Molecular characterization of 'Candidatus Borrelia tachyglossi' (family Spirochaetaceae) in echidna ticks, Bothriocroton concolor.

Recently, a novel species of the genus Borreliawas identified in Bothriocroton concolor and Ixodes holocyclus ticks from echidnas. Analyses of 16S rRNA and flaB genes identified three closely related genotypes of this bacterium (Borrelia sp. Aus A-C) that were unique and distinct from previously described borreliae. Phylogenetic analyses of flaB (763 bp), groEL (1537 bp), gyrB (1702 bp) and glpQ (874 bp) gene sequences and concatenated sequences (3585 bp) of three gene loci (16S rRNA, flaB and gyrB) were consistent with previous findings and confirm that this novel species of the genus Borrelia is more closely related to, yet distinct from, the Reptile-associated (REP) and Relapsing Fever (RF) groups. At the flaB locus, genotypes A, B and C shared the highest percentage sequence similarities (87.9, 88 and 87.9 %, respectively) with B.orrelia turcica (REP), whereas at the groEL and gyrB loci, these genotypes were most similar (88.2-89.4 %) to B.orrelia hermsii (RF). At the glpQ locus, genotypes A and B were most similar (85.7 and 85.4 % respectively) to Borrelia sp. Tortoise14H1 (REP). The presence of the glpQ gene, which is absent in the Lyme Borreliosis group spirochaetes, further emphasises that the novel species of the genus Borrelia characterized in the present study does not belong to this group. Phylogenetic analyses at multiple loci produced consistent topographies revealing the monophyletic grouping of this bacterium, therefore providing strong support for its species status. We propose the name 'CandidatusBorrelia tachyglossi', and hypothesize that this species of the genus Borrelia may be endemic to Australia. The pathogenic potential of this bacterium is not yet known.

The family is classified under the order , belonging to the phylum . This family consists of genera that are of concern to human health, such as and [1], with common pathogenic species including ‘ subsp. pallidum', the causative agent of syphilis worldwide [2], and ‘ subsp. pertenue', the bacterium responsible for yaws [3]. The genus is a member of the family and through convention is divided into three major clades: Lyme disease/Borreliosis (LB) caused by members of the sensu lato complex, the Relapsing Fever (RF) borreliae and the Reptile-associated (REP) borreliae [4]. The LB borreliae currently comprise over 20 recognized species including the primary Northern hemisphere Lyme-disease-causing agents , , sensu stricto, and , along with a newly described genospecies ‘Candidatus mayonii’ that causes LB in the upper Midwestern USA [5-8]. Members of the LB group are vectored by hard ticks (family Ixodidae), with the pathogenic species commonly transmitted to humans and other animals by ticks within the Ixodes ricinus complex: Ixodes ricinus in Europe, Ixodes persulcatus in Europe and Asia, and Ixodes pacificus and Ixodes scapularis in USA [7]. These pathogenic spirochaetes are dependent on wildlife, particularly rodents and birds, which act as asymptomatic reservoir hosts that maintain their life cycles and transmission [9].

within the RF group have been reported throughout a number of continents, including Africa [10], Eurasia [11] and North America [12]. These borreliae, in contrast to the LB group, are generally transmitted by soft ticks (family Argasidae), with the exceptions of identified in I. persulcatus [13], I. ricinus [14], I. pacificus [15] and I. scapularis [16]; ‘' in Amblyomma americanum [17]; in Rhipicephalus microplus [18]; and ‘Candidatus texasensis’ in Dermacentor variabilis [19].

The third major clade of this genus, the REP borreliae, was identified after the discovery of in Hyalomma aegyptium ticks collected from tortoises in Turkey [20, 21], followed by subsequent addition of REP-related species of the genus identified in various reptiles [4, 22]. While the LB and RF spirochaetes consist of zoonotic pathogens and are of significant public health concern in many countries, the pathogenicity and zoonotic potential of the REP group are not yet known.

Although LB borreliae have never been identified in Australian ticks, wildlife or people [23], two RF borreliae are recognized: and that are transmitted by Rhipicephalus (Boophilus) australis [24, 25] and Argas persicus [26, 27], respectively. In addition, ‘' was the first species of the genus to be reported from native long-haired rats, Rattus villosissimus, in north-west Queensland, Australia. While the soft tick, Ornithodorus gurneyi, was considered to be the vector of this species due to its presence in the region, transmission experiments were not successful, and molecular characterization was never conducted to reliably identify the species of the genus [28, 29].

