THESIS MOLECULAR BARCODING OF FIFTEEN VENOMOUS SNAKES, AND IDENTIFICATION EIGHT SNAKE GROUPS IN THAILAND USING MULTIPLEX PCR ARJAREE SUPIKAMOLSENI

THESIS MOLECULAR BARCODING OF FIFTEEN VENOMOUS SNAKES, AND IDENTIFICATION EIGHT SNAKE GROUPS IN THAILAND USING MULTIPLEX PCR ARJAREE SUPIKAMOLSENI GRADUATE SCHOOL, KASETSART UNIVERSITY 2016

Master of Science (Genetics) DEGREE Genetics FIELD Genetics DEPARTMENT TITLE: Molecular barcoding of fifteen venomous snakes, and identification eight snake groups in Thailand using multiplex PCR. NAME: Miss Arjaree Supikamolseni THIS THESIS HAS BEEN ACCEPTED BY ( Assistant Professor Kornsorn Srikulnath, Ph.D. ) THESIS ADVISOR ( Associate Professor Surin Peyachoknagul, Ph.D. ) THESIS CO-ADVISOR ( Miss Lawan Chanhome, Ph.D. ) THESIS CO-ADVISOR ( Associate Professor Arinthip Thamchaipenet, Ph.D. ) DEPARTMENT HEAD APPROVED BY THE GRADUATE SCHOOL ON ( Associate Professor Gunjana Theeragool, D.Agr. ) DEAN

THESIS MOLECULAR BARCODING OF FIFTEEN VENOMOUS SNAKES, AND IDENTIFICATION EIGHT SNAKE GROUPS IN THAILAND USING MULTIPLEX PCR ARJAREE SUPIKAMOLSENI A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science (Genetics) Graduate School, Kasetsart University 2016

Arjaree Supikamolseni 2016: Molecular Barcoding of Fifteen venomous snakes, and identification eight snake groups in Thailand using multiplex PCR.. Master of Science (Genetics), Major Field: Genetics, Department of Genetics. Thesis Advisor: Assistant Professor Kornsorn Srikulnath, Ph.D. 57 pages. DNA barcodes of mitochondrial COI and Cytb genes were constructed from 54 specimens of 16 species for species identification. Intra- and interspecific sequence divergence of COI gene (10 times) were greater than those of Cytb gene (4 times), which suggests that the former gene may be a better marker than the latter for species delimitation in snakes. COI barcode cut-off scores differed by more than 3% between most species, and the minimum interspecific divergence was greater than the maximum intraspecific divergence. Clustering analysis indicated that most species fell into monophyletic clades. These results suggest that these species could be reliably differentiated using COI DNA barcodes. Moreover, a novel species-specific multiplex PCR assay was developed to distinguish between Naja spp, Ophiophagus hannah, Trimeresurus spp, Hydrophiinae snakes, Daboia siamensis, Bungarus fasciatus, and Calloselasma rhodostoma. Antivenom for these species was produced and distributed by the Thai Red Cross for clinical use. Our novel PCR assay could easily be applied to venom and saliva samples and could be used effectively for rapid and accurate identification of species in forensic work, conservation study, and medical research. Student s signature Thesis Advisor s signature / /

ACKNOWLEDGEMENTS I would like to express my sincere gratitude to my thesis advisor Assistant Professor Dr. Kornsorn Srikulnath for patience, motivation and valuable suggestion in this study. I would like to grateful thank to Associate Professor Dr. Surin Peyachoknagul and Dr. Lawan Chanhome at Queen Saovabha Memorial Institute, Thai Red Cross Society for serving on my co-advisor and for her advice and much helpful discussion to my research. A very special thanks goes out to Dr. Lawan Chanhome and all staff in the Laboratory of Department of Research and Development, Queen Saovabha Memorial Institute, Thai Red Cross for providing the specimens, had helped and encouragement in this research work. I would like to express my special gratitude to all members in the Laboratory of Animal Cytogenetics & Comparative Genomics, Kasetsart University, Bangkhen for their great helps not only about laboratory but also in daily life all the time. Special thank to my seniors, my friends and my juniors for their kind helps during conducting the field experiment. This work was financially supported by Kasetsart University Research and Development Institute (KURDI), and National Research Council of Thailand (NRCT) Finally I express deep thanks to my family for continuous encouragement. Arjaree Supikamolseni October 2016

i TABLE OF CONTENTS Page TABLE OF CONTENTS i LIST OF TABLES ii LIST OF FIGURES iii LIST OF ABBREVIATIONS viii INTRODUCTION 1 OBJECTIVES 4 LITERATURE REVIEW 5 MATERIALS AND METHODS 16 Materials 16 Methods 17 RESULTS AND DISCUSSION 24 CONCLUSION 49 LITERATURE CITED 50 CURRICULUM VITAE

ii LIST OF TABLES Table Page 1 Classification and accession numbers of species used in sequence analyses 19 2 Primers used for the amplification of mitochondrial COI and Cytb genes in this study 21 3 Number of haplotypes of COI and Cytb sequences 27 4 Average sequence divergence (p-distance) of COI gene within species and between pairs of venomous snake species. Diagonal values are p-distance for intra-specific comparisons 28 5 Average sequence divergence (p-distance) of Cytb gene within species and between pairs of venomous snake species. Diagonal values are p-distance for intra-specific comparisons 29 6 Summary of species-specific multiplex PCR assay 40

iii LIST OF FIGURES Figure Page 1 Ramphotyphlops australis 6 2 Cylindrophis ruffus ruffus 7 3 Xenopeltis unicolor 8 4 Python molurus 9 5 Acrochordus javanicus 10 6 Ahaetulla nasuta 11 7 Naja kaouthia 12 8 Trimeresurus albolabris 13 9 Phylogenetic tree showing currently accepted hypotheses of snake relationships 14 10 Representative species used in this study. (a) Ophiophagus hannah, (b) Naja siamensis, (c) Naja sumatrana, (d) Naja kaouthia, (e) Bungarus candidus, (f) Bungarus fasciatus, (g) Daboia siamensis, (h) Rhabdophis subminiatus, (i) Boiga cynodon, (j) Trimeresurus albolabris, (k) Trimeresurus macrops, (l) Calloselasma rhodostoma, (m) Hydrophis brookii, (n) Hydrophis obscurus, (o) Enhydrina schistosa, and (p) Python molurus. 18 11 Agarose gel eletrophoresis of PCR products using COI primer (a) and Cytb primer (b). Lane 1 - O. hannah, Lane 2 - N. kaouthia, Lane 3 - B. fasciatus, Lane 4 - B. candidus, Lane 5 - D. siamensis, Lane 6 - T. albolabris. 24 12 Phylogenetic tree shows relationship between 16 species of snake COI gene. 31 13 Phylogenetic tree shows relationship between 16 species of snake Cytb gene. 32

iv LIST OF FIGURES (Continued) Figure Page 14 Dot plot analysis of sequence divergence with COI barcode (a) and Cytb barcode (b). A straight line represents 1:1 ratio between minimum congeneric sequence divergence and maximum intraspecific sequence divergence. Sequence divergences of species which was represented by more than one sample are shown here. 35 15 Agarose gel eletrophoresis of PCR products using multiplex PCR assays with the primer set 1 (panel A) (a) and 2 (panel B) (b) and DNA extracted from blood. Lane 1 H. brookii. Lane 2 E. schistosa. Lane 3 H. obscurus. Lane 4 T. albolabris. Lane 5 T. macrops. Lane 6 D. siamensis. Lane 7 B. candidus. Lane 8 B. fasciatus. Lane 9 O. hannah.lane 10 N. siamensis. 11 N. sumatrana. Lane 12 N. kaouthia. Lane 13 C. rhodostoma. Lane 14 Mus musculus. Lane 15 - Homo sapiens. 37 16 Agarose gel eletrophoresis of PCR products using primer COI Forward and (a) HYD Reverse, (b) TC Reverse, (c) DRS Reverse and (d) DB Reverse.Lane 1 H. brookii. Lane 2 E. schistosa. Lane 3 H. obscurus.lane 4 T. albolabris. Lane 5 T. macrops. Lane 6 D. siamensis.lane 7 B. candidus. Lane 8 B. fasciatus. Lane 9 O. hannah.lane 10 N. siamensis. Lane 11 N. sumatrana. Lane 12 N. kaouthia. Lane 13 C. rhodostoma. Lane 14 Mus musculus. Lane 15 - Homo sapiens. 38

