International Union of Basic and ClinicalPharmacology. LXXV. Nomenclature, Classification,
and Pharmacology of G Protein-CoupledMelatonin Receptors
Margarita L. Dubocovich, Philippe Delagrange, Diana N. Krause, David Sugden, Daniel P. Cardinali, and James Olcese
Department of Pharmacology and Toxicology, School of Medicine and Biomedical Sciences, University at Buffalo, State University of NewYork, Buffalo, New York (M.L.D.); Department Molecular Pharmacology and Biological Chemistry, Feinberg School of Medicine,
Northwestern University, Chicago, Illinois (M.L.D.); Experimental Sciences Department, Institut de Recherches Servier, Suresnes, France(P.D.); Department of Pharmacology, College of Medicine, University of California, Irvine, California (D.N.K.); Division of Reproduction
and Endocrinology, School of Biomedical and Health Sciences, King’s College London, United Kingdom (D.S.); Department of Teaching &Research, Faculty of Medical Sciences, Pontificia Universidad Católica Argentina, Buenos Aires, Argentina (D.P.C.); and Department of
Biomedical Sciences, Florida State University College of Medicine, Tallahassee, Florida (J.O.)
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344I. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
A. Melatonin physiology and function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344B. Melatonin receptor discovery: historical background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347C. International Union of Basic and Clinical Pharmacology criteria for receptor
nomenclature and drug classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348D. Current melatonin receptor nomenclature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
II. G protein-coupled melatonin receptor family. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349A. Protein structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349B. Gene structure and chromosomal localization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350C. Melatonin receptor polymorphisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351D. Molecular structure of MT1 and MT2 melatonin receptor ligand binding pockets . . . . . . . . . . . 352
1. MT1 melatonin receptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3522. MT2 melatonin receptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
III. Cellular signaling of MT1 and MT2 melatonin receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356A. MT1 melatonin receptor signaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357B. MT2 melatonin receptor signaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358C. Melatonin receptor regulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
IV. MT1 and MT2 melatonin receptors: structure-activity relationships and selective ligands. . . . . . . 359A. Ligand selectivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359B. Structure-activity relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359C. Selective MT1 and MT2 melatonin ligands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363D. Ligand efficacy in native tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
1. Agonists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3642. Antagonists/partial agonists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3643. Inverse agonists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3654. Dimers/heterodimers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
V. MT1- and MT2-mediated functional responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365A. Melatonin receptor expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366B. Melatonin receptor function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368
1. Central nervous system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3682. Hypothalamic-pituitary-gonadal axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3703. Cardiovascular system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
Address correspondence to: Dr. Margarita L. Dubocovich, Department of Pharmacology and Toxicology, 102 Farber Hall, School ofMedicine and Biomedical Sciences University at Buffalo State University of New York (SUNY), 3435 Main Street, Buffalo, NY 14214. E-mail:[email protected]
This article is available online at http://pharmrev.aspetjournals.org.doi:10.1124/pr.110.002832.
0031-6997/10/6203-343–380$20.00PHARMACOLOGICAL REVIEWS Vol. 62, No. 3Copyright © 2010 by The American Society for Pharmacology and Experimental Therapeutics 2832/3594882Pharmacol Rev 62:343–380, 2010 Printed in U.S.A.
343
by guest on June 28, 2021D
ownloaded from
4. Immune system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3715. Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3716. Cancer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
VI. Melatonin receptors as therapeutic targets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372A. Agomelatine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372B. Ramelteon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373C. PD 6735 (LY 156735) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373D. Circadin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
VII. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
Abstract——The hormone melatonin (5-methoxy-N-acetyltryptamine) is synthesized primarily in the pi-neal gland and retina, and in several peripheral tissuesand organs. In the circulation, the concentration of mel-atonin follows a circadian rhythm, with high levels atnight providing timing cues to target tissues en-dowed with melatonin receptors. Melatonin recep-tors receive and translate melatonin’s message toinfluence daily and seasonal rhythms of physiologyand behavior. The melatonin message is translatedthrough activation of two G protein-coupled recep-tors, MT1 and MT2, that are potential therapeutic
targets in disorders ranging from insomnia and cir-cadian sleep disorders to depression, cardiovasculardiseases, and cancer. This review summarizes thesteps taken since melatonin’s discovery by AaronLerner in 1958 to functionally characterize, clone,and localize receptors in mammalian tissues. Thepharmacological and molecular properties of the re-ceptors are described as well as current efforts todiscover and develop ligands for treatment of a num-ber of illnesses, including sleep disorders, depres-sion, and cancer.
I. Introduction
Melatonin (5-methoxy-N-acetyltryptamine) is an im-portant hormonal output of the circadian system, and itcirculates nightly to provide timing cues to any tissuethat can read the message. Melatonin receptors, ofcourse, are the entities that receive and translate thismessage to influence daily and seasonal rhythms ofphysiology and behavior. Our evolving understanding ofmelatonin receptors is providing new insights intowhere and how this hormone exerts its physiologicaleffects in the body as well as how these receptors may beuseful therapeutic targets in disorders ranging from in-somnia and jet lag to depression, cancer, and cardiovas-cular disease.
The field of mammalian melatonin receptors got off toa slow start; the first pharmacological characterizationof a functional mammalian melatonin receptor (Duboco-vich, 1983) and the cloning of the first human melatoninreceptor (Reppert et al., 1994) came 25 and 36 years,respectively, after melatonin itself was discovered(Lerner et al., 1959). Now, however, the field is movingquickly, and there are sufficient data upon which toorganize and classify the known mammalian melatoninreceptors. That topic is the focus of this review.
A. Melatonin Physiology and Function
In mammals, melatonin is secreted primarily by thepineal gland during the dark period of the daily light/dark cycle (for reviews, see Cardinali, 1981; Reiter,1991; Borjigin et al., 1999; Klein, 1999; Olcese, 1999).
The circadian rhythm of pineal melatonin synthesis andrelease is driven by circadian pacemaker cells (the “mas-ter clock”) located in the suprachiasmatic nucleus(SCN1) of the hypothalamus that project to the pinealvia a multisynaptic pathway (Fig. 1). Melatonin is alsosynthesized in the retina and a clock mechanism withinthe retina itself seems to drive the melatonin rhythm inthis tissue (Tosini and Menaker, 1998; Tosini et al.,2007). In both cases, the clock rhythm is entrained to a24-h period by environmental light (the photoperiod)that is sensed by a subset of retinal ganglion cells con-taining the photopigment melanopsin, which conveysphotic stimuli to the SCN via the retinohypothalamictract (Berson et al., 2002).
The melatonin rhythm is a consequence of the regu-lation of the hormone’s synthetic enzymes, which arehighly active at night (see Fig. 2). Melatonin is synthe-sized from serotonin through two enzymatic steps. First,
1 Abbreviations: 4P-ADOT, 4-phenyl-2-acetamidotetraline; 4P-PDOT, 4-phenyl-2-propionamidotetraline; 5-HEAT, 5-hydroxye-thoxy-N-acetyltryptamine; AA-NAT, arylalkylamine N-acetyl-transferase; AFMK, N1-acetyl-N2-formyl-5-methoxykynuramine;AMK, N1-acetyl-5-methoxykynuramine; AMMTC, N-acetyl-4-aminomethyl-6-methoxy-9-methyl-1,2,3,4-tetrahydrocarbazole;BKCa, calcium-activated potassium channel; CHO, Chinese ham-ster ovary; CREB, cAMP responsive element binding protein; CT,circadian time; GLP-1, glucagon-like peptide; GPCR, G protein-coupled receptor; H-89, 5-isoquinolinesulfonamide; HIOMT, hy-droxyindole-O-methyltransferase; KO, knockout; PKC, proteinkinase C; PTX, pertussis toxin; RT-PCR, reverse transcription-polymerase chain reaction; SCN, suprachiasmatic nucleus; TM,transmembrane; WT, wild type.
344 DUBOCOVICH ET AL.
serotonin N-acetyltransferase [arylalkylamine N-acetyl-transferase (AA-NAT)] acetylates serotonin to yield N-acetylserotonin. The second step involves transfer of amethyl group from (S)-adenosylmethionine to the 5-hy-droxy group of N-acetylserotonin via the enzyme hy-droxyindole-O-methyltransferase (HIOMT). The regula-tion of these enzymes is fascinating and much studied(Wurtman and Axelrod, 1968; Cardinali, 1981; Klein,1999, 2007; Olcese, 1999). Most recently, molecular ap-proaches are revealing the transcriptional and post-transcriptional mechanisms responsible for daily fluctu-ations in AA-NAT (Borjigin et al., 1995, 1999; Roseboomet al., 1996; Gastel et al., 1998; Klein, 1999; Fukuhara etal., 2001) and regulation of HIOMT (Gauer and Craft,1996). Use of sensitive RT-PCR techniques suggests thatlow levels of AA-NAT and HIOMT, and therefore localmelatonin synthesis, may also occur in other tissues,such as gut, testis, spinal cord, raphe nucleus, and stri-atum (Stefulj et al., 2001). In mammals, most of thenocturnal rhythm in circulating melatonin is abolishedby pinealectomy (Simonneaux and Ribelayga, 2003).However, sporadic reports suggested that as much as20% of circulating melatonin is derived from tissuesother than the pineal gland (Ozaki and Lynch, 1976).
The lipophilicity of the melatonin molecule allows itto diffuse passively across cell membranes as well ascell layers, and thus it can diffuse from the pinealo-cytes as soon as it is synthesized. Early studies by
Cardinali et al. (1972) showed melatonin binding tohuman plasma albumin. This observation was laterconfirmed by Pardridge and Mietus (1980). These au-thors also reported that albumin-bound melatonincrosses the blood-brain barrier. Melatonin binding toplasma albumin was confirmed by Morin et al. (1997)and provided evidence of high-affinity melatonin bind-ing to �1-acid glycoprotein (Cardinali et al., 1972;Morin et al., 1997). An excellent correlation was dem-onstrated between saliva melatonin in humans andlevels of unbound serum melatonin, suggesting for thefirst time that melatonin binding to plasma proteinsmay affect the levels of free melatonin and hencephysiological responses (Kennaway and Voultsios,1998). Because the levels of albumin and �1-acid gly-coprotein may vary with age and disease, particularlyduring inflammatory processes, the level of free mel-atonin and drugs in human plasma cannot be pre-dicted and has to be considered on a case-by-case basis(Viani et al., 1992; Morin et al., 1997; Waldhauser etal., 1988).
