1. Introduction
Dwarf genes have been utilized extensively in plant breeding to improve lodging resistance. Their applications have been associated with increased yields, higher fertility, early maturity and high tillering capacity. The popularization of dwarf cultivars was a major factor in the success of the ‘Green Revolution’ in rice and wheat (Khush, Reference Khush2001; Hedden, Reference Hedden2003; Muangprom & Osborn, Reference Muangprom and Osborn2004). There are various reasons for the dwarf phenotype in plants, but defects in biosynthesis and the perception of gibberellin (GA) is the important determinant of plant height (Sasaki et al., Reference Sasaki, Ashikari, Ueguchi-Tanaka, Itoh, Nishimura, Swapan, Ishiyama, Saito, Kobayashi, Khush, Kitano and Matsuoka2002). GAs are essential endogenous regulators of plant growth and development that affect many aspects of a plant's life cycle, including seed germination, leaf expansion, stem elongation, flower initiation, and flower and fruit development (Harberd et al., Reference Harberd, King, Carol, Cowling, Peng and Richards1998). The important dwarf genes used in agriculture are mutations of genes in the GA biosynthesis or response pathways (Sakamoto et al., Reference Sakamoto, Miura, Itoh, Tatsumi, Ueguchi-Tanaka, Ishiyama, Kobayashi, Agrawa, Takeda, Abe, Miyao, Hirochika, Kitano, Ashikari and Matsuoka2004). Mutations of genes in the biosynthesis pathway cause GA deficiency and dwarf phenotypes, and exogenous GA application can restore the wild-type phenotype in these mutants (Phillips, Reference Phillips1998). The predominant dwarf gene in rice cultivars, semi-dwarf1 (sd1), affects the GA biosynthesis pathway (Monna et al., Reference Monna, Kitazawa, Yoshino, Susuki., Masuda, Maehara, Tanji, Sato, Nasu and Minobe2002; Sasaki et al., Reference Sasaki, Ashikari, Ueguchi-Tanaka, Itoh, Nishimura, Swapan, Ishiyama, Saito, Kobayashi, Khush, Kitano and Matsuoka2002; Spielmeyer et al., Reference Spielmeyer, Ellis and Chandler2002; Muangprom & Osborn, Reference Muangprom and Osborn2004).The semi-dwarf stature of d35 Tan-Ginbozu is caused by a defective early step of GA biosynthesis, which is catalysed by ent-kaurene oxidase (KO) (Itoh et al., Reference Itoh, Tatsumi, Sakamoto, Otomo, Toyomasu, Kitano, Ashikari, Ichihara and Matsuoka2004). Two GA 3β-hydroxylase genes, OsGA3oxl and OsGA3ox2, corresponding to the d18 locus, encoded proteins showing 3β-hydroxylase activity for the steps GA20 to GA1, GA5 to GA3, GA44 to GA38, and GA9 to GA4 (Itoh et al., Reference Itoh, Ueguchi-Tanaka and Yutaka2001).
Dwarf mutants in the GA response pathway display a similar phenotype to the GA biosynthesis mutants, although they fail to respond to exogenous GA treatment (Sun, Reference Sun2000). GA-insensitive dwarf1 gene GID1 encodes a soluble receptor mediating GA signalling in rice (Ueguchi-Tanaka et al., Reference Ueguchi-Tanaka, Ashikari, Nakajima, Itoh, Katoh, Kobayashi, Chow, Hsing, Kitano, Yamaguchi and Matsuoka2005; Hartweck et al., Reference Hartweck and Olszewski2006), GID2 encodes a rice F-box protein, which is essential for GA-mediated DELLA protein degradation (Gomi et al., Reference Gomi, Sasaki, Itoh, Ueguchi-Tanaka, Ashikari, Kitano and Matsuoka2004). Rice GA-insensitive dwarf mutant gene Dwarf 1 encodes the subunit of GTP-binding protein; dwarf mutant d1, which is defective in a subunit of the heterotrimeric G protein, affects GA signal transduction (Ueguchi-Tanaka et al., Reference Ueguchi-Tanaka, Fujisawa, Kobayashi, Ashikari, Iwasaki, Kitano and Matsuoka2000; Ashikari et al., Reference Ashikari, Wu, Yano, Sasaki and Yoshimura1999). SLR1 is an intermediate of the GA signal transduction pathway; slender mutant (slr1-1) results in a constitutive GA response phenotype (Ikeda et al., Reference Ikeda, Ueguchi-Tanaka, Sonoda, Kitano, Koshioka, Futsuhara, Matsuoka and Yamaguchi2001). Most modern commercial wheat cultivars contain Rht mutant alleles; Rht-B1b and Rht-D1b are mutations in the GA response pathway (Silverstone & Sun, Reference Silverstone and Sun2000; Muangprom & Osborn, Reference Muangprom and Osborn2004).
