White Collar-1, a DNA Binding Transcription Factor and a Light Sensor

See allHide authors and affiliations

Science  02 Aug 2002:
Vol. 297, Issue 5582, pp. 840-843
DOI: 10.1126/science.1072795


Blue light regulates many physiological processes in fungi, but their photoreceptors are not known. In Neurospora crassa, all light responses depend on the Per-Arnt-Sim (PAS) domain–containing transcription factor white collar–1 (wc-1). By removing the WC-1 light, oxygen, or voltage domain, a specialized PAS domain that binds flavin mononucleotide in plant phototropins, we show that light responses are abolished, including light entrainment of the circadian clock. However, the WC-1–mediated dark activation of frq remains normal in this mutant, and the circadian clock can be entrained by temperature. Furthermore, we demonstrate that the purified NeurosporaWC-1–WC-2 protein complex is associated with stoichiometric amounts of the chromophore flavin-adenine dinucleotide. Together, these observations suggest that WC-1 is the blue-light photoreceptor for the circadian clock and other light responses in Neurospora.

Blue and near-ultraviolet light are known to regulate many physiological processes in a large number of organisms, including circadian clock functions from fungi to mammals. Currently, phototropins, cryptochromes, and a flavin-adenine dinucleotide (FAD)–containing adenylyl cyclase inEuglena are the three known types of eukaryotic blue-light photoreceptors, all of which are flavin-containing proteins (1–3). Most fungal photoresponses, ranging from growth responses to phototropism to carotenoid induction to circadian clock entrainment, are mediated by blue light (4). For Neurospora crassa, previous evidence has suggested that a flavin-containing blue-light photoreceptor is responsible for mediating all known light responses [reviewed in (4)].

The white collar–1 (wc-1) and wc-2genes are required for all aspects of the known Neurosporablue-light responses, including induction of the circadian clock gene frequency (frq), which has been shown to mediate the light entrainment of the circadian clock (4–10). WC-1 and WC-2 are transcriptional activators that can bind to promoter elements upstream of light-inducible genes via their GATA-type zinc-finger DNA binding domains (5, 6, 11). Both proteins also contain Per-Arnt-Sim (PAS) domains, which are important for the in vivo formation of a nuclear complex between WC-1 and WC-2 (WCC) (9, 12). In addition to their critical roles in light responses, WC-1 and WC-2 are also essential components of theNeurospora frq-wc–based circadian feedback loops because they activate the transcription of frq in the dark (7, 13–16).

Sequence analysis of the WC-1 protein revealed that one of its three PAS domains (Fig. 1A) belongs to a specialized class of these domains known as a light, oxygen, or voltage (LOV) domain. These LOV domains have been best characterized in plant phototropins, in which each LOV domain binds a flavin mononucleotide (FMN) molecule and is able to undergo fully reversible photocycles (1, 17–19). The crystal structure of a LOV-FMN complex revealed 11 residues in the vicinity of FMN that are highly conserved in LOV domains (20), including that of WC-1, suggesting that WC-1 may bind FMN and function as a blue-light photoreceptor.

Figure 1

The WC-1 LOV domain is required for light responses, including the light entrainment of the circadian clock. (A) The domain architecture of WC-1. NLS, nuclear localization signal; ZN, zinc-finger DNA binding domain; aa, amino acid. (B) Western blot analysis showing the amounts of WC-1 and FRQ. wt, wild-type; wc-1 , awc-1 null strain. (C) Northern blot analysis showing the amounts of frq and al-3 mRNA. LP indicates that cultures were given 15 min of light treatment at the 24th hour of complete darkness. The middle panel is a longer exposure of the top panel. (D) Western blot analysis of WC-1 for cultures with (+) or without (–) 15 min of light pulse. WC-1p indicates the hyperphosphorylated WC-1 forms. (E and F) Race tube analysis in constant dark after a single LD transition (E) or LD cycles (F). wc1-2, thewc-1 strain rescued by a wild-typewc-1 gene.

