Eur. J. Biochem. 96, 295-300 (1979)

Quantitative Fluorimetric Determination of Cell-Surface Glycoconjugates with Fluorescein-Substituted Lectins

Michel MONSIGNY, Claude SENE, and Angele OBRENOVITCH

Centre de Biophysique Moleculaire, Centre National de la Recherche Scientifique and Universite d'orledns

(Received November 16, 1978)

Fluorescein-substituted lectins, which can be used to visualize cell surface glycoconjugates, are shown to be usable in the quantitative determination of the number of receptor sites and of their association constant. The fluorescence measurements of the fluorescein-substituted lectins released from the cell surface with the related inhibitor, give quantitative data in a large range of fluorescein- substituted lectin concentration (0.1 to 100 pg/ml).

Using fluorescein-substituted concanavalin A or [3H]acetyl-concanavalin A, it was found that baby hamster kidney cells (BHK 21, wild-type) bind 10 k 2 x lo6 lectin molecules per cell with an apparent association constant of 1.8 or 1.7 x lo6 1 x mol-', respectively. Using the fluoresceinyl and [3H]acetyl-substituted wheat germ agglutinin, we found 40 f 5 x lo6 sites per cell with an apparent binding constant of 1 and 1.3 0.3 x lo6 1 x mol-', respectively. When fluorescein-substituted succinyl wheat germ agglutinin was used instead of the unsuccinylated wheat germ agglutinin derivatives, the number of binding sites was reduced 7 times, while the binding constant was very slightly lowered. Concanavalin A derivatives gave monotonic Scatchard plots; on the opposite, wheat germ agglutinin derivatives gave biphasic Scatchard plots suggesting that wheat germ agglutinin binds to two classes of receptors.

Lectins are proteins able to bind specific carbo- hydrate moieties, and are currently used to investigate properties of cell surface glycoconjugates [I -31. Quantitative determination of the number of glyco- conjugates which bind a given lectin are usually carried out using lectins radioactively labeled by substitution with ['4C]acetic anhydride or [3H]acetic anhydride or by iodination with 1251 or 1 3 1 1 [4-61. Radioactively labeled lectins may also be used to visualize their glyco- conjugates receptors by the means of autoradiography techniques [ 7 - 91.

Fluoresceinyl derivatives of lectins are currently used to visualize and to study the dynamics of cell surface glycoconjugates [lo- 131. In this paper, we present data showing that such fluorescent lectins may be also used to determine the number of cell surface glycoconjugate receptors and the apparent bind- ing constants. The results concerning the binding of fluoresceinyl derivatives of wheat germ agglutinin and of concanavalin A on baby hamster kidney cells (BHK 21 C 13) are compared with those obtained with the [3H]acetyl derivatives of these lectins.

MATERIALS AND METHODS

Cell Culture

Baby-hamster-kidney fibroblasts (BHK 21 C 13, wild type) were obtained from Dr R.C. Hughes (Nation- al Institute for National Research, Mill Hill, London) and grown at 37°C in Glasgow modified minimal essential medium supplemented with 10 (v/v) foetal bovine serum (Rehatuin, F.S. Armour, Pharm. Comp. U.S.A.) tryptose phosphate broth, NaHC03 (2 g/l) and gentamycin (50 pg/ml) [14]. Cells were removed from culture vessels by incubation for 1-2 min at 37 "C in 0.02 7; (v/v) Na2 EDTA in phosphate-buffered saline (pH = 7.4). Before use, suspended cells were washed three times with phosphate-buffered saline containing Ca2+ and Mg2+ ; their viability was evalu- ated by the trypan blue exclusion.

Fixation of' Cells tvith Glutaualdehyde

3 volumes of cell suspension (10' cells per ml) were incubated for 30 min with 1 volume of 40; glutaral-

296 Quantitative Fluoresceinyl-Lectin Binding to Cell-Surface Glycoconjugates

dehyde in phosphate-buffered saline, pH 7.4. After incubation, cells were washed twice with phosphate- buffered saline and resuspended in 10 mM glycin- amide in phosphate-buffered saline. Before any bind- ing experiments with lectins, cells were washed twice with phosphate-buffered saline.

