Cell patterning in the Drosophila segment: spatial ... - Development

Development 110, 291-301 (1990) Printed in Great Britain © The Company of Biologists Limited 1990

291

Cell patterning in the Drosophila segment: spatial regulation of the segment polarity gene patched

ALICIA HIDALGO and PHILIP INGHAM Molecular Embryology Laboratory, ICRF Developmental Biology Unit, Department of Zoology, South Parks Road, Oxford, OX1 3 PS, UK

Summary

Intrasegmental patterning in the Drosophila embryo requires the activity of the segment polarity genes. The acquisition of positional information by cells during embryogenesis is reflected in the dynamic patterns of expression of several of these genes. In the case of patched, early ubiquitous expression is followed by its repression in the anterior portion of each parasegment; subsequently each broad band of expression splits into two narrow stripes. In this study we analyse the contribution of other segment polarity gene functions to the

evolution of this pattern; we find that the first step in patched regulation is under the control of engrailed whereas the second requires the activity of both cubitus interruptus0 and patched itself. Furthermore, the products of engrailed, wingless and hedgehog are essential for maintaining the normal pattern of expression of patched.

Introduction

even-skipped (eve) act as positive regulators of engrailed (en) and negative regulators of wingless (wg). As a result of this regulation, en expression is initiated in a series of single-cell-wide stripes at the cellular blastoderm stage (stage 5), which mark the anterior boundaries of the primordia of each parasegment (Lawrence et al. 1987) and persists in this location throughout embryogenesis. Similarly, wg expression appears in fourteen stripes at the same stage (Baker, 1987), a pattern maintained in the ventral epidermis, to occupy a narrow domain immediately anterior to that of en, within each parasegment. Several lines of evidence suggest that local cell interactions play an important role in the positional specification process after the activity of the pair-rule genes has ceased. Observations of the altered patterns of wg and en expression in embryos mutant for these and other segment polarity genes provided the first indications that such interactions occur (Martinez-Arias et al. 1988). Although en and wg are expressed in adjacent groups of ectodermal cells, the continued expression of either gene is dependent upon the activity of the other, implying that they mediate inductive interactions between neighbouring cells. This interpretation is supported by the finding that wg shows significant sequence similarity to the mouse proto-oncogene int-1 (Rijsewijk et al. 1987), which encodes a secreted glycoprotein found associated with the extracellular matrix of cultured cells (Bradley and Brown, 1990; Papkoff and Schryver, 1990). Moreover, studies of the distribution of the wg protein in Drosophila embryos show that it is closely associated with the membrane of

The epidermis of the Drosophila segment presents an attractive system for investigating the mechanism of cellular positional specification. Mutational analysis has identified a class of genes, the segment polarity class, that play a major role in this process. Lesions in these loci produce localised segmentally repeated defects (Niisslein-Volhard and Wieschaus, 1980; Niisslein-Volhard et al. 1984; Jurgens et al. 1984; Wieschaus et al. 1984) suggesting that they are specifically required for the generation of particular elements of the segmental pattern. Moreover, molecular analysis has revealed that the products of some segment polarity genes are similarly expressed in spatially restricted patterns (Fjose et al. 1985; Kornberg et al. 1985; DiNardo et al. 1985; Baker, 1987; Baumgartner et al. 1987; Cot6 et al. 1987; Nakano et al. 1989; Hooper and Scott, 1989; van den Heuvel et al. 1989). Elucidating the mechanisms that control these spatial patterns of expression is therefore an important step in understanding the cellular basis of patterning. Segmentation of the Drosophila embryo is accomplished by a cascade of gene regulation, starting with maternal products deposited in the developing oocyte and culminating in the activation of the pair-rule genes, which subdivide the embryo into fourteen periodic domains of gene activity (reviewed in Akam, 1987; Ingham, 1988). At least two of the segment polarity genes have been shown to respond directly to positional cues generated by the pair-rule genes (Howard and Ingham, 1986; DiNardo and O'Farrell, 1987; Ingham etal. 1988). For instance, fushi-tarazu (ftz) and

Key words: Drosophila, segment polarity genes, cell interactions, patched.

292

A. Hidalgo and P. Ingham

cells transcribing the wg gene and is apparently endocytosed by neighbouring cells (van den Heuvel et al. 1989). The response of cells to the wg protein is thought to depend crucially upon the activity of the patched (ptc) gene, which encodes a putative transmembrane protein of unknown biochemical function (Nakano et al. 1989; Hooper and Scott, 1989). In this study, we analyse the regulation of ptc transcription during embryogenesis; we present evidence that supports the proposed role of ptc in modulating cell competence with respect to the wg signal and provides further indications of cell interactions during the elaboration of pattern in the body segment. Materials and methods

Drosophila stocks Embryos were collected from flies heterozygous for the appropriate mutations, generated by crossing the balanced stocks to the wild-type laboratory stock Oregon R; this procedure eliminates balancer chromosomes from the parental flies (and, therefore, balancer homozygotes from the progeny) and ensures maximum fecundity. The following stocks were used: Df(2R)enB, Df(2R)SBl,gsb- Kr~, Df(2R)IIX62,gsb-zip-, ptc""08 wg™4, ptJIB928, ptc"022, ptc"099, ptc1^08, ptc1™, ptc"R87, ptP1C34, ptc"xm, ptc"wm, U112 B28 6 59 51 4 ptc" , ptc? , ptc ^, ptc™ , pfc* , wg™ , all balanced over CyO, ciD/Df(4)G, hhIJ35/TM3, and nkd ™16/TM3. in situ hybridisation Embryos were dechorionated in sodium hyperchlorite and fixed for 20 min as previously described (Ingham et al. 1985). Transcripts were detected by in situ hybridisation on whole embryos with digoxigenin (Boehrmger)-labelled DNA probes, following a modification of the protocol of Tautz and Pfeifle (1989) (C. Oh and B. Edgar, personal communication). Probes were prepared from purified plasmid inserts obtained from the following digests: ptc; 1.0+1.2 kb BamHl£coRI fragments from pc2 (Nakano et al. 1989): wg, 1.3+0.7 kb EcoRI-Hircdlll fragments from D-intl (Rijsewijk et al. 1987); en; 2.4kb Hindlll-EcoKl fragment from pE2C (Poole et al. 1985). Hybridisation was in 50 jdfinalvolume at a probe concentration of 0.2ng^l -1 . Samples were incubated for l h at room temperature in anti-digoxigenin antibody (1:2000) which had been preabsorbed against 50/i volume of fixed embryos for 1 h. Stained embryos were mounted in JB4 (Polysciences) A+C:B 25:1, after dehydrating in 95% EtOH for 1-2 min and infiltrating in JB4 A+C for 5-10 min. Results

Evolution of the wild type pattern of ptc Expression of ptc is first clearly detectable at the end of the blastoderm stage (stage 5, of Campos-Ortega and Hartenstein, 1985), occurring throughout most of the embryo with the exception of the pole cells and an anterodorsal region (Nakano et al. 1989; Hooper and Scott, 1989). Although expression at this stage is essentially ubiquitous, a weak modulation can be observed in some preparations, notably with a pair-rule periodicity (data not shown). After gastrulation (stage 8), ptc transcript becomes restricted to segmentally

repeated domains in the presumptive epidermis. These occupy about three fourths of the width of each parasegment and are reciprocal to the narrower domains of en (Fig. 1A). By stage 11 (Fig. IE), transcript accumulation is further restricted to two narrow stripes of cells in each segment, which arise in a stereotyped manner: at late stage 10, ptc transcripts show a heterogeneous distribution within each parasegment, the highest level of expression being in the most posterior cells (Fig. 1B-C). Subsequently, transcript disappears first from the cells located most dorsally in the middle of each parasegment (Fig. ID and 3A) and then from the more ventrally located cells. Concomitantly, first a distinct anterior stripe and then a posterior stripe start to appear; the anterior stripe reaches a higher level of transcription than the posterior one, first dorsally (Fig. IE) and then continuously along the dorsoventral axis. The posterior stripe coincides with the domain of expression of wg whereas the anterior one is immediately posterior to the en domain (data not shown). ptc is negatively regulated by en in the early extended germ band

Because the initial segmental pattern of ptc appears reciprocal to that of en, we monitored the effects of en mutations on ptc expression. As shown in Fig. 2B, ptc expression remains ubiquitous in the ectoderm of a stage 10 en~ embryo, suggesting that the normal pattern of expression is generated by the repression of ptc transcription by en. Although the ectodermal stripes do not form, some modulation of expression is still apparent (Fig. 4E). This is due to the periodic expression in the underlying mesoderm which is out of phase with the normal ectodermal stripes (Nakano etal. 1989) and therefore presumably independent of en function. The expression domain of ptc is reduced in naked embryos In a stage 10 naked (nkd) embryo, the bands of ptc transcript are significantly narrower than in wild type, covering about half rather than three quarters of each parasegment (the band corresponding to parasegment 2 being of weaker intensity than the rest) (Fig. 2C and 4B). The parasegmental boundaries are shifted in nkd embryos, usually forming posterior rather than anterior to each en domain; the orientation of the ptc bands is correspondingly opposite to that in wild type. As embryogenesis proceeds, the abnormally narrow ptc expression domains persist, and during stage 11, somewhat later than normal, each band splits into two narrow stripes, which are closer together than in wild type (data not shown). Fusions of stripes are frequently observed at the shortened germ band stage. Maintenance of the late phase of ptc expression requires wingless, hedgehog and engrailed Alterations in the expression of ptc in wg embryos are first apparent around stage 11, when the striped pattern begins to appear more diffuse than in wild type

Cell patterning in the Drosophila segment

293

Fig. 1. Detail of the evolution of the normal transcriptional pattern of ptc\ each panel shows three adjacent parasegments from embryos between stages 9 and 11, (staging according to Campos-Ortega and Hartenstein (1985)). Orientation in this and all subsequent figures is anterior to the left, dorsal surface uppermost, unless stated otherwise. (A) Stage 9, ptc transcripts are essentially uniformly distributed anteroposteriorly in the non-enexpressing cells of each parasegment. (B) Stage 10 to early 11; expression is highest in the most ventral posterior cells of each ptc domain (arrowhead indicates the ventral midline). (C) Ventral and (D) lateral aspects of stage 11 embryos. High levels of expression are seen in a row of ventral posterior cells (C) (arrowhead indicates the ventral midline); simultaneously, high level expression is seen dorsally at the anterior margin of each parasegment (D) and transcripts disappear gradually from the central region. (E) By late stage 11, ptc bands have completely split, but still show strongest signal anterodorsally (arrow).

I en '"#

A,,

A

nkd

Fig. 2. Expression of ptc in wild-type (A), en (B) and nkd (C) embryos. (A and B) Stage 10, (C) early stage 11, just before the splitting of the ptc domains takes place. Note the continuous ectodermal expression in B (dots indicate where ptc is normally repressed) and the narrower stripes in each parasegment in C. The parasegment boundary shifts in nkd embryos (C) (arrow).

294

A. Hidalgo and P. Ingham

A

B

Fig. 3. Gene activities required for the maintenance of the late phase of ptc expression. The panels on the left show stage 11 embryos, those on the right stage 12. (A,D) ptc expression in wild-type embryos. (B,E) In wg embryos, the narrow stripes form but rapidly decay except in the maxilliary, labial and prothoracic segments (arrows); this reflects the expression of en which also persists in the gnathal segments in embryos of the same genotype (Martinez-Arias et al. 1988; DiNardo et al. 1988) (C) In hh, ptc expression is confined most ventraUy and posteriorly at stage 11 and has completely disappeared by stage 12 (F). (G) Most expression similarly disappears from en homozygotes by stage 12 except in the head, anal pads and mesoderm (arrows).

(Fig. 3B). In late stage 11 the only periodicity is manifested in ellipses of cells located most ventraUy (Fig. 3B and 4D) and circles of cells in the approximate location of the tracheal placodes (not shown), while levels of transcript decrease in the remaining regions. After stage 11, expression begins to disappear, and by the end of germ band retraction is absent from most cells, with the exception of irregularly repeated lateral circles of cells and a dorsoventral stripe in the first thoracic segment and two patches in the maxillary and labial segments (Fig. 3E). Mutations of hedgehog (hh) have similar effects on

the late expression of ptc. Expression is normal up to the early extended germ band stage, but subsequently ptc transcripts start to disappear, being detectable only most ventrally and posteriorly, within each parasegment at stage 11 (Fig. 3C and 4C). By stage 12, there is no evibence of ptc expression anywhere in the embryo (Fig. 3F). In en mutant embryos, transcripts also begin to disappear from the ectoderm as the germ band starts to retract, (though persisting in the mesoderm for some time); by stage 14 ptc expression is confined to a small region in the head around the stomodaeum and pos-

Cell patterning in the Drosophila segment

295

Fig. 4. Detailed view of the ventral region of late stage 10/early stage 11 embryos showing expression of ptc in three adjacent parasegments. ptc stripes are narrower in nkd embryos (B) than in wild type (A), and are localised most ventrally and posteriorly in hh embryos (C). At the same stage, expression of ptc in wg embryos is limited to ventral and lateral circles (D) and is ubiquitous in en embryos (E). Note the underlying periodic modulation (arrow) in the mesoderm of en mutants. The ventral midline is indicated by arrowheads.

*l

V.

teriorly in structures probably corresponding to the anal pads (Fig. 3G). We also analysed the expression of ptc in embryos

mutant for gooseberry (gsb) but found no significant differences from the wild-type pattern (data not shown). Establishment of the late pattern of ptc transcription depends upon selective repression by cubitus interruptus Dominant and ptc In embryos mutant for the cubitus interruptus Dominant (ciD) allele (which is equivalent to a null mutation of the locus (Orenic et al. 1987)), alterations in ptc expression are first evident at stage 11; the broad bands of transcript persist, though with highest intensity in their posterior half ventrally, and in the anterior cells, dorsally (Fig. 5C). Simultaneously, expression disappears completely from the lateral region of each parasegment (Fig. 6C), creating a discontinuous pattern of expression similar to that typical of wg in normal embryos at the same stage (compare Fig. 5F and K). Slightly later, a narrow stripe of non-expressing cells appears ventrally, splitting each domain into a narrow anterior stripe and a broad posterior stripe (Fig. 5F and I)All of the EMS-induced alleles of ptc (with one exception) express ptc transcript at elevated levels and in a similarly aberrant pattern, the transition to the two narrow stripes of expression in each parasegment failing to occur normally at stagell (Fig. 5B,E,H and 6B). At stage 10, prior to the splitting, an ectopic ectodermal groove appears within each band of ptc expression. Subsequently, these initial domains do split, but into two stripes of unequal width, both of which are broader than in wild type, the posterior member of each pair being broader than the anterior one. The gap between the stripes is quite narrow and they frequently fuse at several points in stage 13-14 embryos. , the alterations In the one exceptional case, ptc1 in expression are initially similar, two broad bands forming instead of the normal narrow stripes typical of stage 11-12. The levels of transcript, however, are not as high as in the other ptc mutations and, after stage 11, the bands narrow somewhat resembling more the wildtype pattern; these, however, seem to fuse across segments and look sloppier than wild type. In some ptc mutant embryos we observed a differential intensity of the ptc stripes along the dorsoventral axis: highest signal intensities are localised to the most ventral area of both stripes within each parasegment, while the most anterior of these stripes, that is, the one that does not overlap the wg domain, is more intense dorsally.

296

A. Hidalgo and P. Ingham

A

D

ptc-wg Fig. 5. (A-C) ptc expression in late stage 11 embryos. (A) Wild type; (B) ptc 1N108; (C) ciD. Note the broadened stripes in B and C compared to A. In the ptc mutant splitting does not occur but a narrow stripe of repression appears presumably where en is ectopically induced (arrow). Note the lateral gap in each broad band in C. (D-F) ptc expression in stage 12 embryos. (D) Wild type; (E) ptc IN108; (F) cf. Note the narrow stripes of repression in cf (F) (arrow), which in contrast to those in ptc mutants, are not associated with ectopic en. (G-I) ptc expression in stage 14. (G) Wild type; (H) ptc W1G8; (I) cf. The broad domains of ptc in ptc and ciD persist during embryogenesis. (J) Expression of ptc in wg0*4 ptc"^108 double mutants, stage 14. Note the high levels, and ubiquitous, expression oiptc. (K) Expression of wg in wild-type stage 12 embryo. The lateral gap in each band of wg expression resembles that in the bands oiptc expression in ci (F,I) (L) wg in stage 12 ciD mutant.

To determine whether the derepression of ptc in ptc mutants is dependent upon wg activity, we analysed ptc expression in wg-ptc double mutant embryos. In such embryos, transcript is distributed uniformly and at higher levels than in wg or wild-type embryos. After stage 12, when ptc starts to disappear in wg embryos, expression of ptc is maintained in wg-ptc, persisting through all the late stages to the fully retracted germ band (Fig. 5J). Expression of wg in ciD, gsb and hh mutant embryos Because ptc and wg are coexpressed in some cells in each parasegment, we analysed the expression of wg in the same mutant embryos to investigate the relationship between the regulation of the two genes. In ciD and gsb embryos wg expression is initiated normally, but by stage 11 diverges significantly from the wild-type pattern. In ciD mutants, the stripes of expression appear slightly broader and less regular than in wild type. From stage 12 onwards, expression becomes sporadic, and is frequently absent from some segments (Fig. 5L). In stage 11 gsb mutant embryos, transcripts disappear from the ventral region, whilst persisting dorsally (Fig. 7D,G). At this stage in wild-type em-

bryos, wg expression normally splits dorsoventrally into two distinct domains (Baker, 1988; van den Heuvel et al. 1989) whilst gsb evolves from expanding dorsoventrally to occupying only the most ventral cells (Baumgartneref a/. 1987). In embryos homozygous for a strong hh allele, wg expression also begins normally, but starts to decay at stage 10 and by stage 11 is completely absent from the ectoderm of the trunk region of the embryo, persisting only in the terminal domains (Fig. 7B). Discussion

The emergence of pattern in the Drosophila body segments requires the activity of the segment polarity genes and is reflected in their dynamic patterns of transcription during embryogenesis. Although cell patterning is a continuous process, it is helpful in thinking about the problem to distinguish three distinct phases as revealed by segment polarity gene expression: an initiation phase, when transcription begins, just prior to gastrulation; a consolidation phase during the extension of the germ band, when the initial anteroposterior asymmetries are reinforced uniformly around the dor-

Cell patterning in the Drosophila segment

Fig. 6. Expression of ptc in the lateral region of three adjacent parasegments of wild-type (A), ptc (B) and ciD (C and D) late stage 11 embryos. Note the narrow stripes of repression and the ectopic segment boundaries (arrow) associated with the ectopic expression of en in the ptc mutant (B). A lateral band (arrows) lacking ptc transcripts appears in ciD embryos (C). Note the broad posterior stripe in D and the domain lacking ptc transcripts (arrow), which is not associated with ectopic en expression.

soventral axis; and an elaboration phase, beginning towards the end of the proliferative phase of embryogenesis, when more complicated patterns arise along both the anteroposterior and dorsoventral axes. Whereas control of the first phase depends upon the pair-rule genes, and, through them, the well-characterised segmentation gene hierarchy, the subsequent expression patterns are regulated by interactions mediated by the segment polarity genes themselves. In

297

contrast to the simple transcriptional cascade that characterises the segmentation gene hierarchy, these interactions involve the complex processes of intercellular signalling and signal transduction. The emergence of spatial differences within each developing parasegment is particularly well illustrated by the dynamic pattern of ptc expression (Fig. 1) and accordingly it is the transcriptional regulation of this gene upon which we have focussed our attention in this study. The principle conclusions of our analysis are depicted schematically in Fig. 8 and can be summarised as follows. During the early phase of germ band extension, ptc is down regulated by en in all en expressing cells, an event which is crucial for the maintenance of en expression by their wg neighbours. Towards the end of the extended germ band stage there is a switch in the requirement for en, the maintenance of ptc expression in cells flanking en-expressing cells depending upon en activity. This later non-autonomous role of en implies a signalling mechanism between the two cell types, a mechanism which we suggest is mediated by the hh gene. At the same time, cells in the middle of each parasegment cease transcribing ptc, a process which requires the function of ptc itself, implying a response of cells to some other signal of unknown identity. In what follows, we discuss the data supporting each of these conclusions in some detail, and consider their implications for the cell patterning process. In contrast to en, wg and gsb, each of which is activated in response to pair-rule genes in precisely defined domains in the blastoderm, transcription of ptc at this stage is almost ubiquitous, showing only a transient modulation. This modulation may reflect a fortuitous interaction between some pair-rule gene products and the ptc promoter, but does not appear to have any consequence for the later function or expression of the gene. The first significant spatial regulation of ptc occurs after the onset of gastrulation, with the repression of transcription in the anterior-most quarter of each presumptive parasegment (Nakano et al. 1989; Hooper and Scott, 1989). This domain corresponds to those cells that express en, and two lines of evidence suggest that en itself is responsible for this repression. First, in en" embryos, the repression does not occur and ptc continues to be expressed ubiquitously throughout most of the ectoderm. Second, in nkd~ embryos, the expansion of the en domain of expression described by Martinez-Arias et al. (1988) is accompanied by a corresponding decrease in the size of the ptc expression domain. Although both effects could be independent consequences of the absence of nkd function, the simplest interpretation seems to be that the ectopic expression of en results in the ectopic repression of ptc. The functional significance of this repression is revealed by the changes in en expression in ptc mutant embryos. At around stage 11, an additional domain of en is initiated anterior to the wg domain in each parasegment, apparently in direct response to the expression of wg (DiNardo et al. 1988; Martinez-Arias et al. 1988). Since such induction is blocked in wild-type

298

A. Hidalgo and P. Ingham

I §+ # *-*%* i

% *

» .-* hh Fig. 7. Expression of wg in wild-type and mutant embryos. (A). Wild type; (B) hh, both late stage 10; (C) wild type; (D) gsbSB1/gsb"X62; (E) nkd all stage 11. (F) wild type; (G) gsb58*/gsb"X6i, both stage 14. Note the absence of expression uin B as indicated by the dots. In D and G, expression is absent only ventrally from the regions indicated by dots. In nkd (E), there is ectopic expression (dots); the signal covers most of the parasegment dorsally.

embryos, it follows that the expression of ptc influences the response of cells to wg. If reception of the wg signal is necessary for maintaining the expression of en, it is clearly essential that the en cells repress ptc transcription if they are to continue to express en. Thus, the negative regulation of ptc by en can be regarded as part of a feedback loop which ensures the continued expression of en only in the cells posterior to the wg domain, and hence maintains the polarity of the parasegment (see Fig. 9). Whereas simple transcriptional autoregulation of en would allow the en cell population to maintain itself independently of size and position, this somewhat more complex mechanism imposes specific constraints on the en cells, ensuring that their initial position and proportion is maintained during the

major cell rearrangements that occur during germ band extension (see also discussion in Mohler, 1988). A reciprocal dependence of the wg-expressing cells on their en neighbours has also been described by Martinez-Arias et al. (1988). Given the predominantly nuclear localisation of the homeodomain protein encoded by en (DiNardo et al. 1985), it seems unlikely that this interaction is directly mediated by en itself. Instead, en probably controls the expression of some other signalling molecule that acts in a manner analogous to wg. A good candidate for such a molecule is the product of the hedgehog gene. Mutations of hh have been shown to act non-autonomously in genetic mosaics (Mohler, 1988) and our results show that the expression

i

r

hh

Fig. 8. Schematic representation of the interactions influencing the evolution of the ptc expression pattern during embryogenesis. Each circle represents one quarter the width of a parasegment (indicated by square brackets): at the blastoderm stage (A), this is equivalent to a single cell, but at subsequent stages (B and C) a circle represents several cell diameters. W and E denote domains of wg and en expression, respectively; note however, that the wg domain does not increase in size proportionately so that the designation of an entire circle as a wg domain in B and C is a simplification. (A) At blastoderm, all cells express ptc (shading) whereas expression of wg and en is established in single cell wide domains. (B) by stage 10, expression of ptc is restricted to three quarters of each parasegment, most likely through direct repression by en. At this stage, mutual interactions between the W and E cells, mediated by wg and hh maintain the expression of en and wg respectively. (C) By late stage 11, the ptc expression domains are further restricted to two narrow stripes in each parasegment. Establishment of these stripes depends upon ptc expression itself; this could imply either that neighbouring cells require ptc to send a signal to the cells which down regulate expression; alternatively, the down-regulating cells may themselves require ptc protein to respond to such a signal. The identity of such a hypothetical signalling molecule is unknown (?). The repression depends at least in part upon the activity of ciD. Maintenance of the ptc stripes depends on their neighbouring E cells and is possibly mediated via hh as shown.

Cell patterning in the Drosophila segment

Fig. 9. Schematic representaion of the interactions between en, wg and ptc. Receival of the wg signal by a putative receptor R (1) leads to the induction of en. en, nuclear protein, represses in those cells the expression of ptc (2), which would otherwise block the inductive effect of wg on en (3). In this way, the domain of en is restricted to only those cells that had originally expressed en, thus maintaining the initial polarity of the parasegment.

of wg at the posterior margin of each parasegment disappears in hh embryos by stage 11. Taken with the observation that hh is probably expressed in the enexpressing cells of each parasegment (J. Mohler, personal communication) these findings strongly suggest that hh is involved in a signalling process between the en and wg cells. Whatever the identity of the signal, such a process raises a similar question to that posed by the putative wg signalling function, namely what determines the polarised response of cells to the signal? In this context, the effects of nkd mutations may be significant. Whereas in ptc mutant embryos, expansion of the wg domain is accompanied by the ectopic induction of en expression, in nkd embryos, there is an expansion of the en domain accompanied by the ectopic induction of wg expression in adjacent cells (MartinezArias et al. 1988). These reciprocal effects of the two mutations imply that the wild-type genes may perform complementary functions, expression of nkd blocking or modifying the response of cells to hh in a manner analogous to that proposed above for the relationship between wg andpfc (see also DiNardo and Heemskerk, 1990). We can thus identify at least five genes involved in consolidating the polarity of each parasegment established at the blastoderm by the pair-rule dependent activation of wg and en, with ptc, nkd and hh being involved in signalling processes that maintain wg and en expression in the appropriate spatial domains. These interactions apparently occur between neighbouring cells and are responsible for the maintenance of the boundaries of parasegments. We now consider the factors involved in the subsequent elaboration of pattern within each parasegment. At the end of stage 10, shortly after the interactions between wg- and en-expressing cells take place, the pattern of ptc transcription undergoes a further refinement; cells immediately anterior to the wg domain, ' which we postulate require ptc activity to prevent them

299

inducing en expression in response to wg, now cease transcribing ptc. Expression of the gene thus becomes restricted to two narrow stripes in each parasegment, one marking the anterior border of the future segment, the other coinciding with the wg domain (see Fig. 9). This pattern of expression is never established in embryos mutant for en, hh or wg, implying that the maintenance of ptc transcription after stage 10 requires the expression of all three genes. Such a positive requirement for en at this stage contrasts with its postulated negative regulatory role prior to stage 11. Moreover, whilst the repressing activity of en is a cell autonomous function, the subsequent maintenance of ptc expression is limited to two narrow stripes of cells which flank the en expression domain of each parasegment. This apparently non-autonomous effect of en on ptc expression therefore implies an indirect interaction between the two genes, en presumably regulating the expression of some signalling molecule which in turn stimulates ptc transcription. Although the evidence that wg encodes a signalling molecule is compelling, it seems unlikely that it plays such a direct role in the maintenance of ptc expression. While ptc expression more or less ceases by stage 11 in wg mutant embryos, we find it to be normal in gsb~ embryos despite the fact that such embryos lack wg expression ventrally. This suggests that the cessation of ptc expression in wg mutants is due to some earlier effect of wg, presumably the maintenance of en expression in neighbouring cells described above. Once this function is complete, however, wg seems to have no further role in the maintenance of ptc. The possibility that hh encodes a signalling molecule has already been discussed above; our finding that ptc expression also disappears in hh mutants is thus consistent with a role for hh in signalling to cells neighbouring the en domain to maintain ptc expression. Unlike the postulated interaction between hh and wg, this response of cells to hh would not be polarised and should therefore be independent of nkd activity, consistent with our finding that the narrow stripes of ptc expression are indeed established and maintained in nkd embryos. These considerations might suggest that the evolution of the ptc expression pattern depends on its negative regulation by en followed by positive regulation in cells adjacent to the en domain. Our analysis indicates however, that ptc expression is also actively repressed in cells in the middle region of each of the initial broad domains; the elimination of this repression results in constitutive expression of the gene which is independent of its putative positive regulators. In all of the ptc mutations we analysed, ptc transcript is initially normally expressed, but after stage 10 persists at much higher levels than in wild type in a broad domain, interrupted only by a thin line of non-expressing cells which probably correspond to those in which en expression is ectopically induced. In ptc mutants which also lack wg expression, transcription is virtually ubiquitous, suggesting that this release from repression is independent of the positive regulatory mechanism outlined above.

300

A. Hidalgo and P. Ingham

This persistent expression of ptc in the cells anterior to the normal wg domain can be interpreted in a number of ways. One possibility is that expression of ptc in the two narrow stripes of cells at the edges of the initial broad expression domains is required for those cells to produce a further signal that represses ptc expression in the central cells of each domain. This model, while seemingly unecessarily complex, has the virtue of providing a function for the continued expression of ptc in cells whose differentiation appears unaffected in its absence (see Hooper and Scott, 1989). Moreover, the results of mosaic analysis of ptc mutants in imaginal disc cells provide a precedent for such nonautonomous effects of ptc, ptc~ cells affecting the differentiation of their genetically wild-type neighbours (Phillips et al. 1990). A simpler alternative, however, is to suppose that ptc expression is autonomously required for its own repression. According to this interpretation, presence of the ptc protein in the cell membrane would enable cells to respond to a signal, one aspect of this response being to repress further ptc transcription, an effect in some ways analogous to the down regulation of growth factor receptors elicited by ligand binding (e.g. Russell et al. 1987). Whether or not ptc itself functions as the receptor for such a signal or instead modulates its transduction, for instance, by coupling the receptor to a particular second messenger pathway or modulating the internal physiology of the cell, remains to be determined. Discrimination between these models requires the analysis of genetic mosaics which we are now undertaking; in either case, the identity of the postulated signal to which the p/c-repressing cells respond remains obscure. Clearly it could emanate from the wg-expressing cells; however, it is unlikely to be wg itself since ptc transcription, rather than persisting, disappears from wg~ embryos by stage 11. Embryos homozygous for the ciD mutation, by contrast, show an alteration in the distribution of ptc transcripts which is in some ways reminiscent of the effect of ptc mutations. Dorsally and ventrally, the broad domains of ptc expression persist, although ventrally, a narrow strip of cells does cease expression; laterally, however, the ptc transcript completely disappears. This latter effect is intriguing not leastbecause it mirrors the lateral repression of wg expression which normally occurs at the same stage in wild-type embryos. It is difficult to interpret these changes in terms of indirect effects of the ciD mutation on the expression of other segment polarity genes: en expression is essentially unaffected by the mutation (unpublished observations) whereas the expression domains of wg decay prematurely after broadening slightly. While our data do not allow us to determine whether ciD acts upstream or downstream of ptc to control its transcription, the finding that the ci gene encodes a zinc finger gene (Orenic etal. 1990) is consistent with a direct interaction between the genes. Why ciD mutations should have apparently opposite effects on ptc expression at different positions around the dorsoventral axis remains

unclear; however, these effects provide a further indication of the extent of the heterogeneity around this axis at the end of the extended germ band stage, most clearly exemplified by the wild-type patterns of expression of wg and gsb. Conclusion

Our results suggest that the evolution of the pattern of ptc transcription during embryogenesis depends upon three distinct control mechanisms. Whilst thefirstof these appears to be a cell-autonomous function, the latter two seem to involve intercellular communication. Although these conclusions require testing by mosaic analysis, our data provide strong support or the notion that local cell interactions play a major role in the specification of positional information in the developing embryo. We would like to thank all the members of our laboratory and especially Yoshiro Nakano for helpful discussions. A.H. gratefully acknowledges the award of a Research Bursary by the Imperial Cancer Research Fund. References AKAM, M. E. (1987). The molecular basis for metameric pattern in the Drosophila embryo. Development 101, 1-22. BAKER, N. E. (1987). Molecular cloning of wingless, a segment polarity gene in Drosophila'. The spatial distribution of a transcript in embryos. EMBO J. 6, 1765-1773. BAKER, N. E. (1988). Localization of transcripts from the wingless gene in whole Drosophila embryos. Development 103, 289-298. BAUMGARTNER, S., BOPP, D . , BURRI, M. AND NOLL, M. (1987).

Structure of two genes at the gooseberry locus related to the paired gene and their spatial expression during Drosophila embryogenesis. Genes and Dev. 1, 1247-1267. BRADLEY, R. S. AND BROWN, A. M. C. (1990). The proto-oncogene

int-1 encodes a secreted protein associated with the extracellular matrix. EMBO J. 9, 1569-1575. CAMPOS-ORTEGA, J. AND HARTENSTEIN, V. (1985). The Embryonic

Development of Drosophila melanogaster Berlin: Springer COTE, S., PREISS, A., HALLER, J., KJENLIN, A., SEIFERT, E. AND

JACKLE, H. (1987). The gooseberry-zipper region of Drosophila: five genes encode different spatially restricted transcripts in the embryo. EMBO J. 6, 2793-2801. DINARDO, S. AND HEEMSKERK, J. (1990). Molecular and cellular interactions responsible for intrasegmental patterning during Drosophila embryogenesis. Seminars in Cell Biology 3, in press. DINARDO, S., KUNER, J. M., THEIS, J. AND O'FARRELL, P. H.

(1985). Development of embryonic pattern in D. melanogaster as revealed by accumulation of the nuclear engrailed protein. Cell 43, 59-69. DINAJUX), S. AND O'FARRELL, P. H. (1987). Establishment and

refinement of segmental pattern in the Drosophila embryo: spatial control of engrailed expression by pair-rule genes. Genes and Dev. 1, 1212-1225. DINARDO, S., SHER, E., HEEMSKERK-JONGENS, J., KASSIS, J. A. AND

O'FARRELL, P. H. (1988). Two tiered regulation of spatially patterned engrailed gene expression during Drosophila embryogenesis. Nature 332, 604-609. FJOSE, A., MCGINNIS, W. J. AND GEHRJNG, W. J. (1985). Isolation

of a homeobox containing gene from the engrailed region of Drosophila and the spatial distribution of its transcripts. Nature 313, 284-289. HOOPER, J. AND SCOTT, M. P. (1989). The Drosophila patched gene encodes a putative membrane protein required for segmental patterning. Cell 59, 751-765.

Cell patterning in the Drosophila segment HOWARD, K. AND INGHAM, P. W. (1986). Regulatory interactions

between the segmentation genes fushi-tarazu, hairy, and engrailed in the Drosophila blastoderm. Cell 44, 949-957. INGHAM, P. W. (1988). The molecular genetics of embryonic pattern formation in Drosophila. Nature 335, 25-34. INGHAM, P. W., BAKER, N. E. AND MARTINEZ-ARIAS, A . (1988).

Regulation of segment polarity genes in the Drosophila blastoderm by fushi-tarazu and even-skipped. Nature 331, 73-75. INGHAM, P. W., HOWARD, K. R. AND ISH-HOROWICZ, D . (1985).

Transcription pattern of the Drosophila segmentation gene hairy. Nature 318, 439-445. JURGENS, G . , WlESCHAUS, E . , NOSSLEIN-VOLHARD AND KLUDING,

H. (1984). Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster. U Zygotic loci on the third chromosome. Wilhelm Roux' Arch, devl Biol. 193, 283-295. KORNBERG, T . , SlDEN, I . , O ' F A R R E L L , P . AND SlMON, M . (1985).

The engrailed locus of Drosophila: in situ localisation of transcripts reveal compartment-specific expression. Cell 40, 45-53. LAWRENCE, P. A., JOHNSTON, P., MACDONALD, P. AND STRUHL, G.

(1987). Borders of parasegments in Drosophila embryos aie delimited by thsfushi tarazu and even-skipped genes. Nature 328, 440-442. MAKTINEZ-ARIAS, A., BAKER, N. E. AND INGHAM, P. W. (1988).

Role of segment polarity genes in the definition and maintenance of cell states in the Drosophila embryo. Development 103, 157-170. MOHLER, J. (1988). Requirements for hedgehog, a segmental polarity gene, in patterning larval and adult cuticle of Drosophila. Genetics 120, 1061-1072. NAKANO, Y., GUERRERO, I., HIDALGO, A., TAYLOR, A. M., WHITTLE, J. R. S. AND INGHAM, P. W. (1989). The Drosophila

segment polarity gene patched encodes a protein with multiple potential membrane spanning domains. Nature 341, 508-513. NOSSLEIN-VOLHARD, C. AND WlESCHAUS, E. (1980). Mutations affecting segment number and polarity in Drosophila. Nature 287, 795-801.

301

melanogaster. I. Zygotic loci on the second chromosome. Wilhelm Roux. Arch, devl Biol. 193, 267-282. ORENIC, T., CHTDSEY, J. AND HOLMGREN, R. (1987). Cell and

cubitus interruptus dominant: two segment polarity genes on the fourth chromosome in Drosophila. Devl Biol. 124, 50-56. ORENIC, T. V., SLUSARSH, D. C., KROLL, K. L. AND HOLMGREN,

R. (1990). Cloning and characterisation of the segment polarity gene cubitus interruptus Dominant of Drosophila. Genes and Dev. (in press). PAPKOFF, J. AND SCHRYVER, B. (1990). Secreted int-1 protein is

associated with the cell surface. Molec. cell. Biol. 10, 2723-2730. PHILLIPS, R. G., ROBERTS, I. J. H., INGHAM, P. W. AND WHITTLE,

J. R. S. (1990). The Drosophila segment polarity gene patched is involved in a position-signalling mechanism in imaginal discs. Development, in press POOLE, S. J., KAUVAR, L. M., DREES, B. AND KORNBERG, T. (1985).

The engrailed locus of Drosophila: structural analysis of a transcript. Cell 40, 37-43. RlJSEWLJK, F . , SCHUERMANN, M . , W A G E N A A R , E . , P A R R E N , P . , WEIGEL, D. AND NUSSE, R. (1987). The Drosophila homolog of

the mouse mammary oncogene int-1 is identical to the segment polarity gene wingless. Cell 50, 649-657. RUSSELL, D . S., GHERZI, R., JOHNSONE, L., CHOU, C. AND ROSEN,

O. M. (1987). The protein-Tyrosine kinase activity of the insulin receptor is necessary for insulin-mediated receptor downregulation. / . biol. Chem. 262, 11833-11840. TAUTZ, D. AND PFELFLE, C. (1989). A non-radioactive in situ

hybridisation method for the localisation of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback. Chromosoma 98, 81-85. VAN DEN HEUVEL, M., NUSSE, R., JOHNSTON, P. AND LAWRENCE, P.

A. (1989). Distribution of the wingless gene product: a protein involved in cell-cell communication. Cell 59, 739-749. WlESCHAUS, E . , NOSSLEIN-VOLHARD, C . AND JuitGENS, G . (1984).

Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster: Zygotic loci on the X-chromosome and the fourth chromosome. Wilhelm Roux' Arch, devl Biol. 193, 296-307.

NOSSLEIN-VOLHARD, C , WlESCHAUS, E . AND KLUDING, H . (1984).

Mutations affecting the pattern of the larval cuticle in Drosophila

(Accepted 20 June 1990)