The accessibility of certain proteins on embryonic chick neural retina cells to iodination and tryptic removal is altered by calcium

The inclusion of Ca²⁺+ during extensive trypsinization of embryonic chick neural retinae permits the preparation of cells that have afunctional Ca²⁺-dependent adhesive system (Takeichi, Ozaki, Tokunaga & Okada, 1979; Grunwald, Geller & Lilien, 1980; Magnani, Thomas & Steinberg, 1981; Brackenbury, Rutishauser & Edelman, 1981). These cells are able to form histo-typically organized aggregates in the absence of detectable protein synthesis (Grunwald et al. 1980). In contrast, cells prepared by extensive trypsinization in the absence of divalent cations require protein synthesis for the formation of intercellular adhesions (Moscona & Moscona, 1965; Grunwald et al. 1980); thus, they lack a functional adhesive system.

Both Chinese hamster V79 cells (Takeichi, 1977) and BHK cells (Urushihara, Ueda, Okada & Takeichi, 1977) also retain a functional Ca²⁺-dependent adhesive system following tryptic digestion in the presence of Ca²⁺+. Thus, this effect of Ca²⁺ appears to be a general one. One possible interpretation of these results is that Ca²⁺ prevents the tryptic digestion of certain cell surface polypeptides. Consistent with this interpretation, a protein of M_r approx. 130 ×10³ has been identified on V79 cells that, in the presence of Ca²⁺, is relatively inaccessible to both tryptic removal and iodination (Takeichi, 1977; Urushihara & Takeichi, 1980).

To gain further insight into the effects of Ca²⁺ on cell-surface architecture and to identify cell-surface^ proteins unique to cells trypsinized in the presence of Ca²⁺+, as possible candidates for components of the Ca²⁺-dependent adhesive system, we have used cell-surface-specific labelling techniques and two-dimensional gel electrophoresis to identify proteins on embryonic chick neural retina cells and to investigate the effects of Ca²⁺ on their accessibility to labelling and tryptic removal.

Nonidet P40 was obtained from Bethesda Research Labs. (Rockville, Md, U.S.A.). 1,3,4,6-tetrachloro-3α,6α-diphenylglycoluril (lodogen) was purchased from Pierce Chemical Co. (Rockford, Ill., U.S.A.). Acrylamide (electrophoresis grade), N-N’-methylenebisacrylamide (electrophoresis grade) and sodium dodecyl sulphate (SDS, electrophoresis grade) were obtained from Bio Rad Labs. (Richmond, Calif., U.S.A.). Urea (ultra pure) was supplied by Schwarz/Mann Div., Becton, Dickinson & Co. (Orangeberg, N.Y., U.S.A.). AG 2–11 Servalytes were obtained from Serva Feinbiochemica (Heidelberg, FRG). Antipan and leupeptin were gifts of Dr Philip Cohen and the US-Japan Cooperative Cancer Research Programme through Dr Walter Troll. Sodium diatrizoate (Hypaque) was purchased from Winthrop Labs. Div. of Sterling Drug Inc. (New York, U.S.A.). Plasma albumin (bovine) was obtained fromReheis Chemical Co. (Chicago, Ill., U.S.A.), β-galactosidase (Escherichia coli) was supplied by Worthington Biochemical Corp. (Freehold, N.J., U.S.A.). Neuraminidase (Vibrio cholerae) was obtained from Calbiochem (Los Angeles, Calif.). Phenylmethylsulphonyl fluoride (PMSF), N-2-hydroxythyl-piper-azine-N’-2-ethanesulphonic acid (Hepes), ethylene glycol-bis(β-aminoethyl ether)-N,N’-tetraacetic acid (EGTA), ovalbumin (chicken), ribonuclease A (RNase, bovine pancreas, protease-free), galactose oxidase (Dactylium dendroides) and lactoperoxidase (bovine) were obtained from Sigma Chemical Co. (St Louis, Mo., U.S.A.). Trypsin (bovine pancreas, three times crystallized, Ca²⁺-free) was purchased from Miles Labs., Inc. (Elkhart, Ind., U.S.A.). Deoxyribonuclease (DNase, bovine pancreas) was obtained from Calbiochem-Behring Corp. (San Diego, Calif., U.S.A.); contaminating proteolytic activity was abolished by treatment with PMSF as previously reported (Grunwald et al. 1980). Carrier-free Na¹²⁵I (iodination grade) and flTlpotassium borohydride (1·13 Ci/mmol) were purchased from New England Nuclear, (Boston, Mass. U.S.A.). X-ray film (SB-5 and XR-5) was supplied by Eastman Kodak Co. (Rochester, N.Y., U.S.A.).

HBSGK contains 137mM-NaCl, 5 mM-KCl, 10 mM-glucose and 10 mM-Hepes and is adjusted to the indicated pH with NaOH. HBSGKMg refers to HBSGK supplemented with 1 mM-MgCl₂. HBSGKCa refers to HBSGK supplemented with 1 mM-CaCl₂. HBSGKMgCa refers to HBSGK supplemented with 1 mM-MgCl₂ and 1 mM-CaCl₂. HBSKMgCa refers to HBSGKMgCa lacking glucose.

Dissection of neural retinae from 9-day chick embryos was performed in Tyrode’s solution (Tyrode, 1910) adjusted to a pH of approx. 6·5 with CO₂. Prior to Ca²⁺-free iodination, retinae were freed of Ca²⁺ by three washes with 0·5 ml per retina of ice-cold HBSGKMg (pH 7·4) followed by a 10 min incubation at 37 °C in 0·5 ml per retina of HBSGKMg (pH 7·4) containing 2 mw-EGTA and two washes with 0·5 ml per retina of ice-cold HBSGKMg (pH 7·4). Iodination was performed for 20 min on ice in 0·2 ml per retina of HBSGKMg (pH 7·4) containing 2 μm-Na¹²⁵(5 mCi/ml) in a glass pibe bearing 10 μg per retina of 1,3,4,6-tetrachloro-3α,6α-diphenylglycoluril (lodogen) as a thin film; the retinae were periodically resuspended by gentle swirling. (lodogen is virtually insoluble in water; films were prepared by evaporation from a 1 mg/ml solution in methylene chloride.) Iodination was terminated by transfer of the retinae to a fresh tube containing 1·0 ml per retina of ice-cold HBSGKMgCa (pH 7·4). After three washes with 0·5 ml per retina of ice-cold HBSGKMgCa (pH 7·4),the retinae were re-equilibrated with Ca*+ by a 30 min incubation at 37 °C in 0·5 ml per retina of HBSGKMg (pH 7·4) containing 2 mM-Ca²⁺ and washed three times with 0·5 ml per retina of ice-cold HBSGKMgCa (pH 7·4).

Retinae were iodinated in the presence of Ca*²⁺ by a modification of the above procedure. After two washes with 0·5 ml per retina HBSGKMgCa (pH 7·4), iodination was performed, otherwise as above, in HBSGKMgCa (pH 7·4). The iodinated retinae were washed three times with 0·5 ml per retina of ice-cold HBSGKMgCa (pH 7·4). lodogen was prepared and shown to mediate cell-surface-specific iodination by Fraker & Speck (1978). Markwell & Fox (1978) have verified the surface specificity of labelling.

Retina cell suspensions were prepared as described below. Prior to Ca²⁺-free iodination, cells were washed twice with 20 vol. of cold HBSGKMg (pH 7·0) containing 5 μg/ml cycloheximide, 2 μg/ml colchicine and 50 μg/ml antipain. The cells were then pelleted and resuspended in 2 vol. of an ice-cold solution of lactoperoxidase at 2 mg/ml in HBSGKMg (pH 7·0) containing the above additives. The cell suspension was rapidly warmed to 25 °C and 1 vol. of the same medium containing 8 μM-Na^l25I (20 mCi/ml) and 200 μM-H₂,O₂, (previously warmed to 23° C) was added with rapid mixing. After 1·5 min at 25 °C, labelling was terminated by dilution of the reaction mixture into 20 vol. ice-cold HBSGKMgCa (pH 7·0) containing 50 μg/ml antipain and 5 mM-KI. Iodinated cells were washed three times with 20 vol. ice-cold HBSGKMgCa (pH 7·0) containing 50 μg/ml antipain and 5 mM-KI. Iodination of single cell suspensions in the presence of Ca²⁺ was performed as above with 1 mM-Ca*⁺ present in all solutions.

Prior to their labelling, retina cells, prepared as described below, were washed twice with 20 vol. HBSGKMgCa (pH 7·0) containing 5 μg/ml cycloheximide, 2 μg/ml colchicine and 50 μg/ml antipain. Following incubation for 30 min at 25 °C in 20 vol. of the above medium containing 2 mM-borohydride, the cells were washed three times with 20 vol. of medium and incubated for 30 min at 25 °C in 10 vol. of medium containing 200 μg/ml galactose oxidase. Cells were freed of bound galactose oxidase by a wash with 20 vol. of ice-cold 0·3 M-galactose containing 10 mM-Na.HEPES (pH 7·0), 1 mM-CaCl,, 5 μg/ml cycloheximide, 2 μg/ml colchicine and 50 μg/ml antipain and three washes with 20 vol. of ice-cold medium. Labelling was accomplished by a 30 min incubation at 25 °C in 10 vol. of medium containing approx. 1·8 mM-[³H]borohydride (approx. 2 mCi/ml). After being labelled, the cells were washed three times with ice-cold medium.

After washing twice with 20 vol. of HBSKMgCa (pH 7·0) containing 5 μg/ml cycloheximide, 2 μg/ml colchicine and 50 μg/ml antipain, cells prepared as described below were incubated for 30 min at 25 °C in 20 vol. of the above medium containing 2 mM-borohydride. The cells were then washed three times with 20 vol. of ice-cold medium and resuspended in 20 vol. of ice-cold medium containing O’1 mM, 1 mM or 10 mM-periodate. After 10 min on ice, oxidation was terminated by three washes with 20 vol. ice-cold medium containing 10 mM-glucose. Periodate-generated aldehydes were reduced by a 30 min incubation at 25 °C in 10 vol. medium containing 10 mM-glucose and approx. 1·8 mM-[³H]borohydride (approx. 2 mCi/ml). After labelling, cells were washed three times with 20 vol. of medium containing 10 mM-glucose.

Cell suspensions (TRP cells) were prepared by tryptic digestion of neural retinae from 9-day chick embryos following published procedures (Grunwald et al. 1980), with slight modifications. Retinae were washed three times, incubated 10 min at 37 °C and washed twice, in each case with 1 ml per retina of HBSGK (pH 7·4). Retinae were then digested for 25 min at 37 °C in 1 ml per retina of HBSGK (pH 7·4) containing 4 × 10³ National Formulatory units/ml (approx. 1·1’3 mg/ml) trypsin. After three gentle washes with 1 ml per retina of ice-cold HBSGKMgCa (pH 7·0), the trypsinized retinae were dissociated in 1 ml per retina of HBSGK-MgCa (pH 7·0) containing 2 mM-PMSF and 5 ×10³ Domase units/ml (approx. 50 μg/ml) DNase and washed twice with 20 vol. HBSGKMgCa (pH 7·0).

Cell suspension were also prepared by trypsinization of retinae in the presence of Mg;³⁺ or Ca*+ (Grunwald et al. 1980). Either 1 mM-Mg²⁺ or 1 mM-Ca²⁺ was included during predigestion manipulation and tryptic digestion of the tissues. Cells prepared by tryptic digestion in the presence of Ca*²⁺ are referred to as CaT cells.

Cells were washed twice with 20 vol. of ice-cold HBSGKCa (pH 7·4) and digested for 5 min at 37 °C in 10 vol. of HBSGKCa (pH 7·4) containing 4 × 10³ National Formulatory units/ml trypsin (approx, ci mg/ml). After digestion, cells were washed once with 20 vol. HBSGKMgCa (pH 7·0) containing 5 ·10³ Domaseunits/ml (approx. 50 μg/ml) DNase and 2 mM-PMSF and twice with 20 vol. HBSGKMgCa (pH 7·0) containing 50 μg/ml antipain.

Prior to digestion, the cells were washed with 20 vol. of ice-cold Co₂-saturated Tyrode’s solution (pH ∽ 6·0) containing 5 μg/ml cycloheximide, 2 μg/ml colchicine and 50 μg/ml antipain. Digestion was performed for 30 min at 25 °C in 10 vol. of this medium containing 50 units/ml neuramidase. After digestion, cells were washed three times with 20 vol. of ice-cold HBSGKMgCa (pH 7·0) containing 5 μg/ml cycloheximide, 2 μg/ml colchicine and 50 μg/ml antipain.

Cells were resuspended in 20 vol. (500 μl) of ice-cold HBSGKMgCa (pH 7·0) and layered over 3’0 ml of an ice-cold solution containing 85 mg/ml bovine serum albumin, 46 mg/ml (72 him) sodium diatrizoate, 65 mM-NaCl, 5 m-KI, 1 mM-MgCl,, 1 mM-CaCl₂, 10 mM-glucose and 10 mM-Na.HEPES (pH 7·0). After centrifugation at 2× 10³g for 10 min at 4 °C, the cells that had banded at the interface and those that had pelleted were removed separately and washed three times with 20 vol. of ice-cold HBSGKMgCa (pH 7·0).

Cells were washed three times with 20 vol. of ice-cold HBSGKMg (pH 7·4) containing 5 μg/ml cycloheximide and 2 μg/ml colchicine and incubated for 15 min at 37 °C in 20 vol. of the above medium. As noted for each experiment, either 1 mM-Ca*²⁺ or 50 μg/ml antipain was included in the medium. When called for, Ca²⁺-free incubated cells were re-equilibrated with Ca²⁺ by a 30 min incubation at 37 °C in 20 vol. of HBSGKMgCa (pH 7·0) containing 5 μg/ml cycloheximide, 2 μg/ml colchicine and 50 μg/ml antipain.

Approximately one half-retina or an equivalent number of cells (approx. 3×10⁷) was washed with 500 μl ice-cold HBSGKMgCa (pH 7·0) containing 50 μg/ml antipain and 50 μg/ml leupeptin, and lysed on ice in 100μl of the above medium containing 1% Nonidet P40, too μg/ml DNase, 50 μg/ml RNase and 2 mM-PMSF. After 1 h on ice, the lysate was precipitated with 500 μ1 ice-cold ethanol and held at— 20 °C until analysed. Samples were analysed within 2 weeks of preparation and generally within 1 week. No changes in the protein patterns during storage were detected.

Two-dimensional electrophoretic analysis of cell proteins was performed by a modification of published procedures (O’Farrell, 1975). Isoelectric focusing (IEF) gels contained 2·85% acrylamide, 0·15% methylenebisacrylamide, 9·5 M-urea, 4% Nonidet P40 and 2% ampholytes (pH 2–11). Before use, the gels were pre-focused for 2 h at 100 V with an overlay of sample buffer (9 M-urea, 4% Nonidet P40, 5% 2-mercaptoethanol, 2% ampholytes (pH 2–11), 2 mM-PMSF, 50 μg/ml antipain and 50 μg/ml leupeptin).

Ethanol-precipitated samples were centrifuged (2 min at 10⁴g) and each pellet was resuspended in 100μl IEF sample buffer (see above) containing 20 μg/ml bromphenol blue. After being heated for 5 min at 90–100 °C, the samples were centrifuged (5 min at 10⁴g) and the supernatants applied to the pre-focused gels. Focusing was performed for a total of 6·5 × 10³ V h.

The isoelectric focusing gels were removed from the tubes and equilibrated at 25 °C for 30 min each in three 5-ml changes of 63 mM-Tris-HCl (pH 6·8) containing 5% 2-mercaptoethanol and, successively, 5%, 4% and 3% SDS. After equilibration, the gels were held at— 20 °C under 2-merceptoethanol-saturated N₂.

SDS/polyacrylamide slab gels using the discontinuous buffer system of Laemmli (1970) were employed for the second-dimensional separations. The slabs were cast with 30:1, acrylamide/methylenebisacrylamide, increasing from approx. 5% polyacrylamide at the top to 10% at the bottom; the stacking gels contained 3% polyacrylamide. Focusing gels were thawed, heated for 5 min at 100 °C and sealed to the stacking gels with 63 mM-Tris-HCl (pH 6·8) containing 1% agarose, 3% SDS and 5% 2-merceptoethanol. Electrophoresis was performed at 10 mA per slab.

The slabs were stained for 1 h at 25 °C in 250 ml 0·05% Coomassie brilliant blue in 5:1·5, methanol/acetic acid/water and destained in 10% acetic acid. Gels of iodinated samples were soaked for 30 min in 500 ml water, 1 h in 200 ml 10% glycerol and 15 min in 500 ml water prior to drying under heat and vacuum on Whatman 3 MM paper. Gels of tritiated samples were soaked for 30 min in 500 ml water and 40 min in 500 ml 1 M-sodium salicylate prior to drying; this has been shown to be an effective procedure for fluorographic enhancement (Chamberlain, 1979). Dried slabs were exposed to X-ray film (Kodak SB-5 or, in later experiments, XR-5) at— 80 °C in a press.

Three molecular weight standards were employed; these were E. coli /β-galactosidase at 116 × 10³M_r (Fowler & Zabin, 1977), bovine plasma albumin at 66 × 10³M_r (Brown, 1976) and chicken ovalbumin at 43 × 10³M_r (Castellino & Barker, 1968). The parallel analysis of a mixture of these proteins and a sample of iodinated retinae by SDS/polyacrylamide gel electrophoresis permitted positioning of the standard proteins on two-dimensional gels relative to retina proteins detected by autoradiography or Coomassie brilliant blue staining.

The pH calibration was performed in the following way. Three gels of unlabelled retinae were included in an isoelectric focusing run. After focusing the gels of the unlabelled retinae were cut into 5-tnm sections; each section was equilibrated for 5 h under N, in 0-5 ml of 50 mM-KCl. The KC1 solution was prepared with H,O that had been boiled and stored under N₃. Determination of the pH of each section defined the pH versus distance relationship for that focusing run. After the second-dimension slab gels of the iodinated retina cells were run, stained and autoradiographed, pH values were marked on the stained gels and on tracings of the autoradiograms. In this way, the positions of defined pH values were determined relative to retina proteins detected by autoradiography or Coomassie brilliant blue staining.

Particular labelled proteins were identified on each autoradiograph by their positions relative to other labelled proteins and those proteins detected on the gels by Coomassie brilliant blue staining. The stained slab was aligned with its autoradiograph on the basis of the glycine/chloride boundary, the material that remained at the origin in both dimensions and, when present, the labelled proteins that were also detected by Coomassie brilliant blue staining. Tracings of the patterns on clear plastic sheets were used to facilitate comparisons.

Quantitative comparisons were made only within experiments; i.e. among samples from the same preparation of retinae, which were labelled, electrophoressed and autoradiographed as a set. Relative intensities were judged visually; minor differences in intensity were not considered significant.

The resolution of iodinated retinae by two-dimensional gel electrophoresis reveals that labelled proteins are, in general, minor components not detected on the gel by protein staining (Fig. 1, compare c with a). Tryptic digestion of iodinated retinae removes the majority of the labelled proteins (Fig. 1, compare d with c) but has only minor effects on the protein staining pattern (Fig. 1, compare B with a). Thus, it appears that retina cell-surface proteins are preferentially iodinated by this procedure.

The inclusion of Ca²⁺+ during the trypsinization of iodinated retinae alters the pattern of labelled proteins subsequently seen (Fig. i, compare E with d), but has no detectable effect on the protein staining pattern (data not shown). The primary effect of including Ca²⁺ during trypsin digestion is the retention of labelled d and I. Other consequences of the inclusion of Ca²⁺+ during the trypsinization of iodinated retinae are summarized in Table 1. Labelled h, I and n appear to be generated during Ca²⁺ trypsinization of iodinated retinae; they are presumably tryptic fragments of other labelled proteins.

The consequences of including Mg²⁺’ during tryptic digestion of iodinated retinae are distinct from those seen when Ca²⁺ is present (Fig. i, compare f with E and both with d). The inclusion of Mg²⁺ does not significantly increase or decrease the retention of any protein detectably iodinated in intact retinae (Fig. i, compare f with E and both with c). Those proteins retained during trypsinization when it is performed in the presence of Mg²⁺ thus appear to be fragments of labelled proteins whose further cleavage or loss from the cells is blocked by Mg²⁺.

To ensure that iodination of intact retinae did not result in the labelling of a select subset of cell surface proteins, retina cell suspensions were labelled and analysed. Two techniques were used for surface-specific labelling: lactoperoxidase-mediated iodination and oxidation with galactose oxidase followed by reduction with [³H]-borohydride.

Lactoperoxidase-mediated iodination of retina cell suspensions preferentially labels certain minor cell proteins to high specific activity when a high enzyme concentration and short reaction time are employed (compare Fig.1 with Fig. 2 B). Separation of iodinated cells on the basis of permeability to sodium diatrizoate revealed that label was associated with proteins that were detected by protein staining only in permeable cells (data not shown). Thus, this method appears to be largely surface-specific.

Comparison of proteins iodinated on cells prepared from retinae by Ca²⁺+-free trypsination (TRP cells) with those iodinated on cells prepared by tryptic digestion in the presence of Ca²⁺ (CaT cells) reveals a number of differences (Fig. 2, compare Awith C). The major difference when iodination is performed after digestion is the presence of much more labelled d on CaT cells than on TRP cells; other proteins labelled to a greater extent on CaT cells than on TRP cells are listed in Table i. The sensitivity of these proteins to tryptic cleavage in the absence of Ca²⁺ further confirms their surface localization.

Few sugar residues oxidized by galactose oxidase (presumably galactose or N-acetyl-D-galactosamine) are accessible to this enzyme on either TRP cells or CaT cells; however, one glycoprotein, t, is labelled to a certain extent on both types of cells (data not shown). Prior digestion of the cells with neuraminidase, however, permits the labelling of a number of glycoproteins; of these, only d is uniquely labelled on CaT cells (Fig. 3, compare b with a). None of the labelled glycoprotein is detected by protein staining of the gels.

On the basis of the experiments described above, the major cell-surface species whose removal during trypsinization is blocked by Ca²⁺ (<) is a glycoprotein of approx. 130 × 10³M_r with a pl of approx. 4·8. Its oligosaccharides contain galactose and/or 1V-acetyl-D-galactosamine residues that are not accessible to galactose oxidase prior to digestion of the cells with neuraminidase. Terminal sialic acid residues have been found to prevent the oxidation of subterminal galactose residues by galactose oxidase (Gahmberg, Itaya & Hakomori, 1976). However, periodate oxidation of CaT cells followed by reduction with [³H]borohydride labels d only when a high concentration (10 mM versus 1 mM) of periodate is employed (data not shown). This finding implies that d is not heavily sialated, inasmuch as sialic acid residues are preferentially oxidized by periodate at low concentration (Gahmberg & Andersson, 1977).

The inclusion of Ca²⁺ during the iodination of intact retinae alters the efficiency with which certain species are labelled (Fig. 4, compare B with a). The most obvious alteration is a reduction in the amount of labelled d when the retinae are iodinated in the presence of Ca²⁺+. Other consequences of including Ca²⁺ during iodination are summarized in Table 2.

Similarly, the inclusion of Ca²⁺ during the iodination of CaT cells reduces the efficiency with which d and I are labelled (Fig. 2, compare d with c). Other consequences of including Ca²⁺ during the iodination of CaT cells are summarized in Table 2. The inclusion of Ca²⁺+ during the iodination of TRPcells reduces the efficiency with which I and the small amount of residual d are labelled (Fig. 2, compare B with A).

Thus, certain retina cell surface proteins, most particularly the approx. 130 × 10³ M_rglycoprotein d, are less accessible to both iodination and tryptic cleavage in the presence of Ca²⁺. While digestion of CaT cells with neuraminidase renders galactose and/or TV-acetyl-D-galactosamine residues on d accessible to galactose oxidase, it has no effect on the accessibility of d to lactoperoxidase (data not shown). Conversely, the accessibility of the galactose and/or 7V-acetyl-D-galactosamine residues of d on CaT cells to galactose oxidase does not depend on the presence or absence of Ca²⁺, whether or not the cells are first digested with neuraminidase (data not shown).

Labelled d, g and n are lost when CaT cells derived from Ca²⁺-free iodinated retinae are incubated at 37 °C in the absence of Ca²⁺ (Fig. 5, compare B with a). The inclusion of antipain during Ca²⁺-free incubation prevents the loss of labelled d, but not that of labelled g or n (Fig. 5, compared c with b). It appears that an antipainsensitive protease is associated with CaT cells and is responsible for the loss of labelled d when these cells are incubated at 37 °C in the absence of Ca²⁺or antipain. Indeed, it has been shown that trypsin remains associated with cells prepared by these techniques (Grunwald et al. 1980).

Ca²⁺-free incubation of CaT cells at 37 °C in the presence of antipain permits the efficient labelling of several species, most notably d and I, during subsequent iodination of the cells in the presence of Ca⁸⁺ (Fig. 6, compare A with b; also refer to Fig. 2C and d). In other words, the inclusion of Ca²⁺ during the iodination of CaT cells does not reduce the efficiency with which d and I are labelled if, prior to their iodination, the cells are incubated at 37 °C in the absence of Ca²⁺ (with antipain present to block proteolysis). However, a second incubation of the cells at 37 °C in the presence of Ca²⁺(and antipain) restores the ability of Ca²⁺+ to block the efficient labelling of d during subsequent iodination (Fig. 7, compare c with b and a). Thus, the presence of Ca²⁺during the iodination of CaT cells is not sufficient to block the efficient labelling of d rather, some time- and/or temperature-dependent interaction of the cells with Ca²⁺ is apparently required.

Analogous results are obtained when the accessibility of d on CaT cells is probed with trypsin. When CaT cells are iodinated, incubated at 37 °C and then digested with trypsin, all in the continual presence of Ca²⁺,labelled d, g and n are retained (Fig. 8 a). However, the exclusion of Ca²⁺+ and the inclusion of antipain during incubation of the cells at 37 °C results in the loss of labelled d, g and n during subsequent trypsinization in the presence of Ca²⁺-(Fig. 8, compare b with a). A second incubation of the cells at 37 °C in the presence of Ca²⁺+ (and antipain) restores the ability of Ca²⁺ to block subsequent tryptic removal of labelled d and g (Fig. 8, compare c with b).

We have detected a number of proteins on cells in the neural retinae of 9-day chick embryos that, in the presence of Ca²⁺+, are relatively inaccessible to tryptic removal (see Table 1) or iodination (see Table 2). Among these, an approx. 130 × 10³M_r glycoprotein (d) with a pl of approx. 4·8 is the major species detected by iodination. The removal of Ca²⁺ renders this glycoprotein much more accessible to both procedures. Furthermore, its accessibility to these probes decreases subsequent to the re-addition of Ca²⁺

It is interesting and of possible significance that all of the polypeptides whose iodination and/or tryptic removal is blocked by Ca²⁺ have approximately the same pl. However, the accessibility of other species with similarly low pl values (e.g., a and j) is relatively independent of the presence of Ca²⁺.

While the physical basis for these changes in accessibility to tryptic removal and iodination is not clear, there are two obvious alternatives: a generalized change in surface architecture or changes specific to individual polypeptides. With respect to a more generalized effect, ligatin, an ubiquitous approx. 10 x io³M_r protein that interacts with Ca²⁺ and is present on embryonic chick neural retina cells (Jakoi & Marchase, 1979), may play a role. Whatever the mechanism, the behaviour of d provides an interesting case in point.

The possibility of a reversible Ca²⁺-dependent internalization of d is excluded by the Ca²⁺d sensitive to galactose oxidase, does not change the accessibility of d to iodination in the presence or absence of Ca²⁺.

The behaviour of this cell-surface glycoprotein implies that it may be a component of the Ca²⁺-dependent adhesive system of embryonic chick neural retina cells. The characteristics of this adhesive system are reviewed below and summarized in Table 3 together with the relevant behaviour of d.

The inclusion of Ca²⁺ during trypsinization of embryonic chick neural retinae permits the preparation of cells (referred to here as CaT cells), which have a functional Ca²⁺+-dependent adhesive system (Takeichi et al. 1979; Grunwald et al. 1980; Magnani et al. 1981; Brackenbury et al. 1981). Incubation of CaT cells at 37 °C in the absence of Ca²⁺ renders their subsequent aggregation dependent on protein synthesis (Grunwald et al. 1980). This loss of adhesive competence does not occur in the presence of antipain, an inhibitor of trypsin-like proteases (Grunwald et al. 1980). One aspect of this requirement for antipain may be the inhibition of trypsin that remains associated with these cells following dissociation (Grunwald et al. 1980).

CaT cells that have never been freed of Ca²⁺+ retain the Ca²⁺+-dependent adhesive system during tryptic digestion in the presence of Ca²⁺+ (Grunwald et al. 1980). However, Ca²⁺-free incubation of the cells at 37 °C in the presence of antipain renders the adhesive system sensitive to subsequent trypsinization in the presence of Ca²⁺+ (Grunwald et al. 1980). Thus, it is not the presence of Ca²⁺per se, but rather some consequence of its interaction with CaT cells, that maintains the adhesive system in a trypsin-insensitive state.

The presence of the 130 ×10³M_r glycoprotein (d) on embryonic chick neural retina cells correlates with their possession of a functional Ca²⁺-dependent adhesive system (see Table 3). Indeed, immunological data presented in the accompanying paper (Grunwald et al. 1982) and summarized below strongly suggest that d is a component of the Ca-²⁺dependent adhesive system present on CaT cells. The ability of Fab’ fragments of an anti-CaT cell immunoglobulin G (IgG) preparation to inhibit the aggregation of CaT cells is blocked by medium conditioned by embryonic chick neural retinae. Extensive purification of this blocking activity has yielded a fraction that specifically blocks the recognition by anti-CaT cell IgG of both d and I on CaT cells. The major component of this fraction that is recognized by anti-CaT cell IgG is an approx. 85 × 10³M_r protein with a pl of approx. 4·8. Preliminary one-dimensional peptide mapping by limited proteolysis of iodinated d and the 85 ×10³ M_r/4·8 protein with Staphlococcus V8 protease (Cleveland, Fischer, Kirschner & Laemmli, 1977) suggests that the 85 ×10³ M_rprotein is a fragment of d (Cook, unpublished).

This work was supported by a grant from the N.S.F. io J.L., and J.H. C. was partially supported by an N.I.H. training grant in Cellular and Molecular Biology.