Gene Structure and Sequence Polymorphism of the Coat Color Gene, Mc1r, in the Black-Bellied Vole (Eothenomys melanogaster)

Abstract

Yung-Chih Lai, Shiao-Wei Huang, and Hon-Tsen Yu (2016) Color polymorphism is a long-standing issue in ecological and evolutionary biology. The black-bellied vole (Eothenomys melanogaster) with complete melanic and brown forms provides an outstanding opportunity to study the genetic polymorphism underpinning color variation. Mutations in the coding region of melanocortin 1 receptor (Mc1r) have been shown to cause color variation in a wide range of species. However, the contribution to color variation produced by the Mc1r regulatory regions have rarely been studied in wild animals. To this end, the Mc1r promoter sequence in black-bellied voles was cloned and characterized in this study. At least 11 distinct transcription initiation sites were identified using 5’-RACE. Furthermore, a candidate core promoter region in the upstream GC-rich sequence was identified based on key transcription factor binding motifs. The black-bellied vole Mc1r coding region was conserved with that found in the house mouse and demonstrated characteristics that are consistent with the structure of a G-protein coupled receptor, e.g. seven transmembrane domains. We found a negative association between coat color variations and polymorphisms of either regulatory or coding regions. This implies that Mc1r might reflect geographic cline rather than adaptive evolution. Although we found a negative association, the extra information we obtained in the Mc1r promoter of the black-bellied vole can be beneficial to other studies in exploring the association between regulatory mutations and adaptive phenotypes in wild animals.

BACKGROUND

The striking color polymorphism in animals has held a longstanding attraction to ecologists and evolutionists. Many color variations reflect an adaptive function in a wide range of animals, such as rodents (Burtt 1981; Caro 2005). The rodent genus Eothenomys is distributed in the Oriental Region, and contains eight species (Luo et al. 2004). Within this genus, black-bellied voles (Eothenomys melanogaster) are endemic to Taiwan, occur in the motane areas (1500-2500 m in altitude) and prefer understory habitats in the mist forests (Yu 1993, 1994). The voles show a wide variation in dorsal coat color from brown to black across its geographic range (Fig. 1). In general, the gradation of color from dark on the dorsum to light on the ventrum, named countershading, is widespread in animals to avoid detection by predators (Rowland 2009). Consequently, the dorso-ventral differentiation of coat colors often features a demarcation line in the flanks between the dark dorsum and light ventrum, especially in mammals. In striking contrast, the dark black- bellied vole form shows uniformly dark dorsal and ventral coat colors (complete melanism) without a clear demarcation between dorsal and ventral regions (Fig. 1a). This pigmentation pattern appears to be similar to that found in melanic laboratory mouse strains, e.g. C57BL/6J. The complete melanic form is rarely found in other species of mammals. Moreover, the dark black- bellied vole form is the dominant color morph in some localities while in other places none of the voles are dark, providing an excellent model in which to explore the sequence variation of coat color genes in different morphs.

Fig. 1.

Fig. 1. Variation in coat color of black-bellied voles (Eothenomys melanogaster). (a) A black color form (Yu2045) sampled at Alishan, location I in the map (b) A brown color form (Yu2014) sampled at Wuling, location IV in the map (c) Alishan habitat (I in the map) with darker soil color (d) Wuling habitat (IV in the map) with lighter soil color (e) Map of Taiwan showing sites of sample collection: Alisan (I), Tataka (II), Guanwu (III), and Wuling (IV); the map inset shows the geographical location of Taiwan (red color) in East Asia.