So far, five genes (ga1, ga2, ga3, ga4 and ga5) encoding key enzymes in the GA biosynthesis pathway and one gene in the GA response pathway (GAI) have been cloned in Arabidopsis (Sun et al., Reference Sun, Goodman and Ausubel1992; Yamaguchi et al., Reference Yamaguchi, Sun and Kawaide1998; Helliwell et al., Reference Helliwell, Sheldon and Olive1998; Chang et al., Reference Chang, Hwang and Goodman1995; Xu et al., Reference Xu, Li and Wu1995; Peng et al., Reference Peng, Richards, King, Cowling, Murphy and Harberd1997). The GA biosynthesis orthologous genes have been cloned in several plant species, such as ls (Ait-Ali et al., Reference Ait-Ali, Swain and Reid1997) and le (Martin et al., Reference Martin, Proebsting and Hedden1997) in pea, D8 in maize and Rht1 in wheat (Peng et al., Reference Peng, Richards and Hartly1999).
More dwarf mutant genes related to GA in rice have also been cloned, such as d18 (Itoh et al., Reference Itoh, Ueguchi-Tanaka and Yutaka2001), D35 (Itoh et al., Reference Itoh, Tatsumi, Sakamoto, Otomo, Toyomasu, Kitano, Ashikari, Ichihara and Matsuoka2004) and sd-1 (Monna et al., Reference Monna, Kitazawa, Yoshino, Susuki., Masuda, Maehara, Tanji, Sato, Nasu and Minobe2002; Sasaki et al., Reference Sasaki, Ashikari, Ueguchi-Tanaka, Itoh, Nishimura, Swapan, Ishiyama, Saito, Kobayashi, Khush, Kitano and Matsuoka2002; Spielmeyer et al., Reference Spielmeyer, Ellis and Chandler2002) with a deficiency in the GA biosynthesis pathway, and D1 (Ueguchi-Tanaka et al., Reference Ueguchi-Tanaka, Fujisawa, Kobayashi, Ashikari, Iwasaki, Kitano and Matsuoka2000; Ashikari et al., Reference Ashikari, Wu, Yano, Sasaki and Yoshimura1999) and Gid2 (Gomi et al., Reference Gomi, Sasaki, Itoh, Ueguchi-Tanaka, Ashikari, Kitano and Matsuoka2004 ) with a deficiency in the GA response pathway.
In 1997, a novel semi-dwarf mutant rice was isolated in our research (Tong et al., Reference Tong, Wu, Wu, Zheng and Yu2001). In this paper, the genetic characterization, GA response and molecular mapping of the semi-dwarf mutant gene are reported.
2. Materials and methods
(i) Plant materials and field planting
A semi-dwarf mutant was found in the F6 generation of a medium japonica rice cross between M9056 and R8018 XUAN in 1997. In 1998, the seeds harvested from this mutant were planted in the field and the plant heights of their progenies were recorded and analysed. To study the genetic character and map the mutant gene, plant materials used in these researches were as follows: the semi-dwarf mutant, tall wild-type, and the F1, F2 progenies derived from reciprocal crosses between the semi-dwarf mutant and wild-type; Hua-jing-xian74 (an elite indica rice cultivar); a residual heterozygous line (RHL) population derived from RHL63–146; and a recombinant inbred line (RIL) (F6) originated from an inter-subspecific cross between the semi-dwarf mutant (japonica cv.) and Hua-jing-xian74 (indica cv.).
These plant materials were planted in the field during the rice-growing seasons from 2000 to 2004 at the experiment stations at Hefei (31°N, 117°E), Anhui province, and Sanya (18°N, 109°E), Hainan province, China. The planting density was 13·3 cm between plants in a row, and 16·7 cm between rows, with 11 plants per row. Field management followed normal agricultural practices. Irrigation of the field was maintained to avoid drought stress. Plant heights were recorded at maturity.
