Large-Scale Identification of Putative Exported Proteins in Candida albicans by Genetic Selection

Abstract

In all living organisms, secreted proteins play essential roles in different processes. Of special interest is the construction of the fungal cell wall, since this structure is absent from mammalian cells. The identification of the proteins involved in its biogenesis is therefore a primary goal in antifungal research. To perform a systematic identification of such proteins in Candida albicans, we carried out a genetic screening in which in-frame fusions with an intracellular allele of invertase gene SUC2 of Saccharomyces cerevisiae can be used to select and identify putatively exported proteins in the heterologous host S. cerevisiae. Eighty-three clones were selected, including 11 previously identified genes from C. albicans as well as 41 C. albicans genes that encode proteins homologous to already described proteins from related organisms. They include enzymes involved in cell wall synthesis and protein secretion. We also found membrane receptors and transporters presumably related to the interaction of C. albicans with the environment as well as extracellular enzymes and proteins involved in different morphological transitions. In addition, 11 C. albicans open reading frames (ORFs) identified in this screening encode proteins homologous to unknown or putative proteins, while 5 ORFs encode novel secreted proteins without known homologues in other organisms. This screening procedure therefore not only identifies a set of targets of interest in antifungal research but also provides new clues for understanding the topological locations of many proteins involved in processes relevant to the pathogenicity of this microorganism.

Fungal infections are currently an important source of morbidity and mortality in several countries. Fungi cause a range of diseases, ranging from relatively moderate superficial infections to severe systemic diseases that are often life threatening. Their treatment is difficult because of the close similarity between mammalian and microbial cells and is mainly based on polyenes and azoles (6). Polyenes have several side effects which can be partially overcome with the introduction of new liposome-based formulations. In contrast, emerging resistance to azoles may seriously compromise their usefulness in the near future (67, 85). All the foregoing aspects have prompted the development of new strategies for the search for novel antifungal targets. Because of its absence in mammalian cells, the fungal cell wall, like that of bacteria, is the most attractive multienzymatic target, conferring potential selectivity to antimicrobial agents through its inhibition (21). The cell wall is a highly dynamic structure that is composed mainly of glucan (β-1,3 and β-1,6 type), chitin, and cell wall mannoproteins. These components are essential to fungal cells either as cross-linking enzymes that covalently join different proteins to the β-1,3, β-1,6, or chitin fractions of the wall or as structural components. They are also involved in morphogenesis and may regulate the exchange of compounds with the extracellular medium. In pathogenic fungi, cell wall proteins play a key role in the relationship between the fungal cell and the host, participating in pathogenesis through adhesion phenomena and modulation of the immune response (15).

The cell wall is not the only important source of potential targets, and other exported proteins can be also considered in this context. Membrane proteins, for example, play an essential role in fungal physiology because they are involved in nutrient transport, energy generation, and signal transduction pathways, ultimately leading to growth and host adaptation. Extracellular enzymes encode hydrolytic enzymes, such as lipases (47, 81), proteases (36, 37), and cell wall-hydrolyzing enzymes (32, 50), that are involved in pathogenicity and virulence, mainly through their contribution to invasion. The identification of these secreted proteins is thus of enormous interest because their inhibition may ultimately lead to cellular death, either as a direct consequence of cellular processes—such as the inhibition of cell wall assembly—or because they are novel virulence factors whose inhibition enables the host immune response to control and eradicate infection.

The identification of exported proteins has remained elusive for several reasons. Biochemical separation of cell wall proteins is complicated by their high level of glycosylation and cross-linking to other cell wall components (17, 43). Genetic methods are therefore an alternative approach to analyzing the localizations of these proteins. Although there are no intrinsic specific domains characteristic of exported proteins, most of them have a signal peptide: a 10- to 30-amino-acid-long amino-terminal domain that has a characteristic hydrophobicity profile, that is cleaved off during secretion, and that is responsible for introducing the protein into the secretory pathway (18, 78). Although its existence is neither necessary nor sufficient to unambiguously determine whether a given protein will be present at the cell surface, it can be used as an attractive topological signal to devise genetic screenings with gene reporters and/or gene tags whose localizations can be efficiently traced inside the cell (66).

In this work, we undertook a genetic approach to the identification of exported proteins in Candida albicans; the most frequently isolated fungal pathogen in clinical samples. We made use of genetic selection based on the Saccharomyces cerevisiae invertase gene, SUC2, by means of a strategy in which in-frame fusions with an intracellular allele of this gene can identify putative export signals. This strategy has already proved successful in a mammalian system (45). We thus identified several exported proteins—with several putative new cell wall components—constituting an important source of potential antifungal targets.

MATERIALS AND METHODS

Strains and DNA manipulations.

Due to the poor transformation efficiencies obtained with the original S. cerevisiae suc2 mutant strain, SEY2101 (MATa ura3-52 leu2-3,112 ade2-1 suc2-Δ9) (27), an Ade⁺ spontaneous revertant, named SS10, was obtained from this strain. C. albicans 1001 is a wild-type strain from the Spanish Type Culture Collection (ATCC 64385) (30) and was used as the source of genomic DNA for the preparation of the genomic libraries. C. albicans SC5314 is a wild-type strain (31) frequently used in C. albicans genetic studies. All DNA manipulations were carried out by standard procedures (2, 72). Escherichia coli DH5α [K-12 Δ(lacZYA-argF)U169 supE44 thi-1 recA1 endA1 hsdR17 gyrA relA1 (ø80lacZΔM15) F′] (33) was used in all molecular biology procedures. DNA-modifying enzymes were obtained from Boehringer Mannheim (Mannheim, Germany).

