Kasetsart J. (Nat. Sci.) 44 : 424 - 435 (2010) Genetic Relationship of Three Butterfly Lizard Species (Leiolepis reevesii rubritaeniata, Leiolepis belliana belliana, Leiolepis boehmei, Agamidae, Squamata) Inferred from Nuclear Gene Sequence Analyses Kornsorn Srikulnath1, 2, Kazumi Matsubara3, Yoshinobu Uno2, Amara Thongpan1, Saowanee Suputtitada1, Chizuko Nishida2, 3, Yoichi Matsuda2, 3, 4 and Somsak Apisitwanich1* ABSTRACT The genetic relationship was investigated of three butterfly lizard species (Leiolepis reevesii rubritaeniata, L. belliana belliana and L. boehmei) selectively inhabiting Thailand. The findings were based on RAG1 and C-mos gene analyses. The DNA sequences were also compared with the other squamate reptiles. The analysis strongly supported that L. reevesii rubritaeniata was related more closely to L. belliana belliana than to L. boehmei. The phylogenetic position of Leiolepis spp., however, was contentious with regard to its relationship among the Leiolepidinae, Agaminae and Chamaeleonidae, which suggested that their phylogeny remains uncertain. Keywords: butterfly lizard, Leiolepidinae, phylogeny, RAG1, C-mos INTRODUCTION inhabit Southeast Asia. They show a great variety of karyotypes and sexual systems. In Thailand, The Squamata is the most diverse there are three species, which barely can be reptilian order that has been classified traditionally discriminated from other congeneric species by into three suborders: Serpentes (snakes), their typical scale and skin coloration (Peters, Amphisbaenia (worm lizards) and Lacertilia 1971). Bisexualism has been described in Leiolepis (lizards). The extant lizards can be further belliana belliana (2n=2x=36), which is widely categorized into five infraorders (the Iguania, found throughout the country, L. belliana ocellata Gekkota, Scincomorpha, Diploglossa, Dibamia, (2n=2x=34) found in upper northern, and L. Platynota) (Uetz, 2009). Butterfly lizards reevesii rubritaeniata (2n=2x=36) which is (Agamidae, Iguania) are burrow diggers and distributed only in the northeast (Aranyavalai, 1 Department of Genetics, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 2 Biosystems Science Course, Graduate School of Life Science, Hokkaido University, North 10 West 8, Kita-ku, Sapporo 060- 0810, Japan. 3 Department of Biological Sciences, Graduate School of Science, Hokkaido University, North 10 West 8, Kita-ku, Sapporo 060-0810, Japan. 4 Laboratory of Animal Genetics, Department of Applied Molecular Biosciences, Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan. * Corresponding author, e-mail: [email protected] Received date : 14/10/09 Accepted date : 30/11/09 Kasetsart J. (Nat. Sci.) 44(3) 425 2003; Srikulnath et al., 2009). The putatively 46 other squamate reptiles, using RAG1 and C- unisexual diploidy has been reported also in L. mos as individual and combined data sets to boehmei (2n=2x=34), in which all individuals are speculate on the relationships of the genus female in Songkhla and Nakhon Si Thammarat Leiolepis in Thailand and confirm its phylogenetic provinces, southern Thailand (Aranyavalai et al., position in the Squamata. 2004). However, their genetic relationship has not been revealed. Even though their morphology MATERIALS AND METHODS (body color, pattern and shape) were previously explored as the key to species identification Specimen collection (Aranyavalai, 2003), this morphology might be One adult female each of L. reevesii misleading for constructing the correct phylogeny rubritaeniata and L. boehmei was collected from due to homoplasy. A consideration of molecular Nakhon Ratchasima and Songkhla provinces, information and nuclear and mitochondrial gene respectively. An adult male of L. belliana belliana sequences is appropriate for studies on both was captured in Chon Buri province. All shallow and deep genetic relationships. However, experimental procedures conducted on the animals the mitochondrial DNA sequences commonly conformed to the guidelines established by the demonstrate saturation at basal nodes and deeper Animal Care Committee, Hokkaido University. nodes. Probably, only the nuclear data can be best Although L. belliana ocellata used to be found in assessed to rebuild nodes at the deepest level of Thailand, it was not available for this study. the squamate tree (Townsend et al., 2004). C-mos (cellular moloney murine DNA extraction sarcoma), a candidate nuclear gene, is a proto- Whole genomic DNA, used as a template oncogene encoding a serine/threonine kinase for PCR, was extracted from the livers of all expressed at high levels in germ cells, in which individuals, following a standard phenol- the protein regulates cell maturation and tubulin chloroform-isoamylalcohol protocol (Sambrook formation (Yew et al., 1993). RAG1 (recombina- and Russell, 2001), with slight modification. tion activating gene-1) is a nuclear gene encoding Briefly, after homogenization, the tissue was component of the recombinase enzyme, which is digested at 37°C overnight using 25 µg/µL involved in the V(D)J recombination of the T- proteinase K in 0.5% (w/v) SDS in STE buffer receptor and immunoglobulin genes (Schatz et al., (0.1 M NaCl, 50 mM Tris and 1 mM EDTA, pH 1989). Both C-mos and RAG1 genes are single- 8.0). Then, the mixture was extracted with phenol- copy, without introns. Besides a few insertions and chloroform-isoamylalcohol (25:24:1) and the deletions, there are no repetitive sequences that DNA was precipitated with 0.05 volume of 0.2 M can cause any complication of sequence alignment NaCl and 2.5 volume of 100% ethanol. After among species. They have also been found in the washing in 70% ethanol, the genomic DNA was genome of vertebrates. These attributes make them air-dried and resuspended in TE buffer. particularly useful for reconstructing deep phylogenetic relationships within a number of PCR amplification vertebrate groups, especially in the Squamata PCR primers and conditions for the (Saint et al., 1998; Townsend et al., 2004; Vidal RAG1 gene and C -mos gene were taken from San and Hedges, 2004). In this study, phylogenetic Mauro et al. (2004) and Godinho et al. (2006), trees were documented for L. reevesii respectively. The standard PCR reaction was rubritaeniata, L. belliana belliana, L. boehmei, and performed using 1× ThermalPoll reaction buffer 426 Kasetsart J. (Nat. Sci.) 44(3) 2 containing 1.5 mM MgCl2, 0.2 mM dNTPs, 5 pM (χ ) test of base heterogeneity was calculated for specific primers and 0.25 U of NEB Taq individual and all codon positions, as implemented polymerase (New England Biolabs, Ipswich, in PAUP* v. 4.0b10. Nucleotide saturation was England) and 25 ng genomic DNA in a final analyzed for individual and all codon positions in reaction volume of 20 µL. each nucleotide data set by plotting the total number of transitions (Ts) + transversions (Tv) Cloning and DNA sequencing against genetic distance values, which were based Amplified products were examined by on alternative models implemented with Modeltest electrophoresis on 1% agarose gel; the DNA version 3.7 (Posada and Crandall, 1998), using the fragments were subsequently extracted from the program MEGA4 (Kumar et al., 2004) and PAUP* ethidium bromide-stained gel and were then v. 4.0b10. The level of incongruence between the ligated to pGEM-T Easy Vector System I two genes was examined using PAUP*. This (Promega, Madison, WI, USA). The ligated approach used the incongruence length difference plasmids were transformed into Escherichia coli (ILD) test with parsimony criterion (Farris et al., DH5· competent cells. Nucleotide sequences of 1995); 100 randomizations were performed. the DNA fragments were determined by 1stBase DNA sequencing service (Malaysia). Then, a Phylogenetic analysis nucleotide sequence comparison against the The phylogenetic trees were National Center for Biotechnology Information reconstructed by four different methods: maximum (NCBI) database was performed using the blastx likelihood (ML), maximum parsimony (MP), and the blastn program (http:// neighbor-joining (NJ) and Baysian inference (BI). blast.ncbi.nlm.nih.gov/Blast.cgi). All nucleotide The ML trees were generated with PHYML v.2.4.4 sequences were deposited in DDBJ (DNA data (Guindon and Gascuel, 2003) using non- bank of Japan, http://www.ddbj.nig.ac.jp/ parametric bootstrapping with 1,000 submission-e.html) and accession numbers are pseudoreplicates. The model and parameters shown in Table 1. indicated by Modeltest 3.7 were used, based on the Akaike Information Criterion (AIC) (Posada Sequence analysis and data set and Crandall, 1998). For BI, MrBayes v3.0b4 The RAG1 and C-mos gene nucleotide (Huelsenbeck and Ronquist, 2001) was used with sequences of L. reevesii rubritaeniata, L. belliana the same model and parameters as mentioned belliana and L. boehmei were aligned, using the above. The Markov Chain Monte Carlo (MCMC) default parameters of clustalX (Thompson et al., process was set to run four chains simultaneously 1997), to 46 other squamate reptiles, and 6 other for 1 million generations. After the log-likelihood reptilian and avian species as an outgroup taken value reached stationarity, sampling was carried from the NCBI database (Table1). Phylogenetic out at every 100th generation to get 10,000 trees analyses were conducted with three data sets to provide a majority-rule consensus tree with (RAG1, C-mos and the combined data set of the averaged branch lengths. All sample points prior two genes). All unalignable sites and gap-