Iqg1p is a component of the actomyosin contractile ring that is required for actin recruitment and septum deposition. Cells lacking Iqg1p function have an altered bud-neck structure and fail to form a functional actomyosin contractile ring resulting in a block to cytokinesis and septation. Here it is demonstrated that increased expression of the actin cytoskeleton associated protein Bsp1p bypasses the requirement for contractile ring function. This also correlates with reduced bud-neck width and remedial septum formation. Increased expression of this protein in a temperature-sensitive iqg1-1 background causes remedial septum formation at the bud neck that is reliant upon chitin synthase III activity and restores cell separation. The observed suppression correlates with a restoration of normal bud-neck structure. While Bsp1p is a component of the contractile ring, its recruitment to the bud neck is not required for the observed suppression. Loss of Bsp1p causes a brief delay in the redistribution of the actin cytoskeleton normally observed at the end of actin ring contraction. Compromise of Iqg1p function, in the absence of Bsp1p function, leads to a profound change in the distribution of actin and the pattern of cell growth accompanied by a failure to complete cytokinesis and cell separation.
CELL polarity and morphology are intimately linked in eukaryotic cells and both are dependent upon the actin cytoskeleton. Growth polarity requires that a polarized actin cytoskeleton directs secretion and delivery of plasma membrane constituents to sites of de novo membrane synthesis. In budding yeast, this means that polarized growth is directed to the bud tip in late G1/S phase cells, to the entire periphery of the bud in G2 cells, to the surface of both mother and daughter cells in mitosis, and subsequently to the bud neck late in the cell cycle to permit cytokinesis, septation, and cell separation (Bretscher 2003). This pattern of new membrane and cell wall deposition is precisely mirrored by the polarization of the actin cytoskeleton (Amberg 1998). The promotion and maintenance of polarized growth requires both exocytosis and endocytosis coupled to slow plasma membrane diffusion rates (Valdez-Taubas and Pelham 2004). The most distinct filamentous actin structures in yeast cells are actin patches. These form at sites of endocytosis undergoing distinct phases of protein recruitment and movement that directs clathrin-dependent endocytic vesicle formation and movement (Kakosanen et al. 2003, 2005). Actin also polymerizes into cable structures, which template myosin-based vesicular movement outward toward the cell periphery, again in a polarized manner (Adams and Pringle 1984; Pruyne et al. 2004).
A large number of proteins have been found to associate with actin patches although definition of individual functions has proved difficult, partly as a result of apparent genetic redundancy but also because of a considerable plasticity or buffering capacity allowing for tolerance of reduced cytoskeletal function. One component of actin patches to which no function beyond an ability to bind synaptojanins has been ascribed is Bsp1p. The bsp1 null mutant retains viability and has no obvious defect in polarized growth or cellular morphology. Bsp1p is unusual among actin patch proteins in that it also localizes to the bud neck late in the cell cycle (Drees et al. 2001; Wicky et al. 2003). Interestingly, Bsp1p localization to patch structures is actin dependent whereas its appearance at the bud neck is independent of actin. Actin recruitment to the bud neck is dependent on the yeast IQGAP protein, Iqg1p, both proteins being components of the contractile ring in conjunction with Myo1p and many other proteins (Huh et al. 2003). Contractile ring function is required for cytokinesis when contraction is believed to guide new membrane deposition and septum formation although there is a clear interdependency between contractile ring activity and primary, Chs2p-dependent, septum deposition (Schmidt et al. 2002; VerPlank and Li 2005). Cell separation is ultimately achieved by synthesis of secondary septa flanking the primary septum, creating a trilaminate chitin structure, and the subsequent dissolution of the primary septum through the activity of the chitinase Cts1p (Kuranda and Robbins 1991). A second chitin synthase, Chs3p, is thought to be involved in cell wall remodeling and is required for remedial septum deposition in the absence of normal contractile ring function or septum formation (Shaw et al. 1991; Schmidt et al. 2002). Chs3p translocates to the bud neck late in the cell cycle via a Chs5p/Chs6p-dependent vesicle transport pathway (Santos and Snyder 1997; Ziman et al. 1998; Trautwein et al. 2006). Failure to properly synthesize cell wall structures potentially damages the cell wall. Damage to the cell wall results in activation of a protein kinase C dependent mitogen activated protein kinase signaling pathway (PKC-MPK), linked to upstream cell wall component sensors, that generates a specific pattern of gene expression and concomitant remedial cell wall synthesis (Levin 2005).
