Hague (Hag): A New Mouse Hair Mutation With an Unstable Semidominant Allele

Corresponding author: Atsushi Yoshiki, BioResource Ctr., RIKEN Tsukuba Institute, 3-1-1 Koyadai, Tsukuba, Ibaraki 305-0074, Japan., yoshiki{at}rtc.riken.go.jp (E-mail)

Communicating editor: D. KINGSLEY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A spontaneous mouse hair mutation was identified in a C3H/HeN colony. The mode of inheritance of the mutation was semidominant, with incomplete penetrance when heterozygous. The trait is controlled by a single locus hague (Hag), which was mapped to the telomeric region of chromosome 15. This mutation was shown to be unstable, since its transmission could be switched from semidominant to recessive. To identify the causative gene and the nature of the mutation, hague was introduced into a high-resolution and high-density molecular genetic map. Over 2000 meioses were analyzed and the mutation was mapped to the keratin 2 complex genes. A YAC and BAC physical map of the critical region was then constructed and the gene involved was located in a 600- to 800-kb-long segment. Fourteen genes were mapped to this region; of these, 11 were expressed in the skin (5 epidermic cytokeratin and 6 hard keratin genes), but none were mutated in hague mice.

HAIR follicles are relatively simple structures, with their own pool of stem cells that, once differentiated during embryonic development, undergo repeated cycles of degeneration/regeneration throughout life (COSTARELIS et al. 1990 ; HARDY 1992 ). One possible approach to the understanding of the molecular events involved in the development and cyclic life of these tiny organs might be to generate knockout mice for the genes that are known to be expressed in the hair follicle (YAMANISHI 1998 ) and then to observe carefully the consequences, if any, of these engineered mutations. Another approach is to identify at the molecular level, by positional cloning, the genes that are affected in the many mouse hair mutations available. These two approaches are complementary and, at least in two cases, the engineered null alleles have been shown to mimic and failed to complement a spontaneous mutation [for example, fibroblast growth factor 5 (Fgf5) and angora (go) or transforming growth factor-(Tgfa) and waved 1 (wa1; LUETTEKE et al. 1993 ; MANN et al. 1993 ; HEBERT et al. 1994 )]. The transgenic approach leading to the production of knockout or null alleles is powerful but, by definition, it is limited to those genes whose DNA sequence is already known and often results in unexpected and sometimes extreme consequences. The positional cloning of mutant alleles, even if it requires the breeding of a large quantity of mice, is more general and results in many instances in the identification of previously unknown genes. Using such a strategy allowed, for example, the identification of Foxn1, a gene encoding a transcription factor, as being altered in the nude (nu) mouse mutation (NEHLS et al. 1994 ). Similarly, a mutation affecting the tyrosine kinase activity of the gene encoding for the receptor of the epidermal growth factor (Egfr) was found to be responsible for the waved 2 (wa2) mutation (LUETTEKE et al. 1994 ). Finally, a retroviral insertion helped in the molecular identification of the gene involved in the hairless (hr) mutation (CACHON-GONZALEZ et al. 1994 ). In this article we describe a new mutation affecting the fur that was found segregating in the C3H/HeN inbred strain of mouse. This new mutation, which exhibits unusual inheritance and produces curling of the hair or balding similar to caracul (Ca), is genetically controlled by a single locus on chromosome 15. We gave it the provisional name of hague (bald, in Japanese), with Hag as a symbol.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mouse strains:
C3H/HeN mice were purchased from Charles River Japan. BALB/c and C57BL/6J mice were purchased from CLEA Japan. The B6-Krt2^SEG strain (lab code IRCS 119), an interspecific recombinant congenic strain homozygous for the chromosomal segment spanning the interval D15Mit41 to D15Mit16 and containing the keratin-2 (Krt2) complex from strain SEG/Pas (Mus spretus) in a C57BL/6 genetic background was obtained from Xavier Montagutelli, Institut Pasteur (Paris). The wild-derived strains CAST/Ei (M. musculus castaneus) and PWK (M. musculus musculus) were gifts from the late V. M. Chapman, Roswell Park Memorial Institute (Buffalo). Strain MSM (Mus musculus molossinus) was imported from the National Institute of Genetics (Mishima, Shizuoka, Japan). C57BL/6By-Ca, Kitl^Sl mice were purchased from the Jackson Laboratory (Bar Harbor, ME). All the experiments reported in this article conformed to protocols approved by the Institutional Animal Care and Use Committee of RIKEN.

Preparation of hair samples:
Hair samples were plucked from the middle area of the dorsal skin and fixed in 70% ethanol, dehydrated with 90 and 100% ethanol. Then the samples were immersed three times in xylene, mounted with Marinol (Muto-Pure Chemicals, Ltd., Tokyo) and observed under a microscope (BH-2, Olympus, Tokyo) equipped with a digital camera (Sony digital photo camera model DKC-5000, Sony, Tokyo).

Genomic DNA:
Genomic DNA samples were prepared from tail clips using a classical procedure (MILLER et al. 1988 ).

