Identification and Biotyping of Pythium insidiosum Isolated from Urban and Rural Areas of Thailand by Multiplex PCR, DNA Barcode, and Proteomic Analyses

Pythium insidiosum causes pythiosis, a fatal infectious disease of humans and animals worldwide. Prompt diagnosis and treatment are essential to improve the clinical outcome of pythiosis. Diagnosis of P. insidiosum relies on immunological, molecular, and proteomic assays. The main treatment of pythiosis aims to surgically remove all affected tissue to prevent recurrent infection. Due to the marked increase in case reports, pythiosis has become a public health concern. Thailand is an endemic area of human pythiosis. To obtain a complete picture of how the pathogen circulates in the environment, we surveyed the presence of P. insidiosum in urban (Bangkok) and rural areas of Thailand. We employed the hair-baiting technique to screen for P. insidiosum in 500 water samples. Twenty-seven culture-positive samples were identified as P. insidiosum by multiplex PCR, multi-DNA barcode (rDNA, cox1, cox2), and mass spectrometric analyses. These environmental strains of P. insidiosum fell into Clade-II and -III genotypes and exhibited a close phylogenetic/proteomic relationship with Thai clinical strains. Biodiversity of the environmental strains also existed in a local habitat. In conclusion, P. insidiosum is widespread in Thailand. A better understanding of the ecological niche of P. insidiosum could lead to the effective prevention and control of this pathogen.

Thailand is an endemic area of human pythiosis, and the most affected individuals are farmers [2,4]. Human pythiosis is associated with several hematological disorders (especially thalassemia), in which the underlying mechanism is unknown [2,4,28]. Several case series of human pythiosis (mainly ocular infection) were recently reported from India [29][30][31]. Pythiosis in animals (i.e., horses and dogs) has been mostly diagnosed in other countries, especially Brazil and the United States [1,3,32]. Due to the marked increase in case reports, pythiosis has become a public health concern. As a part of its life cycle, P. insidiosum produces an infective unit, called biflagellate zoospore, in water [33]. Plant materials and animal hairs can attract the organism [34][35][36]. When a swimming zoospore comes in direct contact with an individual, it germinates as hyphae and causes tissue pathology [33,34]. Better understanding the ecological niche of P. insidiosum could lead to the effective prevention and control of this pathogen, especially for those individuals at risk.
Several investigators can isolate P. insidiosum from agricultural or non-residential areas in northern Thailand, Australia, the United States, and Brazil [34][35][36][37][38][39]. To obtain a complete picture of how the pathogen circulates in the environment, we aim to survey the presence of P. insidiosum in a crowded city like Bangkok, as well as rural areas of central and southern Thailand. We employed the hair-baiting technique to isolate P. insidiosum from 500 water samples. Colony morphology was used as a high-throughput screening method, and the culture-positive samples were confirmed by several molecular assays, including multiplex PCR, multi-DNA barcode, and proteomic analyses [4,19,23,[40][41][42][43]. We successfully isolated P. insidiosum from Bangkok and other provinces, in which some areas had a notably-high prevalence of the organism. We explored biodiversity, proteomic feature, and phylogenetic relationship of the environmental and clinical strains of P. insidiosum and proposed the use of multi-DNA barcodes for the identification of this pathogen.

Sample Collection and Culture Condition
A total of 500 water samples were collected from 100 sample-collection sites in 23 sampling locations (i.e., zoo, public parks, rice fields, and ponds) across 7 provinces of Thailand, which included Bangkok (10 locations; 48 sites; 240 samples), Chonburi (1 location; 3 sites; 15 samples), Chachoengsao (3 locations; 8 sites; 40 samples), Nakhon Pathom (1 location; 12 sites; 60 samples), Kanchanaburi (3 locations; 11 sites; 55 samples), Ratchaburi (4 locations; 13 sites; 65 samples), and Trang (1 location; 5 sites; 25 samples) ( Table 1). Five water samples (500 mL/sample) were collected from each sample-collection site using a clean disposable plastic bucket (sampling position: 50-100 cm away from the bank; 5-10 cm depth from the water surface). Each water sample was transferred to a sterile plastic bag containing 5-10 autoclaved 10-cm-long human hairs and left at the ambient temperature overnight. The hairs were removed from the bag by sterile forceps and incubated on Sabouraud dextrose agar (pH 7.2) supplemented with penicillin and streptomycin (100 µg/mL each; Sigma-Aldrich, St. Louis, MO, USA) at 25 • C for 5 days. A growing, submerged, white-to-colorless colony, which is compatible with P. insidiosum, was subculture onto a freshly-prepared Sabouraud dextrose agar (with or without an overlayed sterile cellophane membrane) and subject to the downstream genomic DNA (gDNA) extraction.

