Rediscovery of Ixodes confusus in Australia with the first description of the male from Australia, a redescription of the female and the mitochondrial (mt) genomes of five species of Ixodes

We: (i) report the rediscovery of Ixodes (Sternalixodes) confusus Roberts, 1960 in Australia; (ii) redescribe the male and female of I.confusus; (iii) describe the mitochondrial (mt) genome of I. confusus from five ticks from four localities in Far North Queensland; and (iv) present the first substantial phylogeny of the subgenera of the Ixodes. The mt genomes of I.confusus, I. cornuatus, I. hirsti, I. myrmecobii and I. trichosuri are presented here for the first time. In our phylogeny from entire mt genomes (ca. 15 kb), the subgenus Endopalpiger was the sister-group to subgenera Sternalixodes plus Ceratixodes plus Exopalpiger whereas Exopalpiger was the sister to Sternalixodes plus Ceratixodes. [i.e. ((Endopalpiger) (Sternalixodes, Ceratixodes and Exopalpiger))]. Finally, we show that Ixodes anatis, the kiwi tick, may be closely related to the ticks of marsupials of Australia and Papua New Guinea.


Introduction
The taxonomy of Australian ticks has received renewed interest during the past decade, invigorated by rigorous morphological and molecular techniques, and through public health concerns. Furthermore, part of this resurgence is a realisation that the work of Roberts (1970) was not the final word on Australian ticks and that much remained to be done. The focus of recent investigations has been the genus Ixodes Latreille, 1795, with new species described, either from novel collections (Ash et al., 2017;Barker 2019) or from analyses of existing species (Heath and Palma 2017). However, good taxonomy also requires careful updating of previous descriptions, especially for taxa of uncertain taxonomic status. Among Ixodes of Australia, the species Ixodes confusus Roberts (1960) is a prime example.
Ixodes confusus was described from male and female ticks taken from wallabies in Sogeri, Papua New Guinea. Roberts (1955) initially considered the ticks to be Ixodes cordifer Neumann, 1908, a species with adults that typically parasitise cuscuses and possums in northern Australia, Papua New Guinea and Sulawesi (Roberts 1970;Wilson 1972). Later, Roberts (1960) realised that his specimens were a new species and, due to this confusion, named it I. confusus. The species is known from few specimens, yet Roberts (1970) considered the species to be common in Papua New Guinea on wallabies and cattle. However, in Australia I. confusus was known from just one female specimen taken from a human at Etty Bay near Innisfail, Queensland, in 1949. Thus, without further records in the following 70 years, the species may not be endemic, and the record could be from a human visitor from Papua New Guinea.
Ixodes confusus is one of eight species in the subgenus Sternalixodes Schulze, 1935(Camicas et al., 1998. All eight of these species are endemic to the zoogeographic region Australasia. Five species occur only in Australia (I. cornuatus Roberts, 1960, I. hirsti Hassall, 1931I. holocyclus Neumann, 1899;I. myrmecobii Roberts, 1962; I. trichosuri Roberts, 1960), one species only in Papua New Guinea (I. dendrolagi Wilson, 1967), and two species in both Australia and Papua New Guinea (I. confusus, I. cordifer). All of the species of Sternalixodes have a sternal plate in the female and nymph or only in the nymph (Roberts 1970;Wilson 1967). In addition, the females have a large scutum with strong lateral carinae; the palps are long and slender and the hypostome is lanceolate; the coxae are armed but lack syncoxae; and the anal grooves in the female meet at a point behind the anus. In the male, the anal plate is pointed posteriorly and the scutum has lateral grooves (Roberts 1970;Wilson 1967). Despite the morphological distinctiveness of the subgenus Sternalixodes, in general the subgeneric classification of Ixodes is controversial (Kwak and Heath, 2018;Barker et al., 2021) and requires further testing, particularly with genomic data. Nevertheless, some subgenera seem readily identifiable (Roberts 1960(Roberts , 1970, and at least Endopalpiger Schulze, 1935 has both morphological and genomic support . Herein, we show that Ixodes confusus is endemic to northern Queensland and provide a full description of the male and female. We also provide mitochondrial genomes for I. confusus and four other species of Sternalixodes (I. cornuatus, I. hirsti, I. myrmecobii and I. trichosuri) which further support the recognition of Sternalixodes as a distinctive subgenus of Ixodes.

