Use of concanavalin a as a topographical probe for protein-protein interaction Application to lactose synthase

90

Biochimica et Biophysica Acta, 745 (1983) 90-96 Elsevier

BBA 31613

U S E OF CONCANAVALIN A AS A TOPOGRAPHICAL PROBE FOR PROTEIN-PROTEIN INTERACTION

APPLICATION TO LACTOSE SYNTHASE SHAN S. WONG, THOMAS E. MALONE and TIMOTHY K. LEE

The Biochemistry Program, Department of Chemistry, University of Lowell, Lowell, MA 08154 (U.S.A.) (Received January 5th, 1983)

Key words: Concanavalin A binding," Galactosyltransferase; Lactose synthase; Protein-protein interaction; Carbohydrate group; (Bovine milk)

Galactosyltransferase (EC 2.4.1.38) has been shown to bind to Con A-Sepharose. Concentrations of methyl-a-mannoside greater than 0.7 M were required to release the enzyme from the immobilized lectin. Molecular weight determination by gel filtration revealed that galactosyitransferase formed a 1 : 1 complex with concanavalin A. Preincubation of the enzyme with excess concanavalin A did not affect its catalytic activity either in the presence or absence of a-lactalbumin. The galactosyltransferase-concanavalin A complex was retained on an a-lactalbumin-Sepbarose column in the presence of N-acetylglucosamine and manganese chloride and was eluted from the column in their absence. Galactosyltransferase immobilized onto a Con A-Sepharose was still active either in the presence or absence of a-lactalbumin. Lactose synthase activity was also observed when rite galactosyltransferase-concanavalin A complex was assayed with et-lactalbumin immobilized on Sepharose. These data indicate that the carbohydrate moiety of galactosyltransferase is involved in neither the catalytic process nor the binding of et-lactalbumin and must be linked to the enzyme at a location where it does not present any steric hindrance on the binding of concanavalin A, either free or immobilized on Sepharose.

Introduction

Lactose synthase (EC 2.4.1.22) from bovine milk consists of two hetero-subunits: galactosyltransferase (EC 2.4.1.38) (A protein) and alactalbumin (B protein) [1]. In the absence of the B protein, galaetosyltransferase catalyzes the transfer of galactose from UDPgalactose to N-acetylglucosamine, either free or as the terminal residue of glycoproteins [2]. Glucose is a poor acceptor with a K m above 1 M. However, in the presence of a-lactalbumin, this binding is lowered to the millimolar range, which allows the synthesis of lactose under physiological conditions [3]. The presence of the B protein also inhibits the transfer of galactose 0167-4838/83/$03.00 © 1983 Elsevier Science Publishers B.V.

to other acceptors [4]. Thus, the interaction between galactosyltransferase and a-lactalbumin is a unique control mechanism in which a-lactalbumin acts as a galactosyl acceptor specifier. Although the role of a-lactalbumin in the kinetic mechanism of lactose synthase has been thoroughly investigated [5-7], little is know about the site of interaction between the two proteins, aLactalbumin with an extensive homology to lysozyme [8] showed no carbohydrate-binding capability [9] and lysozyme did not possess any specifier activity [10]. Based on the X-ray model of lysozyme, O'Keefe et al. [11] suggested that the 'cleft' region of a-lactalbumin was not involved in the the acceptor substrate binding and was located

91 some distance from the acceptor binding site on the transferase. Chemical modification studies on the B protein provided evidence that Trp-118 and His-32 of a-lactalbumin were involved in its interaction with the A protein [12,13]. The only amino acid on the transferase that has been implicated in the interactions between the two proteins is tryptophan [14]. Since galactosyltransferaseis a glycoprotein [15], it is of interest to examine whether its carbohydrate moiety is involved in the interaction with alactalbumin. In this communication we report the use of concanavalin A as a probe for the significance of the carbohydrate moiety in lactose synthase. Concanavalin A is a hemagglutinating protein isolated from jack bean and has been shown to form complexes with many polysaccarides [16,17] and glycoproteins [18]. Recently, the structural requirement of an oligosaccharide or a glycopeptide for the binding to Con A-Sepharose was elucidated [19,20]. Affinity chromatography on a Con A-Sepharose column has also become a useful tool for structural studies of glycopeptides [21,22]. We demonstrate here the binding of concanavalin A to galactosyltransferase and show that such interaction does not interfere with lactose and N-acetyllactosamine synthesis activities of the enzyme.

