Examination of Sarcocystis spp. of giant snakes from Australia and Southeast Asia confirms presence of a known pathogen – Sarcocystis nesbitti

We examined Sarcocystis spp. in giant snakes from the Indo-Australian Archipelago and Australia using a combination of morphological (size of sporocyst) and molecular analyses. We amplified by PCR nuclear 18S rDNA from single sporocysts in order to detect mixed infections and unequivocally assign the retrieved sequences to the corresponding parasite stage. Sarcocystis infection was generally high across the study area, with 78 (68%) of 115 examined pythons being infected by one or more Sarcocystis spp. Among 18 randomly chosen, sporocyst-positive samples (11 from Southeast Asia, 7 from Northern Australia) the only Sarcocystis species detected in Southeast Asian snakes was S. singaporensis (in reticulated pythons), which was absent from all Australian samples. We distinguished three different Sarcocystis spp. in the Australian sample set; two were excreted by scrub pythons and one by the spotted python. The sequence of the latter is an undescribed species phylogenetically related to S. lacertae. Of the two Sarcocystis species found in scrub pythons, one showed an 18S rRNA gene sequence similar to S. zamani, which is described from Australia for the first time. The second sequence was identical/similar to that of S. nesbitti, a known human pathogen that was held responsible for outbreaks of disease among tourists in Malaysia. The potential presence of S. nesbitti in Australia challenges the current hypothesis of a snake-primate life cycle, and would have implications for human health in the region. Further molecular and biological characterizations are required to confirm species identity and determine whether or not the Australian isolate has the same zoonotic potential as its Malaysian counterpart. Finally, the absence of S. nesbitti in samples from reticulated pythons (which were reported to be definitive hosts), coupled with our phylogenetic analyses, suggest that alternative snake hosts may be responsible for transmitting this parasite in Malaysia.


Introduction
The genus Sarcocystis comprises a group of more than 200 species of protozoan parasites that infect mammals, reptiles, birds, and fish. Sarcocystis spp. usually require two hosts to complete their life cycle [1,2]. Typically, a carnivorous definitive host preys on an infected omnivorous or herbivorous intermediate host, and the parasite undergoes sexual development within the definitive host's intestine [3]. Sarcocystis parasites using snakes as definitive hosts are of particular interest for several reasons. For example, a high infectious dose of S. singaporensis is pathogenic to its intermediate host, and has been used effectively as a biological rodenticide [4]. In this species, natural intermediate hosts include rats of the genera Rattus and Bandicota, while the giant reticulated python (Malayopython reticulatus) serves as the definitive host [5][6][7][8].
Other snake-host Sarcocystis spp. also are presumed to be pathogenic to humans. Recent outbreaks of disease among tourists in Malaysia were attributed to S. nesbitti, which was identified as the causative agent based on molecular identification of its 18S rRNA gene isolated from muscle cysts (sarcocysts) of patients suffering from acute infection accompanied by myositis [9][10][11]. An earlier record of Sarcocystis-infection and disease among US military personnel in Malaysia in 1993 could have been caused by the same pathogen, although S. nesbitti was not implicated at that time [12]. When searching for the definitive host of S. nesbitti, phylogenetic analyses of its 18S rRNA gene predicted snakes as likely candidates, since the sequence clustered with those from other snake-specific Sarcocystis spp. [13]. The studies that followed this discovery examined a range of snake species, in Malaysia and elsewhere [14,15]. Based on molecular analyses of total DNA extracted from faecal samples using the 18S rDNA sequence as a PCR probe, S. nesbitti was reported to be present in the reticulated python and a monocled cobra (Naja kaouthia) [14,16]. The authors concluded that the snakes were infected with S. nesbitti and served as definitive hosts. However, in none of the faecal samples were sporocysts or oocysts-evidence of infection-identified or examined. Thus, in our view the identity of the definitive host of S. nesbitti remains uncertain.
The findings above implicate snakes as potential carriers of human pathogens, prompting the need for more thorough knowledge of the distribution and host-parasite relationships of Sarcocystis spp. in snakes. In this study, we screened extensively for Sarcocystis spp. within pythons from Southeast Asia, New Guinea (i.e., green tree pythons from Indonesian Papua), and Australia, by detecting sporocysts in faecal samples through a flotation technique and, subsequently, employing a PCR protocol that allows amplification of the 18S rRNA gene from a single sporocyst (the stage released by the definitive host; [17]). This molecular approach directly targets the most relevant parasite stage, without requiring animal experimentation typical of other Sarcocystis life cycle and taxonomic studies. Our results expand our knowledge of Sarcocystis spp. in the region and identify a definitive host of S. nesbitti.

