Chncn Lio ,Wei Yn ,,Zhufeng Chen ,Gng Xie ,Xing Wng Deng ,c,*,Xioyn Tng ,,*

a Guangdong Provincial Key Laboratory of Biotechnology for Plant Development,School of Life Sciences,South China Normal University,Guangzhou 510631,Guangdong,China

b Shenzhen Institute of Molecular Crop Design,Shenzhen 518107,China

c Institute of Plant and Food Sciences,Department of Biology,Southern University of Science and Technology,Shenzhen 518055,China

Keywords:Hybrid rice Nuclear male sterile line The third-generation hybrid rice technology Pollen inactivation Seed sorting

ABSTRACT The breeding and large-scale application of hybrid rice contribute significantly to the food supply worldwide.Currently,hybrid seed production uses cytoplasmic male sterile(CMS)lines or photoperiod/thermo-sensitive genic male sterile(PTGMS)lines as female parent.Despite huge successes,both systems have intrinsic problems.CMS systems are mainly restricted by the narrow restorer resources that make it difficult to breed superior hybrids,while PTGMS systems are limited by conditional sterility of the male sterile lines that makes the propagation of both PTGMS seeds and hybrid seeds vulnerable to unpredictable climate changes.Recessive nuclear male sterile(NMS)lines insensitive to environmental conditions are widely distributed and are ideal for hybrid rice breeding and production,but the lack of effective ways to propagate the pure NMS lines in a large scale renders it impossible to use them for hybrid rice production.The development of‘‘the third-generation hybrid rice technology”enables efficient propagation of the pure NMS lines in commercial scale.This paper discusses the establishment of‘‘the thirdgeneration hybrid rice technology”and further innovations.This new technology breaks the limitations of CMS and PTGMS systems and will bring a big leap forward in hybrid rice production.

1.Introduction

Heterosis indicates the phenomenon that the F1hybrid between two parental lines of different genetic backgrounds is often superior to the parental lines in stress tolerance,disease resistance,adaptability,growth,or yield.Heterosis is widely present in the biological world and is the foundation of hybrid crops.The utilization of heterosis has been proved to be highly fruitful in crop improvement.Maize was the first commercialized hybrid crop.Subsequently,hybrid production systems have been developed in crops such as rice,sorghum,rapeseed,and many others[1].

Rice is an important food crop in the world.The successful deployment of hybrid rice is a significant achievement in crop production.The hybrid rice breeding was initiated in China in the 1970s,and now hybrid rice is planted in most of the rice growing countries[2-5].Cultivation of hybrid rice leads to a great improvement in rice productivity.In general,it increases the grain yield by over 20%compared to the inbred rice varieties[2-5].The hybrid varieties are generated through either the‘‘three-line system”or the‘‘two-line system”[2-5].The core of these systems is the large-scale propagation of pure male sterile seeds that can be used for the large-scale production of hybrid seeds.

The‘‘three-line system”,which is also known as‘‘the firstgeneration hybrid rice technology”,consists of a CMS line,a maintainer line,and a restorer line(Fig.1A)[5,6].The CMS lines are usually caused by abnormal mitochondrial genes encoding cytotoxic proteins[2,6].There are at least seven CMS genes identified in rice,all of them encoding the mitochondrial proteins[2,6].Three CMS genes,including CMS-WA,CMS-HL,and CMS-BT,have been used in the breeding of hybrid rice[2,6].CMS-WA was originally discovered in a wild rice in Hainan Island in 1970[2,6].It encodes a mitochondrial protein that interacts with the nuclearencoded mitochondrial transmembrane protein OsCOX11 to induce male sterility[7].Most of the indica CMS lines currently commercialized in hybrid production use the CMS-WA gene[2-6].The maintainer line has the same nuclear genes as the CMS line,but it does not have the abnormal mitochondrial gene[6].The maintainer line has normal fertility and can self-pollinate to reproduce itself[6].Cross-pollination of the CMS line by the maintainer line can propagate the CMS line[2-6].The restorer line contains specific restorer of fertility(Rf)gene(s)that can inhibit the function of thecorresponding CMS gene by acting on the RNAor protein of the CMS gene[6].For example,CMS-WA can be specifically inhibited by Rf3 and Rf4 genes,CMS-HL can be inhibited by Rf5 and Rf6,and CMS-BT can be inhibited by Rf1[7-12].Because the CMS gene function is suppressed by the Rf gene,the fertility is restored in the hybrid between the CMS line and the restorer line[6-12].The first hybrid rice Nanyou-2 was derived from CMS-WA and was released in 1973.Since then,the CMS hybrid varieties had been quickly deployed for commercial production in China because of their obvious yield advantages over the inbred varieties[2,5,13].Despite the huge success,CMS systems suffer from several intrinsic problems.The major problem is that the Rf genes exist in only 2%-5%of rice germplasms,therefore only a very small number of rice germplasms can be explored as restorer lines for heterosis[2-3,13-14].Besides,it is also tedious to breed new restorer lines.Because the restorer line needs to carry the very specific Rf genes for the CMS genes,breeders not only need to select for the desired traits for plant development and stress tolerance,etc.,but also need to make sure not losing the Rf genes and the combining ability from generation to generation,and these increase very much the workload and difficulties during breeding.In addition,it is also difficult to breed a new CMS line because it requires a maintainer line for propagation[2-4,13-14].Thus breeders have to breed all the desired traits into the maintainer line and then change the cytoplasm through several generations of crossing.Furthermore,CMS genes have negative impact on hybrid performance and are unstable under certain environmental conditions[2-4,12-15].Because of these problems,it is difficult to obtain CMS hybrids of strong heterosis.This is believed to be the main reason that CMS varieties stayed stagnant in yield increase in the last 20 years[2-4,14].

