Morphology and ASAP analysis of the important zoonotic nematode parasite Baylisascaris procyonis (Stefahski and Zarnowski, 1951), with molecular phylogenetic relationships of Baylisascaris species (Nematoda: Ascaridida)

Abstract Abstract Species of Baylisascaris (Nematoda: Ascarididae) are of great veterinary and zoonotic significance, owing to cause Baylisascariosis or Baylisascariasis in wildlife, captive animals and humans. However, the phylogenetic relationships of the current 10 Baylisascaris species remain unclear. Moreover, our current knowledge of the detailed morphology and morphometrics of the important zoonotic species B. procyonis is still insufficient. The taxonomical status of B. procyonis and B. columnaris remains under debate. In the present study, the detailed morphology of B. procyonis was studied using light and scanning electron microscopy based on newly collected specimens from the raccoon Procyon lotor (Linnaeus) in China. The results of the ASAP analysis and Bayesian inference (BI) using the 28S, ITS, cox1 and cox2 genetic markers did not support that B. procyonis and B. columnaris represent two distinct species. Integrative morphological and molecular assessment challenged the validity of B. procyonis, and suggested that B. procyonis seems to represent a synonym of B. columnaris. Molecular phylogenetic results indicated that the species of Baylisascaris were grouped into 4 clades according to their host specificity. The present study provided new insights into the taxonomic status of B. procyonis and preliminarily clarified the phylogenetic relationships of Baylisascaris species.

The raccoon roundworm B. procyonis, is widely distributed in North America, and can cause severe clinical disease in humans and animals, due to extensive larval migration through host tissues (Kazacos and Boyce, 1989;Kazacos, 2001;Gavin et al., 2002Gavin et al., , 2005;;Graeff-Teixeira et al., 2016).Although the morphological characters of B. procyonis have been reported by some previous studies (Stefanski and Zarnowski, 1951;Hartwich, 1962;Sprent, 1968;Overstreet, 1970;Kikuchi and Oshima, 1977;Kazacos and Turek, 1982;Snyder, 1989), our current knowledge of the detailed morphology and morphometrics of B. procyonis is still insufficient.Furthermore, in the genus Baylisascaris, B. procyonis is highly similar to B. columnaris morphologically and genetically.Some recent studies based on different genetic data considered that B. procyonis and B. columnaris are closely related, but distinct species (Franssen et al., 2013;Choi et al., 2017); however, the results of other molecular studies did not support the current species partition of the 2 species (Camp et al., 2018).The hypothesis that B. procyonis and B. columnaris represent 2 separate species still needs to be further tested using different methods or based on different genetic data and broader samples collected from different localities.
In the present study, several adults of B. procyonis were collected from the raccoon Procyon lotor (Linnaeus) (Mammalia: Carnivora) in the Zoo of Kunming, Yunnan Province, China.The detailed morphology of B. procyonis was further studied using light and scanning electron microscopy.The ASAP (Assemble Species by Automatic Partitioning) analysis and Bayesian inference (BI) were employed for delimitation of B. procyonis and B. columnaris based on different nuclear [large ribosomal DNA (28S) and internal transcribed spacer (ITS)] and mitochondrial [cytochrome c oxidase subunit 1 (cox1) and 2 (cox2)] genetic markers.Moreover, to evaluate the evolutionary relationships of Baylisascaris species, phylogenetic analyses including the most comprehensive taxa sampling of Baylisascaris to date, were performed based on the ITS and 28S + ITS + cox1 + cox2 sequence data using maximum likelihood (ML) and Bayesian inference (BI) methods, respectively.

Specimen collection and morphological study
Single Procyon lotor died naturally in the Zoo of Kunming, Yunnan Province, China, which was opportunistically dissected for parasites.Some nematode specimens were isolated from the small intestine of this raccoon.Nematodes were fixed and stored in 70% ethanol until the study.For light microscopy studies, nematodes were cleared in glycerine for examination using a Nikon® optical microscope.For scanning electron microscopy (S.E.M.), specimens were re-fixed in a 4% formaldehyde solution, post-fixed in 1% O s O 4 , dehydrated via an ethanol series (50, 70, 80, 90, 100, 100%) and acetone (100%), and then critical point dried.Samples were coated with gold at about 20 nm and examined using a Hitachi S-4800 scanning electron microscope at an accelerating voltage of 20 kV.Measurements (range, followed by mean in parentheses) are given in millimetres (mm) unless otherwise stated.Voucher specimens were deposited in College of Life Sciences, Hebei Normal University, Hebei Province, China.
PCR products were checked on GoldView-stained 1.5% agarose gels and purified with Column PCR Product Purification Kit (Shanghai Sangon, China).Sequencing for each sample was carried out on both strands.Sequences were aligned using ClustalW2.The DNA sequences obtained herein were compared (using the algorithm BLASTn) with those available in the National Centre for Biotechnology Information (NCBI) database (http://www.ncbi.nlm.nih.gov).The 28S, ITS, cox1 and cox2 sequences data of Baylisascaris procyonis were deposited in the GenBank (http://www.ncbi.nlm.nih.gov).

