Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Rice is a major human food crop that feeds over 50 % of the world’s population. During the past decades applications of the “green revolution” and hybrid rice technologies have rapidly improved rice productivity. However, the previous breeding goal in last century was over shed light on yield; accordingly, the rice quality was not paid more attention to and the grain quality was not increased as fast as yield. The rice qualities include the appearance, milling quality, nutritional quality, and eating and cooking quality (ECQ) [1]; appearance and ECQs have received more attention than other qualities in rice from consumers. Given that the rice endosperm is the major eating part and starch is the predominant component in dehulled grains (composed of 76.7–78.4 % in polished rice), the rice ECQs are thus thought to be mainly influenced by starch properties [2]. The improvement in rice grain quality has been increasingly demanded by consumers and has become a priority for rice breeders and geneticists [3–5]. Studies on the elucidation of the molecular mechanisms underlying appearance and ECQs of rice have made significant advances recently. This chapter will review current progress in understanding the genetics and molecular biology of rice grain quality, focusing on ECQs and appearance of rice grains.

2 ECQs of Rice Grains

2.1 Properties Affecting Rice Grain ECQs

Three physicochemical properties of starch have been considered as determinants of rice grain ECQs: amylose content (AC), gel consistency (GC), and gelatinization temperature (GT) [6, 7].

Amylase content (AC), also called AAC (apparent amylose content) because of the measurement method used to assay amylose content by iodine staining, can detect both amylose and long chain amylopectin [8]. Based on AC values, rice varieties can be divided into two classes: glutinous rice (extremely low AC, 1–2 %) and non-glutinous rice (AC >2 %). Non-glutinous rice can be further classified into four types: very low AC (2–10 %), low AC (10–20 %), medium AC (20–25 %), and high AC (>25 %). In general, japonica rice contains low to very low AC values and tends to be sticky, moist, and tender when cooked. Generally, japonica rice is highly preferred in northern-east Asia, including China, Japan, and Korea. In contrast, indica rice contains high AC and cooks soft and fluffy in texture and is favored in southern and southeastern Asian regions [9].

Gel consistency (GC), as a good index of cold paste-viscosity of cooked rice, refers to the gel running distance of digested grains by KOH in a flattened tube [10]. Based on gel length, rice varieties can be divided into three classes: hard (<40 mm), medium (40–60 mm), and soft (>60 mm). Cooked rice with hard GC tends to be dry and fluffy after cooling, whereas soft GC rice remains moist and continues to remain soft after cooling. In general, soft GC rice is more desirable by consumers [11].

Gelatinization temperature (GT) is the temperature at which starch granules start to lose crystallinity and birefringence by irreversible expansion that alters the starch surface from polarized to a soluble state [12]. To measure rice grain GT, several approaches have been developed, such as differential scanning calorimetry (DSC), rapid visco analyser (RVA), and alkali spreading value (ASV). Based on the ability of seed resistant to alkali digestion, varieties can be classified into three categories: high GT (>74 °C), intermediate GT (70–74 °C), and low GT (<70 °C) [12]. Given that resistance of seeds to alkali digestion is antagonistic with GT, high GT varieties are substantially hard to be digested when eaten, and vice versa.

In general, rice varieties with fine ECQs can be characterized as having medium AC (15–17 %), soft GC (>60 mm), and low GT (<70 °C) [13].

2.2 Genetic Molecular Studies of AC

2.2.1 Inheritance of AC

Studies on domestic varieties have shown that long grain types (indica rice) can be characterized by relatively high amylose content (24–28 %), whereas typical short and medium grain varieties (japonica rice) have relatively low amylose content (15–20 %) [14], which has accelerated the application of AC as a selection criterion in rice breeding programs [14]. At present, three inheritance models of AC have been proposed. According to the first model, high AC is controlled by a single dominant major gene, along with some minor genes and/or modifiers [15, 16]. The second model supports the existence of two dominant or complementary genes that control the AC trait [17]. The third model hypothesizes that AC is a quantitative trait controlled by multiple genes [18]. In general, the first model is more popularly accepted by geneticists and breeders [15, 16, 19–22] and is supported by the transgressive segregation in F2 populations derived from low AC and intermediate AC parents [9, 16].

