Hypothalamic opsin 3 suppresses MC4R signaling and potentiates Kir7.1 to promote food consumption

Abstract

Mammalian opsin 3 (OPN3) is a member of the opsin family of G-protein-coupled receptors with ambiguous light sensitivity. OPN3 was first identified in the brain (and named encephalopsin) and subsequently found to be expressed in other tissues. In adipocytes, OPN3 is necessary for light responses that modulate lipolysis and glucose uptake, while OPN3 in human skin melanocytes regulates pigmentation in a light-independent manner. Despite its initial discovery in the brain, OPN3 functional mechanisms in the brain remain elusive. Here, we investigated the molecular mechanism of OPN3 function in the paraventricular nucleus (PVN) of the hypothalamus. We show that Opn3 is coexpressed with the melanocortin 4 receptor (Mc4r) in a population of PVN neurons, where it negatively regulates MC4R-mediated cAMP signaling in a specific and Gα_i/o-dependent manner. Under baseline conditions, OPN3 via Gα_i/o potentiates the activity of the inward rectifying Kir7.1 channel, previously shown to be closed in response to agonist-mediated activation of MC4R in a Gα_s-independent manner. In mice, we found that Opn3 in Mc4r-expressing neurons regulates food consumption. Our results reveal the first mechanistic insight into OPN3 function in the hypothalamus, uncovering a unique mechanism by which OPN3 functions to potentiate Kir7.1 activity and negatively regulate MC4R-mediated cAMP signaling, thereby promoting food intake.

Keywords: G-protein coupled receptor (GPCR); Opsin 3 (OPN3); cAMP signaling; melanocortin 4 receptor (MC4R).

Fig. 1.

Opn3 and Mc4r are coexpressed in a subset of PVN neurons. (A and B) Expression of Opn3 in anterior (A) and posterior (B) regions of the PVN in P180 Opn3^CreERT2; Ai14 reporter mice (pseudocolored green). (Scale bars: 100 µm.) (C and D) Expression of Opn3 and Mc4r in P35 mice visualized with Opn3-eGFP (green) and Mc4r^2a-Cre; Ai14 (red) reporters in anterior (C) and posterior (D) PVN. (Scale bars: 100 µm.) (E and F) Higher magnifications of the regions identified with corner brackets in C and D. The cells that coexpress both Opn3 and Mc4r reporters appear yellow. (Scale bars: 50 µm.) (G) Higher magnification of the region identified with corner brackets in E, separated into red and green color channels to exclude false positives. The arrows correspond to two cells that appear yellow—marked by dashed ellipses in E—but do not coexpress Opn3-eGFP and Mc4r^2a-Cre; Ai14. (Scale bars: 25 µm.) (H and I) Higher magnifications of the regions identified with corner brackets in F, separated into red and green color channels. The dashed outline in F identifies a domain of the PVN with a high proportion of Opn3 and Mc4r coexpressing neurons. The dashed outlines in the red channel (Mc4r reporter, bottom panels) superimposed to the green channel (Opn3 reporter, top panels) identify individual neurons that do (H) or do not (I) coexpress Opn3 and Mc4r. (Scale bars 25 µm.) (J and K) Quantification of Opn3-eGFP and Mc4r^2a-Cre; Ai14 reporter coexpression indicates that ~13% of Opn3 positive (Opn3+) cells are Mc4r+ and ~11% of Mc4r+ cells are Opn3+ in the anterior (J) and ~28% and ~32% respectively in the posterior (K) PVN. Circles represent individual mice; n = 5 mice with n = 5 - 9 sections analyzed for each location; lines represent mean ± SEM. (L) DAPI-labeled cryosection from an adult C57BL/6J mouse showing (dashed line) the region of the PVN quantified for Opn3 and Mc4r expression using RNAscope. (Scale bar: 100 µm.) (M) Higher magnification of a region of PVN showing the distribution of probes for Opn3 (green) and Mc4r (red) transcripts. (Scale bar: 10 µm.) (N–P) Representative PVN neuron labeled for Opn3 (O, green puncta), Mc4r (P, red puncta), and DAPI (N, merged color channels). (Scale bar: 10 µm.) (Q) Quantification of Opn3 and Mc4r transcript coexpression in the PVN showed ~39% of Opn3+ cells expressed Mc4r and ~25% of Mc4r+ cells expressed Opn3. Circles represent individual regions of interest (ROIs) from n = 19 fields analyzed and n = 4 mice; lines represent mean ± SEM. (R) Representative image of the posterior PVN from a Bregma 0.0 to −1.0 image stack showing nuclei (DAPI, blue), neurons expressing Opn3 (Opn3-eGFP, green), neurons expressing Mc4r (Mc4r^2a-Cre; Ai14), and those labeled with the retrograde tracer cholera toxin (CTB, cyan) from an injection in the lateral parabrachial nucleus (LPB). (Scale bar: 50 µm.) (S–V) Magnified region of the image shown in r separated into the color for each label (S–U), and a merged image (V). (Scale bars: 50 µm.) (W) Venn diagram showing the average number of neurons (n = 3) for Bregma 0.0 to −1.0 of the PVN, for each of the overlapping labeled category.

