First detection of Wolbachia in the New Zealand biota

Wolbachia is one of the most widespread intracellular bacteria on earth, estimated to infect between 40 and 66% of arthropod species in most ecosystems that have been surveyed. Their significance rests not only in their vast distribution, but also in their ability to modify the reproductive biology of their hosts, which can ultimately affect genetic diversity and speciation of infected populations. Wolbachia has yet to be formally identified in the fauna of New Zealand which has high levels of endemic biodiversity and this represents a gap in our understanding of the global biology of Wolbachia. Using High Throughput Sequencing (HTS) of host DNA in conjunction with traditional molecular techniques we identified six endemic Orthoptera species that were positive for Wolbachia infection. In addition, short-sequence amplification with Wolbachia specific primers applied to New Zealand and introduced invertebrates detected a further 153 individuals positive for Wolbachia. From these short-range DNA amplification products sequence data was obtained for the ftsZ gene region from 86 individuals representing 10 host species. Phylogenetic analysis using the sequences obtained in this study reveals that there are two distinct Wolbachia bacteria lineages in New Zealand hosts belonging to recognised Wolbachia supergroups (A and B). These represent the first described instances of Wolbachia in the New Zealand native fauna, including detection in putative parasitoids of infected Orthoptera suggesting a possible transmission path. Our detection of Wolbachia infections of New Zealand species provides the opportunity to study local transmission of Wolbachia and explore their role in the evolution of New Zealand invertebrates.


Introduction
The bacterium Wolbachia [1,2] is estimated to infect between 40 and 66% of arthropod species worldwide [3][4][5] making it among the most abundant intracellular bacterial genera. Wolbachia is a maternally inherited endosymbiont that can induce a range of host phenotypic responses, including cytoplasmic incompatibility, male death, feminization, and parthenogenesis [6][7][8][9][10]. Wolbachia infections can therefore have long-term evolutionary effects on their host lineages, in addition to immediate reproductive modifications, by providing pathways to rapid reproductive isolation and influencing the evolution of sex-determining mechanisms [6,7,[9][10][11]. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Wolbachia is also being trialled as a biocontrol agent of invasive and disease transmiting insects including medflies [12] and mosquitos as part of the Eliminate Dengue Program [13][14][15]. As Wolbachia can become an obligate parasite of parasitic worms it is also the target of research into antimicrobial drugs by the Anti-Wolbachia Consortium, with the goal of preventing the growth and reproduction of the worms and preventing the diseases they induce [16,17].
Wolbachia prevalence differs among species and among populations of the same species, ranging in infection frequency between 30 and 100% of individuals within a population [18,19]. Infection rates are a complex issue not yet well understood, but they are likely to be dynamic, and involve host dispersal and the nature of the host-parasite relationship. For example it has been suggested that the degree of infection may be a result of Wolbachia acting as a mutualistic secondary symbiont rather than an exclusive reproduction parasite [20][21][22].
The mechanism(s) by which Wolbachia moves between host populations has yet to be confirmed, but is unlikely to be solely via vertical transmission. Genetic similarity of Wolbachia found in parasitoids and parasitoid hosts suggest horizontal transmission [23][24][25]. It has been shown that microinjection of Wolbachia infected cells can facilitate transfer [26], and this indicates how Wolbachia might be transferred by ovipositing parasitoids. Should the parasitoid egg fail to develop, Wolbachia may move in to the host and persist into further generations. Alternatively, the bacterium may be transferred through the digestive system of invertebrates feeding on Wolbachia infected hosts as in some other endoparasites (e.g. Gordian worms). Horizontal transmission via the digestive track has been shown to be effective in whiteflies. Wolbachia was observed to persist in leaves for up to 50 days, which if fed upon by un-infected whiteflies, resulted in Wolbachia infections in the majority of whiteflies [27].
Phylogenetic studies have identified 16 globally distributed supergroups of Wolbachia [10,[28][29][30][31][32]. Incongruence between Wolbachia and host phylogenies suggests many episodes of horizontal transfer resulting in unrelated hosts in the same region sharing similar strains of Wolbachia [25,33]. However, inference of phylogenetic relationships is also complicated by recombination among Wolbachia strains [28,34,35], and host-parasite coevolution [36]. For this reason, a Multilocus Sequence Typing (MLST) system is now widely used as it allows differentiation between even closely related strains of Wolbachia [37].
The New Zealand invertebrate fauna has many distinctive features including high levels of species endemicity [38]. As a large continental island, physically isolated from neighbouring terrestrial ecosystems for many millions of years, the biota has had opportunities to evolve in novel ways and it is frequently posited that the biota has been strongly influenced by their ancient isolation [39,40]. If so, this predicts that distinctive species interactions could have evolved including unique strains of endosymbionts such as Wolbachia. However, to date no Wolbachia infections have been reported from any New Zealand native invertebrate species, reflecting few targeted attempts at their discovery. We tested whether Wolbachia could be detected and if so whether there was evidence of distinctive evolutionary lineages in endemic New Zealand insects.

