Biochimica et Biophysica Acta, 1040(1990) 145-152

145

Elsevier BBAPRO33709

Chemical modification of a fl-glucosidase from Schizophyllum commune: evidence for essential carboxyl groups Anthony John Clarke Guelph-Waterloo Centrefor Graduate Work in Chemistry, Department of Microbiology, Universityof Guelph, Guelph (Canada)

(Received12 March1990)

Key words: fl-Glucosidase;Chemicalmodification;Carbodiimide;EAC;Carboxyl;Catalyticmechanism;(S. commune) The fl-glucosidase from SchizophyUum commune was purified to homogeneity by a modified procedure that employed Con A-Sepharose. The participation of carboxyl groups in the mechanism of action of the enzyme was delineated through kinetic and chemical modification studies. The rates of fl-glucosidase-catalyzed hydrolysis of p-nitropbenyl-flD-glucoside were determined at 27°C and 70 mM ionic strength over the pH range 3.0-8.0. The pH profile gave apparent pK values of 3.3 and 6.9 for the enzyme-substrate complex and 3.3 and 6.6 for the free enzyme. The enzyme is inactivated by Woodward's K reagent and various water-soluble carbodiimides; chemical reagents selective for carboxyl groups. Of these reagents, 1-ethyl-3-(4-azonia-4,4-dimethylpentyl)carbodiimide iodide in the absence of added nucleophile was the most effective and a kinetic analysis of the modification indicated that one molecule of carbodiimide is required to bind to the fl-glucosidase for inactivation. Employing a tritiated derivative of the carbodiimide, 44 carboxyl groups in the enzyme were found to be labelled while the competitive inhibitor deoxynojirimycin protected three residues from modification. Treatment of the enzyme with tetranitromethane resulted in the modification of five tyrosine residues with approx. 28% diminution of enzymic activity. Titration of denatured enzyme with dithiobis(2-nitrobenzoic acid) indicated the absence of free thiol groups. Reaction of the enzyme with diethyl pyrocarbonate resulted in the modification of four histidine residues with the retention of 78% of the original enzymatic activity. The divalent transition metals Cu2+ and Hg 2+ were found to be potent inhibitors of the enzyme, binding in an apparent irreversible manner.

Introduction Saccharification of cellulose, the world's most abundant glucose polymer, is naturally achieved by the synergistic action of a combination of enzymes that principally includes cellulase (1,4-(1,3;1,4)-fl-D-glucan 4glucanohydrolase; EC 3.2.1.4), cellobiohydrolase (1,4-flD-glucan cellobiohydrolase; EC 3.2.1.91) and fl-glucosidase (fl-D-glucoside glucohydrolase; EC 3.2.1.21). The latter enzyme does not act directly on cellulose but it is essential for the efficient hydrolysis of the macromolecule. fl-Glucosidase splits the fl-l,4-glucosidic bond of cellobiose, the end product and inhibitor of cellobiohydrolase, to release two equivalents of glucose and

Abbreviations: CMC, 1-cyclohexyl-3-(2-morpholinoethyl)carbodiimide hydrochlodde;EAC, 1-ethyl-3-(4-azonia-4,4-dimethylpentyl) carbodiimide iodide; EDC, 1-ethyl-3[3-(dimethylamino)propyl]carbodiimide; Mes,4-morpholineethanesulfonicacid. Correspondence: A.J. Clarke, Department of Microbiology,University of Guelph, Guelph,Ontario, Canada,N1G 2W1.

thereby increases the overall rate of cellulose biodegradation [1,2]. The cellulolytic enzymes from the fungi Phanerochaete chrysosporium, Trichoderma reesei, Trichoderma koningii and more recently Schizophyllum commune have been particularly well characterized (for example, see Refs. 3-8). This cellulolytic system of enzymes have attracted considerable attention in recent years in view of their great biotechnological potential. However, while the production, isolation and synergism of the various fungal enzymes have been the subject of intensive investigations, limited information concerning their structure and function relationship is available. Subsite mapping of the fl-glucosidase from T. reesei indicated that the active site of this enzyme comprises three subsites with subsite 1 contributing the greatest proportion of the binding energy [9]. Kinetic and chemical modification studies on the enzyme from the tropical saprophyte Botryodipiodia theobromae have implicated the participation of both a carboxyl and histidine residue at its catalytic site [10,11]. Similar studies indicated the essential role of only carboxyl residues in the active site of

0167-4838/90/$03.50 © 1990 ElsevierSciencePublishers B.V.(BiomedicalDivision)