(ii) GA response experiment design
In agriculture, exogenous GA3 treatment has usually been used to stimulate panicle emergence in male sterile (MS) lines to gain greater yield in hybrid rice seed production in China. In order to study the response of the semi-dwarf mutant to exogenous GA, GA3 solution was sprayed on both the semi-dwarf mutant and the tall wild-type at a dosage of 5·6×10−4 g per plant at different rice developmental periods including the seedling stage, tillering stage, heading stage and milking ripe stage. From the jointing stage to the heading stage in 2002, the same dosage of GA3 solution was sprayed on the semi-dwarf mutant and the tall wild-type on 17 July, 25 July and 2 August, respectively. The plant height, panicle and elongation internode lengths were recorded at maturity.
(iii) Mapping population development
To dissect the genetic factors underlying the semi-dwarf mutant, temporarily designated Sdt97, and to construct a population for mutant gene Sdt97 mapping, an inter-subspecific cross was carried out between the mutant, a semi-dwarf japonica rice line, and Hua-jing-xian74, a semi-dwarf indica rice cultivar.
A line with a heterozygous segment surrounding a gene is denoted as a residual heterozygous line (RHL), and can be used for gene mapping in map-based cloning (Yamanaka et al., Reference Yamanaka, Watanabe, Hayashi, Fuchigami and Harada2005). In the Sdt97 mapping, a segregated F6:7 progeny derived from a single line (named RHL63-146), which derived from the 186 RILs (F6) and was identified to be heterozygous around locus Sdt97 (Yamanaka et al., Reference Yamanaka, Watanabe, Hayashi, Fuchigami and Harada2005; Tuinstra et al., Reference Tuinstra, Ejeta and Goldsbrough1997), was selected for Sdt97 mapping (see Section 3 below).
(iv) DNA extraction, BSA and RCA
Genomic DNA was extracted individually from fresh leaves of the parental F6:7 plants, using the modified CTAB method (Doyle & Doyle, Reference Doyle and Doyle1987; Ragers & Bendich, Reference Ragers and Bendich1998). PCR amplification was performed with the Programmable Thermal Controller PTC 100 (MJ Research, Watertown, MA). PCR reactions consisted of 2·5 μl of 10× reaction buffer (with (NH4)2SO4, 100 mM), 2·0 μl MgCl2 (25 mM), 1·0 μl dNTPs (10·0 μM), 1·0 unit Taq DNA polymerase, 100 ng template DNA and 1·5 μl of primer (10 μM), made up to 25·0 μl with distilled water, and then covered with a drop of mineral oil. Amplifications were performed using the following profile: 94°C for 5 min, 35 cycles of 94°C for 1 min, 55°C for 1 min and 72°C for 1 min, with a final extension at 72°C for 5 min. Amplification products were analysed on 4% agarose gels stained with ethidium bromide, and photographed using the Gel Doc 2000 system. When necessary, the amplification products were further analysed on 6% polyacrylamide gel stained with 0·1% silver nitrate.
Bulked segregant analysis (BSA) (Michelmore et al., Reference Michelmore, Paran and Kesseli1991) combined with recessive class analysis (RCA) (Chen et al., Reference Chen, Wang, Que, Pan and Pan2005; Pan et al., Reference Pan, Hu, Tanisaka and Wang2003; Zhu et al., Reference Zhu, Wang and Pan2004; Zhang et al., Reference Zhang, Shen, Dai, Mei, Saghai Maroof and Li1994) was used to identify molecular markers linked to the mutant gene Sdt97 in this study. Genomic DNA from 30 semi-dwarf individuals and 30 dwarf individuals in the F6:7 segregated progenies was pooled to create the semi-dwarf and dwarf DNA bulks, respectively. The parental DNA and the two bulks were used for BSA. Markers were examined for polymorphism between the mutant and Hua-jing-xian74.
Polymorphic markers from the two parents were screened against the two DNA bulks, and polymorphic markers between the two DNA bulks were screened against the recessive individuals. The marker that was linked to the target gene Sdt97 was screened against the entire F6:7 segregated population. Polymorphic markers were used for co-segregation analysis with the plant height genotype. The formula for recombination fraction calculation is:
where r is recombination fraction. The plant height genotype of the individuals in the F6:7 populations was validated by progeny testing of F7:8.
