Circadian Transcription Depends on Limiting Amounts of the Transcription Co-activator nejire/CBP*

The circadian clock orchestrates physiological and behavioral activities, including metabolism, neuronal activity, and cell proliferation in synchrony with the environmental cycle of day and night. Here we show that the Drosophila ortholog of the CBP/p300 family of transcription co-activators, nejire (nej), is an intrinsic component of the circadian clock that performs regulatory functions for circadian controlled transcription. Screening of overexpression mutants revealed that gain of nej function was associated with a loss of behavioral and molecular rhythms. Overexpression of NEJ suppresses the long period phenotype of a mutation in the clock gene period (per). NEJ physically interacts through two binding sites with CLOCK and the CLOCK·CYCLE (CLK·CYC) complex. Induction of CLK·CYC-dependent transcripts upon induction of nej expression from a heat-shock promoter showed that NEJ is limiting. Reduced CLK·CYC-mediated transcription in a nej hypomorphic mutant indicates an essential function of NEJ/CBP for CLK·CYC activity and a regulation of circadian transcription by availability of the co-activator. Competition for recruitment of NEJ/CBP provides a potential mechanism for cross-talk between circadian transcription and other CBP-dependent physiological processes.

The circadian clock orchestrates physiological and behavioral activities, including metabolism, neuronal activity, and cell proliferation in synchrony with the environmental cycle of day and night. Here we show that the Drosophila ortholog of the CBP/ p300 family of transcription co-activators, nejire (nej), is an intrinsic component of the circadian clock that performs regulatory functions for circadian controlled transcription. Screening of overexpression mutants revealed that gain of nej function was associated with a loss of behavioral and molecular rhythms.

Overexpression of NEJ suppresses the long period phenotype of a mutation in the clock gene period (per). NEJ physically interacts through two binding sites with CLOCK and the CLOCK⅐CYCLE (CLK⅐CYC) complex. Induction of CLK⅐CYCdependent transcripts upon induction of nej expression from a heat-shock promoter showed that NEJ is limiting. Reduced CLK⅐CYC-mediated transcription in a nej hypomorphic mutant indicates an essential function of NEJ/CBP for CLK⅐CYC activity and a regulation of circadian transcription by availability of the co-activator. Competition for recruitment of NEJ/CBP provides a potential mechanism for cross-talk between circadian transcription and other CBP-dependent physiological processes.
The circadian clock controls genome wide transcription of many key regulatory components in a diverse selection of vital pathways (1-3) that ultimately allow a coordination of physiological and behavioral activities and their synchronization with the environmental cycles of day and night. The analogous and homologous clock mechanisms in Drosophila and mammals are based on two interconnected feedback loops (4,5). In Drosophila, the heterodimeric complex of transcription factors CLOCK (CLK) 3 and CYCLE (CYC) (BMAL1 in mammals) acti-vates expression of its own inhibitors PERIOD (PER) and TIME-LESS (TIM) forming the first feedback loop. This loop is interconnected with CLK⅐CYC-mediated expression of the transcription repressor vrille (vri) and the activator par-domain protein 1 (pdp1). VRI and PDP1 control the rhythmic transcription of Clk and contribute to the robustness of molecular oscillations (6,7).
Oscillations in cyclic nucleotide, calcium, and MAPK signaling (8 -10) likely contribute to a circadian control of physiological processes such as cell proliferation (11) and the sleep/wake cycle, which is important for memory formation (12). However, these pathways also feedback on the molecular oscillator at least in part through control of CLK⅐CYC activity (13). Cross-talk between circadian and cell signaling may increase the robustness of circadian oscillations and allow a coordination of circadian transcription with physiological requirements. Previous studies showed that recruitment of the CREB-binding protein (CBP) from a limiting cellular pool mediates cross-talk between the transcription factors E2F, JAK/STAT, AP1, and nuclear hormone receptors (14 -16) that control e.g. entry into the cell cycle and the immune response.
