Eur. J. Biochem. 93, 263-270 (1979)
Mitochondria1 GTP-AMP Phosphotransferase 2. Kinetic and Equilibrium Dialysis Studies Alfred0 G. TOMASSELLI and Lafayette H. NODA Department of Biochemistry, Dartmouth Medical School, Hanover, New Hampshire (Received July 17, 1978)
Kinetic and equilibrium dialysis substrate binding studies have been done to investigate the properties of mitochondria1 GTP-AMP phosphotransferase. The results show that the enzyme has a specific requirement for divalent metal ions, namely Mg2+, Mn2+ or Ca2' (Ca2+ is active only in the forward direction, the direction of formation of ADP). The reaction rate depends upon the ratio [Mg2+]: [substrate] rather than on the metal ion concentration alone. The enzymatic activity is influenced by NaCl (or KCl) and optimum pH occurs at 11.5 and 9.5 for guanosine and inosine nucleotides respectively. Examination of binding of substrates to the enzyme showed that there is one binding site (GTP site) for MgGTP, GTP, MgGDP or GDP per molecule of enzyme, with dissociation constants of 4.5, 4.4, 3.0, 2.2 pM respectively and one binding site (AMP site) for AMP, ADP or ATP per molecule of enzyme with dissociation constants of 20.9, 33.4 and 33.4 pM respectively. Since, within the limitations of equilibrium dialysis used in the present studies, AMP binding to one site of the enzyme could be detected only when GDP or GTP is present, the mechanism of the forward reaction may be assumed to be nearly ordered. For the reverse reaction there is no requirement of order of binding of the two nucleotides and so the mechanism of reaction may be assumed to be random. In the preceding paper [l]we describe the purification of the enzyme, GTP-AMP phosphotransferase, from beef heart mitochondria, its amino acid composition and some physical and chemical properties. Kinetic studies were undertaken to explore in detail a variety of physical and chemical properties, including activity dependence on divalent cations, on salt, on the pH, on the temperature and substrate specificity of the enzyme. Equilibrium dialysis studies were made of the interaction of the enzyme with a variety of substrates in order to measure dissociation constants and binding stoichiometries. MATERIALS AND METHODS
Substrates and Reagents Unlabeled nucleotides were obtained from Sigma Chemical Co. Concentrations of stock solutions of nucleotides were .determined spectrophotometrically and stored in a deep freeze. Glycine, glycylglycine, triethanolamine, Mops, Hepes and arginine hydroAbbreviations. Mops, 3-(N-morpholino)propanesulfonic acid; Hepes, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid. Enzymes. GTP-AMP phosphotransferase (EC 2.7.4.10); ATPAMP phosphotransferase, adenylate kinase, myokinase (EC 2.7.4.3); GTPase (EC 3.6.1.-); nucleosidediphosphate kinase, NTP-NDP kinase (EC 2.7.4.6).
chloride were purchased from Sigma Chemical Co. Poly(ethy1ene glycol)-400 ( M , about 400) was purchased as Carbowax from Fisher Scientific Co. Buffers were prepared daily or stored at 4 "C for short periods of time. Sodium hydroxide added in adjusting pH of buffers was taken into account in calculating Na' concentration. Salts and other reagents were of analytical grade. A stock 1.O M solution of MgC12 was standardized against recrystallized 8-hydroxy-quinoline [2] and stored at 4°C. All water used was distilled and deionized. Macherey-Nagel poly(ethy1ene imine)-cellulose on plastic sheets (Polygram Cel 300 PEI/UV254) were 20 x 20-cm fluorescent sheets from Brinkmann Instruments Co. Membrane for dialysis equilibrium, Spectra/Por 2, molecular weight cut-off 12000 - 14000 was purchased from Spectrum Medical Industries, Inc., Los Angeles. Biofluor scintillation fluid (emulsifier cocktail) and most 14C-labeled nucleotides of adenine and guanine were purchased from New England Nuclear, Boston. The [U-14C]GDP was obtained from Amersham (Arlington Heights, Ill., U.S.A.).
General Method of Assay The time course appearance of a ['4C]nucleotide upon incubation of the enzyme in the presence of appropriately labeled and unlabeled nucleotide was measured after separation of the nucleotides by thin-
Mitochondria1 GTP-AMP Phosphotransferase
264
layer chromatography on poly(ethy1ene imine)-cellulose sheets. The reaction mixtures were prepared from stock buffer/salt mixtures at the desired pH containing double the final Na' concentration to allow for further additions of reagents as required for each particular experiment. A 0.5-ml aliquot of the mixture was transferred to a 3-ml vial provided with a magnetic stirrer and placed in a circulating water bath. A calculated amount of ['4C]nucleotide (0.1 - 1.0 pCi) was added to the reaction mixture, when it was significant the increase in volume due to this addition was taken into account. Two zero-time aliquots (5 -25 pl) were applied to the poly(ethy1ene imine)-cellulose 3 cm from one edge. Then, as the clock was started 5 p1 of diluted enzyme was added to the reaction mixture. At l-min intervals, 50-pl aliquots were added to 5 pl 0.5 M EDTA pH 8.2 held in tubes in an ice bath. Aliquots of the EDTA mixture (5 - 25 pl) were applied to the sheet of poly(ethy1ene imine)-cellulose along the line of zero-time samples. Spots were dried in a stream of cold air. The chromatogram was developed at 4 "C for 5 min in 0.3 M LiCl, 30 min in 1.0 M LiCl and in 1.6 M LiCl until the solvent front was at least 13 cm from the origin. The well-resolved spots containing [14C]nucleotides were located with an ultraviolet lamp. Spots were cut out and radioactivity counted in a Packard Tri-Carb model 3385 liquid scintillation spectrometer. Toluene scintillation fluid [3] was made fresh daily. The radioactivity was plotted against time and, from the slope of the best line drawn through the points, the initial reaction rate was calculated. Equilibrium Dialysis
Dialysis was carried out in chambers made from matching paired plexiglass plates with six shallow cylindrical depressions cut into each (6.5-mm diameter, 1.5-mm depth, about 50-pI capacity). A tiny hole cut at the top of each chamber allowed access to each side; 14 small stainless steel bolts were used to compress the plates together. Individual Spectra/Por 2 membranes cut from dialysis tubing were placed over each depression of one plate and the second plate was screwed into place over the first plate to create two chambers separated by membrane. The membranes were prepared by boiling in a 20 mM Na~C03,0.5mM EDTA solution followed by rinsing and boiling four or five times in deionized distilled water. The membrane was cut into small rectangles and stored in distilled water at 4°C. Before each experiment the membranes were soaked in buffer for a minimum of 15 min, then blotted dry and placed in position. Aliquots of 40 pl of a solution having a final composition of 0.02 M buffer, 0.15 M NaC1, 2% poly(ethy1ene glycol)-400 and 1 mg/ml of protein
were placed in each of the six dialysis chambers located on one side of the membrane. An equal volume of the same solution in which the enzyme was replaced by increasing concentrations (from 0.01 mM to 0.6mM) of a specified 14C-labeled nucleotide was placed in each of the six chambers on the opposite side of the membrane. Each concentration of nucleotide was run in duplicate. To prevent loss by evaporation the delivery holes of the chambers were sealed by tape. The apparatus was placed on a wrist-action shaker and allowed to come to equilibrium overnight (16 - 17 h, usually at 4°C). Under the conditions used in our experiments, no loss of enzymatic activity was observed even when samples were left in the shaker for 24 h at 25 "C. After equilibration, 25 pl of sample was taken from each chamber, delivered to 5 ml of Biofluor scintillation fluid and 14C was counted. Equilibrium data were analyzed by Scatchard plots [4]. In many experiments 10 pl of sample from several chambers were chromatographed on poly(ethy1ene imine)-cellulose sheets to see if nucleotides different from those expected after equilibration were present. RESULTS Stoichiometry and Linearity of the Reaction
Complete assays were made for samples at l-min intervals. For the forward reaction when 1 ml of reaction mixture composed of 50 mM glycylglycine, 75 mM Na', 5 mM [14C]GTP, 5 mM [I4C]AMP and 5 mM MgC12 at pH 8.0 was incubated at 30 "C with 1.45 pg enzyme at 7 min the following changes were found: GTP decreased 0.859 pmol (k 0.4 %) and GDP increased 0.816 pmol (k 0.3 %);AMP decreased 0.820 pmol (i0.8 %) and ADP increased 0.780 pmol (k 0.6 %). Or, overall changes in substituents were 1 : 1 within 4.8 % averaging all seven determinations. In the reverse direction (conditions as above except initial substituents were 5 mM [14C]GDP, 5 mM [14C]ADP and 0.54pg enzyme) changes found for sample at 7 min were ADP decreased 0.980 pmol (f 1.0 %) and AMP increased 1.01 pmol (f 0.2 %); GDP decreased 0.950 pmol ( f 0.5%) and GTP increased 0.990 pmol (+ 0.6%). Thus, in the reverse direction substituent changes were 1 : 1 within k 3.4 averaging all seven determinations. Time course assays of this and other reactions showed that the rate of reaction was linear until over 10 % of the substrates present were consumed. Requirement for Divalent Cations
Enzymatic activity has been detected only in the presence of specific divalent metal ions, namely Mg", Mn2+ and Ca2'- (effective only in the forward direc-
265
A. G. Tomasselli and L. H. Noda Table 1. Metal ions and specific activity Chromatographic assays were conducted after incubation with various metal chlorides ( 5 mM) in 0.05 M glycylglycine, 0.075 M Na’, pH 8.0, 30°C and 0.257 pg of enzyme. Forward reaction: GTP and [14C]AMP were 5.63 mM. Reverse reaction: GDP and [14C]ADP were 5.0 mM. Results are given in U/mg enzyme and in parentheses, as a percentage of the activity with MgZ+
I% 500 E
.
