ANALYTICAL

207,321-328

BIOCHEMISTRY

(19%)

Chromogenic Redox Assay for ,&Lactamases Water-Insoluble Products

Yielding

I. Kinetic Behavior and Redox Chemistry Christopher

Bieniarz,

Diagnostics

Received

Division,

May

’ Douglas

Department

F. Young,

of Immunochemistry,

and Michael Abbott

@-Lactamases are major determinants of bacterial resistance to J?-lactam antibiotics (1,2). These enzymes catalyze the hydrolysis of the @-lactam ring of several antibiotics, complicating the therapeutic course of action in the treatment of bacterial infections. There is

oao3-2697/92

Copyright All rights

Abbott

Park, North

Chicago, Illinois

60064

26, 1992

We describe a chromogenic detection system for /3lactamase which yields water-insoluble colored products. The assay is based on kinetic measurement of the appearance of color due to the &lactamase-initiated redox reaction. The substrates are C3’ thiolate-substituted cephalosporins, which, after enzyme-catalyzed hydrolysis of the &lactam ring, undergo elimination of the thiolate ion. This thiolate, in a postenzymatic step, reduces the tetrazolium salts, which are water-soluble colorless compounds, to a colored water-insoluble precipitate of formazan. Our model in this study was a &lactamase Enterobacter cloacae P-99~catalyzed reaction of thiolacetate cephalosporin with several tetrazolium salts. We found that the reaction rate is dependent on the concentration of the electron carrier 5-methyl phenazinium methyl sulfate, the p&, of the C3’ thiolate substituent of the cephalosporin substrate, and the reduction potential of the tetrazolium salts. A kinetic study of this system yielded a rate law for the reaction. We present a mechanism of the reaction and determination of the kinetic parameters for the process. The sensitivity of this kinetic assay is very high; we detect 3 X 10-l’ M &lactamase P-99, which is approximately 30 mIU. The assay times are very short, lasting from 2 to 5 min. The new assay system is particularly suitable for a rapid detection of &lactamases in bacterial colonies and in enzyme immunoassays where &lactamase may be used as the label. @ ioo2 Academic PEWS, 1~.

i To whom

J. Cornwell

Laboratories,

correspondence

should

be addressed.

$5.00

G 1992 by Academic Press, Inc. of reproduction in any form reserved.

therefore a continuing interest in development of methods for the detection of B-lactamases in biological media. Several methods of detection of /I-lactamases have been described in the literature (3). In particular, colorimetric methods are most convenient, and several chromogenic substrates for these enzymes have been described in recent years (4-6). Virtually all of these methods rely on electronic spectral changes accompanying the enzyme-catalyzed cleavage of the /I-lactam ring and the subsequent elimination of the chromogen from the C3’position of the cephalosporins. Alternatively, coordination of the hydrolyzed p-lactam to a suitable transition metal and the consequent change in color has been exploited (7). To our knowledge, there have been no reports of colorimetric measurements of @-lactamases based on chromogenic substrates yielding colored precipitates. We report here a novel method of measuring p-lactamase activity based on the principle alluded to above. The method consists of exposing to the action of /I-lactamase certain cephalosporins suitably substituted at C3’ atom with a reducing functionality. This is carried out in the presence of an appropriate tetrazolium salt and electron carrier. Subsequent to the enzyme-catalyzed ,&lactam ring cleavage, the reducing substituent at C3’ is eliminated, a redox process during which the colorless, water-soluble tetrazolium salt is reduced to a highly chromogenic, water-insoluble formazan occurs. This process may be accelerated by the use of electron carriers, e.g., 5-methylphenazinium methyl sulfate (PMS)’ or 5-ethylphenazinium ethyl sulfate (PES). * Abbreviations used PMS, 5-methylphenazinium methyl sulfate; PMSH, reduced 5-methylphenazinium methyl sulfate; PES, 5-ethylphenazinium ethyl sulfate; TAC, 7-thiophenylacetamido-3-thioacetoxymethyl-3-cephem-4-carboxylate sodium salt INT, 2-p-iodophenyl-3-p-nitrophenyl-5-phenyl-2FZ-tetrazolium chloride; MTT, 3-(4,5-~methyl-2-thiazolyl)-2,5-~phenyl-2~-tetrazolium bromide; 321

322

BIENIARZ,

YOUNG,

This new assay for &lactamases is particularly suited to the estimation of /3-lactamase activity on a solid-phase, bibulous material, i.e., filter paper, cellulose, or other immunochromatographic media. While this work addresses the mechanistic aspects of the process summarized above, the accompanying paper describes the adaptation of the underlying principle described in this work to enzyme immunoassays. MATERIALS

