Macromolecular organization of F1-ATPase isolated from Clostridium thermoaceticum as revealed by electron microscopy

Membrane vesicles and the F1-ATPase from Clostridium thermoaceticum were examined by electron microscopy. F1-ATPase particles projecting from the vesicles have a diameter of 10 to 12 nm. The F1-ATPase has an alpha 3 beta 3 gamma delta structure. The alpha and beta subunits are most likely arranged in an alternating sequence around a central protein mass consisting of the gamma and delta subunits.

with an Mr of about 370,000 consisting of four different subunits in the apparent molar ratio of a33_Y8. The quaternary structure of this enzyme appears to be similar to that of the Fl-ATPases isolated from several respiratory organisms, (4,13,14), but different from that of the enzymes isolated from two nonrespiratory bacteria (3,17). In this report, we describe an electron microscopic study of the three-dimensional structure of the Fl-ATPase from C. thermoaceticum.
The preparation of membrane vesicles and purified F1-ATPase from C. thermoaceticum was as described earlier (6), except that ATP (1 mM) was included in all buffers used for purification. Also, the final DEAE step was replaced by high-performance liquid chromatography with a TSK DEAE .5PW column (8 by 75 mm; LKB Instruments, Inc., Rockville, Md.). Fl-ATPase was eluted with a linear gradient from 0 to 1 M KCl in a buffer solution containing 50 mM Tris hydrochloride (pH 7.6), 1 mM MgC92, 1 mM ATP, and 10% (vol/vol) glycerol. The specific activity of the purified enzyme was 12.4 ,umol of Pi released min-' mg-' at pH 8.5 and 58°C.
Membrane vesicles with attached Fl-ATPase particles (100 ,ug of protein ml-') and purified Fl-ATPase particles (10 ,ug ml-') were negatively stained with an aqueous uranyl acetate solution (3.5%, [wt/vol], pH 4.8) by the diffusion technique of Valentine et al. (18). Deep stain conditions were applied as specified by Johannssen et al. (7). Samples were observed, tilt series were performed, and micrographs were taken with a Philips EM 400 ST electron microscope operated at 80 keV in the conventional transmission mode. Calibration of instrument magnifications was done as described previously (12). Micrographs were taken at magnifications ranging from 46,000x to 60,000x. Tilt series were done at 46,000x. Measurements of particle sizes were made from prints at calibrated magnifications ranging from 143,000x to 200,000x. * Corresponding author.
Membrane vesicles with attached F1-ATPase particles are shown in Fig. 1. The number of F1-ATPase particles per vesicle is relatively low, often not exceeding three for a vesicle of 50-nm diameter. Frames 3 to 5 of Fig. 1 clearly demonstrate a stalk connecting the F1-ATPase particle with the vesicle. The attached F1-ATPase particles have diameters between 10 and 12 nm, which is similar to the diameter obtained for F1-ATPase from mitochondria (16). The projections representing F1-ATPase particles are not seen in vesicles stripped of F1-ATPase (6), and rebinding of F1-ATPase to stripped vesicles results in the reappearance of the projections.
Micrographs of purified F1-ATPase are shown in Fig Figure 2 is an overview of a negatively stained sample, whereas Fig. 3 and 4 depict results of tilting experiments. Figure 2 shows that the F1-ATPase sample used for the electron microscopic analysis was homogeneous. The diameter of the particle types 1 to 4, assumed to be face views of F1-ATPase, is 11 + 1.5 nm. Type 6 particle projections are approximately 11 nm in the direction parallel to the long axis and 8 to 9 nm perpendicular to that axis and are interpreted to be projections of F1-ATPase particles attached to the support film in a tilted fashion. Type 5 projections are also interpreted to be views of attached F1-ATPase particles slightly tilted with respect to the most commonly seen face-view. Type 1 projections in Fig. 2 reveal six intensity maxima arranged in a circle at average intervals of approximately 60°. The center of these projections appears to be white, indicating the presence of a proteinaceous mass therein. The difference between type 1 and type 2 projections is that type 2 projections exhibit negative stain trapped in the center of the particles, causing a dark central dot. Type 3 is very similar to types 1 and 2; however, the six intensity maxima arranged in a circle in types 1 and 2 appear to be arranged into two groups, each group being made up of three intensity maxima arranged in a plane at 1200 intervals. One plane appears to be rotated 600 with respect to the other, thus giving rise to a threefold rotational symmetry. The existence of that type of symmetry is confirmed by the existence of type 4 projections, clearly indicating the subdivision of the particle mass into three submasses arranged at angles of 1200 (Fig. 2).
The four tilt series presented in Fig. 3 demonstrate that the different views of F1-ATPase depicted in Fig. 2 4, 7, 9, 10, and 12 of Fig. 3) exhibits a smaller diameter than that meas'ured from face views (fra'mes 2, 5, 8, and 11 of Fig.   3)..The! central mass of the Fl-ATPase is clearly visible in most face views shown in Fig. 3 (frames 2, 5, 6 , 7, 8, 9, 11, and 12). This indicates that the failure to see the central mtass observed in several cases (types 2 and 3 in Fig. 2) is probably not caused by the absence of such a central mass, but rather by a superimposition of trapped negative-staining salt in the central area of the Fl-ATPase. Thus, the imaging of a central protein mass is hindered.
In the tilt series in Fig. 4 (frames A to C) the central mass vi'sible in C moves to the periphery of the projection when the particle is tilted. This indicates that the central mass is located in a plane different from that of the surrounding major masses. The latter are considered to represent the three ax and three P subunits, whereas the central mass may represent the -y and 8 subunits that may form a stalklike structure. Frames D to G of Fig. 4 show that the central mass  Fig. 3. Bar, 20 nm. Frames A to C constitute the tilt series of one projection (lower left corner) of frames 10 to 12 from Fig. 3, and D to G are selected projections from frames 2, 5, 7, and 8 from Fig. 3. appears to be connected to the peripheral masses with spokes. It is likely that each a and subunit is so connected with the central mass, since spokes are visibly connecting neighboring peripheral masses. According to Lunsdorf et al. (11), the neighboring subunits of the peripheral structure of Fl-ATPase of Escherichia coli are never the same. Thus one at subunit is flanked by two subunits and vice versa.
The results presented here, when compared with previous findings obtained with F1-ATPase isolated from mitochondria (16), chloroplasts (1), and E. coli (1,11), strongly suggest that the quaternary structure of the C. thermoaceticum enzyme is of the form a3P33Y8, with the a and subunits arranged in alternating sequence around the periphery of the particle. In addition, the angles between the three a subunits, as in the mitochondrial Fl-ATPase (16), seem to be 1200. The three a subunits are located in one layer. In a second layer the 3 subunits, also separated from one another by 1200, are each offset relative to each ax subunit by 600. This appears to fit with the model of Tiedge et al. (16) for the mitochondrial Fl-ATPase. The central mass usually seen in face views is assumed to represent the y and 8 subunits.
The structural similarity between the Fl-ATPase isolated from C. thermoaceticum and those isolated from mitochondria, chloroplasts, and E. coli implies a functional similarity as well. We feel that the enzyme from C. thermoaceticum, like its counterparts in respiratory organisms, can synthesize ATP by using an electrochemical gradient of protons as the driving force. The mechanism by which this bacterium produces an electrochemical gradient is unknown, but probably involves the membrane-bound carriers cytochrome b and menaquinone (5) and perhaps one or several unknown additional electron acceptors that carry electrons to the pathway of the synthesis of acetate from CO2 (9)-This work was supported by Public Health Service grant AM-27323 from the National Institute of Arthritis, Diabetes, Digestive and Kidney Diseases and by project DE-AS09-79ER10499 from the U.S. Department of Energy. Support for F.M. as visiting scientist at the University of Georgia was from Georgia Power Co. in support of a Distinguished Professorship in Biotechnology. The electron microscopy was performed at the Center for Advanced Ultrastructural Research, University of Georgia.