Melatonin rapidly disappears from the blood, with ahalf-life of approximately 30 min, depending on the speciesexamined (Cardinali et al., 1972; Waldhauser et al., 1984).In humans, most melatonin in the general circulation isconverted to 6-hydroxymelatonin by the liver, which clears92 to 97% of circulating melatonin in a single pass (Tetsuoet al., 1980; Young et al., 1985). Then 6-hydroxymelatonin
FIG. 1. Regulation of melatonin production and receptor function. Melatonin is synthesized in the pineal gland and in the retina. In the pinealgland, melatonin (MLT) synthesis follows a rhythm driven by the suprachiasmatic nucleus, the master biological clock. Neural signals from the SCNfollow a multisynaptic pathway to the superior cervical ganglia. Norepinephrine released from postganglionic fibers activates �1- and �1-adrenoceptorsin the pinealocyte, leading to increases in second messengers (i.e., cAMP and inositol trisphosphate) and the activity of AA-NAT, the rate-limiting stepin melatonin synthesis. The system is dramatically inhibited by light, the external cue that allows entrainment to the environmental light/dark cycle.The photic signal received by the retina is transmitted to the SCN via the retinohypothalamic tract, which originates in a subset of retinal ganglioncells. Pineal melatonin thus serves as the internal signal that relays day length, allowing regulation of neuronal activity (MT1) and circadian rhythms(MT1, MT2) in the SCN (Dubocovich, 2007), of neurochemical function in brain through the MT1 and MT2 receptors (Dubocovich, 2006), of vasculartone through activation of MT1 (constriction) and MT2 receptors (dilation) in arterial beds (Masana et al., 2002), and seasonal changes in reproductivephysiology and behavior through activation of MT1 receptors in the pars tuberalis (Duncan, 2007). The pars tuberalis of the pituitary gland interpretsthis rhythmic melatonin signal and generates a precise cycle of expression of circadian genes through activation of MT1 receptors. Melatonin synthesisin the photoreceptors of the retina follows a similar circadian rhythm generated by local oscillators (Tosini et al., 2007). Activation of MT1 and MT2melatonin receptors regulate retina function and hence transmission of photic information to the brain (Dubocovich et al., 1997). [Adapted fromDubocovich ML and Masana M (2003) Melatonin receptor signaling, in Encyclopedia of Hormones and Related Cell Regulators (Henry H and NormanA eds), pp 638–644, Academic Press, San Diego, CA. Copyright © 2003 Academic Press. Used with permission.]
MELATONIN RECEPTORS 345
is conjugated and excreted into the urine, approximately50 to 80% as the sulfate derivative and 5 to 30% as theglucuronide (Ma et al., 2008). It is important to point outthat the most abundant metabolite in mouse urine is 6-hy-droxymelatonin excreted as the glucuronide conjugate (75–88%) (Kennaway et al., 2002; Ma et al., 2008). Melatoninmetabolism in the brain, however, may involve oxidativepyrrole-ring cleavage. No 6-hydroxymelatonin is detectedafter melatonin injection into the cisterna magna (Hirataet al., 1974). This pathway may be particularly importantbecause melatonin is also released via the pineal recessinto the cerebrospinal fluid as well as into the circulation(Tricoire et al., 2002). The primary cleavage product isN1-acetyl-N2-formyl-5-methoxykynuramine (AFMK),which is deformylated, either by arylamine formamidaseor hemoperoxidases, to N1-acetyl-5-methoxykynuramine(AMK). Surprisingly, numerous enzymatic (indoleamine2,3-dioxygenase, myeloperoxidase), pseudoenzymatic (oxo-ferryl hemoglobin, hemin), photocatalytic, or free-radicalreactions lead to the same product: AFMK (Hardeland2005). Some estimates indicate that pyrrole ring cleavagecontributes to approximately one third of the total catabo-lism of melatonin (Ferry et al., 2005), but the percentagemay be even higher in certain tissues. Other oxidative
catabolites are cyclic 3-hydroxymelatonin, which can alsobe metabolized to AFMK, and a 2-hydroxylated analogthat does not cyclize but turns into an indolinone (Hard-eland 2005). Additional hydroxylated or nitrosated metab-olites have been detected, but they seem to be present inminor quantities only. AFMK and AMK also form metab-olites by interactions with reactive oxygen and nitrogenspecies (Tan et al., 2007). Antioxidative protection, safe-guarding of mitochondrial electron flux, and, in particular,neuroprotection have been demonstrated in many experi-mental systems to be mediated by melatonin and its en-dogenous metabolites. This effect is not mediated by theknown G protein-coupled melatonin receptors and thuswill not be reviewed in this article.
The melatonin metabolites produced in the liver (e.g.,6-hydroxymelatonin) and in the brain (e.g., AFMK andAMK) are known to modulate a variety of functional re-sponses possible through activation of the G-protein-coupled melatonin receptor. 6-Hydroxymelatonin com-peted for 2-[125I]iodomelatonin binding to both the MT1and MT2 melatonin receptors (Dubocovich et al., 1997).This metabolite also decreases in a concentration depen-dent manner the calcium dependent release of [3H]dopam-ine from rabbit retina (Dubocovich, 1985). AFMK and
FIG. 2. Melatonin synthesis. Melatonin (MLT) is synthesized from serotonin through two enzymatic steps. First, serotonin is acetylated by NATto yield N-acetylserotonin (NAS). The second step involves transfer of a methyl group from (S)-adenosylmethionine to the 5-hydroxyl group ofN-acetylserotonin via the enzyme HIOMT. The rhythms of melatonin and serotonin have opposite phase during subjective night and day (Klein, 1999).
346 DUBOCOVICH ET AL.
AMK decrease sexual development in a prepubertal ratmodel (Kennaway et al., 1988). It is noteworthy thatAFMK accelerates re-entrainment of the 6-hydroxymela-tonin rhythm, when given at the new dark onset, after aphase advance of the light/dark cycle in male rats (Ken-naway et al., 1989).
The melatonin rhythm is an important efferent hor-monal signal driven by the endogenous clock, which cantherefore be used as an internal synchronizer (or “internalZeitgeber”) (Dawson and Armstrong, 1996). Melatonin canimpose circadian rhythmicity upon target structures, andit also is known to act directly on the SCN to modulate theclock itself. Exogenous melatonin, given at the same timeof day, can entrain physiological and behavioral rhythms(e.g., body temperature and rest-activity cycles) (Arendtand Skene, 2005). Melatonin has been reported to modu-late a myriad of other functions, including visual, neuroen-docrine, reproductive, neuroimmune, and vascular physi-ology (Arendt, 2000; Cagnacci et al., 2000; Castrillón et al.,2000; Monti and Cardinali, 2000; Maestroni, 2001; Shar-key et al., 2009). Because the duration of nocturnal mela-tonin secretion is directly proportional to the length of thenight, this hormone also provides a signal for seasonalchange. Melatonin is, indeed, the critical parameter forphotoperiod integration and the induction of particularphysiological responses, such as those observed in seasonalbreeders (Karsch et al., 1988; Pitrosky et al., 1991; Cardi-nali and Pévet, 1998; Malpaux et al., 1999).
In addition to the circadian rhythm of the endogenousligand, target tissues and physiological responses can alsodisplay daily variations in receptivity to melatonin. Forexample, optimal entrainment of the activity rhythm inrodents housed in constant darkness occurs when melato-nin is administered at the onset of activity; this induces aphase advance in the animal’s rhythm (Cardinali et al.,1997; Redman, 1997). It is possible that melatonin trans-mits photoperiodic information by regulating its own re-ceptors, perhaps by altering receptor density, transductionmechanisms, and/or trafficking (Tenn and Niles, 1993;Gerdin et al., 2004a,b). Further discussion in this area isbeyond this review, because it focuses primarily on theaction of exogenous melatonin and drugs with therapeuticpotential mediating actions through activation of MT1and/or MT2 melatonin receptors.
B. Melatonin Receptor Discovery:Historical Background
The study of melatonin receptors can be traced back to1917, when McCord and Allen demonstrated lighteningof Rana pipiens tadpole skin by bovine pineal extracts.This bioassay was used to isolate melatonin from pinealextracts, which led to the elucidation of its chemicalstructure (Lerner et al., 1959). The action of melatoninto aggregate pigment granules (melanosomes) of am-phibian dermal melanophores was used to 1) postulatethe presence of melatonin receptors (Heward and Had-ley, 1975); 2) establish the first structure-activity rela-
tionships of melatonin analogs (Heward and Hadley,1975); and 3) demonstrate that melatonin receptors arecoupled to a pertussis toxin-sensitive G-protein (Whiteet al., 1987). It is noteworthy that the first melatoninreceptor (Mel1c) to be cloned was found using an expres-sion cloning strategy of mRNA from Xenopus laevismelanophores (Ebisawa et al., 1994); this particular re-ceptor has no known mammalian ortholog, but its dis-covery led to the cloning of several melatonin receptorsfrom mammals, including two human receptors (Rep-pert et al., 1994, 1995a).
The first attempts to identify brain melatonin receptorsemployed [3H]melatonin as a radioligand to label bindingsites in membranes from bovine hypothalamus, cerebralcortex, and cerebellum (Cardinali et al., 1979). This wasfollowed by the discovery of the first functional melatoninreceptor in a neuronal mammalian tissue, the rabbit retina(Dubocovich, 1983, 1985, 1988a). Melatonin, acting via apresynaptic heteroreceptor, inhibits retinal dopamine re-lease, and this bioassay was used to establish the relativeorder of potency for a series of agonists and putative an-tagonists as well as to discover the first competitive mela-tonin antagonist, luzindole (Dubocovich, 1988a,c). Vakkuriet al. (1984) introduced the radioligand 2-[125I]iodomelato-nin as a tracer for use in melatonin radioimmunoassays.This molecule turned out to be the silver bullet of melato-nin receptor research in that its selectivity and high spe-cific activity allowed the field to move forward. Soon after-ward, several laboratories simultaneously established theuse of 2-[125I]iodomelatonin as a radioligand for receptorlocalization (Vanecek et al., 1987) and receptor character-ization in native tissues (Laudon and Zisapel, 1986; Dubo-covich and Takahashi, 1987) (for reviews, see Krause andDubocovich, 1990; Krause and Dubocovich, 1991; Morganet al., 1994b; Sugden, 1994; and Dubocovich, 1995).