AND

METHODS

Instrumentation Electronic spectra were recorded on a Hewlett-Packard 8452A diode array spectrophotometer. ‘H magnetic resonance spectra were obtained on a Magnachem A200 instrument. Infrared spectra were obtained on a Perkin-Elmer 298 ir infrared spectrophotometer. Materials Thiolacetic acid and PMS were purchased from Aldrich (Milwaukee, WI). Cephalothin, INT, MTT, and NBT were obtained from Sigma Chemical Co. (St. Louis, MO) and were used as received. /3-Lactamases Type III from Enterobacter cloacae P-99 and RTEM-2 from Escherichia coli were obtained from Porton Products, Ltd., CAMR (Porton Down, UK SP4 OJG). ,&Lactamase Type I from Bacillus cereus was purchased from Sigma. BSA Abbott Code 96363 was from Abbott Laboratories (North Chicago, IL). All buffer components were from Sigma. Silica gel 60 was Merck 70-230 mesh and was purchased from Aldrich. Synthesis of 7-Thiophenylacetamido-3thioacetoxymethyl-3-cephem4carboxylate Salt (TAC)

Sodium

Thiolacetic acid, 0.776 g (0.0102 mol), and sodium bicarbonate, 0.857 g (0.0102 mol), were dissolved in 30 ml of water and warmed to 5O’C. The warm solution was filtered through silica gel and added to a solution of cephalothin, 1.43 g (0.0034 mol), in 10 ml of water. The reaction mixture was stirred for 31 h at 5O’C. After cooling to room temperature, the water was removed by vacuum, and the residue was dissolved in methyl alcohol. Crude product was precipitated by addition of diethyl ether, and the solid was triturated in ethyl acetate. Precipitation and trituration were repeated until pure 7thiophenylacetamido-3-thioacetoxymethyl-3-cephem4-carboxylate sodium salt, 1.16 g (0.0027 mol), was obtained mp 207-210°C; ‘H NMR (200 MHz, DZO, d) 7.30 (lH, t), 7.00 (2H, d), 5.60 (lH, d), 5.05 (lH, d), 3.90 (2H, NBT, 3,3’-(3,~-~methoxy-4,4’-biphenylylene)bis(2-~-nitrophenyl-5phenyl-2Zf-tetrazolium chloride); BSA, bovine serum albumin; mIU, milli-International Unit (1 U will hydrolize 1.0 pmol of cephaloridine per minute at pH 7.0 at 25°C).

AND

CORNWELL

s), 3.85 (2H, dd), 3.45 (2H, dd), 2.30 (3H, s); ir (KBr, cm-‘) 3290, 1750,1650, 1600, 1530, 1400,1355.

Determination of the Dissociation Constant Kdtia of 3Thioacetylmethyl Cephalosporoate ABHThe dissociation constant of ABH- to A- and B-, shown in Fig. 6, was determined by ‘H NMR spectroscopy, in a 0.20 M phosphate-buffered DZO solution at 25’C at pD 7.0. The ratio of bound to unbound thiolacetate B- at equilibrium was determined by measuring the ratio of ‘H NMR signal integrations at 8 2.29 and 6 2.37 for bound and unbound thiolacetate. TAC, 1.99 x lop3 g (4.59 X lop5 mol), was dissolved in 2.00 ml of phosphate-buffered DzO and 0.100 ml of 1 X lop5 M P-99 &lactamase was added. The solution was allowed to equilibrate. The dissociation constant was established by the expression Kdiss = [A-]. K=. [ABH&]/[ABH-] . (1 + Ka/[H+]), where Ka is the acid dissociation constant of thiolacetic acid and ABHLt is the total concentration of 3-thioacetylmethyl cephalosporoate ABH-. Kinetic Measurements @-Lactamase Type III from E. cloacae P-99 was used in the kinetic study. This enzyme had activity of 600 U/mg of protein with cephaloridine as substrate. All electronic spectral changes were measured in buffered pH 7.0 solutions in thermostated cells at 25.O“C under aerobic conditions. Action spectra were acquired by periodic scanning of the appropriate regions of the electronic spectra. All kinetic measurements were done at enzyme concentrations of 3.2 X 10m6 or 3.2 X lo-’ M. At these concentrations the rate of the redox reaction is independent of the enzyme concentration. PMS, 3.23 X 10m5 M, was used in all kinetic runs as electron carrier (8). All kinetic runs were performed in the presence of a minimal, 0.3% (w/v), concentration of BSA in order to facilitate the colloid formation of the water-insoluble formazans (9) and thus allow kinetic study of the spectral changes. The initial rates were determined by measuring the tangents to the absorbance vs time curves during the first 5% of the reaction progress. The limits of linearity were established and measurements of initial rates were done within a time span such that the velocity was linearly related to the enzyme concentration. Only measurements showing correlation coefficients greater than 0.990 were considered. The kinetic measurements were done at absorbance maxima of the respective formazans-INT, 500 nm; NBT, 530 nm; MTT, 564 nm-by measuring the rate of the appearance of the formazans. In each case, the reactions were initiated by addition of the appropriate amounts of stock enzyme solution to a magnetically stirred solution of /3-lactam substrate and tetrazolium salt in 0.1 M phosphate buffer, BSA and PMS in a l-cm path quartz cuvette. The rates of the blank reactions of BSA in ab-

