The Flowering Times of lhy-11 cca1-1, lhy-1, and 35S:GI Plants.
The flowering time of lhy-11 cca1-1 (A) or lhy-1 and 35S:GI-A (B) plants, with or without gi-3, co-2, and ft-1, was measured in LDs (left-hand column for each genotype) or in SDs (right-hand column for each genotype). Flowering time was scored by counting the number of rosette (bottom box in each column) and cauline (top box in each column) leaves on the main stem. Mean leaf number is shown ± se. Each experiment was done at least twice with similar results.

Abundance of the mRNAs of Flowering Time and Circadian Clock–Regulated Genes in lhy-11 cca1-1 Plants Grown under SDs.
The expression of the GI (A), CO (B), FT (C), CCR2 (D), and FKF1 (E) genes was analyzed by RT-PCR in lhy-11 cca1-1, gi-3 lhy-11 cca1-1, gi-3, or Ler plants grown in SDs. Results are presented as a proportion of the highest value after standardization with respect to TUBULIN2 levels (TUB). Open and closed bars along the horizontal axis represent light and dark periods, respectively; these are measured in hours from dawn (zeitgeber time [ZT]). Each experiment was done at least twice with similar results.

Abundance of the mRNAs of Flowering-Time Genes in lhy-1 and 35S:GI Plants Grown under SDs.
The expression of flowering-time gene mRNAs GI (A), CO (B), FT (C), and FKF1 (D) was analyzed by RT-PCR in lhy-1, 35S:GI-B, lhy-1 35S:GI-B, and Ler plants grown in SDs. Results are presented as a proportion of the highest value after standardization with respect to TUB levels. Open and closed bars along the horizontal axis represent light and dark periods, respectively. These are measured in hours from dawn (ZT). Each experiment was done at least twice with similar results.

Circadian Clock–Regulated Gene Expression in 35S:GI and gi-3 Plants under SDs, LL, or DD.
(A) and (C) LHY and CCR2:LUC expression was analyzed in gi-3, 35S:GI-A, 35S:GI-B, or Ler plants. The expression of the LHY gene was analyzed by RNA gel blotting of RNA isolated from plants grown under SDs (8 h light/16 h dark) (A) or LL (C). Results are presented as a proportion of the highest value after standardization with respect to TUB levels. Numbers on the horizontal axis represent the time in hours after dawn (ZT) in SD (A) and after the start of the LL treatment (C). Open and closed boxes on the horizontal axis indicate light and dark, respectively (A), and subjective day and subjective night, respectively (C).
(B), (D), and (F) The expression of the CCR2 gene was followed by the luminescence of transgenic plants carrying the CCR2:LUC transgene and grown under SDs (8 h light/16 h dark) (B), LL (D), or DD (F). The results are presented as normalized luminescence. Data are the means ± se of the luminescence of ∼20 individual seedlings. Error bars are shown every fifth data point for clarity. Five independently transformed wild-type and mutant lines were analyzed under LL and DD with similar results, and under SDs two transformants were analyzed. Numbers on the horizontal axis represent the time in hours after dawn (ZT) in SD (B), after the start of the LL treatment (D), and in hours in darkness in DD treatment (F). Open and closed boxes on the horizontal axis indicate light and dark, respectively (B), subjective day and subjective night, respectively (D), and light and dark boxes on horizontal axis represent subjective day and subjective night, respectively (F).
(E) and (G) Plots showing the FFT-NLLS analysis of the CCR2:LUC data plotted in (D) and (F), respectively. A strong circadian expression of CCR2:LUC is reflected by the clustering of data points with low relative amplitude error values, which indicate robust rhythms. Scattered data points with relative amplitude error values closer to 1 indicate weaker rhythms. All plants in (D) were rhythmic, whereas in (F), more wild-type Ler seedlings were rhythmic (15/18) than gi-3 (12/19) and 35S:GI-B (11/20) seedlings.
Each experiment was done at least twice with similar results.

The Hypocotyl Length of lhy-11 cca1-1 and 35S:GI Plants under Different Intensities of Red Light and under SDs.
(A) and (B) Red light fluence response curves of the hypocotyl length of 35S:GI (A) or lhy-11 cca1-1 and gi-3 lhy-11 cca1-1 (B) seedlings. Hypocotyl length of seedlings grown under red light was measured and the results expressed as a percentage of the mean hypocotyl length of seedlings grown in DD. The mean value from three independent red light experiments was calculated as described in Methods and is presented ±se. On the x axis, light intensity is represented on a logarithmic scale.
(C) Hypocotyl length of 35S:GI seedlings and lhy-11 cca1-1 seedlings, with or without gi-3 grown under SDs (8 h light/16 h dark).

Cellular Localization of GI:GFP in Transgenic Arabidopsis.
Confocal microscope images of cells of 35S:GI:GFP transgenic plants. (A) to (F) illustrate the same hypocotyl epidermal cells and (G) to (I) the same stomatal guard cells. The composite images ([C] and [I]) show the GFP fluorescence channel ([A] and [G]) overlaid with the red ([B] and [H]) and transmission channels. (A) shows a strong green fluorescence in the chloroplast and the nucleus; however, this is not detected by emission fingerprinting of GFP (true GFP signal; [D]). The signal from the red channel (E), the true GFP signal (D), and of the background are overlaid in the composite image (F). In stomatal guard cells, strong green fluorescence was only detected in nuclei (G). These data indicate that in hypocotyl epidermal and stomatal guard cells, fluorescence of GI:GFP was only detected in nuclei. Bar = 10 μm.

Dual Role for GI in Regulating Circadian Rhythms and Flowering Time.
The central oscillator of the Arabidopsis circadian clock was proposed to consist of a negative feedback loop comprising LHY/CCA1 and TOC1 (Alabadi et al., 2002). Within this loop, TOC1 acts in the evening to promote expression of LHY/CCA1 in the morning, and LHY/CCA1 repress TOC1 expression. LHY and CCA1 are also shown as negative regulators of GI based on the earlier phase of GI expression detected in a lhy-11 cca1-1 double mutant (Mizoguchi et al., 2002); however, overexpression of CCA1 causes an increase in GI expression, which may suggest a more complex pattern of regulation (Fowler et al., 1999). GI may play a role in the evening related to that of TOC1 because it is also required for high amplitude expression of LHY/CCA1, is expressed in a similar phase as TOC1, and both gi mutations as well as 35S:GI have effects on circadian phase and period length. In the control of flowering time, GI increases the amplitude of CO and FT expression, which are both increased by 35S:GI and decreased by gi mutations. In addition, 35S:GI and gi mutations have opposite effects on flowering time. GI is therefore proposed to play dual roles acting within the circadian clock to regulate period length and circadian phase, while also more directly promoting expression of a circadian clock output pathway that includes CO and FT and promotes flowering. The effect of GI on flowering probably includes another pathway, indicated with an X, because co mutations only partially suppress the early flowering caused by 35S:GI or lhy cca1. FT activates SOC1 downstream of CO (Schmid et al., 2003; Michaels et al., 2005). In the diagram, the square illustrates the circadian oscillator that generates circadian rhythms, white illustrates daytime, and gray shading illustrates nighttime. The flowering pathway is one of many output pathways controlled by the circadian clock, and three other pathways expressed at different times of the day are illustrated. The genes shown in gray on the right-hand side of the figure are those that promote flowering in response to LDs and delay flowering when inactivated.