Correspondence Sabine Melzer, Department of Zoology, University of Otago, PO Box 56, Dunedin 9054, New Zealand. Email: firstname.lastname@example.org
Over the past few decades, amphibian populations have undergone drastic declines on a global scale. Declines in many anuran populations have been linked to the emergent skin-invasive amphibian chytrid fungus Batrachochytrium dendrobatidis (Bd). Antimicrobial peptides in the skin are thought to act as important components of the innate immune system that may protect some species from infectious diseases. The four archaic species of Leiopelma in New Zealand are of great conservation concern and a severe population crash of Leiopelma archeyi between 1996 and 2001 has been tentatively linked with the outbreak of Bd. Here, we investigated the in vitro activity of skin secretions of six frog species in New Zealand against Bd zoospore growth. The activity of skin secretions produced by frogs in the wild varied significantly between species, with those of Le. archeyi being the most active. The skin secretions of native Leiopelmatid species showed greater Bd zoospore inhibition (31.0–71.9%) than the naturalized Litoria species (17.4–18.2%). Leiopelma archeyi has the most active peptides, even though it is the only native species with known susceptibility to Bd infections.
While the isolated peptides Aurein 2.1 (present in Li. aurea and Li. raniformis) and Caerin 1.1 (present in Li. ewingii) have been tested for their anti-chytrid activity (Woodhams et al., 2006a), the activity of the natural mixtures of skin peptides of these species remain to be quantified. A minimum of 50 μM of Caerin 1.1 is needed to completely inhibit the growth of both zoosporangia and zoospores of Bd isolate JEL 197 (Woodhams et al., 2006a). Litoria aurea and Li. raniformis produce between 16 and 17 aurein peptides, 10 of which are found in both species. Aurein 2.1 has a minimum inhibitory concentration of 200 μM for both mature cells and zoospores of Bd isolate JEL 197 (Woodhams et al., 2006a).
In this study, we establish the relative effectiveness of natural skin peptide mixtures of New Zealand frogs in inhibiting Bd zoospore growth.
Skin secretion collection
Skin secretions containing peptides were sampled in January and February 2007 from adult/subadult individuals of six species of anurans from the North Island and Maud Island of New Zealand under Department of Conservation permits (WK-20068-RES, NM-19892-RES) and in accordance with procedures approved by the Otago University Animal Ethics Committee (AEC no. 63/06). The species sampled included three naturalized Litoria species from Australia, Li. aurea (n=16), Li. ewingii (n=7) and Li. raniformis (n=9) as well as three native Leiopelma species, Le. archeyi (n=17), Leiopelma hochstetteri (n=19) and Leiopelma pakeka (n=9). All frogs were dry swabbed (Medical Wire & Equipment Co. MW 100-100, Corsham, Wiltshire, UK) as described by Hyatt et al. (2007). An in-house SYBR green qPCR assay (S. Herbert et al., unpubl. data), which has been validated against the Taqman qPCR assay (Bishop et al., 2009), was used to diagnose Bd infection. Skin secretions were collected by the non-invasive method of mild transdermal electric stimulation, which has been widely used since 1992 and has no adverse effects on frogs (Tyler, Stone & Bowie, 1992; Apponyi et al., 2004; Smith et al., 2004). Each frog was captured by hand, held individually in new plastic bags for no longer than 20 min and weighed to the nearest 0.1 g before collecting skin secretions. All frogs were handled using fresh latex gloves for each individual and strict hygiene protocols were adhered to at all times to prevent the potential spread of Bd. To collect skin secretions, individual frogs were held by their back legs, their skin moistened with ultra-purified, deionized water (hereafter referred to as MQW; Millipore, Molsheim, France) and a bipolar platinum electrode was gently applied to the dorsal glands. Stimulus strength was adapted to the size of the frogs, ranging from 1 to 1.4 V (AC), and was administered three times for 10 s each. Skin secretions were washed into a clean polypropylene collection beaker with a total of 12 mL MQW per frog and acidified with 0.1% glacial acetic acid to inactivate endogenous peptidases, if present (Resnick et al., 1991; Steinborner et al., 1997b). The acidified collection solution was kept at 4 °C and transported to the lab. Each sample was passed over C-18 Sep-Pak cartridges (Waters Corporation, Milford, MA, USA) and eluted with 70% acetonitrile, 29.9% water and 0.1% trifluoroacetic acid (v/v/v). The volume of the elute was recorded and 1 mL removed for subsequent calculation of peptide concentration. A micro BCA assay (Pierce, Rockford, IL, USA) was performed to determine the total amount of peptides present in each sample using bradykinin (RPPGFSPFR; Sigma Chemical Co., St Louis, MO, USA) to establish a standard curve (Rollins-Smith et al., 2002c). The remaining sample was then centrifuged to dryness, reconstituted with MQW to a standard concentration of 100 μg mL−1 and sterilized using 0.2-μm syringe filters (Corning Inc., Corning, NY, USA) to be used for in vitro growth inhibition assays.
