- Email: [email protected]
Lectin affinity bioassay: an assay method for glycoprotein enzyme
53
Biochimica et Biophysica Acta, 990 (1989) 53-58
Elsevier BBA 23036
Lectin affinity bioassay: an assay method for glycoprotein enzyme Asgar Electricwala Dioision
0/ Biotechnology,
PHLS Centre/or Applied M icrobiology and Research, Porton Dawn, Salisbury (U.K.)
(Received 23 May 1988)
Key words: Bioassay; Plasminogen activator; Lectin
A simple and sensitive chromogenic microtitre plate assay for glycoprotein enzymes is described, using melanoma tissue plasminogen activator (t-PA) as a model enzyme. The assay is based on the binding of t-PA to immobilised lectin and quantitating the bound enzyme with plasminogen, fibrinogen fragments and chromogenic substrate S-2251 on an ELISA plate reader. Seven different lectins were examined for the binding of t-PA, and of these, concanavalin A was chosen for subsequent studies. The specificity of this binding can be inhibited dose-dependently in the presence of n-mannose and methyl a-D-mannoside, but not by D-glucose and n-Iactose, The lower limit of the sensitivity of this assay is about 0.5 IV ImI. Comparison of the dose-response curves indicates that the sensitivity af this assay method is very similar to that of bioimmunoassay using anti-t-PA IgG to capture the antigen. The applicability of this method to other glycoprotein enzymes was also evaluated using alkaline phosphatase from bovine mucosa. The specificity of this method was related to the choice of substrate and this was shown by analysis of a mixture of t-PA and alkaline phosphatase. It is suggested that this assay can be adapted for the analysis of complex glycoprotein mixtures with the appropriate choice of lectin and substrate.
Introduction Over the past few years, several assay methods have been developed for the quantitation of plasminogen activators in body fluids and other media [1-5]. The majority of these assays are some form of immunoassays, as these have the advantage of being selective and sensitive, but they rely on the availability of specific polyclonal or monoclonal antibodies to plasminogen activator. As an alternative to these assay methods and one which does not have the prerequisite of antibodies to plasminogen activator, or to any glycoprotein enzyme in general, an assay method was developed based on the binding reaction between a lectin and a glycoprotein. Lectins have the advantage in that they are easily available commercially, they are of non-immune origin and are specific in their interaction with a particular sugar residue on the glycoproleins. In view of this specificity, leetins have been widely used for the purification of various glycoproteins [6-8], cell agglutination studies, mitogenic assays and in studies involving cell
Abbreviations: t-PA, tissue plasminogen activator: Con A, concanavalin A. Correspondence: A. Electricwala, Div ision of Biotechnology, C.A.M .R., Porton Down, Salisbury, wn«, SP4 OJO, U.K. 0304-4165/89/S03.50
~
surface receptors and cell development (for review see Refs. 9, 10). Using the specificity of the lectins, an assay method was developed for the estimation of glycoproteins which have an enzyme activity. Th is assay has been termed lectin affinity bioassay, or LAB in short. In the present study, t-PA was chosen for study as it is a glycoprotein and a serine proteinase that catalyses the conversion of plasminogen to plasmin [11], the reaction rate being increased several-fold in the presence of fibrinogen fragments [12]. The method described here is a simple and sensitive technique applicable to other glycoprotein enzymes in general. An abstract of this method will be presented at the IXth International Congress on Fibrinolysis, Amsterdam. Materials and Methods Bowes melanoma t-PA and polyclonal goat anti-t-PA IgG were purchased from Biopool, Sweden. The specific activity of the enzyme was 500 IU//Lg of t-PA based on the International Reference preparation of human t-PA (83/517) . Flat-bottomed 96-well polyvinyl microtitre plates were purchased from Flow Laboratories. Foetal calf serum and Eagle's medium were obtained from Imperial Laboratories, Salisbury. Human plasminogen, human fibrinogen, chromogenic substrate S-2251 were obtained from Kabi, Sweden. All the different lectins,
1989 Elsevier Science Publishers B.V. (Biomedical Division)
54 i.e, concanavalin A (Con A), wheatgerm agglutinin, phytohaemagglutinin, garden pea lectin, ricin lectin, pokeweed mitogen and lentil lectin, bovine intestinal alkaline phosphatase, p-nitrophenyl phosphate, Tween20 and Tween-80 were from Sigma Chemical Co., Poole. The specific activity of alkaline phosphatase was 900 unitsy'mg, as quoted by the supplier, and was used without further purification. All other reagents used were of the highest purity available. Cyanogen bromide fragments of human fibrinogen were prepared according to the method of Verheijen et al. [13J.
Cell culture Bowes melanoma cells were cultured in Eagle's minimum essential medium containing 10% foetal calf serum. On attaining 70% confluency, the cell monolayer was washed twice and further incubated at 37 0 C with serum-free culture medium containing 0.01% Tween-BO. The supernatant was harvested after 24 hand t-PA activity in International Units was determined by the clot lysis method as described by Electricwala et al. [14]. Assay protocol for t-PA The assay was carried out in 96-well flat-bottomed microtitre plates. Each well was filled with 200 ILl solution of lectin dissolved in 0.1 M sodium bicarbonate buffer, pH 9.4. After overnight incubation at room temperature in a humid covered box, the lectin solution was emptied and the wells washed four times by emptying and filling with phosphate-buffered saline (PBS) containing 0.1% Tween-20 and shaken dry. The wells were then charged with 200 ILl of varying concentrations of t-PA in the range of 0.1-10 IV/m!. The dilutions were carried out in 20 mM sodium phosphate buffer, pH 7.5 containing 0.1 M NaCI and 0.01% Tween-80. The plate was incubated for between 1 and 3 h at room temperature with constant shaking. It was washed again four times, shaken dry and then 200 ILl of the substrate solution was added to each row of wells with the aid of a multichannel pipette. The substrate solution consisted of a mixture of 0.27 ILM plasminogen, 0.6 mM S-2251 and 120 ILg/ml of CNBr fragments of human fibrinogen in 50 mM Tris-HCI buffer, pH 7.4. The plate was then incubated for 1 h at 37 0 C for the colour to develop and the amount of p-nitrophenol released was measured at 405 om using a Titertek twinreader. Bioimmunoassay for t-PA The assay was carried out essentially as described by Mahmoud and Gaffney [2] with some modifications. Anti-t-PA IgG was bound to the plate and varying concentrations of t-PA were allowed to react with the immobilised antibody. The bound enzyme was then measured with the substrate mixture in a similar manner
to that described above for LAB, so that direct comparison of results could be made between the two assay methods.
Assay protocol for other glycoprotein enzyme Another glycoprotein enzyme, alkaline phosphatase from bovine intestinal mucosa, was chosen to obtain a dose-response curve using LAB. Wens of microtitre plates were coated with Con A as described before and 200 ILl of enzyme solution in the concentration range 2-40 unitsyrnl was incubated for 90 min at room temperature. After washing, the bound enzyme was quantitated by adding 200 ILl of the substrate solution consisting of 4 rngy'rnl of p-nitrophenyl phosphate in 0.1 M glycine, 1 mM MgCI 2 , I mM ZnCI 2 , pH 10.5. After incubation at 37 0 C, the amount of p-nitrophenol released was monitored at 405 nm. Results Effects of various lectins on the binding of topA In order to establish the affinity of t-PA for various lectins and the amount of each lectin necessary to coat the wells for maximal binding of t-PA, rows of wells were coated with different concentrations of seven lectins and the binding of t-PA to these was evaluated. The lectins studied were Con A, wheatgerm lectin, phytohaemagglutinin, garden pea lectin, ricin lectin, pokeweed mitogen and lentil lectin in the concentration range of 0.2-2000 ILg/well. After washing the wells, a fixed amount of t-PA (10 IU/ml) was added to all the wells, incubated for 90 min and the amount of bound enzyme was measured with the substrate mixture as described in Materials and Methods. The results obtained are shown in Table I and indicate that a relatively higher binding of t-PA was observed with Con A, lentil lectin and garden pea lectin, with moderate binding to pokeweed mitogen, ricin lectin and wheatgerm lectin and relatively little binding to phytohaemagglutinin, compared to non-lectin-contain-
TABLE I Comparison of various lee/ins in LAB Lectin
A 40s/h (/-lg/well)
Con A Garden pea Lentil Pokeweed Ricin Wheatgerm Phytohaernagglutinin
0.2
1
10
100
500
1000
2000
0.69 0.87 0.94 0.79 0.69 0.76
0.70 1.53 1.53 1.13 0.86 0.91
0.88 1.71 1.64 1.40 1.19 0.94
1.16 1.98 1.79 1.22 0.91 0.83
1.78 2.15 1.93 1.16 0.97 0.78
1.97 2.20 1.87 1.18 n.d. • 0.64
2.05 2.20 1.08 n.d. 0.52
0.52
0.55
0.40
0.28
0.30
0.31
0.22
• n.d, = not determined.
1.46
55 A
B
15
j
!
.A.
//
1,0
O'-------.l----+---:,--+----l..-----+----/.,-_ _+_+---' 2 4 6 2 4
6
t-PA (IU/ml)
Fig. 1. Comparison of dose-response curves of t-PA by LAB (A) and bioimmunoassay (B). Various dilutions of t-PA in the concentration range of 0.1-10 IU/ml were incubated for 90 min either on immobilised Con A (e) or anti-t-PA IgO (.A.). After washing, the bound enzyme was determined with a mixture of plasminogen, S-2251 and fibrinogen fragments.
ing control wells (A 4os/h = 0.15). It was found that the optimum amount of the various lectins needed for maximal binding of t-PA was different. The amount of bound t-PA increased with increasing concentrations of Con A, garden pea and lentil leetins, and for these lectins, concentrations greater than 500 /Lg/well were preferred. For pokeweed mitogen, ricin and wheatgerm lee tins, the optimal concentration was in the range of 10-500 /lg/well. In contrast, the response obtained with phytohaemagglutinin was negligible in the entire concentration range studied. Although Con A, garden pea and lentil leetins all provided the maximum dose-dependent response, Con A was chosen for all subsequent studies as it is similar to garden pea and lentil lectin in its specificity for sugar residues and it is relatively inexpensive. Thus, a solution of 5 mgyrnl of Con A was used to coat the wells of microtitre plates in further studies.
no significant binding of t-PA was observed to the control wells treated with coating buffer only. Fig. 1B shows the dose-response curve of t-PA obtained by bioimmunoassay under similar conditions. The results indicate that the sensitivity of the bioimmunoassay is very similar to that obtained by LAB, with a lower detection limit of 0.5 IU/ml under the assay conditions.
