Trimorphic stepping stones pave the way to fungal virulence.

T he fungal kingdom encompasses 1.5 million species (1) as diverse as single-celled yeasts, pathogens of animals/plants, and plant root symbionts. Fungi are eukaryotic, closely aligned with metazoans (2, 3). Animals and fungi diverged 1 billion years ago; their last common ancestor was unicellular, motile, and aquatic. Some fungi grow as unicellular yeasts, but most are filamentous multicellular organisms. Importantly, some fungi are dimorphic, growing as both yeast and filamentous forms (i.e., Saccharomyces cerevisiae). Yet others are trimorphic and grow as yeast, hyphae, and pseudohyphae (i.e., Candida albicans). S. cerevisiae pseudohyphal growth long eluded detection—it requires special conditions/strains and was lost during domestication (4). How pseudohyphae are related to yeast and hyphae (as a distinct fate or a continuum) was unknown until the report of Carlisle et al. (5) in this issue of PNAS. They reveal that pseudohyphae are intermediate between yeast and hyphae, with implications for pathogen–host interactions and fungal evolution. C. albicans is a trimorphic fungus from the Ascomycota, diverged from S. cerevisiae 200 million years ago (6, 7), and a commensal of the human microbiota of the gastrointestinal tract, mucous membranes, and skin. C. albicans infects humans, causing oropharyngeal disease, vaginitis, and systemic life-threatening infection. Serum induces C. albicans yeast to produce germ tubes, and growth in defined media elicits two filamentous growth modes: pseudohyphae and hyphae (Fig. 1A). Pseudohyphae resemble hyphae, but are morphologically distinct. In both modes the normally ovoid yeast cells are elongated, but in pseudohyphae each cell–cell junction is constricted, and the diameter between cell walls is wider in the middle than the ends. In contrast, hyphal cell walls are parallel with no mother–daughter or septal junction constrictions. Pseudohyphae and hyphae also differ with respect to septin localization and nuclear division (7). Given these differences, it has been unclear whether pseudohyphae are an intermediate between yeast and hyphae or are an alternative fate (Fig. 1). This question, and implications for virulence, were addressed by Carlisle et al. with engineered strains (5). The C. albicans dimorphic yeast–hyphae transition is thought to underlie its success as a pathogen. Mutants locked as yeasts (lacking Cph1 and Efg1 transcription factors or cyclin Hgc1) or filaments (lacking the Tup1 repressor) are both avirulent, linking both forms to pathogenesis (8–10). Subsequently, a strain was engineered in which morphogenesis is controlled by regulated expression of a filamentation repressor, Nrg1, with the tet promoter (11). Cells grown without doxycycline express Nrg1 and grow as yeast, whereas growth with doxycycline repressed Nrg1 and filamentous growth ensued. Animals infected with yeast remained healthy yet harbored a significant latent fungal burden in the kidney. Adding doxycycline to drinking water activated filamentation, with progression to lethal infection. These studies provide robust support for the concept that dimorphic transitions underlie C. albicans virulence, and they show that yeast can penetrate tissues, whereas hyphae are necessary for progression to lethal infection. In cultured macrophages S. cerevisiae is killed after phagocytosis, whereas C. albicans yeast switch to hyphae, killing and escaping macrophages (12). The studies of Carlisle et al. (5) involve controlling expression of Ume6, a zinc-finger transcription factor downstream of Nrg1 in circuits for dimorphic transition and virulence (13, 14). In their studies, Ume6 was expressed from the doxycycline-regulatable promoter (tet-UME6), leading to controlled consequences: yeast with the promoter off (dox ), and hyphae in response to high-level Ume6 (dox ), even under non-filamentinducing conditions. Intermediate expression elicited a third fate: pseudohyphae. Increasing Ume6 levels converted pseudohyphae to hyphae, and hybrid hyphal/pseudohyphal filaments were observed. When the tetUME6 strain was grown as hyphae (dox ) and then shifted to repress (dox ), filaments produced a majority of pseudohyphae by 3 h and yeast by 7 h. Thus both filamentous modes can be reversibly evoked by expressing a single regulatory element in a dosagedependent fashion, providing evidence that pseudohyphae are intermediate between yeast and hyphae, rather than a distinct fate. When animals were infected with yeast of the tet-UME6 strain (dox ), increased Ume6 expression led to enhanced hyphal growth and tissue invasion and more rapid demise; animals given doxycycline survived much longer. Hence, hyphal growth (or hyphalspecific gene expression) promotes C. albicans virulence. The findings of Carlisle et al. (5) have broad implications for dimorphism in fungal pathogenesis (Fig. 2). S. cerevisiae and C. albicans both undergo yeast–pseudohyphal transitions, but only C. albicans develops hyphae. As C. albicans evolved into a successful commensal of the mammalian gastrointestinal tract, formation of hyphae in biofilms was likely necessary to compete with bacteria, or to form cooperative multispecies biofilms (15). Hyphae are also critical for C. albicans to survive and escape host macrophages. As a result, the trimorphic species C. albicans is a successful commensal and pathogen, whereas the dimorphic yeast S.