v LIST OF FIGURES (Continued) Figure Page 17 Agarose gel eletrophoresis of PCR products using primer COI Forward and (a) BF Reverse, (b) OH Reverse, (c) NK Reverse, (d) CR Reverse. Lane 1 H. brookii. Lane 2 E. schistosa. Lane 3 H. obscurus. Lane 4 T. albolabris. Lane 5 T. macrops. Lane 6 D. siamensis. Lane 7 B. candidus. Lane 8 B. fasciatus. Lane 9 O. hannah.lane 10 N. siamensis. Lane 11 N. sumatrana. Lane 12 N. kaouthia. Lane 13 C. rhodostoma. Lane 14 Mus musculus. Lane 15 Homo sapiens. 39 18 Agarose gel eletrophoresis of PCR sensitivity using primer COI Forward and (a) HYD Reverse, (b,c) TC Reverse, (d,e) DRS Reverse. Lane 1 25ng.Lane 2 1ng. Lane 3 0.5ng. Lane 4 0.1 ng. Lane 5 0.05 ng. Lane 6 0.01 ng. Lane 7 0.005 ng. Lane 8 0.001 ng. Lane 9 0.0005 ng. Lane 10 0.0001 ng. 41 19 Agarose gel eletrophoresis of PCR sensitivity using primer COI Forward and (a,b) DB Reverse, (c) BF Reverse, (d) OH Reverse, (e,f,g) NK Reverse, (h) CR Reverse. Lane 1 25ng. Lane 2 1ng. Lane 3 0.5ng. Lane 4 0.1 ng. Lane 5 0.05 ng. Lane 6 0.01 ng. Lane 7 0.005 ng. Lane 8 0.001 ng. Lane 9 0.0005 ng. Lane 10 0.0001 ng. 42

vi LIST OF FIGURES (Continued) Figure Page 20 Agarose gel eletrophoresis of PCR products using multiplex PCR sensitivity assays with the primer set 1 (panel A) (a) and 2 (panel B) (b) and DNA concentration 0.5 ug. Lane 1 H. brookii. Lane 2 E. schistosa. Lane 3 H. obscurus. Lane 4 T. albolabris. Lane 5 T. macrops. Lane 6 D. siamensis. Lane 7 B. candidus. Lane 8 B. fasciatus. Lane 9 O. hannah. Lane 10 N. siamensis. Lane 11 N. sumatrana. Lane 12 N. kaouthia. Lane 13 C. rhodostoma. Lane 14 Mus musculus. Lane 15 - Homo sapiens. 43 21 Agarose gel eletrophoresis of PCR products using multiplex PCR assays with the primer set 1 (panel A) (a) and 2 (panel B) (b) and DNA extracted from venom. Lane 1 T. albolabris. Lane 2 T. macrops. Lane 3 D. siamensis. Lane 4 B. candidus. Lane 5 B. fasciatus. Lane 6 O. hannah. Lane 7 N. siamensis. Lane 8 N. sumatrana. Lane 9 N. kaouthia. Lane 10 C. rhodostoma. Lane 11 Mus musculus. Lane 12 Homo sapiens. 44 22 Agarose gel eletrophoresis of PCR products using multiplex PCR assays with the primer set 1 (panel A) (a) and 2 (panel B) (b) and DNA extracted from saliva. Lane 1 T. albolabris. Lane 2 T. macrops. Lane 3 D. siamensis. Lane 4 B. candidus. Lane 5 B. fasciatus. Lane 6 O. hannah. Lane 7 N. siamensis. Lane 8 N. sumatrana. Lane 9 N. kaouthia. Lane 10 C. rhodostoma. Lane 11 Mus musculus. Lane 12 Homo sapiens. 45

vii LIST OF FIGURES (Continued) Figure Page 23 Agarose gel eletrophoresis of PCR products using multiplex PCR assays with the primer set 1 (panel A) (a) and 2 (panel B) (b). DNA extracted from bite-site swabs with DNA concentration 2 ug. Lane 1-3 T. albolabris. Lane 4-6 T. macrops. Lane 7-9 D. siamensis. Lane 10-12 B. candidus. Lane 13-15 B. fasciatus. Lane 16-18 O. hannah. Lane 19-21 N. siamensis. Lane 22-24 N. sumatrana. Lane 25-27 N. kaouthia. Lane 28-30 C. rhodostoma. 47

viii LIST OF ABBREVIATIONS A, G, C, and T = adenine, guanine, cytocine and thymine AFLP = amplified fragment length polymorphism AIC = Akaike Information Criterion BI = Bayesian inference BLASTn = Basic Local Alignment Search Tool-nucleotide bp = base pairs C = degree Celcius COI = cytochrome c oxidase I Cyt b = cytochrome b DDBJ = DNA Data Bank of Japan DNA = deoxyribonucleic acid dntp = deoxynucleotide triphosphate EDTA = ethylenediamine tetraacetic acid EST = expressed sequence tag HCl = hydrochloric acid h = hour M = molar MCMC = Markov Chain Monte Carlo MEGA4 = Molecular Evolutionary Genetics Analysis 4 MgCl2 = magnesium chloride min = minute ml = milliliter mm = millimolar mtdna = mitochondrial DNA NaCl = sodium chloride NCBI = National Center for Biotechnology Information ng = nanogram nm = nanometer PCR = Polymerase Chain Reaction ph = Potential of Hydrogen ion

ix LIST OF ABBREVIATIONS (Continued) pm = Picomolar RAPD = randomly amplified polymorphic DNA RFLP = restriction fragment length polymorphism RNA = ribonucleic acid s = second SDS = Sodium dodecyl sulfate SNP = single nucleotide polymorphism ssp. = subspecies Tris = tris (hydroxyl methyl) aminomethane U = Unit USA = United states of America UV = ultra violet WI = Wisconsin w/v = weight of a substance of the total volume μg = microgram μl = microliter

1 MOLECULAR BARCODING OF FIFTEEN VENOMOUS SNAKES, AND IDENTIFICATION EIGHT SNAKE GROUPS IN THAILAND USING MULTIPLEX PCR INTRODUCTION Snakes (Serpentes) are carnivorous squamate reptiles that exhibit phenotypically diverse radiation (Secor and Diamond, 1998; Castoe et al., 2008, 2009) with extant snakes over 3,000 known species worldwide (Uetz, 2014). Snakes form the most divergent group in Squamata and their speciation occurred over a relatively short period. These characteristics make snakes difficult to place among iguanians and anguimorphs, to which they are closely related (Srikulnath et al., 2010; Pyron et al., 2013). It is also difficult for evolutionary biologists to classify all snakes with conserved morphology within the sublevels of Serpentes. However, snake biodiversity is decreasing globally owing to hunting and trading for health food, medicinal products, and pets. This issue requires serious attention in the context of conservation biology, and has led to an effort to produce a collection of entire diversity of snake alongside a modern, accurate taxonomy. The Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) regulates the trade of certain snakes. It is essential to develop reliable methods to forensically identify snake products in order to enforce trading laws. Conventional morphology-based taxonomic procedures for snakes are well established (Cox et al., 2012), but are time-consuming due to the limited availability of snake specimens and differences between life stages and sexes, which can lead to misidentification. This suggests that modern techniques, such as molecular approaches, are needed in addition to traditional taxonomic methods in order to identify species. DNA sequences of a standardized region from an unknown species can be compared to sequences available in databases in order to identify them as belonging to a particular species. This technique is known as DNA barcoding. The reference sequence library is constructed from a known species and becomes the barcode. The

2 degree of nucleotide divergence between individuals can facilitate identification of species. The gene most commonly used as a marker for the barcode is the mitochondrial COI gene because it has been studied in many vertebrates and exhibits interspecific nucleotide divergence that is greater than its intraspecific nucleotide divergence (Chaves et al., 2008). Other mitochondrial genes, such as Cytb and 16S rrna (Xia et al., 2012; Nicolas et al., 2012), have also been employed as barcodes with varying levels of success. Snake DNA barcodes based on COI, Cytb, and 12S rrna genes are well established in India, China, and the USA (Wong et al., 2004; Pook and McEwing, 2005; Dubey et al., 2011; Gaur et al., 2012). However, these successes were based on studies of only those species that are abundant in a given country, and there is no single standard gene for the reliable identification of snakes at the species level. Global expansion of the snake DNA barcode library is therefore necessary to aid work in conservation biology, medicine, and forensic science. In Thailand, venomous snakebites remain a serious daily occurrence. Even though several specific antivenoms are available, bites from venomous snakes that cause morbidity and mortality often occur. There are seven effective antivenoms available from the Thai Red Cross for clinical use. These include antivenoms to Naja kaouthia, Ophiophagus hannah, Trimeresurus albolabris, Daboia siamensis, Bungarus fasciatus, Bungarus candidus, and Calloselasma rhodostoma. Clinically, the most serious problem for antivenom therapy is snake identification. The classification of snakebites is first performed by investigation of the clinical signs and symptoms or examination of the carcass of the snake and the local ecology. However, this is timeconsuming and must be performed by a specialist. Therefore, several approaches have been developed to the examination of snakebites, such as radioimmunoassay, agglutination assay, enzyme-linked immunosorbent assay, and DNA fingerprinting (Selvanayagam and Gopalakrishnakone, 1999; Suntrarachun et al., 2001). However, there remain problems with false positives, low sensitivity, and availability that is limited to local-species-specific markers (Tibballs, 1992; Mead and Jelinek, 1996). In this study, we generated COI and Cytb barcodes for 53 venomous snakes and one non-venomous snake (Python bivittatus bivittatus) found in Thailand, and

3 identified snakes using the degree of nucleotide divergence between barcodes and clustering analysis with a phylogenetic tree. Species-specific multiplex PCR markers were then developed from barcodes to rapidly delimit eight groups of venomous snakes. Effective antivenom is available from the Thai Red Cross for most of these groups.

