Allelic effects on starch structure and properties of six starch biosynthetic genes in a rice recombinant inbred line population

Grouping of recombinant inbred lines

Six allele groups were selected from recombinant inbred lines corresponding to the
6 major starch synthetic enzyme genes, and parental lines (IR64 and Nipponbare) differing
in their alleles (i vs j) for each of the 6 genes. Thus we segregated the RILs into six allele groups according
to the allele combinations (Figure 1) to analyse the effects of i vs j alleles of different starch biosynthetic genes on starch structure and properties.
The allele types of these six genes of each RIL are described in Table 1, and are designated as follows: line 3-14-12 and 3-14-13 as Wxi and Wxj; line 3-5-15-11 and 3-6-20 as SSIi and SSIj; line 3-6-9 and 3-6-1 as SSIIai and SSIIaj; line 3-5-1-2 and 3-5-1-8 as SBEIi and SBEIj; line 3-6-20 and 3-6-9 as SBEIIai and SBEIIaj; and line 3-5-2-2 and 3-5-2-1 as SBEIIbi and SBEIIbj. Meanwhile, we also separated RILs in Wxi and Wxj allele groups in some of the following analyses, because Wx alleles were well known in determining various starch characteristics.

Figure 1. Generation of RILs for each of two alleles of six starch synthetic genes. Six genes for genotyping of alleles are Wx, SSI, SSIIa, SBEI, SBEIIa, SBEIIb. Parental lines are IR64 and Nipponbare. The arrows indicate descendants of each
line. The numbers with slash line indicate the name of each line. The numbers following
‘F’ imply the generation on the left of the figure. The italic letters followed by
‘i’ or ‘j’ (indicating indica and japonica, respectively) indicate alleles of the RILs.

Table 1. Genotypes of RILs of six starch biosynthetic enzyme groups

Amylose content

Although the Wxi allele played a major role in the determination of increased AC, other alleles also
showed impacts on AC. The range of AC of selected RILs was 11?~?35% approximately
of rice grain starch (Figure 2). Six allele groups can be classified into two AC groups depending on the Wx allele. The Wxi allele group contained 17.7?~?35.3% amylose in starch (including Wxi allele and SSI, SSIIa and SBEIIa groups), whereas Wxj allele group contained 11.1?~?18.4% (including Wxj allele and SBEI and SBEIIb groups), with parental lines IR64 (21.2%) and Nipponbare (11.8%) fitting in their
respective Wx groups. Between the Wxi allele group (containing average 22.0% amylose) and the Wxj allele group (containing average 13.1% amylose), the former contained ~9% significantly
higher AC than the latter. Within Wxi allele group, starch samples from Wxi allele lines contained significantly higher AC (35.3%) than all other allele groups,
and starch from SSIi (22.8%) and SSIIaj (20.9%) alleles contained higher AC than other four alleles (SSIj, SSIIai, SBEIIai and SBEIIaj) (Figure 2). Within Wxj allele group, SBEIi allele lines (18.4%) contained significantly higher AC than those from other four
allele groups, among which AC was not significantly different.

Figure 2. Comparison of amylose content among 12 alleles of six starch synthetic genes using
the SEC method. Starches from five plants of each allele were isolated and analysed seperately. Two
replicates were set up for each sample. The columns indicate starch AC of RIL grains.
Black columns indicate those allelles containing Wxi allele and grey columns indicate those allelles containing Wxj allele. The error bars show the standard error of the mean. The identity of each
column is indicated underneath. Columns with different letters are significantly different
at p??0.05.

Chain length distribution (CLD) of debranched starch

In this study, linear glucan chains of DP???24 comprised up to 80% mole of molecules
in amylopectin of rice grains (Table 2). SSIIai played a major role, while SBEIIbi played a minor role in accumulating intermediate chains (DP12?~?22). The differences
in % normalised distribution were obtained by subtracting the CLD of japonica allele from indica allele in each gene group, respectively (Figure 3). The profound difference in CLD was obtained in SSIIa allele group (Figure 3C). SSIIai allele starch contained fewer short chains at DP6?~?DP11, more intermediate chains
at DP12?~?22 and long chains at DP30?~?45 compared with that of SSIIaj allele starch. Limited variations were found in the other five gene groups, however,
SBEIIbi allele showed up to 0.5% normalised distribution of fewer chains at DP10?~?13 than
SBEIIbj allele (Figure 3F).