Culture and maintenance of Bd
The Bd type isolate JEL 197, which was isolated from Dendrobates azureus from the National Zoological Park, Washington, DC (Longcore, Pessier & Nichols, 1999) was obtained from a cryo-archive (Boyle et al., 2003) held by R. Poulter (University of Otago). The fungus was maintained using standard methods (Rollins-Smith et al., 2002b,c) on T-agar plates (1% tryptone), sub-cultured every 7 days by streaking and incubated at 23 °C.
Bd growth inhibition assays
Zoospores were harvested by flooding plates with 3 mL of sterile T-broth for 20 min and gently tipping the plate to collect the liquid containing the motile zoospores, without displacing the mature zoosporangia. The total number of viable cells per millilitre was estimated by counting dilutions of zoospores in Lugol's solution (5% iodine, 10% potassium iodide, 85% MQW) using a haemocytometer.
We tested the ability of natural skin secretion mixtures to inhibit Bd zoospore growth as described previously (Rollins-Smith et al., 2002b,c; Woodhams et al., 2006b). For growth inhibition of zoospores, 5 × 104 zoospores in 50 μL of T-broth were plated in replicates of five in a 96-well microtitre plate (Costar 3596, Corning Inc., Corning, NY, USA). Fifty microlitres of each skin secretion sample was added at a standard concentration of 100 μg mL−1 at a pH of 6.5–7.0 to give a final concentration of 50 μg mL−1. Positive control wells (50 μL of live zoospores/50 μL of sterile MQW), negative control wells (50 μL of heat-killed zoospores/50 μL of sterile MQW) and blank wells (50 μL of T-broth per 50 μL of sterile MQW) were included in each plate in replicates of five. Zoospores were heat-killed at 65 °C for 10 min in a water bath (Johnson et al., 2003). Plates were covered, wrapped in thin plastic film (GLAD® wrap, Auckland, New Zealand) to limit moisture loss, and incubated at 23 °C. Optical density was measured daily for a week at 492 nm using a Fluostar Omega spectrophotometer (Alphatech systems, Auckland, New Zealand). The relative effectiveness of skin secretions was defined as the quantity of peptides produced and their ability to inhibit Bd zoospore growth. This was calculated by multiplying the per cent growth inhibition at 50 μg mL−1 by the peptide concentration in μg g−1 bw produced by each frog (Woodhams et al., 2006b; Tennessen et al., 2009).
To explore if there were differences in Bd growth inhibition, the amount of peptides produced (in μg per gram body weight) and the relative effectiveness of peptides across the six different species, Kruskal–Wallis tests and subsequent post hoc tests (Wilcoxon signed rank tests using a Bonferoni-adjusted α value of P=0.01) were used. All statistics were performed using spss version 17 (SPSS Inc., Chicago, IL, USA, 1999).
