An Approach to the “Grain-Filling Problem” in Rice through the Improvement of Its Sink Strength
Tsuneo Kato
Faculty of Biology-Oriented Science and Technology, Kindai University (930 Nishimitani, Kinokawa, Wakayama, 649-6493 Japan)
Summary: Many rice cultivars with numerous spikelets in a panicle (extra-heavy panicle types, EHPTs) have been developed to attain higher grain yield. However, there is a grain-fi lling problem for these EHPTs: they did not attain expected higher yield in many cases because of their low degree of grain fi lling, resulting from the increase in inferior spikelets showing poor grain fi lling. This short review addresses an approach toward the grain-fi lling problem of EHPT rice, particularly for the genetic improvement of its sink strength, and proposes two strategies for the resolution of this problem. The fi rst strategy is the improvement of the poor grain fi lling in the increased inferior spikelets. This may be realized by utilizing suggested alleles for good grain fi lling at APS2, APL2 (they correspond to ADP-glucose pyrophosphorylase subunits), and SUT1 (sucrose transporter) loci. These alleles are expected to accelerate the unloading and metabolism of sucrose in developing endosperm, and to maintain a gradient of sucrose concentration from source to sink. The second strategy is the construction of a novel EHPT, in which its spikelet number is increased by increasing superior spikelets. A selection experiment strongly indicated that this strategy may be attained by the selection to increase the number of primary branches per panicle. Although these two strategies are needed to evaluate further, it is suggested that they are available to future breeding of EHPTs with good grain fi lling, and they can be combined with other approaches to this grain-fi lling problem.
Keywords: extra-heavy panicle types, grain fi lling, high yield, rice, sink strength.
Introduction
Rice (Oryza sativa L.) cultivars of heavy or extra-heavy panicle types (EHPTs), such as super rice and hybrid rice in China, New Plant Type rice in the International Rice Research Institute, etc., have been developed worldwide to enlarge yield potential. These cultivars have attained very large sink capacity through increasing the number of spikelets per panicle ( Peng et al. 1999, Peng 2009, Khush 2013). Additionally in Japan, many EHPT cultivars, such as Akenohoshi, Takanari, Habataki, Hokuriku 193, Momiroman, etc., have been released from the Super high-yielding project supported by the Ministry of Agriculture, Forestry and Fishery of Japan, since 1982 (Sakai and Ishii 2016).
There are accumulated problems in rice cultivars with EHPTs that do not realize the expected high grain yield in many cases, even if they clearly exhibit large sink capacity to import photoassimilates from source organs (Peng et al. 1999, Yang and Zhang 2010, Panigrahi et al. 2019). Yang and Zhang (2010) proposed a grain-fi lling problem in rice EHPTs to explain this situation. The major source of this problem is derived from the poor grain-filling ability of EHPTs to fill up the enlarged sink
capacity. This is because the increase in spikelets per panicle in EHPTs mainly depends upon an increase in spikelets on secondary branches in a panicle, particularly on the lower part of a panicle. These spikelets on secondary branches generally show an inferior degree of grain filling, compared with those on primary branches (Fig. 1) ( Kato and Takeda 1996, Ishimaru et al. 2005, Yang et al. 2006). Therefore, in this context, we must recognize that yield sink capacity (expressed as a product of panicle number, spikelet number per panicle and maximum single grain weight) and its filling efficiency (degree of grain fi lling) should be genetically increased both, not each of them, in future breeding, because yield is a product of both parameters.
In the course of resolving this grain-fi lling problem to realize sustainable higher yield in rice as a practical demand, many interesting subjects in physiology, biochemistry, genomics, etc., for source-sink relationship in higher plants should be combined. These interactive approaches from diff erent fi elds may provide novel and informative insights in the respective areas. This short review addresses an approach toward the grain-fi lling problem of rice with special references to the genetic improvement of its sink strength, and puts forward two kinds of strategy for the resolution of this problem (Fig. 2).
Acccepted: September 30, 2019
Corresponding author: Tsuneo Kato ([email protected])
Review Article
Improvement of Poor Grain Filling of Inferior
Spikelets in EHPTs: A Strategy
The first strategy should be the improvement of inferior spikelets (spikelets on secondary branches) in conventional EHPTs with poor grain fi lling (Strategy I in Fig. 2). An essential resource for this genetic improvement is the existence of a wide
genetic variation in grain-filling degree among EHPTs. Kato (2010) examined the variation in the degree of grain filling in inferior and superior spikelets among several rice cultivars, including EHPTs, in seven environments of different growing years and locations.