Experimental procedure

according to Fitzgerald et al. [24]. The assay solution contained 75 mM glycylglycine, pH 8.5, 5 mM MnCI 2, 0.26 mM UDPgalactose, 1.5 mM phosphoenolpyruvate, 0.21 mM NADH, I1 units pyruvate kinase, 10.4 units lactate dehydrogenase and either 15.7 mM N-acetylglucosamine or 19.3 mM glucose and 0.1 mg a-lactalbumin. The reaction was initiated with appropriate amounts of galactosyltransferase. The a-lactalbumin-Sepharose used in the study of the protein-protein interactions was prepared according to the method of Royer et al. [25]. The protein was immobilized onto CL-Sepharose 4B via a cleavable spacer (-NH(CH2)6NHCO(CH2)2SS(CH2)2CO-). In all experiments involving concanavalin A, 50 mM N-ethylmorpholineHC1 buffer, pH 7.5, containing 1 mM MnCI 2, 1 mM CaC12 and 1 mM 2-mercaptoethanol, was used, unless otherwise noted. The molecular weight of the A protein-concanavalin A complex was determined on a column of Bio-Gel P-200 (1.0 x 47 cm), equilibrated with 50 mM N-ethylmorpholine-HC1 buffer (pH 7.5)/1 m M C a C 1 2 / I mM M n C 1 2 / 1 mM 2mercaptoethanol, at a flow rate of 3 ml per h. Galactosyltransferase was incubated with a 100fold molar excess of concanavalin A before application to the column. The molecular weight standards used were: ovalbumin, M r 43000; bovine serum albumin, M r 67000; concanavalin A, M, 110000; and aldolase, M r 158000.

Materials Galactosyltransferase was purified from bovine milk according to the method of Geren et al. [23]. The average specific activity was 8.7 units/mg at 25°C under the assay condition of suboptimal concentration of UDPgalactose, as described below. Homogeneous a-lactalbumin was also purified from the same source according to Brodbeck et al. [1]. Concanavalin A was purchased from Pharmacia, whereas Con A-Sepharose, pyruvate kinase (type 1), UDPgalactose, methyl-a-mannoside and N-acetylglucosamine were products of Sigma Chemical Company. Other chemicals used were of reagent grade.

Methods Galactosyltransferase and lactose synthase activities were followed spectrophotometrically

Results

Galactosyitransferase-ConA-Sepharose complex Since concanavalin A has a relatively broad specificity for monosaccharides [16,17], it is likely that it will bind to the carbohydrate moiety of galactosyltransferase. In order to investigate this possibility, concanavalin A immobilized on Sepharose was used. When a solution of galactosyltransferase was applied to such a column, the elution profile under various saccharide concentrations suggested that there was interaction between the two proteins and that the A protein-Con ASepharose complex was very stable. As shown in Fig. 1, the dissociation of this complex could not be effected by glucose up to 1 M concentration. The transferase, however, was released from the

92

X

a6 10C

F'roction number

o

i

,~

'

~

'

1/UOP-Gal mM"~

Fig. 1. Elution of galactosy|transferase from a Con A-Scpharose column. Galactosyltransferase (0.05 units) was applied

to a column of concanavalinA (0.3 ml) and washed with l0 ml of 50 mM N-ethylmorpholine-HClbuffer, pH 7.5, containing 1 mM MnCi2/1 mM CaCI2/l mM 2-mercaptoethanol. The column was then eluted with a 35 ml finear gradient of 0-1 M glucose followedby an additional 10 ml of 1 M glucose(solid line). The column was finally washed with 1.1 M methyl-amannoside as indicated by the arrow ( ~). The flow rate of the column was maintained at 28 ml per h. The first activity peak was due to overloading of the column.

lectin by 1.1 M methyl-a-mannoside, of which a minimum concentration of 0.7 M was required as found by a linear gradient elution. In another experiment, using N-acetylglucosamine instead of glucose, similar results were observed. We infer from these data that, under the subsequent experimental conditions where the concentrations of glucose and N-acetylglucosamine were only 50 mM or below, the A protein and concanavalin A would remain associated whether in the soluble or immobilized form. Results from latter experiments also support this conclusion.

Stoichiometry of the A protein.concanavalin A complex Because the strong interaction between galactosyltransferase and Con A-Sepharose may be due to the binding of several lectin molecules to a transferase, the stoichiometry of the complex was examined. Concanavalin A is known to form insoluble complexes with various macromolecules, e.g., dextran [16]. However, no precipitation was observed when galactosyltransferase was added to concanavalin A in molar ratios of up to a hundred. Since concanavalin A exists as a tetramer with four binding sites at pH values above 6 [26], this

Fig. 2. Effect of concanavalin A on N-acetyllactosamineand lactose synthesis activities. The reciprocals of the initial rates were plotted against the reciprocals of various UDP-galactose concentrations. The enzymes used were: A, galactosyltransferase; ©, galactosyltransferase preincubated with a #0-fold molar excess concanavalin A; A, lactose synthase; e, lactose synthasewith the A protein preincubated with a #o-fold molar excess of concanavalin A. The assay conditions used were as described in Experimental procedure.

implies that each transferase molecule can only interact with one concanavalin A. This speculation is substantiated by a molecular weight determination of the complex by gel filtration through a column of Biol-Gel P-200. In the presence of excess concanavalin A, the complex was found to be about 160000 daltons, which is the sum of one molecule each of concanavalin A and A protein. A similar molecular weight was obtained in another experiment where 1 M methyl-a-mannoside was included in the elution buffer. Thus, only one concanavalin A molecule is bound per molecule of galactosyltransferase, even in the presence of methyl-a-mannoside.

Effect of concanavalin A on enzyme activities The ability of galactosyltransferase and concanavalin A to form a complex in solution allows us to study the effect of concanavahn A on the N-acetyllactosamine and lactose synthesis activities under the assay conditions where either N-acetylglucosamine or glucose is used, respectively. As shown in Fig. 2, the presence of excess concanavalin A did not affect the enzymatic activities. Since the assay mixtures contained less than 20 mM monosaccharide, the A protein-concanavalin A complex should remain intact. With the stability of the complex even in the presence of

93

VALVE A

3

IMENZYME MO61LIZ--~r6 ED ~ ~'J

:-2

I 00 STIRRER_ 10^

20

FRACTIONNUMBER

Fig. 3. Elution of galactosyltransferase.concanavalin A complex from an a-lactalbumin-Sepharose column. A column of 0.5 ml a-lactalbumin-Sepharose was equilibrated with 50 mM Nethylmorpholine-HCl buffer, pH 8.0, containing 50 mM Nacetylglucosamine/I mM 2-mercaptoethanol/ I mM manganese chloride. A solution of galactosyltransferase with a 40-fold molar excess of concanavalin A was applied and the column was further washed with 5 ml of the same buffer. The column was developed with 50 mM N-ethylmorpholine-HCI buffer, pH. 8.0, containing 5 mM EDTA and 1 mM 2mercaptoethanol, at the arrow (J,). The initial activity peak is due to overloading of the column.

1 M methyl-a-mannoside, it is expected that no dissociation would occur on dilution of the complex into the assay mixture. It is therefore noteworthy that a-lactalbumin is capable of interacting with the soluble A protein-concanavalin A complex to form lactose synthase. This is further supported by the results show in Fig. 3, in which the effect of concanavalin A on the interactions between galactosyltransferase and a-lactalbumin was examined by chromatographing the A protein-concanavalin A complex on an a-lactalbumin-Sepharose column. In the presence of manganese chloride and N-acetylglucosamine, the A protein binds to the B protein immobilized on the gel, a fact employed in the purification of the A protein [27,28]. Under such conditions the A protein-concanavalin A complex was retained on the column until N-acetylglucosamine and manganese chloride were excluded from the elution buffer. Such retention of activity indicates that concanavalin A does not interfere with the interactions between galactosyltransferase and ot-lactalbumin.

Enzymatic activities of the immobilized proteins To characterize further the effect of the binding

SPECTROPHOTOMETRIC"

FLOWCELL

Fig. 4. Diagrammatic representation of a flow system for the assay of enzymatic activities of immobilized proteins. The solution is pumped from the beaker through the flow cell placed in the spectrophotometer into the column. The flow rate is controlied by the pump and the column can be bypassed at valve A.

of concanavalin A to galactosyltransferase, the enzymatic activities of the immobilized proteins were studied under various conditions. We have devised a flow system shown diagrammatically in Fig. 4. A column of immobilized enzyme is equilibrated with the assay solution, which is circulated by a peristaltic pump through a flow cell placed in a spectrophotometer to monitor the decrease of absorbance at 340 nm. The rate of the decrease in absorbance is proportional to the contact time of the assay solution in the column which can be regulated by the pump. The column can be bypassed at Valve A to divert the solution from the column directly into the beaker. Assays for the N-acetyllactosamine and lactose synthesis activities of an A protein-saturated Con A-Sepharose column in the flow system are illustrated by the tracings A and B in Fig. 5, respectively. As an assay mix"ture was pumped through the column at position 1, a decrease in absorbance was noted, showing that the enzyme was active. When the column was bypassed at position 2, the absorbance leveled off, indicating that there was no leakage of the A protein from the column. As soon as the column was re-engaged at position 3,

94

¢ q

0

20

40 TIME

cate that a-lactalbumin is still capable of interacting with the transferase in its complexed form with concanavalin A and that the lectin is not displaced from the complex by the B protein. In another experiment to confirm further that a-lactalbumin can interact with the A protein-concanavalin A complex, a column of a-lactalbumin linked to Sepharose through a spacer was used. The result is shown in tracing C of Fig. 5. When an assay mixture containing the A protein and glucose was directed through the column, the absorbance decreased, indicating the production of lactose. Addition of concanavalin A (position 4) had no effect on the rate of the reaction. From the data presented earlier, it is likely that the A protein-concanavalin A complex exists in the assay solution. Again, when the column was bypassed at position 2, the absorbance leveled off. When the solution was pumped through the column again at position 3, the absorbance decreased at the same rate.

IMINI

Fig. 5. Enzymatic activities of the immobilized proteins. A column of Con A-Sepharose (0.2 ml) saturated with A protein was washed with 50 mM N-ethylmorpholine-HCl/1 mM CaCI 2 / 1 mM MnCI 2 / 1 mM 2-mercaptoethanol before use for tracings A and B. For tracing C, 1.3 ml a-lactalbumin-Sepharose were used. Arrows 1 and 3 indicated when the assay solutions were pumped through the column, and arrow 2, when the column was bypassed. Tracing A: 5 ml assay solution containing N.acctylglucosamine for the galactosyltransferase activity were used. At arrow 4, 0.6 mg a-lactalbumin was added. Tracing B: 5 ml assay solution containing a-lactalbumin and glucose for the lactose synthase activity were used. At arrow 4, 17 mg N-acetylglucosamine was added. Tracing C: 5 ml of the assay solution containing the A protein and glucose were used. At arrow 4, 5 mg of concanavalin A were added. The pulsating.part of the tracing is due to the initial equilibrating of the column.

the absorbance continued to decrease. At position 4 of tracing A, a-lactalbumin was added to the assay solution. Since the rate of decrease in absorbance was retarded, the B protein must have inhibited the N-acetyllactosamine synthesis activity, a fact consistent with the behaviour observed in solution. The same inhibition was seen when N-acetylglucosamine was added to the lactose synthase assay mixture as depicted by tracing B at position 4. The results of these experiments indi-

Discussion Galactosyltransferase has been shown to contain up to 12% of carbohydrate [15]. The elucidation of the carbohydrate composition by Lehman et al. [29,30] suggested the presence of a sialic acid-galactose-N-acetylglucosamine-glycopeptide. While we do not know what is the terminal monosaccharide that binds to concanavalin A, the affinity of galactosyltransferase for Con A-Sepharose is certainly very strong (Fig. 1). Concanavalin A has been shown to bind specifically to a-D-mannopyranosyl, a-o-glucopyranosyl and a-D-fructofuranosyl residues in polysaccharides, as well as to monosaccharides bearing C-3, C-4 and C-6 hydroxyl groups of the D-mannopyranose or Dglucopyranose ring configuration [16-19]. In an experiment where the A protein was eluted from Con A-Sepharose using a linear gradient of methyl-a-mannoside, the enzyme peak was found to be associated with 0.7 M methyl-a-mannoside (data not shown). Therefore, it takes at least that concentration of this monosaccharide to dissociate the A protein-concanavalin A complex. Saturated solutions of either glucose or N-acetylglucosamine were not capable of releasing the A protein from the complex. In most cases where Con A-Sep-

95 harose was used as an affinity column for the isolation of glycopeptides, a concentration of 0.1 M methyl-a-mannoside would suffice for the purpose [20,21,31]. Even in the binding of dextran to soluble concanavalin A, a concentration of 0.1 m M methyl-a-mannoside was sufficient to inhibit 50% of the complex formation [17]. The association between galactosyltransferase and Con ASepharose must therefore involve other interaction forces in addition to the simple sugar-binding specificity of the lectin [32]. The affinity between galactosyltransferase and soluble concanavalin A is even stronger. A concentration of 1.0 M methyla-mannoside was not able to dissociate the complex on gel filtration. The difference between soluble concanavalin A and Con A-Sepharose could be due to conformational changes of concanavalin A on immobilization. The molecular weight determination of the A protein-concanavalin A complex reveals that, with excess lectin, only one mole of concanavalin A is bound per mole of the A protein. According to its carbohydrate content [15], galactosyltransferase has about 30 monosaccharide units, this polysaccharide may exist in several chains or branches [33] which may interact mutually exclusively with one molecule of concanavalin A or may interact simultaneously with several binding sites on a single concanavalin A molecule. Kinetics studies of the A protein-concanavalin A complex in solution reveal that the binding of galactosyltransferase to concanavalin A does not affect its activity in either the presence or absence of a-lactalbumin (Fig. 2). Thus, the carbohydrate moiety is not involved in the catalytic process of the enzyme. Geren et al. [30] have also shown that the enzymatic activity of galactosyltransferase is retained after partial removal of the oligosaccharide. Considering the molecular weight of concanavalin A, which is twice the size of galactosyltransferase, we believe that the carbohydrate chain on the A protein must be quite far removed from the active site, so that the binding of concanavalin A does not interfere with the catalytic process. Although no attempt was made to quantitate the enzymatic activity of the immobilized galactosyltransferase on Con A-Sepharose because of the complication due to diffusion limitations in the matrix of a solid support [34], the

retention of its catalytic activity is significant. This is particularly so when a-lactalbumin is present. The lactose synthase activity is not affected by the binding of concanavalin A, either in solution (Fig. 2) or on a solid support (Fig. 5). The ability of the B protein to interact with the galactosyltransferase-concanavalin A complex is further supported by the results of the affinity chromatography using immobilized a-lactalbumin, as shown in Fig. 3. In summary, these data indicate that (a) the carbohydrate moiety of the A protein is not involved in the interaction between the A and B proteins of lactose synthase, and (b) that the carbohydrate chain is removed from the interaction site. These conclusions are consistent with the observation that the B protein does not possess any carbohydrate-binding capability [9]. While the data presented here do not allow us to specifically locate the carbohydrate moiety on the A protein, it must be sufficiently far away from both the active site and the a-lactalbumin interaction site so that the binding of concanavalin A, either free or immobilized on a solid support, does not present any steric hindrance to the activities of the enzyme. Although O'Keefe et al. [11] have proposed that a-lactalbumin does not interact at the acceptor substrate binding site on galactosyltransferase, it is possible that the a-lactalbumin binding site and the catalytic site on the A protein may partially overlap. Acknowledgement

This work was supported by United States Public Health Service Grant GM26623. References

1 Brodbeck, U., Denton, W.L., Tanahashi, N. and Ebner, K.E. (1967) J. Biol. Chem. 242, 1391-1397 2 Brew, K., Vanaman, T.C. and Hill, R.L. (1968) Proc. Natl. Acad. Sci. U.S.A. 59, 491-497 3 Fitzgerald, D.K., Brodbeck, U., Kiyosawa, I., Mawai, R.,

Colvin, B. and Ebner, K.E. (1970) J. Biol. Chem. 245, 2103-2108 4 Schanbacher, F.L. and Ebner, K.E. (1970) J. Biol. Chem. 245, 5057-5066 5 Bell, J.E., Beyer, T.A. and Hill R.L. (1976) J. Biol. Chem. 251, 3003-3013 6 Morrison, J.R. and Ebner, K.E. (1971) J. Biol. Chem. 246, 3992-3998

96 7 Khatra, B.S., Herries, D.G. and Brew, K. (1974) Eur. J. Biochem. 44, 537-560 8 Brew, K., Vanaman, T.C. and Hill, R.L (1967) J. Biol. Chem. 242, 3747-3749 9 Burkhandt, A.E., Russo, S.O., Rinehardt, C.G. and Loudon, G.M. (1975) Biochemistry 14, 5465-5469 10 Browne, W.J., North, A.C.T., Phifips, D.C., Brew, K., Vanaman, T.C. and Hill, R.L. (1969) J. MoL Biol. 42, 65-86 11 O'Keefe, E.T., Mordick, T. and Bell, J.E. (1980) Biochemistry 19, 4962-4966 12 Bell, J.E., CasteUino, F.J., Trayer, I.P. and Hill, R. L. (1975) J. Biol. Chem. 250, 7579-7585 13 Prieels, J.P., Bell, J.E., Schindler, M., Castellino, J. J. and Hill R.L. (1979) Biochemistry 18, 1771-1776 14 Takashi, K. and Ebner, K.E. (1981) J. Biol. Chem. 256, 7269-7276 15 Trayer, I.P. and Hill, R.L. (1971) J. Biol. Chem. 246, 6666-6675 16 Goldstein, l.J., Hollerman, C.E. and Merrick, J.M. (1965) Biochim. Biophys. Acta 97, 68-76 17 Goldstein, I.J., Hollerman, C.E. and Smith, E.E. (1965) Biochemistry 4, 876-883 18 Clarke, A.E. and Denborough, M.A. (1971) Biochem. J. 121,811-816 19 Ogata, S., Muramatsu, T. and Kobata, A. (1975) J. Biochem. (Tokyo) 78, 687-696 20 Krusius, T., Finne, J. and Rauvala, H. (1976) FEBS Lett. 71, 117-120

21 Hodges, L.C., Laine, R. and Chart, S.K. (1979) J. Biol. Chem. 254, 8202-8212 22 Yamamoto, K., Tsuji, T., Irimura, T. and Osawa, T. (1981) Biochem. J. 195, 701-713 23 GEren, C.R., Magee, S.C. and Ebner, K.E. (1976) Arch. Biochem. Biophys. 172, 149-155 24 Fitzgerald, D.K., Colvin, B., Mawal, R. and Ebner, K.E. (1970) Anal. Biochem. 36, 43-61 25 Royer, G.P., Ikeda, S. and Aso, K. (1977) FEBS Lett. 80, 89-94 26 Senear, D.F. and Teller, D.C. (1981) Biochemistry 20, 3076-3083 27 Barker, R., Olsen, K.W., Shaper, J.H. and Hill, R.L. (1972) J. Biol. Chem. 247, 7135-7147 28 Andrews. P. (1970) FEBS Lett. 9, 297-300 29 Lehman, D.E., Hudson, B.G. and Ebner, K.E. (1975) FEBS Lett. 54. 65-69 30 Geren, C.R., Geren. L.M., Lee, D. Ebner, K.E. (1977) Biochim. Biophys. Acta 497, 128-132 31 Marquis, J.K., Hilt, D.C. and Mautner, H.G. (1980) J. Neurochem. 34, 1071-1076 32 Ohoa, J-L. (1981) J. Chromatogr. 215, 351-360 33 Beyer, T.A., Sadler, J.E., Rearick, J.I., Paulson, J.C. and Hill, R.L (1981) Adv. Enzymol. 52, 23-175 34 Katchalski, E., Silman, I. and Goldman, R. (1971) Adv. Enzymol. 34, 445-536