Molecular motors: kinesin, dynein, and myosin mechanics

Intuition Beginner

Picture a delivery worker who walks a tight rail and never lets go with both hands at once. A molecular motor is a protein machine that does this inside a cell. It has a track (a filament), a fuel (ATP), and a body that turns one fuel molecule into one fixed step. There are three families. Myosin walks on thin actin filaments. Kinesin and dynein walk on hollow microtubule tubes — kinesin toward the growing plus end, dynein the other way. The survey of these motors and their tracks is in the companion unit 17.12.01; here we open the bonnet and watch the wheels turn.

Each motor is an ATPase: it grabs ATP, splits it into ADP and phosphate, and releases the pieces, and the energy of that reaction drives a shape change in the motor head. The chemical cycle and the mechanical stride are locked together — this chemomechanical coupling is the whole trick. Burn one ATP, take one step. Kinesin-1 takes 8-nanometre steps, one ATP each. Myosin V takes giant 36-nanometre strides on actin. Dynein's step is irregular, sometimes 8, sometimes 32 nanometres. The three motors solve the same problem with different moving parts.

These tiny walkers do the cell's hauling. Kinesin drags vesicles and mitochondria outward from the cell centre. Dynein hauls them back. In muscle, armies of myosin II pull on actin to shorten the cell; as a lone cargo carrier, myosin V walks a single organelle along actin. In cilia and flagella — the waving hairs that sweep the lung and propel sperm — arrays of dynein slide neighbouring microtubules past each other to bend the whole whip. One chemistry, three architectures, every kind of movement.

Visual Beginner

The three motor families differ in their track, their direction, their step size, and how much of each fuel cycle they spend gripping the track. That last quantity — the duty ratio — is what separates a muscle motor (built for relay in large teams) from a cargo motor (built to walk alone).

Motor	Track	Direction	Step size	Duty ratio	Typical role
Myosin II	Actin	Plus end	5–10 nm	~0.05 (low)	Muscle shortening; many motors in relay
Myosin V	Actin	Plus end	36 nm	~0.7 (high)	Processive vesicle/cargo transport
Kinesin-1	Microtubule	Plus end	8 nm	~0.5 (high)	Long-range axonal and cargo transport
Cytoplasmic dynein	Microtubule	Minus end	8–32 nm (variable)	~0.4	Minus-end cargo haul; mitotic spindle
Axonemal dynein	Microtubule (axoneme)	Slides doublets	Variable	—	Ciliary and flagellar beating

Worked example Beginner

Three numbers fix how a motor is tuned for its job: the stride length, the duty ratio, and how long a lone motor stays on its track.

Step 1 — a giant stride. Myosin V walks on actin, whose twisted ladder repeats every $36\text{nm}$ (13 subunits in 6 turns). Myosin V's two heads span exactly this distance: one head holds while the other swings forward by $36\text{nm}$ and lands on the next matching site. So each stride is $36\text{nm}$. At about $5$ strides per second, a single myosin V moves at $36\times 5=180\text{nm/s}$.

Step 2 — how long a head holds on. A myosin II head passes through one ATP cycle in about $40$ milliseconds, but the power stroke — the part where it grips actin and pulls — lasts only about $2$ milliseconds. The fraction of time spent gripping is the duty ratio: $2/40=0.05$, or $5\%$. A lone myosin II head is detached $95\%$ of the time. Muscle works because hundreds of heads act in relay, each grabbing in turn, so the actin is never released.

Step 3 — a long lone run. Kinesin-1 takes $8$-nm steps and keeps going for about $100$ steps on average before falling off. Total distance: $8\times 100=800\text{nm}=0.8\mu \text{m}$. Unlike myosin II, kinesin is processive: its two heads are gated so one is always attached. That is what lets a single motor carry one vesicle a useful distance without letting go.

What this tells us: motors are tuned for their task — muscle myosin trades grip for speed by working in huge teams, while kinesin and myosin V trade speed for the ability to walk alone.

Check your understanding Beginner

Formal definition Intermediate+

A molecular motor is an ATPase enzyme that couples a cyclic nucleotide chemistry — ATP binding, hydrolysis to ADP and inorganic phosphate, then product release — to a directed displacement along a polar filament. The companion unit 17.12.01 defines the three filament systems and the cross-bridge cycle; here the object of study is the comparative mechanics of the motor itself: how the chemistry is locked to the stride, how many heads a motor needs to walk alone, and how the stride changes under load.

The load-bearing quantity is the duty ratio. Let ${\tau }_{\mathrm{c}}$ be the total duration of one ATPase cycle (from one ATP binding to the next) and ${\tau }_{\mathrm{on}}$ the time within that cycle during which the head is strongly bound to the filament and generating or bearing force. The duty ratio is

$$r=\frac{{\tau }_{\mathrm{on}}}{{\tau }_{\mathrm{c}}}=\frac{{\tau }_{\mathrm{on}}}{{\tau }_{\mathrm{on}}+{\tau }_{\mathrm{off}}},$$

where ${\tau }_{\mathrm{off}}={\tau }_{\mathrm{c}}-{\tau }_{\mathrm{on}}$ is the weakly bound or detached portion of the cycle. A motor with duty ratio $r$ spends fraction $r$ of its time attached. Muscle myosin II has $r\approx 0.05$; kinesin-1 and myosin V have $r\approx 0.5$–$0.7$.

Processivity is the ability to take many steps before dissociating, achieved only when a motor carries two or more heads gated so that at least one is bound at every instant. A motor with $N$ independently acting heads each of duty ratio $r$ keeps the track attached in expectation when $Nr\ge 1$; the minimal head count for processive walking is $N_{\min }=\lceil 1/r\rceil$. This single inequality is the structural reason a lone myosin II (with $r\approx 0.05$, $N_{\min }=20$) cannot walk alone while a kinesin dimer ($N=2$, $r\approx 0.5$) can. The run length $L$ is the mean distance travelled before detachment; the stall force $F_s$ is the largest opposing load the motor can sustain without sliding backward; the step size $d$ is the centre-of-mass displacement per ATP (kinesin $d=8\text{nm}$, myosin V $d\approx 36\text{nm}$).

The three families solve the coupling problem with three different moving parts. Kinesin-1 is a homodimer of P-loop NTPase motor domains; a short peptide called the neck linker lies along the motor domain and docks (zippers onto the motor core) when ATP binds the leading head, throwing the tethered partner head forward. Myosin uses a rigid lever arm — a stiff α-helical extension stabilised by calmodulin-like light chains (IQ motifs) — whose tilt sets the stroke length: myosin II carries two short lever arms (small $d$, low $r$), myosin V carries two long six-IQ lever arms (large $d\approx 36\text{nm}$, high $r$). Dynein is a $1$–$2\text{MDa}$ homodimer built around a ring of six AAA+ domains: ATP hydrolysis at the AAA1 site swings a linker element across the ring face, producing a power stroke toward the microtubule minus end. One chemistry, three architectures.

Key mechanism Intermediate+

The central mechanism of this unit is chemomechanical coupling through a nucleotide-gated conformational stroke, read out as a discrete stepping pattern on a polar filament. State it for kinesin, where the stepping geometry is cleanest, then contrast with the myosin lever arm and the dynein ring.

Mechanism (kinesin neck-linker gating and hand-over-hand stepping). Kinesin-1 advances in $8$-nm centre-of-mass steps, one ATP per step, by a hand-over-hand cycle in which each head alternately detaches, swings forward $16\text{nm}$ past the bound partner, and rebinds.

1. Both heads are bound to successive tubulin dimers $8\text{nm}$ apart on one protofilament; the leading head carries ATP and the trailing head carries ADP (or nucleotide-free). Inter-head strain through the stalk biases the trailing head to release its nucleotide slowly.
2. ATP binds the leading head; the leading neck linker docks forward along its motor domain, dragging the tethered rear head forward through $16\text{nm}$ of diffusion to the next binding site, $8\text{nm}$ beyond the still-bound head.
3. The now-leading head binds tubulin and releases ADP; the trailing head hydrolyses its ATP and releases phosphate, weakening its grip.
4. The trailing head detaches (now carrying ADP), the strain relieves, and the cycle repeats with roles exchanged. One ATP was consumed; the centre of mass advanced $8\text{nm}$.

Hand-over-hand versus inchworm. Two models were proposed for how the two heads move. In the inchworm model the heads keep their leading/trailing order and each shifts $8\text{nm}$ per step. In the hand-over-hand model the heads alternate: each moves $16\text{nm}$ per step while the centre of mass advances $8\text{nm}$. The two predict identical centre-of-mass stepping but different single-head trajectories. Yildiz and colleagues settled this by attaching a single fluorophore to one head and localising it to nanometre precision: the labelled head alternated between positions $0$ and $17\text{nm}$, with no intermediate dwells at $8\text{nm}$ [ref: TODO_REF Yildiz2004]. The result is hand-over-hand. (Myosin V, by the same assay, also walks hand-over-hand, each head swinging $72\text{nm}$ while the cargo advances $36\text{nm}$ [ref: TODO_REF Mehta1999].)

The myosin lever arm and the dynein ring. Myosin transduces hydrolysis through a tilt of the lever arm rather than a neck-linker zipper: phosphate release on actin-bound myosin releases the lever from its cocked to its relaxed angle, a rotation amplified by the lever-arm length ($\sim 3\text{nm}$ per IQ motif, giving myosin V its $36$-nm stride). Dynein, structurally unrelated, hydrolyses ATP in its AAA1 ring; the linker swings across the ring face by $\sim 15\text{nm}$ toward the microtubule, and — uniquely among the three — the step size and even the direction can be modulated by the dynactin complex and cargo adaptors, producing the variable $8$–$32\text{nm}$ strides of cytoplasmic dynein [ref: TODO_REF Reck-Peterson2006].

The force-velocity curve. An optical trap clamps a motor against a fixed load $F$ and records the stepping rate. For kinesin-1 the curve is close to linear from $v_0\approx 800\text{nm/s}$ at zero load down to zero at the stall force $F_s\approx 6$–$7\text{pN}$; above stall the motor steps backward, and the backward rate grows with load [ref: TODO_REF SchnitzerBlock1997] [ref: TODO_REF Visscher1999]. Load acts on the rates through the Bell model: a transition whose transition state lies a distance $\delta$ along the reaction coordinate has rate

$$k(F)=k_0\exp \left(\frac{F\delta }{k_{\mathrm{B}}T}\right),$$

accelerated when load favours the transition and slowed when it opposes it. With forward and backward stepping distances ${\delta }_{+}$, ${\delta }_{-}$ this single exponential predicts both the linear force-velocity regime and the stall point where forward and backward rates balance.

Bridge. The chemomechanical stroke is the foundational reason a single protein turns chemical energy into directed work: the nucleotide state gates the conformation, and the conformation gates the filament interface, so that one ATP buys one stride of fixed size and direction. This builds toward the mesoscale mechanics of muscle and cilia, where ensembles and arrays of these same heads are coordinated, and the ATPase stepping appears again in the force balance of the mitotic spindle 17.08.01 and in kinesin-driven axonal transport 17.09.01. The duty ratio is dual to the motor's architecture: a short lever arm and weak gating give the low duty ratio of muscle myosin, while neck-linker and lever-arm gating give kinesin and myosin V their high duty ratios. Putting these together, the bridge is from the ATP chemistry of 17.04.02 to the contractility of muscle 17.03.02.

Exercises Intermediate+

Advanced results Master

The Intermediate tier presented each motor as a clean cycle producing a fixed stride. The Master picture is that the stride is the output of a load-dependent stochastic machine, that the same hydrolysis chemistry supports two opposed engineering strategies — ensemble speed versus processive solo walking — and that the dynein ring generalises the motor concept to a variable-stride, directionally switchable ATPase.

Force-velocity curves and the load-dependent chemomechanical cycle. The kinesin-1 force-velocity curve, measured under a molecular force clamp by Visscher, Schnitzer, and Block [ref: TODO_REF Visscher1999], is approximately linear from $v_0\approx 800\text{nm/s}$ at zero load to $v=0$ at $F_s\approx 6$–$7\text{pN}$, with a convex downturn near stall and robust backward stepping above it. The slope is set by two load-dependent transitions: ATP binding to the leading head (opposed by load through the distance to its transition state, ${\delta }_{+}\approx 0.8$–$1.0\text{nm}$) and the load-favoured backward step from a one-head-bound intermediate. With forward and backward stepping rates $k_{+}(F)=k_{+}^0{e}^{-F{\delta }_{+}/k_{\mathrm{B}}T}$ and $k_{-}(F)=k_{-}^0{e}^{F{\delta }_{-}/k_{\mathrm{B}}T}$, the velocity $v(F)=d[k_{+}(F)-k_{-}(F)]$ reproduces the measured curve and yields the stall force $F_s=(k_{\mathrm{B}}T/({\delta }_{+}+{\delta }_{-}))\ln (k_{+}^0/k_{-}^0)$, of order a few pN for kinesin's ${\delta }_{\pm }$ and $k^0$ values [ref: TODO_REF SchnitzerBlock1997]. Myosin V and cytoplasmic dynein obey the same formalism but with different characteristic loads: myosin V stalls near $1.5$–$3\text{pN}$ [ref: TODO_REF Mehta1999], and dynein, with its variable step, shows a force-velocity curve that is markedly non-monotonic and load-direction-dependent [ref: TODO_REF Reck-Peterson2006].

Myosin II versus myosin V: two strategies from one stroke. The myosin superfamily repurposes the same lever-arm tilt for opposite engineering goals. Myosin II has a short lever arm (two IQ motifs), a small stroke ($\sim 5$–$10\text{nm}$), and a duty ratio $r\approx 0.05$: each head spends most of its cycle detached, so a single molecule cannot walk, but hundreds of heads arrayed on a thick filament act in relay, each grabbing actin in turn, to sustain continuous contraction. This is the muscle design — speed and collective force through numbers. Myosin V has a long lever arm (six IQ motifs per head, $\sim 24\text{nm}$), a large $36$-nm stride that matches the $36$-nm pseudo-repeat of the actin helix so the two heads land on stereochemically equivalent sites, and a duty ratio $r\approx 0.7$ that makes the dimer processive. The lever-arm length is not incidental: it is tuned to the actin helical repeat, and a single IQ-motif deletion collapses the stride and the processivity [ref: TODO_REF Spudich2001]. The two motors are the clean illustration that $r$, $d$, and $N$ are not free parameters but a coordinated design triple.

Dynein: a ring ATPase with variable stride and regulated direction. Cytoplasmic dynein is structurally unrelated to kinesin and myosin: a $\sim 500\text{kDa}$ heavy chain folds into a ring of six AAA+ domains, with the microtubule-binding stalk projecting from one side and a linker spanning the ring face. Hydrolysis at AAA1 swings the linker by $\sim 15\text{nm}$, producing the minus-end-directed power stroke [ref: TODO_REF Reck-Peterson2006]. Unlike kinesin's fixed $8$-nm stride, dynein steps are $8$, $16$, $24$, even $32\text{nm}$, sometimes sideways between protofilaments, and occasional plus-end steps occur. Directionality and processivity are set not by the motor alone but by the dynactin complex and cargo adaptors (Bicaudal-D, Lis1, NudE), which recruit two dynein heavy chains into a processive, minus-end-locked complex — an instance of a motor whose stepping geometry is externally regulated.

Axonemal dynein and ciliary beating. The ciliary and flagellar axoneme is a cylinder of nine outer microtubule doublets surrounding a central pair (the "$9+2$" arrangement). Outer and inner dynein arms sit on the A-tubule of each doublet and reach toward the B-tubule of the next; their ATP-driven, baseward walking slides neighbouring doublets axially. In a free pair this would telescope the doublets apart; in the intact axoneme the nexin links and radial spokes forbid telescoping, so the sliding is converted to bending. Summers and Gibbons demonstrated the sliding mechanism directly by digesting the radial spokes and nexin links with trypsin and watching ATP cause the doublets to extrude [ref: TODO_REF SummersGibbons1971], a result that established that the axoneme is an array of motors whose coordinated, geometry-gated sliding produces the macroscopic beat.

