Abstract
A large collection of yeast actin mutations has been previously isolated and used in numerous studies of actin cytoskeletal function. However, the various mutations have been in congenic, rather than isogenic, backgrounds, making it difficult to compare the subtle phenotypes that are characteristic of these mutants. We have therefore placed 27 mutations in an isogenic background. We used a subset of these mutants to compare the degree to which different actin alleles are defective in sporulation, endocytosis, and growth on NaCl-containing media. We found that the three phenotypes are highly correlated. The correlations are specific and not merely a reflection of general growth defects, because the phenotypes are not correlated with growth rates under normal conditions. Significantly, those actin mutants exhibiting the most severe phenotypes in all three processes have altered residues that cluster to a small region of the actin crystal structure previously defined as the fimbrin (Sac6p)-binding site. We examined the relationship between endocytosis and growth on salt and found that shifting wild-type or actin mutant cells to high salt reduces the rate of α-factor internalization. These results suggest that actin mutants may be unable to grow on salt because of additive endocytic defects (due to mutation and salt).
ACTIN has been implicated in numerous cellular processes, including muscle contraction, cell locomotion, the intracellular movement of organelles and mRNA, the generation of cell polarity and shape, endocytosis, and translation. The role of actin in muscle contraction is well understood, but the role of actin in many other processes remains enigmatic.
The budding yeast Saccharomyces cerevisiae has emerged as a model organism for the study of actin in cellular processes, largely due to the tremendous classical and molecular genetics possible with this organism (Botsteinet al. 1997) and the fact that yeast has just a single actin gene ACT1 (Ng and Abelson 1980). For example, it has been possible to generate mutations that affect residues all over the surface of actin and therefore likely affect diverse interactions of actin with other cellular structures (Wertmanet al. 1992). These mutants have been useful in identifying roles of actin in the cell and have revealed a wide variety of phenotypes, including defects in polarized growth, endocytosis, secretion, sporulation, growth under hypertonic conditions, bud-site selection in diploids, mitochondrial organization, vacuolar inheritance, and nuclear segregation (Shortleet al. 1984; Novick and Botstein 1985; Adams and Botstein 1989; Dunn and Shortle 1990; Cooket al. 1992; Wertmanet al. 1992; Drubinet al. 1993; Kubler and Riezman 1993; Broweret al. 1995; Hillet al. 1996; Botsteinet al. 1997). However, very little is known of the detailed mechanisms by which actin normally functions in each of these processes. Indeed, the role of actin in many of these processes remains completely unknown.
We sought to further characterize three common phenotypes of actin cytoskeletal mutants: the defects in sporulation, endocytosis, and growth on high NaCl-containing media. We chose to compare these phenotypes because we observed that, of the numerous phenotypes described in the literature, these three are frequently found in the same mutants (Table 1). This observation led us to suggest that these phenotypes may be causally related. For our studies, we generated an allelic series of actin mutant strains. This collection of mutants was designed to be isogenic except at the actin locus, so that differences observed between mutants could be attributed solely to the actin alleles carried. [The large collection of mutations obtained previously by Wertman et al. (1992) is in congenic, rather than isogenic, backgrounds, making detailed comparison of phenotypes more difficult.] We characterized these mutants quantitatively for defects in endocytosis, sporulation, and growth on salt. Comparison of the extent of the defects in each mutant revealed that the three phenotypes are strongly correlated. From an analysis of the localization of altered residues on the actin crystal structure we found the alleles exhibiting the most severe defects in all three assays change residues in a small region of subdomains 1 and 2. Strikingly, these residues form part of the fimbrin (Sac6p)-binding site on actin (Hontset al. 1994; Hanein et al. 1997, 1998; Sandrocket al. 1997). These data suggest that α-factor endocytosis, sporulation, and growth on high salt all require an explicit interaction between actin and fimbrin. The relationship between endocytosis and salt sensitivity was examined, and it was found that salt inhibits endocytosis in yeast, as in mammalian cells (Daukas and Zigmond 1985; Carpentieret al. 1989; Heuser and Anderson 1989). This finding suggests that the failure of actin mutants to grow on salt may be caused by exacerbated defects in endocytosis.
