Genetic Diversity and Differentiation of Dendrocalamus membranaceus (Poaceae: Bambusoideae), a Declining Bamboo Species in Yunnan, China, as Based on Inter-Simple Sequence Repeat (ISSR) Analysis

Dendrocalamus membranaceus Munro is a woody bamboo with a high economic and ecological value that often occurs as natural stands, such as in the large-scale forested areas of China’s Yunnan Province. Due to its overexploitation, the habitat of D. membranaceus in Yunnan has been dramatically reduced, and the quality of the stands has declined. As a preliminary analysis in considering the effective protection for these germplasm resources, we assessed the genetic diversity of 12 natural populations in Yunnan, using inter-simple sequence repeat (ISSR) markers. From 10 ISSR primers, we generated 155 bands, of which 153 were polymorphic (98.71%). Compared with other species in the genus, this species demonstrated a greater genetic diversity (S = 0.349) and lower genetic differentiation (GST = 0.252). Our analysis of molecular variance revealed that the genetic differentiation among the populations is significant. A large proportion of the genetic variation (78.95%) resides among the individuals within populations, whereas only 21.05% are found among populations. Mantel tests indicated no significant correlation between genetic and geographic distances among the populations. Given the low sexual reproducibility and characteristics of monocarpic plants, we recommend implementing in situ conservation measures for all of the D. membranaceus populations in Yunnan and collecting sufficient samples for ex situ conservation. Furthermore, the conservation area should be extended to its main natural habitats, the Lancang-Mekong River Valley.

polymorphism and displays acceptable reproducibility, it is widely used in current studies of population genetics [21][22][23]. In this study, we employed ISSRs to estimate the degree of genetic variation and differentiation among 12 populations of D. membranaceus in Yunnan Province, China (Table 1). Our objective was to use this information to help understand the genetic background of D. membranaceus and also to provide a reference for the incorporation of genetic resources when devising programs for species protection and breeding.

Genetic Diversity within Populations of Dendrocalamus membranaceus
Of the 79 ISSR primers used, 10 produced polymorphism; overall, the band reproducibility of these 10 primers is 93% (Table 2). Figure 1 presents the polymorphic fingerprinting pattern for the Mengyang (MY) population, as generated by ISSR primer UBC810. The 10 primers produced 155 reproducible bands in 240 clumps, for an average of 15.5 bands per primer. We considered each clump as a potential genet and the culms within as ramets of a clone, according to McClure [24]. Of the 155 bands, 153 were polymorphic (99%) at the species level. The average percentage of polymorphic bands (PPB) was 48.06%, ranging from 39% (Xingping, XP and Guanlei, GL) to 55% (Gengma, GM) at the population level (Table 3). Assuming a Hardy-Weinberg equilibrium, the expected diversity was estimated to be 0.164 at the population level and 0.219 at the species level. Shannon's indices of diversity were 0.249 (population) and 0.349 (species). Among our 12 populations, GM exhibited the highest degree of variability (expected heterozygosity, H E = 0.186; Shannon's diversity index, S = 0.279), whereas GL had the lowest (H E = 0.146, S = 0.216; Table 3).   Table 2). Lane C represents the blank control and lanes 1-20 represent template DNA for each individual from MY. Furthermore, a high level of genetic diversity was found for this species (PPB = 99%, H E = 0.219, S = 0.349). By comparison, the mean within-population gene diversity among monocotyledons is 0.144 for Nei's expected heterozygosity [25].
The high genetic diversity detected for D. membranaceus may be a consequence of its evolutionary history and the geological development of South Yunnan. D. membranaceus is most likely an ancient hexaploid [9,11], and polyploid taxa typically display a great genetic diversity, both in terms of heterozygosity and the number of bands amplified [13]. However, the habitat of the species has not suffered major geological disasters since the Tertiary period. The region is also warm and humid and provides the appropriate conditions for the reproduction of tropical bamboos [26]. Moreover, the natural populations were not, as a whole, disturbed by human activities until recently [5,6]. Therefore, we speculate that (1) the ancestors of D. membranaceus have accumulated abundant genetic diversity within an extensive and continuous distribution in Yunnan; (2) the existing population retains the genetic diversity from their ancestors; and (3) the habitat fragmentation caused by human activities does not perturb the genetic base of the existing D. membranaceus populations.

Genetic Structure and Differentiation Among Populations
The results from our analysis of molecular variance (AMOVA) showed that a large proportion of the genetic variation (78.95%) existed among individuals within the populations, and that 21.05% resided among the populations. The genetic differentiation among the populations was significant (p < 0.001) ( Table 4). Consistently, Nei's [27] estimator of population substructure also indicated a moderate level of population differentiation (coefficient of gene differentiation, G ST = 0.252). Furthermore, 24 private bands were found in eight populations (Table 3), implying that some genetic difference existed among the populations. Many factors can determine the genetic structure of plant populations, including the reproductive biology [28], and gene flow. Many RAPD-and sequence tagged microsatellite sites (STMS)-based analyses showed that long-lived, out-crossing taxa retained most of their genetic variability within populations [29]. The woody bamboos have a long vegetative phase of 20-150 years [1][2][3], and are typical long-lived species of the grass family. As one of those critical influences, the out-crossing of a plant species tends to explain 10 to 20% of the genetic variation among populations, whereas the selfing of a species leads, on average, to 50% variation between populations [25]. D. membranaceus can reproduce via seed in the wild, although this phenomenon is rare, and the rate of seed set is low [3,16]. We found a few young seedlings in the Jinghong (JH) population ( Figure 2). Furthermore, studies on the floral biology have indicated that D. membranaceus is likely anemophilous and prone to be an out-crosser [16], which also was supported by the genetic differentiation (G ST = 0.252) that was similar to the average of out-crossing species (G ST = 0.22) [29]. Therefore, extensive longevity and predominant out-crossing may be two important factors for the genetic structure of the existing D. membranaceus in Yunnan.

Genetic Distances and UPGMA Analysis
Our UPGMA tree (Figure 3), based on the values for the genetic distance D, revealed that the 12 populations could be separated into two clusters: XP plus Mengman (MM) and the other 10. We found no agreement with the geographic distance, which was supported by our Mantel test results, demonstrating that D was not significantly correlated with the geographic distance (r = 0.197, p = 0.184) (Table 5, Figure 4). This outcome could have reflected the possibility that the examined populations of D. membranaceus still represented the abundant genetic variation of their ancestors. In fact, this tendency is reinforced by the naturally long vegetative phase and low seed set over the life cycle, which are typical features of woody bamboos [11,30,31]. Another possibility may be that, as an economic importance species, D. membranaceus most likely has undergone human-mediated movement of genotypes among partial or complete distribution areas, especially within those populations comprising small areas, such as XP, GM and Yingjiang (YJ).   Figure 4. Correlation between geographic distance and genetic distance for 12 populations of Dendrocalamus membranaceus.

Implications for the Development of D. membranaceus
The natural resources of D. membranaceus in Yunnan Province have been overexploited for many years, leading to the deterioration of its habitat [6,8]. The same situation has also occurred in other regions of the Lancang-Mekong River Valley, such as Myanmar and Northern Vietnam [7]. Although bamboo is of important significance for the local ecological and agricultural systems, the increasing demand for its culms and shoots and the dramatic declines in its native habitats and stand quality indicate that further research that focuses on its germplasm collection and conservation is urgently needed [3,6].
This new information about the diversity in its population genetics demonstrates the degree of success these plants have had in adapting to their altered environments over time. Therefore, it is of critical importance that this knowledge be applied in assessing their conservation value and the status of particular populations when developing management plans [32,33]. Despite dramatic reductions in natural habitat, based on our results, the high genetic diversity is still maintained within populations of D. membranaceus in Yunnan, China. However, the genetic differentiation among the populations and the limited reproductive capability of D. membranaceus in the wild indicate that it is still necessary for us to protect all of the existing natural populations and their habitat in Yunnan, China. At the same time, germplasm collection and conservation in the Lancang-Mekong River Valley, the distributional center of this species, should be placed at the first and foremost position. A total of 12 natural populations were sampled from almost all of the main growing regions of Dendrocalamus membranaceus in Yunnan Province, China (Table 1, Figure 5). Leaves were collected from 240 clumps (20 per population) that were at least 100 m apart at each site. Using the method of McClure [24], we regarded each clump as a potential genet and the culms within as ramets of a clone, and we assumed that an assortment of genetic material existed within each population. All of the samples were dried with silica gel and stored at 4 °C prior to DNA extraction. Vouchers for each population were deposited at the Herbarium of the Southwest Forestry University (SWFC), Kunming, Yunnan, China.

