Subcellular dynamics of overexpressed PER2 and CRY1

(A) PER2-mNG exhibits rapid nucleocytoplasmic shuttling with a preference for the cytoplasm at early timepoints. DOX-induced expression of PER2-mNG in absence (DMSO) and presence of LMB. Scale bar 25 µm. n = 3. (B) Expression of PER2-mNG supports accumulation of endogenous CRY1 with kinetics correlating with nuclear accumulation of PER2-mNG. Western blot analysis of U2OStx cells transiently transfected and DOX-induced with PER2-mNG sampled at various time points post-DOX induction. (The accumulation of endogenous CRY1 in PER2-mNG-transfected cells is much higher than it appears because the transfection efficiency of U2OStx cells is <20%.) n = 3. (C) Induction of stably-integrated V5-PER2 supports accumulation of endogenous CRY1 with kinetics correlating with nuclear accumulation of PER2-mNG. Western blot analysis of V5-PER2 and CRY1 from DOX-induced U2OStx_V5-PER2 cells at the indicated time points post induction. n = 3. (D) CRY1 and PER2 mutually affect their subcellular dynamics and localization. DOX-induced transiently expressed mK2-CRY1 (top row) and PER2-mNG (bottom row) individually and in combination was analyzed over 28 h. Co-expression was assessed at mK2-CRY1 : PER2-mNG ratios of 3:1 (second row), 1:1 (third row), and 1:3 (fourth row). Scale bar 25 µm. n = 3. (E) Model of subcellular localization of PER2, CRY1, and PER2:CRY1 complexes based on their relative stoichiometries. Left: PER2 dimers, which undergo rapid nuclear–cytoplasmic shuttling, are predominantly localized in the cytoplasm, whereas CRY1 is mainly nuclear. Right: PER2 dimers bound to a single CRY1 molecule remain mainly cytoplasmic, while PER2 dimers bound to two CRY1 molecules localize to the nucleus.

PERs and CRYs in nuclear foci retain their interaction specificity

(A) Schematic of the experimental setup: U2OStx_mK2–CRY1 cells were transfected with PER2 and induced with DOX. Cells were analyzed for mK2–CRY1 expression and by immunofluorescence (IF) using specific antibodies. (B) IF for PER2: mK2–CRY1 and PER2 co-accumulate in nuclear foci. (C) IF for PER1: nuclear foci contain endogenous PER1. (D) IF for BMAL1: nuclear foci contain endogenous BMAL1. (E) IF for CK1δ: nuclear foci contain endogenous CK1δ. Expression levels of endogenous clock proteins are elevated in transfected cells compared to surrounding untransfected cells. Scale bars: 10 µm. n = 3.

CK1δ is stability regulated in an activity-dependent manner

(A) qPCR analysis of CK1δ mRNA in U2OStx cells and stable U2OStx cell lines expressing DOX-inducible CK1δ1, CK1δ1-K38R, and CK1δ1-R178Q. n = 3, Students t-test, ns: not significant. n = 3. (B) Immunoblot analysis of DOX-induced CK1δ variants. Upper panel: DOX-induced expression of indicated FLAG-tagged CK1δ1 variants in stable U2OStx cell lines. Lower panel: Expression of endogenous CK1δ (open circle) and CK1δ1-FLAG variants (filled circle) analyzed with an antibody detecting CK1δ1 and CK1δ2 but not CK1ε (Figure EV2). n = 3. (C) Overexpressed CK1δ1 is unstable with a half-life of about 6 min. CK1δ1-K38R and CK1δ1-R178Q are stable over 90 min. U2OStx_CK1δ1 cells were DOX-treated for 24 h and then treated with CHX. Samples were collected after the indicated time periods (CHX chase) and analyzed by immunoblotting with FLAG antibody. (D) Endogenous CK1δ is stable over 90 min post CHX chase. (E) Quantification of (C) and (D). n = 3, solid and dashed lines: mean ± SD. (F) Left panel: Inhibition by PF670 stabilizes overexpressed FLAG-tagged CK1δ1. Right panel: 1h PF670 treatment of control cells has no effect on endogenous CK1δ levels. More protein (4x) was loaded on the gel for the analysis of endogenous CK1δ. n = 3. (G) Proteasomal inhibition by MG132 (4 h) stabilizes overexpressed (left panel) CK1δ1 and, (middle panel) to a lesser extent, the already stable CK1δ1-K38R. Right panel: MG132 has no effect on endogenous CK1δ levels in U2OStx control cells within the 4 h treatment window. n = 3.