CHROMOGENIC

REDOX

ASSAY

sence of enzyme were substracted from the ,&lactamase-catalyzed reactions. The value of kz, apparent second-order rate constant for the reduction of tetrazolium salt MTT by thiolacetate anion released, was obtained from the values of Vmm and K,,, , derived from the direct nonlinear fit of the Michaelis-Menten-type equation of the rate law and the steady-state concentration of thiolacetate ion B- (see Fig. 4 and equation of the reciprocal plot of u vs concentration of PMS derived below). Equating Vmm and the product of kl times the concentration of cephalosporoate ABH- gave the value of klfirst-order rate constant for the elimination of thiolacetate from cephalosporoate. The value of the extinction coefficient of MTT formazan in a 0.3% (w/v) BSA solution, pH 7.0, phosphate buffer at 564 nm was found to be 13,000 M-l cm-‘. The reaction scheme shown in Fig. 6 may be represented as ABHB-+PMSL

ks fast

4Fl 4tl

F+PMS

= kJB-][PMS].

VI

Applying the steady-state condition expression for the rate law

W44~~-1 ’ = kJA-][H+]

+ kJPMS]

to [B-]

yields the

which has the form of Michaelis-Menten netics

v” = K,,, + [PMS]

1 ’ [PMS]

PM%

PI

saturation

ki-

[PMS].

1 + kJABH-]

Sensitivity

of the iMethod

Determination of the sensitivity of the method was accomplished by plotting the initial rate differences in the presence and absence of /?-lactamase, as a function of enzyme concentration. Since the rate measurements were done by the initial rate method, the kinetic runs were short enough to ensure that only 5% or less substrate was utilized. The sensitivity is based on the ability to discern the initial rate difference between the blank reaction and the enzyme-catalyzed reaction so that the latter is at least 10% greater than the former.

PMS stock solutions, 3.1 X lop3 g in 10 ml of 0.1 M phosphate, pH 7.0, buffer, were freshly prepared before each kinetic run. Stock solutions of tetrazolium salts INT, NBT, and MTT were kept in 0.1 M phosphate buffer solutions, pH 7.0, in the presence of 0.5% (w/v) BSA. These solutions were stable for 6 months at 228’C. All /3-lactamase enzyme solutions were prepared immediately before use, as 0.1 M phosphate buffer solutions.

p-Lactamase-Induced Thiocephalosporins

L31

Inverting both sides of Eq. [2] and making the substitution K&J[B-] = [A-][H+]/[ABH-] yields 1 1 ; = m

employing SAS STAT Program Version 6.03 and an IBM PC. Since steady-state concentrations of B- and ABH- at pH 7.0 can be easily calculated from our measurement of Kass, substitution of the pertinent values into Eq. [4] and equating terms of [4] and ]5] yield the values of kl and kz.

RESULTS

- [PMS]

kl [ABH-] = (k-l/kZ)[A-][H+] + [PMS]

323

B-LACTAMASES

Reagents

PMSH

slow

PMSH+Tu = -

2 A- + B- + H+ h-1

FOR

’

i41

which has the form

The values of V- and K,,, were obtained from the direct fit of the Michaelis-Menten-type equation [3] above,

Redox System of C3’ and Tetrazolium Salts

The reduction of tetrazolium salts by thiols was described several years ago (10,ll). The colored, water-insoluble formazans produced in this reaction are of considerable importance as histochemical reagents for sulfhydryl groups. Our preliminary experiments indicated that the rates of reductions of the tetrazolium salts depended on the pK= of the conjugate acid of the thiolate anion, among other factors. The least basic thiolates were most efficient as reducing agents. We found that thiolacetate ion, pK,, 3.33 (12), is a very efficient reductant of tetrazolium salts to their corresponding formazans. It occurred to us that cephalosporins functionalized at the C3’position with suitable thiolates, when exposed to the action of the appropriate B-lactamases, should readily reduce tetrazolium salts to the corresponding formazans as a consequence of the elimination of the thiolate ion. We synthesized C3’thiolacetic acid cephalosporin TAC and investigated the reactivity of this cephalosporin as a model system in the redox process depicted schematically in Fig. 1. Our choice of TAC as model C3’

324

BIENIARZ,

I\ P

I s

P /

YOUNG,

-0

TEl-RAZOLRJM ELECIRON

FIG.

1.

SALT, * CARRIER

Minimal

scheme

FOWZAN-water chromogenic.

of the

CORNWELL

order MTT > INT > NBT or 3.0 X lo-‘, 2.4 X 10e6, and 1.6 X lO-7 M mine*, respectively. This order of reducibility is identical to that of the previously reported reducibility of tetrazolium salts (15).

S

GO

AND

redox

insduble,

process.

thiocephalosporin was dictated not only by the low pK,, of the leaving group, but also our realization that the nucleofugal ability of the C3’ substituent and, thus, the concentration of the reductant in solution are inversely related to the pKa of the substituent (13,14). Enzyme-catalyzed hydrolysis of the fl-lactam ring generates cephalosporoate enamine ABH-. We found that under experimental assay conditions of 1.3 X 10e3 M TAC and 3.2 X 10e7 M /3-lactamase P-99, the hydrolysis of the &lactam ring of TAC is very fast, so that cephalosporoate ABH- accumulates. This was determined by measuring rate of the p-lactam hydrolysis of TAC substrate at 270 nm (13). Thus, the kinetic considerations that follow refer to the postenzymatic phase of the process. The cephalosporoate intermediate (13,14), in a postenzymatic step, eliminates the C3’ thiolate substituent, which is in equilibrium with the resulting enimine carboxylate A-. The eliminated thiolate B-, in a redox step which is catalyzed by a variety of electron carriers, reduces the tetrazolium salt T, to the water-insoluble, highly chromogenic formazan F. The rate of the reaction is dependent on pK= of the thiolate in the C3’ position of the cephalosporin, the reduction potential of the tetrazolium salt, and other factors presented below. Figure 2 shows time-dependent changes of the electronic spectrum of MTT in the process of being reduced to the corresponding formazan by the thiolacetate ion released from the cephalosporoate ABH- in the postenzymatic step. Clearly, p-lactamase-catalyzed ring opening of the cephalosporin and the concomitant release of the reducing thiolate anion from the C3’ position of the cephalosporoate cause a rapid generation of a chromogenic species at A,,,= 564 nm. The inset to Fig. 2 compares the reactivities of the thiolacetate anion released by cephalosporoate ABHtoward three tetrazolium salts of different reduction POtentials. The reactions were carried in the presence of 0.3% (w/v) BSA, which affects the colloidal dispersion of the formed formazans and thus allows the kinetic study of the system by uv-vis spectrophotometry. The reducibility of the three tetrazolium salts follows the

Dependence of the Reaction Rate on the Type of /3Lactamuse The rate of the redox process described in this work is dependent on the type of /3-lactamase used. Comparison of the initial rates of hydrolysis of the /3-lactam ring of TAC cephalosporin catalyzed by three ,5?-lactamases shows the following order of reaction rates: Type III from E. cloucae P-99 > Type I from B. cerezu P-0389 > RTEM-2 from E. coli. Since we were interested in the reactivity of different p-lactamases toward the TAC substrate, these reactions were monitored by observing the decrease of absorbance of the cephalosporins at 270 nm. At this wavelength the cleavage of the @-lactam ring may be monitored directly (16,17) (Fig. 3). Interestingly, cephalothin, which differs from TAC only in having at its C3’ position an oxygen atom instead of sulfur, has shown the same order of reactivity as TAC. Since the cleavage of the b-lactam ring, measured by the decrease of the absorbance at 270 nm, leads in the first step to the formation of the crucial cephalosporoate ABH--a precursor of the thiolate which is eliminated from species ABH--the overall rate of the process depicted in the Fig. 1 may depend on the efhciency of the enzymatic cleavage of the p-lactam ring. Blank Reaction A complicating TAC-tetrazolium

factor in the redox kinetic study of the system was the blank reaction of for-

FIG. 2. Time-dependent spectral changes of the solution of MTT, TAC, PMS, and &lactamase P-99. A l-cm path cuvette containing 1.25 X low3 M MTT, 1.56 X 10e3 M TAC, 3.13 X 10-s M PMS, and 3.13 X 10-r M fl-lactamase P-99 was scanned every 60 s for 30 min. The solution was pH 7.0 phosphate buffer and 0.3% (w/v) BSA to effect colloidal dispersion of MTT. The inset shows the comparison of the initial rates for the formation of formazans when 1.25 X 10e3 M MTT, INT, and NBT were reduced under conditions described in Fig. 1. The electronic spectral changes were recorded at wavelengths of 500 nm (INT), 530 nm (NBT), and 564 nm (MTT). The ordinate axis shows the differences of the overall formation of formazan minus the blank reaction in absence of the ensyme.

CHROMOGENIC

REDOX

ASSAY

FOR

325

&LACTAMASES

free to bound thiolacetate was 1.24. Substitution of this ratio into the equation for Kdiss shown under Materials and Methods yields the value Ksss = 8.76 X lo-” M’. This value is in excellent agreement with the values of dissociation constants for similar cephalosporins reported by Pratt and Faraci (13). The role of the Electron Carrier PMS

FIG. 3.

Reactivity of three fl-lactamases toward TAC substrate, at pH 7.0,25.0°C. Reactions were initiated by addition of 0.100 ml stock solution of @-lactarnase to 1.900 ml of TAC solution to yield initial concentrations of 9.5 X lo-’ M TAC and 5.0 X lo-’ M enzyme.

mation of formazan reduction products in the absence of ,&laetamase. We found that BSA, as well as several other proteins used in our study, was capable of inducing this blank reaction. BSA was used throughout this study to keep the formazan in colloidal suspension and thus make possible the observation of the spectral changes accompanying the reaction. The blank reaction rates were of considerable magnitude. Thus, under the conditions of our experiments, in a 0.3% (w/v) BSA, pH 7.0, buffer solution of TAC substrate and in the absence of /I-lactamase, the initial rates of reduction of MTT, INT, and NBT were 28.5, 34.1, and 82.3% of the total initial rates (sum of enzyme and blank reactions). The very large contribution of the BSA blank to the overall initial rate in the case of NBT is a consequence of the intrinsic slowness of the redox reaction of thiolacetate and NBT, as shown in the inset to Fig. 2. The following results must be considered in the analysis of the origin of the blank reaction: (1) buffered or unbuffered solutions of tetrazolium salts are stable for months in the presence of various concentrations of BSA as well as other proteins; (2) the rate of the reaction of Ellman’s reagent with pH 7.0 buffered, 0.3% (w/ v) BSA solutions of TAC is very similar to the rates of the blank reactions; (3) buffered and unbuffered solutions of cephalosporins substituted at the C3’ position with various chromogenic groups, i.e., pyridinium 2azo-p-dimethylaniline cephalosporin (18) and Z-nitro5-mercaptobenzoic acid cephalosporin (19), upon incubation with BSA, develop color due to the elimination of their chromogenic C3’ substituent, albeit at rates at least 1000 times slower than the blank rates reported here.

The reductions of tetrazolium salts to colored formazans by thiolates released from cephalosporins are significantly accelerated by electron carriers, i.e., PMS, PES, Meldola Blue, or diaphorase enzyme (15,ZO). Despite the notorious photolability of PMS, we selected it as electron carrier, because of its superior performance over that of other electron carriers. Figure 4 shows the reduction rate of the tetrazolium MTT by thiolacetate eliminated from TAC as a function of PMS concentration. The concentration of PMS was varied over a 250fold range at 25.O’C in a pH 7.0 phosphate buffer. The reaction shows a leveling off at higher concentrations of PMS, suggesting a saturation effect. A double-reciprocal plot of the data in Fig. 4 is shown in the inset. The fit shows an excellent correlation coefficient R2 = 0.999. Direct fit of the saturation curve shown in Fig. 4 yielded the values of Vmm = 1.34 X lob2 AU/s, or 1.03 X 10e6 M s-l, and Km = 1.76 X 10m4M PMS. The values /r2 = 1.35 M-l s-l and kl = 4.84 X lop4 s-l were determined as described under Materials and Methods. Since the value of the dissociation constant Kass has been determined to be 8.76 X lo-” M2 and Ktiss = kllk-l, kvl = 5.53 X lo5

M-2

0

s-l.

2 PMS

Determination

of Kdtis for TAC Cephulosporoate

ABI%-

Determination of the dissociation constant of TAC cephalosporoate ABH- was accomplished in a pD 7.0 phosphate buffer at 25OC by ‘H NMR spectroscopy, as described under Materials and Methods. The ratio of

FIG. 4.

4 6 8 CONCENTRATION@l~xI04

10

Plot of the initial velocities of the reduction of MTT by thiolacetate produced in P-99 b-lactamase-catalyzed hydrolysis of TAC, as a function of PMS concentration. In each experiment, initial velocities of the reaction of 1.29 X lo-’ M MTT, 6.45 X 10e3 M TAC, and PMS concentrations ranging from 3.23 X 10V6 to 8.26 X 10m4 M were initiated by 3.23 X lo-’ M P-99 enzyme. Solutions were at pH 7.0 and 0.3% (w/v) BSA to keep formazans in colloidal suspension.

