Background

Nematodes (roundworms) are common parasites which inhabit in the gastrointestinal tract of humans, and domestic and wild animals, and most of them can cause significant economic losses and public health problems [1]. Toxascaris leonina (Nematoda: Ascarididae) is a common nematode of various animals, including dogs, cats, wolves, tigers, lions and foxes. Although it has a diverse range of definitive hosts, to date, it has been the only species described in the genus Toxascaris. Adult T. leonina parasitize in the small intestines of the definitive hosts and can cause serious disease in young animals [2]. Importantly, larvae of T. leonina can also infect humans, posing a potential public health problem [3].

The species identification of parasites is of significance for studying epidemiology, systematics, diagnostics and population genetics of parasites [4]. Usually, nematodes are identified based on morphological features, hosts, and geographical distributions. However, it may not always be possible to identify to the species level nematodes from different hosts and geographical locations based on these criteria. To overcome the limitation of morphological taxonomy, molecular/DNA markers have been extensively used for species identification and differentiation of various nematodes [5,6,7]. Recently, our preliminary comparative study has indicated that the mitochondrial (mt) cox1 gene sequence of T. leonina from cheetahs and dogs differed by 6.8% [8]. In addition, substantial nucleotide difference in part of the mt nad1 (9.0%) and nad4 (10.8%) were also detected between T. leonina from canid and felid hosts [9]. These findings raised the possibility that T. leonina may represent a species complex. However, this hypothesis was proposed based on a small molecular dataset. Therefore, this hypothesis should be tested by using more datasets. Mt genome (containing 36–37 genes) has been considered a suitable marker for the species identification and differentiation of many nematodes [10]. The mt genomes of T. leonina from canids (such as the dog) and felids (such as the South China tiger) have been sequenced and characterised previously [11, 12], but the T. leonina mt genome sequence from the South China tiger has not yet been deposited in the GenBank NCBI database. Moreover, it is yet unknown whether there are significant differences in the entire mt nucleotide and amino acid sequences between T. leonina from the dog and those from the South China tiger [12].

Therefore, in the present study, we (i) characterized the entire mt genome sequence of T. leonina from the cheetah, (ii) compared it with that of T. leonina from the dog, and (iii) then tested the hypothesis that T. leonina represents a species complex.

Methods

Parasites and DNA isolation

All adult worms (n = 14; coded TC1–TC14) of T. leonina were obtained from a cheetah (Acinonyx jubatus) in Changsha Ecological Zoo, China. This cheetah was dewormed at importation to Changsha in 2012. This cheetah died in 2018 and 14 adult nematode specimens were collected by the Zoo veterinarians at necropsy. Adult nematode specimens were washed separately in physiological saline, identified preliminarily to the species level based on morphological characters [13], fixed in 70% (v/v) ethanol and stored at − 20 °C until further study was performed. Some representative examined specimens were deposited in the Specimen Museum and Resource Bank of Parasites, Lanzhou Veterinary Research Institute, Chinese Academy of Agricultural Sciences (accession no. SMRBP-LVRI-2018001-5), for retrieval and comparative study. Total genomic DNA was isolated from individual worms using sodium dodecyl sulphate/proteinase K treatment, followed by spin-column purification (Wizard® SV Genomic DNA Purification System, Promega). Single specimen was further identified as T. leonina by PCR-based sequencing of ITS1 and ITS2 rDNA [14], and both regions had 100% nucleotide homology with those of T. leonina from Lynx lynx deposited in GenBank (GenBank accession no. JF837179). However, these sequences differed by 3% in ITS1 compared with the sequence of T. leonina from a wolf in China (GenBank: JF837174), and by 8% in ITS2 compared with the sequence of T. leonina from a golden jackal in Croatia (GenBank: MF495479).

Long-PCR amplification and sequencing

The primers (Additional file 1: Table S1) were designed according to well-conserved regions of the mt genome sequences of T. leonina from the dog (GenBank: NC023504) [11]. These primers were then used to amplify the entire mt genome by long-PCR as seven overlapping fragments (2–4 kb each) from the genomic DNA. The PCR reactions contained ~ 20 ng of genomic DNA and were carried out in 50-μl reaction volumes containing 25 μl 2× Phusion Emweald Amp MAX HS PCR Master Mix (TaKaRa, Dalian, China), 2 μl of each primer, 2 μl DNA and 19 μl of ddH2O in a thermocycler (Bio-Rad, California, USA). Cycling amplification was executed under the following conditions: 94 °C for 2 min (initial denaturation), then 94 °C for 30 s (denaturation), 50–55 °C for 45 s (annealing), 66 °C (~ 1 kb region) for 5 min (extension) for 9 cycles, followed by 94 °C for 30 s, 50–55 °C for 45 s, and 66 °C for 5 min for 21 cycles, and a final extension at 68 °C for 10 min. Each PCR reaction yielded a single band detected in a 1.0% (w/v) agarose gel upon ethidium-bromide staining. PCR products were column-purified and then sequenced from both directions.

