Genetic analysis of a 66-kDa protein-encoding gene of Angiostrongylus cantonensis and Angiostrongylus malaysiensis

The rat lungworm Angiostrongylus cantonensis is globally known to be the cause of oeosinophilic meningitis in humans. Another congener, Angiostrongylus malaysiensis, is closely related to A. cantonensis and has been described as a potential human pathogenic parasite. These 2 worms are similar in terms of life cycle, host range and morphological and genetic information. However, there are limited studies about their genetic diversity based on the 66-kDa protein-encoding gene. The objective of this study was to explore the 66-kDa protein sequence variation of A. cantonensis and A. malaysiensis collected from Thailand. Two adult and 53 third-stage larval specimens of Angiostrongylus from 4 geographic locations in Thailand were molecularly identified using the 66-kDa protein gene. The phylogenetic trees (Bayesian inference tree and maximum-likelihood tree) showed that Angiostrongylus formed a monophyletic clade with a clear separation between A. cantonensis and A. malaysiensis. The genetic distance between A. cantonensis and A. malaysiensis varies from 0.82 to 2.86%, with a total of 16 variable sites. The analysis of genetic diversity revealed 1 and 5 new haplotypes of A. cantonensis and A. malaysiensis, respectively, and showed genetic differences between the populations of A. cantonensis and A. malaysiensis. The haplotype networks of A. cantonensis and A. malaysiensis populations in Thailand are similar to those of populations in some countries, indicating the range expansion of genomic origin between populations in different areas. In conclusion, the 66-kDa protein gene was a good genetic marker for studying genetic diversity and discriminating between A. cantonensis and A. malaysiensis.


Introduction
Angiostrongylus Kamensky, 1905 or the lungworm is a parasitic nematode in the superfamily Metastrongyloidea (Spratt, 2015). To date, over 20 species of this genus have been reported around the world, and Angiostrongylus cantonensis and Angiostrongylus costaricensis have been reported to be causative agents of neurological and abdominal angiostrongyliasis in humans, respectively (Spratt, 2015;Barratt et al., 2016). Another species found in Asian countries, Angiostrongylus malaysiensis, is also a potential human pathogenic parasite (Ansdell and Wattanagoon, 2018). In Thailand, A. cantonensis and A. malaysiensis are the most common species . Angiostrongylus cantonensis is a well-known pathogen that causes oeosinophilic meningitis associated with angiostrongyliasis in humans in Thailand, whereas A. malaysiensis is increasingly reported in many provinces in the country (Eamsobhana, 2013;Watthanakulpanich et al., 2021).
Angiostrongylus cantonensis and A. malaysiensis are similar in terms of life cycle, host range, host habitat and morphological and genetic information (Bhaibulaya, 1979;Eamsobhana et al., 2015;Chan et al., 2020). These 2 species of Angiostrongylus can infect the same species of definitive and intermediate hosts (Bhaibulaya and Techasoponmani, 1972). In addition, mixed infections of A. cantonensis and A. malaysiensis in snail intermediate hosts and rodent definitive hosts have been widely reported (Eamsobhana et al., 2016;Watthanakulpanich et al., 2021). This may lead to difficulty in discriminating between A. cantonensis and A. malaysiensis. Adults of A. cantonensis and A. malaysiensis can be morphologically differentiated by the minute protrusion at the posterior end of females and the bursal rays of males (Bhaibulaya, 1979;Thiengo et al., 2010). Nevertheless, the morphological variance between the 2 species can confound identification. Moreover, differences between the morphological characteristics of the larval stages, especially the infective stage, have not yet been clarified. Therefore, the identification of Angiostrongylus species based on morphological characters is difficult due to vague and similar descriptions of size and body shapes among species (Robles et al., 2008;Monte et al., 2014).
Molecular phylogeography analysis of A. cantonensis and A. malaysiensis may provide insight into specific genetic variation and population formation (Avise, 2000). The phylogeography based on COI sequences of A. cantonensis from Thailand, Taiwan, China and Japan revealed that the geographical distribution of A. cantonensis probably reflects multiple independent origins that were likely to have been influenced by human activities (Tokiwa et al., 2012). Likewise, Monte et al. (2012) analysed the phylogenetic relationship of COI sequences for A. cantonensis and revealed that some haplotypes from Brazil clustered with isolates from Asia, while the rest formed distinctly divergent clades, indicating multiple origins of A. cantonensis in Brazil. In addition, the phylogeny based on the 66-kDa protein gene revealed distinct clades among A. costaricensis, A. cantonensis and A. malaysiensis. However, no clear separation of the conspecific taxa between A. cantonensis and A. malaysiensis from different geographical regions was reported. Greater sample sizes of the conspecific taxa from each locality may provide a conclusive inference of distinct phylogeographic patterns (Eamsobhana et al., 2019). Moreover, there have been no reports on the molecular identification of third-stage larvae of A. cantonensis and A. malaysiensis using the 66-kDa protein gene. Therefore, we further analysed the genetic diversity of A. cantonensis and A. malaysiensis in Thailand using 66-kDa protein gene sequences. The phylogeny and haplotype network of a 66-kDa protein gene in A. cantonensis and A. malaysiensis were also analysed to determine the relationship between hosts and parasites.

