Actin age orchestrates myosin-5 and myosin-6 run lengths

Abstract

Unlike a static and immobile skeleton, the actin cytoskeleton is a highly dynamic network of filamentous actin (F-actin) polymers that continuously turn over. In addition to generating mechanical forces and sensing mechanical deformation, dynamic F-actin networks serve as cellular tracks for myosin motor traffic. However, much of our mechanistic understanding of processive myosins comes from in vitro studies in which motility was studied on pre-assembled and artificially stabilized, static F-actin tracks. In this work, we examine the role of actin dynamics in single-molecule myosin motility using assembling F-actin and two highly processive motors, myosin-5 and myosin-6. These two myosins have distinct functions in the cell and travel in opposite directions along actin filaments [1-3]. Myosin-5 walks toward the barbed ends of F-actin, traveling to sites of actin polymerization at the cell periphery [4]. Myosin-6 walks toward the pointed end of F-actin [5], traveling toward the cell center along older segments of the actin filament. We find that myosin-5 takes 1.3- to 1.5-fold longer runs on ADP•Pi (young) F-actin, whereas myosin-6 takes 1.7- to 3.6-fold longer runs along ADP (old) F-actin. These results suggest that conformational differences between ADP•Pi and ADP F-actin tailor these myosins to walk farther toward their preferred actin filament end. Taken together, these experiments define a new mechanism by which myosin traffic may sort to different F-actin networks depending on filament age.

Figure 1

In vitro reconstitution of myosin-5 and myosin-6 motility on assembling F-actin. (A) Schematic of the experiments. Fluorescently labeled myosin-5 (top) and myosin-6 (bottom) walk along two kinds of actin tracks: phalloidin stabilized F-actin (left) and assembling F-actin (right). Nucleotide turnover on actin is illustrated by the transition from pink to blue subunits. (B) Time-lapse fluorescence micrographs of a single myosin-6 motor (green) moving along a single growing actin filament (red). Yellow arrowheads mark the growing F-actin barbed end, white arrowheads mark a single myosin traveling away from the growing end. Time stamp is in s. (C) Representative kymographs showing processive motility of 5 nM myosin-5 and myosin-6 on growing F-actin. Processive runs of myosins appear as green diagonal lines. Actin is shown in red, illustrating elongation of the barbed end toward the kymograph top.

Figure 2

Myosin-5 and myosin-6 runlengths respond to actin nucleotide state in opposite ways. Myosin-5 prefers young filaments, while myosin-6 prefers old. (A) Experimental timecourse. Actin polymerization begins at time zero. Experimental conditions are listed that yield: ADP F-actin (i, ii), mixed ADP and ADP•P_i F-actin (iii, iv), and uniform ADP•P_i F-actin (v). See Supplemental Experimental Details for exact conditions. (B–G) Runlengths of myosins along the F-actins listed in (A). The nucleotide state of the actin and the actin stabilizer are indicated. (B) Runlengths of myosin-5 on filaments assembled without stabilizer (iii) or stabilized with phalloidin after aging (i). Myosin-5 runs 1.4-fold farther on assembling actin (p = 2 × 10⁻⁸). (C) Runlengths of myosin-6, as in (B). Myosin-6 runs 1.7-fold farther on aged, phalloidin-stabilized actin (p = 0). (D) Runlengths of myosin-5 on capped (iv) or capped and aged (ii) F-actin. Myosin-5 runs 1.3-fold farther on younger filaments (p = 7 × 10⁻⁴). (E) Runlengths of myosin-6, as in (D). Myosin-6 runs 2-fold farther on the older filaments (p = 1 × 10⁻⁷). (F) Myosin-5 runlengths on F-actin co-polymerized with phalloidin to trap the ADP•P_i state (v). Myosin-5 runs 1.5-fold farther on the trapped ADP•P_i F-actin than on the ADP phalloidin F-actin (p = 3 × 10⁻⁹). Runlength curves from (B) are shown for comparison. (G) Runlengths of myosin-6, as in (F). Myosin-6 runs 3.6-fold farther on the ADP phalloidin F-actin (p = 0). All curves show the Kaplan-Meier estimator of the runlength survivor function; bands report the 0.95 CI. Events are left truncated at 400 nm and are right censored at filament ends. Reported fold-differences apply to mean runlengths, and p-values report the log-rank test. See Table S2 for summary statistics.

Figure 3

Myosin-5 and myosin-6 runlengths respond to nucleotide state gradients within growing filaments. (A) A schematic kymograph of myosin-6 runs on a growing filament (Figure 2Aiii), based on the kymograph in Figure 1C. Black lines indicate motor runs, contour lines indicate the nucleotide state probabilities along the actin filament. The P(ADP•P_i) values decay from one to zero from the barbed to the pointed end. (B) Myosin-5 runlengths, separated into two classes of P(ADP•P_i) values. Myosin-5 runs farther along stretches of F-actin in the upper third of P(ADP•P_i) values (p = 0.05, log-rank test). (C) Myosin-6 runlengths, separated into two classes. Myosin-6 moves farther along the stretches of actin in the lower third of P(ADP•P_i) values (p = 0.002, log-rank test).

Figure 4

Myosin-5 lands more frequently on ADP•P_i F-actin, while myosin-6 lands more frequently on ADP F-actin. Landing rates (the rate of initiating a processive run) are shown for myosin-5 (A) and myosin-6 (B). Myosin-5 lands significantly more often on ADP•P_i, Phalloidin-Copoly F-actin vs. ADP, Phalloidin F-actin (Figure 2Av vs. 2Ai, p = 5 × 10⁻⁶, Wilcoxon rank-sum test). Conversely, myosin-6 lands significantly more often on ADP, Phalloidin F-actin vs. ADP•P_i, Phalloidin-Copoly F-actin (Figure 2Ai vs. 2Av, p = 2 × 10⁻¹⁰, Wilcoxon rank-sum test).