|Home | About | Journals | Submit | Contact Us | Français|
Budding yeast grow asymmetrically by the polarized delivery of proteins and lipids to specific sites on the plasma membrane. This requires the coordinated polarization of the actin cytoskeleton and the secretory apparatus. We identified Rho3 on the basis of its genetic interactions with several late-acting secretory genes. Mutational analysis of the Rho3 effector domain reveals three distinct functions in cell polarity: regulation of actin polarity, transport of exocytic vesicles from the mother cell to the bud, and docking and fusion of vesicles with the plasma membrane. We provide evidence that the vesicle delivery function of Rho3 is mediated by the unconventional myosin Myo2 and that the docking and fusion function is mediated by the exocyst component Exo70. These data suggest that Rho3 acts as a key regulator of cell polarity and exocytosis, coordinating several distinct events for delivery of proteins to specific sites on the cell surface.
Delivery of newly synthesized proteins and lipids to specific sites on the cell surface is critical for the polarized growth observed in many eukaryotic cells. Budding yeast show a high degree of polarity and polarized growth during several stages of their cell cycle (Welch et al., 1994 ; Lew and Reed, 1995 ). They are polarized to the presumptive bud site in unbudded cells. Polarized growth occurs at the bud tip in small budded cells, then expands in larger buds, and finally is directed to the mother–daughter neck in cells about to undergo cytokinesis (Lew and Reed, 1995 ). Protein secretion is also highly polarized and corresponds to the pattern of bud growth (Tkacz and Lampen, 1973 ; Field and Schekman, 1980 ). This polarization of a yeast cell involves distinct rearrangements of the actin cytoskeleton and the transport of vesicles to the site of polarized growth (Novick and Botstein, 1985 ).
Genetic analysis in yeast has identified a number of gene products required for the terminal stages of exocytic trafficking to the cell surface in yeast. These include the unconventional type V myosin Myo2 (Johnston et al., 1991 ), the rab GTPase Sec4 (Salminen and Novick, 1987 ), a number of gene products that form a multisubunit complex known as the exocyst (TerBush et al., 1996 ; Finger et al., 1998 ; Guo et al., 1999 ), and the SNAP receptor (SNARE) proteins Snc1/2, Sso1/2, and Sec9 (Aalto et al., 1993 ; Protopopov et al., 1993 ; Brennwald et al., 1994 ). Myo2 is thought to be involved in the polarized delivery of vesicles from the mother cell to the bud (Govindan et al., 1995 ; Pruyne et al., 1998 ), whereas the rab, exocyst, and SNARE components all appear to function at the docking and fusion step (Novick and Brennwald, 1993 ).
The rho family of GTPases is thought to have a central role in the polarized growth process (Ridley, 1995 ; Drubin and Nelson, 1996 ). Yeast cells contain five rho family members: RHO1, RHO2, RHO3, RHO4, and CDC42. These genes, similar to their mammalian counterparts, play a variety of roles in cell polarity and polarized cell growth (Johnson, 1999 ; Kroschewski et al., 1999 ). RHO1 is essential and is thought to be involved in regulating the synthesis and integrity of the cell wall (Drgonova et al., 1996 ). RHO2 is not essential and has overlapping functions with RHO1 (Madaule et al., 1987 ). RHO3 and RHO4 may have some partially overlapping functions (Matsui and Toh-e, 1992a ,b ; Kagami et al., 1997 ) and are thought to play a role in bud formation (Imai et al., 1996 ). CDC42 and its guanine nucleotide exchange factor CDC24 are thought to be the most upstream components in the pathway for polarized bud formation (Pringle et al., 1995 ; Richman et al., 1999 ).
A previous report (Imai et al., 1996 ) suggested a possible connection between Rho3 and the exocytic apparatus by identifying SEC4 as a multicopy suppressor of a rho3Δ strain. However in the analysis of a temperature-sensitive allele of RHO3 reported in this same study, they were unable to detect any significant defect in the secretion of the exocytic marker invertase (Imai et al., 1996 ). Here we show a large number of additional genetic interactions between Rho3 and components of the exocytic apparatus, and we demonstrate that Rho3 has a direct role in exocytosis that is independent of its role in regulating actin polarity. Furthermore, analysis of a conditional mutant indicates that the role of Rho3 in exocytosis likely has two components: 1) a myosin (Myo2)-mediated function in the transport of post-Golgi vesicles from the mother cell to the bud and 2) an exocyst (Exo70)-mediated function in the docking and fusion of vesicles with the plasma membrane.