Angiostrongylus worms
The total 55 Angiostrongylus samples consisted of 14 specimens of A. cantonensis and 41 specimens of A. malaysiensis. Within the 14 specimens of A. cantonensis, 2 female worms were collected from a definitive rodent host (Bandicota sp.; n = 1) in Kamphaeng Phet province, central Thailand, and 12 third-stage larvae (L3) were previously collected from an intermediate land snail host (Achatina fulica; n = 30) from Chaiyaphum province, northeastern Thailand (Dumidae et al., 2019). Of the 41 A. malaysiensis samples, 38 L3 specimens were collected from 21 A. fulica in Chiang Rai province, and the remaining 3 L3 specimens were collected from 1 A. fulica in Phrae province in northern Thailand (Dumidae et al., 2019), as shown in Fig. 1. Adult worms were fixed in absolute alcohol and stored at −20°C until DNA extraction. The genomic DNA samples of Angiostrongylus larvae were stored at −20°C.

DNA extraction
Before performing the DNA extraction, adult worms were dried by placing on filter paper for 15 min at room temperature. Individual worms were excised into small pieces, which were then placed in 1.5 mL microcentrifuge tubes, crushed and digested in lysis buffer. Genomic DNA extraction was performed using the NucleoSpin® Tissue kit (Macherey-Nagel, Duren, Germany) according to the manufacturer's protocol. Genomic DNA samples of Angiostrongylus larvae that had been collected previously (Dumidae et al., 2019) from A. fulica were used as samples in this study. The genomic DNA was checked by running it on a 0.8% agarose gel in 1× TBE buffer at 100 V. The gel was stained with ethidium bromide, destained with distilled water and photographed under ultraviolet light. The DNA solution was stored at −20°C prior to further processing.

Polymerase chain reaction (PCR) and sequencing
The DNA fragment (300 bp) of the 66-kDa protein gene was amplified using primers AC1 5 ′ -CTCGGCTTAATCTTTGCGAC-3 ′ and AC2 5 ′ -AACGAGCGGCAGTAGAAAAA-3 ′ (Eamsobhana et al., 2019). The PCR components (30 μL final reaction volume) contained 15 μL of EconoTaq® PLUS 2× Master mix (1×; Lucigen Corporation, Middleton, WI, USA), 1.5 μL of each primer at 5 μM (0.25 μM), 9 μL of distilled water and 3 μL of the DNA template (20-200 ng). The PCR cycle was initial denaturation at 94°C for 3 min, followed by 35 cycles of denaturation at 94°C for 2 min, annealing at 58°C for 1 min and extension at 72°C for 3 min, with a final extension at 72°C for 7 min. All PCRs were performed in a Biometra TOne Thermal Cycler (Analytik Jena AG, Jena, Germany). The amplified products were analysed using 1.2% agarose gel electrophoresis. Purification of the PCR products was performed using a NucleoSpin® Gel and PCR Clean-Up Kit (Macherey-Nagel, Germany) following the manufacturer's instructions. The purified PCR products were run on a 1.2% agarose gel at 100 V in 1× TBE buffer. The PCR products were sequenced in both the forward and reverse directions at Macrogen Inc., Seoul, Korea.

66-kDa protein sequences from GenBank
The nucleotide sequences of a 66-kDa protein-encoding gene of A. cantonensis and A. malaysiensis from Thailand, China, Japan, Malaysia and the United States downloaded from GenBank were included in the present study (Table S1). In addition, the sequences from Ancylostoma caninum and Heterorhabditis bacteriophora were used as outgroup taxa.

