Creating high‐resistant starch rice by simultaneous editing of SS3a and SS3b

The conventional starch from cereals is usually high in energy and easily digested, absorbed and converted into blood sugar in human small intestine, while resistant starch (RS) is hardly enzymatically hydrolysed into glucose in human small intestine and only fermented into beneficial short‐chain fatty acids in large intestine (Jukanti et al., 2020; Shen et al., 2022). Thus, RS‐rich foods can not only effectively reduce the glycaemic index, increase satiety and prevent blood sugar‐related diseases but also contribute to the prevention of intestinal‐related diseases by improving the intestinal micro‐environment and decreasing colonic pH (Zaman and Sarbini, 2016). Rice is an important staple‐food crop and a major source of starch for majority of the population, especially in Asia. However, the RS content in conventionally cultivated rice is usually <1%, which is far below the daily recommended RS intake for humans (Yadav et al., 2010). Therefore, breeding rice rich in RS is an important target for rice variety improvement.

Rice endosperm is the chemical reservoir storing more than 80% starch with varied amounts of amylose and amylopectin, which is mainly controlled by several key starch synthase‐related genes (SSRGs; Huang et al., 2021). Currently, there have been a few successful cases of altering the composition and structure of starch to enhance RS content in rice endosperm by modifying the expression of two kinds of SSRGs, SBE and SS3/SSIII, encoding starch branching enzyme and soluble starch synthase III, respectively (Butardo Jr. et al., 2012; Guo et al., 2020; Miura et al., 2022; Zhou et al., 2016; Zhu et al., 2012). Loss of function of SBE3/SBEIIb resulted in increased RS in rice endosperm, while there is no significant effect of SBE1/SBEI mutation on RS level, but the sbe1sbe3 double mutation could further enhance RS content substantially based on sbe3 single mutation, indicating functional redundancy of these two isoforms in RS formation (Zhu et al., 2012). In rice, there are two isoforms of SS3/SSIII, SS3a/SSIIIa/SSIII‐2 and SS3b/SSIIIb/SSIII‐1 (Huang et al., 2021). Zhou et al. (2016) reported that the ssIIIa/ss3a loss‐of‐function mutant had a high RS content in cooked rice. However, the role and genetic effects of most other SSRGs, such as SS3b, in the formation of RS are not clear.

In rice, SS3a and SS3b, two genes encoding the SS3 isoforms, and their gene structure, protein domain and amino acid sequence were quite conserved (Figure S1a–c; Table S2). SS3b was expressed in low abundance in developing rice grains, but its expression trend was consistent with that of SS3a (Figure S1d,e). It implied that SS3b may also have a similar function as SS3a in regulating starch synthesis and RS formation in rice endosperm. Thus, we created SS3a and SS3b mutants by CRISPR/Cas9 genome editing technology in the japonica rice cultivar Nipponbare (WT). After several generations of genotype and phenotype screening, we obtained several homozygous mutant lines, including SS3a single mutants (ss3a), SS3b single mutants (ss3b) and their pyramiding double mutant lines (ss3a‐ss3b) (Figure 1a,b). There was no significant changes in the seedling growth and plant development of all edited rice lines compared with their wild type (WT; Figure S2a,b).

Figure 1.

Generation and characteristics of ss3 mutants and their wild type (WT). (a) Grain morphology. Scale bar = 5 mm. (b) The schematic diagram of target sites and mutation sites. Font marked in red indicates the target site or the changed base sequence. (c, d) The contents of and in rice flour. (e) In vitro digestion rate of cooked rice. (f–h) The contents of apparent amylose (f), triglyceride (g) and amylose–lipid complex (h) of rice flour. (i–p) Starch granule morphology of grain cross section (i–l) and purified leaf starches (m–p). From left to right are WT, ss3a, ss3b and ss3a‐ss3b mutants, respectively. The red box (i–l) represents the viewing area of the SEM. The observation multiples are 2000 (i–l) and 5000 (m–p), respectively. (q, r) Amylopectin chain‐length distribution of purified starches from endosperm (q) and leaves (r). Different lower‐case letters indicate statistically significant differences at P < 0.05, n = 3. “**” indicates statistically significant differences between ss3a, ss3a‐ss3b mutants and WT at the P < 0.01 level, n = 3.