Ethics statement
Our data were gathered from snakes brought to animal rescue centers and public research institutions, kept by local traders, or captured by ourselves in the wild. We stress that no snakes were harmed for the purpose of our study. In case of direct capture, we released all animals unharmed at the site of collection after examination.

Origin of parasite isolates and sampling procedures
We originally isolated oocysts/sporocysts of Sarcocystis spp. from Southeast Asia from faeces collected from adult wild-caught pythons brought to research institutes and animal rescue centres (e.g., Thailand, Borneo) or from animal traders (e.g., Sumatra, from python skin processing facilities). We collected samples from 75 reticulated pythons (Malayophyton reticulatus), five green tree pythons (Morelia viridis) and three Burmese pythons (Python bivittatus). Only scats that had been freshly discharged by pythons and could be unequivocally assigned to a specific individual were considered. Scats were stored in sterile plastic bags in the refrigerator at 4˚C until further examination (usually no longer than 3 months).
Australian samples were collected from wild pythons captured by hand near the town of Bamaga at the northern tip of Cape York Peninsula, Queensland, Australia. We retrieved faecal samples by gently palpating the snake's hind-gut until defecation occurred. Faeces were kept in sterile plastic vials in the refrigerator until further use. We released all animals unharmed at the site of capture after examination. In total, we obtained faecal material from 23 scrub pythons (Simalia amethistina), six spotted pythons (Antaresia maculosa) and three black-headed pythons (Aspidites melanocephalus).

Screening for oocysts/sporocysts of Sarcocystis spp. in faecal samples
Oocysts/sporocysts were extracted from faecal pellets and concentrated using the standard zinc sulphate flotation technique (33%, w/v, zinc sulphate in distilled water). Briefly, one gram of faecal sample was emulsified in a porcelain beaker in 15-16 ml zinc sulphate solution and the suspension passed through brass wire into a 12.5 ml glass centrifuge tube. Each centrifuge tube was filled until a reverse meniscus was formed. Then, a glass coverslip was gently added to the top of the meniscus and positioned so it would not interfere with the centrifuge bucket. Using a swing-out rotor we centrifuged the samples with coverslips at 1,600 g for 2.5 min without applying a brake. We then lifted the coverslips and placed the adherent concentrates onto glass slides for examination under the microscope. All samples were also screened by direct (wet) smears under a microscope at 400-fold magnification. Length and diameter of sporocysts were measured using a calibrated microscope.

Purification of oocysts/sporocysts in preparation for PCR
Our main subject of examination was the sporocyst; the stage of the genus Sarcocystis that develops in the intestine of the definitive host. Although development of sporocysts (sporulation) takes place inside a typical coccidian oocyst, the thin wall of the latter can rupture easily, releasing two sporocysts once in the lumen of the host's gut [17]. Thus, oocysts usually occur shortly after infection, while only sporocysts may be seen later on. We developed a PCR approach that allowed amplification of the 18S rRNA gene from a single sporocyst. This had the advantage that multiple infections with different Sarcocystis spp. could be distinguished.
Furthermore, we suspected that if total DNA was extracted from scats, contamination with Sarcocystis-stages in prey tissue might occur, since rDNA molecules appear sufficiently stable to be detectable by environmental DNA surveys (e.g., [18]). According to experiences of one of the authors, partly digested prey tissue can occur in scats of giant snakes, particularly if bigger prey items are swallowed (T. Jäkel, unpubl. observations).
Oocysts/sporocysts of Sarcocystis-positive samples were purified via zinc chloride flotation. For this purpose, we diluted 1 or 2 cm 3 of faecal material with 10 ml phosphate-buffered saline (PBS) containing 0.3% Tween 20, centrifuged for 10 min at 1600 g, and discarded the supernatant. The pellet was re-suspended in 15 ml zinc chloride solution with a specific density of 1.45 g per cm 3 and centrifuged for 30 min at 400 g. Then, all of the supernatant was passed through a 25 μm mesh, diluted with tap water (1:10), centrifuged for 10 min at 1600 g and the supernatant discarded. The pellet was re-suspended in 1 ml distilled water and stored at 4˚C until further use.
In preparation for PCR analysis, single sporocysts were aspirated with a pipette under an inverted microscope, transferred to 0.2 ml reaction tubes containing 5 μl of 0.02 M NaOH, and subsequently lysed at 95˚C for 15 min. The lysate was used directly as template for the following PCRs.

