Molecular identification of natural hybridization between Melastoma malabathricum and Melastoama beccarianum in Sarawak, Malaysia

Abstract Hybridization is very common in flowering plants and it plays a significant role in plant evolution and adaptation. Melastoma L. (Melastomataceae) comprises about 80–90 species in tropical Asia and Oceania, among which 41 species occur in Borneo. Natural hybridization is frequently reported in Melastoma in China, but so far there have been no confirmed cases of hybridization in Southeast Asia (including Borneo), where most species occur. Here, we identified a case of natural hybridization between Melastoma malabathricum L. and Melastoma beccarianum Cogn. in Sarawak, Malaysia, by using sequence data of three nuclear genes and one chloroplast intergenic spacer. Melastoma malabathricum is the most widespread species of this genus, occurring in almost the whole range of this genus, while M. beccarianum is a local species endemic to northern Borneo. Our results showed that natural hybridization and introgression occur between M. malabathricum and M. beccarianum, and the introgression was asymmetrical, mainly from M. malabathricum to M. beccarianum. As adaptive traits can be transferred by introgression, our study suggests that natural hybridization should be a significant mechanism for the evolution and adaptation of Melastoma in Southeast Asia. However, introgression from the common species M. malabathricum to the relatively rare species M. beccarianum may cause the decline of M. beccarianum, incurring conservation concern. With a large number of species of Melastoma and almost year‐around flowering in Southeast Asia, more cases of natural hybridization are expected to be found and identified in near future.


| INTRODUC TI ON
Hybridization is referred to a process through which there is interbreeding between species of two genetically distinct populations or species (Harrison & Larson, 2014). Hybridization is very common in flowering plants and it plays a significant role in plant evolution and adaptation (Abbott et al., 2013;Arnold, 1997). The evolutionary outcomes of hybridization include hybrid speciation, reinforcement of prezygotic isolation, transfer of genetic materials between species (introgression), and so on (Abbott et al., 2013;Arnold, 1997).
Introgression can have both positive and negative effects on genetic diversity and adaptation. On one hand, introgression from other species can contribute to the increase of genetic diversity and rapid adaptation to novel environments by transfer of adaptive alleles (Abbott, Barton, & Good, 2016;Lamichhaney et al., 2016;Zhang, Dasmahapatra, Mallet, Moreira, & Kronforst, 2016), on the other hand, introgressive hybridization between a common species and a rare species can cause genetic assimilation of the rare species by the common species and thus incurs the risk of extinction of the rare species (Huxel, 1999;Rhymer & Simberloff, 1996;Todesco et al., 2016). Therefore, identification of the extent of hybridization and determination of the direction of introgression are critical to understand the evolutionary roles of hybridization in plants.
Melastoma has undergone rapid species radiation in the past 1 million years (Renner & Meyer, 2001). Hybridization is frequently reported amongst species pairs in this genus (Dai et al., 2012;Zhou et al., 2017;Zou et al., 2017). However, all these reported cases of hybridization in Melastoma were confined to species in China. As the species distribution center of Melastoma, Southeast Asia possesses about 70-80 species (Wong, 2016). So far, there have been no confirmed cases of hybridization in Melastoma in Southeast Asia, although several putative hybrids have been proposed based on morphology (Wong, 2015(Wong, ,2016. Borneo is the third biggest island in the world and the most in-depth research in Melastoma has been conducted here (Meyer, 2001;Wong, 2016 coexist and the putative hybrid shows morphological intermediacy between the two species ( Figure 1). Hypanthium indumentum is the most important trait to distinguish species of Melastoma (Meyer, 2001;Wong, 2016) and the two species indeed differ markedly in this aspect. Melastoama beccarianum and M. malabathricum are covered by penicillate emergences with long spreading bristles and short imbricate or appressed lanceolate scales on their hypanthiums, respectively, while the putative hybrid is covered by long branched appressed scales on its hypanthiums. Moreover, M. beccarianum and M. malabathricum have lanceolate and ellipse leaves, respectively, while the putative hybrid has lanceolate-ellipse leaves. As hybrid identification based on morphology alone is not always reliable, we aimed to test the hypothesis of natural hybridization between M. malabathricum and M. beccarianum by molecular means. Previous studies showed that sequences of low-copy nuclear genes were very useful for hybrid identification in Melastoma (Dai et al., 2012;Zhou et al., 2017;Zou et al., 2017), and we used this approach in this study.

| Plant materials
We sampled 19 individuals of M. malabathricum , 16 of M. beccarianum, and 16 of their putative hybrid in 2016 along a roadside near Lambir Hills National Park, Miri, Sarawak, Malaysia. In addition, 23 individuals of M. malabathricum were sampled from a natural population in University Malaysia Sabah, Sabah, Malaysia, and considered as a "pure" population because there is no other species occurring there. However, no "pure" populations of M. beccarianum were found, given that they are always sympatric with M. malabathricum or other congeneric species. We aimed to use this "pure" population to distinguish introgression from incomplete lineage sorting (ILS) of ancestral polymorphisms. For each individual, we collected one or two leaves and dried them with silica gel for subsequent DNA extraction.

