- Split View
-
Views
-
Cite
Cite
Kimberley A. Beaumont, Richard A. Newton, Darren J. Smit, J. Helen Leonard, Jennifer L. Stow, Richard A. Sturm, Altered cell surface expression of human MC1R variant receptor alleles associated with red hair and skin cancer risk, Human Molecular Genetics, Volume 14, Issue 15, 1 August 2005, Pages 2145–2154, https://doi.org/10.1093/hmg/ddi219
- Share Icon Share
Abstract
The human melanocortin-1 receptor gene (MC1R) encodes a G-protein coupled receptor that is primarily expressed on melanocytes, where it plays a key role in pigmentation regulation. Variant alleles are associated with red hair colour and fair skin, known as the RHC phenotype, as well as skin cancer risk. The R151C, R160W and D294H alleles, designated ‘R’, are strongly associated with the RHC phenotype and have been proposed to result in loss of function receptors due to impaired G-protein coupling. We recently provided evidence that the R151C and R160W variants can efficiently couple to G-proteins in response to α-melanocyte stimulating hormone. The possibility that altered cellular localization of the R151C and R160W variant receptors could underlie their association with RHC was therefore considered. Using immunofluorescence and ligand binding studies, we found that melanocytic cells exogenously or endogenously expressing MC1R show strong surface localization of the wild-type and D294H alleles but markedly reduced cell surface expression of the R151C and R160W receptors. In additional exogenous expression studies, the R variant D84E and the rare I155T variant, also demonstrated a significant reduction in plasma membrane receptor numbers. The V60L, V92M and R163Q weakly associated RHC alleles, designated ‘r’, were expressed with normal or intermediate cell surface receptor levels. These results indicate that reduced receptor coupling activity may not be the only contributing factor to the genetic association between the MC1R variants and the RHC phenotype, with MC1R polymorphisms now linked to a change in receptor localization.
INTRODUCTION
There have been major advances in the understanding of the genetics of human pigmentation in the past decade (1,2), and important genes involved in skin, hair and eye colour have been identified. The melanocortin-1 receptor (MC1R) is a key determining gene for skin and hair colour variation between individuals, however, there is still much unknown about MC1R function.
Human cutaneous pigmentation is dependent on the production of the visible melanin pigment within the melanosome by epidermal and follicular melanocytes (3). Skin and hair colour is primarily a result of the relative amounts of the black/brown eumelanin and the red/yellow pheomelanin pigment production by melanocytes, as well as the number, size and distribution of melanin-filled melanosomes within the surrounding keratinocytes (4). MC1R plays an important role in determining the ratio of eumelanin and pheomelanin production (5). The rate of eumelanin and pheomelanin syntheses is largely controlled by the availability of the tyrosine and cysteine substrates and activity of three melanosomal enzymes; tyrosinase (TYR), tyrosinase-related protein 1 (TYRP1) and DOPAchrome tautomerase (DCT) (6). The expression of these enzymes can be regulated by MC1R signalling. Stimulation of MC1R by the natural ligand α-melanocyte stimulating hormone (α-MSH) mediates activation of adenylate cyclase, subsequent elevation of cAMP levels (7) and activation of the microphthalmia transcription factor (MITF) (8). MITF binds to the M box, a conserved region found in the promoters of the TYR, TYRP1 and DCT genes, and stimulates the transcriptional up-regulation of these proteins with concomitant maturation of the melanosome (9). This ultimately results in increased eumelanin production and darkening of the skin or hair.
Valverde et al. (10) first identified a link between polymorphism within the human MC1R gene and red hair, fair skin and poor tanning ability, known as the RHC phenotype. It was hypothesized that variant alleles with altered MC1R function would result in poorer MC1R signalling and subsequently less eumelanin synthesis. The MC1R variants most strongly associated with the RHC phenotype include D84E, R151C, R160W and D294H, which have been designated as ‘R’ alleles with an odds ratio for red hair of 63.3 (95% CI 31.9–139.6). The V60L, V92M and R163Q variant alleles have a relatively weak association with RHC and are designated as ‘r’ alleles with a significantly lower odds ratios of 5.1 (95% CI 2.5–11.3) (11–13). Although there is a lack of statistical data, the infrequent R142H and I155T alleles have been shown to have a strong familial association with RHC (13,14). The functional differences between the consensus wild-type receptor and the RHC-associated variants have yet to be resolved. The R151C, R160W and D294H polymorphisms were initially proposed to result in impaired G-protein coupling ability, and this was thought to be the mechanistic basis of their association with the RHC phenotype (15–17) and skin cancer risk (1,18). However, we recently provided evidence that the R151C and R160W variants, in contrast to D294H, can efficiently couple to G-proteins in response to α-melanocyte stimulating hormone (19). We postulated that a difference in the localization of the human wild-type, R151C- and R160W-associated variant receptors in melanocytes might be responsible for the observed phenotype. To investigate this possibility, we have examined the relative cell surface expression of these MC1R alleles through heterologous expression as well as endogenously in primary cultures of genotyped melanocytes.
