1. Cell Biology
Download icon

Scaffold-mediated gating of Cdc42 signalling flux

  1. Péter Rapali
  2. Romain Mitteau
  3. Craig Braun
  4. Aurèlie Massoni-Laporte
  5. Caner Ünlü
  6. Laure Bataille
  7. Floriane Saint Arramon
  8. Steven P Gygi
  9. Derek McCusker  Is a corresponding author
  1. University of Bordeaux, CNRS, European Institute of Chemistry and Biology, IBGC, UMR 5095, France
  2. Harvard Medical School, United States
Short Report
  • Cited 18
  • Views 1,299
  • Annotations
Cite this article as: eLife 2017;6:e25257 doi: 10.7554/eLife.25257

Abstract

Scaffold proteins modulate signalling pathway activity spatially and temporally. In budding yeast, the scaffold Bem1 contributes to polarity axis establishment by regulating the GTPase Cdc42. Although different models have been proposed for Bem1 function, there is little direct evidence for an underlying mechanism. Here, we find that Bem1 directly augments the guanine exchange factor (GEF) activity of Cdc24. Bem1 also increases GEF phosphorylation by the p21-activated kinase (PAK), Cla4. Phosphorylation abrogates the scaffold-dependent stimulation of GEF activity, rendering Cdc24 insensitive to additional Bem1. Thus, Bem1 stimulates GEF activity in a reversible fashion, contributing to signalling flux through Cdc42. The contribution of Bem1 to GTPase dynamics was borne-out by in vivo imaging: active Cdc42 was enriched at the cell pole in hypophosphorylated cdc24 mutants, while hyperphosphorylated cdc24 mutants that were resistant to scaffold stimulation displayed a deficit in active Cdc42 at the pole. These findings illustrate the self-regulatory properties that scaffold proteins confer on signalling pathways.

https://doi.org/10.7554/eLife.25257.001

Introduction

Scaffold proteins are multivalent binding proteins that assemble and regulate the fidelity of signal transduction cascades, conferring properties ranging from signal amplification, specificity, localisation and switch-like behaviour (Langeberg and Scott, 2015). However, the mechanisms by which scaffolds control signalling pathways spatially and temporally are not well understood.

The budding yeast Saccharomyces cerevisiae provides a useful model for understanding the spatio-temporal control of scaffold signalling, since the scaffold protein Bem1 plays an important, yet incompletely defined role in the Cdc42-dependent establishment of polarised growth that the organism displays during the cell cycle. Early in the cell cycle, Cdc42 is activated at a specific site on the plasma membrane by the GEF Cdc24. Site-specific activation of Cdc42 involves landmark proteins and the scaffold Bem1, which binds Cdc24 and Cdc42 GTP (Peterson et al., 1994; Zheng et al., 1995; Bender et al., 1996; Bose et al., 2001; Butty et al., 2002; Yamaguchi et al., 2007). The interaction of the scaffold with the active GTPase and its upstream GEF led to the idea that Bem1 might constitute part of a positive feedback loop: activated Cdc42 could bind Bem1, recruiting the GEF to activate additional Cdc42, thus building up a local pool of the active GTPase (Butty et al., 2002; Johnson et al., 2011). This autocatalytic positive feedback loop underpins numerous mathematical models that have been developed to understand the mechanisms that establish polarity (Goryachev and Pokhilko, 2006, 2008; Howell et al., 2012; Savage et al., 2012; Jose et al., 2013).

The scaffold Bem1 also interacts with the PAK Cla4, which phosphorylates Cdc24 on multiple sites (Gulli et al., 2000; Bose et al., 2001; Wai et al., 2009). While an initial study found that phosphorylation of Cdc24 may reduce its interaction with Bem1 (Gulli et al., 2000), a parallel study did not reach the same conclusion (Bose et al., 2001). Phosphorylation of Rho-family GEFs in Ustilago maydis and Schizosaccaromyces pombe regulates their function by inducing GEF degradation or sequestration by 14-3-3 proteins, respectively (Frieser et al., 2011; Das et al., 2015). However, the function of Cdc24 phosphorylation in budding yeast was elusive, since the mutation of multiple phosphorylation sites did not generate an obvious phenotype in vivo (Wai et al., 2009). Recently, biochemical analysis reported that phosphorylation of Cdc24 reduced its GEF activity, providing a source of negative feedback during Cdc42-mediated polarisation, although the underlying mechanism is unknown (Kuo et al., 2014). Thus, Bem1 has been proposed to play roles in positive and negative feedback that regulate Cdc42 activity. However, it is not presently understood how Bem1 mediates these effects. Indeed, while previous studies did not find a role for Bem1 in Cdc24 GEF activity (Zheng et al., 1995), a recent report using a biosensor of Cdc42 activity in vivo proposed that Bem1 may boost Cdc24 GEF activity rather than acting as part of the autocatalytic positive feedback loop. It is presently unknown whether this scaffold effect on GEF activity is direct, or how this effect might contribute to Cdc42 dynamics (Smith et al., 2013).

Here, we reconstitute Cdc42 activity to directly test the role of the scaffold Bem1 in Cdc42 regulation. Our results suggest that the scaffold Bem1 gates signalling flux through Cdc42 by activating GEF activity in a reversible manner. This effect of the scaffold impinges upon active Cdc42 levels at the cell pole and the global organisation of cellular polarity.

Results

Bem1 directly increases the rate of Cdc24 GEF activity towards Cdc42

Cdc42 activity was reconstituted by purifying the GTPase from the yeast Pichia pastoris (Figure 1—figure supplement 1A, left panel), while Bem1 and Cdc24 were purified from bacteria as full-length proteins (Figure 1—figure supplement 1A, right panel). Working towards the goal of reconstituting Cdc42 activation, we next assessed the capacity of the eukaryotic-expressed Cdc42 to bind fluorescent mant-GDP and mant-GTP. Mant-GDP (Figure 1—figure supplement 1B) or mant-GTP (Figure 1—figure supplement 1C) were incubated with increasing concentrations of purified nucleotide-free Cdc42, and FRET between the mant moiety and Cdc42 was measured (Figure 1—figure supplement 1B and C insets). By fitting a simple binding model to the maximal fluorescence intensity changes over a range of Cdc42 concentrations, estimated Kd of 1 µM and 0.5 µM were obtained for mant-GDP and mant-GTP, respectively, in accordance with previous reports (Figure 1—figure supplement 1B and C) (Zhang et al., 2000).

The rate of GEF-dependent Cdc42 activation was quantified by monitoring the displacement of GDP from Cdc42 in the presence of mant-GTP and unlabelled GMP-PNP, a non-hydrolysable GTP analogue. GDP to GTP exchange was measured both by the fluorescence intensity change of mant-GTP upon Cdc42 binding and by FRET between Cdc42 and mant-GTP. While both approaches gave comparable results, the FRET approach provided a higher signal-to-noise ratio. While neither Cdc42 alone nor Cdc42 and Bem1 displayed significant exchange of GDP for mant-GTP, the addition of Cdc24 provided robust GEF activity to Cdc42 (Figure 1A, blue line). The addition of Bem1 to the GEF reaction accelerated the rate of Cdc24 GEF activity further (Figure 1A, red line). The effect of Bem1-accelerated GEF activity was concentration dependent. Bem1 increased the observed rate constant of Cdc24 by 1.8x when present at 3 µM, and by 2.6x when present at 5 µM (Figure 1B). The C-terminal PB1 domain of Bem1, which interacts with the C-terminal PB1 domain of Cdc24 (Ito et al., 2001), was required for the stimulation of Cdc24 GEF activity, since a Bem1 mutant lacking this domain failed to augment GEF activity (Figure 1—figure supplement 1D and E). However, the PB1 domain of Bem1 alone was not sufficient to increase GEF activity (Figure 1—figure supplement 1D and E). These results demonstrate that Bem1 can directly increase Cdc24 GEF activity towards Cdc42 by interacting with the PB1 domain of Cdc24.

Figure 1 with 1 supplement see all
The scaffold Bem1 directly stimulates Cdc24 GEF activity in a reversible manner via PAK-dependent phosphorylation.

(A) Fluorescence intensity change associated with the nucleotide exchange of GDP-Cdc42 for mant-GTP Cdc42. Fluorescence was measured after the addition of GDP-Cdc42 (9 µM) to reactions containing Mant-GTP (100 nM) and GMP-PNP (100 µM) in the absence (blue curve) and presence (red curve) of Bem1 (5 µM) and Cdc24 (60 nM). (B) Observed kinetic rate constants were obtained by fitting trace data to a single exponential equation. Error bars show SD and confidence where *p<0.05 and ***p<0.001. (C) In vitro kinase reactions in which the indicated proteins were incubated with Cdc24-6xHis. At the times indicated, samples were removed and analysed by SDS-PAGE and Western blotting using anti-His antibody to detect the electrophoretic mobility shift of Cdc24-6xHis. (D) A Western blot showing the phosphorylation of Cdc24 in the samples used for subsequent GEF assays. (E) The observed kinetic rate constants of Cdc24 GEF activity obtained by fitting trace data to a single exponential equation. The reactions indicated on the right had additional Bem1 added to 5 µM.

https://doi.org/10.7554/eLife.25257.002

Bem1 increases the rate and extent of Cdc24 phosphorylation by the PAK Cla4

Since Bem1 interacts with the PAK Cla4, which phosphorylates Cdc24 (Gulli et al., 2000; Bose et al., 2001), we next tested if Bem1 influenced the rate of Cdc24 phosphorylation by Cla4. Kinase reactions were established in vitro to follow the kinetics of Cdc24 phosphorylation. While Cdc24 incubated with kinase assay buffer alone showed no phosphorylation-dependent electrophoretic mobility shift, addition of Cla4 resulted in a marked Cdc24 electrophoretic mobility shift in which three resolved forms of Cdc24 were apparent (Figure 1C top and middle panel). The inclusion of Bem1 in the kinase reaction accelerated the rate at which Cdc24 phosphorylation occurred, concomitant with the appearance of additional phosphorylated species of Cdc24 (Figure 1C bottom panel and Figure 1—figure supplement 1F). The electrophoretic mobility shift exhibited by Cdc24 in the presence of Cla4 and Bem1 in vitro was comparable to the electrophoretic mobility shift at the time of bud emergence in vivo, where Cdc24 is most abundantly phosphorylated (McCusker et al., 2007). These results indicate that Bem1 accelerates the rate and extent of Cdc24 phosphorylation by Cla4, generating hyperphosphorylated Cdc24.

Phosphorylation of Cdc24 inhibits the scaffold-dependent increase in Cdc24 GEF activity in vitro

Our results identified three effects of Bem1 on Cdc24: Bem1 augments Cdc24 GEF activity, while also increasing the rate and extent of GEF phosphorylation by the PAK Cla4. Importantly, a recent study reported that phosphorylated Cdc24 displays reduced GEF activity (Kuo et al., 2014); however, this was interpreted to resolve the competition for polarity factors during polarity establishment, rather than being part of a Bem1-dependent regulatory mechanism. Scaffold-mediated augmentation of Cdc42 GTP production by the GEF may activate PAK activity to hyper-phosphorylate and subsequently inhibit the GEF, serving as a self-regulating mechanism. We therefore tested the effect of Cdc24 phosphorylation on scaffold-dependent Cdc24 GEF activity. Different samples of Cdc24 were generated: Cdc24 alone, Cdc24 phosphorylated by Cla4 and Cdc24 hyper-phosphorylated by Cla4 in the presence of Bem1. The Cdc24 samples incubated with kinase displayed a quantitative shift in electrophoretic mobility compared to the control, non-phosphorylated sample, indicating that the majority of Cdc24 is phosphorylated in these samples. In addition, the Cdc24 sample incubated with Cla4 and Bem1 migrated more slowly than that incubated with Cla4 alone and the electrophoretic mobility shifts in Cdc24 were dependent upon the addition of ATP (Figure 1D). The samples of Cdc24 were next assayed for GEF activity. Cdc24 that had been phosphorylated by Cla4 showed similar GEF activity to Cdc24 alone. However, Bem1 no longer stimulated GEF activity when Cdc24 was phosphorylated (Figure 1E). Importantly, the ability of Cla4 to antagonise Bem1 stimulation of Cdc24 GEF activity was dependent upon the presence of ATP in the reaction, indicating that Cla4's effect on GEF activity is mediated by phosphorylation rather than through a protein-protein interaction. Consistent with our previous observations, addition of Bem1 to 5 µM augmented the GEF activity of the non-phosphorylated Cdc24 sample; however, the GEF activity of the phosphorylated or hyper-phosphorylated Cdc24 were resistant to further Bem1 stimulation (Figure 1E).

Phospho-regulation of Cdc24 impacts polarised growth in vivo

Our in vitro reconstitution experiments revealed that Bem1 stimulated Cdc24 GEF activity, while also attenuating this activation by inducing Cdc24 phosphorylation via the PAK Cla4. These results predict that non-phosphorylated Cdc24 would be active; as it would be constantly amenable to Bem1-stimulation, while constitutively phosphorylated Cdc24 would be less active, since it would be resistant to Bem1 stimulation. To begin to address these predictions, we first mapped phosphorylated residues in Cdc24, before mutating these sites in vivo. Samples of Cdc24 were quantitatively phosphorylated in vitro by Cla4 alone or by Cla4 in the presence of Bem1, and then analysed by mass spectrometry. Cla4 phosphorylated 39 sites on Cdc24, of which 15 were only detected in the presence of Bem1. Since Cdc24 is also a target of Cdk1 (Moffat and Andrews, 2004; McCusker et al., 2007), we also mapped these sites after incubating Cdc24 with Cdk1-Cln2. This identified an additional seven sites. Thus, in total, we mapped 46 phosphorylation sites on Cdc24, while a previous study mapped 35 sites in vivo, 23 of which overlapped with those that we identified (Figure 2—figure supplement 1A) (Wai et al., 2009).

The mapped phosphorylation sites were mutated to alanine or aspartate and the constructs, together with a wild-type control, were integrated at the endogenous CDC24 locus in vivo. Homologous recombination generated a phospho-mutant in which all 46 sites were mutated to alanine (cdc24-46A) and a phospho-mimetic mutant in which 28 sites were mutated to aspartate (cdc24-28D) (Figure 2—figure supplement 1B). We also generated mutants in which only the 15 Bem1-dependent phosphorylation sites in CDC24 were mutated to alanine or aspartate. We refer to these mutants as cdc24-15A and cdc24-15D. We analysed Cdc24 protein mobility by Western blotting in the mutants and control cells during the cell cycle. Wild-type Cdc24 is phosphorylated at time zero due to the cell synchronisation with mating pheromone (Gulli et al., 2000). A portion of Cdc24 is then dephosphorylated until 15–30 min, at which point it undergoes additional phosphorylation at around 60 min after release from the G1-arrest, around the time at which cells form a bud (Figure 2A). Cell-cycle-dependent phosphorylation of Cdc24 was greatly diminished in the cdc24-46A mutant, indicating that many of the phospho-sites mapped by mass spectrometry in vitro are relevant in vivo. Conversely, the cdc24-28D mutant displayed constitutively reduced mobility by SDS-PAGE during the cell cycle that was similar to that of hyperphosphorylated wild-type Cdc24. The cdc24-15A and −15D mutants showed an intermediate level of phosphorylation between wild type and the cdc24-46A/ cdc24-28D mutants. In keeping with our in vitro biochemical observations, the cdc24-46A and the cdc24-15A mutants appeared to be active, since cells did not display an obvious defect in the rate of colony formation at 37°C compared to control cells, while the cdc24-28D and the cdc24-15D mutants exhibited a strong defect in colony formation at 37°C (Figure 2B). The growth defect of cdc24-28D cells at 37°C was not due to degradation of the cdc24-28D protein, consistent with the idea that the mutations had not induced non-specific cdc24 mis-folding (Figure 2—figure supplement 1C). The temperature-dependent growth defect of the cdc24-28D mutant cells at 37°C was partially suppressed by the addition of sorbitol to the growth media. Sorbitol acts as an hyperosmotic stabilising agent and has previously been used to support the viability of cdc24 mutants, which are defective in the polarised secretion of cell wall material (Figure 2—figure supplement 1D) (Bender and Pringle, 1989).

Figure 2 with 1 supplement see all
Phospho-regulation of Cdc24 is required for normal cellular polarity.

(A) Phosphorylation of Cdc24-3xHA, cdc24-46A-3xHA, cdc24-28D-3xHA, cdc24-15A-3xHA and cdc24-15D-3xHA during the cell cycle. Cells were synchronised in mating pheromone and samples were removed at the times shown (top) and analysed by SDS-PAGE and Western blotting using anti-HA antibody to detect Cdc24 and anti-Fip1 antibody as a loading control. (B) cdc24-28D and cdc24-15D cells display a temperature-sensitive growth defect at 37°C. Serial dilutions of the cells indicated were spotted onto YPD plates and grown at 25°C and 37°C for 2 days. (C) The cells indicated were grown to mid-logarithmic phase at 25°C then fixed and stained with Alexa 546-phalloidin to visualise F-actin. DIC images are shown in the top and F-actin on the bottom panels. The numbers indicate the percentage of cells displaying polarised F-actin where more than 150 cells were counted for each strain. Scale bar: 2 µm. (D) The phenotype of the cdc24 phosphorylation mutants is dependent on BEM1. Serial dilutions of the cells indicated were spotted onto YPD plates and grown at the temperatures indicated for 2 days. (E) DIC images showing the effect of deletion of BEM1 in the cdc24 phosphorylation mutants. Scale bar: 2 µm.

https://doi.org/10.7554/eLife.25257.006

We also characterised polarised growth during the cell cycle, which is dependent upon cell cycle cues emanating from Cdk1 and Cdc42 activity (Moseley and Nurse, 2009; McCusker and Kellogg, 2012). After synchronous release into the cell cycle, samples were periodically removed and scored for the appearance of small buds, an indicator of polarised growth. By 60 min after release from the arrest, ~60% of wild-type cells had formed buds, while ~35% cdc24-46A mutants and only ~25% cdc24-28D mutants had formed buds, consistent with a defect in polarised growth due to defective regulation of Cdc24 phosphorylation (Figure 2—figure supplement 1E left panel). While it is presently unknown why the cdc24-46A mutant displays a delay in bud emergence, the reduced rate of bud emergence in the cdc24-28D mutant did not appear to stem from the mutation of Cdk1-dependent sites in CDC24, since the cdc24-15D mutant in which only Bem1-dependent Cla4 sites were mutated also displayed a delay in bud emergence (Figure 2—figure supplement 1E right panel). Inspection of the morphology of the cdc24-28D mutant cells at 25°C revealed a significant proportion of cells that were large and unbudded, or displayed mis-shapen buds, indicative of defects in polarised growth due to aberrant Cdc24 phospho-regulation. The defects in morphology of the cdc24-28D mutant cells were reflected in their cell cycle doubling times (CDC24 99 min, cdc24-46A 111 min and cdc24-28D 140 min at 25°C in YPD medium). The morphological defects were also evident after F-actin visualisation (Figure 2C). While the actin cytoskeleton was polarised in ~75% of control and cdc24-46A cells, this was reduced to 46% in cdc24-28D cells. Moreover, the effect of the two cdc24 phospho-mutant alleles was Bem1-dependent in vivo, since deletion of BEM1 resulted in indistinguishable phenotypes in CDC24, cdc24-46A and cdc24-28D strains (Figure 2D and E).

We next examined the localisation of Cdc24 phospho-mutant proteins. As previously reported, wild-type Cdc24 localises to the nucleus in early G1 of the cell cycle, whereupon cell cycle cues trigger its export to the cytoplasm and localisation to the site of polarity establishment (Figure 3A)(Toenjes et al., 1999; Nern and Arkowitz, 2000; Shimada et al., 2000). Importantly, the cdc24-46A protein was enriched at the site of polarity establishment compared to the wild-type protein (Figure 3A and B), while the cdc24-28D, although polarised, displayed less enrichment at the pole and prominent nuclear localisation at points in the cell cycle when the wild-type protein had been exported from the nucleus to the cytoplasm (Figure 3A and B). Cdc24 is sequestered in the nucleus during early G1 of the mitotic cell cycle via its interaction with Far1 (Nern and Arkowitz, 2000; Shimada et al., 2000), a Cdk1 inhibitor that is degraded late in G1 by rising cyclin-Cdk1 activity and the proteasome, thus releasing Cdc24 into the cytoplasm (Peter et al., 1993; Henchoz et al., 1997). We have not yet successfully deleted FAR1 in the cdc24 phospho-mutant background to test the effect on the phenotype of the cdc24 phospho-mutants. The expression of the mutants and wild-type protein were comparable (Figure 3—figure supplement 1A). The mutation of the phosphorylation sites did not appear to result in complete Cdc24 delocalisation, but rather altered its dynamics at sites to which the wild-type protein normally localises. Consistently, Cdc42 was also enriched at the cell pole in the cdc24-46A mutant (Figure 3C). As a more direct test of whether Cdc42 GTP levels were altered in the cdc24 phosphorylation mutants, we imaged gic21-208-yEGFP, which includes a CRIB motif that has previously been used as a marker for Cdc42 GTP (Tong et al., 2007). Consistent with the localisation of Cdc24 and Cdc42, we found higher levels of the Cdc42 GTP reporter in the cdc24-46A and the cdc24-15A mutants compared to control cells, while the reporter was less enriched at the pole in cdc24-28D and cdc24-15D mutants (Figure 3D).

Figure 3 with 1 supplement see all
Phospho-regulation of Cdc24 is required for normal GEF localisation at the cell pole in vivo and Bem1-stimulated GEF activity in vitro.

(A) Maximum projection images of deconvolved z-stacks. Scale bar: 5 µm. (B) The average fluorescence intensity signal of Cdc24 at the cell pole was plotted for 48–66 cells of the indicated strains. Error bars show SD and confidence where ****p≤0.0001. (C) Maximum projection images of deconvolved z-stacks showing mEOS-Cdc42 fluorescence and quantification of the average fluorescence intensity. The significance of error bars is the same as in (B). (D) Maximum projection images of deconvolved z-stacks showing gic21-208-yEGFP fluorescence and quantification of the average fluorescence intensity. The significance of error bars is the same as in (B) and (C). (E) Observed kinetic rate constants of Cdc24 GEF activity for cdc24-46A and cdc24-28D +/− Bem1. Note how Bem1 stimulates the rate of cdc24-46A GEF activity, but not that of cdc24-28D. Error bars show SD and confidence where *p<0.05 and **p≤0.01. (F) DIC images showing the effects of RGA1 deletion on the indicated strains. Scale bar: 5 µm.

https://doi.org/10.7554/eLife.25257.009

These results indicate that the phospho-blocking and phospho-mimetic mutations confer distinct phenotypes and are thus likely to affect Cdc24 in a specific manner. As a means of corroboration, we tested the GEF activity of purified cdc24-46A and −28D proteins in the presence and absence of Bem1. While both mutants exhibited similar basal GEF activity, Bem1 was able to stimulate the cdc24-46A mutant, but not that of the cdc24-28D mutant (Figure 3E and Figure 3—figure supplement 1B). The lower effect of Bem1 on cdc24-28D did not appear to stem from markedly reduced affinity of cdc24-28D for Bem1, since similar levels of Bem1 associated with Cdc24, cdc24-46A and cdc24-28D, as quantified from pull-down experiments (Figure 3—figure supplement 1C and D). In summary, our in vivo data indicate that the cdc24-46A mutant that displays reduced phosphorylation is viable and is enriched at the cell pole, together with its cognate GTPase Cdc42. Conversely, the cdc24-28D mutant, which displays constitutively reduced electrophoretic mobility, results in a temperature-sensitive phenotype in vivo, reduced recruitment of the mutant protein to the cell pole and aberrant cell polarity. We reasoned that the increased recruitment of cdc24, Cdc42 and the Cdc42 GTP reporter to the cell pole in the cdc24-46A mutant might reflect increased Cdc42 GTP production, but that the mutant might not exhibit morphological defects because Cdc42 GAPs may buffer the GTPase module. We therefore deleted the GAP RGA1 in control and cdc24 phospho-mutants to test whether loss of GAP buffering would result in morphological defects. Indeed, cdc24-46A mutants were sensitive to a reduction in GAP activity compared to CDC24 and cdcd24-28D cells. The cdc24-46A rga1△double mutants were highly elongated, consistent with excessive Cdc42 GTP production, while the cdc24-28D rga1△ mutant cells were not elongated (Figure 3F). Consistently, the deletion of Cdc42 GAPs and CLA4 also results in highly elongated cells (Caviston et al., 2003). These in vivo results support our biochemical model derived from in vitro reconstitution experiments in which phosphorylation of Cdc24 attenuates the scaffold-mediated augmentation of GEF activity.

Discussion

Scaffold proteins contribute to dynamic aspects of signalling, including signal amplitude, feedback and localisation. Here, we show that the scaffold Bem1 in budding yeast plays a direct role in stimulating GEF activity during Cdc42 activation. This spike in Cdc42 GTP production is likely to contribute to the increase in Cdc42 activation that ensues during polarity establishment in vivo (Figure 4) (Gulli et al., 2000; Wai et al., 2009; Atkins et al., 2013; Smith et al., 2013). The augmentation of GEF activity occurs in a reversible fashion, since the scaffold increases the rate of GEF phosphorylation by the PAK Cla4, which attenuates scaffold-mediated GEF activation. Thus, phosphorylation may provide the GEF with something akin to a molecular memory of previous scaffold encounters, generating a pulse of GEF activity that contributes to flux through the Cdc42 GTPase module. It will be important to identify the phosphatase that antagonises Cdc24 phosphorylation by the PAK, Cla4, and to understand the mechanisms that determine the kinase-phosphatase activity during polarity establishment.

A working model depicting scaffold-mediated gating of Cdc42 signalling flux.
https://doi.org/10.7554/eLife.25257.013

Previous studies in U. maydis and S. pombe reported that phosphorylation regulates Cdc42 GEFs by promoting its degradation, or by its 14-3-3-dependent sequestration, respectively (Frieser et al., 2011; Das et al., 2015). While a previous study found that the PAK Cla4 played a positive role in polarity establishment in S. cerevisiae (Kozubowski et al., 2008), subsequent work found that phosphorylation of Cdc24 by Cla4 resulted in reduced GEF activity, while increased Cdc42 recruitment to the cell pole was observed in Cdc24 phosphorylation mutants (Kuo et al., 2014; Wu et al., 2015). Our study supports these latter observations, additionally linking phosphorylation to a scaffold-dependent mechanism that promotes the localised activation of Cdc42 via Cdc24 regulation, an enigmatic problem in the field (Smith et al., 2013; Woods et al., 2015). We found that phosphorylation of Cdc24 appeared to reset GEF activity to a basal state, rather than inhibiting it altogether, potentially allowing the system to remain responsive to activation from other signalling sources. Previous studies had not reported strong morphological defects when Cdc24 phosphorylation sites were mutated (Wai et al., 2009; Kuo et al., 2014). This may stem from our observation that Cdc42 GAPs play an important role in counteracting the increased Cdc42 GTP levels that accumulate when Bem1 stimulates unphosphorylated Cdc24. Presently, it is unknown if Bem1 augments Cdc24 GEF activity by tethering constituents of the GTPase module, hence increasing the local concentration of Cdc42 GDP and Cdc24, or by acting allosterically on the GEF; however, the report that Bem1 interacts preferentially with Cdc42 GTP rather than Cdc42 GDP would be consistent with the latter model (Bose et al., 2001). It will also be important to test whether any scaffold that increases the local Cdc24-Cla4 concentration would stimulate Cdc24 phosphorylation, or whether Bem1 exerts specific effects.

Our observation that the scaffold directly stimulates GEF activity is consistent with two recent reports. First, that scaffold-deleted cells follow an evolutionary pathway in which selective pressure down-regulates Cdc42 GAP activity (Laan et al., 2015). We propose that the evolutionary-driven down-regulation of GAP activity is likely to counteract the diminution in Cdc42 GTP levels that would ensue from loss of the scaffold. Second, it was reported that scaffold-deleted cells display reduced activity of a Cdc42 GTP biosensor in vivo (Smith et al., 2013). Our results are consistent with these observations and indicate that Bem1 is able to directly stimulate Cdc24 GEF activity towards Cdc42.

Materials and methods

Plasmid construction

Request a detailed protocol

BEM1, bem1ΔPB1 (residues 1 to 466) and bem1-PB1 (residues 469 to 550) were amplified by PCR with NdeI and XhoI restriction sites and cloned into a modified pGEX6P-2 vector, in which the BamHI site in the multiple cloning site was changed to NdeI. This generated pDM256, pDM516 and pDM502, respectively, in which the Bem1 constructs were tagged at the N-terminus with GST that could be subsequently removed by digestion with rhinovirus 3C protease.

CDC24 was amplified by PCR with NdeI and XhoI sites and ligated into pET21a to generate pDM272, in which Cdc24 is C-terminally tagged with 6xHis. The nucleotide sequence encoding amino acids 75–854 were synthesised with HindIII and KpnI restriction sites (Bio-Basic, Markham, Canada). Four constructs were synthesised: one in which the 46 serines or threonines mapped by mass spectrometry were mutated to alanine (pDM615), one in which the 46 phosphosites were mutated to aspartic acid (pDM616), one in which only the 15 Bem1-specific phosphosites were mutated to alanine (pDM780) and one in which the 15 Bem1-specific sites were mutated to aspartic acid (pDM781). The HindIII-KpnI fragments were cloned into pDK51, a URA3 yeast vector enabling the in-frame fusion of the cdc24 mutant fragments with three copies of the HA epitope to generate pDM630, pDM632, pDM782 and pDM784 (Carroll et al., 1998). Oligonucleotides with homology to CDC24 were then used to generate cdc24-46A-3xHA::URA3, cdc24-46D-3xHA::URA3, cdc24-15A-3xHA::URA3 and cdc24-15D-3xHA::URA3 PCR products that were transformed into yeast. The cdc24 mutants were sequenced in the resulting yeast transformants, indicating that homologous recombination at the CDC24 locus had generated a cdc24-46A-3xHA mutant (DMY2151), a cdc24-28D-3xHA mutant (DMY2154), a cdc24-15A-3xHA mutant (DMY2333) and a cdc24-15D-3xHA mutant (DMY2134). Since the cdc24-46D mutant was not recovered, this mutant may not be viable. The cdc24-46A and cdc24-28D mutants were amplified by PCR from genomic DNA of DMY2151 and DMY2154 yeast strains, respectively, introducing NotI and XbaI sites, in addition to a ribosome-binding site. These PCR products were cloned into pET21a, generating pDM645 and pDM647, respectively, in which the cdc24 phospho-mutants were C-terminally tagged with 6xHis.

To monitor the in vivo localisation of the Cdc24 phospho-mutants, CDC24, cdc24-46A and cdc24-28D were amplified by PCR from genomic DNA prepared from DMY2147, DMY2151 and DMY2154 yeast, introducing HindIII and SpeI sites. The PCR products were cloned into a modified pRS416 plasmid, consisting of a CYC1 promoter between KpnI and SalI sites. Subsequently, mEOS, in addition to an amino-terminal GAGAGG linker and a carboxyl-terminal GAGAG linker-6xHis-tag, were added to the C-terminus of the constructs using SpeI and NotI sites. Finally, an ADH1 terminator sequence was introduced using a NotI site, generating pDM700, pDM701 and pDM704 for CDC24, cdc24-46A and cdc24-28D, respectively.

To monitor the in vivo level and localisation of Cdc42 GTP in the control and cdc24 phospho-mutant strains, we generated a gic21-208-yEGFP fusion, containing a CRIB motif that was cloned into a modified pRS413 plasmid consisting of a CYC1 promoter (pDM842) (Curran et al., 2014). Full-length wild-type CDC42 was amplified with oligonucleotides that introduced EcoRI and AgeI sites, in addition to an N-terminal 6xHis-6xGly linker and a Strep-tag II sequence. The PCR product was cloned into a modified pPICZa vector (Thermo Fisher Scientific, Waltham, USA) to generate pDM401 for expression of the fusion from the AOX1 promoter in Pichia pastoris. The pRS315 CDC42prom-mEOS-GAGA-CDC42 plasmid (pDM303) was constructed in three steps: the CDC42 promoter was cloned into pRS315 on a SalI-BamHI fragment. The CDC42 ORF and terminator were then ligated into this vector using an NdeI site that was engineered 5' of the BamHI site and a 3' SacII site. Finally, mEOS was ligated into the NdeI digested plasmid and checked for the correct orientation. All plasmids were verified by sequencing.

Yeast growth conditions

Request a detailed protocol

Yeast strains were generated and grown using standard genetic techniques. Cells were grown in selective or rich media, depending upon the experiment. For cell viability assays, yeast were grown in rich medium at 25°C until mid-log phase. Ten-fold serial dilutions were spotted onto YPD plates, which were incubated at the indicated temperatures for 2 days.

Protein expression and purification

Request a detailed protocol

Nucleotide-free and GDP-loaded 6xHis-Strep-tag-II-Cdc42 was purified in a two-step affinity purification method by Co2+-immobilised metal affinity chromatography (IMAC), followed by Strep-Tactin affinity chromatography. Ground yeast powder was generated using a Retsch PM100 grinding mill (Retsch, Haan, Germany) and a 125 ml steel bowl with 20 mm ball bearings chilled with liquid nitrogen. Greater than 85% lysis was typically observed after eight disruption cycles. For Cdc42 purification, room temperature lysis buffer (20 mM Tris-HCL (pH = 8.0), 1 M NaCl, 5% glycerol, 0.5% CHAPS supplemented with EDTA-free Protease inhibitor cocktail (Roche, Basel, Switzerland) and 1 mM PMSF) was added when the yeast powder showed signs of melting. The lysate was then stirred immediately for 10 min at 4°C and remained at 4°C for the subsequent purification. The cell lysate was centrifuged at 48,000 x g for 30 min in a JA25.5 rotor (Beckman, Brea, USA), and the supernatant was loaded into a Co2+-IMAC column. After extensive washing (20 mM Tris-HCL (pH = 8.0), 1 M NaCl, 5% glycerol and 0.05% Tween-20), 6xHis-Strep-tag-II-Cdc42 was eluted in Co2+-IMAC elution buffer (20 mM Tris-HCL (pH = 8.0), 150 mM NaCl, 5% glycerol and 250 mM imidazole and 0.05% Tween-20). In the second purification step, Strep-Tactin Superflow Plus (Qiagen, Venlo, The Netherlands) was added to the elution and agitated for 30 min at 4°C. The sample was loaded onto a column and then washed extensively with Strep-Tactin wash buffer (20 mM Tris-HCL (pH = 8.0), 150 mM NaCl, 5% glycerol and 0.05% Tween-20). Nucleotide-free Cdc42 was generated by washing the column with 50 column volumes of Strep-Tactin wash buffer supplemented with 25 mM EDTA, followed by an additional washing step using Strep-Tactin wash buffer. GDP-bound Cdc42 was generated by washing the column with Strep-Tactin wash buffer supplemented with 200 μM GDP and 10 mM MgCl2, then incubating for 1 hr at room temperature. The non-bound GDP and Mg2+ were removed in an additional washing step. GDP-loaded 6xHis-Strep-tag-II-Cdc42 was eluted in Strep-Tactin elution buffer (20 mM Tris-HCL (pH = 8.0), 150 mM NaCl, 5% glycerol supplemented with 5 mM d-Desthiobiotin and 0.05% Tween-20). Nucleotide-free Cdc42, which was produced for nucleotide binding experiments, was eluted in elution buffer before the GDP loading step. Protein concentration was estimated by absorption at 280 nm using ε = 21430 M−1cm−1 as the calculated molar excitation coefficient. Samples were aliquoted in 50 μl volume and flash frozen in liquid nitrogen for storage.

Active Cla4-TAP was purified from 4 L of budding yeast after lysis in liquid nitrogen using a rotary mill, as described for the Cdc42 purification. Chilled, lysed yeast powder was resuspended in three volumes of lysis buffer (10 mM Tris-HCL (pH = 8.0), 1 M NaCl, 0.1% IGEPAL, 50 mM NaF, 2 mM Na3VO4, 100 mM β-glycerol phosphate, 200 mM potassium acetate and 1 mM PMSF). After centrifugation at 110,000 x g for 90 min, the supernatant was incubated with 300 µl IgG Sepharose for 2 hr. The beads were loaded onto a 1 ml column and washed three times in 10 ml lysis buffer then in 10 ml TEV buffer (10 mM Tris-HCL (pH = 8.0), 150 mM NaCl, 0.1% IGEPAL, 0.5 mM EDTA and 1 mM DTT). TEV protease was then added overnight at 4°C. The eluate was diluted three-fold in calmodulin binding buffer (10 mM Tris-HCL (pH = 8.0), 150 mM NaCl, 0.1% IGEPAL, 1 mM magnesium acetate, 1 mM imidazole, 2 mM CaCl2, and 10 mM β-mercaptoethanol), then incubated with 300 µl calmodulin Sepharose for 2 hr. The column was washed three times in 10 ml calmodulin binding buffer then eluted in kinase assay buffer (50 mM Hepes (pH = 7.6), 2 mM MgCl2, 0.05% Tween-20, 1 mM DTT, 10% glycerol and 2 mM EGTA). The eluted kinase was dialysed in kinase assay buffer lacking EGTA then flash frozen in liquid nitrogen for storage.

Cdc24-6xHis and GST-tagged Bem1 were expressed in Bl21-CodonPlus (DE3) cells. Briefly, cells were grown in terrific broth at 37°C until an OD600 ~2 (Tartoff and Hobbs, 1987). Expression was induced by the addition of IPTG to 0.2 mM, after which cells were grown for 12 hr at 18°C. Cells were harvested and flash frozen in liquid nitrogen. For Cdc24-6xHis purification, room temperature lysis buffer (20 mM Tris-HCL (pH = 8.0), 1 M NaCl, 5 mM imidazole supplemented with EDTA-free Protease inhibitor cocktail (Roche, Basel, Switzerland) and 1 mM freshly prepared PMSF) was added to chilled, ground bacterial powder. The lysate was immediately stirred at 4°C for 10 min and sonicated for 30 s three times. The cell lysate was centrifuged at 48,000x g for 30 min, and the supernatant was loaded into a Co2+-IMAC column. After extensive washing (20 mM Tris-HCL (pH = 8.0), 1 M NaCl, 5 mM imidazole), Cdc24-6xHis was eluted (20 mM Tris-HCL (pH = 8.0), 150 mM NaCl, 500 mM imidazole). The protein was dialysed twice (20 mM Tris-HCL (pH = 8.0), 150 mM NaCl) and flash frozen in liquid nitrogen for storage.

For the Bem1 purification, cells were lysed by adding room temperature lysis buffer (50 mM Tris-HCL (pH = 7.5), 1 M NaCl, 0.1% Tween-20 and 5 mM DTT, an EDTA-free Protease inhibitor cocktail tablet and 1 mM freshly prepared PMSF) to ground bacterial powder. The lysate was processed by sonication and centrifugation in the same manner as the Cdc24 −6xHis lysate. The supernatant was loaded onto a glutathione column. The column was extensively washed (50 mM Tris-HCL (pH = 7.5), 250 mM KCl, 0.1% Tween-20 and 5 mM DTT), and then GST-Bem1 was eluted (50 mM Tris-HCL (pH = 7.5), 250 mM KCl and 5 mM reduced glutathione). For GST pull-down experiments, GST-Bem1 was dialysed twice against 20 mM Tris-HCL (pH = 7.5), supplemented with 150 mM NaCl and flash frozen in liquid nitrogen for storage. For enzymatic experiments, GST-Bem1 was dialysed extensively in protease buffer (20 mM Tris-HCL (pH = 7.5), 150 mM NaCl, 1 mM DTT and 1 mM EDTA), the GST was digested using rhinovirus 3C protease then removed by re-loading the protein onto a glutathione agarose column. The flow-through, which contained Bem1, was dialysed extensively in 20 mM Tris-HCL (pH = 7.5), 150 mM NaCl then flash frozen in liquid nitrogen for storage.

Cdc24 GEF assay

Request a detailed protocol

Förster resonance energy transfer (FRET) between Cdc42 and N-methylanthraniloyl-GTP (mant-GTP) was measured to monitor the Cdc24-mediated GDP to mant-GTP exchange reaction on Cdc42 in real time. Trp97 of Cdc42, which is in close proximity to the GTP binding site, was excited using 280 nm wavelength light using a 5 nm bandwidth. The FRET signal was detected at the emission peak of mant-GTP, at 440 nm using an 8 nm bandwidth. All fluorescence measurements were performed at 27°C on a Tecan Infinite M1000PRO plate reader (Tecan Group, Männedorf, Germany) in 384-well, non-binding microplates (Greiner Bio-One, Courtaboef, France), in a 10 μl reaction volume. The final buffer conditions were 20 mM Tris-HCl (pH = 8.0), 150 mM NaCl, 1 mM DTT, 5 mM MgCl2, 100 nM mant-GTP, supplemented with 100 μM GMP-PNP nucleotide to maintain a low concentration of the mant-GTP fluorophore. Cdc24 was used at a final concentration of 60 nM after 30 min room temperature pre-incubation with Bem1 at 5 µM. The reaction was started by adding GDP-bound Cdc42 to 9 µM final concentration and exchange was monitored for at least 2000 s with 15 s intervals. For each sample, a mock reaction was used in the absence of GDP-loaded Cdc42 to normalise for bleaching and to subtract possible sources of background noise such as Cdc24-mant-GTP interaction. The intrinsic GDP to mant-GTP exchange rate of Cdc42 was determined in the absence of Cdc24.

A single exponential equation (Equation 1) was fitted to each kinetic trace using GraphPad Prism 6 version 6.05 and the observed kinetic rate constants were compared.

(1) I=IMax+(IMaxIMin)(1ekobst),

where I is the fluorescence intensity change, IMax is the maximal fluorescence intensity, IMin is the minimal fluorescence intensity, kobs is the observed kinetic rate constant and t is the time in seconds.

All reactions were performed at least three times with proteins from two different purifications.

Mant-nucleotide binding

Request a detailed protocol

A 100 nM final concentration of mant-GDP or mant-GTP were titrated with increasing concentration of nucleotide-free Cdc42 in buffer containing 20 mM Tris-HCl (pH = 7.5),150 mM NaCl and 1 mM DTT at 27°C. The interaction was monitored for 30 min by FRET at 440 nm using 280 nm wavelength light for excitation. Maximal fluorescence changes (-) were obtained by fitting a single exponential onto each trace:

I=IMax+(IMaxIMin)(1ekobst),

where I is the fluorescence intensity change, IMax is the maximal fluorescence intensity, IMin is the minimal fluorescence intensity, kobs is the observed kinetic rate constant and t is the time in seconds.

Maximal fluorescence changes of each trace were plotted against the nucleotide-free Cdc42 concentration and the dissociation constants were determined by fitting a quadratic binding equation:

I=(Kd+[Cdc42]+[mant])(Kd+[Cdc42]+[mant])24[Cdc42][mant]2[mant],

where I is the maximal fluorescence change observed in each trace, Kd is the dissociation constant, [Cdc42] is the total Cdc42 concentration and [mant] is the total mant-GDP or total mant-GTP concentration.

All reactions were performed at least three times with proteins from two different purifications.

In vitro kinase assays

Request a detailed protocol

For the time-resolved kinase reactions shown in Figure 1C, a 100 nM final concentration of Cdc24 was incubated with 500 pM Cla4 in the presence or in the absence of 100 nM Bem1 in kinase buffer (20 mM Tris (pH = 8.0), 2 mM MgCl2, 10% glycerol and 1 mM ATP) at 30°C. SDS loading buffer was added to samples at the indicated times and analysed by SDS-PAGE and Western blotting using mouse anti-His antibody (Euromedex, Souffelweyersheim, France) and AlexaFluor532 goat anti-mouse IgG(H+L) (Life technology, Carlsbad, USA) as primary and secondary antibodies, respectively. Fluorescence was detected using a Typhoon TRIO+ scanner (GE Healthcare). The result was analysed by ImageQuant TL 1D version 8.1 software (GE Healthcare). The time-resolved kinase reactions were repeated at least three times.

In vitro kinase reaction for the GEF assay

Request a detailed protocol

A 600 nM final concentration of Cdc24 was treated with 3 nM Cla4 in the presence or absence of 2 μM Bem1 - under these conditions more than 80% of Cdc24 is estimated to exist in a complex with Bem1 - in kinase buffer (20 mM Tris (pH = 8.0), 2 mM MgCl2, 10% glycerol and 1 mM ATP) at 30°C for 2 hr. Phosphorylation of Cdc24 was validated by Western blot using anti-His antibody. A 60 nM final concentration of phosphorylated samples were used directly in the GEF assay, as described above.

Cell synchronisation, Western blotting and actin staining

Request a detailed protocol

For cell synchronisation, 0.5 μg/ml α-factor was added to mid-log phase yeast cultures at 25°C for 3 hr, as described previously (McCusker et al., 2007). Cells were then washed three times in YPD lacking pheromone for synchronised release into the cell cycle at 25°C. To determine the budding index, samples were removed at the times indicated, fixed by the addition of formaldehyde to 3.7%, washed with PBS then at least 200 cells were scored per time point to ascertain the percentage of budded cells.

To characterise the mobility of Cdc24 by SDS-PAGE during the cell cycle, 1.5 ml cells were removed at the indicated time points. After centrifugation, small glass beads were added to the pellet and the samples were flash frozen in liquid nitrogen. Samples were vigorously agitated with glass beads in 100 μl SDS loading buffer supplemented with 1 mM PMSF, 50 mM NaF and 75 mM β-glycerol phosphate. Samples were immediately boiled and analysed by SDS-PAGE and Western blotting. For actin staining, yeast were grown to early log phase, fixed and then stained with Alexa 546-phalloidin, as previously described (Jose et al., 2015).

GST-pull down

Request a detailed protocol

Equal amounts of Cdc24, cdc24-46A and cdc24-28D were added to excess amounts of GST or GST-Bem1 immobilised on glutathione agarose beads (Sigma Aldrich, St. Louis, UA). After incubation on ice for 1 hr, beads were washed extensively (20 mM Tris pH = 7.5, 150 mM NaCl and 0.1% Tween-20) and proteins were eluted (20 mM Tris pH = 7.5, 150 mM NaCl, 0.1% Tween-20 and 5 mM reduced glutathione). Proteins were then TCA precipitated, and analysed by SDS-PAGE and Western blotting. All pull-down experiments were repeated three times.

Phosphorylation site mapping

Request a detailed protocol

In order to identify phospho-sites on Cdc24, phospho-peptides of Cdc24 after trypsin digestion were enriched by strong cation exchange chromatography, followed by IMAC, and then analysed by mass spectroscopy. For kinase reactions, a 600 nM final concentration of Cdc24 was treated with 3 nM Cla4 in the presence or absence of 2 μM Bem1 in kinase assay buffer (20 mM Tris (pH = 8.0), 2 mM MgCl2, 10% glycerol and 1 mM ATP) in a 25 μl reaction volume at 30°C for 45 min. Under these conditions in the presence of Bem1, maximal Cdc24 phosphorylation was observed. After the kinase reaction, disulphide bonds in the proteins were reduced by adding DTT to a final concentration of 5 mM at 56°C for 25 min. Samples were then cooled to room temperature before alkylating cysteines at room temperature for 30 min in the dark in 14 mM of freshly prepared iodoacetamide. Excess iodoacetamide was quenched by adding DTT to 5 mM and incubating at room temperature for 15 min. In order to isolate phosphorylated Cdc24, 0.5 μg protein was analysed by 12% SDS-PAGE and Coomassie-stained Cdc24 bands were excised for trypsin digestion.

LC-MS/MS

Request a detailed protocol

Excised gel bands were sliced into approximately 1 mm cubes, and destained by incubation in 50% acetonitrile/50 mM ammonium bicarbonate at 37°C. Destained gel slabs were dehydrated by incubation in neat acetonitrile, and re-hydrated in digestion buffer (50 mM ammonium bicarbonate, 10 µg/ml trypsin). Digestions were incubated overnight at 37°C. The digestion supernatant was decanted to a fresh tube, and the gel pieces were washed two times with 50% acetonitrile/50 mM ammonium bicarbonate. The washes were combined with the digestion supernatant, and evaporated via SpeedVac (Thermo Fisher Scientific, Waltham, MA). Digests were resuspended in 5% formic acid/5% acetonitrile, and cleaned via c18 microextraction (Rappsilber et al., 2003). Samples were loaded onto a 40-cm fused silica column (100 µM ID) packed in-house with c18 resin (1.8 µm particle size) and analysed on an Orbitrap Fusion Tribrid mass spectrometer operating in data dependent mode, using a 90 min gradient over which acetonitrile in 0.1% formic acid was increased from 3% to 40%. Proteins present within each sample were identified by SEQUEST search against the yeast proteome (Uniprot). Peptide and protein results were filtered such that each contained less than 2% false positive identifications using a target decoy approach (Elias and Gygi, 2007). The list of proteins detected in each sample was used to construct a smaller database containing only these proteins, and the data were re-searched against this database using SEQUEST considering phosphorylation at serine, threonine, and tyrosine as variable modifications. Identifications were filtered to a false positive detection rate of no more that 5%. The discrete location of phosphorylation sites within each phosphopeptide detected were validated using the A-Score algorithm (Beausoleil et al., 2006).

Imaging and analysis of Cdc24-mEOS, mEOS-Cdc42 and gic21-208-yEGFP

Request a detailed protocol

In order to visualise the Cdc24 phospho-mutant proteins in vivo, the DMY570 strain, containing Cdc24 under the control of a galactose-inducible promoter as the sole source of Cdc24 was transformed with pDM700, pDM701 and pDM704 plasmids, encoding Cdc24-mEOS, cdc24-46A-mEOS and cdc24-28D-mEOS, respectively. To visualise Cdc42 in the cdc24 phospho-mutant strains, DMY2147, DMY2151 and DMY2154 were transformed with pDM303, encoding mEOS-CDC42. To detect Cdc42 GTP in the cdc24 phospho-mutant strains, DMY2147, DMY2151, DMY2154, DMY2333 and DMY2334 were transformed with pDM842, encoding gic21-208-yEGFP. Transformants were plated on selective minimal media containing dextrose and were grown at 25°C for imaging in mid-logarithmic phase.

All Images were analysed and processed using ImageJ software. To calculate the enrichment of Cdc24, Cdc42 and gic21-208 at the pole, the average fluorescence intensity of the pole (AFIP) and the cytosol (AFIC) were determined for each cell using an empirically determined threshold value that enabled the cell pole to be identified. Next, the AFIP was normalised as follows: normalised fluorescence intensity of the pole= (AFIP-AFIC)/AFIC). The values were plotted using Prism software.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
    Improved media for growing plasmid and cosmid clones
    1. K Tartoff
    2. C Hobbs
    (1987)
    Bethesda Research Laborotories Focus, 9.
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45

Decision letter

  1. Mohan K Balasubramanian
    Reviewing Editor; University of Warwick, United Kingdom

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

[Editors’ note: a previous version of this study was rejected after peer review, but the authors submitted for reconsideration. The first decision letter after peer review is shown below.]

Thank you for submitting your work entitled "Scaffold-mediated gating of Cdc42 signaling flux contributes to GTPase nanoclustering" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by Mohan Balasubramanian as Reviewing Editor and a Senior Editor. The reviewers have opted to remain anonymous.

Our decision has been reached after extensive consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife.

The referees agree that the reconstitution of Cdc42-Cdc24-Bem1, using Pichia expressed Cdc42, is well done and linking the phosphorylation status to self-limitation of Cdc42 activation (i.e. GEF activity of phosphorylated Cdc24 is not activated further by Bem1) is an exciting conclusion (with the proviso that there are a number of controls that need to be done). The genetic experiments and associated imaging with the phosphomimetic mutants are interesting, but need to be strengthened.

The problem is in Figures 6 and 7 (super-resolution microscopy) and this seems like a separate paper that has been appended to the rest of the work, which has technical and interpretational issues, all listed below.

In light of this, we will be unable to consider this paper further for publication in eLife. However, the referees believe that the in vitro reconstitution and the associated genetics could be submitted as a short-article (please take a look at eLife contribution types for more information). If you choose this approach, please address the substantive points raised by the referees pertaining to Figures 1-5 (most of which are control experiments /new experiments that should strengthen the findings). The referees’ comments are appended verbatim.

Reviewer #1:

In this manuscript, Rapali et al. study the regulation of an important complex that regulates cell polarization through Cdc42 GTPase. Cdc42 is activated by its GEF Cdc24, and also binds the scaffold Bem1. The authors show that Bem1 acts in vitro in two opposing ways: it promotes Cdc24 GEF activity towards Cdc42 and also increases the rate of phosphorylation of Cdc24 by the PAK kinase Cla4. Phosphorylated Cdc24 still activates Cdc42, but cannot be promoted by Bem1. This thus likely limits the positive effect that Bem1 has on Cdc42 activation, which the authors interpret as Bem1 contributing to signaling flux through Cdc42. They also identify all phosphorylation sites on Cdc24 and analyze the effects of phospho-mutants and phospho-mimetic mutants in vivo. In a second part, the authors use super-resolution imaging approaches to probe the diffusion and cluster-forming ability of Cdc42, probing both the function of Bem1 and of Cdc24 phosphorylation sites in these processes.

In my view, the first half of the manuscript is very interesting, clean and refreshing. The biochemistry is performed to high standards and largely convincing (a few missing controls are mentioned below). The interpretation that the contrasting roles of Bem1 increase flux is possible, but only one of several interpretations. For instance, another interpretation could be that Bem1 contributes to reaching rapidly a high, but not excessive Cdc42 activity. By contrast, the second half of the manuscript, including the analysis of Cdc24 phosphomutants and especially the super-resolution imaging, is less convincing and does not fit very well with the first half. It reads more like two stories lumped together. The consequence is that I do not feel the statement made by the title of the paper is well supported. Specific comments are below.

Regarding the effect of Cdc24 phosphorylation status on Bem1-dependent activity enhancement, one missing control in Figure 1C is to test the effect of 0.2 µM Bem1, which is the concentration remaining in the reaction in Figure 3A-B. The expectation is that it should not have an effect at this concentration also in absence of Cla4-mediated phosphorylation. Another important missing control is to test whether the effect of Cla4 is solely through phosphorylation or also through protein binding. This should be addressed by using a kinase-dead allele in reaction as in Figure 3C.

The interpretation of the Cdc24 phospho-mutants is complicated and difficult to link with the biochemistry shown in the first part of the paper. They appear to conclude from their in vivo analysis "in keeping with their in vitro biochemical observations" that the 46A is active and the 28D less active (subsection “Phospho-regulation of Cdc24 impacts polarized growth in vivo” and further). If this is "in keeping with the biochemical observations", the difference in activity in the two alleles should also be Bem1-dependent in vivo, not intrinsic to Cdc24 GEF activity (i.e., there should not be a difference between the two alleles in a bem1 mutant), but this is not tested. More generally, all effects of the 46A mutant are interpreted relative to Bem1-Cla4 function. However, this mutant also includes the CDK sites, and thus some effects, such as delay in budding may be due to CDK1 action. It is also unclear in the text how the 28D sites were selected. The text could also sometimes be clearer. For instance, the double-negation in paragraph five of the section is convoluted. What the authors may mean is that the phospho-blocking and phospho-mimetic mutations have distinct phenotypes and thus are likely to affect Cdc24 each in a specific manner. Clarification on all these points would help the reader.

The mobility shift of Cdc24 from synchronized cultures is also difficult to interpret. The authors claim that "wild type Cdc24 undergoes a dramatic reduction in electrophoretic mobility due to phosphorylation at around 60 minutes after synchronous release from a G1-arrest". However, the Figure 4A shows very high level of phosphorylation at time 0, slightly coming down and then up again around 60min. The description of the result thus does not correspond to what is shown on the figure. This is likely due to the α-factor treatment use for synchronization, which may cause polarization responses. For the analysis here, it would likely be much better to synchronize cells by elutriation to probe this without interfering with cell polarization. Secondly, the authors show the reduced mobility is indeed due to phosphorylation by performing phosphatase treatment of wildtype and mutant Cdc24 alleles.

The effect of Cdc24 phosphorylation on Cdc42 distribution and activity in vivo are not clear. The authors interpret the enrichment of Cdc24 and Cdc42 at the polar cap as sign of Cdc42 activity. This could be more directly evaluated by using probes for Cdc42-GTP, such as CRIB, which would be predicted to be in increased amounts in the 46A mutant, and decreased in the 28D mutant. Pull-down of Cdc42-GTP could also be performed to probe this point.

Regarding the super-resolution analysis of Cdc42, one major question I have is which version of mEos is being used? Some versions have the tendency to dimerize, which would make fusion proteins prone to oligomerization. For instance, mEos2 has been shown to still dimerize, which raises questions about mEos, from which mEos2 is derived. mEos3.1 and 3.2 appear to be truly monomeric. Because the authors do not have a Cdc42 mutant that blocks cluster formation, it is difficult to be very conclusive about the validity of the observed nanoclusters. I also feel that the link between nanocluster size and Cdc42 activity status is unclear and not directly supported by evidence. The 46A and 28D mutants show changes in Cdc42 cluster sizes at cell poles, which the authors interpret as a link between Cdc42 activity and cluster size, but (as mentioned above) they have not directly tested Cdc42 activity in these mutants in vivo. To probe for a link between activity and cluster size, the authors should look at the distribution of constitutively active and inactive Cdc42 mutants.

I have similar difficulties with the analysis of diffusion coefficients. The differences between wt and bem1∆ are also very slight, which does not help to affirm the position that the observed nanoclusters are physiologically relevant. I find for instance surprising that, in Figure 6A-B, the diffusion coefficients seem to be altered in bem1∆ both at the pole and in the rest of the cell compared to wildtype (although statistics between bem1∆ and wt are not provided, so I am not sure this is statistically supported). I would expect a faster diffusion at the pole of bem1∆ cells, as shown (but again not statistically tested), but not a slower diffusion elsewhere. More generally, the analysis of Cdc42 trajectories is not particularly convincing. To convince that there are actually four distinct classes (as opposed to a continuum), the authors should show a plot of frequencies versus α value, which should reveal four distinct peaks/modes. From the examples shown, I am for instance not convinced that there is any directed movement: the examples show only two data points in a line before a more confined-looking behavior.

Reviewer #2:

In this manuscript the authors study the role of the scaffold protein Bem1 and its interactions with the Cdc42 GEF Cdc24 and the PAK kinase Cla4, in the context of budding yeast polarization. The first part of the manuscript focuses on biochemical experiments with purified proteins and shows that Bem1 increases GEF activity of Cdc24 on Cdc42 and at the same time supports phosphorylation of Cdc24 by the PAK kinase Cla4. Interestingly, the authors then find that phosphorylated Cdc24 can no longer be stimulated by Bem1, providing a possible self-limiting loop to Cdc42 activation – a critical feature for efficient cell polarization that has remained elusive on the molecular level in all existing models of spontaneous symmetry breaking. The authors then confirm the relevance of the identified reactions in vivo, using a Cdc24 mutant lacking 46 MS-mapped phosphorylation sites. In particular, using single particle tracking of Cdc42 in cells they propose that Bem1 and Cdc24-phosphorylation contribute to dynamics of Cdc42 by affecting nanocluster formation at the polarization site.

The topic of this study connects central aspect of signal transduction and cell polarization research. Bem1 has been proposed to be a key factor for a positive feedback loop driving symmetry breaking and the earliest steps in yeast polarization. However, the specific mechanisms and consequences of Bem1-Cdc24-Cla4 interactions have been subject to a long and controversial debate in the field. The current study clarifies a few important molecular aspects of this debate in a convincing manner, and provides elegant in vivo support for their main biochemical findings. I therefore think that this work is of great relevance and provides the kind of combination between conceptual advance and technical innovation aimed for at eLife. However, there are a couple of issues that definitely need to be clarified or improved before publication (see major points below). There are also a couple of technical concerns that can be easily fixed.

1) The demonstrations of Bem1 effects on Cdc24 activity and phosphorylation are convincing but they mostly confirm or incrementally improve on previous studies. The idea of a self-limitation and negative effect of phosphorylation on GEF-stimulation is however a really new and therefore very important part of this study. As I interpret the current results the negative effect on Bem1-stimulation could either be due to the phosphorylation itself (as the authors suggest) or due to the presence of Cla4 (which is always present in the relevant assays from Figure 3). Can the GEF assays in 3B/C be performed with a kinase dead version of Cla4 as control? Alternatively, the kinase reaction could be performed without phosphate in the medium. This would support the idea of a central effect of the Cdc24 phosphorylation sites. In the same direction: Bem1 seems to stimulate additional phosphorylation of very specific sites (not just more but different pattern). It would good to include a mutant where only the Bem1-specifi sites are mutated and include this in the assays for Figure 3. Finally, it should be possible to combine Dcla4 and Drga1 to show that this leads to a similar phenotype as the 46A mutant + Dcla4 (this might have been already shown, in this case refer to the relevant paper).

2) The second key point of this study is the proposed effect of Bem1 on Cdc42 nanocluster formation. While I have no doubts that such clustering occurs for Cdc42 – as it probably does for every protein of the PM (Spira et al., NCB 2012) – the proposed link to Bem1 and phosphorylation is less convincing to me. My main concern stems from the single molecule localization analysis. The signal density on the cell surface (for both Cdc42 and Cdc24) seems very low, especially in the non-polarized regions. This could be due to the limitations of mEOS2 (permanent bleaching, low quantum yield) compared to chemical dyes used in STORM applications. How can the authors exclude that the cluster results are affected by labeling density (larger cluster in the denser polarized cap area)? This would perfectly correlate with the observed effects (increase at pole for 42A). An additional concern is the current resolution limit (40 nm), which is very close to the identified cluster size (76 nm). Considering additional factors such as mobility of the fluorophore (GAGA linker) and variable polarization of fluorescence (Shivanandan et al., FEBS letters 2014) clusters might not be readily resolvable any more. In summary, at the moment the shown PALM data definitely needs some sort of additional validation. Possible experimental controls with/without predicted clustering would be Cdc42 in LatA treated cells, constitutive active Cdc42 or markers of known clusters such a eisosomes (Pil1, Sur7) or endocytic patches (Ede1, Abp1).

3) A majority of the track in Figure 6 seem to occur at the edge of cells. If using HiLo (beter to use TRIRF) the focus should be on the actual cell surface so that all tracks can be followed in 2D. At the moment I see a bias to shorter tracks as the movement in z can no longer be tracked. Also the observed differences (pole vs. rest) are not very strong and the effect of Bem1 deletion seems to be rather increasing motility in non-polarized areas than decreasing the polar motility. To support the single particle tracking it would therefore be important to provide an additional approach. The authors could perform careful FRAP experiments in cap vs. non-polarized regions of the cell. This should allow detection of the observed effects. After all, the average MSD that was obtained is quite similar to the effective FRAP recovery rate.

4) The current data support a reduction in stimulation of Cdc24 by Bem1 after Cla4-mediated phosphorylation. To really support the sequence of events proposed in the model (Figure 8), Cla4 would need to act after Bem1 recruitment (fits the idea of Bem1 as recruitment platform for Cla4). Can the authors perform such an experiment with their reconstituted system? 1. Measure basal GEF activity, 2. Add Bem1 for 30 min, 3. Measure stimulated GEF activity, 4. Add Cla4 (high levels) for 1 h, 5. Measure GEF activity again to see basal levels again. This would provide a perfect final experiment and nicely relate to the proposed model.

Reviewer #3:

This manuscript investigates how the Cdc42 scaffold Bem1 regulates Cdc42 activation via positive and negative regulation in the budding yeast model system. Reports have shown that Cdc42 activation undergoes oscillations/fluctuations during polarization in the fission yeast (Das et al. 2012) and budding yeast (Howell et al., 2012) model systems. These findings indicated that Cdc42 undergoes both positive feedback as well as a time delayed negative feedback. Further Das et al., 2012 suggested that the negative feedback was mediated by the Cdc42 target Pak1 kinase and in Pak1 kinase mutants Cdc42 activation and GEF-scaffold complex (Scd1-Scd2) is increased at the cell tips leading to increased Cdc42 activation and dampening of Cdc42 pulsing at these sites. Later Kuo et al., 2014 demonstrated that Cdc24 is phosphorylated by the kinase Cla4 and this leads to a decrease in Cdc24 GEF activity. This manuscript provides further evidence that Cla4 phosphorylates Cdc24 and thus leading to dampening of GEF activity. The authors claim that Bem1 promotes Cla4 mediated phosphorylation of Cdc24 and describe this as a "Bem1 dependent mechanism". They also claim that this leads to desensitization of Cdc24 to Bem1 interaction. The evidence in support of these claims is not strong. Further, in support of previous reports, the authors use single particle tracking to show that the Cdc42 complex at the cell tips is less dynamic in the presence of Bem1. The authors suggest that Cdc24 phosphorylation decreases clustering of Cdc42 complexes at sites of polarization. While the findings presented here do not provide a novel insight into how Cdc42 is negatively regulated, it does provide additional evidence for previous findings.

1) The authors state that phosphorylation of Cdc24 reduces its activity, but do not directly test whether this is due to decreased affinity between Cdc24 and Cdc42. While the authors test the GEF activity of Cdc24 in various mutants, they fail to use the most direct (and biologically relevant) approach of measuring Cdc42 activity in vivo. Later they present phenotypes of these mutants, but only indirectly link them to Cdc42 by observing cell morphology and cytoskeleton organization (Figure 4C).

2) Previous reports (Kozubowski et al., 2008) have shown that the Pak kinase forms a complex with Cdc42-Cdc24-Bem1 to positively regulate Cdc42 activation and constitutes a positive feedback loop that leads to symmetry breaking. The authors have not discussed their findings in the context of this report. Moreover, Cdc24-Cla4 fusion was sufficient to normally activate Cdc42 suggesting that Bem1 acts as a scaffold and may stabilize the Cdc42 ternary complex. This could explain increased Cdc24 phosphorylation in the presence of Bem1. Thus it is possible that Bem1 does not specifically influence Cla4 phosphorylation but rather any scaffold capable to bringing together Cdc42-Cdc24-Cla4 in a complex would work.

3) The authors include Cdk1 phosphorylation sites in their D and A cdc24 mutants, but do not describe whether these residues are phosphorylated during bud emergence, which would indicate their relevance. Inclusion of these residues in the S and A cdc24 mutants is problematic, because it prevents the changes in protein activity and localization to be solely attributed to the proposed Bem1-Cdc24-Cla4-dependent mechanism. Additionally, the authors do not provide rationale stating why only 28 residues were mutated in the D mutants, while 46 were used in the A mutants.

4) The authors do not mention time-delayed negative feedback regulation of Cdc42 as reported by Das et al. 2012 and Howell et al., 2012. Could their findings provide a molecular explanation for this feedback regulation? Kuo et al. 2014 have shown that Cdc24 non-phosphorylatable mutants show stabilized Cdc42 activation during bud emergence. Do the authors expect a similar phenotype with their Cdc24-46A mutants.

5) How does Cdc42 dynamics change with cdc24-28D and 46A mutants? This would further explain the outcome of the Cdc24 phosphorylation on Cdc42 dynamics and polarization.

6) If Pak mediated Cdc24 phosphorylation does not alter GEF activity or Cdc24-Bem1 interaction what exactly does Cdc24 phosphorylation lead to? The authors claim that Cdc24 phosphorylation renders it insensitive to Bem1. However, they also show that Cdc24 non-phosphorylatable Cdc24 requires Bem1 for enhanced GEF activity. While these findings are not necessarily contradictory, they do lead to questions about how Cdc24 responds to Bem1 interaction. The authors claim that Cdc42 phosphorylation does not disrupt Cdc24-Bem1 interaction. Could it be that Cdc24 phosphorylation destabilizes that Cdc24-Cdc24-Bem1-Cla4 ternary complex instead?

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for resubmitting your work entitled "Scaffold-mediated gating of Cdc42 signaling flux" for further consideration at eLife. Your revised article has been favorably evaluated by Tony Hunter as the Senior editor, Mohan Balasubramanian as Reviewing editor, and three reviewers.

The referees have found your revised short format article significantly improved and are satisfied that you have provided a compelling biochemical reconstitution of Cdc42 GTPase signaling pathway, especially around the negative feedback regulation of the pathway and the role of Cla4 phosphorylation of Cdc24-GEF in a Bem1 dependent manner.

However, a single major and several minor points have been raised. I would like you to address the major point highlighted below, without fail. I will leave it to your judgement about the rest of the minor points, although several of these are very easily fixed with writing, doing which will improve the readability.

Essential revisions:

The authors have also added a more direct measure of Cdc42-GTP local levels by use of the Gic2 CRIB localization. They report that CRIB levels are higher in the Cdc24-A mutant and undetectable in the Cdc24-D mutants (Figure 3D). This is a good addition, but I am not entirely convinced by this rather dramatic observation. Indeed, the CRIB signal is measured apparently in non-budded cells, but the authors report that the cdc24-D mutants have a delay in polarization (reported in Figure 2—figure supplement 1E). What is the control that the cells that do not show a signal (28D and 15D) actually have a polarized patch of Cdc24? It would be important to quantify only cells that have actually polarized to test whether the local levels of active Cdc42 are indeed lower. One simple way would be to measure CRIB levels in small-budded cells. Another better way would be to co-image Cdc42-GFP and mCherry Gic2 CRIB. In Figure 3C, the authors show that Cdc42 levels are unchanged in cdc24-D mutants, and so co-imaging Cdc42 and CRIB would be a powerful way of demonstrating that the localized Cdc42 is actually less active. If this is the case, this would be I think a first such instance of uncoupling Cdc42 localization and local activation.

https://doi.org/10.7554/eLife.25257.015

Author response

[Editors’ note: the author responses to the first round of peer review follow.]

Reviewer #1:

In this manuscript, Rapali et al. study the regulation of an important complex that regulates cell polarization through Cdc42 GTPase. Cdc42 is activated by its GEF Cdc24, and also binds the scaffold Bem1. The authors show that Bem1 acts in vitro in two opposing ways: it promotes Cdc24 GEF activity towards Cdc42 and also increases the rate of phosphorylation of Cdc24 by the PAK kinase Cla4. Phosphorylated Cdc24 still activates Cdc42, but cannot be promoted by Bem1. This thus likely limits the positive effect that Bem1 has on Cdc42 activation, which the authors interpret as Bem1 contributing to signaling flux through Cdc42. They also identify all phosphorylation sites on Cdc24 and analyze the effects of phospho-mutants and phospho-mimetic mutants in vivo. In a second part, the authors use super-resolution imaging approaches to probe the diffusion and cluster-forming ability of Cdc42, probing both the function of Bem1 and of Cdc24 phosphorylation sites in these processes.

In my view, the first half of the manuscript is very interesting, clean and refreshing. The biochemistry is performed to high standards and largely convincing (a few missing controls are mentioned below). The interpretation that the contrasting roles of Bem1 increase flux is possible, but only one of several interpretations. For instance, another interpretation could be that Bem1 contributes to reaching rapidly a high, but not excessive Cdc42 activity. By contrast, the second half of the manuscript, including the analysis of Cdc24 phosphomutants and especially the super-resolution imaging, is less convincing and does not fit very well with the first half. It reads more like two stories lumped together. The consequence is that I do not feel the statement made by the title of the paper is well supported. Specific comments are below.

We thank the reviewer for their encouraging comments. The title of the manuscript has now been modified to reflect the focus on the biochemistry. We have also performed numerous additional experiments and controls to analyze the cdc24 phospho-mutants in light of the comments raised during review. These are explained in detail below.

Regarding the effect of Cdc24 phosphorylation status on Bem1-dependent activity enhancement, one missing control in Figure 1C is to test the effect of 0.2 µM Bem1, which is the concentration remaining in the reaction in Figure 3A-B. The expectation is that it should not have an effect at this concentration also in absence of Cla4-mediated phosphorylation. Another important missing control is to test whether the effect of Cla4 is solely through phosphorylation or also through protein binding. This should be addressed by using a kinase-dead allele in reaction as in Figure 3C.

As requested, we performed GEF reactions with 0.2 µM Bem1 and have included the data in new Figure 1B. We found a very modest stimulatory effect on GEF activity at this concentration. Additionally, we performed GEF assays to distinguish whether the effect of Cla4 on Bem1-stimulated GEF activity originates from phosphorylation or from protein binding. To do so, we performed GEF assays in the presence of Cla4, while omitting ATP. In new Figure 1E, we report that Cla4 was unable to dampen Bem1 stimulation of GEF activity unless ATP was present, indicating that Cla4 exerts its effect on Cdc24 GEF activity via phosphorylation. These results are described in subsection “Phosphorylation of Cdc24 inhibits the scafford-dependent increase in Cdc24 GEF activity in vitro” of the resubmitted manuscript.

The interpretation of the Cdc24 phospho-mutants is complicated and difficult to link with the biochemistry shown in the first part of the paper. They appear to conclude from their in vivo analysis "in keeping with their in vitro biochemical observations" that the 46A is active and the 28D less active (subsection “Phospho-regulation of Cdc24 impacts polarized growth in vivo” and further). If this is "in keeping with the biochemical observations", the difference in activity in the two alleles should also be Bem1-dependent in vivo, not intrinsic to Cdc24 GEF activity (i.e., there should not be a difference between the two alleles in a bem1 mutant), but this is not tested. More generally, all effects of the 46A mutant are interpreted relative to Bem1-Cla4 function. However, this mutant also includes the CDK sites, and thus some effects, such as delay in budding may be due to CDK1 action. It is also unclear in the text how the 28D sites were selected. The text could also sometimes be clearer. For instance, the double-negation in paragraph five of the section is convoluted. What the authors may mean is that the phospho-blocking and phospho-mimetic mutations have distinct phenotypes and thus are likely to affect Cdc24 each in a specific manner. Clarification on all these points would help the reader.

