VIP re-programmes SCN circadian rhythmicity at the cell and circuit level. a Representative PER2::LUC bioluminescence rhythms of vehicle-, 100 nM VIP- or 10 µM VIP-treated slices. VIP was bath-applied at CT10 of the 5^th cycle (marked by plus). Bioluminescence has been normalised to the first peak. b Scatterplot of phase shifts vs. PER2::LUC fold induction following VIP treatment. Line represents computed linear regression. n = 4–6 slices per concentration of VIP, ****P < 0.0001, Pearson’s correlation. c CCD camera images of PER2::LUC bioluminescence rhythms of representative slice before and after 10 µM VIP. Each set of images has been separately adjusted for maximum contrast. d Mean centre of mass (CoM) path vectors for 4 independent SCN slices; means were calculated over 5 days in the given time window. e CoM path vectors over time for representative SCN before and after VIP. f Group perimeter measurements of CoM path vectors for SCN slices before and with VIP (paired t-test, ***P < 0.001). g Regions of interest (ROI) determined by SARFIA mapped onto phase image of representative SCN. h, i Representative raster plot (h) and associated Rayleigh plot (i) of PER2::LUC bioluminescence rhythms of SCN slices treated with 10 µM VIP imaged on CCD camera (slice in g) VIP treatment marked by arrow. j Length of Rayleigh plot vector (n = 4 slices, mean ± SEM, example in g–i) before and after VIP treatment. k Periods for all ROIs across the SCN slices (n = 4 slices, mean per slice ± SEM; paired t-test, *P < 0.05). l Relative cumulative frequency (n = 4 slices, mean ± SEM, %) of ROIs with a given period before and after VIP treatment. m Amplitude of all ROIs across the 4 SCN slices (mean per slice ± SEM; paired t-test, **P < 0.01)

VIP re-programmes the electrically silenced SCN network. a Representative PER2::LUC bioluminescence rhythms of SCN slices treated with 10 µM VIP at CT10 (marked by plus), after vehicle or 1 µM tetrodotoxin (TTX) 24 h earlier (marked by asterisk). Bioluminescence was normalised to the first peak. b, c Group data for phase shift (b) and acute fold induction of PER2::LUC (c) responses relative to predicted values after treatments (mean ± SEM; Veh/Veh (n = 6), TTX/Veh (n = 5), Veh/VIP (n = 5), TTX/VIP (n = 7). d Peak-to-peak period following treatments (mean ± SEM). VIP was added between cycles 1 and 2. e, f Group data for period (e) and amplitude (f) responses after treatments (mean ± SEM; n as in b and c. g Representative PER2::LUC bioluminescence rhythms of SCN slices treated with 10 µM VIP at CT22 (marked by plus), after vehicle or 1 µM tetrodotoxin (TTX) 24 h earlier (marked by asterisk). h–j Group data for phase shift (h), period change (i) and amplitude (j) responses after VIP at CT22 preceded by TTX/vehicle (mean ± SEM; Veh/Veh (n = 3), TTX/Veh (n = 3), Veh/VIP (n = 5), TTX/VIP (n = 3). All tests two-way ANOVA with Sidak’s (d) or Tukey’s (b, c, e, f) multiple comparisons test, ns = not significant, ****P < 0.0001

VIP-induced gene networks and cellular signalling pathways in SCN. a Hierarchically clustered heatmap of top 300 most significantly altered genes in CT12 VIP vs. CT12 Veh comparison across time and treatment using globally normalised probe expression levels. Two regions are enlarged: the top displays genes acutely upregulated 2 h after VIP treatment before returning to baseline levels, the bottom shows genes which continue to be upregulated after 6 h. Within each gene row, transcript levels have been normalised to the minimum (blue) and maximum (red) values for that given gene. b, c Functional annotation of genes significantly altered by VIP treatment after 2 h (b; 738 genes) and after 6 h (c; 342 genes) using DAVID functional annotation and visualised using the enrichment map plugin for Cytoscape. d Enrichment of CREs in the promoters of significantly regulated genes, displayed as observed/expected frequency, with expected values based on ratios of the appearance of promoter elements across all genes detected on the microarray at CT12 VIP vs. CT12 Veh. CRE-TATA: CRE within 300 bp of TATA box; CRE-NoTATA: CRE in promoter further upstream; Others: No CRE identifiable in promoter (promoter defined as 3 kb upstream to 300 bp downstream of the transcription start site). Chi-squared test, ****P < 0.0001

