- Email: [email protected]
A fluorimetric study of the lanthanides binding to Concanavalin A
0020.71 IX/84 $3.00 + 0.00 Copyright fQ 1984 Pergamon Press Ltd
fnr. J. Biorhem. Vol. 16. No. 12, pp. 1409~1413, 1984 Printedin Great Britain. All rights reserved
A FLUORIMETRIC BINDING
STUDY OF THE LANTHANIDES TO CONCANAVALIN A
LUCIANA AVIGLIANO, PATRIZIA
ADuccI,
PATRIZIA SIRIANNI
and
ALESSANDRO FINAZZI-AGR~*
Istituto di Chimica Biologica, Universiti di Roma “La Sapienza”,
P. le A. Moro
5, 00185. Rome,
Italy
(Receiaed 10 November 1983) Abstract-l. The binding of Tb3+ and other lanthanides to Con A has been studied by sensitized Tb?+ luminescence, by quenching of intrinsic fluorescence and by activity measurements. 2. In all the experimental conditions tested, it was found that holo and apo Con A bind lanthanide ions at a site different from the binding sites of the constitutive metals, Mn2+ and Ca2+. 3. The bound lanthanide did not affect the saccharide binding ability and the hemoagglutinating ability of Con A. 4. The intrinsic fluorescence of Con A is quenched by the binding of Tb3+ and Gd3+. The same quenching is obtained by shifting the pH of Con A from pH 6.5 to 4.5. 5. It is proposed that H+ and Ln3+ completely quench a tryptophan, perhaps the residue 88 or 182.
INTRODUCTION Concanavalin A (Con A) is a lectin isolated from Canavalia ensijormis, which selectively binds carbohydrates with definite configuration (Goldstein et al., 1965). The native protein shows two metal binding sites, called S, and S,, for Mn2+ and Ca2+ respectively. The first site can also accept various transition metals (Ni’+, Co’+, Zn2+, Cd2+) without losing the sugar binding ability (Brown et al., 1977; Christie et al., 1979; Pandolfino et al., 1980; Sherry et al., 1975). Lanthanide ions also bind Con A but their binding sites and their effects on protein activity are not yet fully defined. In fact Sherry et al. (1975), on the basis of a partial activation of saccharide binding ability, suggested the binding of Tb3+ to both S, and S2 sites, while Barber et al. (1975) and Richardson and Behnke (1978) showed the presence of a site S, distinct from both Mn*+ and Ca2+ sites. We have reinvestigated the interaction between this lectin and the lanthanide Tb3+ by spectroscopic and activity measurements, also in light of recent reports indicating that the molecular properties and the metal requirement for activation change with pH (Christie et al., 1978; Christie et al., 1979; Pandolfino et al., 1980), and in the presence of glycides (Brewer et al., 1983; Koenig et al., 1983). MATERIALS AND
METHODS
Con A was either obtained from Pharmacia (Uppsala Sweden) or purified from Jack beans meal according to Agrawal and Goldstein (1967). The methyl-a-o-mannopyranoside (a-MDM) was purchased from Fluka. The fluorescent sugar, 4-methylumbelliferyl-a-n-mannopyranoside (MUM) was obtained from Serva Feinbiochemica (Heidelberg, F.R.G.). Protein concentration was determined spectrophotometrically using an extinction coefficient *Author
to whom
correspondence
should
be addressed.
El&,, = 12.4 (Brown et al., 1977) and is expressed as binding equivalents on the basis of monomer mol. wt = 27,000 (Edelman ef al., 1972). Demetallized Con A (apo Con A) was prepared according to the method of Brown et al. (1977). Reconstituted protein was obtained by incubating, over a period of 12 hr, ape Con A with 20 eauivalents of MnCI, and CaCI, followed bv dialysis against appropriate buffer containing 500 PM EDTA to remove non-bound ions. When apo Con A was incubated in the presence of only one metal ion, namely CaCl,, MnCI, (2&50 equivalents) or TbC1, (5&200 equivalents) EDTA treatment was avoided, since metal dissociates from Mn-Con A or Ca-Con A upon addition of excess EDTA (Brown et al., 1977). In some experiments apo Con A was equilibrated with I-MDM (150 mM) in order to obtain the “locked” form (Brewer et al., 1983; Koenig et al., 1983). Parallel experiments were done with holo Con A. Mn2* content of protein samples was checked by EPR comparing the signal obtained after bringing the protein solution to pH 1 with HCI with that of a MnQ standard. The saccharide binding activity of Con A was tested following the quenching of MUM fluorescence upon addition of protein (Christie er al., 1978). The hemoagglutinating activity of the lectin was tested according to Minetti et al. (1976) using human erythrocytes from healthy donors. Red blood cells were washed three times in saline and suspended to a 1-3 x IO8 cells/ml concentration 25 ~1 of this suspension were added to serial dilutions of protein samples in a V-shaped plate of Takatsy’s microtitrator. Optical absorption spectra were recorded with a Beckman model 5230 spectrophotometer; fluorescence spectra with a FICA 55 corrected fluorimeter. Samples of apo or holo Con A (5520~M) were titrated with TbCI, in acetate or imidazolate buffer in the pH range 46.5. The experiments cannot be performed at higher pH values due to the formation of highly insoluble rare earth hydroxides (Prados et al., 1974). Possible displacement of Mn*+ from holo Con A was checked by X-band EPR using a V4502 Varian Spectrometer operating at 77 K. Comuetition between Tb3+ and Gd’+ was determined bv the lossof luminescence intensity of a Tb3+-saturated ConA samples upon addition of increasing amounts of GdCl,.