4 OBJECTIVES 1. To develop DNA barcodes of venomous snake species in Thailand. Thailand. 2. To develop DNA markers for differentiating venomous snake species in

5 LITERATURE REVIEW Snakes Snakes are reptiles that have been systematically classified in order Squamata. as follows:- Kingdom Animalia Phylum Chordata Subphylum Class Order Suborder Vertebrata Reptilia Squamata Serpentes In Thailand, snake can be divided into 8 families (Jintakun and Chanhome, 1995) as follows:- 1. Family Typhlopidae (typical blind snakes): They are mostly found in the tropical region. There are 2 genera in Thailand consisting of Ramphotyphlops such as Ramphotyphlops braminus (ง ด นธรรมดา) and Typhlops. Such as Typhlops khoratensis (ง ด นโคราช). Morphology: The body is nearly cylindrical and is remarkably rigid in cross section; both characteristics are mechanical adaptations for a subterranean life. Adults, especially of larger species, are slightly thinner anteriorly than posteriorly. Young (or small) individuals have stouter bodies than old (or large) individuals of the same species. The body is darker above and on the sides. The colour of the head and snout shields is consistent within each species. The tail may be characteristically pigmented in some species and in some species the anterior body is slightly paler.

6 Distribution: Found throughout Thailand. The snakes are often found in termite hills and also in flower pots and other damp and dark areas. They eat termite eggs. (Ehmann and Michael, 1993) Figure 1 Ramphotyphlops australis Source: Ehmann and Michael. (1993). 2. Family Cylindrophiidae (pipesnakes): There is only 1 species, Cylindrophis ruffus (ง ก นขบ), in Thailand. Morphology: Small head not distinct from neck, covered with large symmetrical shields; the nostril in a single nasal, which forms a suture with its fellow behind the rostral, with no loreal or preocular scale; a small postocular, a mental groove present; tail short and blunt Distribution: Throughout Thailand in humid regions and flat land. Some have also been found at heights of up to 1700 Meters (Hoser, 2013).

7 Figure 2 Cylindrophis ruffus ruffus Source: Bulian (1999). 3. Family Xenopeltidae (sunbeam snakes): There is only 1 species, Xenopeltis unocolor (ง แสงอาท ตย ). Morphology: The ground color of the snake's dorsal side is a dark brown, almost black, with the ventral side being a white or cream color. In the sunlight, or under strong artificial light, the scales scatter the illumination like a prism, showcasing a breathtaking display of rainbow coloration. There is no snake in the world that can rival the iridescent of the sunbeam snake, this species most enchanting feature (Vitt et al., 2009). Distribution: Can be met in the whole of Thailand.

8 Figure 3 Xenopeltis unicolor Source: Bulian (1999). 4. Family Pythonidae (pythons): There are 3 species as Python reticulates (ง เหล อม), P. molurus (ง หลาม) and P. curtus (ง หลามปากเป ด). Morphology: Pythons (family Pythonidae) represent a family of nonvenomous and now restricted to the warmer regions of the Old World. Distribution: The Python is to be found throughout Thailand. The animals can even be regularly found in the inner city areas of Bangkok (Schleip and Mark, 2010).

9 Figure 4 Python molurus Photo by: Deekrachang (2014). 5. Family Acrochordidae (wart snakes): There are 2 species as Acrochordus javanicus (ง งวงช าง) and A. granulates (ง ผ าข ร ว). Morphology: Small to medium-sized aquatic snakes with blunt heads not distinct from the neck, dorsally directed small eyes and nares, flabby and roughened skin bearing small, spinate scales. Body stout and capable of lateral compression for swimming (Lillywhite and Harvey B, 1996). Distribution: In the coastal areas in the whole of Thailand.

10 Figure 5 Acrochordus javanicus Source: Bulian (1999). 6. Family Colubridae (typical snake): It is the largest snakes family consisting of 5 subfamily, as follows:- - Subfamily Calamariinae such as Calamaria lumbricoidea (ง พงอ อ ลาย), Calamaria pavimentata (ง พงอ อท องเหล อง). - Subfamily Colubrinae such as Ahaetulla fasciolata (ง เข ยวห วจ งจกลาย จ ด), Oligodon joynsoni (ง ป แก วใหญ ). - Subfamily Natricinae such as Amphiesma deschauenseei (ง ลายสาบท อง สามข ด), Opisthotropis maculosus (ง ลายสอจ ดเหล อง). - Subfamily Pseudoxenodontinae such as Plagiopholis nuchlis (ง ห ว ศร), Pseudoxenodon macrops (ง ลายสาบตาโต). - Subfamily Xenodontinae such as Gongylosoma baliodeira (ง สายทอง ลายแถบ). Morphology: Their ventral scales are well developed, usually as broad as the belly. The head is usually oval shaped with systematically arranged shields. The tails are normally cylindrical and pointed (Ehmann, 1993).

11 Distribution: Throughout Thailand. Figure 6 Ahaetulla nasuta Source: Bulian (1999). 7. Family Elapidae: There are 4 subfamilies as follows:- - Subfamily Elapinae as Bungarus candidus (ง ท บสม งคลา), Bungarus fasciatus (ง สามเหล ยม), Naja kaouthia (ง เห าไทย), Naja siamensis (ง เห าพ นพ ษสยาม), Naja sumatrana (ง เห าพ นพ ษส ทอง) and Ophiophagus hannah (ง จงอาง). - Subfamily Hydrophiinae as Hydrophis brookii (ง แสมร งท องเหล อง), Hydrophis obscurus (ง แสมร งระนอง). - Subfamily Psammophiinae as Psammophis indochinensis (ง ม าน ทองอ นโดจ น). - Subfamily Xenodermatinae as Xenodermus javanicus (ง ขอนไม ). Morphology: Elapids show a variety of body forms. The head is usually not much broader than the neck and tail tip is pointed. All are venomous. Short front deeply grooved fangs to administer venom. The fangs are fixed fangs (Shea et at., 1993).

12 Distribution: Throughout Thailand. Figure 7 Naja kaouthia Photo by: Deekrachang (2014). 8. Family Viperidae is venomous snakes found all over the world. It is commonly known as vipers or viperids. There are 2 subfamily: as follows- - Subfamily Crotalinae as Calloselasma rhodostoma (ง กะปะ), Trimeresurus macrops (ง เข ยวหางไหม ตาโต), Ovophis monticola (ง หางแฮ ม ภ เขา), Parias hageni (ง แก วหางแดง) and Popeia fucata (ง เข ยวหางไหม ท องเข ยว ถ นใต ). - Subfamily Viperinae as Daboia russelii (ง แมวเซา). Morphology: All are venomous. The fangs are long, hollow and moveable. They administer primarily a hemotoxin venom. Pit viper has two highly developed facial pits to detect infrared radiation (Nilson and Gutberlet, 2004). Distribution: Throughout Thailand.

13 Figure 8 Trimeresurus albolabris Photo by: Deekrachang (2014). According to most paleontologists, reptiles evolved from the large group of ancient amphibians known as Labrynthodonts, which received their names from the distinctive structure of their teeth. Based on these fossil finds, as well as on anatomical study of modern reptiles, scientists have concluded that the snakes probably evolved from a family of lizards during the time of the dinosaurs. Snakes and lizards share a number of distinct features in the structure of their skull; both, for instance, possess a moveable quadrate bone at the back of the jaw, and both are missing the quadratojugal bone at the rear of the skull (Lenny Flank.Jr, 1997). Pythons and boas (primitive groups among modern snakes) have vestigial hind limbs: tiny, clawed digits known as anal spurs, which are used to grasp during mating. The families Leptotyphlopidae and Typhlopidae (Figure 9) also possess remnants of the pelvic girdle, appearing as horny projections when visible.

