Genes and QTLs for Rice Grain Quality Improvement

As a major cereal crop, rice (Oryza sativa L.) is crucial to food security for at least half the world population. New varieties with high yield potential, good quality and high resistance to biotic and abiotic stresses are needed in order to meet the demand for more food arising from the rapid human population growth and concurrent decrease in arable land. Improvement of rice quality has now become a foremost consideration for rice buyers and breeding programs.


Introduction
As a major cereal crop, rice (Oryza sativa L.) is crucial to food security for at least half the world population. New varieties with high yield potential, good quality and high resistance to biotic and abiotic stresses are needed in order to meet the demand for more food arising from the rapid human population growth and concurrent decrease in arable land. Improvement of rice quality has now become a foremost consideration for rice buyers and breeding programs.
Quality is defined as "the totality of features and characteristics of a product or service that bears its ability to satisfy stated or implied needs" (International Standard Organization (ISO) 8402 1986). Features are identified properties of a product which can be related to the quality characteristics. Grain quality of rice is the totality of features and characteristics of rice or rice product that meets the demand of end-user. The concept of grain quality covers many features ranging from physical to biochemical properties, and includes milling efficiency, grain shape and appearance, cooking easiness, eating palatability, and nutrition. Thus, rice grain quality generally includes four classes, i.e. milling quality, appearance quality, cooking and eating quality, and nutritional quality ( Figure 1). Many countries have set up their own protocols to assess the respective quality. International organizations such as ISO, Association of Analytical Communities International (AOAC), and American Association of Cereal Chemists International (AACCI) have set up methods to evaluate some quality parameters, for example, apparent amylose content (AAC). Rice is consumed mainly as milled, so eating quality mentioned in this article generally relates to the cooked milled rice. However, due to the impact of the western life style, whole grain rice or brown rice becomes popular worldwide, so that the nutritional quality has expanded to the nutrients of brown rice.
Grain quality and its assessment are not only important to consumers, end-users, processors, but also to rice breeders who are engaged in creating rice varieties haboring new features such as high quality, high yield potential, highly resistant to abiotic or biotic stresses. It is necessary for rice breeders to understand how the quality traits are inherited from their parents. Genetic studies have revealed many genes and quantitative trait loci (QTL) for grain quality, though the grain quality traits are complex. Some major genes have been cloned, and their functions in a specific pathway, such as starch, protein, lipid, and flavonoids biosynthesis, have been characterized. Some QTLs have been finely mapped for further map-based cloning and functional characterization. The known genes or QTLs have been successfully applied in breeding programs for marker-assisted selection (MAS) to improve the breeding and selection efficiencies.
This chapter highlights the genes and QTLs available for grain quality of rice, summarizing how many QTLs and genes have been mapped or characterized, and how many could be used in marker assisted selection (MAS), which could help breeders to in-deep understand the genetics of grain quality of rice and apply the knowledge in their breeding practices. Milling quality determines the final yield and the broken kernel rate of the milled rice, which is of concern for consumers and farmers. Three main parameters, brown rice recovery (the percentage of brown rice to rough rice), milled rice recovery (the percentage of milled rice to rough rice), and head rice recovery (the percentage of head rice to rough rice) are used to evaluate the quality and efficiency of the milling process. Brown rice is the de-hulled rice with the palea and lemma removed. Brown rice itself is a type of whole grain that could be used for cooking and eating. Removing all of the bran which consists of the aleurone and pericarp, and germ or embryo from brown rice results in white (or milled) rice. Some milled grains are broken during milling, head rice is a standard term for the whole milled grain. In calculation of head rice recovery, kernels longer than or equal to 3/4 full length of a kernel were considered as whole grains. Among all three parameters to determine the milling quality, head rice recovery is the main factor determining rice market value and one of the most important criteria of milled rice.

Appearance quality
Appearance is one of the crucial properties of rice grain affecting its market acceptability. After milling, the appearance of the grain is associated with size, shape (long vs. round), chalkiness, and translucency. Grain length, width, thickness are used to describe the physical dimensions of rice kernels, while the grain shape is expressed as the ratio of length to width. Grain appearance is also largely determined the clarity, the vitreousness, and the translucency of the endosperm, which is specifically required by most segments of the rice industry. According to the location of the chalkiness in the endosperm, it could be classified into three groups, white belly (chalkiness on the dorsal side of the grain), white back (chalkiness on the ventral side) and white core (chalkiness in the center). Generally, the great the chalkiness, the lower the market acceptability. Percentage of chalky grain is the proportion of grains having a chalky spot on (or in) the endosperm. Chalkiness is measured visually with scales for 0 for none, 1 for small (<10%), 5 for medium (10-20%) and 9 for large (>20 % of the area). Grain transparency may be measured using a light permeation instrument or with an image analyzer, with which the size and shape may be measured simultaneously.