1409
LUUANA
1410
AVIGLIANO
Wavelength
et al.
Inm)
Fig. 1. Terbium binding to Con A. Fluorescence spectra of IOpM holo Con A: (A) (A’) in the presence of 200pM TbCl, in 100mM imidazole-HCI buffer (pH 6.3); (B) (B’) in the presence of 2 mM TbCI, in 100 mM acetate buffer (pH 5.3); (C) 2 mM TbCI, buffer (pH 5.3). The insert shows the difference spectrum obtained subtracting Tb3+-containing (A’) from Tb j+-free (A) Con A. Excitation wavelength: The binding curves were fitted by a Bogart non-linear least-squares fitting program. For competition experiments. the association constant for Gd” was calculated according to De Jersey and colleagues (1980). A Hewlett-Packard 9866A was used for the computation.
RESULTS
Stoichiometry of Th 3’ hinding to holo, ape und recon stituted Con A The addition of TbCI, to native, metal depleted and reconstituted Con A is accompanied by an increase of Tbi+ emission and by a decrease of the protein intrinsic fluorescence (Fig. I). The proteinsensitized Tb3+ emission allows to calculate the number of binding sites present on Con A and their affinity for the lanthanide. Identical results were obtained using commercial ConA or samples of the protein purified in our laboratory. Computation of the binding curves showed that one major binding site per Con A monomer (mol. wt = 27,000) was present. A few sites with lower affinity (Richardson and Behnke, 1978) were also present but their contribution to the overall fluorescence was negligible under our experimental conditions. The major binding site was present in holo, apo and reconstituted Con A, thus being quite distinct from both Mn-binding site (S,) and Cabinding site (S). Its affinity for Tb3+ was pHdependent. The relative values of K, were 2.28 t (SE) 0.24 x IO’M-’ at pH 5.3 and 1.8 t_ (SE) 0.17 x lo4 Mm’ at pH 6.3. In no case a competition between Tb3+ and MnZ+ or Ca2+ was observed. Neither the “locked” form of apo ConA obtained by dialysis against r-MDM was able to bind Tb3+ in the place of Mn2+ or Ca*+. A quenching induced by excess Mn’+
on the bound
Th-‘+ luminescence
was observed.
in the absence and in the absence and in 100 mM acetate the spectrum of 285 nm.
Control experiments showed that in acetate buffer (but not in imidazole) Mn’+ quenches the luminescence of Tb3+ even in the absence of protein. A competition was instead observed between Gd3+ by Tb3+ emission, that and Tb3’ as determined allowed to calculate the affinity for the binding of the former to Con A (Ku = 7.38 + (SE) 0.85 x IO4 Mm’ at pH 6.3).
Wavelengrh
Inm)
Fig. 2. Sugar binding ability of metal-containing Con A. Fluorescence of l5pM MUM in the absence and in the presence of various Con A derivatives (30 PM): (A) MUM in 100 mM imidazole-HCI buffer (pH 6.3). (B) plus apo Con A, (C) plus apo Con A + Tb’+, (D) plus apo Con A + Ca”, (E) plus apo Con A + Mn”, (F) plus reconstituted Con A, (G) plus native Con A, (H) plus native Con A + TbZ+. Each determination was made immediately after the addition of the protein, Protein samples were obtained as described in “Materials and Methods.” Excitation wavelength: 320 nm.
Table
Protein
Lanthanides
binding
I. Hemoagglutinating
activity
A A + Tb’+ + Tb’+ + Ca’+ + Ma’+ Con A Con A + Tb’+
2.31 2.37 14.80 14.80 6.95”-5.05h 2.31 2.31 2.37
Protein samples were obtained as described in Materials and “Incubation at pH 5.3; ‘incubation at pH 6.3. No significant two incubation buffers used was observed with any other Each experiment was done in duplicate and the data are the experiments
Effect of lanthanides
on the activity of Con A
Con A samples in the presence or absence of Tb’+ did not show any significant difference in the binding of MUM (Fig. 2) or in the erythrocyte agglutination (Table 1). Addition of Tb3+ to demetallized Con A did not restore the biological activity measured with the same tests. The saccharide binding activity of the protein has been tested both in the presence of excess metal ions and after exhaustive dialysis. The hemagglutinating activity has been measured only after dialysis since the presence of free bivalent cations (Zakai et al., 1976) or free Tb 3f induces red blood cell agglutination. Quenching of Con A intrinsicjluorescence H’
by Ln’+ and
As reported above, the binding of lanthanides to Con A is accompanied by a decrease of intrinsic protein fluorescence. This quenching at pH values < 5.5 was a function of lanthanide ions bound and levelled off when the binding sites were saturated with Ln3+ ions. The extent of quenching induced upon titration allowed to quantitate the binding of Tb’+ or Cd’+ to the protein. The affinity constants obtained in this way were in good agreement with those for the major
A
of apo- and metal-containing
Minimal agglutinating concentration (pg/ml)
samples
native Con native Con apo Con A apo Con A apo Con A apo Con A Reconstituted Reconstituted
to Concanavalin
1411 Con A
“b of hemoagglutinatmg activity 100 100 16 16 34”47h 100 100 100 Methods. differences between the Con A derivative. results of at least three
binding site calculated using the sensitized Tb’+ luminescence. On the other hand the quenching induced by Eu’+ did not level at high ion/protein ratio, but the protein fluorescence was completely quenched by excess Eu’+. The quenching effect is accompanied by a blue shift in the emission wavelength from 338 to 330 nm while the excitation spectrum was not affected. The difference emission spectrum of Con A vs Tb-‘+saturated Con A is shown in the insert of Fig. 1. At pH values below 5.5 the quenching effect of Tb’+ become less evident until it completely disappeared at pH 4.5. It is interesting to recall that Tb’+ still binds to the protein at this pH value as determined by Tb’+ sensitized emission. These experiments showed that the intrinsic fluorescence of Con A is pH dependent. As reported in Fig. 3 both the intensity and the energy of this fluorescence show a titration like behaviour with an apparent pK = 5.5. The fitting of experimental points to a titration curve is less satisfactory at higher pH that indicates the presence of other phaenomena (ionization of different groups, aggregation of the protein). The difference fluorescence spectrum induced by low pH is similar to that shown in Fig. I. Both holo and apo Con A are quenched in the same way by H+ and Ln’+ furtherly indicating that lanthanides ions bind at a site different from those of Mn2+ and Ca’+. DISCUSSION
Fig. 3. Effect of pH on the intrinsic fluorescence of Con A. Fluorescence intensity (A) and wavelength of emission peak (0) of holo Con A as a function of pH. Solid line is calculated assuming a pK value of 5.5. Excitation wavelength = 285 nm.