Results showed that an EHPT cultivar, Milyang 23, exhibited higher degrees of grain filling in both inferior and superior
Inferior sink (spikelets on secondary branch) Superior sink (spikelets on primary branch)
ExtraͲheavypanicletype
Sourceorgan Yield=Yieldsinkcapacity㽢Filling efficiency Photoassimilates Increase Decrease inconventionalextraͲheavypanicletypesOrdinarytype
Sourceorgan Photoassimilates Inferior sink (spikelets on secondary branch) Superior sink (spikelets on primary branch)StrategyI
StrategyII
Improvementofthepoorgrainfilling ininferiorsink IncreaseinsuperiorsinkFig. 1. The diff erences between ordinary panicle type and extra-heavy panicle type rice in their panicle structure, grain fi lling and yield. A scheme.
spikelets compared with another EHPT cultivar, Akenohoshi, in all diff erent environments. Relatively good grain fi lling has been reported also in other EHPTs, i.e., Nanjing 11 and Takanari ( Kato et al. 2013, Kato and Horibata 2015), and others (Yamamoto et al. 1990). Kato (2010), Kato et al. (2013) and Kato and Horibata (2015) also detected that EHPTs showing higher degrees of grain fi lling also showed higher rates of grain growth after anthesis in both inferior and superior spikelets. These apparent genetic variations in grain-fi lling characteristics can open the possibility to develop EHPTs with good grain fi lling. The next step for the improvement of grain filling should be understanding of the causes underlying these genetic variations.
Source and Sink Strengths
One of the key terms for understanding the grain-filling problem is the source-sink relationship in plants, particularly in the grain-filling stage. The terms of source and sink were first introduced by Manson and Maskel (1928) in cotton, to describe the dynamic movement of photoassimilates within a plant. The source indicates the origin of photoassimilates, and the sink denotes the receipt of these photoassimilates, in a manner like a vessel under the water spout in a kitchen. The abilities of both source and sink organs are frequently called strength (dw/dt) , which is a product of organ size (w) and specifi c activity (dw/dt/ w) (Warren-Wilson 1967). The sizes of source and sink organs are obvious, and source activity should correspond to photosynthetic activity. On the other hand, there have been arguments what the general sink activity is ( Farrar 1993). The present review discusses this subject particularly for rice plants after anthesis, although only for carbon dynamics.
In a rice plant after anthesis, the source organs are mainly active leaf blades on the upper part of the stem, and sink organs are panicles with developing grains, or particularly developing endosperm cells. Before anthesis, or in the vegetative growth period, actively growing organs such as top leaves are sink organs, indicating that a dynamic transition is frequently observed for the source-sink relationship (Hirose et al. 2006). A stem along phloem, including leaf sheaths and a culm, also plays a role as a sink organ before anthesis, and both source and sink after anthesis. The movement of materials between tillers is generally rare in an adult rice plant. Therefore, a tiller or a main culm including a panicle may be regarded as a unit of source-sink relationship in rice.
A n i m p o r t a n t f a c t i s t h a t t h e s e m o v a b l e c a r b o n s (photoassimilates or non-structural carbohydrates (NSC)) are produced only, or mostly, in source organs and should be transported from source to sink. Therefore, long-distant transport through the phloem from source to sink is a key subject in the grain-filling problem. What is the driving mechanism of the
photoassimilates in long-distance transport from source to sink in a rice plant, which has no circulatory organs in higher animals?
Pressure-Flow Hypothesis and Its Application to
the Grain-Filling Problem
A German plant physiologist, Münch, proposed a plausible hypothesis for the long-distant transport through phloem (Münch 1930). This hypothesis gives a parsimonious system for this transport ( Minchin and Lacointe 2005): there is a phloem with two termini, a terminus of source side and the other at the sink side, both of which consist of active cells surrounded by a semi-permeable membrane (Fig. 3). In a rice plant after anthesis, photoassimilates produced by photosynthesis are converted to sucrose in parenchyma cells near source terminus, which is loaded into the source terminus by sucrose transporter on the membrane (Fig. 3A). As a result, water permeates, or is probably passed through aquaporin (Luu and Maurel 2013), from the lower concentration site of sucrose (outside the cell) to the higher concentration site (inside), which inevitably generates higher osmotic pressure within the cell, and also turgor pressure as its counterpart. As a consequence, water at higher turgor (at the source terminus) begins to move toward the sites of lower turgor through the phloem (toward the sink terminus in this case) (Fig. 3B). Consequently, sucrose in source terminus also moves toward the sink terminus along with the directed water fl ow (Fig. 3B), and is then unloaded into the sink apoplast via sucrose transporter (Fig. 3C). This is the Münch s pressure flow hypothesis for movable substances in phloem solute. This pressure flow of sucrose terminates when the source terminus and sink terminus have the same turgor pressure (the same sucrose concentration), resulting in the termination of water fl ow (Fig. 3D). In addition, plants have their own mechanisms with this minimum platform, e.g., fl ow resistance in phloem system (Minchin and Lacointe 2005).