Sequence analyses

Contiguous sequences were assembled manually and aligned with previously published mt genome sequence of T. leonina from the dog using Clustal X program v.1.83 [15] to identify gene boundaries. The open-reading frames (ORFs) and each protein-coding gene (PCG) were predicted using the program MEGA 5.0 [16], then the translation start and stop codons were identified based on comparison with those reported previously [11]. Putative secondary structures of 22 tRNA genes were identified using tRNAscan-SE [17] with manual adjustment, and two rRNA genes were predicted by comparison with those of T. leonina from the dog [11].

Phylogenetic analyses

The deduced amino acid sequences of 12 PCGs were aligned individually with those of published mt genomes of other selected species of the Ascaridoidea, using Thelazia callipaeda as the outgroup [18]. All amino acid sequences were aligned using MAFFT 7.122 [19] and then concatenated into a single alignment. The regions and sites of ambiguous alignment were eliminated using Gblocks online server [20]. Bayesian inference (BI) was used for phylogenetic analyses as described previously [6, 11]. Phylograms were drawn using the program FigTree v.1.4.

Sequencing of cox1 and analysis

The primers JB3/JB4.5 were used for PCR amplification and then sequencing of a portion of the cox1 gene from multiple T. leonina samples from the cheetah. These newly generated sequences and previously published cox1 sequences of T. leonina were aligned using the software MAFFT 7.122. Phylogenetic analysis was performed using the same methods as described above.

Results and discussion

General features of the mt genome of T. leonina from the cheetah

The entire mt genome of T. leonina from the cheetah is 14,685 bp in size. This mt genome contains 12 PCGs, 22 tRNA genes and two rRNA genes (Additional file 1: Table S2). The nucleotide content of the mt genome sequence of T. leonina from the cheetah is biased toward A+T (71.1%), in accordance with mt genomes of other nematodes sequenced to date (such as Ascaris suum, Ascaris lumbricoides, Toxocara canis, Toxocara cati) [6, 21]. All mt genes of T. leonina from the cheetah are transcribed in the same direction. Gene content, arrangement and composition of the mt genome of T. leonina from the cheetah are the same as those of other nematodes, such as T. canis, T. cati and Toxocara malaysiensis [21]. Additionally, the genes in the mt genome of T. leonina from the cheetah overlap by 1 bp between tRNA-Trp and tRNA-Glu (Additional file 1: Table S2). There are 12 intergenic regions (from 1 to 10 bp in length) in the mt genome of T. leonina from the cheetah (Additional file 1: Table S2). In addition, the mt genome of T. leonina from the cheetah contains two non-coding (AT-loop) regions (1450 bp) (Additional file 1: Table S2).

Annotation

The mt genome of T. leonina from the cheetah encodes 12 PCGs. The most common start codon for the mt genome of T. leonina from the cheetah is TTG (7 of 12 PCGs), followed by ATT (3 of 12 PCGs), GTG (1 of 12 PCGs) and GTT (1 of 12 PCGs) (Additional file 1: Table S2). In this mt genome, the PCGs use TAG, T and TA as a stop codon (Additional file 1: Table S2), consistent with previous studies [11, 12]. Twenty-two tRNA genes in this mt genome range from 52 bp to 62 bp in length. The rrnL and rrnS genes of the mt genome of T. leonina from the cheetah were 959 bp and 700 bp in size, respectively (Additional file 1: Table S2), and the long and short non-coding regions were located between tRNA-SerUCN and tRNA-Asn, and nad4 and cox1, respectively (Additional file 1: Table S2).