Genetic constructions.

Three different suc2 alleles lacking the signal peptide and Met²¹ (suc2m^ic1, suc2m^ic2, and suc2m^ic3) were constructed by PCR amplification of the SUC2 gene in order to achieve in-frame fusions with fragments obtained from the genomic DNA. The three alleles were amplified by using an Expand High-Fidelity PCR system from Boehringer Mannheim and pLC1 (a centromeric plasmid [obtained from P. Sanz] that comprises the wild-type SUC2 gene) (23) as a template. Three pairs of oligonucleotides were used: SUC2^lo (5′-AAGAATTCCTGTTACAGATAGCTTGG-3′) and SUC2^up1 (5′-AAGAATTCAGATCTCGCGACAAACGAAACTAGCG-3′) for suc2m^ic1 amplification, SUC2^lo and SUC2^up2 (5′-AAGAATTCAGATCTCGCGACAAACGAAACTAGCG-3′) for suc2m^ic2 amplification, and SUC2^lo and SUC2^up3 (5′-ATAGATCTCCCGGGACAAACGAAACTAGC-3′) to obtain suc2m^ic3. A BglII restriction site was introduced within the oligonucleotides at the 5′ end to allow the construction of in-frame fusions in all three reading frames with Sau3A-derived fragments (Fig. 1A). After PCR amplification, suc2m^ic1 was rendered blunt ended with the Klenow fragment of DNA polymerase and inserted into the SmaI site of the E. coli-yeast shuttle vector YEp352 (34) to generate pSE1. The SacI-HindIII fragment of pSE1, which carries suc2m^ic1, was subcloned into the SacI-HindIII sites of YCplac33B (a centromeric yeast vector obtained as a YCplac33 [29] derivative from which the BglII site had been removed) to construct pSC1. Amplified suc2m^ic2 and suc2m^ic3 alleles were first subcloned into vector pGEM-T (Promega), excised as SacI-SacI or EcoRI-PstI fragments, and finally inserted into SacI- or EcoRI-PstI-opened YEp352 in order to generate pSE2 and pSE3, respectively. The SacI-SacI fragment carrying suc2m^ic2 was also subcloned into the SacI site of YCplac33B to obtain pSC2.

FIG. 1.

Genetic constructions and testing of the system. (A) Schematic representation of the S. cerevisiae SUC2 gene and the suc2 intracellular alleles constructed. The cleavage site of the signal sequence is indicated by a vertical arrow. Sequence details of the fusion zone are given in the box. The names of the centromeric (pSC series) or episomal (pSE series) versions of the gene library vectors carrying the three different alleles are given. aa, amino acids. (B) Complementation of the Suc2⁻ phenotype with the library vectors and genetic constructions including the SUC2 promoter and signal sequence. (C) Analysis of the functionality of the C. albicans HEX1 promoter and signal sequence.

The primers sucpup (GGCGAGCTCATTTTATCATGTTTCGTTTGT) and sucplo [CGAGGATCC(A/G)(C/T)TGATGCAGATATTTTGGGC] were designed and used for the amplification of a DNA fragment carrying the S. cerevisiae SUC2 promoter and signal sequence from plasmid pLC1. The PCR fragment was cloned into vector pT7Blue-T (Novagen) to make pT7psuc. sucplo had been designed in order to obtain a set of fragments encoding different amino acids at position 21 (Met in SUC2), although sequencing eventually revealed that the fragment in pT7psuc encoded Thr²¹. After the amplified DNA was excised from this plasmid with BamHI-BamHI, it was subcloned into the BglII sites of plasmids pSE1, pSC1, and pSE3, resulting in pSE1psuc, pSC1psuc, pSC1psuci, pSE3psuc, and pSE3psuci. A DNA fragment containing the promoter, the signal sequence, and the coding sequence up to amino acid 37 of the HEX1 gene from C. albicans was amplified from SC5314 genomic DNA with the primers up-hex1 (ACCCAAGCTTCTATATTGACAGTAAAAGCGTTT) and lo-hex1 (GAGCTCGAGGATCCTTTCCCAGGTTACTGACTAT). After the PCR product was cloned into vector pT7Blue-T, it was excised as a BamHI-BamHI fragment and inserted into the BglII sites of pSC2 and pSE2 to produce plasmids pSE2phex, pSE2phexi, and pSC2phex. pSC2phex and pSE2phex bear the HEX1 promoter and signal sequence fused in frame with the suc2m^ic2 allele, while pSE2phex1 contains the same DNA insert cloned in the opposite orientation.

For assessing the functionality of these constructs (Fig. 1B), 10⁴ cells of S. cerevisiae SS10 transformants carrying various plasmids were spotted onto YEP-sucrose solid medium (1% yeast extract, 2% peptone, 2% sucrose, 2% agar, 1 μg of antimycin A/ml) and grown at 24°C. pSC1psuc, pSE1psuc, and pSE3psuc bear the SUC2 promoter and signal sequence fused in frame with the corresponding suc2m^ic allele, while in pSC1psuci and pSE3psuci the same DNA insert is cloned in the opposite orientation. pLC1 is a centromeric plasmid that contains the SUC2 gene.

Library construction and screening.