RESULTS
Immunofluorescence detects reduced cell surface expression of R151C and R160W MC1R variant alleles
To investigate the localization of the consensus wild-type and variant MC1R, green fluorescent protein (GFP) tagged MC1R wild-type and R151C DNA constructs were transiently transfected into human MM418 and MM96L metastatic melanoma and HEK293 embryonic kidney cells. Similar high levels of intracellular cytoplasmic staining were seen for both fusion proteins (Fig. 1, middle rows). Cell surface expression on unpermeabilized cells was detected with an MC1R antibody directed against the N-terminus of MC1R (N-19). The specificity of antibody staining was demonstrated by the complete overlap of GFP and MC1R antibody immunofluorescence in permeabilized cells, with only a low level of dim background staining with the MC1R antibody (data not shown). Cell surface staining was strong for the MC1R wild-type-GFP fusion proteins, however, the R151C-GFP variant receptor showed almost no detectable cell surface expression in this assay (Fig. 1). These results were reproduced using a time course of 6, 24 and 48 h after transfection, suggesting that there is no temporal regulation of this difference in cell surface expression.
Similarly, untagged wild-type, R151C, R160W and D294H receptors were transiently transfected into MM418 and MM96L melanoma and HEK293 cell lines. The wild-type and D294H forms of the receptor showed strong cell surface expression, whereas the R151C and R160W allelic variants both showed markedly reduced cell surface receptor staining (Fig. 2A–C). Quantification of R151C and R160W cell surface fluorescence indicated an 80–95% reduction when compared with wild-type in the melanoma cell lines tested, HEK293 showed a 60–70% reduction in cell surface expression of the R151C and R160W receptors. In contrast, the D294H receptor was consistently higher than the wild-type (Fig. 2D). Equivalent amounts of MC1R wild-type and variant MC1R protein were detected by western blot analysis in cells prepared in parallel with immunofluorescence experiments (Fig. 2E), showing that a difference in the total expression levels is not responsible for the observed reduction of receptor surface expression.
Radioligand binding confirms reduced cell surface expression of R151C and R160W MC1R variants in both heterologous and endogenous expression systems
The reduction in cell surface expression for the R151C and R160W variant receptors was confirmed using binding of radiolabelled MC1R ligand, 125I-NDP-MSH, to whole cells at room temperature. Previous studies with 125I-α-MSH have shown very low association/dissociation rates at 4°C (20), hence room temperature was used to achieve steady-state conditions compatible with melanocytic cell viability. MM418 melanoma cells transiently transfected with the R151C and R160W variant allele constructs showed <20% the level of surface binding displayed by the wild-type receptor (Fig. 3A) while still harbouring a similar level of total receptor expression (Fig. 3B). A similar finding was seen in HEK293 cells stably expressing each of these constructs (Fig. 4A), although a greater proportion of R151C and R160W receptor expression was detected at the cell surface of these non-melanocytic cells, which likely explains the ability to detect substantial functional responses mediated by these two variants in this system (19). The response, however, was not dramatically different to that mediated by cells expressing the wild-type receptor. This is likely a consequence of lower total MC1R expression levels in the wild-type expressing cell lines, clonal variability in the relationship between receptor expression and cAMP production (21) and differences during the kinetics of the cAMP response during the 15 min stimulation period limiting the assessment to the detection of large differences in response (19).
It was also apparent that the D294H variant consistently demonstrated the highest level of surface expression, which may be a consequence of its poor ability to couple to adenylate cyclase activation (19,22), hence imparting a resistance to receptor internalization mechanisms dependent on efficient G-protein coupling (23).
