Analysis on genetic background of materials and grouping of BILs and NILs
Polymorphism of eighteen SSRGs had been detected between two parents and they differed in six genes (AGPlar, SSI, SSIII-1, SSIV-2, SBE3 and PUL) (Fig. 1). Then, two BILs were constructed to study the effects of SSIV-2 and its interaction with AGPlar and PUL on rice ECQs. The one BIL with 184 plants differed only in AGPlar and SSIV-2 loci, another with 93 plants differed only in PUL and SSIV-2 loci under the background of CG173R, which came from the BC1F12 generation. In order to further validate the effects of SS IV-2, seeds of 9 plants from BIL 529 that only had polymorphism at SSIV-2 locus were identified and selected to develop the BC1F13, as a near isogenic line at the SSIV-2 locus with CG173R genetic background (NIL-SS IV-2) (Fig. 2). Various genotypes were obtained according to different alleles among BIL and NIL populations, and were designated as follows: to the AGPlar locus, as type I (same as GZ63S), type II (same as CG173R) and type III (heterozygous), respectively; to the SS IV-2 allele, as type G (came from GZ63S), type C (came from CG173R) and type H (heterozygous), respectively; to the PUL allele, as type 1 (same as GZ63S), type 2(same as CG173R) and type 3(heterozygous), respectively(Additional file 3:Fig. S1).
Fig. 1. Polymorphism gene loci of starch synthesis-related genes between two parents.
G and C indicate homozygous for GZ63S and CG173R, respectively
Fig. 2. Generation of BILs and NIL for starch synthesis-related genes.
The italic letters followed by ‘c’ denote possessing the same alleles as CG173R, by “s” mean the alleles heterozygous
|
Phenotypic variation analysis on physicochemical properties of BILs from BC1F12 and their parents
A t test was conducted to compare the difference of the two parents on physicochemical quality characters. The results clearly showed that the parents significantly differed on a majority of quality characters (Fig. 3), which were caused by their various starch biosynthesis genes. The difference of GT between GZ63S and CG173R was not obvious because of same SSII-3 allele for them. However, there was a significant difference on AAC (P = 0.02) although the parents had same Wx allele and belonged to low AAC (range 10-20%). This indicated that there were still other minor genes affecting AAC. Most of the ECQs and all the RVA parameters, such as GC, PKV, HPV and CPV, varied widely among the BC1F12. For example, the GC ranged from 36.00 mm to 87.00 mm, with an average of 57.03 mm, and with a variation coefficient of 15.82%. In contrast, the variations of GT and PeT were relatively narrow. GTs were very close to 71℃ for most of the plants and the difference of PeT between the maximum and minimum is only 0.73 min. These results showed that the main ECQs characters were qualified for following genetic analysis.
Fig. 3. Distribution of the quality characters in BILs from BC1F12.
GZ, Guangzhan 63S; CG, CG173R; a) Gel consistency (GC); b) Apparent amylose content (AAC); c) Gelatinization temperature (GT); d) Peak viscosity (PKV); e) Hot paste viscosity (HPV); f) Cool paste viscosity (CPV); g) Breakdown viscosity (BDV); h) Setback viscosity (SBV); i) Consistence viscosity (CSV); j) Peak time (PeT); k) Pasting temperature (PaT); l) Percent of retrogradation (R). The P values, generated by student’s tests, denote the differences between Guangzhan 63S and CG173R
|
Correlation analysis among physicochemical properties
In order to investigate the relationships among the tested physicochemical properties in BC1F12 population, a pairwise correlation analysis was carried out. The correlations between GC and all the RVA profile parameters, except BDV and SBV, were significantly positive (p < 0.01); And AAC was significantly correlated with HPV, CPV, SBV, PaT and PeT; Even though correlation analysis revealed that R was significantly correlated with most of parameters, but all the correlation coefficients were small (r < 0.2) (Additional file 4:Fig. S2). On the other hand, significantly positive correlations were found between GT and PeT, PaT, and their correlation coefficients were 0.94 and 0.99, respectively.
Effects of SSIV-2 on rice ECQs
Among these investigated 184 plants from BC1F12 line 529, SSIV-2 alleles were divided three genotypes: type G, type C and type H. Significant differences on GC, PKV and SBV among them were detected by analysis of variance (ANOVA). Multiple comparisons (Fig. 4) showed that GC of type G was the softest (65.50 mm), and obviously higher than that of type C and type H, whose GC was the hardest (57.41 mm) and intermediate (59.77 mm), respectively. And same phenomenon was observed in SBV. However, the PKV of type C was significantly higher than that of type G and type H (p<0.01), but there was no notable difference between the PKV of type G and type H.
Next, NIL-SSIV-2(BC1F13) with 341 plants was constructed in order to further confirm the effects of SSIV-2 on rice ECQs. Consistently, GC, PKV and SBV of type G and type C significantly differed from each other. In addition, significant differences were also observed among different genotypes on BDV, GT and R. Moreover, the samples with type G had significant lower PKV, in contrast, GC, SBV, BDV, GT and R of type G significantly increased by 15.05%, 11.55%, 4.81%, 1.12% and 36.04%, respectively, compared with that of type C (p < 0.05 or p < 0.01) (Fig. 4). In order to make the results more powerful, comparisons between the plants of NIL-SSIV-2 with GZ63S genotype (NIL-SSIV-2-GZ) and CG173R (recurrent parent) on these parameters were conducted. GC, PKV, CPV, SBV and R of NIL-SSIV-2-GZ altered significantly, nevertheless obvious changes of the other traits didn’t occurred in NIL-SSIV-2-GZ (Table 1), which was basically consistent with the effects of SSIV-2 in BC1F12 population. These results suggested that the SSIV-2 allele derived from GZ63S could significantly improve GC, SBV, BDV, GT and R, but an opposite effect on PKV of starch. Therefore, SSIV-2 allele derived from CG173R (type C) had better quality traits with lower GT, SBV and R, but inferior traits with lower GC and BDV. Thus SSIV-2 polymorphism has a great influence on rice ECQs under the same major genes (Wx and SSII-3) background.