Species delimitation
The ASAP method (Puillandre et al., 2021) and Bayesian inference were used for species delimitation of Baylisascaris procyonis and B. columnaris based on the 28S, ITS, cox1, and cox2 sequences, respectively.The BI trees were inferred using MrBayes 3.2.7 (Ronquist et al., 2012) under the K81UF model for cox2, MTART + I + F for ITS, HKY + F + I for 28S and TRN for cox1 (two parallel runs, 1 000 000 generations).Baylisascaris laevis was chosen as the out-group.The ASAP analyses was conducted using the ASAP online server (https://bioinfo.mnhn.fr/abi/public/asap) under the Kimura (K80) ts/tv model.The results of ASAP with lowest scores were considered as the optimal group number, except the optimal results recommended by ASAP.

Phylogenetic analyses
Phylogenetic analyses were performed based on the ITS and ITS + 28S + cox1 + cox2 sequence data using maximum likelihood (ML) inference with IQTREE v2.1.2(Minh et al., 2020) 1.The nucleotide sequences were aligned in batches using MAFFT v7.313 under iterative refinement method of E-INS-I (Katoh and Standley, 2013), poorly aligned regions were excluded using BMGE v1.12 (h = 0.4) (Criscuolo and Gribaldo, 2010).Furthermore, partially ambiguous bases were manually inspected and removed.Substitution models were compared and selected according to the Bayesian Information Criterion by using ModelFinder (Kalyaanamoorthy et al., 2017).The HKY + F + I and HKY + I model were identified as the optimal nucleotide substitution model for the ML and BI inference of ITS sequences.The partitioning schemes and the optimal nucleotide substitution model selected for each combination of partition for the ML and BI inference of ITS + 28S + cox1 + cox2 sequences are shown in Table 2. Reliabilities for maximum likelihood inference were tested using 1000 bootstrap replications and Bayesian Information Criterion analysis was run for 5 × 10 6 MCMC generations.
In the ML tree, the bootstrap (BS) values ≥90 were considered to constitute strong nodal support, whereas BS values ≥70 and <90 were considered to constitute moderate nodal support.In the BI tree, the Bayesian posterior probabilities (BPP) values ≥0.90 were considered to constitute strong nodal support, whereas BPP values ≥0.70 and <0.90 were considered to constitute moderate nodal support.The BS values ≥70 and BPP values ≥0.70 were shown in the phylogenetic trees.

