Fertility restoration for cytoplasmic male sterility (CMS)
The annual planting area of the FBZ-series hybrids accounts for 35% of the hybrid rice area in China (Fig. 2b). Therefore, the ability to restore the fertility of CMS is very important for the application of FBZ-series varieties to hybrid rice system. Three-line hybrid rice system comprises CMS line, CMS maintainer line, and CMS restorer line. CMS-WA was derived from a male sterile plant of wild rice O. rufipogon discovered by team of Prof. Longping Yuan, and it is the most widely used type of CMS lines in China. Currently, three CMS systems are mainly used, including Boro II CMS (CMS-BT), Wild Abortive CMS (CMS-WA) and Hong-Lian CMS (CMS-HL). CMS is controlled by interactions between CMS genes and nucleus-encoded restorer of fertility (Rf) genes (Chen et al. 2021). In the CMS-HL system, the fertility restoration is controlled by Rf5/Rf1a and Rf6, while CMS-BT line carries Rf5/Rf1a and Rf1b. Fertility restoration of CMS-WA is controlled by two dominant nuclear loci, Rf3 and Rf4 (Tang et al. 2014). It should be noted that the fertility restoration QTL Rf3 has not yet been cloned, and the causative variant (Chr1P5568692) provided by the previous study (Wei et al. 2021) should be a linkage SNP. In this study, we found that landraces and CMS-WA carried the ‘T’ allele at this locus, by contrast, the ‘C’ allele was present in the varieties after semi-dwarf breeding (Additional file 2: Table S6), suggesting that Rf3 is likely has been selected in the process of semi-dwarf breeding. In particular, both CMS-WA and iconic restorer lines carried ‘G’ allele at the causative variation site (Chr10P19058375) of Rf1b, however, the ‘A’ allele was present in FBZ-series varieties. The possible explanation is that Rf4, Rf1a and Rf1b are closely linked genes, the favorable alleles of Rf1a and Rf1b were selected along with the application of Rf4. Since Rf4, Rf1a and Rf1b all encode PPR proteins (Chen et al. 2021), whether they have an additive effect in restoring CMS fertility remains unclear. In general, the favorable alleles of Rf3 and Rf4 were selected in the early period of the FBZ pedigree, making most of the FBZ-series varieties can restore the fertility of CMS-WA.
Fitness trade-off between defense and growth
Biotic and abiotic stresses are key factors threatening rice growth, leading to decreases in yield and quality. In this study, we found that the selection of Hd1-Pi2-Pid3-Pid2 locus significantly improved the rice blast resistance, which directly contributed to the breeding of FBZ-series varieties (Additional file 1: Figure S4). Moreover, among 43 QTGs involved in semi-dwarf breeding and hybrid rice breeding, 19 QTGs were associated with resistance. However, by comparing the variant information of these QTGs, we found that only 3 QTGs (Pi21, Xa13, Xa21) are beneficial genes for FBZ-series varieties, while the gene effects of 8 QTG locus (Bph9, Xa27, Xa23, OsPsbS1, Pb1, Pik-1/Pik-2, PSTOL1, SNORKEL1/SNORKEL2) cannot be determined due to lack of functional variation information (Additional file 2: Table S6). On the contrary, 8 out of 19 QTGs showed unfavorable genetic effects in FBZ-series varieties, which may cause susceptibility to sheath blight (Xa1, Xa25), UV-B (OsUGT707A2), herbicides (CYP72A31), drought (DRO1), low temperature (qLTG3-1), and nutrient deficiency (OsHKT2;1, ARE1) (Additional file 2: Table S6). We are curious why most of the resistance-related genes play an unfavorable function in semi-dwarf breeding and hybrid rice breeding. One possible explanation is that the adverse effects of these genes are supplemented by other favorable genes. Another possibility is that the adverse effects of these genes were obscured in growth environment without stress. For instance, DRO1-kp allele enables rice to produce more grains under drought-induced stress. However, under non-drought conditions, it could not show a yield advantage because DRO1 alters only root growth angle and does not decrease either shoot or root biomass (Uga et al. 2013). On the other hand, it is well known that plant immune activation often leads to fitness cost, this makes plants trade-off between plant growth and immunity to ensure the best use of limited resources (Chen et al. 2021; Wang et al. 2021). For example, high-quality rice varieties usually have lower yields and poor resistance, such as KDML105, Pusa Basmati 1121 and Koshihikari (Kobayashi et al. 2018; Singh et al. 2018; Vanavichit et al. 2018). Recently, several genes (ipa1-1D, Pigm, OsGRF6, WRKY45) has been reported fine-tunes the balance between disease resistance and grain yield in rice (Wang et al. 2021). However, for most resistance genes, their adverse effects on yield or quality remains unclear. Given that semi-dwarf breeding and hybrid rice breeding are mainly aimed at improving yield and rice quality, we speculate that the selection of unfavorable alleles of these resistance genes may contribute to rice growth. To answer this question, it is essential to further elucidate molecular mechanism and coordination of complex agronomic traits, especially the relationship and balance among grain yield, grain quality, immunity, nutrient use efficiency, and stress tolerance.
Foreign germplasm resources contributed to breeding progress
Guangdong is the province with the largest consumption of high-quality rice in China, which makes many varieties with good quality have been developed before FBZ was bred, such as Mabayouzhan and Zhengchengsimiao. However, before FBZ was bred, no varieties could have both excellent rice quality and blast resistance (Zhou et al. 2007). The lower amylose content is the basic requirement of high-quality rice. In this study, 28zhan carries the same genotype as most FBZ-series varieties at locus Chr6:1.6–23.3 Mb (Additional file 2: Table S7), indicating that the low amylose content and blast resistance of FBZ were inherited from its parent, 28zhan. 28zhan was derived from the cross of Qingxiangzao301×IR37704-131-2-1-3-2. Since two parents of 28zhan are all have low amylose content, it is difficult to determine the parent source of low amylose content traits through pedigree tracing. Previous studies have shown that compared with Nipponbare, there exist polymorphisms in the 5’UTR region of Huanghuazhan (Zhou 2015). Our research showed that FBZ carried T5 haplotype of Waxy gene, which was different from Wxb in the coding region (Additional file 1: Figure S1a). We noticed that the T5 haplotype is a low-frequency variant according to MBKbase database (Peng et al. 2020) statistics. Given IR37704-131-2-1-3-2 showed good rice quality and resistant to rice blast in our field trial (Zhou et al. 2007), we speculate that the favorable allele Pi2 and Waxy of 28zhan were inherited from IR37704-131-2-1-3-2. In the history of rice blast resistance breeding in Guangdong province, three resistance sources were mainly used, including Waixuan35, Qingliuai1 and Gengxian89, however, the linkage drag between rice quality and blast resistance limited their breeding applications (Zhu et al. 2003). Therefore, International Rice Institute variety IR37704-131-2-1-3-2 provides a very rare germplasm resource for rice blast resistance breeding. Aizizhan was the core parent of semi-dwarf breeding in China, and it was introduced to China before 1949. Recent studies have shown that most modern indica rice varieties carry the same sd1 allele as DGWG (Chen et al. 2020; Wang et al. 2020), supporting the view that Aizizhan was introduced from southeast Asian countries. In addition, foreign germplasm resources are also the main donors of restorer genes for wild abortive hybrid rice systems (Xie and Zhang 2018). We think that foreign germplasm resources carry very precious beneficial alleles, or allele combinations, which is an important reason why they can be applied to rice breeding in China.
Rice core germplasm breeding theory
The breeding of the FBZ-series varieties shows that the ‘Rice Core Germplasm Breeding Theory’ (Zhou and Ke 1998) has achieved remarkable success. Based on this theory, the ideal gene system will appear in the future, and the improvement of rice varieties is essentially a trajectory of gradual optimization from the original gene system to the ideal gene system.
