- Email: [email protected]
Physicochemical characterization of Cajanus cajan lectin: effect of pH and metal ions on lectin carbohydrate interaction
Biochimica et Biophysica Acta 1427 (1999) 378^384 www.elsevier.com/locate/bba
Physicochemical characterization of Cajanus cajan lectin: e¡ect of pH and metal ions on lectin carbohydrate interaction Shama Ahmad a
a;
*, Rizwan Hasan Khan b , Aftab Ahmad
1;c
Department of Biochemistry, Faculty of Life Sciences, Aligarh Muslim University, Aligarh, India b Interdisciplinary Biotechnology Unit, Aligarh Muslim University, Aligarh 202002, India c Center for Biotechnology, Hamdard University, New Delhi, India Received 20 January 1999; received in revised form 1 March 1999; accepted 2 March 1999
Abstract The association constant of Cajanus cajan lectin for methyl K-D-mannopyranoside was studied by equilibrium dialysis method. An attempt was also made to understand the metal ion requirements and to establish that ionizable groups are responsible for lectin^carbohydrate interaction. The N-terminal sequence up to 27 amino acid residues was found to be more than 80% homologous with other mannose-specific legume lectins of the tribe Viceae. Like concanavalin A and pea lectin it also exhibits high affinity for the sugar K-methyl mannose and at 37³C the association constant was found to be 1.4U104 M31 . The lectin required one Ca2 and one Mg2 per mole and during the lectin sugar interaction two ionizable groups with pK of 3.75 and 8.3 are ionized. Whether the secondary structure is similarly affected with pH changes and presence or absence of metal ion was investigated by circular dichroism studies. Results suggested that changes in carbohydrate binding properties of the Cajanus cajan lectin due to change in pH and addition of metal ions are not accompanied by any significant change in secondary structure. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Legume lectin; Cajanus cajan; Lectin^sugar interaction; Metal ion; pH e¡ect; N-Terminal sequence
1. Introduction Although biological interactions such as protein^ Abbreviations: BSA, bovine serum albumin; CCL, Cajanus cajan lectin; Con A, concanavalin A; Tris, tris (hydroxy methyl) amino methane * Corresponding author. Present address: Department of Pediatrics, National Jewish Medical and Research Center, D201 Neustadt Building, 1400 Jackson Street, Denver, CO 80206, USA; E-mail: [email protected] 1 Also corresponding author. Present address: Department of Pediatrics, National Jewish Medical and Research Center, D201 Neustadt Building, 1400 Jackson Street, Denver, CO 80206, USA.
protein, protein^nucleic acid and nucleic acid^nucleic acid are well understood, the interaction between protein and carbohydrate is an aspect which needs to be looked at in further detail. E¡orts in this regard have been made by a number of workers [1^3]. Lectins which have a high degree of carbohydrate speci¢city make a useful model for protein carbohydrate interactions [4^9]. They have been implicated in cell to cell interaction, ligand^receptor recognition, blood group typing, transport of biological macromolecules and, to some extent, in the immuno recognition process [10^13]. Plant lectins have also been demonstrated to be determinants for the symbiosis of Rhizobia with legumes [14,15]. In this context, they were recently shown to interact with bacterial cell
0304-4165 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 9 9 ) 0 0 0 3 5 - 5
BBAGEN 24794 11-5-99
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
wall components such as muramic acid, N-acetyl muramic acid and muramyl dipeptide [16,17]. Although these reports are convincing, it is important that we should not forget that lectins are present in bacteria too. The function of lectin as a receptor for sugar recognition seems identical to recognition between enzyme and substrate and antibody and antigen. Moreover, the three dimensional structure of the carbohydrate binding site has similarities with the three-dimensional structure of the antigen binding site of an antibody [5]. Thus studies on lectin carbohydrate binding may provide an even better understanding of antibody^antigen binding [18]. We have therefore made an attempt to study the carbohydrate binding, metal ion requirements and pH-dependent changes in a hitherto less investigated mannose/glucose speci¢c seed lectin of pigeon pea (Cajanus cajan) [19]. 2. Materials and methods 2.1. Materials Cajanus cajan lectin was prepared from seed as described previously [19]. The lectin is a glycoprotein with two similar subunits of molecular weight 38 000. Methyl K-D-mannopyranoside was a product of Sigma, St. Louis, MO, USA. The bu¡er components and salts of metal ions were crystallized thrice before use. All other reagents were of analytical grade and were used without further puri¢cation. 2.2. Methods Protein concentration was determined routinely by the method of Lowry et al. [20], using crystalline BSA as a standard and occasionally by measuring absorbance at 280 nm using the measured value of the speci¢c extinction coe¤cient of the lectin [19]. The neutral carbohydrate concentration was estimated by the method essentially due to Dubois et al. [21] using methyl K-mannopyranoside as a standard. Accordingly, to 1.0 ml of carbohydrate solution (0.1^10.0 Wg) in 10 mM Tris-HCl bu¡er pH 7.0 containing 0.5 M NaCl was added 0.1 ml of 5% phenol. After proper mixing, 0.5 ml of concentrated H2 SO4 was added. The color developed was read against an
379
appropriate blank at 490 nm. The determination of N-terminal sequence of the lectin was done in the laboratory of Professor I.J. Goldstein, The University of Michigan, Ann Arbor, MI. 2.2.1. Measurement of lectin activity The hemagglutinating activity of the Cajanus cajan lectin against rabbit erythrocytes and/or goat IgM precipitating activity was determined as described before [19]. For the determination of e¡ect of metal ions on the carbohydrate binding properties of the lectin, it was ¢rst demetallized by dialyzing against 1 M acetic acid at 4³C followed by overnight dialysis against 10 mM EDTA. 2.2.2. Equilibrium dialysis The binding of methyl K-D-mannopyranoside to Cajanus cajan lectin was quantitatively studied in 20 mM Tris-HCl bu¡er pH 7.0 plus 0.5 M NaCl by equilibrium dialysis in a dialysis bag (capacity 3.0 ml) made from Sigma cellulose membrane (in£ated diameter 6 mm, width 10 mm). The dialysis bag containing 1.0 ml of the lectin solution was placed in plastic vials containing 1.0 ml of Me-mannose solution. After capping the vial, the dialysis bag was mechanically shaken for 24 h at a given temperature (16 and 37³C). After attainment of equilibrium, the decrease in Me-mannose concentration was estimated in the dialysate, outside the dialysis bag by the method of Dubois et al. [21]. The amount of Memannose bound per mole of lectin, was calculated. It should be noted that no detectable binding of Memannose by the dialysis bag was observed. Furthermore, at a salt concentration as high as 0.5 M NaCl used in these studies, the Donnan e¡ect would be insigni¢cant. Cajanus cajan lectin is a dimer. It would readily react with the speci¢c carbohydrate, according to the following reaction. L nS ! L Sn
1
L Sn L Sn
2
K ass !
