Bacteria Isolated from Bats Inhibit the Growth of Pseudogymnoascus destructans, the Causative Agent of White-Nose Syndrome

Emerging infectious diseases are a key threat to wildlife. Several fungal skin pathogens have recently emerged and caused widespread mortality in several vertebrate groups, including amphibians, bats, rattlesnakes and humans. White-nose syndrome, caused by the fungal skin pathogen Pseudogymnoascus destructans, threatens several hibernating bat species with extinction and there are few effective treatment strategies. The skin microbiome is increasingly understood to play a large role in determining disease outcome. We isolated bacteria from the skin of four bat species, and co-cultured these isolates with P. destructans to identify bacteria that might inhibit or kill P. destructans. We then conducted two reciprocal challenge experiments in vitro with six bacterial isolates (all in the genus Pseudomonas) to quantify the effect of these bacteria on the growth of P. destructans. All six Pseudomonas isolates significantly inhibited growth of P. destructans compared to non-inhibitory control bacteria, and two isolates performed significantly better than others in suppressing P. destructans growth for at least 35 days. In both challenge experiments, the extent of suppression of P. destructans growth was dependent on the initial concentration of P. destructans and the initial concentration of the bacterial isolate. These results show that bacteria found naturally occurring on bats can inhibit the growth of P. destructans in vitro and should be studied further as a possible probiotic to protect bats from white-nose syndrome. In addition, the presence of these bacteria may influence disease outcomes among individuals, populations, and species.


Introduction
Emerging infectious diseases can have devastating impacts on wildlife, and they currently threaten many species with extinction [1][2][3][4]. With an increase in anthropogenic disturbance and rise in global trade and transportation, the threat posed by wildlife disease is likely to and a probiotic may be an effective way to reduce WNS impacts if it could at least partially inhibit P. destructans growth and delay mortality long enough for bats to survive hibernation.
We cultured bacteria isolated from the skin of four species of hibernating bats from eastern North America to determine whether naturally occurring bacterial species might exist within the skin microbiome of bats that could inhibit growth of P. destructans. We then quantified the anti-fungal efficacy of these bacteria across a range of concentrations in two challenge experiments.

Sampling and isolating cutaneous microbes
We conducted sampling for cutaneous bacteria on hibernating bats at two hibernacula in New York and two in Virginia (exact locations of study sites are not provided to protect sensitive wildlife habitat). We rubbed sterile polyester swabs dipped in sterile water back and forth five times along each bat's forearm and muzzle. Swabs were frozen in 20% glycerol stock for later culturing. We collected swabs from ten individuals from each of four species Eptesicus fuscus, Myotis leibii, M. lucifugus, and M. sodalis.
Epidermal swab sample collection protocols for this study were approved by the University of California, Santa Cruz IACUC (protocol # FrickW1106). Sample collection was permitted by authorized state biologists from the New York Department of Environmental Conservation and Virginia Department of Game and Inland Fisheries. Handling and sampling of endangered species (Myotis sodalis) was conducted under the appropriate state and federal permits.
Each swab was plated on two types of general media, Reasoner's 2A agar (R2A) and sabouraud dextrose agar (SDA), and plates were incubated at 9°C for three weeks. We classified bacteria on each plate by morphotype, using color, growth form, and gram staining techniques. We isolated one colony from each sample by morphotype (to reduce repeat sampling of the same isolate) using a sterile inoculating loop and re-plated each isolate on R2A media and grew them for 2-5 days at 9°C. Each isolate was cryo-banked by sampling from each of these colonies with a sterile inoculating loop, placing the sample in 30% glycerol, and freezing it at -80C for later use.
Pre-screen for bacteria with anti-P. destructans properties We determined whether isolates could inhibit the growth of P. destructans using a challenge protocol adapted from the National Committee for Clinical Laboratory Standards. All culturing was done on SDA. [31]. A suspension containing 1.7 x 10 7 P. destructans conidia/ml (quantified using a hemocytometer) was prepared by flooding a 3 week-old culture of P. destructans grown on SDA with 20 ml of 1X phosphate buffered saline with tween 20 (PBST 20 ). Colonies were submersed for 5 minutes, and then gently rubbed with a sterile inoculating loop to free the conidia. The supernatant was drawn off and placed into a 50ml falcon tube and vortexed to homogenize the suspension. Each 90mm plate was inoculated with 200ul of the P. destructans suspension and allowed to air dry for 10 minutes. We added bacteria on the plate already inoculated with P. destructans from a growing culture using pinpoint inoculation at three equally spaced points on top of the dried P. destructans suspension. Bacteria cultures were grown from frozen stock 24 hours earlier on SDA. Plates were placed into incubators at 9-10°C, which is within the range that bats hibernate [32], and growth was monitored every other day for 14 days and on day 14 any bacteria that produced a zone of inhibition (a visible reduction of P. destructans growth surrounding the bacterial colony) were included in subsequent challenge experiments described below In addition to bacterial isolates from bats, a Pseudomonas fluorescens isolate PfA506 commonly used in biocontrol of agricultural fungal pathogens was included as a positive control [33], and two types of negative controls, 1) a sham inoculation with 20% sterile glycerol stock, and 2) two bacteria isolated from bats in the genera Chryseobacterium and Sphingomonas (both gram-negative rod shaped bacteria) that are not known to produce anti-fungal compounds [31].