Iqg1p is thought to act as a platform for coordination of actin ring assembly and contraction and has various conserved domains: IQ domains defined as myosin light-chain-binding sites in the neck region of myosin heavy chains (Cheney and Mooseker 1992; Xie et al. 1994); a GAP domain thought to be required for activation of a GTPase (Lippincott and Li 1998; Shannon and Li 1999); and a calponin homology domain that is required for interaction with actin (Epp and Chant 1997; Shannon and Li 1999). More recent evidence suggests a role for Iqg1p in establishing axial bud patterning and secretion through direct physical interactions with Bud4p and Sec3p (Osman et al. 2002). A temperature-sensitive mutation that maps to a single IQ domain of Iqg1p, iqg1-1, fails to recruit actin to the contractile ring, preventing cytokinesis, septum formation, and cell separation (Boyne et al. 2000). Here we report that increased BSP1 dosage is able to restore viability in the iqg1-1 mutant at the restrictive temperature. Suppression, however, does not result from reconstitution of contractile ring formation and is independent of the cell wall integrity signaling pathway. A synthetic genetic interaction between iqg1-1 and a bsp1 null mutation causes disorganization of the actin cytoskeleton, altered cellular morphology, and loss of both actomyosin ring and septum formation. Increased levels of Bsp1p expression restore normal bud-neck structure to iqg1-1 cells, resulting in Chs3p-dependent remedial septum formation.
MATERIALS AND METHODS
Yeast strains, methods, and media:
The yeast strains used in this study were isogenic with W303a and are listed in Table 1. Growth was in YEPD (1% yeast extract, 2% bacto-peptone, 2% glucose) at the stated temperature. Strains for fluorophore visualization were grown in complete synthetic medium (SC; 2% glucose, 0.7% yeast nitrogen base without amino acids, 40 mg/liter adenine, 30 mg/liter leucine, 20 mg/liter histidine, tryptophan, and uracil). Strains carrying plasmids were propagated in selective synthetic medium (Sherman 1991). Solid media was made to the same composition, but with the addition of 2% agar. Temperature shifts during growth in liquid media were carried out via the addition of an equal volume of appropriately prewarmed media to exponentially growing cultures. Yeast transformations were carried out by a standard lithium acetate method (Gietz and Woods 2002). GAL-CDC20, cdc20Δ cells were grown to exponential phase in SC Gal/Raf (as above, but substituting 2% galactose and 4% raffinose for glucose), before being transferred to SC containing 2% glucose to induce metaphase arrest. Induction of constructs carrying the GAL1 promoter was performed by growing cells to exponential phase in SC media with appropriate auxotrophic selection. Cells were washed in 10 culture volumes of selective SC Gal/Raf before being resuspended in 1 vol of media. Induction of MET3 promoters was performed by growing cells in selective synthetic media containing 20 mg/liter methionine before washing with 10 vol of media without methionine and resuspension in one culture volume of the same media.
Plasmids and oligonucleotide primers used in this study are listed in Tables 2 and 3, respectively. Primers CLI264 and CLI266 were used to introduce a 13-Myc tag linked to the kanMX6 G418 resistance gene at the C terminus of BSP1, using the plasmid pFA6a-13Myc-kanMX6 as a template, by standard PCR amplification and subsequent homologous recombination into the yeast genome (Longtine et al. 1998). The module was integrated into diploid W303 (SSC3) and linkage of the insert to the BSP1 locus was tested by PCR using the primer CLI266, which lies 5′ of the insertion site, and CLI 70, which lies within the G418 resistance marker. Further tagging and deletion of genes was similarly achieved using the plasmids pFA6a-GFP-kanMX6, pFA6a-HIS3-kanMX6, pFA6a-CFP-kanMX6, pFA6a-YFP-kanMX6, and pFA6a-hphMX6 as templates (Longtine et al. 1998; Sheff and Thorn 2004) with appropriate primers as listed in Table 3. All primer sequences are available on request.