BIENIARZ,

Rz

YOUNG,

AND

CORNmLL

because at such high concentrations of /3-lactamase the formation of the cephalosporoate ABH- is extremely fast, resulting in accumulation of ABH-. The nonlinear regression fit of the sigmoidal plot of Fig. 5 to the expression

= 0.994

Vmx{ -log[P-99]}n v = {-log[P-99]}n + Km gave the inflection point value of &?,, = -log[P-991 8.52, corresponding to 5.62 X lo-’ M P-99, and V,,1.82 X lo-’ AU/min, when n = 1.59.

= =

7

10

9

8

-Log[P-991

7

6

5

(M)

FIG. 5. Plot of the initial velocities vs concentrations of B-lactamase P-99 in pH 7.0 solution of 1.25 X 10e3 M INT, 1.56 X 10e3 M TAC, 3.13 X 10m6 M PMS, 0.3% (w/v) BSA. The reactions were initiated by addition of varying amounts of P-99 enzyme to make final concentrations in the cuvette ranging from 3.13 X 10-i’ to 1.25 X lo-’ M. The progress of the reactions was followed by recording an increase at Am for the first 5% of the reaction. At each point, the blank reaction in the absence of enzyme was substracted from the overall rate.

Sensitivity

of the Method

The assay method described here detects 3 x lo-” M /3-lactamase P-99 at pH 7.0,25’=‘C, within 5 min of incubating solutions of INT, PMS, and TAC substrate. Since we found the turnover numbers for TAC and cephaloridine to be very similar, this sensitivity translates to 30 mIU of /3-lactamase. A very similar result is achieved when MTT is used in place of INT. This is not surprising in view of the very similar rates of reduction of these tetrazolium salts by the thiolacetate released in the enzymatic reaction. The sensitivity is based on the ability to discern the initial rate difference between the blank reaction and the enzyme-catalyzed reaction so that the latter is at least 10% greater than the former. Figure 5 depicts the dependence of the initial velocities of the redox process on the concentration of P-99 /I-lactamase. It is important to note that at enzyme concentrations in excess of 1 X lo-’ M the initial velocities level off. Nonlinear regression yielded excellent sigmoida1 fit, with initial velocities reaching asymptotic zeroorder dependence on the enzyme concentration at an initial velocity of 18.2 X 103 AU/min. This occurs despite the fact that we ensured that the rates of the enzymatic @-lactam ring cleavage measured at 270 nm are perfectly within the limits of linearity in the plot of velocities vs enzyme concentrations, when these exceed 1 X lo-’ M. It is important to emphasize that at 500 nm, as shown in Fig. 5, the kinetics of the postenzymatic steps no longer depend on the overall enzyme concentration,

DISCUSSION

The redox process described in this work yields chromogenic, water-insoluble precipitates. Consequently, this process lends itself to immunochromatographic applications for the detection of /3-lactamase, as well as for use in solid-phase enzyme immunoassays as labels. We undertook kinetic study of this system in order to gain understanding of the mechanisms and limitations of the reductive process described above. A mechanism consistent with the experimental findings described above is shown in Fig. 6. In the first step, &lactamase catalyzes the hydrolysis of the fi-lactam ring of the TAC substrate. Under our experimental conditions, this is a very fast step. The dependence of the rate on the concentrations of P-99 enzyme is shown in Fig. 5. The plot is clearly sigmoidal, suggesting a complex dependence of the rate on the enzyme concentration. Our kinetic study was performed at a 3.23 X lo-’ M concentration of P-99. Figure 5 shows a clear zero-order dependence of the rate at enzyme concentrations exceding 1 X lo-’ M. However,

~wH 3

COW

co-

TAC

ABH-

+ -scocH~ + H+z k2 slow

BCOO-

A-

T

F FORMAZ4N e&wed water insoluble precipitate.

TETRAz0LruM SALT-colodess, watex soluble.

FIG.

6.

Proposed

mechanism

of the

redox

process.

