Erated at the interface between the lower portion of subunit c 
and the conserved DELSEED-portion in subunit b [8?2]. It has been experimentally established that ATP hydrolysis drives the rotation of the Cterminus of subunit c relative to the hydrophobic bearing formed by the pseudohexagon (ab)3 (Fig. 1) [13?6]. On the other hand, an engineered cross-link between the 
rotor (C-terminus of subunit c) and the stator ((ab)3) of the F1-ATPase has neither impeded ATP hydrolysis, nor the ATP-driven rotation of the non-fixed portion of subunit c [16,17]. This observation has been interpreted to reveal the unfolding of the C-terminal a-helix to generate aswivel joint between neighboring residues. Moreover, in several experiments in F1 of various organisms the C- and N-termini of subunit c were deleted without inactivating the ATPase activity [8?2]. It seems that only a small portion of subunit c is necessary for torque generation. Here, we extended the former work of our group and aimed to identify the domain on subunit c, which is prone to being unfolded by the enzyme-generated torque. Six mutants of Escherichia coli F1ATPase (EF1) were compared (Fig. 1). In this context the original cysteine-free (ab)3c-complex KH7 served as the `wild type’ enzyme. Each mutant contained two engineered cysteines for a rotor-to-stator cross-link formation, namely (-)-Calyculin A chemical information cA285C/aP280C (MM10), cG282C/aP280C (GH54), cI279C/aP281C (FH4), cL276C/aE284C (GH19), cL262C/aA334C (PP2), and cA87C/bD380C (SW3). In the former four mutants (MM10, GH54, FH4, GH19) the cross-link is located at the C-terminal end of subunit c (top), while in the latter two mutants the cross-link is located in the middle of the C-terminal a-helix (PP2) and the bottom (SW3) near the PD1-PDL1 inhibitor 1 globular portion of subunit c (i.e. towards FO), respectively. The top portion of subunit c consists of a single a-helix, while in the middle the C-terminal a-helix encounters its N-terminal counterpart. At the bottom subunit c interacts with the DELSEED region of the b subunits, which serves as a lever to open its nucleotide-binding site. The activity of all mutants was monitored, both under reducing (no cross-link) and oxidizing (closed disulfide bridge between rotor and stator) conditions (Tab. 1), i.e. the rate of ATP hydrolysis in bulk solution, the cross-Unfolding of Subunit Gamma in Rotary F-ATPaseFigure 1. Model of E. coli F-ATPase. The model shows subunits a (green), b (red), and c (blue), and the cross-link positions between subunits c and a/b. For the sake of clarity subunit c and only one copy each of subunit a and b are shown. Subunits d and e are omitted. The white lines show the positions of 24786787 the cross-link sites in the respective mutants. Due to the three-fold symmetry subunit c can form a cross-link with any of the three copies of a/b (original named E, DP, and TP [1]), the exact a/b subunit is not relevant for our experiments. The arrows indicate the positions of the lever with the DELSEED-sequence in subunit b, and the cranked shaft of the coiled coil in subunit c. Throughout the text the C-terminus of subunit c with cross-link position MM10 is referred as the top. The membrane embedded FO with the c-ring (not shown) connect to the globular portion of subunit c at the bottom. In between is the middle region with the coiled coil a-helices. The structural model is based on PDB ID: 3OAA [3]. doi:10.1371/journal.pone.0053754.glink yield in SDS-gels, and the rate of c-rotation of single molecules was determined.Ma.Erated at the interface between the lower portion of subunit c and the conserved DELSEED-portion in subunit b [8?2]. It has been experimentally established that ATP hydrolysis drives the rotation of the Cterminus of subunit c relative to the hydrophobic bearing formed by the pseudohexagon (ab)3 (Fig. 1) [13?6]. On the other hand, an engineered cross-link between the rotor (C-terminus of subunit c) and the stator ((ab)3) of the F1-ATPase has neither impeded ATP hydrolysis, nor the ATP-driven rotation of the non-fixed portion of subunit c [16,17]. This observation has been interpreted to reveal the unfolding of the C-terminal a-helix to generate aswivel joint between neighboring residues. Moreover, in several experiments in F1 of various organisms the C- and N-termini of subunit c were deleted without inactivating the ATPase activity [8?2]. It seems that only a small portion of subunit c is necessary for torque generation. Here, we extended the former work of our group and aimed to identify the domain on subunit c, which is prone to being unfolded by the enzyme-generated torque. Six mutants of Escherichia coli F1ATPase (EF1) were compared (Fig. 1). In this context the original cysteine-free (ab)3c-complex KH7 served as the `wild type’ enzyme. Each mutant contained two engineered cysteines for a rotor-to-stator cross-link formation, namely cA285C/aP280C (MM10), cG282C/aP280C (GH54), cI279C/aP281C (FH4), cL276C/aE284C (GH19), cL262C/aA334C (PP2), and cA87C/bD380C (SW3). In the former four mutants (MM10, GH54, FH4, GH19) the cross-link is located at the C-terminal end of subunit c (top), while in the latter two mutants the cross-link is located in the middle of the C-terminal a-helix (PP2) and the bottom (SW3) near the globular portion of subunit c (i.e. towards FO), respectively. The top portion of subunit c consists of a single a-helix, while in the middle the C-terminal a-helix encounters its N-terminal counterpart. At the bottom subunit c interacts with the DELSEED region of the b subunits, which serves as a lever to open its nucleotide-binding site. The activity of all mutants was monitored, both under reducing (no cross-link) and oxidizing (closed disulfide bridge between rotor and stator) conditions (Tab. 1), i.e. the rate of ATP hydrolysis in bulk solution, the cross-Unfolding of Subunit Gamma in Rotary F-ATPaseFigure 1. Model of E. coli F-ATPase. The model shows subunits a (green), b (red), and c (blue), and the cross-link positions between subunits c and a/b. For the sake of clarity subunit c and only one copy each of subunit a and b are shown. Subunits d and e are omitted. The white lines show the positions of 24786787 the cross-link sites in the respective mutants. Due to the three-fold symmetry subunit c can form a cross-link with any of the three copies of a/b (original named E, DP, and TP [1]), the exact a/b subunit is not relevant for our experiments. The arrows indicate the positions of the lever with the DELSEED-sequence in subunit b, and the cranked shaft of the coiled coil in subunit c. Throughout the text the C-terminus of subunit c with cross-link position MM10 is referred as the top. The membrane embedded FO with the c-ring (not shown) connect to the globular portion of subunit c at the bottom. In between is the middle region with the coiled coil a-helices. The structural model is based on PDB ID: 3OAA [3]. doi:10.1371/journal.pone.0053754.glink yield in SDS-gels, and the rate of c-rotation of single molecules was determined.Ma.