- Email: [email protected]
Amino acid residues involved in the catalytic mechanism of NAD-dependent glutamate dehydrogenase from Halobacterium salinarum
Biochimica et Biophysica Acta 1426 (1999) 513^525
Amino acid residues involved in the catalytic mechanism of NAD-dependent glutamate dehydrogenase from Halobacterium salinarum Francisco Pe¨rez-Pomares, Juan Ferrer, Mo¨nica Camacho, Carmen Pire, Francisco LLorca, Mar|¨a Jose¨ Bonete * Divisio¨n de Bioqu|¨mica, Facultad de Ciencias, Universidad de Alicante, Ap. 99, E-03080 Alicante, Spain Received 5 August 1998; received in revised form 27 October 1998; accepted 10 December 1998
Abstract The pH dependence of kinetic parameters for a competitive inhibitor (glutarate) was determined in order to obtain information on the chemical mechanism for NAD-dependent glutamate dehydrogenase from Halobacterium salinarum. The maximum velocity is pH dependent, decreasing at low pHs giving a pK value of 7.19 þ 0.13, while the V/K for L-glutamate at 30³C decreases at low and high pHs, yielding pK values of 7.9 þ 0.2 and 9.8 þ 0.2, respectively. The glutarate pKis profile decreases at high pHs, yielding a pK of 9.59 þ 0.09 at 30³C. The values of ionization heat calculated from the change in pK with temperature are: 1.19U104 , 5.7U103 , 7U103 , 6.6U103 cal mol31 , for the residues involved. All these data suggest that the groups required for catalysis and/or binding are lysine, histidine and tyrosine. The enzyme shows a time-dependent loss in glutamate oxidation activity when incubated with diethyl pyrocarbonate (DEPC). Inactivation follows pseudo-first-order kinetics with a second-order rate constant of 53 M31 min31 . The pKa of the titratable group was pK1 = 6.6 þ 0.6. Inactivation with ethyl acetimidate also shows pseudo-first-order kinetics as well as inactivation with TNM yielding second-order constants of 1.2 M31 min31 and 2.8 M31 min31 , and pKa s of 8.36 and 9.0, respectively. The proposed mechanism involves hydrogen binding of each of the two carboxylic groups to tyrosyl residues; histidine interacts with one of the N-hydrogens of the L-glutamate amino group. We also corroborate the presence of a conservative lysine that has a remarkable ability to coordinate a water molecule that would act as general base. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Archaea; NAD-glutamate dehydrogenase; Chemical modi¢cation; (Halobacterium salinarum)
1. Introduction An enzyme must interact with substrates; so, understanding enzyme mechanisms requires underAbbreviations: NAD-GDH, NAD-dependent glutamate dehydrogenase; DEPC, diethyl pyrocarbonate; TNM, tetranitromethane * Corresponding author. Fax: +34-96-590-3464; E-mail: [email protected]
standing those interactions taking place between an enzyme and a substrate that lead to transition-state stabilization, as well as of the residues involved in the catalytic process. Useful information can be obtained by organic chemistry and kinetic methods, including the use of structurally related compounds, which affect the function of the enzyme, with the aim of characterizing the binding sites and the possible role, played by the residues involved in the catalytic pathways.
0304-4165 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 9 8 ) 0 0 1 7 4 - 3
BBAGEN 24760 28-1-99
514
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
Glutamate dehydrogenases (GDH) (EC 1.4.1.2^4) are a family of enzymes that catalyze the reversible oxidative deamination of L-glutamate using NAD , NADP or both as coenzymes [1]. Although the metabolic importance and regulatory properties of the GDH species have attracted considerable interest, the three-dimensional structure of any GDH became available only few years ago [2]. The NAD-dependent GDH from Clostridium symbiosum was the ¢rst GDH to yield a crystal structure [3]. The common thread running through the chemical-modi¢cation studies is the involvement of an essential lysine residue in all the GDH species examined, and, in particular, a lysine with an unusually low pK [2]. This residue, totally conserved in all GDHs reported, was lys126 in beef liver [4,5], lys113 in NADP-speci¢c GDH from Neurospora crassa [6] and lys128 in NAD-dependent GDH from Clostridium symbiosum, one of the cluster of three essential lysines (92, 116, 128) in the active site cleft [2]. There are, however, con£icting views as to whether this residue is involved in catalysis, the binding of substrate or coenzymes or in some other function. Our study focuses on the NAD-dependent GDH of the Archaeon extreme halophile H. salinarum. Comparison with other related halophilic dehydrogenases is not possible since there are no reports about pH and chemical modi¢cation studies in other archaeal dehydrogenases except for our previous report [7]. The kinetic mechanism of halophilic NADGDH was reported [8] to be ordered, NAD binding before L-glutamate does, after which an ordered release of ammonium, 2-oxoglutarate and NADH takes place, with the formation of the non-productive complex between the enzyme and K-ketoglutarate (E-K-ketoglutarate). In the present report, a more detailed study is achieved that con¢rms the previous suggestions about the existence of at least one residue of histidine implied in binding and/or catalysis and the possible existence of a residue of lysine playing a role in the chemical mechanism. A complete modi¢cation of the histidyl residues has been carried out with DEPC, a widely used modifying reagent for histidine, ¢nding the existence of three residues, one of which, at least, is essential. Modi¢cation with ethyl acetimidate, a highly speci¢c reagent for lysyl residues, has also shown the existence of an essential residue of lysine, a lysyl residue
that displays an `unusually' low pK, although it seems to be very usual in this enzyme. We also carried out the modi¢cation of tyrosyl groups, which yielded the likely participation of two reactive tyrosines in the chemical mechanism of halophilic enzyme. Finally, some protecting or accelerating modi¢cation reaction metabolites (the substrates: L-glutamate and NAD ; a competitive inhibitor: glutarate; and an activator: L-leucine) were added to the modi¢cation assays and their e¡ects on the modifying reaction studied. All data allowed us to test a chemical mechanism in agreement with experimental ¢ndings reported here. This mechanism has to do with the peculiarities that life at high salt concentrations shows in the characteristics of these groups of organisms, and so in their enzymes, added to the wide regulation patterns that GDHs exhibit. 2. Materials and methods 2.1. Materials L-Glutamic
acid was supplied by Sigma. Modifying reagents diethyl pyrocarbonate, ethyl acetimidate and tetranitromethane were purchased from Fluka. K-Ketoglutarate, NADH and NAD were obtained from Boehringer-Mannheim, Germany. All other chemicals were of analytical grade and were products of Merck. NAD-glutamate dehydrogenase from H. salinarum was puri¢ed using the method of Bonete et al. [9]. 2.2. Enzyme assay and pH studies General conditions are described in a previous paper [9]. Bu¡ers normally used for pH pro¢les were: sodium phosphate pH from 6.5 to 7.5, Tris^HCl from pH 7.5 to 9.0, diethanolamine^HCl from pH 9.0 to 10.0 and CAPS^NaOH from pH 10.0 to 11.0 (data not shown). All bu¡er solutions (0.1 M) contained 1 M NaCl and 3.25 mM EDTA, and were adjusted to pH with NaOH or HCl. The possible di¡erences due to speci¢c bu¡er e¡ect were checked, and they were within the experimental error. pH studies were made by measuring the appearance or disappearance of NADH (forward or reverse reac-
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
tion) at 340 nm with a Hitachi spectrophotometer, maintaining the temperature constant. NADH breakdown at high temperature and pH was also taken into account. The reaction was initiated by the addition of the coenzyme. The pH was checked after each reaction, and showed no change with respect to the initial values. The concentration of Lglutamate was varied for each pH from 4.44 mM to 40 mM for the di¡erent concentrations of the control, without inhibitor and for 2.5, 5.0, 7.5 and 10.0 mM of glutarate. 2.3. Data processing Reciprocal initial velocities were plotted versus reciprocal substrate concentrations; all plots and their replots were linear. Data were ¢tted using the appropriate rate equation and the FORTRAN programs developed by Cleland [10], which also gave the standard error of the mean for V, K and Kis values. Initial velocities (v), obtained at each pH by varying the substrate concentration (S), were substituted in Eq. 1 to yield values for maximum velocity (V), the Michaelis constant (K) for that substrate, and the apparent ¢rst-order rate constant for the interaction of enzyme and substrate (V/K). When the concentration of a competitive inhibitor (I) was varied, the data at each pH value were ¢tted to Eq. 2 to yield a value for the inhibition constant (Kis ). v VS= K S
1
v VS=K 1 I=K is S
2
The variations with pH of the values for V, V/K and Kis were ¢tted to the appropriate equations with the computer program Sigma Plot: log Y logC= 1 H =K 1
3
log Y logC= 1 H =K 1 K 2 =H
4
log Y logC= 1 K 2 =H
5
In these equations Y represents the value of V, V/K, Kis or kobs at a particular pH, [H ] is the hydrogen ion concentration and C is the pH-independent value of the parameter at the optimum state of protonation. K1 and K2 are dissociation constants associated with ionizing groups that, to show activity, must be
515
deprotonated or protonated, respectively. The best ¢t to the data is always chosen according to the standard error of the ¢tted parameter and the lowest value of c, which is the sum of the squares of the residuals divided by the degrees of freedom. The values for pKa obtained at di¡erent temperatures (T) were ¢tted to Eq. 6 in order to determine values for the enthalpy of ionization (vH ion ) and vSion pK a vH = 2; 303RT3vS = 2; 3R
6
To calculate KA for L-leucine, similar assays were performed using concentrations of 0.05, 0.1, 0.5 and 2.0 mM of L-leucine, and data were ¢tted to Eq. 7 (non-essential activation): v V 1 LA=KK A = 1 A=KK A S=K 1 A=K A = 1 A=KK A S 7 yielding KA , K and L, V and K values with their related standard errors. 2.4. Chemical modi¢cation The chemical modi¢cation of NAD-glutamate dehydrogenase by di¡erent group speci¢c reagents was explored through the addition of the reagent, in the adequate concentration, to an enzyme solution containing 0.122 mg/ml, and incubated at 20³C. Aliquots of these solutions were withdrawn at various time intervals and assayed for enzymatic activity. Before adding the reagent, the sample was divided in two equal portions. Then, equal volumes of bu¡er with and without the modi¢er reagent, respectively, were added (adjusting the ¢nal pH if necessary). The second portion served as the reference to check the loss of activity. All kinetic data were analyzed with the aid of computer programs of Cleland [10] and the computer program Sigma Plot. The extent of residue modi¢cation (histidine, lysine and tyrosine) was also explored spectrophotometrically, allowing us to make an estimate of the number of histidyl, lysyl or tyrosyl residues. The di¡erence spectrum between modi¢ed enzyme and the reference in the same conditions, except for the addition of modi¢cation reagent, was obtained for each modi¢cation carried out. Reactives
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
516
were prepared immediately before use in each assay. The kobs value (pseudo-¢rst-order rate constant) for each concentration of the reactive was obtained from the slope of the plot of ln[relative glutamate dehydrogenase activity] versus incubation time. The inactivation of NAD-glutamate dehydrogenase with each reactive was also studied over the pH range indicated in each ¢gure. The time-dependent loss in activity with DEPC for each pH value was measured, and the related pseudo-¢rst-order rate constant was calculated like above. Then, the pKa value for the modi¢cation of the related residue was determined by ¢tting the data to Eq. 3. The modi¢cation of histidine residues was performed by adding a few microliters of a solution of diethyl pyrocarbonate (DEPC). The concentrations of DEPC tested were 0.5, 1.0, 2.5, 5 and 10 mM in 0.1 M triethanolamine bu¡er, pH 8.17. Lysine modi¢cation was carried out with ethyl acetimidate. Ethyl acetimidate is considered to be highly selective for modi¢cation of lysyl residues (K- and O-amino groups), no other groups being modi¢ed at pH from 7 to 10. Amidine derivatives formed in the reaction with the amino groups retain the positive charge of the original amino group [11]. The concentrations of ethyl acetimidate tested were 10, 20, 40, 75 and 150.0 mM, prepared in 0.1 M Tris^HCl bu¡er, pH 8.67. Tetranitromethane (TNM) modi¢es tyrosine residues forming 3-nitrotyrosine. Tetranitro-
methane, although speci¢c for tyrosyl groups, normally oxidizes ^SH groups, and also may modify methionine and tryptophan residues [11]. TNM solutions were made up in ethanol immediately before they were used in each assay. Nitrotyrosine may be quanti¢ed by absorbance at Vmax of 428 nm with an O = 4200 M31 cm31 and Vmax = 360 nm with an O = 2790 M31 cm31 , before removing of the nitroformate ion, which is also formed in the reaction and strongly absorbs at 350 nm. So, the modi¢ed enzyme was ultra¢ltrated using a Millipore Biomax 30K ¢lter. Time-dependent loss of activity was followed by the incubation of the enzyme with the adequate concentration of TNM, adding aliquots of freshly prepared ethanol solution of the reagent to the enzyme solution in 0.1 M Tris^HCl bu¡er, pH 8.0. 3. Results 3.1. pH dependence of V and V/K for L-glutamate The e¡ect of pH on the deamination reaction catalyzed by halophilic GDH, was studied in a pH range of 6.5^10.5 by varying L-glutamate concentrations at a ¢xed level of NAD (2.5 mM). Initial velocity patterns were measured in this range of pH and experiments were carried out at 30³C, 40³C, 50³C and 60³C. As seen in Fig. 1B, log (V/K)Glu
Table 1 pK values for the kinetic parameters for NAD-glutamate dehydrogenase from H. salinarum obtained for the oxidative deaminating (forward) reaction, with the presence of the competitive inhibitor glutarate Temperature (³C)
Pro¢le
30 30 30 40 40 40 50 50 50 60 60 60
log log log log log log log log log log log log
VL-Glu V/KL-Glu KL-Glu VL-Glu V/KL-Glu KL-Glu VL-Glu V/KL-Glu KL-Glu VL-Glu V/KL-Glu KL-Glu
C1 þ S.E.