Material examined
Only field-collected ticks were available for study. The specimens were from new collections made by two of us (DB, SCB) from the Barker and Barker Collection (B&B) at the University of Queensland (UQ), the Queensland Museum (QM), the Australian National Insect Collection (ANIC) and the United States National Tick Collection (USNTC aka USNMENT, the US National Museum Entomology Collection) (Table 1).

Microscopy methods
Ticks were studied using a stereoscopic light-microscope (Nikon SMZ800N, Nikon Corporation, Tokyo, Japan and Olympus SZX16, Olympus Corporation, Tokyo, Japan) and a scanning electron microscope (TM4000 Plus Hitachi High-Technologies Corporation, Tokyo, Japan). Measurements are in millimetres and are given as the range followed by the mean and the number of specimens measured (n) in parentheses. Colour digital images were taken with a Canon 6D camera (Canon Corporation, Tokyo, Japan).

Sequencing and assembly of mitochondrial genomes
Tick DNA was extracted and prepared for sequencing at the University of the Sunshine Coast and UQ. Groups of ticks were cut in half and then incubated at 56 • C for 62 h with Proteinase K to lyse the cells. The QIAGEN DNeasy Blood and Tissue kit was used to extract genomic DNA. The amount of DNA recovered was measured with Nanodrop and Qbit instruments. Groups of ticks that yielded more than 200 ng of DNA were sent to Novogene Singapore for de novo library construction and next-generation Illumina sequencing. Groups of ticks with less than 200 ng were combined with DNA from a distinctly different organism, usually a bird, to reach the minimum threshold of 200 ng of DNA required by Novogene Singapore. At Novogene Singapore, DNA was sonicated to fragments, then fragments were end-polished, A-tailed and ligated with Illumina adaptors. DNA fragments were amplified with PCR, using P5 and P7 oligos, to create genomic libraries which were purified with AMPure XP system. The Illumina Novaseq 6000 sequencing platform was used to generate two giga-bases of nucleotide sequence data (PE 150). De novo contig assemblies of Illumina sequences were then constructed with Geneious Prime (Kearse et al., 2012). Blast-searches of contigs revealed mt genes of ticks; these gene sequences were then assembled until entire mt genomes had been assembled.

Annotation of mitochondrial genomes
Mitochondrial genomes were annotated with Geneious Prime. Protein-coding genes were identified by searches with BLAST (Chen et al., 2015) for open reading-frames. Regions between protein-coding genes were searched with BLAST (Chen et al., 2015) to find rRNA genes, tRNA genes and control regions. The tRNA that we expected to find but were not found with BLAST were found with the aid of the tRNAscan-SE Search Server v1.21 (Lowe and Chan 2016) and the MITOS Web Server (Bernt et al., 2013). The nucleotide sequences of tRNA genes were confirmed by studying the putative secondary structure of transcripts, as implemented in Geneious Prime (Kearse et al., 2012).

Phylogenetic methods
Phylogenies were inferred by both Maximum Likelihood (ML) and Bayesian Inference (BI) methods implemented in the RAXML-HPC2 v 8.2.12 (Stamatakis 2014) and MrBayes v3.2.2 (Ronquist et al., 2012) respectively. The sequence-alignment was put though Gblocks to remove regions with alignment gaps. JmodelTest2 v2.1.6 (Diego et al., 2012) was used to find the optimal substitution model for the nucleotide dataset. The GTR + G + I model was found to be the best fit for our dataset. In all ML and BI runs (experiments), genes were partitioned. Rapid-bootstrapping of 1000 replicates of our data was executed in RAXML-HPC2 v 8.2.12 (Stamatakis 2014). There were two simultaneous BI runs: 10 million generations sampled every 1000 MCMC steps. For every BI run, four MCMC chains (three heated and one cold) were executed. The first 25% of steps were discarded as burn-in. Tracer v 1.5 (Rambaut, 2009) was used to observe the effective sample size (ESS) and convergence of independent runs. Phylogenetic trees were displayed in FigTree v 1.4.4 (Rambaut, 2012). Branch support in the phylogenetic trees generated by RAXML-HPC2 v 8.2.12 (Stamatakis 2014) and MrBayes v 3.2.2 (Ronquist et al., 2012) was assessed by the bootstrap values and posterior probability values, respectively. All phylogenies were inferred through the CIPRES Science Gateway v.3.3 (Miller et al., 2010). Ixodes (Ixodes) pavlovskyi Pomerantzev, 1946, a species from the "other Ixodes" clade (sensu Barker et al., 2021) was the out-group.

