Research ArticleVision

Heteromeric MT1/MT2 Melatonin Receptors Modulate Photoreceptor Function

See allHide authors and affiliations

Science Signaling  08 Oct 2013:
Vol. 6, Issue 296, pp. ra89
DOI: 10.1126/scisignal.2004302

Abstract

The formation of G protein (heterotrimeric guanine nucleotide–binding protein)–coupled receptor (GPCR) heteromers enables signaling diversification and holds great promise for improved drug selectivity. Most studies of these oligomerization events have been conducted in heterologous expression systems, and in vivo validation is lacking in most cases, thus questioning the physiological significance of GPCR heteromerization. The melatonin receptors MT1 and MT2 exist as homomers and heteromers when expressed in cultured cells. We showed that melatonin MT1/MT2 heteromers mediated the effect of melatonin on the light sensitivity of rod photoreceptors in mice. This effect of melatonin involved activation of the heteromer-specific phospholipase C and protein kinase C (PLC/PKC) pathway and was abolished in MT1−/− or MT2−/− mice, as well as in mice overexpressing a nonfunctional MT2 mutant that interfered with the formation of functional MT1/MT2 heteromers in photoreceptor cells. Not only does this study establish an essential role of melatonin receptor heteromers in retinal function, it also provides in vivo support for the physiological importance of GPCR heteromerization. Thus, the MT1/MT2 heteromer complex may provide a specific pharmacological target to improve photoreceptor function.

Introduction

G protein (heterotrimeric guanine nucleotide–binding protein)–coupled receptors (GPCRs), also called “seven-transmembrane receptors,” form the largest protein family encoded by the human genome with about 800 members. GPCRs sense the extracellular environment and are involved in many cellular processes. The structural resolution of several GPCRs confirmed the high degree of conservation of their overall structure despite well-known ligand diversity ranging from photons, metabolites, lipids, and peptides to proteins (1). In addition, GPCRs are major drug targets accounting for up to 30% of currently marketed drugs (2).

Many reports indicate that GPCRs have the potential to interact with themselves (homomers) and with other GPCRs (heteromers). Structural studies have shown that some GPCRs crystallize as homodimers with several putative dimer interfaces, and the formation and function of these homodimers await confirmation in a physiologically relevant cellular environment (3). Although monomeric GPCRs represent the minimal signaling unit (4, 5), GPCR oligomerization, in particular heteromerization, may provide additional pharmacological and functional properties distinct from those of the individual receptors of which they are comprised (68). GPCR heteromers could provide additional pharmaceutical targets leading to improved drug selectivity by acting only on those cells coexpressing both receptors (9). Whereas there is compelling evidence for the existence of a number of GPCR heteromers in transfected cells, in vivo evidence is still lacking in most cases (10, 11), and their physiological relevance remains a matter of intense debate (12). Selected examples, for which strong in vivo evidence for GPCR heteromerization exist, underscore the great potential of GPCR heteromers as future therapeutic targets (1317).

Two members of the melatonin receptor subfamily in humans, melatonin receptor type 1 (MT1) and melatonin receptor type 2 (MT2), tend to homomerize and heteromerize in a constitutive manner when transfected in human embryonic kidney (HEK) 293T cells at physiological concentrations (18). Moreover, the propensity for homomer and heteromer formation differs between the two types of receptors. Whereas the propensity of human MT1/MT2 heteromer and MT1 homomer formation is similar, that of MT2 homomer formation is three- to fourfold lower, suggesting that the MT2 receptor preferentially exists as a heteromeric complex with MT1 (19). MT1 and MT2 receptors bind melatonin with similar affinity, and both inhibit the adenylyl cyclase pathway through Gi proteins (20, 21). The functional consequences of melatonin receptor heteromerization are currently unknown. The formation of MT1/MT2 heteromers has been proposed to occur in the retina and in other tissues where both receptors are detected (22). In humans, both melatonin receptors are located on rod photoreceptors and on ganglion cells, making these cells likely candidates for MT1/MT2 heteromer formation (2326).

Melatonin is synthesized during the night in the mammalian retina and reaches concentrations in the picomolar to low nanomolar range (27, 28). In the retina, melatonin plays an important role in the regulation of retinal physiology and pathophysiology [for review, see (29)]. Indeed, melatonin increases photoreceptor light sensitivity at night (3032) and may be implicated in the pathogenesis of age-related macular degeneration and glaucoma. Low concentrations of melatonin may increase the chance of developing age-related macular degeneration, and administration of melatonin improves age-related macular degeneration symptoms (3335).

The electroretinogram (ERG), consisting mainly of an a-wave and a b-wave, is a commonly used method to assess retinal functioning. In the dark-adapted ERGs, the a-wave represents the response of the photoreceptors to a flash of light, and the b-wave represents the response of the bipolar cells (36). The amplitudes of the a-wave and b-wave can be used to determine the effects of genetic mutations or pharmacological treatments on specific retinal cell types (36, 37). Administration of exogenous melatonin to mice during the day increases the amplitudes of the a-wave and b-wave of the scotopic ERG to values observed at night under control conditions (32); animals deficient in MT1 lack this response to exogenous melatonin (32). Furthermore, mice deficient in MT1 lacked the daily and circadian rhythms in the ERG responses (32,38) and exhibited decreased photoreceptor and ganglion cell viability during aging (32, 35).

Results

Genetic deficiency in MT2 replicates the ERG phenotype of MT1 deletion in mice

If MT1/MT2 heteromers are the functionally relevant signaling unit in photoreceptor cells, then deletion of either component of the heteromer should result in the same phenotype. We generated C3H-f+/+ MT2−/− mice and compared the scotopic ERGs of these mice with those of previously described C3H-f+/+ MT1−/− mice (32). As previously reported (32) in C3H-f+/+ wild-type mice, the amplitudes of the a-wave and b-wave were significantly greater at midnight (ZT18) than at midday (ZT6) (Fig. 1, A to C), whereas in C3H-f+/+ MT2−/− mice, no differences in the amplitude of the a-wave and b-wave were observed between ZT18 and ZT6 (Fig. 1D). Thus, a daily rhythm in the amplitude of the a-wave and b-wave was observed in wild-type mice but not in MT2−/− mice.

Fig. 1 Absence of MT2 replicates the ERG phenotype of MT1 deletion in mice.

