Ultrastructural Localization of Wheat Germ Agglutininbinding Sites on Surfaces of Chick Embryo Cells during Early Differentiation E. J. SANDERS AND A. R. ANDERSON Department of Physiology, University of Alberta, Edmonton, Alberta, Canada T6G 2H7
ABSTRACT The objective of this work was to examine changes in a surface component of cells from the chick embryo during morphogenetic migrations of gastrulation. Two electron microscope techniques were used to localize cellbound wheat germ agglutinin (WGA), a lectin which specifically binds N-acetyl glucosamine residues. One technique involved conjugation of peroxidase to WGA before reaction with the cells; the other technique used glucose oxidase to mark WGA which was already cell-bound. In both cases, binding was revealed using diaminobenzidine. Before formation of the primitive streak, all surfaces of the two-layered embryo bound WGA. After migration of cells through the streak, to form the three-layered embryo, not all cell surfaces bound WGA equally. Epiblast cells generally bound WGA lateral to the primitive streak but not during passage through the streak. Mesenchyme cells, after passage through the streak, bound WGA increasingly as they migrated away from the streak. A WGA-binding matrix was observed in the vicinity of the mesenchyme cells and on the dorsal surface of the endoblast. The ventral surface of the endoblast bound the lectin very poorly. In some instances, a peroxidase reaction product was consistently seen on certain surfaces which was not removable by addition of the simple hapten N-acetyl glucosamine. In these cases, the density of the deposit was lessened by use of diacetyl chitobiose as a hapten. This result, together with the reduction of reaction product following certain hyaluronidase treatments, suggests that WGA may be binding to hyaluronic acid as well as membrane glycoproteins. There is increasing evidence that specific components of the cell surface are changed in various ways during early embryonic development. Some of this evidence comes from studies of sea urchin development, where early embryonic stages are characterized by the appearance of new cell surface antigens (McClay et al., '77) and by changes in t h e complement of lectin receptor sites (Krach et al., '74). The results of investigations of several aspects of early avian development have also suggested t h a t such changes occur. In particular, cell surface alterations involving lectin binding to a variety of tissues have been demonstrated in embryos of both chick (Kleinschuster and Moscona, '72; Zalik and Cook, '76; Hook and Sanders, '77) and quail (Sieber-Blum and Cohen, '78). The rationale for pursuing this type of inJ. CELL.
PHYSIOL. (1979)99: 107-124.
vestigation is the probability that the complex carbohydrate moieties of membrane glycoproteins are involved in cellular recognition and adhesive phenomena, thereby providing a controlling mechanism for the morphogenetic migrations of early embryogenesis and differentiation. Studies on early chick embryo cells in culture have shown that various tissues display behavioural differences associated with intercellular recognition and adhesion t h a t may be related to cell surface characteristics (Zalik and Sanders, '74; Sanders and Zalik, '76; Sanders et al., '78). Interpretation of these results is facilitated by the early state of differentiation of the embryos used, which have only a few identifiably-distinct tissues interacting with one another. Received Aug. 3, '78. Accepted Nov. 7 , '78
107
108
E. J. SANDERS AND A. R. ANDERSON
A consideration closely related to the above is that of the role of extracellular matrix compounds in directed cell movements during morphogenesis (for review see Manasek, '75). The predominant extracellular matrix macromolecule present in the gastrulating chick embryo appears to be the glycosaminoglycan, hyaluronic acid (Solursh, '76; Fisher and Solursh, '771, as it also is during neural crest (Pratt et al., '75) and sclerotome (Kvist and Finnegan, '70) migration. Hyaluronic acid is rich in N-acetyl glucosamine residues, as are cell surface glycoproteins (Cook and Stoddart, '73). Since it has been established that wheat germ agglutinin (WGA) will bind specifically to N-acetyl glucosamine residues (see review of Nicolson, '74), it would appear that this lectin might be useful in labelling both cell surface glycoproteins (Shimizu and Yamada, '76) and extracellular matrix molecules which contain this hexosamine. In the present work, we have localized tissue-bound wheat germ agglutinin by means of two different electron microscope histochemical techniques in early chick embryos up to and including the primitive streak (gastrulation) stage of development. It is demonstrated that this lectin does not bind equally to all cell surfaces a t all stages and that matrix materials bind the lectin a t stages when hyaluronic acid is thought to be present. MATERIALS AND METHODS
Embryo preparation Chick embryos were used a t two different stages of early development. The first stage selected was stage 1 of Hamburger and Hamilton ('51), approximately equivalent to stage XI1 of Eyal-Giladi and Kochav ('76), in which an upper layer of cells (epiblast) is underlaid by a lower layer (hypoblast) which is forming as a sheet of cells which will entirely cover the ventral surface of the embryo. The second developmental stage corresponded to stage 5 of Hamburger and Hamilton ('51), in which the primitive streak is fully formed and gastrulation is resulting in the formation of mesenchyme cells between the epiblast and the new lower layer, the endoblast (Vakaet, '70; Rosenquist, '72). Embryos were removed from the vitelline membrane and cleared of yolk in Pannett and Compton's saline. The blastoderms were rinsed in 0.1 M Sorensen's phosphate buffer, pH 7.4,at 4°C before primary fixation for two
hours in 2.5% glutaraldehyde in the same buffer, also a t 4°C. In some experiments, glutaraldehyde fixation was performed after histochemical manipulation, as described later. Histochemical reactions Two histochemical detection techniques for wheat germ agglutinin were used. Firstly, the coupling procedure of Avrameas ('69) was used to conjugate wheat germ agglutinin to horseradish peroxidase (WGA-HRP) using glutaraldehyde. The conjugate is applied to the cells and the cell-bound peroxidase activity is localized by Graham and Karnovsky's ('66) technique using 3,3' diaminobenzidine hydrochloride (DAB). Previously published methods were used (Huet and Garrido, '72; Garrido et al., '74) with modifications. Fifty micrograms of WGA (lectin obtained from Sigma Chemical Co. and Miles Laboratories Inc. behaved in the same way) and 100 p g of HRP (Sigma Chemical Co.) were dissolved in 1 ml of 0.1 M phosphate buffer, pH 6.8. To this was added glutaraldehyde (Ladd Research Industries) to a final concentration of 0.01%and the mixture was left for three hours a t room temperature. The glutaraldehyde concentration was found to be critical. When coupling was performed using 0.03%glutaraldehyde, the resulting conjugate bound non-specifically to cell surfaces, producing a deposit which was not hapten removable. Following the conjugation reaction, the mixture was dialyzed against a large volume of phosphate buffer, containing 2 mg/ml glycine, overnight a t room temperature with two changes of buffer . In control experiments the conjugate was mixed with the hapten, 0.5 M N-acetyl glucosamine (Sigma Chemical Co.), for a t least two hours at room temperature. Other sugars were used as haptens as follows: 0.5 M N-acetyl galactosamine (Sigma Chemical Co.); 0.5 M Nacetyl glucosamine and 0.5 M N-acetyl galactosamine in equal proportions; and 0.05 M NN-diacetylchitobiose (Sigma Chemical Co.). After incubation with the WGA-HRP conjugate for 30 minutes at room temperature, the embryos were rinsed in phosphate buffer and immersed in a solution containing 2 mg/ml DAB (J.T. Baker) and 0.01% hydrogen peroxide in 0.05 M Tris-HC1 buffer, pH 7.6. The DAB mixture was allowed to react for two hours a t room temperature before use on the embryos for 30 minutes. The tissue was then rinsed three times for ten minutes each in buffer
WGA BINDING TO EARLY CHICK EMBRYO CELLS
before post-fixing in 1%phosphate buffered osmium tetroxide for one hour at room temperature. Controls were carried out in which HRP was reacted with glutaraldehyde in the absence of WGA. The result of applying this to the tissues was that no cell surface deposit was produced by DAB. Therefore, the normal reaction products were not due to non-specific adsorption of HRP to cell surfaces. An alternative method of using HRP to visualize WGA was attempted. This technique (Franqois et al., '72) involved binding unconjugated WGA to the cell surface followed by incubation with unconjugated HRP. The peroxidase was then localized using DAB as above. This technique proved unsuitable in our hands, resulting in randomly distributed heavy deposits which were not hapten removable. The second histochemical reaction employed to detect WGA-binding sites differed from the WGA-HRP method in that WGA was allowed to bind to the cell surface without prior conjugation, thus obviating the possibility that the lectin-binding specificity was altered by glutaraldehyde conjugation. This technique used glucose oxidase as a marker (Kuhlman and Avrameas, '71; Franqois and Mongiat, '77). Glucose oxidase, a glucosaminecontaining glycoprotein, is allowed to bind to cell-boundWGA. The bound enzyme is allowed to convert glucose to gluconate with the production of hydrogen peroxide. The latter is reacted with peroxidase in order to catalyze the oxidation of DAB. Wheat germ agglutinin a t 50 pg/ml in phosphate buffer was incubated with the embryos for 30 minutes a t room temperature. In controls, 0.1 M N-acetyl glucosamine was added to the WGA. This was followed by three buffer rinses for ten minutes each and incubation with glucose oxidase (Sigma Chemical Co.) at 1mg/ml in phosphate buffer for 15 minutes at room temperature. After buffer rinses and glutaraldehyde fixation as above, the cell-bound glucose oxidase was revealed by incubation in a mixture of 15 mg glucose, 0.5 mg DAB and 1mg HRP per ml of 0.1 M phosphate buffer, pH 6.8, for 30 minutes at 37OC. Control experiments in which either WGA or glucose oxidase was omitted resulted in no reaction product. Enzyme treatment In some experiments, embryos were treated with enzymes before fixation and histochem-
109
ical procedures. The enzymes used were: neuraminidase (from V. cholerae, SchwartzMann) a t 20 unitdm1 of saline, pH 7.4, for one hour a t 37°C; hyaluronidase (from ovine testes, Sigma Chemical Co., type 111) a t 22.5 units/ml, pH 5.0, for 30 minutes a t 37°C; hyaluronidase (from bovine testes, Sigma Chemical Co., type VI) a t 40 units/ml, pH 5.0 and 7.4, for 30 minutes a t 37°C; hyaluronidase (from Streptomyces, Calbiochem) a t 12.5 or 16.6 unitdm1 in 0.01 M sodium acetate, pH 7.4, plus 0.1 M sodium chloride for 60 minutes at 37°C. Controls were performed for each treatment by incubating tissue under identical conditions but substituting Pannett and Compton's saline for enzyme preparations. Preparation for electron microscopy Post-fixation in 1%phosphate buffered osmium tetroxide was followed by thorough rinses in buffer and dehydration in graded concentrations of ethanol. Embryos were embedded in araldite using propylene oxide as the transition agent. Unstained sections were examined using a Philips 300 electron microscope. RESULTS
Stage 1 embryos (fig. 1) were sectioned in the centre of the area pellucida and no differences were found when different regions of this area were examined. Stage 5 embryos (fig. 2) were usually sectioned just posterior to Hensen's Node. Departures from this sectioning procedure are noted below. Since some degree of variation in staining intensity from sample to sample is inevitable in this type of experiment, care was taken to repeat each experiment several times, with a minimum of ten embryos per experiment. Each embryo was sectioned many times in various regions and the results presented are typical examples. Stage 1 embryos When the WGA-HRP conjugate was applied to prefixed stage 1 embryos in the presence of the hapten N-acetyl glucosamine, no reaction product was found on any cell surfaces (fig. 3). In the absence of the inhibitor, a deposit was observed on all surfaces of the glutaraldehydefixed embryo, i.e., the dorsal, lateral and ventral surfaces of both the epiblast and the hypoblast (figs. 4-7). When the conjugate was applied to unfixed embryos, the result was similar to the above in that all surfaces showed binding. However,
110
E. J. SANDERS A N D A. R. ANDERSON
the reaction product was typically clumped or patched (fig. 8) and was seen within cytoplasmic vacuoles, apparently having been endocytosed from the cell surface (fig. 9). The glucose oxidase reaction carried out in the presence of N-acetyl glucosamine resulted in no surface deposit (fig. 10). In the absence of the inhibitor, all cell surfaces showed an electron dense deposit (figs. 11-13),and the result was therefore the same as that obtained with HRP. The results were also similar in that patching of the surface deposit was observed in unfixed tissue.
The dorsal surface of the epiblast showed a similar amount of binding in the primitive streak region and in regions remote from the streak. The binding of WGA-HRP to the mesenchyme cells varied according to the region examined: cells appeared t o bind more lectin as they migrated further away from the primitive streak. The most striking non-uniformity in surface binding was that shown by endoblast cells. The dorsal surface of this cell layer bound appreciably more conjugate than the ventral surface (figs. 15, 16). It is the dorsal surface upon which the mesenchyme cells migrate. Stage 5 embryos The glucose oxidase reaction, as in the case This stage of development is characterized of stage 1,produced a dense deposit on stage 5 by the presence of mesenchyme cells migrat- cells which was completely abolished in the ing from the primitive streak into the cavity presence of N-acetyl glucosamine. In the abbetween the epiblast and endoblast. Migration sence of N-acetyl glucosamine, a consistent occurs within the cavity (Vakaet, '70; Eben- binding t o various cell surfaces was observed. dal, '76; England and Wakely, '77), presumab- The dorsal surface of the epiblast showed ly on an extracellular matrix, and on the ven- binding in all regions of the area pellucida tral surface of the epiblast and dorsal surface (fig. 17). Lateral surfaces of epiblast cells of the endoblast (figs. 14, 15). Although cell within the primitive streak generally did not surface material and intercellular matrix are bind WGA (fig. lo), although they did elsecontinuous structures, the former is distin- where. The prominent basement membrane on guished by its close association with the the ventral surface of the epiblast showed plasma membrane (up to a thickness of about some deposit (fig. 181, as did surfaces of the 80 nm in these cells). Extracellular material mesenchyme cells, particularly those in close more remote from the plasma membrane is association with the dorsal surface of the entermed matrix in this work. doblast (fig. 19). These latter deposits were In stage 1 embryos, the WGA-HRP conju- both cell surface-bound and matrix-bound. gate resulted in a deposit that was entirely Those that were matrix-bound presented a hapten removable. However, when the same loose fluffy appearance and tended to increase conjugate was applied to stage 5 embryos, a further away from the primitive streak. Aldeposit was formed consistently on certain though the staining reaction on this matrix is cell surfaces which was not eliminated in the not strong, such matrix-type material in assopresence of N-acetyl glucosamine. The deposit ciation with migrating mesenchyme cells is was also not eliminated in the presence of N- rarely seen in conventionally stained sections acetyl galactosamine or by combination of the at this stage of development. As was the case two sugars. These surfaces were: the dorsal with the WGA-HRP reaction, the ventral sursurface of the epiblast; the dorsal and ventral faces of the endoblast bound glucose oxidase surfaces of the endoblast; and the mesen- very poorly or not at all. chyme cell surfaces, particularly ventrally. Enzyme treatment This binding was similar in the presence or abIn view of the fact that WGA can be bound sence of the hapten inhibitor, leaving only the ventral surface of the epiblast and the lateral by surface sialic acid (Greenaway and LeVine, cell surfaces of both epiblast and endoblast '73), embryos were treated with neuraminwhich were hapten inhibited. This binding idase before carrying out the WGA-HRP propattern was consistently observed, indicating cedure. The result was not different from unthat the WGA was binding to the cell surfaces treated specimens. Similarly, embryos were too strongly to be prevented by the presence of pretreated with several grades of testicular the simple sugar hapten, N-acetyl glucosa- hyaluronidase and fungal hyaluronidase. Of mine. The use of N-N-diacetylchitobiose as an the testicular preparations, Sigma type I11 inhibitor reduced the deposit but did not re- had some effect in reducing WGA binding on all surfaces. The Streptomyces hyaluronidase move it entirely.
