Microsatellite marker-based genetic variability in Spanish rice cultivars and landraces

The genetic relationships among japonica rice cultivars and landraces were studied. Most of them are of Spanish origin, and were developed and cultivated for varying time periods over more than a century. To uncover genetic diversity within each cultivar, we analysed 30 plants per cultivar or accession using 10 fluorescently labelled primer pairs for SSR markers. Six cultivars were included in the study, with accessions from four different regions of Spain of the traditional cultivar Bomba. A total of 37 alleles were detected with a mean of 3.7 alleles per locus. Polymorphism information content (PIC) ranged from 0 to 0.78 with an average of 0.51 per locus. Genetic diversity for cvs. Albufera, Bahia, Taipei 309 and Jsendra ranged between 0.008 and 0.062. Cultivar Bomba accessions displayed higher number of alleles and genetic diversity, with Tarragona the lowest (0.08) and Valencia the highest (0.16) genetic diversity. Although infrequent, some heterozygous individuals were detected in cvs. Bomba, Bahia and Taipei 309. Cluster analysis enabled to categorize four accessions of cv. Bomba and cv. Albufera in one group and cvs. Bahia, Taipei 309, Jsendra and Senia in a second group supporting previously known relationships among varieties. These results show the utility of SSR markers for characterization of rice cultivars. In addition, the use of high number of individuals per cultivar, enabled to evaluate both interand intra-varietal genetic diversity. Therefore, this investigation should be useful in rice breeding programmes for genotype identification and for the evaluation of germplasm purity. Additional key words: japonica rice; molecular characterization; Oryza sativa L.; SSR markers.


Introduction
Rice (Oryza sativa L.) is a major cereal crop of high agronomic and nutritional importance as global rice production is over 650 million tonnes (http://www.fao.org/).It is highly polymorphic with wide geographical and genetic differentiation (Sarla et al., 2005).Rice landraces, maintained through traditional farming practices, possess high genetic diversity and specific traits such as disease resistance, environmental constraint tolerance and nutritional quality which are often used in crop improvement (Camacho-Villa et al., 2005).Furthermore, landraces are adapted to local agro-environmental conditions which contributes to yield stability and hence, they continue playing an important role in traditional and subsistence farming (Camacho-Villa et al., 2005).However, extensive efforts to improve rice productivity have led to large-scale cultivation of highyielding genetically uniform varieties, the replacement of local cultivars and the concomitant decrease in rice genetic resources that created a widespread concern to promote conservation of traditional cultivars/landraces (Kohli et al., 2004;Camacho-Villa et al., 2005;Barry et al., 2007).
Conventionally, genetic diversity has been estimated on the basis of the morphological and physiological markers.Molecular markers offer additional advantages such as high polymorphism and independence from effects related to environmental conditions and the physiological stage of plant.Molecular characterization can be assessed again after several years of maintenance and new accessions can be related to existing collections which provides useful information for different breeding programmes (Bolaric et al., 2005).In addition, molecular markers provide information on possible genetic mechanisms for observed evolutionary patterns (Bautista et al., 2001).