Fig.1.Diagram of the first-generation and the second-generation hybrid technology in rice.(A)‘‘The first-generation hybrid rice technology”consists of a cytoplasmic male sterile(CMS)line,a maintainer line,and a restorer line.The maintainer line and restorer line self-pollinate to reproduce themselves.Crosspollination of the CMS line by the maintainer line propagates the CMS line.Crosspollination of the CMS line by the restorer line propagates the F1 hybrids.(B)‘‘The second-generation hybrid rice technology”consists of a photoperiod/thermosensitive genic male sterile(PTGMS)line and a paternal line.The paternal line self-pollinates to reproduce itself.PTGMS line propagates itself through selfpollination under short day and/or low temperature restoring the male fertility,and outcrosses with the paternal line to produce the F1 hybrid under long day and/or high temperature inhibiting the male fertility.

The‘‘two-line system”,which is also known as‘‘the secondgeneration hybrid rice technology”,depends on the male sterile lines controlled by the recessive photoperiod/thermo-sensitive genic male sterile(PTGMS)genes(Fig.1B)[1-5].There are at least 13 PTGMS loci identified in rice,but only three PTGMS genes have been cloned[16].PMS1 and PMS3 encode phased small-interfering RNAs and long noncoding RNA,respectively,and both mutants are responsive to both photoperiod and temperature[17-19].TMS5 encodes RNase ZS1,and tms5 mutant is responsive mainly to temperature[20].Over 70%of commercial two-line hybrid rice cultivars in China were bred with the tms5-containing male sterile lines[20].Compared with the three-line systems,the PTGMS systems have the following advantages.First,the PTGMS systems do not need a maintainer line for propagation of the male sterile lines,which simplifies the breeding and production procedures.Because male fertility of the PTGMS lines is reversible in response to environmental conditions,PTGMS lines can propagate themselves through self-pollination under conditions restoring the male fertility.Under conditions that inhibit the male fertility,PTGMS lines outcross with paternal lines to produce hybrid seeds[1-4].Second,because the male sterility in PTGMS lines is controlled by recessive nuclear genes,they can cross with any plants carrying the wild type fertility gene to restore the fertility in hybrids.Therefore,most rice germplasms can be explored for breeding of hybrids of superior heterosis[1-4].In addition,the two-line systems are free of the cytoplasmic negative effects imposed by the abnormal mitochondrial genes.Because of these advantages,the PTGMS systems were quickly adopted for farming since the initial application in the 1990s,and hundreds of environment-sensitive genic male sterile lines and two-line hybrids have been released for commercial production[2,16,21].At present,two-line hybrids exceed threeline hybrids in cultivation acreage,and the highest yielding varieties currently cultivated in China are mostly PTGMS hybrids[16,21].Despite the huge success,the PTGMS systems also have intrinsic problems.Most of the commercial PTGMS lines are sterile when grown at long day with temperature above 25°C,but fertile when grown at short day with temperature below 23°C during the booting stage[2,21-22].To meet these requirements,strict environmental conditions must be used for propagation of the PTGMS seeds as well as production of hybrid seeds.The small window of the critical temperatures for fertility transformation(CTFT)also brings high risks to the production of PTGMS seeds and hybrid seeds under unpredictable weather conditions[2-3,21-22].For example,during propagation of the PTGMS seeds,an increase of environmental temperature can reduce the fertility of PTGMS lines and decrease the yield of PTGMS seeds,while during the hybrid seeds production,a drop in temperature can render the PTGMS line fertile,and self-pollination can cause impurity of the hybrid seeds[2-3,21-22].The failure in hybrid seeds production happened quite frequently in recent years in China,which brought huge losses[2,21].In addition,the CTFT in PTGMS lines often shifts up after a few generations of propagation[22].Because the increase in fertile temperature brings higher risk to hybrid seeds production,PTGMS individuals of suitable CTFT must be re-isolated repeatedly during production[3,22].Furthermore,the CTFT trait is influenced by minor QTL,and this significantly increases thedifficulty and uncertainty to breed the PTGMS genes into different genetic backgrounds using marker-assisted selection[16,22].

