A phenotypic profile of the Candida albicans regulatory network

Abstract

Candida albicans is a normal resident of the gastrointestinal tract and also the most prevalent fungal pathogen of humans. It last shared a common ancestor with the model yeast Saccharomyces cerevisiae over 300 million years ago. We describe a collection of 143 genetically matched strains of C. albicans, each of which has been deleted for a specific transcriptional regulator. This collection represents a large fraction of the non-essential transcription circuitry. A phenotypic profile for each mutant was developed using a screen of 55 growth conditions. The results identify the biological roles of many individual transcriptional regulators; for many, this work represents the first description of their functions. For example, a quarter of the strains showed altered colony formation, a phenotype reflecting transitions among yeast, pseudohyphal, and hyphal cell forms. These transitions, which have been closely linked to pathogenesis, have been extensively studied, yet our work nearly doubles the number of transcriptional regulators known to influence them. As a second example, nearly a quarter of the knockout strains affected sensitivity to commonly used antifungal drugs; although a few transcriptional regulators have previously been implicated in susceptibility to these drugs, our work indicates many additional mechanisms of sensitivity and resistance. Finally, our results inform how transcriptional networks evolve. Comparison with the existing S. cerevisiae data (supplemented by additional S. cerevisiae experiments reported here) allows the first systematic analysis of phenotypic conservation by orthologous transcriptional regulators over a large evolutionary distance. We find that, despite the many specific wiring changes documented between these species, the general phenotypes of orthologous transcriptional regulator knockouts are largely conserved. These observations support the idea that many wiring changes affect the detailed architecture of the circuit, but not its overall output.

Figure 1. Overview of experimental design.

(A) 166 transcriptional regulators were knocked out using homologous recombination. Signature-tagged markers (stars) were included for use in future analyses. (B) The phenotyping screen consisted of a large primary screen and several smaller supplemental screens. (C) Photographs of phenotyping media were processed, archived, and scored using custom Java software. Phenotyping scores were then compiled for each combination of knockout and phenotyping medium and exported to a spreadsheet for further analysis.

Figure 2. Phenotypes of C. albicans transcriptional regulator knockout strains.

(A) Nutrient and base media growth phenotypes. (B) Stress media growth phenotypes. (C) Colony morphology phenotypes. The data shown are predominantly from the primary screen, and in some cases scores from two or more similar media have been condensed to a single column (as described in Text S2). Phenotype scores are represented by the color of the indicated circle, and the intensity of the color represents the strength of the phenotype. White circles represent a phenotype that is indistinguishable from wild-type. Blue and red circles represent either an enhancement or reduction, respectively, of growth (A and B) or colony morphology (wrinkling or invasion; (C) relative to wild-type. Dark red circles (‘severe’ phenotype; growth phenotypes only) indicate that the TRKO strain failed to show any growth, while under the same conditions the wild-type strain exhibited significant growth. A grey ‘X’ represents missing data due to poor growth on the underlying base medium, phenotypic inconsistencies, or absence of testing for the given combination of strain and medium. The size of the circles reflects the strength of the specificity score for the given phenotype. High specificity scores (large diameter) suggest that the TR is a primary regulator of the circuits probed by the medium and the TR (see text and Text S2).

Figure 3. A model of the differences in iron homeostasis regulation between S. cerevisiae and C. albicans.

(A) Each ellipse is colored to reflect whether the indicated gene(s) are believed to play a role in iron homeostasis in C. albicans (blue), S. cerevisiae (burgundy), or both species (green). (B) C. albicans phenotyping data that informed the model are displayed in the bottom panel. These phenotypes include sensitivity to the iron chelator BPS, copper-sensitivity/resistance, and sensitivity to alkaline pH. For each growth condition, only mutants with strong phenotypes are shown (ΔΔmac1 was not scored on alkaline medium due to poor growth on the YEPD base). Each square phenotype panel displays the wild-type phenotype (1× and 5× dilutions) above the mutant phenotype (the image coloration is inverted to better highlight growth). The panels display the timepoint and medium concentration that showed the most dramatic difference between wild-type and mutant. BPS phenotypes were also assayed for S. cerevisiae mutants, and were identical to those of the C. albicans orthologs, with two exceptions: (1) the CCAAT-complex mutants were BPS-sensitive in C. albicans, but not S. cerevisiae, and (2) Δaft1 and Δaft2 were BPS-sensitive in S. cerevisiae but the likely C. albicans ortholog (ΔΔorf19.2272) was not BPS-sensitive. Although the C. albicans ΔΔcsr1 mutant is BPS-sensitive, we have omitted Csr1 because the phenotype is only evident at high BPS concentrations and is likely to reflect a role in zinc homeostasis (see BPS entry in Text S1).

Figure 4. Single-cell–derived colony morphology phenotypes of C. albicans transcriptional regulator knockout strains.

(A) Wild-type. The arrows highlight three morphological features in a single colony: peripheral invasive filaments (white arrow), a smooth colony section (black arrow), and a wrinkled colony section (grey arrow). (B) Mutants lacking both colony wrinkling and peripheral invasion. (C) Mutants with enhanced colony wrinkling on Spider at 30°C. (D) Mutants with complex morphological phenotypes. Colony morphologies of strains that exhibited a colony morphology phenotype on Spider medium in the primary screen were further analyzed using colonies grown from single cells. Colonies were photographed after 3 and 7 days of growth on Spider medium at 30°C and 37°C. Several colonies were grown on each plate, and the colony photographed on day 3 is not necessarily the same colony photographed at day 7.

Figure 5. Conservation of transcriptional regulator phenotypes.

(A) C. albicans and S. cerevisiae genes with 1-to-1 orthology. (B) Genes with 1-to-2 orthology where the two S. cerevisiae orthologs arose from a whole genome duplication event. ‘Conserved’ indicates that one or more phenotypes are shared between the indicated orthologs. ‘Mixed’ indicates that some, but not all, phenotypes were shared. ‘Diverged’ indicates that the phenotype(s) of the orthologs differed. The asterisk (*) indicates that orthology may be more complex than 1-to-1. Bold text for the 1-to-2 orthologs indicates a TR that differed in phenotype from the other two genes. Descriptions of the specific phenotypes compared are provided in the supporting materials (Text S5).