Cells were grown in YPD media containing 1% bacto-yeast extract, 2% bacto-peptone, and 2% glucose, all from Difco (Sparks, MD). For all assays performed, 25°C was the permissive temperature, whereas 14°C was used as the restrictive temperature. Room temperature (RT) was ~22°C.
Sorbitol, sodium azide (NaN3), N-ethylmaleimide (NEM), β-mercaptoethanol, o-dianasidine, glucose oxidase, peroxidase, Calcofluor, tetramethylrhodamine isothiocyanate (TRITC)-conjugated phalloidin, Triton X-100, and 3-amino-1,2,4-triazole were obtained from Sigma Chemical (St. Louis, MO). Zymolyase (100T) was from Seikagaku (Tokyo, Japan). BSA and yeast nitrogen base were from US Biologicals (Swampscott, MA). Formaldehyde (37%), gluteraldehyde, and Spurr’s resin were from Electron Microscopy Sciences (Ft. Washington, PA).
Transformations for suppression analysis were performed using the lithium acetate method described by Ito et al. (1983) . Strain crossing, tetrad dissection, diploid sporulation, and mating-type determination were performed as described by Guthrie and Fink (1991) .
Oligo-directed mutagenesis was performed to create the GTP- and GDP-bound forms of RHO3, as well as the pool of effector domain mutations (Kunkel et al., 1987 ). Fourteen synthetic oligonucleotides were used to change each of the 14 amino acids in the effector domain to other random amino acids. Each oligonucleotide had a mixture of bases at the position where different amino acids were desired and was designed to contain at least 10 bp of homology on both sides of the mixed residue. Resulting mutants were then sequenced to determine the amino acid sequence.
To determine the phenotype of each of the rho3 mutants as the only source of Rho3, we transformed the constructs into a diploid yeast strain heterozygous for a chromosomal deletion of RHO3. BY506 (a/α; GAL+/GAL+; ura3-52/ura3-52; leu2-3112/leu2-3112; his3-Δ200/his3-Δ200; rho3Δ::LEU2/RHO3) has one copy of RHO3 disrupted by the insertion of the LEU2 marker. The plasmids containing the rho3 mutant were cleaved within the URA3 gene to target chromosomal integration at the URA3 locus. The transformants were colony purified and sporulated, and tetrads were dissected with a micromanipulator on YPD plates. The plates were grown at 25°C. The haploid progeny were then analyzed by replica plating for the presence of the integrated mutants (scored as ura+), for the presence of a disrupted copy of rho3 (scored as leu+), and for conditional growth at 37 or 14°C.
Cells were grown in YPD to midlog phase, diluted back, and preshifted to the restrictive temperature for 1 h (or kept at the permissive temperature). From these cultures, two aliquots of 0.8 OD599 U were pelleted. One-half was resuspended in 10 mM NaN3 and kept on ice, while the other one-half was resuspended in low-glucose (0.1%) YPD and shifted to 14°C for 5 or 10 h or to 25°C for 1.5 h. After the shift, the fractions were resuspended in NaN3, and aliquots from both fractions were spheroplasted for 30 min at 37°C in a 1.4 M sorbitol and Tris-buffered solution containing β-mercaptoethanol and 0.1 mg/ml 100T zymolyase. Spheroplasted internal fractions were then resuspended in 0.5% Triton X-100.
All internal and external samples were assayed at 37°C in a sodium acetate and NEM buffer by adding 25 μl of 0.5 M sucrose and then stopping the reaction after 15 min with 150 μl of 0.2 M K2HPO4. The addition of an assay mix containing NEM, o-dianasidine, glucose oxidase, and peroxidase in a 0.1 M KPO4 buffer causes a color change in the presence of glucose. This end-point assay was fully stopped and developed by the addition of 6 N HCl. Spectrophotometric readings were taken at A540, and units of activity were calculated and reported as micromoles of glucose released per minute per OD599 of cells.
