TOR complex 2 localises to the cytokinetic actomyosin ring and controls the fidelity of cytokinesis

Abstract

The timing of cell division is controlled by the coupled regulation of growth and division. The target of rapamycin (TOR) signalling network synchronises these processes with the environmental setting. Here, we describe a novel interaction of the fission yeast TOR complex 2 (TORC2) with the cytokinetic actomyosin ring (CAR), and a novel role for TORC2 in regulating the timing and fidelity of cytokinesis. Disruption of TORC2 or its localisation results in defects in CAR morphology and constriction. We provide evidence that the myosin II protein Myp2 and the myosin V protein Myo51 play roles in recruiting TORC2 to the CAR. We show that Myp2 and TORC2 are co-dependent upon each other for their normal localisation to the cytokinetic machinery. We go on to show that TORC2-dependent phosphorylation of actin-capping protein 1 (Acp1, a known regulator of cytokinesis) controls CAR stability, modulates Acp1-Acp2 (the equivalent of the mammalian CAPZA-CAPZB) heterodimer formation and is essential for survival upon stress. Thus, TORC2 localisation to the CAR, and TORC2-dependent Acp1 phosphorylation contributes to timely control and the fidelity of cytokinesis and cell division.

Keywords: Acp1; CAPZA; Myosin II; Myosin V; Rictor; Schizosaccharomyces pombe; TOR; TORC2.

Fig. 1.

TORC2-deficient mutants display defects in cytokinesis and cell fission. (A) Early exponential wild-type (wt) and Rictor^ste20Δ (ste20Δ) prototroph cells were stained with Calcofluor to visualise the division septa. Scale bar: 10 µm. These cells, as well as sin1Δ cells and cells treated with Torin1, were (C) processed for flow cytometry analysis to measure DNA content. Mean cell width (B) and nuclear diameter (D) were each determined from counting 300 wild-type and Rictor^ste20Δ cells in mid-log culture. ***P<0.001, ns, not significant (Student's t-test). (E) A montage of time-lapse images showing red fluorescence from myo2–mCherry sid4–tdTomato cells undergoing cell division. Cells illustrating timing of CAR formation and constriction in relation to SPB segregation in wild-type cells are shown. (F,G) Equivalent montages of time-lapse images of mCherry (magenta) and GFP (green) fluorescene from myo2–mCherry cut12–gfp Rictor^ste20Δ cells. In a large proportion of anaphase Rictor^ste20Δ cells the CAR was seen to either collapse (arrowhead, F) or drift along the cortex towards one end of the cell (arrows, G). Scale bars: 5 µm.

Fig. 2.

TORC2 interacts with and localises to the CAR during ring constriction. (A) Summary of mass spectrometry analysis of three independent experiments of control (Ctrl, no anti-Tor1 antibodies added) and Tor1 immunoprecipitations (Exp), each purified from 20-l cell cultures. (B) Micrographs of RICTOR^Ste20 (green) and Myo2 or Myo51 (magenta) signal from mitotic RICTOR^Ste20–3GFP myo2–mCherry cells or RICTOR^Ste20–3GFP myo51–mCherry cells. (C,D) Maximum projections of 21-slice z-stacks from a timecourse of mitotic myo2–mCherry sid4–tdTomato RICTOR^Ste20–3GFP cells reveals that RICTOR^Ste20 foci (green) are recruited to the cell equator after SPB (magenta) separation and Myo2 ring (magenta) formation has occurred (frame 1–4 with non-separated SPBs represent interphase cells) (C) and coalesce to form a ring during CAR constriction. (D) Micrograph of mCherry and GFP signal from a RICTOR^Ste20–3GFP myo2–mCherry cell illustrating TORC2 association with the CAR. In B and C, arrowheads highlight RICTOR^Ste20 localisation at cell tips. (E) The timing of key events during CAR formation and constriction [(I) Myo2 foci (empty circles) recruit to the cell equator; (II) Myo2 foci coalesce to form a CAR (filled circles); (III) the CAR constricts until, (IV) it reaches a diameter of 0.5 µm or less] were determined in relation to nuclear division (crosses, distance between SPBs) in wild-type. (F) The timing of the appearance of Myp2 ring (red bar) and RICTOR^Ste20 medial foci (light blue bar) or RICTOR^Ste20 ring recruitment (dark blue) were determined in wild-type strains in relation to the events determined in E. Scale bars: 10 µm.

Fig. 3.

Myosin II and V interact with and regulate RICTOR^Ste20 localisation at the CAR. (A) Extract and subsequent anti-GFP and anti-HA (control) immunoprecipitates from S. pombe cells expressing a GFP-tagged Myo51 cargo-binding-tail domain fusion protein were subject to anti-Tor1 (upper panels) and anti-GFP (lower panels) antibody western blot analysis. The asterisks denote background bands. (B–E) Micrographs of mCherry and GFP signals in cells with the indicated genotype. The asterisk highlights rings split in two. Arrowheads highlight RICTOR^Ste20 localisation at cell tips. (F) The timing of the appearance of RICTOR^Ste20 medial foci (light blue bars) or RICTOR^Ste20 ring recruitment (dark blue) were determined in wild-type and myp2Δ strains in relation to the events determined in Fig. 2E. Scale bars: 5 µm.