e ! n K ass S=1 K ass S
3
where n is the number of carbohydrate binding sites in the lectin dimer and Kass is the association con-
BBAGEN 24794 11-5-99
380
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
stant for the interaction of S (or Me-mannose) with L (or lectin). Eq. 3 can be rearranged to Eq. 4. e=S ! 3K ass e nK ass
4
The amount of Me-mannose bound per mole of lectin i.e., e, was determined in triplicate and plotted according to Eq. 4. 2.2.3. CD measurements Circular dichroism measurements were carried out at 25³C on a Jasco spectropolarimeter model J-720 using a Sekonic X^Y plotter (model SPL-430), NESLAB water bath model RTE 110 and quartz cell of path length 0.1 or 1.0 cm. Other experimental conditions were, sensitivity 5 mdeg, response time 2 s, scan speed 20 nm/min and a resolution of 0.1 nm. The mean residue ellipticity [a] of a protein was calculated using the expression a a= 10 nUC p Ul Where a is the observed ellipticity in mdeg, n the number of amino acid residues in the protein whose molar concentration is Cp and l is the path length in cm. 3. Results The N-terminal sequence up to 27 amino acid residues of CCL was established. The N-terminal residue of the lectin was threonine as reported before [19]. An attempt was also made to establish homology of the N-terminal sequence with the N-terminal
Fig. 1. N-Terminal amino acid sequence of legume lectins showing homologies within the Viceae tribe and Con A. Here CCL represents the N-terminal amino acid sequence of Cajanus cajan lectin, LCL (Lens culinaris lectin), LOL (Lathyrus ochrus lectin) and PSL (Pisum sativum lectin). The top numbering is for FBL (fava bean lectin) [22].
sequence of few other mannose/glucose speci¢c legume lectins of the same tribe (Fig. 1) [22]. 3.1. Interaction of Me-K-mannose to CCL The Scatchard plot of the equilibrium dialysis data of CCL in 20 mM Tris HCl bu¡er containing 0.5 M NaCl (pH 7.0) with Me-K-mannose at 37³C is shown in Fig. 2. A linear plot was obtained, the slope of which gave the binding constant Kass . A similar plot was obtained at 16³C (¢gure not shown for clarity). The results are summarized in Table 1. 3.2. Metal ion requirements A ¢xed concentration of the demetallized lectin was incubated with increasing concentration of metal ions in 10 mM Tris HCl bu¡er pH 7.0 and the lectin
Table 1 Binding data for lectin sugar interaction Lectin
Sugar
Kass (T³C)
No. of binding sites
Reference
CCL
Me-K-D-mannopyranoside
1.61U104 M31 (16) 1.40U104 M31 (37)
2 2
Present study Present study
Con A
Me-K-D-mannopyranoside
2.06U104 1.65U104 0.83U104 1.70U104 1.20U104 1.25U104
(2) (8) (27) (8) (20) (37)
2 2 2 2 2 2
[22] [22] [22] [23] [23] present study
Pea
Me-K-D-mannopyranoside
0.33U104 M31 (9) 0.19U104 M31 (19)
2 2
[23] [23]
M31 M31 M31 M31 M31 M31
BBAGEN 24794 11-5-99
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
Fig. 2. Scatchard plot for the binding of methyl K-D-mannopyranoside to CCL. A ¢xed concentration of CCL (37 WM in dimer) in 20 mM Tris-HCl bu¡er pH 7.0 containing 0.5 M NaCl in a dialysis bag was incubated with 4^400 WM methyl KD-mannopyranoside in the same bu¡er at 37³C for 24 h. After the attainment of equilibrium the decrease in Me-mannose concentration was estimated in the dialyzate. Analysis of results according to Eq. 4 given in text yielded the value of association constant Kass and the number of sugar binding sites, X. Each circle (b) represents the mean of three independent experiments performed in triplicate.
