Spatial Coordination of Cytokinetic Events by Compartmentalization of the Cell Cortex

See allHide authors and affiliations

Science  16 Jul 2004:
Vol. 305, Issue 5682, pp. 393-396
DOI: 10.1126/science.1099892


During cytokinesis, furrow ingression and plasma membrane fission irreversibly separate daughter cells. How actomyosin ring assembly and contraction, vesicle fusion, and abscission are spatially coordinated was unknown. We found that during cytokinesis septin rings, located on both sides of the actomyosin ring, acted as barriers to compartmentalize the cortex around the cleavage site. Compartmentalization maintained diffusible cortical factors, such as the exocyst and the polarizome, to the site of cleavage. In turn, such factors were required for actomyosin ring function and membrane abscission. Thus, a specialized cortical compartment ensures the spatial coordination of cytokinetic events.

In animals and fungi, cytokinesis begins with the assembly of the actomyosin ring (13), which starts contracting during anaphase. Consequently, the cleavage furrow ingresses until it contacts the central spindle. Subsequently, daughter cells, which are still connected by the mid-body (2, 3), undergo abscission; the plasma membrane is resolved into two distinct membranes, and the daughter cells are irreversibly separated (4). Septins, a family of conserved GTPases that assemble into membrane-associated filaments (57), are among the proteins that localize to both the cleavage furrow and the mid-body (8, 9). However, their function in cell cleavage is unclear. In Saccharomyces cerevisiae, septins form a filamentous ring at the neck of the growing bud. During the G2 phase, this ring acts as a diffusion barrier to maintain polarity factors in the bud (5, 10, 11) and as a scaffold to recruit actomyosin ring components (12). At cytokinesis, the septin ring splits (13), and the actomyosin ring contracts between the two septin rings (12). We investigated the role of yeast septins during cytokinesis.

We first asked whether septins act as a scaffold to recruit factors to the bud neck during cytokinesis, as they do in G2. Factors recruited to the bud neck at cytokinesis include the polarizome, a complex involved in actin filament nucleation and signaling (14); the exocyst, involved in exocytosis (15); and chitin synthase II, a transmembrane protein required for primary septum formation (16). Cells coexpressing green fluorescent protein (GFP)-labeled Spa2 (polarizome), Sec3 (exocyst) or Chs2 (chitin synthase II) and cyan fluorescent protein (CFP)-labeled septin Cdc3 were analyzed by three-dimensional fluorescence microscopy (17). None of Spa2, Sec3, and Chs2 colocalized with septins at any point of the cell cycle. In small- and medium-budded cells, septins formed an hourglass-shaped structure at the bud neck (Fig. 1A; fig. S1) (13); Spa2 and Sec3 localized to the bud cortex; but Chs2 was not expressed (Fig. 1A; table S1). In the large-budded cells where Spa2, Sec3, and Chs2 localized to the bud neck, the septin ring had split (Fig. 1, A and B; fig. S1, table S1) and the mitotic spindle, when visualized by CFP-tubulin coexpression, was disassembled. In these cells, Spa2, Sec3, and Chs2 localized not on, but between, the split rings, as do the actomyosin ring components Myo1 (myosin II), Hof1 (an actin-associated protein), and Iqg1 (IQGAP) (13, 18). Thus, during cytokinesis, most factors at the cleavage site did not colocalize with septins, which suggests that the split septin rings might not act as scaffolds for their localization. Instead, septin rings appeared to delineate the boundaries of the cleavage area.

Fig. 1.

Protein dynamics during cytokinesis. (A) Localization of Spa2-GFP and the septin CFP-Cdc3 during G2/M (upper lane) and cytokinesis (lower lane). Enlargements of the bud neck are shown (insets). (B) Localization of Iqg1-, Hof1-, Myo1-, Spa2-, Sec3-, and Chs2-GFP (green) and the septin CFP-Cdc3 (red). (C) Analysis of Spa2, Sec3, and Cdc12 dynamics at the bud neck using FRAP. The photobleached region is indicated (orange box). Fluorescence intensities are shown over time for the bleached (light green), and unbleached area (dark). (D and E) Analysis of Spa2 and Myo2 diffusion out of the cytokinetic area using fluorescence loss in photobleaching (FLIP). A region of the cell cortex (orange circle) was repetitively bleached, and fluorescence intensity at the bud neck was monitored. (F and G) Actomyosin rings of diploid cells, coexpressing Myo1-CFP and Myo1-YFP, were subjected to FRAP. Only YFP was bleached. The overlay between CFP and YFP (green), the signal in the CFP channel (red) are shown during cytokinesis (F) and G2 (G). Elapsed time is indicated starting right before bleaching. YFP intensities over time are shown as in (C). Scale bars, 2 μm. All movies and imaging methods are available (17).