Comparisons of the previously described phenotypes of actin cytoskeletal mutants
MATERIALS AND METHODS
Construction of a set of isogenic yeast strains: The yeast strains used in these experiments are derivatives of S288C and are listed in Table 2 and, except for GPY60.1, are all isogenic except at the loci indicated. To generate this collection of isogenic strains, IGY6 (DBY4975 from the Botstein lab) of mating-type α was first transformed with a HO plasmid [Ycp50-HO (Herskowitz lab); a.k.a. pAAB166] to generate an isogenic MATa strain. The lys2-801 allele from this MATa strain was replaced with lys2::HIS3 (in which the LYS2 sequence is replaced by HIS3) to create IGY2. IGY4 was derived from IGY6 by disrupting SAC6 with pAAB123 as described previously (Adamset al. 1991). IGY196 was derived from mating IGY2 and IGY4.
The bar1Δ deletion was introduced into our strains so that α-factor internalization assays could be carried out in the absence of the α-factor protease. The bar1::LYS2 mutation was introduced into IGY196 by transformation with EcoRI-digested pEK3 (obtained from Dr. Greg Payne), which disrupts the BAR1 locus with LYS2. Lys+ transformants were selected and sporulated, and tetrads were dissected. LYS2 segregated 2:2 as expected, and Lys+ segregants were subjected to a bar1Δ barrier assay (Sprague 1991). One segregant (IGY43) that showed the Bar– phenotype was chosen and hereafter will be described as the wild-type haploid strain. Transformation of IGY43 with the HO plasmid pAAB166 and subsequent mating of strains of opposite mating type resulted in generation of the isogenic diploid strain IGY153. Another Bar– segregant (IGY191) from the bar1::LYS2/+ diploid above was mated to IGY43, resulting in the diploid strain IGY123.
Yeast strains used in this study
act1 and sac6 mutants: The IGY strains with numbers 48–63 were derived from strain IGY123 using mutant actin plasmids as described previously (Wertmanet al. 1992). The IGY strains with numbers 87–119 were obtained after first recovering the various act1 mutations from the genomes of appropriate strains and then manipulating the mutant genes as described previously (Sandrocket al. 1999). Other IGY act1 mutant strains generated in this study, but not used in the phenotypic analyses described in this article, are IGY85 (act1-8::HIS3), IGY91 (act1-9::HIS3), IGY93 (act1-10::HIS3), IGY95 (act1-133::HIS3), IGY103 (act1-116::HIS3), IGY106 (act1-21::HIS3), IGY111 (act1-115::HIS3), IGY114 (act1-123::HIS3), IGY116 (act1-2::HIS3), IGY117 (act1-7::HIS3), and IGY121 (act1-102:: HIS3). In addition to these act1 a strains, an isogenic set of act1 α-strains are available (not listed).
IGY149–162, IGY193, and IGY213 were created through transforming haploid strains IGY48–109 with pAAB166 (HO URA3 CEN) to generate the corresponding isogenic a/α-diploid strains.
Finally, a sac6Δ::LEU2, bar1::LYS2 lys2::HIS3 strain (IGY44) was obtained as a segregant from IGY196 that was transformed with pEK3 (as described above).
Yeast growth and media: Growth of yeast was performed as described previously (Shermanet al. 1974). Hypersensitivity to NaCl was determined by monitoring growth of cells on YPD media containing 0–1000 mm NaCl at 25°. Growth was determined by plating suspensions of cells using a 32-point inoculator. Sporulations were performed in liquid media containing 2% potassium acetate plus any auxotrophic requirements at 23°.
Purification of 35S-labeled α-factor, binding, and internalization assays: 35S-labeled α-factor was prepared and purified as described previously using strain GPY60.1 overexpressing α-factor from the 2μ plasmid pDA6300 (MFα1, STE13, LEU2; Dulicet al. 1991; Tanet al. 1993). The binding and internalization assays were performed as described previously (Tanet al. 1993; see also legends for Figures 4 and 5), except in the experiments measuring α-factor internalization under hypertonic conditions, cells were resuspended in binding buffer containing 600 mm NaCl, 1 m sorbitol, or 1.8 m sorbitol after the removal of unbound α-factor. Experiments were performed two to four times with each strain. Typical data are shown in Figure 4. For each strain, quantitation of the rate of endocytosis (Table 3) was obtained by determining the percentage uptake per minute 5 to 15 min after the initiation of uptake. These values were then divided by the percent uptake per minute for wild-type cells to obtain a value relative to wild type.