14 Figure 9 Phylogenetic tree showing currently accepted hypotheses of snake relationships. Source: Lee et al. (2007) DNA barcoding DNA barcoding is a taxonomic method that uses a short genetic marker in an organism's DNA to identify it as belonging to a particular species. It differs from molecular phylogeny in that the main goal is not to determine classification but to identify an unknown sample in terms of a known classification (Kress et al., 2005). A desirable locus for DNA barcoding should be standardized in most taxa of interest and sequenced without species-specific PCR primers. The amplicons are short enough to be easily sequenced with current technology (Kress and Erickson, 2008). They also provide a large variation among species but a relatively small amount of variation within a species. The gene most commonly used as a marker for the barcode is the mitochondrial COI gene because it has been studied in many vertebrates and exhibits interspecific nucleotide divergence that is higher than its intraspecific nucleotide divergence (Chaves et al., 2008; Vargas et al., 2009). As of 2009, databases of COI sequences included at least 620,000 specimens from over 58,000 species of animals, larger than databases available for any other gene (Ausubel, 2009). Other

15 mitochondrial genes have also been employed with varying levels of success, such as Cytb and 16S rrna (Feng et al., 2011; Xia et al., 2012; Nicolas et al., 2012). Snake DNA barcodes based on COI, Cytb, and 12S rrna genes have been well established in India, China, and the USA (Wong et al., 2004; Pook and McEwing 2005; Dubey et al., 2011; Gaur et al., 2012). DNA markers The development of DNA-based markers has had a revolutionary impact on gene mapping and, more generally, on all of animal and plant genetics. With DNAbased markers, it is theoretically possible to exploit the entire diversity in DNA sequence that exists in any cross (Dodgson et al., 1996). Molecular genetic markers are powerful approaches to examine genetic traits of individuals, populations, or species (Avise, 1994; Linda and Paul, 1995). These markers could also determine genetic diversity, leading to the application for management of the natural resources and genetic improvement programs (Hillis et al., 1996; Liu and Cordes, 2004). Molecular markers can be classified into two groups: group I markers are based on investigation of various functional gene products, such as allozyme marker, and group II markers are based on detection of the variation of nucleotide sequences from anonymous genomic regions through PCR (polymerase chain reaction) or hybridization techniques (O Brien, 1991). Species-specific PCR assay is developed to identify species, genus, of any organisms based on nucleotide polymorphism in each lineage. This assay has been reported for species identification in term of biological conservation and forensic identification, such as mammalian and poultry species (Candrian and Luethy, 1994; Cocolin et al., 2000; Arslan et al., 2006; Magnussen et al., 2007). The advantages of this species-specific method are relatively fast, precise, sensitive, and cost effective, comparing to other DNA markers based assay [restriction fragment length polymorphism (RFLP), randomly amplified polymorphic DNA (RAPD), amplified fragment length polymorphism (AFLP), microsatellite typing, single nucleotide polymorphism (SNP), and expressed sequence tag (EST) markers] (Mane et al., 2007). Therefore, the species-specific PCR assay is a good technique to identify venomous snakes in this study.

16 MATERIALS AND METHODS Materials 1. Snake samples collection A total of fifty four specimens from sixteen species of snakes were used (O. hannah, N. siamensis, N. sumatrana, N. kaouthia, B. candidus, B. fasciatus, D. siamensis, R. subminiatus, B. cynodon, T. albolabris, T. macrops, C. rhodostoma, H. brookii, H. obscurus, E. schistosa, and P. molurus.) as described in Table 1. All specimens were classified into four families (Elapidae, Viperidae, Colubridae and Pythonidae) which were mainly found in Bangkok. 2. Computer and programs for nucleotide sequence analysis 2.1) Basic Local Alignment Search Tool - nucleotide; Blastn and Blastx from http://blast.ncbi.nlm.nih.gov/blast.cgi 2.2) Multiple Sequence Alignment; ClustalW from http://www.ebi.ac.uk/tools/msa/clustalw2/ 2.3) OligoAnalyzer 3.1; OligoAnalyzer from https://sg.idtdna.com/analyzer/applications/oligoanalyzer/ 2.4) Reverse Complement; Reverse Complement from http://www.bioinformatics.org/sms/rev_comp.html 2.5) ORF Finder (Open Reading Frame Finder); ORF Finder from http://www.ncbi.nlm.nih.gov/gorf/ 2.6) ExPASy-tranlate tool; ORF translate from http://web.expasy.org/translate/ 2.7) MEGA4 (Kumar et al., 2004) 2.8) PUAP (Swofford D.L. 2002) 2.9) Mr.BAY (Huelsenbeck JP and Ronquist F 2001.)

17 Methods 1. Specimen collection Blood samples were collected from sixteen snake species (54 individual ) of venomous snakes (O. hannah, N. kaouthia, B. fasciatus, B. candidus, N. sumatrana, D. siamensis, T. albolabris, T. macrops, H. brookii, E. schistosa, H. obscurus, C. rhodostoma, N. siamensis, B. cynodon, R. subminiatus, and P. molurus) (Figure 10 and Table 1). These samples were available in Snake Farm of Queen Saovabha Memorial Institute, Thai Red Cross Society (Bangkok, Thailand). Blood sample was individually collected from each snake through the ventral tail vein with a 10 mm EDTA treated 25-gauge needle attached to a 1 ml disposable syringe and kept at -20 C Snake venom and saliva samples were collected from ten snake species (30 individual) (O. hannah, N. kaouthia, B. fasciatus, B. candidus, N. sumatrana, D. siamensis, T. albolabris, T. macrops, C. rhodostoma, and N. siamensis) using swab snake mouth, and kept in normal saline solution (0.9 % w/v). Because the antivenom of ten venomous species in this study were available. Bite-site swabs were collected by inducing venomous snakes to bite Swiss albino mice (Mus musculus), and then kept in normal saline solution (0.9 % w/v) for DNA extraction. 2. Genomic DNA Extraction Whole genomic DNA was extracted from blood, snake venom, saliva and bitesite swabs samples using Wizard Genomic DNA Purification Kit (Promega, Madison, WI, USA), and/or following a standard salting-out protocol, and used as a template for polymerase chain reaction (PCR). Briefly, the blood cells were digested at 55 C for 1 h using 1.6 μg/μl proteinase K in extraction buffer (50 mm Tris-HCl (ph 8.0), 20 mm EDTA, and 1% (w/v) sodium dodecyl sulfate (SDS). The mixture was then extracted using a salt solution (0.05 volumes of 5 M NaCl) and the DNA was precipitated using two volumes of isopropanol. After washing in 70% ethanol, genomic DNA was air-dried, resuspended in 10 mm Tris-HCl (ph 8.0), and kept at -

18 80 C. DNA quality and concentration were determined by 1% agarose gel electrophoresis and nucleic acid quantitation was carried out by spectrophotometer at 260 and 280 nm. Figure 10 Representative species used in this study. (a) Ophiophagus hannah, (b) Naja siamensis, (c) Naja sumatrana, (d) Naja kaouthia, (e) Bungarus candidus, (f) Bungarus fasciatus, (g) Calloselasma rhodostoma, (h) Rhabdophis subminiatus, (i) Boiga cynodon, (j) Trimeresurus albolabris, (k) Trimeresurus macrops, (l) Daboia siamensis, (m) Hydrophis brookii, (n) Hydrophis obscurus, (o) Enhydrina schistosa, and (p) Python molurus.

19 Table 1 Classification and accession numbers of species used in all sequence analyses Family Species Sex Locality Code COI Accession number Cytb Accession number Elapidae Ophiophagus hannah female Bangkok OHA1 AB920180 AB920234 Elapidae Ophiophagus hannah female Bangkok OHA2 AB920181 AB920235 Elapidae Ophiophagus hannah female Bangkok OHA3 AB920182 AB920236 Elapidae Naja kaouthia male unknow NKA1 AB920183 AB920237 Elapidae Naja kaouthia female Bangkok NKA2 AB920184 AB920238 Elapidae Naja kaouthia male Bangkok NKA3 AB920185 AB920239 Elapidae Naja sumatrana female unknow NSU1 AB920186 AB920240 Elapidae Naja siamensis male unknow NSI2 AB920187 AB920241 Elapidae Naja siamensis male Rayong NSI3 AB920188 AB920242 Elapidae Bungarus fasciatus female unknow BFA1 AB920189 AB920243 Elapidae Bungarus fasciatus female unknow BFA2 AB920190 AB920244 Elapidae Bungarus fasciatus female unknow BFA3 AB920191 AB920245 Elapidae Bungarus candidus male NakhonRatchasima BCA1 AB920192 AB920246 Elapidae Bungarus candidus female unknow BCA2 AB920193 AB920247 Elapidae Bungarus candidus female unknow BCA3 AB920194 AB920248 Elapidae Hydrophis brookii male unknow HBR1 AB920212 AB920266 Elapidae Hydrophis brookii female unknow HBR5 AB920213 AB920267 Elapidae Hydrophis brookii female unknow HBR6 AB920214 AB920268 Elapidae Hydrophis obscurus male unknow HOB2 AB920215 AB920269 Elapidae Hydrophis obscurus male unknow HOB3 AB920216 AB920270 Elapidae Hydrophis obscurus male unknow HOB4 AB920217 AB920271 Elapidae Hydrophis obscurus female unknow HOB5 AB920218 AB920272 Elapidae Hydrophis obscurus female unknow HOB8 AB920219 AB920273 Elapidae Hydrophis obscurus female unknow HOB15 AB920220 AB920274 Elapidae Enhydrina schistosa female unknow ESC4 AB920221 AB920275 Elapidae Enhydrina schistosa female unknow ESC5 AB920222 AB920276 Elapidae Enhydrina schistosa female unknow ESC6 AB920223 AB920277 Elapidae Enhydrina schistosa male unknow ESC7 AB920224 AB920278 Elapidae Enhydrina schistosa female unknow ESC9 AB920225 AB920279 Elapidae Enhydrina schistosa male unknow ESC12 AB920226 AB920280 Elapidae Enhydrina schistosa female unknow ESC13 AB920227 AB920281 Viperidae Daboia siamensis male Bangkok DSI1 AB920195 AB920249

