Circadian clocks of taxonomic groups as diverse as plants, fungi and animals are composed of multiple interlocked feedback loops with positive and negative components [3] and many of the components of these clocks are encoded by members of gene families. The Arabidopsis circadian clock, an example of this common design principle, is composed of at least four interlocked feedback loops (Figure ​(Figure1).1). In the central loop (blue in Figure ​Figure1),1), TIMING OF CAB EXPRESSION 1 (TOC1), the founding member of a family of five PSEUDO-RESPONSE REGULATOR (PRR) genes, is a positive regulator of CIRCADIAN AND CLOCK ASSOCIATED 1 (CCA1) and LATE ELONGATED HYPOCOTYL (LHY). CCA1 and LHY are members of a small family of REVEILLE genes that encode single Myb domain transcription factors. Others members of this family have been shown to play roles in clock function as well as in regulation of clock output pathways [4,5]. To complete the first loop, CCA1 and LHY bind to the TOC1 promoter to inhibit its expression. In a second loop (green in Figure ​Figure1)1) within the central loop, CCA1 and LHY also repress expression of CCA1 HIKING EXPEDITION (CHE), which encodes a TCP transcription factor that binds to and represses expression from the CCA1 promoter [1]. In the third loop (yellow in Figure ​Figure1),1), termed the 'morning' loop based on the time of peak mRNA accumulation of its constituents, CCA1 and LHY are positive regulators of two TOC1 relatives, PRR7 and PRR9, that are negative regulators of CCA1 and LHY [4,5]. In a fourth loop (gray in Figure ​Figure1),1), termed the 'evening' loop based again on the time of peak mRNA accumulation, TOC1 represses a component, 'Y', that includes GIGANTEA (GI) and possibly PRR5. This component in turn positively regulates TOC1 expression at least in part through modulation by GI of proteasomal degradation of TOC1 mediated by the F-box protein ZEITLUPE (ZTL) [4,5]. In addition, proper regulation of CCA1 and LHY requires other clock genes, including EARLY FLOWERING 4 (ELF4), which encodes a protein of unknown function, and LUX ARRHYTHMO/PHYTOCLOCK1 (LUX/PCL), which encodes a Myb domain transcription factor; these and other clock components have yet to be fully incorporated into current clock models [4,5]. The number of interlocked feedback loops will undoubtedly increase as the regulatory relationships among clock components are more fully described.

Takata et al. [9] have conducted a parallel analysis of angiosperm CCA1/LHY genes, and their observations are consistent with those obtained in their analysis of the PRR genes; the common ancestor of monocots and eudicots had one CCA1/LHY gene and there has been independent duplication of the LHY/CCA1 genes in the monocots and eudicots. Within the eudicots, there has been independent duplication in poplar and Arabidopsis.

In Arabidopsis, the PRR3 gene offers an example of acquisition of a novel function. PRR9, PRR7, and PRR5 all have a similar role in negatively regulating CCA1 and LHY, suggesting that this represents the ancestral function (Figure ​(Figure2).2). PRR3 appears, instead, to have acquired a novel and specialized function in the vascular tissue, where PRR3 binds to TOC1 and, in doing so, blocks the interaction of TOC1 with ZTL, the F-box protein that targets TOC1 for proteasomal degradation [12]. Thus, PRR3 exhibits a restricted domain of expression and has acquired a novel function, the regulation of TOC1 stability through protein-protein interaction (Figure ​(Figure2).2). In Arabidopsis, loss of PRR3 function confers only a very small shortening of circadian period [13], which is consistent with the apparent loss of PRR3 in poplar, without concomitant perturbation of clock function.