classified as members of the genus Janthinobacterium. For both IH and DIH treatments, the median relative abundance of Janthinobacterium increased in relative abundance by day 3 of the experiment in comparison to preswab day -1 and went back down by day 17 (Figure4a). However only IH treatment showed significant differences across time in comparison to their controls (IH dif[-1/3] = F(1,12) = 7.57, p = 0.01, IH dif[3/17] = F(1,12) = 7.28, p = 0.01), but not DIH treatment with respect to their controls (DIH dif[-1/3] = F(1,10) = 3.15, p = 0.10, DIH dif[3/17] = F(1,10) = 3.31, p = 0.09). Moreover, the results obtained with culturedependent and culture-independent methods indicate that the presence of J. lividum on the tadpoles' skin in both treatments decreased after day 3 independently of differences found between treatments, which suggests that most of the Janthinobacterium detected in the 16S amplicon analysis corresponded to the inoculated rifampicin resistant J. lividum strain J. lividum inoculation did not alter the diversity of skin microbial communities. 16

17 The skin bacterial community of L. clamitans tadpoles was dominated by OTUs from the genus Pseudomonas (Figure 4c). Out of the OTUs identified in these communities 174 belong to the genus Pseudomonas, and they accounted for the majority of the total relative abundance (Mean= 79.42%, SD = +/- 2.76, Figure 4c). More importantly, within the Pseudomonas genus there were two dominant OTUs that were present in all individuals and were both classified as Pseudomonas veronii. Mean relative abundance of OTU across individuals was Mean = 64.69% (SD = +/- 8.23) and OTU was Mean = 10.88% (SD = +/- 1.23). The next more abundant genera were Sanguibacter (Mean = 3.3%, SD = +/- 1.0) and Xanthomonas (Mean = 2.76%, SD = +/- 1.2; Figure 4c). Moreover, OTUs from the Janthinobacterium genus were naturally present on the tadpole's skin (day -1) but at lower relative abundances (Mean = 0.16%, SD = +/- 0.01) We analyzed the community structure of L. clamitans tadpoles across treatments and time (Figure 4b), and we found no significant differences in alpha diversity (PD and Shannon): A two-factor analysis of variance showed no significant effect either of treatment (PD: F(3,2) = 0.72, p=0.54; Shannon F(3,2) = 0.23, p = 0.87) or of swabbing day (PD: F(3,2) = 0.44, p = 0.64; Shannon F(3,2) = 0.30, p = 0.73). There was no significant interaction between treatment and swabbing day (PD: F(3,2) = 1.19, p = 0.32; Shannon F(3,2) = 0.91, p = 0.48)

18 We calculated the beta diversity among samples based on weighted unifrac distances and we found no significant differences among treatments within each swabbing day (Day -1: F(3,22) = 1.21, R2= 0.14, p=0.26; Day 3: F(3,22) = 1.81, R2= 0.19, p=0.07; Day 17: F(3,22) = 1.31, R2= 0.15, p=0.19). In sum, the results obtained for alpha (phylogenetic diversity) and beta diversity (weighted unifrac distances) indicate that the inoculation of J. lividum did not significantly influence the diversity of the tadpoles' skin communities even when the probiotic bacterium reached a high relative abundance detected by cuture-dependent and culture-independent methods Discussion In this study we examined the ability of the antifungal probiotic bacterium J. lividum to colonize and persist on the skin of L. clamitans tadpoles. We found that transmission from inoculated to un-inoculated tadpoles occurred in both treatments, but was more effective when direct contact between tadpoles was not allowed (Indirect horizontal treatment, IH) or when a net was included, than when they were allowed to swim freely and to contact each other (direct plus indirect horizontal treatment, DIH) without the presence of a net. Even though J. lividum was able to colonize the tadpoles skin, the relative abundance of this bacterium decreased over time in both treatments, and in the case of DIH it was not able to persist after 17 days. Moreover, the inoculation of J. lividum did not significantly alter the skin bacterial diversity of L. clamitans tadpoles during the course of the experiment, which was dominated by gram-negative Pseudomonas OTUs. 18

