Fig 1.
An RNAi screen identifies genes associated with the nucleo-cytoplasmic translocation machinery important for circadian dynamics.
U-2 OS Bmal1-luciferase reporter cells were transduced with RNAi constructs targeting 62 genes associated with nucleo-cytoplasmic translocation and bioluminescence rhythms were recorded for several days. Significant period alteration compared to non-silencing controls (ns, blue) was determined by one-way ANOVA of all constructs targeting one gene; a Dunnetts posttest assessed the effects of individual constructs. Multiple testing was corrected for by a Bonferroni-Holm correction. Horizontal dotted lines indicate two standard deviations of the ns control period. Targeting Fbxl3 was used as positive control (green). Genes for which at least two shRNA constructs led to a significant period alteration in the same direction are labeled (red dots); non-significant values are depicted in grey; dot size indicates the significance of period deviation: smallest: 0.05 > q > 0.005, medium: 0.005 > q > 0.001, largest: q < 0.001). When only one RNAi construct was available for gene knockdown, period deviation is depicted as empty circle (error bars = SD, n = 3).
Fig 2.
Knockdown and knockout of TNPO1 shortens the circadian period.
(A) Left: Representative time series for the effect of three different shTnpo1 constructs compared to a non-silencing (ns) control shRNA. Right: Mean period deviation and relative Tnpo1 mRNA and protein levels. Bottom: representative western blot of TNPO1 protein levels in ns control and Tnpo1 depleted cells; lanes are from the same blot with identical exposure time. Error bars = SD, nperiod deviation = 7 to 14 individual recordings, nmRNA and protein = 4–10 individual cell lysates; one-sample t-test to test whether period deviations are different from zero, *** p < 0.001. (B) Top: Schematic illustration of the CRISPR/Cas9-mediated genome editing of the Tnpo1 gene. Bottom left: Representative time series of U-2 OS reporter cells transduced with Cas9 expression vector and indicated guide RNA (gRNA). Middle: Mean periods of independent bioluminescence recordings (error bars = SD, n = 3, one-way ANOVA with Bonferroni posthoc test comparing sgRNAs to wild-type: *** p < 0.001, n.s. = non- significant. Right: Representative western blot of residual TNPO1 protein levels in genome-edited cells.
Fig 3.
(A) Scatterplot of the results of two different anti-CLOCK immunoprecipitations (IPs) from unsynchronized U-2 OS cell nuclei followed by mass-spectrometry. Depicted is the enrichment of identified proteins relative to the starting cell lysate for two CLOCK-antibodies against each other after normalization to an IgG control IP (n = 3 independent IPs, Welch’s t-test). (B) Representative MYC- or V5-CoIPs (empty bars) normalized to IgG control IPs (filled bars) of HEK293 lysates (mean ± SD, n = 3 independent IPs, Student’s t-test: n.s. = non-significant; * p < 0.05; ** p < 0.01; *** p < 0.001). To control for efficient IP, western blots of either MYC-TNPO1 or PER1/2-V5 were performed using the specific anti-MYC or anti-V5 IPs as well as the unspecific control IgG IPs. Experiments were repeated two to seven times. (C) Co-immunoprecipitation from U-2 OS cells stably expressing a PER1-LUCIFERASE fusion protein with either an antibody targeting endogenous TNPO1 (IPTNPO1) or an IgG control (IPC). Shown are luciferase intensities of αTNPO1 IPs (empty bar) normalized to counts from IgG control IPs (filled bar). Given are means ± SD, n = 3 independent IPs (** p < 0.01, Student’s t-test). Representative western blots to control for efficient IPs are shown below. (D) TNPO1 interaction region is located in the C-terminal part of PER1. Top: Schematic illustration of the primary structure of human PER1. Shown are nuclear export sequences (NES, green), the classical nuclear localization sequence (cNLS, blue), the three PY-motifs (i.e. putative non-classical nuclear localization motifs, red), the CRY interaction domain (brown) as well as 25 cysteine residues that are conserved within human and mouse PER1 (yellow and orange). Different grey/black shades indicate the PER1 fragments used. Bottom left: Co-immunoprecipitation from HEK293 cells expressing MYC-TNPO1 and indicated fragments of PER1-LUCIFERASE with anti-MYC (empty bars) or control IgG (filled bars). For comparison, data for full-length PER1 are re-plotted from panel B. Given are means ± SD from nine independent IPs performed at three experimental days. One-way ANOVA with Bonferroni corrected posthoc test revealed indicated significance; ** p < 0.01; n.s. = non-significant). Controls for efficient IP are shown below. Bottom right: Co-immunoprecipitation from HEK293 cells expressing MYC-TNPO1 and a mutant version of PER1-LUCIFERASE (PER1PY/Cys_mut), in which both C-terminal PY-motifs (PY883/834 and PY935/936) as well as seven C-terminal cysteine residues (orange) are exchanged to alanine residues, with anti-MYC (empty bars) or control IgG (filled bars). Given are means ± SD of five independent immunoprecipitations. Representative western blots to control for efficient IPs are shown below.
