The SUMO E3 ligase, AtSIZ1, regulates flowering by controlling a salicylic acid-mediated floral promotion pathway and through affects on FLC chromatin structure

Abstract

Loss-of-function siz1 mutations caused early flowering under short days. siz1 plants have elevated salicylic acid (SA) levels, which are restored to wild-type levels by expressing nahG, bacterial salicylate hydroxylase. The early flowering of siz1 was suppressed by expressing nahG, indicating that SIZ1 represses the transition to flowering mainly through suppressing SA-dependent floral promotion signaling under short days. Previous results have shown that exogenous SA treatment does not suppress late flowering of autonomous pathway mutants. However, the siz1 mutation accelerated flowering time of an autonomous pathway mutant, luminidependens, by reducing the expression of FLOWERING LOCUS C (FLC), a floral repressor. This result suggests that SIZ1 promotes FLC expression, possibly through an SA-independent pathway. Evidence indicates that SIZ1 is required for the full activation of FLC expression in the late-flowering FRIGIDA background. Interestingly, increased FLC expression and late flowering of an autonomous pathway mutant, flowering locus d (fld), was not suppressed by siz1, suggesting that SIZ1 promotes FLC expression by repressing FLD. Consistent with this, SIZ1 facilitates sumoylation of FLD that can be suppressed by mutations in three predicted sumoylation motifs in FLD (i.e. FLDK3R). Furthermore, expression of FLDK3R in fld protoplasts strongly reduced FLC transcription compared with expression of FLD, and this affect was linked to reduced acetylation of histone 4 in FLC chromatin. Taken together, the results suggest that SIZ1 is a floral repressor that not only represses the SA-dependent pathway, but also promotes FLC expression by repressing FLD activity through sumoylation, which is required for full FLC expression in a FRIGIDA background.

Figure1

SIZ1 represses transition to flowering. (a) Wild-type (Col-0) and siz1 (siz1-2 and siz1-3) plants were grown under long days (LD) for 32 days. (b) Wild-type (Col-0) and siz1 (siz1-2 and siz1-3) plants were grown under short days (SD) for 113 and 50 days, respectively. (c) The number of rosette leaves at flowering of wild-type and siz1 plants. Plants were grown under LD or SD. (d) Phenotypic comparison of plants: wild-type (Col-0), siz1-2 and transgenic Pro_SIZ1:SIZ1:GFP expressing siz1-2 (SSG1,8,12,14,15) and Pro_SIZ1:GUS:GFP expressing siz1-2 (SGG) plants. (e) The number of rosette leaves at the flowering of plants, as in (d), which were grown under SD. Data illustrated in (c) and (e) are means ± SE of 15–20 plants per analysis.

Figure2

siz1 partially suppresses late-flowering photoperiodic-pathway mutants. The flowering times of wild-type (Col-0), siz1-2, gi-2 siz1-2, gi-2, co-1 siz1-2, co-1, ft-1 siz1-2 and ft-1 plants were estimated under long days (LD). Data are means ± SE of 15–20 plants per analysis.

Figure3

Early flowering of siz1 under short days (SD) is mainly caused by elevated salicylic acid (SA) levels. (a) Wild-type (Col-0), siz1-2, nahG siz1-2 and nahG plants were grown under long days (LD). (b) Number of rosette leaves at flowering of wild-type, siz1-2, nahG siz1-2 and nahG plants. Plants were grown under LD or short days (SD). Data are means ± SE of 15–20 plants per analysis.

Figure4

SIZ1 represses SOC1 expression but activates MAF4 expression. (a) FLM/MAF1, SVP, MAF2, MAF3, MAF4, MAF5 and SOC1 mRNA levels in wild-type (Col-0) and siz1-2 plants were determined by RT-PCR. RNA was isolated from 14-day-old seedlings grown under long days (LD) or short days (SD). TUBULIN was used as a control for loading. (b, c) Flowering time of wild-type (Col-0), siz1-2, flc-3 siz1-2, flc-3, soc1-2 siz1-2 and soc1-2 plants were estimated under LD and SD. Data are means ± SE of 15–20 plants per analysis.

Figure5

siz1 partially suppresses the late flowering of FRI and ld-1, but not of fld-6. (a) The flowering times of wild-type (Col-0), siz1-2, ld-1 siz1-2, ld-1, fld-6 siz1-2 and fld-6 plants were analyzed under short days (SD). Stars indicate that flowering had not occurred after producing more than 100 rosette leaves in fld-6 and fld-6 siz1-2 plants. (b) Relative FLC mRNA levels were determined in 14-day-old SD-grown wild-type (Col-0), siz1-2, ld-1 siz1-2, ld-1, fld-6 siz1-2 and fld-6 seedlings by quantitative PCR. (c) Flowering times of wild-type (Col-0), siz1-2, FRI siz1-2 and FRI plants were estimated under long days (LD). (d) Relative FLC mRNA levels in wild-type (Col-0), siz1-2, FRI siz1-2 and FRI seedlings were determined by quantitative PCR. RNA was isolated from 10-day-old seedlings grown under long days (LD). Data illustrated in (a) and (c) are means ± SE of 15–20 plants per analysis.

Figure6

SIZ1 mediates SUMO modification of FLD. (a) HA:FLD or HA:FLDK3R (K287R, K693R and K770R) and T7:AtSUMO1 translational fusions were co-expressed in wild-type (Col-0) or siz1-2 protoplasts (Jin et al., 2001). T7:AtSUMO1 and HA:FLD or HA:FLDK3R were co-immunoprecipitated (IP) from extracts, and were then detected on the western blot (WB) with anti-HA. No IP is protoplast lysate before IP. The star indicates the position of sumoylated FLD proteins (right panel). HA:FLD and HA:FLDK3R transient expression levels were similar in siz1 and wild-type (Col-0) plants (left, No IP panel); vector, total protein extract from protoplasts transformed with the empty vector. (b) A WB with anti-T7 was used to determine the expression level of T7:AtSUMO1 in the No-IP samples from (a). Free T7:AtSUMO1 transient expression levels were similar in siz1 and wild-type (Col-0) plants (lower arrow). T7:AtSUMO1 conjugates (upper arrow) were nearly undetectable in siz1-2 (lanes 11 and 12) compared with wild-type (Col-0) (lanes 9 and 10) protoplasts. No IP and vector are as in (a).

Figure7

SIZ1-mediated SUMO modification of FLD represses deacetylation of histone H4 in FLC chromatin. (a) Relative FLC mRNA levels were determined by quantitative PCR in fld-6 protoplasts expressing vector, HA:FLD or HA:FLDK3R. Data are means ± SD (n = 4). (b) The acetylation state of H4 in FLC chromatin was assessed by chromatin immunoprecipitation (ChIP) analysis in fld-6 protoplasts expressing FLD or FLDK3R. Input is fld-6 chromatin before immunoprecipitation that was isolated from protoplasts transformed with empty vector; No AB and ACTIN are as in (b). The fold enrichment in H4 acetylation of fld-6 protoplasts expressing FLD or FLDK3R overexpressing the empty vector is shown. (c) HA:FLD and HA:FLDK3R were expressed equally in fld-6 protoplasts in this experiment.