MECHANISM OF STERILITY OF PHOTOPERIOD/THERMO-SENSITIVE GENIC MALE STERILE LINE. The transition from sterile to fertile was the result

Rice Science, 2010, 17(3): Copyright 2010, China National Rice Research Institute. Published by Elsevier BV. All rights reserved DOI: Mechanism of Sterility and Breeding Strategies for Photoperiod/Thermo- Sensitive Genic Male Sterile Rice CHEN Li-yun, XIAO Ying-hui, LEI Dong-yang (Rice Research Institution, Hunan Agricultural University, Changsha 410128, China) Abstract: To understand the male sterility mechanism of the photoperiod/thermo-sensitive genic male sterile line in rice, the research progress of genetics of photoperiod and/or temperature sensitive genic male sterility [P(T)GMS] in rice was reviewed. A new idea was proposed to explain the sterility mechanism of P(T)GMS rice. The transition of sterile to fertile is the result of cooperating regulation of major-effect sterile genes with the photoperiod and/or temperature sensitive genes, but not the so-called pgms gene in P(T)GMS rice. The minor-effect genes, which exhibited cumulative effect on sterility, were the important factors for the critical temperature of sterility transion. The more minor-effect genes the sterile line holds, the lower the critical temperature of sterility transition is. The critical temperature of sterility transition will be invariable if all the minor-effect genes are homozygous. The strategies for breeding photoperiod/thermo-sensitive male genic sterile rice were also proposed. The selective indices of critical photoperiod and temperature for sterility transition should be set according to the exact varieties and ecological region. Imposing selection pressures was the key technology to breed P(T)GMS rice with lower critical temperature for sterility. Improving the comprehensive performance of the whole traits and combining ability was vital for photoperiod/thermo-sensitive genic male sterile rice line. Key words: rice; P(T)GMS; mechanism of sterility; breeding strategies INTRODUCTION Two-line hybrid rice system was established by using a photoperiod-sensitive genic male sterile (PGMS) mutant discovered in 1973 by SHI Ming-song from a japonica cultivar Nongken 58 (Shi, 1985). According to Yuan s assumption, in China, hybrid rice would shift from three-line to two-line system (Yuan, 1987). At present, two-line hybrid rice production system is based on the use of photoperiod and/or temperature sensitive genic male sterility [P(T)GMS]. This kind of P(T)GMS has wider restorers; therefore, its combination can be more easily obtained. The frequency of heterotic hybrids is higher in two-line hybrids than in three-line hybrids, thereby increasing hybrid breeding efficiency. Since there is no need for restorer genes in the male parents of two-line hybrids, this system is ideal for developing indica/japonica hybrids because most japonica lines do not possess restorer genes. There is no need for a maintainer line for seed multiplication, thus making seed production simpler and more cost-effective. In the aspects of rice quality, yield performance and resistance and so on, two-line hybrids have obvious superiority. Under the support of the National High-tech Research and Development Program, two-line hybrids have been used in large-scale production. At the end of the 10th five-year program, 27 two-line hybrid rice cultivars were released in various rice-growing regions covering more than 3 million hm 2, resulting in increased rice production of up to 3 billion Kg (http://www.863.org.cn/15year/biology/bly26.html). Great success has been achieved in the application of two-line hybrid rice in China. The related basic theoretical research has also made great development and the mechanism of male-sterility of TPGMS has been a hot topic. P(T)GMS is a typical phenomenon of eco-genetic heredity which is controlled both by gene(s) (internal) and ecological factors (external) such as photoperiod and temperature. But the studies on the sterility mechanism could not reach consistent results. Various genetic models or hypotheses were put forward, including one or two pairs of genes controlling the sterility (Shi and Deng, 1986; Yang, 1997; Xie et al, 2000). And it was also reported that the photoperiod sensitive genic sterility in Nongken 58S exhibited characteristics of qualitative-

Rice Science, Vol. 17, No. 3, 2010 quantitative heredity. The results differed a lot even using the materials with the same sterility gene source (Xue and Deng, 1991). There are four antithetical types of reports on the mapping of PGMS sterility gene: (1) On chromosome 5: Zhang et al (1990) first reported one pair of photoperiod gene pms in Nongken 58S linked to marker gene d-1 on rice chromosome 5, which was further validated by Qian et al (1995) and Lin et al (1996); (2) Wu and Wan (1991) located two pairs of photoperiod genes on chromosome 3 and 11, respectively; (3) On or not on chromosome 7: Zhang et al (1994) screened pms1 and pms 2 on chromosome 7 and 3, respectively. Wang et al (1995) also identified molecular marker PGMS0.7 linked to PGMS sterility gene; Nevertheless, according to the research by Wang et al (1997), the mutated gene, which caused Nongken 58 to Nongken 58S, did not exist in the region of pms1 on chromosome 7; (4) On chromosome 12: Wang et al (1995) screened markers associated with sterility gene of PGMS; Mei et al (1999) detected