19 Despite our expectations, the combination of direct plus indirect horizontal transmission was a less effective method of transmission. This result could be explained by several hypotheses: 1) tadpoles in constant contact could lead to the loss of bacteria from the skin through friction, 2) contact between tadpoles could cause higher stress levels on the hosts, and this could in turn alter the skin community (57), and 3) the presence of fiberglass netting in the IH treatment could have served as a reservoir for bacteria, and it could therefore promote re-colonization of J. lividum through the course of the experiment (32, 58). Considering that the skin microbial community was not drastically affected during the course of the experiment, we consider the first two hypotheses to be the least feasible since they both predict changes in the skin microbial community due to either sloughing or stress responses. Considering the third hypothesis, it is possible that the fiberglass netting separating the tadpoles could be an additional substrate for J. lividum growth, allowing this bacteria to become more abundant in the environment and possibly promoting re-colonization on the tadpoles skin. This last hypothesis could also explain why the IH treatment led to longer persistence of J. lividum on partner tadpoles. Even though this is an artificial setting, tadpoles in their natural habitat are surrounded by organic material such as sediment and leaves that could serve as a bacterial reservoir. Therefore, further experiments that incorporate natural substrates (instead of a fiberglass netting) will be needed to help explain the differences between treatments in this experiment

20 An effective probiotic should be able to colonize and persist on the amphibian skin indefinitely or at least long enough for the host s adaptive immune system to become effective against the pathogen (31; Figure 1b). With amphibians, this means that the probiotic needs to persist until the adaptive immune system becomes fully functional, which is likely to be the juvenile or adult stage. Our results show that in the case of L. clamitans tadpoles, J. lividum was able to colonize but it decreased through time in both treatments. The community was dominated by OTUs from the genus Pseudomonas, and it s possible that these taxa and other bacterial species might have eventually excluded J. lividum by secreting antibiotics or by another competitive mechanism (59). It is unlikely that a host immune response caused the reduction in J. lividum since the tadpole s adaptive immune system is not fully developed at this development stage (60). It is also possible that J. lividum does not contain the necessary adaptations to colonize and persist on the skin of the host species, however, based on our culture-independent data, native strains of Janthinobacterium were naturally present on the tadpoles skin but at very low relative abundances. Moreover, J. lividum has been naturally found in other North American amphibian species like Plethodon cinereus (61), Hemidactylium scutatum (34, 62), Notophthalmus viridescens (63) and Rana muscosa (29, 30). In addition, J. lividum has been found in aquatic and soil environments (64, 65), and the production of violacein has been detected in liquid and solid media as well as in biofilms (65, 66). These capabilities make J. lividum a good probiotic candidate that could be inoculated directly in amphibians or in the environment

21 Tracking the addition of probiotic bacteria in amphibians has been done through culturedependent methods (32) and through culture-independent techniques including DGGE (36), T-FRLPs (37) and qpcr (33-35). In this study we tracked the presence and abundance of the probiotic bacterium J. lividum through time using two different techniques: 1) bacterial culturing using selective media for rifampicin resistant strains (32) and 2) 16s amplicon sequencing which is a culture-independent technique increasingly used to describe the bacterial community structure in amphibians (13, 45, 58, 67, 68). Despite the great advantages of using 16S amplicon sequencing for community profiling, one caveat is that this method does not distinguish between the natural J. lividum and the rifampicin resistant inoculated strain. Despite this, we were able to detect an increase of Janthinobacterium that was consistent with the results obtained using the culturedependent method. We conclude that both techniques are effective for detecting and quantifying J. lividum. In particular, 16S amplicon sequencing not only allowed us to detect changes in OTUs that belonged to the Janthinobacterium genus but also we were able to describe the community structure in L. calmitans tadpoles before and after the addition of J. lividum This study is one of the few that has examined the skin community structure in amphibians at the larval stage. Larval skin communities differ from their aquatic environment (45) and are generally dominated by one taxon (45, 46, 69). Species like Ambystoma tigrinum, Pseudacris triseriata and Bufo boreas boreas are dominated by Betaproteobacteria (46, 69), whereas Rana cascadae tadpoles are dominated by 21

