Cammie F. Lesser is a Senior Fellow in Infectious Diseases and Samuel I. Miller is a Professor in the Department of Microbiology, University of Washington, Seattle.
Article Links:
Links to Other ASM Pages:
Tractable S. cerevisiae Is Model System for Studying Pathogenic Bacterial Proteins
Introducing bacterial virulence proteins into yeast provides valuable insights about their interactions with host components
Cammie F. Lesser and Samuel I. Miller
How do bacteria cause disease? One means is by secreting toxins that poison eukaryotic cells. Also, and perhaps more importantly, many bacterial pathogens interact intimately and in a more complex manner through the intracellular action of multiple bacterial proteins on host cell targets. For example, many gram-negative pathogenic bacteria of both plants and animals inject, or translocate, proteins directly into eukaryotic cells by a process called type III secretion.
Type III secretion systems consist of more than 20 different proteins that form a complex structure spanning both the inner and outer membranes of the gram-negative bacterial envelope, which inserts into the eukaryotic host cell membrane (Fig. 1). Once assembled, this structure apparently forms a channel through which sets of virulence proteins from the bacteria are translocated into eukaryotic cells. Researchers first identified type III secretion systems in Yersinia, Salmonella, and Shigella species, but subsequently have found similar systems in many other gram-negative pathogenic bacteria.
Recognizing that so many bacterial pathogens likely directly affect or interfere with basic cellular processes of eukaryotic hosts, we decided to characterize these interactions by using the yeast Saccharomyces cerevisiae as a model system. This yeast, unlike higher eukaryotic systems, is genetically tractable, thereby providing a means to overcome a previous major limitation in determining functions of bacterial virulence proteins. In addition, many mammalian cells used for these purposes are transformed immortalized cells, meaning that results from such studies might not be applicable to in vivo situations. Moreover, since the sequence of the S. cerevisiae genome was determined in the mid-1990s, researchers have been developing and making available a plethora of yeast-specific reagents—such as DNA arrays, a complete set of individually tagged yeast deletion strains, a set of expression plasmids encoding GST (glutathione-S-transferase) fusion genes, and numerous overexpression libraries—that should prove to be extraordinarily useful for analyzing host-bacterial pathogen interactions.
S. cerevisiae Offers Advantages in Analyzing Type III Secreted Proteins
Type III secretion systems translocate at least 60 bacterial proteins into eukaryotic cells. More will likely be identified as additional bacterial genomes are sequenced. Each pathogen translocates a unique set of proteins that promote survival of the pathogen in the host and help promote (or establish) disease.
Few clues exist as to the functions of these translocated virulence proteins. In limited cases, bacterial virulence proteins contain regions of homology to mammalian enzymes, including serine-threonine kinases, tyrosine phosphatases, and adenylate cyclases. In several of these cases, the purified virulence protein carries out the biochemical activity that was proposed for it based on sequence homology, although how such mimicry of specific host biochemical functions contributes to pathogenesis is not entirely clear.
Researchers are taking several approaches to study how specific translocated proteins contribute to bacterial virulence. One approach involves deleting genes that encode translocated proteins from specific pathogens and observing how their loss affects virulence or specific steps in pathogenesis. Another approach entails introducing individual virulence proteins into eukaryotic cells—either by adding genes encoding such bacterial proteins or adding the proteins directly to such cells—and then observing phenotypic changes. In general, these studies provide insights about gross morphologic changes in such cells. Depending on the proteins being introduced, they indicate, for instance, that phagocytosis is inhibited, the cytoskeleton rearranged, or either cytotoxicity or apoptosis is induced.
Long used for studying eukaryotic cell processes, S. cerevisiae offers advantages over more complex mammalian cells for studying how bacterial virulence proteins affect cellular functions. Even though mammalian cells are generally more complex than yeast, there is a high degree of overlap of their respective cellular pathways and proteins, including the actin cytoskeleton, cell cycle, pre-mRNA processing, and protein sorting. Similarly, more than 70 human cDNA clones are known to complement S. cerevisiae mutants, and 21% of positionally cloned genes linked to various human diseases carry a high degree of homology to genes of S. cerevisiae. Because many bacterial virulence proteins likely target cellular processes or signaling cascades that are common to all eukaryotes, we decided to study bacterial virulence proteins in yeast, expecting to speed our efforts to identify mechanisms of host-pathogen interactions that could be subsequently tested further in more complex eukaryotic systems.
Translocated Bacterial Proteins Localized Similarly in Yeast and Mammals
Although many type III secreted bacterial proteins are initially translocated into the cytoplasm of mammalian cells, several translocated proteins of Yersinia pseudotuberculosis subsequently relocate to specific sites, including YpkA to the plasma membrane and YopM to the mammalian cell nucleus. Presumably, the redistribution of these proteins reflects the roles that they play during pathogenesis. We first determined whether these proteins end up in the same subcellular sites in mammals and yeast.
Our tests entailed separately fusing each of the translocated bacterial virulence proteins of Y. pseudotuberculosis—YopE, YopJ, YopH, YopM, and YpkA—to green fluorescent protein (GFP) of Aqueoria victoria. Each of the fusion genes was placed under control of an inducible promoter in case expression of any was toxic to yeast. Yeast carrying plasmids encoding each of the fusion genes were grown under conditions that promote expression of the GFP fusion proteins and observed under a fluorescent microscope.
Remarkably, the localization patterns in yeast paralleled those previously identified in mammalian cells. YopM localized to the nucleus, YpkA to the plasma membrane, and YopH, YopJ, and YopE to the cytoplasm (Fig. 2). (The subcellular localization of YopJ has never been studied in mammalian cells). Notably, the route of delivery of the bacterial virulence proteins in yeast and mammals is completely different. While the bacteria normally inject their virulence proteins across the bacterial and mammalian plasma membranes directly into the mammalian cytoplasm, in yeast the bacterial proteins are expressed from fusion genes. Thus, in yeast, the bacterial proteins were correctly translated, folded, and potentially modified, such that they maintained their distinctive subcellular localization patterns. Since YopM lacks a known nuclear localization signal and YpkA lacks a transmembrane domain, these proteins likely interact with cellular components that are conserved from yeast to mammals. Rapid identification of these components in yeast should provide additional clues as to the roles of these proteins in pathogenesis. In addition, the strong conservation of subcellular localization patterns observed from yeast to mammals for the well-studied Yersinia proteins suggests that subcellular localization patterns of previously uncharacterized translocated proteins in yeast will provide potential clues as to their function in higher eukaryotes.
Some of the Yersinia Virulence Proteins Prove Toxic when Expressed in Yeast
We were encouraged by the observation that the subcellular localization patterns in yeast paralleled those previously observed in mammalian cells. However, the main advantage of using yeast to study bacterial virulence proteins is the genetic tractability of Saccharomyces cerevisiae. Since the most straightforward way to conduct genetic selections or schemes is to exploit growth phenotypes, we next determined if expression of any of the individual Yersinia GFP fusion proteins affected yeast growth.
Plasmids encoding each of the Y. pseudotuberculosis translocated proteins fused to GFP under the control of the galactose-inducible GAL1 promoter were introduced into yeast. Under inducing conditions, yeast cells expressing either YopH or YopJ grow similarly to wild-type cells, whereas those expressing YopE or YpkA grow very poorly or not at all, and those expressing YopM grow at an intermediate rate. These growth phenotypes can be exploited to identify functional domains of the virulence proteins as well as the cellular pathways or proteins that these virulence proteins target.
Soon after we observed that YopE expression in yeast is toxic, Yixin Fu and Jorge Galan at Yale University, New Haven, Conn., reported that the Salmonella YopE homolog, SptP, has RhoGAP (GTP activating protein) activity. Moreover, changing an (ordinarily) invariant arginine to alanine in a short consensus sequence (GxxRxxGS) shared by YopE, SptP, and their Pseudomonas homolog, ExoS, eliminates the RhoGAP activity of purified SptP.
To test if YopE toxicity in yeast is due to its RhoGAP activity, we introduced a mutation to convert the conserved arginine of the Rho GAP consensus sequence of YopE to alanine. Expression of this mutant YopE allele in yeast only minimally inhibited growth, suggesting to us that YopE toxicity in yeast is linked to RhoGAP activity. Indeed, we and Ulrich von Pawel-Rammingen and his collaborators at Umea University in Sweden demonstrated that purified YopE has RhoGAP activity in vitro.
This approach can be used to identify additional functional domains of YopE or other bacterial proteins whose expression confers a growth phenotype in yeast. For example, by selecting for YopE mutants that are still expressed as a stable protein in yeast but whose presence is not toxic to the cells, one could determine the G-protein binding domain of YopE. Candidate proteins can then be purified and characterized by biochemical assays.
Mammalian Rho G-proteins mediate rearrangements of the actin cytoskeleton of cells. For example, when Rho is activated, actin stress fibers form. When YopE is injected into mammalian cells, stress fibers are disrupted—presumably because RhoGAPs, like YopE, inactivate Rho G-proteins. Disrupting the actin cytoskeleton likely enables Yersinia to evade the mammalian immune system by blocking uptake of the bacterial cells into professional phagocytic cells such as macrophages.