Research on genetic diversity and phylogenetic relationships within both cultivated and wild rice species is necessary for better understanding of rice evolution and for aiding plant breeders in selecting appropriate materials for further genetic improvement of cultivars.In this respect, microsatellites or simple sequence repeat (SSR) markers are abundant and well-distributed throughout the 12 rice chromosomes.More than 2,200 SSR markers had been described in rice by McCouch et al. (1997McCouch et al. ( , 2002a,b),b).
Microsatellite markers has been used to study interspecific relationships between Oryza sativa and either wild (O.rufipogon and O. nivara) or cultivated (O.glaberrima) species (Ni et al., 2002;Barry et al., 2007).Ni et al. (2002) also determined the pattern of diversity in indica and japonica rice subspecies and, within the japonica group, in both temperate japonica and tropical japonica types.Additionally, the genetic diversity of a number of rice cultivars and landraces from several countries/regions have been studied by the use of SSR markers (Luce et al., 2001;Brondani et al., 2006;Ram et al., 2007;Tu et al., 2007).
Although rice is a major food crop in Europe, with an annual production of over 3.5 million tonnes, its cultivation is restricted to few countries.In this respect, efforts are directed to exploit and characterize the European and international rice germplasm from the temperate rice area for their conservation and sustainable use in agriculture (http://www.eurigen.net).Spain is the second European rice producer with estimations for 2008 of over 660 thousand tonnes harvested in approximately 96 thousand hectares (http://www.mapa.es).Some rice cultivars of agronomic importance in Spain have been developed in different time periods that range from the end of the XIX to the XXI century.Thus, the traditional cv.Bomba was one of the only two cultivars extensively used by the 1890s.Although the high plant size hinders productivity, cv.Bomba is still highly appreciated due to the grain organoleptic properties.Cultivars Bahia and Senia (Bahia × Sequial) have been the most widely cultured in the 80s and 90s, respectively.Cultivars Jsendra (Senia × M202) and Albufera (Bomba × Senia) were registered during 2005 and 2007, respectively, and Jsendra is already extensively cultivated.
Cell lines and seedlings of cvs.Bomba and Bahia were used in a number of experiments related to abiotic stress tolerance in our previous research (Perales et al., 2005;Wankhade et al., 2010), and Taipei 309, a cultivar with high competence for genetic transformation, has long been utilized in our laboratory as a source of cell lines (Moukadiri et al., 1999(Moukadiri et al., , 2002;;Bahaji et al., 2002;Perales et al., 2008).
In this research SSR markers were used to evaluate in detail the genetic variability and relationships among five japonica rice cultivars (Bomba, Bahia, Senia, Jsendra and Albufera) that have been extensively cultivated in Spain.Cultivar Taipei 309 was also analyzed for comparative purposes.The application of SSR markers to a large number of individuals per cultivar or accession has allowed the reliable characterization of both varieties and landraces.

Plant materials
The following six japonica rice cultivars were used in this investigation: cvs.Bahia, Senia, Bomba, Jsendra, Albufera and Taipei 309.Cultivars Bahia, Senia, Albufera and Jsendra were developed, maintained and kindly provided by the Rice Department of the Instituto Valenciano de Investigaciones Agrarias (IVIA) at Sueca (Valencia).Due to phenotypic variability, four accessions of cv.Bomba from different Spanish rice growing regions were evaluated: Murcia, Sevilla, Tarragona and Valencia.They were kindly provided by cooperatives Virgen de la Esperanza, Calasparra, Murcia; Veta la Mora, Isla Mayor, Sevilla; and the Rice Department, IVIA (Tarragona and Valencia accessions).
Rice plants were grown in the greenhouse with day/ night temperatures of 27/18°C and relative humidity of 60-80%.Leaf samples were collected when seedlings were in their vegetative growth stage and stored at -80°C.

DNA extraction, fragment amplification and analysis
Genomic DNA extractions from a total of 270 plants, 30 plants per cultivar and accession, were performed according to CTAB protocols described by Doyle and Doyle (1987).Leaf fragments of approximately 5 cm in length from each genotype were ground in liquid nitrogen in 1 mL extraction buffer using mortar and pestle.The purity and concentration of extracted DNA was evaluated by horizontal electrophoresis on 0.8% (w/v) agarose gel in 1x TBE buffer using a known ladder of Lambda DNA/EcoRI+HindIII (Fermentas Life Sciences).The stock DNA samples were stored at -20°C and working DNA samples, diluted with sterile distilled water (1:100 v:v) at 4°C for PCR use.
Primer sequences were obtained from Chen et al. (1997) andMcCouch et al. (2002a,b).They were selected according to data on variability reported by Coburn et al. (2002).Initial SSR amplifications were performed with 11 non-labelled primers and based on satisfactory amplification results, 10 fluorescently labelled primers (Applied Biosystems) were selected (Table 1).
DNA amplification was carried out in a 25 µL reaction mixture containing 2.5 µL 10x reaction buffer with 1 to 2 mM MgCl 2 , 0.625 U of Taq DNA polymerase (Netzyme Molecular Netline Bioproducts), 0.2 µM forward and reverse primer (Applied Biosystems), 20 mM dNTP mix (Fermentas Life Sciences) and 10 µL DNA.The Polymerase Chain Reaction (PCR) was performed in PTC-100 (MJ Research) as follows: 4 min at 95°C, 35 cycles of 30 sec at 94°C, 45 sec at 55°C, 1 min at 72°C, and 5 min at 72°C for the final extension.Annealing temperature was adjusted for each primer according to melting temperature (Chen et al., 1997;McCouch et al., 2002a,b).The amplified PCR products were subjected to electrophoresis in 2% (w/v) agarose gel in 1x TBE buffer at 100-130 V along with a known 100 bp DNA ladder (Fermentas Life Sciences).DNA bands were stained with ethidium bromide for 5 min and visualised under UV light (Gibco BRL UV Transilluminator).The 1.5 µL of PCR products were dissolved in 30 µL of sterile distilled water for fragment analysis.

Data analysis
Different size of amplification products from the same primer pair were considered as different alleles.As RM248 did not show variability, a matrix with 270 individuals and 18 character states was generated.Polymorphism Information Content (PIC) or expected heterozygosity (H) for each SSR marker was calculated based on the formula: where p i is the allele frequency for the i-th allele (Nei, 1973).Three different pairwise distances (F ST , Slatkin, 1995) among cultivars implemented in ARLEQUIN vs. 3.00 (Excoffier et al., 2005) were computed and Genetic variability of rice cultivars different clustering methods (single, complete, weighted and un-weighted pair-group method arithmetic average) were checked using NTSYS 2.2 (Rohlf, 2002) to select the best quality indicator (the cophenetic value).The combination of F ST and un-weighted pair-group method arithmetic average (UPGMA) was selected.

Results
All SSR loci amplified consistently size fragments, yielding a total of 37 alleles in the six cultivars and accessions studied.The number of alleles per locus varied widely among 10 primer pairs, ranging from 1 (RM248) to 8 (RM206) with an average of 3.7 alleles per locus.The overall size of amplified PCR products ranged from 81 to 343 bp, with most of the sizes between 125 and 260 bp (Table 1).
No relationships were found among the number of alleles and nucleotide motif composition or the number of repetitions in the motifs (Table 1).The only inva-riable, monomorphic locus (RM248) showed the lowest size of 81 bp, while the most variable loci (RM206 with 8 alleles and RM333 with 7 alleles) ranged between 172 to 196 and 171 to 198 bp respectively.On the other hand, only two alleles were found in RM171, a locus with the largest bp size.Two loci, RM224 and RM464 with bi-nucleotide (AT) repeat motifs detected low number of alleles i.e. 3 and 2 respectively, while RM206, RM225, RM234 and RM248 with CT repeat motifs ranged from 1 to 8 alleles.RM333, RM400 and RM481, loci with tri-nucleotide repeat motifs detected 7, 4 and 5 alleles respectively; while the number of alleles was 2 for the only tetra-nucleotide motif locus (RM171).
alleles were widespread among cultivars and loci, being cv.Bomba the variety with higher number of them and RM206 and RM333 the loci showing the higher number of alleles (Table 2).Low frequency alleles were observed in 7 out of 10 SSR loci studied.Loci RM206, RM333, RM481 and RM224 showed 6, 4, 3, and 2 rare alleles respectively; while 1 allele each was detected in loci RM225, RM234 and RM400.As expected, loci showing greater number of alleles detected more low frequency alleles too.Frequencies of alleles are available upon request.

Genetic diversity among varieties
Genetic diversity was estimated through total number of alleles, average number of alleles and average genetic diversity (Table 3).All these parameters ranged widely as a function of the cultivar.Cultivar Senia showed no variation in any of the loci.Four cultivars (Albufera, Taipei 309, Bahia and Jsendra) showed low levels of variation with total number of alleles ranging between 11 to 13 with an average of 1.11 to 1.33 alleles per locus and average genetic diversity between 0.008 and 0.062.Conversely, high levels of variation were found in cv.Bomba whose total number of alleles among accessions ranged between 17 to 21 with an average of 1.66 to 2.22 alleles per locus.Within cv.Bomba accessions, Tarragona showed the lowest number of alleles and levels of diversity, while Sevilla had the highest number of alleles and Valencia the highest levels of genetic diversity.A total of 127 alleles were estimated for ten microsatellite markers across six rice varieties including accessions of cv.Bomba from four different regions (Table 3).
Ten individuals, representing a 2.7% of the total, showed different alleles in heterozygosis.Two heterozygous individuals belonged to cvs.Bahia and Taipei 309, while the remaining eight were detected in accessions of cv.Bomba, i.e. 4 in Tarragona, 2 in Sevilla and 1 each in Valencia and Murcia.Heterozygotes were detected in five out of ten SSR loci (RM206, RM224, RM234, RM333 and RM481) representing five heterozygous individuals for locus RM333, two for RM481 and one each for loci RM206, RM224 and RM234.

Cultivars relationships and characterization
Affinities among cultivars and accessions are shown in the dendrogram obtained using F ST distances, forming two different groups (Fig. 1).The first cluster included four accessions of cv.Bomba and cv.Albufera where cv.Bomba accessions were closer among them than any of them to cv.Albufera.The second cluster was composed of cvs.Bahia, Taipei 309, Jsendra and Senia where cvs.Bahia and Taipei 309 were closer than any of them to cvs.Jsendra or Senia (Fig. 1).
Taking together the combinations of alleles from all nine variable loci, all the studied cultivars can be differentiated as follows: cvs.Jsendra and Senia, varieties that shared all alleles except for locus RM206.Similarly, cvs.Bahia and Taipei 309 showed different more frequent alleles for loci RM225 and RM333.Likewise, cvs.Bomba and Albufera shared most of the alleles in the studied loci except an allele RM400-318 which was shared with Senia (Table 2), showing its hybrid origin.Cultivar Bomba show higher similarity than cv.Senia with respect to cv.Albufera (Bomba × Senia).Among cv.Bomba accessions different combinations of more frequent alleles, mainly at locus RM206, allowed the identif ication of their origin (Table 3).

Discussion
The genetic relationships among five Spanish rice cultivars that were developed within a long timeframe as well as their genetic diversity were investigated.In this respect, we focused on the diversification of the traditional cultivar Bomba into several landraces.
Our results were obtained with relatively few primers, showing the ability of SSR markers to characterize these Spanish cultivars, as well as cv.Taipei 309.Therefore, they corroborate previous investigations (Nagaraju et al., 2002;Ravi et al., 2003;Sarla et al., 2005;Brondani et al., 2006;Lapitan et al., 2007;Singh et al., 2010) in which SSR markers had been able to discriminate among cultivars.
The size variations of alleles are within the range reported in previous research (Temnykh et al., 2000;Coburn et al., 2002).The PIC values across the variable loci ranged between 0.40 (RM225) and 0.78 (RM333) (Table 1), being lower than those reported by several other authors (Coburn et al., 2002;Nagaraju et al., 2002;Pessoa-Filho et al., 2007).Thus, high levels of heterogeneity of PIC values indicate clear differences among the studied cultivars.
Among cultivars, relationships had been shown through a cluster analysis based on F ST distances, showing that cvs.Bomba and Albufera are genetically closer than any of the studied varieties, while the remaining varieties are in a second group where cvs.Bahia and Taipei 309 are closer between them than to cvs.Jsendra and Senia.Cultivar Albufera is a hybrid of cvs.Bomba and Senia that was generated with the aim of maintaining the grain properties of cv.Bomba but with increased productivity (Carreres et al., 2006), therefore, the short genetic distance between cvs.Bomba and Albufera should be expected.On the other hand, the close relation between cvs.Taipei 309 and Bahia is very interesting at the cellular level as both show higher morphogenic ability in vitro than a number of japonica cultivars we tested (unpublished observations).
Among cv.Bomba accessions, the distance between Tarragona and Murcia was very low (0.008) therefore it appears as 0 in the dendrogram.Interestingly, Valencia and Sevilla were closer in the cluster while farther geographically.This might be related to the migration of Valencian rice growing farmers, during 1930s-40s, to Sevilla.The high number of alleles in cv.Bomba accessions together with the fact that several of them were shared with other varieties agrees with a low fixation of characters of this traditional cultivar of unknown origin.
Self-pollination (autogamy) results in restricted gene flow between landraces of rice and hence, heterozygosis is considered to be rare in these plants (Garris et al., 2005;Wang and Lu, 2006).However, Kohli et al. (2004) reported that low rates of heterozygosity sometimes exist in rice varieties.Our results suggest that five out of the six varieties studied, were highly homogeneous and the heterozygosity found in cvs.Bahia and Taipei 309 should be considered as residual.

Genetic variability of rice cultivars
On the contrary, the levels of heterozygosity detected in cv.Bomba accession were relatively high, representing a 6.66%.Since this traditional cultivar was originated more than a century ago the possibility of crosspollination prior to proper ex situ conservation cannot be excluded.
Although Spain is the second rice producer in Europe, few rice varieties have been developed.Identification of each of the studied cultivars as well as different provenances of cv.Bomba is possible through alleles and characteristic combinations for each one of them, which has implications in germplasm conservation.This is particularly important for cv.Bomba accessions which should be studied deeply to gain a better knowledge on the genetic composition of its landraces for an accurate preservation.The hybrid origin of cv.Bomba is supported by the fact that it shares two alleles for locus RM225 as well as several others in low frequency present within other varieties.Presence of low frequency alleles among different rice varieties were also observed by several other researchers and can be useful for breeding purposes (Jain et al., 2004;Siwach et al., 2004;Lapitan et al., 2007;Luan et al., 2008).Among European cereals, the diversification of traditional cultivars into landraces had been frequent (Zeven, 1998) as there were no individual institutions in charge of their preservation.As far as we know, all the cv.Bomba collections maintain the grain properties highly appreciated by the consumers and the high size of plants that hinders their productivity.These findings are useful tools to detect adulteration of rice varieties, particularly of cv.Bomba and can support their utility in quality assurance of agro-economically important traits (Siwach et al., 2004).However, some diversity has been noticed in additional phenotypic characteristics and they appear to be reflected in the data presented here.At present, cereal landraces and traditional varieties are highly valued in Europe as the source of ingredients for traditional food and drinks (http://www.niab.com/research/pgbe/genetics/diversity-genomics-roup/ research/ projects/ landraces.html)and their preservation in germplasm collections or in situ is organized by Biodiversity International (http://www.bioversityinternational.org).This is the first report on the genetic variability of cv.Bomba, a traditional cultivar, as well as on the genetic characterization of rice cultivars cultured along time periods covering more than a century.The results obtained should also be useful in the development of breeding programmes as they provide a valuable tool for the precise identif ication of rice varieties and landraces.

Table 1 .
Chromosome location (CL), size range (SR), fluorescent marker, primer sequences, motifs, number of alleles (NA) and polymorphism information content (PIC) for ten simple sequence repeat (SSR) markers used in the study

Table 2 .
Data on the alleles of the different simple sequence repeat (SSR) for each cultivar and source.Alleles with less than 10% frequencies are indicated in bold digits (3 individuals or 6 copies)

Table 3 .
Total number of alleles (T NA ) for 10 simple sequence repeat (SSR) loci, average number of alleles (A NA ), average genetic diversity and alleles allowing cultivars and source identification (indicated by the primer name followed by allele size).Low frequency alleles were not considered for cultivar characterization purposes