signals transmitted along retinal axons in drosophila ... · lamina target ﬁeld at the anterior...

3753Development 125, 3753-3764 (1998)Printed in Great Britain © The Company of Biologists Limited 1998DEV9614

Signals transmitted along retinal axons in Drosophila : Hedgehog signal

reception and the cell circuitry of lamina cartridge assembly

Zhen Huang and Samuel Kunes*

Department of Molecular and Cellular Biology, Harvard University, 7 Divinity Avenue, Cambridge, MA 02138, USA*Author for correspondence (e-mail: [email protected])

Accepted 19 July; published on WWW 7 September 1998

The arrival of retinal axons in the brain of Drosophilatriggers the assembly of glial and neuronal precursors intoa ‘neurocrystalline’ array of lamina synaptic ‘cartridges’.Hedgehog, a secreted protein, is an inductive signaldelivered by retinal axons for the initial steps of laminadifferentiation. In the development of many tissues,Hedgehog acts in a signal relay cascade via the inductionof secondary secreted factors. Here we show that laminaneuronal precursors respond directly to Hedgehog signalreception by entering S-phase, a step that is controlled by

the Hedgehog-dependent transcriptional regulator Cubitusinterruptus. The terminal differentiation of neuronalprecursors and the migration and differentiation of gliaappear to be controlled by other retinal axon-mediatedsignals. Thus retinal axons impose a program ofdevelopmental events on their postsynaptic field utilizingdistinct signals for different precursor populations.

Key words: Drosophila, Lamina, Photoreceptor axon, hedgehog,Cubitus interruptus

SUMMARY

;

e

o,91;niornstols

aesse a

lar;onetheptic

is

etrog

INTRODUCTION

The survival and differentiation of postsynaptic cepopulations often depends on the arrival of afferent axonstheir target fields. For example, arriving retinal axons anecessary for the establishment of the retinorecipient layersthe tectum (Kelly and Cowan, 1972; Yamagata et al., 1995).Manduca, the establishment of glomeruli units of the olfactobulb depends on sensory innervation (Oland and Tolbe1996). Thus growing axons not only detect environmental cufor guidance and target recognition, but may emit signals telicit the survival or maturation of cells in their prospectivtarget fields. The molecular identities of such signals havesome cases been determined. Agrin (Rupp et al., 1991; Smet al., 1992; Tsim et al., 1992) and Aria (Falls et al., 1993), example, are transported to motor axon termini where thmediate the assembly of the postsynaptic apparatus at neuromuscular junction.

The visual system of Drosophilaoffers a unique opportunityto investigate these kinds of axon-target interactions. TDrosophila visual system is a complex and finely orderestructure composed of tens of thousands of neurons and (reviewed by Meinertzhagen and Hanson, 1993). It consiststhe compound eyes and the optic ganglia, which are the visprocessing centers of the brain. Each of the approximately 8ommatidial units of the compound eye contains photoreceptor neurons (R-cells) that project retinotopicainto distinct ganglion layers of the brain. The R1-Rphotoreceptors send their axons to destinations in the lamwhere they establish synaptic connections in

ll inre of Inryrt,es

hate inith

foreythe

hedglia ofual008lly6ina,a

‘neurocrystalline’ field of precisely repeating ‘cartridge’ units(Braitenberg, 1967; Meinertzhagen and O’Neil, 1991Meinertzhagen and Hanson, 1993).

In Drosophilaand other arthropods, the development of thvisual portion of the brain is tightly regulated by the ingrowthof axons from the compound eye (Power, 1943; Macagn1979; Fischbach and Technau, 1984; Selleck and Steller, 19reviewed by Kunes and Steller, 1993). As R-cell differentiatioand ommatidial assembly progress in a posterior-to-anterorder across the eye disc epithelium, the arrival of retinal axoin the brain triggers cartridge neuron precursors (LPCs) complete a final cell division and commence neuradifferentiation (Fig. 1; Selleck and Steller, 1991). Retinal axonalso elicit the migration of glial precursors into the lamintarget field, where they mature into glia of several distinct typ(Fig. 1; Winberg et al., 1992; Perez and Steller, 1996). Theevents result in the assembly of a retinal axon fascicle andset of neuronal and glial precursors into a stereotyped celluensemble known as a ‘lamina column’ (Fig. 1C,DMeinertzhagen and Hanson, 1993). The choreography of axfascicle, target cell interactions yields a one-to-oncorrespondence of ommatidial and cartridge units, and sets stage for the subsequent establishment of precise synacircuitry.

One of the signals that retinal axons deliver to the laminathe secreted product of the hedgehoggene (hh; Huang andKunes, 1996; Nusslein-Volhard and Wieschaus, 1980; Leeal., 1992; Mohler and Vani, 1992; Tabata et al., 1992; Tashiet al., 1993). Hedgehog (Hh) is expressed by differentiatinphotoreceptor cells immediately posterior of the

3754

ofvae0ced

6)

d

ned

r)

l.,ti-ofy 5-,-

esndM

et amnnd.

hewhetralandareinlarralaarely.-L4

orion.,ral

Z. Huang and S. Kunes

morphogenetic furrow in the eye imaginal disc (Lee et a1992) and is a signal to more anterior cells to enter the pathwof photoreceptor cell determination (Ma et al., 1993; Heberleet al., 1993, 1995; Dominguez and Hafen, 1997). We harecently provided evidence that Hh is transported from the einto the brain along retinal axons and triggers the initial steof LPC maturation (Huang and Kunes, 1996). Hedgehactivity in the eye is both necessary and sufficient for the eaevents of lamina neurogenesis, including the terminal cdivision of LPCs and the onset of their expression of eadifferentiation markers. Thus Hh acts as a dual signal tcouples neural development between the eye and brain.

In the development of many Drosophila tissues, Hh acts ina signal relay cascade via the induction of secondary secrfactors, such as decapentaplegicand wingless(Basler andStruhl, 1994; Capdevila and Guerrero, 1994; Tabata aKornberg, 1994). In the lamina, Hh might be transmittedirectly by retinal axons to G1-phase LPCs to trigger their cellcycle progression and subsequent differentiation. AlternativeLPC maturation might depend all or in part on secondasignals delivered by an intermediate Hh target. To betresolve these events, we have examined the Hh-dependenglia cell maturation and the autonomy of Hh signaling pathwcomponents in different lamina cell populations. Here we shthat LPCs respond directly to Hh signal reception by enterS-phase, and that this step is controlled by the Hh-dependtranscriptional regulator Cubitus interruptus. Glia ceprecursors, though Hh responsive, apparently depend on sother retinal axon-mediated signal for their migration andifferentiation in the lamina target field. Thus retinal axoncommunicate directly to lamina precursors by utilizindifferent signals for distinct precursor populations.

MATERIALS AND METHODS

Mosaic analysisThe appropriate crosses were set up as described below and, folloegg collection, larvae were heat shocked at about 6 hours phatching for 70 minutes at 37˚C to induce hsFLP transgeneexpression. The larvae were grown at 25˚C (unless otherwise nobefore dissection at late third instar stage. Brain samples from thlarvae were then stained with the appropriate antibodies. Whmosaic analysis was done in the hhts2 background, egg collection wascarried out at 18˚C. The larvae were shifted to 28˚C after heat shin the first instar to block later photoreceptor cell development in eye. The following crosses were used for analyzing the correspondmutations:

• smo: y, hsFLP122; P[arm-lacZ]28A, P[FRT]40A/P[y+], CyO ×yw/Y; smo3, P[FRT]40A/P[y+], CyO

• pka-DC0h2: y, hsFLP122; P[arm-lacZ]28A, P[FRT]40A/P[y+],CyO; hhts2/TM6B, Tb × y/Y; DC0h2, P[FRT]40A/P[y+], CyO;hhts2/TM6B, Tb

• ptc6p: y, hsFLP122; P[FRT]43D, P[arm-lacZ]51A/P[y+], CyO;hhts2/TM6B, Tb × y/Y; P[FRT]43D, ptc6p/P[y+], CyO; hhts2/TM6B, Tb

Ectopic gene expressionMisexpression of hh with a flp-out hh+ transgene (Basler and Struhl1994) was performed in animals made ‘eyeless’ by the use of the hhts2

strain background, as described above (see also Huang and Ku1996). Misexpression of Ci was accomplished by using a P[UAS-Ci]line (Alexandre et al., 1996) and a ‘flp-out’ GAL4 line,P[Tub>CD2>GAL4] (Pignoni and Zipursky, 1997). This experimen

l.,ayinveyepsogrlyellrlyhat

eted

ndd

ly,ryterce ofayowingentllomeds

g

wingost-

ted)eseen

ocktheing

,

nes,

t

was also performed in the ‘eyeless’ condition achieved by the usethe hhts2 strain background. Crosses were set up at 18˚C, and larwere shifted to 28˚C after hatching to prevent R-cell formation. A 3minute heat shock at 37˚C during the early second instar stage induthe expression of the hsFLPtransgene, and activated GAL4expressionin somatic clones. The dominant-negative truncated form of Ci (Ci7was misexpressed using a P[UAS-Ci76] line (Aza-Blanc et al., 1997)and the ‘flp-out’ GAL4driver. The corresponding crosses are listebelow:

• hh: y, hsFLP122; hhts2/TM6B, Tb × y/Y; P[Tub>y+>hh]/+;hhts2/TM6B, Tb

• Ci: y, hsFLP122; hhts2, P[UAS-Ci]/TM6B, Tb × yw/Y;P[Tub>CD2>GAL4]/P[y+], CyO; hhts2/TM6B, Tb

• DN-Ci: y, hsFLP122; P[UAS-Ci76]/TM6B, Tb × yw/Y;P[Tub>CD2>GAL4]/P[y+], CyO

ImmunohistochemistryThe CNS was dissected from late third instar larvae, fixed and staias described previously (Kunes et al., 1993). The ombP1 enhancer trapline (Sun et al., 1995) and Omb antibody (gift of G. O. Pflugfeldewere used interchangeably, with similar results. patchedgeneexpression was monitored with the ptcpromoter-lacZfusion line, FE3(Alexandre et al., 1996). Anti-Repo polyclonal antibody (Halter et a1995), anti-OMB antibody (Grimm and Pflugfelder, 1996) and anCi antibody (Motzny and Holmgren, 1995) were used at dilutions 1:200, 1:400 and 1:20, respectively. S-phase cells were labeled bbromo-2′-deoxy-uridine (BrdU) incorporation (Truman and Bate1988) and detected with anti-BrdU antibody (1:50, BectonDickinson). Conditions for other primary and secondary antibodiwere described previously (Kaphingst and Kunes, 1994; Huang aKunes, 1996). Confocal microscopy was performed on a Zeiss LS410 microscope equipped with a Krypton/Argon laser.

RESULTS

Formation of precartridge ensembles in thedeveloping lamina fieldAs an ommatidial axon fascicle arrives in the lamina targfield, it assembles with neuronal and glial precursors intostereotyped precartridge ensemble known as a lamina colu(Fig. 1; Meinertzhagen and Hanson, 1993). The neuronal aglial precursors have distinct origins within the optic lobeMost neuronal precursors (LPCs) are incorporated into tlamina target field at the anterior margin of the lamina furro(Fig. 1A,C). The precursors of several glia cell types, on tother hand, arise in anlagen adjacent to the dorsal and venmargins of the lamina (Fig. 1A,B) and migrate into the laminalong an axis perpendicular to that of LPC entry (Perez aSteller, 1996). The neuronal and glia precursor populations interwoven into cell-type specific layers at the anterior margof the lamina, so that a lamina column harbors a particuneuronal or glial cell-type precursor at a specific mediolateposition (Fig. 1C,D). Lamina neurons L1-L4 form a stack in superficial layer, while L5 neurons reside in a medial layer nethe R1-R6 axon termini. Epithelial and marginal glia arlocated above and below the R1-R6 termini, respectiveSatellite glia are interspersed among the neurons of the L1layer.

A number of markers distinguish glial and neuronal precurscells from the corresponding mature cell types. The expressof optomotor-blind(omb; Heisenberg, 1972; Pflugfelder et al1990) labels both glial precursors in the dorsal and vent

3755Axonal cues for lamina cartridge assembly

wa

onavead

ile,

anlagen and mature glia that have migrated into the lamtarget field (Fig. 1B,D; Poeck et al., 1993). The glia cell markRepo (Fig. 2A,D; Xiong et al., 1994; Halter et al., 1995) athe enhancer-trap lacZinsertion 3-109 (not shown; Winberg eal., 1992) are expressed by glia once they have enteredlamina target field. Cubitus interruptus (Ci), a transcriptionmediator of Hh signaling (Eaton and Kornberg, 1990; Orenical., 1990; Domiguez et al., 1996; Alexandre et al., 1996; Hep

Fig. 1. Lamina development at the late third instar larval stage. (Afrom the eye imaginal disc trigger the assembly of neuronal and gtarget field. (A) As shown schematically, photoreceptor cells assemfurrow (mf). Two such clusters (green color) in the retina are showcrescent-shaped lamina (lam; blue color). As indicated by the arroaxon target field at its anterior margin, which is demarcated by a mprecursor cells (GPCs) are generated in two domains (red color) t(B) Confocal micrograph shows the photoreceptor cells and their a(anti-HRP) antibody. Postmitotic LPCs within the lamina axon targstaining (blue color). Glial cells in the lamina as well as their precucolor) from the ombP1 enhancer-trap line (Sun et al., 1995). An arroand B, anterior is to the left, dorsal up. Scale bar in B, 40 µm. (C,D) Lalamina differentiation occurs in a temporal progression on the anteommatidial R-cell clusters arrive at the anterior margin of the lami(blue colors) and glia precursors (red color) in a vertical lamina colamina, LPCs await a retinal axon-mediated signal in G1-phase (gray cphase (black cell) at the posterior margin of the furrow. Postmitoticmargin of the furrow. In older columns at the posterior of the laminblue color) as they become specified as the lamina neurons L1-L5vertical axis. Lamina glial cells (red color) take up cell-type positio(E-glia) layers sandwich the R1-R6 axon termini, whereas the satmedulla neuropil (medn), which serves as the target for R7/8 axonlamina cartridge assembly described schematically in C are reveaglial precursors (red color), and neuronal membranes, including pstained as described for B. In C and D, anterior is to the left, latera

inaer

ndt theal etker

et al., 1997) is expressed by LPCs anterior of the lamina furroand by the postmitotic neuronal precursors within the lamin(see Fig. 4D,F). The nuclear protein Dachshund (Dac; Mardet al., 1994) is expressed only by neuronal precursors that hbegun terminal differentiation and lie posterior of the laminfurrow (Fig. 1B,D; Huang and Kunes, 1996). Thus, Omb anCi label the glial and neuronal precursors, respectively, whthe mature cells, following their interaction with retinal axons

,B) The lamina as viewed from the lateral perspective. Retinal axons arrivinglial precursors into precartridge ensembles in the crescent-shaped lamina

ble into ommatidial clusters behind the anteriorly moving morphogeneticn to highlight the topographic projection of their axon fascicles into thews, neuronal precursor cells of the lamina (LPCs) are incorporated into theorphological depression known as the lamina furrow (lf in B and D). Glia

hat lie at the dorsal and ventral margins of the prospective lamina.xons (green color) as revealed by staining with anti-horseradish peroxidaseet field express the nuclear protein Dac, as revealed by anti-Dac antibodyrsors in the two posterior domains are detected by lacZexpression (redw and adjacent bar in A and a bar in B mark the dorsoventral midline. In Amina cartridge assembly from the horizontal perspective. Like the eye,rioposterior axis. (C) As shown schematically, axon fascicles from ‘new’

na (adjacent to the lamina furrow; lf in C and D) and associate with neuronallumn assembly (two such columns are shown in C). At the anterior of theells at the trough of the lamina furrow (lf)) and enter their terminal S- (Dac-positive; blue color) LPCs assemble into columns at the posteriora, a subset of postmitotic LPCs express definitive neuronal markers (dark. These neurons arise at cell-type specific positions along the column’sns in the precartridge assemblies. The marginal (Ma-glia) and epithelial gliaellite glia (S-glia) are interspersed within the neuronal precursor layer. Thes, is separated from the lamina by the medulla (Md) glia (D). The features ofled in the confocal micrograph shown in D. Neuronal precursors (blue color),hotoreceptor cell axons (green color; arriving via the optic stalk (os)) arel is up. Scale bar in D, 15 µm.

3756

insns

-s

e

hC

yfig.

ar

orsf


tion in glia and neuronal precursors. Upregulation of ptcexpression inral perspective (A-C) or a horizontal perspective (D-F). The specimens arey to visualize glia (blue color; shown separately in A,D), anti-β-

ze tc-lacZ(FE3) reporter expression (red color; shown separately in B,E)al membranes, including retinal axons (green color; all panels). the epithelial glial layer, though some neuronal precursor cells are present.both ptcand Repo, while the β-galactosidase-positive, Repo-negative cellsorizontal perspective, the satellite (S-glia), epithelial (E-glia), marginal

lia) express the ptc-lacZreporter. Subretinal glia, located on the lateral not express the ptc-lacZreporter. Cells of the neuronal precursor layerositive. Significant levels of ptc-lacZexpression are first detectable in of the lamina furrow (lf; see E), where G1-phase LPCs apparently receive

at triggers entry into S-phase. In A-C, anterior is to the left, dorsal is up. Inis up. Scale bars, 30 µm in A for A-C; 10 µm in D for D-F.

additionally express Repo and Dac. In the lamina target field‘eyeless’ mutants, such as eyes absent(eya) or sine oculis(so),Dac expression is not detected and Repo expression is grediminshed (Fig. 3E; Perez and Steller, 1996; Huang and Kun1996).

The migration and early differentiation of lamina gliaare independent of HhEnhanced transcription of the putative Hh receptor, patched(ptc;Nusslein-Volhard and Wieschaus, 1980; Hooper and Scott, 19Nakano et al., 1989; Ingham et al., 1991; Marigo et al., 199Stone et al., 1996) is a universal characteristic of Hh sigreception (Hidalgo and Ingham, 1990; Phillips et al., 199Capdevila et al., 1994; Tabata and Kornberg, 1994; Heberleial., 1995; Goodrich et al., 1996; Marigo et al., 1996b; Vortkamet al., 1996). All classes of glia in the lamina region upregulptc expression in a hh-dependent fashion (Fig. 2; Huang anKunes, 1996). These cells are thus Hh-responsive. As showFig. 2F, all three classes of lamina glia, as well as medulla gthat express a ptc-lacZreporter construct are in close proximityto Hh-bearing retinal axons. We have previously shown that gcell ptc reporter gene expression is not observed in hh− animals(data not shown; Huang and Kunes, 1996). This raises question of whether Hh signalreception is responsible for themigration and/or subsequentmaturation of glia cells.

To determine whether themigration of glial precursorsinto the lamina target field isHh-dependent, we examinedthe distribution of Omb-positive cells in hh− animals.In the wild type, a ‘trail’ ofOmb-positive cells delineatesa path of glia migration fromthe dorsal and vental anlagen(arrowheads in Fig. 3A). Theclonal relationship betweenthe Omb-positive cells in thelamina and in the glial anlagewas confirmed by lineagestudies (S. K., unpublishedobservations) in whichclones were marked by a ‘flp-out’ lacZ reporter construct(Basler and Struhl, 1994).Omb-positive cells of thelamina, the glia anlagen andthe putative migrationpathway were includedwithin the same clones. In aneya1 animal, Omb-positivecells are largely absent fromthe lamina target field (Fig.3B), with the exception of themedulla glia, which aredistinguished by their largersize and medial locationadjacent to the medullaneuropil (not shown). Omb

Fig. 2. Derepression of ptc transcripthe lamina, as viewed from a lateco-labeled with anti-Repo antibodgalactosidase antibody to visualipand anti-HRP to visualize neuron(A-C) The focal plane is through In C, purple cells are expressing are red. (D-F) As seen from the h(Ma-glia) and medulla glia (Md-gsurface of the lamina (see D), do(LPCs) are also β-galactosidase-pneuronal precursors at the trougha retinal axon-mediated signal thD-F, anterior is to the left, lateral

of

atlyes,

89;6a;nal0;n et

patedn inlia,

lia

the

expression in the dorsal and ventral glia anlagen is normal eya animals (Fig. 3B), consistent with previous observationusing other glia precursor markers and ‘eyeless’ mutant strai(e.g. sine oculis1; Perez and Steller, 1996).

We determined whether glia precursor migration is Hhdependent by examining the distribution of Omb-positive cellin hh1 animals. hh1 is a regulatory mutation that specificallyaffects hh expression in the visual system. In hh1 animals,approximately 12 columns of ommatidia initiate differentiationin the eye imaginal disc before the anterior progression of thmorphogenetic furrow ceases (Heberlein et al., 1993). hh1

retinal axons lack Hh immunoreactivity by the time they reacthe lamina target field and thus the Hh-dependent steps of LPmaturation fail to occur in hh1 animals (Huang and Kunes,1996). As shown in Fig. 3C, Omb staining reveals a relativelnormal number of glia precursors in the lamina target field ohh1 animals, despite the absence of Dac induction (see also F3F). The Omb-positive cells are distributed uniformly along thedorsoventral axis among the retinal axon fascicles, but appemore closely spaced than in the wild type. A likely explanationfor this spacing defect is the absence of the neuronal precurs(Huang and Kunes, 1996) that would constitute the majority olamina cells at this point of development.


atalngalse

by, ofalts

rnamtehe

al axon-mediated signal is required for glia migration and differentiation.or cells into the lamina was monitored by anti-Omb antibody stainingB) and hh1 (C) animals. In the wild type (A), Omb-positive gliaen at the prospective dorsal and ventral margins of the lamina (seeths indicated by the arrowheads and disperse among the retinal axonsRP antibody) in the crescent-shaped lamina target field (i.e., the regionlf). (Some cells in the lobula (lob) also express Omb). In the absence ofals (B), Omb-positive cells seem to remain in the glial progenitorter the lamina field appear to stall along the migratory pathway (arrowhead retinal axon fascicles reach the brain but lack active Hh, Omb-positivearge numbers, despite the absence of LPC maturation (as shown in F).ina glia was examined by staining with anti-Repo antibody in wild type. In the wild type (D), Repo-positive glia (red color) are dispersed amongrs (blue color) in the lamina target field. In eyaanimals (E), which lack

-positive cells are found in the lamina target field (as indicated by thePCs are completely absent. In hh1 mutants (F), despite the complete, retinal axons are surrounded by a large number of Repo-positive glia (red these Repo-positive cells is likely due to the absence of the neuronalindicate the lamina furrow. In all panels, anterior is to the left, dorsal is up.0 µm in D for D-F.

To determine whether the glial precursors that enter lamina target field in hh− animals express a retinal innervationdependent marker, we examined their expression of Reponoted above, Repo-positive cells are largely absent in eyaanimals (Fig. 3E). Occasionally, a small number of Reppositive cells are observed at ventral or dorsal lamina positi(the arrowhead in Fig. 3E). In hh1 animals, the Omb-positivecells within the lamina also express Repo (Fig. 3F). Moreovthe Repo-positive cells occupy proper layers above and bethe R1-R6 axon termini expected for satellite, marginal aepithelial glia, though the lack of markers specific for thethree glia types precludes an unambiguous determinationglial cell type. The presence of marginal and epithelial gliaconsistent with the observation that R1-R6 growth conterminate in their proper positions between these layers in hh−

animals (not shown; Huang and Kunes, 1996).The ectopic expression of Hh in the brains of ‘eyeles

animals is sufficient to inducethe initial steps of LPCmaturation in the absence ofretinal axons (Huang andKunes, 1996). For example,photoreceptor differentiationcan be prevented by shifting ahhts2 animal to thenonpermissive condition at anearly larval timepoint. hh+

somatic clones generated by a‘flp-out’ hhconstruct (Zecca etal., 1995) in the brains of theseanimals induce the LPCterminal division and the onsetof Dac expression. Todetermine whether either Hhor the Hh-mediated events ofLPC maturation are sufficientfor glia cell migration andmaturation, we examined suchanimals for the presence ofRepo-positive cells in thevicinity of hh+ clones withinthe lamina target field. Ectopichh+ clones in various positionsthroughout the lamina targetregion, as well as other siteswithin the optic lobe, failed toinduce Repo expression (datanot shown). Thus neither Hhnor a secondary signalprovided by the maturation ofLPCs is sufficient to induceglia cell maturation.

LPCs requiresmoothened for cell cycleprogression and neuronaldifferentiationThe activities of a number ofhh signal transductionpathway components are nowwell characterized (reviewed

Fig. 3. A Hh-independent retin(A-C) Migration of glial precurs(red color) in wild type (A), eya(precursors migrate from anlagschematic in Fig. 1A) along pa(green color; stained by anti-Hposterior of the lamina furrow (retinal innervation, as in eyaanimdomains. The few cells that enin B). In hh1 animals (C), whereglia migrate into the lamina in l(D-F) The differentiation of lam(D), eya(E) and hh1 (F) animalsDac-positive neuronal precursoretinal axons, only a few Repoarrowhead) and Dac-positive Labsence of Dac-positive LPCscolor). The close apposition ofprecursors. Arrows in C and F Scale bars, 40 µm in A for A-C; 4

the-. As

o-ons

er,lowndse of ises

s’

in Ingham, 1995; Hammerschmidt et al., 1997). Mutations these loci have been shown to either mimic or block Hh signreception in a cell-autonomous fashion (see below). Examinithe cellular requirements for these genes in mosaic animshould help illuminate the cellular circuitry that mediates thHh-dependent events of lamina development.

The seven-pass transmembrane protein encoded smoothened(smo; Nusslein-Volhard et al., 1984; Alcedo et al.1996; van den Heuvel and Ingham, 1996) acts downstreamthe Hh receptor Patched as a positive effector of Hh signreception (Chen and Struhl, 1996). If Hh exerts its effecdirectly on LPCs, we would expect that loss of smofunctionshould block the entry of G1-phase LPCs into S-phase and/oprevent the expression of Hh-dependent markers of lamidifferentiation such as Dac. We used the FLP/FRT syste(Golic and Lindquist, 1989; Xu and Rubin, 1993) to generamosaic animals harboring somatic clones homozygous for t

3758

rCsi

l

ior

otild

el

gheishelseus

nlseset is


Fig. 4. smoothenedis required cell-autonomously for the Hh-dependentsteps of LPC maturation. Somaticmosaic clones were generated usingthe FLP/FRT system (as described inMaterials and Methods) and markedby the expression of an arm-lacZreporter gene on the smo+

chromosome arm (blue color,revealed by anti-β-galactosidaseantibody staining). Homozygousmutant cells lack β-galactosidaseexpression. (A-C) Dac expression(red color) was examined in somaticclones homozygous for thesmo3mutation. The boxed areaharboring a smo3 clone in A is shownat higher magnification in B and C. Asmo3 clone that spans the laminafurrow (lf) and includes theanteroventral portion of the laminalacks Dac expression (red color in Aand C). The smo3 cells, visualized bythe absence of anti-β-galactosidasestaining (shown exclusively in B andoutlined in C), are Dac-negative.smo3/smo+ cells immediately bordering the clone and within the lamina field are Dac-positive. The dashed portion of the outline at the posteriorof the clone in C indicates uncertainty at the border, which is obscured by lacZ-positive retinal axons in that area. In addition, it is evident that theanterior progression of the lamina furrow (visualized by anti-HRP staining (green color) as a dorsoventral line running through the clone) isretarded in the region of the clone. (D-F) Ci immunoreactivity was monitored in somatic clones homozygous for the smo3mutation. Ciimmunoreactivity (red color in D and F) was detected with an antibody against its C terminus (see text) and reflects stabilization of the protein byHh signal reception (Aza-Blanc et al., 1997). The boxed area harboring a smo3 clone in D is shown at higher magnification in E and F. smo3 cellsare visualized by the absence of anti-β-galactosidase staining (shown exclusively in E and outlined in F). Two smo3 clones that span the laminafurrow (lf) and include anteroventral portions of the lamina have lower levels of Ci expression (red color in A and C) than smo3/smo+ cellsoutside of the clone. The border between high-level and low-level Ci immunoreactivity precisely coincides with the clonal bordery. Thus smoacts cell autonomously with regard to the stabilization of Ci. lob, lobula. Scale bars, 40 µm in A for A and D; 20 µm in B for B, C, E and F.

null allele smo3 (Nusslein-Volhard et al., 1984). Homozygousmo clones were visualized by the absence of an arm-lacZmarker residing on the smo+ chromosome of a smo3/smo+

animal (blue color in Fig. 4; Vincent et al., 1994). The majoriof smo clones recovered included the region anterior immediately posterior of the lamina furrow. smo cloneslocalized to the anterior of the lamina furrow often appearto slow or block its anterior progression, as indicated relative posterior position of the lamina furrow in the vicinitof such clones (Fig. 4A). Clones spanning the lamina furrwere particularly informative, as smo cells posterior of thefurrow (i.e., within the lamina) were in all cases Dac-negat(Fig. 4A,C). In all such cases examined, smo+ cells adjacent tothese clones within the lamina were Dac-positive. Thus, wrespect to lamina differentiation, smoacts cell autonomously.smoclones that extended to the posterior of the lamina wrare. It is possible that LPCs that cannot respond to Hh arereadily incorporated into the lamina and displaced by smo+

LPCs. LPCs that are unable to respond to Hh might eliminated by cell death, as occurs in ‘eyeless’ mutant anim(Selleck et al., 1992).

A hallmark of Hh signal reception in many Drosophiltissues is an increase in immunoreactivity to the C-termiportion of the protein Ci, a transcriptional mediator of Hsignaling (Motzny and Holmgren, 1995; Johnson et al., 19Slusarski et al., 1995). This enhanced Ci immunoreactivity

s

tyor

edbyyow

ive

ith

ere not

beals

analh

95; is

due to inhibition of Ci proteolytic processing, a cellularesponse to Hh signal reception (Aza-Blanc et al., 1997). LPposterior of the lamina furrow display the enhanced Cimmunoreactivity that would be predicted for Hh signareception by LPCs (see Fig. 4D,F). In animals in which hh−

retinal axons innervate the lamina target field, cells posterto the lamina furrow display a level of Ci immunoreactivityequivalent to the basal level detected anterior to the furrow (nshown), indicating that the increased Ci observed in the wtype is Hh-dependent. In smomosaic animals, smocells eitheranterior or posterior of the lamina furrow display a basal levof Ci immunoreactivity, (Fig. 4F) while smo+ cellsimmediately adjacent to the portion of smoclones within thelamina display the high Hh-dependent level.

The initial response of LPCs to the arrival of Hh-bearinretinal axons would appear to be entry into S-phase at tlamina furrow. To determine whether cell cycle progression directly dependent on Hh signal reception, we examined tincorporation of bromodeoxyuridine (BrdU) into S-phase celin smo mosaic animals. In the wild type, LPCs that haventered their terminal S-phase form a discrete and continuoband at the posterior margin of the lamina furrow (Fig. 5A). I‘eyeless’ animals lacking photoreceptor innervation or animain which photoreceptor axons lacking functional Hh enter thlamina target field, only a low background of scattered S-phacells are detected (see Fig. 7H; Selleck and Steller, 1991). I


).h

ofandnd

or

st

etor

edal

d agw.

unclear whether the products of these scattered divisionsincorporated into the lamina (i.e., that these cells are indLPCs; see Selleck et al., 1992, for further discussion of thcells). In smo3 mosaic animals, mutant clones that included tposterior margin of the lamina furrow lacked S-phase LP(Fig. 5B). In contrast, the scattered S-phase cells anterior tolamina furrow, and the distribution of S-phase cells in othproliferation centers, such as the OPC, were unaffected byloss of smo function. At the lamina furrow, smo+ cellsbordering smoclones were often found in S-phase. Thus, sum, smo+ behaved as a cell-autonomous requirement LPCs to initiate the Hh-dependent steps of lamidifferentiation.

Activation of the hh signaling pathway results incell-autonomous LPC maturationThe Hh receptor Ptc, a multiple-pass membrane protein, the cAMP-dependent protein kinase (PKA) normally maintathe Hh signal transduction pathway in a repressed s

antfnion6).oth

a

use

n.ellt

is-

oes

isCi

6;. Ci of

al.,).toithn

of

Fig. 5. smois required cell-autonomously for LPCs to enter S-phasCells in S-phase were identified by incubating whole-mount tissueimmediately after dissection in culture medium containing BrdU, adescribed in Materials and Methods. (A) In the wild type, LPCs intheir terminal S-phase are visualized as a discrete band (red coloanti-BrdU antibody staining) that runs along the posterior margin the lamina furrow (lf). Retinal axons and other neuronal membranare visualized by anti-HRP antibody staining (green color). S-phacells of the outer proliferation center (OPC) and of the innerproliferation center (adjacent to the lobula, lob) are also visible to anterior and posterior, respectively, of the lamina. (B-B″) In a smo3

mosaic specimen, homozygous smo3 cells are marked by the absencof arm-lacZreporter gene expression (blue color in B′and B″; seethe legend to Fig. 4 for details). In smo3 clones that span the laminafurrow, LPCs are not detected in S-phase at the posterior margin the lamina furrow (red color in B and B″), in contrast to smo3/smo+

cells immediately bordering the mutant clones. Other cellproliferation centers, such as the OPC, are not affected by the lossmofunction. In all panels, anterior is to the left, dorsal is up. Scalbar, 20 µm in A for all panels.

areeedeseheCs theer the

inforna

andintate

(reviewed in Ingham, 1995; Hammerschmidt et al., 1997Loss-of-function mutations in either of these genes mimic Hsignal reception and result in the cell autonomous activationHh target genes in many tissues (Ingham et al., 1991; Jiang Struhl, 1995; Lepage et al., 1995; Li et al., 1995; Pan aRubin, 1995; Strutt et al., 1995). If LPC differentiation isindeed induced directly by Hh, LPCs harboring mutations feither pka or ptc should undergo differentiation cell-autonomously and independently of retinal innervation. To tethis prediction, we generated somatic clones for pka or ptc ina hhts2 mutant background, where an early shift to thnonpermissive temperature was used to block photorecepcell differentiation. Clones homozygous for the pkaalleleDC0h2 or the ptcallele, ptc6p, behaved similarly. In both cases,cells within homozygous clones in the lamina region expresslamina differentiation markers despite the absence of retinaxons, while neighboring pka+ or ptc+ cells did not (Fig. 6A-C and 6D-F, respectively). The marked clones often spanneregion including the OPC and the LPCs, containinpopulations both anterior and posterior of the lamina furroIn these clones, cell-autonomous induction of the lamindifferentiation marker Dac was observed exclusively in mutacells posterior of the lamina furrow. Mutant cells anterior othe furrow did not differentiate precociously. This observatiois consistent with the consequences of ectopic Hh expressin an ‘eyeless’ mutant background (Huang and Kunes, 199Hh expression in regions anterior of the lamina furrow did ninduce precocious lamina differentiation, as thoug‘competence’ to respond to Hh is acquired by G1-phase LPCsat the anterior margin of the lamina furrow. Within the lamintarget field, wild-type cells neighboring the DC0h2 or ptc6p

mutant cells were never observed to express Dac. Thactivation of the Hh pathway by loss-of-function in either genresults in a strictly autonomous induction of LPC maturatioThese results permit the conclusion that the terminal cdivision and differentiation of LPCs both require the direcreception of the Hh signal.

LPC cell cycle progression and cell fatedetermination are jointly controlled by thetranscriptional regulator, Cubitus interruptusIn a number of instances, pattern formation mediated by Hhaccompanied by cell division. The well-defined pattern of Hhinduced cell division in the lamina provides an opportunity tdetermine the point at which the Hh signal reception engagthe cell cycle machinery. Biochemical and epistasexperiments have placed the zinc finger molecule downstream of all other hh signaling pathway components(Motzny and Holmgren, 1995; Sanchez-Herrero et al., 199Robbins et al., 1997; Sisson et al., 1997; Chen et al., 1998)has been shown to bind directly to the regulatory sequencesHh-responsive genes (Alexandre et al., 1996; Von Ohlen et 1997; Von Ohlen and Hooper, 1997; Aza-Blanc et al., 1997Should all Hh-mediated events of LPC maturation be found depend on Ci function, we could conclude that, at least wregard to cell proliferation and the expression of differentiatiomarkers, there is no ‘branchpoint’ within the signalingpathway.

To examine the requirement for Ci, we utilized tworecombinant constructs that result in either dominant Cigain-of-function or loss-of-function phenotypes. Overexpression

e.ss

r;ofesse

the

e

of

s ofe

3760

dctea

hltse

erte,e

eslt

e’

anecndCs,s,inr,


ivation of the Hh signaling pathway triggers cell-autonomous LPC. Dac expression (red color) was examined in somatic clonesus for pka-DC0h2 (A-C) or ptc6p (D-F). Mosaicism was induced by theystem in a homozygous hhts2 background to block R-cell formation (asin Materials and Methods). These animals thus lack retinal axons (asthe absence of anti-HRP-positive photoreceptor axons (green color) ind E). Mutant somatic clones, marked by the absence of arm-lacZpression (blue color) extend into the lamina field and are outlined in B animals harboring either pka-DC0h2 (A-C) or ptc6p (D-F) clones,us mutant cells within the lamina target field express Dac (red color;ntibody staining). Homozygous mutant cells outside of the lamina targetgative, as explained in the text. The effects of the mutations are strictlymous, as wild-type cells bordering the clones, but within the lamina, do not express Dac. lob, lobula. In all panels, anterior is to the left,p. Scale bar, 20 µm in A for all panels.

the wild-type Cigene results in a gain-of-function phenotypthat mimics activation of the Hh signaling pathway (Alexandet al., 1996; Domiguez et al., 1996; Hepker et al., 199Expression of an amino terminal fragment of Ci (hereaftreferred to as DN-Ci) results in a dominant loss-of-functiophenotype, as the normal in vivo function of this portion of thmolecule appears to be transcriptional repression of Hh targenes (Aza-Blanc et al., 1997; Hepker et al., 1997). Boththese constructs employ the yeast UASpromoter, so that theycan be expressed in somatic clones of GAL4-expressing cells(Brand and Perrimon, 1993) generated by a ‘flp-out’ GAL4construct (Pignoni and Zipursky, 1997).

With either construct, ectopic expression in the laminregion resulted in the expected phenotype with respect to lamina differentiation marker Dac. Dac expression in ceposterior of the lamina furrow was strongly reduced undetectable in cells expressing DN-Ci (Fig. 7B). Conversethe ectopic expression of wild-type Ci resulted in the inductiof Dac-positive cells in the lamina target field of ‘eyelesanimals (Fig. 7I). The effects observed with either construwere strictly cell autonomous (Fig. 7B,C; data not shownThus our results with ectopic Ci and DN-Ci expression aconsistent with the expectation that Ci modulates Hh signalactivity directly in LPCs.

To determine whether Hh signaling acts via Ci to regulathe G1- to S-phase transition of LPCs at the lamina furrow, wexamined the incorporation of BrdU into S-phase cells animals harboring clones expressing either of thetwo constructs described above. As shown in Fig.7E, cells expressing DN-Ci at the posterior marginof the lamina furrow failed to enter S-phase.Where clones of DN-Ci-expressing cells traversedthe lamina furrow, S-phase LPCs are absent (Fig.7E), while S-phase LPCs are observedimmediately outside of the clone (arrowhead inFig. 7E). Moreover, the effect on cell division waslimited to the LPCs at the lamina furrow. Nodefects were observed in other proliferation zonessuch as the OPC or IPC, the other majorproliferation centers of the optic lobe when theycontained DN-Ci-expressing cells. Conversely, theinduction of lamina differentiation by ectopic Ciexpression in a hhts2 (‘eyeless’; as described inMaterials and Methods) background wasaccompanied by the entry of LPCs into S-phase atthe lamina furrow (Fig. 7I). As when laminadifferentiation was induced in the absence ofretinal axons by ectopic Hh expression (Huang andKunes, 1996), ectopic Ci expression triggers aposterior-to-anterior pattern of differentiation suchthat S-phase LPCs are found at the anterior margin(arrow in Fig. 7I). In sum, these observationsindicate that the induction of cell division by Hhoccurs via the transcriptional regulation of Hhtarget genes.

DISCUSSION

The assembly of a lamina cartridge, an intricatestructure composed of five lamina neurons and six

Fig. 6. ActmaturationhomozygoFLP/FRT sdescribed shown by A, B, D, anreporter exand E. Forhomozygoanti-Dac aare Dac-necell-autonotarget fielddorsal is u

ere7).erneget of

athellsorly,ons’ct).reing

teein

photoreceptor cell axons (reviewed in Meinertzhagen anHanson, 1993), can be viewed as occurring in two distinstages. Initially, an arriving ommatidial axon fascicle elicits thformation of a precartridge cellular ensemble (a lamincolumn; Fig. 1C) by triggering the terminal differentiation ofboth glial and neuronal precursors. The activity of eacommatidial fascicle in establishing a cartridge ensemble resuin a one-to-one correspondence of ommatidial and cartridgunits. The second stage of cartridge formation occurs latwhen the six R1-R6 axons of an ommatidial fascicle separamigrate laterally and form their synaptic connections with thneurons of six neighboring cartridges (Trujillo-Cenoz andMelamed, 1973). In the adult lamina a cartridge thus receivits complement of R1-R6 axons from six different ommatidiaunits, none of which were members of the fascicle thaoriginally established the cartridge cell ensemble. Thmechanism by which this remarkable feat of ‘axon-shufflingoccurs is unknown.

Here we have focused on the signaling events by which ommatidial axon fascicle elicits the assembly of a cartridgensemble. It is clear that this process involves specificommunication events between the axons and neuronal aglial precursors. Retinal axon-mediated signals trigger LPmaturation (Selleck and Steller, 1991; Huang and Kune1996), the specification of L1-L5 neurons (Huang and Kune1996), and the migration and subsequent maturation of glia the lamina target field (Winberg et al., 1992; Perez and Stelle


tus (Ci) regulates both LPC cell division and differentiation. (A-C) Dacn the lamina was examined in the wild type (A) and specimens harboringing a dominant-negative Ci (DN-Ci) transgene, UAS-Ci76 (B,C; Aza-In the wild type, Dac expression (red color) is detected in LPCsf the lamina furrow (lf). (B,C) In this specimen, clones expressing DN-bsence of CD2 immunoreactivity (blue color shown in C; region left ofe LPCs expressing DN-Ci adjacent to the lamina furrow (arrow) fail tos that are not included in the DN-Ci-expressing clones are Dac-positive.

tonomously. (D-F) The effect of DN-Ci expression on LPC cell cycleored by analyzing the incorporation of BrdU into whole-mount tissues inthe wild type, S-phase LPCs form a discrete band on the posteriorrrow (lf), just anterior to the crescent-shaped array of retinal axons by anti-HRP staining). (E,F) In somatic clones expressing DN-Cilack of CD2 immunoreactivity – blue color, shown in F, to the left of these LPCs are entirely absent, while cells that do not express the transgenerdU-positive. (G-I) Ectopic expression of a wild-type Ci transgenen in an ‘eyeless’ animal. Dac expression (blue color) and BrdUase cells (red color) were monitored in the wild type (G), an eyaanimal), and an animal expressing a wild-type Ci transgene (I). In the wild

ntered S-phase at the posterior margin of the lamina furrow (lf) are purpler; Dac-positive, blue color). Postmitotic LPCs within the field of retinali-HRP antibody staining) are Dac-positive. In an eyeless animal (H), S-xpression are completely absent. In an ‘eyeless’ hhts2 animal expressing aac expression in the lamina field occurs in the absence of retinal axons.f the lamina (indicated by an arrow), LPCs are observed in S-phasectopic Ci expression also results in abnormalities in the organization of

by the unusual pattern of S-phase cells in that region. lob, lobula. In alle left, dorsal is up. Scale bar, in A, 40 µm for all panels.

1996). Here we have addressed the question of whether tevents involve the same or different retinal-axon mediatsignals, and whether these signals actdirectly or via the induction ofsecondary signals from resident cellpopulations.

Hedgehog is one of the signals thatretinal axons deliver to the lamina(Huang and Kunes, 1996). Hhimmunoreactivity is found on retinalaxons at the time of their ingrowthinto the lamina. Hh is required forlamina differentiation (see Fig. 3, forexample) and is sufficient for theonset of LPC maturation whenexpressed ectopically in the brain.Nonetheless, in many tissues Hhexerts at least some of its effectsthrough the induction or maintenanceof secondary signals. In the embryo,for example, Hh maintains winglessexpression, thus exerting an indirecteffect on anterior compartmentpattern (Hidalgo and Ingham, 1990;Ingham and Hidalgo, 1993). In thewing imaginal disc, Hh-mediatedactivation of dppexpression along theAP compartment border yields amorphogen gradient with activity inboth compartments (Lecuit et al.,1996; Nellen et al., 1996; Singer et al.,1997). Finally, in the eye, ommatidialdevelopment can take place in cellslacking smoothened function,suggesting that the involvment of Hhis indirect (Strutt and Mlodzik, 1997).This raises the question of whether Hhis a direct signal for cartridge neurondifferentiation or acts via theinduction of a secondary signal,perhaps in some other cell population.

Our data indicate that neuronalprecursor (LPC) differentiation in thelamina is due to direct Hh signalreception. A mosaic analysis with fourHh signaling pathway components(smoothened, patched, ProteinKinase-A and cubitus interruptus)revealed in all cases a cell-autonomous requirement for Hhsignal reception in LPCs. Sincelamina neurons and glia derive fromdistinct lineages (Winberg et al. 1993;Perez and Steller, 1996), theseobservations rule out a requirementfor Hh signal reception in glia for LPCmaturation. The requirement forsmoothenedfor LPCs to enter S-phaseand express lamina differentiationmarkers was observed in relativelysmall somatic clones that coincided

Fig. 7. Cubitus interrupexpression (red color) isomatic clones expressBlanc et al., 1997). (A) immediately posterior oCi are marked by the athe line drawn in B). Thexpress Dac, while LPCDN-Ci thus acts cell auprogression was monitculture medium. (D) In margin of the lamina fu(green color; visualized(clones marked by the line drawn in E), S-pha(arrowhead in E) are Binduces LPC maturatioincorporation into S-phlacking retinal axons (Htype, LPCs that have e(BrdU-positive, red coloaxons (green color, antphase LPCs and Dac eUAS-Ci transgene (I), DAt the anterior margin o(purple colored cells). Ethe OPC, as indicated panels, anterior is to th

heseed

with the location of G1-phase LPCs (Figs 4, 5). Wild-typeLPCs immediately adjacent to those smo cells commenced

3762

in aedn,s-blediset of

a

ne,etd

ll

.

l

ed.

,

s


lamina differentiation normally. Conversely, in the absenceretinal axons, activation of the Hh signaling pathway via loof function in patchedor Protein Kinase-Atriggered LPCmaturation in a cell-autonomous fashion (Fig. 6). Thesomatic clones did not induce differentiative events in adjacwild-type LPCs, which failed to undergo differentiation due the absence of retinal axons. This observations argue stroagainst the notion that Hh signal reception triggers texpression of a secondary diffusible signal that is sufficient LPC maturation.

Hh signal reception is in a number of instances accompanby cell proliferation. In the chick, Sonic Hedgehog triggers cproliferation in conjunction with its patterning activity in thsomitic mesoderm (Fan et al., 1995). In the mouse, losspatchedfunction is associated with enhanced cell proliferatioin the neural tube and a higher incidence of a proliferatdisease, medulloblastoma (Goodrich et al., 1997). Drosophila, morphogenesis induced by Hh in the developeye includes two waves of cell division, one anterior and oposterior of the morphogenetic furrow (Ready et al., 197Heberlein et al., 1995). Our data indicate that, at least in case of the lamina, Hh acts via the transcription factor Ciregulating cell cycle progression (Fig. 7). Both geneepistasis and biochemical experiments place Ci at a termstep in Hh signal transduction, as a nuclear effector of upstream components Ptc, Smo, PKA, and Fused (Motzny Holmgren, 1995; Sanchez-Herrero et al., 1996; Robbins et1997; Sisson et al., 1997; Chen et al., 1998). Thus we concthat, at least with regard to cell proliferation and the expressof differentiation markers, there is no ‘branchpoint’ within thsignaling pathway. Rather, cell proliferation would appear be controlled by a transcriptional target of Hh. This might a component of the cell cycle regulatory machinery, but ledirect avenues are also possible.

There is evidence for the activity of at least two additionretinal axon-mediated signals in lamina cartridge assemFirst, retinal axons appear to provide at least one sigdownstream of Hh that mediates the final differentiation cartridge neurons (Huang and Kunes, 1996). Only a subsethe postmitotic LPCs generated by Hh-dependent evebecome cartridge neurons. Five postmitotic LPCs associawith each retinal axon fascicle begin to express neuromarkers (Fig. 1C); the remaining LPCs are eliminated apoptosis. Hh is not sufficient for the terminal differentiatioof cartridge neurons. Our recent observations (Z. unpublished data) indicate that a second retinal axon-medisignal is required and that it acts via the Drosophila EGFreceptor. A third retinal axon-mediated signal is apparenrequired for lamina glial differentiation. Though responsive retinal axon-borne Hh (Fig. 2), the satellite, marginal aepithelial glia depend on retinal axons, but not Hh, in ordermigrate into the lamina target field and undergo subsequdifferentiation (Fig. 3). Because the hh− retinal axons that canpromote these events cannot trigger LPC maturation, suppose that this third signal is independent of the Hh-mediaevents of LPC maturation (Fig. 3). These events are aindependent of EGF receptor activity (our unpublishobservations). Thus we suppose that at least three retinal amediated signals orchestrate the assembly of a lamina cartri

The transmission of signals from the eye to the brain imechanism by which the precision of ommatidial assembly

ofss

seenttonglyhefor

iedelle ofn

iveIn

ingne6;the inticinaltheand al.,ludeioneto

bess

albly.naloft ofntstednalbyn

H.,ated

tlytond toent

wetedlso

edxon-dge.s a in

the eye is imposed on the developing postsynaptic field. As the case of the eye (Ready et al., 1976), the assembly ofcartridge cellular ensemble proceeds according a stereotypchoreography of cellular events (Meinertzhagen and Hanso1993). At the anterior of the lamina, a retinal axon fascicle ijoined first by a cell destined to become the L1 neuron. Lneurons appear to be added to the fascicle-associated ensemin a cell-type-dependent sequence (Meinertzhagen anHanson, 1993; S. Kunes, unpublished observations). Thprocess yields a precise geometric array of retinal axon, targcell associations that set the stage for the remarkable eventssynaptic cartridge formation that follow. One might supposethen that to some extent the presynaptic axons serve as‘template’ for the establishment of order in the lamina.Conversely, it is clear that target field cells play some role ithe establishment of precise connectivity, for example in thinitial guidance of photoreceptor axons (see, for exampleMartin et al., 1995) and the maintenance of growth cone, targcell associations via the synthesis of nitric oxide (Gibbs anTruman, 1998). Thus, in the Drosophilavisual system, there isa complex interplay between pre- and postsynaptic cepopulations, many aspects of which are as yet unresolved.

We are indebted to M. Mlodzik, Y. H. Sun, R. Holmgren, D.Kalderon T. Kornberg, G. Pflugfelder, G. Technau, P. Ingham, GMardon, G. Rubin, L. Zipursky, K. Basler, G. Struhl and K. Mathews(Drosophila Stock Center, Bloomington) for strains and antibodyreagents. We thank members of the Kunes laboratory for criticareading of the manuscript. This work was supported by a PewScholars Award and NIH-NEI grant EY10112 to S. K.; Z. H. wassupported by a Corning Costar Fellowship Fund award during thcourse of this work. We are especially indebted to Mr Jeff Tarr anthe Markey Foundation for their support of the MCB Imaging Center

REFERENCES

Alcedo, J., Ayzenzon, M., Von Ohlen, T., Noll, M. and Hooper, J. E.(1996).The Drosophila smoothenedgene encodes a seven-pass membrane proteina putative receptor for the hedgehogsignal. Cell 86, 221-232.

Alexandre, C., Jacinto, A. and Ingham, P. W. (1996). Transcriptionalactivation of hedgehogtarget genes in Drosophila is mediated directly bythe Cubitus interruptus protein, a member of the GLI family of zinc fingerDNA-binding proteins. Genes Dev. 10, 2003-2013.

Aza-Blanc, P., Ramirez-Weber, F. A., Laget, M. P., Schwartz, C. andKornberg, T. B. (1997). Proteolysis that is inhibited by hedgehogtargetsCubitus interruptus protein to the nucleus and converts it to a repressor. Cell89, 1043-1053.

Basler, K. and Struhl, G. (1994). Compartment boundaries and the controlof Drosophila limb pattern by hedgehogprotein. Nature368, 208-214.

Braitenberg, V. (1967). Patterns of projection in the visual system of the fly.I. Retina-lamina projections. Exp. Brain Res.3, 271-298.

Brand, A. H. and Perrimon, N. (1993). Targeted gene expression as a meanof altering cell fates and generating dominant phenotypes. Development 118,401-415.

Capdevila, J., Estrada, M. P., Sanchez-Herrero, E. and Guerrero, I.(1994).The Drosophila segment polarity gene patchedinteracts withdecapentaplegicin wing development. EMBO J. 13, 71-82.

Capdevila, J. and Guerrero, I.(1994). Targeted expression of the signallingmolecule decapentaplegic induces pattern duplications and growthalterations in Drosophilawings. EMBO J.13, 4459-4468.

Chen, Y. and Struhl, G. (1996). Dual roles for patchedin sequestering andtransducing Hedgehog. Cell 87, 553-563.

Chen, Y., Gallaher, N., Goodman, R. H. and Smolik, S. M.(1998). Proteinkinase A directly regulates the activity and proteolysis of Cubitusinterruptus. Proc. Natl. Acad. Sci. USA95, 2349-2354.

Dominguez, M., Brunner, M., Hafen, E. and Basler, K.(1996). Sending and


s

,

r.

receiving the hedgehogsignal: control by the DrosophilaGli protein Cubitusinterruptus. Science272, 1621-1625.

Dominguez, M. and Hafen, E.(1997). Hedgehog directly controls initiationand propagation of retinal differentiation in the Drosophilaeye. Genes Dev.11, 3254-3264.

Eaton, S. and Kornberg, T. B. (1990). Repression of ci-D in posteriorcompartments of Drosophilaby engrailed. Genes Dev. 4, 1068-1077.

Falls, D. L., Rosen, K. M., Corfas, G., Lane, W. S. and Fischbach, G. D.(1993). ARIA, a protein that stimulates acetylcholine receptor synthesisa member of the neu ligand family. Cell 72, 801-815.

Fan, C. M., Porter, J. A., Chiang, C., Chang, D. T., Beachy, P. A. andTessier-Lavigne, M. (1995). Long-range sclerotome induction by Sonihedgehog: direct role of the amino-terminal cleavage product amodulation by the cyclic AMP signaling pathway. Cell 81, 457-465.

Fischbach, K. F. and Technau, G.(1984). Cell degeneration in the developingoptic lobes of the sine oculisand small-optic-lobesmutants of Drosophilamelanogaster. Dev. Biol.104, 219-239.

Gibbs, S. M. and Truman, J. W.(1998). Nitric oxide and cyclic GMP regulateretinal patterning in the optic lobe of Drosophila. Neuron20, 83-93.

Golic, K. G. and Lindquist, S. (1989). The FLP recombinase of yeascatalyzes site-specific recombination in the Drosophilagenome. Cell59,499-509.

Goodrich, L. V., Johnson, R. L., Milenkovic, L., McMahon, J. A. and Scott,M. P. (1996). Conservation of the hedgehog/patchedsignaling pathwayfrom flies to mice: induction of a mouse patchedgene by hedgehog. GenesDev. 10, 301-312.

Goodrich, L. V., Milenkovic, L., Higgins, K. M. and Scott, M. P. (1997).Altered neural cell fates and medulloblastoma in mouse patchedmutants.Science277, 1109-1113.

Grimm, S. and Pflugfelder, G. O.(1996). Control of the gene optomotor-blind in Drosophila wing development by decapentaplegicand wingless.Science271, 1601-1604.

Halter, D. A., Urban, J., Rickert, C., Ner, S. S., Ito, K., Travers, A. A. andTechnau, G. M. (1995). The homeobox gene repo is required for thedifferentiation and maintenance of glia function in the embryonic nervosystem of Drosophila melanogaster. Development121, 317-332.

Hammerschmidt, M., Brook, A. and McMahon, A. P. (1997). The worldaccording to hedgehog. Trends Genet. 13, 14-21.

Heberlein, U., Wolff, T. and Rubin, G.(1993). The TGF-βhomolog dppandthe segment polarity gene hedgehogare required for propagation of amorphogenetic wave in the Drosophilaretina. Cell75, 913-926.

Heberlein, U., Singh, C. M., Luk, A. Y. and Donohoe, T. J.(1995). Growthand differentiation in the Drosophilaeye coordinated by hedgehog. Nature373, 709-711.

Heisenberg, M. (1972). Behavioural diagnostics: a way to analyze visumutants in Drosophila. In Information Processing in the Visual System. (ed.R. Wehner), pp. 265-268. Berlin: Springer Verlag.

Hepker, J., Wang, Q. T., Motzny, C. K., Holmgren, R. and Orenic, T. V.(1997). Drosophila cubitus interruptusforms a negative feedback loop withpatchedand regulates expression of Hedgehog target genes. Development124, 549-558.

Hidalgo, A. and Ingham, P. (1990). Cell patterning in the Drosophilasegment: spatial regulation of the segment polarity gene patched.Development110, 291-301.

Hooper, J. E. and Scott, M. P.(1989). The Drosophila patchedgene encodesa putative membrane protein required for segmental patterning. Cell 59, 751-765.

Huang, Z and Kunes, S.(1996). Hedgehog, transmitted along retinal axontriggers neurogenesis in the developing visual centers of the Drosophilabrain. Cell 86, 411-422.

Ingham, P. W. (1995). Signalling by hedgehogfamily proteins in Drosophilaand vertebrate development. Curr. Opin. Genet. Dev.5, 492-498.

Ingham, P. W. and Hidalgo, A.(1993). Regulation of winglesstranscriptionin the Drosophilaembryo. Development117, 283-291.

Ingham, P. W., Taylor, A. M. and Nakano, Y.(1991). Role of the Drosophilapatchedgene in positional signalling. Nature353,184-187.

Jiang, J. and Struhl, G.(1995). Protein kinase A and Hedgehog signaling Drosophila limb development. Cell80, 563-572.

Johnson, R. L., Grenier, J. K. and Scott, M. P. (1995). patchedoverexpression alters wing disc size and pattern: transcriptional and ptranscriptional effects on hedgehogtargets. Development121, 4161-4170.

Kaphingst, K. and Kunes, S.(1994). Pattern formation in the visual centerof the Drosophilabrain: winglessacts via decapentaplegicto specify thedorsoventral axis. Cell 78, 437-448.

, is

cnd

t

us

al

s,

in

ost-

s

Kelly, J. P. and Cowan, W. M.(1972). Studies on the development of thechick optic tectum. 3. Effects of early eye removal. Brain Res.42, 263-288.

Kunes, S. and Steller, H.(1993). Topography in the Drosophilavisual system.Curr. Opin. Neurobiol. 3, 53-59.

Kunes, S., Wilson, C. and Steller, H.(1993). Independent guidance of retinalaxons in the developing visual system of Drosophila. J. Neurosci. 13, 752-767.

Lecuit, T., Brook, W. J., Ng, M., Calleja, M., Sun, H. and Cohen, S.M.(1996). Two distinct mechanisms for long-range patterning byDecapentaplegic in the Drosophilawing. Nature381, 387-393.

Lee, J. J., von Kessler, D. P., Parks, S. and Beachy, P.A.(1992). Secretionand localized transcription suggest a role in positional signaling for productof the segmentation gene hedgehog. Cell71, 33-50.

Lepage, T., Cohen, S. M., Diaz-Benjumea, F. J. and Parkhurst, S. M.(1995). Signal transduction by cAMP-dependent protein kinase A inDrosophila limb patterning. Nature373, 711-715.

Li, W., Ohlmeyer, J. T., Lane, M. E. and Kalderon, D.(1995). Function ofprotein kinase A in Hedgehog signal transduction and Drosophilaimaginaldisc development. Cell 80, 553-562.

Ma, C., Zhou, Y., Beachy, P. A. and Moses, K.(1993). The segment polaritygene hedgehogis required for progression of the morphogenetic furrow inthe developing Drosophilaeye. Cell 75, 927-938.

Macagno, E. R. (1979). Cellular interactions and pattern formation in thedevelopment of the visual system of Daphnia magna (CrustaceaBrachiopoda). I. Interactions between embryonic retinular fibers andlaminar neurons. Dev. Biol. 73, 206-238.

Mardon, G., Solomon, N. M. and Rubin, G. M.(1994). dachshundencodesa nuclear protein required for normal eye and leg development inDrosophila. Development120, 3473-3486.

Marigo, V., Davey, R. A., Zuo, Y., Cunningham, J. M. and Tabin, C. J.(1996a). Biochemical evidence that Patched is the Hedgehog receptoNature384, 176-179.

Marigo, V., Scott, M. P., Johnson, R. L., Goodrich, L. V. and Tabin, C. J.(1996b). Conservation in hedgehogsignaling: Induction of a chickenpatchedhomolog by Sonic hedgehogin the developing limb. Development122, 1225-1233.

Martin, K. A., Poeck, B., Roth, H., Ebens, A. J., Ballard, L. C. andZipursky, S. L. (1995). Mutations disrupting neuronal connectivity in theDrosophilavisual system. Neuron14, 229-240.

Meinertzhagen, I. A. and Hanson, T. E.(1993). The development of the opticlobe. In The Development of Drosophila melanogaster. (ed. M. Bate and A.Martinez-Arias), pp. 1363-1491. Cold Spring Harbor, New York: ColdSpring Harbor Laboratory Press.

Meinertzhagen, I. A. and O’Neil, S. D.(1991). Synaptic organization ofcolumnar elements in the lamina of the wild type in Drosophilamelanogaster. J. Comp. Neurol. 305, 232-263.

Mohler, J. and Vani, K. (1992). Molecular organization and embryonicexpression of the hedgehoggene involved in cell-cell communication insegmental patterning of Drosophila. Development115, 957-971.

Motzny, C. K. and Holmgren, R.(1995). The DrosophilaCubitus interruptusprotein and its role in the winglessand hedgehogsignal transductionpathways. Mech. Dev. 52, 137-150.

