Assessment of genetic diversity in Venezuelan rice cultivars using simple sequence repeats markers

In Venezuela, pedigree analyses indicate that the rice varieties currently under cultivation are closely related. Effective breeding programs, based on knowledge of the genetic diversity of cultivars, are needed to broaden the genetic bases of rice germplasm in the country. In this study, we used a set of 48 simple-sequence-repeat (SSR) markers to assess the genetic diversity of 11 Venezuelan rice cultivars, released by the National Rice Breeding Program between 1978 and 2007. A total of 203 alleles were detected, the number of alleles ( NA ) per marker ranged from 2 to 9, with an average of 4.23. The average genic diversity ( H ) over all SSR loci for the 18 genotypes was 0.524, ranging from 0.105 to 0.815. Positive correlations were found between H at each locus, NA , the allele size range and the maximum number of repeats. Venezuelan cultivars showed lower H (mean = 0.37) and NA (total = 124, mean = 2.58) than the whole sample. UPGMA-cluster-analysis based on genetic distance coefficients clearly separated all the genotypes , and showed that the Venezuelan rice varieties are closely related. Molecular identification of 7 Venezuelan cultivars could be done with 9 primers pairs which produced 10 genotype-specific-alleles. Although the genetic diversity was low, SSRs proved to be an efficient tool in assessing the genetic diversity of rice genotypes. Implications of the low genetic diversity detected and relatedness of Venezuelan cultivars are discussed.

The commercial exploitation of a reduced genetic base and the prevalence of a small set of landraces in the breeding process had been the general approach for several crops species (Souza and Sorrells, 1989;Dilday, 1990;Cuevas-Perez et al. 1992).Rice (Oryza sativa L.) is one of the most important staple food crops supporting the world population.Compared with other crop species, the genetic diversity in the world rice germplasm is quite large.Three subspecies, i.e., indica, japonica, and javanica, compose a large reservoir of rice germplasm including a variety of local landraces and cultivars (Khush, 1997;Lu et al. 2005;Garris et al. 2005).In addition, there are a number of wild relatives that provide potentially valuable resources for the improvement of cultivated rice (Khush, 1997;Ren et al. 2003).
Despite the richness of genetic resources, only a small proportion of the world rice germplasm collections have been used in breeding programs.As a consequence a high genetic similarity is found within several commercial rice germplasms around the world.The limited use made of the rice genetic diversity available worldwide has been a concern in Latin America since the late 1980s.Cuevas-Perez et al. (1992) analyzed a total of 143 commercial varieties released in the region from 1971 to 1989, it was found that 101 different landraces were involved in the crosses that produced the varieties, however only 14 ancient cultivars contributed 70 percent of the genes.A similar situation has been reported for the upland varieties cultivated in the region.Six native varieties make up the base for the upland varieties released up to 1992.Forty ancestors were involved in crossing to develop varieties, but only 11 of them accounted for 81 percent of the genes for the varieties released between 1971 and 1993 (Guimaraes et al. 1995).Although differences in genetic diversity and relatedness are observed within rice germplasms of different countries, the general feature is a very close relationship among cultivars (Cuevas-Perez et al. 1992;Guimaraes et al. 1995;Rangel et al. 1996;Fuentes et al. 1999).In Venezuela, for example, irrigated rice varieties currently under cultivation are closely related among them and with cultivars from Colombia, Brazil and Ecuador (Cuevas-Perez et al. 1992).
Knowledge regarding the amount of genetic variation in germplasm accessions and genetic relationships between genotypes are important considerations for designing effective breeding programs.In the past, the characterization of germplasm diversity was carried out by means of morphological and biochemical markers which, in many cases, did not have the resolution power for revealing polymorphisms in genetic analyses and/or for differentiating between closely related genotypes.Advances in plant genetics and molecular biology have led to the development of many types of molecular markers which can be used to characterize germplasm.Different types of DNA markers are available nowadays, each method differing in principle, application, type and amount of polymorphism detected, and cost and requirement.These include random amplification of polymorphic DNA (RAPDs), restriction fragment length polymorphisms (RFLPs) amplified fragment length polymorphisms (AFLP) and simple sequence repeats (SSRs).SSRs are codominant, abundant and highly reproducible and exhibit a high degree of allelic variation (Panaud et al. 1996;Temnykh et al. 2000).SSRs are an excellent molecular marker system for many types of genetic analyses, including linkage mapping, germplasm surveys, and phylogenetic studies.They have been used for characterizing genetic diversity in several crop species including sorghum (Dean et al. 1999;Smith et al. 2000), maize (Senior et al. 1998), cotton (Liu et al. 2000) and wheat (Prasad et al. 2000).In rice, SSRs have been used to assess the genetic diversity of both wild and cultivated species (Yang et al. 1994;Gealy et al. 2002;Ni et al. 2002;Ren et al. 2003;Sarla et al. 2003;Yu et al. 2003;Siwach et al. 2004;Xu et al. 2004;Brondani et al. 2005;Jeung et al. 2005;Neeraja et al. 2005).These studies showed that SSR markers are efficient in detecting genetic polymorphisms and discriminating among genotypes.
The objectives of this study were to use SSR markers to estimate the genetic diversity within a core of 11 Venezuelan rice cultivars released by the National Rice Breeding Program between 1978 and 2006, to reveal genetic relationships among them, and to distinguish different accessions by means of specific SSR alleles.

SSR assay
DNA was extracted from fresh seedling leaves of each accession following the method of Dellaporta et al. (1983) with slight modifications (Dellaporta et al. 1983;Posso-Duque and Ghneim-Herrera, 2008).Forty-eight SSR markers covering all the twelve chromosomes at about a 30 cM intervals were selected for the genetic diversity analysis based on the published rice SSR framework maps (Temnykh et al. 2000;Coburn et al. 2002).The loci, chromosome position, repeat motifs and primer sequences for these markers are presented in Table 2. SSR primers were obtained from Integrated DNA Technologies, Inc. (Skokie, IL, USA).
PCR amplification was carried out in 20 ml reaction mixtures, each containing 20 ng of template DNA, 0.13 mM of each primer, 250 mM of each dNTPs, 1X PCR buffer (100 mM Tris-HCl, pH 8.8, 500 mM KCl, 2 mM MgCl 2 , 1% Triton X-100, 1% BSA) and 1 U of Taq DNA polymerase.An MJ Research (PTC-100 TM 96V) thermocycler was used along with the following PCR profile: an initial denaturation step of 3 min at 94ºC, followed by 30 cycles of 30 sec at 94ºC, 45 sec at 48-65ºC, 1 min at 72ºC, and a final extension at 72ºC for 5 min.
The PCR products were resolved on a 6% sequencing gel followed by silver staining following the standard protocol (Temnykh et al. 2000).Silver-stained gels were scanned to capture digital images of the gels after air drying.The allele size of the amplified band for each SSR locus was determined based on its migration relative to the 10 bp DNA ladder (Invitrogen Corp., CA, USA) and three check varieties ('Azucena', 'IR64' and 'Nipponbare') that were loaded in each gel as references.
For genetic distance estimates and cluster analysis, the allele molecular sizes were used as codes for the different alleles detected in each locus.

Data analyses
The allelic diversity of the SSR was calculated according to the diversity index, H, described by Nei (1987), in the following formula: Where p ij is the frequency of jth allele for the marker i and summation extends over n alleles.For each marker locus, the total numbers of alleles and allele sizes were calculated using Powermarker (version 3.25).The most frequent alleles and genotype-specific alleles were found using MICROSAT (version 1.5d) (Minch et al. 1997).The "maximum repeat count" for each SSR locus was calculated as described by Cho et al. (2000) using the following formula, Max repeat count = [(max allele MWreference allele MW) / x] + reference repeat count (x = 2 or 3 for dinucleotide and trinucleotide repeats, respectively).The reference repeat count was taken directly from the known sequence of 'Nipponbare'.Pearson correlation coefficients were used to evaluate the relationships between the genetic diversity, the number of alleles, the maximum repeat count or the size range; these calculations were made using SPSS (version 10.0.1).
Data from all polymorphic SSR markers were used for phylogenetic analysis to determine genetic relationships.Genetic distances (Rogers, 1972) between two entries were computed as RD ij = ½ [∑ (X ai -X aj ) 2 ] ½ , where X ai is the frequency of the allele a for individual i, and X aj is the frequency of the allele a for the individual j.A dendrogram showing genetic relationships of the 18 accessions was constructed based on these distances using the unweighted pair-group method with the arithmetic mean (UPGMA).The strength of the dendrogram nodes was estimated with a bootstrap analysis using 1000 permutations with Phylip software (version 3.65) (Felsenstein, 2004).

