Distribution of mitochondrial plasmid-like DNA in cultivated rice (Oryza sativa L.) and its relationship with varietal groups

Genetic variation in the chloroplast genome suggests multiple domestication of cultivated Asian rice (Oryza sativa L.)

Two hundred and seventy-five accessions of cultivated Asian rice and 44 accessions of AA genome Oryza species were classified into 8 chloroplast (cp) genome types (A-H) based on insertion-deletion events at 3 regions (8K, 57K, and 76K) of the cp genome. The ancestral cp genome type was determined according to the frequency of occurrence in Oryza species and the likely evolution of the variable 57K region of the cp genome. When 2 nucleotide substitutions (AA or TT) were taken into account, these 8 cp types were subdivided into 11 cp types. Most indica cultivars had 1 of 3 cp genome types that were also identified in the wild relatives of rice, O. nivara and O. rufipogon, suggesting that the 3 indica cp types had evolved from distinct gene pools of the O. rufipogon - O. nivara complex. The majority of japonica cultivars had 1 of 3 different cp genome types. One of these 3 was identified in O. rufipogon, suggesting that at least 1 japonica type is derived from O. rufipogon with the same cp genome type. These results provide evidence to support a polyphyletic origin of cultivated Asian rice from at least 4 principal lineages in the O. rufipogon - O. nivara complex.

Polymorphic distribution and molecular diversification of mitochondrial plasmid-like DNAs in the genus Oryza

Four kinds of circular plasmid-like DNA, designated B1, B2, B3 and B4, have been found in the mitochondria of rice (Oryza sativa L.). We analyzed the distribution of families of plasmid-like DNAs homologous to those of O. sativa in 40 strains of the genus Oryza with AA, BB, BBCC, CC, CCDD and EE genomes. Plasmid-like DNAs were observed only strains having AA, CC and CCDD genomes. The distribution patterns of strains with AA genome were highly polymorphic. We amplified the plasmid-like DNAs from strains with the AA genome by PCR and examined restriction fragments length polymorphisms (RFLPs). RFLPs were detected among families of plasmid-like DNA amplified from different strains. This result indicated that some mutations, such as base substitutions and the insertion or deletion of a small fragment of DNA, had occurred and had accumulated during the differentiation of strains with an AA genome.

Homologous recombination mediated by two palindromic repeated sequences in the mitochondrial genome of Oryza

Palindromic repeated sequences (PRSs) are distributed in at least ten regions of the mitochondrial (mt) genome of rice and are, apparently, mobile. In the present study, we examined the possibility of homologous recombination via some PRSs during the course of evolution of Oryza. We first performed Southern hybridization of the DNA from 11 species (18 strains) of Oryza in order to identify the distribution of PRSs in the mitochondrial genome of Oryza. The hybridization patterns revealed genome type-specific and/or species-specific variations. We speculated that homologous recombination via some PRSs might have made a contribution to such variations. After subsequent polymerase chain reaction, Southern hybridization and sequencing, we concluded that homologous recombination mediated by two PRSs occurred in the mtDNA of Oryza after divergence of the BB genome type and the other genome types of Oryza. Evidence was obtained that some PRSs were involved in both insertion and recombination events during the evolution of Oryza. Our results indicate, therefore, that PRSs have contributed considerably to the polymorphism of Oryza mtDNAs.

Enigmatic double-stranded RNA in Japonica rice

We have found a linear, 16 kb, double-stranded RNA (dsRNA) in symptomless Japonica rice (Oryza sativa L.) that is not found in Indica rice (Oryza sativa L.). The dsRNA was detected in every tissue and at every developmental stage, and its copy number was approximately constant (about 20 copies/cell). Double-stranded RNA was also detected in two strains of Oryza rufipogon (an ancestor of O. sativa). Hybridization experiments indicated that the dsRNA of O. rufipogon was homologous but not identical to that of O. sativa. The sequence of about 13.2 kb of the dsRNA was determined and two open reading frames (ORFs) were found. The larger ORF (ORF B) was more than 12,351 nucleotides long and encoded more than 4,117 amino acid residues.

Comparative study on the chloroplast, mitochondrial and nuclear genome differentiation in two cultivated rice species, Oryza sativa and O. glaberrima, by RFLP analyses

Restriction fragment length polymorphisms of chloroplast (ct), mitochondrial (mt) and nuclear DNA were investigated using eight cultivars of Oryza sativa and two cultivars of O. glaberrima. Relative variability in the nuclear and cytoplasmic genomes was estimated by a common measure, genetic distance. Based on the average genetic distances among ten cultivars for each genome, the evolutionary variabilities of the mitochondrial and nuclear genomes were found to be almost the same, whereas the variability of the chloroplast genome was less than half that of the other two genomes. Cluster analyses on ct and mt DNA variations revealed that chloroplast and mitochondrial genomes were conservative within a taxon and that their differentiations were well-paralleled with respect to each other. For nuclear DNA variation, an array of different degrees of differentiation was observed in O. sativa, in contrast with little variation in O. glaberrima. As a whole, differentiation between O. sativa and O. glaberrima was clearly observed in all three genomes. In O. sativa, no notable difference was found between the cultivars 'Japonica' and 'Javanica', whereas a large differentiation was noticed between 'Japonica' (including 'Javanica') and 'Indica'. In all three genomes, the average genetic distances within 'Indica' were much larger than those within 'Japonica' (including 'Javanica'), and almost similar between 'Japonica' (including 'Javanica') and 'Indica'. These facts indicate that differentiation in O. sativa was due mainly to 'Indica'.

