The ATC1 Gene Encodes a Cell Wall-linked Acid Trehalase Required for Growth on Trehalose in Candida albicans *

After screening a Candida albicans genome data base, the product of an open reading frame (IPF 19760/ CA2574) with 41% identity to Saccharomyces cerevisiae vacuolar acid trehalase (Ath1p) was identified and named Atc1p. The deduced amino acid sequence shows that Atc1p contains an N-terminal hydrophobic signal peptide and 20 potential sites for N -glycosylation. C. albicans homozygous mutants that lack acid trehalase activity were constructed by gene disruption at the two ATC chromosomal alleles. Analysis of these null mutants shows that Atc1p is localized in the cell wall and is required for growth on trehalose as a carbon source. An Atc1p endowed with acid trehalase activity was obtained by an in vtro transcription-translation coupled system. These results strongly suggest that ATC1 is the structural gene encoding cell wall acid trehalase in C. albicans . Determinations of ATC1 mRNA expression as well as acid trehalase activity in the presence and absence of glucose point out that ATC1 gene is regulated by glucose

The non-reducing disaccharide trehalose (␣-D-glucopyranosil ␣-D-glucopyranoside) has been identified and characterized in a wide variety of organisms ranging from bacteria and fungi to plant and mammals (1)(2)(3). Trehalose plays a set of different functions, depending on the specific biological system analyzed. Thus, in prokaryotes (4), it can be used as an external carbon source (Escherichia coli, Bacillus subtilis) (5,6), compatible solute (Ectothiorhodospira halochloris) (7), or structural component of the "cord factor" in mycobacteria (8). However, in yeasts and fungi, trehalose is mainly used as a reserve carbohydrate as well as a cell protectant against several physiological or environmental types of stress (2,3,9,10). In recent years, biotechnological interest in trehalose has greatly increased, and it is currently used as a stabilizer for a number of foods, pharmaceuticals, and cosmetics and as a preservative for frozen embryos (11).
In yeasts, trehalose synthesis takes place in a sequential two-step reaction: trehalose 6-phosphate is synthesized from UDP-glucose and glucose 6-phosphate, a reaction catalyzed by a magnesium-dependent trehalose 6-phosphate synthase (coded by the TPS1 gene). Subsequent dephosphorylation by a specific trehalose phosphatase (coded by the gene TPS2) renders free trehalose (3,12,13). In turn, hydrolysis of the disaccharide points to ample differences among the species of fungi studied. As a general rule, two different trehalases are present, which differ in location, catalytic properties, and regulation: (i) a cytosolic enzyme, which exhibits maximal activity at neutral pH (7.0) and which is activated by Ca 2ϩ or Mn 2ϩ and regulated by cAMP-dependent protein kinases; (ii) an acidic trehalase (optimum pH about 4.5), located inside the vacuoles or associated to the cell wall, whose activity is independent of divalent cations (2,12,14,15). Whereas neutral trehalases (Nth1p) hydrolyze the intracellular pool of trehalose, the role of acid trehalases (Ath1p) remains more elusive. Ath1p appears to be required for growth on exogenous trehalose (16) and its regulation by catabolite repression has been postulated (17,18).
In Saccharomyces cerevisiae and Candida utilis, both kinds of cytosolic and vacuolar trehalases are present (2,3,10). However, in fungi such as Neurospora crassa and Mucor rouxii, acid trehalase activity is linked to the cell wall of ascospores and conidiospores (19,20). In Emericella nidulans, an acid trehalase has been purified, and the corresponding treA gene has been cloned, being the TreAp located in the cell wall of conidiospores (21).
Candida albicans is an opportunistic pathogen fungus in humans, which can cause either septicaemic or mucosal infections (22)(23)(24). The incidence of candidiasis has dramatically increased in the last few decades, due to the rise of the immunocompromised population (25,26). Furthermore, C. albicans is a dimorphic organism, which exists as unicellular budding yeast (blastoconidia) or as mycelial structures (hyphae and/or pseudohyphae). Morphological transition has been associated with pathogenicity, and the mycelial forms seem to be predominant during host tissue colonization (22,(27)(28)(29). In recent years, trehalose metabolism has been intensively investigated in C. albicans in connection with dimorphism and as a possible contributory factor to its virulence. In this respect, disruption of TPS1 indeed impairs hypha formation and decreases infectivity (30), whereas TPS2 is involved in cell integrity and pathogenicity but is not required for dimorphic transition (31,32). However, the neutral trehalase (Ntc1p) lacks a specific role during dimorphism and pathogenicity (33).
Preliminary enzymatic data suggest that acid trehalase in C. albicans is associated to the external surface (34) and is not strictly necessary for serum-induced dimorphism (35). The whole picture of trehalose biosynthesis and hydrolysis in C. albicans has been filled in by the present paper. We have chosen a sequence-dependent approach to identify ORFs 1 that have homology with Ath1p of S. cerevisiae by an in silico analysis. From the total ORFs, we identified one that has homology with Ath1p (15), which we called Atc1p. Atc1p appears to be localized on the cell wall. Disruption of the two ATC copies is not lethal; the atc1⌬/atc1⌬ null mutant lacks acid trehalase activity and is unable to grow on trehalose as the sole carbon source. To our knowledge, among yeasts this is the first acid trehalase located as an external cell-linked enzyme.

Strains, Growth Conditions, and Transformations-The C. albicans
and Escherichia coli strains used in this study are listed in Table I. C. albicans cells were grown in YP medium (1% yeast extract, 2% bactopeptone) (36) or MM medium (0.7% yeast nitrogen base without amino acids) supplemented with the appropriate nutrients in amounts specified by Sherman (36) and with the appropriate carbon source (2% glucose or 2% trehalose). YPD, MMD, and MMT media were solidified with 2% agar. C. albicans transformation was performed according to Gietz et al. (37). E. coli DH5␣, which was grown routinely in LB medium (0.5% yeast extract, 1% tryptone, 0.5% NaCl) supplemented with 100 g/ml ampicillin (LBA) when required, was used for plasmid propagation. E. coli was transformed as described by Hanahan (38).
Nucleic Acid Manipulations and Analysis-Genomic DNA preparation from C. albicans was carried out using the method described by Fujimura and Sakuma (39); total RNA isolation was carried out as previously described (40). Plasmid purification was performed using the FlexiPrep Kit commercial system (Amersham Biosciences). Standard DNA manipulation procedures were followed using standard protocols (41). DNA probes (amplicons JCE12 and JCE35) for Southern analysis were labeled by random primed incorporation of digoxigenin-labeled deoxyuridine triphosphate (DIG-labeled DNA) using a DIG DNA-labeling kit (Roche Applied Science) according to the manufacturer's instructions. Southern hybridization was performed as described by Ramón et al. (40). DNA and RNA concentrations were determined by measuring absorbance (A 260 ) in a GeneQuant II RNA/DNA calculator spectrophotometer (Amersham Biosciences).
Plasmid Construction for Disruption and Reintegration of the ATC1 Gene-The strategy followed for disrupting the ATC gene was carried out by replacing part of the ORF (from amino acid 29 to 128) with an hisG::URA3::hisG (42). The disruption cassette construction was made by PCR amplification in two steps using genomic DNA as a template. In the first step, an amplicon of 619 bp was obtained using the sense primer JCE3 (CACAGGATCCATAACTTTAGAGTACTCAGG) and the antisense primer JCE5 (CCGTGAGCTCGCATTGCATTTCTTTCT-GCC) containing an engineered BamHI (underlined) and SacI restriction site (underlined), respectively, annealing to the ORF. The amplicon obtained (JCE 35) was digested with BamHI and SacI and then subcloned into p5921 (42), containing the hisG::URA3::hisG cassette; the resulting plasmid, which was named pATH41, contained 619 bp of the ORF (posi-tions ϩ1985 to ϩ2602 with respect to the start codon). In the second step, an amplicon of 650 bp of the ORF (positions ϩ238 to ϩ886 with respect to the start codon) was obtained by using the sense primer JCE1 (TCT-CAAGCTTGAGCAATTAGTCAATAGTCC) and the antisense primer JCE2 (CACACTGCAGCAGAAACAATGACTTGTACG) containing engineered restriction sites HindIII and PstI, respectively (underlined). The amplicon obtained (JCE12) was ligated into the HindIII and PstI sites of pATH41 to create plasmid pATH51 in which 1100 bp (29%) of the coding region were deleted.
To reintegrate the ATC1 gene into the genome of the mutant strain, ATC2-2, an integrative plasmid, using CIp10-MAL2p (43), was constructed. The construction of a derivative of CIp10-MAL2p, in which the ATC1 gene was placed under the control of the MAL2 promoter, was as follows. The ATC1 open reading frame was obtained by PCR amplification using genomic DNA as template, and the sense primer YEJ3 (CACAGATATCATGTGTTTAGCTAATTTATTCC) and the antisense primer YEJ2 (CACAACGCGTATTTACTTTTAATAAAAAACAACTTC-AGC) containing engineered restriction sites EcoRV and MluI, respectively (underlined). The amplicon obtained (YEJ32) was ligated into the EcoRV and MluI sites of CIp10-MAL2p to create plasmid CIpMAL2p-ATC1.
To target integration of CIp10MAL2p-ATC1 to the C. albicans RP10 locus, it was first digested with NcoI, which cuts plasmid uniquely within the RP10 sequence. Digested plasmid was used to transform C. albicans atc1⌬ null mutant.
Isolation of ATC1 Null Mutant-Disruption of ATC1 was achieved as described by Fonzi and Irwin (42). C. albicans CAI4 cells were transformed into Ura ϩ prototrophy with 10 g of an HindIII/SacI fragment from plasmid pATH51. Transformed cells were selected as Ura ϩ in an MMD medium lacking uridine and checked for integration of the cassette at the ATC1 locus by Southern blot analysis. One of the heterozygous disruptants recovered (designated C. albicans ATC1-1) was used to select spontaneous Ura Ϫ derivatives in MMD medium containing 5-fluoro-orotic acid (44). These clones were analyzed by Southern blot hybridization to identify those that had undergone intrachromosomal recombination between hisG repeats. One of these Ura Ϫ derivatives (termed C. albicans ATC1-2) was used for replacement of the second ATC1 allele in a similar way, using the HindIII/SacI fragment from plasmid pATH51. Transformed cells were selected as Ura ϩ , and integration into the correct allele was verified by Southern blot analysis. One of these Ura ϩ transformants (designated C. albicans ATC2-1) was used for 5-fluoro-orotic acid selection to Ura Ϫ auxotrophy. Ura Ϫ segregants were screened by Southern blot analysis using digestion with BglII to identify those carrying both disrupted ATC1 alleles. One of these ATC1 null mutants was designated C. albicans ATC2-2.
Cell Wall Purification and Solubilization of Wall Proteins-To purify cell walls, washed cells were resuspended in a small volume of cold 1 mM phenylmethylsulfonyl fluoride. Glass beads were added in a proportion of 1.5 g/mg of dry cells, and the mixture was vortexed at room temperature for eight periods of 1 min each with intermediate periods of 1 min on ice. Using this method, the whole cell population was broken up, as monitored in the phase-contrast microscope. The purification procedure was continued by repeated washing of the cell wall pellet in cold phenylmethylsulfonyl fluoride (1 mM). The pellet was collected and operationally defined as the cell wall.
Conditions for the solubilization of cell wall proteins with SDS or zymolyase 20T have been described (45,46); growth media and ␤-mer-
SDS-PAGE and Western Blot Analysis-Proteins were separated by SDS-PAGE performed basically as described by Laemmli (49) on SDS-8% (w/v) polyacrylamide gels. For Western blot analysis, proteins were electrophoretically transferred from SDS-polyacrylamide gels onto nitrocellulose filters (Hybond-C Extra; Amersham Biosciences), according to Towbin et al. (50). Filters were probed with rabbit antibodies against a synthetic oligopeptide corresponding to 15 amino acid residues (DLDPDSFKTADDVLK, from amino acids 342-356) of Atc1p at a dilution of 1:500 (pAb-Yol1), followed by goat anti-rabbit IgG conjugated to horseradish peroxidase (Bio-Rad). Synthetic oligopeptide and pAb-Yol1 were purchased from Gramsch Laboratories (Germany). Antiserum binding was visualized by using the enhanced chemiluminescence fluorescent labeling kit (Roche Applied Science) following the manufacturer's instructions. Luminescence was recorded by exposing the filter to a radioautographic X-Omat film (Eastman Kodak Co.).
RT-PCR-Oligo(dT)-primed first strand cDNA was prepared from total RNA derived from C. albicans CAI4, growing on glucose or trehalose as a carbon source, using the SuperScript first strand synthesis system for RT-PCR (Invitrogen) following the manufacturer's instructions. The sense and antisense primers were JCE3 and JCE7 (GCAT-TCAGATATGCTATCAGGG), respectively; this combination of primers specifically amplified a 685-bp amplicon. As an approximate confirmation of the relative quantity of total cDNA in each sample, primers EFB1F (ATTGAACGAATTCTTGGCTGAC) and EFB1R (CATCTTCT-TCAACAGCAGCTTG) were added to each PCR; this pair of primers specifically amplified a 526-bp fragment of EFB from C. albicans (58). Since EFB contains an intron, primers EFB1F and EFB1R also served as control to confirm that samples had no genomic DNA contamination (an amplicon of 890 bp appeared if this was the case). All samples contained the same quantity of first strand cDNA (1 g). PCR was performed as follows: EcoTaq polymerase (2 units; Ecogen) was added to 49.4 l of a solution consisting of 16.6 mM (NH 4 ) 2 SO 4 ; 2.5 mM MgCl 2 ; 67 mM Tris-HCl (pH 8.8); 0.2 mM each dATP, dCTP, dGTP, and dTTP; 0.3 M each primers JCE3, JCE7, EFB1F, and EFB1R; and 1 g of cDNA. Amplification was performed in a PCR thermal cycler (PerkinElmer Life Sciences Gene Amp PCR System 2400) by using one cycle at 94°C for 5 min and then 12, 14, 16, 18, 20, 22, 24, 26, 28, or 30 cycles as follows: 30 s of denaturation at 94°C, 30 s of annealing at 50°C, and 1 min of primer extension at 72°C. Amplicons obtained in different cycles were run on 2% agarose gel.