Identification of bacterial isolates
We identified bacterial isolates used in the following inhibition assays using PCR amplification and DNA sequencing. DNA for PCR was obtained by suspending a small amount of a bacterial colony in 100 μl of sterile deionized water (SDW) and lysing the cells at 95°C (10 min). Universal bacterial 16S rRNA gene primers (16S_F (5`-ACC GCG ATA ATA CGT CCC GAT CG-3`) and 16S_R(5`-TGC GGA CGT GAA GTG CTA G -3`)) were used to amplify the~1.5 kb 16S rRNA gene fragment [34]. The following was added to each PCR template: 1 μl of crude lysate DNA template, 1.5 μl of each 0.6 μM forward and reverse primer, and 5 μl of Taq 5X MM (NEB) at 1X concentration, which contains 1.5 mM MgCl2, 2 mM dNTPs, and PCR buffer. Reaction volumes were made up to 25 μl with SDW. The reaction conditions involved an initial denaturation at 95°C for 3 minutes, followed by 35 cycles of denaturation at 95°C for 15 sec, primer annealing for 15 seconds at 49°C, and extension for 90 seconds at 42°C. The 16S rRNA gene sequences were compared with known sequences in the EMBL database using MEGA BLAST (BLASTN 2.1.1, [35]) to identify the most similar sequence alignment. Pseudomonas fluorescens isolate PfA506 was used to assure proper alignment of sequences.

Inhibition Assays
Two separate inhibition assays were performed. In the first inhibition assay, we determined the ability of each bacterial isolate to grow on lawns of different starting concentrations of P. destructans, and to produce a zone of inhibition in which P. destructans growth was either delayed, halted adjacent to the bacteria colony. In the second inhibition assay, we measured the growth of P. destructans on a lawn of different starting concentrations of each bacterial isolate.
In the first inhibition assay, the growth of each bacterial isolate was quantified in media inoculated with four P. destructans concentrations estimated using hemocytometry (10 7 , 10 6 , 10 5 , and 10 4 conidia/ml). After the P. destructans suspension was dry and conidia fixed to the plate, we used a pipette to inoculate the plates with 0.1 μl of a 10 8 cfu/ml suspension of a given bacterial isolate at three evenly spaced points on the plate. The bacterial suspension was prepared by suspending whole colonies in 30% glycerol, and using an inoculating loop to suspend the colony. Each treatment was replicated nine times, and cultures were grown at 9°C for 37 days. Zones of inhibition were quantified by measuring the distance from the edge of the bacterial colony to the edge of the visible P. destructans growth every other day [31]. We also microscopically examined the zones of inhibition on the final day of the experiment (day 43) to characterize the effects of bacteria on the growth of P. destructans.
In the second inhibition assay we determined the ability of each bacterial isolate to prevent growth of P. destructans across a series of six bacterial concentrations. Each bacterial isolate was plated from cryobanked glycerol stock onto a Petri dish with SDA media and allowed to incubate for two days at 9°C before being added to a 30% sterile glycerol suspension. The concentration of each isolate was standardized by making serial ten-fold dilutions of the culturing stock and then counting the number of colony forming units (cfu) per ml. The bacteria glycerol suspension was frozen at -20°C while calculating the concentration. Each stock was standardized to the same concentration of 7.5x10 7 cfu/ml using 30% glycerol. We plated 50 μl of each bacterial glycerol dilution on SDA in 60 mm Petri dishes. We used three replicates per treatment for bacterial concentrations 10 6 , 10 2 , and 10 1 cfu/ml, and five replicates for concentrations 10 5 , 10 4 , and 10 3 cfu/ml. For the control plate, we added 50 μl of sterile 30% glycerol solution to the plates and then inoculated with P. destructans using a pinpoint inoculation. The diameter of the P. destructans colony was measured for a total of 42 days. Measurements were made every other day for the first 14 days, and then once every seven days thereafter until the end of the experiment.
We used cell-free supernatant plated on a lawn of P. destructans to determine if anti-fungal compounds were being produced by the bacteria in the initial culture. We used cultures of fresh bacteria grown in isolation, and co-cultured with P. destructans, in lysogeny broth. Cultures were then centrifuged for 30 minutes and the supernatant was drawn off. We then inoculated three plates with 50 μL of the supernatant on a lawn of P. destructans using the methods described above.