Fig. 4.

Myp2 and TORC2 localisation to the actomyosin ring is co-dependent. (A) The timing of Myp2 ring recruitment (red bars) was determined in wild-type and Rictor^ste20Δ strains in relation to the events determined in Fig. 2E. (B) Composite micrographs of YFP and phase signal in yfp–myp2 Rictor^ste20+ and yfp–myp2 Rictor^ste20Δ cells. (C–E) Kymographs generated from 30 maximum projections of time-lapse images (3 min/frame) of (C) myo2–mCherry YFP–myp2 and (D) myo2–mCherry YFP–myp2 Rictor^ste20Δ cells illustrating that TORC2 is required for Myp2 to remain at the CAR. (E) Time-lapse kymographs of the perpendicular (upper panels) and longitudinal (lower panels) axes of a yfp–myp2 Rictor^ste20Δ cell in which Myp2 signal is lost from the cell equator and the Myo2-containing CAR slides along the cell cortex before constricting. In C–E, cartoons illustrate orientation and origin of kymograph axes. Scale bars: 5 µm (B); 1 µm (C–E).

Fig. 5.

TORC2 regulates CAPZA ^Acp1–CAPZB^Acp2 heterodimer formation. (A) Alignment of the phosphorylated region of S. pombe Acp1 and the human CAPZA homologues. Conserved residues are highlighted in bold and the two TOR-dependent phosphorylation sites are shown in red. The position of the conserved phosphorylated serine 208 (in yellow) is shown on the crystal structure of CAPZA and CAPZB (Yamashita et al., 2003). (B–D) Anti-GFP (upper panels) and anti-HA (lower panels) antibody western blots of CAPZA^Acp1−HA immunoprecipitations (HA-IPs) from indicated strains in the absence or presence of the TOR inhibitor Torin1 (25 µM) (TORC1^R=tor2.G2037D TORC2^R=tor1.G2040D). (B–E) To ensure the measured relative levels of Acp1 and Acp2 within anti-HA immunoprecipitates were quantifiable, western blot membranes of immuno-precipitates were cut in half and the separate portions were subsequently probed with anti-GFP (upper panels) and anti-HA (lower panels) antibodies.

Fig. 6.

CAPZA^acp1-AA mutants disrupt actin dynamics and cytokinesis. (A) The timing of the appearance of Myp2 ring (red bars), RICTOR^Ste20 foci (light blue bars), Acp1 (yellow bars) medial recruitment and RICTOR^Ste20 ring recruitment (dark blue) were determined in wild-type, myp2Δ, Rictor^ste20Δ (ste20Δ) and acp1Δ strains in relation to the events determined in Fig. 2E. Recruitment of these proteins to the CAR were observed in less than 40% (*) or 50% (**) of these deletion strains. CAR dynamics and composition were followed in >20 cells for each strain. Early exponential prototrophs were used in each experiment. (B) mCherry kymographs generated from 100 timeframe maximum projections from 13 z-plane images of CAPZA^acp1+ Lifeact-mCherry (left panels) and CAPZA^acp1-AA Lifeact-mCherry (right panels) cells (0.6 s/frame). (C) Graph showing lifetime kinetics of Lifeact signal from individual (faint lines) actin patches and overall averages (thick lines) of CAPZA^acp1+ Lifeact-mCherry (black lines) and CAPZA^acp1-AA Lifeact–mCherry (red lines) cells. (D) Maximum projections of a mixture of CAPZA^acp1–GFP Lifeact–mCherry and CAPZA^acp1–AA Lifeact–mCherry cells. Overlaying the GFP (green) and mCherry (magenta) signals demonstrate the increase in actin signal at cortical actin patches in the CAPZA^acp1-AA mutant compared to GFP-labelled wild type (arrows) cells. (E) Histograms illustrating mean±s.d. relative Cdc8 at the CAR in wild-type (WT) and CAPZA^acp1-AA cells (n>30 per strain).

Fig. 7.

TORC2 and CAPZA^acp1-AA mutants disrupt CAR and actin localisation. (A) Micrographs of Myo2 (magenta) and YFP (green) signal CAPZA^acp1-AA–HA myo2–mCherry YFP–myp2 cells. (B) Micrograph of mCherry signal from Rictor^ste20Δ Lifeact–mCherry cells illustrating the cytokinesis defect and lack of polarised actin signal. Scale bar: 10 μm. (C) The CAPZA^Acp1 phosphorylation site mutant CAPZA^acp1−AA is sensitive to heat stress at 37°C.