Sequence and phylogenetic analysis
All nucleotide sequences were edited and assembled using Seq-Man II software (DNASTAR, Madison, WI, USA). Subsequently, multiple-sequence alignment with ClustalW and trimmed sequences was performed in MEGA version 7.0 (Kumar et al., 2016). The 66-kDa protein sequence was blasted in the GenBank database (http://blast.ncbi.nlm.nih.gov/Blast.cgi) to confirm species identification of Angiostrongylus (A. cantonensis and A. malaysiensis).
Phylogenetic trees were constructed via the maximumlikelihood (ML) and Bayesian inference (BI) methods. The ML tree with the Tamura-Nei model (Tamura and Nei, 1993) was generated via 1000 bootstrap replicates in MEGA version 7.0 (Kumar et al., 2016). The BI tree was constructed using the MrBayes 3.2.0 program (Ronquist et al., 2012). The Bayesian posterior probabilities (BPPs) were estimated using Markov chain Monte Carlo analysis, which was run for 10 000 000 generations with data sampling every 500 generations, discarding the first 1000 sampled trees as burn-in (Monte et al., 2012). The final phylogenetic trees were viewed and edited in FigTree v.1.4. The nucleotide variation and P distance of A. cantonensis and A. malaysiensis were calculated using the resultant alignment in MEGA version 7.0.

Genetic analysis
The sequences of a 66-kDa protein-encoding gene obtained in the present study together with the sequences downloaded from GenBank were grouped into 3 datasets to analyse the genetic population and the transmission relationships between the host and parasite. Group 1 included all sequences of A. cantonensis (n = 53) and A. malaysiensis (n = 64) populations from different countries. Group 2 consisted of all sequences of A. cantonensis (n = 32) and A. malaysiensis (n = 60) from different regions of Thailand, and group 3 included all currently available sequences of A. cantonensis and A. malaysiensis isolated from intermediate land snail hosts and definitive rodent hosts from different countries.
The genealogical relationships were estimated by using a haplotype network constructed in Network 5.0.1.1 (http://www. fluxus-engineering.com) based on the median-joining algorithm (Bandelt et al., 1999). Apart from the A. cantonensis and A. malaysiensis 66-kDa protein-encoding sequences in this study, we also included the nucleotide sequences of these worms deposited in GenBank by Eamsobhana et al. (2019). The haplotype nomenclature used in this study was the same as that used by Eamsobhana et al. (2019).
Genetic diversity indices, e.g. haplotype number, segregating sites, haplotype diversity and nucleotide diversity, were computed and generated by DnaSp version 5 (Librado and Rozas, 2009) and ARLEQUIN version 3.5.1.2 (Excoffier and Lischer, 2010). Genetic differentiation between Angiostrongylus from different rodent host species was investigated by comparing genetic divergence within rodent host species based on the K2P model in MEGA version 7.0 (Kumar et al., 2016).

Molecular identification of Angiostrongylus spp.
The molecular identification of Angiostrongylus spp. based on the 66-kDa protein gene was congruent with the morphological identification. PCR-based analysis and sequencing of the 66-kDa protein gene were performed together with a BLASTN search. Fourteen samples (245 bp) of Angiostrongylus (GenBank accession nos. OM280392-

Phylogenetic analyses
The phylogenetic trees of A. cantonensis (53 sequences) and A. malaysiensis (64 sequences) were reconstructed using the BI and ML methods. Both methods revealed congruent topologies. Therefore, here we show the BI tree with posterior probabilities  and only the bootstrap values from ML analyses. The phylogenetic trees based on 245 bp of the 66-kDa protein gene showed that Angiostrongylus formed a monophyletic clade with a clear separation between A. cantonensis and A. malaysiensis (Fig. 2). The interspecific distances between the A. cantonensis and A. malaysiensis sequences ranged from 0.82 to 2.86%, with a total of 16 variable sites found (Table 1).

