Abstract
The mucous skin of amphibians provides a habitat for microorganisms which may interact with their hosts and thereby affect their condition and health. Cultivation-independent analyses of the bacterial communities based on the detection of PCR-amplified bacterial 16S rRNA genes provides a direct approach to characterize their diversity. In the present pilot study we utilized this approach in combination with a high-throughput DNA sequencing technology (454 pyrosequencing), to characterize the bacterial community structure of the skin of three newt species (Lissotriton vulgaris, Ichthyosaura alpestris, Triturus cristatus), collected near Braunschweig, Germany. 16S rDNA sequences were obtained from 19 unique samples. On average, 6113 amplicon sequences were obtained per sample and these could phylogenetically be assigned to a total of 1615 different operational taxonomic units (OTUs). Altogether, most samples were rather similar in their dominant bacterial taxa. Most represented were Betaproteobacteria (46%; mostly Janthinobacterium), Gammaproteobacteria (28%; mostly Pseudomonas), Flavobacteria (phylum Bacteroidetes: 19%, mostly Flavobacterium), and Sphingobacteria (Bacteroidetes: 5%, mostly Pedobacter). We found no significant differences between the three newt species, or between hemi-nested vs. non-nested PCR, but a strong difference among sampling dates (15 and 17 April 2013) which might be explained by the ongoing transition of the newts from their terrestrial to aquatic phase which coincided with this period, or by differences between sexes as these were unevenly sampled on the two dates. 16S rRNA gene sequences retrieved in this study in several cases were identical or very similar to those previously found on the skin of North American salamanders.
Introduction
The skin surface of amphibians is characterized by a mucous layer with a diverse array of chemical compounds secreted by specialized skin glands, which provides a potential habitat to numerous microorganisms. This microbial ecosystem, for which Woodhams et al. (2014) coined the term mucosome, is still poorly understood, and there are indications that in fact the microorganisms present in the mucosome can play crucial roles for the health of the respective amphibians. Pathogens such as chytrid fungi, genus Batrachochytrium (e.g., Fisher et al., 2012; Martel et al., 2013), dwell on the amphibian skin, but their growth can in turn be inhibited by chemical compounds such as antimicrobial peptides secreted by the amphibians, or antibiotics produced by amphibian skin bacteria (Woodhams et al., 2005, 2007; Harris et al., 2006, 2009a, 2009b; Rollins-Smith et al., 2006; Lauer et al., 2007; Becker and Harris, 2010; Bletz et al., 2013). Although unstudied, it is plausible that in amphibians, bacteria-host interactions might also have effects on behavior and evolution of amphibians, as it is known from other animals (Ezenwa et al., 2012). Similar to the skin of humans (Findley et al., 2013), the amphibian skin is populated by a diverse community of bacteria and fungi, with complex interactions among them. The microbial communities may harbour pathogens, antagonists, and specialized symbionts. It is an important challenge to identify such taxa of relevance for amphibian biology among the background of transient opportunistic microbial inhabitants.
Cultivation independent analyses of the diversity of the bacterial communities colonizing the skin of amphibians, especially from North America, have detected Bacteroidetes, Gammaproteobacteria, Alphaproteobacteria, Firmicutes, Sphingobacteria and Actinobacteria as taxonomic components of these microbiota (McKenzie et al., 2012; Kueneman et al., 2013). Differences in the composition of skin-associated bacterial communities were found between species, between larval and adult stages, and also, between individuals of the same amphibian species (McKenzie et al., 2012; Kueneman et al., 2013; Loudon et al., 2014).
Studies on the skin-associated bacteria of European amphibians are still scarce. Here, we report on a pilot study using DNA directly extracted from skin mucus of newts, family Salamandridae. The selected newt species reproduce in ponds in spring, and spend the rest of the year on land. Next-generation sequencing of 16S rDNA amplicons was applied to detect the dominant bacterial community members in these samples. Using sterile swabs, we collected skin mucus of three different newt species on two days at the start of the breeding season in early spring and thus shortly after a transition from terrestrial to aquatic phase. Our data provide a first glimpse of the bacterial diversity of the skin microbiota of western Palearctic amphibians.