Comparative analyses between T. leonina from the cheetah and the dog

The entire mt genome of T. leonina from the cheetah is 375 bp and 408 bp longer than that of T. leonina from the dog and the South China tiger [11, 12], respectively. Nucleotide sequence difference across the entire mt genome (except for the non-coding region sequences) between T. leonina from the cheetah and the dog was 7.2%. This level of mt genome divergence is similar to those between different nematode species. For example, the nucleotide sequence difference was 3.8–9.4% among members of the Pseudoterranova decipiens species complex [22], 11.1% between Ancylostoma duodenale and Ancylostoma caninum [23, 24], 12.5% between T. malaysiensis and T. cati [21], and 6.2% between Baylisascaris schroederi and Baylisascaris ailuri [25]. For the 12 PCGs, the sequence difference between T. leonina from the cheetah and the dog, was 5.0–9.7% (nucleotide sequences) and 1.0–7.2% (aa sequences); these divergence levels are similar to those recorded between other congeneric nematode species. For example, the amino acid sequence divergence between Oesophagostomum dentatum and O. quadrispinulatum was 0.4–8.3% [26], and it was 0.4–7.1% between Angiostrongylus mackerrasae and A. cantonensis [27]. For the 22 tRNAs, sequence difference was 6.0% between T. leonina from the cheetah and the dog. For the rrnL and rrnS genes, sequence difference was both 6.1% between T. leonina from the cheetah and the dog (Table 1).

Table 1 Differences in mitochondrial nucleotides and predicted amino acids sequences between Toxascaris leonina from cheetah (Tc) and dog (Td)
Full size table

Nucleotide sequence difference in the mt cox1 gene was also assessed among T. leonina isolates (n = 23; coded TC1–TC23) from canid and felid hosts. Nucleotide differences in cox1 sequences among the T. leonina isolates from felids (n = 17; coded TC1–TC17) occurred at 10 sites (2.7%). Nucleotide differences among the T. leonina isolates from canids (n = 6; coded TC18-TC23) occurred at 21 sites (6.2%). However, comparison of the mt cox1 sequences among all the T. leonina isolates (n = 23) revealed 36 variable sites (10.6%). Our results revealed substantial sequence differences in the cox1 gene among different hosts, consistent with that of previous studies [8, 9].

Taken together, these results support our proposal that T. leonina from different hosts may present a species complex because a previous study has shown that sequence difference in mt genes between closely-related nematode species were 10–20% [28].

Phylogenetic analysis

Phylogenetic analyses based on amino acid sequences of 12 PCGs showed clear genetic distinctiveness between T. leonina from the cheetah and the dog, and they were a sister taxon to a clade containing Ascaris spp., Parascaris spp. and Baylisascaris spp. of the family Ascarididae, with strong support (Bayesian posterior probabilities = 1). Basal to these taxa were the families Anisakidae, Toxocaridae, Heterakidae and Ascaridiidae (Fig. 1). The genetic distance between the two T. leonina isolates is approximately the same (looking at tree topologies and branch lengths) as that among the Baylisascaris spp. and the Pseudoterranova spp. (Fig. 1). In addition, phylogenetic analyses based on the mt cox1 sequences also showed support for the separation of T. leonina from different hosts into three distinct clades (Fig. 2). Our results showed that T. leonina isolates from felid hosts were in the same clade; however, T. leonina isolates from canid hosts formed two clades (Bayesian posterior probabilities > 0.71). Toxascaris leonina from red foxes was sister to the two remaining clades (Fig. 2). These results are consistent with a recent study [29].

Fig. 1
figure 1

Relationships of Toxascaris leonina with other selected ascaridoidea nematodes based on mitochondrial sequence data. The concatenated amino acid sequences of 12 protein-coding genes were subjected to a Bayesian inference analysis using Thelazia callipaeda as the outgroup. Bayesian posterior probabilities values are indicated

Full size image
Fig. 2
figure 2

Inferred phylogenetic relationships among representative Toxascaris leonina samples by Bayesian inference based on mitochondrial cox1 nucleotide sequence data, using Ascaris suum as the outgroup. Bayesian posterior probabilities values are indicated

Full size image

Conclusions

In summary, the present study determined the entire mt genome sequence of T. leonina from the cheetah. Comparative analyses of the mtDNA datasets provided genetic evidence that T. leonina from canid and felid hosts represents a species complex. To further test this hypothesis, additional studies are warranted (i) to explore, in detail, nucleotide variation in rDNA and mtDNA within and among T. leonina populations from different hosts and geographical locations; (ii) to determine the complete mt genomes of T. leonina from other hosts; and (iii) to undertake detailed morphological studies of T. leonina from a range of hosts.