Genetic variation of A. cantonensis and A. malaysiensis populations from different countries
Based on previous studies of A. cantonensis, 13 haplotypes (Ac66-1 to Ac66-13) were classified, and sequences were deposited in GenBank, i.e. haplotypes Ac66-1 to Ac66-4, and Ac66-12 consisted of sequences covering Thailand, China, Japan and the United States, and haplotypes Ac66-8 to Ac66-11 were found only in the United States. In the current haplotype network analyses, our 14 sequences together with 39 sequences from previous studies retrieved from GenBank revealed 14 haplotypes (Ac66-1 to Ac66-14) (Fig. 3). Among the 14 sequence samples obtained in the present study, 9 and 4 sequences belonged to haplotypes Ac66-1 and Ac66-2, respectively. In addition, 1 sequence obtained in the present study was identified as a new haplotype named Ac66-14. A comparison of nucleotide sequences between this new haplotype and 13 previously reported haplotypes is presented in Table 1. The genetic distances among the haplotypes varied from 0 to 0.016 (Table 2). Of these, 10 haplotypes were unique (Ac66-3 to Ac66-9, Ac66-11, Ac66-13 and Ac66-14), and 4 were shared by at least 2 populations (Ac66-1, Ac66-2, Ac66-10 and Ac66-12). The most widely distributed haplotype, Ac66-1, was shared among samples from Thailand, Japan and the United States. Haplotype Ac66-2 was shared between samples from Thailand and Japan. Haplotype Ac66-10 was shared between samples from China and the United States. Haplotype Ac66-12 was shared between samples from Thailand and China. The haplotype diversity in each population ranged from 0.7117 in Thailand to 0.9333 in the United States, with an average of 0.7903. The nucleotide diversity in each population ranged from 0.0035 in China to 0.0076 in the United States, with an average of 0.0054 (Table 3).
Bold letters indicate the new haplotypes obtained in the present study.
Thirteen haplotypes (Am66-1 to Am66-13) of A. malaysiensis were identified from 41 sequences obtained in the present study together with 23 sequences from GenBank. Among the 41 samples from the present study, 10 sequences belonged to haplotype Am66-1, and 31 sequences belonged to the 5 new haplotypes, including Am66-9 (19 sequences), Am66-10 (9 sequences), Am66-11 (1 sequence), Am66-12 (1 sequence) and Am66-13 (1 sequence). A comparison of nucleotide sequences between the 5 new haplotypes identified in the present study and 8 previously reported haplotypes is presented in Table 1. The genetic distances between the haplotypes varied from 0 to 0.016 (Table 4). Of these, 12 haplotypes were unique (Am66-2 to Am66-13), and 1 (Am66-1) was shared between samples from Malaysia and Thailand (Fig. 3). The haplotype diversity in each population ranged from 0 in Malaysia to 0.8096 in Thailand, with an average of 0.7981. The nucleotide diversity in each population ranged from 0 in Malaysia to 0.0056 in Thailand, with an average of 0.0054 (Table 3).

Genetic variation of A. cantonensis and A. malaysiensis isolated from different regions of Thailand
The analysis of 32 A. cantonensis sequences from Thailand identified 6 haplotypes (Ac66-1 to Ac66-4, Ac66-12 and Ac66-14). Among the A. cantonensis haplotypes in Thailand, Ac66-1 was the most common and was widely distributed in 4 regions (central, north, northeast and south). Another common haplotype was Ac66-2, which was found in several geographical localities in 3 regions (central, northeast and south), and haplotype Ac66-4 was found in the northeast and south regions of Thailand. The remaining haplotypes were found to be unique in a particular isolate, such as haplotype Ac66-3, which was found only in the central isolate, Ac66-12, which was unique to the western isolate, and Ac66-14, which was found only in the northeast isolate (Fig. 4). The haplotype diversity in each population ranged from 0 in the population from the north and west regions to 0.7333 in the population from the south, with an average of 0.6613. The nucleotide diversity in each population ranged from 0 in the north and west regions to 0.8205 in the northeast region, with an average of 0.0034 (Table 5).
Thirteen haplotypes (Am66-1 to Am66-13) of A. malaysiensis were identified from 60 Thailand sequences. Among the A. malaysiensis haplotypes, Am66-4 was the most common and was widely distributed in 4 regions (north, northeast, south and west). Another common haplotype was Am66-1, which was found in several geographical localities in 3 regions (central, north and west), and haplotype Am66-6 was found in isolates from the central and south regions of Thailand. The remaining haplotypes were found to be unique in a particular isolate, such as haplotypes Am66-2, Am66-3 and Am66-8, which were found only in the northeast isolate; Am66-5 and Am66-7, which were unique to the south isolate and Am66-9 to Am66-13, which were present only in the north isolate (Fig. 4). The haplotype diversity in each population ranged from 0.6667 in the population from the west region to 1.0000 in the population from the central and northeast regions, with an average of 0.8096. The nucleotide diversity in each population ranged from 0.0027 in the west region to 0.0082 in the central and northeast regions, with an average of 0.0056 (Table 5).

Genetic variation of A. cantonensis and A. malaysiensis isolated from snail intermediate and rodent definitive hosts
This group consisted of 53 and 64 sequences that were isolated from intermediate land snail hosts and definitive rodent hosts, respectively. In the land snail A. fulica, 12 A. cantonensis and Fig. 3. Median-joining haplotype networks of A. cantonensis and A. malaysiensis from Thailand and other geographical regions inferred from 66-kDa protein sequences. Each haplotype is represented by a circle, and circle sizes are proportional to haplotype frequency. Colours indicate the geographic origin of the haplotypes. Each mutation between haplotypes is represented by a bar. Median vectors (small red dots) represent ancestral haplotypes that are either not sampled or missing haplotypes. 41 A. malaysiensis sequences were classified into 3 haplotypes (Ac66-1, Ac66-2 and Ac66-14) and 6 haplotypes (Am66-1, Am66-9 to Am66-13), respectively (Fig. 5). Among the A. cantonensis haplotypes, 3 haplotypes (Ac66-1, Ac66-2 and Ac66-14) were found in A. fulica in Chaiyaphum province of Thailand. Of these, only 2 haplotypes (Ac66-1 and Ac66-2) were found in rodent hosts from different countries, such as haplotype Ac66-1 found in isolates from Thailand, Japan and the United States, and Ac66-2 was present in isolates from Thailand and Japan. In A. malaysiensis, Am66-1 was the most widely distributed in A. fulica in 2 provinces (Chiang Rai and Phrae) in northern Thailand. In addition, this haplotype was found in rodent hosts from the central province (Bangkok), 2 northern provinces (Chiang Rai and Mae Hong Son) and 2 western provinces (Kanchanaburi and Tak) of Thailand and in Pahang in Malaysia.
For the rodent host species, the haplotype diversity of A. cantonensis ranged from 0 in Bandicota sp. to 0.8889 in Rattus species. Nucleotide diversity ranged from 0 in Bandicota sp. to 0.0071 in Rattus norvegicus. Among the 13 haplotypes identified, 8 were unique, and 5 haplotypes were shared by at least 2 rodent host species. Rattus norvegicus possessed the highest number (5 haplotypes) of unique haplotypes (Fig. 5, Table 6). Genetic divergence within the rodent host species based on the K2P model ranged from 0 to 1.66%, with a mean of 0.47% (Table 6). The greatest withinrodent host genetic divergence (1.66%) was found in R. norvegicus. In A. malaysiensis, the haplotype diversity ranged from 0 in Bandicota indica and Rattus tiomanicus to 1.0000 in Bandicota bengalensis, Rattus exulans, Rattus losea and Rattus norvegicus. Nucleotide diversity ranged from 0 in B. indica and R. tiomanicus to 0.0082 in R. exulans. Among the 8 haplotypes identified, 5 were unique, and 3 haplotypes were shared by at least 2 rodent host species. Rattus norvegicus possessed the highest number (3 haplotypes) of unique haplotypes (Fig. 5, Table 6). Genetic divergence within the rodent host species based on the K2P model ranged from 0 to 1.24%, with a mean of 0.41% (Table 6). The greatest within-rodent host genetic divergence (1.24%) was found in R. rattus.