Actin was visualized using tetramethyl rhodamine iso-thiocyanate (TRITC)-conjugated phalloidin (Sigma, St. Louis) performed as described by Adams and Pringle (1991). Chitin was visualized using Fluorescent Brightener 28 (Sigma) (Pringle 1991). Cells were fixed and processed for immunofluorescence according to Ayscough and Drubin (1998). BSP1-13Myc strains were stained using anti-c-myc primary antibody (Covance) and FITC-conjugated goat anti-mouse IgG secondary antibody (Sigma). BSP1-2HA strains were stained using monoclonal anti-HA primary antibody (Covance) and FITC-conjugated goat anti-mouse IgG secondary antibody (Sigma). Strains for GFP visualization were grown in synthetic complete medium supplemented with 40 mg/liter adenine. Cells were fixed for a maximum of 15 min in 0.1 m potassium phosphate buffer, pH 7.0, containing 3.8% formaldehyde. Live cell imaging was carried out by mounting the cells in SC medium with 1% agarose, and the growth temperature was maintained using a Bioptechs Objective temperature control system (Bioptechs). Fluorescence images were collected using either Leica DMRB or Nikon Eclipse E600 microscopes, fitted with a RTEA/CCD-1317-K/2 CCD camera (Princeton Instruments), or Hamamatsu IEEE1394 C4742-95-12ERG digital CCD camera (Digital Pixel). Images were collected and analyzed using Openlab v2.06 software (Improvision) or Simple PCI v5.1.0.0110. z-sectioning was carried out on a Nikon Eclipse E600 microscope fitted with a PI KG E-662 LVPZT Position Servo-Controller and P-721.17 Piezoelectric Translator (Digital Pixel) or Delta Vision RT (Applied Precision).
Fluorescence resonance energy transfer analysis:
Exponentially growing cultures of BSP1-YFP (SSC836), IQG1-CFP (SSC818), and BSP1-YFP, IQG1-CFP (SSC867) in SC were prepared for live cell imaging. Micrographs for each strain were captured with the following excitation (EX) and emission (EM) wavelengths: YFP 500 (EX)/535 (EM), CFP 435 (EX)/470 (EM), fluorescence resonance energy transfer (FRET) 435 (EX)/535 (EM). Deconvolved images were analyzed using the SoftWorRx program from Applied Precision. Background fluorescent intensity within the cell was removed by comparing pixel intensity with an equivalent volume outside the cell.
Cells were fixed according to the protocol of Wright (2000). Exponentially growing culture (9.5 ml) was added to 9.5 ml of 2× fixative (0.2 m PIPES, pH 6.8, 0.2 m sorbitol, 4 mm MgCl2, 4 mm CaCl2, 6% glutaraldehyde), and incubated at room temperature for 5 min. Subsequently, cells were spun at 1500 × g for 5 min 4° prior to resuspension in 10 ml of 1× fixative at 4° overnight to complete fixation. Pellets were washed three times for 10 min each in water with spinning and resuspension steps in between before being resuspended in 2% aqueous potassium permanganate solution for 5 min. Cells were pelleted using centrifugation and the supernatant was removed and replaced with fresh potassium permanganate for a further 45 min. The potassium permanganate was removed from the pellet by gentle washing with water. The pellet was overlaid with 1% uranyl acetate and incubated for 1 hr at room temperature before being dehydrated in a graded ethanol series (30–100% ethanol). Pellets were embedded in Spurr's resin (medium, long-life formula) before sectioning and poststaining with Reynold's lead citrate. Images were collected on a JEOL 1220 TEM (JEOL, Tokyo) operating at 80 kV and fitted with a SiS Morada CCD camera and AnalySIS v3.0 software (Olympus Soft Imaging Solutions).
Analysis of Mpk1p activation was carried out by growing strains W303 (SSC1), iqg1-1 (SSC166), mid2Δ (SSC792), iqg1-1 mid2Δ (SSC790), wsc1Δ (SSC793), and iqg1-1 wsc1Δ (SSC791) carrying YCplac22 MET3 BSP1 to exponential phase in 75 ml of SC −Leu −Met. Duplicate cultures of each strain were grown to exponential phase at 26°, and these cultures were raised to 37° at time (t) 0 by adding an equal volume of 48° media −Met or +Met (30 mg/liter final concentration). Fifty milliliter samples were taken at t0, t120 min, and t240 min for the preparation of total soluble protein extracts, as described below.
