Terminus-2: ε Subunit CTD Role in F1-ATPase

• Terminus-2 is the C-terminal domain (residues 80–139) of the ε subunit, serving as a compact, mechanically active module in F1-ATPase.
• It switches between an 'up' inhibited state and a 'down' active state by sterically modulating rotor rotation to control ATP hydrolysis.
• Single-molecule FRET using an ABEL trap revealed that binding of AMPPNP shifts the CTD conformation and alters the binding affinity within the F1 complex.

Searching arXiv for the specified paper and closely related work on the ε-subunit regulatory switch in F1-ATPase. Terminus-2 denotes the C-terminal domain (CTD) of the $\epsilon$ subunit of E. coli $F_{1}$-ATPase, the soluble portion of the membrane-embedded enzyme $F_{o}F_{1}$-ATP synthase. In the account developed from single-molecule FRET measurements in an Anti-Brownian electrokinetic trap, this domain functions as a mechanical regulatory switch that modulates whether the $F_{1}$ motor is sterically blocked or rotationally permissive. The CTD spans roughly residues 80–139 of the 139-amino-acid $\epsilon$ subunit, and its conformational transition between an “up” inhibited state and a “down” active state provides a structural basis for minimizing non-productive ATP hydrolysis by $F_{1}$ (Bockenhauer et al., 2014).

1. Molecular definition and structural organization

The $\epsilon$ subunit is full length at 139 amino acids, with its C-terminal domain occupying approximately residues 80–139, or about 60 amino acids (Bockenhauer et al., 2014). This CTD is organized into two long antiparallel $\alpha$-helices, each about 20–25 amino acids in length, connected by a short loop. In structural terms, Terminus-2 therefore corresponds not to a disordered tail but to a compact conformational module capable of adopting distinct mechanically relevant arrangements.

Within the reported framework, the CTD is the specific part of $\epsilon$ that mediates inhibition of ATP hydrolysis. Its architecture enables either extension into the rotary core or retraction into a compact hairpin. This dual configurational capacity is central to the switch model because the same polypeptide segment can alternate between a sterically obstructive geometry and a geometry compatible with rotor motion.

A plausible implication is that the regulatory role of Terminus-2 derives less from catalytic chemistry than from geometric coupling to the motor’s central mechanical axis. The experimental design in the study was explicitly constructed to track this CTD rather than the $\epsilon$ subunit as a whole (Bockenhauer et al., 2014).

2. Conformational states and steric mechanism

In the “up” state, identified as the inhibited conformation, the two C-terminal helices extend upward and insert deeply into the central three-fold cleft between the $F_{1}$0 stator and the $F_{1}$1 rotor (Bockenhauer et al., 2014). This insertion sterically blocks rotation of the $F_{1}$2 subunit and thereby prevents ATP hydrolysis. In this description, Terminus-2 acts as a mechanical jam rather than as an allosteric modulator in the narrow sense.

In the “down” state, identified as the active conformation, the same C-terminal helices fold into a hairpin and retract alongside the N-terminal domain as the rotor turns. Under this configuration, the rotor is free to spin (Bockenhauer et al., 2014). The distinction between the two states is therefore both topological and functional: one inserts into the rotor–stator interface, whereas the other withdraws from it.

The paper’s mechanistic interpretation is explicitly two-state. In the absence of inhibitors, the CTD adopts the “up” conformation and prevents wasteful ATP hydrolysis. Binding of AMPPNP, which mimics the post-hydrolysis state, shifts a fraction of $F_{1}$3 into the “down” conformation, enabling rotation (Bockenhauer et al., 2014). This suggests that Terminus-2 is best understood as a nucleotide-sensitive steric gate coupled to the motor’s rotational coordinate.

3. Experimental interrogation by single-molecule FRET in the ABEL trap

The study examined Terminus-2 by site-specific fluorophore labeling and single-molecule FRET in the Anti-Brownian Electrokinetic (ABEL) Trap developed by Cohen and Moerner (Bockenhauer et al., 2014). Freely diffusing $F_{1}$4, with a diameter of about 10 nm, traverses a confocal volume in approximately 3 ms, which is too short to resolve the relevant conformational dynamics. The ABEL trap counteracts Brownian motion by real-time feedback voltages on the microsecond timescale applied across four platinum electrodes in a PDMS/glass microfluidic cell, extending observation times from less than 5 ms to a few times 100 ms, limited by photobleaching of Atto488 and Atto647N.

Excitation and detection conditions were specified as a CW 488 nm laser at an irradiance of about 1.2 kW/cm$F_{1}$5, with donor detection on an APD at 500–580 nm and acceptor detection on an APD at 630–785 nm (Bockenhauer et al., 2014). Typical count rates were approximately 11.5 kHz in the donor channel for $F_{1}$6-bound complexes and about 5 kHz in the acceptor channel. The measurements were performed in a “liposome buffer” containing 20 mM succinate, 20 mM tricine, 80 mM NaCl, 0.6 mM KCl, 2.5 mM MgCl$F_{1}$7, pH 8.0, selected to minimize surface sticking.

This methodology is significant because the regulatory motion of Terminus-2 is not readily accessible in conventional confocal transit experiments. The ABEL-trap implementation made it possible to observe single labeled $F_{1}$8 molecules in solution long enough to compare FRET behavior across biochemical conditions without immobilization artifacts being the primary experimental constraint.

4. Labeling geometry and FRET observables

The donor fluorophore was placed on an engineered $F_{1}$9-subunit mutation, $F_{o}F_{1}$0K108C, at position 108 on the single $F_{o}F_{1}$1-helix, and labeled with Atto488-maleimide (Bockenhauer et al., 2014). The acceptor fluorophore was placed on an engineered $F_{o}F_{1}$2-subunit mutation, $F_{o}F_{1}$3R99C, at position 99 on the first C-terminal $F_{o}F_{1}$4-helix, and labeled with Atto647N-maleimide. Each maleimide dye was coupled to the engineered cysteine through an approximately 1 nm thiol-reactive linker. Reported labeling efficiencies were about 55% for $F_{o}F_{1}$5K108C with Atto488 and about 30% for $F_{o}F_{1}$6R99C with Atto647N.

The distance interpretation was based on the Förster relation

$F_{o}F_{1}$7

where $F_{o}F_{1}$8 is the dye-to-dye distance and $F_{o}F_{1}$9 nm for the Atto488$F_{1}$0Atto647N pair (Bockenhauer et al., 2014). From the crystal structure, the expected $F_{1}$1 distance in the “up” state is approximately 3 nm, corresponding to $F_{1}$2, whereas the “down” state distance is approximately 6 nm, corresponding to $F_{1}$3.

State of Terminus-2	$F_{1}$4 distance	Theoretical FRET efficiency
“Up” (inhibited)	$F_{1}$5 nm	$F_{1}$6
“Down” (active)	$F_{1}$7 nm	$F_{1}$8

These values provide the structural calibration for the switch model. In ensemble cuvette spectra, mixing $F_{1}$9-Atto488 with $\epsilon$0-Atto647N at 15–20 nM produced acceptor emission at 661 nm that was about 0.16 times the donor peak at 520 nm. Addition of 1 mM AMPPNP caused this ratio to decrease progressively to about 0.10 over 15–30 min (Bockenhauer et al., 2014). In the single-molecule measurements, the reported observable was a “proximity factor,” not a fully corrected FRET efficiency, and the principal histograms were therefore discussed in terms of peaks in $\epsilon$1 rather than direct assignment of absolute $\epsilon$2 values.

5. Nucleotide dependence, affinity shifts, and thermodynamic description

Under single-molecule conditions analyzed by change-point analysis with 10 ms bins, two biochemical regimes were especially prominent (Bockenhauer et al., 2014). In the absence of nucleotide, or in the presence of 1 mM Mg$\epsilon$3ATP, the proximity-factor histogram displayed a single dominant peak at $\epsilon$4. In the presence of 1 mM Mg$\epsilon$5AMPPNP, the distribution partially redistributed toward lower states at $\epsilon$6–0.3 and a very high state at $\epsilon$7, while the $\epsilon$8 population remained diminished but still present.

The binding thermodynamics of $\epsilon$9 to $F_{1}$0, as quoted from reference [56] in the paper, showed strong nucleotide dependence. The dissociation constant was approximately $F_{1}$1 nM in the absence of nucleotide and approximately $F_{1}$2 nM in the presence of AMPPNP, representing about a 60-fold weakening (Bockenhauer et al., 2014). The corresponding off-rate increased by about 80-fold with AMPPNP, implying that $F_{1}$3 is more likely to dissociate when locked in the active conformation.

The free-energy change of binding was expressed as

$F_{1}$4

At 298 K, the paper summary gives $F_{1}$5 kJ/mol for $F_{1}$6 nM and $F_{1}$7 kJ/mol for $F_{1}$8 nM (Bockenhauer et al., 2014). These values formalize the observation that nucleotide analog binding changes not only conformational occupancy but also the stability of the $F_{1}$9–$\epsilon$0 complex itself.

A common misconception would be to interpret AMPPNP simply as switching all complexes into a single active state. The reported data do not support that simplification. Instead, AMPPNP produces only a partial redistribution, and a substantial $\epsilon$1 population remains under ABEL-trap conditions (Bockenhauer et al., 2014).

6. Mechanistic interpretation and experimental limits

The study’s overall interpretation is that Terminus-2 operates as a two-state, thermodynamically driven switch between an inhibitory “up” state and an active “down” state, with both equilibrium occupancy and kinetics modulated by nucleotide-dependent changes in $\epsilon$2–$\epsilon$3 affinity (Bockenhauer et al., 2014). In the absence of inhibitors, the “up” state sterically jams the $\epsilon$4 rotor and suppresses wasteful ATP hydrolysis. AMPPNP shifts part of the population toward the “down” state, which permits rotation and is associated with much weaker binding, so that $\epsilon$5 can dissociate and allow catalysis.

At the same time, the ABEL-trap measurements showed that under conditions with nanomolar concentrations of free $\epsilon$6 and $\epsilon$7, the $\epsilon$8 subunit can re-bind and re-adopt the “up” state even in the presence of AMPPNP (Bockenhauer et al., 2014). This explains why the $\epsilon$9 population remained large rather than disappearing. The qualitative energy-landscape description given in the summary is therefore one in which nucleotide binding lowers the barrier and alters $\alpha$0 so that the “down” basin becomes more populated, but re-binding prevents wholesale activation unless $\alpha$1 eventually dissociates.

The experimental limits were also explicit. Single-molecule transition rates within a trap were too rare—only a few events per 100 ms—to extract reliable $\alpha$2 or $\alpha$3 values (Bockenhauer et al., 2014). Accordingly, the work establishes state occupancies, structural assignments, and nucleotide-linked affinity changes more robustly than microscopic switching kinetics. This suggests that Terminus-2 is well supported as a mechanical regulatory element, while the detailed dynamical pathway between its conformers remained unresolved under the reported observation window.