C. albicans genomic DNA was extracted from strain 1001 grown at 30°C overnight in SD medium (2% glucose, 0.67% yeast nitrogen base without amino acids) supplemented with a mixture of amino acids. Genomic DNA was partially digested with Sau3A by use of appropriate dilutions to obtain a high proportion of 0.5- to 4-kb fragments. After preparative electrophoresis, fragments of 0.5 to 2 kb were eluted with dialysis bags; incorporated into episomal vectors pSE1, pSE2, and pSE3; digested with BglII; and treated with calf intestinal alkaline phosphatase to avoid recircularization. After standardization of the optimal vector DNA/genomic DNA ratio for ligations with each of the three vectors, a series of ligation mixtures were pooled. This procedure afforded three fusion libraries in the three different frames, each of which accounted for more than 95% of the recombinant clones. Totals of 231,000, 192,360, and 128,860 colonies were obtained from the pSE1, pSE2, and pSE3 libraries, respectively. The recombinant clones had an average insert size of 1 to 1.3 kb, as determined with several small-scale preparations of plasmid DNA.

For the screening experiments, SS10 was transformed by electroporation (ElectroCell Manipulator 600; BTX Laboratories) with published protocols (4) and with 186 Ω and 1,400 V as optimal experimental conditions. Ura⁺ transformants were selected on solid SD medium supplemented with amino acids and 1 M sorbitol for osmotic protection. They were then replica plated on YEP-sucrose solid medium, and growth was observed after 3 to 10 days at 28°C. Plasmids were recovered from yeast colonies (65) and amplified in E. coli by electroporation (24). All other yeast transformations were carried out by the lithium acetate procedure (40).

Sequencing and in silicio analyses.

Inserts of the clones selected were sequenced at the Automatic DNA Sequencing Unit of the Universidad Complutense de Madrid by using oligonucleotide sucseq (5′-CATTCATCCAGCCCTTGTTG-3′), which hybridizes to the 3′-terminal region of suc2m^ic alleles. Homology searches were made by using the BLASTx and BLASTn programs (www.ncbi.nlm.nih.gov and www.embl-heidelberg.de/Services) against nonredundant databases or by using the BLASTn program against the C. albicans sequence at the Stanford Genome Technology Center. Sequence data for C. albicans were obtained from the Stanford Genome Technology Center website at www.sequence.stanford.edu/group/candida. The putative C. albicans open reading frames (ORFs) and their homologues were detected within assembly 6 of the C. albicans sequence (www.sequence.stanford.edu/group/candida/download). DNA sequence analysis of signal sequences of translated fusion proteins was performed by using the SignalP World Wide Web server (SignalP V1.1 World Wide Web Prediction Server; www.cbs.dtu.dk/services/SignalP) (59); when the results were not very clear, the iPSORT WWW Service at the Human Genome Center, Institute of Medical Science, University of Tokyo (www.HypothesisCreator.net/iPSORT), was also used. Putative glucosylphosphatidylinositol (GPI) anchors were predicted by the PSORTII (Prediction of Protein Sorting Signals and Localization Sites in Amino Acid Sequences) program at the PSORT World Wide Web server (www.psort.nibb.ac.jp/).

RESULTS

Development of a genetic selection scheme for the identification of heterologous exported proteins.

The final goal of the genetic strategy devised was the identification of C. albicans proteins that contain export signals. This was achieved by using the S. cerevisiae invertase encoded by SUC2 as a reporter of protein localization. SUC2 encodes an extracellular invertase able to hydrolyze sucrose into glucose and fructose. It thus enables the utilization by yeast cells of sucrose as a sole carbon source. Targeting of secreted proteins to the secretory pathway occurs via a short amino-terminal sequence known as the signal peptide. Deletion of the native invertase signal peptide (amino acids 1 to 19) prevents secretion and results in accumulation of the enzyme within the cell, impairing its ability to grow in a medium with sucrose or raffinose as the sole carbon source (42). This system is therefore convenient not only because the enzymatic activity of invertase is readily detectable but also because it allows a positive selection scheme. We screened the C. albicans fusion libraries with intracellular suc2 alleles in order to isolate gene fragments able to restore secretion ability to the invertase; these would presumably correspond to normally secreted proteins.

Different intracellular alleles (suc2m^ic1, suc2m^ic2, and suc2m^ic3) were constructed, enabling us to obtain in-frame fusions in the three open reading frames. They start at amino acid 22 and therefore lack the native signal peptide and the two initiator methionines of the SUC2 gene (14) (Fig. 1A). The intracellular alleles were introduced into YEp352 and YCplac33B to obtain the multicopy (pSE1, pSE2, and pSE3) and centromeric (pSC1/pSC2) vector series. Episomal vectors were used for the construction of the genomic libraries. The use of multicopy plasmids facilitates a priori the chance of isolating C. albicans genes poorly expressed in S. cerevisiae in this heterologous screening. The functionality of these constructions was checked by using the native SUC2 promoter and signal peptide in both types of vectors in different frames (suc2m^ic1-pSE1psuc, suc2m^ic1-pSC1psuc, and suc2m^ic3-pSE3psuc). SS10 transformants carrying the SUC2 promoter and signal peptide displayed growth on solid (Fig. 1B) as well as in liquid sucrose-antimycin A medium.