PER2 supports accumulation of endogenous CK1δ

(A) Endogenous CK1δ co-accumulates with overexpressed V5-PER2. Stable U2OStx_V5-PER2 cells were DOX-induced and analyzed after the indicated time periods for expression of (upper panel) V5-PER2 and (lower panel) endogenous CK1δ. n = 3. (B) Overexpressed CK1δ1 is stabilized dependent on PER2 dosage and PER-CK1δ interaction. HEK293T co-expression of CK1δ1-FLAG (200 ng) with increasing V5-PER2 results in V5-PER2 dose-dependent stabilization of CK1δ1-FLAG. Coexpression with V5-PER2ΔCKBD does not substantially stabilize CK1δ1-FLAG. All samples are from the same gel and Western blot with the same exposure time. Dashed vertical lines indicate that gel lanes with lower PER2 levels and size marker lanes were spliced out. n = 3. (C) PER2 stabilizes CRY1 and CK1δ1. HEK293T cell expression of indicated combinations of V5-PER2, mK2-CRY1, and CK1δ1-FLAG. n = 3. (D) Phosphatase inhibition by CalA promotes phosphorylation of PER2 as well as PER-bound CK1δ and affects levels of PER2-bound CRY1. HEK293T cell expression of indicated combinations of V5-PER2, CRY1-FLAG, and CK1δ1-FLAG with and without CalA treatment. n = 3. Upper panel: CalA promotes hyperphosphorylation of PER2. Middle panel: Levels of PER2-stabilized CRY1 correlate inversely with PER2 phosphorylation state. Lower panels: CalA promotes phosphorylation of free CK1δ1 (lanes 1 and 2) as well as of PER2-bound (stabilized) overexpressed CK1δ1 (lanes 3 and 4) and PER2-bound endogenous CK1δ (lanes 5, 6 and 7). Note, the PER2-dependent stabilization of endogenous CK1δ by comparing with untransfected HEK293T cell control.

Active kinase is primarily degraded in the nucleus

(A) Endogenous CK1δ exhibits diffuse cellular localization with one distinct spot likely corresponding to the centrosome55. U2OStx was stained with DAPI (blue) and an antibody against CK1δ (green; AF488). Scale bar 20 µm. (B) Immunofluorescence of U2OStx cell lines overexpressing FLAG-tagged CK1δ1, CK1δ1-K38R and CK1δ1-R178Q. Cells were DOX-treated for 24 h and stained via FLAG antibody (green; AF488) and with DAPI (blue). Scale bar 20 µm. n = 3. (C) Proteasomal inhibition by MG132 results in accumulation of overexpressed CK1δ1 primarily in nucleus. Stable U2OStx_CK1δ1 cells were DOX-induced for 24 h and treated with 5 µM MG132 for 4 h. Left: IF. Right: Quantification of CK1δ1 levels in MG132-treated cells versus vehicle (DMSO) control. n ≥ 50 cells; mean ± SD; ****: p < 0.0001. Scale bar 20 µm. (D) Kinase inhibition results in accumulation of overexpressed FLAG-tagged CK1δ specifically in the nucleus. Stable U2OStx_CK1δ cells were DOX induced for 24 h and treated with 1 µM PF670 for 1 h. Left: IF. Right: Quantification of CK1δ1 levels in PF670-treated cells versus vehicle (H2O) control. n ≥ 50 cells; mean ± SD; **: p < 0.01. Scale bar 20 µm. (E) NES-tagged FLAG-CK1δ1-K38R accumulates at elevated levels in the cytoplasm and localizes to the centrosome. U2OStx cells were transiently transfected with either FLAG-CK1δ-K38R or NES-FLAG-CK1δ1-K38R. Left: FLAG-IF. Right: Quantification of kinase levels. n ≥ 20 cells; mean ± SD; ****: p < 0.0001. Scale bar 20 µm.

