Cell. Vol. 65. 1099-1096, June 28, t991, Copyright 0 1991 by Cell Press
Development of Cell Polarity in Budding Yeast David 0. Drubin Department of Molecular and Ceil Biology University of California, Berkeley Berkeley, California 94720
For a wide variety of cellular processes, many constituents of the cell, such as plasma membrane proteins, organelies, and cytoskeletai filaments, must be organized asymmetrically within the cell. Cell polarity is required for the transmission of a nerve impulse, the development of a fertilized egg into a multicellular organism, the transport of molecules across an epitheiial ceil layer, the crawling of a fibrobiast cell, and the growth of a budding yeast ceil. Two central challenges in studies on the development of cell polarity are the identification of interdependencies between the many different events involved and the determination of the primary inductive events. Fundamental questions, such as how secretory vesicles are delivered to a particular membrane and how a spindle orients correctly to define a division axis, have not been answered. The yeast ceil, while lacking some interesting behaviors of higher cells, possesses an experimental advantage in its simplicity. Furthermore, yeast present an opportunity to genetically dissect the molecular pathways that lead to the development of cell polarity, much as was done for phage assembly pathways. In many cells, the development of polarity appears to begin at the cortex. For example, the site of fertilization on a frog egg determines the dorsal-ventral axis for embryonic development (Gerhart and Keller, 1986). An axis of ceil polarity can be stably defined on the cortex by the assembly of a nondiffusible axis marker. In plants, the cell wall plays a role in this process (Kropf et al., 1988). The asymmetry marked on the ceil cortex must be spread to the cytoplasmic compartment. The cytoskeleton is well suited to this function: the growth and stability of actin filaments and microtubules can be controlled by interac-
Minireview
tions of their fast-growing “plus” ends with particular regions of the cell cortex (Kirschner and Mitchison, 1986) and the inherent polarity of cytoskeietai filaments can spread polar organization from the cortex to the interior of the cell through interactions between motor proteins that bind to organeiies and polar arrays of cytoskeletal filaments. The Phenomenology of Polarized Yeast Cell Growth and Divlslon S. cerevisiae cells grow asymmetrically by localized vesicle fusion and cell wall synthesis, which result in the formation of buds or mating projections (Figures 1 and 2; Sloat et al., 1981; Pringie et al., 1986, and references therein). The asymmetry of yeast growth is also reflected in the arrangement of the cytoskeleton (Barnes et al., 1990) and many organeiies (Baba et al., 1989). In other cell types, polarized membrane traffic is thought to be important for such processes as ceil migration and extracellular matrix secretion. Yeast ceils choose bud sites in a nonrandom pattern determined by their mating type (Chant and Herskowitz, 1991, and references therein). a and a haploid ceils (as well as a/a and a/a diploid ceils, which mate as a and a ceils, respectively) bud from sites near the site of the previous division (axial budding, Figure 2). a/a diploid ceils bud in a bipolar manner: mother ceils bud from a site either near the site used for the previous division or near the opposite pole, and daughters usually bud from the pole opposite to their birth pole (Figure 2). During mating, a and a ceils ignore genetically determined axis formation instructions and orient toward a mating partner (see Cross et al., 1988). A yeast ceil’s ability to choose a mating partner displays the same high degree of ceil-type selectivity that is required of a cytotoxic T ceil when it uses its secretory cargo to kill a target ceil (Jackson and Hartwell, 1990; Kupfer and Singer, 1989). Polarized secretion is also exploited by mating yeast cells to deliver proteins specifically required for mating (such as Fuslp and agglutinins) to the projection tip (Cross et al., 1988). This review will focus on the budding pathway, but the
MATING BUD FORMATION
PROJECTION FORMATION
Figure 1. Cell Polarity in 9 cerevieiae Chitin is located extracellularly, while the products of C/X3, CDC70, CDCll. and CD02 are located intracellularly.
AXIAL BUDDING (amdo cells)
BIPOLAR BUDDING (ah%WIIS)
Figure 2. Orientation of the Yeast Growth Axis during Mating and Bud Formation Although the shapes of zygotes suggest that cells of opposite mating type can grow toward each other, this has not been documented.
C0ll 1094
CONTROL OF BUDDING PATTERN
BUD SITE ASSEMBLY
CDC24, CDC42, CDC43, BEMl, .9Ehf2, (MSBI, MB.?)