Cell membrane expression of MC1R was also examined in the endogenous system of primary cultures of human melanocytes homozygous for wild-type (+/+), R151C (−/−) or R160W (−/−) variant alleles (24). Specific surface binding was profoundly reduced in the three-melanocyte strains homozygous for either R151C or R160W alleles (Fig. 4C) when compared with wild-type levels, despite similar levels of MC1R mRNA expression detected by quantitative PCR (data not shown). This highlights the physiological relevance of this altered cellular localization associated with the R151C and R160W polymorphisms on MC1R-mediated melanocyte function.
Immunofluorescence detection of cell surface expression of exogenously expressed familial RHC and r MC1R variant receptors
To study the expression patterns of other known RHC variants, untagged wild-type, V60L, D84E, V92M, R142H, I155T and R163Q receptors were transiently transfected into MM418 and MM96L melanoma and HEK293 cell lines. The R151C variant receptor was also included for comparison with previous experiments. As before, all variants showed strong intracellular staining (data not shown). Quantification of cell surface fluorescence in MM418 indicated that the V92M and R142H forms of the receptor have similar or slightly higher surface expression to wild-type, but once again the R151C allelic variant showed an 80% reduction in cell surface receptors. The D84E variant showed a similar reduced level of surface receptors, whereas the I155T variant displayed an 85% decrease when compared with wild-type (Fig. 5A). The R163Q and V60L variants showed a smaller reduction in plasma membrane receptor levels of ∼80 and 55% of the consensus allele, respectively (Fig. 5A). These results were comparable in MM96L transient transfections, whereas HEK293 showed similar trends, although the differences were again less marked (data not shown). Equivalent amounts of MC1R wild-type and variant MC1R protein were detected by western blot analysis in cells prepared in parallel with immunofluorescence experiments (Fig. 5B), showing that a difference in the total expression levels is not responsible for the observed reduction of receptor surface expression. Two additional MC1R bands of reduced electrophoretic mobility were seen with the I155T variant, and it is notable that this allele has the lowest cell surface expression level.
Radioligand binding assay of cell surface expression of exogenously expressed familial RHC and r MC1R variant receptors
The cell surface expression of the D84E, V92M, R142H, R151C, I155T and R163Q variant receptors in relation to wild-type was also quantified using binding of the radiolabelled MC1R ligand, 125I-NDP-MSH, to whole cells at room temperature. The most dramatic reductions were seen in MM418 melanoma cells transiently transfected with variant allele constructs for D84E, R151C and particularly I155T, where surface binding was indistinguishable from mock transfected cells (Fig. 6A). Supporting the immunofluorescence results, the R163Q and V60L variants showed smaller reductions in cell surface expression (Fig. 6B). Quantification of percentage maximum binding relative to wild-type (Fig. 6C) demonstrated similar levels of cell surface expression of the RHC variants demonstrated previously by immunofluorescence (Fig. 5A). All cells harboured equivalent levels of total receptor expression (Fig. 5B). It is notable that the most significant reductions could be seen in the stronger R alleles, whereas the r alleles showed either similar levels to wild-type, as with the V92M variant, or a small reduction.
DISCUSSION
The novel finding, in both exogenous and endogenous experimental assay systems, that there is a marked reduction in expression on the external cell membrane of the RHC-associated MC1R R variants, D84E, R151C, I155T and R160W, has important implications for our understanding of this phenotypic association. Until now, it was assumed that loss of function paradigms could be applied to all the RHC-associated variants, that is, either loss of affinity for the ligand α-MSH or altered G-protein coupling was responsible for receptor inactivation (15–17,22). However, functional assays have produced conflicting results. Many of the RHC variant receptors have been reported to demonstrate reduced function compared with wild-type in cAMP assays (16,19,22,25,26). However, R151C and R160W (19) along with V92M and R163Q (22) have also been shown to efficiently couple to adenylate cyclase activation. The R151C, R160W and, to a lesser extent, D294H RHC variants were also able to partially rescue eumelanogenesis in MC1R deficient mice (27).
What is becoming clear is that variant receptors possess differing levels of activity rather than complete loss of function (12,19,22,27). Extending this, our results point towards differential effects of the variant alleles on receptor localization and trafficking within the cell. cAMP production is dependent on the number of surface MC1R sites per cell, and therefore, a lower receptor number would result in decreased cAMP production and down-stream signalling (21). This appears to be reflected in the substantially reduced cAMP response reported for primary melanocytes heterozygous for the R151C/D294H alleles and homozygous for R160W MC1R (28,29), as opposed to the normal response seen in melanocytes homozygous for V92M (29), which is classed genetically as an r allele and is reported here to have normal receptor expression levels on the cell surface. It cannot be ruled out that altered receptor activity played a role in the functional impairment described in these previous studies; however, as there was no quantification of cell surface expression, it is plausible that reduced levels of surface receptors underlie the reduced signalling. It is significant in our own results that the most substantial reductions in cell surface expression occurred with the R alleles strongly associated with RHC, whereas the r alleles either demonstrated levels similar to the consensus wild-type or less substantial reductions. The decreased cell surface expression of the D84E, R151C, I155T and R160W MC1R variant receptors is thus consistent with this reduction being the underlying molecular defect giving rise to their genetic association with the RHC phenotype, although this may not be the only contributing factor.
Although slight differences in the α-MSH binding properties for the V92M variant have been seen (22,30,31), there was no surface expression level changes noted. Similarly, previous ligand binding studies in transfected COS-1 cells have not described substantial differences in whole-cell binding between wild-type and RHC variant receptors (15,16). This may reflect the use of an 125I-NDP-MSH binding protocol at a higher incubation temperature of 37°C for 2 h (16). It has been shown for the endogenous murine MC1R expressed in B16 melanoma cells that a linear uptake of NDP-MSH occurs over 2 h at 37°C (32), which may complicate interpretation of cell surface levels. However, even the use of a room temperature procedure did not appear to identify a difference in radioligand binding between R151C and wild-type MC1R (15). It is therefore possible that COS-1 cells may permit higher surface expression of the R151C and R160W MC1R variants than is the case for HEK293 and particularly melanocytic cells. Such a consideration highlights the potential importance of cellular context in the examination of receptor trafficking. In this respect, it is notable that expression levels of trafficking proteins have been shown to differ dramatically between cell types (33). Indeed, expression of cell type specific accessory proteins has been shown to influence cell surface expression of a number of G-protein coupled receptors (GPCRs) including the melanocortin family member MC2R (34–37).
The mechanism resulting in the reduced cell surface expression of the D84E, R151C, I155T and R160W variant receptors is unknown and may be different in the case of each variant. There are many examples of single amino acid changes affecting the processing, folding, trafficking and internalization of various GPCR, a notable example being the vasopressin 2 (V2) receptor in which many mutations cause reduced cell surface expression and consequently X-linked nephrogenic diabetes insipidus (reviewed in 38). Cell-surface protein mis-localization is also observed in cystic fibrosis where mutations in the cystic fibrosis transmembrane conductance regulator result in defective processing and consequent loss of surface expression (39).
A more relevant example can be seen in the related melanocortin-4 receptor (MC4R), in which a number of allelic variants are associated with obesity (reviewed in 40). A large number of these mutations result in intracellular retention of the receptor, with a corresponding decrease in cAMP response also demonstrated (41–44). The MC4R R165W variant is equivalent to the R160W variant in MC1R, and two independent reports have detected reduced plasma membrane expression recording either 9 (41) or 26% (42) of MC4R wild-type. This corresponds with the 7–28% cell surface expression of R160W compared to MC1R wild-type detected by immunofluorescence (Fig. 2) and maximum binding data (Fig. 3) in the current study. Missense mutations in MC4R provide a parallel example of the affect of single amino acid changes on receptor cell surface expression, giving further credence to the results of the current study.
The D84E MC1R variant occurs in the second transmembrane region, and even though this is a conservative amino acid substitution, it may cause mis-folding of the receptor and consequently accumulation in the ER, as is seen for the rhodopsin receptor after introduction of a charged residue (45).
The topological model for MC1R based on the structure of bovine rhodopsin indicates that the R151C, I155T and R160W polymorphisms all occur in the second intracellular loop (22). Juxtaposition of these missense changes suggests that these variants may be affecting MC1R trafficking in a similar manner. The mechanism that results in reduced cell surface expression of the R151C, I155T and R160W receptors is currently unknown, although it is possible that the region altered by these mutations is needed for trafficking to the cell surface. Mutations introduced into targeting motifs required for adaptor-based sorting of cell surface proteins have been shown to result in their mis-targeting within the cell (46). A mutation in the second intracellular loop of the V2 receptor and MC4R has also been shown to result in defective transport (44,47).
