Molecular Motors

Classification of Molecular Motors

• • Linear motors: Kinesin, dynein, myosin — walk along cytoskeletal filaments
• • Rotary motors: ATP synthase, bacterial flagellar motor
• • Polymerization motors: Actin polymerization, microtubule growth
• • Nucleic acid motors: RNA polymerase, helicase, topoisomerase

Mechanochemical Cycle

Kinesin-1 is a dimeric motor that walks processively along microtubules toward the plus end, taking 8 nm steps (the tubulin dimer repeat). Each step consumes one ATP. The hand-over-hand mechanism involves alternating catalytic cycles of the two heads.

Key experimental measurements:

• • Step size: $d = 8.2$ nm (Block lab, 1993)
• • Stall force: $F_s \approx 5\text{-}7$ pN
• • Velocity: $v_0 \approx 800$ nm/s (at saturating ATP)
• • Processivity: ~100 steps before detachment
• • One ATP per step (tight coupling)

Force-Velocity Relation

The force-velocity relation can be derived from a Kramers' rate theory model. The stepping rate under opposing force $F$ is:

where $\delta$ is the distance to the transition state along the reaction coordinate (~3-4 nm for kinesin). The velocity is:

At stall ($v = 0$ in the linear model), all the free energy from ATP hydrolysis is converted to work against the load: $F_s \cdot d = |\Delta G_{\text{ATP}}|$, giving a maximum efficiency of $\eta = F_s \cdot d / |\Delta G_{\text{ATP}}| \approx 50\text{-}60\%$.

The Cross-Bridge Cycle

Myosin II is the motor responsible for muscle contraction. Unlike kinesin, it is non-processive: a single myosin head detaches from actin after each power stroke. Force generation requires the collective action of hundreds of myosin heads in a thick filament.

The cross-bridge cycle (Lymn-Taylor model):

• • Step 1: Rigor state (M·A, strongly bound, no nucleotide)
• • Step 2: ATP binding $\to$ rapid detachment (M·ATP)
• • Step 3: ATP hydrolysis $\to$ reprimed head (M·ADP·Pi)
• • Step 4: Weak binding to actin, Pi release $\to$ power stroke
• • Step 5: ADP release $\to$ return to rigor state

Hill's Force-Velocity Equation for Muscle

A. V. Hill (1938) showed empirically that the force-velocity relation of whole muscle is a hyperbola:

where $F_0$ is the isometric (stall) force and $a, b$ are constants with $a/F_0 \approx b/v_{\max} \approx 0.25$. The maximum power output occurs at approximately $F \approx F_0/3$ and $v \approx v_{\max}/3$.

Structure and Mechanism

F$_1$F$_0$-ATP synthase is a rotary molecular motor that couples proton translocation across a membrane to ATP synthesis. It consists of:

• • F$_0$ (membrane): c-ring with $n = 8\text{-}15$ subunits, a-subunit with proton half-channels
• • F$_1$ (soluble): $\alpha_3\beta_3$ hexamer with 3 catalytic $\beta$ subunits, central $\gamma$-shaft

The stoichiometry is: $n$ protons per full rotation, 3 ATPs per rotation. Therefore:

For the mitochondrial enzyme with $n = 10$: 3.3 H$^+$/ATP. The thermodynamic efficiency approaches 100% under near-equilibrium conditions.

Torque and Energy Coupling

The torque generated by proton flow is:

where the proton motive force $\text{pmf} = \Delta\psi - (2.3\, RT/F)\Delta\text{pH}$is typically ~200 mV in mitochondria. The measured torque of the F$_1$ motor is ~40 pN·nm, consistent with the thermodynamic prediction.

The binding change mechanism (Boyer, Nobel Prize 1997) describes three catalytic states of the $\beta$ subunits that cycle through: open (O), loose (L), tight (T). Each 120° rotation of $\gamma$ advances all three$\beta$ subunits to their next state, synthesizing one ATP per 120° step.

Free Energy Landscape Derivation

The power stroke can be understood by considering the free energy as a function of both the mechanical coordinate $x$ (motor position) and the chemical state $\sigma$ (nucleotide state). The total free energy surface is:

For a two-state model (pre-stroke A and post-stroke B):

where $x_B - x_A = d$ (step size) and $\Delta G_{\text{chem}} = \Delta G_{\text{ATP}}$. The transition from A to B (power stroke) occurs when $G_B(x) < G_A(x)$ at the transition point, which is favorable when:

This is the thermodynamic stall condition. The maximum force a motor can generate is $F_{\text{stall}} = |\Delta G_{\text{ATP}}|/d$.

The Jarzynski Equality

Christopher Jarzynski (1997) proved a remarkable exact identity relating nonequilibrium work measurements to equilibrium free energy differences:

Derivation sketch: Consider a system driven from state A to state B by varying an external parameter $\lambda(t)$. The work done on the system in a single realization is:

The key identity follows from the ratio of the forward and equilibrium path probability densities in phase space, combined with Liouville's theorem for Hamiltonian dynamics. The result is exact regardless of how far from equilibrium the process is driven.

The second law follows as a corollary via Jensen's inequality:

This has been verified experimentally by pulling on single RNA hairpins with optical tweezers (Liphardt et al., Science 2002).

Crooks Fluctuation Theorem

The Crooks fluctuation theorem (1999) is a stronger result relating the forward and reverse work distributions:

The two distributions cross at $W = \Delta G$, providing a direct measurement of the free energy difference. The Jarzynski equality follows from integrating the Crooks relation over all work values.

Definitions and Physical Significance

The duty ratio $r$ is the fraction of the catalytic cycle during which a motor head is bound to its track:

For a single motor to be processive (walk multiple steps without falling off), $r > 0.5$ is necessary for a two-headed motor, ensuring at least one head is always bound.

The processivity (run length) is the average number of steps before detachment:

The probability of taking at least $N$ steps follows a geometric distribution:

Comparison of Motor Properties

• • Kinesin-1: $r \approx 0.5$, ~100 steps, cargo transport along microtubules
• • Myosin II: $r \approx 0.05$, non-processive, works in ensembles (muscle)
• • Myosin V: $r \approx 0.7$, ~50 steps, vesicle transport along actin
• • Dynein: $r \approx 0.7$, ~100 steps, retrograde transport along microtubules

Multiple Motors on a Single Cargo

In cells, cargo vesicles are typically carried by multiple motors simultaneously. For $N$ identical motors sharing a load $F$:

• • Stall force scales linearly: $F_s(N) = N \cdot F_s(1)$
• • Velocity is approximately independent of N at low loads
• • Run length increases super-linearly because detachment of one motor does not stop the cargo

The detachment rate for a team of $n$ engaged motors under force $F$ is:

where $F_d \approx 3$ pN is the force sensitivity of detachment. The competition between kinesin and dynein on the same cargo gives rise to bidirectional transport with characteristic switching between plus-end and minus-end directed motion.

Brownian Ratchet vs Power Stroke

Two limiting mechanisms for force generation by molecular motors:

Power stroke: A conformational change in the motor protein directly pushes against the load. The work is:

Brownian ratchet: Thermal fluctuations drive the motor forward, and the chemical cycle rectifies the motion (prevents backward steps). The key condition is that the forward rate exceeds the backward rate:

In reality, most motors use a combination: the power stroke provides the main forward displacement, while thermal fluctuations assist in completing the step and allow the motor to explore the free energy landscape.

The Smoluchowski Equation

The starting point for analyzing diffusive transport in a potential is the Smoluchowski equation (overdamped Fokker-Planck equation), which governs the probability density $P(x, t)$ of finding a Brownian particle at position $x$ at time $t$:

Here $D = k_BT/\gamma$ is the diffusion coefficient (Einstein relation),$\gamma$ is the friction coefficient, and $F(x) = -dU/dx$ is the force derived from the potential $U(x)$. This can be written compactly in terms of the probability current $J(x, t)$:

In thermal equilibrium with a static potential $U(x)$, the current vanishes ($J = 0$) and the solution is the Boltzmann distribution:

This is the fundamental no-go result: a static periodic potential, no matter how asymmetric, cannot produce directed motion. The equilibrium current is identically zero. To obtain directed transport, we must break detailed balance by introducing a non-equilibrium driving mechanism.

The Flashing Ratchet

Consider a periodic asymmetric (sawtooth) potential $U(x)$ with period $L$that is switched on and off with rates $k_{\text{on}}$ and $k_{\text{off}}$. The system alternates between two states:

• • ON state: Particle localizes at the potential minimum (Boltzmann distribution)
• • OFF state: Particle diffuses freely ($U = 0$), spreading symmetrically as a Gaussian

When the potential switches back on, the asymmetry of the sawtooth means that the particle is more likely to be captured by the next minimum to the right than to the left. We can compute this quantitatively. For a sawtooth with short tooth of length$a$ and long tooth of length $b = L - a$ (with $a < b$), and free diffusion time $\tau_{\text{off}}$, the probability of a net rightward step is:

Since $a < b$, we have $P_{\text{right}} > P_{\text{left}}$, giving a net rightward drift. The average velocity from the probability current in steady state is:

where $\tau_{\text{on}} = 1/k_{\text{off}}$ and $\tau_{\text{off}} = 1/k_{\text{on}}$are the mean dwell times in each state.

Thermodynamic Consistency: The Second Law is NOT Violated

A common misconception is that the Brownian ratchet extracts work from thermal fluctuations, violating the second law. This is not the case. The switching of the potential requires an external energy input. The work done on the system per switching cycle is:

where $x_{\text{on}}$ and $x_{\text{off}}$ denote the particle position at the instant of switching on and off, respectively. The useful work output per cycle is at most $W_{\text{out}} = F_{\text{load}} \cdot v \cdot (\tau_{\text{on}} + \tau_{\text{off}})$. The efficiency is:

The entropy production rate is always positive: $\dot{S}_{\text{tot}} = (W_{\text{input}} - W_{\text{out}})/T \geq 0$. The ratchet is a non-equilibrium machine that rectifies thermal fluctuations using an external energy source — fully consistent with the second law of thermodynamics.

General Steady-State Solution

For a more rigorous treatment, the steady-state probability current in a periodic potential with period $L$ can be computed exactly using the Stratonovich formula. With periodic boundary conditions $P(x+L) = P(x)$, the steady-state current is:

where $F_{\text{ext}}$ is any external applied force. The average velocity is then$v = JL$. This formula shows explicitly that $J = 0$ when $F_{\text{ext}} = 0$for a static potential — confirming that non-equilibrium driving (such as potential switching) is essential for directed motion.

Quantitative Comparison of Two Mechanisms

Molecular motors generate force through two fundamentally different strategies. Understanding the quantitative distinction between them is crucial for interpreting single-molecule experiments.

Power Stroke Mechanism

In the power stroke, a conformational change in the motor protein directly displaces the load by a distance $d$. The free energy landscape has two distinct states:

The conformational change from pre- to post-stroke moves the equilibrium position by$d$, doing work $W = Fd$ against an external load $F$. The motor stalls when all chemical energy is converted to work:

Examples: Kinesin ($d = 8$ nm,$|\Delta G| \approx 85$ pN·nm, $F_{\text{stall}} \approx 5\text{-}7$ pN) and myosin II ($d = 5.5$ nm, $F_{\text{stall}} \approx 2$ pN). The efficiency is $\eta = Fd/|\Delta G_{\text{chem}}|$, reaching 50-60% for kinesin.

Brownian Ratchet Mechanism

In the Brownian ratchet, the motor does not directly push the load. Instead, thermal fluctuations drive the motor forward, and an irreversible chemical step (e.g., monomer binding in polymerization) prevents backward motion. The chemical step acts as a pawl that rectifies Brownian motion.

For a polymerization motor (e.g., growing actin filament pushing a membrane), each monomer of size $\delta$ adds when thermal fluctuations create a gap of at least $\delta$ between the filament tip and the load. The probability of such a fluctuation against force $F$ is:

The polymerization velocity is therefore:

Setting $v = 0$ gives the stall force for the Brownian ratchet:

where $\Delta G_{\text{polym}} = k_BT\ln(k_{\text{on}}c/k_{\text{off}})$ is the free energy of polymerization per monomer. For actin: $\delta = 2.7$ nm,$F_{\text{stall}} \approx 1\text{-}2$ pN per filament.

Unified Stall Force Formula

Both mechanisms obey the same thermodynamic bound:

The difference lies in how the free energy is transduced:

• • Power stroke: $\Delta G$ drives a conformational change that deterministically moves the motor. The motor works against the load at every step. Operates efficiently even at high loads.
• • Brownian ratchet: $\Delta G$ biases an irreversible binding event that stochastically rectifies diffusion. The motor relies on thermal fluctuations to create gaps. Efficiency decreases rapidly with increasing load.
• • Distinguishing experimentally: In a power stroke motor, the position change occurs during the chemical transition. In a ratchet, the position change occursbefore the chemical transition (diffusion first, then capture).

Markov Model of the Mechanochemical Cycle

The kinesin mechanochemical cycle can be modeled as a discrete Markov process with $N$ chemical states per mechanical step. For the simplest two-state model, the states are:

• • State 1: Waiting for ATP binding (both heads bound, rear head nucleotide-free)
• • State 2: Post-ATP-binding, pre-step (ATP bound, hydrolysis and stepping occur)

The transition rates are:

where $k_{\text{on}}$ is the second-order ATP binding rate and $k_{\text{cat}}$is the first-order catalytic rate (encompasses hydrolysis, Pi release, and the mechanical step).

Michaelis-Menten-Like Velocity

The mean time to complete one full cycle is the sum of the mean dwell times in each state:

Since each cycle produces one step of size $d = 8$ nm, the velocity is:

Dividing numerator and denominator by $k_{\text{on}}$:

where $v_{\max} = d \cdot k_{\text{cat}}$ is the maximum velocity at saturating ATP and $K_M = k_{\text{cat}}/k_{\text{on}}$ is the Michaelis constant. This has the identical mathematical form to enzyme kinetics (Michaelis-Menten equation), which is not coincidental — kinesin is an enzyme (an ATPase) whose "product" is mechanical displacement.

Typical values for kinesin-1: $k_{\text{cat}} \approx 100$ s$^{-1}$,$k_{\text{on}} \approx 2$ $\mu$M$^{-1}$s$^{-1}$, giving $v_{\max} = 800$ nm/s and $K_M \approx 50$ $\mu$M.

Load-Dependent Velocity

An opposing force $F$ slows the motor by raising the free energy barrier for the mechanical step. Using Kramers' rate theory, the load-dependent catalytic rate is:

where $\delta$ is the distance from the initial state to the transition state along the reaction coordinate (typically $\delta \approx 3\text{-}4$ nm for kinesin, reflecting that the transition state is roughly halfway through the 8 nm step). Substituting into the Michaelis-Menten expression:

This expression captures two key experimental observations:

• • At low [ATP]: velocity is limited by ATP binding ($v \propto [\text{ATP}]$), and the force-dependence is weak (the rate-limiting step is chemical, not mechanical)
• • At saturating [ATP]: velocity is limited by $k_{\text{cat}}$, and the force-sensitivity is exponential with characteristic force $k_BT/\delta \approx 1$ pN

The stall force from this model is $F_s = (k_BT/\delta)\ln(k_{\text{cat}}/k_{\text{back}})$where $k_{\text{back}}$ is the backward stepping rate. This gives $F_s \approx 5\text{-}7$ pN, matching optical trap measurements.

Randomness Parameter

The randomness parameter $r$ quantifies the stochasticity of stepping:

For $N$ rate-limiting steps in the cycle, each exponentially distributed,$r = 1/N$. Experimentally, kinesin gives $r \approx 0.5$ at saturating ATP, indicating $N \approx 2$ rate-limiting steps — consistent with the two-state model (ATP binding + hydrolysis/stepping).

Torque from the Proton-Motive Force

The F$_0$ motor converts the electrochemical potential difference of protons across the membrane into rotary torque. Each proton translocated contributes energy:

where $\Delta\psi \approx 140\text{-}180$ mV is the membrane potential,$\Delta\text{pH} \approx 0.5\text{-}1.0$ in mitochondria, and the total pmf $\approx 180\text{-}220$ mV. The c-ring has $n$ identical subunits, each binding one proton. A full 360° rotation translocates $n$ protons, releasing total energy:

This energy drives the rotation of the $\gamma$-shaft. The work done per full rotation equals the torque times $2\pi$:

Therefore the torque generated by the F$_0$ motor is:

For the mitochondrial enzyme with $n = 10$ and pmf = 200 mV:

This is in excellent agreement with the measured torque of ~40-50 pN·nm from single-molecule rotation assays (Noji et al., 1997; Yasuda et al., 1998).

ATP Synthesis: 3 ATP per 360° Rotation

The F$_1$ sector contains three catalytic $\beta$ subunits arranged 120° apart around the central $\gamma$-shaft. Boyer's binding change mechanism (Nobel Prize, 1997) establishes that the three $\beta$ subunits cycle through three conformations:

• • Open (O): Low affinity — releases ATP, binds ADP + P$_i$
• • Loose (L): Moderate affinity — binds substrates loosely
• • Tight (T): High affinity — catalyzes ATP synthesis ($K_{\text{eq}} \approx 1$ on the enzyme!)

Each 120° rotation of $\gamma$ advances all three $\beta$ subunits: O $\to$ L $\to$ T $\to$ O. One ATP is synthesized per 120° substep, giving exactly 3 ATP per full 360° rotation.

The remarkable insight is that the catalytic step itself (forming ATP from ADP + P$_i$in the T site) is nearly isoenergetic ($\Delta G \approx 0$). The energy from the proton-motive force is used not for catalysis, but for product release — prying the tightly bound ATP out of the T site as it converts to the O conformation.

Thermodynamic Efficiency: Nature's Most Efficient Motor

The input energy per rotation comes from $n$ protons:

The output energy per rotation is the synthesis of 3 ATPs:

The thermodynamic efficiency is therefore:

Plugging in numbers for the mitochondrial enzyme ($n = 10$):

• • $|\Delta G_{\text{ATP}}| \approx 54$ kJ/mol (cellular conditions)
• • $\Delta\tilde{\mu}_{H^+} = F \cdot \text{pmf} = 96.485 \times 0.200 = 19.3$ kJ/mol
• • $W_{\text{in}} = 10 \times 19.3 = 193$ kJ per rotation
• • $W_{\text{out}} = 3 \times 54 = 162$ kJ per rotation

Under conditions closer to thermodynamic equilibrium (lower pmf, higher [ATP]/[ADP] ratio), the efficiency approaches $\eta \to 100\%$. Single-molecule experiments by Toyabe et al. (2011) measured efficiencies near 100% at low rotation speeds, confirming that F$_1$F$_0$-ATP synthase is nature's most efficient molecular motor.

The near-perfect efficiency arises because the motor operates close to thermodynamic reversibility: it can run in both directions (synthesis and hydrolysis) depending on the direction of proton flow, with minimal internal dissipation.

Huxley 1957 Cross-Bridge Model

A. F. Huxley's seminal 1957 model describes muscle contraction through the attachment and detachment kinetics of individual myosin cross-bridges to actin. Let $n(x, t)$be the fraction of cross-bridges attached at displacement $x$ from their equilibrium position. The rate equation is:

where $v$ is the filament sliding velocity, $f(x)$ is the attachment rate, and $g(x)$ is the detachment rate. Huxley proposed asymmetric rate functions:

• • Attachment: $f(x) = f_1 x/h$ for $0 \leq x \leq h$, $f(x) = 0$ otherwise
• • Detachment: $g(x) = g_1 x/h$ for $x \geq 0$, $g(x) = g_2$ for $x < 0$

where $h$ is the range of attachment ($\approx 10$ nm), and $g_2 \gg g_1$ensures rapid detachment when the cross-bridge is compressed (negative $x$), preventing the cross-bridge from acting as a brake.

Each attached cross-bridge acts as a linear spring with stiffness $\kappa$, generating force $f_{\text{cb}} = \kappa x$. The total force per half-sarcomere is:

where $m$ is the number of cross-bridges per half-sarcomere that can interact with a given actin site.

Steady-State Force-Velocity Relation

In steady state ($\partial n/\partial t = 0$), the rate equation becomes an ODE in $x$ parameterized by $v$:

For the Huxley rate functions, this ODE can be solved analytically in each region. The solution in the attachment zone ($0 \leq x \leq h$) gives:

Integrating to find the total force and simplifying, the force-velocity relation is well-approximated by the linear form:

where $F_0$ is the isometric force ($v = 0$) and $v_0$ is the maximum shortening velocity ($F = 0$). The isometric force is:

At zero velocity, cross-bridges accumulate at positive displacements, maximizing the average spring extension and hence the force.

Hill's Equation from Cross-Bridge Kinetics

The empirical Hill equation $(F + a)(v + b) = (F_0 + a)b$ arises naturally from the Huxley model when more realistic (nonlinear) rate functions are used. The hyperbolic form can be derived by noting that the cross-bridge cycling rate depends on velocity through the detachment rate. Rearranging:

The constants have physical meaning: $a$ is related to the energy dissipated during detachment, and $b$ is related to the maximum cross-bridge cycling rate. Experimentally, $a/F_0 \approx b/v_0 \approx 0.25$ for fast skeletal muscle.

Power Output Optimization

The mechanical power output is:

To find the maximum, we differentiate with respect to $F$ and set $dP/dF = 0$:

Simplifying, the optimal force is:

For $a/F_0 = 0.25$, this gives $F_{\text{opt}} \approx 0.31\, F_0 \approx F_0/3$. The corresponding optimal velocity is $v_{\text{opt}} \approx v_0/3$. The maximum power is:

This is a universal design principle: maximum power output occurs at approximately one-third of both the maximum force and the maximum velocity. This has been confirmed for skeletal muscle, cardiac muscle, and insect flight muscle.

Intracellular Transport

Kinesin and dynein transport vesicles, organelles, and mRNA along microtubule highways. Kinesin-1 carries cargo toward the cell periphery (plus-end directed), while cytoplasmic dynein carries cargo toward the cell center (minus-end directed). Disruption of motor-based transport is implicated in neurodegenerative diseases (Alzheimer's, Huntington's, ALS), where axonal transport over distances up to ~1 meter must be maintained reliably.

Muscle Contraction

Myosin II powers muscle contraction through the sliding filament mechanism. Each sarcomere contains ~300 myosin heads that work collectively. The non-processive nature of myosin II (duty ratio ~5%) is essential — it allows rapid detachment and reattachment, enabling high shortening velocities. Cardiac muscle myosin has a higher duty ratio (~10%), reflecting its sustained force requirements.

Bacterial Flagellar Motor

The bacterial flagellar motor is a rotary motor driven by the proton-motive force (or sodium-motive force in marine bacteria). It rotates at up to 1000 rev/s and can switch direction in milliseconds. The stator contains ~11 MotA/MotB complexes, each acting as an independent torque-generating unit. The motor produces ~1300 pN·nm of torque, vastly more than F$_1$F$_0$, making it the most powerful biological rotary motor.

Cell Division: The Mitotic Spindle

During cell division, the mitotic spindle uses multiple motor proteins to segregate chromosomes: kinesin-5 slides antiparallel microtubules apart to elongate the spindle, kinesin-13 depolymerizes microtubules at kinetochores to pull chromosomes poleward, and cytoplasmic dynein anchors spindle poles and generates pulling forces on astral microtubules. Errors in this motor-driven process lead to aneuploidy and cancer.

Synthetic Molecular Motors

Inspired by biological motors, chemists have designed synthetic molecular machines. The 2016 Nobel Prize in Chemistry was awarded to Sauvage, Stoddart, and Feringa for the design and synthesis of molecular machines, including light-driven rotary motors and molecular shuttles. Current challenges include achieving the efficiency, directionality, and processivity of biological motors in synthetic systems.

• 1938A. V. Hill publishes the hyperbolic force-velocity relation for muscle, winning the Nobel Prize (shared with Meyerhof, 1922) for work on heat production in muscle.
• 1954H. E. Huxley & J. Hanson / A. F. Huxley & R. Niedergerke independently propose the sliding filament theory of muscle contraction, based on X-ray diffraction and interference microscopy.
• 1957A. F. Huxley publishes the quantitative cross-bridge model with attachment/detachment rate equations, deriving the force-velocity relation from first principles. This remains the foundation of muscle biophysics.
• 1985Ronald Vale, Thomas Reese, & Michael Sheetz discover kinesin as a novel microtubule-based motor responsible for fast axonal transport. This opened the field of single-molecule motor biophysics.
• 1993Karel Svoboda & Steven Block use optical traps to measure single kinesin steps of 8 nm, establishing the single-molecule approach to motor mechanics.
• 1995Toshio Yanagida and colleagues achieve the first single-molecule measurements of myosin force and displacement using nanometer-precision techniques, revealing the stochastic nature of the power stroke.
• 1997Paul Boyer & John Walker share the Nobel Prize in Chemistry for elucidating the binding change mechanism of ATP synthase (Boyer) and determining its crystal structure (Walker). Jens Skou shared the prize for Na$^+$/K$^+$-ATPase.
• 1997Hiroyuki Noji and colleagues directly visualize the rotation of single F$_1$-ATPase molecules by attaching fluorescent actin filaments to the $\gamma$-subunit, providing dramatic confirmation of the rotary mechanism.
• 2003Ahmet Yildiz & Paul Selvin resolve the hand-over-hand stepping of kinesin and myosin V using FIONA (Fluorescence Imaging with One Nanometer Accuracy), definitively ruling out the "inchworm" model.