Rediscovery of Ixodes confusus in Australia
I. confusus is now known at five sites in Far North Queensland: at two sites at Etty Bay near Innisfail, and at Mt Molloy, Cardwell and Mossman River Gorge, Qld (Table 1; Fig. 1).
Legs moderately long, slender. Coxae (Fig. 5H): coxae I-IV with moderately long, broadly triangular external spur with blunt apex; spur on coxa I slightly longer than those on other coxae, spur on coxae II and III subequal, spur on coxa IV nearly twice shorter than on other coxae; coxae I-IV with distinct longitudinal ridges laterally. Trochanters I-IV with distinct but small triangular spur ventrally (not obvious on SEMs but can be seen by light microscopy). Tarsus I: length 0.80-1.03 (0.94 n = 3); tarsus IV length 0.75-0.88 (0.83); tarsi slightly humped subapically. Roberts (1955Roberts ( , 1970 had much trouble distinguishing I. confusus and I. confusus, particularly the males. Since both I. confusus and I. confusus are now known to be present in Far North Queensland there was the possibility that we might inadvertently associate a male of I. confusus (rather than a male of I. confusus) with the females of I. confusus. Therefore, we sequenced the mt genome of a putative male of I. confusus from Mt Molloy so that the sequence of this male could be compared with the sequence of the females we sequenced from Mt Molloy, Etty Bay and Cardwell. The male sequence matched the females well (99.6%) confirming that the collected males were indeed I. confusus. Alas, our attempts to sequence the mitochondrial genome of I. cordifer from Far North Queensland and Papua New Guinea failed. However, our morphological study supported the separation of I. confusus from I. confusus.

Remarks and differential diagnosis
In the male of I. confusus the spur on coxa IV is only a little longer than that on coxa III, and the spurs on trochanters III and IV are stout, whereas in I. confusus the spur on coxa IV is much longer than spur on coxa III, and the spurs on trochanters III and IV are slender. The female of I. confusus can be easily distinguished from that of I. cordifer by the former having ridges on the dorsal (one ridge) and ventral (three ridges) basis capituli and ridges on coxae I-IV. Furthermore, in I. confusus the sternal plate is almost triangular (Fig. 5C) whereas in I. cordifer it is almost rectangular; the males of these species are also distinguished in the keys of Barker (2019).

Mitochondrial (mt) genomes of Ixodes confusus and its relatives
The mt genomes of five species are presented here for the first time: I.  (Fig. 6). The mt genomes of I. confusus, I. cornuatus, I. hirsti, I. trichosuri and I. myrmecobii have the genearrangement that is typical of Ixodes that are endemic to Australia (Barker et al., 2021, Fig. 1). A phylogeny from these mt genomes, together with previously published mt genomes of Ixodes, indicate monophyly of the subgenera Sternalixodes and Endopalpiger (Fig. 7). The mitochondrial genomes published for the first time in this paper have been submitted to GenBank database; accession numbers OL614953 to OL614959.

Rediscovery of Ixodes (Sternalixodes) confusus in Australia
Before our study, I. confusus was known in Australia from a single specimen collected from a human at Etty Bay in 1949 (Table 1; Roberts 1960Roberts , 1970. Thus, it was extremely doubtful that I. confusus was endemic to Australia. Indeed, the single female specimen could have been brought to Australia from Papua New Guinea on a human, since humans often travel between Papua New Guinea and Cairns, and not a single adult I. confusus had since been collected in Australia. However, a concerted collection effort (DB, SCB) demonstrated that this species is present in Far North Queensland, where it is probably widespread (Table 1; Fig. 1). The host records suggest adults have a preference for macropods (27/35 ticks) but the records from a human, horse and cattle show I. confusus occasionally attaches to other hosts. [We recently found another female in the USNTC from Mossman River Gorge, Mossman, Qld, which was collected during an Archibold Expedition in 1948 (Table 1).]

Mitochondrial (mt) genomes of Ixodes confusus, four of its relatives, and the phylogenetic position of the subgenus Sternalixodes
The live I. confusus provided the first opportunity to investigate the phylogenetic relationships of I. confusus to its relatives with large numbers of nucleotides. Previously, only small numbers of nucleotides have been recovered from museum and other specimens (e.g. Ash et al., 2017;Kwak et al., 2017). Thus, we sequenced the entire mitochondrial genome of I. confusus (five individuals from three localities; refer to Fig. 1, Appendix 1), and four of its relatives: I. cornuatus, I. hirsti, I. trichosuri, and I. myrmecobii. We presented the first substantial phylogeny of the subgenera of the genus Ixodes (Fig. 7). The subgenera Sternalixodes (6 species) and Ceratixodes Neumann, 1902(I. uriae White, 1852 were sisters (sister-groups). This is intriguing since I. uriae is exclusively a sea-bird tick whereas the eight known species of Sternalixodes infest mammals in the adult stage, although birds are apparently the main hosts of the nymphs and larvae in at least two species, I. cordifer and I. hirsti .
It is also intriguing that the subgenus Exopalpiger Schulze, 1935 (I. fecialis Warburton and Nuttall, 1909) was the sister to subgenera Sternalixodes plus Ceratixodes [i.e. the arrangement (Exopalpiger, (Sternalixodes, Ceratixodes))] rather than Exopalpiger being closely related to Endopalpiger (Fig. 7). Indeed, Exopalpiger was well-removed from Endopalpiger in our tree (Fig. 7). So, Camicas and Morel (1977) and Camicas et al. (1998) were mistaken when they subsumed Endopalpiger into Exopalpiger. We can only wonder why Camicas and Morel (1977) and Camicas et al. (1998) (Fig. 7). In other words, Sternalixodes, Ceratixodes and Exopalpiger shared a Most Recent Common Ancestor to the exclusion of Endopalpiger. The two subgenera in our tree with more than one species, Endopalpiger and Sternalixodes, were monophyletic (Fig. 7). Regarding Sternalixodes, we do not have mt genomes for two of the eight species of this subgenus: I. cordifer from Australia and Papua New Guinea; and I. dendrolagi from Papua New Guinea. We do not expect, however, that mt genomes from these two species will challenge the hypothesis of a monophyletic Sternalixodes since I. confusus, I. cordifer and I. dendrolagi are morphologically similar and thus likely, closely related. Indeed, Wilson (1967) considered I. confusus, I. cordifer and I. dendrolagi to be so closely related that he designated the I. cordifer (species) group for these three species and the associated subspecies of I. cordifer (I. cordifer cordifer and I. cordifer bibax).
Finally, we took the opportunity to make a phylogeny from all 27 of the entire mt genomes that are now available for Ixodes (Appendix 2). The subgenera Sternalixodes, Endopalpiger and Ixodes were monophyletic in our tree (Appendix 2).

Ixodes anatis Chilton, 1904, the kiwi tick, may be a closely related to the ticks of marsupials of Australia and Papua New Guinea
The present study of I. confusus and its relatives in the subgenus Sternalixodes led us to consider I. anatis, the kiwi tick, since I. anatis was placed in the subgenus Sternalixodes by Clifford et al. (1973) but considered best removed from Sternalixodes by Kwak and Heath (2018). We made a tree with the cox 1 fragment (674 bp) from I. anatis of Kwak et al. (2017), together with cox 1 sequences from the mt genomes sequenced by us in the present study, and some cox1 sequences from GenBank (Appendix 3). Our trees from this short fragment of cox 1 had I. anatis as a member of the "Australian Ixodes" clade. These trees, however, were from only 674 bp of one gene, cox 1, of the mt genome. For instance, the unresolved position of I. uriae and I. woyliei in the cox 1 tree is resolved by entire mt genomes (Fig. 7). Thus, conclusions about the affinities and evolutionary history of I. anatis, the kiwi tick, must await entire mt genome sequences.

Declaration of competing interest
On behalf of my co-authors I declare that we do not have any conflicts of interest to declare.