Genomic DNA Extraction
Up to 200 mg of an obtained colony were harvested for gDNA extraction by adapting the salt extraction protocol described by Lohnoo et al. [21]. A hyphal mat was transferred to a 2-mL sterile plastic screw-cap tube containing 1000 mg of glass beads (710-1180 mm in diameter; Sigma, St. Louis, MO, USA) and combined with the salt homogenizing buffer (0.4 M NaCl, 10 mM Tris-HCl pH 8.0, and 2 mM EDTA; 400 µL buffer per 100 mg hyphae). To remove carry-over culture agar from the harvested hyphae, the sample tube was boiled at 100 • C for 5 min. The hyphal mat was ruptured by a Tissue Lyzer Retsch MM301 (setting: 2 min at 30 Hz; Qiagen, Hilden, Germany) and mixed with 45 µL of 20% SDS and 8 µL of 20 mg/mL proteinase K, before an overnight incubation at 56 • C. After well-mixed with 0.3 mL of 6 M NaCl, the sample was centrifuged at 10,000× g for 30 min. The supernatant was collected, combined with an equal volume of isopropanol, stored at −20 • C for 1 h, and centrifuged (10,000× g) at 4 • C for 20 min. After discarding the supernatant, a resulting pellet was collected, washed with 70% ethanol, air dried, and resuspended in 100 µL of Tris-EDTA (10 mM Tris, 1 mM EDTA; pH 8.0). A NanoDrop 2000 spectrophotometer estimated DNA concentration at 260/280 nm wavelengths (Thermo Scientific, Waltham, MA, USA).

Species Identification by DNA Barcode Analysis
All extracted gDNA samples were recruited for species identification using the rDNA sequence (i.e., the ITS1-5.8S-ITS2 region) as the primary DNA barcode. PCR amplification was carried out in a 50-µL reaction containing 50 ng gDNA template, the universal fungal primers The secondary DNA barcode sequences (i.e., cox1 and cox2) were also amplified from gDNA samples of P. insidiosum and another Pythium spp. A PCR amplification was set up in a 50-µL reaction, containing 100 ng gDNA template, the oomycete-specific cox1 primers [0.2 µM each of OomCox-I_Levup (5 -TCAWCWMGATGGCTTTTTTCAAC-3 ) and OomCox-I_Levlo All PCR products were checked using the QIAxcel advanced system (as mentioned above). The PCR products were cleaned using a PCR purification kit (Qiagen) and sequenced using a corresponding primer set, such as ITS1/ITS4 (rDNA), OomCox-I_Levup/OomCox-I_Levlo (cox1), and FM58/FM66 (cox2). The obtained sequence was BLAST searched against the NCBI nucleotide database (https://blast.ncbi.nlm.nih.gov/Blast.cgi accessed on 16 January 2021). The BLAST search cutoff for species identification was set at 98.5% identity [43].

