Leaf materials of B. gymnorrhiza were sampled from different individuals at the following ten locations (Fig. 1); V (Hanoi, Vietnam) located in the east of the Asian Continent, MK (Kuching, Malaysia) located in the northwest of the Borneo Island in the Southeast Asia, ML (Langkawi Island, Malaysia) located in the northwest side of the Malay Peninsula, BMJ (Bali Island, Indonesia) located in the east of the Java Island in the Southeast Asia, AC (Cairns, Australia) located in the northeast coast of Australia, AB (Ballina, Australia) located in the east coast of Australia, K (Kakinada, India) located in the northeast coast of India, G (Goa, India) located in the southwest of India, SL (Tolanaro, Madagascar) located in the southeast coast of Madagascar, and MO (Morondava, Madagascar) located in the west coast of Madagascar. We call each sampling location as a population in the following sections. Leaf materials of a closely related species, B. sexangula, were also sampled at two locations on the western coast of the Malay Peninsula. The details of our samples, including the locations and numbers of individuals, are summarized in Table 1.
Table 1. Geographical location and sample size in Bruguiera gymnorrhiza and Bruguiera sexangula
|Sampling location||Abbreviation||Geographical coordinates||Number of individuals|
| B. gymnorrhiza |
|Ballina||AB||28°52′33.26″S, 153°32′47.05″E||15 (55)|
|Cairns||AC||16°50′21.47″S, 145°43′25.45″E||15 (32)|
|Bali||BMJ||8°05′45.84″S, 114°28′33.15″E||10 (10)|
|Hanoi||V||21°00′05.62″N, 107°16′35.82″E||15 (45)|
|Kuching||MK||1°38′31.22″N, 110°14′57.25″E||15 (45)|
|Langkawi||ML||6°26′02.61″N, 99°50′23.41″E||11 (11)|
|Goa||G||15°22′23.85″N, 73°57′42.07″E||15 (23)|
|Kakinada||K||16°56′48.14″N, 82°15′23.12″E||15 (20)|
|Morondava||MO||20°16′47.58″S, 44°17′19.81″E||14 (41)|
|Tolanaro||SL||25°03′43.35″S, 46°53′42.07″E||15 (29)|
| B. sexangula |
|Langkawi|| ||6°26′02.61″N, 99°50′23.41″E||7 (8)|
|Ranong|| ||9°84′43.92″N, 98°55′40.58″E||2 (2)|
Figure 1. Sampling locations of the 10 populations. Populations are indicated by abbreviations. Number of individuals used in this study is shown in parentheses.
Download figure to PowerPoint
DNA extraction, PCR, and sequencing
Fifteen individuals were randomly selected from each population. When the number of individuals in a population was smaller than 15, we analyzed all individuals. Genomic DNA was extracted from a piece of a single leaf using modified cetyltrimethylammonium bromide (CTAB) method (Murray and Tompson 1980). In total, genomic DNAs of 140 individuals of B. gymnorrhiza were obtained.
Based on the published sequence information of the expressed sequence tags (ESTs) of B. gymnorrhiza (Miyama et al. 2006), PCR primer sets of eight nuclear genes were designed. Based on the sequence information of the PCR products, we conducted BLAST search against NCBI database (nucleotide collection). According to the blast search results, we tentatively named the eight genes as follows; NAC (BP941102), VVHP (BP944085), PO (BP939063), GM (BP939421), SF (BP939928), EXP2 (BP942725), EPCRF (BP946081), and UNK (BP945250), where the corresponding B. gymnorrhiza ESTs (Miyama et al. 2006) are shown in parentheses. In addition to the aforementioned eight nuclear genes, two nuclear genes, PAL1 and mang-1 (Inomata et al. 2009), and one chloroplast DNA (cpDNA) region, trnS-trnG spacer (Hamilton 1999; Sun and Lo 2011), were also analyzed.
PCR amplification was carried out under the following protocols: an initial denaturation at 95°C for 3 min followed by 30 cycles of denaturation at 95°C for 30 sec, annealing at 50°C, 55°C, or 60°C for 30 sec, polymerization at 72°C for 1.5 min, and a final extension at 72°C for 7 min. The details of the sequence and annealing temperature of PCR primers are listed in Table S1. After the amplification, PCR products were purified. Then we directly determined their sequences using a DNA sequencing kit (BigDye terminator v. 3.1/1.1 cycle sequencing kit, ABI) with the PCR primers and ABI Prism 3100 or 3730 automatic sequencers (Applied Biosystems, Tokyo, Japan).