Thus, we asked whether the septin rings delineated a separate cortical compartment. We defined a cortical compartment as a cortical region in which cortical proteins can diffuse but from which they cannot exit. To address this possibility, the mobility of Sec3-GFP, Spa2-GFP, and Chs2-GFP at the neck of cytokinetic cells was characterized using FRAP (fluorescence recovery after photobleaching) (17). For all three proteins, fluorescence recovery was rapid (6 s < t1/2 < 15 s; Fig. 1C; table S2), and for Sec3 and Spa2, it was comparable to that observed at the bud cortex during G2 (fig. S2, A and B; table S2). By contrast, septins showed very slow dynamics in cytokinetic cells (Fig. 1C; table S2). Thus, Spa2, Sec3, and Chs2 moved freely and independently of septins at the bud neck. Furthermore, no or little fluorescence loss was observed at the bud neck for either Spa2-GFP or Sec3-GFP when a cortical region in the mother or bud of cytokinetic cells was constantly bleached (Fig. 1D). In contrast, Myo2-GFP fluorescence was lost within 90 s in similar experiments (Fig. 1E). The myosinV, Myo2 shuttles between the bud neck and the cytoplasm during cytokinesis (19). Thus, Spa2 and Sec3 did not escape the neck region. Together these data suggested that the split septin rings delineate an independent cortical compartment.

FRAP experiments on Myo1-GFP, Hof1-GFP, and Iqg1-GFP indicated that none of them diffused within the cleavage area during cytokinesis (table S2). For example, in cells coexpressing Myo1-CFP and Myo1-YFP (yellow fluorescent protein), the entire ring remained visible in the CFP channel, while half of it was bleached in the YFP channel (Fig. 1, F and G). In cytokinetic cells, the bleached fraction of the actomyosin ring failed to recover fluorescence (Fig. 1F), even during ring contraction. In contrast, fluorescence recovery was rapid (t1/2 = 19 ±7 s, Fig. 1G) during G2. Thus, during cytokinesis the actomyosin ring is a self-contained, rigid structure, potentially due to the recruitment of actin cables and myosin II filament assembly.

The cdc12-6 septin allele was used to investigate whether split septin rings served as barriers to maintain diffusible molecules to the bud neck. This allele leads to rapid septin ring disassembly at 35°C (20) (fig. S3A, table S3). When cytokinetic cdc12-6 cells expressing either Spa2-GFP or Chs2-GFP were shifted to 35°C under the microscope (17), fluorescence rapidly decreased at the bud neck (Fig. 2A; table S5). Spa2-GFP foci started to leave the neck region (75% of the cells, N = 19; movie S2 and Fig. 2A) 2 to 3 min after the shift, i.e., after septin ring disassembly. In wild-type cells treated in the same manner, Spa2-GFP and Chs2-GFP remained at the bud neck until after cells separated (Fig. 2A, movie S1). Furthermore, maintenance of Spa2 and Sec3 to the bud neck required the split septin rings to remain continuous. Indeed, in rts1-Δ cells grown at 37°C, septin rings became discontinuous during cytokinesis (20) (Fig. 2B), and Spa2-GFP and Sec3-GFP, but not Myo1-GFP, failed to localize to the cleavage site (17). These markers localized to the bud neck of all post-anaphase wild-type cells treated similarly. Rts1 is the yeast B′ regulatory subunit of protein phosphatase 2A (PP2A) and targets it to the bud neck at cytokinesis (20). Thus, continuous septin rings were required for the maintenance of diffusible markers to the bud neck, which supports the idea that they acted as diffusion barriers.

Fig. 2.

Septin rings maintain diffusible bud neck proteins at the cleavage area. (A) Spa2 localization in wild-type and cdc12-6 cytokinetic cells after shift from permissive (22°C) to restrictive temperature (35°C). GFP-labeled dots moving away from the neck are indicated (open arrows). Elapsed time since the temperature shift is indicated. (B) Localization of Cdc12, Spa2, Sec3, and Myo1 in rts1-Δ (top) and wild-type cells (bottom) after 3 hours at 37°C. Coexpression of GFP-Nop1 (arrow), a nucleolar marker, and CFP-tubulin (asterisk) facilitated the identification of post-anaphase cells. (C) Myo1 localization and contraction in wild-type and cdc12-6 cells shifted to 35°C. Kymographs of ring contraction are shown. Genotypes and cell cycle stages are indicated. Where appropriate, an asterisk indicates the beginning of contraction.

Actomyosin ring components moved little during cytokinesis, which suggests that their maintenance might not require diffusion barriers. Accordingly, actomyosin rings remained stable when the septin ring was disrupted during cytokinesis (Fig. 2C, d and e; fig. S4, A and B, table S4) but, as reported (21, 22), not during G2 (Fig. 2C, a and b). In cdc12-6 cytokinetic cells, Myo1-GFP, Hof1-GFP, and Iqg1-GFP rings even contracted (Fig. 2C, d and e), indicating that they remained functional. However, in half of the cases contraction was significantly slower in the cdc12-6 mutant than in the wild*type (Fig. 2C, c and d). Furthermore, Myo1 rings frequently failed to disassemble after contraction (Fig. 2C, e; table S4). Thus, the situation is similar to that in fission yeast, C. elegans, and mammalian cells (1, 6, 23); namely, budding yeast septins were dispensable for the maintenance of actomyosin components to the cleavage site during cytokinesis. Instead, they facilitated actomyosin ring contraction and disassembly.