2.2.2 QTLs for AC

With the development of high-density marker linkage maps in rice, a series of studies for QTL analysis of rice grain quality have been conducted. Table 16.1 summarizes QTL that have been identified in rice for AC over the past decades. A doubled-haploid (DH) population, derived from the anther culture of an indica/japonica hybrid, was first utilized to identify QTL for AC [1] which identified two QTL on chromosomes 5 and 6. The major QTL on chromosome 6 explained 91.1 % of the variance and was closely linked with the previously identified Wx gene, which had been shown to control AC in both maize and rice [23, 24]. Thereafter, several groups have detected the Wx locus [6, 25–31] as the major AC QTL, as well as other minor effecter loci in diverse populations [1, 26, 28, 32–34]. Besides the popular DH population [6, 29–31], other populations, like F2, BC2, BC3, and CSSL, have also been developed and applied for positioning of AC loci [25, 27, 33, 34]. In addition, two pairs of epistatic QTL involving QTL-by-environment interactions (QEs) of AC have been detected as well [6]. These data indicate that the regulation of AC is complex and associated with a number of heritable factors and environmental conditions.

Table 16.1 Identified QTL for amylose content in rice
Full size table

2.2.3 Genes Regulating Amylase Content

Since the identification of two classical maize mutants defective in amylase in kernels [23] and endosperm in the last century [35], mutants with (Beijing) a similar waxy phenotype have been subsequently identified in rice, barley, wheat, potato, sorghum, and amaranths [24, 36–41]. The maize Wx gene was cloned in 1983 [42] and used subsequently as a probe to identify the homologous gene in rice [43], resulting in the identification of a 2.4-kb transcript that has been fully characterized [22, 44]. Surprisingly, besides the normal 2.4-kb transcript, an extra aberrant 4.0-kb transcript has also been found in glutinous rice from cultivar PI291667. Given that the translation of the maize and barley Wx genes begins in exon 2, raising the possibility that the extra-long Wx sequence may be attributed to the retention of the first intron [41, 45]. Sequencing the aberrant fragment of the Wx transcript from rice cultivar Hanfeng revealed that the entire intron 1 was indeed present in the aberrant long cDNA, and that the 3′ end of Wx cDNA included proper termination features of Poly(A) and an AATAAT sequence. In addition, it shares exactly the same sequence as its counterpart in non-glutinous rice [24, 44]. Furthermore, Northern-blot and Western-blot analyses of multiple varieties showed that low AC cultivars accumulate substantial amounts of un-spliced long Wx transcripts, including intron 1, and that high AC cultivars are depleted of the un-spliced transcript [24]. All of these data support the 5′ end retention hypothesis. Actually, previous data have shown that two Wx proteins, Wxa and Wxb, were present in rice. Wxa is characteristic of indica rice with high AC and Wxb is mainly found in japonica rice with intermediate AC [37]. In contrast to the Wxa-type rice, a T to G change in the 5′ splice site of intron 1 was detected in Wxb and a significant reduction of Wx a transcript amount was also observed [46]. These data further prove that the role of the G to T change in the first intron of the Wx gene affects transcription levels and final amylose biosynthesis. In the rice du1 mutant, Wx b transcription [47] and protein accumulation were reduced significantly [48]. Therefore, Du1 may participate in both Wx transcriptional and translational regulation. A recent study revealed that Du1 encodes a Prp1 protein, a component of spliceosome. The defect of Prp1 in the du1 mutant leads to a specific decrease of the splicing efficiency of Wx rather than other starch biosynthesis-related genes (SSRGs) [49]. Furthermore, the AC level of a du1/wx double mutant is almost the same as that of Du1/wx mutant, which is much lower than that of the du1/wx b mutant. In mammals, the U5-102kD Prp1 protein interacts with U4/U6 snRNPs and bridges the two particles through its TPR elements [50]. However, the molecular basis of why a Du1 mutant specifically affects the splicing of the Wx b pre-mRNA, and how Du1 recognizes the first intron of the Wx transcript remains to be elucidated [49]. In addition, the MYC (for v-myc avian myelocytomatosis viral oncogene homolog) protein has been shown to interact with EREBP (for ethylene responsive element binding protein) and bind to the Wx gene promoter, which results in enhanced transcription of the Wx gene [51, 52]. All these results show that transcriptional and posttranscriptional regulation of the Wx gene is crucial for the expression of Wx and amylose biosynthesis.

