The Mitochondrial Genome of Toxocara canis

Toxocara canis (Ascaridida: Nematoda), which parasitizes (at the adult stage) the small intestine of canids, can be transmitted to a range of other mammals, including humans, and can cause the disease toxocariasis. Despite its significance as a pathogen, the genetics, epidemiology and biology of this parasite remain poorly understood. In addition, the zoonotic potential of related species of Toxocara, such as T. cati and T. malaysiensis, is not well known. Mitochondrial DNA is known to provide genetic markers for investigations in these areas, but complete mitochondrial genomic data have been lacking for T. canis and its congeners. In the present study, the mitochondrial genome of T. canis was amplified by long-range polymerase chain reaction (long PCR) and sequenced using a primer-walking strategy. This circular mitochondrial genome was 14162 bp and contained 12 protein-coding, 22 transfer RNA, and 2 ribosomal RNA genes consistent for secernentean nematodes, including Ascaris suum and Anisakis simplex (Ascaridida). The mitochondrial genome of T. canis provides genetic markers for studies into the systematics, population genetics and epidemiology of this zoonotic parasite and its congeners. Such markers can now be used in prospecting for cryptic species and for exploring host specificity and zoonotic potential, thus underpinning the prevention and control of toxocariasis in humans and other hosts.


Introduction
Roundworms (nematodes) belong to a large phylum (Nematoda) in the animal kingdom. This phylum contains a wide range of species with exceptionally varied life histories [1]. Many nematodes are parasites of plants or animals [2], causing significant diseases and major socio-economic losses globally [3][4][5]. Central to the control of such parasites is knowledge of their population genetics, which has important implications for understanding many areas, including systematics, epidemiology and ecology [6][7][8][9]. The basis for investigating population structures is the accurate analysis of genetic variation, which is known to be widespread in many parasitic nematodes [6][7][8]10], utilizing molecular markers with sufficient levels of intraspecific sequence variability.
Mitochondrial DNA markers are particularly suited to population genetic and systematic investigations due to their high mutation rates and proposed maternal inheritance [6,7,[10][11][12][13][14][15]. In spite of the availability of advanced DNA technologies, there is still a paucity of knowledge of mitochondrial genomes for many parasitic nematodes of socio-economic importance [16], such as members of the Ascaridida, which is a key group of nematodes that parasitizes humans and a range of other vertebrates. Although complete mitochondrial genome sequences are available for Anisakis simplex [17] and Ascaris suum [18], this is not the case for other parasites within this order, such as Toxocara canis, the common roundworm of dogs. This latter nematode parasitizes (at the adult stage) the small intestine of canids (definitive host) and is also transmissible to a range of other mammals, including rodents and humans, in which (after the oral ingestion of infective eggs) the larvae of Toxocara canis invade the tissues and cause covert toxocariasis, ocular larva migrans (OLM), visceral larva migrans (VLM) or neurotoxocariasis [19][20][21][22][23]. Although there has been a significant acquisition of knowledge about the biology of T. canis, there are still major gaps in our knowledge of the genetics, ecology and epidemiology of this enigmatic parasite [24]. In addition, the detection of a cryptic species of ''T. canis'' from cats in Malaysia [25,26], its subsequent genetic characterization [27] and its description as a new species -Toxocara malayensis [28], emphasize the need for detailed molecular genetic studies of T. canis populations using suitable genetic markers [29]. Exploring the mitochondrial genome of T. canis would provide such markers, as a foundation for molecular epidemiological and ecological studies, detecting cryptic species and assessing relationships of related species of Toxocara [29]. Furthermore, the sequencing of the mitochondrial genome of T. canis provides a useful, comparative dataset to those of Anisakis simplex [17] and Ascaris suum [18] within the order Ascaridida. Building on recent progress in long PCR-coupled, automated sequencing [30], the present study determined the sequence and structure of the mitochondrial genome for a representative individual of T. canis from Australia and compared it with those available for Anisakis simplex and Ascaris suum as well as the sequences from related nematode groups (Spirurida and Strongylida), as a foundation for systematic, population genetic and epidemiological studies of T. canis.

Materials and Methods
An adult, male specimen of Toxocara canis (sample code Tcn2; ref. [31]) was collected (during a routine autopsy) from the small intestine of a fox from Victoria, Australia under the Scientific Procedures Premises License for the Faculty of Veterinary Science, University of Melbourne (SPPL045). Initially, the morphological identification of the worm was based on the presence of a postoesophageal bulbus, the length and shape of the alae and the lengths of the spicules [32]. Total genomic DNA was extracted from a small portion (0.5 cm) of the specimen by sodium dodecyl-sulphate/ proteinase K treatment, phenol/chloroform extraction and ethanol precipitation [33], and purified over a spin column (Wizard Clean-Up; Promega) [34]. In order to independently verify the identity of the specimen, the second internal transcribed spacer (ITS-2) of nuclear ribosomal DNA was amplified by the polymerase chain reaction (PCR) and sequenced according to an established method [27]. The ITS-2 sequence obtained was a perfect match with that of T. canis (accession number Y09489; ref. [34]).