activity was measured. The six metal salts were CaCl2 , MnCl2 , ZnCl2 , CoCl2 , CrCl3 and MgSO4 . The demetallized lectin showed barely detectable activity in the presence of increasing concentration of any of the six metal salts. However, when the incubation mixture contained ¢xed concentration of the lectin (2.15 mM), IgM (0.45 WM), CaCl2 (4 WM) and increasing concentration of Mn2 , a steep rise in glycoprotein-precipitating activity occurred at a in£ection point ratio [Mn2 ]/[lectin] of 2.0 (Fig. 3). When increasing concentration of Ca2 were added to the mixture of Cajanus cajan lectin (2.15 WM) and goat IgM (0.45 WM) containing ¢xed amount of MnCl2 (4 WM) similar in£ection point was observed. Further analysis of the results showed that each subunit of the lectin bound one Ca2 and one Mn2 . 3.3. E¡ect of pH The e¡ect of pH on the interaction of CCL with rabbit erythrocytes, dextran and goat IgM was investigated in three bu¡er solution in the pH range of 2.0^10.5. The three bu¡er solutions were of 10 mM
381
concentration and contained 1 mM each of Ca2 and Mn2 . These were sodium citrate bu¡er (pH 2.0^6.0) sodium phosphate bu¡er (pH 7.0^9.0) and Tris-HCl (pH 9.5^10.5). In a total volume of 2.0 ml of a given bu¡er were taken 8.23 WM of the lectin with 1.74 WM goat IgM (at this molar ratio the precipitation of goat IgM with CCL was maximum) and the mixture was incubated overnight at 25³C. The amount of protein precipitated at each pH value was determined in triplicate. Both the dextran and glycoprotein precipitation by the seed lectin was measurable in the pH range of 2.0^10.5. In both these cases, the amount of protein precipitated increased steeply with increase in pH from pH 2.0 to 6.0. Further increase in pH of the bu¡er beyond pH 6.0 caused marked reduction in the precipitation. It should be noted that the speci¢c bu¡er ion e¡ect was negligible here, as similar lectin activity values were obtained in two bu¡er solutions of identical pH values (e.g. bu¡er solution, each of pH 6.0 were prepared with sodium citrate as well as sodium phosphate and pH 9.0 with sodium phosphate and Tris-HCl). The bellshaped pH pro¢le can be described in terms of the two types of ionizable groups. M M =1 H K 1 K 2 =H where M* represents the maximal attainable value of
Fig. 3. E¡ect of metal ions on the IgM precipitating activity of CCL. Curve (b) representing the glycoprotein precipitating activity was obtained when lectin (2.15 WM) was incubated with IgM (0.45 WM) and CaCl2 (4 WM) and increasing concentration of MnCl2 . Similar curves with increasing concentration of either Ca2 (a), Mn2 (R) or Mg2 (O) alone are shown.
BBAGEN 24794 11-5-99
382
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
at pH 6.0 and 9.0 were similar with peak at 196^198 nm and trough at 220^225 nm. At pH 2.0, the trough shifted to lower wavelength, i.e. 215 nm (Fig. 5). CD spectra of demetallized, metallized and metallized along with sugar methyl K-D-mannopyranoside were also studied in the far UV region at pH 6.0 taking 10 mM sodium citrate bu¡er (Fig. 6). Interestingly, the spectral features remain the same in all the three conditions. 4. Discussion Fig. 4. E¡ect of pH on the interaction of CCL with goat IgM (b) and dextran (R). The solid theoretical curve was drawn with appropriate pK1 and pK2 values. The ¢lled circles and triangles represent the experimental values of M (protein precipitated) at di¡erent pH values. For other details see text.
Cajanus cajan, a major pulse crop of India belongs to the subfamily Papilionoidea and tribe Viceae. It
M (protein precipitated), K1 and K2 are the dissociation constants of the two ionizable groups of the lectin. The maximum amount (M*) of the glycoprotein lectin which was precipitated by the dextran at pH 6.0 was determined to be 0.547 mg/ml. Interestingly, the pH dependence of both the precipitin reaction viz. lectin dextran and lectin IgM interactions were found to be similar. The theoretical curve drawn with pK1 and pK2 of 3.75 and 8.3, respectively, was found to account for the observed pH dependence of the two precipitin reactions [23] (Fig. 4). Further analysis of the results on pH dependence of lectin activity could be performed by plotting log M against pH. The initial values of the slope of the straight line between log M and pH can be computed both for the `acid' and `basic' arms of the pH pro¢le given in Fig. 4. A rough value for the slope, i.e. dlog M/dpH was estimated to be not more than one in both the curves. 3.4. CD measurements CD measurements of the Cajanus cajan lectin was carried out at three di¡erent pH values. The lectin solution was dialyzed against 10 mM HCl+10 mM KClO4 (pH 2.0), 10 mM sodium phosphate bu¡er, pH 6.0 and 10 mM sodium borate bu¡er, pH 9.0. The solutions were ¢ltered through a Millipore ¢lter (0.45 Wm) and ¢nally deaerated. The spectral features
Fig. 5. E¡ect of pH on the circular dichroic spectra of CCL. Circular dichroic spectra of CCL was measured in 10 mM HCl+10 mM KClO4 at pH 2.0 (^ c ^), 10 mM sodium phosphate bu¡er pH 6.0 (- - -) and 10 mM sodium borate bu¡er at pH 9.0 (999) in the wavelength region 190 to 250 nm. For experimental details see text.
BBAGEN 24794 11-5-99
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
Fig. 6. Circular dichroic spectra of demetallized (- - -), metallized (^ c ^) and metallized plus sugar bound CCL (999). For experimental details see text.
contains 3% Con A reactive carbohydrate and the protein is predominantly acidic in nature [24]. As expected, the 27 amino acid residues at the N-terminal of the lectin molecule exhibit more than 80% homology with the other lectins of the same tribe (Fig. 1) [20]. The lectin has a similar amino acid content per 100 residues to that of the pea lectin and Con A [19,22]. Analysis of the carbohydrate binding properties of the lectin suggested that the lectin is glucose/mannose-speci¢c having greater a¤nity towards the Kmethyl substituted sugars [19]. Results on equilibrium dialysis showed that the lectin binds methyl K-mannopyranoside with an association constant of 1.4U104 M31 at 37³C. This is comparable to the a¤nity constant of Con A [25,26] for the same sugar, but is more than that of pea lectin [26] (Table 1). The lectin seems to have one binding site per subunit for carbohydrate residues. The transition metal ion requirement of CCL was similar to Con A [27] and other legume lectins as it required one previously bound metal ion Ca2 or Mn2 and another in the incubation mixture (Ca2 and Mn2 ) for the lectin to be fully active. These results suggested that the simultaneous presence of both Ca2 and Mn2 is neces-
383
sary and that the lectin has one Ca2 and one Mn2 binding site per subunit in addition to the carbohydrate binding site. Other transition metal ions tested had little or no e¡ect on the binding properties of the lectin. Available data on the pH dependence of the lectin-mediated dextran and IgM precipitation (Fig. 4) suggest the involvement of two ionizable groups with pK values near 3.75 and 8.3. The fact that the initial slope of the `acid' and `basic' arms of the two curves did not exceed 1.0 shows that the number of each type of ionizable groups would be one. Further comparison of the pK values with those found for ionizable groups in proteins [28] indicate that most likely the two groups are aspartic acid and K-amino group of the N-terminal amino acid. Thus, the bellshaped pH pro¢le can be explained in terms of the deprotonation of an aspartic acid residue, which is one of the conserved residue at the active site of most legume lectins [9], and a charged amino group of the N-terminal threonine [19] lying near the active site. The marked change in the lectin activity produced by changing the pH from 3.0 to 9.0 cannot be attributed to denaturation of the lectin molecule since the CD spectra (Fig. 5) of the lectin at pH 6.0 and 9.0 were found to be similar except at pH 2.0. However, the results con¢rm that the decrease in lectin activity with change in pH from 6.0 to 9.0 is not due to changes in the secondary structure of the protein, but a consequence of ionization of ionizable groups present at or near the carbohydrate binding site. However, in the lower pH range, slight changes in protein conformation along with the titration of functional group near the carbohydrate binding site could be attributed to be the cause of decreased lectin activity. The CD spectra in the presence of metal ion and metal ion plus sugar (Me-K-mannopyranoside) suggest that ligand binding in presence of metal ion is accompanied by undetectable changes in the secondary structure of the lectin. Binding of sugar to the combining site of the lectin does not seem to accompany any major structural rearrangement. This argument is consistent with the high resolution X-ray crystallographic ¢ndings of several legume lectins in complex with their speci¢c saccharide [9]. Thus, the sugar combining sites of legume lectins are quite similar and partly superimposable. The different topographies of the combining site for binding di¡erent sugars arise from the ring of hypervariable
BBAGEN 24794 11-5-99
384
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
amino acid residues found around the perimeter of the carbohydrate binding site [8,29]. The metal binding site is, however, located close to the carbohydrate binding site and the amino acid residues, which are involved in the hydrogen bond formation with sugar also coordinate the metal ions. Acknowledgements This paper is dedicated in the memory of the late Prof. A. Salahuddin. The authors are grateful to Prof. I.J. Goldstein, The University of Michigan, Ann Arbor, MI for the N-terminal sequence determination.
References [1] Y.C. Lee, FASEB J. 6 (1992) 3193^3200. [2] D. Solis, P. Fernandez, T. Diaz Murino, J. Jimenez Barbero, M. Martin Lomas, Eur. J. Biochem. 214 (1993) 677^685. [3] E.J. Toone, Curr. Opin. Struct. Biol. 4 (1994) 719^728. [4] R. Loris, J. Hamelryck, J. Bouckaert, L. Wyns, Biochim. Biophys. Acta 1383 (1998) 9^36. [5] I.J. Goldstein, C.E. Hayes, Adv. Carbohydr. Chem. Biochem. 35 (1978) 127^334. [6] N.M. Young, R.P. Oomen, J. Mol. Biol. 228 (1992) 924^934. [7] N. Sharon, Trends Biol. Sci. 18 (1993) 221^226. [8] R. Banerjee, K. Das, R. Ravishanker, K. Suguna, A. Surolia, M. Vijayan, J. Mol. Biol. 259 (1996) 281^296. [9] V. Sharma, A. Surolia, J. Mol. Biol. 267 (1997) 433^445. [10] N. Sharon, H. Lis, Essays Biochem. 30 (1995) 59^75.
[11] I.E. Liener, N. Sharon, I.J. Goldstein (Eds.), The Lectins : Properties, Functions and Applications in Biology and Medicine, Academic Press, London, 1996. [12] L.A. Lasky, Science 258 (1992) 964^969. [13] K. Sastry, R.A. Ezekowitz, Curr. Opin. Immunol. 5 (1993) 59^66. [14] N.J. Brewin, I.V. Kardailsky, Trends Plant Sci. 2 (1997) 92^ 99. [15] I.M. Skvortsov, V.V. Ignatov, FEMS Microbiol. Lett. 165 (1998) 223^229. [16] A. Ayouba, D. Martin, P. Rouge, FEMS Microbiol. Lett. 92 (1992) 41^46. [17] Y. Bourne, A. Ayouba, P. Rouge, C. Cambillau, J. Biol. Chem. 269 (1994) 9429^9435. [18] D. Gupta, H. Kattner, X. Dong, H.J. Gabius, C.F. Brewer, Glycobiology 6 (1996) 843^851. [19] S. Siddiqui, S. Hasan, A. Salahuddin, Arch. Biochem. Biophys. 319 (1995) 426^431. [20] O.H. Lowry, N.J.M. Rosebrough, A.L. Farr, R.J. Randall, J. Biol. Chem. 193 (1951) 265^275. [21] M. Dubois, K.A. Gilles, J.K. Hamilton, P.A. Rebers, F. Smith, Anal. Chem. 28 (1956) 350^356. [22] N. Sharon, H. Lis, FASEB J. 4 (1990) 3198^3208. [23] A. Ahmad, S. Shah, S. Ahmad, Med. Sci. Res. 24 (1996) 107^109. [24] A. Salahuddin, R.H. Khan, J. Protein Chem. 17 (1998) 181^ 185. [25] W. Bessler, J.A. Shafer, I.J. Goldstein, J. Biol. Chem. 249 (1974) 2819^2822. [26] F.P. Schwarz, K.D. Puri, R.G. Bhat, A. Surolia, J. Biol. Chem. 268 (1993) 7668^7677. [27] J.W. Becker, G.N. Reeke, J.L. Wang, B.A. Cunningham, G.M. Edelman, J. Biol. Chem. 250 (1974) 1513^1524. [28] T.E. Creighton, Proteins, Freeman, New York, 1993, pp. 1^ 47. [29] K. Yamamoto, Y. Konami, T. Osawa, T. Kimura, J. Biochem. 111 (1992) 89^95.
