A Strategy for the Determination of Enzyme Kinetics Using

the need of a calibration curve. Kinetic parameters, such as Km, Vmax and Ki (for thyroxine), obtained by electro- spray mass spectrometry agreed with...
0 downloads 0 Views 57KB Size
Download PDF
Anal. Chem. 2001, 73, 5078-5082

A Strategy for the Determination of Enzyme Kinetics Using Electrospray Ionization with an Ion Trap Mass Spectrometer Xue Ge, Tammy L. Sirich, Martin K. Beyer, Heather Desaire, and Julie A. Leary*

Department of Chemistry, University of California, Berkeley, California 94720

A simple and rapid means of enzyme kinetic analysis was achieved using electrospray ionization mass spectrometry and a one-point normalization factor. The model system used, glutathione S-transferase from porcine liver, is a two-substrate enzyme catalyzing the conjugation of glutathione with a variety of compounds containing an electrophilic center. An internal standard that is structurally similar to the product was added to the reaction quench solution, and a single-point normalization factor was used to determine the product concentration without the need of a calibration curve. Kinetic parameters, such as Km, Vmax and Ki (for thyroxine), obtained by electrospray mass spectrometry agreed with those obtained from traditional UV-vis spectroscopy, and competitive vs noncompetitive inhibition reactions could be delineated via mass spectrometry. These results suggest that our method can be applied to enzymatic processes in which spectrophotometric or spectrofluorometric assays are not feasible or when the relevant substrates do not incorporate chromophores or fluorophores. This new method is competitive with traditional UV assays in that it is facile and it involves very little analysis time. Sequencing of the human genome, and numerous pathogen genomes, has resulted in the explosion of studies aimed at identifying potential drug targets. One of the accompanying challenges is the characterization of disease-associated proteins and enzymes. Coupled to this is, therefore, the required investigation of the mechanisms of enzymatic reactions, because this is one crucial aspect of the development and optimization of lead molecules. Enzyme kinetic parameters, such as the Michaelis-Menten constant, Km, and the maximum velocity Vmax, are commonly determined spectrophotometrically, provided the product of the enzymatic reaction exhibits strong absorption or fluorescence at a characteristic wavelength that is different from the reactants. However, many natural substrates do not provide the necessary chromophores, and either special chromophoric substrates have to be used or the product is converted to a chromophore in another reaction via a coupling enzyme. Multistep synthesis of the artificial substrate can be labor-intensive and time-consuming, and the kinetic behavior of the synthetic chromophoric substrates may differ significantly from the native substrate.1-4 The presence of a coupling enzyme with additional substrates complicates the 5078 Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

assay system and may interfere with the inhibition studies. Similarly, deconvoluting the different behaviors may be tedious, if not impossible. Other assay methods such as measuring radioactivity or circular dichroism also have their limitations.5 With the broad availability of soft ionization techniques, such as electrospray ionization (ESI), matrix-assisted laser desorption (MALDI), and fast atom bombardment (FAB), mass spectrometry would appear to be an excellent technique to complement the more traditional methods for determining enzyme kinetic parameters. In fact, a minireview by Northrop and Simpson6 portrays mass spectrometry as a technique that holds vast potential. Since the time of the Northrop review, several investigators have studied different aspects of enzyme kinetics using mass spectrometry.6-11 Zechel et al. have studied pre-steady-state kinetics using electrospray ionization (ESI) mass spectrometry, 7 and Reilly and co-workers have devised a novel MALDI technique allowing them to do much the same.8 Hsieh et al. 9 have investigated kinetic parameters by applying HPLC coupled to ESIMS to monitor the hydrolysis of dinucleotides with RNase A and lactose with galactosidase. They determined Km and Vmax by following the consumption of the substrate, presumably because of the relatively low ionization efficiency of mononucleotides and monosaccharides, respectively. Newton et al. 10 utilized fast-atom bombardment in the kinetics analysis of phosphodiesterases, and Bothner et al. 11 employed ESI-MS to determine Km and Vmax of glucosidase with p-nitrophenyl glucopyranoside. The latter obtained reasonable agreement of the mass spectrometric and spectrophotometric data, as did Henion and co-workers 9 for the (1) Bakken, A. P.; Hill, C. G., Jr.; Amundson, C. H. Appl. Biochem. Biotechnol. 1991, 28, 741. (2) Keilin, D.; Hartree, E. F. Biochem. J. 1948, 42, 230 (3) Cohn, M.; Monod, J. Biochim. Biophys. Acta 1951, 7, 153 (4) Huggert, A. G.; Nixon, D. A. Biochem. J. 1957, 66, 12P (5) Ge, X.; Campbell, R. E.; van de Rijn, I.; Tanner, M. E. J. Am. Chem. Soc., 1998, 120, 6613 (6) Northrop, D. B.; Simpson, F. B. Bioorg. Med. Chem. 1997, 5, 641 (7) Zechel, D. L.; Konermann, L.; Withers, S. G.; Douglas, D. J. Biochemistry 1998, 37, 7664 (8) Houston, C. T.; Taylor, W. P.; Widlanski, T. S.; Reilly, J. P. Anal. Chem. 2000, 72, 3311 (9) Hsieh, F. Y.; Tong, X.; Wachs, T.; Ganem, B.; Henion, J. Anal. Biochem. 1995, 229, 20 (10) (a) Newton, R. P.; Bayliss, M. A.; Khan, J. A.; Bastani, A.; Wilkins, A, C.; Games, D. E.; Walton, T. J.; Brenton, A. G.; Harris, F. M. Rapid Commun. Mass Spectrom. 1999, 13, 574. (b) Newton, R. P.; Evans, A. M.; Langridge, J. I.; Walton, T. J.; Harris, F. M.; Brenton, A. G. Anal. Biochem. 1995, 224, 32 (11) Bothner, B.; Chavez, R.; Wei, J.; Strupp, C.; Phung, Q.; Schneemann, A.; Siuzdak, G. J. Biol. Chem. 2000, 275, 13455-13459 10.1021/ac0105890 CCC: $20.00