(v) Marker development, linkage analysis and fine-mapping of Sdt97
Only PCR-based SSR markers were used in the primary map study. A set of 832 SSR markers (data not shown) evenly distributed throughout the 12 chromosomes was used. Their map locations, primer sequences and other details are available online at http://gramene.org.ricemicrosat.html. The primer sets were adopted from the International Rice Microsatellite Initiative (IRMI, http://www.gramene.org; McCouch et al., Reference McCouch, Teytelman, Xu, Lobos, Clare, Walton, Fu, Maghirang, Li, Xing, Zhang, Kono, Yano, Fjellstrom, DeClerck, Schneider, Cartinhour, Ware and Stein2002), and the detection procedures are as described by Zhu et al. (Reference Zhu, Wang and Pan2004). SSR primers were synthesized by Shanghai CASarray Co. Ltd.
For further linkage analysis, position-specific microsatellite (PSM) and sequence-tagged site (STS) markers were developed in the target region defined by the SSR markers through bioinformatics analysis (BIA) using the publicly available reference sequences of the entire rice genome of two subspecies: japonica (cv. Nipponbare; http://rgp.dna.affrc.go.jp) and indica (cv. 93–11; http://www.genomics.org.cn). In particular, primer sets for PSM markers were designed based on the reference sequence of cv. Nipponbare using the software tools SSRIT (http://www.gramene.org/microsat/ssrtool) and Primer Premier 5.0 (PREMIER Biosoft International, http://www.premierbiosoft.com). Primer sets for STS markers were designed according to the sequence comparison between the two subspecies in the target region using a software tool, Pairwise BLAST (http://www.ncbi.nlm.nih.gov/blast/bl2seq/bl2.html). When a large deletion existed between the two subspecies, the sequence was considered a candidate from a putative STS marker.
To determine the linkage relationship between the Sdt97 locus and molecular markers, the genotype of plant height was combined with DNA marker data for linkage analysis. Linkage analysis was conducted using the Mapmaker/Exp 3.0 program (Lincoln et al., Reference Lincoln, Daly and Lander1992) at a LOD threshold of 3·0 to construct a local genetic map for the Sdt97 genomic region. Map distance between marker and semi-dwarf gene was estimated by the Kosambi mapping function:
where x is the map distance and r is the recombination fraction (Kosambi, Reference Kosambi1944).
3. Results
(i) Identification and genetic analysis of the semi-dwarf mutant
A semi-dwarf mutant was isolated from the tall F6 progenies which derived from the cross between M9056 and R8018 XUAN in 1997 (Fig. 1). In 1998, seeds harvested from this mutant were planted in the field. Individuals in this population could be divided into two groups: semi-dwarf and tall. All the individuals in this population were self-bred and field- planted in 1999. Among 38 progenies, 9 exhibited semi-dwarf non-segregation, 21 exhibited continued segregation and 8 exhibited tall non-segregation. The ratio of non-segregation semi-dwarf progenies to segregation progenies to non-segregating tall progenies was 1:2·333:0·889 (χ2=0·4737, P>0·05). Among 21 segregating progenies, there were 532 semi-dwarf and 172 tall individuals; the ratio of semi-dwarf to tall individuals equalled 2·993 (χ2=0·0473, P>0·05). These results revealed that there was only one dominant gene locus involved in the control of the semi-dwarfism of the semi-dwarf mutant.
Fig. 1. Plant height, panicle and elongation internode length of the tall wild-type (left) and semi-dwarf mutant (right).
In order to analyse the genetic basis of semi-dwarfism of the mutant further, the plants derived from the non-segregating semi-dwarf progeny (Y98149) and the non-segregating tall progenies (Y98148) were selected as parents, and reciprocal crosses between them were made. In 2000, Y98149, Y98148 and the reciprocal F1 were field-planted.
The plant height of Y98149 was 65·8±4·613 cm (mean±SD; 20 plants were measured, for these and subsequent measurements), the plant height of Y98148 was 84·3±4·814 cm, that of Y98149×Y98148/F1 was 65·0±6·486 cm, and that of Y98148×Y98149/F1 was 68·3±2·517 cm. It was clear that the reciprocal F1 exhibited the same plant height as that of the semi-dwarf mutant (Y98149), and they showed a significant difference compared with the height of the tall wild-type (Y98148).