(A and B) Representative ERG traces recorded at ZT18 and ZT6 in C3H-f+/+ [wild-type (WT)] mice. (C and D) Quantification of dark-adapted ERG responses to flashes of light recorded in the middle of the day (ZT6, • and ▪) and middle of the night (ZT18, ○ and □) in C3H-f+/+ (C) and MT2−/− (D) mice. Mice (3 to 4 months old) were dark-adapted for at least 30 min before the recordings were performed. Data are presented as means ± SEM; n = 6 for each time point and genotype. Differences in the a-wave and b-wave at ZT6 and ZT18 for the WT mice are statistically different [two-way analysis of variance (ANOVA), P < 0.01], but not for the MT2−/− mice (two-way ANOVA, P > 0.1). Luminance is expressed in candela-seconds per meter squared (cd*s/m2).

We then tested whether intraperitoneal injection of melatonin (1 mg/kg) during the day (ZT6) affected the dark-adapted ERG. C3H-f+/+ MT2−/− mice were injected with melatonin and dark-adapted for 1 hour before recording the ERG. Contrary to what was observed in C3H-f+/+ mice (Fig. 2A), melatonin injection during the daytime did not alter the ERG of MT2−/− mice (Fig. 2B). Thus, the ERG data obtained with the MT2−/− mice were identical to those previously obtained with the MT1−/− (32) and, thus, are compatible with our hypothesis of the involvement of MT1/MT2 heteromers in photoreceptor function.

Fig. 2 Administration of melatonin does not affect ERGs in MT2−/−.

(A and B) Quantification of the dark-adapted response of the a-wave (upper panels) and b-wave (lower panels) to flashes of light recorded after 1 hour of dark adaptation and intraperitoneal injection of melatonin (1 mg/kg) or vehicle in C3H-f+/+ (A) and MT2−/− (B) mice. Data are presented as means ± SEM; n = 5 to 8 for each group. Melatonin injection did not induce any significant changes in the amplitude of the a-wave or b-wave in MT2−/− mice (two-way ANOVA, P > 0.1). Luminance is expressed in cd*s/m2. ERGs in this experiment were performed at ZT6.

Murine MT1/MT2 heteromers form in transfected HEK293T cells

To directly assess the possible interaction between murine MT1 and MT2 receptors, we performed coimmunoprecipitation experiments with tagged forms of MT1 and MT2 expressed in HEK293T cells (Fig. 3, A to C). We observed several diffuse bands typical for glycosylated and hydrophobic seven-transmembrane proteins and detected homomeric interactions between each of the receptors (Fig. 3, A and B), as well as heteromeric interactions when the two receptors were expressed in the same cell (Fig. 3C). We also performed bioluminescence resonance energy transfer (BRET) donor saturation experiments with the Renilla luciferase 8 (Rluc8) energy donor fused to the C terminus of MT1 or MT2 and the yellow fluorescent protein (YFP) energy acceptor fused to the C terminus of MT1 or MT2. Cotransfection of a fixed amount of MT1-Rluc8 expression plasmid and increasing amounts of MT1-YFP or MT2-YFP expression plasmids resulted in the expected hyperbolic saturation curve with increasing YFP/Rluc ratios for all receptor combinations, reflecting a specific interaction between BRET donor and acceptor pairs (Fig. 3, D to F). MT1 homomers exhibited a 50% maximal BRET response (BRET50) at 0.71 ± 0.15, and the MT1/MT2 heteromers had a BRET50 of 1.67 ± 0.42 (n = 5). In cells expressing a fixed amount of MT2-Rluc8 and increasing amounts of MT2-YFP, the BRET50 for the MT2 homomers was 1.07 ± 0.24 (n = 5) (Fig. 3E), which is similar to that of the heteromers. Expression of the MT1-Rluc8 donor with the negative control vasopressin V2 receptor fuse to YFP resulted in a linear, nonsaturable BRET increase, characteristic of random interactions. Overall, coimmunoprecipitation and BRET results showed that murine MT1 and MT2 receptors can form homomers and heteromers in HEK293T cells as previously shown for their human homologs (18, 19).

Fig. 3 Formation of murine MT1/MT2 heteromers in transfected HEK293T cells.

(A to C) Coimmunoprecipitation of MT1 and MT2. Lysates from HEK293T cells expressing the indicated Rluc8 and YFP fusion proteins were immunoprecipitated (IP) with antibodies recognizing GFP (green fluorescent protein), and the presence of coprecipitated Rluc fusion proteins was detected with antibodies against Rluc [upper Western blot (WB)]. Expression of fusion proteins in lysates was assessed by immunoblotting with indicated antibodies (lower Western blot). Data are representative of three experiments. Anti-GFP and anti-Rluc: antibodies recognizing GFP and Rluc, respectively. The GFP antibody also recognizes YFP. (D to F) BRET donor saturation experiments with HEK293T cells coexpressing a fixed amount of MT1-Rluc8 (D and F) or MT2-Rluc8 (E) in the presence of increasing amounts of MT1-YFP, MT2-YFP, or vasopressin receptor V2-YFP (negative control). The saturation curves were obtained from five independent experiments.

MT1 and MT2 receptors are expressed in the mouse photoreceptor cells where they heteromerize

By fluorescence in situ hybridization, we detected MTNR1A (MT1) mRNAs in the outer nuclear layer (ONL) and the inner nuclear layer (INL) and in the ganglion cell layer of mouse retina (Fig. 4A), whereas MTNR1B2 (MT2) transcripts were detected in the ONL and INL (Fig. 4B). No signal was detected in sections treated with the sense probes (Fig. 4, C and D). In the ONL, MT1 and MT2 mRNAs were detected in most of the nuclei of the photoreceptors.

Fig. 4 MT1 and MT2 receptors are expressed in mouse photoreceptor cells where they heteromerize.

(A and B) MTNR1A (encoding MT1) (A) and MTNR1B (encoding MT2) (B) transcripts in the retina were detected by in situ hybridization using a fluorescein-labeled probe. (C and D) Signal from the sense MT1 probe (C) or sense MT2 probe (D). (E) Coimmunoprecipitation of MT1 and MT2 from the retina of WT C57/bl6 mice or transgenic C57/bl6 mice expressing either Flag-MT1 (MT1) or Flag-MT1 and Myc-MT2 (MT1, MT2) in photoreceptor cells under the control of the rhodopsin promoter. Data are representative of three experiments. (F and G) Visualization of MT1 and MT2 heterodimerization by in situ PLA on retinal sections of transgenic C57/bl6 mice expressing Flag-MT1 and Myc-MT2 (F) or control lacking the primary antibodies (G) in photoreceptor cells. MT1/MT2 heterodimers were visualized by staining cells with proximity probes directed against Flag and Myc, followed by ligation and rolling circle amplification. The hybridization probes were labeled with a fluorophore that is visualized with the Texas red filter. The nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole) (blue). Data are representative of three experiments. OS, photoreceptors’ outer segments; INL, inner nuclear layer; GCL, ganglion cell layer.