SSR diversity
All 48 SSR markers were found to be polymorphic among the 18 rice genotypes.These markers revealed a total of 203 alleles (Table 3); the average number of alleles per locus was 4.23, ranging from 2 to 9. The overall size of PCR products amplified using 48 primer pairs ranged from 77 to 295 bp.The molecular size difference between the smallest and the largest allele for a given locus varied from 3 to 95 bp (Table 3).SSR markers with ATT, GA, TAA and GTT motifs showed the maximum variation in allele size.There was a considerable range in allele frequency (27.8-94.4%), on average 60.2% of the varieties shared a common allele (Table 3).
The average genic diversity over all SSR loci for the 18 genotypes was 0.524, ranging from 0.105 (RM272 on chromosome 1) to 0.815 (RM547 on chromosome 8) (Table 3).The genic diversity at each SSR locus was significantly correlated with the number of alleles detected (r = 0.739, P < 0.01) and with the allele size range (r = 0.402 P < 0.01) (Table 4).Loci with two-nucleotide repeat motifs (mean number of alleles 4.8, n = 24) and complex repeat motifs (mean number of alleles 4.8, n = 4) were more polymorphic than those with tri-nucleotide repeat motifs (mean number of alleles 3.9, n = 15) or tetra-nucleotide motifs (mean number of alleles 2.2, n = 5).The maximum number of repeats within the SSRs was significantly correlated with the genic diversity (r = 0.362, P < 0.05) and with the number of alleles at a locus (r = 0.505, P < 0.01) (Table 4).
Analyses of genetic diversity and allele number were also performed considering only the Venezuelan accessions (Table 3).For these accessions, only thirty-six SSR markers showed polymorphism.Analysis based only on polymorphic markers showed a total of 112 alleles (Table 3).The average number of alleles decreased to 3.11, ranging from 2 to 5, while the average genic diversity was reduced to 0.49, ranging from 0.14 (RM169 on chromosome 5) to 0.73 (RM110 on chromosome 2).As for the whole set of accessions, the genic diversity observed at each locus for the Venezuelan accessions was significantly correlated with the number of alleles detected (r = 0.861, P < 0.01), and the maximum number of repeats (r = 0.323, P < 0.05).

Genetic relationships among rice cultivars
A dendrogram was constructed based on the Rogers genetic distance calculated from the 203 SSR alleles generated from the 18 rice accessions.All 18 rice cultivars could be easily distinguished.The UPGMA cluster tree analysis led to the grouping of the 18 varieties in three major groups (Figure 1).Group I is comprised of indica varieties, which includes the Venezuelan and Colombian cultivars and 'IR64'.Group II is constituted by the japonica varieties 'Nipponbare', 'Azucena' and 'Caiapó'.The African wild species Oryza glaberrima is clearly differentiated from all other accessions in Group III.The bootstrap analysis supported the consistence of these groups; japonica and indica cultivars were clearly separated 100% of the times, while O. glaberrima was out-grouped 98% of the times (Figure 1).