RHO3, rho3-V51, and sec4-P48 cells were grown overnight to midlog phase at 25°C in YPD. Strains were preshifted to the restrictive temperature of 14°C for 1 h before being transferred to YP media containing low (0.1%) glucose for 5 h. In addition, the late-secretory temperature-sensitive mutant sec18-1 was preshifted to 37°C for 1 h before a 3-h shift into low glucose. Whole-cell glass bead lysates were prepared and boiled with 2× sample buffer. Samples were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with affinity-purified α-invertase antibody at 1:200 or affinity-purified α-CPY antibody at 1:1000. For Bgl2 blots, cells were spheroplasted, and the internal and external fractions were separated. Internal fractions were boiled in 2× sample buffer, and the external fractions were boiled in 6× sample buffer. Samples were subjected to SDS-PAGE, transferred to nitrocellulose, and probed with affinity-purified α-Bgl2 antibody at 1:100.
For actin staining, cells were grown overnight in YPD media to ~0.5 OD599. They were then either shifted to the restrictive temperature of 14°C for 5 h or kept at the permissive temperature before fixative (formaldehyde to 3.7%) was added directly to the media. A second round of fixation was performed in KPO4 buffer (pH 6.5), and the cells were transferred to a sorbitol buffer for overnight storage at 4°C.
The next day cells were permeabilized for 10 min in 0.1% Triton X-100 and washed twice with PBS. Cells were resuspended in 100 μl of PBS and stained in the dark for 25 min with 35 μl of 3.3 μM TRITC-phalloidin dissolved in methanol. Cells were washed six times with PBS and then resuspended in the appropriate volume of mounting media (90% glycerol with 4′,6-diamidino-2-phenylindol to visualize DNA and o-phenylenediamine to retard photobleaching). To study the deposition of chitin in the bud scars, cells from the same round of fixation were not permeabilized, and 3 μl of 1 mg/ml Calcofluor was added instead of labeled phalloidin. Stained cells was viewed on a Nikon Eclipse E600 microscope (Garden City, NY); images were captured with a Princeton Instruments charge-coupled device camera (Trenton, NJ) and Metamorph Imaging software (Universal Imaging, West Chester, PA).
Cells were grown overnight in YPD media to midlog phase and diluted back the next day, and then one-half of the culture was shifted to the restrictive temperature 14°C and grown for 5 h to an OD599 of 0.8. Ten optical density units were harvested on a 115-ml, 0.45-μm Nalgene (Rochester, NY) filter, washed with 0.1 M cacodylate, and resuspended in a 0.1 M cacodylate and 3% gluteraldehyde fixative solution. Cells were then allowed to fix for 1 h at RT and then overnight at 4°C. Cells were spheroplasted for 40 min at 37°C in a KPO4-buffered solution of 0.3 mg/ml zymolyase and then washed with cold cacodylate. Then 1.5 OD U was pelleted in a 1.7-ml microfuge tube and incubated on ice for 1 h at RT with a solution of 2% osmium tetroxide and 0.1 M cacodylate. Pellets were washed three times each with dH2O and incubated in a 2% aqueous solution of uranyl acetate for 1 h at RT in the dark. Each pellet was rinsed twice with dH2O and taken through the following 10-min ethanol dehydration steps: 50, 70, 85, and 95% and then four rinses in 100% ethanol. The pellet was rinsed once with 100% acetone and then covered with a 50% acetone and 50% Spurr’s resin mixture. This remained at RT for 5 h, the acetone and Spurr’s resin mixture was replaced with 100% Spurr’s resin, and these remained at RT overnight. The 100% Spurr’s resin was changed the next day, and the pellets were baked in Spurr’s resin for 24 h at 80°C. Thin sections were cut and layered onto an uncoated copper grid and poststained with lead citrate and uranyl acetate before viewing. Cells were viewed on a JEOL 100CXII Electron microscope (Tokyo, Japan) and photographed at 60 kV.
The constructs were prepared by a recombinational cloning method (Hudson et al., 1997 ) in which the indicated Rho3 fragments were PCR amplified with 70 bp extensions on each end. These extensions then directed the in vivo incorporation of the fragment into a NcoI/PvuII-digested pOBD.CYH vector. Rho3 and activated Rho3-V25 were amplified by PCR, whereas the Rho3-V25,V51 double mutant was generated by fusion PCR, and all were confirmed by DNA sequence analysis. The C-terminal CAAX box was deleted by placing a termination codon in place of the cysteine residue. The GAL4AD constructs were made in a similar way and inserted into the activating domain vector pOAD by recombinational cloning. All constructs were tested to verify that they did not interact when paired with the opposing empty vector. Transformants were shown to express the expected size of the GAL4BD-RHO3 fusion by Western blotting and then transformed with the GAL4AD-MYO2 or GAL4AD-EXO70 constructs. Constructs were transformed into the strain PJ694α that has the HIS3 gene as a reporter (James et al., 1996 ). To reconfirm that similar amounts of Rho3 fusion proteins were present, immunoblotting of the two-hybrid transformants was performed on whole-cell lysates as described above. Blots were probed with affinity-purified Rho3 antibodies and detected with 125I-protein A.