Preparation of Labeled Lectins

Concanavalin A and wheat germ agglutinin (Pharmindustrie, I.B.F. RCactifs, Clichy, France) were labeled with [3H]acetic anhydride (800 Ci/mol) as described by Riordan and Vallee [15]. After purifica- tion by affinity chromatography (see below) labeled concanavalin A and wheat germ agglutinin were in a dimeric form, according to ultracentrifugation deter- minations (molecular weights of 52000 and 36000, respectively). Specific activity of [3H]concanavalin A was 18000 counts x min-' x pg-', and of 3H-labeled wheat germ agglutinin was 5000 counts x min-' x pg protein-'.

Fluorescein-substituted lectins were purchased from Pharmindustrie or prepared in our laboratory by using fluorescein-isothiocyanate following the method described by Loor [15]. The sedimentation constants of purified fluoresceinyl-lectins correspond to the tetrameric concanavalin A and dimeric wheat germ agglutinin.

Succinyl-fluoresceinyl-concanavalin A was pre- pared from fluoresceinyl-concanavalin A as described by Gunther et al. [17] for concanavalin A. Succinyl- fluoresceinyl-wheat germ agglutinin was prepared from fluoresceinyl-wheat germ agglutinin as for wheat germ agglutinin [ 181. Succinylated fluoresceinyl-con- canavalin A and fluoresceinyl-wheat germ agglutinin had molecular weights of 52000 and 36000, respec- tively as estimated by ultracentrifugation analysis.

Before use, the labeled lectins were purified by affinity chromatography, concanavalin A on a Seph- adex G-50 column [19] and wheat germ agglutinin on a column of Ultrogel A-4 substituted with p-amino- benzyl- 1 - thio-2-acetamido- 1,2-dideoxy-fi-~-gluco- pyranoside [20]. Concentrations of lectins used in the experiments were calculated on the basis of A: Fm = 11.4 at 280 nm for concanavalin A and At Fm = 15 for wheat germ agglutinin.

Fluorescence Measurements

Fluorescence measurements were performed using a high resolution differential spectrofluorimeter FICA M K I1 (Fica, France) equipped with a Houston 2000 recorder. Fluorescein derivatives were excited at 495 nrn and their emission was detected at 520 nm. All the measurements were carried out at 25 "C. The base line was adjusted at 650 nm.

Determination of the Number of Binding Sites

Fluorescent Lectins. Binding of fluoresceinyl-lectins was performed on sub-confluent cells at a final concen- tration of 2 x lo6 cells/ml phosphate-buffered saline. The incubation with fluoresceinyl-lectins was carried out at different concentrations of fluoresceinyl-lectins varying from 0.5 to 100 pg/ml, for 1 h at 4°C. At the end of the incubation, cells were centrifuged (1000 x g, 5 min) and the concentration of free fluoresceinyl- lectins in the supernatant was estimated by fluores- cence measurements.

Labeled cells were washed three times at room tem- perature with phosphate-buffered saline. Specifically bound lectin was released by incubating the cells for 1 h at 4°C with 1 ml of a suitable sugar inhibitor (0.3 M) : a-0-methyl-D-mannopyranoside with conca- navalin A, and N-acetyl-D-glucosamine with wheat germ agglutinin. Concentration of fluoresceinyl-lectins released from the cells was calculated from its emission fluorescence intensity ; standard curves were deter- mined in the same experimental conditions with a large range of fluoresceinyl-lectins concentrations. The ap- parent number of binding sites per cell was then evalu- ated from a Scatchard plot [21].

Radioactively Labeled Lectins. These determina- tions were simultaneously conducted with the same batch of cells, by applying the above method. Aliquots of supernatants (3 x 0.3 ml) were mixed with scintisol- complet solvant (Scintix) and the radioactivity was counted in a Beckman Tri-Carb liquid scintillation counter.

Fluorescence Microscopy

Fluorescence microscopy was carried out with a Zeiss microscope equipped with Osram HBO Super Pressure Mercury Lamp.

RESULTS AND DISCUSSION

The concentration of fluoresceinyl-lectins was determined on the basis of their absorbances at 280 nm and at 495 nm. The ratio of the absorbances measured at 495 nm and at 280 nm of fluoresceinyl thiocarbamyl ethanolamine was found to be 3. There- fore, the concentration to fluoresceinyl-lectins was calculated on the basis of the protein contribution of the 280 nm absorbance as follows :

,4280n.m - ,4280.nm - protein - lectin 1 -3 (A22i:m).

The excitation and emission wavelengths selected on the basis of the excitation and emission spectra of fluoresceinyl-lectin (Fig. I), were 495 nm and 520 nm, respectively. The base line was adjusted at 650 nm where neither light scattering nor fluorescence emis- sion occurred.

M. Monsigny, C. Sene, and A. Obrenovitch

400 450 500 550 600 650 X (nm)

Fig. 1. Absolute r.ucitation (----) and emission (- -) spectsa of fluoresceinyl thiocasbamyl derivative of wheat germ agglutinin (50 pglml) in phosphate-buffered saline, p H 7.4 at 2S"C. I f : relative intensity of fluorescence; i: wavelength in nm. Excitation spectrum was run with an emission wavelength centered at 520 nm. Emission spectrum was run with an excitation wavelength centered at 495 nm. The width of the bandpass was 7.5 nm. Base line was adjusted at 650 nm

The 520 nm fluorescence intensity of fluoresceinyl- lectin was used to determine the lectin concentration. The fluorescence intensity was found to be depend on fluoresceinyl-lectin concentration in a very large range of concentrations: 0.01 to 100 pg/ml (Fig. 2). Furthermore, it was found that the fluorescence inten- sity was very little affected by the presence of the sugars used as lectin inhibitors (Fig. 2). Nevertheless, the standard plots of fluorescence intensities versus fluo- resceinyl-lectins concentration were done in the pres- ence of the suitable sugar inhibitor.

BHK 21 C 13 cells were collected from a sub-con- fluent cell culture after treatment with EDTA in phos- phate-buffered saline, and three washings with EDTA- free phosphate-buffered saline. Preliminary experi- ments showed that the cell-binding of fluoresceinyl- concanavalin A and fluoresceinyl-wheat germ aggluti- nin was completed within 30 min at 4 "C as it was with [3H]acetyl-concanavalin A and [3H]acetyl-substituted wheat germ agglutinin. All the binding experiments were conducted at 4 "C in order to avoid capping and endocytosis [ S ] which occurred when the experiments were conducted at 25 "C or 37 "C. After incubation of BHK cells in the presence of fluoresceinyl-lectins, cells were washed three times with phosphate-buffered saline for 5 min. No appreciable amount of fluores- ceinyl-lectin were found in the supernatant of the last washing, showing that the bound lectins are not in true equilibrium with the surrounding medium or at least that the releasing process is very slow.

Attempts to determine the amount of cell-bound fluoresceinyl-lectins using a direct fluorescence mea- surements of the cell suspension did not give accurate

291

I

q 10

5

0 0 50 DO

Concentration of derivative, c (pgi rnl)

Fig. 2. Dqwndence of the fluorescence intensity on the conwitrution of the ,fluosesceinyl tliiocarhamyl derivative of' wheat gesm agglutinin. c, Concentration of fluoresceinyl thiocarbamyl derivative of wheat germ agglutinin in pglml. If: relative intensity of fluorescence. Buffer: phosphate-buffered saline, pH 7.4 at 25 "C (0-0) in the absence of. ('I -'I) in the presence of 0.3 M N-acetyl-D- glucosamine. Inscri: Data obtained at low concentration of the fluoresceinyl-thiocarbdmyl derivative of wheat germ agglutinin

results. The light scattering of the cell suspension was too high to lead to consistent results. Therefore, all the experiments were conducted by measuring the fluores- ceinyl-lectins released by incubation of washed cells in the presence of the sugar inhibitors.

The binding of fluoreseinyl-lectins to the cells and the release of fluoresceinyl-lectins from the cells were also analyzed in a qualitatively way using a fluores- cence microscope. The fluorescence of BHK cells incubated in the presence of fluoresceinyl-concana- valin A and washed with phosphate-buffered saline was concentrated in clusters at the cell periphery (Fig. 3 A). After incubation in the presence of cr-methyl- mannopyranoside for 1 h, cells were slightly fluores- cent, showing that some of the lectin was not released by the inhibitor (Fig. 3B). This cannot be attribute to an endocytosis mechanism because the experiments were conducted at 4 "C and because cells pretreated with glutaraldehyde and control cells exhibited the same quantitative binding capacity. The non-release could be attributed to unspecific binding or to the presence of receptors with a very high binding constant. Indeed, when cells were preincubated with unlabeled concanavalin A in the presence of a-methylmanno- pyranoside, washed with phosphate-buffered saline and finally incubated with fluoresceinyl-concanavalin A (Fig. 3C), the cell fluorescence completely dis- appeared upon subsequent incubation in the presence of the inhibitor. Similar results were obtained with fluoresceinyl-wheat germ agglutinin.

Using 3H-labeled concanavalin A and 3H-labeled wheat germ agglutinin after a pretreatment with the corresponding unlabeled lectin, it was found that the

298 Quantitative Fluoresceinyl-Lectin Binding to Cell-Surface Glycoconjugates

I I

0.0 5 rnm Fig. 3 . .tlic i-o.vc.opc' ohsc~rvatioris of the Jluorescence of' huh). hamstc~r kirlncy cc~lls trewtcvl wirh the ,f~uorc~.sceiny/-thiocurhamj~/ derivative of c~oncunuvnlin A under various conditions. Magnification: 630. (A) Total binding: baby hamster kidney cells were incubated for 1 h at 4 °C in the presence of the fluoresceinyl thiocarbamyl derivative of concanavalin A (50 pg/ml) and then washed three times with phosphate- buffered saline, pH 7.4 at 4°C. (B) Unspecific binding: baby hamster kidney cells were prepared as under A but were then incubated in the presence of 0.3 M O-methyl-cc-D-mannopyranoside, for 1 h at 4 'C. (C) Specific binding: baby hamster kidney cells were incubated (1 h at 4 C) with the unlabeled concanavalin A (50 pgjml) in the presence of 0.3 M O-methyl-sc-D-mannopyranoside, washed once in the presence of the same inhibitor, twice with phosphate-buffered saline, pH 7.4 at 4°C and then treated as under A. When the procedure B was applied to pretreated cells, no trace of fluorescence could be detected on the cells

labeled lectin not released upon treatment with the inhibitory sugar, was less than 5 % of the amount adsorbed, whereas when the cells were not pretreated with unlabeled lectins, the non-released material was about 25 with concanavalin A and 40 with wheat germ agglutinin. These findings are in good agreement with the results of other authors [22].

Quantitative binding experiments were conducted in parallel with 3H-labeled and fluoresceinyl-lectins. The experimental data, plotted according to Scatchard [21] are shown in Fig. 4 with fluoresceinyl-concanava- lin A and in Fig.5 with fluoresceinyl wheat germ agglutinin.

