Local and global Cdc42 guanine nucleotide exchange factors for fission yeast cell polarity are coordinated by microtubules and the Tea1–Tea4–Pom1 axis

ABSTRACT The conserved Rho-family GTPase Cdc42 plays a central role in eukaryotic cell polarity. The rod-shaped fission yeast Schizosaccharomyces pombe has two Cdc42 guanine nucleotide exchange factors (GEFs), Scd1 and Gef1, but little is known about how they are coordinated in polarized growth. Although the microtubule cytoskeleton is normally not required for polarity maintenance in fission yeast, we show here that when scd1 function is compromised, disruption of microtubules or the polarity landmark proteins Tea1, Tea4 or Pom1 leads to disruption of polarized growth. Instead, cells adopt an isotropic-like pattern of growth, which we term PORTLI growth. Surprisingly, PORTLI growth is caused by spatially inappropriate activity of Gef1. Although most Cdc42 GEFs are membrane associated, we find that Gef1 is a broadly distributed cytosolic protein rather than a membrane-associated protein at cell tips like Scd1. Microtubules and the Tea1–Tea4–Pom1 axis counteract inappropriate Gef1 activity by regulating the localization of the Cdc42 GTPase-activating protein Rga4. Our results suggest a new model of fission yeast cell polarity regulation, involving coordination of ‘local’ (Scd1) and ‘global’ (Gef1) Cdc42 GEFs via microtubules and microtubule-dependent polarity landmarks.

. Growth of scd1 low and scd1 low tea1∆ cells expressing wild-type Cdc2. Cell morphology and CRIB-3mCitrine localization in the indicated genotypes. These cells also express Lifeact-mCherry (not shown). Thiamine was added to repress nmt81:3HA-scd1 expression just after the 0 min time-point; therefore at early time-points, cells have relatively higher levels of Scd1 and thus more detectable CRIB at cell tips (see also Fig. 3A). Effects of scd1 repression are apparent from 270 min onwards. Note increased wide/round cell shape in scd1 low cells over time (top panels), and isotropic-like growth and extremely round shape in scd1 low tea1∆ cells (bottom panels). Binucleate cells in scd1 low tea1∆ are likely related to defects in cytokinesis caused by round shape. Some scd1 low tea1∆ cells are binucleate even before repression, but not all scd1 low tea1 ∆ cells become binucleate after repression (see also Fig. S3A and Methods). Bar, 10 µm. See also Movie 3.

Replica on YE5S+hygromycin
Replica on YE5S+G418 = scd1∆::hphMX6 gef1-EANA-3mCherry:kanMX6 = scd1∆::hphMX6 gef1∆::kanMX6 Cross: scd1∆::hphMX6 X gef1-EANA-3mCherry:kanMX6 Cross: scd1∆::hphMX6 X gef1∆::kanMX6 Tetrad: 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 B D E A Figure S3. Supporting data for restoration of polarized growth in scd1 low tea1∆ cells after further loss of gef1 function (A) Cell morphology in the indicated genotypes during exponential growth, after 24 hr scd1 repression. 3-BrB-PP1 was not added to cultures, and thus cdc2-asM17 was not inhibited. CRIB-3mCitrine signal is shown here as a marker for cell volume and cell nuclei. Note that tea1∆ cells are round and often binucleate (see Methods), while tea1∆ gef1∆ cells are more similar to gef1∆ and control cells. (B,C) Gef1-EANA-mCherry is a loss-of-function mutation. In septating cells, both wild-type Gef1-3mCherry (B) and mutant Gef1-EANA-3mCherry (C) localize to the division site (black arrowheads). Wild-type Gef1-3mCherry promotes Cdc42-GTP (CRIB-3mCitrine) accumulation at the division site during early stages of septation (red arrowhead), but Gef1-EANA-3mCherry does not (blue arrowhead). In later stages of septation, CRIB-3mCitrine at the division site is more diffuse and weak and does not correlate with Gef1-EANA-3mCherry (green asterisks); this later localization is known to be independent of Gef1 (Wei et al., 2016). (D, E) Tetrad analyses showing synthetic lethality of gef1-EANA-3mCherry with scd1∆ (D) and confirming synthetic lethality of gef1∆ with scd1∆ (E) (Coll et al., 2003). Spores were germinated on YE5S and replica-plated as indicated. Boxes indicate inferred position of non-viable double mutants. Bars, 10 µm.  is cytosolic, and during cell division, Gef1 localizes to the division site. In all cases, cells were grown in YE5S to mid-log phase and imaged under conditions that minimize stress (Mutavchiev et al., 2016;see Methods). In some cases, high exposures were used to confirm absence of Gef1 from cell tips; as a result, mitochondrial autofluorescence is apparent in images of Gef1-GFP and Gef1-3GFP. Numbers of cells with detectable Gef1 at interphase tips are shown below the representative images. (B) Gef1-3mCitrine localization in cells imaged within 10 min after adding to uncoated glass-coverslip dishes. Under these conditions, Gef1 is also cytosolic during interphase. This demonstrates that pretreatment of coverslips with soybean lectin (normally used for longer-term imaging; see Methods) does not alter Gef1 localization. Inset shows one interphase cell with Gef1 at cell tips (arrowhead). (C) Recruitment of Gef1-3YFP from the cytosol to cell tips after treatment with the microtubule depolymerizing drug thiabendazole (TBZ; 150 µg/ml). TBZ has off-target effects that lead to cell depolarization independently of disrupting microtubules (Sawin and Snaith, 2004). TBZ was added just after imaging the 0 min time-point. After TBZ treatment, Gef1-3YFP transiently localizes to cell tips (red arrowheads) and later localizes more weakly to patches on cell sides (yellow arrowheads), which move towards cell middle. (D) Recruitment of Gef1-3mCherry from the cytosol to cell tips after TBZ treatment in tea4∆ cells. This demonstrates that Tea4 is not required for TBZ-induced Gef1 cell-tip localization. In (C) and ( Fig. 5D. A small percentage of tea1+ cells have detectable CRIB at cell tips, but this is much lower than in tea1-2FKBP12 cells (see also panels in (C)). Differences were highly significant (p<0.0001; Fisher's exact test). (C) Cell morphology and CRIB-3mCitrine localization in the indicated genotypes after 16 hr treatment with rapamycin or DMSO. Note polarized cell shape and CRIB localization to cell tips in rapamy-   Figure 6, but to avoid photobleaching, no images were acquired prior to those shown here. This demonstrates that membrane-associated Pom1-as-tdTomato is much more homogeneously distributed after inhibition by 3-BrB-PP1 in these cells and that loss of signal from cell tips is not simply due to photobleaching. (B) Deletion of gef1 restores polarized growth to scd1 low pom1∆ cells. Movie time-points showing cell morphology and mCherry-Bgs4 distribution in indicated genotypes. scd1 expression was repressed 24 hr before imaging. 3-BrB-PP1 was added 30 min before imaging. Diagrams show cell outlines at beginning and end of movies; outlines were aligned slightly to account for limited cell movement. Bars, 10 µm. See also Movie 8.  Figure S7. Single-channel images of Pom1-tdTomato, Pom1-as1-tdTomato and Rga4-3GFP after 3-BrB-PP1 addition. Single-channel images corresponding to the merged images shown in Figure 6C. Note that at some time-points, some Rga4-GFP signal appears to internal (i.e. not on the plasma membrane). Bars, 10 µm.  Figure S8. Polarity defects in rga4∆ scd1 low cells during extended interphase are rescued by gef1∆. (A) Cell morphology, mCherry Bgs4-localization and CRIB-3mCitrine distribution in the indicated genotypes after 3-BrB-PP1 treatment. scd1 expression was repressed for 24 hr before addition of 3-BrB-PP1. Cells were imaged 5 hr after addition of 3-BrB-PP1. CRIB-3mCitrine signal shows cell dimensions and was used to measure cell width in C. Note that rga4∆ cells in scd1 low background are wider/rounder than other genotypes, although polarity defects are not as strong as in scd1∆ background (see Figure 7A). (B) Quantification of mCherry-Bgs4 at cell tips in the indicated genotypes, from experiments in A. (C) Cell width for the indicated genotypes from images as in A, 5 hr after addition of 3-BrB-PP1. Median and interquartile ranges are shown. All pairwise differences were highly significant (p<0.0001; Mann-Whitney test), except gef1∆ vs. rga4∆ gef1∆ (p=0.20). n indicates number of cells scored. Bar, 10 µm. Bgs4 on the plasma membrane indicates sites of growth. Cells were pretreated with 3-BrB-PP1 60 min prior to start of imaging, to inhibit Cdc2 kinase activity, and then treated with either DMSO or MBC at start of imaging (still in presence of 3-BrB-PP1). For DMSO treatment, cell at lower right corresponds to cell shown in Fig. 2A. For MBC treatment, cell at mid-lower center corresponds to cell shown in Fig. 2A. Time interval during acquisition, 10 min; total elapsed time, 420 min; time compression at 15 frames per second playback, 9000X.
Movie 2. When scd1 is expressed at very low levels, tea1∆ and tea4∆ cells show PORTLI growth. mCherry-Bgs4 distribution and cell morphology in control cells, tea1∆, and tea4∆ cells, all in a scd1 low cdc2-asM17 mCherry-bgs4 genetic background. Bgs4 on the plasma membrane indicates sites of growth. Cells correspond to those shown in Fig. 3B. scd1 expression was repressed by thiamine addition 24 hr prior to start of imaging. Cdc2 kinase activity was inhibited by 3-BrB-PP1 addition 30 min before imaging. Time interval during acquisition, 10 min; total elapsed time, 420 min; time compression at 15 frames per second playback, 9000X.
Movie 3. Growth of scd1 low and scd1 low tea1∆ cells expressing wild-type Cdc2. CRIB-3mCitrine distribution and cell morphology in exponentially-growing scd1 low and scd1 low tea1∆ cells expressing wild-type Cdc2. Thiamine was added to repress scd1 expression just after the first time-point, which is paused in the movie. Cells correspond to those shown in Fig. S2. Effects of scd1 repression, including isotropic-like growth of scd1 low tea1∆ cells, become apparent about half-way through the movie. See Fig. S2 legends and Methods for further details. Time interval during acquisition, 9 min; total elapsed time, 540 min; time compression at 15 frames per second playback, 8100X.
Movie 4. Loss of gef1 function restores polarized growth to scd1 low tea1∆ cells. mCherry-Bgs4 distribution and cell morphology in tea1∆, gef1∆, tea1∆ gef1∆ and tea1∆ gef1-EANA cells, all in a scd1 low cdc2-asM17 mCherry-bgs4 genetic background. Bgs4 on the plasma membrane indicates sites of growth. Cells correspond to those shown in Fig. 4A. Movie 5. Gef1-3YFP is transiently recruited to the cell tips upon TBZ treatment. TBZ was added just after the first time-point, which is paused in the movie. Prior to TBZ addition, Gef1-3YFP in dividing cells is present at the division site and in the cytoplasm, and Gef1-3YFP in interphase cells is uniformly distributed in the cytoplasm, without any visible enrichment at the cell tips. Upon TBZ addition, interphase Gef1-3YFP signal is transiently observed at cell tips and later appears to move along the cell cortex towards the cell middle. Three of the cells in the movie correspond to those shown in Fig. S4C. Time interval during acquisition, 9 min; total elapsed time, 81 min; time compression at 15 frames per second playback, 8100X.
Movie 6. Rapamycin addition to gef1-Frb-GFP tea1-2FKBP12 cells leads to recruitment of Gef1-Frb-GFP to cell tips and increased cell polarization. Gef1-Frb-GFP localization and cell morphology in tea1-2FKBP12 cells and in control cells expressing untagged Tea1 (tea1+), all in scd1∆ tor2-S1837E fkh1∆ genetic background, after addition of rapamycin or control DMSO. Cells correspond to those shown in Fig. S5A. Rapamycin or DMSO were added just after the sixth time point. Time interval during acquisition, 5 min; total elapsed time, 150 min; time compression at 15 frames per second playback, 4500X Movie 7. Inhibition of Pom1 kinase activity in scd1∆ cells leads to PORTLI growth. Cell morphology and distribution of Pom1-tdTomato and GFP-Bgs4, or Pom1-as1-tdTomato and GFP-Bgs4, in scd1∆ cdc2-asM17 genetic background after 3-BrB-PP1 treatment. Bgs4 on the plasma membrane indicates sites of growth. Cells correspond to those shown in Fig. 6A. 3-BrB-PP1 inhibits activity of both Cdc2-asM17 and Pom1-as1-tdTomato and was added just after the first time-point. Note that 3-BrB-PP1 treatment depolarizes Pom1-as1-tdTomato, and this leads to PORTLI growth. Time interval during acquisition, 20 min; total elapsed time, 240 min; time compression at 15 frames per second playback, 18,000X.
Movie 8. Deletion of gef1 restores polarized growth to scd1 low pom1∆ cells. mCherry-Bgs4 distribution and cell morphology of pom1∆ and pom∆ gef1∆ mutants in scd1 low cdc2-asM17 mCherry-bgs4 genetic background. Bgs4 on the plasma membrane indicates sites of growth. Cells correspond to those shown in Fig. S6B. scd1 expression was repressed by thiamine addition 24 hr prior to start of imaging. Cdc2 kinase activity was inhibited by 3-BrB-PP1 addition 30 min before imaging. Time interval during acquisition, 10 min; total elapsed time, 350 min; time compression at 15 frames per second playback, 9000X.
Movie 9. Deletion of rga4 leads to PORTLI growth in scd1∆ cells. mCherry-Bgs4 distribution and cell morphology of scd1∆ and scd1∆ rga4∆ mutants in cdc2-asM17 mCherry-bgs4 background. Bgs4 on the plasma membrane indicates sites of growth. Cells correspond to those shown in Fig. 7A, with slightly larger fields. Cdc2 kinase activity was inhibited by 3-BrB-PP1 addition 30 min before imaging. Time interval during acquisition, 12 min; total elapsed time, 480 min; time compression at 15 frames per second playback, 10,800X. Click here to Download Table S1