Hong Go,Zuoin M,Yunzheng Wng,Mnli Zhng,Xinju Wng,Chnghu Wng,Zhiqing Tng,Liying Zhng,Ling Fu,N He,Hui Wng,Yongn Yin,Yunjun Bi,Guomin Sui,*,Wenjing Zheng,*

a Liaoning Rice Research Institute,Shenyang 110101,Liaoning,China

b Liaoning Academy of Agricultural Sciences,Shenyang 110161,Liaoning,China

c Xinjiang Tianye Group Ltd.,Shihezi 832000,Xinjiang,China

Keywords:Japonica hybrid rice Genetic distance Indica-genotype proportion Heterosis

ABSTRACT The relationship between parental genetic differences and the quality and yield of japonica hybrid rice strongly influences japonica hybrid rice breeding.In this study,137 parental lines of japonica hybrid rice were genotyped using 8K rice SNP-Chips to characterize their genetic diversity,population structure,and indica-genotype proportion.The genetic diversity of total parental lines averaged 0.264,with values of 0.287 for restorer lines and 0.148 for the sterile lines.The introduction of indica lineage increased the genetic diversity of restorer lines relative to that of sterile lines.By model-based population structure analysis,the 137 lines were divided into 14 groups.According to the grouping results,eight restorer lines and five sterile lines were selected from different groups for cross breeding,yielding 40 japonica hybrid rice combinations (F1).Investigation of the yield and quality of these combinations showed that highyield combinations could be obtained by increasing parental genetic distance to 0.8–0.9,a result accomplished largely by the introduction of indica genomic components of restorer lines.To further improve grain quality,the genetic distance between parents should be reduced to 0.4–0.5,suggesting an indicagenotype proportion of 30%–40% for restorer lines.This study may provide a reference for breeding of japonica hybrid rice.

1.Introduction

Indica hybrid rice cultivars produce approximately 15%–20%higher yields than inbred cultivars.The development of indica hybrid rice has progressed rapidly in China,so that its current planting area accounts for ＞80% of indica rice area and ＞50% of total rice production area.In contrast,progress in japonica hybrid rice breeding has been relatively slow,with its planting area accounting for ＜3%of japonica rice area[1,2].This limited development is attributed mainly to the weaker heterosis in japonica than in indica hybrid rice,owing to the short genetic distance between the parental lines.Analysis of the genetic background of the parental lines may permit selecting elite combinations of japonica hybrid rice,and is suggested [3–7] as a strategy for improving breeding efficiency.

The genetic diversity and population structure of parental lines are a desirable reference for breeding selection.In the past,genealogical information along with the origin or the phenotypic variation has been used to evaluate genetic diversity,but the methods had low reliability[8,9].The development of rice genomics has made it possible to accurately characterize genetic differences between parents of hybrid rice [10–15].Although in recent years,progress has been made in the analysis of parental genetic structure and the heterosis mechanism of indica hybrid rice,there are few reports on the genetic diversity and population structure of the parental lines of japonica hybrid rice in northern China [16–21].Previous studies were conducted mostly using restrictionfragment length polymorphism or simple sequence repeat markers[22–25].No published reports describe the deep assessment of genetic diversity among the parental lines of japonica hybrid rice using single-nucleotide polymorphisms (SNP) as more efficient DNA markers.

For japonica hybrid rice in northern China,higher proportions of indica genomic components in restorer lines and large parental genetic distance may cause lower seed setting rate and poor grain quality.The relationship between the genetic differences of the parental lines and the performance of the F1hybrid is more complex than in indica hybrid rice,and systematic analysis is still lacking.The purpose of the present study was to characterize genetic diversity,population structure,and indica-genotype proportion in 137 parental lines of japonica hybrid rice in northern China using 8 K SNP-Chips and to use the yield and quality of the F1hybrids to clarify the genetic differences and selection strategy of the parental lines of japonica hybrid rice.Characterizing the genetic differences among these cultivars was expected to provide guidance and reference for the breeding of strongly heterotic combinations of japonica hybrid rice.