Mass Spectrometric Analysis and Dendrogram
Selected organisms (i.e., a colony without bacterial contamination) isolated from the environment (i.e., P. insidiosum, Pythium catenulatum, and Pythium rhizo-oryzae) (Tables 1 and 2), P. insidiosum strain Pi35 isolated from a patient with pythiosis (positive control) and Candida parapsilosis strain ATCC 22019 (negative control) were recruited for MALDI-TOF MS analysis. Each organism was subcultured in a glass Petri dish containing 10 mL Sabouraud dextrose broth and incubated at 37 • C for 5-7 days. Proteins were extracted from these organisms using the established protocol [19] with some modifications. Briefly, a harvested organism (100 mg) was transferred to a sterile microtube, washed 5 times with 1 mL of liquid chromatography-mass spectrometry (LC-MS) grade water (Merck, Kenilworth, NJ, USA), and centrifuged (16,600× g) at 4 • C for 5 min to remove the supernatant. The organism was mixed with 300 µL of LC-MS grade water before vortexing and adding 900 µL of absolute EtOH (Merck). The sample mixture underwent another round of vortexing, centrifugation, and supernatant removal. A resulting pellet was dried at 37 • C for 30 min and resuspended with equal volumes (up to 100 µL) of 70% formic acid (Merck) and acetonitrile (Merck). The supernatant (containing extracted proteins) was collected by centrifugation and kept at −30 • C until use. Table 2. Twenty-seven strains of P. insidiosum successfully isolated from 5 provinces in Thailand. Sampling locations (with GPS coordinates), water sources, strain IDs, sequence homology analyses, and species-level identification and biotyping (based on DNA barcodes, multiplex PCR, and mass spectrometry) are summarized in the The extracted protein (0.5 µL) was placed on a ground steel target plate (Bruker Daltonics, Billerica, MA, USA) (8 replicates), air-dried, and layered with 0.5 µL of 5 mg/mL α-cyano-4-hydroxycinnamic acid in 70% acetonitrile and 0.1% trifluoroacetic acid. An ultra-fleXtreme mass spectrometer and the FlexControl software version 3.0 (Bruker Daltonics), using the previously-described setting [19], generated mass spectra from the extracted proteins. The MALDI-TOF MS analysis matched the generated mass spectra of each organism against the supplemented Bruker MALDI Biotyper database DB4613 (Bruker Daltonics), containing the main spectral profiles (MSP) of 4274 bacteria, 331 fungi, 7 archaea, 1 green alga, and 13 P. insidiosum strains [19]. Mass spectrum similarity was transformed into an identification score by the MALDI Biotyper software version 3.0 (Bruker Daltonics). A score of 2.00 or higher indicates a reliable species-level identification, while that fall between 1.70-1.99 indicates a reliable genus-level identification. A lower score (<1.70) means an unreliable organism identification. The MATLAB software version 7.1 (MathWorks, Natick, MA, USA) generated a dendrogram of all P. insidiosum isolates tested, by using the distance values (for each pair of P. insidiosum MSPs) calculated by the MALDI Biotyper software [19].

Data Availability
The rDNA, cox1, and cox2 sequences of the P. insidiosum strains used in this study have been deposited in the GenBank/DDBJ databases (see the accession numbers in Tables 2 and 3). The rDNA, cox1, and cox2 sequences of the P. rhizo-oryzae strain RCB01 (accessions: LC556053, LC553639, and LC553641, respectively) and the P. catenulatum strain RM9-06 (accessions: LC556067, LC553640, and LC553642, respectively), used as an outgroup in the phylogenetic analysis, have been deposited in the same databases. Table 3. Twenty-two clinical strains of P. insidiosum used for proteomic and phylogenetic analyses in this study. The table contains strain IDs, reference IDs, affected hosts (i.e., humans or animals), country of origins, mass spectrometry-based prototypes, and GenBank accessions (i.e., rDNA, cox1, and cox2).

Screening the Colony Morphology of P. insidiosum Isolated from Water Samples
Water samples were collected from 100 water collection sites [i.e., rice fields, irrigation channels, and ponds (in a zoo, public recreation parks, and countryside areas); 5 samples/site] in 10 urban (i.e., Bangkok) and 13 rural (i.e., 5 central and 1 southern provinces) sampling locations (Table 1). All 23 locations were depicted in Figure 1 (also available online at https://microreact.org/project/nv2faGXa2rahFjUHKd5QQN access on 16 January 2021 [50]). Human hairs, used to bait P. insidiosum in a water sample, were incubated at room temperature on a Sabouraud agar plate supplemented with the antibacterial agents. From the total of 500 water samples, 446 (89.2%) showed different bacterial growths and various fungal colonies (with or without spores or color pigments), while 54 (10.8%) exhibited no growth on the agar plates. Among them, 64 samples (12.8%) provided a white-to-colorless, non-sporulation, submerged colony, which is compatible with the gross morphology of P. insidiosum. Each suspected P. insidiosum colony was subcultured on a new Sabouraud agar plate for gDNA extraction. The obtained gDNA samples were used for species identification, biotyping, and phylogenetic analysis (see below).

