Transcription factor CBF-1 is critical for circadian gene expression by modulating WHITE COLLAR complex recruitment to the frq locus

Abstract

Transcription of the Neurospora crassa circadian clock gene frequency (frq) is an essential process in the negative feedback loop that controls circadian rhythms. WHITE COLLAR 1 (WC-1) and WHITE COLLAR 2 (WC-2) forms the WC complex (WCC) that is the main activator of frq transcription by binding to its promoter. Here, we show that Centromere Binding Factor 1 (CBF-1) is a critical component of the N. crassa circadian clock by regulating frq transcription. Deletion of cbf-1 resulted in long period and low amplitude rhythms, whereas overexpression of CBF-1 abolished the circadian rhythms. Loss of CBF-1 resulted in WC-independent FRQ expression and suppression of WCC activity. As WCC, CBF-1 also binds to the C-box at the frq promoter. Overexpression of CBF-1 impaired WCC binding to the C-box to suppress frq transcription. Together, our results suggest that the proper level of CBF-1 is critical for circadian clock function by suppressing WC-independent FRQ expression and by regulating WCC binding to the frq promoter.

Fig 1. Deletion of the cbf-1 gene results in long period and low amplitude of circadian rhythms.

(A) Race tube analyses of the cbf-1^KO strains. (B) Luciferase activity of wt, frq-luc and cbf-1^KO, frq-luc strains. The low normalized luciferase signal levels in the cbf-1^KO, frq-luc strain reflect the low-amplitude fluctuation of luciferase activity. Raw data were normalized to subtract the baseline calculated by LumiCycle analysis software. (C) Amino acid sequence alignment of the HLH domain from the Neurospora CBF-1, Saccharomyces CBF-1, Drosophila USF1, mouse USF1, and human USF1. The HLH domain is composed of H₁, Loop and H₂. The asterisk indicates the conserved Glu (E9) in the basic domains of HLH proteins. (D) Western blot analyses of the levels of FRQ protein in the wild-type and cbf-1^KO strains. Asterisks indicate nonspecific bands. Samples were grown in DD for indicated hours before harvest. PVDF membrane stained with Coomassie blue (mem) was used as a loading control. Quantification of the levels of FRQ protein is shown beside the western blot. (E) Northern blot analyses of the levels of frq mRNA. Ribosome RNA (rRNA) bands stained by ethidium bromide shown below the northern blot acted as a loading control for each sample. Quantification of the levels of frq mRNA is shown beside the northern blot.

Fig 2. Overexpression of CBF-1 in the wild-type strain abolishes conidiation rhythmicity.

(A) The leaky expression of CBF-1 can partially rescue the cbf-1^KO phenotype. The highly inducible Myc-CBF-1 caused an arrhythmic conidiation phenotype. (B) The QA-regulated phenotype of wild-type and wt, qa-Myc-CBF-1 strains is shown on race tubes containing variable amounts of the inducer QA (0-10^-2 M). (C) Luciferase reporter assays showing the normalized frq promoter activity of wt, frq-luc and wt, qa-Myc-CBF-1, frq-luc strains. (D) Western blot analyses of the levels of FRQ protein in the wild-type and wt, qa-Myc-CBF-1 strains. Asterisks indicate nonspecific bands. PVDF membranes stained with Coomassie blue (mem) demonstrate equal amounts of protein loaded per lane. Quantification of the levels of FRQ protein is shown beside the western blot. (E) Northern blot analyses of the levels of frq mRNA in the wild-type and wt, qa-Myc-CBF-1 strains. rRNA bands stained with ethidium bromide shown below the northern blot served as a loading control for each sample. Quantification of the levels of frq mRNA is shown beside the northern blot.

Fig 3. CBF-1 rhythmically associates with the C-box of the frq promoter.

(A) EMSA assays showing GST-CBF-1 specifically binds to the C-box. A lower concentration of protein was utilized when protein was incubated with specific and non-specific probe. (B) Western blot analyses showing that polyclonal antibody specifically recognizes CBF-1 protein in the wild-type strain but not in the cbf-1^KO strain. The arrow indicates the CBF-1 protein band detected by CBF-1 polyclonal antibody. WT, wild-type; MW, molecular weight. (C) CBF-1 rhythmically binds to the C-box of frq promoter. Samples were grown for the indicated hours in DD prior to harvest for ChIP. (D) The levels of CBF-1 protein are consistent across circadian time points. (E) Rhythmic binding of WC-2 to the C-box is constantly high in the frq⁹ mutant. (F) Rhythmic binding of CBF-1 to the C-box is disrupted in wc-1, wc-2, and frq⁹ mutants. (G) The levels of CBF-1 protein were not altered in wc-1, wc-2, or frq⁹ mutants. Significance was assessed by using a two-tailed t-test. Data are means ± standard deviations (S.D.) (n = 3; **P<0.01 and ***P<0.001).

