Special Viewpoints

Jekyll and Hyde in the Microbial World

See allHide authors and affiliations

Science  26 Nov 2004:
Vol. 306, Issue 5701, pp. 1509-1511
DOI: 10.1126/science.1104677


Fungi are nonmotile organisms that obtain carbon from compounds in their immediate surroundings. Confronted with nutrient limitation, the yeast Saccharomyces cerevisiae undergoes a dimorphic transition, switching from spherical cells to filaments of adherent, elongated cells that can invade the substratum. A complex web of sensing mechanisms and cooperation among signaling networks (including a mitogen-activated protein kinase cascade, cyclic adenosine monophosphate–dependent protein kinase, and 5′–adenosine monophosphate–activated protein kinase) elicits the appropriate changes in physiology, cell cycle progression, cell polarity, and gene expression to achieve this differentiation. Highly related signaling processes control filamentation and virulence of many human fungal pathogens.

For both the layperson and the molecular biologist, the most familiar fungal cell is baker's yeast (S. cerevisiae). This budding yeast has become a valuable model for examining the eukaryotic way of life at the molecular level. Although relatively benign, S. cerevisiae has shed light on fundamental mechanisms in fungal pathogenesis because it adopts two distinct morphologies: a spherical or ovoid (yeastlike) form, which proliferates by budding, and a filamentous form. For fungal pathogens, such as Candida albicans, the ability to undergo this dimorphic transition is strongly correlated with invasion of host tissue and virulence (1, 2).

On rich medium, S. cerevisiae proliferates as separate spherical cells. However, in response to nitrogen limitation (3), the cells become thin and oblong, and after a daughter buds, it remains adherent and connected end-to-end with its mother, resembling a link in a sausage string (Fig. 1). Elucidating the signaling pathways necessary to elicit such filamentous growth has led to insights about how cells respond to different stimuli and how they share components with other signaling pathways yet evoke unique responses that are physiologically appropriate.

Fig. 1.

The switch from budding (left) to pseudohyphal growth (right) in S. cerevisiae is controlled by at least three signaling inputs: Kss1, a MAPK; Tpk2, a PKA catalytic subunit isoform; and Snf1, a yeast ortholog of AMPK.

As our Connections Map on Science's Signal Tranduction Knowledge Environment (4) indicates, the switch from budding to pseudohyphal growth in S. cerevisiae is controlled by at least three signaling modalities (5). Each input involves the action of different classes of protein kinases. The core of one pathway is a three-tiered mitogen-activated protein kinase (MAPK) cascade (6). Each kinase catalyzes the transfer of phosphate from adenosine 5′-triphosphate (ATP) to serine, threonine, or tyrosine residues in the next kinase, which provides for switchlike amplification of signal propagation. The MAPK cascade operates in concert with a second input that acts by means of cyclic adenosine monophosphate (cAMP)–dependent protein kinase (PKA) (7), best studied in yeast for its role in glucose metabolism in vegetatively growing cells. When the supply of glucose is exhausted, a third input that acts through 5′-AMP-activated protein kinase (AMPK) ensures continued robust filamentous growth on fermentation end products and other nonglucose carbon sources (8, 9). AMPK (Snf1) may act upstream of or in conjunction with the action of the atypical protein kinases, Tor1 and Tor2 (10, 11).

The MAPK module is stimulated by two small guanosine triphosphatases (GTPases) anchored at the plasma membrane, Ras2 and Cdc42. How nutrient limitation promotes conversion of Ras2 and Cdc42 to their active (GTP-bound) state is not understood; however, Cdc42 acts downstream of Ras2. One role of Cdc42 is to activate a p21-activated protein kinase (PAK), Ste20. Ste20 then phosphorylates and activates the first kinase (Ste11) in the MAPK cascade, which in turn phosphorylates the next kinase (Ste7). These gene products are designated ste (for sterile), because they were first identified as required for the signaling that induces haploid yeast cells to mate. Once activated by Ste7, the MAPK (Kss1) phosphorylates two transcriptional repressors, Dig1 and Dig2, in the nucleus, which derepresses the transcription factor, Ste12 (12). Ste12 is targeted to the promoters of genes required for filamentous growth in a complex with a second transcription factor, Tec1 (13).

