![]()
|
|
||||||||
J. Biol. Chem., Vol. 279, Issue 46, 47415-47418, November 12, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

¶
**



From the
Department of Molecular Biosciences, University of Kansas, Lawrence, Kansas 66045 and the Departments of ||Chemical Physics and
Biological Chemistry, Weizmann Institute of Science, Rehovot 71600, Israel
Received for publication, June 9, 2004 , and in revised form, September 14, 2004.
| ABSTRACT |
|---|
|
|
|---|
-rotation assay, that CaATP is capable of sustaining rotational motion in a highly active hybrid photosynthetic F1-ATPase consisting of
and
subunits from Rhodospirillum rubrum and
subunit from spinach chloroplasts (
R3
R3
C). The rotation was found to be similar to that induced by MgATP in Escherichia coli F1-ATPase molecules. Our results suggest a possible long range pathway that enables the bound CaATP to induce full rotational motion of
but might block transmission of this rotational motion into proton translocation by the F0 part of the ATP synthase. | INTRODUCTION |
|---|
|
|
|---|
3
3

. Fluorescent actin filaments attached to the
subunit of engineered surface-immobilized
3
3
subcomplexes of the enzyme, either from a thermophyilic bacterium, TF11 (3), or from Escherichia coli, EF1 (4, 5), were shown to rotate unidirectionally while hydrolyzing MgATP. This rotation was not only dependent on ATP concentration but was shown to have almost 100% efficiency of chemical to mechanical energy conversion. At very low ATP concentration the stepwise motion of the enzyme was exposed (6). Small beads attached to
in place of the actin filament allowed Yasuda et al. (7) to follow the motion at the limit of small load and watch rotation at a rate as high as 130 revolutions per second. Most spectacularly, forced clockwise rotation was shown very recently to lead to ATP synthesis, thus confirming the essential role of
subunit rotation in both directions of the reaction (8). Some single-molecule work also addressed the operation of the F0 part of the F0F1-ATP synthase. It was shown that the F0-c subunit oligomer, which consists of 1014 identical c subunits, is capable of rotating together with
(9, 10).
While much new information has been collected regarding rotational catalysis in the ATP synthase, there are still many open questions, mainly in relation to the coupling between enzymatic activity and rotational motion. The photosynthetic version of the ATP synthase is an attractive system to probe some of these questions, since it presents several distinct functional properties. These include the tight regulation of ATP hydrolysis, which is especially important in photosynthetic cells, where it prevents the depletion of essential ATP pools in the dark (1113). Indeed, plant chloroplasts have a unique regulatory system, termed thiol modulation, which leads to reduction of a disulfide bond formed between Cys-199 and Cys-205, located in the so-called switch region of the CF1-
subunit that does not appear in any other type of CF1-
(14). The oxidation of the disulfide bond inhibits the ATPase activity of the enzyme, while its reduction stimulates it. Thiol modulation was studied indirectly on the single-molecule level by Bald et al. (15) who transferred the switch region from the CF1-
subunit to a TF1-
subunit. A second unique feature of ATP synthases of both chloroplasts (16) and chromatophores (17) is their high sensitivity to inhibition by excess free Mg2+ ions, which results in a drastic reduction of their MgATPase activities at a Mg/ATP ratio exceeding 0.5. Another very interesting phenomenon, found in both chloroplasts (18) and chromatophores (17), is the decoupling of Ca2+-induced ATP hydrolysis from proton translocation, although their membranes do not become leaky to protons. Both systems also show no Ca2+-dependent ATP synthesis.
Understanding the mechanism by which calcium decouples ATP hydrolysis and proton translocation can enhance our knowledge of the interaction between various parts of the F0F1 molecule in relation to its functionality. We provide here the first step toward this goal by demonstrating that the assembled hybrid photosynthetic F1-
R3
R3
C complex, recently described in our laboratory (19), shows rotational motion in the presence of CaATP. This result indicates that the decoupling effect of calcium, which does not interrupt the connection between ATP hydrolysis and
C rotation, might be mediated by a long range effect on the membrane bound F0 subunits.
| EXPERIMENTAL PROCEDURES |
|---|
|
|
|---|
Reconstitution and Biotinylation of the Hybrid F1 EnzymeRecombinant
R containing an N-terminal His6 tag was obtained by amplification of the
R gene from pTZ
(20) and ligation into the pET14b expression vector. This His-tagged
R, as well as the recombinant
R (21), and
C (22) were expressed as inclusion bodies in E. coli BL21(DE3)/pLysS cells grown to steady state at 37 °C. These inclusion bodies were individually solubilized and the expressed proteins were diluted with the refolding buffer (20), assembled into the hybrid
R3
R3
C complex, and purified by high performance liquid chromatography as described by Du et al. (19). The purity of the isolated hybrid complex was evaluated by SDS-PAGE (Fig. 1A).