Fig. 2.

OPN3 negatively regulates MC4R-mediated cAMP signaling. (A) Hypothalamic neuronal cells GT1-7 express similar mRNA levels of Opn3 and Mc4r relative to β-actin (Actb) mRNA. Bars are mean ± SEM; n ≥ 5 independent experiments. (B and C) GT1-7 cells expressing OPN3-targeting shRNA have ~90% less Opn3 mRNA (normalized to Gapdh, B) and ~70% lower OPN3 protein levels (normalized to β-actin, C) compared to CTL shRNA expressing cells. Bars represent mean ± SEM; n = 3 independent experiments; unpaired two-tailed t test, P < 0.0001. (D) Representative live-cell cAMP responses measured with the cADDis fluorescent indicator in GT1-7 expressing OPN3-targeted shRNA (OPN3 KD) or scrambled nontargeting shRNA as a control (CTL). MTII (1 µM) stimulation led to a higher amplitude cAMP response in OPN3 KD compared to NT cells, normalized to the maximal responses elicited by FSK+IBMX (F_norm). (E) Normalized maximal cAMP responses (Fmax) of individual cells (Left) or averaged cells from paired experiments (Right) were higher in OPN3 KD vs. CTL GT1-7 cells. Circles represent individual cells, n ≥ 49 cells/condition; bars represent mean with 95% CI, unpaired two-tailed t test, P < 0.0001; symbols represent average Fmax of all cells in an individual experiment, n ≥ 4 independent experiments. (F) Representative cAMP responses of OPN3 KD or CTL cells treated with PTx (200 ng/µL, 4 h preincubation) exhibited no difference in the amplitude of responses to MTII. (G) Normalized maximal cAMP responses (Fmax) of individual cells (Left) or averaged cells from paired experiments (Right) were similar for OPN3 KD and CTL cells treated with PTx and stimulated with MTII. Circles represent individual cells, n ≥ 40 cells; bars represent mean with 95% CI, unpaired two-tailed t test, P = 0.1064; symbols represent average Fmax of all cells in an individual experiment, n ≥ 5 independent experiments. (H) Activation of endogenous β1-AR with isoproterenol (Iso; 7 µM) evoked cAMP responses with similar amplitudes in representative OPN3 KD and CTL cells. (I) Iso elicited similar cAMP amplitudes in individual cells (Left) or paired experiments (Right) for OPN3 KD and CTL GT1-7 cells. Circles represent individual cells, n ≥ 50 cells; bars represent mean with 95% CI, unpaired two-tailed t test, P = 0.1662; symbols represent averaged experiments, n = 6 independent experiments. (J) Representative cAMP responses exhibiting lower cAMP amplitude of GT1-7 cells expressing OPN3-mCh (OPN3mCh) vs. mCh (control) and stimulated with MTII (1 µM). (K) The means of individual cell responses (Left) or averaged responses in paired experiments (Right) were decreased in OPN3-mCh compared to mCh-expressing cells. Circles represent individual cells, n ≥ 44; bars represent mean with 95% CI, unpaired two-tailed t test, P < 0.0001; symbols represent averaged cells from one experiment, n ≥ 4 independent experiments. (L) Representative single-cell cAMP responses to MTII (1 µM) for OPN3-mCh vs. mCh expressing GT1-7 cells were similar when the cells were treated with PTx (PTx; 200 ng/µL, 4 h preincubation). (M) Mean cAMP amplitudes of PTx-pretreated and MTII-stimulated OPN3-mCh-expressing individual cells r (Left) or averaged responses from paired experiments (Right) were not significantly different compared to mCh-expressing cells. Circles represent individual cells, n ≥ 40; bars represent mean with 95% CI, unpaired two-tailed t test, P = 0.1404; symbols represent averaged cell responses for individual experiments, n ≥ 5 independent experiments. (N) Representative OPN3-mCh and mCh-expressing cells exhibited no difference in the amplitude of cAMP responses to Iso (7 µM)-mediated activation of β2-ARs. (O) Iso (7 µM) stimulation elicited similar averaged normalized maximal cAMP responses of individual cells (Left) or averaged cells from paired experiments (Right) in OPN3-mCh vs. mCh cells. Circles represent individual cells, n ≥ 35; bars represent mean with 95% CI, unpaired two-tailed t test, P = 0.5712; symbols represent averages from n ≥ 5 individual experiments.

Fig. 3.

OPN3 forms a complex with MC4R. (A) Representative bright field and fluorescence confocal images of GT1-7 cells expressing FLAG-tagged OPN3 (OPN3-FLAG) and mCh tagged MC4R (MC4R-mCh) and immunostained with anti-FLAG antibodies showed colocalization at the plasma membrane and in intracellular compartments. (Scale bar: 10 µm.) (B) The calculated Pearson’s correlation coefficient (r) for OPN3-FLAG and MC4R-mCh colocalization was >0.8. Circles represent individual cells; n ≥ 14 cells from ≥3 independent experiments; bars represent mean ± SEM. (C) In HEK293T cells expressing MC4R-HA and OPN3-FLAG, immunoprecipitation (IP) with an anti-FLAG antibody and western blot (WB) with anti-HA revealed a band at ~45 kDa corresponded to MC4R-HA in the IP and lysate lanes for HEK293T cells transfected with MC4R-HA+OPN3-FLAG. Representative of n = 3 independent experiments. (D) In HEK293T cells expressing FLAG-tagged β1-AR (β1AR-FLAG) and HA-tagged OPN3 (OPN3-HA) IP with an anti-HA antibody followed by WB with anti-FLAG revealed a ~70 kDa band corresponding to β1AR-FLAG in the lysates of HEK293T cells expressing β1AR-FLAG alone and OPN3-HA+β1AR-FLAG (Top Left) but not in the IP lanes (Top Right). WB with anti-HA antibodies confirmed the presence of OPN3-HA both in the lysates and IP lanes from cells expressing OPN3-HA alone or OPN3-HA+β1AR-FLAG. Representative of n = 3 independent experiments.

Fig. 4.

Opn3, Mc4r, and Kcnj13 transcripts overlap in the mouse PVN. (A) Schematic of PVN in the hypothalamus. (B) DAPI-labeled (blue) cryosection from an adult C57BL/6J mouse showing the region of the PVN quantified for expression of Opn3 (green), Mc4r (red), and Kcnj13 (magenta) using RNAscope. (Scale bar: 10 µm.) (C–F) Higher magnification of a region of PVN showing the distribution of probes for Opn3 (green), Mc4r (red), and Kcnj13 (magenta) transcripts overlapped (C) or in separate images for each probe (D–F). (Scale bar: 10 µm.) (G and H) Quantification of Opn3, Mc4r, and Kcnj13 transcripts. Of the Opn3+ cells, ~53% contained Kcnj13 transcripts, ~37% expressed Kcnj13 but not Mc4r transcripts, ~15% expressed Mc4r but not Kcnj13, and ~9% contained both Kcnj13 and Mc4r transcripts. Circles represent either individual ROIs (G) from n = 19 fields analyzed or averages of ROIs from an individual animal (H) n = 4 animals; lines represent mean ± SEM.

