Chalky part differs in chemical composition from translucent part of japonica rice grains as revealed by a notched-belly mutant with white-belly.

BACKGROUND: Chalkiness has a deleterious influence on rice appearance and milling quality. We identified a notched-belly mutant with a high percentage of white-belly, and thereby developed a novel comparison system that can minimize the influence of genetic background and growing conditions. Using this mutant, we examined the differences in chemical composition between chalky and translucent endosperm, with the aim of exploring relations between occurrence of chalkiness and accumulation of starch, protein and minerals.
RESULTS: Comparisons showed a significant effect of chalkiness on chemical components in the endosperm. In general, occurrence of chalkiness resulted in higher total starch concentration and lower concentrations of the majority of the amino acids measured. Chalkiness also had a positive effect on the concentrations of As, Ba, Cd, Cr, Mn, Na, Sr and V, but was negatively correlated with those of B, Ca, Cu, Fe and Ni. By contrast, no significant chalkiness effect on P, phytic acid-P, K, Mg or Zn was observed. In addition, substantial influence of the embryo on endosperm composition was detected, with the embryo showing a negative effect on total protein, amino acids such as Arg, His, Leu, Lys, Phe and Tyr, and all the 17 minerals measured, excluding Ca, Cu, P and Sr.
CONCLUSION: An inverse relation between starch and protein as well as amino acids was found with respect to chalkiness occurrence. Phytic acid and its colocalized elements K and Mg were not affected by chalkiness. The embryo exerted a marked influence on chemical components of the endosperm, in particular minerals, suggesting the necessity of examining the role of the embryo in chalkiness formation.

BACKGROUND

Chalkiness is the opaque portion found in an otherwise translucent white endosperm of ass="Species">rice, having a loose structure due to incomplete accumulation of ass="Chemical">pan class="Chemical">starch and protein. It is undesirable because of the substantially negative effect on appearance and milling quality. Worldwide, breeding chalkiness out of rice has been one of the primary goals in rice improvement.1 Grain chalkiness is a complex trait controlled by quantitative trait loci (QTLs). More than 140 QTLs contributing to grain chalkiness have been mapped across all 12 chromosomes of the rice genome. Among them, only a few QTLs have been isolated and functionally analyzed, and a few associated genes have been identified.2, 3 The slow progress in the study of chalkiness partially reflects the complexity of the mechanism underlying chalkiness occurrence, which needs intensive studies.

During grain filling, assimilates such as ass="Chemical">sucrose, ass="Chemical">pan class="Chemical">asparagine and glutamine are translocated from source organs and stored mainly in the endosperm of grains as starch and protein. Similarly, minerals such as phosphorus (P), potassium (K) and magnesium (Mg) are also translocated into grains, but deposit mainly in the aleurone layers. Basically, the rice grain stores these nutrients for the germination of its next generation, while humans use them as principal sources of food and nutrition.4 Thus the chemical composition of rice grains is of both biological and economic significance. In addition, it can serve as a comprehensive index indicating the cumulative effect of gene and environment on accumulation of starch, proteins or minerals during the whole filling stage. Thus research into the chemical compositions of grains should find a way of elucidating the mechanism underlying the accumulation of these chemical components.

There are numerous reports concerning the examination of chemical composition between chalky grains and translucent grains. Examined by scanning electron microscopy (SEM), disorganized ass="Chemical">starch granules was noted as a common feature of the oass="Chemical">paque tissues, indicative of the incomplete development of amyloplast.5, 6 Chemical anaass="Chemical">pan class="Chemical">lysis further showed that chalky grains differ from translucent kernels with respect to amylose content7 and amylopectin fine structure.8, 9, 10 Generally, chalky grains contained lower amylose content and showed a higher ratio of short branch‐chain of amylopectin in comparison with translucent grains, indicating that starch synthesis in chalky grains may slightly favor glucan chain branching over chain elongation. The findings are of value to clarify mechanisms responsible for chalkiness formation and thus provide directions in rice improvement and management.