PCR and sequencing of the 18S rRNA gene of single sporocysts
For each sample, 30 sporocysts were randomly chosen and examined by PCR. Some samples showed no or only few positive reactions. In those cases, the number of sporocysts chosen for examination was increased up to 100.
Because single sporocysts only contain minute amounts of DNA, we chose a nested PCR approach. The 18S rRNA gene was amplified in two overlapping fragments. The first PCR step was performed in a volume of 25 μl, containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 2 mM MgCl 2 , 200 μM of each dNTP, 6.25 pmol of each external primer, and 0.625 U Ampli-Taq Polymerase (Applied Biosystems) and 1.5 μl of the sporocyst lysate. For the nested PCRs (the second step) volumes were increased to 50 μl, containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 2 mM MgCl 2 , 200 μM of each dNTP, 12.5 pmol of each internal primer and 1.25 U Ampli-Taq Polymerase and 3 μl of the first PCR product. The obtained fragment of the front part was approximately 1020 base pairs (bp) long and the back part 900 bp. We designed the primers (Table 1) from the conserved regions of the 18S rRNA gene of S. hirsuta (AF017122; [19]) and S. singaporensis (AF434054; [20]). All amplification reactions started with an initial denaturation at 95˚C for 5 min, followed by 35 cycles of denaturation for 30 s at 95˚C, annealing for 30 s at 50˚C, elongation for 60 s at 72˚C, and a final elongation step for 5 min at 72˚C. The amplicons were sequenced by GATC Biotech AG (Konstanz, Germany). To join the two fragments, the sequences were aligned, the overlapping region identified and concatenated to a single sequence. The two fragments had to be obtained from a single sporocyst. All concatenated sequences identified in this study were regarded as valid when at least two concatenated sequences from different sporocysts showed 100% identity. Joint sequences were submitted to GenBank under the indicated accession numbers. Using the NCBI BLAST online tool the obtained sequences were compared with already existing data in the GenBank.

Phylogenetic analyses
The 18S rRNA gene sequences of the Sarcocystis isolates under investigation (including the variable regions of the gene; [20]) were aligned with published Apicomplexan species available at GenBank [21] using the multiple sequence alignment algorithm of the 'R-Coffee' web server, which takes into account predicted secondary structures of RNA [22]. In the case of S. zamani, where the 18S rDNA sequence had been deposited in conjunction with neighbouring genes (KU2445241), we excised the relevant part after alignment with the homologous sequences of Toxoplasma gondii (M97703) and S. singaporensis (Sin-T; see results). The appropriate model for nucleotide substitution rates was determined using the jModel Test software version 2.0 [23]. Evolutionary distances were computed using the Maximum Composite Likelihood method with Gamma-distributed substitutions (G) and variation among lineages (as observed for the coccidia; [24]) under the Minimum Evolution (ME) algorithm, or the Tamura-Nei-93 model with G and invariant sites (I) under the Maximum Likelihood (ML) algorithm. Substitution rate variation among sites was estimated by ML over five Gamma categories and the number of sites to be analysed. Site coverage for ME and ML was between 79% and 86%; that is, fewer than 14%-21% alignment gaps, missing data, and ambiguous bases were allowed at any position. There were a total of 1544-1565 positions in the final dataset. All phylogenetic trees were tested by 1000 bootstrap iterations. We conducted all analyses using the MEGA 6 software package [25].