Here we show that CLK⅐CYC-mediated transcription is also dependent on CBP, and importantly circadian transcription responds to changes in limiting levels of the co-activator. These findings suggest a novel mechanism for cross-talk between CLK⅐CYC and other CBP-dependent transcription factors.  [1] (nej Q7 ) were obtained from Bloomington Drosophila stock center and EP(X)950, EP(X)1410, EP(X)1149, and EP(X)1179 were obtained from Szeged Drosophila stock center. nej Q7 flies were crossed with Canton S flies to obtain ϩ/Dp(1;Y)FF1, y[ϩ] flies (Dp(1;Y)FF1). hs-nej and Dp(1;Y)FF1 males were used as nej overexpressing flies together with Clk Jrk for arrhythmic and wild-type Canton S and w 1118 for rhythmic controls. The BGluc reporter gene (17) that expressed luciferase from a per promoter was crossed into these backgrounds to generate BGluc and BGluc;; Clk Jrk as well as BGluc/Dp(1;Y)FF1 and BGluc;; hs-nej flies. EP lines were crossed with tim-GAL4 flies to generate EP(X)/Y; tim-GAL4/ϩ as shown in figures.

EXPERIMENTAL PROCEDURES
Locomotor Activity Assays-Flies were entrained during eclosion for 4 days in cycles of 12-h light and 12-h darkness (LD) at 20 or 22°C. 1-4-day old flies were analyzed in locomotor activity assays using the DAM system IV (TriKinetics, Waltham, MA) for 3 days in LD cycles, with or without application of a 37°C heat shock during the last hour in darkness and subsequently for 7 days in constant darkness at 20 or 22°C. Data from the first 5 days in constant darkness were analyzed using Clocklab (Actimetrics, Wilmette, IL) and IandA software (17) to identify rhythmic flies and determine period estimates.
Real-time Bioluminescence Measurements-Bioluminescence from luciferase reporter gene expression in live flies was determined for 3 days in constant darkness as described previously (17) after entrainment of flies for 4 days in 12 h LD cycles. Rhythmic flies were identified by analysis of the first 3 days in constant darkness with IandA software (17).
Quantitative Real-time PCR (RT-PCR)-1-7-day old flies were harvested after incubation for at least 48 h in constant light at either constant 20°C or after application of a 30-min heat shock at 37°C. Heads were isolated and total RNA was purified by homogenization in peqGOLD TriFast (PEQLAB, Erlangen, Germany) following manufacturer's instructions. 1 g of total RNA was reverse transcribed with random hexamer primers using the Quantitect Reverse Transcription kit (Qiagen, Hilden, Germany) following manufacturer's instructions. cDNA products were amplified in an ABI PRISM 7000 (Applied Biosystems, Foster City, CA). Forward (fwd) and reverse (rev) primers and probes for TaqMan quantitative RT-PCR were designed with the ABI PRISM Primer Express software (Applied Biosystems) as follows: n-synaptobrevin (fwd, GGC  GGC GTG TAA GCA ATC; rev, CCC GCT GAA GGA GCA  CAC TA; probe, 6-FAM-CGC TGC CAG GAC GAA AGT  TTC TCG A-TAMRA) . n-synaptobrevin mRNA levels were determined as a constitutively expressed internal control. mRNA levels for individual clock genes were normalized toward n-synaptobrevin transcript levels and quantified by the 2 Ϫ⌬⌬Ct method as described (18,19) and according to the manufacturer's instructions (Applied Biosystems).
Western Blotting-CLK protein levels were determined by Western blot analysis as detailed in the legend of Fig. 4 using an antibody that was raised in rabbit against a peptide composed of the C-terminal 15 amino acids of CLK.