Screening the Colony Morphology of P. insidiosum Isolated from Water Samples
Water samples were collected from 100 water collection sites [i.e., rice fields, irrigation channels, and ponds (in a zoo, public recreation parks, and countryside areas); 5 samples/site] in 10 urban (i.e., Bangkok) and 13 rural (i.e., 5 central and 1 southern provinces) sampling locations (Table 1). All 23 locations were depicted in Figure 1 (also available online at https://microreact.org/project/nv2faGXa2rahFjUHKd5QQN access on 16 January 2021 [50]). Human hairs, used to bait P. insidiosum in a water sample, were incubated at room temperature on a Sabouraud agar plate supplemented with the antibacterial agents. From the total of 500 water samples, 446 (89.2%) showed different bacterial growths and various fungal colonies (with or without spores or color pigments), while 54 (10.8%) exhibited no growth on the agar plates. Among them, 64 samples (12.8%) provided a white-to-colorless, non-sporulation, submerged colony, which is compatible with the gross morphology of P. insidiosum. Each suspected P. insidiosum colony was subcultured on a new Sabouraud agar plate for gDNA extraction. The obtained gDNA samples were used for species identification, biotyping, and phylogenetic analysis (see below).

Figure 4.
Comparison of mass spectra from P. insidiosum and non-insidiosum Pythium species. Four mass spectra are generated from P. insidiosum strains inhabited in the same rice field (IDs: KCB02, KCB03, KCB04, and KCB05). One each of the mass spectra is derived from P. insidiosum strain BKDZ02 (from a zoo in Bangkok), P. rhizo-oryzae strain RCB01 (from a pond in Ratchaburi province), and P. catenulatum strain RM9-09 (from a pond in Bangkok). The P. insidiosum strain Pi35 (from a pythiosis patient) is included as a reference organism. The Y-axis shows mass intensity, while the X-axis represents the mass-to-charge ratio (m/z). Asterisks indicate the prominent m/z peaks that share among different strains of P. insidiosum.

Discussion
We surveyed the presence of P. insidiosum in urban and rural watery areas of Thailand, using the hair-baiting technique [35]. A total of 500 water samples were collected from 100 sites (i.e., ponds and rice fields) in 7 central and southern provinces of the country ( Figure 1). Cultures of most water samples (89.2%) showed growing bacterial and fungal colonies. Among them, 64 samples (12.8%) provided a white-to-colorless, non-sporulation, submerged colony, which is compatible with the gross morphology of P. insidiosum. Such morphologies are not specific to P. insidiosum, as they are observed in several microorganisms. Nevertheless, recognition of the colony characteristics can facilitate the screening of P. insidiosum through a vast number of water samples. Multiplex PCR [23] identified P. insidiosum in 27 out of 64 colony screening-positive samples (Tables 2 and 4). The identity of P. insidiosum was confirmed by DNA barcode analyses (i.e., rDNA, cox1, and cox2) [40][41][42][43]. rDNA is the most common barcode used to identify an organism at the species level [40,43]. However, rDNA failed to assign P. insidiosum in one of 27 PCR-positive samples (sequence identity cutoff: 98.5%) ( Table 2), indicating that the current rDNA database had a limitation in identifying this organism. The secondary barcodes (cox1 and cox2) were then employed [41][42][43]. cox2 identified P. insidiosum in all 27 PCR-positive samples (sequence identity: 100%), whereas cox1 detected the organism only in 19 PCR-positive samples (sequence identity: 99.3-100.0%) ( Table 2). The ineffectiveness of cox1 in the identification of P. insidiosum was due to the limited cox1 database in GenBank, as only 4 cox1 sequences of this species (accessions: JQ305799, HQ708612, HQ708611, and AP014838) were available at the time of analysis. As the final result, the colony screening, multiplex PCR, and DNA barcode analyses co-identified P. insidiosum in 27 out of 500 water samples (detection rate: 5.4%) ( Table 2).
P. insidiosum can be isolated from swampy areas in several countries across the world (i.e., Thailand, Australia, the United States, and Brazil) [34][35][36][37][38][39]. Recently, Jara et al. successfully isolated P. insidiosum throughout the study area in the Chincoteague National Wildlife Refuge in Virginia, the United States [39]. Based on an ecological niche model framework, they predicted that the warm weather during June and August is more suitable for the organism than the cold weather during December and March [39]. In Thailand, Figure 5. Proteomic dendrogram of water-isolated and clinical strains of P. insidiosum. The main spectral profiles (MSP) of 10 water-isolated (4 Clade-II and 6 Clade-III genotype strains; Table 2) and 13 clinical (4 Clade-I, 5 Clade-II, and 4 Clade-III genotype strains; Table 3) strains of P. insidiosum are recruited for the construction of a dendrogram. The distance value of 500 is used as the cut-off value for proteotyping of the organisms. Asterisks indicate water-isolated strains of P. insidiosum.