BBAGEN 24794 11-5-99
Physicochemical characterization of Cajanus cajan lectin: e¡ect of pH and metal ions on lectin carbohydrate interaction Shama Ahmad a
a;
*, Rizwan Hasan Khan b , Aftab Ahmad
1;c
Department of Biochemistry, Faculty of Life Sciences, Aligarh Muslim University, Aligarh, India b Interdisciplinary Biotechnology Unit, Aligarh Muslim University, Aligarh 202002, India c Center for Biotechnology, Hamdard University, New Delhi, India Received 20 January 1999; received in revised form 1 March 1999; accepted 2 March 1999
Abstract The association constant of Cajanus cajan lectin for methyl K-D-mannopyranoside was studied by equilibrium dialysis method. An attempt was also made to understand the metal ion requirements and to establish that ionizable groups are responsible for lectin^carbohydrate interaction. The N-terminal sequence up to 27 amino acid residues was found to be more than 80% homologous with other mannose-specific legume lectins of the tribe Viceae. Like concanavalin A and pea lectin it also exhibits high affinity for the sugar K-methyl mannose and at 37³C the association constant was found to be 1.4U104 M31 . The lectin required one Ca2 and one Mg2 per mole and during the lectin sugar interaction two ionizable groups with pK of 3.75 and 8.3 are ionized. Whether the secondary structure is similarly affected with pH changes and presence or absence of metal ion was investigated by circular dichroism studies. Results suggested that changes in carbohydrate binding properties of the Cajanus cajan lectin due to change in pH and addition of metal ions are not accompanied by any significant change in secondary structure. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Legume lectin; Cajanus cajan; Lectin^sugar interaction; Metal ion; pH e¡ect; N-Terminal sequence
1. Introduction Although biological interactions such as protein^ Abbreviations: BSA, bovine serum albumin; CCL, Cajanus cajan lectin; Con A, concanavalin A; Tris, tris (hydroxy methyl) amino methane * Corresponding author. Present address: Department of Pediatrics, National Jewish Medical and Research Center, D201 Neustadt Building, 1400 Jackson Street, Denver, CO 80206, USA; E-mail: [email protected] 1 Also corresponding author. Present address: Department of Pediatrics, National Jewish Medical and Research Center, D201 Neustadt Building, 1400 Jackson Street, Denver, CO 80206, USA.
protein, protein^nucleic acid and nucleic acid^nucleic acid are well understood, the interaction between protein and carbohydrate is an aspect which needs to be looked at in further detail. E¡orts in this regard have been made by a number of workers [1^3]. Lectins which have a high degree of carbohydrate speci¢city make a useful model for protein carbohydrate interactions [4^9]. They have been implicated in cell to cell interaction, ligand^receptor recognition, blood group typing, transport of biological macromolecules and, to some extent, in the immuno recognition process [10^13]. Plant lectins have also been demonstrated to be determinants for the symbiosis of Rhizobia with legumes [14,15]. In this context, they were recently shown to interact with bacterial cell
0304-4165 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 9 9 ) 0 0 0 3 5 - 5
BBAGEN 24794 11-5-99
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
wall components such as muramic acid, N-acetyl muramic acid and muramyl dipeptide [16,17]. Although these reports are convincing, it is important that we should not forget that lectins are present in bacteria too. The function of lectin as a receptor for sugar recognition seems identical to recognition between enzyme and substrate and antibody and antigen. Moreover, the three dimensional structure of the carbohydrate binding site has similarities with the three-dimensional structure of the antigen binding site of an antibody [5]. Thus studies on lectin carbohydrate binding may provide an even better understanding of antibody^antigen binding [18]. We have therefore made an attempt to study the carbohydrate binding, metal ion requirements and pH-dependent changes in a hitherto less investigated mannose/glucose speci¢c seed lectin of pigeon pea (Cajanus cajan) [19]. 2. Materials and methods 2.1. Materials Cajanus cajan lectin was prepared from seed as described previously [19]. The lectin is a glycoprotein with two similar subunits of molecular weight 38 000. Methyl K-D-mannopyranoside was a product of Sigma, St. Louis, MO, USA. The bu¡er components and salts of metal ions were crystallized thrice before use. All other reagents were of analytical grade and were used without further puri¢cation. 2.2. Methods Protein concentration was determined routinely by the method of Lowry et al. [20], using crystalline BSA as a standard and occasionally by measuring absorbance at 280 nm using the measured value of the speci¢c extinction coe¤cient of the lectin [19]. The neutral carbohydrate concentration was estimated by the method essentially due to Dubois et al. [21] using methyl K-mannopyranoside as a standard. Accordingly, to 1.0 ml of carbohydrate solution (0.1^10.0 Wg) in 10 mM Tris-HCl bu¡er pH 7.0 containing 0.5 M NaCl was added 0.1 ml of 5% phenol. After proper mixing, 0.5 ml of concentrated H2 SO4 was added. The color developed was read against an
379
appropriate blank at 490 nm. The determination of N-terminal sequence of the lectin was done in the laboratory of Professor I.J. Goldstein, The University of Michigan, Ann Arbor, MI. 2.2.1. Measurement of lectin activity The hemagglutinating activity of the Cajanus cajan lectin against rabbit erythrocytes and/or goat IgM precipitating activity was determined as described before [19]. For the determination of e¡ect of metal ions on the carbohydrate binding properties of the lectin, it was ¢rst demetallized by dialyzing against 1 M acetic acid at 4³C followed by overnight dialysis against 10 mM EDTA. 2.2.2. Equilibrium dialysis The binding of methyl K-D-mannopyranoside to Cajanus cajan lectin was quantitatively studied in 20 mM Tris-HCl bu¡er pH 7.0 plus 0.5 M NaCl by equilibrium dialysis in a dialysis bag (capacity 3.0 ml) made from Sigma cellulose membrane (in£ated diameter 6 mm, width 10 mm). The dialysis bag containing 1.0 ml of the lectin solution was placed in plastic vials containing 1.0 ml of Me-mannose solution. After capping the vial, the dialysis bag was mechanically shaken for 24 h at a given temperature (16 and 37³C). After attainment of equilibrium, the decrease in Me-mannose concentration was estimated in the dialysate, outside the dialysis bag by the method of Dubois et al. [21]. The amount of Memannose bound per mole of lectin, was calculated. It should be noted that no detectable binding of Memannose by the dialysis bag was observed. Furthermore, at a salt concentration as high as 0.5 M NaCl used in these studies, the Donnan e¡ect would be insigni¢cant. Cajanus cajan lectin is a dimer. It would readily react with the speci¢c carbohydrate, according to the following reaction. L nS ! L Sn
1
L Sn L Sn
2
K ass !
e ! n K ass S=1 K ass S
3
where n is the number of carbohydrate binding sites in the lectin dimer and Kass is the association con-
BBAGEN 24794 11-5-99
380
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
stant for the interaction of S (or Me-mannose) with L (or lectin). Eq. 3 can be rearranged to Eq. 4. e=S ! 3K ass e nK ass
4
The amount of Me-mannose bound per mole of lectin i.e., e, was determined in triplicate and plotted according to Eq. 4. 2.2.3. CD measurements Circular dichroism measurements were carried out at 25³C on a Jasco spectropolarimeter model J-720 using a Sekonic X^Y plotter (model SPL-430), NESLAB water bath model RTE 110 and quartz cell of path length 0.1 or 1.0 cm. Other experimental conditions were, sensitivity 5 mdeg, response time 2 s, scan speed 20 nm/min and a resolution of 0.1 nm. The mean residue ellipticity [a] of a protein was calculated using the expression a a= 10 nUC p Ul Where a is the observed ellipticity in mdeg, n the number of amino acid residues in the protein whose molar concentration is Cp and l is the path length in cm. 3. Results The N-terminal sequence up to 27 amino acid residues of CCL was established. The N-terminal residue of the lectin was threonine as reported before [19]. An attempt was also made to establish homology of the N-terminal sequence with the N-terminal
Fig. 1. N-Terminal amino acid sequence of legume lectins showing homologies within the Viceae tribe and Con A. Here CCL represents the N-terminal amino acid sequence of Cajanus cajan lectin, LCL (Lens culinaris lectin), LOL (Lathyrus ochrus lectin) and PSL (Pisum sativum lectin). The top numbering is for FBL (fava bean lectin) [22].
sequence of few other mannose/glucose speci¢c legume lectins of the same tribe (Fig. 1) [22]. 3.1. Interaction of Me-K-mannose to CCL The Scatchard plot of the equilibrium dialysis data of CCL in 20 mM Tris HCl bu¡er containing 0.5 M NaCl (pH 7.0) with Me-K-mannose at 37³C is shown in Fig. 2. A linear plot was obtained, the slope of which gave the binding constant Kass . A similar plot was obtained at 16³C (¢gure not shown for clarity). The results are summarized in Table 1. 3.2. Metal ion requirements A ¢xed concentration of the demetallized lectin was incubated with increasing concentration of metal ions in 10 mM Tris HCl bu¡er pH 7.0 and the lectin
Table 1 Binding data for lectin sugar interaction Lectin
Sugar
Kass (T³C)
No. of binding sites
Reference
CCL
Me-K-D-mannopyranoside
1.61U104 M31 (16) 1.40U104 M31 (37)
2 2
Present study Present study
Con A
Me-K-D-mannopyranoside
2.06U104 1.65U104 0.83U104 1.70U104 1.20U104 1.25U104
(2) (8) (27) (8) (20) (37)
2 2 2 2 2 2
[22] [22] [22] [23] [23] present study
Pea
Me-K-D-mannopyranoside
0.33U104 M31 (9) 0.19U104 M31 (19)
2 2
[23] [23]
M31 M31 M31 M31 M31 M31
BBAGEN 24794 11-5-99
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
Fig. 2. Scatchard plot for the binding of methyl K-D-mannopyranoside to CCL. A ¢xed concentration of CCL (37 WM in dimer) in 20 mM Tris-HCl bu¡er pH 7.0 containing 0.5 M NaCl in a dialysis bag was incubated with 4^400 WM methyl KD-mannopyranoside in the same bu¡er at 37³C for 24 h. After the attainment of equilibrium the decrease in Me-mannose concentration was estimated in the dialyzate. Analysis of results according to Eq. 4 given in text yielded the value of association constant Kass and the number of sugar binding sites, X. Each circle (b) represents the mean of three independent experiments performed in triplicate.