To date, a large number of genes have been suggested to affect vertebrate pigmentation. For example, in mice, more than 100 loci and 800 phenotypic alleles have been identified (Bennett and Lamoreux 2003). Among them, melanocortin-1 receptor (Mc1r), a critical regulator in pigmentation synthesis, has been shown to contribute to color polymorphism in numerous species (Majerus and Mundy 2003). The gene encodes a G-protein coupled receptor with seven transmembrane domains that is specifically expressed in melanocytes (melanin-producing cells). In mammalian melanocytes, there are two basic types of melanin: (1) eumelanins, which are black or brown, and (2) pheomelanins, which are yellow or red. Typically, coat color variation among mammals results from the mixing of these two types of pigments in different proportions. The expression of the Mc1r gene is regulated by the agonist, α-Melanocyte-stimulating hormone (α-MSH), and the antagonist, agouti protein. Binding of Mc1r by α-MSH results in an increase in the synthesis of eumelanins, whereas Mc1r bound by agouti protein will increase the production of pheomelanins (Hearing 2000). In addition, Mc1r coding mutations have been shown to be responsible for color differences in a wide range of species (Hoekstra et al. 2006), suggesting that it may be a good candidate gene to reveal the relationship of sequence mutation and color polymorphism in black-bellied voles.

One of the principal goals in evolutionary biology is to elucidate the genetic mechanisms underlying an adaptive trait. However, whether the evolution of an adaptive phenotype mainly results from mutations in the cis-regulatory elements (regulatory theory) or the coding regions remains controversial (Hoekstra and Coyne 2007; Wray 2007). Analysis of colors in vertebrates is currently the most promising approach to explore this question, because variation in adaptive morphological traits in animal colors dramatically demonstrates a clear relationship between phenotypes and Mc1r genotypes (review in Hoekstra 2006; Mundy 2005). However, so far efforts have been mostly focused on finding mutations in the Mc1r coding region, even though Mc1r cis-regulatory mutations may play a role in color variation (Rouzaud and Hearing 2005; Adachi et al. 2000). Since, the promoter region of the Mc1r gene has rarely been cloned in wild mammals, our analysis here can provide extra insights into the phenotype-genotype association underlying coat color variation. Therefore, the aim of the present study was to clone and characterize the black-bellied vole Mc1r gene promoter and coding regions. In addition, we also describe the color polymorphism of black-bellied voles in natural populations and its negative association with Mc1r polymorphism.

MATERIALS AND METHODS

Samples and measurements of coat color

A total of 37 black-bellied voles were collected from 4 locations in Taiwan (Table 1) that belong to two geographical regions in the Central Mountain Range (CMR) - the northern and southern regions, which are approximately 112km apart (Fig. 1e). Alishan (Al, N = 14) and Tataka (Ta, N = 8) are only 9 km apart and located in the southern part of the CMR (referred to as southern samples); Guanwu (Gu, N = 6) and Wuling (Wu, N = 9) are 23 km apart and located in the northern part of the CMR (referred to as northern samples). The average annual precipitation for Alishan, Tataka, Guanwu, and Wuling were 3798 mm, 3136 mm, 3598 mm, and 1473 mm, respectively. Voles were trapped from forested habitats of motane regions and brought to the lab for processing. Within 2-3 days after entering the lab, animals were sacrificed by cervical vertebral dislocation, and tissue samples were preserved from all individuals for DNA molecular analysis. Skin specimens were prepared as voucher specimens for coat color analysis. Because black-bellied voles exhibit coat color polymorphisms, we classified each specimen collected as melanic, intermediate, or brown (Fig. 1).

Table 1. Nucleotide polymorphisms at the Mc1r gene.