Quantification of Bsp1p protein levels was performed by growing W303 (SSC001), BSP1-HA (SSC527), and BSP1-HA transformed with pKO11 GAL-BSP1-2HA cells to exponential phase in SC −URA medium at 30°. Cells were washed in 10 vol of SC Gal/Raf −URA medium before being resuspended in 1 vol of SC Gal/Raf −URA. Half of the BSP1-HA pKO11 GAL-BSP1-2HA culture was maintained in SC −URA medium. The cultures were incubated at 30° for a further 3 hr before being sampled for total soluble protein preparations. Protein extracts were analyzed by Western blotting. Quantification of relative protein levels was performed using a Typhoon 9410 Variable Mode Imager with ImageQuant 5.2 software (Amersham Biosciences).
Samples were harvested by centrifugation and washed once in 1 ml of lysis buffer (50 mm Tris–base, 50 mm NaF, 5 mm EDTA, 1 mm DTT, 80 mm β-glycerophosphate, 15 mm nitrophenylphosphate, 25 mm NaCl, 0.1 mm sodium orthovanadate, 1% v/v NP40, 40 mg/ml pepstatin A, 40 mg/ml aprotinin, 20 mg/ml leupeptin, 200 mg/ml PMSF). The cells were pelleted and the supernatant discarded, 2 vol of acid-washed glass beads were added followed by sufficient lysis buffer to cover the beads. Tubes were then placed in the Ribolyser (Hybaid) and agitated (3 × 10 sec) and then placed immediately on ice. The supernatant was collected after a 5-min 2000 rpm centrifugation and transferred to a new tube for a further 15-min 14,000 rpm centrifugation at 4°. Protein concentrations were determined by Bradford assay (Bradford 1976). In Figure 6B, 200 mg of total protein was loaded per well following a 5-min incubation at 96° in SDS–PAGE loading buffer (50 mm Tris–HCL, 100 mm DTT, 2% SDS, 0.1% bromophenol blue, 10% glycerol). Samples were loaded onto 8% SDS–PAGE gels [Bio-Rad (Hercules, CA) mini-protean II] and proteins were separated at a constant 120 V until the dye front reached the bottom of the gel. Prestained protein markers (New England Biolabs, Beverly, MA) were also loaded. Proteins were then transferred to Hybond-C nitrocellulose membrane (Amersham, Piscataway, NJ) using a wet-blotting system (Bio-Rad Trans-Blot cell) and Western transfer buffer (25 mm Tris–HCL, 192 mm glycine, 20% methanol). Immunoblot analysis was performed using the enhanced chemiluminescence kit (Amersham) and Hyperfilm MP (Amersham) according to the manufacturer's instructions. Anti-Mpk1p (rabbit polyclonal, Santa Cruz Biotechnology) and anti-phospho-p38 Mpk1p (rabbit polyclonal, Cell Signaling Technology) antibodies were used at a dilution of 1:500 and 1:1000, respectively, and detected using HRP-conjugated goat anti-rabbit IgG (DAKO) diluted to 1:5000. Monoclonal anti-HA (Cancer Research UK) antibodies were used at a working dilution of 1:5000 and detected using HRP-conjugated goat anti-mouse IgG (DAKO) at 1:5000.
Determination of total chitin levels:
Strains were grown in SC −LEU medium ±MET for 3 hr before being harvested. Cell walls were prepared by disrupting cells with glass beads (Sigma) using a RiboLyser (Hybaid) before extraction according to the procedure described by Munro et al. (2003). Chitin content was determined by measurement of glucosamine levels after acid hydrolysis of cell wall extracts (Kapteyn et al. 2000). Average chitin levels for a strain were determined from triplicate samples.
We undertook a screen for gene dosage suppressors of the iqg1-1 temperature-sensitive allele and identified a high-copy-number plasmid encoding BSP1. The BSP1 open reading frame was subsequently placed under control of the methionine-regulated MET3 promoter on a high-copy-number yeast plasmid. The data in Figure 1A demonstrate that, in the absence of methionine, conditions under which BSP1 is expressed, growth of the iqg1-1 mutant at 37° is restored, whereas in the presence of methionine, and no increased BSP1 expression, cells retain their temperature-sensitive phenotype. Mutant iqg1-1 cells fail to recruit actin to the actomyosin ring at the bud neck and as a consequence fail to complete cytokinesis and cell separation, leading to a chained cell phenotype (Boyne et al. 2000). Under conditions in which the iqg1-1 temperature-sensitive growth is suppressed by increased BSP1 gene dosage, the appearance of chained cells in the population is severely reduced (Figure 1B). Despite restoring the ability of cells to undergo cell separation, suppression is not dependent upon recruitment of actin to the contractile ring (Figure 1C). The data presented in Figure 1, B and C, are from a single representative experiment, the growth regimes for all strains being carried out simultaneously, and this and subsequent growth experiments have been repeated on at least three independent occasions. Subsequent microscopic analysis of the bud-neck structure revealed the presence of thickened cell walls and the deposition of dense cell wall material within the bud neck of iqg1-1 cells expressing increased levels of Bsp1p (Figure 1D). Increased BSP1 dosage was unable to suppress an iqg1 null mutation (data not shown), implying that the altered iqg1-1-encoded protein retains some function required for BSP1-mediated suppression.
