With the excellent help and painstaking effort of Deb Jones
and the DeVine Farms Pedigree program (money
from it's purchase goes towards Mastiff Rescue), the following
numbers have been generated. The table below (includes stats through December 2001) shows the color and
number of all the pups that resulted from the matings described
in the first two columns. The number in parenthesis indicates
the percent of that color pup of total pups within that mating.
These data show that there were 104 brindle pups (2+22+11+69)
that did not have a registered brindle parent. Those who believe
a brindle pup must have a brindle parent could argue that the
Mastiffs (parent or pup) were registered incorrectly, either by
accident or purposefully. This may be true. However, I'd ask,
"Why are only the brindles misregistered
?" Shouldn't it be just as likely that the fawns and apricots be misregistered
so that these could then 'make up' for the incorrect brindle
registrations ? I don't have the answer, wish I did - I'm only
posing the question.
Within the 104 brindle pups, there were 71 different sires, 76
different dams (three specific sets of parents
were responsible for 14 of these brindle pups) and 273
different grandparents (of the possible 2x 71+76
= 294). Moreover, of the grandparents (five
were of unknown color), there were 42 apricot (16%),
173 fawn (64%) and 53 brindle (20%)
which closely mimics the color distribution within the general
Mastiff database of 3371 apricot (17%),
11907 fawn (60%) and 4440 brindle
(22%). Thus, there does not appear to be any specific
color lineage associated with the 104 brindle pups from non-registered
brindle parents. Moreover, do not be mistaken into believing that these brindle pups are merely ancient history
because, since January of 1998, there have been 18 brindle pups born from non-brindle registered parents according
to the Pedigree database. I have not contacted the owners of these pups to enquire as to their DNA certification.
Not withstanding, the results from this analysis are also internally
consistent. That is, color inheritance, as shown in the table,
is clearly not sex linked. For example, an apricot x fawn or a
fawn x apricot mating produce nearly an identical percentage and
number of fawn, apricot and brindle pups. Furthermore, this can
be demonstrated with all matings between parents of dissimilar
color.
As an aside, there are 4763 mastiffs listed that have no color assigned, and another
72 with some unusual designations including blue brindle (2), apricot fawn (7), fawn apricot (5),
red (4), red fawn (6) and black (4).
The graphs below show the percentage of each color born within the total population (ie. listed)
on a monthly and yearly basis since 1965. Note that there is no good reason to believe that
color and month of birth should show any correlation. I was merely curious.
During embryogenesis specific cells (melanoblasts) migrate from the neural crest [2] into the basal epithelium of the epidermis, hair bulbs of the skin and specific areas of the eye, ear and brain. Migration tends to be directionally from head (rostral) to tail (caudal) and top (dorsal) to bottom (ventral) [3]. One only needs to look at the final color pattern in dogs to see this directional effect. For example, both the vertical stripes of the brindling pattern and the tips coloring pattern (e.g. tail, paws) which may be very different in color from the main body, are both reminiscent of the melanocyte precursor migration pathway. If, during embryogenesis, there is a shortage of melanocytes, particular areas of the body receive priority (e.g. eye, brain and ear). However, the specific survival, proliferation, migration and final differentiation of these melanoblasts into melanocytes is influenced by several paracrine factors, including basic fibroblast growth factor, steel factor [4](a.k.a. mast cell growth factor or stem cell factor or c-kit ligand) and endothelin-1 [5] which may be released by cells (e.g. keratinocytes and fibroblasts) in the epidermis. For color production, the melanocytes synthesize a chemical pigment, melanin, which is deposited onto the structural fibullar component gp100, located in subcellular organelles called melanosomes. However, the follicular microenvironment (i.e. hair bulb area) rather than the melanocyte genotype appear to regulate which type of melanin synthesis (i.e. eumelanogenesis or pheomelanogenesis) predominates [6]. Once the melanosomes are packed with either eumelanin or phaeomelanin, these melanosomes are translocated over the melanocyte cytoskeleton within the developing hair shaft using myosin V through dendritic processes to an area where the melanosomes maybe taken up by keratinocyte phagocytosis [7]. Each melanocyte may have as many as 36 keratinocytes associated with it. Slowly, the keratinocytes are pushed up the hair shaft as cells near the base of the hair fiber (i.e. dermal papilla) continue to multiply. This is a perpetual process employing an array of different proteins so that new melanosomes can be continuously produced and released into adjacent keratinocytes which are then pushed up the shaft. Consequently, color differences are effected, in part, by melanosome movements within the melanocyte. For some excellent figures and description of the developing hair shaft and melanocyte involvement visit the links at Hair Biology and Hair Color.