Immunohistochemistry-Third instar larvae were entrained for 3 days in cycles of 12 h LD, and the brains were dissected during the first day of constant darkness at times indicated in the figures. Circadian time (CT) 0 marks time of subjective "lights on" and CT12 marks time of subjective "lights off". The experiments were repeated twice, and for each time point at least 8 brains were dissected. Brains were fixed in 4% formaldehyde in PEM buffer (100 mM Pipes, pH 6.9, 1 mM EGTA, 2 mM MgSO 4 ) for 2 h, blocked with 10% normal goat serum (NGS) in PT buffer (phosphate-buffered saline plus 0.3% Triton X-100) for 2 h at room temperature, and subsequently incubated for 48 h at 4°C in 50 l of primary antiserum solution containing 1:200 diluted rabbit anti-PER (Alpha Diagnostics, San Antonio, TX) and 1:1000 diluted guinea pig anti-PAP (20) in PT buffer with 10% NGS. Brains were rinsed with PT buffer and PT plus 5% NGS for 20 min each and then incubated at room temperature for 2 h in secondary antiserum solution containing 1:200 diluted TRITC-conjugated donkey anti-guinea pig (Jackson Immuno-Research Laboratories, Inc., West Grove, PA) and 1:200 diluted fluorescein isothiocyanate-conjugated goat anti-rabbit (Calbiochem) in PT buffer with 10% NGS. After incubation, brains were rinsed with PT buffer and PT plus 5% NGS for 20 min each and mounted in the mounting medium (50 mM Tris-Cl, pH 8, 90% glycerol, 2.5% DABCO (Sigma)).
Confocal Microscopy and Quantification of Staining-Optical sections of larval lateral neurons (LNs) were imaged on a Carl Zeiss LSM 510 META confocal microscope. For each CT, LNs from at least 16 brain hemispheres per fly strain were scanned. For each LN sample, PAP staining was used to identify and select an optical section, which was then scanned for PER immunoreactivity. Three images were taken per LN with single laser at 488 nm, 543 nm, and with both lasers together. Images taken with single 488-nm laser were imported to NIH ImageJ 1.34s, and the localization of neuronal PER staining was determined by the double-laser images. Total pixel intensity of PER staining was measured, and the mean Ϯ S.E. of all brains at a particular CT is reported in the graphs.
Expression Constructs and Co-transfection Assays-Clk, E1A(12S), and a E1A(12S) construct that carries a deletion of the first 72 amino acids (E1A(⌬N)) were cloned into XhoI and PmeI sites of pAc5.1/V5-HisA vector (Invitrogen) for expression in Drosophila S2 cells. pHT control vector, reporter plasmids pRLcopia for Renilla luciferase control and pGL3-(4-per-E-box)hs::lucϩ, pGL3-per::lucϩ, and pGL3-tim::lucϩ for expression of firefly luciferase from a minimal heat-shock promoter with four per-Ebox elements, a per promoter, and a tim promoter, respectively, are described previously (13,21). Co-transfection assays were performed as described previously (13) with either 200 ng of pHT control, pAc-E1A(12S), or pAc-E1A(⌬N) vectors. Averages of at least three independent experiments are shown as percent of control. Control is CLK⅐CYC-activated LUC activity set to 100% in the presence of pHT control vector. Expression levels of E1A constructs were determined by Western blot analysis using an anti-E1A(12S) antibody (Santa Cruz Biotechnology, Santa Cruz, CA).
For dsRNA production, approximately 500-bp long DNA fragments of the coding region of target genes Clk, cyc, nej, and EGFP (enhanced green fluorescent protein) control were generated by PCR using primers that contained a 5Ј T7 RNA polymerase-binding site followed by gene specific sequences. Primers were as follows: Clk (fwd, TAA TAC GAC TCA CTA TAG  GGG TGG TCT GCA CCC  To test effects of dsRNAs on reporter gene expression in co-transfection assays, 250 l of Drosophila S2 cells (2 ϫ 10 6 cells/ml) in serum-free Schneider's insect medium (Sigma) were incubated with 7.5 g of dsRNA for 45 min prior to addition of 250 l of Schneider's medium containing 20% fetal bovine serum. Cells were incubated for 24 h, and co-transfection assays were performed essentially as described previously (13) with the difference that another 7.5 g of dsRNA was added 45 min prior to the addition of 20% fetal bovine serum. Averages of at least three independent experiments are shown  OCTOBER 26, 2007 • VOLUME 282 • NUMBER 43 as percent of control. Control is CLK⅐CYC-activated LUC activity in the absence of dsRNA set to 100%.