Bsp1p localization to the contractile ring:
Previous work had indicated that Bsp1p localized to both actin patches and the actomyosin contractile ring (Drees et al. 2001; Wicky et al. 2003). We investigated the timing and dependency of Bsp1p contractile ring localization in more detail using epitope-tagged Bsp1 as shown in Figure 2. Bsp1-GFP clearly localizes to the bud neck (Figure 2A) and undergoes contraction (Figure 2B). In synchronous cultures, Bsp1-myc localizes to the ring structure in early anaphase, prior to the recruitment of actin (Figure 2, C and D). Figure 2D simply demonstrates that the localization pattern of the Bsp1-13myc fusion protein is similar to that of Bsp1-GFP. We next examined the dependency of Bsp1p localization to the bud neck on Iqg1p function. The data in Figure 2E demonstrate that neither Bsp1-GFP nor Bsp1-13myc accumulates at the bud neck in iqg1-1 cells held at the restrictive temperature. Both of the epitope-tagged versions of Bsp1p appear functional in that neither exhibits synthetic genetic interaction with iqg1-1 (see below).
Physical proximity of Bsp1p and Iqg1p:
The above data suggested the possibility that Bsp1p and Iqg1p directly interact. We chose to explore this possibility in vivo using the phenomenon of FRET. Briefly, excitation of the donor molecule (Iqg1-CFP) coupled to excitation and light emission of the acceptor molecule (Bsp1-YFP) indicates that the two proteins lie within an ∼5 Å radius within the cell, strongly indicating a physical interaction. The data presented in Figure 3 demonstrate that, in cells expressing the donor and acceptor molecules, a FRET signal is readily detected within the acceptor emission spectrum when excited at the donor molecule excitation wavelength (Figure 3, bottom row, right). Control experiments demonstrate that no Bsp1-YFP emission is detected when cells are subject either to CFP excitation and emission spectra (Figure 3, top row, middle) or to CFP excitation and YFP emission spectra (Figure 3, top row, right). Similarly, Iqg1-CFP emission is observed only within the appropriate emission spectrum when cells are excited within the CFP excitation spectrum (Figure 3, middle row). Importantly, Bsp1-YFP actin patch association, as seen in the left hand panels, and bottom rows of Figure 3, is not detected by the FRET analysis and this serves as a critical internal control. The data then demonstrate that the two proteins, Iqg1p and Bsp1p, lie in close proximity to each other at the bud neck only in large-budded cells.
Mutation of both iqg1 and bsp1 causes synthetic phenotypes:
Assuming that Bsp1p function is reflected in its localization pattern, it might be predicted that actin ring contraction is somehow altered in bsp1 null mutants. We therefore monitored actin ring contraction and the dynamics of actin patch repolarization to the bud neck, subsequent to ring contraction, in synchronous cultures. The kinetics of actomyosin ring contraction remained the same in bsp1 null mutants as in wild-type cells (Figure 4A). However, we consistently observed a short, but distinct, delay in the repolarization of actin patches to the bud neck following ring contraction (Figure 4, B and C). Figure 4C demonstrates the pattern of actin distribution that was scored as postcytokinetic repolarization. Next we tested for genetic interaction between the bsp1 null and iqg1-1 mutations and observed a 10-fold reduction in viability of double mutants relative to parental strains at 26°, the permissive temperature for the iqg1-1 allele (Figure 4D). When exponentially growing cells were shifted to 37°, the restrictive temperature for the iqg1-1 allele, double mutants rapidly lost viability. In contrast, bsp1Δ mutants continued to grow as wild type and iqg1-1 cells ceased dividing but retained viability over the time course of the experiment (Figure 4E). The phenotype of cells carrying both mutations is distinct from either the parental or the wild-type strains (Figure 4, F–H). Microscopic examination of the double mutant revealed that even at 26°, permissive for iqg1-1, chained cells accumulated and these cells exhibited gross morphological changes with poorly defined, broad bud-neck regions (Figure 4, D and F). Consistent with these observations, actin rings were completely absent in these cells and the overall organization of the actin cytoskeleton was severely compromised (Figure 4, E and F).
