Synergistic Inhibition of the Lethal Fungal Pathogen Batrachochytrium dendrobatidis: The Combined Effect of Symbiotic Bacterial Metabolites and Antimicrobial Peptides of the Frog Rana muscosa
- First Online:
- Cite this article as:
- Myers, J.M., Ramsey, J.P., Blackman, A.L. et al. J Chem Ecol (2012) 38: 958. doi:10.1007/s10886-012-0170-2
- 730 Views
A powerful mechanism for protection against disease in animals is synergy between metabolites present in the natural microbiota of the host and antimicrobial peptides (AMPs) produced by the host. We studied this method of protection in amphibians in regard to the lethal disease chytridiomycosis, which is caused by Batrachochytrium dendrobatidis (Bd). In this study, we show that the AMPs of Rana muscosa, as well as the metabolite 2,4-diacetylphloroglucinol (2,4-DAPG) from Pseudomonas fluorescens, a bacterial species normally found on the skin of R. muscosa, were inhibitory to the growth of Bd in vitro. When both AMPs and 2,4-DAPG were used in growth inhibition assays, they worked synergistically to inhibit the growth of Bd. This synergy resulted in reduced minimum concentrations necessary for inhibition by either 2,4-DAPG or AMPs. This inhibitory concentration of AMPs did not inhibit the growth of a P. fluorescens strain that produced 2,4-DAPG in vitro, although its growth was inhibited at higher peptide concentrations. These data suggest that the AMPs secreted onto frog skin and the metabolites secreted by the resident beneficial bacteria may work synergistically to enhance protection against Bd infection on amphibian skin. These results may aid conservation efforts to augment amphibian skins’ resistance to chytridiomycosis by introducing anti-Bd bacterial species that work synergistically with amphibian AMPs.
KeywordsBatrachochytrium dendrobatidisAntimicrobial peptidesMetabolitesPseudomonas fluorescensSynergy2,4-diacetylphloroglucinolChytridiomycosisAmphibian conservationProbiotic
Amphibian species around the world are undergoing dramatic population declines and local extinctions, and it is estimated that 40 % of amphibian species are vulnerable to extinctions (Simon et al., 2004; Hoffman et al., 2010). Although habitat loss and fragmentation is one critical cause of these declines (Collins and Storfer, 2003), population losses in pristine habitats in Australia, Central and South America, and the western United States have been linked to the emergence of a lethal fungal skin pathogen, Batrachochytrium dendrobatidis (Bd) (Berger et al., 1998; Weldon et al., 2004; Lips et al., 2006). Batrachochytrium dendrobatidis is spread by its waterborne zoospores and colonizes the keratinized layer of juvenile and adult amphibian skins, where it matures into a zoosporangium that is capable of producing new zoospores that can either infect a second host or re-infect the same host (Berger et al., 2005). The resulting disease, chytridiomycosis, is fatal to many, but not all, amphibian species (Berger et al., 1998; Woodhams et al., 2005, 2006, 2007a; Lips et al., 2006). The factors that determine resistance to chytridiomycosis in certain amphibian species are not fully understood. Chytridiomycosis is currently regarded as the greatest disease threat to biodiversity (Kilpatrick et al., 2010), and understanding what makes certain species resistant to lethal Bd infections is crucial for amphibian conservation efforts.
The skin of amphibians is protected by both innate and adaptive immune responses. Innate responses in amphibians include phagocytic cells, complement proteins, lysozyme, and the production and secretion of antimicrobial peptides (AMPs) via granular glands onto the surface of the skin to fight off invading pathogens such as Bd (Daly, 1995; Simmaco et al., 1998; Zasloff, 2002; Conlon et al., 2004). Previous work has shown that the AMP secretions, as well as purified individual peptides, from numerous amphibian species can inhibit the growth of Bd in vitro (Rollins-Smith et al., 2002, 2006; Woodhams et al., 2005, 2006, 2007a; Rollins-Smith and Conlon, 2005; Conlon et al., 2007; Rollins-Smith, 2009). The production of peptide secretions that are effective inhibitors of Bd growth is correlated with both survival of amphibian species in the wild (Woodhams et al., 2005, 2006) and with the overall resistance of certain species when experimentally infected with Bd (Woodhams et al., 2007a). The adaptive immune system also appears to play a role in protection against chytridiomycosis, although the exact mode of protection is not understood (Ramsey et al., 2010). A deficiency of either of these systems could increase the overall susceptibility of a species to chytridiomycosis, leading to fatal infections.
Another defense against Bd infection is the bacterial microbiota resident on amphibian skin. Inhibition of fungal growth has been associated with over 50 bacterial strains in over 20 genera found on the skins of amphibian embryos and adults (Harris et al., 2006; Lauer et al., 2007, 2008; Woodhams et al., 2007b, c). The mutualistic bacterial species, Janthinobacterium lividum, isolated from the yellow-legged mountain frog (Rana muscosa), the four-toed salamander (Hemidactylium scutatum), and the red-backed salamander (Plethodon cinereus), produces anti-Bd metabolites at concentrations lethal to the fungus (Brucker et al., 2008a). Three anti-Bd metabolites secreted by J. lividum and another skin bacterial species, Lysobacter gummosus have been identified: 2,4-diacetylphloroglucinol (2,4-DAPG), indole-3-carboxaldehyde (I3C), and violacein (Brucker et al., 2008a, b). 2,4-DAPG inhibited Bd growth in vitro with an IC50 of 8.7 μM and a minimum inhibitory concentration (MIC) of 136 μM (Brucker et al., 2008b). I3C and violacein also inhibited Bd significantly with MICs of 68.9 μM and 1.8 μM, respectively (Brucker et al., 2008a).
Rana muscosa, a frog species native to the Sierra Nevada Mountains of California, has AMPs that effectively inhibit Bd growth in vitro (Rollins-Smith et al., 2006). Although populations of R. muscosa vary in their AMP compositions, many of the peptides are shared among populations (Woodhams et al., 2007c). Populations that have no experience with Bd usually experience dramatic population declines once Bd arrives. However, there are populations that coexist with Bd. There was a positive relationship between the proportion of individuals with anti-Bd bacteria on the skins of two species of Rana found in the Sierra Nevada and their resistance to Bd (Woodhams et al., 2007c). A second survey conducted after the initial R. muscosa decline showed that the Bd-naive frogs that populated lakes had higher proportion of anti-Bd bacteria on the skin and were able to survive and persist after the emergence of Bd (Lam et al., 2010). It is possible that both AMP defenses and the microbiota present on the skin play a role in dictating whether a population of an amphibian species is resistant to Bd. However, other stressors, such as pesticide or environmental contamination, increased UV radiation, or increases in ambient temperature, could lead to alterations of either AMP defenses or the skin microbiota and affect resistance to Bd.
The potential for synergy between bacterially-produced metabolites and the AMPs secreted onto amphibian skin has not been explored in amphibians, but synergies between bacterial metabolites and AMPs have been found in other systems (Cassone and Otvos, 2010). We hypothesized that metabolites and AMPs present on the skin of amphibians will synergize to more effectively inhibit the growth of Bd as compared to either compound alone. We tested this hypothesis using AMPs from R. muscosa, a species undergoing documented population declines linked to Bd (Briggs et al., 2005; Woodhams et al., 2007c), as well as 2,4-DAPG, one of the metabolites secreted by bacterial species resident on the skin of R. muscosa (Brucker et al., 2008a,b, Lam et al., 2010).
Methods and Materials
Maintenance of Batrachochytrium dendrobatidis Cultures
Stocks of Bd isolate 423, obtained from Dr. Joyce Longcore (University of Maine), were serially transferred monthly with antibiotics to inhibit bacterial contamination. Stocks contained 50 ml 1 % tryptone broth, 1 ml previous stock, 50 μl penicillin (100 mg/ml), and 50 μl streptomycin sulfate (200 mg/ml). All monthly stocks were incubated at 23 °C for 7–10 day without illumination before storage at 4 °C for approximately 1 month. For experiments, 1 ml aliquots of monthly stocks were transferred to 50 ml 1 % tryptone broth and incubated at 23 °C. Experimental stocks were serially transferred weekly by adding 1 ml stock to fresh broth.