Melatonin receptors were first classified according toclassic pharmacological criteria using data obtained fromin vitro bioassays and radioligand binding to native tissues(Cardinali, 1981; Dubocovich, 1988a, 1995; Krause andDubocovich, 1990; Morgan et al., 1994b). The first classifi-cation scheme distinguished two putative receptors,termed ML1 and ML2 receptors, on the basis of kinetic andpharmacological differences observed in 2-[125I]iodomela-tonin binding sites (Cardinali, 1981; Dubocovich, 1988a,1995; Krause and Dubocovich, 1990; Morgan et al., 1994b).The ML1 pharmacological profile (2-iodomelatonin � mel-atonin �� N-acetylserotonin) was exhibited by both2-[125I]iodomelatonin binding in mammalian retina andpars tuberalis and the functional presynaptic receptorcharacterized in rabbit retina (Dubocovich, 1988a, 1995;Krause and Dubocovich, 1990; Morgan et al., 1994b; Ha-gan and Oakley, 1995). In contrast, 2-[125I]iodomelatoninbinding to the ML2 site (later termed MT3) in hamsterbrain membranes was distinguished by another endoge-nous ligand, N-acetylserotonin, that showed equal affinitywith melatonin (ML2: 2-iodomelatonin � melatonin � N-acetylserotonin) (Dubocovich, 1988b, 1995; Krause and
MELATONIN RECEPTORS 347
Dubocovich, 1990; Pickering and Niles, 1990; Molinari etal., 1996).
The next milestone was the cloning of two mammalianG protein-coupled melatonin receptors (GPCRs), nowtermed MT1 and MT2 (formerly Mel1a and Mel1b), (Rep-pert et al., 1994, 1995a,b, 1996). 2-125I-Iodomelatoninbinding to both recombinant hMT1 and hMT2 melatoninreceptors exhibits the general pharmacology of the ML1type (Reppert et al., 1996; Dubocovich et al., 1997).These two melatonin receptors were defined as uniquetypes on the basis of their distinct molecular structureand chromosomal localization (Reppert et al., 1994,1995a,b, 1996; Slaugenhaupt et al., 1995; Barrett et al.,1997); subsequently, distinguishing ligands were identi-fied (Dubocovich, 1995; Dubocovich et al., 1997; Brown-ing et al., 2000; Faust et al., 2000; Audinot et al., 2003).
The mammalian melatonin binding site MT3 (previ-ously referred to as ML2) also has been pharmacologi-cally characterized. Both melatonin and its precursorN-acetylserotonin compete for binding of 2-[125I]iodome-latonin to MT3 melatonin binding sites, which show apharmacological profile distinct from mammalian G pro-tein-coupled melatonin receptors (Dubocovich, 1995;Molinari et al., 1996; Nosjean et al., 2001). Subse-quently, a protein (quinone reductase II; QR2) purifiedfrom hamster kidney was found to have a ligand bindingprofile identical to that of the MT3 binding site of ham-ster brain (Nosjean et al., 2000). In addition, brain andkidney membranes from mice with deletion of the QR2gene demonstrated lack of 2-125I-5-methoxy-carbon-ylamino-N-acetyltryptamine binding to MT3 sites (Mail-liet et al., 2004).
C. International Union of Basic and ClinicalPharmacology Criteria for Receptor Nomenclature andDrug Classification
Melatonin receptors are named and classified on thebasis of operational and structural criteria developed bythe IUPHAR Committee on Receptor Nomenclature andDrug Classification (Vanhoutte et al., 1996; Ruffolo etal., 2000). The operational criteria are fulfilled by apharmacological profile of specific ligands at the recep-tor recognition site, evidence of transduction mecha-nisms beyond the receptor, and demonstration of endo-genously expressed receptors, usually from agonist/efficacy and antagonist dissociation constants obtainedin native tissues. This information, combined with struc-tural data about the protein sequence of the receptor,allows rational classification.
The present classification of melatonin receptorsevolved from deliberations of the IUPHAR Subcommit-tee on Melatonin Receptor Nomenclature and Classifi-cation, formed in 1995, as pharmacological, functional,and structural information about the receptors emerged(Table 1). In accordance with IUPHAR guidelines (Van-houtte et al., 1996; Ruffolo et al., 2000), the receptorswere named for their endogenous ligand melatonin,
TA
BL
E1
IUP
HA
Rm
elat
onin
rece
ptor
nom
encl
atu
rean
dcl
assi
fica
tion
Nom
encl
atu
reM
T1
MT
2M
T3
Pre
viou
sn
ames
ME
L1A,
ML
1A,
Mel
1a
ME
L1
B,
ML
1B,
Mel
1b
ML
2S
tru
ctu
ral
info
rmat
ion
7TM
7TM
Gen
e/ch
rom
osom
eH
um
an35
0aa
,aP
4803
9;ch
r4q
35.1
362
aa,b
P49
286;
chr
11q2
1–22
Mou
se35
3aa
,cQ
6118
4;ch
r8
365
aa,d
U57
554;
chr
9S
elec
tive
agon
ists
S26
284e
IIK
7(1
0.3)
fN
-Ace
tyls
erot
onin
,g–k
5MC
A-N
AT
k
Sel
ecti
vean
tago
nis
tsS
2613
1e4P
-AD
OT
,l,m
4P-P
DO
T(8
.8),
l,m
GR
1281
07l ,
K18
5(9
.3)f
Pra
zosi
nh
Rad
ioli
gan
ds�3
H�M
LT
,n2-
�125I�
IML
To
�3H
�ML
T,n
2-�1
25I�
IML
T,o
�3H
�4P
-PD
OT
p2-
�125I�
5MC
A-N
AT
,k2-
�125I�
IML
Tn
Tis
sue
Fu
nct
ion
sV
ascu
lar
vaso
con
stri
ctio
n,q
–u
inh
ibit
ion
neu
ron
alfi
rin
g,v
phas
e-sh
ift
circ
adia
nac
tivi
tyrh
yth
msu
Vas
cula
rva
sodi
lati
on,t
inh
ibit
ion
reti
nal
dopa
min
ere
leas
e,l
phas
e-sh
ift
circ
adia
nrh
yth
ms
ofn
euro
nal
firi
ngm
,w,v
,x,
Leu
kocy
tead
hes
ion
u
Com
men
tsM
T1
full
-len
gth
cDN
As
wer
ecl
oned
from
hu
man
(350
aa),
a
mou
se(3
53aa
),c
shee
p(3
66aa
),a
rat
(353
aa),
yh
amst
er(3
53aa
),z
mon
key
(352
),a
aan
dpa
rtia
lcD
NA
sfr
omca
ttle
(257
aa)
and
pig
(154
aa).
bb
MT
2cD
NA
was
clon
edfr
omh
um
an(3
62aa
),z
rat
(364
aa),
ym
ouse
(365
aa),
dsh
eep
(376
aa),
ccm
onke
y(3
62aa
),a
aan
dpa
rtia
lcD
NA
from
ham
ster
.z
En
doge
nou
sli
gan
dsar
eM
LT
and
N-a
cety
lser
oton
in.
Are
cen
tre
port
iden
tifi
esan
MT
3bi
ndi
ng
site
inh
amst
erki
dney
asth
een
zym
equ
inon
ere
duct
ase-
2.d
dA
nta
gon
ist
pote
nci
esh
ave
not
yet
been
esta
blis
hed
.
aa,
amin
oac
id(s
);ch
r,ch
rom
osom
e;5M
CA
-NA
T,
5-m
eth
oxyc
arbo
nyl
amin
o-N
-ace
tylt
rypt
amin
e.a
Rep
pert
etal
.,19
94.b
Rep
pert
etal
.,19
95a.
cR
oca
etal
.,19
96.d
Jin
etal
.,20
03.e
Au
din
otet
al.,
2003
.fF
aust
etal
.,20
00.g
Du
boco
vich
,198
8b.h
Lu
cch
elli
etal
.,19
97.i
Eis
onan
dM
ull
ins,
1993
.jP
opov
aan
dD
ubo
covi
ch,
1995
.kM
olin
ari
etal
.,19
96.l
Du
boco
vich
etal
.,19
97.m
Du
boco
vich
etal
.,19
98b.
nB
row
nin
get
al.,
2000
.oD
ubo
covi
chan
dT
akah
ash
i,19
87.p
Du
boco
vich
and
Mas
ana,
1999
.qT
ing
etal
.,19
99.r
Tin
get
al.,
1997
.sM
asan
aet
al.,
2002
.tD
oole
net
al.,
1998
.uL
otu
foet
al.,
2001
.vL
iuet
al.,
1997
.wD
ubo
covi
chet
al.,
2005
.xH
un
tet
al.,
2001
.yA
udi
not
etal
.,20
08.z
Wea
ver
etal
.,19
96.a
aN
ish
iyam
aet
al.,
2009
.bb
Mes
ser
etal
.,19
97.c
cC
ogé
etal
,200
9.d
dN
osje
anet
al.,
2000
.
348 DUBOCOVICH ET AL.
which is abbreviated as “MT” using capital letters, andeach particular type of receptor was denoted by a nu-merical subscript (i.e., MT1, MT2). Species orthologs aredenoted by the recommended lower-case prefix (i.e., h,human; o, ovine; r, rat; m, mouse; e.g., hMT1). Splicevariants, if pharmacologically relevant, would be indi-cated by lowercase, subscript letters in parentheses[e.g., MT1(a)]. No such variants have yet been describedfor either of the cloned melatonin receptors. The originalmelatonin receptor nomenclature and classification in-cluded the MT3 melatonin binding site, which at thetime was thought to be a GPCR.
