Oct 9, 1992 - several temperatures by the method of Chipman et al. (12). The temperature dependence of the association constants was made use.
THEJOURNAL OF BIOLOGICAL CHEMISTRY
Vol. 268, No. 22, Imue of August 5,pp. 16378-16337.1993 Printed in U.S.A.
Q 1993 by The American Society for Biochemistry and Molecular Biology, h e .
Thermodynamic and Kinetic Studies on the Mechanism of Binding of Methylurnbelliferyl Galactosidesto theBasic Agglutinin from Winged Bean (Psophocarpus tetragonolobus)" (Received for publication, October 9, 1992, and inrevised form, March 25, 1993)
Kamal Deep PuriSB,M. I. Khanv, Dipti GuptaS 11, and Avadhesha SuroliaS** From the $Molecutor Biophysics Unit,Indian Instituteof Science, Bangalore 560 012, India and Wiuision of Biochemical Sciences, National Chemical Laboratory, Pune 411 008, India
The binding of winged bean basic agglutinin (WBA
membranes, lectins can affect intercellular recognition, cell growth, and differentiation. They are, therefore, widely used examined by extrinsic fluorescence titrationand in cell biology, biochemistry, and histochemistry to isolate stopped-flow spectrofluorimetry. Upon binding to and/or to characterize cell surface carbohydrates and glycoWBA I, MeUmb a-galactosides show quenching in flu- proteins (1-3). Lectins require configurational and structural orescence intensity, decrease in UV absorbance witha complementarity of sugars for interaction to occur and have concomitant blue shift, and decrease in fluorescence been employed as tools for exploring the structure anddynamexcited-state lifetimes. However, their &analogues ics of cell surfaces (2,4). show enhancement in fluorescence intensity, increase The basic lectin (agglutinin) (PI= 10) isolated from winged in UV absorbance with a red shift, andan increase in bean (Psuphocurpus tetragonulobus) (WBA I)' consists of two fluorescence excited-state lifetimes. This implies that the umbelliferyl groups of a- and &galactosides expe- identical subunits of M, = 29,000 with one binding site/ rience non-polar andpolar microenvironments, re- subunit, The lectin binds with high affinity to D-galactopyspectively, upon binding toWBA I. Replacement of the ranosides (5, 6). Immunochemical analysis of ligand binding anomeric hydroxyl group of galactose by 4-methylum- revealed that the lectin recognizes human blood group A and belliferyl moiety increases the affinity of resulting B but not H-antigenic determinants (7). Khan et ul. (8) have characterized the binding specificity of saccharides. Substitution of C-2 hydroxyl of galactose by an acetamido group leads to increased affinity due WBA I, by thermodynamic and kinetic analysis using N to a favorable entropy change. This suggests that ac- dansylgalactosamine (GalNDns) asthe indicator ligand. They etamido group of MeUmb-a/&GalNAc binds to a rela- observed that a large hydrophobic substituent on the C-2 tively non-polar subsite of WBA I. Most interestingly, carbon of galactopyranoside and a hydrophobic substituent this substitution also reduces the association rate con- in thea-conformation at theanomeric position increases the stants dramatically. Inspection of the activation pa- binding affinity and that theC-4 and C-6 hydroxyl groups are rameters reveals that the enthalpy of activation is the critical for sugar binding. The stopped-flow spectrofluorimelimiting factor for the differences in the forward rate tric studies with GalNDns revealed that the binding mechaconstants for these saccharides and the entropic con- nism is enthalpically driven, and the dissociation rate conis small. tribution to the activation energy stant ( k l = 3.2 X lo-' s" at 25 "C) is the slowest reported for any lectin-carbohydrate complex. To further probe the role of bulkier substituents at the anomeric position and at C-2 carbon of galactose, we have Many biological recognition and adhesion processes involve studied the interaction between 4-methylumbelliferyl glycothe formation of saccharide-protein complexes. To undersides of galactose and N-acetylgalactosamine to WBA I. The stand theselectivity and origin of the association energy, it is 4-methylumbelliferyl group in a- and P-linkage resides in important to know the nature of the forces controlling the relatively non-polar and polar environments, respectively, in saccharide-protein interaction. One of the best characterized the combining site of the lectin. These studiesalso reveal that carbohydrate-mediated recognition of the cell surface receptor the substitution of the hydroxyl group atc-2, as in is the association between lectins, the highly specific carbo- GalaMeUmb and GalBMeUmb, by a bulky acetamido group, hydrate-binding proteins, and thecarbohydrate moiety of the as in GalNAcaMeUmb and GalNAcpMeUmb, reduces the glycolipids and glycoproteins present on the cell membrane. association rate constantsdramatically. Although the binding Because of the integration of carbohydrate moieties in cell
I) to 4-methylumbelliferyl (MeUmb) galactosides was
*This work was supported by a grant from the Department of Science and Technology, Government of India (to A. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18U.S.C. Section 1734 solelyto indicate this fact. Senior research fellow of the Council of Scientific and Industrial Research, India. (1 Research associate in a project funded by the Department of Science and Technology (to A .S.). ** To whom correspondence should be addressed Tel.: 344411 (ext. 2459); Fax: 91 080 341683.
' The abbreviations used are: WBA I, winged bean basic agglutinin; GalNAc, N-acetyl-galactosamine; MeUmb, methylumbelliferyl (7hydroxy-4-methylcoumarin); GalNDns, dansylgalactosamine; dansyl, 5-dimethylaminonaphthalene-1-sulfonyl; GalaMe, methyl a-galactopyranoside; Gal@Me,methyl 8-galactopyranoside; GlcpMe, methyl @glucopyranoside; GlcNAc, 2-acetamido-2-deoxyglucopyranose;all sugars used are D-sugars unless otherwise specified; PBS, phosphatebuffered saline; GalNAcalPMe, methyl-N-acetyl-alp-galactosaminide; GalNAca/@MeUmb, 4-methylumbelliferyl-N-acetyl-a/~-galactosaminide; GalalBMeUrnb, 4-methylumbelliferyl-a/@-galactopyranoside; GlcflMeUmb, 4-methyl umbelliferyl-8-glucopyranoside; GlcNAcBMeUmb, 4-methylumbelliferyl-N-acetyl-~-glucosaminide.