We agree with the reviewer that our working model predicts that there should be no difference between the two cdc24 alleles in a bem1 mutant. In new Figure 2D and E we report that this is the case. The deletion of BEM1 in cdc24-46A or cdc24-28D strains resulted in an indistinguishable temperature sensitive growth phenotype at 37ºC (Figure 2D). In addition, we found identical morphological defects in cdc24-46A bem1, cdc24-28D bem1 and bem1 mutants (Figure 2E).

In order to distinguish the influence of sites phosphorylated by Cla4 in the presence of Bem1 from all of the other sites in Cdc24 that are phosphorylated, we generated mutants in which only the Bem1-dependent sites were mutated to alanine or aspartate, which we refer to as cdc24-15A and cdc24-15D (Figure 2A). The cdc24-15D mutant displayed a temperature sensitive growth defect (Figure 2B) and the cdc24-15D mutant also displayed a delay in bud emergence in synchronized cultures (Figure 2—figure supplement 1E). It therefore appears that the Bem1-dependent phosphorylation sites, rather than the Cdk1-dependent phosphorylation sites, contribute to the observed delay in bud emergence. This point is clarified in subsection “Phospho-regulation of Cdc24 impacts polarized growth in vivo” of the new submission.

The cdc24-28D mutant was generated after homologous recombination of a cdc24-46D construct into the CDC24 locus. This is now explained in subsection “Plasmid construction”. It seems likely that the cdc24-46D mutant was not recovered after transformation because the mutant is not viable.

We apologize for the unclear text and we are grateful for the alternative that you propose. The text referred to has now been modified. In subsection “Phospho- regulation of Cdc24 impacts polarized growth in vivo” paragraph five now read, "These results indicate that the phospho-blocking and phospho-mimetic mutations confer distinct phenotypes and are thus likely to affect Cdc24 in a specific manner". We have re-read the text to try and find other instances of unclear text, but please let us know if we can help clarify any additional points.

The mobility shift of Cdc24 from synchronized cultures is also difficult to interpret. The authors claim that "wild type Cdc24 undergoes a dramatic reduction in electrophoretic mobility due to phosphorylation at around 60 minutes after synchronous release from a G1-arrest". However, the Figure 4A shows very high level of phosphorylation at time 0, slightly coming down and then up again around 60min. The description of the result thus does not correspond to what is shown on the figure. This is likely due to the α-factor treatment use for synchronization, which may cause polarization responses. For the analysis here, it would likely be much better to synchronize cells by elutriation to probe this without interfering with cell polarization. Secondly, the authors show the reduced mobility is indeed due to phosphorylation by performing phosphatase treatment of wildtype and mutant Cdc24 alleles.

We apologize to the reviewer for our imprecise description of the text on this point. We have now edited the text of the new submission to more accurately describe the data. We agree with the reviewer that elutriation would be the optimal method of showing the cell cycle-dependent change in electrophoretic mobility of Cdc24. However, having searched extensively, there are no centrifugal elutriators at the present time in Bordeaux. Moreover, the cell cycle-dependent phosphorylation of Cdc24 has previously been demonstrated in wild type cells (see PMID 17417630 Figure 3C and PMID 11106754 Figure 4A). The main point that we are drawing the reader's attention to in new Figure 2A is the loss of cell cycle-dependent Cdc24 phosphorylation in the cdc24-46A and cdc24-28D mutants, indicating that the phosphorylation sites identified by our mass spectrometry analyses are relevant in vivo. For the last point, the reviewer is correct that >90% Cdc24 migrates as a single species after phosphatase treatment. See, for example, PMID 24631237 Figures 1E and 2E.

The effect of Cdc24 phosphorylation on Cdc42 distribution and activity in vivo are not clear. The authors interpret the enrichment of Cdc24 and Cdc42 at the polar cap as sign of Cdc42 activity. This could be more directly evaluated by using probes for Cdc42-GTP, such as CRIB, which would be predicted to be in increased amounts in the 46A mutant, and decreased in the 28D mutant. Pull-down of Cdc42-GTP could also be performed to probe this point.

In order to clarify the role of Cdc24 phosphorylation on Cdc42 distribution and activity, we followed the reviewer's suggestion and performed quantitative image analyses of Cdc42-GTP using the CRIB domain encoded within Gic2. In new Figure 3D, we find that this reporter of Cdc42-GTP is enriched at the pole of cdc24-46A cells compared to wild type cells, while it was undetectable at the pole of cdc24-28D cells. These results are described in subsection “Phospho- regulation of Cdc24 impacts polarized growth in vivo.” of the new submission. Thus, Cdc24, Cdc42 and a marker for Cdc42-GTP show enrichment in cdc24-46A cells and a diminution in cdc24-28D cells in vivo. We also repeatedly attempted to perform pull down assays of Cdc42-GTP, but the experiment did not work for reasons that we do not currently understand.

Reviewer #2:

[…] 1) The demonstrations of Bem1 effects on Cdc24 activity and phosphorylation are convincing but they mostly confirm or incrementally improve on previous studies. The idea of a self-limitation and negative effect of phosphorylation on GEF-stimulation is however a really new and therefore very important part of this study. As I interpret the current results the negative effect on Bem1-stimulation could either be due to the phosphorylation itself (as the authors suggest) or due to the presence of Cla4 (which is always present in the relevant assays from Figure 3). Can the GEF assays in 3B/C be performed with a kinase dead version of Cla4 as control? Alternatively, the kinase reaction could be performed without phosphate in the medium. This would support the idea of a central effect of the Cdc24 phosphorylation sites. In the same direction: Bem1 seems to stimulate additional phosphorylation of very specific sites (not just more but different pattern). It would good to include a mutant where only the Bem1-specifi sites are mutated and include this in the assays for Figure 3. Finally, it should be possible to combine Dcla4 and Drga1 to show that this leads to a similar phenotype as the 46A mutant + Dcla4 (this might have been already shown, in this case refer to the relevant paper).

We performed additional experiments that are now presented in Figure 1E of the new submission, in which we performed GEF assays with Cla4 and Bem1, but lacking ATP. The results indicate that Cla4 limits GEF activity via phosphorylation rather than via a protein-protein interaction, since Cdc24 GEF activity was stimulated by reactions containing Bem1 and Cla4 when ATP was omitted. However, when ATP was added to the GEF reactions, Cla4 attenuated Bem1 stimulation. The experiment is described in subsection “Phosphorylation of Cdc24 inhibits the scaffold-dependent increase in 135 Cdc24 GEF activity in vitro.”.

We agree with the reviewer that in addition to stimulating the rate of Cdc24 phosphorylation, Bem1 also increases the extent of Cdc24 phosphorylation by promoting the phosphorylation of sites that Cla4 alone does not seem to phosphorylate, even after extensive incubation with Cdc24 in kinase reactions. These results were also borne out by our mass spectrometry analysis. We have therefore clarified this point in the sub-heading “Bem1 increases the rate and extent of Cdc24 phosphorylation by the 116 PAK Cla4” and in the text.

We followed the reviewer's suggestion of generating mutants in which we mutated the Bem1-specific phosphorylation sites in Cdc24 to alanine or aspartate. We refer to these mutants as cdc24-15A and cdc24-15D. We found that, as with the cdc24-28D mutant, the cdc24-15A mutant was temperature sensitive. This data is shown in Figure 2B of the revised submission. We also looked at the levels of Cdc42-GTP in vivoin these mutants. The cdc24-15A mutant showed a higher level of the CRIB construct compared to wild type CDC24 at the pole, but less than the cdc24-46A mutant. In the cdc24-15D mutant, levels of the CRIB construct were below the levels required for reliable quantification. The results are presented in new Figure 3D and explained in subsection “Phospho- regulation of Cdc24 impacts polarized growth in vivo”.

The phenotype of the cla4 rga1 double mutant has previously been reported as being elongated in a similar manner to the cdc24-46A rga1 double mutant. Thus, the data that we are reporting are complementary with previous studies. We have now cited this work, Caviston et al., 2003.

[…] 4) The current data support a reduction in stimulation of Cdc24 by Bem1 after Cla4-mediated phosphorylation. To really support the sequence of events proposed in the model (Figure 8), Cla4 would need to act after Bem1 recruitment (fits the idea of Bem1 as recruitment platform for Cla4). Can the authors perform such an experiment with their reconstituted system? 1. Measure basal GEF activity, 2. Add Bem1 for 30 min, 3. Measure stimulated GEF activity, 4. Add Cla4 (high levels) for 1 h, 5. Measure GEF activity again to see basal levels again. This would provide a perfect final experiment and nicely relate to the proposed model.

We agree that this is a great test of the model, but it has been extremely technically challenging. We perform the assays in 10 µl volumes, so the extended times required to extract the Kobs from the reaction curves meant that the reactions become prone to evaporation. The problems are not remedied by increasing the reaction volumes, since the next problem encountered is judging the effect of phosphorylation. The effect of adding Bem1 to the GEF activity is clear, but the effect of adding kinase is to subtly dampen the rate of GEF activity. There are additional problems here: most of the substrate, Cdc42, has been consumed early in the reaction, resulting in increased signal-to-noise as the reaction proceeds. Moreover, the reaction rate can no longer be fit to a single exponential, but must instead be judged manually. In conclusion, we cannot currently do this experiment, and in fact, it was for this reason that we initially broke this complex reaction into more approachable constituents.

Reviewer #3:

[…] 1) The authors state that phosphorylation of Cdc24 reduces its activity, but do not directly test whether this is due to decreased affinity between Cdc24 and Cdc42. While the authors test the GEF activity of Cdc24 in various mutants, they fail to use the most direct (and biologically relevant) approach of measuring Cdc42 activity in vivo. Later they present phenotypes of these mutants, but only indirectly link them to Cdc42 by observing cell morphology and cytoskeleton organization (Figure 4C).

We have now analyzed a marker for active Cdc42 in vivo, as requested by the reviewer. The results are presented in new Figure 3D and described in subsection “Phospho- regulation of Cdc24 impacts polarized growth in vivo”. We find that the marker of active Cdc42 is enriched at the pole of the cdc24- 46A mutant compared to wild type CDC24, while the levels of active Cdc42 in the cdc24-28D mutant were below the detection limit of our imaging system. Thus, the in vitroGEF assays, actin cytoskeleton staining, Cdc24 localization, Cdc42 localization and active Cdc42 localization are consistent.

2) Previous reports (Kozubowski et al. 2008) have shown that the Pak kinase forms a complex with Cdc42-Cdc24-Bem1 to positively regulate Cdc42 activation and constitutes a positive feedback loop that leads to symmetry breaking. The authors have not discussed their findings in the context of this report. Moreover, Cdc24-Cla4 fusion was sufficient to normally activate Cdc42 suggesting that Bem1 acts as a scaffold and may stabilize the Cdc42 ternary complex. This could explain increased Cdc24 phosphorylation in the presence of Bem1. Thus it is possible that Bem1 does not specifically influence Cla4 phosphorylation but rather any scaffold capable to bringing together Cdc42-Cdc24-Cla4 in a complex would work.

The reviewer raises the complexity of the budding yeast polarity system, which has generated considerable debate in the field (see the comment published by Li & Wedlich-Soldner in the same issue of Current Biology as the Kozubowski paper, PMID 19278629). As the reviewer rightly points out, in the Kozubowski manuscript, the role of Bem1 in symmetry breaking was attributed to its tethering of Cdc24 and Cla4, since a Cdc24-Cla4 fusion bypassed the requirement for Bem1 and Bud1 in polarity establishment. This argued that Cla4 plays a positive role in polarity establishment. However, the role of Cla4 has since been shown to be more complex, as the Lew lab have published two manuscripts in which Cla4 was found to contribute to negative feedback during polarity establishment (PMID 24631237 and 26523396). The two findings are not incompatible; it is possible that Cla4 phosphorylates a subset of Cdc24 sites that boost Bem1-dependent GEF activity, but that the ensuing phosphorylation of all sites attenuates this stimulation. A similar mechanism has been reported for the Cdk1 inhibitor Wee1 in budding yeast (PMID 16096060). The initial phosphorylation of Wee1 by Cdk1 activates Wee1, while subsequent full phosphorylation of Wee1 by Cdk1 is inhibitory. It is not possible for us to perform the exhaustive mutational analyses required to test this hypothesis in the short article format of our resubmission, but we have cited the Kozubowski paper, as requested. In the Discussion section, we have now added that, "While a previous study found that the PAK Cla4 played a positive role in polarity establishment in S. cerevisiae (Kozubowski et al., 2008), subsequent work found that phosphorylation of Cdc24 by Cla4 resulted in reduced GEF activity, while increased Cdc42 recruitment to the cell pole was observed in Cdc24 phosphorylation mutants (Kuo et al., 2014; Wu et al., 2015)."