Melanin is not a simple molecule but actually a polymer synthesized
within the melanosome using two different types of 'building blocks'.
One, eumelanin, responsible for the black to brown coloration
and another, phaeomelanin, responsible for the yellow to red coloration.
Both eumelanin and phaeomelanin are formed from the precursors
dihydroxyphenylalanine (DOPA) and tyrosine. These precursors are
enacted upon by tyrosinase (see C locus), the first enzyme and the rate limiting
enzyme in the melanin synthesis pathway, which catalyzes the formation
of dopaquinone. Whether, dopaquinone proceeds toward eumelanin
(black) or phaeomelanin (red) appears to be dependent upon both
the level of tyrosinase activity and the presence of thiols (e.g.
cysteine and glutathione) [8]. If
tyrosinase activity is low relative to cysteine levels, then phaeomelanogenesis
proceeds; if tyrosinase activity is high, or cysteine levels low,
then eumelanogenesis. Tyrosinase variants tend to be less active than the native form and lessen the
depth of black or red color. For an excellent review of pheomelanogenesis see Wolff [9].
Two other enzymes besides tyrosinase have been described specifically
in the eumelanogenesis pathway, tyrosinase-related protein-1 (TRP-1, see B locus) [10]
and tyrosinase-related protein-2 [11] (TRP-2; a.k.a. dopachrome tautomerase). These enzymes may help to stabilize
tyrosinase by forming a multi-enzyme complex in the melanosome [12]. Thus if any of these proteins are not in
their 'wild-type' or native form, production and depth of black pigment will suffer. This has been shown in a study
by Beermann et al [28] in which mice devoid of TRP-2 resulted in a phenotype similar to the slaty mutation (i.e. greyish).
Melanin synthesis (or melanogenesis) can be modulated primarily by the opposite
action of two intercellular signaling molecules, alpha-melanocyte
stimulating hormone (MSH) and agouti protein (see A locus)[13].
MSH and adrenocorticotropin hormone (ACTH) [14], whose from a family of related molecules cleaved from the
precursor pro-hormone pro-opiomelanocortin, can up-regulate
melanin synthesis by its interaction (i.e. binding) with the melanocortin 1 receptor
(MC1R, see E locus) located on melanocytes. The MC1R is a
G-protein coupled receptor such that the binding of MSH (or ACTH)
triggers a cAMP dependent signal cascade leading to an increase
in the activity of tyrosinase. As discussed above, an increase
in tyrosinase activity results in a preferential increase in the
synthesis of eumelanin and an increase in the eumelanin:phaeomelanin
ratio. So, a wild type MC1R would allow full color development. On the other hand, the switch from eumelanin to
phaeomelanin biosynthesis
can be stimulated by the second intercellular molecule, agouti
protein, which is secreted from dermal papilla cells
and produced by neighboring keratinocytes. In the hair follicle, agouti protein inhibits the binding
of MSH to MC1R, thereby decreasing tyrosinase activity and in effect, increasing the redich pigment. Agouti
protein has also been shown to down-regulate genes necessary for
eumelanin synthesis [15].
In the mouse, more than 100 (~127) different genes have been shown
to affect coat color (see
The Coat Colors of Mice). The proteins encoded by these genes include
DNA transcription factors which regulate the
rate at which specific genes produce their proteins (e.g. MITF locus, microphthalmia transcription
factor which plays a role in melanocyte survival and differentiation),
receptors and their ligands (e.g. MSH and MC1R
or steel factor and cKIT), structural proteins (e.g.
lysosome associated membrane protein), transport proteins (e.g. myosin V)
and enzymes. For example, there is a protein which is necessary
for the transport of cysteine into the melanosome. Since cysteine
is necessary for phaeomelanogenesis, this transport protein would
be required for red pigmentation. Yet, up to now, no one has been
able to identify the gene locus responsible for this transport
protein. Consequently, those proteins described above are an attempt
to simplify melanogenesis (though it might not
seem that way) and to allow the reader to more easily
put the genes (described below), and the proteins they encode,
into the proper perspective as to how they may affect color formation. Generally speaking, in order of phenotypic
expression, proteins which control melanocyte distribution (e.g. S, T and W loci) will have
a more visible effect than those proteins which control melanin synthesis
(e.g. E and C loci) and its regulation (e.g. A locus) which will, in turn,
have a more visible effect than those proteins effecting melanosome
structure and transport (e.g. D locus ?). Thus, different colors and patterns can be achieved by genes controlling
the distribution of melanocytes, the activity of melanin synthesis
and the transport of melanosomes to the keratinocytes.
Gene locus
For the entire dog genome visit the Dog Genome Project .
But for color specifics, there are some very good web pages by Dr. Sheila Schmutz (Genetics of Coat Color in Dogs) and Sue Ann Bowling
(Animal Genetics) which discuss many of the alleles in depth. For more breed specific canine
color inheritance visit the excellent list at Canine Color Genetics Links by Liisa Sarakontu where there are close to 80 different breeds listed.
Here, the journey begins for the English Mastiff. I have seen
as few as 9 and as many as 15 gene loci primarily used
to describe the inheritance of color in the canine.
| A - agouti | F - rufus | P - pink eye |
| B - brown/black | SE - super extension, black mask | R - roan |
| C - concentration | K - brindle (?) | S (or W ?) - spotting; distribution of color |
| D - dilution | G - graying | T - ticking |
| E - extension | M - merle | W - proposed, white |
The three AKC standard acceptable Mastiff colors, fawn, brindle and apricot are shown below. The middle brindle photo below was kindly provided by Gatehouse Mastiffs (Carrie K, copyright 2003).
| A - agouti; encodes the agouti protein which inhibits the interaction of alpha-MSH with MC1R and down regulates genes necessary for eumelanin production. | ||||
| As | - | - | - | |
| Ay | restricts B series and produces phaeomelanin; incompletely dominant to at; ayay could be clear red, dependent upon modifiers while ayat would be, on the average, darker with more black hairs either tipped or interspersed and may darken with age; see A locus red | tan, sable, fawn | Mastiff, Collie, Boxer, Great Dane | |
| at | restricts tan (phaeomelanin) to certain points (e.g. muzzle, over the eyes, chest, all four boots) | black & tan | doberman, rottweiler, Bernese mountain dog | |
| aw | individual hair fibers are banded; wild type, wolf color | salt & pepper | Norwegian elkhounds, Afghan, shih-tzu | |
| asa | saddle tan; restricts dark color to saddle pattern | - | - | |
| a | recessive black; either no agouti or inactive agouti protein may be produced | - | Shetland sheepdog, German shepherd | |
| An all black coat* and the at* pattern has been seen in Mastiffs, though these are not approved by the modern standard. Recently, the agouti locus has been excluded as a possible site for 'dominant black' by Kerns et al. [17] | ||||
| B - brown/black; encodes Tryp1, a 538 amino acid protein located on dog chromosome 11 [18], which helps to stabilize tyrosinase and thus the polymerization of melanin; possibly effects the structure of the melanosomes and the distribution of the eumelanin pigment; may regulate basal levels of melanin formation. Schmutz et al. [18] describe three main sequence variants; S41C, Q331ter and 345delP. Their suggested allele designation is shown in the table below. | ||||
| B | wild type, large, densely packed melanosomes | black | Mastiff | |
| bs | Q331ter, stop codon | brown/liver | chocolate lab | |
| bd | 345delP, proline deletion | brown/liver | chocolate lab | |
| bc | S41C, cysteine replaced by serine | brown/liver | chocolate lab | |
| B or b can be told by the color of the points (i.e. nose leather, lips, eye rims and usually, foot pads). Tyrp1 can make a mask brown. All dogs are either black or liver, regardless of the coat color. Schmutz et al suggest that black dogs have no more than one of these variants and brown dogs have two or more. But, there is no obvious difference in shade of brown in relation to which variant is present. When variants are present melanin is significantly less polymerized and melanosomes are smaller, less densely packed. Can produce what may be described as reds in Australian shepherd, border collie, Doberman, pharaoh hound and Ibizian hound; atatbb = red Doberman; bbcchcch = dead grass color seen in Chesapeake Bay retriever; bb* appeared in Crown Prince. | ||||
| C - concentration; encodes tyrosinase, the critical enzyme in melanin synthesis; primarily noticeable by its affect on phaeomelanin; blacks become silvery gray and reds become cream to off-white. Has been linked to dog chromosome 21 [19]. | ||||
| C | the structural gene for wild type tyrosinase; allows the full depth of pigment | solid color | most breeds, Mastiff | |
| cch | reduces the red (phaeomelanin) pigment to cream with little effect on eumelanin. Very light tan dogs are probably cchcch. | chinchilla | pale yellow lab | |
| cd | inhibits production of red pigment. | white coat, black eyes | Samoyed | |
| ce | extreme dilution or cb (Burmese) | pale gray coat, pale blue eyes | Pekinese, Pomeranian | |
| ca | inactive variant of tyrosinase, lack of melanin; albinism | white coat, pink eyes | not known to occur in dogs | |
| The gene for tyrosinase is very mutagenic. There is a temperature sensitive variant which is displayed in rabbits (Himalayan; equivalent to Siamese in the cat) such that the enzyme works at temperatures below the normal body temperature. ceceee maybe West Highland terrier | ||||
| D - dilution; in mouse encodes the motor protein myosin V, but most likely not myosin V in dogs; responsible for transporting the melanin containing melanosome through the dendritic processes toward the keratinocytes; effects both eumelanin and phaeomelanin intensity of pigmentation | ||||
| D | allows full dense pigmentation | self color | - | |
| d | clumping of melanosomes leading to reduced light absorption; flattening or dulling of color | bluish gray or silver | weimararner, blue Great Dane | |
| This is probably the primary gene that will effect the coloration of the points. Though B, TYRP1, can also show some effects in the mask and BBdd will be bluish gray while bbdd will be silver. If black patterns are present they also will change to blue, e.g. blue mask, blue tipping, blue brindle. dd * may produce blue brindles in mastiffs. | ||||
The following phenotypes (i.e. brindle, black mask and dominant black) are among the most misunderstood as to which gene locus they reside. Early studies by Little [20], and even some present inheritance models, put brindle, black mask and recessive red all at E (extension locus). However, some recent investigators agree that black mask and brindle cannot be in the same series since, for example, a mating of two black masked, brindle mastiffs would produce 25% offspring without a mask, which doesn't happen. S. Pober postulates a K locus where dominant black, brindle and wild-type exist. Yet, other studies involving many other mammals put dominant black at the E locus. Kerns et al.[17] have recently excluded the E locus (MC1R) as a candidate for dominant black in dogs (e.g. black lab, Newfoundland). It's quite likely that there is some confusion with nomenclature and semantics. For 'simplicity' sake, I'll use two loci to describe these phenotypes - even though four maybe incorrect, since E (extension) is the only universally recognized locus at this time.
| E - extension; encodes the expression of MC1 receptor in dog [21]; located on chromosome 5 [22]; | ||||
| E | wild type MC1R (317 residues, sequence in ref. 23), allowing eumelanin over the entire body as the A locus dictates | self coloring | - | |
| EM | melanistic mask, color dependent upon other genes effecting eumelanin; all dogs with mask have at least one copy of a valine substitution for methionine at amino acid 264 (M264V) [24]; demonstrates that a MC1R mutant may be spatially restricted (also seen with Etob, dorsal restriction, in mice) - mechanism unknown | brown or black mask | Mastiff, Boxer, Akita | |
| e | recessive red; encodes an inactive MC1R variant (R306ter); a small number of dark (i.e. eumelanin) hairs can be found dorsally and eumelanin can be found in the skin [21]; dark hairs my occur due to high basal (i.e. unstimulated) activity of eumelanogenesis, thus precluding the need for a functional MC1R ; | red, orange, yellow | Irish setter, Golden retriever, yellow lab | |
| In some breeds (e.g. Great Dane, cocker spaniel, Red chow, Samoyed) reds may be either ayay or ee In general, the ay produces more yellow-tan while the e more reddish tones. The ee will not change as much from birth as the sable. There is also an S90G variant but this is merely a polymorphism and has no known effect on color. In addition to black mask and Etob, two recessive alleles have been postulated which yield a partially distributed pigmentation with the appearance of brindle or tortoiseshell patterns (eb in guinea pig and et in rabbit) by Ibsen in 1919 [25], Chase in 1939 [26] and Punnet in 1924 [27]. | ||||
| K - brindle ??? | ||||
| K | dominant black; same as As in agouti locus as postulated by C. Little | - | - | |
| kbr | allows brindle | - | Mastiff, boxer | |
| k | wild type, no brindling | - | most breeds | |
The next three series (S, T and W) appear to have some phenotypes
which could appear to be similar. Also, trying to equate these
loci and the genes they encode with the mouse (or dog) nomenclature has
been difficult (or impossible). For the sake of consistency,
I try to continually refer to the mouse nomenclature. Mouse nomenclature may take precedence over the nomenclature
suggested for dogs 40 years ago since most future molecular biology
and gene identification will proceed in the mouse or human (my
opinion). In the mouse, the following genes are described:
Each of these genes and their encoded proteins (endothelin-B receptor, steel factor and cKIT) are required for proper melanocyte development and migration. Thus, mutant alleles at any of these loci could be responsible for depigmentation and white color patterns. To add more confusion, even though there is an 's - piebald' locus (endothelin-B receptor), most piebaldism in the mouse results from mutations at the loci encoding either steel factor or cKIT. Consequently, does the S locus (white spotting - described in dogs) encode cKIT or does W (in dogs) encode steel factor or does T encode endothelin-B receptor ?
| S - spotting (or W -white spotting ?) ; controls melanocyte distribution | ||||
| S | an even distribution of color with possible, very minor, white markings on the toes, tail tip and chest | - | - | |
| si | Irish spotting; some what symmetrical; limited to muzzle, neck, chest, feet, legs and tail; no white on rump | - | collie, Boston terrier | |
| sp | piebald; large white spots, generally greater than 50% is white, may cross the back | - | cocker spaniel, shar-pei | |
| sw | extreme white piebald; nearly entire dog is white with only patches of color around the head, eyes and ears | - | white bull terrier, great pyrenees | |
| This series is not completely understood since it is unclear how many alleles occur. The series is greatly affected by other modifying genes. A more extensive discussion of this series can be found at Animal Genetics and Borzois. | ||||
| T - ticking; distribution of melanocytes to specific areas | ||||
| T | allows small areas of self coloring within white areas; ticking is not visible at birth | - | dalmation | |
| t | no ticking | - | - | |
| W - white; proposed to inhibit pigment production | ||||
| W | dilutes gray color of merle to white; lethal in the homozygous state | - | Great Dane | |
| w | non-clearing | - | - | |
Finally, since my investigation into the area of color inheritance in the English Mastiff is an ongoing dynamic project, I fully expect that some of the text here will eventually be proven incorrect. However, I feel that publicly displaying these data are imperative to receiving the feedback essential to altering this working model accordingly. With the recent advances made in the field of molecular biology, it's only a matter of time before we have the answers necessary for a definitive color inheritance model. For example, there are many other proteins and genes recognized that have yet to be fit into this model (e.g. TRP2 (slaty in mouse), MSH, silver locus protein, pink-eyed dilution locus, fibroblast growth factor, endothelin-B). Hopefully, in the not to distant future, we'll have the answers.
Oh, and by the way (if you've read this far), I still don't have the answer for apricot Mastiffs. Furthermore, Bull's stripes seem to be less noticeable in the winter time. Could they possibly be due to a temperature sensitive variant of tyrosinase ? Does brindling tend to be lighter in the cold winter months ? - more food for thought. If you think Bull's stripes are odd, you should visit a web page which shows some more unusual coloring in the mastiff here .
This is not the space to discuss the basics of genetics. There
are many excellent web pages which can educate you about genetics.
I suggest you visit a very good site,
Genetic Science Learning Center.
Moreover, let the reader beware that, even knowing the different
loci for the genes cannot fully explain how they may interact.
For example, there can be:
incomplete (intermediate) dominance - if allele A is 'completely'
dominant, AA and Aa individuals are alike phenotypically; but
if AA produced a more extreme or differential expression than
Aa, the inheritance is 'intermediate'
modifiers - genes which can alter or modify, by suppressing
or enhancing, the visible effects of other genes
epistasis - any gene or gene pair that masks the expression
of another, nonallelic gene, is epistatic to that gene; the gene
suppressed is said to be hypostatic
atavism - the chance combination of genes that allows a
long-suppressed or hidden characteristic to finally come to expression
