Gluconeogenesis in Candida albicans

Abstract

According to different metabolic situations in various stages of Candida albicans pathogenesis the regulation of carbohydrate metabolism was investigated. We report the genetic characterization of all major C. albicans gluconeogenic and glyoxylate cycle genes (fructose-1,6-bisphosphatase, PEP carboxykinase, malate synthase and isocitrate lyase) which were isolated after functional complementation of the corresponding Saccharomyces cerevisiae deletion mutants. Remarkably, the regulation of the heterologously expressed C. albicans gluconeogenic and glyoxylate cycle genes was similar to that of the homologous S. cerevisiae genes. A C. albicansΔCafbp1 deletion strain failed to utilize non-fermentable carbon sources but hyphal growth was not affected. Our results show that regulation of gluconeogenesis in C. albicans is similar to that of S. cerevisiae and that the current knowledge on how gluconeogenesis is regulated will facilitate the physiological understanding of C. albicans.

Gluconeogenesis, CAT8, Saccharomyces cerevisiae, Candida albicans

1 Introduction

During growth on non-fermentable carbon sources yeast cells require sugar phosphates for the synthesis of essential cellular components. The process by which these compounds are synthesized is known as gluconeogenesis. Most of the glycolytic reactions are reversible under physiological conditions with two exceptions. Phosphofructokinase and pyruvate kinase have to be bypassed due to unfavorable thermodynamic balances. In many yeasts including Saccharomyces cerevisiae and Candida albicans this is achieved by phosphoenolpyruvate carboxykinase (Pck1p, building phosphoenolpyruvate), and fructose-1,6-bisphosphatase (Fbp1p, building fructose-6-phosphate). In addition, the enzymes of the glyoxylate cycle, isocitrate lyase and malate synthase, are necessary for gluconeogenesis and in S. cerevisiae the succinate/fumarate transporter, Acr1p, has been shown to be a link between the anaplerotic reactions of the glyoxylate cycle and the TCA cycle [1].

In S. cerevisiae the regulation of most gluconeogenic and glyoxylate cycle enzymes occurs on the levels of transcription, allosteric regulation and proteolytic degradation [2–5]. The genetics of glucose repression have been studied intensively in S. cerevisiae and a number of regulatory elements have been described (for review see [6–9]). Studies on the 5′-regulatory sequences of gluconeogenic and glyoxylate cycle genes, as well as genetic experiments, have pointed to a co-regulatory mechanism acting on common elements within the promoters. Upstream activating sequences (UAS) also named CSRE for carbon source-responsive element (consensus sequence: CGGNBNVMHGGA) in the promoters of PCK1, FBP1, MLS1 and ACR1 have been shown to be essential for release from glucose repression [1,10–13]. The promoters of ICL1, MLS1 and FBP1 also contain upstream repressing sequences (URS) which serve as Mig1p binding sites. Mig1p is a C₂–H₂ zinc finger protein which is a central element for glucose repression of genes required for galactose, maltose or sucrose metabolism. In the presence of glucose Mig1p and its functional homolog Mig2p [14–16] repress transcription by binding to a GC-rich sequence within the promoters of its target genes. Although all Mig1p binding sites within the promoters of the gluconeogenic genes are completely compatible with the binding site requirements for the glucose repressor Mig1p [14] and even the Mig1p binding site of ICL1 confers carbon source-dependent repression of a heterologous test promoter [11], ICL1, MLS1 and FBP1 remain fully glucose-repressible in a Δmig1 null mutation [10,11,17,18].

In addition to the release of Mig1p and Mig2p from cis-acting elements, the transcription of the gluconeogenic and glyoxylate cycle genes needs the binding of the Cat8p gene activator. Cat8p belongs to the family of transcriptional activators that contain a binuclear C6–zinc cluster. The transcription of CAT8 itself is also subject to glucose repression and is regulated by Mig1p [17] and Mig2p (Plummer and Entian, unpublished). Cat8p undergoes a Cat1p (Snf1p)-dependent phosphorylation, which is necessary for transcriptional activation of the gluconeogenic genes [19].

Until now little information is available about gluconeogenesis in C. albicans. According to different metabolic situations in various stages of C. albicans pathogenesis we became interested in the regulation of the genes encoding enzymes of carbohydrate metabolism. Whereas glucose may become available during lymphatic and blood infection, on skin after incorporation in macrophage C. albicans might depend on gluconeogenesis and the glyoxylate cycle for providing carbohydrates for cell wall biosynthesis, which may become of major importance for the morphologic switch from the yeast to the hyphal growth form. For better understanding the physiological role of C. albicans gluconeogenesis and the glyoxylate cycle and its regulation we have isolated CaFBP1, CaMLS1, CaICL1 and CaPCK1 from C. albicans genomic DNA libraries. After heterologous expression in S. cerevisiae the C. albicans genes were still subject to glucose repression and their derepression still depended on the S. cerevisiae Cat8p gene activator, showing that in addition to highly conserved protein sequences, the regulation of gluconeogenic and glyoxylate cycle enzymes is also very similar in S. cerevisiae and C. albicans.

2 Materials and methods

2.1 Strains and media

All S. cerevisiae strains were derived from CEN.PK2 [20] (Table 1). C. albicans strain SC5314, a clinical isolate, was used for enzyme assays. S. cerevisiae and C. albicans strains were grown at 30°C in YEPD medium (1% yeast extract, 2% peptone, 2% glucose), YEPGlyc (1% yeast extract, 2% peptone, 3% glycerol), YEPE medium (1% yeast extract, 2% peptone, 3% ethanol) or in synthetic complete (SC) medium [21]. S. cerevisiae transformants were selected on YEPD plates supplemented with 0.2 mg ml⁻¹ G418 (Gibco BRL, Karlsruhe, Germany). Strains have been deposited at the EUROSCARF strain collection (Frankfurt, Germany) and can be received from there (http://www.uni-frankfurt.de/fb15/mikro/euroscarf/index.html).

1

Yeast strains used

Strain	Genotype	Reference
S. cerevisiae
CEN.PK2	MATa/α leu2-3,112/leu2-3,112 ura3-52/ura3-52 trp1-289/trp1-289 his3-Δ1/his3-Δ1 MAL2-8c/MAL2-8c SUC2/SUC2	[20]
CEN.PK2-1C	MATa leu2-3,112 ura3-52 trp1-289 his3-Δ1 MAL2-8c SUC2	[20]
CEN.PK192-1D	MATa leu2-3,112 ura3-52 trp1-289 his3-Δ1 MAL2-8c SUC2Δpck1::loxP-KanR-loxP	this study
CEN.PK200-4D	MATa leu2-3,112 ura3-52 trp1-289 his3-Δ1 MAL2-8c SUC2Δicl1::loxP-KanR-loxP	this study
CEN.PK201-3B	MATa leu2-3,112 ura3-52 trp1-289 his3-Δ1 MAL2-8c SUC2Δfbp1::loxP-KanR-loxP	this study
CEN.PK202-2C	MATa leu2-3,112 ura3-52 trp1-289 his3-Δ1 MAL2-8c SUC2Δmls1::loxP-KanR-loxP	this study
CEN.PK130-7B	MATα leu2-3,112 ura3-52 trp1-289 his3-Δ1Δcat1::HIS3 MAL2-8c SUC2	[17]
CEN.PK143 -1B	MATa leu2-3,112 ura3-52 trp1-289 his3-Δ1 MAL2-8c SUC2Δcat8::loxP-KanR-loxP	this study
C. albicans
SC5314	Clinical isolate	[51]
CAI-4	Δura3::imm434/Δura3::imm434	[24]
CAE1	FBP1/Δfbp1::URA3Δura3::imm434/Δura3::imm434	this study
CAE2	FBP1/Δfbp1::hisGΔura3::imm434/Δura3::imm434	this study
CAE3	Δfbp1::URA3/Δfbp1::hisGΔura3::imm434/Δura3::imm434	this study
CAE4	RP10/rp10::MET3p-FBP1+URA3Δfbp1::URA3/Δfbp1::hisGΔura3::imm434/Δura3::imm434	this study

C. albicans mutants used in this study are listed in Table 1. For growth of ura⁻ strains, media were supplemented with uridine at a concentration of 80 μg ml⁻¹. For germ tube induction, C. albicans strains were grown overnight at 30°C in SCD and then shifted to SC media with the indicated carbon source in the presence or absence of 10% fetal bovine serum (Gibco) and incubated at 37°C.

2.2 Deletion of ScFBP1, ScPCK1, ScMLS1, ScICL1 and ScCAT8

For polymerase chain reaction (PCR)-mediated deletions in S. cerevisiae the loxP-kanMX-loxP gene disruption cassettes previously described [22] were used and for yeast transformation the protocol of Schiestl and Gietz [23] was followed. Gene-specific deletion cassettes were constructed by PCR from pUG6 as template with primers homologous to both the loxP-kanMX-loxP cassette and the gene of interest (Table 2). After transformation of the deletion cassettes into CEN.PK2, correct gene replacement events in the diploid cells and the derived haploid mutants were verified by diagnostic PCR. Diploid clones heterozygous for the respective deletion were sporulated on 1% potassium acetate agar plates for 3 days at 30°C.