Alternatively, some of the RHC alleles may cause altered receptor internalization and/or recycling. Studies have shown that NDP-MSH is internalized as a receptor complex with MC1R (32). The second intracellular loop has been implicated in receptor internalization in a number of GPCRs (23). The highly conserved DRY sequence, which is nearby the R151C and R160W polymorphisms, contributes to G-protein coupling and GPCR internalization (48). In MC1R, a naturally occurring variation in the DRY motif, R142H (14), has been linked to the RHC phenotype; however, in our assays, the R142H variant appears to have surface levels similar to wild-type receptor. This data therefore indicates that, as in other GPCRs, this central R residue is necessary for the receptor to mediate Gs protein activation. Furthermore, it is unlikely that the cell surface expression of the MC1R variant receptors tested was influenced by altered receptor/ligand complex internalization, as transfections were carried out in reduced serum media to minimize the presence of melanocortin ligands. However, MC1R does have agonist-independent, high basal activity (49), therefore, wild-type MC1R may undergo constitutive endocytosis reducing the surface expression, and it is possible that this constitutive internalization is altered for the R151C, R155T and R160W receptors. In this context, it is also of interest that the D294H RHC-associated variant consistently displayed a higher surface expression than the wild-type receptor. D294H has been linked to impaired G-protein coupling (16,19,22,25), so as a consequence of impaired G-protein coupling, the D294H variant could be expected to have a correspondingly reduced level of constitutive activity and therefore reduced basal internalization, resulting in a higher level of surface expression.
The second intracellular loop is also thought to be involved in GPCR recycling (50), therefore, impaired recycling of the R151C, I155T and R160W receptors could result in decreased surface expression. However, receptor recycling may not be the primary mechanism of re-populating the plasma membrane because wild-type murine MC1R in B16 melanoma cells does not appear to recycle back to the cell surface after internalization, suggesting degradation may be occurring (32). The location of the R151C and R160W polymorphisms also coincides with putative phosphorylation sites of PKA and PKC (26), respectively, raising the possibility that phosphorylation status may play a role in determining levels of surface expression. The infrequent I155T polymorphism lies between these two sites and may also influence phosphorylation status (14).
In conclusion, a molecular difference in the cell surface expression of wild-type MC1R and the RHC-associated MC1R variants, D84E, R151C, I155T and R160W, has been discovered in both transfected melanocytic cells and primary melanocyte strains homozygous for the R151C and R160W variant alleles. The location of the R151C, I155T and R160W substitutions in the second intracellular loop indicates that this domain may be involved in the intracellular trafficking of MC1R. Decreased cell surface expression and therefore decreased signalling ability may explain the genetic link between some of the stronger R alleles with the RHC phenotype seen in the human population, although this may not be the only contributing factor. These results are a significant contribution to our understanding of the role of MC1R in human pigmentation and effects of polymorphisms on MC1R intracellular trafficking. Further, since this study demonstrates several naturally occurring polymorphisms which decrease the level of MC1R surface expression, and previous studies have revealed a large number of missense variants that have a similar effect in MC4R, this may have important implications for all the members of the melanocortin receptor family and the function of other GPCRs.
MATERIALS AND METHODS
Cell culture
All cells were grown at 37°C with 5% carbon dioxide. MM418, MM96L and HEK293 cell lines were maintained in RPMI-1640 media (Invitrogen) supplemented with 10% heat inactivated serum supreme (DKSH), 1% l-glutamine and penicillin/streptomycin antibiotics (40 µg/ml). Melanocytes were maintained in RPMI-1640 media containing 3 mm HEPES buffer supplemented with 10% heat inactivated FBS, 1 mm pyruvic acid, 0.2 mm nicotinamide, 10 ng/ml TPA and 0.6 µg/ml cholera toxin. Queensland foreskin primary melanocyte strains were isolated and propagated from anonymous neonatal foreskin samples (24).
