CAVIN-3 regulates circadian period length and PER:CRY protein abundance and interactions

Abstract

In mammals, transcriptional autorepression by Period (PER) and Cryptochrome (CRY) protein complexes is essential for the generation of circadian rhythms. We have identified CAVIN-3 as a new, cytoplasmic PER2-interacting protein influencing circadian clock properties. Thus, CAVIN-3 loss- and gain-of-function shortened and lengthened, respectively, the circadian period in fibroblasts and affected PER:CRY protein abundance and interaction. While depletion of protein kinase Cδ (PKCδ), a known partner of CAVIN-3, had little effect on circadian gene expression, CAVIN-3 required the PKCδ-binding site to exert its effect on period length. This suggests the involvement of yet uncharacterized protein kinases. Finally, CAVIN-3 activity in circadian gene expression was independent of caveolae.

Figure 1

CAVIN-3 is a new PER2 interaction partner. (A) Schematic representation of TAP-tagged proteins used for the purification. (B) Immunoblot showing the serum induction of PER2-TAP and TAP-LUCIFERASE in the stable cell lines, and endogenous PER2 in NIH3T3 cells. U2AF65 served as a loading control. (C) Silver stain of purified protein complexes separated by 8–16% SDS–polyacrylamide gel electrophoresis. (D) Co-immunprecipitation of V5-tagged PER2 with HA-CAVIN-3. The pCI-HA empty vector was used as a negative control. A negative control IP was performed with beads devoid of antibodies. (E) Co-immunoprecipitation of HA-CAVIN-3 with FLAG-PER2. V5 antibody was used as an irrelevant mouse monoclonal antibody (negative control). (F) CAVIN-3 was immunoprecipitated from Per2::Luc cells with anti-CAVIN-3 serum. Co-immunoprecipitated PER2::Luciferase was assessed as luciferase activity. The fold enrichment represents the luciferase activity normalized to the signal from IPs with pre-immune serum. CMV-luc stable cells served as a control. (G) HA-CAVIN-3 localizes to the cytoplasm. HA-CAVIN-3-expressing synchronized NIH3T3 cells were stained with HA and PER2 antibodies. Nuclei were stained with DAPI. Merge: DAPI and α-HA straining. DAPI, 4,6-diamidino-2-phenylindole; HA, haemagglutinin; IB, immunoblotting; IP, immunoprecipitation; PER, Period.

Figure 2

CAVIN-3 loss- and gain-of-function affects the circadian period length. (A) NIH3T3 cells were transiently co-transfected with the Bmal1-luciferase reporter and either an shRNA (hp1) targeting Cavin-3 mRNA (blue) or the empty plasmid (black). Individual oscillators were synchronized using forskolin. Data from duplicate transfections are shown. (B) Experiment as in A, but using another shRNA targeting Cavin-3 (hp7, red). (C) Bmal1-Luc bioluminescence recordings after co-transfection of an HA-CAVIN-3 expression vector (blue, in triplicates) or an empty vector control (black, duplicates). (D) Bioluminescence rhythms measured from Per2::luc primary cells transduced with HA-CAVIN-3 (blue) or GFP (black, control) expressing lentiviral vectors. The data were filtered with a low-pass filter (filfilt function in Matlab). Raw data were detrended using moving average transformation (window: 24 h). (E) Period length changes in CAVIN-3 loss- (hp1) and gain-of-function experiments normalized to values obtained with control cells. Mean±s.d., n?6. (F) shRNA-knockdown efficiencies on endogenous Cavin-3 mRNA analysed by quantitative PCR after FACS sorting of GFP-positive cells (encoded on shRNA vectors). (G) Bioluminescence recordings of cells expressing Bmal1-luciferase and Cavin-3 hp1 (blue) or the empty vector (black). Cells were synchronized by simulated body temperature rhythms (grey). The result of one representative experiment (out of three) is shown. Raw data were detrended using moving average transformation (window: 24 h). GFP, green fluorescent protein; HA, haemagglutinin; PER, Period; shRNA, short-hairpin RNA.

Figure 3

CAVIN-3 influences PER and CRY expression levels and affects PER:CRY complex interactions. (A) Schematic representation of the mammalian two-hybrid system used to analyse PER:CRY interactions. (B) The PER2:CRY2 interaction measured after serum treatment (black). CAVIN-3 co-expression (blue) dramatically increased the PER2:CRY2 signal. Empty vectors are in grey. (C) Dual luciferase assay quantifying the effect of CAVIN-3 co-expression on the PER2:CRY2 interaction. Values represent FF luciferase signals normalized to RL luciferase (internal control) measured 40 h after synchronization. The Y14:Magoh interaction was used as a specificity control. Mean±s.d., n=10. (D) As in C, but quantifying the effect of Cavin-3 knockdown. Mean±s.d., n=5. (E) Effects of CAVIN-3 co-expression on PER2:CRY1, PER1:CRY2 and PER1:CRY1 interactions. Mean±s.d., n=5. (F) GAL4-V5-PER2 and NF-κB-CRY2 protein expression upon HA-CAVIN-3 overexpression and knockdown, analysed by immunoblotting using V5 and NF-κB antibodies, respectively. U2AF65 served as a loading control. CRY, Cryptochrome; FF, Firefly; HA, haemagglutinin; PER, Period; RL, Renilla.

Figure 4

CAVIN-3 requires its PKC-binding site, but Pkcδ and Caveolin1 seem not to be involved in setting the circadian period length. (A) Bmal1-luciferase reporter rhythms in NIH3T3 cells expressing a dominant-negative PKCδ (PKCδ-DN, green) or the empty vector (black). Raw data were detrended using moving average transformation. (B) Bmal1-luciferase rhythms in cells co-transfected with a vector encoding a Pkcδ-specific shRNA (green) or an empty shRNA vector (black). Raw data were detrended using moving average transformation. (C) Pkcδ-knockdown efficiency measured by quantitative PCR. (D) Expression levels of HA-CAVIN-3 and HA-CAVIN-3-mut assessed by anti-HA immunoblotting. (E) Bmal1-luciferase reporter rhythms in dexamethasone-shocked NIH3T3 cells co-transfected with plasmids encoding HA-CAVIN-3 (dark blue), HA-CAVIN-3-mut (light blue) or the empty plasmid (black). Data from duplicate transfections are shown. (F) Dual luciferase assay quantifying the effect of HA-CAVIN-3 and HA-CAVIN-3-mut co-expression on the PER2:CRY2 interaction. The assay was performed as in Fig 3. Mean±s.d., n=3. (G) Bioluminescence recorded in Bmal1-luciferase lentivector-transduced primary fibroblasts prepared from wt (black), heterozygous (pink) and homozygous (orange) Caveolin1-knockout animals. Raw data were detrended using moving average transformation. CRY, Cryptochrome; HA, haemagglutinin; PER, Period; PKC, protein kinase C; PKCδ-DN, dominant-negative PKCδ variant; shRNA, short-hairpin RNA.