BSP1-mediated suppression of iqg1-1 does not require restoration of Bsp1p bud-neck localization:
As recorded above, Bsp1p localized to the bud neck in an Iqg1p-dependent manner that led us to examine Bsp1p localization under conditions in which the iqg1-1 phenotype was suppressed. Mutant iqg1-1 cells expressing a plasmid-encoded 2HA-Bsp1p under control of the galactose-dependent GAL1-10 promoter exhibit carbon-source-dependent phenotypic suppression (Figure 5A and B). Figure 5C demonstrates the localization of 2HA-Bsp1p to the bud neck and Figure 5D presents the quantitation of induced, full-size Bsp1p levels. The two largest isoforms of HA-tagged Bsp1p indicated by the arrowheads represent hypo- (Figure 5D, bottom) and hyperphosphorylated (Figure 5D, top) forms of the protein, as previously reported Bsp1p is subject to Cdk1p-dependent phosphorylation (Ubersax et al. 2003). Increased expression of 2HA-Bsp1p is accompanied by significantly increased degradation as evidenced by the accumulation of multiple smaller molecular-sized forms of the protein. Despite the observed suppression, Bsp1p fails to localize to the bud neck at the restrictive temperature in an iqg1-1 background. This suggests that suppression is likely to involve a bypass mechanism rather than direct restoration of iqg1-1 function. This interpretation is consistent with the failure to recruit actin to the bud neck (Figure 1C), and the observed excess cell wall deposition at the bud neck when suppression is observed (Figure 1D).
Suppression is dependent upon cell wall damage sensors but not on the downstream MAP-kinase-signaling pathway:
If suppression is mediated via a bypass mechanism, it is likely to involve the cell wall deposition pathway. We tested the dependency of suppression upon the cell wall damage-sensing proteins, Mid2p and Wsc1p (Lodder et al. 1999; Philip and Levin 2001). The data in Figure 6A demonstrate that suppression is dependent upon the presence of both proteins, independently. One possibility is that suppression required remedial cell wall biosynthesis mediated by the Mpk1p map-kinase-signaling pathway. Activation of this pathway correlates with tyrosine phosphorylation of Mpk1p, which was examined in the indicated strains (Figure 6B). Mpk1p phosphorylation did not correlate with Bsp1p-mediated suppression of the iqg1-1 mutation. Indeed, in double mutants carrying the iqg1-1 allele and either mid2Δ or wsc1Δ null mutations, along with the single wsc1Δ mutant, map kinase signaling was seen to be enhanced after prolonged incubation at 37°. This increased level of Mpk1p tyrosine phosphorylation appeared, in the main, to be temperature dependent as, apart from in the wsc1Δ strain, it did not correlate with increased Bsp1p expression. A slight increase in the tyrosine phosphorylation mark is also observed in the iqg1-1 mutant expressing increased Bsp1p levels but this does not correlate with suppression. Taken together, these data indicate that, while suppression is dependent upon Mid2p and Wsc1p function, it does not require Mpk1p-dependent signaling.
Restored bud-neck architecture and chitin synthase III activity are required for suppression:
The dependency of Bsp1p-mediated iqg1-1 suppression on Chs3p function was tested. There are several important points to note from data presented in Figure 7A. Both iqg1-1 and chs3Δ cells exhibit decreased viability at 37° in the presence of methionine, i.e., low levels of BSP1 expression. Wild-type levels of viability are restored in both strains in the absence of methionine when Bsp1p levels are increased. Viability of the double mutant iqg1-1 chs3Δ at 37° is further reduced relative to the parental strains, indicative of a synthetic genetic interaction, and inviability is not rescued by increased BSP1 expression. Potential explanations for the observed genetic interactions were that Chs3p was either inactive or mislocalized in an iqg1-1 mutant and that increased expression of Bsp1p rectified the problem. We next measured overall chitin levels in suppressed and unsuppressed strains and found that these were not altered in cells where BSP1-mediated suppression was observed. Indeed, changes in chitin levels correlated with the presence or absence of methionine in the growth media, but not with suppression (Figure 7B). Examination of the localization of a functional Chs3-GFP fusion protein revealed that the protein was found at the neck of large-budded cells in all strains. However, in the iqg1-1 background in the presence of methionine, the pattern of fluorescence is more diffuse and the bud-neck region appears much wider than in large-budded wild-type cells prior to cytokinesis and septum formation (Figure 7C). Similar observations were made in an iqg1-1 parental strain lacking the BSP1-encoding plasmid (data not shown). Direct measurement of the bud-neck width in wild-type and iqg1-1 strains in the presence and absence of increased BSP1 expression demonstrated that the mutant strain exhibited a statistically significant increase (Student's t-test, P = 0.001). Increased Bsp1p levels restored the wild-type bud-neck width (Student's t-test, P = 0.001) and this was reflected by a less diffuse Chs3-GFP pattern of fluorescence (Figure 7, C and D).
The iqg1-1 mutation confers temperature-dependent growth as a result of failure to complete cytokinesis. A screen for high dosage suppressors of this mutation identified BSP1 as an efficient dosage suppressor capable of effecting suppression even when expressed from a low-copy, centromeric plasmid (Figure 1A and data not shown). Suppression results in the successful reversal of the chained iqg1-1 phenotype and is accompanied by increased deposition of electron dense material at the bud neck (Figure 1, B and D). Septum deposition appears aberrant (Figure 1D) and is not accompanied by restoration of acto-myosin ring formation. Despite these observations, cell separation appears to be relatively efficient as the number of chained cells in the population is reduced fourfold in iqg1-1 cells with increased Bsp1p expression relative to the relevant control cells (Figure 1B). The loss of Iqg1p function can also be compensated for by increased expression of the CYK3, which also does not restore actin ring assembly (Korinek et al. 2000), and superficially it might be speculated that Bsp1p and Cyk3p act in a similar manner to restore cell growth and division. However, in contrast to CYK3, increased BSP1 dosage is unable to support growth of an iqg1 null allele, suggesting that there are at least two possible bypass mechanisms able to compensate for a failure in cytokinesis and septation. Further, it implies that Iqg1p plays a role in the polarization of vesicle transport to the bud neck, consistent with a previously established link between Iqg1p and polarity establishment and secretion (Osman et al. 2002).
Bsp1p has previously been shown to localize to actin patches and the acto-myosin contractile ring (Drees et al. 2001; Wicky et al. 2003). Because of the ability of increased BSP1 expression to suppress the iqg1-1 phenotype, we examined the localization and function of Bsp1p at the bud neck. Bsp1p localizes to a clear ring structure at the bud neck and undergoes contraction (Figure 2, A and B). Recruitment to the bud neck occurs in early anaphase prior to both the presence of fully extended anaphase B spindles and the peak of actin assembly into the contractile ring (Figure 2C). Localization of Bsp1p to the bud neck is dependent upon Iqg1p function, and FRET analysis demonstrates that the two proteins are likely to interact directly (Figure 2D and Figure 3). These data are consistent with previous data relating to the timing of Iqg1p recruitment to the bud neck (Boyne et al. 2000). However, the ability of cells to complete cytokinesis at 26° in the absence of Bsp1p indicates that the protein does not have an essential role in cytokinesis; Figure 4A indicates that actomyosin ring contraction occurs with apparently normal kinetics (compare with Figure 2B). Rather, the data in Figure 4, B and C, demonstrate that repolarization of the actin cytoskeleton to either side of the bud neck following contraction is delayed in bsp1 null mutants, raising the possibility that Bsp1p acts in some way to coordinate contractile ring disassembly and repolarization of the actin cytoskeleton.