POLAR GROWTH
Figure 3. Morphogenetic
Pathway for Bud Formation
similarity of the processes of bud formation and projection formation should be noted (Figure 1; Cross et al., 1988). Molecular Genetic Studies Based on phenotypes, genetic interactions, and the interdependencies and timing of events, the genes required for cell polarity development in a mitotically growing S. cerevisiae cell can be arranged in hierarchical groupings (Figure 3). In this scheme, the budding pattern genes are required for correctly choosing a bud site. The bud site assembly genes are required to assemble components at the chosen bud site so that surface growth will be restricted to the bud. Microtubules, actin filaments, and neck filaments all appear to become organized through interactions with structures at the bud site (Figure 1). Microtu-
BUD SITE ASSEMBLY
BUDDING PAlTERN .._....
bules are required for moving the nucleus to the bud neck and orienting the spindle (see Barnes et al., 1990, for a review of the yeast cytoskeleton), while actin filaments and neck filaments are required for organizing surface growth and for cytokinesis. The BUD7 gene was originally named RSR7 and was identified by its ability, when overexpressed, to suppress a mutation in CDC24 (Bender and Pringle, 1989). Subsequently, BUD7 and four additional BUD genes were identified in a visual screen for mutants with altered budding patterns (Chant and Herskowitz, 1991; Chant et al., 1991). Mutations in BUDI, BUDP, or BUDS cause random budding in cells, irrespective of mating type. Mutations in BUD3 and BUD4 cause bipolar budding in a and a cells, which normally exhibit axial budding; however, these mutations have no effect on bipolar budding in a/a cells. Based on these observations, Chant and Herskowitz (1991) propose that in a and a cells budding would be axial, because all BUD genes are active. In a/a cells, on the other hand, the BUD3 and/or BUD4 gene might be inactive, resulting in bipolar budd!ng. Chant and Herskowitz (1991) point out that the al-a2 repressor present in a/a cells might inhibit either or both of these genes. When a yeast cell commits to a mitotic cell cycle in early Gl , several constituents appear on a specific region of the cell cortex that marks the site from which a bud will emerge, typically about 15 min later. These constituents initially include actin, 10 nm neck filaments, and the product of the gene SPA2 and subsequently chitin (Kim et al., 1991; Ford and Pringle, 1991; Snyder et al., 1991). Mutations in the bud site assembly genes prevent the assembly of these constituents at a bud site. As a result, the cells grow isotropically and become multinucleate because nuclei continue to divide in the absence of bud formation (Adams et al., 1990). Refinement of the scheme presented in Figure 3 will require determination of the location of each gene product in the cell, the biochemical activities and interactions of each protein, and when in the cell cycle the function of
ADDITIONAL INTERACTING GENES
-w
EUDI/RSRI(GTP binding) --.----.BUD2 BUD3 BUD4
BUD5 (GDP-GTP exchange)
)
4
.--....--.) 4------+
Synthetic Lethality Suppression by Overproduction Potential Interectlon Bawd on Sequence Homologlee
Figure 4. Genetic Interactions among Cell Polarity Genes Sequence homologies and the interactions they suggest are also indicated.
yin&view
each protein is required. However, numerous genetic interactions, suggestive sequence homologies, biochemical activities, and experiments that address the interdependence of gene functions all provide compelling evidence that these genes function in a morphogenetic pathway. Two genetic links suggest interactions between the budding pattern genes and the bud site assembly genes (Figure 4). First, overexpression of BUD7 can suppress some mutations in CDC24 (Bender and Pringle, 1989). Second, mutations in BUD5 and a gene involved in bud emergence (BEM7) show a negative synergism, resulting in so-called synthetic lethality in the double mutant (Chant et al., 1991). BEM7, in turn, is linked to CDC24 because both display genetic interactions with the same gene (MSBl; Bender and Pringle, 1991). This chain of interactions ultimately links BUD7 to BUDS. Additional interactions shown in Figure 4 and the similarity of the phenotypes observed for the mutants c&24, cdc42, c&43, beml, and bem2 provide further evidence that the bud site assembly genes function together (Bender and Pringle, 1989, 1991; Adams et al., 1990). Mutations in CDC24 and overproduction of CDC42 can lead to random budding patterns (Sloat et al., 1981; Johnson and Pringle, 1990). Other perturbations (see below) can lead to mild alterations in budding pattern. These observations provide further evidence for interactions between the bud site assembly and budding pattern genes but also blur the distinction between the two classes. However, separate groupings can be