Species delimitation of B. procyonis and B. columnaris
Molecular analysis of B. procyonis and B. columnaris revealed the presence of a very low level of nucleotide divergence between the 2 species in the 28S, ITS, cox1 and cox2 regions (Figures 4, 5; please see Tables 4-7 for the details).ASAP analyses of B. procyonis and B. columnaris using the 28S, ITS, cox1, and cox2 sequence data all did not support the current species partition of these 2 species (Fig. 4).Bayesian inference analyses based on the ITS, cox1, and cox2 sequence data also showed samples of B. procyonis mixed with B. columnaris (Fig. 5), which are accordant with the ASAP results.Although the result of BI analysis based on the 28S sequence data displayed B. procyonis and B. columnaris formed 2 distinct clades with weak support, the present molecular analysis revealed the presence only 1 polymorphic loci between the partial 28S region of these 2 species (Table 4).
Phylogenetic analyses of Baylisascaris spp.
(Figures 6, 7).The results of phylogenetic analyses based on the ITS sequence data using ML and BI methods were more or less identical in topology, which displayed species of Baylisascaris divided into 4 clades (clade I, II, III and IV) (Fig. 6).Baylisascaris tasmaniensis located at the base of the phylogenetic trees representing clade I, which formed a sister relationship with the other species of Baylisascaris.Phylogenetic results based on the 28S + ITS + cox1 + cox2 sequence data using ML and BI methods were slightly different from the phylogenetic results based on the ITS data, which showed B. tasmaniensis is a sister to B. transfuga + B. schroederi + B. venezuelensis (Fig. 7).Phylogenetic relationships of these species B. laevis, B. devosi, B. potosis, B. procyonis and B. columnaris agreed well with that of the phylogenetic trees based on the ITS data (Fig. 7).
The morphology and morphometric data of the present specimens more or less agreed with the previous descriptions of B. procyonis by Stefanski and Zarnowski (1951), Hartwich (1962), Sprent (1968) and Overstreet (1970), including the lengths of body and oesophagus, the morphology of lips and lateral alae, the morphology and length of spicules, the number and arrangement of precloacal papillae, the morphology of cloacal ornamentations, the relative position of vulva, the size of eggs, and the length of tail (See Table 1 for details).In addition, our specimens are also collected from the type of host P. lotor.Consequently, we considered our specimens to be B. procyonis.However, the length of spicules in our specimens is slightly  shorter than that in some previous descriptions (Stefanski and Zarnowski, 1951;Hartwich, 1962; Sprent, but it is accordant with Overstreet's (1970) description.Moreover, Stefanski and Zarnowski (1951) and Overstreet (1970) both considered the tail of male with small spike-like tip.However, we did not observe that in our specimens using S.E.M.The morphology of B. procyonis is very similar to B. columnaris, and some previous phylogenetic studies also supported B. procyonis and B. columnaris have a close affinity (Franssen et al., 2013;Tokiwa et al., 2014;Mata et al., 2016;Camp et al., 2018;Sharifdini et al., 2021).Overstreet (1970) considered the morphology of lips and first pair of postcloacal papillae to be important characters for distinguishing B. procyonis from B. columnaris (lips possessing remarkable medio-apical notch and the first pair of postcloacal papillae being double in B. procyonis vs the medio-apical notch of lips unconspicuous and the first pair of postcloacal papillae not being double in B. columnaris).The present S.E.M. observations confirmed the presence of small obtusely triangular medio-apical notch on each lip and first pair of postcloacal papillae being double in B. procyonis.However, the S.E.M. observations of B. procyonis by Kazacos and Turek (1982) found the absence of medio-apical notch on the lips of their specimens from P. lotor.Kikuchi and Oshima (1977) observed the detailed morphology of B. columnaris based on specimens collected from a skunk using S.E.M., and revealed the presence of small obtusely triangular medio-apical notch on the lips.Additionally, the number and morphology of postcloacal papillae of ascaridoid nematodes often vary between different individuals collected from a same or different hosts, even between the rows of postcloacal papillae in single individual (Uni and Takada, 1981;Li et al., 2016;Zhao et al., 2017).For example, some previous studies reported the presence of 2 closely associated single papillae instead of the postcloacal double papillae in their material of B. transfuga (Baylis and Daubney, 1922;Okoshi et al., 1961;Tenora et al., 1989).Snyder (1989) also pointed out that the first pair of postcloacal papillae of some of his specimens of B. procyonis was not double papillae and appeared as 2 single closely associated papillae based on S.E.M. observations.Moreover, Kikuchi and Oshima (1977) confirmed the first pair of postcloacal papillae of B. columnaris was double papillae in their material.Consequently, it is not reasonable and reliable to use the absence or presence of medio-apical notch on lip and the  Franssen et al. (2013) and Choi et al. (2017) reported some loci of nucleotide polymorphisms in both mitochondrial (i.e.cox1, cox2, ND2, and several tRNA genes) and nuclear markers (ITS) and considered that these loci of nucleotide polymorphisms could be used as a tool to differentiate B. procyonis from B. columnaris.However, the present results of ASAP and BI analyses do not support that B. procyonis and B. columnaris represent 2 distinct species, and indicated that these loci of nucleotide polymorphisms in the 28S, ITS, cox1, and cox2 regions do not represent fixed differences between B. procyonis and B. columnaris.The validity of the important zoonotic species B. procyonis previously supported by morphological features (Hartwich, 1962;Sprent, 1968;Overstreet, 1970) and molecular data (Franssen et al., 2013;Choi et al., 2017) was challenged by the present study.Our results are consistent with the previous study (Camp et al., 2018).Due to the current morphological studies and genetic analyses of B. procyonis and B. columnaris, it is possibly premature to treat B. procyonis as a synonym of B. columnaris.A more rigorous study integrating compared morphological study and molecular analyses with broader samples of B. procyonis and B. columnaris collected from different localities and hosts worldwide is required to solve the taxonomical status of these 2 species.
Among the 9 species of Baylisascaris, B. tasmaniensis showed sister relationship to the remaining Baylisascaris in the phylogenetic trees based on the ITS data, which are accordant with the previous phylogenetic studies (Sharifdini et al., 2021; Barrera et al.,