Peptide yield and activity against Bd
There was a significant difference between the six New Zealand species in the ability of skin secretions to inhibit Bd zoospore growth at 50 μg mL−1 (χ52=33.45; P<0.001; Table 1) and in the concentration of skin secretions produced per gram body weight (χ52=33.47; P<0.001; Table 1). Leiopelma archeyi secreted significantly larger quantities of skin peptides than Le. hochstetteri (U=38, z=−3.9, P<0.001), Le. pakeka (U=19, z=−3.1, P=0.002) and Li. aurea (U=30, z=−43.8, P<0.001). There was no significant difference in the quantities of peptides secreted between either Le. archeyi and Li. ewingii (U=42, z=−1.1, P=2.88) or Li. raniformis (U=61, z=−8.35, P=4.26). Comparisons of Bd inhibition showed that Le. archeyi produced skin secretions that were significantly more active against zoospores (71.9% growth inhibition) than Le. hochstetteri (U=45.5, z=−3.7, P<0.001), Li. aurea (U=18.5, z=−4.2, P<0.001), Li. ewingii (U=9.5, z=−3.2, P=0.001) and Li. raniformis (U=11, z=−3.5, P<0.001). No significant difference between Le. archeyi and Le. pakeka (U=39, z=−2.02, P>0.01) was detected. Despite producing the largest quantities of peptides, Li. ewingii possess the least active secretions of the New Zealand species (Table 1). Overall, the skin secretions of native Leiopelma species (31–71.9% growth inhibition) were more effective at inhibiting zoospore growth than the naturalized Litoria species (17.4–18.2% growth inhibition).
Table 1. Summary of skin secretion defences of New Zealand species
Peptides recovered (μg g−1 body weight)
% growth inhibition of Batrachochytrium dendrobatidis at 50 μg mL−1
Individuals were induced to secrete by mild electric stimulation.
Relative effectiveness of peptide defences against Bd
There was a statistically significant difference in relative peptide defences (skin peptide quantity produced multiplied by their % growth inhibition) among the species tested (χ52=37.39; P<0.001; Fig. 1). Leiopelma archeyi and Li. ewingii have strong peptide defences based on the quantity and quality of their skin secretions (Fig. 1). While all three Litoria species are generally susceptible to chytridiomycosis in New Zealand, Australia and Tasmania (Berger, 2001; Waldman et al., 2001; Carver, 2004; Sadic & Waldman, 2004; Obendorf & Dalton, 2006), individuals of Le. archeyi in the wild are susceptible to Bd infection but able to eliminate the fungus when held in captivity (Bishop et al., 2009). It remains to be determined if Le. pakeka and Le. hochstetteri are susceptible to Bd infection or develop symptoms of chytridiomycosis. The analysis of the skin swabs showed Bd zoospores were present on one Li. aurea (65.2 zoospore equivalents) and the Bd presence of one Li. raniformis was equivocal.
Activity of skin secretions against Bd growth
The present study has provided evidence that the skin secretions of both native and naturalized anuran species in New Zealand can inhibit Bd zoospore growth in vitro. Litoria ewingii and Le. archeyi produce the largest amount of peptides of all six New Zealand species. When comparing the efficacy of skin secretions (at a concentration of 50 μg g−1 bw) to inhibit zoospore growth, the skin peptides of native New Zealand species are among the most active (see Table 1). In addition, Le. archeyi showed great individual variability in both peptide yield and efficacy at inhibiting Bd zoospores in vitro; a pattern that was also found in Centrolene prosoblepon (Woodhams et al., 2006b). However, it is important to note that the total yield of peptides collected may vary dependent on the method of collection. Both mild electric stimulation and Norepinephrine injection cause the contraction of smooth muscle surrounding the granular skin glands and subsequent discharge of secretions, but the effect on peptide yield remains to be quantified. While Le. hochstetteri and Le. pakeka produce lower quantities of peptides, their inhibitory activity is comparatively high. This shows that producing higher quantities of skin peptide mixtures does not necessarily equate to increased inhibition of zoospores growth.