CK1δ facilitates cytoplasmic accumulation of PER2 and release of CRY1

(A) Transient transfection and coexpression of PER2-mNG and mK2-CRY1 in U2OStx_CK1δ1 and U2OStx_CK1δ1-K38R stable cell lines. Expression was DOX-induced 24 h post transfection and cells were then analyzed for 24 h with an Incucyte system. PER2-mNG and mK2-CRY1 accumulate in nuclear foci when coexpressed with inactive CK1δ1-K38R (upper row) and in cytoplasmic foci when coexpressed with active CK1δ1 (lower row). Scale bar 20 µm. (B) Cytoplasmic PER2-mNG foci (CK1δ1 cells) contain less mK2-CRY1 than nuclear foci (CK1δ1-K38R cells). Scale bar 20 µm. Right: Quantification of CRY1 : PER2 ratio in cytoplasmic and nuclear foci. n > 30 foci, mean ± SD; ****: p < 0.0001. (C) PER2 phosphorylation attenuates interaction with CRY1. V5-PER2 and CRY1-FLAG were transiently expressed in HEK293T. Cell lysate was prepared and incubated with and without recombinant CK1δΔtail for 4 h. V5-PER was then immunoprecipitated with V5 beads. Lysate (Input) and an 8x equivalent of the IP were analyzed by Western-blot. FLAG-CRY1 co-immunoprecipitated with hypophosphorylated but not with hyperphosphorylated V5-PER2. n = 3. (D) Cytoplasmic PER:CRY foci persistently release CRY1 into the nucleus in a CK1δ1-dependent fashion. Scale bar 20 µm. Central cell: Representative U2OStx_mK2-CRY1 cell co-transfected with PER2-mNG and CK1δ1, resulting in the formation of cytoplasmic PER:CRY foci. Over time, PER2-mNG levels decrease, while mK2-CRY1 accumulates in the nucleus and remains at persistently high levels, even 36 hours post DOX induction (see Fig EV6 for more examples). Surrounding cells: U2OStx_mK2-CRY1 cells not transfected with PER2-mNG. In these cells, mK2-CRY1 accumulates in the nucleus initially, but its levels decrease over time, likely due to DOX metabolism, with low levels observed 36 hours post DOX induction. (E) CK1δ inhibition by PF670 facilitated the nuclear accumulation of hypophosphorylated PER2. V5-tagged PER2 was transiently expressed in HEK293T cells together with FLAG-tagged CRY1 and CK1δ1. PF670 was added, and the subcellular localization of PER2 was analyzed at the indicated time points following PF670 treatment. (F) Endogenous PER2 subcellular redistribution from the nucleus to the cytosol in U2OStx cells was analyzed over a circadian cycle. U2OStx cells were synchronized with DEX at t = 0 h. After 4 and 8 h, endogenous PER2 was enriched in the nucleus. At later timepoints, a fraction of PER2 was localized in the cytosol.

Nuclear CK1δ affects circadian period length

(A) Schematic of mNG-V5-NLS-CK1δ1 and mNG-V5-NES-CK1δ1 constructs NLS: SV4057; NES: HIV-160. (B) Left: mNG-NLS-CK1δ1 is expressed at a very low level in the nucleus and is stabilized by MG132. B/C-adjusted panel: brightness and contrast adjusted to better visualize the kinase. Right: quantification. n > 30 cells, mean ± SD; ****: p < 0.0001. Scale bar 20 µm. (C) Left: mNG-NES-CK1δ localizes to the cytosol and is stabilized by MG132. Right: quantification. n > 30 cells, mean ± SD; ***: p < 0.001. Scale bar 20 µm. (D) Stable U2OStx lines expressing comparable levels of DOX-induced V5-tagged mNG-NLS-CK1δ1 or mNG-NES-CK1δ1. n = 3. (E) DOX-induced expression of NLS-CK1δ1 results in significant period shortening whereas NES-CK1δ1 has no effect on period. Bmal1 promoter-luciferase assay. Representative bioluminescence traces are shown as a detrended average of 4 technical replicates of PBS- and DOX-treated cells. Resulting period was calculated using cosinor analysis built-in GraphPad Prism 10. n = 3. (F) Circadian period change comparing DOX- vs. PBS-treated mNG-NLS-CK1δ1 or mNG-NES-CK1δ1 cell lines. n = 3, mean ± SD; **: p < 0.01.

Subcellular dynamics of overexpressed PER2 and CRY1.

(A) Constructs used in this study: FLAG–PER2–mNeonGreen (PER2–mNG) and 6×His–mKate2–CRY1 (mK2–CRY1). (B) Individual channels from the middle rows of Fig 1C, showing the localization of mK2–CRY1 (magenta) and PER2–mNG (green), respectively. Scale bar: 20 µm.