- 400 .’g 300 3
A
+
m
Specific activity in
Cations
600
m
E, 200
forward reaction
reverse reaction
U/mg (%I
c N
w
100
C
0
Mg2‘ Mn2’ CaZ BaZt, Cu2+,SrZ+, Fez+,Fe3+ +
124 (100.0) 100 (80.0) 84 (67.7) 0
(0)
397 (100.0) 143 (36.0) 0 (0) 0
(0)
0.2
0.4
0.6
0.8
“a+]
(M)
1.0
1.8
2.0
Fig. 2. Effect of salt concentration on enzymic activity at p H 8.0 in theforwnrd reaction (0) and the reverse reaction (0). Conditions of incubation were as in Fig. 1 except MgCIz was constant at 5 mM and Na’ was varied. The amount of enzyme was 0.257 and 0.128 pg in 0.5 ml for the forward and reverse reactions respectively
380 340
- 300
.
- 260 3
> + .-
5
220
m a,
180
&
140
5.
100
60 C
Fig. 1. Effect of magnesium ion concentration on enzymic activity in theforwardreaction ( 0 )and the reverse reaction (0). The reaction mixture of 0.5 ml contained 5 mM of each nucleotide, 75 mM Na+, 50 mM glycylglycine at pH 8.0 and 30”C, with 0.31 and 0.25 pg enzyme for forward and reverse reactions respectively and with [14C]AMP or [I4C]ADP
tion), while Ba2+, Cu2’-, Sr2‘-, Fez+ and Fe3+ were ineffective (Table 1). Under the conditions used, optimal enzymatic activity was obtained when the ratio of [Mg2+]:[GTP] is 0.6 and when [Mg”] is equal to ’/, ([GDP] + [ADP]) in the forward and reverse directions respectively (Fig. 1). With increasing concentrations of Mg2’ the activity increases to a maximum, and then levels off showing its relation to the ratio [Mg2+]: [nucleotides] rather than to the absolute amount of Mg2+. Effect of Sodium Chloride on Activity andpH Optimum Substitution of 0.1 M Na’ with 0.1 M K + at pH 8.0, 30°C and the usual concentrations used for
the other constituents gave 95% of the activity observed with Na‘.. Hence experiments on salt effects were conducted with NaC1, taking into account the total Na’ ion concentration. Fig. 2 shows that at pH 8.0 a concentration of 0.8 M Na+ gives maximum activity. In similar experiments at pH 11.5 (a pH optimum, see below) it was found that maximal enzymatic activities were at 0,075-0.10 M Na+ and that further increases of Na’ to 0.80 M resulted in slightly decreased activity. Thus the effect of varying pH at various salt concentrations was studied. Effect o f p H at Various Salt Concentrations In the forward direction at Na’ concentrations 0.075-0.80 M the pH optimum is at 11.5 with arginine . HCl as buffer (Fig. 3). In the reverse reaction at Na’ concentration 0.075 M, the pH optimum is at 11.5 (other concentrations of Na+ were not tested). In the forward direction at pH 11- 11.5 additions of NaCl above 0.075 M resulted in slightly decreased activity, but at lower pH the addition of salt above 0.075 M resulted in increased activity. For the reaction ITP + AMP IDP ADP at 0.075 M Na’ and with the same conditions as with the GTP reaction the pH optimum 9.5 instead of at pH 11.5.
+
Substrate Specificity Various combinations of nucleotides were tested for their ability to act as substrates of the purified enzyme. Table 2 summarizes part of the data obtained. Not listed are CMP, GMP, IMP, UMP and CAMP giving no detectable activity under the usual conditions of assay with [14C]GTP as the other substrate.
266
Mitochondria1 GTP-AMP Phosphotransferase r
i
I
I
I
I
I
I
I
I
I
- 500
AMP as the other substrate. Inosine can substitute for guanosine phosphates and a comparison at their best pH values, i.e. 9.5 and 11.5 respectively, are shown in Table 2. However the values obtained are not necessarily strictly comparable, since the optimal conditions of the reaction and details of the kinetics for inosine nucleotides were not thoroughly investigated. dAMP gave 40% of the AMP activity when incubated with GTP at pH 8.0. The enzyme preparation, even at the concentration of 3.1 pg/0.5 ml, was unable to catalyze the following reactions: [14C]ATP ['"ClAMP i2 2 ['4C]ADP and ['"CIGTP HzO + ['"CIGDP + Pi and so it can be considered free of ATPAMP phosphotransferase and GTPase. [14C]GDP either alone or in combination with CDP, IDP or UDP gave no reaction. Since the incubation of GTP and ['4C]ADP with enzyme yielded no ['"CIATP nucleoside diphosphokinase was not present.
+
PH
Fig. 3. Eflect o f p H at various Na' concentrations on enzymic activity in theforward (0,0.075 M Nu'; A, 0.40 M Nai ; x ,0.80 M Na' ) and the reverse reaction (0, 0.075 M Na'). Nucleotides and magnesium chloride were each 5 mM and reactions (0.5 ml) were carried out at 30°C with 0.257 or 0.170 pg enzyme for the forward and reverse reactions respectively. Buffers, 0.05 M, used were: pH 4-6 sodium acetate, pH 6.5 - 7.5 Mops, pH 8 - 9 glycylglycine,pH 9.5 11.O glycine and pH 11.5- 12.5 arginine . HCl/NaOH. [14C]AMP or [14C]ADPwas used as labeled substrate
+
Effect of Temperature Table 2. Substrate speci$city of GTP-AMP phosphotransferase All nucleotides were 5 mM with 5 mM MgClz in 0.05 M buffer, 0.075 M Na' with arginine at pH 11.5, glycine at pH 9.5 and glycylglycine at pH 8.0. All incubations were at 30°C with 0.1290.257 pg enzyme in 0.5 ml assay mixture (with the exception of [I4C]AMP UTP in which 1.55 pg of protein was used)
+
Substrates
PH
_ 14c-
labeled
Product
_
_
unlabeled
(14C~
Specific activity
Relative activity
labeled)
Apparent energies of activation at pH 8.0 and optimal magnesium ion substrate ratios, [Mg2+]/ [GTP] = 0.6 and [Mg2+]/([GDP] [ADPI) = '/2, calculated from Arrhenius plots are 68.2 kJ/mol for the forward reaction and 43.9 kJ/mol for the reverse reaction.
+
Equilibrium Constant U/mg
11.5
AMP ADP
GTP GDP
ADP AMP
436 575
1.o 1.3
9.5
AMP ADP
ITP IDP
ADP AMP
325 437
1.0 1.3
8.0
AMP AMP AMP ADP ADP
GTP ITP UTP GDP IDP
ADP ADP ADP AMP AMP
110 175 6 397 302
1.o 1.6 0.05 3.6 2.8
8.0"
AMP AMP GTP GTP
GTP dGTP AMP dAMP
ADP ADP ADP GDP
126 126 121 51
1.o 1.o 1.o 0.4
* MgC12 concentration was 3 mM instead of 5 mM.