In order to validate this strategy for our heterologous screening, we used the C. albicans HEX1 gene. This gene encodes the secreted hydrolytic enzyme β-N-acetylglucosaminidase (12). The HEX1 gene was selected because of the presence of a clear signal peptide that includes the first 22 amino acids. Expression of the HEX1 gene in S. cerevisiae cells has been shown to occur at moderate levels, although it is not inducible in response to N-acetylglucosamine, as occurs in C. albicans (12). A chimera comprising the sequence encoding the first 37 amino acids of β-N-acetylglucosaminidase and the suc2m^ic2 allele was constructed with centromeric (pSC2) and episomal (pSE2) vectors (generating plasmids pSC2phex and pSE2phex, respectively). pSC2phex showed slight complementation of the Suc⁻ phenotype after 8 days of incubation on sucrose-antimycin A plates, while complementation with the episomal version was faster (Fig. 1C). We observed that insertion of the native SUC2 (plasmids pSC1psuci and pSE3psuci) or C. albicans HEX1 (plasmid pSE2phexi) signal peptides in opposite orientations did not complement the Suc⁻ phenotype. These results indicated that the vectors constructed enabled the functional characterization of signal sequences within C. albicans DNA, even with its own regulatory signals.

Screening.

A set of three libraries was constructed with C. albicans strain 1001 as a source of genomic DNA. Genomic DNA was used because the expression of C. albicans genes in S. cerevisiae is not a major drawback (63). This DNA was partially digested with Sau3A, and fragments of 0.5 to 2 kb—presumably corresponding to promoters and partial open reading frames—were selected. After SS10 transformation, a total of about 107,900 transformants obtained with the three episomal libraries were screened. Transformants were first grown on SD plates to a cell density of about 2,000 transformants/plate. After replica plating on sucrose medium and 3 to 10 days of incubation, 571 positives clones were selected (≃0.5%). Plasmids were rescued from the corresponding S. cerevisiae transformants and, after amplification in E. coli, were transformed back into the host suc2 strain. Some of these plasmids were eliminated because they failed in these recomplementation assays. First, we sequenced 189 clones and obtained 143 different inserts (after eliminating some S. cerevisiae SUC2 genes, some very small inserts, and some repeated clones). Analysis of those inserts identified several proteins lacking a signal peptide. For this reason, new recomplementation assays were carried out, and only the 83 clones that showed the best complementation of the Suc⁻ phenotype were selected (they were able to grow from 3 to 7 days of incubation and, more importantly, they showed homogeneous growth) (Fig. 2).

FIG. 2.

Analysis of the genetic screening. A scheme of the process used in the screening and statistical analysis of the data generated is shown. Tables I, II, III, and IV correspond to Tables 1, 2, 3, and 4, respectively.

Analysis of screening.

Plasmids from the 83 isolated clones (named Cep, for Candida exported protein) included the 5′ regions of 83 C. albicans putative genes in frame with suc2m^ic alleles that encoded different types of proteins (Fig. 2). Eleven of these proteins are already present in public databases (Table 1), while 52 share homology with proteins present in public databases (Tables 2 and 3). These correspond to genes already described (Table 2) as well as to putative ORFs determined by sequencing programs (orphan genes) (Table 3). We also found 5 clones that correspond to novel C. albicans ORFs without homologues in any of the public databases (Table 4) and 15 clones that do not match any C. albicans ORF identified at the Stanford Candida Genome Center (Assembly 6 Contig Index) (Table 5).

TABLE 1.

C. albicans selected proteins described previously

Category	Clone	Gene	aaa	Protein description	Localizationb	Signal peptidec	Reference(s) or source
Related to cell wall, membrane, and secreted proteins	Cep8	LIP4	24	Secretory lipase 4	Secreted	Yes	38
	Cep27	CHT1	30	Hydrolysis of chitin	Cell surface	Yes	49
	Cep33	PLB1	102	Phospholipase B	Secreted	Yesd	35, 47
	Cep58	XOG1	80	Exo-β-1,3-glucanase	Secreted	Yes	16
	Cep74	GSC1	28 (Met648)e	β-1,3-Glucan synthase, catalytic subunit (fragment)e	Plasma membrane, integral	No (fragment: Yes)e	51
	Cep123	SAP9	307	Aspartyl protease, GPI anchored	Cell wall or plasma membrane	Yesd	53
	Cep168	KRE9	38	β-1,6-Glucan synthesis	Cell surface	Yes	48
	Cep213	MNN9	94	N glycosylation of secreted and/or cell wall mannoproteins	Type II membrane protein, Golgi complex	Yesd	79
	Cep246	SAP10	48	Secretory aspartyl proteinase	Secreted	Yesd	AAF66711f
Related to morphological changes	Cep18	ECE1	39	Expression in cell elongation	ND	Yesd	7
	Cep68	OPS4	95	Opaque-phase specific	ND	Yes	55

^aaa, number of amino acid residues in the fusion that are derived from the C. albicans ORF.

^bDetermined experimentally or predicted by in silicio analysis. ND, not determined.

^cDescribed in the reference cited or in the corresponding SwissProt database or YPD entry, unless otherwise indicated.

^dThe signal peptide was predicted by in silicio analysis (see Materials and Methods).

^eThe fragment of GSC1 in clone Cep74 beginning at Met⁶⁴⁸.

^fDatabase entry.

TABLE 2.

