the molecular basis for metameric pattern in the ... · unravel the role played by each of these...

Development 101, 1-22 (1987)Printed in Great Britain © T h e Company of Biologists Limited 1987

Review Article

The molecular basis for metameric pattern in the Drosophila embryo

MICHAEL AKAM

Department of Genetics, Downing Street, Cambridge, CB2 3EH, UK

Summary

The metameric organization of the Drosophila em-bryo is generated in the first 5 h after fertilization. Aninitially rather simple pattern provides the foundationfor subsequent development and diversification of thesegmented part of the body. Many of the genes thatcontrol the formation of this pattern have been ident-ified and at least twenty have been cloned. By combin-ing the techniques of genetics, molecular biology andexperimental embryology, it is becoming possible tounravel the role played by each of these genes.

The repeating segment pattern is defined by thepersistent expression of engrailed and of other genesof the 'segment polarity' class. The establishment ofthis pattern is directed by a transient molecularprepattern that is generated in the blastoderm by theactivity of the 'pair-rule' genes.

Maternal determinants at the poles of the eggcoordinate this prepattern and define the anteropos-terior sequence of pattern elements. The primary

effect of these determinants is not known, but genesrequired for their production have been identified andthe product of one of these, bicoid is known to belocalized at the anterior of the egg. One early conse-quence of their activity is to define domains along theA-P axis within which a series of 'cardinal' genes aretranscribed.

The activity of the cardinal genes is required bothto coordinate the process of segmentation and todefine the early domains of homeotic gene expression.Further interactions between the homeotic genes andother classes of segmentation genes refine the initialestablishment of segment identities.

Key words: Drosophila melanogaster, segmentation,homeotic genes, pattern, in situ hybridization,embryogenesis.

Introduction

In this review, I discuss the patterns of gene ex-pression underlying the metameric organization ofthe early Drosophila embryo. I am concerned withtwo features of this embryonic pattern; the gener-ation of repeating units along the anteroposterior axisand the regional specification of different identitieswithin this array. These features first become appar-ent in the morphology of the animal 5 h or more afterfertilization, but the underlying molecular patternsare generated earlier, during the formation of theblastoderm and gastrulation.

This early period of development culminates in theelaboration of the 'segmented germ band', a morpho-logical stage that is strikingly conserved throughoutthe range of insects (Anderson, 1973). It is in thesegmented germ band that the general and, presum-ably, most ancient features of the insect body plan aredisplayed most clearly, and that the homology of

metameric units is most apparent. As we shall seebelow, this clarity extends to a molecular level. Genescontrolling segmentation and segment identity estab-lish a metameric pattern of activity in the early germband embryo. Only vestiges of this relatively simplepattern can be discerned in the patterns of geneactivity in later development.

The study of this process has been made posssibleby the identification of mutations in approximately 50genes (see Table 1). These disrupt the pattern ofmetameric units (the segmentation genes), the ident-ity of segments (the homeotic genes), or both (e.g.mutations affecting the polarity of the whole em-bryo). Virtually all were first identified by the pheno-types of mutant alleles. Many of the homeotic mu-tations were identified fortuitously, by strikingtransformations of one segment into another (e.g.Bridges & Morgan, 1923; Lewis, 1963). The greatmajority of segmentation genes, however, have beenidentified by systematic, not to say Herculean,

M. Akam

Table 1. Segmentation and homeotic genes in Drosophila

Maternalor

Gene name zygotic Classification(symbol, locus') expression by phenotype

References

MolecularEmbryonic cloning and Pattern ofphenotype structure expression Comments

COORDINATE AND GAP GENES

mbicaudal{bic, 11-67)

Bicaudal-D(Bic-D, 11-52.9)

Bicaudal-C(Bic-C, 11-51)

dicephalic(die, 111-46)

"bicoid(bed, ANT-C)

exuperantia(exu, 11-93)

swallow(sww, 1-14)

torso(tor, 11-57)

trunk(trk, 11-36)

fs(l)Nasrat(fs(l)N, 1-0.0)

oskar(osk, III-48.5)

staufen(stau, 11-83)

tudor(tud, 11-90)

valois(vb, 11-53)

vasa(vas, 11-51)

"hunchback(hb. III-4S)

"Kruppel(Kr, 11-107.6)

knirps(kni, 111-47)

giant(g<- 1-1)

tailless(tit. III-102)

unpaired(upd, 1-59)

hopscotch(KDhop)

m

embryonic polarity 2, 11, 99reversal

2, 12

2, 12

13

anterior pattern defects 9

6

2, 100

anterior and posterior 2, 6pattern defects

10 10

m

m

m

m

m

m

m

m, z

z

z

z

z

z

m, z

"

'posterior group' gapgene

•'

>•

,,

segment gap

Segment gap/irregulardefect

Segment gap/irregulardefect

6

98

7

6

2, 6, 8

6

6

1,2,4, 15,34

1,2,3, 18,19,20

1,2,4

5,57,%

4, 24

5,57

97

—

—

—

—

—

—

—

17

16,21,22

—

—

—

—

—

—

—

—

—

—

—

—

16,17

16,23

—

—

—

—

—

Double-abdomen phenotype, but no duplication ofpole cells. Variable penetrance.

Dominant 'gain of function' phenotype like bicaudal

Dominant haplo-insufficient phenotype like bicaudal

Variably produces bicaudal and dicephalicphenotypes; alters location of nurse cells withinfollicle

Null alleles delete head, gnathal and thoracic regions:telson but not abdomen or pole cells duplicatedanteriorly. See text.

Mutants disrupt anterior head structures; posteriormidgut and proctodaeum duplicated anteriorly.

Phenotype similar to exu; also called fs( 1)1502 ofGans (14)

Mutants disrupt anterior head structures; delete mostposterior derivatives of the fate map (not polecells)

Similar to torso

Maternal phenotype similar to torso;

Segment gap like knirps; pole cells not formed Seetext.

Similar to oskar; also shows head defects and headfold shifted anteriorly.

Similar to oskar; weak alleles affect only pole cells.No null alleles known.

Similar to oskar; cellularization often incomplete.

Similar to oskar

'Cardinal gene'; see Fig. 5; maternal expressionpartially rescues mutant phenotype, but is notrequired for normal segmentation; allelic with Rg(pbx)

'Cardinal gene"; see Fig. 5; mutants also eliminateMalpighian tubules

'Cardinal gene'? see Fig. 5

Mutants disrupt head and A4-8

Zygotic mutant phenotype similar to maternal effectof torso, but limited to ectoderm.

Mutants disrupt principally T2 and A5

Maternal or zygotic function sufficient for normalsegmentation Absence disrupts T2, T3, A4. A5.A8.

Metameric pattern in Drosophila

Table 1. Continued

ReferencesMaternal

orGene name zygotic Classification(symbol, locus") expression by phenotype

MolecularEmbryonic cloning and Pattern ofphenotype structure expression Comments

PAIR-RULE AND SEGMENT POLARITY GENES

z pair rule 1,2,3,33 25,27 25,26,27"even-skipped(eve, 11-55)

"hairy(h, 111-27)

runt(run, 1-65)

"fushi-tarazu(fiz, ANT-C)

"paired(prd, 11-45)

odd-paired(opa, 111-48)

odd-skipped(odd, 11-08)

sloppy-paired(sip, 11-08)

"engrailed(en, 11-62)

"gooseberry(gsb, 11-104)

"wingless(wg, 11-30)

armadillo(arm, 1-1)

cubitus-interruptus(ciD, IV-0)

fused(Ju, 1-59.5)

hedgehog(hh, 111-90)

naked(nkd, 111-47)

patched(ptc, 11-59)

dishevelled(I(I)dsh, 1-34)

m, z

m, z

1,2,4,28 29,30 31

1,2,5,35, — —36

2, 4, 37 38, 39, 42 40, 41, 45

1,2,3,33 43 43

2,4

1,2,3,33 —

2,3

pair rule/segment 1,2,3,46 46,47,48 45,48,49,polarity 50, 51

m, z

segment polarity

»

»

1,2,3,55, 5259

1,2,3,53, 5455

2,5,55,56, —57

1,2,59 —

1,2,55,57, —58

1,2,4,59 —

4 —

3 —

55 —

52

54

—

—

—

—

—

Null mutants eliminate all segmental periodicity; genecontains homeobox. See Fig. 4.

Pattern deletions out of frame with, but overlap, those offtz. See Fig. 4; function also required for adult bristlepattern

Pattern deletions cover more than full metamere; mutantalso shows segment polarity reversals

Pattern deletions correspond to even numberedparasegments See Fig. 4.Gene contains Antp-\iV.& homeobox.

Pattern deletions similar to hairy.Gene contains homeobox, and sequences homologousto bicoid and gooseberry. See Fig. 4.

Pattern deletion complementary to paired

Pattern deletions cover less than full metamere, alsoshows segment polarity reversals

Mutant shows variable pair-rule fusions and segmentpolarity reversals.Expression defines P compartments

Posterior compartments replaced by anteriorduplications. Phenotype defined only by deficiencies ofat least two genes

Null mutants abolish periodicity of cuticular pattern

Polarity phenotype similar to gooseberry, autonomous inclones. Lack of maternal function causes egg shapedefect.

Polarity phenotype of null allele similar to gooseberry;CiD shows gain of function phenotype.Allelicwithif4J«

Polarity phenotype similar to gooseberry; maternal orzygotic expression sufficient for normal segmentation

Null mutant phenotype like wingless, but head normal.

Eliminates denticle belts

Null mutants show duplication of segment boundaryregion in reversed polarity.

Maternal or zygotic function sufficient for normalsegmentation. Absence of function results inphenotype like wingless

M. Akam

Table 1. Continued

References

Gene name(symbol, locus")

Maternal Classificationor by phenotype

zygotic affected segmentsexpression (parasegments)

Embryonicphenotype

HOMEOT1C GENES SHOWING REGION SPECIFIC EXPRESSION11

proboscipedia(pb, ANT-C)

Deformed(Dfd, ANT-C)

Sex combs reduced(Scr, ANT-C)

Anlennapedia{Antp, ANT-C)

Ullrabithorax(Ubx, BX-C)

Abdominal-A(abd-A, BX-C)

Abdominal-B(Abd-B, BX-C)

caudal{cad, II-38E)

F90-2(Ant-C)

z (adult transformation)

z Segment transformationMandible/maxilla (PS 0/1)

z Segment transformationLabial/Tl (PS 2/3)

z Segment transformationT1-T3 (PS 3-5)

z Segment transformationT3-A7 (PS 5-13)

z Segment transformationA2-A8 (PS 7-14)

z Segment transformationA5-A8/9 (PS 10-14)

m, z Loss/transformationA10/11 (PS 15) or telson

z Uncertain

37

60, 61

4, 37, 66

4, 37, 66, 69

66, 79, 80

79, 87, 88, 89

79, 87, 88, 91

92

—

Molecularcloning and

structure

—

61,62

67

70-73

81-83

90

90

93, 94

95

Pattern ofexpression

—

63, 64, 65

57, 65, 67, 68

74-78

84-86

68

63, 68, 95

92

95

Comments

Mutants show no detectable embryonicphenotype

Equivalent to iab-2 or iab-2/iab-4 regionof BX-C

Equivalent to iab-7 or iab-5/tab-8 regionof BX-C

Absence of zygotic function alone affectsanal pads; absence of both maternaland zygotic function also variablydisrupts segmentation

Mutant phenotype not described, butprobably corresponds to labial (101).Expressed anterior to Dfd and posteriorto Abd-B

Genes have been grouped somewhat arbitrarily according to the classification of their phenotypes. Within each group, genes for which molecularinformation is available have been placed first. These are marked with an asterisk.

'Gene loci are shown by chromosome (Roman numeral) and genetic map position, except for genes in the ANT-C (III, 47.5), the BX-C (III. 58.8)or, in the case of caudal, where only chromosome hybridization data are available.

bOnly the homeotic genes of the ANT-C and BX-C are listed, together with one other, caudal, for which molecular and mutant data are available.All of these genes contain a homeobox, and all are known to be expressed at restricted locations along the AP axis. (The inclusion of proboscipedia isbased on the unpublished data of M. Pulz and T. Kaufman). Other genes mutate to homeotic phenotypes, but no others are clearly analogous to the'segment selector' genes of the ANT-C and BX-C.

Key to references

(1) Niisslein-Volhard & Wieschaus, 1980; (2) Nusslein-Volhard et al. 1982; (3) Nusslein-Volhard el al. 1984; (4) Jurgens et al. 1984; (5) Wieschaus ei al1984a; (6) Schupbach & Wieschaus, 1986; (7) Lehmann & Nusslein-Volhard. 1986; (8) Boswell & Mahowald, 1985; (9) Frohnh6fer & Nusslein-Volhard.1986; (10) Frigerio et al. 1986; (11) Bull, 1966; (12) Mohler & Wieschaus, 1986; (13) Lohs-Schardin, 1982; (14) Gans et al. 1975; (15) Lehmann &Nusslein-Volhard, 1987; (16) Jackie et al. 1986; (17) Tautz et al. 1987; (18) Gloor, 1950; (19) Wieschaus et al. 1984b; (20) Seifert et al. 1986; (21) Preisset al. 1985; (22) Rosenberg et al. 1986; (23) Knipple et al. 1985; (24) Strecker et al. 1986; (25) Macdonald et al 1986; (26) Harding et al. 1986; (27) Fraschet al. 1987; (28) Ingham et al. 1985c; (29) Holmgren, 1984; (30) Ish-Horowicz et al. 1985; (31) Ingham et al. 1985c; (33) Nusslein-Volhard et al. 1985;(34) Bender et al. 1987; (35) Gergen & Wieschaus, 1985; (36) Gergen & Wieschaus, 1986a; (37) Wakimoto & Kaufman, 1981; (38) Laughan & Scott.1984; (39) Kuroiwa et al. 1984; (40) Hafen et al. 1984a; (41) Carroll & Scott, 1985; (42) Weiner et al. 1984; (43) Kilcherr et al. 1986; (45) Weir &Kornberg, 1985; (46) Komberg, 1981; (47) Poole et al. 1985; (48) Fjose et al. 1985; (49) Kornberg et al. 1985; (50) Ingham et al. 19856; (51) Dinardo et al.1985; (52) Bopp et al. 1986; (53) Baker, 1987a; (54) Baker, 1987b; (55) Perrimon & Mahowald, 1987; (56) Wieschaus & Riggleman, 1987; (57) Gergen &Wieschaus, 1986ft; (58) Martinez-Arias, 1985; (59) Martinez-Arias & Ingham, 1985; (60) Kaufman, 1983; (61) Regulski et al. 1987; (62) Regulski et al1985; (63) McGinnis et al. 1984; (64) Chadwick & McGinnis, 1987; (65) Martinez-Arias et al. 1987a; (66) Struhl, 1983; (67) Kuroiwa et al. 1985;(68) Harding et al. 1985; (69) Kaufman & Abbott, 1984; (70) Garber et al. 1983; (71) Scott et al. 1983; (72) Schneuwly et al 1986; (73) Laughon et al.1986; (74) Hafen et al. 1983; (75) Levine et al. 1983; (76) Martinez-Anas. 1986; (78) Carroll el al. 1986a; (79) Lewis, 1978; (80) Lewis, 1981; (81) Benderet al. 1983; (82) Hogness et al. 1985; (83) Weinzierl et al. 1987; (84) Akam & Martinez-Arias, 1985; (85) White & Wilcox, 1985; (86) Beachy et al. 1985;(87) Sanchez-Herrero et al. 1985; (88) Tiong et al. 1985; (89) Morata et al. 1983; (90) Karch el al. 1985; (91) Casanova et al. 1986; (92) Macdonald &Struhl, 1986; (93) Mlodzik et al. 1985; (94) Mlodzik & Gehring, 1987; (95) Hoey et al. 1986; (96) Petschek et al. 1987; (97) Perrimon & Mahowald. 1987:(98) Degelman et al. 1986; (99) Nusslein-Volhard, 1977; (100) Zalokar, M. et al. 1975; (101) Kaufman, T. C. (1983); R. Diederich and T. Kaufman,personal communication.


searches for mutations affecting the pattern ofstructures in the cuticle of dying embryos (Niisslein-Volhard & Wieschaus, 1980; Wieschaus, Niisslein-Volhard & Jurgens, 1984a; Jiirgens, Wieschaus,Nusslein-Volhard & Kluding, 1984; Nusslein-Vol-hard, Wieschaus & Kluding, 1984b; Perrimon, Eng-strom & Mahowald, 1984; Schupbach & Wieschaus,1986).

Mutations in the segmentation genes have a varietyof effects, which fall into three broad classes (Nuss-lein-Volhard & Wieschaus, 1980; Nusslein-Volhard,Weischaus & Jurgens, 1982). Mutations in the 'gapgenes' cause embryos to develop with large gaps inthe array of segments. Mutations in the 'pair-rule'genes cause embryos to develop with only half thenormal number of segments, and the characteristicsof the remaining segments suggest that elements ofeach alternate segment have been deleted. Mutationsin the 'segment polarity' genes affect the sequence ofpattern elements within segments, resulting in exten-sive pattern deletions, duplications and reversals ofpolarity within each segment.

engrailed as a marker for segmentationThe segmentation gene engrailed has played a uniquerole in our understanding of metameric pattern inDrosophila. It was first identified by the effects onadult flies of what we now recognize to be a ratherunusual mutant allele. This mutation allows adultsurvival, but in several segments it results in thepartial replacement of posterior structures by an-terior ones (Tokunaga, 1962; Garcia-Bellido & Santa-maria, 1972). Morata & Lawrence (1975) showed bygenetic analysis that engrailed was required to main-tain the lineage boundary that separates, and defines,the anterior and posterior compartments within eachsegment primordium of the developing adult. Sub-sequently, more typical engrailed mutations werefound to display an embryonic segmentation defect,resulting both in the fusion of segments and inalterations of segment polarity (Kornberg, 1981).

engrailed and many other of these genes have nowbeen cloned. In some cases, antibodies are availableto detect the proteins that they encode. Thus, it ispossible to examine the process of pattern generationwith molecular probes. The result has been a remark-able triumph for developmental genetics. Virtually allof the genes identified by mutant phenotypes in lateembryos or adults have proved to be involved in theearly processes of embryonic patterning; many ap-pear to play key roles. Expression of the engrailedgene, in particular, has provided a molecular markerto visualize the metameric pattern of developingembryos as it is first established.

The metameric organization of the early germband

I shall use the term germ band to refer to the wholeregion of the developing egg that will give rise to theembryo proper (Anderson, 1973). In Drosophila, cellsfated to form the germ band cover most of the surfaceof the blastoderm and about two-thirds of these willgive rise to the metameric or segmented part of thebody (Poulson, 1950; Campos-Ortega & Hartenstein,1985; Technau, 1987). The remaining cells generatethe procephalon, or head, the endoderm and theamnioserosa, an extraembryonic membrane (Fig. 1).

The characteristic morphology of the early germband is generated from the blastoderm by the pro-cesses of gastrulation, without any cell division(Fig. 2). At first, the region that will give rise to thesegmented part of the body appears as a uniformdouble-layered structure, mesoderm within and ecto-derm without. The first morphological sign of seg-mentation appears in the mesoderm, as a series ofrepeated thickenings. Shortly thereafter, grooves ap-pear in the outer surface of the ectoderm (Poulson,1950; Turner & Mahowald, 1977). These grooves donot demarcate the future segments, but parasegments(Martinez-Arias & Lawrence, 1985), metameric unitswhich include cells in the posterior part of onesegment and the anterior part of the next. 14 paraseg-ments are clearly visible, though at least some regionsof parasegments 0 and 15 are probably specified in theinitial pattern. Cells in the first four parasegmentsrapidly rearrange to generate the mandibular, maxil-lary and labial lobes of the mouthparts. The followingthoracic and abdominal parasegments remain verysimilar in structural organization until much later inembryogenesis (Turner & Mahowald, 1979; Pets-chek, Perrimon & Mahowald, 1987).

The metameric organization of the germ band isdisplayed most beautifully by the pattern of ex-pression of the engrailed gene. In situ hybridization(Kornberg, Siden, O'Farrell & Simon, 1985) andantibody staining (DiNardo, Kuner, Theis & O'Far-rell, 1985) reveal that the engrailed gene product isexpressed in 15 evenly spaced rows of cells within thegerm band. These rows, each only two or three cellswide, lie immediately posterior to the grooves thatdefine each parasegment (Ingham, Martinez-Arias,Lawrence & Howard, 1985b), suggesting that theparasegments are identical to the lineage units de-fined by a P-A pair of compartments (Fig. 3).

Metamerism of the germ band is also apparent inthe pattern of expression of homeotic genes and, hereagain, parasegments appear to be the metameric unitsdefined by the earliest patterns of gene expression. Atleast seven homeotic selector genes, four in theAntennapedia complex (ANT-C; Kaufman, 1983)

M. Akam

ant4 8 12 cellsI I I

100 ̂ m

Fig. 1. A fate map of the Drosophila blastoderm (from Campos-Ortega & Hartenstein, 1985). The shape is aplanimetric reconstruction of the blastoderm surface. All parts of the egg surface contribute to the embryo proper,except the narrow dorsal primordium for the amnioserosa (as). Hatched areas will invaginate at gastrulation. Cells thatwill generate metameric structures are enclosed by a thick line. Abbreviations: amg, anterior midgut; as, amnioserosa;cl, clypeolabrum; dEpi, dorsal epidermis; dr, dorsal ridge; es, oesophagus; mt, Malpighian tubules; MS, mesoderm;ol, optic lobe; ph, pharynx; pmg, posterior midgut; pNR, procephalic neurogenic region;pr, proctodeum; sg, salivary gland; vNR, ventral neurogenic region; M, mandibular segment; Mx, maxillary segment;La, labial segment; T1-T3, thoracic segments; A1-A10, abdominal segments; ant, anterior; dors, dorsal.

Fig. 2. Early development of Drosophila melanogaster.Diagrams on the left show the morphology of stagesduring early development. (Timings are given as hours at25CC after fertilization.) Corresponding panels on theright illustrate patterns of gene activity established ateach of these stages. (Numbers locate the anterior ofparasegments.)

(A) Cleavage stage 7-8 of Foe & Alberts (1983).Nuclei are shown migrating to the periphery of the egg.Cytoplasmic transplantation experiments show thatdeterminants are localized at the poles of the egg.Crosses - location of maternally derived transcripts fromthe bicoid gene; dots - location of polar granules and ofposterior determinants dependent on the oskar groupgenes. (The polar granules subsequently segregate to thepole cells.) Maternal transcripts of the genes hunchback(pink) and caudal (not shown) are initially uniformlydistributed, but graded distributions are establishedduring early cleavage.

(B) Syncytial blastoderm (cleavage stage 12). Mostnuclei reach the perimeter of the egg and becometranscriptionally active. Pole cells have formed. Yolksegregates to the middle of the egg, leaving a peripherallayer of clear cytoplasm. Localized transcription of thecardinal genes hunchback and then Kruppel (brown) isestablished from the zygotic genome. The transcriptionalpatterns of these genes become more elaborate at laterstages (not shown); both are subsequently expressed in anadditional zone at the posterior of the cellular blastoderm.

(C) Cellular blastoderm (cleavage stage 14A/B;embryonic stage 5 of Campos-Ortega & Hartenstein,1985). Cell membranes are extending down from the eggcortex, but cells are not yet closed. The activity of

maternal determinants falls rapidly. Expression of thepair-rule genes, first apparent at about cleavage stage 13,resolves into a well-defined pattern of overlapping stripesshowing double segment periodicity, (fushi-tarazudomains - yellow/green; even-skipped domains,purple/pink; each of these domains is shown assubdivided by the superimposed expression of one othergene (blue) but the actual sequence of cell states definedby all of the pair-rule genes is more complex. See Fig. 4.)

(D) Gastrulation (embryonic stage 7). Mesoderminvaginates ventrally; anterior and posterior midgutinvaginations form the endoderm. Pole cells lie within theposterior midgut invagination and are carried dorsallyand forward by the beginning of germ band elongation.The transient cephalic furrow demarcates an approximateboundary between head and body regions. Cellsexpressing high levels of engrailed protein (orange) beginto define the definitive metameric repeat. Expression ofthe pair-rule genes continues to evolve (not shown).

(E) After germ band extension (embryonic stage 10).The metameric region forms a uniform double-layeredstructure extending around the posterior pole of the eggand the most posterior segments are apposed to the head.Cell division is underway, most blastoderm cellsundergoing two or three rounds of mitosis beforedifferentiation. Cells expressing engrailed (orange) andwingless (gTeen) define the definitive metameric patternand flank presumptive parasegment borders. Products ofthe pair-rule genes are decaying rapidly.

References for the patterns of gene expression arelisted in Table 1. Abbreviations as Fig. 1 andect, ectoderm; yk, yolk; yn, yolk nuclei; pp, polar plasm;pc, pole cells; cf, cephalic furrow; st, stomodeum.

Morphology

A. Nuclear migration (1-25h. ~ 128 nuclei)

Gene activity

B. Syncytial blastoderm (2h. M5(X) nuclei)

C. Cellular blastoderm (2-5h, ~5000cells)

D. Gastrulation (3 h, -5000 cells)

amg

E. Extended germ band (4-5 h, >5000 cells).

4 6 8

14 12

amg

dors

ant

A

Cel

lula

r bl

asto

derm

Segm

ents

Par

aseg

men

ts

Def

orm

ed

Sex

com

bs r

educ

ed

Ant

enna

pedi

a P

I

Ant

enna

pedi

a P

2

Ull

rabi

thor

ax

abdo

min

al-A

Abd

omin

al-B

caud

al

1 xT

3^A

l tt lA2 i

A

3 t A

4 t A

5 tA

6 % A

7 % A

S A

9/10

11

' 2

' 3

' 4

' 5

' 6

' 7

8 '

9 '

10

' 11

' 121 1

31 14

'

Ext

ende

d ge

rm b

and

W//

////

////

////

////

A

14 13 12 11 10

3 4

B. e

ngra

iled

15 14 13 12 11

10 V2

3 "

••

;-

5 h

C-

Ultr

abith

orax

12

11

10

JD.

Def

orm

ed

14

13

••

/

'/

••

;'

;'

,.

8 M. Akam

and three in the bithorax complex (BX-C; Lewis,1978) are expressed in specific regions of the germband (Fig. 3; see also Harding, Wedeen, McGinnis &Levine, 1985). Ubiquitous and high level expressionof Sex combs reduced (Scr) defines parasegment 2(Kuroiwa, Kloter, Baumgartner & Gehring, 1985;Martinez-Arias, Ingham, Scott & Akam, 1987a), ofAntennapedia (Antp), parasegment 4 (Carroll et al.1986a; Martinez-Arias, 1986; ) and of Ultrabithorax(Ubx), parasegment 6 (Akam & Martinez-Arias,1985; White & Wilcox, 1985). Deformed (Dfd) isexpressed in the region immediately anterior toparasegment 2, including parasegment 1 but includ-ing also more anterior structures (Chadwick &McGinnis, 1987; Martinez-Arias et al. 1987a). Otherparasegments are characterized by the expression ofcombinations of homeotic genes.

As the germ band develops, this metameric patternof homeotic gene expression is rapidly obscured. Cellmovements disrupt the initial patterns (DiNardo et al.1985). Within cells of a single parasegment, theexpression of each homeotic gene is modulated ac-cording to position and cell type and, in somestructures, homeotic genes are activated in domainsthat are not obviously related to the metamericspecification of the animal (White & Wilcox, 1985;Akam & Martinez-Arias, 1985; Carroll et al. 1986a;Martinez-Arias, 1986). It is probably an oversimplifi-cation to suggest that the homeotic selector genessimply define the identity of metameres at any stagein development, but their observed patterns of ex-pression fit this model most closely in the early germband stages.

A scheme for pattern formation along the A-Paxis

A picture is emerging of the processes that generatethe metameric pattern in the germ band (Fig. 2).There is no evidence for any periodic pattern pre-formed in the egg at the time of fertilization. The onlylocalized maternal determinants are probably at thepoles of the egg (Sander, 1984; Frohnhofer, Lehmann& Niisslein-Volhard, 1986; Frohnhofer & Niisslein-Volhard, 1986; Lehmann & Niisslein-Volhard, 1986).During the early stages of cleavage, these determi-nants coordinate pattern and polarity throughout theembryo. The earliest signs of this activity are thegeneration of a gradient in the distribution of at leastone maternally encoded product (Mlodzik, Fjose &Gehring, 1985; Macdonald & Struhl, 1986) and thelocalized activation of a small set of genes, (thesegment gap genes) at defined positions along theA-P axis of the egg (Knipple et al. 1985; Tautz et al.1987). Cross regulatory interactions between thesegenes sharpen the boundaries between the regions

within which each is active (Jackie et al. 1986) and so arelatively precise set of spatial domains are estab-lished along the axis of the egg. I shall refer to theseas the cardinal domains, a term suggested by Mein-hardt (1986; see below).

At almost the same time, another set of genes (thepair-rule genes) are activated within the segmentedbody region. Initially the transcription of these genesis either uniform, or simply localized, but complexperiodic patterns rapidly evolve (Hafen, Kuroiwa &Gehring, 1984a; Ingham, Howard & Ish-Horowicz,1985a; Carroll & Scott, 1985; Harding et al. 1986;Macdonald, Ingham & Struhl, 1986; Kilcherr et al.1986). These patterns characteristically exhibit astage showing a double segment periodicity. Localorder in this evolving pattern probably depends onsome reaction diffusion mechanism, but long-rangeorder requires an interaction between the productsof the genes that define cardinal domains and thepair-rule genes or their products (Carroll & Scott,1986; Carroll, Winslow, Schiipbach & Scott, 19866;Ingham, Ish-Horowicz & Howard, 1986).

The pattern generated by the gap and pair-rulegenes is transient and can appropriately be describedas a prepattern (Stern, 1968). It is elaborated duringcellularization of the blastoderm and decays rapidlyduring gastrulation and formation of the germ band.It serves, however, to establish a series of qualitat-ively different cell states (the 'singularities' of Stern)that define the patterns of expression of engrailed(Howard & Ingham, 1986) and of other genes thatmediate the definitive segment pattern.

Recent results suggest that the same cardinaldomains serve to define the initial regions withinwhich each homeotic gene is potentially active (White& Lehmann, 1986; Martinez-Arias and M. Akam,unpublished results). Subsequently, the evolving pat-terns of homeotic gene expression reflect their inter-actions with both the pair-rule generated segmentalprepattern (Ingham & Martinez-Arias, 1986; Dun-can, 1986), with each other (Hafen, Levine &Gehring, 19846; Struhl & White, 1985), and withengrailed and other persistent components of thesegmental pattern (Martinez-Arias & White, 1987).

In the following sections I elaborate on variousaspects of this scheme. I progress backwards indevelopmental time, because the analysis of mu-tations affecting each developmental stage relies onthe interpretation of altered patterns appearing laterin development.

The appearance of the definitive segmentpattern

The expression of the engrailed gene provides amolecular assay for the developing segment pattern.

By the time the germ band is fully extended, theengrailed stripes are remarkably similar throughoutthe segmented region of the body (Fig. 3), but theydo not all appear simultaneously. The first cells toexpress high levels of engrailed protein do so just asgastrulation is starting (DiNardo et al. 1985). Thesecells lie ventrally, in the region of the future maxillarysegment. They will come to form the engrailed stripeof parasegment 2. In the next 30min, the patternspreads dorsally and posteriorly. A similar timecourse for the evolution of pattern has been observedfor all of the genes that stripe in the blastoderm, andprobably indicates that equivalent molecular de-cisions are taken at slightly different times in differentregions of the embryo. As we shall see below, eventsleading to the activation of engrailed in odd- and ineven-numbered parasegments are not strictly equival-ent, but depend on different elements of the prepat-tern. Probably for this reason, even-numbered stripesappear before the odd ones, with both sets showingthe same progressive spread (Weir & Kornberg,1985).

After the germ band has extended, engrailed pro-tein begins to accumulate in discrete regions of thehead and in parts of the hindgut. The engrailed genemay therefore play a role in the development ofpattern in these parts of the body, but obvious signs ofmetamerism are now obscured, even at the molecularlevel. The regions of engrailed expression are patches,not stripes, even when they are first apparent; theyappear later than in the metameric region and theyappear in regions where the underlying molecularprepattern is quite different from that in the meta-meric region (DiNardo et al. 1985).

It is clear that genes other than engrailed arenecessary from the time of gastrulation onwards toestablish or maintain normal pattern within eachsegment. The most obvious candidates are the knownsegment polarity genes (Table 1). Only two of thesegenes have yet been isolated; wingless (Baker,1987a,/)) and gooseberry (Bopp et al. 1986; Cot6 et al.1987). Both of these show a pattern of expressioncomparable with that of engrailed. Wingless is ex-pressed from gastrulation until late in embryogenesis.In the early germ band embryo, wingless transcriptsaccumulate in a narrow stripe of cells at the posteriormargin of each parasegment (Fig. 2). Thus the wing-less and engrailed stripes presumably abut and definethe cells either side of each parasegment boundary(Baker, 1987a,b). The putative gooseberry transcriptsare also expressed in narrow stripes from gastrulationonwards, but the registration of these with theengrailed stripes has not been determined. It remainsto be seen how many different regions within eachsegment are similarly defined in the early germ bandby the activity of specific genes.

Metameric pattern in Drosophila 9

The prepattern underlying segmentation

Detailed observations of the way in which theengrailed stripes first appear suggest that their lo-cation must be controlled by a prepattern. Alongeach of the presumptive stripes, which initially will beonly one or at most two cells across, individual cellsaccumulate relatively high levels of engrailed protein.Neighbouring cells along the stripe may transientlyshow very different levels of engrailed protein andcells between stripes show none at all. This suggeststhat each cell turns the engrailed gene on indepen-dently and that the precise timing of this event isa stochastic response to an underlying prepattern(DiNardo et al. 1985).

The products of the pair-rule genes are directlyimplicated in the generation of this prepattern. Mu-tation in any one of them alters the pattern ofengrailed expression in the extended germ band,(Howard & Ingham, 1986; Martinez-Arias & White,1987; S. DiNardo and P. H. O'Farrell, personalcommunication; see Fig. 4) implying that all of themdisrupt the initial metameric organization of theembryo.

Many of the pair-rule genes have now been clonedand the patterns of expression of four have beendescribed (see Table 1). In all of these cases, thegenes are first transcribed in the syncytial blastoderm,at or very soon after the time that transcription of thezygotes own genes is first activated (Edgar & Schu-biger, 1986) and about 40min before the engrailedpattern first appears. A striped pattern of pair-ruletranscript distribution evolves while the cellularblastoderm is forming. Characteristically, the patternat its climax in the cellular blastoderm is one of sevenannuli or stripes of transcript accumulation encirclingthe embryo. These are uniform in size, except at theends, and spaced at two-segment intervals through-out the segmented body region (Figs 2, 4).

At gastrulation, the cells of the blastoderm arefinally closed off from the yolk syncytium (Rickoll,1976). Within these cells, the distributions of the pair-rule gene products define a series of different cellstates in a repeating pattern along the A-P axis. Mostcomponents of this pattern have a two-segmentperiodicity. However, during and after gastrulationthe patterns of pair-rule gene expression continue toevolve. In some cases, a pattern with single-segmentperiodicity appears, either by the splitting of existingstripes {paired, Kilcherr et al. 1986) or by the inter-calation of additional ones {even-skipped, eve, Mac-Donald et al. 1986). Finally, by the time the germband is fully extended, transcripts of the pair-rulegenes are virtually undetectable in the metamericregion. At least in the case offushi-tarazu {ftz, read as

10 M. Akam

futz) and eve, the protein products are rapidly disap-pearing (Carroll & Scott, 1985; Frasch et al. 1987).Some of the pair-rule genes are expressed again as thenervous system develops, but this later expressionshows no trace of a pair-rule repeat (DiNardo et al.1985), and is under quite different regulation from theearly expression in the blastoderm (Hiromi, Kuroiwa& Gehring, 1985).

It is not yet clear how the transient blastodermprepattern regulates the activity of engrailed. Elimi-nating the product of any pair-rule gene alters thepattern of engrailed expression, but these effects maybe direct or indirect, for the pair-rule genes them-selves interact (see below). In some cases, the effectsof pair-rule mutations on engrailed expression areconceptually simple, suggesting a fairly direct interac-tion. For example, in mutant embryos that lack afunctional ftz gene, those stripes of engrailed ex-pression that normally overlie a ftz stripe are absent,but the alternate stripes persist. In paired mutants,only these alternate stripes persist. In other cases,however, more complicated pattern alterations re-sult (Fig. 4). It seems likely that both hierarchical(Ingham & Howard, 1985; Ingham & Martinez-Arias,1987) and combinatorial (O'Farrell & Scott, 1986;Gergen, Coulter & Wieschaus, 1986) interactionsdefine cell states that activate the engrailed gene.

Generation of the blastoderm prepatternWhen the pair-rule genes are first transcribed, theirexpression shows no trace of periodic patterning.hairy is expressed uniformly throughout virtuallythe whole egg (Ingham, Howard & Ish-Horowicz,1985a); ftz and eve throughout the segmented bodyregion only, and paired within a localized regionaround the cephalic furrow (the centre from whichthe segmental pattern spreads). This suggests that theperiodicity of the segment pattern is generated afterthe genes are initially activated and raises the possi-bility that the system of pair-rule genes and theirproducts constitute a system that generates pattern denovo, using a mechanism analogous to that proposedby Turing (1952), in which reaction and diffusioncoefficients define many properties of the pattern (seeMeinhardt, 1982). In such a system, it would be theaccumulating products of the pair-rule genes them-selves that regulate the activity of other members ofthe set.

In the case of one of the pair-rule genes, ftz, weknow that regulation must take place at the level oftranscription or RNA processing; protein degra-dation and transport can at most play a secondaryrole, for when the ftz promoter and untranslatedleader sequences are fused to heterologous protein-coding sequences (e.g. bacterial /3-galactosidase) the

resulting RNA and protein distributions are indis-tinguishable from those of the ftz products themselves(Hiromi et al. 1985).

One prerequisite for a reaction-diffusion model isthat the products of the pair-rule genes should have alife time that is short in comparison to the timescale ofpattern evolution. This appears to be the case. In thepresence of or-amanitin, ftz RNA in the blastodermhas a half-life of about lOmin (Edgar, Weir, Schu-biger & Kornberg, 1986). Moreover, in the syncytialblastoderm, the proteins that regulate ftz transcrip-tion also have a very short lifetime; injection ofcycloheximide into eggs just as the ftz stripes areforming results in the re-establishment of uniformtranscription. However, when the stripes are wellestablished in the cellular blastoderm, the transcrip-tion pattern becomes stable to cycloheximide injec-tion (Edgar et al. 1986).

Fig. 4. Expression of segmentation genes in theDrosophila blastoderm.

(A) Embryo at cleavage stage 14A/B hybridized witha 3H-labelled probe for transcripts of the pair-rule genefushi-tarazu. The autoradiograph maps the distribution offtz transcripts within the embryo. Compare Fig. 2C. Theenlarged panel at left shows the location of nuclei (nuc),lying peripheral to the bulk of the cytoplasm (cyt).Hybridization lies over the cortex of the egg, not over thedeeper cytoplasm. Bar, 50pm; for enlargement, 10/im.

(B) Approximate registration of pair-rule stripes,engrailed expression and metameric units. The patterns ofexpression are shown for four of the pair-rule genes atabout the same stage as the embryo shown above. Thelater patterns of engrailed expression have been projectedonto the same diagram, even though at this mid-blastoderm stage hybridization reveals only a single well-defined engrailed stripe (stripe 2).

The bands of engrailed expression define Pcompartments and so lie at the anterior margin of eachparasegment. Stripes of even-skipped and fushi-tarazuexpression are each approximately four cells wide at midblastoderm, and appear to lie out of phase with eachother. Double-labelling experiments in later embryossuggest that the anterior margins of both/?z and evestripes coincide precisely with the engrailed stripes, andhence define parasegment boundaries (Lawrence,Johnston, Macdonald & Struhl, 1987). Hairy stripes areabout the same width, but are displaced slightly withrespect to parasegments and overlap those of ftz. Pairedstripes are broader than a single metameric repeat, butthe seven stripes split into fourteen before gastrulation.

The bottom part of the figure shows the effect that nullmutations in each of the same four pair-rule genes haveon the subsequent expression of engrailed. In paired andftz mutant embryos, it appears that alternate engrailedstripes have simply been deleted. (Autoradiographprovided by P. Ingham; see Ingham et al. 1985a forfurther details. Other data taken from references inTable 1.)

% Egg Length

Segments

Parasegments

hairy

even-skipped

pairedGeneexpressioninnormalembryos

engrailedexpressioninpair-rulemutants

fushi-turazu

engrailed(later)

hairy

even-skipped

paired'

fushi-larazu~

Mn Mx .La 77 72 T3 Al A2 A3 A4 A5 A6 . A7 . AS A9/10

1 2 3 4 : 5 ; 6 : 7 8 9 10 11 12 13 14

I I I I I

I II I

LZ1 • • • •

G D D D D D D L 1

I I I I


A second expectation of any dynamic model is thatthe pattern of any one component will depend on theactivities of other elements of the system. Suchinteractions have been observed by monitoring thedistribution of transcripts or proteins from one pair-rule gene in embryos that lack the products ofanother. These experiments reveal a hierarchy amongthe pair-rule genes. For example, the pattern of ftzstripes is abnormal in embryos that lack the productof the hairy gene, but the hairy pattern is normal in ftzmutant embryos (Howard & Ingham, 1986). Of theset of pair-rule genes, only hairy, runt and eve arenecessary for the normal pattern of ftz expression(Carroll & Scott, 1986); the ftz protein itself is notrequired to establish the normal pattern of ftz tran-scription. Thus ftz cannot form an essential part of themechanism that generates the periodic pattern buthairy, runt and eve may play such a role, ftz must playa secondary role in stabilizing the pattern or me-diating its effects on subsequent development (seebelow).

The molecular details of these interactions are notknown. Since most of the process takes place in asyncytium, the communication that defines the spatialelements of the pattern must be between genes inadjacent nuclei, rather than between cells. Remark-ably, the ftz pattern is unaffected when the size andspacing of nuclei in the blastoderm are altered (Sulli-van, 1987). In a classical reaction-diffusion model,spatial communication would be mediated by freelydiffusible factors. In the context of the syncytialblastoderm, it is likely that the cytoskeletal architec-ture will play an important role (see Foe & Alberts,1983). Indeed, it has been noted that the transcriptsof several of the pair-rule genes accumulate in thecytoplasm immediately adjacent to the egg cortex(Fig. 4; Ingham et al. 1985a; Weir & Kornberg, 1985;Macdonald et al. 1986), in contrast to other RNAspecies, which at the same time are uniformly distrib-uted throughout the peripheral cytoplasm of thesyncytial blastoderm (MEA, unpublished results).Specific attachment to the cytoskeleton perhapsserves to restrict the diffusion of these RNAs beforecellularization, allowing the synthesis of their proteinproducts to remain tightly localized.

The terminal stripes of the blastoderm prepatternhave unique characteristics, as we would expect forany pattern generated by the dynamic interactions ofits components. The most posterior ftz stripe, forexample, is approximately twice as wide as theothers, presumably because cells on its posteriorflank are not subject to the same set of interactions asthose in the more anterior stripes (Figs 2,4). If ftz andother components of the pair-rule pattern define thestructure of metameric units, we would not expect theunits defined at the boundaries of the metameric

region (parasegments, or pseudoparasegments, 0 and15) to be equivalent, or in morphological parlancetruly 'homologous', to those generated internally. Wemight, however, expect the internal elements of thepattern to be strictly homologous, both in terms offinal structure and in the sense that their generationwould depend on the same gene products. This is notthe case. Additional gene products are required inspecific regions of the egg to allow normal generationof the pair-rule pattern. These include the products ofthe segment gap genes.

Segment gap genes and the initial subdivisionof the egg

The genes that mutate to segment gap phenotypes fallinto two broad classes; five that are required princi-pally after fertilization, during development of thezygote, and at least eight others that function largelyor exclusively during oogenesis (Table 1). Theseclasses may themselves embrace genes with veryvaried roles in development, but at least three of thezygotically acting gap genes, hunchback, Krilppel andknirps, appear to play analogous roles, each in adifferent region of the embryo. It is useful to refer tothese three as cardinal genes (Meinhardt, 1986, seebelow).

Mutations in each of these three genes deleteoverlapping regions of the segment pattern (Fig. 5;Nusslein-Volhard & Wieschaus, 1980; Nusslein-Vol-hard et al. 1982). These pattern deletions result from aprimary effect on segment specification. Forexample, in the cuticle of embryos that lack theKriippel gene product, the head and gnathal segmentsare normal, but the last gnathal segment is appendeddirectly to posterior abdominal segments. Three tho-racic and five abdominal segments are missing, andreplaced by one abdominal segment in reversedpolarity (Wieschaus et al. 19846). The localized effectof the Kriippel mutation is already apparent in blasto-derm stage embryos. Cells form normally within theprimordia for the deleted segments, but within thisregion the pattern of ftz and hairy transcription neverresolves into stripes and engrailed expression is neverestablished (Ingham et al. 1986; Carroll & Scott,1986).

The genes Kriippel and hunchback have beencloned (Preiss et al. 1985; Tautz et al. 1987). Both aretranscribed in the early syncytial blastoderm, shortlybefore the pair-rule genes are active, hunchback isinitially transcribed throughout the anterior half ofthe egg, but by early stage 14 (when the pair-rulestripes are beginning to appear), hunchback andKriippel are transcribed only within sharply definedA-P zones of the egg, located within but smaller thanthe region within which each gene affects segment

12 M. Akanx

pattern (Tautz et al. 1987; Knipple et al. 1985). Thus,although the segment deletion zones of hunchbackand Kruppel overlap, the blastoderm transcriptiondomains appear to abut. Expression of one of thesetwo genes probably represses the other, for in hunch-back mutants the transcription of Kruppel extendsanteriorly, and in Kruppel mutants the transcriptionof hunchback extends posteriorly (Fig. 5; from Jackieet al. 1986). Meinhardt (1986) predicted such arelationship for the genes denning each of a set ofadjacent cardinal domains. If such a relationship doesexist, then the expression of these cardinal genes canbe seen to define the first precise spatial subdivisions

along the A-P axis of the egg. These cardinaldomains probably play a major role not only in thecontrol of metamerization, but also, by interactingwith homeotic genes, in the control of regionalidentity (see below).

Some interaction between the pair-rule system andthe cardinal domains is evident not only from the gapsin the segment pattern, but also from alterations inthe spacing of the remaining pair-rule stripes in gapmutant embryos (Carroll & Scott, 1986; Carroll et al.1986; Mahoney & Lengyel, 1987). Both of theseobservations indicate that the pattern-generatingproperties of the pair-rule system are dependant on,

Ma Mx La Tl T2 T3 Al A2 A3 A4 A5 A6 A7 A8 A9/10i i i i i i i i i i i i i i

2 ' 3 ' 4 ' 5 ' 6 ' 7 ' 1 9 ' 10 ' 11 ' 12 ' 13 ' 14

Segmentdeletions

hb~

km

Normalembryos

hb~mutant

Kr~mutant

kni~mutant

hb

Kr

km

hb

Kr

kni

hb

Kr

km

hb

Kr

km

Fig. 5. Phenotypes and transcriptional domains of the cardinal genes.The extent of segment deletions are shown for null mutations at each of the three genes hunchback, Kruppel and

knirps. Solid bars show the deletions due to loss of zygotic gene function. The longer open bar shows the greaterdeletion resulting from the loss of both maternal and zygotic activity of the hunchback gene.

Domains of expression for the two cloned genes, hunchback and Kruppel, are. shown in the wild type and in each ofthe three mutant classes. Both the early, anterior transcription zone (compare Fig. 2B) and the slightly later posteriordomains are shown. Note that both anterior and posterior transcription zones correlate with regions of pattern deletion,but only the anterior zones are affected in the mutant genotypes. Dotted lines show the uncertainty in the location ofboundaries. (Redrawn from Jackie et al. 1986.)


and can be modified by, a set of different proteins indifferent regions of the egg. The extent of auton-omous pattern propagation by the pair-rule systemalone, for example by a reaction-diffusion mechan-ism, must be strictly limited.

Essentially nothing is known about this interaction.Both the Kriippel and the hunchback proteins havehomology to transcription factor Ilia, in the putativenucleic acid binding domain of the metal fingers(Rosenberg et al. 1986; Tautz et al. 1987), but thisalone should probably not be taken as evidence thateither acts as a transcription factor. There are furtherhomologies between the two proteins, however,which suggest that their equivalent developmentalroles are mediated by similar molecular activities(Tautz et al. 1987).

The nature and role of maternal information

The activity of the gap and segmentation genes afterfertilization elaborates the segment pattern, but it isthe structure of the egg established during oogenesisthat must establish the fixed relationship between theaxes of the egg and the blastoderm fate map. Screensfor maternal effect mutations affecting embryonicpattern have identified components of two mechan-isms, one of which determines the dorsal-ventralpolarity of the egg (for reviews see Anderson &Niisslein-Volhard, 1984; Anderson, 1987) and asecond defining the A-P axis.

The phenotypes of these mutations suggest that atleast three aspects of the A-P pattern are specifiedindependently. One group of maternally actinggenes, typified by oskar, are necessary for the forma-tion of certain posterior structures, including theprimordial germ cells and the whole abdominal re-gion of the embryo (Boswell & Mahowald, 1985;Lehmann & Niisslein-Volhard, 1986; Schiipbach &Wieschaus, 1986). Other maternal genes, best exem-plified by bicoid, are necessary only for the formationof anterior structures (Frohnhofer & Niisslein-Vol-hard, 1986). A third group, typified by torso, arenecessary for the formation of structures at both endsof the embryo (Schiipbach & Wieschaus, 1986). Atthe posterior end, mutations of the oskar and torsogroups affect different sets of structures {torso affectsthe 'tail' (Jiirgens, 1987) but not most of the abdo-men). The effects of the torso and oskar group genesare independent and additive in double mutant com-binations, suggesting that they interfere with twoindependent processes.

The results of cytoplasmic transplantation exper-iments strongly suggest that bicoid, and genes of theoskar group, are directly involved in the synthesis ofspecific anterior and posterior determinants which

are localized before fertilization. This predicted local-ization has been visualized for transcripts of the bicoidgene by in situ hybridization (Frigerio et al. 1986).

In the case of oskar, and possibly also bicoid, thesedeterminants have two activities. One is to specifydirectly the fate of those cells to which they aresegregated (e.g. for oskar, the germ cells). Thesecond activity is to generate an influence or gradientwhich affects the organization of the blastoderm fatemap in a large region of the embryo, bicoid mu-tations, for example, eliminate head structures, butalso shift the location of the cephalic furrow, whichseparates head from gnathal structures (Frohnhofer& Niisslein-Volhard, 1986; Lehmann & Niisslein-Volhard, 1986).

It is certainly possible that other, as yet unident-ified, maternal determinants specify those features ofthe blastoderm fate map that lie anterior or posteriorto the metameric region. These regions give rise to nocuticular derivatives in the embryo and so suchmutations would be difficult to detect. It is probablysignificant, however, that none of the maternal effectmutations identified to date result in interstitial gapsin the segment pattern. Experimental manipulation,of Drosophila and of many other insect eggs, suggeststhat only the anterior and posterior extremes of theA-P pattern are determined at the time of fertiliz-ation (Sander, 1976). All elements of the metamericpattern other than those at the extreme poles aresensitive to manipulations that, like ligation or ultra-violet irradiation, alter the influence of these polardeterminants (Schubiger, Moseley & Wood, 1977;Schubiger & Newman, 1982). This is in markedcontrast to the situation at cellular blastoderm, bywhich time the egg behaves as a mosaic within whichmost pattern elements are determined (Simcox &Sang, 1983).

Mutations in bicoid, and in the oskar class of genes,result in gaps in the segment pattern that are similarto those generated by mutations in some of thezygotically acting gap genes. For example, the effectsof oskar mutations on the segment pattern are similarto those of mutations in the cardinal gene knirps. Thissuggests that one role of the oskar and bicoid determi-nants is to specify the initial domains within whicheach of the cardinal genes is activated — and hencespecify the sequence of cardinal domains.

We do not know how these determinants act. Onesuggestion as to how they might act comes fromstudies of two other maternal transcripts, those fromthe genes hunchback and caudal. Both the cardinalgene hunchback and the homeotic gene caudal areexpressed zygotically in specific regions of the egg,but both are also expressed during oogenesis, so thatat fertilization the egg contains maternally derivedtranscripts which are uniformly distributed along the

14 M. Akam

A-P axis (Mlodzik et al. 1985; Mlodzik & Gehring,1987; Tautz etal. 1987). During early cleavage stages,before syncytial blastoderm formation and beforezygotic transcription of the cardinal genes, thesematernal transcripts become differentially distrib-uted, probably by selective degradation. In the caseof caudal, the protein translated from the maternaltranscripts also shows a graded distribution, accumu-lating in the posterior part of the egg. This proteingradient is apparently established before any asym-metric distribution of RNA and so may depend on theselective translation of the RNA (Macdonald &Struhl, 1986; Mlodzik & Gehring, 1987). Whateverthe mechanism, this early sign of asymmetric activitythroughout the length of the egg is presumablyinitiated by the localized polar determinants, for itoccurs in unfertilized eggs (Macdonald & Struhl,1986), in the absence of nuclear replication andzygotic gene activity.

Intriguing as these observations are, it is unlikelythat the maternal transcripts of either caudal orhunchback play a critical role in initiating the zygoticpattern of gene activity. Oocytes lacking either ofthese gene products can develop normally, providedthat they are fertilized by sperm bearing an activecopy of the gene.

Another class of maternal effect mutations resultnot in gaps in the segment pattern, but in polarityreversals affecting the whole anterior {bicaudal) orposterior {dicephalic) region of the egg (Bull, 1966;Nusslein-Volhard, 1977; Lohs-Schardin, 1982;Mohler & Wieschaus, 1986). It is likely that thesemutations are altering the maternal localization ofdeterminants, rather than eliminating them.

The specification of regional identity

Establishing the pattern of homeotic gene expressionModels for segmentation and for the control ofsegment identity have been developed largely inisolation, but the two processes must be intimatelylinked (Akam, 1985; Meinhardt, 1986). States ofhomeotic gene expression change abruptly at para-segment boundaries, and experimental manipulationgenerally perturbs segmentation and segment identitycoordinately. Normal segmentation can be estab-lished in the absence of any segment diversity when,for example, an embryo lacks much of the Antenna-pedia and bithorax complexes, so it must presumablybe the homeotic genes that take their cues from thesegmentation mechanism.

It is now clear that segmentation genes at severallevels in the regulatory hierarchy are intimatelyinvolved in establishing the spatial activity of homeo-tic genes. Mutations in the maternal and zygotic gap

genes and in some of the pair-rule genes affect thevery earliest patterns of homeotic gene expression.

Mutations in hunchback and oskar have particularlydramatic, and largely reciprocal effects, on the earlyexpression of Antp and Ubx. Hunchback activityappears to repress Ubx in the head and thorax (White& Lehmann, 1986), but to be necessary for theactivation of Antp (A. Martinez-Arias, personal com-munication). The posterior determinants dependanton oskar are necessary for the initial activation of Ubxin the abdomen, and also for the repression of Antp inthe same region (Martinez-Arias and ME A, unpub-lished). These interactions affect protein distributionsas soon as they can be determined in the extendedgerm band, but they can be visualized as alterations inthe pattern of transcript distribution as early as theblastoderm.-

Within its wide domain of activity, a homeotic genemay play very different roles in adjacent segments.Ubx, for example, serves quite different functions(Lewis, 1978; Casanova, Sanchez-Herrero & Morata,1985) and is expressed in strikingly different patternsin parasegments 5 and 6 (Akam & Martinez-Arias,1985; Beachy, Helfand & Hogness, 1985; White &Wilcox, 1985). This differential regulation in adjacentparasegments depends on the activity of some but notall of the pair-rule genes.

Mutations in one pair-rule mutation, oddpaired(opa), fuse parasegments 5 and 6, but have littleeffect on the expression of Ubx (Ingham & Martinez-Arias, 1986). Other pair-rule genes, however, clearlyperturb both segmentation and homeotic gene ex-pression, and one in particular, ftz, appears to play amajor role in coupling the spatial regulation ofhomeotic genes to the evolving segment pattern(Duncan, 1986; Ingham & Martinez-Arias, 1986). Innormal embryos, the domains of ftz expressioncoincide with even-numbered parasegments (seeFig. 4). In the blastoderm of ftz mutant embryos, theearly peaks of- Scr, Antp and Ubx expression areabolished. These peaks normally lie within and prob-ably define the primordia for parasegments 2,4 and 6.Later in development, the uniform and high levels ofhomeotic gene expression that normally characterizethese even-numbered parasegments are not ob-served. Instead, the double-segment units generatedin ftz mutant embryos behave as single entities withrespect to the expression of these homeotic genes,and most closely resemble odd-numbered paraseg-ments in their levels and patterns of homeotic geneexpression. This behaviour is in marked contrast withthe bipartite pattern of homeotic gene expressionobserved in the morphologically similar units of theopa embryo (Ingham & Martinez-Arias, 1986).

A direct interaction between the ftz protein andgenes of the Antennapedia and bithorax complexes


seems particularly likely, ftz is the only segmentationgene to contain a 'class 1' homeobox (i.e. one with asequence that is very closely related to those of othergenes in the Antennapedia and bithorax complexes(McGinnis etal. 1984; Scott & Weiner, 1984)). In laterdevelopment, the products of the Antennapedia andbithorax complex genes regulate one-anothers' tran-scription (see below); at early stages, the ftz proteinmay play a similar role.

Elaborating the pattern of homeotic gene expressionOnce the germ band is defined, a new set of controlsare established to maintain and further elaborate thepattern of homeotic gene expression. This must bethe case, for the gap and pair-rule genes that establishthe early pattern are expressed only transiently, andat least in the case of ftz, their protein products decayrapidly (Carroll & Scott, 1985).

These later controls involve at least three factors.One set of products is required to maintain therepression of many homeotic genes in those segmentswhere each is not normally active. These are theproducts of the genes Poly comb (Lewis, 1978),extra sex combs {esc, Struhl, 1981, 1983) and otherswith similar phenotypes (Duncan and Lewis, 1982;Ingham, 1984; Jiirgens, 1985; Dura, Brock & Santa-maria, 1985). Mutations in all of these genes allowinappropriate expression of both ANT-C and BX-Cgenes in all body segments and even in the head. Thisphenotype originally suggested that their productsmight be involved in the initial positional activation ofthe homeotic genes, perhaps serving to define agradient (see Ingham, 1985). This cannot be the case,at least for Polycomb or esc, as mutations in thesegenes have no effect on homeotic gene expressionuntil after the germ band is established (Struhl &Akam, 1985; Wedeen, Harding & Levine, 1986).

A second, and independent, set of controls involveinteractions between the homeotic genes themselves.Among the genes Antp, Ubx, abdominal-A (abd-A)and Abdominal-B (Abd-B), each of the more pos-teriorly expressed genes represses the expression ofother genes initially active within its own domain(Hafen et al. 19846; Harding et al. 1985; Struhl &White, 1985). This effect is seen most clearly in thenervous system during later stages of embryogenesis,but even here it is not complete. Ubx, for example, isexpressed in most or all neural cells of parasegment 6throughout embryogenesis and, in the absence of theabd-A and Abd-B genes, it is similarly expressed inparasegments 7 through 13, where it was initiallyactive. In normal development, however, the activityof abd-A and Abd-B in parasegments 7 to 13 limitsUbx expression to only a defined subset of neural cellsin each segment (Struhl & White, 1985).

This cell-specific pattern of expression within indi-vidual metameres implies that Ubx, and presumablyother homeotic genes, are also regulated by a thirdclass of factors - those which identify positions or celltypes in each metamere. The differential expressionof homeotic genes within regions of a single paraseg-ment is seen in many tissues and becomes moreelaborate from early germ band stages onwards(White & Wilcox, 1985). For Ubx, it clearly dependson genes that define the anteroposterior organizationof the parasegment, including engrailed (Martinez-Arias & White, 1987) and wingless (A. Martinez-Arias & N. Baker, personal communication). Itprobably also depends on genes that define thedifferences between cell types, but these have yet tobe identified.

An evolutionary perspective

Cells fated to give rise to the segmented region of theDrosophila embryo occupy a large fraction of theblastoderm (Fig. 1) and the metameric pattern isessentially defined by the time of gastrulation (seeTechnau, 1987, for review). In this respect, Dros-ophila and other long-germ insects are exceptionalwithin the annelid-arthropod lineage. In less-special-ized insects, the abdominal segments are generatedafter gastrulation, or even postembryonically, bymitotic division within a growth zone. (Anderson,-1972; Jura, 1972). The full pattern of the germ band isnot determined in the blastoderm. In the annelid- andmyriapod-like forms believed to be ancestral to theinsects, sequential growth of the metameric region ofthe body is generally observed and this frequentlyoccurs postembryonically (Anderson, 1973). Indeed,the archetypal pattern of annelid development exem-plified by primitive polychaetes is characterized byquite the opposite strategy of development. Here, theprimary pattern-forming processes of embryogenesisare concerned principally with the generation of thoseparts of the body that are not metamerically seg-mented. Of the 64 cells generated by the first sixdivisions of the stereotyped spiral cleavage pattern, itis typically only the 4d cell and four progeny of the 2dcell that give rise to the entire metameric region of thebody. These form stem cells or teloblasts, which forma growth zone within the trochophore larva, fromwhich the segmented body develops. The remainingcells of the blastula are otherwise specified to formpre-oral structures including the brain, the gut andspecialized larval cells (Wilson, 1892; Anderson,1973).

Since the pattern of spiral cleavage is essentiallyindistinguishable in annelids and molluscs, and ap-pears in slightly modified form in flatworms (Mac-bride, 1914), we must suspect that it is an ancient

16 M. Akam

developmental mechanism, and not a secondary ad-aptation to generate specialized larval structures. So,if we accept a common origin for metamerization inannelids and arthropods, then the mechanism ofsegmentation in Drosophila must have evolved from abudding process that originally occurred later indevelopment. It is perhaps significant in this regardthat localized maternal determinants, which appearto play such a major role in the specification ofembryonic lineages in the spirally cleaving embryos,probably have no direct role in the specification ofindividual segments or of regional differences withinthe metameric region of the insect embryo.

At present, we do not know what elements of thesegmentation mechanism have been conserved be-tween annelids and insects. We might expect thatgenes like engrailed, directly involved in maintainingthe segment pattern throughout development, wouldplay equivalent roles in all those organisms that sharewith Drosophila a common origin of metamerism. Onthe other hand, the transient pair-rule prepattern maybe an invention of the insects that relates specificallyto the subdivision of an extended metameric region inthe blastoderm. This need not be the case, however.Interacting systems that generate extended patternsin space can, with little change in regulatory net-works, generate patterns that have both temporal andspatial components (Meinhardt, 1982).

The evolving role of homeotic genes

It is generally assumed that homeotic genes serve todefine segment identities. This seems to be an appro-priate description of their role in most segments ofthe Drosophila adult, all of which are unique (Lewis,1978). It is probably not, however, appropriate tothink in this way about homeotic genes in primitivemembers of the annelid-arthropod lineage. In manyannelids, segment number is not defined and manysegments have no unique identities.

I find it more likely that homeotic genes of theAntennapedia-Bithorax class originally served todistinguish between the presumptive metameric re-gion and other parts of the embryo. We have littleidea what the role of such an ancestral Antennapedia-like gene might have been within these cells, though itcould perhaps have been related to the requirementfor continued postembryonic growth of the seg-mented region. Subsequently, as segment diversityarose, these same genes must have been utilized todefine different parts of the metameric region (seeMartinez-Arias, 1987), but it is hard to imagine thatprecise boundaries between uniquely defined seg-ments arose immediately.

In the early embryo of even such an advancedarthropod as Drosophila, the expression of homeotic

genes suggests that all parasegments are not qualitat-ively distinguished. The regions of the mouthpartsand the 'tail' (Jurgens, 1987) seem to be 'hard-wired',in that the Deformed and caudal genes are initiallyactivated in precisely the correct regions of theblastoderm (see Fig. 3), but definition of the thoracicand abdominal segments has to evolve, by interac-tions and modulations of the set of genes expressed inparasegments 3-13. Parasegments 3, 4, 5 and 6rapidly come to have unique identities, but paraseg-ments 7 to 12 are remarkably similar throughoutmuch of embryogenesis. The catalogue of structuresin their musculature and peripheral nervous system isidentical (Ghysen et al. 1986; Campos-Ortega &Hartenstein, 1985; Hooper, 1986) and in the earlygerm band they show the same qualitative patterns ofhomeotic gene expression (see Fig. 3).

With this-perspective, it is interesting to revive theviews of a classical arthropod morphologist. Snod-grass (1935) selected two features to characterize thearthropods; the possession of jointed limbs (a criter-ion familiar to all elementary students of zoology)and the subdivision of the body into regions or'tagmata', each composed of structurally and hencedevelopmentally similar segments. The particulararray of tagmata serves to define the major classes ofarthropods - insects having 3 gnathal, 3 thoracic and8-11 abdominal segments; arachnids and Crustaceapossessing different tagmatal organizations.

Any model that seeks to explain segment diversitywithin the arthropods must account for both thesimilarity of segments within a tagma and the differ-ences between them. It may be that early in theevolution of the arthropods, a single homeotic genedefined, not segment identity, but the developmentalcharacteristics of segments throughout each tagma.Expression of the Antennapedia gene would define allthoracic parasegments, and expression of a prototypi-cal bithorax complex gene would define abdominalparasegments. This correlation between tagmaticidentity and homeotic gene expression is no longerclear in the adult roles of homeotic genes in theDiptera, but it is strongly suggested by their patternsof expression in the early germ band.

In form and content this essay owes much to colleagues inCambridge. I thank Michael Ashburner, Nick Baker,Michael Bate, Phil Ingham, Vivian Irish, Peter Lawrence,Annie Rowe, Ernesto Sanchez-Herrero and Rob White fortheir advice, and most particularly Alfonso Martinez-Ariasfor innumerable discussions that have played a large partin shaping my thoughts. P.I., V.I., A.R., E.S.H. andA.M.-A. have generously allowed me to use unpublishedmaterial to prepare the figures.

It is a pleasure to acknowledge my debt to C. Nusslein-Volhard and her colleagues. They have laid a foundationfor this work by pioneering the fusion of systematic genetics


with experimental embryology. My conception of earlyevents in Drosophila embryogenesis was enormously clari-fied in a superb seminar given by Nusslein-Volhard atLeuenberg in 1986.1 thank also Hans Meinhardt for helpingme to understand the implications of reaction-diffusionmodels, Herbert Jackie for enthusing about the gap genes,Ruth Lehmann for correcting at least one of my misunder-standings, and the many others who provided manuscriptsprior to publication. Geoffrey North cast a professional eyeon an early version of this manuscript, and his editorialcomments were much appreciated. John Rodford preparedthe illustrations and Hossein Majidi the colour photo-graphs.

This work was supported by the Medical ResearchCouncil of Great Britain.

References

AKAM, M. (1985). Segments, lineage boundaries and thedomains of expression of homeotic genes. Phil, trans.Roy. Soc. Lond. B 312, 179-187.

AKAM, M. & MARTINEZ-ARIAS, A. (1985). The

distribution of Ultrabithorax transcripts in Drosophilaembryos. EMBO J. 4, 1689-1700.

ANDERSON, D. T. (1972). The development ofhemimetabolous insects. In Developmental systems:Insects, vol. 1 (ed. S. J. Counce, & C. H.Waddington), pp. 96-165. London: Academic Press.

ANDERSON, D. T. (1973). Embryology and Phylogeny inAnnelids and Arthropods. Oxford: Pergamon.

ANDERSON, K. (1987). Dorsal-ventral embryonic patterngenes of Drosophila. Trends in Genetics 3, 91-97.

ANDERSON, K. & NUSSLEIN-VOLHARD, C. (1984). Geneticanalysis of dorsal-ventral embryonic pattern inDrosophila. In Pattern Formation (ed. G. M. Malacinski& S. V. Bryant), pp. 269-290. New York: Macmillan.

BAKER, N. (1987a). Embryonic and imaginal requirementsfor wingless, a segment polarity gene in Drosophila.Devi Biol. (in press).

BAKER, N. (1987fc>). Molecular cloning of sequences fromwingless, a segment polarity gene in Drosophila: thespatial distribution of a transcript in embryos. EMBOJ. 6, 1765-1774.

BEACHY, P. A., HELFAND, S. & HOGNESS, D. S. (1985).

Segmental distribution of bithorax complex proteinsduring Drosophila development. Nature, Lond. 313,545-551.

BENDER, M., TURNER, F. R. & KAUFMAN, T. C. (1987). A

developmental genetic analysis of the gene Regulator ofpostbithorax in Drosophila melanogaster. Devi Biol 119,418-432.

BENDER, W., AKAM, M. E., KARCH, F., BEACHY, P. A.,

PEIFER, M., SPIERER, P., LEWIS, E. B. & HOGNESS, D.

S. (1983). Molecular genetics of the bithorax complexin Drosophila melanogaster. Science 221, 23-29.

BOPP, D., BURRI, M., BAUMGARTNER, S., FRJGERJO, G. &

NOLL, M. (1986). Conservation of a large proteindomain in the segmentation gene paired and infunctionally related genes of Drosophila. Cell 47,1033-1040.

BOSWELL, R. E. & MAHOWALD, A. P. (1985). tudor, agene required for assembly of the germ plasm inDrosophila. Cell 43, 97-104.

BRIDGES, C. B. & MORGAN, T. H. (1923). The thirdchromosome group of mutant characters of Drosophilamelanogaster. Carnegie Inst. Wash. Publ. No. 327,137-146.

BULL, A. L. (1966). Bicaudal, a genetic factor whichaffects the polarity of the embryo in Drosophilamelanogaster. J. exp. Zool. 161, 221-242.

CAMPOS-ORTEGA, J. A. & HARTENSTEIN, V. (1985). The

Embryonic Development of Drosophila melanogaster.Berlin: Springer-Verlag.

CARROLL, S. B., LAYMON, R. A., MCCUTCHEON, M. A.,

RILEY, P. D. & SCOTT, M. P. (1986a). The localizationand regulation of Antennapedia protein expression inDrosophila embryos. Cell 47, 113-122.

CARROLL, S. B. & SCOTT, M. P. (1985). Localisation ofthe fushi-tarazu protein during Drosophilaembryogenesis. Cell 43, 47-57.

CARROLL, S. B. & SCOTT, M. P. (1986). Zygotically activegenes that affect the spatial expression of the fushi-tarazu segmentation gene during early Drosophilaembryogenesis. Cell 45, 113-126.

CARROLL, S. B., WINSLOW, G. M., SCHUPBACH, T. &

SCOTT, M. P. (19866). Maternal control of Drosophilasegmentation gene expression. Nature, Lond. 323,278-280.

CASANOVA, J., SANCHEZ-HERRERO, E. & MORATA, G.

(1985). Prothoracic transformation and functionalstructure of the Ultrabithorax gene in Drosophila. Cell42, 663-669.

CASANOVA, J., SANCHEZ-HERRERO, E. & MORATA, G.

(1986). Identification and characterization of aparasegment specific regulatory element of theAbdominal-B gene of Drosophila. Cell 47, 627-636.

CHADWICK, R. & MCGINNIS, W. (1987). Temporal andspatial distribution of transcripts from the Deformedgene of Drosophila. EMBO J. 6, 779-789.

COTE, S., PREISS, A., HALLER, J., SCHUH, R., KIENCIN,

A., SEIFERT, E. & JACKLE, H. (1987). The gooseberry-zipper region of Drosophila: five genes encode differentspatially restricted transcripts in the embryo. EMBO J.(in press).

DAMBLY-CHAUDIERE, C. & GHYSEN, A. (1986). The senseorgans in the Drosophila larva and their relation to theembryonic pattern of sensory neurons. Wilhelm RouxArch devl Biol 195, 222-228.

DEGELMAN, A., HARDY, A., PERRIMON, N. & MAHOWALD,

A. P. (1986). Developmental analysis of the torso-Yikephenotype in Drosophila produced by a maternal effectlocus. Devi Biol. 115, 479-489.

DINARDO, S., KUNER, J. M., THEIS, J. & O'FARRELL, P.

H. (1985). Development of embryonic pattern inDrosophila melanogaster as revealed by accumulation ofthe nuclear engrailed protein. Cell 43, 59-69.

DUNCAN, I. (1986). Control of bithorax complex functionsby the segmentation gene fushi-tarazu of Drosophilamelanogaster. Cell 47, 297-309.

DUNCAN, I. & LEWIS, E. B. (1982). Genetic control ofbody segment differentiation in Drosophila. In

18 M. Akam

Developmental Order: Its Origin and Regulation (ed. S.Subtelny). New York: Liss. Symp. Soc. Devi Biol. 40,533-554.

DURA, J-M., BROCK, H. W. & SANTAMARIA, P. (1985).Polyhomeotic: A gene of Drosophila melanogasterrequired for correct expression of segmental identity.Molec. gen. Genet. 198, 213-220.

EDGAR, B. A. & SCHUBIGER, G. (1986). Parameterscontrolling Transcriptional Activation during EarlyDrosophila Development. Cell 44, 871-877.

EDGAR, B. A., WEIR, M. P., SCHUBIGER, G. & KORNBERG,T. (1986). Repression and turnover pattern olfushi-tarazu RNA in the early Drosophila embryo. Cell 47,747-754.

FJOSE, A., MCGINNIS, W. J. & GEHRING, W. J. (1985).Isolation of a homeobox containing gene from theengrailed region of Drosophila and the spatialdistribution of its transcripts. Nature, Lond. 313,284-289.

FOE, V. E. & ALBERTS, B. M. (1983). Studies of nuclearand cytoplasmic behaviour in the five mitotic cyclesthat precede gastrulation in Drosophila embryogenesis.J. CellSci. 61,31-70.

FRASCH, M., HOEY, T., RUSHLOW, C , DOYLE, H. &LEVINE, M. (1987). Characterization and localization ofthe even-skipped protein of Drosophila. EMBO J. 6,749-759.

FRIGERIO, G., BURRI, M., BOPP, D., BAUMGARTNER, S. &NOLL, M. (1986). Structure of the segmentation genepaired and the Drosophila PRD gene set as part of agene network. Cell 47, 735-746.

FROHNHOFER, H. G., LEHMANN, R. & NUSSLEIN-VOLHARD, C. (1986). Manipulating the anteroposteriorpattern of the Drosophila embryo. /. Embryol. exp.Morph. 97 Supplement, 169-179.

FROHNHOFER, H. G. & NUSSLEIN-VOLHARD, C. (1986).The organization of anterior pattern in the Drosophilaembryo by the maternal gene bicoid. Nature, Lond.324, 120-125.

GANS, M., AUDIT, C. & MASSON, M. (1975). Isolationand characterization of sex linked lethal mutations ofDrosophila melanogaster. Genetics 81, 683-704.

GARBER, R. L., KUROIWA, A. & GEHRING, W. J. (1983).Genomic and cDNA clones of the homeotic locusAntennapedia in Drosophila. EMBO J. 2, 2027-2036.

GARCIA-BELLIDO, A. & SANTAMARIA, P. (1972).Developmental analysis of the wing disc in the mutantengrailed of Drosophila melanogaster. Genetics 72,87-104.

GERGEN, J. P., COULTER, D. & WIESCHAUS, E. (1986).Segmental pattern and blastoderm cell identities. InGametogenesis and the Early Embryo (ed. S. Subtelny).New York: Liss. Symp. Soc. Devi Biol. 43, 195-220.

GERGEN, J. P. & WIESCHAUS, E. F. (1985). The localizedrequirements for a gene affecting segmentation inDrosophila: analysis of larvae mosaic for runt. DeviBiol. 109, 321-335.

GERGEN, J. P. & WIESCHAUS, E. F. (1986a). Dosagerequirements for runt in the segmentation of Drosophilaembryos. Cell 45, 289-299.

GERGEN, J. P. & WIESCHAUS, E. F. (19866). Localizedrequirements for gene activity in segmentation ofDrosophila embryos: analysis of armadillo, fused, giantand unpaired mutations in mosaic embryos. WilhelmRoux Arch, devl Biol. 195, 49-62.

GHYSEN, A., DAMBLY-CHAUDIERE, CH., ACEVES, E., JAN,

L. Y. & JAN, Y. N. (1986). Sensory neurons andperipheral pathways in Drosophila embryos. WilhelmRoux Arch, devl Biol. 195, 281-289.

GLOOR, H. (1950). Schadigungsmuster eines Letalfaktors(A>) von Drosophila melanogaster. Arch. Jul. KlausStiftung 29, 277-287.

HAFEN, E., KUROIWA, A. & GEHRING, W. J. (1984a).

Spatial distribution of transcripts from thesegmentation gene fushi-tarazu during Drosophilaembryonic development. Cell 37, 833-841.

HAFEN, E., LEVINE, M., GARBER, R. L. & GEHRING, W.

J. (1983). An improved in situ hybridization methodfor the detection of cellular RNAs in Drosophila tissuesections, and its application for localizing transcripts ofthe homeotic Antennapedia gene complex. EMBO J. 2,617-623.

HAFEN, E., LEVINE, M. & GEHRING, W. J. (19846).

Regulation of Antennapedia transcript distribution bythe bithorax complex of Drosophila. Nature, Lond. 307,287-289.

HARDING, K., RUSHLOW, C , DOYLE, H. J., HOEY, T. &

LEVINE, M. (1986). Cross-regulatory Interactionsamong Pair-rule Genes in Drosophila. Science 233,953-959.

HARDING, K., WEDEEN, C , MCGINNIS, W. & LEVINE, M.

(1985). Spatially regulated expression of homeoticgenes in Drosophila. Science 229, 1236—1242.

HIROMI, Y., KUROIWA, A. & GEHRING, W. J. (1985).Control elements of the Drosophila segmentation genefushi-tarazu. Cell 43, 603-613.

HOEY, T., DOYLE, H. J., HARDING, K., WEDEEN, C. &

LEVINE, M. (1986). Homeo box gene expression inanterior and posterior regions of the Drosophilaembryo. Proc. natn. Acad. Sci. U.S.A. 83, 4809-4813.

HOGNESS, D. S., LIPSHITZ, H. D., BEACHY, P. A.,

PEATTIE, D. A., SAINT, R. B., GOLDSCHMIDT-CLERMONT,

M., HARTE, P. J., GAVIS, E. R. & HELFAND, S. L.

(1985). Regulation and products of the Ultrabithoraxdomain of the bithorax complex. Cold Spring HarbourSymp. Quant. Biol. 50, 181-195.

HOLMGREN, R. (1984). Cloning sequences from the hairygene of Drosophila. EMBO J. 3, 833-841.

HOOPER, J. (1986). Homeotic gene function in themuscles of Drosophila larvae. EMBO J. 5, 2321-2329.

HOWARD, K. & INGHAM, P. (1986). Regulatoryinteractions between the segmentation genes fushi-tarazu, hairy and engrailed in the Drosophilablastoderm. Cell 44, 949-957.

INGHAM, P. (1984). A gene that regulates the bithoraxcomplex differentially in larval and adult cells ofDrosophila. Cell 37, 815-823.

INGHAM, P. (1985). The regulation of the bithoraxcomplex. Trends in Genetics 1, 112-116.


INGHAM, P. W., HOWARD, K. R. & ISH-HOROWICZ, D.

(1985a). Transcription pattern of the Drosophilasegmentation gene hairy. Nature, Lond. 318, 439-445.

INGHAM, P., ISH-HOROWICZ, D. & HOWARD, K. (1986).

Correlative changes in homeotic and segmentation geneexpression in Kriippel mutant embryos in Drosophila.EMBOJ. 5, 1659-1665.

INGHAM, P. W. & MARTINEZ-ARIAS, A. (1986). Thecorrect activation of Antennapedia and bithoraxcomplex genes requires the fushi-tarazu gene. Nature,Lond. 324, 592-597.

INGHAM, P., MARTINEZ-ARIAS, A., LAWRENCE, P. A. &

HOWARD, K. (1985£>). Expression of engrailed in theparasegment of Drosophila. Nature, Lond. 317,634-636.

INGHAM, P. W., PINCHIN, S. M., HOWARD, K. R. & ISH-

HOROWICZ, D. (1985c). Genetic analysis of the hairylocus in Drosophila melanogaster. Genetics 111, 463-486.

ISH-HOROWICZ, D., HOWARD, K. R., PINCHIN, S. M. &

INGHAM, P. W. (1985). Molecular and Genetic analysisof the hairy locus in Drosophila. Cold Spring HarborSymp. Quant. Biol. 50, 135-144.

JACKLE, H., TAUTZ, D., SCHUH, R., SEIFERT, E. &

LEHMANN, R. (1986). Cross regulatory interactionsamong the gap genes of Drosophila. Nature, Lond. 324,668-670.

JURA, C. (1972). Development of apterygote insects. InDevelopmental Systems: Insects, vol. 1 (ed. S. J. Counce& C. H. Waddington), pp. 49-95. London: AcademicPress.

JURGENS, G. (1985). A group of genes controlling theexpression of the bithorax complex in Drosophila.Nature, Lond. 316, 153-155.

JURGENS, G. (1987). Segmental organization of the tailregion in the embryo of Drosophila melanogaster. Ablastoderm fate map of the cuticle structures of thelarval tail region. Wilhelm Roux Arch, devl Biol. 196,141-157.

JURGENS, G., WIESCHAUS, E., NUSSLEIN-VOLHARD, C. &KLUDING, H. (1984). Mutations affecting the pattern ofthe larval cuticle in Drosophila melanogaster. II. Zygoticloci on the third chromosome. Wilhelm Roux Arch, devlBiol. 193, 283-295.

KARCH, F., WEIFFENBACH, B., PEIFER, M., BENDER, W.,DUNCAN, I., CELNIKER, S., CROSBY, M. & LEWIS, E. B.(1985). The abdominal region of the bithorax complex.Cell 43, 81-%.

KAUFMAN, T. C. (1983). The genetic regulation ofsegmentation in Drosophila melanogaster. In Time,Space and Pattern in Embryonic Development (ed. R. A.Raff & W. Jefferys), pp. 365-383. New York: AlanLiss.

KAUFMAN, T. C. & ABBOTT, M. K. (1984). Homeoticgenes and the specification of segment identity in theembryo and adult thorax of Drosophila melanogaster. InMolecular Aspects of Development (ed. G. M.Malacinski & W. H. Klein), pp. 182-218. New York:Plenum Press.

KILCHERR, F., BAUMGARTNER, S., BOPP, D., FREI, E. &NOLL, M. (1986). Isolation of the paired gene of

Drosophila and its spatial expression during earlyembryogenesis. Nature, Lond. 321, 493-499.

KNIPPLE, D. C , SEIFERT, E., ROSENBERG, U. B., PREISS,A. & JACKLE, H. (1985). Spatial and temporal patternsof Kriippel gene expression in early Drosophilaembryos. Nature, Lond. 317, 40-44.

KORNBERG, T. (1981). engrailed: a gene controllingcompartment and segment formation in Drosophila.Proc. natn. Acad. Sci. U.S.A. 78, 1095-1099.

KORNBERG, T., SIDEN, I., O'FARRELL, P. & SIMON, M.(1985). The engrailed locus of Drosophila: In situlocalization of transcripts reveals compartment specificexpression. Cell 40, 45-53.

KUROIWA, A., HAPEN, E. & GEHRING, W. J. (1984).Cloning and transcriptional analysis of thesegmentation gene fushi-tarazu of Drosophila. Cell 37,825-831.

KUROIWA, K., KLOTER, U., BAUMGARTNER, P. & GEHRING,W. J. (1985). Cloning of the homeotic Sex combsreduced gene in Drosophila and in situ localization of itstranscripts. EMBO J. 4, 3757-3764.

LAUGHON, A., BOULET, A. M., BERMINGHAM, J. R.,LAYMON, R. A. & SCOTT, M. P. (1986). The structureof transcripts from the homeotic Antennapedia gene ofDrosophila: two promoters control the major protein-coding region. Molec. Cell. Biol. 6, 6476-6481.

LAUGHON, A. & SCOTT, M. P. (1984). The sequence of aDrosophila segmentation gene: Protein Structurehomology with DNA binding proteins. Nature, Lond.308, 25-31.

LAWRENCE, P. A., JOHNSTON, P., MACDONALD, P. &STRUHL, G. (1987). The. fushi-tarazu and even-skippedgenes delimit the borders of parasegments inDrosophila embryos. Nature, Lond. (in press).

LEHMANN, R. & NUSSLEIN-VOLHARD, C. (1986).Abdominal segmentation, pole cell formation andembryonic polarity require the localized activity ofoskar, a maternal gene in Drosophila. Cell 47, 141-152.

LEHMANN, R. & NUSSLEIN-VOLHARD, C. (1987).hunchback, a gene required for segmentation of ananterior and posterior region of the Drosophilaembryo. Devl Biol. 119, 402-417.

LEVTNE, M., HAFEN, E., GARBER, R. L. & GEHRING, W.J. (1983). Spatial distribution of Antennapediatranscripts during Drosophila development. EMBO J.2, 2037-2046.

LEWIS, E. B. (1963). Genes and developmental pathways.Am. Zool. 3, 33-56.

LEWIS, E. B. (1978). A gene complex controllingsegmentation in Drosophila. Nature, Lond. 276,565-570.

LEWIS, E. B. (1981). Developmental genetics of thebithorax complex in Drosophila. In DevelopmentalBiology using Purified Genes. ICN-UCLA symposia 23(ed. D. D. Brown & C. F. Fox), pp. 189-208. NewYork: Academic Press.

LOHS-SCHARDIN, M. (1982). Dicephalic - a Drosophilamutant affecting polarity in follicle organization andembryonic patterning. Wilhelm Roux Arch, devl Biol.191, 28-36.

20 M. Akam

MACBRIDE, E. W. (1914). Textbook of Embryology, vol 1.Invertebrata. London: Macmillan.

MACDONALD, P. M., INGHAM, P. & STRUHL, G. (1986).Isolation, structure and expression of even-skipped: asecond pair rule gene of Drosophila containing ahomeobox. Cell 47, 721-734.

MACDONALD, P. M. & STRUHL, G. (1986). A moleculargradient in early Drosophila embryos and its role inspecifying the body pattern. Nature, Lond. 324,537-545.

MAHONEY, P. A. & LENGYEL, J. A. (1987). The zygoticsegmentation mutant tailless alters the blastoderm fatemap of Drosophila embryos. Devi Biol. (in press).

MARTINEZ-ARIAS, A. (1985). The development of fused"embryos of Drosophila melanogaster. J. Embryol. exp.Morph. 87, 99-114.

MARTINEZ-ARIAS, A. (1986). The Antennapedia gene isrequired and expressed in parasegments 4 and 5 of theDrosophila embryo. EMBO J. 5, 135-141.

MARTINEZ-ARIAS, A. (1987). On the developmental andevolutionary role of some genes from theAntennapedia complex. In Molecular Approaches toDevelopmental Biology (ed. R. A. Firtel & E. H.Davidson) 1CN-UCLA symposia. Mol. Cell Biol. 51,131-146.

MARTINEZ-ARIAS, A. & INGHAM, P. W. (1985). The originof pattern duplications in segment polarity mutants ofDrosophila melanogaster. J. Embryol. exp. Morph. 87,129-135.

MARTINEZ-ARIAS, A., INGHAM, P. W., SCOTT, M. &AKAM, M. (1987a). The spatial and temporaldeployment of Dfd and Scr transcripts throughoutdevelopment of Drosophila. Development 100, 673-683.

MARTINEZ-ARIAS, A. & LAWRENCE, P. A. (1985).Parasegments and compartments in the Drosophilaembryo. Nature, Lond. 313, 639-642.

MARTINEZ-ARIAS, A. & WHITE, R. A. H. (1987).Ultrabithorax and engrailed expression in Drosophilaembryos mutant for the pair rule class of segmentationgenes. Development (submitted).

MCGINNIS, W., LEVINE, M. S., HAFEN, E., KUROIWA, A.& GEHRING, W. J. (1984). A conserved DNA sequencein homoeotic genes of the Drosophila Antennapediaand bithorax complexes. Nature, Lond. 308, 428-433.

MCGINNIS, W. M. (1985). Homeobox sequences of theAntennapedia class are conserved only in higher animalgenomes. Cold Spring Harbor Symp. Quant. Biol. 50,263-270.

MEINHARDT, H. (1982). Models for Biological PatternFormation. London: Academic Press.

MEINHARDT, H. (1986). Hierarchical inductions of cellstates - a model for segmentation in Drosophila. J. CellSci. Suppl. 4, 357-381.

MLODZIK, M., FJOSE, A. & GEHRING, W. J. (1985).Isolation of caudal, a Drosophila homeobox containinggene with maternal expression, whose transcripts forma concentration gradient at the pre-blastoderm stage.EMBO J. 4, 2961-2969.

MLODZIK, M. & GEHRING, W. J. (1987). Expression ofthe caudal gene in Drosophila: Formation of an RNA

and protein gradient during early embryogenesis. Cell48, 465-478.

MOHLER, J. & WIESCHAUS, E. F. (1986). Dominantmaternal effect mutations of Drosophila melanogastercausing the production of double-abdomen embryos.Genetics 112, 803-822.

MORATA, G., BOTAS, S., KERRIDGE, S. & STRUHL, G.(1983). Homeotic transformations of the abdominalsegments of Drosophila caused by breaking or deletinga central portion of the bithorax complex. J. Embryol.exp. Morph. 78, 319-341.

MORATA, G. & LAWRENCE, P. A. (1975). Control ofcompartment development by the engrailed gene inDrosophila. Nature, Lond. 255, 614-617.

NUSSLEIN-VOLHARD, C. (1977). Genetic analysis ofpattern formation in the embryo of Drosophilamelanogaster: Characterization of the maternal effectmutant bicaudal. Wilhelm Roux Arch, devl Biol 183,249-268.

NUSSLEIN-VOLHARD, C , KLUDING, H. & JURGENS, G.(1985). Genes affecting the segmental subdivision ofthe Drosophila embryo. Cold Spring harbor Symp.Quant. Biol. 50, 145-154.

NUSSLEIN-VOLHARD, C. & WIESCHAUS, E. (1980).Mutations affecting segment number and polarity inDrosophila. Nature, Lond. 287, 795-801.

NUSSLEIN-VOLHARD, C , WIESCHAUS, E. & JURGENS, G.(1982). Segmentierung bei Drosophila; eine genetischeanalyse. Verb. Dtsch. Zool. Ges. 1982, 91-104.

NUSSLEIN-VOLHARD, C , WIESCHAUS, E. & KLUDING, H.(1984). Mutations affecting the pattern of the larvalcuticle in Drosophila melanogaster. I. Zygotic loci onthe second chromosome. Wilhelm Roux Arch, devl Biol.193, 267-282.

O'FARRELL, P. & SCOTT, M. P. (1986). Spatialprogramming of gene expression in early Drosophilaembryogenesis. Ann. Rev. Cell Biol. 2, 49-80.

PERRIMON, N., ENGSTROM, L. & MAHOWALD, A. P.(1984). The effects of zygotic lethal mutations onfemale germ line functions in Drosophila. Devl Biol.105, 404-414.

PERRIMON, N. & MAHOWALD, A. P. (1986). l(l)hopscotch,a larval pupal zygotic lethal with a specific maternaleffect on segmentation in Drosophila. Devl Biol. 118,28-41.

PERRIMON, N. & MAHOWALD, A. P. (1987). Multiplefunctions of segment polarity genes in Drosophila. DevlBiol. 119, 587-600.

PETSCHEK, J. P., PERRIMON, N. & MAHOWALD, A. P.(1987). Region specific defects in l(l)giant embryos ofDrosophila melanogaster. Devl Biol. 119, 175-189.

POOLE, S., KOUVAR, L. M., DREES, B. & KORNBERG, T.(1985). The engrailed locus of Drosophila: structuralanalysis of an embryonic transcript. Cell 40, 37-43.

POULSON, D. F. (1950). Histogenesis, organogenesis, anddifferentiation in the embryo of Drosophilamelanogaster Meigen. In The Biology of Drosophila (ed.M. Demerec), pp. 168-274. New York: Hafner.

PREISS, A., ROSENBERG, U. B., KIENLIN, A., SEIFERT, E.& JACKLE, H. (1985). Molecular genetics of Kruppel, a


gene required for segmentation of the Drosophilaembryo. Nature, Lond. 313, 27-32.

REGULSKI, M., HARDING, K., KOSTRIKEN, R., KARCH, F.,LEVINE, M. & MCGINNIS, W. (1985). Homeo Boxgenes of the Antennapedia and bithorax complexes ofDrosophila. Cell 43, 71-80.

REGULSKI, M., MCGINNIS, N., CHADWICK, R. &MCGINNIS, W. (1987). Developmental and molecularanalysis of Deformed; a homeotic gene controllingDrosophila head development. EMBO J. 6, 767-777.

RICKOLL, W. L. (1976). Cytoplasmic continuity betweenembryonic cells and the primitive yolk sac during earlygastrulation in Drosophila melanogaster. Devi Biol. 49,304-310.

ROSENBERG, U. B., SCHRODER, C , PREISS, A., KIENLIN,A., COTE, S., RIEDE, I. & JACKLE, H. (1986). Structuralhomology of the product of the Drosophila Kriippelgene with Xenopus transcription factor IIIA. Nature,Lond. 319, 336-339.

SANCHEZ-HERRERO, E., VERNOS, I., MARCO, R. &MORATA, G. (1985). Genetic organization of Drosophilabithorax complex. Nature, Lond. 313, 108-113.

SANDER, K. (1976). Specification of the basic bodypattern in insect embryogenesis. Adv. Insect Physiol. 12,125-238.

SANDER, K. (1984). Embryonic pattern formation ininsects: Basic concepts and their experimentalfoundations. In Pattern Formation, a Primer inDevelopmental Biology (ed. G. M. Malacinski & S. V.Bryant), pp. 245-268. New York: Macmillan.

SCHNEUWLY, S., KUROIWA, A . , BAUMGARTNER, P . &GEHRING, W. J. (1986). Structural organization andsequence of the homeotic gene Antennapedia ofDrosophila melanogaster. EMBO J. 5, 733-739.

SCHUBIGER, G., MOSELEY, R. C. & WOOD, W. J. (1977).Interaction of different egg parts in determination ofdifferent body regions in Drosophila melanogaster. Proc.natn. Acad. Sci. U.S.A. 74, 2050-2053.

SCHUBIGER, G. & NEWMAN, S. M. (1982). Determinationin Drosophila embryos. Am. Zool. 22, 47-56.

SCHUPBACH, T. & WIESCHAUS, E. (1986). Maternal effectmutations affecting the segmental pattern ofDrosophila. Wilhelm Roux Arch, devl Biol. 195,302-307.

SCOTT, M. P. & WEINER, A. J. (1984). Structuralrelationships among genes that control development:sequence homology between the Antennapedia,Ultrabithorax and fushi-tarazu loci of Drosophila. Proc.natn. Acad. Sci. U.S.A. 81, 4115-4119.

SCOTT, M. P., WEINER, A. J., HAZELRIGG, T. I., POLISKY,B. A., PlRROTTA, V., SCALENGHE, F. & KAUFMAN, T. C.(1983). The molecular organization of the Antennapedialocus of Drosophila. Cell 35, 763-776.

SEIFERT, E., ROSENBERG, U. B., PREISS, A. & JACKLE, H.

(1986). Spatial distribution and local requirement forKriippel gene activity in Drosophila embryos. Nature,Lond. 344, 668.

SIMCOX, A. & SANG, J. H. (1983). When doesdetermination occur in Drosophila embryos? Devi Biol.97, 212-221.

SNODGRASS, R. E. (1935). Principles of Insect Morphology.London: McGraw-Hill.

STERN, C. (1968). Developmental genetics of pattern. InGenetic Mosaics and Other Essays, pp. 130-173.Cambridge MA.: Harvard Press.

STRECKER, T. R., KONGSUWAN, K., LENGYEL, J. A. &MERRJAM, J. R. (1986). The zygotic mutant taillessaffects the anterior and posterior ectodermal regions ofthe Drosophila embryo. Devi Biol. 113, 64-76.

STRUHL, G. (1981). A gene product required for thecorrect initiation of segment determination inDrosophila. Nature, Lond. 293, 36-41.

STRUHL, G. (1983). Role of the esc+ gene product inensuring the selective expression of segment specifichomeotic genes in Drosophila. J. Embryol. exp. Morph.76, 297-331.

STRUHL, G. & AKAM, M. (1985). Altered distributions ofUltrabithorax transcripts in extra sex combs mutantembryos of Drosophila. EMBO J. 4, 3259-3264.

STRUHL, G. & WHITE, R. A. H. (1985). Regulation of theUltrabithorax gene of Drosophila by other bithoraxcomplex genes. Cell 43, 507-519.

SULLIVAN, W. (1987). Independence of fushi tarazuexpression with respect to cellular density in Drosophilaembryos. Nature, Lond. 327, 164-167.

TAUTZ, D., LEHMANN, R., SCHNURCH, H., SCHUH, R.,SEIFERT, E., KIENLIN, K. & JACKLE, H. (1987). Fingerprotein of novel structure encoded by hunchback, asecond member of the gap class of Drosophilasegmentation genes. Nature, Lond. 327, 383-389.

TECHNAU, G. M. (1987). A single cell approach toproblems of cell lineage and commitment duringembryogenesis of Drosophila melanogaster. Development100, 1-12.

TIONG, S. Y. K., BONE, L. M. & WHITTLE, J. R. S.

(1985). Recessive lethal mutations within the bithoraxcomplex in Drosophila melanogaster. Molec. gen. Genet.200, 335-342.

TOKUNAGA, C. (1962). Cell lineage and differentiation inthe male foreleg of Drosophila melanogaster. Devi Biol.4, 489-516.

TURING, A. (1952). The chemical basis of morphogenesis.Phil. Trans. Roy. Soc. Lond. B 237, 37-72.

TURNER, F. R. & MAHOWALD, A. P. (1977). Scanningelectron microscopy of Drosophila embryogenesis.II. Gastrulation and segmentation. Devi Biol. 57,403-416.

TURNER, F. R. & MAHOWALD, A. P. (1979). Scanningelectron microscopy of Drosophila embryogenesis.III. Formation of the head and caudal segments. DeviBiol. 68, 96-109.

WAKJMOTO, B. T. & KAUFMAN, T. C. (1981). Analysis oflarval segmentation in lethal genotypes associated withthe Antennapedia gene complex in Drosophilamelanogaster. Devi Biol. 81, 51-64.

WEDEEN, C , HARDING, K. & LEVINE, M. (1986). Spatialregulation of Antennapedia and bithorax geneexpression by the polycomb locus. Cell 44, 739-748.

WEINER, A. J., SCOTT, M. P. & KAUFMAN, T. C. (1984).A molecular analysis of fushi-tarazu, a gene inDrosophila melanogaster that encodes a product

22 M. Akam

affecting segment number and cell fate. Cell 37,843-851.

WEINZIERL, R., AXTON, J. M., GHYSEN, A. & AKAM, M.

(1987). Ultrabithorax mutations in constant and variableregions of the protein coding sequence. Genes andDevelopment 1, 386-397.

WEIR, M. P. & KORNBERG, T. (1985). Patterns of engrailedand fushi-tarazu transcripts reveal novel intermediatestages in Drosophila segmentation. Nature, Lond. 318,433-439.

WHITE, R. & LEHMANN, R. (1986). A gap gene,hunchback, regulates the spatial expression ofUltrabithorax. Cell 47, 311-321.

WHITE, R. & WILCOX, M. (1985). The distribution ofUltrabithorax proteins in Drosophila. EMBO J. 4,2035-2043.

WlESCHAUS, E . , NOSSLEIN-VOLHARD, C. & JuRGENS, G.(1984a). Mutations affecting the pattern of the larvalcuticle in Drosophila melanogaster. I. Zygotic loci onthe X chromosome and the fourth chromosome.Wilhelm Roux Arch, devl Biol. 193, 267-282.

WlESCHAUS, E . , NUSSLEIN-VOLHARD, C. & KLUDING, H .(1984t>). Kriippel, a gene whose activity is requiredearly in the zygotic genome for normal embryonicsegmentation. Devi Biol. 104, 172-186.

WIESCHAUS, E. & RIGGLEMAN, R. (1987). Autonomousrequirements for the segment polarity gene armadilloduring Drosophila embryogenesis. Cell 49, 177-184.

WILSON, E. B. (1892). The cell lineage of Nereis.J. Morph. 6, 361-480

ZALOKAR, M., AUDIT, C. & ERK, I. (1975).Developmental defects of female sterile Mutants ofDrosophila melanogaster. Devi Biol. 47, 419-432.

the molecular basis for metameric pattern in the ... · unravel the role played by each of these...

Documents