We next tested whether septin function is essential during cytokinesis. At 30°C, cdc12-6 cells rapidly lose septin structures during, but not before, cytokinesis (20). Wild-type and cdc12-6 cells expressing GFP-tagged septin Cdc3 were shifted from room temperature to 30°C (Fig. 3, A and B). Wild-type cells that underwent cytokinesis under the microscope separated within 5 to 10 min after ring splitting, as shown by the “nicking” movements of the separating cells (Fig. 3A). In contrast, 60% (N = 19) of the cdc12-6 cells, which rapidly lost septin structures when the ring split, never separated (Fig. 3B). Thus, septin rings are required during cytokinesis for either abscission or digestion of the primary septum.

Fig. 3.

Septins are required for the completion of cytokinesis. (A and B) Septin ring splitting and cell separation were monitored in wild-type (A) and cdc12-6 (B) cells shifted from 22°C to 30°C. The elapsed time since ring splitting is indicated. (C to E) Role of septins in neck closure, assayed by FLIP. Cells coexpressed cytosolic GFP and Myo1-GFP (arrow), which indicates cytokinetic progression. The photobleaching area is indicated (orange circle). (C and D) Wild-type cell before (C) and after (D) Myo1 contraction. (E) cdc12-6 cell after contraction. In (C), t = 0 was set to the end of the movie. In (D and E), t = 0 corresponds to actomyosin ring disassembly.

To differentiate between these possibilities, we assayed closure of the bud neck when septin rings were disrupted. Wild-type and cdc12-6 strains were constructed that coexpressed soluble GFP in the cytoplasm and Myo1-GFP as a cytokinetic marker; the cells were shifted to 35°C, and the loss of cytoplasmic fluorescence in the bud was monitored while a small region of the mother cell was bleached (Fig. 3, C to E). In wild-type cells, fluorescence decayed in the bud only when photobleaching was applied before actomyosin ring contraction (Fig. 3, C and D). Thus, in wild type, the bud neck closed when the actomyosin ring contracted and disassembled. In contrast, when photobleaching was applied on cdc12-6 cells after actomyosin ring contraction and disassembly, fluorescence decayed in the bud and the mother, with similar kinetics for 80% of the cells (Fig. 3E). Thus, septin rings were required for abscission when the actomyosin ring contracted.

A simple explanation for septin functions in cytokinesis could be that compartmentalized factors are required for both actomyosin ring contraction and abscission. Accordingly, genetic analysis indicated a cytokinetic role for the exocyst. The temperature-sensitive allele sec3-4, but not sec3-2 (24, 25), was synthetic lethal with both myo1-Δ and hof1-Δ, which suggests that Sec3 acts in parallel to Myo1 and Hof1. Furthermore, 50% of mitotic sec3-4 cells shifted to the restrictive temperature failed to undergo cytokinesis (Fig. 4A). Wild-type and sec3-2 cells separated properly. The sec3-4 allele did not affect septin organization (Fig. 4B) but slowed down actomyosin ring contraction (Fig. 4C). Furthermore, although the actomyosin ring completed contraction and disassembly in all sec3-4 cells, many of these cells never completed cytokinesis. Thus, the exocyst is required for both efficient actomyosin ring contraction and abscission, consistent with the requirement of vesicle fusion for plasma membrane extension during contraction (26) and for plasma membrane resolution (27). Therefore, maintenance of cortical factors to the site of cleavage is likely to be a major role of septins during cytokinesis.

Fig. 4.

The exocyst is required for actomyosin ring contraction and completion of cytokinesis. (A) Wild-type, sec3-4, and sec3-2 cells were arrested before mitosis with hydroxy urea (HU) (left) and released at 37°C for 3 hours (right). The percentage of cells in the indicated cycle stages is shown. (B) Septin localization (GFP-Cdc3) in cytokinetic cells of the indicated genotype. (C) Actomyosin localization and contraction in sec3-4 and wild type as in Fig. 2B.

Cell cleavage requires the convergence of independent molecular pathways involved in cortical contractility, vesicle fusion, and plasma membrane fission. Our data indicate that maintenance of the different molecules underlying these processes to the cleavage site is ensured by the compartmentalization of the cortex around the actomyosin ring. In fission yeast and mammalian cells, septins also localize to each side of the cleavage site and are required for late cytokinetic events (7, 28). Thus, compartmentalization of the cortex may represent a conserved role of septins during cytokinesis.

Supporting Online Material

Materials and Methods

Figs. S1 to S5

Tables S1 to S5

References and Notes

Movies S1 to S24

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article