Bacteria motility and chemotaxis
Bacteria motility experiments were conducted to assess whether the Pseudomonas isolates from bats preferentially move towards P. destructans. A 0.3% agar SDA media was prepared and a sterile inoculating loop was dipped into a 7.5x10 7 cfu/ml of bacterial suspension and then stabbed~5 cm into the soft agar for all nine bacterial isolates. To determine whether or not the bacteria preferentially moved towards P. destructans, we repeated the same methods described above, but included a small colony of P. destructans that was stabbed into the media on one side of the tube. The tubes were incubated for 1 week at 10°C, and then stabs were visually inspected for signs of whether bacteria moved away from the initial stab or moved towards the P. destructans stab.

Statistical analysis
We fit linear mixed effects models (function glmer in package lme4 [36] in R v. 3.02 [37]) with day as a categorical random effect and bacteria type and concentration as fixed effects to examine the influence of bacteria type and serial dilution on the zone of inhibition (first inhibition assay, S2 Table) and diameter of fungal colony (second inhibition assay, S3 Table). We fit five a priori models including additive and interactive effects (S4 Table) and compared models using Akaike's Information Criterion (AIC) [38].

Pre-screen for bacteria with anti-P. destructans properties
We isolated a total of 133 bacterial morphotypes from the 40 bats we swabbed. Four isolates from E. fuscus (from 3 bats) and two isolates from two individual M. lucifugus inhibited P. destructans growth in standard challenge assays (Table 1). These six bacteria were selected for further testing in the two inhibition assays. All were identified as members of the genus Pseudomonas, with five of the six isolates most closely related to the Pseudomonas fluorescens group and the other isolate (PA6) being most closely related to Pseudomonas abietaniphila (HF952541) (S1 Table).

Inhibition assays
In the first inhibition assay, bacterial colonies initially expanded quickly and then plateaued in size, with growth continuing for longer at lower initial concentrations of P. destructans in the media (S1 Fig., S4 Table). Some bacterial isolates formed much larger colonies than others, with PF1, PF2, and PF4 forming the largest colonies (S1 Fig.). The size of bacterial colonies increased with decreasing initial concentrations of P. destructans (S1 Fig., S4 Table).
Zones of inhibition could not be visualized until P. destructans growth was visible on days 9-11 ( Fig. 1). At this time, zones of inhibition already differed significantly among bacterial isolates and initial concentrations (Fig. 2, S2 Fig., and S5 Table). Two bacterial isolates, PF1 and PF2, generated larger zones of inhibition across most initial concentrations of P. destructans by the end of the experiment (S2 Table). Three isolates (PF1, PF2, and PF7) established two zones, one where growth of P. destructans was suspended immediately upon germination (Fig. 1g), and another outside of this zone where growth was arrested, but only after the mycelial mat had begun to develop (Fig. 1f). Zones of inhibition on the last day of the experiment (day 37) increased with increasing initial concentrations of P. destructans for the Pseudomonas isolates showing the strongest inhibition ( Fig. 2; PF1, PF2, and PF4; all concentration slopes were significantly negative, all p-values <0.03). For the other four Pseudomonas isolates, the zone of inhibition was either variable across concentrations ( Fig. 2; PF3, PF5, PA6) or increased with decreasing initial P. destructans concentration (PF7; concentration coef. 1.72 ± 0.64, p = 0.008). Two isolates, PF1 and PF2, out-performed the reference P. fluorescens strains (PF7; PfA506) at all initial concentrations with at least a two-fold difference in zone of inhibition (Fig. 2). The two control bacteria (Chryseobacterium sp. and Sphingomonas sp.) and the shaminoculated control produced no zones of inhibition (Fig. 2).
In the second inhibition experiment, P. destructans grew optimally in the absence of bacteria, and on media with low initial concentrations of the control bacteria (Fig. 3). By the end of the experiment, the size of P. destructans colonies differed between bacterial isolates and initial concentrations and the effect of bacterial isolate varied among initial concentrations (S3 and S6 Tables). At the highest initial concentration (10 6 cells/ml), all bacteria (including the two control bacteria) formed lawns and all reduced growth of P. destructans. As the starting concentration of the bacteria lawn decreased, fewer isolates significantly reduced the growth of P. destructans. At the three highest initial bacterial concentrations (10 6 -10 4 cfu/ml), only isolates PF1, PF2, and PF5 completely suppressed P. destructans growth for the duration of the experiment (day 42; S6 Table). At the three lowest initial concentrations of the bacteria, where there were relatively few colonies, two Pseudomonas isolates, PF1 and PF2 performed significantly better than other isolates in reducing P. destructans growth and prevented P. destructans from growing for the full duration of the experiment (Fig. 3, S6 Table). In both experiments, isolates PF1 and PF2 produced the maximum reduction of mycelial growth across all concentrations, regardless of the way the isolates and P. destructans were co-cultured.
Cell-free supernatant drawn from liquid bacterial cultures had no effects on the growth of P. destructans. Pseudogymnoascus destructans grew uniformly across all plates, regardless of whether the supernatant added to the plates was from bacteria co-cultured with P. destructans or cultured alone.

Bacteria motility and chemotaxis
All seven Pseudomonas isolates exhibited signs of motility but no chemotaxis towards P. destructans colonies was observed. Using microscopy, two of the Pseudomonas isolates, PF1 and PF2, were observed dispersing along P. destructans hyphae. The two control bacteria showed no signs of motility (Table 1). . We used gram staining techniques to help better visualize conidia (purple) and hyphae (pink) (scale bars, 10 μm). Within the first zone, indicated by the dark ring surrounding the yellow bacteria colony (PF2), the bacteria either arrested or delayed conidia growth, (g) which can be seen by the small hyphael extension from the conidia. Outside of this first zone, the growth of P. destructans was much more extensive (f), producing a mycelial network before its growth was arrested.  Table). CHR and SPH are isolates in the genus Chryseobacterium and Sphingomonas that are not known to produce antifungal compounds. The control is an inoculation of 30% glycerol stock. PF1-5,7 and PA6 are all isolates in the genus Pseudomonas.
doi:10.1371/journal.pone.0121329.g002 Second inhibition assay measuring the diameter of P. destructans colonies grown on a lawn of bacteria. Pseudogymnoascus destructans was plated with nine bacterial isolates at six different concentrations (highest to lowest, left to right). Lines denoted by the same letter did not differ significantly on the last day of the experiment. CHR and SPH are isolates in the genus Chryseobacterium and Sphingomonas that are not known to produce antifungal compounds. The Control is a sham inoculation of 30% glycerol stock. PF1-7 and PA6 are all isolates in the genus Pseudomonas.

Discussion
As the threat of emerging infectious disease grows with increased global travel and trade [5], new ways of managing wildlife disease must be considered [8]. Traditionally, fungal pathogens have been managed using chemical fungicides [39], but toxicity effects on non-target organisms, and application challenges makes it difficult for broad-scale use on wildlife fungal pathogens [29]. The results from our two sets of experiments demonstrate that in vivo, bacteria cultured from bats can inhibit the growth of P. destructans. Our results suggest that augmentation prior to P. destructans exposure might reduce colonization, whereas bacterial augmentation after exposure could displace P. destructans. Our results also suggest that a key challenge for successful treatment is applying bacteria such that they will persist on bat skin at high enough concentrations to limit P. destructans growth below levels that cause lethal disease. The bacteria we isolated from bats, Pseudomonas spp., is ubiquitous in the environment and is well known to have anti-fungal properties [40]. The group of bacteria that these isolates were most closely related to, Pseudomonas fluorescens, has previously been detected on several mammals (including bats), as well as amphibians, fish, and plants [41][42][43][44]. Members of the P. fluorescens group are known to produce a suite of antifungal compounds that can inhibit many plant fungal pathogens [45] as well as the amphibian fungal pathogens, Batrachochytrium dendrobatidis [46]. Some strains in the P. fluorescens group are also capable of producing mycolysing enzymes that can colonize the mycelia and conidia of fungi rendering them no longer viable [47]. All of our Pseudomonas spp. isolates were motile, which might allow them to use the mycelial networks of fungal colonies to aid in dispersal and colonization [48]. All of these attributes make P. fluorescens ideal as a proposed candidate to be tested as a biological control agent for reducing infection intensity and increasing survival of bats exposed to P. destructans.
Whether these antifungal bacteria that naturally occur on bat skin could partially explain differences in mortality from WNS among populations and species is currently unknown. The isolates with strongest inhibitory properties were cultured from E. fuscus, which has lower mortality from WNS compared to other species [2]. However, we also isolated two strains of P. fluorescens (PF3 and PF4) that showed moderate P. destructans inhibition from M. lucifugus, a species that has suffered severe mortality from WNS [2,20]. Further research is needed to determine the relative abundance, distribution, and inhibitory ability of P. fluorescens on wild bats and whether presence and abundance of P. fluorescens influences disease severity.
The next steps in developing a probiotic for WNS should include testing, in vivo, one or more of the P. fluorescens strains that we isolated against P. destructans using a bat species that suffers high disease mortality from WNS, such as M. lucifugus, M. septentrionalis, or Perimyotis subflavus [2]. Studies with live hibernating bats will determine whether interactions observed in vitro have functional significance in disease outcomes for bat species currently threatened by WNS.
Supporting Information S1 Fig. Bacterial colony size during first inhibition assay. Colony size of nine bacterial isolates grown on plates inoculated with four different concentrations of Pseudogymnoascus destructans with fungal concentrations decreasing from left to right. CHR and SPH are isolates in the genus Chryseobacterium and Sphingomonas that are not known to produce antifungal compounds. The Control is a sham inoculation of 30% glycerol stock. PF1-7 and PA6 are bacterial isolates in the genus Pseudomonas. (TIF) S1 Table. BLAST results of 16S rRNA sequence from the National Center of Biological Information database for the six bacterial isolated in this study from bats (Myotis lucifugus and Eptesicus fuscus). (DOCX) S2 Table. AIC values for the first inhibition assay measuring the zone of inhibition produced by bacteria on a lawn of P. destructans. (DOCX) S3 Table. AIC values for the second inhibition assay measuring how different bacterial isolates influenced the diameter of a P. destructans colony. (DOCX) S4 Table. AIC values for the first inhibition assay measuring the diameter of bacteria colonies on different concentrations of P. destructans. (DOCX) S5 Table. Coefficients for linear models of the influence of nine bacterial treatments and a control on the radius of the zones of inhibition of P. destructans produced by bacteria at four different initial concentrations of P. destructans on day 37 for the data shown in Fig. 2. Table. Coefficients for linear models of the influence of nine bacterial isolates on the diameter of P. destructans colonies for each bacterial concentration on day 43 for the data shown in Fig. 3. (DOCX)