We made a construct in which the LOV domain was deleted from the wild-type wc-1 gene (Fig. 1A) and introduced it into awc-1 null strain (wc-1.lov) (21). The apparent molecular weight of mutant WC-1 was smaller than that of wild-type WC-1, and mutant WC-1 was present in a lower amount than wild-type WC-1 (Fig. 1B). The amount of FRQ in the wc-1.lov mutant was near normal in 22 hours of constant darkness (DD22) but did not increase in cultures grown in constant light (LL), indicating that the light induction of frq is lost in the mutant. This result was confirmed with Northern blot analyses of frq andal-3 (a gene required for carotenoid biosynthesis) RNA amounts (Fig. 1C). As for FRQ in the dark, the amounts offrq mRNA were comparable in both strains (Fig. 1C, middle panel). These data indicate that the WC-1 LOV domain is essential for light-activated transcription of frq but not frqtranscription activation in the dark.

A wc-2 mutant strain (234w) that lacks most of the COOH-terminal part of the protein, including the DNA binding domain, is defective in most aspects of the light response (6,9). However, the light-induced hyperphosphorylation of WC-1 remained unaffected in this mutant (22), showing that WC-1 hyperphosphorylation is independent of transcriptional activation, that WC-2 acts downstream of WC-1, and that WC-2 is not the photoreceptor that directly receives the light signal.

Could WC-1 be the photoreceptor? Several “blind” phenotypes of the wc-1.lov mutant are consistent with this hypothesis: (i) there was no WC-1 hyperphosphorylation observed in the wc-1.lov mutant after 15 min of light (Fig. 1D); (ii) the circadian conidiation rhythm could not be synchronized in the wc-1.lov mutant grown under light-dark (LD) cycles (Fig. 1F); (iii) the wc-1.lov mutant was arrhythmic in constant darkness after either a single LD transition or LD cycles (Fig. 1, E and F), suggesting that the clock was not running under these conditions; and (iv) in contrast to the wild-type strain, the amounts of both frq RNA and FRQ protein in the wc-1.lov mutant were unaffected by the LD transition, and the amounts offrq, FRQ, and (ccg-1) stayed relatively constant in LL and in constant darkness (DD) with no circadian fluctuations (fig. S1, A and B).

The PAS domain of WC-2 is involved in WC-1–WC-2 and WCC-FRQ protein-protein interactions and the maintenance of the steady-state level of WC-1 (9). However, both the wild-type and mutant WC-1 proteins could form complexes with WC-2 and FRQ under LL conditions, suggesting that the circadian oscillator is intact in the mutant (Fig. 2). Together, these data demonstrate that in the wc-1.lov mutant, light responses are abolished and the circadian clock cannot be entrained by light, despite the near-normal activation of frq in the dark and the maintenance of WC-2 and FRQ interactions by the mutant WC-1.

Figure 2

Immunoprecipitation assay with WC-2 antiserum. PI, wild-type extract immunoprecipitated with the WC-2 preimmune serum; IP, immunoprecipitation.

If the circadian clock in the wc-1.lov mutant was functional but just blind, then it should be entrained by other nonphotic cues. To test whether temperature could entrain the wc-1.lov mutant as it does for normal Neurospora (23), we grew cultures with temperature cycles (12 hours at 20°C, then 12 hours at 25°C) in DD for 3 days and then switched them to a constant temperature of 25°C (Fig. 3A). After transfer to constant temperature, the wc-1 (the WGlnull) strain became arrhythmic; and the wild-type strain, thewc-1 strain rescued by a wild-typewc-1 gene (wc1-2), and the wc-1.lov strain exhibited a circadian conidiation rhythm, indicating that the clock can be reset by the temperature treatment. The amplitudes of the conidiation rhythm in the wc-1.lov mutant were lower than those of the wild-type, and the conidiation rhythm usually damped out after 3 to 4 days in constant darkness, suggesting that the lack of the LOV domain partially affected its function in the dark. Similar experiments were performed in LL (Fig. 3B). After transfer to constant temperature, the wild-type strain immediately became arrhythmic because of the suppression of the clock function by light, whereas the rhythm of the wc-1.lov strain persisted. These results, plus experiments on the circadian fluctuations of FRQ protein and ccg-1 mRNA after the temperature treatment (fig. S2, A and B), demonstrate that the functions of WC-1 in light responses and in circadian feedback loops can be effectively separated by removing the LOV domain of WC-1. Furthermore, although the resetting of the clock requires changes in FRQ level, the signaling mechanisms of light and temperature entrainment are different, and the temperature entrainment pathway does not require the function of the WC proteins, as compared with the temperature and light entrainment pathways inDrosophila (24).

