Genetic structure and differentiation in Dendrocalamus sinicus (Poaceae: Bambusoideae) populations provide insight into evolutionary history and speciation of woody bamboos

Abstract

Evolutionary processes, speciation in woody bamboos are presently little understood. Here we used Dendrocalamus sinicus Chia & J.L. Sun as a model species to investigate dispersal or vicariance speciation in woody bamboos. Variation in three chloroplast DNA (cpDNA) fragments and eight simple sequence repeat markers (SSR) among 232 individuals sampled from 18 populations across the known geographic range of D. sinicus was surveyed. D. sinicus populations exhibited a high level of genetic differentiation which divided them into two groups that are consistent with different culm types. Eleven haplotypes and two lineages (Straight-culm and Sinuous-culm lineages) were identified from phylogenetic analyses, and a strong phylogeographic structure across the distribution range was found. The demographic and spatial expansion times of the Straight-culm lineage were calculated as 11.3 Kya and 20.8 Kya, respectively. The populations of D. sinicus had experienced dispersal and long-term isolation, although this trace was diluted by contemporary gene flow revealed by SSR data. Our results provide an phylogeographic insight to better understand the speciation processes of woody bamboos.

Figure 1

Haplotype distribution in 18 Dendrocalamus sinicus populations based on combined sequences of three loci. Colored circles represent the population locations. Maps were drawn using the software ArcGIS version 10.2 (http://desktop.arcgis.com) and modified using Photoshop (Adobe Corporation, California, America).

Figure 2

(a) Strict consensus tree obtained by analysis of 11 cpDNA haplotypes of Dendrocalamus sinicus, with D. latiflorus used as outgroup. The numbers on the left of the dash sign (/) indicate posterior probabilities from the Bayesian analysis; the numbers on the right indicate bootstrap values (>50%). (b) The 95% confidence network of 11 cpDNA haplotypes. The size of circles corresponds to the frequency of each haplotype. Each solid line represents one mutational step that interconnects two haplotypes for which parsimony is supported at the 95% level. The small open circles indicate hypothetical missing haplotypes.

Figure 3

Distribution of the number of pairwise nucleotide differences for cpDNA sequence data in: (a) all populations (b) the TZ lineage (c) the WQ lineage of Dendrocalamus sinicus. The green solid line shows observed values, whereas the red dashed line represents expected values under a model of sudden (stepwise) population expansion.

Figure 4

Bayesian clustering results of the STRUCTURE analysis for nSSR data of 18 populations of Dendrocalamus sinicus (232 individuals). Number of clusters (K) was from one to 10 in 20 independent runs. (a) STRUCTURE-estimated genetic clustering (K = 2, 3, 4, 5, 6), each vertical bar represents an individual and its estimated proportion in K clusters. Black lines separate different populations. Population codes are identified in Table 1. (b) Dot plot representing the mean estimate of Ln posterior probability of data for K = 1–10 (20 replicates), the standard deviation of each mean L(K) value is given as a black vertical line; the pink line represents distribution of ΔK.

Figure 5

Neighbor-joining tree representing the genetic relationships among 18 populations of Dendrocalamus sinicus, based on Nei’s (1987) unbiased genetic distance (D) calculated from SSR data. Population codes are identified in Table 1, and branches are color-coded as in Figs 4 and 5. Numbers by nodes are bootstrap values (>50%) from 1000 replicates.