Báo cáo Y học: Loss-of-function variants of the human melanocortin-1 receptor gene in melanoma cells define structural determinants of receptor function doc

Loss-of-function variants of the human melanocortin-1 receptor gene in melanoma cells define structural determinants of receptor function Jesu ´ sSa ´ nchez Ma ´ s 1 , Concepcio ´ n Olivares Sa ´ nchez 1 , Ghanem Ghanem 2 , John Haycock 3 , Jose ´ Antonio Lozano Teruel 1 , Jose ´ Carlos Garcı ´ a-Borro ´ n 1 and Celia Jime ´ nez-Cervantes 1 1 Department of Biochemistry and Molecular Biology, School of Medicine, University of Murcia, Spain; 2 LOCE, Free University of Brussels, Brussels, Belgium; 3 Department of Engineering Materials, University of Sheffield, Sheffield, UK The a-melanocyte-stimulating hormone (aMSH) receptor (MC1R) is a major determinant of mammalian skin and hair pigmentation. Binding of aMSHtoMC1Rinhumanmel- anocytes stimulates cell proliferation and synthesis of pho- toprotective eumelanin pigments. Certain MC1R alleles have been associated with increased risk of melanoma. This can be theoretically considered on two grounds. First, gain- of-function mutations may stimulate proliferation, thus promoting dysplastic lesions. Second, and opposite, loss-of- function mutations may decrease eumelanin contents, and impair protection against the carcinogenic effects of UV light, thus predisposing to skin cancers. To test these possi- bilities, we sequenced the MC1R gene from seven human melanoma cell (HMC) lines and three giant congenital nevus cell (GCNC) cultures. Four HMC lines and two GCNC cultures contained MC1R allelic variants. These were the known loss-of-function Arg142His and Arg151Cys alleles and a new variant, Leu93Arg. Moreover, impaired response to a superpotent aMSH analog was demonstrated for the cell line carrying the Leu93Arg allele and for a HMC line homozygous for wild-type MC1R. Functional analysis in heterologous cells stably or transiently expressing this vari- ant demonstrated that Leu93Arg is a loss-of-function mutation abolishing agonist binding. These results, together with site-directed mutagenesis of the vicinal Glu94, demon- strate that the MC1R second transmembrane fragment is critical for agonist binding and maintenance of a resting conformation, whereas the second intracellular loop is essential for coupling to the cAMP system. Therefore, loss- of-function, but not activating MC1R mutations are com- mon in HMC. Their study provides important clues to understand MC1R structure-function relationships. Keywords: melanocortin 1 receptor; melanoma; loss-of- function mutations; functional coupling; structure-function relationships. G protein-coupled receptors (GPCRs) constitute the largest family of cell surface receptors involved in signal transduc- tion, with over 1% of the human genome encoding for more than 1000 proteins of this type [1]. The structural hallmarks of GPCRs are an heptahelical transmembrane structure, with an N-terminal extension of variable length facing the extracellular side of the cell membrane, and an intracellular C-terminus, which is often post-translationally modified by acylation of conserved Cys residues [2]. Recent evidence supports a role of GPCRs in the control of normal and aberrant cell growth. Indeed, many potent mitogens stimu- late cell proliferation upon binding to their cognate GPCRs [3–5], and the mas oncogene belongs to this superfamily [6]. Moreover, activating mutations of members of the GPCR family cause several dysplatic syndromes such as hyper- functioning thyroid adenoma [7], familial male precocious puberty [8], and Jansen-type metaphyseal chondrodysplasia [9]. The melanocortin 1 receptor (MC1R) is a GPCR involved in the regulation of key aspects of mammalian skin biology, which belongs to a subfamily of the GPCRs comprising five members designated MC1 to MC5 [10]. The genes encoding for the mouse and human MC1R were cloned in 1992 [11]. The human MC1R gene maps to chromosome 16q2 and contains an open reading frame of 954 base pairs corresponding to a 317 amino acids protein. It is highly polymorphic [12], and specific variants have been found associated with red hair and fair skin [13–15]. The preferential natural agonists of MC1R are aMSH and adrenocorticotropic hormone. Both proopiomelano- cortin-derived peptides bind to the MC1R with the same Correspondence to C. Jime ´ nez-Cervantes, Department of Biochemistry and Molecular Biology, School of Medicine, University of Murcia, Apto 4021. Campus de Espinardo, 30100 Murcia, Spain. Fax: + 34 68 830950, Tel.: + 34 68 364676, E-mail: celiajim@um.es Abbreviations: FSK, forskolin; GPCR, G protein coupled receptor; GCNC, giant congenital nevus cells; HMC, human melanoma cells; MC1R, human melanocortin 1 receptor; mc1r, mouse melanocortin 1 receptor; aMSH, a-melanocyte stimulating hormone; NDP-MSH, [Nle4,D-Phe7]-a-melanocyte stimulating hormone; TM, transmembrane. Note: The research group web site is available at http://www.um.es/ bbmbi/melanocitos.htm Note: Nucleotide sequence data are available at the DDBJ/EMBL/ GenBank databases under the accession numbers AF326275 (con- sensus wild-type MC1R sequence), and AF529884 (Leu93Arg MC1R variant). The SWISS-PROT entry for human MC1R is Q01726. (Received 20 August 2002, revised 14 October 2002, accepted 23 October 2002) Eur. J. Biochem. 