Investigating the Diversity of Wolbachia across the Spiny Ants ( Polyrhachis )

: Among insects, Wolbachia is an exceedingly common bacterial endosymbiont with a range of consequences of infection. Despite the frequency of Wolbachia infection, very little is known about this bacteria’s diversity and role within hosts, especially within ant hosts. In this study, we analyze the occurrence and diversity of Wolbachia across the spiny ants ( Polyrhachis ), a large and geographically diverse genus. Polyrhachis samples from throughout the host genus’ phylogenetic and biogeographical range were ﬁrst screened for single infections of Wolbachia using the wsp gene and Sanger sequencing. The multilocus sequence typing (MLST) scheme was then used on these singly infected samples to identify the Wolbachia strains. A Wolbachia phylogeny was inferred from the Polyrhachis samples analyzed in this study as well as other Formicidae MLST proﬁles from the MLST online database. We hypothesized that three key host factors were impacting Wolbachia diversity within the Polyrhachis genus: biogeography, phylogeny, and species level. The results suggest that the phylogeny and biogeography of Polyrhachis hosts have no impact on Wolbachia diversity; however, species level may have some limited inﬂuence. Additionally, Wolbachia strains appear to group according to being either Old World or New World strains. Among the taxa able to form complete MLST allelic proﬁles, all twenty are seemingly new strains


Introduction
Ants (Formicidae) are a highly diverse family of insects with a global distribution.One of the many factors contributing to the overwhelming ecological success of ants is their many associations with symbiotic microbes.Ants associate with microbial eukaryotes, fungi, viruses, and bacteria; further, many of these associations are understood to have contributed to the diversity of diets, occupied niches, and life history in various ant groups [1].For example, the ant genus Cephalotes is able to survive on a nutrient-poor herbivorous diet due to the microbial symbionts present in its gut [2].In addition, the functions of these symbionts are of particular interest to researchers, especially in the case of maternally transmitted bacterial symbionts already known to alter host reproduction, development, nutrition, and defense in many arthropods [3].The Wolbachia bacterial genus is a wellknown example of such a symbiont.It is estimated that Wolbachia infects up to 75% of all insect species [4] and is an incredibly common, heritable maternally transmitted bacterial symbiont of ants [3,5].Some of the most notable consequences of Wolbachia infection in insects are alterations to the host's reproductive abilities-these include parthenogenesis, male killing, male feminization, and cytoplasmic incompatibility [6].Within ants (Formicidae) specifically, Wolbachia has also been found to accelerate the colony life cycle [7] and enhance the host's nutrient uptake [8].Due to the variety of Wolbachia's impacts on its ant hosts, more studies are needed to elucidate the diversity of Wolbachia across Formicidae to understand the consequences of its associations with ants [9,10].
While it is now known that Wolbachia is a widespread symbiont of insects, it was first discovered as a rickettsial symbiont of the mosquito Culex pipiens in the 1920s [11].All Wolbachia strains are divided into supergroups via phylogenetic analysis using one or multiple marker genes (e.g., 16S rDNA, wsp, ftsZ).Currently, there are twenty-one Wolbachia supergroups, ranging from A to U [11][12][13].Further, these Wolbachia supergroups also appear to have set associations to specific host taxa.For instance, it has been found that the strains in Formicidae hosts are from mostly supergroups A and F with the majority being from supergroup A [9], though there has been a single instance where a supergroup B strain was found associated to an ant host from Mexico, Pheidole sciophila [14].
Previously, the standard procedure for sequence typing Wolbachia strains was based upon sequencing the Wolbachia surface protein gene, wsp [15].After it was determined that wsp experiences extensive recombination via swapping of conserved amino acid motifs within hyper-variable regions [16], the Multilocus Sequence Typing (MLST) approach was proposed [6].MLST was introduced alongside an online database of bacterial and host information (https://pubmlst.org/organisms/wolbachia-spp,accessed on 27 February 2023), and Wolbachia sequence types are based upon the allele determination of five different housekeeping genes (coxA, fbpA, ftsZ, gatB, and hcpA) rather than wsp [4].MLST has become the standard method of sequence typing Wolbachia since it provides a more robust approach to assessing Wolbachia diversity across a variety of host taxa due to the reliance on five loci rather than the wsp locus alone.The MLST scheme has been used by researchers studying this bacterial genus within a wide range of hosts including filarial nematodes and ticks [17], butterflies [18] and ants [1,[19][20][21].