justified (Chant et al., 1991) because BUD genes appear to be required for budding pattern but not for bud formation. The bud site assembly genes, on the other hand, are essential for bud formation. Mutations in bud site assembly genes that affect budding pattern are proposed to perturb interactions between proteins encoded by bud site assembly and budding pattern genes. Why yeast should have genes whose sole function is to control budding pattern is not understood, since random budding causes no apparent disadvantage. The observation that budding pattern is perturbed by mild defects in bud site assembly genes suggests a second possibilitythe BUD genes may encode proteins with activities that are required for bud site assembly but are functionally redundant, so that only nonlethal defects are caused by gene disruptions. This interpretation would fit with the observation that synthetic interactions (e.g., between BUD5 and BEMI) can result when proteins function in a common step of a pathway (Salminen and Novick, 1987). Redundant proteins might be identified by the same sort of genetic screen that was used to identify BEM7 (Bender and Pringle, 1991). Some proteins required for the development of cell polarity have sequence homology with proteins having established biochemical activities, which suggests other activities and interactions (Figure 4). For example, the sequence of CDC24 (also known as CLS4) and the calcium sensitivity of certain cdc24 mutants indicate that the activity of Cdc24p might be modulated by calcium (Miyamoto et al., 1987). Interestingly, calcium has also been implicated in the function of Cdc43p (also known as Call p) and
is suggested to play a role in polarity development for certain cell types (Ohya et al., 1991; Schnepf, 1988). Perhaps most suggestive are the sequence homologies discovered for the proteins encoded by CDC42, CDC43, BUDl, and BUD5 The CDC42 and BUDlIRSRl gene products are homologous to small GTP-binding proteins in the ras family (Johnson and Pringle, 1990; Bender and Pringle, 1989). Additional support for the conclusion that small GTP-binding proteins play a role in cell polarity development comes from the discovery that Bud5p is homologous to the CDC25-encoded GDP-GTP exchange protein (Chant and Herskowitz, 1991). Furthermore, Cdc43p is required for S. cerevisiae geranylgeranyltransferase activity (Finegold et al., 1991) and is homologous to the protein encoded by DPRllFlAMl that prenylates the C-termini of small GTP-binding proteins (Ohya et al., 1991). These homologies and activities indicate that Cdc43p and Bud5p might each interact with Budlp and/or Cdc42p (both of which have appropriate C-terminal consensus sequences for modification by Cdc43p). Prenylation of Budlp and/or Cdc42p might be important for assembly of bud sites at the plasma membrane. At present there are no data that address the mechanism by which the two small GTPbinding proteins function in cellular morphogenesis. They might direct or monitor an assembly process, or they might regulate the timing of events. While many interactions have been observed between bud site selection and assembly genes, no such interactions have been observed between bud site assembly genes and cytoskeleton protein genes. The assembly of each filament system seems largely independent of assembly of the other two systems, but proper assembly of all three is dependent on the action of bud site assembly genes (Figure 3; Snyder et al., 1991; Adams et al., 1990; Ford and Pringle, 1991). Abplp is one candidate for a link between bud site assembly and assembly of the actin cytoskeleton; overproduction of this protein results in an aberrant budding pattern (Drubin et al., 1990). Recent studies suggest that microtubules emanating from the spindle pole body are captured in early Gl by a structure (located proximal to a concentration of Spa2p) at the incipient bud site (Snyder et al., 1991). The nucleus then rotates so that the spindle pole body is proximal to the bud site (Snyder et al., 1991). Although SPA2 is placed along the microtubule pathway in Figure 3, it seems not to function exclusively in conjunction with microtubules. Spa2p is required for mating projection formation (Gehrung and Snyder, 1990), but microtubules are not (Hasek et al., 1987). Bud Site Selection The fact that there is aspatial pattern to bud site selection suggests that a structure that marks the previous bud site must persist from one cell cycle to another (Snyder et al., 1991; Chant and Herskowitz, 1991). Axial budding might require that the BUD3 and BUD4 gene products interact with this structure and direct the assembly of a new bud site nearby. During C. elegans development, the midbody from the previous cell division, or a proximal structure, has been proposed to constitute a microtubule capture site that can orient the spindle for the next cell division (Hyman, 1989).