Results
Prevalence of Sarcocystis spp. in faecal samples from pythons and size classes of sporocysts In total, we examined 115 faecal samples from Southeast Asian and Australian pythons for Sarcocystis infection. Using the zinc sulphate flotation technique, Sarcocystis sporocysts could be detected in 78 samples (Table 2), which is an overall prevalence of 68%. Sporocysts were also readily detectable in 66 (85%) direct smears of the flotation-positive samples. However, prevalence varied between python species and geographic origin. The lowest infection rate (17%) was observed in the Australian spotted python (Antaresia maculosa) and the highest prevalence (100%) in reticulated pythons from Sumatra. Based on our measurements of sporocyst dimensions and comparisons with previously published work on Sarcocystis spp. of pythons, we could distinguish four classes of sporocyst sizes among all samples ( Table 2). Measurements from Southeast Asia fell into two size categories. Class 1 (S. singaporensis) and class 2 (S. zamani) were used as reference values and related to previous work that could unequivocally assign mean sporocyst sizes to these two species through experimental infections of reticulated pythons [26,27]. Means and variances of reference values for class 1 and 2 are provided in footnote 'a' of Table 2. Sporocysts of class 2 were significantly smaller in length and diameter than those of class 1 (T-test: length, P 0.001, t = , respectively, which was not statistically different to the class 1 reference values. Hence, they were categorised as class 1. Molecular analysis of randomly chosen subsets of samples confirmed that size class 1 referred to S. singaporensis in Southeast Asia (Table 3; e.g. see samples 9 and 11).
Three sporocyst size classes were present in the Australian samples (Table 2). Sporocysts isolated from the scrub python (designated here as Sarcocystis sp.1) were similar in size to S. singaporensis (class 1) but constituted a genetically different species (see below). Sporocysts isolated from one spotted python measured 11.2 (± 0.5, ±S.D.) x 8.9 (±0.6) μm (n = 30), which was significantly longer and wider than the class 1 reference values (T-test: length, P 0.001,   [26,27]. Sporocysts of even smaller size have been described for S. zamani [27]. Size classes 3 and 4 are defined in the text.

