This paper reports the characterization of two immunologically related proteins that may be involved in cell adhesion during Drosophila development. These proteins, laminin chain A and a 240K component, share the epitope recognized by monoclonal antibody RD3 (Mab RD3). The two antigens show different developmental expression profiles. Laminin is detected only from 6 to 8 h of development onwards; its concentration increases during embryogenesis to reach steady-state value in larvae, pupae and adult flies. By contrast, the 240K antigen, not found in oocytes, is present before blastoderm stages; its concentration increases during gastrulation, decreases at the end of organogenesis and the antigen is no longer detected in third instar larvae. Light and electron microscope immunolocalization in imaginal discs indicates that laminin is distributed apically in the lumen and basally in the basal membrane that surrounds the nonevaginated disc. During morphogenesis laminin is detected at the basal side of the evaginating part of the disc epithelium. Immunolocalization on paraffin sections of early embryos suggests that the 240K antigen is related to (1) cell formation and polarization in association with cytoskeleton components, (2) establishment of cell—extracellular substratum interactions during the blastoderm cell sheet organization and (3) basement membrane deposition during embryonic germ cell layer segregation. This 240K protein is poorly or not glycosylated, is resistant to chondroitinase ABC and collagenase and appears therefore as a new extracellular component that might be specifically involved in early processes of morphogenesis.

The extracellular matrix (ECM) is thought to be important in the regulation of polarization, adhesion, migration and differentiation of cells during embryogenesis. The role and the nature of ECM components are well documented in vertebrates and amphibians. Most of these components are detected early during the embryogenesis as shown in sea urchin (McCarthy & Burger, 1987; Wessel & McClay, 1987), Pleurodeles (Darribère et al. 1986) or mouse (Dziadek & Timpl, 1985). In Drosophila, despite the many advantages provided by developmental genetics in this organism, little is known about these early cell processes and it is only recently that the major ECM components have been characterized, essentially by comparison with vertebrate components. Type IV collagen, entactin, laminin and proteoglycan have been found in the culture medium of established Drosophila cell Unes (Fessler et al. 1984). Amino-acid sequence data of type TV collagen (Blumberg et al. 1987 ; Cecchini et al. 1987) and of laminin Bl chain (Montell & Goodman, 1988) as well as structural analysis of laminin (Fessler et al. 1987) have shown that these Drosophila proteins are very similar to their vertebrate counterparts. Recently, the existence of fibronectin (Gratecos et al. 1988) has also been proved by using an heterologous probe. Moreover, Leptin et al. (1987) have shown that the positionspecific antigens are related to vertebrate integrins, a family of cell-surface glycoproteins that includes cellsurface receptors for laminin and fibronectin.

In contrast to other organisms, none of the ECM components so far described in Drosophila embryos have been detected before the time of organogenesis (Fessler et al. 1984; Campbell et al. 1987), with the exception of fibronectin, which is present before the blastoderm stage (Gratecos et al. 1988). One can ask whether or not specific extracellular components are needed during early Drosophila embryogenesis.

In order to identify protein components involved in early morphogenetic events and study the extracellular substratum formation in early embryos, we have produced monoclonal antibodies (Mab). Here we describe a Mab that recognizes an epitope shared by two distinct proteins which show different developmental accumulation patterns. The first, present in preblastoderm embryos, is related to ECM deposition during embryogenesis and disappears during larval stages. Crossimmunoprecipitation assays, SDS-PAGE analysis and immunolocalization on isolated imaginal discs indicate that the second protein is laminin.

Antibodies

The hybridoma cell line producing Mab RD3 was obtained from a fusion of X63 Ag 8,653 myeloma cells with spleen cells from a Balb C mouse immunized with protein extracts from 0−15 h embryos. Mab RD3 was selected through immunofluorescence assays on cryostat sections of embryos and the corresponding cell line cloned twice by limiting dilution. Mab RD3 has been shown to belong to IgM class. IgG purified from rabbit antisera directed against murine laminin, either the native molecule or the reduced and alkylated chains, isolated from EHS tumor was a gift of Dr J. C. Lissitsky.

Preparation of protein samples and immunoprecipitation procedure

Embryos and adult flies (strain Oregon R) were collected from mass population cages, larvae and pupae from individual bottles. Homogenization was performed in buffer TE (20 HIm-Tris, 1 mm-EDTA, pH 7·5) supplied with a cocktail of protease inhibitors (IIUM-PMSF, 0·2T.I.U. ml-1 aprotinin and 10 μg ml-1 antipain, leupeptin and pepstatin). The homogenate was centrifuged 10 min at 5000g and the supernatant directly analysed in SDS-PAGE or used for immunoprecipitation assays.

Cell membranes were prepared according to Venkatesh et al. (1980) and Piovant (Manuscript in preparation). Briefly, a homogenate of unstaged embryos (10 mm-Tris, 0·5 mm-CaCl2, 0·25 m-sucrose, 0·2% aprotinin, 0·5mm-PMSF, pH 7-5) was centrifuged 10 min at 25000g to remove debris and then for 30 min at 100000g to pellet a crude cell membrane fraction. Cell membranes were purified by centrifugation on a discontinuous sucrose gradient for 18 h at 35000 revs min-1 in a SW 41 Beckman rotor. Samples from the various fractions, i.e. whole homogenate, debris, crude and purified cell membranes and 100000g supernatant, were then treated for SDS-PAGE.

For immunoprecipitations, the protein samples, diluted in lysis buffer (20 mm-Tris, 0·15m-NaCl, Imm-MgCb, Imm-CaC12, 0·5% NPU), pH7·5) supplied with the cocktail of protease inhibitors, were preincubated for 1h at 4°C with protein A-Sepharose beads to reduce unspecific binding. The depleted samples were then incubated overnight at 4°C either with IgG anti-mouse laminin or with Mab RD3 coupled to protein A-Sepharose beads. The coupling of Mab RD3 required an additional intermediate step, i.e. the coupling of rabbit anti-mouse IgM to protein A beads. After a centrifugation at 5000g for 5 min, the Sepharose beads were washed three times each with buffer I (10 mm-Tris, 0·5M-LíC1, 01 % SDS, 2% NP-40, pH 7·4) and with buffer II (50 mm-Tris, pH 8·0). The beads were then collected by centrifugation and the pellet prepared for SDS-PAGE.