Synthesis. The three motor families are the foundational reason the cytoskeleton is a machine rather than a scaffold: each couples ATP hydrolysis to a nucleotide-gated conformational stroke on a polar track, and this is exactly the coupling that kinesin, myosin, and dynein share despite their unrelated folds. The hand-over-hand stepping of kinesin and myosin V generalises to any two-headed gated motor, while dynein's ring ATPase shows that a variable-stride machine can still be processive when its directionality is set by cargo adaptors. The muscle-versus-cargo distinction is dual to the duty ratio: myosin II trades processivity for ensemble speed, myosin V does the reverse, and the lever-arm length is tuned to the actin repeat to make the high duty ratio geometrically viable. The central insight is that directionality lives in the polar track and is read out by a stroke whose timing and length are set by the nucleotide cycle; putting these together, the bridge is from single-molecule stepping to the contractility of muscle 17.03.02 and the force balance of the spindle 17.08.01, and the same cycle appears again in axonal cargo delivery 17.09.01 and the optical-trap methods that first resolved it 17.11.01.

Full proof set Master

Proposition (duty-ratio bound on processive head count). Consider a motor carrying $N$ identical heads that act independently, each with duty ratio $r$ (the fraction of its ATPase cycle spent strongly bound to the filament). Then the expected number of simultaneously bound heads is $Nr$, and a necessary condition for sustained track attachment — processive walking in expectation — is $Nr\ge 1$. The minimal head count for processivity is $N_{\min }=\lceil 1/r\rceil$, and the probability that all heads are simultaneously detached is $(1-r{)}^N\le e^{-Nr}$.

Proof. Each head is strongly bound with probability $r$ and detached or weakly bound with probability $1-r$, independently of the others. Let $X$ be the number of bound heads; then $X$ follows a binomial law, $X\sim \mathrm{Bin}(N,r)$, with expectation $E[X]=Nr$. The motor loses its grip on the track precisely when $X=0$, an event of probability

$$\mathrm{Pr}(X=0)=(1-r{)}^N.$$

For the motor to remain attached in expectation at every instant, at least one head must be bound on average, requiring $E[X]\ge 1$, i.e. $Nr\ge 1$. The minimal integer head count satisfying this is $N_{\min }=\lceil 1/r\rceil$. The detachment probability satisfies the standard exponential bound $(1-r{)}^N\le e^{-Nr}$ (since $1-r\le e^{-r}$), so to make detachment rare one wants $Nr\gg 1$: doubling $Nr$ squares the detachment probability. For muscle myosin II, $r\approx 0.05$ gives $N_{\min }=20$, matching the thick-filament design of many heads per actin track; for kinesin-1, $r\approx 0.5$ gives $N_{\min }=2$, exactly the dimer geometry, and $Nr=1$ predicts a fragile bound, consistent with the observed sensitivity of kinesin processivity to mutations that perturb inter-head gating. $\square$

Proposition (force-velocity and stall from load-dependent rates). Let a motor step forward by $d$ with load-dependent rate $k_{+}(F)=k_{+}^0\exp (-F{\delta }_{+}/k_{\mathrm{B}}T)$ and backward with rate $k_{-}(F)=k_{-}^0\exp (F{\delta }_{-}/k_{\mathrm{B}}T)$, where ${\delta }_{+},{\delta }_{-}>0$ are the distances to the respective transition states and $k_{+}^0>k_{-}^0$. The velocity is $v(F)=d[k_{+}(F)-k_{-}(F)]$, and the stall force at which $v=0$ is

$$F_s=\frac{k_{\mathrm{B}}T}{{\delta }_{+}+{\delta }_{-}}\ln \frac{k_{+}^0}{k_{-}^0}.$$

Proof. The motor advances by $d$ on each forward step and retreats by $d$ on each backward step, so by the definition of the rates as transitions per unit time the net velocity is $v(F)=d{k}_{+}(F)-d{k}_{-}(F)=d[k_{+}(F)-k_{-}(F)]$. The stall force is the load at which forward and backward stepping balance, $v(F_s)=0$, equivalently $k_{+}(F_s)=k_{-}(F_s)$. Substituting the Bell-model rates,

$$k_{+}^0\exp \left(-\frac{F_s{\delta }_{+}}{k_{\mathrm{B}}T}\right)=k_{-}^0\exp \left(\frac{F_s{\delta }_{-}}{k_{\mathrm{B}}T}\right).$$

Taking logarithms, $\ln k_{+}^0-F_s{\delta }_{+}/(k_{\mathrm{B}}T)=\ln k_{-}^0+F_s{\delta }_{-}/(k_{\mathrm{B}}T)$, and collecting the terms in $F_s$,

$$\frac{F_s({\delta }_{+}+{\delta }_{-})}{k_{\mathrm{B}}T}=\ln \frac{k_{+}^0}{k_{-}^0},F_s=\frac{k_{\mathrm{B}}T}{{\delta }_{+}+{\delta }_{-}}\ln \frac{k_{+}^0}{k_{-}^0}.$$

Because $k_{+}^0>k_{-}^0$ the logarithm is positive, so $F_s>0$: the motor's own kinetic asymmetry generates a finite stall force even before any structural detail is specified. For kinesin-1, with ${\delta }_{+}\approx {\delta }_{-}\approx 1\text{nm}$ and $k_{+}^0/k_{-}^0$ of order $10^2$–$10^3$, this gives $F_s\approx 4$–$7\text{pN}$ at room temperature, in the measured range [ref: TODO_REF SchnitzerBlock1997]. Expanding $v(F)$ around $F=0$ recovers the linear regime $v(F)\approx v_0(1-F/F_s)$ with $v_0=d(k_{+}^0-k_{-}^0)$, matching the nearly linear portion of the optical-trap force-velocity curve [ref: TODO_REF Visscher1999]. $\square$

Proposition (load-dependent run length). Suppose a motor of step size $d$ steps forward at rate $k_{\mathrm{step}}(F)$ and enters its one-head-bound vulnerable state with probability $ϵ$ per step, detaching from that state at the Bell-model rate $k_{\mathrm{off}}(F)=k_{\mathrm{off}}^0\exp (F{\delta }_{\mathrm{d}}/k_{\mathrm{B}}T)$. Then the run length is

$$L(F)=\frac{d{k}_{\mathrm{step}}(F)}{ϵk_{\mathrm{off}}(F)},$$

and it falls off exponentially with opposing load $F$ through the denominator.

Proof. The motor detaches only from the vulnerable one-head-bound state, which it reaches with probability $ϵ$ after each completed step, i.e. at rate $ϵk_{\mathrm{step}}(F)$ while it is stepping. Detachment from that state occurs at rate $k_{\mathrm{off}}(F)$. In a competing-rates picture the mean number of steps before detachment is the ratio of the stepping rate to the rate at which stepping terminates, $N_{\mathrm{steps}}=k_{\mathrm{step}}(F)/[ϵk_{\mathrm{off}}(F)]$, and the mean distance is this step count times the step size, $L(F)=d{N}_{\mathrm{steps}}=d{k}_{\mathrm{step}}(F)/[ϵk_{\mathrm{off}}(F)]$. Substituting the Bell-model rate gives the load dependence explicitly:

$$L(F)=\frac{d{k}_{\mathrm{step}}(F)}{ϵk_{\mathrm{off}}^0}\exp \left(-\frac{F{\delta }_{\mathrm{d}}}{k_{\mathrm{B}}T}\right),$$

so the run length decays exponentially as the opposing load grows, consistent with the observed shortening of kinesin's run length from $\sim 1\mu \text{m}$ at zero load to a few hundred nanometres near stall [ref: TODO_REF Visscher1999]. $\square$

These three results fix the quantitative spine of the unit: the first ties processivity to the duty ratio and head count, the second derives the force-velocity curve and stall force from the load-dependent kinetics, and the third links the microscopic detachment transition state to the mesoscopic run length an optical trap measures.

Connections Master

• Cytoskeleton, molecular motors, and cell motility 17.12.01 is the prerequisite survey: it introduces the three filament systems, the actomyosin cross-bridge cycle, and the cell-migration machinery. This unit descends one level into the comparative mechanics of the three motor families — the chemomechanical coupling, the duty ratio, the hand-over-hand stepping geometry, and the force-velocity curves that the survey names but does not derive.

• Cytoskeleton and contractile proteins 17.03.02 is the muscle sibling: it treats the ensemble actomyosin cross-bridge mechanics and the sliding-filament model at the tissue level. The myosin II analysed there is the low-duty-ratio ensemble motor whose single-molecule stepping is the present unit's concern, and the lever-arm and duty-ratio framework derived here is what makes the muscle thick-filament design intelligible.

• Oxidative phosphorylation and ATP synthesis 17.04.02 supplies the fuel: the ATP whose hydrolysis free energy $\Delta {\mu }_{\text{ATP}}\approx 80$–$100\text{pN}\cdot \text{nm}$ powers every step. The stall-force bound and the efficiency $\eta =F_{s}d/\Delta {\mu }_{\text{ATP}}$ computed here depend directly on the ATP/ADP ratio maintained by the chemiosmotic machinery analysed there.

• The cell cycle 17.08.01 consumes these motors wholesale: mitotic kinesins (Eg5, CENP-E) and dynein generate and balance the forces that assemble the spindle, bi-orient chromosomes, and elongate the anaphase spindle, and the load-dependent stepping analysed here is what lets the spindle respond as a force-balanced mechanical structure.

• Cellular neuroscience 17.09.01 depends on kinesin and dynein for axonal transport: kinesin carries synaptic-vesicle precursors and membrane proteins down the axon to the synapse, and dynein returns them, over distances (up to a metre in a human motor neuron) that only processive high-duty-ratio motors can sustain.

• Cell and molecular biology methods 17.11.01 supplies the instruments that produced nearly every number here: the optical trap that clamped kinesin against a fixed load and measured its $8$-nm step and force-velocity curve, and the single-fluorophore localisation that resolved hand-over-hand stepping.

Historical & philosophical context Master

The motor concept was born twice, once in muscle and once in the axon. Hugh Huxley's 1969 sliding-filament and cross-bridge model [ref: TODO_REF Huxley1969] synthesised a decade of X-ray diffraction and electron microscopy into the proposal that muscle contraction is the cyclic interaction of myosin cross-bridges pulling on actin filaments, with each stroke powered by ATP. Huxley's separation of force generation from filament sliding — distinct, ATP-coupled events — is the conceptual ancestor of every chemomechanical coupling discussed here, including the lever-arm hypothesis later articulated for myosin V by Spudich [ref: TODO_REF Spudich2001].

The generalisation from muscle to non-muscle cells required a methodological leap. Michael Sheetz and James Spudich's 1983 in vitro motility assay, in which myosin-coated beads crawled along the actin cables of the alga Nitella, proved that actin-myosin motility needs nothing more than the two proteins and ATP. Ron Vale, Thomas Reese, and Michael Sheetz's 1985 isolation of kinesin from squid giant axon identified a microtubule-based motor and opened the kinesin superfamily. Dynein had been found far earlier: Ian Gibbons discovered dynein in 1965 as the ATPase of tetrahymena cilia [ref: TODO_REF Gibbons1965], and Summers and Gibbons showed in 1971 that ATP drives sliding between axonemal doublets [ref: TODO_REF SummersGibbons1971], establishing the sliding-to-bending mechanism that underlies all ciliary and flagellar motion.