By contrast, some studies did not find a significant difference between chalky and translucent grains in pan class="Chemical">amylose, amylopectin or protein components.11 The discreass="Chemical">pancy between results of these studies may be associated with comass="Chemical">parisons between whole grain samples of different varieties or from chalky grains and translucent grains for a given variety. These comass="Chemical">parisons have limitations in that they cannot avoid the effect of genotype for grains between varieties, and the effect of growing environment for grains even from the same variety, owing to the significant variations in chemical composition among grains from different positions within a ass="Chemical">panicle.12, 13

To minimize the influence of genetic background and growing environment, we developed a novel comparison system based on a notched‐belly mutant with a high ratio of white‐belly grains (Fig. 1), by which a comprehensive survey of endosperm proteomics was performed to unravel the molecular and biochemical basis of grain chalkiness. 8 The diverse but delicately regulated pathways responsible for grain chalkiness were clarified, which need extensive investigations. Using tass="Chemical">his system, the objectives of tass="Chemical">pan class="Chemical">his study were to measure the chemical composition (starch, proteins, amino acids and minerals) of the chalky and translucent parts of the white‐belly mutant, and thereby to uncover clues as to the formation of grain chalkiness.

Figure 1

Experimental scheme for chemical analysis of translucent and chalky grains. T1 and T2 are the upper and bottom half part of translucent grain, respectively. C1 and C2 are the upper and bottom half part of chalky grain, respectively. T1, T2 and C1 are translucent, while C2 is opaque. T1 accounts for 47.06% of the translucent grains, and T2 for 52.94%. C1 consists of 42.48% chalky grains, and C2 57.52%. The embryos of C2 and T2 were removed.

Experimental scheme for chemical anapan class="Chemical">lysis of translucent and chalky grains. T1 and T2 are the upper and bottom half ass="Chemical">part of translucent grain, respectively. C1 and C2 are the upper and bottom half ass="Chemical">part of chalky grain, respectively. T1, T2 and C1 are translucent, while C2 is oass="Chemical">paque. T1 accounts for 47.06% of the translucent grains, and T2 for 52.94%. C1 consists of 42.48% chalky grains, and C2 57.52%. The embryos of C2 and T2 were removed.

METHODS

Plant material

A notched‐belly mutant (DY1102) with high occurrence of white‐belly grains was used in tass="Chemical">his study. Identification and characterization of the mutant were presented in detail in our previous report.8 Briefly, the mutant has a high percentage of notched‐belly grain (83%), of which 85% had white‐belly that occurs only in the bottom ass="Chemical">part proximal to the embryo (Fig. 1). In 2013, a pot experiment was conducted using plastic pots, 30 cm in height and 34 cm in diameter. Each pot was filled with 15 kg clay soil, containing 0.83 g kg−1 total N, 10.72 mg kg−1 available P and 69.15 mg kg−1 exchangeable K. For one pot, five seedlings of DY1102 were hand transplanted. The basal fertilization before transplanting used 1.0 g N, 1.2 g ass="Chemical">pan class="Chemical">P2O5 and 0.9 g K2O per pot. Topdressing at the panicle initiation stage applied 1.0 g N, 0.6 g P2O5 and 0.9 g K2O per pot.

About 300 panicles with similar maturity were harvested and naturally dried. Grains from the middle primary racpan class="Chemical">his of the ass="Chemical">panicle were sampled and then dehusked as brown ass="Chemical">pan class="Species">rice. As shown in Fig. 1, two types of notched‐belly grains –translucent grains (T) and chalky grains (C) – were separated. Subsequently, the embryos of these grains were removed. Along the notched line, the remaining grains were cut into upper parts (T1 and C1) and bottom parts (T2 and C2) using an anatomical knife.

Endosperm microstructure

Endosperm microstructure was observed by scanning electron microscopy (SEM). Completely dried, T1, T2, C1 and C2 were transversely cut with a razor blade, producing a thin and clean ass="Disease">fracture. Tass="Chemical">pan class="Chemical">his fracture was fixed on SEM stubs and sputter‐coated with gold under vacuum. The fixed specimens were observed under a scanning electron microscope (S‐3000N, Hitachi, Japan) with an accelerating voltage of 7 kV.

Chemical analysis

Starch, protein, and phytic acid

Total ass="Chemical">starch concentration was measured by Ewers' polarimetric method (International Standards: ISO 10520, 1997). Briefly, 2.50 ± 0.05 g flour sample was weighed and placed in a 100 mL flask. To tass="Chemical">pan class="Chemical">his flask was added 25 mL HCl solution (0.33 mol L−1) to wet and dissolve the sample, then another 25 mL HCl solution was added to accelerate the decomposition of the sample. The sample was then boiled in a water bath at 100 ° for 20 min and cooled to room temperature in running water. The sample was transferred to a 100 mL volumetric flask, and 5 mL zinc sulfate heptahydrate solution (30%, m/v) and 5 mL potassium ferrocyanide solution (15%, m/v) were added in order, and finally adjusted to 100 mL volume. Next, 25 mL of the suspension was filtrated using filter paper, and polarimetry measured with a polarimeter. Amylose was measured according the method of Perez and Juliano.14 Amylopectin was calculated by subtraction of amylose from total starch. Fine structure of amylopectin was determined using high‐performance size‐exclusion chromatography with pulsed amperometric detection (HPAEC‐PAD), as reported by Lin et al. 8 Briefly, amylopectin was obtained by the alkaline‐steeping method, and then debranched by isoamylase and separated by HPAEC‐PAD. The degree of polymerization (DP) of the linear fractions in debranched amylopectin was calculated as their molecular weight divided by 162.