Equilibrium was approached in both directions under the following conditions: 25"C, pH 8.5 in 0.05 M triethanolamine buffer, 0.075 M KCI, 1 mM EDTA, 20 mM Mg2+, 5 mM GTP and 5 mM AMP in the forward direction and 5 mM GDP, 5 mM ADP in the reverse direction. The amount of enzyme used was 20.6 pg added to 0.5 ml of reaction mixture. The apparent equilibrium constant under the above conditions was Kapp= [ADP] [GDP]/[GTP] [AMP] = 1.54. Michaelis Constants
The values in Table 3 were obtained using the chromatographic method of assay and avoiding the use of auxiliary enzymes. Binding Studies
Thus it may be concluded that the AMP (or ADP) binding site is specific for adenosine. CTP does not serve as substrate with AMP even when the amount of enzyme was raised from the usual 0.128-0.257 pg to 1.55 pg per 0.5 ml of reaction mixture. At similar high enzyme concentration UTP gives only about 5 % of the activity found for GTP. Under the conditions tested, dGTP served as well as GTP as substrate with
All the equilibrium binding data were analyzed by Scatchard plots. Straight lines were drawn through the data points by the method of least squares. Fig.4-6 show some of those plots and provide explanations about the method used to obtain them. Table 4 summarizes the results obtained for both the dissociation constants and binding stoichiometries of nucleotides and Mg2+-nucleotide complexes. Nucleo-
267
A. G. Tomasselli and L. H. Noda Table 3. Michaelis constants of various substrates for GTP-AMP phosphotransferase Michaelis constants were determined at 25 "C in 0.05 M triethanolamine . HCI, 0.075 M KCI, 1 mM EDTA, 5 mM MgClz and with saturating nucleotides at 5 mM. Initial rates of reaction were determined for at least five different concentrations of variable substrate (0.10-1 mM) Substrate
I 1
0.45
t
I
I
0.2
0.4
I
I
I
a8
1.0
K, mM
AMP GTP ADP GDP ri
0.139 0.143 0.074 0.036 I
1
I
I
0 0
1%
0.6 1 [El,
Fig. 5. Scatchard plot for equilibrium ligand binding. (A) [14C]GDP binding obtained under conditions given in Fig. 4
I
Fig. 4. Scatchard plots for equilibrium ligand binding. (0) Mg[14C]GTP binding was obtained using the following conditions: 0.02 M Hepes buffer, 0.15 M NaCI, 2% poly(ethy1ene glycoI)-400 and 1 mg/ml protein on one side of the membrane; on the other side of the membrane 0.02 M Hepes buffer, 0.15 M NaC1, 2 % poly(ethylene glycol)-400, increasing concentrations from 0.01 M to 0.6 mM labeled nucleotides and corresponding concentrations of MgClz equal to twice the concentration of nucleotide. (0)[14C]GTP binding was under the above conditions and omitting MgC12. Plots were made according to the equation
derived by modifying the Eadie-Scatchard equation (or derived directly from the equilibrium expression for KJ where [S]b, the concentration of bound substrate, = [ES], the concentration of occupied sites; [Sir is the concentration of the free substrate, [El, is the total enzyme concentration, K, = [E]f[S]r/[ES], is the intrinsic substrate dissociation constant of a site, [El, is the concentration of the free enzyme, n is the number of identical and independent ligand-binding sites per molecule of enzyme and on the plot is equal to the intercept on the x axis; n/Ks is the intercept on the y axis
tides that gave poor binding or did not bind at all within the limitations of enzyme and substrate concentrations are not listed. This was the case for AMP, GMP, CTP, and UTP for which equilibrium dialysis
I
I
I
I
[Slb/[El,
Fig. 6. Scatchard plots for equilibrium ligand binding. (A) (0)[14C]ATP binding, (0) ['4C]ADP binding obtained under conditions given in Fig. 4; (B) (A) [14C]AMPbinding obtained under conditions given in Fig. 4 and the additional presence of 5 mM GDP
Table 4. Dissociation constants and number of binding sites Conditions for equilibrium dialysis are specified in Fig.4. n = number of binding sites I4C-Labeled substrate
GTP MgGTP GDP MgGDP ATP ADP AMP AMP a
Unlabeled substrate
-
-
GDP" MgGDP"
Concentration of 5 mM.
n
KS
0.92 1.03 0.88 0.86 1.02 1.17 0.80 0.83
PM 4.4 4.5 2.2 3.0 33.4 33.4 20.9 23.1
Mitochondria1 GTP-AMP Phosphotransferase
268
shows a limitation in this type of study, namely that precise experimental data can not be obtained when the binding of a substrate is determined as the small difference between two large numbers, i.e. total substrate minus free substrate. Under the following conditions ['4C]AMP alone with enzyme was not bound within the limitations of experiment: at pH 8 and 4 "C or 25 "C in the presence and in the absence of Mg''., at pH 11.5 and 4 "C or 25 "C, with AMP concentrations ranging between 0.01 mM to 0.6 mM. But in the presence of MgGDP at pH 8 and 25 "C (not reported in Table 4) it was found that the enzyme bound AMP at 0.93 site with K, = 80 yM. Further data shown in Table 4 at pH 8.0 and 4 "C, indicate that AMP in the presence of 5 mM GDP binds to 0.80 site with a K, = 20.9 yM and in the presence of 5 mM MgGDP binds to 0.83 site with a K, = 23.1 pM. The binding of AMP is tighter at low temperature. The K, of AMP in the presence of 5 mM GTP could not be determined because even in the absence of added Mg2' and in the presence of 25 mM EDTA, [I4C]AMP was transformed to [14C]ADP as determined by thin-layer poly(ethy1ene imine) chromatography. ['4C]ADP binds to the enzyme with a K, = 33.4 yM and the number of sites occupied is 1.17; it is noteworthy that no partner substrate is required to bind [14C]ADP.At the end of the experiment the solutions from the dialysis cells were chromatographed and found free from nucleotides different from [14C]ADP indicating that no reaction took place in the process and that the binding observed was due to [14C]ADP. However when [14C]ADPwas run in the presence of 5 mM GTP, no [14C]ADP binding was observed and in the process a small amount of [14C]AMP was formed, about 5 -7% of the total [14C]ADPpresent. [I4C]ATP binds to 1.02 sites of the enzyme with K , = 33.4 yM. No partner substrate was needed for its binding; the presence of GTP or GDP prevents the binding of the [I4C]ATP to the enzyme; neither side reactions nor presence of unexpected nucleotides were encountered in these experiments. GTP binds to 0.92 site of the enzyme with K, = 4.4 pM and MgGTP occupies 1.03 sites with K, = 4.5 yM. The MgGTP binding was also tested at pH 11.5 and 4°C (not shown in Table 4) and the binding observed was 0.90 site with K, = 10.0 pM. GDPwas found to bind at0.88 sitewith%= 2.2yM and MgGDP gives 0.86 binding site with K, = 3.0 pM. UTP and CTP did not exhibit any binding as evidenced by the difference in radioactivity observed being too small to be significant because of the limitations of conditions of the experiments. Thus, the approximate 5 % of reaction observed in the kinetic studies when UTP and AMP were added to the enzyme might be due to a limitation of binding of the substrates, a limitation of reaction mechanism or by some trace contaminating enzyme.
GMP alone binds very poorly or not at all to the enzyme and the presence of ATP or ADP did not increase the binding of GMP. DISCUSSION GTP-AMP phosphotransferase from beef heart mitochondria was reported by Albrecht [5] as having a molecular weight of 52000, optimum activity at pH 8.5, specific requirement for divalent cations and specific for AMP as phosphate acceptor and unspecific for the phosphate donor since GTP, ITP and to a lesser extent UTP, CTP and ATP could fulfill this role. Results obtained in this work show that the enzyme, as purified according to the procedure described in the preceding paper [l], has an M , of 26000, pH optimum of 11.5 and 9.5 when guanosine and inosine phosphates are respectively substrates. The enzyme has a site specific for AMP or dAMP (ADP in the reverse direction) designated the AMP site. The second enzymatic site, designated the GTP site, transfers in the forward direction phosphate from GTP, dGTP or ITP. Since UTP gives only slight activity and CTP and ATP do not serve as phosphate donors in our enzymatic preparation, it appears likely that at least one other enzyme, capable of catalyzing the interaction with nucleotide phosphates, was present in the preparation reported by Albrecht. The Michaelis constants reported in Table 3 agree within about 4-fold of values reported by Albrecht [5] with the exception of K, for GDP which is 30-fold higher. Part of these differences may be due to dependence on coupled enzymes in assays of the earlier work. By kinetic studies of the activity of the enzyme some conclusions can be drawn about the characteristics that nucleotides must possess for enzymatic activity. The fact that the AMP site can accept only AMP, dAMP (or ADP) for reaction indicates that the 6-amino group of the purine base is essential. The importance of the ribose and 2'deoxyribose can be evaluated by observing that with dAMP the enzyme exhibits only 40% of the activity compared to AMP indicating that the 2'-OH group is preferred to the 2'-deoxy group, but is not essential for activity. GTP and dGTP when compared under the conditions reported in Table 2 are equally suitable indicating that the 2'-deoxy group works as well as the 2'-OH group. The further information that ITP is active with the enzyme and ATP is not, indicate that the 6-oxy group in the purine base is indispensable for activity. The -NH2 group in position 2 in guanine compared to a hydrogen in inosine, may be important in determining the pH optimum for activity since it was observed that optimal activity was at pH 13.5 for guanosine nucleotide and at pH 9.5 for inosine nucleotide. The lack of activity observed with CTP and the very poor activity offered by UTP then indicates that the nature
269
A. G. Tomasselli and L. H. Noda
of the base is a very important feature for GTP-AMP phosphotransferase activity. The presence of a divalent cation, namely Mg2+, Mn2', or Ca2',is indispensable for enzymatic activity. The reaction rate is determined by the ratio [Mg2+] : [nucleotide] rather than by the independent concentration of metal ion alone. This behavior has been reported for myokinase [6], adenylate kinase [7,8] and NTP-NDP kinase [9] from bovine liver mitochondria. The requirement of a divalent cation for activity is a property shared by all kinases in general. However, present studies show that the presence of a divalent cation is not required for the binding of the nucleotides of GTP-AMP phosphotransferase. Since the dissociation constants for free nucleotide GTP or GDP compared to Mg-nucleotide from enzyme are very similar, the simplest explanation is that Mg2+ is not directly coordinated to the protein in these ternary complexes and that the metal ion is involved in facilitating the phosphate transfer taking place. Similar metal-enzyme relationships have been reported for a number of enzymes: by Mildred Cohn and associates for porcine muscle adenylate kinase [ 101 and for rabbit muscle creatine kinase [ l l - 131 using nuclear magnetic resonance studies of the binding of nucleotide to the enzyme in the presence and absence of Mn2+, and by Kuby et al. [14] for rabbit muscle adenylate kinase by using a sedimentation gradient procedure in the presence and in the absence of Mg2+. In contrast, in rabbit muscle pyruvate kinase [15- 171 metal ion has been shown by Mildvan and associates to bind to the enzyme and that in the ternary complex it is located between protein and nucleotide. While ADP binds to the AMP site (reverse reaction), the impossibility of ADP binding to the GTP site precludes the possibility of GTP-AMP phosphotransferase having adenylate kinase activity. The fact that neither ['4C]ADP in the presence of 5 mM GTP nor [I4C]ATP in the presence of 5 mM GTP or GDP bind to the enzyme can be explained by steric hindrance by the bulk of extra phosphate groups. When GTP or GDP is present on the GTP site of the enzyme, the obligatory choice of the partner becomes respectively AMP or ADP. UTP and CTP were found within the limitations of these binding studies not to bind to the enzyme so the slight activity observed with UTP AMP at relatively high concentrations of the enzyme might be explained by inherent trace activity with these substrates or a very small amount of contaminating enzyme in the preparation. From kinetic and binding studies a mechanism of reaction is proposed for the reaction catalyzed by GTP-AMP phosphotransferase. In the forward direction the binding of AMP alone to one site is too low to be detected with the method used in the present
+
studies; however the binding of AMP to one site is greatly enhanced by the presence of GDP (or GTP) on the other site and can be easily detected in that way (Table4). So the mechanism in the forward direction may be assumed as nearly ordered. In the reverse direction there is no requirement of substrate bound to one site to detect the presence of the partner substrate on the other site and hence the mechanism may be assumed random according to the Cleland nomenclature [18]. For many enzymatic reactions the optimum of activity at fixed pH is dependent upon the ionic strength of the medium [19 - 231 ; however, at present a complete understanding of the process has not been provided. As shown by data presented in Fig. 2 and 3, GTP-AMP phosphotrqnsferase activity seems to be related to increasing Na' concentration in such a way that slight inhibition occurs at pH values ranging between 11 and 11.5 and stimulation occurs at pH values lower than 10.5. This behavior is consistent with the assumption that increasing Na+ concentration is accompanied by a variation of the ionization constants of the catalytic groups of the active site or of the substrates. The unusually high pH optimum of 11.5 observed in vitro may not apply to the situation in vivo. But it seems possible, in view of the chemiosmotic mechanism of oxidative phosphorylation [24,25] and the mosaic nature of the matrix membrane [26], that a non-uniform distribution of charges on the matrix side of the inner mitochondria1 membrane may result. Zones of high electrical potential may result in a very heterogeneous distribution of protons thus creating compartments of relatively high and relatively low pH (without any influence on the average pH), i.e. zones favorable and unfavorable for GTP-AMP phosphotransferase activity. This research has been supported by Grant GM-23123 from the National Institutes of Health, United States Public Health Service, for which grateful acknowledgment is made. We wish to express appreciation for the careful and capable technical assistance of Brenda Shea during the course of this work.
REFERENCES 1. Tomasselli, A. G., Schirmer, R. H. & Noda, L. H. (1979) Eur. J . Biochem. 93,251 - 262. 2. Kolthoff, I. M. & Sandell, E. B. (1955) Textbook of' Quantitative Inorganic Analysis, 3rd edn, pp. 362 - 363, Macmillan Co., New York. 3. Nossal, N. G. & Singer, M. F. (1968) J . Biol. Chem. 243, 91 3 -922. 4. Scatchard, G. (1949) Ann. N.Y. Acad. Sci. 51, 660-672. 5. Albrecht, G . J. (1970) Biochemistry, 9, 2462-2470. 6. Noda, L. H. (1958) J . Biol. Chem. 232, 237-250. 7. Markland, F. S. & Wadkins, C. L. (1966) J . B i d . Chem. 241, 4136-4145. 8. Glaze, R. P. & Wadkins, C. L. (1967) J . Biol. Chem. 242, 2139- 21 50.
270
A. G. Tomasselli and L. H. Noda: Mitochondria] GTP-AMP Phosphotransferase
9. Goffeau, A., Pedersen, P. L. & Lehninger, A. L. (1967) J. Biol. Chem. 242, 1845 - 1853. 10. Price, N. C., Reed, G. H. & Cohn, M. (1973) Biochemistry, 12, 3322-3327. 11. Cohn, M. & Leigh, J. S. (1962) Nature (Lond.) 193, 10371040. 12. O’Sullivan, W. J. & Cohn, M. (1966) J. Biol. Chem. 241, 3104- 31 15. 13. Reed, G. H. &Cohn, M. (1972)J. Biol.Chem. 247,3073-3081. 14. Kuby, S. A , , Mahowald, T. A. & Noltmann, E. A. (1962) Biochemistry, I , 748 - 762. 15. Mildvan, A. S. & Cohn, M. (1966) J . Biol. Chem. 241, 11781193. 16. Mildvan, A. S., Leigh, J. S. & Cohn, M. (1967) Biochemistry, 6. 1805-1818.
17. Mildvan, A. S. &Cohn, M. (1965) J. Biol. Chem. 240,238-246. 18. Cleland, W. W. (1963) Biochem. Biophys. Acta, 67, 104-137. 19. Kalnitsky, G., Hummel, J. P., Resnick, H., Carter, J. R., Bernett, L. B. & Dierky, C. (1959) Ann. N.Y. Acad. Sci. 81, 542- 566. 20. Irie, M. (1965) J. Biochem. (Tokyo) 57, 355-362. 21. Barton, N. W., Lipovac, V. & Rosenberg, A. (1975) J. Biol. Chem. 250, 8462 - 8466. 22. Maurel, P. & Douzou, P. (1976) J. Mol. Biol. 102, 253-264. 23. Douzou, P. & Maurel, P. (1977) Trends Biochem. Sci. 2,14- 17. 24. Mitchell, P. (1966) Biol. Rev. 41, 445-502. 25. Mitchell, P. (1961) Nature (Lond.) 191, 144-148. 26. Singer, S. J. (1972) Ann. N . Y . Acad. Sci. 195, 16-23.
A. G. Tomasselli and L. H. Noda, Department of Biochemistry, Dartmouth Medical School, Hanover, New Hampshire, U.S.A. 03755
,
Mitochondria1 GTP-AMP Phosphotransferase 2. Kinetic and Equilibrium Dialysis Studies Alfred0 G. TOMASSELLI and Lafayette H. NODA Department of Biochemistry, Dartmouth Medical School, Hanover, New Hampshire (Received July 17, 1978)
Kinetic and equilibrium dialysis substrate binding studies have been done to investigate the properties of mitochondria1 GTP-AMP phosphotransferase. The results show that the enzyme has a specific requirement for divalent metal ions, namely Mg2+, Mn2+ or Ca2' (Ca2+ is active only in the forward direction, the direction of formation of ADP). The reaction rate depends upon the ratio [Mg2+]: [substrate] rather than on the metal ion concentration alone. The enzymatic activity is influenced by NaCl (or KCl) and optimum pH occurs at 11.5 and 9.5 for guanosine and inosine nucleotides respectively. Examination of binding of substrates to the enzyme showed that there is one binding site (GTP site) for MgGTP, GTP, MgGDP or GDP per molecule of enzyme, with dissociation constants of 4.5, 4.4, 3.0, 2.2 pM respectively and one binding site (AMP site) for AMP, ADP or ATP per molecule of enzyme with dissociation constants of 20.9, 33.4 and 33.4 pM respectively. Since, within the limitations of equilibrium dialysis used in the present studies, AMP binding to one site of the enzyme could be detected only when GDP or GTP is present, the mechanism of the forward reaction may be assumed to be nearly ordered. For the reverse reaction there is no requirement of order of binding of the two nucleotides and so the mechanism of reaction may be assumed to be random. In the preceding paper [l]we describe the purification of the enzyme, GTP-AMP phosphotransferase, from beef heart mitochondria, its amino acid composition and some physical and chemical properties. Kinetic studies were undertaken to explore in detail a variety of physical and chemical properties, including activity dependence on divalent cations, on salt, on the pH, on the temperature and substrate specificity of the enzyme. Equilibrium dialysis studies were made of the interaction of the enzyme with a variety of substrates in order to measure dissociation constants and binding stoichiometries. MATERIALS AND METHODS
Substrates and Reagents Unlabeled nucleotides were obtained from Sigma Chemical Co. Concentrations of stock solutions of nucleotides were .determined spectrophotometrically and stored in a deep freeze. Glycine, glycylglycine, triethanolamine, Mops, Hepes and arginine hydroAbbreviations. Mops, 3-(N-morpholino)propanesulfonic acid; Hepes, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid. Enzymes. GTP-AMP phosphotransferase (EC 2.7.4.10); ATPAMP phosphotransferase, adenylate kinase, myokinase (EC 2.7.4.3); GTPase (EC 3.6.1.-); nucleosidediphosphate kinase, NTP-NDP kinase (EC 2.7.4.6).
chloride were purchased from Sigma Chemical Co. Poly(ethy1ene glycol)-400 ( M , about 400) was purchased as Carbowax from Fisher Scientific Co. Buffers were prepared daily or stored at 4 "C for short periods of time. Sodium hydroxide added in adjusting pH of buffers was taken into account in calculating Na' concentration. Salts and other reagents were of analytical grade. A stock 1.O M solution of MgC12 was standardized against recrystallized 8-hydroxy-quinoline [2] and stored at 4°C. All water used was distilled and deionized. Macherey-Nagel poly(ethy1ene imine)-cellulose on plastic sheets (Polygram Cel 300 PEI/UV254) were 20 x 20-cm fluorescent sheets from Brinkmann Instruments Co. Membrane for dialysis equilibrium, Spectra/Por 2, molecular weight cut-off 12000 - 14000 was purchased from Spectrum Medical Industries, Inc., Los Angeles. Biofluor scintillation fluid (emulsifier cocktail) and most 14C-labeled nucleotides of adenine and guanine were purchased from New England Nuclear, Boston. The [U-14C]GDP was obtained from Amersham (Arlington Heights, Ill., U.S.A.).
General Method of Assay The time course appearance of a ['4C]nucleotide upon incubation of the enzyme in the presence of appropriately labeled and unlabeled nucleotide was measured after separation of the nucleotides by thin-
Mitochondria1 GTP-AMP Phosphotransferase
264
layer chromatography on poly(ethy1ene imine)-cellulose sheets. The reaction mixtures were prepared from stock buffer/salt mixtures at the desired pH containing double the final Na' concentration to allow for further additions of reagents as required for each particular experiment. A 0.5-ml aliquot of the mixture was transferred to a 3-ml vial provided with a magnetic stirrer and placed in a circulating water bath. A calculated amount of ['4C]nucleotide (0.1 - 1.0 pCi) was added to the reaction mixture, when it was significant the increase in volume due to this addition was taken into account. Two zero-time aliquots (5 -25 pl) were applied to the poly(ethy1ene imine)-cellulose 3 cm from one edge. Then, as the clock was started 5 p1 of diluted enzyme was added to the reaction mixture. At l-min intervals, 50-pl aliquots were added to 5 pl 0.5 M EDTA pH 8.2 held in tubes in an ice bath. Aliquots of the EDTA mixture (5 - 25 pl) were applied to the sheet of poly(ethy1ene imine)-cellulose along the line of zero-time samples. Spots were dried in a stream of cold air. The chromatogram was developed at 4 "C for 5 min in 0.3 M LiCl, 30 min in 1.0 M LiCl and in 1.6 M LiCl until the solvent front was at least 13 cm from the origin. The well-resolved spots containing [14C]nucleotides were located with an ultraviolet lamp. Spots were cut out and radioactivity counted in a Packard Tri-Carb model 3385 liquid scintillation spectrometer. Toluene scintillation fluid [3] was made fresh daily. The radioactivity was plotted against time and, from the slope of the best line drawn through the points, the initial reaction rate was calculated. Equilibrium Dialysis
Dialysis was carried out in chambers made from matching paired plexiglass plates with six shallow cylindrical depressions cut into each (6.5-mm diameter, 1.5-mm depth, about 50-pI capacity). A tiny hole cut at the top of each chamber allowed access to each side; 14 small stainless steel bolts were used to compress the plates together. Individual Spectra/Por 2 membranes cut from dialysis tubing were placed over each depression of one plate and the second plate was screwed into place over the first plate to create two chambers separated by membrane. The membranes were prepared by boiling in a 20 mM Na~C03,0.5mM EDTA solution followed by rinsing and boiling four or five times in deionized distilled water. The membrane was cut into small rectangles and stored in distilled water at 4°C. Before each experiment the membranes were soaked in buffer for a minimum of 15 min, then blotted dry and placed in position. Aliquots of 40 pl of a solution having a final composition of 0.02 M buffer, 0.15 M NaC1, 2% poly(ethy1ene glycol)-400 and 1 mg/ml of protein
were placed in each of the six dialysis chambers located on one side of the membrane. An equal volume of the same solution in which the enzyme was replaced by increasing concentrations (from 0.01 mM to 0.6mM) of a specified 14C-labeled nucleotide was placed in each of the six chambers on the opposite side of the membrane. Each concentration of nucleotide was run in duplicate. To prevent loss by evaporation the delivery holes of the chambers were sealed by tape. The apparatus was placed on a wrist-action shaker and allowed to come to equilibrium overnight (16 - 17 h, usually at 4°C). Under the conditions used in our experiments, no loss of enzymatic activity was observed even when samples were left in the shaker for 24 h at 25 "C. After equilibration, 25 pl of sample was taken from each chamber, delivered to 5 ml of Biofluor scintillation fluid and 14C was counted. Equilibrium data were analyzed by Scatchard plots [4]. In many experiments 10 pl of sample from several chambers were chromatographed on poly(ethy1ene imine)-cellulose sheets to see if nucleotides different from those expected after equilibration were present. RESULTS Stoichiometry and Linearity of the Reaction
Complete assays were made for samples at l-min intervals. For the forward reaction when 1 ml of reaction mixture composed of 50 mM glycylglycine, 75 mM Na', 5 mM [14C]GTP, 5 mM [I4C]AMP and 5 mM MgC12 at pH 8.0 was incubated at 30 "C with 1.45 pg enzyme at 7 min the following changes were found: GTP decreased 0.859 pmol (k 0.4 %) and GDP increased 0.816 pmol (k 0.3 %);AMP decreased 0.820 pmol (i0.8 %) and ADP increased 0.780 pmol (k 0.6 %). Or, overall changes in substituents were 1 : 1 within 4.8 % averaging all seven determinations. In the reverse direction (conditions as above except initial substituents were 5 mM [14C]GDP, 5 mM [14C]ADP and 0.54pg enzyme) changes found for sample at 7 min were ADP decreased 0.980 pmol (f 1.0 %) and AMP increased 1.01 pmol (f 0.2 %); GDP decreased 0.950 pmol ( f 0.5%) and GTP increased 0.990 pmol (+ 0.6%). Thus, in the reverse direction substituent changes were 1 : 1 within k 3.4 averaging all seven determinations. Time course assays of this and other reactions showed that the rate of reaction was linear until over 10 % of the substrates present were consumed. Requirement for Divalent Cations
Enzymatic activity has been detected only in the presence of specific divalent metal ions, namely Mg", Mn2+ and Ca2'- (effective only in the forward direc-
265
A. G. Tomasselli and L. H. Noda Table 1. Metal ions and specific activity Chromatographic assays were conducted after incubation with various metal chlorides ( 5 mM) in 0.05 M glycylglycine, 0.075 M Na’, pH 8.0, 30°C and 0.257 pg of enzyme. Forward reaction: GTP and [14C]AMP were 5.63 mM. Reverse reaction: GDP and [14C]ADP were 5.0 mM. Results are given in U/mg enzyme and in parentheses, as a percentage of the activity with MgZ+
I% 500 E
.
- 400 .’g 300 3
A
+
m
Specific activity in
Cations
600
m
E, 200
forward reaction
reverse reaction
U/mg (%I
c N
w
100
C
0
Mg2‘ Mn2’ CaZ BaZt, Cu2+,SrZ+, Fez+,Fe3+ +
124 (100.0) 100 (80.0) 84 (67.7) 0
(0)
397 (100.0) 143 (36.0) 0 (0) 0
(0)
0.2
0.4
0.6
0.8
“a+]
(M)
1.0
1.8
2.0
Fig. 2. Effect of salt concentration on enzymic activity at p H 8.0 in theforwnrd reaction (0) and the reverse reaction (0). Conditions of incubation were as in Fig. 1 except MgCIz was constant at 5 mM and Na’ was varied. The amount of enzyme was 0.257 and 0.128 pg in 0.5 ml for the forward and reverse reactions respectively
380 340
- 300
.
- 260 3
> + .-
5
220
m a,
180
&
140
5.
100
60 C
Fig. 1. Effect of magnesium ion concentration on enzymic activity in theforwardreaction ( 0 )and the reverse reaction (0). The reaction mixture of 0.5 ml contained 5 mM of each nucleotide, 75 mM Na+, 50 mM glycylglycine at pH 8.0 and 30”C, with 0.31 and 0.25 pg enzyme for forward and reverse reactions respectively and with [14C]AMP or [I4C]ADP
tion), while Ba2+, Cu2’-, Sr2‘-, Fez+ and Fe3+ were ineffective (Table 1). Under the conditions used, optimal enzymatic activity was obtained when the ratio of [Mg2+]:[GTP] is 0.6 and when [Mg”] is equal to ’/, ([GDP] + [ADP]) in the forward and reverse directions respectively (Fig. 1). With increasing concentrations of Mg2’ the activity increases to a maximum, and then levels off showing its relation to the ratio [Mg2+]: [nucleotides] rather than to the absolute amount of Mg2+. Effect of Sodium Chloride on Activity andpH Optimum Substitution of 0.1 M Na’ with 0.1 M K + at pH 8.0, 30°C and the usual concentrations used for
the other constituents gave 95% of the activity observed with Na‘.. Hence experiments on salt effects were conducted with NaC1, taking into account the total Na’ ion concentration. Fig. 2 shows that at pH 8.0 a concentration of 0.8 M Na+ gives maximum activity. In similar experiments at pH 11.5 (a pH optimum, see below) it was found that maximal enzymatic activities were at 0,075-0.10 M Na+ and that further increases of Na’ to 0.80 M resulted in slightly decreased activity. Thus the effect of varying pH at various salt concentrations was studied. Effect o f p H at Various Salt Concentrations In the forward direction at Na’ concentrations 0.075-0.80 M the pH optimum is at 11.5 with arginine . HCl as buffer (Fig. 3). In the reverse reaction at Na’ concentration 0.075 M, the pH optimum is at 11.5 (other concentrations of Na+ were not tested). In the forward direction at pH 11- 11.5 additions of NaCl above 0.075 M resulted in slightly decreased activity, but at lower pH the addition of salt above 0.075 M resulted in increased activity. For the reaction ITP + AMP IDP ADP at 0.075 M Na’ and with the same conditions as with the GTP reaction the pH optimum 9.5 instead of at pH 11.5.
+
Substrate Specificity Various combinations of nucleotides were tested for their ability to act as substrates of the purified enzyme. Table 2 summarizes part of the data obtained. Not listed are CMP, GMP, IMP, UMP and CAMP giving no detectable activity under the usual conditions of assay with [14C]GTP as the other substrate.
266
Mitochondria1 GTP-AMP Phosphotransferase r
i
I
I
I
I
I
I
I
I
I
- 500
AMP as the other substrate. Inosine can substitute for guanosine phosphates and a comparison at their best pH values, i.e. 9.5 and 11.5 respectively, are shown in Table 2. However the values obtained are not necessarily strictly comparable, since the optimal conditions of the reaction and details of the kinetics for inosine nucleotides were not thoroughly investigated. dAMP gave 40% of the AMP activity when incubated with GTP at pH 8.0. The enzyme preparation, even at the concentration of 3.1 pg/0.5 ml, was unable to catalyze the following reactions: [14C]ATP ['"ClAMP i2 2 ['4C]ADP and ['"CIGTP HzO + ['"CIGDP + Pi and so it can be considered free of ATPAMP phosphotransferase and GTPase. [14C]GDP either alone or in combination with CDP, IDP or UDP gave no reaction. Since the incubation of GTP and ['4C]ADP with enzyme yielded no ['"CIATP nucleoside diphosphokinase was not present.
+
PH
Fig. 3. Eflect o f p H at various Na' concentrations on enzymic activity in theforward (0,0.075 M Nu'; A, 0.40 M Nai ; x ,0.80 M Na' ) and the reverse reaction (0, 0.075 M Na'). Nucleotides and magnesium chloride were each 5 mM and reactions (0.5 ml) were carried out at 30°C with 0.257 or 0.170 pg enzyme for the forward and reverse reactions respectively. Buffers, 0.05 M, used were: pH 4-6 sodium acetate, pH 6.5 - 7.5 Mops, pH 8 - 9 glycylglycine,pH 9.5 11.O glycine and pH 11.5- 12.5 arginine . HCl/NaOH. [14C]AMP or [14C]ADPwas used as labeled substrate
+
Effect of Temperature Table 2. Substrate speci$city of GTP-AMP phosphotransferase All nucleotides were 5 mM with 5 mM MgClz in 0.05 M buffer, 0.075 M Na' with arginine at pH 11.5, glycine at pH 9.5 and glycylglycine at pH 8.0. All incubations were at 30°C with 0.1290.257 pg enzyme in 0.5 ml assay mixture (with the exception of [I4C]AMP UTP in which 1.55 pg of protein was used)
+
Substrates
PH
_ 14c-
labeled
Product
_
_
unlabeled
(14C~
Specific activity
Relative activity
labeled)
Apparent energies of activation at pH 8.0 and optimal magnesium ion substrate ratios, [Mg2+]/ [GTP] = 0.6 and [Mg2+]/([GDP] [ADPI) = '/2, calculated from Arrhenius plots are 68.2 kJ/mol for the forward reaction and 43.9 kJ/mol for the reverse reaction.
+
Equilibrium Constant U/mg
11.5
AMP ADP
GTP GDP
ADP AMP
436 575
1.o 1.3
9.5
AMP ADP
ITP IDP
ADP AMP
325 437
1.0 1.3
8.0
AMP AMP AMP ADP ADP
GTP ITP UTP GDP IDP
ADP ADP ADP AMP AMP
110 175 6 397 302
1.o 1.6 0.05 3.6 2.8
8.0"
AMP AMP GTP GTP
GTP dGTP AMP dAMP
ADP ADP ADP GDP
126 126 121 51
1.o 1.o 1.o 0.4
* MgC12 concentration was 3 mM instead of 5 mM.
Equilibrium was approached in both directions under the following conditions: 25"C, pH 8.5 in 0.05 M triethanolamine buffer, 0.075 M KCI, 1 mM EDTA, 20 mM Mg2+, 5 mM GTP and 5 mM AMP in the forward direction and 5 mM GDP, 5 mM ADP in the reverse direction. The amount of enzyme used was 20.6 pg added to 0.5 ml of reaction mixture. The apparent equilibrium constant under the above conditions was Kapp= [ADP] [GDP]/[GTP] [AMP] = 1.54. Michaelis Constants
The values in Table 3 were obtained using the chromatographic method of assay and avoiding the use of auxiliary enzymes. Binding Studies
Thus it may be concluded that the AMP (or ADP) binding site is specific for adenosine. CTP does not serve as substrate with AMP even when the amount of enzyme was raised from the usual 0.128-0.257 pg to 1.55 pg per 0.5 ml of reaction mixture. At similar high enzyme concentration UTP gives only about 5 % of the activity found for GTP. Under the conditions tested, dGTP served as well as GTP as substrate with
All the equilibrium binding data were analyzed by Scatchard plots. Straight lines were drawn through the data points by the method of least squares. Fig.4-6 show some of those plots and provide explanations about the method used to obtain them. Table 4 summarizes the results obtained for both the dissociation constants and binding stoichiometries of nucleotides and Mg2+-nucleotide complexes. Nucleo-
267
A. G. Tomasselli and L. H. Noda Table 3. Michaelis constants of various substrates for GTP-AMP phosphotransferase Michaelis constants were determined at 25 "C in 0.05 M triethanolamine . HCI, 0.075 M KCI, 1 mM EDTA, 5 mM MgClz and with saturating nucleotides at 5 mM. Initial rates of reaction were determined for at least five different concentrations of variable substrate (0.10-1 mM) Substrate
I 1
0.45
t
I
I
0.2
0.4
I
I
I
a8
1.0
K, mM
AMP GTP ADP GDP ri
0.139 0.143 0.074 0.036 I
1
I
I
0 0
1%
0.6 1 [El,
Fig. 5. Scatchard plot for equilibrium ligand binding. (A) [14C]GDP binding obtained under conditions given in Fig. 4
I
Fig. 4. Scatchard plots for equilibrium ligand binding. (0) Mg[14C]GTP binding was obtained using the following conditions: 0.02 M Hepes buffer, 0.15 M NaCI, 2% poly(ethy1ene glycoI)-400 and 1 mg/ml protein on one side of the membrane; on the other side of the membrane 0.02 M Hepes buffer, 0.15 M NaC1, 2 % poly(ethylene glycol)-400, increasing concentrations from 0.01 M to 0.6 mM labeled nucleotides and corresponding concentrations of MgClz equal to twice the concentration of nucleotide. (0)[14C]GTP binding was under the above conditions and omitting MgC12. Plots were made according to the equation
derived by modifying the Eadie-Scatchard equation (or derived directly from the equilibrium expression for KJ where [S]b, the concentration of bound substrate, = [ES], the concentration of occupied sites; [Sir is the concentration of the free substrate, [El, is the total enzyme concentration, K, = [E]f[S]r/[ES], is the intrinsic substrate dissociation constant of a site, [El, is the concentration of the free enzyme, n is the number of identical and independent ligand-binding sites per molecule of enzyme and on the plot is equal to the intercept on the x axis; n/Ks is the intercept on the y axis
tides that gave poor binding or did not bind at all within the limitations of enzyme and substrate concentrations are not listed. This was the case for AMP, GMP, CTP, and UTP for which equilibrium dialysis
I
I
I
I
[Slb/[El,
Fig. 6. Scatchard plots for equilibrium ligand binding. (A) (0)[14C]ATP binding, (0) ['4C]ADP binding obtained under conditions given in Fig. 4; (B) (A) [14C]AMPbinding obtained under conditions given in Fig. 4 and the additional presence of 5 mM GDP
Table 4. Dissociation constants and number of binding sites Conditions for equilibrium dialysis are specified in Fig.4. n = number of binding sites I4C-Labeled substrate
GTP MgGTP GDP MgGDP ATP ADP AMP AMP a
Unlabeled substrate
-
-
GDP" MgGDP"
Concentration of 5 mM.
n
KS
0.92 1.03 0.88 0.86 1.02 1.17 0.80 0.83
PM 4.4 4.5 2.2 3.0 33.4 33.4 20.9 23.1
Mitochondria1 GTP-AMP Phosphotransferase
268
shows a limitation in this type of study, namely that precise experimental data can not be obtained when the binding of a substrate is determined as the small difference between two large numbers, i.e. total substrate minus free substrate. Under the following conditions ['4C]AMP alone with enzyme was not bound within the limitations of experiment: at pH 8 and 4 "C or 25 "C in the presence and in the absence of Mg''., at pH 11.5 and 4 "C or 25 "C, with AMP concentrations ranging between 0.01 mM to 0.6 mM. But in the presence of MgGDP at pH 8 and 25 "C (not reported in Table 4) it was found that the enzyme bound AMP at 0.93 site with K, = 80 yM. Further data shown in Table 4 at pH 8.0 and 4 "C, indicate that AMP in the presence of 5 mM GDP binds to 0.80 site with a K, = 20.9 yM and in the presence of 5 mM MgGDP binds to 0.83 site with a K, = 23.1 pM. The binding of AMP is tighter at low temperature. The K, of AMP in the presence of 5 mM GTP could not be determined because even in the absence of added Mg2' and in the presence of 25 mM EDTA, [I4C]AMP was transformed to [14C]ADP as determined by thin-layer poly(ethy1ene imine) chromatography. ['4C]ADP binds to the enzyme with a K, = 33.4 yM and the number of sites occupied is 1.17; it is noteworthy that no partner substrate is required to bind [14C]ADP.At the end of the experiment the solutions from the dialysis cells were chromatographed and found free from nucleotides different from [14C]ADP indicating that no reaction took place in the process and that the binding observed was due to [14C]ADP. However when [14C]ADPwas run in the presence of 5 mM GTP, no [14C]ADP binding was observed and in the process a small amount of [14C]AMP was formed, about 5 -7% of the total [14C]ADPpresent. [I4C]ATP binds to 1.02 sites of the enzyme with K , = 33.4 yM. No partner substrate was needed for its binding; the presence of GTP or GDP prevents the binding of the [I4C]ATP to the enzyme; neither side reactions nor presence of unexpected nucleotides were encountered in these experiments. GTP binds to 0.92 site of the enzyme with K, = 4.4 pM and MgGTP occupies 1.03 sites with K, = 4.5 yM. The MgGTP binding was also tested at pH 11.5 and 4°C (not shown in Table 4) and the binding observed was 0.90 site with K, = 10.0 pM. GDPwas found to bind at0.88 sitewith%= 2.2yM and MgGDP gives 0.86 binding site with K, = 3.0 pM. UTP and CTP did not exhibit any binding as evidenced by the difference in radioactivity observed being too small to be significant because of the limitations of conditions of the experiments. Thus, the approximate 5 % of reaction observed in the kinetic studies when UTP and AMP were added to the enzyme might be due to a limitation of binding of the substrates, a limitation of reaction mechanism or by some trace contaminating enzyme.
GMP alone binds very poorly or not at all to the enzyme and the presence of ATP or ADP did not increase the binding of GMP. DISCUSSION GTP-AMP phosphotransferase from beef heart mitochondria was reported by Albrecht [5] as having a molecular weight of 52000, optimum activity at pH 8.5, specific requirement for divalent cations and specific for AMP as phosphate acceptor and unspecific for the phosphate donor since GTP, ITP and to a lesser extent UTP, CTP and ATP could fulfill this role. Results obtained in this work show that the enzyme, as purified according to the procedure described in the preceding paper [l], has an M , of 26000, pH optimum of 11.5 and 9.5 when guanosine and inosine phosphates are respectively substrates. The enzyme has a site specific for AMP or dAMP (ADP in the reverse direction) designated the AMP site. The second enzymatic site, designated the GTP site, transfers in the forward direction phosphate from GTP, dGTP or ITP. Since UTP gives only slight activity and CTP and ATP do not serve as phosphate donors in our enzymatic preparation, it appears likely that at least one other enzyme, capable of catalyzing the interaction with nucleotide phosphates, was present in the preparation reported by Albrecht. The Michaelis constants reported in Table 3 agree within about 4-fold of values reported by Albrecht [5] with the exception of K, for GDP which is 30-fold higher. Part of these differences may be due to dependence on coupled enzymes in assays of the earlier work. By kinetic studies of the activity of the enzyme some conclusions can be drawn about the characteristics that nucleotides must possess for enzymatic activity. The fact that the AMP site can accept only AMP, dAMP (or ADP) for reaction indicates that the 6-amino group of the purine base is essential. The importance of the ribose and 2'deoxyribose can be evaluated by observing that with dAMP the enzyme exhibits only 40% of the activity compared to AMP indicating that the 2'-OH group is preferred to the 2'-deoxy group, but is not essential for activity. GTP and dGTP when compared under the conditions reported in Table 2 are equally suitable indicating that the 2'-deoxy group works as well as the 2'-OH group. The further information that ITP is active with the enzyme and ATP is not, indicate that the 6-oxy group in the purine base is indispensable for activity. The -NH2 group in position 2 in guanine compared to a hydrogen in inosine, may be important in determining the pH optimum for activity since it was observed that optimal activity was at pH 13.5 for guanosine nucleotide and at pH 9.5 for inosine nucleotide. The lack of activity observed with CTP and the very poor activity offered by UTP then indicates that the nature
269
A. G. Tomasselli and L. H. Noda
of the base is a very important feature for GTP-AMP phosphotransferase activity. The presence of a divalent cation, namely Mg2+, Mn2', or Ca2',is indispensable for enzymatic activity. The reaction rate is determined by the ratio [Mg2+] : [nucleotide] rather than by the independent concentration of metal ion alone. This behavior has been reported for myokinase [6], adenylate kinase [7,8] and NTP-NDP kinase [9] from bovine liver mitochondria. The requirement of a divalent cation for activity is a property shared by all kinases in general. However, present studies show that the presence of a divalent cation is not required for the binding of the nucleotides of GTP-AMP phosphotransferase. Since the dissociation constants for free nucleotide GTP or GDP compared to Mg-nucleotide from enzyme are very similar, the simplest explanation is that Mg2+ is not directly coordinated to the protein in these ternary complexes and that the metal ion is involved in facilitating the phosphate transfer taking place. Similar metal-enzyme relationships have been reported for a number of enzymes: by Mildred Cohn and associates for porcine muscle adenylate kinase [ 101 and for rabbit muscle creatine kinase [ l l - 131 using nuclear magnetic resonance studies of the binding of nucleotide to the enzyme in the presence and absence of Mn2+, and by Kuby et al. [14] for rabbit muscle adenylate kinase by using a sedimentation gradient procedure in the presence and in the absence of Mg2+. In contrast, in rabbit muscle pyruvate kinase [15- 171 metal ion has been shown by Mildvan and associates to bind to the enzyme and that in the ternary complex it is located between protein and nucleotide. While ADP binds to the AMP site (reverse reaction), the impossibility of ADP binding to the GTP site precludes the possibility of GTP-AMP phosphotransferase having adenylate kinase activity. The fact that neither ['4C]ADP in the presence of 5 mM GTP nor [I4C]ATP in the presence of 5 mM GTP or GDP bind to the enzyme can be explained by steric hindrance by the bulk of extra phosphate groups. When GTP or GDP is present on the GTP site of the enzyme, the obligatory choice of the partner becomes respectively AMP or ADP. UTP and CTP were found within the limitations of these binding studies not to bind to the enzyme so the slight activity observed with UTP AMP at relatively high concentrations of the enzyme might be explained by inherent trace activity with these substrates or a very small amount of contaminating enzyme in the preparation. From kinetic and binding studies a mechanism of reaction is proposed for the reaction catalyzed by GTP-AMP phosphotransferase. In the forward direction the binding of AMP alone to one site is too low to be detected with the method used in the present
+
studies; however the binding of AMP to one site is greatly enhanced by the presence of GDP (or GTP) on the other site and can be easily detected in that way (Table4). So the mechanism in the forward direction may be assumed as nearly ordered. In the reverse direction there is no requirement of substrate bound to one site to detect the presence of the partner substrate on the other site and hence the mechanism may be assumed random according to the Cleland nomenclature [18]. For many enzymatic reactions the optimum of activity at fixed pH is dependent upon the ionic strength of the medium [19 - 231 ; however, at present a complete understanding of the process has not been provided. As shown by data presented in Fig. 2 and 3, GTP-AMP phosphotrqnsferase activity seems to be related to increasing Na' concentration in such a way that slight inhibition occurs at pH values ranging between 11 and 11.5 and stimulation occurs at pH values lower than 10.5. This behavior is consistent with the assumption that increasing Na+ concentration is accompanied by a variation of the ionization constants of the catalytic groups of the active site or of the substrates. The unusually high pH optimum of 11.5 observed in vitro may not apply to the situation in vivo. But it seems possible, in view of the chemiosmotic mechanism of oxidative phosphorylation [24,25] and the mosaic nature of the matrix membrane [26], that a non-uniform distribution of charges on the matrix side of the inner mitochondria1 membrane may result. Zones of high electrical potential may result in a very heterogeneous distribution of protons thus creating compartments of relatively high and relatively low pH (without any influence on the average pH), i.e. zones favorable and unfavorable for GTP-AMP phosphotransferase activity. This research has been supported by Grant GM-23123 from the National Institutes of Health, United States Public Health Service, for which grateful acknowledgment is made. We wish to express appreciation for the careful and capable technical assistance of Brenda Shea during the course of this work.
REFERENCES 1. Tomasselli, A. G., Schirmer, R. H. & Noda, L. H. (1979) Eur. J . Biochem. 93,251 - 262. 2. Kolthoff, I. M. & Sandell, E. B. (1955) Textbook of' Quantitative Inorganic Analysis, 3rd edn, pp. 362 - 363, Macmillan Co., New York. 3. Nossal, N. G. & Singer, M. F. (1968) J . Biol. Chem. 243, 91 3 -922. 4. Scatchard, G. (1949) Ann. N.Y. Acad. Sci. 51, 660-672. 5. Albrecht, G . J. (1970) Biochemistry, 9, 2462-2470. 6. Noda, L. H. (1958) J . Biol. Chem. 232, 237-250. 7. Markland, F. S. & Wadkins, C. L. (1966) J . B i d . Chem. 241, 4136-4145. 8. Glaze, R. P. & Wadkins, C. L. (1967) J . Biol. Chem. 242, 2139- 21 50.
270
A. G. Tomasselli and L. H. Noda: Mitochondria] GTP-AMP Phosphotransferase
9. Goffeau, A., Pedersen, P. L. & Lehninger, A. L. (1967) J. Biol. Chem. 242, 1845 - 1853. 10. Price, N. C., Reed, G. H. & Cohn, M. (1973) Biochemistry, 12, 3322-3327. 11. Cohn, M. & Leigh, J. S. (1962) Nature (Lond.) 193, 10371040. 12. O’Sullivan, W. J. & Cohn, M. (1966) J. Biol. Chem. 241, 3104- 31 15. 13. Reed, G. H. &Cohn, M. (1972)J. Biol.Chem. 247,3073-3081. 14. Kuby, S. A , , Mahowald, T. A. & Noltmann, E. A. (1962) Biochemistry, I , 748 - 762. 15. Mildvan, A. S. & Cohn, M. (1966) J . Biol. Chem. 241, 11781193. 16. Mildvan, A. S., Leigh, J. S. & Cohn, M. (1967) Biochemistry, 6. 1805-1818.
17. Mildvan, A. S. &Cohn, M. (1965) J. Biol. Chem. 240,238-246. 18. Cleland, W. W. (1963) Biochem. Biophys. Acta, 67, 104-137. 19. Kalnitsky, G., Hummel, J. P., Resnick, H., Carter, J. R., Bernett, L. B. & Dierky, C. (1959) Ann. N.Y. Acad. Sci. 81, 542- 566. 20. Irie, M. (1965) J. Biochem. (Tokyo) 57, 355-362. 21. Barton, N. W., Lipovac, V. & Rosenberg, A. (1975) J. Biol. Chem. 250, 8462 - 8466. 22. Maurel, P. & Douzou, P. (1976) J. Mol. Biol. 102, 253-264. 23. Douzou, P. & Maurel, P. (1977) Trends Biochem. Sci. 2,14- 17. 24. Mitchell, P. (1966) Biol. Rev. 41, 445-502. 25. Mitchell, P. (1961) Nature (Lond.) 191, 144-148. 26. Singer, S. J. (1972) Ann. N . Y . Acad. Sci. 195, 16-23.
A. G. Tomasselli and L. H. Noda, Department of Biochemistry, Dartmouth Medical School, Hanover, New Hampshire, U.S.A. 03755
,