We identified previously a number of genes, including the plasma membrane SNARE SEC9, on the basis of their ability to act as suppressors of the cold-sensitive Rab GTPase mutant sec4-P48 (Brennwald et al., 1994 ). Although SEC9 was the most potent multicopy suppressor, we describe here the analysis of a second suppressor, HSS43, whose strength of suppression was nearly as strong as that of SEC9. Analysis of subclones of HSS43 (Figure (Figure1A)1A) identified the 2.2-kb SalI-XhoI fragment that contained the entire open reading frame of RHO3 as necessary and sufficient to achieve the suppression attributed to HSS43 (Figure (Figure1B).1B). To determine the specificity of Rho3 in this pathway, we looked at the suppression capabilities of all five of the yeast RHO family members. The results shown in Figure Figure2A2A demonstrate clearly that this property is highly specific to Rho3 because none of the other RHO genes, RHO1, RHO2, RHO4, or CDC42, were able to interact genetically with Sec4 in the same way that Rho3 did.
Rho3 is a member of the Ras superfamily of small GTPases that function as molecular switches (Bourne et al., 1991 ). By analogy to the ras oncogene, mutations were made in Rho3 that predict a constitutively active (GTP-bound) or a constitutively inactive (GDP-bound) form. We made two mutations, RHO3-V25, equivalent to the activating ras-V12 mutant, and RHO3-N30, analogous to the dominant-inhibitory ras-N17 mutant. We would predict that if the GTP-bound form is the active form in this pathway, there would be suppression at much lower doses of RHO3-V25 compared with wild type and that the GDP-bound form would fail to suppress. We therefore examined the effect of each of these on single copy (CEN) plasmids and tested their ability to suppress the sec4-P48 mutant (Figure (Figure2B).2B). Interestingly, a single extra copy of wild-type RHO3 results in only partial suppression of the cold sensitivity. However activated RHO3-V25 on a single copy shows suppression equivalent to that observed with wild-type RHO3 on high copy. In contrast, the RHO3-N30 allele, predicted to be in the GDP-bound form, not only fails to suppress but is growth inhibitory.
In addition to testing the ability of RHO3 to suppress sec4-P48, we examined the ability of the original clone HSS43 and the RHO3 subclone to suppress all of the 10 late-acting sec genes. We found that HSS43 and RHO3 show a significant level of suppression of both sec15-1 and sec8-9 (Brennwald et al., 1994 ) (our unpublished results). Both of these genes are components of the exocyst, a peripherally associated plasma membrane complex thought to aid docking and fusion of vesicles (TerBush and Novick, 1995 ).
In addition, a number of late-acting secretory genes have been found to suppress rho3 mutants. Both SEC9 and SEC4 act as high-copy suppressors of rho3 (Imai et al., 1996 ; Lehman et al., 1999 ). SRO7 and SRO77 were first identified as high-copy suppressors of a rho3 mutant (Kagami et al., 1998 ). We have shown recently that Sro7 and Sro77 are Sec9-binding proteins and that loss of these gene products results in a block in Golgi-to-cell-surface transport (Lehman et al., 1999 ).
To look for additional connections between rho3 and the exocytic apparatus, we used a galactose turn-off assay to identify genes that could rescue the extremely slow growth phenotype of rho3Δ cells. Cells with the sole copy of RHO3 under the control of a Gal promoter were transformed with a genomic library prepared in a 2μ vector. Transformants were selected on glucose-containing media to repress expression of Rho3. Plasmids were isolated from colonies showing growth on glucose, retested for suppression, and sequenced. This resulted in the identification of four distinct suppressing loci (in addition to RHO3). Two of these contained genes shown previously to suppress rho3 mutants, SEC4 and BEM1, a gene shown to be involved in bud formation and cell polarity. Three independent clones contained overlapping segments of the t-SNARE SSO2; subcloning revealed this gene to be necessary and sufficient for the suppressing activity. The fourth gene is still being characterized and will be described elsewhere. In summary, four distinct components of the exocytic machinery, Sec4, Sec9, Sro7, and Sso2, can act as suppressors of rho3Δ mutants. This demonstrates that loss of Rho3 function is most readily assisted by increasing the function of the exocytic machinery. These numerous genetic interactions strongly suggest that Rho3 plays a critical role in exocytosis.
