Much of what we know about pigment cell biology comes from studying laboratory mice, yet the conservation of gene action and interaction across domesticated animals, including horses, cattle, pigs, and dogs, is firm evidence that most aspects of color variation are conserved in mammals. The domestic cat, however, is a special case in which variation within and among breeds exhibits a fascinating glimpse into an entirely new area of color variation as exemplified by tabby patterns. Periodic patterns with color markings spaced at non-random intervals are common in nature and evident in all major mammalian orders – stripes on chipmunks, reticulated markings on giraffes and hyenas, tail rings on lemurs, and, of course, tabby stripes on domestic cats. The similar nature but diverse manifestation among these patterns suggests they are formed by a conserved, adaptable, and largely unexplored mechanism. Periodic patterns are conspicuously missing from the laboratory mouse, but the cat stands out as a unique genetic model in which recent work from (Eizirik et al. (2010)) provides new genetic insight.
Domestic cats have four distinct and heritable coat patterns – ticked, mackerel, blotched, and spotted – that are collectively referred to as tabby markings (Figure 1). These patterns are a composite of two features: (i) a light background component resulting from individual hairs with a subapical light-colored band and (ii) a superimposed darker component resulting from unbanded hairs. The ticked phenotype refers to the absence of any superimposed pattern, leaving only the banded or ‘ticked’ background color. Mackerel, blotched, and spotted phenotypes describe variations of the tabby pattern, in which the darker component forms either periodic vertical stripes (mackerel), whorls (blotched), or leopard-like spots (as in the Ocicat or Egyptian mau breeds of domestic cats). The periodicity of tabby markings distinguishes them in a fundamental way from other characteristic but randomly displayed markings, such as the tricolor patches on a calico cat or the black spots on a Dalmatian. These random patterns arise from events that are stochastically initiated (like inactivation of an X chromosome in females or survival of a melanocyte cluster) but stably maintained, either through epigenetic mechanisms as with X inactivation, or a developmental process, as with a limited window for neural crest migration. By contrast, periodic patterns must arise from a mechanism that is specifically programed to be spatially constrained.
What are the molecules and cells that underlie tabby patterns? A genetic approach to this question began nearly a century ago when Phineas Whiting described three tabby ‘banding factors’, noting a simple pattern of inheritance for each: ticked is dominantly inherited relative to mackerel and blotched, whereas blotched is recessively inherited relative to ticked or to mackerel (Whiting, 1918). Whiting posited a single locus (T) with three alleles – ticked (Ta), mackerel (T+), and blotched (tb). That view has now been revised by the recent work of Eizirik et al. (2010), who applied a genome-wide panel of molecular markers to pedigrees segregating the different coat color patterns, and thereby discovered that two loci were involved in determining the difference between ticked, mackerel, and blotched patterns (Figure 1). In this modern view of cat pattern genetics, the Tabby locus determines the type of pattern, with the TaM (Mackerel) allele dominant to the Tab (Blotched) allele, while a second locus, now known as Ticked, determines the absence or presence of pattern, with the TiA allele dominant to the Ti+ allele. The nomenclature has become complicated; similar to Eizirik et al. (2010), we use lower case roman to describe the phenotypes and/or underlying processes, e.g. tabby, ticked, mackerel, and blotched, and capitalized italics to refer to the loci and/or alleles, e.g. Tabby, Ticked, Mackerel, and Blotched. As described later, the allele of Ticked associated with the presence of pattern is probably ancestral, and thus designated Ti+, while the allele of Ticked associated with the absence of pattern, as in Abyssinian cats, therefore TiA, is probably derived.