2022
), but are in conflict with other phylogeny (Camp et al., 2018).The phylogenetic relationships of B. tasmaniensis and the others Baylisascaris spp.can be easily understood, because B. tasmaniensis only parasitises the marsupial carnivores in Australia (i.e.Sarcophilus harrisii, Dasyurus viverrinus and Dasyurops maculatus) and possesses some particular morphological features (i.e. the presence of fan-shaped lips, 3 pairs of postcloacal double papillae and slender spicules) (Sprent, 1970).Phylogenetic construction based on the ITS and 28S + ITS + cox1 + cox2 sequence data all supported these 3 species B. transfuga, B. schroederi and B. venezuelensis all parasitic in the ursid hosts, have a close affinity, that is identical to the previous studies (Mata et al., 2016;Sharifdini et al., 2021).Barrera et al. (2022) showed the systematic status of B. laevis based on molecular phylogeny for the first time.Baylisascaris laevis is the only known species in this genus to parasitise the rodent definitive hosts, e.g.
Asterisks indicate the sequences obtained herein.
Asterisks indicate the sequences obtained herein.
Asterisks indicate the sequences obtained herein.

Figure 4 .
Figure 4. Assemble species by automatic partitioning (ASAP) analyses of Baylisascaris procyonis and B. columnaris based on 4 different nuclear and mitochondrial genetic markers.Abbreviations: cox1, cytochrome c oxidase subunit I; cox2, cytochrome c oxidase subunit II; ITS, internal transcribed spacer; 28S, large ribosomal subunit; OG, out-group;.Asterisk indicated the genetic data of samples obtained in the present study.

Figure 5 .
Figure 5. Bayesian inference analyses of Baylisascaris procyonis and B. columnaris based on 4 different nuclear and mitochondrial genetic markers, respectively.Bayesian posterior probabilities values ≥0.70 were shown on nodes.Asterisk indicated the genetic data of samples obtained in the present study.

Table 1 .
Clade II included B. transfuga, B. schroederi and B. venezuelensis, all parasitic in the ursid hosts.Among them, B. transfuga and B. schroederi showed closer relationship than B. venezuelensis.Clade III contained only B. laevis reported from the rodent definitive hosts.Clade IV comprising B. devosi, B. potosis, B. columnaris and B. procyonis, parasitising the mustelid and procyonid hosts.Among them, Species of Baylisascaris with detailed genetic information included in the phylogenetic analyses Parasitology B. procyonis clustered together with B. columnaris, and B. devosi is a sister to B. potosis.

Table 2 .
The partitioning schemes and the optimal model selected for each combination of partition for the ML and BI inference based on the ITS + 28S + cox1 + cox2 sequences

Table 4 .
Base difference in the partial 28S region between B. columnaris and B. procyonis parasitic in the carnivorous definitive hosts including mustelids and procyonids.Our phylogenetic results also revealed B. devosi is a sister to B. potosis for the first time.However, B. procyonis and B. potosis both hosted the procyonids, did not display a close relationship (the similar situation also occurring in B. devosi and B. columnaris), that possibly indicated that it is not reasonable to use the specific groups of definitive hosts (Procyonidae or Mustelidae) as a criterion for distinguishing procyonis from B. columnaris.The present study revealed some previously unreported morphological features of B. procyonis.The results of morphological study and ASAP and BI analyses all did not support that B. procyonis and B. columnaris represent 2 distinct species and the validity of B. procyonis was challenged.Molecular phylogeny provided new insights into the evolutionary relationships of Baylisascaris spp.

Table 5 .
Base difference in the partial ITS region between B. columnaris and B. procyonis

Table 6 .
Base difference in the partial cox1 region between B. columnaris and B. procyonis

Table 7 .
Base difference in the partial cox2 region between B. columnaris and B. procyonis