Temporal repression of core circadian genes is mediated through EARLY FLOWERING 3 in Arabidopsis

Abstract

The circadian clock provides robust, ∼24 hr biological rhythms throughout the eukaryotes. The clock gene circuit in plants comprises interlocking transcriptional feedback loops, reviewed in [1], whereby the morning-expressed transcription factors CIRCADIAN CLOCK-ASSOCIATED 1 (CCA1) and LATE ELONGATED HYPOCOTYL (LHY) repress the expression of evening genes, notably TIMING OF CAB EXPRESSION 1 (TOC1). EARLY FLOWERING 3 (ELF3) has been implicated as a repressor of light signaling to the clock [2, 3] and, paradoxically, as an activator of the light-induced genes CCA1 and LHY [4, 5]. We use cca1-11 lhy-21 elf3-4 plants to separate the repressive function of ELF3 from its downstream targets CCA1 and LHY. We further demonstrate that ELF3 associates physically with the promoter of PSEUDO-RESPONSE REGULATOR 9 (PRR9), a repressor of CCA1 and LHY expression, in a time-dependent fashion. The repressive function of ELF3 is thus consistent with indirect activation of LHY and CCA1, in a double-negative connection via a direct ELF3 target, PRR9. This mechanism reconciles the functions of ELF3 in the clock network during the night and points to further effects of ELF3 during the day.

Figure 1

ELF3 Affects Clock Outputs and Clock Genes Hypocotyl measurements of 7-day-old seedlings are shown as an average hypocotyl length (Wassilewskija [Ws] n = 12, elf3-4 n = 18, cca1-11 lhy-21 n = 19, and cca1-11 lhy-21 elf3-4 n = 23), with the error being represented as a standard error of the mean (SEM) (A). Data are representative of two biologically independent experiments. qPCR measurements are shown of RNA levels for ELF3 in Ws wild-type plants (filled triangles) and cca1-11 lhy-21 double mutants (filled diamonds) (B), CCA1 (C) and LHY (D) in Ws (filled triangles), and elf3-4 mutants (crosses). Data are all normalized against IPP2 expression [25]. Graphs are an average of two to three biologically independent experiments, with normalized data being used to generate SEM error bars. Seedlings were grown in 12:12 white light:dark cycles and sampled every 2 hr from Zeitgeber time (ZT) = 0. ZT = 0 is defined as the time of lights-on. See also Figures S1 and S2.

Figure 2

ELF3 Regulates the Expression of Core Circadian Genes qPCR measurements of RNA levels for PRR9 (A), PRR7 (B), GI (C), and TOC1 (D) normalized against IPP2 and between replicates in Ws (filled triangles), elf3-4 (crosses), cca1-11 lhy-21 (filled diamonds), and cca1-11 lhy-21 elf3-4 (filled squares). Graphs are an average of three biologically independent experiments, each containing triplicate samples. Normalized data were used to generate SEM error bars. Seedlings were grown and sampled as in Figure 1. See also Figure S3.

Figure 3

ELF3 Binds In Vivo to the Promoter of PRR9 in the Early Night, but Not during the Day (A) Schematic of the PRR9 genomic region tested. The black bar indicates the specific region amplified from ChIP DNA by primer set P1. (B and C) Chromatin of 3-week-old plants was immunoprecipitated using either no antibody (−) or anti-GFP antibody (+). Resultant DNA extracted from 35S::GFP (B and C), 35S::ELF3::YFP (B), and ELF3::ELF3::YFP (C) plants was analyzed by qPCR. Each signal is expressed as a percentage of the signal in nonimmunoprecipitated DNA (input) extracted from the same tissue sample. Data represent the mean of at least six samples taken from three independent ChIP experiments. Error bars represent the SEM. Student's t test showed that only ELF3::ELF3::YFP had significantly different chromatin association between ZT = 6 and ZT = 14, marked with ^∗p < 0.05. See also Figure S4.

Figure 4

ELF3 Is Required for the Control of Circadian-Regulated Light Responses in GI and PRR9 Acute light induction of PRR9 (A) and GI (B) gene expression was measured by qPCR in Ws (black bars), elf3-4 (light gray bars), cca1-11 lhy-21 (white bars), and cca1-11 lhy-21 elf3-4 (dark gray bars). Seedlings were grown for 5 days under white-light 12:12 LD cycles and released into continuous dark from ZT = 12 on day 5. On day 6, samples were either treated with (+) or without (−) a white-light pulse (20 min, 80 μmol m⁻² s⁻¹) 1 hr before sampling on the predicted day at ZT = 30 (white background) and on the predicted night at ZT = 38 (gray background). Error bars indicate the SEM from 4–6 samples. Student's t test was used to compare treated and untreated samples within a time point and genotype. For clarity, only treated samples that differ significantly from their control are marked with ^∗p < 0.05 or ^∗∗p < 0.005.