population	coat color	specimen	88	89	374	750	913	1104	1279	1323	1518	1524	1607
Guanwu (Gu) 33% 3598 mm	melanic	Yu2086	C	C	G	G	C	T	A	C	G	G	A
	melanic	Yu2088	.	.	.	.	.	.	.	.	G/A	.	.
	intermediate	Yu2091	.	.	.	.	.	.	.	.	.	.	.
	intermediate	Yu2092	.	.	.	.	.	.	.	.	.	.	.
	intermediate	Yu2093	.	.	.	.	.	.	.	.	.	.	.
	intermediate	Yu2087	.	.	.	.	C/G	.	.	.	.	.	.
Wuling (Wu) 0% 1473 mm	brown	Yu2036	.	.	.	.	.	.	.	.	.	.	.
	brown	Yu2041	.	.	.	.	.	.	.	T	.	.	.
	brown	Yu2037	.	.	.	A	.	.	.	T	.	.	.
	brown	Yu2014	.	.	.	G/A	.	.	.	T	.	.	.
	brown	Yu2040	.	.	.	G/A	.	.	.	T	.	.	.
	brown	Yu2039	.	.	.	.	.	.	.	C/T	.	.	.
	brown	Yu2044	.	.	.	.	.	.	.	C/T	.	.	.
	brown	Yu2038	.	.	.	G/A	.	.	.	C/T	.	.	.
	brown	Yu2042	.	.	.	G/A	.	.	.	C/T	.	.	.
Tataka (Ta) 50% 3136 mm	melanic	Yu2022	A	G	.	.	.	C	G	.	.	.	G
	melanic	Yu2072	A	G	.	.	.	C	G	.	.	.	G
	melanic	Yu2071	.	.	A/G	.	.	C	G	.	.	G/A	G
	melanic	Yu2064	C/A	C/G	.	.	.	C	G	.	.	G/A	G
	intermediate	Yu2076	.	.	.	.	.	C	G	.	.	A	G
	brown	Yu2082	.	.	A	.	.	C	G	.	.	.	G
	brown	Yu2020	A	G	.	.	.	C	G	.	.	.	G
	brown	Yu2066	C/A	C/G	.	.	.	C	G	.	.	G/A	G
Alishan (Al) 64% 3798 mm	melanic	Yu2045	.	.	A	.	.	C	G	.	.	.	G
	melanic	Yu2046	.	.	A	.	.	C	G	.	.	.	G
	melanic	Yu2050	.	.	A	.	.	C	G	.	.	.	G
	melanic	Yu2051	.	.	A	.	.	C	G	.	.	.	G
	melanic	Yu2053	.	.	A	.	.	C	G	.	.	.	G
	melanic	Yu2058	.	.	A	.	.	C	G	.	.	.	G
	melanic	Yu2059	.	.	A	.	.	C	G	.	.	.	G
	melanic	Yu2049	.	.	.	.	.	C	G	.	.	.	G
	melanic	Yu2055	.	.	A/G	.	.	C	G	.	.	.	G
	intermediate	Yu2056	.	.	A	.	.	C	G	.	.	.	G
	intermediate	Yu2057	.	.	A	.	.	C	G	.	.	.	G
	brown	Yu2048	.	.	A	.	.	C	G	.	.	.	G
	brown	Yu2052	.	.	A	.	.	C	G	.	.	.	G
	brown	Yu2054	C/A	C/G	A/G	.	.	C	G	.	.	.	G
Gene structure			5’UTR			Exon							3’UTR
Amino acid							Arg		Ala
							Gly		Thr
Amino acid position						54	109	172	231	245	310	312

Numbers in the top row correspond to nucleotide positions. Dots represent identity with respect to the first sequence. For heterozygous sites, the genotype is indicated. Gene structure, amino acid positions, and nonsynoymous changes are indicated at the bottom of the corresponding column. % indicates the proportion of the melanic form. The amount of annual precipitation is given in mm.

Rapid amplification of 5’ complementary DNA ends (5’RACE) and primer design

Poly (A)⁺ RNAs from skin and tail tissues were extracted according to the Trizol protocol (Invitrogen). The 5’-UTR sequences and the transcription initiation sites of the Mc1r gene were determined using a 5’ RACE kit (Roche). Initially, we amplified and obtained the exon sequences of the Mc1r gene using a pair of primers (mMc1r- 1-for and mMc1r-2-rev) specific for the house mouse (Mus musculus) (Wada et al. 2005). Based on these exon sequences, we designed three gene-specific primers for 5’RACE: EmMc1r_ SP1 (5’-CCAGACAGCAGATGAAGCAA-3’) for synthesis of first strand cDNA using AMV reverse transcriptase, EmMc1r_SP2 (5’-GAGGCCATCTGGGATAGACA-3’) for first PCR to amplify the adaptor-ligated cDNA, and EmMc1r_ SP3 (5’-GGTAGCCAGTCCAAGGTGAG-3’) for a second PCR reaction. Using these primers, we obtained 5’UTR sequences. To obtain the sequences encompassing 5’UTR and its entire exon, an additional primer pair, EmMc1r_ F1 (5’-CTACGGGGGCTTTGAACAC-3’) and EmMc1r_R1 (5’-TGGTCCCAGGCAGTTTGTG-3’), was designed to amplify a 1714 bp fragment in the voles.

