Bacterial communities and Bd/Bsal screening
A screening for the two chytridiomycete pathogens, Bd and B. salamandrivorans Martel, Blooi, Bossut and Pasmans, 2013 (Bsal), was conducted in L. nigriventer, Pelophylax bedriagae (Camerano, 1882), Hyla savignyi Audouin, 1827 and Salamandra infraimmaculata Martens, 1885 from seven localities in and around the Hula Valley, to identify the possible presence or absence of these two amphibian pathogens from this region in Israel. Skin microbial communities were analysed for L. nigriventer and the syntopic Levant water frogs (P. bedriagae).
DNA was extracted from swabs with the PowerSoil® DNA Isolation Kit (Mo Bio Laboratories, Carlsbad, Ca, USA), following the Earth Microbiome project protocol (http://www.earthmicrobiome.org) except that centrifuge conditions were adjusted to accommodate reduced rotor speed.
Bd/Bsal screening was performed by duplex quantitative PCR (in duplicate) according to the protocol of Blooi et al. (2013).
To characterise skin bacterial communities, we PCR-amplified the V4 region of the bacterial 16S rRNA gene with dual-indexed primers (Kozich et al., 2013). PCRs were completed in duplicate following Sabino-Pinto et al. (2016). Negative controls were included to check for contamination. PCR products were combined for each sample and roughly quantified on 1% agarose gels. Approximately equal concentrations of PCR amplicons from each sample were pooled, gel purified using the MinElute gel extraction kit (Qiagen). DNA concentration was determined using a Qubit 2.0 and equimolar amounts were sequenced on an Illumina MiSeq platform at the Helmholtz Center for Infection Research in Braunschweig, Germany, using paired-end 2x250 v2 chemistry. Sequences were deposited in the NCBI Short Read Database (SRA BioProject PRJNA326938).
Sequences were processed with the Quantitative Insights Into Microbial Ecology (QIIME; v 1.9.1.) pipeline for Linux (Caporaso et al., 2010). Raw forward and reverse reads of each sample were joined using Fastq-join under default settings (Aronesty, 2011, 2013). After quality filtering with default settings to remove low-quality sequences, we further filtered the reads by length (250–253 bp; usegalaxy.org) and removed chimeric sequences on a per sample basis using de novo usearch61 chimera detection within QIIME (http://drive5.com/usearch/usearch_docs.html; Edgar et al., 2011). Sequences were clustered into bacterial operational taxonomic units (OTUs) with a sequence similarity threshold of 97% using an open reference OTU-picking strategy (Rideout et al., 2014, http://qiime.org/tutorials/open_reference_illumina_processing.html). SILVA 119 (24 July 2014 release; https://www.arb-silva.de) served as reference database and UCLUST (Edgar, 2010) was used in the de novo clustering steps. The most abundant sequences of each OTU were selected as representative sequences and aligned using PyNAST (Caporaso et al., 2010). OTUs with less than 0.005% of total reads were removed from the data set following Bokulich et al. (2013). Taxonomy was assigned using the RDP classifier (Wang et al., 2007) with the SILVA 119 taxonomy and representative sequences as reference, and a phylogenetic tree built using FastTree (Price et al., 2010) adhering to QIIME’s standard procedures. We also used QIIME to generate a rarefaction curve to confirm that an asymptote was reached for all samples (see S1 in the Supplement). Lastly, we rarefied the data to 1000 sequences to correct for sample depth heterogeneity and calculated beta diversity using the weighted UniFrac distance metric in QIIME.
We tested for (i) differences in dorsal and ventral skin bacterial communities of L. nigriventer, (ii) seasonal variation of the skin microbial community, and (iii) species-specific differences between L. nigriventer and P. bedriagae. Statistical analyses were done with PRIMER v7 software (Plymouth Routines In Multivariate Ecological Research; Clarke and Gorley, 2015). PERMANOVA analyses were performed using 999 permutations and associated plots were generated by principal coordinates analysis (PCoA). Core bacterial communities, defined as OTUs present in at least 75% of the samples and overlap among core communities was visualised with Venn diagrams drawn with VENNY 2.0 (http://bioinfogp.cnb.csic.es/tools/venny/).