A Brg1-Rme1 circuit in Candida albicans hyphal gene regulation

Abstract

Major Candida albicans virulence traits include its ability to make hyphae, to produce a biofilm, and to damage host cells. These traits depend upon expression of hypha-associated genes. A gene expression comparison among clinical isolates suggested that transcription factor Rme1, established by previous studies to be a positive regulator of chlamydospore formation, may also be a negative regulator of hypha-associated genes. Engineered RME1 overexpression supported this hypothesis, but no relevant rme1Δ/Δ mutant phenotype was detected. We reasoned that Rme1 may function within a specific regulatory pathway. This idea was supported by our finding that an rme1Δ/Δ mutation relieves the need for biofilm regulator Brg1 in biofilm formation. The impact of the rme1Δ/Δ mutation is most prominent under static or "biofilm-like" growth conditions. RNA sequencing (RNA-seq) of cells grown under biofilm-like conditions indicates that Brg1 activates hypha-associated genes indirectly via repression of RME1: hypha-associated gene expression levels are substantially reduced in a brg1Δ/Δ mutant and partially restored in a brg1Δ/Δ rme1Δ/Δ double mutant. An rme1Δ/Δ mutation does not simply bypass Brg1, because iron homeostasis genes depend upon Brg1 regardless of Rme1. Rme1 thus connects Brg1 to the targets relevant to hypha and biofilm formation under biofilm growth conditions.IMPORTANCECandida albicans is a major fungal pathogen of humans, and its ability to grow as a surface-associated biofilm on implanted devices is a common cause of infection. Here, we describe a new regulator of biofilm formation, RME1, whose activity is most prominent under biofilm-like growth conditions.

Keywords: Candida albicans; biofilms; hyphal development; hypoxia; transcriptional regulation.

Fig 1

RME1 and epithelial cell damage. (A) Seventeen clinical isolates were assayed for ability to damage OKF6/TERT-2 oral epithelial cells (21). Results are presented as the % damage relative to that caused by reference strain SC5314. Statistical analysis was conducted using an unpaired t-test, comparing to SC5314. Asterisks denote statistically significant differences. *P value < 0.05, **P value < 0.01, and ***P value < 0.001. (B)Gene expression profiles that correlate with damage ability were identified by clustering RNA-seq data sets of Cravener et al. (30) (C. albicans cells cultured in RPMI + 10% FBS at 37°C for 4 hours with vigorous shaking). Expression data for two high-damage strains (SC5314 and L26) and three low-damage strains (P76067, P78042, and P78048) were used. The analysis yielded 507 correlated/anticorrelated genes (Table S1). The 20 genes shown in the heatmap were prioritized because they encode surface/secreted proteins or transcription factors and had a large expression difference between high- and low-damage strains. The scale is yellow (log₂ fold change from species median of +3) to blue (log₂ fold change from species median of −3). (C) Overexpression mutants or deletion mutants of the 20 prioritized genes were constructed in the high-damage SC5314 and L26 backgrounds and assayed for epithelial cell damage ability. Damage defects were evident in the ece1Δ/Δ and P_TDH3-RME1 mutants. Statistical analysis was conducted using one-way analysis of variance (ANOVA), and asterisks denote statistically significant differences. **P value < 0.01 and ***P value < 0.001. (D) RNA-seq analysis for SC5314 and its P_TDH3-RME1 overexpression derivative (Table S2) yielded 104 genes significantly downregulated in the P_TDH3-RME1 strain. Gene ontology (GO) term enrichment for these genes is shown.

Fig 2

Impact of RME1 overexpression on biofilm formation and filamentation. (A) Wild-type and P_TDH3-RME1 strains in the SC5314 and L26 backgrounds were assayed for biofilm formation ability in RPMI + 10% FBS at 37°C for 24 hours in 96-well plates. Representative apical views are shown. The white scale bars indicate 100 µm. (B) Representative side views are shown for the panel A samples. (C) Biofilm volume was measured for biological triplicates of wild-type and P_TDH3-RME1 strains. (D) Wild-type and P_TDH3-RME1 strains were assayed for hypha formation ability in RPMI at 37°C for 4 hours (planktonic conditions). This medium yielded a clearer phenotypic difference than RPMI + 10% FBS (Fig. S2). The white scale bars indicate 50 µm. (E) Cell lengths were quantified for the panel D samples. At least 4 fields of view and 100 cells were examined. Statistical analysis was conducted using one-way ANOVA, and asterisks denote statistically significant differences. ***P value < 0.001 and ****P value < 0.0001.

Fig 3

Impact of rme1Δ/Δ on biofilm formation and filamentation. (A) Biofilm assays were conducted on the wild type, rme1Δ/Δ and brg1Δ/Δ single mutants, a brg1Δ/Δ rme1Δ/Δ double mutant, and a brg1Δ/Δ rme1Δ/Δ + RME1 complemented strain in the SC5314 reference background. Biofilm formation was assayed in RPMI + 10% FBS at 37°C for 24 hours in 96-well plates. Representative side (above) and apical (below) views are shown. The white scale bars indicate 100 µm. (B) Biofilm volume was measured for biological triplicates of the panel A strains. (C) Filamentation was assayed for the indicated strains in planktonic conditions: RPMI medium, 30 hours, 37°C with vigorous shaking. Representative images are shown. The white scale bars indicate 50 µm in length. (D) Cell length was measured for the panel C strains. (E) Filamentation was assayed for the indicated strains in biofilm-like conditions: RPMI medium, 30 hours, 37°C with sealed lids and no shaking. Representative images are shown. The white scale bars indicate 50 µm in length. (F) Cell length was measured for the panel E strains. For measurements in panels D and E, at least 4 fields of view and 100 cells were examined. Statistical analysis for panels B, (D, and E was conducted using a one-way ANOVA, and asterisks denote statistically significant differences. ****P value < 0.0001.

Fig 4

RNA-seq analysis of the Brg1-Rme1 circuit. RNA-seq analysis was conducted with cells grown under biofilm-like conditions (RPMI + 10% FBS, 37°C, static incubation, 30 hours). Wild-type, rme1Δ/Δ, brg1Δ/Δ, and brg1Δ/Δ rme1Δ/Δ strains from the SC5314 reference background were used. Numerical data are in Table S2. (A) Gene expression changes in the brg1Δ/Δ mutant vs the wild type. (B) Gene expression changes in the rme1Δ/Δ mutant vs the wild type. (C) Gene expression changes in the brg1Δ/Δ rme1Δ/Δ double mutant vs the brg1Δ/Δ single mutant. (D) GO term summary of the 377 genes that are downregulated in the brg1Δ/Δ vs wild type comparison and upregulated in the brg1Δ/Δ rme1Δ/Δ vs brg1Δ/Δ comparison. (E) GO term summary of the 479 genes that are downregulated in the brg1Δ/Δ vs wild type comparison and not upregulated in the brg1Δ/Δ rme1Δ/Δ vs brg1Δ/Δ comparison.

Fig 5

Summary of the Rme1-Brg1 relationship. Our results indicate that Rme1 acts downstream of biofilm master regulator Brg1 to control hypha-associated genes, biofilm formation, and hypha formation. We propose that Brg1 represses RME1, Rme1 represses UME6, and Ume6 activates hypha-associated genes. ChIP-chip data suggest that RME1 is repressed directly by Brg1 (28) and that UME6 is repressed directly by Rme1 (25). Ume6 is known to be an activator of hypha-associated genes (22, 29). Our results also show that Brg1 is also required for full expression of several iron homeostasis genes and that this role is independent of Rme1. ChIP-chip data indicate that activation by Brg1 of ISU1, FET34, and FTR1 may be direct (28) but an unidentified regulator may intercede for Brg1 to stimulate the other iron homeostasis genes. Rme1 acts in parallel with the well-established hyphal repressor Nrg1 (29), and we suggest that Rme1 and Nrg1 act independently, perhaps under distinct environmental conditions. RME1 is induced by hypoxia (37) or, as shown here, under biofilm-like growth conditions, and RME1 expression depends upon the hypoxia regulator Upc2 (38). The other known function of Rme1—activation of chlamydospore formation (25)—also occurs under hypoxic conditions. Thus, the natural function of Rme1 may be exerted mainly during hypoxic growth.