Features and Organization of the Mitochondrial Genome of Toxocara canis
The circular mitochondrial genome of T. canis (Figure 1) was 14162 bp in length (GenBank accession number EU730761) and contained 36 genes: 12 protein-coding genes (adenosine triphosphatase subunit 6 [atp6], the cytochrome c oxidase subunits 1, 2 and 3 [cox1-cox3], cytochrome b (cytb) and the nicotinamide dehydrogenase subunits 1-6 [nad1-nad6 and nad4l]), 22 tRNA genes (two coding for leucine and two coding for serine) and the small [rrnS] and large [rrnL] subunits of rRNA. Each proteincoding gene had an open reading frame (ORF), and all genes were located on the same strand and transcribed in the same direction (59 to 39) (Figure 1), consistent with the mitochondrial genomes of other secernentean nematodes characterized to date [16]. The gene arrangement for the mitochondrial genome of T. canis was consistent with that of GA2 [41]. This gene arrangement has been reported previously for members of the Ascaridida, including Anisakis simplex [17] and Ascaris suum [18] as well as members of the order Strongylida, such as the hookworms Ancylostoma duodenale and Necator americanus [35] as well as the barber's pole worm, Haemonchus contortus [42]. However, consistent with the mitochondrial genomes characterized to date for other Ascaridida (but not the Strongylida), the AT-rich region for T. canis was located between rrnS and nad1, flanked (59) by the genes trnS (UCN) and (39) by trnN and trnY.
All but the two serine tRNAs (AGN and UCN) had a predicted secondary structure containing a DHU arm and loop and a TVreplacement loop instead of the TyC arm and loop ( Figure 2). As reported previously for secernentean nematodes [15,16,43,44], the two serine tRNAs each contained the TyC arm and loop but lacked the DHU arm and loop. The rrnL and rrnS genes were 923 and 693 bp in length, respectively; the predicted secondary structure of each of these two genes are displayed in Figure 3 Author Summary Toxocara canis (Nematoda: Ascaridida) is the common roundworm of canids. This parasite is transmissible to humans as well as a range of other accidental or paratenic vertebrate hosts, in which (after the oral ingestion of infective eggs) the larvae of Toxocara canis invade the tissues and can cause different forms of clinical disease (toxocariasis). Although some aspects of the biology of T. canis are well understood, there are still significant gaps in our knowledge of areas including the molecular genetics, systematics, ecology and epidemiology of this and related parasitic nematodes. The present study elucidates the sequence, structure and organization of the mitochondrial genome of T. canis and provides mitochondrial gene markers for studies in these areas using molecular tools. A greater understanding of the epidemiology of species of Toxocara would improve the prevention and control of toxocariasis in humans and other animals. (rrnL) and Figure 4 (rrnS). The AT-content of the sequences of rrnL, rrnS and the AT-rich (''control'') region were 77.9%, 66.5% and 78.1%, respectively. The relatively low AT-richness exhibited in the mitochondrial genome of T. canis was pronounced for the rRNA genes: The AT-content of the rrnL sequence was 4.2% and 4.9% less compared with Anisakis simplex (76.1%) [17] and Ascaris suum (76.8%) [18], respectively. The AT-content of the rrnS sequence of T. canis was 5.5% and 5.4% less than that reported for Anisakis simplex (72.0%) [17] and Ascaris suum (71.9%) [18], respectively. Interestingly, the AT-content of the T. canis mitochondrial rRNA genes does not alter their predicted secondary structures with respect to those of other secernentean nematodes studied to date [35,39,41,42].
The AT-rich region (Figure 1) was 828 bp in length and was predicted to exhibit a complex secondary structure ( Figure 5). In addition, the AT-rich region contained 13 regions consisting of a varying numbers of the dinucleotide (AT) repeat (n = 3 to 21) between nucleotide positions 307 and 806 (see Figure 5). The presence of multiple AT repeats was similar to that described in the AT-rich region of several parasitic nematodes, including Ascaris suum [18], Anisakis simplex [17], Ancylostoma duodenale and Necator americanus [35], but distinct from the repetitive elements (CR1-CR6) within the AT-rich region of the free-living nematode Caenorhabditis elegans [18].

Comparative Analysis with Other Nematodes
Pairwise comparisons were made among the amino acid sequences inferred from individual protein-coding genes and the nucleotide sequences of the rRNA genes in the T. canis mitochondrial genome with those representing seven other nematodes (of the orders Ascaridida, Spirurida and Strongylida) ( Table 2). The amino acid sequence similarities in individual inferred proteins ranged from 70.3% (NAD2) to 94.4% (COX1) between T. canis and Ascaris suum and from 74.3% (CYTB) to 93.1% (COX1) between T. canis and Anisakis simplex. The amino acid sequence similarities between T. canis and each species of Spirurida (Brugia malayi, Dirofilaria immitis and Onchocerca volvulus) or Strongylida (Ancylostoma duodenale and Necator americanus) included ranged from 21.5% (ATP6) to 51.6% (COX1) and from 49.3% (NAD6) to 90.6% (COX1), respectively. The nucleotide sequence similarities ( Table 2) in rrnS were 80.5-81.1%between T. canis and the two other species of Ascaridida, 59.0-60.0% between T. canis and the three members of the order Spirurida, and 71.8-72.4% between T. canis and the two species of Strongylida. In addition, the nucleotide sequence similarities in rrnL were 74.8-78.0%, 59.2-61.0% or 65.2-67.0% between T. canis and individual species representing the order Ascaridida, Spirurida or Strongylida, respectively ( Table 2).

Implications for Testing the Phylogeny of Parasitic Nematodes
Pairwise comparisons of the amino acid sequences conceptually translated from the protein-coding genes as well as the nucleotide sequences of the ribosomal RNA genes indicated that the mitochondrial genome of T. canis most closely resembles those of selected members of the order Ascaridida. However, based on pairwise comparisons of sequence data, the next most similar nematode group is the Strongylida, but not the Spirurida. This finding is consistent with previous phylogenetic analyses of mitochondrial datasets, such as concatenated amino acid sequences for all 12 protein-coding genes [15] or gene arrangements [41]. These studies placed the Spirurida in a strongly supported clade separate from all other secernentean nematodes for which data were available, being in accordance with the evolutionary relationships based on traditional, taxonomic data [32,[45][46][47]. This placement of the Spirurida relative to the Ascaridida and Strongylida contrasts the classification of the Nematoda based on phylogenetic analysis of sequence data for the small subunit (SSU) of nuclear rRNA [48], inferring that members of the order Strongylida belong to ''clade V'', whereas those of the orders Ascaridida and Spirurida are within ''clade III''. The distinct taxonomic placement of the Spirurida relative to the Ascaridida and Strongylida is further evidenced by the variation in anti-codon usage in some of the mitochondrial tRNA genes of members of the Spirurida (i.e. Brugia malayi, Dirofilaria immitis and Onchocerca volvulus) as compared with the Strongylida and Ascaridida studied to date [15,39]. The incongruence in the inferred relationships of these three nematode orders (i.e. among the Ascaridida, Strongylida and Spirurida) challenges the proposed molecular phylogeny for the Nematoda based on SSU sequence data [48] and stimulates further investigation of a broader range of nematodes.

Implications for Systematic, Population Genetic, Epidemiological and Ecological Studies
There is major significance in the use of mitochondrial DNA markers for investigating the genetic make-up of species of the Toxocara, particularly given that there are no morphological features which allow the specific identification of some stages (e.g., larvae) [29] and given that cryptic species have been detected within the Ascaridoidea [27,[49][50][51][52][53][54][55][56][57][58][59][60][61][62]. In nematodes, mitochondrial DNA is proposed to be maternally inherited (cf. [16]) and is usually more variable in sequence within a species than nuclear ribosomal DNA [9]. Various different mitochondrial gene regions are suited to studying the population genetics of parasitic nematodes [8,10,15,[63][64][65][66][67]. However, surprisingly, there has been a paucity of information on the mitochondrial genomes of ascaridoid nematodes [16], which appeared to have related mainly to technical limitations and the cost associated with mitochondrial genome sequencing. To overcome this constraint, Hu et al. [30,36] developed the long PCR approach applied herein to T. canis, which has broad applicability to a range of ascaridoids, including other species of Toxocara, Toxascaris, Baylisascaris, Lagochilascaris and members of the Anisakis complex [29,68]. The characterization of the first complete mitochondrial genome sequence for T. canis, in the present study, provides a foundation for addressing ecological and epidemiological questions regarding this and related species. Conserved primers can be rationally and selectively designed to relatively conserved regions flanking ''variable tracts'' in the mitochondrial genome considered to be most informative (based on sequencing from a small number of individuals from particular populations, or genetic variants detected using nuclear ribosomal markers; refs. [29,67]). Using such primers, single-strand conformation polymorphism (SSCP) analysis [69] can be applied to pre-screen large numbers of individuals representing different populations and, based on the 'pre-screen', samples representing the entire spectrum of haplotypic variability can be selected for subsequent sequencing and analyses. Such an approach has been applied effectively, for example, to explore the genetic make-up of the Ascaris populations in humans and pigs in six provinces in China [70]. This study indicated restricted gene flow between human Ascaris and porcine Ascaris, and supported the conclusions from other previous epidemiological and experimental investigations [71,72] that pigs  are not a significant source of Ascaris infection to humans in endemic regions. Utilizing a range of mitochondrial gene markers (with differing degrees of intraspecific variability), such a mutation scanningtargeted approach is also readily and directly applicable to species of Toxocara. This is particularly relevant now, given that population variation and cryptic species have been detected within Toxocara [27,28,31] and that almost nothing is known about the transmissibility of these or other, as yet undetected, variants and/or cryptic species to humans and other hosts. For instance, early reports from Malaysia [25,26] described the occurrence of a parasite in cats, which was identified as T. canis, based on the presence of an oesophageal ventriculus and spear-shaped cervical alae in the adult [73]. This parasite differed from the common species known to parasitize cats [73][74][75][76], such as Toxocara cati, which has arrow-shaped cervical alae, and Toxascaris leonina, which lacks a ventriculus. Because T. canis has been found only rarely in cats elsewhere in the world [76][77][78][79][80][81], the question arose as to the specific identity of this parasite in Malaysian cats. A molecular study, using markers in the first and second internal transcribed spacers (ITS-1 and ITS-2, respectively) of nuclear ribosomal DNA markers, was undertaken to genetically characterize specimens of this parasite, then called Toxocara sp. cf. canis [27]. The molecular investigation indicated clearly that Toxocara sp. cf. canis from Malaysian cats was genetically distinct from T. canis and T. cati, a conclusion which was supported by a subsequent morphological study of a number of ascaridoids from Malaysia [28]. Three morphological features (for lips, alae and spicules) were identified which consistently differentiated Toxocara sp. cf. canis from T. canis, T. cati and other congeners, such as T. tanuki (from canids), T. apodemi and T. mackerrasae (from rodents), T. paradoxura and T. sprenti (from viverrids), T. vajrasthirae (from mustelids) and T. pteropodis (from bats). Hence, the findings from the molecular and classical systematic studies supported the conclusion that Toxocara sp. cf. canis represented a distinct species, subsequently named T. malaysiensis [28].
Although T. canis is well recognized as the causative agent of toxocariasis in humans, including ocular larva migrans (OLM) and/or visceral larva migrans (VLM), other congeners, such as T. malaysiensis, T. cati and T. vitulorum, may have greater zoonotic importance than assumed [82,83]. T. malaysiensis is of particular interest as a potential zoonotic pathogen, given its high prevalence (11%) in cats [26]. The transmissibility of this species to other host species (e.g., mouse, rat, rabbit and pig) warrants assessment, together with epidemiological surveys utilizing molecular tools employing genetic markers from the mitochondrial genome of T. canis as well as specific nuclear markers in the ITS-1 and/or ITS-2. The discovery of T. malaysiensis in cats in Malaysia [27] also raises important questions as to the identity and zoonotic potential of ascaridoids considered to represent T. canis in cats in other geographical regions, including South Africa, Panama, the USA and Czech Republic [76][77][78][79][80][81]84], which provides a stimulus for the genetic characterization of additional Toxocara isolates from a broad range of hosts and geographical origins and to subsequently evaluate their potential to infect humans and/or other hosts.
From epidemiological and ecological perspectives, it would be interesting, utilizing mitochondrial genomic data, to confirm or refute the involvement of Toxocara in human VLM cases in Japan, currently considered to be caused by Ascaris suum based on serological evidence [85,86], as there has been considerable controversy as to the specific identity of the causative agent of the disease in these instances [87]. It would also be particularly relevant to explore whether specific genotypes/haplotypes of Toxocara canis have a particular affinity to the human host and/or predilection sites in tissues to cause different types of toxocariasis and whether there are specific subpopulations of T. canis that undergo arrested development in tissues. Using molecular tools, in combination with traditional parasitological and serological methods, it should also be possible to characterize in detail experimental infections in ''model host systems'' (e.g., mouse, rabbit or pig) [24,[88][89][90][91][92][93]. Furthermore, mitochondrial markers would be useful for exploring the zoonotic risk of paratenic hosts, particularly those commonly encountered in an agricultural setting (e.g. chickens, ducks or pigs [94][95][96][97][98]), and determining the specific identity of eggs in the environment [99].
In conclusion, the present study emphasizes the relevance of the mitochondrial genome of T. canis defined herein, should provide a foundation for a range of systematic, population genetic, epidemiological, ecological and biological studies. Although the PCR-based sequencing-cloning approach used herein was effective, the PCR-coupled 454 technology platform [100], constructed recently for the direct sequencing of mitochondrial genomes from single nematodes [42], provides perhaps the most exciting development for large-scale, high throughput population genetic and mitochondrial genomic studies of nematodes and other organisms.