activity was measured. The six metal salts were CaCl2 , MnCl2 , ZnCl2 , CoCl2 , CrCl3 and MgSO4 . The demetallized lectin showed barely detectable activity in the presence of increasing concentration of any of the six metal salts. However, when the incubation mixture contained ¢xed concentration of the lectin (2.15 mM), IgM (0.45 WM), CaCl2 (4 WM) and increasing concentration of Mn2 , a steep rise in glycoprotein-precipitating activity occurred at a in£ection point ratio [Mn2 ]/[lectin] of 2.0 (Fig. 3). When increasing concentration of Ca2 were added to the mixture of Cajanus cajan lectin (2.15 WM) and goat IgM (0.45 WM) containing ¢xed amount of MnCl2 (4 WM) similar in£ection point was observed. Further analysis of the results showed that each subunit of the lectin bound one Ca2 and one Mn2 . 3.3. E¡ect of pH The e¡ect of pH on the interaction of CCL with rabbit erythrocytes, dextran and goat IgM was investigated in three bu¡er solution in the pH range of 2.0^10.5. The three bu¡er solutions were of 10 mM
381
concentration and contained 1 mM each of Ca2 and Mn2 . These were sodium citrate bu¡er (pH 2.0^6.0) sodium phosphate bu¡er (pH 7.0^9.0) and Tris-HCl (pH 9.5^10.5). In a total volume of 2.0 ml of a given bu¡er were taken 8.23 WM of the lectin with 1.74 WM goat IgM (at this molar ratio the precipitation of goat IgM with CCL was maximum) and the mixture was incubated overnight at 25³C. The amount of protein precipitated at each pH value was determined in triplicate. Both the dextran and glycoprotein precipitation by the seed lectin was measurable in the pH range of 2.0^10.5. In both these cases, the amount of protein precipitated increased steeply with increase in pH from pH 2.0 to 6.0. Further increase in pH of the bu¡er beyond pH 6.0 caused marked reduction in the precipitation. It should be noted that the speci¢c bu¡er ion e¡ect was negligible here, as similar lectin activity values were obtained in two bu¡er solutions of identical pH values (e.g. bu¡er solution, each of pH 6.0 were prepared with sodium citrate as well as sodium phosphate and pH 9.0 with sodium phosphate and Tris-HCl). The bellshaped pH pro¢le can be described in terms of the two types of ionizable groups. M M =1 H K 1 K 2 =H where M* represents the maximal attainable value of
Fig. 3. E¡ect of metal ions on the IgM precipitating activity of CCL. Curve (b) representing the glycoprotein precipitating activity was obtained when lectin (2.15 WM) was incubated with IgM (0.45 WM) and CaCl2 (4 WM) and increasing concentration of MnCl2 . Similar curves with increasing concentration of either Ca2 (a), Mn2 (R) or Mg2 (O) alone are shown.
BBAGEN 24794 11-5-99
382
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
at pH 6.0 and 9.0 were similar with peak at 196^198 nm and trough at 220^225 nm. At pH 2.0, the trough shifted to lower wavelength, i.e. 215 nm (Fig. 5). CD spectra of demetallized, metallized and metallized along with sugar methyl K-D-mannopyranoside were also studied in the far UV region at pH 6.0 taking 10 mM sodium citrate bu¡er (Fig. 6). Interestingly, the spectral features remain the same in all the three conditions. 4. Discussion Fig. 4. E¡ect of pH on the interaction of CCL with goat IgM (b) and dextran (R). The solid theoretical curve was drawn with appropriate pK1 and pK2 values. The ¢lled circles and triangles represent the experimental values of M (protein precipitated) at di¡erent pH values. For other details see text.
Cajanus cajan, a major pulse crop of India belongs to the subfamily Papilionoidea and tribe Viceae. It
M (protein precipitated), K1 and K2 are the dissociation constants of the two ionizable groups of the lectin. The maximum amount (M*) of the glycoprotein lectin which was precipitated by the dextran at pH 6.0 was determined to be 0.547 mg/ml. Interestingly, the pH dependence of both the precipitin reaction viz. lectin dextran and lectin IgM interactions were found to be similar. The theoretical curve drawn with pK1 and pK2 of 3.75 and 8.3, respectively, was found to account for the observed pH dependence of the two precipitin reactions [23] (Fig. 4). Further analysis of the results on pH dependence of lectin activity could be performed by plotting log M against pH. The initial values of the slope of the straight line between log M and pH can be computed both for the `acid' and `basic' arms of the pH pro¢le given in Fig. 4. A rough value for the slope, i.e. dlog M/dpH was estimated to be not more than one in both the curves. 3.4. CD measurements CD measurements of the Cajanus cajan lectin was carried out at three di¡erent pH values. The lectin solution was dialyzed against 10 mM HCl+10 mM KClO4 (pH 2.0), 10 mM sodium phosphate bu¡er, pH 6.0 and 10 mM sodium borate bu¡er, pH 9.0. The solutions were ¢ltered through a Millipore ¢lter (0.45 Wm) and ¢nally deaerated. The spectral features
Fig. 5. E¡ect of pH on the circular dichroic spectra of CCL. Circular dichroic spectra of CCL was measured in 10 mM HCl+10 mM KClO4 at pH 2.0 (^ c ^), 10 mM sodium phosphate bu¡er pH 6.0 (- - -) and 10 mM sodium borate bu¡er at pH 9.0 (999) in the wavelength region 190 to 250 nm. For experimental details see text.
BBAGEN 24794 11-5-99
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
Fig. 6. Circular dichroic spectra of demetallized (- - -), metallized (^ c ^) and metallized plus sugar bound CCL (999). For experimental details see text.
contains 3% Con A reactive carbohydrate and the protein is predominantly acidic in nature [24]. As expected, the 27 amino acid residues at the N-terminal of the lectin molecule exhibit more than 80% homology with the other lectins of the same tribe (Fig. 1) [20]. The lectin has a similar amino acid content per 100 residues to that of the pea lectin and Con A [19,22]. Analysis of the carbohydrate binding properties of the lectin suggested that the lectin is glucose/mannose-speci¢c having greater a¤nity towards the Kmethyl substituted sugars [19]. Results on equilibrium dialysis showed that the lectin binds methyl K-mannopyranoside with an association constant of 1.4U104 M31 at 37³C. This is comparable to the a¤nity constant of Con A [25,26] for the same sugar, but is more than that of pea lectin [26] (Table 1). The lectin seems to have one binding site per subunit for carbohydrate residues. The transition metal ion requirement of CCL was similar to Con A [27] and other legume lectins as it required one previously bound metal ion Ca2 or Mn2 and another in the incubation mixture (Ca2 and Mn2 ) for the lectin to be fully active. These results suggested that the simultaneous presence of both Ca2 and Mn2 is neces-
383
sary and that the lectin has one Ca2 and one Mn2 binding site per subunit in addition to the carbohydrate binding site. Other transition metal ions tested had little or no e¡ect on the binding properties of the lectin. Available data on the pH dependence of the lectin-mediated dextran and IgM precipitation (Fig. 4) suggest the involvement of two ionizable groups with pK values near 3.75 and 8.3. The fact that the initial slope of the `acid' and `basic' arms of the two curves did not exceed 1.0 shows that the number of each type of ionizable groups would be one. Further comparison of the pK values with those found for ionizable groups in proteins [28] indicate that most likely the two groups are aspartic acid and K-amino group of the N-terminal amino acid. Thus, the bellshaped pH pro¢le can be explained in terms of the deprotonation of an aspartic acid residue, which is one of the conserved residue at the active site of most legume lectins [9], and a charged amino group of the N-terminal threonine [19] lying near the active site. The marked change in the lectin activity produced by changing the pH from 3.0 to 9.0 cannot be attributed to denaturation of the lectin molecule since the CD spectra (Fig. 5) of the lectin at pH 6.0 and 9.0 were found to be similar except at pH 2.0. However, the results con¢rm that the decrease in lectin activity with change in pH from 6.0 to 9.0 is not due to changes in the secondary structure of the protein, but a consequence of ionization of ionizable groups present at or near the carbohydrate binding site. However, in the lower pH range, slight changes in protein conformation along with the titration of functional group near the carbohydrate binding site could be attributed to be the cause of decreased lectin activity. The CD spectra in the presence of metal ion and metal ion plus sugar (Me-K-mannopyranoside) suggest that ligand binding in presence of metal ion is accompanied by undetectable changes in the secondary structure of the lectin. Binding of sugar to the combining site of the lectin does not seem to accompany any major structural rearrangement. This argument is consistent with the high resolution X-ray crystallographic ¢ndings of several legume lectins in complex with their speci¢c saccharide [9]. Thus, the sugar combining sites of legume lectins are quite similar and partly superimposable. The different topographies of the combining site for binding di¡erent sugars arise from the ring of hypervariable
BBAGEN 24794 11-5-99
384
S. Ahmad et al. / Biochimica et Biophysica Acta 1427 (1999) 378^384
amino acid residues found around the perimeter of the carbohydrate binding site [8,29]. The metal binding site is, however, located close to the carbohydrate binding site and the amino acid residues, which are involved in the hydrogen bond formation with sugar also coordinate the metal ions. Acknowledgements This paper is dedicated in the memory of the late Prof. A. Salahuddin. The authors are grateful to Prof. I.J. Goldstein, The University of Michigan, Ann Arbor, MI for the N-terminal sequence determination.
References [1] Y.C. Lee, FASEB J. 6 (1992) 3193^3200. [2] D. Solis, P. Fernandez, T. Diaz Murino, J. Jimenez Barbero, M. Martin Lomas, Eur. J. Biochem. 214 (1993) 677^685. [3] E.J. Toone, Curr. Opin. Struct. Biol. 4 (1994) 719^728. [4] R. Loris, J. Hamelryck, J. Bouckaert, L. Wyns, Biochim. Biophys. Acta 1383 (1998) 9^36. [5] I.J. Goldstein, C.E. Hayes, Adv. Carbohydr. Chem. Biochem. 35 (1978) 127^334. [6] N.M. Young, R.P. Oomen, J. Mol. Biol. 228 (1992) 924^934. [7] N. Sharon, Trends Biol. Sci. 18 (1993) 221^226. [8] R. Banerjee, K. Das, R. Ravishanker, K. Suguna, A. Surolia, M. Vijayan, J. Mol. Biol. 259 (1996) 281^296. [9] V. Sharma, A. Surolia, J. Mol. Biol. 267 (1997) 433^445. [10] N. Sharon, H. Lis, Essays Biochem. 30 (1995) 59^75.
[11] I.E. Liener, N. Sharon, I.J. Goldstein (Eds.), The Lectins : Properties, Functions and Applications in Biology and Medicine, Academic Press, London, 1996. [12] L.A. Lasky, Science 258 (1992) 964^969. [13] K. Sastry, R.A. Ezekowitz, Curr. Opin. Immunol. 5 (1993) 59^66. [14] N.J. Brewin, I.V. Kardailsky, Trends Plant Sci. 2 (1997) 92^ 99. [15] I.M. Skvortsov, V.V. Ignatov, FEMS Microbiol. Lett. 165 (1998) 223^229. [16] A. Ayouba, D. Martin, P. Rouge, FEMS Microbiol. Lett. 92 (1992) 41^46. [17] Y. Bourne, A. Ayouba, P. Rouge, C. Cambillau, J. Biol. Chem. 269 (1994) 9429^9435. [18] D. Gupta, H. Kattner, X. Dong, H.J. Gabius, C.F. Brewer, Glycobiology 6 (1996) 843^851. [19] S. Siddiqui, S. Hasan, A. Salahuddin, Arch. Biochem. Biophys. 319 (1995) 426^431. [20] O.H. Lowry, N.J.M. Rosebrough, A.L. Farr, R.J. Randall, J. Biol. Chem. 193 (1951) 265^275. [21] M. Dubois, K.A. Gilles, J.K. Hamilton, P.A. Rebers, F. Smith, Anal. Chem. 28 (1956) 350^356. [22] N. Sharon, H. Lis, FASEB J. 4 (1990) 3198^3208. [23] A. Ahmad, S. Shah, S. Ahmad, Med. Sci. Res. 24 (1996) 107^109. [24] A. Salahuddin, R.H. Khan, J. Protein Chem. 17 (1998) 181^ 185. [25] W. Bessler, J.A. Shafer, I.J. Goldstein, J. Biol. Chem. 249 (1974) 2819^2822. [26] F.P. Schwarz, K.D. Puri, R.G. Bhat, A. Surolia, J. Biol. Chem. 268 (1993) 7668^7677. [27] J.W. Becker, G.N. Reeke, J.L. Wang, B.A. Cunningham, G.M. Edelman, J. Biol. Chem. 250 (1974) 1513^1524. [28] T.E. Creighton, Proteins, Freeman, New York, 1993, pp. 1^ 47. [29] K. Yamamoto, Y. Konami, T. Osawa, T. Kimura, J. Biochem. 111 (1992) 89^95.
BBAGEN 24794 11-5-99