DISCUSSION
The assessment of genetic diversity is an essential component in germplasm characterization and conservation.The results derived from analyses of genetic diversity at the DNA level could be used for designing effective breeding programs aiming to broaden the genetic bases of commercially grown varieties.A narrow genetic base has been reported for Latin American commercial rice cultivars, mainly based on pedigree and/or biochemical analyses (Cuevas-Perez et al. 1992;Guimaraes et al. 1995;Rangel et al. 1996;Fuentes et al. 1999;Aguirre et al. 2005).
To our knowledge this work represents the first study of genetic diversity and molecular characterization of Venezuelan rice germplasm using SSR markers.Furthermore, there are few published reports of DNA fingerprinting of Latin American commercial rice cultivars (Fuentes et al. 1999;Aguirre et al. 2005;Giarrocco et al. 2007).
SSRs were chosen for the analysis of genetic diversity of Venezuelan rice cultivars because several works have showed these markers are very powerful for differentiating individual germplasm accessions, particularly when they are closely related (Bligh et al. 1999;Xu et al. 2004;Jeung et al. 2005).Additionally, SSRs show a series of advantages when compared with other DNA-based markers, such as abundance in the genome, high level of polymorphism, repeatability, co-dominance and costeffectiveness (Ni et al. 2002).
In this study we evaluated 48 SSR markers in 18 rice cultivars, eleven of these genotypes represent indica varieties commercially cultivated in Venezuela.The other seven cultivars represent 3 indica and 3 japonica cultivars and the African wild species Oryza glaberrima.All 48 were polymorphic across the 18 genotypes.A total of 203 alleles were detected with an average number of alleles of 4.23 per locus (range 2 to 9 per locus).This value is quite low compared with those reported for the worldwide collection (range 2 -11, mean = 6.3) and other large scale studies (range = 3-17, mean = 7.4) (Olufowote et al. 1997;Yu et al. 2003) but quite comparable to values reported for studies performed on smaller germplasm sets (Cho et al. 2000;Hashimoto et al. 2004;Siwach et al. 2004).A set of 12 SSR markers (25%) resulted monomorphic when evaluated exclusively in the set of Venezuelan accessions, the average number of alleles for the 36 SSRs that showed polymorphism was 3.11 (range 2 to 5), a value significantly lower than previously reported for SSRs.
The structure and length of simple sequence repeats are considered to be the major factors affecting microsatellite variability (Brinkmann et al. 1998).In general, SSR loci with more repeats tend to be more polymorphic and have larger amplitude of variation (Panaud et al. 1996;Innan et al. 1997;Schug et al. 1998;Cho et al. 2000).The results obtained in this study are in concordance with this general pattern, SSR markers with many repeat units (RM110, RM144, RM262, and RM547) exhibited higher H values, higher number of alleles and larger size differences among alleles, thus high correlation coefficients were found between these parameters.
The UPGMA cluster analysis showed that all 18 rice cultivars could be easily distinguished based on the information generated by the 48 SSR markers.As expected, cultivars were separated in three clear groups corresponding to the indica and japonica subspecies and O. glaberrima.The 48 SSR markers also allowed the distinction among indica accessions; all 14 indica cultivars were clearly distinguished even though a high relatedness or similarity was measured between cultivar pairs.Our results support the contention that SSR marker systems can distinguish genetically close breeding lines and cultivars, and validate their use in the characterization of rice germplasm accessions.
Venezuelan cultivars used in the present study have not been examined previously in terms of genetic relatedness using molecular markers.The low genetic diversity found among the Venezuelan accessions evidence the narrow genetic bases used in our breeding programmes.References to the narrow genetic base of cultivated rice varieties are available from several regions, including Latin America (Cuevas-Pérez et al. 1992;Guimaraes et al. 1995;Fuentes et al. 1999;Aguirre et al. 2005), Japan (Hashimoto, 2004), USA (Dilday, 1990;Xu et al. 2004), Korea (Song et al. 2002) and Taiwan (Lin, 1991) Our results indicate that it is essential to broaden the genetic base of the rice varieties cultivated in the country to reduce its vulnerability to diseases and insect pests.Recent studies carried out by the International Rice Research Institute showed there is still a tremendous amount of unexploited genetic diversity in the primary gene pool of rice that can be used for enhancing the diversity in local germplasms and their performance under diverse agroecological conditions (Guimaraes, 2000;Ali et al. 2006;Lafitte et al. 2006).Wild species of Oryza also represent a potential source of new alleles for improving yield, quality, and stress resistance in rice cultivars (Xiao et al. 1998;Moncada et al. 2001;Ahn et al. 2002;Thomson et al. 2003;McCouch, 2004;Kovach and McCouch, 2008).
Several studies report improvements in performance because the introgression of valuable genes from wild germplasm into elite rice cultivar.Lines derived from crossing the wild species Oryza rufipogon with Oryza sativa cultivars showed higher yields than their progenitors and are tolerant to several abiotic stresses (Moncada et al. 2001;Nguyen et al. 2003;Tian et al. 2006;Xie et al. 2006;McCouch et al. 2007).Yield and grain quality enhancing alleles have also been identified from O. glaberrima (Aluko et al. 2004;Li et al. 2004;Sarla and Swamy, 2005) and O. glumaepatula (Brondani et al. 2002;Rangel et al. 2005).Utilisation of these "novel" gene sources is underway in rice breeding programs of several Latin American countries including Venezuela.

Figure 1 .
Figure 1.Dendrogram of 18 rice genotypes clustered based on Rogers' genetic distance calculated from 48 SSR markers (203 alleles) by UPGMA analysis.The strength of dendrogram nodes was estimated with a bootstrap analysis using 1000 permutations.

Table 2 . List of the 48 simple sequence repeat (SSR) markers used in this study. Table 3. Data on the number of alleles, allele size range, highest allele frequency and genic diversity (H) at each 48 loci found among 18 rice genotypes.
Allele number and genic diversity for each SSR marker were also calculated considering only the Venezuelan accessions (in brackets).

Table 4 . Correlations (Pearson's coefficient) calculated for genic diversity (H), the number of alleles, the number of repeats and the allele size range.
Correlations were calculated based on data derived from the analysis of 48 SSRs on the 18 genotypes.
included to show their relationship with Venezuelan cultivars.

Table 6 . Genotype-specific alleles identified for 13 rice cultivars analyzed in this study.
. In the (Cuevas-Perez et al. 1992ce germplasm, an extensive study based on pedigree analysis carried out in 1992 indicated that a group of only 14 landraces accounted for nearly 70 irrigated rice cultivars released in the region during the period1971-1988(Cuevas-Perez et al. 1992).The commercial rice varieties analyzed in this study represent 79% of the rice cultivars liberated by the Venezuelan Rice Breeding Program since 1978.Our results indicate that, despite breeding efforts, the varieties under current cultivation ('Fundarroz PN-1', 'Venezuela 21', 'Centauro', and 'Cimarrón') are closely related among them and even with cultivars released more than 20 years ago ('Araure 1' and 'Araure 4').A recent study byArnao et al. (2008)using AFLP markers showed similar results, even when cultivars and experimental lines from other national breeding programs were included in the study.A pedigree analysis included in this work showed a strong contribution of only few progenitors to main cultivars and breeding lines(Arnao et al. 2008)