Methods
Two different approaches were employed to survey potential hosts for Wolbachia infection; bioinformatics and molecular ecology. The first approach made use of bioinformatic tools to search for evidence of Wolbachia 'contamination' in High Throughput Sequencing (HTS) data (reads and assembled contigs) from various insects. These low coverage DNA sequence datasets were produced to infer molecular phylogenies of the invertebrate species using multicopy markers (e.g. [41,42]. The second approach used the MLST primer sets [37] to search for evidence of Wolbachia in a wide range of target templates representing multiple host species and populations.

Mining next generation DNA sequences
As part of a phylogenomic study of endemic New Zealand Orthoptera that have distinctive regional diversity, we carried out HTS of genomic DNA isolated from members of three families; Acrididae, Anostostomatidae, and Rhaphidophoridae (Table 1). The DNA libraries were sequenced on one lane of an Illumina HiSeq2000 by BGI [43]. Approximately 1-4 Gigabytes of 100bp paired end sequencing data was generated for each of the 21-sequenced species.
HTS data was analysed with PAUDA [44], and MEGAN5 [45] in order to identify Wolbachia sequences found within each HTS dataset. MEGAN5 [45] visually displays what organisms are detected in the HTS datasets and indicates the number of sequences associated with each species. If Wolbachia matches were among the 16 most frequently recorded organisms detected at the level genus, the respective invertebrate host sample was treated as positive for Wolbachia infection.
Wolbachia sequences from the positive samples were extracted and mapped against the genome of the Wolbachia endosymbiont of Drosophila melanogaster (accession number NC002978) using Geneious. v. 6 (http://www.geneious.com) [46]. Mapping was performed using the medium sensitivity setting, which equates to a minimum overlap of 25bp, at least 80% overlap identity, with a maximum of 30% mismatches allowed per read. Mapping was iterated five times using the consensus sequence of the reads and repeating the mapping process. This allows more reads to be mapped to variable regions or regions that differ from the reference sequence and reduces the likelihood of reads mapping incorrectly. Once sequences were aligned to the reference genome, coverage at the five core gene loci of the MLST system (ftsZ, coxA, fpbA, hpcA, gatB [37]) was determined, by identifying the conserved PCR primer binding sites on the reference gene. Targeting primer binding sites allowed direct comparison between the HTS sequences and those obtained via PCR. Where there was sufficient HTS coverage the consensus sequence of mapped reads at MLST loci was included in subsequent phylogenetic analysis.