DNA constructs and transfections
The cDNA for consensus or variant human MC1R alleles were generated as XhoI–XbaI tagged Pfu-amplified PCR fragments and cloned into the pcDNA3.1 expression vector (Invitrogen) (19). The cDNA for the consensus and R151C variant MC1R alleles were also cloned into the XhoI–KpnI site of the pEGFP–N1 fluorescence-enhanced GFP vector (BD Biosciences Clontech) by replacing the MC1R stop codon with an inframe KpnI site (CACCGCGGTACCCAGGAGCATGTCAGCAC).
DNA constructs were transfected into cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer's guidelines. Cells were harvested 24 h after transfection for analysis. Transfection efficiency was ∼10% for melanoma cell lines and 50% for HEK293. Stably transfected HEK293 cell lines were obtained as previously described (19).
Immunofluorescence and western blotting
Cells were fixed in 4% paraformaldehyde in PBS for 60 min, permeabilized using 0.1% Triton X-100 for 5 min and then stained as previously described (46) using the following antibody titres: MC1R antibody (N-19, Santa Cruz Biotechnology) 1:100, Cy3-conjugated anti-goat secondary antibody (Jackson ImmunoResearch Laboratories) 1:400, and DAPI (Molecular Probes) 1:1000. Cells were viewed using an Olympus Provis AX-70 microscope with a 60× objective lens. Quantification of cell surface fluorescence was performed by selection of representative unpermeabilized cells expressing each variant receptor and taking pictures on a digital camera with the same settings optimized for wild-type MC1R expression. Mean fluorescence intensity (greyscale) was measured in Adobe photoshop and expressed as percentages relative to wild-type. Western blotting was performed as described previously (19).
Radioligand binding
Cells were washed once with binding buffer (RPMI with 20 mm HEPES, 0.25% BSA and 1 mm 1,10-phenanthroline), then incubated for 2 h at room temperature with binding buffer containing 0.1–0.2 nm directly labelled 125I-NDP-MSH obtained from either a commercial source (2000 Ci/mmol, Amersham Biosciences) or radioiodination of NDP-MSH (125I, 2574 Ci/mmol) and HPLC purification. To determine maximal binding and to verify similar ligand binding affinity for the variant receptors, homologous competition was carried out by including a range of concentrations of unlabelled NDP-MSH. Homologous competition analysis was facilitated by the fact that the affinity of directly labelled NDP-MSH, as measured by saturation binding analysis (data not shown), is similar to the Ki of the cold ligand, as has been reported previously (22). Unbound radioligand was removed by washing monolayers twice with binding buffer at 4°C, cells were then solubilized in 0.2 m NaOH and bound radioactivity quantified on a gamma counter. Curve fitting and analysis of the radioligand binding data was performed using Prism 4.0 (Graph Pad).
ACKNOWLEDGEMENTS
We thank Denise Adams and Peter Cassidy for synthesis and provision of radioiodinated NDP-MSH. The work was funded partly by grants from the National Health and Medical Research Council of Australia (301173) and Australian Research Council (DP0451738). R.A.S. is an NHMRC Senior Research Fellow. The Institute for Molecular Bioscience incorporates the Centre for Functional and Applied Genomics as a Special Research Centre of the Australian Research Council.
Conflict of Interest statement. None declared.