2

Oligonucleotides used for gene deletion and plasmid constructs

Primer	Sequence
ICL1-S1	ATG CCT ATC CCC GTT GGA AAT ACG AAG AAC GAT TTT GCA GCA GCT GAA GCT TCG TAC GC
ICL1-S2	TTC TTT ACG CCA TTT TCT TTG AAT TGA TCT TCT GTG ACA CGC ATA GGC CAC TAG TGG ATC TG
MLS1-S1	ATG GTT AAG GTC AGT TTG GAT AAC GTC AAA TTA CTG GTG GCA GCT GAA GCT TCG TAC GC
MLS1-S2	ATC AGT GGG CGT CGC CTT AGT GGA CAC AAT TTC GTC GTA CGC ATA GGC CAC TAG TGG ATC TG
FBP1-S1	ATG CCA ACT CTA GTA AAT GGA CCA AGA AGA GAC TCT ACC GCA GCT GAA GCT TCG TAC GC
FBP1-S2	GCT TTT CCC CCT GCT TGT TCC ATT AAG AAA GCC ATT GGG AGC ATA GGC CAC TAG TGG ATC TG
PCK1-S1	ATG TCC CCT TCT AAA ATG AAT GCT ACA GTA GGA TCT ACT TCA GCT GAA GCT TCG TAC GC
PCK1-S2	TTA CTC GAA TTG AGG ACC AGC GGC TAA TAC ATC TGG TGT GGC ATA GGC CAC TAG TGG ATC TG
CAT8-S1	GAA TTT ACA AGT TGA GTA AAA GAC AAG ACA CAA AAT TAT GCA GCT GAA GCT TCG TAC GC
CAT8-S2	ATT CCG TTT TGA ATA TAT TAC ACT ATG AAA TAA AGA ATT AGC ATA GGC CAC TAG TGG ATC TG
CaFBP1-S1	TAT CCG CCA TTC GGG CGA ACC AGG GTT GGG CGC ACC AAG AAG TAT AAA AAT TCT GGG GTA CTC ATG TTT GAC AGC TTA TC
CaFBP1-S2	TGC ATG ATG GTG TGC TAC AAA AAA AAA AAA AAA GTC GAC AAT ATT TTA TTT TTT TTT GTT CGG TCG GAC AGT GCT CCG AG
hisG-S1	TAT CCG CCA TTC GGG CGA ACC AGG CTT GGG CGC ACC AAG AAG TAT AAA AAT TCT GGG ATA AGA TCT TCC AGT GGT GCA TG
hisG-S2	TGC ATG ATG GTG TGC TAC AAA AAA AAA AAA AAA GTC GAC AAT ATT TTA TTT TTT TTT GTT GAT CCC TGT TCC AGT CAA TC
5′-CaFBP1	AAA CAA CTG CAG AAC ATG AGT GGC CCA CTT AAT TC
3′-CaFBP1	GAG TGG CTG CAG AGG ACT TTA TGG TAT GGG AAA C

2.3 Deletion of CaFBP1 and integration of MET3-CaFBP1

A CaURA3 gene disruption cassette was amplified from pCUB6 [24] with primers homologous to both the CaURA3 gene and CaFBP1 (CaFBP1-S1/CaFBP1-S2; Table 1). After transformation of the deletion cassette into CAI4 [24] correct gene replacement of uridine prototrophic colonies was verified by diagnostic PCR. The resulting strain CAE1 was used for a second round of transformation with a hisG sequence amplified with primers hisG-S1/hisG-S2 and pCUB6 as template. The hisG sequence contained the same 5′- and 3′-flanking regions as the CaURA3 marker for replacing the CaURA3 gene in CAE1 by the hisG cassette. Transformants were selected by plating on 5-fluoro-acetic acid (5FOA). The resulting FOA⁺ strain (CAE2) was used for the next round of transformation with the CaURA3 cassette generating strain CAE3.

A repressible MET3p-CaFBP1 gene was constructed by amplifying the complete open reading frame of CaFBP1 using primers 5′-CaFBP1/3′-CaFBP1 and pCaFBP1 as template. The PCR product was digested with PstI and ligated into pCaEXP [25] resulting in plasmid pDE110. For integration the MET3p-CaFBP1 construct into CAE3, pDE110 was linearized with StuI before transformation.

2.4 Cloning of CaMLS1, CaICL1, CaPCK1 and CaFBP1 by functional complementation

S. cerevisiae deletion strains Δpck1 (CEN.PK192-1D), Δicl1 (CEN.PK200-4D), Δmls1 (CEN.PK202-2C) were transformed with a C. albicans genomic DNA library inserted into shuttle vector YRp7. CaFBP1 was isolated by complementing the deletion of ScFBP1 in CEN.PK201-3B (Δfbp1) by using a C. albicans genomic library in the multicopy shuttle vector YEplac195, because we were unable to isolate a complementing clone from the YRp7 library. Both libraries were kindly provided by J. Ernst (Düsseldorf, Germany). For each cloning procedure approximately 30 000 Trp⁺ or Ura⁺ transformants were collected from the selection plates (SCD-trp or SCD-ura) with liquid SC medium and plated on selective SC agar plates containing 3% glycerol, 2% acetate and 2% ethanol as carbon sources (SCEGA plates). After 4 days of incubation at 30°C several colonies were observed. From transformants grown on selective medium, plasmids were isolated. Functional complementation was confirmed after retransformation of the S. cerevisiae deletion mutants with the corresponding plasmids. The sizes of the genomic C. albicans inserts in the complementing plasmids were estimated by restriction analysis. For further analysis (e.g. sequencing) the ΔScicL1-complementing insert in YRp7 was digested with NsiI/SalI and a 3.2 kb fragment (2.9 kb C. albicans genomic DNA and 0.3 kb vector YRp7) was subcloned into pRS315 cleaved with PstI/SalI. This subclone complements ScICL1 and was named pCaICL1 (Fig. 1). The ScMLS1-complementing plasmid was digested with NcoI/XbaI and the resulting 4.2 kb fragment (3.7 kb C. albicans genomic DNA and 0.5 kb vector YRp7) was subcloned into pSAL (a CEN6/ARSH4, LEU2-containing vector; P. Koetter, unpublished) resulting in pCaMLS1. The plasmid pCaFBP1 complementing the ΔScfbp1 contains an insert of 3.3 kb. The growth phenotypes of all transformants and the respective Δ recipient strains are shown in Fig. 1.

1

Functional complementation of S. cerevisiae deletion mutants. S. cerevisiae deletion strains CEN.PK192-1D (Δpck1), CEN.PK200-4D (Δicl1), CEN.PK201-3B (Δfbp1) and CEN.PK202-2C (Δmls1) transformed with plasmids containing complementing C. albicans sequences were grown on SCE (3% ethanol) agar plates. As control the corresponding deletion strains were transformed with empty vectors.

2.5 DNA sequencing

Template DNA was prepared and sequencing was done by the dideoxy chain-termination method [26] by SRD GmbH (Oberursel/Frankfurt, http://www.srd-biotec.de). The sequences of CaICL1, CaFBP1 and CaMLS1 have been deposited at GenBank under accession numbers AF222905, AF222906 and AF222907. The deduced protein sequences were used for homology searches with the BLAST network service (National Center for Biotechnology Information).

2.6 Preparation of cell-free extracts and enzymatic assay

Crude extracts were prepared with glass beads [27], and protein concentration was determined by the microbiuret method [28] at OD₂₉₀ with bovine serum albumin as a standard. For enzyme assays yeast strains were grown to early-log phase in the presence of glucose and either harvested for crude extract preparation or washed with sterile water, shifted overnight to a medium with a non-fermentable carbon source before harvesting and crude extract preparation. Fructose-1,6-bisphosphatase activity was measured by the method described by Gancedo [29] and phosphoenolpyruvate carboxykinase activity according to Hansen et al. [30]. Isocitrate lyase and malate synthase activity were determined by the protocol of Dixon and Kornberg [31].

3 Results

3.1 Sequence analysis of C. albicans gluconeogenic genes

The nucleotide sequence of pCaFBP1 was determined. An open reading frame of 960 bp encoding a protein of 320 residues was found which shows strong homology to fructose-1,6-bisphosphatases of other species (Table 3). The codon adaption index (CAI; [32]) of ScFBP1 (0.180) is lower than the CAI of CaFBP1 (0.308). Within the sequence is one CTG at amino acid position 250 which encodes a serine instead of a leucine in C. albicans[33]. In S. cerevisiae the corresponding amino acid is an isoleucine at position 259. For S. cerevisiae it was previously shown that the N-terminal proline of the enzyme is necessary for the ubiquitination and degradation of ScFbp1p. CaFbp1p lacks this N-terminal proline and is therefore probably no substrate for rapid glucose-induced degradation in S. cerevisiae. pCaFBP1 contains 1365 bp promoter sequence of CaFBP1.

