Bd/Bsal screening, skin bacterial community, and defensive skin peptides
The pathogen Bd was detected in 32% of the tested amphibian individuals (n = 87) from northern Israel, while none were positive for Bsal. We found Bd in two amphibian species (L. nigriventer and P. bedriagae) and in three of the seven examined locations within the Hula Valley (Hula Nature Reserve, Kiryat Shmona and Yesod HaMa’ala). Infection loads for Bd-positive individuals ranged between 1–311 genomic equivalents of zoospores per swab (Table 4).
Table 4. Number of individuals per species tested for Batrachochytrium dendrobatidis including mean (range) of genomic equivalents of zoospores per swab for positive tested specimen.
Bacterial communities of L. nigriventer were comprised of Proteobacteria (56.7%) with a high representation of Gammaproteobacteria (33.4% of the overall community), Bacteroidetes (25.3%) and Firmicutes (6.7%) (Fig. 9 A). The 20 most abundant OTUs found on the skin of L. nigriventer represented 42% of the total reads (Fig. 9 B; S6 in the Supplement). The most abundant OTU (7% of the total sequences) was assigned to an unspecified Chryseobacterium and present in 100% of the samples, although in varying abundance (< 1–24% of the reads). Comparisons based on weighted UniFrac distances did not reveal significant differences between (i) microbial communities from the ventral versus dorsal skin of L. nigriventer (PERMANOVA: N = 27; p = 0.117; Fig. 10 A), (ii) ventral surfaces of females versus males (PERMANOVA: N = 15; p = 0.646; Fig. 10 B) or (iii) ventral surfaces of Bd-positive versus Bd-negative individuals (PERMANOVA: N = 22; p = 0.283; Fig. 10 C). However, we observed a significant shift in the ventral skin microbial community over time.
While no significant changes were observed between the ventral skin samples taken in mid-February and mid-April (N = 13; p = 0.799) or between those collected in mid-April and the end of June (N = 15; p = 0.0.093), significant differences were detected for all other time-associated comparisons: mid-February – late June (N = 15; p = 0.012); mid-February – mid-September (N = 15; p = 0.001); late June – mid-September (N = 17; p = 0.001) (Fig. 10 D).
The results obtained for the skin-associated bacterial communities of syntopic P. bedriagae were similar to those of L. nigriventer: no significant differences between ventral versus dorsal surfaces of the same individuals (N = 14; p = 0.898) nor between ventral surfaces of Bd-positive versus Bd-negative individuals (N = 22; p = 0.366).
A comparison of the ventral skin-associated communities of L. nigriventer and P. bedriagae from the same location and same time-point revealed differences between the two species (N = 17; p = 0.001; Fig. 10 E). The core bacterial communities contained 30 OTUs (88% of the core skin microbiota of L. nigriventer and 57% of that of P. bedriagae) that were present on the ventral skin of at least 75% of the individuals of both species (Fig. 10 F).
The skin secretions collected from two different individuals and examined for peptide composition had significant amounts of hydrophobic peptides recovered after C18 enrichment. We detected a number of common peptide mass signals shared by both frog individuals. The mass ranges are suggestive of possible antimicrobial peptides (Table 5; S7 in the Supplement). In a growth inhibition assay, the mixture of peptides inhibited the growth of two different Bd isolates (JEL 197 and ‘Section Line’; Fig. 11). At the highest concentration tested (500 μg/ml), Bd growth inhibition ranged from 51% to 91.5% against the Section Line isolate and 70–82% inhibition against the original type isolate JEL 197. Both isolates are among the global panzootic lineages (Schloegel et al., 2012; Piovia-Scott et al., 2015). Furthermore, the direct skin secretion solution was found to inhibit Bd by 35–36%.