WGA BINDING TO EARLY CHICK EMBRYO CELLS
hadlittle detectable effect on stage 1embryos, but a t stage 5 the matrix surrounding the mesenchyme cells was clearly reduced (fig. 21). In the latter case, staining remained on the dorsal surface of the epiblast and the basement membrane on the ventral surface of the epiblast remained detectable but fragmented. DISCUSSION
The main points to be taken from this work are: (a) cells in the stage 1 embryo all bind WGA relatively uniformly; (b) in the stage 5 embryo, lateral surfaces of epiblast cells bind WGA remote from the primitive streak but not as they pass through the streak; (c) mesenchyme cells, after passage through the streak, bind WGA with increasing avidity as they move away from the streak; (d) there is a WGA-binding matrix in the vicinity of the mesenchyme cells and on the dorsal surface of the endoblast, while the ventral surface of the endoblast binds the lectin very poorly. The HRP and glucose oxidase techniques gave similar results although the nature of the deposit was not always comparable in density. For example, both techniques showed heavy binding by the dorsal surface of t h e endoblast, which serves as a substrate for the movement of mesenchyme cells, although HRP gave a dense compact deposit while glucose oxidase gave a loose fluffy deposit. The deposit obtained on stage 5 embryos by HRP was not always removable by the presence of the simple sugar hapten, N-acetyl glucosamine. This may reflect an alteration of the WGA specificity as a result of the glutaraldehyde coupling procedure, since the glucose oxidase deposits were hapten removable. An alternative explanation is discussed below, based on the possible binding to hyaluronic acid. The HRP deposits were, nevertheless, considered to be a useful observation for several reasons. Firstly, the deposits occurred consistently on certain surfaces and never on others, and secondly, the hapten removable glucose oxidase deposits occurred on these same surfaces. The general agreement between the results given by in vitro labelled lectin (WGA-HRP) and lectin labelled whilst bound at the cell surface (glucose oxidase) confirms this general conclusion made earlier for concanavalin A (Temmink et al., '75). The cell surface site which binds WGA has been shown to include N-acetyl glucosamine. This was determined on the basis of the haptenic inhibition of WGA binding by this sugar
111
(Burger and Goldberg, '67). Furthermore, the lectin can react with the sugar residue whether the latter is in a core or a terminal position in the saccharide chain (Nicolson, '74). Since lectins bind more strongly to oligosaccharides than to simple sugars (Nicolson, '741, it might be expected that hyaluronic acid, which contains glucosamine in a repeated disaccharide (Cook and Stoddart, '731, would be a likely binding site for WGA. Moreover, the difference in binding strength would prevent N-acetyl glucosamine from inhibiting WGAhyaluronic acid binding. Chitobiose, a disaccharide of N-acetyl glucosamine, was shown here to be more effective in reducing binding than the simple sugar, although it did not eliminate binding altogether. This disaccharide has been shown to be a powerful inhibitor of agglutination by WGA (Burger and Goldberg, '67). It is possible that the use of chitotriose, the trisaccharide, would have further reduced binding since it inhibits WGA even more effectively than does chitobiose (Goldstein et al., '75). Hyaluronic acid itself would not be the inhibitor of choice since the large size of the molecule restricts its effectiveness (Burger and Goldberg, '67). There is substantial evidence from light microscope histochemistry and biochemical studies that hyaluronic acid is present extracellularly in the region of mesenchyme cells and other migrating cells in the early chick embryo (Kvist and Finnegan, '70; Pratt et al., '75; Solursh, '76; Fisher and Solursh, '77). Scanning electron microscope studies also have produced ample demonstration that the extracellular spaces in which these cells move are supplied with matrix fibrils which may provide a substrate for locomotion (Bancroft and Bellairs, '75; Ebendal, '76; England and Wakely, '77). In contrast, transmission electron micrographs prepared by conventional techniques (Low, '68)or using lanthanum (Sanders and Zalik, '72) or ruthenium red (Mestres and Hinrichsen, '74) rarely show the presence of extensive matrix fibrils between the upper and lower layers of the chick embryo, except for the basement membrane on the ventral surface of the epiblast. It is possible that prolonged incubation of embryos in aqueous solutions leaches out such materials, although there has been a recent demonstration of a matrix particularly associated with the leading edge of the advancing mesenchyme (Mayer and Packard, '78). The present results indicate that a ma-
112
E. J. SANDERS AND A. R. ANDERSON
trix can be demonstrated after WGA-glucose oxidase binding, and that it is extensive on the dorsal surface of the endoblast and around the mesenchyme cells that migrate thereon. This matrix seems to correspond to that demonstrated by scanning electron microscopy. With regard to enzyme treatments prior to WGA, the absence of any effect of neuraminidase suggests that, subject to the sensitivity of the technique, surface sialic acids do not bind appreciable amounts of lectin. The reduction in labelling after testicular hyaluronidase suggests t h e presence of hyaluronic acid andlor chondroitin sulphate. Similar reductions in surface labelling after treatment with this enzyme have been reported several times previously (for example Kelley and Lauer, '75; Vogel and Kelley, '77). The effect of fungal hyaluronidase, which digests hyaluronic acid only, was to reduce the extracellular matrix surrounding the mesenchyme cells a t stage 5 while not affecting staining a t the dorsal surface of the epiblast. This might be expected if the mesenchyme matrix was particularly rich in hyaluronate in comparison with surface material more intimately associated with the cell membrane elsewhere. Patching of lectin receptors on unfixed cells as shown here has been reported previously (Temmink et al., '75). Differences in the uniformity of binding have been related to cell transformation (Nicolson, '74), and although it might be possible to relate the patching capacity of cells to their state of differentiation, consideration must be given to the possibility that in this case patching is a non-specific response of cells to WGA binding. Previous studies have shown that when the distribution of concanavalin A-binding sites is compared with that for WGA-binding sites, the result depends on the cell type used. The distributions may be similar to one another (Garrido et al., '74) or dissimilar (Roos and Temmink, '75). In the chick embryo a t stage 1, the concanavalin A-binding sites (Hook and Sanders, '77) showed a similar distribution to the WGA sites demonstrated here. The asymmetric distribution of concanavalin A sites on the individual hypoblast cells, however, was not seen with WGA. At stage 5, the dorsal surface of the epiblast (heavy binding) and the ventral surface of the endoblast (little or no binding) reacted similarly to both lectins. In contrast to this, there was no evidence of
strong binding on the dorsal surface of the endoblast with concanavalin A or of a matrix surrounding the mesenchyme cells. An interesting difference between concanavalin A and WGA binding relates to the composition of the basement membrane on the ventral surface of the epiblast a t stage 1.This surface stained intensely after concanavalin A but much less so after WGA when both lectins are visualized by HRP and DAB and are therefore comparable. This would indicate a preponderance of mannoside or glucoside residues in this structure in comparison with glucosamine residues, The lack of detectable WGA-binding sites on cells in the process of migration through the primitive streak does not seem to be due to a lack of intercellular penetration of reagents since other lateral surfaces with equally close contact show deposit. By contrast, cells within the streak show increased binding of lanthanum ions (Sanders and Zalik, '72) which non-specifically detect negatively-charged groups a t the cell surface. A possible explanation for the absence of WGA-binding in the primitive streak is that the N-acetyl glucosamine residues on apposed cell surfaces are interacting in some manner, which results in their inaccessibility for the WGA molecule. Such interaction in connection with a glycosyltransferase system in tissues of the more advanced chick embryo has been examined by Shur ('77). Indeed, this author makes the suggestion that actively migrating cells in particular show a high degree of participation in such surface activity. The individually migrating mesenchyme cells, after passage through the streak, show a gradually more intense deposit as they move towards the blastoderm periphery. The same observation has been made for ruthenium red staining by these cells (Mestres and Hinrichsen, '74). Such observations suggest that these cells are undergoing progressive modification to their surfaces after the major transformation from the epithelial to fibroblast-like appearance which occurs with their invagination through the primitive streak. These types of changes in cell surface and matrix composition in correlation with various stages of development lend support to the possibility that such extracellular materials may play a controlling role in cell migration and differentiation (Toole, '73). ACKNOWLEDGMENTS
This work was supported by a n operating
WGA BINDING TO EARLY CHICK EMBRYO CELLS
grant from the Medical Research Council of Canada to E. J. Sanders. LITERATURE CITED Avrameas, S. 1969 Coupling of enzymes to proteins with glutaraldehyde. Use of conjugates for the detection of antigens and antibodies. Immunochem., 6: 43-52. Bancroft, M., and R. Bellairs 1975 Differentiation of t he neural plate and neural tube i n the young chick embryo. A study by scanning and transmission electron microscopy. Anat. Embryol., 147: 309-335. Burger, M. M., and A. R. Goldberg 1967 Identification of a tumor-specific determinant on neoplastic cell surfaces. Proc. natn. Acad. Sci. (U.S.A.), 57: 359-366. Cook, G. M. W., and R. W. Stoddart 1973 Surface Carbohydrates of the Eukaryotic Cell. Academic Press, New York. Ebendal, T. 1976 Migratory mesoblast cells i n t he young chick embryo examined by scanning electron microscopy. Zoon, 4: 101-108. England, M. A., and J. Wakely 1977 Scanning electron microscopy of the development of the mesoderm layer in chick embryos. Anat. Embryol., 150: 291-300. Eyal-Giladi, H., and S. Kochav 1976 From cleavage to primitive streak formation: a complementary normal table and a new look a t the first stages of the development of t h e chick. I. General morphology. Develop. Biol., 49: 321-337. Fisher, M., and M. Solursh 1977 Glycosaminoglycan localization and role in maintenance of tissue spaces in the early chick embryo. J. Embryol. exp. Morph., 42: 195-207. Franqois, D., and F. Mongiat 1977 An ultrastructural study of wheat germ agglutinin binding sites using glucose oxidase as a marker. J. Ultrastruct. Res., 59: 119-125. Franqois, D., V. Van Tuyen, H. Febvre and F. Haguenau 1972 Etude a u microscope electronique de l a fixation de lectines marquees a la peroxydase de Raifort sur des cellules embryonnaires humaines transformks in uitro par le virus du sarcome du Rous (RSV), souche Bryan. C.R. Acad. Sci. Paris %r. D., 274: 1981-1984. Garrido, J., M-J. Burglen, D. Samolyk, R. Wicker and W. Bernhard 1974 Ultrastructural comparison between th e distribution of concanavalin A and wheat germ agglutinin surface receptors of normal and transformed hamster and rat cell lines. Cancer Res., 34: 230-243. Goldstein, I. J., S. Hammarstrom and G. Sundblad 1975 Precipitation and carbohydrate-binding specificity studies on wheat germ agglutinin. Biochem. Biophys. Acta, 405: 53-61. Graham, R. M., and M. J. Karnovsky 1966 The early stages of absorption of injected horseradish peroxidase in the proximal tubules of t h e mouse kidney: ultrastructural cytochemistry by a new technique. J. Histochem. Cytochem., 14: 291-302. Greenaway, P. J., and D. LeVine 1973 Binding of N-acetylneuraminic acid by wheat-germ agglutinin. Nature New Biol., 241: 191-192. Hamburger, V., and H. L. Hamilton 1951 A series of normal stages in the development of t he chick embryo. J. Morph., 88: 49-92. Hook, S.L.,and E. J. Sanders 1977 Concanavalin A-binding by cells of t h e early chick embryo. J. Cell. Physiol., 93: 57-68. Huet, C., and J.Garrido 1972 Ultrastructural visualization of cell-coat components by means of wheat germ agglutinin. Exptl. Cell Res., 75: 523-526. Kelley, R. O., and R. B. Lauer 1975 On the nature of the ex-
113
ternal surface of cultured human embryo fibroblasts: a n ultrastructural and cytochemical analysis utilizing stain and lectin probes. Differentiation, 3: 91-97. Kleinschuster, S. J., and A. A. Moscona 1972 Interactions of embryonic and fetal neural retina cells with carbohydrate-binding phytoagglutinins: cell surface changes with differentiation. Exptl. Cell Res., 70: 397-410. Krach, S. W., A. Green, G. L. Nicolson and S. B. Oppenheimer 1974 Cell surface changes occurring during sea urchin development monitored by quantitative agglutination with plant lectins. Exptl. Cell Res., 84: 191-198. Kuhlmann, W. D., and S. Avrameas 1971 Glucose oxidase as an antigen marker for light and electron microscopic studies. J. Histochem. Cytochem., 19: 361-368. Kvist, T. N., and C. V. Finnegan 1970 The distribution of glycosaminoglycans in the axial region of the developing chick embryo. I. Histochemical analysis. J. Exp. Zool., 175: 221-240. Low, F. N. 1968 Extracellular connective tissue fibrils in t he chick embryo. Anat. Rec., 160: 93-108. Manasek, F. J. 1975 The extracellular matrix: a dynamic component of the developing embryo. Curr. Top. Develop. Biol., 10; 35-102. Mayer, B. W., and D. S. Packard 1978 A study of the expansion of the chick area vasculosa. Develop. Biol., 63: 335-351. McClay, D. R., A. F. Chambers and R. H. Warren 1977 Specificity of cell-cell interactions in sea urchin embryos. Appearance of new cell-surface determinants a t gastrulation. Develop. Biol., 56: 343-355. Mestres, P., and K. Hinrichsen 1974 The cell coat in the early chick embryo. Anat. Embryol., 146: 181-192. Nicolson, G. L. 1974 The interactions of lectins with animal cell surfaces. Int. Rev. Cytol., 39: 89-189. Pratt, R. M., M. A. Larsen and M. C. Johnston 1975 Migration of cranial neural crest cells in a cell-free hyaluronate-rich matrix. Develop. Biol., 44: 298-305. Roos, E., and J. H. M. Temmink 1975 Cytochemical comparison between wheat germ agglutinin and concanavalin A bound to mouse fibroblasts in uitro. Exptl. Cell Res., 94: 140-146. Rosenquist, G. C. 1972 Endoderm movements in the chick embryo between t h e early short streak and head process stages. J. Exp. Zool., 180: 95-104. Sanders, E. J., R. Bellairs and P. A. Portch 1978 In uiuo and in uitro studies on the hypoblast and definitive endoblast of avian embryos. J. Embryol. exp. Morph., 46: 187-205. Sanders, E. J., and S. E. Zalik 1972 Studies on the surface of chick blastoderm cells. 11. Electron microscopy of surface binding characteristics. J. Cell. Physiol., 79: 235-247. Sanders, E. J., and S. E. Zalik 1976 Aggregation of cells from early chick blastoderms. Differentiation, 6.-1-11. Shimizu, S., and K. Yamada 1976 Distribution and nature of saccharides of plasma membranes of rat ascites hepatoma cells as detected by ferritin-conjugated lectins and dialysed iron. Exptl. Cell Res., 97: 322-328. Shur, B. D. 1977 Cell-surface glycosyltransferases in gaatrulating chick embryos. I. Temporally and spacially specific patterns of four endogenous glycosyltransferase activities. Develop. Biol., 58: 23-39. Sieber-Blum, M., and A. M. Cohen 1978 Lectin binding to neural crest cells. Changes of the cell surface during differentiation in uitro. J. Cell Biol., 76: 628-638. Solursh, M. 1976 Glycosaminoglycan synthesis in the chick gastrula. Develop. Biol., 50: 525-530. Temmink, J. H. M., J. G. Collard, H. Spits and E. Roos 1975
114
E. J. SANDERS AND A.
A comparative study of four cytochemical detection methods of concanavalin A binding sites on the cell membrane. Exptl. Cell Res., 92: 307-322. Toole, B. P. 1973 Hyaluronate and hyaluronidase in morphogenesis and differentiation. Amer. Zool., 13: 1061-1065. Vakaet, L. 1970 Cinephotomicrographic investigations of gastrulation in the chick blastcderm. Arch. Biol. (Liege), 81: 387-426. Vogel, K. G., and R. 0. Kelley 1977 Cell surface gly-
R. ANDERSON
cosaminoglycans: identification and organization in cultured human embryo fibroblasts. J. Cell. Physiol., 92: 469-480. Zalik, S. E., and G. M. W. Cook 1976 Comparison of early embryonic and differentiating cell surfaces. Interactions of lectins with plasma membrane components. Biophys. Biochim. Acta, 419: 119-136. Zalik, S. E., and E. J. Sanders 1974 Selective cellular affinities in t h e unincubated chick blastcderm. Differentiation, 2: 25-28.
PLATE 1 EXPLANATION OF FIGURES
1 Light micrograph of a section through t he area pellucida of a stage 1embryo, showing the columnar epiblast cells overlying the flattened hypoblast cell layer. Araldite section, methylene blue stained. X 1,000.
2 Light micrograph of a section through the primitive streak of a stage 5 embryo, showing t he dorsal layer of columnar epiblast cells with the loosely packed mesenchyme cells below. Ventrally, a continuous sheet of flattened endoblast cells is present. x 300. 3 This and all succeeding illustrations are unstained electron micrographs. Stage 1. Control for t he WGA-HRP technique. The dorsal surface of the epiblast shows no reaction product. x 37,100. 4 Stage 1. Dorsal surface of t he epiblast, WGA-HRP.
X
5 Stage 1. Lateral surfaces of epiblast cells, WGA-HRP.
37,100. X
37,100.
WGA BINDING TO EARLY CHICK EMBRYO CELLS E. J . Sanders and A. R. Anderson
PLATE 1
115
PLATE 2 EXPLANATION OF FIGURES
6 Stage 1. Ventral surface of the epiblast, WGA-HRP. x 48,600 7 Stage 1. Ventral surface of the hypoblast, WGA-HRP. x 37,100 8 Stage 1. Ventral surface of the hypoblast, WGA-HRP technique on an unfixed embryo. The deposit is patchy in comparison with t h a t seen on fixed tissue (fig. 7). X 37,100. 9 Stage 1. Ventral surface of t h e epiblast, WGA-HRP technique on an unfixed embryo. Note t h e result of extensive pinocytotic activity. x 37,100.
116
WGA BINDING TO EARLY CHICK EMBRYO CELLS E. J. Sanders and A. R. Anderson
PLATE 2
PLATE 3 EXPLANATION OF FIGURES
10 Stage 1. Control for the glucose oxidase technique. Dorsal surface of t h e epiblast showing no reaction product. X 37,100.
11 Stage 1. Dorsal surface of the epiblast, glucose oxidase technique. 12 Stage 1. Ventral surface of the epiblast, glucose oxidase technique.
37,100.
X
X
13 Stage 1. Ventral surface of the hypoblast, glucose oxidase technique.
37,100. X
37,100.
WGA BINDING TO EARLY CHICK EMBRYO CELLS E. J. Sanders and A. R. Anderson
PLATE 3
119
PLATE 4 EXPLANATION OF FIGURES
14 Section through a stage 5 embryo showing epiblast cells (Ep) and underlying mesenchyme cells (MI. Note the mesenchyme cell adhering to the undersurface of the epiblast. X 4,600. 15 Section through a stage 5 embryo, showing mesenchyme cells (MI, one of which is adhering to the upper surface of t h e endoblast (En). Note that the WGA-HRP deposits on the surface of the mesenchyme cells and on the upper surface of t h e endoblast are heavy relative t o the deposit on the ventral surface of the endoblast. X
6,600.
16 Stage 5.Endoblast cell, WGA-HRP technique. Note the dense deposit on the dorsal surface in comparison with t h e ventral surface. X 18,100.
120
WGA BINDING TO EARLY CHICK EMBRYO CELLS E. J. Sanders and A. R. Anderson
PLATE 4
121
PLATE 5 EXPLANATION OF FIGURES
17 Stage 5. Dorsal surface of t h e epihlast, close to t h e primitive streak, glucose oxidase technique. X 37,100. 18 Stage 5.Ventral surface of the epiblast, glucose oxidase technique. x 37,100. 19 Stage 5. The dorsal surface of t h e endoblast (En) showing matrix fibrils (arrow) after the glucose oxidase technique. x 37,100.
20 Stage 5. Lateral surfaces of epiblast cells in the primitive streak region showing the absence of WGA binding. Glucose oxidase technique. X 37,100. 21 Stage 5. Mesenchyme cell surface after treatment with hyaluronidase prepared from Streptomyces. Little or no WGA-binding material is seen at t h e cell surface or in the region where a matrix would be expected. Glucose oxidase technique. X 37,100.
122
WGA BINDING TO EARLY CHICK EMBRYO CELLS E. J. Sanders and A. R. Anderson
PLATE 5
ABSTRACT The objective of this work was to examine changes in a surface component of cells from the chick embryo during morphogenetic migrations of gastrulation. Two electron microscope techniques were used to localize cellbound wheat germ agglutinin (WGA), a lectin which specifically binds N-acetyl glucosamine residues. One technique involved conjugation of peroxidase to WGA before reaction with the cells; the other technique used glucose oxidase to mark WGA which was already cell-bound. In both cases, binding was revealed using diaminobenzidine. Before formation of the primitive streak, all surfaces of the two-layered embryo bound WGA. After migration of cells through the streak, to form the three-layered embryo, not all cell surfaces bound WGA equally. Epiblast cells generally bound WGA lateral to the primitive streak but not during passage through the streak. Mesenchyme cells, after passage through the streak, bound WGA increasingly as they migrated away from the streak. A WGA-binding matrix was observed in the vicinity of the mesenchyme cells and on the dorsal surface of the endoblast. The ventral surface of the endoblast bound the lectin very poorly. In some instances, a peroxidase reaction product was consistently seen on certain surfaces which was not removable by addition of the simple hapten N-acetyl glucosamine. In these cases, the density of the deposit was lessened by use of diacetyl chitobiose as a hapten. This result, together with the reduction of reaction product following certain hyaluronidase treatments, suggests that WGA may be binding to hyaluronic acid as well as membrane glycoproteins. There is increasing evidence that specific components of the cell surface are changed in various ways during early embryonic development. Some of this evidence comes from studies of sea urchin development, where early embryonic stages are characterized by the appearance of new cell surface antigens (McClay et al., '77) and by changes in t h e complement of lectin receptor sites (Krach et al., '74). The results of investigations of several aspects of early avian development have also suggested t h a t such changes occur. In particular, cell surface alterations involving lectin binding to a variety of tissues have been demonstrated in embryos of both chick (Kleinschuster and Moscona, '72; Zalik and Cook, '76; Hook and Sanders, '77) and quail (Sieber-Blum and Cohen, '78). The rationale for pursuing this type of inJ. CELL.
PHYSIOL. (1979)99: 107-124.
vestigation is the probability that the complex carbohydrate moieties of membrane glycoproteins are involved in cellular recognition and adhesive phenomena, thereby providing a controlling mechanism for the morphogenetic migrations of early embryogenesis and differentiation. Studies on early chick embryo cells in culture have shown that various tissues display behavioural differences associated with intercellular recognition and adhesion t h a t may be related to cell surface characteristics (Zalik and Sanders, '74; Sanders and Zalik, '76; Sanders et al., '78). Interpretation of these results is facilitated by the early state of differentiation of the embryos used, which have only a few identifiably-distinct tissues interacting with one another. Received Aug. 3, '78. Accepted Nov. 7 , '78
107
108
E. J. SANDERS AND A. R. ANDERSON
A consideration closely related to the above is that of the role of extracellular matrix compounds in directed cell movements during morphogenesis (for review see Manasek, '75). The predominant extracellular matrix macromolecule present in the gastrulating chick embryo appears to be the glycosaminoglycan, hyaluronic acid (Solursh, '76; Fisher and Solursh, '771, as it also is during neural crest (Pratt et al., '75) and sclerotome (Kvist and Finnegan, '70) migration. Hyaluronic acid is rich in N-acetyl glucosamine residues, as are cell surface glycoproteins (Cook and Stoddart, '73). Since it has been established that wheat germ agglutinin (WGA) will bind specifically to N-acetyl glucosamine residues (see review of Nicolson, '74), it would appear that this lectin might be useful in labelling both cell surface glycoproteins (Shimizu and Yamada, '76) and extracellular matrix molecules which contain this hexosamine. In the present work, we have localized tissue-bound wheat germ agglutinin by means of two different electron microscope histochemical techniques in early chick embryos up to and including the primitive streak (gastrulation) stage of development. It is demonstrated that this lectin does not bind equally to all cell surfaces a t all stages and that matrix materials bind the lectin a t stages when hyaluronic acid is thought to be present. MATERIALS AND METHODS
Embryo preparation Chick embryos were used a t two different stages of early development. The first stage selected was stage 1 of Hamburger and Hamilton ('51), approximately equivalent to stage XI1 of Eyal-Giladi and Kochav ('76), in which an upper layer of cells (epiblast) is underlaid by a lower layer (hypoblast) which is forming as a sheet of cells which will entirely cover the ventral surface of the embryo. The second developmental stage corresponded to stage 5 of Hamburger and Hamilton ('51), in which the primitive streak is fully formed and gastrulation is resulting in the formation of mesenchyme cells between the epiblast and the new lower layer, the endoblast (Vakaet, '70; Rosenquist, '72). Embryos were removed from the vitelline membrane and cleared of yolk in Pannett and Compton's saline. The blastoderms were rinsed in 0.1 M Sorensen's phosphate buffer, pH 7.4,at 4°C before primary fixation for two
hours in 2.5% glutaraldehyde in the same buffer, also a t 4°C. In some experiments, glutaraldehyde fixation was performed after histochemical manipulation, as described later. Histochemical reactions Two histochemical detection techniques for wheat germ agglutinin were used. Firstly, the coupling procedure of Avrameas ('69) was used to conjugate wheat germ agglutinin to horseradish peroxidase (WGA-HRP) using glutaraldehyde. The conjugate is applied to the cells and the cell-bound peroxidase activity is localized by Graham and Karnovsky's ('66) technique using 3,3' diaminobenzidine hydrochloride (DAB). Previously published methods were used (Huet and Garrido, '72; Garrido et al., '74) with modifications. Fifty micrograms of WGA (lectin obtained from Sigma Chemical Co. and Miles Laboratories Inc. behaved in the same way) and 100 p g of HRP (Sigma Chemical Co.) were dissolved in 1 ml of 0.1 M phosphate buffer, pH 6.8. To this was added glutaraldehyde (Ladd Research Industries) to a final concentration of 0.01%and the mixture was left for three hours a t room temperature. The glutaraldehyde concentration was found to be critical. When coupling was performed using 0.03%glutaraldehyde, the resulting conjugate bound non-specifically to cell surfaces, producing a deposit which was not hapten removable. Following the conjugation reaction, the mixture was dialyzed against a large volume of phosphate buffer, containing 2 mg/ml glycine, overnight a t room temperature with two changes of buffer . In control experiments the conjugate was mixed with the hapten, 0.5 M N-acetyl glucosamine (Sigma Chemical Co.), for a t least two hours at room temperature. Other sugars were used as haptens as follows: 0.5 M N-acetyl galactosamine (Sigma Chemical Co.); 0.5 M Nacetyl glucosamine and 0.5 M N-acetyl galactosamine in equal proportions; and 0.05 M NN-diacetylchitobiose (Sigma Chemical Co.). After incubation with the WGA-HRP conjugate for 30 minutes at room temperature, the embryos were rinsed in phosphate buffer and immersed in a solution containing 2 mg/ml DAB (J.T. Baker) and 0.01% hydrogen peroxide in 0.05 M Tris-HC1 buffer, pH 7.6. The DAB mixture was allowed to react for two hours a t room temperature before use on the embryos for 30 minutes. The tissue was then rinsed three times for ten minutes each in buffer
WGA BINDING TO EARLY CHICK EMBRYO CELLS
before post-fixing in 1%phosphate buffered osmium tetroxide for one hour at room temperature. Controls were carried out in which HRP was reacted with glutaraldehyde in the absence of WGA. The result of applying this to the tissues was that no cell surface deposit was produced by DAB. Therefore, the normal reaction products were not due to non-specific adsorption of HRP to cell surfaces. An alternative method of using HRP to visualize WGA was attempted. This technique (Franqois et al., '72) involved binding unconjugated WGA to the cell surface followed by incubation with unconjugated HRP. The peroxidase was then localized using DAB as above. This technique proved unsuitable in our hands, resulting in randomly distributed heavy deposits which were not hapten removable. The second histochemical reaction employed to detect WGA-binding sites differed from the WGA-HRP method in that WGA was allowed to bind to the cell surface without prior conjugation, thus obviating the possibility that the lectin-binding specificity was altered by glutaraldehyde conjugation. This technique used glucose oxidase as a marker (Kuhlman and Avrameas, '71; Franqois and Mongiat, '77). Glucose oxidase, a glucosaminecontaining glycoprotein, is allowed to bind to cell-boundWGA. The bound enzyme is allowed to convert glucose to gluconate with the production of hydrogen peroxide. The latter is reacted with peroxidase in order to catalyze the oxidation of DAB. Wheat germ agglutinin a t 50 pg/ml in phosphate buffer was incubated with the embryos for 30 minutes a t room temperature. In controls, 0.1 M N-acetyl glucosamine was added to the WGA. This was followed by three buffer rinses for ten minutes each and incubation with glucose oxidase (Sigma Chemical Co.) at 1mg/ml in phosphate buffer for 15 minutes at room temperature. After buffer rinses and glutaraldehyde fixation as above, the cell-bound glucose oxidase was revealed by incubation in a mixture of 15 mg glucose, 0.5 mg DAB and 1mg HRP per ml of 0.1 M phosphate buffer, pH 6.8, for 30 minutes at 37OC. Control experiments in which either WGA or glucose oxidase was omitted resulted in no reaction product. Enzyme treatment In some experiments, embryos were treated with enzymes before fixation and histochem-
109
ical procedures. The enzymes used were: neuraminidase (from V. cholerae, SchwartzMann) a t 20 unitdm1 of saline, pH 7.4, for one hour a t 37°C; hyaluronidase (from ovine testes, Sigma Chemical Co., type 111) a t 22.5 units/ml, pH 5.0, for 30 minutes a t 37°C; hyaluronidase (from bovine testes, Sigma Chemical Co., type VI) a t 40 units/ml, pH 5.0 and 7.4, for 30 minutes a t 37°C; hyaluronidase (from Streptomyces, Calbiochem) a t 12.5 or 16.6 unitdm1 in 0.01 M sodium acetate, pH 7.4, plus 0.1 M sodium chloride for 60 minutes at 37°C. Controls were performed for each treatment by incubating tissue under identical conditions but substituting Pannett and Compton's saline for enzyme preparations. Preparation for electron microscopy Post-fixation in 1%phosphate buffered osmium tetroxide was followed by thorough rinses in buffer and dehydration in graded concentrations of ethanol. Embryos were embedded in araldite using propylene oxide as the transition agent. Unstained sections were examined using a Philips 300 electron microscope. RESULTS
Stage 1 embryos (fig. 1) were sectioned in the centre of the area pellucida and no differences were found when different regions of this area were examined. Stage 5 embryos (fig. 2) were usually sectioned just posterior to Hensen's Node. Departures from this sectioning procedure are noted below. Since some degree of variation in staining intensity from sample to sample is inevitable in this type of experiment, care was taken to repeat each experiment several times, with a minimum of ten embryos per experiment. Each embryo was sectioned many times in various regions and the results presented are typical examples. Stage 1 embryos When the WGA-HRP conjugate was applied to prefixed stage 1 embryos in the presence of the hapten N-acetyl glucosamine, no reaction product was found on any cell surfaces (fig. 3). In the absence of the inhibitor, a deposit was observed on all surfaces of the glutaraldehydefixed embryo, i.e., the dorsal, lateral and ventral surfaces of both the epiblast and the hypoblast (figs. 4-7). When the conjugate was applied to unfixed embryos, the result was similar to the above in that all surfaces showed binding. However,
110
E. J. SANDERS A N D A. R. ANDERSON
the reaction product was typically clumped or patched (fig. 8) and was seen within cytoplasmic vacuoles, apparently having been endocytosed from the cell surface (fig. 9). The glucose oxidase reaction carried out in the presence of N-acetyl glucosamine resulted in no surface deposit (fig. 10). In the absence of the inhibitor, all cell surfaces showed an electron dense deposit (figs. 11-13),and the result was therefore the same as that obtained with HRP. The results were also similar in that patching of the surface deposit was observed in unfixed tissue.
The dorsal surface of the epiblast showed a similar amount of binding in the primitive streak region and in regions remote from the streak. The binding of WGA-HRP to the mesenchyme cells varied according to the region examined: cells appeared t o bind more lectin as they migrated further away from the primitive streak. The most striking non-uniformity in surface binding was that shown by endoblast cells. The dorsal surface of this cell layer bound appreciably more conjugate than the ventral surface (figs. 15, 16). It is the dorsal surface upon which the mesenchyme cells migrate. Stage 5 embryos The glucose oxidase reaction, as in the case This stage of development is characterized of stage 1,produced a dense deposit on stage 5 by the presence of mesenchyme cells migrat- cells which was completely abolished in the ing from the primitive streak into the cavity presence of N-acetyl glucosamine. In the abbetween the epiblast and endoblast. Migration sence of N-acetyl glucosamine, a consistent occurs within the cavity (Vakaet, '70; Eben- binding t o various cell surfaces was observed. dal, '76; England and Wakely, '77), presumab- The dorsal surface of the epiblast showed ly on an extracellular matrix, and on the ven- binding in all regions of the area pellucida tral surface of the epiblast and dorsal surface (fig. 17). Lateral surfaces of epiblast cells of the endoblast (figs. 14, 15). Although cell within the primitive streak generally did not surface material and intercellular matrix are bind WGA (fig. lo), although they did elsecontinuous structures, the former is distin- where. The prominent basement membrane on guished by its close association with the the ventral surface of the epiblast showed plasma membrane (up to a thickness of about some deposit (fig. 181, as did surfaces of the 80 nm in these cells). Extracellular material mesenchyme cells, particularly those in close more remote from the plasma membrane is association with the dorsal surface of the entermed matrix in this work. doblast (fig. 19). These latter deposits were In stage 1 embryos, the WGA-HRP conju- both cell surface-bound and matrix-bound. gate resulted in a deposit that was entirely Those that were matrix-bound presented a hapten removable. However, when the same loose fluffy appearance and tended to increase conjugate was applied to stage 5 embryos, a further away from the primitive streak. Aldeposit was formed consistently on certain though the staining reaction on this matrix is cell surfaces which was not eliminated in the not strong, such matrix-type material in assopresence of N-acetyl glucosamine. The deposit ciation with migrating mesenchyme cells is was also not eliminated in the presence of N- rarely seen in conventionally stained sections acetyl galactosamine or by combination of the at this stage of development. As was the case two sugars. These surfaces were: the dorsal with the WGA-HRP reaction, the ventral sursurface of the epiblast; the dorsal and ventral faces of the endoblast bound glucose oxidase surfaces of the endoblast; and the mesen- very poorly or not at all. chyme cell surfaces, particularly ventrally. Enzyme treatment This binding was similar in the presence or abIn view of the fact that WGA can be bound sence of the hapten inhibitor, leaving only the ventral surface of the epiblast and the lateral by surface sialic acid (Greenaway and LeVine, cell surfaces of both epiblast and endoblast '73), embryos were treated with neuraminwhich were hapten inhibited. This binding idase before carrying out the WGA-HRP propattern was consistently observed, indicating cedure. The result was not different from unthat the WGA was binding to the cell surfaces treated specimens. Similarly, embryos were too strongly to be prevented by the presence of pretreated with several grades of testicular the simple sugar hapten, N-acetyl glucosa- hyaluronidase and fungal hyaluronidase. Of mine. The use of N-N-diacetylchitobiose as an the testicular preparations, Sigma type I11 inhibitor reduced the deposit but did not re- had some effect in reducing WGA binding on all surfaces. The Streptomyces hyaluronidase move it entirely.
WGA BINDING TO EARLY CHICK EMBRYO CELLS
hadlittle detectable effect on stage 1embryos, but a t stage 5 the matrix surrounding the mesenchyme cells was clearly reduced (fig. 21). In the latter case, staining remained on the dorsal surface of the epiblast and the basement membrane on the ventral surface of the epiblast remained detectable but fragmented. DISCUSSION
The main points to be taken from this work are: (a) cells in the stage 1 embryo all bind WGA relatively uniformly; (b) in the stage 5 embryo, lateral surfaces of epiblast cells bind WGA remote from the primitive streak but not as they pass through the streak; (c) mesenchyme cells, after passage through the streak, bind WGA with increasing avidity as they move away from the streak; (d) there is a WGA-binding matrix in the vicinity of the mesenchyme cells and on the dorsal surface of the endoblast, while the ventral surface of the endoblast binds the lectin very poorly. The HRP and glucose oxidase techniques gave similar results although the nature of the deposit was not always comparable in density. For example, both techniques showed heavy binding by the dorsal surface of t h e endoblast, which serves as a substrate for the movement of mesenchyme cells, although HRP gave a dense compact deposit while glucose oxidase gave a loose fluffy deposit. The deposit obtained on stage 5 embryos by HRP was not always removable by the presence of the simple sugar hapten, N-acetyl glucosamine. This may reflect an alteration of the WGA specificity as a result of the glutaraldehyde coupling procedure, since the glucose oxidase deposits were hapten removable. An alternative explanation is discussed below, based on the possible binding to hyaluronic acid. The HRP deposits were, nevertheless, considered to be a useful observation for several reasons. Firstly, the deposits occurred consistently on certain surfaces and never on others, and secondly, the hapten removable glucose oxidase deposits occurred on these same surfaces. The general agreement between the results given by in vitro labelled lectin (WGA-HRP) and lectin labelled whilst bound at the cell surface (glucose oxidase) confirms this general conclusion made earlier for concanavalin A (Temmink et al., '75). The cell surface site which binds WGA has been shown to include N-acetyl glucosamine. This was determined on the basis of the haptenic inhibition of WGA binding by this sugar
111
(Burger and Goldberg, '67). Furthermore, the lectin can react with the sugar residue whether the latter is in a core or a terminal position in the saccharide chain (Nicolson, '74). Since lectins bind more strongly to oligosaccharides than to simple sugars (Nicolson, '741, it might be expected that hyaluronic acid, which contains glucosamine in a repeated disaccharide (Cook and Stoddart, '731, would be a likely binding site for WGA. Moreover, the difference in binding strength would prevent N-acetyl glucosamine from inhibiting WGAhyaluronic acid binding. Chitobiose, a disaccharide of N-acetyl glucosamine, was shown here to be more effective in reducing binding than the simple sugar, although it did not eliminate binding altogether. This disaccharide has been shown to be a powerful inhibitor of agglutination by WGA (Burger and Goldberg, '67). It is possible that the use of chitotriose, the trisaccharide, would have further reduced binding since it inhibits WGA even more effectively than does chitobiose (Goldstein et al., '75). Hyaluronic acid itself would not be the inhibitor of choice since the large size of the molecule restricts its effectiveness (Burger and Goldberg, '67). There is substantial evidence from light microscope histochemistry and biochemical studies that hyaluronic acid is present extracellularly in the region of mesenchyme cells and other migrating cells in the early chick embryo (Kvist and Finnegan, '70; Pratt et al., '75; Solursh, '76; Fisher and Solursh, '77). Scanning electron microscope studies also have produced ample demonstration that the extracellular spaces in which these cells move are supplied with matrix fibrils which may provide a substrate for locomotion (Bancroft and Bellairs, '75; Ebendal, '76; England and Wakely, '77). In contrast, transmission electron micrographs prepared by conventional techniques (Low, '68)or using lanthanum (Sanders and Zalik, '72) or ruthenium red (Mestres and Hinrichsen, '74) rarely show the presence of extensive matrix fibrils between the upper and lower layers of the chick embryo, except for the basement membrane on the ventral surface of the epiblast. It is possible that prolonged incubation of embryos in aqueous solutions leaches out such materials, although there has been a recent demonstration of a matrix particularly associated with the leading edge of the advancing mesenchyme (Mayer and Packard, '78). The present results indicate that a ma-
112
E. J. SANDERS AND A. R. ANDERSON
trix can be demonstrated after WGA-glucose oxidase binding, and that it is extensive on the dorsal surface of the endoblast and around the mesenchyme cells that migrate thereon. This matrix seems to correspond to that demonstrated by scanning electron microscopy. With regard to enzyme treatments prior to WGA, the absence of any effect of neuraminidase suggests that, subject to the sensitivity of the technique, surface sialic acids do not bind appreciable amounts of lectin. The reduction in labelling after testicular hyaluronidase suggests t h e presence of hyaluronic acid andlor chondroitin sulphate. Similar reductions in surface labelling after treatment with this enzyme have been reported several times previously (for example Kelley and Lauer, '75; Vogel and Kelley, '77). The effect of fungal hyaluronidase, which digests hyaluronic acid only, was to reduce the extracellular matrix surrounding the mesenchyme cells a t stage 5 while not affecting staining a t the dorsal surface of the epiblast. This might be expected if the mesenchyme matrix was particularly rich in hyaluronate in comparison with surface material more intimately associated with the cell membrane elsewhere. Patching of lectin receptors on unfixed cells as shown here has been reported previously (Temmink et al., '75). Differences in the uniformity of binding have been related to cell transformation (Nicolson, '74), and although it might be possible to relate the patching capacity of cells to their state of differentiation, consideration must be given to the possibility that in this case patching is a non-specific response of cells to WGA binding. Previous studies have shown that when the distribution of concanavalin A-binding sites is compared with that for WGA-binding sites, the result depends on the cell type used. The distributions may be similar to one another (Garrido et al., '74) or dissimilar (Roos and Temmink, '75). In the chick embryo a t stage 1, the concanavalin A-binding sites (Hook and Sanders, '77) showed a similar distribution to the WGA sites demonstrated here. The asymmetric distribution of concanavalin A sites on the individual hypoblast cells, however, was not seen with WGA. At stage 5, the dorsal surface of the epiblast (heavy binding) and the ventral surface of the endoblast (little or no binding) reacted similarly to both lectins. In contrast to this, there was no evidence of
strong binding on the dorsal surface of the endoblast with concanavalin A or of a matrix surrounding the mesenchyme cells. An interesting difference between concanavalin A and WGA binding relates to the composition of the basement membrane on the ventral surface of the epiblast a t stage 1.This surface stained intensely after concanavalin A but much less so after WGA when both lectins are visualized by HRP and DAB and are therefore comparable. This would indicate a preponderance of mannoside or glucoside residues in this structure in comparison with glucosamine residues, The lack of detectable WGA-binding sites on cells in the process of migration through the primitive streak does not seem to be due to a lack of intercellular penetration of reagents since other lateral surfaces with equally close contact show deposit. By contrast, cells within the streak show increased binding of lanthanum ions (Sanders and Zalik, '72) which non-specifically detect negatively-charged groups a t the cell surface. A possible explanation for the absence of WGA-binding in the primitive streak is that the N-acetyl glucosamine residues on apposed cell surfaces are interacting in some manner, which results in their inaccessibility for the WGA molecule. Such interaction in connection with a glycosyltransferase system in tissues of the more advanced chick embryo has been examined by Shur ('77). Indeed, this author makes the suggestion that actively migrating cells in particular show a high degree of participation in such surface activity. The individually migrating mesenchyme cells, after passage through the streak, show a gradually more intense deposit as they move towards the blastoderm periphery. The same observation has been made for ruthenium red staining by these cells (Mestres and Hinrichsen, '74). Such observations suggest that these cells are undergoing progressive modification to their surfaces after the major transformation from the epithelial to fibroblast-like appearance which occurs with their invagination through the primitive streak. These types of changes in cell surface and matrix composition in correlation with various stages of development lend support to the possibility that such extracellular materials may play a controlling role in cell migration and differentiation (Toole, '73). ACKNOWLEDGMENTS
This work was supported by a n operating
WGA BINDING TO EARLY CHICK EMBRYO CELLS
