- Email: [email protected]
A lectin-immunofluorometric assay using an immobilized Bandeiraea simplicifolia II lectin for the determination of galactosylation variants of glycoproteins
177,383-387
ANALYTICALBIOCHEMISTRY
(1989)
A Lectin-lmmunofluorometric Assay Using an Immobilized Bandeiraea simplicifolia II Lectin for the Determination of Galactosylation Variants of Glycoproteins Jaakko
Parkkinen
and Ulla
Oksanen
Department of Obstetrics and Gynecology, Helsinki University of Helsinki, SF-00290 Helsinki, Finland
Received
September
University
Central Hospital
Most proteins secreted by animal cells contain covalently linked oligosaccharide chains. An individual glycoprotein often carries variable oligosaccharide structures at different sites of the polypeptide but even the same glycosylation site may contain several types of oligosaccharide chains (1,2). This heterogeneity of glycosylation creates molecular variants of an individual glycoprotein (glycosylation variants). In a glycoprotein synthetized in various tissues, the heterogeneity of glycosylation is apparently due to tissue-specific differ0003-2697/89 Copyright
of Medical
Chemistry,
19,1988
The sandwich-type immunometric assay was modified by replacing the solid phase-bound antibody with a lectin for the determination of glycoproteins carrying terminal N-acetylglucosamine residues. Microwells were coated with Bandeiraea simplicifolia II lectin and incubated with glycosylation variants of human serum glycoproteins. The bound glycoproteins were detected by time-resolved fluorometry using europium-labeled antibodies. Agalacto-derivatives of al-acid glycoprotein and transferrin obtained by neuraminidase and j3galactosidase treatment bound to the immobilized lectin, whereas the native or desialylated glycoproteins showed no binding. The measuring range of the method for agalacto-q-acid glycoprotein was 0.01 to 10 fig/ml and for agalacto-transferrin 1 to 300 pg/ml. The binding of the agalacto-glycoproteins was totally inhibited with 1 to 10 mM N-acetylglucosamine which confirmed the specificity of the method for glycoproteins containing terminal N-acetylglucosamine residues. The results indicate that the novel lectin-immunofluorometric method is sensitive and has a wide measuring range for the determination of glycosylation variants of glycoproteins. 8 1989 Academic Press, Inc.
$3.00 0 1989 by Academic
All rights of reproduction
and Department
Press,
Inc.
in any form reserved.
encies in the glycosylation process (3-6). Other factors modifying the oligosaccharide structures of newly synthetized glycoproteins include the differentiation stage of the tissue (7-9) and external regulators such as monokines (10). Various pathological conditions have also been shown to influence the glycosylation process. Glycoproteins excreted by malignant tissues display aberrant sugar structures as compared to those secreted by the corresponding normal tissues (11-13). In rheumatoid arthritis, a specific galactosylation deficiency of IgG’ has been described (14,15). Furthermore, inflammation and excessive alcohol intake result in distinct structural changes in the oligosaccharide chains of (Yeacid glycoprotein and other glycoproteins synthetized in the liver (16-18). Thus, by analyzing glycosylation variants of glycoproteins, a great deal of information on the tissue origin and possible pathological processes affecting it can be obtained. The analysis of glycosylation variants of a glycoprotein requires methods for their separation. Lectins have been used for this purpose, either in affinity chromatography or affino-electrophoresis (19). The glycosylation variants reactive and nonreactive with the lectin can be quantitated immunochemically which makes the analysis of biological samples without prior purification of the glycoprotein possible (16-18). However, these methods are rather elaborate and time-consuming to be adopted, e.g., for routine diagnostics. Sandwich-type immunometric assays have become very popular recently because of their rapidity, sensitivity, and specificity. These assays employ a solid ’ Abbreviations used: LIFMA, lectin-immunofluorometric assay; BS-II lectin, Bandeiruea simplicifoliu II lectin; AGP, cu,-acid glycoprotein; TF, transferrin; TBS, Tris-buffered saline; BSA, bovine serum albumin; IgG, immunoglobulin G, LEIA, lectin-enzyme immunoassay. 383
384
PARKKINEN
phase-bound antibody that captures the analyte and a detecting antibody carrying an enzyme, radioactive, or fluorescent label (20,21). In the present study we have modified the sandwich-type immunometric assay for the detection of glycosylation variants of glycoproteins by replacing the solid phase-bound antibody with a lectin. This lectin-immunofluorometric assay (LIFMA) employs solid phase-bound Bandeiraea simplicifolia II (BSII) lectin which possesses binding specificity for terminal N-acetylglucosamine residues (22). The bound glycoproteins are detected by time-resolved fluorometry using a europium-labeled antibody which has proved to be a highly sensitive and dynamic nonradioactive method (23,24). Terminal N-acetylglucosamine was chosen as the sugar structure to be studied for the following reasons. Heterogeneity of oligosaccharide chains of glycoproteins often involves loss of terminal sugar residues, such as sialic acid and galactose. Whereas glycoproteins expressing terminal galactose residues are readily removed from circulation, those containing terminal N-acetylglucosamine have only a slightly decreased half-life in circulation (25). The agalacto-IgG observed in serum of patients with rheumatoid arthritis carries terminal N-acetylglucosamine residues in its saccharide chains (14) and has been suggested to be a valuable marker for the diagnosis and follow-up of the disease activity (15). In the present study we have applied LIFMA for studies on two human serum glycoproteins, al-acid glycoprotein (AGP) and transferrin (TF). Both of these glycoproteins contain only N-linked oligosaccharide chains typical of plasma glycoproteins. TF carries two biantennary N-linked chains (26), whereas AGP contains on an average five tri- to tetra-antennary chains (27). In both cases, the terminal branches contain the sequence sialylgalactosyl-N-acetylglucosamine. We show here that LIFMA specifically and sensitively measures AGP and TF derivatives containing terminal N-acetylglucosamine residues but does not detect the native glycoproteins or their desialylated derivatives. MATERIALS
AND
METHODS
Materials. IgG fractions of rabbit antisera against human transferrin and a,-acid glycoprotein were obtained from Dakopatts (Copenhagen, Denmark). B. simplicifolia II lectin, human transferrin, N-acetylglucosamine, and N-acetylgalactosamine were from Sigma (St. Louis, MO). Vibrio cholerea neuraminidase was obtained from Calbiochem Behring Diagnostics (La Jolla, CA). Purified Aspergillus niger P-galactosidase was kindly provided by Dr. Jonathan Knowles (Technical Research Center of Finland, Helsinki, Finland). Human al-acid glycoprotein was kindly supplied by Dr. Gunnar Myllyla (Finnish Red Cross Blood Transfusion Service, Helsinki, Finland). Isothiocyanatophenyldiethylene-
AND
OKSANEN
triaminepentaacetic acid complexed with europium (europium chelate) and “enhancement solution” containing 0.1% (w/v) Triton X-100,6.8 mM potassium hydrogen phthalate, 100 mM acetic acid, 50 pM tri-n-octylphosphine oxide, and 15 pM 2-naphtoyltrifluoroacetone Wallac Biochemical Laboratories (23) were from (Turku, Finland). Labeling of antibodies. Labeling of the antibodies with Eu-chelate was carried out essentially as described before (23). IgG was precipitated with 18% Na2S04, dissolved in 0.1 M NaHCO,, pH 9.5, and reacted with a 200fold molar excess of the Eu-chelate overnight at 4°C. The labeled antibodies were separated from unreacted chelate by gel filtration on a column (1 X 20 cm) of Sephacryl S-200 (Pharmacia Fine Chemicals, Uppsala, Sweden). The elution buffer was 20 mM Tris-HCl buffer, pH 7.4, containing 0.15 M NaCl and 5 mM NaN3 (TBS). On an average, six molecules of europium were incorporated into one molecule of IgG. Preparation of glycosylation variants of glycoproteins. The glycoproteins (2 mg/ml) were dissolved in 50 mM Na acetate, pH 4.5, containing 1 mM CaCl, and 5 mM NaN3, filtrated through a 0.22-pm sterile filter, and incubated in the presence of V. cholerae neuraminidase alone, 0.05 U/ml, or in combination with A. niger &galactosidase (0.1 U/ml) for 24 h at 37°C. After incubation the glycoproteins were dialyzed against TBS. Lectin-immunofEuorometric assay. Polystyrene microtiter strip wells (Titertek, Labsystems, Helsinki, Finland) were coated with BS-II lectin, 10 /*g/ml, in 0.1 M NaHCO,, pH 9.0, for 16 h at 4°C. The wells were emptied and incubated with TBS containing 5 mg/ml bovine serum albumin (BSA), 0.1 mM CaClz, 0.1 mM MnClz, and 5 mM NaN, for 3 h at room temperature. After the wells were washed twice with washing solution (containing, per liter, 9 g of NaCl, 0.2 g of Tween 20, and 0.5 g of NaNa), the glycoproteins were incubated in the wells in 200 ~1 of Buffer A (TBS containing 0.1 mM CaCl,, 0.1 mM MnC&, 1 mg/ml BSA, 0.05% (v/v) Tween 20, and 5 mM NaN, ) for 2 h at room temperature with continuous shaking. The wells were washed thrice with washing solution and incubated with the Eu-labeled antibodies, about 100 ng IgG/well, in 200 ~1 of Buffer B (50 mM Tris-HCI buffer, pH 7.7, containing, per liter, 9 g of NaCl, 0.5 g of NaN,, 5 g of bovine serum albumin, 0.5 g of bovine immunoglobulin, and 0.1 g of Tween 40) (23) for 1 h at room temperature with continuous shaking. The wells were washed four times with washing solution and 200 ~1 of enhancement solution was added. After incubation for 15 min, fluorescence was measured with an LKB 1230 Arcus fluorometer (Wallac Biochemical Laboratories). Immunofkorometric assay. Microwells were coated with IgG from the rabbit antisera against AGP or TF, 10 pg/ml, in 200 ~1 of 0.1 M NaHC03, pH 9.0, for 16 h at
LECTIN-IMMUNOFLUOROMETRIC
102 10-4
10-3
10'2
10-l
loo
a, -acid glycoprotein
10'
102
385
ASSAY
103
2 10'4
10-3
(pg/ml)
FIG. 1. Binding of or,-acid glycoprotein and its agalacto-derivative to microwells coated with AGP-specific antibody or BS-II lectin. AGP (0) and agalacto-AGP (0) were incubated at the different concentrations shown in microwells coated with AGP-specific antibody (- - -) or BS-II lectin (-). The bound glycoproteins were detected by Eu-labeled AGP-specific antibody.
4°C. The wells were emptied, incubated with TBS containing 5 mg/ml BSA for 3 h at room temperature, and washed twice with washing solution. The glycoproteins were incubated in the wells in 200 ~1 of Buffer B for 1 h at room temperature with continuous shaking. The wells were washed thrice with washing solution and incubated with the Eu-labeled antibodies as described for LIFMA. RESULTS
AGP and TF were treated with neuraminidase alone or in combination with P-galactosidase to obtain glycosylation variants lacking terminal sialic acid (asialo-derivative) or both sialic acid and galactose (agalacto-derivative), respectively. In the immunofluorometric assays employing solid-phase bound anti-AGP or anti-TF antibody, both the native glycoproteins and their agalacto-derivatives gave similar binding curves (Figs. 1 and 2). This indicated that desialylation and degalactosylation of the glycoproteins did not affect the binding of the Eu-labeled antibodies. The detection limit for both glycoproteins in the immunofluorometric assays was about 1 rig/ml. The Eu-labeled antibodies showed only moderate background binding to microwells coated with the BS-II lectin, about 3 to 4 X lo3 cps, which was similar to the binding of microwells coated with BSA only. This indicated that the labeled antibodies did not carry sugar structures recognized by the immobilized BS-II lectin. Incubation of agalacto-AGP and agalacto-TF in the lectin-coated microwells resulted in dose-dependent binding as detected by the Eu-labeled antibodies (Figs. 1 and 2). Binding of agalacto-AGP could be detected at a concentration of 0.01 ,ug/ml and the binding curve was nearly linear up to l-10 pg/ml (Figs. 1 and 3). Native
10-2
10-l
Transferrin
100
10'
10'
lo3
(pg/ml)
FIG. 2. Binding of transferrin and its agalacto-derivative to microwells coated with TF-specific antibody or BS-II lectin. TF (0) and agalacto-TF (0) were incubated at different concentrations in microwells coated with TF-specific antibody (---) or BS-II lectin (-). The bound glycoproteins were detected by Eu-labeled TF-specific antibody.
AGP and asialo-AGP showed no binding at concentrations below 100 fig/ml (Fig. 3). Native AGP showed very slight binding at a concentration of 100 @g/ml, i.e., 0.2% of the binding of agalacto-AGP (Fig. 1). The detection limit of LIFMA for agalacto-TF was about 1 pg/ml and the binding curve was linear to the highest standard tested, i.e., 300 pg/ml (Fig. 2). Native TF showed slight binding at a concentration of 300 pg/ml which was 0.4% of the binding of agalacto-TF (Fig. 2). At lower concentrations native and asialo-TF showed no binding. The binding of agalacto-AGP and agalacto-TF could be totally inhibited by 1 to 10 mM N-acetylglucosamine (Fig. 4), a specific hapten sugar of BS-II lectin (22). Nacetylgalactosamine did not inhibit the binding even at 100 mM concentration (Fig. 4). This confirmed that the
102
m 1 o-3
10-2
10-l
a, -acid glycoprotein
100
IO'
10"
@g/ml)
FIG. 3. Binding of glycosylation variants of cu,-acid glycoprotein to microwells coated with BS-II lectin. Native AGP (O), asialo-AGP (A), and agalacto-AGP (0) were incubated at different concentrations in the microwells and the bound glycoproteins were detected by Eu-labeled AGP-specific antibody.
386
PARKKINEN
Inhibitor concentration
(mM)
Inhibitor concentration
(mM)
AND
FIG. 4.
Inhibition of binding of the agalacto-derivatives of a,-acid glycoproteins (A) and transferrin (B) to microwells coated with BSII lectin. Agalacto-AGP (10 pg/ml) and agalacto-TF (50 rg/ml) were incubated in the microwells with different concentrations of N-acetylglucosamine (0) or N-acetylgalactosamine (0). The bound glycoproteins were detected by Eu-labeled antibodies.
binding of the agalacto-derivatives to the lectin-coated microwells was due to a lectin-carbohydrate interaction specific for terminal N-acetylglucosamine residues. DISCUSSION
The present paper describes a novel lectin-immunofluorometric assay for the determination of glycoproteins containing terminal N-acetylglucosamine residues. The method sensitively detects the desialylated and degalactosylated derivatives of AGP and TF but does not measure the native or desialylated glycoproteins. The specificity of the method for glycoproteins containing terminal N-acetylglucosamine residues was confirmed by the ability of free N-acetylglucosamine to totally inhibit binding of the agalacto-glycoproteins to the lectin-coated wells. The measuring range of LIFMA is large, nearly lOOOfold, which is typical for immunofluorometric assaysemploying time-resolved fluorescence as the detection system (23,24). LIFMA was less sensitive in the detection
OKSANEN
of the agalacto-glycoproteins than the immunofluorometric assay using solid-phase bound antibody. For agalacto-AGP, the difference was about lo-fold and for agalacto-TF about IOOO-fold. This is evidently due to the lower affinity of the lectin used on the solid phase as compared to the avidity of the polyclonal antibodies. However, the degree of desialylation and degalactosylation of the agalacto-derivatives studied in the present study was not determined and it is possible that the sensitivity of LIFMA for glycoproteins totally lacking galactose is higher than that observed. Interestingly, agalacto-AGP could be detected in LIFMA with about loo-fold higher sensitivity than agalacto-TF. This is probably due to the higher degree of polyvalency of the terminal sugar residues present in AGP as compared to TF (26,27). Polyvalency of the sugar residues has earlier been shown to remarkably affect the affinity of lectins toward glycoproteins (28). Thus, LIFMA obviously detects with higher sensitivity heavily glycosylated glycoproteins such as AGP than it detects those containing less carbohydrates. Recently, a lectin-enzyme immunoassay for the detection of sialovariants of TF has been described (29). This method, called LEIA, employed a solid phase-bound anti-TF antibody and a biotin-conjugated Ricinus communis lectin complexed with avidin-conjugated alkaline phosphatase for detection, In addition to the detection method and the lectin used, the major difference between LEIA and LIFMA is the side of the sandwich on which the lectin is used. A theoretical problem in both systems is the capacity of the solid phase. When antibody is immobilized on the solid phase, other glycosylation variants of the same glycoprotein compete for binding to the immobilized antibody and may limit the binding of the glycosylation variant studied. On the other hand, when lectin is used on the solid phase, other proteins containing similar glycan chains may cause interference. In this respect, an obvious advantage of LIFMA is the high sensitivity of the detection system and wide measuring range which should allow proper dilution of the sample so that interfering glycoproteins would not limit the capacity of the solid phase. In conclusion, the lectin-immunofluorometric assay described in the present paper is a promising method for the rapid and sensitive quantification of glycosylation variants of glycoproteins. Application of the method for clinically important glycoproteins is currently in progress. ACKNOWLEDGMENTS The authors thank Dr. Ulf-Hikan Stenman, cussions and critical reading of the manuscript. ported by the Sigrid Jusilius Foundation.
M.D., for inspiringdisThis work was sup-
REFERENCES 1. Kobata, Robbins, York.
A. (1964) in Biology P. W., Eds.), Vol.
of Carbohydrates (Ginsburg, V., and 2, Chap. 2, pp. 87-162, Wiley, New
LECTIN-IMMUNOFLUOROMETRIC 2. Swiedler, and Hart,
S. J., Freed, J. H., Tarentino, G. W. (1985) J. Biol. Chem.
A. L., Plummer, 260.4046-4054.
T. H., Jr.,
3. Krusius, T., and Ruoslahti, E. (1982) J. Biol. Chem. 257, 34533457. 4. Williams, D. B., and Lennarz, W. J. (1984) J. Biol. Chem. 259, 5105-5114. 5. Sheares, B. T., and Rohbins, P. W. (1986) Proc. Natl. Acad. Sci. USA 83,1993-1997. 6. Parekh, R. B., Tse, A. G. D., Dwek, R. A., Williams, A. F., and Rademacher, T. W. (1987) Embo J. 6,1233-1244. 7. Feizi, T. (1985) Nature (London) 314,53-57. 8. Finne, J. (1985) Trends Biochem. Sci. 10,129-132. 9 Gahmberg, G. C., Ekblom, M., and Andersson, L. C. (1984) Proc. Natl. Acad. Sci. USA 8 1,6752-6756. 10 Mackiewicz, A., Ganapathi, M. K., Schultz, and Kushner, I. (1987) J. Exp.Med. 166,253-258. 11 Yamamoto, K., Tsuji, T., Tarutani, O., and Osawa, T. (1984) Eur. J. Biochem. 143,133-144. 12. Yamashita, K., Totani, K., Kuroki, M., Matsuoka, Y., Ueda, I., and Kobata, A. (1987) Cancer Res. 47,3451-3459. 13. Endo, T., Nishimura, R., Kawano, T., Mochizuki, M., and Kobata, A. (1987) Cancer Res. 47,5242-5245. 14. Parekh, R. B., Dwek, R. A., Sutton, B. J., Fernandes, D. L., Leung, A., Stanworth, D., Rademacher, T. W., Mizuochi, T., Taniguchi, T., Matsuta, K., Takeuchi, F., Nagano, Y., Miyamoto, T., and Kobata, A. (1985) Nature (London) 316,452-457. 15. Parekh, R. B., Roitt, I. M., Isenberg, D. A., Dwek, R. A., Ansell, B. M., and Rademacher, T. W. (1988) Larzcet i, 966-969. 16. Mackiewicz, A., Marcinkowska-Pieta, R., Ballou, S., Mackiewicz, and Kushner, I. (1987) Arthritis Rheumat. 30,513-518. *A.
387
ASSAY 17. Hachulla, 911-915.
E., Laine,
A., and Hayem,
A. (1988)
18. Jezequel, M., Seta, N. S., Corbic, M. M., Feger, G. M. (1988) Clin. Chim. Acta 176.49-57. 19. Beg-Hansen,
T. C. (1973)
20. Wide, L., Bennich, 1105-1107. 21. Miles, L. E. M., 186-189.
Anal.
Biochem.
C. N. (1968)
Nature
Stenman, U.-H., J., and Sepplli, 736.
25. Ashwell, eas Mol.
Alftan, H., Ranta, M. (1987) J. Clin.
G., and Morrell, Biol. 41,99-128.
219,
in The Lectins, ProperBiology and Medicine I. J., Eds.), pp. 91-93,
Adv.
26. Spik, G., Bayard, B., Fournet, B., Strecker, Montreuil, J. (1975) FEBS Lett. 50,296-299.
28. Stowell,
ii,
H., and
Liiv-
T., Vartiainen, E., Jalkanen, Endocrinol. Metab. 64, 730-
A. E. (1974)
B., Montreuil, J., Strecker, 27. Fournet, J., Vliegenthart, J. F. G., Binette, Biochemistry 17,5206-5214.
Lancet
(London)
23. Hemmila, I., Dakubu, S., Mukkala, V.-M., Siitari, gren, T. (1984) Anal. Biochem. 137,335-343. 34
43,
J. M., and Durand,
S. G. 0. (1967)
22. Goldstein, I. J., and Pore& R. D. (1986) ties, Functions, and Applications in (Liener, I. E., Sharon, N., and Goldstein, Academic Press, Orlando, FL.
I-’
Chem.
66,480-488.
H., and Johansson, and Hales,
Clin.
Enzymol.
Related
G., Bouchelet,
Ar-
S., and
G., Dorland, L., Haverkamp, J. P., and Schmid, K. (1978)
C. P., Lee, R. T., and Lee, Y. C. (1980)
Biochemistry
19,
4904-4908. 29. Pekelharing, (1987) Anal.
J. M., Vissers, P., Peters, Biochem. 165,320-326.
H. A., and
Leijnse,
B.
ANALYTICALBIOCHEMISTRY
(1989)
A Lectin-lmmunofluorometric Assay Using an Immobilized Bandeiraea simplicifolia II Lectin for the Determination of Galactosylation Variants of Glycoproteins Jaakko
Parkkinen
and Ulla
Oksanen
Department of Obstetrics and Gynecology, Helsinki University of Helsinki, SF-00290 Helsinki, Finland
Received
September
University
Central Hospital
Most proteins secreted by animal cells contain covalently linked oligosaccharide chains. An individual glycoprotein often carries variable oligosaccharide structures at different sites of the polypeptide but even the same glycosylation site may contain several types of oligosaccharide chains (1,2). This heterogeneity of glycosylation creates molecular variants of an individual glycoprotein (glycosylation variants). In a glycoprotein synthetized in various tissues, the heterogeneity of glycosylation is apparently due to tissue-specific differ0003-2697/89 Copyright
of Medical
Chemistry,
19,1988
The sandwich-type immunometric assay was modified by replacing the solid phase-bound antibody with a lectin for the determination of glycoproteins carrying terminal N-acetylglucosamine residues. Microwells were coated with Bandeiraea simplicifolia II lectin and incubated with glycosylation variants of human serum glycoproteins. The bound glycoproteins were detected by time-resolved fluorometry using europium-labeled antibodies. Agalacto-derivatives of al-acid glycoprotein and transferrin obtained by neuraminidase and j3galactosidase treatment bound to the immobilized lectin, whereas the native or desialylated glycoproteins showed no binding. The measuring range of the method for agalacto-q-acid glycoprotein was 0.01 to 10 fig/ml and for agalacto-transferrin 1 to 300 pg/ml. The binding of the agalacto-glycoproteins was totally inhibited with 1 to 10 mM N-acetylglucosamine which confirmed the specificity of the method for glycoproteins containing terminal N-acetylglucosamine residues. The results indicate that the novel lectin-immunofluorometric method is sensitive and has a wide measuring range for the determination of glycosylation variants of glycoproteins. 8 1989 Academic Press, Inc.
$3.00 0 1989 by Academic
All rights of reproduction
and Department
Press,
Inc.
in any form reserved.
encies in the glycosylation process (3-6). Other factors modifying the oligosaccharide structures of newly synthetized glycoproteins include the differentiation stage of the tissue (7-9) and external regulators such as monokines (10). Various pathological conditions have also been shown to influence the glycosylation process. Glycoproteins excreted by malignant tissues display aberrant sugar structures as compared to those secreted by the corresponding normal tissues (11-13). In rheumatoid arthritis, a specific galactosylation deficiency of IgG’ has been described (14,15). Furthermore, inflammation and excessive alcohol intake result in distinct structural changes in the oligosaccharide chains of (Yeacid glycoprotein and other glycoproteins synthetized in the liver (16-18). Thus, by analyzing glycosylation variants of glycoproteins, a great deal of information on the tissue origin and possible pathological processes affecting it can be obtained. The analysis of glycosylation variants of a glycoprotein requires methods for their separation. Lectins have been used for this purpose, either in affinity chromatography or affino-electrophoresis (19). The glycosylation variants reactive and nonreactive with the lectin can be quantitated immunochemically which makes the analysis of biological samples without prior purification of the glycoprotein possible (16-18). However, these methods are rather elaborate and time-consuming to be adopted, e.g., for routine diagnostics. Sandwich-type immunometric assays have become very popular recently because of their rapidity, sensitivity, and specificity. These assays employ a solid ’ Abbreviations used: LIFMA, lectin-immunofluorometric assay; BS-II lectin, Bandeiruea simplicifoliu II lectin; AGP, cu,-acid glycoprotein; TF, transferrin; TBS, Tris-buffered saline; BSA, bovine serum albumin; IgG, immunoglobulin G, LEIA, lectin-enzyme immunoassay. 383
384
PARKKINEN
phase-bound antibody that captures the analyte and a detecting antibody carrying an enzyme, radioactive, or fluorescent label (20,21). In the present study we have modified the sandwich-type immunometric assay for the detection of glycosylation variants of glycoproteins by replacing the solid phase-bound antibody with a lectin. This lectin-immunofluorometric assay (LIFMA) employs solid phase-bound Bandeiraea simplicifolia II (BSII) lectin which possesses binding specificity for terminal N-acetylglucosamine residues (22). The bound glycoproteins are detected by time-resolved fluorometry using a europium-labeled antibody which has proved to be a highly sensitive and dynamic nonradioactive method (23,24). Terminal N-acetylglucosamine was chosen as the sugar structure to be studied for the following reasons. Heterogeneity of oligosaccharide chains of glycoproteins often involves loss of terminal sugar residues, such as sialic acid and galactose. Whereas glycoproteins expressing terminal galactose residues are readily removed from circulation, those containing terminal N-acetylglucosamine have only a slightly decreased half-life in circulation (25). The agalacto-IgG observed in serum of patients with rheumatoid arthritis carries terminal N-acetylglucosamine residues in its saccharide chains (14) and has been suggested to be a valuable marker for the diagnosis and follow-up of the disease activity (15). In the present study we have applied LIFMA for studies on two human serum glycoproteins, al-acid glycoprotein (AGP) and transferrin (TF). Both of these glycoproteins contain only N-linked oligosaccharide chains typical of plasma glycoproteins. TF carries two biantennary N-linked chains (26), whereas AGP contains on an average five tri- to tetra-antennary chains (27). In both cases, the terminal branches contain the sequence sialylgalactosyl-N-acetylglucosamine. We show here that LIFMA specifically and sensitively measures AGP and TF derivatives containing terminal N-acetylglucosamine residues but does not detect the native glycoproteins or their desialylated derivatives. MATERIALS
AND
METHODS
Materials. IgG fractions of rabbit antisera against human transferrin and a,-acid glycoprotein were obtained from Dakopatts (Copenhagen, Denmark). B. simplicifolia II lectin, human transferrin, N-acetylglucosamine, and N-acetylgalactosamine were from Sigma (St. Louis, MO). Vibrio cholerea neuraminidase was obtained from Calbiochem Behring Diagnostics (La Jolla, CA). Purified Aspergillus niger P-galactosidase was kindly provided by Dr. Jonathan Knowles (Technical Research Center of Finland, Helsinki, Finland). Human al-acid glycoprotein was kindly supplied by Dr. Gunnar Myllyla (Finnish Red Cross Blood Transfusion Service, Helsinki, Finland). Isothiocyanatophenyldiethylene-
AND
OKSANEN
triaminepentaacetic acid complexed with europium (europium chelate) and “enhancement solution” containing 0.1% (w/v) Triton X-100,6.8 mM potassium hydrogen phthalate, 100 mM acetic acid, 50 pM tri-n-octylphosphine oxide, and 15 pM 2-naphtoyltrifluoroacetone Wallac Biochemical Laboratories (23) were from (Turku, Finland). Labeling of antibodies. Labeling of the antibodies with Eu-chelate was carried out essentially as described before (23). IgG was precipitated with 18% Na2S04, dissolved in 0.1 M NaHCO,, pH 9.5, and reacted with a 200fold molar excess of the Eu-chelate overnight at 4°C. The labeled antibodies were separated from unreacted chelate by gel filtration on a column (1 X 20 cm) of Sephacryl S-200 (Pharmacia Fine Chemicals, Uppsala, Sweden). The elution buffer was 20 mM Tris-HCl buffer, pH 7.4, containing 0.15 M NaCl and 5 mM NaN3 (TBS). On an average, six molecules of europium were incorporated into one molecule of IgG. Preparation of glycosylation variants of glycoproteins. The glycoproteins (2 mg/ml) were dissolved in 50 mM Na acetate, pH 4.5, containing 1 mM CaCl, and 5 mM NaN3, filtrated through a 0.22-pm sterile filter, and incubated in the presence of V. cholerae neuraminidase alone, 0.05 U/ml, or in combination with A. niger &galactosidase (0.1 U/ml) for 24 h at 37°C. After incubation the glycoproteins were dialyzed against TBS. Lectin-immunofEuorometric assay. Polystyrene microtiter strip wells (Titertek, Labsystems, Helsinki, Finland) were coated with BS-II lectin, 10 /*g/ml, in 0.1 M NaHCO,, pH 9.0, for 16 h at 4°C. The wells were emptied and incubated with TBS containing 5 mg/ml bovine serum albumin (BSA), 0.1 mM CaClz, 0.1 mM MnClz, and 5 mM NaN, for 3 h at room temperature. After the wells were washed twice with washing solution (containing, per liter, 9 g of NaCl, 0.2 g of Tween 20, and 0.5 g of NaNa), the glycoproteins were incubated in the wells in 200 ~1 of Buffer A (TBS containing 0.1 mM CaCl,, 0.1 mM MnC&, 1 mg/ml BSA, 0.05% (v/v) Tween 20, and 5 mM NaN, ) for 2 h at room temperature with continuous shaking. The wells were washed thrice with washing solution and incubated with the Eu-labeled antibodies, about 100 ng IgG/well, in 200 ~1 of Buffer B (50 mM Tris-HCI buffer, pH 7.7, containing, per liter, 9 g of NaCl, 0.5 g of NaN,, 5 g of bovine serum albumin, 0.5 g of bovine immunoglobulin, and 0.1 g of Tween 40) (23) for 1 h at room temperature with continuous shaking. The wells were washed four times with washing solution and 200 ~1 of enhancement solution was added. After incubation for 15 min, fluorescence was measured with an LKB 1230 Arcus fluorometer (Wallac Biochemical Laboratories). Immunofkorometric assay. Microwells were coated with IgG from the rabbit antisera against AGP or TF, 10 pg/ml, in 200 ~1 of 0.1 M NaHC03, pH 9.0, for 16 h at
LECTIN-IMMUNOFLUOROMETRIC
102 10-4
10-3
10'2
10-l
loo
a, -acid glycoprotein
10'
102
385
ASSAY
103
2 10'4
10-3
(pg/ml)
FIG. 1. Binding of or,-acid glycoprotein and its agalacto-derivative to microwells coated with AGP-specific antibody or BS-II lectin. AGP (0) and agalacto-AGP (0) were incubated at the different concentrations shown in microwells coated with AGP-specific antibody (- - -) or BS-II lectin (-). The bound glycoproteins were detected by Eu-labeled AGP-specific antibody.
4°C. The wells were emptied, incubated with TBS containing 5 mg/ml BSA for 3 h at room temperature, and washed twice with washing solution. The glycoproteins were incubated in the wells in 200 ~1 of Buffer B for 1 h at room temperature with continuous shaking. The wells were washed thrice with washing solution and incubated with the Eu-labeled antibodies as described for LIFMA. RESULTS
AGP and TF were treated with neuraminidase alone or in combination with P-galactosidase to obtain glycosylation variants lacking terminal sialic acid (asialo-derivative) or both sialic acid and galactose (agalacto-derivative), respectively. In the immunofluorometric assays employing solid-phase bound anti-AGP or anti-TF antibody, both the native glycoproteins and their agalacto-derivatives gave similar binding curves (Figs. 1 and 2). This indicated that desialylation and degalactosylation of the glycoproteins did not affect the binding of the Eu-labeled antibodies. The detection limit for both glycoproteins in the immunofluorometric assays was about 1 rig/ml. The Eu-labeled antibodies showed only moderate background binding to microwells coated with the BS-II lectin, about 3 to 4 X lo3 cps, which was similar to the binding of microwells coated with BSA only. This indicated that the labeled antibodies did not carry sugar structures recognized by the immobilized BS-II lectin. Incubation of agalacto-AGP and agalacto-TF in the lectin-coated microwells resulted in dose-dependent binding as detected by the Eu-labeled antibodies (Figs. 1 and 2). Binding of agalacto-AGP could be detected at a concentration of 0.01 ,ug/ml and the binding curve was nearly linear up to l-10 pg/ml (Figs. 1 and 3). Native
10-2
10-l
Transferrin
100
10'
10'
lo3
(pg/ml)
FIG. 2. Binding of transferrin and its agalacto-derivative to microwells coated with TF-specific antibody or BS-II lectin. TF (0) and agalacto-TF (0) were incubated at different concentrations in microwells coated with TF-specific antibody (---) or BS-II lectin (-). The bound glycoproteins were detected by Eu-labeled TF-specific antibody.
AGP and asialo-AGP showed no binding at concentrations below 100 fig/ml (Fig. 3). Native AGP showed very slight binding at a concentration of 100 @g/ml, i.e., 0.2% of the binding of agalacto-AGP (Fig. 1). The detection limit of LIFMA for agalacto-TF was about 1 pg/ml and the binding curve was linear to the highest standard tested, i.e., 300 pg/ml (Fig. 2). Native TF showed slight binding at a concentration of 300 pg/ml which was 0.4% of the binding of agalacto-TF (Fig. 2). At lower concentrations native and asialo-TF showed no binding. The binding of agalacto-AGP and agalacto-TF could be totally inhibited by 1 to 10 mM N-acetylglucosamine (Fig. 4), a specific hapten sugar of BS-II lectin (22). Nacetylgalactosamine did not inhibit the binding even at 100 mM concentration (Fig. 4). This confirmed that the
102
m 1 o-3
10-2
10-l
a, -acid glycoprotein
100
IO'
10"
@g/ml)
FIG. 3. Binding of glycosylation variants of cu,-acid glycoprotein to microwells coated with BS-II lectin. Native AGP (O), asialo-AGP (A), and agalacto-AGP (0) were incubated at different concentrations in the microwells and the bound glycoproteins were detected by Eu-labeled AGP-specific antibody.
386
PARKKINEN
Inhibitor concentration
(mM)
Inhibitor concentration
(mM)
AND
FIG. 4.
Inhibition of binding of the agalacto-derivatives of a,-acid glycoproteins (A) and transferrin (B) to microwells coated with BSII lectin. Agalacto-AGP (10 pg/ml) and agalacto-TF (50 rg/ml) were incubated in the microwells with different concentrations of N-acetylglucosamine (0) or N-acetylgalactosamine (0). The bound glycoproteins were detected by Eu-labeled antibodies.
binding of the agalacto-derivatives to the lectin-coated microwells was due to a lectin-carbohydrate interaction specific for terminal N-acetylglucosamine residues. DISCUSSION
The present paper describes a novel lectin-immunofluorometric assay for the determination of glycoproteins containing terminal N-acetylglucosamine residues. The method sensitively detects the desialylated and degalactosylated derivatives of AGP and TF but does not measure the native or desialylated glycoproteins. The specificity of the method for glycoproteins containing terminal N-acetylglucosamine residues was confirmed by the ability of free N-acetylglucosamine to totally inhibit binding of the agalacto-glycoproteins to the lectin-coated wells. The measuring range of LIFMA is large, nearly lOOOfold, which is typical for immunofluorometric assaysemploying time-resolved fluorescence as the detection system (23,24). LIFMA was less sensitive in the detection
OKSANEN
of the agalacto-glycoproteins than the immunofluorometric assay using solid-phase bound antibody. For agalacto-AGP, the difference was about lo-fold and for agalacto-TF about IOOO-fold. This is evidently due to the lower affinity of the lectin used on the solid phase as compared to the avidity of the polyclonal antibodies. However, the degree of desialylation and degalactosylation of the agalacto-derivatives studied in the present study was not determined and it is possible that the sensitivity of LIFMA for glycoproteins totally lacking galactose is higher than that observed. Interestingly, agalacto-AGP could be detected in LIFMA with about loo-fold higher sensitivity than agalacto-TF. This is probably due to the higher degree of polyvalency of the terminal sugar residues present in AGP as compared to TF (26,27). Polyvalency of the sugar residues has earlier been shown to remarkably affect the affinity of lectins toward glycoproteins (28). Thus, LIFMA obviously detects with higher sensitivity heavily glycosylated glycoproteins such as AGP than it detects those containing less carbohydrates. Recently, a lectin-enzyme immunoassay for the detection of sialovariants of TF has been described (29). This method, called LEIA, employed a solid phase-bound anti-TF antibody and a biotin-conjugated Ricinus communis lectin complexed with avidin-conjugated alkaline phosphatase for detection, In addition to the detection method and the lectin used, the major difference between LEIA and LIFMA is the side of the sandwich on which the lectin is used. A theoretical problem in both systems is the capacity of the solid phase. When antibody is immobilized on the solid phase, other glycosylation variants of the same glycoprotein compete for binding to the immobilized antibody and may limit the binding of the glycosylation variant studied. On the other hand, when lectin is used on the solid phase, other proteins containing similar glycan chains may cause interference. In this respect, an obvious advantage of LIFMA is the high sensitivity of the detection system and wide measuring range which should allow proper dilution of the sample so that interfering glycoproteins would not limit the capacity of the solid phase. In conclusion, the lectin-immunofluorometric assay described in the present paper is a promising method for the rapid and sensitive quantification of glycosylation variants of glycoproteins. Application of the method for clinically important glycoproteins is currently in progress. ACKNOWLEDGMENTS The authors thank Dr. Ulf-Hikan Stenman, cussions and critical reading of the manuscript. ported by the Sigrid Jusilius Foundation.
M.D., for inspiringdisThis work was sup-
REFERENCES 1. Kobata, Robbins, York.
A. (1964) in Biology P. W., Eds.), Vol.
of Carbohydrates (Ginsburg, V., and 2, Chap. 2, pp. 87-162, Wiley, New
LECTIN-IMMUNOFLUOROMETRIC 2. Swiedler, and Hart,
S. J., Freed, J. H., Tarentino, G. W. (1985) J. Biol. Chem.
A. L., Plummer, 260.4046-4054.
T. H., Jr.,
3. Krusius, T., and Ruoslahti, E. (1982) J. Biol. Chem. 257, 34533457. 4. Williams, D. B., and Lennarz, W. J. (1984) J. Biol. Chem. 259, 5105-5114. 5. Sheares, B. T., and Rohbins, P. W. (1986) Proc. Natl. Acad. Sci. USA 83,1993-1997. 6. Parekh, R. B., Tse, A. G. D., Dwek, R. A., Williams, A. F., and Rademacher, T. W. (1987) Embo J. 6,1233-1244. 7. Feizi, T. (1985) Nature (London) 314,53-57. 8. Finne, J. (1985) Trends Biochem. Sci. 10,129-132. 9 Gahmberg, G. C., Ekblom, M., and Andersson, L. C. (1984) Proc. Natl. Acad. Sci. USA 8 1,6752-6756. 10 Mackiewicz, A., Ganapathi, M. K., Schultz, and Kushner, I. (1987) J. Exp.Med. 166,253-258. 11 Yamamoto, K., Tsuji, T., Tarutani, O., and Osawa, T. (1984) Eur. J. Biochem. 143,133-144. 12. Yamashita, K., Totani, K., Kuroki, M., Matsuoka, Y., Ueda, I., and Kobata, A. (1987) Cancer Res. 47,3451-3459. 13. Endo, T., Nishimura, R., Kawano, T., Mochizuki, M., and Kobata, A. (1987) Cancer Res. 47,5242-5245. 14. Parekh, R. B., Dwek, R. A., Sutton, B. J., Fernandes, D. L., Leung, A., Stanworth, D., Rademacher, T. W., Mizuochi, T., Taniguchi, T., Matsuta, K., Takeuchi, F., Nagano, Y., Miyamoto, T., and Kobata, A. (1985) Nature (London) 316,452-457. 15. Parekh, R. B., Roitt, I. M., Isenberg, D. A., Dwek, R. A., Ansell, B. M., and Rademacher, T. W. (1988) Larzcet i, 966-969. 16. Mackiewicz, A., Marcinkowska-Pieta, R., Ballou, S., Mackiewicz, and Kushner, I. (1987) Arthritis Rheumat. 30,513-518. *A.
387
ASSAY 17. Hachulla, 911-915.
E., Laine,
A., and Hayem,
A. (1988)
18. Jezequel, M., Seta, N. S., Corbic, M. M., Feger, G. M. (1988) Clin. Chim. Acta 176.49-57. 19. Beg-Hansen,
T. C. (1973)
20. Wide, L., Bennich, 1105-1107. 21. Miles, L. E. M., 186-189.
Anal.
Biochem.
C. N. (1968)
Nature
Stenman, U.-H., J., and Sepplli, 736.
25. Ashwell, eas Mol.
Alftan, H., Ranta, M. (1987) J. Clin.
G., and Morrell, Biol. 41,99-128.
219,
in The Lectins, ProperBiology and Medicine I. J., Eds.), pp. 91-93,
Adv.
26. Spik, G., Bayard, B., Fournet, B., Strecker, Montreuil, J. (1975) FEBS Lett. 50,296-299.
28. Stowell,
ii,
H., and
Liiv-
T., Vartiainen, E., Jalkanen, Endocrinol. Metab. 64, 730-
A. E. (1974)
B., Montreuil, J., Strecker, 27. Fournet, J., Vliegenthart, J. F. G., Binette, Biochemistry 17,5206-5214.
Lancet
(London)
23. Hemmila, I., Dakubu, S., Mukkala, V.-M., Siitari, gren, T. (1984) Anal. Biochem. 137,335-343. 34
43,
J. M., and Durand,
S. G. 0. (1967)
22. Goldstein, I. J., and Pore& R. D. (1986) ties, Functions, and Applications in (Liener, I. E., Sharon, N., and Goldstein, Academic Press, Orlando, FL.
I-’
Chem.
66,480-488.
H., and Johansson, and Hales,
Clin.
Enzymol.
Related
G., Bouchelet,
Ar-
S., and
G., Dorland, L., Haverkamp, J. P., and Schmid, K. (1978)
C. P., Lee, R. T., and Lee, Y. C. (1980)
Biochemistry
19,
4904-4908. 29. Pekelharing, (1987) Anal.
J. M., Vissers, P., Peters, Biochem. 165,320-326.
H. A., and
Leijnse,
B.