The reciprocal F2 were planted in 2001, and obvious segregation in plant heights occurred. In the Y98148×Y98149/F2 population, distribution of plant heights appeared to be definitely bimodal, breaking at the point in the apparent valley at 99·5 cm: of all 552 individuals, 427 were semi-dwarf and 124 were tall. The ratio of semi-dwarf individuals to tall individuals was 3·4436 (χ2=1·6993, P>0·05). In the Y98149×Y98148/F2 population, the distribution of plant height appeared to be definitely bimodal also, breaking at the point in the apparent valley at 101·5 cm; of all 544 individuals, 412 were semi-dwarf and 132 were tall. The ratio of semi-dwarf individuals to tall individuals was 3·1212 (χ2=1·2010×10−1, P>0·05) (Fig. 2). These results illustrate that a single dominant nuclear gene locus is involved in the control of semi-dwarfism of the mutant, and that the semi-dwarfism expression of the mutant is not affected by its cytoplasm. Similar results were also obtained in 2002 and 2003 (data not shown). It was deduced that the semi-dwarfism of the mutant did not result from the cross-fertilization of the tall wild-type to other dwarf or semi-dwarf rice cultivars, but that a spontaneous mutation had occurred in the tall wild-type gene locus.
Fig. 2. Plant height distribution of the 544 plants in the mutant×wild-type/F2 population.
(ii) Responses of the mutant to exogenous GA
In agriculture, exogenous GA3 treatment has usually been used to stimulate panicle emergence in MS lines to gain greater yield in hybrid rice seed production in China. To research the responses of the semi-dwarf mutant to exogenous GA, exogenous GA3 solution was sprayed on the semi-dwarf mutant at different developmental periods. The results showed that the semi-dwarf mutant was sensitive to exogenous GA3 only during the period from the stem elongation stage to the heading stage. During this time, the semi-dwarfism of the mutant could be partly restored to normal phenotype by exogenous GA3.
At the seedling stage, the plant heights of semi-dwarf mutants treated with GA3 and those treated with water were 69·7±7·5 cm and 68·2±2·7 cm, respectively; at the milking ripe stage they were 69·4±3·5 cm and 67·6±3·0 cm, respectively. At the seedling stage and at the milking grain stages, the semi-dwarf mutants treated with GA3 showed the same plant height as those treated with water. But at the stem elongation stage and heading stages, the plant height of mutants treated with GA3 was significantly different from that of the plants treated with water. At the stem elongation stage, the heights of GA-treated and untreated plants were 80·9±8·2 cm and 72·0±4·3 cm, respectively, and at the heading stage they were 82·9±5·0 cm and 71·6±3·5 cm, respectively.
The above results for the semi-dwarf mutant imply that the mutant gene Sdt97 might be involved in the GA synthesis pathway and not the GA response pathway in rice.
(iii) Mapping population development
To map the semi-dwarf mutant gene Sdt97, a cross between the semi-dwarf mutant and the tall wild-type was made. A total of 680 SSR markers distributed evenly throughout 12 chromosomes were selected for polymorphism scanning between the semi-dwarf mutant and the wild-type; however, no polymorphism was detected. This result implied that the F2 population derived from the cross between the semi-dwarf mutant and tall wild-type is difficult to use for the mapping of Sdt97.
To dissect the genetic factors underlying the semi-dwarf mutant, and to construct a population for Sdt97 mapping, an inter-subspecific cross was carried out between the mutant, a semi-dwarf japonica rice line with the genotype Sdt97Sdt97 Sd-1Sd-1, and Hua-jing-xian74, a semi-dwarf indica rice cultivar with the genotype sdt97sdt97sd-1sd-1. There is a two-locus difference relating to dwarfism between these two varieties. However, the plant height of the F2 population was more continuously distributed, Plant height phenotypes of the individuals in this F2 population cannot easily be separated into discrete classes. This being the case, we attempted to develop another mapping population.
In the segregated F2 progeny, extreme dwarf individuals with a genotype of Sdt97_sd-1sd-1 were selected. In the segregated F2:3 (hereafter F2:3 means the F3 population derived from one of the F2 individuals) progenies, dwarf individuals with the same genotype were selected. Among the segregated F3:4 progenies, one of the F3:4 progeny populations had two discrete classes of plant height: semi-dwarf individuals and dwarf individuals. The ratio of dwarf individuals to semi-dwarf individuals fitted the expected Mendelian segregated ratio of 3:1. This suggested that only one dominant Mendelian factor was involved in controlling the segregation of plant height in the population. Because no tall individuals occurred, the impact of sd-1 can be eliminated and the genetic effects of Sdt97 can be investigated.
