Abstract
A formin Bni1p nucleates actin to assemble actin cables, which guide the polarized transport of secretory vesicles in budding yeast. We identified mutations that suppressed both the lethality and the excessive actin cable formation caused by overexpression of a truncated Bni1p (BNI1ΔN). Two recessive mutations, act1-301 in the actin gene and sla2-82 in a gene involved in cortical actin patch assembly, were identified. The isolation of sla2-82 was unexpected, because cortical actin patches are required for the internalization step of endocytosis. Both act1-301 and sla2-82 exhibited synthetic growth defects with bni1Δ. act1-301, which resulted in an E117K substitution, interacted genetically with mutations in profilin (PFY1) and BUD6, suggesting that Act1-301p was not fully functional in formin-mediated polymerization. sla2-82 also interacted genetically with genes involved in actin cable assembly. Some experiments, however, suggested that the effects of sla2-82 were caused by depletion of actin monomers, because the temperature-sensitive growth phenotype of the bni1Δ sla2-82 mutant was suppressed by increased expression of ACT1. The isolation of suppressors of the BNI1ΔN phenotype may provide a useful system for identification of actin amino-acid residues that are important for formin-mediated actin polymerization and mutations that affect the availability of actin monomers.
THE actin cytoskeleton plays essential roles in a diverse set of cellular processes, including cell polarization, cytokinesis, cell adhesions, and endocytosis. The dynamic organization of the actin cytoskeleton is spatially and temporally regulated. How the actin cytoskeleton assembles and functions, including how its assembly relates to its function, are fundamental questions in cell biology.
The budding yeast Saccharomyces cerevisiae is an excellent model system for studies of the actin cytoskeleton dynamics because yeast cells have a relatively simple actin cytoskeleton and offer powerful experimental tools. Throughout the yeast cell cycle, precisely choreographed changes in the organization of the actin cytoskeleton underlie spatial control of cell-surface growth and thereby determine cell morphology. Extension of the cell surface is preceded by the polarized organization of two actin-filament-containing structures: actin cables and cortical actin patches.
Actin cables, long bundles of actin filaments, are highly dynamic structures containing actin (Act1p), fimbrin (Sac6p), and tropomyosin (Tpm1p and Tpm2p). They originate from polarized sites such as a bud tip and extend throughout the mother cell along the cell axis. One major function of actin cables is to serve as tracks for the polarized transport of secretory vesicles by Myo2p, a type V myosin (Govindan et al. 1995; Pruyne et al. 1998). Bni1p and Bnr1p, members of the formin protein family, play an essential role in actin cable assembly (Evangelista et al. 2002; Sagot et al. 2002a) by stimulating actin nucleation (Pruyne et al. 2002; Sagot et al. 2002b). Formins are a family of highly conserved eukaryotic proteins implicated in a wide range of actin-based processes. These proteins are characterized by the presence of the juxtaposed formin homology (FH) domains, FH1 and FH2 (Evangelista et al. 2003). The proline-rich FH1 domain binds to the actin-monomer-binding protein profilin (Pfy1p), whereas the FH2 domain is sufficient for actin filament nucleation in vitro. Moreover, profilin–actin enhances nucleation by the FH1–FH2 domain but not the FH2 domain alone, suggesting that the interaction between the FH1 domain and profilin is required for increased nucleation (Pring et al. 2003). A 12S complex termed the polarisome comprises Bni1p, Spa2p, Pea2p, and Bud6p (Sheu et al. 1998), all of which are required for apical growth. In their absence, cells fail to confine a growth site to a small region during initial bud emergence and bud growth, resulting in a widened mother-bud neck and spherical, rather than ellipsoid, cell morphology (Sheu et al. 2000). Bud6p, like profilin, interacts with actin monomers and can enhance actin nucleation by Bni1p (Moseley et al. 2004).
Cortical actin patches and their associated proteins are involved in the internalization step of endocytosis (Engqvist-Goldstein and Drubin 2003). Cortical patches are associated with invaginations of the plasma membrane (Mulholland et al. 1994), have life times of only 5–20 sec (Smith et al. 2001), and are highly motile (Doyle and Botstein 1996; Waddle et al. 1996; Carlsson et al. 2002). The formation and reorganization of cortical actin patches are regulated by cortical-patch-like protein structures, including the Arp2/3 complex and several of its activators, and endocytic adaptors and scaffolds (Pruyne and Bretscher 2000). Sla2p, one of the patch components, contains an AP180 N-terminal homology (ANTH) domain that interacts with inositol phospholipids (Sun et al. 2005). In addition, Sla2p contains a central coiled-coil region and a talin homology domain at its C terminus (Yang et al. 1999). The talin homology domain binds to filamentous actin and is localized to actin structures when it alone is expressed in vivo (Yang et al. 1999). Real-time analyses of live sla2Δ mutant cells expressing enhanced green fluorescent protein-tagged patch assembly proteins revealed the formation of actin comet tail-like structures and inhibition of endocytic internalization, demonstrating that Sla2p negatively regulates assembly of actin filaments associated with endocytic vesicles (Kaksonen et al. 2003).
In this study, we found two new mutant alleles, act1-301 and sla2-82, which suppressed growth defects caused by overexpression of an N-terminally truncated (dominant active) Bni1p. Both mutations suppressed the accumulation of the actin-cable-like structures induced by the truncated Bni1p. Our system provides a useful method to identify genes that affect the assembly or dynamics of actin cables.
MATERIALS AND METHODS
Strains, media, and genetic techniques:
Yeast strains used in this study are listed in Table 1 with their genotypes. Strains were grown in rich YPDA medium [1% yeast extract (Difco Laboratories, Detroit), 2% bacto-peptone (Difco), 2% glucose, and 0.01% adenine] or YPGA medium (1% yeast extract, 2% bacto-peptone, 3% galactose, 0.2% sucrose, and 0.01% adenine). The lithium acetate method was used for introduction of plasmids into yeast cells (Elble 1992; Gietz and Woods 2002). Yeast strains carrying complete gene deletions (bni1Δ, bud6Δ, end3Δ, mti1Δ, and vrp1Δ); GAL1 promoter-inducible 3HA-BNI1Δ1-1226, 3HA-BNR1Δ1-755, and 3HA-BNI1; and green fluorescent protein (GFP)-tagged MYO2 were constructed by PCR-based procedures as described (Longtine et al. 1998). The pfy1-116 and las17-11 mutants were constructed by three successive backcrosses into the YEF473 background strain. To construct YEF1159 (sla2-Δ1 PGAL1-BNI1ΔN), the PGAL1-BNI1(1227-1953) fragment was integrated at the LEU2 locus of YEF186 (sla2-Δ1). Strains carrying plasmids were selected in synthetic medium (SD) containing the required nutritional supplements. When indicated, synthetic galactose medium containing 0.5% casamino acids and 20 μg/ml tryptophan (SGA–Ura) was used. Standard genetic manipulations of yeast cells were performed as described previously (Sherman 2002).
S. cerevisiae strains used in this study
Molecular biological techniques:
PCR amplification was performed using a GeneAmp PCR System 9700 (Applied Biosystems, Foster City, CA). The DNA sequences of all constructs containing amplified PCR products were confirmed using an ABI PRISM BigDye Terminator Cycle sequencing ready reaction kit (Applied Biosystems) according to the manufacturer's protocol. DNA sequences were obtained using an ABI PRISM 310 DNA sequencer (Applied Biosystems). The plasmids used in this study are listed in Table 2. Schemes for the construction of plasmids and the sequences of PCR primers are available upon request. Two-hybrid analysis was performed using strain L40 containing the LexA–DNA-binding domain fusion plasmid and the Gal4-transcriptional activation domain fusion plasmid. The extent of the two-hybrid interaction was assayed by growth on SD–Trp–Leu–His plates.
Plasmids used in this study
Isolation of mutations that suppress the lethality caused by the overexpression of Bni1p(1227-1953):
To isolate spontaneous mutants that overcome the growth inhibition caused by overexpression of the dominant active formin, YKT380 cells—in which BNI1Δ1-1226, the sole copy of BNI1 in these cells, was expressed under the control of a galactose-inducible GAL1 promoter (HIS3∷PGAL1-3HA-BNI1Δ1-1226)—were plated on 30 YPGA plates (1 × 107 cells/plate) and grown at 25° for 3 days. Eighty-seven revertants were picked up. To eliminate mutants in which the GAL1 promoter was not induced, we transformed these mutants with pSMA111 (pRS316-PGAL10-GIN11) or pRS316-PGAL7-LacZ, which contained the GIN11 gene that inhibits growth when overexpressed (Kawahata et al. 1999) and the LacZ gene under the control of galactose-inducible promoters, respectively. β-Galactosidase activity was measured by a 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside filter assay (Vojtek et al. 1993). Mutations in nine mutants were judged to specifically suppress the growth inhibition caused by Bni1p(1227-1953) overexpression. Tetrad analysis revealed that three of these nine clones carried a single recessive mutation that was responsible for the suppression. We found that the growth of two of these mutants was inhibited at 37° on glucose-containing medium. This temperature-sensitive growth phenotype cosegregated with the suppression of growth inhibition by Bni1p(1227-1953) overexpression and was suppressed by a single copy of BNI1. Another clone also showed a temperature-sensitive growth phenotype on glucose-containing medium, but this phenotype was not suppressed by a single copy of BNI1. To clone the mutated genes, we transformed the mutants with a YCp50–LEU2-based genomic library and isolated transformants that grew on glucose-containing plates at 37° but not on galactose-containing plate at 25°. Library plasmids were isolated, subcloning was performed by standard methods, and, using the linkage analysis, one allele each of act1, pat1, and sla2 was isolated. During the course of cloning, we found that a plasmid harboring ACT1 (YCp50–LEU2-2.4.1) suppressed the growth defect of the bni1Δ sla2-82 mutant (see results).
Cloning of act1-301 and sla2-82:
To construct pRS316-ACT1, a BamHI–EcoRI ACT1 genomic fragment isolated from YCp–LEU2-2.4.1 was subcloned into the BamHI–EcoRI gap of pRS316. pRS316-act1-301 was constructed by the gap-repair method using pRS316-ACT1 linearized with MunI. To construct pRS313-SLA2gap, a 574-bp EcoRI–AatII fragment and a 502-bp AatII–SalI fragment containing 5′- and 3′-flanking regions of the SLA2 gene, respectively, were amplified by PCR and cloned into the EcoRI–SalI gap of pRS313. SLA2 and sla2-82 were cloned by the gap-repair method using pRS313-SLA2gap linearized with AatII.
