| Edited by Minoru Murata. Nam-Soo Kim: Corresponding author. E-mail: kimnamsu@kangwon.ac.kr |
Oryza is an agronomically important genus with a highly diverse morphology. The genus consists of approximately 24 diploid and tetraploid species with a basic chromosome number of 12 and with genome constitutions of AA, BB, CC, EE, FF, BBCC and CCDD, (Aggarwal et al. 1999). Morishima and Oka (1970) have grouped the various Oryza species into three complexes, Oryza sativa, Oryza officinalis, and Oryza ridleyi, depending on their genomic constitutions. The O. sativa complex numbering seven to eight depending on the classification, contains the AA diploid species and includes the two cultivated species, O. sativa and O. glaberrima, (Vaughan 1994). O. sativa is grown worldwide and contains two different ecotypes, O. sativa spp. Japonica and O. sativa spp. Indica that have been reproductively isolated by a sterility barrier in the F1 hybrids (Oka 1988; Second 1982). Oryza rufipogon and Oryza nivara are the direct wild relatives of O. sativa (Khush 1997). However, these two species were often classified into a single species, O. rufipogon, since they are interfertile and variation between them is continuous (Oka 1988). O. glaberrima is grown in limited areas in West Africa. Oryza longistaminata and Oryza barthii are the wild relatives of the O. glaberrima. Oryza glumaepaetula and Oryza meridionalis are wild AA diploids from Latin America and Australia, respectively (Vaughan 1994).
Studies on genetic diversity and species relationships are important not only for understanding the genetic structure of the genus but also for breeding purposes such as introgression of useful genes from wild relatives to cultivated crops. In Oryza species, genomic relationships among the species have been delineated by various kinds of genetic and cytogenetic analyses including chromosome pairing (Katayama 1982), morphology (Morishima and Oka 1970), isozymes (Glaszmann 1987), and DNA polymorphisms (Wang et al. 1992; Aggarwal et al. 1999; Ge et al. 1999; Ren et al. 2003). Analyses with isozyme and DNA polymorphisms, including both nuclear and chloroplast DNA, have shown that the species in the O. sativa complex are clearly separated from the other Oryza species. This supports the idea that these species have undergone different evolutionary rates in concerted fashion (Morishima 2001). In a study of genetic variation among the AA genome diploid Oryza species by analysis of MITEs (miniature inverted-transposable elements), Park et al. (2003a) demonstrated that the species from Asia, Africa and Australia formed separated lineages, congruent with the previous Restriction Fragment Length Polymorphism (RFLP) (Wang et al. 1992) and Amplification Fragment Length Polymorphism (AFLP) (Vos et al. 1995; Aggarwal et al. 1999) analyses.
Transposon display (TD) is a modification of the conventional AFLP technique that uses consensus sequence domain in transposons as primer binding sites in PCR reactions (Casa et al. 2000). Instead of the double-restriction digestions that are used in AFLP, TD utilizes single-restriction digested DNA fragments that are ligated with appropriate adaptors before amplification. Then, the single-restriction digested fragments are amplified with primers complementary to the transposon consensus site and adaptor. Using a Tourist MITE, Casa et al. (2000) demonstrated high allelic diversity in a segregating mapping population of maize. Wessler et al. (2001) also showed high allelic variation in MITE-TD in rice using a MITE family Olo element. Park et al. (2003a, b) demonstrated that Stowaway MITE-TD was suitable for studying genetic variation and species relationships in the genus Oryza. CACTA is a class 2 type transposon (Wicker et al. 2003) that is very abundant in most plant genomes and comprises 0.5% of the total transposons in rice (Turcotte et al. 2001). Based on the results from cloning and data mining from a 230 Mb of rice genome sequence, Wang et al. (2003) estimated that approximately 600–700 Rim2/Hippa CACTA transposons were present in the rice genome, suggesting that there would be several thousands of CACTA elements in the rice genome. Recently, Kwon et al. (2005) reported a Rim2/Hipa-TD protocol for a new genetic marker system in Oryza species. In the present study, we exploited the Rim2/Hipa-TD for studying genetic diversity and phylogenetic relationships in species of rice with AA genomic constitutions.