Extraction and amplification of Wolbachia from invertebrate DNA
We focused our host sampling on multiple individuals of the Orthopteran genus Hemiandrus (Family Anostostomatidae) that had yielded positive results in HTS analysis and for which we had suitable material [47]. DNA was extracted from leg or abdomen tissue of 204 individual Hemiandrus ground weta. We used a modified salting-out method incorporating an ice-cold ethanol washing step before addition of room temperature ethanol and allowing the ethanol to evaporate, leaving the DNA to be eluted in 50μl water [48]. Extracted DNA was tested for the presence of Wolbachia DNA using the MLST primer combinations (  [49].
Wolbachia infection rates in species and populations was tested using PCR targeting 5 MLST loci [37], and the variable WSP locus that has potential for distinguishing Wolbachia lineages [11,50]. In addition to the absence of Wolbachia in a sample, several technical issues could explain false negatives where amplification failed. Therefore, we used positive PCR controls with universal insect mitochondria primers LCO1490 and HCO2198 [51] to target host DNA. DNA from Nasonia vitripennis wasps known to be infected with Wolbachia was used to verify the specificity of the MLST primers used in this study.
We also searched for Wolbachia infection in DNA from 45 Rhaphidophoridae (cave weta) from the genera Pachyrhamma (n = 39) and Isoplectron (n = 6). To increase taxonomic range, we screened 40 individuals of 24 other invertebrate species ( Table 2). Sixteen of these were exotic species and eight were New Zealand native or endemic panarthropoda species. We included nine samples of the parasitoid wasp Archaeoteleia because Rhaphidophoridae are their hosts [52], and this parasitic interaction is a potential means of horizontal transmission of Wolbachia. A subset of individuals that produced a DNA fragment for the Wolbachia ftsZ region were sequenced using the forward ftsZ [37] primer (Macrogen Inc., Korea). DNA sequences were checked for quality and aligned to published Wolbachia sequences (Table 3) and our sequences extracted from the HTS samples (Table 1) using Geneious v. 6 [46].
Eighty-six Wolbachia DNA sequences were aligned and trimmed to produce an alignment of 211-438bp of the ftsZ locus. Phylogenetic relationships were inferred using Bayesian phylogenetic analysis (MrBayes; HKY, chain length 1100000, subsampling frequency 200, burn-in length 100000, random 22500) [53,54] (Fig 1). Incorporating a published dataset [28] with a subset of data from the present survey (n = 11) allowed us to determine which supergroup the New Zealand Wolbachia sequences were most similar to (Fig 2). Minimum spanning network [55] (epsilon 0) analysis was performed using PopART [56] (Fig 3).
To determine the spatial distribution of the newly discovered Wolbachia infections in New Zealand, ground weta and cave weta collection locations were mapped using QGIS (QGIS Development Team, 2015). Individual locations were coloured according to whether the  insects collected there were infected with Wolbachia or not (Fig 4). Isolate information was processed through the QGIS dataset to display the distribution of the hosts found carrying each isolate and determine if hosts of differing isolates were likely to be found in sympatry or allopatry.

High throughput DNA sequencing of New Zealand Orthoptera
High throughput DNA sequence data was generated for 21 species of New Zealand Orthoptera. Using the metagenomics tools PAUDA and MEGAN to search for evidence of Wolbachia sequences within the unassembled insect shotgun sequencing data, six species were found to have Wolbachia infections (Table 1). We found similar high levels of Wolbachia DNA sequence in a cave weta (Macropathus sp.) and a ground weta (Hemiandrus brucei). To ascertain the level of genome coverage represented by Wolbachia reads in these two species, the short reads were mapped to the complete genome of the Wolbachia endosymbiont of Drosophila melanogaster. This revealed that the Macropathus sp. reads covered 30% of this reference genome, whilst the reads from Hemiandrus brucei mapped over 33.6% of the Wolbachia reference. Average pairwise nucleotide similarities between the reference Wolbachia genome and reads from Macropathus sp. and Hemiandrus brucei were 90.2% and 92%, respectively. However, due to uneven coverage across the genome only one of the MLST genes (ftsZ) had sufficient sequence information to recover its complete consensus sequence from the HTS data. Four other host species samples that contained significant levels of Wolbachia DNA were all endemic Rhaphidophoridae (cave weta), but read numbers were lower compared to those isolated from the Macropathus sp. and Hemiandrus brucei specimens (<2500 DNA sequences) ( Table 1). Low read number corresponded with lower coverage of the Wolbachia genome (<5%) when mapped against the reference.

