To rotate. Only these filaments were considered further. We checked whether enzymes, that stopped rotating after formation of a cross-link, were still active. We could reactivate oxidized enzymes by re-reducing the cysteines. In multiple cycles of reduction, oxidation, and re-reduction enzymes were active, inactive, and active again, respectively [24]. When changing to oxidizing conditions by addition of 4 mM DTNB (8 mM DTNB for FH4) for 10 minutes, 16 out of 29 (55 ) molecules of the mutant GH54 continued to Benzocaine rotate while 13 stopped, typically after 2 minutes. The rotation in the still active subset could be observed significantly longer (for at least 10 minutes). The observation time was limited only by the bleaching of the fluorescent dyes under oxidizing condition. For FH4 only 3 out of 11 (27 ) molecules remained active. In addition, in both cases the rotational rate decreased, typically by 60 (see Fig. 4). These data indicate that not only a fraction of the molecules is inactive, but that the activity of each active molecule is lowered, i.e. the movement of the rotor shaft is hampered. The above single molecule activity data do not immediately relate to bulk activity data as commonly observed. Single molecule experiments are highly selective as they focus on active enzymes only, while in bulk phase measurements an unknown fraction of inactive enzymes can distort the data. 26001275 Although these figures are different from those of the bulk measurement (Tab. 1) they corroborate the finding that the bulk portion of subunit c carries out ATP driven rotation despite of a rotor-stator cross-link in the mutants MM10, GH54, and FH4.DiscussionWe found that a cross-link between the top of the rotor (subunit c) and the stator ((ab)3) of F1 does not necessarily totally inhibit its ATP hydrolysis activity, but gradually reduces the rate up to fourfold (GH19), provided that the lock site on subunit c is notFigure 3. SDS-gel of the mutant GH54 under rotation assay conditions. Two samples of GH54 were oxidized (ox.) with 4 mM or 8 mM DTNB for 12 minutes, and afterwards re-reduced (re-red.) with 20 mM DTT for 12 minutes, to simulate the conditions in the rotation assay. doi:10.1371/journal.pone.0053754.gFigure 4. Rotary trajectories of reduced and oxidized F1 molecules. Trajectories of three active single molecules of GH54 driven by ATP hydrolysis both in the reduced (dashed line) and in the oxidized (dotted line) state, respectively. The mean trajectories for each of both states are shown by the solid lines. doi:10.1371/journal.pone.0053754.gUnfolding of Subunit Gamma in Rotary F-ATPasefarther than nine residues away from its C-terminal end. A crosslink at the penultimate residue of the C-terminal end (c285C, MM10) was even without any effect on the activity. In contrast, a cross-link of residues c262C (PP2, middle) or c87C (SW3, bottom) with the stator subunits practically extinguished the hydrolysis activity of F1. Three different lines of evidence support our observation. First, Peptide M site SDS-gels showed a cross-link yield of .85 . Second, bulk phase experiments revealed an activity of cross-linked mutants of at least 26 compared to wild type EF1 that could be restored after rereducing the samples. Third, rotation assay experiments support our conclusions on a single molecule level. Not only did we find single molecules still rotating despite oxidation, but furthermore was the rotational rate reduced by 60 , indicating that rotation was impaired by the cross-l.To rotate. Only these filaments were considered further. We checked whether enzymes, that stopped rotating after formation of a cross-link, were still active. We could reactivate oxidized enzymes by re-reducing the cysteines. In multiple cycles of reduction, oxidation, and re-reduction enzymes were active, inactive, and active again, respectively [24]. When changing to oxidizing conditions by addition of 4 mM DTNB (8 mM DTNB for FH4) for 10 minutes, 16 out of 29 (55 ) molecules of the mutant GH54 continued to rotate while 13 stopped, typically after 2 minutes. The rotation in the still active subset could be observed significantly longer (for at least 10 minutes). The observation time was limited only by the bleaching of the fluorescent dyes under oxidizing condition. For FH4 only 3 out of 11 (27 ) molecules remained active. In addition, in both cases the rotational rate decreased, typically by 60 (see Fig. 4). These data indicate that not only a fraction of the molecules is inactive, but that the activity of each active molecule is lowered, i.e. the movement of the rotor shaft is hampered. The above single molecule activity data do not immediately relate to bulk activity data as commonly observed. Single molecule experiments are highly selective as they focus on active enzymes only, while in bulk phase measurements an unknown fraction of inactive enzymes can distort the data. 26001275 Although these figures are different from those of the bulk measurement (Tab. 1) they corroborate the finding that the bulk portion of subunit c carries out ATP driven rotation despite of a rotor-stator cross-link in the mutants MM10, GH54, and FH4.DiscussionWe found that a cross-link between the top of the rotor (subunit c) and the stator ((ab)3) of F1 does not necessarily totally inhibit its ATP hydrolysis activity, but gradually reduces the rate up to fourfold (GH19), provided that the lock site on subunit c is notFigure 3. SDS-gel of the mutant GH54 under rotation assay conditions. Two samples of GH54 were oxidized (ox.) with 4 mM or 8 mM DTNB for 12 minutes, and afterwards re-reduced (re-red.) with 20 mM DTT for 12 minutes, to simulate the conditions in the rotation assay. doi:10.1371/journal.pone.0053754.gFigure 4. Rotary trajectories of reduced and oxidized F1 molecules. Trajectories of three active single molecules of GH54 driven by ATP hydrolysis both in the reduced (dashed line) and in the oxidized (dotted line) state, respectively. The mean trajectories for each of both states are shown by the solid lines. doi:10.1371/journal.pone.0053754.gUnfolding of Subunit Gamma in Rotary F-ATPasefarther than nine residues away from its C-terminal end. A crosslink at the penultimate residue of the C-terminal end (c285C, MM10) was even without any effect on the activity. In contrast, a cross-link of residues c262C (PP2, middle) or c87C (SW3, bottom) with the stator subunits practically extinguished the hydrolysis activity of F1. Three different lines of evidence support our observation. First, SDS-gels showed a cross-link yield of .85 . Second, bulk phase experiments revealed an activity of cross-linked mutants of at least 26 compared to wild type EF1 that could be restored after rereducing the samples. Third, rotation assay experiments support our conclusions on a single molecule level. Not only did we find single molecules still rotating despite oxidation, but furthermore was the rotational rate reduced by 60 , indicating that rotation was impaired by the cross-l.