The homeotic genes of Drosophila encode transcription factors that specify morphological differences between segments. To identify the genes that they control, we developed a chromatin immunopurification approach designed to isolate in vivo binding sites for the products of the homeotic gene Ultrabithorax. Here, we report the analysis of one immunopurified binding site. This 110 bp fragment maps within a regulatory region of a gene under homeotic control, connectin. A 4 kb DNA fragment, including the immunopurified binding site, is suf-ficient to reproduce the appropriate homeotic control within a subset of the full tissue distribution of connectin. Analysis of the role of the 110 bp binding site indicates that it mediates transcriptional controls by Ultrabithorax and other homeotic genes. This is the first report of a functional in vivo binding site isolated using the chromatin immunopurification method. We also show that the protein product of the connectin gene is predicted to be a cell-surface molecule containing leucine-rich repeats. The protein, connectin, can mediate cell-cell adhesion thus suggesting a direct link between homeotic gene function and processes of cell-cell recognition.

Differences between metameres in Drosophila are con-trolled by a small group of homeotic selector genes, expressed and required in specific anatomical domains (Lewis, 1963; Garcia-Bellido, 1975; Lewis, 1978; Akam, 1987). For example, the Ultrabithorax gene (Ubx) is expressed predominantly in a domain extending from parasegment (PS)5 to PS13 and is required primarily in the third thoracic and first abdominal segments (Lewis, 1978; Akam and Martinez-Arias, 1985; White and Wilcox, 1985). As long ago as 1975, it was proposed that the homeotic loci select particular developmental pathways by regulating a battery of subordinate target genes and that these target genes are required to coordinate the building of structures characteristic of each segment (Garcia-Bellido, 1975). More recently, the wealth of molecular data on homeotic genes have demonstrated that they encode sequence-specific DNA-binding proteins (Beachy et al., 1988; Muller et al., 1988) capable of activating or repressing transcription from a variety of reporter gene constructs in vitro (Thali et al., 1988; Krasnow et al., 1989; Winslow et al., 1989). How-ever, the in vivo target genes have been difficult to identify (Andrew and Scott, 1992).

The classical genetic approaches, which have been so successful in identifying the early pattern-forming genes in Drosophila development, have so far failed to identify genes that act immediately downstream of the homeotics in the developmental genetic hierarchy, although Kennison and Tamkun (1988) have isolated a number of loci inter-acting with dominant homeotic mutations. Recently, Wagner-Bernholtz and colleagues have used P-element-mediated enhancer detection to screen for regulatory elements that respond, either directly or indirectly, to ectopic expression of Antennapedia (Antp). Using this approach, a previously unidentified gene, spalt major, has been identi-fied as a target of negative regulation by Antp (Wagner-Bernholz et al., 1991). In addition, a few promising candi-date targets have been identified from amongst the many genes that have previously been cloned. Perhaps the best characterised so far is the decapentaplegic (dpp) gene, which plays a role in both dorsoventral and proximodistal pattern formation (St. Johnston and Gelbart, 1987). The protein encoded by dpp is related to a vertebrate growth factor and functions in the visceral mesoderm in Drosophila to induce the expression of another homeotic gene, labial, in the midgut endoderm (Immergluck et al., 1990; Pangani-ban et al., 1990). Genetic experiments show that the dpp gene is positively regulated by Ubx and negatively regulated by another homeotic gene, abdominal-A (Immergluck et al., 1990; Reuter et al., 1990). However, none of the approaches described so far have been able to demonstrate a direct interaction between homeotic proteins and the regulatory DNA of a candidate target gene.

In an effort to identify direct targets of homeotic regulation, we have developed an immunopurification approach that enriches for short chromatin fragments through their association with endogenous Ubx protein (Gould et al., 1990). In brief, embryonic nuclei are digested with a restric-tion enzyme and then lysed; the soluble chromatin is affinity-purified against a matrix containing antibodies to Ubx proteins; the DNA fragments from the immunopurified chromatin are then cloned. Using this method, at least two candidate Ubx in vivo binding sites were isolated. Both target sites (clones 35 and 48) are capable of highly specific binding to Ubx proteins in vitro and are located in genomic DNA close to transcription units regulated by Ubx in vivo. These two transcripts were therefore presented as good candidates for genes directly regulated by Ubx (Gould et al., 1990). We have continued the analysis of these candidates in order to ask, firstly, whether the immunopurified target sites identify regulatory elements mediating homeotic control and, secondly, what can be learnt of the developmen-tal function of these putative target genes.

Here we report the molecular characterisation of the gene, connectin, found adjacent to clone 35 and we demon-strate that the short (110 bp) immunopurified fragment is an important component of a connectin regulatory element under homeotic gene control. Nose et al. (1992) have inde-pendently isolated this gene from a screen of enhancer trap insertions that are expressed in muscle cells. The protein produced by this locus has been named connectin as it may have a role in the formation of connections betweeen certain motor neurons and their target muscles (Nose et al., 1992).

Fly strains

Flies were maintained at 25°C on cornmeal-agar medium and embryos were collected onto yeasted apple-juice agar plates. Stan-dard genetic markers are as described by Lindsley and Zimm (1985, 1990). Wild-type embryos were of the Oregon R strain. Homeotic null mutations used are as follows: Df(3R)bxd100; a deficiency of the entire Ubx gene (Lewis, 1978), abd AM1 and UbxMX6 abd-AM1 (Sanchez-Herrero et al., 1985) and AntpW10 (Wakimoto and Kaufman, 1981). Recipients for P-element transformation were homozygous for cn and ry.

General DNA techniques and sequencing

Standard techniques used are as described by Sambrook et al. (1989). An EMBL3 genomic library made from the iso-1 strain (gift of J. Tamkun) was used to accomplish a genomic walk of about 65 kb at the cytological region 64C. The initial isolation of clones from this region has been described previously (Gould et al., 1990).

A 10 kb SalI genomic fragment, detecting the embryonic transcript distribution, was used to screen a λgt10 cDNA library constructed from 3-12 hour embryonic poly(A)+ RNA. Subsequently the longest hybridising clone, gt35.8 (2.6 kb), was used to screen two independent size-selected plasmid cDNA libraries, one from 8-12 hour embryos and the other from 12-24 hour embryos (Brown and Kafatos, 1988); 40 primary isolates were characterised and all were found to be of one class but with differing 5′ and 3 ′ extents. Sequence analysis shows that the 5′ end of pNB35.8 is identical with that of two other independent cDNA clones, one isolated from a separate library. However, the 3′ extents of sev-eral clones sequenced were found to be unique and in no case was a poly(A) tract found. It therefore seems likely that additional (but short) 3 ′ message sequences are not included within pNB35.8.

cDNA and genomic clones were sequenced by the chain ter-mination method (Sanger et al., 1977) using double-stranded tem-plates and Sequenase (USB). Large regions were covered using nested deletions together with primer synthesis. Nested deletions of gt35.8 were created using the exonuclease III/mung bean nucle-ase system (Stratagene) and sequenced as single-stranded M13 templates using Klenow polymerase.

Sequences were analysed using the Staden (Staden, 1982) and GCG (Devereux et al., 1984) software.

Whole-mount in situ hybridisations using digoxigenin-labelled DNA probes were carried out according to Tautz and Pfeifle (1989) with some modifications (Gould et al., 1990). For hybridizations using RNA probes, the digoxigenin-labelled RNA probes were made according to manufacturers instructions (Boehringer-Mannheim, RNA labelling kit) and the in situ hybridization carried out as for the DNA probes except that the prehybridizations, hybridizations and post-hybridization washes were at 70°C.

Transformation constructs

P-element mediated germ-line transformation of cn ry embryos was according to Spradling (1986) but using pUChsπΔ2-3 (gift of Ernst Hafen) as a constitutive source of transposase. Regulatory regions of the connectin gene were cloned into the blunted XbaI site of pHZ50PL (Hiromi and Gehring, 1987).

The 10 kb SalI genomic fragment was inserted into pBluescript (Stratagene). From this 10 kb subclone, a 4 kb EcoRI fragment (representing the 4 kb EcoRI-SalI genomic interval) was excised, blunted with Klenow polymerase, and then inserted into pHZ50PL to give the 4 kb construct. The 1 kb Pst-SalI genomic region was subcloned into pBluescript and excised as a blunt SmaI-HincII fragment for insertion into pHZ50PL yielding the 1 kb construct. Precise deletion of clone 35 sequences (a 110 bp HaeIII frag-ment; (Gould et al., 1990)) was accomplished from a 0.7 kb PstI-SacI genomic subclone using the Inverse Polymerase Chain Reac-tion with Pfu DNA polymerase (Stratagene) at high stringency. Primers were chosen to generate an additional GC dinucleotide at the deletion junction, thus converting the two HaeIII half-sites into a diagnostic ApaI site. Sequencing confirmed that no inadvertant mutations had been introduced. A 0.4 kb SacI-SpeI fragment, containing the deletion of clone 35, was exchanged with the wild-type 0.5 kb SacI-SpeI fragment in both the 4 kb and 1 kb constructs to give the 4 kb-35 and 1 kb-35 constructs, respectively.

X-gal staining

X-gal staining was performed as described (Protocol 76; Ash-burner, 1989b), except that 5% glutaraldehyde in PBS was used as the fixative. After the staining reaction, embryos were mounted in PBS, 10% glycerol and 0.1% Triton X-100 and photographed using differential-interference or bright-field optics.

Cloning and sequencing of the connectin gene

The immunopurified 110 bp clone 35 was used as a probe to initiate a genomic walk at the cytological location 64C. Screening by in situ hybridization with genomic probes from this region revealed a transcription unit under homeotic gene control (Gould et al., 1990; Fig. 1). To characterize this 35 transcription unit, we isolated many cDNAs covering the cloned interval, the longest being approximately 4.5 kb in length (pNB35.8). This is close to the size of the major polyadenylated RNA species (approximately 5 kb) detected on blots of 0-24 hour embryonic RNA using probes from this region (data not shown). Using this cDNA as a probe in embryo whole-mount in situ hybridisation, the same transcript distribution was obtained as had been seen previously with genomic probes.

Fig. 1.

Organisation of the 35 transcription unit. The positions of restriction sites for HindIII (H), EcoRI (R), and SalI (S) are indicated above a line representing the genomic walk (scale above line in kb). Restriction fragments hybridising to cDNA clone pNB35.8 (see text and Materials and methods) are shown as solid bars beneath the walk with the 5′ to 3′ direction indicated. The location of the 110 bp clone 35 used to initiate this genomic walk is indicated, as are the extents of the lambda phage covering the cloned region. The position of the EcoRI-SalI fragment used in the 4 kb transformation construct is also shown.

Fig. 1.

Organisation of the 35 transcription unit. The positions of restriction sites for HindIII (H), EcoRI (R), and SalI (S) are indicated above a line representing the genomic walk (scale above line in kb). Restriction fragments hybridising to cDNA clone pNB35.8 (see text and Materials and methods) are shown as solid bars beneath the walk with the 5′ to 3′ direction indicated. The location of the 110 bp clone 35 used to initiate this genomic walk is indicated, as are the extents of the lambda phage covering the cloned region. The position of the EcoRI-SalI fragment used in the 4 kb transformation construct is also shown.

To define the 5′ and 3′ limits of any open reading frame (ORF) present, several different cDNA clones were transcribed and translated in vitro. This indicated that both gt35.8 (2.6 kb) and pNB35.8 (4.5 kb) cDNAs encode a similarly sized polypeptide of about 72×103Mr (data not shown). Sequencing of the gt35.8 cDNA revealed an ORF of 2046 nucleotides (ORF:35), encoding a polypeptide of 682 amino acids (76 ×103Mr; Fig. 2). The codon usage found within this ORF:35 agrees well with that predicted for a Drosophila gene (Ashburner, 1989a). The second methionine codon of this reading frame is predicted to be slightly more favourable as a translation start than the first methionine codon, lying 27 nucleotides upstream (Cavener, 1987). Hydropathy analysis (Fig. 3A) indicates that both the amino and carboxy termini are predicted to be highly hydrophobic. The amino-terminal hydrophobic region of 26 amino acids is characteristic of a signal sequence (von Heijne, 1986). There is no likely transmembrane domain but the carboxy-terminal hydrophobic region of 25 residues is consistent with the presence of a glycosyl-phosphatidyl inositol linkage to the cell membrane (GPI-anchor; Fergu-son and Williams, 1988). The existence of a GPI anchor has been confirmed by the phosphatidylinositol specific-phospholipase C induced release of this protein from trans-formed Drosophila S2 cells expressing this ORF:35 (L. Meadows, A. P. G., R. A. H. W., unpublished results).

Fig. 2.

Nucleotide and predicted amino acid sequences of the 35 cDNAs. The sequence of nucleotides 1 to 343 is from cDNA pNB35.8, whilst nucleotides 344 (arrowhead) to 2973 are from the overlapping clone gt35.8 (see Materials and methods). Conceptual translation is initiated at position 496 but the upstream ATG at position 469 is indicated by double underlining. Amino and carboxy terminal hydrophobic stretches of amino acids are underlined and potential N-linked glycosylation sites are shown in bold. This sequence has been submitted to the EMBL sequence data base (accession number X68701).

Fig. 2.

Nucleotide and predicted amino acid sequences of the 35 cDNAs. The sequence of nucleotides 1 to 343 is from cDNA pNB35.8, whilst nucleotides 344 (arrowhead) to 2973 are from the overlapping clone gt35.8 (see Materials and methods). Conceptual translation is initiated at position 496 but the upstream ATG at position 469 is indicated by double underlining. Amino and carboxy terminal hydrophobic stretches of amino acids are underlined and potential N-linked glycosylation sites are shown in bold. This sequence has been submitted to the EMBL sequence data base (accession number X68701).

Fig. 3.

Analysis of the connectin protein precursor. (A) Hydropathy plot using the method of Kyte and Doolittle (1982). Significant regions of hydrophobicity (scoring positively) are found at both the amino and carboxy termini. (B) Alignment of individual LRRs showing highly conserved amino acids boxed. Bracketed numbers to the right indicate the position of each 24 amino acid LRR within the protein precursor. A consensus sequence for the 10 LRRs is shown below. (C) Schematic representation of the amino acid sequence of the protein precursor. Amino (N) and carboxy (C) terminal hydrophobic regions are indicated by black boxes. Shaded boxes show the position of individual LRRs and asterisks mark potential N-linked glycosylation sites.

Fig. 3.

Analysis of the connectin protein precursor. (A) Hydropathy plot using the method of Kyte and Doolittle (1982). Significant regions of hydrophobicity (scoring positively) are found at both the amino and carboxy termini. (B) Alignment of individual LRRs showing highly conserved amino acids boxed. Bracketed numbers to the right indicate the position of each 24 amino acid LRR within the protein precursor. A consensus sequence for the 10 LRRs is shown below. (C) Schematic representation of the amino acid sequence of the protein precursor. Amino (N) and carboxy (C) terminal hydrophobic regions are indicated by black boxes. Shaded boxes show the position of individual LRRs and asterisks mark potential N-linked glycosylation sites.

The most striking feature of the predicted polypeptide is the presence of 10 copies in tandem of a 24 amino acid leucine-rich repeat (LRR; Fig. 3B,C). This motif has a char-acteristic spacing of hydrophobic residues, in particular leucine and is found in many evolutionarily and function-ally diverse proteins (Takahashi et al., 1985; Kataoka et al., 1985; Titani et al., 1987; Fisher et al., 1989; Reinke et al., 1988; Hashimoto et al., 1988; Rothberg et al., 1990). The consensus motif for the LRRs of ORF:35 is shown in Fig. 3B. In several molecules containing blocks of LRRs, there are also conserved flanking regions containing cysteine clusters (Lopez et al., 1987; Hickey et al., 1989; Schneider and Schweiger, 1991). Mutations affecting cysteine residues in a carboxy-terminal flanking region of the Drosophila LRR-protein Toll give rise to dominant Toll mutations (Schneider et al., 1991). ORF:35 shows only very limited homology to these conserved flanking sequences; however, there are cysteines at positions 100, 110, 112 and 126 followed by the sequence DTT ocurring just on the amino-terminal side (137-139) of the LRR block and ICDCRL on the carboxy-terminal side (409-413). The other notable feature of the sequence of ORF:35 is that the amino-terminal region preceding the LRR block is serine-rich.

Comparison of the sequence of ORF:35 with that from a gene isolated by Nose et al. (1992) revealed identity and we refer to this gene as connectin.

The expression of the connectin gene

Before considering the regulation of the connectin gene, we first discuss its expression pattern in more detail than has been presented previously (Gould et al., 1990). Transcripts are first detected by in situ hybridization in the extended germ band (stage 10; stages according to Campos-Ortega and Hartenstein, 1985) in a subset of the splanchnopleura, the cell layer that will subsequently give rise to the visceral mesoderm (Fig. 4A). There is expression in the presumptive foregut and hindgut mesoderm and in a segmentally repeated subset of the presumptive midgut visceral mesoderm. The labelling in the mesoderm around the foregut and hindgut persists but the expression in the midgut visceral mesoderm subsets narrows to small segmentally repeated patches by the end of stage 11 (Fig. 4B) and dis-appears after germband retraction. In late stage 11, con -nectin RNA appears in the ventral nervous system and in the somatic mesoderm (Fig. 4B). The labelling in the ven-tral nervous system is initially in two longitudinal bands of cells with little segment-to-segment modulation in intensity. In contrast, the initial expression in the somatic mesoderm is highly modulated, being strong in the three thoracic seg-ments and relatively weak in abdominal segments. There are differences, too, between the thoracic segments with the labelling in T1 being slightly weaker and having a different arrangement of labelled cells (Fig. 4C,D). By stage 14 (Fig. 4E,F), the intensity differences between segments in the somatic mesoderm are less pronounced but there are clear differences in the arrangement of mesodermal cell groups between thoracic and abdominal segments. From studies with antibodies against the connectin protein, the thoracic ventrolateral clusters correspond to the precursors of the pleural external transverse muscles, whilst in the abdomen the ventral clusters correspond to the ventral superficial oblique muscles and the lateral clusters to the pleural external transverse group (L. Meadows, A. P. G., R. A. H. W., unpublished observations; see also Nose et al., 1992). The labelling in the mesoderm declines shortly after stage 14. In the ventral nervous system, the initially uniform labelling is succeeded by a more modulated pat-tern. Differences between the thoracic and abdominal neuromeres appear at stage 14 and, by stage 16, there is strong labelling in the anterior central nervous system, including the brain lobes and the thoracic neuromeres but little labelling in most of the abdominal neuromeres. Labelling is also seen associated with the gnathal sense organs from stage 11 onwards.

Fig. 4.

Expression of connectin RNA in wild-type embryos. (A) Stage 11; lateral view. Labelling is seen in segmentally repeated subsets of the splanchnopleura, which will form the visceral mesoderm of the midgut, and in the visceral mesoderm of the foregut (fgm) and hindgut (out of focus). (B) Later stage 11; lateral view. Strong labelling is seen in the visceral mesoderm of the foregut (fgm) and hindgut (hgm) and the arrowheads indicate two of the segmentally repeated subsets of labelling in the presumptive visceral mesoderm of the midgut. The predominant labelling in the three thoracic segments T1, T2 and T3 is in the somatic mesoderm. There is also weaker labelling in cells of the somatic mesoderm in the abdominal segments. The arrow indicates labelling in one of the longitudinal tracts of cells in the ventral nervous system. (C) Stage 12; lateral view focussing on the somatic mesoderm. Labelling is predominantly in specific cell clusters in the thoracic somatic mesoderm. (D) Stage 12; ventral view. The two longitudinal tracts of cells in the ventral nervous system are stained with uniform intensity along the anteroposterior axis. In the somatic mesoderm, there is a clear difference in labelling intensity between the thoracic and abdominal segments. Also, T1 is weaker than T2/T3 and shows a different arrangement of labelled cells. The arrow indicates labelling associated with gnathal sense organ precursors. (E) Stage 14; dissected preparation. The labelling in the ventral nervous system (vns) shows the beginnings of segment-to-segment modulation with the thoracic neuromeres showing slightly higher intensity labelling than the abdominal. In the somatic mesoderm, the intensity differences between thoracic and abdominal segments are less pronounced than at earlier stages. There are clear differences in the arrangement of the labelled cell clusters between thoracic and abdominal segments. (F) Stage 14; lateral view focussing on the somatic mesoderm showing the positions of the labelled cell clusters. The labelling in the T1 somatic mesoderm is rather obscured by out-of-focus labelling in the foregut visceral mesoderm. (G) Stage 16; labelling is predominantly in the brain lobes and in the thoracic neuromeres of the ventral nervous system. The arrows indicate labelling associated with head and gnathal sense organs. DNA probes were used for the in situ hybridizations in A, E and G; RNA probes were used for the others.

Fig. 4.

Expression of connectin RNA in wild-type embryos. (A) Stage 11; lateral view. Labelling is seen in segmentally repeated subsets of the splanchnopleura, which will form the visceral mesoderm of the midgut, and in the visceral mesoderm of the foregut (fgm) and hindgut (out of focus). (B) Later stage 11; lateral view. Strong labelling is seen in the visceral mesoderm of the foregut (fgm) and hindgut (hgm) and the arrowheads indicate two of the segmentally repeated subsets of labelling in the presumptive visceral mesoderm of the midgut. The predominant labelling in the three thoracic segments T1, T2 and T3 is in the somatic mesoderm. There is also weaker labelling in cells of the somatic mesoderm in the abdominal segments. The arrow indicates labelling in one of the longitudinal tracts of cells in the ventral nervous system. (C) Stage 12; lateral view focussing on the somatic mesoderm. Labelling is predominantly in specific cell clusters in the thoracic somatic mesoderm. (D) Stage 12; ventral view. The two longitudinal tracts of cells in the ventral nervous system are stained with uniform intensity along the anteroposterior axis. In the somatic mesoderm, there is a clear difference in labelling intensity between the thoracic and abdominal segments. Also, T1 is weaker than T2/T3 and shows a different arrangement of labelled cells. The arrow indicates labelling associated with gnathal sense organ precursors. (E) Stage 14; dissected preparation. The labelling in the ventral nervous system (vns) shows the beginnings of segment-to-segment modulation with the thoracic neuromeres showing slightly higher intensity labelling than the abdominal. In the somatic mesoderm, the intensity differences between thoracic and abdominal segments are less pronounced than at earlier stages. There are clear differences in the arrangement of the labelled cell clusters between thoracic and abdominal segments. (F) Stage 14; lateral view focussing on the somatic mesoderm showing the positions of the labelled cell clusters. The labelling in the T1 somatic mesoderm is rather obscured by out-of-focus labelling in the foregut visceral mesoderm. (G) Stage 16; labelling is predominantly in the brain lobes and in the thoracic neuromeres of the ventral nervous system. The arrows indicate labelling associated with head and gnathal sense organs. DNA probes were used for the in situ hybridizations in A, E and G; RNA probes were used for the others.

Transcription driven by a 4 kb region encompassing clone 35

The connectin gene was isolated as a transcription unit neighbouring the immunopurified clone 35, a candidate in vivo binding site for Ubx protein. We have previously shown that connectin expression is under the control of Ubx and other homeotic genes (Gould et al., 1990). If the immunopurification has identified a functional Ubx protein binding site, it should be possible to demonstrate that the 110 bp clone 35 sequence mediates Ubx-dependent regula-tion of connectin. The approach used was first to ask whether the immunopurified fragment had led us to a con -nectin regulatory element under homeotic gene control. We considered it unlikely that the 110 bp sequence would function efficiently by itself as an enhancer element and so larger genomic restriction fragments spanning the clone 35 region were tested for their ability to drive expression of a lacZ reporter gene from an hsp70 minimal promoter in transformed embryos (Hiromi and Gehring, 1987; Fig. 5). In Fig. 6, we show that the 4 kb SalI-EcoRI fragment encompassing clone 35 is sufficient to give a consistent pat-tern of lacZ expression corresponding to a subset of the total connectin expression pattern (compare Fig. 4D-F and Fig. 6A-C). (We note that this element failed to give the same staining pattern when placed upstream of the hsp70 promoter in the opposite orientation and therefore does not behave like a true enhancer.) High levels of reporter gene expression were detected, from the extended germ band stage onwards, in small groups of cells in the gnathal and thoracic segments and in a posterior segment, probably A9 (Fig. 6A-C). The pattern of gnathal staining correlates with the developing gnathal sense organs, which express con -nectin RNA, and the thoracic labelling correlates with the ventrolateral connectin expression in the somatic mesoderm described above for stage 13 embryos (Fig. 4E,F). One high-expressing line, 4.40, also showed some abdominal labelling (Fig. 6A) which correlated well with the positions of the ventral and lateral clusters of connectin expression in the abdominal somatic mesoderm. In marked contrast to the endogenous connectin RNA distribution, no staining is seen in the central nervous system (Fig. 6B). Staining in gnathal segments was first detectable in stage 10 embryos whilst the thoracic and A9 labelling was not observed until slightly later at stage 11. Thus the 4 kb construct gives a subset of the complete connectin pattern with predominant expression in gnathal and thoracic segments.

Fig. 5.

Structure of transformation constructs containing connectin regulatory regions. For each of the four constructs (4 kb, 1 kb, 4 kb-35, and 1 kb-35; see text), the connectin DNA is shown boxed whilst the minimal hsp70 promoter-lacZ sequences of the vector are shown (see Materials and methods) as a line. The arrow indicates the start and direction of lacZ transcription from these constructs. S, P and R indicate the location of SalI, PstI and EcoRI restriction sites, respectively (see Fig. 1). The position of clone 35 (in the 4 kb and 1 kb constructs) is indicated by a filled box, and its absence in the 4 kb-35 and 1 kb-35 is indicated by a line. The “no. of lines” indicates the total number of independent transformant lines analysed for lacZ expression (denominator) and the number of these showing a consistent pattern of expression (numerator).

Fig. 5.

Structure of transformation constructs containing connectin regulatory regions. For each of the four constructs (4 kb, 1 kb, 4 kb-35, and 1 kb-35; see text), the connectin DNA is shown boxed whilst the minimal hsp70 promoter-lacZ sequences of the vector are shown (see Materials and methods) as a line. The arrow indicates the start and direction of lacZ transcription from these constructs. S, P and R indicate the location of SalI, PstI and EcoRI restriction sites, respectively (see Fig. 1). The position of clone 35 (in the 4 kb and 1 kb constructs) is indicated by a filled box, and its absence in the 4 kb-35 and 1 kb-35 is indicated by a line. The “no. of lines” indicates the total number of independent transformant lines analysed for lacZ expression (denominator) and the number of these showing a consistent pattern of expression (numerator).

Fig. 6.

Expression from the 4 kb construct in wild-type and homeotic mutant embryos. (A) 4 kb construct, line 4.20 in wild-type background; stage 13; ventrolateral view. The most prominent labelling is in the thoracic mesoderm. Labelling is also seen in the gnathal segments and in the abdominal mesoderm. The pattern of expression in the mesoderm closely mimics the endogenous connectin expression (B) 4 kb construct, line 4.40 in wild-type background; stage 13; ventral view. In this line, the labelling is more tightly restricted to gnathal and thoracic segments with little or no labelling of abdominal segments except A9. Note absence of expression in the central nervous system. C-F show comparable ventrolateral views of the 4 kb construct, line 4.40, in the various genetic backgrounds. (C) 4 kb construct, line 4.40 in wild-type background; stage 13. (D) Ubx mutant (Df bxd100) stage 13 embryo showing staining anteriorly in gnathal and thoracic segments as in wild type, but with additional labelling in A1 and A2. (E) Ubx abd-A mutant, stage 13. Labelling extends strongly posteriorly to A3, and there is some staining in abdominal segments between A3 and A9. (F) Antp mutant, stage 13. Labelling is restricted to gnathal segments, T1 and A9.

Fig. 6.

Expression from the 4 kb construct in wild-type and homeotic mutant embryos. (A) 4 kb construct, line 4.20 in wild-type background; stage 13; ventrolateral view. The most prominent labelling is in the thoracic mesoderm. Labelling is also seen in the gnathal segments and in the abdominal mesoderm. The pattern of expression in the mesoderm closely mimics the endogenous connectin expression (B) 4 kb construct, line 4.40 in wild-type background; stage 13; ventral view. In this line, the labelling is more tightly restricted to gnathal and thoracic segments with little or no labelling of abdominal segments except A9. Note absence of expression in the central nervous system. C-F show comparable ventrolateral views of the 4 kb construct, line 4.40, in the various genetic backgrounds. (C) 4 kb construct, line 4.40 in wild-type background; stage 13. (D) Ubx mutant (Df bxd100) stage 13 embryo showing staining anteriorly in gnathal and thoracic segments as in wild type, but with additional labelling in A1 and A2. (E) Ubx abd-A mutant, stage 13. Labelling extends strongly posteriorly to A3, and there is some staining in abdominal segments between A3 and A9. (F) Antp mutant, stage 13. Labelling is restricted to gnathal segments, T1 and A9.

Homeotic regulation of the 4 kb region

The clear differences in expression between thoracic and abdominal segments observed with the 4 kb construct suggested that the 110 bp immunopurified sequence had led us to a regulatory element under homeotic gene control. The reduced levels of expression of the 4 kb construct in abdominal segments A1-A8 suggest a down regulation by Ubx and the abdominal homeotic genes. In order to test this, we examined the expression in embryos homozygous for Ubx or abd-A null mutations and also in embryos homozygous for the Ubx abd-A double mutant combina-tion. In Ubx mutants, expression is derepressed in the abdominal segments, A1 and A2 (Fig. 6D). This is consis-tent with the requirement for Ubx function in both A1 and A2 somatic musculature (Hooper, 1986). In abd-A mutants, there were no consistent changes in the pattern of expression of the 4 kb construct (data not shown) but, in the Ubx-abdA double mutant, the derepression is more dra-matic than that seen in embryos mutant for Ubx alone: high levels of expression are now seen in A3 and, in late embryos, intermediate levels of expression are detected fur-ther posteriorly (Fig. 6E). This indicates that, in the wild type, both Ubx and abd-A repress this construct. The expla-nation for the lack of effect of the abd-A single mutant on the expression pattern is likely to be transregulation between abd-A and Ubx. The abd-A gene normally represses Ubx expression in the posterior abdominal seg-ments and, in abd-A mutants, Ubx expression in the abdomen is derepressed (Struhl and White, 1985). Thus, in the abd-A single mutant, high levels of Ubx expression are likely to repress the 4 kb construct in the posterior abdominal segments. In all the mutants described above, the posterior domain of expression in A9 appeared normal.

To investigate potential activating homeotic loci, we examined the expression of the 4 kb construct in embryos homozygous for an Antp null mutation. In these mutant embryos, expression is strongly reduced in the thoracic seg-ments, T2 and T3, but unaffected in T1, the gnathal region, and in A9 (Fig. 6F). This indicates that Antp is required for the high levels of expression found in T2 and T3 of the wild-type embryo. Consistent with this, the segments T2 and T3 are known to be the principal segments affected by Antp loss-of-function mutations (Wakimoto and Kaufman, 1981) and correspond to the domain of Antp expression in the mesoderm (Boulet and Scott, 1988; Martinez Arias, 1986).

Transcription driven by a 1 kb region encompassing clone 35

In order to examine the regulatory role of sequences more immediately surrounding clone 35, we next tested a 1 kb PstI-SalI fragment in which the 110 bp clone 35 is centrally located (Fig. 5). Like the 4 kb fragment, this 1 kb fragment was placed upstream of the hsp70 minimal promoter in both orientations. Consistent staining, related to the connectin gene expression pattern, was obtained only when the 1 kb element was in the same orientation as the functioning 4 kb element. Expression is confined to a subset of the 4 kb pattern with most labelling in the gnathal region (Fig. 7A). Compared to the 4 kb pattern, the staining observed in T1-T3 and in A9 appears much reduced in intensity relative to the gnathal expression (and is completely absent in several transformant lines). Thus, the 1 kb element lacks some of the sequences required for the high levels of expression in T1-T3 and A9 that are characteris-tic of the 4 kb construct and presumably these sequences reside in the 3 kb remainder fragment.

Fig. 7.

Expression from the 1 kb construct. (A) Wild-type stage 12 embryo, showing gnathal expression (G) in the labial segment, together with weak expression in the thorax, in T1 and T2. (B) Wild-type stage 15 embryo. Labelling is seen anteriorly in the gnathal segments and the arrow indicates the expression domain in the visceral mesoderm. (C) Wild-type stage 16 embryo. Expression occurs anteriorly in gnathal segments and posteriorly in A9. The arrow indicates labelling in the visceral mesoderm at the second midgut constriction. (D) Ubx mutant (Df bxd100) stage 17 embryo. The gnathal labelling is present, but there is no labelling in the visceral mesoderm. The arrow indicates the approximate position corresponding to the second midgut constriction in the wild type.

Fig. 7.

Expression from the 1 kb construct. (A) Wild-type stage 12 embryo, showing gnathal expression (G) in the labial segment, together with weak expression in the thorax, in T1 and T2. (B) Wild-type stage 15 embryo. Labelling is seen anteriorly in the gnathal segments and the arrow indicates the expression domain in the visceral mesoderm. (C) Wild-type stage 16 embryo. Expression occurs anteriorly in gnathal segments and posteriorly in A9. The arrow indicates labelling in the visceral mesoderm at the second midgut constriction. (D) Ubx mutant (Df bxd100) stage 17 embryo. The gnathal labelling is present, but there is no labelling in the visceral mesoderm. The arrow indicates the approximate position corresponding to the second midgut constriction in the wild type.

We were surprised to observe an additional pattern of expression in the visceral mesoderm of the midgut with the 1 kb construct that is not observed with the 4 kb construct. As described above, endogenous connectin RNA is expressed in a segmentally repeated subset of the visceral mesoderm. However, whilst the expression of connectin RNA in the presumptive visceral mesoderm of the midgut is seen transiently from stage 11-13, β-galactosidase activity from the 1 kb construct does not appear until stage 15. The visceral mesoderm expression from the 1 kb con-struct is restricted to the developing second midgut con-striction (Fig. 7B,C) and corresponds to a subset of paraseg-ment 7. This is particularly interesting as Ubx is expressed in PS7 in the visceral mesoderm and Ubx function is required for the formation of the second midgut constric-tion (Bienz et al., 1988; Bienz and Tremml, 1988). The vis-ceral mesoderm expression of the 1 kb construct is also dependent on Ubx as it is abolished in Ubx mutant embryos, whilst the gnathal pattern remains unaffected (Fig. 7D).

Deletion of clone 35 sequences from 4 kb and 1 kb regulatory regions

To test directly the requirement for the clone 35 sequence in the regulation conferred by the 4 kb or the 1 kb ele-ments, constructs containing a precise deletion of this 110 bp region but similar in every other respect were made (called 4 kb-35 and 1 kb-35 constructs).

With the 4 kb-35 construct, the gnathal expression is little affected, but the expression in T1-T3 is weak and variable (Fig. 8A) indicating that the 110 bp region is required for the normal high levels of expression in T1-T3. As described above, comparison between the expression from the 4 kb and 1 kb constructs led to the conclusion that sequences outside the 1 kb region are also required for high levels of thoracic expression. Taken together, these observations imply that multiple sites of control are involved.

Fig. 8.

Expression from 1 kb and 4 kb constructs lacking the clone 35 sequences. (A) 4 kb-35 construct viewed from above. Stage 14 embryo showing strong gnathal labelling but weak and variable expression in the thoracic segments. (B) 1 kb-35 construct, stage 17 embryo. The gnathal expression is present anteriorly, but there is no staining in the visceral mesoderm.

Fig. 8.

Expression from 1 kb and 4 kb constructs lacking the clone 35 sequences. (A) 4 kb-35 construct viewed from above. Stage 14 embryo showing strong gnathal labelling but weak and variable expression in the thoracic segments. (B) 1 kb-35 construct, stage 17 embryo. The gnathal expression is present anteriorly, but there is no staining in the visceral mesoderm.

The 1 kb-35 construct gave an expression pattern identical to the 1 kb construct within the gnathal segments. However, whereas the 1 kb construct gave expression in the visceral mesoderm, staining in this tissue was com-pletely absent in the 1 kb-35 construct (Fig. 8B). As the visceral mesoderm expression from the intact 1 kb element is also absent in a Ubx genetic background, we conclude that clone 35 sequences are required in vivo to mediate Ubx-dependent activation in PS7 of the visceral mesoderm.

In this paper, we have shown that the immunopurified clone 35, which was isolated as a potential in vivo Ubx protein binding site, appears to be an important component of a regulatory element under homeotic gene control. This regulatory element is part of the control sequences regulating the expression of connectin, a gene encoding a cell-surface member of the LRR family.

Connectin

The presence of multiple tandem copies of the LRR motif in the connectin molecule does not immediately assign the molecule to a specific functional class. LRRs were first identified in the leucine-rich alpha2-glycoprotein (Taka-hashi et al., 1985) and have since been found in a diverse set of intracellular, membrane and extracellular proteins including yeast adenylate cyclase (Kataoka et al., 1985), the von Willebrand factor receptor (Titani et al., 1987) and extracellular matrix proteoglycans (Fisher et al., 1989). There is evidence that the LRRs are involved in protein-protein interactions. LRR peptides appear to adopt an amphipathic β-sheet presenting polar and apolar potential interaction surfaces (Gay et al., 1991; Krantz et al., 1991).

Deletion analysis of the LRR region in adenylate cyclase implicates LRRs in complex formation with RAS proteins (Field et al., 1990). Also, several cell surface LRR mole-cules, including the Drosophila molecules chaoptin and Toll protein, share the common property of mediating cell-cell adhesion (Krantz and Zipursky, 1990; Keith and Gay, 1990). Chaoptin shows additional similarity to connectin as it is also linked to the membrane via a GPI-anchor (Krantz and Zipursky, 1990). Mutants in chaoptin disrupt the organ-ization of photoreceptor membranes consistent with a role for chaoptin as a cell adhesion molecule required for normal photoreceptor morphogenesis. Expression of connectin in tissue culture cells reveals that it can also function as a homophilic cell-cell adhesion molecule (Nose et al., 1992; L. Meadows, A. P. G. and R. A. H. W., unpublished results). This supports the connection between cell surface LRR-containing molecules and adhesion, and further impli-cates the LRR sequences themselves in protein-protein interactions.

Regulation of the connectin gene by multiple homeotic loci

Previously, we have demonstrated that connectin RNA expression is regulated in the central nervous system by Ubx and the abdominal homeotic genes (Gould et al., 1990). Here we have examined the control of a 4 kb connectin regulatory element that directs expression predominantly in the somatic mesoderm and not in the ventral nervous system. We have shown that this 4 kb region, located at least 45 kb 3′ of the transcription start site, is sufficient to drive a subset of the connectin expression pattern that is under the control of multiple homeotic genes, including Ubx. Genetic experiments have shown that Antp is the homeotic gene required to activate expression of the 4 kb construct in T2 and T3. However, in Antp mutants, expression remains in gnathal segments, in T1 and in A9. The expression in T1 is likely to be dependent on Sex-combs reduced (Scr), whilst Deformed, labial and pro -boscipedia may be required for the more anterior activa-tion (Mahaffey and Kaufman, 1988).

The 4 kb construct is strongly expressed in the thoracic segments but shows very little expression in abdominal seg-ments. We have shown that Ubx and abd-A are required for this repression in the abdomen. However, the possibility exists that this regulation could be mediated through the influence of Ubx and abd-A on the expression of Antp. Both Ubx and abd-A normally repress Antp expression in the abdominal segments (Hafen et al., 1984). Thus, in Ubx or Ubx abd-A mutants, Antp expression in the abdomen is derepressed and these higher levels of Antp protein might directly enhance expression from the 4 kb construct. How-ever, studies of transformed cuticular phenotypes induced by ectopic Antp expression indicate that Antp is not able to influence cuticle phenotype in regions expressing Ubx or the abdominal homeotic genes (Gibson and Gehring, 1988). This may suggest that Ubx and abd-A exert a dominant direct repression of target genes, blocking their activation by Antp. This phenomenon has been called “phenotypic sup-pression” (Gonzalez-Reyes and Morata, 1990) and it is likely to be important for our understanding of homeotic gene function. The analysis of the molecular basis of phe-notypic suppression has been impeded by the lack of iden-tified target genes. Now, with the identification of specific target gene regulatory elements, such as the connectin 4 kb element, we can hope to clarify this phenomenon.

Deletion of either the clone 35 sequence or of the 3 kb PstI-EcoRI fragment leads to a similar reduction in the intensity of staining in T1-T3. Therefore, it is probable that there are at least two separable elements required for high levels of expression in this region. Their requirement in T2 and T3 indicates that elements within the 3 kb fragment and also within the 110 bp clone 35 both respond to Antp regulation. Another homeotic gene, most probably Scr (Mahaffey and Kaufman, 1988), may activate expression from these elements in T1.

With the 1 kb construct, we observed expression in PS7 of the visceral mesoderm, although this is not seen with the 4 kb construct. There may be two possible explanations for this; either repressing elements are present in the deleted 3 kb region or the visceral mesoderm expression element in the 1 kb construct may not function correctly when placed too far upstream of the hsp 70 minimal promoter. The con -nectin RNA is expressed in repeating subsets of the devel-oping visceral mesoderm including PS7, whereas the 1 kb construct appears to give only the PS7-specific subset of this pattern. However, we first detect expression from this construct only after connectin RNA levels in this region have declined (compare Figs 4B and 7B). Hence, in this particular instance, the 1 kb construct does not faithfully reproduce the timing of onset of connectin expression. Nev-ertheless, the PS7 domain does allow a specific connection to be made between the 110 bp immunopurified fragment and Ubx control as expression is dependent on the presence of both the clone 35 sequence and Ubx function. Given that clone 35 was originally immunopurified using an antibody directed against Ubx proteins and given that it interacts directly with Ubx protein in vitro (Gould et al., 1990), it seems likely that this 110 bp region mediates direct regu-lation by Ubx.

For the 1 kb construct, deletion of the 110 bp fragment has the same effect on expression in the visceral mesoderm as the removal of Ubx gene function. This is what would be expected if the 110 bp sequence identifies a specific Ubx-dependent regulatory site. However, for the 4 kb construct, a Ubx mutation results in increased expression in the abdominal somatic mesoderm whereas deletion of the 110 bp results in reduced thoracic expression. Our interpreta-tion of these different effects is that the deletion of the 110 bp fragment removes, not only sites required for Ubx and abd-A repression, but also the sites required for activation by Antp and Scr.

The DNA sequence of the 110 bp region contains six copies of TAAT, the core homeodomain binding sequence (Ekker et al., 1991; Kissinger et al., 1990). Using muta-genised versions of the 1 kb construct, it should now be possible to assay each of the TAAT cores for a direct and functional interaction with Ubx, abd-A, Antp and Scr. As the 110 bp sequence appears to be recognised by multiple homeotic loci, it will be interesting to determine whether the different homeotic gene products act on separate binding sites or whether they compete for the same sites.

In summary, the immunopurification procedure has successfully led us to a regulatory element under homeotic gene control. We are currently investigating whether other binding sites from the immunopurified DNA library are also associated with elements regulated by homeotic genes. In principle, this method should be generally applicable to searches for in vivo targets of characterised transcription factors or DNA-binding proteins.

Properties of genes regulated by the homeotic loci

Garcia-Bellido proposed that the homeotic genes would exert their effects on morphogenesis by controlling the activities of a group of cytodifferentiation genes. These cytodifferentiation genes would specify a number of cellu-lar characteristics including, mitotic rate, preferential mitotic orientations, cell affinities and final cellular differ-entiation (Garcia-Bellido, 1975). The isolation of connectin as a homeotic target gene fits well with this scheme. Con-nectin is a cell surface molecule capable of mediating cell-cell adhesion and the analysis of its protein distribution in detail reveals that it is expressed in a subset of motoneu-rons and also in their specific target muscle cells both prior to and after the formation of neuromuscular junctions (Nose et al., 1992). Thus it is tempting to speculate that connectin is a cell-adhesion or recognition molecule involved in the development of specific neuromuscular connections. Inter-estingly, recent studies in C. elegans have implicated an Antp homologue mab-5, in the control of cell surface mol-ecules that affect region-specific migratory cell behaviours in the nervous system (Salser and Kenyon, 1992). It has also been suggested that the neural cell adhesion molecule (N-CAM) is likely to be regulated by mouse Hox-gene products (Jones et al., 1992).

Our observation that connectin is likely to be regulated directly by homeotic genes is interesting with regard to the overall command chain of developmental gene controls.

The complex pattern of homeotic gene expression is largely the result of a hierarchy of regulation by transcription fac-tors extending from bicoid, through gap genes, pair-rule genes and segment polarity genes (reviewed Ingham, 1988). It had always been possible that several more layers of tran-scriptional control might necessarily lie between the homeotic selector genes and their cytodifferentiation tar-gets. Here, with connectin, this appears not to be the case and suggests that at least some genes involved in the common processes of morphogenesis are directly under the transcriptional control of homeotic genes.

We thank Nick Brown and Sarah Bray for guidance with cDNA library screening, Shashidhara L. S. for the RNA-probe connectin in situ hybridizations and Peter Lawrence for help with the trans-formations. We are grateful to Sarah Bray, Lisa Meadows and David Strutt for their comments on the manuscript. This work was supported by the MRC and the Wellcome Trust.