CHROMOGENIC

REDOX

ASSAY

at lower enzyme concentrations the reaction shows a first-order dependence and at concentrations below -log[P-991 = 8.52 or 5.62 X lo-’ M P-99, a second-order dependence is quite apparent. The second-order dependence of the rate on the enzyme concentration in the lower range of the plot implies the requirement of two enzyme turnovers and, consequently, the release of two thiolacetate ions in or before the rate-determining step. That this should be so is quite reasonable, in view of the stoichiometric requirement for two thiolates to form a disulfide oxidation product concomitant to the reduction of a single tetrazolium salt to a formazan. Clearly, at lower concentrations of the enzyme the intermediate ABH- no longer accumulates, and the rate of the enzymatic cleavage of the fl-lactam ring impacts the overall rate. We synthesized several C3’ thiolate-substituted cephalosporins and found that the overall rate of the reaction depends on the reduction potential and pK,, of the thiolate substituents. This result agrees with the previously described dependence of the rate of elimination of C3’ substituents on the pK= of the conjugate acid of the leaving group (13). The rate also depends on the reduction potential of the tetrazolium salt, as well as the concentration of PMS. We undertook the kinetic study of the redox system using a thiolacetate-substituted cephalosporin substrate TAC. Due to low pKa of the thiolacetate, this ion is a relatively good leaving group, as reflected by our determination of the value of Kdss = 8.76 X lo-” M2. This value is consistent with the values determined by Pratt and Faraci of Kdis8 = 1.55 X lo-’ Mz and Kdis8 = 1.21 X lo-l2 M2 for 3,4-lutidinium and thiophenol, respectively (13). From this value of the dissociation constant and the concentration of TAC used in our study, the value of free thiolacetate was calculated to be 4.32 X lop3 M. Consequently, sufficient concentration of thiolacetate is attained to carry out the next step of reduction of PMS to PMSH. The reaction rate clearly depends on the concentration of the electron carrier PMS, as demonstrated in experiments represented in Fig. 4. The standard reduction potential of tetrazolium salts is very important. All three tetrazolium salts used in our experiments were reduced to formazans. However, we found that tetrazolium salts with significantly more negative half-wave potentials (15) than those used in our work, i.e., neotetrazolium, blue tetrazolium, or 2,3,5-triphenyl-2H-tetrazolium, were reduced only with great difficulty or not at all. The blank reaction-reduction of MTT, INT, and NBT by aqueous 0.3% (w/v) BSA solutions of TAC at pH 7.0 in the absence of b-lactamase-was a complicating factor in our study. In view of our finding that pH 7.0 0.3% (w/v) BSA solutions of tetrazolium salts may be stored for long periods without appearance of color, we feel that the relatively rapid decomposition of TAC cephalosporin, catalyzed by proteins, i.e., BSA, is the cause of the blank reaction. The thiolacetate produced

FOR

fl-LACTAMASES

327

in the putative decomposition is then the reducing agent. The thiolacetate could in principle result from aminolysis of the @-lactam ring and subsequent elimination of C3’ thiolacetate. In this putative process, the amines on the surface of BSA would act as nucleophiles, triggering the opening of the ring (16,21,22). In our case this is not very likely, however, because other C3’ chromogenic substituents of similar pKa, when exposed to identical conditions as our blanks, react at least lOOOfold slower than the rate of the blank reaction discussed here. The fact that under blank conditions Ellman’s reagent elicits an increase in absorbance at 410 nm in the presence of TAC at essentially the same rate as the blank redox reaction, in conjunction with the other considerations, indicates the aqueous hydrolysis of thiolacetate at C3’ as the cause of the blank. The reaction of transfer of acyl group from a thiolate to water molecule could be general base catalyzed with amines on the surface of BSA acting as general base (23). We synthesized several other C3’ thiol ester-substituted cephalosporins and found that the blank reaction could be suppressed to a great extent in cases of more basic thiolates at C3’. However, such substitution also resulted in a lower reaction rate due to a smaller Knin and, consequently, a lower nucleofugal capability of such thiolates. Our data are consistent with the rate-determining transfer of two electrons from thiolacetate to electron carrier PMS to form phenazine PMSH. In fact, the plot of the dependence of the reaction velocity on the concentration of PMS shows a leveling of the rate at higher PMS concentrations. The rate becomes zero order in PMS at higher concentrations of the electron carrier. The mechanism proposed above and the derived mathematical expression of the dependence of the rate on PMS concentration are consistent with conversion of PMS to PMSH as the rate-determining step, followed by the rapid electron transfer from PMSH to the tetrazohum salt. In our mechanistic scheme, the oxidation of PMSH by MTT is fast, as it occurs after the rate-determining step. Moreover, the reaction in our experimental setting was carried out under aerobic conditions. In general, the more positive the redox potential, the more readily the system is reduced. The second-order rate constant for the oxidation of PMSH to PMS by molecular oxygen was reported to be 180 Mm1 se1 (24), or 124 times greater than our measured value of k2, second-order rate constant for the reduction of PMS by the thiolacetate released in the enzymatic reaction. Since the reduction potential for oxygen is -0.076 V (12,24) as compared to -0.090 V for INT and -O.llV for MTT (15), it is reasonable to expect that the rate of oxidation of PMSH by MTT is also similarly faster than the value of k2. Moreover, the air oxidation of PMSH is almost certainly competitive with the reoxidation of PMSH by the tetrazolium salt, and this dual mode of PMSH reactivity should

328

BIENIARZ,

YOUNG,

be reflected in a lower sensitivity of our assay system under aerobic conditions as compared to that in anaerobic experiments. We confirmed this point by comparison of the maximum absorbances in the experiments carried out in the presence and absence of oxygen, although we made no effort to study these reactions under anaerobic conditions. Most of the existing methods for the detection of p-lactamases in a biological medium are based on measuring the development of color changes after a chromogenic assay reaches its maximum absorbance. The method described in this paper is a kinetic assay, since the rate of the appearance of color is measured. The technique described here is very sensitive compared to other methods reported in the literature (25). It allows the detection of 3 X 10-l’ M @-lactamase E. cloacae P-99 after only 5 min of incubation. This translates to approximately 30 mIU of @-lactamase, compared to 0.27 mIU for the microiodometric method, 13 mIU for the iodometric methods, or 400 mIU for alkalimetric assays (25). Since the unit of fl-lactamase is defined in terms of its reactivity toward cephaloridine as substrate and we found turnover numbers for cephaloridine and TAC to be very similar, the comparison of the sensitivity of our method with that of other assays is valid. However, it is worthwhile to note that while many of the existing assays require lengthy incubation periods of 30 to 60 min (7), this assay is typically read within the first 2 to 5 min of incubation. Also, unlike the iodometric technique and some other chromogenic detection methods, the substrates described here generate color starting from a colorless solution. ACKNOWLEDGMENTS We thank Professor W. L. Mock of University of Illinois at Chicago for reading and valuable critique of the manuscript. We also appreciate the help we received from Dr. David LeBlond in the use of SAS STAT computer program.

REFERENCES 1. Bush,

K. (1988)

2. Frere,

J. M.,

Rev.

and Joris,

Znfect.

Dis.

B. (1988)

10,681-690. Ann.

Biol.

Clin.

46,

151-156.

AND

CORNWELL

3. Ross, G. W., and O’Callaghan, mology (Hash, J. H., Ed.), New York. 4. Jorgensen, Antimicrob.

J. H., Crawford, Agents Chemother.

5. Jones, R. N., Wilson, Thornsberry, C. (1982)

8. Singer,

C. H. (1975) in Methods in Enzy43, pp. 69-85, Academic Press,

S. A., and Alexander, 22,162-164.

G. A. (1982)

H. W., Novick, W. J., Barry, A. L., J. Clin. Microbial. 15, 954-958.

6. O’Callaghan, C. H., Morris, (1972) Antimicrob. Agents 7. Cohenford, Biochem.

Vol.

A., Kirby, Chemother.

M. A., Abraham,

S. M.,

and Shingler,

and A. H.

1,283-288.

J., and Medeiros,

A. A. (1988)

Anal.

168,252-258. T. P., and

Kearney,

E. B. (1954)

B&him.

Biophys.

Actu,

15,151. 9. Nachlas, M. M., Margulies, Biol. Chem. 235,499-503. 10. Findley, 11. Deguchi,

S. I., and

G. H. (1955) J. Histochem. Y. (1964) J. Histochem.

Seligman,

Cytochem. Cytochem.

12. Weast, R. C. (Ed.) (1984) CRC Handbook ics, CRC Press, Boca Raton, FL. 13. Pratt, R. F., and Faraci, W. S. (1986) 5328-5333. 14. Faraci, W. S., and Pratt, R. F. (1984) 1489-1490. 15. Altman, F. P., (1976) Prog. Histochem. 16. Page, M. I., and Proctor, 3825. 17. Faraci, W. S., and Pratt, 18. Barlam, 189.

T., and Neu,

R., Quayle,

H. W.,

and

J. R. (1979)

3,331-338.

Chem.

Sot.

108,

J. Am.

Chem.

Sot.

106,

Cytochem.

9, l-56.

Chem.

J. Clin.

Novick,

Anal.

and Phys-

J. Am.

106,3820-

Sot.

Biochemistry Eur.

25,2934-2941. Microbial.

Biochem.

in Chemistry York.

24. Halaka, F. G., Babcock, 257, 1458-1461.

G. T., and Dye,

25. Citri, N., and Pollock, Biochem. 28,237-323.

M.

R. (1966)

99,

J. Clin.

112-117. J. Chem.

II,

Truns.

and Enzymology,

J. L. (1982) Adu.

3, 185-

W. J. (1982)

21. Proctor, P., Gensmantel, N. P., and Page, M. I. (1982) Sot. Perkin Trans. II, 1185-1192. 22. Morris, J. J., and Page, M. I. (1980) J. Chem. Sot. Perkin 212-219. 23. Jencks, W. P. (1969) in Catalysis pp. 517-523, McGraw-Hill, New

J.

12,261-265.

of Chemistry

J. Am.

R. F. (1986) H. C. (1984)

19. Jones, R. N., Wilson, Microbial. 15,954-958. 20. Ghosh,

P. (1984)

A. M. (1960)

Enzymol.

J. Biol.

Chem.

Relut.

Subj.