2.Materials and methods

2.1.Plant materials

A set of 137 parental lines of japonica hybrid rice from northern China were used,comprising 90 restorer and 47 sterile lines(Table S1).

2.2.Genomic DNA extraction and SNP genotyping

Twenty days post-transplantation,a leaf sample was collected from one plant of each accession,followed by DNA extraction using the modified cetyltrimethyl ammonium bromide (CTAB) method.SNP genotyping was performed using 8 K SNP-Chips developed by China Golden Marker Biotechnology Limited Company,Beijing,China[26].Monomorphic SNPs and those with ＞50%missing data were excluded,leaving 4172 high-quality SNP loci for further analyses.

2.3.Genetic difference analysis

MEGA X [27] was used to compute the polymorphic information content(PIC)as well as other diversity parameters of the samples.Genetic diversity,an estimate of mean evolutionary divergence over sequence pairs within groups,was evaluated using the maximum composite likelihood model.An R package for landscape and ecological association studies (LEA) [28] was used for inference of individual admixture coefficients using sNMF to infer population structure and assign individuals to them.We considered models with approximately K=1–20 groups with admixture and correlated allele frequencies and applied 100 iterations for each K.The number of groups was determined by the crossentropy method [29,30].

The indica-genotype proportion of the parental lines of japonica hybrid rice was calculated based on 3190 SNP loci with polymorphism between Nipponbare (japonica reference genome) and 93-11(indica reference genome)[22].Loci consistent with Nipponbare origin were recorded as jj;loci consistent with 93-11 origin were recorded as ii,heterozygous loci were marked as ij;and the number of loci was N.The indica-genotype proportion(Pi)of each sample was calculated as follows:

2.4.Mating scheme and agronomic evaluation of the hybrids

Eight restorer lines and five sterile lines were selected from different groups of the 137 parent lines and 40 combinations were generated by crossing.Each F1progeny was planted in five-row plots (4 m row length,20 plants per row) with three replications by single-seedling transplanting.Cultivation management was consistent with that of other fields in the experimental base of Liaoning Rice Research Institute.At the maturity stage,five plants with panicle numbers near the mean were selected from each plot.Total grain number per panicle,seed setting rate,and 1000-grain weight were recorded.Grain yield was measured based on the whole harvest in the plot.After harvest,2-kg grain samples of each plot were stored at room temperature for three months for assessment of grain quality by Hunan Hybrid Rice Research Center,Changsha,Hunan,China.Brown rice rate,milled rice rate,head milled rice rate,chalky grain rate,chalkiness,grain length,width,and length-width ratio were determined based on the agricultural-industry standards of the People’s Republic of China NY/T 83-2017 and NY/T 2334.Amylose content was determined based on the agricultural industry standard NY/T 2639 of the People’s Republic of China.Determination of crude protein content of rice was by near-infrared method GB/T 24897-2010.Assays of taste and palatability in cooked rice were performed with an STA1A rice taste analyzer(SATAKE MultiMix Corporation,Saitama,Japan).

2.5.Data analysis

Excel 2016 (Microsoft Corporation,Redmond,WA,USA) and SPSS19 software (IBM Company,Chicago,IL,USA) were used to analyze the variance of the grain yield and quality of japonica hybrid rice.ANOVA model was used to identify significant differences between means at the level of P ＜0.05.

3.Results

3.1.Characterization of SNPs and genetic differentiation of the parental lines of japonica hybrid rice in northern China

A total of 4172 high-quality SNPs were used for genetic background analyses (Fig.1a).The mean PIC value of the 137 parental lines was 0.215,within a range of 0.014–0.375.The mean PIC values for the restorer and sterile lines were 0.231 and 0.124,respectively,within the range of 0–0.375 (Fig.1b;Table 1).The mean gene diversity of the 137 parental lines was 0.264 within a range of 0.014–0.5,and the gene diversities of the restorer and sterile lines were 0.287 and 0.148,respectively,within the range of 0–0.5 (Fig.1c;Table 1).Thus,the mean values of PIC and genetic diversity of the restorer lines were higher than those of the sterile lines.The mean indica-genotype proportion of the 137 parental lines was 23.7% within a range of 6.2%–62.3%;29.7% for the restorer lines (6.5%–62.3%) and 12.1% for the sterile lines (6.2%–29.5%) (Fig.1d;Table 1).Thus,the indica-genotype proportion of the restorer lines was higher than that of the sterile lines.

Table 1 Summary statistics of genetic diversity and indica-genotype proportion.

3.2.Population structure of the parental lines of japonica hybrid rice in northern China

Fig.1.Genetic background analyses of the parental lines of japonica hybrid rice in northern China.(a) Distribution of SNP markers on 12 chromosomes.The red lines represent positions of the SNPs on the chromosome,and the color intensity indicates the polymorphism information content(PIC)of the SNP in the parental lines of japonica hybrid rice.The deeper the color,the higher the PIC.(b)The PIC of the parental lines of japonica hybrid rice.(c)Genetic diversity of the parental lines of japonica hybrid rice.(d)The indica-genotype proportion of the parental lines of the japonica hybrid rice.(e)The cross-entropy statistics for each given K.(f)The indica-genotype proportion of each individual parental line of japonica hybrid rice.(g,h) Clustering analysis of the 137 parental lines via the model-based method when K=2 and K=14.The groups are distinguished by dotted lines;bars indicate the restorer lines (blue) and the sterile lines (red).

In the model-based grouping analysis,when K=2,the parental lines of japonica hybrid rice showed apparent differentiation;42.2%of the restorer lines(38 lines)were in a single cluster and had high indica-genotype proportion averaging 44.2%.In another cluster containing 57.8% of restorer lines (52 lines) and sterile lines (47 lines),the indica-genotype proportions of the restorer and sterile lines were 19.1% and 12.1%,respectively (Fig.1g;Tables 1,S2).With an increase in the K value,additional population differentiation was detected.The minimum value of the cross-entropy criterion at K=14 suggested the presence of 14 genetic clusters in this parental-line population (Fig.1e),including seven independent restorer line groups (groups 1–4,6,7,and 9) with a total of 63 restorer lines and seven mixture groups (groups 5,8,and 10–14)containing varying proportions of restorer and sterile lines(Fig.1h;Table S1).Group 5 was dominated by three restorer lines and one sterile line (53A),which was derived from glutinous rice and had high indica-genotype proportion.Groups 8 and 10 included two restorer lines and five or four sterile lines,respectively.Groups 11 and 13 were dominated by sterile lines,but there was a single restorer line with low indica-genotype proportion in each group.Groups 12 and 14 had 17 and 22 lines,respectively,and the ratios of sterile lines to restorer lines were 9:8 and 6:5,respectively.Thus,70% of the restorer lines showed independent genetic structure and high indica-genotype proportion averaging 34.8% while the other 30% of the restorer lines showed a close genetic relationship with sterile lines and lower indica-genotype proportion (17.8%) (Table S3).

3.3.Genetic diversity of the parental lines of japonica hybrid rice in northern China

The genetic diversity estimated by model-based analysis was ordered as follows:restorer lines ＞all parental lines ＞group 2 ＞group 1 ＞group 3 ＞group 5 ＞group 6 ＞sterile lines ＞group 11 ＞group 14 ＞group 10 ＞group 8 ＞group 12 ＞group 4 ＞group 7 ＞group 13 ＞group 9.Groups 1 and 2 comprised restorer lines and showed high mean genetic diversity as well as indicagenotype proportion (Table 1).Group 11,which contained mainly sterile lines,showed the highest genetic diversity(0.147),and thus could assist in selecting sterile lines in the future.The genetic diversity of group 14,composed of 12 sterile lines and 10 restorer lines,was 0.129 with low indica-genotype proportion averaging 10.3% (restorer lines was 11.7%),indicating that the genetic background of many restorer lines gradually evolved into japonica rice type,providing the prospect of breeding japonica-japonica hybrid rice (Tables 1,S1,S3).