nej/CBP Co-activates CLK⅐CYC-dependent Transcription
Co-immunoprecipitation Experiments-FLAG-tagged CLK or CLK Jrk as well as CYC and domains of NEJ were expressed from a pAc5.1 vector, which carried a SP6 promoter insertion in the KpnI site, in reticulocyte lysate using the SP6 quick-coupled transcription/translation system (Promega) as instructed by the manufacturer. NEJ domains include amino acids 1-897 (NEJ1), 898 -1667 (NEJ2), 1668 -2653 (NEJ3) and 2677-3190 (NEJ4). FLAG-tagged CLK or CLK Jrk was expressed with unlabeled methionine to reduce background, whereas CYC and NEJ fragments were labeled with [ 35 S]methionine. The immunoprecipitation was performed as described previously (22). In brief, 10 l of NEJ-expressing lysate was pre-incubated with 10 l of anti-FLAG-M2-conjugated-agarose (Sigma) in 150 l of IB solution (5 mM Tris-Cl, 10 mM HEPES, pH 7.5, 10% glycerol, 50 mM KCl, 0.05% Triton X-100, 1 mM EDTA, 1 mM dithiothreitol, 1ϫ complete protease inhibitor (Roche Diagnostics)) to eliminate nonspecific interaction. The agarose was then removed from the lysate. The NEJ lysate was added to different interaction reactions either with or without 10 l of CLK or CLK Jrk lysate in the absence or presence of 10 l CYC lysate (as indicated in Fig. 3), and the final volume of the interaction mixture was adjusted to 500 l with IB solution. The protein mixture was incubated at 25°C with gentle rotation for 30 min. Subsequently, 10 l of anti-FLAG-M2-conjugated-agarose, previously pre-blocked in IB solution with 1% milk, was added and incubated for 2 h at 25°C. The beads were then collected and washed three times with 500 l of IB solution for 10 min each. Proteins were eluted with 30 l of 2ϫ sodium dodecyl-sulfate (SDS) sample buffer, and radiolabeled proteins were analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography.

RESULTS AND DISCUSSION
We identified a circadian function of nejire (nej), which is the CBP ortholog in Drosophila, by screening of duplication mutants for phenotypes in circadian behavioral rhythms. A duplication of the X-chromosomal nej locus on the Y chromosome in Dp(1;Y)FF1 caused a large increase in the number of arrhythmic flies ( Fig. 1A; Table 1). Specific overexpression of nej from a heat-shock promoter (hs-nej) resulted in a similar phenotype. Interestingly even a moderate increase in nej levels by about 45%, as observed in hs-nej flies in the absence of heat shock (Fig. 1B), was associated with a loss of behavioral rhythms in 71% of the flies (Table 1, Fig. 1A, and supplemental Fig. S1). The remaining rhythmic flies displayed a wild-type period length. This phenotype was reminiscent to the effects observed for the overexpression of Clk, which also increased the number of arrhythmic flies without affecting period length (23). A similar increase in arrhythmicity of nej overexpressing flies was observed after entrainment in temperature cycles under subsequent free-running conditions (Table 1). This finding indicated that nej did not affect a specific entrainment pathway but the circadian clock itself. To further test specificity of the behavioral phenotype, we overexpressed nej from a UAS enhancer after transcription of a GAL4 driver from a tim promoter (tim-GAL4). Several EP lines carry an insertion of the UAS enhancer in the 5Ј-untranslated region of nej (24). EP(X)1179 overexpressed nej by the tim-GAL4 driver, whereas two lines, EP(X)950 and EP(X)1410, carried the UAS enhancer in the inverse orientation, serving as a control that lacked nej overexpression (Fig. 1, C and D). Overexpression of nej by tim-GAL4 in EP(X)1179 resulted in an increase of behavioral arrhythmicity, without significant effects on general activity levels ( Table  1). In contrast, EP(X)950 and EP(X)1410 showed wild-type rhythms. These results demonstrate that nej gain of function causes an increase in the number of behaviorally arrhythmic flies.