Polyrhachis Smith, 1857 is a large ant genus (over 700 species) that inhabits Africa, Asia, Australia, and Oceania [22,23].They are commonly called "spiny ants" due to the spinescence of most species, which can vary in shape, length, and numbers; this spinescence is hypothesized to be a defense characteristic against vertebrate and invertebrate predators [24,25].They exhibit a large variety of nesting techniques including the use of larval silk to weave their nests (a trait limited to few ant genera), nesting inside hollow bamboo, and attaching nests to stones [26,27].In addition, Polyrhachis belongs to the Camponotini tribe, which is well known for their symbiotic relationships with bacteria-in particular Blochmannia [28]-and an association with Wolbachia has been previously found in Camponotini as well [20,29,30].Polyrhachis's broad biogeographical range spanning across Africa, Asia, Australia, and Oceania [26] makes it a useful host for studying the impacts of host biogeography on Wolbachia diversity.
In past studies, specific ant species have been studied for their associations to Wolbachia [19,23], and the evolutionary association of Wolbachia was evaluated across the entire Formicidae family [9].Additionally, Wolbachia has been studied in other social insects such as bees, termites, and wasps [10].In one species-specific study, the diversity of Wolbachia was analyzed across the geographically diverse giant turtle ant species (Cephalotes atratus), and results suggested that Wolbachia diversity is affected by geography [19].In a broader study that analyzed Wolbachia across Formicidae, the evolutionary origins of Wolbachia infection in ants were illuminated and the biogeographical origin of the symbiosis was inferred to be in Asia [9].Our intention with this work, investigating Wolbachia infection across the Polyrhachis genus, is to further explore the notion that Wolbachia diversity can be impacted by geography, as well as the evolutionary association between host and microbe via phylogenetic and species level analyses.
In the following study, our primary objective was to analyze and observe patterns of Wolbachia infection in Polyrhachis.We hypothesized that three factors related to the Polyrhachis host will impact the observed diversity of Wolbachia: phylogeny, species level, and biogeography.If these factors do impact Wolbachia diversity, we would anticipate seeing significant correlations in increases (or decreases) of Polyrhachis host diversity with that of its Wolbachia symbionts.If, for example, the Polyrhachis host phylogeny impacts the observed diversity of Wolbachia, we will see phylogenetic signal and potential evolutionary co-diversification between Polyrhachis and Wolbachia.If species level within Polyrhachis impacts the observed diversity of Wolbachia, we may observe different kinds of Wolbachia infecting different Polyrhachis species in statistically significant ways.If biogeography of the Polyrhachis host impacts the observed diversity of its Wolbachia, we may see a variance in Wolbachia infection that is correlated with the different locations where each Polyrhachis sample was collected.

Materials and Methods
Samples from 102 Polyrhachis species (237 Polyrhachis samples) were screened for their associated Wolbachia strains.These samples were collected from 29 countries (Table 1 shows samples positive for Wolbachia; all samples are listed in Supplementary Material File S1).