2.Establishment of the third-generation hybrid rice technology

Male sterility caused by nuclear genes that are insensitive to environmental conditions is common in flowering plants.However,commercial application of these mutants is limited because of the difficulty to propagate the pure male sterile lines in a large quantity.In 1990,Rao et al.[23]presented a number of ways for propagation and selection of nuclear male sterile lines in various crop species based on the genetic linkages between male sterility and special phenotypes such as leaf morphology,flower morphology,anther color,seedling color,endosperm color,and shrunken endosperm,etc.However,most of these methods have not been widely deployed for agricultural practices because of the late expression of the phenotype in plant development or the poor linkage between the male sterility and the phenotype.In 1993,Williams and Leemans[24]proposed an idea to obtain a transgenic maintainer by transforming a fertility-restoration gene linked with a pollen-lethality gene.In 2002,Perez-Prat and van Lookeren Campagne[25]discussed two strategies to obtain maintainer lines and pure nuclear male sterile lines.One strategy is to transform the fertility restoration gene linked with a seed-color gene into the male sterile plant to obtain a color maintainer.Cross-pollination of the color-maintainer to the male sterile line would generate 50%of male sterile seeds and 50%of color maintainer seeds that can be separated based on the seed color.The other strategy is to transform the fertility-restoration gene linked with a pollen-lethality gene into the male sterile plant.Cross-pollination of the pollenlethality maintainer to the male sterile line would generate pure male sterile seeds.

Based on these ideas,DuPont-Pioneer devised Seed Production Technology(SPT)in 2006 to produce the transgenic maintainer line in maize by transforming the male sterile mutant ms45 with the wild type gene MS45 linked with the maizeα-amylase gene ZmAA1 to disrupt the transgenic pollen and the red florescence protein gene DsRed to mark the transgenic seed(Fig.2)[26].ZmAA1 was placed under the pollen-specific promoter PG47 to prevent the accumulation of starch in the transgenic pollen grains,which specifically inactivated the pollen grains carrying the transgene[26-27].DsRed was placed under the aleurone-specific LTP2 promoter to make the transgenic seeds produce red florescence[26-27].The hemizygous transgenic plant recovered the male fertility and produced two kinds of pollen grains,the viable pollen grains without the transgene and the nonviable pollen grains with the transgene,in 1:1 ratio[27].Because the transgene had no effect on the female organ,self-pollination of the transgenic plant produced two kinds of seeds,those with the transgene and those without the transgene,in 1:1 ratio[27].The transgenic seeds were labeled with red florescence because of the function of DsRed gene,and the non-transgenic seeds were color-free.The two kinds of seeds could be mechanically separated based on the seed color.The non-transgenic seeds are the male sterile line,and the transgenic seeds are the maintainer line.Cross-pollination of the male sterile line by the maintainer produced the pure male sterile seeds in a large quantity[27].The SPT has been applied in hybrid corn production in the United States since 2012,and the produced hybrid corn was certified as non-GMO(genetically modified organisms)by regulatory agencies in the United States,Australia and Japan[27].Application of SPT in maize saves the costs for mechanical detasseling,which is the predominant method for commercial maize hybrid seed production.

Although the SPT technology was first developed in maize,it is fundamentally more useful for crops that have bisexual flowers not amenable to manual emasculation,including rice,wheat,sorghum,and rapeseeds,etc.In 2010,Wang et al.[28]adopted the same idea and transformed the rice MS26(OsCYP704B2)gene linked with the maizeα-amylase gene ZmAA1 under the pollen-specific promoter PG47 and the red florescence protein gene DsRed under the aleurone-specific LTP2 promoter into the oscyp704b2 male sterile mutant derived from Wuyungeng,a japonica rice variety[29-30].As expected,the resulted transgenic line carrying a single transgene insertion produced 1:1 florescent fertile seeds and non-florescent male sterile seeds,which verified the application potential of this technology in rice[28-29].However,because Wuyungeng is a japonica rice that is neither suitable for hybrid rice production nor a good material for commercial breeding of new male sterile lines,and the red florescence marker was not highly stable for seed sorting,the transgenic line was not further deployed for application[28].

