Are there multiple circadian clocks in plants?

Abstract

We have reported that Arabidopsis might have genetically distinct circadian oscillators in multiple cell-types. Rhythms of CHLOROPHYLL A/B BINDING PROTEIN2 (CAB2) promoter activity are 2.5 h longer in phytochromeB mutants in constant red light and in cryptocrome1 cry2 double mutant (hy4-1 fha-1) in constant blue light than the wild-type. However, we found that cytosolic free Ca²⁺ ([Ca²⁺]_cyt) oscillations were undetectable in these mutants in the same light conditions1. Furthermore, mutants of CIRCADIAN CLOCK ASSOCIATED1 (CCA1) have short period rhythms of leaf movement but have arrhythmic [Ca²⁺]_cyt oscillations. More important, the timing of cab1-1 (toc1-1) mutant has short period rhythms of CAB2 promoter activity (~21 h) but, surprisingly, has a wild-type period for circadian [Ca²⁺]_cyt oscillations (~24 h). In contrast, toc1-2, a TOC1 loss-of-function mutant, has a short period of both CAB2 and [Ca²⁺]_cyt rhythms (~21h). Here we discuss the difference between the phenotypes of toc1-1 and toc1-2 and how rhythms of CAB2 promoter activity and circadian [Ca²⁺]_cyt oscillations might be regulated differently.

The plant circadian clock controls a multitude of physiological processes such as photosynthesis, organ and stomatal movements and transition to reproductive growth. A plant clock that is correctly matched to the rhythms in the environment brings about a photosynthetic advantage that results in more chlorophyll, more carbon assimilation and faster growth.3 One of the first circadian clock mutants to be described in plants was the short period timing of cab1-1 (toc1-1), which was identified using the rhythms of luciferase under a CHLOROPHYLL A/B BINDING PROTEIN2 (CAB2) promoter as a marker for circadian period.4

Circadian rhythms of both CAB2 promoter activity and cytosolic-free Ca²⁺ ([Ca²⁺]_cyt) oscillations depend on the function of a TOC1, CIRCADIAN CLOCK ASSOCIATED1 and LATE ELONGATED HYPOCOTYL (TOC1/CCA1/LHY) negative feedback loop.5 In tobacco seedlings, CAB2:luciferase (CAB2:luc) rhythms and circadian [Ca²⁺]_cyt oscillations can be uncoupled in undifferentiated calli.6 In Arabidopsis, we reported that toc1-1 has different periods of rhythms of CAB2 promoter activity (∼21 h) and circadian [Ca²⁺]_cyt oscillations (∼24 h). The mutant allele toc1-1 has a base pair change that leads to a full protein that has an amino acid change from Ala to Val in the CCT domain (CONSTANS, CONSTANS-LIKE and TOC1).7 On the other hand, the mutant toc1-2 has short period of both rhythms of CAB2 promoter activity and circadian [Ca²⁺]_cyt oscillations (∼21 h).1,7 This allele has a base pair change that results in changes to preferential mRNA splicing, resulting in a truncated protein with only 59 residues.7 Thus, the mutated CCT domain in toc1-1 might lead to the uncoupling of rhythms of CAB2 promoter activity and circadian [Ca²⁺]_cyt oscillations while the absence of TOC1 in toc1-2 causes the shortening of the period of both rhythms. Indeed, zeitlupe-1 (ztl-1) mutants, that have higher levels of TOC1, have long periods of both rhythms of CAB2 promoter activity and circadian [Ca²⁺]_cyt oscillations.1 The biochemical function of the CCT domain is unknown but it is predicted to play an important role in protein-protein interactions8 and nuclear localization.9

One model to explain the period difference of CAB2:luc expression and circadian [Ca²⁺]_cyt oscillation is that the toc1-1 mutation has uncoupled two oscillators in the same cell. Uncoupled oscillators are a predicted outcome of certain mutations in the recently described three-loop mathematical model.10–11 However, both rhythms of TOC1 and CCA1/LHY expression, which would be in uncoupled oscillators accordingly to the model, are described as short-period in toc1-1.5 Thus, we have favored the model in which CAB2:luc expression and circadian [Ca²⁺]_cyt oscillation are reporting cell-types with different oscillators that are affected differently by toc1-1.

It is possible that TOC1 could interact with a family of cell-type specific proteins. The interaction of TOC1 with each member of the family could be affected differently by the mutation in the CCT domain (Fig. 1). Two-hybrid assays have shown that TOC1 interacts with PIF proteins (PHYTOCHROME INTERACTING FACTOR3 and PIF4) and related PIL proteins (PIF3-LIKE PROTEIN 1, PIL2, PIL5 and PIL6).8 In fact, TOC1 interaction with both PIF3 and PIL1 is stronger when the N-terminus receiver domain is taken out and the CCT domain is left intact.8 Thus, it is possible that TOC1 and different PIF/PIL proteins interact to regulate the central oscillator. This interaction could be impaired by the Ala to Val change in the toc1-1 mutation, leading to the period shortening. However, lines misexpressing PIF3, PIL1 and PIL6 showed no changes in their circadian rhythms.12–16

One possible explanation for the absence of alterations in the period of circadian rhythms in lines misexpressing PIF/PIL is that they only have roles in certain cell-types. As an example, PIL6 and PIF3 are involved with flowering time and hypocotyl growth in red light12–15 while PIL1 and PIL2 are involved with hypocotyl elongation in shade-avoidance responses.16 Both hypocotyl growth and flowering time require cell-type specific regulation: vascular bundle cells in the case of the flowering time17 and the cells in the shoot in the case of the hypocotyl elongation.16 If TOC1 interaction with certain PIF/PIL is indeed cell-type specific, the mutated CCT domain found in the toc1-1 mutant could affect the clock in different ways, depending on the type of PIF/PIL protein expressed in each cell-type. Therefore, a question that arises is: which cell-types are sensitive to the toc1-1 mutation?