The single-molecule era closed the gap between ensemble biochemistry and discrete mechanical events. Svoboda and Block's 1994 optical-tweezer measurement resolved the $8$-nm step of kinesin [ref: TODO_REF SvobodaBlock1994], the first single step of any motor. Schnitzer and Block's 1997 force clamp fixed the stoichiometry at one ATP per $8$-nm step and mapped the force-velocity curve [ref: TODO_REF SchnitzerBlock1997], and Visscher, Schnitzer, and Block (1999) established the load-dependent stepping rate under constant force [ref: TODO_REF Visscher1999]. Yildiz and Selvin's 2004 nanometre-localisation of one kinesin head demonstrated hand-over-hand walking [ref: TODO_REF Yildiz2004], distinguishing it from inchworm. Vale's 2003 review [ref: TODO_REF Vale2003] collected these results into the "molecular motor toolbox" framing used here: a common chemomechanical logic realised by three structurally divergent protein families. The motors invoke no new force law; they are ATPases whose conformational cycles, read out on polar filaments and resolved by optical trapping, transduce thermal-scale chemical energy into directed mechanical work.

Bibliography Master

@article{Huxley1969,
  author = {Huxley, H. E.},
  title  = {The mechanism of muscular contraction},
  journal= {Science},
  volume = {164},
  pages  = {1356--1366},
  year   = {1969}
}

@article{Gibbons1965,
  author = {Gibbons, I. R. and Rowe, A. J.},
  title  = {Dynein: a protein with adenosine triphosphatase activity from cilia},
  journal= {Science},
  volume = {149},
  pages  = {424--426},
  year   = {1965}
}

@article{SummersGibbons1971,
  author = {Summers, K. E. and Gibbons, I. R.},
  title  = {ATP-induced sliding of tubules in trypsin-treated flagella},
  journal= {Proceedings of the National Academy of Sciences USA},
  volume = {68},
  pages  = {3092--3096},
  year   = {1971}
}

@article{SvobodaBlock1994,
  author = {Svoboda, K. and Block, S. M.},
  title  = {Force generation and step size: what can we learn from single-molecule experiments?},
  journal= {Cell},
  volume = {77},
  pages  = {773--784},
  year   = {1994}
}

@article{SchnitzerBlock1997,
  author = {Schnitzer, M. J. and Block, S. M.},
  title  = {Kinesin hydrolyses one {ATP} per 8-nm step},
  journal= {Nature},
  volume = {388},
  pages  = {386--390},
  year   = {1997}
}

@article{Visscher1999,
  author = {Visscher, K. and Schnitzer, M. J. and Block, S. M.},
  title  = {Single kinesin molecules studied with a molecular force clamp},
  journal= {Nature},
  volume = {400},
  pages  = {184--189},
  year   = {1999}
}

@article{Mehta1999,
  author = {Mehta, A. D. and Rock, R. S. and Rief, M. and Spudich, J. A. and Mooseker, M. S. and Cheney, R. E.},
  title  = {Myosin-{V} is a processive actin-based motor},
  journal= {Nature},
  volume = {400},
  pages  = {590--593},
  year   = {1999}
}

@article{Yildiz2004,
  author = {Yildiz, A. and Tomishige, M. and Vale, R. D. and Selvin, P. R.},
  title  = {Kinesin walks hand-over-hand},
  journal= {Science},
  volume = {303},
  pages  = {676--678},
  year   = {2004}
}

@article{Reck-Peterson2006,
  author = {Reck-Peterson, S. L. and Yildiz, A. and Carter, A. P. and Gennerich, A. and Zhang, N. and Vale, R. D.},
  title  = {Single-molecule analysis of dynein processivity and stepping behavior},
  journal= {Cell},
  volume = {126},
  pages  = {335--348},
  year   = {2006}
}

@article{Spudich2001,
  author = {Spudich, J. A.},
  title  = {The myosin swinging cross-bridge model},
  journal= {Nature Reviews Molecular Cell Biology},
  volume = {2},
  pages  = {387--392},
  year   = {2001}
}

@article{Vale2003,
  author = {Vale, R. D.},
  title  = {The molecular motor toolbox for intracellular transport},
  journal= {Cell},
  volume = {112},
  pages  = {467--480},
  year   = {2003}
}

@book{Howard2001,
  author    = {Howard, J.},
  title     = {Mechanics of Motor Proteins and the Cytoskeleton},
  publisher = {Sinauer Associates},
  year      = {2001}
}