In other words, a higher concentration gradient of sucrose from source to sink should be maintained, if the sucrose supply is to be maintained to sink organs, which is essential for the sustainable growth of sink organs (Patrick and Offl er 2001). This could lead to the improvement of grain filling in EHPT, particularly in inferior spikelets. Of course, strong and sustainable source strength producing photoassimilates and supplying sucrose to source terminus is important for maintenance of the sucrose and pressure gradient from source to sink. Therefore, sink and source strengths should both be improved, not just one. It will be necessary to explore the best balance between sink and source strengths to resolve this problem. This review concentrates on the sink strength.
Metabolism of Sucrose into Starch in Rice
Developing Endosperm
Figure 4 shows the pathway for the metabolism of sucrose to starch in developing endosperm in rice after anthesis (after Perez et al. 1975). Transported sucrose from source organs is unloaded by sucrose transporters at the sink terminus into the sink apoplast, and again loaded into endosperm cells. In developing endosperm cells, sucrose is metabolized by either sucrose synthase (SuSy, EC 2.4.1.13) or (acid) invertase (EC 3.2.1.26). In the very early stages after anthesis, invertase activity is rather high, but declines soon after ( Kato 1995). Therefore, the pathway directed by SuSy is prevalent in the most active stages in the developing endosperm (Kato 1995, Counce and Bravois 2006, Mohapatra et al. 2009): sucrose is metabolized fi nally to starch, via UDP-glucose, glucose-1-phosphate and ADP-glucose. Starch is insoluble in ordinary conditions, and it is deposited into amyloplast, resulting in matured starchy rice grain. The conversion of soluble sucrose to insoluble starch in developing endosperm cells clearly contributes to the decrease in the sucrose concentration within cells, resulting in maintenance of the sucrose gradient, leading to an improvement in grain filling. Alternatively, sucrose gradient may not be maintained in earlier stages of grain fi lling, if this conversion does not work well, as well as ineffi cient unloading by sucrose transporter, all of which should impair grain fi lling.
Kato et al. (2007) examined the variation in activities of two
enzymes functioning in the sucrose-starch metabolism pathway, SuSy and ADP-glucose pyrophosphorylase (AGPase, EC 2.7.7.27), in the developing rice endosperm of several EHPT c u l t i v a r s . O n e o f t h e o t h e r e n z y m e s , U D P - g l u c o s e pyrophosphorylase (EC 2.7.7.9), generally showed much higher activity than those two enzymes in rice developing endosperm ( Nakamura et al. 1989, Kato 1995), suggesting that this enzyme may not be a limiting step of this pathway. Results showed that an EHPT cultivar, Milyang 23, exhibited higher AGPase activity, particularly for inferior spikelets, than another EHPT cultivar, Akenohoshi, consistently under different environments. As already shown, Milyang 23 had a higher degree of grain fi lling than Akenohoshi (Kato 2010). The activity of SuSy was higher in Milyang 23 than Akenohoshi in several cases, but not so evident compared with AGPase. Although Kato et al. (2007) did not examine the activities of other enzymes such as soluble starch synthase (EC 2.4.1.21), etc., this result indicates that the activity of AGPase, rather than SuSy, in developing rice endosperm may be a limiting factor in the metabolic pathway from sucrose to starch, and it participates in the regulation of the decrease in sucrose concentration at the sink terminus.
AGPase is known as an important enzyme in starch biosynthesis in higher plants. This enzyme is a heterotetramer consisting of two large and two small subunits (Tetlow et al. 2004). Many natural and induced variants for AGPase have been reported, and some of them contributed the increase in grain and tuber yield in several crops (Simdansky et al. 2002, Simdansky
Phloem Sucrose Phloem Sucrose Water Phloem Sucrose Water Phloem
A
B
C
D
TransporterFig. 3. Münch s pressure flow hypothesis. A, Loading sucrose into the source terminus of the phloem via sucrose transporters. B, Water fl ow derived from the pressure gradient from the source (high sucrose concentration) and sink (low sucrose concentration) termini. Sucrose fl ow into the same direction as water is derived. C, Unloading the transported sucrose from the sink terminus to the apoplast. D, Termination of water flow and, thus, of sucrose transport, resulted from the breakdown of pressure gradient.
et al. 2003, Nakatani and Komeichi 1992, Hou et al. 2017, Mohapatra et al. 2009, Oiestad et al. 2016, Wang et al. 2007, Sakulsingharoj et al. 2004). These facts imply the signifi cance of this enzyme in sink development in many monocot and dicot species.
