Analysis of Genetic Diversity and SSR Allelic Variation in Rubber Tree (Hevea brasilensis)

Rubber tree, Hevea brasiliensis, belongs to the family of Euphorbiaceae, originated from Amazon Trends. The family has ten varieties (H. brasiliensis, H. nitida, H. pauciflora, H. spruceana, H. benthamiana, H. camporum, H. microphylla, H. rigidifolia, H. guianensis, H. comargcana) and four variation varieties (H. guianensis var. luter, H. guianensis var. marginata, H. parciflora var coriacea, H. nitida Mart var. toxicadendroides). H. brasiliensis is the most economically important member of the genus Hevea, because its economic importance and its sap-like extract (as latex) can be collected and is the primary source of natural rubber. There are many rubber tree germplasm resources in Brazil, Malaysia, Indonesia, India, French, and China.


Introduction
Rubber tree, Hevea brasiliensis, belongs to the family of Euphorbiaceae, originated from Amazon Trends. The family has ten varieties (H. brasiliensis, H. nitida, H. pauciflora, H. spruceana, H. benthamiana, H. camporum, H. microphylla, H. rigidifolia, H. guianensis, H. comargcana) and four variation varieties (H. guianensis var. luter, H. guianensis var. marginata, H. parciflora var coriacea, H. nitida Mart var. toxicadendroides). H. brasiliensis is the most economically important member of the genus Hevea, because its economic importance and its sap-like extract (as latex) can be collected and is the primary source of natural rubber. There are many rubber tree germplasm resources in Brazil, Malaysia, Indonesia, India, French, and China.
Several researchers have investigated the genetic diversity of rubber tree by using molecular markers (Lekawipat et al., 2003;Saha et al., 2005;Lam et al., 2009;Gouvêa et al., 2010;Oktavia et al., 2011), but there was no report about the polymorphism of lower repeats SSR markers and the loci and the flanking area variation of SSR markers in rubber tree. SSR marker used to detect the alleles by PAGE gel after PCR has slight restrictions in distinguishing the fragments as length or size homoplasy (Estoup et al., 1995;Grimaldi & Crouau-Roy, 1997;Angers & Bernatchez, 1997). However, sequencing of repeats and flanking regions can help detect the difference of the alleles exactly (Xie et al., 2006;. In this study, 16 primer pairs from genome and EST-SSR amplified across the popular cultivars cultivated in China, wild accessions and interspecies. There were three main objectives: (1) detect the genetic diversity and relationships between cultivars and wild accessions in rubber tree, (2) to investigate the polymorphism of low repeat SSR marker, and (3) analyze the loci variations in rubber tree.

Plant materials and SSR markers
Forty-five cultivars which are the main cultivars in China, 11 wild accessions from Brasil and 3 related species were used in this study (Table1). Fresh leaves were collected in bronze period from Rubber Research Institute, Chinese Academy of Tropic Agricultural Science (Danzhou) and stored at -20 after washing by pure water. Leaf genomic DNA was extracted following the CTAB protocol (Venkatachalam et al., 2002).

PCR amplification and detection of fragments
All primers were amplified in the TaKaRa PCR Thermal Cycler Dice, each PCR reaction consisted of: 2 µl of 10×buffer, 0.25 µl of 10 M dNTPs, 1 µl each of forward and reverse primer (20 µmol), 2µl of template leaf genomic DNA (20 ng/µl), 0.15 µl of Taq polymerase (5 U/µl) (TAKARA Biotechnology (Dalian) Co. Ltd), ddH 2 O added to a total reaction volume of 20 µl. The PCR reaction profile was pre-denatured at 94 for 2 min followed by 30 cycles of 94 for 30 sec, annealing temperature for 45 sec and 72 for 1 min and finally, 72 for an extension of 5 min. To ensure precision and reproducibility of fragments, DNA samples were amplified and analyzed at least twice from each individual sample.

