Genomic Structure, Sequence, and Mapping ofHuman FGF8 With No Evidence for Its Role inCraniosynostosis/Limb Defect Syndromes

Koh-ichiro Yoshiura,1 Nancy J. Leysens,1 Jenny Chang,3 Deeann Ward,3,4 Jeffrey C. Murray,1,2* andMaximilian Muenke3,4

1Department of Pediatrics, The University of Iowa, Iowa City2Department of Biological Sciences, University of Iowa, Iowa City3Division of Human Genetics and Molecular Biology, The Children’s Hospital of Philadelphia,Philadelphia, Pennsylvania

4Departments of Pediatrics and Genetics, University of Pennsylvania School of Medicine, Philadelphia

Fibroblast growth factor-8 (Fgf8) is a re-cently identified growth factor that stimu-lates the androgen-dependent growth ofmouse mammary carcinoma cells. Evidencefrom mouse development also shows thatFgf8 may play an important role in growthand patterning of limbs, face, and the cen-tral nervous system. We describe here thehuman FGF8 genomic sequence and demon-strate conservation between the human andmouse sequences, including alternativelyspliced exons in the mouse. Mapping ofFGF8 by FISH using an FGF8-containingbacterial artificial chromosome and by ge-netic linkage using a SSCP variant identi-fied in this study is also reported and re-fines the FGF8 map location to 10q24. SinceFGF8 maps to the same chromosomal regionas FGFR2, has indeed been shown to be aligand for FGFR2, and has an expression pat-tern consistent with limb and craniofacialanomalies, we have screened two kindredswith Pfeiffer syndrome that were previ-ously linked to markers from 10q24-25 and alarge number of individuals with craniosyn-ostosis and limb anomalies for mutations inthe coding sequence of FGF8. While no suchmutations were identified, a rare polymor-phic variant, consisting of an 18-base-pair(six-amino-acid) duplication in exon 1c, isreported that apparently has no clinical ef-fect. Our exclusionary data suggest that mu-tations in FGF8 would be, at best, an infre-

quent cause of such disorders. Am. J. Med.Genet. 72:354–362, 1997. © 1997 Wiley-Liss, Inc.

KEY WORDS: fibroblast growth factor-8;FGF8; craniosynostosis ;Pfeiffer; FISH; alternativesplicing; FGFR2

INTRODUCTIONThe fibroblast growth factors (Fgf) are signaling mol-

ecules that are secreted or localized to the nucleus[Acland et al., 1990; Bugler et al., 1991]. Most FGFs actas ligands that bind to fibroblast growth factor recep-tors (Fgfrs) with high and/or low affinity [Schlessingeret al., 1995]. The Fgf family consists of at least tenmembers [Fgf1–Fgf10; Muenke and Schell, 1995; Ohu-chi et al., 1997] that have a conserved amino acid coresequence. Fgfs are attractive as oncogenes because oftheir transforming activities [MacArthur et al.,1995a,c; Tanaka et al., 1992], but recent studies showthat Fgfs also play an important role in the developingembryo [Tanaka et al., 1992; Cohn et al., 1995; Slack,1995] and in specific cell type development. For ex-ample, the Fgf3 knockout mouse has a short tail andinner ear developmental defects [Mansour et al., 1993],Fgf4 is essential for post-implantation proliferation ofthe inner cell mass [Feldman et al., 1995], and Fgf5regulates the hair follicular cell cycle [Hebert et al.,1994].

Fgf8 was originally isolated as an androgen-inducedgrowth factor in a mouse mammary carcinoma cell line[Tanaka et al., 1992] and is expressed in the gonad inthe adult mouse [Lorenzi et al., 1995; MacArthur et al.,1995c]. In the developing embryo, Fgf8 is expressed inthe developing limb, primitive streak, heart, branchialarch, olfactory epithelium, pharyngeal region, andvarious regions in the central nervous system [Heikin-heimo et al., 1994; Crossley and Martin, 1995]. How-ever, we do not know which regions require Fgf8 ex-pression. Fgf8 consists of at least seven isoforms in the

Contract grant sponsor: NIH; Contract grant numbers: DE-08559, DE-09170, HD28732, HD29862.

*Correspondence to: Dr. Jeffrey Murray, Department of Pedi-atrics, University of Iowa, 200 Hawkins Drive, 2613 JCP, IowaCity, IA 52242-1083.