Materials and methods
Field sites and sampling
Samples were collected in April 2012 (20th) and 2013 (15th & 17th) at night in a pond near Lelm (close to Braunschweig, Germany; geographical coordinates: 52.213336 latitude, 10.83031 longitude; WGS84 datum). Specimens were caught by hand from shallow areas of a pond using a fresh pair of nitrile gloves for each specimen. Specimens were rinsed by spraying 20-30 ml of distilled water using a laboratory wash bottle over their body surface to remove transient bacteria, taking care that all remains of mud or other small particles, as well as all pond water, was fully washed off. Subsequently they were gently swabbed for about 30 s on dorsal and ventral side (avoiding the cloacal region but including the dorsal crests of males) using MW113 swabs (Medical Wire & Equipment, Corsham, UK). Daily minimum and maximum water temperatures in the pond were recorded using an iButton data logger. Swabs were preserved in sterile 1.5 ml tubes with 50 μl TE buffer and frozen upon 2 hours after collection. Although it cannot be excluded that microbial communities on swabs may have slightly changed during this short pre-freezing interval, such changes are very unlikely given previous studies (Lauber et al., 2010) and would likely have affected all samples in a similar way, thus not biasing our comparisons. DNA was extracted from the entire swabs using the Fast DNA Spin Kit for Soil (MP) following manufacturer’s instructions.
Standard PCR protocol
The hypervariable V1 and V2 region of the 16S rRNA gene (16S rDNA) was amplified using the composite forward (5′-CTATGCGCCTTGCCAGCCCGC TCAG TC AGAGTTTGATCCTGGCTCAG-3′) and reverse (5′-CGTATCGCCTCCCTCGCGCCA TCAG ########## CA TGCTGCCTCCCGTAGGAGT-3′) primers. These primers contained the 454 Life Sciences Adaptor B (forward) and A (reverse) (in italics) and the broadly conserved bacterial primers 27F and 338R (underlined). A two-base linker sequence (TC/CA) and four-base key (TCAG) were added as recommended by Roche (454). The reverse primer furthermore contained a unique 10 base multiplex identifier (MID; in the primer sequence above designated as ##########) to tag each PCR product. For PCR, 1 μl of DNA template was added to 25 μl PCR reactions. These were performed using Phusion® Hot Start DNA Polymerase II (Fisher Scientific, Schwerte, Germany). The cycling conditions were as follows: initial denaturation for 90 s at 98°C; 25 cycles of 10 s at 98°C, 30 s at 55°C, and 30 s at 72°C; final extension for 10 min at 72°C.
Hemi-nested PCR protocol
Because no PCR products could be obtained from some of our samples, we explored for the 2013 sample set whether reproducible results could be obtained using hemi-nested PCRs. For a set of samples (from the same DNA extractions used in the standard PCR protocol) we carried out a hemi-nested PCR approach, with a first amplification using primers 16S-27F (AGAGTTTGATCCTGGCTCAG) and 16S-1378R (CGGTGTGTACAAGGCCCGGGAACG) (94°C for 4 min, and 45 cycles of 94°C for 45 s, 45°C for 40 s, 72°C for 2 min; final elongation for 10 min) and a second PCR with the composite forward and reverse primers as above (thermocycling protocol as in first PCR but with 35 cycles), the forward primer corresponding to the same annealing site as the 16S-27F primer.
454 sequencing
All PCR reactions were performed in duplicate, and subsequently combined. PCR products were extracted with the Qiagen MiniElute Gel Extraction Kit and quantified with Quant-iT PicoGreen dsDNA reagent (Life Technologies GmbH, Darmstadt, Germany) on a Mithras LB 940 multimode microplate reader (Berthold Technologies, Bad Wildbad, Germany). Equimolar amounts of purified PCR product were pooled and further purified using Ampure Beads (Agencourt). The complete library was run on an Agilent Bioanalyzer prior to emulsion PCR and sequencing as recommended by Roche. Amplicon libraries were subsequently sequenced on a 454 GS-FLX using Titanium sequencing chemistry.
Cloning of PCR products
For a series of samples collected in 2012, we amplified 16S rDNA using 16S-27F and 16S-1378R primers and a hotstart PCR protocol, ligated the purified PCR products into the vector pGEM®-T, and transformed them into competent cells of E. coli JM109. Transformants were further selected by blue/white screening and sequenced on an automated capillary sequencer (ABI 3730xl). Newly determined sequences were submitted to Genbank (accession numbers KM817575-KM817592). Original 454 sequences are available from the authors and will be made available at the ENA repository (http://www.ebi.ac.uk/ena/).
