Role of Acetyl Coenzyme A Synthesis and Breakdown in Alternative Carbon Source Utilization in Candida albicans

Abstract

Acetyl coenzyme A (acetyl-CoA) is the central intermediate of the pathways required to metabolize nonfermentable carbon sources. Three such pathways, i.e., gluconeogenesis, the glyoxylate cycle, and β-oxidation, are required for full virulence in the fungal pathogen Candida albicans. These processes are compartmentalized in the cytosol, mitochondria, and peroxosomes, necessitating transport of intermediates across intracellular membranes. Acetyl-CoA is trafficked in the form of acetate by the carnitine shuttle, and we hypothesized that the enzymes that convert acetyl-CoA to/from acetate, i.e., acetyl-CoA hydrolase (ACH1) and acetyl-CoA synthetase (ACS1 and ACS2), would regulate alternative carbon utilization and virulence. We show that C. albicans strains depleted for ACS2 are unviable in the presence of most carbon sources, including glucose, acetate, and ethanol; these strains metabolize only fatty acids and glycerol, a substantially more severe phenotype than that of Saccharomyces cerevisiae acs2 mutants. In contrast, deletion of ACS1 confers no phenotype, though it is highly induced in the presence of fatty acids, perhaps explaining why acs2 mutants can utilize fatty acids. Strains lacking ACH1 have a mild growth defect on some carbon sources but are fully virulent in a mouse model of disseminated candidiasis. Both ACH1 and ACS2 complement mutations in their S. cerevisiae homolog. Together, these results show that acetyl-CoA metabolism and transport are critical for growth of C. albicans on a wide variety of nutrients. Furthermore, the phenotypic differences between mutations in these highly conserved genes in S. cerevisiae and C. albicans support recent findings that significant functional divergence exists even in fundamental metabolic pathways between these related yeasts.

Candida albicans is the most important fungal pathogen of humans. Its incidence has risen sharply in recent years, and it is now first in associated mortality among common nosocomial infections (41). C. albicans is normally a benign commensal of the mammalian gastrointestinal tract, and the primary risk factor for the disseminated disease is a defective innate immune system. Thus, there has been significant interest in understanding the molecular interactions between C. albicans and phagocytes of the innate immune system. These studies have revealed very complex interactions between phagocytes and C. albicans at the cell surface (25). No less complex are the changes in Candida physiology during phagocytosis: microarray analysis of C. albicans gene expression during phagocytosis by macrophages or neutrophils reveals extensive metabolic remodeling that reflects adaptation to nutrient-poor environments (10, 17, 32). We and others have shown that some of the alternative carbon assimilation pathways that are upregulated inside macrophages are required for full virulence, including gluconeogenesis, the glyoxylate cycle, and β-oxidation of fatty acids (2, 18, 27, 29), validating the genomic approach to gene discovery in this system. The glyoxylate cycle, in particular, is also necessary for virulence in a diverse array of pathogens, including Mycobacterium tuberculosis and Magnaporthe grisea (19, 21, 39).

The pathways discussed above converge on acetyl coenzyme A (acetyl-CoA) as a central intermediate in carbon metabolism. Phagocytosed cells also upregulate other aspects of acetyl-CoA homeostasis, including carnitine acetyltransferases for intracellular transport and alcohol and aldehyde dehydrogenases (17, 42). Also induced are genes encoding acetyl-CoA synthetases (ACS1 and ACS2) and acetyl-CoA hydrolase (ACH1), catalyzing the conversion between acetyl-CoA and acetate; only the latter form can be transported across intracellular membranes, via the carnitine shuttle. Saccharomyces cerevisiae ach1Δ strains grow poorly on some alternative carbon sources and have been reported to have defects in pseudohyphal development (6, 16). ACS1 is required for growth on acetate, but not ethanol or glucose, in S. cerevisiae (8) and is necessary for glyoxylate cycle induction in Yarrowia lipolytica (13). Conversely, acs2Δ mutants are unviable on glucose-containing media but will grow on ethanol or acetate; the acs1Δ/acs2Δ double mutant is unviable (37). We and others have shown that the C. albicans carnitine acetyltransferases of the carnitine shuttle are important for growth on nonfermentable carbon sources (28, 34a, 42), but little else is known about the in vitro or in vivo roles of acetyl-CoA metabolism in C. albicans.

In this study we demonstrate that acetate and acetyl-CoA metabolism play a central role in C. albicans growth on both glucose and nonglucose carbon sources. We constructed C. albicans mutants lacking ACH1, ACS1, or ACS2. Using a regulated allele of ACS2, we show that ACS2 is required for growth not only on glucose but on many nonglucose carbon sources. ach1Δ mutants have mild growth defects on some alternative carbon sources, while acs1Δ strains have no apparent phenotype, and neither mutant is attenuated for virulence in a mouse model of disseminated candidiasis. The phenotypes of the C. albicans acs2Δ mutant are more pronounced than those of the S. cerevisiae homolog, indicating some divergence of function, as has been seen previously (27, 29, 42), though CaACS2 fully complements an Scacs2Δ mutation. This work continues our efforts to define critical carbon metabolic processes in the important pathogen C. albicans.

MATERIALS AND METHODS

Strains and media.

The C. albicans strains used in this study are listed in Table 1. Standard yeast media were used (34), including YNB (0.17% yeast nitrogen base, 0.5% ammonium sulfate) and YPD (1% yeast extract, 2% peptone, 2% dextrose). YNB was supplemented with carbon sources as indicated at a final concentration of 2%. 5-Fluoroorotic acid medium (YNB plus 2% glucose, 0.2 mM uracil, 0.2 mM uridine, and 0.1% 5-fluoroorotic acid) (3) was also used. Candida albicans strains were transformed either by electroporation (30) or by use of a modified lithium acetate method (5).

TABLE 1.

Fungal strains

Species and strain	Genotype	Source or reference
C. albicans
SC5314 (wild type)	Prototroph	9
RM1000 (ura3 his1)	ura3::λimm434/ura3::λimm434 his1::hisG/his1::hisG	24
ACC16 (ach1Δ/Δ [no. 1])	ura3::λimm434/ura3::λimm434 his1::hisG/his1::hisG ach1Δ::hisG/ach1Δ::HIS1 rps10::CIp10-URA3/RPS10	This study
ACC17 (ach1Δ/Δ+ACH1)	ura3::λimm434/ura3::λimm434 his1::hisG/his1::hisG ach1Δ::hisG/ach1Δ::HIS1 rps10::CIp10-URA3-ACH1/RPS10	This study
ACC20 (ach1Δ/Δ [no. 2])	ura3::λimm434/ura3::λimm434 his1::hisG/his1::hisG ach1Δ::hisG/ach1Δ::HIS1 rps10::CIp10-URA3/RPS10	This study
ACC21 (ach1Δ/Δ+ACH1)	ura3::λimm434/ura3::λimm434 his1::hisG/his1::hisG ach1Δ::hisG/ach1Δ::HIS1 rps10::CIp10-URA3-ACH1/RPS10	This study
ACC9 (acs1Δ/Δ [no. 1])	ura3::λimm434/ura3::λimm434 his1::hisG/his1::hisG acs1Δ::hisG/acs1Δ::HIS1 rps10::CIp10-URA3/RPS10	This study
ACC10 (acs1Δ/Δ+ACS1)	ura3::λimm434/ura3::λimm434 his1::hisG/his1::hisG acs1Δ::hisG/acs1Δ::HIS1 rps10::CIp10-URA3-ACS1/RPS10	This study
ACC13 (acs1Δ/Δ [no. 2])	ura3::λimm434/ura3::λimm434 his1::hisG/his1::hisG acs1Δ::hisG/acs1Δ::HIS1 rps10::CIp10-URA3/RPS10	This study
ACC14 (acs1Δ/Δ+ACS1)	ura3::λimm434/ura3::λimm434 his1::hisG/his1::hisG acs1Δ::hisG/acs1Δ::HIS1 rps10::CIp10-URA3-ACS1/RPS10	This study
ACC24 (acs2Δ/ACS2-MET3p [no. 1])	ura3::λimm434/ura3::λimm434 his1::hisG/his1::hisG acs2Δ::HIS1/URA3-MET3p-ACS2	This study
ACC25 (acs2Δ/ACS2-MET3p [no. 2])	ura3::λimm434/ura3::λimm434 his1::hisG/his1::hisG acs2Δ::HIS1/URA3-MET3p-ACS2	This study
MRC10 (icl1Δ/Δ)	ura3::λimm434/ura3::λimm434 icl1Δ::hisG/icl1Δ::hisG RPS10/rps10::URA3	29
MRC11 (icl1Δ/Δ+ICL1)	ura3::λimm434/ura3::λimm434 icl1Δ::hisG/icl1Δ::hisG RPS10/rps10::ICL1-URA3	29
S. cerevisiae
BY4741	his3 leu2 met1 ura3 MATa	4
BY4741 ach1Δ (library)	his3 leu2 met1 ura3 MATaach1Δ::G418	40
BY4741 acs1Δ (library)	his3 leu2 met1 ura3 MATaacs1Δ::G418	40
BY4741 acs2Δ (ACY1)	his3 leu2 met1 ura3 MATaacs2Δ::G418	This study

The Saccharomyces cerevisiae strains used in this study are listed in Table 1 and are derived from strain BY4741(4). Strains were transformed using a modified lithium acetate method (33).

Candida albicans mutant construction.

All C. albicans mutants were constructed in RM1000 (His⁻ Ura⁻). For the ach1Δ and acs1Δ mutants, we constructed two disruption constructs, one using the hisG-URA3-hisG cassette (9) and the other using the Candida dubliniensis HIS1 gene (26). Briefly, 300 bp of the 5′ and 3′ untranslated regions flanking the gene were amplified by overlapping PCR into one fragment, with a BamHI site in the middle. This was cloned as a HindIII/SacII fragment into pBSKII+. The hisG-URA3-hisG cassette was removed from pCUB-6 (9) by a BamHI/BglII/PvuII digest and then ligated into the fusion plasmids cut with BamHI. In parallel, the CdHIS1 gene was amplified as a BamHI cassette and inserted into the BamHI site in the fusion plasmids. The resulting disruption plasmids are listed in Table 2. Complementing constructs were made by cloning PCR-amplified genes into plasmid CIp10 (22) (Table 2).

TABLE 2.

Plasmids

Plasmid	Description	Source or reference
pFA6KanMX	Contains G418 resistance marker	38
pCaDis	MET3 promoter	7
pCIp10	RPS10 integrating plasmid	22
pCUB6	hisG-URA3-hisG	9
pAC1	pCIp10-CaACS1	This study
pAC3	pCIp10-CaACH1	This study
pAC7	pBSK-acs1Δ::hisG-URA3-hisG	This study
pAC12	pBSK-ach1Δ::hisG-URA3-hisG	This study
pAC24	pBSK-ach1Δ::HIS1	This study
pAC25	pBSK-acs1Δ::HIS1	This study
pAC26	pBSK-acs2Δ::HIS1	This study
pAC48	pCaDis-ACS2-MET3p	This study
p415-GPD	pRS415-GPD1p LEU2 2μ	20
pAC60	pRS415-GPD1p-ScACH1	This study
pAC61	pRS415-GPD1p-ScACS1	This study
pAC62	pRS415-GPD1p-ScACS2	This study
pAC63	pRS415-GPD1p-CaACH1	This study
pAC64	pRS415-GPD1p-CaACS1	This study
pAC65	pRS415-GPD1p-CaACS2	This study

The first allele of ACS2 was knocked out using a HIS1 disruption construct as described above (pAC26). Correct heterozygous mutants were confirmed by PCR. The second allele of ACS2 was put under the control of the MET3 repressible promoter. The first 971 bp of ACS2 were cloned into pCaDis (7). An overlapping PCR strategy was used to introduce a unique EcoRV cut site into the middle of the fragment. The resulting plasmid (pAC48) was linearized with EcoRV and transformed into the ACS2 heterozygote mutant. Correct recombinants were confirmed by PCR.

Saccharomyces cerevisiae mutant construction.

The S. cerevisiae strains used in this study are listed in Table 1 and are derived from BY4741. The ach1Δ and acs1Δ mutants were obtained from the haploid deletion collection made by the Genome Deletion Project (40). The acs2Δ mutant is not found in this library because it is unviable in standard glucose media and so was made using a PCR-mediated disruption protocol (38). BY4741 was transformed with the PCR product, and the cells were grown for 4 h at 30°C in YP-ethanol (2%) and then plated to YP-ethanol plus 200 μg/ml G418. Potentially correct recombinants were incubated on YP-ethanol and YP-glucose agar plates. Mutants that failed to grow on YP-glucose, the expected phenotype (37), were checked by PCR for the presence of the disruption construct.

Saccharomyces cerevisiae plasmid construction.

S. cerevisiae overexpression constructs are listed in Table 2. S. cerevisiae ACH1, ACS1, and ACS2 were amplified by PCR and cloned into plasmid p415-GPD (20) between SpeI and XhoI sites to produce the plasmids pAC60 (p415-GPD-ScACH1), pAC61 (p415-GPD-ScACS1), and pAC62 (p415-GPD-ScACS2), respectively. For the heterologous complementation experiments, C. albicans ACH1, ACS1, and ACS2 were cloned into the p415-GPD vector. CaACS1 and CaACS2 were inserted between SpeI and XhoI sites to produce plasmids pAC64 (p415-GPD-CaACS1) and pAC65 (p415-GPD-CaACS2), while CaACH1 was inserted between SmaI and PstI sites to produce plasmid pAC63 (p415-GPD-CaACH1).

In vitro growth assays.

For spot dilution assays, strains were grown in liquid YNB-glucose at 30°C to mid-log phase, washed twice with water, and transferred to a 96-well plate at an optical density at 600 nm (OD₆₀₀) of 1.0. Cells were then serially diluted fivefold and spotted using a multipipetter to solid YNB medium containing 2% glucose, potassium acetate, ethanol, lactate, glycerol, citrate, or oleate and incubated at 30°C for 3 to 7 days as indicated in the figure legends.

For liquid growth assays, strains were grown in YNB-glucose at 30°C overnight. The next day the cells were collected by centrifugation, washed twice with water, and diluted into fresh YNB medium containing the appropriate carbon source at an OD₆₀₀ of 0.05.

Northern analysis.

Candida albicans SC5314 cells were grown overnight in YNB-glucose, collected by centrifugation, and washed twice with water. Cells were then added to fresh YNB medium containing 2% glucose, potassium acetate, ethanol, glycerol, lactate, citrate, or oleate, grown for 1 h at 30°C, collected by centrifugation, and then quickly frozen on dry ice-ethanol. As controls, the ach1Δ, acs1Δ, and acs2Δ/MET3p-ACS2 mutants were grown in YNB-glucose in the same manner as for the wild type (with addition of 5 mM methionine and 5 mM cysteine to the acs2Δ/MET3p-ACS2 culture). RNA was isolated from the cells using the hot acid-phenol method (1). A total of 15 ng RNA per sample was run on a 1% MOPS (morpholinepropanesulfonic acid)-agarose gel and transferred to a nylon membrane. Probes of ∼400 bp specific for each gene were amplified by PCR and labeled with [³²P]dCTP with the RadPrime DNA labeling kit (Invitrogen) and then purified with Roche Quick Spin columns. The blots were incubated in prehybridization buffer (5× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 50% formamide, 5× Denhardt's solution, 0.1% sodium dodecyl sulfate [SDS] [1], and 100 μg/ml single-stranded DNA) for 2 h at 42°C. Blots were then transferred to fresh prehybridization buffer containing the appropriate labeled probe and incubated at 42°C overnight. The next morning the blots were washed and exposed to film for autoradiography. Images were processed using a Storm PhosphorImager. Blots were then stripped by incubation in stripping solution (0.2% SDS in Tris-EDTA) for 30 min at 65°C. Blots were rehybridized with a probe specific for 18S rRNA as a control.

Reverse transcription-PCR (RT-PCR).

C. albicans SC5314 cells and acs2Δ/MET3p-ACS2 cells were grown overnight in YNB-glucose, collected by centrifugation, and washed twice with water. Cells were then added to fresh YNB medium containing glucose alone or glucose plus 5 mM methionine and cysteine. The cells were grown for 1, 2, or 5 h and then collected by centrifugation and quickly frozen on dry ice-ethanol. RNA was extracted using hot acid-phenol (1). cDNA was then made using reverse transcriptase (Invitrogen), and 300-bp internal fragments of either ACS2 or ACT1 were PCR amplified from 10-fold serial dilutions of cDNA.

Immunoblot analysis of histone H3 and histone H4 acetylation.

SC5314, acs1Δ, and asc2Δ/MET3p-ACS2 cells were grown in 25 ml YNB with 2% glucose overnight, collected by centrifugation, washed twice with phosphate-buffered saline (PBS) and diluted 1:10 in fresh YNB with 2% glucose to an OD of 1.0; then 50 mg/ml cysteine-methionine was added to the cells and 10-ml aliquots were taken at 0, 10, 20, 30, 60, and 120 min and frozen immediately at −80°C. Cells were resuspended in histone extraction buffer (12) supplemented with protease inhibitors, followed by glass bead lysis. Extracts were centrifuged at 7,000 × g for 7 min to isolate the cytosolic fraction. Equal amounts of total protein (10 μg) were separated using 10% SDS-polyacrylamide gel electrophoresis, transferred to polyvinylidene difluoride membranes, and probed with either anti-acetyl-histone H3 polyclonal antibody (AR-0143; LP Bio) or anti-acetyl-Lys5 histone H4 polyclonal antibody (AR-0119; LP Bio). The blots were developed according to the manufacturer's recommendations using the ECL kit (Pierce Biotechnology). After the anti-acetyl-H3 and -H4 antibodies were stripped, the membranes were reprobed with anti-histone H3 (AR-0144; LP Bio) or anti-histone H4 (ab10158; Abcam Inc.) polyclonal antibodies, respectively, and developed as described earlier.

In vivo virulence assays.

Mouse virulence assays were performed as previously described (29) using adult (21- to 25-g) female outbred ICR mice (Harlan). C. albicans strains were grown in YPD to mid-log phase and collected by centrifugation. Cells were washed twice with PBS and resuspended in PBS, and 10⁶ yeast form cells were injected via the tail vein in 0.1 ml PBS. Ten mice were infected per group. Mice were monitored and were euthanized when moribund. Survival data were analyzed using Prism5 (Graphpad Software) and the log rank test. Statistical significance was defined as a P value of less than 0.05. All animal assays were conducted in accordance with protocols approved by the University of Texas Health Science Center Animal Welfare Committee.

RESULTS

Mutant strain construction.

Candida albicans has close homologs of the S. cerevisiae ACH/ACS genes. ACH1 (orf19.3171) encodes a putative acetyl-CoA hydrolase which catalyzes the hydrolysis of acetyl-CoA to free acetate and CoA and is required for acetate utilization in S. cerevisiae (6, 14, 16). ACS1 encodes a 675-residue acetyl-CoA synthetase, one of two C. albicans enzymes responsible for catalyzing the formation of acetyl-CoA from free acetate and CoA. ACS1 (orf19.1743) is 64% identical to the S. cerevisiae ACS1 gene product at the amino acid level. The C. albicans ACS2 gene (orf19.1064) encodes the second putative acetyl-CoA synthetase, a 671-amino-acid protein with 68% identity to the S. cerevisiae Acs2 enzyme. Each of these genes is induced in macrophage-phagocytosed cells (ACS1, 3.3-fold; ACS2, 2.8-fold; ACH1, 6.1-fold) (17), and we were thus interested in their in vitro and in vivo functions.

We constructed mutants with mutations in each gene, as shown in Fig. 1. Homozygous deletions of ACH1 and ACS1 were constructed, along with complemented strains in which the wild-type gene was reintegrated at the RPS10 locus using plasmid CIp10 (Fig. 1) (see Materials and Methods). Attempts to construct a homozygous acs2Δ/Δ strain were unsuccessful; this was expected since the S. cerevisiae gene is essential for growth in glucose (37). Instead, we constructed a strain in which one allele was deleted and the other was under the control of the methionine-repressible MET3 promoter (7) (see Materials and Methods). For each gene, multiple independent deletion strains were generated and analyzed.

FIG. 1.

Construction of C. albicans mutant strains. (A and B) Both alleles of ACH1 (A) and ACS1 (B) were disrupted in RM1000 sequentially by replacing one allele with C. dubliniensis HIS1 and the other with hisG-URA3-hisG. After selection on 5-fluoroorotic acid medium, URA3 was reintroduced at the RPS10 locus using plasmid CIp10 either unlinked (mutant strains) or linked (complemented strains) to a wild-type copy of the gene. (C) The first allele of ACS2 was replaced with the HIS1 marker, and the MET3 promoter linked to URA3 was integrated upstream of the second allele, replacing the native ACS2 promoter.

ACS2 is essential for growth in most carbon sources.

The addition of methionine and cysteine (5 mM each) strongly represses transcription from the MET3 promoter (7). RT-PCR analysis of our acs2Δ/MET3p-ACS2 strain indicated that ACS2 message was undetectable 1 hour after addition of Met-Cys to a logarithmically growing culture (Fig. 2A). ACT1, a control message, was unaffected. This treatment also resulted in the rapid cessation of growth: the acs2Δ/MET3p-ACS2 strain stopped growing within 1 hour of addition of Met-Cys to cultures in YPD, while the wild-type strain continued exponential growth (Fig. 2B). These data indicate that repression of the acs2Δ/MET3p-ACS2 construct effectively depletes ACS2 mRNA and, as a result, inhibits growth.

FIG. 2.

ACS2 depletion inhibits growth in glucose. (A) The wild-type (SC5314) and acs2Δ/MET3p-ACS2 (ACC24) strains were grown for 1 h in glucose with or without 5 mM cysteine plus 5 mM methionine. RNA was subjected to semiquantitative RT-PCR using primers designed to amplify ∼300-bp fragments of either ACS2 (left) or ACT1 (right) as a control. (B) Methionine and cysteine (5 mM each) were added to cells growing logarithmically in YNB-glucose (arrow). Growth was monitored by OD₆₀₀. Error bars indicate standard deviations.

S. cerevisiae acs2Δ mutants cannot grow on media containing glucose (37); because of this, ACS2 was reported to be an essential gene by the genome-wide functional profiling project (11, 40). However, earlier work showed that Scacs2Δ strains could grow in the presence of ethanol or acetate (37). We tested the C. albicans acs2Δ/MET3p-ACS2 strain by plating serial dilutions on several different carbon compounds in the presence and absence of 5 mM methionine and cysteine (Fig. 3). As expected from Fig. 2B and previous work with S. cerevisiae, this strain failed to grow when glucose was the sole carbon source. Surprisingly, however, depletion of acs2 also blocked utilization of acetate, ethanol, lactate, and citrate. In each case, growth was similar to wild type when methionine and cysteine were omitted from the medium (Fig. 3). Repression of the acs2Δ/MET3p-ACS2 construct did not affect growth when either glycerol or oleate was the sole carbon source, indicating that ACS2 is not absolutely essential in C. albicans but is required for growth on a wider range of carbon sources than is the S. cerevisiae homolog.

FIG. 3.

ACS2 is required for growth on diverse carbon sources. Serial 1:5 dilutions of the wild-type (SC5314; upper strain in each pair) or acs2Δ/MET3p-ACS2 (ACC24; lower strain) strain were spotted to YNB plates containing the indicated carbon sources without (left) and with (right) 5 mM methionine-cysteine. Growth was observed after 3 days, except for citrate plates (5 days), at 30°C.

ACH1 and ACS1 mutants have limited in vitro phenotypes.

We sought to test the homozygous ach1Δ/Δ and acs1Δ/Δ mutants for carbon utilization phenotypes using a spot dilution assay. Growth of the acs1Δ mutant on glucose, potassium acetate, or ethanol (Fig. 4) or on oleate (data not shown) was unchanged compared to that of the wild type. In contrast, loss of ACH1 conferred a mild retardation of growth in the presence of ethanol and acetate, phenotypes that were complemented by the restoration of a single copy of ACH1 (Fig. 4). A mutant lacking isocitrate lyase (ICL1), a strain with well-documented growth defects on ethanol and acetate (27, 29), was used as a control.

FIG. 4.

ach1Δ/Δ mutants have mild growth defects on nonglucose carbon sources. Serial 1:5 dilutions of wild-type (SC5314), ach1Δ/Δ (ACC16), and acs1Δ/Δ (ACC9) strains and the complemented strains were spotted to YNB plates containing the indicated carbon sources. Growth was observed after 3 days at 30°C.

In vitro expression analysis of ACH1, ACS1, and ACS2 on different carbon sources.

We examined the expression of ACH1, ACS1, and ACS2 during growth in different carbon sources. RNA was obtained from wild-type and mutant strains incubated for 1 h at 30°C in fresh YNB containing different carbon sources (see Materials and Methods). As shown in Fig. 5A, ACH1 was present at low levels in glucose-containing medium and was highly induced when acetate, ethanol, or glycerol was the carbon source. This agrees with previous findings that S. cerevisiae ACH1 is glucose repressed (15). Interestingly, ACH1 expression was almost undetectable when lactate was the sole carbon source. ACS1 and ACS2 had largely opposite expression patterns (Fig. 5B and C). In glucose, acetate, ethanol, glycerol, and citrate, ACS2 is the more highly expressed of the two. ACS1 was most highly expressed in oleate, while ACS2 expression appeared lower, perhaps explaining why the acs2Δ/Δ mutant is able to grow in the presence of oleate. Probe specificity was confirmed using the acs1Δ/Δ and acs2Δ/MET3p-ACS2 strains.

FIG. 5.

Carbon source-specific gene expression. Expression of ACH1 (A), ACS1 (B), and ACS2 (C) in different carbon conditions was determined by Northern analysis. RNA was prepared from cells grown for 1 hour in glucose (G), potassium acetate (A), ethanol (E), sodium citrate (C), oleate (O), or glycerol (Y) and probed with specific DNA fragments for each gene (top row in each panel). rRNA abundance by ethidium bromide staining (middle row) or Northern analysis with an 18S rRNA probe (bottom row) was used as a loading control. The “Δ” in each panel indicates a probe specificity control, which was RNA prepared from the cognate deletion strain (for ACH1 or ACS1) or from the MET3p-ACS2/acs2Δ strain under repressing conditions.

Functional conservation and complementation between C. albicans and S. cerevisiae.

The ACH and ACS genes of C. albicans were classified as such based on sequence homology. We addressed functional conservation by heterologously expressing the C. albicans genes in S. cerevisiae strains lacking ACH1 or ACS2 (acs1Δ mutants have no phenotype in S. cerevisiae [see Materials and Methods]). The deletion strains were transformed with plasmids containing the S. cerevisiae or C. albicans genes under the control of the strong, constitutive GPD1 promoter.

As shown in Fig. 6, the C. albicans genes complement the S. cerevisiae deletions, indicating conservation of function. Deletion strains transformed with these plasmids were serially diluted onto minimal medium containing either 2% glucose or potassium acetate as the sole carbon source. Expression of CaACH1 or ScACH1 restored growth to a comparable level in the Scach1Δ strain (Fig. 6A). Similarly, CaACS2 complemented the Scacs2Δ strain (Fig. 6B). Overexpression of either gene in BY4741 did not affect growth on glucose or potassium acetate.

FIG. 6.

Cross-species complementation of ACH1 and ACS2. S. cerevisiae strains transformed with the indicated plasmids were grown overnight in SD-Ura, serially diluted 1:5, and spotted to YNB with the indicated carbon source (2%, wt/vol). (A) The wild-type strain (BY4741) and an ach1Δ strain were transformed with ACH1 from either S. cerevisiae or C. albicans under the control of the GPD1 promoter (see Materials and Methods). (B) The wild-type strain and an acs2Δ strain were transformed with ACS2 from either species. (C) The acs2Δ strain was transformed with plasmids expressing ACS1 or ACS2 from either species.

Additionally, we show that heterologous expression of C. albicans ACS1 does not restore growth of the Scacs2Δ strain (Fig. 6C). Overexpression of ScACS1 also does not compensate for the absence of ACS2. Because these genes were expressed from the constitutive GPD1 promoter, this indicates that carbon source-based expression differences do not explain the radical difference in phenotypes between acs1 and acs2 mutants in the two species. The S. cerevisiae Acs1 and Acs2 enzymes are known to have distinct kinetic properties (36), and these, or possibly localization differences, likely explain the observed phenotypes.

acs mutations do not affect global histone acetylation.

It was recently reported that ACS2 is a primary source of acetyl-CoA used for nuclear histone acetylation and, thus, global transcriptional regulation in S. cerevisiae (35). We tested whether the C. albicans Acs2 enzyme performs a similar function by assaying N-terminal acetylation of histones H3 and H4 using antibodies specific for the acetylated proteins (see Materials and Methods). When we repressed transcription of the acs2Δ/MET3p-ACS2 allele by adding Met-Cys, there was no change in acetylation of H3 or H4 over the course of 3 hours, far longer than necessary to see the cessation of growth (data not shown). We also tested the acs1Δ/Δ mutant and, likewise, found no difference in acetylation (data not shown). We conclude that acetyl-CoA synthesis via the Acs enzymes is not required for histone acetylation in C. albicans, in contrast to published data for S. cerevisiae.

ACH1 and ACS1 are not required for virulence in a mouse model of disseminated candidiasis.

Previous work had shown that C. albicans makes use of nonglucose carbon sources during systemic infection in animal models of disseminated infection (2, 18, 27, 29). However, not all mutations that impair growth on alternative carbon sources are necessary in vivo; we have shown that deletion of carnitine acetyltransferases does not attenuate virulence (34a, 42). To determine the role of acetyl-CoA metabolism in vivo, we tested the acs1Δ/Δ and ach1Δ/Δ strains in the standard tail vein injection mouse model of disseminated hematogenous candidiasis (Materials and Methods). We found that neither mutation significantly reduces virulence compared to complemented or wild-type strains (Fig. 7). The ach1Δ/Δ strain was somewhat variable in this assay, so the assay was repeated several times; we concluded that if there was an effect, it was minor and not statistically significant. This is perhaps not surprising, since the in vitro phenotypes are mild.

FIG. 7.

ach1Δ/Δ and acs1Δ/Δ are not attenuated in a mouse model of disseminated infection. Mice were inoculated via tail vein injection with 10⁶ mid-log-phase yeast form cells. Animals were monitored for signs of morbidity, and moribund mice were euthanized according to approved protocols. The graphs represent cumulative data from two independent experiments totaling 20 mice per strain.

We did not evaluate loss of ACS2 in vivo. This experiment would have required reconstructing the conditional ACS2 allele using a tetracycline-regulated promoter (23), as the MET3 promoter is not appropriate for in vivo studies. However, the growth defects of ACS2-depleted strains are so severe that it is extremely unlikely that they would retain virulence; it has been shown that other mutants that cannot grow on glucose are avirulent in vivo (2, 31).

DISCUSSION

In this work, we have examined the in vitro and in vivo roles of acetyl-CoA/acetate interconversion in C. albicans by focusing on acetyl-CoA hydrolase (ACH) and acetyl-CoA synthetase (ACS). C. albicans acs2Δ mutants, while not completely inviable, are unable to assimilate a wide variety of carbon compounds, including glucose, ethanol, and acetate. In contrast, deletion of C. albicans ACS1 produced no observable phenotype but could support growth as the sole Acs enzyme in the presence of fatty acids or glycerol. Expression analysis is consistent with these findings, with ACS2 as the dominant isoform in most conditions. Deletion of ACH1 conferred a mild reduction in growth on some nonfermentable carbon sources, notably acetate and ethanol. Neither acs1Δ nor ach1Δ mutants, however, were attenuated in a mouse model of candidiasis.

Interest in carbon metabolism in pathogens has increased recently due to observations that suggest that at least some environments within the host are deficient in glucose, the preferred nutrient for fungi and many bacteria. Many genes encoding key steps of alternate carbon metabolism are strongly induced, including those of the glyoxylate cycle (e.g., ICL1), β-oxidation (e.g., FOX2), and gluconeogenesis (e.g., FBP1) and these pathways are required for full virulence in mouse models of disseminated infection (2, 18, 27, 29). Similarly, the glyoxylate cycle is required for virulence in M. tuberculosis (19, 21). Because these pathways are compartmentalized in eukaryotic cells, intermediates (acetyl-CoA being the most important) must be transported across organellar membranes. For acetyl-CoA this occurs by converting it to acetate via acetyl-CoA hydrolase (Ach1), conjugating it to carnitine via carnitine acetyltransferases, causing it to cross the membrane by an unknown mechanism, then reversing the process on the other side, using acetyl-CoA synthetase (Acs1, Acs2) to regenerate this molecule. Similar to the findings here, we and others have shown that the carnitine acetyltransferases are induced in phagocytosed cells and are required for the assimilation of nonfermentable carbon sources (17, 28, 34a, 42).

Mutations that disable β-oxidation of fatty acids, the glyoxylate cycle, or gluconeogenesis impair growth on alternative carbon sources and attenuate virulence, moderately to severely depending on the mutation (2, 18, 27, 29). In contrast, mutations in carnitine acetyltransferases and in ACH1 do not compromise virulence despite causing in vitro phenotypes (34a, 42). What is the reason for the difference? For ACH1, the mutant has very mild in vitro phenotypes, and one might not expect this to reduce virulence. In contrast, the importance of acetyl-CoA metabolism can be seen with ACS2, a gene that is essential for viability under most conditions. Similarly, to date we have been unable to construct a strain lacking all three carnitine acetyltransferases (CTN1, CTN2, and CTN3), suggesting that this gene family may also be essential (42).

In the case of both ACS2 and CTN, the phenotypes of C. albicans differ from those of the model yeast S. cerevisiae. Mutation of acs2 in C. albicans confers a much more extensive phenotype than in S. cerevisiae, in which ACS2 is required only for growth in sugars. Similarly, C. albicans single ctn mutants have carbon utilization defects that are more extensive than those of the cognate S. cerevisiae mutants (28, 34a, 42). Finally, C. albicans fox2 deletion strains are deficient in β-oxidation, as expected from S. cerevisiae precedents, but also do not grow in the presence of ethanol or acetate (27, 29). Taken together, these findings suggest that the regulation and function of carbon metabolic pathways have diverged between these two yeasts with vastly different natural environments. Given the importance of carbon metabolism in vivo, we continue efforts to understand these differences.

Finally, nutrient acquisition is a fundamental challenge for all organisms, pathogen or otherwise. While it has been appreciated for many years that mammalian hosts effectively sequester some nutrients, such as iron, from microbes, it is becoming increasingly clear that nutrient deprivation in the host is a general condition of pathogenic species, one that must be overcome as a prerequisite to disease progression. We note that many of these alternative carbon pathways are highly conserved among microorganisms but are absent from mammals, and thus they make attractive drug targets. Further research will be required to determine whether inhibitors of these processes have value as antimicrobial agents.