Cell 1096
In the absence of Bud3p or Bud4p, a bud site might form at the opposite pole from the last division because that is where the spindle pole body points after chromosome separation. This would be the default budding pattern. However, this scenario has the following problems. First, it does not explain how random budding is generated. Second, it is not consistent with the finding that microtubules do not appear to be important for axial or bipolar budding (Jacobs et al., 1988) but appear to be captured at a bud site after its assembly is initiated (Snyder et al., 1991). Finally, it does not account for the observation that, during bipolar budding, the mother cell often buds near the previous bud site, requiring a 1 80° rotation of the spindle pole body with respect to the cell surface. The fact that mutations can cause random budding suggests that the default budding pattern is random and that bipolar budding requires a special mechanism, the nature of which is obscure. Summary and Conclusions The development of cell polarity involves virtually every aspect of cell biology. Yeast are less complex than cells traditionally used for studies on cell polarity and are amenable to sophisticated genetic analysis. This has resulted in a growing number of molecular markers for yeast cell polarity and an increasingly well-defined progression of molecular events required for bud formation. Together, these factors provide a favorable context in which to understand how the interplay between a large number of processes can polarize a cell. Many genes required for morphogenesis have been identified, and genetic interactions provide evidence that the products of these genes function together. Studies on cell polarity development in S. cerevisiae have demonstrated a requirement for small GTP-binding proteins and have established functional relationships between temporally coincident events. With the continued identification and analysis of genes required for morphogenesis, and the pursuit of these studies on a cytological and biochemical level, studies on yeast will continue to contribute to our understanding of cell polarity development. References Adams, A. E., Johnson, D. I., Longnecker, R. M., Sloat, B. F., and Pringle, J. R. (1990). J. Cell Biol. 117, 131-142. Baba, M., Baba, N., Ohsumi, Y., Kanaya, K., and Osumi, M. (1989). J. Cell Sci. 94, 207-216. Barnes, G., Drubin, D. G., and Stearns, T. (1990). Curr. Opin. Cell Biol. 2,109-115. Bender, A., and Pringle, J. R. (1989). Proc. Natl. Acad. Sci. USA 86, 9976-9980. Bender, A., and Pringle, J. Ft. (1991). Mol. Cell. Biol. II, 1295-1305. Chant, J., and Herskowitz,
I. (1991). Cell 65, this issue.
Chant, J., Corrado, K., Pringle, J. R., and Herskowitz, 65, this issue.
I. (1991). Cell
Cross, F., Hartwell. L. H., Jackson, C., and Konopka, J. B. (1988). Annu. Rev. Cell Biol. 4, 429-457. Drubin, D., Zhu, Z.. Mulholland, 343, 288-290.
J., and Botstein, D. (1990). Nature
Finegold, A. A., Johnson, D. I,, Farnsworth, C. C., Gelb, M. H., Judd, S. R., Glomset, J. A., and Tamanoi, F. (1991). Proc. Natl. Acad. Sci. USA 88,4448-4452. Ford, S. K.. and Pringle, J. R. (1991). Dev. Genet., in press.
Gehrung,
S.. and Snyder, M. (1990). J Cell Brol. 771. 1451-1464
Gerhart, J., and Keller, R. (1986). Annu
Rev. Cell Brol. 2, 201-229.
Hasek, J., Rupes. I., Svobodova, J., and Streiblova, E. (1967). J. Gen. Microbial. 133, 3355-3363. Hyman, A. A. (1989). J. Cell Biol. 109. 1185-1193. Jackson, C L., and Hartwell. L. H. (1990). Mol. Cell. Biol. 10, 2202221 3. Jacobs, C. W.. Adams, A. E. M., Staniszlo, P. J.. and Pringle. J. R. (1988). J. Cell Brol. 707, 1409-1426. Johnson,
D. I., and Pringle, J. R. (1990). J. Cell Biol. 171, 143-152.
Kim, H. 8.. Haarer, B. K., and Pringle, J. R. (1991). J. Cell Biol. 772. 535-544. Kirschner. M., and Mitchison, T. (1986). Cell 45, 329-342. Kropf, D. L.. Kloareg, B., and Quatrano, 187-190.
R. S. (1988). Science 239.
Kupfer, A., and Singer, S. J. (1989). Annu. Rev. Immunol. 7, 309-337. Miyamoto, S.. Ohya, Y., Ohsumi, Y., and Anraku, Y. (1987). Gene 54, 125-l 32. Ohya, Y., Goebl, M., Goodman, L. E., Petersen-Bjorn, S., Fnesen, J. D., Tamanoi. F., and Anraku, Y. (1991). J. Biol. Chem., in press, Pringle. J. R.. Lillie, S. H., Adams, A. E M., Jacobs, C. W., Haarer, 8. K., Coleman, K. G., Robinson, J. S., Bloom, L., and Preston, R. A. (1986). In Yeast Cell Biology, J. Hicks, ed. (New York: Alan R. Liss, Inc.), pp. 47-80. Satminen, A., and Novick, P. (1987). Cell 49, 527-538. Schnepf, E. (1986). Annu. Rev. Plant Physiol. 37, 23-47. Sloat, B., Adams, A., and Pringle, J. (1981). J. Cell Biol. 89, 395-405. Snyder, M., Gehrung, S.. and Page, B. D. (1991). J. Cell Biol.. in press.