Relative effectiveness of skin secretions and chytridiomycosis
Of all New Zealand species, Le. archeyi and Li. ewingii have the highest potential to reduce zoospore colonization of the skin or infection intensity based on the amount of skin secretions produced (μg g−1) and their ability to inhibit Bd zoospore growth in vitro (Fig. 1). However, Bd infections of Le. archeyi have been detected in the Coromandel region, where populations suffered major declines (Bell et al., 2004). In 2006, Le. archeyi from a second population in the Whareorino, were also diagnosed with Bd infections (Smale, 2006), even though no population declines were recorded for the area. Although Le. archeyi is susceptible to infection and reinfection with Bd (Bishop et al., 2009; S. Shaw et al., unpubl. data), no clinical signs of chytridiomycosis have ever been reported for this species and individuals are able to self-cure within 2–6 weeks. Similarly, Litoria wilcoxi can eliminate Bd in the wild (Kriger & Hero, 2006) and Taudactylus eungellensis persists with stable Bd infections in Australia (Retallick, McCallum & Speare, 2004). It is possible that the Le. archeyi population in the Whareorino are in an endemic phase of chytridiomycosis and declines in the epidemic phase remained undetected when no monitoring occurred. Alternatively, this species may not be very susceptible to Bd and the population crash in the Coromandel may not have been related to the amphibian chytrid. We suggest that skin peptides might control Bd growth as a first line of defence when initially encountering Bd, until further immunological responses can be mounted. Although all of our study animals, except one Li. aurea and one Li. raniformis, tested negative for Bd, we cannot assume they were naive to the pathogen given that individuals can apparently clear themselves of Bd infections.
The relative peptide defences of the three Litoria species are variable (Fig. 1) but all are susceptible to Bd infections and show clinical signs of chytridiomycosis in the wild and laboratory, both in Australia and New Zealand (Bell et al., 2004; Carver, 2004; Sadic & Waldman, 2004; Speare et al., 2005; R. Poulter et al., unpubl. data). Interestingly, Le. hochstetteri produces only small amounts of a relatively effective peptide mixture. This semi-aquatic species is sympatric and syntopic with the fully terrestrial Le. archeyi, and therefore likely to have also been exposed to the pathogen. However, a comprehensive New Zealand-wide Bd survey of 420 Le. hochstetteri has failed to detect any symptoms of chytridiomycosis or Bd infection using qPCR (Thurley & Haigh, 2008), despite studies showing that species breeding in permanent water bodies are often more affected by the waterborne amphibian chytrid than terrestrial species (Kriger & Hero, 2007). While the skin peptides of Le. pakeka are among the most effective at inhibiting zoospore growth, the peptide yield of this species is relatively low (Table 1). As a result, the overall peptide defences of Le. pakeka are comparatively low and might not reduce zoospore growth significantly. Leiopelma pakeka might be seriously affected if the fungus was introduced into these populations. Although, due to its restriction to several off-shore islands with highly controlled access, the chance of an encounter with Bd is relatively low.
