Plant regeneration methods for rapid generation of a large scale Ds transposant population in rice

To mutagenize rice genomes, a two-element system is utilized. This system comprises an immobile Ac element driven by the CaMV 35S promoter, and a gene trap Ds carrying a partial intron with alternative splice acceptors fused to the GUS coding region. Rapid, large-scale generation of a Ds transposant population was achieved using a plant regeneration procedure involving the tissue culture of seed-derived calli carrying Ac and Ds elements. During tissue cultures, Ds mobility accompanies changes in methylation patterns of a terminal region of Ds, where over 70% of plants contained independent Ds insertions. In the transposon population, around 12% of plants expressed GUS at the early seedling stage. A flanking-sequence-tag (FST) database has been established by cloning over 19,968 Ds insertion sites and the Ds map shows relatively uniform distribution across the rice chromosomes.

A Ds-insertion mutant of OSH6 (Oryza sativa Homeobox 6) exhibits outgrowth of vestigial leaf-like structures, bracts, in rice

OSH6 (Oryza sativa Homeobox6) is an ortholog of lg3 (Liguleless3) in maize. We generated a novel allele, termed OSH6-Ds, by inserting a defective Ds element into the third exon of OSH6, which resulted in a truncated OSH6 mRNA. The truncated mRNA was expressed ectopically in leaf tissues and encoded the N-terminal region of OSH6, which includes the KNOX1 and partial KNOX2 subdomains. This recessive mutant showed outgrowth of bracts or produced leaves at the basal node of the panicle. These phenotypes distinguished it from the OSH6 transgene whose ectopic expression led to a "blade to sheath transformation" phenotype at the midrib region of leaves, similar to that seen in dominant Lg3 mutants. Expression of a similar truncated OSH6 cDNA from the 35S promoter (35S::DeltaOSH6) confirmed that the ectopic expression of this product was responsible for the aberrant bract development. These data suggest that OSH6-Ds interferes with a developmental mechanism involved in bract differentiation, especially at the basal nodes of panicles.

Analysis of gene-trap Ds rice populations in Korea

Insertional mutagen-mediated gene tagging populations have been essential resources for analyzing the function of plant genes. In rice, maize transposable elements have been successfully utilized to produce transposant populations. However, many generations and substantial field space are required to obtain a sufficiently sized transposant population. In rice, the japonica and indica subspecies are phenotypically and genetically divergent. Here, callus cultures with seeds carrying Ac and Ds were used to produce 89,700 lines of Dongjin, a japonica cultivar, and 6,200 lines of MGRI079, whose genome is composed of a mixture of the genetic backgrounds of japonica and indica. Of the more than 3,000 lines examined, 67% had Ds elements. Among the Ds-carrying lines, 81% of Dongjin and 63% of MGRI079 contained transposed Ds, with an average of around 2.0 copies. By examining more than 15,000 lines, it was found that 12% expressed the reporter gene GUS during the early-seedling stage. GUS was expressed in root hairs and crown root initials at estimated frequencies of 0.78% and 0.34%, respectively. The 5,271 analyzed Ds loci were found to be randomly distributed over all of the rice chromosomes.

OsCSLD1, a cellulose synthase-like D1 gene, is required for root hair morphogenesis in rice

Root hairs are long tubular outgrowths that form on the surface of specialized epidermal cells. They are required for nutrient and water uptake and interact with the soil microflora. Here we show that the Oryza sativa cellulose synthase-like D1 (OsCSLD1) gene is required for root hair development, as rice (Oryza sativa) mutants that lack OsCSLD1 function develop abnormal root hairs. In these mutants, while hair development is initiated normally, the hairs elongate less than the wild-type hairs and they have kinks and swellings along their length. Because the csld1 mutants develop the same density and number of root hairs along their seminal root as the wild-type plants, we propose that OsCSLD1 function is required for hair elongation but not initiation. Both gene trap expression pattern and in situ hybridization analyses indicate that OsCSLD1 is expressed in only root hair cells. Furthermore, OsCSLD1 is the only member of the four rice CSLD genes that shows root-specific expression. Given that the Arabidopsis (Arabidopsis thaliana) gene KOJAK/AtCSLD3 is required for root hair elongation and is expressed in the root hair, it appears that OsCSLD1 may be the functional ortholog of KOJAK/AtCSLD3 and that these two genes represent the root hair-specific members of this family of proteins. Thus, at least part of the mechanism of root hair morphogenesis in Arabidopsis is conserved in rice.

A phosphate starvation-induced acid phosphatase from Oryza sativa: phosphate regulation and transgenic expression

A phosphate starvation-induced acid phosphatase cDNA was cloned from the rice, Oryza sativa. The cDNA encoding O. sativa acid phosphatase (OsACP1) has 1100 bp with an open reading frame of 274 amino acid residues. The deduced amino acid sequence of OsACP1 cDNA showed 53% identity to tomato acid phosphatase and 46-50% identity to several other plant phosphatases. OsACP1 expression was up-regulated in the rice plant and in cell culture in the absence of phosphate (Pi). The induced expression of OsACP1 was a specific response to Pi starvation, and was not affected by the deprivation of other nutrients. OsACP1 expression was responsive to the level of Pi supply, with transcripts of OsACP1 being abundant in Pi-deprived root. The OsACP1 cDNA was expressed as a 30 kDa polypeptide in baculovirus-infected insect Sf9 cells. In addition, the OsACP1 gene was introduced into Arabidopsis via Agrobacterium-mediated transformation. Functional expression of the OsACP1 gene in the transgenic Arabidopsis lines was confirmed by Northern blot and Western blot analyses, as well as phosphatase activity assays. These results suggest that the OsACP1 gene can be used to develop new transgenic dicotyledonous plants able to adapt to Pi-deficient conditions.

Analysis of intragenic Ds transpositions and excision events generating novel allelic variation in rice

Even though Ac/Ds gene-tagging systems have been established in many higher plants, maize is the only major plant in which short-distance transposition of Ac/Ds has been utilized to probe gene function. This study was performed to evaluate the efficiency of obtaining new alleles and functional revertants from Ds insertion loci in rice. By analyzing 1,580 plants and the progeny of selected lines, the insertion sites and orientations of Ds elements within 16 new heritable alleles of three rice loci were identified and characterized. Intragenic transposition was detected in both directions from the original insertion sites. The closest interval was 35 bp. Three of the alleles had two Ds elements in cis configuration in the same transcription units. We also analyzed the excision footprints of intragenic and extragenic transpositions in Ds-inserted alleles at 5 loci. The 134 footprints obtained from different plants revealed predominant patterns. Ds excision at each locus left a predominant footprint at frequencies of 30-75%. Overall, 66% of the footprints were 7-bp additions. In addition, 16% of the excisions left 0-, 3-, 6-, and 9-bp additions with the potential of conserving reading frame.

Mapping of quantitative trait loci for salt tolerance at the seedling stage in rice

Salt tolerance was evaluated at the young seedling stage of rice (Oryza sativa L.) using recombinant inbred lines (MG RILs) from a cross between Milyang 23 (japonica/indica) and Gihobyeo (japonica). 22 of 164 MG RILs were classified as tolerant with visual scores of 3.5-5.0 in 0.7% NaCl. Interval mapping of QTLs related to salt tolerance was conducted on the basis of the visual scores at the young seedling stage. Two QTLs, qST1 and qST3, conferring salt tolerance, were detected on chromosome 1 and 3, respectively, and the total phenotypic variance explained by the two QTLs was 36.9% in the MG RIL population. qST1 was the major QTL explaining 27.8% of the total phenotypic variation. qST1 was flanked by Est12-RZ569A, and qST3 was flanked by RG179-RZ596. The detection of new QTLs associated with salt tolerance will provide important information for the functional analysis of rice salt tolerance.

Enhanced expression of a gene encoding a nucleoside diphosphate kinase 1 (OsNDPK1) in rice plants upon infection with bacterial pathogens

A cDNA library was constructed using mRNA extracted from rice leaves infected with Xanthomonas oryzae pv. oryzae (Xoo), a bacterial leaf blight pathogen, to isolate rice genes induced by Xoo infection. Subtractive hybridization and differential screening of the cDNA library led to the isolation of many induced genes including a nucleotide diphosphate kinase 1 (OsNDPK1) and a pathogenesis-related protein 1 (OsPR1) cDNA. Nucleoside diphosphate kinases (NDPKs) are key metabolic enzymes that maintain the balance between cellular ATP and other nucleoside triphosphates (NTPs). Three other OsNDPK genes (NP922751, OsNDPK2 and OsNDPK3) found in databases were obtained by RT-PCR. Three different programs for predicting subcellular targeting indicated that OsNDPK1 and NP922751 were non-organellar, OsNDPK2 plastidic, and OsNDPK3 mitochondrial. Only transcripts of OsNDPK1 accumulated strongly after infection with Xoo. When rice plants were infected with Burkholderia glumae, a bacterial grain/seedling rot pathogen, the pattern of expression of the rice NDPK genes was similar to that following infection with Xoo. OsNDPK1 gene expression was also strongly induced in response to exposure to salicylic acid, jasmonic acid, and abscisic acid, although the level of transcripts and their pattern of expression depended on the inducer.

Rapid, large-scale generation of Ds transposant lines and analysis of the Ds insertion sites in rice

Rapid, large-scale generation of a Ds transposant population was achieved using a regeneration procedure involving tissue culture of seed-derived calli carrying Ac and inactive Ds elements. In the F(2) progeny from genetic crosses between the same Ds and Ac starter lines, most of the crosses produced an independent germinal transposition frequency of 10-20%. Also, many Ds elements underwent immobilization even though Ac was expressed. By comparison, in a callus-derived regenerated population, over 70% of plants carried independent Ds insertions, indicating transposition early in callus formation. In the remaining population, the majority of plants carried only Ac. Most of the new Ds insertions were stably transmitted to a subsequent generation. An exceptionally high proportion of independent transposants in the regenerated population means that selection markers for transposed Ds and continual monitoring of Ac/Ds activities may not necessarily be required. By analyzing 1297 Ds-flanking DNA sequences, a genetic map of 1072 Ds insertion sites was developed. The map showed that Ds elements were transposed onto all of the rice chromosomes, with preference not only near donor sites (36%) but also on certain physically unlinked arms. Populations from both genetic crossing and tissue culture showed the same distribution patterns of Ds insertion sites. The information of these mapped Ds insertion sites was deposited in GenBank. Among them, 55% of Ds elements were on predicted open-reading frame (ORF) regions. Thus, we propose an optimal strategy for the rapid generation of a large population of Ds transposants in rice.

Rice mutant resources for gene discovery

With the completion of genomic sequencing of rice, rice has been firmly established as a model organism for both basic and applied research. The next challenge is to uncover the functions of genes predicted by sequence analysis. Considering the amount of effort and the diversity of disciplines required for functional analyses, extensive international collaboration is needed for this next goal. The aims of this review are to summarize the current status of rice mutant resources, key tools for functional analysis of genes, and our perspectives on how to accelerate rice gene discovery through collaboration.

Reprogramming of the activity of the activator/dissociation transposon family during plant regeneration in rice

Many aspects of epigenetic phenomena have been elucidated via studies of transposable elements. An active transposable element frequently loses its ability to mobilize and goes into an inactive state during development. In this study, we describe the cyclic activity of a maize transposable element dissociation (Ds) in rice. In rice genome, Ds undergoes the spontaneous loss of mobility. However, an inactive state of Ds can be changed into an active state during tissue culture. The recovery of mobility accompanies not only changes in the methylation patterns of the terminal region of Ds, but also alteration in the steady state level of the activator (Ac) mRNA that is expressed by a constitutive CaMV 35S promoter. Furthermore, the Ds-reactivation process is not random, but stage-specific during plantlet regeneration. Our findings have expanded previous observations on Ac reactivation in the tissue culture of maize.

Marker-assisted selection for identification of plant regeneration ability of seed-derived calli in rice (Oryza sativa L.)

Quantitative trait loci (QTL), associated with the ability of plant regeneration from seed-derived callus of rice, were mapped using a recombinant inbred (RI) population from Milyang 23/Gihobyeo. Each flanking marker, RZ474 and RZ575, tightly linked to two QTLs (qSGR-3-1 and qSGR-3-2) that are located on chromosome 3 was used in marker-assisted selection (MAS). These markers were tested on IR 36/MG RI036 (F3), Milyang 23/MG RI036 (F3), and forty-one rice cultivars. A restriction fragment length polymorphism (RFLP) marker, RZ575, that is located on chromosome 3 could effectively differentiate lines with high and poor regeneration ability, based on marker genotypes. This marker might be applicable for screening rice germplasms with high regeneration ability. Its introgression into elite lines might also be valuable in breeding programs to develop highly responsive genotypes to tissue culture.

Quantitative trait loci mapping associated with plant regeneration ability from seed derived calli in rice (Oryza sativa L.)

Quantitative trait loci (QTLs), which are associated with the ability of plant regeneration from seed derived calli, were detected using a recombinant inbred (RI) population from a cross between 'Milyang 23 (toingil)' and 'Gihobyeo (japonica)' in rice (Oryza sativa L.). A tongil type cultivar, 'Milyang 23', has a lower frequency of callus induction and plant regeneration than those of japonica 'Gihobyeo'. Transgressive segregations were observed for the callus induction rate and plant regeneration ability from seed derived calli of the RI population. An interval mapping analysis was used to identify the QTL controlling the plant regeneration ability. Two QTLs for the callus induction rate were detected on chromosomes 1 and 2, explaining the 10.9% total phenotypic variation. Four QTLs that are associated with the plant regeneration ability were located on chromosomes 2, 3, and 11, accounting for 25.7% of the total phenotypic variation.

Analysis of expressed sequence tags from Brassica rapa L. ssp. pekinensis

Non-redundant expressed sequence tags (ESTs) were generated from six different organs at various developmental stages of Chinese cabbage, Brassica rapa L. ssp. pekinensis. Of the 1,295 ESTs, 915 (71%) showed significantly high homology in nucleotide or deduced amino acid sequences with other sequences deposited in databases, while 380 did not show similarity to any sequences. Briefly, 598 ESTs matched with proteins of identified biological function, 177 with hypothetical proteins or non-annotated Arabidopsis genome sequences, and 140 with other ESTs. About 82% of the top-scored matching sequences were from Arabidopsis or Brassica, but overall 558 (43%) ESTs matched with Arabidopsis ESTs at the nucleotide sequence level. This observation strongly supports the idea that gene-expression profiles of Chinese cabbage differ from that of Arabidopsis, despite their genome structures being similar to each other. Moreover, sequence analyses of 21 Brassica ESTs revealed that their primary structure is different from those of corresponding annotated sequences of Arabidopsis genes. Our data suggest that direct prediction of Brassica gene expression pattern based on the information from Arabidopsis genome research has some limitations. Thus, information obtained from the Brassica EST study is useful not only for understanding of unique developmental processes of the plant, but also for the study of Arabidopsis genome structure.

Hormonal cross-talk between auxin and ethylene differentially regulates the expression of two members of the 1-aminocyclopropane-1-carboxylate oxidase gene family in rice (Oryza sativa L.)

Two cDNA clones, pOS-ACO2 and pOS-ACO3, encoding 1-aminocyclopropane-1-carboxylate (ACC) oxidase were isolated from rice seedling cDNA library. pOS-ACO3 is a 1,299 bp full-length clone encoding 321 amino acids (Mr=35.9 kDa), while pOS-ACO2 is 1,072 bp long and is a partial cDNA clone encoding 314 amino acids. These two deduced amino acid sequences share 70% identity, and display a high degree of sequence identity (72-92%) with previously isolated pOS-ACO1 of deepwater rice. The chromosomal location studies show that OS-ACO2 is positioned on the long arm of chromosome 9, while OS-ACO3 on the long arm of chromosome 2 of rice genome. A marked increase in the level of OS-ACO2 transcript was observed in IAA-treated etiolated rice seedlings, whereas the OS-ACO3 mRNA was greatly accumulated by ethylene treatment. Results of ethylene inhibitor studies indicated that auxin promotion of the OS-ACO2 transcription was not mediated through the action of auxin-induced ethylene. Thus, it appears that there are two groups of ACC oxidase transcripts in rice plants, either auxin-induced or ethylene-induced. The auxin-induced OS-ACO2 expression was partially inhibited by ethylene, while ethylene induction of OS-ACO3 transcription was completely blocked by auxin. These results indicate that the expression of ACC oxidase genes is regulated by complex hormonal networks in a gene specific manner in rice seedlings. Okadaic acid, a potent inhibitor of protein phosphatase, effectively suppressed the IAA induction of OS-ACO2 expression, suggesting that protein dephosphorylation plays a role in the induction of ACC oxidase by auxin. A scheme of the multiple regulatory pathways for the expression of ACC oxidase gene family by auxin, ethylene and protein phosphatase is presented.

Structure and expression of the AWI 31 gene specifically induced by wounding in Arabidopsis thaliana

Differential screening of an Arabidopsis cDNA library constructed from the plant tissues harvested 1 h after wounding resulted in the isolation of wound-inducible cDNA clones (Kim et al., 1994). The cDNA clones could be broadly classified into two groups according to the expression time of their transcripts. Nine clones from the 10 different wound-inducible cDNAs were rapidly induced, reaching a maximum level in approximately 1-1.5 h and then were progressively reduced after wounding. The cDNA clone AWI 31 showed steady accumulation of the transcripts and reached the maximum value at a later time point of 2.5 h and then started to decline. The corresponding gene of the AWI 31 in which the coding region was interrupted by an intron, had an open reading frame that predicted a protein of 386 amino acids. However, the gene product did not show any significant homology to other known proteins in the database. Northern hybridization study using the cDNA probe revealed that the gene was not regulated by other environmental stresses such as drought, high salt, low temperature, or a DPE herbicide treatment, indicating that the cDNA clone AWI 31 was specifically induced by wounding.

The semidwarf gene, sd-1, of rice (Oryza sativa L.). II. Molecular mapping and marker-assisted selection

To establish the location of the semidwarf gene, sd-1, the anthocyanin activator (A), purple node (Pn), purple auricle (Pau), and the isozyme locus, EstI-2, in relation to DNA markers on the molecular linkage map of rice, 20 RFLP markers, previously mapped to the central region of chromosome 1 (McCouch et al. 1988), were mapped onto an F2 population derived from the cross Taichung 65 (A,Pn,Pau)/Taichung 65 (sd-1). sd-1 and EstI-2 were determined to be linked most tightly to RFLP markers RG 109 and RG 220, which cosegregated with each other. The distance between these RFLP markers and sd-1 was estimated to be 0.8 cM, based on an observed recombination value of 0.8%. The order of genes and markers in this region of chromosome 1 was determined to be sd-1 - (EstI-2 - RG220 - RG109) - RG381 - A - Pn - Pau. To test the efficacy of selection for sd-1 based on these linked markers, 50-day-old F2 seedlings derived from another cross, Milyang 23/Gihobyeo, were analyzed for marker genotype. At this age, the semidwarf character could not be clearly detected based on phenotype. In addition, plant height was normally distributed in this population, making it difficult to unambiguously identify plants carrying sd-1. Thirteen seedlings homozygous for the sd-1-associated allele at EstI-2, RG220 and RG109, and 13 seedlings homozygous for the Sd-1-associated allele at all three marker loci were selected for further genetic analysis. At 20 days after heading, the culm lengths of these 26 plants were measured and the expected phenotype was confirmed in every case. These 26 plants were then selfed for four generations and F6 lines were again evaluated to determine whether any recombination among the three molecular markers, or between these markers and the sd-1 gene, could be detected. No recombinants were identified, confirming the tight linkage of these loci and the usefulness of genotypic selection for this recessive semidwarf character prior to the time when it can be evaluated based on phenotype.

The semidwarf gene, sd-1, of rice (Oryza sativa L.). I. Linkage with the esterase locus, Estl-2

The linkage relationship between the semidwarf gene (sd-1) and the isozyme locus EstI-2 was elucidated using segregating populations derived from crosses between several semidwarf testers and tall rice varieties. Bimodal distributions for culm length were observed in F2 populations of three cross combinations, including Shiokari/Shiokari (sd-1), Taichung 65 (A,Pn,Pau)/Taichung 65 (sd-1), and Milyang 23/Kasalath. Taking the valley of the distribution curves as the dividing point, two height classes were apparent with a segregation ratio of 3 tall∶1 short, demonstrating this character to be under the control of a single recessive gene. An inheritance study of esterase isozymes, based on isoelectric focusing (IEF), showed that the EstI-2 locus had two active allozymes of monomeric structure and one null form, which were designated "a", "b", and "n", respectively (Eun et al. 1990). Semidwarf testers such as Shiokari (sd-1), Taichung 65 (sd-1) and Milyang 23 have an active allozyme designated as EstI-2(aa), while the tall parents, Shiokari and Taichung 65 (A,Pn,Pau), have the active allozyme, EstI-2(bb), and Kasalath has a null form of the allozyme, EstI-2(nn). By dividing F2 populations based on EstI-2 allozyme patterns, culmlength distributions exhibited trimodal curves. Most of the short plants had the homozygous EstI-2(aa) pattern of the short parents, most of the tall plants had the homozygous pattern, EstI-2(bb) or EstI-2(nn), and most of the intermediate plants had the heterozygous EstI-2(ab) or EstI-2(an) banding pattern. Einkage analysis indicated that sd-1 and EstI-2 were tightly linked. These findings were also confirmed by segregation analyses in F3 progenies. No recombinants among 171 F3 families from the Shiokari/Shiokari (sd-1) combination, five recombinants among 267 F3 families from Taichung 65 (A,Pn,Pau)/Taichung 65(sd-1), and only two recombinants out of 237 F3 families from Milyang 23/Kasalath, were found. The recombination values were 0, 1.87 and 0.8%, respectively.