Light-independent Phosphorylation of WHITE COLLAR-1 Regulates Its Function in the Neurospora Circadian Negative Feedback Loop*

Phosphorylation is a major regulatory mechanism controlling circadian clocks. In the Neurospora circadian clock, the PER-ARNT-SIM (PAS) domain-containing transcription factor, WHITE COLLAR (WC)-1, acts both as the blue light photoreceptor of the clock and as a positive element in the circadian negative feedback loop in constant darkness, by activating the transcription of the frequency (frq) gene. To understand the role of WC-1 phosphorylation, five in vivo WC-1 phosphorylation sites, located immediately downstream of the WC-1 zinc finger DNA binding domain, were identified by tandem mass spectrometry using biochemically purified endogenous WC-1 protein. Mutations of these phosphorylation sites suggest that they are major WC-1 phosphorylation sites under constant conditions but are not responsible for the light-induced hyperphosphorylation of WC-1. Although phosphorylation of these sites does not affect the light function of WC-1, strains carrying mutations of these sites show short period, low amplitude, or arrhythmic conidiation rhythms in constant darkness. Furthermore, normal or slightly higher levels of frq mRNA and FRQ proteins were observed in a mutant strain containing mutations of all five sites despite its low WC-1 levels. Together, these data suggest that phosphorylation of these sites negatively regulates the function of WC-1 in the circadian negative feedback loop and is important for the function of the Neurospora circadian clock.

Eukaryotic circadian oscillators consist of autoregulatory transcription/translation-based negative feedback loops (1,2). In these negative feedback loops, the positive elements activate the transcription of the negative elements, whereas the negative elements inhibit their own transcription by inhibiting the activity of the positive elements. In the Neurospora circadian negative feedback loop, such as those in Drosophila and mammals, the positive element is a heterodimeric complex made of two PER-ARNT-SIM (PAS) 1 domain-containing transcription factors, WHITE COLLAR (WC)-1 and WC-2. In the dark, the two WC proteins form a heterodimer through their PAS domains and activate the transcription of the frequency (frq) gene by directly binding to its promoter (3)(4)(5)(6)(7). When the amount of FRQ protein reaches a certain level, the homodimeric FRQ in complex with FRH, a FRQ-interacting RNA helicase, represses frq transcription by interacting with the WC complex and preventing its binding to the frq promoter, thus closing the negative feedback loop (3, 8 -14).
In addition to their essential role in the circadian negative feedback loop in the dark, WC-1 and WC-2 are required for all known light responses in Neurospora, including the entrainment of the circadian clock (4,7,(15)(16)(17)(18)(19)(20)(21). WC-1 binds to chromophore FAD through its photosensory LOV (light, oxygen, or voltage) domain, a specialized PAS domain, and functions as the blue light photoreceptor for light responses (22)(23)(24). Light triggers the formation of a large WC complex and its binding to the promoters of light-inducible genes (23), resulting in lightinduced transcription and light responses.
Regulations of circadian clock proteins by phosphorylation are critical for clock functions in all eukaryotic systems examined (9,(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35). In Neurospora, FRQ, WC-1, and WC-2 proteins are all phosphorylated in vivo (9,36,37). Extensive studies of FRQ phosphorylation revealed that its phosphorylation status is determined by both kinases and phosphatases. Casein kinases I and II, and a calcium/calmodulin-dependent kinase have been shown to be the kinases that phosphorylate FRQ (38 -41), whereas protein phosphatase 1 and protein phosphatase 2A counter the effects of the kinase by dephosphorylating FRQ (42). Phosphorylation of FRQ promotes its degradation through the ubiquitin/proteasome pathway mediated by FWD-1, an F-box/WD40 repeat-containing protein, which is part of an SCF-type ubiquitin E3 ligase (29,43). In addition, the phosphorylation of FRQ regulates the FRQ-WC interaction and is important for the role of FRQ to repress the activity of the WC complex (39,40,42). Thus, regulation of FRQ phosphorylation states by its kinases and phosphatases is critical for circadian periodicity and proper function of the clock.
