Quantitative analysis of regulatory flexibility under changing environmental conditions

Abstract

The circadian clock controls 24-h rhythms in many biological processes, allowing appropriate timing of biological rhythms relative to dawn and dusk. Known clock circuits include multiple, interlocked feedback loops. Theory suggested that multiple loops contribute the flexibility for molecular rhythms to track multiple phases of the external cycle. Clear dawn- and dusk-tracking rhythms illustrate the flexibility of timing in Ipomoea nil. Molecular clock components in Arabidopsis thaliana showed complex, photoperiod-dependent regulation, which was analysed by comparison with three contrasting models. A simple, quantitative measure, Dusk Sensitivity, was introduced to compare the behaviour of clock models with varying loop complexity. Evening-expressed clock genes showed photoperiod-dependent dusk sensitivity, as predicted by the three-loop model, whereas the one- and two-loop models tracked dawn and dusk, respectively. Output genes for starch degradation achieved dusk-tracking expression through light regulation, rather than a dusk-tracking rhythm. Model analysis predicted which biochemical processes could be manipulated to extend dusk tracking. Our results reveal how an operating principle of biological regulators applies specifically to the plant circadian clock.

Figure 1

Dawn- and dusk-dominant rhythms show flexible timing in Ipomoea nil. Peak times are shown for rhythms of LHCB expression (filled symbols), transpiration rate (shaded symbols) and maximum inhibition of flowering by a red light pulse (NBmax, open symbols), measured in darkness after different light intervals in I. nil. Shaded area of plot, darkness; open area, light.

Figure 2

Arabidopsis clock gene expression changes with photoperiod. Transcript abundance measured with 2 h time resolution by Q-PCR relative to an ACT2 standard, for clock genes CCA1 (A, B), GI (C, D) and TOC1 (E, F, I, J) after entrainment to 24-h light:dark cycles (LD), including a photoperiod of 3 h (red), 6 h (orange), 9 h (yellow/black), 12 h (green) or 18 h (blue). Samples were taken during one diurnal cycle and after release into constant light (LL; A, C, E) or darkness (DD; B, D, F). Time-points 0–22 h are identical for LL and DD. Time stamps below (H, J) apply to all panels. Error bars represent the range of biological duplicates (A–F) or the SE of triplicates (I, J). Light conditions for three photoperiods are shown (open bar, light interval; shaded bar, darkness, with colours orange for 6 h photoperiod, green for 12 h, blue for 18 h). Simulations of LHY (G) and TOC1 (H) RNA levels in the interlocking-loop model illustrate the large phase changes predicted by a dusk-responsive model under this range of photoperiods (3 h, red; 6 h, orange; 9 h, broader yellow; 12 h, green; 15 h, blue). Arrowheads in (G, H) highlight the 3-h phase shift between 3- and 9-h photoperiods. Time series for TOC1 expression in the 3-h (I) and 9-h (J) photoperiod followed by DD are shown with 1 h time resolution at the peaks, together with equivalent data replotted from (F). Arrowheads in (I, J) mark the complex peak waveform observed in the samples with higher time resolution. Source data is available for this figure at www.nature.com/msb.

Figure 3

Predicted and experimentally measured entrainment patterns in the clock of Arabidopsis. The one-loop (A, B), two-loop (C, D) and three-loop (E, F) models of the Arabidopsis clock were analysed to calculate the dusk sensitivity measure (A, C, E) for the peak (upward triangle) and trough (downward triangle) times of mRNA (m) and bulk protein (P) variables of all the genes in the model under a 12-h photoperiod. Dusk sensitivity close to 1 indicates dusk-dominant entrainment; close to 0, dawn-dominant entrainment. Where an expression profile has multiple peaks or troughs, its dusk sensitivity is plotted from left to right, in chronological order after dawn, with the convention that in each peak/trough pair the trough follows the peak in time. The models were solved numerically under a range of simulated photoperiods, resulting in the simulated RNA profiles plotted in Supplementary Figure 3. Times of the peak abundance for each simulated RNA during light:dark cycles are shown (B, D, F; see inset key for gene identity). For comparison, the peak expression time for six clock genes was measured in individual seedlings using mFourfit analysis of in vivo imaging data (Supplementary Figures 5 and 7) from transgenic plants carrying LUC reporter fusions under the same range of photoperiods (G), or following transfer to LL (H) or DD (I; see the inset key for gene identity). PRR9 is absent from I, because very low expression levels in DD prevented phase estimation. Shaded chart areas represent darkness. Error bars indicate the standard error of the mean, calculated as described in the Supplementary information. Two-way ANOVA on LD, LL and DD peak times showed a highly significant interaction between gene and photoperiod in each case (P≪0.001), indicating that the genes responded to photoperiod in significantly different patterns. Source data is available for this figure at www.nature.com/msb.

Figure 4

Dawn- and dusk-dominant regulation of Arabidopsis gene expression. RNA abundance of Arabidopsis clock genes and clock-controlled genes were measured by Q-PCR data under light:dark cycles with photoperiods of 6, 12 and 18 h (see Figure 2, Supplementary Figure 8). Times of the half-maximum rising (triangles) and falling (diamonds) phases are shown in (A) for CCA1, GI and FKF1, and in (B) for CO, SEX4 and DPE2. See inset keys for gene identification. Shaded areas of plots represent darkness, and open areas, light. Source data is available for this figure at www.nature.com/msb.

Figure 5

Circadian clocks show flexible regulation in light–dark cycles. (A) Ipomoea nil shows a dusk-tracking rhythm in the peak sensitivity of flowering to a night-breaking light pulse (NBmax), but dawn-dominant peaks in the rhythms of transpiration and LHCB gene expression. (B) The three-loop model of the clock in Arabidopsis predicts that the peak times of clock mRNAs will not track dusk, except for Y in 6–9-h photoperiod cycles. (C) Luciferase reporter genes in wild-type Arabidopsis plants show peak expression consistent with predictions of the three-loop model under a range of photoperiods. Clock components marked by colours (see legend), with component names in the model above those in the data; the model lacks CCR2. Time runs up each panel; light–dark cycles with different light intervals are denoted by the photoperiod (h) and by the shading, which indicates the time of darkness.