Cooking and eating quality
Cooking and eating quality determines the easiness of cooking, as well as the firmness and stickiness of the cooked rice. Rice cooking and eating quality is highly related to some easily measurable physicochemical properties: apparent amylose content (AAC), gel consistency, gelatinization temperature (GT) and pasting viscosity. All these parameters are related to the properties of starch that makes up 90% of milled rice. Starch consists of two kinds of molecules, the linear and helical amylose and the branched amylopectin. Amylose content is measured with a simplified procedure using I 2 -KI solution. Due to the binding ability of long chain of amylopectin with I 2 , the amylose content measured with I 2 -KI solution is also termed as apparent amylose content (AAC). The AAC of milled rice may be classified as waxy (1-2%), very low (5-12%), low (12-20%), intermediate (20-25%) and high (>25%). Gelatinization is the disruption of molecular orders within the starch granule manifested in irreversible changes in properties such as granular swelling, native crystallite melting, loss of birefringence, and starch solublization. The gelatinization temperature determines the time and energy input required for cooking. Gel consistency was developed as a parameter to index the tendency of cooked rice to harden on cooling, and is normally classified as hard, medium, and soft. Pasting viscosity is another useful parameter to differentiate rice with similar AAC, and is popularly measured by a Rapid Visco-Analyser (RVA) developed by Newport Scientific Pty Ltd., Australia. RVA records the viscosity continuously as the temperature is increased, held constant for a time, and then decreased.
The above mentioned are objective parameters for the cooking and eating quality. However, eating quality is quite subjective and thus is difficult to define as it depends on consumer preferences. Sensory quality of cooked rice could be evaluated by a trained sensory panel (Champagne et al. 2010). Four steps are used to evaluate the cooked rice texture (Table 1). In addition to texture, the flavor (aromatics, taste, mouthfeel) of cooked rice can also be evaluated by the sensory panel.

Nutritional quality
As one of the most important staple food in the world, nutritional quality is closely related to human health, and thus is highly valued by consumers. Protein is the second most abundant constituent of milled rice, following starch. Lysine is the first limiting essential amino acid in rice based on the human requirements. Protein and lysine content are two important parameters determining nutritional value of rice. With social development, diverse people eating rice as staple food may require rice with distinct nutritional quality. For those in the underdeveloped region where micronutrient deficiency (Vitamins and minerals, such as iron and zinc) is apparent, genetics study for and biofortification of micronutrients by breeding are necessary to improve the nutritional quality of rice. For those with improved living standards, consuming of brown rice as one kind of whole grains becomes popular to combat chronic diseases, such as diabetes. Whole grain rice (brown rice) provides more minerals, vitamins, dietary fibers, and phenolics to human health than milled rice (Bao 2012a).

Milling quality
Milling quality is assessed by brown rice recovery, milled rice recovery and head rice recovery, which is one kind of complex quantitative trait whose genetic control is poorly understood. Up to date, no major gene has been genetically identified and functionally characterized. However, many studies have been carried out to search quantitative trait locus (QTL) for the milling quality (Table 2). These researches improve our understanding of the genetic control of milling quality, and provide molecular markers that are useful in breeding for improvement of milling quality in rice.   from the cloned genes would lead to precise phenotype in breeding. QTL mapping studies also suggest that many QTLs exhibit pleiotropic effects; they control not only grain length, but also grain width, grain shape or grain yield    Table 3. The cloned genes for grain appearance quality (grain shape and endosperm chalkiness)

Grain length
A total of 47 QTLs for grain length have been detected in different populations. Among them, the Chromosome 3 harbors more QTLs than other chromosomes ( Figure 2). Up to date, two QTLs have been map-based cloned (Table 3), and seven QTLs have been finely mapped (Table 4).
GRAIN SIZE 3 (GS3) is the first QTL that has been map-based cloned for grain length. It was detected in the RIL population derived from Minghui 63 and Chuan 7, displaying a major role for grain length and weight and a minor role for grain width and thickness and functioning as a negative regulator for grain size (Fan et al. 2006;  . Genetic analysis of a near-isogenic line (NIL) for qGL3-1 revealed that the allele qGL3-1 from CW23 has an additive or partly dominant effect, and is suitable for use in molecular marker-assisted selection . A new variety containing the new allele shows increased grain yield, which indicates that GL3 is a powerful tool for breeding highyield crops (Qi et al. 2012).