Like the positive elements in the circadian negative feedback loops of Drosophila and mammals (31,44), WC-1 and WC-2 proteins are phosphorylated under constant conditions (36,37). Similar to other blue light photoreceptors, such as the plant cryptochromes and phototropins (32,45,46), the WC proteins become hyperphosphorylated after light exposure. The functions of these WC phosphorylation events are not known. Previously, pharmacological studies and in vitro phosphorylation assay suggested that protein kinase C (PKC) might be a kinase that phosphorylates WC-1 (47). Treatment of Neurospora cells by PKC inhibitors led to increased and prolonged light-induced transcription, suggesting that PKC may be a negative regulator of light responses in Neurospora. On the other hand, the transient nature of WC-1 hyperphosphorylation after light exposure and faster degradation rate of WC-1 under light than in the dark suggest that light-induced phosphorylation may lead to WC-1 degradation (36,48).
WC-1 and WC-2 form WC complexes through the PASC region of WC-1 and the PAS domain of WC-2 (6,7). Even though the WC complexes activate gene transcription both in the dark and after light exposure, these two functions can be separated molecularly. The deletion of the WC-1 LOV domain abolishes the light function of the protein, but not its dark function (22). On the other hand, deletion of the zinc finger DNA binding domain of WC-1 eliminates the dark function of the protein, but not its light function (7). Thus, WC-1 uses different domains of the protein to carry out its roles in gene activation in the dark and after light exposure. Therefore, regulation of different domains of WC-1 is likely to have a different impact on these two functions.
In this study, to understand the role of WC-1 phosphorylation in the regulation of circadian clock and light responses, we identified five in vivo WC-1 phosphorylation sites by mass spectrometry analyses. These sites are located immediately downstream of the zinc finger DNA binding domain. Mutation of these phosphorylation sites showed that although they are not required for the light function of the protein, they negatively regulate the activity of WC-1 in the dark and are important for the function of the circadian clock.

EXPERIMENTAL PROCEDURES
Strains, Culture Conditions, and Race Tube Assay-The bd, a (containing a wild-type clock) strain was used as the wild-type strain in this study. A wc-1 RIP strain was used as the host strain for various his-3targeting wc1-2 constructs (22). Liquid culture conditions were as described previously (8). Race tube assay media contained 1ϫ Vogel's, 0.1% glucose, 0.17% arginine, and 50 ng/ml biotin. Densitometric analyses of race tubes and calculations of period length were performed as previously described using Chrono II version 11.1 (49). Cultures were harvested by filtration either under red safety light or after light treatments (1600 LUX). For rhythmic experiments, the Neurospora cultures were moved from LL to DD at time 0 and were harvested in constant darkness at the indicated times (hours).
Identification of in Vivo WC-1 Phosphorylation Sites by Tandem Mass Spectrometry-Purification of the WC complex was as previously described (22). The SDS-PAGE gel was stained with colloidal blue (Invitrogen). Phosphorylation sites were identified by a combination of precursor ion scanning and nanoelectrospray tandem mass spectrometry (MS/MS). The colloidal blue-stained WC-1 protein bands were excised from SDS gels and subjected to in-gel digestion with trypsin essentially as described previously (50). The dried protein digests were dissolved in 5% formic acid and loaded onto a pulled capillary filled with POROS R2 resin. After washing three times with 5% formic acid, the peptides were eluted into a nanoelectrospray needle with 1-2 l of nanoelectrospray sample solutions for either precursor ion scanning in negative ion mode or MS/MS in positive ion mode.
All mass spectrometry analyses were performed on a QSTAR Pulsar-i quadrupole time-of-flight tandem mass spectrometer (Applied Biosystems/MDS Sciex, Toronto, Canada) equipped with a nanoelectrospray ion source (MDS Proteomics, Odense, Denmark). For precursor ion scanning experiments, the instrument was set in negative ion mode, with the quadrupole Q2 pulsing function turned on, to detect the PO 3 Ϫ fragment ion at m/z Ϫ79. The optimum collision energies were determined for each experiment by gradually increasing the voltage of Q0 in steps corresponding to one-twentieth of the m/z value of the precursor ion. After data acquisition by precursor ion scanning, the instrument was switched to positive ion mode, and the phosphopeptide sequence and sites of phosphorylation were identified by nanoelectrospray MS/ MS. In the MS/MS scan mode, precursor ions were selected in quadrupole Q1 and fragmented in the collision cell (q2), using argon as the collision gas.
Protein and RNA Analyses-Protein and RNA analyses were as previously described (8,9,11). Equal amounts of total protein (50 g) were loaded in each protein lane, and the blots were developed by chemiluminescence (ECL; Amersham Biosciences). For Northern blot analysis, equal amounts of total RNA (20 g) were loaded onto agarose gels for electrophoresis, and the gels were blotted and probed with RNA probe specific for vvd, al-3, and frq. Densitometric analysis of the data was performed using National Institutes of Health IMAGE 1.61. For phosphatase treatments, 50 g of total protein was diluted in 1ϫ phosphatase buffer and treated with 1000 units of phosphatase (New England Biolabs) for 30 min at 30°C.