| DNA extraction, PCR, and sequencing
We used CTAB method (Doyle & Doyle, 1987) to extract DNA from dried leaves. Six nuclear genes adopted in previous studies Dai et al., 2012;Huang et al., 2017;Zou et al., 2017) were tested in the three taxa and three nuclear genes (gbss, vr, and tpi) showed successful PCR amplification in all three taxa. Gbss, tpi, and vr encode granule-bound starch synthase, triose phosphate isomerase, and vacuolar invertase, respectively. We conducted PCR amplification with the EasyTaq DNA polymerase (Transgen Biotech, Beijing, China) or the KOD-FX DNA polymerase (TOYOBO, Osaka, Japan).
PCR was conducted with the following conditions for all three nuclear genes: 94°C (4 min); 30 cycles of 94°C (40 s), 55°C (1 min), and 72°C (1 min); and a final extension of 8 min at 72°C The PCR products were purified by electrophoresis through a 1.2% agarose gel followed by use of the Pearl Gel Extraction Kit (Pearl Bio-tech, Guangzhou, China).
For sequences that contained more than one polymorphic site, cloning sequencing was performed using the pMD-18T Vector Kit (Takara, Dalian, China). At least six positive clones were sequenced to phase the haplotypes of each sample. In addition, a chloroplast intergenic spacer (trnH-psbA) was also amplified and sequenced using the same protocol as the nuclear genes. All sequences in this study have been deposited in GenBank with the accession numbers MH910371-MH910491.

| Sequence analyses
SeqMan (DNASTAR, Madison, WI, USA) was used to assemble and edit the sequences. These sequences were further aligned using Clustal X (Thompson, Gibson, Plewniak, Jeanmougin, & Higgins, 1997). DNASP 5.0 (Librado & Rozas, 2009) was applied to phase the haplotypes and calculate the nucleotide diversity as well as Tajima's D for each population. Haplotype network for each gene was constructed by Network v.4.6 (www.fluxus-engineering.com) with the median-joining algorithm (Bandelt, Forster, & Röhl, 1999). Genomic admixture proportions of all individuals of the three taxa were assessed using the program Structure (Pritchard, Stephens, & Donnelly, 2000) with the default settings and employing the admixture model with correlated allele frequencies, as performed in Liao et al. (2015).

| Sequence analysis of the partial tpi gene in M. malabathricum , M. beccarianum, and their putative hybrid
The

| Sequence analysis of the partial gbss gene in M. malabathricum, M. beccarianum, and their putative hybrid
The aligned sequence of the partial gbss gene of the three taxa was

| Sequence analysis of the chloroplast intergenic spacer trnH-psbA in M. malabathricum, M. beccarianum, and their putative hybrid
The chloroplast intergenic spacer trnH-psbA of the three taxa was 285 bp in length, containing two nucleotide substitutions and one two haplotypes, respectively ( Figure 2d). As shown in the haplotype network, M. beccarianum, M. malabathricum shared three haplotypes, among which the two most common haplotypes were also shared with their putative hybrid.

| Nucleotide diversity and structure analysis
At each of the three nuclear genes, the putative hybrid harbored higher nucleotide diversity (π) than either M. malabathricum or M. beccarianum (Table 1)

| Molecular evidence for natural hybridization between M. malabathricum and M. beccarianum
In this study, we tested the hypothesis of natural hybridization be-

| Bidirectional introgression between M. malabathricum and M. beccarianum
We F I G U R E 2 Median-joining networks of (a) partial tpi gene, (b) partial vr gene, (c) partial gbss gene, and (d) chloroplast intergenic spacer trnH-psbA. Haplotypes of each taxon are denoted using the first letter of its species name ("Mm" and "Mb" refer to Melastoma malabathricum and Melastoma beccarianum from Lambir Hills National Park, respectively; while "PH" and "MmS" represent the putative hybrid and allopatric M. malabathricum from University Malaysia Sabah) followed by a number ordered by quantity each population owned. Small black circles represent hypothetical haplotypes. Mutational steps are shown by the number near the connecting lines, and the number is omitted for those with only one mutational step introgression happens between M. beccarianum and M. malabathricum (tpi) and introgression is asymmetrical, mainly from M. malabathricum to M. beccarianum (vr and gbss).

| Factors contributing to natural hybridization between M. beccarianum and M. malabathricum
Several factors may contribute to natural hybridization between M. beccarianum and M. malabathricum. First, species of Melastoma diverged in a relatively short evolutionary time and reproductive isolation between them is still incomplete. It is reported that Melastoma has undergone a rapid species radiation, with about 80-90 species formed in the past 1 million years (Renner & Meyer, 2001). This is also supported by our sequence data, for example, M. beccarianum   year, and pollinators like bumble bees are largely shared among species of Melastoma (Gross, 1993;Loh, 2008;Luo, Zhang, & Renner, 2008), offering ample opportunities for hybridization.

| Conservation implications for rare species in Melastoma
While M. malabathricum is the most widespread species of Melastoma, M. beccarianum is endemic to northern Borneo, and found across west Sabah, Brunei, and north Sarawak (Wong, 2016). Hybridization is a double-edge sword: it may drive rare taxa to extinction through genetic and demographic swamping (Huxel, 1999;Rhymer & Simberloff, 1996;Todesco et al., 2016); conversely, a net fitness can be gained to one or both taxa without loss of species integrity by adaptive trait transfer between species by introgression (Abbott et al., 2016;Lamichhaney et al., 2016;Zhang et al., 2016). Mb2 Mb8 (

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
RW, PZ, and RZ designed the study. GT, ZH, YW, ZN, SH, and YL collected materials, which was then performed experiments by RW and PZ. PZ, GT, ZH, YW, and ZN guided the experiments. RW and PZ analyzed and interpreted the data and wrote the manuscript with guidance of RZ, WW, YL, and SH. All authors read and approved the final manuscript.

DATA ACCE SS I B I LIT Y
The haplotype sequences of our study involved are deposited in GenBank with accession numbers MH910371-MH910491.