Discussion
The advantage of the 66-kDa protein gene as a genetic marker to determine genetic diversity and phylogeny between and within Angiostrongylus populations (A. cantonensis, A. malaysiensis and A. costaricensis) was previously noted (Eamsobhana et al., 2010(Eamsobhana et al., , 2019. In this study, identification of Angiostrongylus based on a partial sequence of a 66-kDa protein-encoding gene was confirmed by 99-100% sequence identity after BLASTN searches. Angiostrongylus cantonensis and A. malaysiensis are closely related in terms of morphological and genetic characteristics and also share similarities in their life cycles (Chan et al., 2020;Watthanakulpanich et al., 2021). Both species utilize the same definitive and intermediate host species (Bhaibulaya and Techasoponmani, 1972). In addition, mixed infections with both A. cantonensis and A. malaysiensis have been widely recorded in snail intermediate hosts and rodent definitive hosts Watthanakulpanich et al., 2021). The 66-kDa protein gene is undoubtedly suitable for discrimination between A. cantonensis and A. malaysiensis because the phylogenetic trees clearly place these 2 species into separate clades. Our findings are similar to those of previous reports (Eamsobhana et al., 2010(Eamsobhana et al., , 2019. This suggests the advantage of the 66-kDa protein sequence for the identification of A. cantonensis and A. malaysiensis. Interestingly, the use of the 66-kDa protein gene as the genetic marker was successful in discriminating the third-stage larvae of A. cantonensis and A. malaysiensis. The level of genetic divergence of the 66-kDa protein sequences between A. cantonensis and A. malaysiensis ranged     Eamsobhana et al. (2019), who reported that the genetic distance between A. cantonensis and A. malaysiensis using 66-kDa protein sequences was approximately 3.27% (Eamsobhana et al., 2019). A lower genetic divergence (0-1.0%) in the nuclear 18S rRNA gene was also noted between A. cantonensis and A. malaysiensis (Chan et al., 2020). In contrast, the genetic divergence between A. cantonensis and A. malaysiensis was relatively high based on COI (9.8-16.4%), cytb (10.9-12.2%), 12S rRNA (6.8-7.9%), 16S rRNA (7.9-10.0%) and ITS2 (15.1-15.7%) sequences (Chan et al., 2020). However, several reports have shown that the nuclear 18S rRNA gene can be used to discriminate between A. cantonensis and A. malaysiensis, which are clearly distinguished into their clades using this marker despite the low interspecies genetic distance (Fontanilla and Wade, 2008;Tokiwa et al., 2012;Rodpai et al., 2016), which is comparable to the results we obtained using the 66-kDa protein gene. Previous studies of third-stage larvae of A. cantonensis isolated from A. fulica snails in 8 provinces of Thailand using 8 random-amplified polymorphic DNA-PCR markers revealed high levels of genetic diversity and low levels of gene flow in A. cantonensis populations (Thaenkham et al., 2012). The worms from these 8 localities were divided into 2 groups with statistically significant genetic differentiation of the 2 populations: group 1 contained A. cantonensis from Chanthaburi, Chiang Mai, Khon Kaen, Narathiwat Nong Khai and Prachuap Khiri Khan provinces, and group 2 contained A. cantonensis from Kanchanaburi and Lop Buri provinces. Similarly, Vitta et al. (2016) reported third-stage larvae of A. cantonensis from freshwater and land snails in 19 distinct geographical areas of Thailand using COI sequences, revealing 2 different origins of A. cantonensis in Thailand: group 1 contained A. cantonensis from Kamphaeng Phet, Phetchabun, Tak and Thailand ac4, and group 2 contained A. cantonensis from Kalasin, Kamphaeng Phet, Phitsanulok, Tak and AC Thai. Nonetheless, haplotypes of groups 1 and 2 were found in the same areas (Kamphaeng Phet and Tak provinces). The results indicate the occurrence of restricted gene flow between localities.
A recent study reported the presence of 13 distinct 66-kDa protein sequence haplotypes (Ac66-1 to Ac66-13) from A. cantonensis in several parts of the world (Eamsobhana et al., 2019). Haplotype Ac66-1 was the most common haplotype in A. cantonensis from Thailand, Japan and the United States. Haplotype Ac66-5 was reported in Japan, while haplotypes Ac66-8, Ac66-9 and Ac66-11 were reported in the United States. Haplotype Ac66-10 was reported in China and the United States, and haplotype Ac66-13 was found only in China. In Thailand, A. cantonensis did not cluster unequivocally according to their geographical origin, as 7 haplotypes (Ac66-1 to Ac66-4, Ac66-6, Ac66-7 and Ac66-12) from 10 geographical regions of Thailand were found. The A. cantonensis haplotypes from Bangkok and Phitsanulok provinces in the central region, Surat Thani province in the south, and the Thailand laboratory strain (originating from Khon Kaen in the northeast) were variable. Moreover, 4 haplotypes were found confined to a single locality: Ac66-3 in Phitsanulok province (central region); Ac66-6 and Ac66-7 in the Thailand laboratory strain and Ac66-12 in Prachuap Khiri Khan province (west region) (Eamsobhana et al., 2019). In the present study, haplotype Ac66-1 was found in Kamphaeng Phet province (2 specimens) in the central region and Chaiyaphum province (7 specimens) in the northeast region of Thailand. Additionally, haplotype Ac66-1 was previously reported in 3 central provinces (Bangkok, Lop Buri and Phitsanulok), 2 southern Fig. 4. Median-joining haplotype networks of A. cantonensis and A. malaysiensis from different regions of Thailand inferred from 66-kDa protein sequences. Each haplotype is represented by a circle, and circle sizes are proportional to haplotype frequency. Colours indicate the geographic origin of the haplotypes. Each mutation between haplotypes is represented by a bar. Median vectors (small red dots) represent ancestral haplotypes that are either not sampled or missing haplotypes. provinces (Ranong and Surat Thani) and in the northern province (Chiang Mai) (Eamsobhana et al., 2019). This Ac66-1 haplotype was dominant in Thailand. The other 4 specimens (haplotype Ac66-2) from Chaiyaphum province in the present study were also reported from 2 central provinces (Bangkok and Samut Prakan) and the south province (Surat Thani). Importantly, 1 new haplotype (Ac66-14) was identified in 1 specimen from Chaiyaphum province of Thailand. Incorporation of the genetic data of the 66-kDa protein gene from A. cantonensis obtained in the present and previous studies (Eamsobhana et al., 2019) revealed sharing of dominant haplotypes (Ac66-1 and Ac66-2), suggesting a common origin. Based on the 66-kDa protein sequence from A. malaysiensis, previous studies reported 8 haplotypes (Am66-1 to Am66-8) in Thailand and Malaysia (Eamsobhana et al., 2019). In this study, 13 haplotypes were identified, with 5 haplotypes being new. One haplotype (Am66-1) of A. malaysiensis from Phrae (3 specimens) and Chiang Rai (7 specimens) provinces in the north region was previously reported from Thailand and Malaysia (Eamsobhana et al., 2019). Five new haplotypes (Am66-9, Am66-10, Am66-11, Am66-12 and Am66-13) were reported from 31 specimens in Chiang Rai province of Thailand. The most common haplotype detected in Thailand and Malaysia was Am66-1. In Thailand, 66-kDa protein haplotypes (Am66-1 to Am66-8) were distributed at random throughout the country, and Am66-1 was the most widely distributed in 2 northern provinces (Chiang Rai and Mae Hong Son), 2 western provinces (Kanchanaburi and Tak) and in a central province (Bangkok). In addition, the A. malaysiensis haplotypes from Bangkok in central Thailand, Mae Hong Son in north Thailand, Nong Khai in northeast Thailand, Satun in south Thailand and Tak in west Thailand showed variable 66-kDa haplotype diversity. Moreover, 5 haplotypes were found confined to a single locality: Am66-2, Am66-3 and Am66-8 in Nong Khai province (northeast region); Am66-5 in Phang Nga province (south region) and Am66-7 in Satun province (south region) (Eamsobhana et al., 2019). Therefore, it was difficult to conclude the relationship between the haplotype of A. malaysiensis and geographic areas in Thailand. A larger sample size of the 66-kDa protein-encoding gene sequence may reveal a clearer relationship between haplotypes and localization in Thailand.
High haplotype diversity in a 66-kDa protein gene for A. cantonensis (14 haplotypes) and A. malaysiensis (13 haplotypes) has also been observed in other genetic markers. Twenty cytb haplotypes have been reported for A. cantonensis from several parts of the world (Dusitsittipon et al., 2015;Yong et al., 2015;Dumidae et al., 2019). For the COI gene, 16 haplotypes were observed in A. cantonensis globally (Eamsobhana et al., 2017), and 9 haplotypes were observed in A. malaysiensis from Laos, Malaysia, Myanmar and Thailand (Rodpai et al., 2016;Eamsobhana et al., 2018). For the 12S rRNA gene of Angiostrongylus in Thailand, the average genetic variation was 0.5% with 6 haplotypes in a population of A. cantonensis (Chan et al., 2020). Similarly, using the 16S rRNA gene, an average genetic variation of 2.2% with 6 haplotypes within 1 population of A. cantonensis and an average genetic variation of 1.7% with 6 haplotypes within 4 populations of A. malaysiensis were found (Chan et al., 2020). Comparatively, we found that the 66-kDa protein gene resulted in 3 haplotypes within 2 populations of A. cantonensis (Chaiyaphum and Kamphaeng Phet provinces). Moreover, 6 haplotypes were found in 2 populations of A. malaysiensis (Phrae and Chiang Rai provinces). Higher intraspecific genetic variation levels and more haplotypes might be observed if more A. cantonensis and A. malaysiensis specimens are sampled from other localities.