Protein fractions including albumin, globulin, prolamin and glutelin were extracted and measured by the methods of Ning et al. 15 Total protein concentration was calculated as the sum of the four protein fractions. ass="Chemical">Phytic acid (PA) concentration was determined by the Fe precipitation method as reported by Ning et al. 15 The P concentration in ass="Chemical">pan class="Chemical">phytic acid (PA‐P) was calculated by multiplying phytic acid by 0.2815. All measurements were replicated in triplicate.

Amino acid

Amino acid anapan class="Chemical">lysis was performed using a high‐speed amino acid analyzer (model L‐8900; Hitachi, Jaass="Chemical">pan). Samples of ∼100 mg were dissolved in 10 mL of 6 mol L−1 ass="Chemical">pan class="Chemical">HCl at 110 °C for 24 h. The samples were adjusted to a final volume of 100 mL with purified water after hydrolysis. Next, 1 mL of each sample was vacuum dried and redissolved in 1 mL of 0.2 M HCl. Then, 20 μL samples were injected into the analyzer and data were acquired using EZChrom Elite for Hitachi AAA Operation software. Seventeen amino acids were measured by this method, with sample preparation by acidic hydrolysis.15 Each sample was measured in duplicate.

Minerals

Minerals were measured by inductively coupled plasma–optical emission spectroscopy (ICP‐OES; Optima 8000DV, PekinElmer, Waltham, MA, USA), with three replications. Samples (∼0.50 g) were wet digested with 10 mL HNO3–pan class="Chemical">HClO4 mixed acid (3:1, guaranteed reagent). ass="Chemical">pan class="Chemical">Argon was used as the make‐up gas, and ICP‐OES parameters were employed as follows: nebulizer flow rate, 0.80 L min−1; radio frequency power, 1300 W; sample flow rate, 1.50 mL min−1; flush time, 15 s; delay time, 30 s; read time, 10 s; wash time, 60 s. The ion standard solutions (1000 ppm, Merck, Germany) were used to obtain the calibration curve for 17 mineral elements, with correlation coefficients more than 0.999.

Comparison system

Using DY1102, a particular comparison system was used to precisely evaluate chalkiness trait as described in our previous report.8 As shown in Fig. 1, notched‐belly grain without white‐belly (T) was used as the control. The effect of embryo was quantified by comparison between the bottom part (T2) and upper part (T1). Comparison of the bottom part (C2) and upper part (C1) of the grain with white‐belly indicated the compound effect of embryo and chalkiness. Therefore, the effect of chalkiness can be evaluated through further comparison between C2/1 and T2/1 (C2/1/T2/1), by which the embryo effect was eliminated.

Statistical analysis

The data presented are averages of triplicate observations, except that amino acids values are means of duplicate measurements. The values of chemical components of chalky grains (Cg) and translucent grains (pan class="Chemical">Tg) are weighted averages according to the dry matter weights of the upper ass="Chemical">parts (T1 or C1) and bottom ass="Chemical">parts (T2 or C2), as shown in Fig. 1. Statistical anaass="Chemical">pan class="Chemical">lysis was performed with SPSS (v. 19.0, IBM) statistical software. Multiple comparisons were operated by Duncan's multiple range test (P < 0.05).

RESULTS

Morphology of the chalky and translucent parts

SEM images show contrasting differences in the microstructure between the translucent parts and chalky parts. For translucent parts from both the upper and bottom parts of translucent grains (Fig. 2A, B) and from the upper translucent part of the white‐belly grains (Fig. 2C), compound ass="Chemical">starch granules are tightly ass="Chemical">packed, with no air sass="Chemical">paces within or between them. Single ass="Chemical">pan class="Chemical">starch granules consisting of the compound granules are polygonal in shape, as shown in Fig. 2A, B, C. By contrast, the typical characteristics of chalky parts are observed, with the starch granules loosely packed in the opaque endosperm cells (Fig. 2 C, D, E). In comparison with the compound granules, the single starch granules are spherical but smaller in size (Fig. 2D, E), and are observed only in the opaque part, indicating the perturbation of development of starch granules. In addition, micro‐pores are detected on the surface of the starch granules (Fig. 2F), which was viewed as the evidence of α‐amylase attack.16

