Control of WHITE COLLAR localization by phosphorylation is a critical step in the circadian negative feedback process

Abstract

Reversible protein phosphorylation has critical functions in the eukaryotic circadian negative feedback loops. In Neurospora, the FREQUENCY protein closes the circadian negative feedback loop by promoting the phosphorylation of its transcription activator, the WHITE COLLAR complex (WCC) and consequently inhibiting WCC activity. Here we show that protein phosphatase 4 is a novel component of the Neurospora clock by regulating both processes of the circadian negative feedback loop. The disruption of pp4 results in short period rhythms with low amplitude. In addition to its role in regulating FRQ phosphorylation and stability, PP4 also dephosphorylates and activates WCC. In contrast to PP2A, another phosphatase that activates WCC, PP4 has a major function in promoting nuclear entry of WCC. PKA, a WC kinase, inhibits WC nuclear localization. Furthermore, the FRQ-dependent WC phosphorylation promotes WCC cytosolic localization. Together, these results revealed WCC nucleocytoplasmic shuttling as an important step in the circadian negative feedback process and delineated the FRQ-dependent WCC inhibition as a two-step process: the inhibition of WCC DNA-binding activity followed by sequestration of WCC into the cytoplasm.

Figure 1

Disruption of pp4 results in a short period and low amplitude circadian conidiation rhythm. Race tube assays showing the conidiation rhythms of the indicated strains in DD (A) and LD 12/12 (B) conditions. The period length or phase of the strains are indicated on the right.

Figure 2

Low amplitude or arrhythmic FRQ expression in the pp4^KO strains. (A) Western blot analysis showing that FRQ oscillates with a low amplitude and short period in the pp4^KO strain. (B) Western blot analysis showing the loss of FRQ oscillation in the pp4^KO ppp-1^RIP strain. The densitometric analyses of these experiments are shown below. A constantly expressed protein was indicated by the asterisks.

Figure 3

FRQ is hyperphosphorylated and rapidly degraded in the pp4^KO strains. (A) Western blot analysis showing that FRQ levels are low and hyperphosphorylated in the pp4^KO and pp4^KO ppp-1^RIP strains. (B) Mobility shifts of FRQ protein in SDS–PAGE could be reversed by λ phosphatase treatments. (C) Western blot analysis showing the degradation of FRQ after a LD transition in the indicated strains. The densitometric analyses from three independent experiments are shown below. (D) Immunoprecipitation assay showing that FRQ associates with PP4 in vivo. IP using preimmune (PI) serum was used as the negative control.

Figure 4

Hyperphosphorylation of WC-2 and reduced WCC activity in the pp4^KO strain. (A) Northern blot analyses showing the frq mRNA expression levels in DD in the indicated strains. (B) Western blot analysis showing that WC-2 is hyperphosphorylated in the pp4^KO strain. To analyse the phosphorylation profile of WC-2, 10% SDS–PAGE gels containing a ratio of 139:1 acrylamide/bisacrylamide was used. (C) Western blot analysis showing that WC-1 and WC-2 are hyperphosphorylated in the pp4^KO strain. Strains were cultured in LL. λ phosphatase treatment showed that the mobility shifts were due to phosphorylations. (D) The results of ChIP assays showing that WC-2 binds to the frq C-box at low levels in the pp4^KO strain. Three independent experiments were performed and the error bars indicate the standard deviations (^*P<0.05).

Figure 5

The loss of nuclear enrichment of WCC in the pp4^KO strain. Western blot analyses showing the levels of WC-1 and WC-2 in the total extracts, nuclear or cytosolic fractions in the pp4^KO (A) and rgb-1^RIP (B) strains. The tubulin was used to show that the nuclear fractions were free of cytosolic proteins. The asterisk indicates a protein band presented in both nuclear and cytosolic fractions non-specifically detected by our WC-1 antibody. Densitometric analyses from three independent experiments were shown for pp4^KO strain (^*P<0.05, ^**P<0.01). The error bars indicate the standard deviations. (C) Western blot analyses showing the levels of Myc-tagged PPP-1, PP4 and PP2A catalytic subunits in different cellular fractions. Cultures were grown in LL. (D) Western blot analysis comparing the levels of WC-1 and WC-2 in different cellular fractions between the wild-type and pkaC^KO strains in LL. Note that pkaC^KO has much lower levels of WC proteins in the total extract than the wild-type, but they had comparable WC amounts in the nuclear fractions. The densitometric analyses from three independent experiments were shown (^**P<0.01).

Figure 6

Phosphorylation and temporal-regulated WCC cellular localization. (A, B) Western blot analyses showing the levels of WC-1 and WC-2 in different cellular fractions. In these gels, the amounts of the nuclear extracts loaded were 1/3 of those in total and cytosolic fractions. Wild-type cultures grown in LL and DD14 were used in (A). In (B), wild-type cultures grown in DD14 and DD24 were compared. The densitometric analyses from three independent experiments were shown below (^*P<0.05). The error bars indicate the standard deviations. (C) Western blot analyses comparing the levels of WC-1 and WC-2 in different cellular fractions between the wild-type and frq¹⁰ strains. The densitometric analyses from three independent experiments were shown below (^*P<0.05). (D) ChIP assay using WC-2 antibody comparing the levels of WCC binding to the frq C box in the wild-type, WT,qa-FRQ and frq⁹,qa-FRQ strains. For experiments in (C) & (D), cultures were harvested at DD24.

Figure 7

A current model of the Neurospora circadian negative feedback loop. WCC functions as the positive element of the loop, whereas FRQ and FRH forms the FFC complex to function as the negative element. FFC inhibits WCC activity by recruiting casein kinases to phosphorylate WCC, resulting the removal of WCC from the DNA and its sequestration in the cytoplasm. To restart a circadian cycle, PP4 promotes the nuclear entry of WCC, whereas PP2A activates WCC DNA-binding activity. The casein kinases also phosphorylate FRQ, which promotes FRQ degradation through the ubiquitin-proteasome pathway. PP1 and PP4 counter the role of the casein kinases and stabilize FRQ. The stability of FRQ determines the period length of the clock. A full-colour version of this figure is available at The EMBO Journal Online.