Identification of Five in Vivo WC-1 Phosphorylation Sites by
Tandem Mass Spectrometry-Previously, we created a Neurospora strain (Myc-His-WC-2) in which WC-2 was tagged by both c-Myc and His 6 epitope tags (22). We showed that the Myc-His-WC-2 can form a functional complex with WC-1 and can rescue the light and circadian clock defects of a wc-2-null strain. Using this strain, we developed a biochemical purification protocol that can obtain pure WC complex. To identify WC-1 phosphorylation sites in vivo, large-scale purifications were carried out using extracts of the Mys-His-WC-2 strain grown in DD. In DD, the phosphorylation profile of WC-1 is similar to that of the LL condition, lacking the hyperphosphorylated WC-1 species induced by a light pulse (36,37). The left panel of Fig. 1A shows the colloidal blue-stained SDS-PAGE gel of the purification products. Similar to our previous result, the WC complex was purified to near homogeneity. The minor bands are mostly degradation products of the WC proteins, except for one contaminating band (labeled with an asterisk). To confirm that the purified WC proteins are still phosphorylated after the purification procedure, a small portion of the purification products were subjected to -phosphatase treatment. As shown in the right panel of Fig. 1A, the phosphatase treatment led to the disappearance of the slow mobility WC-1 species, indicating that they were phosphorylated WC-1. No significant gel mobility difference was observed with Myc-WC-2 after the treatment, suggesting that the majority of WC-2 was not phosphorylated under this condition (36). Thus, we only used the purified WC-1 protein for phosphorylation site mapping studies.
The colloidal blue-stained protein bands corresponding to WC-1 were excised from SDS gels and subjected to trypsin digestion. The resulting peptides were then analyzed by nanoelectrospray MS/MS (50). Most of the identified phosphopeptides match to one trypsin-digested fragment, corresponding to amino acids 988 -999 of WC-1. Fig. 1B shows a representative result of the MS/MS analysis for one of the phosphopeptides, which carries a single phosphorylated serine. In this peptide, Ser-990 was identified as the phosphorylation site. Additional mass spectrometry results indicate that, in some of the phosphopeptides, all five serines in this fragment are phosphorylated. Thus, Ser-988, Ser-990, Ser-992, Ser-994, and Ser-995 are WC-1 phosphorylation sites in vivo. The protein fragment containing these five phosphorylation sites is located immediately downstream of the zinc finger DNA binding domain (Fig. 1C). Interestingly, two of the phosphorylated serines are followed by a proline residue, suggesting that these phosphorylation events may be mediated by a proline-directed kinase.
Phosphorylation of These Sites is Light-independent-Although mass spectrometry can identify in vivo phosphorylation sites, it does not provide information on the extent of phosphorylation of the identified sites in vivo. To confirm that these five residues represent major WC-1 phosphorylation sites in vivo under different conditions, constructs were created in which these sites were mutated to alanines either singly or in combination, and the constructs were transformed into a wc-1-null strain (wc-1 RIP ) (22). As shown in Fig. 2, A and B, mutation of three of the sites (3A; S988A/S990A/S992A) or all five sites (5A; all five sites were mutated to alanines) led to mostly hypophosphorylated WC-1 forms in LL and DD, as indicated by the significantly reduced phosphorylated WC-1 species and the result of the phosphatase treatment. These data indicate that these sites are major WC-1 phosphorylation sites under constant conditions.
To examine whether the phosphorylation of these sites is responsible for the light-induced hyperphosphorylation of WC-1, phosphorylation profiles of WC-1 of the wild-type and 5A strains were monitored after a transition from DD to LL. As shown in Fig. 2C, after 15-30 min of light treatment, the light-induced hyperphosphorylation of WC-1 was similar in both the wild-type and 5A strains. Together with the results in Fig. 2, A and B, these data suggest that these five phosphorylation sites are not involved in the light-induced hyperphosphorylation of WC-1 and that their phosphorylation is light-independent.
To exclude the possibility that the normal light-induced WC-1 hyperphosphorylation profile observed in the 5A mutant was due to phosphorylation of other unidentified sites in this region, we monitored the light-induced WC-1 phosphorylation in two additional mutant strains, bNLS (deletion of the putative nuclear localization signal and the WC-1 C-terminal region) and NLSa (deletion of the zinc finger and the C-terminal region) (7), in which the entire zinc finger region and C-terminal part of the protein were deleted. As shown in Fig. 2D, the light-induced hyperphosphorylation of WC-1 was maintained in these two strains, indicating that the light-dependent hyperphosphorylation of WC-1 is due to phosphorylation of unidentified sites in the N-terminal part of WC-1. Because the size of WC-1 in these deletion mutants is smaller than the wild-type WC-1, their lightinduced phosphorylation appeared to be more extensive.
In addition to the hypophosphorylation of WC-1 in the 3A and 5A mutants, their WC-1 levels were significantly reduced compared with the wild-type strain. The levels of WC-1 in the 3A and 5A mutants were about 30% and 20% of the wild-type levels, respectively, suggesting that the phosphorylation of The arrow indicates the phosphorylated WC-1 forms. B, identification of WC-1 phosphorylation sites by tandem mass spectrometry. Following identification of phosphorylated peptides by precursor ion scanning in negative ion mode, peptide ions at the corresponding m/z values were analyzed by tandem mass spectrometry in positive ion mode. The mass spectrometry of the product ions resulting from collision-induced dissociation of one singly phosphorylated WC-1 peptide is shown. The product ion peaks (either b series or y series) that were most diagnostic of the phosphorylation site are indicated along with the sequence of the peptide. The fragments representing individual b or y ions are indicated by lines between the corresponding amino acids. C, a diagram showing the WC-1 domain structure and the peptide containing the five identified phosphorylation sites. The two arrows indicate the trypsin digestion sites.

FIG. 2. Phosphorylation of these sites is light-independent.
A and B, Western blot analysis showing that WC-1 is hypophosphorylated in DD (top panels) and LL (bottom panels) in the 3A and 5A strains. The arrows indicate the phosphorylated WC-1 forms that were mostly absent in the mutant strains. The asterisk indicates a Neurospora protein cross-reacted with our WC-1 antiserum. In B, some of the samples were treated with phosphatase before analysis. C and D, Western blot analysis showing that the light-induced transient WC-1 hyperphosphorylation is normal in the 5A strain (C) and in strains with deletion of the WC-1 C terminus (bNLS and NLSa, which have the zinc finger and its downstream part of the protein deleted). The arrows indicate the hyperphosphorylated WC-1 species induced by light. E, densitometric analysis of the Western blot results showing the stability of WC-1 of the wild-type strain and the 5A mutant after the addition of cycloheximide (CHX). The strains were grown in DD for 1 day before the addition of cycloheximide (10 g/ml) and the transfer into LL. these sites may be important for maintaining the steady-state levels of WC-1. The low levels of the WC-1 proteins in the mutants are not due to their expression at the his-3 locus because the wild-type wc-1 construct (wc1-2) was expressed at normal level ( Fig. 2A). In addition, Northern blot analysis showed that the expression levels of wc-1 mRNA in the mutants were comparable with that of the wild-type strain (data not shown). When WC-1 stability was measured in the presence of the protein synthesis inhibitor cycloheximide, the degradation rates of WC-1 were not significantly different between the wild-type and the 5A strains (Fig. 2E), suggesting that the phosphorylation of these sites does not promote WC-1 degradation. Thus, it is likely that these mutations may affect WC-1 protein folding after its synthesis (improperly folded protein would be quickly degraded), leading to low levels of WC-1 without affecting the stability of the properly folded protein.
Light-induced Transcription in the 5A Mutant-To examine whether the phosphorylation of these five sites regulates the light function of WC-1, light-induced transcription in the 5A mutant was compared with that in the wild-type strain. Light induction of albino-1 (al-3), vivid (vvd), and frq was examined by Northern blot analysis due to their different WC-1 level requirements for light-induced transcription (7, 19 -21, 24). Light induction for genes such as al-3 is very sensitive to changes in WC-1 levels, whereas near normal light induction of frq can be observed even when the WC-1 level is very low. The WC-1 requirement for vvd light induction is between those of al-1 and frq. In the 5A mutant, as shown in Fig. 3, the light induction of al-3 was significantly reduced (Fig. 3A), whereas the light induction of vvd was only modestly affected (Fig. 3B). For frq, induction in the mutant was similar to that of the wild-type strain. Because of the different WC-1 requirements for light induction of these three genes, the changes in lightinduced transcription observed for al-1 and vvd are most likely due to the reduction of WC-1 levels in the 5A mutant rather than changes in WC-1 activity. Thus, phosphorylation of these five sites does not appear to regulate the light function of WC-1. This interpretation is consistent with our previous results that normal light-induced transcription was observed in strains lacking the WC-1 zinc finger domain and the C-terminal region including these sites (7).