22 Pseudomonas (45). In this study we identified 5,981 bacterial OTUs on L. clamitans skin, and these communities were dominated by one Pseudomonas OTU. Several members of the genus Pseudomonas are known to produce antifungal and antibiotic compounds, and their dominance on the amphibian skin may indicate a defensive function (13, 30, 45, 68, 70). A recent study on Bufo boreas boreas tadpoles that compared bacterial taxa with a Bd-inhibitory bacterial OTUs database found that tadpoles hosted significantly more inhibitory OTUs than other life stages or environmental samples (46). In this respect, it is possible that Pseudomonas OTUs from L. clamitans tadpoles' skins are playing a defensive function at the larval stage. To our knowledge there are no published studies on L. clamitans tadpoles' susceptibility to Bd or the presence of antifungal bacteria on their skin. Thus, the protective function of Pseudomonas in this host system remains to be explored As one caveat, the high relative abundance of Pseudomonas in this study was detected after the tadpoles were housed in a laboratory for 12 days using the acclimation period and kept under controlled conditions. Previous studies on adult frogs and salamanders in captivity have shown a reduced diversity on the skin microbial community structure compared to the situation in nature (58, 71). Despite the changes due to captivity, both of these studies suggest that the dominant members of the community can be preserved through time under artificial conditions. Even though there are not similar studies on tadpoles it is likely that the same effect occurs in the larval stage

23 In conclusion, our results show that indirect horizontal transmission could be effectively used to deliver probiotics in a natural environment. This transmission method would use a few individuals as vectors to transmit the probiotic to other amphibians, including potentially other species from the natural community. These vector tadpoles may be able to direct probiotics to the appropriate habitat where other tadpoles are present, as opposed to an indiscriminant environmental application of probiotics. Moreover, we showed that bioaugmentation of J. lividum on L. clamitans tadpoles' skin was successful but it decreased through time. While this study offers proof of concept, the future challenge is to determine the best conditions in which an effective bacterial probiotic will be able to colonize and more importantly persist on the skin community. Bd and Bsal pose grave threats of amphibian biodiversity, and with further optimization of techniques, bioaugmentation is likely to be a viable strategy to mitigate chytridiomycosis in amphibians Funding information This project was funded by the National Science Foundation Dimensions in Biodiversity Program: DEB to Reid N. Harris Acknowledgements We thank Molly C. Bletz for technical assistance in the laboratory. We thank the Virginia Department of Game and Inland Fisheries for issuing us a collection permit. The animal care IACUC protocol A03-12 was approved for conducting this experiment. 23

24 References Wake DB, Vredenburg VT Are we in the midst of the sixth mass extinction? A view from the world of amphibians. Proc Natl Acad Sci U S A 105: Blehert DS, Hicks AC, Behr M, Meteyer CU, Berlowski-zier BM, Buckles EL, Coleman JTH, Darling SR, Gargas A, Niver R, Okoniewski JC, Rudd RJ, Stone WB Bat White-Nose Syndrome: An emerging pathogen? Science 323: Fisher MC, Garner TWJ, Walker SF Global emergence of Batrachochytrium dendrobatidis and amphibian chytridiomycosis in space, time, and host. Annu Rev Microbiol 63: Fisher MC, Henk DA, Briggs CJ, Brownstein JS, Madoff LC, McCraw SL, Gurr SJ Emerging fungal threats to animal, plant and ecosystem health. Nature 484: Price SJ, Garner TWJ, Nichols RA, Balloux F, Ayres C, Mora-Cabello de Alba A, Bosch J Collapse of Amphibian Communities Due to an Introduced Ranavirus. Curr Biol Schrope M Sea star wasting. Proc Natl Acad Sci U S A 111: Martel A, Spitzen-van der Sluijs A, Blooi M, Bert W, Ducatelle R, Fisher MC, Woeltjes A, Bosman W, Chiers K, Bossuyt F, Pasmans F Batrachochytrium 24

25 salamandrivorans sp. nov. causes lethal chytridiomycosis in amphibians. Proc Natl Acad Sci U S A 110: Martel A, Blooi M, Fisher MC, Farrer RA, Schmidt BR, Tobler U, Goka K, Lips KR, Muletz C, Zamudio KR, Bosch J, Wombwell E, Garner TWJ, Cunningham AA, Salvidio S, Ducatelle R, Nishikawa K, Nguyen TT, Kolby JE, Van Bocxlaer I, Bossuyt F, Pasmans F Recent introduction of a chytrid fungus endangers Western Palearctic salamanders. Science 346: Skerratt LF, Berger L, Speare R, Cashins SD, McDonald KR, Phillott AD, Hines HB, Kenyon N Spread of chytridiomycosis has caused the rapid global decline and extinction of frogs. Ecohealth 4: Lips KR, Diffendorfer J, Mendelson JR, Sears MW Riding the wave: reconciling the roles of disease and climate change in amphibian declines. PLoS Biol 6:e Crawford AJ, Lips KR, Bermingham E Epidemic disease decimates amphibian abundance, species diversity, and evolutionary history in the highlands of central Panama. Proc Natl Acad Sci U S A 107: Rebollar EA, Hughey MC, Harris RN, Domangue RJ, Medina D, Ibáñez R, Belden LK The lethal fungus Batrachochytrium dendrobatidis is present in lowland tropical forests of Far Eastern Panamá. PLoS One 9:e

