A -defensin mutation causes black coat color in domestic dogs

Abstract

Genetic analysis of mammalian color variation has provided fundamental insight into human biology and disease. In most vertebrates, two key genes, Agouti and Melanocortin 1 receptor (Mc1r), encode a ligand-receptor system that controls pigment type-switching, but in domestic dogs, a third gene is implicated, the K locus, whose genetic characteristics predict a previously unrecognized component of the melanocortin pathway. We identify the K locus as beta-defensin 103 (CBD103) and show that its protein product binds with high affinity to the Mc1r and has a simple and strong effect on pigment type-switching in domestic dogs and transgenic mice. These results expand the functional role of beta-defensins, a protein family previously implicated in innate immunity, and identify an additional class of ligands for signaling through melanocortin receptors.

Fig 1

Genetic mapping of the K locus. (A) Initial linkage studies [phase 1, (14)] defined a 12-Mb critical region for K^B; ascertainment and characterization of additional kindreds narrowed the interval to 3.8 Mb (phase 2, figs. S1 to S4). Association analysis for 60 markers in brindle (n = 12) versus yellow (n = 10) Boxers, and for 51 markers in black (n = 9) versus yellow (n = 10) Great Danes, was carried out as described in the text. (B) Candidate genes in the 320-kb region of greatest association in Great Danes; this region includes 12 β-defensin genes (shown in red). (C and D) Significance, plotted as −log of P values from a chi-square test of allele counts, is shown as a function of distance along CFA16 (only for SNPs present at greater than 10% frequency and genotyped in at least 75% of the samples). The dashed red line indicates a Bonferroni-corrected 5% significance level; these regions are indicated by hatched and black bars for Boxers and Great Danes, respectively, in (A). Annotation is based on the Non-dog RefSeq track in the UCSC Genome Browser, except for CBD102, identified by Patil et al. (16).

Fig 2

Resequencing and recombinant haplotype-based mutation analysis for K^B-bearing versus k^y-bearing chromosomes. (A) A 20-kb region surrounding CBD103ΔG23 was resequenced (except for repetitive regions) in 10 dogs from 7 breeds, and haplotypes were inferred for 28 high-frequency biallelic polymorphisms. Blue and yellow squares represent the major and minor alleles in K^B-bearing chromosomes, respectively, and allow some haplotypes to be designated as “k^y-parental,” “K^B-parental,” or “proximal recombinant,” as indicated. White squares represent missing data. Genotypes for five Great Danes (denoted by asterisks) were determined in a second resequencing round targeted specifically for distal recombinants as described in the SOM text. Within the 9.1-kb interval defined by recombinant haplotype analysis, three polymorphisms are completely associated with K^B versus k^y, as indicated in the upper part of the figure. (B) Exon structure of transcripts within the maximal candidate interval and alignment of selected CBD103 orthologs (38).

Fig 3

Expression of β-defensin mRNA and protein in skin and in cultured keratinocytes. (A) Levels of Agouti, CBD1, or CBD103 mRNA from black or yellow dog skin, as indicated, determined relative to Bactin by quantitative RT-PCR, and expressed as percentage of mRNA present in the yellow samples. Results shown represent the mean ± SEM of four different animals. (B) Expression of epitope (V5)–tagged CBD103 (+) or epitope (V5)–tagged CBD103ΔG23 (ΔG23) in cell layer and media after transfection of mouse keratinocytes as determined by Western blotting with antisera against the V5 epitope. Representative results are shown for one of four experiments; for each experiment, the two constructs were transfected in triplicate or quadruplicate.

Fig 4

Pigmentary effects of CBD103 in transgenic mice. Photographs of transgenic (Tg) and non-transgenic littermates, representative of 2/2 independent founders for Tg.CBD103ΔG23 and 20/21 independent founders for Tg.CBD103.

Fig 5

Pharmacology of β-defensin action on melanocortin receptors. (A) Ability of NDP-MSH or CBD103 to stimulate cAMP accumulation in cultured melanocytes. (B) Saturation binding of Eu-NDP-MSH to HEK293 cells transiently transfected with the dog Mc1r. (C to F) Competition binding assays in which varying amounts of unlabeled synthetic β-defensins were added together with Eu-NDP-MSH tracer (at 1.8 to 3 nM) to HEK293 cells transiently transfected with dog (d), mouse (m), or human (h) melanocortin receptors, as indicated. In all panels, ligand concentration {either in nM (B) or log of the molarity [(A) and (C) to (F)]} is plotted on the abscissa; amount of cAMP formed (A) or Eu-NDP-MSH bound [(B) to (F)], measured, respectively, as relative light units (RLUs) or relative fluorescent units (RFUs), is plotted on the ordinate. Each curve represents a single experiment carried out in triplicate; error bars represent SEM.