Sequence analysis of multiple loci offers the advantage over morphological characterization of being highly discriminatory, therefore serving as a reliable method for accurate identification, characterization and population, and epidemiological analyses in numerous bacterial studies [30, 31]. This technique was first used on burgdorferi in 2008 [32] and has become an increasingly common technique for taxonomic and epidemiological studies of this genus [33-36].

Recently, a novel species of the genus was detected in a number of echidna ticks, Bothriocroton concolor, collected [37]. An additional representative was detected in an Ixodes holocyclus tick [38]. Based on the 16S rRNA (1097 bp) and flagellin (flaB, 400 bp) gene phylogenetic analyses, this species of the genus formed a distinct clade from other well-described borreliae, indicating this organism, designated ‘ sp. Aus’, to be unique [37]. In the present study, we conducted sequence analyses of the flaB, groEL, glpQ and gyrB genes, in addition to the 16S rRNA and short flaB loci reported previously [37], to study the relationship of this novel species to other of the genus . Phylogenetic analysis confirmed its species status, and we hereby propose to designate this species as ‘Candidatus tachyglossi’.

This study was conducted under the compliance of the Australian Code for the Responsible Conduct of Research, 2007 and Australian Code for the Care and Use of Animals for Scientific Purposes, 2013. Tick collection was carried out opportunistically with the approval from the Murdoch University Animal Ethics Committee.

Genomic DNA was extracted previously from 97 Bothriocroton concolor ticks, with 38 (39 %) ticks testing positive at the -specific flaB locus [37]. These 38 -positive ticks were included in this study. Nested- and hemi-nested PCRs were conducted using primers targeting the housekeeping genes flaB (763 bp), groEL (1537 bp), glpQ (874 bp) and gyrB (1702 bp) (Table S1, available in the online Supplementary Material). Primers were designed to amplify short fragments of each gene with overlapping regions in order to obtain maximum coverage of the genes analysed for accurate characterization. PCR cycling conditions were as described by Loh et al. [37] (see Table S1 for respective annealing temperatures), with the exception of groEL primers: initial denaturation at 95 °C for 5 mins, 35 cycles of denaturation at 95 °C for 30 s, annealing at 48 °C for 40 s, extension at 72 °C for 2 mins, and a final extension at 72 °C for 7 mins. Amplification products of the targeted DNA products were electrophoresed in 1–2 % agarose gel with blue light using safe (Invitrogen), and positive samples were purified and sent for Sanger sequencing.

DNA sequences generated at flaB, groEL, glpQ and gyrB loci were aligned and analysed together with sequences representing species of the genus retrieved from GenBank. All sequences were aligned using mafft v7.017 [39] and then refined using muscle [40]. The best-fit model for each locus was assessed using mega6 [41] and was selected based on the Bayesian Information Criterion (BIC). Bayesian phylogenetic reconstructions using sequence alignments of all four loci were generated using MrBayes 3.2.6 [42], and concatenated phylogenetic reconstructions using concatenated sequence alignments were produced using the CIPRES Science Gateway V.3.3 [43]. GTR and HKY substitution models were selected, with gamma categories of five, MCMC length of 1 100 000, burn-in length of 10 000 and subsampling frequency of 200.

In the present study, -specific flaB (763 bp), groEL (1537 bp), gyrB (1702 bp) and glpQ (874 bp) DNA sequences were successfully amplified and sequenced from 38 -positive Bothriocroton concolor ticks described by Loh et al. [37]. Thirty samples were positive for all flaB fragments; 24 samples were positive for all glpQ fragments; 13 samples were positive for all groEL fragments; and 10 samples were positive for all gyrB fragments.

