Cytoplasmic dynein transports cargos via load-sharing between the heads

Abstract

Cytoplasmic dynein is a motor protein that walks along microtubules (MTs) and performs mechanical work to power a variety of cellular processes. It remains unclear how a dynein dimer is able to transport cargos against load without coordinating the stepping cycles of its two heads. Here by using a DNA-tethered optical trapping geometry, we find that the force-generating step of a head occurs in the MT-bound state, while the 'primed' unbound state is highly diffusional and only weakly biased to step towards the MT-minus end. The stall forces of the individual heads are additive, with both heads contributing equally to the maximal force production of the dimer. On the basis of these results, we propose that the heads of dynein utilize a 'load-sharing' mechanism, unlike kinesin and myosin. This mechanism may allow dynein to work against hindering forces larger than the maximal force produced by a single head.

Figure 1. Domain organization and mechanochemical cycle of cytoplasmic dynein

(a) The dynein heavy chain consists of an N-terminal cargo-binding tail domain, an AAA+ ATPase ring attached to the tail via the mechanically active linker, and a microtubule binding domain (MTBD) separated from the AAA+ ring by a ~15 nm coiled-coil stalk. Individual AAA subunits are numbered 1 through 6. (b) Following the binding of ATP at the principal ATPase site (AAA1, colored dark blue), the dynein head releases from the MT and its linker undergoes a priming stroke. The primed linker exits the ring at AAA2 rather than AAA4, and a dynein monomer attains an extended conformation. After the head re-binds the MT, its linker undergoes a power stroke, returning to its initial conformation exiting the ring at AAA4. The AAA1 site then releases ADP, completing the mechanochemical cycle.

Figure 2. Measuring forces produced by a single dynein head

(a) Schematic of the head-tethered optical trapping assay for measuring forces produced by a single dynein head. One head of a GST-dimerized motor (GST shown in orange) is labeled with a DNA tether at the C-terminal HT tag (yellow), and tethered to a trapped bead through a biotin-streptavidin (green) linkage. (b) Cy5-DNA labeled dynein migrates in two distinct bands in a denaturing gel. Fluorescent image identifies the labeled fraction. Labeling efficiency was 31% in the gel shown here. (c) Kymograph shows that Cy3-DNA labeled dynein moves processively towards the MT minus end. Scale bars: 2 μm (horizontal), 10 s (vertical). (d) Measurement of the stall force of head-tethered dynein. Valid stall events are marked with red ticks. Trap stiffness (k_trap) is 0.037 pN nm⁻¹. (e,f) Stall force histograms of head- and tail-tethered dynein motors, respectively (mean ± s.e.m.; N = 97 and 123, respectively).

Figure 3. Observation of steps taken by a dynein head under load

(a) Stepping traces of head-tethered dynein in force-feedback mode under 2.5 pN hindering load. Position traces (gray lines) are decimated to 330 Hz. The output of the step-fitting algorithm is shown in blue (k_trap = 0.025 pN nm⁻¹). (b) Step size distributions of head- and tail-tethered motors under 2.5 pN hindering load (mean ± s.d.; N = 2300 and 3293, respectively). (c) Stepping rates (k) of head-tethered and tail-tethered motors under 2.5 pN hindering load, obtained by fitting cumulative probability distributions (solid black lines) to single exponentials (dashed red lines). Reported values are mean ± 95% confidence interval boundaries.

Figure 4. Behavior of a mutant head under load

(a) Schematic of the mutant head-tethered optical trapping assay. The DNA tether is attached to an AAA1_K/A mutant head, which cannot bind ATP at the AAA1 site. The mutant head is heterodimerized with a WT head through rapamycin-induced FRB/FKBP binding. (b) Stall force measurement of a mutant-head tethered heterodimer at 1 mM ATP (k_trap = 0.017 pN nm⁻¹). (c) The mutant head-tethered construct stalls at half the stall force of WT (mean ± s.d.; N = 48) (d) Stall force of the mutant head-tethered construct is independent of ATP concentration. Error bars represent s.e.m. N = 48 to 88 stalls for each data point.

Figure 5. Characterization of a WT head of a WT/AAA1_K/A heterodimer under load

(a) Schematic of the WT head-tethered optical trapping assay. The DNA tether is attached to WT head, which is heterodimerized with an AAA1_K/A mutant head. (b) Stall force measurement of the WT-head tethered heterodimer (k_trap = 0.008 pN nm⁻¹). (c) WT-head tethered heterodimers stall at lower forces (0.5 ± 0.1 pN, mean ± s.d.; N = 39) than the mutant-head tethered motors. (d) Despite having a very low stall force, the live-head tethered construct is able to move processively over long distances against sub-stall (0.4 pN) hindering loads (k_trap = 0.008 pN/nm). (e) Model illustrating dynein force-production in the MT-bound state. Upon releasing from the MT, a dynein head undergoes a priming stroke. This component of the step is highly diffusional and can be biased backwards by external load. The force-generating step occurs after the head binds to a new site on the MT lattice.

Figure 6. WT and AAA1_K/A force-velocity relationships

Force-velocity curves for head-tethered WT/WT homodimer (a) and mutant head-tethered AAA1_K/A heterodimer (b) at 1 mM ATP are fitted with a one state model in which the backward stepping rate is load independent (black curve). The parameters obtained from the fit are the unloaded velocity V₀, the characteristic length L, and the minimum velocity V_min. Error bars represent S.E.M (N_beads = 4-10). Velocities at “0 pN force” were obtained from single molecule fluorescence measurements (N_motor >60).

Figure 7. Load-sharing model for dynein force production

(i) When dynein is in a two-head bound state, the external load (F) is distributed between the two heads due to the elastic nature of the tail (stretched springs) and linker (bent green rods) regions. Because each panel represents an equilibrium state, the sum of the three forces (F_cargo, F₁, and F₂) is always equal to zero. (ii) ATP binding to the AAA1 site of one head triggers its release from the MT and priming of the linker (straight green rod). As this head diffuses to a new binding site, the external load is transferred in its entirety to the DNA-tethered head (iii) The stepping head rebinds to MT, with the linker still in the primed configuration. (iv) The linker returns to its original state via a powerstroke, the tail moves forward and the load is redistributed between the heads. Load sharing allows a dynein dimer to work against hindering forces larger than the forces that can be produced by a single head.