gene pms3 and Li et al (1999), Chen et al (2000) and Li et al (2001; 2002) also found some markers related to this gene. What s more, Lu et al (2005) located the gene pms3 to the region of 28.4 kb on chromosome 12, Many reasons can be used to explain why the results of genetic studies on PGMS sterility gene differed greatly. For one thing is that most of the researchers used F 2 population. Owing to the heavy segregation of growth period in F 2 population, the light and temperature conditions encountered in different stages during the fertility sensitive period, consequently, the recognition of its genetic model would be misled. On the other hand, that the expression of sterile gene was affected by genetic background was always used as an explanation (Zhang and Zhu, 1991). So, what is the genetic background? Yuan et al (1988) conceived the hypothesis, viz. two photoperiodic responses. The first one induces transition of vegetative growth to reproductive growth and subsequently the second one induces the happening of fertility alternation. As we know, there are still many phenomena encountered during the two-line hybrid rice breeding which can not be explicated by photoperiod sensitive male sterility gene. A PGMS line can be bred to an early indica TGMS line (e.g. W6154S was developed from the progeny of Nongken 58S ); then where has the original sterility gene gone? On the contrary, a PGMS line can be bred from the crosses between a TGMS line and photo-sensitive variety; so, where does the photoperiod-sensitive sterility gene come from? In addition, from the long term breeding practice, we discover that there is close relationship between the phototonus and photoperiod-sensitive sterility. In the investigation on the characteristics of the photoperiod sensitivity and sterile characteristic of C815S and its homologous plant lines, we found that the different plant lines from the same hybrid combination differed a lot. Those with strong phototonus had strong photoperiod-sensitive sterility characteristics, those with neutral phototonus had moderate photoperiod-sensitive sterility and those with weak phototonus presented thermo-sensitive sterility characteristics (He et al, 2007). Since P(T)GMS line is related to photoperiod and temperature, how do the light and temperature combined with genetic male sterile gene that causes the male sterility? Therefore the elucidation of the mechanism of photo-thermo sensitive sterile will play a significant role in breeding and propagation of P(T)GMS lines and in hybrid seed production technology. MECHANISM OF STERILITY OF PHOTOPERIOD/THERMO-SENSITIVE GENIC MALE STERILE LINE The transition from sterile to fertile was the result of cooperating regulation of major-effect sterile genes with the photoperiod and/or temperature sensitive genes The so-called pgms gene does not exist in P(T)GMS rice. There is only one or two major-effect sterile genes like the common nuclear male sterility gene (like the nuclear male sterile line Dong A of cotton) which is usually obtained from mutation. The phenomenon of sterility of photoperiod/ thermo-sensitive genic male sterile line in rice was the result of cooperating regulation of major-effect sterile genes with the photoperiod and/or temperature

CHEN Li-yun et al. Mechanism of Sterility and Breeding Strategies for Photoperiod/Thermo- Sensitive Genic Male Sterile Rice sensitive genes. The sterility/fertility changes because of the effect or control of the photoperiod and/or temperature sensitive genes. The induction of photoperiod and/or temperature sensitive genes to male sterile genes should be from quantitative change to qualitative change to determine the duration of the transition from sterile to fertile. The minor-effect genes affect the critical temperature of sterility transfer on P(T)GMS The minor-effect genes, which exhibit cumulative effect for sterility, also exist in conventional rice varieties. It should be the main reason why some rice varieties especially the hybrid rice developed from typical indica and japonica rice have poor adaptability for high or low temperatures. The number of minor-effect genes can affect high or low temperature tolerance. The more minor-effect genes the rice variety holds, the poorer the high or low temperature tolerance, adaptability and stability of seed setting rate are. The critical temperature of sterility transfer in P(T)GMS would be changed to the effect of the minor-effect and major-effect sterility genes. The more minor-effect sterility genes pyramided, the more difficult to stabilize the fertility of P(T)GMS lines. When the minor-effect gene is homozygous, the critical temperature of sterility transfer will be lower and the fertility of P(T)GMS will be stable. Different minor-effect genes have different effect on the fertility of P(T)GMS lines, and the fertility of P(T)GMS will be more stable when it is controlled by genes with bigger effects. Inference derived from new ideas The transition from sterile to fertile resulted from the development of photoperiod and/or temperature sensitive genes reacting on the sterility gene. The degrees of photosensitivity and thermosensitivity determine the sterile degree of P(T)GMS, and the stronger photoperiodism of P(T)GMS, the higher degree photosensitive sterility it has. The photosensitive rice varieties should also be temperature sensitive varieties. This may be