- Email: [email protected]
Effects of pressure on deuterium isotope effects of yeast alcohol dehydrogenase using alternative substrates
ABB Archives of Biochemistry and Biophysics 433 (2005) 335–340 www.elsevier.com/locate/yabbi
Effects of pressure on deuterium isotope effects of yeast alcohol dehydrogenase using alternative substratesq Hyun Park, Gene Kidman, Dexter B. Northrop* Division of Pharmaceutical Sciences, School of Pharmacy, 777 Highland Avenue, University of Wisconsin-Madison, Madison, WI 53705, United States Received 11 August 2004, and in revised form 28 September 2004
Abstract Hydrostatic pressure causes biphasic effects on the oxidation of alcohols by yeast alcohol dehydrogenase as expressed on the kinetic parameter V/K which measures substrate capture. Moderate pressure increases capture by activating hydride transfer, whose transition-state must therefore have a smaller volume than the free alcohol plus the capturing form of enzyme, with DVà = 30 mL mol1 for isopropanol. A comparison of these effects with those on the oxidation of deutero-isopropanol generates a monophasic decrease in the intrinsic isotope effect; therefore, the volume of activation for the transition-state of deuteride transfer must be even more negative, by 7.6 mL mol1. The pressure data extrapolate and factor the kinetic isotope effect into a semi-classical reactant-state component, with a null value of kH/kD = 1, and a transition-state component of QH/QD = 4, suggestive of hydrogen tunneling. Pressures above 1.5 kbar decrease capture by favoring a minor conformation of enzyme which binds nicotinamide adenine dinucleotide (NAD+) less tightly. This inactive conformation has a smaller volume than active E-NAD+, with a difference of 74 mL mol1 and an equilibrium constant of 93 between them, at one atmosphere of pressure. These results are virtually identical to those obtained with benzyl alcohol [Cho and Northrop, Biochemistry 39 (2000) 2406] and give credence to this method of analysis. Moreover, qualitatively similar results with greater pressure sensitivity but less precision are obtained using ethanol as a substrate, only with pressure driving the value of the isotope effect to a value less than Dk = 1.03 directly, without extrapolation. The ethanol data verify the most surprising finding of these studies, namely that the entire kinetic isotope effect arises from a transitionstate phenomenon. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Yeast alcohol dehydrogenase (EC 1.1.1.1); Nicotinamide adenine dinucleotide
Effects of pressure on the steady-state kinetics of enzymatic reactions are numerous and complex [1] to an extent thought to be insoluble [2]. In addition to changing the rate constants associated with the a catalytic turnover, pressure can weaken ligand binding and cause denaturation of enzyme proteins by unfolding polypeptide chains and by dissociating subunits. However, changes in catalytic rate constants cause modulation of catalytic turnovers and may contain useful information about chemical and kinetic mechanisms, q
This project was supported by NSF Grant MCB 0211290. Corresponding author. Fax: +1 608 262 3397. E-mail address: [email protected] (D.B. Northrop).
*
0003-9861/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2004.09.033
whereas denaturation and weakened binding cause removal of enzyme from catalytic turnovers and have no mechanistic significance. Because isotope effects only report back from active enzyme, a distinction between modulation and removal is possible when the kinetics of pressure effects are combined with isotope effects [3]. In the first study of the effects of pressure on the kinetics of an enzymatic reaction using isotopically labeled and unlabeled substrates, both phenomena were clearly delineated [4]. In the oxidation of deuterated benzyl alcohol by yeast alcohol dehydrogenase, modulation was present in the form of a volume change assigned to the transition-state of hydride transfer, the isotopically sensitive step, and removal was present in the form
336
H. Park et al. / Archives of Biochemistry and Biophysics 433 (2005) 335–340
of a volume change assigned to the conformational change associated with the binding of NAD+ to the free enzyme, which obviously is isotopically independent. Confirming these assignments is one goal of the present study, and if successful, it then becomes possible to explore a new form of structure/activity relationship based on a physical characteristic of a specific transition-state: the volume change of the isotopically sensitive step. How that volume change differs with the isotope is the first of its kind in this new form; how it differs with an alternative substrate is the second and a subject of this paper.
Materials and methods Yeast alcohol dehydrogenase, NAD+,1 isopropyl alcohol, and ethyl alcohol were purchased from Sigma. 2-Propanol-2-d (98% D) was purchased from Cambridge Isotope Laboratories, ethanol-1-d (99.5% D) was from Isotech, and 2,4,6-collidine (trimethylpyridine) was from Aldrich. The isopropyl alcohol and ethyl alcohol were varied from 70 to 500 mM (Km = 124 mM) and from 2 to 25 mM (Km = 12 mM), respectively, at saturating concentrations of 10 mM NAD+ (Km = 1 mM). The concentration of enzyme was 0.024 lM. Trimethylpyridine buffer was used to minimize pressure-dependent changes in pH at high pressure, because it has a desirable pKa (7.39) and small volume (0.2 mL mol1) of ionization [5], at pH 6.85 to minimize pressure-induced denaturation of the enzyme [6]. High pressures were generated with a computer-controlled, automated, screw-drive pump with feedback pressure sensor obtained from Advanced Pressure Products. Samples were placed in a 1.5 mL cuvette bottle inside an ISS HP-200 high pressure cell, mounted within an Olis model DW2 Dual Beam Spectrophotometer. Assays of enzymatic activity were performed by following changes in absorbance due to formation of reduced coenzyme. NADH was followed continuously at 340 nm using a reference at 420 nm. Because several minutes were required to mix reagents and bring them to a desired pressure, a limiting level of enzyme was used such that reaction velocities could be measured over a period of 10 min or more with only a nominal change in the substrate concentration prior to actual data collection. Partial progress curves were collected at pressure and fit to an integrated form of the Michaelis–Menten equation [7] to compute the initial velocities extrapolated back to time of mixing. Sets of initial velocities in which the alcohol substrates were varied were then fit to the rate equation for a Sequential Bi Bi kinetic mechanism. Only V/K values were extracted from these fittings because, due to the 1
Abbreviations used: YADH, yeast alcohol dehydrogenase (EC 1.1.1.1); NAD+, nicotinamide adenine dinucleotide.
