A High-Resolution Genetic, Physical, and Comparative Gene Map of the Doublefoot (Dbf) Region of...

Articledoi:10.1006/geno.2001.6657, available online at http://www.idealibrary.com on IDEAL

INTRODUCTION

A High-Resolution Genetic, Physical, and Comparative GeneMap of the Doublefoot (Dbf) Region of Mouse Chromosome 1 and the Region of Conserved

Synteny on Human Chromosome 2q35Christopher Hayes,1,*,§,# Andreas Rump,2,†,§ Matthew R. Cadman,1 Mark Harrison,1

Edward P. Evans,1,‡ Mary F. Lyon,1 Gillian M. Morriss-Kay,3André Rosenthal,2,† and Steve D. M. Brown1

1Medical Research Council, Mammalian Genetics Unit and UK Mouse Genome Centre, Harwell, Didcot, Oxon., OX11 0RD, UK2Genome Sequencing Centre, Institute of Molecular Biotechnology, Beutenbergstrasse 11, 07745 Jena, Germany

3Department of Human Anatomy and Genetics, University of Oxford, South Parks Road, Oxford, OX1 3QX, UK*Present address: Department of Molecular Biology, Neuroscience Research Centre, Merck, Sharp and Dohme,

Terlings Park, Eastwick Road, Harlow, Essex, CM20 2QR, UK†Present address: MetaGen Pharmaceuticals, Ihnestrasse 63, 14195 Berlin, Germany

‡Present address: Department of Zoology, University of Oxford, South Parks Road, Oxford, OX1 3PS, UK

§These authors contributed equally to this work.

#To whom correspondence and reprint requests should be addressed. Fax: +44 (0) 1279 440390. E-mail: [email protected].

The mouse doublefoot (Dbf) mutant exhibits preaxial polydactyly in association with cran-iofacial defects. This mutation has previously been mapped to mouse chromosome 1. We haveused a positional cloning strategy, coupled with a comparative sequencing approach usingavailable human draft sequence, to identify putative candidates for the Dbf gene in the mouseand in homologous human region. We have constructed a high-resolution genetic map of theregion, localizing the mutation to a 0.4-cM (± 0.0061) interval on mouse chromosome 1.Furthermore, we have constructed contiguous BAC/PAC clone maps across the mouse andhuman Dbf region. Using existing markers and additional sequence tagged sites, which wehave generated, we have anchored the physical map to the genetic map. Through the com-parative sequencing of these clones we have identified 35 genes within this interval, indi-cating that the region is gene-rich. From this we have identified several genes that are knownto be differentially expressed in the developing mid-gestation mouse embryo, some in thedeveloping embryonic limb buds. These genes include those encoding known developmen-tal signaling molecules such as WNT proteins and IHH, and we provide evidence that thesegenes are candidates for the Dbf mutation.

Key words: Doublefoot, genetic mapping, physical mapping, gene identification, limb development, comparative sequencing

zeugopodal elements: the tibia is grossly malformed, causing

The Doublefoot (Dbf) mutation arose spontaneously in the3H1 (C3H/HeH 3 101/H F1 hybrid) stock at MRC Harwell,UK. The mutation is inherited as a single, mendelian, auto-somal dominant character, and affected animals have extradigits on all four limbs [1]. The extra digits all arise from thepreaxial (anterior) aspect of the limbs and are all triphalangeal[2]. The limbs also show skeletal abnormalities within the

GENOMICS Vol. 78, Number 3, December 2001Copyright © 2001 by Academic Press. All rights of reproduction in any form rese0888-7543/01 $35.00

luxation of the distal limb skeleton, manifest as a talipesequinovarus deformity of the limb. The Dbf/+ mice alsoexhibit craniofacial abnormalities, the skull being broadenedand bulbous due to an increase in size of the calvarial sutures.Homozygotes, which are not recoverable alive beyond 14.5days post coitum (dpc), also exhibit a midline facial cleft [3].

Molecular analysis of mutant animals has revealed severalother aspects of the Dbf phenotype. Genetic and functionalanalyses of the mutant limb buds revealed that they possessed

197rved.

Article doi:10.1006/geno.2001.6657, available online at http://www.idealibrary.com on IDEAL

i

Dbf/+
ectopic polarizing activityand the mutation upregu-lated the transcriptional tar-gets of the Hedgehog path-way. These data indicatedthat the mutation might rep-resent a novel element of themammalian Hedgehog path-way [3]. A further demon-stration that the mutationcauses the aberrant upregu-lation of the Indian hedgehoggene (Ihh), which residesclose to the Dbf mutation onmouse chromosome 1, raisedthe possibility of the involve-ment of Ihh in the pathogen-esis of the Dbf phenotype. In the same report, however, the recomb Dbf mutation recombined with the Ihh locus, being mappedgenetically 1.3 cM distal to Ihh on mouse chromosome 1 [4].
As a first step towards the identification of the Dbf gene,we have created a high-resolution genetic map of the regionbased on the analysis of 989 meioses, narrowing the criticalDbf-containing region to a 0.4-cM (± 0.0061) interval.Furthermore, we have constructed a bacterial clone contigacross the mouse and homologous human Dbf region.Analysis of the human draft sequence, as well as samplesequencing clones from the mouse and human criticalregions, has identified 35 putative genes, 3 of which areexpressed in the developing embryonic limbs and face, mak-ing them candidate genes for the Dbf mutation. Two of thesegenes are members of the Wnt gene family encoding devel-opmentally regulated signaling molecules, Wnt6 andWnt10a, and the third is Ihh. This report details the effortsmade in construction of the high-resolution genetic andphysical map in the vicinity of the Dbf mutation in order toidentify candidate genes using approaches involving com-parative sequencing to aid in gene identification.

RESULTS

Karyotypic Analysis of the Dbf RegionTo assess the possibility of gross chromosomal rearrange-ments, we karyotyped Dbf/+ (3H1 3 Mus musculus castaneus)F1 hybrids. The Dbf (3H1) chromosome was distinguishedfrom the wild-type (M. m. castaneus) chromosome by virtueof a C-band polymorphism. We observed no karyotypic dif-ferences associated with the Dbf mutation in any of the chro-mosomal preparations analyzed (Fig. 1).

Generation of a High-Resolution Genetic MapA small, interspecific backcross segregating the Dbf mutationallowed it to be mapped to a relatively small region of mousechromosome 1 [1]. To increase the resolution of the mapping ofthe Dbf locus, an intersubspecific backcross to M. m. castaneus

Copyright © 2001 b198

was established, increasing the number of backcrossprogeny to 989 animals. Concomitantly, we testedadditional microsatellite markers for linkage to theDbf locus to refine the genetic map. One microsatel-lite marker, D1Mit24, mapped close to the mutationon the proximal (centromeric) side of the chromo-some, and was separated from the mutation by tworecombinants, whereas D1Mit46 was placed tworecombination events distal (telomeric) to the muta-tion. These markers represented the closest flanking

nant markers with the mutation, placing the Dbf muta-

FIG. 1. Karyotype of chromosome 1 of a 3H1 /Mus musculus cas-taneus F1 hybrid. The 3H1 (Dbf) and M. m. castanteus (wild type)chromosomes can be distinguished by a C-band polymorphism.No chromosomal alterations were detected to be associated withthe Dbf region of chromosome 1.

tion into a 0.4-cM (± 0.0061) interval, which equates to approx-imately 800 kb on the physical map. A further microsatellitemarker, D1Mit7, was found to be non-recombinant with themutation in all of the 989 meioses analyzed, giving the fol-lowing marker order: Cen–D1Mit24–Dbf,D1Mit7–D1Mit46–Tel (Fig. 2). This marker order does not differ sig-nificantly in either order or distance from those in theEuropean Interspecific Backcross (EUCIB; [5]). Notably, withmarkers that were polymorphic between C3H and 101/H, onlythe C3H allele was detected in haplotypes around the Dbf locus,indicating that the original mutation had occurred on the ances-tral C3H chromosome of the founder 3H1 individual, whichwas an F1 hybrid between C3H/HeH and 101/H. Suppressionof recombination was not observed in association with the Dbfmutation, and the genetic distances and order of markers werein general agreement with published sources. These data,together with the cytogenetic analysis, suggest that the Dbfmutation is not due to a chromosomal rearrangement of mousechromosome 1.

Construction of a Bacterial Clone Contig across the Dbf RegionAs a first step towards the identification of the Dbf gene, webegan the construction of a bacterial clone contig across theregion. Clone contigs were generated across the mouse andsyntenic human regions (Fig. 3). Sample sequencing of thesecontigs in association with the analysis of the burgeoningdraft human sequence data has lead to the rapid identifica-tion of candidate genes for the mutation.

To aid in this process of gene identification within the Dbfregion, and also to make the best use of already establishedresources such as human genomic sequence, we undertook acomparative approach to interrogate the data to best effect.Using the genic sequence tagged sites (STSs) described above,the “End-sequence” database and the high-throughputgenomic sequence database for human were searched(http://www.ncbi.nlm.nih.gov/blast/blast.cgi) to identifyclones that contained a portion of the STS used as probe.

GENOMICS Vol. 78, Number 3, December 2001y Academic Press. All rights of reproduction in any form reserved.


A

Identification and Analysis of Transcription Units from
the Dbf Critical RegionSeveral distinct transcription units were identified from theDbf critical region through a variety of approaches. Theseapproaches included end-sequencing clones, the develop-ment of STSs identified from yeast artificial chromosomes(YACs) in the region based on database searches, positionalcandidate approaches, and sample sequencing of the contig.This approach was substantially complemented through theanalysis of draft sequence of the human genome and end-sequence data for human clones, as well as the comparativeanalysis of sequence from the mouse and human regions, asa means to aid in gene identification [6]. The major criteria fordetermining whether or not the genes were transcribed weredetermining if the genes were represented in EST databasesand indicating that they were indeed transcribed. If this wasnegative, the genes were subjected to RT-PCR analysis acrossa range of embryonic and adult tissues to determine if theywere transcribed, as was the case for GS1876. Those putativepredicted genes with no EST hits or expression by RT-PCR inadult or embryonic RNA populations are excluded from thisreport. A summary of the genes identified is given in Table1, and a brief description of each follows.
The mouse homologue of human ARPC2 (ARP2/3 proteincomplex subunit; also known as P34) has not previously beendescribed, but ESTs corresponding to this gene were identi-fied in homology searches from a wide range of mouse tis-sues and from a wide range of taxa, from zebrafish to human.The encoded protein has a role in actin filament assembly [7].

This gene AAMP (angio-associated cell migratory protein)has been implicated in the formation of endothelial tubes [8],but its expression pattern has not been investigated in mouse.A search of the Mouse Genome Informatics database(http://www.informatics.jax.org/) revealed that a geneknown as Aamp-rs (angio-associated cell migratory protein,

GENOMICS Vol. 78, Number 3, December 2001Copyright © 2001 by Academic Press. All rig

FIG. 2. Genetic mapping of the Dbf mutation. (A) Genetic map across the Dbf region of mouse chromosome 1,based on the analysis of 989 backcross progeny. Distances shown are in centimorgans (cM). (B) Haplotype analy-sis across the Dbf region. Open boxes represent the 3H1 (Dbf) allele and filled boxes, the inter(sub)specific allele.Uninformative haplotypes are not shown.

B

hts of reproduction in any form res

related sequence) was genetically mapped to mouse chro-mosome 1. It is worth noting that AAMP is present on thedraft sequence from 2q35. Therefore, we suggest that the genewe have identified in the Dbf contig is the mouse ortholog ofAAMP, which is expressed in the developing vasculature ofthe mammalian embryo (Fig. 4A).

The KIAA1184 sequence displayed significant similarity tothe hydroxyacylglutathione hydrolase gene (glyoxalase;HAGH). This gene seems to have a role in “housekeeping”metabolic processes within the cell and its expression inembryos was not characterized.

The gene Slc11a1 (formerly known as Nramp1) was alsoidentified. This locus has a role in regulating the susceptibil-ity to systemic infection by microbiological agents.

The villin gene (Vil) has been shown to play an importantpart in the function of epithelia, in particular brush borderepithelia [9]. Disruption of the Vil locus results in perturba-tion of the epithelia of the intestine [10].

The KIAA1154 sequence initially showed homology to abovine testis-specific transcript (GenBank acc. no. Z86040).The predicted translation product of this gene suggested thatit showed similarity to ubiquitin carboxy-terminal hydrolaseenzymes, and possibly represents another member of thisfamily of proteins. It was not expressed in the developing10.5-dpc embryo (data not shown).

The gene Rqcd1 (required for cell differentiation,Schizosaccharomyces pombe, homologue 1) is conserved through-out eukaryotes, suggesting a conserved role in differentiationmechanisms [11].

Plcd4 (phospholipase C, d4 subunit) is important in thegeneration of biochemical second messenger systems.

In silico analysis of the ZNF142 clone revealed that itsencoded protein had strong similarity to zinc-finger proteins,with 36 zinc-finger domains predicted, suggesting that thisprotein functions as a DNA-binding protein.

The human gene BCS1L encodes a nuclear-encoded mito-chondrial protein [12] and is important in the respiratorymetabolism of mitochondria. It therefore represents an unlikelyfunctional candidate for the Dbf mutation.

erved.199


FIG. 3. Transcript map acrossthe Dbf region. BAC clones areshown as boxes, with genes asclosed circles, and the relativecentromeric-telomeric orienta-tion is shown. The con-tiguated physical map is alsoshown covering the mouseDbf region on chromosome 1and the region of conserved synteny in human, on chro-mosome 2q35.

The gene Rnf25 encodes one of the RING finger proteins,
which are associated with the ubiquitination of proteins, lead-ing to the destruction of the ubiquitinated protein [13].
The fused gene corresponds to FUH. The protein prod-uct is a positively acting factor serine/threonine kinase inthe Hedgehog pathway [reviewed in 14]. The identificationof a human homologue of this gene has been described [15],although the mouse homologue has not previously beenreported. This gene is expressed in a variety of fetal andadult tissues (C.H., unpublished data; [15]). Therefore,given the previous finding that this mutation exerts its phe-notypic consequences through the upregulation of theHedgehog pathway [3], this gene is a candidate for themutation. (This locus should not be confused with themouse fused gene, originally Fu, now known as AxinFu, onmouse chromosome 17.)

The gene corresponding to human EST KIAA0173 showedhighly significant homology to the porcine tubulin tyrosineligase gene, therefore we designated it to be the mouse andhuman ortholog of this gene, tubulin tyrosine ligase-1 (Ttll1).Furthermore, no expression of this gene was found in 10.5 dpcmutant or wild-type embryos (data not shown).

The encoded peptide product of Cyp27 catalyzes the ini-tial steps in the oxidation of sterol intermediates, leading tothe formation of cholic acid, and would thus appear to beinvolved in basic cell metabolism.

The protein encoded by Ampk3 (cyclic AMP activated pro-tein kinase, g3 subunit) belongs to a family of proteins foundin all eukaryotic cells. This protein family has an importantrole in the regulation of cellular metabolism, particularly inresponse to cues such as stress or vigorous exercise [16].

Expression of Wnt6 in the developing limbs has beenreported [17], and therefore it is regarded as a strong candi-

Copyrig200

date for the mutation. Analysis of the expression pattern ofWnt6 revealed that it is expressed in the developing limb(Figs. 4C and 4E).

The map position of Wnt10a in the mouse was not knownbefore this study. Wnt10a is expressed in the developing limbs(Fig. 4D) and faint staining was observed in the facialprocesses of some embryos (data not shown), which is beingfurther investigated. Therefore, we consider Wnt10a a strongcandidate for the mutation based on its expression pattern inmid-gestation mouse embryos.

Cdk5r2 (formerly p39) encodes one of the cyclin-depen-dant kinases, which are known to have a crucial role in theregulation of the cell cycle, in particular in the G1 to S-phasetransition in the cell [18].

The protein encoded by Pet1 functions in the H+-depen-dant uptake of oligopeptides by the intestinal epithelium.

The gene Cryba2, encoding crystallin-b a2, was excludedas a candidate gene for the mutation as crystallins are unlikelyto have a role in the early patterning events mediating limbdevelopment.

The gene GS1876 (for GeneScan predicted protein, 1876aa) was identified by the comparative sequencing approach.RT-PCR with cDNAs from the mouse Origene panel, whichrepresents a total of 12 cDNA populations, revealed that thisgene was expressed in several tissues, inlcuding thymus,lung, and kidney (C.H. and A.R., unpublished data).

As has been reported previously, the expression of Ihh isdisturbed in the developing limbs and facia of Dbf embryos,consistent with a putative gain-of-function mutation at theIhh locus [4] (Fig. 4A). Therefore, mouse Ihh is also a verystrong candidate for the Dbf mutation. Ihh had previouslybeen mapped proximal to the Dbf locus [4]. This report cor-rects that discrepancy in map position.

GENOMICS Vol. 78, Number 3, December 2001ht © 2001 by Academic Press. All rights of reproduction in any form reserved.


TABLE 1: Genes identified within the mouse and human Dbf region, show-ing cellular localization and biological function

Protein Compartment Function

Arc34 Cytoplasm Cytoskeleton

Aamp Cell surface NK

Hagh Cytoplasm Metabolism

Slc11a1 Cell surface Iron transporter/immune response

Vil Cytoplasm actin binding cytoskeletal structural protein

1154 Plasma membrane* NK

Rqcd1 Plasma membrane/ER* Cellular differentiation

Plcd4 Cytoplasm Second messenger systems

Znf142 Plasma membrane* DNA binding

Bcs1 Mitochondrion Respiratory metabolism

Rnf25 Cytoplasmic* Protein degredation/ubiquitination

Fused Cytoplasm Intracellular signaling

Ttll1 Cytoplasm Cytoskeleton

Cyp27 Mitochondrion Metabolism

Ampkg3 Cytoplasm Metabolism

Wnt6 Secreted Intercellular signaling

Wnt10a Secreted Intercellular signaling

Cdk5r2 Plasma membrane Cell cycle regulation

Pet1 Cell surface Oligopeptide transport

Cryba2 Cytoplasm Structural protein of eye lens

Gs1876 Nuclear* NK

Ihh Secreted Intercellular signaling

Flj10145 Plasma membrane* DNA binding

Stk16 Plasma membrane* NK

Tuba1 Cytoplasm Cytoskeleton

Ptprn Cell surface Neuronal function NO 10.5

Dnpep Plasma membrane* Cellular metabolism

Des Cytoplasm Cytoskeleton/muscle development

Apeg1 Nucleus Vascular system

1279 Cytoplasm* NK

Asic3 Cell surface Proton gated cation channel

Gmppa Plasma membrane Cellular respiration

0657 Plasma membrane* NK

Inha Extracellular Cell-cell signaling

Slc4a3 Cell surface Anion exchanger*Cellular compartment predicted from sequence analysis by PSORT.

Although sequence analysis of the predicted protein of the pituitary, spin
FLJ10145 locus (GenBank acc. no. AK001277) did not reveal anysignificant hits within the non-redundant databases, a putativezinc-finger motif was predicted at amino acids 74 to 96, sug-gesting that this protein may have DNA binding capabilities.
GENOMICS Vol. 78, Number 3, December 2001Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.

A fatty acid acylated serine/threonine pro-tein kinase gene [19] had previously beengenetically mapped to mouse chromosome 11[20]. The human gene STK16 forms part of aUnigene cluster (Hs.153003) and was mappedto 2q34–q37, which we have refined to 2q35.As we have also identified this gene in a clonecontig that covers mouse chromosome 1, weconclude that Stk16 does not map to chromo-some 11, but chromosome 1. A potential inter-pretation of this result is that in the previousstudy, the authors may have mapped anotherclosely related gene or a processed Stk16pseudogene.

The genetic map position of the mousetubulin a4 gene (Tuba4) had not previouslybeen determined.

Ptprn (protein tyrosine phosphatase recep-tor-N) was found not to be expressed inembryos of wild-type or mutant genotype at10.5 dpc; therefore, it can be excluded as a can-didate gene for the Dbf mutation.

Dnpep (aspartyl aminopeptidase) was iden-tified from the EST mapping effort at MRCHarwell (http://www.mgc.har.mrc.ac.uk/est_mapping.html). It was also found to bepresent in the draft sequence from the Dbfregion and the genetic map position of thegene in either mouse or human had yet to bedetermined.

The gene Des encodes an intermediate fil-ament protein, which is expressed in skeletal,visceral, and smooth muscle cells [21].

Apeg1 (aortic preferentially expressedgene-1) is a nuclear gene that is preferentiallyexpressed in smooth muscle cells and is downregulated in response to vascularinjury [22].

Searches of the non-redundant nucleic acidand peptide databases did not reveal any sig-nificant similarities for the KIAA1279 clone,making the inference of a biological role forthe predicted translation product of this geneextremely difficult. Nevertheless, sequenceanalysis of the predicted peptide suggested thepresence of a left-handed coil in the protein, atamino acid position 161–189 of the translatedfragment of KIAA1279.

ASIC4 encodes a non-proton gated acidsensing gated ion channel [23]. This geneshows strongest expression in the developingal cord, and inner ear [24].

The enzyme encoded by Gmppa (GDP mannose pyrophos-phorylase A) catalyzes the reaction that converts mannose-1-phosphate and GTP to GDP-mannose, which is involved inthe production of N-linked oligosaccharides.

201


FIG. 4. Expression of Dbf candidate genes. (A) Expression of the mouse Aamp gene in the developing cardiovascular system, note the lack of expression in thedeveloping limb buds. (B) Expression of Ihh in wild-type and Dbf/+ limb buds at 11 dpc, showing the upregulation of Ihh in the precartilaginous mesenchymeof the mutant. (C) Expression of Wnt6 in the developing limb buds of 10.5 dpc wild-type and Dbf/+ embryos. (D) Expression of Wnt10a in the progress zone of10.5 dpc wild-type and Dbf/+ embryos. (E) RT-PCR analysis of Wnt6 transcripts from wild-type and Dbf/Dbf embryos at 12.5 dpc, showing the relatively lowlevel of the larger isoform of Wnt6 in homozygous mutant RNA.

A

B

C

D

E

A novel gene was also identified that corresponded to thehuman EST KIAA065, derived from adult brain. A secondarystructure prediction program predicted the correspondinghuman protein to be located within the plasma membrane.

Inha (inhibin-a) was found to be contained within the non-recombinant interval for the mutation. Targeted disruption inthis gene by homologous recombination in embryonic stemcells resulted in mice that developed stromal tumors [25].

Slc4a3, which is expressed in neurons and cardiac muscle[26], was found to be recombinant with the mutation, andhence residing outside the Dbf critical region. Thus it can beexcluded as candidate for the mutation.

Expression of Developmentally Important Signaling MoleculesWe assessed the expression of the three developmentallyimportant signaling molecules, IHH, WNT6, and WNT10A,by whole-mount in situ hybridization to further delineate the

Copyrig202

candidacy of their genes for the mutation. In accordance withprevious reports [4], the expression of Ihh was disturbed inthe developing limb buds of Dbf mutant animals, such that itwas upregulated in the precartilaginous mesenchyme of themutant limb buds (Fig. 4B).

Both of the Wnt genes were also found to be expressed inthe developing limb buds of the mutant and wild-typeembryos. Wnt6a was expressed in the developing progresszone mesenchyme and overlying ectoderm of both mutantand wild-type mice (Fig. 4C), with Wnt10a showing a similarprogress zone expression pattern in the limb buds (Fig. 4D).As a further step to identify a putative mutation within thecoding sequences of these genes, the exons of all of the geneswere sequenced (including at least 50 bp of flanking intronicsequence) to rule out the possibility of a sequence change injuxtaposition to a splice site. No changes were identifiedwithin the coding sequence of any of these candidate genes.RT-PCR analysis of Wnt6 revealed that the larger of the two



isoforms of this gene [17] was consistently present in lowerabundance in Dbf/Dbf embryos than the corresponding tran-script in total RNA prepared from wild-type control litter-mates (Fig. 4E). This raises the possibility that the mutationwas affecting either the splicing or the relative abundance ofthe isoforms of this gene, although in this regard no putativesequence changes at the nucleic acid level were identified inWnt6 (C.H. et al., unpublished data), suggesting that if this dif-ference is indeed functional then it is probably not due to asimple splice-site mutation per se.

DISCUSSION

In a previous report the dominant Dbf mutation had beenlocalized to a 13-cM interval on mouse chromosome 1,between the microsatellite markers D1Mit22 and D1Mit8 andnon-recombinant with the markers D1Mit77 and D1Mit24 [1].We have carried out standard karyotype analysis of Dbfchromosomes, suggesting there are no gross chromosomalaberrations associated with the Dbf mutation. Genetic analy-sis of the region indicated that no suppression of recombina-tion was observed, further suggesting that there were no chro-mosomal rearrangements associated with the Dbf region.These data suggest that the mutation is not due to a large-scale rearrangement across this region, although this datadoes not preclude the possibility that smaller rearrangementshave occurred within the region of the Dbf locus. We haveexpanded the backcross to 989 progeny in order to generatenew informative recombination events, as well as to refine thegenetic distances between markers. This expanded backcrosshas also allowed us to construct a physical map of the Dbfregion.

This work describes the construction of an integratedgenetic and physical map around the Dbf locus on mousechromosome 1. In this study we presented data on the iden-tification of 35 genes within the Dbf region of mouse chro-mosome 1, many of which are novel or have not been mappedto the region before. Furthermore, we have identified severalgenes as strong candidates for this developmental mutation.

In a previous report [4], a putative Dbf candidate gene, Ihh,was genetically mapped through a Dbf interspecific backcross.Ihh was reported to be recombinant with the Dbf mutation in4 of 351 meioses analyzed, placing the Ihh locus 1.3 cM prox-imal to the Dbf mutation, a distance equating to a physical dis-tance of approximately 2.5–3 Mb in the mouse genome. Inthat same study the authors reported that there was ectopicexpression of Ihh in the limbs and facia of Dbf embryos. Theaberrant expression of a Hedgehog gene fitted precisely withthe functional characterization of the polarizing activity ofthe Dbf limb mesenchyme [3], making this a highly attractivecandidate gene for the mutation. Yet the large genetic dis-tances involved precluded the possibility that the primaryaffect of the mutation was on the Ihh locus, suggesting thatthe mutation was either exerting an unprecedented long-range effect on the Ihh locus, or that the mutation was acting

GENOMICS Vol. 78, Number 3, December 2001Copyright © 2001 by Academic Press. All rights of reproduction in any form res

on a factor which in turn was causing an alteration in theexpression of Ihh.

The present study sheds further light on the potential dis-parity between the genetic position of the mutation and theIhh locus. As a result of our genetic mapping and samplesequencing efforts, we have now definitively positioned thefunctional copy of mouse Ihh within the Dbf non-recombinantinterval. The data demonstrate that the Ihh locus is non-recombinant with the Dbf mutation and that the mutationcould be acting in cis through the Ihh locus, acting in effect asa gain-of-function regulatory mutation of Ihh. Therefore weconsider Ihh to be a strong candidate gene for the Dbf muta-tion. Given the discrepancy between our data and previousdata [4], we suggest that the most likely explanation for thepreviously reported genetic map position of Ihh would bethat a pseudogenic copy of Ihh was mapped in lieu of thefunctional copy.

Database searching of the Mouse Genome Informaticsresource (http://www.informatics.jax.org) revealed thatanother member of the vertebrate family of Hedgehog genes,Desert hedgehog (Dhh), lies in close proximity to twoWingless genes on mouse chromosome 15, Wnt1 and Wnt10b,just as Ihh lies near two Wnt genes, Wnt6 and Wnt10a. Thissame relationship can be seen in a sequenced cosmid from thepufferfish, Fugu rubripes, containing dhh, wnt1, and wnt10b[27]. The most parsimonious explanation is that a Hedgehoggene and two Wnt genes have undergone duplication in thevertebrate lineage (A.R. and C.H., manuscript in preparation;[28]).

This effort has identified several putative candidates forthe Dbf mutation. Wnt6 and Wnt10a are also strong candi-dates for the mutation as the expression pattern of both ofthese genes correlates with sites of pathology in the mutant.Furthermore, mutation of another member of this family(Wnt7a) has been shown to underlie limb defects includingpolydactyly in the classical mutation postaxial hemimelia(Wnt7apx) [29]. Therefore, mutations in a member of the Wntfamily of signaling molecules expressed in the developinglimb could underlie a polydactyly.

Given that in our previous analyses of this mutant themutant tissue seemed to be the source of a positively actingsecreted factor [3], it is notable that we have identified threegenes that would readily fit this criterion: Ihh, Wnt6, andWnt10a.

It is interesting that all of the genes identified as candi-dates for the mutation are components of developmentallysignificant signal transduction pathways. A further findingwas the identification of a “cassette” of genes that has beeninherited as a group throughout the evolution of the verte-brates, raising the possibility that the phenotypic effect of themutation might be due to a cumulative effect on the misreg-ulation of one or more of these loci. Clearly, further analysisof these candidate genes is required to further identify the Dbfgene and the causative mutation, including further expressionanalysis and sequencing of these genes in mutant and wild-type mice.

erved.203


Could the Dbf mutation be a regulatory mutation affect-ing the expression of Ihh, Wnt6, or Wnt10a? Our data lendconsiderable weight and support to this hypothesis, whileexcluding the possibility that the mutation is due to a large-scale genomic rearrangement. As we and others have demon-strated, the expression of Ihh, which we have affirmativelypositioned with the non-recombinant critical interval for themutation, is upregulated in the precartilaginous mesenchymeof Dbf mutant limb bud mesenchyme [4] (Fig. 4B). We havealso been able to demonstrate that in Dbf/Dbf embryos thelarger of the two isoforms of Wnt6 is present at consistentlylower levels than the smaller transcript, which is generatedby alternative splicing. One explanation for these observa-tions is that the Dbf mutation represents a regulatory muta-tion that affects the differential expression of both Ihh andWnt6. To this end, it is worth pointing out that the upregula-tion of Ihh in the progress zone mesenchyme of the mutantlimb buds mimics the endogenous expression of Wnt6 in thewild-type limb bud. Furthermore, given that these genes havebeen inherited as a transcriptional cassette throughout verte-brate evolution, it is possible that they may indeed share com-mon cis-regulatory elements.

Another interesting candidate is the fused gene. This geneis a positively acting component of the Hedgehog pathway[14]. Analysis of the Dbf mutant revealed that the mutationwas acting through the ectopic upregulation of the Hedgehogpathway [3,4]. Therefore this locus merits further, in-depthscrutiny to assess its candidacy for the mutation. The infor-mation presented here, in terms of both gene content and theintegrated mouse and human physical maps, provides toolsand reagents that will be useful functional genomics tools inthe post-genome era.

MATERIALS AND METHODS

SSLP analysis and genotyping. Backcross progeny were analyzed using SSLPmarkers obtained from Research Genetics (Huntsville, AL). Genomic DNA wasprepared from a tail biopsy, 20 hg of DNA was amplified in a 10 ml reactionvolume containing 1.5 mM magnesium chloride with a thermal cycle profileof 948C for 2 min, and then 35 cycles of 948C for 30 s, 558C for 1 s, 728C for 5 swith a final extension step of 728C for 2 min were carried out. PCR productswere electrophoresed and scored on Visigel separation matrix (Stratagene, LaJolla, CA).

Haplotype analysis. Mapping of genomic markers isolated from the YAC andBAC contigs as well as of coding fragments identified from the non-recombi-nant interval (see below) was performed by combining SSCP and heteroduplexanalysis in silver-stained gels [30].

Isolation of BAC and PAC clones. The RPCI-21 PAC library supplied by theMRC HGMP-RC (Hinxton, Cambs, UK) by hybridization with oligonucleotidesdesigned against the non-repetitive regions of SSLP markers and genic mark-ers. Fingerprint data for clones from the RPCI23/24 BAC libraries constructedby Pieter de Jong were taken from http://www.bcgsc.ca/projects/mouse_mapping. Contigs were constructed using the FPC software(C. Soderlund et al., unpublished data), using a tolerance of 5 and a cutoff of10–08.

Isolation of STSs from BAC and PAC clones. End-clones from BACs (ResearchGenetics, Huntsville, AL) were isolated by the TAIL-PCR method [31]. Thenested primers for the vector pBELOBAC were as described [32]. End clones

Copyrig204

from PAC clones were isolated by direct sequencing. Briefly, DNA was pre-pared from 3 ml bacterial culture and resuspended in 5 ml water. This wasadded to 16 ml BigDye Terminator (PE Biosystems), 5 ml T7 or Sp6 sequencingprimer (10 mM), 1 ml of 1 mM Mg2+, and 13 ml water.

Sequencing of clones. The BAC and PAC clones were sheared with a nebulizerand fragments in the range of 1.2 to 1.5 bp were subcloned in M13mp18 vec-tor [33]. At least 1000 plaques were selected from each clone library, and M13ssDNA was prepared and sequenced using dye-terminators, ThermoSequenase(Amersham), and universal M13-primer (MWG-Biotech). The gels were run onABI-377 sequencers and data were assembled and edited using the GAP4 pro-gram [34]. Genomic DNA sequence analysis was performed using the auto-mated sequence annotation system RUMMAGE [35].

ACKNOWLEDGMENTSThis work is supported by the MRC. C.H. is in receipt of an MRC Research Fellowship.G.M.M.-K. is supported by the BBSRC. RPCI-21 PAC genomic libraries and clonesand IMAGE cDNA clones were obtained from the MRC HGMP-RC, Hinxton, UK.

RECEIVED FOR PUBLICATION APRIL 4; ACCEPTED SEPTEMBER 20, 2001.

REFERENCES1. Lyon, M. F., et al. (1996). Doublefoot: a new mouse mutant affecting development of limbs

and head. Genet. Res. 68: 221–231.2. Hayes, C., Lyon, M. F., and Morriss-Kay, G. M. (1998). Morphogenesis of Doublefoot

(Dbf), a mouse mutant with polydactyly and craniofacial defects. J. Anat. 193: 81–91.3. Hayes, C., Brown, J. M., Lyon, M. F., and Morriss-Kay, G. M. (1998). Sonic hedgehog is

not required for polarising activity in the Doublefoot mutant mouse limb bud. Development125: 351–357.

4. Yang, Y., Guillot, P., Boyd, Y., Lyon, M. F., and McMahon, A. P. (1998). Evidence thatpreaxial polydactyly in the Doublefoot mutant is due to ectopic Indian Hedgehog signal-ing. Development 125: 3123–3132.

5. Rhodes, M., et al. (1998). A high-resolution microsatellite map of the mouse genome.Genome Res. 8: 531–542.

6. Mallon, A. M., et al. (2000). Comparative genome sequence analysis of the Bpa/Str regionin mouse and man. Genome Res. 10: 758–775.

7. Welch, M. D., DePace, A. H., Verma, S., Iwamatsu, A., and Mitchison, T. J. (1997). Thehuman Arp2/3 complex is composed of evolutionarily conserved subunits and is local-ized to cellular regions of dynamic actin filament assembly. J. Cell Biol. 138: 375–384.

8. Beckner, M. E., and Liotta, L. A. (1996). AAMP, a conserved protein with immunoglob-ulin and WD40 domains, regulates endothelial tube formation in vitro. Lab. Invest. 75:97–107.

9. Robine, S., Jaisser, F., and Louvard, D. (1997). Epithelial cell growth and differentiation.IV. Controlled spatiotemporal expression of transgenes: new tools to study normal andpathological states. Am. J. Physiol. 273: 759–762.

10. Ferrary, E., et al. (1999). In vivo, villin is required for Ca(2+)-dependent F-actin disrup-tion in intestinal brush borders. J. Cell Biol. 146: 819–830.

11. Okazaki, N., et al. (1998). Novel factor highly conserved among eukaryotes controls sex-ual development in fission yeast. Mol. Cell. Biol. 18: 887–895.

12. Folsch, H., Guiard, B., Neupert, W., and Stuart, R. A. (1996). Internal targeting signal ofthe BCS1 protein: a novel mechanism of import into mitochondria. EMBO J. 15: 479–478.

13. Freemont, P. S. (2000). RING for destruction? Curr. Biol. 10: 84–87.14. Hammerschmidt, M., Brook A., and McMahon, A. P. (1997). The world according to

hedgehog. Trends Genet. 13: 14–21.15. Murone, M., et al. (2000). Gli regulation by the opposing activities of fused and sup-

pressor of fused. Nat. Cell Biol. 2: 310–312.16. Kemp, B. E., et al. (1999). Dealing with energy demand: the AMP-activated protein kinase.

Trends Biochem. Sci. 24: 22–25.17. Parr, B. A., Shea, M. J., Vassileva, G., and McMahon, A. P. (1993). Mouse Wnt genes

exhibit discrete domains of expression in the early embryonic CNS and limb buds.Development 119: 247–261.

18. Malumbres, M., Ortega, S., and Barbacid, M. (2000). Genetic analysis of mammaliancyclin-dependent kinases and their inhibitors. Biol. Chem. 381: 827–838.

19. Berson, A. E., et al. (1999). Identification and characterization of a myristylated andpalmitylated serine/threonine protein kinase. Biochem. Biophys. Res. Commun. 259:533–538.

20. Stairs, D. B., et al. (1998). Cloning and characterization of Krct, a member of a novel sub-family of serine/threonine kinases. Hum. Mol. Genet. 13: 2157–2166.

21. Li, Z. L., Mattei, M. G., Mattei, J. F., and Paulin, D. (1990). Assignment of the mousedesmin gene to chromosome 1 band C3. Genet. Res. 55: 101–105.

22. Hsieh, C. M., et al. (1996). APEG-1, a novel gene preferentially expressed in aortic smoothmuscle cells, is down-regulated by vascular injury. J. Biol. Chem. 271: 17354–17359.

23. Akopian, A. N., Chen, C. C., Ding, Y., Cesare, P., and Wood, J. N. (2000). A new



member of the acid-sensing ion channel family. Neuroreport 11: 2217–2222
24. Grunder, S., Geissler, H. S., Bassler, E. L., and Ruppersberg, J. P. (2000). A new member
of acid-sensing ion channels from pituitary gland. Neuroreport 11:1607–1611.25. Matzuk, M. M., Finegold, M. J., Su, J. G., Hsueh, A. J., and Bradley, A. (1992). a-Inhibin

is a tumour-suppressor gene with gonadal specificity in mice. Nature 360: 313–319.26. Linn, S. C., Kudrycki, K. E., and Shull, G. E. (1992). The predicted translation product of

a cardiac AE3 mRNA contains an N terminus distinct from that of the brain AE3Cl–/HCO3– exchanger. Cloning of a cardiac AE3 cDNA, organization of the AE3 gene,and identification of an alternative transcription initiation. J. Biol. Chem. 267: 7927–7935.

27. Gellner, K., and Brenner, S. (1999). Analysis of 148 kb of genomic DNA around the wnt1locus of Fugu rubripes. Genome Res. 9: 251–258.

28. Nusse, R. (2001). An ancient cluster of Wnt paralogues. Trends Genet. 17: 443.29. Parr, B. A., Avery, E. J., Cygan, J. A., and McMahon, A. P. (1998). The classical mouse

mutant postaxial hemimelia results from a mutation in the Wnt 7a gene. Dev. Biol. 15:228–234.

GENOMICS Vol. 78, Number 3, December 2001Copyright © 2001 by Academic Press. All rights of reproduction in any form rese

30. Joensuu, T., et al. (1996). Refined mapping of the Usher syndrome type III locus on chro-mosome 3, exclusion of candidate genes, and identification of the putative mouse homol-ogous region. Genomics 38: 255–263.

31. Liu, Y. G., and Whittier, R. F. (1995). Thermal asymmetric interlaced PCR: automatableamplification and sequencing of insert end fragments from P1 and YAC clones for chro-mosome walking. Genomics 25: 674–681.

32. Blanco, G., et al. (1998). A STS content physical and transcription map across the ky,kyphoscoliosis, nonrecombinant region. Genomics 54: 415–423.

33. Yanisch-Perron, C., Vieira, J., and Messing, J. (1985). Improved M13 phage cloning vec-tors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene33: 103–119.

34. Bonfield, J. K., Smith, K. F., and Staden, R. (1995). A new DNA sequence assembly pro-gram. Nucleic Acids Res. 23: 4992–5009.

35. Taudien, S., et al. (2000). RUMMAGE—a high-throughput sequence annotation system.Trends Genet. 16: 519–521.

rved.205

A High-Resolution Genetic, Physical, and Comparative Gene Map of the Doublefoot (Dbf) Region of...

Documents

Transcript of A High-Resolution Genetic, Physical, and Comparative Gene Map of the Doublefoot (Dbf) Region of...