Western blotting procedure

Protein samples were fractionated by SDS-PAGE on 3−10 % linear gradient slab gels. Proteins were electroblotted according to Towbin et al. (1979) and the remaining binding sites on nitrocellulose were saturated following the ‘blotto’ procedure. Fab sheep anti-mouse IgG (H+L) (Biosys, Pasteur, Paris) or protein A labelled with 125I according to Greenwood et al. (1963) were used as the secondary antibody to probe, after autoradiography (Kodak, X Omat film), the binding of the primary antibodies.

Immunohistochemistry

Permeabilization, fixation and devitellinization of embryos were carried out according to Mitchison & Sedat (1983). Rehydrated embryos were equilibrated in hybridoma culture medium (DMEM, 10% FCS) containing 0·1 % Triton X-100. Embryos were incubated overnight at 4 °C with Mab RD3, washed and incubated with rabbit anti-mouse IgM coupled to rhodamine (Jackson, Arondale, USA) for two hours at room temperature. After an extensive wash, embryos were mounted in 80% glycerol in PBS, and viewed with a Zeiss photomicroscope equ ipped with epifluorescence.

Frozen sections of about 5 μm were performed on OCTembedded embryos or imaginal discs manually isolated from third instar larvaé and white pupae. Sections were then fixed in 4% formaldehyde.

Paraffin sections of about 5 μm were obtained from embryos fixed.in 4 %. formaldehyde buffered in PBS and dehydrated through an ethanol series before embedding in paraffin at 60°C. After clearing in xylene and progressive rehydration, the sections were incubated with antibodies as described above.

Ultrathin sections of imaginal discs prefixed in 2 % formaldehyde and placed on nickel grids were first incubated with hybridoma culture medium (DMEM, 10 % FCS) and then transferred for 1 h to a drop of culture supernatant of RD3 hybridoma cell line. After washing in PBS they were labelled for 1 h with rabbit anti-mouse IgM complexed with colloidal gold (size 18 nm, Janssen life sciences). Sections were contrasted with uranyl acetate and viewed in a Philips EM 200 electron microscope.

Identification of RD3 antigens and comparison with Drosophila and mouse laminin

The different electrophoretic mobilities of RD3 antigens, Drosophila and mouse laminins are shown on the Western blots presented in Fig. 1A. The IgG used to reveal the laminins were purified from an antiserum raised against native mouse laminin isolated from EHS tumor. These antibodies (kindly donated by Dr J. C. Lissitsky) reacted..very strongly with the three chains of mouse Jaminin (Fig. 1A, lanes 1 and 2) and also displayed a weak cross-reactivity with Drosophila laminin (Fig. 1A, lanes 5 and 6). The nonreduced proteins comigrate as a single band with a relative molecular mass (Afr) of about 900 ×103. After reduction the apparent molecular weights of murine laminin chains were 400, 220 and 210×103 and those of Drosophila 400, 215 and 185 × 103 (the 185x 103 chain is more clearly seen in Fig. IB, lane 1); these values are in agreement with the values recently published by Fessler et al. (1987) for Drosophila laminin chains.

Fig. 1.

Comparison of RD3 antigens with mouse and Drosophila laminins. Immunoblot analysis of Mab RD3 and of purified anti-mouse laminin IgG from embryo (0·15 h) extracts or purified mouse laminin (A) and from immunoprecipitated antigens (B). Proteins were fractionated by SDS-PAGE on 3-10 % linear gradient gels and visualized on nitrocellulose with anti-mouse laminin IgG and 125I-protein A or with Mab RD3 and 125I-F(ab’)z sheep anti-mouse IgG (H+L). (A) Mouse laminin (0·4pg) revealed with anti-mouse laminin IgG (lanes 1, 2); proteins (7·5 ftg) from embryos (0·15 h) revealed with Mab RD3 (lanes 3, 4) or with anti-mouse laminin IgG (lanes 5, 6). Proteins were treated in sample electrophoresis buffer in the absence (lanes 1, 3, 5) or in the presence (lanes 2, 4, 6) of dithiothreitol. (B) Antigens from 0·15 h embryos immunoprecipitated by Mab RD3 and visualized with antimouse laminin IgG (lane 1) or with Mab RD3 (lane 2); the blot shown in lane 1 was further postrevealed using Mab RD3 (lane 3). Immunoprecipitates obtained with antimouse laminin IgG and visualized with Mab RD3 (lane 4) or with anti-mouse laminin IgG (lane 5). Molecular weight standards are shown at the left.

Fig. 1.

Comparison of RD3 antigens with mouse and Drosophila laminins. Immunoblot analysis of Mab RD3 and of purified anti-mouse laminin IgG from embryo (0·15 h) extracts or purified mouse laminin (A) and from immunoprecipitated antigens (B). Proteins were fractionated by SDS-PAGE on 3-10 % linear gradient gels and visualized on nitrocellulose with anti-mouse laminin IgG and 125I-protein A or with Mab RD3 and 125I-F(ab’)z sheep anti-mouse IgG (H+L). (A) Mouse laminin (0·4pg) revealed with anti-mouse laminin IgG (lanes 1, 2); proteins (7·5 ftg) from embryos (0·15 h) revealed with Mab RD3 (lanes 3, 4) or with anti-mouse laminin IgG (lanes 5, 6). Proteins were treated in sample electrophoresis buffer in the absence (lanes 1, 3, 5) or in the presence (lanes 2, 4, 6) of dithiothreitol. (B) Antigens from 0·15 h embryos immunoprecipitated by Mab RD3 and visualized with antimouse laminin IgG (lane 1) or with Mab RD3 (lane 2); the blot shown in lane 1 was further postrevealed using Mab RD3 (lane 3). Immunoprecipitates obtained with antimouse laminin IgG and visualized with Mab RD3 (lane 4) or with anti-mouse laminin IgG (lane 5). Molecular weight standards are shown at the left.

RD3 antibody recognizes at least two proteins present in the lysate of Drosophila embryo (Fig. 1A, lanes 3 and 4). The first has a relative molecular mass of 240×103 and does not change its electrophoretic mobility upon reduction. The other comigrates with native laminin (900×103) under nonreducing conditions and with the A chain (400X103) under reducing conditions.

To obtain more precise information about a possible relationship between laminin and RD3 antigens, immunoprecipitations of a lysate of dechorionated Drosophila embryos have been performed using Mab RD3 and anti-murine laminin IgG. The immunoprecipitates were loaded on SDS-PAGE gels, electroblotted and stained as shown in Fig. 1B. The three chains of laminin are detected by anti-mouse laminin IgG in the immunoprecipitate obtained with Mab RD3 (Fig. 1B, lane 1). This indicates that the Mab precipitates the whole native laminin molecule. The 240K RD3 antigen (the abbreviation K is used to denote size of antigen) migrates slightly slower than laminin chains B, and the 400K RD3 antigen comigrates with laminin chain A (lane 2). This is supported by comparing the results reported in lane 1 and lane 3 of Fig. 1B: a blot of an immunoprecipitate obtained with Mab RD3 was labelled in a first step with IgG anti-mouse laminin and 125I-protein A (lane 1) and in a second step with Mab RD3 and 125I-Fab fragments (lane 3); an additional band corresponding to the 240K antigen appears in lane 3.

IgG against mouse laminin immunoprecipitates the three chains of Drosophila laminin (Fig. 1B, lane 5). In this immunoprecipitate, Mab RD3 recognizes only the band at 400 ×103 (lane 4). Taken together with the previous result, this latter observation confirms that the 400K RD3 antigen corresponds to the laminin chain A.

As a conclusion, these data allow us to assume that the 400K RD3 antigen corresponds to the chain A of laminin. As will be discussed in the following paragraphs, the developmental expression of this antigen and its localization in imaginal discs agree with this conclusion. The 240K RD3 antigen shares a common epitope with chain A but cannot be related to laminin chain B.

Developmental profile of RD3 antigens

RD3 antigens are differentially expressed during Drosophila development. The Western blots presented in Fig. 2 have been performed by running the same amount of reduced and alkylated proteins in each lane. Laminin A chain is not detected before 6-8 h of development. Its concentration rises rapidly and a maximum is reached at around 10 h of development. The concentration then remains roughly constant during embryogenesis, larval, pupal and adult lives. These data, particularly the time of the first detection, are consistent with the observations reported bv Fessler et al. (1984, 1987).

Fig. 2.

Developmental expression of RD3 antigens. The same amount (7·5 μg) of reduced and alkylated proteins isolated from 0·2 h embryos (lane 1), 3·6 h (lane 2), 6·8 h (lane 3), 10·16h (lane 4), 16·22h (lane 5), first instar larvae (lane 6), third instar larvae (lane 7), pupae (lane 8) and adult flies (lane 9), were fractionated by SDS-PAGE on 3−10% linear gradient gels, blotted on nitrocellulose and probed with Mab RD3. Molecular weight standards are shown at the left.

Fig. 2.

Developmental expression of RD3 antigens. The same amount (7·5 μg) of reduced and alkylated proteins isolated from 0·2 h embryos (lane 1), 3·6 h (lane 2), 6·8 h (lane 3), 10·16h (lane 4), 16·22h (lane 5), first instar larvae (lane 6), third instar larvae (lane 7), pupae (lane 8) and adult flies (lane 9), were fractionated by SDS-PAGE on 3−10% linear gradient gels, blotted on nitrocellulose and probed with Mab RD3. Molecular weight standards are shown at the left.

The developmental profile of the 240K RD3 antigen stands in contrast with that of laminin. A small, but significant, amount of antigen is present before the blastoderm stage. The protein amount increases notice-20 V. Garzino, H. Berenger and J. Pradel ably during gastrulation, the stage at which transcription of the zygotic genome becomes activated (Zalokar, 1976). The antigen concentration reaches a maximum at around 10 h of development and then begins to decrease. Only a small trace can be detected in the first instar larvae and the protein is then definitely missing in the subsequent developmental steps.

The two RD3 antigens, laminin A chain and the 240K antigen, are simultaneously present during a period of developmental time only, from 6-8 h to the end of embryogenesis. Before the onset of the organogenesis, i.e. during preblastoderm, blastoderm, gastrula and germ band extension steps, only the 240K antigen is detected in embryos, whereas laminin alone is recognized in larvae, pupae and adults.

Distribution of laminin in imaginal discs

We undertook this study to get an additional evidence that the 400K RD3 antigen is a component of ECM, and to determine the accurate localization of laminin in imaginal discs. We have checked that the RD 3 antibody only reacts with the laminin A chain and that the 240K antigen is completely absent on Western blots raised with manually isolated imaginal discs (not shown).

The ECM organization in Drosophila is relatively well documented in late embryo and in larval imaginal discs. In mature third instar larvae, each disc is organized as a single undifferentiated epithelium, folded from an invagination of the epidermis. Thus, the outside of the disc corresponds to the basal surface of epidermis, and the disc is surrounded by a basal lamina. ECM components also accumulate apically as fibrous material in the lumen of the disc (Fristrom & Rickoll, 1982; Brower et al, 1987).

Fig. 3 depicts the immunofluorescence staining observed on sagittal sections of isolated wing discs, from third instar larvae and from white prepupae, i.e. 4h after pupariatión. Most of the antigen is concentrated in the basement membrane that surrounds the unevaginated wing disc (Fig. 3A,B). The basal fluorescence is not localized to particular regions but is found along the entire basal lamina, including the squamous peripodial membrane epithelium. A weaker labelling is observed associated with the apical surface of the epithelium facing the lumen and also with adepithelial cells. As development proceeds, the disc evaginates and reeverts through the stalk, whereas the peripodial membrane contracts and become greatly stretched but does not rupture (Milner et al. 1984). Laminin then accumulates particularly at the basal side of the elongating appendage whereas the fluorescence around the disc progressively vanishes together with the peripodial membrane dispersion. Fig. 3D,F show the antigen distribution in a partly evaginated wing disc. In the region of the wing pouch, when the évagination is initiated, laminin is detected at the basal side of the epithelium, now facing the interior of the disc. A strong fluorescence is also observed in the basal lamina lying under adepithelial cells. At the outside of the disc, the antigen is found adhering to the remaining peripodial membrane.

Fig. 3.

Immunolocalization of laminin in imaginal discs. Phase-contrast micrographs (A,C,E) and laminin localization (B,D,F) on sagittal sections of wing discs isolated from mid third instar larvae. (A,B) and from white pupae (C,D,E,F). Laminin is found in the basement membrane that surrounds non-evaginated discs; a faint fluorescence is observed around adepithelial cells and in the lumen of the disc (arrows in B). In evaginating discs, laminin localizes at the basal surface of the elongating wing pouch and in the basal lamina under adepithelial cells (arrows in D,F). Residual peripodial membrane at the outside of the disc is also stained, ac, adepithelial cells; WP, wing pouch; N, presumptive notum. Bars, 25 μm.

Fig. 3.

Immunolocalization of laminin in imaginal discs. Phase-contrast micrographs (A,C,E) and laminin localization (B,D,F) on sagittal sections of wing discs isolated from mid third instar larvae. (A,B) and from white pupae (C,D,E,F). Laminin is found in the basement membrane that surrounds non-evaginated discs; a faint fluorescence is observed around adepithelial cells and in the lumen of the disc (arrows in B). In evaginating discs, laminin localizes at the basal surface of the elongating wing pouch and in the basal lamina under adepithelial cells (arrows in D,F). Residual peripodial membrane at the outside of the disc is also stained, ac, adepithelial cells; WP, wing pouch; N, presumptive notum. Bars, 25 μm.

The genera] ultrastructure of the disc epithelium has been described in detail. Microvilli and zonula adherens are characteristic structures of the apical area; the basal surface is relatively smooth, lined by a basal lamina and occasional ‘dense plaques’ which are filamentous patches of microfilaments found near the basal plasma membrane (Fristrom & Fristrom, 1975; Fristrom & Rickoll, 1982). Electron micrographs in Fig. 4 depict the localization of laminin on sections showing structures typical of the apical and basal area, respectively, of unevaginated disc epithelium. The extracellular material excreted by the apical surface appears in Fig. 4A as long filaments decorated by immunogold particles. The basal membrane that underlies the basal surface is strongly stained by immunogold particles. As suggested by Fig. 4B, laminin lies immediately adjacent to the basal plasma membrane and in the extracellular network under the basal lamina. Gold particles, regrouped in a reduced number of sites near the plasma membrane, are also visible inside the cell (Fig. 4A,B). This may be related to observations of Fristrom & Rickoll (1982) that ECM material to be excreted from the cell surface is organized in filamentous strands and suggests that laminin is associated with other ECM components before its secretion.

Fig. 4.

Electron microscopic immunolocalization of laminin in imaginal disc from third instar larvae. Ultrathin sections were stained with Mab RD3 and with a colloidal gold complexed anti-mouse IgM. Immunostaining is localized on the apical cell surface and on filaments of fibrous material in the lumen (A) and at the basal side, in the extracellular network under the basal lamina (B). za, zonula adherens; mv, microvilli; mi, mitochondria; mf, microfilaments; dp, dense plaque. Bars, 0·5 μm.

Fig. 4.

Electron microscopic immunolocalization of laminin in imaginal disc from third instar larvae. Ultrathin sections were stained with Mab RD3 and with a colloidal gold complexed anti-mouse IgM. Immunostaining is localized on the apical cell surface and on filaments of fibrous material in the lumen (A) and at the basal side, in the extracellular network under the basal lamina (B). za, zonula adherens; mv, microvilli; mi, mitochondria; mf, microfilaments; dp, dense plaque. Bars, 0·5 μm.

Therefore laminin appears to be an abundant component of ECM in imaginal discs. Extracellular material secreted apically in the lumen is, however, poorly reactive in immunofluorescence assays. This suggests either that laminin is indeed a minor component in this extracellular material or that the actual antigen conformation reduces the accessibility of the Mab to the cognate epitope. The latter possibility is favoured by the results obtained through immunoelectron microscopy, since in this experimental context the apically secreted extracellular material reacted with Mab RD3 (Fig. 4A). Laminin distribution in the basement membrane surrounding the nonevaginated discs is in contrast to that of fibronectin (Gratecos et al. 1988) and of CN.6D10 antigen, a proteoglycan recently described (Brower et al. 1987). Laminin localization at the basal side of the evaginating wing pouch resembles that of CN.6D10 proteoglycan and suggests that laminin is needed to create the matrix network necessary for elongation of the disc epithelium during évagination.

Localization of the 240K antigen during early development

The 240K RD3 antigen is detected very early in preblastoderm embryos and in the following stages during cellularization, gastrulation and germ band extension, before the onset of laminin synthesis (Fig. 2). Using immunofluorescence assays on whole mounts or embryo sections, we failed to detect the antigen before the syncytial blastoderm stage. Similarly, no immunofluorescence signal was observed in oocytes in ovary sections stained with Mab RD3; moreover the 240K antigen was not found on Western blots of adult female abdomens (data not shown). This indicates that the antigen is probably not stockpiled in oocytes but rather translated in early embryos from maternal transcripts. We have performed a detailed study of the distribution of the 24OK RD3 antigen during early development, by immunofluorescence assays on sections of paraffin-embedded embryos and on whole mounts of permeabilized embryos.

During the syncytial blastoderm stage, cells are not present, except the pole cells, and the nuclei are enclosed within cytoplasmic domains forming cell-like surface caps on the embryo surface. Each cap is marked out by an organized cytoskeleton including microtubules, F-actin microfilaments and intermediate filaments (Foe & Alberts, 1983; Warn, 1986; Karr & Alberts, 1986). The 24OK RD3 antigen colocalizes with the cortical network that surrounds each cytoplasmic island around nuclei (Fig. 5C,D). During syncytial blastoderm stages, the cap undergoes cleavages and becomes progressively smaller; accordingly the rings delineated by the RD3 antigen become smaller. This appears clearly in Fig. 5A,B which are surface views of whole embryos at the 10th and the 13th nuclear cycles, respectively (stage 11 and 14, Foe & Alberts, 1983).

Fig. 5.

Immunolocalization of the 240K antigen in syncytial blastoderm embryo. Whole mounts of permeabilized embryos were stained with Mab RD3 and with rhodamine-labelled anti-mouse IgM. Surface views of syncytial blastoderm embryos at the 10th and the 13th nuclear cycles are shown in A and B, respectively. Higher magnifications of the embryo in A are presented in C and D. Staining with Hoechst 33342 dye allows the visualization of nuclei in C. The 240K antigen localizes in the cortical caps that surround nuclei. Bars, 100 μm.

Fig. 5.

Immunolocalization of the 240K antigen in syncytial blastoderm embryo. Whole mounts of permeabilized embryos were stained with Mab RD3 and with rhodamine-labelled anti-mouse IgM. Surface views of syncytial blastoderm embryos at the 10th and the 13th nuclear cycles are shown in A and B, respectively. Higher magnifications of the embryo in A are presented in C and D. Staining with Hoechst 33342 dye allows the visualization of nuclei in C. The 240K antigen localizes in the cortical caps that surround nuclei. Bars, 100 μm.

This honeycomb pattern over the embryo surface remains after cellularization and during gastrulation. This indicates an antigen localization at the newly formed cell junctions. During these stages, however, the 240K antigen preferentially concentrates in new embryonic structures related to membrane growth and to the organization of embryonic cell layers. This is shown in Figs 5-8 and discussed below.

The blastoderm cell formation requires the active involvement of cytoskeleton elements and particularly of F-actin microfilaments. The 240K antigen distribution can be described correlatively to the cell membrane growth and to the F-actin localization. Before the onset of membrane extension, the antigen colocalizes with F-actin in the cortical caps (Fig. 5). During cellularization, the 240K antigen localizes at the tips of the growing membranes, in the canal furrows which are visualized as fluorescent spots on embryo sections stained with Mab RD3. As membrane growth proceeds, the fluorescent spots progressively move inward into the embryo and reach the yolk. Fig. 6B shows, for example, the antigen accumulation when the cell membranes have grown up to a median position with regard to the nuclei. The furrow canals appear like fluorescent spots between adjacent nuclei. During this process of membrane extension, F-actin microfilaments are also associated with the canal furrows (Fullilove & Jacobson, 1971; Foe & Alberts, 1983). Upon cellularization, the basal cell membrane and the yolk sac are formed, and the antigen becomes concentrated in a narrow band beneath the blastoderm cells (Fig. 6D,F). However, the antigen is not uniformly distributed in this band and local variations in its concentration can be observed under each individual nucleus. Such a discontinuity is probably related to the incompleteness of the cellularization process in the cellular blastoderm and to the existence of a residual connection with the yolk at this stage.

Fig. 6.

Immunolocalization of the 240K antigen during blastoderm cellularization. Longitudinal paraffin sections were stained with Mab RD3 in B,D,F and with Hoechst 33342 dye in A,C,E to visualize nuclei. During cell membrane growth, the 240K antigen is abundant in the furrow canals (arrows in B). Upon cellularization the antigen concentrates beneath the blastoderm cells (D,F). Note also the faint label of lateral cell membranes in D and F (arrows). Bars, 25 μm.

Fig. 6.

Immunolocalization of the 240K antigen during blastoderm cellularization. Longitudinal paraffin sections were stained with Mab RD3 in B,D,F and with Hoechst 33342 dye in A,C,E to visualize nuclei. During cell membrane growth, the 240K antigen is abundant in the furrow canals (arrows in B). Upon cellularization the antigen concentrates beneath the blastoderm cells (D,F). Note also the faint label of lateral cell membranes in D and F (arrows). Bars, 25 μm.

During gastrulation, the epitope recognized by Mab RD3 is also detected on the basal surface of cells; however, a higher accumulation of antigen can be observed in the area involved in morphogenetic movements. Fig. 7A,B is a ventral view of an embryo at the beginning of gastrulation when the presumptive mesoderm invaginates to form the ventral furrow. Mab RD3 strongly stains the ventral furrow and more slightly the cephalic furrows. The two fluorescent lines that parallel the ventral furrow (arrows in Fig. 7B) probably stain the limits of the band of cells that are involved in the furrow and change their basolateral orientation (Poulson, 1950; Fullilove et al. 1978). An oblique section of an early gastrula embryo (Fig. 7C,D) demonstrates the detailed localization of the 240K antigen in the ventral and cephalic furrows. The antigen is present in a continuum that extends along the basal side of the cell monolayer.

Fig. 7.

Immunolocalization of the 240K antigen during gastrulation. Whole mount of permeabilized embryos (A,B) and an oblique cryostat section (C,D) were stained with Mab RD3 (B,D) or with Hoechst 33342 dye (C). The focusing plane in A (phase contrast) and B (immunostaining) is on the surface of the embryo, vf, ventral furrow; cf, cephalic furrow. The antigen is present beneath the cell monolayer (D), and especially at the level of the ventral furrow (B,D). Arrows point in B to the limits of cells involved in the ventral furrow and in D to lateral cell membranes that remain faintly stained at this stage. Bars, 100 μm.

Fig. 7.

Immunolocalization of the 240K antigen during gastrulation. Whole mount of permeabilized embryos (A,B) and an oblique cryostat section (C,D) were stained with Mab RD3 (B,D) or with Hoechst 33342 dye (C). The focusing plane in A (phase contrast) and B (immunostaining) is on the surface of the embryo, vf, ventral furrow; cf, cephalic furrow. The antigen is present beneath the cell monolayer (D), and especially at the level of the ventral furrow (B,D). Arrows point in B to the limits of cells involved in the ventral furrow and in D to lateral cell membranes that remain faintly stained at this stage. Bars, 100 μm.

During germ cell layer segregation and germ band extension, the immunofluorescence staining of embryo sections (not shown) has shown that the 240K antigen accumulated on the basal side of each cellular epithelium, defining a continuous line beneath each cell layer (ectoderm, mesoderm and endoderm). The preferential ventral distribution seen earlier during gastrulation (Fig. 7) was not observed again.

The developmental distribution of the 240K antigen under the embryonic cell layers, the blastoderm sheet as well as the three germ layers, is highly reminiscent of that of the earliest of the ECM components to appear during embryogenesis of the mouse (Dziadek & Timpl, 1985) or sea urchin (McCarthy & Burger, 1987; Wessel & McClay, 1987). This suggests that this protein acts as an extracellular component. The location in the membrane furrows during cellularization and below the cells later is also consistent with an extracellular location at all times, the antigen being brought into the interior with the advancing membrane. Alternatively the antigen may be a transmembrane component.

In order to examine this point further, cell membranes prepared from unstaged embryos have been probed with Mab RD3 by Western blot. The 240K antigen was not detected in membrane preparations and only minute amounts of laminin copurified with cell membrane fractions (Fig. 8; lanes 3 and 4).The distribution of the RD3 antigens in the other fractions merits some additional comments. Laminin was detected as an insoluble form in the pellet of the initial low speed centrifugation (lane 2) and in lesser amounts as a soluble form in the final 100000g supernatant (lane 5). By contrast, the 240K antigen was essentially recovered in the 100 000g supernatant.

Fig. 8.

RD3 antigens are not transmembrane components. Distribution of RD3 antigens in the various fractions obtained during cell membrane preparation. The detail of the purification procedure is given in Materials and methods. The same amounts (7·5 μg) of reduced and alkylated proteins from whole homogenate (lane 1), 25000g pellet (lane 2), crude (lane 3) and purified (lane 4) cell membranes and 100000g supernatant (lane 5) were fractionated by SDS-PAGE on 3·10% linear gradient gel, blotted on nitrocellulose and probed with Mab RD3. Molecular weight standards are shown at the left.

Fig. 8.

RD3 antigens are not transmembrane components. Distribution of RD3 antigens in the various fractions obtained during cell membrane preparation. The detail of the purification procedure is given in Materials and methods. The same amounts (7·5 μg) of reduced and alkylated proteins from whole homogenate (lane 1), 25000g pellet (lane 2), crude (lane 3) and purified (lane 4) cell membranes and 100000g supernatant (lane 5) were fractionated by SDS-PAGE on 3·10% linear gradient gel, blotted on nitrocellulose and probed with Mab RD3. Molecular weight standards are shown at the left.

In this study, we characterized two proteinaceous components which share the epitope recognized by Mab RD3. We do not know yet what the chemical nature of this epitope is. It cannot be related to a particular (native) conformation since Mab RD3 reacts with the denatured antigens on Western blots. Furthermore. the epitope is probably not a carbohydrate group since the periodate oxidation procedure described by Woodward et al. (1985) did not impair the recognition of the antigens by the Mab (not shown). In this context, it is tempting to assume that the differential reactivity observed during development (Fig. 2) does not correspond to conformational changes of the antigens or to post-translational maturation processes, but rather to differential developmental expression patterns of the cognate antigens.

We took advantage of the fact that the 240K protein is absent from third instar larvae to localize laminin in imaginal discs. The prominent observations concern the laminin specificity in the evaginating disc. During evagination, as the cells in the centre of the disc pull away from the basal lamina leaving a loose network of basal lamina material at the open end of the appendage (Fristrom & Rickoll, 1982), laminin is still associated with the basal epithelium, a localization reminiscent of that of CN.6D10 proteoglycan (Brower et al. 1987). This suggests that laminin, and probably other extracellular components are required, together with CN.6D10 proteoglycan, to provide, as proposed by Brower et al. (1987), a new extensible matrix for the dynamics of epithelium elongation and the large shape changes occuring during evagination.

During the development of many organisms (Leivo et al. 1980; Cooper & MacQueen, 1983; McCarthy & Burger, 1987), laminin is synthesized very early, before the appearance of well-defined basement membranes. By contrast, Drosophila laminin (Fessler et al. 1987; Montell & Goodman, 1988; present study) is not expressed in early embryos and cannot therefore be involved in the first morphogenetic movements occurring during gastrulation and germ cell layer segregation. With the exception of fibronectin (Gratecos et al. 1988), all the other major components of basement membranes reported so far (Fessler et al. 1984; Campbell et al. 1987) are also missing. Consequently, if an extracellular matrix is present in Drosophila during gastrulation and the following developmental steps, it presumably has an unusual composition, containing fibronectin and other unknown components.

The 240K antigen appears to be a new ECM component specifically involved in early processes of morphogenesis. The results presented in this paper suggest it might have a role in cell formation and polarization as well as in cell-extracellular substratum interactions after blastoderm cellularization.

Detected for the first time during syncytial blastoderm stages, the antigen colocalizes with egg cytoskeleton components that delimit the territories of putative cells, in the cortical caps (Fig. 5) and in the tips of the growing membranes (Fig. 6). Taken together with the asymmetrical basolateral staining of blastoderm cell surfaces, this suggests the involvement of the 240K RD3 antigen in the organization of the cytoskeleton of the individual cells that adopt an elongated columnar shape. There is much evidence, from organisms other than Drosophila, of dynamic interactions between extracellular matrix, membrane receptors and cytoskeleton in determining cell shape and polarity and also epithelium formation and stabilization (Sugrue & Hay, 1981; Watt, 1986; McCarthy & Burger, 1987).

Surprisingly, few ultrastructural data are available concerning the possible existence of an organized ECM in the early Drosophila embryo. The only structure so far described, in the early gastrula, is a thin band of microfilaments that extends beneath the yolk sac membrane and is present in the remaining connections between embryonic cells and the yolk sac (Rickoll, 1976). This structure may serve as a template for the establishment of a basement membrane. The 240K RD3 antigen is a good candidate for participating in such a process since it might be associated before and during cellularization with cytoskeleton elements and might be involved in cell-extracellular substratum interactions after the cellularization.

The distribution of the RD3 antigens in the various fractions obtained during cell membrane preparation (Fig. 8), shows that the 240K antigen was recovered in the 100000g supernatant, by contrast to laminin that remained as an insoluble form in the pellet of the initial low speed centrifugation (as mentioned in Materials and methods, homogenates of unstaged embryos were prepared in the presence of free Ca, a condition that prevents the solubilization of laminin from organized ECM; Paulsson et al. 1987). Hence, it is tempting to assume that the distribution of the RD3 antigens noted in Fig. 8 is related to ECM organization in embryos, i.e. that the 240K antigen is a component found in early embryos when ECM is not yet completely organized and that laminin is essentially found later in more firmly organized ECM.

We obtained few data about the biochemical nature of the 240K antigen. Lectins (concanavalin A and wheat germ agglutinin have been assayed, not shown) did not react with the immunoprecipitated antigen, indicating that the protein is poorly, or not at all, glycosylated. Moreover, the antigen cannot be metabolically labelled with [35S]methionine and is not digested by chondroitinase ABC (not shown), an enzyme that degrades vertebrate proteoglycans, which indicates that the antigen does not belong to the proteoglycan family. Neither did collagenase treatment modify the electrophoretic migration pattern of the antigen. Laminin-related components involved in embryogenesis have recently been described in several organisms (Yu-Jui Wan et al. 1984; Darribère et al. 1986). The 240K antigen is immunologically related to Drosophila laminin and shares the epitope recognized by Mab RD3 on the chain A. However, these two components do not seem closely related. First, the polyclonal antibodies we assayed, i.e. IgG purified from sera against reduced and alkylated mouse laminin chains and against native laminin, did not react with the immunoprecipitated 240K antigen. Second, and more convincingly, the developmental Northern blot analysis performed by Montell & Goodman (1988) indicates that the laminin genes are coordinately expressed and encode for three transcripts only. Consequently, the 240K antigen encoding gene is distinct from that of laminin chains. At this time, the relationship between 240K antigen and laminin chain A is restricted to the common epitope recognized by Mab RD3. A molecular and genetical approach would be necessary to look for further homologies between the corresponding genes.

We thank Dr C. Mirre from the collagen team in the Laboratoire de la Différenciation Cellulaire (Marseille) for his help in immunoelectron microscopic experiments. We are indebted to Dr M. Piovant for providing us with cell membrane preparations.