Strategies to identify the Tabby and Ticked genes are discussed further in the following paragraphs, but before doing so, it may be helpful to consider what types of developmental and cellular mechanisms might be involved. (Eizirik et al. (2010)) suggest that mammalian coat patterns are formed by two distinct processes: (i) a ‘spatially oriented’ mechanism that establishes a prepattern in the skin and (ii) a ‘pigmentation-oriented’ mechanism that regulates expression of pigmentary genes in response to the prepattern. Building on this idea, we propose that distinct ‘spatially oriented’ and ‘pigmentation-oriented’ mechanisms provide an appealing conceptual framework to explain why coat patterns occur commonly but not ubiquitously in mammals. Perhaps, a complex, ‘spatially oriented’ mechanism developed prior to or early in mammalian evolution, and diverged to give rise to various types of prepatterns. In contrast, a ‘pigmentation-oriented’ mechanism could have occurred independently several times during mammalian evolution. In other words, white and black stripes on a zebra might depend on the same ancient prepattern as do yellow and black stripes on a tiger, but the specific pigmentary genes that interpret the prepattern could be completely different.
Most ideas about spatial prepatterns in large animals stem from the mathematical work of Alan Turing on reaction–diffusion processes, which have applications not only in biology but also in chemistry and physics. Interaction of two diffusible substances to yield predictable and periodic patterns during skin development forms the basis for theoretical work that could explain a large range of mammalian coat patterns (Bard, 1981). Practically speaking, melanoblasts constitute only a small fraction of the cell population in the developing skin, and therefore seem an unlikely substrate for propagating a reaction–diffusion signal. (Furthermore, tabby stripes occasionally appear to cross areas of skin devoid of melanocytes.) Instead, a putative ‘spatially oriented’ mechanism seems more likely to be mediated by fibroblasts and/or keratinocytes in the developing skin, in which widely expressed components of a reaction–diffusion system come to delimit prepattern boundaries. Genetic variants affecting this process—perhaps alleles of the Tabby gene—could change the shape and/or location of the boundaries, thereby giving rise to stripes or blotches (Figure 2).
What about a ‘pigmentation-oriented’ mechanism in tabby striping? Dark areas in a patterned animal, whether stripes, spots, or blotches, are caused by conversion of a light band to a darker surrounding color on individual hairs. At first glance, prime candidates for this phenomenon are the paracrine signaling molecule Agouti protein and its target the melanocortin 1 receptor (Mc1r), which controls the temporal production of eumelanin and pheomelanin during hair growth. However, the persistence of tabby markings in cats with an Agouti loss-of-function mutation (a/a) is strong evidence that Agouti does not mediate pattern formation in domestic cats. These ‘ghost’ tabby markings are sometimes visible on young, non-agouti cats as darker patterns on a black background. In fact, non-agouti cats can be orange in color because of mutation of the as-yet-unidentified gene sex-linked orange, and in this background, persistence of tabby stripes on a non-agouti background is even more pronounced (Searle, 1968). Recent evidence also points away from Mc1r as a candidate for mediating tabby stripes.(Peterschmitt et al. 2009) report that amber coat color in Norwegian forest cats is caused by an Mc1r loss-of-function and, furthermore, that tabby markings on amber cats appear solid black on an otherwise amber background. The idea that tabby markings persist when either Agouti or Mc1r signaling is disrupted leaves an intriguing alternative hypothesis – that an as-yet-unidentified signaling system, with the capacity to override the pigment-type switching pathway, can regulate melanogenesis in response to a ‘spatially oriented’ prepattern. If so, identification of the Ticked gene might reveal a novel set of inputs that affect pigment cell behavior (Figure 2).
(Eizirik et al. (2010)) suggest that Tabby may be a component of the ‘spatially oriented’ mechanism, because Tabby alleles change the shape of the pattern, and that Ticked may be a component of either mechanism, because the loss of the pattern could result from disruption of either the ‘spatially oriented’ mechanism or the ‘pigmentation-oriented’ mechanism (Figure 2). Using linkage mapping in pedigree individuals, they identify discrete genomic intervals for both genes, containing 40 and 16 genes, respectively. There are no ‘smoking gun’ candidate genes in either interval, which is bad news for those hoping for a quick and simple answer, but good news to those hoping to apply pigmentary variation to fundamental questions in developmental biology.