To test whether MT1 and MT2 form functional heteromers in the mouse photoreceptors, we developed a series of transgenic mice by transducing mouse fertilized eggs with lentiviral vectors containing Flag-MT1 or Myc-MT2 wild-type receptor-coding regions under the control of the rhodopsin promoter to drive expression of these receptors in photoreceptor cells. Tagging the receptors was necessary to detect them properly because reliable antibodies that recognize the wild-type rodent MT2 receptors are currently lacking. Immunofluorescence in these mice was only detected in the photoreceptors, which are present in the outer segment, using antibodies recognizing the Flag or Myc epitopes to detect the tagged MT1 or MT2 receptors (fig. S1). We crossed Flag-MT1 with Myc-MT2 mice to produce Flag-MT1/Myc-MT2 mice and performed coimmunoprecipitation experiments. Retinal cell lysates were prepared from wild-type and transgenic mice expressing either Flag-MT1 or coexpressing Flag-MT1 and Myc-MT2 (Fig. 4E). Flag-MT1 receptors were precipitated, and Myc-MT2 receptors were only coprecipitated in lysates prepared from double transgenic mice, consistent with the formation of MT1/MT2 heteromers in photoreceptors. With in situ proximity ligation assay (PLA) (39) on retinal sections obtained from Flag-MT1/Myc-MT2 mice, a fluorescence signal was observed indicating that MT1 and MT2 receptors exclusively interact in the outer segment, the part of the retina where the photoreceptors are located, as predicted (Fig. 4, F and G).

Pharmacological characterization identifies a specific pathway in retinal MT1/MT2 heteromers analyzed by scotopic ERGs

To further decipher the involvement of MT1 and MT2 in the effect of melatonin on scotopic ERGs, we performed intravitreal injections into the eyes of C3H-f+/+ mice. Intravitreal injection of melatonin induced a dose-dependent increase in the amplitude of the a-wave and b-wave of the scotopic ERGs as expected (Fig. 5, A and B). We then injected the IIK7 compound, which has been reported to behave as a selective agonist of the human MT2 receptor (40). Competition of 2-[125I]iodomelatonin binding at mouse MT1 and MT2 showed also high affinity and selectivity of IIK7 for murine MT2 receptors with Ki (inhibition constant) values of 73 ± 23 nM for MT1 and 0.052 ± 0.015 nM for MT2 (fig. S2A). When injected intravitreally at a concentration that selectively activates MT2 receptors (50 nM), IIK7 did not increase the amplitude of the a-wave or b-wave (Fig. 6A), whereas a higher concentration (5 μM) that activates MT1 and MT2 receptors significantly increased the amplitude of the a-wave and b-wave (Fig. 6B). Conversely, injection of luzindole or cis-4-phenyl-2-propionamidotetralin (4P-PDOT), two melatonin receptor antagonists with high preference for human MT1/MT2 heteromers compared to MT2 homomers (19), prevented the increase in the amplitude of the a-wave and b-wave that followed intraperitoneal administration of melatonin, which is consistent with the involvement of MT1/MT2 heteromers and validates the experimental protocol (Fig. 6, C and D). 4P-PDOT had a higher affinity for murine MT2 than MT1 (fig. S2, B to D). Injection of luzindole or 4P-PDOT alone did not produce any significant changes in the amplitude of the a-wave and b-wave excluding any agonistic and inverse agonistic effects of these compounds in the retina (fig. S3). Collectively, these data confirmed that the effect of melatonin on the amplitude of the a-wave and b-wave of the scotopic ERGs occurred with the expected pharmacology and depended on the simultaneous activation of both the MT1 and MT2 subunits of the MT1/MT2 heteromer.

Fig. 5 Dose response to intravitreal injection of melatonin.

(A) Quantification of the dark-adapted response of the a-wave (upper panel) and b-wave (lower panel) to flashes of light recorded after 1 hour of dark adaptation and intravitreal injection of melatonin (Mel) or vehicle in C3H-f+/+ mice. Intravitreal injection of melatonin (0.93 ng) at ZT6 induced a significant increase of the a-wave and b-wave amplitude (two-way ANOVA, P < 0.001). (B) Quantification of the dark-adapted response of the a-wave (upper panel) and b-wave (lower panel) to different doses of melatonin injected intravitreally. All data are presented as means ± SEM; n = 5 to 8 for each group. Luminance is expressed in cd*s/m2. ERGs in this experiment were performed at ZT6. *P < 0.05 with respect to vehicle injection (t tests).

Fig. 6 Effect of subtype-specific melatonin receptor agonists and antagonists on the ERG.

(A and B) Effect of intravitreal injection of two different concentrations of the MT2-specific agonist IIK7 on the amplitude of the a-wave and b-wave. At the lower concentration (50 nM), which is MT2-selective, there was no effect (A; two-way ANOVA, P > 0.1). At the higher concentration (5 μM), activating MT1 and MT2, IIK7 significantly increased the amplitude of the a-wave and b-wave (B; two-way ANOVA, P < 0.05). (C and D) Effect of intravitreal injections of luzindole (Luz) (C) or 4P-PDOT (D) on melatonin-induced changes in the amplitude of the a-wave and b-wave. Controls received an intravitreal injection of vehicle in the left eye. Melatonin (1 mg/kg) was injected intraperitoneally. Luzindole prevented the melatonin-induced changes in the amplitude of the a-wave and b-wave at the concentrations of 5 and 50 μM (two-way ANOVA, P < 0.05), whereas 4P-PDOT only at the concentration of 50 μM (two-way ANOVA, P < 0.05). All data are presented as means ± SEM; n = 6 for each time point. Luminance is expressed in cd*s/m2. ERGs in this experiment were performed at ZT6.

Expression of the nonfunctional MT2-P95L mutant in mouse photoreceptors blocks the effects of melatonin on the dark-adapted ERGs

In an attempt to interfere with the formation of functional heteromers, we generated transgenic mice expressing a nonfunctional murine MT2-P95L mutant that has been designed on the basis of the naturally occurring loss-of-function MT2-P95L mutant identified in diabetic patients (41). In vitro experiments performed in HEK293T cells confirmed that the murine mutant was expressed at similar amounts as the Myc-tagged wild-type MT2 and was devoid of any 2-[125I]iodomelatonin binding and signaling capacity (fig. S4, A to C), as described for the human mutant (41). However, the MT2-P95L mutant maintained its oligomerization capacity because this mutant readily formed homomers and heteromers with MT1 and MT2 wild-type receptors as demonstrated by coimmunoprecipitation (Fig. 7, A and B) and BRET donor saturation experiments (Fig. 7C). MT2-P95L had similar BRET50 values as the fluorescent protein–tagged wild-type MT2, indicating similar propensities to form oligomers [BRET50 = 2.61 ± 0.71 for the MT2-P95L homomers, 3.48 ± 0.73 for the MT2-P95L/MT1 heteromers, and 1.52 ± 0.40 for the MT2-P95L/MT2 heteromers (n = 3)]. Consistently, the MT2-P95L mutant interfered with MT1/MT2 heteromer formation as illustrated in BRET competition and coimmunoprecipitation experiments (Fig. 7, D and E, upper middle Western blot).

