Creating high‐resistant starch rice by simultaneous editing of <scp><i>SS3a</i></scp> and <scp><i>SS3b</i></scp>

Abstract

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). 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, that is they had comparative levels of total starch content (TSC), apparent amylose content (AAC), triglyceride content, gelatinization temperature (GT) and viscosity properties with its wild type (Figure 1f,g and Figure S3a–c; Tables S3 and S4). But, the SS3b mutation could further strengthen the change of grain physicochemical phenotype combined with the SS3a mutation, leading to further increase in AAC and triglyceride content, significant decrease in TSC, GT and RVA (Rapid Viscosity Analysis) profiles (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 rice grain length and width were similar among different edited lines and WT (Figure 1a and Figure S2c–e), but the chalky grain percentage and chalkiness degree increased significantly in the cases of SS3a mutation, including ss3a and ss3a-ss3b mutant lines, and the 1000-grain weight decreased significantly due to the remarkable decrease in grain thickness (Figure S2f–i). Interestingly, there was no significant change in grain morphology and appearance of the ss3b mutants, but it exacerbated the negative effect of ss3a mutants on grain thickness, leading to further decrease in grain thickness and 1000-grain weight in ss3a-ss3b than that of the ss3a (Figure S2f,g). 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 prop