Cloning and sequencing of SSR alleles
Five EST-SSR markers (HBE008, HBE063, HBE164, HBE187, and HBE199) were selected to investigate SSR loci variation. Of the five primer pairs, the repeat motif of HBE008 and HBE187 was (CT) n, HBE164 and HBE199 was (AG) n. The selected alleles were amplified, recovered, purified, cloned and sequenced.
Alleles from these SSR loci were cut from the dried PAGE gels and used as templates for a new round of PCR amplifications. Each of these alleles was directly cloned into the pGEM-T Easy Vector (Promega, USA) according to the manufacturer's instructions, and transformed into Escherichia coli DH5 α cell. The positive clones were sequenced using the ABI PRISM 3730 sequencer. To obtain reliable sequences, at least three clones per allele were sequenced. The nucleotide sequences were aligned using Clustal X (http://www.ftp-igbmc.ustrasbg.fr/pub/ClustalX/) to compare the amplified SSR alleles with the SSR-containing ESTs to investigate the loci variation.

Data analysis
Sixty-two genomic DNA of cultivars and wild accessions were amplified with 16 primer pairs, and visualized on PAGE gel with silver stain. These bands were recorded as "1" for presence, "0" for absence, "999" for missing data. Number of alleles, observed heterozygosity (Ho) and power of discrimination (PD) were calculated for each locus. Ho was calculated as the number of genotypes which were heterozygous at a given locus divided by the total number of genotypes surveyed at that locus. PD was calculated as 1-∑G 2 ij (Kloosterman et al., 1993), where Gij is the frequency of the jth genotype for the ith locus summed across all alleles at that locus. Genetic similarity (GS) between any two pairs of the 56 cultivars and wild accessions was calculated from the alleles across the 16 SSR loci using the Jaccard similarity coefficients (Sneath & Sokal, 1973). A dendrogram was constructed with the un-weighted pair group method with arithmetic averages (UPGMA) on the basis of the similarity coefficients. All these analyses were performed with NTSYS-pc 2.10 software package.

Polymorphic analysis of EST-and gSSR markers
Sixteen SSR primer pairs, of which ten from EST-SSRs and six from genome, could successfully amplify, expected products across 45 cultivars, 11 wild accessions and 3 related species (Table   www.intechopen.com 2). A total of 43 alleles were obtained from 10 EST-SSR primer pairs across cultivars and wild accessions, with an average of 4.3 alleles and H o = 0.488. Ten alleles were amplified by HBE280, followed by HBE199 with 5 alleles, and HBE164 and HBE316 with 2 alleles each, respectively. Six alleles were amplified by HBE008 and HBE280 in wild accessions, respectively, followed by HBE199 with 5 alleles and HBE316 with 2 alleles. A total of 30 alleles were obtained from 6 gSSR primer pairs, with an average of 5 alleles and H o =0.743. Six alleles were amplified by M197 and MnSOD, respectively in cultivars, 6 alleles by MnSOD as well as in wild accessions. And the other primer pairs were amplified 4 or 5 alleles in cultivars and wild accessions, respectively. HBE280, M197 and MnSOD were the most informative.   (Table 2). For the wild accessions, the ten EST-SSR markers produced a total of 38 alleles with an average of 3.8 alleles, H o = 0.455 and PD = 0.616 per locus; and the six gSSRs produced a total of 28 alleles with an average of 4.7 alleles, H o = 0.758 and PD = 0.685 per locus (Table 2). HBE280 was the most informative among the EST-SSRs, and MnSOD was the most informative among the genomic SSR markers.

www.intechopen.com
Of the 16 primer pairs, only PD value of HBE280 was larger than 0.8 (0.834). PD values of the ten EST-SSRs ranged from 0.420 to 0.834 and the average was 0.597; and the PD values of the six genomic-SSRs ranged from 0.587 to 0.772 and the average was 0.689.