3

Homologies given in % identity of C. albicans gluconeogenic and glyoxylate enzymes

	CaIcl1p	CaMls1p	CaPck1p	CaFbp1
C. tropicalis	94%	94%	–	–
Yarrowia lipolytica	71%	–	–	–
Emericellus nidulans	67%	57%	–	–
Kluyveromyces lactis	–	–	70%	–
Kluyveromyces marxianus	–	–	–	65%
S. cerevisiae	67%	55%	69%	64%
Schizosaccharomyces pombe	–	–	–	65%
Neurospora crassa	63%	56%	–	–
Coprinus cinereus	59%	–	–	–
Homo sapiens	–	–	14%	48%
Escherichia coli	32%	46%	43%	45%

Within the nucleotide sequence of pCaICL1 an open reading frame of 1650 bp encodes a putative protein of 550 residues. A BLASTp analysis of the deduced protein revealed high homology to isocitrate lyases from other species (Table 3). We therefore named this ORF CaICL1 and the corresponding protein CaIcl1p. The codon adaptation index (CAI) for CaICL1 is 0.652 (ScICL1: 0.251). CaIcl1p contains a conserved Icl1p signature sequence [34] from amino acid 208 to 213 (KKCGHM) and a C-terminal peroxisomal targeting signal [35] at amino acids 548–550 (AKA). The plasmid contains a 1033 bp promoter region of CaICL1.

The nucleotide sequence of pCaMLS1 contains an open reading frame of 1653 bp which showed high homology to known malate synthase enzymes (Table 3). This ORF was named therefore CaMLS1 and the corresponding protein CaMls1p. The deduced CaMls1p contains a conserved malate synthase signature sequence from amino acids 274–289. The CAI of 0.574 for CaMLS1 is higher than the CAI for ScMLS1 (0.222). The plasmid contains a 768 bp 5′-regulatory sequence from CaMLS1.

The recombinant plasmid complementing the ScPCK1 deletion (named pCaPCK1) contained an insert of 10 kb. The existence of CaPCK1 on pCaPCK1 was confirmed by digestion with restriction endonucleases and the appearance of characteristic restriction fragments (data not shown). The promoter of CaPCK1 on pCaPCK1 is at least 866 bp in length.

The sequences of CaFbp1p and CaPck1p showed no signal sequences for any distinct localization, suggesting that these proteins are also cytoplasmic when expressed in S. cerevisiae. CaIcl1p contains a C-terminal microbody-targeting signal from amino acids 548–550 so that peroxisomal localization is possible. CaMls1p has no known peroxisomal targeting signature like the C-terminal SKL motif in ScMls1p. The peroxisomal localization of ScMls1p is not essential for a functional glyoxylate cycle in S. cerevisiae[36]. Therefore the cytoplasmic localization of CaMls1p in S. cerevisiae is also possible.

3.2 Carbon source-dependent regulation of C. albicans gluconeogenesis

As in S. cerevisiae Fbp1p, Pck1p, Icl1p and Mls1p activities were not detectable in C. albicans after growth on glucose but they were induced after overnight shift to ethanol medium (Fig. 2). The activities of Icl1p and Fbp1 were further analyzed in media containing both ethanol and glucose as carbon source. For both enzymes no activity was measurable in crude extracts from cells grown in this mixed carbon source medium, clearly proving that the gluconeogenesis in C. albicans is also repressed by glucose (data not shown).

2

Activity of C. albicans gluconeogenic enzymes. Activity of gluconeogenic enzymes of crude extracts from C. albicans strain SC5314 under glucose repression and derepression conditions.

After transformation, the C. albicans genes encoding gluconeogenic and glyoxylate cycle enzymes could restore growth completely on ethanol of S. cerevisiae mutants deleted for the respective genes. After growth on glucose we could measure only residual activity for the heterologously expressed gluconeogenic and glyoxylate cycle enzymes of C. albicans. Shifting cells overnight to ethanol results in enzyme activities in the respective transformants which were several-fold increased (Fig. 3).

3

Regulation of heterologously expressed C. albicans genes in S. cerevisiae. Enzyme activity of Pck1p, Icl1p, Fbp1p and Mls1p under glucose repression and derepression conditions in crude extracts prepared from several S. cerevisiae strains expressing the corresponding C. albicans genes.

3.3 Genetic elements involved in the regulation of C. albicans gluconeogenesis

Carbon source-dependent expression of C. albicans gluconeogenic genes in S. cerevisiae indicated that these genes are also subject to the glucose repression system of S. cerevisiae. Therefore we tested their dependency on the main regulatory element of gluconeogenic and glyoxylate cycle genes, the CAT8-encoded transcriptional activator. After transformation of the C. albicans gluconeogenic and glyoxylate cycle genes into a ScΔcat8 mutant (CEN.PK143-1B), no enzyme activities of CaFbp1p, CaMls1p and CaIcl1p were found, neither on glucose nor after a shift to non-fermentable carbon sources (Fig. 3). These results show that the expression of CaFbp1p, CaMls1p and CaIcl1p depends on a functional Cat8p transcription activator and suggest that a homologous gene might also exist in C. albicans. In contrast to CaFBP1, CaMLS1 and CaICL1 an S. cerevisiaeΔcat8 mutant expressing CaPCK1 resulted in considerable PEP carboxykinase activity. Obviously CaPCK1 expression was independent of the Cat8p transcriptional activator in S. cerevisiae (Fig. 3).

Cat8p needs post-transcriptional phosphorylation for its activation [19]. The phosphorylation of Cat8p depends on the Cat1p protein kinase which is, in addition to the regulation of gluconeogenesis and glyoxylate cycle genes, the key regulatory element in glucose repression/derepression. CAT1 (also known as SNF1 or CCR1) protein kinase is required for the derepression of genes encoding enzymes of alternative sugar utilization [37–39]. Besides its function in activation of Cat8p, the Cat1p kinase also phosphorylates the repressors Mig1p and Mig2p, which triggers their release from the respective promoters and their transport out of the nucleus [15]. To analyze whether the heterologously expressed CaPCK1 gene whose expression is Cat8p-independent is regulated by the Cat1p kinase system we measured CaPck1p enzyme activity in CEN.PK130-7B (Δcat1). We transformed pCaPCK1 in CEN.PK130-7B (Δcat1) and prepared crude extracts from glucose-grown cells or cells shifted overnight to a non-fermentable carbon source. No Pck1p enzyme activity was detectable in both crude extracts, showing that CaPck1p activity is CAT1-dependently regulated in S. cerevisiae (data not shown).

Several putative UAS/CSRE (upstream activating sequence/carbon source-responsive element) consensus sequences (CGGN₆GGA) were identified by sequence analysis of the C. albicans promoters (Table 4). The promoter of CaMLS1 is the only one which has no UAS element, but enzyme activity of CaMls1p is Cat8p-dependent. The Cat8p-dependent transcriptional activation is therefore UAS-independent. A similar phenomenon was described previously for ScACR1[1]. Additionally, consensus sequences for Mig1p binding sites are found in all promoters of C. albicans gluconeogenesis genes (Table 4).

4

Putative URS (upper part) and UAS (lower part) elements within the promoters of C. albicans gluconeogenic genes

Within promoter of	Type	URS sequence	Location
CaFBP1	URS	ATAAAAATTCTGGGGTA	−71–−55
CaICL1	URS	AAAAAAAATCTGGGGAA	−261–−245
CaMLS1	URS	AAAATAAATCTGGGGTA	−302–−286
CaMLS1	URS	TGCAAAAAAAGGGGGAA	−106–−120
CaPCK1	URS	CCAAAAAATCTGGGGTA	−219–−235
CaICL1	URS	TAAATTATCCTGGGGTC	−734–−750
CaICL1	UAS	CGG ATAGATGGA	−337–−348
CaPCK1	UAS	CGG GGGAACGGA	−310–−321
CaPCK1	UAS	CGG GACAAGGGA	−188–−177
CaFBP1	UAS	CGG CACTTTGGA	−214–−204
CaFBP1	UAS	CGG ATAAAAGGA	−109–−120

The promoter of CaMLS1 further contains a putative ‘peroxisome box’[40] with the consensus sequence CGGNNNTNA (−354 to −346) which is found in many peroxisomal protein-encoding genes in S. cerevisiae.

3.4 Deletion and regulated expression of CaFBP1

We have deleted CaFBP1 in C. albicans to determine its physiological role as an example for the gluconeogenic pathway. The most widely used gene disruption strategy involves the Ura-blaster cassette [24,41]. One limitation of the Ura-blaster is the difficulty to amplify the large cassette by PCR to high amounts which are necessary for gene disruption with short homology regions in C. albicans[42]. So we decided to generate a PCR product of CaURA3 with short homology regions of CaFBP1 flanking 3′ and 5′ of CaURA3 (fbp1-CaURA3-fbp1; Table 2). For generation of CAE1 (where the first allele of CaFBP1 was deleted) we transformed approximately 30 μg PCR product into CAI4. We obtained over 100 URA⁺ colonies and confirmed in several of them the correct integration of CaURA3 into the CaFPB1 locus by diagnostic PCR. For recycling the CaURA3 marker in CAE1 we transformed a hisG sequence into CAE1 for replacing the CaURA3 marker. Transformants were selected on 5-fluoro-orotic acid plates resulting in strain CAE2 which was chosen as parental strain for the second round of transformation with the fbp1-CaURA3-fbp1 PCR product. Due to the fact that a strain without Fbp1p is not able to grow on non-fermentable carbon sources, we replica-plated the approximately 200 uracil prototrophic colonies from the second round of transformation on SC glycerol. Only two of them were not able to grow on SC glycerol, indicating that the second allele of CaFBP1 was deleted. We named those strains CAE3. To confirm the deletion phenotype we measured Fbp1p activity in CAE3. It turned out that CAE3 showed no Fbp1p enzyme activity under derepression conditions (Fig. 4a).

4

Deletion analysis and regulated expression of CaFBP1 in C. albicans. a: Enzyme activity of CaFbp1p in crude extracts prepared from SC5314, CAE1 and CAE3 under glucose repression and derepression conditions. b: CAE4, CAE3, CAE1 and SC5314 on SC glycerol without methionine or SC glycerol with 2.5 mM methionine and cysteine after 5 days at 30°C. YEPD overnight cultures were diluted to a final OD of 0.5, and then spotted onto medium as indicated.

To confirm the fbp1 deletion we integrated the CaFBP1 gene behind a regulatable CaMET3 promoter into the RP10 locus of CAE3. Because we had no auxotrophic marker to select for the CaFBP1 integration we selected for colonies which were again able to grow on SC glycerol without methionine. Here the CaMET3 promoter in front of the integrated CaFBP1 is induced but is tightly repressed when growing on a medium containing a mixture of cysteine and methionine [25]. From the 200 appearing colonies all tested colonies were able to grow on SM, SCD+2.5 mM cysteine/methionine, SC glycerol, without a difference in comparison to wild-type, but all were unable to grow on SC glycerol plates containing 2.5 mM cysteine/methionine (Fig. 4b).

3.5 C. albicansΔfbp1/Δfbp1 switch from yeast to filamentous form

C. albicans can switch from a unicellular yeast into a filamentous form. This switch is thought to be important for virulence [43] and is induced by many environmental cues, perhaps the most critical for pathogenicity being the induction by serum or by macrophages. When yeasts are internalized into macrophages the environment of the ingested yeast is glucose-deficient [44] and gluconeogenesis provides the only opportunity for generating glucose, which is required for the synthesis of many macromolecules (e.g. cell wall biosynthesis). If this hypothesis is correct we expect that a C. albicans strain which lacks gluconeogenesis (e.g. Δfbp1/Δfbp1) has lost or has a reduced ability in comparison to wild-type to form germ tubes under conditions where glucose is depleted. To examine the glucose/carbohydrate dependence of filamentous growth in C. albicans wild-type and CAE3 (Δfbp1/Δfbp1) we tested their ability to build germ tubes in minimal synthetic complete media at pH 6.3 with different carbon sources. For germ tube induction the strains were pre-cultured in SCD overnight at 30°C, washed with sterile water, resuspended in test media and incubated at 37°C for 2 h. Under these conditions both strains failed to form germ tubes with a non-fermentable carbon source as sole carbon source so we decided to add 10% serum in the same media for stronger germ tube induction. We have to remark that serum is a complex mixture containing 5 mM glucose and so the glucose concentration in all non-fermentable carbon source induction media used was approximately 0.5 mM. Although we only have strongly reduced the glucose concentration and not completely depleted glucose, this is probably the most realistic situation that happens to C. albicans after incorporation into a macrophage. The morphology of the cells was analyzed microscopically 2 h after germ tube induction (Fig. 5).

5

Germ tube induction of strains SC5314 and CAE3 under glucose-limited conditions. Overnight cultures grown at 30°C were harvested, washed and incubated in different media for germ tube induction. The morphology was examined after 2 h incubation at 37°C under the microscope.

Filamentous growth and flocculation of wild-type and CAE3 (Δfbp1/Δfbp1) occurred in all media with serum except the ethanol-containing media. Even in SC without a carbon source germ tubes were visible. Although the germ tubes induced in media with non-fermentable carbon sources were significantly smaller than the germ tubes from glucose-containing induction media, there was no difference between wild-type and CAE3 (Δfbp1/Δfbp1) indicating that the Fbpase and hence gluconeogenesis is not necessary for the induction of germ tubes and filamentous growth under the chosen glucose-depleted conditions.