The numbers of binding sites and the apparent binding constant of fluoresceinyl-concanavalin A were very close to those of 3H-labeled concanavalin A (Table l), and are in good agreement with the number of binding sites (9 x lo6) found by Meager et al. [14]. So, while the molecular weight of fluoresceinyl-con- canavalin A is twice that of 3H-labeled concanavalin A (Table 2), both labeled lectins have the same binding properties. However, when succinyl-fluoresceinyl-con- canavalin A which has the same molecular weight as the [3H]acetyl-concanavalin A was used, the number of binding sites was about half, while the apparent binding constant was similar. This low number of binding sites may be related to the more acidic nature of succinylated-concanavalin A and to the presence of a considerable amount of bound sialic acids on the cell surface of BHK cells [23]. The repulsion of succi- nyl-concanavalin A from the negatively charged cell surface can explain its decreased binding capacity.

Scatchard plots obtained with fluoresceinyl-con- canavalin A, [3H]acetyl-concanavalin A and succinyl-

r

m

0 1 2 3 4 Bound (pg1 2-IO6 cel ls )

Fig. 4. Scatchard plot o/ tlrc . s i w c ~ i / i c ~ hiruling of' thi, .fliiorescc~i~7yl thiocarhamyl clerivutivc, of c~oncunmulin A to huh). liunzsrev kidnc.1, cells. Bound: amount (pg/2 x 10" cells) of the fluoresceinyl-thio- carbamyl derivative of concanavalin A bound to the cells and specifically released in the presence of 0.3 M O-methyl-r-I,-manno- pyranoside. Free: concentration (pgjml) of the free fluoresceinyl- thiocarbdmyi derivative of concanavalin A. Cells (2 x 10') were incubated (1 h, 4 ' C in 1 ml of phosphate-buffered saline pH 7.4 containing the fluoresceinyl-thiocarbdmyl derivative of concana- valin A (concentration range 5 - 100 pg/ml). The bound lectin was released by incubation of the labeled cells in the presence of0.3 M 0-methyl-a-u-mannopyranoside (1 h, 4 C). Line is fitted by the least square method

fluoresceinyl-concanavalin A, were always monotonic, in the range of lectin concentrations used (Fig. 4), sug- gesting a single class of binding sites. N o evidence for any negative or positive cooperativity was found with BHK cells in contrast to the finding of Bornens et al. [24] in the case ofconcanavalin A binding to rat thymo-

M. Monsigny, C. Sene, and A. Obrenovitch 299

I \ 5

0 2 4 6 Bound (pg /2-106 cells)

Fig. 5 . Scatchard plot of tlir 5pec+c binding ~~fthe,~uore.sceinvl-tliio- curhmmjl derivative of wheat germ agglutinin. Conditions are described under Fig. 4 except the bound lectin was released from cells by using 0.3 M N-acetyl-D-glucosamine

Table 1 . Concanavalin A binding sites and their association constants on BHK cells The number of binding sites per cell and the values of apparent asso- ciation constants, were evaluated by Scatchard plot [21] following the conditions described in Materials and Methods. Molecular weights. evaluated from ultracentrifugations were 1 x l o 5 for fluoresceinyl-concanavalin A and [3H]acetyl-concanavalin A and 5 x 10" for succinyl-concanavalin A

Concanavalin A Concanavalin A lo-' x x derivative bound Siteslcell Association

constant

pg/106 cells I x m o l - '

Fluoresceinyl 1.7 k 0.30 10 k 2 1.8 [3H]Acetyl 0.85 & 0.15 10 k 2 1.7 Succinyl-

fluoresceinyl 0.35 & 0.05 4.2 k 0.6 2.0

Table 2. Setlinirtitrrtioti velocity e.uperiment.s Sedimentation coefficients (SZO.~) of unlabeled and labeled concana- valin A and wheat germ agglutinin dissolved in phosphate-buffered saline (pH 7.4) with an absorbance of A z x o ~ , , , = 1, were obtained at 20 C from ultracentrifugation analysis at the velocity of 60000 rev:min.

Lcctin S ~ O . of derivative

fluoresceinyl- fluoresceinyl- ["Hlacetyl- succinyl-

Concanavalin A 6.1 4.26 4.2 Wheat germ

agglutinin 3.5 3.5 3.6

' The A4q5r!A2~() ratio was 0.15. At 4 ' C , the sedimentation coef- ficient of fluoresceinyl-concanavalin A obtained was 6.06 in good agreement with a tetrameric state, contrarily to that of unlabeled concanavalin A which is dimeric at 4 C

Table 3. Wheat germ agglutinin binding sites and their as.sociation constants on BHK cells Binding studies were performed in the same conditions as those described for fluoresceinyl-concanavalin A. The molecular weight of [3H]acetyl-substituted wheat germ agglutinin and of succinyl- fluoresceinyl-wheat germ agglutinin was 3.6 x lo4. Data are the mean values of experiments in triplicate

Wheat germ lo-' x Sites/cell lo-' x Association agglutinin derivative constant

1 x mol-'

Fluoresceinyl 42 k 4 1 * 0.2 [3H]Acetyl 40 5 1.3 0.3 Succinyl-fluoresceinyl 6 2 0.7 & 0.2

cytes, or of Feller et al. [25] in the case of concanavalin A binding to human fibroblasts.

The effect of charges on binding data is even more pronounced in the case of wheat germ agglutinin. Indeed, results obtained with fluoresceinyl wheat germ agglutinin and [3H]acetyl-substituted wheat germ agglutinin which have the same molecular weights (Table 2) are very similar (Table 3). Succinylated- fluoresceinyl wheat germ agglutinin has a similar molecular weight [18], but has a p l of 4.5 instead of 8.5 for wheat germ agglutinin [18]. The number of binding sites of succinylated fluoresceinyl wheat germ agglutinin is dramatically lower (7 times) than that of fluoresceinyl wheat germ agglutinin or [3H]acetyl- substituted wheat germ agglutinin. Because the sugar affinity of succinylated wheat germ agglutinin was close to that of unsuccinylated wheat germ agglutinin [18], it is assumed that the very large binding site difference is related to the large change of wheat germ agglutinin isoelectric point.

I t is quite interesting to notice that the Scatchard plots obtained using fluoresceinyl wheat germ agglu- tinin, [3H]acetyl-substituted wheat germ agglutinin and also succinyl-fluoresceinyl-wheat germ agglutinin were biphasic (Fig. 5). This behaviour is generally attributed to heterogeneity of receptor sites with two or several classes of sites which differ in their binding affinity.

Concluding Remarks

The quantitative data obtained with fluorescein- substituted lectins are in good agreement with the data obtained with radioactively labeled lectins. However, the use of fluoresceinyl-lectin lead to several advan- tages. With fluoresceinyl-lectins, one avoids the use of radioactively labeled products and radioactive con- tamination problems. Fluoresceinyl-lectins are com- mercially available and may be kept for months at 4 C; they are more stable than radioactively labeled lectins. Fluoresceinyl-lectins, which are visualized under fluo-

300 M. Monsigny, C. Sene, and A. Obrenovitch : Quantitative Fluoresceinyl-Lectin Binding to Cell-Surface Glycoconjugates

rescence microscope, are suitable to detect the pres- ence of labeled and unlabeled cell populations and differences of binding levels within the labeled popu- lation, to monitor the distribution of sites on cells, and also, to detect unspecific bindings. Owing to the fact that the fluorescent labeling can detect the labeled and unlabeled populations, quantitative measurements may be related to the labeled population.

The set up of the technique is readily achieved, specially to define the best conditions of lectin release with a sugar inhibitor, and to detect artefacts such as high level of non-specific binding or internalisation of the lectin material.

This work was partly supported by grants C.L. 75.4.074.3 from lnstitut National de la Sante el de la Recherche Mbdicale, ACC 77.7.0252 from Delegation Generule a la Recherche Scientifique et Technique, and ATP 76.2302 from Centre National de la Recherche Scientlfique. A.O. is Cliargee de Recherche lnstitut National de la Sunre et de la Recherche Medicale. We wish to thank Mrs M. Mit- terand for her valuable technical assistance.