Figure 2

SEM images of endosperm transverse sections from the upper and bottom half parts of grains with white‐belly (C1 and C2) and without white‐belly (T1 and T2). (A, B, C) Translucent tissues of T1, T2, and C1, respectively. (D, E F) Chalky tissues of bottom half part of C2. Large and small white arrows indicate compound starch granules and single starch granules, respectively. Large black arrows indicate traces of smaller protein bodies (PBI) removed during sample preparation. Small black arrows indicate that micro‐pores occur on the surface of the compound starch granules, which show evidence of α‐amylase attack.

SEM images of endosperm transverse sections from the upper and bottom half parts of grains with white‐belly (C1 and C2) and without white‐belly (T1 and T2). (A, B, C) Translucent tissues of T1, T2, and C1, respectively. (D, E F) Chalky tissues of bottom half part of C2. Lass="Chemical">arge and small white arrows indicate compound ass="Chemical">pan class="Chemical">starch granules and single starch granules, respectively. Large black arrows indicate traces of smaller protein bodies (PBI) removed during sample preparation. Small black arrows indicate that micro‐pores occur on the surface of the compound starch granules, which show evidence of α‐amylase attack.

Fractions of starch and protein

Fractions of ass="Chemical">starch and protein in the chalky grains (Cg) and translucent grains (ass="Chemical">pan class="Chemical">Tg) were calculated by weighted averages of C1 and C2, and T1 and T2, respectively. As shown in Table 1, amylose and globulin were significantly higher, whereas amylopectin and albumin were lower, in chalky grains than translucent grains. No significant differences were observed for total starch and total protein between the two kinds of grains. In addition, chalky grains contained fewer short chains but more long chains of amylopectin in comparison with translucent grains (Fig. 3). This finding is contrary to that of Patindol and Wang.10

Table 1

Concentrations of starch and protein fractions in the upper and bottom half parts of translucent grains (T) and chalky grains (C) of the notched‐belly mutant with white‐belly (DY1102)

Nutrients	T1	T2	C1	C2	Tg	Cg	T2/1	C2/1	C2/1/T2/1
Amylose (%)	16.0c	16.5b	17.1a	17.3a	16.3	17.2*↑	1.03↑	1.01	0.98
Amylopectin (%)	60.0a	58.7a	55.7b	60.6a	59.3	58.5*↓	0.98	1.09↑	1.11↑
Amylose/Amylopectin	0.27	0.28	0.31	0.29	0.27	0.29	1.04	0.93↓	0.88↓
Total starch (%)	75.9ab	75.2b	72.8c	77.9a	75.6	75.7	0.99	1.07↑	1.08↑
Albumin (%)	0.52a	0.53a	0.53a	0.51a	0.53	0.52*↓	1.03	0.96	0.93↓
Globulin (%)	0.44a	0.42b	0.43ab	0.44a	0.43	0.44*↑	0.95↓	1.00	1.05↑
Prolamin (%)	0.69b	0.69b	0.73a	0.67b	0.69	0.69	1.00	0.92↓	0.92↓
Glutelin (%)	5.66a	5.26b	5.76a	5.15b	5.45	5.41	0.93↓	0.90↓	0.96
Total protein (%)	7.30a	6.90b	7.45a	6.76b	7.09	7.05	0.95↓	0.91↓	0.96

C1 and C2 are the upper and bottom half part of the translucent grain, respectively. T1 and T2 are the upper and bottom half part of the chalky grain, respectively. Tg and Cg are expressed as weighted average according the dry matter weight of the upper part ( T1 or C1) and the bottom part (T2 or C2).

T2/1 and C2/1 are comparisons between T2 and T1, and C2 and C1, which show the effect of embryo, and the combination effect of embryo and chalkiness, respectively. C2/1/T2/1 is a comparison between T2/1 and C2/1, showing the effect of chalkiness.

Data in a row with different lower‐case letters are significantly different (P < 0.05). Data in a row with an asterisk are significantly different between Tg and Cg (P < 0.05). Bold entries indicate upregulation (↑), downregulation (↓) or no marked differences. For comparisons of T2/1 and C2/1, marked difference is according to Duncan's multiple range test (P < 0.05). For comparisons of C2/1/T2/1, marked difference is defined as those ≤ 0.95 (downregulation) or ≥ 1.05 (upregulation).

