Understanding the diversity and phylogeny of whipworms is important; false classification limits our acuity of biogeography and conceals the zoonotic potential of trichurids (Callejón et al., 2015; Doležalová et al., 2015). Certain species, e.g. T. suis (Schrank, 1788) and T. trichiura (Linnaeus, 1771), are problematic in developing countries and have vast socioeconomic impacts via human or livestock infections (nearly 1 billion human trichuriasis infections are reported globally each year) (Jex et al., 2014). Other species, such as T. muris (Schrank, 1788; Hall, 1916), have been gathering attention in biomedical research for potential use in immunosuppression therapy (Feliu et al., 2000). A more comprehensive understanding of relationships within this group would enable predictions about how close relatives interact with their host(s).
Relationships within Trichuridae have not been well resolved using genetic approaches; results differ depending on the gene(s) sequenced and the approach used for phylogenetic reconstructions (Callejón et al., 2015). Mitochondrial data, primarily cox1, have been commonly used and have allowed for high resolution of closely related lineages; however, it may be less credible to use with Trichuris species due to the degree of hybridization and maternal mitochondrial heredity seen in this genus (Callejón et al., 2015; Doležalová et al., 2015). Nuclear data have provided higher support for relationships than mitochondrial data (Doležalová et al., 2015). The nuclear ITS1-ITS2 genes offer markers that allow closely related species to be detected (Eberhardt et al., 2019) and ITS1-5.8S-ITS2 has been used to show relations among ruminant- and rodent-infecting species (Doležalová et al., 2015). However, the number of variants of RNA genes (including the ITS2 region) makes their utility in disentangling the phylogeny of Trichuris less opportune, particularly given that the amount of ploidy is unknown (Doležalová et al., 2015). The 18S rRNA gene has been used to infer the placement of trichurids within Nematoda as well as to elucidate relationships within Trichuridae and is less prone to result in unclear multiple alignments (Callejón et al., 2013, 2015; Guardone et al., 2013; Doležalová et al., 2015). To date, both nuclear and mitochondrial data have suggested that Trichuris may be a polyphyletic genus; species or groups within the genus, e.g. T. trichiura and T. suis, may also be polyphyletic (Doležalová et al., 2015).
Trichuris fossor (Hall, 1916) has been reported only from hosts belonging to the genus Thomomys (Wied-Neuwied 1839) (Rodentia: Geomyidae) (Todd and Lepp, 1972; Gardner, 1985; but see Falcón-Ordaz, 1993). Descriptions have been based on morphology and host preference and Trichuris from geomyid hosts have never been sequenced (Eberhardt et al., 2019). The aim of this study was to serve as the first molecular report for T. fossor. The 18S rRNA gene was sequenced for four specimens identified putatively as T. fossor collected from separate Thomomys (western pocket gophers) species hosts.
Materials and methods
Thomomys bulbivorus (Richardson, 1829; Brandt, 1855) were salvaged from trappers working in Yamhill County, Oregon, in April 2018. Thomomys bottae (Eydoux and Gervais, 1836), T. mazama (Merriam, 1897), and T. talpoides (Richardson, 1828) were collected from Curry, Jackson, and Grant County, respectively, Oregon in August 2019. Complete intestinal tracts were examined following procedures outlined by Gardner and Jasmer (1983). Collected parasites were stored in 95% ethanol. Based on morphology and previous records for the hosts, nematodes found in the ceca were tentatively identified as T. fossor (Chandler, 1945; Todd and Lepp, 1972; Gardner, 1985).
DNA extraction, amplification, and sequencing were performed on individuals from each host species. Before beginning isolation, nematodes were transferred to 1.5 mL microcentrifuge tubes and repeatedly rinsed with DI water (5 rinses of 1 mL dH2O) to remove all traces of ethanol. Nematodes were then transferred into fresh PCR tubes and mechanically homogenized before extracting with the DNeasy Blood and Tissue Kit (Qiagen) following the manufacturer’s protocols. Overlapping fragments of 18S rRNA were amplified using the primers G18S4F (5′-GCTTGTCTCAAAGATTAAGCC-3′), 136R (5′-TGATCCTTCTCGCAGGTTCACCTAC-3′), 652F (5′-GCAGCCGCGGTAATTCCAGCTC-3′), and 647R (5′-CATTCTTGGCAAATGCTTTCGC-3′) (Callejón et al., 2013). After PCR products were visualized on a 1.5% agarose gel, they were SPRI-purified (Elkin et al., 2001) and prepared for direct end sequencing. Sequencing reactions were processed by the Center for Genome Research and Biocomputing (CGRB; Oregon State University, Corvallis, OR).