Previously, three closely related genotypes were distinguished in the Bothriocroton concolor ticks using the 16S rRNA gene sequences, tentatively given the designations sp. Aus A, sp. Aus B and sp. Aus C [37]. However, in the present study, these genotypes are referred as ‘Candidatus tachyglossi’ genotypes A, B and C, respectively. At the flaB locus, genotypes A, B and C consisted of nine, three and five identical samples, respectively. The flaB gene alignment (787 bp) between ‘Candidatus tachyglossi’ genotypes and other described species of the genus showed that the novel genotypes shared highest percentage sequence identities with from the REP group (87.9–88 %); similarities to the LB group ranged from 82.1–83.2 %; with the least similarity to from the RF group (79.8–80 %). The percentage identities within the ‘Candidatus tachyglossi’ genotypes ranged from 99.6–99.9 %. The percentage nucleotide identities between ‘Candidatus tachyglossi’ genotypes and (87.9–88 %) were higher than that between and (86.7 %). In contrast, the percentage nucleotide identities between ‘Candidatus tachyglossi’ genotypes and (84.6–84.8 %) were lower than that between and (86.7 %) (Table S2). Phylogenetic analyses of the flaB gene locus showed that the ‘Candidatus tachyglossi’ genotypes clustered most closely with with a high posterior probability (Fig. 1).

Fig. 1.

Phylogenetic reconstructions based on flaB, groEL, glpQ and gyrB gene sequences, and concatenated gene sequences of ‘Candidatus tachyglossi’ genotypes A, B, and C identified in Bothriocroton concolor ticks from echidnas. (AY241832), (NZ_CP010566), (X56907) and (JNUP01000064) were used as outgroups for each gene, respectively. Sequences determined in this study are indicated in bold type. Node labels represent posterior probabilities. Bars, substitutions per nucleotide position.

At the glpQ locus, genotypes A and B consisted of eight and two identical samples respectively. The amplification of this gene was not successful for genotype C. The glpQ nucleotide alignment (947 bp) between ‘Candidatus tachyglossi’ genotypes exhibited 83.9–85.7 % similarity with the REP species of the genus and lower similarities (80.6–84.8 %) with the RF species of the genus . The percentage similarity within ‘Candidatus tachyglossi’ genotypes was 98.6 %. Phylogenetic analysis confirmed the closer relationship of the ‘Candidatus tachyglossi’ genotypes from the present study with the REP group (100 % bootstrap support) (Fig. 1). The percentage nucleotide identities between ‘Candidatus tachyglossi’ genotypes and turcica (84.1–84.2 %) and (RF) (84.1–84.8 %) were higher than that between turcica and coriaceae (83.3 %) (Table S3).

At the groEL locus, ‘Candidatus tachyglossi’ genotypes A and B were identical, and both shared 99.9 % similarities with ‘Candidatus tachyglossi’ genotype C. Nucleotide alignment (1540 bp) between ‘Candidatus tachyglossi’ genotypes and other described species of the genus showed that the novel isolates had the least similarity with the LB group (83.9–85 %) and were most similar to from the RF group (89.3–89.4 %). Phylogenetic analysis of the groEL locus showed that ‘Candidatus tachyglossi’ genotypes from the present study clustered with the RF group with strong posterior probability support (Fig. 1). The percentage nucleotide identities between ‘Candidatus tachyglossi’ genotypes and (89.3–89.4 %) were higher than that between and (84.8 %); whereas the percentage identity between ‘Candidatus tachyglossi’ genotypes and (84.3 %) was slightly lower (Table S4). The genetic distance between ‘Candidatus tachyglossi’ and at this locus remains to be determined until the groEL gene is characterized in .

At the gyrB locus, ‘Candidatus tachyglossi’ genotypes A and B were identical. Nucleotide alignment (1712 bp) revealed that the ‘Candidatus tachyglossi’ genotypes exhibited 80.8–82.0 %, 80.8–82.0 % and 84.8–88.2 % sequence similarity with the LB Borreliae, the RF group and turcica, respectively. The percentage similarities within ‘Candidatus tachyglossi’ genotypes ranged from 99.6 to 99.9 %. Phylogenetic analysis showed that ‘Candidatus tachyglossi’ genotypes formed their own monophyletic clade, separate from turcica and the RF group, with high bootstrap support (98 %) (Fig. 1). The percentage nucleotide identities between ‘Candidatus tachyglossi’ genotypes and turcica (87.8–88 %) and hermsii (88.1–88.2 %) were higher than that between turcica and hermsii (86.7 %) (Table S5).