Analysis of genetic diversity
The Jaccard similarity coefficient for the 16 SSR markers was used to analysis the genetic similarities (

Detection of loci variation
Substantial sequence variation was found in the 61 sequences of 5 EST-SSR loci. The insertion, deletion, transition and transversion can be found in these sequences (Fig. 2).
The original EST sequences (the RefEST) ID of HBE164 is EC603146.1, and the SSR repeat loci is (AG) 6 . In comparison with EC603146.1, no variation was observed from all repeat regions, and transversions were observed from the flanking regions of all sequences and insertions were found in the flanking sequences of Dafeng117 and AC/T/15/114. The original EST sequences (the RefEST) ID of HBE187 is EC601817.1, and the SSR repeat loci is (CT) 6 . In comparison with EC601817.1, no variation was observed from all repeat regions, but deletions were observed from the flanking sequences of RRIM712,     Feng et al. (2009) developed 87 EST-SSR primer pairs by analyzing the NCBI database, and the criteria of searching SSR-ESTs was using mononucleotide repeats ≥ 10, dinucleotide to hexanucleotide repeats ≥ 6. But in this study, the search criteria of EST-SSR was changed and 108 novel EST-SSR primer pairs were developed by the same method previous mentioned, of which 4 were selected to analysis the genetic diversity of cultivated varieties and wild accessions. HBE280 appeared to have the most informative despite its repeat number was 4 and the other three novel EST-SSR primer pairs (HBE301, HBE316) also showed different polymorphisms, which were similar to the results of Steinkellner et al. (1997) studied in oak. The four novel primer pairs could be useful as molecular markers in the future study for rubber tree.

Polymorphism of SSR marker
Generally, higher-order repeat motifs, which refers to repeat motif more than three bases as well as the compound repeat, have lower polymorphism than lower-order repeat ones (Dreisigacker et al., 2004;Feng et al., 2009). In this study, higher polymorphism was observed in HBE280 [(gaaa) 4 ] which contained only four times repeats, and this may be due to more A/T content in the repeat unit GAAA/CTTT which deduced a replication origin during DNA replication, and the mismatch cannot be repaired easily because of the existence of rows of (A/T) n.

The polymorphism of EST-and gSSR
SSR molecular markers have been widely used to distinguish crop genotypes (Sun et al., 1999;Virk et al., 1999;Eujayl et al., 2001;Eujayl et al., 2002) and study genetic diversity (Song et al., 2003;Hao et al., 2006;Liu et al., 2007;Caniato et al., 2007). In recent years, a large number of EST-and gSSR molecular markers for many crops were developed (Davierwala et al., 2001;Varshney et al., 2005;Aggarwal et al., 2007). However in rubber tree study, the earliest SSR markers developed by Lespinasse et al. (2000) were from the genomic DNA. Recently, Feng et al. (2009) developed 87 EST-SSR markers from NCBI database. Because of the conservative sequences of ESTs, The level of polymorphism in EST-SSR was lower compared to that of genomic-SSR marker (Eujayl et al., 2002;Leigh et al., 2003;Gonzalo et al., 2005;Yang et al., 2005;Pinto et al., 2006). In this study, gSSRs produced more polymorphisms than EST-SSR as well as PD was slightly higher, despite HBE280 was the most informative marker which belonged to EST-SSRs.
All the SSR primer pairs could be amplified successfully through the 3 related species, and there were no significant differences for transferability between EST-and gSSR, and this was different from other reports (Liewlaksaneeyanawin et al., 2004;Feng et al. 2009). This may due to the relevant lower number of SSR primer pairs.