To analyze whether nej has a role in the molecular oscillator itself, we investigated the circadian profile of clock gene expression by real-time monitoring of CLK⅐CYC-dependent expression of a luciferase reporter from a per promoter in individual BGluc flies (25). After entrainment of BGluc flies in cycles of 12-h light and 12-h darkness, CLK⅐CYC-dependent luciferase bioluminescence was monitored for 3 days in constant darkness. Luciferase expression oscillated in BGluc control flies, whereas the antimorphic Clk Jrk mutation caused molecular arrhythmicity ( Fig. 2A). A moderate overexpression of nej from the heat-shock promoter at 20°C induced arrhythmicity in 40% of hs-nej flies ( Fig. 2A). Similarly, duplication of the nej locus caused molecular arrhythmicity in 64% of the flies (Fig. 2A). These results show that overexpression of nej affects the ability to maintain free running molecular rhythms, consistent with the observed increase of behaviorally arrhythmic flies.
We also analyzed PER protein expression in PDF positive pacemaker lateral neurons of larval brains from EP lines that carried a tim-GAL4 driver (Fig. 2, B and C and data not shown). Although EP(X)950 and EP(X)1410 control flies displayed normal PER oscillations with trough levels at CT 10 and a peak at CT 24, the EP(X)1179 line showed a strongly reduced amplitude of PER protein oscillations. For both time points PER  (Fig. S1). c NA, not analyzed.

nej/CBP Co-activates CLK⅐CYC-dependent Transcription
staining intensities in nej overexpressing EP(X)1179; tim-GAL4 flies were similar to trough PER levels in the EP(X)950; tim-GAL4 control line (supplemental Fig. S2). In summary the data show that gain of nej function is associated with a strong increase in molecular and behavioral arrhythmicity.
In contrast to gain of function, a nej loss of function cannot be analyzed because of embryonic lethality. Two heterozygous nej loss of function mutants (nej 3 and nej Q7 ) did not display a phenotype because of the remaining wild-type allele (data not shown). The lack of a period phenotype for nej partial loss of function is reminiscent to the effects of decreased CLK⅐CYC activity in partial Clk loss of function flies, which show a reduced amplitude but wild-type period length of circadian oscillations (26). However, the insertion of the UAS enhancer in the 5Ј-untranslated region of nej in EP(X)1149 flies has also been shown to cause a nej partial loss of function, when investigated in the absence of a GAL4 driver (24). We observed a severe long period phenotype in EP(X)1149 flies (Table 1). However, this phenotype did not segregate with the EP-element insertion. Sequencing revealed that this line carried an additional mutation in the per locus that substitutes serine 45 to tyrosine and has previously been coined per SLIH (27). The per SLIH mutation causes a temperature-sensitive phenotype with a long period oscillation of about 27.6 h at low temperature and a shorter period length at higher temperatures. Because nej can be overexpressed in EP(X)1149 flies by a tim-GAL4 driver (Fig. 1D), we analyzed the effects of nej overexpression on the per SLIH phenotype. In the presence of the tim-GAL4 driver, overexpression of nej in EP(X)1149 caused a strong increase in arrhythmic flies, similar to the observations in EP(X)1179 (Fig. 1A, Table 1). The remaining rhythmic flies showed, however, a wild-type period length at low temperatures demonstrating that nej overexpression suppressed the temperature-sensitive effect of the per SLIH mutation. The genetic interaction between per and nej supports an intrinsic clock function of nej as a co-activator for CLK⅐CYCdependent transcription.