Rice is a daily staple food crop for over 10 billion people, although breeding goals will be different in various periods and regions, the demand for high-yield, good-quality, multi-resistance, and mechanization will last a long time. In a certain period of economic and population development, regional development will be more balanced, and the demand for rice varieties will tend to be consistent. There will be a combination of genotypes that can adapt to this universal demand, and we define it as an ideal gene system. In this study, varieties derived from FBZ have been widely planted in all indica rice growing areas of China (Fig. 2a), indicating that broad adaptability is not the main limiting factor in the construction of ideal gene systems. In addition, 64 of 83 varieties derived from FBZ share the same genotype at locus Chr6:1.6–23.3 Mb (Fig. 4a), suggesting that the demand for rice breeding in recent years has begun to converge.
In this study, favorable genes were gradually optimized during the breeding of FBZ-series varieties (Fig. 5), which is consistent with the conjecture of the "Rice Core Germplasm Breeding Theory". Since Qingliuai1 was bred in 1990, it took our team fifteen years to bred the elite variety Huanghuazhan. We think it is incredible to breed a variety with an ideal genetic system through a single breeding event, mainly due to the following three reasons. (i) Favorable genes are usually distributed in different countries and subgroups, it is hardly to screen a germplasm resource that carries multiple favorable genes. For example, rice bacterial blight disease resistance gene Xa21 was derived from Oryza longistaminata in West Africa (Song et al. 1995); (ii) Many adjacent genes tended to co-occur within the same interval of 2cM , which may lead to genetic drag, and the introgression of superior alleles at some genes may also introduce inferior alleles at linked loci during pedigree breeding (Wei et al. 2021). For example, blast resistance gene Pi21 is tight linkage with genes that cause poor eating quality (Fukuoka et al. 2009), while chalkiness QTL Chalk5 is linked to grain width effect QTL GW5 (Li et al. 2014); (iii) The breeding goal is continuously optimized according to cultivation technology and market demand. For instance, before chemical fertilizers were widely used in production, the semi-dwarf sd1 allele was an unfavorable gene due to its low nitrogen efficiency (Li et al. 2018). In addition, China’s demand for high-quality rice has only been increasing since the 1980s, which makes slender (qSW5, GS3), non-chalky (Chalk5), low amylose content (Waxy), and fragrance (BADH2) alleles become beneficial genes.
The renovation of Chinese indica rice varieties
In the past several decades, indica rice breeding has experienced three important technological advances in China, namely semi-dwarf breeding, three-line hybrid rice breeding and two-line hybrid rice breeding (Qian et al. 2016). In 1955, Prof. Yaoxiang Huang introduced the semi-dwarf variety Aizizhan from Guangxi Province, China. This allowed Guangchangai to be bred in 1959, which was 7 years earlier than the bred of the miracle rice IR8 (Qian et al. 2016). During the semi-dwarf breeding period, Chinese breeders bred many elite varieties, including Zhenzhuai, Shuanggui1, Guangluai4, Guichao2 and Teqing (Zeng 2018). The earliest male parents of the three-line hybrid rice were mostly imported from the International Rice Research Institute (IRRI), including Taiying 1, IR24, IR661, IR26 and IR36. Minghui63 was one of the most successfully developed male parents and played a key role for the promotion of hybrid rice production in China from the late 1980s (Xie and Zhang 2018). The widespread use of two-line hybrid rice in China began in 2000s, and 9311 was the most representative restorer lines in this period. Using 9311 and its derivative lines as restorer lines, many well-known hybrid rice varieties have been bred, including Liangyoupeijiu, Yangliangyou 6, Fengliangyou 1, and Shenliangyou 5814 (Zeng 2018). From 2010 to 2015, both conventional rice and restorer lines have undergone renovation. Huanghuazhan, Xiangzaoxian45 and Zhongjiazao17 were the three rice varieties with the largest annual promotion area of conventional rice, while the most widely used restorer lines became Huazhan, Bing4114 and Guanghui308 (Zeng 2018). In addition, recent studies have also shown that Huanghuazhan and Huazhan have now become important backbone parents or main recommended varieties of indica rice in China (E et al. 2019; Fang et al. 2020).