Recently, it has been suggested that the role of anuran skin peptides in protecting species from Bd infections in the wild is not clear-cut (Conlon, Iwamuro & King, 2009). For example, Woodhams et al. (2006b) provided correlative evidence suggesting that peptide defences are linked to the persistence of Xenopus laevis and Rana pipiens in the wild, while Bufo boreas, which has weak peptide defences, is endangered. However, species with highly active skin peptides, such as Rana taharumarae or Xenopus tropicalis can still suffer from Bd-related declines (Ali et al., 2001; Parker et al., 2002; Rollins-Smith et al., 2002c). Additionally, Litoria caerulea is highly susceptible to Bd infections in the laboratory (Pessier et al., 1999; Berger, Speare & Skerratt, 2005; Woodhams et al., 2007a) and in the wild (Speare & Berger, 2005), despite the fact that it produces moderately effective skin secretions that can inhibit Bd zoospore growth at concentrations of 271 μg mL−1 (Woodhams et al., 2006a). The skin peptides of Litoria lesueuri, Litoria genimaculata, Litoria nannotis, Nyctimystes dayi and Litoria rheocola vary in their effectiveness to inhibit Bd in vitro (Woodhams et al., 2006a) and all of them have been found with Bd infections in the wild (Speare & Berger, 2005). Centrolene prosoblepon and Hylomantis lemur were predicted to be comparatively resistant to chytridiomycosis based on the quality and quantity of their skin peptides, but are suffering Bd-related declines in highland sites while populations in lowland sites seem stable (Lips et al., 2006; Woodhams et al., 2006b). Furthermore, there are several species of anurans that do not produce cytolytic peptides in their skin secretions (Conlon et al., 2009; C. Shaw, pers. comm.) and it would be of interest to determine their susceptibility to chytridiomycosis.
Factors influencing infection outcome
The level of virulence a pathogen expresses in its host is the result of complex interactions between host immunity, pathogen and environmental context and is not always easily explained by one factor alone (Poulin & Combes, 1999; Wolinska & King, 2009). Many factors are likely to influence the susceptibility to Bd, including developmental stage (Smith et al., 2005), behaviour (Parris, Reese & Storfer, 2006) and environmental conditions (Rohr et al., 2008). Both host and pathogen are highly affected by temperature and moisture; thus changes in these factors can have direct impacts on disease dynamics (Lips et al., 2008). Temperature severely impacts growth rates and infectivity of Bd (Piotrowski, Annis & Longcore, 2004; Woodhams et al., 2008), as well as causing physiological stress to anurans (Reading, 2006). In addition, peptide synthesis and secretion is dependent on a variety of factors such as bacterial flora present on the frog's skin, temperature and exposure to pollutants (Matutte et al., 2000; Harris et al., 2006; Davidson et al., 2007). Symbiotic bacteria on the skin of frogs and salamanders alone can inhibit growth of pathogenic fungi (Harris et al., 2006; Woodhams et al., 2007b; Banning et al., 2008; Lauer et al., 2008). Certain bacteria resident on amphibian skin produce anti-Bd metabolites, which can reduce morbidity and mortality of infected frogs (Harris et al., 2009). Thus, there seems to be an interaction between the presence of microorganisms and the immune response. The induction of defensive peptide production is altered by the presence of bacteria, with peptide production completely inhibited in a sterile environment (Mangoni et al., 2001). Furthermore, there are indications that peptide synthesis can in turn be impacted by Bd infection (Woodhams et al., 2007b). The complex interactions between temperature, bacteria and antimicrobial peptide production in amphibians could explain why we see Bd-associated declines of species with skin peptides that showed high in vitro efficacy. We show that all three native Leiopelmatid species produce skin secretions that are effective in inhibiting Bd zoospore growth in vitro but only small amounts are secreted onto the skin when induced by mild electric stimulation. Based on our results, we therefore recommend to continue the intense conservation management of all three native species.
Future work on the interactions of peptide defences and chytridiomycosis could include studies where susceptibility to Bd of frogs with full skin glands is compared with frogs with emptied glands. Additional studies are needed to determine the quantity of peptides released onto the skin in undisturbed individuals and in response to a Bd infection and finally, to document the complex interactions between peptide production, symbiotic skin bacteria and temperature.
Funding of this work was provided by grants from the James Sharon Watson Conservation Trust, the Society for Research on Amphibians and Reptiles in New Zealand and a University of Otago Post Graduate Scholarship to S. Melzer. We are grateful to the University of Otago for work space and technical support. For help and ongoing advice on the Bd growth inhibition assays, we are greatly indebted to L. Rollins-Smith, D. Woodhams and L. Reinert. I would also like to thank the Frog Research Group at the University of Otago for helpful comments on this paper.