We hypothesized that the above genetic evidence suggests an additional critical role for Rho3 in exocytosis, independent of its role in actin polarity maintenance. This is consistent with the rapidly growing body of evidence that Rho GTPases often have a large number of effectors that can cover a diverse range of functions (Tapon and Hall, 1997 ; Aspenström, 1999 ). In an attempt to reveal a direct role of Rho3 in exocytosis and to distinguish this role from the role of RHO3 in actin polarity, we initiated an extensive mutagenesis of the effector domain. This highly conserved region has been shown to be critical for interactions between ras GTPase superfamily members and their downstream effectors (Bourne et al., 1991 ).
The region of Rho3 known as the effector domain can be aligned with the same conserved domains from other proteins such as H-ras and Cdc42 (Figure (Figure3A).3A). The results of the oligo-directed mutagenesis are listed in Figure Figure3B.3B. Mutant alleles were introduced into yeast and tested for their ability to function as the only copy of RHO3. Of the 46 mutants examined, all were recessive, and 30 were indistinguishable from wild type under the conditions tested. Eleven mutants failed to rescue the severe growth defect associated with the rho3Δ deletion, and 1 mutant rescued slightly but gave constitutively slow growth under all conditions tested. Four mutants showed relatively normal growth at 25°C but showed significantly reduced growth at 14°C (Figure (Figure3,3, these last four are shown in bold type). Figure Figure3C3C shows the growth of four cold-sensitive mutants at 14°C as compared with wild type and compared with their growth at the permissive temperature of 25°C. None of the mutants showed sensitivity to growth at high temperature (37°C) in this study. Bright-field analysis of all four of the mutants revealed an increase in cell size, a phenotype often associated with defects in cell polarity (Novick and Botstein, 1985 ). But this enlargement was not dependent on the temperature shift in any of the cells, and the most dramatic size increase was seen in the mutant rho3-C40 (our unpublished results).
Loss of Rho3 function has been shown previously to cause an altered distribution in the normally highly polarized actin cytoskeleton (Imai et al., 1996 ). We therefore examined the status of the actin cytoskeleton in the four cold-sensitive rho3 mutants by rhodamine phalloidin staining. Both actin filaments and cortical actin patches show a very high degree of polarity in yeast (Welch et al., 1994 ; Lew and Reed, 1995 ; Karpova et al., 1998 ). In wild-type cells, cortical actin is highly polarized and found almost exclusively in the bud (Welch et al., 1994 ; Lew and Reed, 1995 ). Only as the cell nears cytokinesis do the cortical patches become significantly apparent in the mother cell. Actin cables are normally oriented along the mother–bud axis.
Cells were examined after growth at 25°C or after a 5-h shift to 14°C. As expected, several of the effector mutants demonstrated pronounced defects in actin polarity (Figure (Figure4).4). Three of the mutants show a loss of the normal actin patterns. In particular the most severe defects were seen in rho3-C40 and rho3-G46. They showed almost complete loss of cables and random distribution of cortical patches. rho3-G49 showed a partial loss of cables and a disruption of cortical patch polarization. In all three of these mutants, the actin polarity defects appeared to be similar at both temperatures.
In contrast, the rho3-V51 mutant had a normal and polarized actin cytoskeleton that was indistinguishable from the polarized actin in wild-type cells at both 25 and 14°C. As with wild-type cells, the cortical actin patches were well polarized to the bud. The cables were clearly present, well organized, and of normal length relative to cell size (i.e., because rho3-V51 cells are larger, the cables are proportionally longer). We also examined diploids homozygous for the rho3-V51 mutation and found they also showed staining patterns identical to that of wild-type diploids at both the permissive (our unpublished results) and restrictive temperatures (Figure (Figure44).
The proper formation of chitin scars at bud sites is dependent on the polarization of actin (Chant and Pringle, 1991 ). We therefore looked for bud site selection defects in a subset of these rho3Δ mutants by staining chitin with the fluorescent dye Calcofluor. The results demonstrate that rho3-V51 exhibited normal budding patterns in both haploids and homozygous diploids, and in contrast, the rho3-C40 mutant and, to a somewhat lesser extent, the other mutants showed aberrant bud site selection and delocalized chitin deposition (our unpublished results). This provides additional evidence that out of the four cold-sensitive effector mutants, only rho3-V51 is fully functional in regulation of the actin cytoskeleton.