the reason why we cannot find pure PGMS lines which is not affected by temperature. There is interdynamic relation between day-length and temperature on PTGMS. The critical temperature of sterility becomes lower under long day-length condition. For TGMS, the critical temperature of sterility is relatively stable. Long day-length and high temperature as well as short day-length and low temperature naturally occur together, therefore, most of the P(T)GMS is sterile in long day-length/high temperature or short day-length/ low temperature. By cloning of the major sterile gene in P(T)GMS and transferring of the gene to the strong photosensitive rice varieties, PGMS would be bred because of the interaction of photoperiod sensitive genes and sterile gene, otherwise, TGMS would be bred if the major sterile gene was transferred to strong temperature sensitive rice varieties. TGMS can bred by crossing P(T)GMS and the variety which has no photoperiodism; Similarly, PGMS can bred by crossing TGMS and rice variety with strong photo-sensitive character. Early rice type TGMS line may not only have thermo-sensitive character. The photoperiod sensitivity gene would affect the sterile gene and it showed the photo-sensitive sterility phenomenon because the recessive photoperiod sensitivity gene or both photoperiod sensitivity gene and inhibitor gene for photoperiod sensitivity exit in some early rice varieties. The minor-effect genes determine the critical temperature of sterility in P(T)GMS, so the heterozygosity of minor-effect genes is the main reason for the drift of the critical temperature of sterility. The P(T)GMS seed production is the process of developing homozygous plants, once all the minor-effect genes are homozygous, the critical temperature of sterility should be stable. A P(T)GMS with low critical temperature of sterility can be developed by crossing P(T)GMS with high critical temperature of sterility and conventional rice variety through increasing selection pressure. It s difficult to choose P(T)GMS with low critical temperature of sterility when the conventional rice variety and the P(T)GMS line have close genetic relationship. Increasing selection pressure also has little effect because there is not much difference on the sterile gene loci between the parents. Using indica

Rice Science, Vol. 17, No. 3, 2010 and japonica rice as the parents or if there is large genetic distance between the P(T)GMS and the conventional variety, the new P(T)GMS with low critical temperature of sterility can be developed by increasing selection pressure. Increasing selection pressure and utilizing the effect of other minor-effect sterile genes, we could develop the P(T)GMS with more stable sterility character by pyramiding the major sterile gene from different sources. Many studies indicated that the frequency distribution of individual plant fertility is different in different F 2 populations derived from the crosses between the same P(T)GMS and different male parents. There are three types of distribution, single-peak curve, double-peak curve and multi-peak curve (Xue et al, 1995). Why the expression of photoperiod and/or temperature sensitive genes much differed in different genetic backgrounds? The reason is that the male s photo-sensitive character and thermo-sensitive character are different. The different interaction modes among photo-sensitive sterile genes, thermo-sensitive sterile genes and major sterile gene together with the effect of the minor-effect genes cause the differences in sterility of F 2 segregation ratio. In general, the seed setting rate of F 1 from two conventional rice varieties is more stable than the F 1 from P(T)GMS and another conventional rice variety, which may be due to the many minor-effect sterile genes except the major effect sterile gene in P(T)GMS. The F 1 hybrid would pyramid recessive minor-effect sterile genes. These genes would be homozygous when the male had the same minor-effect sterile genes as the P(T)GMS, and the hybrid s seed setting rate would be easily affected by climatic conditions. For example, some two-line hybrid varieties had very low seed setting rate when the flowering stage or meiosis stage encountered with low temperature. For the same reason, the seed setting rate of some hybrid rice varieties which are widely planted in China decreased quickly when encountered with high temperature. The super hybrid rice would have strong heterosis and wide adaptability when the parents have moderate genetic distance. Also the breeders should pay special attention to increase selection pressure under abiotic stresses in order to raise the adaptability of hybrids. THE STRATEGY OF P(T)GMS BREEDING The breeding strategies for P(T)GMS with different development patterns Breeding genic male sterile line with strong photoperiodism The seed production of this kind of P(T)GMS can be safe, and it is also easy to achieve high and stable yield. The critical sterility inducing temperature of this kind of P(T)GMS is low (about 21ºC) under long photoperiod and high (about 25ºC 26ºC) under short photoperiod. For the reverse PGMS, critical sterility inducing temperature is high (about 31ºC) under short photoperiod, and relatively low (about 25ºC 26ºC) under long photoperiod. In order to breed the PGMS with strong photoperiodism, it needs crosses between the varieties with strong phototonus. However, the growth