In the present paper it has been definitely shown that Con A shows specific binding sites for rare earth distinct from those occupied, in the native protein, by Mn2+ and Ca2+ (Barber et al., 1975; Richardson and Behnke, 1978). This conclusion was reached on the basis of spectroscopic and activity measurements showing that the binding of Tb’+ to Con A is quite unaffected by the presence or subsequent addition of Mn2+ from the protein as determined by EPR, nor affects at any extent the ability of Con A to bind MUM or to agglutinate erythrocytes, contrary to previous evidence (Sherry et al., 1975). Lanthanide ions appear to bind to a main site per monomer and to some--4 according to Richardson and Behnke (1978)-secondary sites. The major site binds Tb3+ in competition with other lanthanides and in a pH-dependent way. A peculiar feature of this
LUCIANA AVIGLIANO et al.
1412
binding is the concomitant quenching of Con A intrinsic fluorescence. This quenching could either be due to direct or indirect lanthanide effects on the emitting residues. Indirect (long range) effects can be either conformational change(s) or non-radiative energy transfer (FGrster-type) (Fiirster, 1965) which deactivate the excited state of tryptophanyl residues. Singlet-singlet energy transfer has been proposed by Horrocks and colleagues (1981) as the mechanism underlaying the sensitized emission of Tb3+ bound to proteins. It is worth noting that at pH lower than 5 rare earth ions still bind to Con A without affecting the intrinsic fluorescence of the protein. On the other hand H+ is able to quench the Con A intrinsic fluorescence like lanthanides. Here the possibility of energy transfer is unlikely since no new absorbing species are formed. An interesting feature of the quenching induced by Ln3+ or H+ is the difference spectrum generated in the emission profile by the presence of the cation (Fig. I, insert) which indicates the disappearance of a fluorophore with maximum emission at 350 nm. This species contributes to the overall fluorescence in unquenched samples for 20-25”/ In a protein containing several tryptophans, like Con A (4 residues per monomer) (Cunningham et al., 1975), it is often difficult to resolve the individual contribution of each of them. The overall fluorescence emission of such proteins, F, is: I: F=
n, ti G,) di
s iI where i, and i., are the minima at shorter and longer wavelength side of the emission peak, n is the number of emitting tryptophans and 4 is the quantum yield at each wavelength 1. The contribution of each tryptophan is a function of its micro-environment. Burnstein et ul. (1973) proposed to classify the fluorescent tryptophanyl residues of proteins into three groups depending on the respective maximum of emission. Assuming that every tryptophan contained in Con A has the same quantum yield, one would suggest that Ln3+ or H+ binding to Con A causes the complete quenching of one out of the four tryptophans present. Furthermore this tryptophan should belong to the Burnstein’s class III, i.e. fully solvated and protruding into the solvent. An inspection on the sequence and tridimensional structure of Con A (Becker et al., 1975; Reeke et al., 1975) indicates that only two tryptophanyl residues meet this requirement, at least in the crystalline state, namely Try 88 and Try 182. The former has been shown to participate to the monomer-monomer contact region and therefore might be at least in part prevented from interaction with the solvent. Interestingly these two tryptophanyl residues are located in the same region of the protein and close to the binding site for Ln3+ and heavy metals (e.g. Pb2+) detected by crystallography. This site is formed by Glu 87 and Asp I36 in a symmetry-related monomer belonging to the same dimer. His 180 was also indicated as a third ligand (Becker et al., 1975; Recke et al., 1975). It is tempting to speculate that the protonation of, or Ln’+ coordination to, one of these side chains could cause a non-radiative deactivation
of a nearest tryptophanyl moiety. The quenching effect of protonated carboxylated or imidazole is well known. This effect has been attributed to the electron withdrawal effect of the proton (Cowgill, 1963). Analogously Ln’+ ions coordinated to Con A might act as electron trap deactivating the tryptophan excited state (Ricci and Kilichowski, 1974). Alternatively energy transfer may take place very efficiently given the proximity between the donor tryptophan and the acceptor lanthanide causing the complete quenching of the former. Whatever the quenching mechanism operating, the experimental findings suggest that the conformation and properties of crystalline Con A are mostly retained by the protein in solution. Acknowledgement-The
sideri for performing
authors wish to thank Dr A. DeEPR controls on the protein. REFERENCES
Agrawal B. B. L. and Goldstein I. J. (1967) Isolation of Concanavalin A by specific adsorption on cross-linked dextran gels. Biochim. hiophys. Acra 147, 262-271. Barber B. H., Fuhr B. and Carver J. P. (1975) A magnetic resonance study of Concanavalin A. Identification of a lanthanide binding site. Biochemistry 14, 4075-4082. Becker J. W.. Reeke G. N. Jr, Wang J. L.. Cunningham B. A. and Edelman G. M. (1975) The covalent anld threedimensional structure of Concanavalin A. III Structure of the monomer and its interactions with metal and saccharides. J. hiol. Chem. 250, 1513~1524. Brewer C. F., Brown R. D. and Koenig S. H. (1983) Kinetics of conformational transitions of demetalized Concanavalin A. Biochem. hioplly,s. Res. Commun. 112, 595-60 1. Brown R. D., III, Brewer C. F. and Koenig S. H. (1977) Conformation states of Concanavalin A. Kinetics of transitions induced by interaction with Mn’+ and Ca” ions. Biochemistry 16, 3883-3896. Burnstein E. A., Vedenkina N. S. and Ivkova M. N. (1973) Fluorescence and the location of tryptophan residues in protein molecules. Photo&em. Pholohiol. 18, 263-279. Christie D. J., Alter G. M. and Magnuson J. A. (1978) Saccharide binding to transition metal ion fret Concanavalin A. Biochemiswy 17, 44254430. Christie D. J., Munske G. R. and Magnuson J. A. (1979) Activation of saccharide binding in demetalized Concanavalin A by transition metal ions. Biochemistry 18, 4638-4644. Cowgill R. W. (1963) Fluorescence and the structure of proteins. I Effects of substituents on the fluorescence of indole and phenol compounds. Archs Biochem. Biophys. loo,
36-44.
Cunningham B. A., Wang J. L.. Waxdal M. I. and Edelman G. M. (1975) The covalent and three-dimensional structure of Concanavalin A. II Amino acid sequence of cyanogen bromide fragment FT. J. hiol. Chem. 250, 1503-1512. De Jersey J., Lahue R. S. and Martin R. B. (I 980) Terbium luminescence as a probe of the calcium binding site of trypsin and cx-chymotrypsin. Arch.7 Biochem. Biophys. 205, 536-542.
Edelman G. M., Cunningham B. A., Reeke G. N. Jr, Becker J. W., Wadal J. W. and Wangh J. L. (1972) The covalent and three-dimensional structure of Concanavalin A. Proc. natn. Acad.
Sci. U.S.A.
69, 25862584.
Fiirster T. (1965) Delocalized excitation and excitation transfer. In Modern Quantum Chemistry (Edited by Sinanoglu 0.) Part III, pp. 93-137. Academic Press, New York.
Lanthanides
binding
Goldstein I. J., Hollerman C. E. and Smith E. E. (1965) Protein-carbohydrate interaction. II Inhibition studies on the interaction of Concanavalin A with polysaccharides. Biochemistry 4, 876-883. Horrocks W. Dew Jr and Collier W. E. (1981) Lanthanide ion luminescence probes. Measurement of distance between intrinsic protein fluorophores and bound metal ions: quantitation of energy transfer between tryptophan and terbium (III) or europium (III) in the calcium-binding protein parvalbumin. J. Am. them. Sot. 103, 2856-2862. Koenig S. H., Brown R. D. and Brewer C. F. (1983) Binding of saccharide to demetalized Concanavalin A. Biochemistry 22, 6221-6226. Minetti M., Aducci P. and Teichner A. (1976) A new agglutinating activity from wheat flour inhibited by tryptophan. Biochim. hiophys. Acta 437, 505-517. Pandohino E. R., Christie D. J., Munske G. R., Fry J. and Magnuson J. A. (1980) Activation of Concanavalin A by Cd’+. J. hiol. Chem. 225, 8772-8775. Prados R., Stadtherr L. G., Donato H. Jr and Martin R. B.
to Concanavalin
A
1413
(1974) Lanthanide complexes of aminoacids. J. inorg. nucl. Chem. 36, 689-693. Reeke G. N. Jr, Becker J. W. and Edelman G. M. (1975) The covalent and three-dimensional structure of Concanavahn A. IV Atomic coordinates, hydrogen bonding and quaternary structure. J. biol. Chem. 250, 1525-1547. Ricci R. W. and Kilichowski K. B. (1974) Fluorescence quenching of the indole ring system by lanthanide ions. J. phys. Chem. 78, 1953-1956. Richardson C. E. and Behnke W. D. (1978) Physical studies of lanthanide binding to Concanavalin A. Biochim. biophys. Aria 534, 261-274. Sherry A. D., Newman A. D. and Gutz C. G. (1975) The activation of Concanavahn A by lanthanide ions. Biochemistry 14, 2191-2196. Zakai N., Kulka R. G. and Loyter A. (1976) Induction of fusion of intact human erythrocytes or erythrocyte ghosts by the combined action of Ca” and phosphate. Sixfh European Congress on Electron Microscopy, Jerusalem (Edited by Ben Sham) Vol. II, pp. 324326. Tal Intern.
fnr. J. Biorhem. Vol. 16. No. 12, pp. 1409~1413, 1984 Printedin Great Britain. All rights reserved
A FLUORIMETRIC BINDING
STUDY OF THE LANTHANIDES TO CONCANAVALIN A
LUCIANA AVIGLIANO, PATRIZIA
ADuccI,
PATRIZIA SIRIANNI
and
ALESSANDRO FINAZZI-AGR~*
Istituto di Chimica Biologica, Universiti di Roma “La Sapienza”,
P. le A. Moro
5, 00185. Rome,
Italy
(Receiaed 10 November 1983) Abstract-l. The binding of Tb3+ and other lanthanides to Con A has been studied by sensitized Tb?+ luminescence, by quenching of intrinsic fluorescence and by activity measurements. 2. In all the experimental conditions tested, it was found that holo and apo Con A bind lanthanide ions at a site different from the binding sites of the constitutive metals, Mn2+ and Ca2+. 3. The bound lanthanide did not affect the saccharide binding ability and the hemoagglutinating ability of Con A. 4. The intrinsic fluorescence of Con A is quenched by the binding of Tb3+ and Gd3+. The same quenching is obtained by shifting the pH of Con A from pH 6.5 to 4.5. 5. It is proposed that H+ and Ln3+ completely quench a tryptophan, perhaps the residue 88 or 182.
INTRODUCTION Concanavalin A (Con A) is a lectin isolated from Canavalia ensijormis, which selectively binds carbohydrates with definite configuration (Goldstein et al., 1965). The native protein shows two metal binding sites, called S, and S,, for Mn2+ and Ca2+ respectively. The first site can also accept various transition metals (Ni’+, Co’+, Zn2+, Cd2+) without losing the sugar binding ability (Brown et al., 1977; Christie et al., 1979; Pandolfino et al., 1980; Sherry et al., 1975). Lanthanide ions also bind Con A but their binding sites and their effects on protein activity are not yet fully defined. In fact Sherry et al. (1975), on the basis of a partial activation of saccharide binding ability, suggested the binding of Tb3+ to both S, and S2 sites, while Barber et al. (1975) and Richardson and Behnke (1978) showed the presence of a site S, distinct from both Mn*+ and Ca2+ sites. We have reinvestigated the interaction between this lectin and the lanthanide Tb3+ by spectroscopic and activity measurements, also in light of recent reports indicating that the molecular properties and the metal requirement for activation change with pH (Christie et al., 1978; Christie et al., 1979; Pandolfino et al., 1980), and in the presence of glycides (Brewer et al., 1983; Koenig et al., 1983). MATERIALS AND
METHODS
Con A was either obtained from Pharmacia (Uppsala Sweden) or purified from Jack beans meal according to Agrawal and Goldstein (1967). The methyl-a-o-mannopyranoside (a-MDM) was purchased from Fluka. The fluorescent sugar, 4-methylumbelliferyl-a-n-mannopyranoside (MUM) was obtained from Serva Feinbiochemica (Heidelberg, F.R.G.). Protein concentration was determined spectrophotometrically using an extinction coefficient *Author
to whom
correspondence
should
be addressed.
El&,, = 12.4 (Brown et al., 1977) and is expressed as binding equivalents on the basis of monomer mol. wt = 27,000 (Edelman ef al., 1972). Demetallized Con A (apo Con A) was prepared according to the method of Brown et al. (1977). Reconstituted protein was obtained by incubating, over a period of 12 hr, ape Con A with 20 eauivalents of MnCI, and CaCI, followed bv dialysis against appropriate buffer containing 500 PM EDTA to remove non-bound ions. When apo Con A was incubated in the presence of only one metal ion, namely CaCl,, MnCI, (2&50 equivalents) or TbC1, (5&200 equivalents) EDTA treatment was avoided, since metal dissociates from Mn-Con A or Ca-Con A upon addition of excess EDTA (Brown et al., 1977). In some experiments apo Con A was equilibrated with I-MDM (150 mM) in order to obtain the “locked” form (Brewer et al., 1983; Koenig et al., 1983). Parallel experiments were done with holo Con A. Mn2* content of protein samples was checked by EPR comparing the signal obtained after bringing the protein solution to pH 1 with HCI with that of a MnQ standard. The saccharide binding activity of Con A was tested following the quenching of MUM fluorescence upon addition of protein (Christie er al., 1978). The hemoagglutinating activity of the lectin was tested according to Minetti et al. (1976) using human erythrocytes from healthy donors. Red blood cells were washed three times in saline and suspended to a 1-3 x IO8 cells/ml concentration 25 ~1 of this suspension were added to serial dilutions of protein samples in a V-shaped plate of Takatsy’s microtitrator. Optical absorption spectra were recorded with a Beckman model 5230 spectrophotometer; fluorescence spectra with a FICA 55 corrected fluorimeter. Samples of apo or holo Con A (5520~M) were titrated with TbCI, in acetate or imidazolate buffer in the pH range 46.5. The experiments cannot be performed at higher pH values due to the formation of highly insoluble rare earth hydroxides (Prados et al., 1974). Possible displacement of Mn*+ from holo Con A was checked by X-band EPR using a V4502 Varian Spectrometer operating at 77 K. Comuetition between Tb3+ and Gd’+ was determined bv the lossof luminescence intensity of a Tb3+-saturated ConA samples upon addition of increasing amounts of GdCl,.
1409
LUUANA
1410
AVIGLIANO
Wavelength
et al.
Inm)
Fig. 1. Terbium binding to Con A. Fluorescence spectra of IOpM holo Con A: (A) (A’) in the presence of 200pM TbCl, in 100mM imidazole-HCI buffer (pH 6.3); (B) (B’) in the presence of 2 mM TbCI, in 100 mM acetate buffer (pH 5.3); (C) 2 mM TbCI, buffer (pH 5.3). The insert shows the difference spectrum obtained subtracting Tb3+-containing (A’) from Tb j+-free (A) Con A. Excitation wavelength: The binding curves were fitted by a Bogart non-linear least-squares fitting program. For competition experiments. the association constant for Gd” was calculated according to De Jersey and colleagues (1980). A Hewlett-Packard 9866A was used for the computation.
RESULTS
Stoichiometry of Th 3’ hinding to holo, ape und recon stituted Con A The addition of TbCI, to native, metal depleted and reconstituted Con A is accompanied by an increase of Tbi+ emission and by a decrease of the protein intrinsic fluorescence (Fig. I). The proteinsensitized Tb3+ emission allows to calculate the number of binding sites present on Con A and their affinity for the lanthanide. Identical results were obtained using commercial ConA or samples of the protein purified in our laboratory. Computation of the binding curves showed that one major binding site per Con A monomer (mol. wt = 27,000) was present. A few sites with lower affinity (Richardson and Behnke, 1978) were also present but their contribution to the overall fluorescence was negligible under our experimental conditions. The major binding site was present in holo, apo and reconstituted Con A, thus being quite distinct from both Mn-binding site (S,) and Cabinding site (S). Its affinity for Tb3+ was pHdependent. The relative values of K, were 2.28 t (SE) 0.24 x IO’M-’ at pH 5.3 and 1.8 t_ (SE) 0.17 x lo4 Mm’ at pH 6.3. In no case a competition between Tb3+ and MnZ+ or Ca2+ was observed. Neither the “locked” form of apo ConA obtained by dialysis against r-MDM was able to bind Tb3+ in the place of Mn2+ or Ca*+. A quenching induced by excess Mn’+
on the bound
Th-‘+ luminescence
was observed.
in the absence and in the absence and in 100 mM acetate the spectrum of 285 nm.
Control experiments showed that in acetate buffer (but not in imidazole) Mn’+ quenches the luminescence of Tb3+ even in the absence of protein. A competition was instead observed between Gd3+ by Tb3+ emission, that and Tb3’ as determined allowed to calculate the affinity for the binding of the former to Con A (Ku = 7.38 + (SE) 0.85 x IO4 Mm’ at pH 6.3).
Wavelengrh
Inm)
Fig. 2. Sugar binding ability of metal-containing Con A. Fluorescence of l5pM MUM in the absence and in the presence of various Con A derivatives (30 PM): (A) MUM in 100 mM imidazole-HCI buffer (pH 6.3). (B) plus apo Con A, (C) plus apo Con A + Tb’+, (D) plus apo Con A + Ca”, (E) plus apo Con A + Mn”, (F) plus reconstituted Con A, (G) plus native Con A, (H) plus native Con A + TbZ+. Each determination was made immediately after the addition of the protein, Protein samples were obtained as described in “Materials and Methods.” Excitation wavelength: 320 nm.
Table
Protein
Lanthanides
binding
I. Hemoagglutinating
activity
A A + Tb’+ + Tb’+ + Ca’+ + Ma’+ Con A Con A + Tb’+
2.31 2.37 14.80 14.80 6.95”-5.05h 2.31 2.31 2.37
Protein samples were obtained as described in Materials and “Incubation at pH 5.3; ‘incubation at pH 6.3. No significant two incubation buffers used was observed with any other Each experiment was done in duplicate and the data are the experiments
Effect of lanthanides
on the activity of Con A
Con A samples in the presence or absence of Tb’+ did not show any significant difference in the binding of MUM (Fig. 2) or in the erythrocyte agglutination (Table 1). Addition of Tb3+ to demetallized Con A did not restore the biological activity measured with the same tests. The saccharide binding activity of the protein has been tested both in the presence of excess metal ions and after exhaustive dialysis. The hemagglutinating activity has been measured only after dialysis since the presence of free bivalent cations (Zakai et al., 1976) or free Tb 3f induces red blood cell agglutination. Quenching of Con A intrinsicjluorescence H’
by Ln’+ and
As reported above, the binding of lanthanides to Con A is accompanied by a decrease of intrinsic protein fluorescence. This quenching at pH values < 5.5 was a function of lanthanide ions bound and levelled off when the binding sites were saturated with Ln3+ ions. The extent of quenching induced upon titration allowed to quantitate the binding of Tb’+ or Cd’+ to the protein. The affinity constants obtained in this way were in good agreement with those for the major
A
of apo- and metal-containing
Minimal agglutinating concentration (pg/ml)
samples
native Con native Con apo Con A apo Con A apo Con A apo Con A Reconstituted Reconstituted
to Concanavalin
1411 Con A
“b of hemoagglutinatmg activity 100 100 16 16 34”47h 100 100 100 Methods. differences between the Con A derivative. results of at least three
binding site calculated using the sensitized Tb’+ luminescence. On the other hand the quenching induced by Eu’+ did not level at high ion/protein ratio, but the protein fluorescence was completely quenched by excess Eu’+. The quenching effect is accompanied by a blue shift in the emission wavelength from 338 to 330 nm while the excitation spectrum was not affected. The difference emission spectrum of Con A vs Tb-‘+saturated Con A is shown in the insert of Fig. 1. At pH values below 5.5 the quenching effect of Tb’+ become less evident until it completely disappeared at pH 4.5. It is interesting to recall that Tb’+ still binds to the protein at this pH value as determined by Tb’+ sensitized emission. These experiments showed that the intrinsic fluorescence of Con A is pH dependent. As reported in Fig. 3 both the intensity and the energy of this fluorescence show a titration like behaviour with an apparent pK = 5.5. The fitting of experimental points to a titration curve is less satisfactory at higher pH that indicates the presence of other phaenomena (ionization of different groups, aggregation of the protein). The difference fluorescence spectrum induced by low pH is similar to that shown in Fig. I. Both holo and apo Con A are quenched in the same way by H+ and Ln’+ furtherly indicating that lanthanides ions bind at a site different from those of Mn2+ and Ca’+. DISCUSSION
Fig. 3. Effect of pH on the intrinsic fluorescence of Con A. Fluorescence intensity (A) and wavelength of emission peak (0) of holo Con A as a function of pH. Solid line is calculated assuming a pK value of 5.5. Excitation wavelength = 285 nm.
In the present paper it has been definitely shown that Con A shows specific binding sites for rare earth distinct from those occupied, in the native protein, by Mn2+ and Ca2+ (Barber et al., 1975; Richardson and Behnke, 1978). This conclusion was reached on the basis of spectroscopic and activity measurements showing that the binding of Tb’+ to Con A is quite unaffected by the presence or subsequent addition of Mn2+ from the protein as determined by EPR, nor affects at any extent the ability of Con A to bind MUM or to agglutinate erythrocytes, contrary to previous evidence (Sherry et al., 1975). Lanthanide ions appear to bind to a main site per monomer and to some--4 according to Richardson and Behnke (1978)-secondary sites. The major site binds Tb3+ in competition with other lanthanides and in a pH-dependent way. A peculiar feature of this
LUCIANA AVIGLIANO et al.
1412
binding is the concomitant quenching of Con A intrinsic fluorescence. This quenching could either be due to direct or indirect lanthanide effects on the emitting residues. Indirect (long range) effects can be either conformational change(s) or non-radiative energy transfer (FGrster-type) (Fiirster, 1965) which deactivate the excited state of tryptophanyl residues. Singlet-singlet energy transfer has been proposed by Horrocks and colleagues (1981) as the mechanism underlaying the sensitized emission of Tb3+ bound to proteins. It is worth noting that at pH lower than 5 rare earth ions still bind to Con A without affecting the intrinsic fluorescence of the protein. On the other hand H+ is able to quench the Con A intrinsic fluorescence like lanthanides. Here the possibility of energy transfer is unlikely since no new absorbing species are formed. An interesting feature of the quenching induced by Ln3+ or H+ is the difference spectrum generated in the emission profile by the presence of the cation (Fig. I, insert) which indicates the disappearance of a fluorophore with maximum emission at 350 nm. This species contributes to the overall fluorescence in unquenched samples for 20-25”/ In a protein containing several tryptophans, like Con A (4 residues per monomer) (Cunningham et al., 1975), it is often difficult to resolve the individual contribution of each of them. The overall fluorescence emission of such proteins, F, is: I: F=
n, ti G,) di
s iI where i, and i., are the minima at shorter and longer wavelength side of the emission peak, n is the number of emitting tryptophans and 4 is the quantum yield at each wavelength 1. The contribution of each tryptophan is a function of its micro-environment. Burnstein et ul. (1973) proposed to classify the fluorescent tryptophanyl residues of proteins into three groups depending on the respective maximum of emission. Assuming that every tryptophan contained in Con A has the same quantum yield, one would suggest that Ln3+ or H+ binding to Con A causes the complete quenching of one out of the four tryptophans present. Furthermore this tryptophan should belong to the Burnstein’s class III, i.e. fully solvated and protruding into the solvent. An inspection on the sequence and tridimensional structure of Con A (Becker et al., 1975; Reeke et al., 1975) indicates that only two tryptophanyl residues meet this requirement, at least in the crystalline state, namely Try 88 and Try 182. The former has been shown to participate to the monomer-monomer contact region and therefore might be at least in part prevented from interaction with the solvent. Interestingly these two tryptophanyl residues are located in the same region of the protein and close to the binding site for Ln3+ and heavy metals (e.g. Pb2+) detected by crystallography. This site is formed by Glu 87 and Asp I36 in a symmetry-related monomer belonging to the same dimer. His 180 was also indicated as a third ligand (Becker et al., 1975; Recke et al., 1975). It is tempting to speculate that the protonation of, or Ln’+ coordination to, one of these side chains could cause a non-radiative deactivation
of a nearest tryptophanyl moiety. The quenching effect of protonated carboxylated or imidazole is well known. This effect has been attributed to the electron withdrawal effect of the proton (Cowgill, 1963). Analogously Ln’+ ions coordinated to Con A might act as electron trap deactivating the tryptophan excited state (Ricci and Kilichowski, 1974). Alternatively energy transfer may take place very efficiently given the proximity between the donor tryptophan and the acceptor lanthanide causing the complete quenching of the former. Whatever the quenching mechanism operating, the experimental findings suggest that the conformation and properties of crystalline Con A are mostly retained by the protein in solution. Acknowledgement-The
sideri for performing
authors wish to thank Dr A. DeEPR controls on the protein. REFERENCES
Agrawal B. B. L. and Goldstein I. J. (1967) Isolation of Concanavalin A by specific adsorption on cross-linked dextran gels. Biochim. hiophys. Acra 147, 262-271. Barber B. H., Fuhr B. and Carver J. P. (1975) A magnetic resonance study of Concanavalin A. Identification of a lanthanide binding site. Biochemistry 14, 4075-4082. Becker J. W.. Reeke G. N. Jr, Wang J. L.. Cunningham B. A. and Edelman G. M. (1975) The covalent anld threedimensional structure of Concanavalin A. III Structure of the monomer and its interactions with metal and saccharides. J. hiol. Chem. 250, 1513~1524. Brewer C. F., Brown R. D. and Koenig S. H. (1983) Kinetics of conformational transitions of demetalized Concanavalin A. Biochem. hioplly,s. Res. Commun. 112, 595-60 1. Brown R. D., III, Brewer C. F. and Koenig S. H. (1977) Conformation states of Concanavalin A. Kinetics of transitions induced by interaction with Mn’+ and Ca” ions. Biochemistry 16, 3883-3896. Burnstein E. A., Vedenkina N. S. and Ivkova M. N. (1973) Fluorescence and the location of tryptophan residues in protein molecules. Photo&em. Pholohiol. 18, 263-279. Christie D. J., Alter G. M. and Magnuson J. A. (1978) Saccharide binding to transition metal ion fret Concanavalin A. Biochemiswy 17, 44254430. Christie D. J., Munske G. R. and Magnuson J. A. (1979) Activation of saccharide binding in demetalized Concanavalin A by transition metal ions. Biochemistry 18, 4638-4644. Cowgill R. W. (1963) Fluorescence and the structure of proteins. I Effects of substituents on the fluorescence of indole and phenol compounds. Archs Biochem. Biophys. loo,
36-44.
Cunningham B. A., Wang J. L.. Waxdal M. I. and Edelman G. M. (1975) The covalent and three-dimensional structure of Concanavalin A. II Amino acid sequence of cyanogen bromide fragment FT. J. hiol. Chem. 250, 1503-1512. De Jersey J., Lahue R. S. and Martin R. B. (I 980) Terbium luminescence as a probe of the calcium binding site of trypsin and cx-chymotrypsin. Arch.7 Biochem. Biophys. 205, 536-542.
Edelman G. M., Cunningham B. A., Reeke G. N. Jr, Becker J. W., Wadal J. W. and Wangh J. L. (1972) The covalent and three-dimensional structure of Concanavalin A. Proc. natn. Acad.
Sci. U.S.A.
69, 25862584.
Fiirster T. (1965) Delocalized excitation and excitation transfer. In Modern Quantum Chemistry (Edited by Sinanoglu 0.) Part III, pp. 93-137. Academic Press, New York.
Lanthanides
binding
Goldstein I. J., Hollerman C. E. and Smith E. E. (1965) Protein-carbohydrate interaction. II Inhibition studies on the interaction of Concanavalin A with polysaccharides. Biochemistry 4, 876-883. Horrocks W. Dew Jr and Collier W. E. (1981) Lanthanide ion luminescence probes. Measurement of distance between intrinsic protein fluorophores and bound metal ions: quantitation of energy transfer between tryptophan and terbium (III) or europium (III) in the calcium-binding protein parvalbumin. J. Am. them. Sot. 103, 2856-2862. Koenig S. H., Brown R. D. and Brewer C. F. (1983) Binding of saccharide to demetalized Concanavalin A. Biochemistry 22, 6221-6226. Minetti M., Aducci P. and Teichner A. (1976) A new agglutinating activity from wheat flour inhibited by tryptophan. Biochim. hiophys. Acta 437, 505-517. Pandohino E. R., Christie D. J., Munske G. R., Fry J. and Magnuson J. A. (1980) Activation of Concanavalin A by Cd’+. J. hiol. Chem. 225, 8772-8775. Prados R., Stadtherr L. G., Donato H. Jr and Martin R. B.
to Concanavalin
A
1413
(1974) Lanthanide complexes of aminoacids. J. inorg. nucl. Chem. 36, 689-693. Reeke G. N. Jr, Becker J. W. and Edelman G. M. (1975) The covalent and three-dimensional structure of Concanavahn A. IV Atomic coordinates, hydrogen bonding and quaternary structure. J. biol. Chem. 250, 1525-1547. Ricci R. W. and Kilichowski K. B. (1974) Fluorescence quenching of the indole ring system by lanthanide ions. J. phys. Chem. 78, 1953-1956. Richardson C. E. and Behnke W. D. (1978) Physical studies of lanthanide binding to Concanavalin A. Biochim. biophys. Aria 534, 261-274. Sherry A. D., Newman A. D. and Gutz C. G. (1975) The activation of Concanavahn A by lanthanide ions. Biochemistry 14, 2191-2196. Zakai N., Kulka R. G. and Loyter A. (1976) Induction of fusion of intact human erythrocytes or erythrocyte ghosts by the combined action of Ca” and phosphate. Sixfh European Congress on Electron Microscopy, Jerusalem (Edited by Ben Sham) Vol. II, pp. 324326. Tal Intern.