Dissecting the Developmental and Molecular Defect in Split Handl Split Foot Malformation

Michael A. Crackower

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Molecular and Medical Genetics University of Toronto

O Copyright by Michael A. Crackoser, 1999 NationaI Library Bibliothèque nationale If1 of Canada du Canada Acquisitions and Acquisitions et Bibliographie Services services bibliographiques

395 Wellington Street 395, nie Wellington OttawaON K1AON4 Onawa ON KIA ON4 Canada Canada

The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sell reproduire, prêter, distribuer ou copies of ththesis in microfonn, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/film, de reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fkom it Ni la thèse ni des extraits substantiels may be p~tedor otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation. Thesis Abstract

Dissecting the Developrnental and Molecular Defect in Split Hand/ Split Foot Malformation

Doctoral thesis of Michael A. Crrickower

Depanment of Molecular and Medical Genrtics

University of Toronto

Split hand/ split foot malformation is a hurnan drreloprnental lirnb drfect that affects the central rays of the hands and fret, resulting in rnissing digits and claw-like extremities.

Cornmonly scgegatins as an autosomal dominant disorder. this genetically heterogeneous

disrase displays a number of distinctive genctic fcatures including reduced penetrance.

variable expressivity. and segrqation distonion. where affectcd males transmit the diseasc

to thcir male olfspring in cxcrss of 50 percent.

.A SHFM locus (SHFM 1) has bcen mapped to humrin chromosome 7 on the basis of a large

number of patients with visible cytogenetic narranpernents. A physical map spanning this

region has been generated and a 1 Mb SHFht I critical interval has been detined at 7qZ 1.3- q17.1.

This thesis describes the identification and charactrrization of five genes from the SHFM 1

cntical intend. DSSI,the first gene identified, encodes a putative novel acidic protein and

has a dcvelopmental expression profile consistent with a rolr in the disease process.

DNCII. a cytoplasrnic dynein gene. is liksly not involved in the basic defecr but it may be

implicated in the non-Mendelian segegation obsrrved in this disease. Two homeobox

transcription factors. DLX5 and DLX6. are rxpressed during limb development and the

known funciion of the related gene in Drusoplriln further implicates them in limb development. Lastly. DSS2. a putative mitochondrial ATP transporter. is expressed in a specific pattern during limb development highly suggestive of role in the disease process.

What is more important, DSS2 is the only gene identified that is directly interrupted by disease associated balanced translocation breakpoints, making it an interesting candidate.

In addition. a phenotypic analysis of a mouse mode1 for SHFM. called Dacrylapkzsia. was undertrikm to funher understand the disease process. !t was determined that the phenotypc rcsulrs from the failurc to maintain the central portion of the apical ectodermal ridge, a key signaling centre in the drveloping limb bud. It is funher show that this tissue is lost duc to a lack of cellular prolifcration.

The data presented in this thesis provide a lramework from which to bcgin to understand the dsvclopmental and molecular defect in SHFM. Acknowledgments I am forever grriteful for the guidance that my supervisor, Dr. Lap-Chee Tsui, has given me throughout the time in his laboratory. He has always provided support. encouragement and understanding, while at the same time allowing me the freedom to explore my ideas and dçvelop my research. This, 1 feel, has had the greatest impact on rny growth as a scientist.

1 wouid like to thank the mernbers of my supervisory committce. Drs. Petrr Ray and Janet

Rossant for their intellectuai contribution to this project and the interest they havc show in my tùrure endeavours.

1 ihank Dr. Stephen Schercr. who has aiways played an integral part in my work and brought mc into the projecr that hr had initiated. Steve has always playcd the unofficial role of a CO-supervisorand as such has providcd me with valuable guidance and support. 1

have bcen foriunate to have benetited from the knowledgs and experience of two great

scientists.

I would likc to express my sincere appreciation to Dr. Chi-chung Hui. C.C. has provided

me wirh invaluable and never cnding assistance in rnuch of my work. His support has

rnablrd me to study in the field of developmental biology. I will always cherish the great

number of informa1 discussions we have had about my work and I am extremely thankful

for the friendship C.C. has shown to me. 1 attribute much of my success to him.

A special thanks to Iohanna Rornmens. Jacques Michaud. Marie-Andree Akirnenko. David

Sinasnc. Jun kIotoyama. and the entire Tsui Lab for their contribution to this projrct borh

scirntitically and ernotiondly. 1 would nlso like to thank rny farnily for their love, encouragement. patience and interest they have provided durinp my time as a graduate student. They have always been a key component of rny happiness and success.

Lasrly. I would liks to express my undying appreciation to my wife Sue Anne and my son

Broderick. They have shown geai patience. love. and support and have iillowrd me to kcep my goals in sight and rny priorities in check. Without them. this work would have little mraning. Dedication

This ihesis is dedicated to rny wife Sur Anne and my son Brodenck

and

to the rnernory of Dr. Linx Kilius.

Chapter 1 General Introduction and Thesis Outline ...... 1

1 Introduction ...... 9

1.B Split Handl Split Foot bIalFormation ...... 4 I.B. 1 SHFM Phenotypç ...... 5 1.B.L Genetics of SHFM ...... 8 1.8.3 Syndromic Ectrodactylies ...... 11 I.B.4 SHFM loci ...... 14

1.C Animal Models of SHFM ...... -21 l.D Embryology of the Vertebrate Limb ...... 73 I.D. 1 Venebrate limb induction ...... -24 I.D.2 Pattcming of ihr early venebrare limb ...... 15 I.D.3 Patterning of the limb skelcron ...... 35 1.E Pathogenesis of SHFM ...... 35 1.F Identification of Candidate Genes in a Positional Cloning Strategy ...... 38

1.G Thesis Overview ...... 41

Chapter II Further Refinement of the SHFMl Locus and Analysis of a Candidate Gene for its Expression During Development ...... -M

II Introduction ...... 45

II Materials and Methods ...... 46 1I.B. 1 FISH mapping of chromosome rearrangements in SHFM and SE patients ...... 46 II.B.3 Gene identification ...... -46 II.B.3 cDNA clones and libraries ...... 47 1I.B.4 Grne structure analysis...... -47 II.B.5 RNA in sin< hybndization and Fin Amputation ...... -18 II.B.6 Subcellular Locdization of DSS 1 ...... 45 1I.C Results ...... 49 1I.C. 1 The minimal critical region for SHFM 1 ...... 49

vii Identification of genes in the SHFM 1 critical region ...... 50 Characterization of DSSl and its rnurine homolog (Dssl)...... 51 Expression of Dssl during development ...... 60 Localized Dssl expression in developing limhs ...... 60 Cloning and analysis of Dssl from zebrafish...... 63 Subcetlular localization of Dss 1 ...... 64

Discussion ...... -68

Chapter 111 Cloning and characterization of two cytoplasmic dynein intermediate chain genes in mouse and human ...... 71 1II.A Introduction ...... -72

1II.B Materials and Methods ...... 73 1II.B. 1 Gene Identification. cDNA Clones and Libraries ...... ,...... 74 III.B.2 Gcnornic Structure of DhrCI 1 ...... 75 III.B.3 Northern blot analysis ...... 75 III.B.4 RT-PCR...... -75 III.B.5 RNA in siru hybridization ...... 76

1II.C Results ...... -77 1II.C. 1 Cloning of human DNCI 1 ...... 77 III.C.2 Isolation of murine D~tciland Dnci2 ...... 78 III.C.3 Transcript analysis of hcil and Dnci? ...... 53 111.C.4 Analysis of expression of D12cil and DnciZ during mouse embryogenesis ...... 88 1II.D Discussion ...... 93

Chapter IV Phenotypic Characterization of the Dactylaplasia Mouse. a Model for Human SHFM ...... 98

1V.A Introduction ...... -99

1V.B Materials and Methods ...... 101 1V.B. 1 Mice and Genotyping ...... 101 IV.B.2 Skeletal analysis ...... 101 IV.B.3 RNA in siru hybridization analysis ...... 101 IV.B.4 Scanning electron microscopy ...... 102 IV.B.5 BrdU labeling of embryos ...... 103 IV.B.6 TUNEL Staining of Apoptotic Cells ...... 103 1 .C Results ...... 103 1V.C. 1 Loss of anterior and central bones of the autopod in Dac mice ...... 103 IV.C.2 Onset of the Dac phenotype prior to chondrogenesis in the autopod ...... 107 IV.C.3 Degeneration of the AER in Dac limb buds ...... 110 IV.C.4 Loss of expression of AER markers in mutant limb buds at E 10.5 ...... 110 IV.C.5 Normal mesenchymal gene expression in Doc limb buds ...... I14 IV.C.6 Normal prograrnrned ce11 death in Dac limb buds ...... 115 IV.C.7 Lack of ce11 proliferation in Dac limb buds ...... 117 1V.D Discussion ...... 120 1V.D. 1 Ce11 death and ce11 prolifention in Dac mice and AER maintenance ...... 120 IV.D.2 AER maintenance and the ZPA ...... 122 EV.D.3 Dac as a mode1 for AER maintenance ...... 123 IV.D.4 Candidates for the AER maintenance hctor(s)...... 126 IV.D.5 Dac mice and human split hand/ split foot malformation ...... 127

Chapter V Assessrnent of DLX5. DLX6. and DSSZ as candidate genes for SHFMl ...... 12s VA Introduction ...... 129 V.B &laterials and Methods ...... 130 V.B.1 Gene Identification. cDNA Clones and Libraries ...... 130 V.B.7 RNA il1 riru hybridization ...... 131 V.C Results ...... 131 V.C. 1 Identification of DLY5. DLY6. and DSS2 from the SHFM 1 criticri1 intérval .... 131 V.C.2 Identification of murinc DM. Dl.r6. and DssZ ...... 135 V.C.3 Analysis of the expression of DLyj and Dls6 during mouse limb bud development ...... 136 V.C.4 Antilysis of the expression of Dss2 during mouse embryogenesis ...... 136

V.D Discussion ...... 141

Chapter VI Discussion and Future Directions ...... iu VIA Possible Molecular Etiology of SHFN1...... 145

VI.B The Role of DSS2 in SHFMl ...... 147 V1.C The Biochemical Function of DSSI ...... 149 V1.D Position Effect in SHFbIl? ...... 150 V1.E Generation of blouse Models for SHFM1...... 152 V1.F Other SHFM Loci ...... 153 V1.G Further characterization of the Dac mouse ...... 155

References ...... 158 List of Tables

Table 1- 1 . Clinical and genetic features of the syndrornic ectrodactylies ...... 13

Table 1.2 . Characteristics of SHFM Loci ...... 18 List of Figures Figure I- 1 . The SHFM Phenotype ...... 7

Figure 1-2 . Physical map of the SHFM 1-critical region ...... 15

Figure 1.3 . Development of the Vertebrate Limb ...... 31

Figure 11- 1 . The location of genes in the SHFM 1 critical rezion ...... Figure 11.2 . Nuc lcoiidc and amino acid scquence of human and mouse DSS 1 ...... -33

Figure 11.3 . Evolutionary conservation of DSS I...... 56

Figure 114. hnalysis of the human DSS 1 tninscript...... 57

Figure 11.5 . Genomic structure of human DSSl ...... 58

Figure 11.6 . Expression of Dssl during mouse embryogenesis...... -61

Fisure 11.7 . Whole mounr RNA in sitir hybridizaiion of Dssl during mous<:limb bud morphogenesis ...... 65 Fiyrc 11.8 . Analysis of Dssl expression in the dwrlopin_pand regenerating Zebrafish ...... -66

Figure 11.9 . Subcellular localization of DSS 1 ...... 67

Figure III- I . The localizaiion of DNCII within rhc SHFhl 1 critical region ...... 79 Figure 111.2 . Nuclcotide and amino acid sequençc of human DNCZI ...... Y1 Fisure 111.3 . Comparative spccirs analysis of alternative splicing in DNCII and protein sequence alignment of mouse Dncil and Ditci2 ...... 81 Figure III4. RT-PCR analysis of Dncilin mice and Nonhrm blot analysis of murine Dncil and Dnci2 ...... 86

Figure 111.5 . Expression of Dncil and DriciZ from E 10.5 -E 1 1.5 of mouse development ...... 90

Fisure 111.6 . Analysis of expression of murine Dticil and DnciZ at E13.5 ...... 91 Fisure 111.7 . Expression of Dnci2 in the developing mouse limb ...... 93

Figure IV- 1 . Skelrtal analysis of Dac lirnb phenotype ...... 105

si Figure IV.2 . Whole mount alcian blue staining of E 12.5 and E 13.5 Duc limbs ...... 108

Figure IV.3 . Scanning electron microscopy of Dm limbs...... 112

Figure IV4 Analysis. of. expression of AER marken using whole mount in sim hybridization ...... 113

Figure IV3. Analysis of AP patterning and FgflO expression in Dac limb buds ...... 116

Figure IV.6 . .A nalysis of BrdU incorporation and prognmmed ce11 death in Drzc limb buds ...... 1 18

Fisure IV.7 . Model for AER maintenance in the mammalinn limb bud ...... 174

Figure V- 1 . Amino acid sequence and evolutinary conservation of DSS2...... 133

Figure V.2 . DSS7 motif structure and hydropathy plot ...... 134

Figure V.3 . The expression of DLrj and DLr6 during rnouse limb development ...... 138

Figure V.4 . D.5.d expression in the mouse ernbryo from EIO.j.El 1.5...... 139

Fisure V.5 . DssZ expression in the rnouse ernbryo at E 13.5...... 1 40

xii Ab breviations

AER apical ectodermal ridge AERMF AER maintenance factor

ANT adenine nucleotide transporter

AP anterior-posterior bp base pairs B tdU Bromo-deoxyuridine cDNA complimentq DNA

CM cen timorgans

CMV cyiornegalovims CNS central nervous sy stem cku. D

DNA deoxynbonucleic acid

DNCI cytoplasmic dynein intermediate chain DRG dorsal root _eanglion

DSS I human "deleted in split band split foot malfomtion" gene 1

Dss l mouse DSS 1 homologue (onholog) DSS2 human "dcleted in split handf split foot malformation" gene 2 Dss2 mouse DSSZ homologue

DV dorsal-ven trai E embryonic day (e.g.. E 1 OS) ECP ectrodactyly. cleft lip/palate syndrome EE rctrodacty ly. ectodermal dysplasia syndrome

EEC ectrodactyly. ectodermal dysplasia cleft lip/palate syndrome EST expressed sequence tag FGF fibroblast zrowth factor FGFR fibroblast growth factor receptor

FISH fluorescence in situ hybridization FITC fluorescei n isothiocyanate

GCPS Greig's cephalopolysy ndactyly HFG hand foot genital syndrome

11 11 hedgeho? hr hours

Kb kilobase pairs kDa kilo Daltons

MCF mitochondrial camer family nrDac modifier of Dm

MM# "bIendelian Inheritance in Man" reference number mRNA rnessenger RNA

NES nuclear expon sequencr

NLS nuclex localization sequence nM nanornolar nt nucleotids

PAC Pi rtrtificiril chromosome PCD prognmmçd ce11 death

PCR polyrneme chain reaction

PD proximal-distal

PFGE pulse field gel electrophoresis

PNS peripheral nervous system

PZ prosress zone Eu retinoic acid RNA ribonucleic acid

S'RACE 5' npid amplification of cDNA ends

xiv RT reverse transcriptase

SD segregation distoner SDS sodium dodecyl sulphate

SE sy ndromic ecuodacty ly SEM scanning electroo microscopy SHFM split handJsplir foot malformation

SHFMI split handlspiit foot malformation locus 1 (chromosome 7)

SHFM2 split handlsplit loor mal formation locus 2 ( X chromosome) SHFM3 sptit handlsplit foot malformation locus 3 (chromosome 10)

Shh Sonic hsdgehog sni sydic~lisnirnouse

SPD synpolydaccyly

SSC codium citrate/ sodium chloridc UTR untnnslated region

YAC yrast anificial chromosome

ZPA zone of polarizing activity Chapter 1

General Introduction and Thesis Outline LA Introduction

Descriptions of hurnan limb malformations are as old as recorded rnedical history. The earliest record of an inhet-ited limb malformation is likely found in the Old Testament,

where polydactyly in Goliath and his family is described (7 Samuel xxi, 20). Aristotle clearly was oware of hand and foot deficiencies in humans (cited by Warkany. 1959). but it

was in the 16th century that the first clenr account of several varieties of hand

mal formations was presented by Ambroise Paré ( 1575).

Human limb malformations arc relativsly common. Limb deficiency defects. which

comprise a signifiçant goup of malformations. have an rstimatsd incidence of greater than

I in 2000 live binhs (Evans et 01.. 1991; Buss. 1994). Historically. rcsearchers had placed

little effort on studying limb defccts of senetic origin. since it was believed. and is still

ncnerally truc. that many conipita! limb abnormalities havc non-genctic etiologies. It is. t

in pan. for this reason thrit the majority of genes currently known to be involved in limb

devclopment have been identified from studies in mode1 organisms. as will be discussed in

dctliil bclow. Nevcnheless. with the advancement of the human genome project and the

increasing spced at which positional clonin? of diseüse senes is progressing. there has been

a rcipid expansion of the numbcr of genes identitied that are responsible for human Iimb

abnorrnalities (reviewed in Innis and Monlock. 1998).

The Tint causativr gene identified in a human limb malformation was the zinc finger protein

GLI3 which was shown to br interrupted by translocations in patients with Greig's

cephalopolysyndactyly (Vortkamp et al.. 199 1 ). Rrcently. mutations in CL13 have been

found in two additional human syndromes (reviewed in Biesecker. 1997). [t has also been

show that the hornologous genr in mouse (Gli3)is panially deleted in the exmames mouse

mutant (Hui and Joyner. 1993). GLI3 is a downstream mediator of Sonic Hedgehog (SHH) signaling. a vertebrate homologue of the Drosophila segment polarity gene Hedgehog (for reviews of the SHH pathway see Johnson and Tabin. 1997; Tabin and McMahon. 1997). As such. other members of the SHH pathway have also been shown to play a role in human limb defects. Mutations in Patched (PTC), the putative receptor for

SHH (Marigo es al., 1996a: Stone er al.. 1996). are found in basal ceIl nevus syndrome

(also known as Gorlin syndrome: McKusick. No. 109400). which is characterized by increased neoplasia. dong with polydactyly and syndactyly (Johnson. er al.. 1996; Hahn er al.. 1996). In addition. mutations in the 7-cholesterol reductase sene cause polydactyly found in Smith-Lemli-Opitz syndrome (McKusick No. 770400: Wassif er al.. 1998). This phenotype is likely due to the aberration of SHH signaling since cholesterol metabolism has been shown ro play a role in this signaling pathway (Porter er cil.. 1996).

Rrcently. the senes responsible For a group of skrletal dysplasias characterized by craniosynostosis and associated limb abnorrnalities have been determined (reviewed in

Webster and Donoghue. 1997). In these syndromes gain of function mutations have been

identified in members of the fibroblast growth factor receptor gene family (FGFR). Both FGFRI and FGFRl mutations have been found in patients with Pfeiffer syndrome. the

characteristics of which include broad. rnedially deviated great toes and thumbs. FGFR2

mutations have ülso been associated with syndactyly found in jackson-Weiss and Apen

syndromes. These syndromes represent an interesting case where the same mutation in the

same pene causes more than one clinically distinct syndromes.

Other genes of note found to be mutated in human limb malformations include HOX413.

and HOXDI3. which are responsible for Hand-foot-genital (HFG) syndrome. and Synopolydactyly (SPD). respectively (Akarsu er al.. 1996; Muragaki es al.. 1996: Monlock

and Innis. 1997). The rnouse mutant Hypoducoly (Hd)has been shown to represent a

mode1 for HFG ris it too has mutation in Ho.ral3 sirnilar to those found in the human disease (Monlock et ai., 1996), while HoxdI3 transgenic rnouse models have been shown to mimic the phenotype found in SPD (Zakiny et al.. 1996).

To date. the genes involved in approximately 25 human limb malformation syndromes have been identified (Innis and Monlock. 1998). and that number is surely to nse exponentially within the next few years. A goal of the research presented hm. is to identify the gene(s) responsible for the limb phenotype in human split hanci/ split foot malformation (SHFM) at chromosome 7q7 1.3-qll.1.

Pnor to the onset of this work. positional clonin_ostrategies had been initiated to define the critical interval of this locus and isolate potential transcribed sequençes from this region

(Schcrcr et d.,1994a; 1994b). This thesis describes the work that has been carried out to fully clone and characterize these transcribed scquences. and to evaluate their potential involvement in the disease. To funher undersrand the disease process and thus retïnr the candidate senr criteria. the phenotype of a mousr modd for SHFM has been characterized and the developmental defect that gives nse to this disorder has been dctermined. The data presented here leads to the better undentanding of the drvelopmentai and molecular defect in SHFM.

1.B Sdit Hand/ S~litFoot Malformation

Split handl split foot malformation (SHFM. McKusick 183600). also known as ectrodactyly or lobster claw deformity. is a human developmental Iirnb defect that affects the centml rays of the hands and fert resulting in missing digits and claw like extremities

(see Fig. 1-1). The designation of split hand (Spalthmd) was first suggested by Meller

( 1593) and has since been used extensively in the literature. Long before the term split hand was adopted, however. Saint-Hilaire (1832) described this hand deformity as ectrodactyly and to this day the two terms are used synonymously. True ectrodactyly, however, implies an absence of digits and does not involve fusion, or a median cleft as is obsemed in split hanci/ split foot malformation.

I.B. I SHFM Phenotype

The earlisst report of SHFM was presented by Par6 in 1575 in his description of n 9-year old boy with right split hmd and absence of the ions bones of the legs. The report of "two finzered Negroes" living in Dutch Guinea described by Hansnick ( 17 16- 1779). provided the first insight into the hereditary transmission of the disordrr (in Barsky. 1963). In 1937

Lange describrd an atypical form of split hand drlormity. while others established a clrissit'ication for SHFM as follows (Lewis and Embelton. 1908; Birch-Jensen. 1949):

Typical SHFM

Type 1: This represcnts the classical form of the disordrr characterized by the lobster claw

appearance. where there is a cone shaprd cleft dividing the hand into two parts. which

corne together like a lobster's claw. Typically syndactyly (fusion) of the remaining fingers

on both sides of the cleft is observed. There is variability in the type 1 classification

ranging from missing digit 3 in its mildest fom to a greatrr loss of skeletal elements both

latenlly and proximally.

Type 2: The most severe fom of this disorder chancterized by monodaciyly with only

digit 5 present. Of course no cleft is present.

These anatomic classifications are valuable in describing the phenotype. but they clearly do

not represent distinct genetic abnormalities. Most commonly. gradations between the 2 types are observed (Maisels, 1970). Interestingly, the monodactyly found in type 2 SHFM closely resembles the homozygous phenotype found in the mouse mutant Dacvlaplasia

(Dac) which is a mode1 for the human condition (See below).

Atypical SHFh.1

The ntypical phenotype as proposcd by Lange ( 1937) has similanties to type 1 SHFM as it too is characterized by absent central rays. In addition. however. in atypical SHFM rrmaining digits are hypoplastic. and the cleft is sornetirnes çovered by a skin bridge. to which rudimentary digits may be attached. Frrqucntly seen with atypical SHFM is unilateral involvement of the hands and no involvement of the ket. This tom of SHFM has an incidence of approximately 1: 150000 (Birch-Jensen. 1919) and rrirly investigators believcd this not to be an inhrrited trait (Temtamy and iMcKusick. 1978).

SHFhI with associated limb abnormalities The variability found in SHFM cm also be exacerbated by the presence of other digital and limb malformations. Cases of associated polydactyly. triphalangeal thumb. camptodactyly. clinodricty ly. absent radius. absent ulna. absent humerus and ulna have al1 beçn described

( reviwed in Temtamy and McKusick. 1975). Recently. Zlotogora ( 1994) suggested an alternative classification for SHFM into distinct genetic types based partly on the association with additional limb malformations. Based on a literature review, it was suggested that type I SHFM is a fully penetrant disorder. with no long bone deficiency

whilr type 2 shows reduced penetrance (discussed later) and frcquently associated long

bone deficiency. Type 3 would represent a defect at one locus called SHFM 1 (see below).

while type 1 is associated with a defect at a second locus SHFM3 (see below). WhiIe it

appears that there is some consistrncy with this hypothesis. is has been argued that this classification is ovrrsimplified since there is clear genetic evidence that not al1 cases of type

2 SHFM represent a defeci at the SHFM 1 locus (Evans et al.. 1995). Figure 1-1. The SHFM Phenotype. Hands and feet of mother (A.C) and daughter (B,D)with typical SHFM. (A) Mother has a relatively severe hand presentation with only digits 1 and 5 on the left hand. Right hand shows splaying of the remaining digits. (C) The mother's feet display the classical type 1 typical SHFM with a lobster claw appearance. (B) The daughter's hands are only very mildly affected with clinodactyly of the 2nd and 3rd digits (D) The daughter's feet are affected similar to her rnother's. These patients represent a good example of the variable expressivity that can be found between related individuals and between different limbs of the sarne individual. (Photos obtained from Dr. J.P Evans) I.B.2 Genetics of SHFM

SHFM has been reported to have an incidence of 1:90000 (Birch-Jensen, 1949). however. more recent studies suggest that the incidence is rnuch closer to 1:10000 (Froster and

Baird. 1989: Calzolan et ul.. 1990: Czeizel er aL. 1993). While this disorder is frequently found as sporadic cases. familial forms ty pically show an autosornal dominant pattern of inheritance (Lewis and Embleton, 1908: Trmtamy and McKusick. 1978). Nevertheless. clear examples exist for X-linked (Ahmad et al.. 1987: Faiyaz ul Haqur rr d.19 93) and autosornal recessive (Friere Maia. 197 1: Verma et al.. 1976; Zlotogon and Nubani, 1988) toms of the diseass.

A hallmark feature of this disorder is the pronounced variable expressivity (Maisels. 1970).

As mentioned above. the severity of the phrnotype can range frorn a mild type I CO type ?

(monodactyly). In addition. one can obsenle a defect in a single limb ranging to al1 four limbs. These variations are frequently found cven betwern mernbers of the same farnily. and the variability in severity is frequently observed between different limbs of the same

individual (Trmtamy and McKusick. 1978). Thus. it would be expected that this variable rxpressivity is an epigenctic phrnomrnon.

Reduced penetrance in SHFbI Reduced penetrance is also observed frequently in familial SHFM. It has been calculated

that. in 608 of families analyzed. 30% of obligate carriers have no discemible phenotype (Terntamy and McKusick. 1978). The reason for this dramatic reduced penetrance is

unknown. However. it has been postulated that a second gene at a separate locus may

protect obligatr carriers from devrloping this abnormaiity (David. 1972). Interestingly.

this mechanism has been clcarl y established in the Dac mouse mutant (refer to page 2 1).

An additional phenornenon that has been frequently observed in SHFM pedigrees is the appearance of the disease in more than one sibling born to unaffected parents (David,

1972). Originally. deiayed mutation was suggested as an explanation for this observation

(Auerbach, 1955/56). but this mode1 has not been supported (Vogel. 1957/58). A more acceptable hypothesis is that germinal mosaicism occurred in one of the parents of the affected sibs (David. 1972).

Segregation distortion in SHFhI

Owall. it is genenilly considrred that the sex ratio for SHFM is 1: 1. Howsvcr, it is clear that segregation distortion is present in numrrous SHFM families, where there is an excessive maIr to male ~ansrnission.The tirst reports of this phenomenon in SHFM were actually used to argue against the applicability of Mendelian genrtic principles to human genetics at the turn of ihe ccntury (Pearson. 1908; iLlcMullan and Pearson, 19 13). In 1960

Stevenson and Jennings analyzed segregation ratios in a number of families and round the segrcgation distortion to hold truc. Thsre was a bias in this study, however. since they includrd the original farnily in which this distortion was presented (Pearson. 1908). Subscquently. Jarvick et al. (1994) performed an unbiased analysis of recently identified

families. and conclusively showed thrit in these families there was an excess (greater than

508) of affectrd male offsprin~born io affected males. Interestingly. while it was shown

that affected males do not have a highrr proportion of male childrcn than affected fernales.

both affected male and fernale parents have more male offspring than unaffected parents. The observation of segregation distonion in SHFM is the first example of this in a human

developmen ta1 disorder.

The phenomenon of segregation distortion has been well characterized in Drosophilu,

where moles heterozygous for segrrgation disiorter (SD) second chromosome produce a

gross excess of SD-bcaring progeny. in some stocks excerding 99% (reviewed in Crow,

1991). In this case chromosomal inversions dcfine the locus, which serve to inhibit recombination (Hartl. 1975). This reduced recombination is necessary to maintain haplotypes of an enhancer locus that functions most effectively when in cir with the SD allele (Ganetsky. 1977; Brittnacher and Ganeisky. 1984). It has also been shown that this non-Mendelian transmission of the SD chromosome is related to sperm dysfunction (Han1 er al., 1967).

In mammals segregation distonion has been best characterized in the t-haplotypes of the mouse (reviewed in Silver, 1993). Located on mouse chromosome 17, at the T locus

(brachyury gene). t-haplotypes are defined by 4 non-overlapping chrornosomal inversions. which act to inhibit recornbination dong this region in heterozygous anirnals. Males heterozygous for the t-haplotype transmit the allele to Sreater than 99% of thrir offspring at the expense of the wild-type allelr. Males hornozygous for the t-haplotype are sterile.

Grnetic studies have suggrsted that. in rnice. this transmission ratio distonion is derived from thc combined effects of 1 to 5 gene products. referred to as distorters. that interact with a responder protein which is expressed late in spermatogenesis after the completion of rneiosis (Lyon. 1981). Al1 of these genes are found within the t-cornplex. A srnall nurnber of candidate grnes have bren identified from this locus. which are thought to represent some of the distorter loci (Lrtder et 01.. 1989: Pildsr et al.. 1993; Huw er al., 1995).

Recently. insight into the molecular mcchanisrn of segregation distortion in the t- haplotypes. has become available with the assignment of two of these genes Terex-1 and Tetex-2 (l-omplex ~stise~pressed) as subunits of the microtubule based molecular rnotor protein dynein (King et al.. 1996: Patel-King er al.. 1997). This has lead to a mechanistic mode1 of segregation distortion where differential incorporation of wild-type and mutant dyneins into the sperm axonemes lads to the depanure from Mendelian ratios (Outlinrd in

Patel-King er al., 1997). It has been suggested that the mutant t-haplotype responder protein would likely be membrane bound and thus the responder type (mutant or wild-type) present in a developing spermatid would be dictated by the genotype of that spermatid. The distorter proteins, however, would be freely diffusible and thus mutant and wild-type protein would be present in al1 spermatids. The role of the responder proiein would be to control what type of distorter protein could incorporate into the growing sperm tail. It foIlows that mutant responder would allow only wild-type axonemal dyneins to incorporate into the growing sperrn tail. Conversely, wild-type respondcr protein wou ld al low both wild-type and mutant dyneins into the sperm rail resulting in defective motility of wild type spermatids.

This mode1 does not exclude cytoplasmic dynsins frorn also encoding distoner proteins

(Tctrx-1 is a cytoplasrnic subunit) (King et

While SHFM is itself a distinct clinical disorder. the ectrodactyly phenotype present in

SHFM is frequently found to be one component of a more complex syndrome. These syndromes have been trou ped under the headi ne of s y ndromic ectrodacty lies (SE) (Table

1-1). The prototypical SE is EEC syndrome which. includes ectrodactyly, ectodermal dy splasia. cleft liplpalette ( McKusick No. 129900; Rodini and Ric hieri-Costa, 1990).

Ectodermal dysplasia is characterized by sparse opaque hair. absence of lacrimal ducts, keratitis. and tooth and nail abnormalities. Ectrodactyly. ectodermal dysplasia syndrome

(EE. McKusick No. 1298 10; Wallis. 1988). rcrrodacty ly-cleft palate syndrome (ECP,

McKusick No. 119830: Opitz et al.. 1980). and acro-dermato-ungual-lacrimal-tooth

syndrome (ADULT. iMcKusick No. 103235). which has been described only once (Propping and Zerres. 1993), are other examples of SE which rnay simply represent variations of EEC syndrome. Other SES include, ectrodactyly-hearin; loss syndrome (McKusick No. 120600: Raas-Rothschild et al.. 1989). Karsch-Neugebauer syndrome characterized by rctrodactyly with congenital nystagmus. fundal changes. and cataracts

(McKusick No. 183500; Pilarski er al.. 1985). Patterson-Stevenson-Fontaine syndrome

(McKusick No. 183700: Fontaine et al.. 1974) also referred to as split-foot deformiiy with

mandibulofacial dysostosis. and acral-renal-mandibular syndrome an unusual recessive

form (McKusick No. 200950: Hala1 er 01.. 1980). In genrral terms. it can be said that

these syndromes consist of craniofacial abnormalitiss. sctoderrnal dysplasias. genitourinary

abnormali tics. and ht.cirin_oloss.

In addition to these cases of SE. the rctrodactyly phenotype has also bcen observed in association with numerous other syndromes. cxamples of which include Smith Lemli Opitz

syndrome (de Jong er d..1998) and Wolf-Hirschhorn syndrome (Bnmshad rr ni.. 1998).

Reports of these. however. are exceedin_oly rare. and are in thcmsclves not clinically

de fined syndromes.

Similar to isolatcd SHFM the sy ndromic ecrrodactylies are typically found to be autosomal

dominant disorders that display variable expressivity and reducrd penetrance. Studies on

one locus for SHFM (that comprise much of this thesis and will be outlined below). clearly

show cases of isolatrd SHFM and syndromic rctrodactylies mapping to the same

chromosomal locus. Thus. it is generally believed that in some cases thrse disorders share

the same molecular etiology. The phenotypic variation rnay be a function of genetic

background or epigenetics.

I.B.4 SHFM loci

SHFM1: 7qZ1.3-q22.1 SHFM is a genetically hrterogeneous disorder. and to date a number of loci have been clearly drfined (Outlined in Table 1-2). The first of these loci is found on the long arm of chromosoms 7. This locus was first identikd becriuse of a consistent association between

SHFM and chromosomal deletions of bands 7q2 1 -q22 (Del Porto et al.. 1% 3; Pfeiffer.

1984; Tajarii et (11.. 1989; Rivera er a!.. 199 1 ; Roberts et d..199 1 ). The assi~nrnentof this locus. desipnated SHFMl (Tsui and Farrall. 1991). was further supported by the identification of numerous cytogenctically visible chromosomal translocation brcakpoints in the sarnç ara in SHFM patients (Sharland et ul.. 199 1 : Neri et al.. 1992; Naritorni et al..

1993: Scherer er al., 1 994b).

Prior to the strirt of this thesis. Scherer er al. (1994a) crirried out the first rnoleculrir characterization of this region. Somatic cell hybrid lines were establishrd from 3 SHFM patients with chromosomal deletions and from 3 patients with balanced translocations. as weli as five lines previously derivcd from unaffected individuals with cytogenetic abnormalitirs. Using. PCR and Southrrn hybridization a large number of DNA markers

were assigned to IO intervals defined by these chromosorna1 rearrangernrnts. Four DNA

markers were round within the region deleted in a single SHFM deletion patient. Two of

these markers were found to be distal to the 3 translocation breakpoints and two were

found to be proximal. thus placing these translocation breakpoints within the smallest

commonly dcleted region. Pulse field gel electrophoresis was funher employed to order

these four mnrkers with respect to each other and the translocation breakpoints. Figure 1-2. Physical map of the SHFMl-critical region. (A) The deletion breakpoints defining the SR0 and the translocation or inversion breakpoints locatrd within the S RO (represented by filled boxes) are shown. Patient T3 has 2 breakpoints (T2.1 and T2.2) located within and a third (T2.3) distal to the SRO, see Figure 4.6 for a more detailed e'rplanation. (B) YAC contig spanning the SHFM 1 SR0 region. Open boxes represent individual Y AC clones with the centromere of each YAC (if known) indicated by a circle and the position of rare-cutting restriction enzymes shown. (C) The location of gens and marker DNA fragments (and their corresponding phage or cosmid clones) and rare-cutting restriction enzymes are shown with their location interval representrd by a solid bar; the stippled boxes mark the location of presumed full-length cDNA clones. B=BssHIt. C=ClaI. M=MluI, N=NotI, Nr=NruI, S=SalI, Sf=SfiI, X=Xhot. (From Scherer et al.. 1994b) Subsequently Scherer et al. (1994b) further chancterized this region by using a previously genented YAC based genomic library to construct a physical map chat spanned this region.

Using fluorescence in sirid hybridization, somatic cell hybrids. and grnomic phage clones. a

1.5 Mb criticai interval for SHFMl was defined. The precise location of 6 disease associated chromosomal breakpoints were mapped within this region (Fig. 1-3).

Followin= the mapping of this locus, several methods of genr identification have been rmployed to isolate the genes chat map within this interval (outlined below). Much of the work presented in this thesis focuses on the characterization of these genes idcntified from this locus.

While al1 the patients andyzed in these mapping studies present with the rctrodactyly phenotype characteristic of SHFM. some of these patients were ciinically diagnosed with

SE (Scherer et al.. 199Jb). It was evident from the mapping of the chromosomal breakpoints in these SHFM patients that the position of SE breakpoints with respect to isolated SHFM breakpoints appeared to be randoni. This has allowed one to argue that there is likely a common rtiology of SHFM and SE in some cases (Scherer ct cd.. 199Jb).

SHFCI.12: X-linked inheritance.

The second definrd SHFM locus maps to Xq26-Xq36.1. A large Pakistani pedigree was described which clearly displayed a pattern of X-linked inheritance. where affected males strongly outnumbered fernales (33:3). These individuals displayed monodactylous or bidactylous hands. while the feet displayed the more typical lobster claw phenotype. Other females appeared either normal or displayed mild deformities of the hands andfor feet

(Ahmad et ni.. 1987). The sex ratio and clinical manifestation of the disease in this farnily lerid invrstigators to look for genetic linkage of this phenotype to polymorphic markers on the X chromosome. A Lod score of 5.13 at a recombination fraction of 0.00 contïrrned that the gene defective in this pedigree was located at Xq26-Xq26.1, close to the HPRT locus (Faiyaz ul Haque er ai., 1993). In addition. in the same study linkage to SHFMI locus on chromosome 7 was excludcd. As a result it has been suggested that there may be some similarity between genes on chromosome 7 and the X chromosome.

Additional autosomal dominant loci for SHFM

Conclusive rvidence rhat cm autosomal dominant locus for SHFM maps to chromosome 7 does not preclude t hc existence of additional autosomal loci. Accordingly. preliminary evidence for additional loci was provided by the clinical identification of a smail numbcr of patients with the rctrodactyly phenorype with visible cytogenetic rearrangements rnapping to chromosornal locations other than chromosome 7. A single large interstitial delrrion and a single balanced translocation associated with SHFM have been mapprd to 6q16.3-qE.3

(Braverman er ai.. 1993: Viljcien and Smart. 1993). As well. a small number of cases of atypical SHFM likely rcsult from large interstitial deletions found in the region of chromosome 2q3 1-qN (Ramer et cd.. 1989: 1990: Boles et al.. 1995).

In addition to this cytogenetic information. srveral linkage studies have been carrird out with pedigrees which segrezate SHFM. but in which no visible cytogenetic rearrangenients are present (Gumeri et cri.. 1994; Illarinoni er ni.. 1994: Palmer et al.. 1994). Al1 of these studies have resultçd in exclusion of linkage to the SHFMl locus. In fact. to this date. there are no SHFM families drscribed with no visible cytogenetic rearrangernents which show definitive linkage to the SHFMI locus. These exclusion results provided compelling support to the notion of genetic heterogeneity in autosomal dominant SHFM. Talh 1-2. Chariicteristics of SHFM Loci. -Locus Nsmc Map Location Mcans of 1,ociis Indcntification SHFM Subclriss SHFM 1 Nuiiierous cytogenetic Type 2 (?) recirrange trienis -rcduced penct rance -I;irgc iiitcrst itinl deleiioiis -iil>inlIiypopliisia -hnlanccd iranslocations -inversions

Linkage -single large pedigree

1 Oq24-q25 Linkagel single iiiibnlniiccd Type 1 triiii~10c;itiori -Full pciictriince -syiiienic to reg ion of riioiisc -rio long borie riitit:iiil DM- tleficicricy

single balniiced tr:iiislocotioii ?? single lnrgc intcrsii tinl delet ion

2q3 1-q34 n Ièw large interstitial dclciions

Liiikngc in one pedigree EEC wiili urogenitnl abiiormnl iiy SHFM3: 10q24 The SHFM loci on chromosomes 2 and 6, were the next two potential disease loci identified following the assignment of SHFM 1. Nevenheless, the best studied autosomal dominant SHFM locus next to SHFMl is SHFM3 which maps to chromosome lOq24.

Preliminary evidence for this locus was providrd by a single stillborn infant with sy ndromic rctrodacty ly . w hich possessed nn unbalanced translocation w i th karyotypç 46.

XX. -4. +der(l)t(l;lO)(p 15.1 :qlj.l)pat. Using microsatellitr markrrs that span chromosome 1Oq. linkagc was cstablished in a single pedigree with type I SHFkl. This study mapped the SHFM3 locus to a 9 CM interval in the 1OqX area (Nunes er d..1995).

Shortly following the identitication of the rnap location of SHFM3. the genr responsible for the mouse mutant Dm was mapped to mousc chromosome 19 in an ûrea syntcnic to the

SHFM3 locus (Johnson ci ul.. 19%). Using this information a second group independently idrntitïed a second pedigree linked to this locus (Gurricri er cil.. 1995). The similar mrip location and phenoty pic siniilarities sirongly sugrsted that Duc was thc mousc mode1 for SHFM3.

While the SHFM3 critical region is quite large ri number of genes thought to play a role in embryonic development map to this location. PAX-?. is a paired box transcription factor. thac hûs bren implicnted in human and mouse disordrrs which affect the development of numerous organ sy stems (Keller et al.. 1994: Sanyanusin er ai.. 1995). This gene is not cxpressed in the deveioping limb (Nornrs rr ni.. 1990; Dressler er al.. 1990), and mutations in this gene consistently result in normal limb development. Thus PAX-2 is not considered a good candidate for the SHFM3 phenotype. FGFRZ. as mentioned previously, is mutated in several craniosynostosis syndromes with associated limb abnormalities (reviewed in Webster and Donoghuc. 1997). Whilr the limb abnormalities do not reflect the SHFM phenotype. these are likely gain of Function mutations. The possibility that a loss of function mutation could result in ectrodactyly cmnot be excluded. In support of this, hypornorphic mutations in mouse FGFRI result in limb abnormalities

(Partanen et al.. 1998). Other possible candidate genes include the transcription factors

HOXI I, ZNF32, and HMX2 (Nunes er al., t995; Gurrieri et al., 1996). The role of ZNF32 and HMX? during development is not known (Cannizarro et ai.. 1993; Stadler et al., 1995). while HOXI 1 has been shown to play a rolr in T-cd1 tumorigenesis (Lu er al..

1991: Raju er ai., 1993). RBP4, a gene encoding a retinol binding protein which maps to the SHFM3 location (Rocchi et 01.. 1989). is intriguing since it is known that rctinoic acid dcrivatives play a role in normal limb development (discussed brlow). Probably the most interestinp of the candidate genes that map to the SHFM3 locus is FGF8 (Crossley and

Martin. 1995. As will be described below. this gene has been shown io play a critical role in the outgrowth and patteming of the vertebrate limb: the dekct in SHFM likely results frorn n defect in limb bud out_orowth(see chapter IV).

.A second locus for SE Most rrcently. a nrw autosomal locus for a syndromic ectrodactyly has been identified on human chromosome 19 by linkage within a single pedigree (O'Quinn et al.. 1998). This family displays EEC with an rissociated urogenitial abnorrnality. Mapping close to this locus is TGF-PI. It has been argued that this may be an interesting candidate gene since ectopic placement of TGF-P I protein suggests an endogenous function of this gene in digit formation (Ganan er al., 1996).

Autosorna1 recessive SHFbI.

While most pedigrees identified with SHFM or SE appear to segregate the disorder as a dominant condition there are sevenl reports in the litrrature of families in which SHFM is segregatiog as an autosomal recessive condition (Friere Maia. 1971: Verma et aL. 1976:

Zlotogon and Nubani. 1988). In a11 these cases consanguinity is apparent. Interestingly. the phenotype present in these individuals clearly represents atypical forms of SHFM, which, histocicaIly, was believed to be non-heritable (Temtnrny and McKusick. 1978). Digital hypoplasia and the absence of a true cleft are the characteristic features (Lange,

1937). 1t is a possibility, therefore. that recessive forms of ectrodactyly represent a clinically and genetically distinct disorder From typical SHFM. In addition. the developrnental rnechanism of disease may also be distinct in these two phenotypes.

1.C Animal Models of SHFM

SHFM phenotypes are not only found in man but haïe brçn dcscribed in a numbçr of other species inciuding the chimpanzee. the Macacu rhesus (Pearson. 193 1). the domestic car

(Searle. 1952). the Manatee (Watson and Bonde, 19861, and the mouse (Chai, 1981). Of these rnodels. the mousc mutant has served as the sprcirs of choiçc for which to further study the detëct in SHFM.

The mouse model for SHFM is called Dncryinplasi~z(Dac). This spontaneous mutant arose at the Jackson laboratory. and was first drscribrd by Chai in 1981. In this first report. heirrozygous mice were described as typically rnissing the rhree rniddle digits giving the linib a claw-like appeannce. This phenoiype was due to the absence of missing phalanges with or without metacarpals and/or metatarsals. It was also found that frequently there was fusion of the metatarsal/metac~als.and distonion of the t;usal/carpals at times.

Of interest in this original study was the realization that the Doc mouse represents a two- locus model of inherited disease. Tt was deterrnined that in order for the heterozygous phenotype to presrnt the mice must be homozygous for a second unlinked locus. This locus was temed mDac for modifier of Dac, and the dominant and recessive allele appeared to be relatively equally represented in the repertoire of inbred mouse lines exarnined.

Recently. detailed rnapping studies have been camed out with these mice. and the Dac locus was found to map to mouse chromosome 19 (Johnson et al.. 1995). As stated previously. this region of mousr chromosome 19 is syntenic to human chromosome

lOq24. in the SHFM3 region (Poirier and Guenet, 1994). On the basis of the phenotype and this mapping data it brame clear that this mouse was not only a phenotypiç model of the human disease but most likely a genetic model. The same genes that were found near

SHFM3 were found ro rnap within or near the Dac critical region (Nunrs rr cil., 1995;

Johnson et ni.. 1995: Gurrieri ri cil., 1996). The same üuthors funhcr miipped the niDac

locus to rnousc chromosome 13. The possible candidate gencs in this region are Fgf~-4

(Avraham et al., 1994). and MsxZ (Bell et al., 19931, a horneobox transcription fxtur

irnplicated in limb drvelopment (Davidson er ai.. 199 1 ).

To date. Dac and r~zDacrepresent the only characterizcd two-locus model system for a

naturdly occun-ing congenitcil malformation. Sugested mechrinisrns of action for this two

locus model include direct protein interaction. rithcr cooperritively. or as a ligand receptor interaction for rxample. Functional redundancy bctween these two loci has also been

suggested. These genes would function in a dose dependent manner. such that the phenorype would only resuli if both copies of one of the genes and at least one copy of the

other gene were inactivated (Johnson et al.. 1995). It is likely that the actual mechanism

uill not be revealed until the molecular nature of thesr gene products are detenined.

This two locus rnodel found in the Dac mouse raiscs an interesting alternative explanation for the abnormal inheritancc patterns found in human SHFM (Temtamy and ZvlcKusick. 1978). What was previously ascribed to reduced penetrance. gonadal mosaicism. or multiple alleles. could now be explained by a two-locus model of inheritance. In this case the phenotype would not be present. for example, in an obligate carrier unless a specific allele of a second epistatic gene was present in both copies.

Some work has been carried out with the Dac mouse to identify the molecular defect.

Treating F'pfS as the strongest candidate gene, Seto et al. (1997) atrsmpted to idrnrify a causative mutation in the coding region of this genc. This search did not uncover any such alteration. but this does not exclude the possibility of a regulatory mutation in F@. In support of a possible role for F,@ in the Dac phrnotypc. it was shown that these rnice display an increase in programmed ce11 deaih in the apical ectodermal ridge (AER. to be discussed below). the tissue of F@ expression.

1.D Embrvologv of the Vertebrate Limb

For the pater part of this century. the venebrate limb has served as a model organ system

in which to study the process of patrern formation during developrnent. As such, the study

of venrbrate limb development has lrad to significmt contributions to the advancement of

our understrindin_pof many of the fundamental principles of developmental biology such as

axis determination. organizer activities. morphogenetic gradients, and epithelial

rnesenchymal interactions (reviewed in Tickle and Eichele. 1994).

Pnor to the advent of rnany rnolccular biological techniques, the study of limb development

relied on one's ability to alter the position and orientation of various embryonic tissues.

Classically. the chicken embryo has served as the model organism in which to investigate

vertebrate limb development since the it is most amenablr to in ovo manipulation of embryonic tissues without significant compromise to the survival of the embryo. More recently, with the advent of molecular biology and transgenic animal technologies. the mouse has become an increasingly useful model for which to study gene function during limb development; a model organism that allows for genetic manipulation.

Below is an overview of the development of the venebnte limb. I have chosen to combine the classical ernbryological data with the more rrcent molecular and genetic data. in the hopes of simplifying the discussion. While much of the knowledge has been obrained using chick. it is increasingly clear that at least the sarly phases of limb development are sufficirntly similx in al1 vertebrates such that one may extrapolate to other species.

It should be noted rhat the field of vertebrate limb development is indebted to the rrsearchers who study signaling pathways and pattern formation in the fruit tly Drosopltila nrelano,qnsrrr. A discussion of this area of study is out of the scope of this thesis (reviewed in Ingham. 1995: Irvine and Vogt. 1997: Serrano and O'Farrell. 1997). however. molecular identification of mltny key factors coritrolling vrrtcbnte limb development has been dependent on the pnor identification of these grnes in the tly.

I.D. 1 Vertehrate limb induction

The venebrate limb is derïved from the larenl plate mesoderm. Initial limb bud outgrowth is dependent. not on an increase in the rate of ceIl proliferation in the prospective limb mesoderm. but on a selective decrease in proliferation in the tissue surrounding this region t Searls and Janners. 197 1). Studies using foi1 barriers (Stephens and McNulty. 1981;

Strecker and Stephens. 1983) and tissue ablation (Geduspan and Solursh. 1992) have clçarly shown that the initial development of the vertebrate limb bud is drpendent on inductive signals from the intermediate mesoderm (the mesonepheros). Recently. fibroblast growth factors have emerged as likel y candidates for the endogenous limb bud inducers. Specifically. fibroblast growth factors 8 (Fg(8) and FgfIO are both expressed in the intermediate mesoderm at a time prior to limb induction and when they are ectopically applied to the flank mesenchyme, fully formed ectopic limbs are generated

(Cohn er ai.. 1995: Crossley et al.. 1996; Ohuchi et ai., 1997). The flank is the region in between the prospective fore- and hind-limb territories that is compctent to form limbs.

FgfZO is also expressed in the latenl plate rnesoderm of the prospective limb field. The timing of expression of FgjS and FgflO has lcad to the suggestion that Fu8 in the intermediate mesoderm induces Fa0 in the prospective limb field mesoderm which leads to the initial outgrowth of the lirnb bud (Ohuchi et ai. 1997). Genetic evidrnce for the role of Fds in limb bud induction has yet to be obtained.

I.D.2 Patternin? of the eariv vertebrate Iimb

Axes determination

The early limb is a simple structure histologically consisting of mesenchymal tissue

surrounded by epitheliurn. The vertebrate limb develops dong three axes: proximal-distal

(PD). dorsal-ventral (DV) and. anterior-posterior (AP). In humans. dorsal refers to the

back of the hand or the top of the foot. and antrrior refers to the rhumb or great toe.

Experiments in which the orientation of the limb primordium was rotated. with respect to

the body mis. and where ectoderrn was rotated with respect to mesenchyme have shown

that both the AP and DV axis of the limb are detemined long before the appeannce of the

bud itself and. at these early stages. the polarity of these axes is dictated by the

mesenchymal cells (Harrison. 19 18: S wett, 1937; Hamburger, 1938 Chaube. 1959; Zwilling. 1956: Reuss and Saunders. 1965). At a later stage the control of DV polanty is

subsequently transkrred to the epithelial cells (Saunders and Gasseling. 1968: MacCabe

and Saunders. 197 1 ). however. AP polarity remains under the control of the mesenc hy me. In conuast. the PD axis is not detemined at this early stage. and it would seem that it is reversible even following initial stages of limb bud outgrowth (Saunders. 1972).

The apical ectodermûl ridge (AER) Shonly after the initial outgrowth. signals emanating from the mesenchymal ceils induce the surface sctodsrmal cells at the distal apex of the lirnb bud to change rnorphology and form a thickened ridge. known as the apical ectoderrnal ridge (AER) (Saunders, 1948;

Kieny. 1960: Saunders and Reuss, 1974). In the chick embryo the AER is hisiologically ü pseudosrratified columnar epithelium wtiile in mammals a stratified cuboidal spithclium is the norm (Fallon and Kelley. 1977).

The role of the AER as drtermined experimentally is to signal the undrrlying mesenchymal cells to rnaintain thern in an undiffsrentiated proliferative state, thereby rnaintaining continued limb bud outgrowth. Classical experiments showed that rernoval of the AER durinz early lirnb developrncnt leads to distal truncations of the ensuing appendage: the earlier in developrnent the ridge is removed the more proximal the level of the truncation

(Saunders. 1949: Sumrnerbell et of., 1973: Summerbell. 1974) (Fig. 1-38).

A widely accrpted mode1 of limb development was derived from these experirnents. known as the progress zone model (Summerbell et al.. 1973). The progress zone (PZ) is the sub- ridge mesenchymal cells under the influence of the AER signals. In this model. it is thought that as the limb bud outgrows, cells leave the progress zone at the proximal side.

Once cells are no longer under the influence of the AER their fate is deterrnined and they,

subsequently. begin to differentiate. The time in development at which the cell laves the

progress zone. and the position of that ceIl with respect to the AP and DV axis determines

the idsntity of that ce11 in the adult limb. In this model it is inferred that the structures of the limb are determined in a proximal to distal direction. This corresponds well with fate mapping studies of the vertebnte limb (Saunden, 1948).

The endogenous secreted AER activity is probably encoded again, by Fgfs. FLdZ.1. and

8 are expressed in the AER; Fg/2 and Fg/8 are found throughout the AER while Fd4 is

expressed in a graded pattern bring strongest in the posterior and absent from the anterior

one third (Niswandrr Cr ciL. 1993; Fallon et ai., 1994; Crossley et al.. 1996; Vogel er al.. 1996). AI1 three of thesr molecules. when applied to the distal limb bud following surgical AER removal are capable of rescuing essentially complete limb bud outgrowth. Currently.

genetic dissection of the fiinctions of these genes is being punued to clearly establish their endogenous role in limb development. Mouse knockout models of F@ would suggest some functional rrdundancy between thrse genes since limb development is nomal in these

mice (Ortega er al.. 1998). Howevrr. an AER specific Fgf8 knockout rnodrl clearly

establishes an endogenous role for this gene in AER signaling (Lewandoski el al.. 1997).

The zone of polarizing activity (ZP.4)

While the key role of the XER is to maintain limb bud outgrowth. it is also essential for the

maintenance of the zone of polririzing activity (ZPA). The ZPA is a rnorphologically

indistinguishablc region of rnesenchymai cells located in the posterior distal limb bud (Hinchliffe and Sansom. 1985: Honig and Summerbell. 1985). The discovery of the ZPA was serendipitous. when Mary Gasseling in her attempts to study the role of the

mesenchyme to maintain the thickness of the AER (to be discussed later) grafted posterior

rnesenchymal cells undrrneath the AER at the apex of the limb bud. This expenment

resulted in the presence of a supernumerary limb tip anterior to the graft (Saunders and Gasseling. 1968). Subsequent experiments on the role of the ZPA showed that when the

posterior mesenchyme is grafted io the anterior side of a contralateral limb bud. a rnirror

image duplication of distal limb elements results (Tickle et ai.. 1975) (See Fig. 1-3C). A mode1 for AP patterning suggests that the function of the ZPA is to control the AP polarity of the developing lirnb bud. The closer to the ZPA the more posterior the fate of the cells, and as you move funher away from the ZPA the more anterior the structure. It is thought that the ZPA is the source of a morphogen. a secreted molecule that has a dose dependent effect on the inductive properties (Tickle et al.. 1975; Wolpert. 1989). For a long time the key candidate molccule for the ZPA activity was retinoic acid (RA). It was shown that ectopic application of RA soakrd beads to the anterior mesenchyme of a limb bud would mimic the affect of transplanting a ZPA (Tickle et al.. 1982. 1985; Summerbell.

1983). Lack of support for RA as the actual endogenous ZPA factor. however. cornes from the fact that the concentration of cxogenously applied RA required to generate these mirror image duplications induces its receptor CO a much higher level than is encountered in the posterior limb bud during normal limb drveloprnent (Noji et al.. 1991). In light of this discrepancy it is thousht that ectopically applied RA has the ability to induce an cctopic

ZPA but not replace the ZPA.

A much stronger candidate for the endogenous ZPA signal is the vertebrate homologue of the Drosophiio lzrdgehog gene. Sonic liedgehog (Slili). Shh is rxpressed in the limb bud in a pattern which precisely overlaps that of the ZPA. and when ectopically applied to the anterior mesenchyme of the vertebrate limb gives rise to a mirror image duplication (Riddle et cd.. 1993). The level of duplication is dose dependent. It is accepted that Shlz irnpans some or possibly al1 the function of the ZPA. Genetic ablation of Shh through gene knockout experiments in mice results in a limb thar is completely devoid of an autopod

(hand or foot) (Chiang et ai.. 1996). It is clear, therefore, that the role of Shh on limb devrlopment stems far beyond patterning the AP axis of the limb. Other gcnes are also expressed predorninantiy in the posterior mesenchyme of the developing limb bud. In many cases these genes have been shown to be induced by ectopic Shh and are likely downstream. Most notable is bone nrorphogeneric protein 2 (BmpZ). Bntpl. and Prc. Pfc is likely the receptor for Shh. Prc expression overlaps with

Shh expression in the posterior limb bud mesenchyme and is induced by ectopic Shh

(Goodrich et ui. 1996: Marigo et al.. l996b). Bt~rpZand BmpJ are also expressed in the posterior limb bud. Both are expressed in the AER and BnipJ is further expressed in the anterior mesenchyme (Lyons et al.. 1990: Jones rr cd.. 1991: Francis et al.. 1994).

Expression of BtnpZ has been show to be induced by ectopic ZPA grafts or Shh (Francis et ni.. 1994; Laufer er al.. 1994). Initially. it was reasoned chat BnipZ may bi: the actual morphogen mediating the patterning of the AP axis and that Slih may serve to activate BnipZ. However. ectopic expression of Bi~lpZalone does not yield a mirror image duplication but instead lrads to ectopic Fgf4 expression in the anterior AER (Duprez et al..

1996). Thus. BnipZ may translate the SIih signal to the posterior AER thrreby inducing

F,$4 expression. but does not directly impact AP patteming.

As mentioned one role of the AER is to maintain the ZPA. In hct it has been demonstrated thiit there is a positive fecd back loop betwren Slilt of the ZPA and F,S4 in the posterior AER. such that Shli induces and maintains Fs/J expression (meniioned above). which subsequenti y maintains proper Shh expression in the posterior limb bud mesenchyme

(Laufer et al.. 1994; Niswander et al.. 1994). Ectopic expression experiments with BmpZ would suesest chat this interaction is mediated by this gene. Furthemore. mutant analysis of the lintb de fur mi^ mouse shows thnt the furmin gene is also required for this feedback interaction (Chan et al., 1995a; Haramis et al., 1995: KuhIman and Niswander, 1997). Patterning of the dorsal-ventral axis Patteming of the dorsal ventral (DV) ais during limb development is also coordinated with

AP and PD patterning. As mentioned. prior to the onset of limb bud outgrowth, control of

DV polarity resides in the pre-limb mesenchyme. When the limb bud outgrows. this control is tnnsferred to the ectoderm. Wnr7a. a member of the Wntlwg family of signaling molecules is expressed in the dorsal ectoderm and is required to pattern the dorsal limb

(Dealy et al.. 1993; Riddle et al., 1995; Parr and McMahon, 1995). W11t7a in the dorsal ectoderm induces the expression of bul,a lim domain protrin, in the dorsal rncsenchyme of the lirnb bud (Riddle et al.. 1995: Vogel et al.. 1993). Part of the dorsal patterning activiiy of Wnt7a is mrdiated by Ltrisl (Chen er al.. 1998). Wnt7a is also required to maintain Shh properly in the posterior mesenchyme (Riddlr et al.. 1995; Parr and

McMahon, 1995).

En-I. a venebrate homologue of the Drosophila engrcrilrd homeobox transcription factor is expressed in the ventni ectoderm and is required for ventral patterning as well as inhibiting dorsal patterning by inhibitins Wrtt7

Rodriguez-Esteban et al.. 1997).

Induction of the AER requires proper formation of the DV axis. such that the AER foms at the DV boundary of the limb bud. Ectopic expression of the srcreted molecule Radical fringe. a homoIoeue of the Drosopkila genr friiige (known to be involved in Norch

signaling) argues for a role for this grne in determining the DV boundary during limb bud.

and. hence. critical for AER induction (Laufer er al.. 1997; Rodriguez-Esteban et al..

1997). Studies cmied out by Michaud et al. ( 1997) have shown that the prirnary signal for

DV polarity iikely resides in the somites and the somatopleure. prior to limb bud

outgrowth. Expression of afiinge homologue in the somites may mediate this process. Figure 1-3. Development of the Vertebrate Limb. (A) Two signaling regions. the apical ectodermal ndge (AER) and the zone of polarizing activity (ZPA). are thought to control the developrnent of limb pattern, dong the AP and PD axes. Signals from the AER. which are likely FGFs. iict upon a region of mesenchyme beneath the ridge known as the progress zone (PZ) maintaining this tissue as an undifferentiated actively prolifenting region. thus allowins for limb bud outgrowth. The AER also serves to maintain the ZPA in the posterior mesenchyme. Shh in the ZPA induces Fg/? in the posterior AER which in turn maintains Shh expression. The main role of the ZPA is exened on the progress zone in the specification of AP identity. The PZ is a source of the AER maintenance factor (AERMF) which is ncccssary to maintain the AER through normal development. (B) Whrn the AER is rernoved t'rom an early limb bud. a truncated lirnb dwdops. Simultaneous application of Fgf frorn heparin beads to the posterior limb bud rnargin and the apex of the bud essrntially restores normal limb development. (C) Exprrimcntal manipulations that respecify the AP limb pattern are erafting of a second ZPA. local release of retinoic acid from an ion-exchange bead, and ;ctopic expression of Shh. In al1 instances a complete minor image duplication of the digits can result. This argues for the role of the ZPA in determining the AP pattern of the lirnb. (modified from Tickle and Eichele. 1994) leads to Iimb with ~rooer - thé humerus

I I 1 Proximal i

ulna r AER [Posrenor] Progress Zone

limb truncation

rescues normal developrnent

beads soaked in Fgfs -b

craft ZPA, CIretinoic acid soaked bead, or a SHH expressing cells to antenor mesenc hyme duplication of digits Maintenance of the AER As stated, formation of the AER is dependent on inductive signals from the underlying rnesenchymal cells. FoIlowing AER induction. and its differentiation, it has been shown that signals, also emanating from the mesenchymal cells. are required to maintain the AER. until the stage at which it regresses (Sce Fig. L3A) The first evidence for this maintenance activity came from classical exprnments that showcd that blocking rnesenchyme/AER interactions with a thin mica barrier for 1.5 hours or more resulted in the tlattening of the

AER (Saunders. 1949). It was subsequently shown that non-limb rnesenchyme placed under an AER also results in the flattening of the AER (Searls and Zwillirig, 1964).

Experimrnts where pieces of boih limb and non-limb rnesenchyme were placed under the

AER showed that a thickrned AER was maintained only where limb mesenchyme was juxtaposed to the AER (Zwilling, 196 1).

Thcse rxperimenis Ird Zwilling and Hansborough (1956) to coin the term "AER maintenance factor" (AERMF). They further observed that the AER is thicker on the posterior side than on the anterior side of the chick limb bud. reasoning that. in chick. the

AERMF is asymmetrically distnbuted king more prevalent on the posterior side.

In their work with polydactyly mutants Zwilling and Hansborough (1956) also showrd that the defect residrd in the mesenchymal cells. The onset of the defect could be marked, however. by a posterior-like AER thickening in the anterior ridge. From this they suggested that polydactyly is caused by increased levels of AERMF in the anterior rnrsenchy me. Recent studies in polydactylous mouse mutants have revealed that ectopic anterior digits are likely caused by ectopic Shh andlor Indian hedgehog (Ihh),a highly related rnolecule with similar activity, in the anterior mesenchyme of the developing limb bud (Chan er nL. 1995b: Buscher et al.. 1997: Masuya et al.. 1997: Yang et al.. 1998).

Thus these naturally occurring mutants essentially mimic the experiments of ZPA transplantation in chick. This may suggest that ectopic AERMF may not be the pnmary cause of post-axial polydactyly. Nevenheless these expenments in no way rule out that ectopic ZPA induces ectopic AER maintenance activity. it is clear however from early studies of the endogenous role of the ZPA. that ZPA and AERMF activities are separable

(Summrrbell. 1979: Kaprio. 198 1 ).

It shouid bc poinred out that whilr the AER in the chick limb bud is asymmetnc. rhis is not the case in the n~amn~alianlimb bud. Thus it may be suggested that the AERMF is symmetrical in the mouse limb mesenchyme.

Following these rnrly experiments on the AERMF, a paucity of experiments have been performed to funher investigate this activity. Important. however. are the experiments carried out by Saunders and Gasseling (1963) which showed. using semi-permeable bamers srparihg AER liom the mesodrrrn. that the AERMF is a secreted factor. Several experiments ünalyzin_iAER maintenance in culture drfinitively showed that the chick AER will rapidly undergo cell death in culture, uniess limb bud mesenchyme is placed in close contact with the ridge (Bourin and Fallon. 1984: Fallon et al.. 19%). These studies also showed that insulin applied to the culture medium was also able to maintain the ridge.

Xevenheless. the endogenous role of insulin on AER maintenance is not known.

While the endogenous AERMF has not been identified, FdlO stands out as an interesting candidate for this factor. As mentioned. this gene likely has an endogenous function in limb induction. but following limb bud outgrowth, FgflO is expressed in the SUD-ridge mrsenchyme (Ohuchi es cil.. 1997). Furthemore. ectopic expression of Fgf'l0 is able to induce and maintain FgfS expression and a rudimentary AER in non limb ectoderm.

Inierestingly. F@ can in turn induce FgflO expression. Recently it was show that PZ cells grown in culture in the presence of Fgfs retain their ability to maintain the AER (Ham et al., 1998). If they are cultured without Fgfs this ability is lost. It was speculated that F@ in the AER may induce FgflO expression in the mesenchyme which in tum maintains the AER.

i.D.3 Patterninr of the limb skeleton.

Followin_othe initial stages of limb bud outgrowth the distal aspect of the bud expands in the AP axis to form a paddlc shaped hand/foot plate. It is during this stage that the first morphological differentiation during limb development occurs. as mesenchymal cells condense to form the skeletal blastemas (Muneoka et al.. 1989). The sequence of condensation occurs in a proximal to distal sequence and is associated with branching and

segmentation of thesr blastemas to form the precursor limb skeleton (Forsthoefel. 1963). Like all rndochondral bone formation (where a canilaginous structure is formed prior to

bone formation). following condensation. the mesenchymal cells differentiate into chondroblasts and secrete hyaluronidase and chondroitin sulfate. A subsequent switch

from type 1 collagen to type II collagen. in association with the changes in proteoglycan composition marks the drfiniir stage of cartilage differentiation. iMuch later in development

cartilage is replaced with bone (Carlson. 1988).

Hox genes in limb development

The molrcular mechanisms controlling skeletal patteming in the limb are not as well

unders tood as early patterning. Nevenheless, i t is becoming clear that homeobox genes

found in the Hus clusrers play key roies in the patterning and the formation of limb skeleton (reviewed in Rijli and Chambon. 1997). Ho+ genes, particularly those of the 5'

Hom and Hoxd clustcrs are expressed in overlapping patterns in a spatial-temporal colinearity in the developing limb, similar to Ho*. gene expression dong the body anis in vertebrates and invertebratrs (Nelson es al., 1996). The members of the Ho.m cluster are expressed in nested domains dong the PD axis (Yokouchi et ai., 1991). Hoxal3 is the most distally restricted. In keeping with the temporal colinearity, onset of the expression is 40-al3. Hoxd genes are expressed in nested dornains in the limb bud that are progressively restricted along the AP ais (Dolle et ai., 1989).

Based on the expression domuins it was initially thought these YHos genes served to delimit segment idrniity in the venebrate limb. It was thought that the Hoxa clustrr would determine PD identity. such that expression of A9 alone would derermine shoulder girdle identity. 9 and 10 sty lopod (upper arm). 9.10.1 1 zeugopod ( forearm). and 9.10.1 1. and

13. autopod (Tabin. 1992). Similarly it was believed that the Hosd genes determined the identity of the 5 digits of the mammalian limb (Morgan and Tabin. 1993). This rolr for

Hoxd genes was funher supportcd by experiments demonstnting that thrse genes çould be ectopically activatrd by ectopic ZPA activity in a nested pattern sirnilar to that found endogenously ( Izpisua-Bclmonte et al.. 199 1 ; Nohno et cil.. 199 1). This provided intrîguing yet circurnstantiril evidence that 5' Hard genrs mediüted the pded affect of the

ZPA. It was thought. therefore. that the absence, or ectopic expression of these genes would rcsult in homcotic transformations. The first study. misexpression of Hoxdll. appeared to contkm this hypothrsis (Morgan et al.. 1992).

Since these first expcrirncnts many of the Ho.r genes expressed in the vertebrate limb have been knocked out. and it is becoming increasingly clear that the role of these genes in vertebrate limb development is exceedingly complex and cannot be explained by the initial eloquent hypotheses (revirwed in Rijli and Chambon. 1997). While there does appear to be some tmth to rhe idea of segment identity. (eg. affected structures fa11 into the expected segments). t here are corn plex genetic interactions between genes of the same clus ter and paralogous members of the different clusters. that control recruitment of cells in cartilage condensations, and later during the growth phase of these elements. It also appears that some of these genes rnay be necessary for the proper timing of the endochondral ossification as seen in the HoxdI3 knockout mice (Dolle et al.. 1993). Therefore, the Hox genes do play key roles in growth and differentiation of the developing limb skeleton: complex genetic dissection of the precise roles is required to gain a complete understanding of skeletai patteming.

Interdigital ce11 death A final stage in the general patteming of the venebrate limb is the onset of apoptosis in the regions between the regions of digital condensation shonly following the formation of the cartilaginous digital blastemas. It is this process of' programrned ce11 death (PCD)that

separates the digits (when applicable) to give rise to the final limb pattern (Saunders er al..

1962). Molecular studies have identified Bone morphogenetic proteins (BMP)as key

regulators of the proccss of interdigital ce11 death (reviewed in Hogan. 1996). Bmp2.J and

7 al1 are expressed in the interdigital region (Francis et al., 1994: Jones et al.. 199 1; Lyons

et al.. 1995). and careful expression analysis has shown that the pattern of expression is

intirnately associated with the pattern of PCD (Zou and Niswander. 1996; Macias er aL. 1997). Experiments using beads soaked in B~rzpshas shown their ability to induce PCD

(Ganan et al.. 1996; Mrrino Er d..199s). While the obstruction of BMP-signaling with a

dominant negativc receptor resultrd in decreased interdigital apoptosis (Zou and

Niswander. 1996). It is ihought chat during interdigital ce11 death, the Bnzp signal may bt.

rnediated by Msx genes (Ganan rr

Ferrari et al.. 1998). 1.E Pathogenesis of SHFM

Pnor to the work carried out for this thesis. the precise pathogenesis of SHFM was not known. An explanation based on suppression of the formation of the central nys has been proposed. but this lacks any mechanistic information. Sorne attempts were made to suggest a reasonable mechanism of the disease based on the available knowledge of lirnb developmrnt. One such suggestion was a central defect in the AER or the PZ that would result in the restriction of the phenotype to the central rays (Watson and Bonde. 1986: Roberts and Tabin, 1994). At the tirne of'this suggestion. no gene had bern identified that

was centrally expressed in the developing lirnb bud. However. a similar phenotype is

found when ooly the central portion of the AER is rernoved from a developing chick limb

bud (Saunders. 1948: Surnrnerbell. 1974). Other explanations include a late defect in

patterning dong the AP anis. Digits do not form in a strict AP sequence (Burke and

Alberch. 1985). and thus a "field defect" could affect the developrnrnt of some digits while

not affecting others (Roberts and Tabin. 1994). A defect in the branching of the

cartilaginous elements has also been suggestrd (Robens and Tabin. 1994).

1.F Identification of Candidate Genes in a Positional Cloning: Strate~v

Positional cloning describes the identification of a disease gene based on its chromosoma1

location without any prior knowledge of the biochemical function of the gene product.

Widely ernployed now. this methodology of gene identification was first irnplemented in the 1980's with the cloning of the genes for chronic gnnulomatous disease (Royer-Pokora

et nL. 1986). and DuchennelBecker muscular dysrrophy (Monaco et al.. 1986). Like the work on the SHFM 1 locus, the identification of these genes was greatly aided by the identification of visible cytogenetic rearrangements in patients. A great advancement in positional cloning was later realized with the identification of the gene for cystic fibrosis

(Rommens et ai.. 1989). In this example no cytogenetic "flag"was available to direct the researchers to the disease locus.

The steps of a positional cloning endravor include Jetermination of the grneral chromosomal position of the locus. refinement of that position, î-ollowed by seneration of a physical rnap of the locus. Once a physical map has been rstablished for the locus. rnethods of gene identification are thrn cmployed to isolüte potrntial candidate genes frorn this locus.

There are a nurnber or widely usrd methods for genr identification that arc used to identify oenes from a specific chromosomal location. These include. exon-trapping (Duyk et al.. C

1990: Buckler er al., 199 1). direct cDNA selecrion (Rornmens et al., 1'394). the search for conserved gcnomic segments (Monaco cr al.. 1986). and genomic sequencing (Uberbücher and Mural. 199 1 ). Exon-trapping utilizes endogenous introdexon boundary sequences in a genomic DNA clone to isolate small pieces of sene coding regions from within the clone.

Direct cDNA selection consists of hybridizing pools of çDNAs from niimerous tissue sources. to irnmobilized genomic clones from a given locus. Bound cDNAs are subsequently eluted and cloned. to once again $ive small pieces of possible genes. Searching for conserved sequence across species is a valuable way to identify segments in genomic DNA clones that may encodr for genes. Gene coding regions genenlly display high srquence conservation bctween closely related species. whereas non-coding regions genenlly have little or no sequence similarity. Currently. the most widely applied method for gene identification is through total genomic sequencing. With the advent of the human genome project and the sequencing of the entire genome. it is possible to analyze the availablr genomic sequence for patterns refirctive of functional gene segments. Potential coding segments can then be used to screen cDNA libraries or gene sequence databases to identify cDNAs corresponding to the genomic sequence.

Prelirninary analysis of gene function

Following gene identification, it is ot'ren prudent to cary out functional analysis of the potential candidate genes to further refine the list of potential diseass genes. Currently much is known about sequence/function relationships for nurnerous genes. By cornparing amino acid sequrnce it is possible to gain insight into the potential biochemicai function of newly identified grnes. In many cases these genrs may have been previously cloned.

While sequence cornparison can prove to be useful. rnosi of the information obtained will be speculative. based on the propertirs of relnted senes. A more direci approach to refine the list of candidate genes is to detemine the patterns of gene expression for the various

ÇDNA clones. Northern blotting is a widely usrd method for analyzing the expression profiles of genes. and can be helpful in refining candidate genr lists. For example. if the disease in question affects kidney function. one would expect to observe gene expression in kidney tissue. The absence of detectablr kidney expression may serve to rule out the prospective gene.

A more specific approach to address the potential candidacy of a gene is to analyze the expression of a the gene i~isitu. by using RNA in siric hybridization (Wilkinson. 1992). In this case an antisense RNA probe is synthesizcd in iitro using a cDNA clone as a template.