Filtering and exclusion of samples
Sequences from 454 sequencing were processed with the software mothur v.1.32 and v.1.33 (Schloss et al., 2009), following the mothur 454 SOP version from February 2014 including qfile, with qwindowaverage = 35, qwindowsize = 50, reference = silva.bacteria.fasta, minlength = 150 and average neighbor algorithm for clustering. Chimeric sequences (21.7%) were identified with chimera.uchime (http://drive5.com/uchime) and removed from the dataset, as well as four chloroplast sequences. Sequences were grouped in operational taxonomic units (OTU) at similarities of at least 97% and classified in accordance to the RDP taxonomy with trainset9_032012. After filtering and removal of chimeric sequences, a total of 385 965 16S rDNA sequences were retrieved (normal plus hemi-nested PCR approach). Six samples from the hemi-nested PCR approach showed an unusual and strongly deviant composition of sequences compared to the other samples, with a high representation of sequences of some unusual taxa (e.g., Actinobacteria and Ralstonia). Since similar sequences were also found in small numbers in one of the negative controls (no sequences from two other negative controls), we considered these data as a result of contamination and excluded the respective samples from further analysis.
Proportion of sequences of main bacterial taxa obtained by 454 sequencing from all swabs of newts analyzed. Only genera with proportions of >1.5% representation in the entire data set are included. Number of OTUs per genus as obtained by the chosen algorithm; note that many of these OTUs were represented by only very few (often only single) sequences and might represent sequencing errors, inherent to next generation sequencing technology; OTU number should therefore be taken only as tentative indication of the true diversity of phylotypes per genus.
Proportion of the five most common genera of bacteria (proportion > 3%) plus Ralstonia among 16S rRNA sequences obtained from skin swab samples of three species of newts (from 15 and 17 April 2013). Numbers with asterisk indicate those samples amplified with hemi-nested PCR. A capital F or M at the end of the name indicates samples taken from female or male newts, respectively. Note that especially in the alpine newt (I. alpestris) but also in the other species, a distinct reduction of the percentage of Janthinobacterium sequences and an augmentation of Pseudomonas is observed from 15 to 17 April 2013, concordant with increasing water temperatures in the pond. This figure is published in color in the online version.
Citation: Amphibia-Reptilia 36, 1 (2015) ; 10.1163/15685381-00002970
Statistical analysis
To compare the phylogenetic diversity in sequences obtained by 454 sequencing, an approximately-maximum-likelihood phylogenetic tree was calculated with FastTree Version 2.1.3 SSE3 (Price et al., 2010) and used for the calculation of a weighted Unifrac distance matrix (Lozupone and Knight, 2005) with mothur. This matrix was used for nonmetric multidimensional scaling (NMDS) with mothur and for analysis of similarities (ANOSIM) in R (vegan package). For Unifrac comparisons the dataset was rarefied to 583 sequences per sample. Considering that analyses of all sequences revealed similar results, i.e., Unifrac distances differed by 0.3%, this procedure was justified, and was applied to balance sample sizes.
Results
Summarizing the 24 skin swab samples collected in 2013 from all three newt species, a total of 146 753 16S rDNA sequences were used for analysis. Some of these samples were repeats of the same individuals and excluded from statistical comparisons (ANOSIM; see below), but are included in the summary statistics in the following paragraph.
Per sample, 583-13401 sequences were obtained (average 6113). A total of 1615 OTUs were identified, but most of these were relatively or very rare, and some might represent sequencing errors as inherently recovered by next generation sequencing technology. Each sample was dominated by a limited number of common OTUs. Altogether, only 10 genera were found in proportions > 1.5% (table 1), and only five genera in proportions > 3% (table 1; fig. 1). At the phylum level, 76% of the sequences were Proteobacteria (869 OTUs) and 23% Bacteroidetes (570 OTUs), with only minor proportions of a series of other phyla. The most represented genera were Janthinobacterium, Pseudomonas, and Flavobacterium (detailed proportions in table 1).
The 24 samples from 2013 included (i) 15 unique samples amplified with the normal hotstart protocol, (ii) one repeated sample with the same protocol, (iii) four samples repeated with hemi-nested PCR and thus allowing a comparison among the methods and (iv) an additional four samples obtained by hemi-nested PCR from samples that did not amplify with the normal protocol. After excluding repeated samples, data of 19 newt individuals (15 obtained by normal PCR, four obtained by hemi-nested PCR) were available for comparisons between sampling days, species, sex, and PCR method. In all three newt species, and especially in I. alpestris, specimens captured on 17 April show different proportions of dominant skin bacteria, with an increased proportion of Pseudomonas and a decreased proportion of Flavobacterium. This pattern was largely reproducible in repeats of the same samples (see online supplementary material). Comparing the Unifrac distance matrix (fig. 2) among samples, a significant influence on community composition by sampling date was observed (ANOSIM R = 0.45; ) whereas no such influence was detected for host species (R = −0.07; ) or PCR method (R = −0.17; ). Due to an uneven distribution of males and females among the samples on the two days (see Discussion), also the difference between sexes was statistically significant (R = 0.42; ).
NMDS plots of samples of skin bacterial composition of three newt species, based on a weighted Unifrac distance matrix that considers phylogenetic relatedness among bacterial sequences. Only samples from 2013 included. NMDS stress: 0.117, R-squared for configuration: 0.959. As supported by ANOSIM analysis, sampling date and sex are significant predictor of the bacterial community (due to uneven sampling of sexes on the two dates) while host species and PCR method have no significant influence.
Citation: Amphibia-Reptilia 36, 1 (2015) ; 10.1163/15685381-00002970
Sequences of 500-700 bp (single-stranded sequences cloned from 16S rDNA amplicons) were available for skin bacteria isolated from two specimens of L. vulgaris (19 and 17 sequences) and one specimen of T. cristatus (12 sequences), collected in 2012 at Kleiwiesen near Braunschweig. We aligned these sequences to those obtained by Lauer et al. (2008) from North American salamanders (Hemidactylium scutatum, Plethodon cinereus). Sequence comparison revealed that 28 of the 48 sequences obtained by us corresponded to nine clones from the North American dataset, with identical or almost identical 16S sequences. Extending this analysis to the 2013 samples obtained by 454 pyrosequencing, the vast majority of Janthinobacterium from all samples belonged to a single OTU (27% of all sequences) which differed by only 2 mutations in 270 aligned nucleotides (99% sequence similarity) from clone 15K obtained by Lauer et al. (2008) from Plethodon cinereus.
Discussion
The skin of salamandrid newts contains a considerable amount of toxic substances, among which tetrodotoxins (Yotsu-Yamashita et al., 2007; Hanifin, 2010) and steroidal alkaloids (Vences et al., 2014). Although some amphibian alkaloids are known to have antibacterial effects (Macfoy et al., 2005), it is uncertain whether this is true for newt toxins. Still, it is likely that in combination with other factors these substances generate a selective environment for specifically adapted microorganisms. It therefore is remarkable that a considerable number of dominant bacteria in European newts are identical or very similar in their 16S rDNA sequences (table 2) to those determined from salamanders of the genera Hemidactylium and Plethodon (Lauer et al., 2008). These not only are geographically and phylogenetically distant (North America vs. Europe; family Plethodontidae vs. Salamandridae) but also, as far as known, have secretions of a different chemical composition.
Comparison of DNA sequences of the 16S rRNA gene obtained by amplicon cloning from samples of Triturus cristatus (Tc) and Lissotriton vulgaris (Lv) collected at Kleiwiesen near Braunschweig in 2012, with sequences of skin bacteria from North American salamanders (Hemidactylium and Plethodon: Lauer et al., 2008). The majority of skin bacteria sequences from Germany were identical or differed by only 1-2 substitutions from those of the North American salamanders.
On the other hand, the overall composition of the bacterial communities determined herein differs from that observed on the skin of other North American amphibians, largely in their aquatic phase (McKenzie et al., 2012; Kueneman et al., 2014): similar to these authors we found a high representation of Bacteroidetes and Gammaproteobacteria, but the dominant role of Betaproteobacteria in our data set (largely caused by a single species of Janthinobacterium) is a marked difference. Janthinobacterium is known to exert anti-chytrid activity (Harris et al., 2009a) and has been reported in North America especially from largely or exclusively terrestrial salamanders (Hemidactylium scutatum, Plethodon cinereus). According to Lauer et al. (2008), about half of the bacterial genera and families displaying antifungal activity were shared among Hemidactylium and Plethodon, but there was virtually no overlap at the species level. In contrast, according to our data, 16S sequences of Janthinobacterium that are almost identical to those found on Plethodon (clone 15K) and Hemidactylium (clone CH26-6), in Europe occur on the same species of newt (L. vulgaris) in the same pond.
