sign in sign up
<<
Home , Homeosis, Homeotic gene, Segmentation (biology), Ultrabithorax, William McGinnis

Copyright 0 1994 by the Genetics Society of America Perspectives

Anecdotal, Historical and Critical Commentaries on Genetics Edited by James F. Crow and William F. Dove

A Century of , A Decade of

William McGinnis

Department of Molecular Biophysics and Biochemistry, , New Haven, Connecticut 06520-8114

NE hundred years ago, while the science of genet- ing mammals, and were proposed to encode DNA- 0 ics still existed only in the yellowing reprints of a binding homeodomainsbecause of a faint resemblance recently deceased Moravian abbot, WILLIAMBATESON to mating-type transcriptional regulatory proteins of (1894) coined the term homeosis to define a class of budding yeast and an even fainter resemblance to bac- biological variations in whichone elementof a segmen- terial helix-turn-helix transcriptional regulators. tally repeated array of organismal structures is trans- The initial stream of papers was a prelude to a flood formed toward the identity of another. After the redis- concerning and homeodomain pro- coveryof MENDEL’Sgenetic principles, BATESONand teins, a flood that has channeled into a steady river of others (reviewed in BATESON1909) realized that some homeo-publications, fed by many tributaries. A major examples of homeosis in floral organs and skel- reason for the continuing flow of studies is that many etons could be attributed to variation in genes. Soon groups, working on disparate lines of research, have thereafter, as the discipline of genetics was found themselves swept up in the currents when they born and was evolving into a formidable intellectual found that their favorite protein contained one of the force enriching many biologicalsubjects, it gradually be- many subtypes of homeodomain. This was in part be- came clear that fruitflies contained multiple “homeotic” cause the definition of what proteins belonged to the genes (e.g., bithorax,aristapedia and ) homeodomain family expanded to include proteins that (BRIDGESand MORGAN1923; BALKASCHINA 1929;BRIDGES had only marginal amounts of sequence similarity to the and DOBZHANSKY1933), some of which appeared to be founder members in the Drosophila and loosely clustered on the third chromosome. These ge- Bithorax complexes. Many of the proteins that netic studies culminated in the systematic analyses of have homeodomains have nothing todo with BATESON’S LEWIS(1978) and KAUFMAN et al. (1980),which provided version of homeosis, although there is a loosely defined preliminary definitions of the many homeotic genes of structural subgroup of homeodomains that is closely the Bithorax and Antennapedia complexes, and also linked to homeotic genetic functions in . showed that the mutant for most of these The initial stream of reports immediately explained genes could be traced back to patterning defects in the (or purported to explain) some of the burning ques- embryonic . tions concerning homeotic genes. They seemed to be a Ten years ago, a sudden stream of papers (MCGINNIS fairly closely conserved gene family, and the sequence et al., 1984a,b,c; SCOTTand WEINER1984; LAUGHON and that validated their family membership, the homeobox, SCOTT1984; SHEPHERDet al. 1984; CARRASCO et al. 1984; provided a plausible biochemical function for their ac- LEVINEet al. 1984) introduced the homeobox to devel- tion. They were likelyto be DNA-binding transcriptional opmental genetics and sketched its basic outlines. In regulators that would modulate the expression of many retrospect, each of these studies contained relatively few downstream genes. The conservation of very similar ho- data for theimpact they had. Putting thebest face on it, meobox sequences in other animals suggested that one could claim they are reports of exemplary brevity. homeotic-like genetic functions might exist in structur- These reports defined homeoboxes as members of a ally homologous genes other than in Drosophila. That highly conserved familyof DNA sequences that ap is, perhaps a conservation of developmental genetic cir- peared to be preferentially associated with homeotic cuitry could be detected at themolecular level that was and genes of Drosophila. Homeobox se- invisible at the level of comparative embryological mor- quences were highlyconserved in other animals, includ- phology. All of these things had been suggested before

Genetics 137: 607-611 uuly, 1994) 608 W. McCinnis

in either explicit or vague terms by those with insight, to clone genes. This was especially true in Drosophila, prescience, and/or theoretical leanings (WOLPERT 1969; which already had a mother lode of genetic and cyto- GARCIA-BELLIDO 1977; GARCIA-BELLIDOet al. 1979; LEWIS genetic studies as a biological treasure. And this rich 1978; RAFF and KAUFMAN 1983), but a bit of molecular lode ofgenetics was indeed minedfor all it was worth by evidence goes a long way toward swaying opinion (e+ anyone with a homeobox probeand a hypothesis (e.g., pecially the opinion of molecular biologists), so that FJOSEet al. 1985; LEVINEet al. 1985; REGULSKI et al. 1985; much was made of the homeobox discovery. MACDONALD et al. 1986).For those of us who weredoing In the original set of reports, the evidence for any of something with homeoboxes as students or postdocs the conclusions was incomplete at best, which did not with WALTER GEHRINGin Basel, Switzerland (which in- prevent the original authors from discussing them as cluded MICHAELLEVINE, ATSUSHIKUROIWA, ERNST HAFEN, quite likely to be true. There was even more hope ex- ANDERS FJOSE,MAREK MLODZIK,and me),it will be difficult pressed (and a bit of metaphorical hyperbole, at least in to forget the feeling of guiltypleasure when we realized the titles) in a variety of review articles that suggested how incredibly easy it might be to clone andidentify the variously thatthe homeobox might be a biological coding regions of the Drosophila homeotic genes and equivalent of the Rosetta stone, theuniversal genetic key many of the segmentation genes. That this suspicion to body plan, and so on (e.g.,STRUHL 1984; SLACK1984). wasn’t entirely a Swiss chocolate-inspired delusion was There were even articles in newspapers and popular fortified by a chanceconversation with GINESMORATA at magazines announcing that something important had a Swiss-USGEB meeting. Some of us, in collaboration happened in that might be rel- with FRANCOIS mCHand WELCOMEBENDER, had found evant even to those sophisticated mammals that perform only three homeoboxes in Bithorax complex DNA a daily perusal of The New York Times. All of this at- (REGULSKI et al., 1985). At the time, thenumber of tention jump-started homeobox gene research in Dro- proteincoding transcription units in the Bithorax com- sophila, where it would be defined and enrichedbeyond plex was thought to be eight or more, but MORATA, anyone’s wildest dreams by the rich genetics of that ani- ERNESTO SANCHEZ-HERRERO,and their co-workers hadjust mal. But perhaps themost hope, and themost rapid and discovered that the Bithorax complex contained only concerted jump outof the homeobox research starting three lethal complementationgroups (SANCHEZ- gate, occurred in laboratories studying development in HERREROet al. 1985),suggesting correctly that the three those that had a rich history of descriptive bithorax homeoboxes corresponded to those three le- and experimentalembryology, but rudimentary genetic thal genes, now known as Ubx, abd-A and Abd-B. tool kits compared with Drosophila. Herethe ho- One of the most exciting outcomes of the early ho- meobox seemed to provide a toehold halfway up what meobox research in Drosophila was the general way it had seemed to be a slippery and impassable barrier of confirmed some of E. B. LEWIS’Sspeculations about the developmental genetics. evolution of the Bithorax complex. In an article that is Not all concurred with the blinkered enthusiasm over oft cited but rarely read in its complex entirety, LEWIS the meaning and utility of the homeobox . (1978) proposed that the Bithorax complex genes were Some fancifully suggested that many developmental bi- members of a gene family, having duplicated and di- ologists were inthe gripof “homeobox madness” or “ho- verged from a common ancestor and in the process hav- meobox fever” (RAFF and RAFF 1985; ROBERTSON1985; ing acquired divergent functions that accounted for WILKINS1986), apparently a horizontally transmitted dis- some of the morphological differences that distinguish ease that caused a loss of one’s critical faculties. Some the Drosophila body plan from that of more primitive geneticists and evolutionary biologists were thought to . Luckily for some of us,LEWIS put thatspecu- be immune to this syndrome. Many of those with cool lation in the first paragraph instead of in the middle of heads who read theoriginal homeobox papers carefully, the article among the terse and tortuous genetics. T. C. and interpreted them critically, found some of the ar- KAUFMAN (RAFF and KAUFMAN 1983) had also proposed an guments specious. And some were, if the results within extension of this to embrace the homeotic genes of the a particular paper are considered in isolation. In those Antennapedia complex controlling head and thoracic days, however,the results were coming along so fast that development. It is still unclear how much the variation by the time one paper was written, the results for the next, in homeotic protein function or expression pattern can or the next two or three,were already in one’snotebook. account for evolutionary changes in arthropods, but it So the temptation was to “speculate”rather boldly on be- was eventually shown that the eight homeotic genes of half of some of the early general conclusions described the Antennapedia and Bithorax complexes (now con- above, this being much safer than it looked since addi- ceptually grouped as the Homeotic Complex, or tional evidence to support them was already in hand. HOM-C genes; AKAM 1989) contained eight structurally Many people working on a variety of developing ani- similar homeobox sequences, sometimes designated as mals quickly realized that the homeobox, whatever its the Antpclass of homeoboxes (GEHRINGet al. 1990; ultimate meaning, should be exploited as a useful tool MCCINNISand KRUMLAUF 1992). Perspectives 609

Conservation of anterior-posterior axial patterning: Also adding to the same intellectual picture are the One important threadin homeobox gene research has highly influential studies indicating that both beetles been the studies on the Antelass Hox genes in other and nematodes encode an important partof the genetic animals, particularly in the mouse. It was found early on circuitry that controls their anterior-posterior axial pat- that Hox genes were expressed in discrete anterior- terning in clusters of HOM/Hox-type homeobox genes posterior regions of ( e.g., AWGULEWITSCHet al. (BEEMANet al. 1989; WANGet al. 1993). With the finding 1986; GAUNTet al. 1986), that some of the Hox genes that some of the most primitive animals like hydra have mapped in clusters (LEVINEet al. 1985;HART et al. 1985), Antpclass homeobox genes that areexpressed in local- and that some Drosophila genes were muchmore ized body regions (SHENKet al. 1993), itseems possible closely related in structure to certain mammalian Hox that many or all animals use Antpclass genes in HOM/ genesthan to other Drosophila homeoboxgenes Hox clusters to assign positional identities on the (REGULSKI et al. 1987). But it took the comprehensive anterior-posterior axis (or oral-aboral axis where head is and insightful studies ofBONCINELLI et al. (1988), more difficult to define). Thus,only one hundredyears GRAHAMet a2. (1989), and DUBOULEand DOLL; (1989) after BATESONfinished analyzing some bizarre variations to put it all together. All three groups provided con- in skeletons, and insightfully grouped a class of them as vincing evidence that individual Hoxgenesmapped in the homeotic variations, we now have plausible molecular same relative positions inone of the four Hox complexes explanationsfor the homeotic defects, and a near certainty as did (some of) their homologs in Drosophila. In addi- that many of the variations that he originally noticed in a tion, the latter two groups showed that the embryonic ex- variety of invertebrates and vertebrates are due to varia- pression boundaries of many of the mouse Hox complex tions in the same basic underlying genetic circuitry. genes mimicked their map order within the complexes, Homeodomain proteins as transcriptionfactors: again strikingly similar ain general sense to the properties Much of the currentresearch that concerns homeobox of the HOMtype homeobox genes of Drosophila. gene function has been substantially enriched by the Though many found all this to be compelling evi- work on homeodomain proteinsas transcription factors. dence that theHox genes must be doing somethingsimi- The Antp-type homeodomain proteins are a relatively lar to the Drosophila HOM-C genes, it was still correla- small subset of the total spectrum of proteins grouped tive molecular evidence. The first strong biological in the homeodomain family. The only criterion for ad- evidence as to the role of the Hox genes came from mission to this family isthe conservation of a few crucial inducing their expression anterior to their normallim- amino acid residues that tend to reside in the same po- its, or artificially reducing theirlevels ofexpression, both sitions in the 60-amino-acid primary sequence of known of which caused some interpretable and some uninter- homeodomains. Structural studies of highly divergent pretable defects in the development of more anterior homeodomains suggest that most of thefamily members regions of thefrog or mouse (WRIGHTet al. 1989; defined by these criteria willhave extremely similar RUIZI ALTABA and MELTON1989; KESSEL et al. 1990). By threedimensional structures and similar interactions expressing Hox proteinsin developing Drosophila, one with DNA binding sites (GEHRINGet al. 1990; KISSINGER could also get mouse and human Hox proteins to phe- et al. 1990; WOLBERGERet al. 1991). There are many nocopy specific Drosophila HOM gain-of-function mu- hundreds of homeodomain proteins, in many separate tations (WICKIet al. 1990; MCGINNIS et al. 1990),which subclasses (Scorn et al. 1989). It seems likely that hun- indicated that theHox proteins certainly had homeotic dreds exist even within a single -the current genetic functions in the context of Drosophila cells, count in Drosophila is >60 and climbing (UIONISand though still saying little or nothing about their role in O’FARRELL1993; DESSAINand MCGINNIS1993)-and these mouse or human cells. However,in many recent studies proteins are surely involved in a myriad of biological performed over the past few years (e.g., CHISAKAand control circuits. Many of these are understood quite CAPECCHI 1991; LUFKINet al. 1991;LEMOUELLIC et al. poorly at the genetic level. 1992; RAMIREZSOLISet al. 1993), mouse Hox genes have Evidence was not long in coming that homeodomain been mutatedby gene targeting, and many ofthese loss- proteins actually did have the ability to bind specific of-function result in either loss of axialsmc- DNA sites and that proteins with different homeodo- tures or subtle to obvious homeotic transformations of main sequences had different preferred binding sites skeletal elements and/or rhombomere elements of the (DESPLANet al. 1985,1988;Howand LEVINE1988). How hindbrain. These studies have represented one of the much these different DNA binding preferences have to principal success stories for the practice of “reverse ge- do with their functional specificity is still rather myste- netics,’’ a discipline that has resulted in a “reversal of rious. The most widely accepted model (or class ofmod- fortune” for more than onelong-suffering graduate stu- els) explaining homeodomain proteinfunctional speci- dent or postdoc who has not been so fortunate as to have ficities is largely derived from thebiochemical studies on the mouse Hox genes as the focus of his or her mutant yeast Mata2, mammalian Octl andOct2 and otherPOU screen. proteins (e.g., Tmmet al. 1992; POMERANTZet al. 1992; 610 W. McGinnis

VERSHONand JOHNSON 1993; CLEARYet al. 1993), with spite the obvious involvement of some other homeo- much support from genetic studies on chimeric HOM domainprotein subgroups in mouse (andhuman) proteins in Drosophila embryos (e.g., LINand MCGINNIS heritable morphological abnormalities. 1992; FUROKUBO-TOKUNAGAet al. 1993; CHANand MANN Disclaimer: This essay isdefinitely not intendedto be 1993). This model has the homeodomain presenting a scholarly review of homeobox gene research over the one face to DNA and acquiring abit of its specificityfrom past 10 years, just anadmittedly biased look back at what that interaction.The otherface, a sociable but discrimi- happened 10 years ago, and to look at how a few of the nating face, is free to interact with one or many other questions that were interesting then have either been proteins either on or off DNA. Only when the right set answered (or not)in the ensuing period.I’ve benefited of interactions takes place on both faces is a given ho- enormously from talking to all of the people working on meodomain protein interpreted as part of an active or HOM and Hox genes and proteins and to many of those inactive transcriptional regulatory complex that is ca- working on otherclasses of homeobox genes, and I have pable of flipping a developmental switch. been influenced by nearly everyone. Thus, they allbear Some questionsthat might have been answeredin 10 a highly diffuse responsibility for the opinionsexpressed years, but might take another10 (or 100, until BATESON’S here,though certainly no blame forthe manner in second centennial anniversary):Though the amountof which they are expressed. research that has beendone on HOM/Hox-type LITERATURE CITED homeodomain proteins is enormous, it is still unknown how many genetic or cofactor inputs are required for a AKAM, M., 1989 Hox and HOM: homologous gene clusters in and vertebrates. Cell 57: 347-349. homeotic switch to be thrown that changescells (or even AWGUI.EWITSCH,A,, M. F. UTSET,C. P.HART and F. H. RUDDLE, a single gene for that matter)from being assigned to a 1986 Spatial restriction in expression of a mouse homeo box head, thoracic, or abdominal fate. Anotherway to look locus within the central nervous system. Nature 320: 328-335. BAI.WSCHINA,E. I., 1929 Ein Fall der Erbhomeosis (die Genovarition at this is that the genetic andmolecular interactions be- “aristopedia”) beiDrosophila melanogaster. Wilhelm Roux’ Arch. tween the homeotic proteins and theproteins that con- Entwicklungsmech. Org. 115: 448-463. trol other equally (or more) important developmental BATESON,W., 1894 Materials for theStudy of Variationlreated with Especial Regard to Discontinuity in the Origin of Species. decisions such as sex determination, muscle or nerve Macmillan, London. cell identity, the timing of developmental events, or con- BATESON,W., 1909 Mendel’s Principles of Heredity. University Press, served signal transduction pathways are largely unex- Cambridge, England. BEEMAN,R. W., J. J.STUART, M. S. HAASand R. E. DENELL,1989 Genetic plored and mysterious. analysis of the complex (HOM-C) in the beetle Westill don’t really understand why the HOM/Hox Tribolium castaneum. Dev. Biol. 133 196-209. genes tend to be arranged in a colinear arraythat (usually) BONCINELLI,E., R. SOMMA,D. ACAMPORA,M. PANNESE, M. F. A. D’ESPOSITO et al., 1988Organization of humanhomeobox genes. Hum. correlates with the order of their domains of expression Reprod. 3: 880-886. and function in embryos. There aresome appealing ideas BOTAS,J., 1993 Control of and differentiation by HOM/Hox genes. Curr. Opin. Cell Biol. 5: 1015-1022. about how these clustersmight have arisenand theforces BRIDGES,C. B., andT. DOBZHANSKY, 1933The mutant “proboscipedia” that might tend to keep them together, involving shared in Drosophilamelanogaster-a case of hereditary homeosis. regulatory regions (e.g., CELNIKERet al. 1990), but there is Wilhelm Roux’ Arch. Entwicklungsmech. Org. 127: 575-590. BRIIXES,C. B., and T. H. MORGAN,1923 The third-chromosome not enough evidence as yet to provide a convincing ex- group of mutant characters ofDrosophila melanogaster. Carnegie planation for the persistent colinear arrangements. Inst. Washington Publ. 327: 93. We haveonly a primitive understanding ofhow CARRASCO,A. E., W. M(:GINNIS,W. J. GEHRING andE. M. DEROBERTIS, 1984 Cloning of a Xenopus laevis gene expressed during early HOM/Hox proteins, or any other homeodomain pro- embryogenesis coding for a peptide region homologous to Dro- teins for that matter, mighthave the wholesale but CO- sophila homeotic genes. Cell 37: 409-414. ordinated effects that they exert on morphogenesis. In CELNIKER,S. E., S. SHARMA,D. J. KEELAN and E. B. Lewis, 1990 The molecular genetics of the bithorax complex of Drosophila: cis- Drosophila, the HOM proteins are known to regulate regulation of the Abdominal-B domain. EMBO J. 9: 4277-4286. the expression of other genes that encode other tran- CHAN,S., and R. S. WN,1993 The segment identity functions of scription factors, growth factors, homophilic membrane arecontained within its homeo domain and carboxy-tprminal sequences. Genes Dev. 7: 796-81 1. proteins, and proteins of unknown function (reviewed CHISAW,O., and M. R. CAPECCHI,1991 Regionally restricted devel- in BOTAS1993), buthow is this all coordinated to result opmental defects resultingfrom targeted disruption ofthe mouse in an antenna instead of a leg, or even a gut constriction? homeobox gene hox-1.5. Nature 350: 473-479. CLEARY,M. A,, S. STERN,M.T.4NAKA and W. HERR,1993 Differential One interesting curiosity is that despite the impor- positive control by Oct-1 and Oct-2: activation of a transcription- tance that the human Hox genes must have during de- ally silent motif through Oct-1 and W16 co-recruitment. Genes velopment, there is surprisingly little direct or indirect Dev. 7: 72-83. DESPWN,C., J. THEIS andP. H. O’FAKRELL,1985 The Drosophila de- evidence that their proper functionis relevant to known velopmental gene, , encodes a sequence-specific DNA human heritable developmentaldefects or human tera- binding activity. Nature 318 630-635. tology (e.g., WOLCEMUTHet al. 1989). In addition,to my DESPIAN,C:., J. THEISand P. H. O’FARRELL,1988 The sequence speci- ficity of homeodomain-DNA interaction. Cell 54: 1081-1090. knowledge there areas yetno naturally occurring mouse DESSAIN,S., and W. MCGINNIS,1993 Drosophila homeobox genes. developmental defects that mapto the Hox clusters, de- Adv. Dev. Biochem. 2: 1-55. Perspectives 61 1

DUBOULE,D., and P. DOLLE, 1989 The structural and functional or- MCGINNIS,W., R. L. GARBER,J. WIRZ,A. KUROWAand W. J. GEHRINC, ganization of the murine HOXgenefamily resembles that of Dro- 1984b A homologous proteincoding sequence in Drosophila sophila homeotic genes. EMBO J. 8 1497-1505. homeotic genes andits conservation in othermetazoans. Cell 37: FJOSE, A, W. McGImand W.J. GEHRING,1985 Isolation of a homeobox- 403-408. containing gene from the engrailed region of Drosophila and the MCGINNIS,W., C. P. HART,W. J. GEHRINCand F. H. RUDDLE,1984c Mo- spatial distribution of its transcripts. Nature 313 284-289. lecular cloning and chromosome mapping of a mouse DNA se- FUROKUBO-TOKUNAGA,K., S. FLISTER and W. J. GEHRINC,1993 Func- quence homologous to homeotic genes of Drosophila. Cell 38: tional specificity of the Antennapedia homeodomain.Proc. Natl. 675-680. Acad. Sci. USA 90: 6360-6364. POMERANTZ,J. L., T. M. KFUSTIE and P. A. SHARP,1992 Recognition of GARCIA-BELLIDO,A,, 1977 Homeotic and atavic mutations in insects. the surface of a homeo domain protein.Genes Dev. 6: 2047-2057. Am. Zool. 17: 613-629. RAFF, E. C., and R. A. RAFF, 1985 Possible functions of the homeobox. GARCIA-BELLIDO,A,, P. A. LAWNCE and G. MORATA,1979 Compart- Nature 313: 185. ments in animal development. Sci. Am. Uuly): 102-110. RAFF,R. A,, and T. C. KAUFMAN,1983 Embryos, Genes, and Evolution. GAUNT,S. J., J. R. MILLER,D. J. POWELLand D. DUBOULE,1986 Homeo Macmillan, New York. box in mouse embryosvarieswith position by the RAMIREZ-SOLIS,R., H. ZHENG,J. WHITING,R. KRUMLAUF and A. BRADLEY, primitive streak stage. Nature 324: 662-664. 1993 Hoxb-4 (Hox-2.6)mutant mice show homeotic transfor- GEHRING,W. J., M. MULLER,M. AFFOLTER,A. PERCIVAL-SMITH, M. BILLETER mation of acervical vertebra and defects in closure of the sternal et al., 1990 The Structure of the homeodomain and its func- rudiments. Cell 73: 279-294. tional implications. Trends Genet. 6 323-329. REGUISKI,M., K HARDING, R.KOSTRIKEN, F. KARCH, M. LEVINEet al., GRAHAM,A,, N. PAPALOPULUand R. KRUMLAUF,1989 The murine and 1985 Homeo box genes of the antennapedia and bithoraxcom- Drosophila homeobox genecomplexes have common features of plexes of Drosophila. Cell 43: 71-80. organization and expression. Cell 57: 367-378. REGULSKI,M., N. MCGINNIS,R. CHADWICKand W. MCGINNIS,1987 De- HART,C. P.,A. AWCULEWITSCH,A. FAINSOD, W. MCGINNISand F. H. velopmental and molecular analysis ofDeformed a homeotic gene RUDDLE,1985 Homeo box genecomplex on mousechromo- controlling Drosophila head development.EMBO J. 6: 767-777. some 11: molecular cloning, expression in embryogenesis, and ROBERTSON,M., 1985 Mice, mating types and molecular mechanisms homology to a human homeo box locus. Cell 43: 9-18. of morphogenesis. Nature 318: 12-13. HOEY,T., and M. LEVINE,1988 Divergent homeo box proteins recognize RUIZI ALTABA,A,, and D. A. MELTON,1989 Involvement of the Xe- similar DNA sequences in Drosophila. Nature 332 858-861. nopus homeobox gene Xhox3 in along the KALIONIS, B., and P. H. OFARRELL,1993 A universal target sequence is anterior-posterior axis. Cell 57: 31 7-326. bound in vitro by diverse homeodomains. Mech. Dev. 43: 57-70. SANCHEZ-HERRERO,E., I. VERNOS, R. MARCOand G. MORATA,1985 Ge- huw,T. C., R Lnns and B. WAKIMOTO,1980 Cytogenetic analysis of netic organization of Drosophila bithorax complex. Nature 313: chromosome 3 in Drosophila mlanoptm the homeotic gene com- 108-1 13. plex in polytene chromosome interval 84A-B. Genetics 94: 115-133. Scon, M. P., and A. WEINER,1984 Structural relationships among KESSEL, M., R. BALLINGand P. GRUSS,1990 Variations of cervical ver- genes that controldevelopment: sequence homology between the tebrae after expression of a Hox-1. I transgene in mice. Cell 61: Antennapedia, Ultrabithorax, and fushi tarazu loci ofDrosophila. 301-308. Proc. Natl. Acad. Sci. USA 81: 4115-4119. KISSINGER,C. R., B. LIU,E. MARTIN-BLANCO,T. B. KORNBERGand C. 0. SCOT, M. P., J. W. TAMKUNand I. G.W. HARTZELL,1989 The structure PABO,1990 Crystalostructure of an engrailed homeodomain- and function of the homeodomain. Biochim. Biophys. Acta 989: DNA complex at 2.8 A resolution: a framework understandingfor 25-48. homeodomain-DNA interactions. Cell 63: 579-590. SHENK,M. A,, H. R. BODEand R. E. STEELE,1993 Expression of Cnox-2 LAUGHON,A., and M. P. Scon, 1984 Sequence of a Drosophila seg- a HOM/HOX homeobox gene in hydra, is correlated with axial mentation gene: protein structure homology with DNA-binding pattern formation. Development 117: 657-667. proteins. Nature 310: 25-31. SHEPHERD,J. C. W., W. MCG~NNIS,A. E. -0, E. M. DEROBERnS and LE MOUELLIC,H., Y. LALLEMAND and P. BRUWT,1992 Homeosis in the W. J. GEHRINC,1984 Fly and frog homeo domains show homologies mouse induced by a null in theHox-3.1 gene. Cell 69 with yeast mating type regulatory proteins. Nature 310 70-71. 251-264. SLACK,J., 1984 A Rosetta stone for pattern formation in animals? LEVINE,M., G. RUBIN andR. TIJAN,1984 Human DNA sequences ho- Nature 310: 364-365. mologousto a protein coding regionconserved between ho- STRUHL,G., 1984 A universal genetic key to body plan? Nature 310 meotic genes of Drosophila. Cell 38667-673. 10-11. LEVINE,M., K. HARDING,C. WEDEEN,H. DOYLE,T. HOEYet al., 1985 Ex- TREACY, M. N.,L. I. NEILSON,E. E. TURNER,X. HE andM. G. ROSENFELD, pression of the homeo box gene family in Drosophila.Cold 1992 Twin of I-POU: a two amino acid difference in the I-POU Spring Harbor Symp. Quant. Biol. 50: 209-222. homeodomain distinguishes an activator from an inhibitor of LEWIS, E. B.,1978 A gene complex controlling segmentation in Dro- transcription. Cell 68: 491-505. sophila. Nature 276 565-570. VERSHON, D.,and A. D. JOHNSON, 1993 A short, disordered protein LIN,L., and W. MCGINNIS,1992 Mapping functional specificity in the region mediates interactions between the homeodomain of the Dfd and Ubx homeodomains. Genes Dev. 6: 1071-1081. yeast a2 protein and the MCMl protein. Cell 72: 105-112. LUFKIN,T., A. DIERICH,M. LEMEUR,M. MARK and P. CHAMBON, WANG,B. B., M. M. MULLER-~MMERGI.UC,K,J. AUSTIN, N. T. ROBINSON,A. 1991 Disruption of the Hox-1.6 homeobox gene results in de- CHISHOLMet al., 1993 A homeotic gene cluster patterns the an- fects in a region corresponding to its rostral domain of expres- teroposterior body axis of C. elegans. Cell 74 29-42. sion. Cell 66: 1105-1 119. WILKINS,A. S., 1986 Homeobox fever, extrapolation and develop MACDONALD,P. M., P. INGHAMand G. STRUHL,1986 Isolation, structure mental biology. Bioessays 4: 147-148 and expression of even-skipped a second pair-rule gene of Dro- WOLBERGER,C., A. K. VERSHON, B.LIU, A. D. JOHNSON and C. 0. PABO, sophila containing a homeo box. Cell 47: 721-734. 1991 Crystal structure of a MATa2 homeodomainaperator IMALIcKI,J., K. SCHUCHARTand W. MCGINNIS,1990 Mouse Hox 2.2 complex suggests a general model for homeodomain-DNA in- specifies thoracic segmental identity in Drosophila embryos and teractions. Cell 67: 517-528. larvae. Cell 63961-967. WOLGEMUTH,D.J., R. R.BEHRINGER, M. P. MOSTOLLER,R. L. BRINSTERand MCGINNIS,N., M. A. KUZIORA,M. RECULSKI and W. MCGINNIS,1990 Hu- R. D. PALMITER,1989 Transgenic mice overexpressing the mouse man Hox-4.2 and Drosophila Deformed encode similar regulatory homoeoboxcontaining gene Hox-1.4 exhibit abnormal gut de- specificities in Drosophila embryos and larvae. Cell 63: 969-976. velopment. Nature 337: 464-467. MCGINNIS,W., and R. KRUMLAUF,1992 Homeobox genes and axial WOLPERT,L., 1969 Positional information and the spatial pattern of patterning. Cell 68: 283-302. . J. Theor. Biol. 25: 1-47. MCGINNIS, W., M. LEVINE,E. HAFEN. A. KUROWAand W. J. GEHRING, WRIGHT,c. V. E., K. W. Y. CHO,J. ~WICKE,R. H. COLLINSand E. M. 1984a A conserved DNA sequence found in homeotic genesof DE ROBERTIS,1989 Interference with function of a homeobox the Drosophila Antennapedla and Blthorax complexes. Nature gene in Xenopus embryos produces malformations of the ante- 308: 428-433. rior spinal cord. Cell 59: 81-93.