2.3 Genetic Molecular Studies of Gel Consistency (GC)

2.3.1 Inheritance of GC

Gel consistency (GC) is a good index of cooked rice texture for cold paste-viscosity, especially among rice varieties with high amylose content and varieties with hard, medium, and soft GC levels have been selected for by rice breeders with their breeding goals. To explore the genetic inheritance of the three types of GC levels, bulked F2 and F3 seeds were analyzed and the hard GC was found to be controlled by a single dominant gene [19]. Subsequent investigations using a single grain analysis resulted in a similar conclusion [11, 53]. The inheritance of GC was further explored by utilizing various populations, such as F2, B1F1, and B2F1, derived from parents with hard and soft, hard and medium, and medium and soft GC properties as well, and a major gene with multiple alleles was identified [54]. These studies all suggested that the hard GC is controlled by a single locus/gene. However, analyses using a 6X6 diallel set excluding reciprocals and involving contrasting parents revealed the predominance of additive gene action in the regulation of GC trait expression [55]. A similar result was also observed by Yi and Chen [56], implying that the GC is unlikely controlled by a major gene.

2.3.2 QTL for GC

Like AC, a number of studies have been conducted to understand the genetic basis of GC [25–29, 31–34, 57, 58]. Because of the negative correlation between AC and GC, it was suggested that GC is controlled by Wx, or another gene closely linked to the Wx locus [5, 25]. To identify QTL responsible for GC, an RIL population was developed and a single QTL with a large MS effect was detected on chromosome 6, which corresponded well with the Wx gene [25]. Furthermore, using two additional RIL populations, major QTL for GC on chromosome 6 had been detected as well, accounting for 57 % and 53 % of the phenotypic variation, respectively [26, 57]. The Wx locus responsible for GC was also identified with different DH populations [5, 29, 31]. All of these results indicate that the Wx locus is the major candidate gene controlling GC. Besides Wx, several minor effect QTL (Table 16.2), located on different chromosomes, were also identified from different populations [25–29, 31–34, 57, 58]. Therein, two loci identified using a DH population from similar GC parents were found not to overlap with known SSRG genes [32]; therefore, it should be interesting to clone these novel GC regulatory genes.

Table 16.2 Identified QTL for gel consistency in rice
Full size table

2.3.3 Genes Regulating GC

Based on genetic studies, numerous reports suggested that Wx was the primary determinant of GC; however, no direct molecular evidence was available until recently. The major QTL on chromosome 6, qGC-6, was represented with a DH population and this locus was finally characterized by using chromosome segment substitution lines [5]. qGC-6 encodes a granule-bound starch synthase (Wx), which has been well-documented for its role in AC. A comprehensive comparison revealed several polymorphic sites, including a previously known G/T transition between CJ06/TN1 parents. Although the complementation experiment had confirmed the role of the Wx gene in GC, which or how many SNP or InDel sites in the Wx gene are pivotal for GC and starch biosynthesis remains to be determined. Besides Wx, other starch-related genes, like ALK and SBE3, have been shown to play a role in regulating GC as well. When a functional ALK gene from low GC rice (Shuangkezao, indica type) was transferred into the GC intermediate Nipponbare (japonica type), transformants showed a decreased GC value [59]. However, the opposite phenotype was also observed in rice plants over-expressing the ALK gene, in which the GC value increased significantly [7]. Therefore, it will be interesting to elucidate the molecular mechanism of ALK in regulating GC in future. In addition, transformation of a functional SBE3 gene into japonica cv. WYJ7 decreased the GC value significantly [7], indicating that SBE may function as a minor gene contributing to GC.