Determination of growth rates: Individual colonies were selected and grown overnight in YPD to a density of ∼5 × 107 cells/ml at 23°. Cultures of 100 ml were inoculated with an appropriate number of cells to yield a starting density of 1 × 105 to 1 × 106 cells/ml. Cell density readings were determined using a Klett meter (Klett Mfg. Co., New York) and were taken just after inoculation and every hour afterward for 7 hr. Density readings for each strain were plotted on semilog graph paper and the doubling times were calculated from the linear regions of the graphs.
RESULTS
Creation of an isogenic collection of actin mutants: Previously, a large collection of act1 mutant alleles was generated by alanine-scanning mutagenesis (Wertmanet al. 1992). These alleles were placed in a diploid strain that was generated from two congenic haploid strains, and haploid segregants carrying the various act1 mutations were isolated (Wertmanet al. 1992). However, because the haploid parents were congenic rather than isogenic, it is difficult to be certain that differences in phenotype between the various act1 alleles are due to the actin mutations rather than genetic background effects. Moreover, actin alleles generated in other studies (e.g., Shortleet al. 1984; Adams and Botstein 1989; Johannes and Gallwitz 1991; Cooket al. 1992; Chenet al. 1993; Miller and Reisler 1995; Miller et al. 1995, 1996) have been in yet other genetic backgrounds. Therefore, it has been difficult to compare the various phenotypes among the alleles because of genetic background differences. We therefore generated a set of act1 mutant alleles in a single strain background and further characterized the mutant phenotypes (Tables 2 and 3).
As described in the Introduction and Table 1, many strains harboring mutations in genes encoding components of the actin cytoskeleton are defective in sporulation, endocytosis, and growth under hypertonic conditions, whereas many other actin cytoskeletal mutant strains are not defective in any of these processes. To test the strength of this correlation quantitatively, we examined 16 otherwise isogenic actin mutants for abilities to endocytose α-factor, sporulate, and grow under a wide range of NaCl concentrations. In addition, we determined the growth rates for the various mutants. We then used the data to determine the extent to which defects in endocytosis, sporulation, and growth under hypertonic conditions are correlated and to determine whether any of these phenotypes is associated with a general defect in cell growth.
Characterization of the ability of actin mutants to endocytose α-factor: Previous studies indicated that act1-1 and act1-2 mutants are defective in the internalization of α-factor (Kubler and Riezman 1993). We have now quantitatively examined 16 otherwise isogenic actin alleles for their ability to internalize radioactively labeled α-factor. Several of the actin mutants have a temperature-sensitive growth defect; therefore, we measured uptake of α-factor at both 25° and 37° (Table 3). After temperature shift, α-factor internalization was followed for up to 1 hr. To obtain quantitative information as to the defect exhibited by each mutant, we calculated the initial rate of α-factor uptake/minute, i.e., from 5 to 15 min (see materials and methods). The initial internalization rates for each mutant relative to wild type at 25° and 37° are summarized in Table 3. The actin mutants exhibited a wide range in their abilities to endocytose α-factor at these two temperatures. The observed defects in α-factor uptake result from the inability of act1 mutants to internalize α-factor and not from their inability to bind α-factor, because all act1 mutants tested bind approximately the same amount of α-factor as wild-type cells (data not shown).
Phenotypes of wild-type and various act1 mutant alleles
Characterization of the salt sensitivity of actin mutants: A large collection of congenic actin mutants were originally characterized by Wertman et al. (1992) for their ability to grow on 900 mm NaCl between 14° and 37°. We wished to extend this analysis by determining the salt concentration at which the various isogenic actin mutants were able to grow at 25°. Growth profiles were developed for each actin mutant, and the lowest concentration of NaCl at which each mutant fails to grow is summarized in Table 3. Clearly, the mutants exhibit wide variation in the salt concentrations at which they can grow.
Characterization of the ability of actin mutants to sporulate: It was previously shown that act1-1 and act1-2 are defective in sporulation (Novick and Botstein 1985). We have now quantitatively examined the sporulation ability of 14 otherwise isogenic actin mutants. As shown in Table 3, these mutants display a wide range in their abilities to sporulate at 23°.
Growth rates of actin mutants: To determine the degree to which the various actin mutations cause a general growth defect, we measured the growth rates for each of the mutants. As shown in Table 3, there is a wide range in growth rates.
Correlation between defects in endocytosis, sporulation, and growth on NaCl: We have compared the degree to which actin mutants can endocytose α-factor, sporulate, and grow on NaCl-containing media. To this end, we generated scatterplots comparing pairwise combinations of each of the phenotypes and drew a bestfit line for each plot (Figure 1). From these graphs, it is clear that there is a strong correlation between defects in sporulation and growth on salt, defects in internalization and growth on salt, and defects in sporulation and internalization. The correlation coefficients and probability values for each of the pairwise combinations were calculated and found to be highly significant (Table 4).
As seen in Table 3 and Figure 1B, act1-4 (and to a lesser extent act1-113) is a striking exception to the observed correlations. In particular, the act1-4 mutant allele causes a strong defect in endocytosis, but is only mildly sensitive to NaCl. Frequently, mutant phenotypes are osmoremedial; i.e., they can be suppressed by hypertonic conditions (Adamset al. 1997). We therefore tested whether NaCl treatment suppresses the defects associated with the act1-4 mutant allele, thus resulting in growth on high concentrations of NaCl. To this end, we examined the ability of act1-4 mutant cells to internalize α-factor in the presence or absence of NaCl. Analysis of α-factor uptake in act1-4 mutant cells treated with NaCl revealed that act1-4 mutants are endocytosis defective in the presence of 600 mm NaCl (data not shown). As the endocytosis defect observed in the absence of NaCl is not suppressed (and is even exacerbated) by NaCl, the act1-4 mutant is not osmoremedial. This mutant is considered further in the discussion.
Correlation coefficients (r) for correlations between phenotypes associated with actin mutant alleles
—Scatterplots of the pairwise comparisons of salt sensitivity at 25°, sporulation at 23°, and α-factor internalization at 25°. Scatterplots of sporulation and salt sensitivity (A); salt sensitivity and α-factor internalization (B); and sporulation and α-factor internalization (C).
The observed correlations are not merely a reflection of general growth ability: To determine whether the observed correlations were due simply to the sickest alleles having the most pronounced phenotypes, we measured the growth rate of each mutant and determined whether there was a correlation between doubling time and the defect in sporulation, endocytosis, or salt sensitivity. As shown in Figure 2 and Table 4, there was no significant correlation compared with those seen among endocytosis, sporulation, and salt sensitivity. These data suggest that the correlations between defects in sporulation, endocytosis, and growth on salt are not merely a reflection of general growth defects.
The most defective actin alleles have amino acid substitutions that map to the Sac6p-binding site on actin: The availability of an atomic structure of actin (Kabschet al. 1990) allows us to map defective residues causing similar phenotypes to the crystal structure to determine if the residues cluster to a particular region of actin and implicate a specific binding site for an actin-binding protein. The six mutant alleles showing the least severe defects at 23–25° in sporulation, α-factor internalization, and growth on high NaCl-containing media are act1-101, act1-104, act1-105, act1-117, act1-119, and act1-135 (Figure 1 and Table 3). As shown in Figure 3, these alleles alter residues located on the surfaces of both the large and small domains of actin. In contrast, the four actin mutant alleles displaying the most severe defects at 23–25° in sporulation, α-factor internalization, and growth on high NaCl-containing media (act1-20, act1-120, act1-122, and act1-125) change residues that cluster to the small domain of actin comprised of subdomains 1 and 2 (Figure 3). This region of the actin crystal structure has previously been described to bind fimbrin or Sac6p, an actin-bundling protein (Hontset al. 1994; Hanein et al. 1997, 1998; Sandrocket al. 1999). Specifically, biochemical cross-linking experiments indicate that act1-120, act1-125, and act1-20 are each defective in their interaction with Sac6p (Holtzmanet al. 1994; Hontset al. 1994; Sandrocket al. 1997). Consistent with this finding is that sac6Δ cells, which lack the actinbundling protein fimbrin, also exhibit severe defects in sporulation (Broweret al. 1995), α-factor internalization (Kubler and Riezman 1993), and growth on conditions of high NaCl (our unpublished data). This correlation of the localization of actin residues to the Sac6p-binding site implicates the function of Sac6p in each of these processes. We suggest that efficient internalization of α-factor, sporulation, and growth on conditions of high NaCl may all require bundled actin filaments induced by Sac6p.
—Scatterplots comparing the growth rates at 23° of actin mutants with salt sensitivity and α-factor internalization at 25°. Scatterplots of doubling times and salt sensitivity (A) and doubling times and α-factor internalization (B).
—Location of the act1 mutations to the threedimensional atomic structure of rabbit muscle actin. (A) Ribbon diagram of actin displaying the “front view” of actin with subdomain 1 (dark gray residues 1–32, 70–144, and 338–375) and subdomain 2 (black residues 33–69) emphasized (Botsteinet al. 1997). (B) Ribbon diagram of actin displaying the altered amino acids in spacefill mode. Residues depicted in black are those altered by mutations that cause the most severe defects at 23–25° in sporulation, growth on high NaCl-containing medium, and α-factor uptake. Residues depicted in white are altered by mutations that cause the least defects at 23–25° in sporulation, growth on high NaCl-containing medium, and α-factor uptake. Residues shown in black are act1-20, act1-120, act1-122, and act 1-125. Residues shown in white are act1-101, act1-104, act1-105, act1-117, act1-119, and act1-135.
Effect of NaCl on endocytosis in wild-type and actin mutant cells: To analyze the basis of the correlation between salt sensitivity and endocytosis, we first asked whether salt has an effect on the rate of endocytosis. Experiments carried out in mammalian cells have demonstrated that incubating cells in high external osmolarity decreases endocytosis (Daukas and Zigmond 1985; Carpentieret al. 1989; Heuser and Anderson 1989). It therefore seemed plausible that the failure of actin cytoskeletal mutants (which already have an endocytosis defect) to grow under hypertonic conditions may be a consequence of an exacerbated defect in endocytosis, perhaps to a level that is insufficient for growth. To test this idea, we first asked whether incubation of wild-type yeast cells in hypertonic conditions results in decreased rates of endocytosis, as in mammalian cells. The uptake of 35S-labeled α-factor was measured in the presence or absence of 600 mm NaCl or 1 m sorbitol. As seen in Figure 4A, NaCl and sorbitol both decrease the internalization rate. Further experiments examining α-factor uptake in the presence of increasing concentrations of NaCl (300, 600, 900 mm) demonstrates that endocytosis rates decrease with increasing salt concentrations (data not shown). This apparent reduction in the ability of wild-type cells to internalize α-factor under hypertonic conditions could result from a reduction in the affinity of α-factor for its receptor. However, binding experiments demonstrated that shifting cells to 600 mm NaCl after initial binding of α-factor does not alter the amount of α-factor bound to the cells (Figure 4B). Therefore, the observed defect must be due to an inability to endocytose α-factor. These data indicate that α-factor internalization is a salt-sensitive process. One possible explanation for this result is that the actin cytoskeleton becomes disorganized following osmotic shock leading to decreased endocytosis rates. If this is true, normal endocytosis rates would be restored when the actin cytoskeleton reorganizes after the initial osmotic shock, ∼1 hr after treatment (Chowdhuryet al. 1992). However, experiments examining α-factor internalization after prolonged incubations in NaCl do not result in a restoration of normal endocytosis rates (data not shown).
—(A) Internalization of 35S-labeled α-factor in wild-type (IGY43) cells under normal conditions or after shift to 600 mm NaCl or 1 m sorbitol at 25°. IGY43 cells containing bound α-factor were transferred to media lacking or containing NaCl or sorbitol at 4°. Immediately after transfer, samples were incubated at 25° for 5 min, after which 2% glucose was added to initiate α-factor internalization. At each time point, a sample was removed and the percentage of α-factor internalized relative to the amount of α-factor bound was determined as described in materials and methods. (B) Bar graph representing the percentage α-factor bound to wild-type cells after shifting cells with bound α-factor to 0 mm NaCl, 600 mm NaCl, or 1.8 m sorbitol.
To test whether salt exacerbates the existing endocytosis defects of actin mutants, we examined the ability of two mutants to internalize α-factor in the presence or absence of 600 mm NaCl. We chose to examine an actin mutant (act1-20) that exhibited an intermediate defect in α-factor internalization at 25°, so that we could detect either an increase or decrease in the rate of endocytosis. In addition, we examined sac6Δ cells because those actin alleles with the most severe phenotypes are defective in binding Sac6p (Holtzmanet al. 1994; Hontset al. 1994; Sandrocket al. 1997). As shown in Figure 5, both act1-20 and sac6Δ cells have lower endocytosis rates in the presence than in the absence of NaCl. This finding suggests that these mutants may be unable to grow under hypertonic conditions because of the additive effects of the salt and the mutations on their rates of endocytosis.
DISCUSSION
Numerous studies have addressed the role of actin in cells. To a large extent, these studies have employed drugs, antibodies, or mutations that disrupt the actin cytoskeleton. In yeast, at least 40 actin mutations have been identified through site-specific mutagenesis (e.g., Johannes and Gallwitz 1991; Wertmanet al. 1992; Chenet al. 1993; Miller and Reisler 1995; Miller et al. 1995, 1996), random mutagenesis (Shortleet al. 1984; Dunn and Shortle 1990), or classical genetic approaches such as suppressor (Adams and Botstein 1989; Sandrocket al. 1999) or noncomplementation (Welchet al. 1993) screens. Analysis of these actin mutations has revealed a diversity of phenotypes. In some cases, the mutations specifically interfere with just a subset of actin's functions, leading to the possibility of dissecting actin's functions at the molecular level.
Utility of a set of otherwise isogenic actin mutants: In this study, we have used a set of 27 actin mutations isolated previously to generate a collection of mutant strains containing identical genetic backgrounds. This collection of mutants has allowed us to compare alleles phenotypically with the certainty that differences observed are due to the actin allele being examined, rather than genetic background effects. Comparison of our salt-sensitivity data with those of Wertman et al. (1992) reveals differences that are attributable to genetic background effects. For example, we found that at 25° our act1-3, act1-124, and act1-125 strains fail to grow on 900 mm NaCl, whereas the congenic strains of Wertman et al. (1992) were able to grow at this concentration of NaCl. This observation is particularly striking for act1-125 as this mutant fails to grow on media containing concentrations of NaCl at or above 400 mm (Table 3). Additionally, we have examined the congenic actin alleles generated by Wertman et al. (1992) for their ability to grow on varying concentrations of NaCl and at different temperatures and have found significant differences between the congenic vs. isogenic actin strains (data not shown). On the basis of these observations, we stress the importance of using isogenic strains to compare physiological processes in different strains. The existence of such a collection of mutants should be extremely useful for all future studies that aim to compare the phenotypes of the different mutant alleles. We anticipate that these strains will be particularly useful for comparison of phenotypes that are easily influenced by genetic background effects.
—Internalization of 35S-labeled α-factor in (A) act1-20 (IGY87) and (B) sac6Δ (IGY44) cells in the absence of (solid lines) or after shift to 600 mm (dashed lines) NaCl at 25°. IGY44 and IGY87 cells were incubated in binding buffer for 1 hr at 4° to allow α-factor binding. Unbound α-factor was removed and α-factor-bound cells were transferred to buffer containing 0 mm or 600 mm NaCl at 4°. Immediately following transfer, samples were divided into separate tubes for each time point and incubated at 25°. After 5 min, internalization was initiated in each tube by addition of dextrose to 2%. The amount of α-factor internalized relative to α-factor bound at each time point was determined as described in materials and methods.
The salt-sensitive, endocytosis-, and sporulation-defective mutations change residues that map to the fimbrin-binding site: Our finding that actin mutations exhibiting the most severe defects in endocytosis, sporulation, and the ability to grow on high salt have altered residues that cluster to a small domain of actin suggest that this region of actin is likely important for efficient endocytosis, sporulation, and salt-resistant growth. Both genetic and biochemical cross-linking studies have demonstrated that this region of actin (subdomains 1 and 2) interacts with the actin-bundling protein fimbrin (Holtzmanet al. 1994; Hontset al. 1994; Sandrocket al. 1997). Moreover, the three-dimensional structure of actin decorated with the NH2-terminal actin-binding domain of fimbrin has implicated actin residues that are consistent with both the biochemical and genetic data (Hanein et al. 1997, 1998). We have found that not only do the most defective mutations map to the fimbrin interaction site but that actin mutations that do not exhibit defects in all three processes map to regions far away from the fimbrin interaction site (Figure 3). Considering cells lacking fimbrin also exhibit the same severe phenotypes we suggest that the actin-fimbrin interaction is the critical parameter underlying these phenotypes. A plausible model for why fimbrin is required might be that either (1) actin-bundling activity is required for each of the three processes and/or (2) the disruption of fimbrin binding leads to the improper regulation and/or binding of other actin-binding proteins. Further experiments that examine the competition between actin-binding proteins for actin may help address this issue.
