4149Development 121, 4149-4160 (1995)Printed in Great Britain © The Company of Biologists Limited 1995DEV7438

headcase, an imaginal specific gene required for adult morphogenesis in

Drosophila melanogaster

Thomas A. Weaver* and Robert A. H. White

Department of Anatomy, University of Cambridge, Downing Street, Cambridge CB2 3DY, UK

*Author for correspondence (email: taw1001@mole.bio.cam.ac.uk)

The majority of adult organs of a holometabolic insect likeDrosophila melanogaster are derived from specific imaginalcells. These cells differ from their larval equivalents inmany important cellular characteristics, ranging from thenature of the cell cycle to the timing and pattern of cellulardifferentiation. Here we describe the cellular, molecularand genetic characterization of a gene, headcase (hdc),which is required for imaginal cell development. hdc is thefirst gene to be described which is specifically expressed inall imaginal cells; this has allowed us to identify manyimaginal primordia in the embryo and follow their mor-phogenesis throughout embryonic and larval development.The Hdc protein is an extremely basic (pI 9.6) cytoplasmicprotein with no obvious sequence similarities or conservedmotifs. Interestingly, the spatial-temporal pattern of hdcexpression prefigures imaginal cell re-entry into the mitotic

cell cycle and persists until the final cell divisions. hdc nullalleles have been isolated and found to cause pupallethality, with dead pharate adults exhibiting defects in thedifferentiation of many adult tissues, most notably in headdevelopment. Ectopic expression of hdc, provided by a hdc-minigene, rescues the pupal lethality. Imaginal disc mor-phology in null mutants appears normal, therefore loss ofhdc expression does not affect imaginal cell growth, butinstead interferes with the ability of the imaginal primordiato differentiate properly during pupal development, sug-gesting that hdc may be involved in hormonal responsive-ness during metamorphosis.

Key words: Drosophila, imaginal primordia, adult morphogenesis,proliferation, hormone, headcase

SUMMARY

INTRODUCTION

During the life cycle of a holometabolic insect like Drosophilamelanogaster the body plan is laid down in two forms, onelarval and one adult (imago). The vermiform larval plan isestablished during embryogenesis and, for the most part, doesnot change significantly in cell number or overall organizationthrough three successive larval instars. In contrast, the majorityof adult organs are derived from distinct lineages of imaginalcells, best seen in the mature third instar (Fig. 1A). These cellsundergo a strict program of proliferation and patterningthroughout postembryonic stages and, through hormonal sig-nalling, are instructed to initiate a series of defined morpho-genetic processes that result in the histolysis of larval organsand the genesis of the adult body shape. (For reviews ofimaginal development see Bodenstein, 1950; Bryant, 1978;Cohen, 1993; Fristrom and Fristrom, 1993.)

What are the embryonic origins of imaginal tissue? Classictransplantation studies in the embryo have shown that theimaginal lineage is allocated sometime between cellular blas-toderm (Chan and Gehring, 1971; Simcox and Sang, 1983) andthe onset of gastrulation (Meise and Janning, 1993). Morpho-logical studies for the presence of the imaginal cells showedthat the thoracic imaginal disc primordia can be detected atprecise positions in the early embryo. For example, each leg

disc primordium first emerges as a cluster of cells whichinvaginate underneath the larval epidermis, adjacent to theCNS and towards the middle of each thoracic hemisegment(Bate and Martinez-Arias, 1991); whereas the wing and halteredisc primordia are located more dorsally, and invaginatedeeper underneath the epidermis (Anderson, 1963; Bate andMartinez-Arias, 1991). Morphological criteria alone are insuf-ficient to identify the other imaginal primordia. Therefore,gene(s) or P-lacZ enhancer trap lines expressed in particularimaginal tissues have been used as molecular markers to tracethe embryonic origins of some primordia including adepi-thelial cells, histoblasts and imaginal discs (reviewed in Cohen,1993; Martinez-Arias, 1993).

One of the most important aspects of the imaginal develop-mental program is the regulation of the cell cycle. Following the16th embryonic mitosis, both imaginal and larval cells (with theexception of certain neural lineages) undergo G1 arrest(reviewed in Foe et al., 1993). Importantly, larval cells never re-enter the mitotic cycle, but instead follow a strict spatiotem-poral pattern of endoreplication, leading to polyploidy. Incontrast, imaginal cells remain committed to a future programof mitotic growth, the timing of which is tightly regulated andvaries between the different imaginal primordia. For example,the onset of division in the eye disc begins 14 hours into the firstlarval instar (Madhavan and Schneiderman, 1977), whereas his-

4150 T. A. Weaver and R. A. H. White

toblast cells remain quiescent throughout larval development, butupon pupariation actively proliferate and replace the histolyzinglarval epidermis by 40 hours postpupariation (Madhavan andSchneiderman, 1977; Roseland and Schneiderman, 1979).

What are the mechanisms that maintain the proliferativecapacity of quiescent imaginal cells and what regulates thetiming of their re-entry into the mitotic cell cycle? Recentprogress on the escargot gene has demonstrated that diploidyof histoblast cells (and possibly other imaginal tissues) is main-tained by the suppression of endoreplication (Hayashi et al.,1993), most likely through transcriptional repression of subor-dinate genes (Fuse et al., 1994). Studies of the anachronismgene (ana), mutations in which cause premature activation ofimaginal neuroblast proliferation in the larval brain, havedemonstrated that ana acts non-autonomously to suppress pro-liferation of imaginal neuroblasts by an unknown mechanism(Ebens et al., 1993). This suppression of neuroblast prolifera-tion is overcome by the activity of the trol gene (‘terriblyreduced optic lobes’; Datta and Kankel, 1992), which actsdownstream of ana possibly by inactivating the ana repressoror bypassing its requirement (Datta, 1995).

Strict control of imaginal cell proliferation is necessary for syn-chronizing tissue growth with the hormonal signals initiatingmetamorphosis. This link between imaginal proliferation and theendocrine system is best illustrated in tumor suppressor genemutations that result in imaginal disc overgrowth (Bryant andSchmidt, 1990). In general, these mutants exhibit delayed pupari-ation and an associated reduction in ecdysteroid titer (Sehnal andBryant, 1993). A similar developmental delay is observed whenimaginal disc tissue regeneration is induced in situ by heat-sensitive cell-lethal mutations (Simpson et al., 1980). These andmany other observations have led to the model in which prolif-erating imaginal disc cells control the timing of differentiation byinterfering with ecdysone release (Poodry and Woods, 1990);however, the molecular mechanisms remain unclear.

Imaginal development is a powerful model system in whichto study the fundamental relationships between primordial cellgrowth regulation and morphogenesis during organogenesis.Our interests are focused on the problems of how imaginal andlarval cell growth are differentially regulated, and how this islinked with the timing of adult organ differentiation. Here wereport the molecular genetic characterization of a gene calledheadcase (hdc), which is expressed in all proliferating imaginallineages. This specificity has allowed us to identify theembryonic origin of different imaginal primordia and followtheir lineage throughout embryonic and larval stages. hdcexpression prefigures imaginal cell re-entry into the mitotic cellcycle and continues until the end of primordia growth. Nullalleles have been generated and found to cause a pupal lethalphenotype associated with the developmental arrest of manyadult organs, all aspects of which can be rescued by express-ing Hdc protein from a transgene. Interestingly, the mor-phology of mutant discs appear normal, suggesting thatimaginal growth regulation is unaffected; instead, the defectappears to be in the ability of the imaginal cells to undergoproper differentiation during metamorphosis.

MATERIALS AND METHODS

Stocks The wild-type stock used was Oregon-R, obtained from John Roote.

TM2Z is a third chromosome ‘blue balancer’ harbouring an enhancertrap that expresses lacZ uniformly in the epidermis following germband retraction (called C40.1S3 in Bellen et al., 1989), obtained fromCahir O’Kane. The P[lacZ; ry+] transformant line, B5, was originallyselected by Alex Gould from a collection of lacZ enhancer traps,generated by John Merriam using the strategy of O’Kane and Gehring(1987). Marker mutations are as described in FlyBase (FlyBase,1994).

DissectionsEmbryo fillets were performed as described in White and Wilcox(1984). Embryonic stages of development were scored according toCampos-Ortega and Hartenstein (1985). Larvae were collected asnewly hatched first instars and then aged on yeasted apple juice plates,and appropriate larval stages were scored by morphology of mouthhooks as described in Ashburner (1989). Internal tissues weredissected from bisected and inverted half-larvae. For epidermis prep-arations larvae were heat killed 5 seconds at 55°C before dissectionin order to relax the body wall musculature. Pupae were collected aswhite puparia (stage P1, stages described in Bainbridge and Bownes,1981), then aged 4-5 days at 25°C on water-saturated Whatman paperkept in a Petri dish. Pharate adults (stage P14/15) were dissected frompupal cases pinned out on a Sylgard-filled Petri dish.

X-gal stainingEmbryo whole-mount and flat preparations were X-gal stained asdescribed in O’Kane and Gehring (1987). Devitellinization wasperformed using ethanol rather than methanol to avoid inhibition ofβ-gal activity. Larval tissues were stained as described in Prokop andTechnau (1991).

Cloning and sequencing the hdc geneA 450 bp junction fragment (JF, Fig. 3A) flanking one end of theP[lacZ; ry+] element was isolated from a B5 genomic library, madein lambda Zap (Stratagene), and then used to initiate a 65 kb chromosome walk surrounding the insertion site (EMBL3 genomiclibrary kindly provided by John Tamkun). Overlapping DNAfragments of the walk were used for mapping of the hdc transcript byin situ hybridization (see below). Three probes (grey boxes, Fig. 3A)detected a transcript with an identical distribution pattern to B5-lacZ.Probe G′ (Fig. 3A) was used to isolate two overlapping hdc cDNAs,hdc.NB40 and hdc.K, from two different 12-24 hour embryoniclibraries (plasmid library kindly provided by N. Brown, described inBrown and Kafatos (1988); gt10 library kindly provided by L.Kauvar). Double-stranded sequence analysis was compiled usingStaden and GCG software. Searches for nucleotide and protein simi-larities between the full-length hdc sequence and known genes in theGenEMBL, OWL and Prosite databases were performed usingFASTA, FLASH, BLAST, BLOCKS and MotifFinder programs.

In situ hybridizationDigoxigenin-labelled DNA probes (Boehringer Mannheim) weremade from gel-purified restriction fragments (varying in length from2-8 kb) spanning the chromosome walk. Digoxigenin-labelled RNAprobes (Boehringer Mannheim) were transcribed in vitro fromphdc.NB40 using T7 (anti-sense probe) and SP6 (sense probe) poly-merases. Hybridizations were performed as described in Gould andWhite (1992), with DNA probes at 0.5 µg/ml and RNA probes at 0.1µg/ml (with respect to starting template).

Generating hdc alleles and heat-shock transformantsAll hdc deficiencies were derived from an isogenic B5 line byimprecise excision of the P-element. This was achieved by crossingin a source of transposase (Robertson et al., 1988) and resultingdysgenic male progeny were individually crossed to ry506/TM3 Sb rytester females. 150 individual ry progeny were back-crossed to testerfemales in order to isogenize each excision chromosome. 8 lethals

4151Adult morphogenesis in Drosophila

were obtained and complementation tests revealed that all 8 lethalsfell into the same complementation group. Breakpoints in hdc alleles43, 50, 97 and 117 were mapped by genomic Southern blot hybridiz-ation as described in Daniels et al. (1985) using radioactive probesmade from the chromosome walk. The HS-hdc vector was constructedby subcloning the hdc.NB40 cDNA downstream of the heat-induciblehsp70 promoter in the P-element vector pCaSpeR-hs (kindly providedby Carl Thummel) which has a mini-white+ selectable marker. Trans-formants were made using P-element-mediated germline transforma-tion of y w host embryos as described in Spradling (1986). Six inde-pendent white+ transformant lines were isolated, three of which (lines38.5, 77.1 and 87.1) mapped to the second chromosome and werecrossed into a hdc43 ry/TM3 Sb ry background. One of these lines,87.1, completely rescued the pupal lethality, even without heat shock,as measured by the presence of ry Sb+ progeny from the cross:

y w; P[wm+; HS-hdc]87.1; hdc43 ry/TM3 Sb ry males ×y w; P[wm+; HS-hdc]87.1; hdc43 ry/TM3 Sb ry females

Analysis of mutantsPupal lethalityOvernight egg lays were collected in yeasted agar vials, allowed topupate, and then scored for the number of eclosed or dead pupae onceno more emerging flies were observed. Alternatively, individualcohorts of newly pupariated animals from large bottle cultures weretransferred to Petri dishes as above and then scored for pupal lethalityafter 1 week.

HistologyDissected pharate adult heads were wax embedded and 8 µm serialhistological sections were cut with a Reichert manual microtome.Sections were stained with Harris’ haematoxylin and eosin.

Scanning EMDissected pharate adults were fixed overnight at 4°C in 0.1 M Pipes(pH 7.4)/2% glutaraldehyde/2% formaldehyde. Fixed specimens weretreated with 1% osmium tetroxide, dehydrated in a graded series ofalcohols and critical point dried. Specimens were then coated with 20nm of gold and viewed in a JEOL JSM 35 microscope.

Monoclonal antibody preparation and antibody stainingA 1.25 kb fragment from phdc.NB40, encoding 417 amino acids ofHdc protein (underlined residues in Fig. 3B), was subcloned intopRSETC (Invitrogen) and pGEX-3X (Pharmacia) bacterial expressionvectors in order to generate unique fusion proteins of relativemolecular mass (Mr) 54×103 and 71×103 respectively. The 54×103

protein was used to generate mouse hybridomas as described in Whiteand Wilcox (1984). Supernatants were then strip western blot screenedfor reactivity against Hdc-specific epitopes present in the 71×103

fusion protein. Antibody staining was as described in White andWilcox (1984), and involved: mouse anti-β-gal (Promega); mouseanti-Hdc (this report); rabbit anti-Twist (kindly provided N. Brown)rabbit anti-Snail (kindly provided by J. Castelli-Gair); mouse anti-Grainyhead (kindly provided by S. Bray). Secondary antibodies wereobtained from Jackson ImmunoResearch Laboratories. Fluorescencemicroscopy was carried out on a Leica TCS confocal microscope.

RESULTS

The B5 enhancer trap is a marker of imaginaltissues of the developing Drosophila larvaIn an enhancer trap screen for genes potentially regulated byhomeotic genes of the bithorax complex (Gould, 1990), wediscovered one transformant line, B5, where lacZ is specifi-cally expressed in imaginal cell populations of the maturelarva (Fig. 1B-N). At this stage, the imaginal cells can be dis-

tinguished from their larval counterparts by a characteristicsmall cell size, tissue morphology and positioning within thelarval body plan (see schematic, Fig. 1A; Bodenstein, 1950;Bryant, 1978). In addition to the classic imaginal ‘discs’, theB5-lacZ pattern also highlights the imaginal precursors ofgut, genitalia, CNS, respiratory system and epidermis. Weconsider each aspect of this expression pattern in detailbelow.

The B5-lacZ pattern in the ventral nerve cord correspondsto the distribution of imaginal neuroblasts (Truman and Bate,1988), as well as their postembryonic lineages, with manymore expressing cells in the thoracic than in the abdominalneuromeres (in, Fig. 1N). The posterior tip of the ventral nervecord also has many more lacZ-expressing cells than inabdominal neuromeres. Interestingly, it is this region wheresexually dimorphic neuroblasts are proliferating (Truman andBate, 1988) and most likely represent the precursors of neuronsdestined to innervate the genitalia. The larval pattern dimin-ishes during the early stages of pupal development, and isreplaced by strong expression in the developing central brainand optic centers (data not shown).

In the larval respiratory system, B5-lacZ expression isobserved in imaginal tracheoblasts (Whitten, 1980; Manningand Krasnow, 1993), which are easily identified by their smallsize compared to neighboring larval tracheal cells and closeassociation with the larval tracheal network. Two types ofclusters are seen, one round and one elongate. Round clustersare found in thoracic segments only, and are situated next tothe dorsal longitudinal trunk (Fig. 1D). Elongate clusters arestretched along the visceral and spiracular branches, whichemanate from the transverse connectives (Fig. 1F) and are seg-mentally repeated down the length of the larval trunk. Theother respiratory-associated cell types represented in the B5-lacZ pattern include the dorsal prothoracic discs associatedwith the anterior spiracles (Fig. 1B) and the spiracular histo-blast nest found in the epidermis of abdominal segments A1-A7 (Fig. 1I).

Larval glands that express B5-lacZ include the imaginal ringof the salivary gland (Fig. 1C), the ring gland (Fig. 1N), thelymph gland and pericardial cells flanking the length of theaorta (both not shown). Alimentary expression includes theforegut and hindgut imaginal rings, and midgut imaginalislands (Fig. 1J,M). B5-lacZ is also expressed in the larvalgonads. In the male, B5-lacZ is found in apically associatedcells of the testis (Fig. 1G). These cells correspond to the mitot-ically dividing gonial cells (Cooper, 1950; Fuller, 1993) andremain positive in the adult male, where expression is con-centrated at the anterior tip of the testis (data not shown). Theterminal cells (Fig. 1H) at the posterior tip are of unknownidentity; however, they continue expressing B5-lacZ duringpupation as this group of cells grow out towards the develop-ing genital disc, presumably forming the seminal duct epithe-lium (Cooper, 1950). In the female, B5-lacZ is expresseduniformly throughout the ovary (Fig. 1L).

The embryonic development of imaginal primordiacan be followed using the B5 markerSince the lacZ expression pattern of the B5 line is a uniquemarker for larval imaginal cells, it should allow us to identifythe embryonic origins and morphogenesis of these tissues. Wefocus here on the development of imaginal structures in

4152 T. A. Weaver and R. A. H. White

4153Adult morphogenesis in Drosophila

Fig. 1. Imaginal tissue specificity of the enhancer transformant lineB5. Wandering third instar larvae from the transformant line B5 werestained with X-gal in order to detect the nuclei (blue) of lacZ-expressing cells (B-N). (A) Schematic illustration of a third instarlarva dissected along the ventral midline and laid flat on the right,with tracheal trunk pulled towards the right hand edge. Imaginaltissues are highlighted blue. (B) Dorsal prothoracic imaginal discsurrounded by the anterior spiracle. (C) Salivary imaginal ring at thejunction between the salivary duct and salivary gland. (D) Clusters ofimaginal tracheoblasts clinging to branch points along the larvaltracheal tree in thoracic hemisegments. (E) Imaginal discs: w, wing;l3; metathoracic leg and h, haltere. (F) Imaginal tracheoblasts, as in

D, illustrating the segmentally repeated nature of the elongatedclusters. (G) Gonial cells and (H) terminal cells of the testis. (I) Histoblast nests: v, ventral nest; da, dorsal anterior nest and dp,dorsal posterior nests; sp, spiracular nest. (J) Hindgut imaginal ring.(K) Genital imaginal disc. (L) Ovary. (M) Foregut imaginal ring, fr,located anteriorly along the inner surface of the proventriculus, andsurrounding the oesophagus. The midgut imaginal islands, mi, areinterspersed as in J, but not in the gastric caecae, gc. (N) Imaginaltissue associated with the larval CNS and head skeleton: e, a, eye-antennal disc; l1, prothoracic leg disc; l2, mesothoracic leg discs; lb,labial disc and cl, clypeolabral discs; in, imaginal neuroblasts; r, ringgland.

Fig. 2. B5-lacZexpression patternreveals the embryonicorigin of imaginalprimordia. HomozygousB5 embryos were stainedwith anti-βgal antibodiesas whole-mount (A-F) orflat preparations (G,H).(A,B) Ventral and lateralaspects of stage 13embryos highlighting theearliest detectablepattern. Prominentfeatures include: acluster of neural cells ineach thoracichemisegment of theCNS, white arrow, thesepersist throughout earlydevelopment and formthe imaginal neuroblastpopulation; a stripe ofepidermal cells adjacentto the posterior edge ofthe T1 segment, openarrow, which gives riseto the prothoracic disc;three leg disc primordia,arrowheads; a pair ofepidermal cells, asterisk,of unknown identityrepeated from A1-A7.(C,D) Ventral and lateralaspects of stage 15embryos illustrating thedynamic change inexpression pattern,especially in the CNSwhere the number and intensity of staining cells increases. The three leg disc primordia have increased in staining intensity and are shiftinganteriorly as a consequence of head involution, this is much more apparent in F. The genital disc primordia, gd, is detected posterior to theventral nerve cord, slightly beneath this focal plane. The prothoracic disc primordia has invaginated along with the anterior spiracle (also seeF,G). Cells from the T2 and T3 leg primordia have begun migrating dorsally, these cells will populate the wing and haltere disc primordia(arrowheads in D, also see F,G). Intense staining in the brain lobes, oesophageal ganglia, clypeolabral, labial and eye-antennal primordia can bediscerned in deeper focal planes (data not shown). (E,F) Ventral and lateral aspects of stage 17 embryos displaying the mature embryonicpattern of imaginal primordia. The leg discs have migrated further anteriorly, forming tear-drop-like pockets stretching dorsally towards thewing, w, and haltere, h, disc primordia. Each abdominal segment exhibits two additional groups of cells: a line of imaginal tracheoblasts, tr, anda second pair of cells of unknown identity, asterisk, just internal to the epidermal cells seen in B and D. Staining is also seen in pericardial cells,lymph gland, and testis (data not shown). (G) Stage 17 embryo with CNS removed to reveal all thoracic imaginal disc primordia: prothoracic,p; wing, w; haltere, h; prothoracic leg, l1; mesothoracic leg, l2; metathoracic leg, l3. (H) Internal organization of abdominal segments in a stage17 embryo. As in F, each segment contains a string of approximately 8 imaginal tracheoblasts, tr, attached to the visceral branch of the dorsallongitudinal trunk, and two sets of 2 cells, each of unknown identity, one set sitting on the ventral internal oblique muscles, asterisk, and theother pair located in the epidermis underneath this focal plane.

4154 T. A. Weaver and R. A. H. White

1 GAATTCCCCCCCCAGTTTGAATTTCGCTTTCGATCTGGTTGGTTGTGTTTTTTGGCTTCT 61 CTCTGCGTGCGCCGCTCGTGGTGGGAAAAAACGGAAAACAGGAGAACTAGCAAAACCGAA 121 TAAATCGAAGAGGAACACTCCAGTTTACAATTGTCCTGCGACTAAAAACCAATCGGAAAT 181 TTACACTATCCTCTAATCCTTTTTGTTTGCCGGTAAAAAAAAATAAGAGGCGCAATAAAA 241 TAAACGCAGAAAGCGAAGCAAAATGTGAAAAATTGCCGAACATGGCGCCGCGTGTGTGAA 301 AGAGAGCGAACGATCGTGTGAGAGGGAGATAGGGACGAAGCGAAGATAAACAAGATAAAT 361 AATAAAAAAAAAAACAATTTTTGATACTTAAAGTTCTACAACAACAATAAACTCAGCTCA 421 GGTACAACAACAAAACAGAAAATTACGCAAAACACAACAAACTGCATTGACGGGAATTTT 481 TCGAAGTATTTTTGATACACACAGCAAGGAGGTGTACGATTTTTAAAGCAATAAGTGAAA 541 GCAATTACCCATCAAGTTAAAAATAATACCAATCTCGGTAGAAACCCACTCGGAAAATCC 601 ATCTGAAAACAATTGAAAAAGGCAACCGAAAAGTTAAAACCTAAACAGACTAAATCGCTA 661 AAATAACAAGCATTTACTATACACAAAAAAATTTAATTCCGTATAAAATCAGACTTTTGT 721 CTGCAAAATTTCTAAAAATTTACAAAAACCCACTAAAATTCACGAAAACTCACTGGATCT 781 GCCAAAACCATCAGTTGAAATATATTCTGGAATCGGTTCAAAAATAAATAAATATATGCC 841 AGCAGCATTTATATAGAACGCAAATCGCACCATAAATCGCGAAAAATCAACAAGCAATAT 901 TTTGAAATTCATAATCGTTGGCAAAATGCTGATTCGAAATAGAGATTGAGGATTCTCGAG 961 GACAGCCAACCCAGCCAACCAACCAAGTTTTTCGCCGCAACGAGTGCATATTAAGAGGAA

1021 AAAACACAAATTTATGGCTCCGCGTCGCAACAGCAACATCAGCAGCAGCGGCAACAGCAC 1 M A P R R N S N I S S S G N S T 1081 CCAGCAGCAGCACCAGCAACAGTTGCAGCAGCAGCACCAGACGCAGCAGCATCAACTGCT 21 Q Q Q H Q Q Q L Q Q Q H Q T Q Q H Q L L 1141 GCAGTTCTACGCGGAGAACGTGAATTCTTCAGGCGTCCTAATGGCGCCCATGCAGGGAAC 41 Q F Y A E N V N S S G V L M A P M Q G T 1201 CGGTGGCGTTGGTGGCGTTGGCGGCGGGGGCATGGTCCACTGCTGCCTGCCCAGCGGAGA 61 G G V G G V G G G G M V H C C L P S G D 1261 CTGTCGCAAACTGGACACCCTGATACCCCTGAACGAGCTCTCCCTGGCGGACTGCATCCG 81 C R K L D T L I P L N E L S L A D C I R 1321 TGTGCTGTGCAACAACGAGAACTGCAGCGCGGGCCAGTACATGCATCGCGAGTGCTTCGA 101 V L C N N E N C S A G Q Y M H R E C F E 1381 GTGGTGGGAGGCGAGCGTGCTGCAGGCACTGAAGGCCAACGGCCGTGCTCGATCCTGGTC 121 W W E A S V L Q A L K A N G R A R S W S 1441 GGAGCGGCAGCGTCTGCAGCACCTGTGGACCAAGAAGGGCTACGAGCTGGTCCAGAAGGC 141 E R Q R L Q H L W T K K G Y E L V Q K A 1501 CTGCAGCTGCAAGTGCGGCAGGGGACAATTGCGCAAGGACCTCGACTGGGTGCCGCCGAG 161 C S C K C G R G Q L R K D L D W V P P S 1561 CAGCCAGGGCGTCATATACCTCAACGGGTCTGGCAACCGGTCCAATCTGGCCAATGGCAG 181 S Q G V I Y L N G S G N R S N L A N G S 1621 CCTCAGCGAGGACGACGACAAGAAGAAGGCCAAGAAGAAGAGGAACCGCAACAACAACAA 201 L S E D D D K K K A K K K R N R N N N N 1681 CAACAACGGCGGCGGCGGCGGTGGTGGTGGTGGTGTCAATGGCAACACCAAAACCCCCTT 221 N N G G G G G G G G G V N G N T K T P L 1741 GAGCAACAACAACGGCAACCATGACGCCGGACTCACGCCCAATCCCAATGTGGGCGTCAT 241 S N N N G N H D A G L T P N P N V G V I 1801 AGGGTCGGGTTTGCCCCACAACAACGGCAACACCGCCAGCAATGGCAGCAGTGGCAACAA 261 G S G L P H N N G N T A S N G S S G N N 1861 CGGCTCTCCGGCGTGCTGCAGACCAGTGCGCTGGCCACTTTCAGCAACATCCTCAATACG 281 G S P A C C R P V R W P L S A T S S I R 1921 AACAATGTCTTGGCCTGGACCTGCGCGCCAGGGCTGGCAGCTATCCTCCAGTGGAGCTGG 301 T M S W P G P A R Q G W Q L S S S G A G 1981 CAGTGGATCCACCTCGCCCTCGGACTCCCAGTCCAGTGGCGAGATATCCGTGTCGCCGGT 321 S G S T S P S D S Q S S G E I S V S P V 2041 GCAGCAATTGCTCGCACAGCAGCAGCAACAACAGCAGCAGTCGCTCGTGATCCAGCCTCT 341 Q Q L L A Q Q Q Q Q Q Q Q S L V I Q P L 2101 GGGTCCAGCTTTCGGCAGCAACCTTTTGCAAAATGGATTGGGTTTGTCCAAGAATCTGCT 361 G P A F G S N L L Q N G L G L S K N L L

2161 GATTGCTCCCCAGCAGCCGCAGCAGCAGCAGCTGCAACAACAGCAGCAGCAGCAGAATAA 381 I A P Q Q P Q Q Q Q L Q Q Q Q Q Q Q N N 2221 TCTGCCTGCTTTGGCCAATATCAGCAACTTCAAGCCATTGGCCAGCTACGAGCAGCAGCA 401 L P A L A N I S N F K P L A S Y E Q Q Q 2281 ACTTGTGCAGCAGCAGAAGAACAAGGAGGTCGAACTCTATTCCGAGCGAGTGCGCTCCAC 421 L V Q Q Q K N K E V E L Y S E R V R S T 2341 ATCTGGCTGCAATGGGATCTTTTCGCGTCGCCTGGACTTTTCGTCGTTCAATTTGCTGCC 441 S G C N G I F S R R L D F S S F N L L P 2401 AAAGACTCGTTTGAATTCTTATCAAGTCAAGATTGAGGACGAGGGCAACCACGGCAACGA 461 K T R L N S Y Q V K I E D E G N H G N D 2461 TGAGACGCGCCTGTTCATCCTCTCCTCGCTGGCCCAATCCCAGATGAGCCGGGTGGCCTG 481 E T R L F I L S S L A Q S Q M S R V A C 2521 CATTCTGTGCGAGGAGCCGCTGCTCGTATTCGATCGCTATCCACTGGTCGATGGCTCCTT 501 I L C E E P L L V F D R Y P L V D G S F 2581 CTTTCTGAGTCCCAAGCAGCATTCCAGCGGCTGCATCGAGGTTAAATACGAGGGACGCAC 521 F L S P K Q H S S G C I E V K Y E G R T 2641 TCTCTACTTGACCTGTGTTTGCATGAGCTGCCTGGACGGAACCTCCAGCAGCCGGGCCAT 541 L Y L T C V C M S C L D G T S S S R A I 2701 CAACTGCAGATTCTGCAAGGAGCCGTGGGATGGCTCCAGCCTGGTTTTGGGCACCATGTA 561 N C R F C K E P W D G S S L V L G T M Y 2761 CGCCTACGACATCTTTGCGGCCATGCCCTGTTGCGCCGAGCGGTTCAAGTGCAATAACTG 581 A Y D I F A A M P C C A E R F K C N N C 2821 CTTCAAGATGCTAATGCATCCGCAGCAGCGTCTGAGCTTCTACAGTGACTACTCGCACGG 601 F K M L M H P Q Q R L S F Y S D Y S H G 2881 AGTGACCTGTCCGTATTGCAACACGCAGGACACGCACTTCGTTAAGCCGCTGACCTTCTG 621 V T C P Y C N T Q D T H F V K P L T F C 2941 CTACGCCAAGAGCACGGCGACGCGGCTGCCGACGCTCGCATAACATGAGGCCCCGTTGGA 641 Y A K S T A T R L P T L A * 3001 GTCGCCCGTGGGAACAGGTGCCACCACCACTCAGGTGCCAAATGCCCAGGGCTCTCCAAC 3061 GGCCTCCGGCTGCAGCAGCAACACCATCGCCAGCAAGCAGCCACCGAAGCAGCAGCTCTA 3121 CAATGCCGCCATCGATTACTCGGCACCGCTGCAGCAGCAAATTAATCCGGACTTGATTGG 3181 AGCCCATCCGCATGCCCAGCAGCATCAGCAGCAGGTGCGGCAGCAGCAACAGCAGCAGCA 3241 GCAGCAGCAACCACAGCAGCAACAACAACAGCAGCAGCAACAGACACAACAACAACAGTC 3301 ACAACAACAACAGCAGCAGCAGCAGCAGCAACATCAGCCTGTGGTCAGCACTTTTCAGAG 3361 GGGCAACTTCCCTCAACAGTCCCATAAAACCATGAACATGATGGCCATGGCGGATGTGAT 3421 GAGCAAGTACCAGCAGCAGCAGCACCAACAGCAGCAGCAACAGCGCCAGCAGCAACACAA 3481 CCTGCAGCCACAGCAGCAACATGCAACGCAGAAAGGACTCGAGGGCCTTCTATTGACCAA 3541 CACAAACACAAGCAACAACAATAGCAGCAGCAGCAGCAGCAGCAGCAGTATTAGTAACTT 3601 AATCATTAGCCATAATTCAGCCAGCAATACAGGCAACAAAATGCAGCCCACTTGGGGCCA 3661 GAGTGATGCCATCTCGGCATTGTGGGGCTGCTCCAGCAGCAGCAGCGGCTGCTCCTCGGG 3721 CAGTGGATCGCAGCCCTCGTTGAGTCCGACGGCCAGCAGCAATGGCAACGATGGCTCCAA 3781 GATGATGTTGTGGGCGGCCCAAAGTTCGCCGCAAAATGCTGCAACACTCAGGGATGGATT 3841 CAAGGAGTTGCAGCGCCAGCAGCCACCGCAGCAGCAGGTGCCACAGCAACAGCCACATGC 3901 AGCCTCGCCCACAGCTTCGCTGACCAGTTCCAGTTCCTCCTCGAACGGATGGAGCGCCTC 3961 GCATTATGGCAAGCCCAACATCTGGGAGCTGCCAGGATCCACCGGGCAACGGACACCGGC 4021 CAGCAGCCACGACATCTTCACCGATCTCCTGTGCAACCTGAGCATCAGTCAGGACAACAG 4081 CAGCCAGCAGGCCAGCAGCAAGGCCGATGCCTCCGACGTCAGCTCGAATTCCTCCTCGTC 4141 GGCCGGGGAACTCCTGGAGGCGGCCAATATCTGGAGGTTTCCAGAGTACGCCAGTAGTCA 4201 GCTCTACATGGAGGAGCCGACTGGTGGCGCACGTGCCTGCATGCAGCTCTTCAACGACTA 4261 TCTCAACATGAACTAAAACGAAGAATTGGGAGCCACAACAGCCTGCCGCATTTGTTACTC 4321 TAGCTAAATTGAAAAACAAAAATTGAGGAAAAAGGAGTAAGCAGGCGGAGGAGAGGCCGT 4381 ATACGAGCCACTCTCCTGAAGCGAGTGTTTTAGTTTTTTAAAATCATTTGATTTTTTTTA 4441 TATACACACACACAACACACATCTAAAACCGAGAGAAAACAAAGCAAACTAAAAACGAAT 4501 TTCTTGTTATTTTTGTTGAGTTTTTTAGTTCTTTGTTTACACACACCTACCTAAACGAAA 4561 AAAAAAGCAATTTTTAGTTTTATTTTATAAAGCAAATTTTTTATACATTTTACCAAAAAA 4621 AAAAAAACGAATGCTGCGGCCGCCGAATTC

hdc97, hdc117

hdc43, hdc50

G

p3.7

H

F

B B B S B/RRRRSS R

25 30 35 65 kb

P[lacZ;ry+]

0 kb JF

hdc.K

hdc.NB40

(S)

5'

5'

3'

3'

G'

Fig. 3. Molecular analysis of the headcase gene. (A) Schematic summarizing how the headcase genewas identified and mutated. The junction fragment,JF, 65 kb chromosome walk, and B5 enhancer trapelement P[lacZ;ry+] are included on a thick black linemeasured out in 1 kb units (below) and selectedrestriction sites (above). Boxes F, G, G′, p3.7 and Hare representative DNA probes used for mapping theheadcase transcription unit; those filled grey gavepositive signals. Black boxes represent hdc cDNAs.Thin black lines underneath genomic walk representthe breakpoint mapping of four independent hdcdeletion mutants (break in line), allele names areindicated to the right. (B) Nucleotide and theoreticaltranslation of composite hdc cDNA sequences, withhdc.K beginning at nucleotide 14 and ending at 1371,and hdc.NB40 beginning at nucleotide 356 andending at 4627. Two large ORFs are separated by asingle amber stop codon (boxed) beginning atnucleotide 2980 (sequences encoding the downstreamORF are shaded and end in another amber stopcodon, also boxed). Underlined residues from theupstream ORF were used to make Hdc antibodies.Accession number Z50097.B

A

4155Adult morphogenesis in Drosophila

thoracic and abdominal segments, although we have also beenable to identify the eye-antennal, clypeolabral and labial discprimordia (data not shown).

The rudiments of the strong staining in the thoracic segmentsof the CNS are discernible by stage 13 of embryogenesis (Fig.2A) and mature as the ventral nerve cord condenses (Fig.2A,C,E). This pattern prefigures the larval pattern of postem-bryonic imaginal neuroblasts (IN), which is unusual since otherIN markers, such as Grainyhead protein (Bray and Kafatos,1991), do not show segmental modulation until the very lateststages of embryogenesis.

As in the CNS, the pattern of B5-lacZ expression in theepidermis differs significantly between thoracic and abdominalsegments. The reason for this is the presence of the imaginaldisc primordia in thoracic segments, the relative positions ofwhich are shown in Fig. 2G. The ventral discs (arrowheads)and the dorsal prothoracic disc (open arrow) are the firstdetected following germ band retraction, early in stage 13 (Fig.2B). The wing (w) and haltere (h) primordia appear later duringhead involution and dorsal closure, as cells appear to migratedorsally from the ventral discs and form small pockets clingingto tracheal branches (Fig. 2F,G). This migration has also beenobserved for Distalless-lacZ-expressing cells (Cohen et al.,1993). The dorsal prothoracic disc primordia, initially adjacentto the posterior T1 segment boundary, rounds up and invagi-nates with the anterior spiracle during head involution (Fig.2B,D,F).

A very different pattern is seen in abdominal segments.Three groups of strong lacZ-staining cells are reiterated in eachsegment: (1) the imaginal tracheoblasts (tr, Fig. 2F,H), whichare situated along the same position in the embryonic trachealnetwork as the elongate clusters described in the mature larva;(2) two epidermal cells of unknown identity (asterisk, Fig.2B,D,F) and (3) two cells located on the ventral internaloblique muscles (asterisk, Fig. 2H), also of unknown identity.

We also see B5-lacZ expression in the location of the genitaldisc primordium (gd, Fig. 2C,E; Hartenstein and Jan, 1992;Whiteley et al., 1992). Expression appears in three cellclusters: two transverse stripes of cells, one on each side of themidline and one, more posterior, median cluster. These clusterssubsequently associate together presumably to form the fusedgenital disc. We do not yet know the relationship of the threecell clusters to the development of the adult analia and genitaliaor to the somewhat similar three clusters found in other flies(Dubendorfer, 1971).

Identifying and cloning headcase, the generesponsible for the B5-lacZ expression patternThe B5 line harbours a single insertion of the P[lacZ;ry+]transposon at cytological position 99F (data not shown). Inorder to characterize the gene responsible for the B5-lacZpattern, which we have named headcase, (hdc), we screenedneighboring genomic DNA first for hdc transcription units andthen for hdc cDNAs (see Methods). The composite sequenceof two overlapping cDNAs is shown in Fig. 3B and predicts amessage of 4.3 kb. This is consistent with a single 4.3 kb tran-script detected by northern blot analysis of embryonic mRNAwhen probed with the nearly full-length cDNA, phdc.NB40(data not shown). Importantly, this same probe recognizes thecomplete B5-lacZ expression pattern in the embryo (compareFig. 4A and B) and larva (not shown). Southern blot mappingof the cDNA sequences onto the genomic walk (black boxes,Fig. 4A) revealed that the first 2/3 of the transcript mapsdirectly downstream of the insert, whereas the remaining 1/3is distal to the DNA isolated in the walk. The unmapped 3′domain appears to be a bone fide part of the hdc mRNA, notonly because of the correlation with transcript size, but alsobecause in situ probes made from these sequences detect thecomplete B5-lacZ embryonic expression pattern (data notshown).

Fig. 4. Embryonic headcasegene expression mimics B5-lacZ. Flat preparations ofstage 16 embryos. (A) Anti-βgal staining of B5transformant line. (B) In situhybridization of wild-typeusing an antisense RNAprobe transcribed fromhdc.NB40. (C) Anti-Hdcstaining of wild-type. (D) Double labelling of wild-type imaginal primordia withanti-Hdc and anti-Sna, amarker of the wing andhaltere imaginal primordia(Alberga et al., 1991). High-power magnification revealsnuclear staining Snail andcytoplasmically staining Hdcproteins in the same cells ofthe wing, w, and haltere, h,imaginal disc primordia; Hdcstaining alone is present inthe two leg disc primordia,l2, l3.

4156 T. A. Weaver and R. A. H. White

Fig. 5. Confirmation of headcase null alleles. (A) Double stainingwith X-gal and in situ hybridization of a stage 15 embryoheterozygous for hdc43 and the third chromosome blue balancerTM2Z. This embryo stains positive for both β-gal (blue) and hdcRNA (purple); whereas in B a homozygous hdc43 mutant embryo,identified by the lack of X-gal staining, does not express any hdcRNA.

The Headcase protein is a novel cytoplasmic proteinpresent in embryonic imaginal primordiaTheoretical translation of the hdc cDNA is shown in Fig. 3B.Surprisingly, there are two large open reading frames (ORF)separated by a single in-frame stop codon. PCR analysis of thegenomic sequence verified the existence of the stop codon(data not shown). The upstream ORF predicts a protein with aMr of 71×103; whereas translation of the downstream ORF,either by suppressing termination or by internal initiation,predicts proteins of Mr 125×103 or 35×103, respectively.Mouse monoclonal antibodies were raised against a portion ofthe upstream ORF (underlined residues in Fig. 3B). Antibodystaining with this monoclonal detects a protein expressed in theimaginal cells of the embryo (Fig. 4C) and larva (data not

Fig. 6. Pupal lethal phenotypeof headcase mutants. Wild-type and homozygous hdc43

mutant pharate adults weredissected from their pupalcases and assessed for defectsin head structures. (A) Wild-type head and thorax. (B,C)Mutant animals show grosscuticular defects: in B, theleft side of the head capsule isreplaced by an unknownstructure (arrow) and, in C,the whole head capsule ismissing with only theproboscis (arrow) extendingfrom the thorax. (D) Wild-type and (E) mutant animalsviewed by scanning EMreveals that the mutant has athird antenna (ant, arrows)that replaces half the headcapsule. (F) Transverse waxthin section through the headsof wild-type and (G) mutantanimals showing a grossdisruption of internalstructures including: m,musculature; ts, tracheal sacs;cns, central nervous system;oc, optic center.

4157Adult morphogenesis in Drosophila

shown), mirroring the exact patterns observed with B5-lacZand hdc in situ staining (compare Fig. 4C with A,B).

High-power magnification of antibody-stained tissuerevealed that the Hdc protein is located in the cytoplasm. Wetook advantage of this property in order to establish the per-centage of Hdc-expressing cells in embryonic wing and haltereimaginal disc primordia, all of which express the nuclearprotein Snail (Sna) (Alberga et al., 1991). Embryos double-antibody labelled for Hdc and Sna proteins reveal that all Sna-positive cells of the dorsal disc primordia co-express Hdcprotein (Fig. 4D). The additional Hdc-positive and Sna-negative cells seen adjacent to the discs most likely representmesodermally derived adepithelial precursors, as these Hdc-positive cells co-express the imaginal mesoderm-specific Twistprotein (Fig. 7A).

The headcase null phenotype affects imaginal celldifferentiation but not growthIn order to assess the possible role of the hdc gene in imaginaltissue development, deletions removing part or all of themapped hdc transcription unit (Fig. 3A) were generated byimprecise P-element excision (see Methods). All deletionscaused pupal lethality when homozygous and heterallelic com-binations show similar mutant phenotypes suggesting that all

neuroblasts (situated above the arrow) express both hdc and grh. At this late third instar, hdc expression fades in the neuroblast population, but reneuroblasts (indicated by arrow) strongly express grh but not hdc. At thiwill stop dividing over the next 12-24 hours (Truman and Bate, 1988).

alleles are functionally null. We confirmed this for the typicalallele, hdc43, where neither hdc RNA (Fig. 5) nor protein (datanot shown) are detected in embryos or larvae homozygous forthe mutation. We can conclude that the pupal lethal phenotypeof the hdc mutant chromosomes is caused by the removal ofthe hdc gene activity, since Hdc function provided from anheat-inducible hdc transformation construct can completelyrescue the pupal lethality (see Methods).

Although hdc mutants die as pupae, all larval imaginal discsare present, and their size and shape appear normal. Thereforethe loss of hdc does not affect cell growth in any obvious way.The penetrance of pupal lethality is 100%; however, the devel-opmental stage at which the pupae die is variable, ranging frombrown pupae with no obvious differentiated tissue, to deadpharate adults. The most prominent hdc defects involve headdevelopment (Fig. 6) and can result in complete deletion of thehead capsule (Fig. 6C), or duplication of head cuticle orantennae (Fig. 6B,E). Although the majority of pharate adultsexhibit normal external head morphology, they possessmassive defects in internal head structures, including the CNS,muscular and tracheal tissues (Fig. 6G). Other tissues are alsoaffected (including wings, halteres, legs and epidermis) butvary between mutant individuals. Such pleiotropy is consistentwith the broad range of imaginal cell-type specificity of the hdc

Fig. 7. hdc expression correlates withimaginal cell proliferation. Confocalimages of B5 transformantdissections that have been antibodystained with anti-β-gal (red; labelsnuclei of hdc-expressing imaginalcells) as well as a second antibody(green) against either the Twist (A,B;labels adepithelial cells) orGrainyhead (C,D; labels imaginalneuroblasts) proteins. Each image issplit into three panels showing green(top left), red (top right), andcombined (bottom right) fluorescencechannels. (A) A stage 16 embryodissected flat, with anterior towardsthe bottom of figure and the ventraldisc primordia bilaterally reflectedabout the ventral midline. At thisearly stage, hdc expression is foundin a subset of adepithelial cells(appear as yellow cells) associatedwith the thoracic imaginal discs. (B) However, by mid-third instarwhen both imaginal disc andadepithelial cells are dividingrapidly, hdc expression is present inall adepithelial cells (of a leg disc inthis example). (C) Similarly, in theCNS of a second instar, hdcexpression is concentrated along theventral surface of thoracichemisegment and is found in allimaginal neuroblast cells. Forexample, a cluster of three

stage, all neuroblasts are undergoing proliferation. (D) However, bymains on in the neuronal progeny. In this example, a cluster ofs transitional stage, the neuroblast cells, although still proliferating,

4158 T. A. Weaver and R. A. H. White

gene expression; however, the reason for variable expressivityis unclear.

hdc expression prefigures imaginal cell proliferationOne unifying character shared by imaginal cells is their com-mitment to a strict program of diploid growth. Although hdcis expressed in all imaginal lineages, it initiates in the differentimaginal cell groups according to a developmental sequence.Therefore, we were interested in whether hdc expression in aparticular imaginal cell type correlated with that tissue’soverall program of proliferation and differentiation.

It is clear from the embryonic expression in imaginal discprimordia (Fig. 2G) that hdc activation in these cells occurs atleast 24 hours before their re-entry into the mitotic cell cycle,since mitosis in these cells is not observed until late first/earlysecond instar (Madhavan and Schneiderman, 1977). Thisobservation is also true for the adepithelial cells (Fig. 7A,B).These cells constitute the primordia of adult somatic muscula-ture and are found closely associated with the imaginal discsto which they ultimately attach (Reed et al., 1975; Currie andBate, 1991); an association that we can trace back intoembryonic stages using twist as a marker for adepithelial cells(Bate et al., 1991) and hdc as a marker for the disc primordia(Fig. 7A). Interestingly, at this early stage of adepithelialdevelopment, approximately half of the cells express hdc,complete overlap of twist- and hdc-expressing cells is apparentby larval stages (Fig. 7B). Since the muscle precursors aremitotically quiescent until second larval instar (as judged bysusceptibility to a DNA synthesis inhibitor (Broadie and Bate,1991), we can conclude that hdc expression prefigures the acti-vation of imaginal somatic mesoderm cell proliferation.Similarly, for the imaginal histoblast lineage, hdc-positive his-toblast cells are not detected in the embryo (Fig. 2), or in 1st,2nd or early 3rd instar larvae (data not shown), even thoughhistoblast cells are present and can be visualized by several his-toblast-specific markers (Alberga et al., 1991; Hartenstein andJan, 1992; Whiteley et al., 1992; Hayashi et al., 1993). Instead,hdc expression is first detected in wandering third instar (Fig.1I), corresponding to 12-24 hours before the various histoblastnests begin division during pupal stages (Madhavan andMadhavan, 1980).

A correlation also exists between hdc repression and the endof imaginal proliferation and the onset of differentiation (Fig.7C,D). This is best illustrated for imaginal neuroblast (IN) pro-liferation which has been studied extensively by bromod-eoxyuridine incorporation studies (Truman and Bate, 1988;Truman et al., 1993). Double-labelling experiments usingGrainyhead protein as an IN marker (Bray and Kafatos, 1991)demonstrate that during the intense proliferative stages of thesecond (Fig. 7C) and early third instars (data not shown) hdcis expressed in IN, as well as their progeny. This relationshipchanges when the larval CNS prepares for pupariation in latethird instar. Now the levels of hdc expression are severelyreduced in the IN (Fig. 7D), but not their progeny, presumablybecause these cells are competent to continue cell proliferation.The timing of this change precedes final neuroblast divisionwhich, for these neuroblasts, located in the thoracic region ofthe ventral nerve cord, occurs sometime within the first 12hours following pupariation (Truman and Bate, 1988). Thisobservation is broadly supported by the pupal expression ofhdc during disc morphogenesis. Whereas expression is

uniformly strong throughout the proliferative stages, this fadesduring the first 24 hours of pupal development when celldivision is ending. Therefore, the activation of hdc expressioncorrelates with the onset of imaginal proliferation and hdc inac-tivation correlates with its cessation.

DISCUSSION

Molecular properties of hdcThe hdc gene encodes two large open reading framesseparated by a single in-frame amber stop codon. We areconvinced that the first ORF encodes part, if not all, the Hdcprotein for two reasons. Firstly, in vitro translated hdc cRNAproduces a product of approximately 70×103 Mr, the sizepredicted for the first ORF (data not shown). Secondly, a monoclonal antibody raised specifically against the upstreamORF recognizes a protein whose tissue distribution coincideswith the complete B5-lacZ and hdc mRNA expression patterns.What is the significance of the downstream ORF? Unfortu-nately, we cannot detect the endogenous Hdc protein bywestern blotting of embryonic or imaginal disc protein lysatesand so the molecular weight of Hdc in vivo is unknown.However, western blot analysis of imaginal disc protein lysatesmade from heat-shock-treated HS-Hdc transformant line 87.1revealed two bands: a strongly cross-reacting 71×103 Mrprotein corresponding to the upstream ORF, and a weaklycross-reacting protein >116×103 Mr, most probably resultingfrom inefficient translational readthrough into the downstreamORF (data not shown). This intriguing result suggests that thedownstream ORF is potentially translatable, albeit at a reducedefficiency compared to the upstream ORF. Stop codon sup-pression is not without precedent in Drosophila. During thetranslation of the kelch gene mRNA, which encodes acomponent of ring canals linking nurse cells during oocytematuration, partial suppression of a stop codon results insynthesis of a large readthrough polyprotein (Xue and Cooley,1993).

Database searches revealed no sequence similarity orconserved protein motifs within either the upstream or down-stream ORF. However, Hdc is cysteine-rich and has a predictedpI of 9.6. Since the Hdc protein is concentrated in thecytoplasm, the extreme basicity may reflect an affinity forprotein-protein or protein-RNA interactions.

The hdc gene is not required for imaginal cellgrowth, but rather adult differentiationTo address the role that hdc plays in imaginal morphogenesis,we characterized the developmental consequences caused bysmall deletions removing the hdc transcription unit. Theresulting pupal lethality is associated with defects in theimaginal morphogenesis of many tissue types. The lack of aneffect on embryonic and larval development is also consistentwith the imaginal specificity of hdc gene expression. Althoughmany late lethals may result from maternal effects (Perrimonet al., 1989), a maternal source of Hdc is not consistent withour in situ hybridizations and anti-Hdc-staining experiments,which fail to detect hdc expression before stage 13 (data notshown). Even so, we cannot completely rule out a maternalcontribution until germ line hdc nulls are tested for more severeembryonic or larval phenotypes.

4159Adult morphogenesis in Drosophila

Since Hdc is a pioneer protein, we cannot infer a mechanismfor its action based on its amino acid sequence. However, wedo have a detailed understanding of its expression pattern.Because hdc expression is strictly limited to imaginal cells, hdcmay participate in some process shared by all imaginal cellsand distinct from larval cells. We can think of two such cellularbehaviors, the first is maintenance of mitotic growth and thesecond is the response to molting hormones. We are as yetunable to distinguish between these two possibilities.However, our results indicate that hdc expression prefiguresthe time when imaginal cells re-enter the mitotic cell cycle.Expression continues throughout proliferation, but then ceasesjust before the final cell divisions, suggesting a role for hdc inthe timing of proliferation. However, as hdc mutations exhibitneither a small nor overgrown disc phenotype, it is unlikelythat hdc functions in regulating the number of cell divisions orsurvival of imaginal cells.

hdc and hormonal signalling?The mechanisms coordinating imaginal cell proliferation withdifferentiation depend on hormonal control. In particular,juvenile hormone (JH) prevents imaginal tissues from under-going premature metamorphosis in the presence of the moltinghormone 20-hydroxyecdysone (ecdysone) (reviewed inRiddiford, 1993). The relative titer of each hormone is critical:abnormally high levels of JH administered before the onset ofproliferation in the imaginal discs (Riddiford and Ashburner,1991) or histoblasts (Madhavan, 1973; Postlethwait, 1974)interferes with their future ability to differentiate withoutaffecting their ability to divide, a phenotype similar to that seenin hdc mutations. Conversely, a reduction in ecdysone titerresults in delayed or blocked metamorphosis (Garen et al.,1977; Sehnal and Bryant, 1993). Although several ecdysone-inducible genes have been isolated (reviewed in Fristrom andFristrom, 1993), little is known about the molecular mecha-nisms linking hormonal signalling to tissue morphogenesis.Mutational dissection of the ecdysone-inducible early puffgene E74 demonstrates its role in imaginal tissue morphogen-esis during prepupal and pupal development (Fletcher et al.,1995). Most notably, mutations effecting the ‘B’ transcript inE74 fail to undergo head evertion at pupation, a cryptocephalicphenotype reminiscent of that seen in hdc mutants (Fig. 6C).These results raise the possibility that hdc may be involved inthe integration or response to hormonal information byimaginal cells.

We thank the following people for generously providing reagents:Nick Brown, Larry Kauvar, John Tamkun and Carl Thummel forlibraries and vectors; Sarah Bray, Nick Brown and James Castelli-Gair for antibodies; John Roote, Rachel Drysdale, Alex Gould andCahir O’Kane for fly stocks. We also thank for the expert technicalassistance: Rachel Chesterton for illustration, John Bashford for pho-tography, Peter Torok for confocal microscopy, Tony Burgess for EMand Jill King for histology. We especially thank Sarah Bray andRachel Drysdale for critically reading this manuscript. This work wassupported by the Wellcome Trust, and is dedicated to the memory ofHazel Flood and her home-grown wisdom.

REFERENCES

Alberga, A., Boulay, J., Kempe, E., Dennefeld, C. and Haenlin, M. (1991).

The snail gene required for mesoderm formation in Drosophila is expresseddynamically in derivatives of all three germ layers. Mech. Dev. 36, 117-127.

Anderson, D. T. (1963). The embryology of Dacus tryoni 2. Development ofimaginal discs in the embryo. J. Embryol. Exp. Morph. 11, 339-351.

Ashburner, M. (1989) Drosophila. A Laboratory Handbook. Cold SpringHarbor Laboratory Press.

Bainbridge, S. P. and Bownes, M. (1981). Staging the metamorphosis ofDrosophila melanogaster. J. Embryol. Exp. Morph. 66, 57-80.

Bate, M. and Martinez-Arias, A. (1991). The embryonic origin of imaginaldiscs in Drosophila. Development 112, 755-761.

Bate, M., Rushton, E. and Currie, D. A. (1991). Cells with persistent twistexpression are the embryonic precursors of adult muscles in Drosophila.Development 113, 79-89.

Bellen, H. J., O’Kane, C. J., Wilson, C., Grossniklaus, U., Pearson, R. K.and Gehring, W. J. (1989). P-element-mediated enhancer detection: Aversatile method to study development in Drosophila. Genes Dev. 3, 1288-1300.

Bodenstein, D. (1950). The postembryonic development of Drosophila. InBiology of Drosophila, pp. 275-367. New York: Wiley.

Bray, S. J. and Kafatos, F. C. (1991). Developmental function of Elf-1: anessential transcription factor during embryogenesis in Drosophila. GenesDev. 5, 1672-1683.

Broadie, K. S. and Bate, M. (1991). The development of adult muscles inDrosophila: ablation of identified muscle precursor cells. Development 113,103-118.

Brown, N. H. and Kafatos, F. C. (1988). Functional cDNA libraries fromDrosophila embryos. J. Mol. Biol. 203, 425-437.

Bryant, P. J. (1978). Pattern formation in imaginal discs. In The Genetics andBiology of Drosophila, pp. 230-336. New York: Academic Press.

Bryant, P. J. and Schmidt, O. (1990). The genetic control of cell proliferationin Drosophila imaginal discs. J. Cell Sci. Supplement 13, 169-189.

Campos-Ortega, J. A. and Hartenstein, V. (1985) The EmbryonicDevelopment of Drosophila melanogaster. Amsterdam: Springer-Verlag.

Chan, L.-N. and Gehring, W. (1971). Determination of blastoderm cells inDrosophila melanogaster. Proc. Natl. Acad. Sci. USA 68, 2217-2221.

Cohen, B., Simcox, A. A. and Cohen, S. M. (1993). Allocation of the thoracicimaginal primordia in the Drosophila embryo. Development 117, 597-608.

Cohen, S. M. (1993). Imaginal disc development. In The Development ofDrosophila melanogaster, pp. 747-842. Cold Spring Harbor LaboratoryPress.

Cooper, K. W. (1950). Normal spermatogenesis in Drosophila. In Biology ofDrosophila, pp. 1-61. New York: Wiley.

Currie, D. A. and Bate, M. (1991). The development of adult abdominalmuscles in Drosophila: myoblasts express twist and are associated withnerves. Development 113, 91-102.

Daniels, S. B., McCarron, M., Love, C. and Chovnick, A. (1985).Dysgenesis-induced instability of rosy locus transformation in Drosophilamelanogaster - analysis of excision events and the selective recovery ofcontrol element deletions. Genetics 109, 95-117.

Datta, S. (1995). Control of proliferation activation in quiescent neuroblasts ofthe Drosophila central nervous system. Development 121, 1173-1182.

Datta, S. and Kankel, D. R. (1992). l(1)trol and l(1)devl, loci affecting thedevelopment of the adult central nervous system in Drosophilamelanogaster. Genetics 130, 523-537.

Dubendorfer, A. (1971). Untersuchungen zum anlageplan unddeterminationszustand der weiblichen genital- und analprimordien vonMusca domestica L. Wilhelm Roux Arch. EntwMech. Org. 168, 142-168.

Ebens, A. J., Garren, H., Cheyette, B. N. R. and Zipursky, S. L. (1993). TheDrosophila anachronism locus: A glycoprotein secreted by gila inhibitsneuroblast proliferation. Cell 74, 15-27.

Fletcher, J. C., Burtis, K. C., Hogness, D. S. and Thummel, C. S. (1995). TheDrosophila E74 gene is required for metamorphosis and plays a role in thepolytene chromosome puffing response to ecdysone. Development 121,1455-1465.

FlyBase (1994). The Drosophila genetic database. Nucleic Acids Res. 22, 3456-3458.

Foe, V. E., Odell, G. M. and Edgar, B. A. (1993). Mitosis and morphogenesisin the Drosophila embryo: point and counterpoint. In The Development ofDrosophila melanogaster, pp. 149-300. Cold Spring Harbor LaboratoryPress.

Fristrom, D. and Fristrom, J. W. (1993). The metamorphic development ofthe adult epidermis. In The Development of Drosophila melanogaster,pp. 843-898. Cold Spring Harbor Laboratory Press.

4160 T. A. Weaver and R. A. H. White

Fuller, M. T. (1993). Spermatogenesis. In The Development of Drosophilamelanogaster, Cold Spring Harbor Laboratory Press, pp. 71-148.

Fuse, N., Hirose, S. and Hayashi, S. (1994). Diploidy of Drosophila imaginalcells is maintained by a transcriptional repressor encoded by escargot. GenesDev. 8, 2270-2281.

Garen, A., Kauvar, L. and Lepesant, J. (1977). Roles of ecdysone inDrosophila development. Proc. Natl. Acad. Sci. USA 74, 5099-5103.

Gould, A. P. (1990). Homeotic gene function during embryogenesis inDrosophila. PhD Thesis. University of Cambridge.

Gould, A. P. and White, R. A. H. (1992). Connectin, a target of homeotic genecontrol in Drosophila. Development 116, 1163-1174.

Hartenstein, V. and Jan, Y. N. (1992). Studying Drosophila embryogenesiswith P-lacZ enhancer trap lines. Roux’s Arch. Dev. Biol. 201, 194-220.

Hayashi, S., Hirose, S., Metcalfe, T. and Shirras, A. D. (1993). Control ofimaginal cell development by the escargot gene of Drosophila. Development118, 105-115.

Madhavan, K. (1973). Morphogenetic effects of juvenile hormone andjuvenile hormone mimics on adult development of Drosophila. J. InsectPhysiol. 19, 441-453.

Madhavan, M. M. and Madhavan, K. (1980). Morphogenesis of theepidermis of the adult abdomen of Drosophila. J. Embryol. Exp. Morph. 60,1-31.

Madhavan, M. M. and Schneiderman, H. A. (1977). Histological analysis ofthe dynamics of growth of imaginal discs and histoblast nests during thelarval development of Drosophila melanogaster. Wilhelm Roux’s Arch. Dev.Biol. 183, 269-305.

Manning, G. and Krasnow, M. A. (1993). Development of the Drosophilatracheal system. In The Development of Drosophila melanogaster, pp. 609-686. Cold Spring Harbor Laboratory Press.

Martinez-Arias, A. (1993). Development and patterning of the larvalepidermis of Drosophila. In The Development of Drosophila melanogaster,pp. 517-608. Cold Spring Harbor Laboratory Press.

Meise, M. and Janning, W. (1993). Cell lineage of larval and imaginalthoracic anlagen cells of Drosophila melanogaster, as revealed by single-celltransplantations. Development 118, 1107-1121.

O’Kane, C. J. and Gehring, W. J. (1987). Detection in situ of genomicregulatory elements in Drosophila. Proc. Natl. Acad. Sci. USA 84, 9123-9127.

Perrimon, N., Engstrom, L. and Mahowald, A. P. (1989). Zygotic lethalswith specific maternal effect phenotypes in Drosophila melanogaster. I. Locion the X chromosome. Genetics 121, 333-352.

Poodry, C. A. and Woods, D. F. (1990). Control of the developmental timerfor Drosophila pupariation. Wilhelm Roux’s Arch. Dev. Biol. 199, 219-227.

Postlethwait, J. H. (1974). Juvenile hormone and the adult development ofDrosophila. Biol. Bull. Mar. Biol. Lab.,Woods Hole. 147, 119-135.

Prokop, A. and Technau, G. M. (1991). The origin of postembryonicneuroblasts in the ventral nerve cord of Drosophila melanogaster.Development 111, 79-88.

Reed, C. T., Murphy, C. and Fristrom, D. (1975). The ultrastructure of thedifferentiating pupal leg of Drosophila melanogaster. Wilhelm Roux’s Arch.Dev. Biol. 178, 285-302.

Riddiford, L. M. (1993). Hormones and Drosophila development. In TheDevelopment of Drosophila melanogaster, pp. 899-940. Cold Spring HarborLaboratory Press.

Riddiford, L. M. and Ashburner, M. (1991). Role of juvenile hormone inlarval development and metamorphosis in Drosophila melanogaster. Gen.Comp. Endocrinol. 82, 172-183.

Robertson, H. M., Preston, C. R., Phillis, R. W., Johnson-Schlitz, D. M.,Benz, W. K. and Engels, W. R. (1988). A stable source of P-elementtransposase in Drosophila melanogaster. Genetics 118, 461-470.

Roseland, C. R. and Schneiderman, H. A. (1979). Regulation andmetamorphosis of the abdominal histoblasts of Drosophila melanogaster.Wilhelm Roux’s Arch. Dev. Biol. 186, 235-265.

Sehnal, F. and Bryant, P. J. (1993). Delayed pupariation in Drosophilaimaginal disc overgrowth mutants is associated with reduced ecdysteroidtiter. J. Insect Physiol. 39, 1051-1059.

Simcox, A. and Sang, J. H. (1983). When does determination occur inDrosophila embryos? Dev. Biol. 97, 212-221.

Simpson, P., Berreur, P. and Berreur-Bonnenfant, J. (1980). The initiationof pupariation in Drosophila: dependence on growth of the imaginal discs. J.Embryol. Exp. Morph. 57, 155-165.

Spradling, A. C. (1986). P element-mediated transformation. In Drosophila aPractical Approach, pp. 175-198. Oxford: IRL Press.

Truman, J. W. and Bate, M. (1988). Spatial and temporal patterns ofneurogenesis in the central nervous system of Drosophila melanogaster.Dev. Biol. 125, 145-157.

Truman, J. W., Taylor, B. J. and Awad, T. A. (1993). Formation of the adultnervous system. In The Development of Drosophila melanogaster, pp. 1245-1276. Cold Spring Harbor Laboratory Press.

White, R. A. H. and Wilcox, M. (1984). Protein products of the bithoraxcomplex in Drosophila. Cell 39, 163-171.

Whiteley, M., Noguchi, P. D., Sensabau, S. M., Odenwald, W. F. andKassis, J. A. (1992). The Drosophila gene escargot encodes a zinc fingermotif found in snail-related genes. Mech. Dev. 36, 117-127.

Whitten, J. (1980). The tracheal system. In The Genetics and Biology ofDrosophila, pp. 499-540. New York: Academic Press.

Xue, F. Y. and Cooley, L. (1993). Kelch encodes a component of intercellularbridges in Drosophila egg chambers. Cell 72, 681-693.

(Accepted 23 August 1995)