Fig 4.
Nuclear localization of PER1 is regulated by TNPO1.
(A) Depicted are representative fluorescence microscopic images (left) of U-2 OS cells ectopically expressing a version of PER1-Venus fusion protein, in which the classical NLS has been mutated (see also Fig 3D top). Cells were either treated with solvent (ethanol) or the nuclear export inhibitor Leptomycin B (LMB); scale bar: 50 μm. Right: Steady-state ratios of nuclear to cytoplasmic fluorescence intensity in single cells (box: median ± 25 percentile; whiskers: 10–90 percentile; n = 128 to 134 cells, statistics: Mann-Whitney-test; *** p < 0.001). (B) U-2 OS cells ectopically expressing PER1- or PER2-Venus fusion proteins were lentivirally transduced with shRNA constructs targeting Tnpo1 or ns control shRNA. Depicted are steady-state ratios of nuclear to cytoplasmic fluorescence intensity in single cells (box: median ± 25 percentile; whiskers: 10–90 percentile; n = 247 to 744 cells, statistics: Mann-Whitney-test with Bonferroni-Holm posttest, n.s. = non-significant, *** q < 0.001). (C) Kinetics of nuclear import was analyzed using fluorescence recovery after photobleaching (FRAP) of individual nuclei of U-2 OS cells described in (B). Upper part: representative images with arrow heads indicating bleached nuclei (scale bar = 20 μm). For quantification (lower part) nuclear fluorescence intensity (normalized to cytoplasmic intensity) was determined per time point (left panels) and the average recovery time of 25% of initial fluorescence intensity was calculated (right panels). Given are means ± SEM, n = 8 to 18 individual cells. Statistics on import kinetics: two-way-ANOVA, posttest: Sidak‘s multiple comparison test, time points significantly different from ns control (p < 0.05) are indicated as filled circles. Statistics on 25% recovery time: Given are means ± SD, n = 10 to 17 individual cells, One-way ANOVA with Bonferroni-Holm corrected posthoc Student’s t-test (one-sided); * p < 0.05; *** p < 0.001, n.s. = non-significant.
Fig 5.
Oxidative stress leads to increased PER1-TNPO1 binding.
Immunoprecipitation (IP) studies as described in Fig 3B were performed either under normal (filled bars) or oxidative stress (empty bars) conditions (200 μM H2O2). Shown are bioluminescence intensities upon MYC- or V5-CoIPs (red) normalized counts from IgG control IPs (black). Given are means ± SD, n = 3 independent IPs. One-way ANOVA revealed significant difference between columns (p < 0.001) except for the IP condition PER2-V5 TNPO1-LUC; Bonferroni post-test: n.s. = non-significant; * p < 0.05; ** p < 0.01; *** p < 0.001. Additional experiments gave similar results. For the IPs without H2O2 data from Fig 3B are re-plotted for comparison.
Fig 6.
Nuclear localization is of PER1 is modulated by oxidative stress in a TNPO1-dependent manner.
(A) Steady-state subcellular localization of ectopically expressed PER1- or PER2-Venus fusion proteins in TNPO1 depleted (red) or ns control (black) U-2 OS cells with or without H2O2 treatment. Depicted are nuclear to cytoplasmic fluorescence intensity ratios determined in four individually performed experiments (box: median ± 25 percentile; whiskers: 10–90 percentile, n = 185 to 801, One-way-ANOVA, posttest: Tukey’s multiple comparison test: n.s. = non-significant, *** p < 0.001). (B) Steady-state subcellular localization of ectopically expressed truncated versions (see also Fig 3D top) of PER1-Venus fusion proteins in U-2 OS cells with or without H2O2 treatment. Depicted are nuclear to cytoplasmic fluorescence intensity ratios (box: median ± 25 percentile; whiskers: 10–90 percentile, n = 21 to 59 cells, One-way-ANOVA, posttest: Tukey’s multiple comparison test: n.s. = non-significant, *** p < 0.001). (C) Kinetics of nuclear import upon H2O2 treatment was analyzed using fluorescence recovery after photobleaching (FRAP) of individual nuclei. Upper part: representative images with arrow heads indicating bleached nuclei (scale bar = 20 μm). Lower part, left: For quantification nuclear fluorescence intensity (normalized to cytoplasmic intensity) was determined per time point (left panels). Given are means ± SEM, n = 15 to 18 individual cells. For comparison, data for non-silencing control are re-plotted from Fig 4C. Two-way-ANOVA revealed indicated significance between treatment groups (*** p < 0.001, n.s. = non-significant). Lower part, right: The average recovery time of 25% of initial fluorescence intensity was calculated. Given are means of the times for 25% of initial PER1-Venus fluorescence recovery after photobleaching nuclei from Tnpo1depleted (red) or ns control (black) cells (error bars = SD, n = 10–17 cells for each condition, one-way ANOVA with Bonferroni-Holm corrected posthoc Student’s t-test (one-sided); * p < 0.05; n.s. = non-significant). Data of the two leftmost columns are re-plotted from Fig 4C.