In rice, several isoforms in AGPase subunits have been reported in leaf and endosperm ( Lee et al. 2007). Ohdan et al. (2005) indicated that the isoforms OsAGPS2b (a type from alternative splicing of this transcript) and OsAGPL2 are the major forms in developing endosperm. Kato et al. (2010) and Kato and Horibata (2015) examined sequence polymorphisms in the loci responsible for the enzyme subunits, i.e., OsAGPS2 (APS2, Os08g0345800) and OsAGPL2 (APL2, Os01g0633100), as well as the locus for OsSUT1, sucrose transporter 1 in sink terminus (SUT1, Os03g0170900) (Aoki et al. 2003), among six rice cultivars including EHPTs with different degrees of grain filling. The results showed that more than 20 polymorphisms among these six cultivars within each of the three loci involving coding sequences and the 5 and 3 untranslated regions (Fig. 5). Any of these polymorphisms, however, do not alter the amino acid sequence.
Of these polymorphisms, fi ve, fi ve and seven SNPs in APS2, APL2, and SUT1, respectively, can provide molecular markers to discriminate the respective genotypes (Fig. 5, Table 1). Kato and Horibata (2015) and Kato and Horibata (2018) explored alleles or haplotypes at each of the three loci, according to the
genotypes for these SNPs, using a total of 248 cultivars including 69 cultivars of the world rice core collection (Kojima et al. 2005). As a result, eight, seven and six alleles were detected at APS2, APL2, and SUT1, respectively (Table 1). Of course, the numbers of alleles at these three loci are the minimum, because we did not examine the genotypes of all SNPs.
Alleles for Good Grain Filling
Table 2 shows the relationship between the genotype of the three loci and the degree of grain filling among several rice cultivars including EHPTs. Notably, EHPTs showing poor grain filling, Akenohoshi and Momiroman, had commonly a genotype, APS2-1 APL2-1 SUT1-1. On the contrary, EHPTs showing good grain fi lling, Milyang 23, Nanjing 11, Hokuriku 193, Takanari, Habataki, had APL2-2 in common in all cases, and they also have APS2-2 and SUT1-2 in some cases. This strongly suggested that the alleles designated as 2 at these three loci, particularly APS2-2, may be alleles for good grain fi lling. Kato and Horibata (2015) also showed that Milyang 23 and Nanjing 11 (APS2-2 APL2-2 SUT1-2) exhibited a higher degree of grain filling, and also higher AGPase activity in developing endosperm than Akenohoshi and Momiroman (APS2-1 APL2-1 SUT1-1). However, the materials used in the previous studies were all cultivars with different genetic
Fig. 4. A metabolism map and key enzymes from sucrose to starch into a rice developing endosperm cell (referred and arranged from Perez . (1975)).
Sucrose UDPͲglucose GlucoseͲ1Ͳphoshate ADPͲglucose Starch Fructose Glucose GlucoseͲ6Ͳphosphate FructoseͲ6Ͳphosphate UDP UTP PPi ATP ADP PPi ATP ADP ADP Pi A B C D E F G H I A:Sucrosesynthase B:UDPͲglucosepyrophosphorylase C:ADPͲglucosepyrophosphoryrase D:Solublestarchsynthase,etc. E: Nucleotidedephosphokinase F:(Acid)invertase G:Hexokinase H:GlucoseͲphosphateisomerase I:Phosphoglucomutase J :Phospholyrase J Sucrosetransporter Photoassimilates =Sucrose Insolubilize, compartmented inamyloplast Developingendospermcell Perezetal.(1975)(PlantPhysiol.56:579Ͳ583) Fig.4
Table 1. Alleles at the three loci related to grain filling in rice, , , and , and their single nucleotide polymorphisms, as judged with molecular markers..
Position1) * *
-1 CC del CC CC CC CC del CC
550 T G G G G G G G
2751 A G A A G A G A
2838 G A G A A G G A
4462 AT del AT AT del AT del AT
5275 T C T C C C C T Position1) * -3955 CA TG TG CA TG ? TG -3680 G C C C C C C -2549 C C T C T T T -455 AC GT AC AC AC AC 1138 A A G A G A A 1282 C C T C C T C Position1) * -388 A G A G A A -287 1 2 2 2 1 1 1468 A T A T A A 1909 G C C C C G 4744 T A T A T T 4752 A C A A A A 5065 C T C T C ?
* indicates rare alleles which are detected less than 1% of the target population.
1) Nucleotide position from the beginning of the start codon. Negative values indicate upstream.