duration of F 1 will has obvious transgressive inheritance, so the application of this PGMS is limited. The autumn seed production in South China is under relative short day-length, so critical sterility inducing temperature will be higher, and the seed production will be weak. The biggest advantage of the PGMS with strong photoperiodism lies in the facility to get high stable yield, however, it has no advantage in other areas. Now the techniques of consecutive irrigation with cold water for multiplying P(T)GMS with low critical sterility inducing temperature have been solved, so there is no need to deliberately seek nuclear male sterile line with strong photoperiodism. Breeding genic male sterile line with weak photoperiodism It s better to use the P(T)GMS which has a certain degree of photoperiodism for mid-season or later season hybrid rice in Yangtze River valley. Generally, Yangtze River valley is suitable for two-line hybrid seed production in summer. It is better to improve yield characters by choosing the P(T)GMS with a certain degree of photoperiodism and suitable long growth stage. For such P(T)GMS, it is recommended that the critical sterility inducing

CHEN Li-yun et al. Mechanism of Sterility and Breeding Strategies for Photoperiod/Thermo- Sensitive Genic Male Sterile Rice temperature should be about 22ºC under short photoperiod. The minimum temperature for sterility induced physiologically on rice is about 17ºC, and such P(T)GMS will be fertile when it is irrigated with cold water (20ºC 21ºC) for 12 15 d. The critical sterility inducing temperature will be raised to about 23.5ºC because of weak photoperiodism. For this kind of P(T)GMS, the seed production can be done in autumn in Guangdong, Guangxi provinces or in winter in Hainan province, China. It is easy to multiply this kind of P(T)GMS at winter in Hainan province, China. C815, which is bred by the authors, belongs to such weak photoperiodism type (Chen et al, 2007; Tang et al, 2004). Breeding TGMS The P(T)GMS which will be used to make cross combinations of early hybrd rice in Yangtze region should be early-mature rice. The genic male sterile line of early rice type should be TGMS, for there is no or weak photoperiodism in early rice varieties. The ideal critical sterility inducing temperature of TGMS is about 22.5ºC. Take Zhu 1S for an example, the critical sterility inducing temperature is 22.7ºC and it never had problem on seed production during the last eight years. Zhu 1S also demonstrated good characteristics of propagation in the winter in Hainan province or with continuous irrigation of cold water. Increasing selection pressure is the important strategy for breeding ideal PTGMS In the past, the ideal P(T)GMS was selected by the natural condition. But there is very small probability to meet the required critical sterility inducing temperature, especially for breeding P(T)GMS with low critical sterility inducing temperature. Building water temperature processing system is an ideal way to solve this problem. The water in temperature controlled pool and the processing pool is a continuous circulation system in order to ensure the constant range of water temperature and meet with the evaluation of the P(T)GMS materials. The system includes an air conditioner, temperature-controlling instrument, water circulation equipment and so on. The temperature is usually automatically controlled at 22ºC 23ºC depending on the P(T)GMS type. First, choose the P(T)GMS lines with good characters and move to water temperature processing system at the fourth stage of young panicle differentiation. Then materials will be transferred to natural light and temperature conditions after 5 7 d, and the single stems (the distance between flag leaf and second leaf is about±2cm) were marked. Check the pollen abortion rate through microscopic examination for three days and discard the materials with unacceptable fertility. Ratooning plants with good sterility and transfer them to the temperature-controlled facility with cool water at 20ºC for 10 12 d. Repeat the procedure for several generations to ensure high yield and stable sterility of P(T)GMS. Improve comprehensive traits and combining ability of PTGMS Commercial two-line hybrids should not only have high heterosis, quality grain and resistance to biotic stresses but also high stability in its seed multiplication and seed production. So far, there are only few PTGMS released in the regional trials after years, which indicates the importance of practical P(T)GMS breeding (Chen, 2001). Good comprehensive traits 1) Stability of fertility, security of hybrid seed production and high-yielding seed propagation are required for male sterile lines. 2) Outstanding agronomic traits. The male parents which are dwarf or semi-dwarf, resistant to lodging, with good plant type and high biological or biomass yield are emphasized. 3) Outstanding yield traits. The male sterile lines should have high sink capacity equal to or even surpass the control varieties in regional tests, which is an important component of high yield. 4) Good rice quality. High head rice rate (over 50%), low chalky rice rate (lower than 20%), moderate content of amylase (16% 24%) and good tastes. 5) Resistance. Moderate resistance to rice blast, resistances of certain degrees to higher and lower temperatures, sheath blight and the kernel smut are required for super parents.

Rice Science, Vol. 17, No. 3, 2010 6) Excellent outcrossing characters. High exsertion rate (over 70%) and powerful viability of stigma, good outcrossing rate, early and concentrated flowering time, and the well close of lodicules and lemmas after pollination are required for the female parent. Luxuriant growth ability, large anthers, large quantity and vigor of pollen are required for the male parents. Strong general combining ability P(T)GMS with high combining ability is the base of strong heterosis hybrid rice varieties. The yield of two-line hybrid rice could not match the ideal yield for a long period. The most important reason is that the combining ability of P(T)GMS is weaker than CMS lines. Even some strong heterosis two-line hybrid rice varieties were benefited from good combining ability of restorers. Therefore, improving the combining ability of P(T)GMS must be emphasized. The genetic differences between the parents must be moderate The genetic differences between the P(T)GMS and the male must be moderate. However, what needs to be stressed in particular is that we can use only part of the heterosis between the indica and japonica subspecies, but not the full heterosis between typical indica and typical japonica rice or the ones including excessive indica or japonica ingredients. REFERENCES Chen L, Mei M H, Xu C G, Wang W, Cui H. 2000. Identification of AFLP-RFLP makers linked to the photoperiod-sensitive male sterile gene pms3 in rice. Xiamen Agri Univ (Nat Sci), 39(4): 421 425. (in Chinese with Chen L Y. 2001. Theory and technology of two-line hybrid rice. Shanghai: Shanghai science and technology Press, 102 105. (in Chinese) Chen L Y, Xiao Y H, Tang W B, Lei D Y. 2007. Practices and prospects of super hybrid rice breeding. Rice sci, 14(2): 71 77. He Q, Chen L Y, Deng H F, Tang W B, Xiao Y H, Yuang L P. 2007. Fertility photo-thermo characteristics in PTGMS Rice C815S and its homologous plant lines. Sci Agric Sin, 33(2): 262 268. (in Chinese with Li X H, Lu Q, Wang F L, Xu C G, Zhang Q. 2001. Separation of the two-locus inheritance of photoperiod sensitive genic male sterility in rice and precise mapping the pms3 locus. Euphytica, 119: 343 348. Li X H, Wang F L, Lu Q, Xu C G. 2002. Fine mapping of PSGMS Gene pms3 in rice (Oryza sativa L.). Sci Agric Sin, 28(3): 310 314. (in Chinese with Li Z Y, Lin X H, Xie Y F, Zhang D P. 1999. Tagging of a photoperiod-sensitive genic male sterile gene in Nongken 58S via molecular markers. Acta Bota Sin, 41(7): 731 735. (in Chinese with Lin X H, Yu G X, Zhang D P, Xie Y F, Qin F L. 1996. Location of one PGMS gene in Nongken 58S on Chromosome 5 of rice. Central china Agric Univ, 15(1): 1 5. (in Chinese with Lu Q, Li X H, Guo D, Xu C G, Zhang Q. 2005. Localization of pms3, a gene for photoperiod-sensitive genic male sterility, to a 28.4-kb DNA fragment. Mol Genet Genomics, 273: 507 511. Mei M H, Dai X K, Xu C G, Zhang Q. 1999. Mapping and genetic analysis of the genes for photoperiod-sensitive genic male sterility in rice using the original mutant Nongken 58S. Crop Sci, 39: 1711 1715. Qian Q, Zhu X D, Zend D L, Xiong Z M, Min S K. 1995. Linkage analysis of Hubei photoperiod-senstive genic male sterile locus in rice. Zhejiang Agric Sci, 7(6):429 433. (in Chinese