To determine whether the actin-independent growth defect associated with the rho3-V51 mutant was caused by a defect in the exocytic function of these cells as predicted by the abundance of genetic data described above, we examined the ability of these strains to export two periplasmic proteins, invertase and Bgl2. Both of these proteins are dependent on the classical secretory pathway for their delivery to the cell surface (Kaiser et al., 1996 ). RHO3, rho3-V51, and sec4-P48 strains were shifted to low glucose to derepress the production of invertase at either 25 or 14°C. The amount of internal and external enzyme activity was determined in five separate experiments, and the amount of invertase in the internal fraction is shown in Figure Figure5A.5A. As expected, the sec4-P48 mutant shows a defect at 14°C but secretes well at 25°C. In agreement with the prediction that Rho3 has a direct role in exocytosis, the rho3-V51 cells display a pronounced defect in invertase secretion similar to that of sec4-P48. Although wild-type cells were able to secrete 92% of invertase and retained only 8% in the internal fraction at 14°C, the sec4-P48 mutant retained 33%, and rho3-V51 retained 27% in the internal fraction at the restrictive temperature. Surprisingly, when tested at 25°C, the rho3-V51 mutant showed a significant defect in exocytosis at the permissive temperature (21% internal invertase for rho3-V51 compared with only 5% for wild-type cells).
We also examined the ability of the other mutants to secrete invertase after a 5-h shift to 14°C. They were found to be significantly less defective in secretion: 17 ± 9.8% (rho3-C40), 10 ± 6.5% (rho3-G46), and 16 ± 9.6% (rho3-G49). However, because defects in actin have been shown previously to cause defects in secretion (Novick and Botstein, 1985 ), it is unclear whether these partial defects are associated with a direct effect on the secretory capacity versus an indirect effect resulting from a depolarized actin cytoskeleton.
Immunoblot analysis with α-invertase antibodies showed that virtually all of the invertase in the cells migrated as the fully glycosylated form, demonstrating that the defect is likely to be at the post-Golgi stage of secretion (Figure (Figure5C).5C). Consistent with a post-Golgi defect, maturation and transport of CPY appear to be unaffected. Immunoblots of these cells show that only the mature form is present in rho3-V51 as in wild-type and sec4-P48 (Figure (Figure5D).5D). In addition, pulse-chase analysis of CPY maturation demonstrated that there was no defect in the kinetics of transport of CPY to the vacuole (our unpublished results).
Two classes of vesicles have been demonstrated to be involved in post-Golgi delivery to the plasma membrane (Harsay and Bretscher, 1995 ). Invertase is specific to the lighter class of vesicles, and Bgl2, a cell wall endoglucanase, has been shown to be specific to the denser class of vesicles. To determine whether both classes of vesicles were affected, we examined the secretion of Bgl2 in the rho3-V51 and sec4-P48 cells. Immunoblots were used to probe the internal and external fractions at both the permissive and restrictive temperatures. Figure Figure5B5B shows the results of Bgl2 secretion assays that are expressed as the percentage of Bgl2 found internally. In agreement with the results seen for the secretion of invertase above, rho3-V51 shows a pronounced defect in the secretion of Bgl2 at both 25 and 14°C. Although wild-type RHO3 cells show small internal pools of only 13 and 19% at 25 and 14°C, the rho3-V51 cells show much larger internal pools of 38 and 40%, respectively. The quantitation is the mean ± SD of three assays, and a representative blot is shown (Figure (Figure5B,5B, bottom). The identity of the bottom band seen in all the external fractions is unknown and was not included in the quantitation. This defect in Bgl2 secretion indicates that the rho3-V51 mutant is affecting both classes of vesicles. Interestingly, the sec4-P48 mutant appears to be less temperature dependent in its secretion of Bgl2 than invertase because it shows defects at both temperatures.
