Transport efficiency of membrane-anchored kinesin-1 motors depends on motor density and diffusivity

Abstract

In eukaryotic cells, membranous vesicles and organelles are transported by ensembles of motor proteins. These motors, such as kinesin-1, have been well characterized in vitro as single molecules or as ensembles rigidly attached to nonbiological substrates. However, the collective transport by membrane-anchored motors, that is, motors attached to a fluid lipid bilayer, is poorly understood. Here, we investigate the influence of motors' anchorage to a lipid bilayer on the collective transport characteristics. We reconstituted "membrane-anchored" gliding motility assays using truncated kinesin-1 motors with a streptavidin-binding peptide tag that can attach to streptavidin-loaded, supported lipid bilayers. We found that the diffusing kinesin-1 motors propelled the microtubules in the presence of ATP. Notably, we found the gliding velocity of the microtubules to be strongly dependent on the number of motors and their diffusivity in the lipid bilayer. The microtubule gliding velocity increased with increasing motor density and membrane viscosity, reaching up to the stepping velocity of single motors. This finding is in contrast to conventional gliding motility assays where the density of surface-immobilized kinesin-1 motors does not influence the microtubule velocity over a wide range. We reason that the transport efficiency of membrane-anchored motors is reduced because of their slippage in the lipid bilayer, an effect that we directly observed using single-molecule fluorescence microscopy. Our results illustrate the importance of motor-cargo coupling, which potentially provides cells with an additional means of regulating the efficiency of cargo transport.

Keywords: lipid bilayers; molecular motors; motor-cargo coupling; streptavidin-binding peptide; transport efficiency.

Fig. 1.

Membrane-anchored kinesin-1 motors diffuse with about one-half the diffusivity of the free lipids. (A) Time-lapse series of fluorescence recovery after photobleaching (FRAP) images for an SLB with a molar composition of DOPC:DSPE-PEG-2000-Biotin:DOPE-Atto647n of 99:1:0.05. (Scale bar, 10 µm.) (B) Representative curve for the normalized fluorescence intensity after photobleaching vs. time (black; n = 4 SLBs; mean ± SD). The diffusivity of the lipid bilayer was determined to be D_Lipid = 3.0 ± 0.3 µm²/s (mean ± SD; n = 4 independent experiments; dashed red line; fit according to ref. 18). (C) Single-molecule trajectories of freely diffusing rKin430-SBP-GFP anchored on a biotinylated SLB via streptavidin. (Scale bar, 10 µm.) (D) MSD data (black; mean ± SD) of diffusing rKin430-SBP-GFP molecules. The diffusion coefficient was determined to be D_Kin = 1.4 ± 0.2 µm²/s (mean ± 95% confidence interval; n = 196 tracked molecules; red line; linear fit to the first eight points).

Fig. 2.

In vitro reconstitution of a membrane-anchored gliding motility assay. (A) Schematic drawing (not drawn to scale) of the experimental setup: truncated rat kinesin-1 with streptavidin-binding peptide tag (rKin430-SBP) is attached, via streptavidin, to a biotinylated SLB. The motors diffusively anchored on the SLB propel the microtubules. (B) Representative ensemble MSD data for the center positions of the microtubules (mean ± SEM; n ≥ 40 microtubules) at different motor concentrations in 1 mM ATP and 1 mM AMP-PNP (only for 0.13 µM rKin430-SBP). The red arrow indicates increasing motor concentration. To calculate the linear translocation components (i.e., the microtubule velocities ν_MT) the data were fit by Eq. 1. (C) Ensemble-averaged microtubule gliding velocities for different microtubule lengths, binned into 1-µm intervals, at different motor concentrations (mean ± 95% confidence interval; n ≥ 15 microtubules for each data point).

Fig. 3.

Membrane-anchored kinesin-1 motors slip in the lipid bilayer, while propelling a microtubule. (A) Time-lapse images for microtubules driven by membrane-anchored motors (Upper) and surface-immobilized motors (Lower) with schematic experimental setups on the Right. (Scale bar, 5 µm.) The arrows indicate the transport directions of the gliding microtubules. Microtubules propelled by membrane-anchored motors do not cross each other in contrast to microtubules driven by surface-immobilized motors. (B and C) Representative kymographs (inverted contrast) showing the movement of individual rKin430-SBP-GFP motors (dark signals) while propelling microtubules at low motor density (B) and high motor density (C). The red lines mark the trailing ends of the microtubules as guides to the eye. In C, a gliding microtubule collides with another passing microtubule and is temporarily stalled (also shown in the schematics on the Right) until the other microtubule glides away.

Fig. 4.

Theoretical description of the membrane-anchored gliding motility fits well to the experimental observations. (A) Nanoscopic view of the experimental setup illustrating the velocities and frictional forces of microtubules and motors, used to derive the mathematical model. The depicted microtubules glide to the left with respect to the substrate at velocity v_MT. The motor steps on the microtubule with velocity v_Step and thereby moves its anchor in the SLB to the right at velocity v_Kin-slip with respect to the substrate. The frictional forces act on the microtubules and motors in the directions opposite to their motion. (B) Averaged microtubule transport efficiency (mean ± 95% confidence interval; n ≥ 40 microtubules for each data point) for various kinesin-1 surface densities (mean ± SD; n = 5 regions of interest) and different SLB compositions (0% CH, 20% CH, and 60% CH). The data were fitted (solid lines) to the mathematical model (Eq. 13), with one free parameter ω (reach of a diffusing kinesin-1 motor to bind to a microtubule; c = 0.72 µm⁻³⋅s; R² ≥ 0.95).

Fig. S1.

Single-molecule stepping velocity of rKin430-SBP-GFP. (A) Representative kymograph of single rKin430-SBP-GFP molecules moving on a surface-immobilized microtubule. Time is progressing from Top to Bottom, while the motors (dark signals) move along a microtubule from Left to Right. (B) Histogram of single-molecule velocities with an ensemble average velocity of 0.67 ± 0.14 µm/s (mean ± SD; n = 545 molecules). Schematics shown on Top of the histogram.

Fig. S2.

Diffusivity of lipids and membrane-anchored kinesin-1 motors reduces with increasing amount of cholesterol in the SLBs. (A and E) Time-lapse series of fluorescence recovery after photobleaching (FRAP) images for two different SLBs with molar compositions of DOPC:Cholesterol:DSPE-PEG-2000-Biotin:DOPE-Atto647n at 79:20:1:0.05 (20% CH) (A) and DOPC:Cholesterol:DSPE-PEG-2000-Biotin:DOPE-Atto647n at 39:60:1:0.05 (60% CH) (E). (Scale bar, 10 µm.) (B and F) Representative curves of the normalized fluorescence intensity after photobleaching vs. time (black; n = 4 SLBs; mean ± SD). The diffusivity of the lipid bilayer was determined to be D_{Lipid_20%CH} = 2.1 ± 0.4 µm²/s (B) and D_{Lipid_60%CH} = 1.1 ± 0.2 µm²/s (F) (mean ± SD; n = 4 independent experiments; dashed red lines; fits according to ref. 18). (C and G) Single-molecule trajectories of freely diffusing rKin430-SBP-GFP anchored on a 20% CH SLB (C) and 60% CH SLB (G) via streptavidin. (Scale bar, 5 µm.) (D and H) MSD data (black; mean ± SD) of diffusing rKin430-SBP-GFP molecules. The diffusion coefficient was determined to be D_{Kin_20%CH} = 1 ± 0.2 µm²/s (D) and D_{Kin_60%CH} = 0.5 ± 0.1 µm²/s (H) (mean ± 95% confidence interval; n = 48 and 58 tracked molecules, respectively; red line; linear fit to the first eight points).