A Bayesian phylogenetic tree reconstructed using the concatenated alignment (3585 bp) consisting of three genes, 16S rRNA, flaB and gyrB, available for each of the main borreliae (Fig. 1), illustrated that the ‘Candidatus tachyglossi’ genotypes from Bothriocroton concolor ticks grouped separately, with turcica as the closest relative (91.1–91.2 % nucleotide identities) (Table S6). Likewise, the concatenated alignment (5154 bp), which excluded turcica (REP), based on the four loci amplified in the present study (16S rRNA, flaB, groEL and gyrB) also produced a similar tree topology with ‘Candidatus tachyglossi’ genotypes forming a monophyletic clade supported by high posterior probabilities (Fig. S1), sharing the highest percentage identities with hermsii (90.3 %) (Table S7).

All phylogenetic trees reconstructed revealed similar topologies with the REP group species of the genus as the closest sister clade. The recently established REP group has been detected in various reptiles from several countries and in ticks that parasitise them [4, 21]. The recent discovery of a novel species of the genus in Amblyomma varenense collected from the reticulated python (Python reticulatus) showed that the species of the genus identified clustered together with the REP-associated group, along with turcica, based on phylogenetic analyses of 16S rRNA and flaB genes [44]. However, the pathogenic potential of the species of the genus belonging to the REP group is unknown.

A number of members within the genus are well known to cause diseases in humans outside of Australia; nonetheless, borrelial tick-borne disease in humans still remains highly speculative and controversial in this country. ‘Candidatus tachyglossi’ was first reported in 2016 [37], hence the urgency to further characterize this bacterium on the basis of multi-loci gene sequencing in order to confirm the taxonomic position of this new member in the genus . Our results, based on sequence and phylogenetic characterization of multiple loci, provide conclusive evidence that ‘Candidatus tachyglossi’, identified in Bothriocroton concolor ticks from echidnas, is distinct from other described species of the genus and constitutes to a new clade in this genus.

Borrelial spirochaetes are well known to be associated closely with wildlife and utilize tick vectors to maintain a sylvatic life cycle [45, 46]. Australian wildlife are also known to be involved in spill-over of various zoonotic parasites [47]. Therefore, it is plausible that this bacterium is also likely to persist in the environment through circulation among native ticks and native mammalian (including marsupial) hosts. Unlike the Australian paralysis tick I. holocyclus, Bothriocroton concolor is a highly specialized tick, with echidnas as their primary host, and with a geographic distribution known only in Australia and Papua New Guinea [48]. ‘Candidatus tachyglossi’ has previously been identified in one human-biting tick (I. holocyclus) removed from an echidna [38] and its prevalence in I. holocyclus or other human-biting ticks remains to be determined. The morphological characteristics and the pathogenicity of this bacterium are also unknown.

Description of ‘Candidatus tachyglossi’

‘Candidatus tachyglossi′ (ta.chy.glos′si. N.L. gen. n. tachyglossi of Tachyglossus aculeatus, the monotreme host of the ticks in which the bacterium was first identified]).

Species can be differentiated from other borreliae based on sequence and phylogenetic analyses of five genomic loci (16S rRNA, flaB, groEL, gyrB and glpQ). Comparisons of the flaB gene sequences among the ‘Candidatus tachyglossi’ genotypes showed two single nucleotide polymorphisms (SNPs) in genotype A at bases 485 and 519 (GenBank accession no. KY586966); and one SNP in genotype B at base 227 (KY586964). As for the groEL gene, analysis revealed two SNPs in genotype C at bases 656 and 1143 (KY586970). Analysis of the gyrB gene showed two SNPs in genotype B at bases 985 and 1416 (KY586972); and five SNPs in genotype C at bases 405, 757, 1075, 1093 and 1420 (KY586973). At the glpQ locus, genotype A (KY586968) and genotype B (KY586967) showed 12 base differences at bases 3, 43, 103, 112, 264, 515, 520, 530, 532, 702, 819 and 852.

The DNA G+C contents for 16S rRNA, flaB, groEL, gyrB and glpQ genes of ‘Candidatus tachyglossi’ genotype A are 47.3, 40.9, 38.5, 33.8 and 34.1 mol%, respectively. The DA G+C contents for 16S rRNA, flaB, groEL, gyrB and glpQ genes of ‘Candidatus tachyglossi’ genotype B are 47.2, 40.9, 38.5, 33.9 and 34.7 mol%, respectively. The DNA G+C contents for 16S rRNA, flaB, groEL, and gyrB genes of ‘Candidatus tachyglossi’ genotype C are 47.2, 41, 38.5 and 34 mol%, respectively.