Table 2. Chain length fractions of RILs of six starch biosynthetic enzyme allele groups

Figure 3. Comparison of chain length distribution of debranched starch of six allele groups
for six starch synthetic genes. Starches from five plants of each allele were used for the analysis. Two replicates
were set up for each sample. For each allele group, values of chain lengths for japonica allele lines were subtracted from values of chain length for the indica allele lines. A: Wx, B: SSI, C: SSIIa, D: SBEI, E: SBEIIa, F: SBEIIb. Each bar corresponds to the difference of a chain length in mole percentage. The
error bars are the standard errors.

Based on Nakamura’s study (Nakamura et al. 2002]), indica and japonica type amylopectin from the majority of cultivated Asian rice strains can be distinguished
by the ratio of ?DP???10 to ?DP???24 fractions (R_CL10/24) of CLD. Our present study showed the R_CL10/24 value was 0.16 for IR64 and 0.29 for Nipponbare (Table 2), and most of the selected RILs were smaller than 0.14, suggesting that they were
indica type. With the only exception of SSIIaj allele, its R_CL10/24 value was slightly over 0.20. The highest value of SSIIaj resulted from a high percentage of short chain fraction (?DP???10, significantly
higher than that of SSIIai allele, and 43?~?67% higher than the other RILs) and low percentage of short and
intermediate chain fraction (?DP???24, statistically significantly lower than that
of SSIIai allele) in the starch. In terms of the longer chain fraction (?DP??24), SSIIaj was the highest but not significantly different to a few other alleles in the group
(Table 2). Besides SSIIaj, SBEIIbj allele was also detected with more short chains in starch than the counterpart-allele
SBEIIbi.

Starch paste viscosity

By analysing RVA characteristics pasting properties of starch were determined as reported
in previous studies (Sasaki et al. 2000]; Han and Hamaker 2001]; Chen et al. 2003]). The RVA characteristics involved in this study were peak viscosity (PV), trough
(Tr), breakdown (BD), final viscosity (FV), setback (SB) and peak time (PT). The values
of these characteristics were statistically analysed and shown in Table 3. The RVA result revealed that i and j alleles from each of six allele groups contributed to parameters of starch paste
viscosity properties differentially.

Table 3. Viscosity properties of RIL wholemeal flours analysed using RVA

IR64 and Nipponbare showed significant differences in all parameters except for FV
(Table 3). RILs in Wx, SSIIa, SBEIIa and SBEIIb allele groups had significantly distinct PV between i and j alleles, respectively. PV values of Wxj and SSIIai alleles were ~60 RVU higher than Wxi and SSIIaj alleles, whereas the increase in SBEIIaj and SBEIIbi alleles was smaller (31?~?35 RVU) comparing to SBEIIai and SBEIIbj alleles. Tr was the least diverse RVA parameter in this analysis, showing only about
30 and 10 RVU significant increase in SSIj and SBEIIbi alleles comparing to their counterpart-alleles, respectively. Significant differences
in BD were obtained in Wx, SSI, SSIIa and SBEIIa allele groups, with the greatest (~70 RVU) between SSIIai and SSIIaj alleles. Compared to their counterparts, the values of BD were increased by ~40 RVU
in Wxj allele, and ~30 RVU in SSIi and SBEIIaj alleles. In terms of FV, significant variations were found in Wx, SSI and SBEIIb allele groups, with the highest (~40 RVU) between SSIi and SSIj alleles, ~30 and 15 RVU in Wx and SBEIIb allele groups, respectively. The significant variations of SB were 42 RVU in Wx group and 28 RVU in SSIIa group. Less than 10 RVU increase of SB in both SSIj and SBEIj alleles were also significantly different. Significant differences in PT were determined
between SSI alleles (0.6 min) and Wx group (0.2 min). The PT values were mainly clustered into the two Wx alleles groups, while IR64 was the lowest. Notably, the SSIIaj allele starch exhibited remarked differences in paste viscosities from all the others
including parental controls.

Starch thermal properties

Consistent with previous studies, two primary peaks were observed in DSC curves for
rice wholemeal samples: the first one was the initial gelatinization peak at 65?~?85°C
of amylopectin; the second one at around 105°C was the amylose-lipid complex dissociation
peak (Biliaderis et al. 1985]). Onset, peak and end temperatures of the first peak were named as ¹To, ¹Tp and ¹Te, and those of the second peak as ²To, ²Tp, ²Te. ¹?H was named for the amylopectin gelatinization enthalpy, and ²?H for the dissociation enthalpy of amylose-lipid complexes. As shown in Table 4, the DSC result indicated that SSIIaj allele had a major effect, while Wxi allele played a minor role in decreasing ¹Tp. However, SSIj, SSIIaj and Wxi alleles increased ²?H.

Table 4. Thermal characteristics of RIL flours determined using DSC

Regarding amylopectin gelatinization temperatures, IR64 had significantly higher values
in ¹To, ¹Tp and ¹Te than Nipponbare, whereas the RILs exhibited divergent values. Wxj allele group observed ~3°C higher ¹To than those from Wxi allele group. There was barely any variation in ¹To among different alleles within Wxi allele group except SSIIaj allele having ~11°C decrease compared with SSIIai allele. In terms of ¹Tp, Wxj allele group overall was ~3°C higher than Wxi allele group. Within Wxi allele group, SSIIai was ~11°C higher than SSIIaj, while SSIj and SBEIIai were ~1 and 0.5°C higher than SSIi and SBEIIaj alleles, respectively. For ¹Te, Wxj allele group overall was ~2°C higher than Wxi allele group. There was ~9°C increase in SSIIai allele comparing to SSIIaj allele. In terms of ¹?H, Wxj allele group was ~0.4 J/g higher than Wxi allele group, except for SBEIi allele 0.3 J/g lower than the counterpart-allele.

Regarding dissociation temperatures of amylose-lipid complexes, IR64 wholemeal had
significantly higher values of ²To, ²Tp and ²Te than Nipponbare. In terms of ²To, the only significant variations were observed in SSIIaj and SBEIIai alleles which both had ~2°C increase compared with their counterpart-alleles. In
terms of ²?H, SSIIaj and SSIj alleles were up to 0.5 J/g and 0.24 J/g higher than SSIIai and SSIi alleles, respectively. Wxi allele group overall exhibited higher energy than Wxj allele group. Within Wxi group, SSIIaj and Wxi were the highestand lowest, whereas in Wxj allele group SBEIi allele had higher ²?H than the others. Interestingly, similar to the observation in RVA examination,
SSIIaj allele starch also showed remarked differences in thermal properties from others.

Protein analyses in starch granules of mature grains

The analysis of GBPs prepared from purified starch of mature rice grains showed that
four major protein bands with 60 kDa and above were detected in most of the RILs (Figure 4). In parental lines, the top bands at ~88 kDa were identified as SSIIa, ~83 kDa as
SBEIIb, ~75 kDa as SSI and ~60 kDa as GBSSI by immunoblotting using specific antibodies
(Figure 5).

Figure 4. Analysis of starch GBPs in mature rice grain starch of RILs from six allele graoups
of six starch synthetic genes by SDS-PAGE. Starches from five RIL lines of each allele were used. Section A: Wx allele group, B: SSI allele group, C: SSIIa allele group, D: SBEI allele group, E: SBEIIa allele group, F: SBEIIb allele group. The molecular sizes are labelled on the left of protein marker bands
in kDa. The identity of each protein band in the samples is indicated on the right
side of the pictures by an arrow head.

Figure 5. Immunodetection analysis of GBPs of purified starch from mature rice grains. A, parental lines; B and C, selected RIL lines. The names of alleles are labelled on top of each lane for Wxj (3-14-13), Wxi (3-14-12), SSIIaj (3-6-1) and SSIIai (3-6-9). The estimated molecular weight of protein bands are shown on the left. The
protein bands detected by various antibodies are indicated by arrows on the right.’Nip’
is the abbreviation for Nipponbare.

Comparing the two parental lines, IR64 contained a higher amount of GBSSI, SSI, SSIIa
and SBEIIb in the GBPs than Nipponbare, and SSIIa was barely detectable by immunodetection
in Nipponbare (Figure 5A). Consistent with parental lines, GBSSI abundance of Wxi allele was significantly higher than that of Wxj allele, while the abundance of SSI, SSIIa and SBEIIb in Wxi and Wxj alleles remained at same levels, respectively (Figures 4A, 5B). In SSIIa alleles, the amount of SSI, SSIIa and SBEIIb protein was significantly decreased
in SSIIaj allele (Figure 5C). Similarly, only faint bands of SSIIa were detected in SSIIaj RILs (Figure 4C), whereas GBSSI remained at the same level. No significant changes were observed
in abundance of the four proteins in SDS-PAGE gels of other groups.