Akam
,
M.
(
1987
).
The molecular basis for metameric pattern in the Drosophila embryo
.
Development
101
,
1
22
.
Akam
,
M.
and
Martinez-Arias
,
A.
(
1985
).
The distribution of Ultrabithorax transcripts in Drosophila embryos
.
EMBO J
.
4
,
1689
1700
.
Andrew
,
D. J.
and
Scott
,
M. P.
(
1992
).
Downstream of the homeotic genes
.
New Biologist
4
,
5
15
.
Ashburner
,
M.
(
1989a
).
Drosophila
.
A Laboratory Handbook
.
New York
:
Cold Spring Harbor Laboratory Press
.
Ashburner
,
M.
(
1989b
).
Drosophila
.
A Laboratory Manual
.
New York
:
Cold Spring Harbor Laboratory Press
.
Beachy
,
P. A.
,
Krasnow
,
M. A.
,
Gavis
,
E. R.
and
Hogness
,
D. M.
(
1988
).
An Ultrabithorax protein binds near its own and Antennapedia P1 promoters
.
Cell
55
,
1069
1081
.
Bienz
,
M.
,
Saari
,
G.
,
Tremml
,
G.
,
Muller
,
J.
,
Zust
,
B.
and
Lawrence
,
P. A.
(
1988
).
Differential regulation of Ultrabithorax in two germ layers of Drosophila
.
Cell
53
,
567
576
.
Bienz
,
M.
and
Tremml
,
G.
(
1988
).
Domain of Ultrabithorax expression in Drosophila visceral mesoderm from autoregulation and exclusion
.
Nature
333
,
576
578
.
Boulet
,
A.
and
Scott
,
M.
(
1988
).
Control elements of the P2 promoter of the Antennapedia gene
.
Genes Dev
.
2
,
1600
1614
.
Brown
,
N. H.
and
Kafatos
,
F. C.
(
1988
).
Functional cDNA libraries from Drosophila embryos
.
J. Mol. Biol
.
203
,
425
437
.
Campos-Ortega
,
J. A.
and
Hartenstein
,
V.
(
1985
)
The Embryonic Development of Drosophila melanogaster
.
New York
:
Springer-Verlag
.
Cavener
,
D. R.
(
1987
).
Comparison of the consensus sequence flanking translational start sites in Drosophila and vertebrates
.
Nucl. Acids Res
1987
,
1353
1361
.
Devereux
,
J.
,
Haeberli
,
P.
and
Smithies
,
O.
(
1984
).
A comprehensive set of sequence analysis programs for the VAX
.
Nucl. Acids. Res
.
12
,
387
395
.
Ekker
,
S. C.
,
Young
,
K. E.
,
Von Kessler
,
D. P.
and
Beachy
,
P. A.
(
1991
).
Optimal DNA sequence recognition by the Ultrabithorax homeodomain of Drosophila
.
EMBO J
.
10
,
1179
1186
.
Ferguson
,
M. A. F.
and
Williams
,
A. F.
(
1988
).
Cell-surface anchoring of proteins via glycosyl-phosphatidylinositol structures
.
Ann. Rev. Biochem
.
57
,
285
320
.
Field
,
J.
,
Xu
,
H.-P.
,
Michaeli
,
T.
,
Ballester
,
R.
,
Sass
,
P.
and
Wigler
,
M.
(
1990
).
Mutations of the adenylyl cyclase gene that block RAS function in Saccharomyces cerevisiae
.
Science
247
,
464
467
.
Fisher
,
L. W.
,
Termine
,
J. D.
and
Young
,
M. F.
(
1989
).
Deduced protein sequence of bone small proteoglycan I (Biglycan) shows homology with proteoglycan II (Decorin) and several nonconnective tissue proteins in a variety of species
.
J. Biol. Chem
264
,
4571
4576
.
Garcia-Bellido
,
A. G.
(
1975
).
Genetic control of wing disc development in Drosophila
.
In Cell Patterning, Ciba Foundation Symposium
29
(ed.
S.
Brenner
), pp.
161
178
.
New York
:
Associated Scientific Publishers
.
Gay
,
N. J.
,
Packman
,
L. C.
,
Weldon
,
M. A.
and
Barna
,
J. C. J.
(
1991
).
A leucine rich repeat peptide derived from the Drosophila Toll receptor forms extended filaments with a ß-sheet structure
.
FEBS Lett
.
291
,
87
91
.
Gibson
,
G.
and
Gehring
,
W. J.
(
1988
).
Head and thoracic transformations caused by ectopic expression of Antennapedia during Drosophila development
.
Development
102
,
657
675
.
Gonzalez-Reyes
,
A.
and
Morata
,
G.
(
1990
).
The developmental effect of overexpressing a Ubx product in Drosophila embryos is dependent on its interactions with other homeotic products
.
Cell
61
,
515
522
.
Gould
,
A. P.
,
Brookman
,
J. J.
,
Strutt
,
D. I.
and
White
,
R. A. H.
(
1990
).
Targets of homeotic gene control in Drosophila
.
Nature
348
,
308
312
.
Hafen
,
E.
,
Levine
,
M.
and
Gehring
,
W. J.
(
1984
).
Regulation of Antennapedia transcript distribution by the bithorax complex of Drosophila
.
Nature
307
,
287
289
.
Hashimoto
,
C.
,
Hudson
,
K. L.
and
Anderson
,
K. V.
(
1988
).
The Toll gene of Drosophila, required for dorso-ventral polarity, appears to encode a transmembrane protein
.
Cell
52
,
269
279
.
Hickey
,
M. J.
,
Williams
,
S. A.
and
Roth
,
G. J.
(
1989
).
Human platelet glycoprotein IX : An adhesive prototype of leucine-rich glycoproteins with flank-center-flank structures
.
Proc. Natl. Acad. Sci. USA
86
,
6773
6777
.
Hiromi
,
Y.
and
Gehring
,
W. J.
(
1987
).
Regulation and function of the Drosophila segmentation gene fushi tarazu
.
Cell
50
,
963
974
.
Hooper
,
J. E.
(
1986
).
Homeotic gene function in the muscles of Drosophila larvae
.
EMBO J
.
5
,
2321
2329
.
Immergluck
,
K.
,
Lawrence
,
P. A.
and
Bienz
,
M.
(
1990
).
Induction across germ layers in Drosophila mediated by a genetic cascade
.
Cell
62
,
261
268
.
Ingham
,
P. W.
(
1988
).
The molecular genetics of embryonic pattern formation in Drosophila
.
Nature
335
,
25
34
.
Jones
,
F. S.
,
Prediger
,
E. A.
,
Bittner
,
D. A.
,
De Robertis
,
E. M.
and
Edelman
,
G. M.
(
1992
).
Cell adhesion molecules as targets for Hox genes : Neural cell adhesion molecule promoter activity is modulated by cotransfection with Hox-2.5 and -2.4
.
Proc. Natl. Acad. Sci USA
89
,
2086
2090
.
Kataoka
,
T.
,
Broek
,
D.
and
Wigler
,
M.
(
1985
).
DNA sequence and characterization of the S. cerevisiae gene encoding adenylate cyclase
.
Cell
43
,
493
505
.
Keith
,
F. J.
and
Gay
,
N. J.
(
1990
).
The Drosophila membrane receptor Toll can function to promote cellular adhesion
.
EMBO J
.
9
,
4299
4306
.
Kennison
,
J. A.
and
Tamkun
,
J. W.
(
1988
).
Dosage-dependent modifiers of Polycomb and Antennapedia mutations in Drosophila
.
Proc. Natl. Acad. Sci. USA
85
,
8136
8140
.
Kissinger
,
C. R.
,
Liu
,
B.
,
Martin-Blanco
,
E.
,
Kornberg
,
T.
and
Pabo
,
C. O.
(
1990
).
Crystal structure of an engrailed homeodomain-DNA complex at 2.8 A resolution: a framework for understanding homeodomain-DNA interactions
.
Cell
63
,
579
590
.
Krantz
,
D. D.
,
Zidovetski
,
R.
,
Kagan
,
B. L.
and
Zipursky
,
S. L.
(
1991
).
Amphipathic β structure of a leucine-rich repeat peptide
.
J. Biol. Chem
.
266
,
16801
16807
.
Krantz
,
D. E.
and
Zipursky
,
S. L.
(
1990
).
Drosophila chaoptin, a member of the leucine-rich repeat family, is a photoreceptor cell-specific adhesion molecule
.
EMBO J
.
9
,
1969
1977
.
Krasnow
,
M. A.
,
Saffman
,
E. E.
,
Kornfeld
,
K.
and
Hogness
,
D. S.
(
1989
).