In the process of RNA synthesis the probe is labeled. typically with a radioactive nucleotide or a nucleotide conjugated to a substrate that. whcn metabolized. @es a visible signal (cg.. alkaline phosphatase). The RNA probe is then hybridized in a whole mount embryo for example or to a tissue section to reveal the precise expression profile. This technique not only determines the organ in which the sene is expressed but can determine the expression pattern to the resoiution of ceIl type, as well as the spatial and temponl pattern of expression. While it is possible to perform these experiments with human tissues, these tissues may be very difficult to obtain. particularly when it is embryonic expression that is of interest. For this reason analysis of the expression of the murine homologue in mouse ernbryonic tissue is frequently performed and the data is extrapolated to humans. This

method is very valuable when the diseasr of interest is a congenital abnormality. since the appropriate human embryonic tissue is frequently not available for Northrrn blot analysis.

Funhermore, the resolution of RNA in situ hybridization may br rcquired to generate an expression profile of a gene in sufficicnt detail to makc a qualified decision concerning its potrntial as a candidate.

1.G Thesis Overview

A main goal in my laboratory is to idrntify the gene(s) responsiblr for SHFM. specifically on chromosome 7 (SHFM 1). Prior to the onset of this thrsis. Stephen Schsrcr. who ai the

time was a graduate student in the laboratory. idrntified the locus for this disease at

chromosomr 7q21.3-q12.1. and subsequently carried out a detailed molecular characterization of this locus and begm to idenri. genes from this region. The work

presented in this thesis deals with the characterization of these genes to assess their

potential rok in the disease. In addition. the phenotypic characterizaiion of a mouse mode1

for this disease was performed. to elucidate the developmental mechanism of the phenotype

and gain funher insight into possible molecular defects.

Chapter 1 consists of a general overview of the available information regarding SHFM. As

well. a comprehensive overview of the development of the venebrate limb is given, along

with a brief synopsis on the isolation and chmcterization of candidate disease genes. Chapter 2 describes the cloning and chancterization of a candidate gene for the disorder.

DSSI. This first gene characterized from the SHFMl region is a novel gene with unknown biochemical function. Described is the expression of this gene during mouse embryogenesis suggesting an involvement of this gene in distal lirnb development.

Chapter 3 outlines the work done on the cloning and characterization of a second gene froni

the SHFM critical region, DNCII. a cytoplasmic dynein gene. dong with a highly related

.=osne DrVCI2. DIVCII was found not to express during limb bud development thus niling ir out as a candidate gene. On the other hand, identification of a dynein sene in the SHFM 1

critical rcgion suggests a potentiril role for this gene in the transmission ratio distortion frequrntl y encountered.

Chaptcr 4 describrs the phenotypic characterization of the mouse mutant Dm-.A numbrr of methodologiss werc employed to concludr that the defect in thcse mice is a failure to

maintain the XER. in the central portion of the lirnb. This same defect likely occurs in the human disease. This work. iherefore. not only helps to better definr candidate genes. but greatly rnpands the knowledge concrrning AER maintenance.

Chapter 5 further describes the characterizaiion of 3 additional candidate genes from the

SHFM 1 locus, DLXj. DU(6. and DSS2. The DLY genes are good candidates since their

corresponding gene from Drosophiin. disrul-less. is required for proper appendage developrnent. and their expression patterns in mouse are consistent with a role in limb bud development. DSS2. a recently identified gene. may represent the best candidate gene for

SHFM 1. since it is expressed in the AER at a critical time. and is the only gene from this

region found to be directly intempted by translocation breakpoints. In the final chapter a genenl discussion is given concerning the various genes identified and their potential role in the disease, and maintenance of the AER. t use this section to suggest future experirnents that may be carried out to further understand the genetics and developmental biology of SHFM. Chapter II

Further Refinement of the SHFMl Locus and Analysis of a Candidate Gene for its Expression During Developrnent.

Attribution of work:

The majority of work presrnted on the analysis of DSSI was performed by the candidate. The dcterminntion of the subcellulrir distribution of the protein was performed by David Sinasac. a fellow graduate student in Our laboratory. The analysis of Dssl expression during tin re_oenention was carricd out by Dr. Marie-Andree Akirnenko (Loeb Institute. Ottawa Civic Hospital). The rnapping of translocation breakpoints, and preliminary identification and mapping of al1 the genes was carried out in part by Dr. Stephen Scherer, in his capacity as a graduate student. research associate and staffscientist at the Hospital for Sick Children. Direct cDNA selection was performed by Dr. Iohanna Rornmens.

The data presented in this chapter has been published in part in:

Crackower, @I.A., Schercr, S.W.. Rommens, LM.. Hui, C.C.. Poorkaj, P.. Soder, S.. Cobbcn. LM.. Hudgins. L.. Evans. J.P., and Tsui. LX. ( 1996). Characterization of the split hanci/ split foot malformation locus (SHFMI)at 7q21.3-q22.l and analysis of a candidate gene for its expression dunng limb development. Hum. Mol. Genet. 5: 571-579. 1I.A Introduction

A subset of SHFM and SE patients were found to have cytogenetically visible chromosome rearrangements. Molecular studies of these patients have led to the assignrnent of an autosomal dominant form of the disease to 7qZ 1.3-q22.1 (Scherer et al., 1994a: Mannoni et ai.. 1994): the locus has bcen designated SHFiMI (Tsui and FarraIl. 199 1). A physical map consisting of overlapping yeast anificial chromosome (YAC) clones encompassing this entire region has been consticted (Scherer er ul.. 1994b). Together with the mapping of deletion breakpoints. a critical region of - 1.5 Mb could be established for the SHFM 1 locus

(Scherer er cd.. 1 Yg-tb). Furthermore. the breakpoints for 6 patients with translocations or inversion were localized within the cri tical region. Although these patients included both simple SHFM and çomplrx SE. there was no obvious correlation brtwren the sites of chromosomal disruption and clinical severity.

Haploinsufficiency may bc considered the cause of SHFMI (Scherer et ul.. 199Jb). Deletion. translocation or inversion involving the critical region rnay affect the activity of one or a series of enss. through either direct interruption of the gene(s) or their regulatory elrments. To pin funhrr insight into the molecular basis underlying SHFMI. a systematic search for candidate genrs locatrd in the critical region has been conducted.

This chapter describes the identification and rnapping of 6 additional disease causing visible chromosomal rearrangernrnts and the identification of 5 genes that map to the critical interval. A detailed charricterization of one of these genes. DSSI. a highly conserved. novel gene. is provided. RNA in situ hybridization data show that Dssl has a widespread. but dynamic expression partern in the mouse embryo suggesting an important role in mammalian devrlopment. including proper limb morphogenesis. The zebnfish homologue of Dssl was also cloned and a preliminq analysis of its expression during ernbryogenesis and fin regeneration is presented. Finally. an epitope-tagged construct of Dssl was transfected into COS cells, which revealed a predominant nuclear localization of the protein. The data suggest that this gene is a strong candidate that should be considered when delineating the etiology of SHFM and SE.

1I.B Materials and Methods

II.B.l FISH mapping of chromosome rearrangements in SHFM and SE patients

The precise breakpoints of rearrangements determined by FISH were localized within the

çlonrd grnornic DNA sqments on the basis of altered signal intrnsity (in the case of chromosornr delcrion) and splitting signals (for translocation). as drscribed (Scherer et d..

1994b).

II.B.2 Gene identification

Direct cDNA seleciion was performed according to described methodology (Rornmens et nL. 1994). Brirfly. the YAC DNA was purifird from the endogenous yeast chromosomcs by pulsed field gel electrophoresis. transferred to nylon membrane. ana usrd as substrate for rxhaustiw hybridization with PCR-amplified cDNA pools made from RNA isolated from 10 tissues (Romrnens et cil.. 1991). Exon amplification was conducted on "tiling path" phage and cosmid clones using the pSPL3 vector and the protocols described by the supplier (GibcoIBRL). To identify evolutionarily conserved DNA sequences, genomic sub-clones from the phage. cosmid. and YACs were hybridized systematically against

DNA from seven different specirs. The 132 DNA segments retiieved in these experiments that rnapped back to the rxpected genomic clones were analyzed by DNA sequencing.

Northern blot analysis with RNA from various tissues as indicated. and Southem blot analysis with genomic DNA (cloned and uncloned) according to standard protocols

(Sambrook et al.. 1989). The DNA sequences were grouped and aligned. if overlapping. examined for potential open-reading frames, and used to screened GenBank, dbEST and other databases for sequence identity and possible motifs.

II.B.3 cDNA clones and libraries

The full-lcngth DSSI clones FC4C and FB1IA were isolated from cDNA libraries of the frontal cortex and frtal brain. respectively. Five full-length Dssl mousr clones were isolated from an oligo dT-primrd cDNA library constructed in the UAPvector (Stratagrne) with mRNA pools of the limb buds from E9.5 to E13.0 mouse embryos. Zebrafish cDNAs were isolated from an oligo dT/random primed library made from mixed embryonic stage zebrafish RNA (kindly provided by Dr. Choy Hew)

1I.B.J Gene structure analysis Human genomic DNA fragments hybridizing to the cDNA clones were isolated and clonrd into pBluescript (Stratagene) for sequencing analysis. The exon-intron boundaries were identitïed by sequence alignmrnt. The transcription initiation sites for DSSI wsre detrrmined by 5'-RACE (rapid amplification of cDNA ends) (Life Technologies). Caco-I and fibroblast total cellular RNA was reverse tnnscribed with an anti-sense oligonucleotide primer (5'-TATG A AGTCTCC ATCmAT-3'). The firs t round PCR was performed in a total volume of 50 pl for 35 cycles with the primer annealing at 50°C. followed by a 'nested-PCR' with genr specific primer (5'-TCTCTAGTTCAGCTCGTAAC-3'). PCR producü were subcloned into the pCRII- TA cloning vector (Invitrogen). II.B.5 RNA in situ hybridization and Fin Amputation Mouse (CD1 strain) embryos were used in this study. Staging of the developing embryos was determined in reference to the first midday when vaginal plug was detected as E0.5.

Whole-mount in situ hybridization using a digoxigenin-labeled RNA probe and an alkaline phosphatase-coupled anti-digoxigenin antibody was performed as described (Ang and

Rossant. 1993). Paraffin sections of developmentally staged embryos were hybridized with 3%-labeled sense and anti-sense probes as described previously (Hui and Joyner.

1993). Adjacent sections wrre examincd to allow accurate localization of expression patterns. RNA in situ hybridization on wholc-mount zebrafish rmbryos was performed 3s previously described (Püschel el al.. 1997) with minor modification (Akimcnko er ul..

1991). Amputation and regrneration of adult zebrafish fins were performed as descnbed (Westerfield, 1993).

II.B.6 Subcellular Localization of DSS 1

Flag epitope tagged Dss 1 constructs were constructed as follows. For the generation of the construct with Flag on the 5' end, PCR was carried out with an annealing temperature of 55°C using human cDN.4 clone FB2S as a template with oligonucleotide primer 3'-

CCGACAGGATCCATGGACTACA,AGGACGACGATGACAAGATGTCAGAGMAA

(3-3' containing a BamHI site for cloning. the Flag epitope and 17 bp homology to the 5' end of DSSI was used as the forward primer and 5'- GAATGTGAATTCCTATGAAGTCTCCATCTT-3' with a EcoRI site for cloning as the reverse primer. Similarly for the 3' llagged construct 5'-

CCGACAGGATCCATGTCAGAGAAAAAG-3* was used as the Forward primer and 5'- TTTAGGGAATTCCTACTTGTCATCGTCGTCCTTGTAGTCTGAAGT-3' as the reverse. Fdowing PCR amplification the reaction was digested with EcoR1 and BamHI and directionally cloned inro rhe EcoRI and BamHI site of the pcDNA expression vector

(Invitrogen) following standard protocols (Sambrook er ai.. 1989). Construct DNA was transfected into COS cells using lipofectamine (Gibco-BRL) following supplied protocols. 24 hours following transfections, cultures were split and seeded onto microchamber slides (Nunc). After 48 hrs growth, cells were fixed and permeabilized with ethanollacetone Blocked with goat serum (Cedarlane). incubated with prirnary flag monoclonal antibody

M2 (Interscience). washed and incubated with a donkey anti-mouse IgG antibody conjugatsd to FITC (Gibco-BRL). Following washes. cells were treated with DAPI to stain the nucleus. and the tluorescrnce was visualized with an Aristoplan epifluorescence microscope ( Leica). and photopphcd with Kodak Ektachrome P 1600 film.

1I.C Results

II.C.1 The minimal critical region for SHFICII

In attempt to delirnit the critical region of the SHFM 1 locus. characterization of 6 additional patients. namely. D6. D7. T6. T8. T9. and Tl0 with chromosome 7 abnormalities. was carrird out by RSH. sornatic ce11 hybrid. densitometry. or microsatellite analysis, usins the physical mapping reagents shown in Figure II- 1. The clinical and cytogenetic description of al1 patients except D7 has bcen described (D6 in rd. Hudgins et al.. 1994: 16 in ref. Scherer et al.. 1994b: and Ignatius et al.. 1995; T8 in ref. Cobben et (il.. 1995). D7 corresponded to an EEC patient with an interstitial deletion of 7q21.3 deteciable by rnolecular analysis only. The deletion breakpoints of this patient were found to lie just outside the region shown in Figure 11-1. which confirmed the critical genetic region for SHFMI but failed to narrow the position for the disease locus. The proximal deletion breakpoint of patient D6 has. however, allowed the critical interval to be refined to about

1.000 kb. T9 and Tl0 correspond to balancrd translocation breakpoints in patienrs with

EEC and isolated SHFM. respcctively. Positions of the breakpoints were detemined by

FISH. Al1 9 translocation breakpoints could be mapped within a 700-kb region. and 5 of them (TI. T2.2, T4, TS, and T6) within 500 kb of each other. while the remaining 4 (T3,

Tg, T9, and T10) clustered within 100 kb (see Fig. II- 1).

II.C.2 Identification of genes in the SHFMl critical region Direct cDNA selsction, detection of evoIutionarily conserved sequences. exon- amplification. and analysis of genomic sequence were used to isolate putative transcribed sequences from genomic clones covrring the critical region. A total of 97 cDNA fragments. ?O conserved DNA segments. and 15 putative exons were characterized. Thesc clones wcre sequenccd. grouped and alizned. analyzed by nonhern blot with RNA from vürious tissues, cxamined for their correspondence to genomic DNA segments that showed sequence conservarion with other sprcies. and used as probes to screen cDNA librarics. In addition. rhr cDNA sequences were usrd to search dlavailable public databases for identity and hornology, and for coding and splicing potential. From this exhaustive analysis. only 5 'bona tïde' transcription units could be discerned (Fig. 11-1). The transcription orientation of these genss were determined by fine restriction enzyme mapping and blot hybndization analysis with different cDNA segments as probes (data not shown).

Two of the genes. DLXS and DLX6. were reported previously (Schrrer et al.. 1994b:

Simeone er (11.. 1993). The 5' end of DLX6 could not be determined because it was not representrd in any of the selected cDNA clones nor recovered from cDNA libraries made in the conventional rnanner. DNCII. a cytoplasrnic dynein intermediate chain gene was idrniified. mapping close ro the proximal cluster of translocation breakpoints. DSSZ. a novel gene with similariry to an ATP transporter protein. was the last gene identified. and the only one to be directly intempted &y translocation breakpoints. The detailed analysis of

DNCII is descnbed in chapter III. while analysis of Dm.DU6. and DSSZ is presented

in chapter IV. The remaining gene was not previously described and was thus assigned the gene symbol

DSSI (for deleted in the split handhplit foot SHFMl region). As shown in Figure 11-1,

DSSI was mapped towards the proximal end of YAC clone HSC7E57 1. and the gene appeared to be surrounded but not directly interrupted by any of the 4 rearrangernent breakpoints mapped nearby. The closest breakpoint. T6. was Iocalizrd within 10 kb of the

3'-end of the gene while 3 others. TA, T2.1. and T5. are on either side. within 100 kb (Fig.

II- 1). It should br noted that T2.I belonged to a simple SHFM patient and that the other 3. to complrx SE patients.

II.C.3 Churacterization of DSSl and its rnurine homologue (Dssl) Based on the combinrd sequences of the various overlapping clones and two apparent full- length clones, a consensus cDNA of -194 bp could be estnblished for DSSl (Fig. 11-1). A purative open rcading frame coding for 70 amino acids was identified with the first available methionine codon bcing thc initiation codon. which appcared to conform well to the Kozak consensus sequence (Kozak. 1957). A polyadenylation signal (AATAAA) was found 11 bp upstream of the poly(A) tail. The deduced DSSI polypeptide üppeared to br highly acidic with 17 residues being cither aspartic or glutamic acids (-40%). Initial database searches failed to detcçt any signi ticant idrntity with known protrins or protein motifs. but several EST clones isolatrd from a diverse array of sprcies with signifiant similariiy to

DSS 1 were idrnti fird (Fig. 11-3).

Northem blot analysis revea1r.d a single RNA species of about 500 nt in size. The DSSl

mRNA was present in al1 the adult and fetal tissues examined although the level appeared to

be the lowest in adult brain (Fig. II--13). The transcription start site was determined using

the 5'-RACE method (Frohmrin er cd.. 198s). Total RNA was isolated from both fibroblast and Caco-2 (a colon carcinoma ce11 line). and used as template for the 5' extension reaction (DSSI was known to be expressed in these cells). The extension product appeared as a discrete band on the agarose gel (Fig. II-Jb). The band was excisrd and cloned. The DNA sequences were determined for 14 independent cIones and the result indicated that the position for the major transcription start site was located at 104 bp upstream irom the putative initiation codon (Fig. 11-4~). This result was also consistent with the presence of a few minor transcription stan sites. as observed for most TATA-less promoters (Sharp. 1992).

Restriction mapping analysis showed that DSSl consisted of 3 exons spanning approximately 20 kb of gcnomic DNA and that the gene was orirnted wirh its 3'-end centrorneric to the 5'-end (Fig. II-ja). The genomic DNA regions that wrrc detccted by blot hybridization analysis were aiso sequenced. The DNA sequences at the intron-exon boundaries (Fig. II-5b) all apprared to conform to the canonical splice donor and acceptor sequences (Mount. 1982). There was also no apparent TATA box upstream of the major transcription start site (Fig. Il-4c).

To prepare for RNA i!i situ hybridization analysis in developing nice. the corresponding mouse gene. Dssl. was isolatcd by screening a limb bud cDNA librûry with the hurnan gcne. Five apparently full-Icngth mouse cDNA clones were isolated and the DNA sequence of the coding region was round to bs 90% identical to the human. differing only at the wobble positions (Fig. 11-3). The absolutc conservation between the human and mouse senes at the amino acid level suggested critical imponance of the entire polypeptide at the functional level. Figure 11-1. The location of genes in the SHFiMl critical region. Top to bottom. The position of the 8 microsatellite markers. and three YAC clones is shown (Scherer et al., 1994b) . The sentrorneric and tdomeric boundaries of this LOO0 kb region are defined by the proximal deletion breakpoint and distal deletion breakpoint in patient D6 and DS. respcctively. The thick black line represents contigs that have been sequrnced or are in the process of being sequenced. The position of the 9 translocation brenkpoints within the critical region and the cornplexity of phenotype (simple SHFM or syndrornic ectrodactyly) of each patient examined is shown. A composite of the location of the rare-cutter restriction enzymes determined by PFGE-rnappins the YACs is also shown. Thc breakpoints contained in T 1. T2.2. T4. T5. T6. Tg. T9. and Tl O were localized within the critical interval by vanous rnethods. The location and relative orientation (the arrow denotes the 5' to 3' direction) dong the chromosome. of DSSI, DLM. DL=, DSS2, and DNCII is shown. B=BssHII. C= Cial, E= EcoRI. M= Mid, N= NotI, Nr= NrttI, S= Sail. Sf= S'I. X= XlroI.

Human 1 Mouse 1

ttccttgaqgaagagtgagggttccaacttttctgcttatctgggaggtgttgggcgcgg ------g- -- -q--t- ---c------acagtcqagATGTCAGAGAAAAAGCAGCCGGTAGACTTAGGTCTGTTAGAGGmGACGAC --c -----T--A--G------C-----C----- CC-G--A--G------MSEKKQPVDLGLLEEDD 4 GAGTTTGAAGAGTTCCCTGCCGAAGACTGGGCTGGCTTAGCAT

GTCTGGGAGGATMTTGGGATGATGACAATGTAGAGGATGACTTCTCTMTCAGTTACGA4

VWEDNWDDDNVEDDFSNQLR

Figure 11-2. Nucleotide and arnino acid sequence of hurnan and mouse DSS1. Nucleotidr and predicted amino acid sequence of DSSl and alignrnent of the human and mouse sequrnce (GenBank accession numbers U41515 and U41626, respectively). Dnshed lines represent nucleotide identity. In the protein coding region the human and mouse nucleotide sequence differ in the third position only. The putative human poly- adenylation signal is boxed. The sites of the exon-intron boundriries are indicated by R. communi s A. t ha1 i ana mamIi an zebrafi sh Br. mal ayi C.el e ans S.

R-comnunis 45 A-ihaliana 46 mamli an 3 9 zebrafi sh 3 9 Y Br. mal uy i s O C.el eg ans 5 3 if: Q V.A S. po me 4 1 VAAN

Figure 11-3. Evolutionary conservation of DSS 1. Cornpanson of several DSS 1 putaiive proteins frorn a wide spectrum of animal and plant species. Identical amino acids are shaded in black. and similar amino acids are shaded in grey. R. cumniiinis ( Ricinus coninrwtis). A. [haliana (Arabidopsis rltaliano). C. elegans (Cclenorhabditis elegctrisl. Br. rnaiqi (Briigia rnalayi), and rai sequences are found in the dbEST database. with Genbrink accession numbers Tl48 13, T4 1856. D763 10, RSS305. and H3533 1. respectively. Mammalian refers to human. mouse and rat sequence. X X XXX X X X

GCGGTAGTGACGGTGGCGmCCTTGAGWGAGTGAGGGnCC Translation stan

Figure 11-4. Analysis of the human DSSI transcript. (A) Northern blot analysis of DSSI expression in adult and fetal tissues reveals a sinale transcript of 500 bases in rnost tissues examined. (B)Blot-hybridization experirnent of?'- RACE products of fibroblast and Caco-2 RNA probed with human DSSI cDNA A single band of 356 bp is observed in both rxperimental samples. whereas the negative (-ve) control (H20). the sample without reverse transcriptase ( fibroblast-RT). and the sarnple without terminal transferase (fibroblast-TdT) show no signal. .A 273 bp product is observed when a nested 5' primer is used instead of the anchor primer (fibroblastlF1) in the PCR as a positive control for first strand cDNA synthesis. (C) The DNA sequence of 14 independent 5'-RACE products aligned with the human genomic DNA sequence upstream of the initiation codon. 'x' indicates the observed transcription start site of each clone, the rnosr comrnon being 10J nucleotides upstream of the initiation ATG. Figure 11-5. Genomic structure of human DSSI. (A) Restriction map of cenomic region encompassing DSS 1. position of exons are shown below the map. and position of two overiapping cosmid cloncs from which the map was derived are shown above. E=EcoRI, Sa=SalI, X=XbaI, Xh=XhoI, H=HindIII, St=StuI. (B) Characteristics of intron exon boundaries for each exon. Primers tlanking the exons for mutlitional annlysis include: exon If. 5'-CAAGTCTCTATGGTAGCGTCAGC-3' Tm=6O.S. and exonlr. 5'-GAGGCCTGAGTC ACC GTT C-3' Tm= 60.8, product sizs=243 bp; rxon2f. 5'-AATTGTGA GTGTTCAGTAGCAAGC-3' Tm= 59.9. and exon2.r. 5'-TGhACATG.ACAATAA AXXTCTGGC-3' Tm=60.7. product size=233 bp: exon3. f. 5'-TCTGGCATAG CTCATTTATTTG-3'. Tm= 58.1. and exon3.r. 5'- TCTGCCCTAGAATGTATTAAGCA-3' Tm=58.1. product six= 200 bp.

II.C.4 Expression of Dss 1 during development

To investigate the function of DSSl and its relationship with SHFM and SE. we examined

the expression pattern of the gene using RNA in situ hybridization with whole mounts and

sections of developing mouse ernbryos and newborn anirnals. For these studies, the murine cDNA clone. including the 3'-untranslated region. was used to generate sense and

anti-sense riboprobes. As revealed by the results of whole mount studies. expression of

Dssl could be detectrd at V~~OUSstages of embryonic development. At embryonic day 9.5

(E9.5). Dssl was found to express strongly in the mesenchyme of the first branchial arch

and the frontonasal prominencc (Fig. II-6a). As facial drvelopment progressed. the expression was latcrally restricted in the masillary and mandibulary prominence of the firsr

and second branchial arch (data not shown). consistent with a role in the development of

the bones of the lowsr face and jaw. By E 12.0 the expression üppeared to be restricted to

the region of the prospective tooth buds (data not sh«wn). In addition to the facial

primordium. expression of Dssl was dciected in the early genital tubcrcle. but it was not followed aftcr E 12.5 (data not shown).

The results of ernbryo and newborn section analysis also reveûled thlit the expression of

Dssl was widespread at low levels (data not shown). thus in sood agreement with the

nonhrrn blot results. In addition. Dssl was tound to be strongly expressed in the dermis

of new born mice (Fig. II-6b-d). The latter expression pattern seemed relevant to the

ectodermal dysplasia phenotype in the syndromic ectrodactyly patients.

II.C.5 LocaIized Dssl expression in developing limbs

The spatial and temporal pattern of expression of Dssl was examined in the developing

mouse limb bud by using whole rnount RNA i~isirii hybridization (Fig. 11-7). At E9.5.

when the first signs of limb bud outgro~vthwas evident (Wanek rr al.. 1989). Dssl

expression could be detected ubiquitously throughout the mesenchyme of the structure but not in the ectoderm (data not show). As the limb bud elongated. Dssl was no longer expressed in the proximal mesenchyme. At El 1.0. in the fore limb bud. expression was only apparent in the distal mesenchyme and the peripherai regions of the media1 mesenchyme (Fis. II-7a). It was presumed that. at this stage. the mesenchymal cells of the central core of the limb bud were condensing to form the stylopodial or zeugopodial elernents (Wanek er cil.. 1989). In El23 embryos. Dssl expression in thc fore limb was found to be further restricted as the mesenchymal condensations occurred in the initial stages of digit formation (Fis. Il-7b). At this stage. the Dss I peappeared active in the inierdigi ta1 mesenchy me. but not in the condensing mesenchyme.

As limb developrnent proccrded. the expression of Dssl in the core of the interdigital mesenchyme dissipated in a proxirnodistal direction and it became restricted to the mesenchyme just outside the perichondriurn of the devrloping digits by E13.25 (Fi:. II-

7c). At this stage in the hind limbs the bordcr of the staining was more proximal than in the fore limbs obscrved. This difference may be a transient change in exprcssion pattern seen in both the fore and hind limb. or it may be hind limb specific. It may ais0 be just experirnentai variation. Finally. while the digitûi condensations were undergoing segmentation during the formation of phalanges (nt E13.5). Dssl outlined the pattern of the

prospective bones (Fig. II-7d). Therefore. the Dssl expression pattern. which was similar.

if not identical. in the fore- and hind-limb buds with respect to their relative stage of

dcvelopment. was consistent with that of a gene with a role in the early specification of digit

formation (Roberts and Tabin, 1994) Figure 11-6. Expression of Dssl during mouse embryogenesis. (A) Whole mount in situ hybndization of E9.5 whole embryo showing suong expression in the mesenchyme of the fint branchial arch (arrow) and the Frontonasal prorninence (arrow) with some weaker expression in ihe cephalic regions. (B-D) Section in situ hybndization of transverse section of new-born limb. Panel B= bright field, panel C= dark field sense probe (negative control). and panel D= dark field antisense probe reveals the strong expression of Dssl in the dermal layer of the skin. II.C.6 Cloning and analysis of Dssl from zebrafish.

Amino acid sequence conservation between human and mouse Dssl was 100%. TO determine if this level of identity was common throughout the vertebrate species, a zebrafish cDNA library was screened using a full length mouse cDNA as a probe at low stringency. Apparent full length clones were idrntified that showed high similority to the

murinc and rnouse Dssl cDNA sequencr. The predictcd amino acid scquence of thess clones was 974 identical to the mammalian protcin differing in only 2 residues (Fig. 11-3).

This high degree of sirnilarity across vsrtebrate species argues for a critiçal role of Dssl in the function of thc muliicellular organism.

To funher analyzc Dssl function in zebrafish. whole rnount RNA iii siril hybridizlttion was carried out with zebrafish rmbryos. h low levsl expression of Dssl \vas detected

throughout most of the embryo. whilc higher levels wherr found in the pectoral fin buds. the siIl arches. the dorsal midbrain. and the rycs (Fig. II&). Expression in the 18hr fin

bud was bund to br posterior and distally restricted. but absent from the thfold (The fish

equivalcnt of the XER) (Fig. II-Sb).

Zebrafish. likc other tkh. regencrate their fins after injury and rersrablish normal patteming

(GCraudie and Singerm 1990). To begin to determine a possible role for Dssl in fin

regeneration. the expression of this gene was analyzed in adult fish. following amputation

and regeneration of caudal fins. Dssl was in fact found to re-express as early as I days

following amputation at the tip of each fin ray stump suggesting a role in early blasiema

formation (data not show). Stronger expression was found 4 days following amputation

at the distal aspect of the regenerating fin (Fig. II-8c). which was found to localize to the

rnesenchyrnal blastema cells by sectioning (Fig II-Sd). II.C.7 Subcellular localization of Dssl Since the putative DSSl protein is novel. it is not possible to speculate about the biochemical function of this protein. Therefore. as an initial step towards understanding the function of this protein. Dss l -FLAG epitope tagpd constructs were generated to detemine the subcrllular localization of the protein. The Flag-Dss 1 constructs. exprrssed under the control of a CklV promoter. were transfected into COS cells. Following primary anribody labsling. tagged protein was visualized using a secondary antibody conjugated to FITC.

Borh 5' and 3' ragged constructs were frcquently found in the nucleus. with diffuse cytoplasmic staining (Fie. II-9a.b). The frequency of exclusive nuclear staining however. was not 100% and at varying lcvels equal distribution of cytoplûsmic and nuclear stain could bc found (Fig. II-9c). The proportion of mostly nuclear staining varied with the two constmcts. suggesting ihat the 3' Tag may hinder nuclear localization. Figure 11-7. Whole mount RNA in situ hybridization of Dssl during mouse limb bud morphogenesis. (A) E 11 .O fore limb- shows Dss 1 expression in the distal and antenor and postenor media1 mesenchyme. (B) E 12.5 fore limb- shows Dss l expression in the interdigital mesenchyme but not in the condensing mesenchyme. (C) E13.5 hind limb Dss 1 expression is further resuicted to the mesenchyme surrounding the perichondnum of the digit blastemas. It is unclear if the more proximal expression observed in this hind limb is specific to hind limbs or a transient change in the expression pattern. (D) E 13.5 fore limb- the expression signal is resaicted to the mesenchyme surrounding the future skeletal element of the individual phalanges. Figure 11-8. Analysis of Dssl expression in the developing and regenerating Zebrafish. (A) Whole mount RNA in situ of a 36 hi- zebrafish embryo. Arrow shows expression in the developing brain, expression is also detected in the eyes and tail mesenchyme. (B) expression of Dssl in a 48hr zebrafish pectoral fin bud. Expression is found in the rnesenchymal cells (me) but excluded from the finfold (ff'). (C) Arrow points to Dssl expression in the distai mesenchyme of the fin ray blastema 48 hrs following amputation. (D) sectioning on regenerating fui in C shows expression confined to the mesenchymai cclls beneath the wound epithelium (we). Figure 11-9. Subcellular localization of DSS1. (B) Localization of 5Wag tagged DSS1 to the nucleus of COS cells. (D) 3'Flag tagged DSS 1 showing nuclear localization with diffuse cytoplasmic signal. Panels A and C are the DAPI stained photograph of B and D, respectively. 88% of cells examined transfected with the 5' tagged construct showed nuclear staining (n=108), while only 40% of cells transfected with the 3' tagged construct showed nuclear staining (n= 138). 1I.D Discussion

Through the molecular characterization of the genomic rearnngements found in a group of patients with SHFM and SE. we have identified a novel gene that appears to play an important role in limb, craniofacial. skin. and genitourinary developrnent. The gene.

DSSI. which encodes a highly conserved. acidic protein. maps within the criticai region and is closcly surrounded. but not directly intempted. by the chromosomal translocation breakpoints round in the patients. Based on RNA itz situ hybridization analysis with the murinr homologue. Dssl, the genr appears to express predominantly in the limb and facial primordia during days 9.5- 13.5 of mouse development. It is also expressed strongly in the demis of newborn rnice, the exly genital bud and possibly the tooth primordium. Thus. a

reduction in the expression of this gene dunng human embryogenesis may not only explain the phenotype observed in SHFMI patients. but also somc forms of syndromic cctrodxtyly including EEC.

1 have designed oligonucleotide primers flanking the exons of DSSl (Fig. 11-5) to facilitate the search for mutations in SHFM patients. A11 3 rxons have bcen scanned in 60 sporadic

SHFbII patients but. so far. no mutation could be detected (J. Evans and P. Charlion.

unpublished data). As stated previous studies showcd that polyrnorphic markers located

within the SHFM 1 critical region failed to demonstnte linkage to the phenotype in sevenl

large SHFM families (Palmer et al.. 1994: Mannoni et al.. 1994: Gunieri et al.. 1994). It

is possible that the SHFM and SE phenotypes are caused only by deletion or transloca:ion

affrcting the SHFM 1 locus. Therefore. searching for point mutations in DSSl may not be

an effective strategy to demonstrate the role of these genes in limb development. It is difficult to predict the endogenous function of DSSI. However. the expression pattern of Dssl during limb development is similar to rhat of Gli3 which may provide insight into its function. Gli3 encodes a zinc finger protein whose reduced expression is known to cause extra-toes in mouse (Hui and Joyner. 1993) and GCPS in humans (Vortkamp et al.,

1991). GLI3 is thought to play a role in prograrnmed ce11 death as its disruption results in an increase in the number of digits and the lack of di@ separation. Similarly. the SHFM 1 phenotypc may be due to a loss of the overall convol of programmed ce11 death in which an increase could result in the loss of digits and a decreasr could result in the fusion of digits.

Alternatively. it may be proposed that DSSI is required to maintain cells in an undifferentiated state or that it promotes the proliferation of these cells. In suppon of this assumption. Dssl expression is present in regions of rapid ce11 growth (limb bud, branchial arch. genital bud. skin) and cxcluded from regions of ceIl differentiation (e.g., digital condensations). In addition. Dssl expression is widespread early in the developing embryo (not shown) and as developrneni progresses and the fate of ceIl lineages becorne determined Dss 1 expression dissipatrs and is found in late drveloping tissues that remain in an undifferentiated state. Although further studies are required to delineate the biological function(s) of DSSl it remains an excellent candidate for study of limb development.

Expression nnalysis of Dss I in zebrafish funher serves to implicate this gene in embryonic development. particularly of the limbs. The expression of this pene is relatively similar in both the mouse limb bud and the fish fin bud. In both cases expression is restricted to the distal bud but excluded from the AEFUfinfold. Furrhennore. expression appears to be more prevalrnt in the posterior mesenchyme. panicularly in the fin. This similar expression pattern in both species argues for a conserved role for Dssl in vertebnte limb development.

Funher analysis in the zebnfish shows that Dssl is exprsssed in the regrnerating adult tail fin. Studies have shown that the molecular mechrinisms of limb regeneration are not necessarily the same as those found during embryonic development (Akimenko er al.,

1995). Dss l is expressrd in the mesenchymal cells of the developing and regenerating blastemas and thus may play a similar role in these two processes. Nevertheless gene regulation of Dssl is likely different during development and regeneration since no expression was found in the developing caudal fin at the stases sxamined. yet it is expresssd during regencration.

While the biochcmiclil function of Dssl is not known. it appcars that ttiis protein is largely. but not exclusivçly. nuclear localized. It is possible that subcellular localization of Dssl may be influenced by thc state of the cell, such as place in the cell cycle. Alternatively. nuclear localization mriy be a function of its activation by various cxtracellular or intracellular signals, as secn with various signal transduction plithways such as TGFP

(Hoodlsss er d..1996) and hedgehog ( Aza-Blanc er cd.. 1997). Interestingly. Dss l appem to laçk a conscilsus nuclear localization signal (Nigg. 1997). It is possiblc that

Dss 1 contains a novel nuclcar localization domain not previously cncountcrcd: an alternative hypothrsis is thot nuclear accumulation of Dssl is achieved through its interaction with a second protcin that contains a nuclear localization signature. The identification of ESTs with a hizh degree of similarity to DSS 1 from nurnerous specirs across diverse phyla including plants. argues that this gcnr likely plays a role in a basic cellular process comrnon to ail cukaryotrs. The arnino acid sequence of these proteins. howevrr. do not possess any obvious peptide domain that provides insight into the function. While there does appear to be some arnino acid residues conserved across species. the region of these proteins that show stretches of identity appear to be localized to the regions of acidic amino acids.

Unforrunately, at this timr. no known functional dornain has been ascribed to these acidic

stretc hes. Chapter III

Cloning and characterization of two cytoplasmic dynein intermediate chain genes in mouse and hurnan

Attribution of work:

The ori_einal identification and cloning of human DNCIl was performed by Dr. Stephen Scherer and Dr. Johanna Rommens. The determination of genomic structure was panially detennined by Dr. Michal Prochazka (Arizona). Nonhern blot analysis of both DNCII an DNClt was performed by David Sinasac, a graduate student in the laboratory. Al1 the other data presented was obtained through experiments carried out by the candidate.

The data presented in this chapter has been published in pan in :

Crackower, MA.. Sinasac. D.S., Xia. J.. Motoyama. J.. Prochazka. M., Rommens. J.M.. Scherer. S.W.. and LX. Tsui. Cloning and charac terization of two cy toplasmic dynein intermediate chah genes in mouse and human. Genontics. In Press. 1II.A Introduction

Dynein is a large rnultisubunit protein complex involved in microtubule-based rnotility during a diverse array of cellular functions (Holzbaur and Vallee, 1994). There are two classes of dyneins that have been identified. The first class is comprised of axonernal dyneins that genente the sliding force of adjacent microtubules required for the beating of cilia and flagella (Witman, 1992). The second class. cytoplasmic dyneins. are required for the inovement of membrrtnous organelles within the ce11 towards the minus end of microtubules (Schroer. 1994). Specifically. cytoplasmic dyneins are thought to be responsibie for retrograde transport in neuronal axons (Paschal and Vallee. 1987). as well as the positioning of lysosomes. endosornes (Lin and Collins. 1992). the Golgi apparatus

(Conhesy-Theulaz et al.. 1993), and the nucleus (Li et al.. 1993: Eshel et al.. 1993). In addition. thcy are required for proper chrornosoma1 migration during mitosis (Pfarr et al..

t 990).

Cytoplasmic dyneins contain two 530 kDa heavy chains which genente force and catalytic activity (Shpetner et al.. 1985). Three intermediate chains (74 kDa) are also present in each cytoplasmic dynrin complex and are thought to direct the complex to the proper cellular site

(Paschal et 31.. 1992). Light intermediate chains (55-59 kDa) found only in cytoplasmic dynein complexes are thought to play a role in regulating dynein activity (Hughes et al.. 1995). Finally. light chains (8-25 kDa) have been identified but as yet have no function

ascnbed to them (King and Patel-King. 1995).

OnIy recently has there bern some elucidation of the function of dyneins at the level of the

rnulticellular organism. For exarnple, partial loss of function mutations in Drosophila

cytoplasmic dynein light chain (ddlcl)resuit in rnorphogenetic defects in bristle and wing

development as well as oogenesis (Dick et al.. 1996). Complete Ioss of function of this gene results in embryonic lethality and an apparent increase in apoptotic cell death (Dick et al.. 1996). Alterations in the Drosophila cytoplasmic dynein heavy chah gene Dhc64C result in defects in oogenesis (Gepner et al., 1996); this gene is required for spindle orientation during gemline cell divisions and oocyte differentiation (McGrail and Hqs, 1997).

The function of dynrin senes in mammals has also begun ro be dissected. A knockout mouse model for the cytoplasmic dynein heavy chain gene has been generated. Thrsr mice are characterized by early embryonic lethality. a disorganization of the Golgi complex. impropcr distribution of endosornes and lysosomes. and a defect in ce11 proliferation (Harada et al.. 1998). Furthemore. in mammals. dyneins have long been irnplicated in defects in immotile cilia syndrome (.4fzelius. 1976). Frequently associated with this abnomality is heterotaxia (or situs invcrsus) (Afzelius, 1995). Recently it has been show that a heavy chain axonemal dynein is mutaled in the Niverslis iiscrnmi (iv) mouse. a model of situs inversus. thus confirming this causal relationship (Supp et al.. 1997).

Therefore. it is becoming clex that dyneins are important for proper ce11 morphogrnesis and differentiation durhg ernbryonic development.

As part of our effort to determinc the genrtic etiolo_pyof split hand/split bot malformation mapped to human chromosome 7 (SHFM 1) (Scherer et al.. 1994a. 1994b: Crackower et al.. 1996). we have identified a human cytoplasmic dynein intermediate chain gene (DNCII)within the 1 Mb SHFM 1 critical region. SHFM 1 is a clinically and genetically heterogeneous lirnb abnomality. characterized by missing digits, fusion of remaining digits and a deep median cleti (Temtamy and McKusick, 1978). Our previous studies showed ihat a syndrornic form of split handhpiit foot malformation also mapped to the identical locus as SHFM l at iq2 1.3-q??. 1. At least 2 patients with the syndromic forrn of this disease have been descnbed as having a form of heterotaxia (Genuardi et al.. 1993). Given the known involvement of other dynein genes in various developrnental abnormalities,

DNCll could be a candidate for SHFM 1.

This chapter describes the cloning of DNCII and its mouse homologue Dncil, as well as another highly relatrd mouse gene called DiciZ. It is funher show by RNA in sirir hybridization analysis of these genes during embryogenesis that D~tcilhas a highly restricted expression pattern. compared to a more widespread pattern of expression found for LhtciZ. No Dncil cxpression could be detected in the developing limb. thus rxcluding this gene as a candidate for SHFM 1. On the other hand, DrzciZ was found to have a dynamic cxpression pattern during mouse limb bud development. suggesting a role during digit formation. The presence of a dynein gene in the SHFM 1 critical region has potential implications for the abnomal inheritance patterns observed in SHFM

I1I.B klaterials and Mcthods

IILBJ Gene Identification, cDSA Clones and Libraries

Direct cDNA seleciion w3s prrformed on the YAC clone HSC7E1170 as previously described (Crackower sr al.. 1996). DNA sequence analysis of one of the direct selected cDNA clones. El 170cd53. revealed strong similarity to the family of intermediate chain dynein genes. In order to identify the corresponding full-length transcript the cDNA fragment was initially used to screen o cDNA librq made from a colon carcinoma cell line

(Caco-2). and the DNA sequence from the resulting clones was used for 5'-RACE rxperirnents. Mouse D~iciland DnciZ were isolated with El 170cd53 through hybridization screening of an adult brain cDNA library and a h-EXlox mouse embryonic cDNA library

(Novagen). respectively. Low stringency hybridization conditions were used: 55'C hybridization in Church and Gilben Buffer (Sarnbrook et al.. 1989); washing was carried out at SaCto a final srrinpcy of 0.h $SC, O. 1% SDS. II.B.2 Genomic Structure of DNCIl Cosmid clones from the Lawrence Livermore chromosome 7-specific library and Pl- derived ani ficial chromosome clones were identified with the DNCI I gene probe. The cornpiete list of 21 cosmids and 9 Pl clones can be found at our website: http://www.$enet.sicklids.on.ca. Cosmid clones 57d5. 228d2. and 737d7 were used for determining partial exon-intron boundaries by cycle sequencing (Arne rs ham). Sequences were analyzed on an AB1 Prism 373 automated scqusncer (PE Applied Biosystems Inc).

Primcrs were dssigned from cDNA sequencr. The size of the individual introns were estimated by agarosr gel elcctrophoresis of PCR-ampiified products. or determined by direct sequencing of short inlrons. Moreover. the structure of a large portion of the gene was determined and confirmed by cornparison to genomic DNA sequence available at the

Washington University Gcnome Sequencing Ccnter (http://genome.wustl.edu/gschlast/ blast-scrvers.html).

III.B.3 Northern blot analysis Total RNA was isolated using the guanidine isothiocyanate extraction method (Chirgwin et al.. 1979) from various mouse tissues ohined from adult CD- 1 outbred mice and embryonic day 12.5 embryos (E17.5) (Charles River). Northern blots were generated using standard procedures (Sambrook et al.. 1989). 10 pg of total RNA were loaded into each lane.

III.B.4 RT-PCR

RNA was isolated by the samr method used for Northern blot analysis. 5 pg of total RNA from mouse adult brain. testis. and liver was used for first strand cDNA synthesis. with 200 Lr Superscript II reverse transcriptase (Gibco BRL) and random hexamen, in a total volume of JO pl. The reaction mixture was incubated for l hr at 37°C. In al1 cases a minus reverse transcriptase control was used. One rnicroliter of the reaction was used for 32 cycles of PCR at 94'C for 45 sec. 60aCfor 45 sec. and 72'C for 1 min. The sequence of the prirners used were IC LF2. 5'-AGCATTGGCATATCACCG-3'and IC LR2, 5'- GCTTGTGCAGTCGTCTCC-3'. PCR products were resolved on a 3% agarose gel for photography or on a non-denaturing 8% poly-acry lamide gel. Individual bands were cut out of the gel, and DNA was extracted by the cmsh and soak method (Sambrook et al..

1989). Purified DNA was then sequencrd usin= a Thermosequenase kit (Amersham) and prirners IC 1-F? and IC 1 -RI.

III.B.5 RNA in situ hybridization

Mouse CD1 (Charles River) embryos were used in this study. Staging of the developing embryos was drtennined in reference to the first midday when vaginal plug was detected as E0.5. Whole-mount in silu hybridizarion using a digoxigenin-labeled RNA probe and an alkaline phosphatase-coupled anti-digoxigenin antibody was performed ris described (Ang and Rossant. 1993). For these studies partial cDNAs were used to generate anti-sense riboprobes. Clone Dricil- 1 1 -2 rncompassing nucleotides 1 3- 1663. was used for the Dncil probe and clone DnciZ-4- i en compas sin^ nuclcotidrs 1384-2367. was used for the DnciZ probe. Frozen sections of drvelopmentally staprd embryos were hybridized with dipoxigcnin-labeled anti-sense probes as described prrviously (Schaeren-Wierners and Grrfin-Moser. 1993). Adjacent sections were rxarnined to allow accurate localization of expression patterns. 1II.C Results

III.C.l Cloning of human DNCII

To identify candidate genes for SHFMI. direct cDNA selection was conducted on YAC clones covering the cntical region. A cDNA fragment. E 1 l7Ocd53. with sequence identity to rhc. rat intermediate chain cytoplasmic dynein 1 sene (Dncil. also known as Dncicl)

(Paschal et al.. 1992) was identified. Based on the overlapping sequence of various cDNA clones. a c«mposite gene sequencr for DNCllo of 2954 bps was established (Fig. 111-2).

A putative open rcading frame of 645 amino acids was idrntified with the first available

methionine as rhe initiation codon. A polyadenylation signal was found at a position 13

bps upstream of the poly(A) tail. Three of the right isolatcd cDNA clones characterized did

noi includr nucleotides 41 1-461 of the consensus scquencc which would translate into an

open rcading franle of 628 arnino acids. suggestin= the prcsence of alternative splicing

Fi.-3) For discussion we will refer to the larper transcript as DNCIla and the

smaller alternative transcript as DNCIlc. Sequence cornparison of the predicted protein

showed that human genr product of DNCllri was 99% identical to rat Dricilu indicating

this gene was likely the human onholog.

Restriction rnapping of genomic clones and DNA sequcncç analysis indicnted that DNCIl

consisted of 17 exons spûnning approximately 160 kb with the 3' end of the gene towards

iqter (Fig. 111-1). It appeared that the alirrnatively spliced DiVCllc transcript arose as a

result of a cryptic splice-acceptor site loclitcd within rnon 4 (Fig. III-2 and III-3A). The

genr was not round to be interrupted directly by any of the SHFMI inversion or

trmslocation breakpoints (data not show). A breakpoint close to the gene (T3)was a

familial rearrangernent in which some individuais with SHFMI also exhibited situs

inversus (Genuardi et al.. 1993). The rnapping of human DNCll to the SHFM locus on

chroniosorne 7 supports the syntenic reiationship this chromosomal region has with mouse chromosome 6. where mouse Dncil and other genes from this region have been rnapped

(Vaughan et al.. 1996; Crackower et al.. 1997).

III.C.2 Isolation of rnurine Dncil and Dnci2

In order to characterize the DNCll gene during development we chose to identify the rquivalenc mousr grne by low stringency hybridization screening of a mousr srnbryonic cDNA library and a inouse adult bnin cDNA library. The D~icilgene \vas identiried from

the adult brain cDNA library and DNA sequence analysis revealed a composite cDNA

sequence of 2586 bps with a putative open reading frame of 618 amino acids (Fig. III-3B). This isoform was most similar to DNCIIC frorn hurnan and thus named accordingly.

Drrcilc was found to be 97% and 98% identical at the amino acid level to the human and rat

orthologs. respectively. Sequence analysis also identified a second isoform of Dncil

(tentativrly named D~icile)in which amino acids 106-175 of D~rcifcwsrc absent (Fig. III-

3A). Alternative splicing at this same position had also been reponed for the rat ortholog

(Paschal et al.. 1992). To dctermine if additional splice variants similar to thosc found in

rat and hiiman are also present in rnice. RT-PCR was performed with adult mouse brain

RNA whrre the region of alternative splicing was amplified. The presence of J mouse

brin Dm=ilisoforms wrre identitïcd. with varyinz contribution (Fie. III-JA). Sequence

analyses of these bands showed that the predominant species was isoform Dtrcilc. As

well. Dmile was also present. As expected. a splice variant homolo~ousto Dncila was

found. but in a relatively low amount (Fig III-3A and 111-414). howevrr. isoform Dncilb was not present in mouse brain. The founh isoform present from mouse brain. Dncild.

was found to be a novel isoform not previously described (Fig III-3A and 1113A). Figure 111-1. The localization of ONCI1 within the SHFMl critical region. (A). Top ro bortom. The SHFMI smallest region of overlap (SRO)shown as a shaded bar has bren described previously (Crackower et al.. 1996). Three YAC clones (Scherer et al.. 1994) in the SHFM 1-critical interval is represented by thick black lines. The position of the 7 translocation breakpoints within the critical region of each patient examined is shown. The clinical and cytogenetic characteristics of these patients have been described (Crackower et al.. 1996 and references therein). The location and relative orientation (the arrow dsnotrs the 5' to 3' direction) dong the chromosome. of DSSI. DLX6. DLXj. DNCIl. and PDKJ is shown. The position of the 9 microsatellite marken are shown. (B) The imnicdiate region around DNCll is rnagnified below to show the local PAC and cosmid contig and the position of the 17 cxons of DNCII. Marker D7S527 is found in intron 11 of DIVCII. SHFMI SR0

----- cos 122e 1O cos57d5 cosl20dll cos237d7

1Okb .. . Figure 111-2. Nucleotide and amino acid sequence of human DNCII. Consensus nuçleoiidr sequence of hurnan DNCIl (Genbank accession AF063738) with predicted arnino acids sequence beneath the respective codons. The coding region is in capital lrtters whilr 5' and 3' UTRs are in lower case. Polyadenylation signal 13 bp upstrearn of the poly-(A) tail is in bold. Positions of introns within the gene are shown by mows and their respective intron number. A cryptic splice-acceptor site within exon 1 is shown in bold. The AGfr consensus slice site is underlined. A thick line is placed above the nucleotide sequcnce that is absent in the alternatively spliced DNCIla.

Screening of the mouse embryonic cDNA library did not result in any positive clones for Dncil; however. nurnerous clones for a highly related gene Dnci2 were recovered and

sequenced. The DnciZ gene was found to contain a 2486 bp cDNA with a putative open

reading frame of 612 amino acids (Fig. III-3B). Mouse Dnci2 was found to bc 726 identical to human DKII at the nucleotide level and 77% at the amino acid bel. Dicil

and Dncil gcnss are 76% identical over the entire length of the predicted proteins (Fig. III-

3B). Dncil had previously been identitied in nt (Vaughan and Valler. 1995) and sequence

cornparison indicated thcre was 998 amino acid identity with the mouse gene.

Alternatively spliced isoforms for nt Duci2 have also bren reported (Vaughan and Valler.

1995). but only a single cDNA species was identified in our mouse library screening

experimcnts (data not shown). A provisional map location for murinc Duci2 hns bren

reponed on chromosome 2 (Vaughan et al.. 1996).

IiI.C.3 Transcript analysis of Dncil and Dnci2 Nonhern blot analysis of Dticil and Dnci2 revealed expression profiles sirnilar to that

reponed for the rat hornologs (Fig. 111-4~)"'.&au~hanand Vallee. 1995). hcil expression

was prirnarily resrricted to brain and weak expression was found in the testis. Dncil

exhibitcd ri widespread expression profile found in al1 tissues tested with strong sipls in

adult brnin. total embryonic tissue. and testis. Intrrestingly. this analysis revealed the

presence of a unique isoform in testis for DtzciZ not found in nt (Fis. II11B).

Since rhe band obsrwed in mouse testis for Dncil was very weak, RT-PCR was employed to confirm the expression of this sene. The results clearly indicated that this gene is

enpressed in testis as a single isofonn. Dncile (Fig. III-1A). Figure 111-3. Comparative species analysis of alternative splicing in Dncil and protein sequence alignment of mouse Dncil and DnciZ. (A) Amino acid sequence alignment of rat. human and mouse D~icilin the region of alternative splicing. Rat Dlzcilu. human DDCIla and mouse Dncilo sequencrs are identical through this region except for the absence of two bp at position 114 in the rat. Two regions of alternative splicing are common between at least iwo species. The second region of alternative splicing in nt and mouse are identical except for an additional arginine residue in the mouse. Mouse isoform Dtlcild. identified by RT-PCR,has a novel splice-accrptor site which likely resides in exon four. (£3) Amino acid sequence alignment of mouse Dticil and Dnci-1. The two proteins are found to be 76% identical and 88% sirnilar in sequence. Complete cDNA sequence can be found in Grnbank. Accession AF063229. AF063230 and AF06323 1 for mouse Dncilc, Dncile. and Dtici2. respectiveiy. Vertical linrs indicaie residue identity. Double and single dots indicate relative amino acid similarity. Mouse nomt.nclature denotes the mouse genes as Dticicl and DncicZ. aa. 71 I rat Dnci la EPPLVQPLHF LTWDTCYFHY LVPTPMSPSS KSVSTPSEAC SQD..DLCPL TRTLQIIiIM'DP SJLQLQSDSE LCRRLNKLCV rat Dncilb EPPLVQPLHF LTWDTCYFHY LVPTPHSPSS KSVSTPSEAG SQD..DLGPL TR...... RLNKLCV huntan DNCIla EPPLVQPLHF L"i7CDTCYFHY LVPTPMSPSS KSVSTPSEAG SQDSGDLGPL TRTLQWDTDP SVLQLQSDSE LGRRLHKLGV hulaan DNCIlc EPPLV ...... PTPMSPSS KSVSTPSEAG SQDSGDLGPL TRTLQWDTDP S'JLQLQSDSE LGRRLHKXV mause Dncila EPPLVQPLHF LTdDtCYFHY LVPTPYSPSS KSVSTPSDAG SQDSGDLGPL TRTLQWM'DP SVLQLQSDSE LGRRLHKLGV mouse Dnci lc EPPLV ...... PTPHSPSS KSVSTPSDAC SQDSGDLGPL TRTLQWDTDP WLQLQSDSE LGRRLHKLW mouse Dncild EPPLV ...... STPSDAG SQDSGDLGPL TRTLQWDTDP SVLQLQSDSE LGRRLHKLGV mouse Dncile EPPLV ...... PTPKSPSS KSVSTPSDAC SQDSGDLGPL TR...... RRLHKLCV

Dnc i 1

Dnc i 2

Dnc i l

Dnc i 2

Unc r 1

3nc :2 Dnc i i

rjnc :Z

Dnc i i

3nc L 2

Dnc ::

Dnc :Z

Gnc :?

Dnc :2

Dnc ::

5nci 7

Dnc i 1

Dnc 12

Dnc i l

DnclZ

Dnc i l

Dnc 12

Dnc i i Dnci2

Dnc i i Dnci? Figure 111-4. RT-PCR analysis of Dncil in mice and Northern blot analysis of murine Dncil and Dnci2. (A) RT-PCR amplification of Dncil from adult mouse tissues identify four altrrnatively spliced isoform prrsent in brain. and one isoform present in testis. Bands were excised from the gel and sequenced. Putative amino acid sequence of the isofoms are shown in figure 3. Presence of an amplification producl confirms the expression of Dnd in mouse testis. whereas the non-expressing tissue. liver. serves as a control. In each case a minus reverse transcriptase control (-n)was used to show no genomic DNA contamination. (M) rrpresents a 123 bp marker lanr. (B) Nonhcrn blot analysis shows that Dncil is cxpressed in brain and weakly in testis. DnciZ is found to express in al1 tissues rxaminrd with strongrsi signal in brain and testis. Note. that there is an additional isoform (2600 nt) for DnciZ in mouse testes not present in other tissues. A control O-actin probe is used to control for loading. The strong hybridizntion signal in hewt reprcsrnrs a speci fic O-actin iso forrn. ---braln testis liver M RT -RT RT -RT RT -RT -IL.-- III.C.4 Analysis of expression of Dncil and Dnci2 during mouse embryogenesis

To characterize the rnouse genes further. the spatial and temporal expression pattern was examined using RNA in situ hybridization with whole mount and sections of developing mouse embryos. Expression patterns were analyzed from embryonic day (E) 10.5 to

E13.5.

As was observed using Northern blot analysis it was found that Dticil has a highly restrictrd pattern of expression during mouse embryogenesis. At ElO.5. predorninant expression of Dncil wns found in the dorsal root ganglia (DRG) (Fi:. III-5A and C). Of relevance to our studies of SHFM 1. there was no expression observed in the developing limb bud (Fig. III-5A). which would rxcludc this gene as a poteniial candidate for this disease. Analysis of expression on sectioncd cmbryos revealed additional sites of expression at El 1.5 in the prospective motor nrurons of the ventral nruriil tube. the muscle precursors of the myotome. the sympathetic ganglia. and mesenchymr in the proximity of the notochord (Fig. III-5C). Weak expression in migrating myoblasts was also observed

(data not shown). Later during embryogenesis. expression of Dticil remained highly restricrrd to the DRG and skeletal muscle (Fis. III-6A and D). with wrak expression being detecrsd in bone precursors. At El 3.5 very specific expression ol' Duc-il was found in the neural cortex of the forebrain, while the rest of the central nervous system (CNS) lacked drtectable expression (Fig. IIL6A and C).

The expression of DnciZ durhg ernbryogenesis was found to be more widespread. At

E 10.5 expression was detected in the branchial arches. limb buds, frontonasal prominence. brain. neural tube. and DRG using whole mount analysis (Fig. III-5B). Srctioned in sitic analysis indicated that branchial arch and frontonasni prominence expression was found in both the epithelial and the mesenchymal cells (Fis. III-5E and not shown). At El 1.5 Dnci2 was expressed throughout the neural tube with strongest expression in the ventral motor neurons (Fig. III-5D). Strong expression was also found in the DRG. but no signal couid be detected in sympathetic ganglia or rnyotome as observed for Dncil (compare Fig. 111-5C and D). In El 1.5 cranial sections, expression was found throughout the entire brain as well as the fifth trigeminal sanglia. facio-acoustic ganglia and Rathke's pouch (Fig. III- 5F). Analysis at E13.5 revealed extensive expression of DnciZ throughout the CNS and peri pheral nervous system ( PNS ). Expression was also observed in rhe toothbuds. olfactory epithelium. Jacobsen's organ. tongue. palate and whiskrr folliclcs (Fig. III-6B.

E-G). Weaker expression was found in cartilage primordiurn. lung and thymus (Fip. III- 6B).

The expression found in E 10.5 limb buds for Dnci2 was funher analyzed during the early stages of limb rnorphogenesis. Closer analysis at E lO.5 revealcd strong expression in the distai limb bud (Fig. III-7A) and tissue sections revealed that the expression was found in

both mesenchymal and epitheiial cells (Fig. 111-78). At E 12.5 the expression was confined

to the hand/foot plate in the interdigital mesenchyme and surrounding the condensing

mesenchyme of the digit blastemas (Fig. III-7C). At E 13.5 Duci2 expression was found in

the mesenchymal tissue outlining the metacarpal primordium (Fip. III-7D). Expression

patterns in forelimbs and hindlimbs were found to be identicai (data not show). I -

. . . ma.

Figure 9IbE -Expression et DncCl and mouse development. Whole mount in situ hybridization shows tha t at E10.5 Dncil is expressed in the dorsal root ganglia (drg) but is excluded from the limb buds (A). Analysis in tissue sections at El 1.5 show that Dncil is expressed in the drg, Motor neurons (mn), myotome (my). sympathetic ganglia (sg) and mesenchyme in the vicinity of the notochord (me) (C). At E10.5 whole mount in situ hybridization shows that Dnci2 is expressed in the drg, limb buds (Ib), first and second branchial arches, (ba) and (ha), respectively (B). Section in situ at E10.5 shows expression in the branchial arches in both the mesenchyme (me) and epithelium (ep) (E). At El 1.5 DnciZ is expressed in the drg, the motor neuron. and throughout the remaining neural tube (nt) (D). Cranial sections reveai expression of Dnci2 throughout the forebrain (fb) and hindbrain (hb), as well as in the trigeminal ganglia V (tgV), the fascio-acoustic ganglia (fa) and Rathkes pouch (rp) (F). Figure 111-6. Analysis of expression of murine Dncil and Dnci2 at E13.5. Section in situ shows that Dncil is expressed specifically in the neural cortex of the forebrain (fib. nc) (A and C). Expression is also found in the dorsal root ganglia (drg), skeletal muscle (mu), and cartilage primordium (cp) (A and D). Signal is detected in Meckle's cartilage (mc), but may-bekapped probe: Section and whde mount analysis show that Dnci2 is expressed throughout the entire brain (B and F). Expression is also detected in the drg, cartilage primordium, thymus (th), lung (lu), neural tube (nt) (B). Closer analysis finds expression in the olfactory epithelium (oe) surrounding the nasal cavity (nc), as weil as in Jacobsen's organ (jo) and incisor primordium (ip) (E). Whole mount analysis at E13.5 further shows expression in the palate (pa), tongue (to), and whisker foiiicle (wf) (F and G). Figure 111-7. Expression of Dnci2 in the developing mouse limb. Whole mount analysis shows that Dnci2 is restricted to the distal limb bud at E10.5 (A). Analysis in tissue sections show that the expression is found in both the distal limb bud mesenchyme (me) and the epithelium (ep) (B). At E 12.5 expression is found in the interdigital mesenchyme (im) and surrounding the digit blasternas (C). At E 13.5 expression is confined to the mesenchyme surrounding the developing metacarpals (me) (D). Only forelimb bud expression is shown here. Hindlimb bud expression was found to be similar (not shown). I1I.D Discussion

In our effort to identify the gene(s) responsible for split-hand split-foot malformation

(SHFM 1) at chromosome 7qZ 1.3-qlI. 1 we have identified the cytoplasmic dynein intermediate chah 1 sene (DNCI1) using posiiional cloning experiments.

In ût least 2 SHFM 1 patients carryin: balanced chromosome 7q2 1-qX translocations. situs

inversus has been observed (Genuardi et al.. 1993). Since the involvement of dyneins in

abnormal organ migration has previously been demonstrated (Supp et al.. 1997). we hypothesizrd the DNCII might be involved in the heterotûxia observed in ihrse 2 patients.

and that it could also have a broadrr involvement in lirnb development. To test this hypothesis. the mouse Dncil was cloned and RNA in situ hybridization analysis was

performed. Durinp embryogcnesis the gene was not found to be expressed in the

developin~limbs suggestin_oit probably does not contribute to the abnormalities of the

distal extremities in SHFM 1 patients.

Despite the lack of deteciable Dizcil gene expression in the developing lirnb. it remains an

interesting possibility thet this gene is responsible for the observed phenomenon of segre_ration distonion obsened in SHFM 1. where male to male transmission rates are

above the expected Mendelian ratios (Stevenson and Jennings. 1960: Jarvik et al., 1994).

The rnechanism of segregation distortion is poorly understood but the phenomenon has

best been characterized in the t-haplotypes (reviewed in Silver. 1993). T-haplotypes result from the rearrangernent of the T-locus in mice. Males who cany such a haplotype transmit

this allele to 99% or more of their offspring at the expense of the normal allele. Recently.

genes from this locus have bren identified and two of thern. 7cre.r- 2 and Tcte-r-2, were

determined to encode light chain dynrin senes (King et al.. 1996: Patel-Kinz et al.. 1997).

Based on these observations one may speculate that haploinsufficiency of this dynein gene in the SHFM 1 critical region is responsible for the segregation distortion observed in some families,

Alternative splicing has ken reported for both Dncil and Dnci2 in nt (Paschal et al.. 1992; Vaughan and Vallee. 1995). In this study we have identified sevenl alternatively spliced isoforrns for both human and mouse DNCIl. The regions of alternative splicing found in both mouse and human were highly similar to that reported for rat. On the other hand. novel isoforms not reported in rat were found in human. mouse. or both. The differences in identified isoforms likely reflects the methodology employed or the tissue source analysed. Nevertheless, rat isoform Drrcib was not found in human or mouse even following RT-PCR analysis. In this case RT-PCR was performed on mouse brain RNA. the sarne tissue used for nt gene identification (Paschal et ai.. 1992). Thus. it would seem that Dncilh rnay be specific to the rat brain. The alternative spiicing in human DiVCIl is probably the result of a cryptic splice site wirhin exon 4. The same splice isoform was identified in mouse. and is likely a result of a similar cryptic splice site. In addition. based on the human genomic structure of DNCIl. it is likely that mouse Dncild is the result of the utilization of a second cryptic splice-acceptor site in mouse exon 4. It would be nccessary. however. to detrrminr the genomic structure of mouse Dnci 1 to confirm this.

Alternative splicing has also been reported for DnciZ in rat. but only a single isoform was identified in mouse. It is likrly that the latter was also a result of the nature of the tissue analysed since an ernbryonic library was usrd in this study to isolate the rnouse cDNAs. Northern blot analysis of Dnci2 in mouse. however. has provided evidence for an additionai isoform expressed exclusively in testes. This isofom was not observed in nt tissues (Vaughan and Valler. 1995). It has been frequently observed that alternative isoforrns for many genes are expressed post-meiotically during spematogenesis (reviewed in Erickson. 1990). One may speculate that this alternative isoform in mouse may be a result of post-meiotic gene expression. It is reasonable to suggest that there is a requirement for dynein at these stages of spermatogenesis since cytoskeletal reorganization is necessary for the dramatic structural changes that occur dunng these processes (Willison and Ashworth, 1987). Other cytoskeletal components, such as a-tubulin (Distel et al.. 1984) and y-actin (Slaughter et al.. 1989) are also post-meiotically expressed. supponing

this hypothesis. Cloning of this testes-specitïc isofonn may facilitate the analysis of post-

meiotic gene expression of DnciZ and elucidate the importance of dyneins in the process of

spermatogenesis.

Northern blot analysis indicates that DnciZ is expressed in rnost. if not all. adult tissues

exarnined. This is consistent with a rolr for D~iîas the predominant intermediate chain in

mamrnals. In contrast. Dricil has a highly restricted expression domain found only in brain

and testes. Consistent with the Northern bloc data it was found that Dncil has a highly

restrictrd expression pattern. predominantly found in specific regions of the PNS and the

brain. On the other hand. DnciZ was found to express throughout the CNS and PNS

except for the sympathetic ganglia.

The dynamic expression of Dm52 in the developing limb bud suggests that this gene plays

a role in discal limb development and digit patteming. It is known that E10.5 is an active

stage of limb development when the anterior-posterior and dorsal-ventral axes are patterned

(reviewed in Tickle and Eichcle. 1994). It is possible that DnciZ is required to establish or

maintain proper ceIl polarity in the limb bud at thrse early stages. It has been postulated

that cell polarity is likely dictated by an intracellular asymmetrical molecule (Brown et al..

1991). While dynein itself may not be ihis molecule. it could rnediate the role of this

molecule by orienting intncellular componenrs accordingly (Levin and Nascone. 1997). The identification of a mutation in an axonemal dynein gene in the iv mouse (Supp et al.. 1997) supports the hypothesis that these genes are important for establishing proper ce11 polarity. It is apparent chat the products encoded by many dynein genes may function as both axonema1 and cytoplasmic dyneins. Thus, it has been postulated that the situs inversus phenotype in iv mice may be a result of the failure to establish proper ce11 polarity during early embryogenesis essential for proprr left-right asymrnetry during development (Supp et al., 1997).

Previously. a backcross mapping study indicated a putative map location for murine Dncil on chromosome 6 (Vaughan et al., 1996). This assignment is supported by the syntenic conservation of genes mapprd to the SHFM 1 region (Tsui et al.. 1995; Crackower et al..

1997). It has been proposed that Dncil may be a candidate gene for the mouse mutant hop- srerile (Iiop).which also maps to mouse chromosome 6 (Johnson and Hunt. 197 1). Hop niice display male infertility. preaxial-polydactyly. hydrocephaly and a hopping gate

(Johnson and Hunt, 197 1 ). Spermatogenrsis is abnormal in thcse mice due to improper

migration of the centrioles. As a result normal sperm tails rarely develop. Since related abnormalitirs have been obsrrved in cultured cells when dynein function has been disturbed (Echeverri et al.. 1996). a defect in dynein function in these mice has been

proposed (Vaushan et al.. 1996). Nevenheless. no sequence aitcraiion of Dncil could be

identificd in hop mice compared to wildtype littermates (our unpublished observations). In

addition. no expression of Dncil was found in the limb bud. thus excluding this gene from

a possible role in the genention of the preaxial polydactyly phenotype.

The data presented hrrc provide a framework for funher annlysis of the functional role of

Dncil and Dnci2 in mouse and human. I< appears chat these genes should play overlapping and distinct roles during embryogenesis. The generation of "knockout" mice should

provide vduable insight into the precise rolrs that these genes play. In addition. it will further add to the growing understanding of the role of dyneins in the multicellular organrsm. Chapter IV

Phenotypic Characte~izationof the Dactylaplasia Mouse, a Mode1 for Human SHFM

Attribution of work:

The work presented in this chapter was pnmdy carrird out by the candidate. The analysis of ce11 proliferation by BrdU was perfoned in collaboration with Dr. Jun Motoyarna, a postdoctoral fellow at the Hospital for Sick Children.

The data presented in this chapter has been published in pan in : Crackower, b1.A.. Motoyama. J.. and Tsui. L.-C. (1998). Defect in the maintenance of the apical rctodemal ridge in the Dactylnplosin mouse. Dev. Biol. 201: 78-89. 1V.A Introduction

When the vertebrate limb bud fint appears (at E9.5 in mouse). it consists of mesenchyme sunounded by an ectoderm-derived epithelial jacket. Shonly thrreafter. signals from the mescnchyme induce the overlying ectodermal cells at the apex of the limb bud to change morphology and form a thickening called the apical ectodemal ridge (AER) (Saunders.

1948: Kieny. 1960: Saunders and Reuss. 1974). One role of the AER. as determined throush experimental studies carried out primarily in chick rmbryos. is to signal the underlying mesenchymal cells. known as the progress zone (PZ). to remain in an undifferentiated state and promote limb bud outgrowth (Saunders. 1948: Summerbell et al..

1973; Summerbell, 1974).

A second rolr of the AER is to maintain the zone of polarizing activity (ZPA). a region of cells in the posterior mesenchyme of the early limb bud that conirols patteming of the anterior-posterior t AP) axis of the developins appendage (Saundcrs and Gasseling. 1968:

Vosel and Tickle. 1 993).

In the past there have been a number of classical ernbryolo_oicalstudies thût have addressed the maintenance of the AER. These have led to the hypothesis that secreted signals rmanating from the distal limb mesenchyme are required to maintain the AER (Zwilling and

Hansborough. 1956: Saunders and Gasseling, 1963). There has been an absence of direct genetic or molecular data to support this hypothesis. however. and the rnolecular nature of C this factor and the biochemical pathway in which it functions is unknown.

The phçnotype of SHFM represents a unique fom of ectrodactyly (missinp digits) in which central digits are preferentially affected. It has been suggested that a defect in the central .\ER could give rise io this abnorrnaiity (Watson et al.. 1986: Roberts and Tabin. 1994). Furthemore, Zwilling (1961) suggested deficiencies of the AER maintenance factor could result in ectromelia, or loss of some distal skeletal elements. Taken together, it can be suggested that the abnomality in SHFM and Dac rnice may be a result of a defect in AER maintenance.

Presented in this chapter is the data from a sysrematic characterizarion of the semi-dominant mouse mutant Dncplaplasia (Dac).Heterozygous animals display an isoiatrd limb drfect chnracterized by missing phalanges in the central digits and homozygous animais are characicrized by monodactyly (Chai, 198 1: Johnson et al.. 1995). For the Dac phenotype to express. however. heterozygous animals must be homozygous for a recessive allele at an unlinked locus, called rrtDac (Chai, 1981). The molecular nature of neither Dac nor

~tiDacis known but thrir respective genes have been rnapped to chromosomc 19 and 13

(Johnson et al., 1995).

One human locus. SHFM3. has been mapped to human chromosome 10 in a region syntenic to the Duc locus supponing the hypothesis that the Dac rnouse represents an animal mode1 for the human condition (Johnson et al. 1996).

Through cellular and molecular analyses we sought to analyze the Dac phenotype and dctermine the developmental defect giving rise to the abnormality. The data presented in this chapter show that Dac mice have a defect in the maintenance of the AER. Evidence is provided suggesting that the failure to maintain the AER is due to reduction in the amount of ce11 proliferation in the AER. A mode1 for AER maintenance is proposed where additional secondary sources of AER maintenance activity are capable of partially comprnsating for the Duc activity, and that signals are localized to the anterior and posterior mesenchyme of the developing lirnb bud. 1V.B Materials and Methods

IV.B.l Mice and Genotyping Mice heterozygous for the Dac mutation on a SM/Ckc background were obtained from the Jackson Laboratory (Bar Harbor. ME). To obtain FI offspnng of the proper genetic background. the Dac mice were crossed with wild-type BALB1cBy.J mice (Jackson

Laboratory); both SM/Ckc and BALBIcByl are homozygous for the recessive mDac allele that is required to express the Dac phenotype (Johnson et al.. 1995). Affected Fi mice were rhen intercrossed to generaie Fz mice used in the study. Mating was allowed overnight and noon of the day of evidence for a vaginal plug was considered embryonic day (E) 0.5. Since the Dac gene has yet to be cloned. genotypes of Fî animais were inferred by a rnicrosaicllite marker DI9iMirlO which has been shown to be tightly linked to the Dac mutation and informative for the two background strriins used (Johnson et al..

1995). Grnotypins was performed by PCR analysis of tail or yolk sac DNA (Hopet al..

1994). Mice with two BALBIcByJ alleles. one BALBfcByJ allele and one SM/Ckc allele. and two SMICkc allelcs were considered wildtype. Dac/+. and DadDac. respectively.

IV.B.2 Skeletal analysis

The canilaginous skelrton of El 1.5- 13.5 embryos was analyzed by Alcian blue staining (Jegalian and De Robenis. 1992). Skeletal analysis of adult and E18.5 specimens were performed by using AIizarin red and combined Alcian bluelAlizarin red. respectively (Lufiin et al.. 1992.)

IV.B.3 RNA In situ hybridization anaIysis

Whole mount in situ hybndization analysis was perfomed as described (Ang and Rossant.

