Identification of Candida albicans ALS2 and ALS4 and Localization of Als Proteins to the Fungal Cell Surface

Abstract

Additional genes in the growing ALS family of Candida albicans were isolated by PCR screening of a genomic fosmid library with primers designed from the consensus tandem-repeat sequence of ALS1. This procedure yielded fosmids encoding ALS2 and ALS4. ALS2 and ALS4 conformed to the three-domain structure of ALS genes, which consists of a central domain of tandemly repeated copies of a 108-bp motif, an upstream domain of highly conserved sequences, and a domain of divergent sequences 3′ of the tandem repeats. Alignment of five predicted Als protein sequences indicated conservation of N- and C-terminal hydrophobic regions which have the hallmarks of secretory signal sequences and glycosylphosphatidylinositol addition sites, respectively. Heterologous expression of an N-terminal fragment of Als1p in Saccharomyces cerevisiae demonstrated function of the putative signal sequence with cleavage following Ala17. This signal sequence cleavage site was conserved in the four other Als proteins analyzed, suggesting identical processing of each protein. Primary-structure features of the five Als proteins suggested a cell-surface localization, which was confirmed by indirect immunofluorescence with an anti-Als antiserum. Staining was observed on mother yeasts and germ tubes, although the intensity of staining on the mother yeast decreased with elongation of the germ tube. Similar to other ALS genes, ALS2 and ALS4 were differentially regulated. ALS4 expression was correlated with the growth phase of the culture; ALS2 expression was not observed under many different in vitro growth conditions. The data presented here demonstrate that ALS genes encode cell-surface proteins and support the conclusion that the size and number of Als proteins on the C. albicans cell surface vary with strain and growth conditions.

The opportunistic pathogen Candida albicans is a fungus that exists in a diverse range of associations with its human or animal host. C. albicans can survive in the host without overt disease symptoms and, under the appropriate circumstances, can cause disease that varies in site and severity. C. albicans infection can be localized and superficial or systemic and disseminated to a wide range of organs (reviewed in reference 43). This complexity of interactions of C. albicans with its host suggests that the fungus possesses numerous mechanisms to adapt to this diversity of host sites, a versatility undoubtedly controlled by differential gene expression. Much investigative effort has focused on defining molecular mechanisms C. albicans uses for growth and pathogenesis. One attribute of the fungus that is positively correlated with pathogenicity is adherence (reviewed in reference 7). Adherent strains of C. albicans are more virulent than those with a less-adhesive nature. Additionally, a hierarchy exists among Candida species, with the more frequently isolated pathogenic species exhibiting greater adhesive capacity (reviewed in reference 7). Adherence of C. albicans to host surfaces is also involved in the process of colonization, which may occur without accompanying pathogenesis. Investigations to understand C. albicans adhesion have involved characterization of the cell surface, since this is the initial point of contact between fungus and host (reviewed in references 7, 9, 17, 23, and 29). Adherence of C. albicans to numerous cell types, cellular components, and nonliving substances has been examined to further define adhesive relationships. Numerous obstacles exist to hamper the study of C. albicans adhesion, including widely noted growth-medium-dependent effects and differences between C. albicans strains tested in the same assay (reviewed in references 15, 30, and 45). Despite these obstacles, several C. albicans molecules involved in adhesive interactions have been identified (reviewed in references 7, 9, 17 and 23).

Characterization of the C. albicans ALS gene family has yielded data that address the major themes discussed above. The first gene in the ALS family, ALS1, was isolated in a differential screen to identify hypha-specific genes (26). Although subsequent studies demonstrated that ALS1 is not strictly hypha-specific, its sequence has significant identity with the sequence of AGα1 from Saccharomyces cerevisiae, which encodes α-agglutinin, a cell-surface adhesion glycoprotein that facilitates contact between haploid cells during mating (19, 37). Because C. albicans has not been observed to undergo meiosis or mating, it may be less likely that the function of Als1p is directly analogous to that of Agα1p (26). However, conservation between sequences required for the adhesive function of α-agglutinin and those at the N terminus of Als1p raised the intriguing possibility that Als1p is an adhesion glycoprotein (26). Data supporting this conclusion have been published recently (16, 18).

In addition to its potential to encode an adhesion glycoprotein, other features of ALS1 prompted further study. ALS1 encodes a central domain of tandemly repeated copies of a highly conserved 108-bp sequence that, when translated, predicts a highly conserved 36-amino-acid motif (26). The tandem-repeat sequence hybridizes to several genomic fragments from C. albicans, suggesting that ALS1 belongs to a gene family (26). The existence of a gene family defined by the tandem-repeat-hybridizing fragments was demonstrated by the characterization of ALS3 (25). The size of the ALS family is difficult to estimate because of the presence of additional tandem-repeat-hybridizing fragments and of other genomic sequences that hybridize to a probe derived from the 5′ end of ALS1 (25, 26).

Experiments to characterize the ALS genes and to understand their regulation were initially undertaken to lay the groundwork for studying Als protein function in C. albicans. Studying Als protein function in C. albicans is challenging because, assuming redundancy of function among proteins in the family, creation of a truly null mutant requires characterization of the entire family and disruption of many genes. We reasoned that the number of gene disruption steps could be reduced by knowing which genes were expressed under a particular growth condition. Combining a particular growth condition with specific disruptions could effectively create a null mutant. Studies of ALS gene regulation demonstrated that ALS1 and ALS3 are differentially expressed (25, 26). ALS1 expression in vitro is regulated by components of growth media, and ALS3 is hypha-specific (25, 26). In addition, expression of ALS1 was shown to vary among strains of C. albicans (25).

In this study, we present data to further characterize the ALS genes and their encoded proteins. Two new genes, ALS2 and ALS4, are described. Similar to previously characterized ALS genes, ALS2 and ALS4 are shown to be differentially regulated. Comparison of ALS gene sequences yielded a generalized ALS gene structure that fits another C. albicans gene, ALA1 (18). Here, we recognize the place of ALA1 in the ALS family. Analysis of sequence features of the five predicted Als proteins suggests that they are localized on the C. albicans cell surface, a property demonstrated by indirect immunofluorescence with an anti-Als antiserum. Taken together, the data presented here demonstrate that the cell-surface-localized Als proteins could account for a significant portion of the strain- and growth-medium-dependent differences in adhesion commonly noted in the C. albicans literature.

MATERIALS AND METHODS

Media and strains.

All standard growth media and strains were described previously (25, 26). C. albicans B311 used in this study was purchased from the American Type Culture Collection; other isolates of B311 used in previous studies were not used here.

Library screening and DNA sequencing of ALS genes.

Construction of the fosmid library and PCR screening with primers specific for the consensus tandem-repeat sequence were described previously (25). Fosmids that were positive in the PCR screen were grouped on the basis of PCR products (25); a representative fosmid was chosen from each group for subcloning and DNA sequencing of each ALS allele. Fosmids chosen were 19F-1 (ALS2-1), 20F-3 (ALS2-2), 20E-6 (ALS4-1), and 29F-9 (ALS4-2). These fosmids were purified and digested with a variety of restriction enzymes. Southern blots of these digests were probed with an 870-bp KpnI fragment from ALS1 that encodes only tandem-repeat sequences (26) to identify fragments likely to encode related ALS genes. These fragments were subcloned into pUC vectors (57) and transformed into Escherichia coli DH5αMCR (Gibco BRL). Plasmid purification and DNA sequencing were conducted as described previously (26). DNA sequencing of the tandem-repeat regions of ALS genes was limited to that required to determine that the tandem repeats conformed to the consensus pattern derived from ALS1 and that the tandem-repeat domain did not encode anything besides head-to-tail copies of the 108-bp sequence. This was accomplished with a combination of repeat-region subclones and production of nested deletions with a double-stranded nested deletion kit (Amersham Pharmacia Biotech). Sequences were reported as domains 5′ and 3′ of the tandem-repeat domain. Both alleles of ALS2 and ALS4 from C. albicans 1161 were sequenced and deposited separately. Accession numbers for other sequences discussed here include L25902 (ALS1 [26]), U87856 (ALS3 [25]), and AF025429 (ALA1/ALS5 [18]).

ALS2- and ALS4-specific probes.

A 164-bp fragment located immediately 3′ of the tandem-repeat domain was amplified by PCR and found to detect both ALS2 and ALS4. This PCR fragment was amplified with the forward primer 5′ TCCGAGTCCATTCCAGTACTAA 3′ and the reverse primer 5′ GTTACAGCATCACTAGAAGGAATATC 3′. The standard PCR for these primers and others described below included 1 μM (each) primer, 0.2 mM (each) deoxynucleoside triphosphate, 1.5 mM MgCl₂, 1× Promega PCR buffer, 2.5 U of Promega Taq DNA polymerase, and 10 to 100 ng of template DNA.

Since there was a high degree of identity between ALS2 and ALS4 in the domain 3′ of the tandem repeats, this region could not be exploited to define gene-specific probes as it had been for ALS1 (26) and ALS3 (25). Because of the high degree of nucleotide sequence conservation between all ALS genes within the domain 5′ of the tandem repeats, ALS2 and ALS4 could be specifically detected only with oligonucleotide probes. The resulting ALS2-specific probe, 5′ TAGTTCCTTACAAAGTAAGCCGTTCAATTT 3′ (located at nucleotide 897 in the ALS2-coding region), and the ALS4-specific probe 5′ CGCGGCTTCTGTTGATGACTCATTTACTCATACT 3′ (located at nucleotide 900 of the ALS4-coding region) were used in Southern blotting. The reverse complement of each probe was synthesized for use in Northern blotting as described below.

ALS5 probe.

A probe encoding only tandem-repeat sequences from ALA1/ALS5 (18) was synthesized by PCR with the forward primer 5′ GGTACAAGTTCCACTGCCAAA 3′ and the reverse primer 5′ AAGACAGTTCTTCCAATGGATCA 3′. These primers amplified two products from genomic DNA of C. albicans 1177, one at approximately 200 bp and the other at approximately 680 bp. These two products were presumably due to amplification of tandem-repeat regions from each allele of ALS5 in this strain. The 680-bp fragment was cloned into pCR2.1 (Invitrogen) and transformed into E. coli TOP10F′ cells (Invitrogen). The DNA sequence of the cloned PCR fragment conformed to the consensus tandem-repeat sequence of ALS5 (18).

Nucleic acid blots.

Southern blotting was performed as described previously (26) with the digoxigenin nonradioactive nucleic acid labeling and detection system (Boehringer Mannheim). Blots probed with the ALS1 or ALS5 tandem-repeat sequence were hybridized at 65°C overnight. Blots were washed in 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–0.1% sodium dodecyl sulfate (SDS) at room temperature for 30 min and then with 0.5× SSC–0.1% SDS at 65°C for 1 h.

Cultures of strains SC5314 and 3153A were used to demonstrate the growth-stage-associated expression of ALS4 on Northern blots. Cells from an overnight YPD (yeast extract, peptone, dextrose) culture were washed twice in sterile water and inoculated into 500 ml of fresh YPD at 5 × 10⁶ cells/ml. An aliquot of this culture was removed to extract total RNA for the 0 h time point. Cells for RNA preparation were washed twice with diethylpyrocarbonate-water, flash frozen in an ethanol dry-ice bath, and stored at −80°C until RNA was extracted. The freshly inoculated YPD culture was incubated at 30°C, with shaking at 200 rpm. Samples were removed every hour for 8 h and processed for RNA extraction as described above. Three or four independent cell counts were performed at each time point to construct a growth curve.

Total RNA extraction, formaldehyde gel electrophoresis, and Northern blotting were performed as described previously (25). Fifty micrograms of total RNA was loaded into each gel lane. ALS4-specific message was detected with the end-labelled ALS4-specific oligonucleotide. Hybridization with oligonucleotide probes followed the method of Sundstrom et al. (52). Equal loading of total RNA on Northern blots was evaluated with a fragment from the C. albicans TEF1 gene (52) as previously described (26).

Production of anti-Als antiserum.

Hyperimmune anti-Als serum raised in a New Zealand White rabbit was a gift from George Livi (SmithKline Beecham Pharmaceuticals). The anti-Als serum was raised against four 10-mer peptides derived from the N-terminal domain of Als1p (26). These peptides were chosen because they were likely to be in a hydrophilic, surface-exposed region of the mature, folded protein as predicted by secondary-structure algorithms (14). Peptides from the N-terminal region of Als1p were selected because this portion of the protein was predicted to be relatively free of glycosylation compared to other regions of the protein (26). The peptides selected were GWSLDGTSAN (amino acids 53 to 62), FYSGEEFTTF (amino acids 98 to 107), TGSSTDLEDS (amino acids 139 to 148), and NTVTFNDGDK (amino acids 156 to 165). These peptides were linked to keyhole limpet hemocyanin (KLH) and combined in equal quantities prior to emulsification in Freund’s complete adjuvant (Sigma). The emulsion was injected into the rabbit at multiple subcutaneous sites. A blood sample was collected from the marginal ear vein 14 days later. A booster immunization was administered 4 weeks after the initial immunization. The booster immunization was performed with the mixture of four KLH-linked peptides emulsified in incomplete Freund’s adjuvant (Sigma). A blood sample was collected 14 days after the booster immunization, and the anti-Als titer was assayed on a Western blot of a heterologously produced soluble N-terminal fragment of Als1p (see below). Four total-booster injections were performed, with the anti-Als titer increasing following each round. Increasing titer was judged by increasing dilutions of serum required to obtain an equivalent Western blot signal. Serum collected from the rabbit was stored in small aliquots at −80°C. Preimmune serum collected from the same rabbit in which the anti-Als serum was raised and a commercially purchased anti-KLH serum (ICN) were both utilized as negative controls.

Indirect immunofluorescence of C. albicans cells.

Cells of strain SC5314 were grown in YPD until they reached late stationary phase; at this stage of growth, cultures typically have a density of approximately 5 × 10⁸ cells/ml. Cells from this culture were washed twice in phosphate-buffered saline (PBS) (per liter: 10 g of NaCl, 0.25 g of KCl, and 1.43 g of Na₂HPO₄ [pH 7.2 to 7.3]) and counted. A fresh culture of RPMI 1640 (catalog no. 11875-085; Gibco BRL) was inoculated at a density of 5 × 10⁶ cells/ml. This culture was incubated at 37°C and 120 rpm for 1 h 45 min. One hundred microliters of this culture was spread into an area of a cleaned glass slide that had been delineated by etching with a diamond pen. Slides were washed thoroughly in PBS before we proceeded. Slides were blocked with 200 μl of 1.5% normal goat serum (Jackson Research Laboratories) diluted in PBS and incubated for 10 min at room temperature in a humid chamber. Excess normal serum was drained from slides, and 200 μl of a primary antibody solution was added. Two different primary antisera were used, a 1:100 dilution of the anti-Als serum purified on a protein G column according to the manufacturer’s instructions (MAbTrap G II column; Amersham Pharmacia Biotech) and a 1:500 dilution of the preimmune serum from the rabbit in which anti-Als serum was raised. Immunoglobulin G (IgG) concentrations of these preparations were roughly equivalent; a lower dilution of the protein-G-purified serum was used to account for the dilution that occurs during the purification procedure. Primary antiserum was incubated on slides for 1 h at 4°C in a humid chamber. Slides were washed thoroughly in ice-cold PBS. The secondary antibody, fluorescein-isothiocyanate (FITC)-conjugated goat anti-rabbit IgG (heavy plus light chains) (catalog no. 111-095-003; Jackson Research Laboratories) was diluted 1:2,500 from a 0.75 mg/ml stock and incubated on slides for 1 h at 4°C in a humid chamber. Slides were washed thoroughly in ice-cold PBS and stored in darkness at 4°C until they were viewed later the same day.

Immunofluorescence images were obtained with an Olympus BX60 fluorescence microscope with an oil immersion lens at approximately 400× magnification. Images were captured with a Photometrics Ltd. system consisting of a charge-coupled device camera (model CH250), an electronic unit (model CE 200A, equipped with a 50-Hz 16-bit A/D converter), and a controller board (model NU 200). Images were acquired and evaluated with Adobe Photoshop software and a Macintosh Quadra 840 AV computer (Apple Computer, Inc.).

Heterologous expression of an N-terminal fragment of Als1p in S. cerevisiae.

A PCR product encoding the N-terminal 433 amino acids of Als1p was produced with the forward primer 5′ CCCCCCCATGGTTCAACAATTTACATTGTTATTCCTATA 3′ and the reverse primer 5′ CCCCCGTCGACCAGTGGAACTTGTACCACCACTGTGTCA 3′. The PCR product derived from C. albicans B311 genomic DNA was digested with NcoI and SalI and cloned into the NcoI-SalI-digested S. cerevisiae expression vector p138NB. Features of this vector have been described previously (38, 42). Briefly, the expression vector contains the TRP1 selectable marker and partial 2μm sequences for maintenance at high copy number. Expression is driven by the copper-inducible CUP1 gene promoter. A multiple cloning site is present in the vector downstream of the CUP1 promoter and upstream of the CYC1 transcription terminator. Shuttle vector functions allowed a correct construct, pLH109, to be selected in E. coli DH5αMCR (Gibco BRL). Plasmid pLH109 was transformed into the S. cerevisiae strain YPH 274 (a/α ura3-52/ura3-52 lys2-801^amber/lys2-801^amber ade2-101^ochre/ade2-101^ochre trp1-Δ1/trp1-Δ1 his3-Δ200/his3-Δ200 leu2-Δ1/leu2-Δ1 [50]). The resulting S. cerevisiae strain was grown in synthetic complete medium without tryptophan (22). A 500-ml culture was grown to late stationary phase at 30°C. Cells were harvested from this culture, washed twice in fresh synthetic complete medium without tryptophan, and resuspended in 30 ml of fresh medium. Expression from the CUP1 promoter was induced by the addition of 150 μM CuSO₄. Induced cells were grown for 4 h at 30°C with 200 rpm shaking. Cells were collected by centrifugation, and the resulting culture supernatant was brought to 70% saturation with ammonium sulfate to precipitate proteins, including the secreted N-terminal portion of Als1p. The ammonium sulfate precipitation was incubated overnight at 4°C on a rocker platform and then harvested by centrifugation at 15,000 rpm (31,000 × g) for 30 min in a Beckman JA-17 rotor. Supernatant was decanted, and the precipitate was dissolved in PBS. The dissolved precipitate was thoroughly dialyzed against PBS in 12,000- to 14,000- molecular-weight-cutoff dialysis tubing (Spectrum Laboratories, Inc.). The dialysate was concentrated in a Centricon-10 unit (Amicon, Inc.) and run on a 12.5% Tris-glycine polyacrylamide gel (31). S. cerevisiae YPH274 transformed with blank p138NB plasmid was processed concurrently as a negative control. Coomassie blue R-250 (Sigma) and silver staining (Bio-Rad) indicated the presence of a single protein band at approximately 65 kDa in samples from YPH274(pLH109) that was not present in YPH274(p138NB). This band was presumed to be the N-terminal portion of Als1p. Its identity was confirmed by Western blotting with the anti-Als antiserum described above.