Transcriptional activation and repression by Ultrabithorax proteins in cultured Drosophila cells
.
Cell
57
,
1031
1043
.
Kyte
,
J.
and
Doolittle
,
R.
(
1982
).
A simple method for displaying the hydrophobic character of a protein
.
J. Mol. Biol
.
157
,
105
132
.
Lewis
,
E. B.
(
1963
).
Genes and developmental pathways
.
Am. Zool
.
3
,
33
56
.
Lewis
,
E. B.
(
1978
).
A gene complex controlling segmentation in Drosophila
.
Nature
276
,
565
570
.
Lindsley
,
D.
and
Zimm
,
G.
(
1985
)
The genome of Drosophila melanogaster. Part 1: Genes A-K
.
Drosophila Inform. Service
62
.
Lindsley
,
D.
and
Zimm
,
G.
(
1990
)
The genome of Drosophila melanogaster. Part 4: Genes L-Z, balancers, transposable elements
.
Drosophila Inform. Service
68
.
Lopez
,
J.
,
Chung
,
D.
,
Fugikawa
,
K.
,
Hagen
,
F.
and
Papayannapoulou
,
T.
(
1987
).
Cloning of the α-chain of glycoprotein 1b: a transmembrane protein with homology to leucine-rich α2-glycoprotein
.
Proc. Natl. Acad. Sci. USA
84
,
5615
5619
.
Mahaffey
,
J. W.
and
Kaufman
,
T. C.
(
1988
).
The homeotic genes of the Antennapedia complex and the Bithorax complex of Drosophila
.
In Developmental Genetics of Higher Organisms : A Primer in Developmental Biology
(ed.
G. M.
Malacinski
), pp
329
360
.
New York
:
Macmillan
.
Martinez-Arias
,
A.
(
1986
).
The Antennapedia gene is required and expressed in parasegments 4 and 5 of the Drosophila embryo
.
EMBO J
.
5
,
135
141
.
Muller
,
M.
,
Affolter
,
M.
,
Leupin
,
W.
,
Otting
,
G.
,
Wuthrich
,
K.
and
Gehring
,
W.
(
1988
).
Isolation and sequence specific DNA binding of the Antennapedia homeodomain
.
EMBO J
7
,
4299
4304
.
Nose
,
A.
,
Mahajan
,
V. B.
, and
Goodman
,
C. S.
(
1992
)
Connectin: A Homophilic cell adhesion molecule expressed on a subset of muscles and the motoneurons that innervate them in Drosophila
.
Cell
70
,
553
567
.
Panganiban
,
G. E. F.
,
Reuter
,
R.
,
Scott
,
M. P.
and
Hoffman
,
F. M.
(
1990
).
A Drosophila growth factor homolog, decapentaplegic, regulates homeotic gene expression within and across germ layers during midgut morphogenesis
.
Development
110
,
1041
1050
.
Reinke
,
R.
,
Krantz
,
D. E.
,
Yen
,
D.
and
Zipursky
,
S. L.
(
1988
).
Chaoptin, a cell surface glycoprotein required for Drosophila cell morphogenesis, contains a repeat motif found in yeast and human
.
Cell
52
,
291
301
.
Reuter
,
R.
,
Panganiban
,
G. E. F.
,
Hoffman
,
F. M.
and
Scott
,
M. P.
(
1990
).
Homeotic genes regulate the spatial expression of putative growth factors in the visceral mesoderm of Drosophila embryos
.
Development
110
,
1031
1040
.
Rothberg
,
J. M.
,
Jacobs
,
J. R.
,
Goodman
,
C. S.
and
Artavanis-Tsakonas
,
S.
(
1990
).
slit: an extracellular protein necessary for the development of midline glia and commissural axon pathways contains both EGF and LRR domains
.
Genes Dev
.
4
,
2169
2187
.
Salser
,
S. J.
and
Kenyon
,
C.
(
1992
).
Activation of a C. elegans Antennapedia homologue in migrating cells controls their direction of migration
.
Nature
355
,
255
258
.
Sambrook
,
J.
,
Fritsch
,
E. F.
and
Maniatis
,
T.
(
1989
).
Molecular Cloning -A Laboratory Manual. Second edition
.
New York
:
Cold Spring Harbor Laboratory Press
.
Sanchez-Herrero
,
E.
,
Vernos
,
I.
,
Marco
,
R.
and
Morata
,
G.
(
1985
).
Genetic organization of Drosophila bithorax complex
.
Nature
313
,
108
113
.
Sanger
,
F.
,
Nicklen
,
S.
and
Coulson
,
A. R.
(
1977
).
DNA sequencing with chain-terminating inhibitors
.
Proc. Natl. Acad. Sci.USA
74
,
5463
5467
.
Schneider
,
D. S.
,
Hudson
,
K. L.
,
Lin
,
T.-Y.
and
Anderson
,
K. V.
(
1991
).
Dominant and recessive mutations define functional domains of Toll, a transmembrane protein required for dorso-ventral polarity in the Drosophila embryo
.
Genes Dev
.
5
,
797
807
.
Schneider
,
R.
and
Schweiger
,
M.
(
1991
).
A novel modular mosaic of cell adhesion motifs in the extracellular domains of the neurogenic trk and trkB tyrosine kinase receptors
.
Oncogene
6
,
1807
1811
.
Spradling
,
A. C.
(
1986
).
P-element mediated transformation
.
In Drosophila; A Practical Approach
(ed.
D. B.
Roberts
), pp
175
197
.
Washington
:
IRL
.
St. Johnston
,
R. D.
and
Gelbart
,
W. M.
(
1987
).
Decapentaplegic transcripts are localized along the dorso-ventral axis of the Drosophila embryo
.
EMBO J
.
6
,
2785
2791
.
Staden
,
R.
(
1982
).
An interactive graphics program for comparing and aligning nucleic acid and amino acid sequences
.
Nucl. Acids Res
.
10
,
2951
2961
.
Struhl
,
G.
and
White
,
R. A. H.
(
1985
).
Regulation of the Ultrabithorax gene of Drosophila by other bithorax complex genes
.
Cell
43
,
507
519
.
Takahashi
,
N.
,
Takahashi
,
Y.
and
Putnam
,
F. W.
(
1985
).
Periodicity of leucine and tandem repetition of a 24-amino acid segment in the primary structure of leucine-rich a2-glycoprotein of human serum
.
Proc. Natl. Acad. Sci. USA
82
,
1906
1910
.
Tautz
,
D.
and
Pfeifle
,
C.
(
1989
).
A non-radioactive method for the localisation of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback
.
Chromosoma
98
,
81
85
.
Thali
,
M.
,
Muller
,
M.
,
Delorenzi
,
M.
,
Matthias
,
P.
and
Bienz
,
M.
(
1988
).
Drosophila homeotic genes encode transcriptional activators similar to mammalian OTF-2
.
Nature
336
,
598
601
.
Titani
,
K.
,
Takio
,
K.
,
Handa
,
M.
and
Ruggeri
,
Z. M.
(
1987
).
Amino-acid sequence of von-Willebrand factor-binding domain of platelet membrane glycoprotein 1b
.
Proc. Natl. Acad. Sci. USA
84
,
5610
5614
.
von Heijne
,
G.
(
1986
).
A new method for predicting signal sequence cleavage sites
.
Nucl. Acids Res
.
14
,
4683
4690
.
Wagner-Bernholz
,
J. T.
,
Wilson
,
C.
,
Gibson
,
G.
,
Schuh
,
R.
and
Gehring
,
W. J.
(
1991
).
Identification of target genes of the homeotic gene Antennapedia by enhancer detection
.
Genes. Dev
.
5
,
2467
2480
.
Wakimoto
,
B. T.
and
Kaufman
,
T. C.
(
1981
).
Analysis of larval segmentation in lethal genotypes associated with the Antennapedia gene complex in Drosophila melanogaster
.
Dev. Biol
.
81
,
51
64
.
White
,
R. A. H.
and
Wilcox
,
M.
(
1985
).
The distribution of Ultrabithorax proteins in Drosophila
.
EMBO J
.
4
,
2035
2043
.
Winslow
,
G. M.
,
Hayashi
,
S.
,
Krasnow
,
M.
,
Hogness
,
D. S.
and
Scott
,
M. P.
(
1989
).
Transcriptional activation by the Antennapedia and fushi-tarazu proteins in cultured Drosophila cells
.
Cell
57
,
1017
1030