To directly test this hypothesis we first analyzed the effects of decreased nej activity on CLK⅐CYC function in cell culture. CLK⅐CYCdependent luciferase expression from a four per-E-box containing reporter in cell culture (21) was strongly reduced by co-expression of the adenoviral E1A(12S) protein (Fig. 3A), a well characterized inhibitor of CBP/p300 (28). Specificity of NEJ/CBP inhibition was verified by co-expression of E1A(⌬N), a truncated construct that lacked the N-terminal CBP/p300 binding domain, which showed no inhibition of CLK⅐CYC activity. The same results were obtained using a reporter construct that expressed luciferase from a genomic fragment of the tim or per promoter ( Fig. 3B and data not shown). Both E1A(12S) and E1A(⌬N) were expressed to similar levels (Fig. 3C). To confirm the specificity of reduced CLK⅐CYC function by inhibition of nej, we analyzed the effects of a knock-down of nej expression by small interfering RNAs (siRNA) in the co-transfection assay (Fig. 3D). siRNA targeting either Clk or cyc expression strongly reduced CLK⅐CYC activity as expected. In contrast, control siRNA targeting the EGFP gene had no effect on CLK⅐CYC function. When siRNA targeting nej transcription was included in the reporter assay, CLK⅐CYC activity was significantly reduced. The decrease in CLK⅐CYC-dependent transcription was almost proportional to the reduction in nej mRNA levels (compare Fig. 3D and supplemental Fig. S3). An almost complete inhibition of CLK⅐CYC activity (less than 10%) was observed after reduction of nej transcription by siRNA and additional inhibition of residual NEJ by E1A(12S) co-expres-sion (supplemental Fig. S3). These results strongly suggest that nej plays an important role for CLK⅐CYC-dependent transcription.
A direct interaction between NEJ and the CLK⅐CYC complex would be expected, if NEJ acts as a transcription co-activator for CLK⅐CYC-dependent transcription. Yeast two-hybrid assays indicated a binding between CLK and the N terminus of NEJ (data not shown). To map the interaction between Drosophila NEJ and either CLK or CLK⅐CYC, we performed co-immunoprecipitation experiments of reticulocyte lysate-expressed domains of NEJ (Fig. 3E) with FLAG-tagged CLK in the presence or absence of CYC (Fig. 3F). The intensity of the CYC signal served as a control for the efficiency of co-immunoprecipitation with CLK because binding of CYC was specific (Fig.  3G). We identified two domains of NEJ that bound to CLK as well as to CLK⅐CYC complexes (Fig. 3H), the N-terminal fragment from amino acid 1 to 898 (NEJ 1) that also interacted with nuclear hormone receptors such as the retinoic acid (RAR, RXR), estrogen, progesterone, thyroid hormone, and glucocorticoid receptors (29), as well as the middle domain of NEJ from amino acids 1668 to 2653 (NEJ 3) that contained the histone acetyl transferase activity and the zinc finger domains C/H2 and C/H3. Binding of the N-terminal NEJ 1 fragment to CLK was independent of CYC, whereas the interaction of the middle domain NEJ 3 to CLK appeared to be enhanced by the presence of CYC (Fig. 3H). The interaction was independent of the glutamine-rich domain of CLK because both fragments of NEJ bound to the C-terminally truncated CLK Jrk protein that lacked this region (Fig. 3I). The antimorphic nature of the CLK Jrk mutation may be because of effects on the histone acetyl transferase activity of CLK that has previously been identified in mammals (30). Our results show that NEJ physically interacts through two binding sites, the N terminus, and a middle domain, with an N-terminal part of CLK, which is consistent with a role of NEJ as a co-activator for CLK⅐CYCdependent transcription in Drosophila. The interaction of the mammalian CLK⅐BMAL1 complex with CBP and p300 family members suggests that this function is conserved between the invertebrate and vertebrate clock (31)(32)(33), although mutant phenotypes await to be investigated in a mammalian model.
