The VIP-VPAC2 neuropeptidergic axis is a cellular pacemaking hub of the suprachiasmatic nucleus circadian circuit

We characterised mouse lines expressing Cre-recombinase in either VIP or VPAC2 cells, mapping cellular distributions by inter-crossing with R26R-EYFP genomic reporter mice. EYFP-labelled cells exhibited contrasting SCN distributions: VIP^EYFP cells clustered ventrally within the core whereas VPAC2^EYFP cells localised around the dorsal, lateral and medial shell (Supplementary Fig. 1a, b). Immunostaining and in situ hybridisation of SCN sections showed clear segregation of VIP and VPAC2 (or AVP) cells (Supplementary Fig. 1c–f), with minimal evidence for mutual VIP–VIP cell signalling via VPAC2. These lines therefore provided selective genetic access to distinct VIP or VPAC2 cellular compartments.

When compared between circadian day and night, the contrasting regularity of firing in VIP and VPAC2 cells (p < 0.0001) neither changed with time, nor was there an interaction (p = 0.14) between time and cell type (Supplementary Fig. 5d, e). Regularity of firing is, therefore, not circadian, but an intrinsic property distinguishing VIP and VPAC2 neurons. In contrast, the temporal pattern of RMP differed significantly between VIP and VPAC2 cells (Fig. 2c–e) (ANOVA, time: p < 0.05, genotype: p < 0.001, interaction: p < 0.0001). VPAC2^TdTom cells were consistently depolarised compared to VIP^TdTom cells and showed significant hyperpolarisation at night versus day (Fig. 2d, e), whereas VIP^TdTom cells lacked measurable circadian RMP changes (Fig. 2d). In both populations, SFR was significantly higher in circadian day (CT0-8) versus night (CT16-20) (p < 0.0001) (Fig. 2f, g), but with a significant interaction between time and cell type (p < 0.05). Whereas SFR of VIP^TdTom cells peaked at CT0-4 (Fig. 2f), VPAC2^TdTom cells peaked at CT4-8 indicating that electrical activity in VIP cells is phase advanced relative to VPAC2 cells. To obtain complementary analyses of circuit-level electrophysiological activity, we transduced SCN with the voltage sensor AAV synapsin-ArcLight, expressed either pan-neuronally^19,22 (Supplementary Fig. 5f–h) or conditionally in VIP^Cre or VPAC2^Cre cells (Fig. 2h), and phased-registered the fluorescence to simultaneous whole-field Per2::Luciferase recordings. Both VIP^ArcLight and VPAC2^ArcLight neurons demonstrated stable circadian rhythms in electrical activity (Fig. 2i, j). Importantly, the peak circadian phase was significantly phase advanced in VIP^ArcLight cells by ~1 h relative to pan-neuronal and VPAC2^ArcLight cells (Fig. 2k) (Neuronal: CT 7.6 ± 0.2 h; VIP: CT6.8 ± 0.1 h; VPAC2: CT7.4 ± 0.2 h). Thus, VIP and VPAC2 cells are electrophysiologically distinct and exhibit differentially phased (VIP advanced) circadian electrophysiological rhythms.

If resetting by VIP cells cannot be explained solely by increased firing of their VPAC2 targets, molecular and/or metabolic actions of VIP additional to electrical activation likely effect resetting of the TTFLs of VPAC2 cells. The most direct way would be VIP-dependent acute induction of Per1 and Per2, mediated by cAMP/calcium responsive elements (CREs)¹⁵. To explore this, we monitored acute changes in Per2::Luciferase bioluminescence in arrhythmic Cry1,2-null (Cry-null) SCN, which provided a stable baseline to observe TTFL responses induced by activation of VIP cells. In Cry-null SCN, acute pan-neuronal optogenetic activation elevated bioluminescence, peaking after ~3–4 h and returning to baseline after ~10-11 h (Fig. 5g). Stimulation solely of VIP^ChR2::EYFP cells in Cry-null SCN resulted in a comparable bioluminescent induction (Fig. 5h). Furthermore, using ChR2::EYFP to delineate the localisation of VIP cells and cross-correlating maps of bioluminescent change with EYFP intensity revealed that optogenetic activation induced Per2 expression both inside and outside the territory of VIP cells, with the same magnitude, kinetics and distribution of responses (both p > 0.2) (Fig. 5i–l, Supplementary Fig. 8b). Thus, output from VIP cells activates and recruits pan-SCN cellular TTFLs, with serial activation along the VIP/VPAC2 cellular axis re-setting ensemble phase.