Many actin cytoskeleton components are well conserved between yeast and mammals. For example, yeast actin is 89% identical to mammalian actin and, although yeast do not encode as many members of the Rho family of G-proteins, the members they do encode are 40-80% identical to their mammalian homologs. Moreover, yeast cells do not contain actin stress fibers. Instead, the yeast actin cytoskeleton contains two distinct components, cortical patches and actin cables, whose definitions are based on fluorescence microscopy. Cortical patches—actin-rich punctate bodies at the cell cortex—collect at the site of bud formation during cell division. Actin cables are bundles of actin filaments that also localize at the cell cortex. Both components align along the mother-bud axis in a polarized fashion.
Since YopE toxicity in yeast is likely due to its RhoGAP activity, we sought to determine whether its toxicity reflects an ability to disrupt the actin cytoskeleton in yeast. When expressed in yeast, YopE blocks the polarization of both cortical patches and actin cables that ordinarily occurs during bud formation. In fact, when YopE is present, no actin cables can be visualized (Fig. 3). Whether this virulence protein actively disrupts actin cables or inhibits their formation is not known. Nevertheless, in yeast, as in mammalian cells, expression of YopE effectively disrupts the actin cytoskeleton.
Salmonella Manipulates the Actin Cytoskeleton of Higher Eukaryotes
Besides YopE, many other bacterial virulence proteins affect the mammalian actin cytoskeleton. For example, Salmonella manipulate the host cytoskeleton to promote their own uptake into normally nonphagocytic cells, such as epithelial cells, thereby evading the host immune system. Upon contact with nonphagocytic cells, Salmonella promote the formation of protrusions, or "ruffles," along the host cell membrane that engulf the bacteria. Once internalized, Salmonella survive within specialized membrane-bound compartments (phagosomes). Membrane ruffle formation is dependent on an intact type III secretion system, which is encoded on Salmonella pathogenicity island 1 (SPI1).
SPI1 translocates at least seven bacterial proteins into eukaryotic cells, several of which facilitate membrane ruffling. In fact, injecting SopE1 or SopE2, both of which are Rho GTP exchange factors (GEF) that activate Rho G-proteins, triggers diffuse membrane ruffling. Notably, during Salmonella infection membrane ruffling is confined to sites of bacteria attachment. However, injecting SopE1 along with SptP, which is a RhoGAP, blocks membrane ruffling. When these proteins are injected sequentially by the type III secretion apparatus, they may regulate membrane ruffling—first turning it on, then halting it.
Two other translocated Salmonella proteins interact with actin and may direct membrane ruffles to sites where the pathogen attaches to host cells. Purified SspA, a translocated protein, binds to and effectively decreases the concentration of actin required to form actin filaments, and inhibits the rate of actin filament depolymerization. Purified SspC, another translocated protein and a component of the translocase apparatus, nucleates actin filament formation and facilitates actin bundling. Because SspC also is inserted into the plasma membrane, it seems likely that it helps to direct membrane ruffles to sites of bacterial attachment. However, when nonphagocytic cells are infected with Salmonella mutants that encode SspC, but not SspA, membrane ruffles occur diffusely rather than at sites of bacteria contact. Thus, SspA, perhaps in conjunction with SspC, directs membrane ruffles to sites of bacterial attachment.
Salmonella Also Perturbs the Actin Cytoskeleton in Yeast
To study how SspA competes with other endogenous proteins which interact with the actin cytoskeleton in living cells, we introduced this protein into S. cerevisiae. When the GFP-SspA fusion protein is expressed at high levels in yeast, it colocalizes with the entire actin cytoskeleton (Fig. 4A), providing the first in vivo evidence that SspA and actin interact in living cells. Notably, when SspA is expressed in yeast, the actin cables appear thicker than normal and tend to lose their normal cellular polarity, suggesting that SspA binds and reorganizes the actin cytoskeleton. In addition, actin cables that colocalize with SspA are resistant to depolymerization by Latrunculin A, a drug that measures turnover of actin filaments. Thus, SspA also blocks depolymerization of actin filaments within cells.
Another advantage of the yeast system is the ability to regulate expression levels of proteins of interest by using promoters of different strengths or altering the copy number of the fusion gene. These types of experiments proved to be very informative when studying SspA. For instance, when the copy number of the GFP-SspA fusion gene is greatly decreased, thus resulting in relatively low- level expression of SspA, SspA initially associates with only a minor subset—1 or 2 of the 30-50—actin cortical patches in the cell (Fig. 4A). However, as more GFP-SspA is produced, the fusion proteins do not localize to additional cortical patches but rather with actin cables that appear to originate from those cortical patches that initially associated with SspA (Fig. 4A). These actin cables, like those observed in yeast expressing high levels of SspA, lose their normal cellular polarity. Thus, it appears that SspA can promote the growth of actin cables to specific sites in the cell.
These studies in yeast suggest that SspA displays several distinct activities in living cells: (i) binding to actin cytoskeleton, (ii) inhibiting depolymerization of actin cables, and (iii) redirecting actin cable growth to specific cellular locations. These observations provide clues about the role SspA plays during pathogenesis in mammalian cells. For example, we propose that SspA redirects the growth of stable actin cables to sites of bacterial attachment, thus promoting membrane ruffling to specific sites. Moreover, based on recent experiments in yeast, SspA might interact with additional host proteins that could redirect SspA to specific sites in the cell.
New Directions for the Study of Bacterial Virulence Proteins in Yeast
When we set out to test if yeast would serve as a model to study virulence proteins translocated by type III secretion systems, our work focused on relatively well-understood translocated bacterial proteins such as YopE and SspA. However, we now plan on using a variety of tools in yeast to identify functions for less-well-understood bacterial virulence proteins.
For example, researchers in several laboratories, including our own, have identified at least 10 additional proteins that apparently are translocated by the type III secretion system encoded within Salmonella pathogenicity island 2 (SPI2). While the type III secretion system encoded within SPI1 is required for uptake of Salmonella into nonprofessional phagocytes, the SPI2 type III secretion system is essential for survival of Salmonella within macrophages in the phagosomes. Notably, one of these proteins, whose function is undefined, localizes to the yeast vacuolar membrane (Fig. 4B). In addition, this protein is toxic to yeast, thus allowing use of convenient genetic screens to study just how it affects eukaryotic cells.
It is becoming more and more apparent that bacterial pathogens can directly manipulate the basic cellular processes of their eukaryotic hosts to promote their own survival and subsequently cause disease. The yeast Saccharomyces cerevisiae can be used to study host-pathogen interactions in an efficient and informative manner. We hope to continue using yeast to explore the functions of a variety of such proteins and to conduct experiments that would be difficult, if not impossible, to design if we were restricted to using more complex mammalian cells.
ACKNOWLEDGMENTS
Tyler Kimbrough generously provided electron micrographs and the schematic diagram presented in Fig. 1. C.F.L. is currently a Howard Hughes Physician Postdoctoral Fellow and was supported by a Pfizer Postdoctoral Fellowship while this work was conducted. This work was also supported by a grant to S.I.M. from the National Institute of Health RO1 A130479.
SUGGESTED READING
Cornelis, G. R. 2000. Molecular and cell biology aspects of plague. Proc. Natl. Acad. Sci. USA 97:8778-8783.
Fu, Y., and J. E. Galan. 1999. A Salmonella protein antagonizes Rac-1 and Cdc42 to mediate host-cell recovery after bacterial invasion. Nature 401:293-297.
Hayward, R. D., and V. Koronakis. 1999. Direct nucleation and bundling of actin by the SipC protein of invasive Salmonella. EMBO J. 18:4926-4934.
Hueck, C. J. 1998. Type III protein secretion systems in bacterial pathogens of animals and plants. Microbiol. Mol. Biol. Rev. 62:379-433.
Lesser, C. F., and S. I. Miller. 2001. Expression of microbial virulence proteins in Saccharomyces cerevisiae models mammalian infection. EMBO J. 20:1-10.
Miao, E. A., and S. I. Miller. 2000. A conserved amino acid sequence directing intracellular type III secretion by Salmonella typhimurium. Proc. Natl. Acad. Sci. USA 97:7539-7544.
Rosqvist, R., A. Forsberg, and H. Wolf-Watz. 1991. Intracellular targeting of the Yersinia YopE cytotoxin in mammalian cells induces actin microfilament disruption. Infect. Immun. 59:4562-4569.
Rosqvist, R., K. E. Magnusson, and H. Wolf-Watz. 1994. Target cell contact triggers expression and polarized transfer of Yersinia YopE cytotoxin into mammalian cells. EMBO J. 13:964-972.
Von Pawel-Rammingen, U., M. V. Telepnev, G. Schmidt, K. Aktories, H. Wolf-Watz, and R. Rosqvist. 2000. GAP activity of the Yersinia YopE cytotoxin specifically targets the Rho pathway: a mechanism for disruption of actin microfilament structure. Mol. Microbiol. 36:737-748.
Zhou, D., M. S. Mooseker, and J. E. Galan. 1999. Role of the S. typhimurium actin-binding protein SipA in bacterial internalization. Science 283:2092-2095.
Email: webmaster@asmusa.org