2.4 Molecular Genetic Studies of Gelatinization Temperature

2.4.1 Inheritance of GT

The inheritance of GT has been studied extensively, but its mode has not been found to be consistent, not only in the number of genes responsible for GT but also in the nature of dominance-recessive relationships [60, 61]. Puri et al. reported the segregation patterns of GT in five reciprocal cross populations derived from three different GT (high, medium, and low) parents, but and could not identify consistent mode of inheritance, suggesting the lack of a major gene in controlling GT [61]. A similar conclusion was also drawn from studies on crosses between the cultivar 9192 and the mutant mahsuri [60] and an indica cytoplasmic male sterile (CMS) line and its restorer lines, respectively [62]. However, a bimodal frequency distribution was detected in an F2 population between SD7 and 72-3764, indicating that a major locus is governing GT [17]. Moreover, a few additive genes with major effects, along with modifier genes, were proposed as well [17]. Lastly, it should be pointed out that GT inheritance has been shown to be affected by environment, whereby high air temperature after flowering raises GT and lower temperatures have the opposite effect [60, 63].

2.4.2 QTLs for GT

As mentioned above, the inheritance model of GT had been well explored using crosses among diverse cultivars. To identify the genes responsible for GT, a number of QTL mapping populations (e.g., F2, RILs, BILs, DH, and CSSLs) have been developed and are listed in Table 16.3. Among them, DH populations have been widely employed [1, 6, 29, 30, 32] and resulted in the identification of important two QTL [1]. One QTL was found to be a major contributor to GT and was delimited to the CT506-C235 interval on chromosome 6. This QTL was recognized by several labs using diverse population types [6, 26, 28, 29, 58]. The CT506-C235 region contains a known locus, ALK, which is responsible for alkali digestion, indicating the pivotal role of ALK for GT. Besides ALK, the Wx gene, with both major and minor effects, has been identified as well [6, 26, 28, 29, 32, 34].

Table 16.3 Identified QTL for gelatinization temperature in rice
Full size table

2.4.3 Cloning ALK—A Gene That Regulates GT

To clone ALK, a locus that has been shown to regulate GT, segregating F2 populations were utilized to map and fine map ALK to a 9-kb region on chromosome 6 between the genetic markers R2147 and C1478 [12, 64]. BlastX analysis revealed a partial ORF that encoded a soluble starch synthase IIa (SSIIa or ALK) within the 9-kb region and three amino acid substitutions in the ALK genes of the parental lines C Bao (low GT) and Shuangkezao (high GT) were detected [12]. Considering the previous data that ALK is functional in elongation of medium chain-length amylopectin [65], it is therefore interesting to understand the role of different amino acid substitutions in ALK in regulating chain-length elongation. To do that, an elaborate set of expression shuffle constructs with diverse amino acid residue substitutions in the wild-type cv IR36 ALK sequence were established, and the purified proteins were utilized to test their SSIIa activity in vitro [66]. Results showed that the replacement of Val-737 with Met-737 abolished SSIIa activity in chain-length elongation from the degree of polymerization (DP) <12 to DP 13–25, indicating a critical role of the Val-737 site for ALK function. In contrast, substitution of the site Phe-781 only partially affected ALK activity, and double mutations of residues Gly-604 and Phe-781 were shown to enhance the deficiency of ALK activity. These observations suggest that the Phe-781 is an important secondary site and that Gly-604 may interact positively with Phe-781 to determine ALK activity [66].

2.5 Complex Network Regulating Starch Biosynthesis

As rice grain ECQs are triploid endosperm traits, their inheritance patterns are very complicated because the genetic expression of an endosperm trait in cereal seeds is conditioned not only by the triploid endosperm genotype, but also by the diploid maternal genotype, and additional cytoplasmic components [1, 67, 68]. Several studies concerning the inheritance of rice grain ECQs have been conducted over several decades, but the data are not always consistent. The generally accepted model is that nuclear gene expression is the predominant mechanism affecting rice grain quality, though a few studies have suggested that the chloroplast genome may play a role as well [56, 69].

Based on current knowledge, starch structure is determined by four classes of enzymes: ADP glucose pyrophosphorylase (AGPase), starch synthase (SS), starch branching enzyme (BE), and debranching enzyme (DBE) [70, 71]. As previously described, the two SS enzymes, Wx and ALK, have been well documented to affect AC and GT. Accumulating evidences have shown that Wx is involved in amylose biosynthesis, especially in the formation of extra-long chain fractions [72, 73]. In contrast, ALK has been shown to determine the elongation of short- to medium-length starch chains (DP 13–25) [12, 59, 66]. In addition to Wx and ALK, other SSRGs, such as OsSS1 [74], OsSSIIIa [75], isoamylase1 [71], branching enzyme [76, 77], and pullulanase [78], confer their unique or overlapping roles on the formation of fine starch structure and rice quality regulation as well (Table 16.4).

Table 16.4 Characterized starch biosynthesis and regulation-related genes in rice
Full size table

As the major SS isozyme in developing endosperm, SS1 mutations have not been shown to have an influence on the size or shape of grains and starch granules, or on the crystallinity of endosperm starch. However, ss1 mutants do lead to an obvious decrease in DP 8–12 chains and an increase in DP 6–7 and DP 16–19 amylopectin chains [74]. In contrast, the ssIIIa mutant exhibited obvious deficiency phenotypes in grain shape, and the internal amylopectin chains, DP 6–9 and DP 16–19, were shown to decrease, while DP 10–15 and DP 20–25 chains increased [75, 79]. The opposite chain-length deficiency in ss1 and ssIIIa mutants strongly indicates their distinct and overlapping functions in amylopectin biosynthesis. While the double ss1 ssIIIa null mutant is sterile, double mutants with a leaky ss1 and a null ssIIIa allele are fertile. These results further support the role of both SSI and SSIIIa in starch biosynthesis in rice endosperm and seed development [80].

SBE1 and SBEIIb are two well-characterized starch branching enzymes from rice. The mature grain of sbe1 mutant looks like that of the wild type, not only in appearance but also in the size and weight, whereas sbeIIb mutant has significantly smaller kernels with a floury appearance [77]. Biochemical analysis revealed that the AC level of endosperm starch in sbe1 mutant was similar to that of the wild type, while significant decreases in both long chains (DP >37) and short chains (DP12-21) of amylopectin were observed [77]. In contrast, the sbeIIb mutant was specifically reduced in short DP <17 chains, with the greatest decrease in DP 8–12 chains to alter the structure of amylopectin in the endosperm [76].

For starch DBEs, two genes, ISA1 [71] and PULL [78, 81], have been well characterized. An antisense transgenic line of isa1 contains increased levels of short chains (DP <12) and is depleted in intermediate-size chains (DP 13–23) [71, 82]. In contrast, the pull mutant showed an increased level of short chains of DP <13 [78, 81, 83]. The pull-null/mild isa1 double mutant still retained starch in the outer layer of the endosperm tissue, while amounts of short chain amylopectin (DP ≤7) were higher than that of the isa1 mutant. These data indicate that the function of PULL is partially overlapping with that of ISA1 and its deficiency has less impact on the synthesis of amylopectin than that of ISA1 [83].