In this study, Huazhan and Wushansimiao were the two restorer lines with the largest annual promotion area of hybrid rice (Additional file 1: Table S2). The hybrids derived from FBZ were grown with area of 3.8 million hectares, accounting for 35% of the total hybrid rice area in the country (Fig. 2b). It should be pointed out that Shanyou63, a milestone for China’s hybrid rice development, had a large planting area during the years from 1985 to 2001 with an average area of 3.6 million hectares and 28.3% of the national hybrid rice growing areas annually (Xie and Zhang 2018). This shows that the annual planting area of FBZ-derived hybrids has exceeded Shanyou63. Moreover, Huanghuazhan has been grown in China on a cumulative 3.4 million hectares from 1982 to 2019 (Additional file 1: Table S1), and is currently the high-quality indica conventional rice variety with the largest cumulative planting area in China. Hence, FBZ has become a milestone of Chinese modern indica rice, it’s derived conventional rice and hybrids have greatly promoted the replacement of rice varieties in Southern China.
Perspective for breeding next generation rice
Under the guidance of ‘Rice Core Germplasm Breeding Theory’ (Zhou and Ke 1998), we have not only bred the FBZ-series varieties such as Huanghuazhan, Huazhan and Wushansimiao, but also the iconic aromatic rice variety Meixiangzhan2. Meixiangzhan2 was the only indica rice variety that has won the three gold medals in the national committee of evaluation on eating quality of high-quality rice varieties, and its taste quality has even surpassed that of Thai Hom Mali Rice KDML105. Meixianzhan2 was released in 2006 and has been widely planted in China, with an annual promotion area of about 133,000 hectares (NATESC 2019). It has also been introduced to Myanmar, Vietnam, Laos, Thailand, Mozambique and other countries for cultivation (Li et al. 2021). As mentioned above, we believe that the ideal gene system will appear in the future. We are thinking about how to breed a variety that has the similar yield levels, disease resistance and combining ability of FBZ-series varieties, and the same eating quality as Meixiangzhan2. In order to achieve this goal, using molecular marker-assisted selection technology to manipulate some favorable genes, such as Pi2, Waxy, Rf3, Rf4, BADH2, will greatly improve breeding efficiency. In addition, it is necessary to introduce more germplasm resources, especially tropical or temperate japonica rice. This may help us discover more favorable alleles or combinations of favorable genes, thereby creating a new balance between yield, quality, and resistance.
Rice is the staple food for over half the world’s population and approximately 90% of the world’s rice is grown in Asia (Muthayya et al. 2014). China is the largest rice producer in the world, the annual production has stabilized at more than 200 million tons for 10 consecutive years (National Bureau of Statistics). On the contrary, China’s rice planting area is only about half of India’s, indicating that the high yield per unit largely contributes to China's rice production. Except for China, there are six countries in Asia with rice planting areas more than 5 million ha, including India, Bangladesh, Indonesia, Thailand, Vietnam and Myanmar. The total rice planting area in these countries is about 90 million ha (FAO, Food and Agriculture Organization of the United Nations), which is about four times that of China. The current farm yield has reached about 7 metric tons per hectare (t ha-1) in China (National Bureau of Statistics), however, the farm yields in these countries varied from 2.9 to 5.1 t ha-1 (FAO 2019), suggesting there is still a lot of room for improvement in potential yield. With the initiative and funding of ‘Green Super Rice’ (GSR) project, Huanghuazhan was released in Mozambique, Indonesia and India, another 15 varieties derived from Huanghuazhan were also released in some countries in Asia and Africa. Those FBZ-series varieties and other GSR varieties have been promoted in Africa, South Asian and Southeast Asian countries about 6.12 million hectares, which marked significant contributions to rice production and food security in these countries (Wang et al. 2018a; Yu et al. 2020; Yu et al. 2022)( Additional file 1: Table S3). It is estimated that the world food production needs to increase by 70% in 2050, driven primarily by population growth (FAO 2013). We are very willing to help these neighboring countries increase food production, because this is a win-win situation. China is advancing the ‘One Belt, One Road’ strategy. Scientists should try their best to help them increase the yield levels of rice, providing them with theoretical guidance, technical support, extension services, germplasm resources, and modern varieties.