Initiation of ensemble rhythmicity demonstrated the clear pacemaking property of the VIP/VPAC2 cellular axis. We then used CCD recordings to test whether it also directed further network features, including synchrony and spatio-temporal dynamics. CoL measurements revealed the absence of a spatio-temporal wave in Cry1-transduced, single-Cre SCN (Fig. 7e). In contrast, joint activation of cell-autonomous TTFLs in VIP^Cre and VPAC2^Cre cells initiated strong Per2::Luciferase bioluminescent waves, approaching wild-type trajectories and accompanied by tight phase coupling of oscillators across the network (Fig. 7e, f, Supplementary Fig. 10). Furthermore, phase maps of Per2::Luciferase bioluminescence in Cry1-initiated slices recapitulated many of the spatiotemporal dynamics of the wild-type slice, including region-specific phasing (Fig. 1), with the induced activity starting dorsomedially and progressing to the core before moving dorsally, medially and laterally into the phase-delayed shell (Fig. 7g, Supplementary Fig. 10).

All experiments were carried out in accordance with the UK Animals (Scientific Procedures) Act of 1986, with local ethical approval provided by the Medical Research Council Laboratory of Molecular Biology Animal Welfare Ethical Review Board (LMB AWERB) and overseen institutionally by designated animal welfare officers (NACWOs). VIP^Cre mice (Viptm1(cre)Zjh/J) were purchased from the Jackson Laboratory (Bar Harbor, Maine, USA). VPAC2^Cre mice (Tg(Vipr2-cre)KE2Gsat/Mmucd) were purchased from GENSAT (Gene Expression in the Nervous System Atlas) project (Rockefeller University, New York City, USA). R26 floxed STOP EYFP (B6.129×1-Gt(ROSA)26Sortm1(EYFP)Cos/J) mice were kindly provided by A. McKenzie, MRC Laboratory of Molecular Biology, Cambridge, UK. Cry1-null and Cry2-null animals were derived from founders kindly provided by G. van der Horst, Erasmus University Medical Centre, Rotterdam, NL. Per2::Luciferase mice were kindly provided by J. S. Takahashi, University of Texas Southwestern Medical Centre, Dallas, USA. All lines were maintained on a C57BL/6J background with required genotypes bred in-house by intercrossing the lines.

AAVs used were a mix of commercially available AAVs and custom made AAVs. The following AAVs were sourced directly from Penn Vector Core and via Addgene: Cre-conditional tdTomato (AAV1-CAG-FLEX-tdTomato-WPRE (Addgene viral prep # 51503-AAV1)), pan-neuronal GCaMP6f (AAV1.Syn.GCaMP6f.WPRE.SV40 (Addgene viral prep #100837-AAV1)), Cre-conditional neuronal GCaMP6f (AAV1.Syn.Flex.GCaMP6f.WPRE.SV40 (Addgene viral prep # 100833-AAV1)), pan-neuronal ArcLight (AAV1.hSynap.ArcLightD.WPRE.SV40 (Addgene viral prep # 100037-AAV1)), Cre-conditional neuronal ArcLight (AAV1.hSynap.Flex.ArcLightDco.WPRE.SV40 (Addgene viral prep # 100038-AAV1)), pan-neuronal ChR2::EYFP (AAV1.Syn.hChR2(H134R)-EYFP (Addgene viral prep # 26973-AAV1)), Cre-conditional ChR2::EYFP (AAV1.EF1a.double floxed.hChR2(H134R)-EYFP.WPRE.HGHpA (Addgene viral prep # 20298-AAV1)), Cre-conditional neuronal mCherry (AAV8-hSyn-DIO-mCherry (Addgene viral prep # 50459-AAV8)) and Cre-conditional EYFP (AAV1-EF1a-DIO-EYFP (Addgene viral prep # 27056-AAV1)). pAAV-CAG-FLEX-tdTomato-WPRE was a gift from Hongkui Zeng. pAAV.Syn.GCaMP6f.WPRE.SV40 and pAAV.Syn.Flex.GCaMP6f.WPRE.SV40 were gifts from Douglas Kim & Genie Project⁴⁵. pAAV.hSynap.ArcLightD.WPRE.SV40 and pAAV.hSynap.ArcLightDco.WPRE.SV40 were gifts from Vincent Pieribone. pAAV.hSyn.hChR2(H134R).EYFP, pAAV.EF1a.double floxed.hChR2(H134R)-EYFP.WPRE.HGHpA and pAAV.EF1a.DIO-EYFP were gifts from Karl Deisseroth. pAAV-hSyn-DIO-mCherry was a gift from Bryan Roth. The following AAVs backbones were produced in-house and packaged by Penn Vector Core as AAV1 serotype: Cre-conditional pCry1-luciferease (pCry1.DIO.luc) and Cre-conditional Cry1::EGFP (pCry1.DIO.Cry1::EGFP).

For packaging into AAV vectors, the DtR-containing transgenic cassette was cloned into pAAV-EF1a-double floxed-hChR2(H134R)-mCherry-WPRE-HGHpA (Addgene #20297). Gibson cloning was used to insert an mCherry fluorescent protein and the simian DtR separated by a P2A peptide (to generate separate proteins in equimolar amounts) between the four loxP sites contained within the plasmid. hChR2-mCherry was excised from pAAV-EF1a-double floxed-hChR2(H134R)-mCherry-WPRE-HGHpA using BsrGI and NheI, linearising it in the process. mCherry-P2A was amplified from HIV.CMV.mCherry-P2A-Cre using the forward primer 5′-TAACTTCGTA TAGGATACTTTATACGAAGTTATGCTAGCCACCatggtgagcaagggcgagg-3′ and the reverse primer 5′-GCTTCATagggccgggattctcctccacgtc-3′ (capitalised letters represent regions of the primers complementary to the vector backbone and the DtR sequence respectively). The DtR sequence was amplified from pAAV.pCAG.DIO.DtR.WPRE, kindly donated by Dr. Marco Tripodi (MRC Laboratory of Molecular Biology, Cambridge, UK) using the forward primer 5′-GTGGAGGAGAATCCCGGCCCTatga agctgctgccgtcgg-3′ and the reverse primer 5-CATTATACGAAGTTATGGCG CGCCTTACTTGTACAtcagtgggaattagtcatgcccaacttc-3′ (capitalised letters represent regions of the primers complementary to the P2A sequence and the vector backbone respectively).

The resulting PCR fragments were purified and incubated with the linearised vector backbone and Gibson mix (NEB) for 30 min at 50 °C. The reaction was chilled on ice before adding 5 µl of the mix to chemically competent cells, kindly donated by Dr. Ernesto Ciabatti (MRC Laboratory of Molecular Biology, Cambridge, UK). This was left on ice for 2 min and then plated onto LB plates containing ampicillin. Colonies were selected, mini-prepped and confirmed by sequencing. Following validation in cells, the Ef1a.DiO.mCherry-P2A-DtR plasmid was packaged into AAV1 serotype vectors by Penn Vector Core.

Mice (P8-10) were sacrificed according to local and Home Office rules, and the suprachiasmatic nucleus (SCN) was removed and cultured as an explant. Briefly, coronal hypothalamic slices were cut at 300 µm and the SCN was dissected free using a razor blade in ice-cold GBSS supplemented with (in mM): 5 mg/ml glucose, 50 µM D-AP5, 100 nM MK-801 and 3 mM MgCl₂. Slices were maintained in the interface method for 2–3 h in media containing: 50% Eagle’s Basal Medium (Gibco), 25% EBSS (Gibco), and 25% Horse Serum supplemented with 5 mg/ml Glucose, 2 mM GlutaMAX (Gibco), 1:100 dilution of Penicillin/Streptomycin (Gibco), 50 µM D-AP5, 100 nM MK-801, and 3 mM MgCl₂. Following 2–3 h in culture, slices were incubated in the same media without the addition of D-AP5, MK-801 and MgCl₂ for a week. After a week, culture medium was changed, and 1-µl AAVs (between 1 × 10¹² and 1 × 10¹³ GC/ml in PBS) were added drop-wise to the surface of the slice 24 h later. Transduced slices were left for one week before AAVs were washed out by fresh culture medium and, in most cases, successful transduction was assessed by imaging.

For bioluminescent photomultiplier tube (PMT) recordings, slices were transferred to DMEM-based (Sigma-Aldrich) recording medium supplemented with: 4.17 mM NaHCO₃, 5 mg/ml glucose, 1:100 dilution of Penicillin/Streptomycin (Gibco), 10 mM HEPES, 5% FCS, 2 mM GlutaMAX, and 100 µM luciferin in 35-mm dishes. The dishes were then sealed with glass coverslips and vacuum grease before being transferred to a custom built PMT (H9319-11 photon counting head, Hamamatsu) array within a light-tight incubator at 37 °C. Bioluminescent emissions were collected in real time and binned into 6-min intervals before analysis. For bioluminescent imaging via CCD camera, slices were sealed into 35-mm dishes and transferred to the heated stage of an inverted microscope and CCD camera (Hamamatsu) setup. Bioluminescent time-lapse images were taken over 1-h intervals.

For combined bioluminescent and fluorescent imaging, slices were sealed into 35-mm dishes with glass bottoms (Mattek) and transferred to the heated stage of an LV200 microscope system (Olympus) running Olympus proprietary acquisition software (CellM, xcellence rt or cellSens) and equipped with an EM-CCD camera (Hamamatsu). Bioluminescence (PER2::Luciferase and pCry1-luc) and fluorescence (EYFP, GCaMP6f and ArcLight) images were taken once every 30 min, and recorded for at least 5 cycles. Exposure times ranged between 9.5 and 29.5 min for bioluminescence and 25 and 100 ms for fluorescent reporters (EYFP: 25–100 ms; GCaMP/ArcLight: 100 ms) dependent on the configuration of the experiment.

Images were analysed in FIJI⁴⁶, IgorPro (Wavemetrics) using the SARFIA plugin, Excel (Microsoft) and Graphpad Prism (Graphpad). To generate phase maps and maps of bioluminescence/fluorescence intensity, slices were thresholded to remove extra-SCN signals before a custom FIJI plugin was used to extract signals from a continuous grid of ROIs placed over the SCN. In the case of phase maps, signals were detrended and smoothed by a 2.5 h moving average before peaks were identified and the peak-to-peak period of the PER2::LUC signal was used to calculate and normalise the phase of each cycle to allow the cycle to cycle and mean phase maps to be plotted for PER2 and GCaMP signals. For changes in normalised bioluminescence intensity, the change within each ROI was normalised to the baseline bioluminescence intensity within the same ROI. In order to define regions delineated by fluorescent markers, the average intensity projection was overlaid with the same grid and the intensity of the signal was normalised to the mean intensity across all ROIs. Areas with a normalised intensity of <1 were defined as outside the region, while areas with a normalised intensity of >1 were defined as within the region. Continuous phase maps were assembled and coloured in Graphpad Prism using the heat-map plot and were coloured using the “rainbow” colour setting. In the case of cluster phase-maps, phases were classified by manually assigning colours to any ROIs falling within the following phase ranges: <CT10, CT10–11, CT11–12, CT12–13, CT13-14 and >CT14.

CoL was measured using IgorPro scripts utilising the SARFIA toolset. CoL trajectory area was calculated from the perimeter of the CoL excursion and expressed as a percentage of the area of the temporally integrated bioluminescence signal. Circadian bioluminescence and fluorescence signals were analysed using a manual peak identification and peak-to-peak period approach in the case of phase-mapping electrophysiology experiments (Fig. 2), fluorescence/bioluminescence experiments (Fig. 3) and optogenetics experiments (Fig. 6). Differences in phasing comparing pan-neuronal or targeted ArcLight and GCaMP6f were assessed statistically by ordinary one-way ANOVA with Holm-Sidak’s correction for multiple comparisons. Potential differences in aggregate period between Per2::Luciferase and pan-neuronal or targeted (VIP^Cre and VPAC2^Cre) ArcLight/GCaMP6f were assessed statistically by paired t-tests. Differences in phasing, Rayleigh vector length, CoL trajectory, baseline and amplitude for pCry1-luciferase rhythms were assessed by unpaired t-tests. Differences in period for pCry1-luciferase rhythms between pan-neuronal (synapsin-Cre) and targeted (VIP^Cre and VPAC2^Cre) conditions were assessed by ordinary one-way ANOVA with Holm-Sidak’s correction for multiple comparisons. Bioluminescence recordings from slice cultures in long-term PMT recordings were analysed to calculate circadian period length, amplitude, relative amplitude error (RAE) and goodness of fit (GOF) using the Fast Fourier Transform—Non-Linear Least Squares (FFT-NLLS) function in the BioDARE software⁴⁷ (www.biodare2.ed.ac.uk↗).

Mouse brains were post-fixed in 4% PFA (in phosphate buffer) and cryoprotected in 20% sucrose (in PBS) prior to sectioning. Brains were sectioned on a freezing microtome (Bright Instruments, UK) at 40-µm thickness. Sections containing the SCN were blocked with 5% NGS prior to immunostaining. Primary antisera used were: Rabbit anti-AVP (Peninsula, T-4563) (used at 1:1000 dilution), rabbit anti-VIP (Immunostar, 20077) (used at 1:750 dilution) and guinea pig anti-VIP (Peninsula T-5030) (used at 1:1000 dilution). Fixed brain sections and slices were imaged using the Zeiss 780 inverted system (Zeiss, Germany) with either ×10 or ×20 objective for low power images or ×63 objective for high power images, used for cell counting. Cell counting for co-localisation was done manually using the FIJI Nucleus Counter and Cell Counter plugins. Transduction efficiency of AAVs was calculated using AAV-dependent fluorescence cell counting in conjunction co-localisation. Efficiency of AAV1.pCry1.DIO.Cry1::EGFP transduction was expressed as a percentage of Cre+ cell counts, where Cre+ cells were revealed using AAV8.pSyn.DIO.mCherry transduction. AAV8.pSyn.DIO.mCherry transduction efficiency was then assessed by calculating the percentage of Cry1::EGFP-positive cells where mCherry is absent, termed “over-spill” expression.

Due to spatial patterning of VIP-ir in the projections of VIP neurons, it can be technically difficult to perform cell counts. To confirm specificity of Cre-recombinase expression in the VIP-cells and perform cell counts, fluorescent in situ hybridisation (FISH) was performed against Cre-recombinase and VIP. Labelled riboprobes were generated via a one-step reverse-transcriptase PCR (RT-PCR) protocol using gene-specific primers to generate cDNA, followed by an in vitro transcription (IVT) to create the riboprobe. Primer sequences were sourced from the Allen Brain Atlas (VIP forward: CCTGGCATTCCTGATACTCTTC, VIP reverse: ATTCTCTGATTTCAGCTCTGCC; Cre forward: CCAATTTACTGACCGTACACCA, Cre reverse: TATTTACATTGGTCCAGCCACC) and synthesised (Sigma) containing either the bacterial promoter T7 sequence (TAATACGACTCACTATAGGGAGA; forward primer) or T3 sequence (AATTAACCCTCACTAAAGGGAGA; reverse primer) 5′ to the rest of the sequence to generate sense and antisense probes, respectively, via IVT. A one-step reverse-transcriptase (RT)-PCR was carried out (SuperScript III One-Step RT-PCR System with Platinum Taq kit, Invitrogen) to reverse-transcribe RNA extracted from SCN slices (RNeasy Micro, Qiagen) and amplify the resultant cDNA. 50–100 ng RNA was used as a template. Once the resultant cDNA was purified (QIAquick PCR Purification kit, Qiagen), 500 ng was in vitro transcribed using either T3 or T7 polymerase (Roche), in a reaction mix containing 1X transcription buffer (Roche), 0.01 M DTT, digoxigenin (DIG) or fluorescein (FLU) labelling mix (Roche), RNasin, and polymerase. This was left at 37 °C for 2 h in a thermocycler and was subsequently cleaned up using Micro Bio-Spin 6 columns (Bio-Rad) before diluting 1:5 in nuclease-free water and storing at −20 °C.

Mice were culled by cervical dislocation and exsanguination, and dissected brains were immediately frozen on aluminium foil on top of dry ice, SCN facing upwards. Twelve µm sections were taken using a cryostat and allowed to air dry. These sections were post-fixed in 4% PFA made up in DEPC-PBS for 20 min at 4 °C. Following three washes in DEPC-PBS, sections were incubated in triethanolamine (TEA; Sigma) solution (1.4 M TEA in DEPC-H₂O) for 3 min at room temperature before being acetylated in 0.25% acetic anhydride in TEA solution for 10 min at room temperature. Sections were subsequently washed in DEPC-PBS before dehydration by ascending concentrations of ethanol (50, 70, 95 and 100%) for 3 min each. Sections were then left to air-dry before addition of hybridisation buffer (62.5% deionised formamide, 375 mM NaCl, 1.25x Denhardt’s solution, 12.5 mM Tris (pH 8), 1.25 mM EDTA, 12.5% dextran sulphate, supplemented with 250 µg/ml Torula yeast RNA and 12.5 mM dithiothreitol.

RNA probes were diluted 1:100 in hybridisation buffer, heated at 60 °C for 10 min and then immediately chilled on ice for 5 min. Eighty microlitres of probe solution was added to each slide and a coverslip was placed on top of the solution to ensure uniform spreading and reduce evaporation. Slides were placed in a hybridisation chamber (saturated with 50% formamide and 50% PBS) and incubated overnight in a ventilated oven at 58 °C. Following hybridisation, coverslips were removed by washing in 4× saline-sodium citrate buffer (SSC) before treating with 20 µg/ml RNase A for 30 min at 37 °C. The sections were subsequently washed in decreasing SSC concentrations (2×, 1×, 0.5×) before a final wash in 0.1x SSC at 60 °C. Sections were then washed and equilibrated in Tris-NaCl buffer (TNB; 0.1 M Tris (pH 8), 150 mM NaCl) before blocking in 5% normal sheep serum (in TNB) for 30 min. The primary antibody (anti-DIG-POD) was diluted 1:100 in this blocking buffer and 80 µl of this antibody solution to each slide, cover-slipping and incubating at 4 °C overnight. Sections were washed three times in TNB to remove the antibody solution and were then treated with TSA working solution (TSA Plus Cy3 diluted in TSA Diluent 1:50; Perkin Elmer): 80 µl was added to each slide, which were then cover-slipped and incubated in the dark at room temperature for 1 h to produce a fluorescent precipitate. Three washes in TNB followed before incubating the slides in 3% H₂O₂ (in TNB) for 30 min. Sections were again blocked and treated with the remaining antibody (anti-FLU-POD) at a 1:100 dilution and incubated overnight as before. The washes and TSA treatment were also repeated, using TSA plus FLU. Three washes in TNB and a final rinse in water were carried out before immediately cover-slipping with Vectashield Hardset mounting medium with DAPI.

For electrophysiological recordings, SCN slices were maintained in PMTs and their bioluminescent output was monitored to determine phase and extrapolate time-stamped recordings in circadian time. Slices were transferred to the stage of an upright (BX51WI, Olympus) microscope by excision of the supporting membrane and maintained in ACSF (in mM: 125 NaCl, 25 NaHCO₃, 1.25 NaH₂PO₄, 3 KCl, 25 Glucose, 2 CaCl₂, 1 MgCl₂) bubbled with 95%CO₂/5% O₂, constantly perfused at ~1.5 ml/min at 32 °C via an inline heater. For cell-attached recordings, borosilicate glass pipettes (3-6MOhms) were filled with 150 mM NaCl and recordings were made in voltage clamp mode using a Cornerstone BVC-700A amplifier (Dagan). Signals were digitised at 10 kHz and filtered at 1 kHz. For electrophysiological optogenetic characterisation, cells recorded in cell-attached voltage clamp mode were identified by their ChR2::EYFP tag. Each cell recorded was stimulated with a train of 60 equally spaced pulses (15 ms) delivered at 1, 5, 10, 15 and 20 Hz using a computer-controlled TTL-triggered LED array (pE-4000, 470 nm LED, CoolLED) via the widefield standard EGFP fluorescence cube. Firing index was calculated by dividing the number of evoked spikes by 60 (the number of light pulses delivered), with values of <1 representing failure of the stimulation pattern to induce firing at the desired rate.

For whole-cell recordings, borosilicate glass pipettes (4–8 MOhms) were filled with a K-gluconate based internal solution (in mM: 135 K-gluconate, 7 NaCl, 10 HEPES, 2 Na₂-ATP, 0.3 Na₂-GTP, 0.01 Biocytin-Alexa488, 2 MgCl₂) and recordings were made using a Cornerstone BVC-700A amplifier (Dagan). Targeted recordings were made via visualisation of the tdTomato signal excited using widefield illumination via LED illumination (pE-4000, 550 nm LED, CoolLED) via a standard Texas Red fluorescence filter cube. Signals were digitised at 20 kHz and filtered at 3 kHz. Series resistance for whole-cell recordings was monitored at break in and at the end of the recording, and if it changed by >25%, recordings were omitted. Input resistance was measured by applying a series of 500 ms long hyperpolarising current steps (−30pA to 0pA, 10pA steps) to neurons and measuring the steady-state voltage deflection. Input resistance was defined as the gradient of the linear relationship between the voltage deflection and the injected current. All recordings were completed within 3–4 min of break-in to minimise effects of intracellular perfusion, but cells were allowed to perfuse for up to 10 min to confirm successful targeting by registration of the green Alexa-488 signal with the red tdTomato signal and images were acquired using micromanager (ImageJ). For all whole-cell recordings, correction of the junction potential was applied a posteriori.

Analysis was performed in AxoGraphX (AxoGraph), Igor Pro (Wavemetrics) using the Neuromatic plugin⁴⁸, Excel (Microsoft), Graphpad Prism (Graphpad) and R (version 3.6.1, courtesy of R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org↗) using RStudio (version 1.2.1335, courtesy RStudio Team, Boston, MA. http://www.rstudio.com↗). The regularity of firing was assessed quantitatively using the LvR metric (reproduced from ref. ²⁵):1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{LvR}} = \frac{3}{{n - 1}}\mathop {\sum }\limits_{i = 1}^{n - 1} \left( {1 - \frac{{4I_iI_{i + 1}}}{{(I_i + I_{i + 1})^2}}} \right)\left( {1 + \frac{{4R}}{{I_i + I_{i + 1}}}} \right),$$\end{document}LvR=3n−1∑i=1n−11−4IiIi+1(Ii+Ii+1)21+4RIi+Ii+1,where I_i and I_i+1 represent consecutive interspike intervals (ISIs), n is the number of ISIs and R represents the refractoriness constant (defined as 5 ms²⁵) to correct for the refractory period within an ISI directly following a spike and define regularity as a variable independent of firing rate.

For pooled and time-aligned electrophysiological data, a linear mixed model was fitted to the data to take into account differences in the numbers for neurons recorded between slices. For pooled electrophysiological data, to examine the underlying differences between the two genotypes (VIP^Cre or VPAC2^Cre), the linear mixed model was fit using genotype as a fixed factor with slice as a random factor. In the time-aligned datasets, the linear mixed model was fit using genotype, time and an interaction between the two as fixed factors with slice as a random factor. The models were fit using the nlme package in R (nlme version 3.1.143) (courtesy R Core Team nlme: https://CRAN.R-project.org/package=nlme↗). In the case of the pooled electrophysiological data, the significance of the genotype effect was assessed by applying an unpaired, two-tailed Welch’s t-test. For the time-aligned data, the significance of the fixed factors in the model were determined through application of a two-way ANOVA. The two-way ANOVA was followed by pair-wise comparisons made on the estimated marginal means derived from the model and Tukey’s correction was applied to these multiple comparisons made using the emmeans package in R (emmeans version 1.4.3.1) (Estimated Marginal Means, aka Least-Squares Means. R package version 1.4.3.01. https://CRAN.R-project.org/package=emmeans↗).

SCN slices were recorded in specially adapted PMTs with fibre optic cables running into the base of the chamber to allow for optogenetic stimulation (1 h, 10 Hz) in situ via computer-controlled LEDs (in-house hardware and software, Electronics Workshop, MRC Laboratory of Molecular Biology). Stimulation strength of both LED wavelengths (470 nm: M470F, Thor Labs/625 nm: M625F, Thor Labs) was measured to be at least 1 mW/mm². Slices were recorded for 4 cycles before stimulation and this was used to extrapolate the phase of stimulation and to predict the peaks post-stimulation. Slices were recorded for 4 cycles post-stimulation and phase shifts were calculated as the difference between the actual peak and the predicted peak normalised in circadian time to the baseline period and expressed as the mean of the three cycles following stimulation. For the Cry-null optogenetic experiments, stimulation was given after recording a 24 h stable PER2::Luciferase baseline. In the case of direct imaging, slices were checked for a stable baseline in PMTs before being transferred to an LV200 bioluminescence/fluorescence imaging system (Olympus). Bioluminescence and EYFP signals were captured every 30 min before optogenetic stimulation (1 h, 10 Hz) was delivered to the slice via the GFP cube. Slices were then recorded for a further 24 h. Significance was assessed in the PRCs via two-way ANOVA with Sidak’s correction for multiple comparisons. Bioluminescent changes in Cry-null SCN receiving pan-neuronal or VIP-targeted optogenetic stimulation were analysed via paired t-tests or repeated-measures two-way ANOVAs with Sidak’s correction for multiple comparisons.

Following baseline recording of SCN slices for >5 days, Dtx or vehicle (recording media) was added to the bulk media. Dtx (Sigma) was bath-applied to SCN slices at a final concentration of 20 ng/ml. Following 5–7 days of treatment, Dtx or vehicle was washed off via media change and slices were recorded further for >5 days. Recordings were analysed in BioDare, values were calculated based on rhythms 4 days pre-treatment and 4 days post-treatment excepting the first 36 h immediately after treatment application. Data were analysed via either a two-way ANOVA or three-way ANOVA, followed by Sidak’s correction for multiple comparisons.

Cry1-null SCN slices were recorded in PMTs for 7 days to confirm arrhythmicity. They were then transduced with 1 μL of AAV.pCry1.DIO.Cry1EGFP and returned to the PMTs for a further 7 days, followed by a media change and return to PMTs for a final 7 days. Pre- and post-analysis of Per2::Luciferase rhythms used the first and last 7 days respectively. In order to follow transient period changes from cycle to cycle as Cry1::EGFP expression increased following transduction, peak-to-peak period was calculated using the area under a curve feature in Graphpad Prism to automatically identify the time of PER2::LUC peaks. Time intervals between these automatically identified peaks were expressed as the period for that cycle. Period comparisons were made via a repeated-measures two-way ANOVAs with Sidak’s correction for multiple comparisons. Cry1,2-null SCN slices were recorded from in PMTs for 7 days to confirm arrhythmicity. They were then transduced with 1 μL of AAV.pCry1.DIO.Cry1::EGFP and returned to the PMTs for a further 7 days, followed by a media change and return to PMTs for a final 7 days. Pre- and post-analysis of Per2::Luciferase rhythms used the first and last 7 days respectively. After PMT recordings, slices were imaged on CCD camera. Finally, the slices were transduced with AAV8.Syn1.DIO.mCherry to confirm Cre-recombinase expression for one week before being fixed (4% PFA in phosphate buffer) and mounted using Vectashield with DAPI (Vectorlabs, U.S.A) on slides ready for confocal imaging. Cell counting was performed in FIJI using the “nucleus counter” plugin. %EGFP cells was calculated relative to DAPI cell counts. Data were analysed using BioDARE⁴⁹. Comparisons of period between pan-neuronal and VIP^Cre /VPAC2^Cre initiations were made via unpaired two-tailed t-tests. GOF measures were compared via a repeated-measures two-way ANOVA with Sidak’s correction for multiple comparisons. Trajectory area and Rayleigh vector data was compared by ordinary one-way ANOVA with Kruskal–Wallis correction for multiple comparisons.

The sex of SCN slices was not known. Therefore, both male and female adult (aged between 10 to 14 weeks) mice expressing both VIP^Cre and VPAC2^Cre on a Cry1,2-null background from two separate cohorts (n = 4 male, 4 female; N = 8) were individually housed in cages equipped with running wheels with food and water available ad libitum. As part of the breeding strategy, Cry1-null, Cry2-heterozygous animals were also produced and subjected to the same stereotaxic surgery to determine the in vivo pacemaking efficiency the VIP and VPAC2 cells (n = 3 male, 2 female; N = 5). Mice were singly caged in cages equipped with running wheels with environmental conditions complying with NC3Rs guidelines of ambient temperatures of between 20–24 °C and relative humidity between 45 and 65%. Mice were entrained to a 12-h light:12-h dark schedule (lights on at 07:00) for ca. 10 days, followed by 10 days in continuous darkness (DD) to assess their endogenous free-running period. The mice then received bilateral stereotaxic injections of AAV1 pCry1-DIO-Cry1::EGFP (0.3 µl per site) into the region of the SCN (±0.25 mm medio-lateral to Bregma, 5.5 mm deep to dural surface) under halothane anaesthesia. Following surgery, the mice were maintained on a 12-h light:12-h dark schedule for several days for recovery before transferring back to DD for re-assessment of their endogenous behavioural period. To confirm that administration of AAV particles to the SCN led to expression in the VIP and VPAC2 cell populations within the SCN, mice were culled and brain harvested, fixed in 4% PFA, cryo-preserved overnight in 20% sucrose and sectioned (40 µm) on a sledge freezing-microtome. Confocal microscopy of native EGFP signal was used to identify successful targeting of the SCN. Due to inefficient targeting, two animals were removed from further analysis in the Cry1,2-null cohort, and three animals were removed from the Cry1-null, Cry2-heterozygous cohort.

Behavioural data were analysed using ClockLab (Actimetrics, Inc., USA) running within Matlab (Mathworks, USA). Activity profiles were generated by outputting the last 7-days of recordings from ClockLab binned at 12-min intervals across a 27 h day (the average post-surgery period) for both pre-and post surgery. Profiles were then averaged to generate a single 27 h profile for each animal which was then normalised to the highest peak in the activity trace. Traces were then aligned so that their activity onsets were centered at the mid-point of the 27 h day, and averaged across all animals to generate an average activity profile.

For Cry1,2-null cohorts, paired comparisons were made between pre- and post-surgery conditions using a paired one-tailed Student’s t-test. A one-tailed test was deemed appropriate due to a priori assumptions that can be made from the ex vivo SCN data. Comparisons were made between slice and behavioural data using a two-tailed unpaired t-test. For Cry1-null,Cry2-heterozygous cohort, comparisons were not made against slice data because the final number of animals (n = 2) was deemed too low for the statistical analysis to be reliable.

Further information on research design is available in thelinked to this article. Nature Research Reporting Summary

Supplementary Information Peer Review File Reporting Summary Description of Additional Supplementary Files Supplementary Movie 1