Short Period, Low Amplitude, or Arrhythmic Conidiation Rhythms in DD When WC-1 Phosphorylation Sites Are Mutated-Because the zinc finger DNA binding domain of WC-1 is required for its dark function but not its light function (7) and the identified phosphorylation sites are immediately downstream of the zinc finger domain, it is likely that these phosphorylation events regulate the function of WC-1 in the circadian feedback loop in the dark. Thus, we examined the circadian conidiation rhythms of the wc-1 RIP strains carrying a construct with either a single or multiple WC-1 phosphorylation site mutations in DD. In addition to its arrhythmicity on race tubes, our wc-1 RIP strain showed a faster growth rate (ϳ15%) than the wild type on race tubes, similar to other previously described wc-1 mutants (20). As shown in Fig. 4A, a wild-type wc-1 construct (WC1-2) (6) was able to rescue the robust circadian conidiation rhythmicity of the wc-1 RIP strain with a period that was ϳ1 h longer than that of the wild type. Mutation of a single phosphorylation site (1A; S990A) was also able to rescue the conidiation rhythm; however, its period was 1.6 h shorter than that of the WC1-2 strain. A short period conidiation rhythm was also observed in the 2A strain (S988A/ S990A), but its conidiation rhythm became less robust than the 1A strain, as indicated by its broad conidiation bands. When three (3A; S988A/S990A/S992A) or all five sites (5A) were mutated, even though the conidiation bands can be observed in the first 2 days, conidiation became arrhythmic or oscillated with a low amplitude in the following days in DD. The growth rates of these mutants resemble that of the wc-1 RIP strain, probably due to their low WC-1 levels. Previously, we have shown that high WC-1 levels lead to a modestly shorter period length of the clock (5). Thus, the short period rhythms of the phosphorylation site mutants are unlikely to be due to their low WC-1 levels. These results suggest that the phosphorylation of WC-1 at these sites is important for its role as a positive element in the circadian negative feedback loop. Like the wild-type strain, conidiation of all mutant strains could be entrained by light/ dark (LD) cycles (Fig. 4B), further indicating that these WC-1 phosphorylation events are not required for its light function.
Circadian Expression of frq mRNA and FRQ Protein-WC-1 is the limiting protein in the WC complex for its dark and light functions, so that low levels of WC-1 will result in low levels of frq transcription (5,12). Thus, the lack of robust circadian rhythmicity in DD in the 3A and 5A mutants could be due to their low levels of WC-1 (Fig. 2), even though we have previously shown that similar levels of wild-type WC-1 can support robust clock function (5). If so, frq mRNA and FRQ protein levels should be low in these mutants if their WC-1 activities are not altered. To examine this possibility, frq mRNA and FRQ protein levels were monitored in DD for 48 h after the LD transition. As shown in Fig. 5, circadian rhythms of frq mRNA could be observed in the 5A mutant, and the levels of frq RNA are comparable to, if not slightly higher than, those in the wild-type strain. In addition, frq mRNA peaked earlier in the mutant than the wild-type strain, suggesting a short period.
The low levels of WC-1 protein and the normal levels of frq mRNA and its earlier peaks in the 5A strain suggest that its WC-1 activity in activating frq in the dark is increased in this strain compared with the wild-type protein.
At the protein level, FRQ oscillates robustly both in its amount and phosphorylation states in the wild-type strain in DD (Fig. 6). For the 5A mutant, however, low amplitude rhythms of FRQ levels and phosphorylation states were seen for the first 2 days in DD, and the overall levels of FRQ in the 5A mutant were slightly higher than those of the wild-type strain. The low amplitude FRQ rhythm of the 5A strain (especially after 28 h in DD) is consistent with its race tube phenotype in the first 2 days in DD. The comparison of the phosphorylation profile of FRQ between the two strains indicates that the 5A mutant has a shorter period than the wild type (note the ratio between extensively phosphorylated FRQ and newly synthesized hypophosphorylated FRQ at 12 and 28 h). Because the conidiation of the 5A strain became arrhythmic after 2 days in DD, the molecular rhythms observed here are likely due to the light/dark transition. Immunoprecipitation assays suggest that the interaction between FRQ and the WC complex is not affected by the phosphorylation site mutations (data not shown). Together, these molecular data indicate that the lack of robust circadian rhythmicity in the phosphorylation site mutants is not due to their low WC-1 levels, suggesting that the phosphorylation of WC-1 at these sites negatively regulates its activity as a positive element in the circadian negative feedback loop. DISCUSSION Phosphorylation of clock proteins is critical for functions of circadian clocks in eukaryotic systems. WC-1, an essential component of the Neurospora circadian clock, functions both as the blue light photoreceptor mediating light input into the clock and as a positive element in the circadian negative feedback loop in the dark, by activating frq transcription. In this study, using mass spectrometry analyses, we identified five in vivo WC-1 phosphorylation sites near the WC-1 zinc finger DNA binding domain. Mutations of these sites indicate that these phosphorylation events are not required for the light function of WC-1, but they are important for the function of WC-1 as an activator of frq expression in the dark. Strains with one or two of the phosphorylation sites mutated showed short period circadian rhythms, whereas strains with three or five sites mutated exhibited arrhythmic or low amplitude circadian conidiation rhythms. In the mutant with all five sites mutated, despite its low WC-1 levels, the amounts of frq mRNA and FRQ protein in DD are comparable with or slightly higher than those of the wild-type strain. Because the amounts of WC-1 determine the expression level of frq in the dark in a wild-type strain (5), these data suggest that these phosphorylation events negatively regulate the transcription activator activity of WC-1 in the dark. The short period phenotype of the phosphorylation site mutants, therefore, is likely due to their increased activities of WC-1, resulting in earlier activation of frq transcription than seen in the wild-type strain. This interpretation is consistent with the advanced phases of the frq mRNA and FRQ protein observed in the 5A mutant (Figs. 5 and 6). Thus, the regulation of WC-1 activity by phosphorylation of these sites is important for period determination and proper function of the circadian clock. Because the identified phosphorylation sites reside immediately downstream of the WC-1 zinc finger region, which is essential for the dark function of WC-1 but not required for its light function, it is likely that the phosphorylation of this region negatively regulate the DNA binding ability of WC-1 in the dark. This is the first time that in vivo phosphorylation sites and their function have been revealed for a positive element in a eukaryotic circadian negative feedback loop. Like the WC proteins in Neurospora, the positive elements of the circadian negative feedback loop in Drosophila (dCLOCK) and mouse (CLOCK and BMAL1) are phosphorylated in vivo (31,44), but the sites and roles of their phosphorylation are not known.
Hyperphosphorylation of WC-1 was previously observed to be associated with WC-1 degradation in LL (36,48), suggesting that phosphorylation probably promotes WC-1 degradation. However, in strains with mutations of the identified phosphorylation sites, WC-1 levels are low, and the stability of WC-1 is not significantly altered. In addition, the phosphorylation of these sites is light-independent. These data suggest that the phosphorylation of these five sites does not lead to WC-1 degradation. Thus, the likely role of phosphorylation in regulating WC-1 stability is due to light-dependent phosphorylation of WC-1 in the N-terminal part of the protein. The sites of these phosphorylation events remain to be identified.
Pharmacological studies previously suggested that PKC acts as a negative regulator of light responses in Neurospora, probably by regulating the function of the WC complex directly or indirectly (47). In vitro, PKC can phosphorylate the zinc finger region of WC-1, but it is unclear whether this phosphorylation is relevant in vivo. Although the phosphorylation sites we identified here are located in the zinc finger region, their phosphorylation is unlikely to be due to PKC phosphorylation. First, the amino acid sequence surrounding these five sites does not resemble those of the PKC consensus phosphorylation sites (S/TXK/R), although there are putative sites similar to the PKC sites in the zinc finger region. Second, mutations of these sites or deletions of the zinc finger region do not appear to affect the light-dependent WC-1 phosphorylation and the function of WC-1 in light-induced transcription (7). Thus, the function of PKC in regulating light responses is probably not through its phosphorylation of the zinc finger region of WC-1.
The proline residues next to two of the phosphorylation sites suggest that their phosphorylation is likely mediated by a proline-directed kinase. Interestingly, glycogen synthase kinase-3, a known proline-directed kinase (51,52), has been shown to regulate the circadian clock in Drosophila (33). Because of the conservation of known clock kinases from Neurospora to animal circadian systems (27,30,35,39,41,53), it will be interesting to examine to the role of the Neurospora glycogen synthase kinase-3 in WC-1 phosphorylation.