Altogether, in terms of dominant bacterial taxa, all of our samples analyzed by 454 pyrosequencing resulted to be rather similar to each other. Since our 2013 sampling was limited to a single pond, we can exclude additional confounding environmental factors: all studied newts were exposed to the same water and thus the same environmental microbial community, were sampled using the same type of gloves, rinsed with the same distilled water, and kept at environmental temperature for a similar period of time (two hours). We rinsed specimens with abundant distilled water exerting some water pressure during the rinsing process which leaves us confident that transient bacteria were largely reduced; and none of the control samples contained those bacteria most relevant for the differences between sexes/sampling dates (i.e., Janthinobacterium, Flavobacterium, Pseudomonas), indicating that the similarity observed among species is unlikely to have been caused by transient bacteria or contamination. Also, despite some obvious differences between repeated samples (supplementary fig. S1) it is unlikely that our results were heavily influenced by different selective amplification in the two PCR methods, as such major methodological influences would have been detectable by the ANOSIM analysis.
Due to failure of standard PCR in many of the originally collected samples and exclusion of some hemi-nested PCR samples due to probable contamination, our final sampling of 454 data is somewhat opportunistic. Especially the biased distribution of males vs. females on the two sampling dates makes it impossible to statistically distinguish between two alternative hypotheses explaining the encountered variation: these might primarily be caused by temporal fluctuation, or by intersexual differences of skin bacterial communities. In fact, our initial sampling design did not consider the possibility of differences between the two sampling dates as these were only two days apart. If temporal differences are invoked, then these might be related to the fact that our sampling took place in early spring, shortly after melting of an ice layer on the study pond. Central European newts hibernate on land and start migrating into the pond shortly after the ice cover started to melt, during a period of rising water temperatures (supplementary material). During the change from terrestrial to aquatic phase the newt skin undergoes dramatic changes (e.g., Perrotta et al., 2012), and the specimens might also have sloughed their skin which is known to affect the community of skin-associated bacteria (e.g., Meyer et al., 2012; Cramp et al., 2014).
Different from other studies of bacterial communities associated to the amphibian skin (e.g., McKenzie et al., 2012; Kueneman et al., 2013) our ANOSIM analysis does not indicate a strong influence of host identity (species) on community composition but such differences might well have remained undetected by our data. On one hand, the low sample size and strong effect of sampling date and/or sex might have statistically masked possible minor differences between species. In fact, visual inspection of fig. 1 suggests a possible frequency difference in the dominant bacteria between T. cristatus and I. alpestris on one sampling date (17 April) but due to low sample sizes (females, 2 vs. 3 individuals; males, 2 vs. 1 individual) we here refrain from a statistical comparison. On the other hand, the newts tested here are ecologically rather similar part of one subclade in the Salamandridae (Steinfartz et al., 2007) and thus phylogenetically related whereas McKenzie et al. (2012) and Kueneman et al. (2013) compared species belonging to different families and partly different orders (salamanders vs. anurans). Hence, inter-species differences might be detectable more easily between phylogenetically distant and ecologically distinct species.
Although preliminary, our results highlight the importance of large sample sizes across species, seasons, ecological phases and developmental stages to understand the dynamics of amphibian skin microbiota. Studies in progress will analyze large sample sets from the same location, and will thereby be able to better quantify the influences of the various predictors. Whether the rather common PCR failure in our samples was due to PCR inhibition (despite usage of a soil DNA extraction kit supposedly removing inhibitory agents), or to low density of bacteria on some newt individuals, remains as an important field for future studies aimed at methodological optimization. Also, more extensive comparisons over small and large geographical scales can identify those bacteria specializing on the amphibian skin (and maybe other, similar mucous tissues of vertebrates), and distinguish them from others that are only opportunistic and transient in this habitat. Furthermore, the pattern observed herein also encourages experimental manipulation of the system, to better understand if and how environmental changes (terrestrial vs. aquatic or temperature) might induce rapid changes in bacterial community composition.