C. albicans proteins homologous to already known proteins

Category	Clone	ORF (Stanford)	aaa	C. albicans signal peptideb	Gene homologous to:	Speciesc	E valued	Protein description or function	Localizatione	Homologue signal peptidef	Reference
Related to cell wall, membrane, and secreted proteins	Cep2	6.1639	353	Yes	CRH2	Sc	1.3e−103	Cell wall construction	Cell wall	Yes	68
	Cep3	6.8192	90	Yes	FET3	Ca	1.6e−154	Ferroxidase of the multicopper oxidase family	Plasma membrane	Yes	25
	Cep6	6.4841	121	Yes	STA1	Sc	3.1e−88	Glucoamylase precursor	Extracellular	Yes	89
	Cep10	6.1769	22	Yes	Lustrin A	Hr	7.4e−18	Matrix protein from shell and pearl nacre, modular structure	Extracellular	Yes	76
	Cep32	6.3505	313	Yes	CRH1	Sc	2.1e−67	Cell wall construction	Cell wall	Yes	68
	Cep64	6.5059	92	Yes	WSC3	Sc	3.1e−16	Maintenance of cell wall integrity and stress response	Plasma membrane	Yes	86
	Cep69	6.3209	32	Yes	HYR1	Ca	1.5e−93	Hyphally regulated	Cell wall	Yes	3
	Cep77	6.3113	22	Yes	AGA1	Sc	2.3e−11	Cell adhesion receptor, anchor subunit of a-agglutinin	Cell wall	Yes	71
	Cep78	6.3600	184	Yes	PIR3	Sc	3.7e−38	Structural cell wall protein	Cell wall	Yes	57
	Cep103	6.2855	114	Yes	ADP1	Sc	1.4e−286	Membrane transporter (ATP-binding cassette superfamily)	Plasma membrane, integral	Yes	22
	Cep111	6.857	325	No	HYR1	Ca	5.1e−171	Hyphally regulated	Cell wall	Yes	3
	Cep117	6.848	220	Yes	MID1	Sc	3.5e−72	Required for Ca2+ uptake and mating	Plasma membrane, integral	Yes	39
	Cep118	6.5424	315	Yes	YFW1	Sc	4.4e−42	Serine and threonine rich, homologue of pathogenesis-related proteins	Membrane	Yesg	88
	Cep119	6.5509	56	Yes	PHO2	Yl	4.2e−130	Acid phosphatase precursor	Secreted	Yes	84
	Cep125	6.1768	145	Yes	WSC2	Sc	1.2e−11	Cell wall integrity and stress response component 2	Plasma membrane	Yes	86
	Cep131	6.1832	150	Yes	PHO3	Sc	8.7e−108	Acid phosphatase with thiamine-binding activity	Periplasmic space	Yes	60
	Cep142	6.4807	277	No	mtr	Nc	1.8e−65	Amino acid permease	Plasma membrane	No	80
	Cep156	6.4005h	163	Yes	SCW4	Sc	3.1e−73	Potential glucanase	Cell wall	Yes	13
	Cep170	6.3969	361	Yes	PST1	Sc	1.4e−59	Cell wall generation	Cell surface	Yes	61
	Cep221	6.1240	83	Yes	SSP120	Sc	2.0e−46	Secretory protein SSP120 precursor	Secreted	Yes	77
	Cep231	6.4725	90	Yes	HYR1	Ca	1.3e−103	Hyphally regulated	Cell wall	Yes	3
	Cep242	6.5009	44	Yes	FET3	Ca	2.0e−202	Ferroxidase of the multicopper oxidase family	Plasma membrane	Yes	25
	Cep243	6.3942	90	Yes	ALX2	Sc	3.6e−103	Required for axial budding pattern	Plasma membrane	Yes	69
	Cep244	6.8973	76	Yes	KRE1	Ca	1.3e−12	β-1,6-Glucan synthesis	Cell surface	Yes	9
	Cep247	6.7473	66	Yes	YPS1	Sc	1.4e−20	GPI-anchored aspartyl protease	Plasma membrane	Yes	1/PICK>
	Cep258	6.6914	120	Yes	CSA1	Ca	2.5e−50	Mycelial surface antigen precursor, may function in host interaction	Surface	Yes	46
Secretory	Cep1	6.6467	88	Yes	SCJ1	Sc	1e−74	Chaperone, protein folding	ER, transmembrane	No	74
	Cep37	6.8569	68	Yes	ERV25	Sc	7.7e−61	Component of the COPII coat ofER-derived vesicles, complexes with Emp24p	ER, type I	Yes	5
	Cep38	6.7851	30	Yes	SLS1	Yl	5.6e−30	Involved in the protein translocation process	ER	Yes	8
	Cep61	6.230	324	Yes	CPY1CA	Ca	3.1e−148	Carboxypeptidase Y	Lysosome-like vacuoles	Yes	58
	Cep65	6.7407	234	Yes	OST3	Sc	4.4e−42	Oligosaccharyl transferase gamma subunit necessary for N glycosylation	ER	Yes	44
	Cep71	6.5557	270	Yes	GPI16	Sc	2.4e−112	Protein subunit of GPI transamidase complex	ER, transmembrane	Yes	28
	Cep85	6.5860	122	Yes	EMP24	Sc	5.8e−63	Component of the COPII coat of ER-derived vesicles, complexes with Erv25p	ER, type I	Yes	73
	Cep86	6.2244	33	Yes	EPS1	Sc	7.3e−72	Protein disulfide isomerase-related protein	ER, membrane	Yes	87
	Cep107	6.7797	125	Yes	SWP1	Sc	8.7e−21	Oligosaccharyl transferase delta subunit involved in N glycosylation	ER, transmembrane	Yes	83
	Cep109	6.4628	256	Yes	HRD3	Sc	3.6e−58	Protein degradation and translocation	ER, integral membrane	Yes	64
	Cep114	6.4218	58	Yes	CWH41	Sc	4.2e−60	Glucosidase I involved in N-glycan processing	ER, integral membrane	No	41
	Cep132	6.8096	288	Yes	PRB1	Sc	7.3e−159	Cerevisin precursor (protease B)	Vacuoles	Yes	52
	Cep223	6.4056	225	Yes	MPD1	Sc	1.2e−48	Suppresses the loss of protein disulfide isomerase	ER	Yes	82
Other	Cep101	6.7977	200	Yes	UTH1	Sc	2.0e−110	Aging protein also involved in the regulation of mitochondrion biogenesis	ND	No	11
	Cep250	6.9117	263	Not clear	PMP47B	Cb	1.3e−55	Probably encodes a substrate carrier	Peroxisomes, membrane	No	54

^aSee Table 1, footnote a.

