Mitochondria1 and Bacterial ATPases
1 Boyer, P. D. (1989) FASEB J. 3, 2164-2178
2 Boyer, P. D. (1993) Biochim. Biophys. Acta 1140,
215-250
3 Boyer, P. D., Cross, R. L. and Momsen, W. (1973)
Proc. Natl. Acad. Sci. U.S.A. 70,2837-2839
4 Kayalar, C., Rosing, J. and Boyer, P. D. (1977)
J. Biol. Chem. 252,2486-2491
5 Boyer, P. D. and Kohlbrenner, W. E. (1981) in
Energy Coupling in Photosynthesis (Selman, B.
and Selman-Reiner, S., eds.), pp. 231-240,
Elsevier/N. Holland, New York
6 Penefsky, H. S. and Cross, R. L. (1991) Adv.
Enzymol. 64, 173-214
7 Gogol, E. P., Johnston, E., Aggeler, R. and Capaldi,
R. A. (1990) Proc. Natl. Acad. Sci. U.S.A. 87,
9585-9589
8 Lightowlers, R. N., Howitt, S. M., Hatch, L.,
Gibson, F. and Cox, G. (1988) Biochim. Biophys.
Acta 933,241-248
9 Cain, B. D. and Simoni, R. D. (1989) J. Biol.
Chem. 264, 3292-3300
10 Vik, S. B. and Antonio, B. J. (1994) J. Biol. Chem.
269, 30364-30369
11 Abrahams, J. P., Leslie, A. G., Lutter, R. and
Walker, J. E. (1994) Nature (London) 370,621-628
Ila Duncan, T. M., Bulygin, V., Zhou,’Y., Hutcheon,
M. L. and Cross, R. L. (1995) Proc. Natl. Acad.
Sci. U.S.A., in the press
12 Zoller, M. and Smith, M. (1987) Methods
Enzymol. 154, 329-350
13 Lee, R. S., Pagan, J., Wilke-Mounts, S. and Senior,
A. E. (1991) Biochemistry 30, 6842-6847
14 Gibson, F., Downie, J. A., Cox, G. B. and Radik, J.
(1978) J. Bacteriol. 134, 728-736
15 Fraga, D., Hermolin, J., Oldenburg, M., Miller,
M. J. and Fillingame, R. H. (1994) J. Biol. Chem.
269, 7532-7537
16 Wise, J. G. (1990) J. Biol. Chem. 265, 10403-10409
17 Vandeyar, M. A., Weiner, M. P., Hutton, C. J. and
Batt, C. A. (1988) Gene 65, 129-133
18 Chang, A. and Cohen, S. (1978) J. Bacteriol. 134,
1141-1156
19 Downie, J., Langman, L., Cox, G., Yanofsky, C. and
Gibson, F. (1980) J, Bacteriol. 143, 8-17
20 Ysern, X., Amzel, L. M. and Pedersen, P. L. (1988)
J. Bioenerg. Biomembr. 20,423-450
21 Landt, O., Grunert, H.-P. and Hahn, U. (1990)
Gene 96, 125-128
22 Gibson, F., Cox, G. B., Downie, J. A. and Radik, J.
(1977) Biochem. J. 162, 665-670
23 Senior, A. E., Fayle, D. R. H., Downie, J. A.,
Gibson, F. and Cox, G. B. (1979) Biochem. J. 180,
111-118
24 Peterson, G. L. (1977) Anal. Biochem. 83, 346-356
25 Pullman, M. E., Penefsky, H. S., Datta, A. and
Racker, E. (1960) J. Biol. Chem. 235, 3322-3329
26 Ziegler, M., Xiao, R. and Penefsky, H. S. (1994)
J. Biol. Chem. 269,4233-4239
27 Aggeler, R., Haughton, M. A. and Capaldi, R. A.
(1995) J. Biol. Chem. 270,9185-9191
28 Ohta, S., Yohda, M., Ishizuka, M. et al. (1988)
Biochim. Biophys. Acta 933, 141-155
29 Means, G. E. and Feeney, R. E. (1971) in
Chemical Modification of Proteins, pp. 155-157,
Holden-Day, Inc., San Francisco
Received 21 July 1995
Nucleotide-binding sites in F,-ATPase: different pockets for different types of
nucleotide analogues
1. A. Berden and A. F. Hartog
E. C. Slater Institute, Biocentrum, University of Amsterdam, Plantage Muidergracht 12, I0 I8 n/ Amsterdam,
The Netherlands
Introduction
T h e publication of the protein structure of the
mitochondria1 FI-ATPase at 2.8 A resolution [ 13
has provided a very useful framework for the
interpretation of data obtained from ligand-binding and kinetic studies. O n the role of each of
the six nucleotide-binding sites of the enzyme
[2] in the catalytic mechanism, however, the
structural data on one specific conformation do
Abbreviations used: FSBA, 5’-p-fluorosulphonylbenzoyladenosine; SBA, covalently bound FSBA;
TMS, Tris/HCl/MgClJsucrose; TEG, I’ris/HCl/
EDTNglycerol; TDAB, tetradecyltrimethylammonium
bromide.
not provide direct information. Ligand-binding
studies remain appropriate to define the roles of
the various nucleotide-binding sites, both catalytic and non-catalytic, as well as their mutual
interactions. T h e emphasis that is put by the
authors of [l] on the structural possibility of
rotation of the .I/? moieties around the y-subunit
has, however, confused the discussion on the
mechanism of catalysis by isolated F1: speculations about the properties of the native system
are put forward without first considering all available data on the isolated enzyme itself. In this
contribution we will first remind you of some
findings that prove the idea that in isolated FI
only two /?-sites are performing multi-site
I995
74 I
Biochemical Society Transactions
742
catalysis while the other one contains a nonexchangeable nucleotide [3-71, excluding a rotational model for catalysis by this form of the
enzyme. This fact may well be an intrinsic property of the ATP synthase or the consequence of
the disconnection of FI from the membraneembedded part of the native enzyme, and we will
discuss some experimental approaches to find
the definitive answer.
We further report studies with the analogue
5'-p-fluorosulphonylbenzoyladenosine (FSBA) as
a control for our model. This analogue has been
applied in several systems [8-lo], and the data
obtained with isolated FI originating largely from
the group of Allison [ 11-13], show that FSBA
binds to regulatory nucleotide-binding sites only
and induces removal of adenine nucleotides from
these sites [12]. These sites are in essence
located at the a-subunits [ l ] , although a
covalently binding probe may modify an amino
acid of the neighbouring jj-subunit, as has been
reported for most of the adenine nucleotide analogues, including FSBA. According to our model
of FI one non-catalytic tightly bound adenine
nucleotide is located at a /j-site, and the prediction was that this nucleotide should not be
removed from its site on binding of FSBA to the
a-sites. A difference between the effects of FSBA
on our enzyme and those on the Allison enzyme
may be expected, since we have previously shown
[14] that the preparations of F, used by the
groups of Allison and Gautheron are significantly
different from the preparations used by us and
others. These differences imply kinetic properties (hysteretic inhibition) and nucleotide
content.
Results and discussion
Definition of the nucleotide binding sites
From the studies with g-azido-AT(D)P [3,6],
2-azido-AD [7,15,16] and the dissociation-reconstitution experiments by ourselves [4] and by
Miwa et al. [5] a model for the various nucleotide-binding sites is derived that differs from the
commonly accepted one.
Our isolated enzyme, like the enzyme as isolated by Cross and Penefsky, contains three
tightly bound nucleotides (see [14]), of which
one is at a catalytic site (site 2), while the two
others do not exchange during turnover and
during incubation with various analogues. These
are located at sites 1 and 4 in Scheme 1, one
being largely ATP, the other ADP. T h e site at
which ADP, 2 - azido - ADP and 3' -0- (3- [N-(4-
Volume 23
azido-2- nitrophenyl) - amino] propionyl} -2-azidoADP bind quite tightly in the presence of Mg2+
[16-191, causing a partial inhibition of the
ATPase activity, is site 5 . Site 3 is the second
catalytic site and site 6 is again a regulatory site.
When this latter site is occupied with 8-nitrenoAT(D)P the biphasic kinetics of ATP hydrolysis
become monophasic, with disappearance of the
low K,, value [6].
/I-sites: 1 (tight ATP) 2 (tight catalytic) 3 (loose catalytic)
a-sites: 4 (tight ADP) 5 (regulatory)
6 (regulatory)
Scheme 1
Whether, in the intact system, site 1 is
directly involved in multi-site catalysis is being
investigated using two approaches: one is to try
to exchange in submitochondrial particles
specifically the nucleotide at site 1 with 2-azidoATP, the other to replace in isolated F1 the ATP
at site 1 with 2-azido-ATP, using the dissociation-reconstitution procedure described previously [7] and to reconstitute FI-depleted
submitochondrial particles with this F1. In this
latter case the localization of covalently bound
2-nitreno-AT(D)P can be more easily verified.
E@ct of FSBA binding on enzyme activity and
nucleotide content
T h e rate of inhibition of the ATPase activity of
F1 upon addition of FSBA increases with increasing pH. At pH 8.0 FSBA is unstable and
degrades in time (see also [ZO]), but at pH 7.5
the compound is not significantly hydrolysed
within 4 h, so this pH was chosen as the
standard condition under which labelling of
mainly
was expected [ 121. T h e rate also
increases with increasing concentrations of FSBA
up till about 2 mM. As a standard condition we
used 0.8 mM FSBA, similar to the concentration
used by Bullough and Allison [12]. Neither the
presence of Mg2+ nor the presence of a fourth
bound nucleotide had a significant influence on
the time course of the inhibition. This inhibition
reaches about 35% after 1 h and after 2 h at 20°C
about 50-60% inhibition is obtained. T h e
ITPase activity, however, is inhibited by 8 0 4 5 %
after 2 h.
T h e effect of binding of [I4C]FSBA on the
activity of F1 is shown in Figure 1 for preparations of F1 in TMS (Tris/HCl/MgCl2/sucrose) or
T E G (Tris/HCl/EDTfdglycerol). It can be seen
that the binding of the first molecule of FSBA
causes about 50% inhibition of the ATPase
activity, while the binding of additional molecules
Mitochondria1 and Bacterial ATPases
Table I
Analysis of the binding of [“CIFSBA to Fi containing four bound nucleotides
The incubation buffer (20°C) was TMS, the concentration of F, 2 m g h l and [FSBA] 0.8 mM. The incubation was stopped by adding cold
FSBA, followed by three column centrifugation steps. In the sample marked with an asterisk the incubation was stopped with 50 mM
dithiothreitol. followed by column centrifugation. Bound adenine nucleotides were determined with the bioluminescence assay [ I 61.
Covalently bound FSBA (SBA) was determined from TDAB gels.
ATPA i
(mol/mol)
ADPAi
(mol/mol)
AXPA I
(mol/mol)
Inhibition of
ATPase (%)
SBMi
Label
Incubation time (min)
(mol/mol)
(alp)
0
30
60
90
I20
I20*
240
0.64
0.50
0.7 I
0.70
0.79
0.78
0.78
3.38
3.05
2.75
2.63
2.43
2.86
2.39
4.02
3.55
3.46
3.33
3.22
3.64
3.17
0
35
51
57
66
66
77
0
0.59
0.85
1.21
I .45
1.51
I .97
-
induces relatively less inhibition. Nearly all F1
molecules have bound one FSBA before a substantial level of binding of the second FSBA
molecule takes place, as can be concluded from
the inhibition of the ITPase activity: after binding of one FSBNFI the ITPase activity is inhibited by more than 80%. These data are in line
with the results of Allison’s group [21].
This group has reported that FSBA induces
removal of ADP from the non-catalytic sites [ 121.