Even though QTL mapping and cloning of starch synthesis genes have provided useful information for rice grain quality, it has been difficult to isolate QTL/genes with minor effects in order to elucidate the complex network of starch synthesis due to limited germplasms used in single experiments. To gain a broader understanding of the starch synthesis and its regulatory network, 18 SSRG genes were selected as candidates to carry out an association study [7]. With this approach, a fine network of rice starch biosynthesis and regulation was established. As a result 10 of the 18 SSRGs have been shown to be associated with rice grain quality. Both Wx and ALK are two central determining factors affecting all three properties (AC, GC, and GT). Wx functions as the sole major gene for both AC and GC, but as a minor gene affecting GT, consistent with QTL mapping results, whereas ALK was found to be the sole determinant for GT, but as a minor gene affecting AC and GC. In addition, several genes were shown to be associated with minor effect on starch biosynthesis: SSIII-2, AGPlar, PUL, and SS1 for AC; AGPiso, ISA, and SBE3 for GC; and SSIV-2, ISA, and SBE3 for GT [7]. So far, this is the first genome-wide study of how the allelic diversity of SSRGs has collectively been shown to regulate rice grain quality via the starch biosynthesis network.

3 Grain Appearance

3.1 Features Affecting Rice Grain Appearance

The quality of grain appearance is mostly determined by grain shape as specified by grain length (GL), grain width (GW), grain length/width ratio (GS), and grain chalkiness [84]. Although preferences for rice grain appearance vary by consumers, long and slender rice is generally preferred by most of consumers in north-America and Asian countries [84]. Based on the Chinese national criteria for rice quality, the grain length of grade I rice is 6.5–7.5 mm for indica and 5.0–5.5 mm for japonica rice; and the ratio of grain length/width is >3.0 for indica and 1.5–2.0 for japonica rice.

3.2 Inheritance of Grain Shape

Grain shape is determined by grain length, grain width, and/or the ratio of grain length to width. Studies on grain shape have been explored extensively [17, 85–87], not only because of its elegant appearance from a visual sense, but also because of its strong effect on yield improvement due to its positive correlation with grain weight [88]. Genetic studies from different crosses among japonica X japonica and indica X japonica have showed that there are no obvious differences in reciprocal backcross progeny and that continuous distribution patterns were observed in progeny populations, suggesting that grain length is governed by quantitative maternal nucleic genes [89–91]. Similarly, observations using F2 populations derived from crosses between varieties with different grain width also detected a continuous grain width distribution, suggesting a polygene model for grain width [17, 86, 92–94]. However, numbers of genes responsible for grain shape appear to be variable and are probably dependent on their genetic backgrounds. For example, Liu [95] found that the grain length segregation ratio in a specific F2 population was 3:1 suggestive of a single gene model controlling grain shape, whereas Xu et al. found a transgressive segregation which correlated with a QTL gene model that controlled grain shape in a similar F2 population [90].

