Concanavalin A-Binding by Cells of the Early Chick Embryo S. L. HOOK AND E. J. SANDERS Department of Physiology, University ofAlberta, Edmonton, Alberta, Canada T6G 2H7
ABSTRACT The surfaces of cells from the early embryo of the chick were examined using electron microscope techniques for the visualization of concanavalin A-binding sites. Horseradish peroxidase and Ferritin labelled concanavalin A were used to determine the distribution of the binding sites. All surfaces of the epiblast and hypoblast layers which were accessible to concanavalin A showed the presence of binding sites in stage 1embryos. The ventral surface of the epiblast showed a high lectin affinity which may reflect the development of a basal lamina on this surface. The individual hypoblast cells a t this stage showed a non-uniform distribution of binding sites, having a greater affinity on the dorsal surface than the ventral. By the time of primitive streak formation (stage 45) the dorsal surface of the epiblast displayed increased binding sites, while the frequency of sites on the ventral surface of the endoblast was reduced. The latter may reflect a change from one cell population to another, which occurs in the lower layer of the embryo at this time. No consistent correlation could be drawn between changes in motility of cells actually invaginating through the primitive streak and changes in affinity for concanavalin A. An overall increase in affinity of the dorsal surface of the epiblast was revealed by Ferritin and may reflect the changes in surface structure occurring in readiness for the morphogenetic migrations of gastrulation. The cell surface plays an important role in the control of cell growth, cell movement and cell recognition, and is consequently intimately involved in differentiation and morphogenesis. The early chick embryo provides a good system for the study of cell surface changes accompanying differentiation and morphogenetic migration. Previous work on the surface properties of these cells has characterized their electrophoretic mobility (Zalik e t al., '721, the binding of markers to surface ionogenic groups (Sanders and Zalik, '72) and the affinity of the cells for calcium ions (Harris and Zalik, '74). More recent work has dealt with the reaggregation and sorting out of dissociated chick cells taken from early developmental stages. It has been shown that the surfaces of blastoderm cells are sufficiently differentiated to possess characteristics which allow them to sort out into two cell populations, (Zalik and Sanders, '74; Eyal-Giladi e t al., '75; Sanders and Zalik, '76). The molecular basis for the expression of J. CELL. PHYSIOL., 93: 57-68.
cell surface differences appears to lie in surface receptors of a carbohydrate nature (see review of Hughes, '73). Therefore, lectins such as concanavalin A (con A) which bind to specific surface glycoproteins can be used as probes with which to study changes in the distribution of such receptor sites, (Nicolson, '74). That con A binds to, and agglutinates, cells from the early chick blastoderm has been demonstrated, by Zalik and Cook ("76).These authors, and also Kleinschuster and Moscona ('72), have shown that as development proceeds the availability of lectin binding sites is altered, as judged by agglutinability, probably by the masking effects of trypsin-sensitive material at the cell surface. It was the purpose of this study to elucidate the distribution of con A binding sites (CABS) on surfaces of cells from the early chick embryo. The relative numbers of binding sites were compared on surfaces of corresponding cells in prestreak and gastrulating embryos Received Feb. 1, '77. Accepted Apr. 7, '77.
57
58
S. L. HOOK AND E. J. SANDERS
by the use of ultrastructural cytochemical techniques. A change in the binding of con A on any surface was taken to indicate alterations in the complement of cell surface receptor sites. It was considered that such alterations could in turn relate t o the onset of morphogenetic migrations in that particular cell group or be indicative of the presence of different populations of cells within the embryo. Some aspects of the con A affinity of early amphibian embryo cells (O'Dell et al., '74; Sanders and DiCaprio, '76; Johnson and Smith, '76) and chick embryo cells at later stages (Lee e t al., '76) has been described previously. MATERIALS AND METHODS
Stage 1 and stages 4-5blastoderms (Hamburger and Hamilton, '51) were removed from the vitelline membrane and placed in Pannett and Compton's saline. Stage 1 blastoderms were marked by removing a wedge of tissue to facilitate orientation at the time of embedding. Embryos were given a thorough wash with 0.1 M cacodylate buffer, pH 7.4, and fixed for two hours in 2.5% buffered glutaraldehyde. For detection of CABS with horseradish peroxidase (HRP), a modification of Bernhard and Avrameas' ('71) method was employed using Graham and Karnovsky's ('66) diaminobenzidine (DAB) technique. The blastoderms were washed in buffer and incubated in 100 pg/ml con A (Sigma), in cacodylate buffer pH 7.2 for two hours, washed again in buffer and incubated in 50 p g/ml horseradish peroxidase (Sigma Type VI) for 30 minutes. After another buffer wash they were incubated in a saturated aqueous solution of 3,3'diaminobenzidine hydrochloride (J. T. Baker analyzed reagent) and 0.01%hydrogen peroxide in 0.5 M Tris-HC1 buffer, pH 7.6 for 14 to 30 minutes followed by a further wash in buffer. Control embryos were incubated in con A plus 0.5 M a methyl-D-mannoside (aMM, Sigma Grade 111) and HRP with 0.5 M a M M added. All of the above was carried out a t 22°C. Further controls were done in the absence of con A to eliminate the possibility of nonspecific binding of HRP-DAB reaction product to the cell surface. The first group of control blastoderms was incubated directly after glutaraldehyde fixation with 50 pg/ml HRP for 30 minutes rinsed and placed in a saturated solution of DAB with 0.01%hydrogen peroxide
in Tris-HC1buffer for 15 minutes. A second set of embryos was incubated in saturated DAB in Tris-HC1 buffer for 15 minutes and in 3 X M potassium ferricyanide for five minutes. Since potassium ferricyanide readily oxidizes DAB, mimicking the effect of hydrogen peroxide, this control was to determine whether non-specific adsorption of DAB or its oxidation product occurred. A third set of blastoderms was incubated in saturated DAB and 0.01% hydrogen peroxide in Tris-HC1 buffer. Ferritin conjugated covalently to con A (Calbiochem A grade) was also used to detect CABS. After fixation in 2.5%phosphate buffered glutaraldehyde pH 7.4 the blastoderms were thoroughly washed in buffer and incubated in approximately 100 pg/ml ferritin-con A for one hour, and then thoroughly washed in buffer again. All embryos were then post-fixed in 1%phosphate buffered osmium tetroxide for one hour at 4OC. All blastoderms were dehydrated in an ethanol series, followed by propylene oxide and embedded in araldite. Stage 1 embryos were sectioned in two areas: anteriorly and posteriorly. Stage 4 and 5 embryos were sectioned in an anterior region, either anterior to Hensen's node or through Hensen's node, and then through the anterior third of the primitive streak. Some embryos were sectioned in a third region, in the posterior two-thirds of the primitive streak. RESULTS
Con A-HRP-DAB produced a dense reaction product which did not penetrate the plasma membrane. The density of reaction product was evaluated on the basis of the thickness of t h e deposit and whether i t displayed a clumped or even pattern of binding. The scoring of the reaction product was a follows: 0, no reaction product; , a sparsely scattered thin layer; +, a denser layer in which most or all of the membrane has reaction product; and +++, a very thick deposit over the entire membrane. Ferritin-con A was seen as a particulate deposit, limited to the outer surface of the plasma membrane. Generally, these particles displayed a nonuniform distribution, appearing singly or in clumps separated by areas of membrane with no molecules bound. A score of 0 was given if no reaction product was found, if particles of ferritin were found
+
+
+
59
CON A BINDING BY CHICK EMBRYO CELLS TABLE 1
Summary of results Stage 4.5
Stage 1 Anterior
Posterior
Anterior
Concanavalin A - Horseradish peroxidase - Diaminobenzidine + f ++ DE +++ 0 VE 0 0-+ ’ 0 DN 0 +-++’ +-++I VN 0 Ferritin - Concanavalin A ++ ++ DE +++ VE 0-+ or 0 +-+ DN 0-+ or 0 +-+ VN 0- +
++
DE VE DH VH DE VE DH VH
+++ + +-++
+++
+-++ +
Posterior
++’ 0 0 0
+++ 0-+ or 0 0-+ or 0 0-
+
See text for evaluation procedure and explanation of symbols. DE, dorsal surface of the epiblast; VE, ventral surface of the epiblast; DH, dorsal surface of the hypoblast; VH, ventral surface of the hypoblast; DN, dorsal surface of the endohlast; VN, ventral surface of the endoblast. ‘ S e e text for further details.
++
widely spaced on the membrane, if partiif the cles were closely adjacent, and membrane was largely covered by ferritin particles or aggregates. The symbol 0 is used in the table to indicate that although no reaction was present this may have reflected the inaccessibility of the surface to the cytochemical procedure. Generally the control procedures using aMM showed no binding of reaction product (figs. 1, 2).
cells showed a non-uniform distribution of deposit. The ventral surface of individual cells possessed denser deposits than the dorsal surface of the same cell (figs. 7-91. The ventral surface of the hypoblast was always stained but the pattern varied in density (+-++ +). Ferr-con A results are summarized in table 1. Sections taken from anterior regions of stage 1did not usually have a complete hypoblast, but there was no difference between anterior regions and posterior regions in the Stage 1 pattern of ferr-con A binding despite the At this stage of development the embryo difference in the completion of the hypoblast. consists of one layer of dorsally situated cells Ferr-con A appeared to penetrate more easily (epiblast) and a more or less complete lower between the cell layers than the HRP-DAB layer of cells (hypoblast). The hypoblast be- markers. The deposits shown by ferr-con A comes a coherent layer in the posterior region were more consistent than those found with initially and spreads anteriorly by migrating con A-HRP-DAB and in agreement with the on the ventral surface of the epiblast. HRP result, the dorsal surface of the epiblast Although within each experiment some var- showed a less dense deposit ( + + ; fig. 51, than ; fig. iability existed between embryos with respect the ventral surface of the epiblast ( to corresponding surfaces, the following pat- 6). Single hypoblast cells showed a sparse tern resulted with con A-HRP-DAB (table 1). deposit on their surface with no striking In stage 1 embryos all surfaces showed stain- changes in binding intensity between the ing except where a complete layer of hypo- various aspects of individual cells. In sections blast was found in the posterior regions of taken from posterior regions the dorsal and the embryo. In these areas no reaction product ventral surfaces of the hypoblast had ferritin was found on the ventral surface of the epi- binding similar to that in anterior regions blast or the dorsal surface of the hypoblast (0). (+-+ +). Except for this, no differences were found beStage 4-5 tween the embryos sectioned in anterior and This stage of development is characterized posterior regions. The dorsal surface of the epiblast showed a fairly dense deposit ( + + ; by the primitive streak, and a complete lower fig. 31, and the ventral surface of the epiblast layer of cells, (endoblast). showed a very dense clumped pattern in the Stage 4-5 embryos with con A-HRP-DAB anterior region ( + +; fig. 4). If the hypo- (table 1) displayed binding on the dorsal surblast was incomplete, individual hypoblast face of the epiblast in anterior regions ( + +;
+++
+++
+
60
S. L. HOOK AND E. J. SANDERS
fig. 10).No reaction product was found on the ventral surface of the epiblast or the dorsal surface of the endoblast in any area sectioned, (0). The ventral surface of the endoblast did not bind reaction product in either region (0; fig. 12). In areas taken from the primitive streak, two distinct patterns of reaction product were found on the dorsal surface of the epiblast. In one, no reaction product was found on the walls of the groove or on the base of the groove. The absence of reaction product in the primitive groove was not seen in all embryos, where a dense clumped layer of reaction product was observed with no discontinuity. The reason for this variation was not readily apparent. Stage 4-5 embryos with ferr-con A (table 1) displayed a heavier binding on the dorsal surface of the epiblast (+++; fig. l l ) , than stage 1, with ferritin particles often aggregated. No differences in the affinity of this surface for lectin were detected with respect to the region of the blastoderm examined. The surface of cells in the primitive streak (fig. 14) showed the same deposits as cells more peripherally situated. In both areas sectioned the vcntral surface of the cpiblnst and the dorsal surface of the endoblast displayed little or no reaction product, ( O d + or 0 ) . The ventral surface of the endoblast in the anterior region and beneath the primitive streak displayed a pattern with a few isolated particles or no reaction product (fig. 13; 0-+). DISCUSSION
Although two different cytochemical techniques were used in this study, the results given by each were similar. The main discrepancy was that of the pattern seen on individual hypoblast cells, since the non-uniform distribution seen with HRP-DAB did not occur with ferr-con A. The denser reaction product seen on the ventral surface compared to the dorsal surface with HRP-DAB was probably the result of the very high activity of HRP and the amplification effect which occurs during oxidation of the DAB (Temmink et al., '75). Thus, what may have been a small difference in binding capacity between these two surfaces was not detected by Ferr-con A but, owing to the amplification, was demonstrated by HRP-DAB. The denser deposit on the dorsal surface may reflect the presence of an otherwise indistinct basal lamina, not previously observed. The most immediate difference between the
two stages of development examined was the affinity of the ventral surface of the lower layer. In stage 1 this surface showed a consistent binding of the lectin, which was absent a t stage 4-5. This change is probably a reflection of the different populations of cells that make up the lower layer of the blastoderm during early morphogenesis. In stage 1, the lower layer consists of groups of cells which aggregate into a single layer to form the hypoblast, (Vakaet, '70; Eyal-Giladi and Kochav, '76). At stage 5, these cells have moved peripherally, and have been replaced by the definitive endoblast (Vakaet, '701, which develops into the embryonic endoderm (Rosenquist, ' 66). Bellairs et al. ('77) have demonstrated behavioural differences between hypoblast and endoblast cells which may derive from cell surface differences such as those demonstrated here, and their different lectin affinities may provide a further means of distinguishing them. If the definitive endoblast, which did not bind con A, originates by the movement of cells from the epiblast, (Vakaet, TO), which bound con A on its dorsal and ventral surfaces, the receptor sites may be masked or removed from the cell surface after invngination. Wolk et al. ('74) reported that the antigenic specificity of primary hypoblast cells a t stage 1 is not present on cells of the lower layer a t stage 5. Furthermore, these authors showed that the hypoblast cells were not uniformly stained with fluorescent antibody, in that case being most intense on the dorsal surface. This result is similar to the present one, where con A also appears to bind nonuniformly. Differences were also seen in the ventral surface of the epiblast. However, the marked change from stage 1 embryos, with a very dense staining, to no staining in stage 5 may result from inaccessibility to the lectin. During stage 1, a basal lamina is laid down on the ventral surface of the epiblast, which can be seen as a diffuse amorphous deposit, (Low, '67). Martinez-Palomo ('70) reported that this structure was stained with ruthenium red, which marks acid mucopolysaccharides, a t the primitive streak stage. The present study confirmed the presence of mannose andlor glucose residues in this structure as early as stage 1. The possibililty that con A binding sites are more mobile within the membrane of migratory cells than of non-migratory cells, resulting in different patterns of distribution, has been
CON A BINDING BY CHICK EMBRYO CELLS
61
chemistry by a new technique. J. Histochem. Cytochem., 14: 291-302. Hamburger, V., and H. L. Hamilton 1951 A series of normal stages in the development of the chick embryo. J. Morph., 88: 49-92. Harris, H. L., and S. E. Zalik 1974 Studies on the surface of chick blastoderm cells. 111. Calcium binding to ionogenic sites. J. Cell. Physiol., 83:359-368. Hughes, R. B. 1973 Glycoproteins as components of cellular membranes. Prog. Biophys. Molec. Biol., 26: 189268. Johnson, K. E., and E. P. Smith 1976 The binding of concanavalin A to dissociated embryonic amphibian cells. Exptl. Cell Res., 101:63-70. Kleinschuster, S. J., and A. A. Moscona 1972 Interactions of embryonic and fetal neural retina cells with carbohy drate-binding phytoagglutinins: cell surface changes with differentiation. Exptl. Cell Res., 70: 397-410. Lee, H.-Y., J. B. Sheffield, R. G. Nagele and G. W. Kalmus 1976 The role of extracellular material in chick neurulation. I. Effects of concanavalin A. J. Exp. Zool., 198:261266. Low, F. N. 1967 Developing boundary (Basement) membranes in the chick embryo. Anat. Rec., 159:231-238. Martinez-Palomo, A. 1970 The surface coats of animal cells. Int. Rev. Cytol., 29: 29-75. Nicolson, G. L. 1974 The interactions of lectins with animal cell surfaces. Int. Rev. Cytol., 39: 89-190. ODell, D. S., R. Tencer, A. Monroy and J. Brachet 1974 The pattern of concanavalin A binding sites during the early development of Xenopus loeuis. Cell Differentiation, 3: 193-198. Roberson, M., A. Neri and S. B. Oppenheimer 1975 Distribution of concanavalin A receptor sites on specific populations of embryonic cells. Science, 189: 639-640. Rosenquist, G. C. 1966 A radioautographic study of labelled grafts in the chick blastoderm. Contr. Embryol. Carneg. Inst., 38: 71-113. Sanders, E. J., and R. A. DiCaprio 1976 A freeze-fracture and concanavalin A binding study of the membrane of cleaving Xenopus embryos. Differentiation, 7:13-21. 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: 235247. ACKNOWLEDGMENTS Sanders, E. J., and S. E. Zalik 1976 Aggregation of cells from early chick blastoderms. Differentiation, 6: 1-11. This work was supported by an operating Temmink, J. H. M., J. G. Collard, H. Spits and E. Roos 1975 grant from the Medical Research Council of A comparative study of four cytochemical detection Canada to E. J. Sanders. methods of concanavalin A binding sites on the cell membrane. Exptl. Cell Res., 92: 307-322. Vakaet. L. 1970 Cineohotomicroeraohic investieations LITERATURE CITED of gastrulation in the chick blktbderm. Arc(. Biol. Bellairs, R., P. A. Portch and E. J. Sanders 1977 Behaviour (Liege), 81: 387-426. patterns of chick blastoderm cells. J. Anat., in press. Wolk, M., M. Schlesinger and H. Eyal-Giladi 1974 The deBernhard, W., and S. Avrameas 1971 Ultrastructural velopment of and migration of the primary hypoblast visualization of cellular carbohydrate components by cells in the chick embryo studied by immunofluorescent means of Concanavalin A. Exptl. Cell Res., 64: 232-236. technique. Isr. J. Zool.,23: 204-205. Eyal-Giladi, H., S. Kochav and A. Yerushalmi 1975 The Zalik, S. E., E. J. Sanders and C. Tilley 1972 Studies on the sorting-out of thymidine-labelled chick hypoblast cells in surface of chick blastoderm cells. 1. Electrophoretic mixed epiblast-hypoblast aggregates. Differentiation, 4: mobility and pH mobility relationships. J. Cell. Physiol., 57-60. 79:225-234. Eyal-Giladi, H., and S. Kochav 1976 From cleavage to Zalik, S. E., and E. J. Sanders 1974 Selective cellular primitive streak formation: a complementary normal affinities in the unincubated chick blastoderm. Differtable and new look a t the first stages of the development entiation, 2: 25-28. of the chick. Develop. Biol., 49: 321-337. Zalik, S. E., and G. M. W. Cook 1976 Comparison of early embryonic and differentiating cell surfaces. Interaction Graham, R. B., and M. J. Karnovsky 1966 The early stages of lectins with plasma membrane components. Biochim. of absorption of injected horseradish peroxidase in the Biophys. Acta., 419: 119-136. proximal tubules of mouse kidney: ultrastructural cyto-
raised by work on unfixed cells from sea urchin embryos, (Roberson et al., '75). The present results appear to show no difference in the overall lectin affinity of the dorsal surface of the epiblast cells actively invaginating through the primitive streak compared to non-invaginating cells. I t is possible, however, t h a t other lectins with different carbohydrate specificity would show such a difference, which has been indicated by lanthanum (Sanders and Zalik, '72). In the latter study, lanthanum, which binds to cell surface material without a particular glycoprotein specificity, was shown to deposit more heavily on the surface of invaginating epiblast cells in the primitive groove a t stage 5, than on non-invaginating ones. The slight, but consistent, increase in the lectin affinity of the dorsal surface of the epiblast from stage 1 to stage 4-5, shown by ferritin, may reflect the overall increase in motility of these cells as they prepare to undergo the morphogenetic migrations of gastrulation. A similar change in the lectin binding characteristics of cells about to undergo gastrulation has been demonstrated in amphibian embryos (Johnson and Smith, '76) and the relationship between this change and altered cell mobility and adhesiveness was discussed. It is likely that chick embryo epiblast cells undergo a similar surface change in preparation for invagination, and t h e increase in con A binding T to gastrulation may be an expression of this alteration.
N
Q,
6
+
+ +. Microvilli are
+ +. X 64,500. Stage 1. Ventral surface of the epiblast, posteriorly. Ferritin-con A deposit + + +. Note the particles a t -
Stage 1. Ventral surface of the epiblast, anteriorly. Con A-HRP-DAB reaction product + +. Note the penetration of staining between the cells. The lower layer of cells (hypoblast) is incomplete a t this stage. X 37,100.
tached to the basal lamina. X
64,100.
5 Stage 1. Dorsal surface of the epiblast, posteriorly. Ferritin-con A deposit
4
3 Stage 1. Dorsal surface of the epiblast, anteriorly. Con A-HRP-DABreaction product seen and are commonly found on this surface. X 37,100.
2 Control Ferritin-con A reaction carried out in the presence of aMM. Stage 4-5, ventral surface of the endoblast beneath the primitive streak showing no reaction product. X 64,100.
1 Control. Con A-HRP-DAB reaction carried out in the presence of aMM. Stage 4-5, ventral surface of the epiblast anterior region, showing no reaction product. BL, basal lamina. All micrographs unstained. X 37,100.
EXPLANATION OF FIGURES
PLATE 1
w
Q,
CON A BINDING BY CHICK EMBRYO CELLS S . L. Hook and E. J. Sanders PLATE 1
PLATE 2 EXPLANATION OF FIGURES
7 Stage 1. Con A-HRP-DAB reaction on a cell from t h e incomplete hypoblast layer in the anterior region of the embryo. Note that the staining on the ventral surface (single arrow) is greater than that dorsally (double arrow) E-epiblast cell. Since t h e hypoblast layer is incomplete, lack of penetration is not the cause of this non-uniformity. X 6,600.
8 Stage 1.The dorsal surface of a hypoblast cell from the anterior region of the blasComparc with figure 9. X 37,100. toderm. Con A-HRP-DAB rcoction product
+.
9 Stage 1. The ventral surface of a hypoblast cell from the anterior region of the blastoderm. Con A-HRP-DAB reaction product +. Compare with figure 8. X 37,100.
+
10 Stage 4-5. Dorsal surface of the epiblast from a peripheral region of t h e embryo. Con A-HRP-DAB reaction product + +. X 37,100.
++
+, and shows aggregates of particles. A cell process from one epiblast cell is looping over to contact the adjacent one. X 64,500.
11 Stage 4-5. Same as figure 10. Ferritin-con A deposit
64
CON A BINDING BY CHICK EMBRYO CELLS S. L. Hook and E. J. Sanders
PLATE 2
PLATE 3 EXPLANATION OF FIGURES
12 Stage 4-5. Ventral surface of the endoblast below the primitive streak and repre. sentative of all regions of this surface a t this stage. Con A-HRP-DAB reaction product 0. X 18,200. 13 Stage 4-5. Same as figure 12. Ferritin-con A deposit
+. X
64,500.
14 Stage 4-5. Dorsal surface of the epiblast at the base of the primitive groove. Ferritincon A deposit Portions of several epiblast cells are seen in the process of invaginating. X 64,500.
+ + +.
66
CON A BINDING BY CHICK EMBRYO CELLS S. L. Hook and E. J. Sanders
PLATE 3
67
ABSTRACT The surfaces of cells from the early embryo of the chick were examined using electron microscope techniques for the visualization of concanavalin A-binding sites. Horseradish peroxidase and Ferritin labelled concanavalin A were used to determine the distribution of the binding sites. All surfaces of the epiblast and hypoblast layers which were accessible to concanavalin A showed the presence of binding sites in stage 1embryos. The ventral surface of the epiblast showed a high lectin affinity which may reflect the development of a basal lamina on this surface. The individual hypoblast cells a t this stage showed a non-uniform distribution of binding sites, having a greater affinity on the dorsal surface than the ventral. By the time of primitive streak formation (stage 45) the dorsal surface of the epiblast displayed increased binding sites, while the frequency of sites on the ventral surface of the endoblast was reduced. The latter may reflect a change from one cell population to another, which occurs in the lower layer of the embryo at this time. No consistent correlation could be drawn between changes in motility of cells actually invaginating through the primitive streak and changes in affinity for concanavalin A. An overall increase in affinity of the dorsal surface of the epiblast was revealed by Ferritin and may reflect the changes in surface structure occurring in readiness for the morphogenetic migrations of gastrulation. The cell surface plays an important role in the control of cell growth, cell movement and cell recognition, and is consequently intimately involved in differentiation and morphogenesis. The early chick embryo provides a good system for the study of cell surface changes accompanying differentiation and morphogenetic migration. Previous work on the surface properties of these cells has characterized their electrophoretic mobility (Zalik e t al., '721, the binding of markers to surface ionogenic groups (Sanders and Zalik, '72) and the affinity of the cells for calcium ions (Harris and Zalik, '74). More recent work has dealt with the reaggregation and sorting out of dissociated chick cells taken from early developmental stages. It has been shown that the surfaces of blastoderm cells are sufficiently differentiated to possess characteristics which allow them to sort out into two cell populations, (Zalik and Sanders, '74; Eyal-Giladi e t al., '75; Sanders and Zalik, '76). The molecular basis for the expression of J. CELL. PHYSIOL., 93: 57-68.
cell surface differences appears to lie in surface receptors of a carbohydrate nature (see review of Hughes, '73). Therefore, lectins such as concanavalin A (con A) which bind to specific surface glycoproteins can be used as probes with which to study changes in the distribution of such receptor sites, (Nicolson, '74). That con A binds to, and agglutinates, cells from the early chick blastoderm has been demonstrated, by Zalik and Cook ("76).These authors, and also Kleinschuster and Moscona ('72), have shown that as development proceeds the availability of lectin binding sites is altered, as judged by agglutinability, probably by the masking effects of trypsin-sensitive material at the cell surface. It was the purpose of this study to elucidate the distribution of con A binding sites (CABS) on surfaces of cells from the early chick embryo. The relative numbers of binding sites were compared on surfaces of corresponding cells in prestreak and gastrulating embryos Received Feb. 1, '77. Accepted Apr. 7, '77.
57
58
S. L. HOOK AND E. J. SANDERS
by the use of ultrastructural cytochemical techniques. A change in the binding of con A on any surface was taken to indicate alterations in the complement of cell surface receptor sites. It was considered that such alterations could in turn relate t o the onset of morphogenetic migrations in that particular cell group or be indicative of the presence of different populations of cells within the embryo. Some aspects of the con A affinity of early amphibian embryo cells (O'Dell et al., '74; Sanders and DiCaprio, '76; Johnson and Smith, '76) and chick embryo cells at later stages (Lee e t al., '76) has been described previously. MATERIALS AND METHODS
Stage 1 and stages 4-5blastoderms (Hamburger and Hamilton, '51) were removed from the vitelline membrane and placed in Pannett and Compton's saline. Stage 1 blastoderms were marked by removing a wedge of tissue to facilitate orientation at the time of embedding. Embryos were given a thorough wash with 0.1 M cacodylate buffer, pH 7.4, and fixed for two hours in 2.5% buffered glutaraldehyde. For detection of CABS with horseradish peroxidase (HRP), a modification of Bernhard and Avrameas' ('71) method was employed using Graham and Karnovsky's ('66) diaminobenzidine (DAB) technique. The blastoderms were washed in buffer and incubated in 100 pg/ml con A (Sigma), in cacodylate buffer pH 7.2 for two hours, washed again in buffer and incubated in 50 p g/ml horseradish peroxidase (Sigma Type VI) for 30 minutes. After another buffer wash they were incubated in a saturated aqueous solution of 3,3'diaminobenzidine hydrochloride (J. T. Baker analyzed reagent) and 0.01%hydrogen peroxide in 0.5 M Tris-HC1 buffer, pH 7.6 for 14 to 30 minutes followed by a further wash in buffer. Control embryos were incubated in con A plus 0.5 M a methyl-D-mannoside (aMM, Sigma Grade 111) and HRP with 0.5 M a M M added. All of the above was carried out a t 22°C. Further controls were done in the absence of con A to eliminate the possibility of nonspecific binding of HRP-DAB reaction product to the cell surface. The first group of control blastoderms was incubated directly after glutaraldehyde fixation with 50 pg/ml HRP for 30 minutes rinsed and placed in a saturated solution of DAB with 0.01%hydrogen peroxide
in Tris-HC1buffer for 15 minutes. A second set of embryos was incubated in saturated DAB in Tris-HC1 buffer for 15 minutes and in 3 X M potassium ferricyanide for five minutes. Since potassium ferricyanide readily oxidizes DAB, mimicking the effect of hydrogen peroxide, this control was to determine whether non-specific adsorption of DAB or its oxidation product occurred. A third set of blastoderms was incubated in saturated DAB and 0.01% hydrogen peroxide in Tris-HC1 buffer. Ferritin conjugated covalently to con A (Calbiochem A grade) was also used to detect CABS. After fixation in 2.5%phosphate buffered glutaraldehyde pH 7.4 the blastoderms were thoroughly washed in buffer and incubated in approximately 100 pg/ml ferritin-con A for one hour, and then thoroughly washed in buffer again. All embryos were then post-fixed in 1%phosphate buffered osmium tetroxide for one hour at 4OC. All blastoderms were dehydrated in an ethanol series, followed by propylene oxide and embedded in araldite. Stage 1 embryos were sectioned in two areas: anteriorly and posteriorly. Stage 4 and 5 embryos were sectioned in an anterior region, either anterior to Hensen's node or through Hensen's node, and then through the anterior third of the primitive streak. Some embryos were sectioned in a third region, in the posterior two-thirds of the primitive streak. RESULTS
Con A-HRP-DAB produced a dense reaction product which did not penetrate the plasma membrane. The density of reaction product was evaluated on the basis of the thickness of t h e deposit and whether i t displayed a clumped or even pattern of binding. The scoring of the reaction product was a follows: 0, no reaction product; , a sparsely scattered thin layer; +, a denser layer in which most or all of the membrane has reaction product; and +++, a very thick deposit over the entire membrane. Ferritin-con A was seen as a particulate deposit, limited to the outer surface of the plasma membrane. Generally, these particles displayed a nonuniform distribution, appearing singly or in clumps separated by areas of membrane with no molecules bound. A score of 0 was given if no reaction product was found, if particles of ferritin were found
+
+
+
59
CON A BINDING BY CHICK EMBRYO CELLS TABLE 1
Summary of results Stage 4.5
Stage 1 Anterior
Posterior
Anterior
Concanavalin A - Horseradish peroxidase - Diaminobenzidine + f ++ DE +++ 0 VE 0 0-+ ’ 0 DN 0 +-++’ +-++I VN 0 Ferritin - Concanavalin A ++ ++ DE +++ VE 0-+ or 0 +-+ DN 0-+ or 0 +-+ VN 0- +
++
DE VE DH VH DE VE DH VH
+++ + +-++
+++
+-++ +
Posterior
++’ 0 0 0
+++ 0-+ or 0 0-+ or 0 0-
+
See text for evaluation procedure and explanation of symbols. DE, dorsal surface of the epiblast; VE, ventral surface of the epiblast; DH, dorsal surface of the hypoblast; VH, ventral surface of the hypoblast; DN, dorsal surface of the endohlast; VN, ventral surface of the endoblast. ‘ S e e text for further details.
++
widely spaced on the membrane, if partiif the cles were closely adjacent, and membrane was largely covered by ferritin particles or aggregates. The symbol 0 is used in the table to indicate that although no reaction was present this may have reflected the inaccessibility of the surface to the cytochemical procedure. Generally the control procedures using aMM showed no binding of reaction product (figs. 1, 2).
cells showed a non-uniform distribution of deposit. The ventral surface of individual cells possessed denser deposits than the dorsal surface of the same cell (figs. 7-91. The ventral surface of the hypoblast was always stained but the pattern varied in density (+-++ +). Ferr-con A results are summarized in table 1. Sections taken from anterior regions of stage 1did not usually have a complete hypoblast, but there was no difference between anterior regions and posterior regions in the Stage 1 pattern of ferr-con A binding despite the At this stage of development the embryo difference in the completion of the hypoblast. consists of one layer of dorsally situated cells Ferr-con A appeared to penetrate more easily (epiblast) and a more or less complete lower between the cell layers than the HRP-DAB layer of cells (hypoblast). The hypoblast be- markers. The deposits shown by ferr-con A comes a coherent layer in the posterior region were more consistent than those found with initially and spreads anteriorly by migrating con A-HRP-DAB and in agreement with the on the ventral surface of the epiblast. HRP result, the dorsal surface of the epiblast Although within each experiment some var- showed a less dense deposit ( + + ; fig. 51, than ; fig. iability existed between embryos with respect the ventral surface of the epiblast ( to corresponding surfaces, the following pat- 6). Single hypoblast cells showed a sparse tern resulted with con A-HRP-DAB (table 1). deposit on their surface with no striking In stage 1 embryos all surfaces showed stain- changes in binding intensity between the ing except where a complete layer of hypo- various aspects of individual cells. In sections blast was found in the posterior regions of taken from posterior regions the dorsal and the embryo. In these areas no reaction product ventral surfaces of the hypoblast had ferritin was found on the ventral surface of the epi- binding similar to that in anterior regions blast or the dorsal surface of the hypoblast (0). (+-+ +). Except for this, no differences were found beStage 4-5 tween the embryos sectioned in anterior and This stage of development is characterized posterior regions. The dorsal surface of the epiblast showed a fairly dense deposit ( + + ; by the primitive streak, and a complete lower fig. 31, and the ventral surface of the epiblast layer of cells, (endoblast). showed a very dense clumped pattern in the Stage 4-5 embryos with con A-HRP-DAB anterior region ( + +; fig. 4). If the hypo- (table 1) displayed binding on the dorsal surblast was incomplete, individual hypoblast face of the epiblast in anterior regions ( + +;
+++
+++
+
60
S. L. HOOK AND E. J. SANDERS
fig. 10).No reaction product was found on the ventral surface of the epiblast or the dorsal surface of the endoblast in any area sectioned, (0). The ventral surface of the endoblast did not bind reaction product in either region (0; fig. 12). In areas taken from the primitive streak, two distinct patterns of reaction product were found on the dorsal surface of the epiblast. In one, no reaction product was found on the walls of the groove or on the base of the groove. The absence of reaction product in the primitive groove was not seen in all embryos, where a dense clumped layer of reaction product was observed with no discontinuity. The reason for this variation was not readily apparent. Stage 4-5 embryos with ferr-con A (table 1) displayed a heavier binding on the dorsal surface of the epiblast (+++; fig. l l ) , than stage 1, with ferritin particles often aggregated. No differences in the affinity of this surface for lectin were detected with respect to the region of the blastoderm examined. The surface of cells in the primitive streak (fig. 14) showed the same deposits as cells more peripherally situated. In both areas sectioned the vcntral surface of the cpiblnst and the dorsal surface of the endoblast displayed little or no reaction product, ( O d + or 0 ) . The ventral surface of the endoblast in the anterior region and beneath the primitive streak displayed a pattern with a few isolated particles or no reaction product (fig. 13; 0-+). DISCUSSION
Although two different cytochemical techniques were used in this study, the results given by each were similar. The main discrepancy was that of the pattern seen on individual hypoblast cells, since the non-uniform distribution seen with HRP-DAB did not occur with ferr-con A. The denser reaction product seen on the ventral surface compared to the dorsal surface with HRP-DAB was probably the result of the very high activity of HRP and the amplification effect which occurs during oxidation of the DAB (Temmink et al., '75). Thus, what may have been a small difference in binding capacity between these two surfaces was not detected by Ferr-con A but, owing to the amplification, was demonstrated by HRP-DAB. The denser deposit on the dorsal surface may reflect the presence of an otherwise indistinct basal lamina, not previously observed. The most immediate difference between the
two stages of development examined was the affinity of the ventral surface of the lower layer. In stage 1 this surface showed a consistent binding of the lectin, which was absent a t stage 4-5. This change is probably a reflection of the different populations of cells that make up the lower layer of the blastoderm during early morphogenesis. In stage 1, the lower layer consists of groups of cells which aggregate into a single layer to form the hypoblast, (Vakaet, '70; Eyal-Giladi and Kochav, '76). At stage 5, these cells have moved peripherally, and have been replaced by the definitive endoblast (Vakaet, '701, which develops into the embryonic endoderm (Rosenquist, ' 66). Bellairs et al. ('77) have demonstrated behavioural differences between hypoblast and endoblast cells which may derive from cell surface differences such as those demonstrated here, and their different lectin affinities may provide a further means of distinguishing them. If the definitive endoblast, which did not bind con A, originates by the movement of cells from the epiblast, (Vakaet, TO), which bound con A on its dorsal and ventral surfaces, the receptor sites may be masked or removed from the cell surface after invngination. Wolk et al. ('74) reported that the antigenic specificity of primary hypoblast cells a t stage 1 is not present on cells of the lower layer a t stage 5. Furthermore, these authors showed that the hypoblast cells were not uniformly stained with fluorescent antibody, in that case being most intense on the dorsal surface. This result is similar to the present one, where con A also appears to bind nonuniformly. Differences were also seen in the ventral surface of the epiblast. However, the marked change from stage 1 embryos, with a very dense staining, to no staining in stage 5 may result from inaccessibility to the lectin. During stage 1, a basal lamina is laid down on the ventral surface of the epiblast, which can be seen as a diffuse amorphous deposit, (Low, '67). Martinez-Palomo ('70) reported that this structure was stained with ruthenium red, which marks acid mucopolysaccharides, a t the primitive streak stage. The present study confirmed the presence of mannose andlor glucose residues in this structure as early as stage 1. The possibililty that con A binding sites are more mobile within the membrane of migratory cells than of non-migratory cells, resulting in different patterns of distribution, has been
CON A BINDING BY CHICK EMBRYO CELLS
61
chemistry by a new technique. J. Histochem. Cytochem., 14: 291-302. Hamburger, V., and H. L. Hamilton 1951 A series of normal stages in the development of the chick embryo. J. Morph., 88: 49-92. Harris, H. L., and S. E. Zalik 1974 Studies on the surface of chick blastoderm cells. 111. Calcium binding to ionogenic sites. J. Cell. Physiol., 83:359-368. Hughes, R. B. 1973 Glycoproteins as components of cellular membranes. Prog. Biophys. Molec. Biol., 26: 189268. Johnson, K. E., and E. P. Smith 1976 The binding of concanavalin A to dissociated embryonic amphibian cells. Exptl. Cell Res., 101:63-70. Kleinschuster, S. J., and A. A. Moscona 1972 Interactions of embryonic and fetal neural retina cells with carbohy drate-binding phytoagglutinins: cell surface changes with differentiation. Exptl. Cell Res., 70: 397-410. Lee, H.-Y., J. B. Sheffield, R. G. Nagele and G. W. Kalmus 1976 The role of extracellular material in chick neurulation. I. Effects of concanavalin A. J. Exp. Zool., 198:261266. Low, F. N. 1967 Developing boundary (Basement) membranes in the chick embryo. Anat. Rec., 159:231-238. Martinez-Palomo, A. 1970 The surface coats of animal cells. Int. Rev. Cytol., 29: 29-75. Nicolson, G. L. 1974 The interactions of lectins with animal cell surfaces. Int. Rev. Cytol., 39: 89-190. ODell, D. S., R. Tencer, A. Monroy and J. Brachet 1974 The pattern of concanavalin A binding sites during the early development of Xenopus loeuis. Cell Differentiation, 3: 193-198. Roberson, M., A. Neri and S. B. Oppenheimer 1975 Distribution of concanavalin A receptor sites on specific populations of embryonic cells. Science, 189: 639-640. Rosenquist, G. C. 1966 A radioautographic study of labelled grafts in the chick blastoderm. Contr. Embryol. Carneg. Inst., 38: 71-113. Sanders, E. J., and R. A. DiCaprio 1976 A freeze-fracture and concanavalin A binding study of the membrane of cleaving Xenopus embryos. Differentiation, 7:13-21. 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: 235247. ACKNOWLEDGMENTS Sanders, E. J., and S. E. Zalik 1976 Aggregation of cells from early chick blastoderms. Differentiation, 6: 1-11. This work was supported by an operating Temmink, J. H. M., J. G. Collard, H. Spits and E. Roos 1975 grant from the Medical Research Council of A comparative study of four cytochemical detection Canada to E. J. Sanders. methods of concanavalin A binding sites on the cell membrane. Exptl. Cell Res., 92: 307-322. Vakaet. L. 1970 Cineohotomicroeraohic investieations LITERATURE CITED of gastrulation in the chick blktbderm. Arc(. Biol. Bellairs, R., P. A. Portch and E. J. Sanders 1977 Behaviour (Liege), 81: 387-426. patterns of chick blastoderm cells. J. Anat., in press. Wolk, M., M. Schlesinger and H. Eyal-Giladi 1974 The deBernhard, W., and S. Avrameas 1971 Ultrastructural velopment of and migration of the primary hypoblast visualization of cellular carbohydrate components by cells in the chick embryo studied by immunofluorescent means of Concanavalin A. Exptl. Cell Res., 64: 232-236. technique. Isr. J. Zool.,23: 204-205. Eyal-Giladi, H., S. Kochav and A. Yerushalmi 1975 The Zalik, S. E., E. J. Sanders and C. Tilley 1972 Studies on the sorting-out of thymidine-labelled chick hypoblast cells in surface of chick blastoderm cells. 1. Electrophoretic mixed epiblast-hypoblast aggregates. Differentiation, 4: mobility and pH mobility relationships. J. Cell. Physiol., 57-60. 79:225-234. Eyal-Giladi, H., and S. Kochav 1976 From cleavage to Zalik, S. E., and E. J. Sanders 1974 Selective cellular primitive streak formation: a complementary normal affinities in the unincubated chick blastoderm. Differtable and new look a t the first stages of the development entiation, 2: 25-28. of the chick. Develop. Biol., 49: 321-337. Zalik, S. E., and G. M. W. Cook 1976 Comparison of early embryonic and differentiating cell surfaces. Interaction Graham, R. B., and M. J. Karnovsky 1966 The early stages of lectins with plasma membrane components. Biochim. of absorption of injected horseradish peroxidase in the Biophys. Acta., 419: 119-136. proximal tubules of mouse kidney: ultrastructural cyto-
raised by work on unfixed cells from sea urchin embryos, (Roberson et al., '75). The present results appear to show no difference in the overall lectin affinity of the dorsal surface of the epiblast cells actively invaginating through the primitive streak compared to non-invaginating cells. I t is possible, however, t h a t other lectins with different carbohydrate specificity would show such a difference, which has been indicated by lanthanum (Sanders and Zalik, '72). In the latter study, lanthanum, which binds to cell surface material without a particular glycoprotein specificity, was shown to deposit more heavily on the surface of invaginating epiblast cells in the primitive groove a t stage 5, than on non-invaginating ones. The slight, but consistent, increase in the lectin affinity of the dorsal surface of the epiblast from stage 1 to stage 4-5, shown by ferritin, may reflect the overall increase in motility of these cells as they prepare to undergo the morphogenetic migrations of gastrulation. A similar change in the lectin binding characteristics of cells about to undergo gastrulation has been demonstrated in amphibian embryos (Johnson and Smith, '76) and the relationship between this change and altered cell mobility and adhesiveness was discussed. It is likely that chick embryo epiblast cells undergo a similar surface change in preparation for invagination, and t h e increase in con A binding T to gastrulation may be an expression of this alteration.
N
Q,
6
+
+ +. Microvilli are
+ +. X 64,500. Stage 1. Ventral surface of the epiblast, posteriorly. Ferritin-con A deposit + + +. Note the particles a t -
Stage 1. Ventral surface of the epiblast, anteriorly. Con A-HRP-DAB reaction product + +. Note the penetration of staining between the cells. The lower layer of cells (hypoblast) is incomplete a t this stage. X 37,100.
tached to the basal lamina. X
64,100.
5 Stage 1. Dorsal surface of the epiblast, posteriorly. Ferritin-con A deposit
4
3 Stage 1. Dorsal surface of the epiblast, anteriorly. Con A-HRP-DABreaction product seen and are commonly found on this surface. X 37,100.
2 Control Ferritin-con A reaction carried out in the presence of aMM. Stage 4-5, ventral surface of the endoblast beneath the primitive streak showing no reaction product. X 64,100.
1 Control. Con A-HRP-DAB reaction carried out in the presence of aMM. Stage 4-5, ventral surface of the epiblast anterior region, showing no reaction product. BL, basal lamina. All micrographs unstained. X 37,100.
EXPLANATION OF FIGURES
PLATE 1
w
Q,
CON A BINDING BY CHICK EMBRYO CELLS S . L. Hook and E. J. Sanders PLATE 1
PLATE 2 EXPLANATION OF FIGURES
7 Stage 1. Con A-HRP-DAB reaction on a cell from t h e incomplete hypoblast layer in the anterior region of the embryo. Note that the staining on the ventral surface (single arrow) is greater than that dorsally (double arrow) E-epiblast cell. Since t h e hypoblast layer is incomplete, lack of penetration is not the cause of this non-uniformity. X 6,600.
8 Stage 1.The dorsal surface of a hypoblast cell from the anterior region of the blasComparc with figure 9. X 37,100. toderm. Con A-HRP-DAB rcoction product
+.
9 Stage 1. The ventral surface of a hypoblast cell from the anterior region of the blastoderm. Con A-HRP-DAB reaction product +. Compare with figure 8. X 37,100.
+
10 Stage 4-5. Dorsal surface of the epiblast from a peripheral region of t h e embryo. Con A-HRP-DAB reaction product + +. X 37,100.
++
+, and shows aggregates of particles. A cell process from one epiblast cell is looping over to contact the adjacent one. X 64,500.
11 Stage 4-5. Same as figure 10. Ferritin-con A deposit
64
CON A BINDING BY CHICK EMBRYO CELLS S. L. Hook and E. J. Sanders
PLATE 2
PLATE 3 EXPLANATION OF FIGURES
12 Stage 4-5. Ventral surface of the endoblast below the primitive streak and repre. sentative of all regions of this surface a t this stage. Con A-HRP-DAB reaction product 0. X 18,200. 13 Stage 4-5. Same as figure 12. Ferritin-con A deposit
+. X
64,500.
14 Stage 4-5. Dorsal surface of the epiblast at the base of the primitive groove. Ferritincon A deposit Portions of several epiblast cells are seen in the process of invaginating. X 64,500.
+ + +.
66
CON A BINDING BY CHICK EMBRYO CELLS S. L. Hook and E. J. Sanders
PLATE 3
67
