The F-actin bundler α-actinin Ain1 is tailored for ring assembly and constriction during cytokinesis in fission yeast

Abstract

The actomyosin contractile ring is a network of cross-linked actin filaments that facilitates cytokinesis in dividing cells. Contractile ring formation has been well characterized in Schizosaccharomyces pombe, in which the cross-linking protein α-actinin SpAin1 bundles the actin filament network. However, the specific biochemical properties of SpAin1 and whether they are tailored for cytokinesis are not known. Therefore we purified SpAin1 and quantified its ability to dynamically bind and bundle actin filaments in vitro using a combination of bulk sedimentation assays and direct visualization by two-color total internal reflection fluorescence microscopy. We found that, while SpAin1 bundles actin filaments of mixed polarity like other α-actinins, SpAin1 has lower bundling activity and is more dynamic than human α-actinin HsACTN4. To determine whether dynamic bundling is important for cytokinesis in fission yeast, we created the less dynamic bundling mutant SpAin1(R216E). We found that dynamic bundling is critical for cytokinesis, as cells expressing SpAin1(R216E) display disorganized ring material and delays in both ring formation and constriction. Furthermore, computer simulations of initial actin filament elongation and alignment revealed that an intermediate level of cross-linking best facilitates filament alignment. Together our results demonstrate that dynamic bundling by SpAin1 is important for proper contractile ring formation and constriction.

FIGURE 1:

SpAin1 bundles F-actin. (A) Domain organization of fission yeast α-actinin SpAin1 and human α-actinin 4 HsACTN4, with amino acid numbers indicated next to domain boundaries. Dashed lines indicate significant residue identity between specific spectrin repeats (SR). ABD, actin-binding domain. (B) Regions of Coomassie blue–stained gels showing purified proteins used in this study. (C and D) Low-speed (10,000 × g) sedimentation of F-actin preassembled from 5 μM Mg-ATP actin, following incubation with a range of SpAin1 or HsACTN4 concentrations for 20 min at 25°C. (C) Coomassie blue–stained gel of pellets from a representative experiment with SpAin1. (D) Plot of the dependence of actin in the pellets on the concentration of SpAin1 or HsACTN4. Error bars, SE; n = 3. (E) Fluorescence micrographs of preassembled F-actin labeled with rhodamine–phalloidin following incubation with the indicated cross-linkers for 20 min. Scale bar: 5 μm.

FIGURE 2:

SpAin1 makes F-actin bundles of mixed polarity. (A–C) TIRFM observation of the addition of 0.75 μM Mg-ATP actin (33.3% Oregon Green-actin) to preassembled F-actin bundles (red boxes) (10.0% Oregon Green-actin) made by the indicated concentrations of human fascin HsFSCN1, fission yeast fimbrin SpFim1, SpAin1, or HsACTN4. (A) Time-lapse fluorescence micrographs. Scale bar: 5 μm. (B) Time-lapse fluorescence micrographs of representative parallel (top) and antiparallel (bottom) two-filament bundles made by SpAin1, used for quantification in C. Scale bar: 5 μm. Green and blue arrowheads mark F-actin barbed ends. (C) Percentage of two-filament bundles that are parallel and antiparallel.

FIGURE 3:

SpAin1 bundles F-actin less well than HsACTN4. (A) Merged two-color TIRFM time-lapse fluorescence micrographs showing the assembly of 2.5 μM Mg-ATP actin (33.3% Oregon Green-actin) in the presence of 300 nM SpAin1 (20% labeled with TMR). Blue and yellow arrowheads mark elongating barbed ends and SpAin1 decoration of a two-filament bundle. Scale bar: 5 μm. (B–E) Single-color TIRFM of the assembly of 2.5 μM Mg-ATP actin (33.3% Oregon Green-actin) in the absence and presence of the indicated concentrations of HsACTN4 or SpAin1. (B) Representative field of filaments 12 min after reactions were initiated. Fluorescence micrographs (left) and pseudocolored panels (right) with individual and bundled filaments traced with different colors: yellow, 1 filament; red, 2; green, 3; magenta, 4; cyan, ≥ 5. Scale bar: 5 μm. (C) Dependence of the number of filaments per bundle on the concentration of SpAin1 or HsACTN4. Error bars, SE; n = 4. (D) Distribution of Oregon Green-actin pixel intensity of filaments assembled in the absence and presence of SpAin1 or HsACTN4. (E) Polarity of two-filament bundles.

FIGURE 4:

SpAin1 is a highly dynamic F-actin–bundling protein. (A–E) Two-color TIRFM of 2.5 μM Mg-ATP actin (33.3% Oregon Green-actin) assembled in the presence of 900 nM SpAin1, HsACTN4 or SpAin1(R216E) (1% TMR-labeled). (A) Merged fluorescence micrographs of α-actinin on two-filament bundles. Scale bar: 5 μm. (B) Time-lapse fluorescence micrographs of TMR-labeled α-actinin on corresponding bundles in A. Arrowheads mark speckles that remain for more (red) or less (green) than 1.5 s. (C) Kymographs of α-actinin on bundles in A over 20 s. Red boxes mark the 1.5 s of montages shown in B. Scale bars: 5 μm (horizontal); 5 s (vertical). (D) Histograms of the residence time of α-actinin speckles on two- and three-filament bundles. (E) Exponential fits of the fraction of α-actinin bound over time on two- and three-filament bundles, revealing the indicated residence times.

FIGURE 5:

Mutant SpAin1(R216E) bundles F-actin better than WT SpAin1. (A) Domain organization of SpAin1 indicating the position of the R216E mutation in the second of two CH domains (top). ABD, actin-binding domain; CH, calponin homology domain. Alignment of the last α-helix of the CH2 domain of α-actinin from different species, with the 100% conserved residue SpAin1 R216 labeled in red with an asterisk (bottom) (Borrego-Diaz et al., 2006; Weins et al., 2007). (B) Low-speed sedimentation of F-actin over a range of concentrations of SpAin1 or SpAin1(R216E). Error bars, SE; n = 3. (C and D) Single-color TIRFM of the assembly of 2.5 μM Mg-ATP actin (33.3% Oregon Green-actin) in the absence and presence of 300 nM SpAin1 or SpAin1(R216E). (C) Representative field of filaments 10 min after a reaction with SpAin1(R216E). Fluorescence micrograph (left) and pseudocolored panel (right) with individual and bundled filaments traced with different colors: yellow, 1 filament; red, 2; green, 3; magenta, 4; cyan, ≥ 5. Scale bar: 5 μm. (D) Average number of filaments per bundle. Error bars, SE; n = 4. (E) Quantification of bundle polarity of 300 nM WT SpAin1 or mutant SpAin1 (R216E). Error bars, SE; n = 2.

FIGURE 6:

SpAin1(R216) causes severe cytokinetic defects in fission yeast. (A–D) WT cells, SpAin1 null ain1-Δ1 cells, and ain1-Δ1 cells expressing either WT SpAin1-GFP or mutant SpAin1(R216E)-GFP from low- (81nmt1) and medium- (41nmt1) strength promoters integrated at the leu locus. (A and B) General cytokinetic defects 24 h after nmt1-promoter induction by removal of thiamine. (A) Fluorescence micrographs of cells stained with DAPI and calcofluor to visualize nuclei and septa. (B, left) Percentage of cells with 1 (white), 2 (light blue), or > 2 nuclei (dark blue). n ≥ 400 cells. (B, right) Percentage of normal (n, white) and abnormal (a, blue) septa. n ≥ 100 septa. (C) Z-projection fluorescence micrographs of the indicated cells expressing the contractile ring marker Rlc1-tdTomato (left), SpAin1-GFP or SpAin1(R216E)-GFP (middle), and merge (right). Scale bar: 5 μm. (D) SpAin1-GFP and SpAin1(R216E)-GFP whole cell fluorescence for the indicated strains. Error bars, SE; n ≥ 10 cells.

FIGURE 7:

SpAin1(R216) causes defects in both contractile ring assembly and constriction. (A–F) Cytokinetic kinetics of WT cells, SpAin1 null ain1-Δ1 cells, and ain1-Δ1 cells expressing either WT SpAin1-GFP or mutant SpAin1(R216E)-GFP from low- (81nmt1) and medium- (41nmt1) strength promoters integrated at the leu locus. (A–D) Fluorescence micrograph montages of the time course of Rlc1-tdTomato–labeled contractile ring formation and constriction (right), corresponding to the boxed regions in DIC images (left). Cytokinetic stages are marked with red numbers as indicated in Supplemental Figure S7C: 1, Rlc1-tdTomato appearance (start); 2, ring assembly complete; 3, ring constriction begins; 4, ring constriction ends; 5, Rlc1-tdTomato disappears. Scale bars: 5 μm. (E) Average duration of ring assembly. Error bars, SD; n ≥ 10 cells. *p < 0.05. (F) Average duration of ring constriction. Error bars, SD; n ≥ 10 cells. *p < 0.05.

FIGURE 8:

The role of dynamic SpAin1 during contractile ring assembly. (A and B) Cartoon model for SpAin1-mediated contractile ring assembly and constriction. (A) Condensation of pre ring nodes (red and blue circles) into a contractile ring by the SCPR model in a wild-type cell (Vavylonis et al., 2008). Actin filaments assembled from nodes in random orientations are aligned into antiparallel bundles by SpAin1 (α-actinin) (i), which allows filaments to be efficiently captured by myosin-II motors on adjacent nodes (ii). The dynamic nature of SpAin1 allows myosin II to pull actin filaments without impediment, promoting proper ring constriction (iii). (B) Contractile ring formation in a cell expressing the less dynamic bundling mutant SpAin1(R216E). Actin filaments are elongated from nodes in random orientations (i), but get stalled as a result of excessive bundling, preventing actin filaments from reaching neighboring nodes (ii and iii). (C and D) Simulation of node capture at varying cross-linker affinities for F-actin (unitless K_d⁻¹). (C) Examples of node search and capture at low (0.3) cross-linker affinity (top) or high (3.0) cross-linker affinity (bottom). (D) Dependence of node capture rate (k_first-hit (s⁻¹)) over a range of cross-linker affinities (τ_x k_on) at filament elongation angles of 15° (black circles), 25° (red squares), or 35° (blue triangles) from the axis connecting the two nodes.