To examine the exocytic defect in greater detail, we examined rho3-V51 mutant cells by electron microscopy. All of the post-Golgi–blocked secretory mutants show an accumulation of 80- to 100-nm vesicles. These vesicles are only rarely seen in wild-type cells because their rate of fusion with the bud tip plasma membrane is very rapid. Because the rho3-V51 mutants had secretory defects at both ambient and low temperatures, we examined the morphology of cells grown with and without a 5-h shift to 14°C. Consistent with the secretory defect seen at 25°C, the rho3-V51 mutant shows a large accumulation of 80- to 100-nm vesicles. The vesicles appear to be present primarily in the bud and neck region of the cells (Figure (Figure6A,6A, left). In contrast, although rho3-V51 cells grown at 14°C had a slight increase in the total number of vesicles per cell, a dramatic shift was observed in the location of the vesicles within the cell. Although mother cells with large numbers of vesicles were rare at 25°C, an abundant accumulation of vesicles in the mother cells was observed after a 5-h shift to 14°C (Figure (Figure6A,6A, middle). Wild-type RHO3 cells are also shown in Figure Figure6A,6A, right. The quantitation of the vesicle accumulation phenotype is shown as a bar graph (Figure (Figure6B).6B). At 14°C, there is a dramatic shift in the distribution of vesicles within the cell; 60% of the mother cells contained >60 vesicles, whereas <20% of the cells at the permissive temperature fall into this category. This demonstrates that the rho3-V51 mutant has a cold-sensitive defect in the transport of vesicles from the mother cell into the bud.
A recent study identified fragments of Exo70- and Myo2-coding sequences in a two-hybrid screen for potential effectors of Rho3 (Robinson et al., 1999 ). To determine whether the exocytic defects associated with rho3-V51 might be a result of a defect in interaction with these candidate effectors, we examined the effect of the rho3-V51 mutation on this interaction. We made use of the same regions of MYO2 and EXO70 identified by Robinson et al. (1999) , fused to the activation domain of GAL4, whereas the RHO3 alleles were fused to the DNA-binding domain of GAL4. In this assay, we made use of the HIS3 gene as a reporter of activation by the two-hybrid constructs. Activation of this reporter results in growth in the absence of histidine. The strength of interaction can be monitored in part by the sensitivity of growth to the presence of 3-aminotriazole, an inhibitor of the histidine biosynthetic pathway. As can be seen in Figure Figure7A,7A, a portion of the MYO2 neck and tail region (amino acids 871-1204) interacts specifically with the activated allele of RHO3, RHO3-V25. However, the introduction of the V51 effector mutation into RHO3-V25 completely abolishes the interaction with MYO2. Like MYO2, the C-terminal portion of EXO70 shows both a strong interaction with the activated RHO3-V25 as well as a weaker interaction with the wild-type form of RHO3. The strength of the interaction between activated RHO3-V25 and EXO70 is apparent because growth occurs when 5 mM 3-aminotriazole is present and persisted even in the presence of 40 mM 3-aminotriazole (Figure (Figure7A;7A; our unpublished results). As was the case with the MYO2 interaction the presence of the V51 mutation in the activated form of RHO3 leads to a complete loss of growth even on −His media lacking 3-aminotriazole (Figure (Figure7A).7A). Therefore, the rho3-V51 mutation leads to a loss of the ability of Rho3 to interact with both Myo2 and Exo70. This is consistent with the vesicle delivery defects we observe at 14°C being attributable to a loss of interaction with Myo2 and with the docking and fusion defects observed at 25°C being attributable to defects in interaction with the Exo70 component of the exocyst complex (TerBush et al., 1996 ).
Analysis of an extensive collection of mutants in the effector domain of Rho3 allowed us to separate the exocytic function of Rho3 from its role in regulating actin localization. Our results demonstrate that Rho3 acts as a key regulator of cell polarity by coordinating two important cellular processes: 1) maintenance and localization of the actin cytoskeleton and 2) the polarized delivery, docking, and fusion of post-Golgi secretory vesicles with the plasma membrane. Genetic evidence gave us the first clues concerning a possible role for Rho3 in exocytosis, but it was the mutagenesis of the effector domain in particular that allowed us to discriminate clearly its role in exocytosis from its role in actin polarity. Of the four loss-of-function mutations that we studied, three showed clear actin polarity defects, whereas the fourth showed no defects in actin polarity but had an exocytic defect. It was especially important to distinguish between these two possibilities because actin mutants are known to cause defects in post-Golgi trafficking (Novick and Botstein, 1985 ).
