Introduction
Rice (Oryza sativa L.) cultivars having numerous spikelets in a panicle (extra-heavy panicle types, EHPT) have been developed at many institutes as Hybrid Rice and ‘Super Rice’ cultivars in China (Chen et al. 2007, Yang and Zhang 2010), and ‘New Plant Type’ cultivars in the International Rice Research Institute (Peng et al. 1999). Many reports have pointed out that the EHPT do not always realize their high yielding potential, mainly because of their poor grain filling (Peng et al. 1999, Yang
et al. 2002, Zhang 2007, Yang and Zhang 2010, Kato et al.
2010). This poor grain filling is mainly attributable to a general fact that the increase in the number of spikelets per panicle (SP/ P) mostly depended on an increase in inferior spikelets for grain filling in EHPT rice (Kato et al. 2007, Kato 2010, Kato et al. 2010, Yang and Zhang 2010).
Kato (2010) showed a clear genetic variation in the degree of grain filling among EHPT rice. This variation was partly attributable to sink activity of developing endosperm during grain filling, particularly to the activity of ADPglucose pyrophosphorylase, which is a key enzyme for sucrose-to-starch conversion in this organ (Kato et al. 2007). Kato et al. (2010) demonstrated that the same series of nucleotide polymorphisms at OsAGPS2 (ADP-GLUCOSE PYROPHOSPHORYLASE
SMALL SUBUNIT 2) in chromosome 8 is present in common in
two EHPT rice cultivars, Milyang 23 and Nanjing 11, both of which showed good grain filling, high activity of ADPglucose pyrophosphorylase and a high rate of grain filling after anthesis. This means that these two EHPT cultivars have the same allele for good grain filling (AS2-2) at this locus. In addition, one more allele for good grain filling (SUT1-2) was detected at
OsSUT1 (SUCROSE TRANSPORTER 1) in chromosome 3, from
the same exploration for nucleotide polymorphisms (Kato et al. 2010). These alleles for good grain filling tend to associate together among a wide range of rice cultivars, particularly in
indica-type cultivars, but rarely exist in japonica-type cultivars
(Kato and Horibata 2011). Kato and Horibata (2011) emphasized that these alleles should be introduced from indica-type cultivars into EHPT genoindica-types in future breeding programs in order to obtain good grain filling.
The objective of this study was to confirm the availability of these two alleles for good grain filing in much extended sink capacity of rice panicles. First, a cross was made between two EHPT cultivars, Milyang 23 and Nanjing 11. Because both parents have a genotype of AS2-2 SUT1-2, no segregation should occur in the progenies. This study carried out a recurrent directional selection from F2 to F5 generations to increase SP/P
and its related traits. Finally, this study tried to obtain novel genotypes showing higher SP/P, and examined their degree of grain filling. In addition to this prime objective, selection
A Trial to Develop Novel Genotypes of Extra-Heavy Panicle Types with
Good Grain Filling in Rice
Tsuneo Kato
Faculty of Biology-Oriented Science and Technology, Kinki University
（930 Nishimitani, Kinokawa, Wakayama 649−6493, Japan）
Summary: The alleles AS2-2 and SUT1-2 at OsAGPS2 and OsSUT1, respectively, were suggested to improve grain filling in extra-heavy panicle type rice. To confirm the suitability of these alleles, this study sought to generate novel genotypes showing much higher number of spikelets per panicle (SP/P) and examined their degree of grain filling. These genotypes were generated by recurrent selection for higher SP/P from F2 to F5 populations
derived from a cross between two extra-heavy panicle type rice, Milyang 23 and Nanjing 11. Because both parents contain AS2-2 and SUT1-2, all progenies derived from this cross should have a AS2-2 SUT1-2 genotype. From the change in selection efficiency with successive generations, dominance effects were prevalent in most of panicle traits, including SP/P. Selected lines for SP/P in F6 generation exhibited much higher SP/P than the
parents and non-selected lines, which reached >300 SP/P in the best line. These selected lines also showed higher proportions of well-filled grains (specific gravity >1.15) than cultivars with AS2-1 SUT1-1 genotype, indicating clearly that the alleles for good grain filling, AS2-2 and SUT1-2, should sufficiently improve grain filling even with the greatly enlarged sink capacity of rice panicle. The necessity for improvement in source strength etc. was also suggested to improve grain filling with further increased numbers of SP/P.