VIP results in CRE activation and re-programmes the TTFL of the SCN. a Representative CRE-Luc bioluminescence rhythms of SCN slices treated with vehicle (left) or 1 µM VIP (right; treatments marked by plus). A detrended trace is included with the VIP-treated slice to aid visualisation of the oscillation. b, c Group data for acute induction of CRE-Luc bioluminescence (b) and period change (c) following treatment with vehicle or VIP (mean ± SEM, veh, n = 3; VIP, n = 4, unpaired t-test, *P < 0.05). d Representative CRE-Luc bioluminescence and RCaMP Ca²⁺ rhythms from SCN slice treated with VIP ( + ). e Detrended CRE-Luc bioluminescence and RCaMP Ca²⁺ rhythms before and after VIP (slice as in d). Note altered phase relationship following VIP. Rhythms were detrended by 24 h rolling average subtraction, and have been offset on the y-axis to aid visualisation. f, g Group data for phase of CRE-Luc rhythm relative to RCaMP rhythm (f negative values where CRE-Luc peaks later than RCaMP, mean ± SEM, n = 3) and period comparison between CRE-Luc and RCaMP, before and after VIP treatment (g two-way ANOVA with Sidak’s multiple comparisons test, **P < 0.01). h, i Group data for amplitude changes of CRE-Luc (h) and RCaMP (i) rhythms before and after VIP treatment (mean ± SEM, n = 3, paired t-test, *P < 0.05, **P < 0.01). Data were normalised to the highest value in the recording. j Representative PER2::LUC, Per1-Luc and Cry1-Luc bioluminescence rhythms of SCN slices treated with 1 µM VIP (at CT10, marked by plus). Bioluminescence was normalised to the first peak of the recording (not shown). k, l Group data for phase shift (k) and period change (l) responses (mean ± SEM) to VIP applied to either PER2::LUC (n = 4), Per1-Luc (n = 9) or Cry1-Luc (n = 7) slices, with vehicle controls (PER2::LUC: n = 6; Per1-Luc: n = 8; Cry1-Luc: n = 7; two-way ANOVA with Sidak’s multiple comparisons test, *P < 0.05, ***P < 0.001)

ERK1/2 activation is essential for the VIP response. a, b Representative PER2::LUC bioluminescence rhythms of SCN slices treated with vehicle (a) or 1 µM VIP ((b) marked by plus) at CT10, 30 min after 3 µM SP600125 (JNK1/2/3 inhibitor), 100 nM SCH772984 (ERK1/2 inhibitor) or vehicle. Bioluminescence has been normalised to the first peak. c–e Group data for phase shift (c), period change (d) and amplitude (e) responses after treatments (mean ± SEM; Veh/Veh (n = 5), Veh/VIP (n = 4), SP600/Veh (n = 3), SP600/VIP (n = 4), SCH772/Veh (n = 4), SCH772/VIP (n = 4); two-way ANOVA with Dunnett’s multiple comparisons test, *P < 0.05, **P < 0.01, ***P < 0.001)