Acknowledgements
We are grateful to Meike Kondermann for her help with labwork, in particular DNA extraction. Susanne Hauswaldt helped with sampling and supervised part of the labwork. Ariel Rodriguez contributed the data logger readings of pond temperature. This study was supported by Deutsche Forschungsgemeinschaft (grant VE247/9-1 to MV).
References
Becker M.H., Harris R.N. (2010): Cutaneous bacteria of the redback salamander prevent morbidity associated with a lethal disease. PLOS One 5: e10957.
Bletz M.C., Loudon A.H., Becker M.H., Bell S.C., Woodhams D.C., Minbiole K.P., Harris R.N. (2013): Mitigating amphibian chytridiomycosis with bioaugmentation: characteristics of effective probiotics and strategies for their selection and use. Ecol. Lett. 16: 807-820.
Cramp R.L., McPhee R.K., Meyer E.A., Ohmer W.E., Franklin C.E. (2014): First line of defence: the role of sloughing in the regulation of cutaneous microbes in frogs. Conserv. Physiol. 2: cou012.
Ezenwa V.O., Gerardo N.M., Inouye D.W., Medina M., Xavier J.B. (2012): Animal behavior and the microbiome. Science 338: 198-199.
Findley K., Oh J., Yang J., Conlan S., Deming C., Meyer J.A., Schoenfeld D., Nomicos E., Park M., Kong H.H., Segre J.A. (2013): Topographic diversity of fungal and bacterial communities in human skin. Nature 498: 367-370.
Fisher M.C., Henk D.A., Briggs C.J., Brownstein J.S., Madoff L.C., Mccraw S.L., Gurr S.J. (2012): Emerging fungal threats to animal, plant and ecosystem health. Nature 484: 186-194.
Hanifin C.T. (2010): The chemical and evolutionary ecology of tetrodotoxin (TTX) toxicity in terrestrial vertebrates. Mar. Drugs 8: 577-593.
Harris R.N., James T.Y., Lauer A., Simon M.A., Patel A. (2006): Amphibian pathogen Batrachochytrium dendrobatidis is inhibited by the cutaneous bacteria of amphibian species. Ecohealth 3: 53-56.
Harris R.N., Brucker R.M., Walke J.B., Becker M.H., Schwantes C.R., Flaherty D.C., Lam B.A., Woodhams D.C., Briggs C.J., Vredenburg V.T., Minbiole K.P. (2009a): Skin microbes on frogs prevent morbidity and mortality caused by a lethal skin fungus. ISME J. 3: 818-824.
Harris R.N., Lauer A., Simon M.A., Banning J.L., Alford R.A. (2009b): Addition of antifungal skin bacteria to salamanders ameliorates the effects of chytridiomycosis. Dis. Aquat. Organ. 83: 11-16.
Kueneman J.G., Wegener Parfrey L., Woodhams D.C., Archer H.M., Knight R., McKenzie V.J. (2013): The amphibian skin-associated microbiome across species, space and life history stages. Mol. Ecol. 23: 1238-1250.
Lauber C.L., Zhou N., Gordon J.I., Knight R., Fierer N. (2010): Effect of storage conditions on the assessment of bacterial community structure in soil and human-associated samples. FEMS Microbiology Letters 307: 80-86.
Lauer A., Simon M.A., Banning J.L., Andre E., Duncan K., Harris R.N. (2007): Common cutaneous bacteria from the eastern red-backed salamander can inhibit pathogenic fungi. Copeia 3: 630-640.
Lauer A., Simon M.A., Banning J.L., Lam B.A., Harris R.N. (2008): Diversity of cutaneous bacteria with antifungal activity isolated from female four-toed salamanders. ISME J. 2: 145-157.
Loudon A.H., Woodhams D.C., Parfrey L.W., Archer H.M., Knight R., McKenzie V.J., Harris R.N. (2014): Microbial community dynamics and effect of environmental microbial reservoirs on red-backed salamanders (Plethodon cinereus). ISME J. 8: 830-840.
Lozupone C., Knight R. (2005): UniFrac: a new phylogenetic method for comparing microbial communities. Appl. Environ. Microbiol. 71: 8228-8235.