Genomic DNA extraction, PCR cloning, and sequencing

Vole genomic DNA was extracted from frozen muscle, liver or kidney tissues using commercial kits (Lamda). PCR amplification was performed in thermal cyclers (ABI 2720 or Bio-Rad PTC 200) in 100 μl volume containing 1.0 unit of Phusion DNA polymerase (Finnzymes), 20 μl 5× Phusion HF buffer, 0.2 mM for each dNTP and 1.5 mM MgCl2. Thermocycler conditions were as follows: (1) an initial denaturing step of 98°C for 30 s; (2) 30 cycles of the following: 10 s at 98°C, 10 s at 58-61°C, 2 min at 72°C; and (3) a 10 min extension step at 72°C. The 1714 bp PCR product was purified using a QIAquick gel extraction kit (Qiagen). After A-tailing with 10mM dATP at 72°C for 30 min, TA cloning was performed using the pGEM-T vector system (Promega) according to the manufacturer’s instructions. At least five clones were sequenced for each vole.

Data analysis

Sequences were edited and aligned manually by Sequencher, v.4.1.4 (Genecodes). All Mc1r sequences generated in this study have been deposited in GenBank (accession numbers GU001524-GU001573). To define the boundaries of the coding region, E. melanogaster Mc1r sequences were aligned with those of the house mouse (M. musculus) and pocket mouse (Chaetodipus intermedius) by ClustalW. In addition, the transmembrane domains were independently predicted by four web tools: HMMTOP 2.0 (http:// www.enzim.hu/hmmtop/), SOSUI (http://bp.nuap. nagoya-u.ac.jp/sosui/), TMHMM 2.0 (http://www. cbs.dtu.dk/services/TMHMM-2.0/), and TMpred (http://www.ch.embnet.org/software/TMPRED_ form.html). Potential transcription factor binding sites were predicted by the Transcription Element Search System (TESS; http://www.cbil.upenn.edu/ tess). A parsimony haplotype network based on Mc1r sequences from all alleles was constructed by TCS 1.21. Evidence for selection at Mc1r was based on Tajima’s D statistic using the program DnaSP 4.10.9.

RESULTS

Color polymorphism

The coat colors on black-bellied voles are variable between individuals and populations (Table 1). For example, in Alishan, we found three coat colors and the melanic form was dominant. In Wuling, we only found the brown form. In addition, the most melanic population (Alishan) lives in the most humid environment (3798 mm annual precipitation), while the least melanic population (Wuling) inhabits the driest habitat (1473 mm annual precipitation). The relationship between coat color and annual precipitation corresponds to Gloger’s rule, which states that individual endothermic animals are darker in humid habitats than those in drier environments (Gloger 1833; Zink and Remsen 1986). It implies that coat colors may, at least in part, reflect environmental variation.

Characteristics of the 5’UTR and coding regions of the Mc1r gene

A region of 1714 bp of the Mc1r gene, including the entire coding region (954 bp), 5’UTR (588 bp), and 3’UTR (172 bp), was characterized for the black-bellied vole. The 5’UTR sequence shows 71.7% nucleotide identity with that of the house mouse (Adachi et al. 2000). Totally, eleven transcription initiation sites identified by the 5’- RACE were stretched over an ~500 bp GC-rich (57.38%) region upstream of the Mc1r coding region (Fig. 2, positions at 1, 199, 256, 271, 276, 293, 302, 305, 318, 378, and 474). No apparent transcriptional initiation site clusters were noted, in contrast to those reported in humans (Moro et al. 1999). We predict three and one putative transcription factor binding sites for SP-1 and AP-2, respectively, using TESS. Moreover, five CANNTG motifs, which have been experimentally validated in the house mouse Mc1r promoter were also discovered in black-bellied voles (Adachi et al. 2000). Neither the TATA nor CAAT boxes were found in this proximal region.

Fig. 2.

