SUMOylation of rice DELLA SLR1 modulates transcriptional responses and improves yield under salt stress

Abstract

SUMOylation of SLR1 at K2 protects productivity under salt stress, possibly by modulation of SLR1 interactome. DELLA proteins modulate GA signaling and are major regulators of plant plasticity to endure stress. DELLAs are mostly regulated at the post-translational level, and their activity relies on the interaction with upstream regulators and transcription factors (TFs). SUMOylation is a post-translational modification (PTM) capable of changing protein interaction and has been found to influence DELLA activity in Arabidopsis. We determined that SUMOylation of the single rice DELLA, SLENDER RICE1 (SLR1), occurs in a lysine residue different from the one identified in Arabidopsis REPRESSOR OF GA (RGA). Artificially increasing the SUMOylated SLR1 levels attenuated the penalty of salt stress on rice yield. Gene expression analysis revealed that the overexpression of SUMOylated SLR1 can regulate GA biosynthesis, which could partially explain the sustained productivity upon salt stress imposition. Furthermore, SLR1 SUMOylation blocked the interaction with the growth regulator YAB4, which may fine-tune GA20ox2 expression. We also identified novel SLR1 interactors: bZIP23, bHLH089, bHLH094, and OSH1. All those interactions were impaired in the presence of SUMOylated SLR1. Mechanistically, we propose that SUMOylation of SLR1 disrupts its interaction with several transcription factors implicated in GA-dependent growth and ABA-dependent salinity tolerance to modulate downstream gene expression. We found that SLR1 SUMOylation represents a novel mechanism modulating DELLA activity, which attenuates the impact of stress on plant performance.

Keywords: Oryza sativa; DELLA; Gibberellin; Productivity; SLR1; SUMOylation; Salt stress; Transcription factors.

Fig. 1

Representation of the presence of SUMOylated lysines (K2 and K65) in several plant DELLA proteins, following a phylogenetic analysis of DELLA proteins (Fig. S1). SUMOylation motifs focusing on lysine 2 and the correspondent lysine 65 were investigated using GPS-SUMO 2.0 at the DELLA domain. Blue and white rectangles display the presence or absence of SUMOylation on the investigated lysines, respectively

Fig. 2

SLR1 is SUMOylated in a novel SUMO-conjugation canonical motif. a In silico prediction of SLR1 SUMOylation motifs scored by GPS-SUMO2 and comparison with AtRGA. K2 and K60 are marked in green and orange, respectively. b In vitro SUMOylation assays using rice SUMOylation machinery (SCE1, SAE1/2, and SUMO1) with recombinant full SLR1 (~ 65 kDa) and SLR1t (~ 19 kDa) proteins, treated with ULP1 SUMO protease, as indicated. SLR1 and SLR1t were detected by immunoblotting using a custom-made anti-SLR1 antibody, where b corresponds to the full-length SLR1, and c to the SLR1 N-terminal truncated (SLR1t) coding the leading 190 first SLR1 residues, as well as the mutated isoforms SLR1t K2R and SLR1t K60R. d Shoot protein extracts of 14-day-old Nipponbare plants subjected to ethanol (mock), GA₃ (10 µM) and PAC (0.1 µM), immunoblotted with a custom-made anti-SLR1 antibody. ULP1 treatment (right panel). Coomassie Brilliant Blue (CBB) staining for Rubisco (RbcL) was used as a loading control. The band density was measured with ImageJ and the quantification showed beneath the WB analysis

Fig. 3

SLR1 SUMOylation confers salt tolerance at the seedling stage and reduces the impact of salt stress on yield. a Accumulation of SLR1 under salt stress. Shoot protein extracts of 14-day-old Nipponbare plants subjected to 120 mM of NaCl for 8 h. Immunoblotting with a custom-made anti-SLR1 antibody. ULP1 treatment (right panel). Coomassie Brilliant Blue (CBB) staining for Rubisco (RbcL) was used as a loading control. The band density was measured with ImageJ and the quantification showed beneath the WB analysis. b Schematic illustration of gene constructs used to generate rice SLR1-OX and SUMO1SLR1-OX transgenic lines. Zea mays Ubi1 (ZmUbi1) promoter drives the expression of both SLR1 (SLR1-OX) and SLR1-SUMO1 gene fusion (SUMO1SLR1-OX). c SLR1 and SUMO1SLR1 protein accumulation in T1 independent transgenic lines (#1 and #2) of SUMO1SLR1-OX and SLR1-OX, negative segregant (NS) and SLR1 recombinant protein used as positive control. Immunoblot using a custom-made anti-SLR1 antibody. d Stress severity as evaluated by IRRI salinity trial SES score for all lines in 6 and 8 days of stress and after 8 days of recovery period in 14-day-old seedlings (n = 6 and n = 3 for SUMO1SLR1-OX #1) e Representative plants depicting the differences in shoots of 14-day-old Nipponbare, SLR1-OX, and SUMO1SLR1-OX plants after 8 days of salinity imposition with 120 mM NaCl and a recovery period of 8 days. Scale bar = 10 cm. f Shoot height seedlings measurements for all analyzed lines after 8 days of salinity stress. g–k Nipponbare (Nipp), SLR1-OX, and SUMO1SLR1-OX plants under salt stress at the booting stage (80 mM) (n = 6). g Height as in aerial part length (n = 6). h Number of tillers (n = 6). i Panicle number (n > 10). j Percentage of filled seeds per total of seeds (n > 10) (g) 1000-grain weight (n > 10). Error bars correspond to standard error and statistical significance represented as * (P < 0.05) and **(P < 0.01) for comparing mock and stress conditions

Fig. 4

Salt treatment had a major contribution to the differences in gene expression among Nipponbare, SLR1-OX, and SUMO1SLR1-OX. a Principal Component Analysis (PCA) based on gene transcriptional profile obtained with RNA-seq of 14-day-old rice shoot samples from WT (Nipponbare), SLR1-OX, and SUMO1SLR1-OX lines, after 8 h under salt stress and control conditions. b–c Venn diagrams showing the overlaps between SLR1-OX and SUMO1SLR1-OX differentially expressed genes (DEGs) compared to Nipponbare in control conditions (b) and differentially expressed genes found in SLR1-OX, SUMO1SLR1-OX, and Nipponbare lines under salt stress compared to the respective genotypes under control conditions (c). Red: upregulated genes; blue: downregulated genes; yellow: contrasting gene expression. d–e GO functional enrichment analysis of the differently expressed genes in SLR1-OX and SUMO1SLR1-OX in control conditions (d), and SLR1-OX, SUMO1SLR1-OX, and Nipponbare lines under salt stress compared to the respective genotypes under control conditions (e), respectively. BP stands for biological processes

Fig. 5

SLR1 SUMOylation impairs its interaction with key transcription factors. Y2H analysis of both SLR1 and SUMO1_(GG-AA)-SLR1. SUMO1-SLR1 fusion carries two mutations in the SUMO-protease cleavage motif (GG → AA) to avoid its removal by endogenous yeast proteases and guarantee constitutive SUMO attachment to SLR1 during the assay SLR1 and SUMO1_(GG→AA)-SLR1 were fused with the GAL4 activation domain (AD) and bHLH089, bHLH094, bZIP23, IDEF1, OSBZ8, OSH1, PCF1, and YAB4 with the GAL4 DNA binding domain (BD) into appropriate expression vectors before transfer into yeast (Y2HGold strain). The different yeast strains were plated on a synthetic complete selective medium lacking Leu and Trp (SD/-Leu/-Trp) or on a synthetic complete medium lacking Trp, Leu, Ade, and His (SD/-Leu/-Trp/-Ade/-His) for the screening. pAD-WT/pBD-WT (Wild-type fragment C of lambda cI repressor) was used as a positive control (C +). The empty BD vector was used as a negative control (Fig. S5a). The interaction was confirmed by three different clones