The Role of the βDELSEED Motif of F1-ATPase

In F1-ATPase, a rotary motor enzyme, the region of the conserved DELSEED motif in the β subunit moves and contacts the rotor γ subunit when the nucleotide fills the catalytic site, and the acidic nature of the motif was previously assumed to play a critical role in rotation. Our previous work, however, disproved the assumption (Hara, K. Y., Noji, H., Bald, D., Yasuda, R., Kinosita, K., Jr., and Yoshida, M. (2000) J. Biol. Chem. 275, 14260–14263), and the role of this motif remained unknown. Here, we found that the ε subunit, an intrinsic inhibitor, was unable to inhibit the ATPase activity of a mutant thermophilic F1-ATPase in which all of the five acidic residues in the DELSEED motif were replaced with alanines, although the ε subunit in the mutant F1-ATPase assumed the inhibitory form. In addition, the replacement of basic residues in the C-terminal region of the ε subunit by alanines caused a decrease of the inhibitory effect. Partial replacement of the acidic residues in the DELSEED motif of the β subunit or of the basic residues in the C-terminal α-helix of the ε subunit induced a partial effect. We here conclude that the ε subunit exerts its inhibitory effect through the electrostatic interaction with the DELSEED motif of the β subunit.

In F 1 -ATPase, a rotary motor enzyme, the region of the conserved DELSEED motif in the ␤ subunit moves and contacts the rotor ␥ subunit when the nucleotide fills the catalytic site, and the acidic nature of the motif was previously assumed to play a critical role in rotation. Our previous work, however, disproved the assumption ( Here, we found that the ⑀ subunit, an intrinsic inhibitor, was unable to inhibit the ATPase activity of a mutant thermophilic F 1 -ATPase in which all of the five acidic residues in the DELSEED motif were replaced with alanines, although the ⑀ subunit in the mutant F 1 -ATPase assumed the inhibitory form. In addition, the replacement of basic residues in the C-terminal region of the ⑀ subunit by alanines caused a decrease of the inhibitory effect. Partial replacement of the acidic residues in the DELSEED motif of the ␤ subunit or of the basic residues in the C-terminal ␣-helix of the ⑀ subunit induced a partial effect. We here conclude that the ⑀ subunit exerts its inhibitory effect through the electrostatic interaction with the DELSEED motif of the ␤ subunit. F 1 , together with the membrane-embedded proton-conducting unit F o , forms the F o F 1 -ATP synthase that reversibly couples trans-membrane proton flow to ATP synthesis/hydrolysis (1)(2)(3)(4)(5)(6). Isolated F 1 has ATP-hydrolyzing activity, hence is called F 1 -ATPase, and has an ␣ 3 ␤ 3 ␥␦⑀ subunit structure in which ␣ and ␤ subunits have noncatalytic and catalytic nucleotidebinding sites, respectively. The ␥ subunit inserts its coiled-coil structure into the central cavity of the ␣ 3 ␤ 3 hexagonal ring (7) and rotates relative to the ␣ 3 ␤ 3 ring during ATP hydrolysis (8 -13). The conformation of the ␤ subunit with a nucleotidefilled catalytic site is significantly different from that of the ␤ subunit with an empty catalytic site; specifically, the C-terminal helical domain is lifted up to or swung down from the nearly immobile N-terminal domain, respectively. The acidic cluster sequence, known as the DELSEED motif, in the C-terminal domain of the ␤ subunit moves in contact with the ␥ subunit when the catalytic site is filled with nucleotide. This sequence has been well conserved in all F 1 s with minor variations. For example, DELSDED in F 1 from a thermophilic Bacillus PS3 (TF 1 ) 1 is assumed to play an essential role in the efficient coupling between catalysis and transport (14). However, a mutant ␣ 3 ␤ 3 ␥ complex of TF 1 , in which acidic residues in the ␤DELSDED sequence were replaced with alanines, showed the rotary motion of the ␥ subunit with a normal torque value, 40 pN⅐nm (15). Thus, negative charges in this motif do not have a direct role in torque generation, and the role of the ␤DELSEED motif remains unclear.
The ⑀ subunit rotates along with the ␥ subunit (16 -18), and because the ⑀ subunit has been known as an intrinsic inhibitor of the ATPase activity of F 1 and of the ATP-driven proton translocating activity of F o F 1 -ATP synthase (19 -22), the mechanism of how the ⑀ subunit inhibits catalysis while it is rotating is intriguing. We found that the ⑀ subunit of TF 1 can adopt two conformations, an inhibitory state and a noninhibitory state (23). The high resolution structure of the ⑀ subunit of Escherichia coli F 1 -ATPase (EF 1 ) (24 -26) most likely corresponds to the noninhibitory state. When ATP-Mg is added to TF 1 (␣ 3 ␤ 3 ␥␦⑀) or the ␣ 3 ␤ 3 ␥⑀ complex, ATP hydrolysis starts slowly, accelerates with time, and then reaches the final rate. This time-dependent activation was not observed for the ␣ 3 ␤ 3 ␥ and ␣ 3 ␤ 3 ␥␦ complexes (23,27). From the results, we suspected that a slow nucleotide-dependent conformational transition of the ⑀ subunit in TF 1 during the transition from inhibitory to noninhibitory state occurred. It should be mentioned that EF 1 shows an apparently similar time-dependent activation that is derived from a different mechanism: dissociation of the ⑀ subunit from EF 1 during catalysis (19).
We had shown that a mutant ⑀ subunit lacking the C-terminal half had no inhibitory effect on TF 1 (22). In EF 1 , a cysteine residue introduced at the C-terminal end of the helix ␣ 2 (residues from Lys 114 to Lys 133 in TF 1 ; see Ref. 25) of the ⑀ subunit was cross-linked with a cysteine introduced in the ␤DELSEED motif (28 -30). 4 -6 positively charged residues exist in the C-terminal helix ␣ 2 of the ⑀ subunit of almost all F 1 s, and there * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Supported by research fellowships from the Japan Society for the Promotion of Science for Young Scientists.
§ To whom correspondence should be addressed: Masasuke Yoshida Chemical Resources Lab., R-1, Tokyo Inst. of Technology, Nagatsuta 4259, Yokohama, 226-8503, Japan. Fax: 81-45-924-5277; E-mail: myoshida@res.titech.ac.jp. 1 The abbreviations used are: TF 1 , F 1 -ATPase from thermophilic Bacillus PS3 (a soluble portion of F o F 1 -ATP synthase); MF 1 , F 1 from bovine heart mitochondria; EF 1 , F 1 from E. coli; helix ␣ 2 , the C-terminal end of the ␣-helices of the ⑀ subunit (residues from Lys 114 to Lys 133 in TF 1 ; see Ref. 25); wtЈ complex, a mutant ␣(C193S) 3 ␤(10H) 3 ␥(S107C) complex of TF 1 used as a second wild-type complex here (the enzymatic characteristics are nearly unchanged from the real wild-type ␣ 3 ␤ 3 ␥ complex); N 5 A, a mutant complex in which all five acidic residues in the ␤DELSDED sequence of wtЈ complex were replaced with alanines (␤AALSAAA); ⑀ WT , wild-type ⑀ subunit of TF 1 ; ⑀ NCX , ⑀ P1A , ⑀ P3A , and ⑀ P4A , mutant ⑀ subunits of TF 1  is a possibility that the negative charges in the ␤DELSEED motif interact with positive charges in the C-terminal helix ␣ 2 of the ⑀ subunit. Therefore, this electrostatic interaction may be critical for inhibition of ATPase activity by the ⑀ subunit. To examine this possibility, we compared the ATPase activities of the mutant ␣ 3 ␤ 3 ␥ and the ␣ 3 ␤ 3 ␥⑀ complexes of TF 1 in which acidic residues in the ␤DELSDED sequence and basic residues in the ⑀ subunit were replaced with alanines. The results indicate that the interaction between the negative charges in the ␤DELSEED motif and the positive charges in the C-terminal end of helix ␣ 2 of the ⑀ subunit is essential for the ⑀ subunit to play a role as an inhibitor for F 1 -ATPase.