Two haplotype sequences were determined for each individual. After the direct sequencing when we found no or only one heterozygous site in a sequence of a single individual, we inferred sequences of two haplotypes for each individual. On the other hand, when two or more heterozygous sites or indels were found in a sequence of a single individual, purified PCR product was cloned into the pGEM T-easy vector (Promega KK, Tokyo, Japan). Then, we sequenced each clone using T7 and Reverse primers designed at the promoter sites of the vector (Table S1). Four to sixteen clones were sequenced for each individual. Sequences of two haplotypes of a heterozygous individual were determined when we obtained at least two exactly identical sequences for each haplotype.
Sequences of each gene were aligned using CLUSTAL X 2.1 (Larkin et al. 2007) and further edited by hand. Because, we could not amplify EXP2 gene of one individual sampled from the AB population from the east coast of Australia, we treated it as missing data in the following analysis. A continuous alignment gap was treated as a single indel. All sequences obtained in this study were deposited in the DNA Data Bank of Japan (DDBJ). Accession numbers are AB813916-AB817042.
Molecular population genetic analysis of B. gymnorrhiza was conducted using sequences of B. sexangula as an outgroup. Using inter-simple sequence repeat, ribosomal internal transcribed spacer, and chloroplast DNA markers, Sun and Lo (2011) showed that B. gymnorrhiza and B. sexangula are genetically close relatives. Molecular population genetic parameters, π (the average number of nucleotide differences per site: nucleotide diversity; Nei 1987), Ks (number of nucleotide substitutions per silent site between species), and Ka (number of nucleotide substitutions per replacement site between species), were estimated using the DnaSP ver. 5.10.00 (Librado and Rozas 2009). To assess the deviation from the standard neutral model, Tajima's test, Fu and Li's test, Fay and Fu's test, and MK test were performed using the DnaSP version 5.10.00, and the multilocus HKA test was also performed using HKA program downloaded from http://genfaculty.rutgers.edu/hey/software#HKA.
The extent of genetic differentiation among populations was estimated by analysis of molecular variance (AMOVA) approach (e.g., Weir and Cockerham 1984). The Arlequin software version 18.104.22.168 (Excoffier and Lischer 2010) was used to estimate F-statistics. Three hierarchical levels, (i) among populations, (ii) among individuals/within populations, and (iii) within individuals, were examined for nuclear genes by the locus-by-locus analysis, and two levels, (i) among populations and (ii) within populations, were examined for cpDNA. Each haplotype in a gene was regarded as an allele. Overall values of F-statistics for all the ten nuclear genes were obtained by summing variance components over the genes. In the AMOVA analyses, genotypic data of nuclear genes, where gametic phase of genotype is unknown, and haplotypic data of the cpDNA region were used. A haplotype network of the cpDNA region was constructed using the TCS ver. 1.21 (Clement et al. 2000).
Genetic clusters of populations were inferred using the Bayesian clustering approach, STRUCTURE ver. 2.3.3 (Pritchard et al. 2000; Falush et al. 2003; Hubisz et al. 2009). Under the admixture model, five independent runs were performed for each number of clusters (K: K = 1–10) to confirm the convergence of Markov chain Monte Carlo (MCMC) chains. In each run, 500,000 MCMC iterations were performed after a burn-in period of 1,000,000 iterations. To estimate the number of clusters, we applied the method proposed by Evanno et al. (2005); K is identified when delta K, which is the rate of change in the log probability of data between continuous K values, is maximal.
Population history of genetic clusters identified using the STRUCTURE and evolutionary history of the two species, B. gymnorrhiza and B. sexangula, were inferred. Population parameters were estimated under the isolation-with-migration (IM) model using the IMa2 program (Hey and Nielsen 2007; Hey 2010). In this analysis non-recombinant regions were used. Non-recombinant regions were estimated using the four-gametic test (Hudson and Kaplan 1985) implemented in DnaSp ver. 5.10.00. Population splitting (or speciation) times (T = tu), population sizes (θ = 4Nu), and migration rates (2Nm) between genetic clusters (or species) were estimated, where u is the mutation rate per locus per generation, t is the population splitting (or speciation) time, N is the effective population size, and m is the migration rate. The infinite site model (Kimura 1969) was used except VVHP, EXP2, and mang-1. For the three regions, we applied HKY model (Hasegawa et al. 1985) because there were more than two alleles at some variable sites in these three genes. The upper bounds of the parameters are set to include most part of the posterior distribution of preliminary simulations. We confirmed that the burn-in period was long enough so that the MCMC simulation reached the stationary state. Each simulation run was performed with 40 chains with adjusted heating terms according to the IMa2 manual. To check the convergence of the MCMC simulation to the true stationary state, multiple independent runs were performed. The posterior distributions of parameters were estimated using 100,000 genealogies sampled from multiple independent runs. Peaks of the posterior distribution were defined as estimates of parameters.