Fig. 4. Comparison of starch physicochemical properties among different SS IV-2 genotypes in BC1F12 and BC1F13 generation.
a) Gel consistency (GC); b) Peak viscosity (PKV); c) Setback viscosity (SBV); d) Breakdown viscosity (BDV); e) Gelatinization temperature (GT); f) Percent of retrogradation (R). Each column represents the mean ± standard deviation. Lowercase and capital letters above column denote significant at 0.05 and 0.01 levels, respectively. G denotes homozygous for GZ63S, C and H denotes homozygous for CG173R and heterozygote, respectively. Each sample was repeated twice
|
Combined effects of SSIV-2 and AGPlar alleles on rice ECQs
Due to SSIV-2 and AGPlar alleles separated simultaneously in the tested line 529, their interaction on rice ECQs were able to further analyze by a split block design. The results demonstrated that interacting effects of SSIV-2 and AGPlar alleles for ECQs were significant (p < 0.01) (Table S2). The combined effects of SSIV-2 and AGPlar alleles on rice ECQs should be further studied.
To the AGPlar locus, three genotypes, type I, type II and type III, were identified among the 184 plants, which resulted in nine combinations together with SSIV-2 among the population of line 529 (Table 2). Multiple comparisons revealed that GC, AAC, GT and R of the nine combinations significantly differed from each other (p < 0.05 or p < 0.01). The combination I-G and I-C possessed the softest gel (63.49 mm) and hardest gel (53.08 mm), respectively. Moreover, the AAC of I-G was similar with that of II-G, which was significantly higher than that of I-C and II-C. However, the I-C and II-G had similar GT and the GT was significantly higher than that of I-G and II-C, that with no significant difference from each other as well. R of II-C (16.63%) was lower than that of I-G (19.20%), I-C (20.72%) and II-G (19.99%) (p < 0.05), whereas no significant differences among the next three combinations.
Significant differences on RVA profile characteristics were observed among the different combinations (Fig. 5). Specifically, I-C was significant higher in PKV (4166.65 cP) and BDV (2181.00 cP), but lower in HPV (1985.65 cP), CPV (3040.62 cP), CSV (1054.98 cP) and SBV (-1126.02 cP) than other three combinations: I-G, II-C and II-G (p < 0.05 or p < 0.01). Interestingly, it was observed that there were significant differences in GC, PKV and SCV between different SSIV-2 genotypes under the background of type I, but not under the type II. Therefore, it implied that AGPlar had an epistasis effect on SSIV-2. Different combinations of SSIV-2 and AGPlar genotypes played an important role in determining GC, AAC, GT, R and all RVA profile characteristics except for PaT and PeT under the same major genes (Wx and SSII-3) background.
Fig. 5. Comparison of RVA profile characteristics for different combinations of SS IV-2 × AGPlar.
a) Peak viscosity (PKV); b) Hot paste viscosity (HPV); c) Cool paste viscosity (CPV); d) Breakdown viscosity (BDV); e) Consistence viscosity (CSV); f) Setback viscosity (SBV); Type I, II and III indicate homozygous for GZ63S, CG173R and heterozygote in AGPlar locus, respectively; Type G, C and H represent homozygous for GZ63S, CG173R and heterozygote in SSIV-2 locus, respectively; Each column represents the mean ± standard deviation. Lowercase and capital letters above column denote significant at 0.05 and 0.01 levels. Each sample was repeated twice
|
Combined effects of SS IV-2 and PUL alleles on rice ECQs
To the BIL 530, 93 plants were divided into three groups by PUL alleles: type 1, type 2 and type 3, but only five combinations of SS IV-2 and PUL alleles. Results of multiple comparisons revealed that five combinations significantly differed from each other in GC and AAC (Table 3). C-1 and C-3 (52.28 and 55.38 mm, respectively) possessed a median GC (range 41 - 60 mm), and the others had a soft GC (>61 mm). The AAC of G-2 (range 10.68-15.86%) was the highest, which was significantly higher than that of C-1 (range 5.20-13.06%).
The combined effects of SS IV-2 and PUL alleles were further investigated on rice RVA profile characteristics (Fig. 6). Significant difference was not detected on RVA profile characteristics between different SSIV-2 genotypes under the same background of PUL (type 2). However, the PKV (4287.61 cP) and BDV (2185.01 cP) were greatly increased in C-1, while the CSV (972.13 cP) and SBV (-1212.88 cP) were significantly decreased compared to that of G-2 and C-2. The results together indicated that combinations of SSIV-2 and PUL alleles had remarkable impact on GC, AAC, PKV, BDV, CSV and SBV under the same major genes (Wx and SSII-3) background.
Fig. 6. Comparison of RVA profile characteristics for different combinations of SSIV-2 × PUL. a) Peak viscosity (PKV); b) Hot paste viscosity (HPV); c) Breakdown viscosity (BDV); d) Cool paste viscosity (CPV); e) Consistence viscosity (CSV); f) Setback viscosity (SBV); Each column represents the mean ± standard deviation. Type G, C and H represent homozygous for GZ63S, CG173R and heterozygote in SSIV-2 locus, respectively; Type 1, 2 and 3 indicate homozygous for GZ63S, CG173R and heterozygote in PUL locus, respectively; Lowercase and capital letters above column denote significant at 0.05 and 0.01 levels
|