1993). The DM probe was a full length cDNA obtained by probing a mouse ernbryonic cDNA libnry (Novagen) with a full length human cDNA (our unpublished data). A detailed analysis of the mouse DMexpression pattern will be published at a later time. The deiails of the remaining probes have been reported elsewhere: Fgf8 (Crossley and Martin. 1995). Fgl;l (Hébert et al.. 1990). FdIO (Bellusci. et al.. 1997), BmpZ (Lyons et al., 1989). Bnzp4 (Jones et al.. 1991). Hoxdl.3 (Dollé et al., 1991). HoxdlZ (Izpisua-

Belmonte et al., 199 1). Shh (Echelard et al.. 1993). Wnt7a (Parr et al., L993). A minimum of five srnbryos of each genotype were analyzed in more than two separare experiments. for each probe analyzed.

IV.B.4 Scanning electron microscopy

Embryos were dissected in PBS and fixed at 4OC in universai fix (1% glutanldehyde). After post-fixation in 1% osmium tetroxide in 0.1 M phosphate buffer for 1 hour, specimens were deh ydrated in gradrd alcohols. Specimens were then subjected to criticai point drying. rnountcd on metal studs. sputter coated with gold particies. and observed using a JEOL JSM 820 scanning electron microscope.

IV.B.5 BrdU labeling of embryos

In vivo labeling of newiy synthesized DNA and scoring of replicating celis were prrformed as described by Gratzner et ai. (1982). BrdU (20 mgkg of body weight) was injected intraprritonrally into pregnant females 3 hours prior to sacrifice. Embryos were fixed in ethanol. dehydrated and paraffin embedded. Transverse sections (7pm) perpendicular to the body axis were made. From a E10.5 lirnb bud. 60-70 sections were collected and about 100 sections were collected from each El 1.5 limb. The position of anterior. middle and posterior sections was determined depending on the sequence of sections and on the relation to other organs like heart and lung buds. Extreme anterior and posterior sections and sections from the rniddls 20% of sections were used for counting. The AER region was defined by ceIl and tissue morphology. Total cells and BrdU-positive cells were counted in at leasr 1 sections from 2 embryos for each genotype and position. For counting total and BrdU-positive mesenchymal cells, we defined an area of distal mesenchyme which is within 200 pm from the apex of the limb bud. Following detection. tissue sections were counter-stained with Hematoxylin (Sigma).

IV.B.6 TUNEL Staining of Apoptotic Cells

The original TUNEL staining method of Gavrieli et al. ( 1992) was used to detect apoptotic cells on tissue sections of E 10.5 and E 1 1.5 embryos. except that 1nM of biotin- 1 1-dUTP (Boehringer Mannheim) and 0.3 UIp1 of terminal deoxynucleotidyl transferase (Gibco BRL) were used in the nick-translation reaction. Incorporated Biotinylated dUTP was detected by Streptavidin-horss radish peroxidase (Gibco BRL). Following detection. tissue sections were counter-stained with msthy l green (Sigma).

1V.C Results

IV.C.1 Loss of anterior and central bones of the autopod in Dac mice Dac has bern considered a genetic and phenotypic mousc mode1 for split hanci/ split foot malfornation in humans because of rheir resemblance in appearance (Chai. 198 1; Johnson et al.. 1995) and their syntenic map relationships (Johnson et al.. 1995). Prior to this study. however. the morphology of the limbs of adult Dac mice had only been briefly described (Chai. 1981; Johnson et al., 1995). Therefore. to obtain a detailed understanding of the limb phenotype. ihe morphology and skeietal structure of mutant limbs were analyzed in adult and E 18.5 mice.

The results showed chat Dac/+ mice typically exhibited loss of phalanges in the central digits in 1-1 limbs, but digit 1.1 and 5 were prescnt (Fig. IV-1 B). Bone fusion between digits 4 and 5 at the level of the first phalange was genenlly observed (Fig. IV-1 B'). The heterozygous phenotype was highly variable. however. and more severe (Fig. IV- I D.D1: central metatarsa1 absent) and less severe (Fig. IV- 1C.C': soft tissue fusion only) examples were found. This extensive variability has also been seen in SHFM (McKusick. 1992). Other bones of the autopod and the long bones appeared to be unaffected in heterozygous

Dac mice. The skeletal structure and variation of seventy in affected forelirnbs was similar to that seen in hindlimbs (data not shown). Of the limb structures still present. there was no obvious defect in dorsoventnl (DV) patterning as determined by the presence of skin striations and dermal foot pads on the ventral surface and fur on the dorsal surface of the limbs (Fig. IV- I A-D and data not shown). As revealed by combined alcian bluel ûlizann red staining. the skeletal structure of Dac/+ mice at E18.5 was consistent with that observed in the adult (Fig. IV- 1F.1).

The rnajority of Dac/Dcic mice died at binh in the genrtic background used and thus no adults werr availablr for analysis. The reason for this early rnonality was not determined,

but no other skeletal abnormdity could be srcn in thrse mice other thon thosr of the limbs

(data not shown). On the basis of combined alcian blue/alizain red staining, DadDac mice examined at E18.5 consistently had only a single posterior digit similar in size and

morphology to digit 5 (Fig. IV- 1G.J). Only posterior bones of the autopod remained and

rhese were sometimes dysmorphic (Fis IV-1G.J). In addition. it was often found that the

distal radius wns mildly affected (Fig. IV-1J). While premature ossification of the talus

was observed in some cases (Fig. IV-lG). ossification appeared to proceed at a normal rate in general. The structure of mutant limbs was further analyzed by scanning electron

microscopy (SEM). It was apparent that at El45 only digit 5 was present in homozygous mice (Fig. IV-3E.F). These results thrrefore suggested that the defect in these mice was due to the loss of normal structures and not malformation of existing elements. Figure IV-1. Skeletal analysis of Dac limb phenotype. Palmer view (A-D) and skeletal alizarin red stain (AD-D')of WT (A. A') and Dac/+ (B-D, Be-Di)adult hind limbs. Cornbined alizarin redl alcian blue staining of E18.5 WT (E.H) Daclt (F.1) and DadDoc (GJ)limbs. (B.B')Dac/+ adult phenotype commonly missing digits 2 and 3 and digits 1 and 5 fused at the lcvel of the first phalange (small arrow in BI). (C.Ct) Less severe Dac/+ phenotype with only soft tissue fusion between digits 1 & 2 and 4 & 5. (D.D1) More severe Dac/+ phenotype displaying absence of digits and metatarsals 2.3 and 4 (arrow in D'). DuclDac frequently die at birth and are not available for adult analysis. Similar phenotype and variation is found in forelimbs (data not shown). (E. H) WT E18.5 hindlimb (E) and forelimb (H). Digits labeled anterior to posterior, 1-V in H. (F.1) El 8.5 Dac/+ hindlimb (F) and forelimb (1). a similar phenotype is observed to that found in adult limbs. Note the more severe phenotype observed in the E 18.5 Dacl+ forelimb showing absent metacarpals (arrow in 1). DaclDac E18.5 hindlimbs (G)and forelimbs (J) display absence of digits 1-4. digit 5 (labeled V). with only posterior bones of the autopod remaining some of which are malformed. Reduction of the distal radius is also found (arrow in J) and premature ossification of rhr talus has also been observed (mow in G).

IV.C.2 Onset of the Dac phenotype prior to chondrogenesis in the autopod To determine the pattem of chondrogenesis of the preskeletal elernents of the autopod in mutant Iimbs, embryos from El 1.5-E 13.5 were analyzed by whole-mount alcian blue staining. At E13.5 Dad+ mice showed a reduction in the formation of the central digit blastemas. and posterior digit 4 and 5 were frequently hsed at thsir distal ends (Fig. IV- 2H.K). These mice also displayed an altered pattem of chondrogenesis of carpal and tard elements. In E13.5 DaclDuc mice only a single digit and a smail nurnber of carpal and tard elemrnts could be obsenvd and the distal radius and tibia blastemas appeared to be shonened (Fig. IV-2I.L). In addition there appeared to bc a slight drlay in chondrogenesis in the digits. Staining of El23 rmbryos revealed an aberrant pattern and reduced amount of cartilage formation in hetrrozygous mutants (Fig. IV-2B.E). It was apparent that the overail size of thc hand plate was also reduced (Fig. IV-2B.E). This observation was also true for hornozygous mutant rmbryos but the phrnotyps was more scvere (Fig. IV-2C.F).

At this stage distal autopodial condensations were observed but they were abnormül in shape and six (Fig. IV-2B.C). Thsrefore. it appeared that skeletal clements that were absent in the adult lirnbs never formed in the embryo.

At El 1.5 a disruption of the formation of the hand plate was observed. but at this stage cartilage staining was not detectable in the autopod. suggesting that the defect arises prior to chondrogenesis (data not shown). The reduction in the size of the handlfoot plaie suggrstrd that the loss of skeletal elements could bc dur CO the reduction in the number of cells avaiiable for chondrogenesis in the distal autopod. Figure IV-2. Whole mount dcian blue staining of EIZ.5 and E13.5 Dac limbs. WT (.A.D,G.J) Dac/+ (I3,E.H.K) and DaclDac (C,F.I.L) E12.5 (A-F) and Ei3.5 (G-L) forelimbs (A-C. G-1) and hindIimbs (D-F,J-L) are shown. (B.E) Dacl+ El23 limbs display a reduction in the size of the hand plate and abnormal pattern and amount of chondrogenic tissue. (Cf) DuclDac E12.5 limbs display an exacerbation of the Dac/+ phenorype. Asrerisks in B and C pinpoint abnormal digital condensations. (H.K)E13.5 Dac limbs frequently display absence of central digital condensations and show fusion of the posterior disit blastemas. (1.L)DadDac E13.5 limbs display a single digit blastema and condcnsations for a small number of autopodial elements. Chondrogenesis of the distal long bones appear reduced as well ( arrow in 1 and L).

IV.C.3 Degeneration of the AER in Dac lirnb buds The apical ectodermal ridge (AER) is required to maintain the underlying mesenchyme (the progress zone) in a proliferating state. thus promoting limb outgrowth (Sumrnerbell et al..

1973). Moreover, in chick, removal of the AER leads to tmncation of the distal skeletal elements (Saunders. 1948: Summerbell, (974). Since the first observable morphological defect in Dac rnicc was a reduction of the size of the handfooi plate, prior to riny cartilaginous condensation of the distal autopod. we reasoned that Dnc mice müy have a dekct in the AER.

Scanning electron microscopy (SEM) was employed to determine if there was a morphological alteration in the AER. At E10.5 the Dac/+ and the DaclDuc AER appeared morphologically unaltered compared to wild type litter mates (Fig. IV-3A.B and data not shown). At El 1.5. however. there was a sporadic degeneration of the AER in Dac/+ limbs (data not shown). In DaclDac mice most of the .4ER had degenented by El 1.5 leaving only the posterior region intact (Fig. IV-3D). Therefore. prior to E10.5 the AER in both heterozygous and homozygous limbs appeared morphologically normal but by El 1.5 there was degeneration of this structure in animals of both genotyprs. These data suggest chat the AER was properly induced but failed to be rnaintained in Dac mutant limb buds.

IV.C.4 Loss of expression OF AER markers in mutant limb buds at ElO.5 Marker gene analysis was smploycd to characterize funher the defect in the AER. The expression patterns of AER markers were analyzed in limb buds from El0 to El 1.5 of development. At EIO. FQfS expression was found to be normal in the forelimbs and hindlimbs of mutant embryos (data not shown). Analysis of the expression of FgfS at

ElO.5, however. revealed a lower level of expression in the central AER of affected forelimb buds (Fig. IV-4B.C). The expression of Fd3 (Fig. IV-JB) in Dac/+ mice was highly variable nnting from complete absence of expression in the cenual third of the AER (Fig. IV4B) to patchy expression throughout the AER (Fig. IV-4F). This variation in gene expression correlated well with the variability of the heterozygous skeletal phenotype. In DaclDac EL0.5 forelimbs the expression of Fgfs (Fig. IV-4C)was absent from the majonty of the AER excrpt for the most posterior and anterior regions. It was found that the expression of F@ is unaltered in hindlimbs of both heterozygous and hornozygous animals at E 10.5 (Fig. IV-4A-C). There is a 0.5 day developmental delay of the hindlimbs with respect to the forelimbs during mouse embryogenesis and accordingly, an alteration of Fm expression was observed in E 1 1.0 hindlimbs (Fig. IV-4D) similar to that observed in

E 10.5 forelimbs. Furthermore. in E 1 I .O forelimb buds of homozygous mice it was found that only the posterior expression domain of Fgf8 remained (Fig. IV-4E). The expression of Fd4. Bmp2 (Francis et al.. 1991). BmpJ (Jones. et al.. 1991) and Dl.r5 (this thesis) were also examined during these developmental stages and their expression profiles in the AER appeared to be similarly affectcd to F,@ (Fig. IV-lH.I and data not shown). From this it can be concluded that grne expression in the central and anterior AER is affected in Dac mice and that the stage of onset is approximately E 10.5 in forelimbs. This data would further argue thai prior to E 10.5 the function and morphology of the AER was unaffected supportin2 the hypothesis that the AER was induced and differentiated normally in the mutant limb buds. Togcther with the results of the SEM analysis drscribed previously. ii would suggest that Dcic mice have an isolated defect in the maintenance of the AER. Figure IV-3. Scanning electron microscopy of Dac limbs. WT (A,C.E) and DaclDac (B,D,F) limbs were analyzed for morphology. (A,B) At E10.5 the morphology of the AER is the same in WT and mutant limb buds (arrowheads show AER). (C,D)By El 1.5 most of the AER has begun to degenerate in DaclDac limbs (hindlimb shown here) except for the most posterior portion (indicated by arrow in D) , while AER rernains intact in WT (arrowheads in C). (E,F) At E 14.5 it is clear that only digit V remains in DaclDac Limbs. Figure IV4 Analysis of expression of AER markers using whole mount in situ hy b ridization. Expression of FgfB (A-F) and Fgf4 (G-0are shown. (A-C) Lateral view of WT (A) Dac/+ (B) and DaclDac (C) E10.5 showing Fgf8 expression in the AER. At E10.5 strong staining is observed through the AER in forelimb and hindlirnb in WT embryos (mow and arrowhead in A). In Dad+ embryos at this stage, absence of staining in the central AER is frequently observed in forelimbs (arrowhead in B). However, the Dac/+ expression pattern is highly variable. An example is shown in F where an non- homogeneous expression pattern is found where small arrowheads show regions of no or weak expression. (C) In DaclDac ElO.5 embryos Fgf8 expression is absent from most of the AER (arrowheads) except for the most anterior and posterior regions. At E10.5 hindlirnb expression is unaffected (arrow in A-C). (D) At E 11 .O DaclDac hindlimbs display similar expression patterns to that of E10.5 forelimbs with only anterior and posterior AER expressing FM (arrowheads). (E) By El 1.0 Dac/+ forelimbs only retain the postenor domain of Fgf8 expression (arrowhead). (G-I) similar aberrant expression patterns are observed for Fgf4 (arrowheads show regions of absent expression) as well as other AER markers tested (data not shown). IV.C.5 Normal mesenchymal gene expression in Duc limb buds The zone of polarizing activity (ZPA) is known to interact with AER to coordinate PD and AP patteming dunng limb development (Laufer et al.. 1994, Niswander et al.. 1994). To determine if the AER defect in Dac might be due to an alteration in the ZPA. the expression of mesenchymal marker Slih was examinrd in the mutant mice. The result showed that limb bud expression of Shh was unalterrd at EiO.5 (Fig. IV-5A.B) and up to E11.5 (data not shown). Consistent with this observation. the expression of Fsf3 is properly maintainrd towards the posterior extremity of the AER (Fig. IV-4G-1). In addition, wr studied AP pattrming of the mutant limb through examination of the expression profile of the most 5' Hard genes and the results indicatrd that both Hoxdl.3 (Fig. IV-5C.D) and

Ho.rd22 (data not shown) were normally expressed at E 1 OS. Thesr data are therefore consistent with the Duc skeletal phenotype which shows a distinct AP axis.

Also examined was the expression patterns of additional mesenchymal markers. including

Bnip?. a biochemical marker downstrearn of SIiIi and strongly expressed in the posterior mesenchyme (Francis et al.. 1994). Bnip-l. also rxpressed in the posterior mesenchyme but in addition expressed in the anterior mesenchyme (Jones et al.. 199 1). and FgflO, a recently identified member of the Fgf family. rxpressed in the prospective limb bud mesoderm and distal mesenchyme of the rarly limb bud (Ohuchi et al.. 1997). The data showed that the mesenchymal expression of nonr of these rnarkers were affected in Dac limb buds at ElO.5 (Fig. IV-SE-H and data not shown). The result with FgflO is particularly interesting because this distal limb mesenchyme marker has been shown to induce and maintain F@ expression and a rudimentary AER upon ectopic expression in chick limb buds (Ohuchi et al., 1997). Although FdZO rnight be a strong candidate for the endogenous ridge inducing signal and could play a role in AER maintenance. its expression was not affected by the absence of Dac. On the basis of the morphology of Dac limbs (Fi_o. IV-i), we reasoned that genes regulating DV patterning would have a normal expression profile. To confirm this assumption, the expression pattern of Wnt7a. a key regulator of DV polarity dunng limb development was examined (Dealy et al.. 1993: Riddle et al., 1995; Parr and McMahon.

1995; reviewed in Zeller and Duboule, 1997). The result indicated that the dorsal limb cctoderm expression of this marker was indeed unaffected in mutant limb buds at E10.5

(data not shown).

IV.C.6 Normal programmed ce11 death in Dac limb buds

As previously describcd (Milaire and Rooze, 1983; Vaahtokari et ai., 1996). a high level of programmed ce11 death (PCD)could be detected in the wildtype AER at E10.5. This suggested that a significanr amount of ce11 death in the AER at E10.5 was required for its

proper function. Since the degenention of the AER in mutant limb buds by E 1 1.5 could be due to an increase in PCD,TUNEL assay was performed to assess the degree of PCD

in mutant limb buds. The results showed that there was little or no change in the amount

of death in the AER of affected limb buds as compared to wild type littermates (n=4 for

each grnotype) (Fig. IV-6A. B). In addition. very few apoptotic cells could be found in

the distal mesenchyme of wild type and mutant limbs at E10.5 and El 1.5 (n=4 for each

genotype) (Fig. IV-6A-D). Therefore. these results indicated thai the small size of the

mutant limb bud was not due to excessive ce11 dsath in the dista1 limb bud. Figure IV-5. Analysis of AP patterning and FgflO expression in Dac limb buds. The expression of Shh (A,B), Hoxdl3 (C,D),Bmp4 (E,F). and FgflO (G,H)was analyzed in E 10.5 WT (A,C,E,G). Dacl+ (not shown) and DadDac (B,D,F,H) limb buds. No difference in pattern or amount of expression was observed between any of thtse markers. Arrow in A and B point to region of Shh expression in posterior limb buds. Arrow and arrowhead in E and F point to expression of Bmpl in posterior and anterior mesenchyme, respectively . Arrowheads in G and H point to positive deep purple staining for FgflO in the distal limb bu&. IV.C.7 Lack of ce11 proliferation in Duc limb buds By El 1.5, limbs of affected animals were clearly smaller in size as compared with their nomal littemates (see Fig. IV-2). Since the absence of the AER should lead to a reduction in the rate of ce11 proliferation in the underlying mesenchymal cells (Summerbell et al..

1973). analysis of Bromo-deoxyuridine (BrdU) incorporation was performed to confirm this causal relationship. The results showed that there was a high percentage of BrdU- incorporatin_ecells in the distal mesenchyme of WT limb buds at E 10.5, and there was no obsemed difference in Dad+ or DadDac limbs (ndfor each genotype) (Fig. IV-6E,F.K).

At this same stage. however, it was observed that 42% of cells in the central AER of the

wildtype buds were BrdU-positive (n=4). while only 5% were positive in the mutant limb

buds (n=6) (Fig. IV-6G.H.L). Thus. these results suggested that the degeneration of the

AER in mutant limb buds is likrly due to a lack of ce11 proliferation. Furthemore. it was

observed that the mutant AER. while not proliferating. was morphologically normal at this

stage (compare Fie. IV-6G and H) strengthening the argument that the AER developed

norrnally pior to this stage.

Dcperation of AER was clearly evident in DaclDac mice at El 1.5 (Fig. IV-6J) and there

was a sipnificant decrease in the proliferation of underlying mesenchyme in the central limb

bud as compared to that of wildtype littermates (Fig. IV-61.J.K). Taken together. these

results indicate that the Dac mutation results in reduced ce11 proliferation in the AER and

suggcst that the abnormality observed in the mesenchyme of the mutant mice is due to the

lack of a funcrional AER. Figure IV-6. Analysis of BrdU incorporation and programmed ce11 death in Dac limb buds. Analysis of programrned ce11 death by the TUNEL detection method shows that at E10.5 PCD is high in WT and mutant AER (A.B) but no difference is observed. Very low number of ce11 positively label as apoptotic in E 1 1.5 mesenchyme of WT and mutant limb buds (C,D).Black nuclei are positive for apoptosis. AER is highlighted with a bracket. At E10.5 ce11 proliferation as measured by BrdU incorporation in the distal mesenchyme is the sarne in WT (E,K), Dac/+ (K) and DaclDac (F,K). Close observation of E10.5 AER shows a substantial arnount of incorporation in the WT AER (G,L) and vinually a complete absence of incorporation in central AER of mutant limb buds (H,L). Blow up of box in E and F are shown in G and H. respectively. At El 1.5. a flattened AER is found in WT limb buds (bracket in 1) but no AER is found in the ccnter of mutant limb buds (J and data not shown). At this same stage a major reduction in ce11 proliferation of the mesenchyme is found in DaclDac limb buds (J,K) and an intermediate reduction is found in Dacl+ limb buds (K, and data not shown). A cornparison of anterior. middle and posterior sections wrre analyzed and show thrit cell proliferation in E10.5 AER is unaffected in the anterior and posierior lirnits of the mutant limb (L). Brown nuclei are positive for BrdU incorporation. The data shown in K are the percentage of incorporating cells +/- Standard Error Mrasurement. For each data point 4-5 sections were counied from at least two independent rrnbryos. Al1 sections shown are taken from the center one third of the AER with respect to the AP axis.

1V.D Discussion

Classical embryological studies in chick Ied to the hypothesis that there was a secreted factor present in the limb bud mesenchyme that maintains the AER (Saunders, 1949:

Zwilling and Hansborough. 1956; Saunders and Gasseling, 1963; Zwilling, 1964). In most instances these studies did not specifically address the AER maintenance activity; the idea of such an activity was formulated to explain the observed effect that limb and non- limb mesenchyme had on AER rnorphoIo_oyand maintenance. While the data from these studies strongly argued for the existence of an AER maintenance activity it was not possible, based on these indirect observations. to determine if this activity was distinct from the activities of AER induction or differentiation. Therefore. until now, the idea of an isolated AER maintenance activity had remained hypothetical. The data presented here show that DLK mice have ri defect in the maintenance of the AER providing direct genetic

çvidence for the existence of an AER maintenance activity. The data also indicate that the

AER maintenance activity that is defective in thesc mice is distinct from AER induction and di fferen tiation.

IV.D.1 Ce11 death and ce11 proliferation in Dac mice and AER maintenance The fact that a significant reduction in cell proli feration could be observed in the AER pnor to any change in tissue morphology of the Dac mutant strongly suggests that ce11 proliferation in the AER is closely associated with. if not itself. the pnmary defect in the

Dac rnice. Since AER doçs not grow significantly in size during early stages of limb development. the active ce11 prolifention is apparently compensated by the high PCD at the same time (Milaire and Rooze. 1983: Vaahtokari et al., 1996; and this study). A high rate of ceIl turnover in the AER is therefore probably a process required for proper cellular differentiation and morphogenesis of the limb structures. It is interesting to note. that the data presented here suggests that the cells of the AER may not represent a homogeneous population, as some cells appear to act as stem cells while others follow a ce11 death pathway.

The reduction in the rate of ce11 proliferation in the affected regions of the mutant AER was observed prior to any change in tissue morphology. Since the amount of PCD is natunlly high in the AER and unchanged in the Dac mutant AER, it is this failure to proliferate that results in the degeneration of the tissue. The data shed new light on the role of ceIl proliferation in the AER and argues that the function of the AER. maintenance factor is to promote cell prolifention.

In contrast to the findings presented in this study. Seto et al. have recently reported an observed increase in ce11 death in the AER of Dac heterozygote limbs. using Nile blue staining (Seto er al.. 1997). They have reported strong staining in the AER of E10.5 forelimb and hindlimb in 50% of embryos from Duc/+ X wildtype litters. Interestingly. they find little or no staining in the AER of the other 50% of the embryos. From these observations the authors conciude that an increase in ce11 death in the AER is the cause of the Doc phenotype. At this point there is no obvious explanation for the discrepancies between the previous report and the data presented here. Nevenheless, the studies here clearly show that there is no significant increase in ce11 death in the mutant AER as compared to the wildtype.

We reason that the degenention of the AER lads to the decrease in ce11 proliferation in the progress zone and subsequently a lack of cells available to give rise to the skeletal elrments. Previous studies have shown that removal of the AER prior to stage 21

(Hamburger and Hamilton. 195 1) in chick embryos results in ce11 death of sub-ridge mesenchyme (Rowe et al.. 1982). If the AER is rernoved after stage 7 1. however. no ce11 death is observed, but Iimb bud outgrowth is still affected. It was shown in this study that there was little PCD in the mesenchyme of the mutant limb buds following loss of the AER. This obsewation is consistent with the results in chick ernbryos since the AER is lost in Dac limb buds after E10.5 which is developmentally comparable to stage 22/23 in the chick (Hamburger and Hamilton, 1% 1).

IV.D.2 AER maintenance and the ZPA In chick, the AER is thicker in the posterior than the anterior limb bud. This difference has been interpreted to be a result of stronger AER maintenance activity in the posterior limb than in the anterior (see Zwilling, 1961 for review). It has also been postulated that an increase in the AER maintenance activity in the anterior limb bud is the cause of polydactyly (Zwilling and Hansborough, 1956). It is now known that polydactyly in many mouse mutants is likely a result of ectopic Slth expression in the anterior limb bud (Chan et al.. 1995: Büscher et al.. 1997: Masuya et al.. 1997). Together. these observations suggest that the ZPA has a close relationship to the AER maintenance activity.

On the other hand, studies in chick designed to address the endogenous role of the ZPA in

limb development have shown that if the activity of the ZPA is blocked by insening an impermeable barrier through the limb bud that intersects the AER. the AER is still maintained (Summerbell. 1979: Kaprio. 198 1). Funhermore. data presented in this study

have shown that the expression of Shh. a marker for the ZPA. is normal despite the loss of

the AER maintenance activity. Thus. while it is clear that the ZPA and the AER

maintenance activity are distinct it is possible that the ZPA does play a role in the initial induction of this AER maintenance pathway and that ectopic ZPA may induce ectopic

maintenance activity. In addition. it is conceivable that the ZPA may have a more direct local efkct on the maintenance of the AER in the posterior limb bud as will be elabonted below . IV.D.3 Dac as a mode! for AER maintenance In the Dac mice. the anterior and posterior AER remains intact in heterozygous mice while the anterior AER is lost (albeit more slowly than central tissues) with only postenor AER remaining in homozygotes. Therefore. not al1 of the AER is lost in DoclDac mice. The semi-dominant nature of the Dac allele strongly suggests that it is a loss of function mutation. Nevenheless. it is possible that the mutation may not result in a complete loss of function. The data supgest that the requirement for Dac is more stringent in the central

AER thnn in the anterior or posterior AER. We propose a mode1 for AER maintenance

(Fig. IV-7). where the Dac activity plays a role in the maintenance of the entire AER. but additional or seconday secreied signal(s) originating from the posterior and anterior mesenchyme may partially substitute for the Dac pathway. This mode1 requires that the anterior activity be wcakrr than the posterior activity. The secondary maintenance activitics would be at lower levels in the central lirnb bud and thus be below threshold amounts to compensate for the Dac pathway. Hence. the central AER would be more susceptible to the loss of the Dac activity. In heterozygous animals. partial Dac activity combined with modente levels of the additional activities would be enough io maintain the

AER in the intrrmcdiate region between the center and most anteriodposterior limb. In the absence of Dac activity. however. only high lrvels of the secondary factors would be able to rnaintain an AER. An alternative hypothesis is of course possible whrre the presence of the small amount of functional .\ER in the hornozygous mutants may be due to the presence of residual Dac activity. We favor the previous rnodel since the latter would imply that the different requirement for the Duc activity throughout the AER dong its AP mis would be intrinsic to the AER itseIf. Figure IV-7. Mode1 for AER maintenance in the mammalian limb bud. In this model we show Dac present in the distal mesenchyme. Mesenchymal localization of Dac is used to represent the expression of the putative maintenance factor present in the pathway whose function is disrupted in Dac mice. Mesenchymal localization of the Dac gene product. however. is not a prerequisite for this rnodel. In wildtype limb buds. the Dac pathway maintenance activity is expressed in the distal limb mesenchyme undrrneath the AER. Posterior (Slih as a candidate) and anterior (BmpJ as a candidate) localized maintenance ûctivities diffuse from a local source. Gradients of maintenance aciivity are created by overlapping activities from the Dac pathway and diffusing activities from the localized sources. A threshold lrvel of total activity from al1 sources is required to maintain the AER. The posterior activiiy is stronser or more abundant thon the anterior activity. In Dac/+ mice the activity of the Dac pathway is reduced in half. assuming Dac is a loss of function nul1 allele. The level of maintenance activity drops below threshold levels in the central AER leaving only the anterior and posterior ridge. In Daflac rnice complete loss of Dac pathway activity leaves only the most posterior mesenchyme wirh sufficient activity to maintain the AER.

IV.D.4 Candidates for the AER maintenance factor(s) It has recently been suggested that FgflO plays a role in the maintenance of the AER (Ohuchi et al.. 1997). In this study we do not detect any significant change in FgflO expression in Dac mice despite a defect in AER maintenance (Fig. IV-6). It should be stated that an apparent. slight decrease in expression of FgflO in mutant forelimbs is likely a result of the failure of the AER to express F@. It has been shown that the expression of

FgfI0 is dependent on F@ in chick limb experiments (Ohuchi er u1.. 1997). In addition. no change can be observcd in EL03 hindlimb blids (data not shown). whcre therc is no loss of expression of FM.

The rnolccular defect in these mice could be localized to the AER and play a role in receiving. intrirpreting or processing the secreted maintenance signal and thus may na effect expression of the sccreted mesenchymal maintenance factor. Localizing the defect in

Dac mice to the mesenchyme or AER utilizing heterografting studies between Dac ectoderm and chick lirnb bud n~esenchymewill prove to be informative in clarifying what role FgflO may have in this process. Studies of ectopic FgflO in chick sirongly argue for its role in AER induction, as well as maintenance (Ohuchi et al.. 1997). The data presented here show that AER maintenance is distinct from AER induction. and therefore. if FLf10 docs play a rols in maintenance. its expression or function rnay be differentially regulatrd brfore and after E 10.5 in rnice.

Insulin has also been proposed as a possible AER maintenance factor. When the AER is cultured in the absence of mesoderrn. it rapidly dies (Boutin and Fallon, 1984). However, if the culture medium is supplemented with insulin. the AER rernains viable (Boutin and Fallon. 1984). Nevertheless. the endogenous role of insulin or its receptors in AER maintenance is not known. Both Shh and its downstream mediator Bmp2 are expressed in posterior mesenchyme and their expression is unaffected in the Dac mice (Fig. IV-SA,B, and data not shown). They are strong candidates for the additional (or secondary) AER maintenance signal locdized to the posterior limb bud. In support of this, Shh is known to be required for the induction and maintenance of Fg/J expression in the posterior AER (Laufer et al.. 1994 (Niswander et al, 1994). an activity Iikely mediated by Bmp (Duprez et al.. 1996). Sith may also induce the Dac pathway. as suggested by the studies on the polydactyly mice discussed above. Moreover. Slzh may have an additional, more direct influence on AER maintenance because polydacty ly could be the result of a local influence of ectopic ZPA in the anterior limb bud. Therefore. the complete loss of the autopod in Shh mutant mice (Chiang et al..

1996) may be due to a combinrd loss of local AER maintenance in the posterior iimb bud and failure to induce the Duc pathway.

BrnpJ may Se considered a candidate for the anterior maintenance activity since it has a localized expression domain in the anterior mesenchyme (Jones et al.. 1991) and is norrnally expressed in the mesenchyme of Dac limb buds (Fig. IV-5E.F'). However. mice heterozygous for a loss of function allele of Bmp4 have recently been shown to display a limb phenotype of pre-axial polydactyly (Dunn et al.. 1997). suggesting that this gene does not play a role in AER maintenance.

IV.D.5 Dac mice and human split handl split foot malformation

As discussed previously. the Dac mouse is a mode1 for split hanci/ split foot malformation

in hurnans and other sprcirs (Chai. 1981; McKusick, 1992; Johnson et al.. 1995). Our present data would argue strongly that AER maintenance is the general pathway affected in

SHM. This study of Dac hiis provided us with valuable insights into genes that might

play a role in controlling ce11 prolifention in the AER. Chapter V

Assessrnent of DLXS, DLX6, and DSS2 as candidate genes for SHFMl

Attribution of work:

The RNA in siru data presented in this chapter. dong with the cloning of rnouse Dix5 and Dix6 were carried out by the candidate. Cloning of human DLXS and DLX6 was performed by Dr. Stephen Scherer. Cloning of both mouse and human DSS2 was carried out by Dr. Stephen Scherer. David Sinasac. and Jeff Lee. an undergraduate student. V.A Introduction

Despite DSSI being a strong candidate gene for SHFMI. it remains unclear if it plays a direct role in the disease phenotype. Dssl. while being expressed during limb dçvelopment. is not directly intempted by translocation breakpoints. and no point mutation could be detected in the coding region of DSSZ in 60 spondic cases analyzed (see Chapter 1. While this negative data certainly does not exclude DSSZ as a candidate. a funher characterization of the orher gencs from the SHFM critical interval is warranted. As outlined, three additional genes. DLX5. DLX6. and DSS2 have been identified in the

SHFM 1 crtitical interval.

DLYS and DM6 are wtebrate homologurs of the Drosophila Distnl-less (dl0 homeobox gene (Sirneone et al.. 1994). Dl1 has been shown to be required for appendûge development (Vachon et ctL. 1992: Gorfinkicl et al.. 1997). Partial Ioss of function of dl1 leads to an absence of the distal slructures of the insect appendage (Cohen et al.. 1989). Given the degree of functional conservation that has been observed in genes involved in limb development (Tickle and Eichrlr, 1994). it suggests that vertebnte DU( genes would be required for proper limb development. Therefore. the presence of these genes in the SHFMI critical interval sugrsts an involvement for these genes in the disease phenotype.

As shown in figure II- 1. thesr genes map townrds the distal boundary of the SHFM 1 locus and are transcribrd in a tail to tail fashion (Fig. 11-1). While the 5' end of DU6 has not been cloned, it would apperir thlit rhese genes map at least 100 kb from the nearest translocation breakpoint.

The most recent gene to be identified frûm the SHFMI locus is DSSZ. This novel gene is rnost sirnilar to a putative C. rlrgnns protein. and has significant similarity to a family of mitochondrial carrier protrins. Whilr a functional relation of this gene to a congenital abnormality of the limb is not obvious. this gene is directly interrupted by disease associated translocation breakpoints, a feature that cannot be overlooked.

This chapter describes the cloning and analysis of DUS, Dm,and DSS2 from the SHFMI locus. Munne homologues for these three genes were isolated and used to analyze their expression patterns during embryogenesis. The expression of DlxS during mouse limb development is similar to that found in the chick and the rat (Zhao et al.. 1994; Ferrari et al.. 1995)- while Dlx6 expression is found in an overlapping yet distinct pattern in the developing limb. The expression of Dss2 is surprisingly highly specific, found primanly. but not exclusively in epithelial cells. Most interesting. is the observed strong expression of DssZ in the AER. in complete agreement with a potential role in AER maintenance. The data strongly implicate this sene in the disease process.

V.B Materials and Methods

V.B.1 Gene Identification, cDNA Clones and Libraries Isolation of human DM5 and DLX6 cDNAs have been described previously (Scherer et al.. 1994b). Mouse Dk5 and DLI-6 cDNAs were isolated, by low stringency hybridization screrning of a rnouse ernbryonic h-EXlon cDNX library (Novagen) with human cDNA clones FC3 and FB315 (Scherer et uL, 1994b). respectively. Low stringency hybridintion conditions were: 55'C hybridization in Church and Gilbert Buffer (Sambrook et al.. 1989); washing was carried out at 55°C to a final stnngency of 0.2~SSC. 0.1% SDS.

Sequencing of PAC clone H-GSOISMO? (see http://www.genet.sickkids.on.ca) was perfomed (Washington University). Blast searching of this sequence (Genbank) identified several EST clones from this region. corresponding to a previously unidentified transcnpt DSSZ. A 1.3 kb EST. clone #46806 (IMAGE) was used to screen a Caco-2 oligo dT cDNA library. A 3.1 kb clone was identified and sequenced (AM). To clone mouse DSSZ. EST clone W6806 (IMAGE) was used to screen a mouse embryonic h-EXlox cDNA library (Novagen). The full length mouse cDNA sequence was detemined by 5'- RACE (Life Technologies). E12.5 mouse embryo total cellular RNA was reverse transcribed with an anti-sense oligonucleotide primer (S-TATGAAGTCTCCATCTTAT- 3'). The first round PCR was perfonned in a total volume of 50 pl for 35 cycles with the primer annealing at 50°C. followed by a 'nested-PCR' with gene specific primer (5'- TCTCTAGTTCAGCTCGTAAC-3'). PCR products were subcloned into the pCRII TA cloning vector (Invitrogen).

V.B.2 RNA in situ hybridization Mouse CD1 (Charles River) embryos were used in this study. Staging of the developing embryos was detemined in reference to the first midday when vaginal plug was deircted as

E0.5. Whole-rnount in situ hybridization using a digoxigenin-labeled RNA probe and an alkaline phosphatase-coupled anti-digoxigenin antibody was performed as descnbed (Ang and Rossant. 1993). Frozen sections of developmentaily staged embryos were hybridized with digoxigenin-labeled anti-sense probes as descnbed previously (Schaeren-Wiemers and GerFin-Moser, 1993).

V.C Results

V.C.1 Identification of DLXS ,DLX6, and DSSZ from the SHFMl critical interval Independent of this study, DU(5 and DLYo had previously bcen cloned from mouse and

humnn (Simeone et ul.. 1994). FISH mapping showed that these genes mapped to human 7q22 region. A genomic subclone frorn a cosmid containing the DLXS gene was used to screen a frontal cortex cDNA library. A 1.5 kb clone was identified (FC3).that was believed to be full length based on the approximate transcript size found with Northem blot analysis (Simeone et al., 1994). FC3 was found to rnap into the YAC contig generated for the SHFM1 region. and pulse field gel electrophoresis placed this gene approximately 90 kb from the telomeric end of HSC7E571 (refer to Fig. 1-3; Scherer et al., 1993b). Phage clones positive for FC3 were found to span the distal deletion boundary of D3 suggesting that most or al1 of DM5 mapped within the critical interva[ (Scherer et al.. 1994b).

Cross hybridizntion of non-overlapping phage to the DM5 cosmid. suggested the presence of an additional related gene. Srquencing of subclones confirmed the presence of a second

DLX gene on the original cosmid. A genomic subclone containing the new DLX gene. called DLX6. was used to screen cDNA libraries. and one DM-specific clone (FB325) was isolated from a human fetal brain cDNA library (Srratagene). As mentioned above.

DM6 was concurrentiy identitïed by another group (Simeone et al.. 1994). and sequence cornparison confirmed this gene to be DLXo (data not shown). FB325 was only 1.4 kb and was dctermined to be less than full-lrngth on the basis of the size of the mRNA (7.5 kb) and the result of DNA sequencing (Simeone et al., 1994: and data not shown). DM6 was found to map approximately 70 kb from DU5 in a tail to tail orientation (Fig 1-3).

In an effort to identify additional potentiûl candidate genes from the SHFMI locus. A PAC clone from the SHFM 1 cntical region was subject to total genomic sequencing (carried out in collaboration with the Washington University genome sequencing initiative. St. Louis).

Sequence was compared with drpositrd EST sequences frorn the dbEST database identified numerous hurnan ESTs that aligned to the genomic sequence, that were not previously identified. Highly related mouse ESTs were also identified. Database searches also DSS2 Dai2 K02F3

DSS2 Dia2 K0 2 F3

DSS2 De82 K02F3

DSS2 D8.2 X02F3

DSSZ Di 8 2 K02F3

DSSZ 0112 K0 2 F3

DSSZ Der2 K02F3

DSS2 Dei2 K02F3

DSS2 ~8~2 K02F3

DSSZ Dia2 iC02F3

DSS2 Daa2 K02F3

DSS2 Daa2 K02F3

DSS 2 Des2 K02F3

Figure Amino acid sequence and evolutinary conservation of DSS2. Alignrnent of hurnan DSS2. rnouse DSSZ, and C. elego& predicted prorein ~02~3.2. Idenucal amino xids are shaded in blrick. and similar amino acids are shaded in grey. EF Hands Mitochondrial Energy Transfer Motifs

100 amino acid repeat

Mitochondnal Carrier Protein Family

Kyte and Doolittle

Figure V-2. DSS2 motif structure and hydropathy plot. (A) Motif structure of DSS2 with 2 or 3 EF-Hands, 3 Mitochondrial Enrgy Transfer motifs. Each energy transfer motif is pan ofa LOO amino acid repeat, and together these comprise a Mitochondnal CherProtein farnily mernber. Amino acid residue is shown below schematic of protein. (B) Kyte and Doolittle hydopathy plot showing six transmernbrane domains in the 3' end of the protein. showed that these ESTs had 55% identity to a putative C elegans protein K02F3.2 . A human EST clone was used to probe a human Caco-2 cDNA library which resulted in the identification of a full length clone based on sequence and aiignment with the worm homologue (Fig. V-1). This novel gene has been called DSSZ.

Using cornputer based motif identification DSS? appeared to be contain two distinct dornains. the carboxy terminal haif of the protein has hornology to the Mitochondrial

Carrier Farnily (MCF) of proteins (Kuan and Saier. 1993). while the amino terminal possesses at least two possibly 3 EF-hand motifs, which are typically found to be calcium binding domains (Ikura. 1996) (Fig. V-2h) One other protein in addition to DSS2 and

KOZF3.1 has been identified with both of these dornains and that is Efinal. a nbbit protein

(Weber er cil.. 1997). This gene shows weak identity to DSS2. however. and is therefore likely not the truc ortholog (data not shown). Further nnalysis of the structure by hydropathy plot shows that the Carboxy terminal half of the protein contains six trmsmrmbnne domains found in a tripartite organization typical of MCF proteins (Fig. V- 2B).

V.C.3 Identification of murine DM,DM, and DSS2 Mouse clones for DMand Dfx6 were obtained by screening a rnouse embryonic cDNA library at low stringency. using FC3 and FB325 as probes. respectively. Numerous positive clones for DMwcre identified. Sequencing of a 1.5 kb clone determined it to represent the full length DLrS cDNA (Sirneone et al., 1994 and data not shown). Screening of the same libraq for Dlit6 failed io identify the full length sequence of DLr6. but several clones encompassing the horneodomain and the 3' end of the gene were identified (Sirneone er al.. 1993 and data not shown). To obtain full length sequence of mouse DSSZ, a mouse embryo cDNA library was screened with a DSSL EST. A number of cDNAs were identified but none were found to be full length. To obtain the full length sequence of the mouse DSSZ cDNA. S'RACE was carried out usinç mouse tmbryo RNA as a template. Sequence comparison showed that the predicted amino acid sequence of rnouse DssZ is 96% identical to the hurnan protein (Fig. v- 1).

V.C.3 Analysis of the expression of Dix5 and Dix6 during mouse limb bud development

Using wholr mount RNA in sini hybridization the expression of DL5 and DM during mouse limb development was analyzed. At ElO.5 DLr5 was expresscd in the pro,oress zone mesenchyme. with stronger anterior expression (Fig V-3A). More strikin~was the strong expression of this gene in the AER. Expression of DLt5 was also found in the presurnptive

AER prior to E 10.5 (data not show). At E10.5 Dlx6 could also be detected in the distal mesenchyme. howrver. thrre appeared to be no asymmetry. as sren for DLrj (Fig. V-3D).

Expression of DMwas also found in the AER but, unlike Dkj. expression in the rest of the limb bud epithelium was dso detrcted. At €12.5 the expression of DM was found in the distal mesenchymr as well as the posterior necrotic zone (Fig. V-3B). At E13.5 D1x5 expression was restricted to the distal tips of the developing digits and discrete regions of the digital condensations (Fig. V-3C). DMat E12.5 remained distally restricted in the mesenchymr. in addition to the persistence of expression in the epitheiium (Fig. V-3E). At E 1 3.5 Dlx6 expression was restricted to the mesenchyme surrounding the prospective digits (Fig. V-3R

1.4 Analysis of the expression of DssZ during mouse embryogenesis

The expression of Dssl during mouse development was analyzed using RNA in situ hybridization with both whole mount embryos and tissue sections. At E10.5 DssZ was found to express in the limb buds. branchial arches and tailbud. By section analysis it was found that at this stage Dss2 was expressed in the cephalic mesenchyme and the mesenchyme and epithelium of the branchial arches. Most interesting, strong Dss2 expression was detected in the AER of the E10.5 limb bud. Also observed was a weaker expression in the underlying mesenchyme. Later during embryogenesis DssZ expression was found predominanrly in most epithelial cells of the embryo, including the Iungs. intestine. kidnry. pancreas. gall bladder. submandibular gland nasal epithelium. and

Jacobsen's organ (Fig. V-5). Expression u1asalso detected throughout the liver. thymus and the mesenchyme of the tooth buds. Figure V-3. The expression of D1x5 and Dlx6 during mouse limb development. Whole mount RNA in situ hybridization on developing mouse limb buds. (A) At E10.5 DMis expressed in the distal limb mesenchyme and the AER (arrow). (B) At E12.5 D1x5 expression is found throughout the distai handffoot plate and in the posterior necrotic zone (arrow head) and the condensing mesenchyme of the long bone blastema (anow). (C) At E13.5 D1x5 expression is found at the distal tips of the digits and in the developing cartilage of the digits (arrow). (D) At E10.5 D1x6 is found in the distal mesenchyme of the limb bud and throughout the limb bud ectoderm. (E) Expression of D1x6 at E12.5 is sirnilar to that observed at E10.5. (F) At E13.5 D1x6 expression is found in the mesenchyme surrounding the formed digits. Figure V-4. Dss2 expression in the mouse embryo from E10.5-E11.5. (A)Whole mount analysis of Dss2 expression at E 10.5 shows staining in the branchial arches (1st ba), limb bud (lb) and tail bud. (B)Analysis of Dss2 expression on tissue sections shows specif'ic AER staing in the E10.5 limb bud (C) Dss2 expression is also found in the mesenchyme (me) and epithelium cells (ep) of the branchial arches. (D) At El 1.5, Dss2 expression is found in the cephalic mesenchyme (cm) Figure V-5. Dss2 expression in the mouse embryo at E13.5. Analysis of Dss2 expression on tissue sections show that (A) at E13.5, Dss2 is found in most organs predominantly localized to the epithelial cells expression is also &tecred in the CNS. (B) Higher mapification shows expression in the intestinal epithelim (ie), (C) whisker folicles (wf), (D) kidney epithelium, throughout the liver, (E) and lung bud epithelium (le). (fi), forebrain; fi),Liver; (k),kidney; (lu), lung. V.D Discussion

A number of previous studies in mouse (Simeone er al., 1994), rat (Zhao er al.. 1994) and chick (Ferrari et al.. 1995) indicate that DLrj and DLr6 are expressed dunng development in a unique spatial and temporal pattern. The expression pattem of the mouse Dlx5 and

Dl.r6 genes were found to be alrnost identical (Simeone et al., 1994). Both show strong signals in facial and branchial arch mesenchyme. otic vesicles. and frontal cctoderm around olfactory placodes at E8.5-9. and in the developing forebnin in prirnordia of the ynglionic eminence and ventral diencephalic regions at E10. At midgestation (E12.5). they are expressed in almost every developing skeletal element. and in the forebrain. Expression of

Dl.r.5 cm also be detected in regions of ossification. as well as ear ossicles and primordia of teeth. This previous analysis of the expression patterns of Dlxj and Dix6 in mouse. howrver. did not address early limb development (Sirneone et al.. 1991). In the rat. the expression pattern has only brcn studied with DM and the pattern is essentially the same as that of the mouse. but showing an additional site in epidermis of the skin and the apical rctodermal ridge (AER) of limb buds (Zhao et al.. 1994). The study of Dl.r.5 in chick is primarily focused on limb development and the description of its expression pattern is more refined (Ferrari et al.. 1995). In addition to AER. the expression of the chick Dl-~jjene is found in the mrsoderm at the anterior marsin of the limb bud and in a discrete group of mesodrrmal cells at the rnid-proximal posterior rnargin that correspond to the posterior necrotic zone.

The expression pattem for Dlxj dunng limb development presented here is similar to that sren in the chick. except that expression is found throughout the entire distal mesenchyme in rnice. The expression profile of DLr6 during limb development has not been previously drscribed. In addition. the data presented here revealed differences in the expression profile of Dlxj and DM, which is in contnst to the previous study of the expression of these two genes in the developing mouse embryo (Simeone et aL. 1994). which suggested that the expression pattern of these two genes was identical.

Five additional veriebrate homologues of dl!. have been cloned and characterized

(Robinson et al., 199 1 ; Do116 et al., 1992; Ozcelik et al.. 1992; BuIfone et al., 1993;

Robinson and Mahon, 1994; Nakamura et al.. 1996: Quinn et al.. 1998). Interestingly, al1 of these senes analyzed are expressed in the drveloping vertebrate limb bud. Nevenhelsss. two of these genes, DLrl and Dld. have been knocked out in mice. and both single hornozygous mutants as well as double mutants develop normal limbs (Qiu et al.. 1995;

1997: Thomas et al.. 1997). This lack of limb phenotype may suggest that there is a certain degree of functional redundancy between these genes.

Based on the RNA in situ hybridization data. DLXS and DLX6 are equally strong candidates as DSSl in explaining the SHFM and SE phenotypes. Most interesting is their expression in the AER. the structure not maintained in the Duc rnouse. It is of importance io note. howe\w. that. iike DSSI. no mutation could be detected in these genes in sporadic

SHFM patients (Genuardi rr al., 1995). Moreover. although DU5 and DLYo are located within the critical interval defined by the deletion breakpoints. thry are not intempted by the translocation breakpoints. If reduced expression of the 2 genrs is part of the eiiology of

SHFM and SE. they might be regulated by distantly located control elements which becorne sepanted in the translocations.

Dcspitr the identification of three candidate genes from the SHFMI region. the lack of idrntifird point mutation and interrupting translocation breakpoints. suggested the possibility that the causative gene had yet to be identified. Xfter exhausting the possibilities of cDN.4 selection. cxon-trapping. and searching for consencd genomic sections. we resonrd to sequencing the entire SHFM 1 critical region. By doing so. a new gene was identified. named DSS2. that was directly intermpted by four translocation breakpoints.

The expression pattern of DSSZ is also highly consistent with a role in SHFMI since it is expressed in effected tissues, like branchial arches, tooth buds ectodermally derived structures, and rnost importantly the AER. The combination of map location and expression profile, considering what is known about the developmental mechanism of

SHFM, argues strongly for a role of this gene in the disease process. Nevertheless. there have been 5 additional translocation breakpoints mapping to the SHFM 1 critical interval that do not interrupt DSS-7. and thus, other mechanisms such as position effects or distant cis control elements musc stilI bt3 present.

Sequence analysis predicts DSSî to be a mitochondrial carrier protein that can bind calcium ions (D. Sinansac. unpublished data). While this does not lead to an obvious association with SHFM 1. two recently identified mitochondrial carrier proteins, dif- 1 in C. eleguns

(Ahrinzer. 1995). and colt in Drosophib (Hanenstein er al.. 1997). have both been show to be involved in developrnental processes. thus DSSZT putative biochemical function does not exclude it as a candidate gene. Chapter VI

Discussion and Future Directions At the time this project was initiated. it was thought that Our gene identification strategies would lead to the identification of a single gene that was intemipted by most if not ail the translocation breakpoints associated with the disease. The process of identification of the causative genes for SHFMI has proven to be more difficult than expected. however, and the answer is still unclear. We have identified from this region a number of genes whose spatial and temporal expression pattern during mouse limb bud development are al1 consistent with a possible role in the disease process. In addition. the dynein gene from the critical region suggests a role in the atypical inhentance pattern of this disease.

V1.A Possible Molecular Etiology of SHFMl

A number of potential scenarios have been presented to explain how the deletions and translocations identificd at the SHFM 1 locus could cause the SHFM phenotype (Scherer er al.. 1994b). First. there could be a large gene that was disrupted by al1 the translocaiion breakpoints found in the patients. This explanation requires the presence of a gene of at lest 700 kb in size. spanning most. if not dl. of the translocation breakpoints. Despite exhaustive searching. we have not identified such a gene in the region. Second, it was suggested thrit there may be a number of genes present in this region that are essential for limb development and that disruption of expression of any or al1 of them could lead to the observed phenotype. This hypothesis is supponed by the identification of 4 genes that are expressed in the limb bud. A third hypothesis that apprars attractive is based on the notion of gene regulation by distant cis-acting regulatory elements. or long range position effects

(reviewed in Henikoff. 1990). Separation of important promoter/enhancer elements from the coding regions of the causative gene(s) may result in aberrant gene expression leading to the abnormal developrnent. Altention of the gnomic environment of the causative gene may have sirnilar resuits. Obviously, the latter two hypotheses are not mutually exclusive. Interesting is the apparent lack of genetic Iinkage with families without cytogenetic rearrangement to the SHFM 1 locus. Furthemore. no point mutation has been identified in

DSSI. DMand DU6 to date. in 60 individuals with the sporadic disease. This negative data. however. does not exclude these genes as candidates. Currently. a search for mutations in DSSZ is being camed out. The absence of linkage and point mutations leads to the speculation that the oniy way to achieve the disease phenotype at this locus is by cytogenetic rearrangement. Possibly. the expression of more than one gene at rhis locus must be affected. The presence of 4 candidate genes in this region with overlappin_p expression patterns supports this idea.

The presence of the two DLX genes in this region is very interesting with regards to a multigene hypothesis. As presented. good evidence exists for functional redundancy of

DUpnes during limb development (Qiu et al.. 1997). Loss of function of one copy of either DLXS or DMmay not cause a phenotype because of this redundancy. However.

loss of a single copy of both DU5 and DLX6 may result in abnormal limb developmrnc. Despite this attractive hypothesis. knockout mutation in mice for DLrj and DLr6 have failed.

thus far, to support a rolr for these genes in normal limb development. Mutations in DM do not result in a limb phenotype. even when hornozygous (J.L. Rubenstein. persona1

communication). while heterozygous Dix6 and double heterozygous DLrjIDLi-6 mutations. also do not affect limb drvelopment (T. Lufiin. persona1 communication). It is possible

that a hornozy_oous mutation of Dl.r6, or double homozygous mutations. may result in a

limb phenotype. but this would be a different scenario than that found in the human condition. Thus. whilr these genes cannot be completely excluded as candidate genes from these experiments. they are certainly less attractive. A further possibility is that the

mutations in these genes must be in the proper genetic backgound. such as a niDachriDac

background in order for the limb phenotype to be observed (see Chapter 1). Thus. breeding these mutations into these backsrounds may prove informative. A second hypothesis to explain the absence of point mutations concems the presence of a srnall gene like DSSI in the SHFM 1 region. Since this gene is so srnall, the chance of a random DNA polymorphism occumng in the coding region of this gene would be statistically very low. So low. in fact that it may not have occurred in the human population. Subsequently. patients with spondic SHFM may only have mutations in genes at one of the other SHFM loci. Funhermore. a point mutation in the coding region of this gene may not result in a loss of function mutation but in a more severe gain of function or dominant negative mutation. that may lead to more severe, possibly embryonic lethal phenotypes.

V1.B The Role of DSS2 in SHFM1

While one cannot overlook any gene identified from this region. the expression of DSS~in

the AER at €10.5. dong with its disruption by 4 translocation breakpoints. rnakes this gene an intriguing candidate. The biochemical function of Dss2 is not known but the putative protein likely funciions as a calcium dependent ATP transporter in mitochondtia. At first

thought it is difficult to rnvision a role for this gene in a development defect. On the other hand. as deterrnined by the analysis of ihr Dac mice. the failure to maintain the AER may

result from the loss of the balance between ce11 proliferation and ceIl death. suggesting that the two processes in the AER are intimairly connectrd. Mitochondria are known to be

critical organelles in the cellular process of PCD (Reviewed in Kroemer et al., 1998).

The permeability transition. which underlies mitochondrial changes that occur duting the rarly phases of cell death. is a well-characterized phenornenon where a megachannel complex is opened in the mitochondria. releasing numerous factors to the cytoplasm

(Zoratti and Szabo, 1995). It has been shown that this transition occurs prior to any

changes in ceil morphology associated with apoptosis. such as nuclear envelope blebbing and chromatin condensation (Kroemer et nl., 1998). it has also been demonstrated that the factors released from the mitochondria following the penneability transition cause these ceIl alterations (Susin et al., 1996; 1997; Kanuow and Piantadosi, 1997). Thus, it is clear that mitochondna control ce11 death.

While most of the components of the rnegachannel have not been identified. clearly one component is the mitochondrial addennine nucleotide tnnslocator (ANT) (Brustovetsky and

Klingenberg. 1996). Interestingly. DSS? is most closely related to ANT (see Fig. V-2) and it is possible that DSSZ may function as pan of this rnegachannel. as well. Of funher interest is the presence of the EF-hands in DSS7. A role for calcium ions in apoptosis is well documented (Lipton and Nicotera. 1998: Nicotera and Orrenius, 1998). and it is known that an increase in intracellular calcium concen trütion can activate the permeability transition (Kroemer et al., 1998 and references therein). Thus, one may spsculate that

DSS:! rnay serve to sense the level of ~a?+in ihr dl.and thus. control the function of the megachannel.

To begin to address the cellular function of DSS2. its subcellular localization must be deirrrnined either with epitope iagged constructs or with the use of anti-DSS2 mono- or poly-clona1 antibodirs. While mitochondrial localization is a strong possibility it is not absolute. Should DSSZ play a role in the permeability transition. a mutation in this protein may lead to aberrant control of ce11 death and possibly indirectly a decrease in ce11 proliferation leading to a failure to maintain the AER. To test this hypothesis one could express mutant DSS2 in culture. where the calcium binding domains would be deleted or mutated. and analyze its effect on calcium induced apoptosis. One may observe a failure to promote ce11 death. or aitematively. constitutive cell death. V1.C The Biochemical Function of DSSl

DLX5 and DLX6 certainly function as transcription factors. However. the biochemical function of DSSl is not clear. Sub-cellular localization shows that it is preferentially nuclear localized but not exclusively. Recently . two inde pendent groups have-identified a Sacchnroniyces cerevisiae homologue of DSSI. in various screens for different cellular functions. In one case this gene was shown to play an important role in nuclear expon (Dr. C. Dehoratis, personal communication). More recently. DSS 1 in yeast has been identified as a cdc mutant playing a role in cell cycle control (Dr. K. Shirahige, persona1 communication). In the case of nuclear cxport human and/or mouse DSS 1 was able to functionally replace the yeast protein. arguing for functional conservation. Based on these findings it can be suggested that DSS 1 may play a role in similar intracellular processes in mammals. and also may be very important for cellular function across al1 eukaryotic species.

To begin to examine the potential roles of DSS 1 in these cellular processes. we can begin to idrntify proteins with which DSS 1 may interact. This straiegy could potentially identify an interacting protein known to be involved in a specific cellular process and thus quickly allow for the funher analysis of DSS 1 in this process. Recently. we have attempted to screen a mouse embryonic expression library with radioactively labeled DSS 1 produced in bacterial cells. however. positive clones wrre not identified (data not shown). While this sxpenment was not canied out under varying conditions. we may altematively choose to directly test for the interaction of DSSI with proteins known to be involved in nuclear transport.

Nuclear protein impon and export are mediated by receptor proteins that recognize nuclear localiziition sequences (NLS) or nuclear expon sequences (NES) and target the NLS- bearinz or NES-bearing proteins to the nuclear pore cornplex. Translocation of the receptor-cargo complex in both directions requires the GTPase Ran. and it is thought that it is the GTP hydrolysis that drives the movement of complex through the nuclear pore

(reviewed in Gorlich. 1997; Ohno el al.. 1998). A number of Ran binding proteins have been identified which act as these NLS or NES receptors (Deane et al.. 1997; Pemberton et al.. 1997; Yaseen and Blobel. 1997; Nachury et al.. 1998). Upon careful examination of the DSS 1 sequence a smali region of sirnilarity was found with the Ran binding domain of numerous Ran-binding proteins (Coutavas et al.. 1993; Moroianu er al.. 1996; Yaseen and

Blobel. 1997). This may suggest that DSSI rnay also function in nuclear transport by binding to the Ran GTPase.

V1.D Position Effect in SHFMI? The role of position effect at the SHFMl locus is an attractive hypothesis. While DSS2 is interrupted by 4 translocation breakpoints, it rernains hundreds of kilobases away from the other 5 breakpoints. Furthemore. the remaining genes are not interrupted by translocations at all. The expression of DSSI may be panicularly susceptible to the proposed position effect because the gene is closely surrounded by the translocation breakpoints that do not interrupt any other known genes in the region. On the other hand. position effect type mutations have been shown to occur hundreds of Kb away from the causative gene. thus the othrr grnes certainly are not excluded from this mechanism.

Position e ffect. is best describe in Drusuphilci where a chromosomal rearrangement places a euchromatic gene next to a heterochrornatic region. As a result heterochromatinization of this gene can take place thereby altering gene expression (reviewed in Henikoff. 1990). Hrterochromatic regions are condensed hypemethylated regions of the chromosome where Little gene expression occurs. In recent years. position effect has been irnpiicated in numerous examples of mammalian disease. such as SOX9 for campomelic dysplasia (Wagner et al.. 1994). PAX6 for anindia

(Fantes et al., 1995), GLI3 for Greig's cephalopolysyndactyly (GCPS)(Vortkarnp et al,, 199 1). the Steel-locus for ovarian follicle development (Bedell et al., 1995), the agouti coat color gene (Duhl et aL. 1994). POU3F.I for familial X-linked deafness (de Kok et al..

1995). and myosin VI for the Stiell's waitzer deafness (Avraham et ai-. 1995). For some of these genes, expression is affected by rearrangements 50-400 kb away.

Determining the role of position effect in SHFM 1 is not an easy task. Nevenhcless. some indirect methods of establishing an alteration of expression as a result of the translocation are possible. Analyzing the level of expression from patient cell linrs is one possibility.

This rnethod. however. is not lavored since the gene must be expressed in the available cell lines (most likely lymphoblasts), and a negative result does little to answer the question.

Embryonic cells are those concrmed in this disorder. and gene regulation in adult cultured cells may be very different from that found in the embryo. The rffect of a translocation may also diffcr between ernbryonic cells and those of the adult organism. Therefore, it stands to reason that for an experirnent such as this to really be informative, it should be cmied out with embryonic tissue. This, of course. is not possible.

If position effect does occur. one would expect that the influence from the novel surrounding DNA rnay serve to turn off gene expression of the causative gene. While it is not entirely clear. it is known from studies in Drosophila that heterochromatin cm spread into the misplaced euchromatic DNA since buffer elements are absent. thus turning the

DNA from which the gene resides into relatively inactive heterochromatin (Henikoff,

1990). Abnormal rnethylation may be one mechanism for turning off gene expression in this scrnario. Therefore a second way in which to assay for a position effect in SHFMl would bc to determine methylation patterns in and around the candidate genes using DNA from translocation patients. The alteration of methylation of these genes can be tested by analyzing the pattern of restriction enzyme digest using methylation sensitive restriction enzymes. I have in fact attempted this type of analysis in a number of ~nslocationpatients

for DSS 1. however. no alteration of methylation pattern could be detected. Nevertheless, this is an indirect test and a negative result is not conclusive.

It has been shown that allele specific gene inactivation may be caused by an alteration of

replication timing at a given ailele (Chess et al.. 1994). Funhermore. this mechanism of allele specific inactivation has been shown to be caused by altered patterns of DNA methylation, for example on the inactive X chromosome or regions of gene imprinting

(Jablonka et al.. 1985: Kitsberg et al.. 1993) Therefore. a third way to study the affect of a translocation on gene expression is to analyze the replication timing of a given gene. Altered

mcthyiation as a result of heterochromatinization may lead to an asynchrony in replication timing and thus allelic inactivation. Preliminary experiments with DSS 1 have indeed

revealed such an alteration of replication timing in one translocation patient. suggesting the possibility of a position effect (Dr. J. Evans. personal communication).

V1.E Generation of Mouse Models for SHFMl

Dur to the lack of proper clinical material (Le.. n linked farnily). the search for point mutations rnay prove to be unsuccessful in the identification of the gene that causes

SHFM I. In our lab we have chosen to employ a direct analysis of gene function by zenerating knockout mice for the various genes in the hopes of uncovering a SHFM b

phenotype. Much work has been carried to genente Dssl mutant mice without success.

Wr are also working towards the generation of Dss2 mutant mice. It is our hope that these endeavon will provide the key to the molecular defect in this disease. Nevertheless. it remains possible that as sugzested. more than one gene musc be affected to give rise to this disease. If this is tmly the case. double mutant mice may need to be genented. Breeding expenments should allow for this to be investigated. On the other hand, there remains the possibility that a defect in more than one of the genes must occur in cis. This would be statistically impossible to accomplish through breeding strategies since the senes are tightly linked. An alternative approach then is to generate a one Mb deletion in the mouse that corresponds to the SHFM 1 smallest region of overlap. an approach that has bern shown to be technically feasible (Ramirez-Solis et ai.. 1995). This should in effect mimic the deletion found in the patients. Altematively. similar technology could be employed to generate balanced translocations in mice that mimic those found in the human

patients. This. however. is not as favourable as generating a deletion since it would require

prior knowledge concerning the precise location of the second breakpoint on the other

chromosome. While these methods will not directly lead to the identification of the disease

orne. ir will provide a model for which to analyze the phenotype and determine its similarity C to that found in Dac. Furthemore. double mutants made by crossing with Dac mice will

allow one to begin to address the genetic and biochemical interactions between the different

SHFM loci. Ln addition. a model like this may not be completely ineffective in identifying

the causative genes. Once the model is established. one at a time candidate genes could be

reintroduced in to the deleted mice. by simple transgenic protocols. Adding back the real

SHFM I gene may rescue the phenotype.

V1.F Other SHFM Loci

As rnentioned in Chapter 1. an additional autosomal locus for SHFM has been mapped to chromosome 2 by a small of number interstitial deletion patients. The region of

chromosome 7 implicated in SHFMI is one that has maintained its gene content over the

course of vertebrate evolution. such that 4 loci exist where one finds clustered highly

related genes. For exarnple on chromosome 7 we find the HOXX cluster (Boncinelli et al.. 1988) (even though in humans it is on the short arm) two DLX genes, a dynein gene, and

DSS3 (Simeone et ai.. 1995; Vaughan et al.. 1996; and this thesis). On chromosome 17

the HOXB cluster is found dong with DU3/7 (Nakamun et al., 1996). On chromosome

2 we find the HOXD cluster (Boncinelli et al.. 1988). DM],DLXZ (Rossi ei al., 1994). and DNCIZ (Vaughan er al., 1996). In addition we have identified a highly related gene to

DSSZ. MY.that also maps to this locus (our unpublished observation).

The consrrved gene cluster on chromosome 2 has been found to map to 2q3 1.1 (Rossi er

riL. 1994). within the region cornrnonly deleted in the SHFM patients. Thus. it is

intriguing to speculate that the related locus on chromosome 2 may harbor the gene

responsiblr for the disease on this chromosome. It would further aque that the causative

genr on chromosomr 7 would be the one homologous to the causative gene on chromosome 3. In this sccnririo. DSS2 and PKY become the best candidates. Dlxl and

DMhave both been knocked out. and double mutants have also been constructed (Thomas

er dl.. 1997). In each case no limb phenotype is present. While DNCI2 is expressed in the

limb in a pattern consistent with a role in the disease process. DNCII clearly dors not play

a direct rolc in the phenotype. Prelirninary analysis of Pky shows that while its overall

expression pattern is different from DssZ it too is expressed during limb development (Our

unpublished observations).

Since patients with cytogenetic renrrangement are rare for the chromosome 2 locus. it is

critical to identify a SHFM family with a defect at this locus. by linkage analysis.

Identification of such a family would allow for the direct identification of point mutations in

the candidate senes. Prelirninary attempts to identify such a family have. thus far. been

unsuccessful (our unpublished observation). V1.G Further characterization of the Dac mouse The analysis of the Dac mouse has provided a framework from which to further investigate not only the SHFM phenotype but the ce11 and molecular biology, and genetics of the maintenance and function of the AER.

It is important, however. considering the data presented by Seto et al ( 1997) that a closer look be taken at the rate of prognmmed ce11 death (PCD)in the wild type and the mutant

AER. It is possible that at E10.5 little increase in the rate of PCD has occurred. but shonly aftenvard there is a massive increase leading to the absence of the AER at El 1.5 (the next time point analyzed) As well analyzing the rate of ce11 proliferation and ce11 death before ElO.5 will help detïne the precise nature of this balance between the two cellular procrsses.

As mentioned from my analysis of the Dac mice it is not possible to pinpoint the molecular dcfect to the mesenchyme or the epithelium. Cloning of the Duc gene rnay resolvz this problem. but it is possible thot the gene will be expressed in both the AER and the mesenchyme as seen for DssZ. DlxS and Dlr6. Thus, the critical location and the precise biochrmical and cellular function of the protein rnay remain unclrar. A sirnilas case has been seen in the Ld mice where predominant expression is found in the AER. with weak expression in the mesenchyme (Chan er al., 199%; Haramis er al.. 1995. h'owever. experirnents have shown ihat the defect that leads to the abnormality in fact lies in the

rnrsenchyrnr (Kuhlman and Niswander. 1997)

Using the same approach that was employed in the analysis of the U mutation (Kuhlrnan

and Niswandrr. 1997) it is possible to determine the location of the defect in Dac mice by performing hetrrospeciçs grafting experiments with mutant mice and chick embryos. In

this scenario. chi& limb buds are rxcised from stage 19-20 limb buds (Hamburger and

Hamilton. 195 1). and rpithelium is consequently dissociated from mesenchyme and discarded. Similarly. epithelium frorn mutant mouse embryos is isolated. The mouse epithelium is then placed over the chick mesenchyme, and they are allowed to re-associate in culture. Subsequently these reconstituted limbs are grafted to the Rank of a host chick embryo. Should the defect reside in the AER in Dac mutant mice. one should observe a SHFM like phenotype in the chick limb. If the defect resides in the mesenchyme then normal limb deveioprnent should ensue. A dnwback to this system is that one may need to rely on a negative result (Le.. normal limb developrnent to pinpoint the defect). Using this methodology it is not possible to perform the opposite type of experiments (mouse mesenchyme and chi& epithelium) because these limbs will not lead to viable grrifts.

In the analysis of the Duc mice I hypothesized that Shh may indeed be the endo,nenous

AERMF that is localized to the posterior mesenchyme. it may be possible to test this hypothesis. since SliIi KO mice have been generated. Homozygous Shh mutant mice have no autopod at dl. but heterozygous mice appear to be normal (Chiang er al., 1996). If Slih compensates for rhc absence of Dclc in the posterior mesenchyme, however. it ma' be expected that loss of one copy of Slrh may lead to an exacerbation of the Dac phenotype.

For example. a shortened posterior digit or no digit at all. One rnay also see an increasc in the relative posterior limit of the limb defect in Dac heterozygotes in the absence of one copy of SM. since thrrr would be lowcr octivity and diffuse less far. By generating double mutant micr with Dac and the Shh KO it is possible to test this hypothesis.

Nevertheless, we do not fully understand the reguiation of Shh. and it is possible that a single copy of Shli is just as affective as two copies. Alternaiively, transcription of Shh might be upregulated from the remaining allele in the absence of the other.

Recently the defect in the mousr mutant qnd~

ligand Serrate. (Sidow et cil.. 1997) The phenotype in these rnice is a fusion of the middle three digits as a result of a hypcrplastic AER. The AER is much wider in these mice and on cross section appears to expand downwards into the mesenchyme. On first glance it appears that this defect may be opposite to the failure to maintain the AER in Dac mice. It would be prudent to investigate the cellular cause of the expanded AER in srn mice to determine if the hyperplasticity is a result of a decrease in ce11 death. an increase in ce11 proliferation. or both. Nevertheless, both an increase in ce11 proliferation and a decrease in ce11 death. are effectively the opposite cellular phenotype to Dac and thus represent a classic scenario to study the epistatic relationship between Dac and sm. By generating double

mutant mice and analyzing the phenotype the genetic relationship between sm and Dac can

be elucidated. In this experiment if DCK and sni function independently of one another the

resultant phenotype should be neither un-like nor Dac-like but an intermediate or something

different. However. if sin is epistatic to Dm. ihat is sm is a negative regulator of Dac, the

double mutant should have a Dac phenotype. On the other hand if Dac is a negative

regulator of sm we should observe a sm phenotype. Identification of an epistatic

relationship would be usrful to funher definc the rnolccular and biochemical phenotype in SHFM. References Afzelius. B. A. (1976). A human syndrome caused by immotile cilia. Science. 193: 3 17- 3 19. Afzelius, B. A. (1995). Situs invenus and ciliary abnormalities What is the connection? hi. J. Dev. Biol. 39: 839-844.

Akarsu. A.N.. Stoilov. 1.. Yilmaz, E., Sayil, B.S., and Sarfarazi. M. (1996). Genomic structure of HOXDZ3 gene: a nine poiyalanine duplication causes synpolydactyly in two unrelated families. Hum. Mol, Genet. 5: 945-952.

Akimenko. M.-A.. Ekker. M.. Wegner. J.. Lin. W., and Westerfield. M. (1994). combinatonal expression of three zebnfish genes related to distal-less: pan of a homeobox code for the head. J, Nercrosci. 14: 3475-3486.

Akimenko, M.-A., Johnson. S.L.. Westerfield, M.. and Ekker. M. ( 1995). Differential induction of four msx homeobox genes during fin development and regeneration in zebrafish. Deiulopnlent. 121: 347-357. Ahmad. M.. Abbas. H., Haque, S., and Flatz, G. (1987). X-chromosomally inherited split handlsplit foot anomaly in a Pakistani kindred. Him. Genet. 75: 169- 173.

Xhringer. J. ( 1995). Embryonic tissue diflerentiation in Cuenorhabditis eleguns requires difl. a gene homologous to mitochondrial solute chers. EMBO J. 14: 2307-13 16. Ang. S-L. and Rossant. J. (1993). Anterior mesendoderm induces mouse Engrailed genes in explant cuItures. Developmenr. 118: 139- 149.

Auçrbach. C. (195511956). A possible case of delayed mutation in man. Annuls Hum. Genet. 20: 266-269.

Avraharn. K.B.. Givol. D.. Avivi. A.. Yayon, A., Copeland. N.G.. and Jenkins. N.A. ( 1991). Mapping of murine fibroblast growth factor receptors refines rcgions of homology between mouse and human chromosomes. Genomics. 21: 656-658. hvraham, K.B.. Hasson. T.. Steel. K.P.. Kingsley. D.M.. Russell. L.B.. Mooseker. M.S..Copeland. N.G. and Jenkins. N.A. (1995). The mouse Snell's walr;er deafness oene encodes an unconventional rnyosin required for structural integrity of inner ear hüir cells. Natitre Genet. 11: 369-375. AmBlanc, P.. Ramirez-Weber. F.A.. Laget. M.P.. Schwartz. C.. and Komberg, T.B. ( 1997). Proteolysis that is inhibited by hedgehog targets Cubitus intemptus protein to the nucleus and converts it to a repressor. Cell. 89: 1043-1053.

Bamshad. M., O'Quinn. J.R.. and Carey, J.C. (1998). Wolf-Hirschhorn syndrome and a split-hand malformation. Am . J. Med. Genet. 75: 35 1-354.

Barskey. A.J. (1964). Cleft hand: Classification. incidence and treatment. J. Bone Joint Stq. 16A: 1707-1719.

Bedell. M.A.. Brannan. C.I.. Evans, E.P.. Copeland. N.G.. Jenkins. N.A. and Donovan, P.J. ( 1995). DNA rearnngements located over 100 kb 5' of the Steel (S1)-coding region in Steel-panda and Steel-contrasred mice deregdate SI expression and cause fernale sterility by disrupting ovarian follicle development. Genes Der: 9: 455-470. Bell, J.R.. Noveen, A.. Liu, Y.H.. Ma, L., Dobias. S., Kundu, R., Luo, W.. Xia, Y,, Lusis, A.J., Snead, M.L., and Maxson, R. (1993). Genomic structure, chromosomal location, and evolution of the mouse Hm8 gene. Genomics. 16: 123-131. Bellusci. S.. Grindley, J.. Emoto. H.. Itoh. N. Hogan, B.L.M. (1997). Fibroblast growth factor 10 (FGFIO) and branching morphogenesis in the embryonic lung. Development. 124: 4867-4878.

Biesecker, L.G.(1997). Strike three for GLI3- Nat. Genet. 17: 259-260.

Birch-Jensen, A. ( 1949). Congenital defortniries of the upper extremities. Ej nar Munksgaard, Copenhagen. Boles. R.G., Pober, B.R., Gibson. L.H., Willis, CR., McGrath, J., Roberts, D.J., and Yang-Feng, T.L. (1995). Deletion of chromosome 2q24-q3 1 causes characteristic digital anomalies: case report and review. Am J. Med. Gellet. 55: 155- 160. Boncinelli. E.. Somma. R., Acampora, D., Pannese, M., D'Esposito, M.. Faiella. A., and Simeone. A. ( 1988). Organization of human horneobox genes. Hitm. Reprod. 3: 880-886. Boutin. E.L. and Fallon. J.F. (1984). An analysis of the fate of the chick wing bud apical ectodermal ndge in culture. Dev. Biol. 104: 1 1 1- 1 16.

Braverman, N.. Kline. A.. and Pyeritz, R. (1993). Interstitial deletion of 6q associated with ectrodactyly . Ani J. Hum. Genet. 53(suppl.): A4 10. Brittnacher. J.G.. and Ganetsky, B. (1984). On the components of segregation distortion in Drosophila rnefrinogaster.III. Nature of enhancer of SD. Genetics. 107: 423-434.

Brown. N. A.. McCxthy. A., and Wolpcrt. L. ( 199 1 ). Developmrnt of handed body asymmetry in mammals. Ciba Found. Svmp. 162: 182-201. Brustovetsky. N.. and Klingenberg, M. (1996). Mitochondrial ADPIATP camer can be reversibly convrned into a large channel by C& Biochemistry 35: 8483-8488.

Buckler. A.J..Chang. D.D.. Graw. S.L.. Brook. J.D., Haber, D.A.. Sharp, P.A., and Housman. D.E. ( 199 1). Exon amplification: A strategy to isolate mammalian genes based on RNA splicing. froc. Nad Acad. Sci USA. 88: 4005-4009.

Bulfone A.. Kim. H.J.. Puelles, L.. Porteus. M.H.. Grippo. J.F.. and Rubenstein, J.L. i 1993). The rnouse DIX-2 (Tes-1)gene is expressed in spatially restricted domains of the forebrain. face and limbs in midgestation mouse embryos. Mech. Dev. 40: 12% MO.

Burke. AC.. and Alberch, P. (1985). The development and homology of the Chelonian carpus and tarsus. J. hforphol. 186: 1 19- 13 1. Büscher. D.. Bosse. B.. Heymer, J. and Rüther. U. (1997). Evidence for genetic control of Sonic iwdgrliog by G[i3in mouse Iimb development. Mech. Dev. 62: 175- 182.

Buss, P.W. ( 1994). Cleft hand/foot: clinical and developmrntal aspects. J. Med. Genet. 31: 726-730. Calzolari. E.. Manioguigi, O.. Garani, G.P.,Cocchi G.. Magnani, C., and Milan, M. (1990). Limb reduction defects in Emilia Romagna Italy: epiderniological and genetic study in 173 109 consecutive live births. J. Med. Genet. 29: 353-357. Cannizarro, L.A., Aronnson. M.M., and Thiesen, H.J. (1993). Human zinc finger gene ZNF23 (KOX16) maps to a zinc finger gene cluster on chromosome 16q22 and ZNF32 (KOX30) to chromosomç lOq23-q24. Hum. Genet. 91: 383-385.

Carlson. B.M. (1988). Parten's Foiuzdations of Enibryology, 5th Ed. McGraw-Hill. New York. pp.340-354.

Chai. C.K. ( 198 1). Dactylaplasia in mice. a two-locus mode1 for developmental anomalies. J. Hered. 72: 334-237. Chan, D.C.. Wynshaw-Boris, A., and Leder, P. (1995a). Formin isoforms are differentially expressed in the mouse embryo and are required for the normal expression of f5/-4 and shh in the iimb bud. Developmenr. 121: 3 15 1-3 162. Chan. D.C.. Laufer, E.. Tabin, C. and Leder, P. (1995b). Polydactylous limbs in Strong's Liuoid mice results from ectopic polarizing activity. Developmenr. 121: 197 1- 1978. Chaubr, S. (1959). On axiation and symmetry in transplanted wing of the chick. J. Exp. 2001. 140: 29-75.

Chen. H.. Lun. Y.. Ovchinnikov. D., Kokubo. H.. Oberg. K.C.. Pepicrlli. C.V.. Gan. L.. Lee. B.. and Johnson. R.L.(1998). Limb and kidney defects in Lmxlb mutant mice suggest an involvement of LMX 1 B in hurnan nail patella syndrome. NUL Genet. 19: 5 1-

Chess. A.. Simon. 1.. Cedar. H.. and Axel, R. (1994). Allelic inactivation regulates olfactory receptor gene expression. Cd.78: 823-834. Chiang. C.. Litingtung. Y.. Lee. E.. Young, K.E.. Gorden. J.L.. Westphal. H. and Beachy. P.A. (1996). Cyclopia and defective axial patterning in rnice lacking Sonic hedgelrog gene function. Nariire. 383: 407-4 14.

Chirgwin, J. M., Przybyla, A. E.. MacDonald. R. J., and Rutter. W.J. (1979). Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry. 18: 5294-5299. Cobben. J.M., Verheij, J.B.G.M.. Eisma, W.H., Robinson, P.H., Zwierstra, R.P.. Leegte. B. and Castedo. S. (1995). Bilateral split handkplit foot malformation and inv(7)(p22q21.3). J. bled. Genet. 32: 375-378. Cohen. S.M.. Bronner. G.. Kuttner. F., Jurgrns, G,and Jackle. H. (1989). Distal-Iess encodss a homeodomain protein required for limb developrnent in Drosophila. Nature. 338: 432-433.

Cohn. M.J., Izpisua-Bslmontr, J.C.. Abud, H.. Heath. J.K., and Tickle. C. (1995). Fibroblast growth factors induce additional limb development from the flank of chick ernbryos. Cell. 80: 739-746. Corthesy-Theulaz. 1.. Pauloin. A.. and Pfeffer. S. R. (1992). Cytoplasmic dynein participates in the centrosomal localization of the Golgi complex. J. Ce11 Biol. 118: 1333- 1345. Coutavas. E.. Ren, M., Oppenheim. J.D.. D'Eustachio. P.. and Rush. M.G.. (1993). Characterization of proteins that interact with the cell-cycle regulatory protein Ran/TC4. Nature. 366: 585-587.

Crackower. M.A.. Scherer, S.W.. Rommens. J.M.. Hui, C.-c.. Poorkaj, P.. Soder. S.. Cobben. J.M.. Hudpins. L.. Evans. J.P. and Tsui. L.-C. (1996). Characterization of the split handl split foot malformation locus SHFM Z at 7qll.3-qX. 1 and analysis of a candidate gene for its expression dunng limb development. Hia~i.Mol. Genet. 5: 57 1-579.

Crackower. M. A., Heng. H. H. Q.. Shi. X., Scherer. S. W.. and Tsui. L.-C. ( 1997). Fluorescent in situ mapping of the murine deleted in split han& split foot 1 gene (dssl) to chromosome 6. Mutnrn. Genorne. 8: 704. Crossley. P.H. and Martin. G.R. (1995). The mouse Fgf8 gene encodes a family of polypeptides and is expressed in regions that direct outgrowth and patteming in the developing rmbryo. Derutopment. 121: 439-45 1.

Crossley. P.H., Minowada. G.. MacArthur. C.A. and Martin. G.R. (1996). Roles for FGF8 in the induction. initiation. and maintenance of chick limb development. Cell. 84: 127- 136.

Crow. J.F. ( 199 1 ). Why is Mendelian scgregation so exact'? Biorssuys. 13: 305-3 12.

Czeizel. A.E.. Vitez. M.. Kodaj. T.. and Lenz. W. (1993). An epidemiological study of isolated split handlfoot in Hungary. J. Med. Gener. 30: 593-596.

David. T.J. ( 1972). Dominant rctrodactyl y and possible germinal rnosaicisrn. J. Med. Genet. 9: 3 16-320.

Davidson. D.R..Crawley. A.. Hill. R.E.. and Tickle. C. (1991). Position-dependent expression of two related homeobon genes in developing vertebrate limbs. Naiure. 352: 429-43 1. de Jong. G.. Kirby, P.A., and Muller, L.M. ( 1998). RSH (Smith-Lrmli-Opitz) syndrome: "Severe" phenotype with ectrodacty ly. Ani J. Med Gener. 75: 283-257. de Kok. Y.J.M.. Mrrkx. G.F.M., van der Maarel. S.M.. Huber. 1.. Malcolm. S., Ropers, H-H. and Cremers. F.P.M. (1995). A duplication/paracentric inversion associated with familial X-linked deafness (DFN3) suggesü the presence of a regulatory element more than 400 kb upsueam of the POU3F4 gene. Hum. Mol. Genet. 4: 2145-2 150.

Dcaly. C.N.. Roth. A.. Ferrari. D.. Brown, A.M.C. and Koshrr. RA. ( 1993). Wnt-Sa and Wut-7a are expressed in the developing chick limb bud in a manner suggesting roles in pattern formation dong the proximodistal and dorsovenual axes. Mech. Dev. 43: 175- 186.

Deans. R.. Schafcr, W.. Zimmermann. H.-P.. Mueller. L.. Gorlich. D., Prehn. S., Ponstin_ol. H.. and Bischoff. F.R. (1997). Ran-binding protein 5 (RanBPS) is related to the nuclcar transport factor importin-P but interacts differently with RanBP 1. Mol. Cell. Biol. 17: 5057-5096. Del Porto, G., D'Alessandro, E., De Matteis, C., Lo Re, M.L., Invaldi, M., and Di Fusco, C. (1983). Delezione interstiziale del braccio lungo del chromosome 7 e sue correlazone chiche. Pathologica. 75: 268-27 1. Dick, T.. Ray, K., Salz, H. K.. and Chia. W. (1996). Cytoplasmic dynein (ddlcl) mutations cause morphogenetic defects and apoptotic ce11 death in Drosophila melanogaster. Mol. Cell. Biol. 16: 1966- 1977.

Distel. R. J.. Kleene, K. C., and Hecht. N. B. (1984). Haploid expression of a mouse testis alpha-tubulin gene. Science. 224: 68-70.

Dollé. P.. Izpisua-Belmonte, J.C., Falkenstein, H.. Renucci, A. and Duboule. D. (1989). Coordinate expression of the murine Hox-5 complex horneobox-containing genes during limb pattern formation. Nature. 342: 767-772.

DollC, P.. Izpisua-Belmonte. J.-C.. Boncinelli, E. and Duboule. D. ( 199 1 ). The Hox-1.8 gene is localized at the 5' extremity of the Hox-4 complex and is expresscd in the most posterior parts of the body during development. Mech. Dev. 36: 3- 13. Dollé. P., Price. M.. and Duboule. D. (1992). Expression of the murine Dlx-1 homrobox gene during hcial. ocular and lirnb development. Diferentiation. 49: 93-99.

Dollé, P., Dirrich. A., LeMeur, M.. Schirnmang, T., Schuhbaur, B.. Chambon, P.. and Douboulr. D. (1993). Disruption of the Hoxd- 13 gene induces localized heterochrony leading to mice with neotenic limbs. Cell. 75: 43 1-441.

Dressler. G.R..Deutsch. U.. Choudhury, K.. Nomes. H.O.. and Gruss. P. (1995). Pax?. a new murinr paired box containing gene and its expression in the developing rxcretory sy stem. Drrelopntent. 109: 787-795.

Duhl. D.M.J.. Vrieling. H.. Miller. KA.. Wolff, G. and Barsh. G.S. (1994). Nromorphic agorrri mutations in obese yellow mice. Natwe Genet. 8: 59-65. Dunn. N.R.. Winnier. G.E.. Hargett, L.K., Schrick. J.J., Fogo. A.B. and Hogan. B.L. ( 1997). Haploinsufficient phenotypes in Bmp4 heterozygous nul1 mice and modification by mutations in Gli3 and Aix4 Dev. Biol. 188: 235-247.

Duprez. D.M.. Kostakopoulou. K., Francis-West, P.H.. Ticklr. C. and Brickell. P.M. (1996). Activation of FC@ and HoxD gene expression by BMP-2 expressing cells in the devrloping chick limb. Deveiopment. 122: 182 1- 1828.

Duyk. G.M.. Kim. S.. Myers. R.M.. and Cox, D.R. (1990). Exon-trapping: A genetic screen to identify candidate transcribed sequences in cloned rnammaiian genomic DNA. Proc. Natl. Acad. Sci. USA. 87: 8995-8999.

Echelard. Y.. Epstein. D.. St-Jacques, B., Shen, L., Mohler. J.. McMahon, J.A. and McMahon. A.P. (1993). Sonic hedgehog, a mernber of a family of putative signaling molecules. is implicated in the regulation of CNS polarity. Ceil. 75: 14 17- 1430. Echevrrri. C. J.. Paschal, B. M.. Vaughan. K. T., and Vallee. R. B. (1996). Molecular characterization of 50kd subunit of dynactin reveals function for the cornplex in chromosome alignmrnt and spindle orientation during mitosis. J. Ce11 Biol. 132: 6 17-634. Erickson, R. P. ( 1990). Post-meiotic -ne expression. Trends Genet. 6: 264-269. Eshel, D., Urrestarauza, L. A.. Vissers. S.. Jauniaux. J. C.. van Vliet-Reedijk, J. C.. Planta, R. J.. and Gibbons. 1. R. (1993). Cytoplasmic dynein is required for normal nuclear segregation in yeast. Proc. Natl. Acad. Sci USA. 90: 1 1172- 1 1 176.

Evans, J.P.. Vitez, M., and Czeizel, A. (1994). Congenital abnormalities associated with limb deficiency defects: a population study based on cases from the Hungarian Congenital Malformation Registry (1975-1984). Am. J. Med. Gener. 49: 52-66. Evans, J.P., Palmer. S.. Scherer. S.W.. Kukolich. M., Wijsman. E.M.. Tsui. L.-C.. and Stephens. K. (1995). Reply to Zlotogora. Am J. Hitm. Genet. 56: 342-343.

Faiyez-ul-Haque, M., Uhlhnas. S.. Knapp. M.. Schuler. H.. Friedl, W., Ahmad. M. and Propping, P. (1993). Mapping of the gene for X-chromoçomal split handl split foot anomaly to Xq26-q16.1. Him. Genet. 91: 17- 19.

FaIlon. J.F.. Lopez. A.. Ros, M.A., Savage. M.P.. Olwin, B.B. and Simandl. B.K. (1994). FGF-2: Apical ectoderrnal ridge growth signal for chick limb development. Science. 264: 104- 1 07.

Fallon, J.F.. Boutin. E.L.. and Carrington. J.L. (1986). Apical ecrodermnl ridge maintenance in oso and itt vitro. In Cellular Enducrinology: Hoririund Cotrtrol of Embryonic and Cellular Diflerentiation. Alan R. Liss, Inc. pp. 103- 1 13.

Fantes. J.. Rcdeker. B.. Breen. M.. Boyle. S.. Brown, J.. Fletcher. J.. Jones. S.. Bickmore. W.. Fukushima. Y.. Mannens. M.. Danes. S.. van Heyningen. V. and Hanson. 1. (1 995). Aniridia-associated cytogenetic rearrangements suggest that a position effect rnay cause the mutant phenotype. Hlori. Mol. Genet. 4: 415-422.

Ferrari. D.. Sumoy. L.. Gannon. J.. Sun. H.. Brown, A.M.C.. Upholt. W.B. and Kosher. R.A. (1995). The expression pattem of the Distal-less homeobox sene Dl-r-5 in the deveioping chick limb bud suggests its involvement in apical ectodermal ridge activity. pattern formation. and cartilage differentiation. Mech. Dev. 52: 257-264.

Ferrari. D.. Lichtlrr. AC.. Pan, 2.2.. Dealy. C.N., Upholt, W.B.. and Kosher. RA. (1998). Ectopic expression of Msx-2 in posterior limb bud mesoderm impairs limb morphogenesis while inducing BMP-4 expression. inhibiiing ce11 proliferation. and prornoting apoptosis. Del*. Biol. 197: 12-24.

Fontaine. G.. Farriaux, J.P.. Delattre. P.. Gidlecki. Z., Poupard, B., Durieur;. G.. and Piquet. J.J. (1971). Une observation familiale du syndrome ectrodûctylie et dysostose mandibuIo-faciale. J. Genet. Hitm. 22: 289-307.

Forsthoefel. P.F. ( 1963). Observations on the sequence of blastemal condensations in the lirnbs of the mouse embryo. iînar. Rec. 147: 129- 138. Francis, P.H.. Richardson. M.K.. Brickell. P.M. and Tickle. C. (1994). Bone morphogenetic proteins and a signaling pathway that controls patteming of the developing chick limb. Dei*elopniriil. 120: 209-2 18.

Freire-Maia. A. (1971). 4 recessive form of ectrodactyly and its implications in genetic counseling. J. Hered 62: 53. Frohrnan. M.A.. Dush. MX. and Manin. GR. (1988). Rapid production of full-length cDNAs frorn rare transcripts: amplification using a single gene-specific oligonucleotide primer. Proc. Nad. Acad. Sci. USA .85: 8998-9002. Froster. U.G.,and Baird. P. (1989). Limb reduction defects in over 1 million consecutive live births. Teratology. 39: 127- 135.

Ganan. Y.. Macias. D.. Duterque-Coquillard. M.. Ros. M.A.. and Hurle, J.M. (1996). Role of TGFbs and BMPs as signal controlling the position of digits and the areas of interdigital cell drath in the developing chick limb autopod. Developtnent. 122: 2349- 2357. Ganan. Y.. Macias. D.. Basco. R.D.. Merino. R.. and Hurle. J.M. (1998). Morphological diversity of the avian foot is related with the pattern of msx gene expression in the developing autopod. Dm Biol. 196: 33-4 1.

Ganetsky. B. (1977). On the components of segregation distortion in Drosophila rnelanogaster. Genetics. 86: 32 1-355. Gavrieli. Y.. Sherman. Y. and Ben-Sasson. S.A. (1992). Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J. Ce11 Biol. 119: 493-50 1.

Geduspan. J.S.. and Solursh. M. (1992). A growth-promoting influence from the rnesonepheros during limb outgrowth. Der: Biol. 151: 2 12-250.

Genuardi. M.. Pomponi. M. G.. Sammito. V.. Bellussi. A.. Zollino. M.. and Neri. G. ( 1993). Split hand/foot anomaly in a farnily segregating a balanced translocarion with breakpoint on 7q22.1. Ani J. bled. Genet. 47: 823-83 1.

Genuardi. M., Gurrieri, F.. Nanni. L.. Boncinelli. E., Evans, J.P. and Neri, G. (1995). Mutational analysis of distal-less _ornes in split handkplit foot anomaly (abstract). Am. J. N111?1.Genet.. 57:s. A 133. Grpner. J.. Li. M.. Ludmann. S.. Konas, C.. Boylan. K.. Iyadurai. S. J.. McGrail. M.. and Hays. T. S. (1996). Cytoplasrnic dynein function is essential in Drosophila melanogaster. Genetics. 112: 865-878.

Géraudie. J.. and Singer. M. ( 1990). The fish fin regenerate. In Key to Regenerarion (H.W. Sauer). Mono. Dev. Biol. Vol 23. pp. 62-72. Karger. New York.

Goodrich. L.V., Johnson. R.L.. Milenkovic,L., McMahon, LA.. and Scott, M.P. (1996). Conservation of the hed~ehoglpatchedsignaling pathway from nies to mice: induction of mouse patched by Hedgehog. Genes Dev. 10: 301-312. Gorîïnkiel, N.. Morata, G.. and Guerrero. 1. (1997). The homeobox gene Distul-less induces ventral appendage development in Drosophila. Genes Dev. 11: 2259-227 1.

Gorlich. D. ( 1997). Nuclear protein import. Curr. Opin. Ce11 Biol. 9: 4 L 2-4 19. Gratzner. HG. (1982). Monoclonal antibody to 5-bromo and 5-iododeoxyuridine. Science. 218: 474475. Gumeri. F., Genuardi. M.. Chiurazzi. P.. Gillessen-Kaesbach, G. and Neri. G. (1994). Exclusion of linkage between autosomal dominant split handsplit foot and markers from chromosome 7q: further evidence for genetic heterogeneity. Am J. Hum. Genet. 55: 853- 855. Gumeri, F., Prinos, P., Tackels. D.. Kilpatrick, M.W., Allanson, J.. Genuardi, M., Vuckov, A., Nanni, L., Sangiori. E.. Nunes, M.E., Garofalo, G., Neri, G.; Schwartz, C., and Tsipouras, P. (1996). A split hand-split foot (SHFM3) locus is located on chromosome lOq24-25. Am. J. Med. Genec. 62: 427-436. Hahn, H.. Wicking, C.. Zaphiropoulous. P.G.. Gailani. M.R.. Shanley. S.. Chidambaram, A.. Vorechovsky, 1.. Holmberg. E.. Unden. A.B.. Gillies. S.. Negus. K.. Smyth, I., Pressman, C.. Leffell. D.J.. Gerrard, B., Goldstein, A.M., Dean, M., Toftgard, R.. Chenevix-Trench, G.. Wainwright. B.. and Bale, A.E. (1996). Mutations of the human homolog of Drosophiln porclied in the nevoid basal cell carcinoma syndrome. Cell. 85 :84 1-85 1 Halal. F., Desgranges. M.F.. Leduc. B.. Theor'et, G.. and Bettez. P. (1980). Acro-renal- mandibular syndrome. Anr. J. Med. Gener. 5: 277-284.

Hamburger. V. (1938). Morphogenetic and axial self-differentiation of transplanted limb primordia of 2-day chick embryos. J. Erp. Zool. 77: 379-400.

Hamburger. V. and Hamilton. H.L. (195 1). A series of normal stages in the development of the chick embryo. J. Morpltul. 88: 49-92.

Hara. K.. Kirnura. J.. and Ide. H. (1998). Effects of FGFs on the morphogenic potency and AER-maintenance activity of cultured progress zone cells of chick limb bud. Int. J. Dei: Biol. 42: 591-599. Harada, A., Takei. Y., Kanai. Y., Tanaka. Y.. Nonaka. S., and Hirokawa, N. (1998). Golgi vesiculation and lysosome dispersion in cells lacking cytoplasmic dynein. J. Ce11 Biol. 141: 51-59.

Haramis. A.G.. Brown. J.h.1.. and Zeller. R. (1995). The limb defornzify mutation disrupts the SHWFGF-4 feedback loop and regulation of 5' HoxD genes during limb pattern formation. Developrnenr. 121: 42374245. Hamson. R.G. (1918). Experiments on the development of the forelimb of Arnbysroma. a self-differentiating equipotential system. J. E-rp. 2001.25: 4 13-46 1.

Hrirtenstein, K., Sinha. P.. Mishra. A., Schenkel. H. Torok. 1.. and Mechler, B.M. ( 1997). The congested-like tracheae gene of Drosophilu melanoguster encodes a member of the mitochondrial carrier family required for gûs-filling of the tracheal system and expansion of the wings after eclosion. Generics. 147: 1755- 1768.

Hani. D.L.. Hiraizurni. Y.. and Crow. J.F. ( 1967). Evidence of sperm dysfunction as the mechanism of segregation distonion in Drosophila melanoguster. Froc. Nad Acad. Sci. USA. 58: 2240-2245. Hanl. D.L. (1975). Genetic dissection of segregation distonion. II. Mechanism of suppression of distortion by certain inversions. Generics. 80: 539-547. Hébert, J., Basilico, C., Goldfarb, M., Haub. O. and Martin. G. R. (1990). Isolation of cDNAs encoding four mouse FGF family members and characterization of their expression patterns during embryogenesis. Dev. Biol. 138: 454-463. Henikoff, S. (1990). Position-effect variegation after 60 years. Trendr Cenef. 6: 422-426.

Hinchliffe. J.R., and Sansom, A. (1985). The distribution of the polarking zone (ZPA) in the leg bud of the chick ernbryo. J. Enibryol. Exp. Morphol. 86: 169- 175.

Hogan, B., Beddington, R., Constantini. F. and Lacy, E. ( 1994). Manipulating the mouse embryo. Cold Spring Harbour, Second Edition. Holzbaur. E. L. F., and Vallee. R. B. (1994). Dyneins: Molecular structure and cellular function. Annu. Rev. Ce11 Biol. 10: 339-372. Honig, L.S.. and Summerbell. D. (1985). Maps of strength of positional signaling activity in the developing chick wing. J. Embwl. Exp. Morphol. 87: 163- 174. Hoodless. P.A.. Haeny, T.. Abdollah. S., Stapleton. M., O'Connor. M.B., Attisano, L., and Wrana. J.L. (1996). MADRI. a MAD-related protein that functions in BMP2 signaling pathways. Cell. 85: 489-500. Hudgins. L., Massa, H., Disteche. C.. Scherer. S.W.. Tsui. L.-C.. Trask. B. and Evans. J.P. (1994). Identification of a microdeletion at 7q21.3 with fluorescence in situ hy bridization with split handfsplit foot (rctrodactyly)(abstract). Ani. J. Him. Genet. 3: A107.

Hughes. S. M.. Vaughan. K. T.. Herskovits. J. S.. and Vallee. R. B. (1995). Molecular analysis of a cytoplasmic dynein light intermediate chain reveals homology to a family of ATPases. J. Ce11 Sci. 10: 17-24.

Hui. C.-C. and Joyner. A.L. (1993). A rnouse mode1 of Greig cephalopolysyndactyly syndrome: the exma-toesJmutation contains an intrazenic deletion of the Gli3 gene. Nature Genet. 3: 241-246.

Hui. C.C.. Slusarski, D., Platt. KA.. Holmgren. R.. Joyner, A.L. ( 1994). Expression of threr mouse homologs of the Drosophila segment polarity gene cubitus interruptus. Gli. Gli-2. and Gli-3, in ectoderm- and mesoderm-derived tissues suggests multiple roles during postimplantation development. Dei*. Biol. 162: 402-1 13.

Huw, L.-Y.. Goldsborough, A.S., Willison. K., and Artzt. K. (1995). Tctex-2: a sperm tail protein mapping to the t-complex. Dev. Biol. 170: 183- 194. Ignatius, J.. Knuutila, S., Scherer, S.W., Trask, B. and Kere. J. (1996). Split handlsplit foot malformation, deafness and mental retardation with a complex cytogenetic rearrangement involving 7qZ 1.3. J. bled. Genet. 33: 507-5 1 0.

Ikura. M. (1996). Calcium binding and conformational response in EF-hand proteins. Trends Biochent. Sci. 21: 14- 17.

Ingham. P.W. ( 1995). Signaling by hedgehog famil' proteins in Drosophila and vertebrate dsvelopment. Curr. Opin Genet. Dei.. 5: 492-49s. Innis, J.W., and Mortlock, D.P. (1998). Limb development: molecular dysniorphology is at hand! Clin. Genet. 53: 337-348. Irvine, K.D., Vogt, T.F. (1997). Dorsal-ventral signaling in limb development. Curr. Opin. Cell. Biol. 9: 867-876. Izpisk-Belmonte, J.-C., Tickle, C.. Dolle. P., Wolpen. L.. and Duboule. D. (199 la). Expression of the horneobox Hox-4 genes and the specification of position in chick wing development. Nature. 350: 585-589.

Izpisiia-Belmonte. J.-C., Falkenstein, H.. Dollé, P., Renucci. A. and Duboule, D. (1991b). Murine genes related to the Drosophila AdbB horneotic gene are sequentially expressed during development of the postenor part of the body. EMBO J. 10: 2279-2289.

Jablonka. E.. Goitein, R., Marcus. M., and Cedar, H. (1985). DNA hypornethylation causes an increase in DNase-I sensitivity and an advance in the rime of replication of the entire inactive X chromosome. Chrornosonia. 93: 152- 156.

Jarvik, G.P.. Patton, M.A., Homfray. T. and Evans. J.P. (1994). Non-Mendelian transmission in a human developmental disorder: split handsplit foot. Am. J. Hum. Genet.

Jrgalian, B.G. and De Robertis, E.M. (1992). Homeotic transformations in the mouse induced by overexpression of human Hox3.3 transgene. Cell. 71: 9 10-910. Johnson. D. R.. and Hunt. D. M. (1971). Hop-sterilc. a mutant gent affecting sperm tail development in the mouse. J. Embryi. Exp. Morph. 25: 223-236.

Johnson. K.R., Lane. P.W., Ward-Bailey, P. and Davisson, M.T. (1995). Mapping the mousr Dactylaplasia mutation, Dac, and a gene that controls its expression. rnDac. Genomics. 39 : 457-464.

Johnson, R.L., Rothrnan, A.L., Xie. J., Goodrich. L.V., Bare, LW., Bonifas, J.M., Quinn, A.G.. Myers. R.M.. Cox. D.R.Jr.. and Scott. M.P. ( 1996). Human hornolog of pa~hed.a candidate gene for the basal ce11 nevus syndrome. Scierice. 272: 1668- 167 1. Johnson, R.L., and Tabin, C.J. (1997). Molecular models t'or vertebrate limb development. Cell. 90: 979-990.

Jones. CM, Lyone. K.M. and Hogan. B.L.M. (1991). Involvement of Bone Marpliogeneiic Prorein-4 (BMP-4)and Vgr-1 in morphogenesis and neurogenesis in the rnouse. Dei*elopnient. 11 1: 53 1-542.

Justice, M.J. and Stephenson. D.A. ( 1994). Mouse Chromosome 13. Mamm. Genome. 5: S 196-S206.

Kantrow. S.P.. and Piantadosi. C.A. (1997). Release of cytochrome c from liver mitochondria during permeability transition. Bioclzenz. Biopliy. Res. Conmiun. 232: 669- 671.

Kaprio. E.A. ( 198 1 ). Trmsfilter evidence for a zone of polnrizing activity participating in limb morphogenesis. J. Embr-yol. Eip. Morphol. 65: 185- 197. Keller. S.A.. Jones. J.M.. Boyle, A.. Barrow, L.L.. Killen. PD.. Green. D.G.. Kapousta. N.V., Hitchcock, P.F.. Swank, R.T., and Meisler, M.H. (1994). Kidney and retinal defects (Krd), a transgene-induced mutation with a deletion of mouse chromosome 19 chat includes the Pax-2 locus. Genomics. 23: 309-320. Kieny. M. (1960). Role inducteur du mesoderme dans la differenciation du bourgeon de membre chez l'embryon de poulet. J. Embryoi. Lrp. Morph. 8: 457-467. King. S. M.. and Patel-King. R. S. (1995). The M(r)= 8.000 and 11.000 outer arm dynein light chains from Chlamydomonas flagella have cytoplasmic homologues. J. Bioi.Chem. 270: 1 1445- 1 1452. King, S. M.. Dillman, J. F.. Benashski. S. E.. Lye. R. J.. Patel-King. R. S.. and Pfister. K. K. (1996). The mouse t-complex-encoded protein Tcrex-1 is a light chah of brain cytoplasrnic dynein. J. BBiol. Chern. 271: 3228 1-32287.

Kitsberg. D.. Selig. S.. Brandeis. M.. Simon. 1.. Keshet. 1.. Driscoll. D.J.. Nicholls. R.D.. and Cedar. H. ( 1993). Allele-specific replication timing of imprinted gene regions. Natitre. 364: 459-463.

Kozak. M. (1987). An analysis of 5'-nonencoding sequences from 699 vrrtebrate messenger RNAs. Niicl. Acids Res. 15: 8 125-8 148. Kroemer. G.. Dallaporta. B.. and Resche-Rigon, M. (1998). The mitochondrial death/life replator in apoptosis and necrosis. Annu. Rev. Physiol. 60: 619-612. Kuan. J.. and Saier. M.H.Jr. (1993). The mitochondrial carrier family of transport proteins: structural. functional. and evolutionary relationships. Crit. Rev. Biochetri. Mol. Biol. 28: 209-233.

Kuhlman. J.. and Niswander. L. ( 1997). Limb deformity proteins: role in mesodermal induction of the apical ectodermal ridge. Developtrient. 124: 133- 139.

Lader. E.. Ha. HA..O'Neill. M., Artzt, K.. and Bennett. D. ( 1989). tete-r-1: a candidate gene family for a mouse t complex sterility locus. Ceii. 58: 969-979.

Lange. M. ( 1937). Grundsiitzlicher über die beuneilung der rntstehung und brwertung atypischer hand- und fussmissbildungen. Z. Orthop. 66: 80. Laufer. E.. Nelson. C.. Johnson. R.L.. Morgan. B.A. and Tabin. C. (1994). Sonic Iiedgehog and Fgf-4 act through a signaling cascade and feedback Ioop to integrate growth and patterning of the developing limb bud. Ceil. 79: 993-1003. Laufer. E.. Dahn, R.. Orozco, O.E., Yeo, C.-Y.. Pisenti. J.. Henrique, D., Abbott. U.K.. Fallon. J.F.. and Tabin. C. (1997). Expression of Radical fringe in limb-bud rctodcrm regulates apical ectodermal ridge formation. Nature. 386: 366-373. Lsvin. M.. and Nascone, N. (1997). Two moleculrtr models of initial left-right asymmetry _orneration. Med. Hypotheses. 49: 429-435. Lrwandoski. M.. Meyers. E.N.. and Martin. GR. (1997). Analvsis of Fcf8 gene function in vensbnte development. Cold. Spring Harb. Svmp. Quant. ~iol.62: fi9-168. Lewis. T.H.. and Embleton. D. (1908). Split-hand and split-foot deformities. thsir types, origin and transmission. Biometrika. 6: 26. Li, Y. Y.. Yeh, E., Hays, T.. and Bloom, K. (1993). Disruption of mitotic spindle orientation in a yeast dynein mutant. Proc. Nad Acad. Sci. USA. 90: 10096-10100.

Lin. S. X.. and Collins. C. A. (1992). Immunolocalization of cytoplasmic dynein to lysosomes in cultured cells. J. Ce11 Sci. 101: 125-137.

Lipton. S.A., and Nicotera. P. ( 1998). Calcium. free radicals and excitotoxins in neuronal apoptosis. Ce11 Calciir~n.23: 165- 17 1.

Logan. C.. Hombmch. A.. Campbell, 1. and Lumsden, A. (1997). The role of Engrailed in establishing the dorsoventral axis of the chick limb. Developmenr. 124: 23 17-2324.

Loornis. C.A.. Hams. E.. Michaud. J.. Wurst. W., Hanks. M.. and Joyner. A.L. (1996). The mouse Engrailed-1 gene and ventral limb patteming. Nariire. 382: 360-363. Lu, M.. Zhang. N.. and Ho. A.D. (1992). Genomic organization of the putative hurnan homeobox proto-oncogene HOX- 1 1 (TCL-3) and its endogenous expression in T cells. Oncogene. 7: 1325- 1330. Lufkin. T.. Mark M.. Hart. C.P.. Dolle, P.. LeMeur, M. Chambon. P. (1992). Horneotic transformation of the occipital bones of the skull by ectopic expression of a homeobox

Eoene. Narrtre. 359: 835-84 1. Lyon. M.F. (1983). Transmission ratio distonion in the mouse r-haplotypes is due to multiple distoner genes acting on a responder locus. Cd. 37: 67 1-628.

Lyons. KM.. Pçlton. R.W. and Hogan. B.L.M. (1989). Patterns of expression of munne Vgr- 1 and BMP-7a RNA suggest that transforming growth factor-beta-like genes coordinately regulate aspects of embryonic development. Genes Dei*.3: 1657- 1668.

Lyons. KM.. Hogan. B.L.M.. and Robertson, E.J. (1995). Colocalization of Bmp7 and BmpZ RNAs suggest that these factors cooperatively mediate tissue interactions during murine development. Mrch. Der. 50: 7 1-83.

MacCabe. J.A.. Errick. J.. and Saunders. J.W.Jr. (1974). Ectodermal control of the dorsoventral mis in the Itg bud of the chick ernbryo. Der.. Biol. 39: 69-83. Macias. D.. Ganan. Y .. Sampath. T.K., Piedra. M.E.. Ros. MA.. and Hurle, J.M. (1997). Role of BMP-2 and OP- 1 (BMP-7) in programrned ce11 death and skeletogenesis during chick limb development. Deveiopment. 124: 1109-1 1 17.

blaisels. D.O. ( 1970). Lobster-claw deformities of the hands and feet. Br. J. Plasr. Srirg. 33: 369.

Manzzi. G.. Wang. Y.. and Sassoon. D. ( 1997). Msx? is a transcriptional regulator in the BMPCmediatrd programrned ce11 death pathway. Dev. Biol. 186: 127- 138.

Marigo. V.. Davey. RA.. Zuo. Y.. Cunningham, J.M.. and Tabin. C.J. (1996a). B iochemical evidence that patched is the Hedghog receptor. Nntctre. 384: 176- 179. Marigo. V.. Scott. M.P.. Johnson. R.L.. Goodrich. L.V.. and Tabin, C.J. (1996b). Consenation in hedgehog signaling: induction of a chicken patched homolog by Sonic hedgehog in the developing limb. Development 122: 1225-1233. Mannoni, J.-C., Boyd. E.. Sherman. S. and Schwartz, C. (1994). Familial split handsplit foot long bone deficiency does not segregate with markers linked to the SHFDl locus in 7q2 1.3-q22.1. Hum. Mol. Genet. 3: 1355- 1357. Marinoni, J.C., Stevenson. R.E.. Evans, J.P.. Geshuri, D., Phelan, M.C. and Schwartz. C.E. (1994). Split foot and developrnental retardation associated with a deletion of three microsatellite rnarkers in the 7q2 1.7-q22.1. Clin. Genet. 47: 90-95. Masuya. H.. Sagai. T.. Moriwaki. K. and Shiroishi. T. (1997). Multigenic control of the localization of the zone of polarizing activity in limb rnorphogenesis in the mouse. Dev. Bid. 182: 42-5 1 McGrail. M.. and Hays. T. S. (1997). The microtubule rnotor cytoplasmic dynein is required for spindle orientation during sermline ce11 divisions and oocyte differentiation in Drosophila. Developmetit. 121: 2109-24 19.

Mc Kusick. V.A. ( 1992). In Mendeiian Intieritance in Man. Catafogs of Ailtosotnal Domirianf. Aictosoninl Rrcessiw und X-linked Phenotypes. Johns Hopkins University Press, Baltimore, 10th edn.

McMullen. G.L.. and Pearson. K. ( 19 13). On the inheritence of the deformity known as spli t foot or lobster claw. L3iometriX.n. 9: 38 1. Meller. J. (1893). Ein l'dl von anpeborencr spaltbildung der hande und füsse. Bed. Klh. tvsclzr. 30: 232-233.

Merino. R., Ganan. Y.. iMacias. D.. Economides. AN.. Sarnpath. KT., and Hurle. J.M. (199s). Morphogenesis of digits in the avian limb is controlled by FGFs. TGFbetas. and noggin through BMP signaling. Dev. Biol. 200: 35-45. Michaud, J.L.. Lapointe. F.. and Le Douarin. N.M. (1997). The dorsoventral polarity of the presurnptive limb is determined by signals produced by the somites and by the lateral sornstopleure. Deveiopnient. 124: 1133- 1452. Milaire. J. and Rooze, M. (1983). Hereditary and induced modifications of the normal necrotic patterns in the developing limb buds of the rat and mouse: Facts and hypothesis. Arch, Bio. 94: 459-490.

Monaco. A.P., Neve, R.L., Colletti-Feener, C., Bertelson. C.L., Kunit. D.M.. and Kunkel. L.M. (1986). Isolation of candidate cDNA's for portions of the Duchenne muscular dystrophy gene. Nature. 323: 646-650.

Morgan. B.A., Izpisua-Belmonte. J.-C.. Duboule, D.. and Tabin, C. ( 1992). Targeted rniseapression of Hox-J.6 in the avian limb bud causes apparent homeotic transformation. Nature. 358: 236-239.

Morgan. B.A.. and Tabin. C.J. (1993). The role of homeobox genes in limb developrnent. Citrr. Opin. Gener. Dev. 3: 668-674. Moroianu. J., Blobel. G.9 and Radu, A. (1996). Nuclear protein import: Ran-GTP dissociates the karyopherin cr/P heterodimer by displacing a from an overlapping binding site on p. Proc. Nail. Acad. Sci. USA. 93: 7059-7062. Mortlock. D.P., Post. L.C.. and Innis, J.W. (1996). The molecular basis of hypodactyly (Hd): a deletion in Ho.ral3 leads to arrest of digital arch formation. Nat. Genet. 13: 284- 289.

Mortlock. D.P.. and Innis, LW. (1 997). Mutation of HOXA 13 in hand-foot-genital syndrome. Nat. Genet. 15: 179- 180.

Mount. S.M. (1982). A catalogue of splice junction sequences. Nid. Acids. Res. 10: 459-472.

Muneoka, K.. Wanek. N.. and Bryant. S.V. (1989). Mammalian limb bud development: In situ fate maps of early hindlimb buds. J. Ekp. Zool. 249: 50-51.

Muragaki. Y.. h.1undlos. S.. Upton. J.. and Olsen. B.R. (1996). Piltered growth and branching patterns in synpolydactyly caused by mutations in HOXDI3. Science. 272: 548-551.

Nachury. M.V.. Rydrr. U.W., Lamond, AI.. and Weis. K. ( 1998). Cloning and characterization of hSRPlg. a tissue-specific nuclear transport factor. Proc. Narl. Acad. Sci. USA. 95: 582-587.

Nakamura. S., Stock. D.W.. Wydner, K.L.. Bollekens. J.A.. Takrshita. K.. Nagai. B.M., Chiba, S., Kitamura, T., Freeland. T.M.. Zhao. Z.. Minowada, J., Lawrence. J.B.. Weiss. KM., and Ruddle, F.H. (1996). Genomic analysis of a new marnrnalian distal-less gene: DLr7. Genoinics. 38: 3 14-324. Naritorni. K.. Izumikawa. Y.. Tohura. T.. and Hirayana. K. (1993). Invcrted insertion of chromosome 7q and rctrodactyly. Am. J. Med. Genet. 46: 192493.

Nelson. C.E.. Morgan. B.A.. Burke. A.C.. Laufer, E.. DiMambro. E.. Murtaugh. L.C., Gonzalrs. E.. Tessarollo. L.. Parada. L.F.. and Tabin. C. (1996). Analysis of H0.r gene expression in the chick limb bud. Development. 122: lU9- 1466.

Neri. G.. Genuardi. M.. Pomponi. M.G.. Sammito, V.. and Zollino, M. (1992). Funher rvidence for the presence of a split handlfoot deformity gene on chromosome 7q2 1 from a family segregating a 27 translocation. Am. J. Hiini. Gener. Sl(supp1.): A327.

Nicotcra. P.. and Orrenius. S. (1998). The role of calcium in apoptosis. CdCalcico~i. 23: 173- 180.

Nigg. E.A. ( 1997). Nucleocytoplasmic transport: signals mechanisms and regulation. Ncmtre. 386: 779-787.

Niswander, L.. Ticklr. C., Vogel, A., Booth. 1. and Martin, G.R. (1993). FGF-4 replaces the apical ectodermal ridge and directs outgrowth and patterning of the limb. Cell. 75: 579- 587.

Niswander, L., Jeffrey. S.. Manin. G.R. and Tickle, C. (1994). Positive feedback loop coordinates growth and pattemine in the venebnte Iirnb. Nurim. 371: 609-6 13. Nohno, T., Noji, S., Koyarna, E., Ohiyarna, K., Myokai, F., Kuroiwa, A., Saito, T., Taniguchi, S. (1991). Involvement of the Chox-4 chicken homeobox genes in determination of anteroposterior polarity dunng Iimb development. Cell. 64: 1197- 1205. Nomes, H.O., Dressler. G.R., Knapik, E.W.,Deutsch, U., and Gniss, P. (1990). Spatially and temporally restricted expression of Pax-2 during murine neurogenesis. Deveiopment. 109: 797-809. Nunes, M.E., Schutt. G.. Kukolich. M.. Byers. P. and Evans. J.P. (1995). A second autosornal split handl split foot locus maps to chromosome IOq25. Hum. Mol. Genet. 4: 2 165-2 170. Ohno, M.. Fomerod. M.. and Mattaj. LW. (1998). Nucleocytoplasmic transport: the 1ast 200 nanometers. Cell. 92: 327-336.

Ohuchi. H.. Nakagawa. T.. Yamamoto, A., Araga. A.. Ohata, T.. Ishimaru. Y.. Yoshioka, H.. Kuwana, T., Nohno. T.. Yamasaki. M., Itoh. N. and Noji. S. (1997). The mesenchymal factor. FGF10. initiates and maintains the outgrowth of the chick limb bud through interaction with FGFS. an apical ectodermal factor. Developmenr. 121: 2235- 2244. O'Quinn. J.R.. Hennrkam. R.C.M.. Jorde. L.B.. and Bamshad. M. (1998). Syndromic ectrodacty ly wi th severe limb. ectodermal. urogenital, and palatal defects maps to chromosome 19. Am J. Hrtrn. Genet. 62: I3O- 135, Opitz. J.M.. Frias. J.L.. and Cohen. M.M.Jr. (1980). The ECP syndrome. another autosomal dominant cause of monodactylous ectrodactyly. Ew. J. Pediritr. 133: 11 7-220. Ortega. S., Ittmann. M.. Tsang, S.H.. Ehrlich, M., Basilico. C. (1998). Neuronal defrcts and delayed wound hcaling in rnice Iacking fibroblast growth factor 2. Proc. Nad. Acad. Sci. USA. 95: 5672-5677. Ozcelik. T.. Porteus. M.H.. Rubenstein, J.L.. and Francke, U. (1992). DLX2 (TESI).a homeobox sene of the Distal-less family, assigned to conserved regions on human and mouse chromosomes 2. Genonrics. 13: 1 157- 1 16 1. Palmer, S.E., Scherer. S.W.. Kukolich. M.. Wijsman. E.M., Tsui, LX., Stephens, K. and Evans, J.P. (1994. Evidence for locus heterogeneity in human autosomal dominant split handlsplit foot malformation. Am. J. Hlim. Genet. 55: 21-26. Paschal, B. M., and Vallee. R. B. (1987). Retrograde transport by the microtubule- associated protein MAPI C. Nariire. 330: 18 1- 183.

Paschal. B. M.. Mikami. A.. Pfister, K. K.. and Vallee, R. B. (1992). Homology of the 74-kD cytoplasmic dynein subunit with a flagellar dynein polypeptide su,,**ests an intracellular tarseting function. J. Cell Biol. 118: 1 133-1 143. Paré. A. (1573). Les oercwes de M. Ambroise Pare. Gabries Boun. Paris. pp. 5-8. (Quoted by Barskey. 1961. and Stein and Bettman. 1910)

Parr. B.A.. Shea. M.J.. Vassileva. G. and McMahon. A.P. (1993). Mouse WNT genes exhibit discrete domains of expression in early embryonic CNS and Iimb buds. Developmenr. 119: 717-361. Parr, B.A. and McMahon. A.P. (1995). Donalizing signal Wnt-7a required for normal polarity of D-V and A-P axes of mouse limb. Nature. 374: 350-353. Pananen. J.. Schwartz. L., and Rossant, J. (1998). Opposite phenotypes of hypomorphic and Y766 phosphorylation sitemutations reveal a function for Fgfrl in anteropostenor patterning of mouse embryos. Genes Dev. 12: 2332-2344.

Patel-King. R. S.. Benashski, S. E.. Harrison. A.. and King. S. M.. (1997). A Chlamydomonas homologue of the putative murine t complex distoner Tctex-2 is an outer amdynein light chain. J. Cell Biol. 137: 108 1 - 1090. Pearson. K. (1908). On the inheritence of the deformity known as split foot or lobster claw. Biometrika. 6: 69-75.

Pearson. K. (193 1). On the existence of the digital deformity so called "lobster claw" in the apes. Ann. Eupen. 1: 339.

Pemberton. L.F., RosenbIurn, J.S., and Blobel. G. (1997). A distinct and paraIIel pathway for the nuclear import of an rnRNA-binding protein. J. Cell Biol. 139: 1645- 1653.

Pfarr, C. M.. Coue. M., Grisson, P. M.. Hays. T. S., Porter. M. E.. and Mclntosh. J. R. ( 1990). Cytoplasmic dynein is localized to kinetochores during mitosis. Nature. 345: 263- 265.

Pfeiffer. RA. ( 1984). Interstitial deletion of a chromosome 7(q 1 1 .?q2?. 1) in a child with split handlspiit foot malformation. rltin. Gciler. 27: 45-18. Pilarski, R.T.. Pauli. R.M.. Bresnick. G.H., and Lebowitz. R.M. (1985). Karsch- Neugebauer syndrome: split footlsplit hand and congenital nystagmas. Clin. Genet. 27: 97-101.

Pilder. S.H.. Olds-Clarke. P.. Phillips. DM.. and Silver. L.M. ( 1993). Hybrid Sreriliry- 6: a mouse I complex locus controllinp sperm flagel1a.r assembly and movement. Der: Biol. 159: 63 1-64?. Poirier, C.,and Guenet. J.-L. (1994). Mouse chromosome 19. Mamm. Genonze. 5: S266- S275.

Porter J.A.. Young. K.E.. and Beachy. P.A. (1996). Cholesterol modification of hedge hog signaling proteins in animal development. Science. 274: 255-259.

Propping, P, and Zerres. K. ( 1993). ADULT syndrome: an autosornal dominant disorder with pigment anomalies. ectrodacty l y, nail dy splasia. and hypodontia. Am. J. Med. Gener. 45: 634-648.

Püschel. A.W., Gruss. P.. and Westrrfield. M. (1992). Sequence and expression pattern of par-6 are highly conserved between zebnfish and rnice. Development. 114: 643-651.

Qiu. M., Bulfone, A., Maninez, S.. Meneses. J.J.. Shimamura, K., Pedersen. R.A.. and Rubenstein. J.L. (1995). Null mutation of D1.r-2 results in abnormal morphogenesis of proximal first and second branchial arch derivatives and abnormal differentiation in the forebrain. Genes Dev. 9: 2523-2535. Qiu, M., Bulfone. A., Ghattas, I., Meneses, J.J.. Christensen, L.. Sharpe, P.T., Presley, R., Pedenen, R.A., and Rubenstein, J.L. (1997). Role of the Dix homeobox genes in proximodistal patteming of the branchial arches: mutations of DLr-1, DLr-2, and DLr-1 and -2 alter morphogenesis of the proximal skeletal and soft tissue stnictures derived from the first and second arches. Dev. Biol. 185: 165- l84. Quinn. L.M.. Latham. S.E.. and Kalionis. B. (1998). A distal-less homeobox gene, DM4, is a candidate for regulating epithelial-mesenchyrnal ce11 interactions in the human placenta. Placenta. 19: 87-93.

Raas-Rothschild. A.. Aviram. A.. Ben-Ami. T., Berger. 1.. Katznelson. M., and Goodman. R.M. (1989). Newly recognized ectrodactyly/deafness syndrome. J. Craniofac. Genet. Dev. Biol. 9:12 1 - 127. Raju. K.. Tang. S.. Dube, I.D.. Karnel-Reid. S.. Bryce. D.M., and Breitman. M.L. (1993). Characterization and developmental expression of ïlx- I. the murine homolog of HOX 1 1. Mech. Dev. 44: 5 1-64.

Ramer. J.C., Ladda. R.L.. Frankel, C.A., and Beckford. A. (1989)- A review of phenotype-karyotype correlations in individuals with interstitial deletions of the long arm of chromosome 2. Am. J. Med. Genet. 32: 359-363. Ramer. J.C.. Mowrey. P.N., Robins. D.B.. Ligato. S.. Towfighi, J.. and Ladda, R.L. (1990). Five children with del (2)(q3lq33) and one individual with dup (2)(q31q33) from a single family: review of brain. cardiac. and limb malformations. Am. J. Med. Genet. 37: 393-400.

Ramirez-Solis. R.. Liu. P.. and Bradley. A. ( 1995). Chromosome engineering in mice. Nature. 378: 720-724.

Reuss. C.. and Saunders. J.W. ( 1965). Inductive and axial properties of prospective limb mesoderm in the early chick embryo. Am. Zool. (abstr.) 5: 214. Riddlr. R.D.. Johnson. R.L.. Laufcr. E. and Tabin. C. (1993). Sonic hedgehog mediates the polarking activity of the ZPA. Cell. 75: 140 1- 1-116. Riddle, R.D., Ensini. M., Nelson, C.. Tsuchida. T., Jessel. T.M. and Tabin, C. (1995). Induction of the LIM homeobox gene h~rlby WNT7a establishes dorsoventral pattern in the vcrtebrate limb. Ceil. 83: 63 1-6.10.

Rijli. F.M.. and Chambon. P. (1997). Genetic interactions of Hox genes in limb development: learning from îornpound mutants. Cicrr. Opin. Genet. Dev. 77: 48 1-487. Rivera. H.. Sanchez-Corona. J., Burgos-Feueles. U.R.. and Melendez-Ruiz, M.J. ( 199 1). Deletion of 7qZ2 and rctrodactyly. Gener. Colinsel. 2: 27-3 1.

Roberts. D.J. and Tabin. C. ( 1994). The genetics of human limb development. Am. J. Hum. Genet, 55: 1-6. Roberts. S.H.. Hughes. H.E.. Davies. S.J.. and Meredith. A.L. (1991). Bilateral split hand and split foot malformation in a boy with a de >ioio interstitial deletion of 7q21.3. J. Med. Genet. 28: 47938 1. Robinson. G.W., Wray, S., and Mahon, K.A. (1991). Spatially restricted expression of a member of a new family of murine Distal-less homeobox genes in the developing forebrain. New Biol. 3: 1183- 1 194. Robinson, G.W., and Mahon, K.A. (1994). Differential and overlapping expression domains of DLx-2 and DIX-3 suggest distinct roles for Distal-less horneobox genes in craniofacial development. Mech. Dev. 48: 199-2 15.

Rocchi, M., Covone, A., Romeo, G., Faraonio, R.. and Colantouni, V. (1989). regional rnapping of RBP4 to 10q23-q24 and RBPl to 3q21-q22 in man. Somat. Ceil. Mol. Genet. 15: 185-190. Rodini. E.S.O.and Richien-Costa, A. (1990). EEC syndrome: report on 20 new patients, clinical and genetic considerations. Am. J. Med. Genet. 37: 42-53 . Rodriguez-Esteban, C., Schwabe. J.W.R.. de la Pena. J., Foys, B.. Eshelman. B.. and Izpisua-Belmonte, J.C. (1997). Radical fringe positions the apical ectodermal ridge at the dorsoventral boundary of the vertebrate limb. Nanrre. 386: 360-365. Rommens, J.M., Iannuzzi. MC, Karem. B., Drumm, M.L.. Melmer, G., Dean. M., Rozmahel. R., Cole, J.L., Kennedy. D., Hidaka. N., Zsiga, M., Buchwald, M., Riordan. J.R.. Tsui. L.C., and Collins, F.S. (1989). Identification of the cystic fibrosis gene: Chromosome walking and jumping. Science. 245: 1059- 1065.

Rommens, LM., Mar, L., McArthur, J.. Tsui. L.-C. and Scherer, S.W. (1994). In Hochgeschwender, U.. and Gardiner. K. (eds.). The identification of transcribed seqiiences. Plenum Press, pp. 65-79.

Rossi. E.. Faiella, A., Zeviani. M.. Labeit, S.. Floridia, G., Brunelli. S., Cammarata. M., Boncinelli, E., and Zuffardi, 0. (1994). Order of six loci at 2q24q3 1 and orientation of the HOXD locus. Genomics. 24: 34-40. Rowe. D.A.. Cairns, J.M. and Fallon, J.F. (1982). Spatial and temporal patterns of ce11 denth in limb bud mesodem after apical ectodermal ridge removal. Dei.. Biol. 93: 83-9 1.

Rowles. J. C.. Scherer. S. W.. Xi. T.. Majer. M.. Nickle, D. C.. Rommens, J. M.. Popov. K. M.. Harris. R. A.. Riebow. N. L., Xia. J.. Tsui. L.-C.. Bogardus. C., and Prochazka. M. (1996). Cloning and characterizaiion of PDK4 on 7qZ 1.3 encoding a fourth human pyruvate dehydrogenase kinase isoenzyme in hurnan. J. Biol. C'hem. 271: 22376- 22382, Royer-Pokora. B.. Kunkel. L.M.. Monaco, .\.P.. Goff. S.C.. Newburger, P.E.. Baehner. R.L.. Cole, F.S.. Curnutte, J.T.. and Orkin. S.K. (1986). Cloning of the gene for an inherited human disorder-chronic granulomatous disease on the basis of its chromosomal location. Nature. 322: 32-38.

Saint-Hilaire. I.G. ( 1832). Histoire générale et particulierè des anomalies de l'organisation chez l'homme et les animaux. Vol. 1, Paris. Sambrook, J.. Fritsch. E. F. and Maniatis. T. (1989). "Molecular Cloning: A Labontory Manual, 2nd edn.." Cold Spring Harbor Laboratory Press. New York.

Sanyanusin. P.. Schimmenti, L.A.. McNoe. LA. Ward, TA., Pierpont. M.M., Sullivan, M.J.. Dobynd, W.B., and Eccles. M.R. ( 1995). Mutation of the PAX2 gene in a family with optic nerve colobomas. renal anomalies and vesicoureteral reflux. Nat. Genet. 9: 358-364. Saunders. J.W.Jr. (1948). The proximo-distal sequence of origin of the parts of the chick wing and the role of the ectoderm. J. Erp. 2001. 108: 363-404. Saunders. J.W., Jr. (1949). An analysis of the role of the apical ectodermal ridge in the development of the limb bud in the chick. Anat. Rec. 103: 567-568. Saunders. J.W.. Gasseling, MT., and Saunders. L.C. (1962). Cellular death in morphogenesis of the avian wing. Dev. Biol. 5: 147- 178. Saunders. J.W.. Jr. and Gasseling, M.T. (1963). Trans-filter propagation of apical ectoderm maintenance factor in the chick embryo wing bud. Dei.. Biol. 7: 64-78. Saunders, J.W. and Gasseling, M.T.(1968). Ectodermal-mrscnchymai interactions in the origin of lirnb symmetry. In Epithelial-MesenchynInl Inreroctions. (eds. R. Fleischmajer and R.E. Billingham) pp. 78-97. Baltimore: Williams and Wilkins. Saunders. J.W.Jr. (1972). Developrnental control of threr-dimensional polarity in the avian lirnb. Arrn. NY. Acad. Sci. 193: 29-42. Saunders, J.W.. Jr. and Reuss. C. (1974). Inductive and axial propenies of prospective wing-bud mesoderm in the chick embryo. Dev. Biol. 38: 1 L -50.

Schaeren-Wiemcrs. N.. and Gerfin-Moser. A. (1993). A single protocol to derect transcripts of various types and expression levels in neural iissucs and çultured cells: in situ hybridization using digoxigenin-labeled cRNA probes. Hisroclici~tisrn.100: 43 1-440. Sçherer. S.W.. Poorkaj. P., Allen, T., Kim. J.. Geshuri. D.. Nunes. M., Soder, S.. Stephens. K.. Pagon. R.A.. Patton. M.A., Berg. M.A.. Donlon. T.. Rivera. H.. Pfeiffer. R.A.. Naritorni. K.. Hughes. H., Genuardi. M.. Gurrirri. F.. Neri. G., Lovrein, E. Magcnis. E.. Tsui. L.-C.and Evans, J.P. (199Ja). Fine mapping of the autosomal dominant split hand split foot malformation on chromosome 7 band q11.3-q22.1. Am. J. Httm. Genet. 55: 12-20. Scherer. S.W.. Poorkaj. P., Massa. H.. Soder. S.. Allen. T.. Nunes. M.. Geshuri. D.. Wong, E.. Belloni, E.. Little. S., Zhou, L., Becker, D.. Kere. J.. Ignatius. J., Niikawa, N.. Fukushirna. Y.. Hasegawa. T., Weissenbach. J.. BoncineIli. E.. Trask. B.. Tsui, L.- C. and Evans. 1.P. (1994b). Physical mapping of the split handl split foot locus on chromosome 7 and implication in syndromic ectrodaciyly. Him. Mol. Genet. 3: 1345- 1354.

Schroer. T. A. (1994). Srructure. function and regdation of cytoplasrnic dynein. Curr. Opiri. Ce11 Biol. 6: 69-73.

Searle. AG.. ( 1952). Hrreditary "split-hand in the domestic cat. Arin Eugen. 17: 279.

Searls. R.L.. and Zwilling, E. (1964). Regenention of the apical ectodermal ndge of the chick limb bud. Der?.Biul. 9: 38-55.

Srarls. R.L.. and Jmnrrs, M.Y. (1971). The initiation of limb bud outgrowth in the embryonic chick. De\?. Biol. 24: 198-213. Serrano. N., and OiFarrell. P.H.(1997). Limb morphogenesis: connections between patterning and growth. Curr. Biol. 7: R 186-R195. Seto, M.L.. Nunes. M.E.. Macarthur. C.A.. and Cunningham, M.L. (1997). Pathogenesis of ectrodactyly in the Dacrylnplasia mouse: aberrant ce11 death of the apical ectodermal ndge. Teratology. 56: 262-270. Sharland. M., Patton, M.A., and Hill, L. (1991). Ectrodactyly of the limbs and feet in a child with a complex translocation including 7q2 1.2. Am. J. Med. Genet. 39: 4 13-4 14.

Sharp. P.A. ( 1993). TATA-binding protein is a classless factor. Ceil. 68: 8 19-82 1. Shpetner. H. S.. Paschal. B. M.. and ValIee, R. B. (1988). Characterization of the microtubule-activated ATPnse of bnin cytoplasmic dynein (MAPIC). J. Cell. Biol. 107: 1001-1009.

Sidow. A.. Bulotsky. M.S.. Kerrebrock, AN.. Bronson. R.T.. Daiy. M.J.. Reeve. M.P., Hawkins. T.L.. Birren. B.W.. Jaenisch. R., and Lander. E.S. ( 1997). Serrate2 is disrupted in the mouse limb-development mutant syndacty lism. Narure. 389: 7 22-725.

Silver. L. M. (1993). The pcculiar joumey of a selfish chromosome: mouse t hapiotypes and meiotic drive. Trends Genet. 9: 250-254.

Sirneone. A.. Acampora, D.. Pannese. M.. D'Esposito. M.. Stornaiuolo. A.. Gulisano. M.. Maliamaci. A.. Kasiury. K.. Druck. T.. Huebner. K.. and Boncinelli. E. ( 1994). Cloning and characterization of two new members of the venebrate Dlx sene farnily. Proc. ~Vutl.Actrtl. Sci. USA. 91: 2250-2254.

Slaughter. G. R.. Meistrich. M. L.. and Means. A. R. (1989). Expression of RNAs for calmodulin. actins. and tubulins in rat testis cells. Biol. Reprod. JO: 395-105. Stadler. H.S.. Murray. J.C.. Leysens. N.J.. Goodfellow. P.J.. and Solursh. M. (1995). Phylogenetic conservation and physical mapping of mernbers of the H6 homeobox gene family . itinnzrrt. Genome. 6: 383-385.

Stein. H.C., and Bettman, EH. (1940). Rare malformation of the arm. Double humerus with three hands and sixteen fingen. Am. J. Swg. 50: 36.

Stephçns. T.D., and McNaulty. T.R. ( 198 1). Evidence for a metameric pattern in the development of the chick humerus. J. Emb~@. &p. Morphol. 61: 19 1-205. Stevenson. A. C. and Jennings. L. M. (1960). Ectrodactyly-evidence in favour of a disturbrd segregation in the offspring of affected males. Ann. H~irn.Genet. Lund. 24: 89- 96. Stone. D.M.. Hynes. M.. Armanini, M., Swanson, T.A.. Gu. Q.. Johnson. R.L.. Scott. M.P.. Psnnica. D., Goddard. A.. Phillips. H., Noll. M.. Hooper. LE., de Sauvage. F.. and Rosrnthal. A. (1996). The tumour-suppressor gene patchcd encodes a candidate receptor for Sonic hedge hog. Numre. 384: 139- 134. Strecker. T.R.. and Stephens. T.D. (1983). Penphenl nerves do not play a trophic role in limb skeletai morphogenesis. Teratology. 27: 159- 167. Summerbell. D.. Lewis. J.H. and Wolpert. L. ( 1973). Positional information in chick limb morphogenesis. Nature. 244: 492-496. Summerbell. D. (1974). A quantitative analysis of excision of the AER from the chick limb bud. J. Embryol. Erp. Morphol. 32: 65 1-660. Summerbell. D. (1979). The zone of polarizing activity: evidence for a role in normal chick limb morphogenesis. J. Embryol. Eïp. Morphol. 50: 2 17-233. Summerbell. D. (1983). The effect of local application of retinoic acid to the anterior margin of the developing chick limb. J. Ernbyol. Eip. Morphol. 78:269-289. Supp. D. M.. Witte. D. P.. Potter, S. S.. and Brueckner. M. (1997). Mutation of an axonrrnal dynein affects left-right asymrnetry in inversus viscerum mice. Nature. 389: 963-966.

Susin. S.A.. Zamzarni. N., Castedo. M.. Hirsch. T.. Marchetti. P.. Macho, A.. Daugas. E., Geuskens. M., and Kroemer, G. (1996). Bcl-2 inhibits the mitochondrial release of an apoptogenic protease. J. Erp. Med. 184: 133 1 - 1342. Susin. S.A.. Zamzarni. N.. Castedo. M.. Daugas, E.. Wang. H.-G..Geley. S., Fassy. F.. Rrcd. J.. and Kroemer. G. ( 1997). The central rxecutioner of apoptosis. Multiple links be twcen pro teasr acti vaiion and mi tochondria in Fas/Apo- 1lCD95-and ceramide-induced apoptosis. J. E-vp. Med. 186: 25-37.

Swett. F.H. ( 1937). Determination of limb-axes. Qriart. Rrv. Biol. 12: 322-339.

Tabin. C.J. (1992). Why we have (only) five fingers per hand: Hox genes and the evolution of paired limbs. Drvelopment. 116: 289-296.

Tajara. E.H..Varella-Garcia. M.. and Gusson. A. ( 1989). lnterstitial long-am deletion of chromosome 7 and ectrodactyly. Am J. Med. Genet. 32: 192- 194.

Temtamy. S. A.. and McKusick. V. A. (1978). In The Genetics of Hand Malformation. Alan R. Liss. Inc.. New York, pp. 53-7 1.

Thomas. B.L.. Tuckcr. AS.. Qiu. M.. Ferguson. C.A.. Hardcastle, 2.. Rubenstein, J.L.. and Sharpe. P.T. (1997). Role of Dlx-l and DL-2 genes in patterning of the murine dentition. Deidopment. 124: 18 1 1-48 18.

Thomas. K.R. and Capecchi. M.R. ( 1987). Site-directcd mutagenesis by gene targeting in rnouse embryo-derived stem cells. Cell. 51: 503-512.

Thompson. M.W.. Mclnnes, R.R. and Willard, H.F. ( 199 1). Genetics in Medicine. 5th edn. W.B. Saunden. Toronto, pp. 83-85.

Ticklr. C.. Summerbell. D.. and Wolpert. L. (1975). Positional signaling and specification of digits in chick limb morphogenesis. Nature. 251: 199-202. Tickle. C.. Albens. B.. Wolpert, L.. and Lee. J. (1982). Local application of retinoic acid to the limb bud mimics the action of the polarizing region. Nature. 296: 564-566.

Ticklr. C.. Lee. J.. and Eichele. G. (1985). A quantitative analysis of the effect of dl- rrans-retinoic acid on the pattern of chick wing development. Dei: Biol. 109: 82-95. Tickle, C. ( 199 1). Retinoic acid and limb patteming and morphogenesis. In Developmental patteming of the vertebrate litnb. eds. J.R. Hinchliffe. .J. Hurle. D. Summerbell. Plenum, New York. pp. 143-150. Tickle, C., and Eichele. G. (1994). Venebrate limb developrnent. Annu. Rev. Cell Biol. 10: 121-152.

Tsui, L.-C., and Farrall. M. ( 199 1 ). Human gene rnapping 1 1. Qtogenet. Ce11 Genet. 58: 337-38 1.

Tsui. L.-C..Donis-Keller. H.. and Grzeschik. K. H. (1995). Report of the second international workshop on human chromosome 7 rnapping 1994. Cyrogenet. Ce11 Genet. 71: 2-21.

Uberbacher. E.C., and Mural. R.J. (1991). Locating protein-coding regions in human DNA sequences by a multiple sensor-neural network approach. Proc. Natl. Acad. Sci. USA. 88: 11261-t L265.

Vaahtokari. A., Aberg, T. and Thesleff. I. (1996). Apoptosis in the developing (00th: association with an embryonic signaling center and suppression by EGF and FGF-4. Development. 122: 17 1 - 179. Vachon, G.. Cohen, B.. Pfeifle. C., McGuffin. M.E.,Botas, J.. and Cohen. S.M. (1991). Homeotic gençs of the Biihorax complex repress limb developmeni in the abdomen of the Drosophila embryo ihrough the tarpet gene Disral-less. Cd.71: 137450.

Vaughan. K. T.. and Vallc. R. B. (1995). Cytoplasmic dynein binds dynactin through a direct interaction betwern the intermediate chains and p l50Glued. J. Cell Biul. 13 1: 1507- 1516.

Vaughan. K. T.. Mikami. A.. Paschal. B. M.. Holzbaur. E. L. F.. Hughes. S. M.. Echeverri. C. J.. Moore. K. J.. Gilbert, D. J., Copeland. N. Ci.. Jenkins. N. A.. and Vallee. R. B. (1996). hlultiplr mouse chrornosornal loci for dynein-based motility. Genomics. 36: 29-38.

Verma. 1.. Joseph. R.. Bhargava, S. and Mehta, S. (1976). Split-handlsplit-foot defomity inherited as an autosomal recessive trait. Clin. Genet. 9: 8-14.

Viljoen. D.L.. and Sman. R. ( 1993). Split foot anornaly. microphthalmia. cleft lip and cleft palate and mental retardation associated with a chromosome 6: 13 translocation. Clin. D~smorphol.2: 271-277.

Vogel, A. and Tickle, C. (1993). FGF-4 maintains polarizing activity of posterior limb bud cells in vivo and in vitro. Deidopnient. 119: 199-206.

Vogel. A., Rodriguez. C.. Warnken. W., and Izpisua-Belmonte. LC. (1995). Dorsal ce11 fate specified by chick hriduring vertebnte limb development. Norive. 378: 7 16-720.

Vogel. A.. Rodriguez. C. and Izpisua-Belmonte. J.-C. (1996). Involvement of FGF-S in initiation. outgrowth and patterning of the venebraie limb. Developrnent. 122: 1737-1750.

Vogel. F. ( 1957/ 1958). Vrrzogene mutation beim menschen? .4nnols Hum. Ge~ier.22: 132-137. Vortkamp, A.. Gessler, M. and Grzeschik, K.-H. (1991). GL13 zinc-finger gene interrupted by translocations in Greig syndrome families. Nature. 352: 539-540. Wagner, T., Wirth, J., Meyer, J., Zabel, B., Held. M., Zimmer, J., Pasantes, J., Bricarelli, F.D., Keutel, J., Hustert. E., Wolf, U., Tomrnerup, N.. Schempp, W. and Scherer, G. (1994). Autosornal sex reversal and campomelic dysplasia are caused by mutations in and around the SRY-related gene SOX9. Cell. 79: 1 1 1 1-1 120. Wallis. C.E. (1988). Ectrodactyly (SWSF)and ectodermal dysplasia with normal lip and palette in a four generation kindred. Clin. Genet. 31: 257-257.

Wanek. N.. Muneoka. K.. Holler-Dinsmore, G., Bunon. R. and Bryant, S.V. (1989). A staging system for mouse limb bud development. J. Exp. 2001.249: 41-49.

Warkany, J. (1959). Congenital malformations in the past. J. Chronic Dis. 10: 84-

Wassif, C.A.. Maslen, C., Kachilele-Linjewile. S.. Lin. D.,Linck. L.M., Connor. W.E.. Steiner. R.D..and Porter, F.D. (1998). Mutaions in the human sterol 7-reductase gene at I lq 12- 13 cause Smith-Lemli-Opitz syndrome. Anr. J. Nitni. Genet. 63: 55-62. Watson. A.G. and Bonde. R.K. (1986). Congenital malformations of the flipper in three wcst indian manatees, Tricheclii

Wrstrrtkld. M. ( 1993). The zebrafish hool University of Oreson Press. Eugene.

Wilkinson. D.G.. (1992). Iii Sitii H-vbridisciriori: A Pruc~icalApproncli. IRL Press at Oxford University Press. Oxford.

Willison. K., and Ashworth. .A. (1987). Mammalian sprrmatogenic gene expression. Trends Gener. 3: 35 1-355.

Witman. G. B. ( 1992). Axonemal dyneins. Ciirr. Opin. Ceil Biol. 4: 74-79. Wolpen. L. (1989). Positional information and prepattern in the development of pattern. In Ce11 to Celt Signolling: From Experimeiirs to Theoretical Models. ed. A. Goldbetter, Acadrmic Press. New York. pp. 133- 143.

Yasecn, N.R., and Blobel. G. (1997). Cloning and characterizaiion of human kaqopherin b3. Proc. Nutl. Acad. Sci. USA. 94: 445 1-4356. Yang, Y.. Guillot. P.. Boyd. Y.. Lyon. M.F.. and McMahon. A.P. (1998). Evidence that preaxial polydactyly in the Doitblefoot mutant is due to ectopic Indian Hedgehog signaling. Developnient. 125: 3 123-3 132. Yokouchi, Y., Sasaki, H.. and Kuroiwa. A. (1991). Horneobox gene expression correlated with the bifurcation process of limb cartilage development. Nature. 353: 443- 445. Zkiiny. J., and Duboule, D. (1996). Synpolydactyly in mice with a targeted deficiency in the HoxD complex. Nanrre. 384: 69-7 1.

Zeller, R. and Duboule. D. (1997). Dorso-ventral limb polarity and the origin of the ridge: on the fringe of independence? BioEssays. 19: 541-546. Zhao. G.-Q., Zhao. S.. Zhou. X.. Eberspaecher. H.. Solursh. M. and de Crombrugghe, B. (1994). rDlx, a novel distal-less-like homeoprotein is expressed in developing cartilages and discrere neuronal tissues. Dev. Biol. 164: 37-5 1.

Zlotogora. 1.. and Nubani. N. (1988). 1s there an autosomal recessive form of split hand and split foot malformation? J. Med. Genet. 25: 138- 140.

Zlotogon. J. (1994). On the inheritance of the split handhplit foot malformation. Am. J. Mcd. Genet. 53: 29-32.

Zoratti. M.. and Szabii. 1. (1995). The mitochondriai pcrmeability transition. Biochem. Biophy. Acta Rer*. Bionrenibranes. 124 1: 139- 176.

Zou, H.. and Niswander. L. ( 1996). Requirement for BMP si~nalingin interdigital apoptosis and scale formation. Science. 272: 738-741.

Zwilling. E. ( 1956). Interaction between limb bud ectoderm and mcsodcrm in the chick embryo. 1. Axis establishment. J. Erp. 201. 132: 173- 188.

Zwilling. E. and Hansborough. L. (1956). Interaction between limb bud rnesoderm and ectodenn in the chick embryo. III. Experiments with polydnctylous limbs. J. Erp. 201. 132: 2 19-239.

Zwilling. E. ( 196 1 ). Limb morphogenrsis. Advan. Morphogenesis. 1: 301-330.

Zwilling. E. ( 1964). Developrnent of fragmented and of dissociated limb bud rnesoderm. Dev. Biul. 9: 20-37.