APS2 (OsAGPS2,Os08g0345800) 1kbp
APL2 (OsAGPL2,Os01g0633100)
SUT1 (OsSUT1,Os03g0170900)
InDel CAPS CAPS CAPS CAPS CAPS
CAPS CAPS InDel CAPSCAPS
CAPS CAPS CAPS CAPS CAPS CAPS SSR CAPS
Fig. 5. Structure of three loci related to grain fi lling of rice, , , and . Open arrows indicate the positions of single nucleotide polymorphisms (SNPs) detected in several rice cultivars involving extra-heavy panicle types (Kato and Horibata 2015). Closed arrows indicate the position of molecular markers designed to detect these SNPs in Table 1. Two vertical lines at each locus indicate the start and end positions of transcription, respectively.
backgrounds, other than the three loci, which might aff ect their grain filling. Further examinations are needed to confirm the eff ects of these alleles for good grain fi lling in a common genetic background.
An interesting fact was that the alleles at the three loci, APS2, APL2, and SUT1, distributed among 248 rice cultivars including the world rice core collection in an associated manner, not at random (Kato and Horibata 2011, Kato and Horibata 2015, Kato and Horibata 2018): APS2-2, APL2-2, and SUT1-2, plausible alleles for good grain fi lling, tended to be associated with each
other only in indica-type cultivars, although these three loci are located on different chromosomes (Fig. 6). This association among these three alleles might be the result of unintended selection in the course of previous breeding, but only for indica-type cultivars. These plausible alleles for good grain fi lling can be introduced easily from indica-type cultivars. In fact, second-generation NPT cultivars were partially improved for their poor grain filling, by means of the introduction of indica-type germplasm into the first-generation NPT through crossing of indica-type cultivars (Peng et al. 1999).
Table 2. Genotypes for the three loci related to grain fi lling in rice, , , and , for some cultivars including extra-heavy panicle types (EHPTs) with diff erent grain fi lling
Cultivar Notes
Milyang 23 EHPT, good grain fi lling
Nanjing 11 EHPT, good grain fi ling
Hokuriku 193 EHPT, good grain fi lling
Takanari EHPT, good grain fi lling
Habataki EHPT, good grain fi lling
Akunohoshi EHPT, poor grain fi lling
Momiroman EHPT, poor grain fi lling
Taichung Native 1 non-EHPT
Nipponbare non-EHPT
Koshihikari non-EHPT
Fig. 6. Distribution of genotypes for three loci related to grain fi lling of rice, , , and among 248 cultivars including a world rice core collection. Closed and open bars indicate -type and -type cultivars, respectively. 0 5 10 15 20 25 30 35 111 121 131 135 141 145 221 222 223 224 241 242 243 244 252 254 311 321 322 331 333 341 343 345 353 361 371 411 421 423 424 431 441 442 443 444 445 455 522 523 524 543 611 641 722 841
indica
japonica
Genotypes(arrayofallelesatAPS2, APL2,andSUT1)
Pe rc en ta ge of cultiv ar s within indica Ͳtype or japonica ͲtypeAnother Strategy toward the Grain-Filling Problem
An additional strategy for the grain-fi lling problem other than the improvement of inferior spikelets is the construction of a novel panicle architecture in which superior spikelets are prevalent (Strategy II in Fig. 2). Kato (2013) conducted an upward directional selection experiment in the segregating populations derived from Milyang 23/Nanjing 11, both of the parents have a common genotype, APS2-2 APL2-2 SUT1-2. From an F2 population, individuals showing higher values of top two or
four percent were selected successively up to F5 generation to
examine the selection responses for the number of spikelets per panicle and its components. Target component traits were the number of primary branches (PBs) per panicle, the number of spikelets on PBs (SPBs) per PB, the number of secondary branches (SBs) per PB, and the number of spikelets on SBs (SSBs) per SB, because the number of spikelets per panicle is expressed as PBs/panicle × {SPBs/PB+(SBs/PB × SSBs/SB)} (Kato and Takeda 1996).
Fig. 7 shows the performances of F8 lines resulting from
selection for the respective traits. Because of the high positive correlation between SBs/PB and SSBs/SB (Kato and Takeda 1996), simultaneous selection was done for these two traits. The
upward directional selection could effectively increase the respective traits (data not shown). The selection for PBs/panicle, SBs/PB and SSBs/SB, and spikelets/panicle resulted in an increase in the number of spikelets/panicle, but the selection for SPBs/PB did not (Fig. 7A). This ineff ectiveness in the upward selection for SPBs/PB to increase spikelets/panicle might be a narrow variation in this trait. The number of spikelets on SBs/ PB, inferior spikelet number, increased by means of the selection for SBs/PB and SSBs/SB, and spikelets/panicle, but not for PBs/ panicle and SPBs/PB (Fig. 7B). In contrast, the selection for SBs/PB and SSBs/SB, and spikelets/panicle did not increase SPBs/panicle, superior spikelet number, although the selection for PBs/panicle signifi cantly increased the same trait (Fig. 7C).
The results of these selection responses clearly indicated two important facts. First, simple upward selection for spikelets/ panicle was eff ective for a direct increase in its own trait. This increase in spikelets/panicle was apparently accompanied by an increase in inferior spikelet number, and also resulted from the indirect selection for SBs/PB and SSBs/SB. This situation in the panicle is similar to conventional EHPTs. Secondly, the selection for PBs/panicle obviously increased spikelets/panicle, without a signifi cant increase in the number of inferior spikelets/ panicle (SSBs/PB in Fig. 7B). This strongly suggested that we might
0 50 100 150 200 250 300 Randomly selected (control) PBs/panicle SPBs/PB SBs/PB andSSBs/SB Spikelets /panicle No. of spik ele ts /panicle
A
F=4.16 P=0.097 F=0.59 P=0.464 F=5.88 P=0.052 F=5.78 P=0.037 Selectedfor 0 2 4 6 8 10 12 14 16 18 Randomly selected (control) PBs/panicle SPBs/PB SBs/PB andSSBs/SB Spikelets /panicle No. of SSBs /PBB
F=0.87 P=0.394 F=0.73 P=0.418 F=22.57 P=0.001 F=11.90 P=0.014 Selectedfor 0 20 40 60 80 100 Randomly selected (control) PBs/panicle SPBs/PB SBs/PB andSSBs/SB Spikelets /panicle No. of SPBs /panicleC
F=7.40 P=0.042 F=0.46 P=0.523 F=17.82 P=0.002 F=0.19 P=0.680 SelectedforFig. 7. Performance of F8 selected lines for primary branches (PBs) per panicle, spikelets on primary branches (SPBs) per
PB, secondary branches (SBs) per PB and spikelets on SBs (SSBs) per SB, and spikelets per panicle after directional upward selection from F2 to F5 of the segregating populations from Milyang 23/Nanjing 11 (Kato 2013).
Randomly selected lines are also shown as control. -value and -value indicate the results of ANOVA for the difference between randomly selected lines and the lines selected for the respective traits and its probability, respectively. Means ± S.E.
construct novel panicle architecture with an increase in superior spikelets and without an increase in inferior ones, by means of the upward selection for PBs/panicle. In fact, several quantitative trait loci for this trait have already been identified in the rice genome (Kato 2004, Terao et al. 2010). Further investigation is needed to evaluate this strategy.
Concluding Remarks
The present short review proposed two strategies for improving poor grain filling generally found in conventional EHPTs in rice: (1) improvement of poor grain fi lling, particularly for inferior spikelets which are prevalent in the panicle of conventional EHPTs. Plausible alleles for good grain filling, APS2-2, APL2-2, and SUT1-2 have been already detected. Additionally, (2) construction of novel panicle architecture in which superior spikelets are prevalent. Though both strategies will be needed to evaluate their practical availability, it is strongly expected to contribute to the future breeding of EHPTs.
Other major remaining problems to be understood are: (1) underlying mechanisms causing the diff erences in grain fi lling between inferior and superior spikelets, and (2) non-random association among alleles at APS2, APL2, and SUT1. For the former problem, it is necessary to note that the difference between inferior and superior spikelets within a panicle might not directly relate to the genetic differences in grain filling among EHPTs, because inferior and superior spikelets have the same genotype. Some physiological and/or morphological differences between the two categories should be considered. Many factors, e.g., phytohormone levels (Yang et al. 2006, Panigrahi et al. 2019), grain-filling rates (Kato 2010), gene expression levels (Ishimaru et al. 2005, Panigrahi et al. 2019), enzyme activity (Kato 1995), etc., have been reported to be different between inferior and superior rice spikelets. It is possible, however, that these differences may be derivative results from other fundamental causes. For the latter association, it should be understood why this association among three alleles for good grain filling is detected only in indica-type cultivars, not in japonica-type ones, if this association resulted from unintended selection.
One of the approaches toward the same problem from cultivation practices, for example, is moderate soil drying during the grain fi lling period (Yang and Zhang 2006, Yang and Zhang 2010, Wang et al. 2019). This drying procedure aimed at appropriate acceleration of senescence of the whole rice plant. This and other cultivation approaches can be combined with genetic approaches in this review toward comprehensive resolution of the grain-fi lling problem.
Acknowledgements
I sincerely thank Emeritus Professor P. K. Mohapatra, Sambalpur University, India, for valuable discussions and suggestions on this study, Dr. A. Horibata, Faculty of Biology-Oriented Science and Technology, Kindai University, and Dr. N. Aoki, Graduate School of Agricultural and Life Sciences, the University of Tokyo, for technical supports. This study was supported in part from the Grants-in-Aid of MEXT (04660010, 22580020, 17K07627), and the Grant from Sapporo Bioscience Foundation.