Fig. 2. The 5’ UTR nucleotide sequence of the Mc1r gene. The maximum length of the 5’UTR is 588 bp. The first ATG codon is boxed. Transcriptional initiation sites are indicated by arrows. Consensus binding sites for SP-1 and AP-2 transcription factors are underlined. The parentheses refer to mouse (M) or rat (R). The CANNTG motifs based on reports from mouse experiments are heavily underlined.

From start codon (ATG) to stop codon (TGA), the nucleotide positions 589-1542 (954 bp) corresponds to the gene exon sequence in the house mouse and pocket mouse. The exon sequence in the black-bellied vole shows 89.0% and 82.8% nucleotide identity to that of the house mouse (Wada et al. 2005) and the pocket mouse (Nachman et al. 2003), respectively. All four bioinformatics tools independently predict that the Mc1r coding region we sequenced contains seven putative transmembrane domains, a key feature of the G-protein coupled receptor.

Mc1r sequence polymorphism

Eleven polymorphic sites were detected, three in the 5’UTR, seven in the exon, and one in the 3’UTR (Table 1). Within the exon, five of the substitutions were synonymous and two were nonsynonymous. The two nonsynonymous mutations (amino acid positions 109 and 231) occurred in the second extracellular region and third intracellular region, respectively. None of the eleven substitutions showed a significant association with coat color variation (Table 1), i.e., we did not observe variants that were unique to melanic, intermediate, or brown coat color in the four populations. This suggests that Mc1r is not a principal determinant underlying coat color polymorphism in the black-bellied vole, at least in the four populations we sampled.

Phylogenetic relationships and testing for selection

Sequence variation at Mc1r was largely consistent with the phylogeographic relationship inferred from mitochondrial cytochrome B sequence variation (Chang 2007). Three polymorphic sites (Table 1: 1104, 1279, and 1607) support a split between the northern and southern vole populations, which correspond to the basal split inferred from mtDNA data. In addition, the genealogy of Mc1r haplotypes (Fig. 3) is similar in topology to those generated from mitochondrial cytochrome B for black-bellied voles (Chang 2007). Only two nonsynonymous substitutions were observed in the Mc1r sequence, one in a heterozygote (Table 1: specimen: Yu2087, nucleotide position 913, amino acid position 109), and the other separating the northern and southern vole populations (nucleotide position 1279, amino acid position 231). Although Tajima’s D statistic was positive in our analysis, it did not significantly differ from zero (0.87142, P > 0.10), suggesting that the coat color related gene, Mc1r, is not subject to selection.

Fig. 3.

Fig. 3. Statistical parsimony network based on Mc1r haplo- types. Each line between haplotypes indicates one substitution, with the positions of the nucleotide substitutions noted above. * indicates a nonsyonymous substitution. Al, Ta, Gu, and Wu: alleles from the Alishan, Tataka, Guanwu, and Wuling populations, respectively. AlTa: alleles from the Alishan and Tataka populations. WuGu: alleles from the Guanwu and Wuling populations. The stippled background indicates the two northern populations, and the white background indicates the two southern populations.

DISCUSSION

Gene expression in most eukaryotes is under the control of regulatory elements, especially promoter regions, which are usually located upstream of transcription initiation sites. We cloned and characterized the promoter region of Mc1r in black-bellied voles. Our results show that the ~ 500 bp GC-rich region upstream of the Mc1r coding region seems to be the main promoter. Not only are all of the detected transcription initiation sites located in this region, but also the features within this region (including GC-rich, TATA-less, and SP-1 transcription factor binding sites) are consistent with a G-protein-coupled receptor. In addition, the first two CANNTG motifs in the black- bellied vole are completely identical in sequence to the promoter region in the house mouse, as recognized by the transcription factor of the basic- helix-loop-leucine zipper (bHLH-Zip) protein family (Adachi et al. 2000). Therefore, we suspect that the two CANNTG motifs upstream from most of the transcription initiation sites are important candidate core promoters.

