Inactivation Studies of Acetylcholinesterase with Phenylmethylsulfonyl Fluoride

Materials and Methods

Chemicals.

PMSF and BSF were purchased from Aldrich Chemical Company (Milwaukee, WI). All other chemicals were purchased from Sigma Chemical Company (St. Louis, MO) or Fisher Chemical Company (Pittsburgh, PA).

Enzymes.

The following enzymes were used: wild-type mouse AChE; four mutants of this wild-type enzyme (Y330A, Y330F, F288L, and F290I) (Radic et al., 1993); AChE from T. californica; wild-type mouse BChE; recombinant wild-type AChE from T. californica; and the double mutant F288L/F290V AChE from T. californica (Harel et al., 1992). AChE from T. californica was a kind gift of Drs. Jean Massoulié and Suzanne Bon (Laboratorie of Neurobiologie, Ecole Normale Supérieure, Paris, France).

AChE Assay.

The assay was conducted as described in Ellman et al. (1961), using the following final reagent and enzyme concentrations: 0.5 mM acetylthiocholine chloride (ATC), 0.1 mM 2,2′-dithionitrobenzoic acid in 0.1 M potassium phosphate buffer, pH 7.4, and enzyme with an activity of about 0.1 ΔAbs₄₁₂/min/ml. The absorbance at 412 nm was measured for 1 min on a Cary Varian UV/Vis spectrophotometer. Units correspond to ΔAbs₄₁₂/min.

AChE Inactivation in the Absence of Substrate.

The enzyme was diluted in 0.1 M phosphate buffer to an activity of approximately 2 ΔAbs₄₁₂/min/ml immediately before the experiment. Solutions of PMSF and BSF were prepared in methanol on the day of use. Inactivation reactions were carried out by adding 2.5 μL of inactivator solution to 0.5 ml of enzyme solution and incubating the mixture at 25°C. Aliquots (50 μl) were removed and diluted into assay medium (0.95 ml of assay medium in a 1.0 ml cuvette) every 2 to 4 min thereafter. Residual enzyme activity of each aliquot was measured by absorbance over 1 min at 412 nm.

Analysis of Data for AChE Inactivation in the Absence of Substrate.

The mechanism of inactivation of AChE by BSF is assumed to involve the formation of a reversible enzyme-inhibitor complex followed by the covalent modification of the enzyme (SchemeFS1).

The fractional activity (FA) for this scheme is:where v₀ and v_t are the velocities of the enzymatic reaction at times 0 and t, andEquation 1The natural logarithm of FA versus time was plotted and the value of k′ determined.

Then by plotting k′ versus [I], the values ofk₂ and K_I were determined for the inactivation of the enzyme by a nonlinear least-squares fit of the data to eq. 1 using KaleidaGraph 3.0 (Synergy Software, Reading, PA).k₂/K_I is the second-order rate constant for the inactivation of free enzyme by inhibitor.

AChE Inactivation in the Presence of Substrate.

The reaction mixture included 0.01 to 0.6 mM ATC and 1.0 mM 2,2′-dithionitrobenzoic acid in 0.1 M phosphate buffer, pH 7.4. Enzyme solution made with 0.1 M phosphate buffer, pH 7.4, was added to the reaction mixture to give an approximate activity of 0.005–0.05 ΔAbs/min/ml. PMSF and BSF solutions were freshly prepared in methanol with final concentrations in the reaction mixture ranging from 0.006 to 1.5 mM. The final methanol concentration was 0.1 to 0.5%. The reactions were carried out at 25°C. The absorbance at 412 nm was measured for 15 to 60 min using a Cary Varian UV/Vis spectrophotometer.K_m values for ATC were determined using ATC without inhibitor present.

Analysis of Data for AChE Inactivation in the Presence of Substrate.