^bThe signal peptide was predicted by in silicio analysis (see Materials and Methods).

^cCa, C. albicans; Cb, Candida boidinii; Nc, Neurospora crassa; Hr, Haliotis rufescens; Sc, S. cerevisiae; Yl, Yarrowia lipolytica.

^dThe expected value (E value) is a parameter that describes the number of hits that one can expect to see just by chance when searching a database of a particular size.

^eSee Table 1, footnote b.

^fDescribed in the reference cited or in the corresponding SwissProt database or YPD entry, unless otherwise indicated.

^gThe signal peptide was predicted by in silicio analysis (see Materials and Methods).

^hThe 5′ fragment isolated in clone Cep156 is nearly the same as orf6.4005 found in assembly 6 of the C. albicans sequence, but it has three repeats of an 18-bp sequence instead of the two repeats found in this ORF (L. Monteoliva, unpublished results).

TABLE 3.

C. albicans proteins homologous to unknown or putative proteins

Clone	ORF (Stanford)	aaa	C. albicans signal peptideb	Gene homologous toc:	Speciesd	E valuee	Protein description or functionf
Cep39	6.2232	22	Yes	YPR157w	Sc	4.4e−98	Unknown function
Cep90	6.1449	185	No	IMB2 (fragment)	Sp	1.6e−25	Importin-like protein (fragment)
Cep91	6.2324	54	Yes	YGL139w	Sc	1.6e−187	Strong similarity to YPL221w and YAL053w proteins
Cep102	6.6369	82	No	YHL017w	Sc	2.5e−78	Unknown function
Cep104	6.2554	90	No	YBR220c	Sc	2.9e−137	Similar to acetyl coenzyme A transporters
Cep112	6.9066	258	Yes	YHR151c	Sc	9.1e−44	Unknown function
Cep113	6.3650	127	Yes	YOR154w	Sc	1.1e−76	Unknown function
Cep130	6.8185	706	Yes	CG3376	Dm	3.1e−30	High homology to several sphingomyelin phosphodiesterases
Cep144	6.975	144	Yes	PRY1	Sc	3.2e−39	Similar to plant pathogenesis-related proteins
Cep211	6.1174	65	Yes	YAL053w	Sc	1.1e−193	Strong similarity to YOR365c, YGL139w, and YPL221w proteins
Cep254	6.6421	15	Yes	YDL072c	Sc	4.6e−38	Unknown function

^aSee Table 1, footnote a.

^bSee Table 2, footnote b.

^cw, Watson; c, chick.

^dDm, Drosophila melanogaster; Sc, S. cerevisiae; Sp, Schizosaccharomyces pombe.

^eSee Table 2, footnote d.

^fData were obtained from databases.

TABLE 4.

New C. albicans proteins without known homologues

Clone	ORF (Stanford)	aaa	C. albicans signal peptideb	GPI anchorc	Localizationd
Cep22	6.7675	87	Yes		Extracellular (including cell wall)
Cep60	6.4379	50	Yes		Extracellular (including cell wall)
Cep79	6.701	65	Yes	Yes	Plasma membrane or extracellular (including cell wall)
Cep106	6.7284	188	Yes	Yes	Plasma membrane or extracellular (including cell wall)
Cep248	6.8516	13	Yes		Extracellular (including cell wall)

^aSee Table 1, footnote a.

^bSee Table 2, footnote b.

^cSeems to be GPI anchored (predicted by the PSORTII program at the PSORT World Wide Web server).

^dMost likely location (predicted by the PSORTII program at the PSORT World Wide Web server).

TABLE 5.

N-terminal fragments of new putative C. albicans proteins

Clone	N-terminal sequence in fusiona	aab	Signal peptidec
Cep7	MMTILLTTTAAMTHHFLLLIAQYCHLN	27	Yesd
Cep9	MIVLLLYLLLFNIKNSKF	18	Yes
Cep14	MYLYPIPLLIGML	13	Not clear
Cep20	MLKLILQIKTLNCLLVYLLLLEP	23	Yes
Cep63	MSIIIQYLLLLLRCSIHLTLGELACGKQ	28	Yes
Cep80	MLSRFALVTFNFICSKISDMNRER	24	Yesd
Cep82	MIWLIELSLLVAMI	14	Yesd
Cep84	MYLCFYCSYLYLIGLF	16	Yesd
Cep88	MLFSLLMCWVHMMFMMVQEI	20	Yesd
Cep92	MLDRVTKVEQFTIIWGISLLGIILGNQFR	28	Nod
Cep229	MLFIYVQWLSSMLCIYFLLNRYSCQEI	27	Yes
Cep230	MKVFIKATSLSTLAVAISMSSMYMKIG	27	Yesd
Cep241	MIRSNIGKMVSGLLCKPVSKLPTTICLISIFTMLSHSIDDL	41	Mitochondrial
Cep249	MQIILFKYLLAACASNVYK	18	Mitochondrial
Cep251	MIGILMILQPLVVAVVVVVGILAI	24	Yesd

^aN-terminal sequence of the putative peptide encoded in frame with the suc2 allele, as determined with PCGENE software.