Underlying basis of the correlated phenotypes: In this study, we have identified strong correlations between three phenotypes that were not previously associated with each other: defects in endocytosis, sporulation, and ability to grow on salt. The finding of a correlation between defects in endocytosis, sporulation, and growth on salt suggests that either (i) the actin-fimbrin interaction has a common role in all three of these processes and those alleles that are most defective in some common function are therefore most defective in endocytosis, sporulation, and growth on salt or (ii) the actinfimbrin interaction causes a defect in one process, such as endocytosis, which then causes a defect in the others.
A possible common role for the actin-fimbrin interaction might be in cell surface integrity, which in turn might affect the efficiency of endocytosis, sporulation, and ability to grow in the presence of elevated salt. It has been suggested that the actin cytoskeleton may stabilize the cell during cell membrane and wall growth (Mulhollandet al. 1994), and it is not hard to imagine how such stabilization might in turn affect each of the three processes of interest. In the case of act1-4 (and to a lesser extent act1-113) it is possible that this mutant does not affect cell surface integrity, but rather has a separate, somewhat more specific effect on endocytosis and very little effect on growth on salt.
The alternative possibility—that the actin-fimbrin interaction is required for endocytosis and defects in endocytosis affect the ability of cells to both grow on salt and sporulate—is equally plausible. Consistent with this view, we demonstrated that salt decreases the rate of endocytosis both in wild-type and actin mutant cells, suggesting that salt may exacerbate existing defects in endocytosis, perhaps to a level that is too low for viability. However, the finding that act1-4 mutant cells are not salt sensitive, despite having a severe defect in endocytosis, indicates that the relationship between the two phenotypes is not quite so simple. If a defect in endocytosis does affect the ability of cells to grow on salt, a possible explanation for act1-4 is that the mechanism by which endocytosis is inhibited by this particular mutation is not exacerbated by salt.
If the defect in endocytosis does lead to the defect in sporulation and growth on salt, then endocytosis must have a role in sporulation. In preliminary studies, we have demonstrated that endocytosis does occur during sporulation (Davis 1998); however, the role of endocytosis in sporulation remains completely unknown.
The correlation between endocytosis and salt sensitivity is seen in other organisms also: The association between hypersensitivity to NaCl and reduced endocytosis in actin cytoskeletal mutants is not unique to S. cerevisiae. In Dictyostelium discoideum, deletions in the genes encoding the actin-binding proteins α-actinin and gelation factor result in cells that are defective in phagocytosis and hypersensitive to osmotic stress (Riveroet al. 1996). Furthermore, the conventional myosin in Dictyostelium is required for the osmotic response, as myosin II mutants are hypersensitive to osmotic stress (Kuwayamaet al. 1996). In addition, myosin is phosphorylated and relocalized following osmotic stress. These studies indicate that the actin cytoskeleton plays a fundamental role under conditions of osmotic stress, and understanding the molecular mechanisms underlying the function of actin in yeast will more than likely apply to higher eukaryotes.
Acknowledgments
We thank Jerry Honts, Bruce Patterson, Mani Ramswami, Scott Selleck, and Ted Weinert for helpful discussions, and Greg Payne, David Drubin, Ken Wertman, and Susan Michaelis for strains, plasmids, and advice. We are grateful to the staff at the DNA sequencing facility at the University of Arizona for assistance with DNA sequencing. This work was supported by grants from the National Institutes of Health (GM45288) and the Council for Tobacco Research-U.S.A., Inc. (grant no. 4630) and by a Flinn Foundation Fellowship and Cancer Biology Training Grant to J.L.W., a Physiology Training Grant to D.A.D., and a National Science Foundation Predoctoral Minority Fellowship to K.A.T.
Footnotes
-
Communicating editor: F. Winston
- Received August 31, 2000.
- Accepted October 30, 2000.
- Copyright © 2001 by the Genetics Society of America