Inactivation assays were performed on mutant mouse andT. californica AChE using an in situ procedure for measuring the inactivation rates as described by Hart and O'Brien (1973). This procedure is based on using a system containing enzyme, inhibitor, and substrate. The general scheme for this system is as shown in SchemeFS2, where I is the inhibitor, which in this case is PMSF or BSF. Kinetic analysis of the reactions gives the following equations (Hart and O'Brien, 1973)Equation 2orEquation 3where k₂ is the rate constant of sulfonylation of enzyme by PMSF, K_I is the dissociation constant of PMSF, k′ is the pseudo–first-order rate constant for the disappearance of active enzyme, and α is given by the following equation:Equation 4where [S] is the concentration of substrate andK_m is the Michaelis constant of the enzyme for ATC.

The values ofk₂/K_I were obtained from a plot of 1/[I] versus 1/k′ according to eq.2 or from a plot of 1/(1 − α) versus [I]/k′ according to eq. 3.

k′ was determined by one of two methods. Method 1: the derivatives of the graph of concentration of final product (thionitrobenzoate) with respect to time in both the presence and absence of inactivator were calculated. These derivatives are equal to the enzyme activity as a function of time, and yield FA. Plotting ln(FA) versus t yields k′ values for each concentration of inhibitor, I. Method 2: integrating the FA equationgives the equation:where C is a constant of integration. If the rate of nonenzymic hydrolysis of ATC (k_h) is taken into account, then the equation becomesEquation 5A₄₁₂ is plotted versus time and, using KaleidaGraph 3.0, a nonlinear regression to eq. 5 yields the value for k′, with an R²typically greater than 0.99.

If substrate concentration is kept constant and inhibitor concentration is varied, a plot of 1/[I] versus 1/k′ yields a straight line with a slope equal to [k₂ (1 − α)]/K_I and a y-intercept of −(1− α)/K_I, as depicted in eq. 2. Similarly, if inhibitor concentration is kept constant but substrate concentration is varied, a plot of 1/(1 − α) versus [I]/k′ yields a straight line with slopek₂/K_I and intercept −[I]/K_I, as depicted in eq. 3.

The value of K_m for ATC was determined under the conditions described in the AChE assay above, but with [ATC] varying from 0.01 to 0.6 mM. The initial rate of substrate hydrolysis was plotted against [ATC], and the data were analyzed using a nonlinear curve fit to the Michaelis-Menten equation.

Using eq. 2 or eq. 3, k′, and the measured value ofK_m to determine α, it was possible to determine k₂ andK_I values and the second-order rate constant for inactivation (k₂/K_I).

Results and Discussion

The K_m values we obtained for each of the enzymes studied is shown in Table 1, along with their reported relative turnover numbers.

Figure 3 shows a representative plot for the inactivation of mouse AChE by 4 mM BSF or 0.5 mM PMSF in the absence of substrate. The reactions follow pseudo–first-order kinetics with rate constants, k′, indicated on the graph. PMSF is the more reactive inactivator, as shown by the 8-fold higher BSF concentration necessary to achieve even a 6-fold slower inactivation than that using PMSF.

Figure 4 shows a representative time plot for the inactivation of T. californica AChE by 4 mM BSF or 4 mM PMSF in the absence of substrate. The time course for inactivation of the enzyme in the presence of 0.5% methanol is shown for comparison.

As first reported by Fahrney and Gold (1963), Fig. 4 shows that BSF is clearly the more reactive inactivator of T. californicaAChE. In fact, PMSF does not inactivate the T. californicaenzyme significantly more than does 0.5% methanol, even under the extremely high concentrations used (4 mM).

The inactivation by BSF of T. californica AChE, mouse wild-type AChE, and several mouse mutants was examined in the absence of substrate. Figure 5 shows the plot ofk′ versus [BSF] for wild-type mouse AChE. This inactivation shows saturation kinetics and the data were analyzed using eq. 1 as described under Materials and Methods.

BSF inactivation of AChE was also examined in the presence of substrate for wild-type T. californica, mouse, and both T. californica and mouse mutants, as will be described below. Table2 gives values ofk₂, K_I, andk₂/K_I for BSF inhibition in both the presence and absence of substrate. BSF reacts measurably with all enzymes assayed.

The two mutants assayed in the absence of substrate, mouse Y330F and mouse Y330A, havek₂/K_I values of 50 and 100 M⁻¹ min⁻¹, respectively. Again, these values are not significantly different from the original mouse AChEk₂/K_I. These results confirm previous reports that BSF inhibits both mammalian andT. californica AChE (Barnett and Rosenberry, 1978) and show that mouse mutations at position 330 to the residues of either T. californica AChE (Y330F) or mouse BChE (Y330A) do not affect the inactivation significantly.

BSF inactivation of AChE in the presence of substrate (acetylthiocholine) was measured as follows: product formation was measured as a function of time for AChE in the presence of BSF and the resulting data were analyzed according to eq. 5 as described underMaterials and Methods to obtain k′ andk₂/K_I values, as shown in Fig. 6.

BSF inactivation in the presence of substrate givesk₂/K_I values with smaller associated error terms than for the absence of substrate assay for both wild-type T. californica and mouse, but in both cases, the two sets of data are consistent. Wild-type T. californica and mouse AChE have fairly slow rates of inactivation by BSF (k₂/K_I = 44 and 110 M⁻¹ min⁻1, respectively). Mutations in F288 and F290 have been shown to convert the acyl pocket of AChE to that of BChE (Harel et al., 1992; Vellom et al., 1993). All such acyl pocket mutations tested (Mouse F288L and F290I and T. californica F288I/F290V) greatly accelerate the rate of inactivation of the enzyme by BSF. Although the data are associated with larger errors for the individual rate and dissociation constants, it seems that at least some of the rate enhancement comes from better binding (K_I), whereas some may also come from an accelerated sulfonylation reaction (k₂), perhaps because of better BSF positioning in the mutant enzymes. Either the larger pocket can better accommodate the bulky benzene group or better position it to attack or the active site gorge as a whole is rendered more flexible, allowing BSF in more readily.

PMSF inactivation of various AChEs was studied both in the absence and presence of substrate. The plot of k′ versus [PMSF] for wild-type mouse AChE in the absence of substrate is shown in Fig.7. Although the concentration of PMSF was not high enough to saturate the enzyme, the data could still be analyzed as above for BSF to obtain k₂ andK_I values. Data from inactivation in the presence of substrate were analyzed as for BSF as well.

The k₂/K_Ivalues for the inactivation by PMSF of wild-type T. californica and mouse AChEs, mouse mutants Y330F, Y330A, F288L, and F290I, T. californica double mutant F288L/F290V, and mouse BChE are shown in Table 3. Each enzyme was assayed in both the absence and presence of substrate.

Comparison of the reactivities of BSF and PMSF in inactivating mouse AChE (Tables 2 and 3) shows that BSF second-order inactivation constants for mouse AChEs are 10-fold less than PMSF second-order inactivation constants for the same enzyme. Despite the increased reactivity of PMSF over BSF for the mouse AChE and the close similarity in structure of the two enzymes, wild-type T. californicaAChE is inactivated by BSF, whereas PMSF has no effect on this enzyme. The object of this study was to assess the influence of various residues on this difference in reactivity.

The k₂K_Ivalues confirm the fact that wild-type T. californica AChE is not inactivated by PMSF at all. Mouse AChE, on the other hand, is inactivated by PMSF, with ak₂/K_I value of 960 M⁻¹ min⁻¹. Thek₂/K_I values for mouse Y330A and Y330F are of the same magnitude as the wild-type mouse AChE; this indicates that mutating the mouse enzyme at position 330 does not significantly affect its PMSF sensitivity. This is in contrast to the large change in specificity found for huperzine binding to the enzyme on making these same mutations (Saxena et al., 1994). PMSF must bind to the enzyme at a location different than that of huperzine.

PMSF inactivates mouse BChE with ak₂/K_I value approximately 8- to 9-fold greater than that of mouse AChE (Table 3), indicating that the large acyl pocket of BChE is important for accommodating PMSF. We therefore investigated the effects on PMSF inactivation of mutating positions 288 and 290 in both mouse andT. californica AChEs. The mutations F288L and F290I convert mouse AChE residues to those normally present in mouse BChE. If these positions are those primarily responsible for the difference in inactivation between AChE and BChE, then we would expect that mutating mouse AChE to residues present in BChE at these positions would increase inactivation rates. Indeed, thek₂/K_I values of F288L and F290I, which are 16,000 and 56,000 M⁻¹min⁻¹ in the presence of substrate, respectively, show this expected increase in rate of inactivation (Table 3). Unexpectedly, however, T. californica F288L/F290V showed a much more dramatic increase in inactivation by PMSF. In fact, these mutations caused the enzyme to change its behavior from completely insensitive to PMSF to ak₂/K_I value of 6000 M⁻¹ min⁻¹, behavior in the same range as the mouse acyl pocket mutants. We expected that the difference in sensitivity of mouse and T. californicaAChE to PMSF would be a result of differences in amino acid sequence. However, mutation of residues that were the same in both species resulted in loss of the difference in sensitivity.

PMSF Inhibition of BSF Inactivation.

It remains unclear whyT. californica AChE is insensitive to PMSF while it is inactivated by BSF. The difference in inactivation might be caused by one of two factors: 1) PMSF may be unable to bind in the T. californica active site gorge or 2) PMSF may be unable to orient within the gorge to sulfonylate the T. californica AChE. To determine which factor is responsible, T. californica AChE was assayed with both PMSF and BSF simultaneously to observe the ability of PMSF to competitively inhibit BSF inactivation. BSF (1.5 mM) and various concentrations of PMSF were added to the enzyme solution at time zero, after which the same procedure (described above) was followed to measure BSF inactivation in the absence of substrate. If PMSF cannot bind at all in the active site gorge, we would expect thek′ values for BSF to remain constant regardless of the concentration of PMSF. This is exactly what happens (data not shown), indicating that no competitive inhibition is occurring.

The results of the PMSF/BSF competition studies seem to indicate that PMSF fails to inactivate T. californica AChE because of an inability to bind in the active site gorge. This is a surprising result, because there are many hydrophobic inhibitors of AChE that are larger than PMSF that are able to enter and bind in the active site gorge. Nevertheless, if PMSF is unable to bind in the gorge, then one of the residues lining the gorge that differ between T. californica and mouse AChE should be responsible for blocking the active site. However, altering residues 288 and 290, which lie within the gorge but are the same for mouse and T. californica, results in an enzyme that is now able to accommodate PMSF productively within the gorge. This observation suggests that there may be “breathing” motions that are different between mouse and T. californica that restrict PMSF orientation in the native T. californica enzyme. These breathing motions may be less restricted in the F288L/F290V T. californica mutant, thus permitting accomodation of PMSF. The source of the differences in these motions between native mouse and T. californica enzymes may be difficult to find, because they need not be the result of differing amino acid residues in the active site gorge itself, but in more distant parts of the protein. Although we previously thought that a complete study of the active-site residues that are different in the mouse and T. californica enzymes would clarify the structural basis for the difference in PMSF specificity, we now think that looking at mutations that change the breathing characteristics of the enzyme may be more productive. Morel et al. (1999) have reported that a mutation in the T. californica enzyme, L282A (a residue that is the same in Torpedo and mouse), has decreased temperature stability (decreased enthalpy of activation for heat denaturation) that also confers increased reactivity of C231 with thiol reagents. This seems to suggest an increase in breathing of the mutant so as to allow thiol reagents access to C231 (a buried residue). We have preliminary data showing that T. californica L282A is inactivated by PMSF (manuscript in preparation). This supports the idea that the thermal instability reflects increased breathing motions and that these motions are connected with PMSF sensitivity.