The DNA extractions was performed on whole ant specimens following the DNeasy Blood and Tissue (Qiagen) protocol.The DNA was stored at −20 • C. The sampled Polyrhachis species were taken to be representative of the entire host genus and spanned across the entire Polyrhachis biogeographical range.To screen for Wolbachia and determine which samples contained single infections, sequencing of Wolbachia's wsp gene was performed.The wsp gene was PCR amplified using Taq DNA Polymerase, primers wsp81f and wsp69r (at 1 µM each), and 1 µL of DNA [17,31] for 36 cycles with an annealing temperature of 59 • C [4].The thermocycler program was set to the following: the cycle began with denaturation at 94 • C for 30 s, annealing for 45 s, 72 • C for 1.5 min, an elongation step at 70 • C for 10 min, and a hold at 4 • C. Annealing temperatures varied by gene: coxA was annealed at 55 • C, fbpA at 59 • C, hcpA at 53 • C, and both ftsZ and gatB at 54 • C. The PCR products were first evaluated using gel electrophoresis [32] wherein the presence of a band indicated the infection of at least one Wolbachia strain for that sample.Wolbachia-positive PCR products were purified using ExoSap (Cleveland, OH, USA) with the manufacturerrecommended thermocycler settings.BigDye Terminator (Applied Biosystems, Waltham, MA, USA) was used to prepare the samples for Sanger sequencing, which was carried out by the Cornell Institute of Biotechnology (Ithaca, NY, USA).The resulting sequence electropherograms were evaluated in Geneious Prime 2022.1 (https://www.geneious.com,accessed on 15 August 2022) to determine whether samples were infected with single or multiple strains of Wolbachia.Only singly infected Polyrhachis samples (n = 34) were subjected to the Multilocus Sequence Typing (MLST) process wherein the five MLST genes (coxA, fbpA, ftsZ, gatB, and hcpA) were amplified and sequenced according to the same procedure as was done for the wsp gene.Since Sanger sequencing of multiple strains at once creates indecipherable electropherograms (due to the sequences for each strain overlaying each other), multi-infected Polyrhachis samples were excluded from the MLST process.Sequence alignments for each locus were created in Geneious Prime 2022.1 (https://www.geneious.com,accessed on 15 August 2022) then checked against reference sequences in the MLST online database (https://pubmlst.org/organisms/wolbachia-spp,accessed on 15 August 2022) to determine closest matching allele types.The closest matching sequence type (ST) for each Wolbachia strain able to produce clear electropherograms for all five loci (n = 20) was determined based upon these five alleles.
The five MLST genes were concatenated (2098 bp total length, order: coxA, fbpA, ftsZ, gatB, hcpA) for each of the 20 remaining samples, then added to a pool of 70 MLST database sequences from other Formicidae-associated Wolbachia strains [33].A Wolbachia phylogeny was inferred with these 90 MLST sequences via the IQ-Tree web server 1.6.12[34] to infer a phylogenetic tree by maximum likelihood and generate bootstrap values.The best fit model of substitution for each locus was determined by the ModelFinder [35] and partition model [36] features available through IQ-Tree web server.Partitions and their best-fit models are shown in Table 2. Wolbachia strains for ST124 and ST557 (both supergroup F) from the host species Ocymyrmex picardi and Paratrechina, respectively, formed the outgroup of the phylogeny.The haplotype network for each MLST gene was constructed with Network 4.5.1.0[37] using the median joining parameter.Two mantel tests were performed using the R package vegan [38].The first of these tested the correlation between phylogenetic distance between Wolbachia strains and geographical distances between latitudes and longitudes of collection sites.The second test examined correlations between phylogenetic distances between Polyrhachis host species (Blanchard and Moreau, in press) and phylogenetic distances of Wolbachia strains.Both the Wolbachia and Polyrhachis phylogenies were pruned down to seven tips, representing ant host or Wolbachia from seven different Polyrhachis species: P. bihamata, P. cephalotes, P. carbonaria, P. thrinax, P. shixigensis, P. illaudata, and P. hexacantha.

Results
Of the initial 237 Polyrhachis samples screened for Wolbachia using the wsp gene, 112 (47%) tested positive (Table 3).Positive samples represent 69 of the 102 tested Polyrhachis species.To test the hypothesis of host phylogeny influence on Wolbachia diversity, in the subsequent analyses we kept only the Wolbachia-positive samples of host species present in the Polyrhachis phylogeny generated by Mezger and Moreau [24]; this reduced the sample size to 73 samples.There were 43 different Polyrhachis species across the 73 samples.After analyzing the electropherograms to evaluate if the positive samples were single or multiple infections of Wolbachia, 34 of the 73 samples (47%) were determined to be single infections.Single and multiple infections of Wolbachia in Polyrhachis occurred in the same six biogeographical regions (Figure 1A); the singly infected samples were collected from 15 different countries (Figure 1B).Of the 34 singly infected samples, there were 21 different Polyrhachis species represented (Figure 1C).Table 3.The allele, ST, and Polyrhachis host information for the 34 Polyrhachis samples.Allele numbers are included for each MLST gene that was able to be sequenced; blue-shaded alleles are close matches i.e., new allele variants for those loci.A dash (-) in an MLST gene column indicates that no sequence was able to be produced and thus no allele determination was made, and a dash in the ST column indicates a sample that was unable to be assigned to a sequence type.An asterisk (*) indicates the closest matching ST.The strains from samples CSM2738, MJ9280, MJ9287, MS1177, and SUL02 all had multiple "close matching" STs according to the MLST database, indicated by two asterisks (**).The country of origin and host species are shown in the last two columns.Twenty of the singly infected samples were able to produce viable sequences for all five MLST loci.Table 3 shows the allele and ST determinations, as well as host information, for those 20 samples.Further, Figure 2 illustrates nucleotide differences in the form of a haplotype network.Loci with no exact matches to sequences in the MLST database were considered to have new allele variants-every strain identified had at least two loci with new variants.The 14 singly infected samples that were unable to produce complete MLST alignments each had at least one locus with indeterminable Sanger resultstwo samples, KATE02 and MJ8291 (from Polyrhachis shistacae in South Africa and Polyrhachis sp. in Papua New Guinea, respectively) were unable to produce sequences for any of the five loci (Table 3).
The coxA and gatB loci were seemingly the most stable MLST loci for Polyrhachisassociated Wolbachia strains.Of the five loci, they had the most samples with exact matches to allele variants currently registered in the MLST database, with only five possible new allele variants found at both loci.The ftsZ and hcpA loci presented a greater number of new allele variants than either the coxA or gatB loci: seven and eight new allele variants were found at the ftsZ and hcpA loci, respectively.The fbpA locus presented the most genetic change of all five loci when compared to references in the Wolbachia MLST database-17 of the 20 samples with complete MLST profiles presented new allele variants, each of which appear to be unique.Additionally, 13 samples with incomplete MLST allelic profiles produced indeterminable sequences for the fbpA locus, and for 10 of these samples, fbpA was the only locus unable to be properly sequenced (Table 3).Ultimately, due to each strain having at least one new MLST allele variant, it appears that all 20 samples present new Wolbachia STs not yet seen in the MLST database.Twenty of the singly infected samples were able to produce viable sequences for all five MLST loci.Table 3 shows the allele and ST determinations, as well as host information, for those 20 samples.Further, Figure 2 illustrates nucleotide differences in the form of a haplotype network.Loci with no exact matches to sequences in the MLST database were considered to have new allele variants-every strain identified had at least two loci with new variants.The 14 singly infected samples that were unable to produce complete MLST alignments each had at least one locus with indeterminable Sanger results-two samples, KATE02 and MJ8291 (from Polyrhachis shistacae in South Africa and Polyrhachis sp. in Papua New Guinea, respectively) were unable to produce sequences for any of the five loci (Table 3).
The coxA and gatB loci were seemingly the most stable MLST loci for Polyrhachisassociated Wolbachia strains.Of the five loci, they had the most samples with exact matches to allele variants currently registered in the MLST database, with only five possible new allele variants found at both loci.The ftsZ and hcpA loci presented a greater number of new allele variants than either the coxA or gatB loci: seven and eight new allele variants were found at the ftsZ and hcpA loci, respectively.The fbpA locus presented the most genetic change of all five loci when compared to references in the Wolbachia MLST database-17 of the 20 samples with complete MLST profiles presented new allele variants, each of which appear to be unique.Additionally, 13 samples with incomplete MLST allelic profiles produced indeterminable sequences for the fbpA locus, and for 10 of these samples, fbpA was the only locus unable to be properly sequenced (Table 3).Ultimately, due to each strain having at least one new MLST allele variant, it appears that all 20 samples present new Wolbachia STs not yet seen in the MLST database.The phylogeny inferred with 90 Wolbachia MLST sequences (20 are the Polyrhachisassociated from this study, 70 are other Formicidae-associated strains from the MLST database) is shown in Figure 3. Bootstrap values ≤ 70% were hidden.No samples exhibited close relationships to any Wolbachia strains from the outgroup, supergroup F. Thus, all Wolbachia found in Polyrhachis belong the Supergroup A. The Polyrhachis strains from this study were organized into 13 genotypes, seven of which contain only one Polyrhachis-associated strain (either independently or with another Formicidae-associated strain).Only two of these genotypes contain samples from the same country of origin: P. (Myrmhopla) sp. and P. (Polyrhachis) sp.from Cambodia, and the two P. illaudata samples from Laos.Additionally, the samples from Laos are also the only grouping which contains Wolbachia strains from the same host species (Polyrhachis illaudata).In addition, all Polyrhachis-associated Wolbachia strains grouped with other Polyrhachis-associated strains, which suggests that there is a specificity of Wolbachia for Polyrhachis species.Six biogeographical ranges are represented in the phylogeny by the Polyrhachis-associated strains and all Polyrhachis-associated strains were grouped together with other Old World samples.Additionally, distinct clades formed to separate Wolbachia into Old World and New World groupings.The clades "a" and "c" contain several samples from the same biogeographical region-the Old World.Clade "b" are mixed, however contain two subclades: "b1" with samples from the Old World, and "b2" with samples from the New World (Figure 3).Clades "a," and "c" are the Old World clades that formed within the supergroup A taxa.Within clade "b," "b1" represents Old World taxa while "b2" represents the New World taxa that seemingly evolved from Old World taxa.The red-highlighted taxon (ST111_Odontomachus_clarus_USA) is the only strain to not group according to being from an Old World or New World sample.
Results of the Mantel tests indicated no correlation between both Wolbachia phylogenetic distance and geographic distance (Mantel statistic r: −0.030; p-value: 0.531) and Polyrhachis phylogenetic distance and Wolbachia phylogenetic distance (Mantel statistic r: 0.117; p-value: 0.302).

Discussion
By using such a large and biogeographically diverse host genus like Polyrhachis, we were able to study whether host geography, phylogeny, and species level have any observed impact on Wolbachia diversity.Although 12 of the strains with complete allelic profiles best matched to ST61, they are all seemingly unique since their ST determination is based upon apparently new allele variants at multiple loci.For instance, RO122 (Polyrhachis [subgen.Myrma] sp. from Tanzania) was best matched to ST61 while having a possible variant at the ftsZ locus, but CSM1854 (Polyrhachis cephalotes from Malaysia) was also best matched to ST61 while having possible new variants at the coxA, ftsZ, and hcpA loci (Table 3).Indeed, these two samples were divided into their own clades in the Wolbachia phylogeny and there appears to be no tendencies for other strains with the same best matching STs to be grouped into clades.Therefore, our results suggests that each strain is a new ST (for a total of 20 new Wolbachia strains being found across the Polyrhachis genus), implying that across Polyrhachis there is an incredible diversity of Wolbachia.
The inferred Wolbachia phylogeny indicates that all strains identified in Polyrhachis aare from supergroup A, since there were no ant samples from this study that nested within the outgroup clade.Since the 20 strains included in the phylogeny span across the entire Polyrhachis geographic range, this phylogeny also suggests that the Wolbachia found within this host genus will likely belong to supergroup A, independent of the host's geographic range.
Some studies have seen that Wolbachia strains may group according to being Old World or New World [7,16], and it appears that the inferred phylogeny follows this trend as well.The blue boxes in Figure 3 represent Old World clades ("a," "b," and "c") that formed among the Supergroup A taxa-taxa not included in these boxes are strains from New World samplings.Clade "b" was further divided into clades "b1" and "b2"-"b1" being Old World taxa that seemingly evolved from Old World taxa, and "b2" being New World taxa that evolved from Old World taxa.Both taxa within the supergroup F outgroup are from Old World hosts.The only taxon that did not group according to the New World and Old-World clades is an ST111 strain from another study from an Odontomachus clarus host in the United States (highlighted red in Figure 3).This New World taxon grouped most closely into Old World clade "b1" and closest to Polyrhachis clade six.All other strains sourced from New World hosts formed exclusive New World clades.To understand why this O. clarus strain best fit into an Old World clade-and close to a Polyrhachis clade-rather than with other New World samples, more sampling of Wolbachia from that host genus would be necessary.Clades "a" and "b" appear to share a more recent common ancestor than they do with clade "c."Interestingly, all Polyrhachis-associated strains fell into the more closely related "a" and "b" clades; however, the majority were grouped into clade "b" with only clades one and two being part of "a."Ultimately, all but one taxon grouped according to being New or Old World, but it was not a perfect split-grouping since there were multiple clades of either type.Regardless, this still supports the trend seen in previous studies of Wolbachia [7,16] wherein strains will form clades according to Old or New World geography.