Indica-Japonica differentiation of rice cultivars viewed from variations in key characters and isozymes, with special reference to landraces from the Himalayan hilly areas

To assess the extent of differentiation into the Indica and Japonica types of Asian rice in its diversity center, we investigated landraces collected from the hilly areas of Nepal, Sikkim and Assam. We examined variations in four key characters and six isozyme loci known to be diagnostic for classifying Indica and Japonica types, and compared the results with those from a control set of rice cultivars representing the whole of Asia. The hill cultivars showed a high level of genetic diversity in key characters as well as in isozymes. A marked feature found in their character variation was the occurrence of various atypical cultivars that were intermediate between the Indica and Japonica types. With respect to isozymes, however, the hill cultivars could be classified mostly as either Indicas or Japonicas, although the patterns of allelic association were more random than in the control cultivars. Indica-Japonica variation in key characters and in isozymes corresponded well with each other in the controls, but not in the hill cultivars. This means that nonrandom association in characters as well as in genes (gametic disequilibrium) is not fully developed in the diversity center. Populations of hill cultivars were highly polymorphic genetically, but did not show a trend to Indica-Japonica differentiation within their populations. The process of Indica-Japonica differentiation is discussed in view of these observations.

Analysis of nuclear sequences homologous to the B4 plasmid-like DNA of rice mitochondria; evidence for sequence transfer from mitochondria to nuclei

Nuclear sequences homologous to the plasmid-like DNA, B4, were analyzed in the Japonica rice variety, Fujiminori. Homologous sequences existed at several positions in the nuclear genome, but each contained only a portion of the B4 sequence. It was impossible to reconstruct the entire sequence of B4 even by collating all the homologous sequences. Overlaps between some of the B4 sequences present in the nuclear genome resulted in parts of the sequence being represented more than once. These features indicate that nuclear sequences homologous to B4 are not the origin of B4 and that they have been transferred from mitochondria and integrated into the nuclear genome. Five other foreign sequences originating in the chloroplast or mitochondrial genome were found within 1 kb of the B4-homologous sequences. Structural analysis is consistent with the hypothesis that the DNA sequences were transferred via RNA.

Linkage analysis of the nuclear homologues of mitochondrial plasmid-like DNAs in rice

A genetical study on the nucleotide sequences of the nuclear DNAs which share homology with rice mitochondrial plasmid-like DNAs, B1, B2, B3 and B4 was carried out. Restriction fragments of the nuclear DNAs hybridized with these plasmid-like DNAs showed polymorphisms in their length between Indica and Japonica rice cultivars. The hybridized signals found specifically in Indica or Japonica cultivars segregated in the F2 population derived from a cross between these two subspecies. The observed ratio of the nuclear homologues in the F2 population demonstrated that they were transmitted according to the Mendelian inheritance. The co-segregation of homologues was examined and the linkage was detected between the B1-nuclear homologue of Japonica and the B4-nuclear homologue of Indica, and also between the nuclear homologues of B2 and B3 of Indica. The linkage between the B1-nuclear homologue of Japonica and the B4-nuclear homologue of Indica was conserved in the different rice cultivars.

RFLP analysis of nuclear DNAs homologous with mitochondrial plasmid-like DNAs in cultivated rice

B1 and B2 are small, circular, mitochondrial plasmid-like DNAs found in male-sterile cytoplasm (cms-Bo) of rice. In this study, nuclear sequences homologous to these DNAs were investigated among a number of rice cultivars. Several copies of nuclear B1-and B2-homologous sequences were detected in all examined cultivars, regardless of the presence or absence of the B1 and B2 DNAs in mitochondria, indicating that the existence of the B1- and B2-homologous sequences in the rice nuclear genome was widespread. A restriction fragment length polymorphism (RFLP) was detected for both sequences, and we propose that these DNAs could be useful RFLP markers for the rice nuclear genome. To analyze these nuclear homologues genetically, segregation analysis of the RFLP was carried out in the F2 progenies of an Indica-Japonica rice hybrid. Of the B1 homologues, there were two nonallelic fragments, one specific to the Indica parent and the other to the Japonica. These results indicate that the B1 and B2 homologues were dispersed in the nuclear genome. The integration of B1-homologous DNA into the nuclear DNA may have occurred independently after sexual isolation of the Indica and Japonica rice varietal groups, or a intranuclear transposition of these sequences took place during the process of rice differentiation into the varietal groups.

A chimeric gene containing the 5' portion of atp6 is associated with cytoplasmic male-sterility of rice

Three ATPase subunit 6 (atp6) genes of rice mitochondria were isolated, one from normal and two from cms-Bo male-sterile cytoplasms, in order to determine whether the extra atp6 gene in cms-Bo rice plays a role in cytoplasmic male-sterility (CMS). The nucleotide sequences of all three genes were determined and analysis showed a chimeric atp6 gene (urf-rmc) as well as a normal atp6 gene in cms-Bo cytoplasm, but only the normal atp6 gene in normal cytoplasm. The urf-rmc gene is completely homologous to the normal atp6 gene from at least position -426 in the 5' flanking region to position +511 downstream from the initiation codon ATG: however, the following downstream sequence shows no homology with the normal rice atp6 gene, or any other reported sequence. Introduction of the restorer of fertility gene altered transcription of the urf-rmc gene but not the atp6 gene, indicating participation of the chimeric gene in the expression of CMS. Southern blot analysis showed that the urf-rmc gene was generated by an intramolecular recombination event in mitochondrial DNA, and the homologous recombination point between the atp6 gene and the opposite ancestral sequence was identified as 5'-TTCCCTC-3'.