Fig. 5.

OPN3 negatively regulates MC4R-Kir7.1 functional interaction in a Gα_i/o-dependent manner. (A) Representative images of live-cell Tl⁺ flux through potassium channels at different time points after addition to extracellular Tl⁺. (B) The change in fluorescence (∆F/F₀) of a single cell before and after addition of extracellular Tl⁺ was represented as a function of time and used to calculate the slope of the initial fluorescence increase (dotted line) that is proportional to the K⁺ flux through the membrane. (C) The initial slope of the Tl⁺ flux was significantly smaller in GT1-7 treated with MTII (0.5 µM) and further decreased by incubation with the Kir7.1 channel blocker VU590 (7 µM), compared to vehicle-treated cells. Treatment with vehicle control is indicated by (−) and treatment with MTII or VU590 is indicated by (+). Circles represent individual cells; n ≥ 75 cells per condition from ≥3 independent experiments; lines represent mean with 95% CI. The data were analyzed using two-way ANOVA with post hoc Tukey’s multiple comparisons test; P < 0.05 or P < 0.0001. (D) In unstimulated cells the initial slope of the Tl⁺ flux responses was significantly lower in OPN3 KD cells compared to CTL cells and further decreased in response to MTII (0.5 µM). Circles represent individual cells; n ≥ 129 cells from ≥3 independent experiments; lines represent mean with 95% CI. Data were analyzed using two-way ANOVA with post hoc Tukey’s multiple comparisons test; P < 0.05 or P < 0.0001. Treatment with vehicle control is indicated by (−) and treatment with MTII is indicated by (+). (E) WB analysis showed higher Kir7.1 protein levels in OPN3 KD vs. CTL GT1-7 cells (Top). Densitometry analysis of WBs probed with Kir7.1 and β-actin antibodies showed 2.68-fold higher Kir7.1 protein levels in OPN3 KD compared to CTL cells. Bars represent mean ± SEM; n = 3 independent experiments; unpaired two-tailed t test, P = 0.0224). (F) PTx treatment reduced the Tl⁺ flux response in CTL cells to levels similar to OPN3 KD cells. OPN3 KD cells treated with PTx or vehicle did not exhibit a significant change in the Tl⁺ flux response (n ≥ 110 cells from ≥ 3 independent experiments). (G) CTL and MC4R KD cells have similar basal Tl⁺ flux responses that are comparably reduced by PTx treatment. Circles represent individual cells, n ≥ 192 cells from ≥3 independent experiments; lines represent mean with 95% CI. Tl⁺ flux initial slope data were analyzed using two-way ANOVA with post hoc Tukey’s multiple comparisons test, P < 0.05 or P < 0.0001. Pretreatment with vehicle control is indicated by (−) and pretreatment with PTx is indicated by (+). (H) Representative whole-cell current traces from HEK293T cells transfected with Kir7.1(M125R), Kir7.1(M125R)+OPN3, and nontransfected in response to the voltage step protocol indicated. (I) Average current-voltage (I-V) relationships showed significantly increased currents in cells expressing Kir7.1(M125R)+OPN3. Data were analyzed using repeated measures one-way ANOVA with post hoc Tukey’s multiple comparisons test, P = 0.0008. *Corresponds to the comparison between Kir7.1(M125R) vs. Kir7.1(M125R)+OPN3, P = 0.0012, and # corresponds to the comparison between nontransfected vs. Kir7.1(M125R)+OPN3, P = 0.0066.

Fig. 6.

OPN3 can form a complex with Kir7.1. (A) Representative fluorescence confocal images of GT1-7 cells expressing OPN3-FLAG and Myc tagged Kir7.1 (Kir7.1-Myc) and immunostained with anti-FLAG and anti-Myc antibodies showed colocalization at the plasma membrane and in intracellular compartments. (Scale bar: 10 µm.) (B) The calculated Pearson’s correlation coefficient (r) for OPN3-FLAG and Kir7.1-Myc colocalization was ~0.75. Circles represent individual cells; n ≥ 14 cells from ≥3 independent experiments; bars represent mean ± SEM averaged across three individual experiments. (C) In cells expressing Kir7.1-Myc and OPN3-FLAG IP with anti-Myc antibodies revealed a band at ~36 kDa corresponding to OPN3-FLAG in the IP (Top Right) and lysate lanes (Top Left) of HEK293T cells expressing Kir7.1-Myc and OPN3-FLAG. As a positive control, Kir7.1-Myc was detected in the corresponding lanes in the lysates (Middle Left) and also IP-ed by the anti-Myc antibody (Middle Right). The ~55 kDa band in all four lanes of the Middle Right panel corresponds to the heavy chain of the IP antibodies. Representative of n = 5 independent experiments. (D) HEK293T cells expressing MC4R-HA, OPN3-FLAG, and Kir7.1-Myc were treated with MTII (1 µM) or vehicle for 10 min before IP with anti-HA antibodies. The lysates and immunoprecipitates were blotted with anti-FLAG (Top), anti-Myc (Middle-Left) or anti-Kir7.1 (Middle-Right), and anti-HA (Bottom) antibodies. We detected similar bands for OPN3-FLAG (Top Right) and Kir7.1-Myc (Middle Right) when immunoprecipitating MC4R-HA (Bottom Right), regardless of MTII treatment (lanes 4 vs. lanes 5 Right blots). Representative of n = 4 independent experiments.

Fig. 7.

Conditional deletion of Opn3 from MC4R-expressing neurons decreases food intake and locomotion. (A) Averaged 72 h food consumption for Mc4r^2a-Cre; Opn3^fl/fl and Mc4r^2a-Cre; Opn3^fl/+ mice was significantly decreased compared to Opn3^fl/+ mice. Lines represent mean ± SEM averaged across animals; n ≥ 11; data were analyzed using two-way ANOVA, P < 0.0001. (B and C) Average 12 h food intake for individual mice was not significantly different between Mc4r^2a-Cre; Opn3^fl/fl and Mc4r^2a-Cre; Opn3^fl/+ compared to Opn3^fl/+mice in both light phase (B) and dark phase (C). Circles represent individual mice; n ≥ 9 mice; lines represent mean ± SEM; data were analyzed using one-way ANOVA or Kruskal–Wallis, P > 0.05. (D) Averaged 24 h food consumption was significantly decreased in Mc4r^2a-Cre; Opn3^fl/fl and Mc4r^2a-Cre; Opn3^fl/+ compared to Opn3^fl/+mice. Lines represent mean ± SEM averaged across animals; n ≥ 11; data were analyzed using two-way ANOVA, P < 0.0001. (E) Locomotor activity of Mc4r^2a-Cre; Opn3^fl/fl and Mc4r^2a-Cre; Opn3^fl/+ was significantly lower than Opn3^fl/+mice. Data were analyzed using two-way ANOVA, P < 0.0001. (F) Body weight of Opn3^fl/+, Mc4r^2a-Cre; Opn3^fl/fl, and Mc4r^2a-Cre; Opn3^fl/+ mice at 5 mo of age were not significantly different. Circles represent individual mice; lines represent mean ± SEM; data were analyzed using one-way ANOVA, P > 0.05.

Fig. 8.

Schematic of OPN3 signaling and function in the hypothalamus. (A) On a molecular level, OPN3 can form a complex with Kir7.1 and constitutively potentiates Kir7.1 channels in a Gα_i/o-dependent manner. (B) OPN3 negatively regulates MC4R-mediated cAMP signaling in response to αMSH. (C) Building on feeding circuitry in the hypothalamus (–, , –66), our findings reveal that Opn3 and Mc4r significantly overlap in neurons of the PVN. Opn3 functions in Mc4r-expressing neurons to regulate food consumption. Abbreviations: POMC: proopiomelanocortin, AC: adenylyl cyclase, ARC: arcuate nucleus, 3V: third ventricle, PBN: parabrachial nucleus. This figure was created using Biorender.com and anatomy was adapted from the Allen brain atlas.