26 Belden LK, Hughey MC, Rebollar EA, Umile TP, Loftus SC, Burzynski EA, Minbiole KPC, House LL, Jensen R V., Becker MH, Walke JB, Medina D, Ibáñez R, Harris RN Panamanian frog species host unique skin bacterial communities. Front Microbiol 6: Colombo BM, Scalvenzi T, Benlamara S, Pollet N Microbiota and Mucosal Immunity in Amphibians. Front Immunol 6: Conlon JM Structural diversity and species distribution of host-defense peptides in frog skin secretions. Cell Mol Life Sci 68: Macfoy C, Danosus D, Sandit R, Jones TH, Garraffo HM, Spande TF, Daly JW Alkaloids of anuran skin: Antimicrobial function? Zeitschrift fur Naturforsch - Sect C J Biosci 60: Harris RN, James TY, Lauer A, Simon MA, Patel A Amphibian pathogen Batrachochytrium dendrobatidis is inhibited by the cutaneous bacteria of amphibian species. Ecohealth 3: Brucker RM, Harris RN, Schwantes CR, Gallaher TN, Flaherty DC, Lam BA, Minbiole KPC Amphibian chemical defense: antifungal metabolites of the microsymbiont Janthinobacterium lividum on the salamander Plethodon cinereus. J Chem Ecol 34:

27 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 34: Becker MH, Walke JB, Murrill L, Woodhams DC, Reinert LK, Rollins-Smith LA, Burzynski EA, Umile TP, Minbiole KPC, Belden LK Phylogenetic distribution of symbiotic bacteria from Panamanian amphibians that inhibit growth of the lethal fungal pathogen Batrachochytrium dendrobatidis. Mol Ecol 24: Woodhams DC, Alford RA, Antwis RE, Archer H, Becker MH, Belden LK, Bell SC, Bletz M, Daskin JH, Davis LR, Flechas SV, Lauer A, Gonzalez A, Harris RN, Holden WM, Hughey MC, Ibáñez R, Knight R, Kueneman J, Rabemananjara F, Reinert LK, Rollins-Smith LA, Roman-Rodriguez F, Shaw SD, Walke JB, McKenzie V Antifungal isolates database of amphibian skin-associated bacteria and function against emerging fungal pathogens. Ecology 96: Fites JS, Ramsey JP, Holden WM, Collier SP, Sutherland DM, Reinert LK, Gayek AS, Dermody TS, Aune TM, Oswald-Richter K, Rollins-Smith LA The Invasive Chytrid Fungus of Amphibians Paralyses Lymphocyte Responses. Science 342: McMahon TA, Sears BF, Venesky MD, Bessler SM, Brown JM, Deutsch K, Halstead NT, Lentz G, Tenouri N, Young S, Civitello DJ, Ortega N, Fites JS, Reinert LK, Rollins- 27

28 Smith LA, Raffel TR, Rohr JR Amphibians acquire resistance to live and dead fungus overcoming fungal immunosuppression. Nature 511: Rollins-Smith LA, Conlon JM Antimicrobial peptide defenses against chytridiomycosis, an emerging infectious disease of amphibian populations. Dev Comp Immunol 29: Rollins-Smith L, Reinert L, Burrowes P Coqui frogs persist with the deadly chytrid fungus despite a lack of defensive antimicrobial peptides. Dis Aquat Organ 113: Conlon JM The contribution of skin antimicrobial peptides to the system of innate immunity in anurans. Cell Tissue Res 343: Loudon A, Holland J, Umile T, Burzynski E, Minbiole KPC, Harris RN Interactions between amphibians symbiotic bacteria cause the production of emergent anti-fungal metabolites. Front Microbiol 5: Lam BA, Walton DB, Harris RN Motile zoospores of Batrachochytrium dendrobatidis move away from antifungal metabolites produced by amphibian skin bacteria. Ecohealth 8: Lam BA, Walke JB, Vredenburg VT, Harris RN Proportion of individuals with anti-batrachochytrium dendrobatidis skin bacteria is associated with population persistence in the frog Rana muscosa. Biol Conserv 143:

29 Woodhams DC, Vredenburg VT, Simon M-A, Billheimer D, Shakhtour B, Shyr Y, Briggs CJ, Rollins-Smith LA, Harris RN Symbiotic bacteria contribute to innate immune defenses of the threatened mountain yellow-legged frog, Rana muscosa. Biol Conserv 138: Bletz MC, Loudon AH, Becker MH, Bell SC, Woodhams DC, Minbiole KPC, Harris RN Mitigating amphibian chytridiomycosis with bioaugmentation: Characteristics of effective probiotics and strategies for their selection and use. Ecol Lett 16: 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 152: 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 75: Harris RN, Brucker RM, Walke JB, Becker MH, Schwantes CR, Flaherty DC, Lam BA, Woodhams DC, Briggs CJ, Vredenburg VT, Minbiole KPC Skin microbes on frogs prevent morbidity and mortality caused by a lethal skin fungus. ISME J 3: Becker MH, Harris RN, Minbiole KPC, Schwantes CR, Rollins-Smith LA, Reinert LK, Brucker RM, Domangue RJ, Gratwicke B Towards a better understanding of 29

30 the use of probiotics for preventing chytridiomycosis in Panamanian golden frogs. Ecohealth 8: Woodhams DC, Geiger CC, Reinert LK, Rollins-Smith L a, Lam B, Harris RN, Briggs CJ, Vredenburg VT, Voyles J Treatment of amphibians infected with chytrid fungus: learning from failed trials with itraconazole, antimicrobial peptides, bacteria, and heat therapy. Dis Aquat Organ 98: Küng D, Bigler L, Davis LR, Gratwicke B, Griffith E, Woodhams DC Stability of microbiota facilitated by host immune regulation: Informing probiotic strategies to manage amphibian disease. PLoS One Belden LK, Harris RN Infectious diseases in wildlife: the community ecology context. Front Ecol Environ 5: Naik S, Bouladox N, Wilhelm C, Molloy MJ, Salcedo R, Kastenmuller W, Deming C, Quinones M, Koo L, Conlan S, Spencer S, Hall JA, Dzutsev A, Kong H, Campbell DJ, Trinchieri G, Segre JA, Belkaid Y Compartmentalized Control of Skin Immunity by Resident Commensals. Science 127: Belkaid Y, Segre JA Dialogue between skin microbiota and immunity. Science 346: Brubaker BSW, Monack DM Microbial metabolite triggers antimicrobial defense. Science 348:

31 Edel-Hermann V, Brenot S, Gautheron N, Aimé S, Alabouvette C, Steinberg C Ecological fitness of the biocontrol agent Fusarium oxysporum Fo47 in soil and its impact on the soil microbial communities. FEMS Microbiol Ecol 68: Bajsa N, Morel MA, Braña V, Castro-Sowinski S The effects of agricultural practices on resident soil microbial communities: Focus on biocontrol and biofertilization. Molecular Microbial Ecology of the Rhizosphere, 2: Fierer N, Ferrenberg S, Flores GE, González A, Kueneman J, Legg T, Lynch RC, McDonald D, Mihaljevic JR, O Neill SP, Rhodes ME, Song SJ, Walters WA From Animalcules to an Ecosystem: Application of Ecological Concepts to the Human Microbiome. Annu Rev Ecol Evol Syst 43: 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 23: 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 Wyngaard GA, Chinnappa CC In: Harrison FW, Cowden RR (eds). General Biology and Cytology of Cyclopoid. Developmental biology of freshwater invertebrates. A.R.: Liss, New York, NY, pp

32 R Core Team R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL project.org/ Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA, Turnbaugh PJ, Fierer N, Knight R Global patterns of 16S rrna diversity at a depth of millions of sequences per sample. Proc Natl Acad Sci U S A 108 Suppl : Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, Fierer N, Peña AG, Goodrich K, Gordon JI, Huttley GA, Kelley ST, Knights D, Jeremy E, Ley RE, Lozupone CA, Mcdonald D, Muegge BD, Reeder J, Sevinsky JR, Turnbaugh PJ, Walters WA QIIME allows analysis of high-throughput community sequencing data. Nat Methods 7: Edgar RC Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26: McDonald D, Price MN, Goodrich J, Nawrocki EP, DeSantis TZ, Probst A, Andersen GL, Knight R, Hugenholtz P An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME J 6: Wang Q, Garrity GM, Tiedje JM, Cole JR Naïve Bayesian classifier for rapid assignment of rrna sequences into the new bacterial taxonomy. Appl Environ Microbiol 73: