Roux's Archives of

velopmental Biology

Roux's Arch Dev Biol (1988) 197:507-512

9 Springer-Verlag 1988

Expression of a reporter gene resembles that of its neighbour: an insertion in the hairy gene of Drosophila Laurent Fasano, Nathalie Cork, and Stephen Kerridge LGBC, CNRS, Centre Universitaire de Marseille-Luminy, Case 907, F-13288 Marseitle Cedex 9, France Summary. Random insertions of a promotor fused to a reporter gene, such as Lac-Z, reveal regulatory sequences that confer temporal and spatial patterns of gene expression in eukaryotes. These patterns may reflect the activity of a neighbouring gene and thus lead to the isolation of new genes essential for normal development. Here, we demonstrate that this hypothesis is true for an insertion into the well characterized segmentation gene, hairy, in Drosophila. The insertion is homozygous lethal and fails to complement other hairy alleles, giving the phenotype described for hairy mutations. The insertion is located at 66D on the polytene chromosome map, is within 300 600 bp 5' to the first hairy exon, and is orientated in the same sense (5'-3') as the hairy transcription unit. Expression of /~-galactosidase (]?-gal), deriving from the insertion, follows closely the spatio-temporal patterns of expression of hairy gene product during embryogenesis. In addition, other sites of fl-galactosidase expression are shown in the third larval instar stage and in the adult ovary. The results show that some insertions, giving restricted patterns of reporter gene expression, will reflect the temporo-spatial activity of a nearby gene. Key words: Drosophila - Reporter gene - hairy - Segmentation

Introduction The random insertion of reporter genes in prokaryotes and single celled eukaryotes, has led to the isolation of new genes, the examination of gene regulation and the analysis of protein structure and function (for review see Silhavy and Beckwith 1985). More recently, this type of approach has been applied to higher eukaryotes (Teeri et al. 1986; O'Kane and Gehring 1987; Allen et al. 1988; Fasano and Kerridge 1988). An important question for the analysis of such reporter genes is whether their pattern of spatio-temporal expression resembles that of the adjacent gene. In Drosophila, a number of genes have been isolated which are of importance for the construction of the basic body plan (Lewis 1978; Wakimoto and Kaufman 1981; Nfisslein-Volhard and Wieschaus 1980). Many of these have been characterized at the molecular level (for reviews see Scott and O'Farrell 1986; Akam 1987) and their patterns of expression during

Offprint requests to: L. Fasano

development have been described. One way to ask whether the random integration of a reporter gene will reflect that of its neighbour would be to isolate insertions in these, or other well characterized, genes. Here we report the expression of a reporter gene (O'Kane and Gehring 1987), consisting of a weak promoter and the structural gene for/%galactosidase (fl-gal), inserted into the segmentation gene, hairy, of Drosophila. Both genetic (Niisslein-Volhard and Wieschaus 1980; Ingham et al. 1985a) and molecular (Ish-Horowicz et al. 1985; Ingham et al. 1985b; Carroll et al. 1988) aspects of the hairy gene have been described. The spatio-temporal expression of/% gal, with minor modifications, resembles that of the normal products of the hairy gene during embryogenesis. Also we describe the spatial expression of fl-gal during other stages of development.

Materials and methods

Fly stocks and cuticle preparations A P(Lac, ry +)A insertion (O'Kane and Gehring 1987) was found, whilst screening a collection of about 180 insertions (Fasano and Kerridge 1988), within the hairy gene. By classical genetic crosses, the insertion was localised to chromosome 3 and balanced with TM3, Sb ry gK (see Lindsley and Grell (1968) for genetic terminology). Cuticle preparations, from aged eggs derived from this stock, were carried out as described by Van der Meer (1977). Alleles of the hairy gene used w e r e Df(3L)h i22, h 8Kl15 and h 2 (Ingham et al. 1985a). Isolation of Genomic DNA Genomic DNA, from ry 5~ and P(Lac, ry+)A L43a/ry 5~ flies, was isolated as described by Junakovic (1980), digested separately by EcoRl, HindIII and Xba I, electrophoresed on an agarose gel, and blotted onto nitrocellulose. Probes labelled by random primer extension (Feiuberg and Vogelstein 1983) were hybridised to the blot.

Detection of fl-galactosidase Embryos were collected, dechorionated, permeabilized and fixed as described before (Hiromi et al. 1985; O'Kane and Gehring 1987). Embryos were collected overnight, dechorionated in bleach and fixed and permeabilised for 25 rain

H

R

R

[,,.k . . . . . .

B

S

~

Sn~

HR

R

S Se H

PROBES

d

H

t.....................................................................!........................................1,,.

K PX~

Not B

$

RV

(X'~)

HB

rSFS//S/S~F/F//F/#F//A

RO

1Kb

Fig. 1. Cuticle preparations of wildtype a and homozygous P(Lac,ry+)A L43a b mutant first instar larvae. Thoracic (1 3) and abdominal (A1 8) segments are indicated. The phenotype in b is typical of the pair-rule phenotype described for amorphic mutations at the hairy locus (Ingham et al. 1985a). The homozygous phenotype is variable; in 2 ~ 3 0 % of homozygotes fusion of abdominal belts A3~4 and A5-6 (arrow heads') occurs, e shows hybridisation of an EcoRl fragment from the hairy gene (RO in Panel d) to genomic D N A isolated from ry 5~ (lanes 1, 2 and 3) and P(Lac,ry+)A L43a/ry 5~ (lanes 4, 5 and 6) flies. The RO probe hybridizes to a 3.2 kb fragment in ry s~ D N A on EcoR I digestions (lane 3). On L43a/ry 5~ D N A (lane 6) the same fragment hybridizes (due to the ry 5~ chromosome) plus a band at 3.5 kb (arrow). The band at 3.2 kb consists of two hybridizing fragments; when the 500 bp 5' P element is used as a probe, this same fragment (3.1 kb) lights up on L43a D N A (not shown). On Xba I digestions two fragments plus a large hybridize (lanes 2 and 5) of 13 kb and 7.3 kb respectively. On L43a D N A (lane 5), these same fragments (due to ry5~ (arrow) fragment hybridize. The EcoRI fragment R + I (see Panel d) hybridizes only to the large 13 kb fragment in both populations insertion to the left of the Xba I site covered by the probe (see Panel d). On Hind III of flies (not shown). This places the P(Lac,ry§ digestions (lanes 1 and 4) the RO probe lights up a 10.7 kb and a 7.2 kb fragment on ry 5~ DNA. On L43a D N A the same fragments (due to ry 5~ plus fragments of 5.9 kb and 1.1 kb (lane 4, arrows) hybridize. Since the EcoRI fragment R + 1 hybridizes to the 5.9 kb and 7.2 kb fragments (data not shown), this places the insert about 1.1 kb to the right of the HindIII site, located in RO. Orientation insertion (see Panel d). The 3' EcoRI fragment of the insertion is possible because of the assymmetry of EcoRI sites in the P(Lac,ry§ transposon is 3 kb in size; if this were adjacent to the Hind III site, a fragment of 5.6 kb should light up in of the P(Lac,ry§ lane 6 (3 kb of insertion plus about 2.6 kb of D N A to the left of the insertion). The absence of a fragment of 5.6 kb shows that the orientation is as shown in Panal d. d Sites of restriction enzymes on P(Lac,ry+)A (O'Kane and Gehring 1987) and the hairy gene region (Rushlow and Ish-Horowicz personal communication), in addition to the approximate location of the insertion into the hairy gene. For ry s~ genomic DNA, we found a difference in the position of an Xba I site, compared to that given to us by Rushlow and Ish-Horowicz (indicated by brackets). Instead we find an Xba I site further to the right (Xba I*)

509 in a 1 : 1 mixture of heptane and 4% paraformaldehyde in 50 m M Pipes at p H 7.5. Eggs were then tested directly for /%galactosidase activity using 5-bromo-4-chloro-3-indolyl/~-galactopyranoside (X-gal) as a substrate (Hiromi et al. 1985) or using a monoclonal antibody directed against/% galactosidase (Promega biotec), following devitellinization (Mitcheson and Sedat 1983), using the techniques described by M a c D o n a l d and Struhl (1986). Both primary anti-/%gal and secondary alkaline phosphatase conjugated anti-mouse antibodies (Promega Biotec) were diluted 500 times in BBT buffer (see M a c D o n a l d and Struhl 1986) for the reaction with the embryos. Oocytes and larval tissues, from the balanced L43a insertion stock, were fixed and stained as described by Fasano and Kerridge (1988). Results and discussion

Amongst about 180individual P(Lac,ry+)A insertions (O'Kane and Gehring 1987; Fasano and Kerridge 1988), incorporated into the Drosophila genome via P element transformation (Rubin and Spradling 1982), one insertion (L43a) was found in the hairy (h) gene (Niisslein-Volhard and Wieschaus 1980; Ingham et al. 1985a). Homozygotes for the L43a insertion, and trans heterozygotes of L43a with Df(3L) h 122"or h TM 1s (lngham et al. 1985 a), are lethal during embryogenesis; all three combinations give a strong pair rule phenotype (Fig. 1 b), as described by Ingham et al. (1985a). L43a in trans with the homozygous viable allele, h 2, gives the extra chaetae phenotype described previously (Lindsley and Grell 1968; Ingham et al. 1985a). These phenotypes support the idea that homozygotes for the insertion possess little or no hairy gene product in the segmented part of the embryos. The cytological position of the P(Lac,ry+)A insertion is at 66D (not shown), the site of the h gene (Ingham et al. 1985 a). Finally, comparison of southern blot restriction patterns of L43a D N A versus ry 5~ parental D N A reveals that the L43a insertion is localized close to the 5' end of the gene (Figs. 1 c, d), in the same orientation as that of the h gene itself (Figs. 1 c, d). The pattern of ~-gal activity during embryogenesis, in line L43a, is shown in Fig. 2. Messages from the hairy gene are first detected at the beginning of interphase two cell cycles before cellularization of the blastoderm (Ingham et al. 1985 b). At this stage, transcripts are uniformly distrib-

Fig. 2a-e./~-galactosidase activity of P(Lac,ry +)A L43a insertion during embryogenesis, a fl-gai pattern in a blastoderm embryo. Seven stripes of /~-gal activity are detected (1-7). At this stage the head patch (see b), where hairy products are normally produced, is not mimicked by the enzyme activity, b Stage 5 6 embryo (Campos-Ortega and Hartenstein 1985);/~-gal activity persists and is more intense in the seven stripes. The stripes are wider than that of hairy products; note also the tendency for the first stripe to label cells posterior to the cephalic furrow (cf). Products of the hairy gene usually are restricted to cells anterior to the cephalic furrow. The head patch is clearly evident (O). Cells of the ventrally located mesoderm (ms) also stain, e Hindgut (hg), foregut (fg) and tracheal pits (arrow heads) possess fl-gal activity in stage 11 (Campos-Ortega and Hartenstein 1985) embryos. Weak residual enzyme activity is present in mandibular, labial, mesothoracic (between arrows), and abdominal segments 1, 3, 5, and 7. d Stage 16 (Campos-Ortega and Hartenstein 1985) embryo; fi-gal is expressed in cells of the fore- (fg) and hing-gut (hg), in addition to the anal

pads (ap). e Stage 11 embryo homozygous for the P(Lac,ry+)A insertion. Enzyme activity persists in the seven stripes seen at blastoderm in addition to the cells making up the mesoderm. Cells normally showing no hairy products such as the amnioserosa (as), posterior head (ph) exhibit/~-gal activity. For a b c and e monoclonal antibodies directed against fl-galactosidase were used; for d X-gal was used as the substrate (see Materials and methods). With the exception of e, embryos resulted from an outcross of the insertion stock to ry 5~ flies; the embryo in e came from an egg lay from the balanced stock

510 uted throughout the egg cortex (Ish-Horowicz et al. 1985; Ingham et al. 1985b). Using a monoclonal antibody against fl-galactosidase, enzyme is visible by the late cellular blastoderm stage (Fig. 2 a); the earlier uniform pattern is not detectable. The spatial pattern closely matches that described for the h messages at blastoderm; seven stripes of enzyme activity, from 20 to 75% egg length along the anteriorposterior axis of the egg, are detected. Recently, a weaker 8th stripe of h message, posterior to the 7th and presumably corresponding to the hindgut primordia, has been shown (Ingham personal communication). This expression has not been detected in our experiments, suggesting either that this is too weak to detect or the insertion has disrupted the regulatory elements required for normal initiation of this spatial expression. In addition, the transcripts (Ingham et al. 1985b) and protein (Carroll et al. 1988; Hooper and Ish-Horowicz personal communication) corresponding to the h gene are present in a group of cells in an anterior dorsal (or head) patch. This head patch is not reproduced by the L43a insertion (Fig. 2 a) in cellular blastoderm embryos, but activity in the corresponding cells can be seen in stage 6-7 embryos (Campos-Ortega and Hartenstein 1985) this site of fl-gal activity augments (Fig. 2b). Relevant and consistent to this observation is that the regulatory unit necessary for the head patch expression (Howard et al. 1988) is close ( - 2 . 5 kb to - 5 . 5 kb upstream of transcription start site), and upstream to, the site of integration of P(Lac,ry+)A L43a. One possible explanation for the delay in the expression of fl-gal in the head patch may be due to the altered distance of the head patch regulatory element with respect to the normal h and reporter gene promoters; for example, it is known that the efficiency of transcription of genes is dependent on the distance of regulatory elements and promotors with respect to the turn of the DNA helix (Takahashi et al. 1986). Alternatively, the difference between the hairy and P element promoters may explain the different efficiency of fl-gal expression with respect to normal hairy products. During gastrulation, the first stripe lies just anterior to the cephalic furrow (Fig. 2b), as described for h messages (Ish-Horowicz et al. 1985; Ingham et al. 1985b) and anti-h antibodies (Carroll et al. 1988) showing that fl-gal messages receive the correct spatial information normally given to the h messages. With respect to intensity, stripes 1, 2 and 7 seem to be more intense than the others. Following invagination of the ventral furrow, cells of the mesoderm show significant fl-gal activity (Fig. 2b). Similar observations have been observed for the normal h protein (Carroll et al. 1988; Hooper and Ish-Horowicz personal communication). The relative widths of the stripes of/~-galactosidase are larger than those reported for normal hairy gene products. Messages and proteins are localised in about 3-5 cells at the cellular blastoderm stage (Ish-Horowicz et al. 1985; Ingham et al. 1985b; Carroll et al. 1988), whereas up to 6 or 7 cells can be labelled in our line (Fig. 2a). Similarly, some staining of the yolk nuclei (Fig. 2a) of blastoderm stage embryos (a site where h products have not been reported) is sometimes observed. These differences may be due to the absence of sequences in the fl-galactosidase messages required for localisation to the apical cytoplasm, that may be present in normal h messages (Ish-Horowicz et al. 1985; Ingham et al. 1985b; Carroll et al. 1988). This difference may thus result in diffusion of h controlled fl-gal transcripts, before cellularization.

During later stages, the stripes of fl-gal persist when normal h products have disappeared (Fig. 2c). These are not the strongest sites of expression; by the end of germ band extension (Fig. 2c), strong levels of fl-galactosidase are found in the developing fore- and hind-gut primordia (or proctodaeal and stomodaeal invaginations). Furthermore, segmentally repeated patterns of enzyme activity are found around the primordia of the tracheal pits (Fig. 2c). Messages from the h gene (Ingham, personal communication), as well as the h protein (Carroll et al. 1988; Hooper and Ish-Horowicz personal communication), have been shown to be present in these groups of cells. Upon retraction of the germ band and during later embryonic stages, cells of the hind- and fore-gut, in addition to the anal pads, exhibit enzyme activity (Fig. 2d). Muscles also exhibit weak fl-gal expression. Anti-h antibodies also label the same sets of cells late in embryogenesis (Hooper and Ish-Horowicz personal communication). Amongst the population of embryos deriving from P(Lac,ry+)A L43a/TM3 flies, about one quarter showed variant patterns of fl-gal activity during gastrulation and later stages. These abnormal patterns were shown to be homozygotes for the insertion, using an anti-fushi tarazu antibody (not shown), which shows an altered pattern of expression with respect to wildtype during embryogenesis (Carroll and Scott 1986). At the cellular blastoderm stage, the homozygous and heterozygous patterns of fl-gal are indistinguishable. During germ band extension however, the stripes of fl-gal expression, seen earlier, persist in homozygotes for the insertion (Fig. 2e). By the end of embryogenesis, all cells of L43a homozygotes express fl-gal (not shown). These observations indicate that normal h products are required for repression of its own regulatory units during gastrulation. This repression is presumably indirect, perhaps via other genes essential for the normal segmentation process (Scott and O'Farrell 1986; Akam 1987), since cells that normally never express h products now express fl-gal. Direct influences of h products on their own regulatory elements, however, cannot be excluded. An alternative possibility, to explain the ectopic expression of fl-gal in insertion homozygotes (Fig. 2e), might be that sequences essential for the repression of the gene lie 3' to, or even span, the site of integration. fl-gal activity was also tested at other stages of development. Specific labelling of parts of all of the imaginal discs (Fig. 3 b), the salivary glands (Fig. 3 a), the larval epidermis, the trachea and isolated ceils of the larval central nervous system was found (Fig. 3 b). Adult ovaries also possessed fl-gal activity; this was restricted to the follicle cells, throughout oogenesis (Fig. 3 c). Recently messages from the h gene have been detected in the salivary glands using in situ hybridisation to tissue sections (Pinchin and Ish-Horowicz personal communication), supporting the idea that these reporter gene expressions reflect the activity of the adjacent h gene. Other molecular evidence for expression of the h gene has been reported for these moments in development (Ish-Horowicz et al. 1985); h messages are detectable during larval and pupal stages and "almost undetectable" in the follicles of ovaries on Northern blots (IshHorowicz et al. 1985). The serendipitous discovery of an insertion in the hairy gene suggests that promoter-reporter gene fusions can be used to isolate genes in eukaryotes and that the expression of fl-gal will, in some cases, follow accurately that of the

511

Fig. 3. fl-galactosidase activity of P(Lac,ry+)A L43a in larval tissues (Panels a and b) and oogenesis (Panel c). a /?-gal expression in the salivary glands; enzyme is restricted to the nuclei since the P(Lac,ry+)A insertion carries the first exon of the tranposase made by the P element (O'Kane and Gehring 1987). b /?-gal activity in leg discs (ld), eye discs (ed), and larval central nervous system (cns). Note that two rings of cells (arrows) light up in the cerebral hemispheres and isolated cells in each segment of the ventral ganglia. e/~-gal expression in the ovaries of an adult female. Only follicle cells exhibit enzyme activity; both anteriorly-located columnar (cfc) and nurse cell associated (naf) follicle cells stain blue, in stage 10 follicles (King 1970). In younger follicles (to the "left), all follicle cells possess enzyme activity especially at the poles

adjacent gene. In the case o f this h insertion, the site o f integration is close to the transcriptional start site and receives much of the s p a t i o - t e m p o r a l information that hairy does normally. The insertion is also in the same orientation as the adjacent transcript. At present, we cannot say whether the site and orientation o f the insertion with respect to the gene is o f importance. The 5' regulatory regions o f eukaryotic genes seem to consist o f several distinct units required for independent spatio-temporal expressions (Hiromi et al. 1985; D i N a r d o e t al. 1988; B ienz et al. 1988; How-

ard et al. 1988). Presumably, if an insertion were in the middle o f an a r r a y o f such units, only the units 5' to the reporter gene p r o m o t e r would be detected; the others, 3' to the insertion, might be too far for these regulatory elements to act in the presence o f a relatively large insertion (11 k b in this case). I f so, such patterns will be o f use to analyse parts o f the regulatory regions o f genes next d o o r to the reporter gene insertions. In Drosophila, P-elements often lie close to the 5' end o f the gene (Ish-Horowicz et al. 1985; T s u b o t a et al. 1985;

512 Searles et al. 1986; Kelley et al. 1987; Chia, personal communication). However, these elements cause m u t a n t phenotypes; about 80-90% of homozygous P element insertions give no phenotype ( O ' K a n e and Gehring 1987; Cooley et al. 1988; Fasano and Kerridge 1988). Further work is required to see whether reporter gene expression, especially for those insertions giving no phenotype, will reflect the activity of a subset, or the entire set, of the regulatory units of a neighbouring gene.

Acknowledgements. We thank C. O'Kane and W. Gehring for their PLacA92 plasmid, D. Ish-Horowicz for communicating unpublished data and materials from the hairy gene region and D. IshHorowicz and P. Ingham for helpful discussions and ideas. References Akam M (1987) The molecular basis for metameric pattern in the Drosophila embryo. Development 101 : 1 22 Allen ND, Cran DG, Barton SC, Hettle S, Reik W, Azim Surani M (1988) Transgenes as probes for active chromosomal domains in mouse development. Nature 333:852-855 Bienz M, Saari G, Tremml G, Miiller J, Zfist B, Lawrence PA (1988) Differential regulation of Ultrabithorax in two germ layers of Drosophila. Cell 53 : 567 576 Campos-Ortega JA, Hartenstein V (1985) The Embryonic Development of Drosophila melanogaster. Springer, Berlin Heidelberg New York Tokyo Carroll SB, Scott MP (1986) Zygotically active genes that affect the spatial expression of thefushi-tarazu segmentation gene during early Drosophila embryogenesis. Cell 45:113 126 Carroll SB, Laughton A, Thalley BS (1988) Expression, function, and regulation of the hairy segmentation protein in the Drosophila embryo. Genes Dev 2:883 890 Cooley L, Kelley R, Spradling A (1988) Insertional mutagenesis of the Drosophila genome with single P elements. Science 239:1121 1128 DiNardo S, Sher E, Heemskerk-Jongens J, Kassis JA, O'Farrell PH (1988) Two-tiered regulation of spatially patterned engrailed gene expression during Drosophila embryogenesis. Nature 332:604-609 Fasano L, Kerridge S (1988) Monitoring positional information during oogenesis in adult Drosphila. Development 104:245-254 Feinberg AP, Vogelstein B (1983) A technique for radiolabeting DNA restriction endonuclease fragments to high specific activity. Anal Biochem 132:6-13 Hiromi Y, Kuroiwa A, Gehring WJ (1985) Control elements of the Drosophila segmentation genefushi-tarazu. Cell 43:603-613 Howard KR, Ingham PW, Rushlow C (1988) Region specific alleles of the Drosophila segmentation gene hairy. Genes Dev 2:1037 1046 Ingham PW, Pinchin SM, Howard KR, Ish-Horowicz D (1985a) Genetic analysis of the hairy gene in Drosophila. Genetics 111 :463-486 Ingham PW, Howard KR, Ish~Horowicz D (1985b) Transcription pattern of the Drosophila segmentation gene hairy. Nature 318 :439-445