The digestive characteristics of the grains from the above mutants were firstly determined. The results showed that the RS content in ss3a mutants increased significantly to 4.76%–5.01% while that in WT was only 0.58% (Figure 1c). Moreover, the total digestible starch (TDS) content and digestion rate in ss3a mutants decreased significantly (Figure 1d,e), which was consistent with previous report (Zhou et al., 2016). In ss3b mutants, there was no significant change in the contents of RS and TDS as well as digestion rate compared with those in wild type. But SS3b mutation could significantly enhance the RS level in the background of SS3a mutation, resulting in the RS content reaching 9.54%–9.73% in the ss3a‐ss3b double mutants, and further decreased TDS content and digestion rate (Figure 1d,e).

Except digestion properties, the SS3b mutation had no significant effect on other grain physicochemical properties, could further strengthen the change of grain physicochemical phenotype combined with the SS3a mutation (Figure 1f,g and Figure S3a–c; Tables S3 and S4). These results suggested that SS3b can only have a significant effect on starch synthesis and RS formation in rice endosperm based on SS3a mutation, indicating ss3b mutant has synergistic effects on ss3a mutant in increasing RS content.

The starch granule morphology of ss3a and ss3a‐ss3b mutants was obviously abnormal, showing most compound starch grains composed of only a few starch granules, instead, dozens of starch granules formed a compound starch grain in the wild type (Huang et al., 2021), and some starch granules were spherical (Figure 1i–l and Figure S3e–h). Furthermore, the starch granules in each compound starch granule complex were closely linked, but loosely arranged between the adjacent compound starch granules in ss3a and ss3a‐ss3b mutants (Figure 1i–l). These results explained the high chalkiness phenotype in ss3a and ss3a‐ss3b mutants. The starch granule morphology of ss3b mutant had no obvious change compared with the wild type.

As to starch fine structure, the true amylose fraction (Am) increased significantly and the proportion of amylopectin long chains (Ap2) decreased dramatically in ss3a and ss3a‐ss3b mutants, and the range of changes in ss3a‐ss3b mutants was larger than that in ss3a mutants, but no significant change occurred in ss3b mutants (Figure S5a). Besides, the amylopectin B₂ chains with degree of polymerization (DP) 22–37 increased in ss3a mutants, while the amylopectin A chains with DP 6–9 and B₁ chains with DP 13–21 decreased dramatically in ss3a and ss3a‐ss3b mutants (Figure 1q). The X‐ray diffraction (XRD) results showed that the starch relative crystallinity of ss3a and ss3a‐ss3b mutants was significantly reduced (Figure S3d). Correspondingly, the area ratio of the characteristic peak, which representing the amylose–lipid complex as well as the type 5 RS (RS5), was significantly increased in the ss3a and ss3a‐ss3b mutants, reaching 1.7 and 2.3 times that of the wild type, respectively, consistent with the changes in triglyceride and RS contents (Figure 1h). Expectedly, the SS3b mutation had no significant effect on the relative crystallinity and RS5 content of rice endosperm starch compared with the wild type (Figure 1h and Figure S3d). It implied that the increased contents of amylose and amylose–lipid complex are the main  reasons for the high RS content in ss3a and ss3a‐ss3b mutants.

For transitory starch in leaves, there was no significant change in starch composition, structure, and starch granule morphology after SS3a mutation, while in ss3b mutants, TSC and proportions of amylopectin B chains (DP > 12) were significantly decreased, AAC and true AC were slightly increased, and starch granules were irregular in shape and rough in surface, similar to the effect of SS3a mutation on endosperm starch synthesis (Figure 1m–r, Figures S4 and S5b). The TSC and proportions of amylopectin B chains in leaves of ss3a‐ss3b mutant were further decreased, and the abnormal morphology of starch granules was aggravated, but there was no significant change on AC (Figure 1m–r, Figures S4 and S5b). These results suggested that SS3b plays a major role in leaf starch synthesis and ss3a mutant has synergistic effects on ss3b mutant.

In conclusion, our data indicated that SS3a and SS3b have synergistic effects in starch biosynthesis. In rice endosperm, further mutation of SS3b could strengthen the effect of SS3a mutation on alteration of starch physicochemical properties and RS content. A new rice germplasm with significantly increased RS content and improved digestive properties was created by co‐knockout SS3a and SS3b. These results provide a new strategy to breed novel rice varieties with high RS content and further help to elucidate the functions of SSRGs in cereals.

Conflict of interest

The authors declare no conflict of interest.

Author contributions

Q.Q.L., L.H. and C.Z. conceived the project. L.H., Y.X., W.Z., Y.R., H.S., D.Z., X.F., C.Z. and Q.F.L. carried out the experiments and analysed the data. L.H. and Q.Q.L. wrote the manuscript. All authors read and approved of the manuscript.

Supporting information

Data S1 Supplementary Materials and Methods.

Figure S1‐S5 Supplementary Figures

Table S1‐S4 Supplementary Tables.