^bNumber of amino acid residues in the peptide encoded in frame.

^cThe signal peptide of the peptide encoded in frame was predicted by in silicio analysis.

^dResult of the prediction with the iPSORT WWW Service when this result was not clear with the SignalP program (see Materials and Methods).

(i) Previously identified C. albicans proteins.

As shown in Table 1, many of the previously identified proteins are related to cell wall biogenesis and include enzymes involved either in the hydrolysis or biosynthesis of major cell wall components, mainly chitin (Cht1p) and glucan (Xog1p, Gsc1p, and Kre9p), or in the proper glycosylation of cell wall mannoproteins (Mnn9p). In addition, and as expected, most of the isolated clones correspond to exported proteins whose localization is the cell surface. Some of the proteins are localized at the plasma membrane, such as Gsc1p/Fks1p (involved in glucan biosynthesis), while other clones identify enzymes that are secreted into the medium. Examples are the major glucanase in C. albicans cells, Xog1p, or members of families involved in the pathogenic process, such as the secretory lipase or the secretory aspartyl proteinase families (Lip4p, Sap10p, and Plb1p).

It is important to note that a signal peptide was detected in 10 of these 11 proteins. The remaining one (β-1,3-glucan synthase, Gsc1p) is a cytoplasmic membrane protein with no apparent signal peptide. Interestingly, although the insert isolated in clone Cep74 is short, in silicio analysis of the peptide starting at Met⁶⁴⁸ revealed that the insert could encode a functional signal peptide.

(ii) Proteins homologous to other proteins in databases.

The criterion for the inclusion of genes in this subset was a specific degree of homology determined by the E parameter of the BLAST algorithm. DNA sequences were submitted to the Stanford Candida Genome Center, and putative C. albicans ORFs were identified in the Assembly 6 Contig Index. Among the ORFs identified, we selected those that displayed homology to sequences encoding already described or putative proteins with an E value of less than e⁻¹⁰. These are listed in Table 2 (41 clones) or Table 3 (11 clones), respectively.

Twenty-six of the 41 proteins in Table 2 contain C. albicans peptides reported to be located at the cell surface, either as secreted proteins or as cell wall or plasma membrane components. Some of them are indeed proteins related to cell wall construction and/or to the maintenance of its integrity (i.e., Chr1p, Chr2p, Wsc3p, Wsc2p, PstIp, and Kre1p). Alternatively, they are structural cell wall component proteins (Pir3p). Most of these proteins (24 clones) display a clear signal peptide. An additional group of 13 clones correspond to C. albicans proteins whose homologues are located in the secretory pathway. Some of them are involved in the secretory mechanism that allows other proteins to reach the cell surface. For example, Scj1p, Mpd1p, and Hdr3p are involved in protein folding or are required when unfolded or misfolded proteins are introduced into the endoplasmic reticulum (ER), while Erv25p and Emp24p form a complex constituent of the COPII coat of certain ER-derived vesicles. Some proteins also located in the secretory pathway are involved in the N-glycosylation process of mannoproteins (Ost3p or Swp1p) or in GPI anchor synthesis (Gpi16p).

For 11 of the selected clones, the C. albicans fragment displays homology with novel proteins of unknown function, mainly identified through sequencing projects. These are summarized in Table 3. Interestingly, we cloned a homologue of the S. cerevisiae PRY1 gene—similar to genes encoding plant pathogenesis-related proteins—as well as two clones (Cep211 and Cep91) corresponding to proteins encoded by YAL053w and YGL139w, which are close homologues of each other. Although some of these proteins have been designated putative, it can be assumed that the C. albicans homologue can at least be expressed in this heterologous host.

(iii) New C. albicans proteins without known homologues.

Five cloned fusions contain C. albicans gene fragments that encode proteins without homologues in public databases. These proteins, which are completely novel and which, to date, have been detected only in this fungus, are listed in Table 4. In silicio analysis indicated that all of them do have a signal peptide, while ORF 6.701 (Cep79) and ORF 6.7284 (Cep106) encode proteins that seem to be GPI anchored. This result suggested their membrane or cell wall location. Furthermore, ORF 6.7284 encodes a serine- and threonine-rich protein (a feature characteristic of many cell wall proteins). The location of the other three proteins in Table 4 is also predicted to be extracellular.

Table 5 includes 15 clones with a C. albicans sequence that is present at the Stanford Candida Genome Center (Assembly 6 Contig Index) but for which no ORF was identified. For all of them, a peptide encoded by a sequence in frame with the corresponding suc2m^ic allele was detected, and in some cases a signal peptide was predicted. Further work is needed to determine whether they really encode a functional signal peptide for a C. albicans secreted protein.