N-terminal amino acid sequencing of the Als1p-derived fragment.

The 65-kDa band in the supernatant of the YPH274(pLH109) culture was well separated from the sparse number of protein bands present on the acrylamide gel. To determine the N-terminal amino acid sequence of the 65-kDa Als1p fragment, a 12.5% polyacrylamide gel of supernatant sample was run. Proteins from the gel were electroblotted to a polyvinylidene difluoride membrane (Gelman Sciences, Inc.) with a 10 mM CAPS [3-(cyclohexylamino)-1-propanesulfonic acid] (Sigma) (pH 11.0)–10% methanol buffer in a Bio-Rad TransBlot apparatus. The polyvinylidene difluoride membrane was rinsed thoroughly with deionized water and stained with 0.2% Coomassie blue R-250 in 45% methanol–10% acetic acid. The membrane was destained with 50% methanol–1% acetic acid and air dried, and the Als1p-derived fragment was excised with a new razor blade. N-terminal amino acid sequencing was performed by the Iowa State University Protein Facility with a model 492 Procise protein sequencer and a model 140C analyzer (Applied Biosystems, Inc.).

Nucleotide sequence accession numbers.

Both alleles of ALS2 and ALS4 from C. albicans 1161 were sequenced and deposited separately. GenBank accession numbers are AF024580 (5′ domain) and AF024581 (3′ domain) (ALS2-1), AF024582 and AF024583 (ALS2-2), AF024584 and AF024585 (ALS4-1), and AF024586 and AF024587 (ALS4-2).

RESULTS

ALS2 and ALS4 complete the set of genes that hybridize with the ALS1 tandem-repeat probe.

An 870-bp KpnI fragment derived from within the tandem-repeat domain of ALS1 hybridizes with several genomic fragments from C. albicans and C. stellatoidea; the number of hybridizing fragments depends upon the strain examined (26). The 10 copies of the 108-bp tandem-repeat element from ALS1 in strain B792 were aligned to derive a consensus sequence (26). PCR primers based on the most-conserved regions of the tandem-repeat sequence were used to screen a fosmid library from C. albicans 1161 (25). By this technique, fosmids were noted to yield different PCR product patterns and were grouped accordingly (25). Characterization of a representative fosmid from one of the groups yielded ALS3 (25); fosmids from the other groups encoded alleles of two other ALS genes, designated ALS2 and ALS4. Restriction enzyme analysis and hybridization with ALS-gene-specific probes indicated that ALS2 and ALS4 accounted for the remaining fragments that hybridize with the ALS1 tandem-repeat probe in genomic DNA from C. albicans 1161 (Fig. 1). A similar analysis with 12 C. albicans and two C. stellatoidea strains indicated that, in each strain, all four ALS genes are present and account for all of the restriction fragments that hybridize with the ALS1 tandem-repeat probe (data not shown). Varying numbers of tandem-repeat-hybridizing fragments originally noted in each strain (26) were due to differences between allelic fragments encoding the same gene and to comigration of larger restriction fragments (data not shown).

FIG. 1.

Designation of genomic DNA fragments that encode ALS genes. Southern blots of BglII- or HindIII-digested genomic DNA from C. albicans 1161 were hybridized with an 870-bp KpnI fragment from ALS1 that contained only tandem-repeat sequences (26). ALS alleles corresponding to each of these restriction fragments were identified on separate blots with gene-specific probes developed previously (25, 26) or in this study. Allelic fragments were deduced by data from several experiments and sources, including physical mapping of C. albicans chromosomes (11, 48), restriction mapping of specific fosmid clones, DNA sequence analysis, and gene regulation studies (25, 26). Molecular size markers (in kilobases) are included between the blots.

The DNA sequence was derived for both alleles of ALS2 and ALS4. ALS2 and ALS4 conformed to the basic three-domain structure of ALS genes, which includes a central domain of varying numbers of copies of a tandemly repeated 108-bp sequence, a 5′ domain that is approximately 1.3 kb in length and conserved between ALS genes, and a 3′ domain that is variable in length and sequence (25). Characterization of fosmids encoding ALS2 and ALS4, as well as physical mapping efforts in this lab and others, suggested that ALS2 and ALS4 were located approximately 70 kb from each other on chromosome 6, SfiI fragment C (data not shown; 48).

Allelic differences have been noted previously for other ALS genes, although this has been mainly in the tandem-repeat region, where alleles have been demonstrated to encode different numbers of head-to-tail copies of the 108-bp motif (26) (Table 1). In the current study, alleles of ALS2 and ALS4 were sequenced to study conservation of nucleotides in the non-tandem-repeat domains. In these other two domains, alleles of ALS2 and alleles of ALS4 were more than 97% identical in nucleotide sequence (see below). Although the complete tandem-repeat region in both ALS2 and ALS4 was not sequenced, sufficient sequencing was completed in each allele to determine that the tandem-repeat region conformed to the consensus tandem-repeat sequences derived for ALS1 and ALS3 and to confirm that only tandem-repeat sequences were present (data not shown). Sequencing of large regions of tandemly repeated DNA has been omitted in other studies such as the S. cerevisiae genome project (28).

TABLE 1.

Comparison of Als protein features from predicted amino acid sequences

Als protein feature	Predicted amino acid sequence
	Als1pa	Als2-1p	Als2-2p	Als3p	Als4-1p	Als4-2p	Als5pa
No. of amino acids
N-terminal domain	433	432	432	433	433	433	433
C-terminal domain	470	372	373	293	372	373	770
No. of consensus N glycosylation sitesb
N-terminal domain	0	2	2	0	0	0	0
C-terminal domain	6	4	4	4	4	4	4
% Serine and threonine
N-terminal domain	28	27	27	27	31	31	27
C-terminal domain	45	36	37	41	37	36	45
No. of repeat copiesc	10	35	26	10	33	23	6
Size of nonglycosylated protein (kDa)d	133	218	183	120	209	171	150

^aThe Als1p sequence is from C. albicans B792 (26); the Als5p sequence (18) is from a clinical isolate. All other sequences are from C. albicans 1161. In strain 1161, ALS3 alleles are similar in size, but ALS1 alleles are detectably different lengths (Fig. 1). 

^bAsn-X-Ser/Thr, where X is any amino acid except proline (2). Consensus N glycosylation sites are abundant in the tandem-repeat domain; the consensus tandem-repeat sequence contains one consensus N glycosylation site per copy (26), suggesting that the tandem repeats are highly N and O glycosylated. 

^cValues for the sequences of Als1p, Als3p, and Als5p are exact. Values for Als2p and Als4p were estimated from DNA sequence and detailed restriction mapping information. 

^dValues for Als1p, Als3p, and Als5p are exact. Values for Als2p and Als4p were estimated as follows: (predicted molecular weight of N-terminal domain) + (predicted molecular weight of C-terminal domain) + (number of tandem-repeat copies) (molecular weight of consensus Als1p tandem-repeat sequence). 

Designation of C. albicans ALS5 and its place in the ALS family.

Characterization of ALS1, ALS2, ALS3, and ALS4 yielded a generalized three-domain structure for ALS genes. Another recently characterized C. albicans gene, called ALA1 (18), also fits this basic motif and belongs to the ALS family. Because the name ALA1 has been previously used to denote an alanyl tRNA synthetase in S. cerevisiae (44), and because C. albicans nomenclature follows the S. cerevisiae precedent (48), we propose that the gene described by Gaur and Klotz (18) be called ALS5.

ALS5 was not identified in the original screening of genomic DNA by hybridization with the ALS1 tandem-repeat probe. Therefore, although both genes encode a domain of tandem repeats, the exact tandem-repeat sequences were sufficiently dissimilar to escape detection by cross-hybridization under the experimental conditions employed. Comparison of the consensus tandem-repeat sequences of ALS1 (26) and ALS5 (18) indicated matches between only 39% of the nucleotides (data not shown). Because hybridization between two sequences depends on matches between individual tandem-repeat copies rather than an idealized consensus, Southern blotting was done to see if any genomic fragments hybridized with both probes. An ALS5 tandem-repeat probe was amplified by PCR with primers that flank the tandem-repeat region in this gene (18). The DNA sequence of this probe fragment indicated that it closely matches the consensus sequence of the tandem repeats of ALS5 in the clinical isolate used by Gaur and Klotz (data not shown; 18). Southern blots of BglII-digested genomic DNA from C. albicans and C. stellatoidea strains indicated that the ALS1 tandem repeats hybridized to a different set of genomic fragments than did the ALS5 tandem repeats in most strains (Fig. 2). Based on this result, the small number of fragments of similar size that were detected with both probes most likely encode different genes. Additional ALS genes have been isolated which possess the general ALS three-domain structure but which have more sequence similarities to ALS5 than to ALS1. Analysis of these gene sequences suggested that variability in the tandem-repeat sequences was a potential criterion for the division of the ALS family into subfamilies (24).