Figure 3

Differences in chain length distribution of amylopectin between chalky grains (Cg) and translucent grains (Tg) as revealed by HPAEC‐PAD.

Concentrations of pan class="Chemical">starch and protein fractions in the upper and bottom half ass="Chemical">parts of translucent grains (T) and chalky grains (C) of the notched‐belly mutant with white‐belly (DY1102)

C1 and C2 are the upper and bottom half part of the translucent grain, respectively. T1 and T2 are the upper and bottom half part of the chalky grain, respectively. pan class="Chemical">Tg and Cg are expressed as weighted average according the dry matter weight of the upper ass="Chemical">part ( T1 or C1) and the bottom ass="Chemical">part (T2 or C2).

Data in a row with different lower‐case letters are significantly different (P < 0.05). Data in a row with an asterisk are significantly different between pan class="Chemical">Tg and Cg (P < 0.05). Bold entries indicate upregulation (↑), downregulation (↓) or no marked differences. For comass="Chemical">parisons of T2/1 and C2/1, marked difference is according to Duncan's multiple range test (P < 0.05). For comass="Chemical">parisons of C2/1/T2/1, marked difference is defined as those ≤ 0.95 (downregulation) or ≥ 1.05 (upregulation).

Differences in chain length distribution of amylopn>ectin between chalky grains (Cg) and translucent grains (pan class="Chemical">Tg) as revealed by HPAEC‐PAD.

On the basis of calculation of T2/T1, an embryo effect was detected concerning these chemical components (Table 1). The embryo lowered the accumulation of globulin, glutelin and total protein, but with higher accumulation of pan class="Chemical">amylose.

By comparing C2/C1 and T2/T1, chalkiness effect was qualified (Table 1). Occurrence of chalkiness was associated with higher accumulation of amylopectin, total pan class="Chemical">starch and globulin, but with lower concentrations of albumin and prolamin, but appeared to be not correlated with ass="Chemical">pan class="Chemical">amylose. In addition, glutelin and total protein tended to be lowered by chalkiness.

Amino acid composition

Comparison between chalky grains and translucent grains showed that the former were significantly higher in nine of the 17 amino acids measured: ass="Chemical">alanine, ass="Chemical">pan class="Chemical">arginine, histidine, isoleucine, leucine, phenylalanine, proline, tyrosine and valine (Table 2). Nevertheless, concentration of total amino acids was not significantly different between the two types of grains, consistent with the trend for total protein concentration (Table 1).

Table 2

Amino acid concentrations of the upper and bottom half parts of translucent grains and chalky grains (mg g−1)

Amino acids	T1	T2	C1	C2	Tg	Cg	T2/1	C2/1	C2/1/T2/1
Alanine	4.38b	4.25bc	4.69a	4.16c	4.31	4.39*↑	0.97	0.89↓	0.92↓
Arginine	7.17b	6.07c	7.80a	6.95b	6.59	7.31*↑	0.85↓	0.89↓	1.05↑
Aspartic acid†	7.37ab	7.30ab	7.70a	6.99b	7.33	7.29	0.99	0.91↓	0.92↓
Cysteine	0.85b	0.84b	0.96a	0.85b	0.84	0.90	0.99	0.88↓	0.89↓
Glutamic acid†	15.93b	15.49bc	16.95a	14.94c	15.70	15.79	0.97	0.88↓	0.91↓
Glycine	3.46ab	3.43ab	3.65a	3.25b	3.44	3.42	0.99	0.89↓	0.90↓
Histidine	3.73b	3.08c	3.93a	3.63b	3.39	3.76*↑	0.83↓	0.92↓	1.12↑
Isoleucine	2.89b	2.87b	3.19a	2.87b	2.88	3.01*↑	0.99	0.90↓	0.91↓
Leucine	6.44b	6.26c	6.97a	6.20c	6.34	6.53*↑	0.97↓	0.89↓	0.92↓
Lysine	4.69a	4.50b	4.72a	4.43b	4.59	4.55	0.96↓	0.94↓	0.98
Methionine	1.90b	1.80bc	1.75a	1.65c	1.85	1.69	0.91	0.92↓	1.01
Phenylalanine	5.13b	4.92c	5.41a	4.85c	5.02	5.09*↑	0.96↓	0.90↓	0.93↓
Proline	3.00b	2.97b	3.26a	2.92b	2.98	3.06*↑	0.99	0.90↓	0.91↓
Serine	4.43b	4.38b	4.70a	4.24b	4.40	4.44	0.99	0.90↓	0.91↓
Threonine	3.11b	3.08b	3.26a	2.97b	3.09	3.09	0.99	0.91↓	0.92↓
Tyrosine	3.45b	3.24c	3.96a	3.34bc	3.34	3.60*↑	0.94↓	0.84↓	0.90↓
Valine	4.11b	4.07bc	4.42a	4.03c	4.09	4.20*↑	0.99	0.91↓	0.92↓
Total	82.04	78.55	87.32	78.27	80.19	82.11	0.96	0.90↓	0.94↓

Data of aspartate are sum of aspartate and asparagine, and those of glutamate are sum of glutamate and glutamine. See also note to Table 1.

