Control of cytoplasmic dynein force production and processivity by its C-terminal domain

Abstract

Cytoplasmic dynein is a microtubule motor involved in cargo transport, nuclear migration and cell division. Despite structural conservation of the dynein motor domain from yeast to higher eukaryotes, the extensively studied S. cerevisiae dynein behaves distinctly from mammalian dyneins, which produce far less force and travel over shorter distances. However, isolated reports of yeast-like force production by mammalian dynein have called interspecies differences into question. We report that functional differences between yeast and mammalian dynein are real and attributable to a C-terminal motor element absent in yeast, which resembles a 'cap' over the central pore of the mammalian dynein motor domain. Removal of this cap increases the force generation of rat dynein from 1 pN to a yeast-like 6 pN and greatly increases its travel distance. Our findings identify the CT-cap as a novel regulator of dynein function.

Figure 1. Rat cytoplasmic dynein motor domain constructs.

(a) Domain organization of the native dynein heavy chain (HC) and the engineered constructs, MD-WT (a.a. 1,157–4,644) and MD-ΔCT (a.a. 1,157–4,348). The line indicates the HC dimerization region, which is truncated in MD-WT and MD-ΔCT and replaced with an N-terminal GST for MD dimerization. (b) Sequence alignment of the proximal C-termini of native rat dynein, native yeast dynein and MD-ΔCT. The MD-ΔCT truncation eliminates the hinge region and distal C-terminus, but preserves the proximal H1 helix, as in yeast. (c) Dynein MD structure (PDB entry 3VKH14). The C-terminal elements are represented as tubes in white. ‘L’ and ‘S’ indicate large and small subdomains, respectively, of AAA1 and AAA2. Inset: same view, with AAA1S, the AAA5 extension, and the CT-cap removed. The dashed outlines indicate the positions of AAA1S and the AAA5 extension. Note the AAA1 active site (formed at the interface of AAA1L, AAA1S and AAA2L) and the H1 helix (running between AAA5/AAA6 and the AAA5 extension). (d) Schematic illustrations of the MD-WT and MD-ΔCT constructs, created using PDB entries 3VKH and 1VF4 (see Methods section for additional information). (e) Coomassie-stained gel of MD-WT and MD-ΔCT purified via SpinTrap column. HC: dynein heavy chain; *free GST (see Supplementary Information).

Figure 2. Single-molecule function of MD-WT.

(a) Illustration of the optical trapping assay. GST-dynein is attached via an anti-GST antibody to a 1- μm polystyrene microsphere (‘bead’) that is optically trapped above a MT. As dynein moves along the MT, the trap exerts an opposing force. (b) Fraction of motile beads (those generating forces ≥0.5 pN using a trap stiffness of k=0.01 pN nm⁻¹) versus the relative MD-WT concentration. Error bars are Clopper–Pearson 95% confidence intervals (95% CIs) of the mean. We tested 14–25 beads at each concentration (109 total). The curves are fits assuming processive (solid line) or nonprocessive (dashed line) motors (see Methods). The data are best fit by the processive model (coefficient of determination R²=0.99 versus R²=0.91; F-test P-value=0.02). (c) Representative examples of MD-WT force generation (1 mM ATP) at motor concentrations for which 50% or fewer beads moved. Red bars indicate duration of maximal sustained force. (d) Histogram of stall forces (maximal forces sustained for≥200 ms), with average 1.0±0.5 pN (mean±s.d.). The curve is a Gaussian fit to the data (mean 0.9 pN and s.d. 0.5 pN; 95% CIs (0.8, 1.0) and (0.4, 0.6) pN, respectively). Of 381 MT encounters (derived from 54 beads over 19 experiments), 100 (26%) met the criterion for stalling. (e) Example of large forces produced by MD-WT at high motor concentration (100% fraction of motile beads). Experiments were performed with AC-purified protein (e), with AC-/SEC-purified protein (b,c), and with both AC-purified and AC-/SEC-purified protein (d).

Figure 3. Force generation of single MD-ΔCT molecules.

(a) Processivity analysis, as in Fig. 2b. The data are best fit by the processive model (R²=0.99 versus R²=0.89; F-test P-value=0.01). We tested 11–50 beads at each concentration (154 total). (b) Example of single-motor force generation (k=0.061 pN nm⁻¹). Steps are visible after reaching ~4 pN. The two red bars indicate periods of stalling (here at ~6.5 pN). The red asterisk marks a run terminated by detachment before stalling. (c) Histogram of stall forces (maximal forces sustained for ≥400 ms), with average 5.4±1.1 pN (mean±s.d.) and Gaussian fit (mean 5.5 pN; s.d. 1.0 pN; 95% CI’s (5.3, 5.7) and (0.8, 1.2) pN, respectively). Of 340 total MT encounters (derived from 25 beads over 8 separate experiments), 135 (44%) met the stalling criterion. ATP concentration was 1 mM. Experiments were performed with AC-purified protein (a,b) and with both AC-purified and AC-/SEC-purified protein (c).

Figure 4. Single-molecule function of MD-ΔCT.

(a) Bead movement under constant load (optical trap force clamp). Top: repeated bead displacements by single motors (black). The trap (magenta) follows at a fixed distance to apply a constant force (here, 2 pN). When the bead travels beyond the region of force-clamp operation, it either stalls (black arrow) or detaches. Bottom: detail of the event marked by an asterisk in the top panel, fit with a line to measure the mean velocity (~510 nm s⁻¹). The lower inset shows the applied force, which was held constant at 2.1±0.3 pN (mean±s.d.) during force-clamp operation (yellow region). (b) Velocity versus force. Points are means of repeated measurements under constant force, including only runs ≥50 nm. Error bars span 95% CIs of the mean. The red line is a weighted linear fit (V=672 nm s⁻¹−119 nm s⁻¹ pN⁻¹ × F), with shaded region of 95% confidence that intercepts the abscissa at (575, 769) nm s⁻¹ and the ordinate at (5.1, 6.5) pN. Data are from six beads over two experiments (N=38–233 at each force; 655 events total). (c) Histogram of dwell times between consecutive forward steps under 5 pN load. An exponential fit, y=A exp(−k_cat t), gives k_cat =12.1 s⁻¹; 95% CI (10.2, 13.9) s⁻¹. Inset: empirical cumulative probability density function, with fit y=1−exp(−k_cat (t−t_L)), where t_L is the dwell time detection limit (~6–8 ms) and k_cat =10.6 s⁻¹ (95% CI (10.4, 10.8) s⁻¹). Data from four beads over three separate experiments (211 events). (d) Top: histogram of step sizes under 5 pN load (N=254). Grey and black bars represent forward (95% of steps) and backward steps, respectively. A Gaussian fit (black curve) to the forward steps yields 8.2±1.3 nm (mean±s.d.; 95% CI’s (8.0, 8.4) and (1.1, 1.5) nm, respectively). Backward steps were 7.8±2.7 nm (mean±s.d.). Bottom: example trace showing steps (red line) identified by a step-finding algorithm (see Methods). ATP concentration was 1 mM. Experiments were performed with AC-purified protein.