The juvenile alopecia mutation (jal) maps to mouse Chromosome 2, and is an allele of GATA binding protein 3 (Gata3).

BACKGROUND: Mice homozygous for the juvenile alopecia mutation (jal) display patches of hair loss that appear as soon as hair develops in the neonatal period and persist throughout life. Although a report initially describing this mouse variant suggested that jal maps to mouse Chromosome 13, our preliminary mapping analysis did not support that claim.
RESULTS: To map jal to a particular mouse chromosome, we produced a 103-member intraspecific backcross panel that segregated for jal, and typed it for 93 PCR-scorable, microsatellite markers that are located throughout the mouse genome. Only markers from the centromeric tip of Chromosome 2 failed to segregate independently from jal, suggesting that jal resides in that region. To more precisely define jal's location, we characterized a second, 374-member backcross panel for the inheritance of five microsatellite markers from proximal Chromosome 2. This analysis restricted jal's position between D2Mit359 and D2Mit80, an interval that includes Il2ra (for interleukin 2 receptor, alpha chain), a gene that is known to be associated with alopecia areata in humans. Complementation testing with an engineered null allele of Il2ra, however, showed that jal is a mutation in a distinct gene. To further refine the location of jal, the 374-member panel was typed for a set of four single-nucleotide markers located between D2Mit359 and D2Mit80, identifying a 0.55 Mb interval where jal must lie. This span includes ten genes-only one of which, Gata3 (for GATA binding protein 3)-is known to be expressed in skin. Complementation testing between jal and a Gata3 null allele produced doubly heterozygous, phenotypically mutant offspring.
CONCLUSIONS: The results presented indicate that the jal mutation is a mutant allele of the Gata3 gene on mouse Chromosome 2. We therefore recommend that the jal designation be changed to Gata3jal, and suggest that this mouse variant may provide an animal model for at least some forms of focal alopecia that have their primary defect in the hair follicle and lack an inflammatory component.

Background

The initial assignment of spontaneous hair variants to particular genes can be a crucial first step in the long-term investigation into the role these genes play in the normal (and disrupted) development of the mammalian integument (for example, see refs. [1-9]). Unfortunately, several naturally-occurring hair and skin variants in mice remain out-of-the-mainstream of modern biological investigation, simply because they have not yet been assigned to a causative gene or even, in some cases, to a particular chromosome. One such variant is generated by the recessive juvenile alopecia mutation, abbreviated jal. This variant arose on the standard C3H/HeJ genetic background, and its origin and novel phenotype were described in a single brief paper published by McElwee et al. in 1999 [10]. Homozygous mice exhibit patchy hair loss (see Figure 1), wavy truncal hair, defects in hair follicles, and abnormalities in hair growth cycle regulation. Vibrissae defects are apparent at birth, and focal alopecia is evident as soon as hair develops in the neonatal period. Although McElwee and coworkers suggested that jal is located on mouse Chromosome (Chr) 13 [10], our preliminary backcross analysis [11] clearly showed that jal does not map anywhere on that chromosome.

Figure 1

A three-month-old C3H/HeJ-/J mouse, homozygous for .

Here, we describe the completed molecular-genetic analysis of a pair of large backcross families that allowed us to locate jal on mouse Chr 2, and then restrict its location to a small, defined interval at the centromeric tip. In addition, we describe complementation testing between jal and engineered null alleles of two co-localizing candidate genes, one of which (Gata3, for GATA binding protein 3) we identify as the likely basis of the juvenile alopecia phenotype in mice.

Methods

Mice

Mice from the standard inbred strains C57BL/6 J, C3H/HeJ, A/J, as well as inbred C3H/HeJ-jal/J mice were obtained from The Jackson Laboratory (Bar Harbor, ME, USA). Mice homozygous for the mutant jal allele were most reliably identified by vibrissae defects that are first evident shortly after birth. By two weeks of age, homozygotes present with distinct patches of hair loss (most often on the dorsal surface) that persist throughout life (see Figure 1). The amount of body surface affected varies widely among homozygous individuals (from less than 5% to greater than 95% [see Additional file 1]), even within the inbred C3H/HeJ-jal strain. While both male and female jal/jal homozygotes are fertile, we have maintained the C3H/HeJ-jal line since 2009 by crossing heterozygous females with homozygous males, to produce segregating litters.