The first nomenclature approved by the IUPHAR No-menclature Committee for melatonin receptors was pub-lished in the IUPHAR compendium in 1998, when mel-atonin receptors were designated as mt1, MT2, and MT3.(Dubocovich et al., 1998a). The MT1 melatonin receptorwas denoted in lower case (mt1) because there was noevidence at the time that the native receptor was func-tional in mammals. Subsequent functional, pharmaco-logical, and immunohistochemical studies, as well asstudies with the MT1 knockout (KO) mice, characterizedMT1 melatonin receptor protein in native tissues leadingto the more recent classification, MT1, MT2, and MT3,published in the 2000 IUPHAR Compendium (Dubocov-ich et al., 2000).
MT3 (formerly ML2) was originally included in theclassification on the basis of operational criteria (Dubo-covich, 1995; Molinari et al., 1996; Dubocovich et al.,1998a, 2000). Because the structure of the receptor pro-tein was not yet established, it was referred to in upper-case italics, as dictated by IUPHAR guidelines. The MT3binding site has a distinct pharmacology with selectiveagonists and antagonists and similar affinity for twoendogenous indoles, melatonin and its precursor, N-ace-tylserotonin. The characterization of MT3 as a melato-nin binding site on the enzyme QR2 led the IUPHARNomenclature Committee to remove the MT3 site fromthe Melatonin Receptor Nomenclature and Classifica-tion. We expect the classification of melatonin receptorswill continue to evolve as operational and structuraldata for existing receptors are further defined and pos-sible variants and/or new receptors are characterized.This is the case for the recent cloning of the ovine MT2receptor, where for many years it was suspected thatsheep possessed only one melatonin receptor (Cogé etal., 2009).
D. Current Melatonin Receptor Nomenclature
The current nomenclature classifies the two clonedmammalian melatonin receptors into two types: MT1and MT2. Detailed pharmacological and molecular char-acterization and supporting scientific evidence for thesereceptors are described later in this review (Table 1). Seealso the Melatonin Receptor-IUPHAR database (Dubo-covich et al., 2009).
MT1 (formerly Mel1a, MEL1A, ML1A) refers to the firstcloned mammalian melatonin receptor (Reppert et al.,1994). It is a Gi/o protein-coupled receptor linked, inpart, to pertussis-toxin sensitive G proteins that medi-ate inhibition of cAMP in both recombinant expressionsystems and native tissues. Functional, immunohisto-chemical, and genetic KO studies indicate the presenceof MT1 receptors in various tissues, including the parstuberalis of the pituitary gland (von Gall et al., 2002a)and the SCN of the hypothalamus (Dubocovich et al.,2005; Dubocovich, 2007).
MT2 (formerly Mel1b, MEL1B, ML1B) refers to the sec-ond cloned mammalian melatonin receptor (Reppert etal., 1995a). It is a Gi/o protein-coupled receptor capable ofinhibiting cAMP and cGMP production in recombinantsystems and stimulating PKC activity in a native tissue,the SCN. The pharmacological profile of this receptorwas initially characterized in the retina and was definedby the use of selective MT2 melatonin receptor antago-nists (4P-PDOT and 4P-ADOT) (Dubocovich et al.,1997).
It is important to note that IUPHAR nomenclaturecriteria are applied only to mammalian receptors be-cause they are the most closely aligned with therapeu-tics. The melatonin receptor field, however, actively en-compasses a variety of species and has greatly benefitedfrom the initial characterizations and cloning of melato-nin receptors that occurred using frog melanophores(Sugden, 1989; Ebisawa et al., 1994; Sugden et al.,2004). At this point, there is no official consensus onclassifying nonmammalian receptors, such as the clonedMel1c subtype that is found in birds and amphibians(Reppert et al., 1995b).
II. G Protein-Coupled MelatoninReceptor Family
A. Protein Structure
The MT1 and MT2 melatonin receptors comprise theirown subgroup within the GPCR superfamily. Both mel-atonin receptors have a general structural motif consist-ing of seven transmembrane (TM)-spanning �-helicalsegments connected by alternating intracellular and ex-tracellular loops, with the amino terminus located on theextracellular side and the carboxyl terminus on the in-tracellular side (Fig. 3). These seven �-helical segmentscontain stretches of 20 to 25 predominantly hydrophobicresidues that span the cell membrane. The melatoninreceptors are classified with the rhodopsin/�2-adrener-gic receptor family (Deupi et al., 2007). Within this fam-ily, most of the sequence homology between the melato-nin receptors and other G protein-linked receptorsoccurs within the TM domains (Fig. 4).
The human MT1 and MT2 melatonin receptors encodeproteins of 350 and 362 amino acids, respectively. Theirpredicted mass is 39,374 and 40,188 Da, respectively;however, these numbers do not take into account possi-
MELATONIN RECEPTORS 349
ble postranslational modifications. The amino acid ho-mology for the human MT1 and MT2 melatonin receptorsis approximately 60% overall and 73% within the trans-membrane domains. The amino terminus of the MT1melatonin receptor contains two consensus sites for N-terminal asparagine-linked glycosylation, whereas thatof the MT2 shows only one site. The carboxyl tail of thetwo receptors contain consensus sites for casein kinase1�, casein kinase II, and protein kinase C as well aspostsynaptic density 95/disc-large/zona occludens bind-ing domains that may participate in membrane localiza-tion and trafficking (Hung and Sheng, 2002). Featuresthat distinguish the melatonin receptor family fromother GPCRs include a NRY motif downstream from thethird transmembrane domain and a NAXIY in trans-membrane domain 7, rather than DRY and NPXXY mo-tifs, respectively (Reppert et al., 1994, 1995a; Roca et al.,1996).
B. Gene Structure and Chromosomal Localization
Molecular analyses of genomic clones show that thegenes that encode the human MT1 and MT2 melatoninreceptors are formed by two exons separated by an �13-kilobase intron (Reppert et al., 1995a; Slaugenhaupt etal., 1995; Roca et al., 1996). The intron in the firstcytoplasmic loop of the MT1 and MT2 melatonin receptorgenes could potentially lead to alternative splice formswith distinct receptor structure, as well as operationaland transduction characteristics. Such functional splicevariants, however, have not yet been identified. It isnoteworthy that the rat MT2 receptor is composed ofthree exons, although the last exon contains no openreading frames (Ishii et al., 2009).
The melatonin receptors show distinct chromosomallocalization. The MT1 melatonin receptor was localized
to human chromosome 4q35.1 and mouse chromosome 8(Slaugenhaupt et al., 1995; Roca et al., 1996). Slaugen-haupt et al. (1995) identified a region of syntenic conser-vation between distal chromosome 4 and mouse chromo-some 8 that includes the genes plasma kallikrein(KLK3), mitochondrial uncoupling protein (UPC), andcoagulation factor XI (F11) (Beaubien et al., 1991; Millset al., 1992). By contrast, the MT2 melatonin receptormaps to human chromosome 11q21–22 (Reppert et al.,1995a). Reppert et al., 1995 (Reppert et al., 1995a),pointed out that the hMT2 receptor maps to a regionsyntenic to mouse chromosome 9 in the region of the D2dopamine receptor (Drd2) and thymus cell antigen 1(thy) loci (Seldin et al., 1991; Goldsborough et al., 1993).
The phylogenetic tree of the melatonin receptors(MT1, MT2, and Mel1C) and the melatonin-related recep-tor GPR50 (also known as melatonin-related receptor orH9) sequences revealed that GPR50, which cannot bindmelatonin, is relatively distant to the functional mela-tonin receptors (MT1, MT2, Mel1C). The Mel1C receptor,which is not expressed in mammals, is phylogeneticallycloser to the MT2 receptor than to the MT1 receptor. Thehuman MT1 receptor shows more similarities with therodent MT1 receptors than with the bovine, ovine, andporcine MT2 receptors. As already observed for otherGPCRs, the ovine MT1 receptor shows significant homol-ogy with the bovine MT1 receptor (Fig. 4).
Whereas the Mel1C receptor has been found only infish, birds, and X. laevis, GPR50 has only been found ineutherian mammals and not birds or fish. An in silicoapproach has suggested that GPR50 is the ortholog ofthe Mel1C receptor (Dufourny et al., 2008). This conclu-sion is based on an analysis of the melatonin receptorfamily phylogenetic tree and the conserved synteny ofgenes surrounding the Mel1C and GPR50 genes. It is
FIG. 3. Membrane topology of the hMT1 melatonin receptor showing amino acids conserved in the hMT2 receptor. Gray circles denote amino acidsidentical in the hMT1 and hMT2 melatonin receptors. The two glycosylation sites on the hMT1 receptor are denoted (Y) in the N terminus. [Adaptedfrom Reppert SM and Weaver DR (1995) Melatonin madness. Cell 83:1059–1062. Copyright © 1995 Elsevier Inc. Used with permission.]
350 DUBOCOVICH ET AL.
suggested that rapid evolution of Mel1C into GPR50 ledto the mutation of several critical amino acids and theaddition of a long C-terminal tail resulting in the loss ofaffinity of GPR50 for melatonin. However, formation ofthe GPR50/MT1 receptor heterodimer in recombinantcells significantly reduces the affinity and potency ofmelatonin agonists binding for the MT1 melatonin re-ceptor (Levoye et al., 2006a). Recent evidence supportsthe idea that GPR50 expression in the Siberian hamsterependymal layer is under photoperiod control (Barrett etal., 2006). This orphan receptor may also be important inregulating energy metabolism (Ivanova et al., 2008).
The MT2 melatonin receptor gene cloned from theSiberian or Syrian hamsters seems to be a pseudogenebecause it is endowed with two nonsense mutations inthe coding region of the receptor cDNA. The stop codonsare located in transmembrane domain V and in thesecond extracellular loop (Weaver et al., 1996). The Si-berian and Syrian hamster are considered natural MT2melatonin receptor mutants.
C. Melatonin Receptor Polymorphisms
Genetic polymorphisms have been reported for mela-tonin receptors in human and sheep. In human, poly-morphisms have been compared for both MT1 and MT2in subjects with circadian rhythm sleep disorders andcontrols. Seven mutations were found in the MT1 recep-tor, with two that resulted in amino acid changes: R54Win the first cytoplasmic loop and A157V in the fourthtransmembrane domain (Ebisawa et al., 1999). Al-though the mutations were more common in non-24-hsleep-wake syndrome subjects than in delayed sleepphase syndrome or controls, no significant change inreceptor affinity and/or density was observed when themutants were expressed in heterologous cells. Two mu-tations were also reported for the hMT2: G24E in theN-terminal domain and L66F in the first cytoplasmicloop. However, neither shows altered MT2 receptor bind-ing characteristics (Ebisawa et al., 2000). The effect ofthese mutations in melatonin receptor function has notbeen reported.
