The study of mammalian pigmentation has been fueled by the abundance of mouse coat color mutants, the ease with which they are recognized, and the insight they provide into diverse biological and physiological pathways. Significant progress has been made in piecing together the genetic and molecular puzzle that is mammalian pigmentation, with the basic framework largely defined but the location of many of the ‘inner pieces’ yet to be determined. One also wonders how many pieces might be left in the box – genes not yet identified. Mice and many other mammals produce black hairs with a subapical yellow band. The switch in type of pigment synthesized by melanocytes from black/brown eumelanin to yellow/red pheomelanin involves signaling through the melanocortin-1 receptor (MC1R), a G-protein-coupled receptor (GPCR) (reviewed by Barsh, 2006 and depicted in Figure 1). Its primary agonist, alpha melanocyte stimulating hormone (α-MSH), promotes signaling through the MC1R and adenylyl cyclase to effect eumelanin production. Transient expression of an inverse agonist, agouti signaling protein (ASIP), during the mid portion of the hair growth cycle reduces signaling below constitutive levels and leads to the production of pheomelanin. Recently, a third ligand was added to this list when it was shown that dominant black coat color in dogs is caused by a mutation in β-defensin, which appears to competitively inhibit binding of ASIP to the MC1R (Candille et al., 2007). Further pieces of the pigment-type switching puzzle have been discovered through the study of mice with mutations that affect eumelanin/pheomelanin production, but exactly how all the pieces fit together has remained a matter of speculation. Recent work by Pérez-Oliva et al. provides the first clue as to where one piece, Mahogunin Ring Finger-1 (MGRN1), fits into the bigger picture. Regulation of MC1R signaling. Left panel: model of the active state of MC1R. Binding of α-MSH results in a conformational change in MC1R that stimulates exchange of GDP for GTP by receptor-bound Gαs. Active, GTP-bound Gαs dissociates from Gβγ and stimulates cAMP production by adenylyl cyclase. Note that mouse MC1R exhibits constitutive signaling even in the absence of α-MSH. Right panel: proposed model of the inactive state of MC1R. ASIP binds MC1R and ATRN, blocking α-MSH. Binding of ASIP to ATRN may lead to the association of the ATRN cytoplasmic tail with MC1R, which may promote and/or stabilize the association of MGRN1 with MC1R. Interaction between MGRN1 and MC1R would exclude Gαs from associating with MC1R, preventing GDP–GTP exchange. ASIP binding would thus prevent both constitutive and α-MSH-induced MC1R signaling, resulting in reduction of cAMP below basal levels. β-defensin is shown as a competitive inhibitor of ASIP. Mahogunin Ring Finger-1 is an E3 ubiquitin ligase required for ASIP-mediated production of pheomelanin. Alternative splicing of Mgrn1 produces four isoforms that differ by the presence or absence of a nuclear localization sequence (NLS) and in their carboxy terminal sequences. Mice homozygous for a null allele are black, and genetic studies place Mgrn1 downstream of ASIP and upstream of Mc1r (reviewed by Barsh, 2006). The mechanism by which MGRN1 modulates MC1R signaling has remained an open question. As a ubiquitin ligase, MGRN1 is expected to catalyze the ubiquitination of target molecules, which suggested it might ubiquitinate MC1R to cause increased receptor internalization and turnover. Pérez-Oliva et al. show that this is not the case. They demonstrated that while MGRN1 physically interacted with the MC1R in HEK293T cells, the receptor was not ubiquitinated and MGRN1 overexpression did not increase MC1R internalization or turnover. Instead, MGRN1 overexpression reduced adenylyl cyclase-coupled signaling through MC1R (and its family member, MC4R). The authors dissected this phenomenon to show that the effect of MGRN1 on MC1R could be rescued by overexpression of the stimulatory G protein α subunit, Gαs. MGRN1 and Gαs appeared to bind MC1R in a competitive fashion, leading the authors to propose a model wherein MGRN1 binds MC1R/MC4R to competitively exclude Gαs from binding, inhibiting functional coupling to adenylylcyclase. As this would represent a novel regulatory mechanism for GPCR signaling, it will be of great interest to determine how widely applicable the concept may be: does MGRN1 overexpression have a similar effect on signaling through all five members of the mammalian melanocortin receptor family, and perhaps through even more distantly related receptors? While the model presented by Pérez-Oliva et al. is intriguing, it is not readily apparent how it fits into the existing framework of the pigment-type switching puzzle. Their model predicts that loss of MGRN1 would cause a general increase in MC1R signaling throughout the hair cycle rather than the temporally restricted, ASIP-dependent effect observed in Mgrn1 null mutant mice. Understanding how ASIP and MGRN1 cooperate to modulate MC1R signaling will require studies of ASIP-treated wild-type and Mgrn1 null melanocytes. Studies in melanocytes will also address the unavoidable caveats that arise from studying the consequences of overexpressing proteins in heterologous cell types. While Pérez-Oliva et al. deftly dissected the mechanism by which overexpression of MGRN1 affected MC1R signaling in HEK293T cells (which normally express MGRN1 at high levels), two lines of evidence suggest that MGRN1 is unlikely to have a similar effect on MC1R in its normal physiological context. First, Mgrn1 null mutant melanocytes had a normal response to NDP-MSH and ASIP with respect to cAMP levels (Hida et al., 2009). Second, transgenic mice individually expressing each of the Mgrn1 isoforms failed to demonstrate an inhibitory effect of MGRN1 overexpression on pigment-type switching: while not every isoform rescued the pigmentation phenotype of Mgrn1 null mutant mice, none caused an obvious darkening of fur on a wild-type background (Jiao et al., 2009). Regardless of whether competitive inhibition of Gαs binding is critical for the function of MGRN1 in vivo, the data of Pérez-Oliva et al. may point to a previously unsuspected connection between MGRN1 and the Gαs binding state of MC1R, with MGRN1 perhaps serving to stabilize an inactive conformation of MC1R (thus appearing to compete with Gαs for a binding site) or to sense its Gαs binding state. Mgrn1 mutant mice have an intriguing array of phenotypes in addition to hyperpigmentation that includes curly hair, embryonic patterning defects, mitochondrial dysfunction and spongiform encephalopathy. This suggests that MGRN1 has additional molecular partners, and it is unlikely that they are all melanocortin receptors, or even GPCRs. Pérez-Oliva et al. make the intriguing observation that although MC1R was not ubiquitinated, it coimmunoprecipitated with an unknown ubiquitinated protein. The presence of additional MGRN1 seemed to promote this interaction, suggesting that MGRN1, MC1R, and the unknown ubiquitinated protein may interact in a ternary complex. The mystery molecule is an obvious candidate for MGRN1-mediated ubiquitination. It is tempting to propose attractin (ATRN) for this role, given the phenotypic overlap between Mgrn1 and Atrn mutant mice (both exhibit spongiform neurodegeneration, mitochondrial dysfunction and hyperpigmentation). This possibility is supported by the observations that an ATRN ortholog, attractin-like-1, interacts with the MC4R and can compensate for loss of ATRN when overexpressed. The formation of a ternary complex between ATRN, ASIP and MC1R could recruit MGRN1 and/or stabilize its association with MC1R to prevent binding of the G protein heterotrimer, as depicted in our proposed model (Figure 1). An additional puzzle is whether the four alternative isoforms of MGRN1 have different targets and physiological roles. While Pérez-Oliva et al. observed a similar effect of each isoform on cAMP signaling, they report an interesting behavior of the two isoforms that contain a NLS. All four isoforms were excluded from the nucleus when transfected into HEK293T cells alone, but the isoforms containing the NLS localized to the nucleus in HEK293T cells coexpressing the MC1R and in MC1R-expressing human melanoma cells. This suggests that nuclear localization of MGRN1 is a regulated event that can be triggered by the presence of the MC1R, perhaps providing a novel mechanism of GPCR-to-nucleus signaling. This concept does not necessarily mesh with the authors’ model of MC1R-adenylyl cyclase decoupling by MGRN1, however, as it predicts that inhibition of MC1R by competitive binding of MGRN1 should result in elevated MC1R signaling in cells expressing NLS-containing isoforms of MGRN1 yet all isoforms had similar effects on cAMP levels. Furthermore, our laboratory has observed nuclear localization of a green fluorescent protein fusion protein for one of the NLS-containing isoforms in HEK293T and Neuro2a cells without coexpressing MC1R or any other protein (Jiao et al., 2009 and TMG unpublished results), suggesting that the nuclear import of MGRN1 does not specifically require the presence of the MC1R. It will be interesting to determine whether nuclear import is affected by cAMP levels or some other consequence of MC1R expression. The observation of equivalent effects of all four MGRN1 isoforms on MC1R signaling is also at odds with the fact that transgenes expressing NLS-containing MGRN1 isoforms were much less effective at rescuing the phenotypes of Mgrn1 null mutant mice, including hyperpigmentation (Jiao et al., 2009). The transgenic studies suggest that the NLS-containing MGRN1 isoforms do not play a significant role in pigment-type switching and that their absence does not have overt physiologic consequences. Mahogunin Ring Finger-1 is an interesting puzzle piece as it is appears to fit into several different biological frameworks, including pigment-type switching, neurodegeneration and embryonic patterning. While the Endosomal Sorting Complex Required for Trafficking protein TSG101 was recently shown to be ubiquitinated by MGRN1, it remains unclear whether loss of this interaction underlies any of the phenotypes of Mgrn1 mutant mice. Further studies are needed to determine the relationship, if any, between the role of MGRN1 in endosomal trafficking and its effect on MC1R signaling. Identification of additional MGRN1 interacting partners should reveal more pieces of the pigment-type switching pathway puzzle and provide insight into how all the pieces fit together. Understanding the cellular function of MGRN1 will have implications beyond pigmentation. In fact, a recent article by Chakrabarti and Hegde (2009) demonstrated that aggregated prion protein (PrP) could sequester MGRN1, suggesting that loss of MGRN1 function may underlie spongiform neurodegeneration not only in Mgrn1 null mutant mice but also in prion-related neurodegenerative disorders. It will be interesting to discover what picture the puzzle reveals when MGRN1 is finally put into its place.