FIG. 2.

Southern blots of BglII genomic DNA fragments hybridized with ALS1 and ALS5 tandem-repeat probes. Genomic DNA from a variety of C. albicans and C. stellatoidea strains was digested with BglII, blotted and probed with an 870-bp KpnI ALS1-tandem-repeat-specific probe (left panel). The blot was stripped and reprobed with a PCR-amplified ALS5 tandem-repeat fragment PCR amplified from C. albicans 1177 (right panel). Molecular size markers (in kilobases) are indicated at the left for each blot.

Sequences in the N-terminal domain of Als proteins are highly conserved and encode a secretory signal peptide.

Sequences N-terminal of the tandem-repeat domain were highly conserved in each predicted Als protein, with amino acid identity between different proteins ranging from 68 to 86% (Fig. 3). Identity of the nucleotide sequences encoding the N-terminal domain was 73 to 90%. At the start of each coding region was a hydrophobic sequence with hallmarks of a secretory signal peptide (54). To test whether the hydrophobic N terminus functioned as a signal sequence, a construct expressing the N-terminal 433 amino acids of Als1p under control of the CUP1 promoter was transformed into S. cerevisiae (see Materials and Methods). Supernatants from cultures of the CuSO₄-induced construct and a control strain harboring the blank expression plasmid were run on SDS-polyacrylamide gels (31). Coomassie blue and silver staining of these gels indicated that a major protein species was present at approximately 65 kDa in supernatant from cells with the Als1p construct and absent from cells transformed with blank vector (data not shown). N-terminal amino acid sequencing of the 65-kDa protein yielded the sequence Lys-Thr-Ile-Thr and indicated true secretion of the N-terminal Als1p fragment with resultant loss of the first 17 amino acids (Fig. 3). Cleavage following Ala17 was correctly predicted by the Signalase program (21), which is based on the predictive algorithms of von Heijne (54) and identifies sites for signal peptide cleavage. Analysis of the other Als protein sequences with the Signalase program predicted signal peptide cleavage at the same site, which was conserved in each of the Als amino acid sequences characterized to date (Fig. 3). Because signal sequences have been shown to be processed similarly in S. cerevisiae and C. albicans (39), it is likely that the same processing site is utilized by C. albicans.

FIG. 3.

Alignment of N-terminal amino acid sequences predicted from genes in the ALS family. Amino acid sequences were predicted by the translation of ALS gene sequences 5′ of the start of the tandem-repeat domain. Sequences included are Als1p (26), Als2p and Als4p (this study), Als3p (25), and Als5p/Ala1p (18). Sequences were aligned with default parameters of the PILEUP program of Genetics Computer Group software (14). A consensus sequence (Cons), indicating amino acids conserved in all sequences, is provided below the alignment. The vertical line (designated SSC) between residues 17 and 18 denotes the site of signal sequence cleavage demonstrated biochemically for Als1p and predicted by computer algorithm to be conserved for the remaining proteins. Boxed regions labelled 1, 2, 3, and 4 correspond to the four 10-mer peptides from Als1p used to raise the rabbit polyclonal anti-Als antiserum used in indirect immunofluorescence studies. The boxed Als2p and Als4p sequences (labelled Probe) correspond to the region in the nucleotide sequence from which ALS2- and ALS4-specific oligonucleotides were derived. Boxed sequences between alleles of Als2p or Als4p indicate nonconserved amino acid sequences predicted from allelic nucleotide sequences. Consensus N glycosylation sites (2) are underlined in the Als2p sequences at positions 253 and 315. All CUG codons have been changed from Leu to Ser (46, 55).

The N-terminal domain of the Als protein molecule, particularly within the first 330 amino acids, was predicted to be relatively free of glycosylation. In this region, all predicted Als protein sequences, with the exception of Als2p, lacked consensus sequences for N glycosylation (Fig. 3). After the first 330 amino acids, each predicted Als protein had a threonine-rich region, raising the possibility that O glycosylation may be added; the increased frequency of serine and proline residues also supported this possibility (Fig. 3).

Sequences C-terminal of the tandem-repeat domain are divergent but have a serine-threonine-rich composition.

Of the three domains present in Als proteins, the domain C-terminal of the tandem repeats was the least conserved across the family. This domain varied in sequence and length in the predicted Als proteins but exhibited similar amino acid sequence compositions (Table 1). The serine-threonine richness of this domain was consistent with the possibility of abundant O glycosylation (27). This feature, along with the presence of consensus sites for N glycosylation, predicted that the C-terminal domain was heavily glycosylated (Table 1). Both of these features were also observed in the tandem-repeat domain, which was similarly predicted to be heavily glycosylated.