© 2001 American Chemical Society Published on Web 10/03/2001

hydrolysis of dinucleotides with bovine RNase A, albeit neither mentioned the problem of in-source fragmentation, which we know to be problematic with enzyme systems involving hydrolysis. Turecek and co-workers12 have used ESI-MS to diagnose genetic diseases by profiling the enzyme activity. Having cited these cases, mass spectrometry has still not been shown to be a rigorous or routine method for the determination of enzyme kinetics. This is most probably due to the lengthy procedures required for the generation of calibration curves with standard compounds, many of which probably do not exist for the vast array of different enzyme processes or because sophisticated instrumentation and lengthy sample preparation are involved. Because most of the existing MS studies rely on calibration curves, the number of samples generated by MS is at least twice that of the standard UV assay. At this point, mass-spectrometry-based assay methods for kinetics are far from replacing the spectrophotometer in the enzymologist’s lab, as was originally suggested by Northrop and Simpson. The assay method developed and described herein is rather unique in that a calibration curve is not required; neither are chromophores a necessity. The number of samples generated is no more than that required for the traditional UV assay, and analysis time is approximately the same. The method can be applied to any enzymatic system, provided an internal standard of comparable structure and ionization efficiency to that of the product or substrate is found. An added advantage is the fact that the mass of the product can be monitored directly during the course of the reaction. It is well-known that the UV data from an enzyme assay produces an amount of information in excess of what is needed for a Lineweaver-Burk plot. Kinetic curves usually show reasonable linear behavior up to 10% of substrate conversion. This means that the initial velocity can be determined with a single datapoint, provided it is taken in this linear region. In so doing, a “normalization factor” can be calculated at relatively low substrate concentration, and this factor can be used to determine reaction velocity for each subsequent substrate concentration. For small initial substrate concentrations, [Si], the equilibrium lies on the side of the product, so that the final product concentration, [Pf], equals the initial substrate concentration for that particular reaction. This can be easily substantiated by monitoring the ions corresponding to substrate and product in the initial mass spectrum. The validity of the described approach was tested on the kinetic parameters of porcine liver glutathione S-transferase (GST) activity. As a two-substrate system, GST catalyzes the conjugation of glutathione using a variety of compounds containing an electrophilic center. This renders the toxic electrophile more easily excretable. In the system described herein, GST is used to catalyze 1-chloro-2,4-dinitrobenzene (CDNB) to the product S-dinitrophenylglutathione (SDPG) via glutathione conjugation as shown below.

The intensity of the ion representing the product of the reaction is monitored relative to an internal standard, and calculated reaction velocities are plotted against substrate concentration in order to obtain kinetic parameters such as Km and Vmax. For the studies involving the inhibitor, thyroxine, the same assay is used by varying both the inhibitor and substrate concentrations. Similar to the usual strategy with the spectrophotometer, one or two preliminary runs with a limited number of concentrations are necessary to adjust the reaction time and amount of enzyme used and to determine the range of substrate concentrations necessary to obtain a reliable Km and Vmax. However, the number of solutions that are prepared and analyzed using our technique is the same as with the traditional spectrophotometric method. This study indicates that the ESI-MS method is clearly competitive with traditional spectroscopic methods. EXPERIMENTAL SECTION General Methodology. (a) Calculation of Km and Vmax. Glutathione, 1-chloro-2,4-dinitrobenzene (CDNB), thyroxine, S-nitrobenzylglutathione (SNBG) and glutathione S-transferase (GST) from porcine liver were purchased from Sigma Co. and used without further purification. To prove validity, both the standard UV assay and the proposed MS assay were run in order to compare the measured Km and Vmax values using both methods. For the MS methods, Snitrobenzylglutathione (SNBG) was chosen as an internal standard, because it is close in molecular weight and has a similar ionization efficiency as that of the product. For both assays, stock solutions of GST (0.01 units/µL) and glutathione (10 mM) were prepared in 10 mM ammonium acetate buffer at pH 7.0. SNBG was dissolved to 1 mM in a 1:4 ethanol:buffer mixture, and CDNB was dissolved to 150 mM in pure ethanol. Buffer, glutathione, and CDNB stock solutions were mixed in the appropriate ratios to prepare 990 µL of each reaction mixture. After initialization of the reaction with 10 µL of GST stock, the CDNB concentration was 1 mM; 0.1 units/mL GST; and 0.02, 0.05, 0.07, 0.1, 0.2, 0.5, and 0.8 mM glutathione, respectively. The temperature of the reactions was maintained at 22 °C. After introducing the enzyme to start the reaction, the progress was monitored for 3 min in the UV spectrophotometer by recording the change of absorption at 340 nm using an extinction coefficient of 9.6 mM-1 cm-1. Upon reaching a total reaction time of 3 min, 200 µL of each reaction solution was quenched in 800 µL of methanol, and 20 µL of the SNBG internal standard solution was added prior to analysis. These solutions were analyzed by mass spectrometry, as described later. The remaining 800 µL of the 0.1 M glutathione sample was kept at room temperature and monitored for the production of product (m/z 472) and depletion of substrate (m/z 306), after which time, 200 µL was withdrawn and quenched with 800 µL of methanol and 20 µL of SNBG internal standard solution. This solution was used for determining the one-point normalization factor. At this point, all of the glutathione substrate from the 0.1 mM solution was converted to product, as evidenced by the mass spectrum, and therefore, [P] ) 0.1 mM for the data generated in this study. (12) Gerber, S. A.; Scott, C. R.; Turecek, F.; Gelb, M. H. Anal. Chem. 2001, 73, 1651

Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

5079

The normalization factor is determined from eq 1, shown below.

R)

(IP)[internal std] (IIS)[product]

(1)

IP is the intensity of m/z 472, which is the product of the reaction, and IIS is the intensity of the internal standard. This ratio is multiplied by the respective concentrations of the product and the internal standard in order to obtain R, the normalization factor. This R value is used for all subsequent calculations of [P] and V (velocity) for the seven samples analyzed. New normalization factors are calculated for each new set of samples generated; however, R values change very little from day to day over a onemonth period, provided all of the component concentrations remain consistent. In this study, the R value was determined to be 1.3, and it varied by only ( 0.2 during the time course of this study. For each sample analyzed, the product concentration is then calculated from eq 2, below, using the mass spectrometric data.

[P] )

(IP)[internal std] (IIS)(R)

(2)

Once the product concentration is calculated, it can be used to calculate the velocity (eq 3) at any specific substrate concentration, thus allowing for the determination of Km and Vmax. In eq 3, Tq is the quench time, and [E] is the enzyme concentration.

V)

[P] (Tq)[E]

(3)

Using the above set of equations, the internal standard concentration can be changed as needed, and the product concentration can still be readily determined. The method utilizes the facts that the initial velocity is equal to the product concentration divided by time and that the product ion intensity can be related to concentration through the normalization factor. This is viable, provided the reaction is quenched in the linear region of the reaction. The initial velocities were plotted as a function of substrate concentration, and the kinetic parameters were determined by a direct fit of the data to the Michaelis-Menten equation using the GraFit program. (b) Calculation of Ki. Thyroxine stock solution was prepared in ethanol with 1% HCl. The inhibition analyses were performed at 22 °C in 10 mM ammonium acetate buffer, at pH 7.0. The reaction solution was prepared in 200 µL, total volume. Four concentrations of thyroxine inhibitor were used: 0, 2, 6, and 10 µM. For each of the inhibitor concentrations, four glutathione concentrations ranging from 0.07 to 0.3 mM were used. The concentration of CDNB remained constant at 1 mM. The reaction was initiated by an addition of enzyme and quenched at 3 min. The concentration of GST stock solution was 0.0016 units/µL (10 µL used in the assay). The quenched solution mixtures were analyzed by ESI-MS, and the velocities were determined in the same manner as described above. The mode of inhibition was evaluated by double reciprocal analysis, and the Ki value was 5080

Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

obtained by nonlinear regression analysis of a linear competitive inhibition equation (as determined from the raw data) using the program SAS. Mass Spectrometry. Electrospray ionization mass spectrometry experiments were performed on a Thermo-Finnigan LCQ ion trap mass spectrometer equipped with a liquid chromatographic (LC) pump and Xcaliber version 1.0 software (Thermo-Finnigan MAT, San Jose, CA). The quenched solutions were delivered via the LC pump at a flow rate of 20 uL/min, and samples were analyzed in the negative ion mode using selected ion monitoring (SIM). For all of the experiments, the capillary was heated to 200 °C, and the spray voltage was maintained at 3.2 kV. Optimization of the ion of interest was initially completed using the automatic tuning parameter on the instrument. This tune file was then loaded for every experiment. Approximately 50 µL of the quenched solution was injected into the pump while the instrument was set to continual scan. The two ions monitored using SIM were the product of the reaction, SDNG, at m/z 472, and the internal standard ion, SBNG, at m/z 441. The signal intensity for each injection typically increases to approximately 3 × 106 detector counts and then decreases to ∼1 × 104 detector counts once all the sample is consumed, at which point the next quenched solution is injected into the pump. This procedure was repeated for all of the quenched solutions of various substrate and inhibitor concentrations. Both the chromatogram and the spectrum list were used to process the information. The chromatogram was used to monitor the progress of sample consumption over time. Each peak of the chromatogram corresponded to a different reaction mixture that was monitored. The intensities of the product and internal standard ions were generated by summing the intensities within 0.4 mass units around the center of each ion (m/z 472 and 441). For example, the intensities that were 0.2 mass units above 472 (m/z 472.2), the intensities that were 0.2 mass units below 472 (m/z 471.8), and the intensity of the central peak (m/z 472) were summed. This was repeated for the internal standard ion. These intensity sums were then inserted into eq 2 discussed above in order to determine the molar concentration of the products from which velocities were then determined using eq 3. RESULTS AND DISCUSSION Km and Vmax. In this study, we show that a two-substrate enzymatic reaction can be easily studied by mass spectrometry. To determine Km and Vmax for one of the substrates, such as glutathione, the concentration of this substrate was varied in each reaction while the other substrate, CDNB, was kept at a fixed saturating concentration. One or two runs were carried out using a small number of glutathione concentrations to estimate the Km value and to determine the proper substrate concentration range, which was generally between 0.2 and 5 times the estimated Km value. Provided the reaction is quenched at the linear kinetic curve region, less than 10% of substrate conversion, the initial velocity can be obtained. 1-Chloro-2,4-dinitrobenzene was chosen as the nucleophilic substrate, because it produces a chromophoric product, and therefore, a standard UV assay could be carried out in order to compare the kinetic parameters obtained from both assays.

Figure 1. Velocity vs substrate concentration data obtained from the MS assay.

Figure 2. Velocity vs substrate concentration data obtained from the UV assay Table 1. Kinetic Constants Calculated from Both Standard Ultraviolet Absorption Assay and the Mass Spectrometric Assay.

UV MS

Km, mMa

SDb

Vmax, mM/min/unita

SD

0.11 0.13

0.02 0.03

0.18 0.21

0.03 0.04

a Data generated from eight independent experiments deviations

b

Standard

By introducing the internal standard as part of the quenching solution, it becomes possible to use an internal standard that is similar to the product or substrate without interfering with the enzyme activity. The internal standard should be similar in ionization efficiency to the product or substrate in order to ensure a linear response over a large range of concentrations. The ideal standard would be the product or substrate with an isotopic label. S-nitrobenzylglutathione was chosen as the internal standard, because it is close in molecular weight and structure to the product of our model enzymatic system, S-dinitrophenylglutathione, and is expected to have a similar ionization efficiency. Because the product of this particular enzymatic reaction is not commercially available, we have adopted an approach that allows us to monitor the reaction to completion by following the uptake of substrate and production of product at some initial substrate concentration. Thus, the concentration of the product is equal to the substrate concentration when all the substrate has reacted. The completion of the reaction was checked by mass spectrometry to make sure that no residual glutathione was detected. These results suggest that our proposed strategy for the mass spectrometric measurement of Km and Vmax is valid and reproducible. Inhibition Studies. Because GST is thought to be responsible for drug resistance of tumor cells, therapeutic strategies aimed at inhibiting specific GSTs may be useful in extending the efficacy of those anti-cancer drugs. Thyroxine was reported to show competitive inhibition with respect to glutathione for GST from bovine brain, and the inhibition constant, Ki, was determined to be 6.6 µM.5 Duplication of the Lineweaver Burk plots in the presence of an inhibitor and comparison to the plot generated without inhibitor allows one to obtain information on whether the inhibitor is competitive or noncompetitive (see Figure 3). Once the velocities of the reaction are determined from eq 3 using our mass spectrometric method, we should then be able to calculate Ki from eq 4 for competitive inhibitors or from eq 5 for noncompetitive inhibitors.13

1/v ) (Km/Vmax){1 + ([I]/Ki)}(1/[S]) + 1/Vmax (4) 1/v ) (Km/Vmax){1 + ([I]/Ki)}(1/[S]) + (1/Vmax) {1 + ([I]/Ki)} (5)

Figure 1 is a plot of velocity vs substrate concentration obtained from the ESI-MS data of the reaction of glutathione with CDNB to SDPG catalyzed by GST, in comparison with the corresponding results from the UV assay (Figure 2). These data are for one experiment from a total of eight experiments generated using both assays. In both cases, the velocity approaches a maximum at a glutathione concentration of ∼1-2 mM. The ESI-MS data reflect very accurately the behavior of the results from the optical measurement. The average Km value calculated from eight replicate experiments was determined to be 0.13 mM for the ESI-MS assay and 0.11 mM for the UV assay (Table 1). In addition, the Vmax values of 0.21 mM min-1 unit -1 and 0.18 mM min-1 unit -1 for the MS and UV methods, respectively, are in excellent agreement. The R2 value of the corresponding double reciprocal linear Lineweaver Burk plots was 0.997 for both the ESI-MS and the UV assays.

The strategy reported here was to study the inhibition of thyroxine for GST from porcine liver and determine if the kinetic data generated by mass spectrometry supported competitive inhibition previously shown for bovine brain. Figure 4 is the double reciprocal plot of the initial velocity data generated using SAS. Raw data generated from the double reciprocal plots indicated that all four concentrations of inhibitor crossed at approximately the same 1/v on the y axis. Thus, the equation for a competitive inhibitor was used to generate the best fit plot using SAS. It clearly reveals a pattern consistent with competitive inhibition ,with a Ki of approximately 1.4 µM. This result is consistent with the inhibition study with glutathione S-transferase from bovine brain and clearly shows how (13) Walsh, C. Enzymatic Reaction Mechanism; W. H. Freeman and Co.: San Francisco, 1979

Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

5081

Figure 3. Typical Lineweaver-Burk plots for competitive vs noncompetitive inhibitors.

Figure 4. Lineweaver-Burk plot indicating competitive inhibition for thyroxine inhibitor. Data generated using MS assay.

the mass spectrometry assay can reveal information as to the competitive or noncompetitive nature of the inhibitor. The data, which were generated from five replicate experiments, were collected on five different days, and individual Ki’s calculated were 1.7, 1.2, 1.5, 1.4, and 1.1 µM. These ESI assays are now being used to calculate kinetic parameters from a series of inhibitor libraries for both estrogen sulfotransferase and Nod sulfotransferase and will be published separately.14 CONCLUSIONS A straightforward strategy has been developed for enzyme kinetic assay and inhibition studies. Addition of an internal standard to the enzymatic reaction quench solution made it (14) Verdugo, D. E.; Cancilla, M. T.; Ge, X.; Gray, N. S.; Chang, Y. T.; Schultz, P. G.; Leary, J. A.; Bertozzi, C. R. J. Med. Chem. 2001, 44, 2683-2686.

5082

Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

possible to use a standard that is chemically and structurally similar to the reaction components. Utilization of a one-point normalization factor simplified the SIM-MS analysis, which was performed on an ion trap instrument. The kinetic constants for GST obtained from the MS method agreed with those determined by traditional UV-vis spectroscopy. This method can be applied to rapidly characterize enzymes in which no simple spectrophotometric or spectrofluorimetric assay is feasible. This technique can also be used to quickly determine inhibition constants for inhibitors identified by high-through-put screening methods.

Received for review May 29, 2001. Accepted August 21, 2001. AC0105890