Molecular and phylogenetic analysis of 18S rDNA from Sarcocystis sporocysts
Out of the 78 Sarcocystis-positive faecal samples, 18 (11 from Southeast Asia, 7 from Northern Australia) were randomly chosen for each snake species at each location and subjected to further molecular analysis. We did not include samples from the green tree python (Morelia viridis) of Papua (Indonesia) here, because molecular investigations are still ongoing. In total, we identified four different Sarcocystis spp., one in Southeast Asia and three in the Australian sample set ( Table 3). The only Southeast Asian Sarcocystis sp. detected by PCR was S. singaporensis from the reticulated python (Thailand isolate Sin-T: Accession No. KY513624, Table 3, Fig 1). Although there were two size classes present in mixed infections as observed under the microscope, the larger S. singaporensis-like (class 1) and the smaller S. zamani-like (class 2) sporocysts (Table 2), we could not amplify 18S rDNA of S. zamani, because only one mixed infection was chosen for analysis and concentration of S. zamani-like sporocysts was too low (sample 6 from Mindanao). We obtained partial 18S rDNA sequences of S. singaporensis from a wide range of locations in Southeast Asia (Table 3), which were all highly similar (99%-100%) to sample 11 from Thailand (Sin-T) in pair-wise comparisons using the posterior fragment of the 18S rRNA gene. We observed three different Sarcocystis species in the samples from Northern Australia. Molecular level differences corresponded closely with the different size classes of sporocysts. For instance, sporocysts of size class 3 from the spotted python (A. maculosa) revealed an 18S rDNA sequence of a so far unknown species, which we designate here as Sarcocystis sp.2 ( Table 3, KY513629). Our phylogenetic analyses placed this species close to S. lacertae (branch C , Fig 1), which cycles between snakes and lizards [28]. The two other Sarcocystis spp. were both found in scrub pythons (Simalia amethistina). Interestingly, molecular analysis of class 1 sporocysts (Fig 2A) revealed that these sporocysts contained an 18S rDNA sequence that matched up to 99% with an isolate of S. zamani from a bandicoot rat in Thailand (clade S1 in Fig 1; [29]). Although this indicates that the Australian isolate could be S. zamani, we address this species here as Sarcocystis sp.1 (Table 3), because mean sizes of sporocysts were significantly larger than the reference values of class 2. There existed three different molecular variants of Sarcocystis sp.1 (Table 3). Pair-wise comparison of sequences S.sp.1-1/2 (KY513625) and S. Most intriguingly, we isolated a sequence from sporocysts of size class 4 that showed a high degree of similarity (98%) with the 18S rDNA of S. nesbitti from a muscle biopsy of a patient from Pangkor island in Malaysia (Fig 1) and another closely related human sample from Pangkor [9]. This similarity value is within the range of sequence variability observed among isolates of S. nesbitti from humans, macaques, and snakes (98%-100%), which we determined by pair wise BLAST comparison of S. nesbitti sequences (human: HF544323, HF544324, JX426074, JX661499; macaques: HM021725, HM021724; snakes: KC878476, KC878478, KC878479). Based on the high degree of sequence similarity with S. nesbitti-specific 18S rDNA and the results of our phylogenetic analysis, we address this isolate and the corresponding sporocysts (Fig 2) from the Australian scrub python as Sarcocystis cf. nesbitti, awaiting further characterization as discussed below. Although we could only retrieve sequences of S. nesbitti of one sample (sample 18) from class 4 sporocysts, this size class and morphology was observed three times among the 23 samples from scrub pythons ( Table 2), suggesting that prevalence of infection may have been up to 13%.
When we employed Minimum Evolution (ME) and Maximum Likelihood (ML) algorithms for phylogeny reconstruction, all resulting trees consistently showed two monophyletic clades containing the bulk of the snake-host Sarcocystis spp.: here indicated as S1 and S2 and exemplified by an ME tree (Fig 1). While in this example clades S1 and S2 shared a common ancestor, one has to note that in some of our analyses, regardless of the substitution model or algorithm for phylogeny reconstruction, the Sarcocystis spp. of ruminants (as intermediate hosts) became a direct sister group of S2, which rendered the snake-host Sarcocystis spp. paraphyletic. It appeared that this split occurred more frequently once additional taxa were included in the tree. However, branch support for this scenario remained relatively weak. Isolates of S. nesbitti, including the new sequence from Australia (KY513626), formed a well-supported sub-clade of S2, which also contained Sarcocystis spp. from African, Asian, and American vipers and a colubrid snake. In contrast, clade S1 contained Sarcocystis spp. of colubrid and pythonid snakes distributed in Europe and or Asia, in particular those associated with the reticulated python (S. singaporensis, S. zamani) and the green tree python (KC201639, KC201640). Hence, considering that Sarcocystis sp.2 from the Australian spotted python clustered with taxa of branch C (Fig 1), our phylogenetic results indicated the existence of three snake-host lineages.

Distribution range of S. singaporensis
Our results on the prevalence of Sarcocystis infection among pythons and molecular characterization indicate that S. singaporensis is widely distributed in the Indo-Australian Archipelago [30], probably also being a frequent coccidian parasite of the green tree python (Morelia viridis) in New Guinea. Although our own molecular analysis of the samples from green tree pythons was still ongoing at the time of writing and could not be included here, More et al. [31] isolated an 18S rDNA sequence of Sarcocystis from this snake that showed 97% sequence identity with S. singaporensis. Whether or not S. singaporensis is also present in Australia could not be determined by our study, since molecular sampling did not yield a S. singaporensis-specific sequence despite sporocysts of appropriate size (class 1) being present. Considering that half the pythons in the study area were infected with Sarcocystis spp., and given that S. singaporensis can reach high infection rates in its definitive host (see Table 2 and [27]), it is unlikely that our sample sizes were too small to detect an infection. It is more likely that, if present, S. singaporensis uses different snake-hosts in Australia-possibly the green tree python that is also native there [32], or carpet pythons (Morelia spilota). Previous work has shown at the light microscopical level that infection of laboratory rats using sporocysts isolated from the carpet python resulted in tissue cysts that resembled S. singaporensis [33]. Furthermore, examination of Sarcocystis-infected muscle tissue of Australian bush rats (Rattus fuscipes) revealed the typical ultrastructure of sarcocysts of S. singaporensis [34]. Focusing future research on these hosts should help clarify the status of S. singaporensis in Australia. These tests should also examine if isolates from New Guinea and Australia are infective for reticulated pythons in Southeast Asia. In the other direction, patent infections could be successfully induced in the black-headed python from Australia by feeding rats infected with S. singaporensis from Thailand [8].

New Sarcocystis spp. and host associations in Australian pythons
The discovery of a Sarcocystis species (Sarcocystis sp.1) in the Australian scrub python that is closely related to S. zamani is interesting, because this suggests that S. zamani is present in Australia and uses a different snake-host there. This is not surprising, since muscle cysts that resemble S. zamani have been described from potential intermediate hosts like Australian bush rats [34]. Furthermore, the recent publication of a Sarcocystis 18S rDNA sequence from sporocysts in green tree pythons (presumably of Papua New Guinean origin; [31]) that clustered with S. zamani in our phylogenetic analyses (Fig 1; KC201639) renders it plausible that S. zamani also occurs in New Guinea, linking Southeast Asia with Australia. A sequence variation of up to 2% divergence that was present among our samples of Sarcocystis sp.1 indicates a higher degree of variability than the variation observed among isolates of S. singaporensis throughout large parts of its distribution range. However, this is under the restriction that we used smaller fragments of 18S rDNA for alignment of sequences of S. singaporensis compared to those of Sarcocystis sp.1. Our observations are in line with phylogenetic analyses of the 28S rRNA gene of S. zamani, where two genetically distinct sub-populations of this species were described [29]. However, that study did not link the genetic findings with histological or ultrastructural evidence for presence of S. zamani, so it remains unclear whether or not genetic variability correlated with morphological disparities. The existence of two sub-populations of S. zamani sporocysts from Southeast Asia that were different in size [26,27], and our discovery of even larger sporocysts of a putative S. zamani in Australia, also suggest a relatively high degree of phenotypic variability. Taken together, these uncertainties prevent us from confirming the presence of S. zamani in Australia.
Sarcocystis sp. 2 was clearly distinct to the other snake-host Sarcocystis spp. of this study (taxa of clades S1 and S2). The 18S rRNA gene sequence of Sarcocystis sp. 2 clustered with a group of Sarcocystis spp. that includes a diverse range of definitive hosts (branch C of Fig 1), which is regarded as poorly sampled [24]. Evidence is now accumulating that this group of taxa not only contains S. lacertae, which cycles between the colubrid snake Coronella austriaca as definitive host and the common wall lizard Podarcis muralis as intermediate host [28], but also other snake-host Sarcocystis spp.; for instance those that use neotropical tree boas of the genus Corallus as definitive hosts [15]. This implies the existence of a third lineage of snakehost Sarcocystis spp. that is distinct from the taxa of clades S1 and S2. We suggest that the hypothesis for a single lineage of snake-specific Sarcocystis spp. that has co-evolved with their definitive hosts is in need of revision [35,36].
Because most of the Apicomplexan protozoa are only known from sequences of the 18S rRNA gene [37]-and this is particularly true for the snake-host Sarcocystis spp.-our reconstruction of the tree of different taxa reflects the phylogeny of the gene rather than the species involved. To construct a species phylogeny, multiple gene sequences, best coupled with phenotypic characters should be analysed [37]. To that end, additional genes should be sequenced, such as the mitochondrial cytochrome oxidase 1 (cox1) of the new Sarcocystis isolates examined here. This gene has been found useful for discriminating closely related Sarcocystis spp. [38,39]. Additionally, such investigations should be complemented with cross-infection experiments as outlined for S. singaporensis to clarify hosts and distributions beyond Australia.

