Morphospecies diversity and endosymbiont infection frequencies:
A total of 3509 individual arthropods were collected and sorted into 390 different morphospecies. Out of these 198 morphospecies were exclusively obtained from pitfall traps, 123 from leaf litter sampling and 69 morphospecies were obtained from both. To evaluate whether the sampling method employed yielded a significant proportion of the community diversity, we computed rarefaction analysis with EstimateS. These provided estimates ranging from 858 (±0), obtained from Incidence coverage estimator to 600 (±32.97), obtained through Jack1 (Table S1). This indicates our sampling could capture 45-65% of the possible morphospecies in the community (Figure 1A). This is within expectations when compared to similar studies (Rhoades et al., 2017; Weller & Bossart, 2017).
The taxonomic identification of the 390 morphospecies obtained were primarily done by visual inspection and confirmed by using theirCO1 sequences in BOLD and NCBI database (Table S3). These belonged to seven classes, 24 orders, 118 families and 198 genera of arthropods. We were able to amplify CO1 gene for 314 morphospecies. This was probably due to nucleic acid degradation as they were brought out of storage many times for visual identification, sorting and photography. Most of these samples were of single individuals (190 morphospecies) which prevented DNA extraction from additional samples.
Out of 390 morphospecies screened, approximately 47 (12.05%) morphospecies were found to be infected with Wolbachia . Among these, 38.30% of them belonged to Hymenopterans, 25.53% to Hemiptera, 12.77% to Diptera, 8.51% each to Araneae and Coleoptera, 4.26% to Orthoptera and 2.13% to Sarcoptiformes (Figure 1B). Two morphospecies, morph0081 and morph0085 (both Hymenoptera- Platygastridae) had multipleWolbachia infections and therefore not included for further analysis. There were nine infected morphospecies for which we were unable to amplify all the five MLST genes probably because of the above-mentioned DNA quality issues. We proceeded with 36 unique host-Wolbachia combinations and 34 unique ST’s for which we had amplified all the five MLST genes. When resultant 180 allele profiles were compared with existing sequences in PubMLST database, we found 77 new allelic profiles (14 each for gatB and coxA , 27 forhcpA , 12 for ftsZ and 10 for fbpA ) with 30 new ST’s (Table 1). For the strains labelled ST-N1 and ST-N2, unique ST could not be assigned through PubMLST, as due to DNA quality issues as only one strand of gatB (ST-N1, ST-N2) and ftsZ (ST-N1) could be amplified. As PubMLST requires chromatogram information from both strands, these were manually labelled as ST-N1 and ST-N2.
Phylogenetic analysis of MLST data using ClonalFrame showed 17Wolbachia strains to cluster with known Wolbachiasupergroup A and 15 with B supergroup while two clustered with supergroup F (Figure 2A). Supergroup A infections were predominantly found in Hymenoptera (70.5%) whereas Hemipterans (73.3%) had mostly B supergroup infections. Such taxonomic bias of Wolbachiasupergroup has been noted previously in dipterans (Stahlhut et al., 2010), bees (Gerth, Röthe, & Bleidorn, 2013), ants (Russell et al., 2009) and in lepidopterans (Ilinsky & Kosterin, 2017).
Eleven (2.82%) of the morphospecies had Cardinium infections with four (33%) each from Araneae and Hymenoptera, and one each from Entomobryomorpha, Mesostigmata and Psocodea (Figure 2B). All 11Cardinium strains found in this study clustered with group ACardinium strains (Nakamura et al., 2009). Three morphospecies,i.e . morph0085 (Hymenoptera- Platygastridae), morph0152 (Hymenoptera- Dicroscelio sp.), morph0171 (Hymenoptera-Trichopria sp.) were found to be infected with bothWolbachia and Cardinium . Eight morphospecies (2.05%) hadArsenophonus infections with two each from Hemiptera and Hymenoptera and one each from Diptera, Entomobryomorpha, Psocodea and Thysanoptera (Figure 2C). Two morphospecies, morph0294 (Hymenoptera- Platygastridae) and morph0329 (Hemiptera- Balclutha ) were found to be infected with both Wolbachia and Arsenophonus.Morph0085 had multiple Wolbachia as well as also hadCardinium infections, whereas Morph0328 (Psocodea-Embidopsocus ) had both Cardinium and Arsenophonus . Therefore, multiple endosymbionts can use the same host to spread across different communities (Russell et al., 2012; Zhao, Chen, Ge, Gotoh, & Hong, 2013).
Horizontal Transfer of endosymbiont strains:
To reveal the extent of horizontal transfer events of endosymbionts across their host taxa, a qualitative assessment of phylogenetic congruency was done with host and their corresponding bacterial infections. As figure 3 reveals there is extensive horizontal transfer of the endosymbionts within the soil arthropods. A Mantel test (r ) of correlation between pairwise distance of host and their corresponding endosymbiont also showed no significant correlations (Figure S5).
If endosymbionts are first moving around the host taxa of this particular community then very similar bacterial strains would be found in taxonomically distant soil arthropods. This is precisely what we found with two distinct Wolbachia strains. ST-541 and ST-559 were each found in two distinct taxonomically unrelated hosts (Table 1). Morph0001 (Orthoptera- Neonemobius ) and morph0098 (Hemiptera-Phorodon ) were found to be infected with Wolbachia ST-541, whereas ST-559 was found in both morph0213 (Hemiptera-Heteropsylla ) as well as morph0220 (Hemiptera- Delphacidae). Again, the possibility remains that these transfers could have happened independently and not correlated with the hosts being members of a particular community. But this assumes a non-parsimonious explanation that two independent events would converge on the transfer of the sameWolbachia ST in two different hosts.
As these bacteria are transferred around to different hosts, they are also coming in contact with each other. Whether this leads to stable multiple infections is not known, but this obviously creates opportunities for genetic exchange where the two interacting bacteria are now in a single host cytoplasm. Moreover, such co-infections can trigger selection whereby only a single endosymbiont can remain within a host. Such flux seems to be a key feature of endosymbiont dynamics, especially with Wolbachia , where loss is 1.5 times higher than acquisition of new infections (Bailly-Bechet et al., 2017).
Thus, this phenomenon of horizontal transfer should also create another opportunity where endosymbionts can potentially undergo recombination with each other since they are now in the same host cytoplasm.