Molecular markers:
All D15Mit microsatellite markers used in our experiments were from the MIT database (DIETRICH et al. 1996 ). To maximize polymorphisms, PCR primers for the analysis of most keratin genes, except Krt2-4 and Krt2-5, were chosen from the 3' untranslated region. For Krt2-4, the primers were chosen in the first exon and for Krt2-5 the primers were chosen in the fourth and fifth exons; all amplified products were specific of a keratin isoform. The sequences for all primers used in our experiments are listed in Table 1.

All PCR reactions were carried out in 25-µl volume under standard conditions. For microsatellites D15Mits, D15Kus2, D15Kus4, Krt2-8, and Krt1-18, the parental origin of the different alleles was estimated on the basis of simple sequence length polymorphism analysis performed in 4% agarose gel; for markers D15Kus1, Krt2-17, and Krt2-7 the allelic forms were analyzed by single-strand conformation polymorphism using the Clean DNA analysis kit (Pharmacia, Piscataway, NJ).

Yeast artificial and bacterial artificial chromosome clones:
Yeast artificial chromosome (YAC) 183h3 was retrieved from the MIT database (HALDI et al. 1996 ) and purchased from Research Genetics (Birmingham, AL). YACs 70m8a9, 53m3d3 (LARIN et al. 1991 ), and b22m5a1 (HALDI et al. 1996 ) were identified after screening the GENOSCOPE database (Evry, France). For each address, 10 independent clones were isolated and the sizes of the cloned DNA were determined using pulsed-field gel electrophoresis standard procedures (CARLE and OLSON 1987 ). YAC ends were isolated by three rounds of degenerated oligoprimer (DOP)-vector PCR (WU et al. 1996 ) using DOP-JUN1 and DOP-6MW (TELENIUS et al. 1992 ; WU et al. 1996 ) as degenerated primers for the insert and the primers specific to the YAC arms (HERRING et al. 1998 ). Reactions were performed using 100 ng of total DNA extracted from the yeast clones as a template under the conditions described by HERRING et al. 1998 and WU et al. 1996 . Bacterial artificial chromosome (BAC) ends were isolated by two rounds of DOP-vector PCR. Internal sequences from YACs were isolated by inter-B1 PCR (HUNTER et al. 1994 ). The PCR products were cloned in pCR2.1 (Invitrogen, San Diego) and sequenced using an ALF sequencer (Pharmacia).

Expression of keratin genes in the skin by reverse transcriptase-PCR:
Total RNA was extracted from the dorsal skin tissues of mice of each genotype at 10 days old by using the total RNA extraction kit (Amersham Pharmacia Biotech, Little Chalfont, Buckinghamshire, England) according to the manufacturer's instruction. Reverse transcription of messenger RNA into cDNA was performed by incubating the total RNA with SuperScript II RNase H-reverse transcriptase (GIBCO BRL, Rockville, MD) and oligo(dT) primers (GIBCO BRL). The cDNA was purified using the PCR purification kit (QIAGEN, Hilden, Germany). PCR reaction was carried out by using Advantage cDNA polymerase M (CLONTECH, Palo Alto, CA) with the oligonucleotide primers for each keratin gene. The PCR products were subjected to agarose gel electrophoresis and stained by ethidium bromide to determine the presence and size of the products. The PCR products were also sequenced to detect the mutation.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Description of phenotypes and inheritance of the traits:
The original mutant phenotype was identified, because of extremely curly fur, in the C3H/HeN inbred strain at the RIKEN Institute and the mutation was kept in this background. When affected mice of this kind were crossed to wild-type partners of the same strain, the mutation appeared to be transmitted as a semidominant or dominant allele with both curly and normal-furred offspring. When mutant mice were intercrossed, a more extreme phenotype was observed in their progeny, with some mice being extensively depilated from 10 days of age. We refer to this severe phenotype as phenotype2 hereafter, phenotype1 being the first observed phenotype (Fig 1). As is the case for several other fur mutations (rex, caracul, etc.), curling of the hair in mice exhibiting phenotype1 was obvious in young mice and became less obvious (at least more difficult to recognize) after 2 months.

View larger version (153K): In this window In a new window Download PPT slide   Figure 1. Phenotypes segregating in the C3H/HeN colony. Four-week-old C3H/HeN mice from left to right: phenotype1, phenotype2, and wild type.

Mice with either phenotype1 or phenotype2 have the classical four types of hair: monotrich, awl, auchene, and zigzag, although each hair shaft is severely malformed and curled. In mice with phenotype2, zigzag hairs developed poorly and broken hair shafts are commonly observed under the microscope, probably indicating an increased fragility.

To study the transmission of the mutant phenotypes, different types of crosses between phenotype1, phenotype2, and wild-type (wt) mice were set as shown in Table 2. All these crosses were between mice of the same coisogenic strain, C3H/HeN.

When intercrossed, mice with phenotype2 appeared to breed true since only phenotype2 offspring were observed. Conversely, no phenotype2 was ever observed in the progeny of phenotype2 mice mated to wild type (+/+). In these crosses, no significant bias related to the sex of the partner contributing the mutant allele was observed. Surprisingly, however, in these progeny not all the offspring exhibited the same phenotype, indicating that the hague (Hag) mutation is not a classical bona fide semidominant mutation with all Hag/+ mice exhibiting phenotype1. In fact, in the above-mentioned cross (phenotype2 x wt), 28% of the offspring exhibited phenotype1 while the other 72% exhibited a wild-type phenotype. Similarly, when phenotype1 mice (supposed to be Hag/+) were mated to wild-type partners (+/+), 42% of the offspring exhibited phenotype1 while the remaining 58% had a wild phenotype. From the analysis of the different crosses, we postulated that the phenotypic heterogeneity among Hag/+ mice was the result of either incomplete penetrance or a polygenic control of the phenotype involving modifier genes.

Genetic localization of the trait:
One way to clarify the situation reported above for the inheritance of the phenotypes was to localize precisely the locus (Hag) determining the hague mutation and then to perform a complete genetic analysis of the progeny born from different crosses involving the same Hag mutation. To achieve this, intersubspecific matings between females with phenotype2 (Hag/Hag) and males of the CAST/Ei strain (+/+) were set. The F₁ born from these crosses again appeared heterogeneous with curly haired (phenotype1) and wild phenotypes. To recover all three kinds of genotypes, we intercrossed F₁ animals of the two kinds (i.e., wild type x wild type and curly haired x curly haired) and produced 413 F₂ mice. Among these offspring, 120 exhibited phenotype2, 73 exhibited phenotype1, and 220 were wild type.

A complete genome scanning was then performed on a sample of 15 phenotype1 and a sample of 15 phenotype2 mice and evidence for linkage was found only with the telomeric region of mouse chromosome 15. This preliminary data being taken into account, we then typed all the 193 affected mice (120 phenotype2 and 73 phenotype1 mentioned above) for all polymorphic markers in this region and analyzed the data with MAP MANAGER (MANLY 1993 ). The localization of the Hag locus on chromosome 15 was confirmed, where it appeared flanked by microsatellite markers D15Mit173 and D15Kus2 at the centromeric edge and by microsatellite markers D15Mit79 and D15Mit35 at the telomeric edge (Fig 2A).

View larger version (101K): In this window In a new window Download PPT slide   Figure 2. Genotype of the F2 progeny. (A) The 193 affected F2 mice (phenotype1 and phenotype2). (B) The wild-type F2 mice. Black, white, and gray boxes indicate homozygosity for the CAST allele, homozygosity for the C3H allele, and heterozygosity, respectively. The distance is indicated as centimorgans (cM) with the 95% confidence interval in brackets.

Analysis of penetrance:
Genotyping the 220 F₂ offspring with a wild phenotype (i.e., noncurly hairs), we found 125 of them to be heterozygous for both D15Kus2 and D15Mit35, as well as for all markers in between (Fig 2B). These mice obviously had the Hag/+ genotype, indicating that heterozygosity at this locus was not sufficient to generate phenotype1. In fact, in the F₂, there were two kinds of Hag/+ mice: some exhibiting the expected curly haired phenotype and others being wild type.

Considering this observation, we had an opportunity to test for a parental effect in the occurence of phenotype1 in genetically Hag/+ heterozygous mice. We compared the proportion of phenotype1 vs. wild type in heterozygous mice at Krt2-17, which is tightly linked to hague, between the F₂ populations originating from either F₁ mice with phenotype1 (cross 1) or F₁ mice with a wild-type phenotype (cross 2; Table 3). Surprisingly, we found a strong bias in the phenotype distribution, with only 3 phenotype1 mice out of 74 (4.1%) heterozygotes in cross 1, compared to 52 out of 84 (61.8%) in cross 2 .

To search for modifier genes, a second genome scan was performed on a selected sample of 63 F₂ mice born from cross 1 (phenotype1 x phenotype1) and all heterozygous for the markers between D15Mit97 and D15Mit16. From these 63 mice, 36 were affected (phenotype1) while 27 had a wild phenotype. However, we did not find any evidence for a statistically significant linkage disequilibrium and accordingly concluded that the trait is controlled by a single locus only and that the phenotype of the heterozygous Hag/+ parents has a strong influence on the phenotype of Hag/+ offspring.

Loss of semidominance by inbreeding:
From its origin on an inbred mouse strain and the mapping experiments above, we concluded that the Hague phenotype is controlled by a single major locus. However, in the progeny of heterozygous (Hag/+) wild-type mice, most of the genetically heterozygous (Hag/+) mice appeared with a wild-type phenotype, which suggested that the mutant allele might now behave as a recessive mutation. To test this hypothesis we intercrossed for five generations Hag/Hag mice exhibiting phenotype2, in their original inbred genetic background (C3H/HeN), the one where the mutation first occurred, and then we crossed some of these Hag/Hag coisogenic mice with wild-type (+/+) C3H/HeN mice. Surprisingly, we found no phenotype1 among 80 progeny of such a cross! Since no modifier genes were identified in the previous crosses, this suggested that the original semidominant mutation was unstable and could turn into a recessive allele. From this point on, we use the symbols Hag for the semidominant allele and hag for the recessive one.

To check this hypothesis, we decided to generate mice carrying only one mutated allele, either the semidominant original Hag allele or its new recessive form hag. For this experiment, some B6-Krt2^SEG mice were crossed with affected homozygous mice (phenotype2) from the mutant colony. At the same time, some mice homozygous for the "new recessive" allele were crossed with C57BL/6 mice. The affected F₁ (phenotype1) from the first cross were mated with the wild-type F₁ from the second cross; 181 affected F₂ mice were born from this cross, among which 88 were phenotype1 and 93 phenotype2. This allele segregation of the two markers around the Hag locus (D15Mit246 and D15Kus2) was analyzed in the affected progeny (Fig 3). At both markers, all the phenotype2 mice were homozygous for the C3H allele and all the phenotype1 mice carried a C3H allele and a C57BL/6 allele. As expected, all phenotype1 mice carried a C3H haplotype inherited from the affected F₁ parent, and no SEG alleles were detected. This experiment demonstrated that the original semidominant allele was indeed unstable and could occasionally be turned into a recessive one, which suggested that the incomplete penetrance was caused by a change in the mode of transmission from semidominant to recessive. Using data from Table 3, we could obtain a raw estimation of the switch from semidominant to recessive, 32/84 = 38.1%, and its reversion from recessive to semidominant, 3/74 = 4.1%, after one generation on a mixed genetic background (C3H x CAST). A second mutant colony was then established with phenotype2 mice carrying the recessive allele.

View larger version (23K): In this window In a new window Download PPT slide   Figure 3. Loss of semidominance by inbreeding. S, SEG alleles; H, C3H alleles from the mutant colony; B, C57BL/6 alleles; h, C3H alleles carried by the haplotype originating from mice that have lost the ability to produce affected heterozygous mice.

We observed that the expression of the phenotype in affected heterozygous (Hag/+) mice varied from one animal to the next (Fig 4A), suggesting a variegated expression of the semidominant allele, Hag. To test this variegation hypothesis, both alleles were transferred into a genetic background increasing hair fragility. To achieve this, phenotype2 mice homozygous for either the semidominant allele (Hag) or the recessive allele (hag) were mated with mice homozygous for the caracul mutation (Ca) and the fur of the offspring was observed. Mice carrying the recessive allele in association with the caracul mutation Ca +/+ hag displayed a homogeneous caracul phenotype. The fur of Hag +/+ Ca mice appeared heterogeneous with hairy and bald patches, suggesting that in the bald regions the Hag allele was expressed and interaction between Hag and Ca products was increasing the fragility of the hair shafts, while in hairy patches similar to the fur of hag +/+ Ca mice Hag expression was similar to hag expression (Fig 4B).

View larger version (231K): In this window In a new window Download PPT slide   Figure 4. The semidominant allele is variegated. (A) Ten-day-old mice of Hag/+ genotype (a, b, and c) and wild type (d). Curly whiskers and dorsal hairs are more evident in the following order: a > b > c. (B) Eight-week-old mice. From left to right, three mice heterozygous for the Caracul mutation and the semidominant allele (Hag +/+ Ca), one mouse heterozygous for both the Caracul mutation and the recessive hague (hag +/+ Ca), and one C3H wild-type mouse.

Physical mapping of the Hag locus:
To refine the mapping of the mutation, a series of backcrosses and intercrosses were set between C3H/HeN mice heterozygous for the mutant allele Hag and wild-derived mice. Up to 2091 meioses were analyzed and three recombinant mice were identified between the Hag locus and the cluster D15Mit246-Krt2-8-Krt1-18, placing this locus centromeric to these three markers. Two YAC libraries (LARIN et al. 1991 ; HALDI et al. 1996 ) were screened by PCR for markers D15Mit246, Krt2-1, Krt2-17, Krt2-16, and Krt2-7. The positive clones (183h3, 70m8a9, b22m5a1, and 53m3d3) were tested by PCR for the rest of the markers mapped to the critical region (Fig 5). Two internal deletions were detected in b22m5a1 and 53m3d3. The ends of the above-mentioned YACs were cloned and four of them were reverse mapped to the contig (Fig 5): 183h3-T3 end (D15Kus1), 70m8a9-L end (D15Kus2), 70m8a9-R end (D15Kus5), and b22m 5a1-T7 end (D15Kus3). Some polymorphisms were detected for D15Kus1 and D15Kus2 and these two additional markers were then introduced into the genetic map, where D15Kus2 appeared the closest centromeric landmark to Hag and no recombination event was found between Hag and D15Kus1 (Fig 5). Among the inter-B1 sequences isolated from the YACs, only one, D15Kus4, was mapped back to the region. To complete a BAC contig, the MIT 129/Sv strain library and a C57BL/6 library (OSOEGAWA et al. 1998 ) were screened by PCR for Krt2-1, Krt2-4, Krt2-6a, Krt2-8, Krt2-11, and Krt2-16 markers. From these screenings 13 clones were retrieved (Fig 5). The 51l7Sp6 end contained a part of the first exon and the first intron of Krt2-6a. Since this BAC clone was negative for the 3' ends of both Krt2-6 genes, Krt2-6b was located centromeric to Krt2-6a. The 254k21Sp6 end (AQ929432) contained two internal exons and the flanking intronic sequences of Krt2-1 but did not contain the 3' end of Krt2-1 and Krt2-17; Krt2-1 was then located telomeric to Krt2-17 (Fig 5).

