High-throughput screening of enzyme inhibitors: automatic determination of tight-binding inhibition constants

Determination of tight-binding inhibition constants by nonlinear least-squares regression requires sufficiently good initial estimates of the best-fit values. Normally an initial estimate of the inhibition constant must be provided by the investigator. This paper describes an automatic procedure for the estimation of tight-binding inhibition constants directly from dose-response data. Because the procedure does not require human intervention, it was incorporated into an algorithm for high-throughput screening of enzyme inhibitors. A suitable computer program is available electronically (http://www.biokin.com). Representative experimental data are shown for the inhibition of human mast-cell tryptase.     

Dual functional roles of ATP in the human mitochondrial malic enzyme

Human mitochondrial malic enzyme is a regulatory enzyme with ATP as an inhibitor. Structural studies reveal that the enzyme has two ATP-binding sites, one at the NAD(+)-binding site in the active center and the other at the exo site in the tetramer interface. Inhibition of the enzyme activity is due to the competition between ATP and NAD(+) for the nucleotide-binding site at the active center with an inhibition constant of 81 microM. Binding of the ATP molecule at the exo site, on the other hand, is important for the maintenance of the quaternary structural integrity. The enzyme exists in solution at neutral pH and at equilibrium of the dimer and tetramer with a dissociation constant (K(TD)) of 0.67 microM. ATP, at a physiological concentration, shifts the equilibrium toward tetramer and decreases the K(TD) by many orders of magnitude. Mutation of a single residue Arg542 at the tetrameric interfacial exo site resulted in dimeric mutants. ATP thus has dual functional roles in the mitochondrial malic enzyme.     

Determination of accurate KI values for tight-binding enzyme inhibitors: an in silico study of experimental error and assay design

Determination of accurate K(I) values for tight-binding enzyme inhibitors is important from both a basic biochemistry point of view (understanding the differences in affinity of related molecules) and a medicinal chemistry vantage (developing structure-activity relationships (SAR)). It is advantageous to directly fit the quadratic equation describing tight-binding behavior, known commonly as the Morrison equation, to obtain these K(I) values. The results of simulated experiments that examine the effect of assay design and experimental error on the ability to accurately determine K(I) values at several [E]0/K(I-app) ratios are described. Input ("true") values of the uninhibited velocity, inhibition constant, and total enzyme concentration were used to calculate the velocity at various inhibitor concentrations. Gaussian error was introduced into the velocities and the simulated reactions were fit to estimate upsilon0, K(I), and [E]0. Recommendations for optimizing the inhibitor dilutions within the context of a 96-well-plate format and simple serial dilution steps are made. These include using three points to determine the enzyme concentration ([I]=0, 0.5[E]0, and [E]0), using a narrow dilution series with only two or three points to determine the asymptote at high inhibitor concentration, and avoiding fixing [E]0 to a constant value in the fitting if at all possible. The risks and rewards of fixing [E]0 to a constant value, especially the effect on SAR, are also examined.     

Structural analysis of NSAID binding by prostaglandin H2 synthase: time-dependent and time-independent inhibitors elicit identical enzyme conformations

Nonsteroidal antiinflammatory drugs (NSAIDs) block prostanoid biosynthesis by inhibiting prostaglandin H(2) synthase (EC 1.14.99.1). NSAIDs are either rapidly reversible competitive inhibitors or slow tight-binding inhibitors of this enzyme. These different modes of inhibition correlate with clinically important differences in isoform selectivity. Hypotheses have been advanced to explain the different inhibition kinetics, but no structural data have been available to test them. We present here crystal structures of prostaglandin H(2) synthase-1 in complex with the inhibitors ibuprofen, methyl flurbiprofen, flurbiprofen, and alclofenac at resolutions ranging from 2.6 to 2.75 A. These structures allow direct comparison of enzyme complexes with reversible competitive inhibitors (ibuprofen and methyl flurbiprofen) and slow tight-binding inhibitors (alclofenac and flurbiprofen). The four inhibitors bind to the same site and adopt similar conformations. In all four complexes, the enzyme structure is essentially unchanged, exhibiting only minimal differences in the inhibitor binding site. These results argue strongly against hypotheses that explain the difference between slow tight-binding and fast reversible competitive inhibition by invoking global conformational differences or different inhibitor binding sites. Instead, they suggest that the different apparent modes of NSAID binding may result from differences in the speed and efficiency with which inhibitors can perturb the hydrogen bonding network around Arg-120 and Tyr-355.     