Wolbachia infections using specific primers for DNA amplification
To further explore the degree of Wolbachia infection in New Zealand insects we carried out a PCR screen using primers that target Wolbachia MLST or WSP loci [37]. Positive results for Wolbachia infections from six independent Orthoptera lineages were detected through PCR. Wolbachia was detected in a clade of Anostostomatidae ground weta (Hemiandrus brucei, Hemiandrus luna, Hemiandrus maculifrons and Hemiandrus nox) and in the rhaphidophorid genera Pachyrhamma and Isoplectron (Table 2).

Infection rates
Infections rates varied from 30%-93% of individuals per species where > individuals were tested ( Table 2). For Orthoptera species with the largest sample size infection rates were 73% (H. brucei, n = 89), 93% (H. luna, n = 29), and 57% (H. maculifrons, n = 63) positive for amplification using at least one of the MLST primer pairs. Wolbachia was detected in three of four individuals of North Island Hemiandrus nox but absent from the six individuals from the South Island Hemiandrus nox (n = 10) ( Table 2). In addition, of a further 12 Hemiandrus species tested only a single Hemiandrus nitaweta individual (Table 2)  Pachyrhamma (33%) samples tested positive for Wolbachia at one or more of the Wolbachia specific primers ( Table 2).
As a potential vector for Wolbachia horizontal transmission in Rhaphidophoridae and possibly other insects, the parasitoid wasp Archaeoteleia was tested for infection. Of nine specimens available for testing, four individuals were positive for an infection; two A. gilbertae and an individual each of A. onamata and A. kawere (Table 2). Of the further 40 invertebrate individuals representing twenty-four species collected from New Zealand and tested for Wolbachia using MLST primers few were positive. Wolbachia infection was identified in two species; a native tree-living booklouse Ectopsocus sp. and an exotic grass fly Chlorops sp.
Wolbachia infections in New Zealand Orthoptera are geographically widespread (Fig 4). At many locations individuals with Wolbachia infections were collected alongside individuals that were not infected with Wolbachia, suggesting that where Wolbachia is present it is not at saturation. Wolbachia was detected throughout the North Island and northern South Island (Fig 4), however, Wolbachia was detected at a lower frequency in Hemiandrus ground weta from the southern half of South Island, where only two individuals were positive for a Wolbachia infection (Fig 4). The Wolbachia infection rate in Rhaphidophoridae was highest in samples from central and northern North Island although fewer southern samples were examined (Fig 4).
We sequenced the ftsZ region of Wolbachia infections from 82 Orthoptera hosts, three parasitoid wasps (Archaeoteleia sp.) and one booklouse (Ectopsocus sp.; Table 2). Incorporating representatives of all major global Wolbachia supergroups into a phylogenetic analysis with the New Zealand Wolbachia, DNA sequences revealed that the New Zealand diversity nested within supergroups A and B, based on the ftsZ gene (Fig 2). The New Zealand Wolbachia sequences from Macropathus and Pachyrhamma cave weta, Ectopsocus booklouse and some Hemiandrus ground weta fell within supergroup (clade) B, while the Wolbachia sequences from Isoplectron cave weta and other Hemiandrus ground weta fell within supergroup (clade) A. We refer tentatively to New Zealand Wolbachia samples that are part of supergroup A [28] as isolate A as we currently have DNA sequence from one locus. Isolate A samples included 43 sequences from New Zealand Wolbachia, but the New Zealand representatives did not form a monophyletic group within clade A. However, New Zealand clade B sequences differed from all available published (GenBank) Wolbachia (Table 3) (Fig 1) by a minimum of five substitutions (Fig 3) and formed a monophyletic cluster. The closest match was the sister clade consisting of infections from China, India, and Europe [28] and we tentatively referred to as isolate B. Six host individuals had DNA sequences from both isolates, suggesting that they were infected with two different Wolbachia lineages.
Minimum spanning networks of ftsZ for isolate A (360 bp) and isolate B (228 bp) reveal the diversity within New Zealand Wolbachia (Fig 3). Fourteen distinct sequences were identified within isolate A, differing by 1-3 nucleotides. The parasitoid wasp Archaeoteleia was infected with Wolbachia having the same sequence as that obtained from three different New Zealand orthopteran host species. Seven distinct sequences were identified within isolate B Wolbachia, and these differed by a minimum of five mutations from published Wolbachia sequences (2.2%; Fig 3).