In contrast to most studies focused only on the coding region, we found non-association with color variations both in cis-regulatory and coding regions. Similar to our findings in the black-bellied vole, Mc1r mutations did not associate with color variations in other animals with a wide spectrum of color morphs (Goncalves et al. 2012; Hosoda et al. 2005; McRobie et al. 2014). It is likely that other genes, e.g. Agouti (Linnen et al. 2009) or Tyrp1 (Schmutz et al. 2002), may be responsible for coat color variation for each of these species. In the future, more advanced approaches will be adopted to explore the molecular evolution that underlies these adaptive traits (Supplement Fig. 1). For example, Hoekstra et al. (2006) showed that a single gene can influence several traits. In contrast, a single phenotype also can be regulated by multiple genes (Steiner et al. 2007). Now, the strength of population genomics has grown and will enable us to explore the genetic evolution of color genes at the whole genome scale (e.g. Poelstra et al. 2014). Taking advantage of next- generation sequencing (NGS), scientists are able to explore the whole genome for various adaptive characteristics. Alternatively, causative enhancers of Mc1r could be too far away from the transcription initiation sites to be identified. At this time, we cannot exclude that Mc1r mutations may underlie coat color variation in populations we did not screen. For example, among pocket mice an association between coat color variation and Mc1r coding substitution has been found in only one of four natural populations (Hoekstra and Nachman 2003). Consequently, our data do not completely rule out the potential role of Mc1r in this context. However, the consistent patterns between Mc1r- based and cytochrome B-based phylogenies and the fact that no significant selection was detected by Tajima’s D test suggest that the differences at the Mc1r locus, including regulative elements and the coding region, might reflect a geographic cline, as indicated by TCA analysis, rather than adaptive evolution. This means that genetic drift, rather than natural selection, could be the primary force driving Mc1r sequence polymorphism.

Despite the negative association of Mc1r and pigmentation in the black-bellied vole, feature and promoter region sequence similarities with the house mouse provide may aid in obtaining regulatory sequence information for the Mc1r gene in other species. This could also be applied to previous studies that found no association between colors and Mc1r coding mutations (Cheviron et al. 2006; Herczeg et al. 2010; MacDougall- Shackleton et al. 2003; Skoglund and Hoglund 2010; Wlasiuk and Nachman 2007). Based on our results, one can re-evaluate the phenotype- genotype association in the regulatory region, especially for core promoters. In addition, the experiments we designed to determine whether cis-regulatory mutations and/or coding mutations play a crucial role in adaptive traits have provided a promising avenue to confirm regulatory evolution. In the future, it might require more time to obtain sufficient evidence to confirm the hypothesis of regulatory evolution (Carroll 2005).

Finally, coat color variation in black-bellied voles may have some adaptive significance as we observed that more melanic forms were from wetter localities (see Table 1). This is in keeping with a long-recognized phenomena concerning mammalian coat color known as Gloger’s Rule in which animals in more humid environments tend to be darker than their counterparts in drier habitats (Gloger 1833; Zink and Remsen 1986). Two interpretations have been given for Gloger’s rule. One is that melanic forms are more cryptic against a darker background caused by higher vegetation density (i.e., shade) and darker soil color (i.e., saturated with moisture) as a result of increased levels of precipitation (Lai et al. 2008). An alternative explanation is that bacteria are more abundant and active in humid environments, and the melanic form with more eumelanin pigments resists bacterial degradation better than brown forms caused by pheomelanin pigments (Burtt and Ichida 2004). Both hypotheses require advanced experiments to confirm or deny their predictions in the future.

CONCLUSIONS

The key pigment gene, Mc1r, has provided a vital proxy to explore molecular evolution based on its protein coding region for various wild animals. Further, we used the black-bellied vole as an example to clone and characterize its regulatory region. The multiple transcription initiation sites and candidate transcription factor binding sites in its promoter provide regulatory potential for animal color patterns. Both the promoter and coding regions we sequenced are consistent with the biogeographical distribution, suggesting other candidate color genes could play a role for the populations we investigated. Whether the adaptive phenotypes are mainly caused by regulatory or coding mutations can be clarified by surveying their homologs in more populations and species.

Supplementary materials

Supplement Fig. 1.

Progression of approaches to analyze phenotype-genotype association for adaptive traits. (download)