20 Table 1 (Continued) Family Species Sex Locality Code COI Accession number Cytb Accession number Viperidae Daboia siamensis female Bangkok DSI2 AB920196 AB920250 Viperidae Viperidae Trimeresurus albolabris female Bangkok Trimeresurus albolabris female Chanthaburi TAL3 AB920197 AB920251 TAL4 AB920198 AB920252 Viperidae Trimeresurus macrops female NongBuaLamphu TMA2 AB920199 AB920253 Viperidae Trimeresurus macrops female Bangkok TMA7 AB920200 AB920254 Viperidae Trimeresurus macrops female Sara Buri TMA14 AB920201 AB920255 Viperidae Trimeresurus macrops female RajchaBuri TMA19 AB920202 AB920256 Viperidae Trimeresurus macrops female RajchaBuri TMA20 AB920203 AB920257 Viperidae Trimeresurus macrops female RajchaBuri TMA21 AB920204 AB920258 Viperidae Trimeresurus macrops female RajchaBuri TMA24 AB920205 AB920259 Viperidae Trimeresurus macrops female NakhonRatchasima TMA31 AB920206 AB920260 Viperidae Trimeresurus macrops male unknow TMA35 AB920207 AB920261 Viperidae Trimeresurus macrops female unknow TMA36 AB920208 AB920262 Viperidae Trimeresurus macrops female unknow TMA37 AB920209 AB920263 Viperidae Trimeresurus macrops - unknow TMA39 AB920210 AB920264 Viperidae Calloselasma rhodostoma female Chon Buri CRH1 AB920211 AB920265 Colubridae Boiga cynodon female unknow BCY4 AB920228 AB920282 Colubridae Boiga cynodon male unknow BCY5 AB920229 AB920283 Colubridae Colubridae Colubridae Rhabdophis subminiatus male NakhonRatchasima Rhabdophis subminiatus female NakhonRatchasima Rhabdophis subminiatus female unknow RSU1 AB920230 AB920284 RSU2 AB920231 AB920285 RSU3 AB920232 AB920286 Pythonidae Python molurus - unknow PMO1 AB920233 AB920287 3. PCR amplification and sequencing The partial DNA fragments of the COI and Cytb mitochondrial genes were then amplified using PCR primers (Table 2). Standard PCR reaction was performed using 1 ThermalPoll reaction buffer containing 1.5 mm MgCl 2, 0.2 mm dntps, 5

21 pm specific primers, 0.25 U of Taq polymerase (Vivantis Technologies, Selangor Darul Ehsan, Malaysia), and 25 ng genomic DNA in a final reaction volume of 20 µl. PCR cycling conditions contained the initial denaturation at 94 C for 3 min, followed by 35 cycles of denaturation at 94 C for 30 s, primer annealing at 56 C of COI primers and 49 C of Cytb primers for 30 s, primer extension at 72 C for 45 s, then postcycling extension at 72 C for 10 min. The PCR products were examined by electrophoresis on 1% agarose gel. The DNA fragments were subsequently be extracted from the ethidium bromide-stained gel using Gel/PCR DNA fragments Extraction Kit (Geneaid, Sijhih, Taipei, Taiwan), and directly sequenced by 1 st Base DNA sequencing service (Seri Kembangan, Malaysia). Table 2 Primers used for the amplification of mitochondrial COI and Cytb genes in this study. Primer name Targeted gene Forward primer (5'-3') Specific snake Reference Amplified sized COI forward COI TCAGCCATACTCCTGTGTTCA all snakes Makowsky et al. 700 (2010) COI reverse COI TAGACTTCTGGGTGGCCAAAGAATCA all snakes Makowsky et al. 700 (2010) H16064 Cytb CTTTGGTTTACAAGAACAATGCTTTA all snakes Palumbi (1996) 1100 Gludg Cytb TGACTTGAARAACCAYCGTTG all snakes Burbrink (2000) 1100 HYD COI AGGGCCCTGAGTGAACAATA snake in Hydrophiinae in this study 411 TC COI GAAAGCCATGTCTGGGGTT Trimeresurus spp. in this study 272 DB COI TAATAGCATAGTAATTGCTGCTGCAAG Daboia siamensis in this study 621 DRS COI CGGATCAAACAAATAGAGGGAAGTTAA Daboia siamensis and in this study 579 Bungarus candidus BF COI AGGACTGGTAGGGCTAGTAAA Bungarus fasciatus in this study 590 OH COI TAAATGCGTGGGCAGTTACTAAA Ophiophagus hannah in this study 181 NK COI AGAGAAGTAGGAGGGATGGAGG Naja spp. in this study 322 CR COI CGAGTGAACTAGGTTTCCGG Calloselasma rhodostoma in this study 414 4. Nucleotide sequence and phylogenetic analysis The COI and Cytb nucleotide sequences of all specimens were used to search the homologies of mitochondrial genes in the National Center for Biotechnology

22 Information (NCBI) database using the BLASTx and BLASTn programs, and were deposited in the DNA Data Bank of Japan (http://www.ddbj.nig.ac.jp/index-e.html). The partial mitochondrial COI and Cytb sequences of 54 venomous snakes and P. molurus were aligned using the default parameters of Molecular Evolutionary Genetics Analysis 4 (MEGA4) software (Kumar et al., 2004). All unalignable sites and gap-containing sites were carefully removed from these data sets. Sequence divergence was conducted with two data sets (COI and Cytb nucleotide sequences) with uncorrected pairwise distances (p-distance) using MEGA4. The phylogenetic trees were reconstructed using Bayesian inference (BI) with MrBayes v3.0b4 (Huelsenbeck and Ronquist, 2001). The corresponding best-fit evolutionary model and parameters indicated by Modeltest version 3.7 (Posada and Crandall, 1998) based on the Akaike Information Criterion (AIC) and PAUP* v. 4.0b10 (Swofford, 2002) were used. The Markov Chain Monte Carlo (MCMC) process was set to run four chains simultaneously for 1 million generations. After the log-likelihood value reached stationarity, sampling procedure was performed at every 100 generations to get 10,000 trees and subsequently provide a majority-rule consensus tree with averaged branch lengths. All sample points prior to reaching convergence were discarded as burn-in, and Bayesian posterior nodal relationship in the sampled tree population was shown in percentages. 5. Species-specific multiplex PCR markers The nucleotide sequences of partial mitochondrial COI gene of sixteen species of venomous snakes (O. hannah, N. siamensis, N. sumatrana, N. kaouthia, B. candidus, B. fasciatus, D. siamensis, R. subminiatus, B. cynodon, T. albolabris, T. macrops, C. rhodostoma, H. brookii, H. obscurus, E. schistosa, and P. molurus.) were aligned using the default parameter of ClustalW. Species-specific primers were designed based on the sites of all mitochondrial COI nucleotide sequences of the venomous snakes in this study that differed between different clusters of snake species. Primers were examined using two different approaches:

23 1. Single primer pairs PCRs were used to examine different snake clades (Naja spp. (1); O. hannah (2); Trimeresurus spp. (3); Hydrophiinae (4); D. siamensis (5); B. fasciatus (6); B. candidus (7); and C. rhodostoma (8)) using the appropriate annealing temperature (Table 2). PCR reaction was performed using 1 ThermalPoll reaction buffer containing 1.5 mm MgCl 2, 0.2 mm dntps, 5 pm specific primers, 0.25 U of Taq polymerase (Vivantis), and 25 ng genomic DNA in a final reaction volume of 15 µl. PCR cycling conditions contained the initial denaturation at 94 C for 3 min, followed by 35 cycles of denaturation at 94 C for 30 s, specific primer annealing temperature for 30 s, primer extension at 72 C for 45 s, then postcycling extension at 72 C for 5 min. The PCR products were examined by electrophoresis on 1% agarose gel. 2. Single tube multiplex PCRs were performed to detect simultaneously different snake species group. Multiplex PCR reaction was performed using 1 ThermalPoll reaction buffer containing 1.5 mm MgCl 2, 0.2 mm dntps, specific primers concentration (Table 3), 0.25 U of Taq polymerase (Vivantis), and 25 ng genomic DNA in a final reaction volume of 20 µl. PCR cycling conditions contained the initial denaturation at 94 C for 3 min, followed by 35 cycles of denaturation at 94 C for 30 s, specific primer annealing temperature for 30 s, primer extension at 72 C for 45 s, then postcycling extension at 72 C for 5 min. The PCR products were examined by electrophoresis on 1% agarose gel. To test the sensitivity of species-specific single PCR markers and multiplex PCR markers, serial dilutions of the DNA template ranging from 0.0001 ng to 5 ng were subjected to PCR. The specificity of the single and multiplex PCR systems was also determined by adding DNA mixture from all snake species to each PCR reaction, and the amplification was performed as mentioned above. Snake venom, saliva and snakebites were also collected from almost all snake species, except for B. cynodon, R. subminiatus, P. molurus, and snakes of Hydrophiinae. Whole-genomic DNA was isolated as described above. Multiplex PCR reactions were performed with 10 15 ng of DNA template, and the PCR products were also examined by electrophoresis on 1% agarose gel.

24 RESULTS AND DISCUSSION 1. DNA Extraction and Gene amplification with PCR Whole genomic DNA was extracted from blood, snake venom, saliva, and snakebites samples. The results showed that the DNA concentrations were 11.58 1296.35 ng/µl, and DNA purities were between 1.80 2.00 in absorbance at 260/280 nm. Then, DNA quality were determined by 1% agarose gel electrophoresis. Genomic DNA was then used as templates to amplify two partial mitochondrial COI and Cytb genes under PCR based method (Table 2). The results showed that size of DNA fragments of COI gene was about 700 bp and size of DNA fragments of Cytb gene was about 1,100 bp in all snake species (Figure 11). Figure 11 Agarose gel eletrophoresis of PCR products using COI primer (a) and Cytb primer (b). Lane 1 - O. hannah, Lane 2 - N. kaouthia, Lane 3 - B. fasciatus, Lane 4 - B. candidus, Lane 5 - D. siamensis, Lane 6 - T. albolabris

25 2. Barcode construction DNA barcodes of mitochondrial COI and Cytb gene sequences were constructed from 54 specimens representing 16 species, belonging to 11 genera (O. hannah, N. siamensis, N. sumatrana, N. kaouthia, B. candidus, B. fasciatus, D. siamensis, R. subminiatus, B. cynodon, T. albolabris, T. macrops, C. rhodostoma, H. brookii, H. obscurus, E. schistosa, and P. molurus.). Most of all species were represented by two or more samples. Good quality sequences of approximately 528 bp for COI gene sequences and 396 bp for Cytb gene sequences were obtained from all samples used in this study. Average nucleotide frequencies for the entire data set were A =26.49%; T =30.51%; C =26.60%; G =16.39% for COI data set and A =31.60%; T =30.22%; C =29.57%; G =8.6% for Cytb data set. There were also no statistically significant proportion differences among snake sequences, indicating that the two data set analyses had no heterogeneity of base frequencies. No stop codons were found in these sequences, indicating that nuclear pseudogenes had not occurred in the analysis. All barcodes represent the functional mitochondrial COI and Cytb gene sequences. 3. Sequence analysis The sequences were deposited in DDBJ (Accession Numbers AB920180 AB920287). Sequence divergences were then conducted with two data sets (COI and Cytb nucleotide sequences), and estimated with MEGA4 using uncorrected pairwise distances (p-distance). Different COI haplotypes were found in N. siamensis, B. fasciatus, B. candidus, T. macrops, T. albolabris, H. brookii, E. schistosa and H. obscurus (Table 3) with the mean intraspecific uncorrected pairwise distance (pdistance) of 0.2%, 0.9%, 0.6%, 3.6%, 1.52%, 0.13%, 0.7% and 0.1%, respectively (Table 4). By contrast, O. hannah, N. kaouthia, D. siamensis, B. cynodon and R. subminiatus had each only one haplotype. The overall mean sequence divergence was 16.7%, and the average distance of intrafamily Elapidae, Viperidae, and Colubridae were 13.1%, 10% and 10.8%, respectively. The analysis of Cytb sequences showed that various Cytb haplotypes were found in O. hannah, N. kaouthia, N. siamensis, B.

26 fasciatus, B. candidus, T. macrops, T. albolabris, E. schistosa, H. obscurus and R. subminiatus (Table 3) with the mean intraspecific uncorrected pairwise distance (pdistance) of 0.4%, 1.1%, 4%, 0.9%, 14.4%, 6.1%, 1.1%, 0.1%, 0.3% and 0.5%, respectively (Table 5). By contrast, D. siamensis, H. brookii and B. cynodon had each only one haplotype. The overall mean sequence divergence was 22.8%, and the average distance of intrafamily Elapidae, Viperidae, and Colubridae were 17.4%, 12.3% and 14.1%, respectively. The results showed that the minimum of sequence divergence for COI gene was 3.1% between N. siamensis and N. sumatrana and maximum of sequence divergence was 21.9% between N. kaouthia and D. siamensis. Generally, the sequence divergence of intraspecific variation is approximately 2% which is low for the COI gene (Avise 2000; Hebert et al., 2003). The variation of sequence divergence of sea turtle in Brazil have been revealed between 6.3 13.9% (mean 8.2%) (Vargas et al., 2009). The minimum of sequence divergence for Cytb gene was 5.1% between H. brookii and H. obscurus and maximum of sequence divergence was 32.7% between B. candidus and P. molurus. Generally, the sequence divergence of intraspecific variation is approximately 3% (Johns and John, 1998). The variation of sequence divergence of genus Natrix between N. maura and N. natrix was about 16.0% and N. maura and N. tessellate was about 18.4% (Guicking et al., 2006). Therefore, the sequence divergence of COI and Cytb barcodes in this study were similar to those of other vertebrates, and can be used for identifying each snake species.

27 Table 3 Number of haplotypes of COI and Cytb sequences Species COI haplotypes Cytb haplotypes Ophiophagus hannah 1 3 Naja. kaouthia 1 3 Naja. siamensis 2 2 Bungarus. fasciatus 2 2 Bungarus. candidus 2 3 Trimeresurus. macrops 6 8 Trimeresurus. albolabris 2 2 Hydrophis. brookii 3 1 Enhydrina. schistosa 2 4 Hydrophis. obscurus 4 5 Rhabdophis. subminiatus 1 2 Daboia. siamensis 1 1 Boiga. cynodon 1 1

28 Table 4 Average sequence divergence (p-distance) of COI gene within species and between pairs of venomous snake species. Diagonal values are p-distance for intra-specific comparisons. Species O. hannah N. kaouthia N. sumatrana N. siamensis B. fasciatus B. candidus H. brookii H. obscurus E. schistosa D. siamensis T. albolabris T. macrops C. rhodostoma B. cynodon R. subminiatus P. molurus O. hannah 0.000 (1) N. kaouthia 0.156 0.000 (1) N. sumatrana 0.150 0.067 a N. siamensis 0.153 0.054 0.031 0.002 (2) B. fasciatus 0.146 0.172 0.165 0.161 0.009 (2) B. candidus 0.165 0.159 0.151 0.145 0.127 0.006 (3) H. brookii 0.191 0.198 0.193 0.186 0.169 0.167 0.001 (2) H. obscurus 0.189 0.189 0.183 0.169 0.166 0.160 0.052 0.001 (2) E. schistosa 0.190 0.179 0.174 0.173 0.170 0.168 0.064 0.064 0.007 (5) D. siamensisi 0.188 0.219 0.207 0.202 0.199 0.194 0.217 0.208 0.203 0.000 (1) T. albolabris 0.151 0.182 0.183 0.184 0.183 0.177 0.214 0.210 0.212 0.181 0.015 (2) T. macrops 0.211 0.205 0.195 0.200 0.196 0.196 0.211 0.202 0.203 0.192 0.131 0.036 (6) C. rhodostoma 0.165 0.177 0.171 0.178 0.186 0.181 0.207 0.206 0.201 0.192 0.164 0.175 a B. cynodon 0.183 0.194 0.181 0.191 0.172 0.179 0.200 0.208 0.196 0.211 0.196 0.205 0.181 0.000 (1) R. subminiatus 0.183 0.203 0.183 0.185 0.175 0.172 0.203 0.189 0.196 0.207 0.207 0.211 0.186 0.181 0.000 (1) P. molurus 0.186 0.190 0.188 0.185 0.189 0.186 0.217 0.209 0.209 0.196 0.190 0.201 0.198 0.196 0.207 a The numbers in parentheses indicate the number of haplotypes per species. a Samples were collected from only one individual.

29 Table 5 Average sequence divergence (p-distance) of Cytb gene within species and between pairs of venomous snake species. Diagonal values are p-distance for intra-specific comparisons. Species O. hannah N. kaouthia N. sumatrana N. siamensis B. fasciatus B. candidus H. brookii H. obscurus E. schistosa D. siamensis T. albolabris T. macrops C. rhodostoma B. cynodon R. subminiatus P. molurus O. hannah 0.004 (2) N. kaouthia 0.194 0.011 (2) N. sumatrana 0.189 0.097 a N. siamensis 0.208 0.112 0.057 0.040 (2) B. fasciatus 0.194 0.244 0.215 0.238 0.009 (3) B. candidus 0.258 0.295 0.260 0.266 0.208 0.144 (3) H. brookii 0.215 0.235 0.214 0.223 0.200 0.260 0.000 (1) H. obscurus 0.205 0.222 0.211 0.226 0.193 0.256 0.051 0.003 (4) E. schistosa 0.228 0.230 0.222 0.242 0.210 0.284 0.081 0.075 0.001 (2) D. siamensis 0.252 0.270 0.274 0.286 0.265 0.288 0.266 0.267 0.263 0.000 (1) T. albolabris 0.226 0.288 0.298 0.307 0.269 0.289 0.288 0.264 0.267 0.240 0.011 (2) T. macrops 0.234 0.277 0.282 0.292 0.262 0.279 0.301 0.286 0.295 0.218 0.139 0.061 (8) C. rhodostoma 0.217 0.263 0.266 0.285 0.233 0.284 0.274 0.267 0.282 0.216 0.214 0.196 a B. cynodon 0.223 0.260 0.261 0.284 0.246 0.314 0.251 0.244 0.243 0.256 0.251 0.258 0.237 0.000 (1) R. subminiatus 0.233 0.286 0.279 0.285 0.256 0.305 0.265 0.251 0.255 0.271 0.281 0.295 0.257 0.233 0.005 (3) P. molurus 0.257 0.293 0.282 0.298 0.285 0.327 0.274 0.263 0.264 0.293 0.272 0.274 0.259 0.253 0.288 a The numbers in parentheses indicate the number of haplotypes per species. a Samples were collected from only one individual..

30 4. Phylogenetic analysis Bayesian inference cladograms were reconstructed based on COI and Cytb data sets. They produced similar topology, and strongly supported the monophyletic clade of the sequences belonging to the same species from both Genbank and this study (Figure 12, 13). This indicates the diagnostic ability of COI and Cytb to correctly identify species. However, almost all individual snakes of T. macrops and T. albolabris were not included within the monophyletic group in either COI or Cytb data sets.

31 Figure 12 Phylogenetic tree shows relationship between 16 species of snake COI gene.

32 Figure 13 Phylogenetic tree shows relationship between 16 species of snake Cytb gene. There were nucleotide sequence variation of partial mitochondrial COI and Cytb genes in most snake species in this study, but the mtdna-based identification enabled differentiation between intraspecific and interspecific variation. The level of intraspecific sequence divergence of COI gene was low (0.59%). This is similar to the

33 intraspecific sequence divergence of 0.30% reported for N. siamensis (Wuster and Thorpe, 1994). In vertebrates, the sequence divergence of intraspecific variation is about 2%, whereas the sequence divergence of interspecific variation ranges from 4% to 32% (mean = 9.6%; Avise 2000; Hebert et al., 2003). By contrast, the level of sequence divergence of Cytb gene was about 2.22% for intraspecific and 10.47% for interspecific distance (Ophiophagus, Naja, Bungarus, Daboia, Trimeresurus, Hydrophis, Enhydrina, Boiga, Rhabdophis, Calloselasma, and Python) in our snake data set. This result agrees with that obtained for African cobras, for which the range is reported to be 0 11% (Wuster et al., 2007). However, these percentages are high compared with the interspecific distances between mammals and reptiles in general, which are approximately 3% (Johns and Avise, 1998). Among turtles, the intraspecific distances are mostly lower than 1%, with an interspecific divergence rate of more than 8% (Shen et al., 2013). This implies that, in the snake lineage, the mutation rate of the COI gene has been slower than that of the Cytb gene over the course of evolution. In terms of the barcoded cut-off scores for snake identification, these might vary depending on the specimen collection, the gene studied, and its length (Dawnay et al., 2007). In this study, we applied three analytical methods: (1) 98% identity (Dove et al., 2008; Eaton et al., 2010), (2) 10X rule threshold values (Hebert et al., 2004), and (3) 3% threshold values (Hebert et al., 2003). First of all, the DNA sequences of the unidentified organism must be at least 98% similar to the most common haplotype of the described taxon. However, using this criterion, more than three snake species were misidentified at the intraspecific level regardless of whether the COI or Cytb genes were considered. To employ 10X rule threshold values, the interspecific distance should be 10 times the intraspecific distance in order to differentiate between species. Our standard COI barcode sequence divergence between species within a genus (average: 5.10%) was about 10 times the average divergence within species (0.59%; Table 4). In contrast, the average interspecific sequence divergence of the Cytb gene within each genus (average: 10.47%) was about four times the intraspecific distance (2.22%; Table 5). This result suggests that the interval divergence at the species and genus levels was greater for the COI gene than for the Cytb gene. This indicates that the COI gene may be a more suitable marker than the Cytb gene for species identification in snakes.

34 However, the barcoded cut-off scores of the COI barcode were insufficient to differentiate species of Naja spp, and the species in Hydrophiinae which suggests that these species recently diverged from one another (Wüster et al., 2007; Pyron et al., 2013.). Hebert et al. (2003) asserted that, in order to differentiate among species, interspecific divergence should be at least 3%. This threshold agrees with our COI barcode, which can be used to differentiate all snake species in this study, except for T. macrops, which has a high degree of intraspecific nucleotide divergence (3.6%). To determine the barcoding gap, which can be used to monitor the difference between intraspecific and interspecific divergence, graphic representations were produced by plotting the minimum interspecific divergence on the y axis and the maximum intraspecific divergence on the x axis for both COI and Cytb barcodes (Rasmussen et al., 2009). There was a distinct barcoding gap between intra- and interspecific divergence, even though the species from Naja and Hydrophiinae were not differentiated using 10X rule threshold values (Figure. 14). However, T. macrops could not be differentiated using the COI barcoding gap, and T. macrops and B. candidus could not be differentiated using the Cytb barcoding gap. One green viper specimen (T. macrops: TMA2) was collected from Nong Bua Lamphu (17 10 0 N, 102 23 0 E) in the northeast of Thailand, whereas the other T.macrops specimens were collected from the central regions. This suggests that the barcoding gap decreased when species were sampled across a broad geographic area, owing to intraspecific divergence. While BI clustering analyses of both COI and Cytb indicated that most species of the same genus and family clustered together, one T. macrops individual (TMA2) was positioned in the clade of T. albolabris (Figure 12, 13). The snakes in the genus Trimeresurus (Asian green pit vipers) are one of the most diverse radiations of pit vipers; they inhabit southern Asia and the Indo Malay Archipelago. The species complex is remarkable for its extreme morphological similarity, occupies a wide range of ecological habitats, and exhibits diverse life histories and behaviors (Malhotra et al., 2011). Our results suggest that the Trimeresurus species complex has frequently been misidentified. Interspecific divergence of the COI gene between TMA2 and two T. albolabris individuals was 1%, whereas divergence between TMA2 and other T. macrops specimens was 13% 15% for the COI gene. These results collectively suggest that TMA2 is a variant of T. albolabris

35 or is a cryptic species that is closely related to T. albolabris. However, morphologic studies and more specimens are needed to examine the barcode of Trimeresurus spp. Figure 14 Dot plot analysis of sequence divergence with COI barcode (a) and Cytb barcode (b). A straight line represents 1:1 ratio between minimum congeneric sequence divergence and maximum intraspecific sequence divergence. Sequence divergences of species which was represented by more than one sample are shown here.

36 5. Species-specific multiplex PCR markers Interspecific nucleotide sequence differences were observed to design eight species-specific reverse primers. These primers were paired with the forward primer (COI Forward), and amplified into eight snake groups according to antivenom: Naja spp. (1); O. hannah (2); Trimeresurus spp. (3); Hydrophiidae snakes (4); D. siamensis (5); B. fasciatus (6); B. candidus (7); and C. rhodostoma (8) (Table 2). Whereas the primer NK, OH, BF, DB, TC, HYD, and CR amplified specifically for group (1) (6) and (8), the primer DRS could amplify the two target species, D. siamensis (5) and B. candidus (7). For more efficient analysis, the multiplex-pcr assay was developed to differentiate all eight snake groups with two sequential panels. Each multiplex PCR panel allows the rapid identification between individuals assigned to each of the groups. Firstly, the multiplex PCR (panel A) was performed using three speciesspecific primers: HYD, TC, and DRS in combination with COI Forward (used as common forward primer) allowing the following diagnostic bands: (i) one bands of 418 bp for snakes in Hydrophiidae (4), (ii) one band of 277 bp for Trimeresurus spp. (3), and (iii) one band of 621 bp for both D. siamensis (5) and B. candidus (7) (Figure 15a). However, cross-reaction was detected in O. hannah with the two DNA bands at 500 and 600 bp, indicating that non-specific band was derived from the combination of multiple primers in the PCR reaction. Secondly, multiplex PCR (panel B) comprised five primers: DB; BF; OH; NK; and CR in combination with COI Forward. The species-specific PCR products of 541 bp, 599 bp, 176 bp, 323 bp, and 400 bp could be observed from D. siamensis (5), B. fasciatus (6), O. hannah (2), Naja spp. (1), and C. rhodostoma (8), respectively, but no PCR products were found from B. candidus (Figure. 15b). However, cross-species amplification was detected in T. macrops with the two DNA bands at 400 and 500 bp, indicating that non-specific bands resulted from the combination of multiple primers in the PCR reaction. Specificity was also performed in both single PCR and multiplex PCR assays by pooling the DNA of all snake species. The results indicated the DNA fragments from all species were specifically detected as mentioned above (Figure 16 and 17). The sensitivity of each species-specific DNA markers and each multiplex PCR markers were examined by detecting the limitation of those specific primers in amplification

