The mouse curly whiskers (cw) mutations are recessive alleles of hephaestin-like 1 (Hephl1)

Abstract

The spontaneous, curly whiskers mutation (abbreviated cw) generates kinky, brittle vibrissae in homozygous mice. Although cw has been mapped to the centromeric end of mouse Chromosome 9, no particular gene has been causally implicated, and this lack of genetic assignment has stymied cw's complete molecular and functional analysis. As a foundation for its positional cloning, we have fine-mapped cw to a small, 0.57 Mb interval that contains only three skin-expressed genes, including hephaestin-like 1 (Hephl1), which encodes a membrane-bound, multi-copper ferroxidase. Sequence analysis of all Hephl1 coding regions in cw/cw mutants revealed a single-base-pair substitution that alters Hephl1 mRNA splicing, and is specific to the cw allele, only. Sequence analysis of a second, independent, re-mutation to curly whiskers (that we verified by complementation testing with cw and have designated cw^2J) revealed a distinct defect in Hephl1 (a frame-shifting, single-base-pair insertion) that is specific to cw^2J. The results presented strongly suggest that defects in the Hephl1 gene are the molecular basis of the classical, curly-whiskers mutant phenotypes.

Highlights

• •Genetic mapping identifies a small number of candidates for the mouse cw mutation.
• •Sequence analysis of one of these candidates, Hephl1, reveals a cw-specific defect.
• •Analysis of skin mRNA indicates that the Hephl1^cw transcript is aberrantly spliced.
• •Complementation testing identifies a distinct re-mutation to curly whiskers, cw^2J.
• •Sequence analysis of Hephl1 reveals a cw^2J-specific, frameshifting insertion.

1. Introduction

The recessive cw mouse mutation—named “curly whiskers” to highlight the most obvious mutant phenotype it controls—was discovered in 1958 in a subline of CBA/Cbi mice at the Chester Beatty Research Institute (London, UK), and its linkage with short ear (Bmp5) on Chromosome (Chr) 9 was first reported by Falconer and Isaacson [11,12]. The strongly curled vibrissae in mutants (see Fig. 1) are easily scored soon after birth, and persist throughout the life span. In 1967, Lyon and Butler [21] located cw very near the centromere on Chr 9, but this spontaneous variant has not been assigned to any causative gene. A distinct recessive mutation, initially named “tail hair depletion” (abbreviated thd), was reported by Les and Roths [20] to affect hair development in homozygotes while also generating a dominant, gain-type barrier to tail-skin graft compatibility. Mice carrying the thd mutation are now extinct, but while still extant this mutation was found (in a standard complementation test conducted by Roths) to be an allele of cw, and was renamed cw^thd in 1978.

Fig. 1.

The mutant curly whiskers phenotype. (A) A 25-day old CWD/LeJ mouse, homozygous for the cw mutation, and (B) a close-up of the mutant vibrissae.

These mutant phenotypes suggest that the gene identified by cw alleles must play important roles in both hair-follicle development and in histocompatibility. However, a detailed analysis of this pleiotropic locus has long been hampered by a lack of molecular probes that might allow interrogation of this gene's normal (and disrupted) structure and function. Here we have taken a positional-candidate approach [6] toward making such probes available. As a basis for this effort, we genetically mapped the original cw mutation to a very small interval on mouse Chr 9, where only three skin-expressed genes are also located. Next, DNA sequence analysis of one of those candidate genes in mice homozygous for cw or for a second, spontaneous, re-mutation to curly whiskers revealed distinct defects that are predicted to impair protein function. Taken together, the evidence we describe strongly suggests that these inherited defects in Hephl1 (for hephaestin-like 1) are the molecular basis of the developmental and immunogenetic phenotypes displayed by the classical curly-whiskers mutants.

2. Materials and methods

2.1. Mice

Animals were housed and fed according to Federal guidelines, and the Institutional Animal Care and Use Committee (IACUC) at Central Connecticut State University (CCSU) approved of all procedures involving mice (Animal Protocol Applications #142, #158, #162 and #163). Mice from the standard inbred strains C57BL/6J (JAX Stock #000664), C57BL/10SnJ (JAX Stock #000666), and DBA/2J (JAX Stock #000671); wild-derived, inbred CAST/EiJ mice (JAX Stock #000928); and inbred CWD/LeJ mice, homozygous for non-agouti (a), cw, and dilute (Myo5a^d)(JAX Stock #000284) were obtained from The Jackson Laboratory (Bar Harbor, ME, USA). Although CWD/LeJ strain embryos were cryopreserved in 1989 and again in 1999 by crossing cw Myo5a^d/+ Myo5a^d with cw Myo5a^d/cw Myo5a^d, upon its most recent reconstitution in 2015, only cw/cw offspring were recovered, suggesting that the CWD/LeJ stock no longer segregates for cw, cw⁺.

Twenty-four males of the strain B10.SM H2^v H2-T18^b/(70NS)Sn/J (JAX Stock #000456) were imported to CCSU (after cryo-recovery from liquid nitrogen storage since 1999) in June of 2018. These mice all displayed curly whiskers. In addition, we obtained (from Jane Ober, DNA Resource, The Jackson Laboratory, Bar Harbor, ME) DNA samples isolated from two, independent B10.SM H2^v H2-T18^b/(70NS)Sn/J mice that were archived in 1983 (without notation as to their whisker phenotype).