In 2016,Chang et al.[31]published another work in rice by using the recessive osnp1 male sterile mutant derived from the indica rice Huanghuazhan(HHZ)and the corresponding wild type gene OsNP1 for construction of the nuclear male sterile(NMS)system.HHZ is an elite indica cultivar that is semi-dwarf with super high yield and good eating-quality and is widely cultivated in diverse geographical regions in China.It is a core germplasm for commercial rice breeding along Yangtze River and southern part of China[32].With the purpose to develop a practical NMS system for commercial breeding of hybrid varieties,the group started from constructing the HHZ mutant library and screening for the male sterile mutants that can be developed into a male sterile line of commercial value[31].From～300 male sterile mutants,they selected the osnp1 mutant for construction of the NMS system,because this mutant had normal vegetative growth,high stigma extrusions,high outcrossing rate,no pollen,and stable male sterility under diverse environment conditions[31].OsNP1 was cloned using the SIMMmethod based on the whole genome resequencing of the mutant plants[31,33].OsNP1 is a novel gene specifically expressed in the tapetum and microspores and is required for pollen exine formation[31].The tissue-specific expression pattern of OsNP1 further ensures that the osnp1 mutation affects only the male fertility but not any other developmental processes.By transforming the OsNP1 gene linked with ZmAA1 under PG47 promoter and DsRed gene under LTP2 promoter plus the 35S enhancer into osnp1 mutant,the team obtained a maintainer line named Zhen18B carrying one copy of the transgene[31].Zhen18B has no difference from the wild type HHZ plant in morphology,growth and development,and seed setting.As expected,self-pollination of Zhen18B propagated itself and the male sterile Zhen18A seeds in 1:1 ratio,while cross-pollination of Zhen18B to Zhen18A propagated pure Zhen18A seeds to a large quantity[31].The transgene in Zhen18B stayed stable in at least 12 generations of millions of plants tested by far.It also stayed stable when crossed into various other rice backgrounds during the breeding of new NMS lines[31].Approximately 85%of the hybrids between Zhen18A and other rice germplasms out-performed their parents in the per-plant yield[31].The seeds of Zhen18A were distributed to many rice breeders in China for test crossing.A number of hybrids with very high yield and excellent grain quality were obtained.In 2018,Zhen18A was officially certified to meet the national standard of commercialization in China;and in 2020,permit for production test was assigned to Zhen18B by the Genetically Modified Organisms Safety Committee in China(Table 1).These marked the successful establishment of the NMS system in rice,which was also called‘‘the thirdgeneration hybrid rice technology”(Fig.2)[29].

Compared with the CMS and PTGMS systems,the NMS system has several obvious advantages.First,since the male sterility phenotype is controlled by a recessive nuclear gene,any rice germplasm containing the wild type gene can restore the fertility ofthe hybrid.Thus,the NMS system allows broader choices of germplasms as paternal lines to breed hybrids of superior heterosis.Second,the male sterility of the NMS line is insensitive to environmental conditions,thus both the male sterile line and the hybrid seeds can be propagated under regular farming conditions.This significantly lowers the demand on specific environmental conditions for seed production,which are often difficult and costly to satisfy with the PTGMS systems.In addition,it can also reduce the risk induced by unpredictable weather changes on hybrid seed production.Third,the male sterility and fertility restoration are each controlled by a single genetic locus,and both traits can be stably inherited in different genetic backgrounds.Thus,both loci can be easily crossed into other cultivars to generate new maintainers and male sterile lines using marker-assisted breeding.Fourth,because the fertility restoration gene is closely linked to the pollen-killer gene,it blocks the transmission of transgenic components into environment through pollen,which significantly increases the environmental safety.Finally,although the technology involves transgenics,only the maintainer line carries the transgene.Both the male sterile seeds and the hybrid seeds are non-transgenic.Thus,transgenic oversight is applicable only to the maintainer line,which requires only a small acreage for cultivation.The production of hybrid seeds and cultivation of hybrid rice do not involve transgenics and thus do not require transgenic oversight.

Fig.2.Diagram of the third-generation hybrid technology in rice.

Table 1Summary of NMS lines derived from the OsNP1-maintainer.

3.Innovations to the third-generation hybrid rice technology

Zhen18B was the first NMS system constructed in rice.Although the resulting Zhen18A line displayed good application potential,there are still traits requiring improvements,such asdisease resistance,stress tolerance,and particularly the yield trait following cross-pollination and the combining ability to produce super heterosis.More importantly,as a technology with a big application potential,genetic diversity and technical perfection of the NMS systems are both necessary to ensure the wide and sustainable application in large scales.Following the publication of Zhen18B in 2016,a number of valuable innovations to the NMS systems have been published.Some of these innovations have been done in maize,but they are also valuable to the improvement of the NMS technology in rice.

3.1.Innovations of the pollen-inactivation function

For NMS technology,inactivation of the transgenic pollen grains is critical,because it not only ensures propagation of pure male sterile seeds during cross-pollination of the maintainer line to the male sterile line,but also prevents the transmission of transgenic components to the environment.The ZmAA1 gene under PG47 promoter in the maize SPT system and rice Zhen18B line can effectively kill the transgenic pollen grains,but crosspollination of the maintainers to the male sterile lines still showed transgene transmission rates of 0.002%-0.518%in different transgenic lines of the maize SPT system and 0.01%-0.08%in Zhen18B line growing at different seasons,based on the numbers of red florescent seeds produced[27,31].Therefore,there are still rooms for improvement of the pollen-killing function.

The deployment of new pollen-killer genes is an important innovation to the NMS technology.Besides ZmAA1,several other genes have been reported with pollen-killing functions in various plants,including a new riceα-amylase[34],the cytotoxin barnase[35-36],and the rice CMS gene orfH79[37].The new riceαamylase gene driven by the PG47 promoter is able to kill the transgenic pollen grains in rice[34].This result implicated that many otherα-amylase genes in other plant species may deliver the pollen-killing function in rice as well.The Barnase gene has been used for genetic engineering of male sterility in various plants either by disrupting the tapetum or pollen grains[38].Because barnase has strong cytotoxic activity,to reduce the impact of its leaky expression,people have tried to split the gene into two parts and then use two different pollen-specific promoters to drive the N-terminal part and C-terminal part in the same T-DNA,or to use two different pollen-specific promoters to drive barnase and its inhibitor gene barstar in the same T-DNA to inactivate the transgenic pollen grains[35-36].orfH79 encodes a mitochondrial protein inducing pollen sterility in HL-CMS rice[39].ORFH79 acts on the electron transport chain in mitochondria,causing abnormal energy supply and abnormal accumulation of reactive oxygen species,which leads to male gamete infertility[40].When the recombinant gene of ORFH79 fusion with the RF1b mitochondrial signal peptide was expressed in indica rice 93-11 under the PG47 promoter,the transgenic plant displayed the 1:1 ratio of normal and abnormal pollen grains,confirming the pollen-killing activity of the transgene[37].This pollen-killer gene was then linked with the seed-marker gene DsRed2 and the fertility-restoration gene OsCYP703A3 and transformed into 93-1103a3,a male sterile mutant of 93-11 with the OsCYP703A3 gene knocked out.A stable maintainer line was obtained and named 93-11-3B[37].The 93-11-3B plant displayed～1:1 ratio of normal pollen grains and abnormal pollen grains,and self-pollination of the 93-11-3B plant produced～1:1 segregation of seeds with red florescence and seeds without florescence,which further verified that orfH79 can be used as a pollen-killer gene in the NMS system[37].Cross-pollination of 93-11-3B to the male sterile lines showed the transgene transmission rates of 0.14%-0.17%,which was＞2-fold higher than that displayed by the ZmAA1 gene in Zhen18B[37].In addition to orfH79,many other CMS genes have been cloned from rice and several other plant species[6].Further exploration of these CMS genes may identify other genes that can be developed into pollenkilling tools for the construction of NMS systems in rice and other crops as well.

Inactivation of transgenic pollen requires a pollen-specific promoter to drive the pollen-killer gene at the late stage of pollen development.Thus,deployment of a very strong promoter for pollen-inactivation can also improve the NMS system in rice.A number of late-stage pollen-specific promoters have been identified from various plant species,however,only a few have been tested for pollen-inactivation thus far,including the maize PG47 promoter[26-28,31,35-37,41-42].The PG47 gene is specifically expressed in pollen grains at the stage of first mitosis to pollen maturity(stage 11-12)[43].To obtain more promoters suitable for construction of the NMS technology in rice,Wang et al.[44]conducted an RNA-seq analysis to identify genes that are specifically expressed in mature pollen grains.They then used qRT-PCR analysis to determine the relative RNA expression levels of these genes in anthers at different developmental stages.They tested six LSP(late-stage pollen-specific)promoters of genes that showed relatively high levels of RNA expression,and found that the top three promoters(OsLSP3,OsLSP5,and OsLSP6)very active at stages 11 and 12 could drive ZmAA1 to inactivate pollen in rice,while the promoters of relative lower activities could not.The promoter of OsLSP4,which showed higher gene expression at stage 12 but lower expression at stage 11,could not drive ZmAA1 to inactivate pollen,indicating that strong promoter activity at stage 11 was critical for pollen inactivation.The OsLSP3,OsLSP5,and OsLSP6 provide additional tools for genetic engineering of the rice NMS systems.It is expected that promoters stronger than these three gene promoters can produce even better pollen-killing results.

With more pollen-killer genes and late-stage pollen-specific promoters identified,combinational use of these components can produce even better outcome in pollen-inactivation,as recently demonstrated by the multi-control sterility(MCS)system in maize[45-46].The MCS system involves co-expression of two pollenkilling genes,ZmAA1 and DNA adenine methylase gene(Dam)in pollen.The ZmAA gene is placed under the promoter PG47 to prevent the starch accumulation in transgenic pollen grains.The Dam gene is constructed under the pollen-specific promoter Zm13 to catalyze the methylation of adenine residues in pollen DNA,which affects the cell viability of transgenic pollen[47-49].The MCS vector containing the fertility restoration gene,two pollen-killer genes,the red fluorescence seed-marker gene,and a herbicide resistance gene was transformed into the male sterile mutant.Compared with transgenic plants expressing only ZmAA1 under the PG47 promoter,the MCS transgenic plants showed 7～8-fold reduced rates of transgene transmission[45-46].This strategy is also expected to reduce the transgene transmission from the maintainer line into the male sterile line through pollination in rice.

3.2.Innovations to the seed sorting function

The NMS systems can propagate the male sterile seeds by selfpollination of the maintainer line.Because this also produces 50%maintainer seeds,a highly efficient seed sorting machine must be used for large scale production.Without an efficient seed sorting machine,male sterile seeds can be propagated by crosspollination of the male sterile lines by the maintainer lines,which would significantly reduce the demand on seed sorting to clean the contaminated maintainer seeds from transgene transmission.However,to achieve a maximal yield of the male sterile seeds,it is necessary to use pure maintainer lines for cross-pollination.The pure maintainer seeds can be obtained if a herbicide resistance gene is added to the transgenic cassette,so the male sterile seedscan be selectively killed by herbicide treatment(Fig.2).This strategy was tested by An et al.[50]who transformed the fertility restorer gene,pollen-killer gene,color selection gene,and the herbicide resistance Bar gene linked in one T-DNA cassette into the maize male sterile ms7 mutant.Because only the transgenic maintainer lines carry the herbicide resistance gene,the mixed male sterile lines were removed by herbicide treatment[50].This strategy is expected to work well in the rice NMS system as well.

3.3.Combination of CRISPR/CAS9 with the NMS technology to create the male sterile line and maintainer line in one step

To convert a rice material into a NMS maintainer line,cross the established NMS maintainer line with this material as the recurrent pollen donor is a good choice for the sake of simplified transgenic regulation.However,it would take at least four generations of backcrossing and then two generations of self-pollination to get a stable NMS maintainer line.This is a time-consuming and labor extensive process.Besides,the resulted NMS maintainer line may contain more or less the genetic materials from the donor transgenic parent.

To solve these problems,Qi et al.[51]proposed a strategy that combines the CRISPR/Cas9 gene editing technology with the NMS technology to create the male sterile line and maintainer line in one step(Fig.2).They constructed two plant transformation vectors.One vector contains the CRISPR/Cas9 gene editing tool targeting two non-coding regions of the maize sterility gene ZmMS26.The function of this vector is to delete an exon of ZmMS26 to create the male sterile mutation.The other vector contains the linked genes of ZmMS26 cDNA under the ZmMS26 native promoter,ZmAA1 gene under PG47 promoter,and DsRed gene under LTP2 promoter.The function of this cassette is to create the NMS maintainer for the zmms26 mutant.They co-transformed the two constructs into a normal maize line.The CRISPR/Cas9 gene editing tool created deletion in ZmMS26 gene causing male sterility,but the plant co-transformed with the NMS construct restored the fertility and set non-florescent male sterile seeds as well as the fertile florescent seeds.Fertile progeny carrying the homozygous zmms26 mutation and the NMS cassette but lacking the CRISPR/Cas9 cassette was identified by PCR as the NMS maintainer line,and the sterile progeny lacking the CRISPR/Cas9 cassette was the zmms26 male sterile line.This method greatly shortens the time for construction of NMS systems based on the known male sterile genes.

3.4.Construction of nuclear female sterile line of normal male fertility

Currently,commercial production of rice hybrid seeds is carried out by planting the two parental lines side-by-side in separate rows,and then using manual assistance to facilitate crosspollination.When pollination is finished,the paternal lines are removed manually to avoid contamination of the hybrid seeds by the paternal seeds.This process is very laborious and unsuitable for mechanized production of hybrid seeds,and it is the major contributor to the high price of hybrid seeds.If the paternal line is female sterile but with normal male fertility,then the two parental lines can be mix-planted for hybrid seed production,which would facilitate the mechanized production of hybrid seeds.

Xia et al.[52]attempted to innovate the technology for propagation of female sterile seeds by using the rice recessive ptb1 mutant and the corresponding wild type gene(Fig.2).PTB1(POLLEN TUBE BLOCKED 1)gene is required for pollen tube growth in transmitting track after pollen germination,and ptb1 mutant blocks the pollen tube growth,resulting in female sterility[53].The PTB1 was constructed together with the pollen-killer gene ZmAA1 and seed-marker gene DsRed,and the construct was transformed into the ptb1 mutant plant[52].The function of PTB1 restored the female fertility of the transgenic plant,but pollen grains carrying the transgene were inactivated by the pollenkiller gene.Self-pollination of the transgenic plant produced seeds carrying the transgene that are fertile with red florescence,and non-transgenic seeds that are female sterile without red florescence.The two kinds of seeds are sorted out based on the red florescence.The female sterile line can be used for hybrid seed production,and the transgenic line can be used as the maintainer line for propagation of the female sterile line through selfpollination.Although the work proved the concept,however,because the ptb1 mutant is not completely sterile and has a seed setting rate of～1.8%[53],this system cannot be used for commercial application.Further improvement of this technology should use a mutant of complete female sterility.

3.5.Synthetic apomixis as a strategy to fix the heterosis

Apomixis is an asexual reproduction process in which clonal seeds are produced without meiosis and fertilization[54].Through apomixis,a hybrid of strong heterosis can reproduce itself without chromosomal recombination and segregation,thus the heterosis can be passed on through generations[54].This method has enormous advantages over the hybrid systems because it needs neither the male sterile seeds nor cross-pollination for production of hybrid seeds.Apomixis does not exist naturally in crops such as rice,maize,wheat,etc.However,it was found recently that CRISPR knockout of three genes important for meiosis(PAIR1,REC8,and OSD1)can generate a MiMe(mitosis instead of meiosis)mutant with the meiosis replaced by a mitosis-like division[55,56].The MiMe mutant can generate diploid gametes identical to the mother genome.Further expression of the BBM1(BABY BOOM1)in egg cells or disruption of MATRILINEAL(MTL)in the MiMe background can induce the production of clonal seeds[55,56].By simultaneous engineering MiMe with egg-expression of BBM1 gene or disruption of MTL gene in the F1hybrid,synthetic apomixis can be established to produce the F1clonal seeds(Fig.3).Although this strategy is very attractive,currently it can only produce a few diploid seeds and thus is still at the research stage.

4.Future perspectives of the third-generation hybrid rice technology

The development of the NMS systems enables the use of recessive nuclear genes for hybrid rice breeding and production,which overcomes the problems of traditional CMS and PTGMS systems.This new technology is a significant breakthrough in the field of hybrid rice breeding and will bring a huge step forward in hybrid rice production.With the rapid development of molecular biology,more genes will be identified that can be used for further improvement and perfection of the NMS systems.Decades of efforts in traditional rice breeding have accumulated numerous excellent rice germplasms with superior traits in yield,disease resistance,stress tolerance,grain quality,and many others[57].These materials plus the abundant rice genomic data and tools for molecular markers assisted selection will enable the rapid breeding of new NMS lines with improved traits based on the established NMS systems through crossbreeding[57].

With the CRISPR/Cas9 technology available and many male sterile genes cloned,it is convenient for de novo construction of new NMS systems.Several factors should be considered in selecting rice materials and male sterile genes for de novo construction of NMS systems.First,it is better to use the elite rice lines amenable to transformation and with big stigma and high seed-setting rates upon cross-pollination,because these reproductive traits are necessary for the high yield during the production of hybridseeds and male sterile seeds[2-5].In addition,high combining ability is also required because this will allow the male sterile line to generate more superior hybrid varieties in combination with other germplasms[2-5].Besides,these excellent traits will also speed up the breeding process when the established NMS maintainer is further used for crossbreeding of new NMS lines[2-5].By far,at least 40 NMS genes have been cloned in rice[58-60].However,most of these genes were cloned from japonica varieties such as Nipponbare,Zhonghua 11,and Wuyungeng,that are easier for molecular genetic researches but not suitable for application as male sterile lines[58,59].Other than these genes,many recessive male sterile genes have been identified in other plant species such as maize and Arabidopsis[45,58,61].The homologues of these genes can be tested in rice for the roles in male fertility as well.It should be noted that many male sterility genes are expressed not only in anthers but also in other tissues[58-61].To avoid the potential negative impact of the mutation on male sterile lines under natural conditions,it is better to select the genes that are specifically expressed in anthers but not in any other tissues.In addition,the male sterility must be stable under different environmental conditions.