Grain width
A total of 48 QTLs for grain width have been detected in different populations with more QTLs on chromosome 3 and 5 ( Figure 2). Up to date, four QTLs have been map-based cloned ( Table 3).
GRAIN WIDTH 2 (GW2) is a major QTL for rice grain width and weight, which was initially detected from a cross between a large-grain japonica rice variety (WY3) and a small-grain indica rice variety (Fengaizhan-1). GW2 encodes a RING-type E3 ubiquitin ligase (Song et al. 2007). WY3 has a 1-bp deletion resulting in the introduction of a premature stop codon in its exon 4, causing the large-grain phenotype. GW2 negatively regulates cell division by targeting its substrates to proteasomes for regulated proteolysis; loss of GW2 function results in an increase in cell number in the spikelet hull and acceleration of the grain-milk filling rate, thus enhancing grain width, weight, and yield. Functional markers developed from these major genes and finely mapped QTL resources allow breeders to efficiently manipulate grain size and shape (Tables 3 and 4).

Grain chalkiness
Chalkiness is a major concern in rice breeding because it is one of the key factors in determining quality and price. The chalky endosperm consists of loosely packed, round and large compound starch granules while the translucent endosperm comprises tightly packed, polyhedral and small single starch granules. The chalky grains show significantly different physicochemical, morphological, thermal, cooking and textural properties from translucent grains. Percentage of grains with chalkiness (PGWC) is one of the main indices of rice-determining appearance quality, which is easily determined.
Many factors contribute to the formation of chalkiness in the rice grain. Environmentally, rice grown at the higher temperature contains more chalky grains. Genetically, defect in genes affecting starch biosynthesis, starch granule structure, and grain filling may lead to endosperm chalkiness. These genes include starch branching enzyme IIb (BEIIb), branching enzyme I (BEI), starch synthase IIIa (SSIIIa), floury and sugary genes, etc. It should be noted that many of the genes characterized show pleiotropic effects on other traits in addition to chalkiness.
A rice genic male-sterility gene ms-h is recessive and has a pleiotropic effect on the chalky endosperm (Woo et al. 2008). Fine mapping and nucleotide sequencing analysis reveal a single nucleotide substitution at the 3'-splice junction of the 14th intron of the UDP-glucose pyrophosphorylase 1 (UGPase1) gene, which causes the expression of two mature transcripts with abnormal sizes caused by the aberrant splicing. Overexpression of UGPase1 in ms-h mutant plants restored male fertility and the transformants produced T1 seeds that segregated into normal and chalky endosperms (Woo et al. 2008).
The grain incomplete filling 1 (gif1) mutant defects in grain-filling capacity, but its grains are with more chalkiness as a result of loosely packed starch granules. A frameshift mutation caused by a 1-bp nucleotide deletion in GIF1 results in premature termination of its open reading frame. GIF1 encodes a cell-wall invertase required for carbon partitioning during early grain filling (Wang et al. 2008a).
Two white-core genes have been characterized with knockout mutants. A floury endosperm-4 (flo4) rice mutant with a floury-white endosperm but a normal outer portion was generated by T-DNA insertion into the fifth intron of the OsPPDKB gene encoding pyruvate orthophosphate dikinase (PPDK) ). Other two additional alleles, flo4-2 and flo4-3 also showed the same white-core endosperm phenotype. OsPPDKB was mainly expressed in the endosperm, aleurone, and scutellum of the developing kernel, suggesting that cytosolic PPDK functions in rice to modulate carbon metabolism during grain filling. Ryoo (Table 4).
qPGWC-8 is a major QTL for the percentage of grains with white chalkiness in the interval G1149-R727 on chromosome 8 which was identified using a chromosome segment substitution line (CSSL). Guo et al. (2011) narrowed down the location of this QTL to a 142 kb region between Indel markers 8G-7 and 8G-9. qPGWC-8 accounted for 50.9% of the difference in PGWC between the parents. qPGWC-7 is a QTL for the percentage of grain with chalkiness (PGWC) on 7 which was identified using a set of chromosome segment substitution lines, made from a cross between PA64s and 9311. Segregation analysis of the F 2 population from the cross between C-51 (a CSSL harboring qPGWC-7 and having a chalky endosperm) and 9311 showed PGWC is a semidominant trait, controlled by a single nuclear gene. Fine mapping of qPGWC-7 with a large F 2 population constructed from the cross C51 × 9311 delimitated it to a 44-kb DNA fragment, containing thirteen predicted genes ).
The markers tightly linked to qPGWC-8 and qPGWC-7 facilitate cloning of the gene underlying the QTLs and is of value for marker-assisted selection for endosperm texture. However, it is still far away from clear understanding the mechanism of formation of the grain chalkiness. First, the QTLs mapping results show low coherence in different genetic populations, suggesting many minor QTLs affecting chalkiness exist in different rice germplasm that we do not know. Second, in addition to the major genes or QTLs we have known, how their interactions with each other, and with the major genes for amylose and protein synthesis Zheng et al. 2012) that may affect chalkiness are unknown. Third, effect of environment on the formation of chalkiness is well known, but how its effect on the gene expression that leads to the formation of chalkiness is largely unknown.