Besides stripes and blotches, how (and why) do related patterns arise, such as cheetah spots or leopard rosettes? Part of the answer could involve hybridization. A famous example is the ‘liger’, the offspring of a male lion and female tiger, whose coat pattern is distinct from either tiger stripes or the rosettes characteristic of juvenile lions, and might best be described as ‘broken mackerel’. Amazingly, (Eizirik et al. (2010)) describe an analogous situation in domestic cats, in which a cross between true-breeding spotted and blotched cats produced ‘broken mackerel’ patterns that were intermediate between stripes and spots (Figure 2). Genetics in ligers is difficult, but when (Eizirik et al. (2010)) backcrossed their broken mackerel F1 animals to blotched cats, they observed an equal ratio of blotched and non-blotched cats. The non-blotched cats, however, displayed a range of patterns extending from fully striped to fully spotted. Thus, while the genetic architecture of spotting remains elusive, we learn something about the interaction of spots, stripes, and whorls from this cross – in the appropriate genetic background, stripes are converted to spots, but whorls are not. This selective pattern conversion indicates that a spotting gene (or genes) is capable of interacting with TaM but not Tab.
Molecular identification of Ticked, Tabby, and spotting gene(s) should be rewarding to the evolutionary as well as the developmental biologist, because mechanisms homologous to those in tabby cats seem likely to generate the stripes, spots, and intricate rosette patterns that distinguish the 37 extant species of wild cats. Their recent evolutionary divergence (within the last ∼12 million years), well-characterized phylogenetic relationships, and range of species-specific coat patterns make the larger cat family an attractive model for comparative genomics once the patterning genes in domestic cats are identified.
Going forward, a population-based mapping strategy provides an attractive approach for narrowing the Ticked and Tabby candidate intervals. In fact, cat populations have been recognized for a long time as being particularly amenable to such approaches. In two studies published more than 50 years ago, the British geneticist A.G. Searle compared traits of ‘alley cat’ populations in different parts of the world where he spent time as a professor (Searle, 1949, 1959). He noted equal ratios of blotched and mackerel but very few ticked cats in London, whereas, in Malaya, Singapore, he noted equal ratios of ticked and mackerel cats but very few blotched cats. Searle took the skewed distributions of tabby patterns as evidence that (i) cats descended from a mackerel-patterned ancestor (the presumed wild ancestors of domestic cats, Felis silvestris or Felis lybica, are both striped) and that (ii) ticked and blotched alleles were derived after cat domestication in different parts of the world. Placed in context of the present study, the TiA and Tab alleles represent derivative variants at their respective loci, and the regions surrounding the causative mutations should be defined by reduced allelic variation relative to ancestral alleles.
Of course, the ultimate goal is not only to identify coat pattern genes but also to learn how they function. The laboratory mouse, with an array of well-characterized pigmentation mutants and a robust toolkit for genetic manipulation, remains the ideal mammalian model for studying gene function. If a ‘spatially oriented’ mechanism is conserved across mammalian species, then engineering a striped or spotted mouse might require simply engaging the right ‘pigmentation-oriented’ pathway. Such a strategy does not seem so impractical, considering that the striped African mouse (Rhabdomys pumilio), a close relative of the laboratory mouse, is…well…striped.
The recent work by (Eizirik et al. (2010)) was motivated by the longstanding interest and focus of O’Brien and colleagues on the domestic cat as a model genetic system and builds on more than a decade of comparative cytogenetics, radiation hybrid maps, BAC libraries, and, more recently, next-generation sequencing for SNP discovery. Lagging a few years behind the domestic dog, genetic and genomic resources for the cat are now coming on line and promise molecular exploration of frontiers in developmental and evolutionary biology that have been widely admired but poorly understood.