Total DNA Extraction
Genomic DNA was extracted according to the CTAB protocol [34]. The quality of DNA was determined using 1.0% agarose gels. The purified genomic DNA was quantified using a BioRad Smartspec3000 UV-Vis spectrophotometer.

ISSR PCR Amplification
Genomic DNA was PCR-amplified using ISSR primers obtained from the Biotechnology Laboratory, University of British Columbia (Vancouver, Canada). Amplifications were performed using an ABI 2720 Thermal Cycler (Applied Biosystems, USA). The 20 μL mixture contained 10 ng of template DNA, 2.0 μL of 10 × PCR buffer, 1.8 mM MgCl 2 , 0.1 mM dNTPs (TAKARA, Dalian, China), 2% formamide, 100 nM of each primer, 1.5 units of Taq polymerase (Fermentas), and double-distilled water. The cycling conditions included an initial denaturation at 95 °C for 3 min; then 40 cycles of 95 °C for 30 s, 50 °C for 30 s, and 72 °C for 2 min; and a final extension at 72 °C for 10 min. The PCR products were electrophoretically separated using 2.0% agarose gels buffered with 0.5× TBE. A 100-bp DNA ladder (Fermentas) was used as a size marker. The PCR reactions were repeated twice to ensure reproducibility. After staining with ethidium bromide (1 μg·mL −1 ), the DNA fragments were identified by software Lab Works Software 3.0 [35] for gel documentation.

Data Analysis
Only distinct, reproducible, well-resolved fragments were scored as present (1) or absent (0) for each marker, and were displayed as part of a binary matrix. The data matrices were analyzed using POPGENE version 1.31 [36]. The following genetic parameters were determined: The percentage of polymorphic bands (PPB), expected heterozygosity (H E ), coefficient of gene differentiation (G ST ), genetic distance (D), and Shannon diversity index (S). An analysis of Molecular Variance (AMOVA program Version 1.55) [37] was performed to describe the genetic structure and variation among the populations and among the individuals. The number of private bands was calculated using GenAlEx 6.1 [38].
To visualize the genetic relationships among the populations, we constructed a dendrogram based on Nei's [39] genetic distance D, implementing an unweighted pair-group method of cluster analysis that used arithmetic averages (UPGMA) and the software program NTSYS-PC V2.1 [40]. To test for putative correlations between the genetic distances and geographic distances among the populations, we performed Mantel tests using TFPGA software [41], computing 1000 permutations.

Conclusions
In summary, high genetic diversity and lower genetic differentiation were detected among the Dendrocalamus membranaceus populations in Yunnan. A large proportion of the genetic variation (78.95%) resides among individuals within the populations, whereas only 21.05% exist among the populations. No significant correlation was found between the genetic and geographic distances among the populations. Considering that D. membranaceus is a monocarpic and has prominently low sexual reproduction, we propose to implement in situ conservation measures for all of the populations and collect sufficient samples for ex situ conservation. Our results are also of important significance for the resource conservation of D. membranaceus in its main natural habitat, the Lancang-Mekong River Valley.