The seeds of the cultivated and wild rice accessions were kindly provided by the International Rice GeneBank Collection (IRGC), IRRI, Los Baňos, Philippines. Genomic DNAs for the cultivated accessions were obtained from the Genetic Resources Center, IRRI, Los Baňos, Philippines. The IRGC accession numbers and species types, variety names, and origins are shown in the Fig. 2. For the wild accessions, genomic DNA was extracted from young leaves using the protocol of Dellaporta et al. (1983).
Rim2/Hipa CACTA primer and adaptors were designed from the consensus sequences from the GenBank database using the basic information of He et al. (2000). The nucleotide sequences of the primers and adaptors are in Table 1. The PCR amplification conditions and the electrophoresis conditions were as reported before (Kwon et al. 2005).
 View Details | Table 1. Primer sequences of the Rim2/Hipa CACTA-TD analysis |
Each amplified fragment was treated as a unit character and scored as a binary code 1 and 0 for presence and absence, respectively. Only the prominent bands were scored for data reliability. Genetic similarity between accessions was calculated with the algorithm of Nei and Li (1979); Sxy = 2Nxy/(Nx+Ny), where Nxy represents the number of shared bands between two accessions x and y. Nx and Ny represent the total number of bands in accession x and accession y, respectively. The phylogenetic dendrogram was constructed by UPGMA (Unweighted Pair-group Method with Arithmetic Average) in the NTSYS-pc program (Rohlf 2000) and bootstrapping was done using WINBOOT (Yap and Nelson 1996). Genetic variation and genetic distances (GDs = 1 – Sxy) within species and between species were calculated using Microsoft Excel PC-program from the NTSYS data set. The degree of CACTA-TD polymorphisms was quantified using Shannons information index for phenotypic diversity (Browman et al. 1971): Hs = –ΣPi ln Pi, where Hs is the phenotypic diversity value of a group S and Pi is the frequency of the ith CACTA-TD phenotype (band) in a specific group S.
Rim2/Hipa-TD revealed a widespread distribution of the Rim2/Hipa motif among Oryza species with the AA genomic constitution (Fig. 1). Rim2/Hipa-TD produced many distinctively amplified fragments ranging in size from 100 bp to 700 bp (Fig. 1). As the amplified fragments were highly polymorphic between accessions in a species, the number of recordable fragments in a given population of each species varied from 67 (-CG primer in O. glaberrima) to 229 (-CG primer in O. rufipogon) (Table 2).
 View Details | Fig. 1. Rim2/Hipa-TD profile of the AA diploid Oryza species using the anchor primer KRMP-AG. |
 View Details | Table 2. Distribution of Rim2/Hipa-TD markers detected for different primers in Oryza species with AA genome |
Table 2 shows the number of polymorphic fragments per amplified number of fragments in each species, revealing a high degree of polymorphism. Table 2 shows the average genetic diversity and percentage of polymorphic markers in each species. The African taxa, O. glaberrima and O. barthii exhibited lower levels of polymorphisms than the Asian taxa, O. sativa, O. rufipogon, and O. nivara. However, O. longistaminata, another African taxon, showed a degree of polymorphism as high as the Asian taxa. The Latin American taxon, O. glumaepatula, and the Australian taxon, O. meridionalis, showed levels of polymorphisms that were intermediate to those of the Asian and African taxa. The lowest polymorphism was observed in O. glaberrima (32.1%) while the highest polymorphism was observed in O. rufipogon (95.7%). Among the geographically grouped taxa, wild species showed higher levels of genetic polymorphism than their counterpart domesticated species. As previously observed with isozyme analysis (Morishima and Oka 1970), the perennial species, O. rufipogon in Asian taxa and O. longistaminata in African taxa, exhibited higher genetic variations than the annual species, O. sativa and O. nivara in Asian taxa, and O. glaberrima and O. barthii in African taxa. Within the O. sativa complex, the Japonica type accessions showed slightly lower levels of polymorphism than the Indica type accessions as revealed by Shannons index of diversity in Table 3.
 View Details | Table 3. The average genetic diversity and percentage of polymorphic markers detected with eight primer combinations among the accessions within Oryza species |
Among the Oryza species with AA genomic constitutions, O. glaberrima and O. barthii showed the closest genetic affinity with a distance value of 0.213 (Table 4). The least genetic affinity was observed between O. meridionalis and the O. sativa complex species – a distance value of 0.752 with O. sativa spp Japonica and 0.757 with O. sativa spp Indica. The American origin, O. glumaepatula showed closer affinities to the African taxa, O. glaberrima (0.622) and O. barthii (0.616) than to the species of the Asian taxa. The Australian taxon, O. meridionalis, showed most distant affinity to other species, and the genetic affinities of O. meridionalis with Asian taxa and African taxa were not significantly different. The genetic affinity between the two ecotypes, Japonica and Indica of O. sativa was 0.536, which was higher than the genetic affinities between the O. sativa spp. Indica and O. rufipogon (0.483) and between O. sativa spp. Indica and O. nivara (0.479), respectively. The genetic affinities between O. sativa spp. Japonica and O. rufipogon and O. nivara were 0.513 and 0.539, respectively.
 View Details | Table 4. Distance matrix showing between- species genetic distances for Oryza species with AA genome |
The phylogenetic tree identified three major groups at the genetic similarity level of 0.409 (Fig. 2). The first group consisted of three Asian taxa: O. sativa, O. rufipogon and O. nivara. The second group comprised the three African taxa, O. glaberrima, O. barthii, O. longistaminata, and an American taxon, O. glumaepatula. The third group contained the Australian taxon, O. meridionalis.
 View Details | Fig. 2. Phylogenetic dendrogram of the AA diploid Oryza species based on the Rim2/Hipa-TD markers. At the right column, the P and A denote perennial and annual, respectively. |
The cluster consisting of the Asian taxa was subdivided into two subclusters, each with a genetic similarity of 0.525. The first subcluster consisted of 19 O. sativa spp. Japonica accessions and three wild accessions derived from Taiwan - two O. rufipogon accessions (acc. 100692, 100678) and one O. nivara accession (acc. 100593). While the O. rufipogon accession 100692 clustered closely with accessions from Korea, Japan, China and Iran, the O. rufipogon accession 100678 and the O. nivara accession 100593 formed an out-group cluster within this subcluster. Three aromatic rice accessions formed a subcluster with a slightly looser relationship with the Japonica accessions. In the second subcluster of the Asian taxa, Aus/Boro type accessions formed a separate subcluster from the O. sativa spp. Indica types. In terms of photoperiod response, photoperiod insensitive accessions were classified as Aus/Boro types while photoperiod sensitive accessions were classified as Aman groups of the O. sativa spp. Indica (Oka 1988). In our analysis, we classified the Aman type as just Indica types and the Aus/Boro types as Indica Aus/Boro types. In the subcluster comprised mostly of O. sativa spp. Indica accessions, four O. rufipogon accessions (acc.100647, 100657, 101193, 100898) grouped together. In the subcluster of Aus/Boro types, three O. nivara accessions (acc. 103835, 103836, 100967) and one O. rufipogon accession (acc. 100926) were closely grouped and two O. nivara accessions (acc. 100918, 104405) formed a subcluster slightly more distant from the others. In the second subcluster of the Asian taxa, two O. rufipogon (acc.100916, 101173) and one O. nivara (acc. 105701) accessions formed an out-group subcluster at the genetic similarity of 0.531. These three wild accessions could be distant ancestors for both Aman and Aus/Boro Indica types of O. sativa. The two species O. rufipogon and O. nivara did not form distinct subclusters of their own, and always occurred with other species in the Asian taxa cluster.
The African taxa, O. glaberrima, O. barthii, and O. longistaminata formed a cluster with the American taxon, O. glumaepatula. The O. glaberrima accessions closely clustered with O. barthii, supporting the claim that the latter is a direct ancestor of the African cultigen, O. glaberrima (Oka 1988). An accession of O. barthii (acc. 100122), out grouped from other O. barthii accessions to cluster with accessions of O. glaberrima, may require further classification. The most distant clustering of the Austrailan taxon, O. meridionalis, in our results agrees with the results of other studies using RFLP (Wang et al. 1992), AFLP (Aggarwal et al. 1999), and RAPD analysis (Buso et al. 2001).