|
R3
R3
C complex was achieved by using the reversible disulfide bond formed between cysteine residues 199 and 205 of the
C subunit, as described by Hisabori et al. (23). Purified
R3
R3
C was oxidized with 200 µM CuCl2, for 1 h at 22 °C, to promote disulfide bond formation before the addition of 1 mM N-ethylmaleimide to block all other exposed cysteines. The
C disulfide of the N-ethylmaleimide-treated complex was then reduced by incubation with 20 mM dithiothreitol for 30 min at 22 °C. After passage through two successive Sephadex G-50 centrifuge columns to remove the excess reductant, the enzyme was labeled with a 3-fold excess of biotin-PEAC5-maleimide for 30 min at 22 °C. The labeling was stopped by the addition of 200 µM cysteine followed by passage through a centrifuge column. The specificity of labeling was confirmed by Western blots (Fig. 1B).
EF1-
3
3
ComplexesThese complexes were prepared according to Noji et al. (5) and Panke et al. (10) and kindly supplied by Drs. W. Junge and S. Engelbrecht.
Polimerization of the Biotinylated G-actinStocks of biotinylated G-actin were maintained at a concentration of 1 mg/ml in 2 mM Hepes-NaOH (pH 7.2), 0.2 mM CaCl2, and 0.2 mM ATP at 80 °C. To obtain fluorescently labeled F-actin, the protein was diluted to 18 µM into the polymerization buffer at a final concentration of 20 mM Hepes-NaOH (pH 7.2), 100 mM KCl, 10 mM MgCl2, 2 mM ATP, and 36 µM phalloidintetramethylrhodamine B isothiocyanate conjugate and polymerized overnight at 4 °C.
Rotational AssaysConstruction of the flow cells, adherence of the His6-tagged and biotinylated
R3
R3
C, and infusion of the actin filaments were carried out essentially as described previously for TF1- and EF1-
3
3
complexes (3, 4, 10). Briefly, flow cells with a chamber volume of 2030 µl were infused with the following series of buffers (2 x 25 µl per step, 5-min incubation): 1) Buffer A (10 mM Hepes-NaOH (pH 7.2), 25 mM KCl, 5 mM MgCl2) with 2.4 µM HRP-Ni-NTA; 2) Buffer B (Buffer A plus 10 mg/ml bovine serum albumin) with a 3 nM concentration of the biotinylated
R3
R3
C complex; 3) Buffer B with 50 nM streptavidin; and 4) Buffer B with 50 nM fluorescently labeled actin filaments. Each step was followed with a wash of 100 µl of Buffer B, except for step 4, after which we performed a wash with 100 µl of Buffer B and 100 µl of Buffer A, both lacking MgCl2. CaATP-dependent rotation was initiated by the infusion of 50 mM Tricine-NaOH (pH 8.0), 4 mM CaCl2, 4 mM ATP, 216 µg/ml glucose oxidase, 36 µg/ml catalase, 25 mM glucose, and 1%
-mercaptoethanol. The EF1-
3
3
ATPase was adsorbed into the flow cell in a similar fashion, but since its MgATP-dependent rotation was tested, the CaCl2 in the reaction mixture was replaced with 2 mM MgCl2.
Immediately following infusion of the reaction buffer, the flow cell was mounted on an inverted microscope (Olympus) and illuminated with a mercury lamp. Fluorescent images were taken using a digital intensified CCD camera (Roper Scientific). Typically, 500 frames were collected from each microscope field of view at a rate of 15 frames per second. Images were processed using custom-written Matlab routines.
Other MethodsCa2+-dependent ATP hydrolysis activities were measured for 5 min at 37 °C in the presence of 50 mM Tricine-NaOH (pH 8.0), 4 mM ATP, 4 mM CaCl2, 50 mM NaCl, and 5 µg of enzyme. For measuring MgATPase activity the 4 mM CaCl2 was replaced by 2 mM MgCl2, and 10 µg of enzyme was used. Pi release was determined as described by Taussky and Shorr (24). Protein concentration was determined by the method of Lowry (25) or according to Bradford (26). SDS-PAGE was carried out on Novex Pre-Cast 1020% Tris-glycine gradient gels.
| RESULTS AND DISCUSSION |
|---|
|
|
|---|
R and
R subunits andaCF1-
C subunit, all expressed in E. coli (19). The isolated hybrid
R3
R3
C complex showed high ATPase activity, particularly in the presence of CaATP as a substrate, reaching 167 s1 even when assayed at 22 °C in its biotinylated state (Table I). The complex fully retained the unique properties of its parent ATPases, which include sensitivity to inhibition by free magnesium (16, 17), and a higher CaATPase than MgATPase activity (Table I), both being modulated by
C oxidation and reduction (19, 27). Single-molecule studies of extracted native CF1 (lacking a
subunit) have already provided some evidence for its
subunit rotation (23). But this native CF1 (-
) showed a very low MgATPase activity of 1.3 s1 at 25 °C and could be fixed to the glass surface only spontaneously. The availability of our highly active recombinant
R3
R3
C enabled us to apply protein engineering tools and other experimental methods similar to those employed with the bacterial TF1 (3, 6) and EF1-ATPases (4, 5, 9, 10).
|
R3
R3
C complex reported here was motivated by the intriguing effects of CaATP on the photosynthetic enzyme. The much higher activity of the enzyme in the presence of CaATP, as compared with MgATP (Table I), suggests that it must be caused by multiple site catalysis (28, 29) and should therefore involve
C subunit rotation within our hybrid complex. We therefore started monitoring rotational motion of individual
R3
R3
C hybrid F1-ATPase molecules in the presence of calcium, using the relatively simple and broadly applicable actin filament assay. A limitation of this assay is the load exerted by the large actin filament on the rotating enzyme, but this limitation is also an asset, as it allows determining the torque developed by the protein through the relation between filament length and rotation speed (6).
To facilitate the attachment of our hybrid complex to the glass surface, via absorption to an HRP-Ni-NTA conjugate, a His6-tag was introduced at the N terminus of the
R subunit, as described by Omote et al. (4). Attachment of the fluorescently labeled actin filament at the opposite end of the molecule was achieved by forming a biotin-streptavidin-biotin linkage between the filament and the specific cysteine pair in
C. Using a streptavidin-conjugated peroxidase we demonstrated that the biotinylation appears specifically in the
C subunit and is dependent on the prereduction of its disulfide (Fig. 1B, lanes 1 and 2). The binding of biotin had no significant effect on either Mg2+-dependent or Ca2+-dependent ATPase activity. However, at 22 °C, the temperature of the rotation assays, MgATPase activity was reduced much more than CaATPase activity (Table I).
Molecules of our
R3
R3
C complex readily adsorbed to a glass surface covered with HRP-Ni-NTA, but no enzyme attachment occurred in the absence of the HRP-Ni-NTA conjugate. The fluorescent actin filaments attached to the immobilized proteins only after addition of streptavidin and were easily washed out if the biotinylated
R3
R3
C complex was not present in the flow cell (data not shown). About 100 fluorescent actin filaments could be observed in each field of view of the microscope, and on the average one molecule in a field showed rotational motion in presence of the CaATP buffer, a percentage that is comparable with previous reports (6, 9). An example of a rotating filament is shown in Fig. 2A, and the rotational trajectories of several molecules are plotted in Fig. 2B.
|
3
3
MgATPase, whose rotational motion was already thoroughly investigated (4, 5, 9, 10). A comparison of the EF1-MgATPase and
R3
R3
C-CaATPase-induced rotation is presented in Fig. 3, in which the rotational velocity (in inverse seconds) is plotted as a function of filament length for the two groups. Only molecules that showed at least 5 full rotations (>70% of the rotating molecules) were included in this figure. The solid lines (see legend to Fig. 3) were calculated based on the following relation between the rotation velocity in Hz,
, and the applied torque N:
= 1/3
cat + 2
/N, in which
cat is the turnover time of the enzyme, and
is the rotational friction coefficient that depends on the filament length L. For a rigid filament
is given by 4
L3/3[ln(L/2r)0.447], in which
is the viscosity of the solution and r is the filament radius, taken as 5 nm (6, 30). The torque developed by each molecule was calculated by inverting the above equation. The average torque was 14.6 ± 1.5 pN·nm for the hybrid
R3
R3
C and 16.8 ± 1.4 pN·nm for the EF1-
3
3
molecules. While these values are smaller than previously reported (46), the high similarity between the two groups of molecules is quite striking and suggests that they rotate in a very similar fashion.
|
R3
R3
C complexes could rotate in the presence of MgATP, very similarly to the CaATP-induced rotation, but only when 100 mM of bicarbonate were added (data not shown). This finding is on par with earlier reports that the MgATPase activity of both CF1 (16) and RF1 (17) is very sensitive to inhibition by free Mg2, and in both cases high rates of catalysis require addition of oxyanions such as bicarbonate or sulfite (19, 31). In addition, at 22 °C under the conditions of the rotation assay, the MgATPase activity of the
R3
R3
C complex was 4-fold lower than its CaATPase activity (Table I). This difference could account for the lack of any MgATP-induced
C rotation in the absence of bicarbonate. Indeed, a strong, nonlinear relation between ATPase activity and the number of rotating molecules was already noted by Müller et al. (32).
How can CaATP-induced rotational motion of the photosynthetic
R3
R3
C complex be explained, in view of the decoupling activity of calcium? Current models of the function of F0F1-ATP synthase assert that the protonmotive force created by proton translocation through the F0 part drives rotation of the c subunit oligomer, which in turn causes rotation of the
subunit, leading to ATP synthesis by the F1 part (33, 34). A reverse sequence of events is expected to occur following ATP hydrolysis, which should drive rotation of the
subunit, leading to rotation of the c subunit oligomer that is coupled to proton translocation. Indeed, it has been shown that
rotation is contemporaneous with rotation of the c subunit oligomer in both TF1 and EF1 (9, 35). Boyer's binding change mechanism recognizes three asymmetric catalytic sites on the three
subunits (1). These three sites interconvert as the
subunit rotates. Efficient ATP synthesis without waste of energy requires coupling between events occurring at the three sites and proton translocation. Thus proton translocation must occur only after ADP and Pi bind at a loose catalytic site, while at the same time an ATP molecule is bound at the tight site (36). This scheme was borne out by experiments carried out by Cross and co-workers (37) who showed that a protonmotive force cannot sustain rotation in the absence of ADP and Pi in the medium.
Consequently, there has to be a switch mechanism that couples or decouples the F1 and F0 parts and transmits information between them in both directions (38, 39). It could involve some form of a conformational change induced by substrate binding at the
catalytic sites and further transmitted through the
subunit to the F0-c subunits. The combined effects of calcium, which include both a modulation of F1 activity and decoupling of F1 and F0, point to its involvement with the same switch mechanism suggested by Cross (36). Since the current study shows that the
R3
R3
C decoupled CaATPase activity does not hinder
C rotation, it is likely that the switch works in parallel to, but outside, the rotational pathway. It might involve, for example, the F1-
subunit, which is a well known inhibitor of both CaATPase and MgATPase activity in the photosynthetic F1 (27, 40).
We cannot currently rule out a direct effect of calcium at a more "downstream" position than suggested here, e.g. at the F0-c subunit. However, we believe that our results, combined with previous indications in the literature (36, 37), do suggest that the calcium effect is related to a long range pathway for allosteric communication between the F0 and F1 parts of the ATP synthase. The experimental system described in this work provides an opportunity to directly check this hypothesis. In particular, the difference between CaATP and MgATP binding at the active sites on the interface between the
and
subunits should be probed. This can now be studied in our photosynthetic complex, since we have already assembled a hybrid
R3
R3
C containing the catalytic site mutant
R-T159S (41), which shows exceptionally high MgATPase but very low CaAT-Pase activities (19). Carmeli and co-workers (42) suggested recently, based on a simulation, that the average bond length of a Ca2+ ion at the active site of F1 is longer by 0.3 Å than that of Mg2+. It is as yet unknown how such local differences in binding can propagate through the structure of the
subunit into the CF1
and the F0 c and a subunits. It will therefore be important to attempt to assemble the c subunit oligomer into the hybrid protein and test rotation of this part of the enzyme in the presence of calcium. These and other future studies will be facilitated by the current availability of a fully active recombinant photosynthetic protein that shows rotational motion.
| FOOTNOTES |
|---|
¶ Present address: Dept. of Physiology, University of Wisconsin, Madison, WI 53706. ![]()
** Present address: Dept. of Pathology, University of Tennessee Health Science Center, Rm. 411, Molecular Science Bldg., Memphis, TN 38163. ![]()

To whom correspondence should be addressed. E-mail: gilad.haran{at}weizmann.ac.il.
1 The abbreviations used are: CF1, EF1, RF1 and TF1, F1 -ATPases from chloroplasts, E. coli, R. rubrum, and thermophilic Bacillus PS3, respectively;
R,
R-
, and
subunits from Rhodospirilum rubrum;
C,
subunit from spinach chloroplast; HRP-Ni-NTA, horseradish peroxidase nickel-nitrilotriacetic acid; Tricine, N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine; N, newton. ![]()
| ACKNOWLEDGMENTS |
|---|
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
H. S. Samra, F. Gao, F. He, E. Hoang, Z. Chen, P. A. Gegenheimer, C. L. Berrie, and M. L. Richter Structural Analysis of the Regulatory Dithiol-containing Domain of the Chloroplast ATP Synthase {gamma} Subunit J. Biol. Chem., October 13, 2006; 281(41): 31041 - 31049. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |