Multi-omic analysis shows REVEILLE clock genes are involved in carbohydrate metabolism and proteasome function

Abstract

Plants are able to sense changes in their light environments, such as the onset of day and night, as well as anticipate these changes in order to adapt and survive. Central to this ability is the plant circadian clock, a molecular circuit that precisely orchestrates plant cell processes over the course of a day. REVEILLE (RVE) proteins are recently discovered members of the plant circadian circuitry that activate the evening complex and PSEUDO-RESPONSE REGULATOR genes to maintain regular circadian oscillation. The RVE8 protein and its two homologs, RVE 4 and 6 in Arabidopsis (Arabidopsis thaliana), have been shown to limit the length of the circadian period, with rve 4 6 8 triple-knockout plants possessing an elongated period along with increased leaf surface area, biomass, cell size, and delayed flowering relative to wild-type Col-0 plants. Here, using a multi-omics approach consisting of phenomics, transcriptomics, proteomics, and metabolomics we draw new connections between RVE8-like proteins and a number of core plant cell processes. In particular, we reveal that loss of RVE8-like proteins results in altered carbohydrate, organic acid, and lipid metabolism, including a starch excess phenotype at dawn. We further demonstrate that rve 4 6 8 plants have lower levels of 20S proteasome subunits and possess significantly reduced proteasome activity, potentially explaining the increase in cell-size observed in RVE8-like mutants. Overall, this robust, multi-omic dataset provides substantial insight into the far-reaching impact RVE8-like proteins have on the diel plant cell environment.

Figure 1

Phenotypic analysis of rve 4 6 8 plant growth and development. A, Iodine starch staining of rosettes at 28-day post-imbibition. Rosettes were photographed at the same time and are part of a single image but were cropped to save space. Scale bars represent 1 cm. Error bars show standard deviation. B, Enzymatic quantification of starch levels in rosettes. All phenotyping was conducted under a 12-h:12-h LD photoperiod using 100 µmol m⁻² s⁻¹ light. Asterisks denote Bonferroni adjusted P-value significance: **P ≤ 0.005 and *P ≤ 0.05. Error bars show standard deviation. C, Starch related genes PGM1, GBSS1, and SEX4. All time points and genotypes were analyzed in biological triplicate (n = 3). Asterisks denote Student’s t test P-value significance: ***P ≤ 0.005, **P ≤ 0.01, and *P ≤ 0.05. Horizontal lines show the median. Primer pairs used in the NanoString analysis are described in Supplemental Table S1.

Figure 2

Diel mRNA gene expression analysis of circadian clock and photoreceptor related genes in rve 4 6 8 and Col-0 plants. Select genes from different circadian/diel implicated plant cell processes were examined for their expression level at EN (ZT0) and ED (ZT12) by NanoString expression analysis (see “Materials and methods”). A, Expression of circadian clock/RVE8-like related genes CCA1, LHY, TOC1, PRR5, ELF4, PIF4, and PIF5. B, Expression of photoreceptors PhyA, PhyB, CRY1, PHOT1, PHOT2, and ZTL. All time points and genotypes were analyzed in biological triplicate (n = 3). Asterisks denote Student’s t test P-value significance: ***P ≤ 0.005, **P ≤ 0.01, and *P ≤ 0.05. Primer pairs used in the NanoString analysis are described in Supplemental Table S1.

Figure 3

Analysis of the quantified proteome and phosphoproteome changes at ZT0 or ZT12 in rve 4 6 8 versus Col-0. The significantly changing (A) proteome and (B) phosphoproteome (n = 4; FC ≥ 1.5, q-value ≤ 0.05) were plotted by Venn diagrams and scatterplots to visualize the overlap and Log2FC between rve 4 6 8 and WT Col-0 plants at ZT0 and ZT12. GO enrichment of biological processes, subcellular localization and molecular function was then performed to systemically contextualize proteome and phosphoproteme changes (see “Materials and methods”).

Figure 4

Association network analysis of significant genotypic proteome fluctuations at ZT0 and ZT12. The association network depicts significant Log2FC in protein abundance between rve 4 6 8 and WT Col-0 at ZT0 and ZT12 (q-value ≤ 0.05; Supplemental Table S2). Networks were generated using Cytoscape and a combination of the STRING-DB and enhancedGraphics App (see “Materials and methods”) using all datatypes and an edge score ≥0.7. Nodes with no edges ≥0.7 were removed. Brighter red (rve 4 6 8) or blue (WT Col-0) coloration indicates the relative increase in measured Log2FC protein abundance in that corresponding genotype. Highlighted node clusters were manually annotated using a combination of ThaleMine (https://bar.utoronto.ca/thalemine/begin.do), TAIR (https://www.arabidopsis.org/), and Cytoscape Functional Annotation.

Figure 5

Association network analysis of significant genotypic phosphoproteome fluctuations at ZT0 and ZT12. The association network depicts significant median Log2FC phosphoproteome changes between rve 4 6 8 and WT Col-0 at ZT0 and ZT12 (q-value ≤ 0.05; Supplemental Table S3). Networks were generated using Cytoscape and a combination of the STRING-DB and enhancedGraphics App (see “Materials and methods”) using all datatypes and an edge score ≥0.7. Nodes with no edges ≥0.7 or only edges ≤0.6 were removed. Brighter red (rve 4 6 8) or blue (WT Col-0) coloration indicates the relative increase in measured Log2FC protein abundance in that corresponding genotype. Highlighted node clusters were manually annotated using a combination of ThaleMine (https://bar.utoronto.ca/thalemine/begin.do), TAIR (https://www.arabidopsis.org/), and Cytoscape Functional Annotation.

Figure 6

Relative Log2FC in diel metabolite levels. Rosette tissue (28 days postimbibition) was sampled at ZT0, ZT6, ZT12, and ZT18 as outlined in the “Materials and methods”. Cell values indicate the relative Log2FC (rve 468 vs. WT Col-0) in metabolite abundance.

Figure 7

Proteasome function is reduced in rve 4 6 8. A, Proteasome activity in relative fluorescence units per minute (RFU/min) measured in crude extracts of 3-week-old rve 4 6 8 and WT Col-0 plants. ****Bonferroni adjusted P < 0.0001; *Bonferroni adjusted P < 0.05). The center line of the box-plot shows the median, with box limits representing the 75th and 25th percentiles and whiskers plotted following Tukey’s 1.5× interquartile range method. B, Seedling growth experiments with proteasome inhibitor treatment (50 µM MG132). Seedlings were germinated, stratified, and then grown for 6 days under 12:12 LD conditions. A total of six replicate plates per condition were assayed (Additional plates depicted in Supplemental Figure S4).