Melatonin secretion follows a circadian rhythm withhigh levels at night. By contrast, insulin release is highduring the day. The drop in insulin levels at night mayresult from endogenous melatonin-mediated inhibi-tion by activation of MT1 and MT2 in pancreatic islets(Peschke et al., 2002; Mulder et al., 2009). Recentstudies have revealed an association of high fastingplasma glucose, early stage impairment of insulin se-cretion, and increased risk of type 2 diabetes in per-sons with genetic variations in the MTNR1B geneencoding the MT2 melatonin receptor (Bouatia-Naji etal., 2009; Lyssenko et al., 2009; Prokopenko et al.,2009). Based on increases in MT2 melatonin receptormRNA expression in human pancreatic islets of sub-jects without diabetes with the risk allele and subjectswith type 2 diabetes led to the suggestion that anincrease in MT2 receptor density may be involved inthe pathogenesis of these conditions (Lyssenko et al.,2009). However, whether increases in mRNA expres-sion reflect increases in MT2 melatonin receptor den-sity in pancreatic islets is not known.
Several polymorphisms have been described in theovine MT1 type, leading to changes in amino acids(A282D, H358R, I361V), one in extracellular loop 3 andtwo in the carboxyl-terminal tail. This variant receptor,which seems fully functional, has not been linked with aspecific phenotype (Barrett et al., 1997). Polymorphismof MnlI restriction sites in exon II of the MT1 receptor
Huma
Sheep
Mouse
Huma
Sheep
Mouse
Cattle
Sheep
Djung
Cattle
Sheep
Djung
Golde
Mouse
Rat M
Golde
Mouse
Rat M
Huma
Mouse
Pig M
Huma
Mouse
Pig M
Rat M
Huma
Frog M
Rat M
Huma
Frog MFrog M
Chick
Frog M
Chick
an H9
H9
e H9
an H9
H9
e H9
e MT1
MT1
garian Hamster MT1
e MT1
MT1
garian Hamster MT1
n Hamster MT1
e MT1
MT1
n Hamster MT1
e MT1
MT1
an MT1
e MT2
T1
an MT1
e MT2
T1
MT2
an MT2
Mel1c
MT2
an MT2
Mel1cMel1c
Mel1c
Mel1c
Mel1c
FIG. 4. MT1 and MT2 melatonin receptor dendrogram. Phylogenetictree of melatonin receptor or melatonin receptor-related (GPR50, mela-tonin-related receptor or H9) sequences. The evolutionary distances be-tween the different sequences were calculated with the matrix of Blosum62 score. The tree was drawn using the Unweighted Pair Group Methodwith Arithmetic mean (UPGMA). GenBank accession numbers and thenumber of amino acids for each receptor are as follows: human H9:U52219, 613; sheep H9: U52221, 613; mouse H9: AF065145, 791; cattleMT1: U73327, 257; sheep MT1: U14109, 366; Djungarian hamster MT1:U14110, 353; golden hamster MT1: AF061158, 325; mouse MT1: U52222,353; rat MT1: AF130341, 326; human MT1: U14108, 350; Pig MT1:U73326, 154; mouse MT2: AY145850, 365; rat MT2: U28218, 120; humanMT2: U25341, 362. The sequences for cattle and pig MT1 receptors havebeen partially cloned.
MELATONIN RECEPTORS 351
was analyzed in the Mérinos d’Arles ewe in relation tothe expression of reproductive seasonality (Pelletier etal., 2000). The MnlI restriction sites show an associationbetween the homozygous genotype for the absence of aMnlI site at position 605 (�/�) and seasonal anovulatoryactivity. Other mutations were observed, not simulta-neously, at positions 706 and 893, which resulted in thesubstitution of a valine by an isoleucine and of an ala-nine by an aspartic acid, respectively. However, in theIle-de-France sheep breed, the two allelic forms of theMT1 receptor gene have no direct effect on the seasonalpattern of various seasonal functions. It was suggestedthat the effect of this polymorphism on seasonal functionseems to be dependent on the breed and/or environmen-tal condition (Hernandez et al., 2005).
D. Molecular Structure of MT1 and MT2 MelatoninReceptor Ligand Binding Pockets
1. MT1 Melatonin Receptor. A rhodopsin-based com-puter model has been used to propose the molecularstructure of the melatonin receptor binding site (Nava-jas et al., 1996). This particular model has advantagesover bacteriorhodopsin-based models for the melatoninreceptor binding site (Sugden et al., 1995; Grol andJansen, 1996), because bacteriorhodopsin is not coupledto G proteins, and its sequence has none of the distinc-tive features of the GPCR family (Baldwin, 1993). Therhodopsin-based molecular model has been investigatedby site-directed mutagenesis studies, revealing that thebinding site of the melatonin receptor has some similar-ities with those of other GPCRs of the rhodopsin/�2-adrenergic receptor family. For example, His195 in pu-tative TM5 is conserved in all of the melatonin receptors,and the position is identical to that used in the ligandbinding site of many other rhodopsin-like GPCRs. Themodel proposes that this His residue can form a hydro-gen bond with the oxygen atom of the 5-methoxy groupof melatonin. Site-directed mutagenesis of the melato-nin receptor (Conway et al., 1997; Kokkola et al., 1998)and studies with sulfur analogs of melatonin (Davies etal., 2004) have given support to this suggestion. Thebinding site model also proposes that Val192, which islocated approximately one helical turn above the His195facing the hydrophobic binding pocket, is important forthe binding of the methyl portion of the methoxy groupof melatonin. Val192 is analogous to a residue in the�2-adrenergic and 5-HT receptors that is important inligand binding (Strader et al., 1989; Ho et al., 1992; Kaoet al., 1992). In addition, Met107 in TM3 and Ser280 andAla284 in TM7, which were proposed to be important forthe binding of the N-acetyl group of melatonin, do notseem to directly participate in melatonin receptor acti-vation (Kokkola et al., 1998). Thus, computer modelinghas revealed one site that is important for melatoninbinding, but the other residues/domain(s) of the melato-nin receptor, which are critical for ligand binding, haveyet to be identified.
A second method for determining residues that areimportant for receptor binding and activation has beento modify amino acid residues conserved in the rhodop-sin-like GPCRs. Protonation of the aspartic/glutamicacid in the highly conserved D/ERY motif at the cyto-plasmic side of TM3 is believed to be involved in activa-tion of the rhodopsin-like GPCRs. The binding of theligand causes the Asp/Glu to become unprotonated, re-sulting in receptor activation, as shown most directly bythe rhodopsin receptor (Arnis et al., 1994). Changing theD/E to a neutral amino acid that mimics the unproto-nated state results in constitutive activation and im-proved coupling of many of the rhodopsin-like receptors.The melatonin receptor is unique in that it has an NRYmotif instead of the D/ERY motif. Changing the NRY toan ARY (mimicking the unprotonated, activated state)actually decreases binding to such an extent that it isimpossible to measure receptor activation (Nelson et al.,2001). Changing the melatonin receptor NRY motif tothe D/ERY motif modestly decreases the binding affinity(2-fold) and decreases the capacity for melatonin to ac-tivate the receptor (Nelson et al., 2001). Thus, unlikeother rhodopsin-like GPCRs, the melatonin receptorseems to not require deprotonation of the NRY motif tobe active, and Asn is needed for optimal ligand bindingand receptor activation.
Pro267 is a highly conserved amino acid in rhodopsin-like GPCRs. This proline residue occurs in the center ofTM6, causing a kink in the center of the �-helix. Mutat-ing the proline to an alanine results in constitutive ac-tivity of the yeast �-factor and �2-adrenergic receptor(Konopka et al., 1996), presumably by making the alphahelix less “kinked.” However, when the correspondingresidue in the MT1 melatonin receptor, �253, is mutatedto Ala, constitutive activity of the melatonin receptor isnot seen and, in fact, the receptor affinity is decreased byseveralfold (Kokkola et al., 1998).
G-protein coupled melatonin receptors have two con-served cysteines (Cys127, Cys130) between helix III andthe second intracellular loop, a region important in re-ceptor/G-protein coupling. Indeed, mutation of Cys127and Cys130 to Ser in the MT1 receptor revealed thatthese cysteines are necessary for normal G protein acti-vation and receptor trafficking (Kokkola et al., 2005).
Finally, there is an NPXXY sequence found at the endof TM7 in rhodopsin-like receptors. The Asn302 is pro-posed to interact with the Asp83, suggesting that TM2and TM7 are in close proximity. Kinking and twisting ofPro303 is proposed to allow these two residues to comein contact. It is noteworthy that the melatonin receptorshave retained the conserved Asp83 (Asp73) and Asn302(Asn291) but have replaced the Pro303 with an alanine(Ala292). How the Ala292 affects the overall structuralmotif is not known (Table 2).
Analysis of rhodopsin binding and of extensive mu-tagenesis data involving the �2 adrenergic receptor sug-gests that TMs 3, 5, 6, and 7, especially TM3 and/or
352 DUBOCOVICH ET AL.
TA
BL
E2
Eff
ect
ofam
ino
acid
mu
tati
ons
onli
gan
dbi
nd
ing
toth
eM
T1
mel
aton
inre
cept
orA
min
oac
ids
are
repr
esen
ted
insi
ngl
e-le
tter
code
wit
hpo
siti
onn
um
ber
show
n.
Su
pers
crip
tsaf
ter
the
seco
nd
amin
oac
idin
dica
teth
atth
esu
bsti
tute
dam
ino
acid
repr
esen
tsth
eam
ino
acid
inth
ede
sign
ated
rece
ptor
atth
ean
alog
ous
posi
tion
.T
he
posi
tion
inth
etr
ansm
embr
ane
dom
ain
isin
dica
ted
usi
ng
the
nu
mbe
rin
gsc
hem
eof
Bal
lest
eros
and
Wei
nst
ein
(199
5).