pK1 þ S.E.
s1
C2 þ S.E.
pK2 þ S.E.
s2
8.7 þ 0.7 1.3 þ 0.3
7.19 þ 0.13 7.9 þ 0.2 ^ 6.9 þ 0.2 7.85 þ 0.12 ^ 6.7 þ 0.2 7.64 þ 0.14 ^ 6.4 þ 0.4 7.56 þ 0.10 ^
0.167 0.294 ^ 0.535 0.426 ^ 0.264 0.198 ^ 0.365 0.196 ^
^ ^ 9 þ4 ^ ^ 0.23 þ 0.05 ^ ^ 0.15 þ 0.02 ^ ^ 0.144 þ 0.014
^ 9.8 þ 0.2 9.59 þ 0.09 ^ 9.45 þ 0.14 9.3 þ 0.2 ^ 9.37 þ 0.17 9.29 þ 0.14 ^ 9.32 þ 0.13 9.12 þ 0.11
^ ^ 0.188 ^ ^ 1.142 ^ ^ 0.250 ^ 0.142 0.196
^ 14.7 þ 1.5 3.5 þ 0.5 ^ 20 þ 3 3.0 þ 0.6 ^ 25 þ 4 3.0 þ 0.4 ^
The kinetic parameters were determined as indicated in Section 2 for di¡erent temperatures from 30³C to 60³C. These values have been obtained by ¢tting data to Eqs. 3, 4 and 5, performed by the computer program Sigma Plot.
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
517
pH was a half-bell with a slope of +1 in the acid side of the pro¢le. Fitting data to Eq. 3 yielded a pK1 value that decreased as temperature increased, reaching a minimal value at 60³C (Table 1). 3.2. pH dependence on binding of L-glutamate analogue
Fig. 1. pH dependence of log V and log V/K, for the reaction catalyzed by NAD-glutamate dehydrogenase from H. salinarum. Open circles (A) and ¢lled circles (B) represent the experimentally determined values of V (a) and V/K (b), respectively. Data were obtained at 30³C at a ¢xed concentration of 120 mM NAD while varying the other substrate (L-glutamate). Curves in both panels are theoretical from a ¢t of Eq. 3A and 4B to the data. The values of the parameters used to draw the curves are given in Table 1.
decreases at both high and low pH values, indicating that the protonation states of two residues associated with the binary complex or L-glutamate are important for binding and/or catalysis. Data for bellshaped pH pro¢les were ¢tted using Eq. 4. The value obtained for the acid side pK1 decreases as temperature increases in the range from 7.9 þ 0.2 at 30³C to 7.56 þ 0.10 at 60³C. For the basic side the values obtained of pK2 (Table 1) were in the range from 9.8 þ 0.2 at 30³C to 9.32 þ 0.13 at 60³C. The data in Fig. 1A show that the variation of log(V) with
In order to determine whether the pK values observed in the V/K pro¢le were true or apparent values we measured the e¡ect of pH on the interaction of glutarate with the enzyme^NAD complex. Glutarate, a competitive inhibitor with respect to L-glutamate [8], gave rise to a pKis pro¢le that had a halfbell shape on the basic side (Fig. 2). Full inhibition patterns for glutarate were obtained at di¡erent pHs for each temperature value. Fitting data to Eq. 5 yielded a pK2 of 9.59 þ 0.09 at 30³C. When temperature was varied, this parameter decreased until a value of 9.12 þ 0.11 at 60³C. All these values are summarized in Table 1, as well as the related pHindependent values for the kinetic parameters (C). pKs observed for the various groups vary tremendously with the environment in which the protein is located. Therefore, the value of a pK is only a very rough guide to the nature of the group. One method to identify the groups, or help us in this purpose, is the study of the temperature dependence of kinetic parameters. The vHion values for various groups differ considerably, and as long as the ionization is not accompanied by conformation changes in proteins, these values can help distinguish the groups. Carboxyl and imidazole groups show little or almost no temperature dependence. Lysine, in turn, gives twice the value of imidazole. To identify the groups whose protonation a¡ects the catalytic activity of halophilic glutamate dehydrogenase, temperature dependence of the pK values
Table 2 Enthalpy and entropy values for the ionization of the groups involved in reaction of oxidative deamination of L-glutamate by NADdependent glutamate dehydrogenase from H. salinarum obtained ¢tting valued for pK1 and pK2 related to V, V/K and Kis for di¡erent temperatures, ranged from 30³C to 60³C according to Eq. 7 from Section 2 Pro¢le VL-Glu V/KL-Glu KL-Glu
vH1 þ S.E. 4
4
1.19U10 þ 0.07U10 5.7 U103 þ 1.0 U103 ^
vS1 þ S.E.
s1
vH2 þ S.E.
vS2 þ S.E.
s2
6þ2 318 þ 3 ^
0.0472 0.0689 ^
^ 7 U103 þ 2 U103 6.6U103 þ 1.6U103
^ 322 þ 7 322 þ 5
^ 0.146 0.111
The vHion is measured in calories/mol and the C related to the ionization entropy is measured in calories/molUdegree.
BBAGEN 24760 28-1-99
518
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
Fig. 2. Variation with pH of pKis for glutarate. The velocity data for inhibition by glutarate were obtained at 30³C at a ¢xed concentration of 120 mM NAD while varying the other substrate (L-glutamate), and ¢t to Eq. 2. The curve for the pKis for glutarate represents the best ¢t of the data to Eq. 5. The values of the parameters used to draw the curves are given in Table 1.
was measured to calculate the enthalpies of ionization (vHion ) and entropy (vSion ). The resulting pK values were ¢tted to Eq. 6, and enthalpies and entropies calculated are summarized in Table 2. Little changes in pK values that are involved in the studies of the dependence on these parameters with temperature may correspond to high changes in vHion . So, although these values can help us to distinguish the groups, other studies such as chemical modi¢cation were carried out to assure this assignation. 3.3. Inactivation kinetics of glutamate dehydrogenase by DEPC, ethyl acetimidate and TNM Inactivation of glutamate dehydrogenase from H. salinarum with DEPC, ethyl acetimidate and TNM results in the rapid loss of enzymatic activity versus time at various concentrations of modi¢er, showing
pseudo-¢rst-order kinetics (Fig. 3). The plots of the ¢rst-order rate constants (Table 3) as a function of DEPC and ethyl acetimidate concentrations also show that these inactivations are bimolecular processes [12,13] (Fig. 3 inset). The second-order rate constant is 53 M31 min31 in contrast with the higher value 24.44 M31 s31 reported for histidine modi¢cation of D-lactate dehydrogenase [13] or 40 þ 5 M31 s31 for phosphotriesterase [14]. The second-order constant obtained for ethyl acetimidate was 1.2 M31 min31 , a value even lower than the one obtained for the modi¢cation of the histidine residue with DEPC. The analysis on the dependence of the rate constant (Table 3) with TNM concentration yields a second-order constant of 2.8 M31 min31 , a value in the same order as the values for histidine and lysine but smaller than the value 1.6 M31 s31 reported for lactate dehydrogenase from Lactobacillus bulgaricus [13]. Reduction of the nitrotyrosine through the addition of excess of sodium dithionite resulted in a full regain of enzyme activity, over 90% of the original enzyme activity. 3.4. pH-dependence on the inactivation of GDH by DEPC, ethyl acetimidate and TNM DEPC modi¢es only unprotonated amino acid residues [15], thus a pH dependence of inactivation can lead to information regarding the nature of the amino acid being modi¢ed. The pseudo-¢rst-order rate constants of DEPC, ethyl acetimidate and TNM modi¢cation increased when the pH was raised (Table 4, Figs. 4 and 5). The higher the value of k the `faster' the modi¢cation is achieved, this parameter being thus an indicative of the improvement of the modi¢cation reaction rate. The k values obtained in
Table 3 Pseudo-¢rst-order rate constant (k) for the concentration-dependent loss of activity in the presence of the respective modifying reagents for halophilic NAD-GDH [DEPC] (mM)
k (min31 )
[Ethyl acetimidate] (mM)
k (min31 )
[TNM] (mM)
k (min31 )
0.5 1.0 2.5 5.0 10.0
0.0054 þ 0.0004 0.021 þ 0.002 0.153 þ 0.007 0.37 þ 0.05 0.50 þ 0.08
10.0 20.0 40.0 75.0 150.0
0.0094 þ 0.0004 0.0247 þ 0.0008 0.0361 þ 0.0012 0.060 þ 0.002 0.174 þ 0.008
1.7 3.3 6.7 15.0
0.0187 þ 0.0007 0.0222 þ 0.0014 0.0277 þ 0.0014 0.055 þ 0.003
These parameters were obtained by ¢tting respective data for each modifying reagent concentration tested to Eq. 8.