Self-association of diisopropylphosphoryl-alpha-chymotrypsin: the involvement of active site of the enzyme in its self-association behaviour

The self-association of diisopropylphosphoryl(DIP)-alpha-chymotrypsin is studied in order to find out whether the active site of the enzyme is involved in its self-association behaviour or not. Sedimentation coefficient as well as the weight-average (Archibald) molecular weight data are obtained as a function of concentration using an analytical ultracentrifugation technique. The analysis indicated that the experimental data fits the model of indefinite self-association. The comparison of the data with earlier data on alpha-chymotrypsin revealed that after the modification at the active site, the association constant for the self-association is reduced by about 47%, and the system deviated from ideality. Results showed further that Ser-195, at the active site, appears to be involved in the self-association behaviour of alpha-chymotrypsin; however, the participation of other groups at the active site is also implicated.     

Mixtures of tight-binding enzyme inhibitors. Kinetic analysis by a recursive rate equation.

When two or more tight-binding inhibitors are present in an enzyme assay, the equation that relates the initial velocity v to the concentration of reactants cannot be written in an algebraically explicit form. Rather, for n inhibitors it is an implicit polynomial equation of degree n + 1 with respect to v. The complexity of the polynomial coefficients dramatically increases with each added inhibitor. Solving the transcendental rate equation by traditional methods of numerical mathematics has proven tedious because of the sensitivity of these methods to initial estimates and because of the existence of multiple roots. However, the equation can be rearranged into a convenient recursive form, one in which the velocity appears on both sides and the solution is found iteratively. The algebraic form of the recursive rate equation is remarkably simple and differs from the rate equation for classical rather than tight-binding inhibition only by an added term. The numerical stability and the speed of convergence were tested on the case of two competitive inhibitors. Initial estimates of velocity that spanned 12 orders of magnitude converged within five iterations. The velocities computed with the recursive method for a single tight-binding inhibitor were identical with the values predicted by the Morrison equation. The method is used to analyze experimental data for the inhibition of rat liver dihydrofolate reductase by mixtures of the anticancer drug methotrexate and its metabolic precursor form, methotrexate-alpha-aspartate (a prodrug).     

Analysis of slow-binding enzyme inhibitors at elevated enzyme concentrations

The improvement in the characterization of slow-binding inhibitors achieved by performing experiments at elevated enzyme concentrations is presented. In particular, the characterization of slow-binding inhibitors conforming to a two-step mode of inhibition with a steady-state dissociation constant that is much lower than the initial dissociation constant with enzyme is discussed. For these systems, inhibition is rapid and low steady-state product concentrations are produced at saturating inhibitor concentrations. By working at elevated enzyme concentrations, improved signal-to-noise ratios are achieved and data may be collected at saturating inhibitor levels. Numerical simulations confirmed that improved parameter estimates are obtained and useful data to discern the mechanism of slow-binding inhibition are produced by working at elevated enzyme concentrations. The saturation kinetics that were unobservable in two previous studies of an enzyme inhibitor system were measured by performing experiments at an elevated enzyme concentration. These results indicate that consideration of the quality of the data acquired using a particular assay is an important factor when selecting the enzyme concentration at which to perform experiments used to characterize the class of enzyme inhibitors examined herein.     