Development of Cell Polarity in Budding Yeast David 0. Drubin Department of Molecular and Ceil Biology University of California, Berkeley Berkeley, California 94720
For a wide variety of cellular processes, many constituents of the cell, such as plasma membrane proteins, organelies, and cytoskeletai filaments, must be organized asymmetrically within the cell. Cell polarity is required for the transmission of a nerve impulse, the development of a fertilized egg into a multicellular organism, the transport of molecules across an epitheiial ceil layer, the crawling of a fibrobiast cell, and the growth of a budding yeast ceil. Two central challenges in studies on the development of cell polarity are the identification of interdependencies between the many different events involved and the determination of the primary inductive events. Fundamental questions, such as how secretory vesicles are delivered to a particular membrane and how a spindle orients correctly to define a division axis, have not been answered. The yeast ceil, while lacking some interesting behaviors of higher cells, possesses an experimental advantage in its simplicity. Furthermore, yeast present an opportunity to genetically dissect the molecular pathways that lead to the development of cell polarity, much as was done for phage assembly pathways. In many cells, the development of polarity appears to begin at the cortex. For example, the site of fertilization on a frog egg determines the dorsal-ventral axis for embryonic development (Gerhart and Keller, 1986). An axis of ceil polarity can be stably defined on the cortex by the assembly of a nondiffusible axis marker. In plants, the cell wall plays a role in this process (Kropf et al., 1988). The asymmetry marked on the ceil cortex must be spread to the cytoplasmic compartment. The cytoskeleton is well suited to this function: the growth and stability of actin filaments and microtubules can be controlled by interac-
Minireview
tions of their fast-growing “plus” ends with particular regions of the cell cortex (Kirschner and Mitchison, 1986) and the inherent polarity of cytoskeietai filaments can spread polar organization from the cortex to the interior of the cell through interactions between motor proteins that bind to organeiies and polar arrays of cytoskeletal filaments. The Phenomenology of Polarized Yeast Cell Growth and Divlslon S. cerevisiae cells grow asymmetrically by localized vesicle fusion and cell wall synthesis, which result in the formation of buds or mating projections (Figures 1 and 2; Sloat et al., 1981; Pringie et al., 1986, and references therein). The asymmetry of yeast growth is also reflected in the arrangement of the cytoskeleton (Barnes et al., 1990) and many organeiies (Baba et al., 1989). In other cell types, polarized membrane traffic is thought to be important for such processes as ceil migration and extracellular matrix secretion. Yeast ceils choose bud sites in a nonrandom pattern determined by their mating type (Chant and Herskowitz, 1991, and references therein). a and a haploid ceils (as well as a/a and a/a diploid ceils, which mate as a and a ceils, respectively) bud from sites near the site of the previous division (axial budding, Figure 2). a/a diploid ceils bud in a bipolar manner: mother ceils bud from a site either near the site used for the previous division or near the opposite pole, and daughters usually bud from the pole opposite to their birth pole (Figure 2). During mating, a and a ceils ignore genetically determined axis formation instructions and orient toward a mating partner (see Cross et al., 1988). A yeast ceil’s ability to choose a mating partner displays the same high degree of ceil-type selectivity that is required of a cytotoxic T ceil when it uses its secretory cargo to kill a target ceil (Jackson and Hartwell, 1990; Kupfer and Singer, 1989). Polarized secretion is also exploited by mating yeast cells to deliver proteins specifically required for mating (such as Fuslp and agglutinins) to the projection tip (Cross et al., 1988). This review will focus on the budding pathway, but the
MATING BUD FORMATION
PROJECTION FORMATION
Figure 1. Cell Polarity in 9 cerevieiae Chitin is located extracellularly, while the products of C/X3, CDC70, CDCll. and CD02 are located intracellularly.
AXIAL BUDDING (amdo cells)
BIPOLAR BUDDING (ah%WIIS)
Figure 2. Orientation of the Yeast Growth Axis during Mating and Bud Formation Although the shapes of zygotes suggest that cells of opposite mating type can grow toward each other, this has not been documented.
C0ll 1094
CONTROL OF BUDDING PATTERN
BUD SITE ASSEMBLY
CDC24, CDC42, CDC43, BEMl, .9Ehf2, (MSBI, MB.?)
POLAR GROWTH
Figure 3. Morphogenetic
Pathway for Bud Formation
similarity of the processes of bud formation and projection formation should be noted (Figure 1; Cross et al., 1988). Molecular Genetic Studies Based on phenotypes, genetic interactions, and the interdependencies and timing of events, the genes required for cell polarity development in a mitotically growing S. cerevisiae cell can be arranged in hierarchical groupings (Figure 3). In this scheme, the budding pattern genes are required for correctly choosing a bud site. The bud site assembly genes are required to assemble components at the chosen bud site so that surface growth will be restricted to the bud. Microtubules, actin filaments, and neck filaments all appear to become organized through interactions with structures at the bud site (Figure 1). Microtu-
BUD SITE ASSEMBLY
BUDDING PAlTERN .._....
bules are required for moving the nucleus to the bud neck and orienting the spindle (see Barnes et al., 1990, for a review of the yeast cytoskeleton), while actin filaments and neck filaments are required for organizing surface growth and for cytokinesis. The BUD7 gene was originally named RSR7 and was identified by its ability, when overexpressed, to suppress a mutation in CDC24 (Bender and Pringle, 1989). Subsequently, BUD7 and four additional BUD genes were identified in a visual screen for mutants with altered budding patterns (Chant and Herskowitz, 1991; Chant et al., 1991). Mutations in BUDI, BUDP, or BUDS cause random budding in cells, irrespective of mating type. Mutations in BUD3 and BUD4 cause bipolar budding in a and a cells, which normally exhibit axial budding; however, these mutations have no effect on bipolar budding in a/a cells. Based on these observations, Chant and Herskowitz (1991) propose that in a and a cells budding would be axial, because all BUD genes are active. In a/a cells, on the other hand, the BUD3 and/or BUD4 gene might be inactive, resulting in bipolar budd!ng. Chant and Herskowitz (1991) point out that the al-a2 repressor present in a/a cells might inhibit either or both of these genes. When a yeast cell commits to a mitotic cell cycle in early Gl , several constituents appear on a specific region of the cell cortex that marks the site from which a bud will emerge, typically about 15 min later. These constituents initially include actin, 10 nm neck filaments, and the product of the gene SPA2 and subsequently chitin (Kim et al., 1991; Ford and Pringle, 1991; Snyder et al., 1991). Mutations in the bud site assembly genes prevent the assembly of these constituents at a bud site. As a result, the cells grow isotropically and become multinucleate because nuclei continue to divide in the absence of bud formation (Adams et al., 1990). Refinement of the scheme presented in Figure 3 will require determination of the location of each gene product in the cell, the biochemical activities and interactions of each protein, and when in the cell cycle the function of
ADDITIONAL INTERACTING GENES
-w
EUDI/RSRI(GTP binding) --.----.BUD2 BUD3 BUD4
BUD5 (GDP-GTP exchange)
)
4
.--....--.) 4------+
Synthetic Lethality Suppression by Overproduction Potential Interectlon Bawd on Sequence Homologlee
Figure 4. Genetic Interactions among Cell Polarity Genes Sequence homologies and the interactions they suggest are also indicated.
yin&view
each protein is required. However, numerous genetic interactions, suggestive sequence homologies, biochemical activities, and experiments that address the interdependence of gene functions all provide compelling evidence that these genes function in a morphogenetic pathway. Two genetic links suggest interactions between the budding pattern genes and the bud site assembly genes (Figure 4). First, overexpression of BUD7 can suppress some mutations in CDC24 (Bender and Pringle, 1989). Second, mutations in BUD5 and a gene involved in bud emergence (BEM7) show a negative synergism, resulting in so-called synthetic lethality in the double mutant (Chant et al., 1991). BEM7, in turn, is linked to CDC24 because both display genetic interactions with the same gene (MSBl; Bender and Pringle, 1991). This chain of interactions ultimately links BUD7 to BUDS. Additional interactions shown in Figure 4 and the similarity of the phenotypes observed for the mutants c&24, cdc42, c&43, beml, and bem2 provide further evidence that the bud site assembly genes function together (Bender and Pringle, 1989, 1991; Adams et al., 1990). Mutations in CDC24 and overproduction of CDC42 can lead to random budding patterns (Sloat et al., 1981; Johnson and Pringle, 1990). Other perturbations (see below) can lead to mild alterations in budding pattern. These observations provide further evidence for interactions between the bud site assembly and budding pattern genes but also blur the distinction between the two classes. However, separate groupings can be