Rim2/Hipa-TD analysis revealed that the Rim2/Hipa CACTA sequence motif is widely distributed in the genomes of all the Oryza species with the AA genomic constitution. The polymorphism, as detected by the Rim2/Hipa-TD, was very high within and between species among the AA diploid Oryza species, and permitted the study of both genetic diversity and species relationships. In our analysis, the Asian taxa showed higher levels of polymorphisms within and between species compared to taxa from Africa, America and Australia. This was consistent with other reports by Aggarwal et al. (1999) using AFLP, Park et al. (2003a) using MITE-TD, and by Ishii et al. (2001) using microsatellites. The presence of low genetic variation within a species implies that these species may have undergone a relatively recent differentiation. The wider geographical distribution of the Asian taxa compared to others may also have contributed to the higher genetic variation (Vaughan 1994). The relatively low genetic variation of the cultivated species, O. sativa and O. glabberimma, compared to the wild species, O. rufipogon, O. nivara, and O. barthii, was also noted by other studies using RFLP (Wang et al. 1992), AFLP (Aggarwal et al. 1999), and ISSR (Joshi et al. 2000) analyses. Based on the analysis of 122 accessions of O. rufipogon and 75 accessions of O. sativa with 44 RFLP markers, Sun et al. (2001) proposed that natural and human selection have caused the loss of many alleles, leading to the low genetic diversity of the cultigens during the course of domestication. We observed the highest level of polymorphism for O. rufipogon among the AA diploid species. This may be attributable to its wide distribution as well as the perennial nature associated with allogamy, as hypothesized by Morishima (2001). Furthermore, the higher level of polymorphism for O. sativa spp. Indica types than for O. sativa spp. Japonica types may be attributed to the fact that the former types are more widely distributed than the latter (Vaughan 1994).
Our phylogenetic dendrogram showed clear distinctions among the AA diploid species according to their geographic origins except for O. glumaepatula, which occurs in Latin America, but is grouped with the cluster of African taxa. O. glumaepatula, and was once considered a subtype of O. rufipogon (Vaughan 1994). However, many studies with molecular markers, including ours, showed that O. glumaepatula is more closely related to the African taxa, O. barthii and O. glabberima, than to the Asian taxon, O. rufipogon (Wang et al. 1992; Aggrawal et al. 1999; Buso et al. 2001). Likewise, Cheng et al. (2002) noted that the O. glumaepatula accessions were more closely related to the O. barthii or O. glaberrima accessions than to the O. rufipogon or O. sativa accessions, based on the analysis of SINE insertion polymorphisms. These observations indicate a clear separation of O. glumaepatula from O. rufipogon. The low bootstrap values in some nodes in the dendrogram might have derived from the chromosomal distribution of the CACTA-TD markers. The CACTA-TD markers were located mostly around the pericentric regions except in a few chromosomes where they are rather evenly distributed (manuscript in preparation). Since centromeric repetitive sequences are hypervariable, the unequal distribution of the CACTA-TD markers along the chromosomes may contribute to the low bootstrap values in some nodes.
The close genetic distance between O. sativa and the species O. rufipogon and O. nivara implies that the latter two are the likely ancestors of O. sativa (Oka 1988). The classification of O. rufipogon and O. nivara is in dispute (Morishima 2001; Cheng et al. 2003; Zhu and Ge 2005a, b). The mixed placement among the accessions of O. rufipogon and O. nivara in the dendrogram was also reported by Zhu and Ge (2005a, b) based on the nucleotide variations of five nuclear loci (Adh1, CatA, Lsh1, Waxy and Os9971), which may be supported by the fact that the two species are sympatrically distributed in southern Asia and show continual variations in their morphologies (Oka 1988; Vaughan 1994). However, analysis of insertion polymorphism of SINEs by Cheng et al. (2003) revealed that perennial and annual strains of O. rufipogon were clearly divided into one annual group and three perennial groups. Although the phenetic dendrogram in the current report, and the results of Zhu and Ge (2005a, b) may both indicate no distinct genetic differentiation between O. rufipogon and O. nivara, studies with more accessions of Asian AA diploid Oryza species are needed for a clear understanding of the differentiation between these two species.