The reviewer asks if any scaffold that tethers Cdc42-Cdc24-Cla4 might suffice to stimulate Cdc24 phosphorylation. We have not tested whether other scaffolds might induce the same pattern of phosphorylation in Cdc24 as Bem1 and we have stated this in the Discussion section of the resubmitted manuscript: "It will also be important to test whether any scaffold that increases the local Cdc24-Cla4 concentration would stimulate Cdc24 phosphorylation, or whether Bem1 exerts specific effects." While Bem1 presumably does increase the local Cdc24 concentration available for phosphorylation by Cla4, our data suggest that it does so in a specific fashion. Even when Cla4 and Cdc24 are incubated together for extensive periods (6 hours), the extent of Cdc24 phosphorylation is markedly reduced compared with the inclusion of Bem1 in the reaction. Since no other proteins are present in these reactions, it seems likely that such extensive incubations in the absence of Bem1 would eventually result in full phosphorylation of Cdc24, but it does not. This point was also borne out by our mass spectrometry analysis in which we identified Bem1-specific phosphorylation sites in Ccd24.

3) The authors include Cdk1 phosphorylation sites in their D and A cdc24 mutants, but do not describe whether these residues are phosphorylated during bud emergence, which would indicate their relevance. Inclusion of these residues in the S and A cdc24 mutants is problematic, because it prevents the changes in protein activity and localization to be solely attributed to the proposed Bem1-Cdc24-Cla4-dependent mechanism. Additionally, the authors do not provide rationale stating why only 28 residues were mutated in the D mutants, while 46 were used in the A mutants.

In our resubmitted manuscript we generated mutants in which only Bem1- dependent sites in Cdc24 were mutated to alanine or aspartate. We refer to these mutants as cdc24-15A and cdc24-15D. These mutants resulted in enrichment of Cdc42-GTP at the pole in the case of the cdc24-15A mutant (Figure 3D) and a temperature sensitive phenotype in the case of the cdc24-15D mutant (Figure 2B). These mutants do not contain the Cdk1-dependent sites that we previously identified.

The cdc24-28D mutant was generated by homologous recombination after transformation of a cdc24-46D construct into yeast. Our interpretation is that mutation of all 46 phosphorylated residues to aspartate is inviable, whereas the cdc24-28D mutant supports viability, albeit with temperature sensitivity. We have explained this more clearly in subsection “Plasmid construction” of the Materials and methods section in the resubmission.

4) The authors do not mention time-delayed negative feedback regulation of Cdc42 as reported by Das et al. 2012 and Howell et al., 2012. Could their findings provide a molecular explanation for this feedback regulation? Kuo et al. 2014 have shown that Cdc24 non-phosphorylatable mutants show stabilized Cdc42 activation during bud emergence. Do the authors expect a similar phenotype with their Cdc24-46A mutants.

This is a very good point and is something that we are currently investigating using standard widefield imaging and SPT-PALM. The reviewer is correct that the current models from the Verde and Lew labs would predict that Cdc42 might be stabilized in the cdc24-46A mutant.

5) How does Cdc42 dynamics change with cdc24-28D and 46A mutants? This would further explain the outcome of the Cdc24 phosphorylation on Cdc42 dynamics and polarization.

We have only looked at Cdc42 dynamics in the mutants by single particle tracking. Cdc42 diffusion was reduced in the cdc24-46A mutant, both at the pole and non-pole, but this data has been removed from the resubmitted manuscript at the request of the editor.

6) If Pak mediated Cdc24 phosphorylation does not alter GEF activity or Cdc24-Bem1 interaction what exactly does Cdc24 phosphorylation lead to? The authors claim that Cdc24 phosphorylation renders it insensitive to Bem1. However, they also show that Cdc24 non-phosphorylatable Cdc24 requires Bem1 for enhanced GEF activity. While these findings are not necessarily contradictory, they do lead to questions about how Cdc24 responds to Bem1 interaction. The authors claim that Cdc42 phosphorylation does not disrupt Cdc24-Bem1 interaction. Could it be that Cdc24 phosphorylation destabilizes that Cdc24-Cdc24-Bem1-Cla4 ternary complex instead?

We apologize for giving the wrong impression on this point. The pulldown experiment that we present only allows us to conclude that phosphorylation of Cdc24 does not markedly influence its affinity for Bem1. Since the sensitivity of a pulldown assay is limited, further biophysical investigation will be required to fully address this point. It is possible that phosphorylation may increase the off-rate of Cdc24 for Bem1 or decrease the on-rate, but our results do not currently enable definitive conclusions on this point. Similarly, we are looking at how phosphorylation may affect the ternary Cdc42-Cdc24-Bem1-Cla4 complex. However, the labile nature of the complex, involving low affinity interactions between most components, makes this a very challenging project.

[Editors' note: the author responses to the re-review follow.]

Essential revisions:

The authors have also added a more direct measure of Cdc42-GTP local levels by use of the Gic2 CRIB localization. They report that CRIB levels are higher in the Cdc24-A mutant and undetectable in the Cdc24-D mutants (Figure 3D). This is a good addition, but I am not entirely convinced by this rather dramatic observation. Indeed, the CRIB signal is measured apparently in non-budded cells, but the authors report that the cdc24-D mutants have a delay in polarization (reported in Figure 2—figure supplement 1E). What is the control that the cells that do not show a signal (28D and 15D) actually have a polarized patch of Cdc24? It would be important to quantify only cells that have actually polarized to test whether the local levels of active Cdc42 are indeed lower. One simple way would be to measure CRIB levels in small-budded cells. Another better way would be to co-image Cdc42-GFP and mCherry Gic2 CRIB. In Figure 3C, the authors show that Cdc42 levels are unchanged in cdc24-D mutants, and so co-imaging Cdc42 and CRIB would be a powerful way of demonstrating that the localized Cdc42 is actually less active. If this is the case, this would be I think a first such instance of uncoupling Cdc42 localization and local activation.

We redesigned the marker used to assay levels of Cdc42-GTP in vivoand we repeated the experiments shown in Figure 3D. By extending the reporter construct that recognises Cdc42-GTP from residues 1-192 to 1-208 of Gic2, changing the promoter and swapping the tag from the N- to the C-terminus of Gic2, we found that the signal from the reporter was brighter, enabling us to monitor Cdc42-GTP levels in all of the cdc24-Ala and cdc24-Asp phosphorylation site mutants. Importantly, the optimization of this Cdc42-GTP reporter enabled us to ensure that quantitative imaging was only performed on polarizing cells, as requested by the reviewers. The conclusion of these experiments is unchanged from our previous submission, but we are now able to present corroborating quantitative analyses for the cdc24-28D and cdc24- 15D mutants.

https://doi.org/10.7554/eLife.25257.016

Article and author information

Author details

  1. Péter Rapali

    University of Bordeaux, CNRS, European Institute of Chemistry and Biology, IBGC, UMR 5095, Pessac, France
    Contribution
    PR, Formal analysis, Investigation, Methodology, Writing—original draft
    Competing interests
    The authors declare that no competing interests exist.
  2. Romain Mitteau

    University of Bordeaux, CNRS, European Institute of Chemistry and Biology, IBGC, UMR 5095, Pessac, France
    Contribution
    RM, Formal analysis, Investigation
    Competing interests
    The authors declare that no competing interests exist.
  3. Craig Braun

    Department of Cell Biology, Harvard Medical School, Boston, United States
    Contribution
    CB, Investigation
    Competing interests
    The authors declare that no competing interests exist.
  4. Aurèlie Massoni-Laporte

    University of Bordeaux, CNRS, European Institute of Chemistry and Biology, IBGC, UMR 5095, Pessac, France
    Contribution
    AM-L, Investigation
    Competing interests
    The authors declare that no competing interests exist.
  5. Caner Ünlü

    University of Bordeaux, CNRS, European Institute of Chemistry and Biology, IBGC, UMR 5095, Pessac, France
    Contribution
    CÜ, Investigation
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0612-3111
  6. Laure Bataille

    University of Bordeaux, CNRS, European Institute of Chemistry and Biology, IBGC, UMR 5095, Pessac, France
    Contribution
    LB, Investigation
    Competing interests
    The authors declare that no competing interests exist.
  7. Floriane Saint Arramon

    University of Bordeaux, CNRS, European Institute of Chemistry and Biology, IBGC, UMR 5095, Pessac, France
    Contribution
    FSA, Methodology
    Competing interests
    The authors declare that no competing interests exist.
  8. Steven P Gygi

    Department of Cell Biology, Harvard Medical School, Boston, United States
    Contribution
    SPG, Formal analysis, Supervision, Methodology
    Competing interests
    The authors declare that no competing interests exist.
  9. Derek McCusker

    University of Bordeaux, CNRS, European Institute of Chemistry and Biology, IBGC, UMR 5095, Pessac, France
    Contribution
    DM, Conceptualization, Formal analysis, Supervision, Funding acquisition, Investigation, Writing—original draft, Writing—review and editing
    For correspondence
    mccusker@iecb.u-bordeaux.fr
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-1455-1711

Funding

Centre National de la Recherche Scientifique

  • Derek McCusker

Agence Nationale de la Recherche (ANR-13-BSV2-0015-01)

  • Derek McCusker

Regional Council of Aquitaine (2012 13 01 012)

  • Derek McCusker

Regional Council of Aquitaine (2015-1R30113)

  • Derek McCusker

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Jean-Louis Mergny for the use of the fluorescence plate reader and Lionel Minvieille-Sébastia for the anti-Fip1 antibody. We also thank Anne Royou, Doug Kellogg and Robert Arkowitz for comments on the manuscript. This work was supported by funding from ANR grant ANR-13-BSV2-0015-01, the Regional Council of Aquitaine, ARC, CNRS and the University of Bordeaux.

Reviewing Editor

  1. Mohan K Balasubramanian, University of Warwick, United Kingdom

Publication history

  1. Received: January 18, 2017
  2. Accepted: March 15, 2017
  3. Accepted Manuscript published: March 17, 2017 (version 1)
  4. Version of Record published: April 10, 2017 (version 2)

Copyright

© 2017, Rapali et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,299
    Page views
  • 395
    Downloads
  • 18
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, Scopus, PubMed Central.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Cell Biology
    2. Developmental Biology
    Eric R Brooks et al.
    Research Article

    Neural tube closure defects are a major cause of infant mortality, with exencephaly accounting for nearly one-third of cases. However, the mechanisms of cranial neural tube closure are not well understood. Here we show that this process involves a tissue-wide pattern of apical constriction controlled by Sonic hedgehog (Shh) signaling. Midline cells in the mouse midbrain neuroepithelium are short with large apical surfaces, whereas lateral cells are taller and undergo synchronous apical constriction, driving neural fold elevation. Embryos lacking the Shh effector Gli2 fail to produce appropriate midline cell architecture, whereas embryos with expanded Shh signaling, including the IFT-A complex mutants Ift122 and Ttc21b and embryos expressing activated Smoothened, display apical constriction defects in lateral cells. Disruption of lateral, but not midline, cell remodeling results in exencephaly. These results reveal a morphogenetic program of patterned apical constriction governed by Shh signaling that generates structural changes in the developing mammalian brain.

    1. Cell Biology
    Tomás Edreira et al.
    Research Article

    In fission yeast, the septation initiation network (SIN) ensures temporal coordination between actomyosin ring (CAR) constriction with membrane ingression and septum synthesis. However, questions remain about CAR regulation under stress conditions. We show that Rgf1p (Rho1p GEF), participates in a delay of cytokinesis under cell wall stress (blankophor, BP). BP did not interfere with CAR assembly or the rate of CAR constriction, but did delay the onset of constriction in the wild type cells but not in the rgf1Δ cells. This delay was also abolished in the absence of Pmk1p, the MAPK of the cell integrity pathway (CIP), leading to premature abscission and a multi-septated phenotype. Moreover, cytokinesis delay correlates with maintained SIN signaling and depends on the SIN to be achieved. Thus, we propose that the CIP participates in a checkpoint, capable of triggering a CAR constriction delay through the SIN pathway to ensure that cytokinesis terminates successfully.