Reinforcement amid genetic diversity in the Candida albicans biofilm regulatory network

Abstract

Biofilms of the fungal pathogen Candida albicans include abundant long filaments called hyphae. These cells express hypha-associated genes, which specify diverse virulence functions including surface adhesins that ensure biofilm integrity. Biofilm formation, virulence, and hypha-associated gene expression all depend upon the transcription factor Efg1. This transcription factor has been characterized extensively in the C. albicans type strain SC5314 and derivatives, but only recently has its function been explored in other clinical isolates. Here we define a principal set of Efg1-responsive genes whose expression is significantly altered by an efg1Δ/Δ mutation across 17 clinical isolates. This principal gene set includes 68 direct Efg1 targets, whose 5' regions are bound by Efg1 in five clinical isolates, and 42 indirect Efg1 targets, whose 5' regions are not detectably bound by Efg1. Three direct Efg1 target genes encode transcription factors-BRG1, UME6, and WOR3 -whose increased expression in an efg1Δ/Δ mutant restores expression of multiple indirect and direct principal targets, as well as biofilm formation ability. Although BRG1 and UME6 are well known positive regulators of hypha-associated genes and biofilm formation, WOR3 is best known as an antagonist of Efg1 in the sexual mating pathway. We confirm the positive role of WOR3 in biofilm formation with the finding that a wor3Δ/Δ mutation impairs biofilm formation in vitro and in an in vivo biofilm model. Positive control of Efg1 direct target genes by other Efg1 direct target genes-BRG1, UME6, and WOR3 -may buffer principal Efg1-responsive gene expression against the impact of genetic variation in the C. albicans species.

Fig 1. Hypha formation ability.

Confocal images of clinical isolate WT (Left) and efg1Δ/Δ (Right) planktonic hyphal formation in liquid RPMI+10% serum at 37°C for 4hr. Samples were fixed in 4% formaldehyde in 1X PBS and stained with calcofluor white. White scale bar indicates 25 μm. Clade number is designated in parentheses.

Fig 2. Biofilm formation ability.

Clinical isolate WT and efg1Δ/Δ strains were assayed for biofilm formation ability in RPMI+10% FBS at 37°C for 24 hrs using the “Silicone substrate–confocal microscopy” method. Specimens were imaged using confocal microscopy and Alexafluor594-conjugated wheat germ agglutinin. Above are sideview images for both WT and efg1Δ/Δ. Scale bar indicates 250 μm.

Fig 3. Relationship between the principal EFG1 targets and the clinical isolate-derived hypha-associated gene set.

Well studied hypha-associated genes including ALS3, BRG1, ECE1, IHD1, HWP1, HGC1, HYR1, UME6, and SAP5 are included in both gene sets.

Fig 4

Panel A. Relative expression levels of selected 17-strain principal direct and indirect targets. Direct targets are genes whose 5’ regions are bound by Efg1; indirect targets are genes whose 5’ regions are not detectably bound by Efg1 [–13]. RNA expression levels were calculated relative to the efg1Δ/Δ for each TDH3-TF strain under strong hypha-inducing conditions (RPMI+10% FBS at 37°C). Selected genes were those with a conserved response to a TDH3-TF mutation in the SC5314, P87, and P75010 backgrounds. Shown here are the relative expression levels for these genes in SC5314 background. Panel B. Expression levels of BRG1, UME6, and WOR3 relative to WT in both the efg1Δ/Δ mutant and in each efg1Δ/Δ TDH3-TF strain in the SC5314 isolate background. P-values for significance are given in S6 Table.

Fig 5. Rescue of efg1Δ/Δ hypha formation.

Fluorescence image of SC5314, P87, and P75010 wild-type (WT), efg1Δ/Δ, efg1Δ/Δ TDH3-BRG1, efg1Δ/Δ TDH3-UME6, and efg1Δ/Δ TDH3-WOR3 strain planktonic hypha formation in RPMI+10%FBS 37°C for 4hrs. Samples were fixed in 4% formaldehyde in 1X PBS and stained with Calcofluor-white. Scale bar indicates 80 μm.