grant from the Medical Research Council of Canada to E. J. Sanders. LITERATURE CITED Avrameas, S. 1969 Coupling of enzymes to proteins with glutaraldehyde. Use of conjugates for the detection of antigens and antibodies. Immunochem., 6: 43-52. Bancroft, M., and R. Bellairs 1975 Differentiation of t he neural plate and neural tube i n the young chick embryo. A study by scanning and transmission electron microscopy. Anat. Embryol., 147: 309-335. Burger, M. M., and A. R. Goldberg 1967 Identification of a tumor-specific determinant on neoplastic cell surfaces. Proc. natn. Acad. Sci. (U.S.A.), 57: 359-366. Cook, G. M. W., and R. W. Stoddart 1973 Surface Carbohydrates of the Eukaryotic Cell. Academic Press, New York. Ebendal, T. 1976 Migratory mesoblast cells i n t he young chick embryo examined by scanning electron microscopy. Zoon, 4: 101-108. England, M. A., and J. Wakely 1977 Scanning electron microscopy of the development of the mesoderm layer in chick embryos. Anat. Embryol., 150: 291-300. Eyal-Giladi, H., and S. Kochav 1976 From cleavage to primitive streak formation: a complementary normal table and a new look a t the first stages of the development of t h e chick. I. General morphology. Develop. Biol., 49: 321-337. Fisher, M., and M. Solursh 1977 Glycosaminoglycan localization and role in maintenance of tissue spaces in the early chick embryo. J. Embryol. exp. Morph., 42: 195-207. Franqois, D., and F. Mongiat 1977 An ultrastructural study of wheat germ agglutinin binding sites using glucose oxidase as a marker. J. Ultrastruct. Res., 59: 119-125. Franqois, D., V. Van Tuyen, H. Febvre and F. Haguenau 1972 Etude a u microscope electronique de l a fixation de lectines marquees a la peroxydase de Raifort sur des cellules embryonnaires humaines transformks in uitro par le virus du sarcome du Rous (RSV), souche Bryan. C.R. Acad. Sci. Paris %r. D., 274: 1981-1984. Garrido, J., M-J. Burglen, D. Samolyk, R. Wicker and W. Bernhard 1974 Ultrastructural comparison between th e distribution of concanavalin A and wheat germ agglutinin surface receptors of normal and transformed hamster and rat cell lines. Cancer Res., 34: 230-243. Goldstein, I. J., S. Hammarstrom and G. Sundblad 1975 Precipitation and carbohydrate-binding specificity studies on wheat germ agglutinin. Biochem. Biophys. Acta, 405: 53-61. Graham, R. M., and M. J. Karnovsky 1966 The early stages of absorption of injected horseradish peroxidase in the proximal tubules of t h e mouse kidney: ultrastructural cytochemistry by a new technique. J. Histochem. Cytochem., 14: 291-302. Greenaway, P. J., and D. LeVine 1973 Binding of N-acetylneuraminic acid by wheat-germ agglutinin. Nature New Biol., 241: 191-192. Hamburger, V., and H. L. Hamilton 1951 A series of normal stages in the development of t he chick embryo. J. Morph., 88: 49-92. Hook, S.L.,and E. J. Sanders 1977 Concanavalin A-binding by cells of t h e early chick embryo. J. Cell. Physiol., 93: 57-68. Huet, C., and J.Garrido 1972 Ultrastructural visualization of cell-coat components by means of wheat germ agglutinin. Exptl. Cell Res., 75: 523-526. Kelley, R. O., and R. B. Lauer 1975 On the nature of the ex-
113
ternal surface of cultured human embryo fibroblasts: a n ultrastructural and cytochemical analysis utilizing stain and lectin probes. Differentiation, 3: 91-97. Kleinschuster, S. J., and A. A. Moscona 1972 Interactions of embryonic and fetal neural retina cells with carbohydrate-binding phytoagglutinins: cell surface changes with differentiation. Exptl. Cell Res., 70: 397-410. Krach, S. W., A. Green, G. L. Nicolson and S. B. Oppenheimer 1974 Cell surface changes occurring during sea urchin development monitored by quantitative agglutination with plant lectins. Exptl. Cell Res., 84: 191-198. Kuhlmann, W. D., and S. Avrameas 1971 Glucose oxidase as an antigen marker for light and electron microscopic studies. J. Histochem. Cytochem., 19: 361-368. Kvist, T. N., and C. V. Finnegan 1970 The distribution of glycosaminoglycans in the axial region of the developing chick embryo. I. Histochemical analysis. J. Exp. Zool., 175: 221-240. Low, F. N. 1968 Extracellular connective tissue fibrils in t he chick embryo. Anat. Rec., 160: 93-108. Manasek, F. J. 1975 The extracellular matrix: a dynamic component of the developing embryo. Curr. Top. Develop. Biol., 10; 35-102. Mayer, B. W., and D. S. Packard 1978 A study of the expansion of the chick area vasculosa. Develop. Biol., 63: 335-351. McClay, D. R., A. F. Chambers and R. H. Warren 1977 Specificity of cell-cell interactions in sea urchin embryos. Appearance of new cell-surface determinants a t gastrulation. Develop. Biol., 56: 343-355. Mestres, P., and K. Hinrichsen 1974 The cell coat in the early chick embryo. Anat. Embryol., 146: 181-192. Nicolson, G. L. 1974 The interactions of lectins with animal cell surfaces. Int. Rev. Cytol., 39: 89-189. Pratt, R. M., M. A. Larsen and M. C. Johnston 1975 Migration of cranial neural crest cells in a cell-free hyaluronate-rich matrix. Develop. Biol., 44: 298-305. Roos, E., and J. H. M. Temmink 1975 Cytochemical comparison between wheat germ agglutinin and concanavalin A bound to mouse fibroblasts in uitro. Exptl. Cell Res., 94: 140-146. Rosenquist, G. C. 1972 Endoderm movements in the chick embryo between t h e early short streak and head process stages. J. Exp. Zool., 180: 95-104. Sanders, E. J., R. Bellairs and P. A. Portch 1978 In uiuo and in uitro studies on the hypoblast and definitive endoblast of avian embryos. J. Embryol. exp. Morph., 46: 187-205. Sanders, E. J., and S. E. Zalik 1972 Studies on the surface of chick blastoderm cells. 11. Electron microscopy of surface binding characteristics. J. Cell. Physiol., 79: 235-247. Sanders, E. J., and S. E. Zalik 1976 Aggregation of cells from early chick blastoderms. Differentiation, 6.-1-11. Shimizu, S., and K. Yamada 1976 Distribution and nature of saccharides of plasma membranes of rat ascites hepatoma cells as detected by ferritin-conjugated lectins and dialysed iron. Exptl. Cell Res., 97: 322-328. Shur, B. D. 1977 Cell-surface glycosyltransferases in gaatrulating chick embryos. I. Temporally and spacially specific patterns of four endogenous glycosyltransferase activities. Develop. Biol., 58: 23-39. Sieber-Blum, M., and A. M. Cohen 1978 Lectin binding to neural crest cells. Changes of the cell surface during differentiation in uitro. J. Cell Biol., 76: 628-638. Solursh, M. 1976 Glycosaminoglycan synthesis in the chick gastrula. Develop. Biol., 50: 525-530. Temmink, J. H. M., J. G. Collard, H. Spits and E. Roos 1975
114
E. J. SANDERS AND A.
A comparative study of four cytochemical detection methods of concanavalin A binding sites on the cell membrane. Exptl. Cell Res., 92: 307-322. Toole, B. P. 1973 Hyaluronate and hyaluronidase in morphogenesis and differentiation. Amer. Zool., 13: 1061-1065. Vakaet, L. 1970 Cinephotomicrographic investigations of gastrulation in the chick blastcderm. Arch. Biol. (Liege), 81: 387-426. Vogel, K. G., and R. 0. Kelley 1977 Cell surface gly-
R. ANDERSON
cosaminoglycans: identification and organization in cultured human embryo fibroblasts. J. Cell. Physiol., 92: 469-480. Zalik, S. E., and G. M. W. Cook 1976 Comparison of early embryonic and differentiating cell surfaces. Interactions of lectins with plasma membrane components. Biophys. Biochim. Acta, 419: 119-136. Zalik, S. E., and E. J. Sanders 1974 Selective cellular affinities in t h e unincubated chick blastcderm. Differentiation, 2: 25-28.
PLATE 1 EXPLANATION OF FIGURES
1 Light micrograph of a section through t he area pellucida of a stage 1embryo, showing the columnar epiblast cells overlying the flattened hypoblast cell layer. Araldite section, methylene blue stained. X 1,000.
2 Light micrograph of a section through the primitive streak of a stage 5 embryo, showing t he dorsal layer of columnar epiblast cells with the loosely packed mesenchyme cells below. Ventrally, a continuous sheet of flattened endoblast cells is present. x 300. 3 This and all succeeding illustrations are unstained electron micrographs. Stage 1. Control for t he WGA-HRP technique. The dorsal surface of the epiblast shows no reaction product. x 37,100. 4 Stage 1. Dorsal surface of t he epiblast, WGA-HRP.
X
5 Stage 1. Lateral surfaces of epiblast cells, WGA-HRP.
37,100. X
37,100.
WGA BINDING TO EARLY CHICK EMBRYO CELLS E. J . Sanders and A. R. Anderson
PLATE 1
115
PLATE 2 EXPLANATION OF FIGURES
6 Stage 1. Ventral surface of the epiblast, WGA-HRP. x 48,600 7 Stage 1. Ventral surface of the hypoblast, WGA-HRP. x 37,100 8 Stage 1. Ventral surface of the hypoblast, WGA-HRP technique on an unfixed embryo. The deposit is patchy in comparison with t h a t seen on fixed tissue (fig. 7). X 37,100. 9 Stage 1. Ventral surface of t h e epiblast, WGA-HRP technique on an unfixed embryo. Note t h e result of extensive pinocytotic activity. x 37,100.
116
WGA BINDING TO EARLY CHICK EMBRYO CELLS E. J. Sanders and A. R. Anderson
PLATE 2
PLATE 3 EXPLANATION OF FIGURES
10 Stage 1. Control for the glucose oxidase technique. Dorsal surface of t h e epiblast showing no reaction product. X 37,100.
11 Stage 1. Dorsal surface of the epiblast, glucose oxidase technique. 12 Stage 1. Ventral surface of the epiblast, glucose oxidase technique.
37,100.
X
X
13 Stage 1. Ventral surface of the hypoblast, glucose oxidase technique.
37,100. X
37,100.
WGA BINDING TO EARLY CHICK EMBRYO CELLS E. J. Sanders and A. R. Anderson
PLATE 3
119
PLATE 4 EXPLANATION OF FIGURES
14 Section through a stage 5 embryo showing epiblast cells (Ep) and underlying mesenchyme cells (MI. Note the mesenchyme cell adhering to the undersurface of the epiblast. X 4,600. 15 Section through a stage 5 embryo, showing mesenchyme cells (MI, one of which is adhering to the upper surface of t h e endoblast (En). Note that the WGA-HRP deposits on the surface of the mesenchyme cells and on the upper surface of t h e endoblast are heavy relative t o the deposit on the ventral surface of the endoblast. X
6,600.
16 Stage 5.Endoblast cell, WGA-HRP technique. Note the dense deposit on the dorsal surface in comparison with t h e ventral surface. X 18,100.
120
WGA BINDING TO EARLY CHICK EMBRYO CELLS E. J. Sanders and A. R. Anderson
PLATE 4
121
PLATE 5 EXPLANATION OF FIGURES
17 Stage 5. Dorsal surface of t h e epihlast, close to t h e primitive streak, glucose oxidase technique. X 37,100. 18 Stage 5.Ventral surface of the epiblast, glucose oxidase technique. x 37,100. 19 Stage 5. The dorsal surface of t h e endoblast (En) showing matrix fibrils (arrow) after the glucose oxidase technique. x 37,100.
20 Stage 5. Lateral surfaces of epiblast cells in the primitive streak region showing the absence of WGA binding. Glucose oxidase technique. X 37,100. 21 Stage 5. Mesenchyme cell surface after treatment with hyaluronidase prepared from Streptomyces. Little or no WGA-binding material is seen at t h e cell surface or in the region where a matrix would be expected. Glucose oxidase technique. X 37,100.
122
WGA BINDING TO EARLY CHICK EMBRYO CELLS E. J. Sanders and A. R. Anderson
PLATE 5