Amino acid concentrations of the upper and bottom half pan class="Chemical">parts of translucent grains and chalky grains (pan class="Chemical">mg g−1)

Data of ass="Chemical">aspartate are sum of ass="Chemical">pan class="Chemical">aspartate and asparagine, and those of glutamate are sum of glutamate and glutamine. See also note to Table 1.

The embryo showed a negative effect on the concentrations of amino acids. The majority of the 17 amino acids were downregulated by the embryo, although there were only six amino acids, i.e. ass="Chemical">arginine, ass="Chemical">pan class="Chemical">histidine, leucine, lysine, phenylalanine and tyrosine, that demonstrated a marked difference. Similarly, chalkiness had a suppressing effect on the majority of the 17 amino acids. However, two of amino acids – arginine and histidine – were upregulated.

Concentrations of minerals

In general, chalky grains had lower concentrations of most of the minerals examined, except for Cr, Cu, K, Mg, Sr, Ni and V (Table 3). Similarly, the embryo showed an obvious negative effect on all the minerals except for Ca, Cu, P and Sr (Table 3). Chalkiness had a positive effect on the concentrations of As, Ba, Cd, Cr, Mn, Na, Sr and V, but negatively correlated with the concentrations of B, Ca, Cu, Fe and Ni. In addition, no significant differences were detected for pan class="Chemical">phytic acid‐P, P, K, Mg and Zn.

Table 3

Concentrations of minerals from the upper and bottom half parts of translucent grains and chalky grains (µg g−1)

Minerals	T1	T2	C1	C2	Tg	Cg	T2/1	C2/1	C2/1/T2/1
As	2.22a	0.71b	0.17c	0.53bc	1.42	0.38*↓	0.32↓	3.12	9.76↑
B	5.40a	3.48b	1.99bc	0.97c	4.38	1.41*↓	0.64↓	0.49	0.76↓
Ba	0.34a	0.20b	0.28ab	0.17b	0.27	0.22*↓	0.57↓	0.61	1.06↑
Ca	118.04a	117.54a	117.61a	110.73a	117.78	113.65*↓	1.00	0.94	0.95↓
Cd	0.19a	0.11b	0.09b	0.08b	0.15	0.08*↓	0.60↓	0.94	1.57↑
Cr	1.77a	0.49c	0.64bc	1.32ab	1.09	1.03	0.28↓	2.07	7.44↑
Cu	4.50ab	3.90ab	5.83a	3.57b	4.18	4.53	0.87	0.61↓	0.71↓
Fe	18.6a	15.99b	18.51a	12.94c	17.22	15.31*↓	0.86↓	0.70↓	0.81↓
K	1173.3a	1078.0b	1142.6ab	1075.7b	1122.9	1104.14	0.92↓	0.94	1.02
Mg	951.99a	764.11b	993.13a	771.34b	852.53	865.56	0.80↓	0.78↓	0.97
Mn	54.02a	46.90b	45.50b	41.77b	50.25	43.35*↓	0.87↓	0.92	1.06↑
Na	113.79a	67.62b	65.84b	58.85b	89.35	61.82*↓	0.59↓	0.89	1.50↑
Ni	1.03b	0.43c	1.44a	0.38c	0.71	0.83*↑	0.42↓	0.27↓	0.64↓
P	2696.8a	2335.0ab	2449.4ab	2194.3b	2505.2	2302.7*↓	0.87	0.90	1.03
PA‐P	1860.5a	1528.4c	1750.6b	1441.8d	1684.7	1573.0*↓	0.82↓	0.82↓	1.00
Sr	0.32a	0.27a	0.31a	0.28a	0.29	0.29	0.85	0.90	1.06↑
V	0.39a	0.03b	0.22ab	0.16b	0.2	0.19	0.08↓	0.72	8.69↑
Zn	20.36a	15.70bc	18.01ab	14.05c	17.90	15.73*↓	0.77↓	0.78↓	1.01

PA‐P, P content in phytic acid (PA), calculated by content of PA multiplied by 0.2815. See also note to Table 1.

