Transcription factors NAC20 and NAC26 interact with RPBF to activate albumin accumulations in rice endosperm

Storage proteins (SPs) and starch deposited in cereal endosperms are the most important sources of human food. SPs are classified into glutelins, prolamins, albumins and globulins according to their solubility in various solvents (Müntz, 1998). Most of these proteins are encoded by multiple genes and deposited specifically in endosperms through the regulation of transcription factors (TFs) (Liu et al., 2022). Among them, albumin is the most abundant water-soluble protein (Müntz, 1998), and the most important allergenic protein in the rice endosperm (Matsuda et al., 1991). Thus, understanding molecular machinery underlying albumin accumulations is critically important. In maize, NAC TFs, NAC128/130, regulate starch and SP accumulations (Zhang et al., 2019). NAC20 and NAC26 (NAC20/26 hereafter) are specifically expressed in developing rice endosperm, and mutations of both genes led to floury endosperm, together with compromised accumulations of starch and SP (Wang et al., 2020). In this study, a NAC20/26 double knockout mutant, named nac20/26-3, was generated in a Japonica rice variety Zhonghua 11 (ZH11; Figure S1), which exhibited similar floury endosperm phenotype as reported (Wang et al., 2020; Figure S2). SDS-PAGE analyses of total proteins extracted from mature nac20/26-3 endosperms showed that a 16-kDa band was disappeared, and a 26-kDa band was attenuated greatly when compared with ZH11 (Figure 1a). Previous studies suggest that the former is prolamin, and the latter is α-globulin (αGlb; Wang et al., 2020). Immunoblotting performed in this study showed that αGlb accumulation was significantly reduced, while prolamin was only slightly reduced in nac20/26-3 (Figure S3), which is inconsistent with the SDS-PAGE result (Figure 1a). Subsequently, mass spectrometry analysis of the 16-kDa band excised from gel showed that abundances of four albumins encoded by Os07g11360 (Alb1), Os07g11380 (Alb2), Os07g11330 (Alb3) and Os07g11410 (Alb4) were decreased sharply in nac20/26-3 (Table S1). Immunoblot analysis using anti-Albs antibody confirmed that albumin accumulations were almost disappeared in nac20/26-3 (Figure 1a). As expected, when water-soluble proteins were extracted and analysed by SDS-PAGE, the albumin band was hardly detectable in nac20/26-3 (Figure 1b). Systematic analyses of reported 16-kDa albumins and allergenic proteins, together with whole-genome sequencing data, showed that 16-kDa albumins in rice are encoded by five genes including Alb1~4 and Os07g11510 (Alb5; Alb1~5 hereafter; Figure S4; Wakasa et al., 2011; Zhou et al., 2017). qRT-PCR analyses revealed that expressions of Alb1~5 were decreased dramatically in nac20/26-3 endosperms (Figure 1c), suggesting that NAC20/26 may regulate albumin deposits in rice endosperm through transcriptional activations. To confirm if NAC20/26 activate directly the expressions of Alb1~5, a dual-luciferase (LUC) transactivation assay was performed in rice protoplasts by co-transformations pairwisely of effector constructs of p35S:NAC20 and p35S:NAC26, and reporter constructs of either pAlb1:LUC, pAlb2:LUC, pAlb3:LUC or pAlb4:LUC or pAlb4:LUC5. Results showed that NAC20/26 activated Alb1~5 expressions, in which NAC26 was more effective than NAC20 (Figure 1d; Figure S5). Domain swapping experiment, together with transactivation assay in rice protoplasts, showed that the intermediate region of NAC26, from 149th to 214th amino acid (AA), attributed to its higher efficiency (Figure 1e; Figure S6). CACG as the core binding motif of NAC TFs is present in promoter regions of 26-kDa globulin, 16-kDa prolamin, GluA-1 and GluB4/5 (Wang et al., 2020; Figure S7). CACG motifs (Figure S8), some with conserved flanking sequences (Figure S9), were identified in 5′ upstream sequences of Alb1~5. Bindings of NAC20/26 to Alb1~2 5′ upstream sequences, as representatives, were confirmed by electrophoretic mobility shift assay (EMSA; Figure 1f). Competition assay with probes carrying single-nucleotide substitutions covering CACG motif and its flanking nucleotides in Alb1 promoter, as a representative, showed that ACG was critical for NAC20/26 