Discussion
We surveyed the presence of P. insidiosum in urban and rural watery areas of Thailand, using the hair-baiting technique [35]. A total of 500 water samples were collected from 100 sites (i.e., ponds and rice fields) in 7 central and southern provinces of the country (Figure 1). Cultures of most water samples (89.2%) showed growing bacterial and fungal colonies. Among them, 64 samples (12.8%) provided a white-to-colorless, non-sporulation, submerged colony, which is compatible with the gross morphology of P. insidiosum. Such morphologies are not specific to P. insidiosum, as they are observed in several microorganisms. Nevertheless, recognition of the colony characteristics can facilitate the screening of P. insidiosum through a vast number of water samples. Multiplex PCR [23] identified P. insidiosum in 27 out of 64 colony screening-positive samples (Tables 2 and 4). The identity of P. insidiosum was confirmed by DNA barcode analyses (i.e., rDNA, cox1, and cox2) [40][41][42][43]. rDNA is the most common barcode used to identify an organism at the species level [40,43]. However, rDNA failed to assign P. insidiosum in one of 27 PCR-positive samples (sequence identity cutoff: 98.5%) ( Table 2), indicating that the current rDNA database had a limitation in identifying this organism. The secondary barcodes (cox1 and cox2) were then employed [41][42][43]. cox2 identified P. insidiosum in all 27 PCR-positive samples (sequence identity: 100%), whereas cox1 detected the organism only in 19 PCR-positive samples (sequence identity: 99.3-100.0%) ( Table 2). The ineffectiveness of cox1 in the identification of P. insidiosum was due to the limited cox1 database in GenBank, as only 4 cox1 sequences of this species (accessions: JQ305799, HQ708612, HQ708611, and AP014838) were available at the time of analysis. As the final result, the colony screening, multiplex PCR, and DNA barcode analyses co-identified P. insidiosum in 27 out of 500 water samples (detection rate: 5.4%) ( Table 2).
P. insidiosum can be isolated from swampy areas in several countries across the world (i.e., Thailand, Australia, the United States, and Brazil) [34][35][36][37][38][39]. Recently, Jara et al. successfully isolated P. insidiosum throughout the study area in the Chincoteague National Wildlife Refuge in Virginia, the United States [39]. Based on an ecological niche model framework, they predicted that the warm weather during June and August is more suitable for the or-ganism than the cold weather during December and March [39]. In Thailand, Supabandhu et al. successfully isolated 59 P. insidiosum strains from 325 water samples collected from agricultural areas (i.e., rice fields, irrigation channels, and water reservoirs) in northern Thailand [35]. They reported the isolate-per-sample (IPS) value of 59/325 or 0.18. The current study reported the IPS value of 27/500 or 0.05, which was calculated based on the P. insidiosum-positive samples collected from urban areas (i.e., zoo and public parks; 7 isolates per 300 samples) and agricultural areas (i.e., ponds and rice fields; 20 isolates per 200 samples) in central and southern Thailand ( Table 1). The IPS value of Supabandhu et al. (0.18) was 3.4-fold higher than that of our study (0.05). In our study, the IPS value of agricultural areas (20/200 or 0.10) was 5-time higher than that of urban areas (7/300 or 0.02). P. insidiosum may be more prevalent in the northern part than the other parts of Thailand. On the other hand, the low prevalence may due to sampling biases, as 48% of the water samples were collected from the urban areas. Taken together, we learned that: (i) P. insidiosum is widespread in Thailand (and perhaps in neighborhood countries where cases are not yet reported); (ii) the organism presents in the crowded city, i.e., Bangkok; and (iii) the pathogen is more prevalent in the agricultural habitats. The higher prevalence of P. insidiosum in the agricultural areas was consistent with the fact that the majority of Thai patients with pythiosis were farmers living all over the country [2,35]. An individual who exposes to the ecological niche of P. insidiosum could become at risk of the infection.