Microscopic observations:
Cells expressing Myo2p–GFP were fixed for 10 min by the direct addition of 37% formaldehyde stock solution (Wako Pure Chemical Industries, Osaka, Japan) to a 3.7% final concentration in medium and observed using a GFP bandpass filter set (excitation, 460–500 nm; dichroic mirror, 505 nm; emission, 510–560 nm). To visualize the actin cytoskeleton, cells were fixed as described above, harvested, and fixed again for 1 hr in phosphate-buffered saline (PBS) containing 3.7% formaldehyde. Fixed cells were incubated for 30 min at room temperature in 1 μm tetramethylrhodamine isothiocyanate (TRITC)-phalloidin (Sigma Chemical, St. Louis). Following three washes with PBS, cells were mounted in 50% glycerol containing n-propyl gallate (Wako). Cells were observed using a G-2A TRITC filter set (excitation, 510–560 nm; dichroic mirror, 575 nm; emission, 590 nm). Microscopic observation was performed as described previously (Mochida et al. 2002). To quantify the polarized localization of Myo2p–GFP, at least 200 small-budded cells were randomly selected and observed. Small buds were identified to be <30% the size of the mother cell.
RESULTS
Identification of act1-301 and sla2-82 mutations that suppress the growth inhibition caused by overexpression of a dominant active Bni1p:
Overexpression of NH2-terminally truncated forms of Bni1p or Bnr1p leads to disorganization of the actin cytoskeleton, including the production of an excessive number of actin cables and the depolarization of actin patches, resulting in inhibition of cell growth (Evangelista et al. 1997; Kikyo et al. 1999). To search for regulators of actin cytoskeletal reorganization, we screened for mutations that suppressed the growth inhibition caused by overexpression of NH2-terminally truncated Bni1p. PGAL1-3HA-BNI1Δ1-1226 (PGAL1-BNI1ΔN) mutant cells overexpress a truncated Bni1p, Bni1p(1227-1953), which contains the FH1 and FH2 domains. This mutant does not grow in galactose medium, in which the GAL1 promoter is induced. We screened for spontaneous revertants that could grow in galactose medium. We isolated three independent revertants that each carried a single recessive mutation and cloned the affected genes. The first gene, ACT1, is the only gene that encodes actin in budding yeast. The second gene, PAT1, is involved in mRNA decapping (Bonnerot et al. 2000). The third gene, SLA2, codes for Sla2p, a protein involved in the assembly of cortical actin patches (Holtzman et al. 1993; Raths et al. 1993). We next investigated how the act1 and sla2 mutant alleles (designated act1-301 and sla2-82, respectively) suppressed the phenotypes caused by overexpression of a truncated Bni1p. Both act1-301 and sla2-82 also suppressed the growth defect of BNR1ΔN-overexpressing cells as well as that of the BNI1ΔN-overexpressing cells (Figure 1A). PGAL1-BNI1ΔN cells carrying act1-301 or sla2-82 showed normal spherical morphology and polarized distributions of cortical actin patches and actin cables (Figure 1B), indicating that these mutations suppressed the excessive formation of actin cables. Interestingly, both act1-301 and sla2-82 showed synthetic growth defects with bni1Δ at higher temperatures (Figures 1C, 3A, and 5B).
Identification of act1-301 and sla2-82. (A) act1-301 and sla2-82 mutations suppress growth defects caused by overexpression of a dominant active BNI1 or BNR1. Strains grown on a YPDA plate were streaked onto a YPGA plate, followed by incubation at 30° for 3 days. WT (wild-type strain), YKT38; PGAL1-BNI1ΔN, YKT380; PGAL1-BNI1ΔN act1-301, YKT967; PGAL1-BNI1ΔN sla2-82, YKT846; PGAL1-BNR1ΔN, YKT542; PGAL1-BNR1ΔN act1-301, YKT991; PGAL1-BNR1ΔN sla2-82, YKT992. (B) act1-301 and sla2-82 suppress the morphological and actin cytoskeletal defects caused by overexpression of a dominant active BNI1. Strains grown in YPDA were shifted to YPGA medium and cultured at 30° for 12 hr. Cells were fixed, labeled with TRITC–phalloidin, and visualized by differential interference contrast or with a TRITC filter. PGAL1-BNI1ΔN/BNI1 (control), YKT379; PGAL1-BNI1ΔN/BNI1 act1-301/act1-301, YKT1091; PGAL1-BNI1ΔN/BNI1 sla2-82/sla2-82, YKT1092. Bar, 5 μm. (C) Both act1-301 and sla2-82 show synthetic growth defects with bni1Δ. Strains were streaked onto YPDA plates, followed by incubation at 25° for 3 days or at 37° for 2 days. WT, YKT38; act1-301, YKT986; sla2-82, YKT847; bni1Δ, YKT446; bni1Δ act1-301, YKT968; bni1Δ sla2-82, YKT853.
Act1-301p seems to impair Bni1p-catalyzed actin polymerization:
We cloned the act1-301 mutant gene and determined its DNA sequence. act1-301 contained a point mutation resulting in the substitution of a Lys residue for a Glu residue at amino-acid position 117. For act1-301 mutant cells, the growth rates at low and high temperatures, cell morphology, polarized distributions of actin patches and cables, and fluid-phase endocytosis as assayed by the uptake of Lucifer yellow were indistinguishable from wild-type cells (data not shown), indicating that the E117K substitution did not severely impair the functions of Act1-301p.
The FH2 domain of Bni1p is required for actin cable assembly in vivo (Evangelista et al. 2002; Sagot et al. 2002a) and stimulates actin polymerization in vitro (Pruyne et al. 2002; Sagot et al. 2002b). In the absence of BNR1, a structural and functional homolog of BNI1 with a minor role, three temperature-sensitive alleles of BNI1 that produce proteins with amino-acid substitutions in the FH2 domain—bni1-116 (V1475A, K1498E, and D1511N) (Kadota et al. 2004), bni1-11 (D1511G and K1601R) (Evangelista et al. 2002; Pruyne et al. 2002), and bni1-FH2#1 (R1528A and R1530A)—cause actin cables to rapidly disappear when the mutant cells are incubated at high temperatures. The Bni1FH2#1 protein also exhibits defective actin polymerization in vitro (Sagot et al. 2002b). We found that the phenotype caused by bni1-11 was the result of a single amino-acid substitution, D1511G (data not shown). Similar to bni1Δ, the bni1 alleles bni1-116, bni1(D1511G), and bni1-FH2#1, when combined with act1-301, caused the temperature-sensitive growth phenotypes (Figure 2A). These results suggest that the actin-polymerizing activity of Bni1p is relevant to the genetic interaction of bni1 mutations with act1-301.
Defects caused by act1-301 are tied to Bni1p-mediated actin polymerization. (A) act1-301 shows synthetic growth defects with mutations in the FH2 domain of Bni1p. The bni1Δ act1-301 strain (YKT968) was transformed with pRS314 (mock), pRS314-BNI1 (BNI1), pRS314-bni1-116 (116), pRS314-bni1(D1511G) (D1511G), or pRS314-bni1-FH2#1 (FH2#1). The transformants were streaked onto a YPDA plate, followed by incubation at 37° for 2 days. (B) Mislocalization of Myo2p–GFP in bni1 act1-301 mutant. YKT662 (MYO2–GFP), YKT1058 (act1-301 MYO2–GFP), YKT1057 (bni1-116 MYO2–GFP, wild-type), and YKT1060 (bni1-116 act1-301 MYO2–GFP) were grown in YPDA medium at 25° and then shifted to 37° for 5 min. Cells were fixed and observed by fluorescence microscopy. (C) A time course of the polarized localization of Myo2p–GFP in small-budded cells. The strains described in B were grown in YPDA medium at 25° and then shifted to 37° for the indicated periods of time. Cells were fixed and at least 200 small-budded cells were observed for each data point by fluorescence microscopy.
The temperature-sensitive growth of bni1Δ act1-301 mutant cells seems to be due to defective actin cable formation. Labeling of actin cables with phalloidin, however, did not allow us to examine this possibility, because almost no actin cables could be detected in the bni1Δ cells (data not shown). Therefore, we visualized a GFP-tagged type V myosin, Myo2p–GFP, whose polarized localization depends on actin cables (Karpova et al. 2000) in bni1-116 act1-301 mutant cells. Because a fraction of Myo2p–GFP is localized at bud tips in an actin-independent manner (Ayscough et al. 1997), we examined the time course of polarized Myo2p–GFP localization after a temperature upshift. act1-301 alone did not affect Myo2p–GFP localization after a 5-min incubation at 37°, but, after 10 and 30 min, slightly enhanced the mislocalization of Myo2p that was seen in wild-type cells. In the bni1-116 single mutant, Myo2p–GFP was slightly mislocalized after a 5-min incubation at 37° and completely mislocalized after 30 min. Myo2p–GFP was completely mislocalized even after a 5-min incubation at 37° in the bni1-116 act1-301 mutant (Figure 2, B and C). These results suggest that act1-301 exacerbates the actin-cable-assembly defects in the bni1-116 mutant.
act1-301 genetically interacts with a mutation in profilin and BUD6:
The genetic interactions between act1-301 and other genes related to Bni1p function were further investigated. act1-301 showed synthetic growth defects with the mutant profilin allele pfy1-116 and bud6Δ, but not with spa2Δ (Figure 3A). Interestingly, the synthetic growth defect with bud6Δ was more severe: the bud6Δ bni1-116 act1-301 mutant did not grow even at 30°. Pfy1p and Bud6p interact with the FH1 and the COOH-terminal domains of Bni1p, respectively (Evangelista et al. 1997). These proteins specifically bind actin monomers to stimulate nucleotide exchange (Moseley et al. 2004). We examined whether the interactions of Bni1p with these proteins are required for act1-301 mutant cells to grow. Truncation mutants of Bni1p, which lacked the domains that interact with Rho/Cdc42p (ΔRBD), Spa2p (ΔSBD), Pfy1p (ΔFH1), or Bud6p (ΔBBD), or lacked the FH2 domain (ΔFH2), were expressed in bni1Δ act1-301 mutant cells and cell growth at 37° was examined (Figure 3B). Proper expression of the mutant proteins was confirmed by immunoblotting for myc-tagged versions of the truncated Bni1ps (data not shown). ΔSBD fully restored growth, whereas ΔRBD only weakly restored it. ΔFH1, ΔFH2, and ΔBBD, however, failed to restore any growth. Our results suggest that act1-301 affected the assembly of actin cables, a process in which Bni1p acts to polymerize actin through interactions with Pfy1p and Bud6p. We examined the physical interaction between Act1-301p and Pfy1p or Bud6p by the two-hybrid method. Act1-301p, however, interacted with both Pfy1p and Bud6p to an extent similar to that of Act1p (Figure 3C). Furthermore, when we examined the interaction between Act1-301p and Bni1p fragments containing the FH1 and FH2 domains, we found that Act1-301p and Act1p interacted with Bni1p(1227-1953) to a similar extent. Bni1p(1227-1750), which lacked the Bud6p-binding domain, also interacted with both Act1-301p and Act1p (Figure 3C), but Bni1p(1348-1953), which lacked the FH1 domain, did not (data not shown). Since Pfy1p binds to the FH1 domain, the two-hybrid interaction of Bni1p with actin may be mediated by the interaction between Bni1p and Pfy1p.
The profilin- and Bud6p-related functions of Act1-301p are defective. (A) act1-301 shows synthetic growth defects with bud6 and pfy1-116. Strains were streaked onto YPDA plates, followed by incubation at the indicated temperatures for 2 days. Symbols indicate relative growth rates from wild type (+++) to no growth (−). (B) The interaction of Bni1p with Pfy1p or Bud6p, but not with Rho/Cdc42p or Spa2p, is required for Bni1p to alleviate the act1-301 phenotypes. (Right) The structures of the truncated Bni1p proteins. The numbers indicate the positions of the amino-acid residues. The represented domains are the Rho-binding domain (RBD), the Spa2p-binding domain (SBD), formin homology domains (FH1 and FH2), and the Bud6p-binding domain (BBD). The bni1Δ act1-301 strain (YKT968) was transformed with pRS314 (mock), pRS314-BNI1 (Full), pRS314-bni1Δ6-642 (ΔRBD), pRS314-bni1Δ826–986 (ΔSBD), pRS314-bni1(1–1750) (ΔBBD), pRS314-bni1Δ1228–1414 (ΔFH1), or pRS314-bni1Δ1553–1646 (ΔFH2). The transformants were streaked onto a YPDA plate, followed by incubation at 37° for 2 days. (C) Act1-301p interacts with Pfy1p, Bud6p, Bni1p(1227–1953), and Bni1p(1227–1750) in the two-hybrid method. DNA fragments encoding ACT1 or act1-301 were cloned into the pACTII vector (Gal4-transcriptional activation domain vector) and the resultant plasmids were introduced into an L40 cell expressing Pfy1p, Bud6p, Bni1p(1227–1953), or Bni1p(1227–1750) fused with LexA–DNA-binding domain. Each transformant was patched onto an SD plate without histidine, followed by incubation at 30° for 2 days. Protein interactions were examined qualitatively by histidine auxotrophy.
The previously isolated act1-101 mutation also acts as a suppressor of the growth defects caused by the BNI1ΔN overexpression:
Our identification of act1-301 suggests that isolation of suppressor mutations of the BNI1ΔN-overexpressing mutant is useful to identify amino-acid residues that are specifically required for the actin cable assembly. Previously, a large collection of act1 mutant alleles, which affect residues over the surface of Act1p protein, was generated by alanine-scanning mutagenesis (Wertman et al. 1992). We examined five of these alleles and four other temperature-sensitive act1 mutant alleles for the suppression of growth defects of the BNI1ΔN-overexpressing cells. Interestingly, act1-101 (D363A, E364A) displayed the suppression (Figure 4). Aspartate 363 and glutamate 364 of Act1p also may be involved in the formin-mediated actin poly-merization.
Effect of various act1 alleles on the growth defect caused by overexpression of a dominant active BNI1. act1 strains were transformed with pRS316-PGAL1-HA (control) or pRS316-PGAL1-HA-BNI1ΔN (PGAL1-BNI1ΔN). The transformants were streaked onto an SGA–Ura plate, followed by incubation at 30° for 2 days. Representative results were shown in A and the results were summarized in B. Strains used were as follows: ACT1, YKT38; act1-301, YKT986; act1-3, DDY335; act1-101, DDY338; act1-120, DDY347; act1-124, DDY349; act1-2, DDY311; act1-4, DDY334; act1-119, DDY346; act1-125, DDY377; act1-159, DDY1492.
sla2-82 genetically interacts with mutations in other actin-cable-related genes:
Phenotypic suppression of PGAL1-BNI1ΔN by sla2-82 and the synthetic growth defect between bni1Δ and sla2-82 are surprising, because Sla2p is a component of the machinery that assembles cortical actin patches. When sla2-82 was combined with bni1-116, bni1(D1511G), or bni1-FH2#1, the resulting double mutants showed a temperature-sensitive growth phenotype (Figure 5A), indicating that the actin-polymerizing activity of Bni1p is relevant to the genetic interaction of bni1 mutations with sla2-82. We further examined the genetic interactions of sla2-82 with mutations in actin-cable-related genes and found that sla2-82 exhibited synthetic growth defects with both pfy1-116 and bud6Δ (Figure 5B). Experiments using the BNI1 deletion constructs schematized in Figure 3B also suggested that the interaction of Bni1p with Pfy1p and Bud6p was required for sla2-82 mutant cells to grow (Figure 5C). Defective assembly of actin cables in the bni1Δ sla2-82 mutant was further examined by the observation of Myo2p–GFP. sla2-82 alone slightly enhanced Myo2p–GFP mislocalization in wild-type cells after incubations of 5–30 min at 37°. In contrast, Myo2p–GFP was completely mislocalized even after a 5-min incubation at 37° in the bni1-116 sla2-82 mutant (Figure 6, A and B). These results suggest that sla2-82 exacerbates the defects in cells carrying mutations in genes related to actin cable assembly.