146 the A. wentii enzyme [12]. Using the affinity label, conduritol-fl-epoxide, the side chain of a specific aspartyl residue in the N-terminal region of this enzyme, Asp-53, was subsequently identified as a catalytic group [131. In contrast to many other fungi, the wood-destroying basidomycete, S. commune is a potent producer of fl-glucosidase [14]. The control, production and isolation of this enzyme has been described in detail [15,16]. Two closely related forms of the fl-glucosidase ( M r 95 700 and 93 800) are secreted by S. commune but they are postulated to arise from the same gene [15]. Clones coding for a partial sequence of the enzyme have been isolated from a cDNA library and comparison of the derived amino acid sequence with that of both a fl-glucosidase from Candida pelliculosa [17] and an active site sequence proposed for the S. commune cellulase [18,19] revealed a region with considerable homology [20]. On this basis, Moranelli et al. [20] postulated that these cellulolytic enzymes share a common catalytic mechanism of action. Furthermore, Glu-160 and Asp-177 of the fl-glucosidase were identified as potential catalytic residues. The present paper describes both kinetic and chemical modification studies that provide experimental evidence for the essential role of carboxyl groups in the mechanism of action of this enzyme.

Materials

and Methods

D-Cellobiose, Con A-Sepharose, 1-cyclohexyl-3-(2morpholinoethyl)carbodiimide hydrochloride (CMC), diethyl pyrocarbonate, 5,5'-dithiobis(2-nitrobenzoic acid), 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide (EDC), D-gluconic acid lactone (glucono-&lactone), Woodward's K reagent (2-ethyl-5-phenylisoxazolium3'-sulfonate), 1-O-methyl-a-D-mannoside, 4-morpholineethanesulfonic acid (Mes) and tetranitromethane were products of Sigma, St. Louis, MO. Iodomethane was purchased fro/n Anachemia Chemicals, Toronto, Ontario, while [3H]iodomethane was a product of Amersham Canada, Oakville, Ontario. p-Nitrophenylfl-o-glucoside and deoxynojirimycin were obtained from Boehringer Mannheim Canada, Dorval, Quebec. BioRad, Richmond CA supplied DEAE-Bio-Gel A, Bio-Gel P-60 and Bio-Gel P-6DG. Ethylenimine, formerly purchased from ICN.K&K Laboratories, Plainsview, NY was carefully re-distilled (b.p. 56°C) and stored at -20°C. This reagent is extremely toxic and consequently it is no longer commercially available. 1-Ethyl3-(4-azonia-4,4-dimethylpentyl)carbodiimide iodomethane (EAC) and its tritiated derivative were synthesized from the free base of EDC and iodomethane by the method of Sheehan et al. [21] and recrystallized from acetone/ether (melting point 94-95°C).

Analytical methods

Concentrations of fl-glucosidase were determined by amino acid analysis [22] assuming Lys = 16 and His = 13 [14] and using a Beckman System Gold Amino Acid Analyzer. Protein samples (0.8-3 nmol) were hydrolysed with 200 /~1 of 5.7 M HC1 in sealed evacuated tubes at 108°C for 24 h. Ultraviolet and visible absorbance measurements were made with a Varian Model 2290 recording spectrophotometer, fl-Glucosidase activity was routinely assayed by following spectrophotometrically the hydrolysis of 0.5 mM p-nitrophenyl-fl-D-glucoside [23] in 0.1 M sodium acetate buffer (pH 5.0 at 27°C). SDS-PAGE was performed with a Tris buffer system, pH 8.8 [24] and protein bands were visualized with Coomassie blue. Radioactivity measurements were made with a Packard Tri-Carb 2000 scintillation counter using Liquiscint scintillation cocktail (National Diagnostics, Manville, N J). Enzyme purification

A freeze dried ethanol precipitate of a 9-day culture of S. commune was kindly provided by M. Paice, Pulp and Paper Research Institute Canada, Pointe Claire, Quebec. The smaller ( M r 93 800) of the two forms of the fl-glucosidase was isolated and purified to electrophoretic homogeneity by a modified method of those previously described [6,14,20] employing a combination of ion-exchange, affinity and size exclusion chromatographies. The fl-glucosidase was separated from the other cellulolytic and xylanase enzymes on DEAE-Bio-Gel A using a linear gradient of 0.30-0.45 M pyridine/acetate buffer (pH 5.1) [14]. Fractions containing enzyme, eluting between 0.38 M and 0.41 M pyridine/acetate, were pooled and lyophilized. The dried material was dissolved in 50 mM sodium acetate buffer (pH 6.0) containing 100 mM sodium chloride and 1 mM each of calcium chloride, manganese chloride and magnesium chloride and applied to a colunm (2.5 × 8 cm) of Con A-Sepharose, previously equilibrated in the same buffer. Elution of the adsorbed fl-glucosidase was achieved with the equilibration buffer containing 300 mM amethyl mannoside. The enzyme preparation, contaminated with the smaller form of the cellulase ( M r 38 000), was desalted by gel filtration on a column (3 × 50 cm) of Bio-Gel P-6DG with water serving as eluent. The two enzymes were separated by size-exclusion chromatography on a column (1.5 × 70 cm) of Bio-Gel P-60, equilibrated and eluted with 10 mM ammonium acetate, pH 5.0. The fl-glucosidase, eluting before the cellulase, was collected and stored as a freeze-dried powder at - 20oc. Kinetic parameters