Preparation of Cell-free Extracts and Enzymic Assays-Cell-free extracts were obtained and acid trehalase activity was determined as described elsewhere (18). Briefly, cell samples were harvested and washed and were resuspended at known densities (10 -15 mg ml Ϫ1 , wet weight) in 10 mM MES, pH 6.0, plus 1 mM phenylmethylsulfonyl fluoride. After vigorous shaking in a vortex for 5 min at 4°C, the pellet was resuspended in the same buffer at initial density.
Acid trehalase was assayed by incubating 50 l of cell wall pellet with 200 l of 200 mM trehalase prepared in 200 mM sodium citrate containing 1 mM EDTA. The reaction was incubated at 30°C for 30 min and terminated by heating in a water bath at 100°C for 5 min. Glucose was measured in the supernatants by the glucose oxidase-peroxidase procedure (18). Specific activity is expressed as nmol of glucose liberated min Ϫ1 (mg of protein) Ϫ1 .
Homology Model for ATC1 from C. albicans-The predicted amino acid sequence (residues 1-1019 of the mature protein) of the C. albicans Atc1 protein was submitted to the JIGSAW 3D Protein Homology Modeling Server (available on the World Wide Web at www.bmm.icnet. uk/servers/3djigsaw) (51). The JIGSAW server returned the Lactobacillus brevis maltose phosphorylase structure, Protein Data Bank code 1H54 (52), as the only successful structural template within its sample space. To better model the Atc1p, the sequence was divided into two 800-amino acid overlapping blocks, which were submitted to the 3D-PSSM server (available on the World Wide Web at www.sbg.bio.ic.uk/ servers/3dpssm) (53).
In Vitro Synthesis of Acid Trehalase-The in vitro transcription-translation coupled system (Rapid Translation System RTS 100 E. coli HY Kit; Roche Applied Science) was utilized for the synthesis of acid trehalase from the PCR-amplified ATC1 gene carrying the T7 promoter sequence. To generate linear DNA, two PCRs were performed using the Linear Template Generation Set, HA tag (Roche Applied Science) following the manufacturer's instructions and the specific primers AT-C-N-FOR (CTTTAAGAAGGAGATATACCATGTGTTTAGCTAATTTA-TTCC) and ATC-N-REV (ATCGTATGGGTAGCTGGTCCCATAAAAA-ACAACTTCAGCTAG), which allow amplification of the entire, HAtagged, ATC1 gene.
Protein synthesis was produced using the RTS100 E. coli HY Kit (Roche Applied Science). Reactions were performed mainly according to the manufacturer's manual, using ϳ200 ng of DNA per reaction. Reactions were incubated at 30°C for 6 h. The resulting lysates were directly used for the acid trehalase assay. To act as a control, GFP was synthesized according to the manufacturer's recipe.

RESULTS
In Silico Screening for Potential Acid Trehalases in C. albicans-Taking as template the acid trehalase from S. cerevisiae (Ath1p), a C. albicans genome data base (available on the World Wide Web at genolist.pasteur.fr/CandidaDB/) was blasted. One ORF protein (IPF 19760/CA2574) was found and presented a significant homology with the S. cerevisiae Ath1p. Comparison of the amino acid sequence with the sequences found in protein data bases using the BLAST search algorithm (54) revealed a significant homology with both the S. cerevisiae Ath1p and E. nidulans TreAp (Fig. 1). CA2574 shares 41% identical and 59% similar amino acids with Ath1p and 29% identical and 47% similar amino acids with TreAp, both comparisons being made over their entire length. These results suggest that IPF 19760 could encode C. albicans acid trehalase, for which reason we called it Atc1p (for acid trehalase of Candida albicans).
Structural Analysis of the Amino Acid Sequence Encoded by ATC1-The ORF encodes a putative polypeptide of 1040 amino acids with a calculated molecular weight of 116,260 and a pI of 4.81. Analysis of the predicted amino acid sequence revealed an N-terminal region with characteristics of a signal peptide (55) and a predicted cleavage site between positions 22 and 23 ( . . . CSG2AP) (Fig. 2B). Hydropathy analysis (56) of the deduced amino acid sequence showed that the hydrophobic signal sequence is followed by a neutral region representing the mature protein ( Fig. 2A). Assuming the cleavage site is at position 22, the mature protein has 1019 amino acid residues with a calculated molecular mass of 113.7 kDa. Nineteen potential N-glycosylation sites (Asn-Xaa-Ser/Thr) were identified at amino acid positions 113, 147, 205, 250, 264, 313, 495, 554, 564, 624, 633, 792, 800, 867, 885, 894, 909, 966, and 1000. The difference between the predicted size of Atc1p (113.7 kDa) and the size deduced from the motility in SDS-PAGE (170 kDa) (Fig. 5) can be accounted for by glycosylation of one or more of the potential N-glycosylation sites.
A search of protein motifs in Atc1p revealed a match to the conserved sequence of glycosyl hydrolase family 65, N-terminal domain (amino acids 89 -350) and glycosyl hydrolase family 65 central catalytic domain (amino acids 351-753) (Fig. 2B). This family of glycosyl hydrolases contains vacuolar acid trehalase and maltose phosphorylase. Maltose phosphorylase is a dimeric enzyme that catalyzes the conversion of maltose and inorganic phosphate into ␤-D-glucose-1-phosphate and glucose (52). This domain is believed to be essential for catalytic activity, although its precise function remains unknown. The catalytic domain also forms most of the dimerization interface.
The JIGSAW server was used to generate a putative homology model for the Atc1p from C. albicans. Following alignment of the amino acid sequence of the L. brevis maltose phosphorylase (LbMP) and C. albicans Atc1p, the C. albicans Atc1p was modeled by the 3DPSSM server (Fig. 2C, 1). There were two gaps in the model; the first occurred between residues 1 and 88, and the second occurred between residues 931 and 1019. It is reasonable to assume that the C. albicans Atc1p model structure should not be threaded against the far N-and C-terminal regions, since they are not present in the LbMP.
An in silico analysis of the hypothetical structure of Atc1p reveals at least five distinct regions, which is consistent with the LbMP structure (52). The N-terminal region was not predicted by the folding server but was presumptively assigned as an ␣-helical region (residues 1-88), except for one short ␤-sheet (residues 72-75) on the basis of secondary structure prediction. The second domain includes residues 89 -350 in a ␤-strand region, entirely made up of antiparallel ␤-strands. This region was recognized as the N-terminal domain for glycosyl hydrolase family 65 (52). The third domain (residues 351-753) is an ␣-helical region including a short ␤-strand linker region (residues 552-578). Basically, this is the glycosyl family 65 central catalytic domain. From residue 754 to 930, the prediction program suggests a ␤-strand region. The C-terminal region was not threaded by the prediction program. It is reasonable to assume that an ␣-helical sheet region (residues 931-1019) exists, as suggested by the secondary structure prediction.
The structural resemblance of the central domain of Atc1p to LbMP leads us to suggest a structural interpretation for that protein. The two crucial catalytic residues, Glu 570 and Asp 442 , are situated at the center of the pocket. Both residues overlap with totally conserved residues in LbMP. Glu 570 (Atc1p) overlaps with Glu 487 (LbMP), and Asp 442 (Atc1p) overlaps with Asp 359 (LbMP) (Fig. 2C, 2). The carboxylate OE1 atom of Glu 570 is hydrogen-bonded to the carboxylate group of Glu 506 , whereas its OE2 atom points toward Lys 639 . The putative active site pocket in Atc1p is also walled by residues that are very well conserved in family 65 (52). Tyr 435 , which is positioned around the rim of the pocket, provides side chains that could interact with hydroxyl groups of the substrate. Asp 570 is also surrounded by a cluster of hydrophobic residues that would be good candidates for interaction with hydrophobic parts of the trehalose molecule (Leu 587 , Lys 639 , Pro 649 , and Leu 650 ).
Disruption of the ATC1 Gene-To investigate the function of the Atc1p, null mutants were constructed by targeted gene disruption, and the resulting phenotypes were analyzed. The ATC1 gene was disrupted by using a strategy originally developed for S. cerevisiae (57) and modified for use in C. albicans (42). This method uses a cassette consisting of the C. albicans URA3 flanked by direct repeats of the Salmonella typhimurium hisG (see "Materials and Methods" for details). This cassette was used to replace 1097 bp of the ATC1 ORF. A linear HindIII/SacI fragment (see "Materials and Methods") including the cassette flanked by ATC sequences was used to transform C. albicans CAI4 to Ura ϩ .