Collection of Batrachochytrium dendrobatidis Zoospores
Bd plates were prepared by adding 1 ml weekly experimental stock onto 1 % tryptone agar plates. Bd zoospores were harvested after a 1 week incubation from plates, and separated from mature sporangia (Harris et al., 2009a). Briefly, plates were flooded with 4 ml tryptone broth for 20 min. The liquid was removed via pipette and passed though a sterile coffee filter to remove mature sporangia. An additional 1 ml tryptone broth was added to wash the plate, removed via pipette, and filtered. Zoospores were counted using a hemocytometer and diluted to 1 × 106 zoospores/ml using tryptone broth.
Amphibian Antimicrobial Peptides
Antimicrobial peptides were obtained from Dr. Louise Rollins-Smith (Vanderbilt University). Natural mixtures of antimicrobial peptides were originally collected from two R. muscosa frogs in 2005 by norepinephrine injection and collection of gland secretions. Although the exact peptide composition of these samples is unknown, all R. muscosa individuals secrete similar peptide profiles and include peptides such as temporin-1M and ranatuerin-2Ma and -2Mb (Rollins-Smith et al., 2006). These natural mixtures were enriched by removing non-peptide components then stored at −20 °C as in Ramsey et al. 2010. After purification, the peptide concentrations were determined by MicroBCA assay (Pierce, Rockford, IL, USA) using bradykinin (Sigma Chemicals, St. Louis, MO, USA) as a standard.
Organic Extraction of Liquid Broth Cultures and Preparation for Analysis
A P. fluorescens strain isolated from a Rana muscosa individual was co-cultured with Bd, then re-isolated from co-culture and incubated in a 1 % solution of tryptone broth at room temperature for 4 days without illumination. The bacterial species was identified by sequencing a portion of its 16 S rRNA gene and comparing the sequence to the GenBank database (Lauer et al. 2007). A10 ml sample of this culture was extracted in triplicate with an equal volume of ethyl acetate (EtOAc). The combined organic layers were dried over Na2SO4, filtered, and evaporated in vacuo. This uncharacterized sample (5.1 mg) was dissolved in high-performance liquid chromatography (HPLC)-grade methanol (1 ml) for analysis by reverse phase (RP)-HPLC (Agilent Technologies, 1200 series, Wilmington, DE, USA). The chromatogram of the crude extract was compared to that of a 2,4-DAPG standard (10 ppm; Sigma Chemicals, St. Louis, MO, USA) with diode array detection at 270 nm, the λmax for 2,4-DAPG (Brucker et al., 2008b). The samples were injected (100 μl) into the HPLC equipped with a C18 reverse phase column (5 μm; 4.6 × 150 mm; Agilent Technologies, Wilmington, DE, USA) and eluted at 1 ml/min. The solvent program was isocratic for 2 min in 10 % acetonitrile/water (v/v) containing 0.1 % acetic acid followed by a linear gradient to 100 % acetonitrile containing 0.1 % acetic acid over 18 min, and a final 3 min isocratic period before returning to the initial conditions.
LC/MS Analysis of Extract
A sample of 10 μl of the resuspended P. fluorescens culture extract was diluted 100-fold in HPLC grade methanol before analysis by LC/MS (Agilent Technologies; 6530 Accurate-Mass Q-TOF). The diluted culture extract or 10 ppb DAPG standard (5 μl) was injected into the LC equipped with an Eclipse Plus C18 column (2.1 × 100 mm) at a temperature of 50 °C, and eluted at 0.3 ml/min. The solvent program was isocratic for 0.25 min in 5 % acetonitrile/water (v/v) containing 0.1 % formic acid, followed by a 3.75 min linear gradient to 95 % acetonitrile containing 0.1 % formic acid and a final 0.5 min isocratic period before returning to the initial conditions. The samples were detected with a quadrupole, time-of-flight [Q-TOF] mass spectrometer using an ESI ion source run in negative mode.
Batrachochytrium dendrobatidis Growth Inhibition Assay
Before each assay, Bd zoospores were harvested as described above. Zoospores were plated (5 × 104 zoospores/50 μl, five replicates) in tryptone broth in 96-well flat-bottom microtiter plates with serial dilutions (50 μl total) of: 1) AMP mixture diluted in sterile water (1.56 μg/ml to 200 μg/ml); 2) purified 2,4-DAPG (Sigma, St. Louis, MO) diluted in sterile water (1.1 μM to 136 μM); 3) a combination of both AMPs and 2,4-DAPG diluted in sterile water or with sterile water as the positive control. Negative control wells contained zoospores that were heat-killed at 60 °C for 1 hr. Plates were incubated for 1 week at 23 °C before assessing growth by measuring optical density at 490 nm (O.D.490) with a microplate reader. Percent growth was calculated as: (1 - [(X-Y)/X])*100, where X is the positive control optical density value, and Y is an experimental well optical density value. This assay was conducted twice, with ten total replicates per sample.
Pseudomonas fluorescens Growth and Challenge Assay
We used P. fluorescens as it entered log phase for challenge assays with AMPs to maximize bacterial growth. The growth curve of P. fluorescens was determined using a standard colony forming unit (CFU) plating technique, with CFUs plotted against optical density at each hour. Logarithmic phase occurred after approximately 5 hr.
The challenge assay was conducted with 50 μl P. fluorescens (1 × 103 CFU/ml, five replicates) plated in tryptone broth in 96-well flat-bottom microtiter plates with or without serial dilutions (50 μl total) of antimicrobial peptides diluted in sterile water or sterile water as the positive control. Negative control wells received 5 μl streptomycin sulfate (200 mg/ml) plus 45 μl sterile water. Plates were incubated for 18 hr at 26 °C before measuring O.D.490. This assay was conducted twice with ten total replicates per sample.
Data Analysis and Statistics
We defined MIC as the lowest concentration that yielded percent growth not significantly different from the negative control (Rollins-Smith et al., 2006). We used a two-tailed t- test with Bonferroni correction for multiple comparisons and an alpha level of 0.05. We determined the separate effects of AMP concentration and 2,4-DAPG concentration and interaction effects using a two-way ANOVA. All analyses were performed using the SAS statistical software.
Confirmation of the Production of 2,4-DAPG by P. fluorescens Strain BOH3
Effect of AMPs on B. dendrobatidis Growth
Effect of 2,4-DAPG on B. dendrobatidis Growth
Purified 2,4-DAPG, produced and released by P. fluorescens and other bacteria, also strongly inhibited the growth (P < 0.01) of Bd in vitro (Fig. 2b). All concentrations ≥ 4.4 μM were inhibitory, and the MIC was 17 μM. These data show that 2,4-DAPG is an effective inhibitor of Bd growth.
Combined Effects of AMPs and 2,4-DAPG on B. dendrobatidis Growth
Effects of AMPs on the Growth of P. fluorescens
We provide evidence for a synergistic interaction between AMPs secreted by the frog R. muscosa and the bacterially-produced metabolite 2,4-DAPG released by P. fluorescens, a bacterial species present in the natural skin microbiota of R. muscosa (Woodhams et al., 2007c, Lam et al., 2010). Each treatment inhibited B. dendrobatidis when tested alone. When combined together, the concentration of both AMPs and 2,4-DAPG needed to inhibit Bd in vitro was reduced four-fold. These finding support our hypothesis that a mutualistic relationship exists between R. muscosa and P. fluorescens. This study evaluated the potential synergy between one metabolite and the AMPs of one amphibian species. Future studies need to be conducted to establish whether synergy exists between other metabolic products or groups of metabolites, and AMPs from other species or individual AMPs.
The concentration of AMPs that occur naturally on the skin of R. muscosa are not known. The South African clawed frog, Xenopus laevis, secretes concentrations of AMPs both in a resting state (3.3 mg/ml in mucus) and after stress (20 mg/ml in mucus) that are inhibitory to Bd (≥ 125 μg/ml) (Ramsey et al., 2010). The northern leopard frog (Rana pipiens), also secretes inhibitory levels of AMPs while resting and after a mock predator attack, and the secretions inhibited Bd growth in vitro (Pask et al., 2012). Experiments to determine whether populations of R. muscosa can produce sufficient AMPs to inhibit Bd or beneficial bacteria in the wild should be conducted. Although previous studies have estimated the concentrations of the bacterial metabolites violacein and indole-3-carboxaldehyde (both produced by J. lividum) on the skin of the amphibian Plethodon cinereus (Brucker et al. 2008b), there are no published results that establish the skin concentrations of 2,4-DAPG in amphibians.
In order to more fully explore the potential mutualism between R. muscosa and P. fluorescens, we evaluated whether P. fluorescens could survive and grow in the presence of AMPs at concentrations that are necessary for synergism. At the lower concentrations that led to synergy with 2,4-DAPG, the AMPs did not inhibit P. fluorescens growth. These data suggest that P. fluorescens may survive on the skin of R. muscosa even when low levels of AMPs are being secreted to defend against pathogens. In this case, a potential mutualism could minimize the cost to the amphibian of AMP secretion, while the bacteria benefit by surviving in an environment with limited competition from other microbes. However, the presence of Bd or the addition of bacteria to the skin may lead to upregulation of AMPs to levels that could disturb the skin microbiota and potentially impair this synergy, although there is no report in the literature of AMP upregulation after microbe introduction. The results of this study suggest that a mutualistic relationship has developed between R. muscosa and P. fluorescens. Future studies are necessary to further test this hypothesis, including determining the concentration of AMPs naturally produced by R. muscosa in the presence or absence of P. fluorescens, as well as evaluating the concentration of 2,4-DAPG released by naturally-occurring P. fluorescens on the skin of R. muscosa. The combined results of this and future studies have implications for our understanding of symbiosis, amphibian immunology, and the disease dynamics of a lethal amphibian pathogen. For example, the effectiveness of AMPs as a mechanism of innate immunity may depend on synergies with bacterial metabolites. Since microbial community structure and, therefore, metabolites are likely to be variable, the ability of skin pathogens to evolve resistance may be difficult.
These results also provide preliminary support for use of a bioaugmentation or probiotic approach to treat susceptible amphibians in the field. We suggest that a critical criterion of an effective probiotic is that the bacterium works synergistically with the host species AMPs. Probiotic treatments using anti-Bd bacterial species on the skin of amphibians increased the overall protection of those amphibians to Bd infections (Harris et al. 2009a, b), highlighting the importance of this approach. An ongoing field study in the Sierra Nevada of California also showed that probiotic treatment increased resistance to Bd in Bd-endemic ponds (Vredenburg et al., 2011 and personal communication). Given the success of probiotic trials in amphibian conservation, we urge continued research in this area in order to optimize efficacy.
Bioaugmentation approaches to combating chytridiomycosis in the wild present an ecological dilemma. It is possible that the probiotic could spread to nontarget species in the wild or alter ecosystem dynamics when introduced. One way to minimize nontarget effects is to use a probiotic species found in the local environment. Previous experiments have determined that the addition of bacteria to control diseases in agricultural contexts are effective and have no adverse effects on nontarget species (Berg et al., 2007; Scherwinski et al., 2008). The decision on whether bioaugmentation would be worthwhile will benefit from using an ecological ethics framework (Minteer and Collins, 2008), which balances the risks between amphibians, ecosystems, and the general public. As humans alter their environments, antifungal microbes may be eliminated from waters and soils. As a result, amphibians may be losing their protective skin microbiota, and the restoring of this microbiota may be critical to an amphibian species’ survival in the future. The addition of bacterial species that produce beneficial metabolites to the skin of amphibians that are threatened by Bd would allow for synergy to occur between metabolites and amphibian AMPs, making the skin a more hostile environment for Bd and thus preventing fatal chytridiomycosis.
This work was supported by a National Science Foundation EAGER grant (DEB-1049699) and an NSF MRI grant (CHE-0958973). The authors wish to thank Dr. Kyle Seifert (James Madison University) for his expertise and for providing laboratory space.