37 of serially diluted DNA templates from 25, 1, 0.5, 0.1, 0.05, 0.01, 0.005, 0.001, 0.0005, and 0.0001 ng (Figure 18, 19). The results showed that the minimum detection limits was 0.001 ng for COI Forward-BF primers. By contrast, the limitation of detection for multiplex PCRs was 0.5 ng with the species-specific PCR products of panel A and panel B (Figure 20). Figure 15 Agarose gel eletrophoresis of PCR products using multiplex PCR assays with the primer set 1 (panel A) (a) and 2 (panel B) (b) and DNA extracted from blood. Lane 1 H. brookii. Lane 2 E. schistosa. Lane 3 H.obscurus. Lane 4 T. albolabris. Lane 5 T. macrops. Lane 6 D. siamensis. Lane 7 B. candidus. Lane 8 B. fasciatus. Lane 9 O. hannah. Lane 10 N. siamensis. Lane 11 N. sumatrana. Lane 12 N. kaouthia. Lane 13 C. rhodostoma. Lane 14 Mus musculus. Lane 15 - Homo sapiens.

38 Figure 16 Agarose gel eletrophoresis of PCR products using primer COI Forward and (a) HYD Reverse, (b) TC Reverse, (c) DRS Reverse and (d) DB Reverse.Lane 1 H. brookii. Lane 2 E. schistosa. Lane 3 H. obscurus. Lane 4 T. albolabris. Lane 5 T. macrops. Lane 6 D. siamensis. Lane 7 B. candidus. Lane 8 B. fasciatus. Lane 9 O. hannah. Lane 10 N. siamensis. Lane 11 N. sumatrana. Lane 12 N. kaouthia. Lane 13 C. rhodostoma. Lane 14 Mus musculus. Lane 15 - Homo sapiens.

39 Figure 17 Agarose gel eletrophoresis of PCR products using primer COI Forward and (a) BF Reverse, (b) OH Reverse, (c) NK Reverse, (d) CR Reverse. Lane 1 H. brookii. Lane 2 E. schistosa. Lane 3 H. obscurus. Lane 4 T. albolabris. Lane 5 T. macrops. Lane 6 D. siamensis. Lane 7 B. candidus. Lane 8 B. fasciatus. Lane 9 O. hannah. Lane 10 N. siamensis. Lane 11 N. sumatrana. Lane 12 N. kaouthia. Lane 13 C. rhodostoma. Lane 14 Mus musculus. Lane 15 - Homo sapiens.

40 Table 6. Summary of species-specific multiplex PCR assay. Family Species name Common name Multiplex PCR set Antivenom panel A DNA banded size (bp) panel B DNA banded size (bp) Elapidae Ophiophagus hannah King Cobra a - + 176 Ophiophagus Elapidae Naja kaouthia Monocled cobra - - + 323 Naja spp. Elapidae Naja sumatrana Sumatran Cobra - - + 323 Naja spp. Elapidae Naja siamensis Indo-Chinese spitting cobra - - + 323 Naja spp. Elapidae Bungarus fasciatus Banded krait - - + 599 Bungarus fasciatus Elapidae Bungarus candidus Malayan krait + 621 - - Bungarus candidus Elapidae Hydrophis brookii Brooke's Sea Snake + 418 - - b Elapidae Hydrophis obscurus Russell's Sea Snake + 418 - - b Elapidae Enhydrina schistosa Beaked sea snake + 418 - - b Viperidae Daboia siamensis Russell's viper + 621 + 541 Daboia siamensis Viperidae Trimeresurus albolabris White-lipped Pit Viper + 277 a - Trimeresurus spp. Viperidae Trimeresurus macrops Big-eyed pit viper + 277 a - Trimeresurus spp. Viperidae Calloselasma rhodostoma Malayan pit viper - - + 400 Calloselasma rhodostoma a Non-specific DNA bands are found in the PCR reaction. b There is no any specific antivenom which is available in Thailand. + It is possible to identify with appearance of a specific DNA band. - Impossible to identify.

41 Figure 18 Agarose gel eletrophoresis of PCR sensitivity using primer COI Forward and (a) HYD Reverse, (b,c) TC Reverse, (d,e) DRS Reverse. (f,g) DB Reverse Lane 1 25ng. Lane 2 1ng. Lane 3 0.5ng. Lane 4 0.1 ng. Lane 5 0.05 ng. Lane 6 0.01 ng. Lane 7 0.005 ng. Lane 8 0.001 ng. Lane 9 0.0005 ng. Lane 10 0.0001 ng.

42 Figure 19 Agarose gel eletrophoresis of PCR sensitivity using primer COI Forward and (a) BF Reverse, (b) OH Reverse, (c,d,e) NK Reverse, (f) CR Reverse. Lane 1 25ng. Lane 2 1ng. Lane 3 0.5ng. Lane 4 0.1 ng. Lane 5 0.05 ng. Lane 6 0.01 ng. Lane 7 0.005 ng. Lane 8 0.001 ng. Lane 9 0.0005 ng. Lane 10 0.0001 ng.

43 Figure 20 Agarose gel eletrophoresis of PCR products using multiplex PCR sensitivity assays with the primer set 1 (panel A) (a) and 2 (panel B) (b) and DNA concentration 0.5 ng. Lane 1 H. brookii. Lane 2 E. schistosa. Lane 3 H. obscurus. Lane 4 T. albolabris. Lane 5 T. macrops. Lane 6 D. siamensis. Lane 7 B. candidus. Lane 8 B. fasciatus. Lane 9 O. hannah. Lane 10 N. siamensis. Lane 11 N. sumatrana. Lane 12 N. kaouthia. Lane 13 C. rhodostoma. Lane 14 Mus musculus. Lane 15 - Homo sapiens. Species-specific multiplex PCR markers were applied to examine with DNA extracted from venom and saliva samples from 10 species (O. hannah, N. kaouthia, B. fasciatus, B. candidus, N. sumatrana, D. siamensis, T. albolabris, T. macrops, C. rhodostoma and N. siamensis) (Figure 21, 22). The results showed that DNA extracted from venom and saliva were consistent with the multiplex PCR with DNA extracted from blood.

44 Figure 21 Agarose gel eletrophoresis of PCR products using multiplex PCR assays with the primer set 1 (panel A) (a) and 2 (panel B) (b) and DNA extracted from venom. Lane 1 T. albolabris. Lane 2 T. macrops. Lane 3 D. siamensis. Lane 4 B. candidus. Lane 5 B. fasciatus. Lane 6 O. hannah. Lane 7 N. siamensis. Lane 8 N. sumatrana. Lane 9 N. kaouthia. Lane 10 C. rhodostoma. Lane 11 Mus musculus. Lane 12 Homo sapiens.

45 Figure 22 Agarose gel eletrophoresis of PCR products using multiplex PCR assays with the primer set 1 (panel A) (a) and 2 (panel B) (b) and DNA extracted from saliva. Lane 1 T. albolabris. Lane 2 T. macrops. Lane 3 D. siamensis. Lane 4 B. candidus. Lane 5 B. fasciatus. Lane 6 O. hannah. Lane 7 N. siamensis. Lane 8 N. sumatrana. Lane 9 N. kaouthia. Lane 10 C. rhodostoma. Lane 11 Mus musculus. Lane 12 Homo sapiens. Species-specific multiplex PCR markers were applied to examine with DNA extracted from snake bite-site swabs (Figure 23). The results showed that DNA extracted from bite-site swabs were not consistent with the multiplex PCR with DNA

46 extracted from blood. This might result from low amont and quality of gdna from snake bite-site swabs. The new primers should be designed to amplify the PCR product of smaller sizes for easier and better amplification.

47 Figure 23 Agarose gel eletrophoresis of PCR products using multiplex PCR assays with the primer set 1 (panel A) (a) and 2 (panel B) (b). DNA extracted from bite-site swabs with DNA concentration 2 ug. Lane 1-3 T. albolabris. Lane 4-6 T. macrops. Lane 7-9 D. siamensis. Lane 10-12 B. candidus. Lane 13-15 B. fasciatus. Lane 16-18 O. hannah. Lane 19-21 N. siamensis. Lane 22-24 N. sumatrana. Lane 25-27 N. kaouthia. Lane 28-30 C. rhodostoma. To develop DNA marker to differentiate among snakes for which antivenom is available in Thailand, Species-distinguishing sites were investigated in the COI barcode for snake identification. Eight DNA markers (HYD, TC, DRS, DB, BF, OH, NK and CR) were designed based on distinctive sites of COI sequences. These markers were successfully used to distinguish Naja spp (1), O. hannah (2),