We next addressed the question of whether recruitment of NEJ to the CLK⅐CYC complex is a constituent of the basal transcription machinery or NEJ is a regulatory component of the circadian clock. NEJ as well as the mammalian ortholog of NEJ, CBP, affects transcription through different modes of action: 1) through histone acetylation; 2) by mediating interactions between transcription factors and the basal transcription machinery (co-activation); 3) by mediating crosstalk between transcription factors (29). Previous studies in mammals showed that the cellular concentration of CBP is limiting, and CBP-dependent transcription factors such as AP1, nuclear hormone receptors, the JAK/STAT pathway, and E2F compete for recruitment of CBP, thereby establishing a CBP-mediated cross-talk between these signaling pathways (14 -16). The molecular arrhythmicity induced by overexpression of NEJ in flies suggests that NEJ is a limiting factor for CLK⅐CYC activity as well. To further analyze a co-activator function of NEJ for CLK⅐CYC-mediated transcription in vivo and to test whether trans-activation by CLK⅐CYC responds to changes in nej levels, we assayed constitutive CLK⅐CYC-dependent transcription in wild-type and nej mutant flies under constant light conditions, when the Drosophila circadian clock did not oscillate. Under such conditions at least the per/tim feedback loop is not functional because of degradation of PER and TIM proteins in light. Therefore feedback mechanisms that may compensate for an increase or decrease of CLK⅐CYC activity are at least partially non-functional, which allows to monitor effects on CLK⅐CYC activity in flies more directly. tim and pdp1 transcript levels were decreased in the nej partial loss of function mutant EP(X)1149 compared with EP(X)950 control, consistent with a co-activator function of nej for CLK⅐CYC-mediated transcription (Fig. 4, A, C, and E). hs-nej flies showed an induc- tion of tim and pdp1 mRNA levels upon heat shock, whereas wild-type flies displayed no significant changes or rather decreased transcript levels (Fig. 4, B, D, and F). Similar results were obtained for per and vri expression (Supplemental Fig. S4). CLK levels were not significantly different in these fly strains and conditions as determined by Western blotting (Fig. 4, G and H). These results show that overexpression of nej increases CLK⅐CYC-dependent transcription by ϳ50% suggesting that under wild-type conditions about one-third of CLK⅐CYC complexes lack NEJ to reach maximal activity. The decrease in CLK⅐CYC-dependent transcripts in EP(X)1149 flies is proportional to the decrease in nej levels (Fig.  4, A, C, and E). These experiments further demonstrate a coactivator function of NEJ for CLK⅐CYC activity, and they show that trans-activation by CLK⅐CYC responds to changes in limiting cellular levels of NEJ.
Consistent with our results, a parallel study by Lim et al. (34) found a role for nej in the Drosophila circadian clock. From the finding of low expression levels of CLK⅐CYC-dependent genes after constitutive overexpression of nej it was concluded that NEJ acted as a repressor on CLK⅐CYC function. We show, however, that a reduction in nej levels results in a proportional decrease in CLK⅐CYC activity in flies as well as in cell culture, and induction of nej expression in flies induces CLK⅐CYC-dependent transcripts, which demonstrates that NEJ acts as a co-activator for CLK⅐CYC function. Reduced levels of CLK⅐CYC-dependent genes observed after constitutive overexpression of nej (see also PER levels in Fig.  2 and supplemental Fig. S2) are likely because of circadian feedback regulation that compensates for hyper-activation of CLK⅐CYC by NEJ.
The involvement of NEJ in the functional organization of the Drosophila circadian clock adds additional modes to the regulation of circadian transcription. The finding that moderately reduced or increased levels of NEJ decrease or enhance CLK⅐CYC-dependent transcription, respectively, demonstrates that CLK⅐CYC activity is sensitive to limiting levels of NEJ, and availability of NEJ regulates circadian transcription ( Figs. 1 and 4). Studies on transcriptional activation by AP1, nuclear hormone receptors, the JAK/STAT pathway, and E2F have shown that multiple transcription factors compete for recruitment of the mammalian ortholog CBP from a limiting pool, thereby establishing negative cross-talk between these pathways (14 -16). Our results suggest that CLK⅐CYC joints this competition for recruitment of NEJ (or CBP in mammals) allowing a cross-talk between circadian transcription and NEJ/CBP-dependent transcriptional regulation of metabolism, development, cell proliferation, memory formation, and other physiological processes. In addition it has been shown that a number of signaling pathways such as cyclic nucleotides, calcium, and Ras/ MAPK regulate NEJ-dependent transcription through both, direct phosphorylation of NEJ as well as phosphorylationdependent recruitment of NEJ to target transcription factors (35). The regulation of CLK⅐CYC-dependent transcription by these signaling pathways, is in addition to direct phosphorylation of CLK (13), likely mediated through the control of NEJ co-activator function. NEJ thereby provides an interface for cross-talk between circadian transcription and vital physiological processes, which likely assists the circadian orchestration of life.