16378
WBA I
4-Methylumbelliferyl Galactoside Reaction with
16379
was expressed in terms of a protomer ( M , 29,000). Fluorescence Measurements-The fluorescence emission spectra of 4-methylumbelliferyl galactosides in the absence of, and at various concentrations of,WBA I were recorded on a Shimadzu RF-5000 fluorescence spectrofluorimeter. The samples were excited at 318 nm and the emission spectra recorded above 330 nm, with slit widths of 5 nm for both the monochromators. Fluorescence titration measurements were made on a Union Giken MATERIALS ANDMETHODS FS 501A fluorescence polarization instrument equipped with photon Ligands-GalaMe, GalNAcaMe, GlcaMe, GlcNAcbMeUmb, counting photomultipliers. Samples were excited at 318 nm with a GlcBMeUmb,GalSMeUmb, GalaMeUmb, and GalNAcBMeUmb were 3.5-nm slit and emission was monitored by means of a metal interpurchased from Sigma. GalNAcaMeUmb was a kind gift from Prof. ference band passfilter (X, = 10 nm) centered at 375 nm along with Irwin. J. Goldstein, University of Michigan, Ann Arbor. All other a 335-nm cutoff filter. Samples in 1 X 1 X 4.5-cm quartz cuvettes reagents used wereof analytical grade. The concentrations of the were placed in a thermostatedcopper holder maintained at a constant fluorescent sugars were determined using the molar extinction coef- temperature (20.1 "C) by means of a constant temperature Lauda ) of 1.36 X lo' M" cm" a t 318 nm (9). All solutions circulating water bath. All measurements were made under continuficient ( ~ ~ 1 8value of MeUmb galactosides were free of 7-hydroxy-4-methylcoumarin as ous stirring with an inbuilt stirrer. The fluorimeter is controlled by a assayed by the absence of fluorescence at 447 nm. microprocessor, which allowsaveraging of several readings. Typically, Preparation of WBA I-WBA I was prepared from defatted and the average values of 10 measurements (S.D. = +0.5%) with 1-s gate dried winged bean seed meal according to the method of Khan et al. time were taken. (8), with a slight modification, that the pHof the supernatantof the For determination of the number of binding sites and association seed meal extract was adjusted to 4.0 with acetic acid and after 1 h constants for the binding of MeUmb galactosides, a fixed concentrathe solution was centrifuged to remove acid insoluble material. The tion of WBA I (2.82 FM) was titrated against MeUmb galactosides pH of the supernatantwas then readjusted to 7.2 with NaOH solution (1-60 p M ) at 25 "C and the resulting changes in fluorescence were prior to the ammonium sulfate fractionation. The protein thus puri- measured after equilibrating the sample for 2 h. A linear standard fied wasfound to be homogeneous by SDS-polyacryamide gel electro- curve for MeUmb galactosides in PBS was obtained in the range of phoresis. Protein concentrations were estimated by the method of (1-60 p M ) concentration at the same temperature. The relative fluoLowry et al. (10) using bovine serum albumin as a standard. The rescence in the absence and presence of WBA I gave the amount of protein concentration was expressed in terms of a dimer ( M , 58,000) free and bound sugar, for analysis of the dataaccording to Scatchard for the Scatchard type of titration and for all other experiments it
of numerous ligands to several lectins has been studied, such a striking effect of a substituent of ligand structure has not been observed before. The enthalpy of activation is the limiting factor for the observed differences in the forward rate constants for these saccharides.
FIG. 1. Quenching and enhancement of the fluorescence of GalaMeUmb and GaUMeUmb, respectively, by WBA I at 20 "C. a, fluorescence spectra of a 2-ml solution of 5.48 pM (curue I ) of GalaMeUmb with 4,11, 15,20, 40,50, and 60 pl of WBA I (1300 pM) are shown in curues 1 to 8, respectively. The controls are 60 p~ WBA I (curue 9) and PBS (curue 10). b, titration of (1.24 p ~ of)GalaMeUmb by . gives KO= 2.82 WBA I (1300 p ~ )Inset X lo' IC'. The straight line is drawn according to the regression equation ( r = 0.9989). c, fluorescence spectra of a 2ml solution of 5.5 p~ GalpMeUmb (curue I ) and with successive additions ) of 5 and 10 plof WBA I (700 p ~ are shown in curues 2-5 and 6-10, respectively. Controls are as in ( a ) . d , titration of GalaMeUmb (1.1 p ~ with ) WBA I (700 pM). Inset gives K, = 3.627 X lo' M-'. The straight line is drawn according to theregression equation ( r = 0.9986).
0
50
100
Proteinaliquotadded
/
1
0.31
Protein aliquot added (?I)
150 200 (PI)
,- 1
I
1
4-Methylumbelliferyl Galactoside Reaction with WBA 1
16380
A fixed volume (2 ml) of MeUmb galactoside was also titrated by addition of small, defined aliquots ofWBA I in the cuvette. After each addition, a 2-min interval was given before measurements were taken. The association constants for the interaction of WBA I and MeUmb galactosides were determined from the titration data at several temperatures by the method of Chipman et al. (12). The temperature dependence of the association constants was made use of to determine the thermodynamic parameters of MeUmb galactosides-WBA I interaction. The binding of nonfluorescent, inhibitory sugars to WBA I was studied by competitive substitution titration. To a fixed volume (2 ml) of MeUmb galactosides (1 PM) a defined aliquot (75-100 pl) of WBA I (750 PM) was added, and the fluorescence intensity was measured before and afteraddition of WBA I. This mixture was then titrated with small, defined aliquots of the inhibitory sugar solution, and after each addition, the fluorescence of the sample was measured. The association constants from these titrations were determined according to Bessler et al. (13). Kinetic Studies-Fast reaction kinetic studies were performed on a Union Giken RA 401 stopped-flow spectrophotometer working in the fluorescence mode. Samples were excited at 318nm and the emission was monitored beyond 360 nm by using a cutoff filter at right angles to the exciting beam. The dead time of the instrument under the experimental conditions was determined to be 0.5 ms. The sample reservoirs and the cell compartments were maintained at a constant temperature ( f O . 1 "C) by circulating water from a Lauda constant temperature bath, through jackets surrounding the sample reservoirs and the cell compartments. For determination of the association rate constants, 5.1 p~ of MeUmb galactosides were mixed with various concentrations (50-475 p M ) of WBA I. The dissociation rate constants of sugar-WBA I complexes were evaluated by dissociating the preformed complex with a 12 mM solution of an inhibitory sugar, GalaMe. UVAbsorbance DifferenceSpectra-MeUmb galactoside difference absorbance spectra were recorded with JASCO UV/VIS spectropho-
TABLE I The relative quuntum yields and the fluorescence changes of MeUmb galactosides upon binding to WBA Z Fluorescence change upon binding Ligand Yield" Quenching Enhancement %
68.90 71.40
4.20 GalaMeUmb 23.90 37.74 2.65 GalNAcaMeUmb 1.45 GalPMeUmb 1.40 GalNAcj3MeUmb Relative quantum yield of the bound ligand.