Nakano, Y., Guerrero, I., Hidalgo, A., Taylor, A., Whittle, J. R. andIngham, P. W. (1989). A protein with several possible membrane-spanningdomains encoded by the Drosophilasegment polarity gene patched. Nature341, 508-513.

Nellen, D., Burke, R., Struhl, G. and Basler, K.(1996). Direct and long-range action of a DPP morphogen gradient. Cell 85, 357-368.

Nusslein-Volhard, C. and Wieschaus, E.(1980). Mutations affectingsegment number and polarity in Drosophila. Nature287, 795-801.

Nüsslein-Volhard, C., Wieschaus, E. and Kluding, H.(1984). Mutationsaffecting the pattern of the larval cuticle in Drosophila melanogaster:zygotic loci on the second chromosome. Roux’s Arch. Dev. Biol.193, 267-282.

Oland, L. A. and Tolbert, L. P. (1996). Multiple factors shape developmentof olfactory glomeruli: insights from an insect model system. J. Neurobiol.30, 92-109.

Orenic, T. V., Slusarski, D. C., Kroll, K. L. and Holmgren, R. A. (1990).Cloning and characterization of the segment polarity gene cubitusinterruptus Dominantof Drosophila. Genes Dev. 4, 1053-1067.

Pan, D. and Rubin, G. M. (1995). cAMP-dependent protein kinase andhedgehogact antagonistically in regulating decapentaplegictranscription inDrosophila imaginal discs. Cell80, 543-552.

3764

n

get

e-

e


Perez, S. E. and Steller, H.(1996). Migration of glial cells into the retinalaxon target field in Drosophila melanogaster. J. Neurobiol. 30, 359-373.

Pflugfelder, G. O., Schwarz, H., Roth, H., Poeck, B., Sigl, A., Kerscher, S.,Jonschker, B., Pak, W. L. and Heisenberg, M.(1990). Genetic andmolecular characterization of the optomotor-blind gene locus in Drosophilamelanogaster. Genetics 126, 91-104.

Phillips, R. G., Roberts, I. J. H., Ingham. P. W. and Whittle, J. R. S.(1990).The Drosophila segment polarity gene patchedis involved in a position-signalling mechanism in imaginal discs. Development110, 105-114.

Pignoni, F. and Zipursky, S. L. (1997). Induction of Drosophila eyedevelopment by decapentaplegic. Development124, 271-278.

Poeck, B., Hofbauer, A. and Pflugfelder, G. O.(1993). Expression of theDrosophila optomotor-blindgene transcript in neuronal and glial cells of thdeveloping nervous system. Development117, 1017-1029

Power, M. E. (1943). The effect of reduction in numbers of ommatidia upothe brain of Drosophila melanogaster. J. Exp. Zool. 94, 33-72.

Ready, D. F., Hanson, T. E. and Benzer, S.(1976). Development of theDrosophilaretina, a neurocrystalline lattice. Dev. Biol. 53, 217-240.

Robbins, D. J., Nybakken, K. E., Kobayashi, R., Sisson, J. C., Bishop, J.M. and Therond, P. P. (1997). Hedgehog elicits signal transduction bmeans of a large complex containing the kinesin-related protein CostCell 90, 225-234.

Rupp, F., Payan, D. G., Magill-Solc, C., Cowan, D. M. and Scheller, R. H.(1991). Structure and expression of a rat agrin. Neuron6, 811-823

Sanchez-Herrero, E., Couso, J. P., Capdevila, J. and Guerrero, I.(1996).The fu gene discriminates between pathways to control dpp expression inDrosophila imaginal discs. Mech. Dev. 55, 159-170.

Selleck, S. B. and Steller, H.(1991). The influence of retinal innervation onneurogenesis in the first optic ganglion of Drosophila. Neuron6, 83-99.

Selleck, S. B., Gonzales, C., Glover D. M. and White, K.(1992). Regulationof the G1-S transition in postembryonic neuronal precursors by axingrowth. Nature355, 253-255.

Singer, M. A., Penton, A., Twombly, V., Hoffmann, F. M. and Gelbart, W.M. (1997). Signaling through both type I DPP receptors is required anterior-posterior patterning of the entire Drosophila wing. Development124, 79-89.

Sisson, J. C., Ho, K. S., Suyama, K. and Scott, M. P.(1997). Costal2, a novelkinesin-related protein in the Hedgehog signaling pathway. Cell 90, 235-245.

Slusarski, D. C., Motzny, C. K. and Holmgren, R.(1995). Mutations thatalter the timing and pattern of cubitus interruptusgene expression inDrosophila melanogaster. Genetics139, 229-240.

Smith, M. A., Magill-Solc, C., Rupp, F., Yao, Y.-M. M., Schilling, J. W.,Snow, I. and McMahan, U. J.(1992). Isolation and characterization of anagrin homolog in the marine ray. Mol. Cell. Neurosci. 3, 406-417.

Stone, D. M., Hynes, M., Armanini, M., Swanson, T.A., Gu, Q., Johnson,R. L., Scott, M. P., Pennica, D., Goddard, A., Phillips, H., Noll, M.,Hooper, J. E., de Sauvage, F. and Rosenthal, A.(1996). The tumour-suppressor gene patchedencodes a candidate receptor for Sonic hedgehoNature384, 129-134.

Strutt, D. I., Wiersdorff, V. and Mlodzik, M. (1995). Regulation of furrowprogression in the Drosophilaeye by cAMP-dependent protein kinase ANature373, 705-709.

e

n

yal2.

on

for

g.

.

Strutt, D. I. and Mlodzik, M. (1997). Hedgehog is an indirect regulator ofmorphogenetic furrow progression in the Drosophilaeye disc. Development124, 3233-3240.

Sun, Y. H., Tsai, C. J., Green, M. M., Chao, J. L., Yu, C. T., Jaw, T. J.,Yeh, J. Y. and Bolshakov, V. N.(1995). Whiteas a reporter gene to detecttranscriptional silencers specifying position-specific gene expressioduring Drosophila melanogastereye development. Genetics141, 1075-1086.

Tabata, T., Eaton, S. and Kornberg, T. B.(1992). The Drosophila hedgehoggene is expressed specifically in posterior compartment cells and is a tarof engrailedregulation. Genes Dev. 6, 2635-2645.

Tabata, T. and Kornberg, T. B.(1994). Hedgehog is a signalling protein witha key role in patterning Drosophilaimaginal discs. Cell76, 89-102.

Tashiro, S., Michiue, T., Higashijima, S., Zenno, S., Ishimaru, S.,Takahashi, F., Orihara, M., Kojima, T. and Saigo, K.(1993). Structureand expression of hedgehog, a Drosophilasegment-polarity gene requiredfor cell-cell communication. Gene124, 183-189.

Trujillo-Cenoz, O. and Melamed, J.(1973). The development of the retina-lamina complex in muscoid flies. J. Ultrastruct. Res. 42, 554-581.

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.

Tsim, K. W., Ruegg, M. A., Escher, G., Kroger, S. and McMahan, U. J.(1992). cDNA that encodes active agrin. Neuron8, 677-689.

van den Heuvel, M. and Ingham, P. W.(1996). smoothenedencodes areceptor-like serpentine protein required for hedgehogsignalling. Nature382, 547-551.

Vincent, J. P., Girdham, C. H. and O’Farrell, P. H. (1994). A cell-autonomous, ubiquitous marker for the analysis of Drosophila geneticmosaics. Dev. Biol. 164, 328-331.

Von Ohlen, T. and Hooper, J. E.(1997). Hedgehog signaling regulatestranscription through Gli/Ci binding sites in the winglessenhancer. Mech.Dev. 68, 149-156.

Von Ohlen, T., Lessing, D., Nusse, R. and Hooper, J. E.(1997). Hedgehogsignaling regulates transcription through Cubitus interruptus, a sequencspecific DNA binding protein. Proc. Natl. Acad. Sci. USA94, 2404-2409.

Vortkamp, A., Lee, K., Lanske, B., Segre, G. V., Kronenberg, H. M. andTabin, C. J. (1996). Regulation of rate of cartilage differentiation by Indianhedgehog and PTH-related protein. Science273, 613-622.

Winberg, M., Perez, S. and Steller, H.(1992). Generation and earlydifferentiation of glial cells in the first optic ganglion of Drosophilamelanogaster. Development115, 903-911.

Xiong, W. C., Okano, H., Patel, N. H., Blendy, J. A. and Montell, C. (1994).repo encodes a glial-specific homeo domain protein required in thDrosophilanervous system. Genes Dev. 8, 981-994.

Xu, T. and Rubin, G. M. (1993). Analysis of genetic mosaics in developingand adult Drosophilatissues. Development117, 1223-1237.

Yamagata, M., Herman, J. P. and Sanes, J. R.(1995). Lamina-specificexpression of adhesion molecules in developing chick optic tectum. J.Neurosci. 15, 4556-4571.

Zecca, M., Basler, K. and Struhl, G.(1995). Sequential organizing activitiesof engrailed, hedgehogand decapentaplegicin the Drosophila wing.Development121, 2265-2278.

signals transmitted along retinal axons in drosophila ... · lamina target ﬁeld at the anterior...

Documents