Key words: directional selection, extra-heavy panicle type, grain filling, spikelet number, rice
Acccepted : November 7, 2012
efficiencies for the panicle traits in every generation after crossing were also elucidated to provide valuable information to the breeding of rice panicle traits.
Materials and Methods
All of the materials used in this study were seeded in nursery boxes in early May, grown in a green house, and transplanted in mid June into the paddy field of the Faculty of Biology-Oriented Science and Technology, Kinki University, Kinokawa, Wakayama, Japan. The planting density was 30 cm inter-row and 15 cm inter-hill, with a single plant per hill. Fertilizers were applied all as basal dressing at the rate of 6:6:6 g m-2 for
N:P2O5:K2O, except for the grain filling experiment, the details
of which are below. Standard cultivation practices including irrigation control, pest management, etc., were applied.
A cross was made in 2005 between two EHPT cultivars, Milyang 23 and Nanjing 11. An F2 population of 176 plants
derived from this cross was cultivated in 2007. Heading dates were recorded for the individual plants. After reaching maturity, the panicle on the longest culm in each hill was collected and counted for primary branches (PB) (PB/P), spikelets on primary branches (SPB), secondary branches (SB), and spikelets on secondary branches (SSB) as a panicle basis. From these raw data, SPB/PB, SB/PB, SSB/SB, and SP/P were calculated (Kato and Takeda 1996). For every panicle trait of SP/P, PB/P, SPB/ PB, SB/PB, and SSB/SB, the plants attaining the highest 4% of values of the population (7 plants out of 176 for each trait) were selected as parents of selected F3 progenies. The plants showing
extremely late heading and consequent poor ripening were excluded from the selected plants. In addition, 100 plants were selected at random as parents of non-selected F3 progenies. In
the F3 and other populations of the succeeding generations,
progeny lines consisting of six plants were cultivated. Up to the F5 generation, the plants attaining the highest 2% of values of the
whole population including all selected lines for every trait, as well as non-selected lines (total of about 690 to 910 plants, depending on the generations) were selected as parents of selected progenies of the next generation. According to the procedure of such recurrent directional selection as above, F6
population was finally produced in 2011. For lines derived from random selection, a method of single-seed decent was applied to proceed to the next generation. Two parents, Milyang 23 and Nanjing 11, were also cultivated and evaluated their panicle traits in each generation. The length of the longest culm in each hill, the length of the panicle on the longest culm, and panicle number per hill were also recorded for every hill.
To evaluate the efficiency of the directional selection for each trait, realized heritability (hR2) was estimated from the following
equation:
hR2 = {Ms’(n+1)− Mp(n+1)} / {Ms(n) − Mp(n)}... (1)
, where Ms(n) and Ms’(n+1) are the means of selected plants in Fn and their progenies in Fn+1, respectively, and Mp(n) and
Mp(n+1) are the means of the parental plants in Fn and in Fn+1,
respectively. The numerator and denominator of the equation (1) correspond to genetic advance and selection intensity, respectively, standardized with non-segregating population (two parents).
