Circadian control of carbohydrate availability for growth in Arabidopsis plants at night

Abstract

Plant growth is driven by photosynthetic carbon fixation during the day. Some photosynthate is accumulated, often as starch, to support nocturnal metabolism and growth at night. The rate of starch degradation in Arabidopsis leaves at night is essentially linear, and is such that almost all of the starch is used by dawn. We have investigated the timer that matches starch utilization to the duration of the night. The rate of degradation adjusted immediately and appropriately to an unexpected early onset of night. Starch was still degraded in an appropriate manner when the preceding light period was interrupted by a period of darkness. However, when Arabidopsis was grown in abnormal day lengths (28 h or 17 h) starch was exhausted approximately 24 h after the last dawn, irrespective of the actual dawn. A mutant lacking the LHY and CCA1 clock components exhausted its starch at the dawn anticipated by its fast-running circadian clock, rather than the actual dawn. Reduced growth of wild-type plants in 28-h days and lhy/cca1 mutants in 24-h days was attributable to the inappropriate rate of starch degradation and the consequent carbon starvation at the end of night. Thus, starch degradation is under circadian control to ensure that carbohydrate availability is maintained until the next anticipated dawn, and this control is necessary for maintaining plant productivity.

Fig. 1.

Starch degradation is adjusted to unexpected changes in the length of the light period. WT plants were grown under 24-h T-cycles for 4 weeks. Some plants were then subjected to an early night (white symbols); controls experienced the full 12-h light period (black symbols). Plants were analyzed for starch content (A) or transcript levels of sugar-repressed genes (B). Background shading: white and dark gray show the light period and night, respectively, striped indicates the premature night period (for plants transferred to an early night only), and light gray indicates an extended night period (imposed only in the photoperiod in which harvesting took place). (A) Starch content of whole rosettes. Error bars (SEM) were smaller than symbols for all data points (n = 6–8 individual rosettes). (B) Transcript levels of sugar-repressed genes At1g76410 and At3g59940 were analyzed by qRT-PCR. Error bars are SD: where not visible, they are smaller than the symbol (n = 4 samples, each from a different plant). The graph shows the last 8 h of an experiment identical to that in A.

Fig. 2.

Skeleton photoperiods do not disrupt the normal pattern of starch degradation. Plants were grown from germination under skeleton photoperiods (black symbols) or 24-h T-cycles (white symbols). Plants were harvested when 3 weeks old over a full light/dark cycle. Background shading: white and dark gray show the light period and night, respectively, and light gray indicates an extended night period (imposed only in the photoperiod in which harvesting took place). Error bars (SEM) were smaller than symbols for all data points (n = 6–8 individual rosettes). (A) Starch content of whole rosettes under skeleton light periods. (B) Starch content during the night as a proportion of the amount of starch present at the end of day.

Fig. 3.

Starch degradation is regulated according to the subjective morning as anticipated by the circadian clock. WT plants (Ws) were grown for 23 days under 28-h T-cycles (A) or 17-h T-cycles (B). Starch content was determined and transcript levels of clock-related genes LHY (squares), GBS1 (diamonds), and two sugar-repressed genes At1g76410 and At3g59940 were analyzed by qRT-PCR. Background colors: dark gray, actual night period; light gray, extended night period (imposed only in the photoperiod in which harvesting took place). For starch, n = 6–8 individual rosettes; the error bars (SEM) were smaller than symbols for all data points. For transcript levels, n = 4 samples, each from a different plant; error bars are SD: where not visible they are smaller than the symbol.

Fig. 4.

The circadian clock mutant cca1/lhy shows an abnormal pattern of starch degradation. WT (Ws: black symbols) and cca1/lhy mutant plants (white symbols) were grown for 23 days under 24-h (A) or 17-h T-cycles (B). Starch content was determined and transcript levels of GBS1 and two sugar-repressed genes At1g76410 and At3g59940 were analyzed by qRT-PCR. Background colors and replication are as in Fig. 3. Where error bars are not visible, they are smaller than the symbol.

Fig. 5.

Growth reduction under 28-h T-cycles is related to depletion of starch reserves rather than rates of CO₂ assimilation. Bars in A to E represent WT plants (black bars) and sex4-3 (gray bars) and lsf1-1 (white bars) mutant plants, grown for 4 weeks on soil under 24-h (T24) or 28-h T-cycles (T28). (A) Fresh weight of whole rosettes (n = 20 individual rosettes; error bars are SEM). (B) Total chlorophyll content of mature leaves (n = 6 independent samples; error bars are SEM). (C) Net CO₂ assimilation of Col WT plants (n = 8 individual rosettes; error bars are SEM). The values for T24 and T28 are statistically significantly different (Student's t-test, P = 0.005). (D) Starch content at the end of night (eon) and after 10 h of darkness (10 h D) under 28-h T-cycles (n = 6–8 individual rosettes; error bars are SEM). (E) Transcript level of the sugar-repressed gene At1g76410 (n = 5 samples, each from a different plant, error bars are SD). (F) Fresh weight of whole rosettes. WT (Col) were grown for 3 weeks on MS plates with (+SUC) or without (−SUC) sucrose (30 mM) under 24-h T-cycles (black bars) or 28-h T-cycles (gray bars). n = 12 individual rosettes; error bars are SEM.