Fig. 7 The nonfunctional MT2-P95L mutant interferes with MT1/MT2 heteromer formation and blocks the effects of melatonin injection on the dark-adapted ERGs in the mouse photoreceptors.

(A and B) Coimmunoprecipitation of MT2-P95L-YFP with Flag-MT1 (A) or MT2-Rluc8 (B) in HEK293T cells. (C) Competition for MT1 by MT2-P95L detected by BRET. BRET donor saturation curves were performed by coexpressing equivalent amounts of MT1-Rluc8, MT2-Rluc8, or MT2-P95L-Rluc8 together with increasing quantities of MT2-P95L-YFP in HEK293T cells. (D) Reduction in BRET between MT1-Rluc8 and MT2-P95L-YFP by increasing the amount of coexpressed Myc-MT1. Expression of the Myc-MT2-P95L mutant was determined by Western blot. (E) Competition between MT2-P95L-YFP and the MT1/MT2 heteromer as monitored by coimmunoprecipitation. (F) Quantification of the dark-adapted response of the a-wave (upper graph) and b-wave (lower graph) to flashes of light recorded in MT2-P95L mutant mice after 1 hour of dark adaptation and intraperitoneal injection of melatonin (1 mg/kg) or vehicle at ZT6. Data are presented as means ± SEM; n = 5 to 8 for each group. Melatonin injection induced no significant changes in the amplitude of the a-wave and b-wave (two-way ANOVA, P > 0.1). Coimmunoprecipitation data are representative of three experiments. BRET data are obtained from three to five independent experiments. **P < 0.01; ***P < 0.001, t test. Anti-GFP, anti-Rluc, and anti-Flag: antibodies recognizing GFP, Rluc, or the Flag epitope, respectively. Luminance is expressed in cd*s/m2. ERGs in this experiment were performed at ZT6.

To examine the consequence of the presence of MT2-P95L on melatonin receptor function in the retina, we generated transgenic mice expressing the MT2-P95L mutant in photoreceptors and tested the effect of melatonin injection (1 mg/kg, intraperitoneally) on the dark-adapted ERGs. Control mice responded to melatonin injection with an increase in the amplitude of the a-wave and b-wave (Fig. 7F), whereas MT2-P95 transgenic mice showed no increase in the amplitude of the a-wave and b-wave when injected with melatonin (Fig. 7F). These results phenocopy those observed in MT1−/− or MT2−/− mice, suggesting that the effect of melatonin on dark-adapted ERGs depends on the presence of functional MT1/MT2 heteromers in photoreceptor cells.

Melatonin affects mouse photoreceptor function through the phospholipase C and protein kinase C signaling pathway

To investigate the intracellular signaling pathway leading to the effect of melatonin on the dark-adapted photoreceptors, we tested the effect of various inhibitors and activators of signaling molecules on the ability of melatonin to increase the amplitude of the a-wave. Because intraperitoneal administration of exogenous melatonin (1 mg/kg) reliably and significantly increased the amplitude of the a-wave 7 ms after the initial flash of light (Fig. 8A), we quantified the effect of intravitreal injection of the modulators of signaling pathways on response at 7 ms, which we considered the point at which the melatonin response was maximal (Fig. 8, B and C). Whereas the pan-inhibitor of heterotrimeric G proteins BIM 46187 (7-[2-amino-1-oxo-3-thio-propyl]-8-cyclohexylmethyl-2-phenyl-5,6,7,8-tetrahydro-imidazo-[1,2a]-pyrazine dimer, hydrochloride) reduced the melatonin-induced response, the Gi-specific pertussis toxin (PTX) had no effect (Fig. 8B and fig. S5). The injection of 8-bromoadenosine 3′,5′-cyclic monophosphate (8-br-cAMP), which is a cell-permeable cAMP analog, had no effect on the melatonin response (Fig. 8B and fig. S5), which was consistent with our observation that intravitreal injection of the protein kinase A (PKA) inhibitor H89 did not mimic the photoreceptor response to melatonin (Fig. 8C and fig. S5). These results exclude the involvement of the cAMP to PKA pathway.

Fig. 8 Melatonin affects mouse photoreceptor function through the PLC/PKC signaling pathway.

(A) Effect of melatonin (Mel, 1 mg/kg, intraperitoneally) on the a-wave of the photoreceptor response. The difference at 7 ms is significant (two-way ANOVA, P < 0.01). (B) Effect of PTX (1 ng, Gi protein inhibitor), BIM 46187 (100 μM, pan-heterotrimeric G protein inhibitor), 2-APB (50 μM, IP3 receptor inhibitor), 8-br-cAMP (50 μM, cAMP analog), or Bisl (25 μM, PKC inhibitor) on the melatonin-induced increase in a-wave amplitude at 7 ms. *P < 0.05, t test. Melatonin was injected intraperitoneally; pharmacological agents were injected intravitreally. Data shown in (A) and (B) represent the means ± SEM (n = 6 to 8). (C) Effect of inhibition of PKA activity by H89 or activation of PKC with PMA on the melatonin-induced increase in a-wave amplitude at 7 ms. *P < 0.05, t test. Data represent the means ± SEM; n = 5 to 7 for each point. For the data in (A) to (C), luminance is expressed in cd*s/m2. ERGs were performed at ZT6. (D) Production of inositol phosphate after stimulation with increasing concentrations of melatonin in HEK293T cells expressing the indicated receptors. Data are presented relative to the amount in unstimulated cells expressing the same receptor. (E and F) PKC activity after stimulation with increasing concentrations of melatonin (E) or 1 μM melatonin (F) in HEK293T cells expressing the indicated receptors. Data are presented relative to the amount in unstimulated cells for each condition (**P < 0.01, t test). (G) Inhibitory effect of melatonin on forskolin (Fsk)–stimulated cAMP production in HEK293T cells expressing the indicated receptors. Data in (D) to (G) represent the means ± SEM of at least four experiments performed in triplicate.

In contrast, inhibiting inositol 1,4,5-trisphosphate (IP3) receptors with 2-aminoethoxydiphenyl borate (2-APB) or inhibiting protein kinase C (PKC) with bisindolylmaleimide II (Bisl) abolished the effect of exogenous melatonin on the photoreceptor response amplitude at 7 ms after a flash of light. Consistent with a role for PKC, activation of PKC with the phorbol ester, phorbol 12-myristate 13-acetate (PMA), also reduced the amplitude of the a-wave in dark-adapted retinas (Fig. 8C), which mimics the effect of exogenous melatonin on the photoreceptor response. Thus, these data indicated that the effect of melatonin on the photoreceptors involves activation of the phospholipase C (PLC)/PKC pathway.

Consistent with this finding, physiological melatonin concentrations stimulated the production of the IP3 precursor, inositol phosphate (IP), in HEK293T cells coexpressing MT1 and MT2 [median effective concentration (EC50) = 2.8 ± 1.1 nM] (Fig. 8D). Melatonin was less potent (EC50 = 51.0 ± 35.2 nM) and less efficacious in stimulating IP production in cells only expressing MT1 or was inactive in cells expressing MT2 alone (Fig. 8D), suggesting a positive allosteric effect of the MT2 subunit on the MT1 subunit in the heteromer. The allosteric behavior was further confirmed in cells coexpressing the wild-type MT1 receptor with the inactive MT2-P95L mutant, which completely abolished the effect of melatonin in the heteromer (Fig. 8D).

To explore the existence of a potential cross-antagonism (the effect of an antagonist binding to one subunit of the receptor on the effect of an agonist binding to another subunit) in MT1/MT2 heteromers, we compared the effect of the MT2-selective 4P-PDOT antagonist (fig. S2C) on melatonin-induced IP production in cells expressing MT1 in the absence or presence of MT2 (fig. S6). Irrespective of the coexpression of MT2, competition curves were monophasic with similar IC50 values 5.8 ± 7.7 nM and 4.0 ± 3.5 nM (n = 3) for cells expressing MT1 or MT1 and MT2, respectively, suggesting the absence of cross-antagonism.

Melatonin dose-dependently increased PKC activity in cells expressing both MT1 and MT2 and was less potent and efficacious in stimulating this response in cells expressing MT1 alone (Fig. 8E). The effect on MT1/MT2 heteromers was inhibited in cells coexpressing the dominant-negative MT2-P95L mutant (Fig. 8F), which is consistent with our in vivo observation in transgenic mice overexpressing the MT2-P95L mutant (Fig. 7F). A possible switch between Gi protein toward Gq coupling from homomers to heteromers could be excluded because the inhibition of cAMP production was enhanced in cells coexpressing MT1 and MT2 receptors compared with those expressing only MT1 or MT2 (Fig. 8G).

Together, we found that MT1/MT2 heteromers could signal through the Gi to cAMP and from the Gq to PLC/PKC pathway. The effect of melatonin on the scotopic ERG involved the activation of the PLC/PKC pathway through MT1/MT2 heteromers, in which the MT2 subunit likely exerts a positive allosteric effect on the MT1 subunit.

Discussion

Establishing whether GPCRs form physiologically relevant functional homomers and heteromers in vivo has been a major biochemical challenge. Apart from obligate class C GPCR dimers (13, 41), only a few studies support the formation of physiologically relevant heteromers (1517, 4244). Here, we provide biochemical and functional in vivo evidence for the formation of MT1/MT2 heteromers in photoreceptors of the mouse retina. The existence of two melatonin receptor subtypes with apparently redundant functional properties (similar affinity for melatonin and signaling properties) and overlapping expression patterns in several tissues led us to hypothesize that the presence of both MT1 and MT2 might result in the formation of MT1/MT2 heteromers with unique functional properties, which we evaluated using mouse retina. In the retina, MT2 had a similar distribution to MT1; transcripts for both were expressed in photoreceptors. We found that melatonin affected the early photoreceptor-specific component of ERGs and that this effect was abolished by the photoreceptor-specific expression of a nonfunctional MT2 mutant or in mice lacking MT2.

Previous in vitro studies of human MT1 and MT2 indicated the formation of MT1/MT2 heteromers (18, 19). We confirmed these findings with murine MT1 and MT2 in transfected cells and provide experimental data suggesting that functional heterodimerization also occurs in mouse photoreceptors in vivo. We generated mice expressing tagged MT1 and MT2 receptors in photoreceptors and crossed these mice to produce double transgenic mice in which both tagged receptors were expressed in photoreceptors (Flag-MT1/Myc-MT2). Coimmunoprecipitation and in situ PLA experiments performed with retinas obtained from these mice indicated that these receptors indeed form a complex in the photoreceptors. Within the retina, melatonin receptor heteromer formation appears restricted to photoreceptors because ganglion cell viability is not affected in animals deficient for MT2 (35), and the circadian rhythms in retinal dopamine are not affected in animals deficient for MT1 (38).

To examine the functional consequences of MT1/MT2 heteromerization in vivo, we examined the responses of photoreceptors lacking one of the melatonin receptor subtypes or expressing a melatonin binding–deficient mutant of MT2 (MT2-P95L) or used selective pharmacological tools injected intravitreally. Photoreceptors lacking MT2 exhibited ERGs similar to those previously reported in the mice lacking MT1 (32), namely, a complete loss of the effect of melatonin on the scotopic ERGs. Furthermore, the effect of melatonin on dark-adapted ERGs in transgenic mice expressing the mutant in photoreceptors was abolished. Thus, the effect of melatonin depends on the presence of both receptor subtypes, most likely by targeting MT1/MT2 heteromers in photoreceptors. Consistent with these results, we found that administration of the 4P-PDOT or luzindole, two antagonists with preference for MT1/MT2 heteromers (19), inhibited the effect of melatonin on dark-adapted ERGs. In addition, IIK7, an MT2-selective agonist at low concentrations, only reproduced the effect of melatonin at concentrations high enough to target both MT1 and MT2, but not at concentrations specific for MT2, consistent with the hypothesis that both subunits in the heteromer have to bind melatonin and be activated to produce the photoreceptor response.

To investigate the pathway involved in melatonin-induced ERG regulation, we used pharmacological reagents to interfere with or promote specific downstream signaling pathways. MT1 and MT2 receptors typically inhibit cAMP formation by activating PTX-sensitive Gi proteins that are negatively coupled to adenylyl cyclase (20, 21). Surprisingly, manipulating the cAMP signaling pathway, by either inhibiting Gi proteins or PKA activity or administrating a cell-permeable cAMP analog, did not block or mimic the melatonin-induced effects in photoreceptors, respectively. Instead, the pharmacological data indicated that melatonin’s effect on photoreceptors was mediated by a PKC and IP3 pathway. Indeed, IP concentrations were increased in cells coexpressing MT1 and MT2 when exposed to physiological melatonin concentrations, whereas cells expressing only MT2 failed to increase IP concentration and cells expressing only MT1 only stimulated an increase in IP concentration at supraphysiological concentrations of melatonin.

Activation of the PLC/PKC pathway is functionally important in rod outer segments (45, 46). This pathway may play an important role in the light-dependent translocation of arrestin, a major regulator of rhodopsin function in rod photoreceptors (47). Whether melatonin regulates arrestin translocation remains a question for future studies.

Activation of the PLC/PKC pathway by melatonin is not unique to photoreceptors. Previous reports showed that MT1, when expressed alone in transfected cells, activates this pathway (48) and that in the suprachiasmatic nucleus (SCN) of the hypothalamus, PKC is also activated (49). Both MT1 and MT2 are present in the SCN (50), but whether PKC activation or any other effect of melatonin in the SCN depends on MT1/MT2 heteromers is unknown. Phenotypic characterization of MT1−/− and MT2−/− mice suggests that the action of melatonin in the SCN is not mediated by MT1/MT2 heteromers, at least in the outputs so far investigated (51). Nevertheless, no study has investigated in detail the presence of these receptors in the SCN and whether these receptors are present in the same neuron and regulated by the circadian cycle and where they may form heterodimers and control a yet to be identified SCN output.

The increased potency and efficacy of melatonin on IP production and PKC activity in cells expressing MT1/MT2 heteromers compared with those expressing the corresponding homomers indicated an allosteric behavior of the heteromer. No activation of this pathway was observed when the MT2 wild-type receptor in the heteromer was replaced by the inactive P95L mutant, further confirming the need for a functional MT2 subunit to allosterically activate the MT1 subunit in the heteromer. In contrast, specific activation of only MT2 by IIK7 was not sufficient to activate the PLC/PKC pathway, which required activation of both MT1 and MT2 subunits. The absence of cross-antagonism or negative allosterism in the MT1/MT2 heteromer is in agreement with previous observations of human MT1/MT2 heteromers (19). Collectively, these data suggest a working model in which both subunits bind melatonin and the MT2 subunit allosterically potentiates or facilitates the activation of the PLC/PKC pathway by the MT1 subunit.

In conclusion, we provide evidence that the modulatory effect of melatonin on mouse photoreceptor light sensitivity is mediated by MT1/MT2 heteromers and involves the activation of the PLC/PKC pathway. The involvement of MT1/MT2 heteromers may have important therapeutic implications because the heteromer complex may provide a unique pharmacological target to improve photoreceptor functioning and to extend the viability of photoreceptors during aging.

Materials and Methods

Animals

C3H MT2−/− knockout mice homozygous for the rd1 mutation, donated by S. M. Reppert and D. R. Weaver (University of Massachusetts Medical School), were backcrossed with C3H-f+/+ mice in which the rd1 mutation had been removed to produce C3H-f+/+ MT2−/−. The genotypes were determined according to the protocols previously described (32, 51). All the experimental procedures were performed in accordance with the Association for Assessment of Laboratory Animal Care policies and approved by the Morehouse School of Medicine Animal Care and Use Committee.

In situ hybridization

Mice were sacrificed by CO2 asphyxiation, and the eyeballs were immediately removed, punctured, and then fixed with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) (pH 7.0) for 6 hours at 4°C. The eyes were transferred to a 30% sucrose solution for 12 to 14 hours, embedded in Tissue-Tek OCT compound (Miles), and cut into 20-μm-thick cryosections. The template for transcription was a cDNA (complementary DNA) fragment of mouse MT2 subcloned into a pZLI vector (GenBank NM145712). The correct orientation of the construct was verified by sequence analysis and restriction enzyme digestion. Antisense and sense cRNA (complementary RNA) probes were generated with fluorescein-12-UTP (PerkinElmer Life Sciences) by in vitro transcription (MT2; forward, 5′-acactcacatagggcgattg-3′; reverse, 5′-agtgtgctggaattcggttc-3′; 512 base pairs). The templates for transcribing RNA probes were made by linearizing recombinant plasmids. Details about the MT1 probes are reported by Baba et al. (32). Sections were immersed in prehybridization buffer containing 50% formamide, 5× Denhardt’s solution, and 5× SSC [1× SSC = 150 mM NaCl, 15 mM sodium citrate (pH 7.0)] for 2 hours at room temperature. The sections were then hybridized with 75 μl of hybridization buffer (1×), covered with a cover slip, and incubated overnight in a humidified chamber at 67°C. The best labeling was obtained at a probe concentration of 1:100. Slides were then washed in 5× SSC/50% formamide at 68°C for 1 hour and in 2× SSC for 1 hour at 68°C, and then incubated in ribonuclease A (20 mg/ml) at 37°C for 30 min followed by 2× SSC for 1 hour and 0.2× SSC for 30 min (twice) at room temperature. Slides were mounted and then viewed with a Zeiss Axioskop microscope equipped with epifluorescence.

In situ PLA

Eyeballs obtained from Flag-MT1/Myc-MT2 were fixed in 4% PFA overnight, transferred to a 20% sucrose solution for 12 to 14 hours, embedded in Tissue-Tek OCT compound (Miles), and then sectioned (10 μm). PLA was conducted with Duolink In Situ Fluorescence kit (Sigma) (39). After blocking (Duolink blocking solution), the sections were incubated with the mouse c-Myc antibody (1:250; Santa Cruz Biotechnology Inc.) and the Flag antibody (1:500; Sigma-Aldrich) overnight. Anti-mouse MINUS PLA probe (1:5; Sigma), anti-rabbit PLUS PLA probe (1:5; Sigma), and Duolink In Situ Detection Reagents Red kit were used to detect protein interactions. Sections were washed with buffer A [NaCl (8.8 g/liter), tris base (1.2 g/liter), Tween 20 (0.5 ml/liter) (pH 7.4)] after the first and second incubation, and the wash buffer B [NaCl (5.84 g/liter), tris base (4.24 g/liter), tris-HCl (26 g/liter) (pH 7.5)] was used after the amplification process. After drying at room temperature, the slides were mounted with cover slip with Duolink In Situ Mounting Medium with DAPI (Sigma). Slides were mounted and then viewed with a Zeiss Axioskop microscope equipped with epifluorescence. Primary antibodies were omitted in control sections.

Scotopic ERG

Mice were anesthetized with ketamine (80 mg/kg) and xylazine (16 mg/kg). The pupils were dilated with 1% atropine and 2.5% phenylephrine (Sigma), and mice were placed on a heating pad set at 37°C with feedback from the rectal temperature probe. The eye was lubricated with saline solution, and a contact lens–type electrode (LKC Technologies model: N1530NNC) was topically applied on the cornea. A needle reference was inserted in other side of cheek, and the ground needle was inserted into the base of tail. All preparation of ERG recordings was conducted under red dim light (<3 lux, 15-W Kodak safe lamp filter 1A, Eastman Kodak).