There is evidence that CAB2 and CATALASE3 (CAT3) are regulated by two oscillators that respond differently to temperature signals.18 These genes might be regulated by two distinct circadian oscillators within the same tissues or a single cell.18 Interestingly, the spatial patterns of expression of CAB2 and CATALASE3 overlap in the mesophyll of the cotyledons.18 Furthermore, rhythms of CAB2 and CHALCONE SYNTHASE (CHS) promoter activity have different periods and they are equally affected by toc1-1 mutation.19 Whereas CAB2 is mainly expressed in the mesophyll cells, CHS is mainly expressed in epidermis and root cells.19 However, rhythms of AEQUORIN luminescence, which reports [Ca²⁺]_cyt oscillation, were insensitive to toc1-1 mutation and appear to come from the whole cotyledon.20 One cell-type which is found in the whole cotyledon but is distinct from either mesophyll or epidermis cells is the vascular tissue and associated cells.

Another approach to determine which cell-types are insensitive to toc1-1 mutation is to compare the toc1-1 and toc1-2 phenotypes. The period of circadian [Ca²⁺]_cyt oscillations is not the only phenotype that is different in toc1-1 and toc1-2 mutants. Rhythms in CAB2 promoter activity in constant red light are short period in toc1-1 but arrhythmic in toc1-2.21,22 COLD, CIRCADIAN RHYTHM AND RNA BINDING 2/GLYCINE-RICH RNA BINDING PROTEIN 7 (CCR2/GRP7) is also arrhythmic in toc1-2 but short period in toc1-1 in constant darkness.7,22 When the length of the hypocotyl was measured for both toc1-1 and toc1-2 plants exposed to various intensities of red light, only toc1-2 had a clear reduction in sensitivity to red light. Therefore, toc1-2 has long hypocotyl when maintained in constant red light while hypocotyl length in toc1-1 is nearly identical to that in the wild-type.22 These differences may allow us to separate which cell-types are sensitive to the toc1-1 mutation and which not.

Hypocotyl growth is regulated by a large number of factors such as light, gravity, auxin, cytokinins, ethylene, gibberellins and brassinosteroids.23 There is also a correlation between the size of the hypocotyl in red light and defects in the circadian signaling network.24,25 The fact that toc1-1 has different hypocotyl sizes from toc1-2 suggests that circadian [Ca²⁺]_cyt oscillations could be involved in the light-dependent control of hypocotyl growth. Circadian [Ca²⁺]_cyt oscillations might encode temporal information to control cell expansion and hypocotyl growth.26–28 toc1-1 have short-period rhythms of hypocotyl elongation, which indicates that the cells in the hypocotyl have a 21 h oscillator.29 However, toc1-1 might also have a wild-type hypocotyl length in continuous red light because cells which generate the signal to regulate hypocotyl growth might have 24 h oscillators.

The toc1-1 mutation was the first to be directly associated with the plant circadian clock, revitalizing the field of study.4 Now, by either uncoupling two feedback loops or by distinct TOC1 protein-protein interaction in different cell-types, toc1-1 has shown new properties of the circadian clock that may deepen our understanding of this system.

Figures and Tables

Figure 1 Models of how the toc1-1 mutation might differently affect cell-type specific circadian oscillators. The single mutant toc1-1 have 21 h rhythms of CAB2 promoter activity and 24 h-rhythms of [Ca²⁺]_cyt oscillations. The toc1-1 mutation is a single amino acid change in the CCT domain. The CCT domain is involved in protein-protein interaction and/or nuclear localization. We have proposed that circadian oscillators with different periods are present in different cell-types. The luminescence generated by CAB2 promoter-drived luciferase (from the CAB2:luc) is probably originated in the epidermis and mesophyll cells. In this model, we propose that the mutation on the CCT domain impairs the mutated TOC1 interaction with the hypothetical protein Z in these cells-types. In contrast, in other cell-types, the mutated TOC1 still interacts with other hypothetical proteins (W), despite the mutation in the CCT domain. In those cell-types, the circadian oscillator could still run with a 24 h period for [Ca²⁺]_cyt rhythms (from the 35S:AEQ construct). One possible identity for Z and W are the members of the PHYTOCHROME INTERACTING FACTOR (PIF) related PIF3-LIKE (PIL) family.

Acknowledgements

This research was funded by the USA National Institute of Mental Health (R01 MH43836) to C.H.J. and the BBSRC UK to A.A.R.W., who is also grateful to the Royal Society of London for the award of a University Research Fellowship. CTH is supported by a CAPES, Brazil Scholarship.

Addendum to: