A Gq-Ca2+ axis controls circuit-level encoding of circadian time in the suprachiasmatic nucleus

Abstract

The role of intracellular transcriptional/post-translational feedback loops (TTFL) within the circadian pacemaker of the suprachiasmatic nucleus (SCN) is well established. In contrast, contributions from G-coupled pathways and cytosolic rhythms to the intercellular control of SCN pacemaking are poorly understood. We therefore combined viral transduction of SCN slices with fluorescence/bioluminescence imaging to visualize GCaMP3-reported circadian oscillations of intracellular calcium [Ca2+]i alongside activation of Ca2+ /cAMP-responsive elements. We phase-mapped them to the TTFL, in time and SCN space, and demonstrated their dependence upon G-coupled vasoactive intestinal peptide (VIP) signaling. Pharmacogenetic manipulation revealed the individual contributions of Gq, Gs, and Gi to cytosolic and TTFL circadian rhythms. Importantly, activation of Gq-dependent (but not Gs or Gi) pathways in a minority of neurons reprogrammed [Ca2+]i and TTFL rhythms across the entire SCN. This reprogramming was mediated by intrinsic VIPergic signaling, thus revealing a Gq/[Ca2+]i-VIP leitmotif and unanticipated plasticity within network encoding of SCN circadian time.

Figure 1

Circadian Activation of Ca²⁺/cAMP-Responsive Elements (CRE) in the SCN (A) Representative SCN slice transduced with LV:CRE-luciferase, showing numerous cells distributed across SCN that exhibit circadian oscillations (inset), reflected in the aggregate bioluminescence plot before (left) and after linear detrending (right). (B) Semiautomatic image analysis (SARFIA) reveals rhythmic CRE activation at single cell level, expressed as raster plot (left). CRE oscillations have “saw-tooth” waveform and slightly different phases across the slice (right). (C) Representative traces of PMT bioluminescence recording of WT, Fbxl3^Afh/Afh and CK1ε^Tau/Tau SCN transduced with the LV:CRE-luc reporter. All data mean + SEM, p < 0.0001, ANOVA with Bonferroni correction. (D) Representative trace of bioluminescence recording from SCN transduced with LV:CRE-luc and treated with TTX (0.5 μM) showing decreased amplitude and robustness (increased RAE values, right panel, mean ± SEM, p < 0.0001, n = 11, paired two-tailed t test). (E and F) De-trended traces and Rayleigh plots of representative WT and VIP^−/− SCN transduced with LV:CRE-luc, showing impaired overall rhythmicity and disrupted internal coupling of CRE oscillators (r mean vector CRE WT = 0.78, n = 120; VIP^−/− = 0.24, n = 72). (G) Cumulative plots reveal increase in period scattering and minimal robustness of the CRE rhythm (high RAE). Scale bars: 50 μm. See also Movie S1.

Figure 2

Selective Roles for Gs, Gi, and Gq Signaling Pathways in Controlling Circadian Rhythms of CRE Activation in SCN (A) Representative bioluminescence recordings from SCN slices transduced with the LV:CRE-luc reporter and LV-DREADDs: Syn-hM3DGq-IRESmCherry, LV:Syn-rM3/β1Gs-IRESmCherry or LV:Syn-hM4DGi-IRESmCherry, respectively. CNO (100 nM) triggered Gq- and Gs-stimulated increases or a Gi-stimulated decrease of CRE activation. Insets show detrended traces to reveal the circadian oscillations before and in the presence of CNO. (B) Relative induction/suppression of CRE activation plotted by day of treatment (mean ± SEM, n = 3–6 per group). For all CNO/DREADD treatments, two-way ANOVA, p < 0.5 = ^*; p < 0.01 = ^**; p < 0.001 = ^***p < 0.001 = ^**** versus initial activity. (C) Gq and Gi activation decreased the amplitude of the CRE rhythm, whereas no significant changes are observed following Gs stimulation (mean + SEM, two-tailed t test). (D) Activation of Gq pathway significantly lengthened period of CRE-luc rhythm (p < 0.01 n = 6). Stimulation of Gs or Gi signaling had no significant effect on period (mean + SEM, n = 3–6 per group, two-way ANOVA, ^** = p < 0.01 versus other measures). See also Figure S1.

Figure 3

Gq Activation Reorganizes Circadian CRE Activation and Per1 Expression across the Entire SCN Circuit (A) Representative SCN transduced with LV:CRE-luc and Syn-hM3DGq-IRESmCherry: merge reveals single and double transduced cells. (B) CNO-stimulated CRE activation was present both in Syn-hM3DGq⁺-cells (red, left) and the hM3DGq⁻ population (gray, middle). Note significantly slower kinetics of induction in DREADD-negative cells (right). (C) Representative traces from Per1-luc SCN transduced with the LV:Syn-hM3DGq-IRESmCherry and treated with either vehicle (black) or CNO (red) (100 nM, CNO) followed by washout (wo) with fresh medium after 5 days. (D) Mean data (+SEM) of circadian period before, during and after treatment with vehicle (black/gray) or CNO (red). Note significant lengthening with CNO (p < 0.05 n = 4, two-way ANOVA with Bonferroni correction), but not vehicle (n = 5) that is sustained after washout (p < 0.01). (E) Mean data (+SEM) of relative amplitude of circadian CRE activation revealed ca. 50% reduction both during CNO (left, p < 0.01 two-tailed t test) and after CNO removal (right, p < 0.001). Scale bar = 50 μm.

Figure 4

Circadian Rhythm of Intracellular Calcium in SCN, Phase-Mapped to the Circadian TTFL (A) Serial images of SCN slice transduced by AAV-Syn-GCaMP3 showing high rates of neuronal transduction and high GCaMP3 expression levels (inset: single cells in the slice). Below is aggregate circadian rhythm of [Ca²⁺]_i before (left) and after (right) de-trending (period = 23.81 ± 0.12 hr, mean ± SEM, n = 14). (B) SARFIA analysis showing rhythms in single cells plotted as raster (left) and graphical (right) plots. Note that the distinctive waveform of the GCaMP3⁺ aggregate trace is evident in individual neurons. (C) Detrended traces and Rayleigh plots of representative WT and VIP^−/− SCN transduced with AAV:Syn-GCaMP3, showing similar results to the CRE (r mean vector GCaMP3: WT = 0.66 n = 191, VIP^−/− = 0.18, n = 112). (E) Cumulative plots of GCaMP3 oscillations displaying increase in period scattering and minimal robustness of the rhythm (high RAE) in VIP^−/− SCN. (F) Serial images from Per2:luc knock-in SCN transduced with AAV-Syn-GCaMP3, showing different peaking times. (G) Circadian rhythms of bioluminescence and fluorescence from SCN transduced with AAV:Syn-GCaMP3 and expressing CRE-luc, Per1-luc, Per2:luc or Cry1-luc reporters, respectively. Data plotted as mean ± SEM (n ≥ 3) for each configuration. Interpeaks distance (ΔCT versus GCaMP3) is used to align the various rhythms around the circadian day. (H) Phase map of circadian cytosolic oscillations and TTFL based on normalized cycles for each component. Scale bars: 50 μm. Inset: 10 μm. See also Movies S2 and S3.

Figure 5

Selective Reprogramming of [Ca²⁺]_i Circadian Rhythms by Gq Signaling across the SCN Circuit (A, F, and K) Representative SCN slices transduced with AAV:Syn-GCaMP3 and LV:Syn-hM3DGq-IRESmCherry, LV:Syn-rM3/β1Gs-IRESmCherry or hM4DGi-IRESmCherry, respectively. Microphotographs showed that only a subset of the GCaMP3⁺ neurons also expressed the DREADD receptors. (B) Detrended trace of the SCN aggregate fluorescence signals from (A) before, during and after CNO treatment, shows irreversible reduction of amplitude of GCaMP3 oscillations. (C) Raster plots and representative traces of GCaMP3 from individual cells in (A) showing both loosened overall synchrony and reduced amplitude and robustness in single cells. (D) Rayleigh plots confirm desynchronization of SCN cells within the slice induced by Gq, an effect not reversed by CNO removal. (E) Cumulative plots showing that Gq activation irreversibly increases period and decreases the amplitude and coherence (RAE) of individual cellular rhythms. (G) Detrended trace of the aggregate fluorescence signals from the culture in (F) shows that activation of Gs signaling did not alter significantly overall GCaMP3-reported [Ca²⁺]_i rhythms. (H) Raster plots and representative traces of cellular GCaMP3 recordings show that Gs stimulation did not affect cellular GCaMP3 fluorescence rhythms. (I) Rayleigh plots reveal slight reduction of the phase coupling of cellular [Ca²⁺]_i rhythms, elicited by Gs stimulation. (J) Cumulative plots demonstrate that stimulating Gs has little effect on cellular period, amplitude, or coherence (RAE). (L) Detrended trace of the aggregate fluorescence signals from the SCN in (K) in the presence of CNO and after drug removal. Stimulation of Gi transiently diminished the overall GCaMP3 rhythm amplitude, but CNO removal reversed the phenotype, with rhythmic amplitude exceeding the pretreatment values. (M) Raster plots and representative traces of rhythms show that Gi reduced [Ca²⁺]_i amplitude in single cells, but this effect was reversed on CNO removal. (N) Rayleigh plots displaying transient reduction of the phase coupling of GCaMP3⁺ oscillators by Gi, reversed by CNO washout. (O) Cumulative plots showing transiently decreased amplitude and increased RAE of the [Ca²⁺]_i cellular rhythms in the presence of CNO, returned to, or even exceeded original pretreatment levels, after drug removal. Scale bars: 50 μm. See also Table S1, Figure S2, and Movies S4, S5, and S6.