Am
ino
Aci
dM
uta
tion
Sch
eme
TM
No.
Spe
cies
Exp
ress
ion
Sys
tem
Ch
arac
teri
zati
onR
efer
ence
R54
W(1
.59)
Hu
man
CO
Sce
lls
Het
eroz
ygou
spo
lym
orph
ism
wit
hn
oph
enot
ype.
Dec
reas
edB
max
(3.5
�)
and
slig
htl
yin
crea
sed
Kd.
Ebi
saw
aet
al.,
1999
S10
3A(2
.28)
Hu
man
CO
S-7
No
chan
gein
Bm
ax
orK
d.
Con
way
etal
.,20
01M
107T
(3.3
2)H
um
anC
OS
-7N
och
ange
inB
max
orK
d.
Con
way
etal
.,20
01;
Kok
kola
etal
.,19
98S
110A
a(3
.35)
Hu
man
CO
S-7
Dec
reas
edB
max
(10�
),in
crea
sed
Kd
(8�
)an
dE
C50
ofcA
MP
prod
uct
ion
(22�
).N
och
ange
inK
iof
luzi
ndo
leC
onw
ayet
al.,
2001
S11
4Aa
(3.3
9)H
um
anC
OS
-7D
ecre
ased
Bm
ax
(4�
),in
crea
sed
Kd
(9�
)an
dE
C50
ofcA
MP
prod
uct
ion
(14�
).N
och
ange
inK
iof
luzi
ndo
le.
Con
way
etal
.,20
01
N12
4A/K
(3.4
9)H
um
anA
tT20
Dec
reas
edB
max
(21�
),te
nds
tobe
reta
ined
inG
olgi
.N
osp
ecif
icbi
ndi
ng.
Nel
son
etal
.,20
01
N12
4A(3
.49)
Hu
man
Sac
char
omyc
esce
revi
siae
Incr
ease
dE
C5
0fo
rm
elat
onin
(230
�).
Kok
kola
etal
.,19
98
N12
4L(3
.49)
Hu
man
AtT
20D
ecre
ased
Bm
ax
(21�
),te
nds
tobe
aggr
egat
edn
ear
surf
ace.
No
spec
ific
bin
din
g.N
elso
net
al.,
2001
N12
4D/E
(3.4
9)H
um
anA
tT20
No
chan
gein
Bm
ax
orK
d.
Mel
aton
inin
duce
din
hib
itio
nof
cAM
P(e
ffic
acy)
and
volt
age-
sen
siti
veC
a2�
chan
nel
s,bu
tn
otK
ir3.
1/3.
2po
tass
ium
chan
nel
acti
vati
on.
Nel
son
etal
.,20
01
A15
7V(4
.55)
Hu
man
CO
Sce
lls
Het
eroz
ygou
spo
lym
orph
ism
wit
hn
oph
enot
ype.
No
chan
gein
Bm
ax
orK
d.
Ebi
saw
aet
al.,
1999
H19
5Aa
(5.4
6)H
um
anS
.ce
revi
siae
Dec
reas
edE
C5
0(3
–6�
).N
-ace
tyls
erot
onin
gave
anap
pare
nt
satu
rabl
ere
spon
se,
wh
erea
sth
ew
ild-
type
rece
ptor
did
not
satu
rate
atth
esa
me
con
cen
trat
ion
s.
Kok
kola
etal
.,19
98
H21
1F/l
(5.4
6)O
vin
eC
OS
-7In
crea
seK
d(6
�)
wit
hm
elat
onin
.D
ecre
ased
Ki
(3–1
5�)
wit
hN
-NE
A.
No
chan
gein
Ki
wit
hN
-ace
tyls
erot
onin
.C
onw
ayet
al.,
1997
V19
2T�
H19
5A(5
.42
�5.
46)
Hu
man
S.
cere
visi
aeN
osp
ecif
icre
spon
seK
okko
laet
al.,
1998
V20
8A(5
.42)
Ovi
ne
CO
S-7
No
chan
gein
Kd
orin
Ki
ofse
vera
lm
elat
onin
anal
ogs.
Con
way
etal
.,19
97V
208L
(5.4
2)O
vin
eC
OS
-7In
crea
sed
Kd
and
Ki
for
seve
ral
mel
aton
inan
alog
s(5
–12�
).C
onw
ayet
al.,
1997
A25
2C(6
.49)
Hu
man
CO
S-7
No
chan
gein
Kd
orB
max.
Con
way
etal
.,20
00H
um
anC
OS
-7N
och
ange
inK
dor
Bm
ax.
Gu
bitz
and
Rep
pert
,20
00G
258T
(6.5
5)H
um
anC
OS
-7S
peci
fic
bin
din
gdr
asti
call
yre
duce
dG
ubi
tzan
dR
eppe
rt,
2000
;C
onw
ayet
al.,
2000
A25
2C�
G25
8T(6
.49
�6.
55)
Hu
man
CO
S-7
No
spec
ific
bin
din
gG
ubi
tzan
dR
eppe
rt,
2000
P25
3A(6
.50)
Hu
man
S.
cere
visi
aeN
osp
ecif
icre
spon
se.
Kok
kola
etal
.,19
98A
202D
,H
342R
,I3
47V
(ext
.lo
op3,
C-t
erm
inal
)O
vin
eL
-cel
lsP
olym
orph
ism
ofpr
evio
usl
ycl
oned
ovin
eM
T1.
No
phen
otyp
ein
vivo
and
full
yfu
nct
ion
alin
mou
se.
Lce
lls
assh
own
byh
igh
affi
nit
ybi
ndi
ng,
com
peti
tion
bin
din
gan
alys
is,
GT
P�
San
din
hib
itio
nof
cAM
P.
Bar
rett
etal
.,19
97
S28
0A(7
.38)
Hu
man
S.
cere
visi
aeN
och
ange
inap
pare
nt
EC
50
Kok
kola
etal
.,19
98S
280F
�A
284G
(7.3
8�
7.42
)H
um
anS
.ce
revi
siae
No
spec
ific
resp
onse
Kok
kola
etal
.,19
98
N-N
EA
,N
-�2-
(1-n
aph
thyl
)et
hyl
�ace
tam
ide.
aA
min
oac
idre
sidu
esim
port
ant
for
mod
ula
tin
gbi
ndi
ng
toth
eM
T1
rece
ptor
(Far
ceet
al.,
2008
).
MELATONIN RECEPTORS 353
TM7, are major players in ligand binding (Kobilka andDeupi, 2007; Rosenbaum et al., 2007). The TM3 domainsof ClassA GPCRs contain a high number of Ser/Thr/Cysresidues (seven residues/TM3). These residues form hy-drogen bonds to the peptide backbone and thereby bendand twist helices (Gray and Matthews, 1984; Ballesteroset al., 2000). Thus, different hydrogen bonding statesmay result in different TM3 conformations that repre-sent different functional states of the same receptor,such as liganded versus unliganded or active versusinactive (Ballesteros et al., 2001). As in other GPCRs,the Ser/Thr/Cys residues in TM3 of MT1 are importantin ligand binding at MT1 receptors. Mutations of Ser110and Ser114, but not Ser103, to alanine reduced melato-nin binding but did not affect luzindole binding (Conwayet al., 2001). These experiments give some molecularsupport to the experimental data from the structure-activity relationships of a series of luzindole analogssuggesting that melatonin receptor agonist and antago-nist-binding sites may differ (Teh and Sugden, 1998)(Table 2). Based on these studies, a two dimensionalmodel of the MT1 melatonin receptor was created com-prising the transmembrane domains and the potentialorientation of melatonin in the binding pocket (Barrettet al., 2003). This model takes into account the interac-tion between melatonin’s methoxy group and the con-served histidine (His195) in TM5.
A third method for determining important residuesfor melatonin receptor binding has been to create chi-meric melatonin MT1 and melatonin-related receptors(GPR50). Despite having 57% amino acid sequenceidentity with the TM domains of the MT1 receptor, theGPR50 does not bind 2-[125I]iodomelatonin or [3H]me-latonin. It is noteworthy that two studies involvingchimeric receptors suggest that TM6, extracellularloop 2, and intracellular loop 2 are critical for mela-tonin receptor binding to the MT1 melatonin receptor.Replacement of transmembrane domains 1, 2, 3, 5, or7 of the hMT1 receptor by the corresponding TM fromGPR50 induces only minor change in 2-[125I]iodome-latonin binding affinity. On the contrary, chimericreceptors with TM 6 from GPR50 receptor display nospecific binding (Conway et al., 2000; Gubitz and Rep-pert, 2000). Gubitz and Reppert (2000) reported thatreplacement of TM4 of the hMT1 with TM4 of GPR50did not alter binding affinity, but Conway et al. (2000)failed to detect specific binding in this mutant. Toconfirm that some specific amino acids of TM 6 play animportant role in binding, both research groups per-formed point mutations in this transmembrane of theMT1 receptor. Mutation of glycine to threonine(G258T) severely reduced both the binding and acti-vation of the MT1 receptor (Conway et al., 2000; Gu-bitz and Reppert, 2000). The mutant A252C displayedbinding affinity close to that of the native hMT1 re-ceptor. Double mutation of glycine to threonine(G258T) and alanine to cysteine (A252C) was found to
completely inhibit binding by Gubitz and Reppert(2000) and to have no effect on binding by Conway etal. (2000). Taken together, these data do support theidea that mutation of glycine 258, which is predictedto face the hydrophilic receptor core, may be impor-tant for maintaining an appropriate MT1 melatoninreceptor structure.
2. MT2 Melatonin Receptor. Melatonin receptors,like most other GPCRs, contain a conserved cysteineresidue in extracellular loop 1 and in extracellular loop2. Mutation of these Cys residues in rhodopsin (Karniket al., 1988), �-opioid (Ehrlich et al., 1998), platelet-activating factor (Le Gouill et al., 1997), and M3 musca-rinic receptors (Zeng et al., 1999) demonstrate the crit-ical importance of this disulfide bond for the properreceptor conformation for ligand binding, receptor acti-vation, and cell surface expression. These conserved cys-teine residues, however, do not always participate indisulfide bonding, as shown previously for the �2-adren-ergic receptor (Noda et al., 1994). The disulfide bondformation between Cys113 and Cys190 residues wasshown to be crucial to maintain a proper hMT2 receptorconformation for melatonin binding without altering cellsurface receptor expression (Mseeh et al., 2002).Whether this disulfide bond occurs within a single mel-atonin receptor or between two melatonin receptorsforming a dimer remains to be determined. N-ethylma-leimide alkylation of Cys140 appears to contribute tochanges in ligand affinity, whereas alkylation of Cys143and Cys219 reduced binding capacity (Mseeh et al.,2002). The cysteines involved in N-ethylmaleimide-in-duced changes in affinity and receptor density are prob-ably located in receptor regions near the melatonin bind-ing site and/or G protein coupling region.