Eating and cooking quality
Great progresses have been made in the understanding of the genetic basis of cooking and eating quality (Bao 2012b; Chen et al. 2012). Starch properties play important role in determining the cooking and eating quality, which is highly associated with starch biosynthesis related genes. Starch biosynthesis pathways and genes or enzymes participating in have been well clarified (Figure 3). Amylose is synthesized mainly by GBSSI, and the amylopectin synthesis process is governed by a combination of multiple isoforms of SS, BE, and DBE to produce a uniform number of chains per amylopectin cluster. Wx encoding GBSSI is mainly responsible for the natural variation of amylose content, gel consistency and RVA pasting viscosity, while the SSIIa is mainly for gelatinization temperature, thermal properties, and amylopectin structure (Bao 2012b  retrogradation, and the rice with SNP allele T gives a short, firm gel, and has a firm texture when freshly cooked with little change in texture over storage. In a cross between two varieties having similar high AAC, but with different paste viscosity properties, Traore et al. (2011) indicated that the exon 10 SNP marker is associated with most RVA pasting measurements and the proportion of soluble to insoluble apparent amylose.

Gelatinization temperature, thermal properties
SSIIa locus on chromosome 6 is a major QTL for gelatinization temperature and amylopectin structure

Contributions of other starch biosynthesis related genes
Cooking and eating quality is a complex trait which is not only determined by the Wx and SSIIa genes, but also other genetic factors, such as other starch biosynthesis related genes. Three

Other traits related to cooking and eating quality
In addition to the amylose content, gelatinization temperature, gel consistency and pasting viscosity, other parameters, such as water absorption, volume expansion and cooked rice elongation have been set up to evaluate the cooking characteristics of rice ).

Nutritional quality
Few molecular genetics studies have been conducted for nutritional quality (Table 5), but many molecular breeding activities through transgenic engineering to improve nutritional quality of rice have been reported (see 4.4).

Protein and amino acid content
There are nice reports about QTL mapping for protein content ( A QTL cluster on chromosome 1 was associated with the content of eight amino acids. The results are useful for candidate gene identification and marker-assisted breeding targeting the development of improved rice amino acid composition for human nutrition.

Fat content
Fat content affects eating quality and nutritional values, and storage stability of rice as well. Apparently, 48 QTLs for fat content have been reported. Chromosome 1, 3 and 6 harbor more QTLs than other chromosomes ( 2: The value in this column indicates chromosome number, the two or three same values in the same line indicate two or three QTLs in the same chromosome. 3: Percentage of total variation explained by a single QTL (%). Table 5. QTLs for protein content and fat content in the rice grain Rice -Germplasm, Genetics and Improvement orthologs of Arabidopsis genes encoding key enzymes in lipid metabolism co-localized with 11 mapped QTLs. A strong QTL for oleic (18:1) and linoleic (18:2) acid is associated with a gene encoding acyl-CoA:diacylglycerol acyltransferase, while another one for palmitic acid (16:0) is possibly associated with the acyl-ACP thioesterase gene. QTLs for these traits via association mapping using a diverse set of rice germplasm including red rice and black rice. Four, six and six QTLs were found associated with phenolic content, flavonoid content, and antioxidant capacity, respectively. Among them, four QTLs for phenolic content were also shared for other two traits. Ra (i.e. Prp-b for purple pericarp) and Rc (brown pericarp and seed coat) were main-effect loci for rice grain color and nutritional quality traits. Association mapping for the traits of the 361 white or non-pigmented rice accessions (i.e. excluding the red and black rice) revealed marker (RM346) is associated with phenolic content.