269, 6133–6141 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03329.x affinity and, upon binding to the receptor, trigger the cAMP cascade, thus activating a variety of intracellular signaling pathways [16]. As a result of MC1R activation, the activity of the rate-limiting enzyme in melanin synthe- sis, tyrosinase, is increased and skin pigmentation is promoted. Melanin pigments are complex heteropolymers, which can be classified into two main groups: the brown to black eumelanins, and the reddish to yellow, sulfur- containing phaeomelanins [17]. Although both types of melanins are found in different relative proportions in all skin types, the phaeomelanin/eumelanin ratio is higher in individuals with red hair and fair skin [18]. Activation of MC1R by its agonists promotes the switch from phaeo- to eumelanogenesis, and increases the eumelanin/phaeomela- nin ratio [10]. Eumelanins are photoprotective pigments, and UV irradiation is considered the main ethiologic factor for skin cancers. Thus, it can be hypothesized that an impairment of the normal function of the aMSH/MC1R system as a result of loss-of-function mutations in the MC1R gene, might lead to an increased risk of melanoma. In keeping with this view, an association of several MC1R allelic variants with increased melanoma and nonmel- anoma skin cancer risk has been demonstrated [19–21]. These variants were subsequently shown to be loss-of- function mutations [22]. In addition to its effect on melanocyte differentiation, it is well established that aMSH stimulates normal melanocyte proliferation [23]. Because of this mitogenic activity, it could also be thought that hyperactivity of the aMSH/MC1R system might promote abnormal melanocyte growth, thus contributing to premalignant or malignant phenotypes. In agreement with this, growth of normal melanocytes and nevus cells in culture, but not of melanoma cells, is dependent on aMSH and/or other stimulators of the cAMP cascade [24]. Most studies on the possible involvement of MC1R gene variants in melanoma reported thus far are case- control studies where MC1R was sequenced from peripheral blood of melanoma patients or healthy indi- viduals. These studies have established an increased risk of melanoma associated with MC1R loss-of-function variants leading to light skin phototypes, and hence to increased UV sensitivity. However, they might have failed to detect somatic gain-of-function mutations of the MC1R gene within melanocytes, leading to increased, aMSH-independent proliferation. In order to establish whether or not activating mutations are present in malignant or premalignant melanocytes, we have sequenced the entire open reading frame of the MC1R gene from cultured HMC and GCNC. The functional properties of the allelic variants found have been analyzed. Moreover, we have compared the aMSH- triggered responses in heterologous cells expressing the receptor variants under study, and the ones of malignant melanocytes of defined MC1R genotype. Our results show that loss-of-function, but not gain-of-function, mutations of MC1R are indeed frequent in HMC lines, and that impairment of signaling through the cAMP cascade occurs in these cells even in the absence of mutations in the coding region of the gene. Finally, our study of the natural MC1R variants highlights several aspects of the structure–function relationships in the MC1R protein, such as the critical importance of the TM2 fragment and the second cytosolic loop, and can target the design of site-directed mutagenesis studies. MATERIALS AND METHODS Cell culture Seven HMC lines were used. HBL and LND1 cell lines were established at the LOCE, Brussels, Belgium. DOR, IC8 and T1C3 cells were a gift from J. F. Dore ´ , INSERM, Lyon, France. The IC8 and T1C3 cells are clones of the same melanoma line and were isolated based on their different metastatic potential. A375-SM cells were a gift from I. J. Fidler (University of Texas M. D., Houston, TX, USA), and C8161 a gift from Prof Meyskens (University of California, Irvine, CA, USA). All HMC lines were cultured as previously described [25] in HAM-F10 medium supple- mented with 5% fetal bovine serum, 5% newborn calf serum, 2 m M glutamine, 100 UÆmL )1 penicillin, 0.1 mgÆmL )1 streptomycin sulfate and 0.1 mgÆmL )1 kana- mycin sulfate (all from Gibco, Paisley, UK). Three primary cultures of GCNC were also analyzed. Isolation and culture of GCNC was performed as described [26]. Briefly, the tissue was incubated overnight in HAM-F10 medium containing 0.6 UÆmL )1 dispase and 0.05 UÆmL )1 collage- nase, supplemented with 200 UÆmL )1 penicillin G, 0.2 mgÆmL )1 streptomycin sulfate, 0.2 mgÆmL )1 kanamycin sulfate and 25 lgÆmL )1 gentamycin. The detached cells were washed, seeded in culture flasks at the density of 10 5 cellsÆmL )1 andincubatedfor48hwith0.1mgÆmL )1 geneticin in the culture medium described, supplemented with 2% Ultroser-G, 16 n M phorbol 12-myristate 13- acetate, and 0.1 m M isobutyl methylxanthine. GCNC were cultured for a maximum of 15 passages. DNA extraction, amplification of the MC1R gene and sequencing Genomic DNA was extracted from cultured cells with the Wizard kit (Promega, Madison, WI). The complete MC1R coding sequence was amplified by PCR, with the forward primer CCT AAGCTTACTCCTTCCTGCTTCCTGG ACA (called universal MC1R forward primer), and the reverse primer CTG GAATTCACACTTAAAGCGCG TGCACCGC. These primers match nucleotides 435–456 and 1418–1439, respectively, according to the sequence reported in [11], and yield a 1023-bp fragment that can be cloned by means of added HindIII and EcoRI restriction sites (underlined). Thirty PCR rounds (denaturation for 1 min at 95 °C, annealing for 2 min at 68 °C and extension for 3 min at 72 °C, followed by a final 10 min extension at 72 °C) were performed using 1 lg of genomic DNA, 0.5 lg of each primer, 200 l M each dNTP and 2.5 U of the proofreading Pfu polymerase. The PCR products were purified by agarose gel electrophoresis and completely sequenced in both strands using internal primers, at the DNA sequencing facility of the ÔCentro de Investigaciones Biolo ´ gicasÕ, Madrid, Spain. When necessary for functional studies, the amplification product was cloned in the expression vector pcDNA3 (Invitrogen, Carlsbad, CA, USA), using the added restriction sites, following standard procedures [27]. The identity of the clones was ascertained by complete resequencing of the cloned inserts. 6134 J. Sa ´ nchez Ma ´ s et al.(Eur. J. Biochem. 