Fig 4. CBF-1 suppresses WC-independent frq transcription by promoting the recruitment of the transcriptional co-repressor RCM-1.

(A, B) ChIP assays showing the enrichment of A) WC-1 and B) WC-2 is decreased and the peak of WC-1 binding is delayed in the cbf-1^KO strain compared to the wild-type strain. (C) The phosphorylation of WC-1 and WC-2 was markedly increased over time in the cbf-1^KO strain compared to the wild-type strain. (D) Western blot analyses were performed using antibodies against FRQ or WC-2 in the wild-type, cbf-1^KO, wc-2^KO, and cbf-1^KO wc-2^KO strains. Samples were grown under constant light (LL) or in DD for indicated hours before harvest. (E) Northern blot analyses of the levels of frq mRNA in wild-type, cbf-1^KO, wc-2^KO, and cbf-1^KO wc-2^KO strains. (F) The enrichment of RCM-1 at the C-box of frq promoter is reduced in cbf-1^KO strains at the DD14 compared to the wild-type strain. (G) The levels of RCM-1 phosphorylation of the cbf-1^KO strain are higher than those in wild-type strain at the indicated times. (H) The RCM-1 protein levels are similar to that in the cbf-1^KO and wild-type strains. The errors bars ±S.D. (n = 3; **P<0.01 and ***P<0.001; two-tailed t-test).

Fig 5. CBF-1 suppresses WCC binding without FRQ-feedback inhibition.

(A) The recruitment of WC-2 to the C-box of the frq promoter in the wild-type, wc-2^KO, frq⁹, cbf-1^KO, and cbf-1^KO frq⁹ strains at DD14. (B) Northern blot analysis of the levels of frq mRNA in wild-type, wc-2^KO, frq⁹, cbf-1^KO and cbf-1^KO frq⁹ strains. (C) The phosphorylation levels of WC-1 and WC-2 do not change between frq⁹ and cbf-1^KO frq⁹ strains. Plotted are means ±S.D. (n = 3; **P<0.01 and ***P<0.001; two-tailed t-test).

Fig 6. Conserved HLH domain of CBF-1 is required for normal circadian gene expression.

(A) Race tube assays showing that the defect of circadian conidiation rhythms in cbf-1 mutants can be rescued by wild-type CBF-1 but not mutant CBF-1 proteins. (B) The enrichment of CBF-1 at the C-box is abolished in mutants of CBF-1 defective in DNA binding. (C, D) ChIP assays of C) WC-1 and D) WC-2 were performed in the indicated strains at DD14. (E, F) The phosphorylation levels of E) WC-1 and F) WC-2 were markedly increased in mutants of CBF-1 defective in DNA binding. Plotted are means ±S.D. (n = 3; **P<0.01 and ***P<0.001; two-tailed t-test).

Fig 7. Overexpression of CBF-1 decreases recruitment of WCC to the C-box.

(A, B) Rhythmic binding of A) WC-1 and B) WC-2 to the C-box is decreased by overexpression of Myc-CBF-1. (C) ChIP assays showing that the binding of CBF-1 to C-box increased in wild-type strain with overexpression of CBF-1. (D) Race tube assays showing that overexpression of the mutant CBF-1 has no effect on circadian conidiation phenotype. (E, F) ChIP assays showing that the binding of E) WC-1 and F) WC-2 decreased in the wild-type strain upon overexpression of the wild-type CBF-1 but not mutant CBF-1 at DD14. All error bars represent the ±S.D. (n = 3; *P<0.05, **P<0.01, and ***P<0.001; two-tailed t-test). (G, H) Immunoprecipitation assays using G) WC-1 antiserum and H) WC-2 antiserum showing that different versions of CBF-1 proteins interact with WC-1 and WC-2 in vivo.

Fig 8. A model of CBF-1-mediated regulation of frq transcription.

In the wild-type strain, WCC and CBF-1 bind to the C-box site in the regulatory region of the frq gene and cooperate with each other to ensure rhythmic expression of FRQ. In the cbf-1^KO mutant, the delay in WCC binding to the C-box may lead to long periods without CBF-1. WC-dependent and WC-independent frq transcription both contribute to the levels of frq mRNA. In the CBF-1 overexpression strain, elevated CBF-1 binding to the C-box decreases the recruitment of WCC, which leads to frq mRNA oscillation at a low level.