REFERENCES I . Nicolson, G . L. (1978) in Advanced Techniques in Biological

Electron Microscopy (Koeler, J. K., ed.) vol. 2, pp. 1-38, Springer Verlag, Berlin, Heidelberg, New York.

2. Sharon, N. & Lis, H. (1975) in Methods in Membrane Biology (Korn, E. D., ed.) vol. 3, pp. 147-200, Plenum Press, New York.

3. Monsigny, M., Roche, A. C. & Kieda, C. (1978) in Structure and Functions of Biological Membranes. Molecular Aspects (Centre de Biophysique Moleculaire, ed.) pp. 161 -234, Commission of the European Communities, Brussels.

4. Cline, M. J. & Livingston, D. C. (1971) Nut. New Biol. 232, 155-156.

5. Noonan, K. & Burger, M. M. (1973) J . Biol. Chum. 248,

6. Hunter, W. M. & Greenwood, F. C. (1962) Nature (Lond.] 134,

7. Unanue, E. R., Perkins, W. D. & Karnovsky, M. J. (1972) J .

8. Weber, T. M., Lindahl-Kiessling, K., Mattsson, A. & Alm,G.

9. Yahara, I. & Edelman, G. M. (1973) Erp. Cell. Res. 81, 143-

10. Comoglio, P. M. & Guglielmone, R. (1972) FEBS Lett. 27.

11. Inbar, M. & Sachs, L. (1973) FEBS Lett. 32, 124- 128. 12. Inbar, M., Ben Bassat, H . & Sachs, L. (1973) Int. J . Cancer, 12,

13. Nicolson, G. L. (1973) Nut. New Biol. 243, 218-220. 14. Meager, A., Ungkitchanukit, A. & Hughes, R. C. (1976) Bio-

15. Riordan, J . F. & Vallee, B. L. (1972) Methods Enzymology, 25,

16. Loor, F. (1973) E-xp. Cell. Res. 82, 415-425. 17. Gunther, G. R., Wang, J. L., Yahara, I., Cunningham, B. A.

& Edelman, G. L. (1973) Proc. Nut1 Acad. Sci. U.S.A. 70, 1012-1016.

18. Monsigny, M., Sene, C., Obrenovitch, A., Roche, A. C., Del- motte, F. & Boschetti, C. Eur. J . Biochem. in the press.

19. Agrawal, B. B. L. & Goldstein, I . J . (1965) Biochcvn. J . Y6, 230. 20. Bouchard, P., Moroux, Y., Tixier, R., Privat, J. P. & Monsigny,

21. Scatchard, G. (1949) Ann. N . Y. Acad. Sci. 51, 660-672. 22. Phillips, P. G., Furmanski. P. & Lubin, M. (1974) ESP. Crll.

23. Vicker, M. G. & Edwards, J. C. (1972) J . Cell. Sci. 10, 759-

24. Bornens, M., Karsenti, E. & Avrameas, S. (1976) Eur. J . Bio-

25. Feller, M., Richardson, C., Behnke, D. & Gruenstein, E. (1977)

4286 -4292.

495 - 496.

E,xP. Med. 136, 88 5 - 906.

V. (1972) Life Sci. 11 , 343-3350,

155.

256 - 258.

93 - 99.

chem. J . 154. 1 I3 - 124.

494 ~ 500.

M. (1976) Biochimie, 58, 1247-1253.

Res. 86, 301 - 308.

768.

chem. 65,61-69.

Biochem. Biophys. Res. Commun. 76, 1027- 1035.

M. Monsigny, C. Sene, and A. Obrenovitch, Centre de Biophysique Moleculaire du Centre National de la Recherche Scientifique, 1A Avenue de la Recherche Scientifique, F-45045 Orleans-Cedex, France