Engineered myosin designs. An artificial rigid lever arm²⁰ composed of spectrin-like repeats (red) is fused to myosin VI after residue 780, retaining only 7 amino acids of the critical redirecting unique insert. This fusion is predicted to generate an exit angle that is intermediate between myosin VI and (+) end directed myosins. During the power stroke, rotation of the converter domain (light blue) leads to a net motion of the tip of the lever arm toward the (−) end of the actin filament for longer lever arms, and toward the (+) end for shorter lever arms. Fixed directionality constructs (a) with lever arms composed of differing numbers of spectrin repeats generate motion toward opposite ends of the actin filament. Calcium-controllable MCaR constructs (b) have chimeric lever arms composed of a single spectrin repeat fused to two IQ repeats (grey). Under low [Ca²⁺] conditions, the IQ domains are stabilized by binding to calmodulin (yellow), yielding a long lever arm and (−) end directed motion. High [Ca²⁺] favors dissociation of calmodulin, inducing a rigid to flexible transition that shortens the effective lever arm and reverses the direction of motion of the lever arm tip. An illustration (c) based on crystal structures shows the expected rigor conformation of MCaR-2IQ bound to the actin filament. In illustrations and block diagrams (d), structural modules are represented as follows: blue, myosin VI catalytic domain; light blue, myosin VI converter domain; red, one (1R) or two (2R) spectrin repeats from α-actinin; gray, IQ repeats from myosin V; ~, (GSG)₄ flexible linker; pink, eYFP. Sequences of junctions between modules are shown in Supplementary Figure 4.

Gliding filament assays of engineered myosins. Filament velocities were determined in the presence of 2.5 μM calmodulin, 2 mM ATP, and the indicated concentrations of free calcium, reported as pCa = −log₁₀([Ca²⁺]). Positive velocities indicate (+) end directed motion. Data points with error bars indicate mean ± SEM across sets of filaments (see Supplementary Table 1 and Supplementary Figure 2). For fixed directionality constructs (a), gliding direction is sensitive to engineered lever arm length but not to calcium (Movie S1). MCaR constructs (b) switch directionality near pCa 6, corresponding to ~μM concentrations of free calcium (Movies S2, S3, S4). A velocity time course (c) shows the results of a dynamic switching assay with MCaR-2IQ. pCa buffer was exchanged three times in the flow chamber while observing the same field of filaments (Movies S6A,B). Conditions alternated between pCa 8 (10 nM free calcium) and pCa 5 (10 μM free calcium). Out of 47 filaments analyzed, 30 filaments (shown here) switched directions three times; see Supplementary Figure 3 for complete tabulation of filament behaviors.

Processive motility of a dimeric controllable motor. Dimers were constructed (a) by fusion of MCaR-2IQ to the medial tail region of myosin VI (black) followed by the GCN4 leucine zipper (green) and a HaloTag (pink). Assays of processive motion on immobilized actin filaments were performed in 50 μM ATP, 2 μM calmodulin, and EGTA (b) or pCa 4 (100 μM free calcium) buffer (c). Compiled traces from single fluorophore tracking (Movies S7–S9) are shown on a polarized axis representing distance traveled along the actin filament. Directionalities were scored by comparing to subsequent assays of control (−) end directed M6DI₈₁₆2R-DIM motors^10,11 on the same filaments (Movies S8, S9). For MCaR-2IQ-DIM, (−) end directed traces (green) are ubiquitous in EGTA but represent a small minority (6%) of traces in pCa 4. A total of 75 runs in EGTA and 36 runs in pCa 4 were analyzed for velocity and processivity, yielding average velocities of −2.9±1.0 nm/s in EGTA and +3.2±2.0 nm/s in pCa 4, and processive run lengths of 550 ± 50 nm in EGTA and 170 ± 30 nm in pCa 4.