bindings (Figure S10). These results together demonstrate that NAC20/26 regulate albumin expressions through binding to ACG motif. We then used NAC20 to identify interacting proteins through yeast two-hybrid. As the full-length NAC20 exhibited auto-activation in yeast, a fragment from 86th to 320th AA was selected as the bait. Among 65 putative interaction clones identified (Table S2), two encoded Rice Prolamin Binding Factor (RPBF) that regulates starch and SP accumulations in rice (Kawakatsu et al., 2009). Firefly luciferase complementation imaging (LCI) and bimolecular fluorescence assays performed in tobacco leaves showed that NAC20/26 interacted with RPBF (Figure 1g; Figure S11). Co-immunoprecipitation (Co-IP) analysis showed that both NAC20 and NAC26 precipitated RPBF (Figure 1h). Two independent mutant alleles, rpbf-1 and rpbf-2, were generated by gene-editing, and their grains exhibited shrunken phenotype as reported (Kawakatsu et al., 2009; Figure S12). SDS-PAGE analysis revealed that intensities of major SPs including albumin were reduced in their endosperms (Figure 1i, upper panel). Immunoblot analysis using anti-Albs confirmed that, compared with ZH11, albumin accumulation was significantly reduced in rpbf-1 and rpbf-2 (Figure 1i, middle panel). qRT-PCR analyses performed in rpbf-2 endosperms confirmed that expressions of Alb1~2 (as representatives) were dramatically decreased (Figure 1j), suggesting that RPBF regulates albumin accumulations through transcriptional activation. It has been showed that RPBF activates Alb1 expression in rice callus protoplasts (Yamamoto et al., 2006). Although a confined RPBF-binding motif AAAG was frequently identified in promoter regions of Alb1~5, no intact P-box (TGTAAAG) has been found (Figure S8). Dual-luciferase transactivation assay performed in seedling protoplasts by co-transformations pairwisely of effector construct p35S:RPBF with reporter constructs of either pAlb1:LUC or pAlb2:LUC showed that RPBF activated expressions of Alb1~2 (Figure 1k). However, when p35S:RPBF was included in transactivation assays of p35S:NAC20 with pAlb1:LUC or pAlb2:LUC, and p35S:NAC26 with pAlb1:LUC or pAlb2:LUC, increased transactivation activities were observed (Figure 1l), suggesting that RPBF may act as a cofactor for NAC20/26. It is plausible that the transcriptional activity observed in callus protoplasts may cause by faint expressions of NAC20 or NAC26, or the shorter promoter sequence used (Yamamoto et al., 2006). In summary, we demonstrate that NAC20/26 act redundantly, with different efficiencies, on ACG motif of Alb promoters to regulate their expressions, while RPBF functions as a cofactor to enhance NAC20/26 on transcriptions (Figure 1m). It has been reported that RPBF regulates the expressions of starch and SP genes in rice endosperms through interactions with RISBZ1, a bZIP-type TF (Kawakatsu et al., 2009). In maize, it has been showed that the RPBF orthologue of PBF functions additively with a bZIP TF O2, regulating expressions of multiple SPs (Zhang et al., 2015), and NAC128/130, maize orthologues of NAC20/26, regulate expression of SPs (Zhang et al., 2019). If PBF/RPBF function as a general cofactor, instead of specific TF, to coordinate activities of other TFs in storage product accumulations in cereal endosperms remain to be investigated. Further, since albumin is the most important allergenic protein in rice (Matsuda et al., 1991), the discovery of this study may provide a unique avenue to cope with the allergenic problem for rice consumers. This work was supported by National Key R&D Program of China (2021YFF1000200), the Innovation Project of Chinese Academy of Sciences (XDA24010402), and National Natural Science Foundation of China (32070348). The authors declare no conflict of interest. C.M.L. designed and supervised the project; M.W.W. performed most of the experiments; X.B. and W.Q.C. performed transactivation and Co-IP assays; M.W.W., Y.R. and W.Q.C. performed immunoblot analyses; other authors provided technical supports; M.W.W., J.L. and C.M.L. wrote the paper. Appendix S1 Materials and methods. Figure S1–S12 Supplementary Figures. Table S1–S3 Supplementary Tables. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.