Discussion
Wolbachia was detected in HTS DNA sequence datasets from six orthopteran individuals that are endemic to New Zealand, representing two families and five genera (Macropathus sp., Hemiandrus brucei, Talitropsis sedilloti, Miotopus diversus, and two Neonetus specimens). Orthoptera elsewhere in the world are known to be hosts of Wolbachia [57,58], but we present the first documented cases of Wolbachia infection of any endemic New Zealand invertebrate. The samples from cave weta Macropathus sp. (Rhaphidophoridae) and ground weta Hemiandrus brucei (Anostostomatidae) provided~30% coverage of the Wolbachia genome which was the largest in our sample. These sequences were unambiguously identified as part of the Wolbachia global diversity.
The Macropathus sp, and Hemiandrus brucei samples yielded approximately twice the number of total DNA reads compared to the other HTS datasets analysed, however, Wolbachia was also detected in a number of other Rhaphidophoridae (cave weta) samples. The level of detection in four samples was lower (<4%), but Wolbachia represented the majority of prokaryote reads detected in the analysis and DNA sequences were close matches to published Wolbachia (Pairwise % Identity and identical sites of the samples of !85% in M. diverus and both Neonetus specimens). In contrast, the sample from Talitropsis sedilloti had bacterial sequences with less similarity to Wolbachia (pairwise 54% and identical 27.1%) which might represent different bacteria.

Wolbachia and the Hemiandrus maculifrons-complex
The genomic DNA sequence datasets provided evidence of infection from single representatives of five different species (Fig 4). To investigate infection rates, we amplified DNA from numerous individuals of the same species using specific primers, targeting host species within the same genus. Wolbachia was detected in all three species of the ground weta species H. brucei, H. luna, and H. maculifrons through the MLST protocol. Infections were detected in the majority of individuals tested, with 73% of H. brucei, 93% of H. luna, and 57% of H. maculifrons. Hemiandrus brucei and H. luna showed the high-level pattern of infection as suggested by Hilgenboecker, et al. [3]. Their metaanalysis indicates that intraspecific Wolbachia infections rates tend to show a 'most or few' infection pattern, as very high or very low infection frequencies were more likely to occur than intermediate rates. Hemiandrus maculifrons had a lower infection rate, well below the high (>90%) infection level but much higher than in low level (<10%) infections. Within the same host, Wolbachia infections can vary among tissue types, tending to be at higher density in female reproductive tissue. Our weta DNA extractions were mostly from femur muscle and as numbers of intracellular bacteria tend to be limited in somatic tissue this may have reduced detectability in our sample [10,34]. Small host sample sizes also make estimates of infection rates less precise, and this can be rectified by expanded sampling now that the Wolbachia target has been recognised. The same Wolbachia strain can produce various reproductive modifications (pathenogenisis, male killing, cytoplasmic incombatibility) in different host lineages [10], that result in dissimilar infection frequency [7]. In the morphologicaly cryptic Hemiandrus species we studied their genetic similarity suggests it unlikely that Wolbachia has caused different reproductive modifications in each species, but further research will reveal if Wolbachia was a contributing factor in their speciation.