Among the 70 database MLST profiles used to infer the Wolbachia phylogeny, there was one strain also sourced from a Polyrhachis host (ST51_Polyrhachis_vindex_Philippines).This strain showed close relation to the strain from sample SKY24 (Polyrhachis sp.from Singapore), and together they form a distinct clade (clade one, Figure 3).However, since these samples are sourced from different hosts and different countries, this clade suggests that Wolbachia diversity is not significantly impacted by host species level or biogeography.Rather, this clade (as well as the other 12 clades) suggests that strains are likely to be more closely related if they are from the same host genus since no clades formed with strains sourced from different host genera.Although they did not form a single, unified clade, the fact that all 13 clades contain exclusively Polyrhachis-associated strains suggests that the host's genus has some degree of influence on the associated Wolbachia diversity.
Of the 13 clades that the Polyrhachis-associated Wolbachia strains formed within the phylogeny, two had biogeographical consistency across the clade-clade 12 with two samples from Laos and clade seven with two samples from Cambodia-with both clades being from the Indomalayan biogeographical range (Figure 3).Interestingly, clade 12 contains the only two representative samples for the host species Polyrhachis illaudata (RA1163, RA1157), but based on their allelic profiles from Table 3 they are perhaps more likely to be closely related STs rather than the exact same STs.Although both strains were flagged as having possible new allele variants at the fbpA and ftsZ loci (and RA1163 with an additional variant at the coxA locus), they are not flagged for the same nucleotide modifications at either loci.Further verification of the genetic alterations that indicate these loci as having new allele variants would need to be conducted in order to distinguish these samples as different STs.
Polyrhachis illaudata was also the only host species able to have multiple samples with complete MLST allelic profiles sequenced.Both P. illaudata-associated strains formed a single clade (clade 12, Figure 3), suggesting that Wolbachia strains from the same host species will be more related than strains sampled from different host species.If all sampled Polyrhachis hosts receive expanded sampling across multiple colonies, it will be possible to determine whether this trend is true to other Polyrhachis hosts beyond P. illaudata.Thus, current results suggest that the species level of Polyrhachis hosts potentially impacts the observed Wolbachia diversity within this host genus.
The samples sourced from Cambodian Polyrhachis hosts present an interesting case.There was a third sample from Cambodia, AS4121 (Polyrhachis [subgen.Myrmhopla] sp.), not included in clade seven (Figure 3) with the other two Cambodian samples, AS4148a (Polyrhachis [subgen.Myrmhopla] sp.) and AS4132b (Polyrhachis [subgen.Polyrhachis] sp.)-this is seemingly because AS4148a and AS4132b both have the same new allele variant at the gatB locus whereas AS4121 has an already documented gatB allele variant (Table 3).Yet the host of AS4121 is more closely related to the host of AS4148a since they both belong to the subgenus Myrmhopla, while the host of AS4132 is subgenus Polyrhachis [24].This instance suggests, then, that neither geography nor host phylogeny impacts the association of Wolbachia strains since more closely related hosts do not share similar Wolbachia strains and strains with hosts from the same country and geographical region do not appear in the same clade.Indeed, the Mantel test results support this since there was no correlation found between Wolbachia phylogenetic distance and geographical distance or between Wolbachia phylogenetic distance and Polyrhachis phylogenetic distance.However, this does not necessarily exclude the possibility that host species level could still be an impactor on Wolbachia diversity as seen in the P. illaudata clade that formed (clade 12, Figure 3).
