COS 144-4
Skin microbiome composition varies among sites, but not co-occurring Plethodon salamander species along environmental gradients

Friday, August 14, 2015: 9:00 AM
326, Baltimore Convention Center
Carly R. Muletz Wolz, Department of Biology, University of Maryland, College Park, MD
Evan H. Campbell Grant, Patuxent Wildlife Research Center, US Geological Survey
Robert Fleischer, Center for Conservation and Evolutionary Genetics, National Zoological Park, Washington, DC
Karen R. Lips, Department of Biology, University of Maryland, College Park, MD

The composition of host-associated microbiota can be shaped by mechanisms acting at the host species level, representing evolutionary time, and at the host environment level, representing recent time. Previous studies on the amphibian skin microbiome indicated that host species is a significant predictor of microbiome composition with less significant site-to-site variation often observed. This suggests that mechanisms acting at the level of host species are more stabilizing than those operating at the environment level. Yet, the host environment influences composition and may do so in a predictable manner. To test this, we characterized skin bacterial composition on three Plethodon salamander species along latitudinal and altitudinal gradients. We collected skin swabs from 74 P. cinereus, 14 P. glutinosus and 13 P. cylindraceus at three protected areas (Catoctin MP, Shenandoah NP, and Mt. Rogers NRA) along a 497 km latitudinal transect in the Appalachians. At Shenandoah NP, P. cinereus was sampled along an altitudinal gradient (702 – 974 m) at 100 m increments with two populations per elevation. We used 454 pyrosequencing of the 16s rRNA gene to identify all bacteria from all field swabs.


We generated 244,154 high quality sequences representing 562 OTUs from 22 described bacterial phyla. Actineobacter and Pseudomonas (phylum Proteobacteria) were numerically dominant representing 40 % and 36% of total sequences, respectively. Furthermore, two Actinetobacter OTUs and one Pseudomonas OTU were present on every salamander sampled. Both genera are known for their antibiotic properties and may provide protection to the host from microbial pathogens. Contrary to expectations, we found no differences among salamander species in OTU richness (ANOVA: p = 0.416; μ = 60 ± 22 OTUs) or community composition (PERMANOVA: p = 0.159). As expected, we found differences in bacterial community composition among sites (PERMANOVA: p = 0.001), with salamanders at Shenandoah NP most dissimilar from salamanders at Catoctin MP and Mt. Rogers NRA. We found a weak correlation between P. cinereus OTU richness and elevation (p = 0.017, R2 = 0.09), and between P. cinereus community composition similarity and elevation (p = 0.001, R2 = 0.15). Our results suggest that Plethodon salamander skin microbiome can be predicted by environmental factors and not by host species. The microbiome similarity of congeneric Plethodon species suggests that the mechanisms driving speciation of the salamanders did not result in bacterial community divergence or selection for a specific bacterial community caused a once diverged community to converge.