In the F6 generation, seven selected lines for SP/P and seven
non-selected lines, chosen at random from the respective populations, were cultivated separately from the above experiment to evaluate their grain filling. As controls, three EHPT cultivars (Milyang 23, Nanjing 11 and Kinangdang Puti) with the well-filling genotype of AS2-2 SUT1-2 (AS2-2 SUT1-2 cultivars) and three EHPT cultivars (Akenohoshi, Takanari and IR65598-112-2) with a poorly filled grain genotype of AS2-1
SUT1-1 (the counterpart alleles of AS2-2 and SUT1-2,
respectively) (AS2-1 SUT1-1 cultivars) were also cultivated. They were planted in the same paddy field, but a top dressing at the rate of 3:0:0 g m-2 and 3:4:4 g m-2 for N:P
2O5:K2O at three
weeks and seven weeks after transplanting, respectively, was applied in addition to the basal dressing. These lines and cultivars consisted of 12 hills (single plant per hill) in a single row. After maturation, one panicle on the longest culm was harvested from each of six hills in the center of the row. After measured the panicle traits described as above, all spikelets were threshed by hand, and were measured for the proportion of well-filled grains (grains with specific gravity >1.15) using NaCl solution (Kato et al. 2010).
Results and Discussion
Selection responses
Figures 1 and 2 show the changes in the means of panicle traits in non-selected and selected populations, respectively. In non-selected populations, the relative values compared with the F2 population decreased with generation advancement for most
of the traits. This might mainly be caused by the prevalence of dominance effects in the early generations after crossing and the decrease in these effects according to the elimination of heterozygosity caused by successive selfing. On the other hand, the trait values of selected populations were sustained or slightly increased compared with the F2 population, indicating that
directional selection was effective in those traits.
Figure 3 shows the changes in the efficiencies of the directional selection for panicle traits, expressed as realized heritability, with generation advancement. For most of the traits, except for SPB/PB, the selection efficiencies appeared to increase up to F4
efficiencies was apparently attributable to the decrease in dominance effect, as deduced from the changes in the non-selected population, and also to a decrease in the degree of segregation. The decrease in efficiency after the F4 generation
was probably caused by genetic homogenization within the population resulting from recurrent directional selection. Notably, the realized heritability for SP/P was lower than its component traits, i.e. PB/P and SB/PB, indicating the availability of indirect selection for SP/P (Kato 1997). An accidental drop in the realized heritability for SSB/SB in the F4 generation might
have been caused by unknown error factors.
Only a few experiments have been conducted for the change in heritability with generation advancement, with the exception of theoretical studies. Kato (1990) reported that the narrow-sense heritability for rice grain size estimated from parent-offspring correlation was high irrespective of generation (F2 to
F4), suggesting that an additive effect was prevalent for this trait.
Oinuma (1971), on the other hand, observed an increase in narrow-sense heritability for yield-related traits of tobacco from the F2 to F4 generation. In this case, large dominance variance
was estimated in the observed population. In the present case, the heritabilities for several panicle traits generally increased from the F2 to F4 generation, also suggesting the prevalence of
dominance effects, except for SPB/PB. An additive effect, not a dominance effect, might be primary in SPB/PB. We, therefore, can expect that additional selection is inefficient and not needed for the panicle traits of these selected populations.
Spikelet number and grain filling of the selected lines
Table 1 shows agronomic performances of several populations examined. The selected lines for SP/P appeared to be similar to other populations for the number of panicles per hill and culm length, slightly late in heading date, and larger panicle size than the others. Figure 4 and Table 2 show that the selected lines had more SP/P than AS2-2 SUT1-2 cultivars and non-selected lines, and similar values to AS2-1 SUT1-1 cultivars, as confirmed by variance analysis (Table 3). SP/P of one of the selected lines exceeded 300, which is generally a very high for SP/P in rice, although extremely high SP/P (more than 600) has been reported (Ikeda et al. 2010). The increase in SP/P in the selected lines was mainly caused by the increase in SB/PB (Table 2).
Figure 4 also shows that the proportion of well-filled grains of the selected lines was much higher than AS2-1 SUT1-1 cultivars, as confirmed by variance analysis (Table 3). This result clearly indicates that the alleles for good grain filling sufficiently improved grain filling, even with the much increased level of SP/P. Although multi-trial experiments will be needed, the availability of these alleles for the improvement of grain filling was confirmed in actually increased sink capacity of a rice panicle, at least in the present study. These alleles could be effectively utilized in a breeding program for a kind of ‘new
Fig. 1 Changes in relative values of panicle traits compared with the F2 population with generation advancement in the non-selected populations derived from a cross between rice extra-heavy panicle types, Milyang 23 and Nanjing 11. See text for the abbreviations for panicle traits.
Fig. 2 Changes in relative values of panicle traits compared with the F2 population with generation advancement in the population from recurrent directional selection derived from a cross between rice extra-heavy panicle types, Milyang 23 and Nanjing 11.
See text for the abbreviations for panicle traits.
Fig. 3 Changes in the realized heritabilities for panicle traits with generation advancement in the population from recurrent directional selection derived from a cross between rice extra-heavy panicle types, Milyang 23 and Nanjing 11. See text for the abbreviations for panicle traits.
1)
Selected lines, F6 lines derived from recurrent directional selection to increase the number of
spikelets per panicle; Non-selected lines, F6 lines from SSD procedure. Both populations were
derived from the cross between rice cultivars, Milyang 23 and Nanjing 11. AS2-1 SUT1-1 and
AS2-2 SUT2-2 cultivars, rice cultivars of extra-heavy panicle types showing poor grain filling and
good grain filling, respectively. Value in parentheses is SD.
Table 1 Means of agronomic traits in four rice populations used in the study
Fig. 4 Relationship between the number of spikelets per panicle and the proportion of well-filled grains in the four rice populations used in the experiment.
See text for the abbreviations for populations.
1)Selected lines, F
6 lines derived from recurrent directional selection to increase the number of spikelets
per panicle; Non-selected lines, F6 lines from SSD procedure. Both populations were derived from the
cross between rice cultivars, Milyang 23 and Nanjing 11. AS2-1 SUT1-1 and AS2-2 SUT2-2 cultivars, rice cultivars of extra-heavy panicle types showing poor grain filling and good grain filling, respectively. Value in parentheses is SD. See text for the abbreviations of panicle traits.
super rice’ with extra-heavy panicles and good grain filling in future.
The selected lines also showed nearly the same degree of grain filling as the non-selected lines, but slightly less than AS2-2
SUT1-2 cultivars (Table 3). We detected a negative correlation
between SP/P and the proportion of well-filled grains among the selected lines only (r = −0.813, P = 0.026), which is generally
found in most cases of EHPT rice, as cited above (Yan and Zhang 2010). The selected lines were directionally selected only for SP/ P in the present experiment. Therefore, this negative correlation suggests that the selected lines showing very high SP/P might be improved in their grain filling through the improvement of source strength etc. in addition to having alleles for good grain filling. See text and Tables 1 and 2 for the four populations.
Values in parentheses for Source are means of the respective populations. Values in parentheses for F-value and P are those when the mean of square for the comparison between groups are tested against that for heterogeneity within group.
Table 3 Analysis of variance for the number of spikelets per panicle and the proportion of well-filled grains (grains with specific gravity >1.15) between selected lines and other lines/cultivars used in this study
Acknowledgements
The author expresses special thanks to A. Horibata, H. Yamamoto and N. Hirooka, the Faculty of Biology-Oriented Science and Technology, Kinki University, for their valuable technical assistance.
References
Chen, S., J. Zhuang, Y. Fan, J. Du and L. Cao (2007) Progress in research and development on hybrid rice: A super-domesticate in China. Ann. Bot. 100 : 959− 966.
Ikeda, M., Y. Hirose, T. Takashi, Y. Shibata, T. Yamamura, T. Komura, K. Doi, M. Ashikari, M. Matsuoka and H. Kitano (2010) Analysis of rice panicle traits and detection of QTLs using an image analyzing method. Breed. Sci. 60 : 55− 64. Kato, T. (1990) Heritability for grain size of rice (Oryza sativa L.)
estimated from parent-offspring correlation and selection response. Jpn. J. Breed. 40 : 313− 320.
