Candida albicans Cannot Acquire Sufficient Ethanolamine from the Host To Support Virulence in the Absence of De Novo Phosphatidylethanolamine Synthesis

Abstract

Candida albicans mutants for phosphatidylserine (PS) synthase (cho1ΔΔ) and PS decarboxylase (psd1ΔΔ psd2ΔΔ) are compromised for virulence in mouse models of systemic infection and oropharyngeal candidiasis (OPC). Both of these enzymes are necessary to synthesize phosphatidylethanolamine (PE) by the de novo pathway, but these mutants are still capable of growth in culture media, as they can import ethanolamine from media to synthesize PE through the Kennedy pathway. Given that the host has ethanolamine in its serum, the exact mechanism by which virulence is lost in these mutants is not clear. There are two competing hypotheses to explain their loss of virulence. (i) PE from the Kennedy pathway cannot substitute for de novo-synthesized PE. (ii) The mutants cannot acquire sufficient ethanolamine from the host to support adequate PE synthesis. These hypotheses can be simultaneously tested if ethanolamine availability is increased for Candida while it is inside the host. We accomplish this by transcomplementation of C. albicans with the Arabidopsis thaliana serine decarboxylase gene (AtSDC), which converts cytoplasmic serine to ethanolamine. Expression of AtSDC in either mutant restores PE synthesis, even in the absence of exogenous ethanolamine. AtSDC also restores virulence to cho1ΔΔ and psd1ΔΔ psd2ΔΔ strains in systemic and OPC infections. Thus, in the absence of de novo PE synthesis, C. albicans cannot acquire sufficient ethanolamine from the host to support virulence. In addition, expression of AtSDC restores PS synthesis in the cho1ΔΔ mutant, which may be due to causing PS decarboxylase to run backwards and convert PE to PS.

Keywords: Candida albicans; mice; oropharyngeal; phosphatidylethanolamine; phosphatidylserine; serine decarboxylase; virulence.

FIG 1

Pathways for synthesizing aminophospholipids in C. albicans. Enzymes are named after their confirmed or predicted homologs in S. cerevisiae. The Kennedy pathway enzymes are shown in blue, and the de novo pathway proteins are shown in green. Arabidopsis thaliana serine decarboxylase (AtSDC) is shown in red. Ser, serine; Eth, ethanolamine; Cho, choline; Eth-P, phosphoethanolamine; Cho-P, phosphocholine; Etn-CDP, cytidyldiphosphate-ethanolamine; Cho-CDP, cytidyldiphosphate-choline; PS, phosphatidylserine; PE, phosphatidylethanolamine; PC, phosphatidylcholine; CDP-DAG, cytidyldiphosphate-diacylglycerol; DAG, diacylglycerol.

FIG 2

cho1ΔΔ and psd1ΔΔ psd2ΔΔ mutants are attenuated for virulence during in vivo and in vitro models of OPC. Mice were inoculated orally with the wild type (WT) or cho1ΔΔ (cho1) or cho1ΔΔ::CHO1 (cho1R) reintegrant strains (A) or the WT or psd1ΔΔ psd2ΔΔ (psd1,2) or psd1ΔΔ psd2ΔΔ::PSD1 (psd1,2R) reintegrant strains (B). After 5 days their tongues were harvested, and the CFU/gram of tissue were measured. *, P < 0.0002 compared to the WT. (C) The strains were coincubated with FaDu cells, and the level of lactase dehydrogenase (LDH) released into the medium was measured as a proxy for FaDu cell damage. P = 0.0007 (**) and P = 0.0005 (***), both compared to the wild type.

FIG 3

Ethanolamine (1 mM) is required to support growth of the cho1ΔΔ and psd1ΔΔ psd2ΔΔ mutants in minimal medium. The cho1ΔΔ (A) and psd1ΔΔ psd2ΔΔ (B) strains were grown overnight in YPD, diluted to an OD₆₀₀ of 0.1, and grown in minimal medium plus various levels of ethanolamine (Eth) at 37°C with shaking, and their growth was measured for up to 48 h. Growth is compared to that of the wild type (WT) with 0 μM ethanolamine.

FIG 4

Expression of AtSDC restores ethanolamine prototrophy in cho1ΔΔ and psd1ΔΔ psd2ΔΔ mutants. (A to C) Mutant and wild-type (WT) strains with or without AtSDC were grown in minimal medium with or without ethanolamine (Etn) at 37°C and a growth curve was plotted. (D) Phospholipids were isolated from each strain grown in minimal medium lacking ethanolamine and then separated by TLC. Lipid spots were visualized with primuline stain under UV light. cho1ΔΔ (cho1), psd1ΔΔ (psd1), and psd1ΔΔ psd2ΔΔ (psd1 psd2) strains were included, all with or without AtSDC.

FIG 5

Expression of AtSDC restores virulence in cho1ΔΔ and psd1ΔΔ psd2ΔΔ strains in a mouse model of systemic candidiasis. (A and B) Mice were infected by tail vein with 5 × 10⁵ cells from each strain, and survival was graphed over time. The numbers of mice per strain are shown beside the strain name in parentheses. The cho1ΔΔ and psd1ΔΔ psd2ΔΔ mutant survival curves were significantly different from that of the wild type (*, P = 0.0004). (C) Mice were infected with 5 × 10⁵ cells by tail vein, and after 5 days they were euthanized, kidneys were removed, and CFU/gram of kidney were measured by plating. *, P < 0.0001 compared to WT or strains expressing AtSDC; **, P = 0.0031 compared to the cho1ΔΔ mutant. cho1 (cho1ΔΔ), psd1 (psd1ΔΔ), psd1,2 (psd1ΔΔ psd2ΔΔ), and SDC (AtSDC) mutant strains were studied.

FIG 6

Expression of AtSDC restores virulence to the cho1ΔΔ or psd1ΔΔ psd2ΔΔ strain for in vivo and in vitro models of OPC. (A) An OPC infection model was performed with mice, and after 5 days they were sacrificed, tongues were removed, and CFU/gram of tongue were measured. *, P < 0.01 compared to the WT. **, P < 0.01 compared to the cho1ΔΔ-AtSDC strain. ***, P < 0.01 compared to the psd1ΔΔ psd2ΔΔ strain. ****, P < 0.01 compared to the psd1ΔΔ psd2ΔΔ-AtSDC strain. (B) The wild-type (WT), cho1ΔΔ, and psd1ΔΔ psd2ΔΔ strains were compared to their respective AtSDC transformants for their abilities to damage FaDu cells based on release of lactase dehydrogenase (LDH). P = 0.0157 (¥) and P = 0.0163 (å), both compared to the WT; ¥¥, P = 0.0071 compared to the cho1ΔΔ strain; åå, P = 0.0201 compared to the psd1ΔΔ psd2ΔΔ strains. cho1 (cho1ΔΔ), psd1 (psd1ΔΔ), psd1,2 (psd1ΔΔ psd2ΔΔ), and SDC (AtSDC) strains were used.

FIG 7

Expression of AtSDC suppresses β(1,3)-glucan unmasking in the cho1ΔΔ strain. (A) Cells were stained with anti-β(1,3)-glucan primary and phycoerythrin-conjugated secondary antibodies, and 100,000 cells were measured by flow cytometry for each replicate sample. Experiments were performed with three replicates per strain. (B) Cells were stained with anti-β(1,3)-glucan primary and Cy3-conjugated secondary antibodies and viewed by immunofluorescence microscopy. *, P < 0.0001; **, P = 0.0067.

FIG 8

Expression of AtSDC in the cho1ΔΔ mutant leads to production of PS. (A) Strains were grown overnight in YPD, and phospholipids were extracted and run on TLC to determine if PS was being synthesized. Lipid spots were compared to purified standards, which are marked on the TLC. The PS spot is marked with a black arrow in the cho1ΔΔ and cho1ΔΔ AtSDC lanes. (B) Cells were grown overnight in YPD and diluted to 10⁵/ml in fresh YPD, and 100 μl of the culture was added to wells containing YPD and PapA or the methanol-water solvent control at the concentrations shown and then incubated overnight without shaking at 30°C.