Figure 3

Circadian rhythms in the wc-1.lov strain could be entrained by temperature. (A andB) The race tube cultures were entrained by three 20°–25°C temperature cycles before they were kept at a constant temperature of 25°C in either DD (A) or LL (B).

To further test whether WC-1 is the photoreceptor that directly receives the light signal, we sought in vivo evidence to show that WC-1 is associated with a flavin chromophore. Using Neurosporaprotein extracts from a strain containing epitope-tagged WC-2 (Myc–His–WC-2) (Fig. 4A), the WCC was purified (21). The final purification product (immunoprecipitation with c-Myc monoclonal antibody–coupled beads) contained four protein bands, including the c-Myc monoclonal immunoglobulin G band (Fig. 4B); the top two bands are WC-1 and the tagged WC-2, respectively, and the minor band at 60 kD is a contaminating band because it could also be immunoprecipitated with hemagglutinin antibody–coupled beads (25). No other protein was found to tightly bind WCC. The relative molar concentration of WC-1 and WC-2 was about 1:1, consistent with the 200-kD dimeric WCC (16).

Figure 4

WCC purified from Neurospora is associated with FAD. (A) Race tube analysis in DD of the Myc–His–WC-2 strain after a LD transition. (B) Silver staining of SDS–polyacrylamide gel electrophoresis showing the purified Neurospora WCC. B, the protein extract before the final immunoprecipitation; asterisk, a contaminating band. (Cthrough E) Fluorescence excitation (C) and emission [(D) and (E)] spectra analyses.

To detect the existence of chromophore in WCC, fluorescence spectroscopic analyses were performed. As shown in Fig. 4, C and D, WCC contained a noncovalently bound fluorescent cofactor, with excitation and emission spectra very similar to those of free FMN and FAD standards. Like both of these, the WCC cofactor exhibited two excitation peaks at 370 and 450 nm and an emission peak at 520 nm, all of which are characteristic of flavins. The emission spectra of the WCC cofactor showed a substantial fluorescence increase at an acidic pH, indicating that it is FAD (Fig. 4E) (26). This identification was confirmed by thin-layer chromatography: The WCC fluorescent cofactor and FAD standard had identical retention factor values (0.08). In addition, comparison of the concentrations of WC-1 and FAD showed that FAD bound WC-1 stoichiometrically, with a molar ratio of about 1:1. Together, these observations suggest that FAD is the chromophore that binds to WC-1 and confirm previous physiological evidence suggesting that the Neurosporaphototransduction mechanism is mediated by a flavin species.

A notable difference between the LOV domains of WC-1 and phototropins is a large extension in the loop connecting the α′A and αC helices in the WC-1 LOV domain (fig. S3A). This extension may be important for accommodating the larger FAD molecule as compared with FMN (21). Similar extensions were also found in the LOV domains of several other proteins, including the Neurosporaprotein VIVID (VVD), a repressor of light-regulated processes (27), and ZTL and FKF1, the two closely relatedArabidopsis proteins that regulate the timing of floral development and circadian periods (28, 29). Phylogenetic analysis showed that the ztl andfkf1 LOV domains are more closely related to those of WC-1 and VVD than to those of the plant phototropins (fig. S3B). Therefore, we suggest that these three proteins may also bind FAD to accomplish their roles in the light responses and circadian clock.

In this study, we have presented physiological, molecular, and biochemical data which strongly suggest that WC-1 is the blue-light photoreceptor for the circadian clock and other light responses inNeurospora. Thus, this fungal photoreceptor functions as both a DNA binding transcription factor and a light sensor. The nuclear localization of WC-1 and its ability to sense light and (together with WC-2) to bind to the promoters of light-inducible genes (11) ensure that the light signal is directly targeted to gene promoters to regulate transcription.

Supporting Online Material

Materials and Methods

Figs. S1 to S3

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: Yi.Liu{at}


View Abstract

Stay Connected to Science

Navigate This Article