Initial velocities of fl-glucosidase catalyzed hydrolysis of p-nitrophenyl-fl-D-glucoside were determined by continuously monitoring the release of p-nitrophenol

147 spectrophotometrically at the p-nitrophenol/p-nitrophenoxide isobestic point, 347.5 nm. At this wavelength, the extinction coefficient was determined to be 5600 M -1. cm -1. The reactions were initiated by addition of enzyme (13.7 nM, final concentration) to a solution of substrate (0.040-0.50 mM) in buffer which had been thermally equilibrated to 27°C. Buffers employed were: 50 mM sodium formate for pH 3.0-3.6; 50 mM sodium acetate for pH 3.6-6.0; and 25 mM sodium phosphate for pH 6.0-8.0, each containing 2 mM EDTA (/~ = 70 mM with potassium chloride). Units of activity are expressed in katals (kat) where 1 kat is defined as 1 mole of glucose released per second. The kinetic parameters for the enzymatic reaction under initial velocity (zero-order) conditions (K m, Vm~,) were determined from computerized least-squares fits of the Lineweaver-Burk equation to the data.

Chemical modifications Modification of the carboxyl groups of Asp/Glu residues was performed using a variety of group-selective reagents that included carbodiimides, Woodward's reagent and ethylenimine. The modification of the enzyme by ethylenimine was performed according to the procedure of Yamada et al. [25]. fl-Glucosidase (2.14 /~M) in 100 mM Mes (pH 5.5) was incubated with ethylenimine (0.26 M) for 24 h at 27°C. Aliquots of the reaction mixture were periodically removed and residual hydrolytic activity was determined using 0.5 mM pnitrophenyl-fl-D-glucoside at pH 5.5 as substrate. Excess reagent was removed by chromatography on BioGel P-6DG with water serving as eluent. In experiments employing carbodiimides and Woodward's K reagent, fl-glucosidase (1.3/tM) in 100 mM Mes buffer (pH 5.5 or 6.0) was incubated at 27°C for 60 rain with EAC (50 mM), EDC (50 mM) or Woodward's K reagent (45 mM), either in the absence of added nucleophile or in the presence of methylamine (20 mM), glycine ethyl ester (50 mM) or glucosamine (100 mM). At the indicated times, 20/~1 aliquots of the various mixtures were withdrawn and added to 80 /~1 of 100 mM sodium acetate buffer (pH 4.5) to quench residual reagent. The remaining activity of the diluted enzyme derivatives was determined and expressed as a percentage of an appropriate control enzyme solution that was treated in a similar manner except for the addition of reagent. The kinetics of the EAC inactivation of the fl-glucosidase were investigated at pH 6.0 over a range of reagent concentrations (10-200 mM) and pseudo-firstorder rate-constants were determined from semi-logarithmic plots of residual activity against time. In a set of parallel experiments, samples of the enzyme were treated with EAC (34 mM) after prior incubation with deoxynojirimycin (0.248-1.73 mM). Inactivation of fl-glucosidase in both the absence and presence of de-