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
519
acetimidate pK1 = 8.36 þ 0.13 was similar to the value for glutamate dehydrogenase from Escherichia coli 7.6 [16] or the pKa values of 7.6 at 25³C and 8.1 at 10³C for both bovine liver [17] and C. symbiosum glutamate dehydrogenase [18,19]. For TNM the value found was pKa = 9.0 þ 0.2. This value is smaller than the pKa in the range of 10 to 10.5 reported for tyrosine [20], although the same decrease was reported for lactate dehydrogenase from L. bulgaricus [13]. 3.5. UV-di¡erence spectrum of modi¢ed GDH versus native GDH
Fig. 3. Inactivation of NAD-GDH from H. salinarum by ethyl acetimidate. The enzyme was incubated with reagent concentrations of: b, 10 mM; F, 20 mM; R, 40 mM; S, 75 mM; 8, 150 mM at a ¢xed pH of 8.67. Aliquots were removed at the indicated time intervals and assayed for enzyme activity, as described in Section 2. The inset shows the e¡ect of ethyl acetimidate concentration on the rate of inactivation of the halophilic NAD-GDH. The line is from a linear regression analysis of the data. The slope of this line yields a second-order constant of 1.2 M31 min31 .
the pH dependence study for inactivation at ¢xed concentration of 2.5 mM DEPC, 75.0 mM ethyl acetimidate and 6.7 mM TNM, were ¢tted to Eq. 8: kobs kobs max = 1 H =K a
8
to calculate the pKa of the modi¢ed residue. The pKa of the titratable group was pK1 = 6.6 þ 0.6, similar to the value for phosphotriesterase histidyl residue from Pseudomonas diminuta 6.1 þ 0.1 [14] or the pKa value 7.1 reported for an histidyl residue from L. bulgaricus NAD-lactate dehydrogenase [13]. The pKa found for the group titrated by ethyl
The di¡erence spectrum of GDH after incubation with DEPC (2.5 mM) for 15 min (0.5% of the remaining activity) shows an absorption peak at 252 nm (A252 = 0.031), in the histidine characteristic range of 230 to 250 nm, indicating modi¢cation of histidine [15]. No absorbance decrease was observed around 278 nm in the assay period of time (15 min), but this sharp decrease was shown out at longer modi¢cation times ( s 30 min), indicating the modi¢cation of tyrosyl residues. So, the modi¢cation of histidine with DEPC is easier to achieve and takes place before the modi¢cation of tyrosine. The di¡erence spectrum between native enzyme and enzyme 25 min after addition of 75.0 mM ethyl acetimidate (3.3% of the remaining activity) shows absorption peaks due to the loss of free amino groups [11] at 340 nm (A340 = 0.025) and at 367 nm (A367 = 0.022). TNM, prior to the removal of nitroformate ion by membrane ¢ltration, showed absorption peaks at Vmax of 428 nm (A428 = 0.012) and at Vmax of 360 nm (A360 = 0.016). The number of histidine residues modi¢ed under these conditions, based upon the maximal absorbance of the di¡erence spectrum, was estimated as 3.3 (i.e., approximately three residues for each enzyme subunit) using an extinction coe¤cient of vO = 3200 M31 cm31 , and ¢xing the molecular weight of the enzyme subunit, electrophoretically calculated, at 59KD. The number of essential histidine residues was determined by ¢tting data to equation: log kobs n logDEPC log k
9
where n represents the number of moles of DEPC
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
520
required to react with essential histidine residues to inactivate the enzyme [27]. Two residues of histidine are needed for the catalytic activity of halophilic NAD-GDH. Incubation of the DEPC-modi¢ed enzyme with neutral hydroxylamine 0.5 M for a long period of time did not result in complete reactivation of enzymatic activity, but only in a slight activity increase, probably showing that the modi¢cation of histidine residues is irreversible, due to the possibility of cleavage of the imidazole ring, which renders modi¢cation of histidine irreversible. The number of lysine residues modi¢ed under these conditions, based upon the maxima absorbance of the di¡erence spectrum, was estimated to be 0.9 (i.e., approximately one residue for each enzyme subunit). As we previously determined for histidine [27], we also have found that only one lysine residue is needed. From the absorbance value of the ¢rst peak, 2.3 tyrosyl residues could be estimated, and 1.6 from the other, so two residues of tyrosine are present in each subunit of the enzyme. Evaluation of number of residues implied in the catalytic reaction as stated before [27] also suggests that two residues are needed. 3.6. Ligand protection of glutamate dehydrogenase against inactivation To investigate the possible e¡ect of di¡erent metabolites, protecting the enzyme against the modifying reaction or enhancing the reaction, a nearly saturated concentration of them was added to the reaction mixture. The metabolites tested were L-glutamate and NAD as substrates, L-glutarate as a
Fig. 4. The inactivation of NAD-GDH from H. salinarum with ethyl acetimidate as a function of pH. Kinetics of the inactivation of NAD-glutamate dehydrogenase, as a function of pH concentration. The enzyme was incubated at a ¢xed concentration of 75 mM ethyl acetimidate at di¡erent pHs: F, 7.16; R, 7.91; b, 8.67; S, 9.30; 8, 10.67. Aliquots were removed at the indicated time intervals, as described in Section 2. The lines are from a linear regression analysis of the data.
competitive inhibitor and L-leucine as an activator [21]. In all cases, the time-dependent loss of activity followed pseudo-¢rst-order kinetics, and the related rate constant was determined for each case. As was previously stated for concentration and pH dependence, this rate constant was taken as a useful parameter to compare the modi¢cation with and without
Table 4 Pseudo-¢rst-order rate constant (k) for the pH-dependent loss of activity in the presence of the respective modifying reagents for halophilic NAD-GDH Diethyl pyrocarbonate 31
Ethyl acetimidate
Tetranitromethane 31
pH
k (min )
pH
k (min )
pH
k (min31 )
6.35 7.14 7.67 8.17 8.67 9.61
0.12 þ 0.02 0.081 þ 0.005 0.139 þ 0.007 0.153 þ 0.007 0.414 þ 0.013 0.57 þ 0.02
7.16 7.91 8.67 9.30 10.67
0.0072 þ 0.0014 0.040 þ 0.003 0.060 þ 0.002 0.125 þ 0.005 0.139 þ 0.006
7.3 8.0 8.3 8.7 9.3
0.0063 þ 0.0005 0.0277 þ 0.0014 0.0447 þ 0.0009 0.062 þ 0.002 0.216 þ 0.011
These parameters were obtained by ¢tting respective data for each pH tested at the ¢xed concentrations of the modifying reagent: 2.5 mM for diethyl pyrocarbonate, 75 mM for ethyl acetamidate and 6.7 mM for tetranitromethane, to Eq. 8.
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
metabolites. A higher constant than the control one would indicate that the modi¢cation is enhanced by the addition of the metabolite. A lower constant with respect to the control value would mean that the addition of that metabolite has a protecting e¡ect against the modifying reaction. The addition of the L-glutamate substrate (at a saturating concentration of 0.25 M) to the reaction mixtures results in protection against modi¢cation in the three di¡erent modifying reactions assayed. The results obtained are shown in Table 5. As can be seen, the rate constants in presence of L-glutamate showed much smaller values than the values of k in absence of L-glutamate. L-Glutamate was more e¡ective protecting the modi¢cation of histidyl residues with DEPC. The addition of NAD (at a concentration of 15 mM) to the modi¢cation reaction of histidyl and tyrosyl residues results in protection against modi¢cation. The results are shown in Table 5. The opposite e¡ect was observed for the modi¢cation of lysyl residues with ethyl acetimidate, yielding a k value of
Fig. 5. The inactivation of NAD-GDH from H. salinarum with TNM as a function of pH. Kinetics of the inactivation of NAD-glutamate dehydrogenase, as a function of pH concentration. The enzyme was incubated at a ¢xed concentration of 6.75 mM TNM at di¡erent pHs: b, 7.3; F, 8.0; R, 8.3; S, 8.7; 8, 9.3. Aliquots were removed at the indicated time intervals, as described in Section 2. The lines are from a linear regression analysis of the data.
521
Fig. 6. Binding site for L-glutamate. The functional group interactions indicated are those inferred from pH and chemical modi¢cation studies. One water molecule is hydrogen bonded to the deprotonated Q-nitrogen of a highly conserved lysine residue whose pK is abnormally low. A base (His) serves to polarize charge through hydrogen bonding that facilitates hydride abstraction by NAD and deamination.
0.096 min31 , since the value for the assay in the absence of NAD was 0.060 min31 , indicative of an increase in the enzyme modi¢cation rate, showing a residual activity of 4.5% at only 15 min. The addition of 6.7 mM L-leucine (an activator) results in a protecting e¡ect against modi¢cation with DEPC. The k for the reaction in the presence of L-leucine was 0.047 min31 , while the k of the control was 0.153 min31 . Residual activity was 36.8% at 10 min. However, the addition of the inhibitor glutarate at 30.0 mM yielded a k value higher than the value for the control. So, the e¡ect was just the contrary. The residual activity was only 0.2% at 15 min in the presence of the inhibitor. The modi¢cation of lysyl residues was accelerated by the addition of L-leucine, showing a k value of 0.168 min31 , since the value for the assay in the absence of L-leucine was 0.060 min31 . The residual activity when L-glutamate was present was only 0.33% at 15 min. On the other hand, the addition of the inhibitor glutarate implied the same e¡ect as that of the activator, although with a much lower increase, with a k value of 0.079 min31 and a residual
BBAGEN 24760 28-1-99
522
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
Table 5 Pseudo-¢rst-order rate constants for the reactions with modifying reagents, with and without the presence of metabolites that in£uence the modi¢cation rates DEPC Ethyl acetamidate Tetranitromethane
Control (min31 )
L-Glu
(min31 )
0.153 þ 0.007 0.060 þ 0.002 0.0277 þ 0.0014
0.00675 þ 0.00014 0.0058 þ 0.0004 0.0145 þ 0.0007
activity of 2.9% at 20 min. Similar results were obtained for the reaction with tetranitromethane for the activator, yielding a k higher than the value of k in absence of L-leucine. The residual activity found was only 1.2% at 30 min in the presence of L-leucine, whereas the presence of glutarate has, on the contrary, a protective e¡ect against modi¢cation, with a smaller k. The residual activity was 31.7% at 45 min in the presence of glutarate. All these k values measured are summarized in Table 5. 4. Discussion Two widely used techniques for the identi¢cation of functional groups, and, hopefully, to identify the role they play in catalysis are pH dependence of the kinetic parameters and chemical modi¢cation [22,23]. The reactions catalyzed by all hexameric glutamate dehydrogenases (and probably by all L-K-amino acid dehydrogenases) appear to involve a very speci¢c mechanistic feature. This feature consists in the nucleophilic attack on the K-iminoglutarate intermediate (which results from the preceding hydride transfer step) by water molecule hydrogen bonded to the deprotonated O-nitrogen atom to a highly conserved lysine whose pK is abnormally low [18,19,24,25]. The group that appears in the pro¢le of V is similar to those reported for all glutamate dehydrogenases. It has an abnormally low pK and must be deprotonated to achieve its function, so we may assume this residue to be this highly conserved lysine. Only one residue of lysine seems to be present in each enzyme subunit, so we may conclude that it should be the residue discussed for the V pro¢le. The role that has always been assigned to this residue is the binding of the water molecule. It would play this role due to its ability to coordinate water molecules, making sure the water nucleophilic attack
NAD (min31 )
L-Leu
(min31 )
0.048 þ 0.002 0.096 þ 0.009 0.0129 þ 0.0004
0.047 þ 0.002 0.168 þ 0.007 0.070 þ 0.002
Glutarate (min31 ) 0.202 þ 0.008 0.079 þ 0.004 0.0119 þ 0.0005
would be in the right position, i.e., very close to the K-carbon. pH dependence of the parameter V/K for L-glutamate provides a pro¢le that yields pKs for groups involved in binding and/or catalysis-isomerizationproduct release. While pKs in V pro¢les can be perturbed in either direction, the pKs in the V/K pro¢les can only be perturbed outward. That is, the pKs for groups that ionize on the acidic limb of the pro¢le can be perturbed to the left on the pro¢le, to a lower pH, while pKs that ionize on the basic limb of the pro¢le can be perturbed to the right on the pro¢le, to a higher pH, but this perturbation is usually 6 1 pH unit [26,28]. The pK estimated for the acidic group at 30³C is 7.9 and decreases slightly with temperature (Table 1). This pK value is close to the values for histine residues that range from 5.5 to 7 [23], although slightly higher probably due to the shape of the pro¢le. The dependence of the pK value with temperature gives an vH ion of 5.7 kcal/mol, which is in good concordance with the assignation of this residue as histidine (enthalpy of ionization in the range from 6 to 7.5). In our assays, the treatment with hydroxylamine slightly increases the activity of the modi¢ed enzyme, but the reactivation of the enzyme was not achieved. The N-carbethoxyimidazole formed can be deacylated with hydroxylamine, histidine being regenerated and this can be used to distinguish whether lysine or cysteine residues have been modi¢ed [29]. However, this does not exclude the modi¢cation of histidine. Avaeva and Krasnova [30] suggested that these results might be due to the occurrence of irreversible steps in the modi¢cation reaction in the presence of excess diethyl pyrocarbonate, it being necessary to include agents for the stabilization of the enzyme [15]. Other possible explanations would be that histidine residue(s) were also involved in the active conformation of the active site, being irreversibly lost
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
when carbethoxylation is performed. Halophilic NAD-GDH is very sensitive to changes such as being frozen, irreversibly losing all its activity; or it may be simply that we have not found the adequate condition to decarboxylate the histidine residues modi¢ed. pH dependence of the modi¢cation reaction yields a pK for the modi¢ed group of pK1 = 6.6 [13,14]. This result is in very good concordance with the modi¢cation of an histidyl group, and di¡erent enough from the value obtained in the modi¢cation of the lysyl group with ethyl acetimidate to allow us to consider them as di¡erent groups. Besides, the reaction with DEPC was very fast and modi¢cation of lysyl or cysteinyl groups is very slow. The decrease in absorbance at 278 nm can only be observed when the enzyme is modi¢ed at times longer than an hour of reaction, times much longer than the assay times for histidine modi¢cation with DEPC. The other group or groups that ionize on the basic limb of the pro¢le and so are preferentially protonated for binding and/or catalysis, exhibit a pK of 9.8 at 30³C, a value that also slightly decreases with temperature, yielding a vH ion of 6.6 kcal/mol. This result leads to the assignation of this group to a residue of tyrosine, whose pK values usually range from 9.8 to 10.5 and shows a vH ion of 6 kcal/mol, close to the enthalpy of ionization found for our NAD-GDH residue, although slightly lower in the range of error. Similar results can be observed in the pH pro¢le for the inhibition constant Kis with glutarate as a dead-end competitive inhibitor. At 30³C, the pK2 obtained is 9.59 and also shows a relatively small decrease with temperature that leads to a low vH ion value of 6.6 kcal/mol. It also does support the interpretation that it is ionization of an enzymic residue that is being observed, and not that the pK de¢ned by the high pH side of the pro¢le arises from glutamate itself (pK matches that of the amino group of glutamate). These results are consistent with the ionization of tyrosine. Addition of a competitive inhibitor and protonation^deprotonation of the enzyme comes to equilibrium, and thus dissociation constants are obtained and the pH dependence of the dissociation constants yields true pK values [20,31]. The Kis pro¢le provides true thermodynamic pK values for groups that are involved in binding [26].
523
All the other groups assigned are deprotonated, and, if the assignment is right, all are neutral. There must be residues involved in binding the two dicarboxylic groups of the substrate, since it exhibits negative charges likely to be positively charged. The residues usually involved in playing this role in other glutamate dehydrogenases, and also for other dehydrogenases such as leucine dehydrogenase from Bacillus stearothermophilus, are lysine residues [1,3,17,19] [32]. But according to our results, the lysine residue would play a catalytic role. Cleland [23] postulated that the 1-carboxyl group was hydrogen bonded to a formally neutral group, the 5-carboxyl being the group bound to lysine. So we postulate that the two tyrosine residues found for NADGDH are involved in the binding of both carboxylic groups on the substrate. The same role we propose for these two tyrosyl groups is that assigned for serine-380 in the glutamate dehydrogenase from Clostridium symbiosum (serine-381 in the bovine liver GDH sequence [1] in the mechanism described by Maniscalco et al. [19]). The rate constant has been taken as a useful parameter to study if the addition of metabolites exerts any in£uence on the modi¢cation reaction. The protection obtained by the addition of L-glutamate against both DEPC and ethyl acetimidate is related to the increase in the rate of modi¢cation in both groups when the inhibitor glutarate is present in the reaction mixture. So, the protection by L-glutamate is equivalent to the contrary e¡ect achieved by glutarate. With DEPC, L-leucine and NAD show a protective e¡ect; however, the opposite e¡ect is seen with ethyl acetimidate. These di¡erences would show that we are modifying di¡erent groups. For both histidine and lysine the presence of glutarate gives rise to an increase in the rate of the modifying reaction, showing that these residues could be not only involved in catalysis but also interacting between the hydrogens of the L-glutamate substrate. Glutarate as well as the other inhibitors were stated to be electron-donating molecules in a previous paper [21]. When an L-glutamate molecule is present, it would ¢t the active site, making the reaction of modi¢cation (protecting e¡ect) more di¤cult. When it is glutarate that binds (carboxylic groups), and negative charge is transferred, the new distribution of charge would facilitate the interactions of the glutarate K-hydro-
BBAGEN 24760 28-1-99
524
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
gens with both lysine and histidine residues. Charge would accumulate in these residues, thus enhancing the oxidation modifying reactions. The presence of glutarate protects the enzyme against TNM reaction, so the binding of glutarate probably involves the tyrosyl residues as previously mentioned. Since glutarate has no amino group, the reaction would not take place in spite of the increase in charge in the other groups. The protection achieved by L-glutamate shows a close relation with the acceleration of the rate of modi¢cation achieved with the activator L-leucine. Activators (L-amino acids) for the halophilic NADGDH are known to be electron-accepting [21]. The protecting e¡ect on the histidyl residue may be due to this role, making the modi¢cation of this residue di¤cult by accepting charge. Since two histidines are involved, it must be the one that binds to NAD where the activator plays its role, facilitating the transfer of hydride. The presence of L-leucine makes lysine and tyrosine more reactive, probably polarizing the charge of these two residues to the abovementioned histidine. The di¡erent behaviour of activators and inhibitors with DEPC and TNM again makes evident that the most directly implied residues in activation and inhibition are histidine and tyrosine. This is also consistent with our previous ¢ndings on the study of activation and inhibition of this halophilic enzyme. These results broadly outline the enzyme active site (Fig. 6) and the main reaction steps. The proposed chemical mechanism consistent with the previous discussion is: 1. Binding of the NAD with a deprotonated histidine (previous results) [7]. 2. Hydrogen binding of each of the two carboxylic groups to either tyrosine. The fact of not binding to a positively charged group would partly explain, together with the high number of essential site-bindings located in di¡erent places, the high speci¢city of the enzyme for L-glutamate. A weak union would require the substrate to be ¢tted better to the site. 3. Interaction of one of the N-hydrogens of the L-glutamate amino group (positively charged) with one deprotonated histidine. The substrate molecule L-glutamate must lose a proton from
4.
5.
6. 7.
its K-amino group prior to the hydride transfer step. Possible interaction between the K-hydrogen of L-glutamate and the deprotonated lysine.This conservative lysine has a remarkable ability to coordinate water molecules, so it could also interact with this hydrogen. The net result of the interaction or binding to these general bases (lysine and histidine) is the accumulation of charge in the vicinity of K-hydrogen. Binding of a water molecule to a deprotonated lysine. This coordinated water molecule would be involved in the subsequent nucleophilic attack. So, deprotonation of lysine may be a previous step as discussed extensively in available literature for glutamate dehydrogenase [18,19]. Hydride transfer to the C4 of the coenzyme nicotinamide ring rendering the related K-imino acid [17,32]. Nucleophilic attack by a water molecule bounded to the lysine residue to yield the intermediate Kcarbinolamine, which then decomposes to the keto acid, liberating ammonia.
A curious fact shown in the previous ¢ndings is that all the residues involved in catalysis are neutral, even those that arise from positively charged amino acids. This fact could be related to the special environment to which this archaeal glutamate dehydrogenase from H. salinarum is adapted. Halophilic proteins are usually highly acidic proteins and, as a result, cations would be required to neutralize the acidic groups in order to keep the protein in the proper conformational state [33]. The presence of a charge is accompanied by salt ions of the opposite charge, so that a neutral zone in the protein would not be coordinated by salt ions. This could be a way to maintain the binding or catalytic sites as free as possible of salt ions that could interfere with interactions. These mechanisms would thus show a somehow `halophilic character'. Acknowledgements We thank M. Aniorte, D. Pastor and V. Pina for help with preparation of the manuscript. This research is supported by CICYT Grant PB95-0695
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
and Generalitat Valenciana Grant GV-1170/93 to M.J.B. References [1] E.L. Smith, B.M. Austen, K.M. Blumenthal, J.F. Nyc, in: P.D. Boyer (Ed.), The Enzymes vol. XI, Academic Press, New York, 1975, pp. 294^366. [2] K.S. Lilley, P.C. Engel, Eur. J. Biochem. 207 (1992) 533^ 540. [3] P.J. Baker, K.L. Britton, P.C. Engel, G.W. Farrants, K.S. Lilley, D.W. Rice, T.J. Stillman, Proteins 12 (1992) 75^86. [4] D. Piszkiewicz, E.L. Smith, Biochemistry 10 (1971) 4538^ 4552. [5] R.B. Wallis, J.J. Holbrook, Biochem. J. 133 (1973) 183^ 187. [6] K.M. Blumenthal, E.L. Smith, J. Biol. Chem. 248 (1973) 6002^6008. [7] M.L. Camacho, M.J. Bonete, E. Cadenas, Int. J. Biochem. 25 (1993) 979^985. [8] M.J. Bonete, M.L. Camacho, E. Cadenas, Biochim. Biophys. Acta 990 (1989) 150^155. [9] M.J. Bonete, M.L. Camacho, E. Cadenas, Int. J. Biochem. 18 (1986) 785^789. [10] W.W. Cleland, Methods Enzymol. 63 (1979) 103^138. [11] R.B. Freedman, in: Techniques in Protein and Enzyme Biochemistry, B117, Elsevier/North-Holland Biomedical Press, 1978, pp. 1^23. [12] F.C. Church, R.L. Lundblad, C.M. Noyes, J. Biol. Chem. 260 (1985) 4936^4940. [13] S. Kochhar, P.E. Hunziker, P. Leong-Morgenthaler, H. Hottinger, J. Biol. Chem. 267 (1992) 8499^8513.
525
[14] D.P. Dumas, F.M. Raushel, J. Biol. Chem. 265 (1990) 21498^21503. [15] E.W. Miles, Methods Enzymol. 47 (1977) 431^442. [16] M.J. McPherson, A.J. Baron, K.M. Jones, G.J. Price, J.C. Wootton, Protein Eng. 2 (1988) 147^152. [17] J.E. Rife, W.W. Cleland, Biochemistry 19 (1980) 2328^2333. [18] N. Singh, S.J. Maniscalco, H.F. Fisher, J. Biol. Chem. 268 (1993) 21^28. [19] S.J. Maniscalco, S.K. Saha, P. Vicedomine, H.F. Fisher, Biochemistry 35 (1996) 89^94. [20] W.W. Cleland, Adv. Enzymol. 45 (1977) 273^383. [21] M.J. Bonete, F. Pe¨rez-Pomares, J. Ferrer, M.L. Camacho, Biochim. Biophys. Acta 1289 (1996) 14^24. [22] A. Fersht, in: Enzyme Structure and Mechanism, 2nd ed., Freeman, New York, 1985, p. 155. [23] W.W. Cleland, Methods Enzymol. 87 (1982) 390^405. [24] N.M.W. Bruhuber, J.S. Blanchard, Crit. Rev. Biochem. Mol. Biol. 29 (1994) 415^467. [25] W.E. Karsten, R.E. Viola, Biochim. Biophys. Acta 1077 (1991) 209^219. [26] D.S. Sem, C.B. Kasper, Biochemistry 32 (1993) 11539^ 11547. [27] D. Bhattacharyya, U. Bandyopadhyay, R.K. Banerjee, J. Biol. Chem. 267 (1992) 9800^9804. [28] I.H. Segel, in: Enzyme Kinetics. Behavior and Analysis of Rapid Equilibrium and Steady-state Enzyme Systems, Wiley, New York, 1975, pp. 884^943. [29] P. Tomme, S. Chauvaux, P. Be¨guin, J. Millet, J.P. Aubert, M. Claeyssens, J. Biol. Chem. 266 (1991) 10313^10318. [30] S.M. Avaeva, V.I. Kransnova, Bioorg. Khim. 1 (1975) 1600. [31] A.J. Berdis, P.F. Cook, Biochemistry 32 (1993) 2041^2046. [32] T. Sekimoto, T. Matsuyama, T. Fukui, K. Tanizawa, J. Biol. Chem. 268, (36) (1993) 27039^27045. [33] H. Larsen, Adv. Microb. Physiol. 1 (1967) 97^132.
BBAGEN 24760 28-1-99
Amino acid residues involved in the catalytic mechanism of NAD-dependent glutamate dehydrogenase from Halobacterium salinarum Francisco Pe¨rez-Pomares, Juan Ferrer, Mo¨nica Camacho, Carmen Pire, Francisco LLorca, Mar|¨a Jose¨ Bonete * Divisio¨n de Bioqu|¨mica, Facultad de Ciencias, Universidad de Alicante, Ap. 99, E-03080 Alicante, Spain Received 5 August 1998; received in revised form 27 October 1998; accepted 10 December 1998
Abstract The pH dependence of kinetic parameters for a competitive inhibitor (glutarate) was determined in order to obtain information on the chemical mechanism for NAD-dependent glutamate dehydrogenase from Halobacterium salinarum. The maximum velocity is pH dependent, decreasing at low pHs giving a pK value of 7.19 þ 0.13, while the V/K for L-glutamate at 30³C decreases at low and high pHs, yielding pK values of 7.9 þ 0.2 and 9.8 þ 0.2, respectively. The glutarate pKis profile decreases at high pHs, yielding a pK of 9.59 þ 0.09 at 30³C. The values of ionization heat calculated from the change in pK with temperature are: 1.19U104 , 5.7U103 , 7U103 , 6.6U103 cal mol31 , for the residues involved. All these data suggest that the groups required for catalysis and/or binding are lysine, histidine and tyrosine. The enzyme shows a time-dependent loss in glutamate oxidation activity when incubated with diethyl pyrocarbonate (DEPC). Inactivation follows pseudo-first-order kinetics with a second-order rate constant of 53 M31 min31 . The pKa of the titratable group was pK1 = 6.6 þ 0.6. Inactivation with ethyl acetimidate also shows pseudo-first-order kinetics as well as inactivation with TNM yielding second-order constants of 1.2 M31 min31 and 2.8 M31 min31 , and pKa s of 8.36 and 9.0, respectively. The proposed mechanism involves hydrogen binding of each of the two carboxylic groups to tyrosyl residues; histidine interacts with one of the N-hydrogens of the L-glutamate amino group. We also corroborate the presence of a conservative lysine that has a remarkable ability to coordinate a water molecule that would act as general base. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Archaea; NAD-glutamate dehydrogenase; Chemical modi¢cation; (Halobacterium salinarum)
1. Introduction An enzyme must interact with substrates; so, understanding enzyme mechanisms requires underAbbreviations: NAD-GDH, NAD-dependent glutamate dehydrogenase; DEPC, diethyl pyrocarbonate; TNM, tetranitromethane * Corresponding author. Fax: +34-96-590-3464; E-mail: [email protected]
standing those interactions taking place between an enzyme and a substrate that lead to transition-state stabilization, as well as of the residues involved in the catalytic process. Useful information can be obtained by organic chemistry and kinetic methods, including the use of structurally related compounds, which affect the function of the enzyme, with the aim of characterizing the binding sites and the possible role, played by the residues involved in the catalytic pathways.
0304-4165 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 9 8 ) 0 0 1 7 4 - 3
BBAGEN 24760 28-1-99
514
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
Glutamate dehydrogenases (GDH) (EC 1.4.1.2^4) are a family of enzymes that catalyze the reversible oxidative deamination of L-glutamate using NAD , NADP or both as coenzymes [1]. Although the metabolic importance and regulatory properties of the GDH species have attracted considerable interest, the three-dimensional structure of any GDH became available only few years ago [2]. The NAD-dependent GDH from Clostridium symbiosum was the ¢rst GDH to yield a crystal structure [3]. The common thread running through the chemical-modi¢cation studies is the involvement of an essential lysine residue in all the GDH species examined, and, in particular, a lysine with an unusually low pK [2]. This residue, totally conserved in all GDHs reported, was lys126 in beef liver [4,5], lys113 in NADP-speci¢c GDH from Neurospora crassa [6] and lys128 in NAD-dependent GDH from Clostridium symbiosum, one of the cluster of three essential lysines (92, 116, 128) in the active site cleft [2]. There are, however, con£icting views as to whether this residue is involved in catalysis, the binding of substrate or coenzymes or in some other function. Our study focuses on the NAD-dependent GDH of the Archaeon extreme halophile H. salinarum. Comparison with other related halophilic dehydrogenases is not possible since there are no reports about pH and chemical modi¢cation studies in other archaeal dehydrogenases except for our previous report [7]. The kinetic mechanism of halophilic NADGDH was reported [8] to be ordered, NAD binding before L-glutamate does, after which an ordered release of ammonium, 2-oxoglutarate and NADH takes place, with the formation of the non-productive complex between the enzyme and K-ketoglutarate (E-K-ketoglutarate). In the present report, a more detailed study is achieved that con¢rms the previous suggestions about the existence of at least one residue of histidine implied in binding and/or catalysis and the possible existence of a residue of lysine playing a role in the chemical mechanism. A complete modi¢cation of the histidyl residues has been carried out with DEPC, a widely used modifying reagent for histidine, ¢nding the existence of three residues, one of which, at least, is essential. Modi¢cation with ethyl acetimidate, a highly speci¢c reagent for lysyl residues, has also shown the existence of an essential residue of lysine, a lysyl residue
that displays an `unusually' low pK, although it seems to be very usual in this enzyme. We also carried out the modi¢cation of tyrosyl groups, which yielded the likely participation of two reactive tyrosines in the chemical mechanism of halophilic enzyme. Finally, some protecting or accelerating modi¢cation reaction metabolites (the substrates: L-glutamate and NAD ; a competitive inhibitor: glutarate; and an activator: L-leucine) were added to the modi¢cation assays and their e¡ects on the modifying reaction studied. All data allowed us to test a chemical mechanism in agreement with experimental ¢ndings reported here. This mechanism has to do with the peculiarities that life at high salt concentrations shows in the characteristics of these groups of organisms, and so in their enzymes, added to the wide regulation patterns that GDHs exhibit. 2. Materials and methods 2.1. Materials L-Glutamic
acid was supplied by Sigma. Modifying reagents diethyl pyrocarbonate, ethyl acetimidate and tetranitromethane were purchased from Fluka. K-Ketoglutarate, NADH and NAD were obtained from Boehringer-Mannheim, Germany. All other chemicals were of analytical grade and were products of Merck. NAD-glutamate dehydrogenase from H. salinarum was puri¢ed using the method of Bonete et al. [9]. 2.2. Enzyme assay and pH studies General conditions are described in a previous paper [9]. Bu¡ers normally used for pH pro¢les were: sodium phosphate pH from 6.5 to 7.5, Tris^HCl from pH 7.5 to 9.0, diethanolamine^HCl from pH 9.0 to 10.0 and CAPS^NaOH from pH 10.0 to 11.0 (data not shown). All bu¡er solutions (0.1 M) contained 1 M NaCl and 3.25 mM EDTA, and were adjusted to pH with NaOH or HCl. The possible di¡erences due to speci¢c bu¡er e¡ect were checked, and they were within the experimental error. pH studies were made by measuring the appearance or disappearance of NADH (forward or reverse reac-
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
tion) at 340 nm with a Hitachi spectrophotometer, maintaining the temperature constant. NADH breakdown at high temperature and pH was also taken into account. The reaction was initiated by the addition of the coenzyme. The pH was checked after each reaction, and showed no change with respect to the initial values. The concentration of Lglutamate was varied for each pH from 4.44 mM to 40 mM for the di¡erent concentrations of the control, without inhibitor and for 2.5, 5.0, 7.5 and 10.0 mM of glutarate. 2.3. Data processing Reciprocal initial velocities were plotted versus reciprocal substrate concentrations; all plots and their replots were linear. Data were ¢tted using the appropriate rate equation and the FORTRAN programs developed by Cleland [10], which also gave the standard error of the mean for V, K and Kis values. Initial velocities (v), obtained at each pH by varying the substrate concentration (S), were substituted in Eq. 1 to yield values for maximum velocity (V), the Michaelis constant (K) for that substrate, and the apparent ¢rst-order rate constant for the interaction of enzyme and substrate (V/K). When the concentration of a competitive inhibitor (I) was varied, the data at each pH value were ¢tted to Eq. 2 to yield a value for the inhibition constant (Kis ). v VS= K S
1
v VS=K 1 I=K is S
2
The variations with pH of the values for V, V/K and Kis were ¢tted to the appropriate equations with the computer program Sigma Plot: log Y logC= 1 H =K 1
3
log Y logC= 1 H =K 1 K 2 =H
4
log Y logC= 1 K 2 =H
5
In these equations Y represents the value of V, V/K, Kis or kobs at a particular pH, [H ] is the hydrogen ion concentration and C is the pH-independent value of the parameter at the optimum state of protonation. K1 and K2 are dissociation constants associated with ionizing groups that, to show activity, must be
515
deprotonated or protonated, respectively. The best ¢t to the data is always chosen according to the standard error of the ¢tted parameter and the lowest value of c, which is the sum of the squares of the residuals divided by the degrees of freedom. The values for pKa obtained at di¡erent temperatures (T) were ¢tted to Eq. 6 in order to determine values for the enthalpy of ionization (vH ion ) and vSion pK a vH = 2; 303RT3vS = 2; 3R
6
To calculate KA for L-leucine, similar assays were performed using concentrations of 0.05, 0.1, 0.5 and 2.0 mM of L-leucine, and data were ¢tted to Eq. 7 (non-essential activation): v V 1 LA=KK A = 1 A=KK A S=K 1 A=K A = 1 A=KK A S 7 yielding KA , K and L, V and K values with their related standard errors. 2.4. Chemical modi¢cation The chemical modi¢cation of NAD-glutamate dehydrogenase by di¡erent group speci¢c reagents was explored through the addition of the reagent, in the adequate concentration, to an enzyme solution containing 0.122 mg/ml, and incubated at 20³C. Aliquots of these solutions were withdrawn at various time intervals and assayed for enzymatic activity. Before adding the reagent, the sample was divided in two equal portions. Then, equal volumes of bu¡er with and without the modi¢er reagent, respectively, were added (adjusting the ¢nal pH if necessary). The second portion served as the reference to check the loss of activity. All kinetic data were analyzed with the aid of computer programs of Cleland [10] and the computer program Sigma Plot. The extent of residue modi¢cation (histidine, lysine and tyrosine) was also explored spectrophotometrically, allowing us to make an estimate of the number of histidyl, lysyl or tyrosyl residues. The di¡erence spectrum between modi¢ed enzyme and the reference in the same conditions, except for the addition of modi¢cation reagent, was obtained for each modi¢cation carried out. Reactives
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
516
were prepared immediately before use in each assay. The kobs value (pseudo-¢rst-order rate constant) for each concentration of the reactive was obtained from the slope of the plot of ln[relative glutamate dehydrogenase activity] versus incubation time. The inactivation of NAD-glutamate dehydrogenase with each reactive was also studied over the pH range indicated in each ¢gure. The time-dependent loss in activity with DEPC for each pH value was measured, and the related pseudo-¢rst-order rate constant was calculated like above. Then, the pKa value for the modi¢cation of the related residue was determined by ¢tting the data to Eq. 3. The modi¢cation of histidine residues was performed by adding a few microliters of a solution of diethyl pyrocarbonate (DEPC). The concentrations of DEPC tested were 0.5, 1.0, 2.5, 5 and 10 mM in 0.1 M triethanolamine bu¡er, pH 8.17. Lysine modi¢cation was carried out with ethyl acetimidate. Ethyl acetimidate is considered to be highly selective for modi¢cation of lysyl residues (K- and O-amino groups), no other groups being modi¢ed at pH from 7 to 10. Amidine derivatives formed in the reaction with the amino groups retain the positive charge of the original amino group [11]. The concentrations of ethyl acetimidate tested were 10, 20, 40, 75 and 150.0 mM, prepared in 0.1 M Tris^HCl bu¡er, pH 8.67. Tetranitromethane (TNM) modi¢es tyrosine residues forming 3-nitrotyrosine. Tetranitro-
methane, although speci¢c for tyrosyl groups, normally oxidizes ^SH groups, and also may modify methionine and tryptophan residues [11]. TNM solutions were made up in ethanol immediately before they were used in each assay. Nitrotyrosine may be quanti¢ed by absorbance at Vmax of 428 nm with an O = 4200 M31 cm31 and Vmax = 360 nm with an O = 2790 M31 cm31 , before removing of the nitroformate ion, which is also formed in the reaction and strongly absorbs at 350 nm. So, the modi¢ed enzyme was ultra¢ltrated using a Millipore Biomax 30K ¢lter. Time-dependent loss of activity was followed by the incubation of the enzyme with the adequate concentration of TNM, adding aliquots of freshly prepared ethanol solution of the reagent to the enzyme solution in 0.1 M Tris^HCl bu¡er, pH 8.0. 3. Results 3.1. pH dependence of V and V/K for L-glutamate The e¡ect of pH on the deamination reaction catalyzed by halophilic GDH, was studied in a pH range of 6.5^10.5 by varying L-glutamate concentrations at a ¢xed level of NAD (2.5 mM). Initial velocity patterns were measured in this range of pH and experiments were carried out at 30³C, 40³C, 50³C and 60³C. As seen in Fig. 1B, log (V/K)Glu
Table 1 pK values for the kinetic parameters for NAD-glutamate dehydrogenase from H. salinarum obtained for the oxidative deaminating (forward) reaction, with the presence of the competitive inhibitor glutarate Temperature (³C)
Pro¢le
30 30 30 40 40 40 50 50 50 60 60 60
log log log log log log log log log log log log
VL-Glu V/KL-Glu KL-Glu VL-Glu V/KL-Glu KL-Glu VL-Glu V/KL-Glu KL-Glu VL-Glu V/KL-Glu KL-Glu
C1 þ S.E.
pK1 þ S.E.
s1
C2 þ S.E.
pK2 þ S.E.
s2
8.7 þ 0.7 1.3 þ 0.3
7.19 þ 0.13 7.9 þ 0.2 ^ 6.9 þ 0.2 7.85 þ 0.12 ^ 6.7 þ 0.2 7.64 þ 0.14 ^ 6.4 þ 0.4 7.56 þ 0.10 ^
0.167 0.294 ^ 0.535 0.426 ^ 0.264 0.198 ^ 0.365 0.196 ^
^ ^ 9 þ4 ^ ^ 0.23 þ 0.05 ^ ^ 0.15 þ 0.02 ^ ^ 0.144 þ 0.014
^ 9.8 þ 0.2 9.59 þ 0.09 ^ 9.45 þ 0.14 9.3 þ 0.2 ^ 9.37 þ 0.17 9.29 þ 0.14 ^ 9.32 þ 0.13 9.12 þ 0.11
^ ^ 0.188 ^ ^ 1.142 ^ ^ 0.250 ^ 0.142 0.196
^ 14.7 þ 1.5 3.5 þ 0.5 ^ 20 þ 3 3.0 þ 0.6 ^ 25 þ 4 3.0 þ 0.4 ^
The kinetic parameters were determined as indicated in Section 2 for di¡erent temperatures from 30³C to 60³C. These values have been obtained by ¢tting data to Eqs. 3, 4 and 5, performed by the computer program Sigma Plot.
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
517
pH was a half-bell with a slope of +1 in the acid side of the pro¢le. Fitting data to Eq. 3 yielded a pK1 value that decreased as temperature increased, reaching a minimal value at 60³C (Table 1). 3.2. pH dependence on binding of L-glutamate analogue
Fig. 1. pH dependence of log V and log V/K, for the reaction catalyzed by NAD-glutamate dehydrogenase from H. salinarum. Open circles (A) and ¢lled circles (B) represent the experimentally determined values of V (a) and V/K (b), respectively. Data were obtained at 30³C at a ¢xed concentration of 120 mM NAD while varying the other substrate (L-glutamate). Curves in both panels are theoretical from a ¢t of Eq. 3A and 4B to the data. The values of the parameters used to draw the curves are given in Table 1.
decreases at both high and low pH values, indicating that the protonation states of two residues associated with the binary complex or L-glutamate are important for binding and/or catalysis. Data for bellshaped pH pro¢les were ¢tted using Eq. 4. The value obtained for the acid side pK1 decreases as temperature increases in the range from 7.9 þ 0.2 at 30³C to 7.56 þ 0.10 at 60³C. For the basic side the values obtained of pK2 (Table 1) were in the range from 9.8 þ 0.2 at 30³C to 9.32 þ 0.13 at 60³C. The data in Fig. 1A show that the variation of log(V) with
In order to determine whether the pK values observed in the V/K pro¢le were true or apparent values we measured the e¡ect of pH on the interaction of glutarate with the enzyme^NAD complex. Glutarate, a competitive inhibitor with respect to L-glutamate [8], gave rise to a pKis pro¢le that had a halfbell shape on the basic side (Fig. 2). Full inhibition patterns for glutarate were obtained at di¡erent pHs for each temperature value. Fitting data to Eq. 5 yielded a pK2 of 9.59 þ 0.09 at 30³C. When temperature was varied, this parameter decreased until a value of 9.12 þ 0.11 at 60³C. All these values are summarized in Table 1, as well as the related pHindependent values for the kinetic parameters (C). pKs observed for the various groups vary tremendously with the environment in which the protein is located. Therefore, the value of a pK is only a very rough guide to the nature of the group. One method to identify the groups, or help us in this purpose, is the study of the temperature dependence of kinetic parameters. The vHion values for various groups differ considerably, and as long as the ionization is not accompanied by conformation changes in proteins, these values can help distinguish the groups. Carboxyl and imidazole groups show little or almost no temperature dependence. Lysine, in turn, gives twice the value of imidazole. To identify the groups whose protonation a¡ects the catalytic activity of halophilic glutamate dehydrogenase, temperature dependence of the pK values
Table 2 Enthalpy and entropy values for the ionization of the groups involved in reaction of oxidative deamination of L-glutamate by NADdependent glutamate dehydrogenase from H. salinarum obtained ¢tting valued for pK1 and pK2 related to V, V/K and Kis for di¡erent temperatures, ranged from 30³C to 60³C according to Eq. 7 from Section 2 Pro¢le VL-Glu V/KL-Glu KL-Glu
vH1 þ S.E. 4
4
1.19U10 þ 0.07U10 5.7 U103 þ 1.0 U103 ^
vS1 þ S.E.
s1
vH2 þ S.E.
vS2 þ S.E.
s2
6þ2 318 þ 3 ^
0.0472 0.0689 ^
^ 7 U103 þ 2 U103 6.6U103 þ 1.6U103
^ 322 þ 7 322 þ 5
^ 0.146 0.111
The vHion is measured in calories/mol and the C related to the ionization entropy is measured in calories/molUdegree.
BBAGEN 24760 28-1-99
518
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
Fig. 2. Variation with pH of pKis for glutarate. The velocity data for inhibition by glutarate were obtained at 30³C at a ¢xed concentration of 120 mM NAD while varying the other substrate (L-glutamate), and ¢t to Eq. 2. The curve for the pKis for glutarate represents the best ¢t of the data to Eq. 5. The values of the parameters used to draw the curves are given in Table 1.
was measured to calculate the enthalpies of ionization (vHion ) and entropy (vSion ). The resulting pK values were ¢tted to Eq. 6, and enthalpies and entropies calculated are summarized in Table 2. Little changes in pK values that are involved in the studies of the dependence on these parameters with temperature may correspond to high changes in vHion . So, although these values can help us to distinguish the groups, other studies such as chemical modi¢cation were carried out to assure this assignation. 3.3. Inactivation kinetics of glutamate dehydrogenase by DEPC, ethyl acetimidate and TNM Inactivation of glutamate dehydrogenase from H. salinarum with DEPC, ethyl acetimidate and TNM results in the rapid loss of enzymatic activity versus time at various concentrations of modi¢er, showing
pseudo-¢rst-order kinetics (Fig. 3). The plots of the ¢rst-order rate constants (Table 3) as a function of DEPC and ethyl acetimidate concentrations also show that these inactivations are bimolecular processes [12,13] (Fig. 3 inset). The second-order rate constant is 53 M31 min31 in contrast with the higher value 24.44 M31 s31 reported for histidine modi¢cation of D-lactate dehydrogenase [13] or 40 þ 5 M31 s31 for phosphotriesterase [14]. The second-order constant obtained for ethyl acetimidate was 1.2 M31 min31 , a value even lower than the one obtained for the modi¢cation of the histidine residue with DEPC. The analysis on the dependence of the rate constant (Table 3) with TNM concentration yields a second-order constant of 2.8 M31 min31 , a value in the same order as the values for histidine and lysine but smaller than the value 1.6 M31 s31 reported for lactate dehydrogenase from Lactobacillus bulgaricus [13]. Reduction of the nitrotyrosine through the addition of excess of sodium dithionite resulted in a full regain of enzyme activity, over 90% of the original enzyme activity. 3.4. pH-dependence on the inactivation of GDH by DEPC, ethyl acetimidate and TNM DEPC modi¢es only unprotonated amino acid residues [15], thus a pH dependence of inactivation can lead to information regarding the nature of the amino acid being modi¢ed. The pseudo-¢rst-order rate constants of DEPC, ethyl acetimidate and TNM modi¢cation increased when the pH was raised (Table 4, Figs. 4 and 5). The higher the value of k the `faster' the modi¢cation is achieved, this parameter being thus an indicative of the improvement of the modi¢cation reaction rate. The k values obtained in
Table 3 Pseudo-¢rst-order rate constant (k) for the concentration-dependent loss of activity in the presence of the respective modifying reagents for halophilic NAD-GDH [DEPC] (mM)
k (min31 )
[Ethyl acetimidate] (mM)
k (min31 )
[TNM] (mM)
k (min31 )
0.5 1.0 2.5 5.0 10.0
0.0054 þ 0.0004 0.021 þ 0.002 0.153 þ 0.007 0.37 þ 0.05 0.50 þ 0.08
10.0 20.0 40.0 75.0 150.0
0.0094 þ 0.0004 0.0247 þ 0.0008 0.0361 þ 0.0012 0.060 þ 0.002 0.174 þ 0.008
1.7 3.3 6.7 15.0
0.0187 þ 0.0007 0.0222 þ 0.0014 0.0277 þ 0.0014 0.055 þ 0.003
These parameters were obtained by ¢tting respective data for each modifying reagent concentration tested to Eq. 8.
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
519
acetimidate pK1 = 8.36 þ 0.13 was similar to the value for glutamate dehydrogenase from Escherichia coli 7.6 [16] or the pKa values of 7.6 at 25³C and 8.1 at 10³C for both bovine liver [17] and C. symbiosum glutamate dehydrogenase [18,19]. For TNM the value found was pKa = 9.0 þ 0.2. This value is smaller than the pKa in the range of 10 to 10.5 reported for tyrosine [20], although the same decrease was reported for lactate dehydrogenase from L. bulgaricus [13]. 3.5. UV-di¡erence spectrum of modi¢ed GDH versus native GDH
Fig. 3. Inactivation of NAD-GDH from H. salinarum by ethyl acetimidate. The enzyme was incubated with reagent concentrations of: b, 10 mM; F, 20 mM; R, 40 mM; S, 75 mM; 8, 150 mM at a ¢xed pH of 8.67. Aliquots were removed at the indicated time intervals and assayed for enzyme activity, as described in Section 2. The inset shows the e¡ect of ethyl acetimidate concentration on the rate of inactivation of the halophilic NAD-GDH. The line is from a linear regression analysis of the data. The slope of this line yields a second-order constant of 1.2 M31 min31 .
the pH dependence study for inactivation at ¢xed concentration of 2.5 mM DEPC, 75.0 mM ethyl acetimidate and 6.7 mM TNM, were ¢tted to Eq. 8: kobs kobs max = 1 H =K a
8
to calculate the pKa of the modi¢ed residue. The pKa of the titratable group was pK1 = 6.6 þ 0.6, similar to the value for phosphotriesterase histidyl residue from Pseudomonas diminuta 6.1 þ 0.1 [14] or the pKa value 7.1 reported for an histidyl residue from L. bulgaricus NAD-lactate dehydrogenase [13]. The pKa found for the group titrated by ethyl
The di¡erence spectrum of GDH after incubation with DEPC (2.5 mM) for 15 min (0.5% of the remaining activity) shows an absorption peak at 252 nm (A252 = 0.031), in the histidine characteristic range of 230 to 250 nm, indicating modi¢cation of histidine [15]. No absorbance decrease was observed around 278 nm in the assay period of time (15 min), but this sharp decrease was shown out at longer modi¢cation times ( s 30 min), indicating the modi¢cation of tyrosyl residues. So, the modi¢cation of histidine with DEPC is easier to achieve and takes place before the modi¢cation of tyrosine. The di¡erence spectrum between native enzyme and enzyme 25 min after addition of 75.0 mM ethyl acetimidate (3.3% of the remaining activity) shows absorption peaks due to the loss of free amino groups [11] at 340 nm (A340 = 0.025) and at 367 nm (A367 = 0.022). TNM, prior to the removal of nitroformate ion by membrane ¢ltration, showed absorption peaks at Vmax of 428 nm (A428 = 0.012) and at Vmax of 360 nm (A360 = 0.016). The number of histidine residues modi¢ed under these conditions, based upon the maximal absorbance of the di¡erence spectrum, was estimated as 3.3 (i.e., approximately three residues for each enzyme subunit) using an extinction coe¤cient of vO = 3200 M31 cm31 , and ¢xing the molecular weight of the enzyme subunit, electrophoretically calculated, at 59KD. The number of essential histidine residues was determined by ¢tting data to equation: log kobs n logDEPC log k
9
where n represents the number of moles of DEPC
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
520
required to react with essential histidine residues to inactivate the enzyme [27]. Two residues of histidine are needed for the catalytic activity of halophilic NAD-GDH. Incubation of the DEPC-modi¢ed enzyme with neutral hydroxylamine 0.5 M for a long period of time did not result in complete reactivation of enzymatic activity, but only in a slight activity increase, probably showing that the modi¢cation of histidine residues is irreversible, due to the possibility of cleavage of the imidazole ring, which renders modi¢cation of histidine irreversible. The number of lysine residues modi¢ed under these conditions, based upon the maxima absorbance of the di¡erence spectrum, was estimated to be 0.9 (i.e., approximately one residue for each enzyme subunit). As we previously determined for histidine [27], we also have found that only one lysine residue is needed. From the absorbance value of the ¢rst peak, 2.3 tyrosyl residues could be estimated, and 1.6 from the other, so two residues of tyrosine are present in each subunit of the enzyme. Evaluation of number of residues implied in the catalytic reaction as stated before [27] also suggests that two residues are needed. 3.6. Ligand protection of glutamate dehydrogenase against inactivation To investigate the possible e¡ect of di¡erent metabolites, protecting the enzyme against the modifying reaction or enhancing the reaction, a nearly saturated concentration of them was added to the reaction mixture. The metabolites tested were L-glutamate and NAD as substrates, L-glutarate as a
Fig. 4. The inactivation of NAD-GDH from H. salinarum with ethyl acetimidate as a function of pH. Kinetics of the inactivation of NAD-glutamate dehydrogenase, as a function of pH concentration. The enzyme was incubated at a ¢xed concentration of 75 mM ethyl acetimidate at di¡erent pHs: F, 7.16; R, 7.91; b, 8.67; S, 9.30; 8, 10.67. Aliquots were removed at the indicated time intervals, as described in Section 2. The lines are from a linear regression analysis of the data.
competitive inhibitor and L-leucine as an activator [21]. In all cases, the time-dependent loss of activity followed pseudo-¢rst-order kinetics, and the related rate constant was determined for each case. As was previously stated for concentration and pH dependence, this rate constant was taken as a useful parameter to compare the modi¢cation with and without
Table 4 Pseudo-¢rst-order rate constant (k) for the pH-dependent loss of activity in the presence of the respective modifying reagents for halophilic NAD-GDH Diethyl pyrocarbonate 31
Ethyl acetimidate
Tetranitromethane 31
pH
k (min )
pH
k (min )
pH
k (min31 )
6.35 7.14 7.67 8.17 8.67 9.61
0.12 þ 0.02 0.081 þ 0.005 0.139 þ 0.007 0.153 þ 0.007 0.414 þ 0.013 0.57 þ 0.02
7.16 7.91 8.67 9.30 10.67
0.0072 þ 0.0014 0.040 þ 0.003 0.060 þ 0.002 0.125 þ 0.005 0.139 þ 0.006
7.3 8.0 8.3 8.7 9.3
0.0063 þ 0.0005 0.0277 þ 0.0014 0.0447 þ 0.0009 0.062 þ 0.002 0.216 þ 0.011
These parameters were obtained by ¢tting respective data for each pH tested at the ¢xed concentrations of the modifying reagent: 2.5 mM for diethyl pyrocarbonate, 75 mM for ethyl acetamidate and 6.7 mM for tetranitromethane, to Eq. 8.
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
metabolites. A higher constant than the control one would indicate that the modi¢cation is enhanced by the addition of the metabolite. A lower constant with respect to the control value would mean that the addition of that metabolite has a protecting e¡ect against the modifying reaction. The addition of the L-glutamate substrate (at a saturating concentration of 0.25 M) to the reaction mixtures results in protection against modi¢cation in the three di¡erent modifying reactions assayed. The results obtained are shown in Table 5. As can be seen, the rate constants in presence of L-glutamate showed much smaller values than the values of k in absence of L-glutamate. L-Glutamate was more e¡ective protecting the modi¢cation of histidyl residues with DEPC. The addition of NAD (at a concentration of 15 mM) to the modi¢cation reaction of histidyl and tyrosyl residues results in protection against modi¢cation. The results are shown in Table 5. The opposite e¡ect was observed for the modi¢cation of lysyl residues with ethyl acetimidate, yielding a k value of
Fig. 5. The inactivation of NAD-GDH from H. salinarum with TNM as a function of pH. Kinetics of the inactivation of NAD-glutamate dehydrogenase, as a function of pH concentration. The enzyme was incubated at a ¢xed concentration of 6.75 mM TNM at di¡erent pHs: b, 7.3; F, 8.0; R, 8.3; S, 8.7; 8, 9.3. Aliquots were removed at the indicated time intervals, as described in Section 2. The lines are from a linear regression analysis of the data.
521
Fig. 6. Binding site for L-glutamate. The functional group interactions indicated are those inferred from pH and chemical modi¢cation studies. One water molecule is hydrogen bonded to the deprotonated Q-nitrogen of a highly conserved lysine residue whose pK is abnormally low. A base (His) serves to polarize charge through hydrogen bonding that facilitates hydride abstraction by NAD and deamination.
0.096 min31 , since the value for the assay in the absence of NAD was 0.060 min31 , indicative of an increase in the enzyme modi¢cation rate, showing a residual activity of 4.5% at only 15 min. The addition of 6.7 mM L-leucine (an activator) results in a protecting e¡ect against modi¢cation with DEPC. The k for the reaction in the presence of L-leucine was 0.047 min31 , while the k of the control was 0.153 min31 . Residual activity was 36.8% at 10 min. However, the addition of the inhibitor glutarate at 30.0 mM yielded a k value higher than the value for the control. So, the e¡ect was just the contrary. The residual activity was only 0.2% at 15 min in the presence of the inhibitor. The modi¢cation of lysyl residues was accelerated by the addition of L-leucine, showing a k value of 0.168 min31 , since the value for the assay in the absence of L-leucine was 0.060 min31 . The residual activity when L-glutamate was present was only 0.33% at 15 min. On the other hand, the addition of the inhibitor glutarate implied the same e¡ect as that of the activator, although with a much lower increase, with a k value of 0.079 min31 and a residual
BBAGEN 24760 28-1-99
522
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
Table 5 Pseudo-¢rst-order rate constants for the reactions with modifying reagents, with and without the presence of metabolites that in£uence the modi¢cation rates DEPC Ethyl acetamidate Tetranitromethane
Control (min31 )
L-Glu
(min31 )
0.153 þ 0.007 0.060 þ 0.002 0.0277 þ 0.0014
0.00675 þ 0.00014 0.0058 þ 0.0004 0.0145 þ 0.0007
activity of 2.9% at 20 min. Similar results were obtained for the reaction with tetranitromethane for the activator, yielding a k higher than the value of k in absence of L-leucine. The residual activity found was only 1.2% at 30 min in the presence of L-leucine, whereas the presence of glutarate has, on the contrary, a protective e¡ect against modi¢cation, with a smaller k. The residual activity was 31.7% at 45 min in the presence of glutarate. All these k values measured are summarized in Table 5. 4. Discussion Two widely used techniques for the identi¢cation of functional groups, and, hopefully, to identify the role they play in catalysis are pH dependence of the kinetic parameters and chemical modi¢cation [22,23]. The reactions catalyzed by all hexameric glutamate dehydrogenases (and probably by all L-K-amino acid dehydrogenases) appear to involve a very speci¢c mechanistic feature. This feature consists in the nucleophilic attack on the K-iminoglutarate intermediate (which results from the preceding hydride transfer step) by water molecule hydrogen bonded to the deprotonated O-nitrogen atom to a highly conserved lysine whose pK is abnormally low [18,19,24,25]. The group that appears in the pro¢le of V is similar to those reported for all glutamate dehydrogenases. It has an abnormally low pK and must be deprotonated to achieve its function, so we may assume this residue to be this highly conserved lysine. Only one residue of lysine seems to be present in each enzyme subunit, so we may conclude that it should be the residue discussed for the V pro¢le. The role that has always been assigned to this residue is the binding of the water molecule. It would play this role due to its ability to coordinate water molecules, making sure the water nucleophilic attack
NAD (min31 )
L-Leu
(min31 )
0.048 þ 0.002 0.096 þ 0.009 0.0129 þ 0.0004
0.047 þ 0.002 0.168 þ 0.007 0.070 þ 0.002
Glutarate (min31 ) 0.202 þ 0.008 0.079 þ 0.004 0.0119 þ 0.0005
would be in the right position, i.e., very close to the K-carbon. pH dependence of the parameter V/K for L-glutamate provides a pro¢le that yields pKs for groups involved in binding and/or catalysis-isomerizationproduct release. While pKs in V pro¢les can be perturbed in either direction, the pKs in the V/K pro¢les can only be perturbed outward. That is, the pKs for groups that ionize on the acidic limb of the pro¢le can be perturbed to the left on the pro¢le, to a lower pH, while pKs that ionize on the basic limb of the pro¢le can be perturbed to the right on the pro¢le, to a higher pH, but this perturbation is usually 6 1 pH unit [26,28]. The pK estimated for the acidic group at 30³C is 7.9 and decreases slightly with temperature (Table 1). This pK value is close to the values for histine residues that range from 5.5 to 7 [23], although slightly higher probably due to the shape of the pro¢le. The dependence of the pK value with temperature gives an vH ion of 5.7 kcal/mol, which is in good concordance with the assignation of this residue as histidine (enthalpy of ionization in the range from 6 to 7.5). In our assays, the treatment with hydroxylamine slightly increases the activity of the modi¢ed enzyme, but the reactivation of the enzyme was not achieved. The N-carbethoxyimidazole formed can be deacylated with hydroxylamine, histidine being regenerated and this can be used to distinguish whether lysine or cysteine residues have been modi¢ed [29]. However, this does not exclude the modi¢cation of histidine. Avaeva and Krasnova [30] suggested that these results might be due to the occurrence of irreversible steps in the modi¢cation reaction in the presence of excess diethyl pyrocarbonate, it being necessary to include agents for the stabilization of the enzyme [15]. Other possible explanations would be that histidine residue(s) were also involved in the active conformation of the active site, being irreversibly lost
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
when carbethoxylation is performed. Halophilic NAD-GDH is very sensitive to changes such as being frozen, irreversibly losing all its activity; or it may be simply that we have not found the adequate condition to decarboxylate the histidine residues modi¢ed. pH dependence of the modi¢cation reaction yields a pK for the modi¢ed group of pK1 = 6.6 [13,14]. This result is in very good concordance with the modi¢cation of an histidyl group, and di¡erent enough from the value obtained in the modi¢cation of the lysyl group with ethyl acetimidate to allow us to consider them as di¡erent groups. Besides, the reaction with DEPC was very fast and modi¢cation of lysyl or cysteinyl groups is very slow. The decrease in absorbance at 278 nm can only be observed when the enzyme is modi¢ed at times longer than an hour of reaction, times much longer than the assay times for histidine modi¢cation with DEPC. The other group or groups that ionize on the basic limb of the pro¢le and so are preferentially protonated for binding and/or catalysis, exhibit a pK of 9.8 at 30³C, a value that also slightly decreases with temperature, yielding a vH ion of 6.6 kcal/mol. This result leads to the assignation of this group to a residue of tyrosine, whose pK values usually range from 9.8 to 10.5 and shows a vH ion of 6 kcal/mol, close to the enthalpy of ionization found for our NAD-GDH residue, although slightly lower in the range of error. Similar results can be observed in the pH pro¢le for the inhibition constant Kis with glutarate as a dead-end competitive inhibitor. At 30³C, the pK2 obtained is 9.59 and also shows a relatively small decrease with temperature that leads to a low vH ion value of 6.6 kcal/mol. It also does support the interpretation that it is ionization of an enzymic residue that is being observed, and not that the pK de¢ned by the high pH side of the pro¢le arises from glutamate itself (pK matches that of the amino group of glutamate). These results are consistent with the ionization of tyrosine. Addition of a competitive inhibitor and protonation^deprotonation of the enzyme comes to equilibrium, and thus dissociation constants are obtained and the pH dependence of the dissociation constants yields true pK values [20,31]. The Kis pro¢le provides true thermodynamic pK values for groups that are involved in binding [26].