Fig.3.Diagram of synthetic apomixis in rice.

By far,two rice NMS systems have been reported with satisfying results,one is the OsNP1-based Zhen18B constructed by Chang et al.in 2016[31],the other is the OsCYP703A3-based 93-11-3B constructed by Song et al.in 2020[37].Zhen18B is at the production testing stage according to the transgenic regulation policy in China.Using Zhen18B as the donor parent,two stable NMS maintainer lines with improved cross-pollination yield,Zhen20B and Zhen21B have been bred through crossbreeding and selection.The corresponding male sterile lines Zhen20A and Zhen21A have also been certified to meet the national standard of commercialization in China(Table 1).In addition,crossbreeding of Zhen18B to an elite inbreed line and two elite PTGMS lines also generated several other new maintainer lines with improved traits,and their corresponding male sterile lines will soon be certified(Table 1).Test crossing of these NMS lines with many different rice germplasms all showed strong combining abilities(Table 1).Several hybrids derived from the OsNP1-based NMS maintainers are in the national variety approval test in China.The male sterile line 93-11-3A derived from 93-11-3B was also used in test crossing[37].One hybrid line‘‘Sanyou No.1”derived from 93-11-3A displayed super high yield in a field test in 2020 in Hunan province,China(http://www.hunan.gov.cn/hnszf/hdjl/xwfbhhd/wqhg/202011/t20201104_13950556.html).These initial results demonstrated an enormous application potential of the NMS systems in hybrid rice breeding and production.

The first two generations of hybrid rice technologies have contributed significantly to food security in China since their deployment.However,with the improvement of inbred rice breeding,hybrid rice gradually loses advantages to inbred rice in the past 20 years,as demonstrated by the gradual decline of hybrid rice planting area from the historical 65%to 45%in year 2019[62].The reason for acreage reduction was mostly due to the stagnant yield improvement,extensive labor requirement for production,high price of hybrid seeds,and poor eating quality of many hybrid varieties,which make hybrid rice unprofitable and unsuitable for mechanized and labor-light cultivation.These problems all root on the limitations of the traditional male sterile systems that slow down the progress in breeding of super parental lines.Overcoming these problems requires technical breakthroughs to break the limitations imposed by the traditional technologies and continuing improvements of the parental lines for high seed yield,suitable for simple and mechanized production of hybrid seeds,strong tolerance to biotic and abiotic stresses,superior combining ability,good grain quality,herbicide tolerance,and suitable for direct seeding and mechanized grain production.With the successful establishment of the NMS technology,it is anticipated that many more superior NMS maintainer lines of different genetic backgrounds will soon be created through crossbreeding or de novo construction.The wide application of the third-generation hybrid rice technology will overcome the limitations imposed by the traditional methods and make the hybrid rice production more profitable.The NMS technology is suitable not only for rice but also for other crop species.A recent publication reported the establishment of this technology in tomato by transforming a male-sterile plant with the fertility-restoration gene linked with a seedling-color gene inducing pigmentation in the seedlings[63].With the rapid development of plant genomics and molecular biology,the NMS systems are likely to be developed in many other crops in the near future.The wide application of the NMS technology in a variety of crops will bring a huge elevation in crop production.

CRediT authorship contribution statement

Chancan Liao,Zhufeng Chen,and Xiaoyan Tangwrote the paper.Gang Xieprepared Table 1.Wei Yan and Xing Wang Dengreviewed and edited the manuscript.All authors read and approved the manuscript.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

We are grateful to colleagues in Shenzhen Institute of Molecular Crop Design for sharing the unpublished results of Zhen18B and other OsNP1-based NMS lines.This work was supported by theNational Natural Science Foundation of China(U1901203),Natural Science Foundation of Guangdong Province(2018B030308008 and 2019A1515110671),Major Program of Guangdong Basic and Applied Research(2019B030302006),Shenzhen Commission on Innovation and Technology Programs(JCYJ20180507181837997),and China Postdoctoral Science Foundation(2019M662957).