Minerals
Pigmented rice accumulates anthocyanins (black rice) and proanthocyanidin (red rice), which are benefit to human health. Genetically, the pericarp color of red rice was controlled by two complementary genes, Rc (brown pericarp) on chromosome 7 and Rd (red pericarp) on chromosome 1. When present together, these loci produce red seed color. Rc in the absence of Rd produces brown seeds, whereas Rd alone has no phenotype (Sweeney et al. 2006;Furukawa et al. 2007). A natural mutation in rc has reverted brown pericarp to red pericarp and resulted in a new, dominant, wild-type allele, Rc-g (Brooks et al. 2008). The color of dark purple pericarp was also controlled by two complementary genes, Pb and Pp, located on chromosome 4 and 1, respectively . Wang and Shu (2007) mapped Pb gene and suggested that this gene may be Ra gene. Markers for these genes may be useful for pigmented rice breeding, especially useful if new rice expects to accumulate both anthocyanins and proanthocyanidin.

Molecular breeding
Molecular breeding is the application of molecular biology tools in plant breeding, which is generally include marker assisted selection (MAS) and genetic engineering (genetic transformation) in addition to QTL mapping or gene discovery. Both of MAS and genetic engineering have been applied in grain quality improvement in rice. MAS has been successfully applied for cooking and eating quality improvement because of available of the excellent markers, while the genetic engineering has been widely used to improve nutritional quality of rice.

Marker assisted selection
QTLs underlying natural occurring variation in grain quality have been widely explored, however, only few of them have been applied in current rice breeding programs. To the best of our knowledge, most of reports in terms of improving grain quality simply mean to improve the eating and cooking quality. The most useful genes are Wx, SSIIa, and fragrance (Table 6). Functional markers developed from GS3 are also available for grain length improvement ). There are two strategies to conduct MAS in the breeding program. One is to improve the grain quality for the rice with high yield potential or high resistance to abiotic or biotic stresses, but with low quality. This is referred to foreground selection, which means that selection of a marker for grain quality trait by MAS denotes a trait obtained. Foreground selection is particularly useful for traits that need laborious or time-consuming phenotypic screening procedures, such as grain quality traits. The other is to improve the yield potential and high resistance for good quality rice, such as basmati or jasmine rice. This is referred to background selection. The markers for grain quality are used as background selection, which is to avoid the loss of good quality traits during introduction of the other traits.

Wx, fgr and SSIIa
Low quality of hybrid rice in China is mainly owing to its poor quality maintainer line.

Combining grain quality with other traits
Breeding is working for not only one trait, but all the traits for the formation of a new variety. In addition to grain quality traits, yield and other agronomic or resistance traits are also very important. For those rice cultivars already have good quality, the objective of MAS is to combine the important quality traits with other traits. There are special cases for basmati and jasmine rices which have premium grain quality, and have been widely accepted by consumers worldwide. MAS has been carried out to introduce bacterial blight resistance (Pandey et

Transgenic engineering
The advantage to conduct MAS is that abundant molecular markers are available for rice and many traits have been tagged with molecular markers. However, the disadvantage is that MAS is only effective when the target traits exist in rice germplasm, and becomes void when the traits of interest are not present in the rice germplasm. In this case, transgenic engineering is useful, which could introduce the new traits into rice by transferring the target gene from other species. Expression of exotic gene in rice could produce the target trait. Transgenic engineering has some successful examples to introduce new nutrient traits into rice grain, such as vitamine a (Va), that confers rice high nutritional and increased benefit to human health.

Resistant starch
Consumption

Protein
Expression of a gene encoding a precursor polypeptide of sesame 2S albumin, a sulfur-rich seed storage protein in transgenic rice plants results in the improvement of the nutritive value of rice; the crude protein content in rice grains was increased by 0.64-3.54%, and the methionine and cysteine contents of these transgenic rice grains were respectively elevated by 29-76% and 31-75% compared with those of wild-type rice grains (Lee et al. 2003). Over-expression of aspartate aminotransferase genes in rice results in altered nitrogen metabolism and increased amino acid content and protein contents in seeds ).