Dwarf individuals with genotype Sdt97_sd-1sd-1 were selected in the segregated F3:4 progeny. A similar process was used to select dwarf individuals in segregated F4:5 and segregated F5:6 progenies. A segregated F6:7 progeny derived from a single line (named RHL63-146), which derived from the 186 RILs (F6) and was identified to be heterozygous around the locus Sdt97 (Yamanaka et al., Reference Yamanaka, Watanabe, Hayashi, Fuchigami and Harada2005; Tuinstra et al., Reference Tuinstra, Ejeta and Goldsbrough1997), was selected for Sdt97 mapping; it comprised 257 individuals.
(iv) Identification of Sdt97 by RHL population
In the F6:7 progenies, populations were categorized into two groups according to plant height. Group 1, derived from homozygous dwarf F6 individuals, exhibited dwarf non-segregation; 44 F6:7 progenies were categorized in this group. Group 2, derived from heterozygote dwarf F6 individuals, showed continued plant height segregation and resulted in both semi-dwarf and dwarf individuals; 82 F6:7 progenies were categorized in this group (Fig. 3). The ratio of group 2 to group 1 progenies was 1·86. The ratio of the segregating F6:7 progenies to the non-segregating dwarf F6:7 progenies did not differ significantly from 2:1 (χ2=0·081, P>0·05). The results confirmed there was only one dominant gene controlling the segregation of plant height in the RHL population and the gene was designated Sdt97.
Fig. 3. Semi-dwarf plant (left) and dwarf plant (right) in RHL63-146, the mapping population of segregating F6:7 progenies.
One segregating F6:7 progeny population, RHL63-94, was investigated further. Distribution of plant height in this population appeared to be bimodal. Of the 693 individuals, 524 were dwarf and 169 were semi-dwarf giving a ratio of dwarf to semi-dwarf plants of 3·10 (χ2=0·1082, P>0·05) (Fig. 4). It was suggested that only one single dominant gene locus was involved in the segregation of plant height in this population. Similar results were obtained in 2004 (data not shown).
Fig. 4. Plant height distribution of the 693 plants in RHL63-94, a segregating F6:7 progeny population.
The parents of the segregating F6:7 progenies were a residual heterozygous line (RHL) (Yamanaka et al., Reference Yamanaka, Watanabe, Hayashi, Fuchigami and Harada2005; Tuinstra et al., Reference Tuinstra, Ejeta and Goldsbrough1997), and the two chromosomes in their nuclear genome showed close chromosomal similarity to each other along the entire genome length with the only heterozygous sequences covering the Sdt97 locus. The heterozygous chromosomal region (approximately 25·5 cM, 6646 kb, starting at RM3430 and ending at RM6395; data not shown) initiated from different parents, one carrying the Sdt97 gene derived from the mutant, and the other carrying the sdt97 gene derived from Hua-jing-xian74. The DNA marker polymorphism was easily detected in this chromosome region.
(v) Molecular mapping of Sdt97
One segregating F6:7 progeny, RHL63-146, comprising 257 individuals, was selected for Sdt97 gene mapping in this study. Six hundred and eighty known SSR markers selected from 12 rice chromosomes with intervals of 2·7 cM were tested in the segregated F6:7 populations, using the BSA approach. One SSR marker (RM340), located on the long arm of chromosome 6, showed positive polymorphisms for the mutant gene in the mapping populations. The RCA approach was then employed to determine the linkage relationships between the semi-dwarf mutant gene and marker RM340. A total of 72 extremely semi-dwarf individuals were subjected to linkage analysis and 16 distinct recombinants were identified. Polymorphic markers were confirmed by testing the plants in the populations individually. The results showed that marker RM340 co-segregated with the mutant gene.
To confirm this result, an additional 77 SSR markers located on chromosome 6 were tested. Results showed that markers RM3138, RM5509, RM3430, RM6395, RM5371, RM5314 and RM5957 co-segregated with the semi-dwarf gene locus in the mapping population. Markers RM3430 and RM6395 were closer to the semi-dwarf locus, and the Sdt97 locus was flanked by RM3430 on the telomeric side and RM6395 on the centromeric side at a distance of 3·6 cM and 2·4 cM, respectively, indicating that the gene involved in the mutation is located on the long arm of chromosome 6.