With Stowaway MITE-TD analysis (Park et al. 2003a), the accessions of the O. rufipogon and O. nivara were also mixed with O. sativa accessions in the phenetic dendrogram. However, the small number accessions of O. sativa (9 accessions) from limited origins as Japonica types from Korea and Indica types from IRRI might have limited the ability to trace the origins of the O. sativa subspecies in their study. Therefore, we selected a wider range of O. sativa accessions from diverse geographical origins for this study. The accessions of O. sativa in our analysis were assorted as a basic set representing O. sativa germplasm in IRRI on the basis of their agronomic characteristics and variety group determined by isozyme analysis. The accessions that closely clustered with Japonica types (O. rufipogon acc. 100692, O. rufipogon acc. 100678, O. nivara acc. 100593) might be the Japonica types of O. rufipogon and O. nivara, respectively. These three wild accessions also clustered with Japonica types in the MITE-TD analysis by Park et al. (2003a). Garris et al. (2005), using extensive SSR analysis, have included the aromatic accessions with those of the Japonica group. The four O. rufipogon accessions (acc. 100647, 100657, 101193, 100898) closely clustered with O. sativa spp. Indica types in the MITE-TD analysis by Park et al. (2003a), suggesting that they are Indica type accessions of O. rufipogon. The four O. nivara accessions (acc. 103835, 103826, 100967, 100918) and an O. rufipogon accession (acc. 100926), that clustered with O. sativa spp. Indica Aus/Boro accessions in our analysis, loosely clustered with O. sativa spp. Indica accessions in MITE-TD (Park et al. 2003a). The Aus/Boro types are photoperiod insensitive Indica types while the Aman types are photoperiod sensitive Indica types (Oka 1988). Therefore, the Aus/Boro types might have originated from distinct accessions of O. nivara and O. rufipogon compared to the photoperiod sensitive Aman Indica types. An accession of O. rufipogon (acc. 100916) and one accession of O. nivara (acc. 105701), which joined to the Indica subgroups lastly in our analysis, showed loosely clustered with both Japonica and Indica clusters in MITE-TD analysis (Park et al. 2003a). On the basis of these results, we were able to conclude that the O. sativa spp. Japonica types and O. sativa spp. Indica types might have originated from different wild accessions; this supports the diphyletic theory of the origin of the O. sativa complex (Oka and Morishima 1982; Second 1982; Glaszmann 1987; Wang et al. 1992; Cheng et al. 2003; Zhu and Ge 2005a, b).
Among African taxa, the close genetic distance between O. glabberima and O. barthii indicates that of the two species, O. barthii is the ancestral species (Oka 1988). The differentiation of the African taxa appears to be monophyletic compared to the diphyletic origin of O. sativa, since each species showed distinct clusters in our analysis. As mentioned earlier, the closer relatedness of O. glumaepatula to O. glabberima and O. barthtii than to O. longistaminata is interesting. O. longistaminata is perennial and propagates by rhizomes (Oka 1988), traits that might be related to the highest level of polymorphism among the three African taxa observed in our analysis. Morishima (2001) proposed four directions of differentiation in the primary gene pool of Oryza species with AA genomic constitutions: (i) from wild to cultivated, (ii) from perennial to annual, (iii) geographical differentiation in wild races, and (iv) varietal differentiation. Although the number of accessions used in our analysis may not be enough to draw definitive conclusions, differentiation from wild to cultivated types hold true for both the Asian and African taxa. Differentiation from perennial to annual types may not hold true for Asian taxa because the O. rufipogon (perennial) and O. nivara (annual) occur in each subcluster in our phylogenetic dendrogram analysis, but does hold true for the African taxa. Based on our analysis, O. glabberima differentiated most recently from O. barthii, while both annuals O. glabberima, O. barthii and perennial O. glumaepatula originated from common source. O. longistaminata and the other three species could have originated from a common source, although O. longistaminata diverged earliest among those.
In summary, the Rim2/Hipa-TD was shown to be a useful marker system for studying genetic diversity and species relationships among Oryza species with diploid AA genomic constitutions. The amplification profiles from the Rim2/Hipa-TD were robust and reproducible between different brands of Thermocyclers, and using different sources of Taq DNA polymerases (Kwon et al. 2005). In the current study, except for O. glumaepatula the AA diploid species were distinctly separated from each other according to their geographical origins. Higher genetic variations were observed in wild and perennial species, compared with those of cultivated and annual species. Finally, the Asian taxa showed higher genetic variations than did the African taxa.
We thank the Ministry of Science and Technology, Republic of Korea, for funding this research to NSK through the Crop Functional Genomics Center (Project Number CG3122), Generation Challenging Program for providing to IRRI and IRRI-RDA collaboration between KLM and YJP. Thanks are also extended to Dr. R. Bhambhani for critical reading and comments for the manuscript.
|
Index