Figure 6

Gq Activation Reprograms the Global Spatiotemporal Dynamics of the SCN TTFL (A and B) Recordings of Per2:luc and Cry1-luc bioluminescence (mean ± SEM, n = 5–13) from SCN transduced with LV:Syn-hM3DGq-IRESmCherry. Addition of Gq stimulation by CNO significantly increased period (p < 0.01, two-way ANOVA with Bonferroni correction) and decreased the rhythms’ amplitude (two-tailed t test). (C and E) Left: Representative Per2:luc and Cry1-luc SCN transduced with LV:Syn-hM3DGq-IRESmCherry and AAV:Syn-GCaMP3. Syn-hM3DGq-IRESmCherry⁺ cells constituted only a relatively small fraction of the neurons within the cultures (Per2:luc: DREADD⁺ 155 ± 21 n = 3; Cry1-luc: DREADD⁺ 134 ± 31 n = 3), when compared to GCaMP3⁺, Per2⁺ and Cry1⁺. Right: Addition of CNO alters GCaMP3, Per2:luc and Cry1-luc oscillations. (D and F) Representative serial images and Poincaré plots depict the progression of the center of luminescence (CoL, white spot) for 3 days before (gray plots, left panels) and after CNO addition (red plots, right panels). Both reporters consistently showed a reduction in the dorsomedial dynamics of the CoL. Plots are representative of at least 3 SCN for each reporter. Scale bars: 50 μm. See also Movies S7 and S8.

Figure 7

Reprogramming the Pacemaker by Gq Signaling Requires the Intrinsic, VIP-Mediated Coupling of the SCN Network (A) Group plots of bioluminescence from VIP^−/−, Per2:luc SCN transduced with LV:Syn-hM3DGq-IRESmCherry. Note weak, damping oscillations (blue plots) restored by WT SCN graft (black plots) and no effect exerted by Gq activation (red plot) before and after CNO washout (gray plot). (B) Group data showing that activation of Gq signaling by CNO (100 nM) did not affect period, amplitude or robustness, neither in the presence of CNO (red plots), nor after washout (gray plots). (C and D) As for (A) and (B) but with SCN transduced with LV:CRE-luc. Grafting caused an immediate increase in baseline and rhythm of CRE activity. Activation of DREADD-mediated Gq by CNO (100 nM) (red), acutely induced CRE but no effects on circadian properties were observed. (All data mean + SEM, n = 5; period: two-way ANOVA, amplitude and RAE ratios: two-tailed t test.) See also Figure S3.

Figure 8

Direct Activation of Gq Signaling in VIPergic Neurons Mirrors the Circuit Reprogramming Elicited by Untargeted Gq Activation (A) Representative microphotographs from VIP:IRESCre⁺/EYFP⁺ SCN cultures stained with polyclonal AVP antiserum, delineating a clear core/shell anatomical partitioning. (B) Representative microphotographs from VIP-IRESCre⁺/ EYFP⁺ SCN slices transduced with AAV:DIO-Syn-hM3DGq:mCherry vectors. hM3DGq:mCherry receptor fusions are localized in the plasma membrane (see insets) of VIP:Cre/EYFP⁺ neurons, thus confirming both effective and specific targeting of the VIPergic SCN subpopulation. (C) Representative trace of Per2:luc⁺/VIP:IRESCre⁺ SCN slices transduced with DIO-Syn-hM3DGq:mCherry⁺ and activated by 100 nM CNO. CNO addition increased the baseline, lengthened the period (1.3 ± 0.3 hr) and reduced amplitude of the oscillations (mean ± SEM, n = 7 p < 0.01 paired two-tailed t test). (D) Microphotographs from representative VIP:IRESCre⁺/EYFP⁺ SCN slice transduced with DIO-Syn-hM3DGq:mCherry. Gq signaling was activated by CNO and network effects analyzed by CoL analysis (right panels), showing dorsomedial compression of Per2:luc dynamics. Plots are representative of at least 3 SCN for each reporter. Scale bars: 50 μm. Inset: 10 μm. See also Movie S9.