To find additional markers flanking the Sdt97 locus, 40 new SSR markers located in the target region were adopted from IRMI and the Rice Genome Sequence Program, Japan (RGP; web site: http://rgp.dna.affrc.go.jp). Among them, the two markers RM20495 and RM20522, where 2 and 11 recombinants from those identified at RM6395 and RM3430 loci, respectively, were detected, indicating that the Sdt97 locus was further defined by the markers on both sides at 0·4 and 2·6 cM, respectively.
For further fine-mapping of the Sdt97 locus, 14 PSM markers and 74 STS markers were developed in the smaller region based on the reference sequences of cv. Nipponbare by BIA. Among these new PCR-based markers, two PSM markers (RT1, RT5) and three STS markers (TX5, TX7, N6) showed polymorphism to the parents. The five polymorphic markers were further tested and the two markers RT1 and RT5, where 2 and 5 recombinants from those identified at RM20495 and RM20522 loci, respectively, were detected. The three STS markers, TX5, TX7 and N6, where 4, 4 and 1 recombinants from those identified at RT5 and RT1 loci, respectively, were detected, and no recombinant was detected at the other loci. These revealed that a total of 15 markers (RM340, RM3138, RM5509, RM3430, RM20522, RT5, TX7, TX5, N6, RM20495, RT1, RM6395, RM5371, RM5314 and RM5957) co-segregated with the Sdt97 locus. The genetic region spanning the Sdt97 locus between TX5 and N6 was estimated to be 1·0 cM in length (Fig. 5).
Fig. 5. Fine mapping of Sdt97.
A contig map covering the Sdt97 locus through Pairwise BLAST analysis was constructed. The physical distance between markers N6 and RT5 is 118 kb on the RGP BAC/PAC contig (Fig. 6). Based on the available sequence annotation database (http://www.rgp.dna.affrc.go.jp; http://www.tigr.org), there are 19 predicted genes in the 118 kb target region of the cultivated rice genome. Of these, 10 protein genes, 7 putative genes and 2 hypothetical protein genes were identified (see Supplementary Table 1). Identification of the candidate gene of Sdt97 is still in progress.
Fig. 6. A contig map covering the Sdt97 allele region. RM20495 and RM20522 are SSR markers, RT1 and RT5 are PSM markers, and N6, TX5, and TX7 are STS markers. The long horizontal line indicates the genomic region encompassing the Sdt97 locus. The short horizontal line represents BAC/PAC clones of cv. Nipponbare with the accession numbers indicated. The vertical lines indicate the relative position of the corresponding marker on BAC/PAC clones
Supplementary Table 1. Candidate genes in the 118 kb target region
4. Discussion
One technique that has been widely used to isolate the genes identified as quantitative trait loci (QTLs) is map-based cloning. To clone genes by this method, investigators adopt a fine-mapping strategy using a series of near-isogenic lines (NILs), introgression lines or chromosome-substitution lines. For the fine-mapping of QTLs, it is necessary to carry out high-resolution linkage analysis using a large number of plants that segregate only around the QTL being investigated. However, The RHL strategy has two main advantages over using an NIL developed by backcrossing. The first is that in developing the population for fine-mapping, only one line from an RIL population need be selected on the basis of its genotype; repeated backcrossings and selections based on DNA markers or phenotypes are not required. The second advantage is that the genomic composition of a RHL can be determined merely by checking the genotype data of the linkage map; it is not necessary to analyse the genotypes of any of the markers (Yamanaka et al., Reference Yamanaka, Watanabe, Hayashi, Fuchigami and Harada2005). Using an RHL derived from an RIL, Yamanaka et al. (Reference Yamanaka, Watanabe, Hayashi, Fuchigami and Harada2005) succeeded in fine-mapping the soybean flowering-time QTL, and using the RHL population, we succeeded in mapping the semi-dwarf mutant gene Sdt97 to the long arm of chromosome 6 in rice.
sd-1, termed the ‘Green Revolution gene’, was first identified in the Chinese variety Dee-geo-woo-gen (DGWG), and developed in the semi-dwarf cultivar IR8, which produced record yields throughout Asia and formed the basis for the development of new high-yielding, semi-dwarf plant types (International Rice Research Institute, 1967). It has contributed substantially to the significant increase in rice production and averted a chronic food shortage, an issue of great concern after the rapid expansion of the world population since the 1960s (Jennings, Reference Jennings1964; Spielmeyer et al., Reference Spielmeyer, Ellis and Chandler2002).
Since the 1960s, more than 60 dwarf genes have been identified in rice; most of these are recessive genes, and only one dominant gene, D53, was reported. These genes are associated with traits such as severe dwarfism, floret sterility, or abnormal plant and grain development; therefore, most of the dwarf mutants identified in rice (d1 to d60) have not been used in crop improvement (Aquino & Jennings, Reference Aquino and Jennings1966; Kinoshita, Reference Kinoshita1995). In recent years, new semi-dwarf genes non-allelic to sd-1 have been identified in rice (Liang et al., Reference Liang, Gu, Pan, Liang and Zhu1994, Reference Liang, Cao, Sui, Zhao, Yan, Yi and Gu2004; Li et al., 2001, Reference Li, Xu, Wang, Yan, Liang and Gu2003; Jiang et al., Reference Jiang, Liang, Zhai, Gu, Lu, Xu and Zhu2002; Zhao et al., Reference Zhao, Liang, Zhou, Yan, Cao and Gu2005), but sd-1 is still the primary semi-dwarf gene used in rice breeding (Kinoshita, Reference Kinoshita1995). About one-half of the stock from the International Rice Research Institute collection is allelic to Dgwg. In southern China, the semi-dwarf gene found in varieties of economic importance is found at the same locus as sd-1 (Gu & Zhu, Reference Gu and Zhu1979; Xiong et al., Reference Xiong, Min, Yu, Lin, Chen and Yang1989).
Frequent usage of the sd-1 gene may reduce genetic diversity and bring about genetic vulnerability to pests and diseases. It is of great significance to develop a new source for broadening the genetic basis of semi-dwarfism. Given that the dwarf or semi-dwarf germ plasm in rice reported previously results mostly from recessive genes, and can not be directly utilized in heterosis, it is obvious that the discovery and utilization of a dominant semi-dwarf gene may be of great significance, not only in genetic theory but also in rice breeding practice (Tong et al., Reference Tong, Wu, Wu, Zheng and Yu2001, Reference Tong, Wu, Wu, Zheng, Zhang and Zhu2003).
The dominant mutant gene Sdt97 reported here is not allelic to the sd-1 gene on chromosome 1 (Monna et al., Reference Monna, Kitazawa, Yoshino, Susuki., Masuda, Maehara, Tanji, Sato, Nasu and Minobe2002; Sasaki et al., Reference Sasaki, Ashikari, Ueguchi-Tanaka, Itoh, Nishimura, Swapan, Ishiyama, Saito, Kobayashi, Khush, Kitano and Matsuoka2002; Spielmeyer et al., Reference Spielmeyer, Ellis and Chandler2002), sd-t (Li et al., 2000; Jiang et al., Reference Jiang, Liang, Zhai, Gu, Lu, Xu and Zhu2002 ), sd-t2 (Zhao et al., Reference Zhao, Liang, Zhou, Yan, Cao and Gu2005) on chromosome 4, sd-g (Liang et al., Reference Liang, Gu, Pan, Liang and Zhu1994, Reference Liang, Cao, Sui, Zhao, Yan, Yi and Gu2004) and sd-n (Li et al., Reference Li, Xu, Wang, Yan, Liang and Gu2003) on chromosome 5, or dwarf gene D53 on chromosome 11 (Wei et al., Reference Wei, Xu, Li, Qian and Zhu2006). It is a new kind of semi-dwarf gene reported in rice.
The semi-dwarf mutant was deduced to result from a spontaneous mutation, and possessed a number of desirable characteristics. In conventional rice breeding programmes the semi-dwarf mutant could be used as an elite parent. If used in inter-subspecific heterosis utilization, it could provide a power genetic tool to resolve the problem of excessive height in the inter-subspecific F1 hybrid. In our research, Sdt97 and pms3 have been recombined together, a series of semi-dwarf PSGMS rice has been bred successfully, and they are now available for two-line hybrid rice breeding programmes.
This work is supported by the National Nature Science Foundation of China (3037863). the Nature Science Foundation of Anhui province (01041103), and the 973 Project of the Ministry of Sciences and Technology of China (2006CB101704).