How does the signal get transmitted from the plasma membrane to the nucleus? Membrane-bound, activated Cdc42 binds to a specific sequence [Cdc42/Rac interactive binding (CRIB) motif] in the initiating kinase Ste20. Membrane association of Ste20 is enhanced by a scaffold protein, Bem1, that is membrane-tethered through its phosphoinositide-binding PX domain and interacts with Pro-rich motifs in Ste20 through tandem SH3 domains. The encounter of activated Ste20 with its substrate, the kinase Ste11, is ensured, given that Cdc42 also binds to Ste50, an adaptor protein that is tightly bound to Ste11. Also, Ste11 associates with a membrane-localized osmosensor, Sho1, and with a MAPK kinase (Pbs2) that is itself bound to Sho1 (14). Consequently, Ste11 gets phosphorylated and activated when Ste20 has been activated. Presumably, dissociation of Ste11 from the membrane permits it to encounter its target in the cytosol, the kinase Ste7. Ste7 forms tight complexes with Kss1, the third kinase in the MAPK cascade, through a MAPK-docking motif at the N terminus of Ste7 (15).

Kss1 also gets phosphorylated and activated, albeit only transiently, when a yeast cell responds to the peptide pheromone that induces mating (16). Yet, pheromone stimulation does not lead to a filamentous growth response. Pathway specificity is maintained, in part, because pheromone response channels the Ste11- and Ste7-dependent signals to a different MAPK, Fus3. Also, activated Fus3 phosphorylates Tec1 at a specific site (which is not a target of Kss1), marking Tec1 for ubiquitin-mediated degradation. The absence of Tec1 prevents expression of the bank of genes necessary for filamentous growth. Embarking on one course of action by the cell limits the other options open to it in a simple and straightforward fashion, which has enormous ramifications for understanding what developmental biologists call “commitment” to a particular cell fate during differentiation in multicellular organisms.

In addition to triggering the MAPK cascade, activated Ras2 stimulates adenylate cyclase (Cyr1), an integral membrane enzyme that converts ATP to cAMP (17), thus increasing the intracellular cAMP concentration. Binding of cAMP to the regulatory subunit (Bcy1) of PKA releases the three catalytic subunit isoforms (Tpk1, Tpk2, and Tpk3) from inhibition, thereby enhancing their activity and perhaps affecting their relative concentrations in the nucleus. Despite their similarity (67 to 76% identity), only Tpk2 is required to fully activate the filamentous growth response, whereas Tpk3 (but not Tpk1) has an inhibitory role. Tpk2 phosphorylates the transcription regulators, Flo8 and Sfl1. Sfl1 antagonizes Flo8-promoted expression of FLO11, which encodes the flocculin required for the cell-cell and cell-substratum adhesion necessary for filamentation. Phosphorylation by Tpk2 promotes Flo8 binding to and activation of the FLO11 promoter and, at the same time, relieves Sfl1-mediated repression by prohibiting its dimerization and DNA binding. Likewise, AMPK (Snf1) promotes sustained FLO11 expression by blocking two other repressors (Nrg1 and Nrg2).

Diploid yeast initiate filamentous growth when deprived of nitrogen. Haploid yeast undergo a similar dimorphic switch (dubbed “invasive growth”) but only when glucose is limiting. In either instance, the MAPK cascade and PKA are involved. How do yeast sense a decrease in the nitrogen or glucose supply? An ammonium permease (Mep2) (18) has been implicated in the former process, but whether its function is coupled to Ras2 activation is unclear. Nitrogen limitation causes increased expression of the GPR1 gene, which encodes a heterotrimeric GTP-binding protein (G protein)–coupled receptor that associates with a G protein α subunit, Gpa2 (19). The ligands for Gpr1 include glucose itself (20). Activated Gpa2 stimulates adenylate cyclase and promotes filamentous growth through PKA but does not need Ras2 to do so. Just as mammalian adenylate cyclase is activated directly by Gαs, Gpa2 may stimulate the cyclase Cyr1. Apparent Gβ subunits for Gpa2 (Gpb1/Krh2 and Gpb2/Krh1) and a candidate Gγ subunit (Gpg1) have been found. These proteins act as negative regulators of PKA signaling, possibly by sequestration of Gpa2, but action on other targets has not been ruled out.

Two integral membrane proteins, Sho1 and Msb2 (perhaps in a complex), are necessary for Kss1 (MAPK) activation under conditions that promote filamentous growth, and Msb2 interacts directly with activated (GTP-bound) Cdc42 (21). Both Sho1 and Msb2 also provide stimulatory inputs in response to hyperosmotic stress. These new insights do not yet reveal what is sensed by Sho1 and Msb2, nor how Ras2 and Cdc42 become activated, to evoke filamentous growth. Interestingly, Msb2 strongly resembles a class of mammalian membrane-anchored glycoproteins, called mucins, implicated in metazoan cell signaling (22). Clearly, more will be learned about environmental and nutrient sensing, integration of signaling pathways, and signaling specificity in complex networks by continued study of the filamentous growth process in yeast.

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article