P. insidiosum is classified into 3 genotypes, in association with its geographic origins (i.e., Clade-I genotype in Americas, Clade-II genotype in Asia and Australia, and Clade-III genotype in Thailand) [52,53]. The multiplex PCR has the ability to not only detect P. insidiosum, but also genotype this organism into Clade-I, -II, or -III strains, simply based on size and number of the amplicons [23]. This amplification technique correctly assigned 27 water-isolated P. insidiosum strains into Clade-II (n = 19) and Clade-III (n = 8) genotypes, which were in agreement with the phylogenetic findings (Table 2; Figure 3). Biodiversity of the Thai water-isolated strains of P. insidiosum (n = 26; Table 2), in relation to the human and animal strains from different geographic areas (n = 22; Table 3), was assessed by phylogenetic and proteomic approaches. Using the rDNA-cox1-cox2 concatenated sequences, P. insidiosum can be grouped into 3 phylogenetic clades, as expected ( Figure 3). The Thai water-isolated strains were restricted to only Clade-II and Clade-III (Table 2; Figure 3). This finding suggests that the major circulating strains of P. insidiosum in the Thai environment are the Clade-II and Clade-III genotypes, which are the typical genotypes of the pathogen isolated from all Thai patients [23,53]. Until 2020, a clade A th (equivalent to Clade-I) strain of P. insidiosum was isolated from the first dog with pythiosis in Thailand [54]. Such information suggests that the Clade-I strains might also circulate in Thailand, but to a much lesser extent than the Clade-II and -III strains.
We initially explored the proteome-based biodiversity of the P. insidiosum isolated from water (n = 10), humans (n = 6), and animals (n = 3) (Tables 2 and 3). Unlike the phylogenetic approach, the mass spectrometry-derived dendrogram divided these isolates into only 2 groups: proteotype-A (comprising Clade-I and -II genotypes) and proteotype-B (comprising only Clade-III genotypes) ( Figure 5). Hence, the proteomic method exhibited less discrimination power for bio-diversifying P. insidiosum than the phylogenetic approach. Nine of the water-isolated strains were from the same sample collection site (Rice Field#3) and can be grouped into 2 subpopulations: proteotype-A/genotype Clade-II (n = 4) and proteotype-B/genotype Clade-III (n = 5) (Table 2; Figure 5). The proteomic ( Figure 5) and phylogenetic (Figure 3) analyses demonstrated the marked biodiversity of the P. insidiosum subpopulation inhabiting a local environment.
In conclusion, we successfully isolated P. insidiosum from the urban and rural areas (including the city of Bangkok), using the hair-baiting technique. The identity of the organism was confirmed by multiplex PCR, DNA barcoding, and proteomic analysis. The combination of rDNA and cox2 barcodes showed superior performance for the identification of P. insidiosum, while the cox1 barcode cannot assign a species to some strains due to the lack of a comprehensive dataset in GenBank. Proteomic and phylogenetic analyses revealed subpopulations and biodiversity (i.e., proteotype-A/genotype Clade-II and proteotype-B/genotype Clade-III) of the water-isolated P. insidiosum strains in a local area. P. insidiosum is ubiquitous in Thailand and only the Clade-II and Clade-III genotypes (the typical genotypes that infect Thai patients) circulate in the environment (i.e., rice fields and ponds). Better understanding the ecological niches of P. insidiosum can lead to a proper measure to reduce the exposure of an individual at risk to the pathogen, and thus prevent pythiosis.