Insights into the transcriptional and post-transcriptional regulation of the rice SUMOylation machinery and into the role of two rice SUMO proteases

Abstract

Background: SUMOylation is an essential eukaryotic post-translation modification that, in plants, regulates numerous cellular processes, ranging from seed development to stress response. Using rice as a model crop plant, we searched for potential regulatory points that may influence the activity of the rice SUMOylation machinery genes.

Results: We analyzed the presence of putative cis-acting regulatory elements (CREs) within the promoter regions of the rice SUMOylation machinery genes and found CREs related to different cellular processes, including hormone signaling. We confirmed that the transcript levels of genes involved in target-SUMOylation, containing ABA- and GA-related CREs, are responsive to treatments with these hormones. Transcriptional analysis in Nipponbare (spp. japonica) and LC-93-4 (spp. indica), showed that the transcript levels of all studied genes are maintained in the two subspecies, under normal growth. OsSUMO3 is an exceptional case since it is expressed at low levels or is not detectable at all in LC-93-4 roots and shoots, respectively. We revealed post-transcriptional regulation by alternative splicing (AS) for all genes studied, except for SUMO coding genes, OsSIZ2, OsOTS3, and OsELS2. Some AS forms have the potential to alter protein domains and catalytic centers. We also performed the molecular and phenotypic characterization of T-DNA insertion lines of some of the genes under study. Knockouts of OsFUG1 and OsELS1 showed increased SUMOylation levels and non-overlapping phenotypes. The fug1 line showed a dwarf phenotype, and significant defects in fertility, seed weight, and panicle architecture, while the els1 line showed early flowering and decreased plant height. We suggest that OsELS1 is an ortholog of AtEsd4, which was also supported by our phylogenetic analysis.

Conclusions: Overall, we provide a comprehensive analysis of the rice SUMOylation machinery and discuss possible effects of the regulation of these genes at the transcriptional and post-transcriptional level. We also contribute to the characterization of two rice SUMO proteases, OsELS1 and OsFUG1.

Keywords: Alternative splicing; Rice (Oryza sativa); SUMO proteases; SUMOylation; T-DNA; cis-elements.

Fig. 1

Analyses of three rice putative OsSUMOs proteins. a Maximum likelihood phylogeny of the SUMO family. Only nodes with bootstrap support > 75% show the correspondent bootstrap value. Os – Oryza sativa, Sc - Saccharomyces cerevisiae, At – Arabidopsis thaliana, Zm – Zea mays and Hv – Hordeum vulgare. b Protein alignment of the three rice putative SUMO proteins OsSUMO1, OsSUMO2 and OsSUMO3. Diglycine motif is highlighted (++). The lysines with a high probability of being SUMOylated are marked with “S” in bold and the ones with low SUMOylation probability are marked with a regular “S”. c Venn diagram showing common CREs in the promoter region of the different OsSUMOs. d Basal expression levels of OsSUMO1/2 and OsSUMO3 genes by qPCR, in shoots (no pattern) and roots (patterned) in LC-93-4 (LC) and Nipponbare (Nipp)

Fig. 2

Analysis of the rice genes encoding the SUMO activating enzyme (SAE). The rice E1 is constituted by OsSAE1 encoding the regulatory subunit and OsSAE2 encoding the catalytic subunit. Maximum likelihood phylogeny for OsSAE1 (a) and OsSAE2 (b). Only nodes with bootstrap support > 75% show the correspondent bootstrap value. Os – Oryza sativa, Sc - Saccharomyces cerevisiae, At – Arabidopsis thaliana, Zm – Zea mays and Hv – Hordeum vulgare. c Venn diagram showing common CREs in the promoter region of both E1 subunit genes. Schematic representation of SAE alternative splicing forms showing three ASF for both OsSAE1 (d) and OsSAE2 (e). White boxes – untranslated regions (UTR); black boxes – exons; lines – introns. The proteins lengths are indicated. The difference in the sequence between OsSAE2.1 and OsSAE2.2/3 is indicated by an arrow. The arginine in OsSAE1 and the catalytic cysteines in OsSAE2 are depicted, as well as the domains and protein length for all ASFs. SAE2 proteins have four domains: adenylation domain where the catalytic cysteine domain is located, followed by the ubiquitin-like (UBL) and the C-terminal domains [108]. f Alignment of SAE2 proteins from different organisms Saccharomyces cerevisiae, Arabidopsis thaliana, and the two rice SAE2 ASF. The five amino acid insertion in OsSAE2.1 is highlighted. g Basal expression levels of the different ASF of OsSAE genes by qPCR in shoots (no pattern) and roots (patterned) in LC-93-4 (LC) and Nipponbare (Nipp). h and (i) Transcript level profile of ASFs of genes OsSAE1 and OsSAE2, respectively, in response to 15 μM ABA or 100 μM of GA

Fig. 3

Analysis of the rice three genes of the E2 SUMO conjugating enzyme, OsSCE1a/b/c. a Maximum likelihood phylogeny of the SCE family. Only nodes with bootstrap support > 75% show the correspondent bootstrap value. Os – Oryza sativa, Sc - Saccharomyces cerevisiae, At – Arabidopsis thaliana, Zm – Zea mays and Hv – Hordeum vulgare. b Protein alignment of OsSCE1a, OsSCE1b and OsSCE1c. The catalytic cysteine in the active center is highlighted with “*”, and the E1 contact residues are highlighted with arrows. The residues that differ in Class II SCE from Class I are highlighted with “+”. c Venn diagram showing common CREs between OsSCEs. d Schematic representation of OsSCE alternative splicing forms, showing two for each OsSCE. White boxes – untranslated regions (UTR); black boxes – exons; lines – introns. The proteins lengths are indicated. High probability SUMOylation residues are indicated with an “S” in bold, whereas low probability SUMOylation residues are indicated with a normal “S”. e Basal expression levels of the different ASF of OsSCE1a/b/c genes by qPCR in shoots (no pattern) and roots (patterned) in LC-93-4 (LC) and Nipponbare (Nipp). OsSCE1b.1 and OsSCE1b.2 were quantified together since they could not be discriminated. f Transcript level profile of ASFs of OsSCE1a in response to 15 μM ABA

Fig. 4

Analysis of the rice E3 SIZ and HPY2 classes. Maximum likelihood phylogeny of the E3 SUMO ligase family performed with the catalytic SP-RING domain: a OsSIZ and (b) OsHPY2. Only nodes with bootstrap support > 75% show the correspondent bootstrap value. Os – Oryza sativa, Sc - Saccharomyces cerevisiae, At – Arabidopsis thaliana, Zm – Zea mays and Hv – Hordeum vulgare. c Venn diagram showing common CREs of E3 SUMO ligases. d Schematic representation of E3 SUMO ligases alternative splicing forms, showing three ASF for OsSIZ1 and two for OsHPY2. The difference between OsSIZ1.1–2 is three nucleotides missing in OsSIZ1.2 (arrow). White boxes – untranslated regions (UTR); black boxes – exons; lines – introns. The different domains and proteins lengths are indicated. In the case of OsHPY2, black indicates a confirmed ASF structure, which is not available for OsHPY2.2 (in gray). Basal expression levels of OsSIZ (e) and OsHPY2 (f) by qPCR in shoots (no pattern) and roots (patterned) of LC-93-4 (LC) and Nipponbare (Nipp). OsSIZ1.1 and OsSIZ1.2 were analysed together. g Transcript level profile of ASFs of OsSIZ1 in response to 15 μM ABA or 100 μM of GA

Fig. 5

The analysed SUMO proteases in rice: Os03g22440 (OsFUG1), Os01g25370 (OsELS1), Os05g11770 (OsSPF1), Os03g29630 (OsELS2) and Os01g53630 (OsOTS3). a Maximum likelihood phylogeny of the SUMO proteases family. Only nodes with bootstrap support > 75% show the correspondent bootstrap value. Os – Oryza sativa, Sc - Saccharomyces cerevisiae, At – Arabidopsis thaliana, Zm – Zea mays and Hv – Hordeum vulgare. Only the C48 peptidase domain was used to perform the alignments. b Venn diagram showing common CREs between five rice SUMO proteases: Os03g22440 (OsFUG1), Os01g25370 (OsELS1), Os05g11770 (OsSPF1), Os03g29630 (OsELS2) and Os01g53630 (OsOTS3). c Schematic representation of the SUMO proteases alternative splicing forms, showing three alternative splicing forms for OsFUG1 and OsELS1, and two for OsSPF1. White boxes – untranslated regions (UTR); black boxes – exons; lines – introns. Proteins lengths are indicated as well as the location of the catalytic triad of the C48 peptidase domain (histidine, aspartate and cysteine). d Basal expression levels of the different ASF of SUMO proteases genes performed by qPCR, in shoots (no pattern) and roots (patterned) in LC-93-4 (LC) and Nipponbare (Nipp)

Fig. 6

Molecular characterization of the rice T-DNA lines. a Expression levels of the gene of interest relative to its background line by qPCR: OsSCE1c.1 expression in 1A-23738 vs wild type Hwayoung (Hway), OsFUG1.1 expression in 2A-20225 vs Hwayoung; OsSAE1.1 expression in 3D-00611 vs wild type Dongjin; OsSCE1a.1 expression in 3A-05464 vs Dongjin; OsSIZ1.1–2 expression in 3A-02154 vs Dongjin; OsELS1.1 expression in 04Z11JY66 vs wild type Zhonghua11 (Zh11). Expression levels of each transcript in the T-DNA lines were normalized relative to its respective wild type (with a normalized value of 1). b-f T-DNA insertion localization site in the respective genes. White boxes – untranslated regions (UTR); black boxes – exons; lines – introns. b T-DNA in OsSCE1a knockdown (KD) line. c T-DNA in OsSCE1c knockout (KO) line. d T-DNA in OsELS1 KO line. e T-DNA in OsFUG1 KO line. f T-DNA in OsSIZ1 KD line is located in the 15th exon according to Wang et al. (2011). g-i Global levels of SUMOylation in the shoots of the T-DNA insertion lines in Western blots with anti-SUMO1, in normal growth conditions. T-DNA lines were compared to both wild type (WT) and negative segregant (NS) plants. Global SUMOylation levels in the KO lines of OsFUG1 and OsELS1 (g), in the T-DNA lines of OsSCE1c and OsSCE1a (h) and in OsSIZ1 KD line (i). HMWSC – High Molecular Weight SUMO Conjugates. LC – Loading Control (Coomassie blue staining). Free SUMO is marked with an asterisk

Fig. 7

Phenotype characterization of the rice T-DNA insertion lines vs. wild type and negative segregant plants. a Plant height measured at reproductive stage. b Phenotype of OsFUG1 knockout (KO) line and the respective negative segregant plants (NS). c Heading date measured as panicle ripening. The phenotype difference is exemplified by OsELS1 KO line vs NS in (d). Seed-related parameters are shown in (e-i). e Percentage of fertility. f Seed weight shown as the weight of 100 seeds. g Panicle length (cm). h Total seeds per panicle. i Panicle phenotype of OsFUG1 KO line versus NS. Asterisks represent statistical significance (p-value < 0.05). Only the significant differences between the T-DNA lines and their respective wild type/negative segregant lines are depicted