Received 9 April 1997; Accepted 9 April 1997

American Journal of Medical Genetics 72:354–362 (1997)

© 1997 Wiley-Liss, Inc.

developing mouse embryo for normal development. Allof these isoforms are created by alternative splicing inthe 58 region of the gene [Tanaka et al., 1992; Crossleyand Martin, 1995; MacArthur et al., 1995b]. Crossleyand Martin [1995] have shown that variant 1 may bethe predominantly expressed form in the developingembryo. Recent evidence also suggests that isoforms 8band 8c may provide specific signals for craniofacial andlimb development [MacArthur et al., 1995b; Crossley etal., 1996].

Fgf8 has been mapped to mouse chromosome 193C-D [Mattei et al., 1995] and human chromosome10q25-26 [Lorenzi et al., 1995; White et al., 1995]. Thelatter chromosomal region also contains the fibroblastgrowth factor receptor-2 gene (FGFR2). It is of interestthat FGF8 is one of the ligands for FGFR2 or FGFR3[MacArthur et al., 1995b]. Mutations in FGFR2 havebeen shown to cause clinically distinct craniosynostosissyndromes–Crouzon [Jabs et al., 1994; Reardon et al.,1994], Jackson-Weiss [Jabs et al., 1994], Pfeiffer [Rut-land et al., 1995; Schell et al., 1995], and Apert [Wilkieet al., 1995] syndromes. In addition, a dominant vari-ant of the split hand/split foot syndrome (SHFM3) wasalso mapped to this region after evaluation of a patientwith abnormalities of the face and limbs who had a10q25 unbalanced translocation [Nunes et al., 1995].Finally, two families with Pfeiffer syndrome weremapped to 10q25 but have not been shown to havemutations in FGFR2 [Schell et al., 1995].

Here we report human FGF8 genomic structure,exon-intron boundary sequences, and screening for mu-tations in this gene in two kindreds and a large numberof individuals with craniosynostosis and limb abnor-malities. FISH and genetic linkage studies adjust theFGF8 location to 10q24 in humans, and a series of poly-morphic variants is demonstrated. No mutations were

identified in 63 families including a range of craniosyn-ostosis types.

MATERIALS AND METHODSCloning and Sequence of the Human FGF8

Genomic RegionA primer set was designed from the FGF8(AIGF) 38

region of the gene (58-CGAGGGCTGGTACATGGCCT-38, 58-GGCCCAAGTCCTCTGGCTGC-38) that ampli-fied two fragments from human genomic DNA. Ampli-fication was carried out in 1 × PCR buffer (Boehringer-Mannheim Biochemicals, Indianapolis, IN) with 5%DMSO using 200 mM dNTPs and 1 mM forward andreverse primers at an annealing temperature of 62°C.One fragment was about 200 bp in length and the otherabout 150 bp. Sequencing showed that the larger frag-ment was the mouse Fgf8 homolog while the shorterhad no homology with Fgf8.

A human bacterial artificial chromosome (BAC) li-brary (Research Genetics, Huntsville, AL) was thenscreened with this primer set. Positive amplificationsfor the 200 bp fragment were seen from a few of theDNA pools, and one of the positive clones was chosenfor further sequence analysis. DNA from the BAC wasextracted with a Qiagen column (Qiagen, Chatsworth,CA) under manufacturer-specified conditions. TheBAC DNA was digested with Sau3AI or EcoRI andcloned in BamHI- or EcoRI-digested pT7T3a-19 plas-mid. The plasmid library was screened with five oligo-nucleotides designed to span the mouse Fgf8 gene se-quence, and positive plasmids were purified and se-quenced with the identifying oligonucleotides and T3and T7 primers.

Once the homologous sequences with mouse Fgf8were identified, oligonucleotides were synthesized fromthose sequences and used for sequence reaction. All

TABLE I. PCR Conditions and Primer Sequences for SSCP Analysis

1× PCR buffer (Boehringer, Germany):10-mM Tris-HCl (pH 8.3),1.5-mM MgCl2, 50-mM KCl

200 mM in each of the fourdNTPs (except as noteda)

1 mM in each primerat final concentration

Primer SequenceAnnealingtemp. (°C)

% DMSOof additive

FGF8.1Fa 58-ACC CGC ACC CTC TCC GCT-38 58 10FGF8.1Ra 58-CCT CCT CAC CTG GGC TTG G-38FGF8.2F 58-GTT GCA CTT GCT GGT CCT CT-38 60 5FGF8.2R 58-CAG GAG GGG TGC TAC CTC T-38FGF8.3F 58-ACC GGG ACT GAC TCT GCG G-38 60 5FGF8.3R 58-CGG CTG GGT GAA CCC TAT G-38FGF8.4F 58-CTC TCA ACA TTT GCT CCG TA-38 60 5FGF8.4R 58-CCC ACC CAC GCA AGT CGG-38FGF8.5F 58-CTG GCG GCT TGG GGC AGT-38 60 5FGF8.5R 58-CCG CTG GTG CGG CTG TAG A-38FGF8.6F 58-GAG CCT GGT GAC GGA TCA GC-38 60 5FGF8.6R 58-CGG CCA GTG CAG TTG GGA C-38FGF8.7F 58-TTT GGA GCA GTT GCT GCT GG-38 60 5FGF8.8R 58-TGT TCA TGC AGA TGT AGA GGC-38FGF8.8F 58-CGT GGA GAC GGA CAC CTT TG-38 60 5FGF8.9R 58-TAC CTT GTT GGG ATC AGA GCC-38FGF8.10F 58-GGG TGC CCA CCT GCT GTC T-38 60 5FGF8.10R 58-TGG TGC TGC CGC GTC TTG G-38FGF8.11F 58-GCT GGT ACA TGG CCT TCA CC-38 60 5FGF8.11R 58-TCT CTG CGG TCT GGC ATT GT-38

aThe specific amplification of FGF8.1F/1R required 500 mM of each of the dNTPs.

Structure, Sequence, and Mapping Human FGF8 355

sequence reactions were performed with the ABI-PRISM dye-terminator cycle sequencing kit and elec-trophoresed on an auto-sequencer (model 373A, Ap-plied Biosystems, Foster City, CA). Sequence reactionswere performed on both strands to verify the sequence.The sequence results are shown in Figure 1b.

Physical and Genetic Mapping of Human FGF8

The human FGF8 gene was regionally localized onchromosome 10 by fluorescence in situ hybridization(FISH) of the FGF8-containing BAC which was identi-fied in this study. The BAC was labeled with biotin bynick translation as recommended by the supplier (On-cor, Inc., Gaithersburg, MD). Metaphase spreads wereprepared from normal blood lymphocytes or a lympho-blastoid cell line by standard methods. FISH using the

biotin-labeled FGF8-BAC and a chromosome 10-specific centromere probe (Oncor, Inc., Gaithersburg,MD) and analysis were done as described previously[Lichter et al., 1990; Ried et al., 1992; Yoshiura et al.,1993]. A total of 18 metaphases was analyzed afterFISH. The distance of the chromosome 10 centromereto the hybridization signal was measured, comparedto the total length of 10q, and plotted on diagramsin which the individual widths of chromosome bandswere as observed on banded chromosomes [Francke,1994].

Genetic mapping by using an SSCP variant de-scribed below was also carried out on the CEPH refer-ence pedigrees. Fifteen CEPH families comprising theprimary-linkage and secondary-linkage trays of the Co-operative Human Linkage Center [CHLC; Murray et

Fig. 1. Genomic structure of human FGF8. a: The lengths of introns and exons are indicated in kb (kilobase pairs) or bp (base pairs). E, exon (exon1a, exon 1b, exon 1c, exon 1d, exon 2, exon 3); I, intron. b: Genomic sequence of human FGF8. Exon sequences are shown in upper case and intron in lowercase. Exon-intron boundary consensus sequence is conserved in all of them. Translated amino acid sequence of human and mouse is shown below thenucleotide sequence. Polymorphic sites are indicated by boldfaced letters.

356 Yoshiura et al.

al., 1994] was carried out. Genotypes were entered intoCEPH database format and linkage carried out usingthe publicly available linkage server of the CHLC(http://www.chlc.org/). Individual and pairwise link-ages as well as bin assignments were provided by thelinkage server results.

SSCP Analysis of Patients WithCraniosynostosis and Limb Anomalies

An initial FGF8 mutation screening by SSCP wasperformed in two large families with Pfeiffer syndrome.They were selected because affected individuals hadcraniofacial anomalies and limb defects linked to mark-ers in 1024-q25 but did not have any detectable muta-tions in FGFR2 [Schell et al., 1995]. In addition, onefamily with craniosynostosis type Philadelphia [Robinet al., 1996], which was compatible with linkage to10q24-25 markers (Muenke, unpublished) wasscreened. Last, we screened a total of 65 DNA samplesfrom patients with sporadic Pfeiffer syndrome or othercraniosynostosis syndromes with anomalies of thehands and feet which were negative for the known mu-tations in FGFR1 [Muenke et al., 1994] and FRFR2[Schell et al., 1995].

Ten primer sets were designed to amplify all of theexons and the exon-intron boundaries. The sequencesof these primer sets and amplification conditions areshown in Table I. Fifty-nanogram genomic DNA of in-dividuals with craniosynostosis and limb anomalieswas used as a template, with 1 mCi (0.1 ml) [a-32P]dCTP (Amersham, Arlington Heights, IL) in 10 ml PCRmixture. The amplified products were mixed with 40 mlformamide loading buffer, heated to 94°C for 3 min-utes, and quenched on ice prior to loading.

Gels, 0.5× MDE-gel (AT Biochem, Malvern, PA),were run in 0.6× TBE buffer with or without 5% glyc-erol. The gels were run at a constant 10 W for glycerol-containing gels and at a constant 5 W for the glycerol-less, with cooling for 12 to 16 hours at room tempera-ture. The gels were transferred to Whatmann 3MMfilter paper, dried thoroughly, and used to expose Ko-dak X-OMAT AR film, without intensifying screen, atroom temperature for 6 to 18 hours.

When the SSCP analysis showed a mobility shift,10 ml PCR amplification was performed again withoutisotope, and 0.1 ml from the 10ml mixture was used forasymmetric PCR (100:1 primer ratio) in 30 ml PCR.Asymmetric PCR products were purified with a QiagenPCR purification column and sequenced with an ABI-PRISM dye-terminator cycle sequencing kit and a 373Aautomated sequencer (ABI) or with a Taq cycle se-quencing kit (Perkin-Elmer, Oak Brook, IL) with

[a-33P] dATP (Amersham, Arlington Heights, IL).When PCR products were to be cloned into plasmids,PCR was performed in 50 ml, and the products werepurified with a Qiagen column and cloned into pKRXT-tailed vector [Pruessner et al., 1995]. Fifty plasmidscontaining PCR products were picked up and culturedall together. This plasmid pool DNA was extracted witha Qiagen column and used as a sequence template.

RESULTSSequence Analysis

Sequence analysis demonstrates that human FGF8is highly conserved with the mouse Fgf8. The putativehuman FGF8 is completely identical to Fgf8b in aminoacid sequence (215/215), and exon 1c in humans has 26amino acids out of 29 (89.6%) identical to the equiva-lent exon [Crossley and Martin, 1995; MacArthur et al.,1995b,c] in mouse. Interestingly, the extended exon 1b,which is used for Fgf8c, Fgf8d, Fgf8g, and Fgf8h, is nothomologous between human and mouse. Not only isthere less homology, but there is also an in-frame ter-mination codon in this region in human FGF8. It maybe that isoforms using the extended exon 1b region arenot expressed in humans and are less important forFgf8 function in the developing mouse embryo. Thisidea is supported by the evidence that Fgf8b more ef-ficiently transforms the NIH 3T3 cells than Fgf8c[MacArthur et al., 1995a] and that Fgf8b is expresseddominantly over other isoforms [Crossley and Martin,1995]. However, the extended region of exon 1b that isnot homologous to the human sequence encodes one ofthe two glycosylation consensus sequences in themouse, so it is possible that isoforms containing theexon1b extended region are important only in mouse,not in humans.

Regional Mapping of FGF8

Linkage analysis using the common FGF8 polymor-phism described above was carried out and providedLOD scores greater than 3 with the following fourmarkers: D10S192, D10S205, D10S222, and D10S184.These markers are located in interval 13 of the CHLCframework maps and support the localization of thismarker to 10q25. In addition, FISH mapping using thebiotin-labeled BAC showed all FGF8-specific signals on10q corresponding to a region near band 10q24.

Mutation Analysis

SSCP analysis of each of the coding exons with aportion of flanking intron sequence identified six vari-

TABLE II. Polymorphisms Found in FGF8

Individual no(s). Location Sequence change Frequency

262 Intron 1a CCCCCGCGC (C/T) CCTC C:0.992, T:0.008a

1419 Exon 1c 18-base-pair duplication Duplication:0.00251002 Intron 1c GGCCACTCG (G/A) GCCC G:0.992, A:0.008a

1655, 1696 Intron 2 AGCAGGCAG (G/A) TGGG G:0.985, A:0.015a

1419 Intron 2 GGGGCC (G/A) CT CGCCC G:0.8, A:0.21766 Intron 2 GGGGCCGCT (C/T) GCCC C:0.992, T:0.008a

aFrequencies as determined in the affected population only.

Structure, Sequence, and Mapping Human FGF8 357

ants in our search of both normal individuals and the65 individuals with limb and craniofacial defects. Thesequences of these SSCP variants are shown in Figure2 and Table II. Five of them are C/G to T/A transitionsthat are seen commonly, and one is an 18-base-pair

duplication in the alternatively spliced exon 1c. SSCPanalysis for a control population showed that one was acommon polymorphic variant, due to a single nucleo-tide change in intron 2, with allele frequencies .2/.8 ina Caucasian population of 100 normal individuals. A

Fig. 2.

358 Yoshiura et al.

second polymorphism was discovered in individual1766, the third one in individual 1002, the fourth one inindividuals 1665 and 1696, and the fifth one in indi-vidual 262. All of these base substitutions occur in theintrons and do not seem to affect the splicing process.The sixth mutation was an 18-base-pair (six-amino-acid), in-frame duplication in exon 1c of the FGF8 gene.This was initially found in an affected father and son(propositus) with craniosynostosis type Philadelphia(Fig. 3). Analysis of additional relatives showed thatthe unaffected paternal grandmother, her unaffectedson, and several of her sibs had the same 18-base-pairinsertion. It has since been identified in an additionaltwo unrelated patients with an allele frequency of.0025 in a Caucasian population. This insertion appar-ently causes no phenotypic effect in these individualsand suggests that heterozygous abnormalities in thisexon are of no functional importance.

DISCUSSION

Recent studies show fibroblast growth factors (FGFs)and their receptors (FGFRs) are crucial in craniofacialmorphogenesis. Several classic craniosynostosis syn-dromes with or without limb abnormalities (Apert,Crouzon, Jackson-Weiss, and Pfeiffer syndromes) arecharacterized by mutations in FGFRs [for a review, seeMuenke and Schell, 1995]. In Apert syndrome, all mu-tations described to date affect two adjacent amino ac-ids within the putative ligand-binding domain ofFGFR2 [Park et al., 1995; Wilkie et al., 1995]. In con-trast, in patients with Pfeiffer syndrome not all havemutations in either FGFR1 or FGFR2. In particular,two kindreds with Pfeiffer syndrome that were linkedto 10q25-q26 do not have any demonstrated FGFR2mutation [Schell et al., 1995]. Thus, these cases wereconsidered to be candidates for a mutation in FGF8,which had been mapped to this region. In addition, one

Fig. 2. Sequence variants in FGF8. a: Individuals 1766 and 1434: polymorphism in intron 2. b: Individual 1002: Polymorphism in intron 1c. c:Individuals 1665 and 1696: polymorphism in intron 2. d: Individual 1419: insertion-deletion polymorphism in exon 1c.

Structure, Sequence, and Mapping Human FGF8 359

form of ectrodactyly, or split hand/foot malformation,which maps to 10q24-25 [Nunes et al., 1995; Gurrieri etal., 1996] could also be a candidate gene for an FGF8mutation. Other limb malformations, including spo-radic or syndromic syndactyly or polydactyly, mightalso be candidates for FGF8 mutations.

Recently, Cohn et al. [1995] showed that a morpho-logically normal limb bud is induced by implantedbeads soaked in Fgf1, Fgf2, and Fgf4. Only Fgf2, Fgf4,and Fgf8 are expressed in the AER (apical ectodermalridge) physiologically [Slack, 1995]. Crossley and Mar-tin [1995], however, showed Fgf8 mRNA is expressedin the ventral ectoderm of the early limb bud, beforethe AER is formed. Crossley et al. [1996] suggestedthat Fgf8 had a role in the induction, initiation, andmaintenance of chick limb development. Thus, FGF8may play an important role in outgrowth of limb budsin humans as well [Tabin, 1995].

We show here the FGF8 genomic sequence, includingpart of the intron sequences, primer sequences, andamplification conditions. This genomic sequence is alsoreported by Gemel et al. [1996]. Most of the introns are

small also, facilitating mutation screens that might in-volve alternate splice-site creation or destruction.These sequences are especially useful in creating prim-ers for mutation analysis in that FGF8 is not expressedin any adult tissue except in the gonads [Lorenzi et al.,1995]. Thus RT-PCR (reverse transcription and PCR)SSCP analysis would be difficult as patient specimenswould be hard to obtain.

Our extensive exclusionary data of mutations in apanel of individuals with craniosynostosis with or with-out limb defect makes it at least unlikely that muta-tions in the coding region of FGF8 are common causesfor such abnormalities. It is still possible that muta-tions, including microdeletions in the promoter and/orenhancer regions, cause these anomalies. For example,in the case of Campomelic dysplasia [Wanger et al.,1994], Rieger syndrome [Semina et al., 1996], and Holt-Oram syndrome [Yi Li et al., 1997], chromosomaltranslocations do not disrupt the coding regions of theetiologic gene. These results suggest that the enhancer/promoter region that controls the time, place, oramount of gene expression is at some distance from

Fig. 3. Transmission of insertion allele in the family. Insertion allele is transmitted from 1554 to 1584 and 1417.

360 Yoshiura et al.

coding sequence and is important for gene function. Avariety of expression patterns would support that thisFGF8 gene should continue to be a candidate consid-ered in at least some limb-defect disorders and thoseinvolving the central nervous system or kidney. Cer-tainly families on which linkage data support the lo-calization to the long arm of chromosome 10 shouldhave FGF8 considered as a candidate for their defects.The series of amplification primers for each of the cod-ing exons will be an extensively useful resource forthese screens.

Of significant interest is the fact that an in-frame,six-amino-acid insertion into one of the coding exonshas no apparent phenotypic effect in the heterozygousstate for individuals carrying this mutation. Whileexon 1c, which contained the 18-bp insertion, is alter-nately spliced, so making it possible that the isoformprotein containing exon 1c is not important for the nor-mal function of FGF8, this at least suggests that suchmutations do not result in any haploinsufficiency-typesyndromes; however, additional studies of the function-ality and expression of the various FGF8 exons wouldserve to suggest further appropriate clinical syndromesthat could be caused by deficiencies or defects in thisgene.

ACKNOWLEDGMENTS

We thank Nancy Newkirk for administrative assis-tance. This work was supported by NIH grants DE-08559 and DE-09170 (to J.C.M.) and HD28732 andHD29862 (to M.M.).

REFERENCES

Acland P, Dixon M, Peter G, Dickson C (1990): Subcellular fate of the Int-2oncoprotein is determined by choice of initiation codon. Nature 343:662–665.

Bugler B, Amalric F, Prats H (1991): Alternative initiation of translationdetermines cytoplasmic or nuclear localization of basic fibroblastgrowth factor. Mol Cell Biol 11:573–577.

Cohn MJ, Izpisua-Belmonte JC, Abud H, Heath JK, Tickle C (1995): Fi-broblast growth factors induce additional limb development from theflank of chick embryo. Cell 80:739–746.

Crossley PH, Martin GR (1995): The mouse Fgf8 gene encodes a family ofpolypeptides and is expressed in regions that direct outgrowth andpatterning in the developing embryo. Development 121:439–451.

Crossley PH, Minowanda G, MacArthur CA, Martin GR (1996): Roles forFGF8 in the induction, initiation, and maintenance of chick limb de-velopment. Cell 84:127–136.

Feldman B, Poueymirou W, Papaioannou VE, DeChiara TM, Goldfarb M(1995): Requirement of FGF-4 for postimplantation mouse develop-ment. Science 267:246–249.

Francke U (1994): Digitized and differentially shaded human chromosomeideograms for genomic applications. Cytogenet Cell Genet 65:206–219.

Gemel J, Gorry M, Ehrlich GD, MacArthur GA (1996): Structure and se-quence of human FGF8. Genomics 35:253–257.

Gurrieri F, Prinos P, Tackels D, Kilpatrick MW, Allanson J, Genuardi M,Vuckow A, Nanni L, Sangiorgi E, Garofalo G, Nunes ME, Neri G,Schwartz C, Tsipouras P (1996): A split hand-split foot (SHFM3) geneis located at 10q24-25. Am J Med Genet 62:427–436.

Hebert JM, Rosenquist T, Gotz J, Martin GR (1994): FGF5 as a regulatorof hair growth cycle: Evidence from targeted and spontaneous muta-tions. Cell 78:1017–1025.

Heikinheimo M, Lawshe A, Shackleford GM, Wilson DB, MacArthur CA(1994): Fgf-8 expression in the post-gastrulation mouse suggests role inthe development of the face, limbs and central nervous system. MechDev 48:129–138.

Jabs EW, Li X, Scott AF, Meyers G, Chen W, Eccles M, Mao J-i, CharnasLR, Jackson CE, Jaye M (1994): Jackson-Weiss and Crouzon syn-dromes are allelic with mutations in fibroblast growth factor receptor 2.Nat Genet 8:274–281.

Lichter P, Tang C-J, Call K, Hermanson G, Evans GA, Housman D, WardDC (1990): High-resolution mapping of human chromosome 11 by insitu hybridization with cosmid clones. Science 247:64–69.

Lorenzi MV, Long JE, Miki T, Aaronson SA (1995): Expression cloning,developmental expression and chromosomal localization of fibroblastgrowth factor-8. Oncogene 10:2051–2055.

MacArthur CA, Lawshe A, Shankar DB, Heikinheimo M, Shackleford GM(1995a): FGF-8 isoforms differ in NIH3T3 cell transforming potential.Cell Growth Differ 6:817–825.

MacArthur CA, Lawshe A, Xu J, Santos-Ocampo S, Heikinheimo M, Chel-laiah AT, Ornitz DM (1995b): Fgf-8 isoforms activate receptor spliceforms that are expressed in mesenchymal regions of mouse develop-ment. Development 121:3603–3613.

MacArthur CA, Shankar DB, Shackleford GM (1995c): Fgf-8, activated byproviral insertion, cooperates with the Wnt-1 transgene in murinemammary tumorigenesis. J Virol 69:2501–2507.

Mansour SL, Goddard JM, Capecchi MR (1993): Mice homozygous for atargeted disruption of proto-oncogene int-2 have developmental defectsin the tail and inner ear. Development 117:13–28.

Mattei M-G, deLapeyriere O, Bresnick J, Dickson C, Birnbaum D, MasonI (1995): Mouse Fgf7 (fibroblast growth factor 7) and Fgf8 (fibroblastgrowth factor 8) genes map to chromosomes 2 and 19 respectively.Mamm Genome 6:196–197.

Muenke M, Schell U, Hehr A, Robin NH, Losken HW, Schinzel A, PulleynLJ, Rutland P, Reardon W, Malcolm S, Winter RM (1994): A commonmutation in the fibroblast growth factor receptor 1 gene in Pfeiffersyndrome. Nat Genet 8:269–274.

Muenke M, Schell U (1995): Fibroblast-growth-factor receptor mutationsin human skeletal disorders. Trends Genet 18:308–313.

Murray JC, Buetow KH, Weber JL, Ludwigsen S, Sherpbier-Heddema T,Manion F, Quillen J, Sheffield VC, Sunden S, Duyk GM, WeissenbachJ, Gyapay G, Dib C, Morrissette J, Lathrop GM, Vignal A, White R,Matsunami N, Gerken S, Melis R, Albertsen H, Plaetke R, Odelberg S,Ward D, Dausset J, Cohen D, Cann H (1994): A human comprehensivelinkage map with centimorgan density. Science 265:2049–2054.

Nunes ME, Schutt G, Kapur RP, Luthardt F, Kukolich M, Byers P, EvansJP (1995): A second autosomal split hand/split foot locus maps to chro-mosome 10q24-25. Hum Mol Genet 4:2165–2170.

Ohuchi H, Nakagawa T, Yamamoto A, Araga A, Ohata T, Ishimaru Y,Yoshioka H, Kuwana T, Nohno T, Yamasaki M, Itoh N, Nogi S (1997):The mesenchymal factor, FGF10, initiates and maintains the out-growth of the chick limb bud through interaction with FGF8, an apicalectodermal factor. Development 124:2235–2244.

Park WJ, Meyers GA, Li X, Theda C, Day D, Orlow SJ, Jones MC, JabesEW (1995): Novel FGFR2 mutations in Crouzon and Jackson-Weisssyndromes show allelic heterogeneity and phenotypic variability. HumMol Genet 4:1229–1233.

Pruessner J, Ranade K, Dracopoli N, Schutte B (1995): High-efficiencycloning of PCR products with the XcmI T-vector pkRX. Am J HumGenet 57(S):1556.

Reardon W, Winter RM, Rutland P, Pulleyn LJ, Jones BM, Malcolm S(1994): Mutations in the fibroblast growth factor receptor 2 gene causeCrouzon syndrome. Nat Genet 8:98–103.

Ried T, Baldini A, Rand TC, Ward DC (1992): Simultaneous visualizationof seven different DNA probes by in situ hybridization using combina-torial fluorescence and digital imaging microscopy. Proc Natl Acad SciUSA 89:1388–1392.

Robin NH, Segel BL, Carpenter G, Muenke M (1996): Craniosynostosis,Philadelphia type: A new autosomal dominant syndrome with sagittalcraniosynostosis and syndactyly of the fingers and toes. Am J MedGenet 62:184–192.

Rutland P, Pulleyn LJ, Reardon W, Baraitser M, Hayward R, Jones B,Malcolm S, Winter RM, Oldridge M, Slaney SF, Poole MD, Wilkie AOM(1995): Identical mutations in the FGFR2 gene cause both Pfeiffer andCrouzon syndrome phenotypes. Nat Genet 9:173–176.

Schell U, Hehr A, Feldman GJ, Robin NH, Zackai EH, de Die-Smulders C,Viskochil DH, Stewart JM, Wolff G, Ohashi H, Price RA, Cohen MM,Muenke M (1995): Mutations in FGFR1 and FGFR2 cause familial andsporadic Pfeiffer syndrome. Hum Mol Genet 4:323–328.

Structure, Sequence, and Mapping Human FGF8 361

Schlessinger J, Lax I, Lemmon M (1995): Regulation of growth factor ac-tivation proteoglycans: What is the role of the low affinity receptors?Cell 83:357–360.

Semina EV, Reiter R, Leysens NJ, Alward WLM, Small KW, Datson NA,Siegel-Bartelt J, Bierke-Nelson D, Bitoun P, Zabel BU, Carey JC, Mur-ray JC (1996): Cloning and characterization of a novel bicoid-relatedhomeobox transcription factor gene, RIEG, involved in Rieger syn-drome. Nature Genet 14:392–399.

Slack JMW (1995): Growth factor lends a hand. Nature 374:217–218.

Tabin C (1995): The initiation of the limb bud: Growth factor, Hox genes,and retinoids. Cell 80:671–674.

Tanaka A, Miyamoto K, Nimamino N, Takeda M, Sato B, Matsuo H, Mat-sumoto K (1992): Cloning and characterization of an androgen-inducedgrowth factor essential for the androgen-dependent growth of mousemammary carcinoma cells. Proc Natl Acad Sci USA 89:8928–8932.

Wanger T, Wirth J, Meyer J, Zabel B, Held M, Zimmer J, Pasantes J,Dagna Bricarelli F, Keutel J, Hustert E, Wolf U, Tommerup N,Schempp W, Scherer G (1994): Autosomal sex reversal and campomelic

dysplasia are caused by mutations in and around the SRY-related geneSox9. Cell 79:1111–1120.

White RA, Dowler LL, Angeloni SV, Pasztor LM, MacArthur CA (1995):Assignment of FGF8 to human chromosome 10q25-26: Mutations inFGF8 may be responsible for some types of acrocephalosyndactylylinked to this region. Genomics 30:109–111.

Wilkie AOM, Slaney SF, Oldridge M, Poole MD, Ashworth GJ, Hockley AD,Hayward RD, David DJ, Pulleyn LJ, Ritland P, Malcolm S, Winter RM,Reardon W (1995): Apert syndrome results from localized mutations ofFGFR2 and is allelic with Crouzon syndrome. Nature Genet 9:165–172.

Yoshiura K, Tamura T, Hong H-S, Ohta T, Soejima H, Kishino T, Jinno Y,Niikawa N (1993): Mapping of the bone morphogenetic protein 1 gene(BMP1) to 8p21: Removal of BMP1 from candidacy for the bone disor-der in Langer-Giedion syndrome. Cytogenet Cell Genet 64:208–209.

Yi Li Q, Newbury-Ecob RA, Terrett JA, Wilson DI, Curtis ARJ, Ho Yi C,Gebuhr T, Bullen PJ, Robson SC, Strachan T, Bonnet D, Lyonnet S,Young ID, Raeburn A, Buckler AJ, Law DJ, Brook JD (1997): Holt-Oram syndrome is caused by mutation in TBX5, a member of theBrachyury (T) gene family. Nat Genet 15:21–29.

362 Yoshiura et al.