Ish-Horowicz D, Howard KR, Pinchin SM, Ingham PW (1985) Molecular and genetic analysis of the hairy locus in Drosophila. Cold Spring Harb Syrup Quant Biol 50:135-144 Junakovic N (1980) Variability in the molecular organization of the 5S RNA genes among strains of Drosophila melanogaster. Nucl Acids Res 8 : 3611 3622 Kelley MR, Kidd S, Berg RL, Young MW (1987) Restriction of P-element insertions at the Notch locus of Drosophila melanogaster. Mol Cell Biol 7 : 1545-1548 King RC (1970) Ovarian Development in Drosophila melanogaster. Academic Press, New York Lewis EB (1978) A gene complex controlling segmentation in Drosophila. Nature 276: 565-570 Lindsley DL, Grell EH (1968) Genetic Variations of Drosophila melanogaster. Carnegie Institute of Washington, Washington DC MacDonald PM, Struhl G (1986) A molecular gradient in early Drosophila embryos and its role in specifying the body pattern. Nature 324:537-545 Mitcheson TJ, Sedat J (1983) Localization of antigenic determinants in whole Drosophila embryos. Dev Biol 99:261-264 N~sslein-Volhard C, Wieschaus E (1980) Mutations affecting segment number and polarity in Drosophila. Nature 287:795-801 O'Kane CJ, Gehring WJ (1987) Detection in situ of genomic regulatory elements in Drosophila. Proc Natl Acad Sci USA 84:9123-9127 Rubin GM, Spradling AC (1982) Genetic transformation of Drosophila with transposable element vectors. Science 218:348 353 Searles LL, Greenleaf AL, Evans Kemp W, Voelker RA (1986) Sites of P element insertion and structures of P element deletions in the 5' region of Drosophila melanogaster RpII215. Mol Cell Biol 6:3312-3319 Scott MP, O'Farrell PH (1986) Spatial programming of gene expression in early Drosophila embryogenesis. Ann Rev Cell Biol 2: 49-80 Silhavy TJ, Beckwith JR (1985) Uses of Lac fusions for the study of biological problems. Microbiol Rev 49 : 398-418 Takahashi K, Vigneron M, Matthes H, Wildeman A, Zenke M, Chambon P (1986) Requirement of stereospecific alignments for initiation from the simian virus 40 early promoter. Nature 319:121 126 Teeri TH, Herrera-Estrella L, Depieker A, Van Montagu M, Palva ET (1986) Identification of plant promoters in situ by T-DNAmediated trancriptional fusions to the npt-II gene. EMBO J 5:1755-1760 Tsubota S, Ashburner M, Schedl P (1985) P-element induced control mutations at the r gene of Drosophila melanogaster. Mol Cell Biol 5 : 2567-2574 Van der Meer JM (1977) Optical clean and permanent wholemount preparation for phase contrast microscopy of cuticular structures of insect larvae. Drosophila Inform Serv 52:160 Wakimoto BT, Kaufman TC (1981) Analysis of larval segmentation in lethal genotypes associated with the Antennapedia gene complex (ANT-C) in Drosophila melanogaster. Dev Biol 81:5164 Received September 7 / Accepted in revised form November 24, 1988