choices in reactant concentrations employed, values of V were accompanied by large standard errors. Standard errors on V/K values averaged less than 5%.
Theory The effect of pressure on the kinetics of capture can be described by Eq. (1) [3]: ! k1 jV =Kjp ¼ 1 þ K G=E eDV G=E p=RT ! z R0 eDV p=RT ; z z 1 þ C f eDV p=RT þ C r eðDV DV eq Þp=RT ð1Þ where k1 is the diffusion-controlled rate constant for the combination of enzyme (E) and substrate (S), R0 is the product ratio of forward and reverse enzymatic rate constants up to and including a designated isotopically sensitive step, DVà is the volume of activation between reactants E + S (e.g., E*-NAD+ and isopropanol in the current experimental design) and the transition-state of the isotopically sensitive step, p is the pressure in bar (0.98692 standard atmospheres), R is the gas constant (82.0578 mL bar mol1 K1), T is the temperature (298 K), KG/E is the equilibrium constant between free enzyme forms E and G (i.e., E*-NAD+ and E-NAD+, respectively, where the asterisk designates the conformational form of YADH that binds the nucleotide more tightly and allows the binding of the alcohol substrate), DVG/E is the volume difference between the forms, Cf and Cr are the forward and reverse commitments to catalysis, respectively, and DVeq is formally the volume change between reactants (E + S) and products (F + P), but assumed to be between S and P only. When DVeq is small, commitments may be approximated by combining them as in Eq. (2): ! ! z k1 R0 eDV p=RT : jV =Kjp ¼ z 1 þ K G=E eDV G=E p=RT 1 þ ðC f þ C r ÞeDV p=RT ð2Þ D
Given an intrinsic isotope effect on hydride transfer, k, =kH/kD, a global form of Eq. (2) takes the form ! k1 jV =Kjp ¼ 1 þ K G=E eDV G=E p=RT ! z R0 eDV p=RT ; z 1 þ C i ½D k 1 þ ðC f þ C r ÞeDV p=RT ð3Þ where Ci is the fraction of isotopic labeling. If the intrinsic isotope effect is pressure-dependent as well, Eq. (3) expands to
H. Park et al. / Archives of Biochemistry and Biophysics 433 (2005) 335–340
jV =Kjp ¼
k1 1 þ K G=E eDV G=E p=RT
!
! z R0 eDV p=RT ; z 1 þ C i ½D kðD Q 1ÞeDV Q p=RT þ D k 1 þ ðC f þ C r ÞeDV p=RT ð4Þ
D
where Q is the ratio of Bell tunneling corrections, QH/QD [8], and represents the portion of an isotope effect originating in the transition-state and DVQ is the apparent volume difference between a hydride and deuteride transfer. When commitments are small, Eq. (4) reduces to ! k1 jV =Kjp ¼ 1 þ K G=E eDV G=E p=RT ! z R0 eDV p=RT : 1 þ C i ½D kðD Q 1ÞeDV Q p=RT þ D k 1 ð5Þ At atmospheric pressure and given a small KG/E, k1R0 will equal the value of V/K in Eq. (5) and it can be replaced by an empirical constant during regression analysis.
Results Fig. 1A shows the biphasic pressure dependence of the capture of isopropanol and deutero-isopropanol. Fig. 1B shows the ratio of the two sets of data points
Fig. 1. Effect of pressure on the capture of isopropanol (m) and deutero-isopropanol (j) by yeast alcohol dehydrogenase and on deuterium isotope effects (d). Values of V/K were fit to Eqs. (4) or (5) to determine the solid lines in both panels, representing pressure effects on hydride transfer, on the availability of enzyme for capture and on the intrinsic isotope effect.
337
which takes the form of a monophasic decrease in the deuterium isotope effect. The data are qualitatively similar to those obtained previously with benzyl alcohol as a substrate [4]. The curves for capture are nearly symmetrically bell-shaped with a maximum near 1.5 kbar, and the curve for isotopic effects decreases to approximately half the starting value at 2.25 kbar. The data in Fig. 1A were subjected to regression analysis using a BASIC computer program employing the nonlinear regression routine of Duggleby [9]. An attempt to fit the data to Eq. (4) did not converge, apparently because of excessive covariance, with the largest error focused on Cf + Cr at 0.04 ± 1.8. Given the small value for the commitment to catalysis, the data were then fit to Eq. (5) which has no commitments. The regression converged to the fitted parameters listed in the left side of Table 1 and these generated the solid curves shown in both panels. Also shown in the right side of Table 1 are the published values of parameters for benzyl alcohol [4]. Notable in a comparison of the kinetic parameters for the two alcohols are the equivalencies of KG/E and DVG/E. Because these precede the binding of the labeled substrate, one might expect them to be independent of the substrate and, indeed, these equivalencies provide evidence that the assignment of a negative pressure function governing substrate capture to a change in the capturing form of enzyme is correct. Assuming the equivalencies are valid, the data were fit to Eq. (4) a second time with KG/E and DVG/E fixed as constants using the values obtained with benzyl alcohol. The regression analysis converged to the values shown in the center of Table 1, none of which were significantly different from those on the left, and verifying the low value for the commitments to catalysis. The results with isopropanol and benzyl alcohol are quantitatively similar as well, but not identical. The isotope effect again extrapolates to a null result at high pressure, with the semi-classical isotope effect, Dk, not significantly different from one, leaving the entire isotope effect attributed to a change in the tunneling correction, DQ. The pressure dependence is somewhat less sensitive with isopropanol, as indicated by the lower DVQ, as is the activation volume, DVà. Fig. 2A shows the biphasic pressure dependence of the capture of ethanol and deutero-ethanol. Fig. 2B shows the ratio of the two sets of data points which again takes the form of a monophasic decrease in the deuterium isotope effect. The data are qualitatively different from those obtained previously, with the curves for capture skewed to the left with a maximum between 0.5 and 1.0 kbar, and the curve for isotopic effects decreasing nearly to one directly, without extrapolation. The data in Fig. 2A were fit to Eq. (4), the regression analysis converged to the fitted parameters listed in the left side of Table 2 and these generated the solid curves shown in both panels. Despite the successful
338
H. Park et al. / Archives of Biochemistry and Biophysics 433 (2005) 335–340
Table 1 Pressure effects and isotope effects on the capture of isopropanol by YADHa D
k Q DV zH DVQ KG/E DVG/E Cf + Cr D
0.97 ± 0.04 4.0 ± 0.3 29.6 ± 2.0 mL mol1 7.7 ± 2.1 mL mol1 0.011 ± 0.003 74 ± 2 mL mol1 —
0.95 ± 0.03 4.0 ± 0.4 29.5 ± 0.5 mL mol1 7.6 ± 2.2 mL mol1 (0.012) (72.5 mL mol1) <0.001 ± 0.3
[0.99 ± 0.03] [4.99 ± 0.37] [38 ± 1 mL mol1] [10.4 ± 1.5 mL mol1] [0.012 ± 0.002] [72.5 ± 1.4 mL mol] [<0.001 ± 10]
a Data on the left were obtained by a fit of Fig. 1A to Eq. (5). Data in the middle were obtained by a fit of Fig. 1A to Eq. (4) assuming values for KG/E and VG/E (shown in parentheses) were constants equal to those obtained using benzyl alcohol. Data on the right were obtained with benzyl alcohol [4] and are listed for comparison purposes (shown in brackets).
DVG/E to values obtained with benzyl alcohol, but this attempt failed to converge, unlike the regression with data for isopropanol. Because the value of the isotope effect approached a measured value of one at high pressure, Dk was fixed at this value and the attempt repeated, but failed to improve the analysis, as did other manipulations. These exercises suggest that while the sum of the commitments is low, they are significant, and the data lack sufficient precision to escape the limitations of covariance when an isotope effect decreases for two reasons. Nevertheless, the remaining parameters for ethanol are significantly different from those obtained with the other alcohols, with the exception of the volume change associated with the conformational change, DVG/E. The value of the tunneling correction, DQ, is smaller, but its sensitivity to pressure is clearly much greater, as indicated by the larger value of DVQ. Most surprising is the larger value for KG/E which describes the conformational change that precedes binding of the alcohol. Fig. 2. Effect of pressure on the capture of ethanol (m) and deuteroethanol (j) by yeast alcohol dehydrogenase and on deuterium isotope effects (d). Values of V/K were fit to Eqs. (4) or (5) to determine the solid lines in both panels, representing pressure effects on hydride transfer, on the availability of enzyme for capture and on the intrinsic isotope effect.
convergence, standard errors were rather large. A fit to Eq. (5) was also successful, returning parameter values with smaller standard errors shown in the middle of Table 2. Repeating the process followed in Table 1, an attempt was made to fit the data to Eq. (4) fixing KG/E and
Discussion The results in Tables 1 and 2 are reassuring, and suggest that this experimental design can return significant measures of new kinetic parameters with accurate assignments. Particularly satisfying are the characterizations of the conformational change associated with the binding of NAD+. The agreement between values of KG/E and DVG/E in Table 1 is particularly good, and serves to validate their assignment. That assignment is further validated by independent studies on the effects of pressure on
Table 2 Pressure effects and isotope effects on the capture of ethanol by YADHa D
k Q DV zH DVQ KG/E DVG/E Cf + Cr D
a
0.74 ± 0.29 3.0 ± 0.8 84 ± 61 mL mol1 27 ± 54 mL mol1 0.37 ± 0.55 97 ± 31 mL mol1 0.12 ± 0.78
0.71 ± 0.13 2.9 ± 0.6 74 ± 31 mL mol1 38 ± 13 mL mol1 0.43 ± 0.23 95 ± 25 mL mol1 —
[0.99 ± 0.03] [4.99 ± 0.37] [38 ± 1 mL mol1] [10.4 ± 1.5 mL mol1] [0.012 ± 0.002] [72.5 ± 1.4 mL mol] [<0.001 ± 10]
Data on the left were obtained by a fit of Fig. 1A to Eq. (4). Data in the middle were obtained by a fit of Fig. 1A to Eq. (5). Data on the right were obtained with benzyl alcohol [4] and are listed for comparison purposes (shown in brackets).
H. Park et al. / Archives of Biochemistry and Biophysics 433 (2005) 335–340
the spectrum of enzyme-bound NADH. Binding to yeast alcohol dehydrogenase results in a hypsochromic shift of the NADH absorbance maximum at 340 nm [10]. Application of high hydrostatic pressure to the enzyme–nucleotide complex returns the absorbance maximum to longer wavelengths [11]. A different value for the equilibrium constant is found with ethanol, the natural substrate for this enzyme, but that does not particularly detract from this finding because the commercial preparations of YADH contain at least two isozymes. Benzyl alcohol is oxidized almost exclusively by the type II isozyme, while ethanol is preferentially oxidized by the type I isozyme (Bryce Plapp, personal communication). The results are also reassuring in that they confirm the total pressure dependence of the isotope effect found in the first report [4]. Moreover, the results with ethanol do so without a long and therefore dubious extrapolation. The absence of a semi-classical component of the isotope effect arising from zero point energy differences takes on greater significance by this agreement with three alcohols. It suggests that the origin of the isotope effect in an enzymatic reaction somehow differs from that of a purely chemical reaction. In the latter, the semi-classical component is present and is larger than the transition-state component, and the isotope effect asymptotically approaches the former value at high pressure [12]. One explanation of the absence is that the isotope effect is not arising from zero point energies and tunneling phenomena, but rather originates in a mechanical model of catalysis that reflects protein motion [4]. In this model, the mass difference between a hydride and a deuteride is expressed in atomic movements throughout the enzyme protein and the volumes of activation are physically dispersed accordingly. The differences in the pressure dependencies with different alcohols are new information that are open to interpretation, but portends great utility as the parameters reflect physical properties of the transition-state heretofore not accessible. However, in the present case, that utility is compromised by the presence of isozymes. Difficulties with the regression analysis on the ethanol data were disappointing. As it seems likely that the commitments to catalysis are small but significant, these difficulties suggest that it may not be possible to use this experimental design to determine intrinsic isotope effects when moderate to large commitments are present. If the isotope effect is decreasing due to a pressure effect on hydrogen tunneling plus an increase in commitments to catalysis, the regression analysis appears unable to sort them out due to extreme covariance. Of course, in the absence of any pressure effect directly on the intrinsic isotope effect, and presence of changes in the isotope effect due solely to changing commitments, then separation and characterization of commitments to catalysis will likely be possible. As an example, similar experiments with formate dehydrogenase returned an intrinsic
339
Fig. 3. Simulation of the effect of pressure on the expression of an intrinsic isotope effect due to commitments to catalysis that increase with increasing pressure. A value of Dk = 5 was assumed using Eq. (3). (see text).
isotope effect that was either pressure-independent or increased slightly with increasing pressure [13]. Changes in the expression of an isotope effect as a result of changing in the commitments to catalysis is a sigmoidal function. As illustrated in Fig. 3 for the condition when commitments to catalysis increase with increasing pressure and thereby decrease the observed isotope effect, then the ability to successfully fit data to Eq. (3) depends upon the starting value of the commitments to catalysis at atmospheric pressure, or in a practical sense, where the data collection begins before pressure is increased. If atmospheric pressure is at point A, then data fitting, and in effect, extrapolation to the intrinsic isotope effect (indicated by the upper dashed line) should converge successfully. If at point B, it is less likely, but at the inflection point C and below it will not be possible for the regression to locate the upper asymptote. Alternatively, if commitments to catalysis decrease with increasing pressure, then the sigmoidal curve will also increase (not shown) and extracting an intrinsic isotope effect will be determined solely by how high a pressure can be reached with laboratory equipment. An example of this latter case has yet to be found.
Note added in proof Support for this model was recently obtained in the form of a pressure dependence of a 13C isotope effect in the oxidation of benzyl alcohol by YADH, as heavy atom isotope effects are thought not to have a tunneling component (H. Park and D.B. Northrop, unpublished results).
References [1] K.L. Laidler, P.S. Bunting, The Chemical Kinetics of Enzyme Action, second ed., Clarendon Press, Oxford, 1973, pp. 220–232.
340
H. Park et al. / Archives of Biochemistry and Biophysics 433 (2005) 335–340
[2] C. Balny, in: J.A. Markely, C.A. Royer, D.B. Northrop (Eds.), High Pressure Effects in Molecular Biology and Enzymology, Oxford University Press, New York, NY, 1996, pp. 196–210. [3] D.B. Northrop, Biochem. Biophys. Acta 1595 (2002) 71–79. [4] D.B. Northrop, Y.-K. Cho, Biochemistry 39 (2000) 2406–2412. [5] Y. Kitamura, T. Itoh, J. Solution Chem. 16 (1987) 715–725. [6] G. Kidman, H. Park, D.B. Northrop, Protein Pept. Lett. 11 (2004) 543–546.
[7] R.G. Duggleby, Biochem. J. 228 (1985) 55–60. [8] R.P. Bell, The Tunnel Effect in Chemistry, Chapman & Hall, New York, NY, 1980. [9] R.G. Duggleby, Comput. Biol. Med. 14 (1984) 447–455. [10] A.J. Sytkowski, Arch. Biochem. Biophys. 185 (1977) 505–517. [11] G. Kidman, D.B. Northrop, Protein Pept. Lett. (2005) in press. [12] D.B. Northrop, J. Am. Chem. Soc. 121 (1999) 3521–3524. [13] D.J. Quirk, D.B. Northrop, Biochemistry 40 (2001) 847–851.
Effects of pressure on deuterium isotope effects of yeast alcohol dehydrogenase using alternative substratesq Hyun Park, Gene Kidman, Dexter B. Northrop* Division of Pharmaceutical Sciences, School of Pharmacy, 777 Highland Avenue, University of Wisconsin-Madison, Madison, WI 53705, United States Received 11 August 2004, and in revised form 28 September 2004
Abstract Hydrostatic pressure causes biphasic effects on the oxidation of alcohols by yeast alcohol dehydrogenase as expressed on the kinetic parameter V/K which measures substrate capture. Moderate pressure increases capture by activating hydride transfer, whose transition-state must therefore have a smaller volume than the free alcohol plus the capturing form of enzyme, with DVà = 30 mL mol1 for isopropanol. A comparison of these effects with those on the oxidation of deutero-isopropanol generates a monophasic decrease in the intrinsic isotope effect; therefore, the volume of activation for the transition-state of deuteride transfer must be even more negative, by 7.6 mL mol1. The pressure data extrapolate and factor the kinetic isotope effect into a semi-classical reactant-state component, with a null value of kH/kD = 1, and a transition-state component of QH/QD = 4, suggestive of hydrogen tunneling. Pressures above 1.5 kbar decrease capture by favoring a minor conformation of enzyme which binds nicotinamide adenine dinucleotide (NAD+) less tightly. This inactive conformation has a smaller volume than active E-NAD+, with a difference of 74 mL mol1 and an equilibrium constant of 93 between them, at one atmosphere of pressure. These results are virtually identical to those obtained with benzyl alcohol [Cho and Northrop, Biochemistry 39 (2000) 2406] and give credence to this method of analysis. Moreover, qualitatively similar results with greater pressure sensitivity but less precision are obtained using ethanol as a substrate, only with pressure driving the value of the isotope effect to a value less than Dk = 1.03 directly, without extrapolation. The ethanol data verify the most surprising finding of these studies, namely that the entire kinetic isotope effect arises from a transitionstate phenomenon. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Yeast alcohol dehydrogenase (EC 1.1.1.1); Nicotinamide adenine dinucleotide
Effects of pressure on the steady-state kinetics of enzymatic reactions are numerous and complex [1] to an extent thought to be insoluble [2]. In addition to changing the rate constants associated with the a catalytic turnover, pressure can weaken ligand binding and cause denaturation of enzyme proteins by unfolding polypeptide chains and by dissociating subunits. However, changes in catalytic rate constants cause modulation of catalytic turnovers and may contain useful information about chemical and kinetic mechanisms, q
This project was supported by NSF Grant MCB 0211290. Corresponding author. Fax: +1 608 262 3397. E-mail address: [email protected] (D.B. Northrop).
*
0003-9861/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2004.09.033
whereas denaturation and weakened binding cause removal of enzyme from catalytic turnovers and have no mechanistic significance. Because isotope effects only report back from active enzyme, a distinction between modulation and removal is possible when the kinetics of pressure effects are combined with isotope effects [3]. In the first study of the effects of pressure on the kinetics of an enzymatic reaction using isotopically labeled and unlabeled substrates, both phenomena were clearly delineated [4]. In the oxidation of deuterated benzyl alcohol by yeast alcohol dehydrogenase, modulation was present in the form of a volume change assigned to the transition-state of hydride transfer, the isotopically sensitive step, and removal was present in the form
336
H. Park et al. / Archives of Biochemistry and Biophysics 433 (2005) 335–340
of a volume change assigned to the conformational change associated with the binding of NAD+ to the free enzyme, which obviously is isotopically independent. Confirming these assignments is one goal of the present study, and if successful, it then becomes possible to explore a new form of structure/activity relationship based on a physical characteristic of a specific transition-state: the volume change of the isotopically sensitive step. How that volume change differs with the isotope is the first of its kind in this new form; how it differs with an alternative substrate is the second and a subject of this paper.
Materials and methods Yeast alcohol dehydrogenase, NAD+,1 isopropyl alcohol, and ethyl alcohol were purchased from Sigma. 2-Propanol-2-d (98% D) was purchased from Cambridge Isotope Laboratories, ethanol-1-d (99.5% D) was from Isotech, and 2,4,6-collidine (trimethylpyridine) was from Aldrich. The isopropyl alcohol and ethyl alcohol were varied from 70 to 500 mM (Km = 124 mM) and from 2 to 25 mM (Km = 12 mM), respectively, at saturating concentrations of 10 mM NAD+ (Km = 1 mM). The concentration of enzyme was 0.024 lM. Trimethylpyridine buffer was used to minimize pressure-dependent changes in pH at high pressure, because it has a desirable pKa (7.39) and small volume (0.2 mL mol1) of ionization [5], at pH 6.85 to minimize pressure-induced denaturation of the enzyme [6]. High pressures were generated with a computer-controlled, automated, screw-drive pump with feedback pressure sensor obtained from Advanced Pressure Products. Samples were placed in a 1.5 mL cuvette bottle inside an ISS HP-200 high pressure cell, mounted within an Olis model DW2 Dual Beam Spectrophotometer. Assays of enzymatic activity were performed by following changes in absorbance due to formation of reduced coenzyme. NADH was followed continuously at 340 nm using a reference at 420 nm. Because several minutes were required to mix reagents and bring them to a desired pressure, a limiting level of enzyme was used such that reaction velocities could be measured over a period of 10 min or more with only a nominal change in the substrate concentration prior to actual data collection. Partial progress curves were collected at pressure and fit to an integrated form of the Michaelis–Menten equation [7] to compute the initial velocities extrapolated back to time of mixing. Sets of initial velocities in which the alcohol substrates were varied were then fit to the rate equation for a Sequential Bi Bi kinetic mechanism. Only V/K values were extracted from these fittings because, due to the 1
Abbreviations used: YADH, yeast alcohol dehydrogenase (EC 1.1.1.1); NAD+, nicotinamide adenine dinucleotide.
choices in reactant concentrations employed, values of V were accompanied by large standard errors. Standard errors on V/K values averaged less than 5%.
Theory The effect of pressure on the kinetics of capture can be described by Eq. (1) [3]: ! k1 jV =Kjp ¼ 1 þ K G=E eDV G=E p=RT ! z R0 eDV p=RT ; z z 1 þ C f eDV p=RT þ C r eðDV DV eq Þp=RT ð1Þ where k1 is the diffusion-controlled rate constant for the combination of enzyme (E) and substrate (S), R0 is the product ratio of forward and reverse enzymatic rate constants up to and including a designated isotopically sensitive step, DVà is the volume of activation between reactants E + S (e.g., E*-NAD+ and isopropanol in the current experimental design) and the transition-state of the isotopically sensitive step, p is the pressure in bar (0.98692 standard atmospheres), R is the gas constant (82.0578 mL bar mol1 K1), T is the temperature (298 K), KG/E is the equilibrium constant between free enzyme forms E and G (i.e., E*-NAD+ and E-NAD+, respectively, where the asterisk designates the conformational form of YADH that binds the nucleotide more tightly and allows the binding of the alcohol substrate), DVG/E is the volume difference between the forms, Cf and Cr are the forward and reverse commitments to catalysis, respectively, and DVeq is formally the volume change between reactants (E + S) and products (F + P), but assumed to be between S and P only. When DVeq is small, commitments may be approximated by combining them as in Eq. (2): ! ! z k1 R0 eDV p=RT : jV =Kjp ¼ z 1 þ K G=E eDV G=E p=RT 1 þ ðC f þ C r ÞeDV p=RT ð2Þ D
Given an intrinsic isotope effect on hydride transfer, k, =kH/kD, a global form of Eq. (2) takes the form ! k1 jV =Kjp ¼ 1 þ K G=E eDV G=E p=RT ! z R0 eDV p=RT ; z 1 þ C i ½D k 1 þ ðC f þ C r ÞeDV p=RT ð3Þ where Ci is the fraction of isotopic labeling. If the intrinsic isotope effect is pressure-dependent as well, Eq. (3) expands to
H. Park et al. / Archives of Biochemistry and Biophysics 433 (2005) 335–340
jV =Kjp ¼
k1 1 þ K G=E eDV G=E p=RT
!
! z R0 eDV p=RT ; z 1 þ C i ½D kðD Q 1ÞeDV Q p=RT þ D k 1 þ ðC f þ C r ÞeDV p=RT ð4Þ
D
where Q is the ratio of Bell tunneling corrections, QH/QD [8], and represents the portion of an isotope effect originating in the transition-state and DVQ is the apparent volume difference between a hydride and deuteride transfer. When commitments are small, Eq. (4) reduces to ! k1 jV =Kjp ¼ 1 þ K G=E eDV G=E p=RT ! z R0 eDV p=RT : 1 þ C i ½D kðD Q 1ÞeDV Q p=RT þ D k 1 ð5Þ At atmospheric pressure and given a small KG/E, k1R0 will equal the value of V/K in Eq. (5) and it can be replaced by an empirical constant during regression analysis.
Results Fig. 1A shows the biphasic pressure dependence of the capture of isopropanol and deutero-isopropanol. Fig. 1B shows the ratio of the two sets of data points
Fig. 1. Effect of pressure on the capture of isopropanol (m) and deutero-isopropanol (j) by yeast alcohol dehydrogenase and on deuterium isotope effects (d). Values of V/K were fit to Eqs. (4) or (5) to determine the solid lines in both panels, representing pressure effects on hydride transfer, on the availability of enzyme for capture and on the intrinsic isotope effect.
337
which takes the form of a monophasic decrease in the deuterium isotope effect. The data are qualitatively similar to those obtained previously with benzyl alcohol as a substrate [4]. The curves for capture are nearly symmetrically bell-shaped with a maximum near 1.5 kbar, and the curve for isotopic effects decreases to approximately half the starting value at 2.25 kbar. The data in Fig. 1A were subjected to regression analysis using a BASIC computer program employing the nonlinear regression routine of Duggleby [9]. An attempt to fit the data to Eq. (4) did not converge, apparently because of excessive covariance, with the largest error focused on Cf + Cr at 0.04 ± 1.8. Given the small value for the commitment to catalysis, the data were then fit to Eq. (5) which has no commitments. The regression converged to the fitted parameters listed in the left side of Table 1 and these generated the solid curves shown in both panels. Also shown in the right side of Table 1 are the published values of parameters for benzyl alcohol [4]. Notable in a comparison of the kinetic parameters for the two alcohols are the equivalencies of KG/E and DVG/E. Because these precede the binding of the labeled substrate, one might expect them to be independent of the substrate and, indeed, these equivalencies provide evidence that the assignment of a negative pressure function governing substrate capture to a change in the capturing form of enzyme is correct. Assuming the equivalencies are valid, the data were fit to Eq. (4) a second time with KG/E and DVG/E fixed as constants using the values obtained with benzyl alcohol. The regression analysis converged to the values shown in the center of Table 1, none of which were significantly different from those on the left, and verifying the low value for the commitments to catalysis. The results with isopropanol and benzyl alcohol are quantitatively similar as well, but not identical. The isotope effect again extrapolates to a null result at high pressure, with the semi-classical isotope effect, Dk, not significantly different from one, leaving the entire isotope effect attributed to a change in the tunneling correction, DQ. The pressure dependence is somewhat less sensitive with isopropanol, as indicated by the lower DVQ, as is the activation volume, DVà. Fig. 2A shows the biphasic pressure dependence of the capture of ethanol and deutero-ethanol. Fig. 2B shows the ratio of the two sets of data points which again takes the form of a monophasic decrease in the deuterium isotope effect. The data are qualitatively different from those obtained previously, with the curves for capture skewed to the left with a maximum between 0.5 and 1.0 kbar, and the curve for isotopic effects decreasing nearly to one directly, without extrapolation. The data in Fig. 2A were fit to Eq. (4), the regression analysis converged to the fitted parameters listed in the left side of Table 2 and these generated the solid curves shown in both panels. Despite the successful
338
H. Park et al. / Archives of Biochemistry and Biophysics 433 (2005) 335–340
Table 1 Pressure effects and isotope effects on the capture of isopropanol by YADHa D
k Q DV zH DVQ KG/E DVG/E Cf + Cr D
0.97 ± 0.04 4.0 ± 0.3 29.6 ± 2.0 mL mol1 7.7 ± 2.1 mL mol1 0.011 ± 0.003 74 ± 2 mL mol1 —
0.95 ± 0.03 4.0 ± 0.4 29.5 ± 0.5 mL mol1 7.6 ± 2.2 mL mol1 (0.012) (72.5 mL mol1) <0.001 ± 0.3
[0.99 ± 0.03] [4.99 ± 0.37] [38 ± 1 mL mol1] [10.4 ± 1.5 mL mol1] [0.012 ± 0.002] [72.5 ± 1.4 mL mol] [<0.001 ± 10]
a Data on the left were obtained by a fit of Fig. 1A to Eq. (5). Data in the middle were obtained by a fit of Fig. 1A to Eq. (4) assuming values for KG/E and VG/E (shown in parentheses) were constants equal to those obtained using benzyl alcohol. Data on the right were obtained with benzyl alcohol [4] and are listed for comparison purposes (shown in brackets).
DVG/E to values obtained with benzyl alcohol, but this attempt failed to converge, unlike the regression with data for isopropanol. Because the value of the isotope effect approached a measured value of one at high pressure, Dk was fixed at this value and the attempt repeated, but failed to improve the analysis, as did other manipulations. These exercises suggest that while the sum of the commitments is low, they are significant, and the data lack sufficient precision to escape the limitations of covariance when an isotope effect decreases for two reasons. Nevertheless, the remaining parameters for ethanol are significantly different from those obtained with the other alcohols, with the exception of the volume change associated with the conformational change, DVG/E. The value of the tunneling correction, DQ, is smaller, but its sensitivity to pressure is clearly much greater, as indicated by the larger value of DVQ. Most surprising is the larger value for KG/E which describes the conformational change that precedes binding of the alcohol. Fig. 2. Effect of pressure on the capture of ethanol (m) and deuteroethanol (j) by yeast alcohol dehydrogenase and on deuterium isotope effects (d). Values of V/K were fit to Eqs. (4) or (5) to determine the solid lines in both panels, representing pressure effects on hydride transfer, on the availability of enzyme for capture and on the intrinsic isotope effect.
convergence, standard errors were rather large. A fit to Eq. (5) was also successful, returning parameter values with smaller standard errors shown in the middle of Table 2. Repeating the process followed in Table 1, an attempt was made to fit the data to Eq. (4) fixing KG/E and
Discussion The results in Tables 1 and 2 are reassuring, and suggest that this experimental design can return significant measures of new kinetic parameters with accurate assignments. Particularly satisfying are the characterizations of the conformational change associated with the binding of NAD+. The agreement between values of KG/E and DVG/E in Table 1 is particularly good, and serves to validate their assignment. That assignment is further validated by independent studies on the effects of pressure on
Table 2 Pressure effects and isotope effects on the capture of ethanol by YADHa D
k Q DV zH DVQ KG/E DVG/E Cf + Cr D
a
0.74 ± 0.29 3.0 ± 0.8 84 ± 61 mL mol1 27 ± 54 mL mol1 0.37 ± 0.55 97 ± 31 mL mol1 0.12 ± 0.78
0.71 ± 0.13 2.9 ± 0.6 74 ± 31 mL mol1 38 ± 13 mL mol1 0.43 ± 0.23 95 ± 25 mL mol1 —
[0.99 ± 0.03] [4.99 ± 0.37] [38 ± 1 mL mol1] [10.4 ± 1.5 mL mol1] [0.012 ± 0.002] [72.5 ± 1.4 mL mol] [<0.001 ± 10]
Data on the left were obtained by a fit of Fig. 1A to Eq. (4). Data in the middle were obtained by a fit of Fig. 1A to Eq. (5). Data on the right were obtained with benzyl alcohol [4] and are listed for comparison purposes (shown in brackets).
H. Park et al. / Archives of Biochemistry and Biophysics 433 (2005) 335–340
the spectrum of enzyme-bound NADH. Binding to yeast alcohol dehydrogenase results in a hypsochromic shift of the NADH absorbance maximum at 340 nm [10]. Application of high hydrostatic pressure to the enzyme–nucleotide complex returns the absorbance maximum to longer wavelengths [11]. A different value for the equilibrium constant is found with ethanol, the natural substrate for this enzyme, but that does not particularly detract from this finding because the commercial preparations of YADH contain at least two isozymes. Benzyl alcohol is oxidized almost exclusively by the type II isozyme, while ethanol is preferentially oxidized by the type I isozyme (Bryce Plapp, personal communication). The results are also reassuring in that they confirm the total pressure dependence of the isotope effect found in the first report [4]. Moreover, the results with ethanol do so without a long and therefore dubious extrapolation. The absence of a semi-classical component of the isotope effect arising from zero point energy differences takes on greater significance by this agreement with three alcohols. It suggests that the origin of the isotope effect in an enzymatic reaction somehow differs from that of a purely chemical reaction. In the latter, the semi-classical component is present and is larger than the transition-state component, and the isotope effect asymptotically approaches the former value at high pressure [12]. One explanation of the absence is that the isotope effect is not arising from zero point energies and tunneling phenomena, but rather originates in a mechanical model of catalysis that reflects protein motion [4]. In this model, the mass difference between a hydride and a deuteride is expressed in atomic movements throughout the enzyme protein and the volumes of activation are physically dispersed accordingly. The differences in the pressure dependencies with different alcohols are new information that are open to interpretation, but portends great utility as the parameters reflect physical properties of the transition-state heretofore not accessible. However, in the present case, that utility is compromised by the presence of isozymes. Difficulties with the regression analysis on the ethanol data were disappointing. As it seems likely that the commitments to catalysis are small but significant, these difficulties suggest that it may not be possible to use this experimental design to determine intrinsic isotope effects when moderate to large commitments are present. If the isotope effect is decreasing due to a pressure effect on hydrogen tunneling plus an increase in commitments to catalysis, the regression analysis appears unable to sort them out due to extreme covariance. Of course, in the absence of any pressure effect directly on the intrinsic isotope effect, and presence of changes in the isotope effect due solely to changing commitments, then separation and characterization of commitments to catalysis will likely be possible. As an example, similar experiments with formate dehydrogenase returned an intrinsic
339
Fig. 3. Simulation of the effect of pressure on the expression of an intrinsic isotope effect due to commitments to catalysis that increase with increasing pressure. A value of Dk = 5 was assumed using Eq. (3). (see text).
isotope effect that was either pressure-independent or increased slightly with increasing pressure [13]. Changes in the expression of an isotope effect as a result of changing in the commitments to catalysis is a sigmoidal function. As illustrated in Fig. 3 for the condition when commitments to catalysis increase with increasing pressure and thereby decrease the observed isotope effect, then the ability to successfully fit data to Eq. (3) depends upon the starting value of the commitments to catalysis at atmospheric pressure, or in a practical sense, where the data collection begins before pressure is increased. If atmospheric pressure is at point A, then data fitting, and in effect, extrapolation to the intrinsic isotope effect (indicated by the upper dashed line) should converge successfully. If at point B, it is less likely, but at the inflection point C and below it will not be possible for the regression to locate the upper asymptote. Alternatively, if commitments to catalysis decrease with increasing pressure, then the sigmoidal curve will also increase (not shown) and extracting an intrinsic isotope effect will be determined solely by how high a pressure can be reached with laboratory equipment. An example of this latter case has yet to be found.
Note added in proof Support for this model was recently obtained in the form of a pressure dependence of a 13C isotope effect in the oxidation of benzyl alcohol by YADH, as heavy atom isotope effects are thought not to have a tunneling component (H. Park and D.B. Northrop, unpublished results).
References [1] K.L. Laidler, P.S. Bunting, The Chemical Kinetics of Enzyme Action, second ed., Clarendon Press, Oxford, 1973, pp. 220–232.
340
H. Park et al. / Archives of Biochemistry and Biophysics 433 (2005) 335–340
[2] C. Balny, in: J.A. Markely, C.A. Royer, D.B. Northrop (Eds.), High Pressure Effects in Molecular Biology and Enzymology, Oxford University Press, New York, NY, 1996, pp. 196–210. [3] D.B. Northrop, Biochem. Biophys. Acta 1595 (2002) 71–79. [4] D.B. Northrop, Y.-K. Cho, Biochemistry 39 (2000) 2406–2412. [5] Y. Kitamura, T. Itoh, J. Solution Chem. 16 (1987) 715–725. [6] G. Kidman, H. Park, D.B. Northrop, Protein Pept. Lett. 11 (2004) 543–546.
[7] R.G. Duggleby, Biochem. J. 228 (1985) 55–60. [8] R.P. Bell, The Tunnel Effect in Chemistry, Chapman & Hall, New York, NY, 1980. [9] R.G. Duggleby, Comput. Biol. Med. 14 (1984) 447–455. [10] A.J. Sytkowski, Arch. Biochem. Biophys. 185 (1977) 505–517. [11] G. Kidman, D.B. Northrop, Protein Pept. Lett. (2005) in press. [12] D.B. Northrop, J. Am. Chem. Soc. 121 (1999) 3521–3524. [13] D.J. Quirk, D.B. Northrop, Biochemistry 40 (2001) 847–851.