To investigate the distribution of haplotypes in hosts, medianjoining networks were constructed to analyse all currently available 66-kDa protein sequences of A. cantonensis and A. malaysiensis isolated from snail intermediate hosts (A. fulica) and Fig. 5. Median-joining haplotype networks of A. cantonensis and A. malaysiensis from different land snail and rodent host species inferred from 66-kDa protein sequences. Each haplotype is represented by a circle, and circle sizes are proportional to haplotype frequency. Colours indicate the geographic origin of the haplotypes. Each mutation between haplotypes is represented by a bar. Median vectors (small red dots) represent ancestral haplotypes that are either not sampled or missing haplotypes. rodent definitive hosts from different countries. In A. cantonensis, 2 haplotypes (Ac66-1 and Ac66-2) were shared between A. cantonensis isolated from A. fulica (from Chaiyaphum province of Thailand) and rodent hosts (from Thailand, Japan and the United States). In A. malaysiensis, Am66-1 was the most widely distributed haplotype in A. fulica in 2 northern provinces (Chiang Rai and Phrae) of Thailand. In addition, this haplotype was found in rodent hosts from the central province (Bangkok), 2 northern provinces (Chiang Rai and Mae Hong Son) and 2 western provinces (Kanchanaburi and Tak) of Thailand and in Pahang in Malaysia. This could be considered the result of range expansion of genomic origin. These findings indicate that lineage-specific A. cantonensis and A. malaysiensis have been spreading across Thailand. The transmission of this nematode has been linked to the dispersal of invasive hosts (Tokiwa et al., 2012). The increased presence of A. cantonensis in the country is likely a result of the rapid spread of its intermediate host, A. fulica, contributing to the dispersion of this parasite and infection of the definitive host Dumidae et al., 2019Dumidae et al., , 2021. Invasion by A. fulica facilitates the establishment of the life cycle of the parasite and thus increases the chances for exposure of native snails to A. cantonensis in existing endemic areas. In addition, this invasive snail accelerates the spread of A. cantonensis to new areas since it rapidly expands the parasite's range (Lv et al., 2009). This phenomenon is described as one of the primary causes of the spread of oeosinophilic meningitis . Invasive snails and rodents are implicated in an increase in the distribution of A. cantonensis in Brazil , Spain (Foronda et al., 2010), China (Yang et al., 2013), Uganda (Mugisha et al., 2012), Japan (Tokiwa et al., 2013) and the United States (York et al., 2015). The hypothesis that Angiostrongylus has achieved global-scale dispersal with various organisms/hosts or vectors is largely influenced by human transportation (Monte et al., 2012;Tokiwa et al., 2012;Dusitsittipon et al., 2015). Comparisons of genetic divergence among rodent host species found unique and shared haplotypes in different rodent species. All rodent host species included in the present study shared at least 1 common 66-kDa protein haplotype. Our findings showed that 8 haplotypes were unique and 5 haplotypes were shared in A. cantonensis, whereas 5 haplotypes of A. malaysiensis were found to be unique, and 3 haplotypes were shared by at least 2 rodent host species. The high degree of haplotype uniqueness suggests that there are some limitations to the spread of genomic origin among rodent host species. Many rodent species serve as definitive hosts for A. cantonensis and A. malaysiensis and are capable of highly promoting the distribution and intraspecific transfer of this parasite (Eamsobhana et al., 2016). The limited dispersal of rodent hosts might be expected to limit the genomic origin of their worm parasites, resulting in genetic structure over small geographical scales (Pocock et al., 2005;Gardner-Santana et al., 2009). However, data on the genetic diversity of Angiostrongylus remain scarce in invaded areas (Simões et al., 2011;Monte et al., 2012;Moreira et al., 2013;Dalton et al., 2017). The pathogenicity of A. cantonensis against laboratory hosts varies between different A. cantonensis genetic strains (Lee et al., 2014). Whether this dominant haplotype presents any difference in pathogenicity compared to other haplotypes remains to be clarified. Further studies are needed to clarify whether this haplotype has a different pathogenicity that may contribute to its evolution.

Conclusion
This study further confirmed the presence of 66-kDa protein genetic diversity in various geographical isolates of A. cantonensis and A. malaysiensis. We demonstrate the utility of the 66-kDa protein-encoding gene as a genetic marker for species discrimination in the larval stage, as it clearly discriminated A. cantonensis and A. malaysiensis into separate clades. Importantly, we identified 1 and 5 new haplotypes from A. cantonensis and A. malaysiensis, respectively. Our findings revealed that the 66-kDa protein gene has sufficient intraspecific genetic variation to be considered a genetic marker for future Angiostrongylus population-level studies.
Supplementary material. The supplementary material for this article can be found at https://doi.org/10.1017/S0031182022001573 Data. All relevant data are within the paper and its supporting information file.