Macfoy C., Danosus D., Sandit R., Jones T.H., Garraffo H.M., Spande T.F., Daly J.W. (2005): Alkaloids of anuran skin: antimicrobial function? Zeitschrift für Naturforschung C-Journal of Biosciences 60: 932-937.
Martel A., Spitzen-van der Sluijs A., Blooi M., Bert W., Ducatelle R., Fisher M.C., Woeltjes A., Bosman W., Chiers K., Bossuyt F., Pasmans F. (2013): Batrachochytrium salamandrivorans sp. nov. causes lethal chytridiomycosis in amphibians. Proc. Natl. Acad. Sci. USA 110: 15325-15329.
McKenzie V.J., Bowers R.M., Fierer N., Knight R., Lauber C.L. (2012): Co-habiting amphibian species harbor unique skin bacterial communities in wild populations. ISME J. 6: 588-596.
Meyer E.A., Cramp R.L., Bernal M.H., Franklin C.E. (2012): Changes in cutaneous microbial abundance with sloughing: possible implications for infection and disease in amphibians. Dis. Aquat. Organ. 101: 235-242.
Perrotta I., Sperone E., Bernabò I., Tripepi S., Brunelli E. (2012): The shift from aquatic to terrestrial phenotype in Lissotriton italicus: larval and adult remodelling of the skin. Zoology 115: 170-178.
Price M.N., Dehal P.S., Arkin A.P. (2010): FastTree 2 – approximately maximum-likelihood trees for large alignments. PLoS ONE 5: e9490.
Rollins-Smith L.A., Woodhams D.C., Reinert L.K., Vredenburg V.T., Briggs C.J., Nielsen P.F., Conlon J.M. (2006): Antimicrobial peptide defenses of the mountain yellowlegged frog (Rana muscosa). Dev. Comp. Immunol. 30: 831-842.
Schloss P.D., Westcott S.L., Ryabin T., Hall J.R., Hartmann M., Hollister E.B., Lesniewski R.A., Oakley B.B., Parks D.H., Robinson C.J., Sahl J.W., Stres B., Thallinger G.G., Van Horn D.J., Weber C.F. (2009): Introducing mothur: Open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl. Environ. Microbiol. 75: 7537-7541.
Steinfartz S., Vicario S., Arntzen J.W., Caccone A. (2007): A Bayesian approach on molecules and behaviour: reconsidering evolutionary patterns in Triturus newts (Amphibia: Salamandridae). J. Exp. Zool. B: Mol. Dev. Evol. 308B: 139-162.
Vences M., Sanchez E., Hauswaldt J.S., Eikelmann D., Rodríguez A., Carranza S., Donaire D., Gehara M., Helfer V., Lötters S., Werner P., Schulz S., Steinfartz S. (2014): Nuclear and mitochondrial multilocus phylogeny and survey of alkaloid content in true salamanders of the genus Salamandra (Salamandridae). Mol. Phylogenet. Evol. 73: 208-216.
Woodhams D.C., Brandt H., Baumgartner S., Kielgast J., Küpfer E., Tobler U., Davis L.R., Schmidt B.R., Bel C., Hodel S., Knight R., McKenzie V. (2014): Interacting symbionts and immunity in the amphibian skin mucosome predict disease risk and probiotic effectiveness. PLOS One 9: e96375.
Woodhams D.C., Vredenburg V.T., Simon M., Billheimer D., Shakhtour B., Shyr Y., Briggs C.J., Rollins-Smith L.A., Harris R.N. (2007): Symbiotic bacteria contribute to innate immune defenses of the threatened mountain yellow-legged frog, Rana muscosa. Biol. Conserv. 138: 390-398.
Woodhams D.C., Voyles J., Lips K.R., Carey C., Rollins-Smith L.A. (2005): Predicted disease susceptibility in a Panamanian amphibian assemblage based on skin peptide defenses. J. Wildl. Dis. 42: 207-218.
Yotsu-Yamashita M., Mebs D., Kwet A., Schneider M. (2007): Tetrodotoxin and its analogue 6-epitetrodotoxin in newts (Triturus spp.; Urodela, Salamandridae) from southern Germany. Toxicon 50: 306-309.
Footnotes
Associate Editor: Jonathon Marshall.