4 Discussion

We have isolated the genes coding for the key enzymes of the glyoxylate cycle and gluconeogenesis from C. albicans: CaICL1, CaMLS1, CaPCK1 and CaFBP1. The deduced amino acid sequences show high similarity to the gluconeogenic enzymes from other fungi, especially to those from the closely related pathogen Candida tropicalis. The activity of gluconeogenic enzymes in C. albicans showed carbon source-dependent regulation indicating that there is also a glucose repression/derepression system in C. albicans comparable to that in S. cerevisiae. The same is true after heterologous expression in S. cerevisiae of the C. albicans genes from their own promoter. Moreover, with the exception of CaPCK1, heterologous expression in ScΔcat8 results in no activity of the corresponding enzyme, indicating that the activity of these C. albicans enzymes in S. cerevisiae is also ScCat8p-dependent. Because the only known function of the ScCat8p is activating the transcription of gluconeogenic genes (derepression) by binding directly to the UAS elements [45–47], we conclude that the loss of activity of C. albicans gluconeogenic enzymes in a Δcat8 S. cerevisiae strain is also due to a loss in transcription of C. albicans gluconeogenic genes. So S. cerevisiae regulates the activity of the C. albicans gluconeogenic enzymes on the level of transcription by the same mechanisms as it uses for its own genes. This strongly suggests that C. albicans controls the activity of gluconeogenic enzymes also on the level of transcription and it should have the same/similar set of regulatory proteins as S. cerevisiae.

To date two proteins have also been identified in C. albicans of which homologs regulate glucose repression in S. cerevisiae. The first one is CaSnf1p [48], which is part of the protein kinase complex essential for glucose derepression. The second one is CaMig1p, a zinc finger protein that binds to a GC-rich recognition sequence upstream of many glucose-repressible genes [49]. Our results strongly suggest that there is also a Cat8p in C. albicans regulating the expression of gluconeogenic enzymes. This protein was quite recently identified (S. Dicken, unpublished).

Although depending on carbon source the activity of CaPck1p in S. cerevisiae is independent of Cat8p. No CaPck1p activity was detected in a ScΔcat1 deletion mutant after shift to non-fermentable carbon sources, therefore its regulation in S. cerevisiae contributes to the general S. cerevisiae CAT1-dependent glucose repression/derepression system. One possible candidate which could be responsible for the carbon source-dependent, ScCAT1-dependent and ScCAT8-independent regulation of CaPck1p in S. cerevisiae could be the Mig repressor complex. This idea is supported by the fact that we found two putative URS elements in the promoter of CaPCK1. These URS elements serve as Mig1p binding sites. Probably there is an additional post-transcriptional component of CaPck1p enzyme activity regulation in S. cerevisiae. Leuker et al. [50] reported that they could measure β-galactosidase activity in S. cerevisiae (which still harbors MIG1/MIG2) carrying a CaPCK1 promoter LAC4 reporter fusion construct, also in crude extract from glucose-grown cells. This means that the promoter of CaPCK1 in S. cerevisiae is to some extent carbon source-independently transcribed. This is probably due to other putative transcription-activating sequences in the promoter near the start ATG of CaPCK1 (data not shown).

To analyze the physiological role of gluconeogenesis and the glyoxylate cycle in C. albicans we deleted CaFBP1, encoding Fbpase, the last enzyme of the gluconeogenic pathway. The deletion strain was not able to grow on any non-fermentable carbon source; an identical phenotype was already described for S. cerevisiae. This indicates that there is only one Fbpase-encoding gene within the genome of C. albicans. This was also confirmed by measuring the remaining Fbpase enzyme activity in this strain.

We used the CaFBP1 deletion strain as a tool to investigate whether gluconeogenesis is involved in germ tube formation under glucose-limited conditions. In its natural environment several situations are possible where a switch between yeast and hyphal growth is induced although the yeast is depleted for long-chain carbohydrates like glucose. Especially incorporation into macrophage induces germ tube formation, perhaps most critical for virulence. After ingestion glucose from the surrounding media is quickly depleted and the yeast switches its metabolism for using other available non-fermentable carbon sources [44]. The gluconeogenesis provides a pathway to use a broad range of non-fermentable carbon sources for further growing and germ tube formation. We therefore initially speculated that gluconeogenesis is probably necessary for hyphal development under glucose-limited conditions. Our results indicate that the gluconeogenesis is neither essential for vegetative growth nor influences hyphal development in C. albicans under the experimental conditions chosen. However, Lorenz et al. showed the same for CaICL1, but an ICL1-deficient strain shows less virulence than a wild-type strain in the mouse model. Therefore the Δfbp1 strain might also show less virulence, what has to be examined in the future.

Acknowledgements

We thank H. Hegeman (pUG6) and P.E. Sudbery (pCaExp) for kindly providing plasmids, F. Mühlschlegel for C. albicans strains SC5314 and CAI4 and K. Melcher for critical reading of the manuscript. D.E. is granted by Graduiertenkolleg 160 (Proteinstrukturen, Dynamik und Funktion). This work was supported by the Deutsche Forschungsgemeinschaft, SFB 474 and the Fonds der Chemischen Industrie.