Literature Cited
Aoki, N., T. Hirose, G. N. Scofi eld, P. R. Whitefi eld and R. T. Furbank (2003) The sucrose transporter gene family in rice. Plant Cell Physiol. 44: 223-232.
Counce, P. A. and K. A. Bravois (2006) Sucrose synthase as a potential indicator of high rice grain yield. Crop Sci. 46: 1501-1507.
Farrar, J. F. (1993) Sink strength: What is it and how do we measure it? Introduction. Plant, Cell Environ. 19: 1015-1016. Hirose, T., T. Ohdan, Y. Nakamura and T. Terao (2006)
Expression profi ling of genes related to starch synthesis in rice leaf sheaths during the heading period. Physiol. Plant. 128: 425-435.
Hou, J., T. Li, Y. Wang, C. Hao, H. Liu and X. Zhang (2017) ADP-glucose pyrophosphorylase genes, associated with k e rn e l w e ig h t, u n d e rw en t s e l ect io n d u ri n g w h ea t domestication and breeding. Plant Biotech. J. 15: 1533-1543. Ishimaru, T., T. Hirose, T. Matsuda, A. Goto, K. Takahashi, H.
Sasaki, T. Terao, R. I. Ishii, R. Ohsugi and T. Yamagishi (2005) Expression patterns of genes encoding carbohydrate-metabolizing enzymes and their relationships to grain fi lling in rice (Oryza sativa L.): Comparison of caryopses located at different positions in a panicle. Plant Cell Physiol. 46: 620-628.
Kato, T. (1995) Change of sucrose synthase activity in developing endosperm of rice cultivars. Crop Sci. 35: 827-831.
Kato, T. (2004) Quantitative trait loci controlling the number of spikelets and component traits in rice: Their main eff ects and interaction with years. Breed. Sci. 54: 125-132.
Kato, T. (2010) Variation and association of the traits related to grain filling in several extra-heavy panicle type rice under diff erent environments. Plant Prod. Sci. 13: 185-192.
Kato, T. (2013) A trial to develop novel genotypes of extra-heavy panicle types with good grain filling in rice. J. Crop Res. 58: 15-20.
alleles for good grain fi lling at OsAGPS2 and OsSUT1 among a wide range of rice (Oryza sativa L.) cultivars. Breed. Sci. 61: 217-220.
Kato, T. and A. Horibata (2015) Alleles for good grain fi lling in rice extra-heavy panicle types and their distribution among rice cultivars. Proceedings of 8th Asian Crop Science Association Conference: 51-61.
Kato, T. and A. Horibata (2018) Distribution of alleles at the three loci relating to grain fi lling of rice among a wide range of cultivars including the world rice core collection. Abstracts of the 246th Meeting of the CSSJ: 43 (in Japanese).
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., D. Shinmura and A. Taniguchi (2007) Activities of enzymes for sucrose-starch conversion in developing rice endosperm and their association with grain filling of 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 fi lling in extra-heavy panicle type of rice. Crop Sci. 50: 2448-2456.
Khush, G. S. (2013) Strategies for increasing the yield potential of cereals: case of rice as an example. Plant Breed. 132: 433-436.
Lee, S. K., S. K. Hwaung, M. Han, J. S. Eom, H. G. Kang, Y. Han, S. B. Choi, M. H. Cho, S. H. Bhoo, G. An, T. R. Hahn, T. W. Okita and J. S. Jeon (2007) Identification of the ADP-glucose pyrophosphorylase isoforms essential for starch synthesis in the leaf and seed endosperm of rice. Plant Mol. Biol. 65: 531-546.
Luu, D. and C. Maurel (2013) Aquaporin trafficking in plant cells: An emerging membrane-protein model. Traffi c 14: 629-635.
Manson, T. G. and E. J. Maskell (1928) Studies on the transport of carbohydrates in the cotton plant: II. The factors determining the rate and the direction of movement of sugars. Ann. Bot. 42: 571-636.
Minchin, P. E. and A. Lacointe (2005) New understanding on phloem physiology and possible consequences for modelling long-distance carbon transport. New Phytol. 166: 771-779. Mohapatra, P. K., P. K. Sarkar and S. R. Kuanar (2009) Starch
synthesizing enzymes and sink strength of grains of contrasting rice cultivars. Plant Sci. 176: 256-263.
Münch, E. (1930) Die Stoff bewegungen in der Pfl anze. Verlag von Gustav Fischer, Jena: 234
Nakamura, Y., K. Yuki, S. Y. Park and T. Ohya (1989) Carbohydrate metabolism in the developing endosperm of rice grains. Plant Cell Physiol. 30: 833-839.
Nakatani, M. and M. Komeichi (1992) Relationship between starch content and activity of starch synthase and
ADP-glucose pyrophosphorylase in tuberous root of sweet potato. Jpn. J. Crop Sci. 61: 463-468.
Ohdan, T., P. B. Francisco, Jr., T. Sawada, T. Hirose, T. Terao, H. Satoh and Y. Nakamura (2005) Expression profi ling of genes involved in starch synthesis in sink and source organs in rice. J. Exp. Bot. 56: 3229-3244.
Oiestad, A. D., J. M. Martin and M. J. Giroux (2016) Overexpression of ADP-glucose pyrophosphorylase in both leaf and seed tissue synergistically increase biomass and seed number in rice (Oryza sativa ssp. japonica). Func. Plant Biol. 43: 1194-1204.
Panigrahi, R., E, Kariali, B. B. Panda, T. Lafarge and P. K. Mohapatra (2019) Controlling the trade-off between spikelet number and grain fi lling: the hierarchy of starch synthesis in spikelets of rice panicle in relation to hormone dynamics. Func. Plant Biol. 46: 507-523.
Patrick, J. W. and C. E. Offler (2001) Compartmentation of transport and transfer events in developing seeds. J. Exp. Bot. 52: 551-564.
Peng, S. (2009) Current status and challenges of rice production in China. Plant Prod. Sci. 12: 3-8.
Peng, S., K. G. Cassman, S. S. Virmani, J. Sheehy and G. S. Khush (1999) Yield potential trends of tropical rice since the release of IR8 and the challenge of increasing rice yield potential. Crop Sci. 39: 1552-1559.
Perez, C. M., A. A. Perdon, A. P. Resurreccion, R. M. Villareal and B. O. Juliano (1975) Enzymes of carbohydrate metabolism in the developing rice grain. Plant Physiol. 56: 579-583.
Sakai, M. and T. Ishii (2016) History and prospect on breeding for Super-High-Yielding Rice in Japan. Abstracts of the 245th Meeting of the CSSJ: 260 (in Japanese).
Sakulsingharoj, C., S. B. Choi, S. K. Hwang, G. E. Edwads, J. Bork, C. R. Meyer, J. Preiss and T. W. Okita (2004) Engineering starch biosynthesis for increasing rice seed weight: the role of the cytoplasmic ADP-glucose pyrophosphorylase. Plant Sci. 167: 1323-1333.
Smidansky, E. D., M. Clancy, F. D. Meyer, S. P. Lanning, N. K. Blake and L. E. Talbert (2002) Enhanced ADP-glucose pyrophosphorylase activity in wheat endosperm increases seed yield. Proc. Natl. Acad. Sci. USA 99: 1724-1729. Smidansky, E. D., J. M. Martin, L. C. Hannah, A. M. Fischer and
M. J. Giroux (2003) Seed yield and plant biomass increases in rice are conferred by degeneration of endosperm ADP-glucose pyrophosphorylase. Planta 216: 656-664.
Terao, T., K. Nagata, K. Morino and T. Hirose (2010) A gene controlling the number of primary rachis branches also controls the vascular bundle formation and hence is responsible to increase the harvest index and grain yield in rice. Theor. Appl. Genet. 120: 875-893.
Tetlow, I. J., M. K. Morrel and M. J. Eems (2004) Recent developments in understanding the regulation of starch metabolism in high plants. J. Exp. Bot. 55: 2131-2145. Warren-Wilson, J. (1967) Ecological data on dry-matter
production by plants and plant communities. In The collection and processing of fi eld data E. F. Bradley and Q. T. Denmead (eds.), Interscience, Sydney. 77-123.
Wang, G., H. Li, L. Feng, M. Chen, S. Meng, N. Ye and J. Zhang (2019) Transcriptomic analysis of grain fi lling in rice inferior grains under moderate soil drying. J. Exp. Bot. 70: 1597-1611. Wang, Z., X. Chen, J. Wang, T. Liu, Y. Liu, L. Zhao and G.
Wang (2007) Increasing maize seed weight by enhancing the cytoplasmic ADP-glucose pyrophosphorylase activity in
transgenic maize plants. Plant Cell Tissue Organ Cult. 88: 83-92.
Yamamoto, Y., T. Yoshida, T. Enomoto and Y. Yoshikawa (1991) Characteristics for the efficiency of spikelet production and the ripening in high-yielding Japonica-Indica hybrid and semidwarf Indica rice varieties. Jpn. J. Crop Sci. 60: 365-372 (in Japanese).
Yang, J., J. Zhang, Z. Wang, K. Liu and P. Wang (2006) Post-anthesis development of inferior and superior spikelets in rice in relation to abscisic acid and ethylene. J. Exp. Bot. 57: 149-160.
Yang, J. and J. Zhang (2010) Grain-filling problem in super rice. J. Exp. Bot. 61: 1-6.