Wolbachia and Rhaphidophoridae
New Zealand has a high diversity of Rhaphidophoridae (cave crickets or cave weta) with at least 19 endemic genera. The orthopteran family is found worldwide and typically cave-dwelling, but several New Zealand species are unusual in that they inhabit forests and the sub-alpine zones. Wolbachia was detected infecting six of these genera; Pachyrhamma, Isoplectron, Neonetus, Talitropsis, Miotopus and Macropathus, with the highest infection rate 33% in Pachyrhamma. As we did not target specific tissue known to have high Wolbachia densities (ovarian follicles) this might underestimate the true infection rate. At least one species of Isoplectron was host to Wolbachia with an infection rate detected of 50%. Further samples will need to be tested to determine the level of infection at both the population and species level. Wolbachia was detected in Macropathus through HTS. Inclusion of this genus in further surveys would be informative.

Transmission
Intracellular bacteria such as Wolbachia are regularly transmitted in egg cells from mother to offspring (vertical transmission [34]). However, Wolbachia is also suspected to be transmitted between species horizontally [25,28,34,59], potentially by an uninfected insect eating an infected one or by multiple species being host to the same parasitoid wasps [60,61]. Archaeoteleia is a genus of parasitoid wasp known for its parasitism of eggs of New Zealand Pachyrhamma cave weta species. However, the typical hosts of two species (A. gilbertae, A. karere) that were positive for Wolbachia is not known. Wolbachia was found in four individuals representing two parasitoid wasp species. The congruence between the Wolbachia infecting weta and the Wolbachia infecting Archaeoteleia may indicate an avenue for further research into a potential interspecies transmission route of Wolbachia in weta. Notably, the Wolbachia DNA sequences from both Archaeoteleia gilbertae and A. karere were identical to Wolbachia sequences from the cave weta Isoplectron (not the Pachyrhamma examined) and two ground weta species (Anostostomatidae: Hemiandrus). The presence of matching Wolbachia in Isoplectron and ground weta rather than Pachyrhamma is interesting because if it is determined that Wolbachia can be transmitted via Archaeoteleia this may be the first indication of new hosts for these parasitoids.
Within the New Zealand insect hosts examined, two distinct clades of Wolbachia were detected. Both isolates of Wolbachia have managed to infect the New Zealand Rhaphidophoridae. The New Zealand Wolbachia lineage that is part of A supergroup clustered with identical Wolbachia DNA sequences from hosts sampled outside of New Zealand (Fig 1). In contrast, other New Zealand Wolbachia ftsZ gene sequences formed a monophyletic group within supergroup B (Fig 1). This might represent a distinct New Zealand lineage of Wolbachia. Further testing of the MLST regions is required because recombination of MLST fragments between strains of Wolbachia occurs. The distribution of 'isolate B' through Hemiandrus sister species was extensive with at least 11 confirmed H. luna hosts and three confirmed H. maculifrons hosts in addition to the 14 confirmed H. brucei hosts. We also detected that some insect hosts were infected with both A and B isolates of Wolbachia.
To our knowledge, this work documents the first cases of Wolbachia infection in endemic New Zealand insects. We detected infection by Wolbachia in endemic species of two families of Orthoptera and in endemic parasitic wasps that attack these Orthoptera. Relatively high observed infections rates, considering our sampling of somatic tissue, in more than one Hemiandrus lineage suggest that Wolbachia is not involved in formation of reproductive barriers between ground weta species, and no definitive pattern of Wolbachia distribution has yet been determined in New Zealand. It was present in all the Hemiandrus species tested spanning both main islands. Further study including analysis of female reproductive tissue will inform on the prevalence of infections across the country and among related species, and reveal what, if any, effect, infections have on reproductive capabilities of the endemic New Zealand insect fauna.