Mice carrying a targeted mutation in the interleukin 2 receptor, alpha chain gene (Il2ra) were also obtained from The Jackson Laboratory. The creation of the Il2ra loss-of-function allele is described by Willerford et al.[12]. In brief, these investigators used homologous recombination to replace a 5.5 kb segment of the Il2ra gene which contains Exons 2 and 3 and encodes the interleukin 2 binding site [13] with a phosphoglycerate kinase (PGK)-neomycin resistance (neo) cassette. Mice carrying a targeted mutation in the GATA binding protein 3 gene (Gata3) were kindly donated by Dr. James Douglas Engel (University of Michigan, Ann Arbor, MI, USA). The creation of the Gata3 loss-of-function/reporter allele is described by van Doorninck et al.[14]. In brief, these investigators replaced 157 bp in Exon 2, including the start codon, with a nuclear localization signal (nls)-lacZ fusion cassette, followed by a PGK-hygromycin resistance (hyg) cassette.

All studies were in compliance with protocols approved by the Institutional Care and Use Committee (IACUC) at Central Connecticut State University (New Britain, CT, USA).

DNA isolation and analysis

Genomic DNA was isolated from 3 mm tail tip biopsies taken from two-week-old mice, using Nucleospin kits from BD Biosciences (Palo Alto, CA, USA). The polymerase chain reaction (PCR) was performed using the Titanium PCR kit from Clontech (Palo Alto, CA, USA). Oligonucleotide primers for PCR were synthesized by Invitrogen (Carlsbad, CA, USA), based on sequence information from online sources [15,16]. In addition to standard, PCR-scorable, microsatellite markers [17], we also assayed 4 markers based on single-nucleotide polymorphisms that have been reported to differ between the A/J and C3H/HeJ strains [15,16]. These markers, designated herein as SNP1-4, are described in detail in Additional file 2 and Additional file 3. To distinguish between Il2ra carriers and wild type mice, we used the 4-primer PCR assay recommended by the mouse supplier (The Jackson Laboratory). Two of these primers (5′CTGTGTCTGTATGACCCACC 3′, and 5′ CAGGAGTTTCCTAAGCAACG 3′) correspond to Exon 2 of Il2ra, which in the mutant has been replaced with a PGK-neo cassette, and yield a 280 bp amplimer with wild type DNA templates. The other two primers (5′ CTTGGGTGGAGAGGCTATTC 3′, and 5′AGGTGAGATGACAGGAGATC 3′) correspond to the neo gene, and direct the amplification of a 280 bp amplimer from mutant DNA templates. To distinguish between Gata3 carriers and wild type mice, we used a 3-primer PCR assay of our own design. For this test, one primer-pair (forward primer, 5′ CCCTAAACCCTCCTTTTTGC 3′, and reverse primer 5′ GATACCTCTGCACCGTAGCC 3′) flanked the site of the engineered disruption in Exon 2, and produced a 399 bp amplimer with wild type templates; that forward primer and second reverse primer (5′ GTTTTCCCAGTCACGACGTT 3′), based on sequences within in lacZ, yielded a 320 bp amplimer that is specific to the Gata3 allele.

PCR products plus 2 ul loading buffer (bromophenol blue in 20% Tris-buffered sucrose) were electrophoresed through 3.25% NuSeive 3:1 agarose gels (Lonza, Rockland, ME, USA). Gels were stained with ethidium bromide (0.5 ug/mL) and photographed under ultraviolet light. For sequence analysis, about 1.5 ug of individual PCR amplimers were concentrated into a 30 ul volume using QIAquick PCR Purification kits (Qiagen, Valencia, CA, USA). Purified amplimers were shipped to SeqWright, Inc. (Houston, TX, USA) for primer-extension analysis.

mRNA analysis

Total RNA was isolated from skin and thymus samples taken from 1-month-old mutant and wild type mice mice using the Nucleospin® RNA L kit by Macherey-Nagel (Easton, PA, USA). cDNA was generated using the SMARTer™ RACE cDNA amplification kit (Clontech Laboratories). To amplify Gata3-specific cDNA, primer pairs that flanked exon junction boundaries were used in “step-down” PCR reactions. The products of this initial reaction were diluted 1:10 in Tricine-KOH buffer (10 mM, pH 8.5) plus 1 mM EDTA, and were amplified again in standard PCR reactions using the same or nested primer pairs. Second-round amplimers were purified (as described above) and shipped to SeqWright, Inc., for primer-extension sequencing.

Results

Mapping jal to a mouse chromosome

To determine if jal might be carried on the mouse X chromosome, we conducted reciprocal crosses of homozygous mutant mice with wild type mice from the C57BL/6 J strain. Since the F1 progeny of both genders were phenotypically wild type [see Additional file 4], we confirm that the jal mutation is recessive, and conclude that it must reside in an autosomal portion of the genome.

To determine an autosomal location for the jal mutation, we crossed (C57BL/6 J × C3H/HeJ-jal)F1jal/+ females back to their jal/jal sire. This cross produced 43 mutants and 60 wild type progeny, not significantly different from the 1 mutant : 1 wild type ratio expected for a testcross (χ2 = 2.81; P > 0.09). DNA samples isolated from these 103 backcross (N2) progeny were analyzed for 93 PCR-scorable microsatellite markers from throughout the mouse genome, including two from the pseudoautosomal region on the X and Y chromosomes. The average spacing of these markers was 16 cM, with the largest gap being a 31 cM interval on Chr 4. Among the markers tested, only those from the centromeric portion of Chr 2 showed an inheritance pattern significantly different from the 1 parental : 1 recombinant ratio predicted if the marker and jal were independently assorted (see Figure 2). The largest deviation (82 parental and 21 recombinant types; χ2 = 36.13; P < 1.85 × 10-9) was observed for marker D2Mit1, which is located 2.23 cM from the centromeric end of Chr 2 [15].

Figure 2

Inheritance of and 93 microsatellite markers, tested for goodness-of-fit with an independent-assortment model. Each microsatellite marker tested is represented by a single bar positioned on the horizontal axis to show its approximate location in the mouse genome. Markers from odd chromosomes are in black, those from even chromosomes are in blue. Results are plotted as negative log-transformed P values calculated by the chi-squared method (with 1 degree of freedom). Bars descend below the baseline for those markers where more recombinant types (i.e., jal inherited from the F1 mother together with a C57BL/6-derived marker allele, or jalinherited with a C3H/HeJ-derived marker) than parental types (jal inherited from the F1 mother together with a C3H/HeJ-derived marker allele, or jalinherited with a C57BL/6-derived marker) were observed in a set of 44 family members initially typed. Additional mice (up to all 103 in the backcross panel) were typed for markers that showed a surplus of parental types such that goodness-of-fit testing with the expected 1:1 ratio gave P < 0.1. Only markers from proximal Chr 2 showed a significant (above the orange line, where P < 0.05) or highly significant (above the red line, where P < 0.01) excess of parental types, indicative of linkage with jal.

Meiotic fine-mapping

To more precisely locate jal on proximal Chr 2, we bred (A/J × C3H/HeJ-jal/J)F1, jal/+ females back to C3H/HeJ-jal/jal males, since this strain combination offered more microsatellite and single nucleotide polymorphisms (SNPs) than the C57BL/6 J and C3H/HeJ strain combination. These N2 mice were typed for jal and six microsatellite markers on proximal Chr 2, as summarized in Figure 3. The 374 progeny from this backcross generation fit well with the expected 1 wild type : 1 mutant ratio expected for a testcross (χ2 = 0.17; P > 0.67), so mutants appear to be equally viable as their wild type, heterozygous littermates. Segregation of markers among this large N2 family indicates that jal is located between D2Mit359 and D2Mit80, a span of about 11 cM that contains some 11.66 Mb of DNA [16].

Figure 3

Segregation of and five microsatellite markers on proximal Chr 2 among 374 backcross mice. The five markers typed are shown to the left of the diagram. The haplotype transmitted by the heterozygous F1 dam is depicted. Open boxes indicate A/J-derived alleles; solid boxes indicate C3H/HeJ-derived alleles. The centromere is indicated by a knob at the top of each haplotype. The number of progeny inheriting each haplotype is shown below it. Genetic distances are shown to the right. The red arrows indicate that in these recombinant mutants, the mutant jal allele must be located below D2Mit359, but above D2Mit80. The blue arrows similarly indicate that in these recombinant wild type mice, the normal jalallele must be located below D2Mit359, but above D2Mit80.

Complementation testing between jal and a targeted mutation in Il2ra

A recent genome-wide association study for alopecia areata (AA, OMIM #104000) in humans has implicated several genes, including IL-2RA (for interleukin 2 receptor, alpha chain) in the development of disfiguring hair loss [18]. Because AA appears similar in at least some ways to the mutant jal/jal phenotype in mice [10], and because IL-2RA is located on human Chr 10p15.1—a region that is orthologous with the D2Mit359 and D2Mit80 interval on Chr 2 in mouse—we decided to test jal for complementation with the recessive Il2ra loss-of-function mouse mutation [12]. Because mice homozygous for the targeted mutation show poor survival, we crossed Il2ra/+ heterozygous females with jal/jal males. If jal were a defect in Il2ra, then the mice that inherit jal and Il2ra could express no wild-type gene product, and would therefore be expected to show some mutant phenotype, perhaps as mild as defective vibrissae (as displayed by all jal/jal mutants) or perhaps as severe as the slower growth and progressive wasting (cachexia) seen in mice homozygous for Il2ra[19]. Alternatively, if jal and Il2ra are distinct genes, then all of the progeny would be phenotypically normal (since both mutations are recessive).

This cross yielded 19 offspring that were typed by PCR for the Il2ra targeted disruption [Additional file 5] and observed for 30 weeks. DNA typing identified 11 Il2ra carriers (5 females and 6 males) and 8 mice without the targeted disruption (7 females and 1 male), not significantly different from the 1:1 ratio expected for a test cross (χ2 = 0.47; P = 0.49). All of these mice (Il2ra carriers and noncarriers) displayed normal vibrissae and body hair. Furthermore, Il2ra carriers and noncarriers showed indistinguishable growth rates (over a period of 30 weeks), with no signs of the cachexia seen in Il2ra/Il2ra controls [Additional file 6]. These data suggest that jal is not an allele of Il2ra.

Refinement of the meiotic map for jal

The 41 mice from the (A/J x C3H/HeJ-jal/J)F1 × C3H/HeJ-jal/J backcross that were recombinant in the D2Mit359 and D2Mit80 interval were next typed for four, single-nucleotide polymorphisms designated SNP1, SNP2, SNP3 and SNP4 (see Additional file 2 and Additional file 3). This analysis identified six crossovers between SNP1 and jal, and one crossover between jal and SNP2, placing the jal mutation between these two markers (see Figure 4a), a 0.55 Mb span that does not include Il2ra. Of the ten genes or predicted genes [16] that do map to this interval, only one—Gata3 (for GATA binding protein 3)—is known to be expressed in skin [20,21].

Figure 4

Physical maps of the region on mouse Chr 2. (a) Molecular markers and genes on mouse Chr 2 that are linked with jal. Segregation data from the 374-member backcross panel shown in Figure 3 placed jal between microsatellite markers D2Mit359 and D2Mit80 (shown in blue), an interval that also includes Il2ra (shown in gray). Single-nucleotide polymorphisms (SNP1-4, see Additional file 2 and Additional file 3) were used to more precisely locate crossovers among the 41 mice recombinant in this interval. The number of crossovers located between the various pairs of adjacent markers are shown on the chromosome map, which is drawn to the 5 Mb scale shown. Seven recombinants located jal between SNP1 and SNP2 (shown in red). The region between SNP1 and SNP2 is expanded below the chromosome map (drawn to the 0.1 MB scale bar shown), to show the locations of the 10 candidate genes (represented by orange boxes) that populate this span. Of these ten genes, only one, Gata3 (shown in yellow), is known to be expressed in skin. (b) The Gata3 gene is expanded to show the arrangement of exons, where taller boxes are coding regions and shorter boxes are the 5′ or 3′ untranslated regions. Gata3 is transcribed from the reverse strand, but is drawn here so that the six exons are shown in ascending numerical order. The length of each exon (in bp) is shown below the corresponding box. The portions of exons shaded green have been sequenced in C3H/HeJ and C3H/HeJ-jal/J DNA, but no differences were found.

Evaluation of Gata3 as the possible genetic basis of the jal mutation

To determine if jal could be a mutant allele of the Gata3 gene, we imported a mouse carrying an engineered Gata3 null allele, Gata3[14], for complementation testing. To create litters of half experimental (doubly heterozygous) and half control offspring (carriers of the jal allele, only), we crossed Gata3/+ heterozygous females with jal/jal males. If jal is the result of a defect in Gata3, then the mice that inherit both jal and Gata3 could express no wild-type gene product, and would therefore be expected to show defective coats and vibrissae. Alternatively, if jal and Gata3 are distinct genes, then the dihybrid progeny (jal/+, Gata3/+) would be phenotypically normal.

This cross yielded 22 offspring that were typed by PCR for the Gata3 targeted disruption. DNA typing identified 11 Gata carriers (6 females and 5 males) and 11 mice without the disruption (10 females, 1 male), as expected for a test cross (Figure 5a). All Gata3 carriers displayed defective vibrissae and body hair (see Figure 5c and e), while those without the targeted mutation in Gata3 appeared phenotypically normal (Figure 5b and d). Thus, jal and Gata3 fail to complement, suggesting that these mutations are allelic.

Figure 5

The recessive and mutations fail to complement in doubly heterozygous mice. (a) Typical results of a 3-primer PCR test designed to identify Gata3carriers. The 320 bp band and a fainter, high-molecular-weight band (marked with an asterisk) are specific to the mutant allele. The size standard shown (MM) is a 50 base pair ladder. A 10-day-old litter from a cross of Gata3/Gata3x jal/jal included pups displaying wild type (b) or mutant (c) hair development. All phenotypically wild type pups showed the 399 bp band only, and the phenotypically mutant pups all carried the targeted mutation. The snouts of one wild type (d) and one mutant (e) pup from the same litter are enlarged to show normal vs. defective vibrissae formation, respectively.

All coding regions of the Gata3 gene, plus the 5′ untranslated regions encoded by two alternative 1st exons (see Figure 4b for transcript diagram and summary) were sequenced in DNA isolated from C3H/HeJ and C3H/HeJ-jal/J mice. However, we found no differences in DNA sequence between these coisogenic wild type and mutant strains. In addition, using total RNA isolated from skin and from thymus, we amplified (and sequenced) only identically-spliced Gata3 cDNA from both wild type C3H/HeJ and C3H/HeJ-jal/J mutant mice (see Additional file 7).

Discussion

The results presented suggest to us that the jal mutation is a mutant allele of the Gata3 gene on mouse Chr 2. We therefore recommend that the jal designation be changed to Gata3. While we have not yet been able to pinpoint a sequence-level change in Gata3, our analysis has mostly been limited to coding regions. We hypothesize that the Gata3 defect is likely to be a regulatory mutation (perhaps located in the promoters, introns, or 3′ untranslated region) that—in some fashion—impacts expression, processing, or degradation of the Gata3 transcript, although we find that the Gata3-001 transcript appears to be normally spliced. (We found no evidence for expression of the alternative Gata3-201 transcript in total RNA isolated from skin or thymus.) Quantitative and qualitative evaluation of Gata3 transcripts or protein in the epidermis and hair follicles of C3H/HeJ-jal mice versus wild type controls could help refine this array of possibilities. This prediction (that the Gata3 defect is likely to be a regulatory mutation) does seem consistent with the variable phenotypic presentation of focal alopecia that we observe in Gata3/Gata3 mice (see Additional file 1). Since at least some patches of normal fur are seen on most if not all mutants (with some mutants showing almost entirely normal coats), we anticipate that a standard primary protein sequence (albeit improperly regulated) is likely to be encoded by the Gata3 allele.

The positional assignment of jal did not reveal (as with our introductory examples, refs. 1–9) an unsuspected function of Gata3 in skin, since the study of mouse strains engineered to carry targeted mutations have previously indicated a role for Gata3 in hair follicle development and skin cell lineage determination. Mice homozygous for germline Gata3 null mutations die around embryonic day 11 [22,23], precluding a detailed assessment of the functional role of Gata3 in hair follicle morphogenesis. However, some investigators have rescued mutant skin by transplantation to athymic hosts [24], or else ablated Gata3 specifically in the epidermis and hair follicles to reveal a crucial role in skin [25]. Since the mouse juvenile alopecia phenotype (patchy hair loss) is distinct from that of these conditionally-targeted mutants (complete baldness)—whatever its molecular basis—we believe that Gata3 likely offers a novel mutant allele, compared to the existing set of engineered Gata3 disruptions. Addition of this viable and phenotypically-unique natural variant to the Gata3 mutational inventory will surely allow new approaches to the functional analysis of this locus, just as the recent assignment of the spontaneous mouse frizzy (fr) and rat “hairless” (fr) mutations to the prostasin gene [26] has productively advanced the in vivo analysis of Prss8 function in mammalian skin [27-29].

Haploinsufficiency of human GATA3 (due to loss-of-function mutation of GATA3) causes a dominantly-inherited syndrome of hypoparathyroidism, sensorineural deafness, and renal disease (HDR, OMIM #146255) also known as Barakat syndrome. Notably, HDR syndrome does not appear to involve immune-related disorders or alopecia [30,31]. The mouse Gata3 mutation has been shown to generate deafness in heterozygotes [32-34], and is considered a model for HDR. It would certainly be interesting to investigate parathyroid, cochlear, and renal function in Gata3 homozygotes and heterozygotes. In any case, a molecular explanation for the distinct modes of inheritance and phenotypic presentations of juvenile alopecia in mice versus HDR in humans will require discovery of the precise structure of the Gata3 allele.

Histological observation of immune cell infiltrates associated with follicular dystrophy in AA [35,36] combined with Petukhova et al.’s linkage of genes involved in both innate and acquired immunity (including IL-2RA) to AA susceptibility [18] seem to firmly establish AA as an autoimmune disorder. Although Gata3 is known to play a crucial role in T cell development [22,37], our elimination of Il2ra as the basis of the mutant phenotype as well as McElwee et al.’s failure to detect any signs of hair follicle inflammation in jal/jal mutants [10] suggest that mouse juvenile alopecia does not provide an ideal model for AA. However, it remains possible that juvenile alopecia could provide an animal model for at least some forms of focal alopecia which may have their primary defect in the hair follicle and lack an inflammatory component, but which may nonetheless be diagnosed as AA based on similar pathophysiology (i.e., patchy hair loss). Indeed, the future study of mouse juvenile alopecia may be helpful in identifying such a homologous human condition, defining approaches for distinguishing that disorder from AA, and in developing appropriate, specialized treatments.

Conclusions

The recessive jal mutation in mice maps to proximal Chr 2, and has been shown by complementation testing to be a variant allele of the Gata3 gene. While further study will be needed to discover the molecular defect in Gata3 that is the basis of the mutant phenotype, this spontaneous mouse variant promises to provide an animal model for some forms of focal alopecia in humans that have their primary defect in the hair follicle and lack an inflammatory component.

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

FR led all aspects of the genome-wide linkage screen, including experimental design, data acquisition and interpretation. AMF led all aspects of the Chr 2 fine-mapping, Gata3 complementation testing, and sequencing of Gata3; including experimental design, data acquisition and interpretation. KMC, LAR and AMF conducted the Gata3 cDNA analysis. NV-S conducted complementation testing between jal and Il2ra. EBA, AS, JMH and DVS made substantial contributions to the genome-wide and regional genetic analyses. LAR, KMC and SRH contributed significantly to the SNP marker analysis. TRK conceived of the study, carried out all procedures involving mice, and drafted the manuscript. All authors read, edited, and approved the final manuscript.

Authors’ information

TRK is a professor in the Department of Biomolecular Sciences at Central Connecticut State University (New Britain, CT). FR was a student in the Master of Arts program in Biomolecular Sciences, and AMF, EBA, AS, NV-S, JMH, LAR, KMC, SRH and DVS were undergraduates majoring in Biomolecular Sciences or Biochemistry at CCSU when they conducted this research.

Additional file 1

Three-month-old mutants from a (C3H/HeJ-/J x C57BL/6 J)F× C3H/HeJ-/J backcross display variable expressivity of the juvenile alopecia phenotype.

Click here for file

Additional file 2

Description of SNP markers referred to in the Ramirez . (2013) text.

Click here for file

Additional file 3

Location of SNP markers referred to in the Ramirez . (2013) text.

Click here for file

Additional file 4

Fdata from reciprocal crosses in mice tests the juvenile alopecia mutation () for X versus autosomal linkage.

Click here for file

Additional file 5

DNA typing for the or alleles among the progeny of a complementation cross, /x

Click here for file

Additional file 6

The recessive and mutations complement in doubly heterozygous mice.

Click here for file

Additional file 7

Sequence analysis of splice junctions in wild-type C3H/HeJ and mutant C3H/HeJ-cDNA.

Click here for file