Kato, T. (1997) Selection responses for the characters related to yield sink capacity of rice. Crop Sci. 37 : 1472− 1475. Kato, T. (2010) Variation and association of the traits related to
grain filling in several extra-heavy panicle type rice under different environments. Plant Prod. Sci. 13 : 185− 192. Kato, T. and K. Takeda (1996) Associations among characters
related to yield sink capacity in space-planted rice. Crop Sci. 36 : 1135− 1139.
Kato, T. and A. Horibata (2011) Non-random distribution of the alleles for good grain filling at OsAGPS2 and OsSUT1 among a wide range of rice (Oryza sativa L.) cultivars. Breed. Sci. 61 : 217− 220.
Kato, T., D. Shinmura and A. Taniguchi (2007) Activities of enzymes for sucrose-starch conversion in developing endosperm of rice and their association with grain filling in extra-heavy panicle types. Plant Prod. Sci. 10 : 442− 450. Kato, T., A. Taniguchi and A. Horibata (2010) Effects of the
alleles at OsAGPS2 and OsSUT1 on the grain filling in extra-heavy panicle types of rice. Crop Sci. 50 : 2448− 2456. Oinuma, T. (1971) Inheritance and selection of quantitative
characters in Burley tobacco. II. Heritability and genotypic correlation in F2∼F4 generations of a varietal cross Burley 1×
Harmony. Jpn. J. Breed. 21 : 269− 274. (In Japanese with English summary)
Peng, S., K.G. Cassman, S.S. Virmani, J. Sheehy and G.S. Khush (1999) Yield potential trends of tropical since the release of IR8 and its challenge of increasing rice yield potential. Crop Sci. 39 : 1552− 1559.
Yang, J. and J. Zhang (2010) Grain-filling problem in ‘super’ rice. J. Exp. Bot. 61 : 1− 5.
Yang, J., S. Peng, Z. Zhang, Z. Wang, R. M. Visperas and Q. Zhu (2002) Grain and dry matter yields and partitioning of assimilates in japonica/indica hybrids. Crop Sci. 42 : 766−772. Zhang, Q. (2007) Strategies for developing green super rice.
Proc. Natl. Acad. Sci. USA 104 : 16402− 16409.
登熟良好な新規極穂重型イネ遺伝子型を開発する試み
加藤恒雄
近畿大学生物理工学部（〒 643 − 6493 和歌山県紀の川市西三谷 930）
極穂重型イネ品種の登熟を向上させることが示唆されているイネ遺伝子座 OsAGPS2 および OsSUT1 上の遺伝子 AS2 − 2 お よび SUT1 − 2 の効果を検証する一環として，上記遺伝子をもち，かつ従来品種の穎花数／穂を超える極穂重型系統を F5ま での定方向選抜によって開発して，これらの登熟程度を検討した．各世代における選抜効率を実現遺伝率として評価したと ころ，穂関連形質では交雑後の分離初期世代で優性効果が優勢であることが示唆された．F6において穎花数／穂について選 抜した系統と無選抜系統を，良登熟型遺伝子をもつ品種と非良登熟型遺伝子をもつ品種とともに栽培し，穎花数／穂および 良登熟籾歩合を比較した．その結果，選抜系統は他群と同等もしくはより多い穎花数／穂をもち，最高値で 300 穎花／穂以 上を記録したものの，非良登熟型遺伝子を持つ品種よりも高い良登熟籾歩合を示すことが分かり，本良登熟型遺伝子は非常 に多い穎花数／穂をもつ遺伝子型においても登熟向上に有用であることが分かった．一方，選抜系統の良登熟籾歩合は良登 熟型遺伝子をもつ品種よりもやや低く，ソース能等をさらに改良する必要性も指摘された． キーワード：イネ，穎花数，極穂重型，定方向選抜，登熟 作物研究 58：15 − 20（2013） 連絡責任者：加藤恒雄（[email protected]）