Acetylcholinesterase. Two types of inhibition by an organophosphorus compound: one the formation of phosphorylated enzyme and the other analogous to inhibition by substrate

1. The kinetics of the reaction of di-(2-chloroethyl) 3-chloro-4-methylcoumarin-7-yl phosphate (haloxon) and related compounds with acetylcholinesterase were studied and found to be unusual. 2. By a progressive reaction haloxon produces a di-(2-chloroethyl)phosphorylated enzyme. The influence of substrate on this reaction leading to a phosphorylated active centre was studied. From competition experiments between inhibitor and substrate values of K(m) for acetylcholine and acetylthiocholine of 0.79mm and 0.23mm respectively were derived. 3. Haloxon also combines with acetylcholinesterase by a non-progressive reaction, producing a complex that is reversible by dilution and by high concentrations of acetylcholine and acetylthiocholine. From this non-progressive reaction the competition between haloxon and substrate was studied, and it was shown that haloxon combines with a site involved in inhibition by substrate. From competition experiments the following dissociation constants were derived: for combination of haloxon and this site K(i) is 4.9mum and for the combination of substrates with this site K(88) values are 12mm and 3.3mm for acetylcholine and acetylthiocholine respectively. 4. The non-phosphorus-containing compound 3-chloro-7-hydroxy-4-methylcoumarin was shown to be a good reagent for the site involved in inhibition by substrate; its dissociation constant for the combination with this site is 30mum. 5. In order to interpret the experimental results, theoretical equations were derived for an enzyme with two binding sites to both of which substrate and inhibitor can combine. The equations correlate the activity of the enzyme with the concentration of substrate and inhibitor, for both progressive and non-progressive inhibition. These equations are applicable to reactions of acetylcholinesterase with organophosphorus compounds, carbamates etc. and may be applicable to other enzymes possessing two binding sites.     

Appendix to the sulfur-cyanolysis sites of serum albumin: metabolite competition studies. Tight-binding inhibitions that yield atypical Henderson plots

Ordinary tight-binding inhibition in steady-state enzyme systems is conveniently evaluated by means of the Henderson plot. This is a linear plotting form that has an ordinate intercept equal to the total enzyme concentration. However, there are two experimental situations that yield deviations from the common Henderson plot form. These are inhibitor binding in a separate, noninhibitory mode that depletes the concentration of free inhibitor, and partial inhibition, i.e., the retention of partial activity by the enzyme-inhibitor complex. Noninhibitory depletion results in Henderson plots with elevated ordinate intercepts. Competitive partial inhibition yields a characteristic pattern of parabolic Henderson plots.     

Activity and dimerization of human immunodeficiency virus protease as a function of solvent composition and enzyme concentration.

The activity of human immunodeficiency virus 1 (HIV-1) protease has been examined as a function of solvent composition, incubation time, and enzyme concentration at 37 degrees C in the pH 4.5-5.5 range. Glycerol and dimethyl sulfoxide inhibit the enzyme, while polyethylene glycol and bovine serum albumin activate the enzyme. When incubated at a concentration of 50-200 nM, the activity of the protease decreases irreversibly with an apparent first-order rate constant of 4-9 x 10(-3) min-1. The presence of 0.1% (w/v) polyethylene glycol or bovine serum albumin in the reaction buffer dramatically stabilizes enzyme activity. In the absence of prolonged incubation of the enzyme at submicromolar concentration, the specific activity of HIV-1 protease in buffers of either high or low ionic strength is constant over the enzyme concentration range of 0.25-5 nM, indicating that dissociation of the dimeric protease, if occurring, can only be governed by a picomolar dissociation constant. Similarly, the variation of the specific activity of HIV-2 protease over the enzyme concentration of 4-85 nM is consistent only with a dimer dissociation constant of less than 10 nM. We conclude that: 1) the assumption of a nondissociating HIV-1 protease is a valid one for kinetic studies of tight-binding inhibitors where nanomolar concentrations of the enzymes are employed; 2) stock protease solutions of submicromolar concentration in the absence of activity-stabilizing compounds may lead to erroneous kinetic data and complicate mechanistic interpretations.     

Inhibition kinetics of mushroom tyrosinase by copper-chelating ammonium tetrathiomolybdate

With a strategy of chelating coppers at tyrosinase active site to detect an effective inhibitor, several copper-specific chelators were applied in this study. Ammonium tetrathiomolybdate (ATTM) among them, known as a drug for treating Wilson's disease, turned out to be a significant tyrosinase inhibitor. Treatment with ATTM on mushroom tyrosinase completely inactivated enzyme activity in a dose-dependent manner. Progress-of-substrate reaction kinetics using the two-step kinetic pathway and dilution of the ATTM revealed that ATTM is a tight-binding inhibitor and high dose of ATTM irreversibly inactivated tyrosinase. Progress-of-substrate reaction kinetics and activity restoration with a dilution of the ATTM indicated that the copper-chelating ATTM may bind slowly but reversibly to the active site without competition with substrate, and the enzyme-ATTM complex subsequently undergoes reversible conformational change, leading to complete inactivation of the tyrosinase activity. Thus, inhibition by ATTM on tyrosinase could be categorized as complexing type of inhibition with a slow and reversible binding. Detailed analysis of inhibition kinetics provided IC50 at the steady-state and inhibitor binding constant (K(I)) for ATTM as 1.0+/-0.2 microM and 10.65 microM, respectively. Our results may provide useful information regarding effective inhibitor of tyrosinase as whitening agents in the cosmetic industry.     

Determination of active enzyme concentration using activity-based probes and direct mass spectrometric readout

Activity-based probes (ABPs) are specific covalent inhibitors developed for different classes of enzymes. We have titrated a serine protease and a lipase with their specific ABPs and measured the extent of inhibition using nanoelectrospray mass spectrometry (nanoESI-MS). Because ABPs only interact with the active enzyme form, the approach allows to accurately measure the active enzyme concentration in solution. This is even possible in the presence of contaminants. The concentrations of the two enzymes were also investigated by UV spectroscopy, which appears to give higher concentrations than those measured with the active site titration method.     

Study of betacyanin-discolorating enzyme]

An enzyme catalyzing the discoloration and breakdown of betacyanins was isolated from beet roots Beta vulgaris by centrifugation in sucrose density gradient (2.5 M, 2.0 M, 1.5 M, 1.0 M, tris-HCl buffer, 0.05 M, pH 7.2), and purified 100-fold. The enzyme activity induced the discoloration of betanin, betanidin. It was found that the beet root enzyme exists in an insoluble state and is firmly bound with subcellular structures, which were isolated by centrifugation in sucrose gradient. The optimal activity of the enzyme was observed at pH 3.4, +40 degrees C. The dependence of the enzymatic reaction on the enzyme concentration showed a linearity. Studies of the enzyme inhibition by sodium azide, sodium diethyldithiocarbamate, thiourea, demonstrated that the active site of the enzyme contains a metal. The enzymatic discoloration of betanin is followed by the oxygen uptake.     

Homotropic effects in aspartate transcarbamoylase: What happens when the enzyme binds a single molecule of the bisubstrate analog N-phosphonacetyl-l-aspartate?