We find, however, that the amount of bound adenine nucleotides does not change upon incubation of FI [F1(3) or Fl(4)] with FSBA. With Fi(4)
(see Table 1) a slight decrease in bound adenine
nucleotides is measured, but this decrease also
occurs in the absence of FSBA, when at different
time intervals the enzyme is passed over three
centrifugation columns [ 16,17,19]. Since in Fi(4)
at least two or-sites contain adenine nucleotides,
minimally one a-subunit can bind both ADP and
FSBA at the same time. Therefore, one subunit
contains pockets for two adenosine moieties.
Efect of pyrophosphate on the inhibition of
F,-ATPase by FSBA
In a recent paper Jault et al. [22] have shown that
the formerly reported [23] increased inhibition
by FSBA in the presence of 1 mM pyrophosphate
is not due to faster binding of FSBA, but to
abolition of the activity-stimulating effect of
pyrophosphate. Under our conditions the stimulatory effect of 1 mM pyrophosphate on the
ATPase activity is very low. This concentration of
pyrophosphate also has no effect on the binding
of FSBA, nor on the nucleotide content of F1.At
3.54
3.35
2.55
2.43
2.46
I .88
Figure I
Inhibition of Fi by covalently bound [“CISBA
F, (2 mg/ml) was incubated for 3 h with 0 8 mM [ 4C]FSBA in
either TMS or TEG buffer Bound radioactivity was determined at
various incubation times after a chase with unlabelled FSBA,
followed by two column centrifugation steps and tetradecyltrimethylammonium bromide (TDAB) gel electrophoresis The labelled
bands of a- and /hubunits were excised and the radioactivity
determined m, ATPase activity, measured with an ATP regenerITPase adivity, measured in experiments in
ating system
I00 mM pyrophosphate
which the medium was TEG buffer
+,
“0
015
I
11.5
2
2.5
3
Bound radioactivity (mol of [14C] SBNmol of F,)
high concentrations, however, the binding is
strongly accelerated: after 30 min in the presence
of 100 mM pyrophosphate and 0.8 mM FSBA, 1.5
mol of FSBNmol of F1 is bound instead of 0.6.
Determination of bound ligand reveals that in the
presence of 100 mM pyrophosphate 3 mol of
FSBNmol of FI are bound at 90% inhibition.
T h e ITPase activity is largely inhibited with one
FSBNFI (see Figure 1).
Pyrophosphate has been shown to bind to
both catalytic and non-catalytic sites [6,24,25],
and Jault et al. [22] report the binding of 3 mol
I995
743
Biochemical Society Transactions
744
of pyrophosphate per mol of F1when nucleotidedepleted Fl is incubated with 1 mM pyrophosphate. Since 1 mM pyrophosphate has no effect
on our enzyme, we may assume that the three
sites that bind pyrophosphate at this concentration in the nucleotide-free preparation of Jault et
al. contain adenine nucleotides in our preparation (sites 1, 2 and 4). Upon incubation with 100
mM pyrophosphate, however, the nucleotide content decreases to 2moVmol of F1 (at sites 1 and
4). Removal of ADP from the catalytic site is
clear from the kinetics of the ATPase reaction:
the lag phase after the addition of ATP, due to
the presence of ADP at the catalytic site [26],
has disappeared. The permanent stimulation of
the ATPase activity must be due to binding of
pyrophosphate at site 5. It should be kept in
mind that binding of ADP at this site induces
inhibition [2]. It is likely that during the incubation in 100 mM pyrophosphate sites 3 and 6 are
also occupied, but these sites lose the ligand
after column centrifugation andor dilution in the
assay mix. Since the stimulation of the binding of
FSBA by pyrophosphate largely disappears after a
column centrifugation step, the stimulation of
FSBA binding is largely due to binding of pyrophosphate to site 6.
A high concentration of pyrophosphate does
not remove the nucleotides from sites 1 and 4,
and at least three FSBA molecules can bind in
the presence of pyrophosphate, so under these
conditions FSBA and ADP can also be bound to
the same a-subunit, assuming again that FSBA
binds to non-catalytic sites only.
Localization of bound SBA
TDAB polyacrylamide gel electrophoresis of F1,
incubated with radioactive FSBA, showed that
both a- and P-subunits were labelled. In the
presence of glycerol the ratio was always about
1:1. This result is principally not much different
from the results of Allison’s group [ 121, although
this group concluded that only binding to P-subunits was specific. Since, however, only one SBAl
F1 induces nearly full inhibition of the ITPase
activity, the modification of the a-subunits has to be as inhibitory as that of the P-subunits. This conclusion is further strengthened by
the finding that in F1 samples incubated with
FSBA in TMS buffer (no glycerol), the a-subunits contained much more label than did the
P-subunits. With F,(4) the alp ratio varied from
3.5:l in the initial phase of the incubation to
about 2:l after 4 h (Table 1). With Fl(3) this
Volume 23
ratio was slightly less.
For determination of the sites of modification we wanted to be sure that no hydrolysis of
bound ligand occurred, so we avoided acid precipitation since much label disappeared after an
acid precipitation step. The FSBA-treated
enzyme was digested with trypsin for a very long
time (24 h), making more cuts than the normal
ones (behind Lys and Arg), in order to further
break down the very large a-fragments obtained
otherwise. The disadvantage of this procedure is
that not all fragments can be predicted.
The separation of the various trypsin-treated
preparations by reverse-phase HPLC (we left out
the anion-exchange chromatography step [6])
resulted in a very reproducible pattern of label
distribution (Figure 2). The obtained fractions
might have contained several peptides, and
therefore we rechromatographed the fractions in
which the label probably originated from proteinbound SBA (22, 24, 26 and 30) with a more
extended gradient. The original fraction 26
appeared to contain two different labelled fragments, named A and B. Amino acid sequencing
showed that indeed several peptides were present
in each fraction, making the analysis more difficult. Since the amount of applied label was
known, the peptides with the wrong intensity
could be disregarded and in most cases the
labelled amino acid could be detected as the
absence in the sequence of a tyrosine. The
labelled peptides appeared to be the peptides
containing
(in fraction 30),
(in
fraction 26A), and P - T Y(in
~ ~fraction
~
26B). No
labelled P-His4” was found. In agreement with
the original distribution of the label over a- and
P-subunits, the two labelled a-fragments contained twice as much label as the P-fragment.
The minor amount of label in fractions 22 and 24
could not be identified as yet and is possibly not
protein-bound.
Since the analysis of the sequence data is
not completely unambiguous, all fractions will be
analysed with electro-spray mass spectrometry as
well. This is not yet completed. The preliminary
conclusion, then, has to be that in our F1 FSBA
binds at positions different from those in the
preparation of Allison’s laboratory, although in
both cases only non-catalytic sites are modified.
The different position may explain why we see
no effect on bound nucleotides, while Allison
does. It seems worthwhile to note that the
a-TyrZMin the preparation of Allison’s group is
modified with 5’-p-fluorosulphonylbenzoyletheno-
Mitochondria1 and Bacterial ATPases
Figure 2
HPLC of ["CISBA-labelled peptides
F, (2 mghl) containing 3 mol of adenine nucleotides per mol of F , , was incubated with 0.8
mM [14C]FSBAfor various times in TMS buffer. The incubations were stopped by addition of
50 mM dithiothreitol. followed by a column centrifugation step. After trypsin treatment ( I ms/
20 mg of F,) for 24 h at 30°C the samples were put on a C4-Vydac RP-HPLC column and
eluted with a gradient of 0-70% B (90% acetonitrile, 0. I % trifluoroacetic acid, 9.9% water).
The flow was I . I ml/min and each fraction contained I , I ml. The absorbance at 2 I 5 and 258
nm was followed (not shown in the Figure) and the radioactivity in each fraction was
determined by liquid-scintillation counting. Shown are the results from the sample that was
incubated for 2 h. Total covalently bound label in the sample (45000 d.p.m., I .73 mol of
SWmol of F,) was determined separately.
16
745
60
I
50
c
.
I
40
..................
......
- ....
sp
. . .
30
ee
20
0
8
4
...............
r(
x
10
0
-10
0
30
20
10
40
Time (min)
I
radioactivity
adenosine (FSBEA) [27]. It is a pity that these
authors did not investigate whether FSBEA also
removes bound ADP from their enzyme.
Emct of ADP and diadenosine tetraphosphate on
the distribution of label
We have not observed a clear competition of
added ADP with FSBA (cf. [12]). But since the
ratio between a- and P-labelling decreased
during the binding process, the second molecule
apparently binds differently from the first one.
We therefore investigated whether the presence
of ADP or A p A (1 mM) had an effect on the alp
distribution of bound label. T h e result of one
experiment is shown in Figure 3. T h e rate of
inhibition by [I4C]FSBAwas slightly decreased in
the presence of ADP or A p A and so was the
labelling of P-subunits, while the rate of labelling
of the a-subunits remained the same. T h e relation between total labelling and inhibition was
not changed. At each level of inhibition, therefore, more a-subunits now were labelled and
1
fewer P-subunits. This proves again that modification of both a- and P-subunits causes inhibition. Since the a-subunits are modified at Tyr3O0
and Tyr244,and the P-subunits at Tyr368,we propose that each a-subunit contains two pockets
where adenosine moieties (and FSBA) can bind:
one below the P-loop region and one above it.
Since FSBA does not contain a di- or tri-phosphate group, the P-loop itself is not involved in
this binding. For ADP or ATP the P-loop is part
of the binding site, and therefore only one
nucleotide with a di- or tri-phosphate group can
bind at each subunit. But principally the position
of the nucleoside moiety may be variable, being
bound above the P-loop or below. T h e latter
position is preferred by AD(T)P itself and the
analogues in the anticonfiguration, such as
2-azido-AD (T)P. This latter analogue modifies
the p-Tyr368,just as FSBA does at this position.
FSBA in the upper pocket modifies a-Tyr244and
a-Tyr3"". When we suppose that the P-subunits
also have these two pockets, many labelling data
I995
Biochemical Society Transactions
Figure 3
Effect of ADP and &,A on the labelling of F, with
[I4C]FSBA
746
F, (2 mg/ml), containing 3 rnol of adenine nucleotides per mol of
F,, was incubated for various times with 0.6 mM [I4C]FSBA in
TMS buffer. ADP and Ap4A,when present, had a concentration of
I mM. The incubations were stopped with 50 rnM dithiothreitol,
followed by column centrifugation. Samples were put on a TDAB
gel and [“C] in the bands of the 2- and /hubunits was determined after excision and solubilization. The ATPase activity was
measured before the column centrifugation step with an ATP
, Total [“C]SBA/Fl
(all samples):
regenerating system.
0,a-bound [“c]SBA/F, in the absence of ADP or Ap4A;
0,P-bound [I4c]SBA/Fi in the absence of ADP or Ap4A A ,
a-bound [I4C]SBA/F, in the presence of ADP or ApA; m,
8-bound [‘4c]SBA/Fl in the presence of ADP or Ap4A.
+
0
0
0.5
I
1
1.5
2
2.5
1
3
Bound radioactivity (mol of [I4C]SBA/molof FI)
can be explained: 2-azido-ATP modifies /3-Tyr345
(lower pocket), but 8-azido-ATP and 4-chloro7-nitrobenzofurazan modify P-Tyr3l1at the upper
site of the P-loop.
Since ADP binds quite strongly at site 5, we
propose that the first FSBA also binds at the
a-subunit that contains site 5. T h e kinetic
effects of the first FSBA are indeed very similar
to those of ADP at site 5 (results not shown).
This binding is, in the absence of glycerol,
mainly at the upper pocket, and when ADP is
bound as well this preference is even stronger,
without a change in the level of inhibition. Binding of both ADP and FSBA (or of two FSBA
molecules) at site 5 induces a stronger inhibition
of enzyme activity than that of one ligand alone.
T h e subunit with site 6 also binds FSBA, but
here the upper and lower pocket become about
equally occupied. Whether the binding at site 6
causes much inhibition is not completely sure,
especially since we know that g-nitreno-AT(D)P
at this subunit does not cause inhibition of
activity [6]. It seems likely that in our experiments the a-subunit with site 4 remains free of
FSBA. This subunit apparently binds FSBA only
in the type of preparation isolated by the groups
of Allison and Gautheron (see [14]). T h e latter
group has indeed reported binding of up to 6 mol
Volume 23
of FSBNmol of F1 [28], but assumed wrongly
that both catalytic and non-catalytic sites were
involved.
Vogel and Cross [29] have shown that Ap+4
inhibits F1 by tight binding of one mol/mol of FI.
Their data can be interpreted nicely on the
assumption that this analogue can occupy both
pockets on the a-subunit at opposite sides of the
P-loop. From their data it is clear that a noncatalytic site is occupied, and according to our
interpretation this has to be site 5. Binding at a
catalytic site is not clear at all and we propose
that the high level of inhibition (80%) is due to
the fact that both pockets are occupied. In our
experiments A p A induces the same inhibition as
ADP (maximally 40-45%) and only modification
of
by FSBA is inhibited, just as with
ADP. In our enzyme, therefore, only one pocket
is occupied by ApA.
This work was supported in part by grants from the
Netherlands Organization for the Advancement of
Scientific Research (NWO) under the auspices of the
Netherlands Foundation for Chemical Research
(SON).
1 Abrahams, J. P., Leslie, A. G. W., Lutter, R. and
Walker, J. E. (1994) Nature (London) 370, 621-628
2 Wagenvoord, R. J., Kemp, A. and Slater, E. C.
(1980) Biochim. Biophys. Acta 593,204-21 1
3 Van Dongen, M. B. M. and Berden, J. A. (1986)
Biochim. Biophys. Acta 850, 121-130
4 Nieboer, P., Hartog, A. F. and Berden, J. A. (1987)
Biochim. Biophys. Acta 894, 277-283
5 Miwa, K., Ohtsubo, M., Denda, K., Hisabori, T.,
Date, T. and Yoshida, M. (1989) J. Biochem.
(Tokyo) 106,679-683
6 Edel, C. M., Hartog, A. F. and Berden, J. A. (1993)
Biochim. Biophys. Acta 1142, 327-335
7 Hartog, A. F., Edel, C. M., Lubbers, F. B. and
Berden, J. A. (1992) Biochim. Biophys. Acta 1100,
267-277
8 Pal, P. K., Wechter, W. J. and Colman, R. F. (1975)
J. Biol. Chem. 250, 8140-8147
9 R. F. Colman (1990) in The Enzymes, Vol. XIX
(Sigman, D. S. and Boyer, P. D., eds.), pp.
283-321, Academic Press, New York
10 Bitar, K. G. (1982) Biochem. Biophys. Res.
Commun. 109, 30-35
11 Esch, F. S. and Allison, W. S. (1979) J. Biol. Chem.
254, 10740- 10746
12 Bullough, D. A. and Allison, W. S. (1986) J. Biol.
Chem. 261,5722-5730
13 Bullough, D. A., Brown, E. L., Saario, J. D. and
Allison, W. S. (1988) J. Biol. Chem. 263,
14053- 14060
Mitochondria1 and Bacterial ATPases
14 Berden, J. A., Hartog, A. F. and Edel, C. M. (1991)
Biochim. Biophys. Acta 1057, 151-156
15 Van Dongen, M.B. M., De Geus, J. P., Korver, T.,
Hartog, A. F. and Berden, J. A. (1986) Biochim.
Biophys. Acta 850,359-368
16 Edel, C. M.,Hartog, A. F. and Berden, J. A. (1992)
Biochim. Biophys. Acta 1101,329-338
17 Kironde, F. A. S. and Cross, R. L. (1987)J. Biol.
Chem. 262,3488-3495
18 Kironde, F. A. S. and Cross, R. L. (1986)J. Biol.
Chem. 261, 12544-12549
19 Edel, C. M.,Hartog, A. F. and Berden, J. A. (1995)
Biochim. Biophys. Acta 1229,103-1 14
20 Colman, R. F. (1983) Annu. Rev. Biochem. 52,
67-91
21 Bullough, D. A. and Allison, W. S. (1986)J. Biol.
Chem. 261, 14171-14177
22 Jault, J. M.,Paik, S. R., Grodsky, N. B. and Allison,
W. S. (1994)Biochemistry 33, 14979-14985
23 Jault, J. M.and Allison, W. S. (1993)J. Biol. Chem.
268, 1558-1566
24 Kalashnikova, T. Y.,Milgrom, Y. and Murataliev,
M. (1988)Eur. J. Biochem. 177,213-218
25 Michel, L., Garin, J., Vincon, M., Gagnon, J. and
Vignais, P. (1995) Biochim. Biophys. Acta 1228,
67-72
26 Drobinskaya, I. Y.,Kozlov, I. A., Murataliev, M. B.
and Vulfson, E. N. (1985)FEBS Lett. 182,419-424
27 Verburg, J. G. and Allison, W. S. (1990) J. Biol.
Chem. 265,8065-8074
28 Di Pietro, A., Godinot, C., Martin, J.-C. and
Gautheron, D. C. (1979) Biochemistry 18,
1738- 1745
29 Vogel, P. D. and Cross, R. L. (1991)J. Biol. Chem.
266,6101-6105
Received 19 June 1995
~~
Catalytic mechanism of Gcherichia coli FI -ATPase
A. E. Senior, J. Weber and M. K. Al-Shawi
Department of Biochemistry, Box 607, University of Rochester Medical Center, Rochester, N Y
The six nucleotide-bindingsites of
Gcherichia coli FI
Native E. coli F1,as purified in MgATP-containing buffer, contains around three bound adenine
nucleotides (moVmol) [l]. It may be depleted
completely of nucleotides [2,3]; the nucleotidefree enzyme binds six [3H]Mg-adenosine
5’-[by -imido] triphosphate (p [N HI ppA) (mol/
mol) [2]. T h r e e of the six sites are classed as
non-catalytic (NC) because nucleotides bound at
these sites exchange with medium MgATP or
MgADP very slowly. Other distinguishing
features of the NC sites are that they bind PP,,
show strong preference for adenine over guanine/
inosine nucleotide and bind free nucleotide
poorly [ 1,431. All current evidence indicates that
the N C sites play no direct role in MgATP
hydrolysis or oxidative phosphorylation [ 11.
T h e other three, potential catalytic (C),
sites show strong negative binding co-operativity
with Mg-nucleotide. For example, the three Kd
values for [3H]Mgp[NH]ppA range from 0.3 pM
(highest affinity) to 20 pM (lowest affinity) ([Z];
see also [6]). In the absence of Mg, the free
nucleotide-binding affinities are weaker [7,8],
such that radioactive nucleotide-binding assays
cannot be effectively used to characterize all
Abbreviations used: NC, non-catalytic; C, catalytic;
p[NH]ppA, adenosine 5’-[/3y-imido]triphosphate.
14642,U.S.A.
three C sites. Also, MgATP-binding parameters
cannot be characterized by radioactivity assays at
concentrations where significant hydrolysis
occurs.
Tryptophan fluorescence spectroscopy has
recently been introduced by our laboratory as a
tool for specific characterization of the properties
and functions of C and NC sites. In particular, it
allows determination of MgATP-binding parameters directly. Mutational analysis of the C sites,
begun in 1979 [9], has been a major approach to
elucidating the catalytic mechanism and the roles
of specific residues in catalysis. There is excellent agreement with regard to residues identified
as critical to catalysis by mutagenesis and the Xray structure of the C site [lo]. For several C
site residues much functional information has
already been documented. In this paper we discuss our recent work using tryptophan fluorescence and mutational analysis.
Mg co-ordination is central to
asymmetry of C sites, catalysis and the
binding change mechanism
In 1992 we demonstrated that in wild-type C
sites residue P-Tyr-33 1 makes direct contact
with the adenine ring of bound substrate and
thereby accelerates catalysis [ 111. T h e X-ray
structure confirms the juxtaposition of the Tyr
and adenine rings [lo]. We introduced T r p at
I995
747