All electrodes were connected to a Universal DC Amplifier (LKC Technologies model UBA-4200), and bands were filtered from 0.3 to 500 Hz. Data were recorded and analyzed by EM for Windows (ver. 8.2.1, LKC Technologies). Core body temperature was maintained in 37°C by a feedback temperature control system (FHC Inc.) throughout the entire ERG recording. In the dark-adapted ERG protocol, seven series of flash intensities between 0.03 and 6.28 cd*s/m2 were presented to the mouse eye. Flashes were generated by 530-nm green LEDs in a Ganzfeld illuminator (LKC Technologies), and the interval of the flashes increased from 0.612 to 30 s as the intensity of the flashes increased. Responses of 3 to 10 flashes were averaged to generate a waveform for each step of light intensity, and the a-wave and b-wave of ERG measurement were analyzed from the trace of waveform. To directly determine the photoreceptors’ response to melatonin, we used an early time point in the a-wave (7 ms after the initial flash of light) [see (36) for details], because at this early time, the amplitude of the a-wave is not affected by the activity of the bipolar cells (37).

Administration of exogenous melatonin, melatonin agonists, antagonists, and inhibitors

Melatonin (Sigma) was dissolved in ethanol (8 mg/ml) and then diluted with sterilized PBS (0.1 mg/ml, 1.25% ethanol). Melatonin was administered by intraperitoneal (1 mg/kg) or intravitreal injection in different dosages according to the experimental designs. We used melatonin receptor agonists or antagonists, and various activators or inhibitors in the experiments as follows: IIK7 [Sigma, dimethyl sulfoxide (DMSO) stock solution (30 mg/ml), 0.005% final DMSO concentration], luzindole [TOCRIS, DMSO stock solution (29 mg/ml), 0.05% final DMSO concentration], 4P-PDOT [Sigma, DMSO stock solution (28 mg/ml), 0.05% final DMSO concentration], 2-APB [Sigma, DMSO stock solution (20 mg/ml), 0.006% final DMSO concentration], BIM 46187 (52) [DMSO stock solution (6.7 mg/ml), 0.23% final DMSO concentration], bisindolylmaleimide I [Sigma, DMSO stock solution (10 mg/ml), 0.1% final DMSO concentration], and PMA [TOCRIS, DMSO stock solution (61 mg/ml), 0.0003% final DMSO concentration] were dissolved in DMSO and diluted with sterilized PBS. Identical DMSO concentrations and sterilized PBS were used for vehicle control for each particular drug treatment. PTX from Bordetella pertussis [PTX, Sigma, stock solution (0.1 mg/ml)], 8-br-cAMP [Sigma, stock solution (25 mg/ml)], and H89 dihydrochloride hydrate (H89, Sigma, 10 mg/ml) were dissolved in sterilized double-distilled water and sterilized PBS. After mice were anesthetized with the isoflurane vaporizer machine (VetEquip), the drugs and vehicle control were intravitreally injected for a volume of 1 μl with a 10-μl Hamilton syringe (Hamilton Company) with 30 gauge needle attached (Becton, Dickinson & Company). A fresh needle was used for each experiment. The dosages of the drugs were calculated according to an estimated vitreal volume of 20 μl in mouse eye (53). All drugs or vehicles, except PTX, were injected 1 hour before the ERG recording, and all animals were placed in a dark isolated chamber just after the drug was administered. PTX was administered 4 hours before ERG measurement. For some experiments, melatonin was administered by intraperitoneal injection after intravitreal administration of the drug at the dosage of 1 mg/kg. Unless differently specified in the text, all the ERGs were performed at ZT6.

Plasmid constructions

The 3.8-kb fragment of the mouse rhodopsin promoter was generated by chemical synthesis (DNA2.0 Inc.) and flanked with a Mlu I (5′) and a Bam HI site (3′) for insertion into the pJ cloning plasmid from DNA2.0 Inc. The coding regions of mouse MT1 (NM_008639.2) or mouse MT2 (NM_145712.2) that were preceded by a Flag or Myc tag sequence, respectively, were generated by chemical synthesis (DNA2.0 Inc.) and flanked with attL1 and attL2 sites and cloned into the pJ cloning plasmid. Flag-MT1 and Myc-MT2 fragments were cloned behind the rhodopsin promoter into the lentiviral pTrip IZI vector to generate the pTrip-rho-Flag-MT1 and pTrip-rho-Myc-MT2 vectors. Flag-MT1 and Myc-MT2 fragments were inserted into the pcDNA3 expression vector behind the CMV (cytomegalovirus) promoter to generate pcDNA3-CMV-Flag-MT1 and pcDNA3-Myc-MT2 vectors. The Myc-MT2-P95L mutant was generated by site-directed mutagenesis from the corresponding MT2 wild-type vector. BRET fusion proteins were generated by inserting the coding regions without the STOP codon of Flag-, HA (hemagglutinin)–, or Myc-tagged mouse MT1, MT2, or MT2-P95L receptors, respectively, in frame with the coding region of Renilla luciferase variant (Rluc8) or YFP: Flag-mMT1-Rluc8, Flag-mMT1-YFP, HA-mMT2-Rluc8, HA-mMT2-YFP, Myc-MT2-P95L-Rluc8, and Myc-MT2-P95L-YFP. All constructs were verified by sequencing.

Coimmunoprecipitation

Transfected cells were solubilized in RIPA (radioimmunoprecipitation assay) buffer [20 mM Hepes (pH 7.4), 120 mM NaCl, 5 mM EDTA, 10% glycerol], supplemented with protease inhibitors [leupeptine (1 μg/ml), pepstatin (1 μg/ml), benzamidine (2 μg/ml), AEBSF (1 μg/ml)] and 1% Triton X-100 for at least 3 hours at 4°C. The soluble fraction was recovered by 1 hour of centrifugation at 13,000g and subjected to immunoprecipitation with antibodies (2 μg/ml) recognizing the Flag (Sigma F7425) or the GFP epitopes (Abcam ab290). Denatured complexes were heated, separated by SDS–polyacrylamide gel electrophoresis, transferred to nitrocellulose, and immunoblotted with antibodies against either the Flag epitope (Sigma F3165, 1:1000 dilution), Rluc (Millipore MAB4400, 1:500), or GFP (Roche 11814460, 1:500). Immunoreactivity was revealed using secondary antibodies coupled to 680- or 800-nm fluorophores with the Odyssey LI-COR infrared fluorescent scanner (ScienceTec). For coimmunoprecipitation experiments from retinal samples, three retinas were solubilized overnight in 1 ml of TEM [75 mM tris (pH 7.5), 2 mM EDTA, 12 mM MgCl2] supplemented with protease inhibitors and 1% Triton X-100 at 4°C. Immunoprecipitation was performed with mouse antibodies against the Flag epitope (F3165, Sigma) (2 μg) and immunoblots with rabbit antibodies against the Myc epitope (sc-789, A14, 1:500).

BRET assay

For BRET donor saturation curves, HEK293T cells seeded in six-well plates were transiently transfected with 0.5 ng of MT1-Rluc, 2 ng of MT2-Rluc, or 1 ng of MT2-P95L-Rluc and 10 to 3000 ng of YFP plasmids. Twenty-four hours after transfection, cells were transferred into a 96-well white OptiPlate (PerkinElmer Life Sciences) precoated with poly-l-lysine (10 mg/ml; Sigma-Aldrich) and incubated for another 24 hours before BRET measurements. BRET measurements were performed as described previously (54) with the Mithras lumino/fluorometer (Berthold Technologies). Results are expressed in milliBRET units (mBU), with 1 mBU corresponding to the BRET ratio values multiplied by 1000.

Radioligand-binding assay

Competition binding experiments were performed as previously described (55) with 200 pM 2-[125I]iodomelatonin (PerkinElmer Life Sciences) and a range of different concentrations of the indicated melatonin receptor ligands. Ki values were determined according to the Cheng-Prussof formula: Ki = IC50/1 + L + Kd.

cAMP assay

Concentrations of cAMP were determined in cell suspensions stimulated with 2 μM forskolin for 30 min at room temperature in the absence or presence of different concentrations of melatonin (0.1 fM to 1 μM) by HTRF (homogeneous time-resolved fluorescence) using the Cisbio cAMP femto Tb kit according to the manufacturer’s instructions.

IP assay

IP concentrations were determined in cell suspensions stimulated for 90 min at 37°C in the absence or presence of different concentrations of melatonin (1 fM to 10 μM) by HTRF with the Cisbio IP-One Tb kit according to the manufacturer’s instructions.

PKC activity assay

PKC activity was measured by a method adapted from Kent et al. (56). Briefly, cells seeded on polylysine-coated 24-well plates were serum-deprived overnight and stimulated with 100 nM melatonin for 15 min. Cells were then put on ice, washed with ice-cold PBS containing 1 mM Na3VO4, and incubated for 20 min in 25 μl of a buffer (137 mM NaCl, 20 mM Hepes, 10 mM MgCl2, 25 mM β-glycerophosphate, 1 mM dithiothreitol, 1 mM Na3VO4, and protease inhibitors) containing 0.05% digitonin. Activity was then measured with the Upstate PKC Assay Kit (Millipore) according to the manufacturer’s instructions.

Lentivirus production

The viral particles were produced by transient transfection of HEK293T cells with the previously described p8.9 and pMD-G plasmids and either pTrip-rho-Flag-MT1, pTrip-rho-Myc-MT2, or pTrip-rho-Myc-MT2-P95L vectors. Supernatants were collected 48 hours after transfection, and high titer stocks were prepared as described (57). The stocks were titrated and normalized for the p24 antigen assayed by enzyme-linked immunosorbent assay.

Generation of transgenic mice

The lentiviral pTrip-rho-Flag-MT1, pTrip-rho-Myc-MT2, or pTrip-rho-Myc-MT2-P95L vectors expressing MT1, MT2, or MT2-P95L under the control of the rhodopsin promoter were injected into the perivitelline space of mouse fertilized (C57/bl6/N) eggs as described previously (58). The injected eggs were next reimplanted into the oviduct of pseudopregnant females 0.5 day after coitus. Flag-MT1/Myc-MT2 double transgenic mice were generated by crossing Flag-MT1 and Myc-MT2 transgenic mice.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/6/296/ra89/DC1

Fig. S1. Flag and Myc immunoreactivity in the photoreceptors Flag-MT1 and Myc-MT2 mice.

Fig. S2. Competition binding of 2-[125I]iodomelatonin on membranes of HEK293 cells expressing mouse MT1 and MT2.

Fig. S3. The effect of luzindole and 4P-PDOT on the ERG.

Fig. S4. Functional characterization of the MT2-P95L mutant in transfected HEK293T cells.

Fig. S5. The effect of different signaling pathway stimulators and inhibitors on the amplitude of the a-wave at 7 ms.

Fig. S6. Inhibition of melatonin-induced IP production by 4P-PDOT.

References and Notes

Acknowledgments: We thank H. Pilet, C. Sarkis, and M.-J. Lecomte (Newvectys Inc., France) for help and advice in the design, construction, and production of lentiviral expression vectors. We thank J. Dam and M. Scott (Institut Cochin, France) for their expert comments on the manuscript, Morehouse School of Medicine Center for Laboratory Animal Resources for housekeeping the transgenic mouse models, and S. M. Reppert and D. R. Weaver (University of Massachusetts Medical School) for providing the C3H MT2−/− knockout mice homozygous for the rd1 mutation. Funding: This work was supported by NIH grants NS43459, EY028821, EY022216, 5U54NS060659, S21MD000101, G12-RR03034, and U54RR026137 and by INSERM, CNRS, and the “Who am I?” laboratory of excellence no. ANR-11-LABX-0071 funded by the French government through its “Investments for the Future” program operated by the French National Research Agency (ANR) under grant no. ANR-11-IDEX-0005-01. A.B.-C. was supported by a doctoral fellowship from the CODDIM 2009 (Région Ile-de-France). Author contributions: K.B. participated in the project design, performed experiments with electroretinography and PLA, interpreted the data, and contributed to the writing of the paper; A.B.-C. designed and performed Western blots and coimmunoprecipitation experiments, participated in IP detection, and interpreted the data; A.-S.J. designed and performed all BRET experiments; M.K. generated the MT2-P95L mutant, performed coimmunoprecipitation from the retina, and constructed lentiviral vectors; J.-L.G. performed IP and cAMP detection and PKC activity; S.D. generated transgenic mice; F.G. constructed BRET fusion protein vectors and supervised BRET experiments; K.Y. performed initial BRET and coimmunoprecipitation experiments; C.L. performed in situ hybridization; S.C.-A. bred and genotyped the various transgenic mice; R.J. was responsible for the project supervision, experiment design, data interpretation, and manuscript writing and provided funding support; and G.T. was responsible for the project supervision, experiment design, data interpretation, and manuscript writing and provided funding support. Competing interests: The authors declare that they have no competing financial interests.
View Abstract

Navigate This Article