Fourteen of the resulting Ura ϩ transformants were analyzed, and nine of them were seen to contain the desired insert at the ATC1 locus (data not shown). Southern analysis of a representative isolate, C. albicans ATC1-1, after digestion with PstI, revealed that the cassette had integrated in the correct allele, originating a 5.88-and a 0.75-kb fragment when amplicons JCE12 and JCE35 were used as a probe and a fragment of 5.88 kb when the cassette was used as a probe (Fig. 3, A, step  2, and B, lanes 2); the 0.75-kb fragment appeared because a new PstI restriction site had been formed in the disruption cassette construction process. This is consistent with the replacement of one allele of the ATC with the transforming DNA. The 1.94-and 1.04-kb PstI fragments correspond to the other allele that was still present in the Ura ϩ transformants. Ura Ϫ segregants of strain C. albicans ATC1-1 were selected on a medium containing 5-fluoro-orotic acid (44) and examined by Southern blot analysis. Five of the eight independent segregants examined had undergone intrachromosomal recombination between the hisG repeats, resulting in excision of the URA3 marker and one copy of hisG, whereas three of them had experienced an interchromosomal recombination event, reverting to the parental genotype (data not shown). Southern blot analysis of a representative intrachromosomal recombinant, strain C. albicans ATC1-2, is shown in Fig. 3B (lanes 3). The 5.88-kb PstI fragment in strain C. albicans ATC1-1 containing the atc1⌬::hisG-URA3-hisG disruption was absent, and a new band of 2.88 kb was present, indicating the loss of both URA3 and one copy of hisG (Fig. 3A, step 3).
The homozygous atc1⌬ null mutant was generated after transformation of strain C. albicans ATC1-2 with the same disruption cassette. Fifteen Ura ϩ transformants were examined by Southern blot analysis, and only four of them showed the hybridization pattern corresponding to the integration of the cassette in the undisrupted allele of ATC1, whereas the other 11 transformants showed replacement of the first disrupted allele. The Southern blot analysis of a representative Ura ϩ isolate that exhibited a hybridization pattern consistent with targeting of the previously undisrupted allele, strain C. albicans ATC2-1, is shown in Fig. 3B (lanes 4). The parental 1.94-and 1.04-kb PstI fragments were absent, with a 5.88-kb fragment appearing instead, indicating correct integration. C. albicans ATC2-1 was plated on 5-fluoro-orotic acid-containing medium to select Ura Ϫ segregants. Five Ura Ϫ segregant isolates were screened by Southern analysis, and each exhibited the correct hybridization pattern. The Southern blot analysis of one of these segregants, strain C. albicans ATC2-2, is shown in Fig. 3A (lanes 5). The 5.88-kb PstI fragment was absent, giving rise to a 2.88-kb fragment, which results from the loss of URA3 and one copy of hisG, indicating that both alleles of the ATC1 gene were disrupted. Northern blot analysis demonstrated that no mRNA was present when hybridized with the amplicons JCE12/JCE35 probe in RNA samples from the null mutant C. albicans ATC2-2.
C. albicans atc1⌬ Lacks Acid Trehalase Activity and Does Not Grow on Trehalose as Carbon Source-Two trehalase ac- tivities, neutral and acid enzymes, have been characterized in S. cerevisiae and related yeasts (3). Neutral trehalase is responsible for the physiological hydrolysis of trehalose in intact cells (12,14,33), and acid trehalase is necessary to grow on trehalose as a carbon source (16,21). This led us to search for conditions under which ATC1 deletion shows a phenotype. C. albicans atc1⌬ null mutant lacked acid trehalase activity, an effect that appears to be gene dosage-dependent (Table II). Of note is the fact that substitution of glucose by trehalose as an exogenous carbon source induced a marked activation of Atc1p, either in exponential or stationary cultures (Table II). In turn, atc1⌬ cells presented the same neutral trehalase activity as the parental strain (data not shown). As Fig. 4A shows, C. albicans strains CAI4, ATC1-2, and ATC2-2 exhibited normal growth on glucose, but when a replica plating was made on a medium containing trehalose as carbon source, the null mutant strain C. albicans ATC2-2 showed no growth.
In addition to the studies using solid medium, growth in liquid medium was also analyzed. No differences were found when the different strains were grown in liquid MMD medium (Fig. 4B). When cells were grown in liquid MMT medium, no growth of the atc1⌬ null mutant strain was observed, and the parental and the heterozygous mutant strain grew as the same rate (Fig. 4B).
To assess whether the absence of acid trehalase activity in the C. albicans atc1⌬ null mutant was a consequence of the ATC1 gene disruption, the ATC1 open reading frame placed under the control of the MAL2 promoter was integrated into the genome of the atc1⌬/atc1⌬ mutant. When these transformed cells were grown on glucose as a carbon source, the activity of acid trehalase was virtually undetectable (Ͻ0.3 nmol min Ϫ1 (mg of protein) Ϫ1 ). However, when maltose was used instead of glucose as carbon source, an acid trehalase activity of 6.5 nmol min Ϫ1 (mg of protein) Ϫ1 was measured in the transformed C. albicans atc1⌬ null mutant.
The ATC1 Null Mutant Lacks a 170-kDa Cell Wall Protein-To identify the gene product encoded by ATC1, the different protein species present in different subcellular fractions were analyzed by Western blotting, using antibodies raised against a synthetic oligopeptide containing 15 amino acid res-  idues of the predicted protein (pAb-Yol1) (see "Materials and Methods"). As shown in Fig. 5, the antibody mainly detected a 170-kDa species in SDS extracts of the cell wall from parental strain but not in the null mutant. A band of 120 kDa appeared in both the parental strain and the null mutant (Fig. 5) that could be the consequence of this protein cross-reacting with the antibody. No differences were observed in the protein pattern of other cell wall extracts, spent media, and cytosolic fraction (data not shown). It was therefore concluded that ATC1, in C. albicans, codes a cell wall protein necessary to grow on trehalose as a carbon source.
Expression Patterns of ATC1 and Activity in Presence of Glucose-The expression of ATC1 was examined by RT-PCR. Fig. 6 shows the RT-PCR amplification of an ATC1-specific DNA fragment (685 bp) from first strand cDNA derived from cells growing on glucose or trehalose as a carbon source. Different samples, containing the same quantity of first strand cDNA, were prepared and subjected to different cycles of amplification. Fig. 6 shows that the quantity of EFB-specific amplicon (526 bp) was approximately the same in each of the first strand cDNA samples. The presence of one intron in the corresponding region of the EFB genomic DNA allows differentiation between bands amplified from cDNA and any potentially contaminating genomic DNA (890 bp) (58). The results indicate that ATC1 shows a differential expression pattern as a function of the carbon source, being more strongly expressed when cells were grown on trehalose as a carbon source than on glucose. In order to elucidate whether acid trehalase activity is subjected to catabolite repression by glucose, cells growing in trehalose as carbon source (MMT medium) were supplemented with 4% glucose. As is shown in Fig. 7A, acid trehalase activity decreased significantly upon glucose addition and remained at a low level after 3 h. To rule out the possibility that the presence of glucose may alter the Atc1p function, but not necessarily its mRNA abundance, an RT-PCR was performed at different times after glucose addition to cells growing in trehalose as the sole carbon source. As is shown in Fig. 7B, the abundance of ATC1 mRNA underwent a significant decrease after the glucose addition. This decrease ran parallel to a rapid loss of acid trehalase activity (Fig. 7). These results strongly suggest that ATC1 is a glucose-repressible gene.
Protein Analysis of in Vitro Translation Products-In order to verify that the ATC1 gene actually encodes a structural acid trehalase and not a transcription factor necessary for expression of another putative acid trehalase gene in the C. albicans genome, Atc1p was in vitro synthesized from a PCR-amplified ATC1 gene carrying the T7 promoter sequence (see "Materials and Methods"). Western blot analysis of the resulting lysates showed that a correct synthesis of Atc1p was performed (Fig. 8,  lane 3), and they were used for the acid trehalase assay. In the reaction mixture containing ATC1 PCR, the acid trehalase activity was 3.7 nmol min Ϫ1 (mg of protein) Ϫ1 , whereas in the reaction mixtures lacking ATC1 PCR product, measurements of acid trehalase activity were below the level of detection (Fig.  8, lanes 1 and 2).  4. A, growth of heterozygous (atc1-2) and homozygous (atc2-2) for atc1⌬ compared with the ATC1/ATC1 CAI4 parental strain. Cells were streaked on MMD (glucose) and allowed to grow for 2 days. After this time, the cells were replica-plated onto a fresh MMT (trehalose) plate and allowed to grow at 28°C for 5 days. B, growth of C. albicans on liquid MMD (glucose) or MMT (trehalose) medium at 28°C. q, C. albicans CAI4 (ATC1/ATC1); ‚, C. albicans ATC1-2 (ATC1/atc1⌬); ࡗ, C. albicans ATC2-2 (atc1⌬/at1c⌬).