Our results suggest that Rho3 regulates cell polarity by simultaneously directing the rearrangements of the actin cytoskeleton and the polarized delivery and fusion of the vesicles to specific sites on the cell surface. This is clearly important because these are both dynamic processes in yeast cells. Although the bud tip is the major destination of polarized growth early in the cell cycle, growth becomes isotropic later in the cell cycle. The site of actin polarity and delivery reorients itself to the mother–bud neck, just before cytokinesis. Rho3 could be acting to maintain the proper synchronized actions of actin and vesicle transport; both in specifying a new site and in maintaining delivery to the current site.
In this view, it is especially interesting to find evidence that Rho3 functions in two separate steps of exocytosis: the delivery of vesicles and the docking and fusion of the vesicles. Rho3 shows a number of strong genetic interactions with gene products involved in this second step of the docking and fusion of secretory vesicles at the plasma membrane. In particular, rho3Δ was suppressed by the two t-SNAREs Sec9 and Sso2, as well as by the Sec9-binding protein Sro7 (Kagami et al., 1998 ; Lehman et al., 1999 ). In addition, there are strong interactions with two subunits of the exocyst complex and the Rab GTPase Sec4. Recent data have suggested that the Rab GTPase and exocyst complex function as tethering factors and work just upstream of SNAREs in driving the vesicles to initiate the fusion reaction (Pfeffer, 1999 ). Presumably the genetic interactions between these components reflect the role of Rho3 in the docking and fusion step.
With respect to the vesicle delivery step, a conditional defect was apparent when electron microscopy revealed that a shift to 14°C resulted in a very clear accumulation of vesicles in the mother cell. Because vesicles are thought to be primarily produced in the mother cell, they must be delivered to their site of fusion in the bud, so an accumulation of vesicles in the mother cell is a result of a defect in this vectorial delivery process. This cold-sensitive delivery phenotype of Rho3 is strikingly similar to the phenotype of an unconventional type V myosin Myo2 mutant (myo2-66) that shows a conditional accumulation of vesicles in the mother cell at the nonpermissive temperature (Govindan et al., 1995 ).
Recently, Robinson et al. (1999) isolated two candidate effectors of Rho3 using a two-hybrid screen. This resulted in the identification of Myo2 and Exo70 as proteins that specifically interact with the GTP-bound form of Rho3. Our analysis of the rho3-V51 mutant gives strong evidence that Myo2 and Exo70 are likely to be bona fide effectors of Rho3 function in exocytosis. In particular the loss of binding interaction, as measured by two-hybrid analysis, between rho3-V51 and both MYO2 and EXO70 strongly supports the idea that it is a defect in interaction with these two factors that is responsible for the exocytic defects we observe in this mutant. In addition the constitutive exocytic defect of the rho3 mutants may be caused by a defect in the activation of the exocyst complex via Exo70. The electron microscopy of the mutant under permissive conditions supports this idea. The accumulation of vesicles in the bud and not in the mother in rho3-V51 cells at 25°C suggests a defect at the docking and fusion stage. This is similar to the phenotype observed in temperature-sensitive mutations in components of the exocyst complex after a brief shift to 37°C (Govindan et al., 1995 ).
Taken together, these data suggest a model, shown in Figure Figure8,8, in which Rho3 acts as a key regulator of polarized trafficking by coordinating three distinct effector pathways. One pathway involves the regulation of the actin cytoskeleton, whereas the other two pathways involve exocytic transport to the cell surface. This includes the regulation of a Myo2-dependent delivery of vesicles from the mother cell to the bud (Figure (Figure8,8, Step 1) and regulation of an Exo70-dependent docking and fusion of vesicles at specific sites on the plasma membrane (Step 2). These data give us an outline of the function of Rho3 in exocytosis and cell polarity and place it as a key step in the coordinate regulation of these two essential processes. Future work will no doubt unravel the precise mechanisms by which Rho3 effects these functions.
The authors are especially grateful to Dr. Geri Kreitzer and Dr. Enrique Rodriguez-Boulan for allowing us use of their microscope and their expertise and assistance with the fluorescence microscopy and image analysis. We also thank Henry Hamilton for generation of the RHO3-V25 and RHO3-N30 mutants, Linda Burg-Freedman and Lee Cohen-Gould for assistance with electron microscopy, Dr. Doug Johnson and Dr. Mike Snyder for plasmids, and Dr. Ruth Collins, Dr. Tim McGraw, Luba Katz, Mayya Maksimova, and Greg St. John for critical reading of the manuscript. This work was supported by grants from the Mathers Charitable Foundation and the Pew Scholars in Biomedical Sciences Program and by the National Institutes of Health grant GM-54712.