Fig 6. Rescue of efg1Δ/Δ biofilm formation.

Restoration of biofilm formation by Efg1 target TF gene expression. Biofilm formation was assayed in RPMI+10% FBS at 37°C for 24 hrs in a 96 well plate for WT, efg1Δ/Δ, and efg1Δ/Δ TDH3-TF strains in the SC5314 background using the “96 well plate—XTT reduction” method. After growth, wells were washed with PBS and 100μL of 1mg/mL XTT in PBS plus 0.025 μL menadione in acetone were added to each well and incubated for 1 hour at 37°C. Absorbance at 492nm was then measured to quantify biofilm formation. Three technical replicates for each strain were compared by one-way ANOVA to the efg1Δ/Δ strain. ** denotes p = 0.0080 and **** denotes p < 0.0001.

Fig 7. Imaging of TDH3-WOR3 biofilms in the SC5314 background.

Biofilms were imaged after 24 hrs at 37°C in RPMI+10% FBS for WT, efg1Δ/Δ, and efg1Δ/Δ TDH3-WOR3 SC5314-derived strains using the “Silicone substrate–confocal microscopy” method. Maximum intensity apical projections of the entire biofilm (Left column) and the top half of the biofilm (Middle column) were generated for each strain. Side view projections (Right column) were also generated. Scale bar indicates 60 μm.

Fig 8. Biofilm formation dependence on WOR3.

Panels A and B. Biofilm formation ability for WT, wor3Δ/Δ, and complemented strains was assessed at 37°C for 24 hrs in RPMI, RPMI+10% FBS, and YPD+10% FBS in both SC5314 (Panel A) and P87 (Panel B) backgrounds using the “96 well plate–Fluorescence microscopy” method. Side-view images of two independent mutant strains were tested for each clinical isolate background for both the wor3Δ/Δ and complemented strains. Scale bars are equal to 285 μm. Panel C. SC5314-derived wild-type, mutant, and complemented strains were tested for biofilm formation in a rat venous catheter infection model [21]. C. albicans cell counts per catheter were determined at 48 hr post-infection. Wild-type and mutant strains differed with p-value = 0.02.

Fig 9. Biofilm formation in brg1Δ/Δ and brg1Δ/Δ TDH3-WOR3 strains.

Wild-type (WT), brg1Δ/Δ, and brg1Δ/Δ TDH3-WOR3 strains in the indicated clinical isolate backgrounds were assayed for biofilm formation in RPMI+FBS, 37°C for 24hrs using the “Silicone substrate–confocal microscopy” method. Biofilm sideview images are above. Scale bar indicates 145 μm.

Fig 10. Size of EFG1 responsive gene set.

Panel A. Regression showing the decay of the number of principal EFG1-responsive genes with the addition of multiple strains. Each box plot represents all possible number of principal response genes (Y-axis) among a given number “n” of strains (X-axis). Panel B. Tabulation of mean number of Efg1-responsive genes as a function of number of strains tested with fold changes of 1 log₂ (as used throughout this study; described by y = 701.93x^-0.656 and R² = 0.9894), or with higher thresholds of 2 log₂ (described by y = 310.83x^-0.717 and R² = 0.9974) or 3 log₂ (described by y = 180.96x^-0.844 and R² = 0.999).

Fig 11. Reinforcement and extension of the Efg1 principal network by BRG1, UME6, and WOR3.

Network of Efg1 direct target transcription factor interactions with other principal Efg1 direct targets (green) and Efg1 indirect targets (blue). Arrows represent conserved activation by the TDH3-TF allele in the SC5314, P87, and P75010 efg1Δ/Δ strain backgrounds. Blunt-end connections represent repression by the TDH3-TF allele in all three clinical isolate efg1Δ/Δ backgrounds. Bolded arrows represent conserved activation among all three clinical isolate efg1Δ/Δ backgrounds of a principal Efg1 direct target TF by one of the other principal Efg1 direct target TFs. Network relationships were deduced from one set of growth conditions and three diverse C. albicans strain backgrounds.