Key conserved amino acids (Table 3) seem to be in-volved in ligand binding to the MT2 melatonin receptorsas determined in binding studies after mutation to ala-nine. Mutation of Asn175 in TM4 or His208 in TM5 ofthe hMT2 melatonin receptor significantly decreased thebinding affinity for melatonin (Gerdin et al., 2003).Asn175 in TM4 seems to facilitate binding of the 5-me-thoxy group of the melatonin molecule to the hMT2melatonin receptor (Gerdin et al., 2003). Thus, His208 inTM5 in both the oMT1 receptor (Conway et al., 1997) andthe hMT2 melatonin receptor (Gerdin et al., 2003) arecritical for melatonin binding. Trp264 or Phe257 in TM6,although not critical for melatonin binding, may interactwith aromatic regions of molecules such as luzindole and4P-ADOT. Mutation of Ser123 or Ser127 in TM3 orSer293 in TM7 of the MT2 receptor did not affect bind-ing affinity, although equivalent serines (Ser110 andSer114 in TM3) were reported to be critical for melato-nin binding to the hMT1 melatonin receptor (Conway etal., 2001). Thus, the binding pockets of the MT1 and MT2melatonin receptors seem to share a common histidineresidue (His195/208 in TM5) but also have distinct res-idues (Ser110/123 and Ser110/127 in TM3) necessary for
354 DUBOCOVICH ET AL.
TABLE 3Effect of amino acid mutations on ligand binding to the hMT2 melatonin receptor
Amino acids are represented in single-letter code with position number shown. Superscripts after the second amino acid indicate that the substituted amino acid representsthe amino acid in the designated receptor at the analogous position. The position in the transmembrane domain is indicated using the numbering scheme of Ballesteros andWeinstein (1995).
Amino Acid MutationScheme TM No.
ExpressionSystem Characterization Reference
HumanG24E NTerm COS-7 Heterozygous polymorphism with no phenotype. When expressed
in COS-7 cells no change in Kd or Ki with melatonin.Ebisawa et al., 2000
L66F 1.58 COS-7 Heterozygous polymorphism with no phenotype. Ebisawa et al., 2000C113A ECL1 HEK293 No specific binding. Mseeh et al., 2002C140A ICL2 HEK293 No change in Kd, slightly increased Ki for melatonin (1.6�),
decreased Bmax (22�).Mseeh et al., 2002
C143A ICL2 HEK293 No change in Kd, slightly increased Ki for melatonin (1.4�),slightly increased Bmax (1.8�).
Mseeh et al., 2002
C190A ECL2 HEK293 No specific binding. Mseeh et al., 2002C219A 5.57 HEK293 No change in Kd or Ki. Decreased Bmax (5�). Mseeh et al., 2002C263A 6.47 HEK293 No change in Kd or Ki. Decreased Bmax (31�). Mseeh et al., 2002C302A 7.47 HEK293 No change in Kd or Ki. Decreased Bmax (4�). Mseeh et al., 2002S123A 3.35 HEK293 No change in Kd or Ki. Decreased Bmax (5�). Gerdin et al., 2003S127A 3.39 HEK293 No change in Kd or Ki. Decreased Bmax (3�). Gerdin et al., 2003N175Aa 4.60 HEK293 No change in Kd, slightly increased Ki for melatonin. No change
in Bmax.Gerdin et al., 2003
H208Aa 5.46 HEK293 Increased Kd and Ki for melatonin. No change in Bmax. Gerdin et al., 2003F257A 6.41 HEK293 No change in Kd or Ki. No change in Bmax. Gerdin et al., 2003W264A 6.48 HEK293 Decreased Kd, no change in Ki. Decreased Bmax (22�). Gerdin et al., 2003S293A 7.38 HEK293 No change in Kd or Ki. No change in Bmax. Gerdin et al., 2003V204Aa 5.42 HEK293 No specific binding. Mazna et al., 2004V205 5.43 HEK293 No change in Kd. No change in Bmax. Mazna et al., 2004F209A 5.47 HEK293 No change in Kd. Decreased Bmax. No change in Ki for melatonin,
luzindole or 4P-PDOT.Mazna et al., 2004
G271T 6.55 HEK293 Not saturable. Mazna et al., 2004L272Aa 6.56 HEK293 No specific binding. Mazna et al., 2004Y298Aa 7.43 HEK293 No specific binding. Mazna et al., 2004M120A 3.32 HEK293 No change in Kd or Bmax. Mazna et al., 2005G121A 3.33 HEK293 No change in Kd or Bmax. Mazna et al., 2005G121I 3.33 HEK293 No change in Kd or Bmax. Mazna et al., 2005V124A 3.36 HEK293 No change in Kd with decreased Bmax. Mazna et al., 2005I125A 3.37 HEK293 No change in Kd or Bmax. Mazna et al., 2005Y188A ECL2 HEK293 No specific binding. Mazna et al., 2005Y188F ECL2 HEK293 No specific binding. Mazna et al., 2005N268Aa 6.52 HEK293 No specific binding. Mazna et al., 2005N268Da 6.52 HEK293 No specific binding. Mazna et al., 2005N268La 6.52 HEK293 No specific binding. Mazna et al., 2005N268Qa 6.52 HEK293 No change in Kd or Bmax. Mazna et al., 2005A275I 6.59 HEK293 No specific binding. Mazna et al., 2005A275Va 6.59 HEK293 No change in Kd or Bmax. Mazna et al., 2005V291Aa 7.36 HEK293 No specific binding. Mazna et al., 2005V291Ia 7.36 HEK293 No specific binding. Mazna et al., 2005L295Aa 7.40 HEK293 No specific binding. Mazna et al., 2005L295Ia 7.40 HEK293 No specific binding. Mazna et al., 2005L295Va 7.40 HEK293 No specific binding. Mazna et al., 2005
HamsterP41A 1.33 CHO-K1 No change in Kd or Bmax. No change in EC50 or Emax for
melatonin or 2-iodomelatonin stimulation of GTP�35S binding.Mazna et al., 2008
P93A 2.57 CHO-K1 No change in Kd or Bmax. No change in EC50 or Emax formelatonin or 2-iodomelatonin stimulation of GTP�35S binding.
Mazna et al., 2008
P95A 2.59 CHO-K1 No change in Kd or Bmax. No change in EC50 or Emax formelatonin or 2-iodomelatonin stimulation of GTP�35S binding.
Mazna et al., 2008
P158A 4.40 CHO-K1 No change in Kd or Bmax. No change in EC50 or Emax formelatonin or 2-iodomelatonin stimulation of GTP�35S binding.
Mazna et al., 2008
P174A 4.59 CHO-K1 No specific binding. Mazna et al., 2008P174G 4.59 CHO-K1 No specific binding. Mazna et al., 2008P212A 5.50 CHO-K1 No change in Kd or Bmax. Decreased Emax for melatonin and 2-
iodomelatonin, stimulation of GTP�35S binding.Mazna et al., 2008
P212G 5.50 CHO-K1 No change in Kd or Bmax. Increased EC50 for 2-iodomelatoninstimulation of GTP�35S binding with no change in Emax.
Mazna et al., 2008
P266A 6.50 CHO-K1 No specific binding. Mazna et al., 2008P266G 6.50 CHO-K1 No specific binding. Mazna et al., 2008A305P 7.50 CHO-K1 No specific binding. Mazna et al., 2008A305V 7.50 CHO-K1 No change for Kd or Bmax. Increased EC50 for melatonin and
2-iodomelatonin stimulation of GTP�35S binding withdecreased Emax.
Mazna et al., 2008
HEK, human embryonic kidney; CHO, Chinese hamster ovary.a Amino acid residues important for modulating binding to the MT1 receptor (Farce et al., 2008).
MELATONIN RECEPTORS 355
ligand binding. Mazna et al. (2004) identified severalamino acids in TM V (Val204), VI (Leu272), and VII(Tyr298) that are involved in melatonin interactionswith the MT2 melatonin receptor binding pocket. In asubsequent studies, this group demonstrated that resi-dues Asn268 and Ala275 in TM6 as well as residuesVal291 and Leu295 in TM7 are essential for 2-iodome-latonin binding to the hMT2 receptor (Mazna et al.,2005). Mazna et al. (2008) assessed the impact of muta-tions on the MT2 melatonin receptor structure by molec-ular dynamic simulations of the receptors embedded inthe fully hydrated phospholipid bilayer and demon-strated that residues Pro174, Pro212, and Pro266 areimportant for the ligand binding and/or signaling of thisreceptor (Table 3). Taken together, the identification ofsequence specific motifs may ultimately provide the mo-lecular basis for the rational design of type specific ther-apeutic compounds.
Farce et al. (2008) have published models showingthe predicted binding site for melatonin on the MT1
and MT2 melatonin receptors based on site-directedmutagenesis analysis and a three-dimensional homol-ogy modeling of the receptors using bovine rhodopsinas a template (Fig. 5). In these models, the bindingspace for melatonin on the MT1 receptor seems to berelatively smaller than the space for the MT2 receptor.The conserved histidine of TM5 (His195 for MT1 andHis208 for MT2), which is predicted to bind to the methoxygroup, seem to be common to both receptor binding sites(Gerdin et al., 2003). The amine moiety interacts withAsn175 of TM4 in the MT2 receptor; however, a corre-sponding amino acid in the MT1 receptor TM3 is not anAsn; rather, serines (Ser110 and Ser114) seem to bind tomelatonin (see Fig. 5; Farce et al., 2008).
III. Cellular Signaling of MT1 and MT2Melatonin Receptors
The best-known signaling pathway for melatonin recep-tors is inhibition of cAMP formation via pertussis toxin-
FIG. 5. MT1 and MT2 melatonin receptor 3 dimensional models and putative mode of binding for melatonin. A and B show critical amino acidsresidues for melatonin binding to the MT1 and MT2 melatonin receptors, respectively. The amino acids labeled in white have been defined bysite-directed mutagenesis to modulate binding affinity (see Table 2 and 3). C and D show interactions among melatonin and key amino acid residuesimportant for binding to the MT1 and MT2 melatonin receptors, respectively. [Adapted from Farce A, Chugunov AO, Logé C, Sabaouni A, Yous S, DillyS, Renault N, Vergoten G, Efremov RG, Lesieur D, and Chavatte P (2008) Homology modelling of MT1 and MT2 receptors. Eur J Med Chem43:1926–1944. Copyright © 2008 Elsevier Masson SAS. Used with permission.]
356 DUBOCOVICH ET AL.
sensitive G proteins. Although first described in frog mela-nophores (White et al., 1987), melatonin-mediateddecreases in cAMP have been observed in a number ofmammalian tissues, including pituitary, SCN, and cere-bral arteries (Capsoni et al., 1994; Morgan et al., 1994b).Pertussis toxin (PTX) sensitivity indicates the involvementof G proteins in the Gi/Go family; however, the identity ofthe specific G proteins that transduce the melatonin signalin native tissues is not known. Using recombinant humanreceptors, adenylate cyclase inhibition has been confirmedas a signaling mechanism for both MT1 and MT2 melato-nin receptor types (Reppert et al., 1995a). Recent studies,however, indicate that melatonin can elicit multiple recep-
tor-mediated intracellular responses. Signal transductionmechanisms shown to be associated with MT1 and MT2receptors are summarized below.
A. MT1 Melatonin Receptor Signaling
MT1 melatonin receptors can couple to both PTX-sensitive (Gi) and insensitive (Gq/11) G proteins (Brydon etal., 1999b; Roka et al., 1999) (Fig. 6A). Activation of MT1melatonin receptors decreases forskolin-stimulated cAMPformation (Reppert et al., 1994; Witt-Enderby and Dubo-covich, 1996; Brydon et al., 1999a; Petit et al., 1999). Pro-tein kinase A activity and phosphorylation of the cAMPresponsive element-binding protein (CREB) (Witt-
FIG. 6. MT1 and MT2 melatonin receptor signaling. A, melatonin (MLT) signals through activation of the MT1 receptor via two parallel pathwaysmediated by the �-subunit (i.e., inhibition of cAMP formation) and the ��-subunits [i.e., potentiation of phosphoinositide turnover stimulated by aGq-coupled receptor (R)] of Gi. B, signaling pathways coupled to MT2 melatonin receptor activation. Melatonin-mediated phase shifts of circadianrhythms through MT2 receptors are mediated by PKC activation (the mechanism leading to PKC activation remains putative, however). DAG,diacylglycerol; PKA, protein kinase A; R, Gq-coupled receptor (i.e., prostaglandin F2� receptor FP and purinergic receptor P2Y) (Masana andDubocovich, 2001). [Adapted from Masana MI and Dubocovich ML (2001) Melatonin receptor signaling: finding the path through the dark. Sci STKE2001:pe39. Copyright © 2001 American Association for the Advancement of Science. Used with permission.]
MELATONIN RECEPTORS 357
Enderby et al., 1998) are also inhibited. Some authors havealso proposed that the �� subunit of a PTX-sensitive Gprotein may mediate the potentiation of phospholipaseactivation by prostaglandin F2�, leading to increase inphosphoinositide turnover (Godson and Reppert, 1997) orin ATP (Roka et al., 1999). In addition, the MEK1/2-ERK1/2 pathway is stimulated by MT1 receptors in non-neuronal cells (Witt-Enderby et al., 2000; New et al., 2003;Radio et al., 2006).
MT1 melatonin receptors can also regulate ion fluxes andspecific ion channels. Activation of endogenous MT1 receptorsin ovine pars tuberalis cells increases intracellular calciumvia PTX-insensitive G proteins (Brydon et al., 1999a) (Fig.6A). In contrast, melatonin acts via PTX-sensitive G proteinsto inhibit calcium influx in neonatal rat pituitary cells (Slanaret al., 2000) and in AtT20 cells expressing MT1 receptors(Nelson et al., 2001). Vasoconstriction seems to be mediatedby decreases in cAMP-mediated phosphorylation of calcium-activated potassium channels (BKCa) through Gi/Go protein-coupled MT1 melatonin receptors present in the smooth mus-cle, although participation of receptors localized in theendothelium cannot be ruled out (Nelson and Quayle, 1995;Geary et al., 1998; Masana et al., 2002). Conversely, melato-nin transiently increases BKCa channel activity in culturedrat myometrial cells (Steffens et al., 2003), an effect that canbe blocked by PTX as well as by inhibition of protein kinase Aactivity. Inward-rectifier potassium channels (Kir) are alsoactivated by melatonin. MT1 melatonin receptors expressedin X. laevis oocytes (Nelson et al., 1996) or AtT20 cells (Nelsonet al., 2001) activate Kir3 inward-rectifier potassium chan-nels through a PTX-sensitive mechanism that may involve�� subunits of Gi proteins. Activation of Kir3 channels mayunderlie melatonin-mediated increases in potassium conduc-tance (Jiang et al., 1995) and may be the mechanism bywhich melatonin inhibits neuronal firing in the SCN (Masonand Brooks, 1988; Shibata et al., 1989; Stehle et al., 1989).Hyperpolarization of neonatal pituitary cells may also bemediated by activation of MT1 melatonin receptors (Vanecekand Klein, 1992). Thus the data available indicate that acti-vation of MT1 melatonin receptors elicits a variety of tissue-dependent signaling responses.
B. MT2 Melatonin Receptor Signaling
Recombinant MT2 melatonin receptors have alsobeen shown to couple to inhibition of cAMP formation(Reppert et al., 1995a; Petit et al., 1999) (Fig. 6B). Inaddition, activation of MT2 melatonin receptors alsocan lead to inhibition of cGMP formation (Petit et al.,1999). In the SCN, melatonin increases PKC activitythrough activation of MT2 melatonin receptors, be-cause this response is blocked by the selective MT2receptor antagonist 4P-PDOT (Hunt et al., 2001). Thisfinding suggests that MT2 melatonin receptors inter-act with the phospholipase C/diacylglycerol signalingpathway (McArthur et al., 1997). In the retina, MT2melatonin receptors inhibit neurotransmitter releasethrough a mechanism that probably involves intracel-
lular calcium regulation (Dubocovich, 1995). Humanmyometrium from both pregnant and nonpregnantwomen expresses both MT1 and MT2 melatonin recep-tors (Schlabritz-Loutsevitch et al., 2003). In thisstudy, 4P-PDOT blocked the melatonin-induced inhi-bition of cAMP signaling in cultured myometrial cellsfrom nonpregnant women, suggesting the involve-ment of the MT2 melatonin receptor. Recent data con-firmed the involvement of MT2 melatonin receptor inthe action of melatonin on human myometrial smoothmuscle cells and further demonstrated the involve-ment of PKC in MT2 melatonin receptor signaling(Sharkey and Olcese, 2007; Sharkey et al., 2009).
C. Melatonin Receptor Regulation
Regulation of signal transduction events is essentialfor maintaining timely and efficient cellular responsesand homeostasis. Activation of GPCRs leads tochanges in receptor sensitivity (desensitization, sen-sitization, internalization) and trafficking, leading tochanges in ligand efficacy (Ferguson, 2001). The MT1and MT2 melatonin receptors are differentially anddistinctly regulated by physiological (30 – 400 pM) andsupraphysiological (1–1000 nM) concentrations ofmelatonin. Physiological concentrations of nocturnalmelatonin (100 – 400 pM) are already well above thepotency (EC50) for the melatonin receptors, which areactivated by picomolar concentrations of melatonin(Reppert et al., 1996; Dubocovich et al., 1997). Day-time concentrations typically fall below 30 pM and yetthey can still induce activation and desensitization ofmelatonin receptors upon prolonged exposure to thehormone (�8 h) (Gerdin et al., 2004b). Blood melato-nin levels after administration of an oral dose of 0.3mg are similar to endogenous levels found in humansat night (Dollins et al., 1994). However, oral doses ofmelatonin or other ligands at �1 mg may increaseblood levels several times above the concentrationnecessary to activate melatonin receptors and there-fore may alter receptor sensitivity (Dollins et al.,1994; Vachharajani et al., 2003; Mulchahey et al.,2004; Karim et al., 2006). hMT1 melatonin receptorsexpressed in heterologous mammalian cells show noobservable changes in melatonin-receptor density, af-finity, or functional sensitivity after exposure to phys-iological concentrations of melatonin for a period oftime that mimics normal nocturnal exposure (i.e., 8 h)(Gerdin et al., 2004b). By contrast, exposure to supra-physiological concentrations of melatonin (100 nM)increases MT1 receptor density and decreases receptoraffinity, but there is no detectable internalization orloss of MT1 melatonin membrane receptors in CHOcells (MacKenzie et al., 2002; Gerdin et al., 2003,2004b). In contrast, rapid arrestin-dependent inter-nalization of the MT1 melatonin receptor was demon-strated in GT1–7 neurons after short-term exposure tomelatonin (Roy et al., 2001). The GT1–7 cells express
358 DUBOCOVICH ET AL.
low levels of endogenous MT1 melatonin receptors andthus the presence of endogenous signaling partnersdifferent from those found in CHO and human embry-onic kidney 293 cells and/or low level of constitutivelyactive MT1 receptors may have facilitated MT1 mela-tonin receptor internalization (Dubocovich andMasana, 1998; Roka et al., 1999; Kokkola et al., 2007).Exposure to melatonin functionally desensitizes MT1-mediated inhibition of cAMP production (Hazlerigg etal., 1993; Witt-Enderby et al., 1998; Jones et al., 2000)and stimulation of PI hydrolysis (MacKenzie et al.,2002). At high concentrations, melatonin decreasescell proliferation and transformation via activation ofeither hMT1 or hMT2 receptors expressed in NIH-3T3cells (Jones et al., 2000). Long-t