/
1.5
1.0 U'1
Dose-response of t-P.A on immobilised Con A and its comparison with bioimmunoassay To evaluate the sensitivity of the binding of t-PA to the immobilised lectin, increasing amounts of t-PA in the concentration range of 0.1-10 IU/ml were incubated for 90 min in wells coated overnight with Con A. The activity of the bound enzyme was estimated by reaction with 8-2251 and plasminogen in the presence of fibrinogen fragments and the absorbance was monitored after 60 min incubation at 37 0 C. A plot of absorbance versus log concentration of t-PA showed a sigmoidal dose-response curve with the lower limit of detection at about 0.5 IU/ml t-PA (Fig. lA). However,
Q
"f
05
•
O'-------+---*---~----k-----'
56
1.5
1.0
! o - - - - v ' - - - - - V ' - - - - - 'vi / - - - - V ...
4 t-PA (IU/ml)
oil.
10
1 2 4 Alkaline phosphatase (units Iml)
10
Fig. 3. Analysis of a mixture of t-PA and alkaline phosphatase by LAB. t-PA. alkaline phosphatase and a mixture of t-PA and alkaline phosphatase were incubated on Con A-coated wells of micro titre plate. After 90 min, the bound enzyme activity was determined with substrate mixture for t-PA. In a duplicate plate. the bound alkaline phosphatase activity was measured with the substrate p-nitrophenyl phosphate. v, t-PA; ..... alkaline phosphatase; o. t-PA+ alkaline phosphatase.
However, the sensitivity of both these assay methods can be increased even further by prolonging the incubation time of the reaction with the substrate mixture beyond 60 min, but for the sake of comparison in these studies, a fixed incubation time of 60 min was adhered to for all the experiments. Assay for other glycoprotein enzymes In order to evaluate the applicability of this assay method to other glycoprotein enzymes, bovine intestinal alkaline phosphatase was chosen to study the dose-response effect by the LAB method. After incubation of various concentrations (2-40 unitsyrnl) of enzyme solution with immobilised Con A, the activity of the bound enzyme was measured after 20 min incubation with the substrate. The results obtained are shown in Fig. 2 as a semi-logarithmic plot of absorbance versus enzyme concentration. The data indicate that under these assay conditions, the sensitivity of the assay was about 2-5 unitsyrnl of enzyme, but can be increased further with prolonged incubation with the substrate. Analysis of a mixture of t-PA and alkaline phosphatase by
LAB In an attempt to study the interference by other glycoproteins during the binding and measurement of t-PA to immobilised lectin, a mixture of t-PA and alkaline phosphatase was investigated by the LAB method. Two separate microtitre plates, precoated with Con A, were each charged with a row of t-PA (0.6-10 IU /ml), alkaline phosphatase (0.6-10 unitsy'ml) and a
mixture of equal amounts of t-PA and alkaline phosphatase. After incubation for 90 min at room temperature, the activity of bound t-PA on one plate was monitored using the substrate mixture for measurement of t-PA activity while bound alkaline phosphatase activity was measured with the substrate, p-nitrophenyl
1.5
1.0
05
/
/ 2 t-PA (IU /mr)
5
10
15
Fig. 4. Dose-response curve of t-PA in the conditioned medium. Various dilutions of serum-free conditioned medium containing t-PA in the range of 0.5-15 IU/ml were incubated on immobilised Con A for 90 min. After washing. the bound activity was determined with the substrate mixture for t-PA.
57 phosphate, on the other plate. Both the plates were monitored for 1 h at 37 0 C in a Titertek twinreader. The results obtained are shown in Fig. 3 and shows that the binding, and activity, of t-PA in the absence and presence of alkaline phosphatase was unaffected except at higher concentrations, where some competition was observed. As expected alkaline phosphatase had no direct effect on the substrate mixture used for t-PA. Similar results were also obtained for the binding and activity of alkaline phosphatase, in the absence and presence of t-PA.
Measurement of t-PA activity in culture supernatant by LAB In an attempt to evaluate whether the present assay method can be used to monitor the activity of t-PA in cell culture supernatant, serum-free conditioned medium from melanoma cell culture was used in the study. t·PA activity in the sample was initially determined in international units by fibrin clot lysis assay [14]. The medium was diluted appropriately to obtain solutions in the range of 0.5-15 IU/ml. These were then analysed by the present assay method, as described previously. The results obtained are shown in Fig. 4 and indicate that the dose-response effect obtained with the conditioned medium is similar to that obtained with the purified enzyme in the same concentration range. Discussion For many years, lectins have been widely used in various studies, one of which being the purification of glycoproteins by affinity chromatography [6-10]. Such applications of lectins have been based on the relative specifici ty of the reaction between a lectin and specific sugar residue on a glycoprotein. This property of lectin was therefore used to develop a simple and sensitive assay method for the estimation of glycoprotein enzymes, the specificity of the assay being achieved by a combination of lectin binding and enzyme activi ty determination. In the present study, t-PA was used as a model glycoprotein enzyme. The optimum conditions of the LAB method were determined using the lectin Con A for the binding of t-PA (see below). It was found that the blocking of non-specific sites on the wells with albumin, before incubation with t-PA, was not necessary as it had little or no effect on the dose-response curve of t-PA. It may be the high concentration of lectin used would occupy all the reactive sites on the well and therefore the blocking step was omitted from the final protocol. Also, the optimum incubation time for the interaction of Con A and t-PA was determined in a separate experiment and was found to be about 90 min. The sensitivity of the LAB method for the determination of t-PA concentration in buffer was about 0.5
IV /m!. This detection limit was found to be similar to that obtained by bioimmunoassay, where instead of lectin, specific anti-t-PA IgO was used to capture the antigen. There are several different assay methods reported in the literature for the determination of t-PA concentration [1-5]. The sensitivity of some of these immunoassays has been shown to be in the range of 0.12-1 ngy'm! [2-4], which when based on the specific activity value of 500 IU per p.g of enzyme would be equivalent to about 0.06-0.5 IU /m!. Thus, the present LAB method is nearly as sensitive as other immunoassays. One of the advantages of the LAB method is that it involves few steps and can be carried out within a short period. This is of special importance when using labile glycoproteins which may be unstable at room temperature over longer periods. However, this assay technique does not discriminate between between different isotypes of a glycoprotein enzyme. A dose-response graph, similar to that shown in Fig. lA, was obtained when urokinase-type plasminogen activator was analysed by this assay method. Several lectins were investigated to study their binding affinity for t-PA by the LAB method. The results obtained indicate that the binding of t-PA was relatively greater for Con A, garden pea and lentil lectins compared to the others. These lectins are specific for a-Omannose and this suggests that the oligosaccharide core on the t-PA molecule might have high-mannose type, instead of complex or hydrid type, structure [15]. Recently, it has been shown that treatment of t-PA with o-mannosidase increases its enzyme activity, and this suggested that the carbohydrate part of the molecule plays an important role in substrate affinity [16]. Other lectins, whieh have sugar specificity for galactose, glucose and/or its derivatives, have relatively lower affinity for t-PA when analysed by the LAB technique. It may be that either these sugar residues are absent from the oligosaccharide structure of t-PA or, if present. then their interaction with the lectin is hindered, possibly due to the presence of terminal sialic acid residue(s). Thus, by the use of the LAB method with different lee tins, one can deduce some information about the sugar residues present on the glycoprotein molecule. In competitive binding experiments, it was observed that the binding of t-PA to Con A was significantly inhibited in the presence of n-mannose and methyl e-n-mannoaide, but these had little or no effect when glucose or lactose was used as a competing ligand (data not shown). This may also explain why Con A-agarose is widely used as an affinity matrix for the purification of t-PA and its subsequent elution with methyl a-Dmannoside [11]. The present assay method offers a useful alternative technique over other immunoassays as it has been shown that the sensitivity of this assay is very similar to that of
58
bioimmunoassay and is little affected when non-relevant glycoproteins are present. Such an assay method would be useful under conditions where other immunoassays are not practicable due to non-availability of antibodies to a particular glycoprotein enzyme. This technique would also be of potential value in 'dissecting' a complex mixture of glycoproteins with the appropriate choice of substrate or in discriminating between glycosylated and non-glycosylated forms of a recombinant protein, but it may not be possible to distinguish between different degrees of glycosylation of the same molecule. Acknowledgements
My sincere thanks to Prof. Tony Atkinson for his constant encouragement and helpful discussions on various aspects of the work described in this paper. Thanks are also due to Miss Z. Moola and Mr. R. Ling for technical assistance. References Ranby, M., Norrman, B. and Wallen, P. (1982) Thromb. Res. 27, 743-749.
2 Mahmoud, M. and Gaffney, P.J. (1985) Thromb. Haemostas. 53, 356-359. 3 Korninger, C.. Speiser, W., Wojta, J. and Binder, B.R. (1986) Thromb. Res. 41, 527-535. 4 Schleef, R.R., Wagner, N.V., Sinha, M. and Loskutoff, D.J. (1986) Thromb. Haemostas. 56, 328-332. 5 Holvoet, P., Boes, J. and Collen, D. (1987) Blood 69, 284-289. 6 Dufau, M.L., Tsuruhara, T. and Catt, K.J. (1972) Biochim. BiDphys. Acta 278, 281-292. 7 Cummings, R.D. and Kornfeld, S. (1982) J. BioI. Chern. 257, 11235-11240. 8 Hayman, M., Skehel, J.J. and Crumpton, M.J. (1973) FEBS Lett. 29, 185-188. 9 Gold, E.R. and Balding, P. (1975) Plant and Animal Lectins, Excerpta Medica, Amsterdam. 10 Bog-Hansen, T.C. (ed.) (1981-1986) Lectins: Biology, Biochemistry, Clinical Biochemistry, Vols, 1-5, W. de Gruyter, Berlin. 11 Rijken, D.C. and Collen, D. (1981) J. Bioi. Chern, 256, 7035-7041. 12 Hoylaerts, M., Rijken, R.C., Lijnen, H.R. and Collen, D. (1982) J. BioI. Chern. 257, 2912-2919. 13 Verheijen, J.H., Mullarrt, E., Chang, G.T.G., Kluft, C. and Wijngaards, G. (1982) Thromb. Haemostas, 48, 266-269. 14 Electricwala, A., Ling, R.J., Sutton, P.M., Grirfiths, B.J., Riley, P.A. and Atkinson, T. (1985) Thromb. Haemostas. 53,200-203. 15 Kornfeld, R. and Kornfeld, S. (1985) Annu. Rev. Biochem. 54, 631-664. 16 Opdenakker, G., Van Damme, J., Bosman, F., Billiau, A. and Seer, P.D. (1986) Proc. Soc. Exp. BioI. Med. 182,248-257.
Biochimica et Biophysica Acta, 990 (1989) 53-58
Elsevier BBA 23036
Lectin affinity bioassay: an assay method for glycoprotein enzyme Asgar Electricwala Dioision
0/ Biotechnology,
PHLS Centre/or Applied M icrobiology and Research, Porton Dawn, Salisbury (U.K.)
(Received 23 May 1988)
Key words: Bioassay; Plasminogen activator; Lectin
A simple and sensitive chromogenic microtitre plate assay for glycoprotein enzymes is described, using melanoma tissue plasminogen activator (t-PA) as a model enzyme. The assay is based on the binding of t-PA to immobilised lectin and quantitating the bound enzyme with plasminogen, fibrinogen fragments and chromogenic substrate S-2251 on an ELISA plate reader. Seven different lectins were examined for the binding of t-PA, and of these, concanavalin A was chosen for subsequent studies. The specificity of this binding can be inhibited dose-dependently in the presence of n-mannose and methyl a-D-mannoside, but not by D-glucose and n-Iactose, The lower limit of the sensitivity of this assay is about 0.5 IV ImI. Comparison of the dose-response curves indicates that the sensitivity af this assay method is very similar to that of bioimmunoassay using anti-t-PA IgG to capture the antigen. The applicability of this method to other glycoprotein enzymes was also evaluated using alkaline phosphatase from bovine mucosa. The specificity of this method was related to the choice of substrate and this was shown by analysis of a mixture of t-PA and alkaline phosphatase. It is suggested that this assay can be adapted for the analysis of complex glycoprotein mixtures with the appropriate choice of lectin and substrate.
Introduction Over the past few years, several assay methods have been developed for the quantitation of plasminogen activators in body fluids and other media [1-5]. The majority of these assays are some form of immunoassays, as these have the advantage of being selective and sensitive, but they rely on the availability of specific polyclonal or monoclonal antibodies to plasminogen activator. As an alternative to these assay methods and one which does not have the prerequisite of antibodies to plasminogen activator, or to any glycoprotein enzyme in general, an assay method was developed based on the binding reaction between a lectin and a glycoprotein. Lectins have the advantage in that they are easily available commercially, they are of non-immune origin and are specific in their interaction with a particular sugar residue on the glycoproleins. In view of this specificity, leetins have been widely used for the purification of various glycoproteins [6-8], cell agglutination studies, mitogenic assays and in studies involving cell
Abbreviations: t-PA, tissue plasminogen activator: Con A, concanavalin A. Correspondence: A. Electricwala, Div ision of Biotechnology, C.A.M .R., Porton Down, Salisbury, wn«, SP4 OJO, U.K. 0304-4165/89/S03.50
~
surface receptors and cell development (for review see Refs. 9, 10). Using the specificity of the lectins, an assay method was developed for the estimation of glycoproteins which have an enzyme activity. Th is assay has been termed lectin affinity bioassay, or LAB in short. In the present study, t-PA was chosen for study as it is a glycoprotein and a serine proteinase that catalyses the conversion of plasminogen to plasmin [11], the reaction rate being increased several-fold in the presence of fibrinogen fragments [12]. The method described here is a simple and sensitive technique applicable to other glycoprotein enzymes in general. An abstract of this method will be presented at the IXth International Congress on Fibrinolysis, Amsterdam. Materials and Methods Bowes melanoma t-PA and polyclonal goat anti-t-PA IgG were purchased from Biopool, Sweden. The specific activity of the enzyme was 500 IU//Lg of t-PA based on the International Reference preparation of human t-PA (83/517) . Flat-bottomed 96-well polyvinyl microtitre plates were purchased from Flow Laboratories. Foetal calf serum and Eagle's medium were obtained from Imperial Laboratories, Salisbury. Human plasminogen, human fibrinogen, chromogenic substrate S-2251 were obtained from Kabi, Sweden. All the different lectins,
1989 Elsevier Science Publishers B.V. (Biomedical Division)
54 i.e, concanavalin A (Con A), wheatgerm agglutinin, phytohaemagglutinin, garden pea lectin, ricin lectin, pokeweed mitogen and lentil lectin, bovine intestinal alkaline phosphatase, p-nitrophenyl phosphate, Tween20 and Tween-80 were from Sigma Chemical Co., Poole. The specific activity of alkaline phosphatase was 900 unitsy'mg, as quoted by the supplier, and was used without further purification. All other reagents used were of the highest purity available. Cyanogen bromide fragments of human fibrinogen were prepared according to the method of Verheijen et al. [13J.
Cell culture Bowes melanoma cells were cultured in Eagle's minimum essential medium containing 10% foetal calf serum. On attaining 70% confluency, the cell monolayer was washed twice and further incubated at 37 0 C with serum-free culture medium containing 0.01% Tween-BO. The supernatant was harvested after 24 hand t-PA activity in International Units was determined by the clot lysis method as described by Electricwala et al. [14]. Assay protocol for t-PA The assay was carried out in 96-well flat-bottomed microtitre plates. Each well was filled with 200 ILl solution of lectin dissolved in 0.1 M sodium bicarbonate buffer, pH 9.4. After overnight incubation at room temperature in a humid covered box, the lectin solution was emptied and the wells washed four times by emptying and filling with phosphate-buffered saline (PBS) containing 0.1% Tween-20 and shaken dry. The wells were then charged with 200 ILl of varying concentrations of t-PA in the range of 0.1-10 IV/m!. The dilutions were carried out in 20 mM sodium phosphate buffer, pH 7.5 containing 0.1 M NaCI and 0.01% Tween-80. The plate was incubated for between 1 and 3 h at room temperature with constant shaking. It was washed again four times, shaken dry and then 200 ILl of the substrate solution was added to each row of wells with the aid of a multichannel pipette. The substrate solution consisted of a mixture of 0.27 ILM plasminogen, 0.6 mM S-2251 and 120 ILg/ml of CNBr fragments of human fibrinogen in 50 mM Tris-HCI buffer, pH 7.4. The plate was then incubated for 1 h at 37 0 C for the colour to develop and the amount of p-nitrophenol released was measured at 405 om using a Titertek twinreader. Bioimmunoassay for t-PA The assay was carried out essentially as described by Mahmoud and Gaffney [2] with some modifications. Anti-t-PA IgG was bound to the plate and varying concentrations of t-PA were allowed to react with the immobilised antibody. The bound enzyme was then measured with the substrate mixture in a similar manner
to that described above for LAB, so that direct comparison of results could be made between the two assay methods.
Assay protocol for other glycoprotein enzyme Another glycoprotein enzyme, alkaline phosphatase from bovine intestinal mucosa, was chosen to obtain a dose-response curve using LAB. Wens of microtitre plates were coated with Con A as described before and 200 ILl of enzyme solution in the concentration range 2-40 unitsyrnl was incubated for 90 min at room temperature. After washing, the bound enzyme was quantitated by adding 200 ILl of the substrate solution consisting of 4 rngy'rnl of p-nitrophenyl phosphate in 0.1 M glycine, 1 mM MgCI 2 , I mM ZnCI 2 , pH 10.5. After incubation at 37 0 C, the amount of p-nitrophenol released was monitored at 405 nm. Results Effects of various lectins on the binding of topA In order to establish the affinity of t-PA for various lectins and the amount of each lectin necessary to coat the wells for maximal binding of t-PA, rows of wells were coated with different concentrations of seven lectins and the binding of t-PA to these was evaluated. The lectins studied were Con A, wheatgerm lectin, phytohaemagglutinin, garden pea lectin, ricin lectin, pokeweed mitogen and lentil lectin in the concentration range of 0.2-2000 ILg/well. After washing the wells, a fixed amount of t-PA (10 IU/ml) was added to all the wells, incubated for 90 min and the amount of bound enzyme was measured with the substrate mixture as described in Materials and Methods. The results obtained are shown in Table I and indicate that a relatively higher binding of t-PA was observed with Con A, lentil lectin and garden pea lectin, with moderate binding to pokeweed mitogen, ricin lectin and wheatgerm lectin and relatively little binding to phytohaemagglutinin, compared to non-lectin-contain-
TABLE I Comparison of various lee/ins in LAB Lectin
A 40s/h (/-lg/well)
Con A Garden pea Lentil Pokeweed Ricin Wheatgerm Phytohaernagglutinin
0.2
1
10
100
500
1000
2000
0.69 0.87 0.94 0.79 0.69 0.76
0.70 1.53 1.53 1.13 0.86 0.91
0.88 1.71 1.64 1.40 1.19 0.94
1.16 1.98 1.79 1.22 0.91 0.83
1.78 2.15 1.93 1.16 0.97 0.78
1.97 2.20 1.87 1.18 n.d. • 0.64
2.05 2.20 1.08 n.d. 0.52
0.52
0.55
0.40
0.28
0.30
0.31
0.22
• n.d, = not determined.
1.46
55 A
B
15
j
!
.A.
//
1,0
O'-------.l----+---:,--+----l..-----+----/.,-_ _+_+---' 2 4 6 2 4
6
t-PA (IU/ml)
Fig. 1. Comparison of dose-response curves of t-PA by LAB (A) and bioimmunoassay (B). Various dilutions of t-PA in the concentration range of 0.1-10 IU/ml were incubated for 90 min either on immobilised Con A (e) or anti-t-PA IgO (.A.). After washing, the bound enzyme was determined with a mixture of plasminogen, S-2251 and fibrinogen fragments.
ing control wells (A 4os/h = 0.15). It was found that the optimum amount of the various lectins needed for maximal binding of t-PA was different. The amount of bound t-PA increased with increasing concentrations of Con A, garden pea and lentil leetins, and for these lectins, concentrations greater than 500 /Lg/well were preferred. For pokeweed mitogen, ricin and wheatgerm lee tins, the optimal concentration was in the range of 10-500 /lg/well. In contrast, the response obtained with phytohaemagglutinin was negligible in the entire concentration range studied. Although Con A, garden pea and lentil leetins all provided the maximum dose-dependent response, Con A was chosen for all subsequent studies as it is similar to garden pea and lentil lectin in its specificity for sugar residues and it is relatively inexpensive. Thus, a solution of 5 mgyrnl of Con A was used to coat the wells of microtitre plates in further studies.
no significant binding of t-PA was observed to the control wells treated with coating buffer only. Fig. 1B shows the dose-response curve of t-PA obtained by bioimmunoassay under similar conditions. The results indicate that the sensitivity of the bioimmunoassay is very similar to that obtained by LAB, with a lower detection limit of 0.5 IU/ml under the assay conditions.
/
1.5
1.0 U'1
Dose-response of t-P.A on immobilised Con A and its comparison with bioimmunoassay To evaluate the sensitivity of the binding of t-PA to the immobilised lectin, increasing amounts of t-PA in the concentration range of 0.1-10 IU/ml were incubated for 90 min in wells coated overnight with Con A. The activity of the bound enzyme was estimated by reaction with 8-2251 and plasminogen in the presence of fibrinogen fragments and the absorbance was monitored after 60 min incubation at 37 0 C. A plot of absorbance versus log concentration of t-PA showed a sigmoidal dose-response curve with the lower limit of detection at about 0.5 IU/ml t-PA (Fig. lA). However,
Q
"f
05
•
O'-------+---*---~----k-----'
56
1.5
1.0
! o - - - - v ' - - - - - V ' - - - - - 'vi / - - - - V ...
4 t-PA (IU/ml)
oil.
10
1 2 4 Alkaline phosphatase (units Iml)
10
Fig. 3. Analysis of a mixture of t-PA and alkaline phosphatase by LAB. t-PA. alkaline phosphatase and a mixture of t-PA and alkaline phosphatase were incubated on Con A-coated wells of micro titre plate. After 90 min, the bound enzyme activity was determined with substrate mixture for t-PA. In a duplicate plate. the bound alkaline phosphatase activity was measured with the substrate p-nitrophenyl phosphate. v, t-PA; ..... alkaline phosphatase; o. t-PA+ alkaline phosphatase.
However, the sensitivity of both these assay methods can be increased even further by prolonging the incubation time of the reaction with the substrate mixture beyond 60 min, but for the sake of comparison in these studies, a fixed incubation time of 60 min was adhered to for all the experiments. Assay for other glycoprotein enzymes In order to evaluate the applicability of this assay method to other glycoprotein enzymes, bovine intestinal alkaline phosphatase was chosen to study the dose-response effect by the LAB method. After incubation of various concentrations (2-40 unitsyrnl) of enzyme solution with immobilised Con A, the activity of the bound enzyme was measured after 20 min incubation with the substrate. The results obtained are shown in Fig. 2 as a semi-logarithmic plot of absorbance versus enzyme concentration. The data indicate that under these assay conditions, the sensitivity of the assay was about 2-5 unitsyrnl of enzyme, but can be increased further with prolonged incubation with the substrate. Analysis of a mixture of t-PA and alkaline phosphatase by
LAB In an attempt to study the interference by other glycoproteins during the binding and measurement of t-PA to immobilised lectin, a mixture of t-PA and alkaline phosphatase was investigated by the LAB method. Two separate microtitre plates, precoated with Con A, were each charged with a row of t-PA (0.6-10 IU /ml), alkaline phosphatase (0.6-10 unitsy'ml) and a
mixture of equal amounts of t-PA and alkaline phosphatase. After incubation for 90 min at room temperature, the activity of bound t-PA on one plate was monitored using the substrate mixture for measurement of t-PA activity while bound alkaline phosphatase activity was measured with the substrate, p-nitrophenyl
1.5
1.0
05
/
/ 2 t-PA (IU /mr)
5
10
15
Fig. 4. Dose-response curve of t-PA in the conditioned medium. Various dilutions of serum-free conditioned medium containing t-PA in the range of 0.5-15 IU/ml were incubated on immobilised Con A for 90 min. After washing. the bound activity was determined with the substrate mixture for t-PA.
57 phosphate, on the other plate. Both the plates were monitored for 1 h at 37 0 C in a Titertek twinreader. The results obtained are shown in Fig. 3 and shows that the binding, and activity, of t-PA in the absence and presence of alkaline phosphatase was unaffected except at higher concentrations, where some competition was observed. As expected alkaline phosphatase had no direct effect on the substrate mixture used for t-PA. Similar results were also obtained for the binding and activity of alkaline phosphatase, in the absence and presence of t-PA.
Measurement of t-PA activity in culture supernatant by LAB In an attempt to evaluate whether the present assay method can be used to monitor the activity of t-PA in cell culture supernatant, serum-free conditioned medium from melanoma cell culture was used in the study. t·PA activity in the sample was initially determined in international units by fibrin clot lysis assay [14]. The medium was diluted appropriately to obtain solutions in the range of 0.5-15 IU/ml. These were then analysed by the present assay method, as described previously. The results obtained are shown in Fig. 4 and indicate that the dose-response effect obtained with the conditioned medium is similar to that obtained with the purified enzyme in the same concentration range. Discussion For many years, lectins have been widely used in various studies, one of which being the purification of glycoproteins by affinity chromatography [6-10]. Such applications of lectins have been based on the relative specifici ty of the reaction between a lectin and specific sugar residue on a glycoprotein. This property of lectin was therefore used to develop a simple and sensitive assay method for the estimation of glycoprotein enzymes, the specificity of the assay being achieved by a combination of lectin binding and enzyme activi ty determination. In the present study, t-PA was used as a model glycoprotein enzyme. The optimum conditions of the LAB method were determined using the lectin Con A for the binding of t-PA (see below). It was found that the blocking of non-specific sites on the wells with albumin, before incubation with t-PA, was not necessary as it had little or no effect on the dose-response curve of t-PA. It may be the high concentration of lectin used would occupy all the reactive sites on the well and therefore the blocking step was omitted from the final protocol. Also, the optimum incubation time for the interaction of Con A and t-PA was determined in a separate experiment and was found to be about 90 min. The sensitivity of the LAB method for the determination of t-PA concentration in buffer was about 0.5
IV /m!. This detection limit was found to be similar to that obtained by bioimmunoassay, where instead of lectin, specific anti-t-PA IgO was used to capture the antigen. There are several different assay methods reported in the literature for the determination of t-PA concentration [1-5]. The sensitivity of some of these immunoassays has been shown to be in the range of 0.12-1 ngy'm! [2-4], which when based on the specific activity value of 500 IU per p.g of enzyme would be equivalent to about 0.06-0.5 IU /m!. Thus, the present LAB method is nearly as sensitive as other immunoassays. One of the advantages of the LAB method is that it involves few steps and can be carried out within a short period. This is of special importance when using labile glycoproteins which may be unstable at room temperature over longer periods. However, this assay technique does not discriminate between between different isotypes of a glycoprotein enzyme. A dose-response graph, similar to that shown in Fig. lA, was obtained when urokinase-type plasminogen activator was analysed by this assay method. Several lectins were investigated to study their binding affinity for t-PA by the LAB method. The results obtained indicate that the binding of t-PA was relatively greater for Con A, garden pea and lentil lectins compared to the others. These lectins are specific for a-Omannose and this suggests that the oligosaccharide core on the t-PA molecule might have high-mannose type, instead of complex or hydrid type, structure [15]. Recently, it has been shown that treatment of t-PA with o-mannosidase increases its enzyme activity, and this suggested that the carbohydrate part of the molecule plays an important role in substrate affinity [16]. Other lectins, whieh have sugar specificity for galactose, glucose and/or its derivatives, have relatively lower affinity for t-PA when analysed by the LAB technique. It may be that either these sugar residues are absent from the oligosaccharide structure of t-PA or, if present. then their interaction with the lectin is hindered, possibly due to the presence of terminal sialic acid residue(s). Thus, by the use of the LAB method with different lee tins, one can deduce some information about the sugar residues present on the glycoprotein molecule. In competitive binding experiments, it was observed that the binding of t-PA to Con A was significantly inhibited in the presence of n-mannose and methyl e-n-mannoaide, but these had little or no effect when glucose or lactose was used as a competing ligand (data not shown). This may also explain why Con A-agarose is widely used as an affinity matrix for the purification of t-PA and its subsequent elution with methyl a-Dmannoside [11]. The present assay method offers a useful alternative technique over other immunoassays as it has been shown that the sensitivity of this assay is very similar to that of
58
bioimmunoassay and is little affected when non-relevant glycoproteins are present. Such an assay method would be useful under conditions where other immunoassays are not practicable due to non-availability of antibodies to a particular glycoprotein enzyme. This technique would also be of potential value in 'dissecting' a complex mixture of glycoproteins with the appropriate choice of substrate or in discriminating between glycosylated and non-glycosylated forms of a recombinant protein, but it may not be possible to distinguish between different degrees of glycosylation of the same molecule. Acknowledgements
My sincere thanks to Prof. Tony Atkinson for his constant encouragement and helpful discussions on various aspects of the work described in this paper. Thanks are also due to Miss Z. Moola and Mr. R. Ling for technical assistance. References Ranby, M., Norrman, B. and Wallen, P. (1982) Thromb. Res. 27, 743-749.
2 Mahmoud, M. and Gaffney, P.J. (1985) Thromb. Haemostas. 53, 356-359. 3 Korninger, C.. Speiser, W., Wojta, J. and Binder, B.R. (1986) Thromb. Res. 41, 527-535. 4 Schleef, R.R., Wagner, N.V., Sinha, M. and Loskutoff, D.J. (1986) Thromb. Haemostas. 56, 328-332. 5 Holvoet, P., Boes, J. and Collen, D. (1987) Blood 69, 284-289. 6 Dufau, M.L., Tsuruhara, T. and Catt, K.J. (1972) Biochim. BiDphys. Acta 278, 281-292. 7 Cummings, R.D. and Kornfeld, S. (1982) J. BioI. Chern. 257, 11235-11240. 8 Hayman, M., Skehel, J.J. and Crumpton, M.J. (1973) FEBS Lett. 29, 185-188. 9 Gold, E.R. and Balding, P. (1975) Plant and Animal Lectins, Excerpta Medica, Amsterdam. 10 Bog-Hansen, T.C. (ed.) (1981-1986) Lectins: Biology, Biochemistry, Clinical Biochemistry, Vols, 1-5, W. de Gruyter, Berlin. 11 Rijken, D.C. and Collen, D. (1981) J. Bioi. Chern, 256, 7035-7041. 12 Hoylaerts, M., Rijken, R.C., Lijnen, H.R. and Collen, D. (1982) J. BioI. Chern. 257, 2912-2919. 13 Verheijen, J.H., Mullarrt, E., Chang, G.T.G., Kluft, C. and Wijngaards, G. (1982) Thromb. Haemostas, 48, 266-269. 14 Electricwala, A., Ling, R.J., Sutton, P.M., Grirfiths, B.J., Riley, P.A. and Atkinson, T. (1985) Thromb. Haemostas. 53,200-203. 15 Kornfeld, R. and Kornfeld, S. (1985) Annu. Rev. Biochem. 54, 631-664. 16 Opdenakker, G., Van Damme, J., Bosman, F., Billiau, A. and Seer, P.D. (1986) Proc. Soc. Exp. BioI. Med. 182,248-257.