sla2-82 interacts with mutations in actin-cable-related genes. (A) sla2-82 shows synthetic growth defects with mutations in the FH2 domain of Bni1p. The bni1Δ sla2-82 strain (YKT853) was transformed with pRS314 (mock), pRS314-BNI1 (BNI1), pRS314-bni1-116 (116), pRS314-bni1(D1511G) (D1511G), or pRS314-bni1-FH2#1 (FH2#1). The transformants were streaked onto a YPDA plate, followed by incubation at 37° for 2 days. (B) sla2-82 shows synthetic growth defects with bud6 and pfy1-116. Strains were streaked onto YPDA plates, followed by incubation at the indicated temperature for 2 days. Symbols indicate relative growth rates from wild-type (+++) to no growth (−). (C) Interaction of Bni1p with Pfy1p or Bud6p, but not with Rho/Cdc42p or Spa2p, is required for Bni1p to suppress the sla2-82 deficiencies. The bni1Δ sla2-82 strain (YKT853) was transformed with plasmids described in Figure 3B. The transformants were streaked onto a YPDA plate, followed by incubation at 37° for 2 days.
Defects in actin-cable-dependent functions in bni1 sla2-82 mutant cells. (A) Mislocalization of Myo2p–GFP in bni1-116 sla2-82 mutant cells. YKT662 (MYO2–GFP, wild type), YKT1059 (sla2-82 MYO2–GFP), YKT1057 (bni1-116 MYO2–GFP), and YKT1061 (bni1-116 sla2-82 MYO–GFP) were grown in YPDA medium at 25° and then shifted to 37° for 5 min. Cells were fixed and observed by fluorescence microscopy. (B) A time course of the polarized localization of Myo2p–GFP in small-budded cells. The strains described in A were grown in YPDA medium at 25° and then shifted to 37° for the indicated periods of time. Cells were fixed and at least 200 small-budded cells were observed for each time point by fluorescence microscopy.
Synthetic growth defects of the bni1Δ sla2-82 mutant may be due to depletion of the actin monomer pool:
Phenotypic suppression of PGAL1-BNI1ΔN by mutations in other components involved in actin patch assembly was examined. arp2-1, end3Δ, and las17-11, however, failed to suppress the growth defect of the PGAL1-BNI1ΔN mutant (Figure 7A). Moreover, synthetic growth defects with bni1Δ were not observed for sla1Δ, abp1Δ, arp2-1, and las17Δ (Figure 7B), suggesting that the genetic interaction with actin-cable-related mutations was specific to sla2-82.
A genetic interaction with bni1 is unique to sla2 among mutant alleles of cortical-actin-patch-related genes. (A) arp2-1, end3Δ, and las17-11 do not suppress growth defects caused by the overexpression of BNI1ΔN. Strains grown on a YPDA plate were streaked onto a YPGA plate, followed by incubation at 30° for 2 days. The strains used were the PGAL1-BNI1ΔN strain (YKT380, control) containing arp2-2 (YKT979, arp2), end3Δ (YKT980, end3), las17-11 (YKT981, las17), or sla2-82 (YKT846, sla2). (B) Among actin-patch-related genes, only the sla2-82 mutation shows synthetic growth defects with bni1Δ. Strains were streaked onto YPDA plates, followed by incubation at the indicated temperatures for 2 days. Symbols indicate the relative growth rate from wild type (+++) to no growth (−).
The functional difference between Sla2p and the actin patch proteins examined above is that Sla2p negatively regulates the assembly of actin patches (Kaksonen et al. 2003). In the sla2Δ mutant, actin is continuously nucleated from nonmotile endocytic complexes, resulting in the formation of actin comet tail-like structures. We reasoned that the sla2-82 mutation also caused the sequestration of actin monomers in the actin comet tails, which might render the impaired actin-cable-assembly machinery incapable of assembling a sufficient number of actin cables to allow polarized growth. Consistent with this hypothesis, increasing the number of copies of ACT1 suppressed the growth defect of the bni1Δ sla2-82 mutant and partially suppressed the Myo2p–GFP mislocalization in the bni1-116 sla2-82 mutant (Figure 8, A and B). Mti1p/Bbc1p, another negative regulator of actin patch assembly, inhibits Las17p from activating the Arp2/3 complex (Rodal et al. 2003). mti1Δ also exhibited a synthetic growth defect with bni1Δ (Figure 8C; Tong et al. 2001). We previously showed that mti1Δ suppressed the temperature-sensitive growth phenotype observed in cells carrying a mutation in VRP1, a positive regulator of actin patch assembly (Mochida et al. 2002). If genetic interactions between genes related to actin cables and actin patches occur depending on the availability of monomeric actin, bni1Δ may have suppressed vrp1Δ, because the size of the actin pool available for actin patch assembly would be increased in the bni1Δ mutant. The temperature-sensitive growth phenotype of vrp1Δ mutants was suppressed by bni1Δ as well as by mti1Δ (Figure 8D).
Differential genetic interactions between bni1 and regulators of actin patch assembly. (A) Suppression of the synthetic growth defects of the bni1Δ sla2-82 mutant by an increased number of copies of ACT1. The bni1Δ sla2-82 strain (YKT853) was transformed with pRS316 (control), pKT1412 (pRS316-ACT1, ACT1), or pKT1227 (pRS316-BNI1, BNI1). The transformants were streaked onto a YPDA plate, followed by incubation at 37° for 2 days. (B) Suppression of Myo2p–GFP mislocalization of the bni1-116 sla2-82 mutant by an increased number of copies of ACT1. The bni1-116 sla2-82 MYO2–GFP strain (YKT1061) was transformed with plasmids used in A. The transformants were grown in YPDA medium at 25° and then shifted to 37° for the indicated periods of time. Cells were fixed and at least 200 small-budded cells were observed for each time point by fluorescence microscopy. (C) Synthetic growth defects of bni1Δ mti1Δ mutant cells. Strains grown on a YPGA plate were streaked onto YPGA (galactose) or YPDA (glucose) plates, followed by incubation at 37° for 2 days. PGAL1-BNI1, YKT414; PGAL1-BNI1 mti1Δ, YKT984. (D) Suppression of the growth defects caused by the vrp1 mutant by bni1 as well as mti1. Strains were streaked onto YPDA plates, followed by incubation at 30° or 35° for 2 days. WT, YKT38; bni1Δ, YKT446; mti1Δ, YKT189; vrp1Δ, YKT680; bni1Δ vrp1Δ, YKT982; mti1Δ vrp1Δ, YKT983.
Deletion analysis of SLA2 for the suppression of growth defects of the BNI1ΔN-overexpressing cells:
Several structural motifs and functional domains have been identified in Sla2p (Yang et al. 1999; Sun et al. 2005) (Figure 9). The NH2-terminal ANTH domain of Sla2p binds to PtdIns(4,5)P2 (Sun et al. 2005), whereas the COOH-terminal talin-like domain binds to filamentous actin (McCann and Craig 1997). In the middle of Sla2p, there are a proline-rich region, a glutamine-rich region, a leucine-zipper domain, and predicted coiled-coil regions. To obtain structural information about Sla2-82p, sla2-82 was cloned and sequenced. sla2-82 contained a nonsense mutation at codon 491, which resulted in the production of Sla2pΔ(491–968). In contrast to the sla2Δ mutant, which shows a temperature-sensitive growth phenotype (Holtzman et al. 1993), the sla2-82 mutant grew normally at 37°. In addition, these cells showed a slight impairment in the polarized organization of cortical actin patches and actin cables and a weakly impaired uptake of the endocytic marker Lucifer yellow (data not shown), suggesting that Sla2-82p was partially functional.
To examine the effect of the deletion of other regions of Sla2p on the growth of the Bni1pΔN-overexpressing cells, we expressed SLA2 deletion mutants constructed previously (Yang et al. 1999) in cells overexpressing BNI1ΔN (Figure 9). sla2Δ502-968, which is deleted for nearly the same region with sla2-82, displayed the suppression. However, sla2Δ768-968, which is deleted for only the talin-like domain, did not display the suppression, suggesting that the deletion of the middle region is responsible for the suppression. Consistently, sla2Δ360-575, which is deleted for the middle region, displayed the suppression. Interestingly, sla2Δ33-359, sla2Δ33-501, and sla2Δ33-750, which are deleted for the NH2-terminal region, did not display the suppression, irrespective of the presence or absence of the middle region, suggesting that the NH2-terminal region is required for the suppression. Consistently, the sla2Δ mutant did not display the suppression (data not shown). However, NH2-terminal deletion mutants as well as sla2Δ were impaired in growth at 30° (data not shown), raising a possibility that the lack of suppression was due to the impaired growth. In conclusion, our results at least suggest that deletion of the middle region of Sla2p is involved in the suppression of the growth inhibition caused by the BNI1ΔN overexpression.
Effect of sla2 deletions on the growth defect caused by overexpression of a dominant active BNI1. C-terminal deletions were examined in YKT859 (sla2Δ502-968) and YKT861 (sla2Δ768-968), which were transformed with pRS316-PGAL1-HA-BNI1ΔN. N-terminal deletions were examined in YKT1159 (sla2Δ PGAL1-HA-BNI1ΔN), which was transformed with pRS313-sla2Δ33-359, pRS313-sla2Δ33-501, pRS313-sla2Δ33-750, or pRS313-sla2Δ360-575. The transformants were examined for growth on a galactose-containing plate (SGA–Ura or YPGA) at 30° for 3 days. The numbers indicate the positions of the amino-acid residues. The represented domains or regions are the ANTH domain (ANTH), the proline-rich region (PR), the glutamine-rich region (QR), the leucine-zipper domain (LZ), predicted coiled-coil regions (CC), and the talin-like domain (talin) (Yang et al. 1999; Sun et al. 2005). An arrowhead indicates the amino-acid residue 491 that was changed to a nonsense codon in sla2-82.
DISCUSSION
act1-301 mutant cells exhibit actin-cable-specific defects:
act1-301 reduced the massive assembly of actin cables induced by overexpression of BNI1ΔN. Subsequent genetic studies suggested that Act1-301p inhibited formin-catalyzed actin assembly. Formin family proteins polymerize actin via a mechanism that is different from that mediated by the Arp2/3 complex. Surface amino-acid residues of actin that are important for the formin-catalyzed polymerization remain to be identified. Our isolation of act1-301 demonstrates that these residues could be uncovered by screening cells with the mutant alleles of ACT1 for suppressors of the BNI1ΔN-overexpression phenotype.
Glutamate 117, which was substituted to lysine in Act1-301p [Act1p(E117K)], is one of the Act1p surface amino-acid residues that were characterized previously by systematic charged-to-alanine mutagenesis (Wertman et al. 1992; Drubin et al. 1993). Mutations in the act1-119 allele cause three consecutive amino-acid substitutions: R116A, E117A, and K118A. However, act1-119 did not suppress the growth defects of the Bni1pΔN-overexpressing cells (Figure 4), suggesting that the glutamate-to-lysine substitution at amino-acid position 117 may play an important role in the suppression. In contrast, the act1-101 mutation exhibited the suppression. act1-101(D363A, E364A) mutant cells were found to have no actin cables at 37° (Drubin et al. 1993), which might be explained by weakened interactions with Pfy1p in the two-hybrid method (Amberg et al. 1995). Pfy1p and Bud6p bind to the FH1 and COOH-terminal domains of Bni1p, respectively, and these proteins promote actin monomers to exchange bound ADP for ATP (Moseley et al. 2004). In our two-hybrid experiments, Act1p(E117K) interacted normally with Pfy1p and Bud6p. act1-301, however, exhibited synthetic growth defects with both pfy1-116 and bud6Δ, raising the possibility that Act1p(E117K) is somewhat impaired in its ability to be acted on by Pfy1p or Bud6p. It is thought that formin FH2 dimers nucleate and processively cap the elongating barbed end of the actin filament, whereas profilin generates a local increase of ATP–actin monomers to promote actin assembly (Moseley et al. 2004; Romero et al. 2004; Otomo et al. 2005). The E117K and D363A E364A substitutions may interfere with a step in these processes.
Genetic interactions among genes involved in the assembly of actin cables and actin patches may reflect the availability of actin monomers:
The isolation of a mutation in SLA2 as a suppressor of the BNI1ΔN-induced hyperaccumulation of actin cables was unexpected, because Sla2p is a regulator of cortical actin patch assembly. Synthetic growth defects of the bni1Δ sla2-82 mutant also suggested that the sla2-82 mutation affected the dynamics of actin cables. Given that Sla2p is involved in the negative regulation of Arp2/3-mediated actin nucleation at endocytic sites (Kaksonen et al. 2003), we hypothesized that sla2-82 interferes with actin cable assembly by depleting the available actin monomers. Consistent with this idea, increased expression of Act1p suppressed the temperature-sensitive growth phenotype of the bni1Δ sla2-82 mutant. Suppression of vrp1Δ by bni1Δ may be similarly explained by an increased actin pool due to low levels of actin cable assembly in bni1Δ cells. In agreement with this hypothesis, an increased number of copies of ACT1 suppressed the phenotype caused by vrp1-1 (Vaduva et al. 1997). Although we cannot exclude the possibility that Sla2p is more directly involved in actin cable assembly, the following two observations that implicated Sla2p in polarized growth can be similarly explained by indirect effects due to actin depletion. First, SLA2 was identified in a collection of mutants that inhibited hyperpolarized growth in cdc34-2 mutant cells (Bidlingmaier and Snyder 2002). Second, electron microscopic studies suggested that Golgi-derived secretory vesicles accumulate in sla2 mutants (Mulholland et al. 1997; Gall et al. 2002). On the basis of the observation that yeast have low levels of free actin monomers, Karpova et al. (1995) speculated that yeast actin cytoskeletons might be very static in comparison to those of motile cells wherein actin filaments can undergo very rapid cycles of assembly and disassembly. Quantitative analysis of the recovery rates of cortical actin patches after photo-bleaching, however, demonstrated that actin assembles in yeast at rates similar to those observed in motile cells (Kaksonen et al. 2003). How two distinct actin-organizing systems for actin cables and actin patches can efficiently recruit and incorporate monomeric actin into actin filaments using a small pool of actin monomers remains an interesting research question.
Finally, we cannot exclude the possibility that Bni1p and Sla2p share a third function that is independent of polarized transport and endocytosis. A mutant allele of SLA2 was identified (mop2; modifier of Pma1p) as an enhancer of the phenotypes caused by a mutation in PMA1, a gene that encodes the plasma membrane H+-ATPase (Na et al. 1995). In the mop2 mutant, the abundance of Pma1p on the plasma membrane was reduced. SLA2 might be involved in plasma membrane integrity, because sla2 mutations increased cell lysis as assayed by staining for extracellular alkaline phosphatase activity. Moreover, the cell-lysis phenotype was more prominent in the bni1Δ sla2-82 mutant (data not shown). Other reports suggested that Sla2p might also be involved in mRNA and DNA metabolism. A sla2 mutant was identified in a collection of mutants that exhibited defects in the decay of several mRNAs (Zuk et al. 1999). In addition, sla1 and sla2 mutants were found to enhance the sensitivity of cells to a self-poisoning mutant allele of DNA topoisomerase I mutant (Fiorani et al. 2004). It remains to be clarified whether these functions of Sla2p are mediated by the actin cytoskeleton.
Acknowledgments
We thank Rinji Akada, David Drubin, and Yoshimi Takai for yeast strains and plasmids; Eriko Itoh for technical assistance; and Konomi Kamada and members of the Tanaka Lab for helpful discussion and advice during the course of this study. This work was supported by grants-in-aid for Scientific Research to T.Y. and K.T. from the Ministry of Education, Culture, Sports, Science, and Technology, Japan.
Footnotes
Communicating editor: T. Stearns
- Received January 3, 2006.
- Accepted March 10, 2006.
- Copyright © 2006 by the Genetics Society of America
