Rapid time-stamped analysis of filament motility

Abstract

The in vitro motility assay is a valuable tool to understand motor protein mechanics, but existing algorithms are not optimized for accurate time resolution. We propose an algorithm that combines trace detection with a time-stamped analysis. By tracking filament ends, we minimize data loss from overlapping and crossing filaments. A movement trace formed by each filament end is created by time-stamping when the filament either first (filament tip) or last (filament tail) occupies a pixel. A frame number vs distance curve is generated from this trace, which is segmented into regions by slope to detect stop-and-go movement. We show, using generated mock motility videos, accurate detection of velocity and motile fraction changes for velocities <0.05 pixels per frame, without manual trace dropping and regardless of filament crossings.Compared with established algorithms we show greatly improved accuracy in velocity and motile fraction estimation, with greatly reduced user effort. We tested two actual motility experiments: 1) Adenosine triphosphate (ATP) added to skeletal myosin in rigor; 2) myosin light chain phosphatase (MLCP) added to phasic smooth muscle myosin. Our algorithm revealed previously undetectable features: 1) rapid increase in motile fraction paralleled by a slow increase in velocity as ATP concentration increases; 2) simultaneous reductions in velocity and motile fraction as MLCP diffuses into the motility chamber at very low velocities. Our algorithm surpasses existing algorithms in the resolution of time dependent changes in motile fraction and velocity at a wide range of filament lengths and velocities, with minimal user input and CPU time.

Molecular-level evidence of force maintenance by smooth muscle myosin during LC20 dephosphorylation

“…Actin filament movement was visualized and recorded using an excitation light source (X-Cite 120Q; Excelitas Technologies), electron-multiplying charge-coupled device camera (KP-E500; 720 × 480 resolution; 30 frames/s; 8-bit grayscale; Hitachi Kokusai Electric), and custom software with a frame grabber (MOR/2VD/84; Matrox). The average velocity ( v ) of all filaments moving faster than a specified cutoff velocity (0.05 μm/s for the in vitro motility assays or 0.10 μm/s for the in vitro motility mixture assays; see Table S1 for all parameters used in the video analysis) and motile fraction ( f mot ; the percentage of filaments moving above this threshold) were measured using customized Matlab (vR2018a) video analysis software ( Ijpma et al, 2018 ). To improve the accuracy of v measurements at low f mot values, any instantaneous filament velocity greater than three times the mean velocity at each time point for all videos of a given condition was ignored.…”

“…S8 shows the Western blots of the phasic and tonic SMM. Table S1 shows the parameter settings of the Matlab software used to analyze the in vitro motility assay videos ( Ijpma et al, 2018 ). Table S2 and Table S3 show the nonnormalized sigmoid model parameters determined by fitting the in vitro motility and mixture assay data, respectively.…”