Overall, the trends seen among the samples from Cambodia appear across the phylogenythere is no consistent grouping of Wolbachia strains according to how related their host species are.For example, SUL02 and RO122 are both from subgenus Myrma of Polyrhachis, and MJ9280 and MJ9243 are both from subgenus Myrmhopla.Yet, in both cases, the two taxa are distantly related into two separate clades (SUL02 clade six, RO122 clade 13; MJ9280 clade 5, MJ9243 clade nine).From the perspective of host geography, there is rarely consistency for samples sourced from the same region to have more closely related strains.The three samples from Sundaland-GM3589b (Polyrhachis [subgen.Myrmothrinax] sp. from Malaysia), CSM1854 (Polyrhachis cephalotes from Malaysia), and SKY24 (Polyrhachis sp.from Singapore)-have perhaps the most distinct case of exhibiting that host geography may have no impact on the strain similarity of associated Wolbachia.Despite two of the three samples being from the same country, the three samples are split into three distant clades (6, 10, and one, respectively) in the phylogeny which, again, suggests that the geography of Polyrhachis hosts is not structuring Wolbachia diversity.Previous studies across genera in butterflies [18] and termites [39] similarly concluded that host geography did not impact which Wolbachia strains would be associated to the host.The study in termites also found that distantly related host species could have more closely related Wolbachia strains [39] as seen in this study, thereby supporting the notion that the phylogeny of Polyrhachis hosts also has no strong impact on Wolbachia associations.In contrast, these results may contradict the results of Kelley et al. [19], which found that the association of Wolbachia to Cephalotes atratus was impacted by host biogeography.Yet, this may not be a true contradiction if it can be confirmed that across a single Polyrhachis species, host biogeography impacts Wolbachia diversity (which is seemingly seen in the P. illaudata clade [clade 12, Figure 3]) since the study by Kelley et al. [19] took place in a single host species.
For the third host factor (host species level), some results suggest that it has an impact on Wolbachia diversity.However, as discussed with the case of P. illaudata, expanded sampling of each Polyrhachis species is required to verify the observed trends.In the initial sample pooling, there were multiple instances where the same host species was sampled from several colonies.However, once removing samples containing multiple strains of Wolbachia the sample pool was reduced by over 50% and many of these multi-colony samplings were lost.These samples were removed because multiple strains in one sample cannot be parsed into individual strains.
We found that Polyrhachis-associated Wolbachia strains will form exclusive clades distinct from strains of other host genera.In other words, Polyrhachis-sourced strains of Wolbachia will only form clades with other Polyrhachis-associated strains.It was also found that samples of the same host species were sometimes grouped into the same clade.This suggests that there is some level of restructuring occurring at the hosts' species level.Beyond the Polyrhachis genus, there also appears to be separation of Wolbachia strains based upon being either Old World or New World, wherein taxa from the Old World will not typically be grouped into a closely related clade with New World taxa and vice versa.
Ultimately, the results of this study suggest that host biogeography and phylogeny do not have any significant impact on which strains of Wolbachia will be associated to the Polyrhachis host species, though our findings suggest that the Polyrhachis species level may have some effects on Wolbachia strain.Further work on the impact of geography of Wolbachia infection would benefit from incorporation of more data on the host's current range and historical biogeography, which were not included in this study.Additionally, horizontal transfer of Wolbachia between hosts is not common, but has been observed, primarily in related hosts [7].Horizontal transfer events may affect the results of phylogenetic analyses, particularly in comparisons of the host's and Wolbachia phylogenies.
Our findings from this study, particularly our observation that some Wolbachia strains may be associated with particular Polyrhachis species, highlight the impacts that microbial diversity can have on ant diversity, and vice versa.The presence of vertically-transmitted symbionts like Wolbachia suggests the possibility of a microbial impact on evolution; coevolution of ants and microbes over long time-scales has already been observed in some ant genera, in some cases allowing the ants to pursue diets or occupy niches not previously available to them.While our findings about Polyrhachis help to elucidate more of the ways that symbionts can impact ant diversity, still, little is known about the microbial partners of the majority of ant genera.Studies of this nature are crucial in understanding the many factors that contribute to present-day ant diversity and may provide insights into the ways that the associates of ants may shape the evolutionary future of their hosts.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/d15030348/s1,File S1: Sample ID, host species, and country of origin for all for each of the initial 237 samples screened for Wolbachia.The presence of Wolbachia is indicated by the wsp column ("+" indicates positive for Wolbachia, "−" indicates negative for Wolbachia).

Figure 1 .
Figure 1.(A) Biogeographic distribution of single-strain Wolbachia infections and multi-strain Wolbachia infections within Polyrhachis hosts.The "Sundaland" and "Wallacean" groups are included here as separate categories to better distinguish their geography from the more northern parts of the Indomalayan realm.(B) Country distribution of the singly infected Polyrhachis samples.(C) Polyrhachis species distribution of the 34 singly infected samples.There were 21 different host species represented in this sample pool.

Figure 1 .
Figure 1.(A) Biogeographic distribution of single-strain Wolbachia infections and multi-strain Wolbachia infections within Polyrhachis hosts.The "Sundaland" and "Wallacean" groups are included here as separate categories to better distinguish their geography from the more northern parts of the Indomalayan realm.(B) Country distribution of the singly infected Polyrhachis samples.(C) Polyrhachis species distribution of the 34 singly infected samples.There were 21 different host species represented in this sample pool.

Figure 2 .
Figure 2. Haplotype network figures for all samples able to be assigned a sequence type.The haplotype size represents the frequency found, and black bars between haplotypes represent the numbers of nucleotide differences between haplotypes.Red dots (labelled "mv" and numbered) were added by the program as a hypothetical haplotype.

Figure 2 .
Figure 2. Haplotype network figures for all samples able to be assigned a sequence type.The haplotype size represents the frequency found, and black bars between haplotypes represent the numbers of nucleotide differences between haplotypes.Red dots (labelled "mv" and numbered) were added by the program as a hypothetical haplotype.

Figure 3
Figure 3 Wolbachia phylogeny.Samples are named here with the following convention: "Sequence-Type_Genus_species_Country".White diamonds indicate bootstrap values ≥ 70%.The bolded taxa are the 20 Wolbachia strains from this study; each of the 13 genotypes they formed are numbered and highlighted in light teal.All 20 taxa belong to the supergroup A clade.The supergroup F clade is the outgroup.Stars next to the taxa within each sample indicate biogeography: Afrotropical (n = 2, blue), Australian/Oceania (n = 4, red), Palearctic (n = 2, green), Indomalayan (n = 8, purple), Sundaland (n = 3, yellow), Wallacean (n = 1, orange).Only two of the 13 genotypes formed by the Polyrhachis-associated strains contain multiple samples from the same biogeographical region (seven and 12).Clades "a," and "c" are the Old World clades that formed within the supergroup A taxa.

Figure 3 .
Figure 3. Wolbachia phylogeny.Samples are named here with the following convention: "Sequence-Type_Genus_species_Country".White diamonds indicate bootstrap values ≥ 70%.The bolded taxa are the 20 Wolbachia strains from this study; each of the 13 genotypes they formed are numbered and highlighted in light teal.All 20 taxa belong to the supergroup A clade.The supergroup F clade is the outgroup.Stars next to the taxa within each sample indicate biogeography: Afrotropical (n = 2, blue), Australian/Oceania (n = 4, red), Palearctic (n = 2, green), Indomalayan (n = 8, purple), Sundaland (n = 3, yellow), Wallacean (n = 1, orange).Only two of the 13 genotypes formed by the Polyrhachisassociated strains contain multiple samples from the same biogeographical region (seven and 12).Clades "a," and "c" are the Old World clades that formed within the supergroup A taxa.Within clade "b," "b1" represents Old World taxa while "b2" represents the New World taxa that seemingly evolved from Old World taxa.The red-highlighted taxon (ST111_Odontomachus_clarus_USA) is the only strain to not group according to being from an Old World or New World sample.

Table 1 .
Sample ID, host species, and country of origin for all samples positive for Wolbachia.A complete list of samples screened is available in Supplementary Material File S1.

Table 2 .
Partitioning and best-fit models for each partition as determined by the IQ-Tree web server ModelFinder.The third and fourth columns detail the length of each partition and its position within the concatenated sequence (with a total length of 2098 bp).

Table 3 .
Partitioning and best-fit models for each partition as determined by the IQ-Tree web server ModelFinder.The third and fourth columns detail the length of each partition and its position within the concatenated sequence (with a total length of 2098 bp).