Concentrations of minerals from the upper and bottom half pan class="Chemical">parts of translucent grains and chalky grains (µg g−1)

PA‐P, P content in pan class="Chemical">phytic acid (PA), calculated by content of PA multiplied by 0.2815. See also note to Table 1.

DISCUSSION

Significance of the notched‐belly mutant

Substantial advances in elucidating the mechanisms governing chalkiness formation have been made, as evidenced by the cloning of controlling genes like Chalk5,2 and the comprehensive anapan class="Chemical">lysis of related enzymes and their regulatory ass="Chemical">pathways by the tools of proteomics8 and transcriptomics.9 However, the molecular and biochemical mechanisms underlying chalkiness formation are still imperfectly understood, owing to the complex interaction between genotype and environment.

In previous work, we identified a notched‐belly mutant (DY1102) with a high ratio of white‐belly grains (83.4%). Interestingly, white‐belly only occurs in the bottom part, whereas the upper part is translucent. In addition, about 15% of the notched‐belly grains are translucent, which can be used to quantify the embryo effect. Using tpan class="Chemical">his mutant, we developed a novel comass="Chemical">parison system that can clarify the effect of chalkiness. The comass="Chemical">parison is performed within the same grain, thus proving a nearly identical genetic background. Using tass="Chemical">pan class="Chemical">his comparison system, we detected marked differences in chemical components between chalky parts and translucent parts, which are of value for exploring the physiological foundation of grain chalkiness.

Notably, some findings of tass="Chemical">his study are contrary to previous studies, which may be associated with the comn>an class="Chemical">ass="Chemical">parison system and the material used. Take ass="Chemical">pan class="Chemical">amylose as an example. Many studies revealed that reduced amylose content is a typical feature for chalky grains ripening under high temperature, which is attributed to the suppression of GBSS.17 Conversely, the chalky grains as a whole had a higher amylose content than did the translucent grains in this study. Further, using the comparison between T1 versus T2 and C1 versus C2, no significant effect of chalkiness on amylose content was detected when the embryo effect was eliminated. This case stresses the importance of selection of the comparison system when exploring the mechanisms underlying grain chalkiness.

Role of the embryo in formation of grain chalkiness

Within the endosperm, formation of chalky tissues is considered as related to ass="Disease">insufficient vascular supply of metabolites such as ass="Chemical">pan class="Chemical">sucrose,11 perturbation of growing of starch granules in amyloplast18 or interruption in starch biosynthesis.10, 19 Similarly, SEM images of this study show the incomplete development of starch granules in the chalky parts, and analysis of amylopectin structure demonstrates that impaired synthesis of long chains occurs in the chalky parts. In addition to starch accumulation, there is growing awareness of the role of starch hydrolysis in the occurrence of grain chalkiness. Tsutsui et al. 20 and Yamakawa et al. 17 revealed that mRNA expression of Amy1A, Amy1C, Amy3D and Amy3E, as well as α‐amylase activity, increased under high‐temperature stress, suggesting a relation of starch degradation to grain chalkiness formation. In this study, SEM images show that many micro‐pores occur on the surface of the compound granule, which indicates evidence of α‐amylase attack. However, little is known about the reasons for incomplete accumulation and hydrolysis of starch, in particular from the perspective of interaction between the endosperm and the embryo.

The ass="Species">rice grain is mainly composed of the embryo, endosperm and pericarp. The embryo and endosperm develop synchronously in terms of cell division and differentiation, and complete morphological differentiation and development about 10 days after fertilization.21 The transportation of nutrients to the embryo is performed through the endosperm ass="Chemical">part. At maturity, two to three layers of endosperm cells behind the scutellum contain no reserve substances and show a state of degeneration, because of the deprivation of their reserves by the embryo.22 The ass="Chemical">pan class="Chemical">phenomenon was also observed in this study, as shown in Fig. 1. Even for the translucent grains, a thin layer of endosperm cells near the scutellum appears opaque due to the disorder of storage substance accumulation, indicating a strong effect of the embryo on endosperm development.

Using translucent grains of the notched‐belly mutant, the effect of embryo on the composition of grains was evaluated by comparison of the upper and bottom half parts. Substantial influence of the embryo was detected. Generally, the embryo lowered the concentrations of the chemical components examined, especially minerals. Except for Ca, Cu, P and Sr, concentrations of all the 17 elements were reduced in the bottom part as compared with the upper part. The strong effect of the embryo suggests that ass="Chemical">starch, protein and minerals in the endosperm are prone to be mobilized by the embryo in order to meet the requirements of embryo development, which is crucial for the ass="Chemical">pan class="Species">rice plant to cope with the growing environments. Therefore, we propose that the embryo is involved in the formation of white‐belly for this notched‐belly mutant, and the role of the embryo should be thoroughly examined in the study of rice chalkiness.

Chemical composition in relation to chalky grain formation

The chemical composition of grains is of biological significance for germination and economic value for producing nutritious ass="Species">rice. It may also function as an integrative index of the cumulative effect of genes and environment on the physiological process involved in accumulation of ass="Chemical">pan class="Chemical">starch, proteins and minerals. Thus, by comparison of the chemical composition between chalky and translucent grains, we can gain comprehensive information concerning the physiology of chalky grain formation. In this study, we used a novel comparison method and found that total protein, as well as amino acids, tended to be lowered by the occurrence of chalkiness, while amylopectin and total starch were increased. The inverse relationship between starch and protein with respect to chalkiness formation needs to be further addressed.

In addition, for the 17 elements measured, higher concentrations of As, Ba, Cd, Cr, Mn, Na, Sr and V, and lower concentrations of B, Ca, Cu, Fe and Ni, were related to chalkiness. Notably, both P and PA‐P were unaffected by chalkiness, and K, Mg and Zn, the colocalized minerals with ass="Chemical">phytic acid,23 were also not clearly influenced by chalkiness. Considering that P was measured by ICP‐AES while PA‐P was measured by the Fe‐precipitation method, these two ass="Chemical">parallel methods together indicate that ass="Chemical">pan class="Chemical">phytic acid, the majority of P stored in rice grains, should not be involved in the formation of chalkiness. Since mineral nutrition plays a critical role in translocation of assimilates from source organs and the synthesis of starch and proteins in grains, the information obtained by this study is of value for future study on the relation between minerals and chalkiness.

Unlike ass="Species">wheat, the grains of which are milled into flour and then processed into different kinds of foods such as bread and noodles, ass="Chemical">pan class="Species">rice is consumed in the form of whole grains. Chemical properties differ between rice flour and grains, as is manifested in the current study. Chemical analysis showed higher starch concentration, but SEM revealed incomplete compound starch granules in the chalky tissues of grains. Disagreement may be associated with the loose structure of the chalky part. As observed by SEM, there are obvious inter‐granule spaces in the chalky part, clearly indicating that the concentration of starch is lower per square millimeter in the chalky part than in the translucent part. On the other hand, in chemical analysis the intact structure of both the chalky part and the translucent part were destroyed and reduced/condensed to powders, thus masking the dilution effect of the loose structure, as was reported by Lisle et al. 11 This could explain why the chalky part (C2) of the white‐belly grains contained higher starch compared with its counterpart, the translucent part (C1), by chemical analysis (Table 1). Therefore, the dilution effect of the loose structure should be considered when explaining the data of chemical analysis of chalky grains. Furthermore, the limitation of flour‐based analysis should be addressed in future studies on rice quality.

CONCLUSIONS

Using a novel comparison system, we compared the chalky part with the translucent part of a notched‐belly mutant with white‐belly, and revealed notable differences in the composition between chalky and translucent parts. Importantly, the effects of the embryo and endosperm were qualified. The embryo showed substantial influence on the composition of grains, especially lowering the accumulation of minerals. By excluding the embryo effect, a significant influence of chalkiness was revealed. The occurrence of chalkiness was associated with elevated concentrations of amylopectin and total ass="Chemical">starch, while it appeared to be negatively correlated with the concentration of the majority of the 17 amino acids. In addition, chalkiness had a synergistic effect on As, Ba, Cd, Cr, Cu, Mn, Na, Sr and V, but an antagonistic effect on B, Ca, Fe and Ni. Notably, ass="Chemical">pan class="Chemical">phytic acid, K, Mg and Zn were not affected by chalkiness, indicating that metabolism of phytic acid should not be involved in the formation of chalkiness. This study provides a useful clue concerning the relation between chalky part formation and the accumulation of starch, proteins and minerals. Further, the role of the embryo in the formation of grain chalkiness was highlighted, which deserves further investigation.

However, tpan class="Chemical">his study shows the static state of the mature grain and its chemical composition, uncovering some clues as to the formation of grain chalkiness. More work concerning the dynamic nature of the chalky tissue is needed, in ass="Chemical">particular the metabolic profiles and genes responsible for C and N metabolism, in order to gain an integrated understanding of the mechanism underlying chalkiness formation.