tometer model 7850. The binding of MeUmb galactosides to WBA I was monitored by measurements of changes in their UV absorbance in the range 300-360 nm a t 25 "C.Each half of the tandem cells were filled with equal volumes (0.8 ml) of WBA I and MeUmb galactoside solutions, respectively. Protein concentration was varied from 63 to 444 pM, keeping a fixed concentration of the MeUmb galactosides (50 pM). The base line was recorded before and after tipping both reference and sample cells. Fluorescence Lifetimes-Fluorescence lifetimes were measured with single-photon counting fluorescence lifetime CD900 Edinburgh Instrument fluorometer with a nanosecond flash lamp. The lamp was filled with N, under a pressure of 1.03 bar and operated at 2.84 kV and 25 kHz. Under these conditions, full-width half-maximum was 0.8 ns. The ratio of stop-to-start pulses reaching the time to amplitude converter was kept below 4% to avoid pulse pileup problems. No significant pulse was detected by the photomultipliers when a blank was performed with buffer alone (absence of scatter). Excitation was at a 316-nm NOband through the excitation monochromator, and emission was monitored at 375 nm for MeUmb galactosides and at 450 nm for 4-methylumbelliferone through an emission monochromator. Fluorescence lifetime of the 2 p~ solutions of 4-methylumbelliferone and MeUmb galactosides in PBS was determined in the absence and atvarious concentrations of WBA I at 24 "C maintained using a constant temperature water bath. Fluorescence lifetimes of 2 pM solutions of MeUmb galactosides at varying concentrations of glycerol in PBS were also determined. RESULTS
When MeUmb galactosides were excited at 318 nm and their fluorescence emission spectra were recorded in the absence of and with increasing concentrations of WBA I, the fluorescence intensity of MeUmb a-galactosides (Xmax = 378 nm) was progressively quenched,whereas the fluorescence intensity of MeUmb@-galactosides ,,X,( = 380 nm) was enhanced. These fluorescence changes were reversed by addition of 0.1 M GalaMe and GalNAcaMe (0.1 M), showing thatthey were dueto sugar-specificbinding. Moreover, Glc@MeUmband GlcNAcBMeIJmb did not show any change in fluorescence intensity upon addition of WBA I (250 p M ) and GlcaMe failed to reverse the fluorescence quenching of GalaMeUmb by WBA I. Representative plotsof the quenchingandenhancement of the fluorescence intensities of GalcvMeUmb and GalpMeUmb upon their titration with increasing concentrations of WBA I a t 20 "C are shown in Fig. 1 a and c, respectively. The quantumyields of MeUmb galactosides bound toWBA
TABLE I1 Association constants and thermodynamic parameters for the binding of sugars to WBA I Values in parentheses indicate S.E. values ( n = 4). X
Sugar
15 "C
20 "C
8.12 (rt0.42) 35.40 (f2.7) 24.02 (f1.5) 110.8 (f6.5) 0.827 49.50 (f2.O)
6.66 (f0.31) 28.20 (f1.5) 19.33 (k1.2) 89.31 (k5.3) 0.704 36.27 (21.5) 0.52 69.39 (22.5)
25 "C
30 "C
-AH
-AG
-AS
27.75
21.45 (f0.66) 24.96 (f0.65) 24.04 (f0.8) 27.77 (k0.25) 15.97 25.57 (522) 15.23 27.16 (f0.27)
21.51 (f1.2) 38.71 (k1.5) 21.94 (f1.2) 16.00
~~~~
GalaMe GalaMeUmb GalNAcaMe GalNAcaMeUmb GalPMe" Galj3MeUmb GalNAcpMe" GalNAcj3MeUmb
97.18 (k2.8)
Values obtained from Khan et al. (8).
72.77 (k4.8) 0.603 26.88 (rtl.1)
4.57 (f0.41) 17.72 (f0.80) 12.79 (k0.6) 60.09 (f3.6) 0.519 18.99 (k0.7)
49.11 (f1.9)
36.27 (21.5)
5.52 (k0.29) 21.91 (f1.1) 15.70 (k1.1)
(k1.1)
36.27 (kO.9) 30.47 (f1.5) 32.45 (f1.1) 22.58 59.14 (f2.5) 49.80 (f1.5)
(k1.8)
22.55 114.43 (f5.6) 77.31 (k3.1)
Reaction with WBA I
4-Methylumbelliferyl Galactoside
I were obtained by extrapolating a plot of F o p 0 - FA versus inverse of protein concentration (1/[P],)where Fo and F, are the fluorescence intensities of free sugar and at a particular concentration of protein, respectively (Table I). From the protein concentration-dependent changes of MeUmb galactoside fluorescence, the K, values for their binding were calculated by the method of Chipman et al. (12) by plotting log\(Fo- F,)/(F, - F,)) versw log[P],
from the equation: logl(Fo - Fc)/(Fc - F m ) )
log\[PIt - (hF/hFrn)[L]r) + log K
O
16381
where Fo and F, have the same meaning as before an? AF = ( F , - Fo) and A F , is the maximum change observed, uiz. when all the ligand molecules are complexed with the protein. [PI,, [ L ] , arethetotal proteinand ligand concentration, respectively, and [PI,is the free protein concentration which is the term in brackets on the right hand side, uiz. log[P]f = log([PI, ( f l / f l m ) [ L l t I . Representative plots are shown for the binding of GalaMeUmb (Fig. I b ) and GalpMeUmb (Fig;. Id) at 20 "C. The values of K , (intercept = pK,) determined by this method is 2.82 x lo4 and 3.627 x lo4 M-', respectively, at 20 "C, assuming a stoichiometry of 1:1 with the WBA I protomer of M , 29,000. Values for other MeUmb galactosides rapge from 8.931 X lo4 M" (GalNAcaMeUmb) to 6.939 X lo4 M" (GalNAcpMeUmb) at 20 "C (Table 11). The data obtained by titration of a fixed concentration of the protein with varying concentration of GalpMeUmb, GalNAcpMeUmb, GalaMeUmb, and GalNAcaMeUmb, analyzed at 25 "C according to Scatchard (ll),gave a straightline with association constants (KO)of 2.78 X IO4,4.33 x lo4, 2.09 x IO4, and 7.11 x lo4 and thevalues of the intercepts at theabscissa, n, equal to 2.00, 2.09, 2.18, and 2.14, respectively, for WBA I with M, of 58,000. From the changes in enthalpies as evaluated from Van't Hoff plots (Fig. 2) and Gibbs free energies (AGO = -RT In K J , the changes in entropies ( A P ) for the association of these ligands were determined from equation:
-
A@ = A€f' - TAS'
The K , values for the interaction of the non-fluorescent reference ligands GalaMe and GalNAcaMe were also determined by substitution titration using MeUmb galactosides as the indicator ligands according to themethod of Bessler et al. 9.3L .r (13), and arealso listed in Table 11. Van't Hoff plots of KOfor these sugars are shown in Fig. 2. It may be noted that the association constant values determined for the indicating ligands, uiz. MeUmb galactosides, by substitution titrations are in the range of values obtained from direct titrationsusing the method of Chipman et ai. (12). Thermodynamic parameters for methyl galactosides and MeUmb galactosides are 3.3 3.4 3.5 compared in Table 111. Kinetic Studies-The elementary steps for the binding of ( ~ - 9 MeUmb galactosides to WBA I were evaluated by stoppedFIG. 2. Van't Hoff plots for the association of various sugars flow spectrofluorimetry. The observed rate constant, kobs, for to WBA I are drawn according to the regression equation. The the fluorescence change of an indicator ligand is related in a symbolsused are: (0)GalNAcaMeUmb ( n = 4, r = 0.9998); (0) concentration-dependent manner on the type of elementary GalaMeUmb ( n = 4, r = 0.9978); (A) GalNAcBMeUmb ( n = 4, r = 0.9980); (A)GalpMeUmb ( n = 4, r = 0.9885); (0)GalNAcaMe ( n = step to which the relaxation belongs(14-16). Three most 4, r = 0.9999); and (B) GalaMe ( n = 4, r = 0.9999). The values of KO likely possibilities for such interactions involve either asingle were determined in quadruplicate, and S.E. values are listed in the bimolecular association step between the protein ( P ) and tables. ligand ( L )or the reaction could proceedthrough the formation
IOYJ
TABLE111 Difference betweenA H o , ASo, and AGO values of methyl-galactosides and corresponding MeUmb galactosides Sugar
3 -91.88
GalaMe 27.75 GalaMeUmb 21.94 GalNAcaMe 30.47 GalNAcaMeUmb 16.00 32.45 GalpMe -36.56 -9.60 GalPMeUmb GalNAcpMe GalNAcpMeUmb 49.80 a Values at 20 "C.
-AG&
-
A
H
Q
-ASo
kJ mol"
J mol" K"
21.44 24.96 36.27 24.04 27.77
21.51 38.71
22.55 15.97 22.58 25.57 59.14 15.23 27.16 -11.93
A A P
AAHO
-3.52
AASO
J mol" K"
kJ mol"
-8.52
-17.20 5.94
114.43 77.30
4-Methylumbelliferyl Galactoside Reaction with
16382
WBA I
(a)
2000 Time (msec) I
0.25
I
1
-I
>
0)
3
0
0
- 0.25
>
'0
I
I
100
200
lo
300
[PI ( P M )
FIG. 4. Determination of the rate constants for the association of WBA I with GaZBMeUmb (0)and GalNAcDMeUmb (0). kobs values determined at 20 "C were plotted against the protein concentration [PI (after mixing). Theconcentration of the sugar was fixed a t 2.55 p M (after mixing) for GalPMeUmb and tothe GalNAcpMeUmb. The slope of the linedrawnaccording regression equation yielded a k, value of 1.182 X IO5M" s-' ( n = 7, r = 0.9985) for GalPMeUmb and 8.9 X lo3 M" s" (n = 7, r = 0.9981) for GalNAcpMeUmb.
0.25
c
Time (msec) I
I
I
-I
plot of hobs uersus [PI used for determination of k, and k-, for thefastest ligand,GalPMeUmb andthe slowestligand, GalNAcpMeUmb a t 20 "C is shown in Fig. 4. The values of k-I were alsodetermined directlybydisplacing indicator ligands from their complexes with WBA I by GalaMe (Fig. 3b). Kinetic experiments for these ligands were also carried out at several temperatures rangingfrom 15 to 30 "C in order to calculate the activation parameters using Arrhenius plot (Fig. 5 ) . Activation enthalpies, entropies, and energies were calculated using thefollowing equations: M=E,-RT
-0.25
1
,
,
,
4
FIG. 3. Stopped-flow fluorescence traces of GalaMeUmbWBA I association and dissociation reactions at 20 'C. a, for ) the association reactions equalvolumes of GalaMeUmb (5.1 p ~and WBA I (200 g ~ were ) mixed in the stopped-flow cell. The samples were excited at 318 nm, and emission was recorded above 360 nm. The spectrum represents the average of 10 measurements. Continuous line is the nonlinear least square fit of the data. Kobe thus obtained was 9.58 0.45. Residuals are shown in thebottom panel. b, stopped) flow trace for the dissociation of a complex of WBA I (100 p ~ and ) 12 mM GalaMe. Continuous line is the GalaMeUmb (5.1 p ~ with nonlinear least square fit of the data which gives k-, = 2.08 ? 0.08 s-I. Residuals are shown in the bottom panel.
ln(k/T) = - AHI/RT A@ =
~
+ AS'/R + ln(k'/h) - TASt
where k is the appropriate rate constant, k' is Boltzman's constant, and h is Planck's constant. Activation parameters thus obtained are listed in Tables IV-VII. The association rate constants for bindingof MeUmb galactosides to WBA I, determined in this study vary from 1.18 X lo5 to 0.89 X lo4 M" s-' a t 20 "C. Values of k-l obtained from Y-intercept of kobsuersus [PI plots are consistentwith those estimated from the first order rate constants for the displacementof MeUmb galactosidescomplexed withWBA I by alargeexcess of GalaMe. The k-, values thus calculated for various MeUmb of an initial complex (PLi)which isomerizes to give rise to galactosides range from 0.128 to 3.29 s-l at 20 "C (Tables IVthe finalcomplex PL* or the transformationof a n inactive to VII) . UV Absorbance Difference Spectra-Upon binding toWBA an active form of protein capable of associating with the ligand ( P $ P*). A representative stopped-flow fluorescence 1, the absorbance spectraof MeUmb a-galactosidesundergoes galactosides show a red shift trace of time-dependent change in fluorescence of a blueshift, whereas the @-linked difference spectrum shows GalaMeUmb upon its rapidmixing with WBA I, thecompo- (data not shown). The absorbance nent always kept in great excess uiz. P >> L is shown in Fig. two extremes at about 323 and 338 nm (Fig. 6, a and b). The absorbance change at 338 nm was monitored to calculate the 3a. The observed rate constants, were evaluated from the nonlinear least square fits of the data to a monoexponential association constants. Theoccurrence of this difference specof carbohydrates aswas reaction. Values of association ( kl) and dissociation ( k - J rate trum is directly related to the binding evidenced by thecompleteinhibition of theabsorbance constants were obtained from the slope andtheordinate intercept of the linear plots of kobsuersus [PI.A representative change by GalaMe. From a series of difference spectra ob-
*
4-Methylumbelliferyl Galactoside Reaction withWBA I
16383
12.4 1
! .1
.3 h
FIG. 5. Arrhenius plots for the association of GawMeUmb (0) and GalNAcBMeUmb (0) and dissociationrate constants for the binding of GalOMeUmb (A) and GalNAcaMeUmb (A) to WBA I, were drawn according to regression equation with n = 4, r = 0.9995; n = 4, r = 0.9971; n = 4, r = 0.9978; and n = 4, r = 0.9992, respectively. kl and k-, plotted are themean values of 9-10 independent runs. S.E. for the parameters thus determined are listed in the tables.
3.5
19 Y
.
-
1.72 7
I
Y
C
1.2
2 .o
3.3
2.8 3 .5
3.4
103/T
(K")
the MeUmb &-galactosides were longer than those of the corresponding /3-anomers (Table VIII). The fluorescence lifetimes of MeUmb galactosides were strongly dependent on WBA I concentration. With an increase in WBA I concentration, MeUmb cy-galactosides showed a decrease, whereas P-galactosides showed an increase log(Ml(aAmax - A A ) l = log{Po- L d U / h A m a x ) 1 + log in fluorescence lifetime; however, that of 4-methylumbelliferin which AA,,.is the maximal absorbance change obtained one showed no change indicating that the fluorescence lifefrom a plot of l/AA against the reciprocal of the totalprotein time change is due to the interaction of the MeUmb galactoconcentration Po; Lois the totalligand concentration, and the sides on the carbohydrate specific binding pocket on WBA I, log term represents the free protein concentration. The value which is further evidenced by the complete inhibition of the of K, determined by this method is 2.163 X lo4 M" at 25 "C, fluorescence lifetime change upon addition of GalaMe to a assuming a stoichiometry of 1:1 with the WBA I protomer of preformed complex of MeUmb galactoside and WBA I (Table M , 29,000. This value is in good agreement with the value of VIII). Fluorescence lifetimes of the MeUmb galactosides in2.199 X lo4M" at 25 "C obtained by the fluorescence titration creased by 10-15% at 40% glycerol concentration, then it method. The association constants for other MeUmb galac- remained more or less constant up to 80% glycerol. This tosides were also in good agreement to thevalues determined increase indicates the increase in the rotation correlation time by fluorescence studies. due to immobilization of the fluorophore with an increase in Fluorescence Lifetimes-Fluorescence excited-state life- glycerol concentration. times of 4-methylumbelliferone and MeUmb galactosides, DISCUSSION determined in the absence and at near saturation concentration of WBA I, are listed in Table VIII. The experimental WBA I quenches the fluorescence of MeUmb cy-galactofluorescence decay curves of 4-methylumbelliferone and sides, but enhancesthose of MeUmb &galactosides. This MeUmb galactosides fitted best to a single-lifetime compo- contrast implies that the anomeric configuration of the galnent model, as indicated by the low values of the mean actoside residue bound to WBA I leads to a significant change weighted residue, x* (17) andby the random oscillation around in the environment of the fluorophore. Thus theenvironment zero of the deviation function (data notshown). of the indicating group (4-methylumbelliferyl) of MeUmb aFluorescence lifetimes were in therange of 0.39-0.57 ns for galactosides in the combining site of WBA I appears to be MeUmb galactosides, compared t o a lifetime of 5.33 ns for relatively non-polar, whereas those of MeUmb @-galactopythe free 4-methylumbelliferone. The lifetimes measured for ranosides experiences a polar environment (9, 17, 18). This tained with a constant concentrationof MeUmb galactosides at different WBA I concentrations (Fig. 6, a and b ) , the maximal change was estimated from adouble-reciprocal plot, assuming a1:1complex; the value of the association constant, K, was calculated according to theequation:
TABLEIV Rate constants and activation parameters for the interaction to WBA GalaMeUmb I with Values in parentheses indicate S.E. values ( n = 4). T "C 15 20 6.99
25 30
WBA I
4-Methylumbelliferyl Galactoside Reaction with
16384
k,, "1
X
10-3 x K.
k-,
10-4
10-3 X K,
(equilibrium) (kinetics) "I
S-1
s-l
4.62 35.54 (k0.35) 5.88 28.26 (k0.45) 21.98 (k0.89) 9.40 17.67 (k1.08)
1.30 1.15 (kO.09) 2.08 (kO.ll) 3.18 (rt0.25) 5.32 (k0.46)
15
T
k,,
"C
1"
lo"
k-,
s-l
S-1
X
35.40
20
28.20
25
21.91
30
TABLEV Rate constants and activation parameters for the interaction to WBA GalNAcaMeUmb I with Values in parentheses indicate S.E. values ( n = 4).
17.72
"1
11.2
11.1
8.90
8.90
7.32
7.28
5.98
6.01
(k0.98) 32.41 kJ mol" (k1.84) = -54.34 J mol" K" (k2.56) AG*+, = 48.33 kJ mol" (k1.18)
E*-l = 85.85 kJ mol"
(k2.5) A H * + , = 65.01 kJ mol" (k2.2) AS", = -17.37 J mol" K" (rt1.7) A@+, = 70.10 kJ mol" (k2.8)
- A H - , = -36.06
(equilibrium)
=
Et-, = 67.45 kJ mol"
AH*+,
10" x K.
K.
E*+,= 34.81 kJ mol"
(k0.8) AH*+, = 28.95 kJ mol" (k1.2) AS*+, = -54.8 J mol" K" (k2.06) AGf+, = 45.01 kJ mol" (k1.08)
AH' =
X
(kinetics)
0.103 (k0.008) 0.166 (k0.011) 0.254 (k0.016) 0.395 (kO.019)
(f0.08) 1.48 (f0.09) 1.86 (k0.15) 2.36 (k0.18)
E*+, = 31.34 kJ mol"
10"
kJ mol"
(k1.44) (k1.64) ASo = AS*+] - AS*-, = -37.43 J mol" K-' (k2.5) AGO = A@+, - AG*-~= -25.09 kJ mol" (k0.8)
argument isalso supported by changes in theirlifetimes upon binding to WBA I. The fluorescence excited-state lifetime of 4-methylumbelliferylgroup isdramatically reducedfrom 5.325 ns to less than 0.5 ns when it is covalently attached to Galor GalNAc. Upon bindingto WBA I the lifetime of MeUmb a-galactosides decreases whereas those of MeUmb @-galactosides increases. It may be noted that the values of polarization ( p ) of GalaMeUmb and GalPMeUmb change from 0.101 and 0.130 to 0.402 and 0.400, respectively, upon their total binding to the protein a t 20 "C. Thus the increase and decrease of the fluorescence lifetimes of MeUmb P-galactosides and MeUmb a-galactosides are related to the increase and decrease of the fluorescence quantum yields of these sugars, as thedegree of immobilization of these ligands in the bound state are quite similar as evidenced by almost identical changes in their polarization ( A p = 0.301 and 0.270 for GalaMeUmb and GalpMeUmb, respectively, at 20 "C) upon binding toWBA I. To furtherprobe the combining region of MeUmb galactosides on WBA I, ultraviolet absorbance difference spectroascopic studies were done.A blue shiftintheMeUmb galactoside absorbance spectra resulted in a decrease in the absorbance difference spectra. However, MeUmb P-galactosides show a red shift and an increase in absorbance in the absorbance difference spectra. This behavior of MeUmb galactosides shows that the 4-methylumbelliferyl group resides
(k3.5) 64.86 kJ mol" (k3.25) AS*-, = -38.34 J mol" K" (k1.5) aG*L1 = 76.07 kJ mol" (k2.45)
AH-,
AHo =
ASo =
AH*+,
=
-
= -32.45 kJ mol"
- AS*-]
= -16.00 J mol"
K"
(k0.76) AG' = AG*+] - AG*-] = -27.74 kJ mol" (k0.71)
in an apolar environment when it is a-linked and in a polar environment when it is @-linked to the galactoside (19, 20). WBA I is able to discriminate GalaMe and GalNAcaMe from their corresponding p-anomers (8).However, when their anomerichydroxylgroups are substituted with a bulky 4methylumbelliferyl group, the lectin loses its anomeric preference, as indicated by the association constants of MeUmb a-galactosides (2.82 X lo4 M" and 8.93 X lo4 M-' at 20 "cfor GalaMeUmb and GalNAcaMeUmb,respectively), which are very close to the association constant values of their corresponding @-counterparts(3.63 X lo4 M" and 6.94 X lo4 M" at 20 "C for GalPMeUmband GalNAcpMeUmb, respectively). GalaMeUmb and GalNAcaMeUmb are 4.24 and 4.62 times stronger ligands than GalaMe and GalNAcaMe, respectively. GalpMeUmb and GalNAcPMeUmb show 51.6- and 133.5-fold greater affinities over GalPMe and GalNAcpMe, respectively, indicating apositive contribution of 4-methylumbelliferyl groupfor the association of MeUmb galactosides. Better affinities of MeUmb a-galactosides over the corresponding methyl a-galactosides is probablydue to a relatively more favorable entropic contributions of the 4-methylumbelliferyl group reflectinghydrophobic interaction between the later and the protein. Although enthalpic contributionalso appears to favor theseinteractionsto some extent. Amarginally favorable change in the entropy for the binding of MeUmb amannopyranoside ascompared to methyl a-mannopyranoside
4-Methylumbelliferyl Galactoside Reaction with WBA I TABLEVI Rate constants and activation parameters for interaction to WBA I with GalSMeUmb Values in parentheses indicate S.E. values ( n = 4). T "C 15
k+l x 10-4 I"
k-1
10-~x
x.
(kinetics)
5-1
s-l
9.52 (f0.57) 20 11.82 (k0.14) (f0.55) 14.58 25 (f0.48) (k1.01) 30 17.67 (f0.69) (f1.19)
1.94
10"
"'
x K.
(equilibrium)
49.07
49.50
3.29
35.93
36.27
5.38
27.10
26.88
9.81
18.01
18.99
(k0.13)
Et,,
29.78 kJ mol" (f1.5) AHt+, = 27.38 kJ mol-' (f1.37) AS*+,= -54.24 J mol" K" (k2.71) A@+, = 42.37 kJ mol" =
(k1.8)
16385
mechanism. An interesting feature noted for the first time in protein-sugar interactions is the dependence of the second order rate constants for the binding of WBA I to MeUmb galactosides on the natureof the substituenta t C-2 carbon of galactose as well as on the natureof the linkage of the aglycon 4-methylumbelliferyl moiety. It is seen that the association rate constant of GalPMeUmb is 13.3 times higher than the corresponding GalNAcPMeUmb, in which C-2 hydroxyl is substituted by an acetamido group. The effect of this substitution on the association rate is also seen in the a-anomer, in ligand than which GalaMeUmb is 4.0 timesfastera GalNAcaMeUmb. Thus, substitution of the hydroxyl group at C-2 of galactose by a bulky acetamido group reduces the association rate constant dramatically. The k, values observed here are in the range of 0.7 X lo4 "1 s-l (GalNAcpMeUmb at 15 "C) to 1.8 X lo5 M" s" (GalpMeUmb at 30 "C). These second order rate constants are slower than a diffusion-controlled process and are similar to those for the binding of fluorogenicfchromogenic ligands to jacalin (22), concanavalin A (23, 24), Ricinus communis agglutinin (25), and soybean lectin (26). The binding for such reactions is presumed to involve a putative intermediate complex (PL,)which isomerizes to the final complex PL*
E*-l = 86.65 kJmol" (f4.33) 84.21 kJmol" (k4.05) AS'-, = 59.69 J mol" K-' (f5.4) AG*-, = 66.72 kJ mol" (k3.1) =
with
and DJio =
AH*+,
ASo = AS*+l
- AHt-, -
= -56.83 kJ mol-' (k2.8) AS*-] = -113.93 J mol" K" (f6.3) AGt-, I: -23.45 kJ mol" (+1.17)
Our inability to detect the formation of PL; complex could be due to an unobservable signal change in the faster step. If the first step is always kinetically uncoupled from the second step, PLi -* PL is rate-controlling ie., k-1>> kz, then the kobs has also been attributed by Landschoot et al. (20) to the non- would not be a linear function of [PI. In such a case, kobs polar interaction between the MeUmb group of the former, should progress from a first orderdependence to a zero-order with the lectin. Quenching of fluorescence intensities of dependence as the concentration of the excess component MeUmb a-galactosides upon binding to WBA I is also con- (which is P in our case) increases from a value much lower sistent with a hydrophobic environment around their 4-meth- than l / K l to P >> l/Kl. Since our plots are linear up to 250 ylumbelliferyl moieties. Increase in fluorescence intensities of PM of P, the association constant of the first step (KI) has to MeUmb @-galactosideson binding to WBA I, together with a belower than 2000 M-'. Despite lower second order rate large favorable enthalpic contribution indicates that the bet- constants observed here than the diffusion controlled reacter affinities of these galactosides compared to their corre- tions, the probability of occurrence of reaction intermediate(s) sponding methyl 0-galactosides reflects the formation of ad- within the dead time of the instrument or otherwise is preditional van der Waals contacts and/or hydrogen bonds. In cluded since the fluorescence changes appearing in the either case, the interaction of the 4-methylumbelliferyl group stopped-flow traces, for both the association and dissociation with the protein promotes the association of GalNAc and Gal reactions, are similar in magnitude to those observed in the residues, thereby imparting them a higher affinity as com- steady state titrations. The agreement between kinetically pared to their corresponding methyl galactosides. determined values of association constants ( k 1 / L l ) and Replacement of the hydroxyl group at C-2 carbon of galac- changes in enthalpies and those determined by fluorescence tose by an acetamido group leads to the stabilization of the titrations indicates that the K, and the enthalpy changes are lectin-MeUmb galactoside complex by 2.81 and 1.59 kJ mol" related tothe total binding process and rule out the occurrence in the a- and@-linked MeUmb galactosides, respectively. of anyintermediate that contributes appreciably to these Comparison of the thermodynamic parameters reveals that parameters for the saccharide binding. Linearity of Arrhenius this stabilization is mainly due to a favorable entropy change, plots also rules out the occurrence of dramatic conformational which suggests a strongfavorable interaction that presumably changes in the lectin molecule inthe temperature range involves dispersion forces. Hence, the acetamido group at C- studied. Thus, kinetics of binding of MeUmb galactosides to 2 of galactose in GalNAcaMeUmb and GalNAcPMeUmb ap- WBA I is both quantitatively and qualitatively consistent pears to bind to a subsite which is considerably non-polar. with a single step bimolecular association reaction. Kinetic Studies-The kinetics follow a one-step binding Analyses of activation parameters of MeUmb galactosides AGO = AG'+1
-
16386
4-Methylumbelliferyl Galactoside Reaction with WBA 1
TABLEVI1 Rate constants and activation parameters for interaction to WBA I with GelNAcflMeUmb Values in parentheses indicate S.E.values ( n = 4). 10-4 X K. 10-4 x K. T k+, X lo"
0.02
~~~~~
(kinetics)
k-1
"C
15 20 25 30
MI
0.70 (f0.04) 0.89 (f0.04) 1.20 (f0.11) 1.57 (k0.12)
0.072 (k0.004) 0.128 (fO.01) 0.245 (20.02) 0.432 (20.03)
9.68
9.72
6.95
6.94
4.90
4.91
3.63
3.63
E*-, = 53.36 kJ mol" (k1.83) 88.56 kJ mol" (k2.75) AS*-, = 40.86 J mol" K-' (k3.9) AGf-, = 76.60 kJ mol" (k2.4)
AH*-, =
- AH*-, = -49.80
kJ mol"
(f1.0) 77.32 J mol" K-' (%2.7) AGO = AG*+~ AG*L = -27.16 kJ mol-' (k0.4)
AS' = AS*+,
- AS*-,
0.0 0.01
-0.031
41.15 kJ mol" (k1.2) = 38.76 kJ mol" (k1.17) ASf+, = -36.46 J mol" K-l (k1.13) AG*+, = 49.44 kJ mol" (k1.34)
AH*+,
4
-0.02
=
AHo =
0.01
"'
S"
s-1
(equilibrium)
=
-
reveal that the energy of activation is the limiting factor for the differences in forward rate constant for these saccharides. Much of the activation energy barrier for the association of MeUmb galactosides due is to a high enthalpic contribution since the entropic barrier is insignificant. The values of activation energy (E1$) for the association reaction of the slowest ligand, GalNAcpMeUmb, is 41.15 kJ mol" and that of the fastest ligand GalBMeUmb is 29.78 kJ mol-', indicating that sufficiently large amounts of energy have to be expended for the binding process to take place. This energeti; barrier could be utilized in overcoming'some steric constraints, such as some conformational changes of the protein and/or in breaking hydrogen bonds between the solvent molecules and thesugar, as well as between the protein and the solvent molecules, and for the formation of newer ones between the sugar and theprotein in the complexes. A comparison of the energy diagram for the binding of GalPMeUmb and GalNAcPMeUmb to WBA I, which are the fastest and the slowest binding ligands, respectively (Figs. 7, a and b), provides some insights into the mechanism of the interaction of WBA I with MeUmb galactosides. The entropy of activation for the association process is small, indicating that theassociation process does not involve a highly ordered transition state. Thiscould mean that thesugar can approach the binding pocket in several ways. This is in contrast with the situation with peanut agglutinin (27), where a highly
300
I
I
322.8 338
1
360
300 323.4 338.1 360
x (nm)
A (nm)
FIG. 6. Difference absorbance spectra for the titration of GalaMeUmb (a)and GaUMeUmb ( b )with increasingconcentration of WBA I at 26 OC. Each half of the tandem cells were filled with equal volumes (0.8 ml) of a fixed concentration of MeUmb galactosides at different concentrations of WBA I. The differenceabsorbance spectrum was recorded after mixing the sample cell components. Base line was recorded before and after tipping both the reference and sample cells. a, series of difference-absorbancespectra obtained with a constant concentration of GaIcvMeUmb (25 p ~after , ; mixing) at various concentrations of WBA I (curve I , 31.5 p ~ curve 2, 38.4 p M ; curve 3, 55.5 pM; curve 4, 77.5 p M ; curve 5, 133.5 pM; and curve 6 , 222 pM) after mixing at 25 "c.Curue 7 represents the base line. b, series of difference-absorbance spectra obtained with a constant concentration of GalpMeUmb (26.2 pM after mixing) at various concentrations of WBA I (curve 1 , 31.5 p M ; curve 2, 38.4 pM; curve 3, 55.5 p ~ curve ; 4, 77.5 p ~ curve , 5, 133.5 p ~ after ) mixing at 25 ' C . Curve 6 represents the base line.
TABLEVI11 Fluorescence excited-state lifetimesof MeUmb galactosides and 4-methylumbelliferone at 24 "C in absence and presence of 106.6 G M WBA I Comuound
Lifetime
Chi-sauare
I?s
X2
GalaMeUmb GalaMeUmb WBA I GalNAcaMeUmb GalNAcaMeUmb WBA I
0.565 0.291
1.28 1.13
0.498
0.378
1.41 1.54
GalaMeUmb + WBA I
0.414 0.684
1.37 1.61
0.394 0.564 o.394
1.35 1.35 1.18
5.325 5.343
1.79 1.89
+
+
GalNAcflMeUmb GalNAcj3MeUmb GalNAcPMeUmb WBA WBA II
+ +
4"eUmbe11iferone 4-Meumbelliferone
+
+ WBA I
GalaMe
ordered transition statehas been implicated in an entropically driven binding reaction (ASl* = -160 J mol" K-I). Binding of concanavalin A to saccharides is also accompanied by considerably large ASIS values (28). On the other hand the noticeable differences in the second order rate constants for the binding of GalpMeUmb over GalNAcpMeUmb can be accounted by differences in their activation enthalpies, which is the principal barrier for WBA I-MeUmb galactoside association. Relatively slower binding of GalNAcbMeUmb and GalNAccuMeUmb is presumably due the requirement of a considerable reorientation of water molecules around the acetamido group and/or thecorresponding loci from the protein. Thus, the energetically costly disruption of the hydrogenbonding network of water molecules in the binding cleft of the lectin by the acetamido group accounts for the slower
16387
4-Methylumbelliferyl Galactoside Reaction with WBA I
""_
IT\ L
.
"
n c
FIG. I . Thermodynamic/kinetic profile for the binding of GaVMeUmb (a) and GalNAcDMeUmb to WBA I (b).
"
a
React ion coord i nate
Acknowledgments-We thank Dr. N. B. Joshi for the fluorescence lifetime experiments, Prof. I. J. Goldstein for the kind gift of GalNAcaMeUmb, and Dr. N. V. Joshi for helpful discussions. REFERENCES 1. Goldstein, I. J.,and Poretz, R.D. (1986) in The Lectim: Properties, Functions, and Applications in Siobgy and Medicine (Liener, I. E., Sharon, N., and Goldstein, I. J., eds) pp. 33-247, Academic Press, New York 2. Lis, H.,and Sharon, N. (1986) Annu. Reu. Biochem. 66.35-67 3. Sharon, N.,and Lis, H. (1989) in Lectins, pp. 66-91, Chapman & Hall Ltd., London 4. Goldstein, I. J., and Hayes, C. E. (1978) Adu. Carbohydr. Chem. Biochem. 3s., ~127-340 ." 5. Appukuttan, P. S., and Basu, D. (1981) Anal. Biochem. 1 1 3 , 253-255 6. Kortt, A. A. (1985) Arch. Biochem. Biophys. 236,544-554
1
-
1
c3
association rate constants of GalNAcMeUmb derivatives as compared to the corresponding galactose derivatives. In any event the marked differences in the values of association rate constants for N-acetylgalactosamine derivativesover that observedfor galactosecompoundsunderscore the hitherto unrecognized roleof sugar structures in determining the overall rates of their reactions withlectins (8, 16, 26, 28-30). In conclusion, we have shown that the anomeric configuration of the galactoside residues boundto WBA I leads to a considerable change in the environment of the fluorophore, and that a large hydrophobic aglycon substituent such as 4methylumbelliferylat the anomeric carbon tends to mask the anomericdiscriminatorypowerofWBA I as compared to methyl group or a glycon group such as glucose. Moreover, the presence of acetamido group at C-2 position of galactose not only increases their affinity toward WBA I but also has considerable influenceon the association rate constant.
7T I S-1
7. Matsuda, T., Kabat, E. A., and Surolia, A. (1989) Mol. Immuml. 2 6 , 189195 8. Khan, M. I., Sastry, M. V. K., and Surolia, A. (1986) J. Biol. Chem. 2 6 1 , 3013-3019 9. Loontiens, F. G., Clegg, R. M., and Jovin, T. M. (1977) Biochemistry 1 6 ,
"_ "_
159-1 fifi
10. Lowry, 0.H.,Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Bwl. Chem. 1 9 3 , 265-275 11. Scatchard, G . (1949) Ann. N . Y. Acud. Sci. 5 1 , 6 6 0 4 7 2 12. Chipman, D. M., Grisaro, V., and Sharon, N. (1967) J. Bwl. Chem. 2 4 2 , 43M-4394 ~ ~ . . 13. Bessler, W., Shafer, J. A., and Goldstein, I. J. (1974) J. Biol. Chem. 2 4 9 , ""
2R19-2R22 - -.--
"
14. Hammes, G.G. (1982) Enzyme Catalysis and Regulotion, pp. 179-185, Academic Press, New York 15. Lin, W.-Y., Lin, S. H., Morris, R. J., and Van Wart, H. E. (1988) Biochemistry 27,5068-5074 16. Acharya, S., Patan'ali, S. R ,Saijan, S. U., Gopalakrishnan, B., and Surolia, A. (1990) J . Bioi Chem. 266,11586-11594 17. Dean, B. R., and Homer, R. B. (1973) Biochim. Biophys. Acta 3 2 2 , 141144 18. Decastel, M., Vincent, M., Matta, K. L., and Frenoy, J.-P. (1984) Arch. Bwchem. Biophys. 232,640-653 19. Landschoot, A. V.,Loontiens, F. G., and De Bruyne, C. K. (1978) Eur. J. Biochem. 83,277-285 20. Landschoot, A. V., Loontiens, F. G., and De Bruyne, C. K. (1980) Eur. J. Biochem. 103,307-312 21. Decastel, M., Tran, A.-T., and Frenoy, J.-P. (1982) Biochem. Biophys. Res. Commun. 106,638-643 22. Gu ta, D., Prasad Rao, N. V. S. A. V., Puri, K. D., Matta, K.L., and iurolia, A. (1992) J. Biol. Chem. 267,8909-8918 23. Gray, R. D., and Glew, R. H.(1973) J. Biol. Chem. 248,7547-7551 24. Brewer, C.F., Sterlicht, H., Marcus, D.M., and Grollman, A. (1973) Biochemistry 12,4448-4457 25. Podder, S. K., Surolia, A., and Bachhawat, B. K. (1978) FEES Lett. 8 6 , 313-316 26. Swamy, ~ M J., . Sastry, M.V. K., Khan, M. I., and Surolia, A. (1986) Biochem. J. 234,515-522 27. Neurohr, K. J., Young, N. M., Smith, I. C. P., and Mantsch, H. H.(1981) Biochemistry 20,3499-3504 28. Farina, R. D., and Wilkins, R. G. (1980) Biochim. Biophys. Acta 631,42843R 29. Lewis, S. D., Shafer, J. A., and Goldstein, I. J. (1976) Arch. Biochem. Biophys. 172,689-695 30. Clegg, R. M., Loontiens, F. G., and Jovin, T. M. (1977) Biochemistry 16, ~~
1C7
lcIK
LYI-II"