EXPERIMENTAL PROCEDURES
Construction of Expression Plasmid for ␣ 3 ␤ 3 ␥ Complex and ⑀ Subunit of TF 1 -The ␣ 3 ␤ 3 ␥ complex of TF 1 containing the mutations (␣C193S, 10 His at N termini of ␤, ␥S107C; wtЈ) was considered to be a wild-type complex in this work. The expression plasmid for the wtЈ complex was pKABG1 (31), and the mutations in the ␤DELSDED sequence were introduced into the plasmid as described previously (15). The expression plasmid for the wild-type ⑀ subunit (⑀ WT ) was pTE2 (27), and that for the mutant ⑀ subunit (S48C/N125C; ⑀ NCX ) of TF 1 was prepared as described previously (23). Alanine mutations (K133A, ⑀ P1A ; K121A/ R122A/R126A, ⑀ P3A ; K121A/R122A/R126A/K133A, ⑀ P4A ) were introduced into pTE2 by the method of Kunkel et al. (32). Using a polymerase chain reaction method with rTaq polymerase (Takara Inc.), NdeI and HindIII sites were introduced into the 5Ј and 3Ј termini of the mutant ⑀ subunit gene, respectively. The polymerase chain reaction products were digested by NdeI and HindIII, and the fragments obtained were cloned into pET21c vector (Novagen Inc.) for efficient expression.
Proteins-The mutant ␣ 3 ␤ 3 ␥ complexes of TF 1 were purified as described (31). The wild-type and mutant ⑀ subunits of TF 1 were prepared as described previously (23,33). The ␣ 3 ␤ 3 ␥⑀ complexes were reconstituted from ␣ 3 ␤ 3 ␥ and the ⑀ subunit by incubation for 30 min at room temperature. To remove the unbound ⑀ subunit, the mixture was concentrated several times with a centrifuge concentrator (Microcon 100; 100,000 M r cut-off; Millipore Amicon), and their purity was analyzed by 6% (w/v) PAGE without SDS (native-PAGE) in the presence of 20 M ATP and 2 mM MgCl 2 .
Measurement of ATPase Activity-ATPase activity was measured spectrophotometrically with an ATP regenerating system at 25°C (34). The assay mixture contained 50 mM Tris-HCl (pH 8.0), 100 mM KCl, 2.5 mM phosphoenolpyruvate, 2 mM MgCl 2 , 0.2 mM NADH, 50 g/ml pyruvate kinase, 50 g/ml lactate dehydrogenase, and 20 M or 2 mM of ATP-Mg. The reaction was initiated by the addition of the enzyme solution into 1 ml of assay mixture. The change in absorbance at 340 nm was monitored in a spectrophotometer V-550 (Jasco, Japan) and stored in an on-line computer. In the experiments shown in Table I, the rates of ATP hydrolysis at 20 M ATP in the period 5-6 min after the addition of the ␣ 3 ␤ 3 ␥ and ␣ 3 ␤ 3 ␥⑀ complexes were compared.
Conformational Transition of the ⑀ Subunit in ␣ 3 ␤ 3 ␥⑀ Complex-Formation of the intramolecular cross-link in the mutant ⑀ subunit in ␣ 3 ␤ 3 ␥⑀ complex was examined in the presence or the absence of 20 M or 2 mM ATP-Mg. The isolated ␣ 3 ␤ 3 ␥⑀ NCX-RED complex (1 mg/ml) was incubated with 40 M or 4 mM ATP and 4 mM MgCl 2 for 10 min at 25°C in 50 mM TES-NaOH (pH 7.0) and 100 mM NaCl. As a control, 4 mM MgCl 2 was added in the absence of ATP. Cross-linking was induced by the addition of equal volumes of 100 M of CuCl 2 solution (in 50 mM TES-NaOH (pH 7.0) and 100 mM NaCl), which gives the final concentration of ATP-Mg at 0 and 20 M and 2 mM, respectively. After incubation for 1 h at 25°C, the reaction was quenched by the addition of 10 mM EDTA. After 10 min, 0.1% (w/v) SDS and 25 M TMR maleimide were added to label cysteines that did not form disulfide bonds, and the solution was incubated at 25°C for 10 min. Then 10 mM N-ethylmaleimide was added to terminate the labeling. Incorporation of TMR maleimide into the ⑀ subunit was analyzed on 15% (w/v) SDS-PAGE without reducing reagent. Other conditions were the same as those described by Kato-Yamada et al. (23).
Other Procedures-Protein concentrations were determined by the method of Bradford (35) using bovine serum albumin as a standard.

Effect of the ⑀ Subunit on ATPase Activity of N 5 A Mutant
Complex-The ATPase activity of the N 5 A mutant complex, in which all of the five acidic residues in the ␤DELSDED sequence of the wtЈ complex were replaced with alanines, was measured and compared with the activity of the wtЈ complex. The reaction was initiated by the addition of the ␣ 3 ␤ 3 ␥ or the ␣ 3 ␤ 3 ␥⑀ complex, designated as the Ϫ⑀ complex and the ϩ⑀ complex, respectively, to the reaction mixture. As reported previously (27), ATP hydrolysis at low ATP concentration by the wtЈ(ϩ⑀) complex was very slow in comparison with the wtЈ(Ϫ⑀) complex, and the inhibitory role of the ⑀ subunit is thus evident (Fig. 1A). It should be noted that the inhibition by the ⑀ subunit was relieved slowly and that the ATPase activity was recovered gradually by the reaction period. At 2 mM ATP, this recovery was much faster, and the inhibitory effect of the ⑀ subunit was only observed at the very initial phase of the reaction (Fig. 1B). As stated previously, this recovery is due to the nucleotide-dependent conformational transition of the ⑀ subunit (23). On the contrary, the N 5 A(ϩ⑀) complex hydrolyzed ATP at the same rate as the N 5 A(Ϫ⑀) complex at both 20 M and 2 mM ATP (Fig. 1, C and D). Apparently, the N 5 A complex lost sensitivity to the inhibition by ⑀ subunit.
The ⑀ Subunit Does Not Dissociate from the N 5 A(ϩ⑀) Complex during ATP Hydrolysis-The results shown in Fig. 1 could be explained even if it were by the dissociation of the ⑀ subunit from the N 5 A complex rapidly during ATP hydrolysis. To confirm that this is not the case, we incubated the N 5 A(ϩ⑀) complex with 20 M ATP-Mg under the same conditions used for the measurement of the ATPase activities. After a 30-min incubation, the complex was isolated by gel filtration HPLC in the presence of 20 M ATP-Mg (Fig. 2). The homogeneity of the N 5 A(ϩ⑀) complex was further demonstrated by native-PAGE, because the complex electrophoresed as a single band (Fig. 2,  inset b). The electrophoretic mobility of the N 5 A(ϩ⑀) complex was slightly slower than that of the N 5 A(-⑀) complex. SDS-PAGE analysis of the isolated complex showed that the N 5 A(ϩ⑀) complex contained all of the ␣, ␤, ␥, and ⑀ subunits with the correct stoichiometry (Fig. 2, inset a). These results strongly suggest that the N 5 A complex lost its sensitivity to inhibition by the ⑀ subunit without dissociation of this subunit.
Conformational Transition of ⑀ NCX in the N 5 A(ϩ⑀) Complex-To confirm the nucleotide-dependent conformational change of the ⑀ subunit in the N 5 A(ϩ⑀) complex, we used a mutant ⑀ subunit, ⑀ NCX , in which two cysteine residues were introduced at the interface of the two domains (S48C/N125C) (23). ⑀ NCX inhibited the wtЈ complex but not the N 5 A complex as wild-type ⑀ (Fig. 1, A and C). When the N-terminal ␤ strand domain and C-terminal ␣-helical domain packed closely in ⑀ NCX ("closed" form) as determined by NMR and x-ray crystallography (24 -26), an interdomain disulfide is readily formed by the assistance of the low concentrations of CuCl 2 . However, if these two domains assume the "open" form, a disulfide bond is no longer formed, and the free sulfide residues are susceptible to labeling by fluorescent maleimide. This efficiency of labeling by TMR maleimide is therefore a good indicator of the existence of the open form of ⑀ NCX in which a disulfide bond is very rarely formed. Electrophoretic mobility of ⑀ NCX containing a disulfide is slightly faster than the ⑀ NCX with TMR-labeled sulfides, and this distinction was also adopted as a means to assess the open and closed forms. In the presence of 2 mM ATP and 50 M CuCl 2 , ⑀ NCX both in the wtЈ and the N 5 A complex was largely unlabeled by TMR maleimide, and the protein bands corresponding to ⑀ NCX with a disulfide were dominant (Fig. 3). On the other hand, disulfide formation in ⑀ NCX in the N 5 A complex did not occur the presence or the absence of 20 M ATP, similar to wtЈ complex. These results imply that the ⑀ subunit in the complex can undergo a nucleotide-dependent conformational transition to the noninhibitory state, although 20 M ATP-Mg was insufficient to this transition .
Effect of Partial Elimination of Negative Charges in the ␤DELSEED Motif-To clarify the important residues in the ␤DELSEED motif for inhibition by the ⑀ subunit, we generated seven mutant ␣ 3 ␤ 3 ␥ complexes in which the acidic residues in the ␤DELSDED sequence were substituted by alanines. Substitution of the last three acidic residues (DELSAAA) efficiently suppressed the propagation of the inhibitory effect of the ⑀ subunit (Table I). The replacement of the first two acidic residues (AALSDED) also resulted in suppression of the ⑀-induced inhibition, although the mutation showed a partial effect (Table I). A single replacement at Asp 390 , Glu 391 , or Asp 396 also has some suppression on the ⑀-induced inhibition. Taken together, nearly all of the acidic residues in the ␤DELSDED sequence cumulatively contribute to the inhibitory effect of the ⑀ subunit, and the last three are the most critical.
Effect of Partial Elimination of Positive Charges in the Cterminal ␣-Helix of the ⑀ Subunit-Next we investigated whether the positively charged residues in the C-terminal ␣-helix of the ⑀ subunit contribute to the inhibitory role of the ⑀ subunit. For this purpose, we designed three new mutant ⑀ subunits of TF 1 in which some positively charged residues in the C-terminal ␣-helix were replaced with Ala, and they were designated ⑀ P1A , ⑀ P3A , and ⑀ P4A according to the number of the substituted Ala from the positively charged residue(s). The ATPase activities of the wtЈ complex with these mutant ⑀ subunits were measured in the presence of 20 M ATP (Fig. 4). The wtЈ (ϩ⑀ P4A ) complex hydrolyzed ATP at the almost same rate as the wtЈ (Ϫ⑀) complex. The wtЈ (ϩ⑀ P4A ) complex was isolated by gel filtration HPLC in the presence of 20 M ATP-Mg (Fig. 5), suggesting that the complex obtained with this mutant ⑀ subunit was stable. The wtЈ (ϩ⑀ P1A ) and wtЈ (ϩ⑀ P3A ) complexes showed partial effects on the inhibitory role of the ⑀ subunit. These results suggested that the ⑀ subunit exerts the inhibitory role through the positive charges in the C-terminal ␣-helix of the ⑀ subunit.  Inhibitory Signal from ⑀ Subunit Is Transmitted through the ␤DELSEED Motif-The ATPase activity of the N 5 A(ϩ⑀) mutant complex is resistant to inhibition by the ⑀ subunit. There are two possible explanations for this observation. Either the mutated ␤DELSEED motif interferes with the conformational transition of the ⑀ subunit to an inhibitory state, or the ␤ subunit with the mutated ␤DELSEED motif is no longer affected by the ⑀ subunit in the inhibitory state. We conclude that the latter is the case for the following reasons. First, nucleotidedependent formation of the interdomain disulfide bond in ⑀ NCX , indicative of the conformational conversion of ⑀ from the inhibitory state to the noninhibitory state, occurred in the N 5 A(ϩ⑀) complex in the same manner as in the wtЈ complex. Second, in the N 5 A(ϩ⑀) complex, the ⑀ subunit did not inhibit ATP hydrolysis at 20 M ATP (Fig. 1), although ⑀ NCX existed in the inhibitory conformation (Fig. 3). Therefore, we here propose that the inhibitory effect of the ⑀ subunit propagates through the interaction with the ␤DELSEED motif.
Large Conformational Transition of the ⑀ Subunit Is Further Supported-Because the resolution of the yeast mitochondrial F 1 ϩF o c structure was limited to 3.9 Å (36), the details of the ⑀ subunit (called ␦ in mitochondrial F 1 ) could not be clearly defined. Nevertheless, the electron density of the ␦ subunit (bacterial ⑀ subunit) can readily accommodate the structure of the isolated ⑀ subunit from EF 1 (24,26). The ␦ subunit (bacterial ⑀ subunit), however, lies on the c subunit ring, far from the ␣ 3 ␤ 3 structure, where a part of the ⑀ subunit cannot interact with the DELSEED motif of the ␤ subunit. When the ⑀ subunit acts as an inhibitor, a large conformational transition must occur in the ⑀ subunit to achieve interaction with the catalytic portion. Cross-linking experiments showed that the C-terminal domain of the ⑀ subunit is close to the ␤DELSEED motif under certain conditions (28 -30). A mutant ⑀ subunit whose C-terminal domain was truncated could not inhibit the ATPase activity of the F o F 1 -ATP synthase (22). A nucleotide-dependent conformational change in the ⑀ subunit of EF 1 was reported by Wilkens and Capaldi (26). We also have reported the nucleotide-dependent formation of an interdomain cross-link in the ⑀ subunit of TF 1 (23). Hence there are several results that strongly suggest the direct interaction of the ␤DELSEED motif with the Cterminal domain of the ⑀ subunit and a large conformational transition of the ⑀ subunit. Finally, these possibilities were supported recently by the newly reported three-dimensional structures of the MF 1 (37) and the ␥'-⑀ complex from EF 1 (38) as described below.
Interaction between the ␤DELSEED Motif and the C-terminal ␣-Helix of the ⑀ Subunit Involves Electrostatic Interactions-Elimination of negative charges in the ␤DELSEED motif cumulatively desensitized F 1 from ⑀-induced inhibition ( Table I), indicating that the ␤DELSEED motif interacts with the ⑀ subunit not by specific interactions between certain residues but through a cluster of negative charges. In addition, elimination of positive charges in the C-terminal ␣-helix of the ⑀ subunit also gave similar results (Fig. 4). The cumulative contribution of the negative charges in the ␤DELSEED motif and the positive charges in the C-terminal region of the ⑀ subunit strongly suggest that the electrostatic interaction between these two regions dominates the inhibition of the F 1 -ATPase activity.
