The melanocortin 1 receptor (mc1r) is a gene that has been implicated in the wide

Introduction The melanocortin 1 receptor (mc1r) is a gene that has been implicated in the wide variety of colors that exist in nature. It is responsible for hair and skin color in humans and the various colors that can be seen in animals across the globe. The extent of the importance of the gene is continually undergoing research, but is already known to be the only gene that controls the observed variations in human pigmentation (Rees 200). One way to determine the importance of a gene is to observe the evolution of it between and among different species. Canis lupus familiaris, or the dog, is an ideal species to use due to the wide variety of colors that exist among the different breeds. However, humans have played an important role in the domestication of the present day dog, and artificial selection may sway the results. However, by comparing many breeds of dogs to other carnivores and omnivores, both closely and distantly related, and using the results to observe the evolution of mc1r across species, we will determine the importance of the gene. It is well known that the wolf is the closest relative of the dog. We will use the sequences for the mc1r gene from multiple wolf breeds to compare the various dog breeds to determine which wolf breed is closest related to the dog. We will also use the mc1r sequences from other carnivores and omnivores to create trees that will tell us how well the gene has been conserved across species. This will tell us how important the gene is, has been, and can be predicted to be throughout the years. The species that were used were all either omnivores or carnivores, the majority of which are also land mammals. The dog breed sequences that were used consisted of a,,, Siberian husky,, Tibetan Mastiff, and the Alaskan Malamute. The dog breeds were then compared to three different breeds of wolf; the Eurasian wolf, Arctic

wolf, and Gray wolf. The other land mammals that were used were a bear, fox, Arctic fox, raccoon dog, human, chimpanzee,, jaguarundi, Tasmanian devil, gopher, and mouse. Most of the species used are very well known. However, the raccoon dog, fisher, jaguarundi, and Tasmanian devil are a little less talked about. The raccoon dog (Nyctereutes procyonoides) is a type of wild dog, about the size of a fox, which has markings that resemble a raccoon. Its habitat and many of its behaviors are very similar to that of a wolf or dog (WAZA 2014). It was included in the analysis due to its similarities between both the wolf and the dog. The fisher (Martes pennanti) is a carnivorous member of the marten family that resides in the forests of North America (Kyle et al. 2002). The jaguarundi (Puma yagouaroundi) is a close relative to the puma and lives in regions of Atlantic Rainforest and eucalypt plantations of southeastern Brazil (Tofoli et al. 2009). It was included in the analysis to provide multiple comparisons between different members of the feline family. The Tasmanian devil (Sarcophilus harrisii) is a carnivorous marsupial that is endemic to the Australian island of Tasmania. It was included in the data analysis because it is unique since it is the largest carnivorous marsupial, but also possesses some qualities that are similar to a dog or a wolf (Jones et al. 2008). Finally, a Weddell seal, alligator, and killer whale were also included. The hypothesis is that there will be positive or diversifying selection among the different dog breeds due to the influence of humans and artificial selection. Artificial selection would allow humans to select for more appealing colors and phenotypes that would alter natural evolution of the gene (Lindblad-Toh et al. 2005). However, across species, the hypothesis is that there will be purifying selection that will confirm the importance and conservation of the mc1r gene. It can also be assumed that the dog breeds will be closest related to the wolf breeds.

Materials and Methods The data analysis consisted of the alignment of the various sequences, the production of phylogenetic trees, and calculations to determine what, if any, kind of selection was present between the various dog breeds. The sequences were found using the gene search for mc1r on the National Center for Biotechnology Information website. The accession numbers for each of the sequences can be found in Appendix A. The first step was to align the sequences using CLUSTALW. Once the sequences were aligned, the next step was to create distance matrices and phylogenetic trees. Two distance matrices were created; one using the Jukes and Cantor model and one using Kimura s model. This was accomplished using the program for Molecular Evolutionary Genetics Analysis, or MEGA. Using MEGA, UPGMA (unweighted pair group method with arithmetic mean), maximum likelihood, and maximum parsimony trees were created. The UPGMA and maximum likelihood trees were created using both the Jukes Cantor model and Kimura model, resulting in two of each of the two kinds of trees. Each tree also used the bootstrap method with 0 replicas. For both UPGMA and maximum likelihood trees, a bootstrap cutoff value of 50 was applied. Multiple bootstrap trees were also created for the maximum parsimony tree; one with 30% certainty and one with a 90% certainty. The various trees with different certainties provide insight into the probably relationships between the sequences, and then a description of how likely the relationships really are. After producing the distance matrices and then the phylogenetic trees, we then analyzed a few of the different relationships among breeds of dog using Tajima s D test for neutral selection, also in MEGA. We tested the relationship between the various dog sequences to determine which, if any, type of selection was acting between the various breeds. Along with the selection test between dogs, we also ran the Tajima D test between the three subspecies of wolves. Using all of the data from

MEGA, we were able to visualize and quantify the relationships between the various species and breeds to answer our questions and determine the accuracy of our hypotheses. Results 41 40 34 44 33 69 85 97 97 57 60 70 83 Arctic 93 Arctic 85 33 93 61 Image 1. Maximum Parsimony Tree with Bootstrap Cutoff at 30% Image 2. Maximum Parsimony Tree with Bootstrap Cutoff at 90% These trees show the results of a maximum parsimony analysis with 0 bootstrap replicates. Image 1 represents a condensed tree with a 30% certainty value while Image 2 is a condensed tree cutting off the bootstrap to include only taxa separation above 90. The numbers indicate the percentage of data sets that group these taxa together. This tree shows the MC1R gene in the seven breeds of dogs being the most closely related to the three different species of wolves. Along this branch of the tree is another subtree shows the relationship between the Siberian husky and the Alaskan malamute. The next closest related species to the dogs and wolves are the raccoon dog, artic fox and fox. Without the 30% cutoff, the tree shows a subtree that relates the gene in the fox to the Arctic fox. The 30% tree shows that the bear and Weddell seal have a similar form of MC1R but because this number is quite low, we refer to the second tree with the cutoff showing that they are just a related to each other as they are to the

jaguarundi, fisher, killer whale and chimpanzee. One surprising result is the relationship between the Tasmanian devil and the alligator. All 0 bootstrap replicates showed a further degree of similarity between the Tasmanian devil and alligator than to anything else. This subtree is also connected to a group containing the human and chimpanzee in 61% of the bootstrap replicates. 99 99 59 62 74 91 74 89 68 67 51 86 Arctic 54 90 Arctic 96 84 69 94 91 64 65 Image 3. Jukes and Cantor Model Maximum Likelihood Tree Image 4. Kimura Model Maximum Likelihood Tree Image 3 and 4 are both maximum likelihood trees one run using the Jukes and Cantor model and the other using Kimura s model. The trees were run with 0 bootstrap samples and have a cutoff value of 50. Both trees group together dogs, wolves, foxes and the raccoon dog. The bear, Weddell seal, jaguarundi, fisher, and killer whale are all on separate branches while the chimpanzee and human are on a separate branch, with the gopher, and mouse sharing their own as well. Like maximum parsimony, these trees also show in all the bootstrap replicates, that the Tasmanian devil and the alligator share a distant common form of the MC1R gene as well as the human and chimpanzee. A deeper look at the dog and wolf branch shows a close relation between the Siberian husky and the Alaskan malamute just as the maximum parsimony tree

showed us. The bootstrap numbers indicate that 89 percent and 91 percent of the replicates placed these two breeds on a separate branch for Kimura model and Jukes and Cantor model respectively. The next closest related to these would be the chowchow however only 74 percent of bootstrap replicates showed this. Another subtree shows the and resulting from a common structure of MC1R in about 600 of the bootstrap replicates. 57 59 56 64 89 90 53 53 91 57 90 93 51 64 Arctic 92 54 67 Arctic 92 61 65 89 60 68 99 99 85 82 Image 5. Jukes and Cantor Model UPGMA Tree Image 6. Kimura Model UPGMA Tree Tree 5 and 6 above are UPGMA trees based on Jukes and Cantor and Kimura s models respectively. Both trees were condensed for bootstrap values lower than 50. The bootstrap numbers correspond to percentages of 0 bootstrap replicates. The results of the UPGMA trees are very similar to both the maximum parsimony and maximum likelihood trees. Dog MC1R genes show a high resemblance to those of wolves with the Siberian husky and Alaskan malamute showing a high similarity to each other. The bootstrap values for the relationship between and are lower than the maximum likelihood trees at only 44 and 52 for the Jukes and Cantor and Kimura models respectively. A subtree shows the jaguarundi, bear and

Weddell seal having similar versions of the MC1R gene however only 60% of the bootstrap replicates showed this making it not very significant. Unlike both maximum parsimony and maximum likelihood, the UPGMA trees do not show the Tasmanian devil and the alligator on their own subtree. Table 1. Jukes And Cantor Model Distance Matrix Table 2. Kimura Model Distance Matrix Table 1 and 2 are distance matrices computed using the Jukes and Cantor Model and Kimura s Model respectively. While the trees show the evolutionary relationship of MC1R in different species, the distance matrices show how similar the gene coding sequences are. Higher numbers correspond to larger difference between the sequences. The upper left section of the matrices shows relatively low numbers which indicate a high similarity between the breeds of

dogs and species of wolves. The comparison between the Arctic wolf, Eurasian wolf, Tibetan Mastiff and the displays all zeros indicating that these sequences are identical. Moving further down the table the numbers elevate slightly but not extremely. The distances between the dogs, wolves, foxes, bear, seal, and jaguarundi are all around a distance of 0.1 or below. The rest of the species have distances that are not only higher compared to the species mentioned above but also between themselves particularly the Tasmanian devil and the alligator. However, the human and the chimpanzee mc1r genes have a low distance between them. m n S π D 7 942 5 0.002831 1.519165 Table 3. Results from Tajima s Neutrality Test on Dogs m n S π D 3 942 0 0 n/a Table 4. Results from Tajima s Neutrality Test on Wolves Table 3 and 4 show the results of Tajima s Neutrality test. The symbols indicate as follows: m- Number of sequences, n- number of sites, S- number of segregating sites, π- nucleotide diversity and D- Tajima Test Statistic. The seven sequences that were run in table 3 were from all seven breeds of dogs; Siberian husky, Tibetan Mastiff, chowchow,, Alaskan malamute,, and. In table 4, the three different sequences were the gray wolf, Arctic wolf and the Eurasian wolf. All three of the wolf sequences were identical leading to a result that was not able to be calculated.

Conclusions The mc1r Gene is Conserved across a Variety of Species Looking at the results from all three tree types we can confirm prior knowledge that dogs and wolves are most similar to each other. The next closest species are the fox, artic fox and the raccoon dog. Across the trees the trend showed large sub grouping of species and within those subtrees, particularly the dog and wolf subtree, the bootstrap values were quite low. The data from the wolf Tajima D test shows a complete conservation of the gene across three subspecies. Adding in data from the distance matrices backs up the close relationship between MC1R in dogs and wolves. These distances between species are also low when also considering the foxes, bear, seal, and jaguarundi. Drawing from this data we can conclude that the MC1R gene is conserved across carnivores with the exception of the fisher, alligator and the Tasmanian devil. As we see from the trees, the alligator and Tasmanian devil resulted from a common ancestor. The and Tibetan Mastiff are the Closest Dog Breeds to Wolves Both the Jukes and Cantor and the Kimura model distance matrices show a section in which there is a complete match between 5 sequences of mc1r. These sequences are from the gray wolf, Arctic wolf, Eurasian wolf, and the Tibetan Mastiff. This data along with the grouping of the taxa on the 6 different trees shows that the and Tibetan Mastiff are the closest breeds of dogs compared to the wolf. mc1r is Under Positive Selection in Dog Breeds Based on the results from Tajima s D test for selection, we see that the mc1r gene in dogs is under positive selection. The result from the test showed a D value of 1.519 which is above 1. Values above 1 are considered to show positive selection for the species. Due to mc1r controlling coat color in dogs we predict that this positive selection may be due to human interaction in selecting the preferable coat color.

Appendix A Species Dog Eurasian Wolf Arctic Wolf Gray Wolf Raccoon Dog Cat Aortic Weddell Seal Tasmanian Devil Killer Whale Dog- Dog- Dog- Dog- Siberian Husky Dog- Dog- Tibetan Mastiff Dog- Alaskan Malamute Accession Number KC332684 JX273639 JX273636 JX273640 JN575070 JX083391 JX083393 AY363625 JF501545 AJ245705 AB725425 AJ786718 XM_006746161 AY237399 XM_003759223 XM_006269398 FJ773303 EF488834 FJ389441 JX273625 JX273612 JX273609 JX273579 JX273599 JX273619 JX273591

References Rees, J.L. (2000) The Melanocortin 1 Receptor (MC1R): More Than Just Red Hair. Pigment Cell Res 13, 135-140 Jones, M., Cockburn, A., Hamede, R., Hawkins, C., Hesterman, H., Lachish, S., Mann, D., McCallum, H., Pemberton, D. (2008) Life-history Change in Disease-ravaged Tasmanian Devil Populations. Proceedings of the National Academy of Sciences of the United States of America 105(29) 23-27 Kyle, C.J., Robitaille, J.F., Strobeck, C. (2002) Genetic Variation and Structure of Fiser (Martes pennanti) Populations Across North America. Molecular Ecology 10(9) 2341-2347 Lindblad-Toh, K., Wade, C.M., Mikkelsen, T.S., Karlsson, E.K., Jaffe, D.B., Kamal, M., Clamp, M., Chang, J.L., Kulbokas, E.J., Zody, M.C., et al. (2005) Genome Sequence, Comparative Analysis and Haplotype Structure of the Domestic Dog. Nature 438, 803-819 Rees, J.L. (2000) The Melanocortin 1 Receptor (MC1R): More Than Just Red Hair. Pigment Cell Res 13, 135-140 WAZA (2014) Raccoon Dog, Tanuki. World Association of Zoos and Aquarieums Tofoli, C.F., Rohe, F., Setz, E.Z.F. (2009) (Puma yagouaroundi) (Geoffroy, 1803) (Carnivor, Felidae) Food Habits in a Mosaic of Atlantic Rainforest and Eucalypt Plantations of Southeastern Brazil. Brazilian Journal of Biology 69(3) 871-87