Va
Vitamin A deficiency has been linked to night blindness, corneal scarring and permanent blindness. Vitamin A deficiency increases infant mortality rates and the incidence and severity of infectious diseases. Carotenoids, a precursor of Vitamin A, is an important lipid-soluble antioxidants in photosynthetic tissues, which are known to be completely absent in rice endosperm. The entire β-carotene biosynthetic pathway in rice endosperm has been introduced into rice by transformation of plant phytoene synthase, Erwinia uredovora carotene desaturase, and lycopene β-cyclase genes via Agrobacterium-mediated transformation. The transgenic rice, Golden Rice 1, can accumulate a maximal level of 1.6 μg/g total carotene in the endosperm. Insertion of the phytoene synthetase gene from maize and the carotene desaturase gene from Erwinia uredovora into rice resulted in the greatest accumulation of total carotenoids and β-carotene. Golden Rice 2 contains as much as 37 μg total carotenoids per gram of dry weight of grain, of which 31 μg/g is β-carotene (Paine et al. 2005).

Folate
Folates are B vitamins (vitamin B9). Humans cannot synthesize folates and have to absorb them from the diet, with plants usually being the main dietary sources. Folates play roles in the prevention of neural tube defects and in reducing the risk of cardiovascular disease, colon cancer, and neuropsychiatric disorders. In the United States, folic acid is added to refined cereals and grain products; these products are major contributors to total folate intake. Rice is a poor source of folates (vitamin B9). Overexpressing two Arabidopsis thaliana genes of the pterin and para-aminobenzoate branches of the folate biosynthetic pathway, Storozhenko (2007) obtained transgenic rice with a maximal folate content enhancement as high as 100 times above wild type, with 100 g of polished raw grains containing up to four times the adult daily folate requirement.

Minerals (Fe)
Iron deficiency is the most widespread micronutrient deficiency world-wide that afflicts an estimated 30% of the world population, especially where vegetable-based diets are the primary food source. Expression of the soybean ferritin gene (Goto et al. 1999) or pea ferritin gene (Ye et al. 2007) in rice produced seeds with greater Fe contents. Especially, Vasconcelos et al. (2003) showed that expression of the soybean ferritin gene under the control of the glutelin promoter in rice has proven to be effective in enhancing grain nutritional levels, not only in brown grains but also in polished grains. Expression of a thermotolerant phytase gene from Aspergillus fumigatus in rice endosperm is expected to decrease the phytic acid and increase iron bioavailability (Lucca et al. 2001).

Flavonoids
Flavonoids are lacking in the endosperm of rice. Expression of maize C1 and R-S regulatory genes driven by an endosperm specific promoter of a rice prolamin gene in rice grain resulted in dark brown pericarp of the C1/R-S homozygous lines, and the major flavonoids, dihydroquercetin (taxifolin), dihydroisorhamnetin (3′-O-methyl taxifolin) and 3′-O-methyl quercetin were identified in the rice grain (Shin et al. 2006). These rice lines have the potential to be developed further as a novel variety that can produce various flavonoids in its endosperm.

Serotonin
Serotonin derivatives such as p-coumaroylserotonin and feruloylserotonin, a family of plant polyphenol compounds, play roles in an array of biological activities including antioxidative activity, but neither their production nor identification has been reported in crop plants.
Transgenic rice expressing the pepper hydroxycinnamoyl-CoA:serotonin N-(hydroxycinnamoyl) transferase gene produced on average 274 ng/g seed weight which was nine-fold higher than wild-type (30 ng/g seed weight) . Chemical treatments such as transcinnamic acid and tyramine increased the serotonin derivatives contents by two-to three fold in both wild-type and transgenic rice. The transgenic rice had higher radical scavenging activities than that of wild-type, suggesting that neutraceutical serotonin derivative could be enriched by transgenic engineering ).

Coenzyme Q
Coenzyme Q (CoQ), also called ubiquinone, is an electron transfer molecule in the respiratory chain. CoQ is also a lipid-soluble antioxidant. Most cereal crops produce mainly CoQ9, which has nine isoprene units, whereas humans produce mainly CoQ10, with 10 isoprene units.
CoQ10 is a very popular food supplement. Takahashi et al. (2009) produced CoQ10-enriched rice plants by introduction of the gene for decaprenyl diphosphate synthase. In CoQ10enriched rice plants, seed CoQ10 content per weight was increased to up to 10 times that of wild-type rice, but its level is still insufficient for practical use. Combination of the transgene with giant embryo mutant lines produced giant embryo line-type CoQ10-enriched rice with seed CoQ10 content per weight increased to up to 1.4-1.8 times. It was found that CoQ was preferentially accumulated in bran and germ of rice seed.

Future directions
Great progress has been achieved in our understanding of the genetic and molecular basis of grain quality of rice. This is especially true for grain appearance and grain shape, since they are not only linked with grain quality, but also with grain yield, a more important trait.
Cooking and eating quality has a strong relation with starch biosynthesis pathway which has been well understood. Markers derived from the starch biosynthesis related genes have been widely applied in MAS. However, there are four major problem areas that challenge researchers working on molecular genetics of grain quality.

Functional genes for milling quality and chalkiness
Genetic understanding of milling quality is quite poor since only limited numbers of QTLs have been detected, and no QTL has been finely mapped or cloned. To make in-deep research into the area of milling quality, (1) rapid and accurate analytical tools are needed to measure the trait; (2) finely dissection of QTLs with large effect should be carried out; (3) because no mutants for milling quality have been reported, the mutants such as those induced by T-DNA insertion may provide a good start to characterize the genes responsible for milling quality. For grain chalkiness, two finely mapped QTLs await further characterization, and transcriptome for chalkiness formation during seed development have been described (Yamakawa et al. 2007;Liu et al. 2010). It looks optimism to see more progress from this area.

Molecular genetics studies for nutritional quality
Nutrition quality of rice will be a new area for further research because people keep increasingly concern about the health benefit of the food they eat. Nutrition quality covers a wide range of traits, for example, protein, amino acids, fat and phenolics. In this area, naturally occurring variation for protein, amino acids, fat and fatty acid compositions have been under exploration, but only few genes have been characterized. Formation of each nutrient in rice grain requires a complex pathway in which many genes or enzymes are involved. Current advances in protein and fatty acid biosynthesis in other crops and Arabidopsis may help understand the pathways in rice.
Phenolics are expected to be an important field because they are proven to benefit human health in many ways (Shao and Bao 2012). Genes for red pericarp formation, Rc (brown pericarp) on chromosome 7 and Rd (red pericarp) on chromosome 1 have been under-stood, but their roles in regulating the flavonoids biosynthesis are unknown. The genes for dark purple pericarp formation, Pb and Pp, wait for finely mapping and functional characterization. In this field, MAS could be conducted to breed rice accumulating not only anthocyanins (a characteristic of black rice) and proanthocyanidin (a characteristic of red rice). Genetic transformation could be conducted to breed rice with accumulation of the anthocyanins or proanthocyanidin in the endosperm, since these phytochemicals accumulate only in the bran layer (Shao and Bao 2012).

Cooking and eating quality of brown rice
As concerns about nutritional quality rise, consumption of brown rice will become popular in the near future. Cooking and eating quality of brown rice will be another issue. The knowledge we have established for milled rice may not be applicable to the brown rice. Needless to say the genetic control of the cooking and eating quality of brown rice, what parameters to assess these qualities should be firstly considered. How to make brown rice appeal to consumers through suitable cooking methods should also be considered as well. At last, the question is how to improve the cooking and eating quality of brown rice.

MAS with more genes/QTLs together
Targeting more traits with more markers, such as Wx, SSIIa, and fragrance (Jin et al. 2010), is increasingly needed in the breeding programs. MAS for quality and yield and resistance traits should be considered together in the future. Strategies for more effective selection should be developed when many markers are used at the same time. In silica molecular breeding is coming into the era, with which alleles of different markers are designed in the computer; the phenotypes of new rice could also be designed and displayed in the computer.

Conclusion
Grain quality of rice as a whole is a complex trait that is comprised of appearance quality, milling quality, eating and cooking quality, and nutritional quality etc. Researches on the genetic control of the quality traits have made a great progress, especially for the appearance quality, cooking and eating quality. More genetic studies are needed for milling quality and nutritional quality.
The progress on the molecular genetics on grain quality has allowed MAS to be conducted more efficiently. However, only MAS for cooking and eating quality and genetic engineering for nutritional quality have made some achievements. More molecular breeding practices are needed for improvement of grain quality.
With social development and improvement of living standards, cooking and eating quality of brown rice will be a new theme that deserves greater attention from researches. Studies including cooking methods, parameters for cooking and eating, genetics, and molecular breeding are among the top priorities.