Blumberg
,
B.
,
Mackrell
,
A. J.
,
Olson
,
P. F.
,
Kurkinen
,
M.
,
Monson
,
J. M.
,
Natzle
,
J. E.
&
Fessler
,
J. H.
(
1987
).
Basement membrane procollagen IV and its specialized carboxyl domain are conserved in Drosophila mouse and human
.
J. biol. Chem
.
262
,
5947
5950
.
Brower
,
D. L.
,
Piovant
,
M.
,
Salatino
,
R.
,
Brailey
,
J.
&
Hendrix
,
M. J. C.
(
1987
).
Identification of a specialized extracellular matrix component in Drosophila imaginai discs
.
Devi Biol
.
119
,
373
381
.
Campbell
,
A. G.
,
Fessler
,
L. I.
,
Salo
,
T.
&
Fessler
,
J. H.
(
1987
).
Papilin: A Drosophila proteoglycan-like sulfated glycoprotein from basement membranes
.
J. biol. Chem
.
262
,
17605
17 612
.
Cecchini
,
J. P.
,
Knibiehler
,
B.
,
Mirre
,
C.
&
Le Parco
,
Y.
(
1987
).
Evidence for a type IV related collagen in Drosophila melanogaster
.
Eur. J. Biochem
.
165
,
587
593
.
Cooper
,
A. R.
&
Macqueen
,
H. A.
(
1983
).
Subunits of laminin are differentially synthesized in mouse eggs and early embryos
.
Devi Biol
.
96
,
467
471
.
Darribère
,
T.
,
Riou
,
J. F.
,
De Li
Shi
,
Delarue
,
M.
&
Boucaut
,
J. C.
(
1986
).
Synthesis and distribution of laminin-related polypeptides in early amphibian embryos
.
Cell Tissue Res
.
246
,
45
-
51
.
Dziadek
,
M.
&
Timpl
,
R.
(
1985
).
Expression of nidogen and laminin in basement membranes during mouse embryogenesis and in teratocarcinoma cells
.
Devi Biol
.
111
,
372
382
.
Fessler
,
J. H.
,
Lunstrum
,
G.
,
Duncan
,
K. G.
,
Campbell
,
A. G.
,
Sterne
,
R.
,
Bàchinger
,
H. P.
&
Fessler
,
L. I.
(
1984
).
Evolutionary constancy of basement membrane components
.
In The Role of Extracellular Matrix in Development
(ed.
R.
Trelstad
), pp.
207
219
.
New York
:
Alan R. Liss, Inc
.
Fessler
,
L. L
,
Campbell
,
A. G.
,
Duncan
,
K. G.
&
Fessler
,
J. H.
(
1987
).
Drosophila laminin: characterization and localization
.
J. Cell Biol
.
105
,
2383
2391
.
Foe
,
V. E.
&
Alberts
,
B.
(
1983
).
Studies of nuclear and cytoplasmic behaviour during the five mitotic cycles that precede gastrulation in Drosophila embryogenesis
.
J. Cell Sci
.
61
,
31
70
.
Fristrom
,
D.
&
Fristrom
,
J. W.
(
1975
).
The mechanism of évagination of imaginai discs of Drosophila melanogaster
.
I
:
General considerations
.
Devi Biol
.
43
,
1
23
.
Fristrom
,
D. K.
&
Rickoll
,
W. L.
(
1982
).
The morphogenesis of imaginai discs of Drosophila
.
In Insect Ultrastructure
(ed.
R. C.
King
&
H.
Akai
), vol.
1
, pp.
247
277
.
New York
:
Plenum Press
.
Fullilove
,
S. L.
&
Jacobson
,
A. G.
(
1971
).
Nuclear elongation and cytokinesis in Drosophila montana
.
Devi Biol
.
26
,
560
577
.
Fullilove
,
S. L.
,
Jacobson
,
A. G.
&
Turner
,
F. R.
(
1978
).
Embryonic development: Descriptive
.
In The Genetic and Biology of Drosophila
,
2C
(ed.
R. M.
Ashburner
&
T. R. F.
Wright
).
London, New York, San Francisco
:
Academic Press
.
Gratecos
,
D.
,
Naidet
,
C.
,
Astier
,
M.
,
Thiery
,
J. P.
&
Sémériva
,
M.
(
1988
).
Drosophila fibronectin: a protein that shares properties similar to those of its mammalian homologue
.
EMBO J
.
7
,
215
223
.
Greenwood
,
F. C.
,
Hunter
,
W. M.
&
Glover
,
J. S.
(
1963
).
The preparation of 13iI-labeled human growth hormone of high specific radioactivity
.
Biochem. J
.
89
,
114
123
.
Karr
,
T.
&
Alberts
,
B.
(
1986
).
Organization of the cytoskeleton in early Drosophila embryo
.
J. Cell Biol
.
102
,
1494
1509
.
Leivo
,
L
,
Vaheri
,
A.
,
Timpl
,
R.
&
Wartiovaara
,
J.
(
1980
).
Appearance and distribution of collagens and laminin in the early mouse embryo
.
Devi Biol
.
76
,
100
114
.
Leptin
,
M.
,
Aebersold
,
R.
&
Wilcox
,
M.
(
1987
).
Drosophila position-specific antigens resemble the vertebrate Gbronectin-receptor family
.
EMBO J
.
6
,
1037
1043
.
Mccarthy
,
R. A.
,
Beck
,
K.
&
Burger
,
M. M.
(
1987
).
Laminin is structurally conserved in the sea urchin basal lamina
.
EMBO J
.
6
,
1587
1593
.
Mccarthy
,
R. A.
&
Burger
,
M. M.
(
1987
).
In vivo embryonic expression of laminin and its involvement in cell shape change in the sea urchin Sphaerechinus granularis
.
Development
101
,
659
671
.
Milner
,
M. J.
,
Bleasby
,
A. J.
&
Kelly
,
S. L.
(
1984
).
The role of the peripodial membrane of leg and wing imaginai discs of Drosophila melanogaster during évagination and differentiation in vitro
.
Wilhelm Roux’s Arch, devl Biol
.
193
,
180
186
.
Mitchison
,
T.
&
Sedat
,
J.
(
1983
).
Localization of antigenic determinants in whole Drosophila embryos
.
Devi Biol
.
99
,
261
264
.
Montell
,
D. J.
&
Goodman
,
C. S.
(
1988
).
Drosophila substrate adhesion molecule: sequence of laminin Bl chain reveals domains of homology with mouse
.
Cell
53
,
463
473
.
Paulsson
,
M.
,
Aumailley
,
M.
,
Deutzmann
,
R.
,
Timpl
,
R.
,
Beck
,
K.
&
Engel
,
J.
(
1987
).
Laminin-nidogen complex. Extraction with chelating agents and structural characterization
.
Eur. J. Biochem
.
166
,
11
19
.
Poulson
,
D. F.
(
1950
).
Histogenesis, organogenesis and differentiation in the embryo of Drosophila melanogaster Meigen
.
In Biology of Drosophila
(ed.
M.
Demerec
), pp.
168
274
.
New York
:
John Wiley
.
Rickoll
,
W. L.
(
1976
).
Cytoplasmic continuity between embryonic cells and the primitive yolk sac during early gastrulation in Drosophila melanogaster
.
Devi Biol
.
49
,
304
310
.
Sugrue
,
S. P.
&
Hay
,
E. D.
(
1981
).
Response of basal epithelial cell surface and cytoskeleton to solubilized extracellular matrix molecules
.
J. Cell Biol
.
91
,
45
54
.
Towbin
,
H.
,
Staehelin
,
T.
&
Gordon
,
J.
(
1979
).
Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedures and some applications
.
Proc. natn. Acad. Sci. U.S.A
.
76
,
4350
4354
.
Venkatesh
,
T. R.
,
Zingle
,
S.
&
Krishnan
,
K. S.
(
1980
).
Isolation and characterization of membranes from Drosophila melanogaster
.
In Development and Neurobiology of Drosophila
(ed.
O.
Siddigi
,
P.
Babu
,
L. M.
Hall
&
J. C.
Hall
) Basic life Sciences. Vol. 16.
New York and London
:
Plenum Press
.
Warn
,
R. M.
(
1986
).
The cytoskeleton of the early Drosophila embryo
.
J. Cell Sci
.
5
(
suppl
.)
311
328
.
Watt
,
F. M.
(
1986
).
The extracellular matrix and cell shape
.
TIBS
11
,
482
485
.
Wessel
,
G. M.
&
Mcclay
,
D. R.
(
1987
).
Gastrulation in the sea urchin embryo requires the deposition of cross-linked collagen with the extracellular matrix
.
Devi Biol
.
121
,
149
165
.
Woodward
,
M. P.
,
Young
,
W. W.
&
Bloodgood
,
R. A.
(
1985
).
Detection of monoclonal antibodies specific for carbohydrate epitopes using periodate oxidation
.
J. Immunol. Methods
15
,
143
153
.
Yu-Jut
Wan
,
Tsung-Chieh
WU
,
Chung
,
A. E.
&
Damjanov
,
I.
(
1984
).
Monoclonal antibodies to laminin reveal the heterogeneity of basement membranes in the developing and adult mouse tissues
.
J. Cell Biol
.
98
,
971
979
.
Zalokar
,
M.
(
1976
).
Autoradiographic study of protein and RNA formation during early development of Drosophila eggs
.
Devi Biol
.
49
,
425
437
.