Sequences were examined using MEGA v. 7.0.26 (Kumar et al., 2016). Forward and reverse sequences for individual segments were combined by alignment using MUSCLE, followed by combining the two overlapping segments. Low-quality ends were trimmed and a BLAST search against the NCBI nr database was performed. Sequence information from the BLAST match for 26 related taxa was incorporated into a phylogenetic comparison. These additional Trichuris spp. included in the analyses infect dogs, humans, pigs, sheep, and other rodents (murids, cricetids, and arvicolids). Sequences were aligned using MUSCLE, ends were trimmed, and 1,644 base pairs remained. The newly generated sequences were submitted to the GenBank database under accession numbers MT071351, MT071352, MT071353, and MT071354.
Phylogenetic analyses were performed in MEGA v. 7.0.26 (Kumar et al., 2016) and in BEAST2 v. 2.6.1 (Bouckaert et al., 2019). MEGA determined that Kimura 2-parameter with invariant sites and a gamma distribution was the best fit substitution model for this data based on Bayesian information criterion. An evolutionary history was inferred based on this model using the maximum likelihood (ML) method and a consensus tree was generated using 1,000 bootstrapping replicates in MEGA. The Bayesian inference (BI) analysis was prepared in BEAUti v. 2.6.0 (Bouckaert et al., 2019) and performed in BEAST2 v. 2.6.0 (Bouckaert et al., 2019). The analysis used the HKY substitution model with equal frequencies (K2P + I + G is not available in BEAST2) and ran for 1×107 generations. Tracer v 1.7.1 (Rambaut et al., 2018) was used to evaluate convergence and ensure that effective sample size values for each parameter were met (all > 1,000). Tree files were combined in LogCombiner v. 2.6.0 (Bouckaert et al., 2019) and a maximum clade credibility (MCC) tree was constructed using TreeAnnotator v. 2.6.0 (Bouckaert et al., 2019) with posterior probabilities limited 50% and a burn-in percentage of 10%. FigTree v. 1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/) was used to visualize the MCC tree.
Results and discussion
Trichuris, Trichinella, and capillariid species represented highly supported (100%) monophyletic groups (Figs. 1, 2). The capillariid and trichurid clades formed sister taxa with 100% support in both analyses. These findings are consistent with previous studies (Feldman and Ramirez, 2014; Borba et al., 2019). Our new sequences fell within the Trichuris clade with 100% node support and formed an independent subclade with 100% support in both analyses. Four subclades were present within Trichuris: (i) T. discolor, T. ovis, T. skrjabini, T. leporis, (ii) T. trichiura, T. suis, unidentified Trichuris sp. (= T. colobae, see Cutillas et al., 2014), (iii) T. vulpis, T. muris, T. arvicolae, and (iv) the new sequences from T. fossor. The composition of previously studied species in subclades 1 to 3 are consistent with results from studies that used nuclear, mitochondrial, and/or concatenated data (Callejón et al., 2013, 2015; Feldman and Ramirez, 2014; Doležalová et al., 2015). In the ML analysis, the relatedness of the Trichuris subclades to one another had low (< 70%) support or were unresolved (Fig. 1). The BI analysis offered better resolution among trichurids (Fig. 2). The T. fossor subclade was most closely related to the T. arvicolae, T. muris, and T. vulpis subclade; the remaining two subclades were more closely related to one another than to the other two subclades.
Figure 1:
Bootstrap consensus tree (1,000 replicates) generated using the maximum likelihood method based on the K2P + I + G model. Bootstrap support values of 70% or greater are indicated next to nodes. Hosts are included in parentheses next to the T. fossor sequences.
Figure 2:
Maximum clade credibility consensus tree generated using the Bayesian inference method under the HKY model. Posterior probabilities for nodes above 70% are displayed. Hosts are included in parentheses next to the T. fossor sequences.
The results of the phylogenetic analyses verify that, based on molecular data, T. fossor is a distinct species. This is the first report of T. fossor from a T. mazama host. Trichuris fossor from T. bulbivorus host was an outgroup to other T. fossor specimens in both analyses. In the BI analysis, T. fossor from T. bottae and T. talpoides were sister taxa with the specimen from T. mazama as an outgroup. This suggests that variability likely exists among T. fossor from different host species, but 18S DNA is not reliable for determining whether genetic distances among Trichuris fall within the range of intraspecific variation (Guardone et al., 2013).
This work represents a preliminary step in investigating the phylogeny of T. fossor. Examining more molecular data and including different genes will likely show increased resolution of the closest relatives of T. fossor and provide a more comprehensive view of this phylogeny. Comparing other markers of nuclear and organellar DNA may be helpful (Doležalová et al., 2015) as well as examining mitochondrial data (e.g. cox1 gene), especially given there is support that it is more reliable when separating closely related species than that of 18S rRNA (Guardone et al., 2013). Incorporating T. fossor from different host species and from different geographic areas will also be valuable as lineages within the T. fossor subclade could be uncovered (Callejón et al., 2010).
