Myosin-II-mediated directional migration of Dictyostelium cells in response to cyclic stretching of substratum

Abstract

Living cells are constantly subjected to various mechanical stimulations, such as shear flow, osmotic pressure, and hardness of substratum. They must sense the mechanical aspects of their environment and respond appropriately for proper cell function. Cells adhering to substrata must receive and respond to mechanical stimuli from the substrata to decide their shape and/or migrating direction. In response to cyclic stretching of the elastic substratum, intracellular stress fibers in fibroblasts and endothelial, osteosarcoma, and smooth muscle cells are rearranged perpendicular to the stretching direction, and the shape of those cells becomes extended in this new direction. In the case of migrating Dictyostelium cells, cyclic stretching regulates the direction of migration, and not the shape, of the cell. The cells migrate in a direction perpendicular to that of the stretching. However, the molecular mechanisms that induce the directional migration remain unknown. Here, using a microstretching device, we recorded green fluorescent protein (GFP)-myosin-II dynamics in Dictyostelium cells on an elastic substratum under cyclic stretching. Repeated stretching induced myosin II localization equally on both stretching sides in the cells. Although myosin-II-null cells migrated randomly, myosin-II-null cells expressing a variant of myosin II that cannot hydrolyze ATP migrated perpendicular to the stretching. These results indicate that Dictyostelium cells accumulate myosin II at the portion of the cell where a large strain is received and migrate in a direction other than that of the portion where myosin II accumulated. This polarity generation for migration does not require the contraction of actomyosin.

Figure 1

Estimation of myosin II-localization and pseudopod dynamics that follow. (A–C) From two sequential images at a time the interval of 40 s, l₁, l₂ and F were defined as shown in A and B (see Materials and Methods for details). F and l₂/l₁ were calculated along the 12 dotted lines in C. F refers to spontaneous localization of myosin II in freely migrating cells. The value l₂/l₁ means the contraction or expansion of pseudopodia. (D–F) From two images before and after four repeats of stretching using the device shown in Fig. 2A, l₁, l₂, F₁ and F₂ were defined as shown in D and E (see Materials and Methods for details). The values l₁, l₂, F₁ and F₂ were calculated along the six dotted lines in F. F₂/F₁ means localization of myosin II in response to the four repeats of stretching of substratum.

Figure 2

Microstretching device. (A) Overview of a device. By pushing a part of the device with a glass microneedle, the chamber is stretched in the horizontal direction. (B) Strain distribution of A when the magnitude of stretch in the entire microchamber is 0.26 (26% stretch). There is a roughly homogeneous strain field.

Figure 3

Crawling migrations of wild-type Dictyostelium cells on elastic substratum with cyclic stretching. Stretching ratio and time cycle were 20% and 5 s, respectively. (A and B) Trajectories of migrating cells under cyclic stretching in the horizontal (0°–180°) direction (A; n = 120 from five experiments) and under no cyclic stretching (B; n = 161 from six experiments). (C and D) Frequencies of migrating direction (θ) calculated from A and B, respectively. (A, inset) θ is defined as an angle between a horizontal line and a straight line connecting the starting and ending points, measured for each migrating cell during a single experiment for 30 min. (E) Average |sinθ| values calculated from C and D. The |sinθ| values on the substratum under cyclic stretching were significantly higher than those for no cyclic stretching (P = 1.2 × 10⁻⁶). Arrows in C indicate two peaks of frequencies at 90° and 270°. There is no significant peak in D.

Figure 4

Crawling migrations of mutant cells on elastic substratum with cyclic stretching. Stretching ratio and time cycle were the same as in Fig. 3. (A and B) Trajectories of migrating mhcA⁻ cells under cyclic stretching in the horizontal (0°–180°) direction (A; n = 141 from five experiments) and under no cyclic stretching (B; n = 132 from six experiments). (C and D) Frequencies of migrating direction (θ) calculated from A and B, respectively. (E) Average |sinθ| values calculated from C and D. There is no significant difference between the stretching and no-stretching columns of |sinθ| values (P = 0.39). (F and G) Frequencies of the migrating direction (θ) of SibA⁻ cells under cyclic stretching in the horizontal (0°–180°) direction (F; n = 114 from six experiments) and under no cyclic stretching (G; n = 116 from six experiments). (H) Average |sinθ| values calculated from F and G. There is a significant difference between the stretching and no-stretching columns of |sinθ| values (P = 0.001).

Figure 5

Myosin II localization caused by stretching of the substratum. (A) Typical localizations of GFP-myosin II by 20% stretching of the substratum using the device featured in Fig. 2A, shown before stretching (left), after two stretches (center), and after four stretches (right). The mhcA⁻ cells expressing GFP-myosin II were dispersed in the chamber of the device. The stretching directions are indicated by the double-headed arrow in the left panel. Localizations of GFP-myosin II were bilateral in the horizontal direction after two stretches (dotted ellipses) and became clearer after four stretches. (B–F) Evaluation of GFP-myosin II localization. (B) The cell area was divided into four portions by dotted lines. The areas 2 μm from the cell boundary in the four portions were defined as a, b, c, and d. F_a, F_b, F_c, and F_d are the maximal fluorescence values for GFP-myosin II in a–d, respectively. F_i|F_j is expressed as the larger of F_i and F_j divided by the smaller. (C) F_a, F_b, F_c, and F_d before stretching. There was no significant difference between F_a and F_c (n = 62, P = 0.57), and F_b and F_d (n = 62, P = 0.83). (D) F_a, F_b, F_c, and F_d after four stretches. There was no significant difference between F_a and F_c (n = 53, P = 0.87), and F_b and F_d (n = 53, P = 0.14). F_a and F_c in D are significantly larger than the corresponding values in C (P = 0.001 for F_a and 0.001 for F_c), although there is no significant difference in F_b and F_d between C and D (P = 0.49 for F_b and 0.06 for F_d). (E) F_a|F_c and F_b|F_d before stretching. There was no significant difference between F_a|F_c and F_b|F_d (P = 0.41). (F) F_a|F_c and F_b|F_d after four stretches. F_a|F_c significantly decreased, although F_b|F_d did not change (P = 1 × 10⁻⁵). (G) Typical localizations of GFP-ABD120k by 20% stretching of the substratum using the device shown in Fig. 2 A. Actin meshwork was shown as fluorescence of GFP-ABD120k in midcell. Actin filaments in the meshwork did not align in any direction, even after four repeats of stretching. In the boundary of the cell, localization of actin filaments was shown. (H and I) Evaluation of localizations of GFP-ABD120k before stretching (H; n = 29) and after four stretches (I; n = 21).

Figure 6

Localization of myosin II variants caused by stretching of the substratum using the device shown in Fig. 2A. (A–C) Typical localization of GFP-3xALA, GFP-3xASP, and GFP-E476K myosin II in mhcA⁻ cells by 20% stretching of the substratum, shown before stretching (left) and after four repeats of stretching (right). The stretching directions are indicated by double-headed arrows under the left images. Localizations of GFP-3xALA and -E476K myosin II were bilateral in the horizontal direction after four repeats of stretching (dotted ellipses), although uniform localization of GFP-3xASP myosin II was not changed by the stretching. (D–I) Evaluation of localizations of GFP-myosin II variants. Data were calculated in the same manner as for Fig. 5, E and F. (D–F) GFP-3xALA, GFP-3xASP, and GFP-E476K myosin II, respectively, before stretching. In all variants, there is no significant difference between F_a|F_c and F_b|F_d (n = 144 (GFP-3xALA), 160 (GFP-3xASP), and 130 (GFP-E476K), P > 0.1). (G–I) GFP-3xALA, GFP-3xASP, and GFP-E476K myosin II, respectively, after four repeats of stretching. In the case of GFP-3xALA and GFP-E476K myosin II, F_a|F_c is significantly smaller than F_b|F_d (n = 142 (GFP-3xALA) and 101 (GFP-E476K), P < 0.01). There is no significant difference in F_a|F_c and F_b|F_d in GFP-3xASP myosin II.

Figure 7

Crawling migrations of mhcA⁻ cells expressing GFP-myosin II variants on elastic substratum with cyclic stretching. Stretching ratio and time cycle were the same as in Fig. 3. (A–C) Frequencies of the migrating direction (θ) of 3xALA cells (A; n = 109 from six experiments), 3xASP cells (B; n = 149 from eight experiments), and E476K cells (C; n = 106 from six experiments) under cyclic stretching in the horizontal (0°–180°) direction. Arrows in A and C indicate two peaks at 90° and 270°. (D–F) Frequencies of migrating direction (θ) of 3xALA cells (D; n = 119 from eight experiments), 3xASP cells (E; n = 128 from eight experiments), and E476K cells (F; n = 83 from five experiments) in the absence of cyclic stretching. (G–I) Average |sinθ| values. The values of the left and right columns in each graph were calculated from the migrations under cyclic stretching (A–C) and with no cyclic stretching in the horizontal (0°–180°) direction (D–F), respectively. In 3xALA and E476K cells, the |sinθ| values were significantly higher in response to cyclic stretching than in the absence of stretching (G and I, P < 0.01).

Figure 8

Localization of myosin II and pseudopod dynamics that follow. (A and B) Freely migrating mhcA^- cells expressing GFP-myosin II (A) and E476K cells (B) on a coverslip. F refers to spontaneous localization of myosin II in freely migrating cells. The value l₂/l₁ means the contraction or expansion of pseudopodia. (C–E) The mhcA^- cells expressing GFP-myosin II (C), 3xALA (D), and E476K cells (E) under cyclic stretching using the device shown in Fig. 2A. F₂/F₁ means localization of myosin II in response to the four repeats of stretching of substratum. Dotted lines in C–E are where F₂/F₁ = 1.