References
Sturm, R., Teasdale, R. and Box, N. (
Quevedo, W. and Holstein, T. (
Sturm, R.A., Box, N.F. and Ramsay, M. (
Naysmith, L., Waterston, K., Ha, T., Flanagan, N., Bisset, Y., Ray, A., Wakamatsu, K., Ito, S. and Rees, J.L. (
Busca, R. and Ballotti, R. (
Goding, C.R. (
Marks, M.S. and Seabra, M.C. (
Valverde, P., Healy, E., Jackson, I., Rees, J.L. and Thody, A.J. (
Box, N.F., Wyeth, J.R., Ogorman, L.E., Martin, N.G. and Sturm, R.A. (
Sturm, R.A., Duffy, D.L., Box, N.F., Newton, R.A., Shepherd, A.G., Chen, W., Marks, L.H., Leonard, J.H. and Martin, N.G. (
Duffy, D.L., Box, N.F., Chen, W., Palmer, J.S., Montgomery, G.W., James, M.R., Hayward, N.K., Martin, N.G. and Sturm, R.A. (
Flanagan, N., Healy, E., Ray, A., Philips, S., Todd, C., Jackson, I.J., Birch-Machin, M.A. and Rees, J.L. (
Frandberg, P.A., Doufexis, M., Kapas, S. and Chhajlani, V. (
Schioth, H.B., Phillips, S.R., Rudzish, R., Birch-Machin, M.A., Wikberg, J.E. and Rees, J.L. (
Sanchez Mas, J., Olivares Sanchez, C., Ghanem, G., Haycock, J., Lozano Teruel, J.A., Garcia-Borron, J.C. and Jimenez-Cervantes, C. (
Sturm, R.A., Duffy, D.L., Box, N.F., Chen, W., Smit, D.J., Brown, D.L., Stow, J.L., Leonard, J.H. and Martin, N.G. (
Newton, R.A., Smit, S.E., Barnes, C.C., Pedley, J., Parsons, P.G. and Sturm, R.A. (
Siegrist, W., Solca, F., Stutz, S., Giuffre, L., Carrel, S., Girard, J. and Eberle, A.N. (
Sanchez Mas, J., Gerritsen, I., Hahmann, C., Jimenez-Cervantes, C. and Garcia-Borron, J.C. (
Ringholm, A., Klovins, J., Rudzish, R., Phillips, S., Rees, J.L. and Schioth, H.B. (
Ferguson, S.S. (
Leonard, J.H., Marks, L.H., Chen, W., Cook, A.L., Boyle, G.M., Smit, D.J., Brown, D.L., Stow, J.L., Parsons, P.G. and Sturm, R.A. (
Robinson, S.J. and Healy, E. (
Siegrist, W., Stutz, S. and Eberle, A.N. (
Healy, E., Jordan, S.A., Budd, P.S., Suffolk, R., Rees, J.L. and Jackson, I.J. (
Kadekaro, A.L., Kanto, H., Kavanagh, R. and Abdel-Malek, Z.A. (
Scott, M.C., Wakamatsu, K., Ito, S., Kadekaro, A.L., Kobayashi, N., Groden, J., Kavanagh, R., Takakuwa, T., Virador, V., Hearing, V.J. et al. (
Xu, X., Thornwall, M., Lundin, L.G. and Chhajlani, V. (
Koppula, S.V., Robbins, L.S., Lu, D., Baack, E., White, C.R., Jr, Swanson, N.A. and Cone, R.D. (
Wong, W. and Minchin, R.F. (
Brown, D. and Stow, J.L. (
Brady, A.E. and Limbird, L.E. (
Saito, H., Kubota, M., Roberts, R.W., Chi, Q. and Matsunami, H. (
Tan, C.M., Brady, A.E., Nickols, H.H., Wang, Q. and Limbird, L.E. (
Metherell, L.A., Chapple, J.P., Cooray, S., David, A., Becker, C., Ruschendorf, F., Naville, D., Begeot, M., Khoo, B., Nurnberg, P. et al. (
Oksche, A. and Rosenthal, W. (
Welsh, M.J. and Smith, A.E. (
Lubrano-Berthelier, C., Cavazos, M., Dubern, B., Shapiro, A., Stunff, C.L., Zhang, S., Picart, F., Govaerts, C., Froguel, P., Bougneres, P. et al. (
Nijenhuis, W.A., Garner, K.M., van Rozen, R.J. and Adan, R.A. (
Lubrano-Berthelier, C., Durand, E., Dubern, B., Shapiro, A., Dazin, P., Weill, J., Ferron, C., Froguel, P. and Vaisse, C. (
Yeo, G.S., Lank, E.J., Farooqi, I.S., Keogh, J., Challis, B.G. and O'Rahilly, S. (
Tao, Y.X. and Segaloff, D.L. (
Sung, C.H., Davenport, C.M. and Nathans, J. (
Miranda, K.C., Khromykh, T., Christy, P., Le, T.L., Gottardi, C.J., Yap, A.S., Stow, J.L. and Teasdale, R.D. (
Tsukaguchi, H., Matsubara, H., Taketani, S., Mori, Y., Seido, T. and Inada, M. (
Arora, K.K., Cheng, Z. and Catt, K.J. (
Sanchez-Mas, J., Hahmann, C., Gerritsen, I., Garcia-Borron, J.C. and Jimenez-Cervantes, C. (