Analysis of untransfected U2OStx_mK2-CRY1 cells for control of PER2 transfected cells shown in Figure 2

Untransfected U2OStx_mK2-CRY1 cells displays predominantly homogeneous nuclear localization of mK2-CRY1 and (A) cytoplasmic and nuclear localization of endogenous PER2. (B) cytoplasmic and slightly enriched nuclear localization of endogenous PER1. (C) nuclear localization of endogenous BMAL1. (D) diffuse nuclear and cytoplasmic localization of endogenous CK1δ as well as localization at the centrosome. Scale bars 10 µm. n = 3.

Endogenous CK1δ levels decrease upon overexpression of transgenic CK1δ

(A) CK1δ antibody (abcam, Cat# ab85320) recognizes CK1δ1 and CK1δ2 but not CK1ε. FLAG-tagged kinases were expressed in HEK293T cells. Untr.: untransfected control. Cell lysates were analyzed by Western-blot with antibody ab85320 (upper panel) and FLAG antibody (lower panel). n = 3. (B) Overexpression of CK1δ1-FLAG and (C) mNG-CK1δ1 reduce expression levels of endogenous CK1δ. Western blots with CK1δ antibody ab85320. n = 3.

PER2 supports stabilization of mutant kinases

PER2 stabilizes CRY as well as CK1δ1-K38R and CK1δ1-R178Q, which are in absence of PER2 already more stable than CK1δ1 (see Fig 4c, from the same blot). HEK293T cell expression of indicated combinations of V5-PER2 and FLAG-tagged mK2CRY1, CK1δ1-K38R and CK1δ1-R178Q. n = 3.

Impact of proteasome inhibition and kinase inhibition on endogenous CK1δ abundance and localization

(A) Proteasome inhibition by 5 µM MG132 for 4 h and (B) Kinase inhibition by 1 µM PF670 for 1 h does not result in accumulation of endogenous CK1δ. Scale bars 20 µm. Right figure parts: Quantifications of (A) and (B) show no significant change in abundance of endogenous CK1δ in inhibitor-treated cells compared to the vehicle control. n ≥ 50 cells, mean ± SD. (C) Schematic of CK1δ homeostasis. For details see main text.

Active CK1δ promotes the formation of cytoplasmic PER2:CRY1 foci, serving as reservoirs for CRY1 release into the nucleus

(A) PER2-mNG alone accumulates slowly in the nucleus when expressed with inactive CK1δ-K38R and in the cytoplasm when expressed with active CK1δ. n = 3. (B) Relates to Fig 6A. Additional fields of view showing cytoplasmic PER:CRY foci in U2OStx_CK1δ1 cells transfected with PER2-mNG and mK2-CRY1. (C) Relates to Fig 6D. Additional fields of view showing that cytoplasmic PER2-CRY1 foci serve as a reservoir for the CK1δ1-dependent release of mK2-CRY1 into the nucleus after degradation of PER2-mNG. Scale bars 20 µm. (D) Relates to Fig 6F. β-tubulin was used as a loading and fractionation control.

Model illustrating the impact of nuclear and cytoplasmic CK1δ on circadian period length.

(A) Endogenous CK1δ continuously shuttles between the cytoplasm and the nucleus. Nuclear CK1δ levels remain low because the kinase is either rapidly degraded in the nucleus or efficiently exported back to the cytoplasm in a kinase activity-dependent manner. (B) Overexpression of NES-tagged CK1δ does not increase the nuclear kinase pool. The NES-tagged kinase enters the nucleus via PER:CRY complexes, with each PER molecule capable of importing only one NES-tagged kinase. Within the nucleus, PER-bound CK1δ equilibrates with the free nuclear pool. Once released from PER, rebinding of NES-tagged CK1δ is antagonized by either degradation or rapid export to the cytoplasm. Consequently, the nuclear kinase concentration remains low, and circadian period length is unaffected by NES-CK1δ overexpression. (C) Overexpression of NLS-tagged CK1δ increases the nuclear kinase pool by counteracting both nuclear export and degradation of CK1δ. This leads to a shortening of period length, as the elevated nuclear CK1δ facilitates PER2 binding and phosphorylation. Blue and gray schematic representations of CK1δ indicate high and low kinase concentrations, respectively.