Diversity analysis and genetic relationship within/between cultivars and wild accessions
Crop genetic diversity is the basis of genetic improvement of crops, and the study was of great significance in the collection, preservation, evaluation and utilization of crop germ plasm resources. In the past few years, genetics investigations of H. brasiliensis have been www.intechopen.com studies (Seguin et al., 1996;Chevallier et al., 1998;Seguin et al., 2002), Lekawipat et al. (2003) used twelve microsatellite markers to detect DNA polymorphism among 108 accessions of H. brasiliensis inclusive of 40 cultivated (Wickham) clones and 68 wild accessions (1981 Amazonian accessions) collected from Amazon forest, and they found wild accessions were more polymorphic than cultivated Wickham clones and could be divided into three clusters, depending on the geographical origin of collection areas such as Acre, Rondonia and Mato Grosso state. In this study, the similar results were also reached and 16 EST-and gSSR molecular markers were used to detect the genetic diversity within/between 45 cultivars planted in China and other countries, of which 33 were cultivars of China, and 11 wild accessions. The results showed that wild accessions were more polymorphic than cultivars, and the analysis of rubber tree germplasm resources by SSR markers consisted with the pedigree analysis approximately, a true reflection of genetic variation and relationships for rubber tree. For example, IAN6645, IAN2904 and FX3899 were clustered together, and they are good clones from Brazil, in which IAN6645 is a descendant from an FX43.655 [FX213 (F4542×AVROS183) ×AVROS183] ×PB86 cross, IAN2904 from FX516 (F4542×AVROS363) ×PB86, FX3899 from F4542×AVROS363, which all have the same parent F4542. The clustering result consisted with the pedigree analysis. PB5/51 and PB5/63 were clustered together, which are high-yielding clones in Malaysia, and they are the descendant from a common PB56×PB24 cross and the clustering result consisted with the pedigree analysis as well; Haiken2, Dafeng99 and Wenchang7-35-11 were clustered together, and the similarity coefficient between Haiken2 Dafeng99 was 0.95, and they are the descendant from a common PB86×PR107 cross. Wenchang7-35-11 from PB5/51×PR107, these three windresistant and high-yielding clone have the same paternal PR107, which is a wind-resistant and high-yielding clone, and the high-yielding maternal PB86 and PB5/51, the clustering result consisted with the pedigree analysis as well. IAN873, GT1, Reken525 and Reyan7-33-97 were clustered together; these results consisted with Feng et al. (2009). At the same time the descendants from the same cross would not been clustered in the same category or nearest position, such as Dafeng99 and Dafeng95 bred by Haiken Dafeng Farm, Hainan, from the same PB86×PR107 cross, and the similarity coefficient between them was 0.72, but they were clustered in group and respectively, more variations would be in F1 hybrid generation. Baoting155, Baoting3410,Baoting911,Wengchang11 and Baoting235 are descendants from the same RRIM600×PR107 cross, but the similarity coefficient between them varied from 0.57 to 0.88, and they were clustered in different group or sub-group. The reasons of inconsistent with the pedigree may be due to that rubber trees are cross-pollinated crops and there is a long-term natural hybridization among the population. The variations between descendants from the different species cross may be from their parents or separation.
With the development of the forest breeding, the consistency of bred varieties increased gradually, resulting in the narrow basis of genetic breeding. Therefore, the reasonable utilization of wild germplasm resources may be an effective way to breeding improvement and widen genetic basis (Benong, 2002;Aidi-Daslin, 2002;Clement-Demange, 2002;Varghese, 2002). In our study, the genetic diversity of wild accessions was higher than the cultivars, which consisted with Besse et al. (1993Besse et al. ( , 1994. Similarity coefficient between cultivars ranged from 0.51 to 0.95, with an average of 0.73, and the variations of the descendants from those crosses with high similarity coefficient would be limited. It was difficult to breed breakthrough varieties. The similarity coefficient between the cultivars and wild accessions was 0.48~0.81, which may provide possibility for the selection of elite parents with improved genetic basis (Varghese, 2002).
It was difficult to distinguish Haiken2 from Dafeng99, GT1 from Reken525 on the level of similarity coefficient of 0.95. Feng et al. (2009) also failed to distinguish GT1 from Reken525 on the level of similarity coefficient of 0.96 by 87 EST-SSR markers, however, according to , there were both the same repeat number and structure in GT1 and Reken525, and the point mutations were found in the flanking regions by sequencing in the HBE156 loci .
The methods of detecting genetic diversity, whether on the level of morphology, cytology (chromosome), physiology, biochemical or even now molecular method, each had its own advantages and limitations in theory or practice, and they cannot be replaced with each other completely. Therefore, using different method reflecting by agronomic characterization or molecular data, the decision should be based on one or several methods selected in detecting and evaluation of the genetic diversity. Analysis of genetic diversity may be help to select the elite and fit parents for improving breeding efficiency.

Detection of loci variation
As exactly as most crops (Fraser et al., 2004;Jung et al., 2005;Aggarwal et al., 2007), Feng et al. (2009) reported that DNRs were the main repeat motif, and AG / TC were the predominant DNRs; In this study, 61 sequences were recovered from 5 loci, of which 44 sequences belonged to AG / TC repeats.
The changes of flanking regions might lead to SSR loci variation, Feng et al. (2009) found that there were point mutations and deletions occurred in the flanking regions in rubber tree, and allelic variations were due to the most frequent InDels (insertions and deletions) in maize flanking regions (Matsuoka et al., 2002). On the contrary, Xie et al. (2006) pointed out that there were no insertion or deletion mutation occurred in AG/CT repeat loci flanking regions for almond, and the similarity results were found in A. thaliana (Symonds & Lloyd, 2003). In this study, of the four AG/CT repeat loci in flanking regions, no insertions or deletions mutation occurred in HBE008, however in HBE164, Dafeng117 and AC/T/15/114 were found AA and GAAA insertion respectively, and deletions occurred in HBE187 and HBE199. Gutierrez et al. (2005) reported that variation was mainly due to the change of the repeat number and insertion, deletion mutation and base substitution in Medicago truncatula, insertion and deletion mutation also led to sequence variation (Feng et al., 2009);Symonds & Lloyd (2003) pointed out that interruptions in the repeat regions of most SSR loci were associated with shortening of the original repeat length in A. thaliana. In rubber tree, substitutions can be found in many alleles, and a complete and long-repeat sequence was divided into several smaller repeats or become relatively short. For example, CT repeats were shortened for C was replaced by G or A in HBE008, and AT repeats were divided into smaller ones for A was replaced by C as well, in HBE199, A was replaced by C or T led the repeat units into several smaller ones.
sequence might lead to a new SSR repeat unit. But in rubber, interestingly, point mutations frequently occurred in flanking sequence of different loci, in HBE063, A was replaced by C, in HBE187, A by G, in HBE87, C by T, and in HBE199, A/T and T/A were replaced with each other. Long repeat sequences are more frequently targets for mutation (Johannsdottir et al., 2000), Symonds & Lloyd (2003) and Xie et al., (2006) have proved that a SSR motif with more repeats should provide an even more efficient substrate for rapid mutation rate in comparison with SSR motifs containing fewer repeats. In rubber tree, more variations occurred in HBE063 supported the above view.

Conclusion
A total of 43 alleles were obtained from 10 EST-SSR and 30 alleles from 6 genomic SSR (gSSR) primer pairs across cultivated and wild accessions; and HBE280, M197 and MnSOD were the most informative SSR markers. All the cultivated and wild accessions were clustered into two big groups. On the level of similarity coefficient 0.68, wild accessions were distinguished from cultivars. Sixty one sequences were sequenced from 5 EST-SSR loci. In comparison with the original EST sequences, insertion, deletion, conversion and transversion mutation occurred in SSR repeats and flanking regions, and long repeat sequences had more variations, and point mutation frequently occurred in flanking regions indicated the new SSR loci in rubber tree.

Acknowledgment
Financial support was provided by the National Non-profit Institute Research Grant of CATAS-ITBB ITBBZD0715 and Post Graduate Study and Academic Leader Grant of Qiongzhou University (QYXB201008).