3.3 QTLs for Grain Shape

Over 20 QTL mapping studies have been conducted to understand the genetics of grain shape and hundreds of responsible loci had been detected: 119 for grain length, 90 for grain width, and 60 for grain length/width ratio [96]. Among these studies, some QTL have been shown to account for major effects (Table 16.5) [1, 84, 97–103]. Lin et al. utilized F2 populations derived from two pairs of indica parents with significant differences in grain shape to detect QTLs affecting grain length, width, and thickness. Consequently, 14 QTLs were detected [99], 5 for grain length, 2 major and 2 minor genes for grain width, and 1 major and 4 minor genes for grain thickness. Huang et al. developed a DH population from IR64 and Azucena parents for QTL mapping. Twelve QTL affecting grain shape were localized onto 5 different chromosomes, among them 4 for grain length, 5 for grain width, and 3 for length/width ratio [104]. Using an F2 population, Redona and Mackill [100] identified 7 QTL for GL, of which two loci on chromosomes 3 and 7 with high LOD values had already been identified by Takeda and Saito [105] and Takamure and Kinoshita [106]. It should be pointed out that the QTL on chromosome 7 for GL also affected GW and grain length/width ratios as well. Xing et al. [107] utilized an RIL population to analyze QTL for grain shape and identified a major QTL (GW5) on chromosome 5 which is responsible for all three features of grain shape indicating that these loci may function as positive regulators to increase grain weight. By using F2:3 and RIL populations derived from crosses between Zhenshan97 and Minghui63, a major QTL for GL on chromosome 3 (GS3) and a major QTL for grain width on chromosome 5 (GW5) were identified [84]. GS3 was also detected by several groups using different mapping populations [28, 84, 104, 108–110], indicating its general role in grain length determination. In addition, several independent studies identified a number of QTL for rice grain width using diverse mapping populations [96]. Among them, WG5 and WG7, which were first described by Lin using an F2 population derived from Tesanai 2 × CB1128 [99], were found to have significant contributions to the total grain width [84, 101, 103]. Recently, by using an F2 population derived from Zhonghua11 × Baodali (a variety with larger grain size), two major QTL for GW located on chromosomes 3 and 6 were identified [97]. In addition, a major QTL on chromosome 8 (GW8) has been detected by several groups in a number of populations [111]. These QTL have laid a solid foundation for further gene cloning and understanding of the regulation of grain width.

Table 16.5 Major QTL identified for grain shape in rice
Full size table

3.4 Genes Affecting Rice Grain Shape

Because of the strong correlation between grain shape and yield, significant efforts have been made to fine map and clone genes that regulate grain shape [110–113]. To clone the major QTL for grain length, GW3.1 or GS3 [84, 102, 114] was first fine-mapped to a 93.8-kb interval on chromosome 3 using a BC2F2 population derived from a cross between Jefferson and O. rufipogon. GS3 was then cloned using an F2 population derived from Minghui63 and Chuan7 as was proposed to be a loss-of-function mutation of a putative transmembrane protein [113]. Protein domain analysis indicated that GS3 may have four putative domains: a plant-specific organ size regulation (OSR) domain at the N terminus, a transmembrane domain, a tumor necrosis factor receptor family cysteine-rich domain, and a von Willebrand factor type C (VWFC) domain at the c-terminus. To elucidate the roles of these domains, a series of transformation assays with different protein truncations were conducted, showing that the OSR domain is essential and sufficient for GS3 to function as a negative grain size regulator [112].

A number of studies have been reported on the mapping of GW QTL (Table 16.5). The GW2 gene was first cloned by using a BC3F2 population derived from a cross between WY3 and Fengaizhan [115]. GW2 encodes a RING-type E3 ubiquitin ligase and WY3 GW2, truncated by 310 amino acids, still possesses intrinsic E3 ligase activity, suggesting that the C-terminal of GW2 is not essential for substrate degradation. Mutations of GW2 result in increased cell numbers and acceleration of grain milk filling rate, which in turn enhances grain width, and yield [115]. Shomura et al. [101] performed a QTL analysis with an F2 population derived from Nipponbare × Kasalath and identified and subsequently cloned qSW5 (GW5), which explained 38 % of the natural variation in the F2 population. Sequence comparison of qSW5 between the two parents revealed a 1,212-bp deletion in (Nipponbare) and several SNPs. Further complementation experiments and sequencing of qSW5 from additional cultivars revealed that the 1,212-bp deletion played an important historical role in rice domestication [101]. To understand the grain width difference between Asominori and IR24, Weng et al. [116] also found the same 1,212-bp deletion in the GW5 gene [116]. Recently, two previously identified GW QTLs, GS5 and GW8, were fine-mapped and cloned [84, 107, 117, 118]. GS5, which encodes a putative serine carboxypeptidase belonging to the peptidase S10 family and has a PF00450 consensus domain, may function as a positive regulator of grain size by affecting grain width, filling, and weight [117]. In addition, sequencing the promoters of 51 rice accessions from diverse geographic regions identified three haplotypes that appear to be associated with grain width [117].

Previous studies revealed a major QTL GW8.1 for grain width [119], which was fine-mapped to a 306-kb region [111] on chromosome 8. However, there is no report to date on the cloning of this gene. Recently, Wang et al. [118] reported the cloning of a major gene (GW8) on chromosome 8, which does not appear to be allelic to GW8.1. GW8 encodes a Squamosa promoter-binding protein-like 16, which belongs to the SBP domain family of transcription factors and shares homology with TGA1, a domestication syndrome gene associated with the formation of naked grains in maize.

4 Perspective for Rice Quality Improvement

Development of new cultivars with improved grain quality for eating, cooking, and grain shape is critical for rice production. Although significant efforts have been made to understand the nature of grain quality, a comprehensive molecular understanding of these phenotypes remains elusive. For example, “Yangzhou fried rice,” a popular food cooked with long grain rice, has better palatability and morphology than short grain rice (Fig. 16.1). Those observations lead to the conclusion that the network of starch biosynthesis and rice quality regulation is complex. In this chapter we summarized what is presently known about the inheritance and molecular basis of grain quality characteristics and outlined the strategies for the development of high-quality rice in the future.

Fig. 16.1
figure 00161

(a) Three major types of rice showing their different morphologies and properties. (b) Yangzhou fried rice, a popular Chinese dish cooked with long grain length (GL) japonica rice

Full size image

QTL mapping has been extensively and successfully applied to clone major genes that affect grain quality, especially for grain shape and ECQ. Unfortunately this approach has achieved limited success toward the cloning of minor grain quality genes. Two recently developed techniques, i.e., co-expression and association mapping, have been shown to be very useful for the identification of ECQ genes [7, 51]. It is therefore expected that fine-scale regulators that control grain quality will soon be identified by employing these methods, and subsequently cloned.

Even though rice supplies about 20 % of the world’s dietary energy and is a good source of thiamine, riboflavin, and niacin [120], its protein content is much lower than other cereal crops (only 6.3–7.1 % in milling rice). More importantly, multiple micronutrient factors for human health like vitamin A, B, C, and D are defective in milled rice grain. For example, vitamin A deficiency in humans exacerbates afflictions such as diarrhea, respiratory disease, and childhood diseases such as measles [121]. Therefore, an important goal in rice grain quality research and breeding programs is to improve micronutrient content in rice endosperm. Efforts to improve vitamin A content in rice endosperm have been highly successful by the use of recombinant DNA technology to introduce 3 essential β-carotene biosynthetic pathway genes (PSY, LCY, CRTL) into the rice genome [121]. In addition, rice endosperm protein content has been significantly increased by transformation of the β-phaseolin seed storage protein gene from common bean [122, 123]. These examples demonstrate that it is possible to significantly improve rice endosperm grain quality and opens the door to engineer additional micronutrient and protein enhancements in the future.

With the world’s population expected to increase from 7 to 9 billion inhabitants by 2050, rice breeders have been challenged to produce new cultivars that can grow with less water, fertilizer, and pesticides and have doubled yields. This challenge must also account for global warming. Preliminary data has shown that high temperatures have adverse effects on rice productivity and quality [124–128], such as the decreased brown rice rate and milling rice rate. The comparison of grain quality gene expression patterns under high and low temperatures showed that the grain quality deterioration pathway may proceed through increased sucrose synthase activity and different SSRG gene expression [127, 129]. Therefore, it is extremely important to investigate the molecular mechanisms of how high temperature signal transduction affects grain quality, which in turn will facilitate the development of new elite cultivars that have higher productivity, better quality, richer nutrients, and greater adaptation to global climate changes.