Bacterial Adaptation to Venom in Snakes and Arachnida

Abstract

Animal venoms are considered sterile sources of antimicrobial compounds with strong membrane-disrupting activity against multidrug-resistant bacteria. However, venomous bite wound infections are common in developing nations. Investigating the envenomation organ and venom microbiota of five snake and two spider species, we observed venom community structures that depend on the host venomous animal species and evidenced recovery of viable microorganisms from black-necked spitting cobra (Naja nigricollis) and Indian ornamental tarantula (Poecilotheria regalis) venoms. Among the bacterial isolates recovered from N. nigricollis, we identified two venom-resistant, novel sequence types of Enterococcus faecalis whose genomes feature 16 virulence genes, indicating infectious potential, and 45 additional genes, nearly half of which improve bacterial membrane integrity. Our findings challenge the dogma of venom sterility and indicate an increased primary infection risk in the clinical management of venomous animal bite wounds. IMPORTANCE Notwithstanding their 3 to 5% mortality, the 2.7 million envenomation-related injuries occurring annually-predominantly across Africa, Asia, and Latin America-are also major causes of morbidity. Venom toxin-damaged tissue will develop infections in some 75% of envenomation victims, with E. faecalis being a common culprit of disease; however, such infections are generally considered to be independent of envenomation. Here, we provide evidence on venom microbiota across snakes and arachnida and report on the convergent evolution mechanisms that can facilitate adaptation to black-necked cobra venom in two independent E. faecalis strains, easily misidentified by biochemical diagnostics. Therefore, since inoculation with viable and virulence gene-harboring bacteria can occur during envenomation, acute infection risk management following envenomation is warranted, particularly for immunocompromised and malnourished victims in resource-limited settings. These results shed light on how bacteria evolve for survival in one of the most extreme environments on Earth and how venomous bites must be also treated for infections.

Keywords: drug resistance evolution; extremophiles; genome analysis; microbiome; multidrug resistance; venom.

FIG 1

Snake venom microbiomes cluster on account of host species. Viperid venom microbiomes cluster separately from N. nigricollis, with the exception of B. atrox, as determined by (A) PCoA, (B) an UPGMA tree (i.e., unweighted pair group method using average linkages), and (C) class-level taxonomic profiling following 16S rRNA phylogenetic analysis. Dots in panels A and B are colored by species (red, B. arietans; black, N. nigricollis; light blue, B. atrox; orange, C. atrox; dark blue, O. scutallatus), represent data from individuals in captivity, are labeled with the short species name, are enumerated for the individual, and are identified for the envenomation number (E1 or E2) of the sample. The 8 wild B. arietans samples (red dots B1 to B8) and the commercially sourced, lyophilized B. arietans sample (red L dot) are independently labeled. Sample B3 was removed from the analysis due to the yield of ~100× lower read depth from this sample compared to all other B. arietans samples. Relative taxonomic diversity profiles in panel C are aligned to the UPGMA tree sample labels, with the Shannon-Wiener index (SWI) of each sample indicated. The geographical origins of the wild B. arietans samples collected in South Africa are shown in panel D.

FIG 2

The intra- and interindividual relationships of venom and oral microbiomes in N. nigricollis. Comparison of the oral and venom microbiomes in three N. nigricollis individuals by (A) PCoA, (B) UPGMA tree, and (C) class-level taxonomic profiling following 16S rRNA phylogenetic analysis indicates separate clustering of the microbiota in the two microenvironments. (D) Within-animal incidence comparisons of OTUs suggest (E) unique taxa exist within the oral but also the venom microenvironments. (F) Between-animal comparisons per niche (E1, E2, and oral) indicate most OTUs are shared, but some OTUs are unique to each animal for each site. Dots in panels A and B represent individual N. nigricollis (N. nig) animal data and are colored/labeled by sample type (black, oral; red, envenomation 1 [E1]; blue, envenomation 2 [E2]). Relative taxonomic diversity profiles in panel C are aligned to the UPGMA tree sample labels, with the Shannon-Wiener index (SWI) of each sample indicated. The “venom” histogram in panel E represents the sum OTU fraction found in the two envenomation samples per individual (± standard deviation).

FIG 3

Whole-genome sequencing identifies viable bacteria in N. nigricollis venom as two animal-specific E. faecalis strains. (A) White punctate colonies were recovered in blood agar (upper panels) and MacConkey agar (lower panels) blind cultures of individual oral swab (O) and two consecutive envenomation samples (E1 and E2) obtained from three captivity N. nigricollis snakes. N/D, none detected. (B) Blind multiple-sequence alignment (ClustalO followed by ClustalW phylogeny) of homologous sequences across the de novo assembled genomes against the E. faecalis V583 katA gene (distance to V583 katA indicated in the V583 track) suggests two sequence groups reflecting the history and housing of the host animals. (C) Blind MST construction based on the MLST of the N. nigricollis-derived isolates against nine E. faecalis reference genomes again separates samples into two distinct clusters that reflect the history and housing of the host animals. Partially available allele data are included in this analysis, and instances of allelic differences between nearest neighbors are annotated in white boxes. (D) Blind complete genome MLST against a custom schema generated using three closely related E. faecalis reference genomes clusters these isolates by animal of origin (animal 1, light blue; animal 2, dark blue; animal 3, orange). The host animal color scheme depicted in panel D is also used in panels B and C.

FIG 4

Comparative genomics of mobile and core genomic chromosomal elements of venom-tolerant E. faecalis. (A) Circos coverage plots of the vancomycin resistance-associated V583 plasmids pTEF1, pTEF2, and pTEF3 in the E. faecalis isolates obtained from oral, envenomation 1 (E1), and envenomation 2 (E2) samples from three N. nigricollis individuals reinforce the two sequence type groupings and highlight within-animal variation (green arcs) indicative of sample-specific variation (lack of reads) across E2 samples in animals 1 and 2. The central plot for each plasmid and animal reflects GC content. All data are represented in 50-nt blocks. (B) Blind maximum likelihood tree of the core genomic alignments for the 6 N. nigricollis-derived E. faecalis isolates against the V583 and OG1RF reference strains, with color coding referring to the origin of the isolates: light blue, animal 1; dark blue, animal 2; yellow, animal 3. (C) Maximum likelihood tree from concatenated nucleotide sequence alignment of 865 core genes (381,319 bp) from 734 genomes after removing the sites with gaps. The best-fit GTR+I+G4 substitution model was used with 100,00 ultrafast bootstraps and SH-aLRT tests. The tree was rerooted on the longest branch, and branch lengths of <0.001 were collapsed. The scale bar shows number of nucleotide substitutions per site. Branches in red, blue, purple, and green show group A, group B, and clades containing strains V853 and OG1RF, respectively.