DISCUSSION
Enzymes involved in trehalose metabolism must be considered as a potential target in the search for new antifungal compounds (31,32), since the sugar is absent from mammal cells, whereas trehalase is located in the brush border membranes of epithelial cells and in the kidney proximal tube (59). Previous work showed that in C. albicans, TPS1 and TPS2 genes are factors of virulence (30 -32). In addition, it is also conceivable that proteins located in the external surface should be preferential targets for antifungal drugs. On the basis of this rationale approach, we have carried out the cloning and functional characterization of a cell wall-linked acid trehalase.
To identify ORFs showing homology with S. cerevisiae Ath1p (acid trehalase) a sequence approach was followed, screening the data base of C. albicans. In this way, we found the prospective protein code for one ORF (IPF 19760/CA2574) that has high homology with Ath1p of S. cerevisiae and TreAp of E. nidulans, which we therefore called Atc1p.
The results obtained revealed that ATC1 encodes a cell wall protein, since its disruption leads to the absence of a 170-kDa protein band only in the material released from isolated cell walls by SDS. The deduced amino acid sequence reveals the presence of a signal peptide at the N terminus of the protein, which is characteristic of proteins that transit through the  12-30) and then run on a 2% agarose gel. The expected amplicons using JCE3/JCE7 (amplicon ATC1, 685 bp) and EFB1F/EFB1R (amplicon EFB1, 526 bp without intron) as a pair of primers are shown on the right (see "Results" for further details).

FIG. 7.
A, repression by glucose of the acid trehalase activity in the parental strain CAI4 (ATC1/ATC1) of C. albicans. Cells were grown in MMT medium up to early stationary phase (A 600 ϭ 1.66), and then 4% glucose was added. The acid trehalase activity was measured at different times after the glucose addition. B, ATC1 mRNA abundance measured by RT-PCR in cells growing in MMT medium at time 0, 5, and 10 h after the glucose addition. The experiment was performed as in Fig. 6 (see the legend to Fig. 6 for further details). secretory pathway. The theoretical molecular mass of mature Atc1p is 113.7 kDa, but by SDS-PAGE it shows an apparent motility of 170 kDa. This experimental motility could be due to N-glycosidic modification, as happens in S. cerevisiae and E. nidulans; Atc1p has 20 potential N-glycosylation sites.
To obtain information about the possible function of Atc1p, we analyzed the phenotype of the atc1⌬ mutant. The only phenotype that we could demonstrate when studying this mutant was the absence of acid trehalase activity as well as its inability to grow on a medium containing trehalose as a single carbon source, suggesting that Atc1p is required for the hydrolysis of exogenous trehalose (Table II, Fig. 4). This phenotype is similar to that observed for the S. cerevisiae ath1⌬ strain (15,16,60) and for E. nidulans treA⌬ strains (21). It is possible that acid trehalases form a family involved in the assimilation of external trehalose. The fact that Atc1p is localized in the cell wall, as was previously shown for E. nidulans (21), that it is probably N-glycosylated, and that it contains a signal peptide (as deduced from the amino acid sequence) is in agreement with this conclusion. In contrast to the E. nidulans and C. albicans acid trehalases, S. cerevisiae Ath1p has been localized in vacuoles (61), a difference that might reflect the distinct mechanisms that fungal species use to assimilate exogenous trehalose.
Comparison of the amino acid sequence of C. albicans Atc1p with E. nidulans TreAp and S. cerevisiae Ath1p showed that it shares 29% identity and 41% similarity with TreAp and shares 41% identity and 59% similarity with Ath1p over their entire length. A more detailed analysis of the alignment between the three proteins reveals stretches of amino acids (15-20 residues) that are up to 75% identical and could constitute the catalytic domain of these enzymes. This region (amino acids 404 -757) of maximum identity matches the central catalytic domain of the glycosyl hydrolase family 65. The two crucial catalytic residues, Glu 570 and Asp 442 that overlap with totally conserved residues in LbMP are present in the three acid trehalases, as predicted from the structural interpretation of the Atc1p.
ATC1 expression was very low when cells were grown on glucose as carbon source, unlike when they were grown on trehalose (Fig. 6), which suggests that ATC1 expression might be repressed by glucose or induced by trehalose. However, the fact that, in the presence of glucose, acid trehalase activity decreased in parallel with a significant reduction of ATC1 mRNA expression (Fig. 7) strongly supports the control of ATC1 by glucose repression. Remarkably, this regulatory mechanism has already been reported for acid trehalase (ATH1 gene) in S. cerevisiae (17,18). Analysis of the promoter region of ATC1 may be required to define the potential binding sites for transcriptional elements that may control the expression of ATC1, depending on the carbon source. Further studies with modified versions of Atc1p should provide insight into the molecular organization of this protein.