justified (Chant et al., 1991) because BUD genes appear to be required for budding pattern but not for bud formation. The bud site assembly genes, on the other hand, are essential for bud formation. Mutations in bud site assembly genes that affect budding pattern are proposed to perturb interactions between proteins encoded by bud site assembly and budding pattern genes. Why yeast should have genes whose sole function is to control budding pattern is not understood, since random budding causes no apparent disadvantage. The observation that budding pattern is perturbed by mild defects in bud site assembly genes suggests a second possibilitythe BUD genes may encode proteins with activities that are required for bud site assembly but are functionally redundant, so that only nonlethal defects are caused by gene disruptions. This interpretation would fit with the observation that synthetic interactions (e.g., between BUD5 and BEMI) can result when proteins function in a common step of a pathway (Salminen and Novick, 1987). Redundant proteins might be identified by the same sort of genetic screen that was used to identify BEM7 (Bender and Pringle, 1991). Some proteins required for the development of cell polarity have sequence homology with proteins having established biochemical activities, which suggests other activities and interactions (Figure 4). For example, the sequence of CDC24 (also known as CLS4) and the calcium sensitivity of certain cdc24 mutants indicate that the activity of Cdc24p might be modulated by calcium (Miyamoto et al., 1987). Interestingly, calcium has also been implicated in the function of Cdc43p (also known as Call p) and
is suggested to play a role in polarity development for certain cell types (Ohya et al., 1991; Schnepf, 1988). Perhaps most suggestive are the sequence homologies discovered for the proteins encoded by CDC42, CDC43, BUDl, and BUD5 The CDC42 and BUDlIRSRl gene products are homologous to small GTP-binding proteins in the ras family (Johnson and Pringle, 1990; Bender and Pringle, 1989). Additional support for the conclusion that small GTP-binding proteins play a role in cell polarity development comes from the discovery that Bud5p is homologous to the CDC25-encoded GDP-GTP exchange protein (Chant and Herskowitz, 1991). Furthermore, Cdc43p is required for S. cerevisiae geranylgeranyltransferase activity (Finegold et al., 1991) and is homologous to the protein encoded by DPRllFlAMl that prenylates the C-termini of small GTP-binding proteins (Ohya et al., 1991). These homologies and activities indicate that Cdc43p and Bud5p might each interact with Budlp and/or Cdc42p (both of which have appropriate C-terminal consensus sequences for modification by Cdc43p). Prenylation of Budlp and/or Cdc42p might be important for assembly of bud sites at the plasma membrane. At present there are no data that address the mechanism by which the two small GTPbinding proteins function in cellular morphogenesis. They might direct or monitor an assembly process, or they might regulate the timing of events. While many interactions have been observed between bud site selection and assembly genes, no such interactions have been observed between bud site assembly genes and cytoskeleton protein genes. The assembly of each filament system seems largely independent of assembly of the other two systems, but proper assembly of all three is dependent on the action of bud site assembly genes (Figure 3; Snyder et al., 1991; Adams et al., 1990; Ford and Pringle, 1991). Abplp is one candidate for a link between bud site assembly and assembly of the actin cytoskeleton; overproduction of this protein results in an aberrant budding pattern (Drubin et al., 1990). Recent studies suggest that microtubules emanating from the spindle pole body are captured in early Gl by a structure (located proximal to a concentration of Spa2p) at the incipient bud site (Snyder et al., 1991). The nucleus then rotates so that the spindle pole body is proximal to the bud site (Snyder et al., 1991). Although SPA2 is placed along the microtubule pathway in Figure 3, it seems not to function exclusively in conjunction with microtubules. Spa2p is required for mating projection formation (Gehrung and Snyder, 1990), but microtubules are not (Hasek et al., 1987). Bud Site Selection The fact that there is aspatial pattern to bud site selection suggests that a structure that marks the previous bud site must persist from one cell cycle to another (Snyder et al., 1991; Chant and Herskowitz, 1991). Axial budding might require that the BUD3 and BUD4 gene products interact with this structure and direct the assembly of a new bud site nearby. During C. elegans development, the midbody from the previous cell division, or a proximal structure, has been proposed to constitute a microtubule capture site that can orient the spindle for the next cell division (Hyman, 1989).
Cell 1096
In the absence of Bud3p or Bud4p, a bud site might form at the opposite pole from the last division because that is where the spindle pole body points after chromosome separation. This would be the default budding pattern. However, this scenario has the following problems. First, it does not explain how random budding is generated. Second, it is not consistent with the finding that microtubules do not appear to be important for axial or bipolar budding (Jacobs et al., 1988) but appear to be captured at a bud site after its assembly is initiated (Snyder et al., 1991). Finally, it does not account for the observation that, during bipolar budding, the mother cell often buds near the previous bud site, requiring a 1 80° rotation of the spindle pole body with respect to the cell surface. The fact that mutations can cause random budding suggests that the default budding pattern is random and that bipolar budding requires a special mechanism, the nature of which is obscure. Summary and Conclusions The development of cell polarity involves virtually every aspect of cell biology. Yeast are less complex than cells traditionally used for studies on cell polarity and are amenable to sophisticated genetic analysis. This has resulted in a growing number of molecular markers for yeast cell polarity and an increasingly well-defined progression of molecular events required for bud formation. Together, these factors provide a favorable context in which to understand how the interplay between a large number of processes can polarize a cell. Many genes required for morphogenesis have been identified, and genetic interactions provide evidence that the products of these genes function together. Studies on cell polarity development in S. cerevisiae have demonstrated a requirement for small GTP-binding proteins and have established functional relationships between temporally coincident events. With the continued identification and analysis of genes required for morphogenesis, and the pursuit of these studies on a cytological and biochemical level, studies on yeast will continue to contribute to our understanding of cell polarity development. References Adams, A. E., Johnson, D. I., Longnecker, R. M., Sloat, B. F., and Pringle, J. R. (1990). J. Cell Biol. 117, 131-142. Baba, M., Baba, N., Ohsumi, Y., Kanaya, K., and Osumi, M. (1989). J. Cell Sci. 94, 207-216. Barnes, G., Drubin, D. G., and Stearns, T. (1990). Curr. Opin. Cell Biol. 2,109-115. Bender, A., and Pringle, J. R. (1989). Proc. Natl. Acad. Sci. USA 86, 9976-9980. Bender, A., and Pringle, J. Ft. (1991). Mol. Cell. Biol. II, 1295-1305. Chant, J., and Herskowitz,
I. (1991). Cell 65, this issue.
Chant, J., Corrado, K., Pringle, J. R., and Herskowitz, 65, this issue.
I. (1991). Cell
Cross, F., Hartwell. L. H., Jackson, C., and Konopka, J. B. (1988). Annu. Rev. Cell Biol. 4, 429-457. Drubin, D., Zhu, Z.. Mulholland, 343, 288-290.
J., and Botstein, D. (1990). Nature
Finegold, A. A., Johnson, D. I,, Farnsworth, C. C., Gelb, M. H., Judd, S. R., Glomset, J. A., and Tamanoi, F. (1991). Proc. Natl. Acad. Sci. USA 88,4448-4452. Ford, S. K.. and Pringle, J. R. (1991). Dev. Genet., in press.
Gehrung,
S.. and Snyder, M. (1990). J Cell Brol. 771. 1451-1464
Gerhart, J., and Keller, R. (1986). Annu
Rev. Cell Brol. 2, 201-229.
Hasek, J., Rupes. I., Svobodova, J., and Streiblova, E. (1967). J. Gen. Microbial. 133, 3355-3363. Hyman, A. A. (1989). J. Cell Biol. 109. 1185-1193. Jackson, C L., and Hartwell. L. H. (1990). Mol. Cell. Biol. 10, 2202221 3. Jacobs, C. W.. Adams, A. E. M., Staniszlo, P. J.. and Pringle. J. R. (1988). J. Cell Brol. 707, 1409-1426. Johnson,
D. I., and Pringle, J. R. (1990). J. Cell Biol. 171, 143-152.
Kim, H. 8.. Haarer, B. K., and Pringle, J. R. (1991). J. Cell Biol. 772. 535-544. Kirschner. M., and Mitchison, T. (1986). Cell 45, 329-342. Kropf, D. L.. Kloareg, B., and Quatrano, 187-190.
R. S. (1988). Science 239.
Kupfer, A., and Singer, S. J. (1989). Annu. Rev. Immunol. 7, 309-337. Miyamoto, S.. Ohya, Y., Ohsumi, Y., and Anraku, Y. (1987). Gene 54, 125-l 32. Ohya, Y., Goebl, M., Goodman, L. E., Petersen-Bjorn, S., Fnesen, J. D., Tamanoi. F., and Anraku, Y. (1991). J. Biol. Chem., in press, Pringle. J. R.. Lillie, S. H., Adams, A. E M., Jacobs, C. W., Haarer, 8. K., Coleman, K. G., Robinson, J. S., Bloom, L., and Preston, R. A. (1986). In Yeast Cell Biology, J. Hicks, ed. (New York: Alan R. Liss, Inc.), pp. 47-80. Satminen, A., and Novick, P. (1987). Cell 49, 527-538. Schnepf, E. (1986). Annu. Rev. Plant Physiol. 37, 23-47. Sloat, B., Adams, A., and Pringle, J. (1981). J. Cell Biol. 89, 395-405. Snyder, M., Gehrung, S.. and Page, B. D. (1991). J. Cell Biol.. in press.