The overall pairwise comparisons between the 14 groups classified by model-based grouping indicated differentiation between the two restorer-line groups (groups 1 and 2) with six groups (8 and 10–14 containing mainly sterile lines),and genetic distance values ranging from 0.67 to 0.79(Fig.2).Groups 3 and 4 (containing only restorer lines)and group 5(containing three restorer lines and one sterile line) showed a modest degree of differentiation with the abovementioned six model-based groups (containing mainly sterile lines) with the pairwise genetic distance values ranging from 0.39 to 0.50.Lower levels of differentiation were observed in the pairwise comparisons of groups 6,7,and 9 (containing only restorer lines) with the six abovementioned groups(containing mainly sterile lines) with pairwise genetic distance values ranging from 0.15 to 0.38.The abundant genetic differences among the parental lines produced strong heterosis;thus,high levels of differentiation were observed in the pairwise comparisons of groups 1 and 2 with the other groups,specifically the groups containing mainly sterile lines.

3.4.Yield and quality of combinations of japonica hybrid rice in northern China

Eight restorer lines (from groups 1–3 and 5–7) and five sterile lines (from groups 10–14) were selected from diverse groups of the 137 parent lines and 40 combinations were generated.Only one combination did not mature,owing to a long growth period;the other combinations matured normally with stable yields.According to the genetic distance between the parents,the hybrid combinations were divided into six groups(Tables 2,S4).The grain yields of the hybrids (F1) reached 650.0 kg per 667 m2when the parental genetic distances were 0.3–0.9.The mean yield reached its maximum(690.5 kg per 667 m2)when the parental genetic distances were 0.8–0.9.The grain yield of the hybrids decreased(637.2 kg per 667 m2)when the parental genetic distance exceeded 0.9.The yield of combinations with parental genetic distances between 0.2 and 0.3 decreased significantly to a mean of only 588.8 kg per 667 m2.The number of effective panicles per plant was greatest when the parental genetic distances were between 0.4 and 0.5,whereas the number of grains per panicle was greatest at parental genetic distances of 0.4–0.9.These results suggested that the high yield of japonica hybrid rice was achieved mainly by increased panicle size.The seed setting rate of japonica hybrid rice combinations in northern China exceeded 80% in general,reaching its maximum values(＞90%)at parental genetic distances of 0.2–0.4 and minimums at parental genetic distances ＞0.9.The 1000-grain weight reached its highest value(25.7 g)when the parental genetic distances were 0.3–0.4.The trend for grain length was similar to that for the ratio of grain length to width,the grain shape was desirable when the parental genetic distances were 0.5–0.8 and 0.2–0.3.However,the total variation of these three traits was small and likely influenced by the parental characteristics.

The taste scores of all 39 hybrids reached 80 points(Table 3),an excellent level.Milling quality was higher when the parental genetic distances were 0.2–0.4.The chalky traits of almost all hybrid combinations were unsatisfactory,suggesting that the selection of parental appearance-quality traits should be emphasized.Combinations with large parental genetic distance (＞0.9)showed the highest amylose and protein contents.When the parental genetic distances were between 0.4 and 0.5,the amylose content and endosperm protein content were moderate,the taste score was highest,and the yield remained at a high level.