with Shi M S. 1985. Study on the photoperiod sensitive genic male sterile lines in rice. Sci Agric Sin, 18(2): 44 48. (in Chinese with Shi M S, Deng J Y. 1986. The discovery, identification and utilization of Hubei photoperiodic sensitive nuclear male sterile rice. Acta Genet Sin, 13(2): 107 112. (in Chinese with Tang W B, He Q, Xiao Y H. 2004. Heterosis analysis of the combinations with dual-purpose genic male sterile rice C815S. J Hunan Agric Univ (Nat Sci), 30(6): 499 502. (in Chinese with Wang B, Xu W W, Wang J Z, Wu W, Zheng H G, Yang Z Y, Ray J D, Nguyen H T. 1995. Tagging and mapping the thermo-sensitive genic male-sterile gene in rice (Oryza sativa L.) with molecular markers. Theor Appl Genet, 91: 1111 1114. Wang F P, Mei M H, Xu C G, Zhang Q F. 1997. Pms1 is not the locus of relevant to fertility difference between the photoperiod-sensitive male sterile rice Nongken 58S and normal rice Nongken 58S. Acta Bot Sin, 39(10): 922 925. (in Chinese with Wang J Z, Wang B, Xu Q F, Yang D C, Zhu Y G. 1995. Tagging rice photoperiod-sensitive genic male sterile gene by RAPD analysis. Acta Genet Sin, 22(1): 53 58. (in Chinese with Wu X Y, Wan B H. 1991. Study on the genetic relationships between photoperiod-temperature sensitive genic male

CHEN Li-yun et al. Mechanism of Sterility and Breeding Strategies for Photoperiod/Thermo- Sensitive Genic Male Sterile Rice sterility genes and isozyme genes and its linkage measurement in rice. South China Agric Univ, 12(1):1 9. (in Chinese with Xie G S, Zhang D P, Xie Y F. 2000. Study of the number and allelism of restoring genes in photrperiod sensitive genic male sterile rice. J Wuhan Bot Res, 18(6): 449 453. (in Chinese with Xue G X, Deng J Y. 1991. Studies on the Hubei photoperiodic sensitive nuclear male sterile rice (Oryza sativa L. subsp. japonica) modification of minorgenes on the phenotypic values of progenies of photoperiodic sensitive male sterile rice. Acta Genet Sin, 18(1): 59 66. (in Chinese with English abstract) Xue G Y, Deng G Y, Zhao J Z, Chen P, Shen Y Z. 1995. Genetical studies on photo-sensitive genic male sterility (PGMS) of rice polymorphism of fertility distribution among different F 2 segregated populations. Sci Agric Sin, 28(1): 33 41. (in Chinese with Yang Z P. 1997. Inheritance of photoperiod sensitive genic male sterility and breeding of photoperiod sensitive genic male-sterile lines in rice (Oryza sativa L.) through anther culture. Euphytica, 94: 93 99. Yuan L P. 1987. Prospects of h ybrid rice breeding. Hybrid Rice, (1): 1 3. (in Chinese) Yuan S C, Zhang Z G, Xu C Z. 1988. Studies on the critical stage of fertility change induced by light and its phase development in HPGMR. Sci Agric Sin, 14(1): 8 9. (in Chinese with Zhang D P, Deng X A, YU G X, Lin X H, Xie Y F, Li Z B. 1990. Chromosomal location of the photoperiod sensitive male genic srerile gene in Nongken 58S. Central china Agric Univ, 9(4):407 419. (in Chinese with Zhang Q, Shen B Z, Dai X K, Mei M H, Saghai Maroof M A, Li Z B. 1994. Using bulked extremes and recessive class to map gene for photoperiod-sensitive genic male sterility in rice. Proc Natl Acad Sci USA, 99(8): 8875 8679. Zhang X G, Zhu Y G. 1991. Genetic analysis of the sterility of Hubei PGMS. Hereditas, 13(3): 1 3. (in Chinese)