Genetic interaction between iqg1-1 and bsp1Δ was tested and a synthetic slow-growth phenotype was coupled to reduced viability observed at 26° (Figure 4D). Double mutants also exhibited an increased rate of cell death when incubated at 37°, the restrictive temperature for the iqg1-1 allele (Figure 4E). Cells exhibited a chained phenotype, the bud-neck structure was altered, and actin rings were never observed at the junction of cell bodies (Figure 4, F–H). At 26° iqg1-1 bsp1Δ cells failed to divide and showed gross morphological alterations, which correlates with disorganization of the actin cytoskeleton although some vestige of polarization was retained (Figure 4H). Bsp1p may be required for actin recruitment to the contractile ring when Iqg1p function is compromised although the failure to assemble an actin ring may be an indirect consequence of the observed morphological changes. Those changes suggest a role for both Iqg1p and Bsp1p in coordinating the reorganization of the actin cytoskeleton during and after cytokinesis. It remains possible that the morphological aspects of the synthetic phenotype may be secondary to initial failure to perform cytokinesis, but the morphology is distinct from that of iqg1-1 held at the restrictive temperature. This interpretation is again consistent with the observation that Iqg1p interacts with components of the secretory and polarity establishment pathways (Osman et al. 2002).
Bsp1p-mediated suppression of the iqg1-1 cytokinetic failure then appears to act through increased deposition of electron-dense material at the bud neck. One possibility was that this bypass mechanism required the action of the PKC–MAP kinase cell integrity pathway (Levin 2005). Consistent with this hypothesis, suppression was dependent upon Wsc1p and Mid2p (Lodder et al. 1999; Philip and Levin 2001), two plasma membrane-associated sensors of cell wall damage required for activation of the PKC–MAP kinase pathway (Figure 6A). However, despite the requirement for these two proteins, the PKC–MAP kinase pathway was not activated during Bsp1p-mediated suppression, as judged by Mpk1p tyrosine phosphorylation (Figure 6B). This result implies that suppression was dependent upon functions of these proteins not associated with detection of cell wall damage. Recent evidence has demonstrated that Wsc1p functions in the regulation of Rho3p/Rho4p-mediated polarized exocytosis (Fernandes et al. 2006). Moreover, the fact that no synthetic genetic interactions are observed between iqg1-1 and either mid2Δ or wsc1Δ suggests that neither of these proteins are normally required for Iqg1p function. One possibility, then, is that increased expression of Bsp1p links to the Rho3p/Rho4p pathway, and the data presented here are a reflection of that. Further detailed study of the regulation of the individual components of this pathway will be required to confirm this hypothesis.
Bsp1p function is not strictly required for either actin ring assembly or contraction. Rather, it is likely to be an accessory factor that serves to coordinate contraction with subsequent reorganization of the actin cytoskeleton. Despite the fact that Bsp1p is an integral component of the acto-myosin contractile ring, suppression of iqg1-1 is not associated with the restoration of actin ring formation at the bud neck. Increased Bsp1p expression allows suppression of iqg1-1 via a bypass mechanism that requires remedial septum formation and, accordingly, was found to be CHS3 dependent (Figure 7A). Despite this, there was no overall increase in chitin levels associated with suppression (Figure 7B). We were unable to test for CHS2 dependency as the null mutant was inviable in our strain background. Interestingly, chs3Δ and iqg1-1 mutations exhibited a synthetic lethal phenotype at 37°, and indeed the double mutant grows more slowly than either parental strain (data not shown). It is possible that the synthetic lethality is indicative of a role in the same pathway for the two proteins. Alternatively, it might simply be that cell survival, in the absence of cytokinesis and either normal or remedial septum formation, still requires chitin deposition in the bud-neck region. BSP1-mediated suppression of iqg1-1, however, is clearly dependent on Chs3p function, consistent with the construction of a remedial septum. Moreover, increased Bsp1p levels and suppression correlate with a reduced bud-neck width. Taken together with the gross morphogenetic changes in the iqg1-1 bsp1Δ cells, we favor the conclusion that Bsp1p is required for the maintenance of normal bud-neck architecture, which could reflect a role in polarized exocytosis. An alternative explanation is that the restoration of septation causes the reduction in bud-neck width, and it is difficult to discern between the two possibilities. Overall, the data suggest that Bsp1p has an auxiliary role in actomyosin ring function and a potential role in polarized growth, and it is the latter function that permits the observed suppression.
The original high-copy suppression plasmid, not described in this manuscript, was isolated by D. Kezenman-Pereira. This work was supported by a project grant (989/C18794) and a graduate stipend award (E.M.) from the Biotechnology and Biological Sciences Research Council.
↵1 These authors contributed equally to this article.
Communicating editor: M. D. Rose
- Received October 2, 2007.
- Accepted February 5, 2008.
- Copyright © 2008 by the Genetics Society of America