The parental genetic distance was determined largely by the indica genomic components of restorer lines,given that the indicagenotype proportion of most sterile lines was ＜20% (Table S1).Thus,the hybrids were divided into three groups according to the indica-genotype proportion of restorer lines(Table 4).The grain yield of hybrid combinations reached 650 kg per 667 m2when the indica-genotype proportion of restorer lines was above 30%,a result due mainly to panicle size.However,the yield level of combinations decreased markedly when the indica-genotype proportion of restorer lines was 20%–30%.There was no significant difference in taste score,but the score was higher when the indicagenotype proportion of restorer lines was 30%–40% (Table 5).In terms of physical and chemical properties,when the indicagenotype proportion of restorer lines was ＞50%,the amylose and protein contents of the combinations were highest,also leading to a decline in grain quality (Table 5).In general,when the indicagenotype proportion of the restorer line was 30%–40%,the yield and quality of japonica hybrid rice could be improved synchronously.

Table 2 Grain yield performance of hybrids (F1) under different parental genetic distances.

Table 3 Quality of hybrids (F1) under different parental genetic distances.

Table 4 Grain yield performance of hybrids (F1) under different indica-genotype proportions of restorer lines.

Table 5 Quality performance of hybrids (F1) under different indica-genotype proportions of restorer lines.

4.Discussion

4.1.Genetic background of the parental lines of japonica hybrid rice in northern China

Although the discovery of sterile lines has promoted the largescale production of hybrid rice;most of the sterile lines have a similar genetic background,so that low genetic diversity has led to a bottleneck in the development of hybrid rice [31,32].The lower genetic diversity among the sterile lines than the restorer lines was due mainly to the low genetic diversity of japonica compared with indica rice.Also,the sterile cytoplasm of japonica hybrid rice in northern China carried only a BT-type pollen abortion,leading to a more similar genetic background.In consequence,it was necessary to introduce new types of cytoplasmic-sterile lines to broaden the genetic diversity of japonica-type sterile lines in northern China.In contrast,given that there were no restorer genes in japonica rice,restorer genes of japonica hybrid rice restorer lines were all from indica rice.For this reason,most restorer lines had more indica components.The introduction of indica lineage increased the genetic diversity of restorer lines [14,21].Breeding restorer lines with indica genetic background was an effective way to expand genetic diversity in northern China.

Fig.2.Genetic distances between groups based on model clustering.

In the present study,70% of the restorer lines had independent genetic structures differentiated from that of the sterile male lines,whereas 30% of the restorer lines were still closely related to the sterile male lines and clustered in the same groups.These results were not consistent with previous results[19,20,33–37]of genetic structure analysis or group classification of the parental lines of indica hybrid rice,in which the indica restorer and male-sterile lines were clustered into different groups.There was a minimal genetic difference between 30% restorer lines and sterile lines to avoid the negative effects of the introduction of the genetic background of indica rice on the ecological adaptability of japonica hybrid rice in low-temperature areas.This choice resulted in a small genetic difference between restorer and the sterile lines[13,38].

4.2.Genetic difference and yield and quality of japonica hybrid rice in northern China

The parental lines that are clustered in different groups produce greater heterosis [16,35,39,40].The correlation between genetic distances and heterosis was low or insignificant,varying across genetic materials and types of DNA marker [41,42].In the present study,the results showed that high grain yield combinations could be obtained by expanding parental genetic distance,a finding consistent with those of previous studies [16–18].For northern japonica hybrid rice,most sterile lines were similar to japonica rice varieties.Thus,the parental genetic distance was determined largely by the introduction of indica genomic components of restorer lines;In the present study,the yield advantage of the F1generation was greatly increased when the indica-genotype proportion of restorer parents was above 30%,suggesting the effective use of heterosis of indica-japonica subspecies by expanding the parental genetic distance.Still,poor seed setting rate may occur when the indica-genotype proportion of restorer parents is above 50%.Under northern climatic conditions,the late growth stage is affected by low temperature.With the further increase of parental genetic distance and indica lineage of restorer lines,japonica hybrid rice often shows differentiation between strong and weak (filled and partially filled) grains and poor grain quality.This observation leads to the dilemma in which the presence of too many indica components hinders its adaptation to the ecological conditions of northern China,whereas too few indica components restrict the expansion of parental genetic distance,resulting in weak yield heterosis [7,15].It is necessary to reduce the genetic distance between parents to improve taste quality.However,this distance should not be less than 0.3,in view of the yield consequences.The combinations with the indica-genotype proportion in restorer parents of 30%–40% showed the highest taste value,moderate amylose content,and lowest endosperm protein content.Thus,the appropriate indica and japonica components of parents are the fundamental starting point for high yield and quality breeding of japonica hybrid rice in northern China.

4.3.Breeding strategy for japonica hybrid rice in northern China

Targeting the high yield potential of japonica hybrid rice in northern China,the breeding strategy involves the use of a typical japonica male-sterile line as the female parent,to be crossed with a restorer line containing an appropriate proportion of indica rice components.This strategy has been supported by practical application.Liaoyou 9906,a japonica hybrid rice with large planting area in Liaoning province,resulted from a cross between restorer line C2106 (in group 1) and the sterile line 99A (in group 13) and showed a high parental genetic distance (0.9).Also,the indicagenotype proportions of the restorer and sterile parental lines were respectively 56.4% and 7.3% and the yield and seed setting rate were respectively 666.2 kg per 667 m2and 87.4% (Table S4).The characteristics and ecological adaptability of the female parent were retained while the genetic distance between the parental lines was increased as much as possible to ensure the seed setting rate.This strategy is a feasible scheme for exploiting the heterosis of indica-japonica subspecies.

For targeting the quality improvement of japonica hybrid rice in northern China,the breeding strategy follows the technical route of‘‘double fitness and double excellence”:the parental genetic distance and indica components of restorer parents were moderate,and the quality of both parental lines was excellent.In recent years,researchers have tried to improve quality while breeding high-yield japonica hybrid rice in northern China and have developed multiple excellent japonica hybrid rice cultivars with high yield and good quality[7].Jingyou 653,a new high-quality japonica hybrid rice,won the special prize of national excellent-taste japonica rice in 2015.The parental lines were restorer line C315 (in group 7)and sterile line 65A(in group 11)and showed a moderate parental genetic distance(0.3).The indica-genotype proportions of the restorer and the sterile parental lines were 24.5% and 11.0%,respectively,and the hybrid showed both excellent taste quality and stable yield.Thus,the quality of japonica hybrid rice could be improved by transferring the sterile line from the abundant good-quality japonica resources and crossing them with restorer lines with good appearance quality (especially chalkiness traits).

5.Conclusions

Breeding restorer lines with indica lineage is an effective way to expand genetic diversity and the heterosis of progenies.High-yield combinations were obtained by increasing parental genetic distance to 0.8–0.9,a distance largely determined by the introduction of indica genomic components of restorer lines.Further improvement of rice quality was associated with a genetic distance between parents of 0.4–0.5,with the indica-genotype proportion of the restorer line suggested to be 30%–40%.The breeding strategy requires a suitable choice of parental genetic distance,moderate proportion of the indica components in restorer parents,and excellent grain quality in both parental lines.

CRediT authorship contribution statement

Hong Gao:Conceptualization,Data curation,Formal analysis,Writing-original draft.Zuobin Ma:Data curation,Formal analysis,Software.Yuanzheng Wang:Data curation,Resources.Manli Zhang:Investigation,Resources.Xianju Wang:Investigation,Resources.Changhua Wang:Investigation,Resources.Zhiqiang Tang:Methodology.Liying Zhang:Software.Liang Fu:Investigation.Na He:Investigation.Hui Wang:Project administration.Yongan Yin:Investigation.Yuanjun Bai:Supervision.Guomin Sui:Funding acquisition,Project administration,Writing– review &editing.Wenjing Zheng:Conceptualization,Funding acquisition,Writing– review &editing.

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

This research was supported by Seed Industry Innovation Special Project of Shenyang Science and Technology Bureau(21-110-3-01),China Agriculture Research System (CARS-01-55),Liaoning Key Agricultural Program (2019JH1/10200001-2),and Hybrid Rice Breeding Subject (2019DD133719).

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2022.01.007.