View larger version (27K): In this window In a new window Download PPT slide   Figure 5. Physical map around Hague. Clone address, insert size, and strain of origin (B, C57BL/6; 1, 129; and H, C3H) are indicated. Black denotes recombinant markers; gray, nonrecombinant markers; and white, nonpolymorphic markers that could not be typed.

Molecular cloning of new intermediate filament genes:
To find new type II hair keratin genes, we downloaded from the GenBank database all mouse expressed sequence tag (EST) sequences containing "Mus musculus type II hair keratin mRNA" as a criterion in the definition field. From this search, 209 EST clones were retrieved; 182 of these clones were arranged into seven contigs using the GCG package (http://www.gcg.com/). Early high-quality sequencing stop occurred for the 27 remaining EST clones, leading to poor-quality sequences and the inability to arrange these clones in contig. Three out of these seven contigs identified previously known hair keratin genes (Krt2-10, Krt2-11, and Krt2-18). Further analysis of the remaining four contigs was carried out by resequencing some of the EST clones and two new type II hair keratin sequences were identified (Krt2-19 and Krt2-20). The first one, Krt2-19 (AF312018), was colocalized with the previously known hard keratin genes, between Krt2-7 and D15Kus1. The other one, Krt2-20 (AY028606), was mapped between Krt2-16 and the Sp6 end of BAC 298p5.

Candidate genes:
Several genes were identified in the critical region: two nonepidermal basic cytokeratin genes, Krt2-4 and Krt2-7; six epidermal cytokeratin genes, Krt2-1, Krt2-5, Krt2-6a, Krt2-6b, Krt2-6h, and Krt2-17; and six basic hard keratin genes, Krt2-10, Krt2-11, Krt2-16, Krt2-18, Krt2-19, and Krt2-20.

Because these 14 genes were potential candidates by their position, a survey of their functions was undertaken to identify the best ones. Three criteria were used: expression in the skin, phenotype of the available knockout mice, and disease associated with the human ortholog (Table 4). In a particular cell type, a type I keratin protein is always found in association with a type II keratin protein and, in human mutations, the coexpressed type I and type II keratin genes were reported to lead to the same disease (FUCHS 1995 ). By reverse transcriptase-PCR, no expression of Krt2-6b and Krt2-7 was detected in the skin of either wild-type or hague mice. Human mutations of two hard keratin genes have been associated with the hair loss disease monilethrix (WINTER et al. 1997A , WINTER et al. 1997B ), but no mutations were found in the hague alleles for the six hard keratin genes. Alopecia and hair anomalies are found in pachyonychia congenita patients, but no fur anomalies were found in mice homozygous for null alleles of Krt2-6a and Krt2-6b (WOJCIK et al. 2000 ; WONG et al. 2000 ). Moreover, no mutation was found in the coding region of Krt2-6a in Hague mice. Krt2-6h, like its human ortholog K6HF (WINTER et al. 1998 ), is expressed only in the companion layer of the hair follicle but no mutation was found in the genome of Hag/Hag mice. Conversely, no hair anomalies have been observed in mice homozygous for a null allele of Krt2-4 (NESS et al. 1998 ). Human mutations of KRT1 (ortholog of Krt2-1), KRT2 (ortholog of Krt2-17), and KRT5 (ortholog of Krt2-5) lead to blistering skin diseases (reviewed by FUCHS 1995 ). No such skin blistering was observed in the hague mice, suggesting that these genes are not candidates.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The transmission of the trait is unusual:
We have identified a semidominant mutation, Hague (Hag), producing hair curling. At first, the mutant allele appeared to be semidominant with incomplete penetrance (Hag/+), and we have shown that the semidominant allele is unstable, that it can be converted from dominant to recessive at an appreciable rate and from recessive to dominant at a lower rate, and that this instability could be due to either epigenetic modification or structural rearrangements occurring in the region. A similar situation has already been reported for several yellow mutant alleles (A^hvy, A^iapy, A^iy, and A^vy) at the A locus, chromosome 2, with an expressivity depending on the parental origin of the mutant allele in the zygote (DICKIE 1962 , DICKIE 1966 ; MICHAUD et al. 1994B ; SIRACUSA et al. 1995 ). All these mutations occurred in a C3H background and were caused by integration of intracisternal A particle sequences in a different region of the gene: in exon 1C for A^hvy (ARGESON et al. 1996 ), between exon 1D and exon 2 for A^iapy and A^iy (DUHL et al. 1994 ; MICHAUD et al. 1994A ), and in exon 1A for A^vy (DUHL et al. 1994 ). Transcription from the intracisternal A particle led to a deregulation of the agouti gene associated with the yellow phenotype.

Methylation of these insertions has been associated with their silencing and, in this case, despite carrying a yellow allele, the mice appear pseudo-agouti. The efficiency of this silencing process has been studied through parental germline transmission. The rate of silencing varied from one allele to another: 15.6% for A^vy (WOLFF 1978 ), 34.8% for A^hvy (ARGESON et al. 1996 ), and 40.6% for A^iapy (MICHAUD et al. 1994A ) upon passage through the paternal germline; these rates decreased to 1.2, 1, and 2.5%, respectively, upon passage through the maternal germline. A survey of 6748 mice has shown that, at least for A^vy, the rate of silencing was dependent on the genetic background (WOLFF 1978 ). Our data showed a high switching rate of 38.2% from semidominant to recessive and a lower reverse switching rate of 4.1% on the mixed (C3H x CAST) genetic background. This data could be explained by methylation of a transposon. In this case, the integration of a mobile element containing promoter-like sequences in a keratin gene could lead to an overexpression of the gene or to the transcription of a chimeric or truncated messenger. Such a mutation would be expected to lead to a disruption of the keratin network with a semidominant effect, and the subsequent silencing of the mobile element, over several generations, would then lead to a stable recessive allele as observed in the case of our hague mutation.

Fine mapping of Hag:
By linkage analysis we have mapped the locus to the subtelomeric region of mouse chromosome 15. In this region, at least four loci associated with wavy hairs (caracul, Ca; crimpy, cpy) or hair loss (naked, N; shaven, Sha) have already been mapped (HUPPI et al. 1998 ). A BAC and YAC physical map of the critical region (600–800 kb), including the mouse keratin complex genes, was constructed. We have mapped four basic hair keratin genes between Krt2-7 and Krt2-6b; in humans the basic keratin genes have been recently located in a 200-kb fragment between KRT7 and the keratin 6 gene (ROGERS et al. 2000 ). In humans, the physical map of the ortholog region at 12q13 has been resolved and the cytokeratin gene order has been found to be conserved during evolution (YOON et al. 1994 ). In both human and mouse, the basic hair keratin gene has been located between the keratin 7 gene and the keratin 6 gene. The structure of several type II keratin genes has been elucidated in different species (mouse, human, and sheep). In all cases but one, the genomic structure was conserved (<10 kb with nine exons); therefore any single type II keratin gene should be entirely within one, single BAC clone. The hague critical region being covered with only five overlapping BAC clones, genetic complementation by BAC transgenesis is now amenable to experimentation that should help in further narrowing the critical region.

Candidates genes:
In humans, the cytokeratin genes have been demonstrated to be involved in several skin diseases (MCLEAN and LANE 1995 ). In mice, to date, no such cytokeratin gene mutations have been found to be associated with any skin disease, but mice expressing a deleterious allele of the human keratin 14 gene displayed nearly all the symptoms of the corresponding human disease (VASSAR et al. 1991 ). In addition, mice carrying mutated Krt2-6a transgenes exhibited severe blistering and neonatal lethality (WOJCIK et al. 1999 ). These data suggested that epidermal cytokeratin gene mutations should lead to skin blistering diseases in mice, which is not observed in hague mice. It has also been reported that mutations in at least two basic hair keratin genes lead to the inherited hair disorder monilethrix (WINTER et al. 1997A , WINTER et al. 1997B ). In mice, overexpression of a sheep hard keratin gene was found to lead to hair loss similar to the phenotype of homozygous Hag/Hag mice (POWELL and ROGERS 1990 ). Since the Hag/Hag or Hag/+ mice display a hair phenotype without any blistering of the skin, the basic hair keratin gene family may be considered top-ranked candidates in the positional cloning of Hag. In humans, this gene family, located in a 200-kb fragment between KRT7 and the keratin 6 genes, contained at least six basic hair keratin genes and four expressed pseudogenes (ROGERS et al. 2000 ).

	FOOTNOTES

¹ Present address: Department of Obstetrics and Gynecology, Baylor College of Medicine, Houston, TX 77030.
³ Present address: Biochemistry Division, National Cancer Ctr. Research Institute, Chuo-ku, Tokyo 104-0045, Japan.
⁴ Present address: ALOKA Co., Ltd., Mitaka-shi, Tokyo 181-8622, Japan.

	ACKNOWLEDGMENTS

The authors are grateful to Patricia Baldacci and Andrew S. McCallion for critical reading of this manuscript. We thank Seiichi Otake and Jun Inoue for taking care of the hague mice colonies. This work was supported by a grant from the Science and Technology Agency of the Japanese government.