The definitive host of S. nesbitti
We report here for the first time the potential presence of S. nesbitti in Australia. We provide molecular and morphological evidence for the stage (sporocyst containing sporozoites) developing in Australian scrub pythons. This discovery is remarkable in several ways. First, it confirms that a snake is indeed the definitive host-as was suggested after phylogenetic analysis of 18S rDNA from sarcocysts of infected monkeys [13]. However, we did not expect S. nesbitti's distribution to extend to Australia, because non-human primates that were identified as intermediate hosts do not occur there [13,[40][41][42]. Therefore, if this parasite is present in Australia (in absence of primates), this immediately leads to a quest for the potential intermediate hosts involved in its life cycle. Australian scrub pythons prefer rainforest habitats, where they prey on birds and mammals, including rodents and wallabies (D. Natusch unpubl. observations; [43]). Therefore, we postulate that S. nesbitti might exhibit an 'ordinary' snake-rodent life cycle, possibly with a lower degree of specificity for the intermediate host (which would account for its presence in primates and humans). Another line of support for this assumption comes from the phylogenetic analysis: If the host-parasite relationships of the Sarcocystis species with snake-rodent life cycle reflect the true evolutionary relationships among these taxa [24], then the prediction that S. nesbitti may use rodents as intermediate host is a relatively robust one. All taxa of clades S1 and S2, except S. nesbitti, are either known to develop in rodents or have been isolated from snakes that are known rodent-eaters. Again, reservations regarding species identification based on a single gene sequence also apply here. Infection experiments with the Australian isolate are required to determine its host range and morphological traits, including further molecular characterization.
Based on previous research, we expected the reticulated python of Southeast Asia to be S. nesbitti's definitive host [14]. However, we did not find S. nesbitti in this snake, despite extensive sampling of animals and locations in the region. Considering that we possibly detected three S. nesbitti infections in 23 scrub pythons (while acknowledging that only one was confirmed by molecular analysis), we would have expected to find at least one or two infected reticulated pythons among our samples if this snake species was a suitable definitive host. Therefore, we do not believe that the reticulated python is a suitable definitive host for S. nesbitti, although we cannot rule out this possibility until infection experiments can provide a clear answer.
Which snake species could be the definitive host of S. nesbitti in Southeast Asia? Our finding of S. nesbitti in Simalia amethistina suggests a close relative to that species as a likely candidate, perhaps among the Afro-Asian pythons (e.g., Python bivittatus or P. brongersmai; [44]). However, our phylogenetic analysis appears to contradict that hypothesis, since S. nesbitti did not cluster with S. singaporensis or S. zamani, which are both associated with pythons. Instead, it grouped with taxa of clade S2, which included Sarcocystis spp. of colubrid and viperid snakes. This could imply that S. nesbitti uses snakes of different families as definitive hosts. For example, S. muriviperae appears to develop in colubrid as well as viperid snakes [45]. On the other hand, clade S2 may still be under-sampled; in that case additional sampling could reveal more python hosts in this group.
With regard to the outbreaks on Tioman Island [10,46] we are reluctant to accept a reticulated python-long-tailed macaque life cycle for S. nesbitti as hypothesized by Shahari and coworkers [47]. Adult reticulated pythons do prey on long-tailed macaques among other larger mammals, but only once they have attained a certain size (3-4 m total length in South Sumatra; [48,49]); smaller pythons mainly feed on rats [49]. Thus, even if macaques occasionally form part of their diet on Tioman, the volume of infectious sporocysts released into the environment would not be sufficient to explain the apparently widespread contamination of village water sources [10]. In order to find the source of infection, we suggest future investigators search the headwaters of streams flowing into areas where significant human infections occurred [10,50], and sample snake species found nearby to these water bodies. Trapping and examination of rodents in these same areas should also be undertaken.
In conclusion, the potential presence of S. nesbitti in the Australian scrub python extends the known range of this pathogen significantly, and warrants further investigations into its relationship with Malaysian isolates and disease-causing properties. If confirmed, this would have implications for human health in the region. Australian authorities and medical personnel should be made aware of the parasite's potential presence and further sampling should be undertaken in more populated parts of the country.