Mice lacking Dusp4 exhibit attenuated circadian responses to light in vivo. a, b Group data for circadian period of wheel-running activity rhythms in 12:12 light:dark (LD) and continuous dim red light (DD) (a) and phase angle of entrainment (b) of Dusp^+/+ (n = 6), Dusp4^+/- (n = 6) and Dusp4^−/− (n = 5) mice (mean ± SEM). Phase angle refers to time from lights off until activity onset. Negative values indicate activity onset after lights off. c Representative double-plotted actograms of wheel-running behaviour of Dusp^+/+, Dusp4^+/- and Dusp4^−/− mice exposed to an 8 h phase advance and an 8 h phase delay in LD cycle. Grey shading represents dim red light. d, e Group data for activity onsets during the phase advance (d) and activity offsets during the phase delay (e) of Dusp^+/+, Dusp4^+/- and Dusp4^−/− mice (mean ± SEM). Two-way ANOVA with Dunnett’s multiple comparisons test. Inset: 50% phase shift values (PS50; mean ± SEM) of phase advances and phase delays, one-way ANOVA with Dunnett’s multiple comparisons test. f Representative double-plotted actograms of wheel-running behaviour of Dusp^+/+and Dusp4^−/− mice exposed to a 30-minute phase delaying light pulse (yellow asterisk) at ZT14 after 12:12 LD, then immediately placed into DD. Grey shading represents dim red light. g, h Group data for phase delays (g) and phase advances (h) of wheel-running activity of Dusp^+/+, Dusp4^+/- and Dusp4^−/− mice following light pulses at CT14 and CT22, respectively. One-way ANOVA with Tukey’s multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Dusp4 knockout shortens circadian period and potentiates the VIP response of SCN slices. a, b Representative PER2::LUC bioluminescence rhythms of slices transduced with Mecp2.SpCas9 alone or in addition to U6.Dusp4g1.hSyn.GFP-KASH before treatment (a) or with 1 µM VIP application (b marked by plus). Bioluminescence has been normalised to the first peak. c Group data for steady-state period of slices transduced as in (a) (mean ± SEM, n = 9 per group, unpaired t-test). d, e Group data for phase shifts (d) and period changes (e) following vehicle or VIP of slices transduced as in a (mean ± SEM, n = 4 for all vehicle-treated groups, n = 6 for Cas9 with VIP, n = 9 for Cas9 + Dusp4 g1 with VIP). f Group data for period change of Dusp4^flx/flx SCN slices transduced with hSyn.GFP::Cre AAV and treated with vehicle or VIP, along with non-transduced controls (mean ± SEM, n = 3 for all vehicle-treated groups, n = 4 for all VIP-treated groups). Two-way ANOVA with Sidak’s multiple comparisons. *P < 0.05, **P < 0.01

Constitutive DUSP4 expression shortens circadian period and reduces the VIP response in SCN slices. a Representative confocal micrographs of an SCN slice transduced with hSyn.GFP::Cre AAV and Ef1a.DiO.mCherry-P2A-Dusp4 (SynCre-DUSP4^Ox). b Group data for steady-state (pre-treatment) period of slices transduced as in a, along with individual AAV controls (mean ± SEM, one-way ANOVA, n = 14 for DUSP4^Ox group, n = 8 SynCre and SynCre-DUSP4^Ox). c Representative PER2::LUC bioluminescence rhythms of SynCre and SynCre-DUSP4^Ox slices treated with 1 µM VIP (marked by plus). d, e Group data for post-treatment relative amplitude (d), and period change (e) responses (mean ± SEM) of slices treated with vehicle or VIP at CT10. n as follows: DUSP4^Ox: veh, 8; VIP, 10; SynCre: veh, 3; VIP, 5; SynCre-DUSP4^Ox: veh, 4; VIP, 6. f Representative confocal micrographs of a Vpac2-Cre SCN slice transduced with Ef1a.DiO.mCherry-P2A-Dusp4 (VpacCre-DUSP4^Ox), immunostained for VIP to demonstrate separate populations of VIP and VPAC2 neurons (close up on right, white boxed region). g Representative PER2::LUC bioluminescence rhythms of VpacCre and VpacCre-DUSP4^Ox slices treated with 1 µM VIP (marked by plus). h, i Group data for post-treatment relative amplitude (h), and period change (i) responses (mean ± SEM) of slices treated with vehicle or VIP at CT10. DUSP4^Ox data replicated from d, e. n as follows: VpacCre: veh, 4; VIP, 6; VpacCre-DUSP4^Ox: veh, 7; VIP, 10. All tests in d, e, h, i were two-way ANOVAs with Sidak’s multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001