DISCUSSION

The availability of the C. albicans genome is having a profound impact on the way in which biological research is being carried out with this fungal pathogen, especially in view of the difficulties involved in C. albicans genetic studies (19, 63). While the amount of information is significant and while several new genes are being identified through global sequencing programs (9,168 ORFs in the Stanford Assembly 6 Contig Index of the C. albicans sequence), there is a significant need to define the functionality of the encoded proteins, especially from the perspective of finding novel cell wall components. Given the ability of several C. albicans genes to be expressed in a heterologous host (63), the complementation of specific S. cerevisiae mutants altered in cell wall functions is an attractive and feasible approach. Global functional strategies have been carried out with S. cerevisiae (for example, using a bacterial β-galactosidase to monitor both gene expression and subcellular localization) (10, 70); however, for C. albicans, only one recent report, using an antisense approach, has addressed the identification of essential functions (20) in this organism. In the present work, by making use of S. cerevisiae genetic tools, we carried out a genetic screening with the primary goal of identifying C. albicans secretory protein-encoding genes and/or export domains within them. We used the invertase encoded by the SUC2 gene, since this strategy combines both a functional domain identification feature and a positive selection scheme that facilitates subsequent cloning steps. We followed a strategy similar to that already described for the isolation of mammalian exported proteins (45). We thus avoided the use of C. albicans genetic tools, which can be time-consuming.

As expected, we selected a significant number of genes encoding extracellular proteins (39 clones, Tables 1, 2, and 3; Fig. 2 shows localization by description, homology, or in silicio analysis). Some of the proteins have already been shown to be located in the fungal cell wall (eight clones) and to be involved either in the biosynthesis of the major cell wall components glucan and chitin or in the maintenance of this cellular structure. Proteins whose final destination is the cytoplasmic membrane (11 clones) or that are destined for secretion (9 clones) were also identified, since these are apparently able to enter the secretory pathway. They include membrane receptors and transporters and extracellular enzymes. In some cases, the true subcellular location of the identified protein is not well defined, although it is known or predicted to be the cell surface (11 clones). The five new C. albicans proteins described in Table 4 are included in this group. Identification of these proteins, which, to date, have been considered only putative in the C. albicans sequencing project at the Stanford Genome Technology Center, would be one of the more relevant results of the screening. We observed that the putative ORFs encode expressed proteins, at least in S. cerevisiae. The fact that these proteins have no homologues in databases suggests that they are species-specific proteins, indicating their involvement in processes specific to C. albicans biology, such as dimorphism or virulence.

A different but important group of cloned genes encode proteins located in the secretory machinery. This finding was expected, since these proteins have a signal peptide responsible for their entry into the ER and for reaching their final cellular location. The homologues of most of the proteins found are located in the ER. This finding can be explained in terms of the notion that they are isolated as truncated chimeras and are therefore devoid of the four amino acids that contain the ER carboxy-terminal retention signal (HDEL) (62). Finally, the clones in Table 5 did not correspond to any ORF identified at the Stanford Candida Genome Center (Assembly 6 Contig Index). Among the possible explanations are that they arise from mistakes in the Stanford database, that they are ORFs able to encode proteins shorter than 100 amino acids, or even that they are spliced genes.

It should be stressed that the assignment of proteins or functions between even closely related microorganisms is risky. The assignment of homology in Tables 2 and 3 was made based on computer algorithms and, in some cases, the C. albicans protein may not represent the counterpart of the described protein. For example, the best result in the homology search with Cep101 was UTH1. However, the protein showed homology with all members of the SUN family (56). This family has four members with many different cellular functions. Sun4p, one of its members, has been described as a cell wall protein with a clear signal peptide, as was also observed for the protein corresponding to the C. albicans clone identified in this screening (Cep101). Therefore, a complete functional analysis may change this in silicio assignment. In addition, the subcellular location of any given protein may vary from host to host.

There are some limitations to this screening. One is that it relies on heterologous expression, and although C. albicans genes are normally expressed in S. cerevisiae (63), the level and/or pattern of the time dependence of expression of these genes may differ between the organisms, resulting in a lack of complementation. A few modifications to this screening might expand the numbers of putative exported proteins identified. For example, larger genomic fragments could be selected to avoid the cloning of genes for small peptides (such as the clones in Table 5), which would hinder interpretation of the results. The construction of new genomic libraries (in which genomic DNA is digested with different restriction enzymes or even randomly) should allow the cloning of more new genes. Expression libraries could also be used. In this case, the entire protein, with most of the location domains (except for C-terminal domains such as HDEL or the GPI anchor), would be expressed.

Despite these limitations, the screening carried out here is extremely useful for several reasons. First, it allowed the selection of 5 novel interesting ORFs from among more than 9,000 ORFs in the C. albicans genome. Second, it is a broad-range screening; a similar screening for secretory protein-encoding genes of S. cerevisiae based on the functional selection of fusions with a gene reporter (PHO5) was reported several years ago (77). In this case, the authors described the isolation of five unique sequences and the complete SSP120 gene. Significantly, the C. albicans SSP120 homologue was also obtained in our approach. Third, the SUC2 system has afforded new data about the localization of previously known proteins relevant to different physiological processes but whose localization (measured as the ability to export the intracellular invertase) was not known. These include some proteins relevant to the dimorphic transition, such as Ece1p, the sequence of which was cloned by using a cDNA differential screening (7), or phenotypic switching, such as Ops4, a protein differentially expressed in the opaque phase (55). Finally, the screening was able to select proteins that showed no obvious signal peptide in their amino-terminal regions but which were indeed extracellular proteins. The presence of this kind of protein—such as glycolytic enzymes—within the fungal cell wall has remained controversial for some time (26, 75), although it was recently shown that yeast enolase, encoded by the ENO2 gene, can be efficiently incorporated into regions external to the plasma membrane (61).The use of this system for other genetic elements, such as transposons, may provide an additional useful tool for C. albicans genetic studies.