The active sites of aspartate transcarbamoylase from Escherichia coli were titrated by measuring the decrease in the enzyme-catalyzed arsenolysis of N-carbamoyl-L-aspartate caused by the addition of the tight-binding inhibitor, N-phosphonacetyl-L-aspartate. Because the enzyme is a poor catalyst for this non-physiological reaction, high concentrations are required for the assays (more than 1000-fold the dissociation constant of the reversibly bound inhibitor) and, therefore, virtually all of the bisubstrate analog is bound. From the endpoint of the titration, 5.7 active sites were calculated, in excellent agreement with the number, six, based on the structure of the enzyme. Simple inhibition was observed only when the molar ratio of inhibitor to enzyme exceeded five; under these conditions, as shown in earlier physical chemical studies, the R-conformational state of the enzyme is the sole or predominant species. At low ratios of inhibitor to enzyme, the addition of inhibitor caused an increase in activity which is attributable to the conversion of the enzyme from the low-activity T-state to the much more active R-state. Comparison of the linear increase in activity as a function of inhibitor concentration at the low molar ratio (0.01, i.e. 1 inhibitor/600 active sites) with the activity lost at the high ratio provided a direct value for the mean number of active sites converted from the T-state to the R-state as a result of the binding of one bisubstrate analog to an enzyme molecule. This number was four with Mg X ATP or carbamoyl phosphate present and 4.7 for the enzyme in the presence of Mg X PPi, values approaching or identical to the theoretical maximum, 4.7, for a concerted transition with all of the active sites of the molecule changing from the T- to R-state upon the formation of a binary complex of hexameric enzyme with a single inhibitor. With the enzyme in the absence of effectors or with Mg X CTP present, the titrations showed that an average of two and one sites, respectively, of 4.7 possible, changed conformation upon ligand binding. These results were interpreted as a manifestation of an equilibrium between a sub-population of T- and R-state enzyme complexes containing one bound inhibitor molecule. The R-state species would represent 40% of the population for aspartate transcarbamoylase in the absence of extraneous ligands.(ABSTRACT TRUNCATED AT 400 WORDS)     

The crystal structure of E.coli 1-deoxy-D-xylulose-5-phosphate reductoisomerase in a ternary complex with the antimalarial compound fosmidomycin and NADPH reveals a tight-binding closed enzyme conformation

The key enzyme in the non-mevalonate pathway of isoprenoid biosynthesis, 1-deoxy-D-xylulose 5-phosphate reductoisomerase (DXR) has been shown to be the target enzyme of fosmidomycin, an antimalarial, antibacterial and herbicidal compound. Here we report the crystal structure of selenomethionine-labelled Escherichia coli DXR in a ternary complex with NADPH and fosmidomycin at 2.2 A resolution. The structure reveals a considerable conformational rearrangement upon fosmidomycin binding and provides insights into the slow, tight binding inhibition mode of the inhibitor. Although the inhibitor displays an unusual non-metal mediated mode of inhibition, which is an artefact most likely due to the low metal affinity of DXR at the pH used for crystallization, the structural data add valuable information for the rational design of novel DXR inhibitors. Using this structure together with the published structural data and the 1.9 A crystal structure of DXR in a ternary complex with NADPH and the substrate 1-deoxy-D-xylulose 5-phosphate, a model for the physiologically relevant tight-binding mode of inhibition is proposed. The structure of the substrate complex must be interpreted with caution due to the presence of a second diastereomer in the active site.     

Kinetics of factor Xa inhibition by recombinant tick anticoagulant peptide: both active site and exosite interactions are required for a slow- and tight-binding inhibition mechanism

Recombinant tick anticoagulant peptide (rTAP) is a competitive slow- and tight-binding inhibitor of factor Xa (FXa) with a reported equilibrium dissociation constant (K(I)) of approximately 0.2 nM. The inhibitory characteristics and the high selectivity of rTAP for FXa are believed to arise from the ability of the inhibitor to specifically interact with the residues of both the active site as well as those remote from the active site pocket of the protease. To localize the rTAP-interactive sites on FXa, the kinetics of inhibition of wild-type and 18 different mutants of recombinant FXa by the inhibitor were studied by either a discontinuous assay method employing the tight-binding quadratic equation or a continuous assay method employing the slow-binding kinetic approach. It was discovered that K(I) values for the interaction of rTAP with four FXa mutants (Tyr(99) --> Thr, Phe(174) --> Asn, Arg(143) --> Ala, and a Na(+)-binding loop mutant in which residues 220-225 of FXa were replaced with the corresponding residues of thrombin) were elevated by 2-3 orders of magnitude for each mutant. Further studies revealed that the characteristic slow type of inhibition by rTAP was also eliminated for the mutants. These findings suggest that the interaction of rTAP with the P2-binding pocket, the autolysis loop, and the Na(+)-binding loop is primarily responsible for its high specificity of FXa inhibition by a slow- and tight-binding mechanism.     

Active-site titration of serine proteases in organic solvents

Calculation of kinetic constants of an enzymatic reaction in organic solvents requires knowledge of the functional active-site concentration in organic solvents, and this can be significantly different than that in water. An experimental method for active-site titration of serine proteases in organic media has been developed based on the kinetics of inhibition by phenylmethanesulfonyl fluoride (PMSF), a serine-specific inhibitor (or suicide substrate). This kinetic approach is fundamentally different from other techniques that require complete titration of all accessible enzyme active sites. This active site titration method was applied to subtilisins BPN' and Carlsberg and alpha-chymotrypsin and resulted in fractions of active sites that ranged from 8 to 62% (of the fraction active in water) depending on the enzyme, the method of enzyme preparation, and the organic solvent used. The active-site concentration of subtilisin BPN' and Carlsberg increased with increasing hydrophobicity of the solvent and with increasing solvent hydration in tetrahydrofuran. The dependence of the fraction of active sites on the nature of the organic solvent appears to be governed largely by solvent-induced inactivation caused by direct interaction of a hydrophilic solvent with the enzyme. (c) 1996 John Wiley & Sons, Inc.     

Inhibition of jack bean urease by p-benzoquinone: elucidation of the role of thiols and reversibility of the process

p-Benzoquinone (pBQ) was studied as an inhibitor of jack bean urease in 20 mM phosphate buffer, pH 7.0, 1 mM EDTA, 25 degrees C. The inhibition was carried out by the use of a preincubation procedure in the absence of substrate. The influence of the inhibitor concentration and the preincubation time on the enzyme activity was elucidated. It was found that increase in pBQ concentration resulted in a linear decrease of urease activity. The dependence of the enzyme activity on the preincubation time showed that the rate of inhibition rapidly decreased at the beginning of the process in order to achieve the constant value. The inhibition became time independent in the studied time range. This observation is characteristic of a slow binding mechanism of inhibition. The protective experiment proved that the urease active site is involved in the binding of pBQ. High effectiveness of thiol protectors against pBQ inhibition indicates the strategic role of the active site sulfhydryl group in the blocking process. There were two methods used for reactivation of pBQ-inhibited urease. The dilution of the urease-pBQ complex in urea solution did not result in a regain of enzyme activity. Alternatively, the addition of dithiothreitol into the urease-pBQ mixture caused the instant and efficient reactivation of the enzyme. The experiments showed that the nature of the urease-pBQ complex is irreversible but the application of a specific thiol reagent can release the active enzyme from the complex.     

An algebraic model for the kinetics of covalent enzyme inhibition at low substrate concentrations

This article describes an integrated rate equation for the time course of covalent enzyme inhibition under the conditions where the substrate concentration is significantly lower than the corresponding Michaelis constant, for example, in the Omnia assays of epidermal growth factor receptor (EGFR) kinase. The newly described method is applicable to experimental conditions where the enzyme concentration is significantly lower than the dissociation constant of the initially formed reversible enzyme–inhibitor complex (no “tight binding”). A detailed comparison with the traditionally used rate equation for covalent inhibition is presented. The two methods produce approximately identical values of the first-order inactivation rate constant (k_inact). However, the inhibition constant (K_i), and therefore also the second-order inactivation rate constant k_inact/K_i, is underestimated by the traditional method by up to an order of magnitude.