A Model for the Conformational Transition of the ⑀ Subunit   FIG. 6. A model for the conformational transition of the ⑀ subunit and the activation of the ␣ 3 ␤ 3 ␥⑀ complex. We propose a model for the conformational transition of the ⑀ subunit and the activation of the ␣ 3 ␤ 3 ␥⑀ complex. When the complex is in the inactive state, the ␣-helical region of the ⑀ subunit extends and entwines around the ␥ subunit to enable the electrostatic interaction between the positively charged residues in the C-terminal ␣-helix of the ⑀ subunit and the negative charged residues in the ␤DELSEED motif (left panel). The addition of ATP induces the conformational transition of the ⑀ subunit and makes the ␣ 3 ␤ 3 ␥⑀ complex active (right panel). This model is based on our results and the reported structures (37,38). and Activation of ␣ 3 ␤ 3 ␥⑀ Complex-Just when we concluded the present study, the crystal structure of the ␥Ј (residues ␥11-␥258)-⑀ complex in EF 1 (38) and that of the ␣ 3 ␤ 3 ␥␦⑀ complex in MF 1 (37) were reported. These two structures revealed that the ⑀ subunit in EF 1 and the corresponding subunit in MF 1 , ␦, adopted very different conformations. The two helices of EF 1 -⑀ extended and entwined the coiled-coil structure of the ␥ subunit, and the C-terminal end reached out toward the ␣ 3 ␤ 3 ring. In contrast, the two ␣-helices of MF 1 -␦ abutted the ␤-sandwich structure and the C termini located on the surface of c subunits.
The interaction between the ␤DELSEED motif and the Cterminal ␣-helix of the ⑀ subunit must be possible if the conformation of the ⑀ subunit in the complex is similar to that reported in the ␥Ј-⑀ complex of EF 1 (38). In contrast, the crosslinking in ⑀ NCX must occur when the structure of the ⑀ subunit in the complex is in the same conformation as that of the ␦ subunit in MF 1 (37). If this is the case, our results should be well explained by the large conformational transition of the ⑀ subunit. The conformation of the ⑀ subunit in the ␥Ј-⑀ complex of EF 1 (38) and the ␦ subunit in MF 1 (37) correspond to the inhibitory state and noninhibitory state, respectively.
Interestingly, the ⑀ subunit of MF 1 lay between the ␥ subunit and the ␦ subunit like a wedge (37). This subunit seems to mimic the structure of the C-terminal ␣-helical domain of the bacterial ⑀ subunit in the complex. The C-terminal part of the MF 1 -⑀ subunit contains four positively charged residues at the counterpart of the C-terminal ␣-helix of the MF 1 -␦ subunit, which is shorter and contains only two positively charged residues. It is therefore an intriguing assumption that the C-terminal region of the ⑀ subunit of MF 1 interacts with ␤DELSEED motif and inhibits the activity in MF 1 -ATPase under certain conditions, although the role of this subunit is not known very well.
Here we propose a model for the conformational transition of the bacterial ⑀ subunit to explain the activation of ␣ 3 ␤ 3 ␥⑀ complex (Fig. 6). When the ⑀ subunit is in the inhibitory form, the C-terminal ␣-helix of the ⑀ subunit interacts with ␤DELSEED motif through their respective charges. The apparent bridge formed between the ⑀ subunit and a ␤ subunit make the ␣ 3 ␤ 3 ␥⑀ complex inactive. When the conformational change is induced in the ␤ subunit by ATP binding, the electrostatic interaction is broken, and the ⑀ subunit resumes the noninhibitory form. Then the ␣ 3 ␤ 3 ␥⑀ complex becomes active. From the present results and our previous reports, it is evident that ATP-Mg stimulates the conformational transition of the ⑀ subunit from the inhibitory state to the noninhibitory state. The information on the trigger for this conformational transition of the ⑀ subunit is requisite for further understanding of the internal regulation of the F 1 -ATPase.