oxynojirimycin (1.9 mM) with [3H]EAC was also performed on a semi-preparative scale by incubating 0.4 ml samples of the enzyme (8.4 /~M) in 50 mM Mes (pH 5.01) with 180 mM [3H]EAC. Following incubation at ambient temperature (23°C) for 120 min, reactions were quenched with the addition of 0.2 M sodium acetate buffer (pH 5.0). Excess reagent and by-products were removed by chromatography on Bio-Gel P6-DG using 10 mM ammonium bicarbonate buffer (pH 7.9) as eluent at a flow rate of 30 ml. h -1. Fractions (1.1 ml) containing protein were pooled and assayed for both fl-glucosidase and [3H]EAC concentrations. Nitration of tyrosine residues was performed employing tetranitromethane according to the method of Sokolovsky et al. [26]. Tetranitromethane (21 mM) in 95% ethanol was added to 1.0 ml samples of the enzyme (4.3 /~M) in 100 mM Tris-HCl (pH 8.0) to give the desired molar excess of reagent over enzyme and incubated at 30°C for 5 h. The number of modified tyrosine residues was continuously estimated spectrophotometrically at 428 nm using the molar extinction coefficient of 4100 M - 1 - cm -1 for 3-nitrotyrosine. At appropriate time intervals, 10/tl of the reaction mixture was withdrawn and added to 100 ttl of 100 mM sodium acetate buffer (pH 4.5). Enzyme samples treated in a similar manner except for the addition of tetranitromethane served as controls. The residual activity of the diluted enzyme samples was determined and expressed as a percentage of a control. The number of free sulfhydryls present in the enzyme was assessed by titration with dithiobis(2-nitrobenzioc acid) according to the method of Ellman [27]. Samples of fl-glucosidase (7.85/xM, 1.0 ml) in 50 mM Tris-HCl buffer (pH 8.0) (with and without 8 M urea) were pipetted into a glass cuvette to which dithiobis(2-nitrobenzoic acid) (1 mM, final concentration) was added. The change in absorbance at 412 nm was continuously monitored at 25°C for 3 h and the number of cysteine residues was estimated using the c4~2 value of 1.36.104 M -1- cm -~ for the released 5-thio-2-nitrobenzoic acid. The participation of histidine residues in the mechanism of action of the enzyme was investigated by treatment with diethyl pyrocarbonate. A 1 ml sample of the /3-glucosidase (3.2 #M) was transferred to a quartz cuvette and diethyl pyrocarbonate in ethanol was added to a final concentration of 27 mM. The conversion of histidine residues to N-carboethoxy derivatives was continuously monitored by following the increase in A242. The number of modified histidine residues was calculated using the molar extinction coefficient of 3200 M -~. cm -~ for the N-ethoxycarbonyl histidyl derivative [28]. The remaining activity of the modified enzyme was periodically determined after dilution of 10 /~1 aliquots with 100 #1 of 100 mM sodium acetate buffer (pH 4.5).

148

Metal-ion inactivation of fl-glucosidase The ability of various divalent metal ions to inactivate fl-glucosidase was tested by incubating samples of the enzyme (1.2/~M) in 50 mM sodium acetate buffer (pH 5.0) with the chloride salts of the metals, the final concentration of which was adjusted to achieve the desired molar excess over that of the enzyme. The progress of the reactions was periodically assessed by the dilution of an aliquot (10/~1) of the reaction mixture with 0.5 ml of 50 mM sodium acetate buffer (pH 5.0) and determining residual activity.

where Kal and K~2 are the dissociation constants for essential ionizable groups in the enzyme and (Vmax)m is the maximum rate when the enzyme is in its optimal ionized form. The data of the log Vmax versus pH plot (Fig. la) fit reasonably well with the curve calculated using Kat and Ka2 values of 5.01 • 10 -4 and 1.26.10 -7, respectively, thus providing pKes values of 3.3 and 6.9 for groups in the enzyme-substrate complex. Values of pK~ were suggested from the log Vm~x/Km versus pH plot (Fig. lb) to be 3.3 and 6.6 for groups on the free enzyme.

Results

Chemical modifications In order to identify the essential ionizable groups in the S. commune fl-glucosidase, chemical reagents selective for the various functional groups of the different amino acids were employed. The participation of carboxyl groups in the mechanism of catalysis was investigated using a variety of reagents including water soluble carbodiimides, Woodward's K reagent and ethylenimine. Treatment of the fl-glucosidase with 260 mM ethylenimine resulted in a progressive inactivation such that after 24 h of incubation at 27°C, the enzyme derivative retained 42% of its original activity. Following the removal of excess reagent by gel filtration, a kinetic analysis of the modified enzyme derivative revealed a change in both K m and kcat for p-nitrophenyl/3-D-glucoside at pH 5.0 from 0.58 mM and 60.2 s -1 (for the native state) to 1.2 mM and 29 s-1, respectively. Incubation of the enzyme with the carbodiimides or Woodward's K reagent in the absence or presence of different nucleophiles for 1 h resulted in a progressive loss of catalytic activity (see Fig. 2 for representative experiments). In most cases, the extent of inactivation that was achieved ranged between 45% and 70%. However, under similar conditions, the water soluble carbodiimide EAC in the absence of added nucleophile proved to be the most effective of the reagents tested. Consequently, further detailed studies of carboxyl group modification focused only on the interaction of EAC with the enzyme. The kinetics of the inactivation of the/3-glucosidase by EAC were analyzed by reaction of the enzyme for 1 h at 27°C with different concentrations of the carbodiimide. Semi-logarithmic plots of residual activity as a function of time were biphasic (Fig. 3), indicating a complexity of the EAC-induced inactivation. However, the course of this inhibition can be resolved into two first-order processes with the slope at longer times determining the rate of inactivation. A plot of the apparent-pseudo-first-order rate constants, determined from the second phase, as a function of EAC concentration was linear (not shown) and provided a second-order rate constant of approx. 2.5 • 10 -3 M -1 • s -1. Analysis of the order of the inactivation with respect to EAC concentration by the method of Levy et al. [30] yielded

Kinetics: p H dependence Initial velocities for fl-glucosidase catalyzed hydrolysis of p-nitrophenyl-fl-o-glucoside were determined as the average of at least three measurements at each substrate concentration and pH. The dependence of initial velocity upon substrate concentration was hyperbolic at each pH studied and all Lineweaver-Burk plots were linear. Values for V ~ and K m were obtained for each pH and are presented graphically in Fig. 1. These semi-log (Dixon [29]) plots indicate the dependence of catalytic activity on the ionization of at least two essential groups at the active site of the enzyme. The lines drawn through the data are the theoretical curves generated using the equation: Vmax =

(V,n.,)m [H + ] Ka2 1+--R-~1 + [H +]

-4.00

-5.00

-6.00

7 :m

B -4.50

o, o

J

-5.50 2

3

4

5 6 pH

7

8

9

Fig. 1. Dependence of kinetic parameters on pH for the fl-glucosidase-catalyzed hydrolysis ofp-nitrophenyl-fl-D-glucoside. Values of K m and Vmax were obtained from Lineweaver-Burk plots of initial rates of hydrolysis determined with five substrate concentrations (0.040-0.50 raM). (A) Effect of pH on log Vmax; (B) effect of pH on log V m ~ / K m. The solid lines represent the theoretical curve calculated using pK a values of 3.3 and 6.9 for (A) and 3.3 and 6.6 for (13). Buffers employed were: (O) 50 mM sodium formate; (A) 50 mM sodium acetate; (ll) 25 mM sodium phosphate; each containing 2 mM EDTA (# = 70 mM with potassium phosphate).

149 100

>.,

.

.

.

.

.

2.00

8O

i

1.8o

1.60

.-2_ 4O rY o~ 20

1.40

~ o

0

'

0

'

10 20 Time (rain)

30

Fig. 2. Inactivation of fl-glucosidase by chemical reagents selective for carboxyl groups. Enzyme (1.3 #M) in 100 mM Mes (pH 5.5) was incubated at 27°C with : Woodward's K reagent (45 mM) in the absence (11) and presence of 20 mM methylamine (ra); EDC (50 mM) in the absence (A) and presence of 20 mM methylamine (zx); CMC (50 mM) and 20 mM methylamine (~); EAC (50 mM) in the absence of added nucleophile (O).

a slope of 0.93 (Fig. 3, inset), indicating that one molecule of EAC binds to one molecule of enzyme when inactivation occurs. A partial-EAC-modified enzyme derivative which retained 30% of its original catalytic activity was prepared on a semi-preparative scale. Chromatography of the reaction mixture on Bio-Gel P 6 - D G affected a clear separation of the modified enzyme derivative from excess reagent and by-products. The kinetic parameters, 2.00

"> ~:

1.50 *'Ill ~

~ a

"-''-'-a

1.00 C12

o-, o

0.50

[~ (JO

Log [EAC]

,2.5 ,

,

0.00 0

20

40

3.

60

Time (min) Fig. 3. Inactivation of fl-glucosidase by the water-soluble carbodiimide, EAC. Enzyme (1.3/~M) in 100 mM Mes, pH 6.0 was treated with EAC at 27°C. At the indicated times, the reactions were quenched with the addition of 100 mM sodium acetate buffer (pH 4.5) and residual enzymic activity was determined using p-nitrophenyl-fl-Dglucoside as substrate. EAC concentrations were: (O) 10 mM; (A) 20 mM; (ll) 50 raM; ( , ) 100 raM; (v) 200 mM. Inset: Apparent order of reaction with respect to reagent concentration. The pseudo-first-order rate constants (k') were calculated from the slopes of the data at longer times in the figure.

-k'

!

'o/[

1.20

i

20

40

Time

(min)

,

L

60

Fig. 4. Effect of a competitive inhibitor, nojirimycin, on the inactivation of fl-glucosidase by EAC. Enzyme (1.3/~M) in 100 mM Mes (pH 5.25) was treated with EAC (34 mM) in the absence (o) and presence of 0.248 mM (O), 0.496 mM (0), 1.24 mM (11) or 1.73 mM (A) deoxynojirimycin. At the indicated times, residual activity of quenched aliquots was determined. Inset: Protection of EAC inactivation of fl-glucosidase by deoxynojirimycin. The apparent pseudo-first-order rate constants obtained in the absence (k'a) and in the presence (k~) of the ligand deoxynojirimycin (L) were plotted according to the method of Scrutton and Utter [31].

g m and turnover number (kca t) were obtained frbm this

derivative from Lineweaver-Burk plots of the data. A value calculated for the Michaelis constant g m of the modified enzyme for p-nitrophenyl-fl-D-glucoside at p H 5.0 was 0.42 mM, which is comparable to that of the native enzyme under similar conditions of p H and temperature, viz. 0.58 mM. On the other hand, the kca t of the modified enzyme for this substrate was 0.29 s-1, a value two orders of magnitude lower than that of the native enzyme (60.2 s - l ) , indicating that catalytic hydrolysis and not substrate binding is affected by the chemical modification. In order to investigate the specificity of the EAC inactivation of fl-glucosidase, experiments were conducted in which the enzyme was pre-incubated with a known competitive inhibitor, deoxynojirimycin, prior to reagent addition. The inclusion of deoxynojirimycin into the reaction medium appeared to decrease the rate of activity loss (Fig. 4). Thus, after incubation with 34 m M EAC for 60 min in the absence of added ligand, the enzyme retained approx. 30% of its catalytic activity, whereas greater than 90% residual activity remained in enzyme derivative modified in the presence of 1.73 m M deoxynojirimycin. Protection afforded by this ligand was concentration dependent, and a plot of inhibitor concentration against protection [31] extrapolated to the origin (Fig. 4, inset) suggesting that EAC inactivation of a true fl-glucosidase-deoxynojirimycin complex is not possible. The stoichiometry of the EAC modification of fl-glucosidase at p H 5.0 was assessed employing [3H]EAC to

150 facilitate quantitation. Following removal of excess reagent by chromatography on Bio-Gel P6-DG, 44.7 + 0.79 residues of a total 184 Asx/Glx residues [14] were modified for complete inactivation of the enzyme. With the inclusion of the protective ligand, deoxynojirimycin, in the reaction medium, 41.9 + 0.65 equivalents of EAC were found to bind to the enzyme complex which retained 32% of its original hydrolytic activity. This suggests that approximately three residues are protected by this competitive inhibitor from modification by EAC. While carbodiimides are highly selective toward the modification of carboxyl residues in proteins under the conditions employed [32,33], the modification of cysteine and tyrosine residues at slightly acidic pH has also been documented in the literature [33,34]. Amino acid analysis of a sample of fl-glucosidase inactivated to 93% by EAC did not reveal any appreciable loss of tyrosine suggesting that an acid stable O-arylisourea derivative of tyrosine was not generated (data not shown). Moreover, modification of the enzyme with the sensitive tyrosyl reagent, tetranitromethane, reduced the initial enzymatic activity by only 28% confirming the absence of accessible catalytically-essential tyrosine residue(s) at the active site. Titration of the enzyme (7.85 #M) in 50 mM Tris-HC1 buffer (pH 8.0) with dithiobis(2-nitrobenzoic acid) in both the absence and presence of 8 M urea did not result in any significant release of 5-thio-2-nitrobenzoic acid, suggesting the absence of cysteine residues. Modification of fl-glucosidase with an 8300-fold molar excess of diethyl pyrocarbonate for 1 h resulted in the conversion of 4 of the 13 histidine residues present in the enzyme to the corresponding Ncarboethoxy derivatives and loss of only 22% of the original activity (data not shown). Consequently, it appears unlikely that the essential ionizable group(s) exhibiting basic pK a values between 6.6 and 6.9 may be ascribed to a histidine residue.

TABLE I

Inhibition of S. commune fl-glucosidase by selected divalent cations Cation

Mn 2 + Fe 2+ Co 2 + Cu 2+ Zn 2+ Hg 2+

Concn. a

% Residual activity after

(mM)

10 min

120 rain

25 12 25 25 0.013 25 25 0.013

98 99 96 0.42 17 97 0.031 5.4

87 81 63 0.10 0.092 63 0.001 0.084

a Values reported are for concentrations of metals in incubation solutions (see Materials and Methods). Final concentration of metals in enzyme assay solutions after dilution were: Concn. ( m M ) × 2 . 10 -3 '

The effect of selected divalent cations on the S. fl-glucosidase activity is shown in Table I. The enzyme appeared to be relatively stable in the presence of Mg 2+, Ca 2+, Fe 2+ and Mn 2+, retaining greater than 80% of its original activity after the 120 min incubation period. The transition metals Zn 2÷ and Co2÷, however, were partially inhibitory whereas both Cu 2÷ and Hg 2+ were very potent inhibitors of fl-glucosidase activity. The final concentration of these latter metals in the assay mixtures was 25 nM suggesting very high affinity of the enzyme for the metals. In a subsequent experiment, the catalytic activity of the enzyme was monitored over a 60 min time period during its reaction with Cu 2÷ and Hg 2+. The loss of activity was observed to be dependent upon time (results not shown) indicating an apparent irreversible inactivation of the fl-glucosidase. commune

Discussion

The nature of the catalytically-essential ionizable groups in the S. c o m m u n e fl-glucosidase was delineated by chemical modification studies. These experiments indicated the role of carboxyl groups in the mechanism of fl-glucosidase action, while clearly discounting the essential participation of histidyl, tyrosine and cysteine residues. The partial loss of activity observed with the enzyme derivatives with modified histidyl or tyrosine residues probably reflects small conformational changes in the enzyme induced by the respective modifications. Of the water-soluble carbodiimides examined, EAC in the absence of added nucleophiles proved to be the most effective reagent by inducing a rapid and complete inactivation of the enzyme. According to a recently proposed mechanism [35], a protonated carboxyl side chain of the enzyme would catalyze its own modification by a carbodiimide to generate an O-acylisourea adduct. With EAC, the O-acylisourea formed is apparently not susceptible to nucleophilic attack by either water or added compounds, since it readily undergoes intramolecular rearrangement to the stable N-acylurea derivative [36]. This, combined with the fact that unlike EDC, EAC cannot undergo the stabilizing tautomerization to an unreactive cyclic form [37], would account for the enhanced reactivity of EAC as a modification reagent. Although the kinetics of the reaction of EAC with the fl-glucosidase indicated that the modification of one carboxyl group is required to abolish catalytic activity, differential labelling of the enzyme with [3H]EAC in the presence and absence of deoxynojirimycin resulted in the protection of three carboxylic residues. It is conceivable that two of these residues constitute the proposed catalytic diad, viz. Glu 160 and Asp 177, with the modification of only one being sufficient for inactivation. The third protected residue probably participates in substrate binding.

151 At present, it is u n k n o w n how m a n y of the 109 Asx and 75 Glx residues observed in the enzyme [14] exist as A s p and Glu residues but it is a p p a r e n t that a large n u m b e r of them are exposed and readily react with EAC. The biphasic nature of the time course of the inactivation clearly suggests that modification of some of these p r o d u c e a less-active derivative of the enzyme. Modification of these latter residues however, which accounts for between 40% and 60% of activity loss, does not seem to impair substrate binding as indicated by the kinetic parameters of a partially modified enzyme derivative. As these readily modified carboxyl groups must be non-catalytic, it follows that their amidation presumably perturbs the catalytic residues through conformational a n d / o r steric alterations. The p h e n o m e n o n of biphasic inactivation kinetics is not u n c o m m o n a m o n g reports of carbodiimide modification o f a variety of enzymes including a n u m b e r of other carbohydrases [38-43]. The apparent irreversible binding of Cu 2÷ or H g 2÷ to the fl-glucosidase and subsequent inactivation of the e n z y m e would suggest that the metals are tightly coordinated either to or near to catalytic residues of the active site. Interestingly, an identical situation has been observed for the interaction between these same metals and the S. c o m m u n e cellulase [44]. In this case, the metal inactivation of the cellulase was p r o p o s e d to proceed by chelation involving carboxyl groups at the active centre of the enzyme in a m a n n e r analogous to that observed by b o t h X-ray crystallography and N M R for metal-hen egg-white lysozyme complexes [45,46 and references therein]. This m a y also occur with the fl-glucosidase in view of b o t h the current observations concerning the chemical modification experiments and the a m i n o acid sequence similarity previously observed between a segment of the enzyme comprising residues Glu-160 and Asp-177 and a catalytic sequence p r o p o s e d for the S. c o m m u n e cellulase [20].

Acknowledgements I thank Mr. Craig Christmann for his excellent technical assistance, These studies were supported by operating grants from b o t h the N a t i o n a l Research and Engineering Council of C a n a d a and the University of G u e l p h and a s u m m e r assistantship to C.C. for the Ontario Ministry of Energy.

References 1. Sternberg, D. (1976) Appl. Environ. Microbiol. 31, 648-654. 2. Sternberg, D., Vijayakumar, P. and Reese, R.T. (1977) Can. J. Microbiol. 23, 139-147. 3. Enari, T.-M. and Niku-Paavola, M.-L. (1987) CRC Crit. Rev. Biotechnol. 5, 67-87. 4. Sadana, J.C. and Patil, R.V. (1985) Carbohydr. Res. 140, 111-120.

5. Henrissat, B., Driquez, W., Viet, C. and Schiilein, M. (1985) Bio/technology 3, 722-726. 6. Willick, G.E., Morosoli, R., Seligy, V.L., Yaguchi, M. and Desrochers, M. (1984) J. Bacteriol. 159, 294-299. 7. Montenecourt, B.S. (1983) Trends Bioteclmol. 1, 157-161. 8. Lamed, R., Setter, E., Kenig, R. and Bayer, E.A. (1980) Biotechnol. Bioeng. Syrup. No. 13, 163-181. 9. Chirico, W.J. and Brown, Jr., R.D. (1987) Eur. J. Biochem. 165, 333-341. 10. Umezurike, G.M. (1987) Biochem. J. 241,455-462. 11. Umezurike, G.M. (1977) Biochem. J. 167, 831-833. 12. Legler, G. and Gilles, H. (1970) Hoppe-Seyler's Z. Physiol. Chem. 351,741-748. 13. Bause, E. and Legler, G. (1974) Hoppe-Seyler's Z. Physiol. Chem. 355, 438-442. 14. Desrochers, M., Jurasek, L. and Paice, M.G. (1981) Appl. Environ. Microbiol. 41,222-228, 15. Willick, G.E. and Seligy, V.L. (1985) Eur. J. Biochem. 151, 89-96. 16. Wilson, R.W. and Niederpruem, D.J. (1967) Can. J. Microbiol. 13, 1009-1019. 17. Kohchi, C. and Toh-e, A. (1985) Nucleic Acids Research 13, 6273-6282. 18. Yaguchi, M., Roy, C., Rollin, C.F., Paice, M.G. and Jurasek, L. (1983) Biochem. Biophys. Res. Commun. 116, 408-411. 19. Paice, M.G., Desrochers, M., Rho, D., Jurasek, L., Roy, C., Rollin, C.F., DeMiguel, E. and Yaguchi, M. (1984) Bio/technology 2, 535-539. 20. Moranelli, F., Barbier, J.R., Dove, M.J., MacKay, R.M. and Seligy, V.L. (1986) Biochem. Int. 12, 905-912. 21. Sheehan, J.C., Crucikshank, P.A. and Boshart, G.L. (1961) J. Org. Chem. 26, 2525-2528. 22. Moore, S. and Stein, W.H. (1963) Methods Enzymol. 6, 810-831. 23. Nisizawa, T., Suzuki, H. and Nisizawa, K. (1971) J. Biochem. 70, 387-393. 24. Weber, K. and Osborn, M. (1975) in The Proteins (Neurath, H., Hill, R.L. and Boeder, C. eds) vol. 1, pp. 179-233, Academic Press, New York. 25. Yamada, H., Imoto, T. and Noshita, S. (1982) Biochemistry 21, 2187-2192. 26. Sokolovsky, M., Riordan, J.F. and Vallee, B.L. (1966) Biochemistry 5, 3582-3589. 27. Ellman, G.L. (1959) Arch. Biochem. Biophys. 82, 70-77. 28. Muhlrad, A., Hegyi, G. and Horanyi, M. (1969) Biochim. Biophys. Acta 181, 184-190. 29. Dixon, M. and Webb, E.G. (1979) Enzymes, 3rd edn. pp. 138-164, Academic Press, New York. 30. Levy, H.M., Leber, P.D. and Ryan, E.M. (1963) J. Biol. Chem. 238, 3654-3659. 31. Scrutton, M. and Utter, M. (1965) J. Biol. Chem. 240: 3714-3723. 32. Hoare, D.G. and Koshland Jr., D.E. (1967) J. Biol. Chem. 242, 2447-2453. 33. Carraway, K.U and Koshland Jr., D.E. (1972) Methods Enzymol. 25, 616-623. 34. Carraway, K.U and Koshland Jr., D.E. (1968) Biochim. Biophys. Acta 160, 272-274. 35. Chan, V.W.F., JOrgensen, A.M. and Borders, Jr., C.L. (1988) Biochem. Biophys. Res. Commun. 151,709-716. 36. Timkovich, R. (1977) Anal. Biochem. 79, 135-143. 37. Tenforde, T., Fawwaz, R.A. and Freeman, N.K. (1972) J. Org. Chem. 37, 3372-3374. 38. Svensson, B., Moiler, H. and Clarke, A.J. (1988) Carlsberg Res. Commun. 53, 331-342. 39. Koyama, T., Inokuchi, N., Iwama, M. and Irie, M. (1985) J. Biochem. 97, 633-641. 40. Inokuchi, N., Iwama, M., Takahashi, T. and Irie, M. (1982) J. Biochem. 91, 125-133.

152 41. Millet, F., Darley-Usmar, V. and Capaldi, R.A. (1982) Biochemistry 21, 3857-3862. 42. Terra, W.R., Terra, I.C.M., Ferreira, C. and De Bianchi, A.G. (1979) Biochirn. Biophys. Acta 571, 79-85. 43. Gray, C.J. and Jolley, M.E. (1973) FEBS Letts. 29, 197-200. 44. Clarke, A.J. and Adams, L.S. (1987) Biochim. Biophys. Acta 916, 213-219.

45. Perkins, S.J., Johnson, L.N., Machin, P.A. and Phillips, D.C. (1979) Biochem. J. 181, 21-36. 46. Dobson, C.M. and Williams, R.J.P. (1977) in Metal-Ligand Interactions in Organic Chemistry and Biochemistry (PuUam, B. and Goldblum, N. eds.) Part 1, pp. 255-282, D. Reidel Publishing Co., Dordrecht, Holland.