Manuscript received July 9, 2002; Accepted for publication July 23, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ARGESON, A. C., K. K. NELSON, and L. D. SIRACUSA, 1996  Molecular basis of the pleiotropic phenotype of mice carrying the hypervariable yellow (A^hvy) mutation at the agouti locus. Genetics 142:557-567.[Abstract]

CACHON-GONZALEZ, M. B., S. FENNER, J. M. COFFIN, C. MORAN, and S. BEST et al., 1994  Structure and expression of the hairless gene of mice. Proc. Natl. Acad. Sci. USA 91:7717-7721.[Abstract/Free Full Text]

CARLE, G. F. and M. V. OLSON, 1987  Orthogonal field alternation gel electrophoresis. Methods Enzymol. 155:468-482.[Medline]

COSTARELIS, G., T.-T. SUN, and R. M. LAVKER, 1990  Label-retaining cells reside in the bulge area of pilosebaceous unit: implications for follicular stem cells, hair cycle, and skin carcinogenesis. Cell 61:1329-1337.[Medline]

DICKIE, M. M., 1962  A new viable yellow mutation in the house mouse. J. Hered. 53:84-86.[Free Full Text]

DICKIE, M. M., 1966  Intermediate yellow. Mouse News Lett. 34:30.

DIETRICH, W. F., J. MILLER, R. STEEN, M. A. MERCHANT, and D. DAMRON-BOLES et al., 1996  A comprehensive genetic map of the mouse genome. Nature 380:149-152.[Medline]

DUHL, D. M., H. VRIELING, K. A. MILLER, G. L. WOLFF, and G. S. BARSCH, 1994  Neomorphic agouti mutations in obese yellow mice. Nat. Genet. 8:59-65.[Medline]

FUCHS, E., 1995  Keratins and the skin. Annu. Rev. Cell Dev. Biol. 11:123-153.[Medline]

HALDI, M. L., C. STRICKLAND, P. LIM, V. VANBERKEL, and X. CHEN et al., 1996  A comprehensive large-insert yeast artificial chromosome library for physical mapping of the mouse genome. Mamm. Genome 7:767-769.[Medline]

HARDY, M. H., 1992  The secret life of the hair follicle. Trends Genet. 8:55-61.[Medline]

HEBERT, J. M., T. ROSENQUIST, J. GOTZ, and G. R. MARTIN, 1994  FGF5 as a regulator of the hair growth cycle: evidence from targeted and spontaneous mutations. Cell 78:1017-1025.[Medline]

HERRING, C. D., C. CHEVILLARD, S. L. JOHNSTON, P. J. WETTSTEIN, and R. RIBLET, 1998  Vector-hexamer PCR isolation of all insert ends from a YAC contig of the mouse Igh locus. Genome Res. 8:673-681.[Abstract/Free Full Text]

HUNTER, K. W., S. D. ONTIVEROS, M. L. WATSON, V. P. STANTON, JR., and P. GUTIERREZ et al., 1994  Rapid and efficient construction of yeast artificial chromosome contigs in the mouse genome with interspersed repetitive sequence PCR (IRS-PCR): generation of a 5cM, >5 megabase contig on mouse chromosome 1. Mamm. Genome 5:597-607.[Medline]

HUPPI, K., D. SIWARSKI, and V. LETTS, 1998  Encyclopedia of the mouse.genome VII. Mouse chromosome 15. Mamm. Genome 8:S292-S306.

LARIN, Z., A. P. MONACO, and H. LEHRACH, 1991  Yeast artificial chromosomes libraries containing large inserts from mouse and human DNA. Proc. Natl. Acad. Sci. USA 88:4123-4127.[Abstract/Free Full Text]

LUETTEKE, N. C., T. H. QIU, R. L. PEIFFER, P. OLIVER, and O. SMITHIES et al., 1993  TGFa deficiency results in hair follicle and eye abnormalities in targeted and waved-1 mice. Cell 73:263-278.[Medline]

LUETTEKE, N. C., H. K. PHILLIPS, T. H. QIU, N. G. COPELAND, and H. S. EARP et al., 1994  The mouse waved-2 phenotype results from a point mutation in the EGF receptor tyrosine kinase. Genes Dev. 8:399-413.[Abstract/Free Full Text]

MANLY, K. F., 1993  A Macintosh program for storage and analysis of experimental genetic mapping data. Mamm. Genome 4:303-313.[Medline]

MANN, G. B., K. J. FOWLER, A. GABRIEL, E. C. NICE, and R. L. WILLIAMS et al., 1993  Mice with a null mutation of the TGF- gene have abnormal skin architecture, wavy hair, and curly whiskers and often develop corneal inflammation. Cell 73:249-261.[Medline]

MCLEAN, W. H. I. and E. B. LANE, 1995  Intermediate filaments in disease. Curr. Opin. Cell Biol. 7:118-125.[Medline]

MICHAUD, E. J., S. J. BULTMAN, M. L. KLEBIG, M. J. VAN VUGT, and L. J. STUBBS et al., 1994a  A molecular model for the genetic and phenotypic characteristics of the mouse lethal yellow (A^y) mutation. Proc. Natl. Acad. Sci. USA 91:2562-2566.[Abstract/Free Full Text]

MICHAUD, E. J., M. J. VAN VUGT, S. J. BULTMAN, H. O. SWEET, and M. T. DAVISSON et al., 1994b  Differential expression of a new dominant agouti allele (A^iapy) is correlated with methylation state and is influenced by parental lineage. Genes Dev. 8:1463-1472.[Abstract/Free Full Text]

MILLER, S. A., D. D. DYKES, and H. F. POLESKY, 1988  A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res. 16:1215.[Free Full Text]

NEHLS, M., D. PFEIFER, M. SCHORPP, H. HEDRICH, and T. BOEHM, 1994  New member of the winged-helix protein family disrupted in mouse and rat nude mutations. Nature 372:103-107.[Medline]

NESS, S. L., W. EDELMANN, T. D. JENKINS, W. LIEDTKE, and A. K. RUSTGI et al., 1998  Mouse keratin 4 is necessary for internal epithelial integrity. J. Biol. Chem. 273:23904-23911.[Abstract/Free Full Text]

OSOEGAWA, K., P. Y. WOON, B. ZHAO, E. FRENGEN, and M. TATENO et al., 1998  An improved approach for construction of bacterial artificial chromosome libraries. Genomics 52:1-8.[Medline]

POWELL, B. C. and G. E. ROGERS, 1990  Cyclic hair-loss and regrowth in transgenic mice overexpressing an intermediate filament gene. EMBO J. 9:1485-1493.[Medline]

ROGERS, M. A., H. WINTER, L. LANGBEIN, C. WOLF, and J. SCHWEIZER, 2000  Characterization of a 300kbp region of human DNA containing the type II hair keratin. J. Invest. Dermatol. 114:464-472.[Medline]

SIRACUSA, L. D., L. L. WASHBURN, D. A. SWING, A. C. ARGESON, and N. A. JENKINS et al., 1995  Hypervariable yellow (A^hvy), a new murine agouti mutation: A^hvy displays the largest variation in coat color phenotypes of all known agouti alleles. J. Hered. 86:121-128.[Abstract/Free Full Text]

TELENIUS, H., N. P. CARTER, C. E. BEBB, M. NORDENSKJOLD, and B. A. PONDER et al., 1992  Degenerate oligonucleotide-primed PCR: general amplification of target DNA by a single degenerate primer. Genomics 13:718-725.[Medline]

VASSAR, R., P. A. COULOMBE, L. DEGENSTEIN, K. ALBERS, and E. FUCHS, 1991  Mutant keratin expression in transgenic mice causes marked abnormalities resembling a human genetic skin disease. Cell 64:365-380.[Medline]

WINTER, H., M. A. ROGERS, M. GEBHARDT, U. WOLLINA, and L. BOXALL et al., 1997a  A new mutation in the type II hair cortex keratin hHb1 involved in the inherited hair disorder monilethrix. Hum. Genet. 101:165-169.[Medline]

WINTER, H., M. A. ROGERS, L. LANGBEIN, H. P. STEVENS, and I. M. LEIGH et al., 1997b  Mutations in the hair cortex keratin hHb6 cause the inherited hair disease monilethrix. Nat. Genet. 16:372-374.[Medline]

WINTER, H., L. LANGBEIN, S. PRAETZEL, M. JACOBS, and M. A. ROGERS et al., 1998  A novel human type II cytokeratin, K6hf, specifically expressed in the companion layer of the hair follicle. J. Invest. Dermatol. 111:955-962.[Medline]

WOJCIK, S. M., S. IMAKADO, T. SEKI, M. A. LONGLEY, and L. PETHERBRIDGE et al., 1999  Expression of MK6a dominant-negative and C-terminal mutant transgenes in mice has distinct phenotypic consequences in the epidermis and hair follicle. Differentiation 65:97-112.[Medline]

WOJCIK, S. M., D. S. BUNDMAN, and D. R. ROOP, 2000  Delayed wound healing in keratin 6a knockout mice. Mol. Cell. Biol. 20:5248-5255.[Abstract/Free Full Text]

WOLFF, G. L., 1978  Influence of maternal phenotype on metabolic differentiation of agouti locus mutants in the mouse. Genetics 88:529-539.[Abstract/Free Full Text]

WONG, P., E. COLUCCI-GUYON, K. TAKAHASHI, C. GU, and C. BABINET et al., 2000  Introducing a null mutation in the mouse k6alpha and K6beta genes reveals their essential structural role in the oral mucosa. J. Cell Biol. 150:921-928.[Abstract/Free Full Text]

WU, C., S. ZHU, S. SIMPSON, and P. J. DE JONG, 1996  DOP-vector PCR: a method for rapid isolation and sequencing of insert termini from PAC clones. Nucleic Acids Res. 24:2614-2615.[Free Full Text]

YAMANISHI, K., 1998  Gene-knockout mice with abnormal epidermal and hair follicular development. J. Dermatol. Sci. 18:75-89.[Medline]

YOON, S.-J., J. LEBLANC-STRACESKI, D. WARD, K. KRAUTER, and R. KUCHERLAPATI, 1994  Organization of the human keratin type II gene cluster at 12q13. Genomics 24:502-508.[Medline]