269) Ó FEBS 2002 Transient expression of MC1R HEK 293T cells were grown in six-well plates with RPMI 1640, 10% fetal bovine serum, 100 UÆmL )1 penicillin and 0.1 mgÆmL )1 streptomycin. Transfection was carried out with the Superfect reagent (Qiagen, Paisley, UK), as per instructions, with 1.5 lg plasmid DNA per well. Three hours after adding the transfection mix, the medium was removed. Cells were gently washed with 500 lLNaCl/P i and 1 mL of fresh medium was added. After 24 h, cells were completely serum-deprived for an additional 24 h before binding or coupling assays. Radioligand binding assay Transfected cells were incubated (1 h, 37 °C) with increas- ing concentrations of [Nle4, D -Phe7]-aMSH (NDP-MSH, Sigma, Saint Louis, MO, USA), ranging from 10 )12 to 10 )6 M and a fixed amount of 125 I-labelled-NDP-MSH (Amersham Pharmacia Biotech, Little Chalfont, Bucking- hamshire, UK), corresponding to 10 )10 M and 0.1 lCi per well, in a final volume of 500 lL RPMI. Cells were then washed twice with RPMI for 5 min at room temperature, and trypsinized. The cell suspension was pipetted into plastic tubes and the associated radioactivity was measured. Non-specific binding was estimated from the radioactivity bound in the presence of 10 )6 M NDP-MSH or to cells transfected with empty vector and incubated with 125 I- labelled-NDP-MSH alone, with similar results. Determination of agonist-induced cAMP increases Cells were serum-deprived for 24 h and then incubated with 10 -13 )10 -7 M NDP-MSH, for 20 min. The medium was aspirated, cells quickly washed with 1 mL ice-cold NaCl/P i ,lysedwith350lLpreheated0.1 M HCl (70 °C), and carefully scrapped. The resulting mix was freeze-dried for 90 min, washed with 100 lLH 2 O, and freeze-dried for another 20 min. Dried samples were dissolved in suitable volumes of 50 m M Tris, 4 m M EDTA, pH 7.5, from 65 to 300 lL, depending on the expected cAMP content. cAMP was measured by radioimmunoassay (Amersham Phar- macia Biotech), as per instructions. Parallel dishes for protein determination were included. Cells were dissolved in 10 m M phosphate buffer pH 7, 1% Igepal CA-640, containing 0.1 m M EDTA, and 0.1 m M phenyl- methanesulfonyl fluoride. The protein concentration in these lysates was determined by the bicinchoninic acid method. Construction and cloning of a MC1R tagged-sequence The wild-type and Leu93Arg variant sequences of MC1R were tagged by adding a (HA) 2 His 6 tag to the C-terminus of the protein. This tag corresponds to the sequence RFYPYDVPDYAGYPYDVPDYAHHHHHH, placed immediately after the C-terminal W317 residue of MC1R. It contains two consecutive influenza virus hemagglutinin epitopes (HA) and a terminal hexahistidine sequence (His6). In order to prepare the constructs, we amplified the tag sequence from the pREP1-cdc2Ha6His vector (a kind gift from Prof Jose ´ Cansado, Department of Genetics and Microbiology, University of Murcia, Spain) by using the primers CGC GAATTCTACCCA TACGAT (forward) and AGC TCTAGATTAGTGGT GATG (reverse), containing EcoRI and XbaIsites (underlined). We also amplified the complete coding sequence of the wild-type and Leu93Arg variant of MC1R with the universal MC1R forward primer and the GCG TCTAGATCAGAATTCCCAGGAGCACGT CAG reverse primer. This primer contains an EcoRI restriction site before a stop codon (shown in bold) and a XbaI site immediately after it. PCR amplification reactions were performed with 30 ng of template DNA, 0.5 lgof each corresponding primer, 200 l M each dNTP and 2.5 U of the proofreading Pfu polymerase. Thirty rounds of amplification were carried out (denaturation 1 min at 95 °C, annealing 2 min at 48 °C and extension for 3 min at 72 °C). The MC1R amplicons were cloned into pcDNA3 by means of the HindIII and XbaIrestriction sites. The tag sequence was then cloned into the MC1R- pcDNA3 constructs, with the added EcoRI and XbaI sites. The identity of the products was verified by complete sequencing. Construction of the Glu94Lys MC1R variant The Glu94Lys variant was generated in two PCR steps. First, we used the universal MC1R forward primer and the reverse primer GACGGCCGTCTTCAGCACGTTGCT, which introduces the desired amino acid change (variant base shown in bold), to yield a 328-bp fragment. This ampliconwasthenusedasaforwardprimerwiththerev- erseCTG GAATTCACACTTAAAGCGCGTGCACCGC (stop codon in bold, EcoRI site underlined). The resulting cDNA was cloned into pcDNA3. The sequence of this artificial variant was ascertained by double strand DNA sequencing. Western blot and related procedures Cells were solubilized in 50 m M Tris/HCl, pH 8.8, 1 m M EDTA, 1% Igepal CA-630 and 10 m M iodoacetamide, and then centrifuged at 105 000 g,for30mininatable- top Beckman TL-100 ultracentrifuge. Electrophoresis was performed in 10% acrylamide gels, under reducing conditions. For immunochemical detection of the tagged MC1R constructs in extracts from transfected cells, 10 lg of protein from each sample were mixed in a 2 : 1 ratio with sample buffer (0.18 M Tris/HCl, pH 6.8, 15% glycerol, 0.075% bromophenol blue, 9% SDS, and 3 M 2 mercaptoethanol). For deglycosylation studies, the extracts were heated at 95 °C for 5 min prior to incuba- tion at 37 °C for 4 h in the presence of 5 U of N-glycosidase F (from Roche, Mannheim, Germany) in a50-m M phosphate buffer, pH 7.0, containing 10 m M EDTA and 0.1% SDS. Control samples were treated under identical conditions, except for the omission of the glycosidase. These samples were processed as previously described for electrophoresis. Gels were transferred to poly(vinylidene difluoride) (PVDF) membranes, blocked with 5% nonfat dry milk and incubated overnight at 4 °C with a monoclonal anti-HA Ig (Sigma) as per instructions. Staining of immunoreactive bands was performed with a chemiluminescent substrate (Amersham Pharmacia Bio- tech), after incubation with a peroxidase-labeled secondary Ó FEBS 2002 Loss-of-function MC1R variants in human melanoma (Eur. J. Biochem. 269) 6135 antibody. Comparable loading and transfer were ascer- tained by cutting the lower portion of the membrane before blocking and staining for total protein with Amido Black. RESULTS AND DISCUSSION Seven HMC lines and three GCNC cultures were selected for genotyping and functional studies, based mostly on previous observations of high MC1R gene expression at the mRNA level [28]. Sequencing of the complete coding region of MC1R after amplification of genomic DNA with suitable primers [29,30] showed that four HMC, and two GCNC cultures harbored MC1R variants (Table 1). The variants found were the well-known Arg151Cys and Arg142His alleles and a new variant not reported so far, to the best of our knowledge, the Leu93Arg allele. That this variant was indeed present in the sample and was not the result of a PCR artifact was ascertained by repeated amplification and sequencing reactions, and by the observation that the same allele was found in the IC8 and T1C3 cell lines, which both come from the same parental cell line and correspond to two clones isolated on the basis of their different metastatic potential (J. F. Dore ´ , personal communication). The posi- tions within the MC1R molecule of the substitutions found in our study are shown in Fig. 1, where several other functionally relevant residues are also highlighted. The Arg151Cys and Arg142His alleles are two well- documented loss-of-function MC1R variants [22,34]. The Arg151Cys allele is strongly associated with red hair [14,15,22,35], and its functional impairment might be related to deficient coupling to the cAMP signaling pathway, as it has been reported that the binding properties of the mutant receptor are very similar, if not identical, to wild-type [22,34]. The Arg142His substitution has a similar functional effect, with conserved binding properties but strongly impaired coupling to the cAMP generation system [22]. Interestingly, both substitutions are located in the second intracellular loop of the receptor protein, thus highlighting the contribution of this region to efficient coupling to G S . Moreover, Arg142 is part of a DRY sequence conserved in most if not all GPCRs, and particularly in all members of the melanocortin receptors subfamily. WenexttriedtocorrelatetheMC1R genotype of several of the HMC lines with their degree of responsiveness to agonists, by studying their cAMP levels after treatment with the superpotent melanocortin analog NDP-MSH (Fig. 2). LND1 and HBL cells, homozygous for the wild-type allele, both responded with strong and concentration-dependent increases in cAMP, although the maximal levels of the second messenger were higher for the former. Surprisingly, both DOR cells, homozygous for wild-type MC1R,andIC8 cells, carrying one variant Leu93Arg allele, failed to increase their cAMP intracellular levels in response to the superpotent agonist. That this failure is not due to a defective adenylyl cyclase was shown by a strong stimulation of the enzyme activity in cells treated with the specific activator forskolin (Fig. 2, B). Conversely, DOR and IC8 cells also failed to respond to the natural agonists aMSH and adrenocortico- tropic hormone, employed at a saturating concentration of 10 )7 M (not shown). It is worth noting that the unrespon- siveness of DOR cells is also unrelated to a lack of MC1R gene expression, as we have previously found that the mRNA levels for MC1R are comparable in DOR, LND1 and HBL cells, and higher in these HMC than in normal human melanocyte cultures [28]. Moreover, both DOR and IC8 cells display a significant number of specific binding sites, in the range of 2500 ± 500 sites per cell. This is approxi- mately 50% as compared to the responsive HBL cells. The lack of functional coupling of these cell lines, particularly of DOR cells homozygous for the wild-type MC1R, is therefore perplexing and will be the subject of further studies. On the other hand, as opposed to the well-known Arg142His and Arg151Cys variants, no functional data are available for the new Leu93Arg allele. Although the unresponsiveness of the IC8 HMC line suggested that it may correspond to a loss-of-function mutation, no clear conclusion could be drawn at this stage, as DOR cells, homozygous for wild-type MC1R, were also defective in signaling through cAMP. Therefore, we performed a func- tional analysis of the Leu93Arg MC1R. For comparison, Table 1. Occurrence of MC1R variant alleles in human melanoma and GCN cells. Cell line Genotype a Variant allele A375-SM 3 Arg151Cys C8161 1 Arg151Cys DOR 0 – HBL 0 – IC8 b 1 Leu93Arg LND1 0 – T1C3 b 1 Leu93Arg GCN1 1 Arg151Cys GCN2 0 – GCN3 3 Arg142His a 0, homozygous wild-type; 1, variant heterozygous; 2, variant compound heterozygous; 3, variant homozygous. b The IC8 and T1C3 cells lines derive from the same patient, and correspond to two different clones selected on the basis of their different meta- static potential. Fig. 1. Structure of MC1R. The topology of the TM fragments is depicted according to the model proposed in [31], and the length of the signal peptide (hatched first 22 amino-terminal residues) was estimated according to [32]. The allelic variants found in this study, as well as other natural loss-of-function mutations described in the literature (reviewed in [33]) are specified (gray circles) and highlighted by an arrow. 6136 J. Sa ´ nchez Ma ´ s et al.(Eur. J. Biochem. 269) Ó FEBS 2002 the Arg151Cys receptor was also included in the study. Figure 3 shows the coupling properties of the wild-type and mutant receptors, as analyzed in clones of stably transfected CHO cells. As expected, Arg151Cys clones failed to respond to saturating concentrations of NDP-MSH with increases in their intracellular cAMP levels, thus confirming results by others [22,34]. A complete lack of functional coupling was also observed for the Leu93Arg receptor. Next, the binding properties of both variants were compared to wild-type, in order to determine whether lack of agonist-induced cAMP generation was related to impaired binding or coupling properties. Clones stably transfected with the Leu93Arg form were unable to specifically bind significant levels of radiolabeled agonist (Fig. 4). However, and as previously reported by others [22], the affinity for NDP-MSH was identical, within experimental error, for the wild-type and the Arg151Cys receptors (not shown). This result was in keeping with the lack of hormone- induced cAMP generation, and suggested that the func- tional impairment was mostly explained by the inability of the mutant receptor to bind melanocortins. However, the possibility still existed that the unresponsiveness of the clones stably transfected with the Leu93Arg variant could be related to other causes, such as low expression or aberrant processing of receptor molecules. To check these possibilities, we performed transient expression experiments using HEK 293T cells. Previous studies have shown that, in Fig. 2. Functional coupling of several HMC lines of defined MC1R genotype. (A) Dose–response curves for intracellular cAMP levels in different HMC lines treated with increasing concentrations of NDP- MSH. Results are shown as mean ± SD (n ‡ 3). (B) Levels of cAMP in human melanoma cell lines treated with a saturating concentration of NDP-MSH (10 )7 M ), or with the adenylyl cyclase stimulator forskolin (FSK, 10 )5 M ), for 20 min. Results are the mean ± SD of two independent experiments, each performed in triplicate dishes. Fig. 3. Lack of functional coupling of the Arg151Cys and Leu93Arg MC1R in stably transfected CHO cells. CHO cells from clones stably transfected with the variants indicated, and selected on the basis of high MC1R expression at the mRNA level, were seeded in six-well plates, grown to semiconfluence and serum-deprived for 24 h. Then, the cells were challenged with vehicle or with 100 n M NDP-MSH for 20 min, before cell lysis and determination of the cAMP contents in the lysates. For each clone, parallel dishes were solubilized for protein determination. Results are the mean ± SD of two independent experiments, each performed in triplicate dishes (n ¼ 6). Similar results were obtained with two independent clones for each variant. Fig. 4. Inability of the Leu93Arg MC1R variant to bind NDP-MSH. Clones of CHO cells stably transfected with the variant or wild-type receptor were seeded in six-well plates, and incubated with a fixed amount of 125 I-labelled-NDP-MSH in 500 lLofserum-freemedium in the presence of increasing concentrations of unlabeled agonist. After washing and harvesting as described in Materials and methods, the radioactivity associated with the cell pellets was counted. Results are the mean ± SD of two independent experiments, each performed in triplicate dishes (n ¼ 6). Ó FEBS 2002 Loss-of-function MC1R variants in human melanoma (Eur. J. Biochem. 269) 6137 this system, the expression of MC1R binding sites is high, with B max values ranging from 2 to 8 pmolesÆmg protein )1 , depending on the construct under study [30]. Figure 5 shows that the Leu93Arg mutant was also unable to mediate cAMP increases (A) or bind significant levels of radiolabeled agonist (B), under conditions of strong over- expression in transiently transfected HEK 293T cells. Under identical conditions, the wild-type construct, which was expressed at a density of 2.2 ± 0.25 pmoles binding sites per mg protein, was highly efficient in eliciting a cAMP response in cells challenged with NDP-MSH. Moreover, MC1R mRNA levels for the wild-type and Leu93Arg alleles were very similar in transiently transfected cells (C). This was expected, as the two constructs are identical except for the single base substitution in codon 93 determining the Arg to Leu mutation. On the other hand, in order to detect Fig. 5. The Leu93Arg MC1R variant is adequately expressed in transiently transfected HEK 293T cells, but fails to bind agonists or elicit agonist- induced cAMP increases. (A) Functional coupling to cAMP production in cells transfected with empty vector (pcDNA), wild-type receptor or the Leu93Arg variant. Cells were serum deprived 24 h before the measurement of cAMP under basal conditions (empty bars), or after stimulation with NDP-MSH (10 )7 M , closed bars). Results are the mean ± SD of quadruplicate independent dishes. (B) Equilibrium binding of 125 I-labelled-NDP- MSH to the transiently expressed wild-type and Leu93Arg variants. Transiently transfected cells seeded in six-well dishes were incubated for 1 h in thepresenceof10 )10 M 125 I-labelled-NDP-MSH, corresponding to 0.1 lCi per well. Data are given as specifically bound c.p.m., after subtraction of the radioactivity bound to cells treated under identical conditions but in the presence of 10 )6 M unlabeled NDP-MSH. Results are the mean ± SD of quadruplicate independent dishes. (C) Comparable levels of MC1R mRNA in cells transfected with the wild-type or Leu93Arg constructs. Cells were seeded in 25 cm 2 flasks and transfected with empty vector (–), or the wild-type (WT) or Leu93Arg (L93R) receptors, under conditions comparable to the transfections for functional studies. Total RNA was extracted and analyzed for MC1R mRNA abundance by Northern blot (upper), as previously described [28]. Comparable loading was ascertained by probing the membranes with a GAPDH probe (lower). Specific signals were detected in a GS-525 phosphorimager from Bio-Rad (Hercules, CA). (D) Upper: Western blot detection of epitope-tagged Leu93Arg (Arg93-T) and wild-type (WT-T) receptors. The electrophoretic mobility of molecular weight markers is shown on the left, and the molecular weight of the specific bands is shown on the right (expressed in kDa). The specificity of the bands is clearly demonstrated by their absence in the lane corresponding to cells transfected with empty vector (–). Lower: effect of treatment with N-glycosidase F on the electrophoretic mobility of the tagged receptor protein. Control and glycosidase F-treated (EndoF) extracts from cells transfected with the WT-T construct were analyzed by Western blot using the anti-HA monoclonal antibody. Upon treatment with the glycosidase, the band of higher molecular size was transformed into a band comigrating with the lower molecular weight form. 6138 J. Sa ´ nchez Ma ´ s et al.(Eur. J. Biochem. 269) Ó FEBS 2002 possible effects on protein stability and/or processing, two constructs were prepared, where an epitope-containing tag was included at the C-terminus of the wild-type and Arg93 proteins, so as to enable the detection of the transiently expressed proteins by Western blots. These constructs, termed WT-T and Arg93-T were identical to the ones used in functional studies, except for the presence of an in frame C-terminal extension coding for 27 amino acids, with two HA and an His6 epitope. The protein levels of the tagged Arg93 protein were easily detectable by Western blot. By means of an antihemagglu- tinin antibody, an identical electrophoretic pattern, with two bands corresponding to apparent molecular weights of 34 and 38.7 kDa were detected for the WT-T and Arg93-T receptors (D). The relationship of these bands was studied by deglycosylation of the samples followed by Western blotting. In samples deglycosylated with N-glycosidase F, only the higher mobility band was observed, thus showing that this band does not correspond to a degradation product of the higher molecular weight protein. Rather, the faster band is very likely the de novo form of the tagged MC1R, which, upon glycosylation and post-translational processing, gives raise to the lower mobility band. In any case, the slightly lower levels of the Arg93-T protein, as compared to the wild-type, do not seem to be sufficient to account for the complete loss of functional response in terms of agonist binding or coupling to adenylyl cyclase, specially considering the high overexpression obtained under our experimental conditions. Moreover, the identical electrophoretic pattern strongly suggests that the Leu93Arg mutation has no effect on the processing of the receptor molecule, at least as studied with the tagged forms of the protein. Taken together, these results show that the efficiency of expression of wild-type and Leu93Arg MC1R is similar at the mRNA and protein levels, yet neither agonist binding or functional coupling can be demonstrated for the Leu93Arg mutant. Therefore, the lack of specific binding of melanocortins to the Leu93Arg receptor most likely reflects an actual impair- ment of the receptor binding properties. Consistent with this hypothesis, a pocket located between TM domains 2, 3 and 6 or 7 is probably responsible for the docking of agonists [32,36,37]. Within this pocket, acidic residues Glu94 in TM2 and Asp117 or Asp121 in TM3 have been shown to be critical for ligand–receptor interactions [37]. These negatively charged residues are presumably spatially adjacent and would provide a hydrophilic binding pocket interacting with the positively charged Arg8 of the D -Phe7-Arg8-Trp9 core sequence of NDP-MSH [37]. Within this context, it appears likely that the introduction of a positively charged residue at position 93 might either interfere with the positively charged Arg8 residue of the agonist, or alternatively, counteract a negative charge in the receptor protein important for electrostatic interaction with the agonist. It is worth noting that a constitutively active variant of the mouse Mc1r, the naturally occurring somber E so)3J allele is also associated with a substitution of a charged residue in TM2, namely Glu92Lys, and that Glu92 in Mc1r is equivalent to Glu94 in human MC1R. Interestingly, the constitutively active som- ber Mc1r might also display a reduced binding affinity [38]. Therefore, the functional properties of the naturally occurring Leu93Arg variant, along with site-directed muta- genesis studies reported by others [37], strongly suggest that the TM2 fragment of human MC1R plays a pivotal role in agonist binding, and might also be an important determi- nant of the receptor coupling properties. To further explore this possibility, we generated an artificial construct homo- logous to the mouse somber E so)3J allele, namely Glu94Lys, and analyzed its functional properties in transiently trans- fected HEK 293T cells. The coupling behaviour of the MC1R mutant was highly reminiscent of the one reported by Robbins et al. [38] for the mouse somber allele (Fig. 6A). After transfection in HEK 293T cells, human Glu94Lys MC1R displayed a noticeable agonist-independent consti- tutive activity, as shown by the induction of high levels of cAMP. These levels corresponded to approximately 50% of the maximal stimulations achieved, and were approximately twice as high as those found in cells transfected with the wild-type receptor. Also in keeping with results reported for themouseE so)3J allele, higher stimulations of adenylyl cyclase were only obtained at high agonist doses (Fig. 6A), consistent with a loss of binding affinity observed in radioligand binding assays (Fig. 6B). This reduced affinity for NDP-MSH of the Glu94Arg variant was evident by a Fig. 6. Functional analysis of an artificial MC1R homologue of the mouse somber E so)3J constitutively active allele. (A) Functional coupling to the cAMP production system. (B) Displacement curves of 125 I-labelled-NDP-MSH (10 )10 M in the incubation medium) by the indicated concentrations of unlabeled ligand. The radioactivity bound by the Glu94Lys variant in the absence of competing ligand was approximately four times lower than for the wild-type receptor. Ó FEBS 2002 Loss-of-function MC1R variants in human melanoma (Eur. J. Biochem. 269) 6139 two logs rightwards shift of the displacement curve and by the lower amount of radioactivity bound in the absence of competitor (in spite of the lack of statistically significant differences in the number of binding sites, B max ). In summary, our results show that impaired signaling through the MSH/MC1R pathway is common in HMC lines, consistent with previous reports of increased melan- oma risk in individuals carrying germline mutations of the MC1R gene [19–21], and with the notion that loss-of- function mutations in the MC1R gene sensitize human melanocytes to the DNA damaging action of UV light [39]. Moreover, functional analysis of naturally occurring vari- ants highlights important features of structure-function relationships within the receptor protein, and points to the extracellular side of TM2 as a major determinant of agonist binding. Within this region, results by others have shown that removal of the negative charge at position 94 strongly diminishes agonist binding affinity [37], without yielding a constitutively active receptor. We have shown here that the change of a neutral side chain for a positively charged amino acid at position 93 completely abolishes agonist binding. Moreover, the nonconservative substitution of Glu94 for a positively charged Lys residue not only decreases binding affinity by two logs, but also confers constitutive activity to the mutant receptor. Therefore, the balance of electric charges in the external side of TM2 (and, most likely, their interactions with adjacent negative charges in TM3) not only determines the ability of the receptor protein to bind ligands with high affinity, but also the prevalence of a resting, uncoupled conformation as opposed to an active conformation able to couple to Gs in an agonist-independ- ent fashion. Again consistent with the crucial role of charge interactions between the external side of TM2 and TM3, the Cys125Arg mutation confers constitutive activity to the fox receptor, and the homologous Cys123Arg mutation in the mouse receptor displays a similar pharmacology, highly reminiscent of the one observed for the mouse Glu92Lys and the human Glu94Lys variants [10, this study]. On the other hand, residues present in the second intracellular loop appear critical for interaction with G proteins. Several naturally occurring mutations in, or near this region, such as the Arg142His, Arg151Cys and Arg160Trp and Arg162Pro variants found by us and others abolish efficient coupling without major effects on binding [22,29,34, this study]. Notably, with the possible exception of the Arg142His allele, all these mutations abolish completely a positive charge in the relatively short second intracellular loop. It is tempting to speculate that the overall positive charge of this loop could be important for binding a negatively charged surface of the Gs protein. In many GPCRs, the third intracellular loop and the cytoplasmic C-terminal extension are the main regions involved in coupling to the G proteins [2]. It is clear that in the case of the MC1R, the contribution on the second intracellular loop is also relevant. ACKNOWLEDGEMENTS This work has been supported by grants PM1999-0138, from the DGI, Ministry of Science and Technology, Spain and PTR1995-0582-OP. Jesu´ sSa ´ nchez Ma ´ s is recipient of a fellowship from the Fundacio ´ n Se ´ neca, Comunidad Auto ´ noma de la Regio ´ ndeMurcia,and Concepcio ´ nOlivaresSa ´ nchez of a FPI fellowship from the Ministry of Education, Spain. REFERENCES 1. Marinissen, M.J. & Gutkind, J.S. (2001) G-protein-coupled receptors and signaling networks: emerging paradigms. Trends. Pharmacol. Sci. 22, 368–376. 2. Probst, W.C., Snyder, L.A., Schuster, D.I., Brosius, J. & Sealfon, S.C. (1992) Sequence alignment of the G-protein coupled receptor superfamily. DNA Cell. Biol. 11, 1–20. 3. Rozengurt, E. (1986) Early signals in the mitogenic response. Science 234, 161–166. 4. Moolenaar, W.H. (1991) Mitogenic action of lysophosphatidic acid. Adv. Cancer Res. 57, 87–102. 5. Gutkind, J.S. (1998) Cell growth control by G protein-coupled receptors: from signal transduction to signal integration. Oncogene 17, 1331–1342. 6. Young, D., Waitches, G., Birchmeier, C., Fasano, O. & Wigler, M. (1986) Isolation and characterization of a new cellular onco- gene encoding a protein with multiple potential transmembrane domains. Cell 45, 711–719. 7. Parma, J., Duprez, L., Van Sande, J., Cochaux, P., Gervy, C., Mockel, J., Dumont, J. & Vassart, G. (1993) Somatic mutations in the thyrotropin receptor gene cause hyperfunctioning thyroid adenomas. Nature 365, 649–651. 8. Shenker, A., Laue, L., Kosugi, S., Merendino, J.J., Minegishi, T. & Cutler, J.B. (1993) A constitutively activating mutation of the luteinizing hormone receptor in familial male precocious puberty. Nature 365, 652–654. 9. Schipani, E., Kruse, K. & Juppner, H. (1995) A constitutively active mutant PTH-PTHrP receptor in Jansen-type metaphyseal chondrodysplasia. Science 268, 98–100. 10. Lu, D., Chen, W. & Cone, R.D. (1998) The Pigmentary System. In Physiology and Pathophysiology (Nordlund, J.J., Boissy, R.E., Hearing, V.J., King, R.A. & Ortonne, J.P., eds), pp. 183–197. Oxford University Press, New York. 11. Mountjoy,K.G.,Robbins,L.S.,Mortrud,M.T.&Cone,R.D. (1992) The cloning of a family of genes that encode the melano- cortin receptors. Science 257, 1248–1251. 12. Harding, R.M., Healy, E., Ray, A., Ellis, N.S., Flanagan, N., Todd, C., Dixon, C., Sajantila, A., Jackson, I.J., Birch-Machin, M.A. & Rees, J.L. (2000) Evidence for variable selective pressures at MC1R. Am.J.Hum.Genet.66, 1351–1361. 13. Valverde, P., Healy, E., Jackson, I.J., Rees, J.L. & Thody, A.J. (1995) Variants of the melanocyte-stimulating hormone receptor gene are associated with red hair and fair skin in humans. Nature Genet. 11, 328–330. 14. Box, N.F., Wyeth, J.R., O’Gorman, L.E., Martin, N.G. & Sturm, R.A. (1997) Characterization of melanocyte stimulating hormone receptor variant alleles in twins with red hair. Hum. Mol. Genet. 6, 1891–1897. 15. Smith, R., Healy, E., Siddiqui, S., Flanagan, N., Steijlen, P.M., Rosdahl, I., Jacques, J.P., Rogers, S., Turner, R., Jackson, I.J., Birch-Machin, M.A. & Rees, J.L. (1998) Melanocortin 1 receptor variants in an Irish population. J. Invest. Dermatol. 111, 119–122. 16. Busca ´ , R. & Ballotti, R. (2000) Cyclic AMP a key messenger in the regulation of skin pigmentation. Pigment Cell Res. 13, 60–69. 17. Prota, G. (1992) Melanins and Melanogenesis. Academic Press, Inc. San Diego. 18. Thody, A.J. & Graham, A. (1998) Does alpha-MSH have a role in regulating skin pigmentation in humans? Pigment Cell Res. 11, 265–274. 19. Palmer, J.S., Duffy, D.L., Box, N.F., Aitken, J.F., O’Gorman, L.E., Green, A.C., Hayward, N.K., Martin, N.G. & Sturm, R.A. (2000) Melanocortin-1 receptor polymorphisms and risk of mel- anoma: is the association explained solely by pigmentation phe- notype? Am.J.Hum.Genet.66, 176–186. 20. Bastiaens, M.T., ter Huurne, J.A.C., Kielich, C., Gruis, N.A., Westendorp, R.G.J., Vermeer, B.J. & Bavinck, J.N.B. (2001) 6140 J. Sa ´ nchez Ma ´ s et al.(Eur. J. Biochem. 269) Ó FEBS 2002 Melanocortin-1 receptor gene variants determine the risk of nonmelanoma skin cancer independently of fair skin and red hair. Am.J.Hum.Genet.68, 884–894. 21. Box, N.F., Duffy, D.L., Irving, R.E., Russell, A., Chen, W., Griffyths, L.R., Parsons, P.G., Green, A.C. & Sturm, R.A. (2001) Melanocortin-1 receptor genotype is a risk factor for basal and squamous cell carcinoma. J. Invest. Dermatol. 116, 224–229. 22. Schioth, H.B., Phillips, S.R., Rudzish, R., Birch-Machin, M.A., Wikberg,J.E.S.&Rees,J.L.(1999)Lossoffunctionmutations of the human melanocortin 1 receptor are common and are associated with red hair. Biochem. Biophys. Res. Commun. 260, 488–491. 23. Abdel-Malek, Z., Swope, V.B., Suzuki, I., Akcali, C., Harriger, M.D., Boyce, S.T., Urabe, K. & Hearing, V.J. (1995) Mitogenic and melanogenic stimulation of normal human melanocytes by melanotropic peptides. Proc. Natl. Acad. Sci. 92, 1789–1793. 24. Shih, I.M. & Herlyn, M. (1993) Role of growth factors and their receptors in the development and progression of melanoma. J. Invest. Dermatol. 100, 196S–203S. 25. Morandini, R., Boeynaems, J., Hedley, S., MacNeil, S. & Gha- nem, G. (1998) Modulation of ICAM-1 expression by alpha-MSH in human melanoma cells and melanocytes. J. Cell Physiol. 175, 276–282. 26. Smit, N., Westerhof, W., Asghar, S., Pavel, S. & Siddiqui, A. (1989) Large-scale cultivation of human melanocytes using collagen- coated sephadex beads (cytodex 3). J. Invest. Dermatol. 92, 18–21. 27. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: A. Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 28. Loir, B., Pe ´ rez Sa ´ nchez, C., Ghanem, G., Lozano, J., Garcı ´ a- Borro ´ n, J.C. & Jime ´ nez-Cervantes, C. (1999) Expression of the MC1 receptor gene in normal and malignant human melanocytes. A semiquantitative RT-PCR study. Cell. Mol. Biol. 45, 1083–1092. 29. Jime ´ nez-Cervantes, C., Olivares, C., Gonza ´ lez, P., Morandini, R., Ghanem,G.&Garcı ´ a-Borro ´ n, J.C. (2001) The Pro162 variant is a loss-of-function mutation of the human melanocortin 1 receptor gene. J. Invest. Dermatol. 117, 156–158. 30. Jime ´ nez-Cervantes,C.,Germer,S.,Gonza ´ lez, P., Sa ´ nchez, J., Olivares,C.&Garcı ´ a-Borro ´ n, J.C. (2001) Thr40 and Met122 are new partial loss-of-function natural mutations of the human melanocortin 1 receptor. FEBS Lett. 508, 44–48. 31. Prusis, P., Schioth, M.B., Muceniece, R., Herzyk, P., Afshar, M., Hubbard, R.E. & Wikberg, J.E.S. (1997) Modeling of the three- dimensional structure of the human melanocortin 1 receptor, using an automated method and docking of a rigid cyclic mela- nocyte-stimulating hormone core peptide. J. Mol. Graph. Model. 15, 307–317. 32. Nielsen, H., Engelbrecht, J., Brunak, S. & von Heijne, G. (1997) Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Engineering 10, 1–6. 33. Sturm, R.A., Teasdale, R.D. & Box, N.F. (2001) Human pig- mentation genes: identification, structure and consequences of polymorphic variation. Gene 277, 49–62. 34. Fra ¨ ndberg, P A., Doufexis, M., kapas, S. & Chhajlani, V. (1998) Human pigmentation phenotype: a point mutation generates nonfunctional MSH receptor. Biochem. Biophys. Res. Commun. 245, 490–492. 35. Flanagan, N., Healy, E., Ray, A., Philips, S., Todd, C., Jackson, I.J., Birch-Machin, M.A. & Rees, J.L. (2000) Pleiotropic effects of the melanocortin 1 receptor (MC1R) gene on human pigmenta- tion. Hum. Mol. Genet. 9, 2531–2537. 36. Haskell-Luevano, C., Sawyer, T.K., Trumpp-Kallmeyer, S., Bikker, J.A., Humblet, C., Gantz, I. & Hruby, V.J. (1996) Three- dimensional molecular models of the hMC1R melanocortin receptor: complexes with melanotropin peptide agonists. Drug Des. Discov. 14, 197–211. 37. Yang, Y., Dickinson, C., Haskell-Luevano, C. & Gantz, I. (1997) Molecular basis for the interaction of [Nle4, D -Phe7] melanocyte stimulating hormone with the human melanocortin-1 receptor (melanocyte alpha-MSH receptor). J. Biol. Chem. 272, 23000– 23010. 38. Robbins, L.S., Nadeau, J.H., Johnson, K.R., Kelly, M.A., Roselli- Rehfuss, L.R., Baack, E., Mountjoy, K. & Cone, R.D. (1993) Pigmentation phenotypes of variant extension locus alleles result from point mutations that alter MSH receptor function. Cell 72, 827–834. 39. Scott, M.C., Wakamatsu, K., Ito, S., Kadekaro, A.L., Kobayashi, N., Groden, J., Kavanagh, R., Takakuwa, T., Virador, V., Hearing, V.J. & Abdel-Malek, Z.A. (2002) Human melanocortins 1 receptor variants, receptor function and melanocyte response to UV radiation. J. Cell. Sci. 115, 2349–2355. Ó FEBS 2002 Loss-of-function MC1R variants in human melanoma (Eur. J. Biochem. 269) 6141 . Loss -of- function variants of the human melanocortin-1 receptor gene in melanoma cells define structural determinants of receptor function Jesu ´ sSa ´ nchez. cAMP cascade occurs in these cells even in the absence of mutations in the coding region of the gene. Finally, our study of the natural MC1R variants highlights