523
All the other groups assigned are deprotonated, and, if the assignment is right, all are neutral. There must be residues involved in binding the two dicarboxylic groups of the substrate, since it exhibits negative charges likely to be positively charged. The residues usually involved in playing this role in other glutamate dehydrogenases, and also for other dehydrogenases such as leucine dehydrogenase from Bacillus stearothermophilus, are lysine residues [1,3,17,19] [32]. But according to our results, the lysine residue would play a catalytic role. Cleland [23] postulated that the 1-carboxyl group was hydrogen bonded to a formally neutral group, the 5-carboxyl being the group bound to lysine. So we postulate that the two tyrosine residues found for NADGDH are involved in the binding of both carboxylic groups on the substrate. The same role we propose for these two tyrosyl groups is that assigned for serine-380 in the glutamate dehydrogenase from Clostridium symbiosum (serine-381 in the bovine liver GDH sequence [1] in the mechanism described by Maniscalco et al. [19]). The rate constant has been taken as a useful parameter to study if the addition of metabolites exerts any in£uence on the modi¢cation reaction. The protection obtained by the addition of L-glutamate against both DEPC and ethyl acetimidate is related to the increase in the rate of modi¢cation in both groups when the inhibitor glutarate is present in the reaction mixture. So, the protection by L-glutamate is equivalent to the contrary e¡ect achieved by glutarate. With DEPC, L-leucine and NAD show a protective e¡ect; however, the opposite e¡ect is seen with ethyl acetimidate. These di¡erences would show that we are modifying di¡erent groups. For both histidine and lysine the presence of glutarate gives rise to an increase in the rate of the modifying reaction, showing that these residues could be not only involved in catalysis but also interacting between the hydrogens of the L-glutamate substrate. Glutarate as well as the other inhibitors were stated to be electron-donating molecules in a previous paper [21]. When an L-glutamate molecule is present, it would ¢t the active site, making the reaction of modi¢cation (protecting e¡ect) more di¤cult. When it is glutarate that binds (carboxylic groups), and negative charge is transferred, the new distribution of charge would facilitate the interactions of the glutarate K-hydro-
BBAGEN 24760 28-1-99
524
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
gens with both lysine and histidine residues. Charge would accumulate in these residues, thus enhancing the oxidation modifying reactions. The presence of glutarate protects the enzyme against TNM reaction, so the binding of glutarate probably involves the tyrosyl residues as previously mentioned. Since glutarate has no amino group, the reaction would not take place in spite of the increase in charge in the other groups. The protection achieved by L-glutamate shows a close relation with the acceleration of the rate of modi¢cation achieved with the activator L-leucine. Activators (L-amino acids) for the halophilic NADGDH are known to be electron-accepting [21]. The protecting e¡ect on the histidyl residue may be due to this role, making the modi¢cation of this residue di¤cult by accepting charge. Since two histidines are involved, it must be the one that binds to NAD where the activator plays its role, facilitating the transfer of hydride. The presence of L-leucine makes lysine and tyrosine more reactive, probably polarizing the charge of these two residues to the abovementioned histidine. The di¡erent behaviour of activators and inhibitors with DEPC and TNM again makes evident that the most directly implied residues in activation and inhibition are histidine and tyrosine. This is also consistent with our previous ¢ndings on the study of activation and inhibition of this halophilic enzyme. These results broadly outline the enzyme active site (Fig. 6) and the main reaction steps. The proposed chemical mechanism consistent with the previous discussion is: 1. Binding of the NAD with a deprotonated histidine (previous results) [7]. 2. Hydrogen binding of each of the two carboxylic groups to either tyrosine. The fact of not binding to a positively charged group would partly explain, together with the high number of essential site-bindings located in di¡erent places, the high speci¢city of the enzyme for L-glutamate. A weak union would require the substrate to be ¢tted better to the site. 3. Interaction of one of the N-hydrogens of the L-glutamate amino group (positively charged) with one deprotonated histidine. The substrate molecule L-glutamate must lose a proton from
4.
5.
6. 7.
its K-amino group prior to the hydride transfer step. Possible interaction between the K-hydrogen of L-glutamate and the deprotonated lysine.This conservative lysine has a remarkable ability to coordinate water molecules, so it could also interact with this hydrogen. The net result of the interaction or binding to these general bases (lysine and histidine) is the accumulation of charge in the vicinity of K-hydrogen. Binding of a water molecule to a deprotonated lysine. This coordinated water molecule would be involved in the subsequent nucleophilic attack. So, deprotonation of lysine may be a previous step as discussed extensively in available literature for glutamate dehydrogenase [18,19]. Hydride transfer to the C4 of the coenzyme nicotinamide ring rendering the related K-imino acid [17,32]. Nucleophilic attack by a water molecule bounded to the lysine residue to yield the intermediate Kcarbinolamine, which then decomposes to the keto acid, liberating ammonia.
A curious fact shown in the previous ¢ndings is that all the residues involved in catalysis are neutral, even those that arise from positively charged amino acids. This fact could be related to the special environment to which this archaeal glutamate dehydrogenase from H. salinarum is adapted. Halophilic proteins are usually highly acidic proteins and, as a result, cations would be required to neutralize the acidic groups in order to keep the protein in the proper conformational state [33]. The presence of a charge is accompanied by salt ions of the opposite charge, so that a neutral zone in the protein would not be coordinated by salt ions. This could be a way to maintain the binding or catalytic sites as free as possible of salt ions that could interfere with interactions. These mechanisms would thus show a somehow `halophilic character'. Acknowledgements We thank M. Aniorte, D. Pastor and V. Pina for help with preparation of the manuscript. This research is supported by CICYT Grant PB95-0695
BBAGEN 24760 28-1-99
F. Pe¨rez-Pomares et al. / Biochimica et Biophysica Acta 1426 (1999) 513^525
and Generalitat Valenciana Grant GV-1170/93 to M.J.B. References [1] E.L. Smith, B.M. Austen, K.M. Blumenthal, J.F. Nyc, in: P.D. Boyer (Ed.), The Enzymes vol. XI, Academic Press, New York, 1975, pp. 294^366. [2] K.S. Lilley, P.C. Engel, Eur. J. Biochem. 207 (1992) 533^ 540. [3] P.J. Baker, K.L. Britton, P.C. Engel, G.W. Farrants, K.S. Lilley, D.W. Rice, T.J. Stillman, Proteins 12 (1992) 75^86. [4] D. Piszkiewicz, E.L. Smith, Biochemistry 10 (1971) 4538^ 4552. [5] R.B. Wallis, J.J. Holbrook, Biochem. J. 133 (1973) 183^ 187. [6] K.M. Blumenthal, E.L. Smith, J. Biol. Chem. 248 (1973) 6002^6008. [7] M.L. Camacho, M.J. Bonete, E. Cadenas, Int. J. Biochem. 25 (1993) 979^985. [8] M.J. Bonete, M.L. Camacho, E. Cadenas, Biochim. Biophys. Acta 990 (1989) 150^155. [9] M.J. Bonete, M.L. Camacho, E. Cadenas, Int. J. Biochem. 18 (1986) 785^789. [10] W.W. Cleland, Methods Enzymol. 63 (1979) 103^138. [11] R.B. Freedman, in: Techniques in Protein and Enzyme Biochemistry, B117, Elsevier/North-Holland Biomedical Press, 1978, pp. 1^23. [12] F.C. Church, R.L. Lundblad, C.M. Noyes, J. Biol. Chem. 260 (1985) 4936^4940. [13] S. Kochhar, P.E. Hunziker, P. Leong-Morgenthaler, H. Hottinger, J. Biol. Chem. 267 (1992) 8499^8513.
525
[14] D.P. Dumas, F.M. Raushel, J. Biol. Chem. 265 (1990) 21498^21503. [15] E.W. Miles, Methods Enzymol. 47 (1977) 431^442. [16] M.J. McPherson, A.J. Baron, K.M. Jones, G.J. Price, J.C. Wootton, Protein Eng. 2 (1988) 147^152. [17] J.E. Rife, W.W. Cleland, Biochemistry 19 (1980) 2328^2333. [18] N. Singh, S.J. Maniscalco, H.F. Fisher, J. Biol. Chem. 268 (1993) 21^28. [19] S.J. Maniscalco, S.K. Saha, P. Vicedomine, H.F. Fisher, Biochemistry 35 (1996) 89^94. [20] W.W. Cleland, Adv. Enzymol. 45 (1977) 273^383. [21] M.J. Bonete, F. Pe¨rez-Pomares, J. Ferrer, M.L. Camacho, Biochim. Biophys. Acta 1289 (1996) 14^24. [22] A. Fersht, in: Enzyme Structure and Mechanism, 2nd ed., Freeman, New York, 1985, p. 155. [23] W.W. Cleland, Methods Enzymol. 87 (1982) 390^405. [24] N.M.W. Bruhuber, J.S. Blanchard, Crit. Rev. Biochem. Mol. Biol. 29 (1994) 415^467. [25] W.E. Karsten, R.E. Viola, Biochim. Biophys. Acta 1077 (1991) 209^219. [26] D.S. Sem, C.B. Kasper, Biochemistry 32 (1993) 11539^ 11547. [27] D. Bhattacharyya, U. Bandyopadhyay, R.K. Banerjee, J. Biol. Chem. 267 (1992) 9800^9804. [28] I.H. Segel, in: Enzyme Kinetics. Behavior and Analysis of Rapid Equilibrium and Steady-state Enzyme Systems, Wiley, New York, 1975, pp. 884^943. [29] P. Tomme, S. Chauvaux, P. Be¨guin, J. Millet, J.P. Aubert, M. Claeyssens, J. Biol. Chem. 266 (1991) 10313^10318. [30] S.M. Avaeva, V.I. Kransnova, Bioorg. Khim. 1 (1975) 1600. [31] A.J. Berdis, P.F. Cook, Biochemistry 32 (1993) 2041^2046. [32] T. Sekimoto, T. Matsuyama, T. Fukui, K. Tanizawa, J. Biol. Chem. 268, (36) (1993) 27039^27045. [33] H. Larsen, Adv. Microb. Physiol. 1 (1967) 97^132.
BBAGEN 24760 28-1-99