Three Novel Pigmentation Mutants Generated by Genome-Wide Random ENU Mutagenesis in the Mouse

Three mutant mice with pigmentation phenotypes were recovered from a genomewide random mouse chemical mutagenesis study. White toes (Whto; MGI:1861986), Belly spot and white toes (Bswt; MGI:2152776) and Dark footpads 2 (Dfp2; MGI:1861991) were identified following visual inspection of progeny from a male exposed to the point mutagen ethylnitrosourea (ENU). In order to rapidly localize the causative mutations, genome-wide linkage scans were performed on pooled DNA samples from backcross animals for each mutant line. Whto was mapped to proximal mouse chromosome (Mmu) 7 between Cen (the centromere) and D7Mit112 (8.0 cM from the centromere), Bswt was mapped to centric Mmul between D1Mit214 (32.1 cM) and D1Mit480 (32.8 cM) and Dfp2 was mapped to proximalMmu4 between Cen and D4Mit18 (5.2 cM). Whto, Bswt and Dfp2 may provide novel starting points in furthering the elucidation of genetic and biochemical pathways relevant to pigmentation and associated biological processes.


Introduction
Pigment arises from the melanosome, an organelle found in melanocytes that fills with melanin. The melanin is subsequently transferred from the melanocyte to surrounding keratinocytes, resulting in pigmentation (see Jackson, 1997). Defects in pigmentation can arise at the site of melanin production as a consequence of abnormal melanosomes, or as a result of developmental failure to deliver melanocytes at the required sites for correct pigmentation to occur. Melanocytes arise from the neural crest; during embryonic development, as the neural tube folds and closes, neural crest cells migrate from the dorsal neural tube to disparate regions of the embryo. Neural crest cells are precursors to a number of different cell types, including the neuronal and glial cells of the peripheral nervous system as well as skeletal and connective tissue components of the head, in addition to melanocytes (Nicholls, 1992). Consequently, mutations in genes affecting pigmentation can cause defects in other cell lineages if such cells share a common precursor with melanocytes.
To date, mouse genetic studies have implicated more than 40 proteins in the biochemical and developmental processes involved in pigmentation, including G-protein coupled receptors and their ligands, receptor tyrosine kinases and their ligands, melanogenic enzymes and transcription factors, to name a few (see Jackson, 1997;Nakamura et al., 2002). Although analysis of mutant mouse phenotypes has made an enormous contribution toward elucidating biochemical and developmental pathways involved in pigmentation, and more broadly in characterizing mammalian gene function in general, there is a clear disparity between the number of mouse phenotypes available and the number of known genes. If the entire spectrum of available mouse mutants for all observed phenotypes is considered, presently it is likely that mutant phenotypes exist for less than 10% of the predicted number of mouse genes (Brown and Hardisty, 2003). In order to fully exploit the mouse as a model organism for genetic research, there is a need to discover new phenotypes that result from mutations at new loci and uncover multiple mutants for the same locus. Several mutagenesis programmes have been established to increase the mutant mouse resource utilizing the mutagen ENU (Balling, 2001;Brown and Balling, 2001). ENU is a potent producer of point mutations; male mice exposed to ENU by intraperitoneal injection can acquire germline base changes at a level of >5 × 10 −4 /kb (Coghill et al., 2002). Because ENU mutagenesis acts in a random, genome-wide fashion, it is necessary to locate mutated genes by linkage analysis and positional cloning following recovery of mutant progeny from ENU-exposed males. The MRC/GSK mouse mutagenesis programme, initiated by the UK Medical Research Council and GlaxoSmithKline Pharmaceuticals (Nolan et al., 2000), has thus far generated over 160 dominant phenotypes by ENU mutagenesis and the chromosomal localization of over 70 of these phenotypes has been elucidated (http://www.mgu.har.mrc.ac.uk/mutabase/).
In this paper we describe three novel mutant mice with dominant coat colour phenotypes that were identified from the MRC/GSK ENU programme, White toes (Whto), Belly spot and white toes (Bswt) and Dark footpads 2 (Dfp2), and report their genetic map locations as determined by backcross linkage analysis.

Animals
Animal studies described here were carried out under the guidance issued by the Medical Research Council in Responsibility in the Use of Animals for Medical Research (July 1993) and Home Office Project Licence No. 30/1517. Details of the mutagenesis programme are described elsewhere (Nolan et al., 2000). Briefly, BALB/c males (Charles River, UK) were injected intraperitoneally with two weekly doses of 100 mg/kg ENU (Sigma) at approximately 10 weeks of age. Visual examination was carried out on F 1 progeny of mutagenized male mice crossed to C3H/HeH females (Charles River, UK) and at 5 weeks of age, the mice were assessed using the SHIRPA behavioural and functional assessment protocol (Rogers et al., 1997). For inheritance testing, F 1 mice were backcrossed to the C3H/HeH strain and progeny classified for the phenotype identified in the founder. Mutant lines were subsequently maintained by backcrossing to C3H/HeH. Three novel mouse pigmentation mutants 125 (Isaacs et al., 2000). Briefly, tail snips were taken from mutant backcross mice and DNA extracted using standard procedures. DNA concentrations were measured by UV absorbance and equimolar aliquots of DNA combined for each strain. Non-mutant backcross pools were also constructed for Dfp2 ; for Bswt and Whto mutant pools only were used. The DNA pools, together with DNA from BALB/c, C3H/HeH and F 1 mice, were screened with 93 simple tandem repeat markers polymorphic between BALB/c and C3H/HeH, spaced at 20 cM (centimorgan) or less, selected from the Whitehead/MIT database (wwwgenome.wi.mit.edu). Genotyping was performed using an Applied Biosystems 377 PRISM system in accordance with the manufacturer's instructions. Genotype data were analysed using a modified version of TrueAllele software (Cybergenetics) and linkage assessed by relative BALB/c;C3H/HeH allele signal strength. Preliminary linkage was confirmed by genotyping individual animals. Crude map locations were refined by genotyping with additional markers.

Results
Three pigmentation mutants were identified in the progeny of BALB/c male mice exposed to ENU and crossed to C3H/HeH females; autosomal dominant inheritance of each phenotype was confirmed by backcross analysis. The resulting mutant lines were named 'White toes', 'Belly spot and white toes' and 'Dark footpads 2' (Table 1). White toes (Whto, MGI:1 861 986) is characterized by a white belly spot and white hind toes, with hydrocephaly in some mice and small litter size. Belly spot and white toes (Bswt, MGI:2 152 776) has white feet and a white belly spot. Dark footpads 2 (Dfp2, MGI:1 861 991) is similar to the previously known dark footpads mutant (Dfp; Kelly, 1968) and, as the name suggests, has markedly increased pigmentation in the footpads. At 5 weeks of age, Bswt and Dfp2 mice were assessed using the SHIRPA behavioural and functional assessment protocol (Rogers et al., 1997). This screen, involving a battery of 40 simple tests, is semiquantitative and based on modifications of earlier screens described by Irwin (1968). No behavioural or functional anomalies were detected in either mutant line. The results of the SHIRPA tests and a full description of the protocol can be viewed at http://www.mgu.har.mrc.ac.uk/mutabase/ In order to rapidly identify regions of linkage, genome-wide scans using 93 simple tandem repeat markers with spacing of 20 cM or less, selected from the Whitehead/MIT database (http://www-genome.wi.mit.edu/: marker details available on request), were performed on pools of equimolar DNA samples from mutant (BALB/c × C3H/HeH) × C3H/HeH backcross animals for Whto (n = 44) and Bswt (n = 81), as described in Isaacs et al. (2000). Both mutant and non-mutant DNA pools were screened for Dfp2 (n = 31 and 20, respectively). Mutant pools only were used for Whto and Bswt because, unlike Dfp2, these lines showed evidence for incomplete penetrance in the ratio of mutant vs. non-mutant backcross animals recovered. From these initial genome scans, Bswt was mapped to centric (Mus musculus; Mmu) chromosome 1, Dfp2 was mapped to proximal Mmu 4 and Whto was mapped to proximal Mmu 7. These crude map positions for Whto and Dfp2 were originally reported in Nolan et al. (2000); Bswt has not been reported previously.
The map positions were confirmed by genotyping individual mice and more precise locations identified by genotyping with additional markers in the critical interval for each mutant. Map positions were refined to D1Mit214 (32.1 cM)-D1Mit480 (32.8 cM) for Bswt, Cen (0.0 cM)-D4Mit18 (5.2 cM) for Dfp2 and Cen (0.0 cM)-D7Mit112 (8.0  (Table 1). Marker positions are taken from the Jackson mouse genetic map (http://www. informatics.jax.org).

Discussion
No previously mapped pigmentation mutants listed in the mouse genome database (www.informatics. jax.org) lie in the genetic intervals bearing Whto or Bswt, other than the reduced pigmentation mutant rp, which maps within the Whto region (Gibb et al., 1981). The rp mutant shows generalized reduction in pigmentation and is recessive, whereas the dominant Whto has a distinct white belly spot and white hind toes; therefore it is unlikely that Whto and rp are allelic. The gene responsible for the rp phenotype has yet to be identified. Although Whto and Bswt appear to be unique, Fitch et al. (2003) recently described Dsk4, a pigmentation mutant strikingly similar to Dfp2. Dsk4 maps to an interval on proximal chromosome 4 that is coincident with Dfp2, strongly suggesting that Dfp2 and Dsk4 are allelic. Combining data from Dfp2 and Dsk4 should therefore expedite mapping and gene identification for this mutant. The map location of the original Dfp mutant (Kelly, 1968) is unknown; therefore it is possible that Dfp and Dfp2/Dsk4 are allelic. However, Fitch et al. (2003) also describe several other dark footpad mutants and it is equally likely that Dfp is an allele at one of these alternative loci. Although no comparable pigmentation mutants map to the same genetic regions as Whto or Bswt, mutants similar to these, either mapping to other regions or as yet unmapped, have been described previously. The splotch (Sp) mutant (Russell, 1947) shares many of the characteristics of the Whto and Bswt phenotypes. Mice that are heterozygous for Sp have a white patch on their abdomen; mouse embryos homozygous for Sp show defects in neural tube closure and neural crest cell development, leading to reduction or absence of a number of neural crest cell derivatives and resulting in death at embryonic day 16 (Franz, 1989;Serbedzija and McMahon, 1997). The Sp phenotype is caused by mutations in the Pax3 gene, located on Mmul (Epstein, 1991). Although Bswt is also linked to Mmul, Pax3 at 44 cM from the centromere according to the Jackson map (www.informatics.jax.org), lies outside the critical interval (Bswt is between 32.1 and 32.8 cM on the Jackson map). Bswt does not therefore appear to be allelic to Sp. Bst (belly spot and tail; Southard and Eicher, 1977) is a second dominant mutant with similarities to Bswt and Whto. Heterozygotes have ventral spotting, a short kinked tail and occasionally white feet, reduced body size, malocclusion, anomalies of the spine and eyes, and polydactyly. Bst maps to Mmul6 (Rice et al., 1995) and therefore is not allelic with either Bswt or Whto.
Significant deviation from the expected ratio of 1 mutant:1 non-mutant for a fully penetrant dominant trait was observed for Bswt and Whto (81 : 115 and 44 : 65, respectively) giving an estimated penetrance of approximately 80% for both lines at the first backcross generation. This reduced penetrance may result from inter-animal variation in genetic background as a consequence of heterogeneity with respect to the BALB/c and C3H/HeH contribution to the genome of each mouse; certain genetic background combinations may mask the effect of the mutant allele. However, it is also possible that increased mortality of mutant mice in utero may be the cause of the skew in favour of non-mutant animals. An alternative explanation for an apparent reduction in penetrance is that two or more ENU-generated point mutations are required for expression of the phenotype and that segregation of these mutations occurs in the backcross; however, in this instance such an explanation is unlikely, as the expected outcome would be an perceived drop in penetrance to 50% or less.
Given the phenotypic similarities between Bst, Bswt, Sp and Whto, it is possible that they operate on the same biochemical or developmental pathway. As described in the introduction, pigmented cells originate in the neural crest; much is known about the migratory pathways taken by neural crest cells and their derivatives, but our knowledge of the genes that control these processes, although increasing rapidly, is still limited (Maschhoff et al., 2000;Wu et al., 2003). Given that Bswt and Whto are phenotypically similar to Sp and show a heterogeneous distribution of pigmentation (white spotting), it is likely that these mutants result from defects of melanocyte development or migration, rather than melanosome dysfunction, and may be useful in the elucidation of genetic and biochemical pathways relevant to Three novel mouse pigmentation mutants 127 developmental processes involving the neural crest. As yet, Bswt and Whto homozygote mutants have not been generated; this is an important next step and may reveal developmental defects and provide further clues to Bswt and Whto gene function.
Although the genetic intervals harbouring Whto and Dfp2, at approximately 8.0 cM and 5.2 cM, respectively, are too large to speculate usefully on potential candidate genes in these regions; the Bswt region at less than 1 cM is more tractable. 19 genes are listed in the interval bounded by the Bswt flanking markers D1Mit214 and D1Mit480 in the February 2002 freeze of the University of California Santa Cruz (UCSC) mouse genome sequence browser (http://genome.ucsc.edu/). Approximately 40 genes are listed in the equivalent human genome interval (November 2002 freeze), which has more complete coverage. Of these, the most intriguing candidate for the Bswt phenotype is the Frizzled-7 gene (Fzd7 ). Frizzled proteins are cell surface receptors for Wnt ligands; activation of the Wnt-Frizzled signalling pathway ultimately results in Wnt-dependent gene expression. Experiments in Xenopus have established a role for Wnt-Frizzled signalling in neural crest formation, and melanocyte induction by Frizzled-3 has been demonstrated (see Wu et al., 2003). Complete ablation of Fzd7 in Xenopus results in severe gastrulation defects (Winklbauer et al., 2001), but the effects of partial or complete loss of Fzd7 function in mammalian systems is unknown. We are currently screening Fzd7 for point mutations in Bswt mice.