A frozen spleen sample (archived in 1999) from a mouse homozygous for the cw-Bmp5^se-tk haplotype—descended separately from the same, linkage-tester stock from which the CWD/LeJ line was derived in the 1960's—was kindly donated to us by Simon Ball and Rachel Summerfield (MRC Harwell, Oxfordshire, UK).

2.2. DNA isolation and analysis

Genomic DNA was isolated from two- to three-mm tail-tip biopsies taken from two- to three-week-old mice using Nucleospin® Tissue kits distributed by Clontech Laboratories, Inc. (Mountain View, CA, USA), as directed. The polymerase chain reaction (PCR) was performed in 13 μl reactions using the Titanium PCR kit from Clontech Laboratories, as directed. Oligonucleotide primers for PCR were designed and synthesized by Integrated DNA Technologies, Inc. (Coralville, IA, USA), based on sequence information available online [9,29]. To score PCR product sizes for microsatellite markers [8], reactions plus 2 μl loading buffer were electrophoresed through 3.5% NuSieve® agarose (Lonza, Rockland, ME, USA) gels. Gels were stained with ethidium bromide and photographed under ultraviolet light.

DNA markers based on single nucleotide polymorphisms (SNPs) previously reported to differ between wild-derived CAST and most standard inbred mouse strains [24] were also scored. These markers (herein designated SNP1–5) are described in detail in Supplementary Table S1, Supplementary Table S2. For DNA sequence analysis, about 1.5 μg of individual PCR amplimers were purified and concentrated into a 30 μl volume using QIAquick® PCR Purification kits (Qiagen Sciences, Germantown, MD, USA) prior to primer-extension sequence analysis performed by the Keck Foundation Resource Laboratory at Yale University (New Haven, CT, USA).

To rapidly determine cw and cw^2J genotypes, especially in phenotypically wild type mice, 0.5 μg of individual PCR amplimers were purified and concentrated into a 30 μl volume using QIAquick® PCR Purification kits. For cw genotyping we used forward (5′ AAACGTGCTCTGAGATGG 3′) and reverse (5′GTTGCCTTGGAAATAAACTCC 3′) primers that flank the Hephl1 splice-acceptor defect we describe in the text. For cw^2J genotyping we used forward (5′ ATCCAAGGCCTTCTGTTAAGG 3′) and reverse (5′CAGGATGGGACAGACTTTGG 3′) primers that flank Hephl1, Exon 14, which contains the single-base insertion we describe in the text. 12 ul samples of the respective purified amplimers were then incubated with 10 units of AluI (for cw genotyping) or MlyI (for cw^2J typing) (New England BioLabs, Inc.; Ipswich, MA, USA) at 37 °C for 1 h prior to electrophoresis through 3.5% NuSieve® agarose gels at 145 V for 1 h.

2.3. RNA isolation and analysis

Total RNA was isolated from tail skin samples taken from 1-month-old mutant and control mice using the Nucleospin® RNA Midi kit by Macherey-Nagel (Bethlehem, PA, USA). From these samples, cDNA was generated using the SMARTer® RACE 5′/3′ kit (Clontech Laboratories, Inc.). To detect Hephl1-specific sequences, primer pairs that annealed in Exon 9 (5′-ACCTACAGGTGGACAGTGCCAGAAAGC-3′) and Exon 12 (5′-GCTGCTGATCTCATAGATCTGACCCATGCC-3′) were used to direct standard PCR amplifications of these cDNAs. Primers that annealed within Exon 4 of the mouse β actin (Actb) gene (5′-CCCAGCCATGACGTAGCCATCCA-3′ and 5′-GAAGCTGTAGCCACGCTCGGTCAG-3′) were used together with Hephl1 primers to provide an internal loading control. The resulting products were visualized in 3% NuSieve® agarose gels. For primer-extension sequencing, these amplification products were purified and concentrated (as described above) and shipped to the Keck Foundation Resource Laboratory at Yale University.

3. Results

3.1. Genetic mapping of the cw mutation on mouse Chr 9

To map the cw mutation with respect to molecular markers at the centromeric end of Chr 9, F₁ heterozygotes made by crossing CWD/LeJ-cw/cw mice with standard C57BL/6J mice were crossed back to CWD/LeJ mutants, producing 168 offspring. This intraspecific backcross (N₂) generation segregated for cw, dilute (Myo5a) and for several PCR-scorable, microsatellite DNA markers [8] on Chr 9. Supplementary Fig. S1 shows the string of markers transmitted by the F₁ parent to each of these 168 N₂ progeny. This haplotype analysis suggested that the cw gene must be located about 10% recombination centromeric to D9Mit64.

To more precisely locate cw in the region between D9Mit64 and the centromere, a new set of F₁ heterozygotes was produced by crossing CWD/LeJ to the wild-derived CAST/EiJ strain (since this strain combination offered more microsatellite and single-nucleotide marker dimorphisms than did the CWD/LeJ and C57BL/6J strain-pair), and these cw/+ heterozygotes were crossed back to CWD/LeJ mutants. The 994 N₂ progeny resulting from this inter-subspecific backcross were typed for curly whiskers, dilute pigmentation, and for 10 microsatellite markers on proximal Chr 9, as summarized in Fig. 2. This haplotype analysis indicated that cw must lie between markers D9Mit61 and D9Mit60, and very close to marker D9Mit220 (which was never meiotically separated from cw in this large backcross panel).

Fig. 2.

Segregation of alleles of cw, Myo5a (dilute, d), and eight dimorphic microsatellite markers among 994 progeny from an inter-subspecific mouse backcross. Heterozygous F₁ mice (CWD/LeJ x CAST/EiJ) were backcrossed to homozygous CWD/LeJ mutants. The resulting progeny were scored for their fur texture and colour, and a DNA sample from each mouse was typed for the microsatellite markers shown at the left. Only the Chr 9 haplotype inherited from the F₁ parent is shown (with a knob at the top of the haplotype representing the centromere), and the number of mice inheriting that haplotype is shown below it. Very similar numbers of wild type and mutant progeny—as expected for a testcross (χ² = 1.95, P > .16)—suggest that the mutant phenotype is both fully penetrant and fully viable. The five wild type and the five mutant recombinants marked with an asterisk show that the cw locus must lie between D9Mit61 and D9Mit60. D9Mit220 was not separated from cw in this backcross analysis. Genetic distances in percentage recombination are shown to the right (± 1 Standard Error).

DNA samples from the ten mice identified as having a crossover between D9Mit61 and cw or between cw and D9Mit60 were typed next for five, single-nucleotide polymorphisms (SNPs) known to lie in the D9Mit61 to D9Mit60 interval. These five SNP markers are described in detail in Supplementary Table S1, Supplementary Table S2, and are designated herein as SNP1–5. This analysis located the eight crossovers that fell centromeric to cw between SNP1 and SNP2, and the two crossovers that fell distal to cw between D9Mit220 and SNP4 (see Fig. 3A), thus restricting the possible location of cw between SNP1 and SNP4 (very near SNP2, SNP3 and D9Mit220, which were not meiotically separated from each other or from cw).

Fig. 3.

Physical maps of the cw-critical region on mouse Chr 9. (A) The relative positions of three microsatellite (D9Mit) markers and five single-nucleotide (SNP) markers that closely flank cw are shown with a 0.1 Mb scale bar. The eight crossovers that fell centromeric to cw (shown in blue, and see Fig. 2) were localized between SNP1 and SNP2, while the two crossovers that fell telomeric to cw (shown in red, and see Fig. 2) were located between D9Mit220 and SNP4. The extent of 17 known genes that lie in the interval between D9Mit61 and D9Mit60 are shown below the line that represents Chr 9. Null alleles of the six genes depicted in brown do not impact hair development in homozygotes. Of the remaining 11 genes, those depicted in blue are not expressed in skin. Because cw must be located in the 0.57 Mb interval between SNP1 and SNP4, the skin-expressed genes (B09Rik, Ankrd4, and Hephl1, depicted in green) are most likely to harbor the causative mutation. (B) The Hephl1 gene is reversed and expanded to show the 20 exons it comprises. Tall green boxes represent coding regions and shorter white boxes represent the untranslated regions. The number below each exon is its length in base pairs. The mutant Hephl1 allele found in cw mice is drawn to show a single-base-pair transition just 5′ to Exon 11 (indicated by the red asterisk) that eliminates a splice acceptor site, resulting in the two variant transcripts diagrammed here (and described further in the text). The mutant Hephl1 allele found in cw^2J mice is drawn to show the insertion of a single A residue in Exon 14 (at the red dagger), that is predicted to generate an out-of-frame stop codon later in Exon 14 (as described further in the text). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.2. Evaluation of Hephl1 as the possible genetic basis of the cw mutation

The 0.57 Mb span from SNP1 to SNP4 (where the cw mutation must lie) includes ten expressed genes (Fig. 3A). Of these ten, at least six (Piwil4, Fut4, Mre11a, Gpr83, Izumo1r and Panx1) have been knocked out genetically with no reported effect on hair morphology ([2,3,14,15,32,33], respectively), making them unlikely to harbor a mutation that causes the curly-whiskers mutant phenotype. Of the remaining four genes, only three (1700012B09Rik, Ankrd49 and Hephl1) are known to be expressed in skin [7], where the cw mutation has its most obvious effect in mouse mutants. Among these three, Hephl1 [5] seemed to us the most likely gene candidate, since: 1) its product is expressed on the cell surface, where Les and Roths [20] suggested the histo-antigenic cw^thd gene product should be found; and 2) its product is a copper-dependent ferroxidase, and generalized copper deficiency (as in humans with defects in the Cu²⁺-transporting ATPase, alpha polypeptide gene, ATP7A) is known to cause clinical features that include kinky hair [27].

To test this prime candidate further, the genomic DNA sequence of all Hephl1 coding regions was determined for CWD/LeJ-cw/cw mutants and for C57BL/6 J control mice. This analysis revealed only a single-base difference in mutants compared to wild type: an A-to-G transition two bases upstream of Exon 11 (see Fig. 4A). Because this mutation appears to destroy a splice-acceptor signal, we generated cDNA from tail-skin transcripts isolated from cw/cw and control wild type tail skin, and PCR-amplified Hephl1 sequences between Exon 9 and Exon 12. As shown in Fig. 4B, amplification of cDNA from wild-type skin yielded a 606 bp amplimer as expected for the normal splicing of Exons 9 through 12 (and see sequence details in Supplementary Fig. S2A). By contrast, amplification of Hephl1 sequences from cw/cw skin-derived cDNA yielded two smaller PCR products of 581 and 393 bp (Fig. 4B). Primer-extension analysis of the 581 bp splice variant (Variant 1) revealed a novel junction between Exon 10 and a cryptic splice-acceptor site within Exon 11. This truncated Exon 11, designated Exon 11(Δ25), is 25 nucleotides shorter than the standard exon. This aberrant splice is predicted to disrupt the reading frame and alter three amino acids before an out-of-frame stop codon would terminate translation (as shown in Supplementary Fig. S2B). Primer-extension analysis of the 393 bp splice variant (Variant 2) showed Exon 10 joined with Exon 12 (see Supplementary Fig. S2C). The skipping of Exon 11 in this cw-specific transcript is predicted to omit 71 amino acids from the mutant protein product but maintain the normal downstream reading frame. The full-length wild-type and two, cw-specific, variant Hephl1 transcripts are diagrammed in Fig. 3B.

Fig. 4.

Comparison of Hephl1 in wild type and cw/cw mutant mice. (A) All coding regions were sequenced in wild type and in cw/cw mutant genomic DNA, but only a single difference, at the Intron 10–11 to Exon 11 boundary was found. Shown here are nucleotides 9:15076971 to 9:15076921 from reference sequence GRCm38.p6 (EMGS 2019); with intronic sequences in lower case, exonic sequences in upper case. The pyrimidine-rich tract and the 5′-AG-3′ dinucleotide that compose the standard splice-acceptor signal are indicated over the sequence labeled wild type. The A-to-G substitution found in cw/cw mutants (at position 9:15076948) is indicated by a red asterisk on the sequence labeled cw. (B) Total RNA isolated from wild type, homozygous mutant and heterozygous tail skin was copied into cDNA, and PCR amplified using primers that annealed within Exons 9 and 12 of Hephl1. The 606 bp length (shown here) and sequence (see Fig. S2A) of the amplimer copied from wild type (+/+) cDNA was that expected for the normal splicing of Exons 9 through 12. By contrast, amplification of cDNA derived from mutant (cw/cw) skin yielded two different-sized products, of 581 bp (Variant 1) and 397 bp (Variant 2), in about an equal abundance. These variant, cw-specific PCR products were isolated and sequenced, as shown in Fig. S2B and 2C, respectively. Notably, since Variant 1 generates an early stop codon, it is likely to be the target of nonsense-mediated decay [4], and so it may be transcribed at an initially higher level than Variant 2. Amplification of Hephl1 sequences from cDNAs derived from heterozygous (+/cw) skin yielded all three product sizes. While the Variant 2 transcript should be stable (since it is predicted to be fully translated), it appears much less abundant than the wild type transcript in the cDNA pool based on heterozygous skin. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Next, we developed an AluI-sensitivity assay (see Supplementary Fig. S3A) that was used to rapidly screen among 57 mouse strains that are not associated with the cw mutation, and found the standard splice-acceptor sequence in all of them. Only one strain, a linkage-tester stock homozygous for the cw-Bmp5^se-tk haplotype (from the MRC Harwell Institute, Oxfordshire, UK), was found to encode the same Hephl1 DNA defect as the CWD/LeJ strain (see Supplementary Fig. S3B & C).

3.3. Rediscovery and genetic characterization of the cw^2J mutation

One strain cryopreserved at The Jackson Laboratory, B10.SM H2^v H2-T18^b/(70NS)Sn/J (a C57BL/10 strain congenic for a recombinant H2 haplotype), was described in some records (but not others) as carrying the cw mutation. We obtained DNA samples from two independent mice from this strain that were archived in 1983 (without notation as to their whisker phenotype), but these both showed the wild type sequence at the splice-acceptor site 5′ to Exon 11 in Hephl1 (see Supplementary Fig. S3B & C). To determine whether B10.SM H2^v H2-T18^b/(70NS)Sn/J stock that was cryopreserved in 1994 might carry the cw mutation, we obtained 24 reconstituted males—which all displayed curly whiskers—and crossed some with known cw homozygotes or heterozygotes (see Table 1). The failure of a recessive defect homozygous in the B10.SM H2^v H2-T18^b/(70NS)Sn/J strain to complement cw (see Fig. 5) suggests that this defect is an allele of cw. Because this congenic strain is unrelated to the original cw mutation (and because it lacks the point mutation we have found to be specifically associated with cw), we hypothesized that this recessive variant might be a spontaneous re-mutation that we designated “curly whiskers; curly whiskers 2 Jackson”, abbreviated cw^2J.

Table 1.

Initial crosses of curly-whiskered, B10.SM H2^vH2-T18^b/(70NS)Sn/J males.

Cross	Female partner strain (cw genotype)	Offspring phenotype
		Wild type	Curly
1	C57BL/10 (+/+)	8	0
2	CWD/LeJ (cw/cw)	0	12
3	CWD.D2 (cw/cw)	0	19
4	CWD.D2 (+/cw)	13	7

Males of the B10.SM H2^vH2-T18^b/(70NS)Sn/J strain, all showing curly whiskers, were crossed with female mice, as indicated. The resulting offspring were assessed for their whisker phenotype at weaning. The result of Cross 1 suggests that the mutant hair phenotype displayed by B10.SM males is generated by homozygosity for a recessive mutation. The results of Crosses 2–4 show that this recessive mutation does not complement the cw mutation. We therefore designate the recessive mutation that arose on the B10.SM background curly whiskers-2 Jackson, abbreviated cw^2J. Typical wild type and mutant pups from Cross 4 are shown in Fig. 5. The number of wild type and mutant offspring from Cross 4 is not different from the 1:1 ratio expected for a test cross (Χ² = 1.8; P > .17).

Fig. 5.

Phenotype and genotype of mice from a cross of +/cw with curly-whiskered B10.SM H2^vH2-T18^b/(70NS)Sn/J mice (i.e., Cross 4 in Table 1). This cross produced 20 offspring, 13 with straight whiskers and 7 with curly whiskers. The snout of a typical straight-whiskered mouse (A) and a typical curly-whiskered mouse (B) are shown at 2 weeks of age. This result suggests that the mutation that causes hair-curling in the B10.SM H2^vH2-T18^b/(70NS)Sn/J strain is recessive, and that this mutation is an allele of cw that we hereby designate cw^2J. (C) To confirm that all non-complementing, mutant offspring from this cross (labeled M) inherited one copy of cw, and that the phenotypically wild type offspring (labeled W) inherited a cw⁺ allele instead, a 203 bp amplimer that included the site of the Hephl1, splice-acceptor defect associated with cw was produced for each mouse, and tested for sensitivity to AluI digestion. By this assay (see Fig. S2), amplimers derived from cw templates are resistant to cleavage, while amplimers based on other templates are cut into 125 and 78 bp fragments.

Sequence analysis of all 20 coding regions of the Hephl1 gene in B10.SM H2^v H2-T18^b/(70NS)Sn/J-cw^2J mutants revealed just one DNA defect vs. C57BL/10SnJ controls: the insertion of a single A residue in Exon 14, codon 820 (see Fig. 6A). This frameshift mutation is predicted to alter 7 amino acids before an out-of-frame stop codon would terminate translation. While normal-sized Hephl1 PCR products may be amplified from cDNAs based on transcripts isolated from cw^2J/cw^2J mutant tail skin, those products appear less abundant than products amplified from cDNAs derived from wild-type skin (see Fig. 6B). This observation may indicate that cw^2J mutant transcripts are unstable, consistent with nonsense-mediated decay [4]. The full-length cw^2J transcript is diagrammed in Fig. 3B.

Fig. 6.

Comparison of Hephl1 in wild type and cw^2J/cw^2J mutant mice. (A) All coding regions were sequenced in wild type and in cw^2J/cw^2J mutant genomic DNA, but only a single difference, in Exon 14, was found. This mutation, the insertion of a single A residue in cw^2J/cw^2J mutants compared to wild type, is indicated with a red asterisk on the sequence labeled cw^2J. The sequence shown here includes nucleotides 9:15067163 to 9:15067125, from reference sequence GRCm38.p6 (EMGS 2019). This insertion is predicted to shift the translational reading frame on mutant transcripts, generating seven novel amino acids (residues 820–826, shown in red) followed by an early, out-of-frame stop codon. (B) Total RNA isolated from wild type, homozygous cw^2J/cw^2J mutant or heterozygous tail skin was copied into cDNA, and PCR amplified using one primer pair specific to Hephl1 together with a second primer pair specific to the β actin gene (Actb). The resulting amplimers were then separated by gel electrophoresis. While the levels of Actb signal amplified from the cDNAs derived from each different genotype is essentially equivalent, Hephl1 transcripts amplified from cw^2J/cw^2J templates appear markedly less abundant compared to amplimers based on wild type cDNAs. It is likely that this variant transcript, which cannot be fully translated, is unstable due to nonsense-mediated mRNA decay [4]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Finally, we developed an MlyI-sensitivity assay (see supplementary Fig. S4A) to rapidly screen among 60 mouse strains for defects in Hephl1, Exon 14, codon 820. The only amplimers resistant to endonuclease treatment (and verified by primer-extension analysis to possess an inserted adenosine residue) were based on our reconstituted B10.SM H2^v H2-T18^b/(70NS)Sn/J-cw^2J mice or on the B10.SM H2^v H2-T18^b/(70NS)Sn/J DNA samples that were archived in 1983 (see Supplementary Fig. S4B).

4. Discussion

4.1. The mouse cw mutations are mutant alleles of Hephl1

We have taken a positional-candidate approach to assign the classic cw mutation in mice to a splice-acceptor defect in the Hephl1 gene. Our identification of a second, distinct, allele-specific Hephl1 defect (a frameshift-inducing, single-base insertion) associated with the cw^2J mutation strongly supports this single-gene assignment. We therefore recommend that these curly-whiskers mutations be formally renamed “hephaestin-like 1; curly whiskers” (abbreviated Hephl1^cw), and “hephaestin-like 1; curly whiskers 2 Jackson” (abbreviated Hephl1^cw-2J). We suggest that the Hephl1^cw-2J mutation occurred spontaneously on the B10.SM H2^v H2-T18^b/(70NS)Sn/J background at least by 1983 (since both the archived DNA samples we analyzed are homozygous for the A insertion in Exon 14), and appears to have become fixed in that strain by its initial cryopreservation in 1999 (since all 39 mice reconstituted for us in 2018 displayed curly whiskers).

Because we restricted our DNA-sequence analysis to Hephl1 coding regions, only, it remains formally possible that the mutant curly-whiskers phenotype could require additional, unidentified defect(s) in the critical region that work in combination with the Hephl1 defects that we have described. While we think that such a “two-hit” mechanism is highly unlikely, we note that this model will be explicitly tested when mice homozygous for discrete, engineered variants of Hephl1, such as have already been produced (see [36,31,25]) or can be generated de novo by gene editing, are phenotypically evaluated.

4.2. Structure and function of Hephl1, and its mutant isoforms

Hephl1 encodes one of three muticopper feroxidases (in addition to hephaestin, Heph, and ceruloplasmin, Cp) that facilitate iron transport in a variety of tissues and display mostly distinct expression patterns in mammals [5]. The membrane-bound Hephl1 ferroxidase has 6 cupredoxin domains, with binding sites for 6 copper ions. Three of the copper ions form a trinuclear center at the interface of Domains 1 and 6, while the other three form mononuclear centers, and are organized in Domains 2, 4, and 6. In addition, Domain 6 includes a predicted iron-binding site, and is followed by a transmembrane domain [5]. The splice-acceptor defect that we have found associated with cw causes omission of Exon 11 from the mature transcript (designated splice Variant 2 in the text), and the resulting mutant Hephl1^cw protein is predicted to lack 71 amino acids normally found in Domain 4, but the downstream reading frame is unaltered. By contrast, the frameshifted variants we described (the Hephl1^cw-specific splice Variant 1 and the Hephl1^cw-2J transcript), since they encode early stop codons, are likely to be eliminated by nonsense-mediated decay [4]. Even if these truncated protein products were produced, they would lack Domains 5 and 6 (including the iron-binding site), and the C-terminal, membrane-spanning domain, and therefore are not expected to contribute any membrane-bound ferroxidase function at all. Hephaestin-like 1's role in iron nutrition and homeostasis is not well understood, but our assignment of two, allelic, curly-whiskers mouse mutants to Hephl1 should help to set-the-stage for the needed molecular and whole-animal studies.

Segregation data shown in Fig. S1 and in Fig. 2 suggest that the Hephl1^cw allele does not negatively impact general viability in homozygotes, at least by weaning age (when our testcross offspring were counted). This may suggest that Hephl1 is not an essential protein, or instead may indicate that the Hephl1^cw isoform translated from transcripts that skip Exon 11—despite lacking 71 amino acids in Domain 4—might be normally located in the cell membrane and might retain some “leaky” ferroxidase function. While we have collected only limited segregation data from Hephl1^cw-2J testcrosses, to date, our current tally does show a small deficit of mutants compared to wild type (perhaps indicating that Hephl1^cw-2J may control a more severe phenotype than Hephl1^cw), but these counts are still modest (N = 106) and the deviation from the 1 wild:1 mutant testcross ratio expected for full viability is not statistically significant (P > .12). Of course, any predictions regarding Hephl1^cw and Hephl1^cw-2J protein structure, location, and function need to be explicitly tested, ideally, side-by-side with a bone fide null mutant, as are readily available from multiple sources (see [36,31,25], for example).

4.3. Hephl1 impairment causes pili torti

Our selection of Hephl1 as a likely positional candidate for cw was largely due to the association of kinky, brittle hair (pili torti) with generalized copper deficiency, as in Menkes disease, an X-linked recessive disorder (resulting from defects in the copper-transport protein ATP7A) whose myriad clinical features are thought to result from the dysfunction of several copper-dependent enzymes (OMIM, entry #309400, [27]). While several cuproenzymes have been proposed to account for the various features of the disorder (tyrosinase for depigmentation of hair and skin, lysyl oxidase for connective tissue defects, cytochrome c oxidase for hypothermia, and ascorbate oxidase for skeletal demineralization, for example; reviewed by [23]), this report is the first to implicate hephaestin-like 1 in, at least, the kinky-hair feature of Menkes disease and the related occipital horn syndrome (OMIM, entry #304150, [26]). Other aspects of these disease syndromes may also be owing to reduced Hephl1 activity, since Hephl1 is expressed in some of the same tissues impacted by these conditions (including the retina [17,19], synovial membranes [1], and in the cardiovascular, connective, urinary and nervous systems [5]), but these tissues have not yet been examined in any detail in the curly-whiskers mouse mutants. In any case, the animal models we have now causally ascribed to Hephl1-deficiency should facilitate the functional dissection of these complex, pleiotropic human disorders, and hopefully advance the development of treatments for at least some aspects of the syndromes they generate.

4.4. Do these curly-whiskers mutants generate skin-graft histoincompatibility?

We remain especially curious to learn whether Hephl1^cw and/or Hephl1^cw-2J, like the now-extinct cw^thd allele, might also generate an immunological “gain” or “loss” of antigenicity that can be detected by skin-graft-exchange assay, for example. Classically, mutations that create such heritable barriers to graft compatibility have identified the so-called minor histocompatibility (H) loci [28], which currently number in excess of 60. However, mutations that affect both transplantation acceptance and a distinct developmental function are quite rare. We know of only one other example, the recessive H-mshi mutation (for “male sterility and histoincompatibility”, [37]) which causes aspermia in homozygous males, but was initially discovered as the loss of a cell-surface antigen, such that tail skin from +/+ mice is rejected by (otherwise genetically-matched) mshi/mshi mutants, which recognize the mshi⁺ antigen as “foreign” [13,22,30,34].

This sort of phenotypic complexity interests us for several reasons. First, we anticipate that the antigenicity of the gene product might easily be exploited to provide another molecular “handle” for gaining access to the molecular physiology controlled by Hephl1. Second, mutations that control seemingly unrelated functions (histocompatibility and sperm production or hair development, for example), offer an opportunity to dissect the cause of such compelling pleiotropy, a phenomenon of enduring interest in formal genetics. Third, it seems possible (even likely) that the study of antigen-altering mutations could reveal much about human disease phenotypes that might involve auto-immune reactions against the orthologous protein products (see [16]). Finally, we believe it is still important to identify and characterize new minor H loci, so that their individual and collective significance in graft transplantation can be formally assessed.

To prepare to conduct such histogenic assessments, we have been developing genetically “pure” mouse stocks that segregate for Hephl1⁺ and each recessive curly-whiskers mutation. These inbred strains should allow the production of homozygous mutant, heterozygous, and homozygous wild-type subjects on the otherwise uniform genetic background that is required for single-gene, allograft analysis. To develop a line that segregates for Hephl1⁺ and Hephl1^cw, we have crossed CWD/LeJ homozygotes, Hephl1^cw Myo5a^d/Hephl1^cw Myo5a^d, with strain DBA/2J (the original source of the Myo5a^d in the CWD/LeJ strain). The resulting Hephl1⁺ Myo5a^d/ Hephl1^cw Myo5a^d F₁ offspring (with straight vibrissae) were then crossed back to CWD/LeJ mutants to produce Hephl1^cw/Hephl1⁺ heterozygotes that were crossed back to CWD/LeJ, and so on. This segregating congenic line, named CWD.D2-Hephl1⁺/Hephl1^cw, is currently at N₁₃. To develop a stock segregating for Hephl1⁺ and Hephl1^cw-2J, we crossed B10.SM H2^v H2-T18^b/(70NS)Sn/J-Hephl1^cw-2J mutant males to C57BL/10SnJ females, and the F₁ offspring were crossed back to C57BL/10SnJ. Heterozygous offspring from this cross, identified by the MlyI-sensitivity test we described, were crossed back to C57BL/10SnJ, and so on, to create a segregating inbred strain named C57BL/10SnJ-Hephl1⁺/Hephl1^cw-2J (currently at N₅) that will be, essentially, co-isogenic. While primarily designed for the immunogenetic analyses that we anticipate performing, these uniform, inbred stocks should also be ideal for making other well-controlled molecular and functional comparisons among the various Hephl1 genotypes.

4.5. Positional cloning: the end of an era

In recent years, modern methods (like whole-genome DNA sequencing) have supplanted positional cloning (sometimes called “reverse genetics”) as an approach for making causative-gene assignments for spontaneous mutations, in both mice and man [10,18,35]. And even if this powerful new method works best when the natural variant under study has had some prior positional characterization, genetic mapping as a means to gain initial access to gene structure and function appears (perhaps with this report) to have become outdated. Indeed, the current ease with which genetic variants can be deliberately and precisely engineered (by gene targeting or gene editing, for example), infers that even spontaneous mutations—once the only source genetic variation—are no longer particularly prized. We would argue, instead, that natural mouse variants (like curly whiskers) often contribute a critically informative part of an allelic series, where they can frequently offer a unique perspective into the pathobiology of inherited human disorders, which similarly result from leaky, pleiotropic and often surprising spontaneous mutations.

The following are the supplementary data related to this article.

Supplementary Fig. S1

Segregation of alleles of cw, Myo5a (dilute), and four dimorphic microsatellite markers among 168 progeny from an intraspecific mouse backcross. Heterozygous F₁ mice (CWD/LeJ x C57BL/6 J) were backcrossed to homozygous CWD/LeJ mutants. The resulting progeny were scored for their fur texture and colour, and a DNA sample from each mouse was typed for the microsatellite markers shown at the left. Only the Chr 9 haplotype inherited from the F₁ parent is shown (with a knob at the top of the haplotype representing the centromere), and the number of mice inheriting that haplotype is shown below it. Very similar numbers of wild and mutant progeny—as expected for a testcross (χ² = 0.024, P > .87)—suggest that the mutant phenotype is both fully penetrant and fully viable. The 11 wild type recombinants and the 6 mutant recombinants marked with an asterisk show that the cw locus must lie centromeric to D9Mit64. Genetic distances in percentage recombination are shown to the right (± 1 Standard Error).

Supplementary Fig. S2

Exon/exon junctions for the Hephl1 PCR products that were amplified from cDNA pools based on wild type or cw/cw mutant skin. PCR products shown in Fig. 4B were isolated and subjected to primer-extension sequence analysis. (A) The 606 bp amplimer based on wild type cDNA demonstrated normal splicing of Exons 9 through 12 (only the Exon 10/11 and Exon 11/12 junctions are shown). (B) The 581 bp product, designated Splice Variant 1, includes the unusual junction shown, with Exon 10 joined to Exon 11(Δ25), which is 25 nucleotides shorter than the standard Exon 11. The cw-specific Exon 11(Δ25) seems to be defined by a cryptic splice-acceptor site composed of the polypyrimidine tract at the 3′ end of Intron 10–11 and the first 5′-AG-3′ dinucleotide within Exon 11 (shown underlined in blue in Fig. 4A). The Variant 1 transcript is predicted to disrupt the normal reading frame, introducing 3 novel amino acids (numbered 623–625, shown in red) followed by an out-of-frame stop codon. It is likely that this variant transcript, which cannot be fully translated, will be unstable due to nonsense-mediated mRNA decay [4]. (C) The 393 bp product, designated Splice Variant 2, is the result of Exon 10 joining with Exon 12. The skipping of Exon 11, which measures 213 nucleotides, generates a variant transcript that can be fully translated, and is predicted to yield a mutant protein product that lacks 71 (of the usual 1159) amino acids.

Supplementary Fig. S3

A splice-acceptor defect located 5′ of Exon 11 in the Hephl1 gene is specific to the cw mutation in mice. (A) The rationale of an AluI-sensitivity test for rapid detection of mutation at the splice acceptor site 5′ of Exon 11 in the Hephl1 gene is shown. The sequence displayed here includes nucleotides 9:15077024 to 9:15076822, from reference sequence GRCm38.p6 [9]. Lower-case sequences are from Intron 10–11; sequences in capital letters are from Exon 11. An AluI, restriction endonuclease target sequence is highlighted in yellow, and a red arrow shows the position of the cleavage site. The primers used to direct amplification of the 203 bp sequence shown are underlined. AluI digestion will yield two fragments of 78 + 125 = 203 bp for amplimers based on wild-type templates. Amplimers that carry any disruption of the AluI target sequence (for example, the A to G transition we have found in cw mutants) will be resistant to AluI digestion. (B) Typical results of AluI-sensitivity testing among various inbred mouse strains. PCR-generated amplimers based on genomic DNA templates were incubated with AluI, and electrophoresed through 3.5% NuSieve agarose. The only amplimers fully resistant to cutting were based on CWD/LeJ and STOCK cw-Bmp5^se-tk/H (marked with a red dagger) templates, strains known to be homozygous for cw. Three known cw carriers (backcross generations N₂, N₈, and N₁₀) show both AluI-resistant and sensitive digestion products. The lane marked with a red asterisk contains an AluI-sensitive amplimer copied from the B10.SM H2^vH2-T18^b/(70NS)Sn/J strain (which has phenotypically curly whiskers) indicating a normal DNA sequence at this site. (C) DNA sequence of B10.SM H2^vH2-T18^b/(70NS)Sn/J and STOCK cw-Bmp5^se-tk/H genomic DNA near the splice acceptor site 5′ to Exon 11 in the mouse Hephl1 gene. A black arrow indicates the position which has mutated to a G in the CWD/LeJ-cw/cw strain (position 9:15076948, from reference sequence GRCm38.p6; EMGS [9]). Consistent with the results of AluI-sensitivity testing, STOCK cw-Bmp5^se-tk/H shows the same defect as CWD/LeJ, whereas B10.SM H2^vH2-T18^b/(70NS)Sn/J strain has no defect at this position.

Supplementary Fig. S4

Insertion of a single A residue in Exon 14 of the Hephl1 gene is specific to the cw^2J mutation in mice. (A) The rationale of an MlyI-sensitivity test for rapid detection of mutation at codon number 820 of the Hephl1 gene is shown. The sequence displayed here includes nucleotides 9:15067278 to 9:15067007, from reference sequence GRCm38.p6 (EMGS [9]). Sequences in capital letters are Exon 14; lower-case sequences are from the flanking introns. Two MlyI restriction endonuclease target sequences are highlighted in yellow, and red arrows show the positions of the respective cleavage sites. The primers used to direct amplification of the 272–273 bp sequence shown are underlined. MlyI digestion will yield three fragments of 69 + 81 + 122 = 272 bp for amplimers based on the wild type sequence. Amplimers that encode a disruption of the 5′ MlyI target sequence (for example, the A insertion we have found in cw^2J mutants) will cut only at the 3′ MlyI target site, yielding an 81 bp and a longer fragment (e.g., 192 bp for a single-base-pair insertion). (B) Typical results of MlyI-sensitivity testing among various inbred mouse strains. PCR-generated amplimers based on genomic DNA templates were incubated with MlyI, and electrophoresed through 3.5% NuSieve® agarose. Only amplimers based on cw^2J/cw^2J or +/ cw^2J templates showed the MlyI-resistant 192 bp band.

Supplementary Table S1

Description of SNP markers referred to in the Eragene et al. (2019) text.

Supplementary Table S2

Location of SNP markers referred to in the Eragene et al. (2019) text.

Competing interests

The authors declare no competing interests.