(Received for publication, January 27, 1997, and in revised form, April 14, 1997)
From the Departments of Biological Chemistry and
§ Biophysics and Biophysical Chemistry, The Johns
Hopkins University, School of Medicine,
Baltimore, Maryland 21205-2185
The chemical mechanism by which the
F1 moiety of ATP synthase hydrolyzes and synthesizes
ATP remains unknown. For this reason, we have carried out studies with
orthovanadate (Vi), a phosphate analog which has the
potential of "locking" an ATPase, in its transition state by
forming a MgADP·Vi complex, and also the potential, in a
photochemical reaction resulting in peptide bond cleavage, of
identifying an amino acid very near the -phosphate of ATP. Upon
incubating purified rat liver F1 with MgADP and
Vi for 2 h to promote formation of a
MgADP·Vi-F1 complex, the ATPase activity of
the enzyme was markedly inhibited in a reversible manner. When the
resultant complex was formed in the presence of ultraviolet light
inhibition could not be reversed, and SDS-polyacrylamide gel
electrophoresis revealed, in addition to the five known subunit bands
characteristic of F1 (i.e.
,
,
,
,
and
), two new electrophoretic species of 17 and 34 kDa. Western
blot and N-terminal sequencing analyses identified both bands as
arising from the
subunit with the site of peptide bond cleavage
occurring at alanine 158, a conserved residue within
F1-ATPases and the third residue within the nucleotide
binding consensus GX4GK(T/S) (P-loop).
Quantification of the amount of ADP bound within the
MgADP·Vi-F1 complex revealed about 1.0 mol/mol F1, while quantification of the peptide cleavage products revealed that no more than one
subunit had been cleaved. Consistent with the cleavage reaction involving oxidation of the methyl
group of alanine was the finding that [3H] from
NaB[3H]4 incorporates into
MgADP·Vi-F1 complex following treatment with
ultraviolet light. These novel findings provide information about the
transition state involved in the hydrolysis of ATP by a single
subunit within F1-ATPases and implicate alanine 158 as
residing very near the
-phosphate of ATP during catalysis. When
considered with earlier studies on myosin and adenylate kinase, these
studies also implicate a special role for the third residue within the
GX4GK(T/S) sequence of many other
nucleotide-binding proteins.
Despite our extensive knowledge about the structure and function
of the F1 moiety of ATP synthases (1-5), sufficient
information is not available to write a chemical mechanism by which ATP
is hydrolyzed and synthesized. In contrast, myosin-ATPase has been successfully studied using orthovanadate,
Vi,1 a phosphate analog which
in the presence of MgADP forms a transition state
MgADP·Vi-myosin inhibitory complex (6-8). Irradiation of this complex with uv light results in the modification of the single
serine within the nucleotide binding consensus GESGAGKT followed by peptide bond cleavage at this site (Fig. 1A) (8, 9). These studies strongly implicated this serine as contacting directly the -phosphate of ATP in the transition state, and were recently confirmed by x-ray structural analysis (10). The reaction pathway of F1-ATPase is believed to be quite similar to
that of myosin-ATPase (11). In addition, within the catalytic sites of
both enzymes resides the nucleotide binding consensus
GX4GKT (P-loop) (12), which in the
-subunit
of F1 (GGAGVGKT), contains alanine in place of
the internal Vi-sensitive serine characteristic of the
myosin consensus (GESGAGKT). As this serine in myosin is known to contact the
-phosphate of ATP in the transition state (10),
it seemed reasonable to assume that in F1 a nearby serine residing outside the consensus region or the terminal threonine within
the consensus region may serve this role in the transition state.
Alternatively, the third position within the consensus region of
F1, although containing an alanine, may play an important role in the transition state of F1 as does the serine in
the same position in myosin. Studies described below were carried out
both to define optimal conditions for trapping F1-ATPase in
a MgADP·Vi-F1 inhibitory transition state
complex, and to establish in this state the identity of an amino acid
residue near the
-phosphate of ATP. Both goals were accomplished and
provide novel insights into the chemical mechanism of ATP
synthases.
Materials
Rats (Harlan Sprague-Dawley, white males) were obtained from
Charles River Breeding Laboratories. ATP, ADP, MgCl2, MOPS,
CAPS, sodium orthovanadate, phosphoenolpyruvate, pyruvate kinase, and lactic dehydrogenase were obtained from Sigma. SDS, acrylamide, and
bisacrylamide were from Bio-Rad. Ammonium sulfate and potassium phosphate were from J. T. Baker Chemical Co. Tuberculin syringes and
Sephadex G-50 used in nucleotide binding assays were from Becton-Dickinson Co. and Pharmacia Biotech Inc., respectively. PVDF
membranes were obtained from Millipore and Western blot reagents from
Amersham. A polyclonal antibody against the rat F1
-subunit was raised in rabbits using the synthetic peptide
KIGLFGGAGVGKCT. [3H]ADP and
NaB[3H]4 were from NEN Life Science Products.
All other reagents were of the highest purity commercially
available.
Methods
Purification of Rat Liver F1-ATPaseThe enzyme
was purified by the procedure of Catterall and Pedersen (13) with the
modification described by Pedersen et al. (14). The purified
enzyme, in 250 mM KPi and 5.0 mM
EDTA, was divided into 100-µl aliquots, lyophilized to dryness, and
stored at 20 °C. Prior to use, aliquots (150-250 µg) of
lyophilized F1 were dissolved at 25 °C in 100 µl of
water and precipitated twice with 3 M ammonium sulfate, 5 mM EDTA, redissolving between precipitations in 200 mM K2SO4, 10 mM
Tris-Cl, pH 7.5, or 50 mM MOPS, pH 8.0.
To
minimize the presence of polymeric species the following protocol was
followed: Na3VO4 powder was dissolved in water
and the pH adjusted with HCl to pH 10 (orange color). The solution was
boiled for 2 min at which time the solution became clear. The pH was
readjusted to pH 10 and the previous boiling repeated 2 times. After
the optical density was determined at 265 nm, the Vi
concentration was calculated using the molar extinction coefficient of
2925 M1 cm
1. The stock
solutions used in this study were 155 mM and were covered
with aluminum foil and stored until use at
80 °C.
F1 (50 µg) was priorly incubated in a 100- or 200-µl system containing 50 mM MOPS, pH 8.5, 10% glycerol (v/v) and, where indicated, also Vi, Vi + ADP, Vi + ADP + MgCl2, Vi + ATP + MgCl2, or MgCl2, all at concentrations indicated in legends. Prior incubations were carried out at 25 °C for the indicated times. In experiments where photoactivation of vanadate was induced with uv light (320 nm), the incubation mixture in an open Eppendorf tube was placed under a 100 watt, long wavelength mercury spot lamp (BLAK-RAY (Model B 100A, 115 V, 2.5 amperes; San Gabriel, CA) at a distance of 7.8 cm. The time in the presence of the light source varied as indicated in the figure legends.
Assay for ATPase ActivityThe spectrophotometric procedure was used in which ADP formed was coupled to the pyruvate kinase and lactic dehydrogenase reactions (13). The reaction mixture contained the following in a volume of 1 ml at pH 7.5 and 25 °C: 0.2 mM ATP, 65 mM Tris-Cl, 4.8 mM MgCl2, 2.5 mM KPi, 0.40 mM NADH, 0.60 mM phosphoenolpyruvic acid, 5 mM KCN, 1 unit of lactic dehydrogenase, 1 unit of pyruvate kinase, and 1.5 µg of F1.
Assay for ADP BindingBinding assays were carried out at
25 °C by incubating F1, or fractions derived therefrom,
for 20 min in a final volume of 100 µl, containing concentrations of
F1, [3H]ADP, MgCl2,
Vi, and buffer as indicated in the legend to Fig. 4. The
entire reaction mixture was loaded onto a Sephadex G-50 "fine"
column (1-cm3 tuberculin syringe with a filter at the
bottom), which had been pre-equilibrated with 50 mM
Tris-Cl, pH 7.6, and priorly centrifuged for 1.5 min at 2,500 rpm in an
IEC model HN-SII clinical centrifuge. Centrifugation of the reaction
mixture was carried out for 1.5 min at 2,500 rpm to separate nucleotide
bound to F1 from free nucleotide (15, 16).
SDS-PAGE
This was carried out in a Bio-Rad Mini-Protean dual slab cell in 15% acrylamide according to the method of Laemmli (17), or in cylindrical glass tubes in 5% acrylamide according to the method of Weber and Osborn (18). Where indicated, densitometric analysis of the Coomassie-stained bands was carried out using a Fujifilm Bas-1500 PhosphorImager and MacBas (V 2.31) software.
Western Blot AnalysisAfter conducting SDS-PAGE, the
proteins on the gel were transferred electrophoretically onto a PVDF
membrane (1 h at 100 volts and 0.2 amp at 4 °C in 10 mM
CAPS, 10% methanol transfer buffer, pH 11). The product was then
blocked for 1 h with 2% bovine serum albumin plus 5% nonfat dry
milk in PBS-T (80 mM Na2HPO4, 20 mM NaH2PO4, 100 mM
NaCl, 0.1% Tween 20, pH 7.5), incubated for 1 h at 23 °C with
a rat liver F1- subunit polyclonal antibody (see "Experimental Procedures"), and then further incubated for 1 h at 23 °C with the secondary antibody (horseradish
peroxidase-conjugated anti-mouse IgG). The immunoreactive bands were
detected by the enhanced chemiluminescene (ECL) system of Amersham Life
Sciences.
F1 -subunit
peptide fragments were transferred from SDS-PAGE gels onto PVDF
membranes by electroblotting. Transfer conditions were for 1 h in
10 mM CAPS buffer, 10% methanol, pH 11, at 4 °C in the
case of the 17-kDa F1-
subunit fragment, and 2 h in
the same buffer at 25 °C in the case of the 34-kDa fragment. Only under the latter conditions could an N-terminal sequence be obtained with the 34-kDa fragment. The peptides were then excised and subjected to N-terminal sequencing (19) using an Applied Biosystems 475A Protein
Sequencing System (20).
Protein was determined by the method of Lowry et al. (21) after first precipitating with 5% trichloroacetic acid.
To promote formation
of an inhibitory MgADP·Vi-F1 transition state
complex, several precautions were taken. First, the Vi stock solution was priorly treated exactly as described under "Experimental Procedures" and maintained at pH 10 in the dark at
80 °C until use to prevent formation of polymeric species. Second,
the zwitterionic buffer MOPS was used in all experiments as it both
stabilizes F1 and prevents or minimizes inhibition by
MgCl2 and the product MgADP (or ADP). Third, a pH of 8.5 was used in all experiments, as this pH is near optimal for the
hydrolysis of ATP catalyzed by rat liver F1. Finally,
equimolar amounts of MgCl2, ADP, and Vi were
used to promote formation of MgADP·Vi at the active site
of F1.
Fig. 2A summarizes results of an experiment where
F1 was priorly incubated in the presence of 200 µM each of MgCl2, ADP, and Vi for
the indicated times and then assayed for ATPase activity as described
under "Experimental Procedures." Here it is clear that under these
conditions, which are optimal for forming a
MgADP·Vi-F1 transition state complex,
F1-ATPase activity is markedly inhibited. Half-maximal
inhibition is reached in about 45 min, and maximal inhibition (~80%)
is reached in about 2 h. In control experiments with
F1 alone, F1 + Vi, F1 + ADP, and F1 + MgADP, inhibition at 2 h is,
respectively, <1, 10, 20, and 20% under conditions described in the
figure legend. In an experiment not presented, F1 prior incubated with MgCl2 alone, was not inhibited after 2 h. Finally, Fig. 2B shows that when F1 is prior
incubated with 200 µM each of MgCl2, ADP, and
Vi, but in the presence of light (320 nm) and atmospheric
oxygen, an essentially identical inhibition profile is observed.
Inhibition of F1-ATPase Activity by MgCl2 + ADP + Vi Is Reversible in the Absence of UV Light but Not in Its Presence
If formation of a
MgADP·Vi-F1 transition state complex is
responsible for the inhibition observed in the presence of
MgCl2 + ADP + Vi, this state should be
reversible. Results presented in Fig. 3A show that this is
the case. Thus, following maximal inhibition of F1-ATPase
activity in the presence of 200 µM each of
MgCl2, ADP, and VI (Fig. 3A, left
panel), an aliquot was removed and subjected to two dilution/wash
cycles in the presence of 50 mM MOPS, 10% glycerol, pH
8.5. This resulted after two such cycles in the restoration of the
original activity to a level near 90% (Fig. 3A, right
panel). Significantly, reversal of ATPase activity could not be
achieved (Fig. 3B, right panel) following inhibition of
F1-ATPase activity under identical conditions but in the
presence of uv light (Fig. 3B, left panel) and with longer
incubation times. The reason for this becomes clear in the description
of other experiments described below.
Vi Induces Inhibition of F1-ATPase Activity under Turnover Conditions
It is expected that if a MgADP·Vi-F1 transition state is formed in the presence of MgCl2, ADP, and Vi as implicated from the above studies, it would be formed also under turnover conditions, i.e. when ATP, the substrate for ATP hydrolysis is present. For this reason, F1 was incubated exactly as described above but with ATP replacing ADP in the prior incubation mixture. Thus, the final prior incubation mixture contained 200 µM each of MgCl2, ATP, and Vi. After 1 h under these turnover conditions, aliquots were removed and assayed for ATPase activity. Results presented in Fig. 4 show that F1-ATPase activity is inhibited about 50% in the absence of uv light and about 70% in its presence. Although after 1 h the degree of inhibition is not as great as that achieved when prior incubation is carried out with ADP rather than ATP in the prior incubation mixture, this is to be expected. Thus, under the latter conditions ATP must first undergo hydrolysis before ADP is available for formation of the MgADP·Vi-F1 complex and MgATP competes with ADP for binding to F1-ATPase active sites.
Polypeptide Chain Cleavage Occurs within the MgADP·Vi-F1 Complex in the Presence of UV LightAs indicated earlier, vanadate is photoreactive and has
been shown in the case of the MgADP·Vi-myosin complex to
modify an active site serine residue within contact distance of
Vi and, in the presence of light and atmospheric oxygen, to
induce peptide bond cleavage at this site (8, 9) (Fig.
1A). For this reason F1 was
priorly incubated in the presence of 200 µM each of
MgCl2, ADP, and Vi in the absence and presence
of uv light exactly as described for Figs. 2 and
3 and then subjected to SDS-PAGE. When the resultant
SDS-PAGE profiles (Fig. 5, A and B) of the two
incubation mixtures (absence and presence of light) are compared, it is
clear that only in the latter case has polypeptide bond cleavage
occurred. Thus, in addition to Coomassie-stained bands corresponding to the known F1 subunits ,
,
,
, and
, bands
distinct from these subunits with apparent molecular masses of 17 and
34 kDa appear (Fig. 5B, lane 6). Polypeptide bond cleavage
within the F1 molecule is highly specific, and is not
observed in the absence of uv light under any condition tested (Fig.
5A), and is observed in the presence of uv light with
MgCl2 + ADP + Vi (Fig. 5B, lane 6).
Fig. 5C shows that the appearance of the 17- and 34-kDa
peptide fragments increases with time as expected and levels off after
about 2 h (lane 9). These findings are consistent with
polypeptide bond cleavage within the nucleotide-binding consensus
region (GGAGVGKT) of the 51.5-kDa
-subunit, as this would give rise
to two fragments with molecular masses near the experimentally
determined values of the 17- and 34-kDa bands (Fig. 5C).
Polypeptide Chain Cleavage within the MgADP·Vi-F1 Complex Induced by UV Light Does Not Alter the Oligomeric State of F1 and Occurs Only within the
Results of native PAGE experiments presented in
Fig. 6A, lanes 1-5, show that, under conditions which
result in cleavage of a polypeptide chain within the
MgADP·Vi-F1 complex in the presence of uv
light and atmospheric oxygen, the oligomeric state of the F1 molecule remains intact. In fact, the F1
molecule remains intact in uv light under all conditions tested
(i.e. alone or with Vi, MgCl2 + ADP,
ADP + Vi, or MgCl2 + ADP + Vi; Fig.
6A, lanes 1-5, respectively). Specifically,
as it applies to the uv light and Vi-dependent
cleavage of a polypeptide chain within the
MgADP·Vi-F1 complex (Fig.
5B, lane 6), these results are consistent
with a very localized reaction which is otherwise without deleterious effect on the remaining part of the F1 molecule.
Identification of the or "catalytic" subunit within
F1 as the source of the 17- and 34-kDa bands was derived
from two separate experiments. In the first, SDS-PAGE gels of the uv
light-treated MgADPVi·F1 complex, after
transfer to PVDF membranes, were probed with a polyclonal, antibody
raised against the synthetic peptide KIGLFGGAGVGKCT, containing the
GX4GKT consensus region of the rat liver
subunit. As shown in Fig. 6B, lane 6, only
the intact
subunit and the 17-kDa band cross-react with the
antibody. This is the expected result if polypeptide bond cleavage
occurs within the nucleotide binding consensus region of the
subunit, as the epitope reactive with the antibody would be largely
retained at the C terminus of the 17-kDa fragment, but missing from the
34-kDa fragment. In the second experiment, N-terminal sequence analysis (Fig. 6C) which identified the 7-amino acid stretch APKAGTA
confirmed the
subunit (Fig. 1B) as the origin of the
17-kDa fragment. (The first six amino acids at the N terminus of the
rat liver
subunit predicted from c-DNA cloning (Fig. 1B)
are not present in the isolated protein.) N-terminal sequence analysis
of the 34-kDa fragment also proved possible by carrying out the
transfer from SDS-PAGE to PVDF membranes for 2 rather than 1 h,
and at 25 °C rather than 4 °C (see "Experimental
Procedures"). This modification in the transfer procedure was done to
promote removal of an oxalyl group (HO-CO-CO-) predicted to be at the N
terminus of the 34-kDa fragment following the
Vi-dependent cleavage reaction (Fig.
1A). Significantly, the N-terminal sequence obtained,
GVGKTVLIMELINN (Fig. 6D), not only confirmed the 34-kDa
fragment as being derived from the
subunit, but placed the site of
cleavage at alanine 158 within the GGAGVGKT consensus
region.
To determine the
extent of involvement of the 3 subunits of F1 in the
formation of the MgADP·Vi-F1 transition state
complex, and in the formation of the 17- and 34-kDa cleavage products, both ADP-binding studies and densitometric analysis of
Coomassie-stained bands were carried out. Fig. 6E shows that
under the conditions used in this study, rat liver F1 binds
only about 1 mol of ADP/mol of F1 and that Vi
and MgCl2 have little or no affect on this stoichiometric ratio. When MgCl2, ADP, and Vi are added
together, each at a concentration of 200 µM to favor
formation of the MgADP·Vi-F1 complex (Fig. 2A), the stoichiometry of ADP binding remains constant.
Thus, formation of the transition state complex appears to be
restricted predominantly to the involvement of 1
subunit. This
correlates well with the estimated number of
subunits involved in
formation of the 17- and 34-kDa cleavage products when the transition
state complex is formed in the presence of uv light. Here, a loss of 33% of the total F1
subunit staining intensity results
which is fully recovered by the sum of the staining intensities of the 17- and 34-kDa products (Fig. 6F).
Studies described above provide rather compelling evidence
that, in the presence of atmospheric oxygen, uv light-induced cleavage of a single subunit within the
MgADP·Vi-F1 complex occurs at alanine 158. If
the
-methyl group of this alanine residue is oxidized, it is
expected to proceed through a series of reactions resulting first in
the formation of serine, then an aldehyde (Fig. 7A), and finally other intermediates before
peptide bond cleavage finally occurs (see Fig. 1A and Refs.
22 and 27). Therefore, it should be possible, via reduction of the
aldehyde, to demonstrate a uv light-dependent incorporation
of tritium [3H] from 3H-labeled sodium
borohydride into F1-ATPase following the enzyme's inactivation by MgADP·Vi. Results presented in Fig.
7B show that such an incorporation does occur in very
significant amounts over control levels obtained in the absence of
MgADP·Vi. Incorporation reaches a maximal level in about
7 min, and then declines as expected because the aldehyde is a
transient intermediate in the overall oxidation process.
Implications for the Chemical Mechanism of ATP Hydrolysis by F1
Studies reported here strongly implicate alanine
158 in the third position of the GGAGVGKT consensus of the
rat liver F1 -subunit as residing very near the
-phosphate group of ATP in the transition state. Significantly,
alanine in this position is conserved in all F1-ATPases
(23), and in Escherichia coli, F1 mutations in this position to valine or proline result, respectively, in a >90%
loss or a 2-fold activation of catalytic activity (24). Moreover, the
comparable "third position" residue, serine 181, in
Dictyostelium myosin subfragment 1, has been shown to be
within contact distance (2.6 Å) of the Vi oxygen atoms in
the x-ray structure of the MgADP·Vi-myosin S1 complex
(10). With these facts, and the extensive data presented in this paper
in mind, a tentative pathway for ATP hydrolysis at the active site of
F1 is proposed. Here, formation of the transition state
(Fig. 8A, center panel) from the pre-ATP
hydrolysis state (Fig. 8A, left panel) is considered to
align the C
atom of alanine in close proximity to the
-phosphorus atom of ATP. In the proposed trigonal bipyramidal transition state complex, the planar
-phosphorus atom, a water molecule, and a potential catalytic base, "B," are considered to be
sufficiently close to facilitate ATP hydrolysis. As only one of the
three
subunits forms a transition state complex in the presence of
MgADP·Vi (Fig. 6, E and F), and
F1 is believed to function by involving sequential
participation of the
subunits in catalysis via
,
-subunit
interactions (5), it is not unreasonable to suggest also that these
interactions may help stabilize the transition state.
The mechanistic role of the third position serine in the
GX4GK(T/S) sequence in myosin (8, 9) involves a
much greater capacity to form hydrogen bonds to the oxygen of the
-phosphate of ATP than does the alanine in the same position in
F1-ATPase. Therefore, the third position serine in myosin
may play a different role in the mechanism of this enzyme than does the
third position alanine in F1-ATPases. Nevertheless, in both
cases, this third position residue may also be critical in determining
the polarity, size, and/or rate of formation of the binding pocket in
which the
-phosphate group of ATP is contained in the transition
state. Significantly, differences in these parameters among different ATP-dependent enzymes may help "set" the catalytic rate
at a value most compatible with an enzyme's physiological role, while
determining in part substrate specificity (see also Refs. 24 and 25). Along these lines, it is interesting to note that Vi in the
presence of uv light and oxygen also cleaves adenylate kinase at the
third position (proline 17) within the nucleotide binding consensus GGPGSGKGT (26). Thus, in support of the view proposed here, three different enzymes, myosin, F1-ATPase, and adenylate
kinase are all cleaved at the same third position despite the fact that the amino acid occupying this position is very different in all cases
(Fig. 8B), but conserved within its specific enzyme
class.
Finally, it should be noted that results presented here, which have
focused on alanine 158 of rat liver F1, do not preclude other amino acids near the -phosphate of ATP in the transition state. Significantly, in a recent intriguing paper Senior and colleagues (28) have summarized the possible roles of three catalytic
site residues in the E. coli F1-ATPase. One of
these, lysine 155 (lysine 162 in rat liver F1) is
considered to be involved in the major functional interaction with the
-phosphate of MgATP in the substrate bound or "ground state,"
but to undergo conformational repositioning during catalysis.
Therefore, when taken together with the novel findings from studies
reported here, it will be of considerable interest to visualize the
precise location and orientation of lysine 162 and alanine 158 in the
transition state when the x-ray structure of the
MgADP·Vi-F1 complex is elucidated.
We are grateful to Dr. Albert Mildvan with whom we had many discussions about this work, Joanne Hullihen for technical assistance, and Jackie Seidl for processing the manuscript for publication.