Although sequences of the C-terminal domain were divergent, those residues within 50 amino acids of the stop codon were highly conserved (Fig. 4). Within these conserved residues was a hydrophobic region with hallmarks of the consensus site for glycosylphosphatidylinositol (GPI) addition. The features of a GPI addition site have been well characterized (reviewed in references 8 and 12); following these rules, cleavage at the conserved Gly or Ser was predicted (Fig. 4). Yeast proteins to which GPI is added can be localized to either the cell membrane or, after truncation of the GPI, cross-linked in the cell wall (reviewed in reference 8). Analysis of amino acid sequences predicted from the S. cerevisiae genome sequencing project identified a dibasic motif immediately preceding the GPI attachment site that is present in proteins localized to the cell membrane (8). The lack of this dibasic motif in the Als protein sequences (Fig. 4) suggested that if C. albicans followed the same rules as S. cerevisiae, Als proteins were likely to be localized in the cell wall.

FIG. 4.

Alignment of the C-terminal amino acid sequences of Als proteins. Approximately the last 50 amino acids of each predicted Als protein sequence were aligned, to demonstrate sequence conservation in this region. A consensus sequence indicating amino acids conserved in every protein is indicated below the multiple alignment. The putative GPI addition sites are indicated by arrows. The larger arrow over the Gly residue suggests that this is the more likely GPI attachment site.

Although the C-terminal domain was the most highly divergent of the three ALS domains, the C-terminal domains of Als2p and Als4p and the nucleotide sequences which encode them were more than 95% identical. Multiple alignment of the nucleotide sequences of the ALS2 and ALS4 alleles indicated that only 3.9% of the sequence positions were mismatched. ALS2-1 and ALS4-1 each had a small gap; in each case, this gap was apparently due to duplication of a trinucleotide sequence in the other allele. Allelic sequences were also examined; ALS2 alleles had 2.8% of sequences mismatched, whereas ALS4 sequences varied in only 0.5% of nucleotides. The mismatch between allelic sequences in the 5′ domain was slightly less, with 0.3% variation between the ALS2 alleles and 0.4% between alleles of ALS4. The nucleotide sequence identity between the 3′ domains of ALS2 and ALS4 extended beyond the coding region; a region approximately 500 bp 3′ of each coding region was sequenced and found to be over 95% identical for each gene (data not shown).

Cell-surface localization of Als proteins by indirect immunofluorescence.

The predicted amino acid sequences of the Als proteins suggested that they were localized on the cell surface. Features such as an N-terminal signal peptide, a C-terminal GPI addition site, repeated sequences, and C-terminal regions rich in serine and threonine have all been noted in other yeast cell-surface proteins (reviewed in reference 8). Cell-surface localization of Als proteins was demonstrated by indirect immunofluorescence with a rabbit polyclonal antiserum raised against four KLH-linked 10-mer peptides from Als1p. Subsequent characterization of additional ALS genes indicated that the 10-mer peptide sequences were highly conserved in the predicted amino acid sequences for other proteins of the Als family (Fig. 3). Recognition of Als1p and Als3p by the anti-Als antiserum was demonstrated by Western blotting of heterologously produced protein fragments (data not shown); other Als proteins remain to be tested.

For immunofluorescence studies, yeast-form cells were grown in YPD medium and transferred to RPMI 1640 to induce germ tube formation. Both the mother yeast and germ tube stained with the anti-Als antiserum (Fig. 5). The specificity of this staining was demonstrated in competition experiments in which staining of both the mother yeast and the germ tube was blocked by the addition of soluble N-terminal Als1p fragment (data not shown). C. albicans cells treated either with preimmune serum from the rabbit in which the polyclonal serum was raised (Fig. 5F) or with commercially purchased anti-KLH antiserum in place of the anti-Als serum did not stain (data not shown).

FIG. 5.

Indirect immunofluorescence of C. albicans SC5314 germ tubes. YPD-grown C. albicans cells of strain SC5314 were induced to form germ tubes by inoculation into RPMI 1640 medium. Panels A, C, and E are light micrographs corresponding to the fluorescent micrographs in panels B, D, and F, respectively. Cells in panels B and D were treated with anti-Als antiserum followed by FITC-labelled goat-anti-rabbit IgG. Cells in panel F were stained with preimmune serum from the same rabbit in which the anti-Als serum was raised, followed by staining with the FITC-labelled secondary antiserum. The arrow in panel B indicates a large mother yeast cell that has lost its fluorescence. Arrows in panel D indicate small mother yeast cells for which fluorescence is still visible.

Although YPD-grown yeast forms stained with anti-Als serum (data not shown), this staining intensity diminished with elongation of the germ tube (Fig. 5B and D). The germ tube length which corresponded to the loss of mother yeast fluorescence differed with the size of the mother yeast cell (Fig. 5A and B [arrow] versus Fig. 5C and D [arrows]). An additional experiment was performed in which cells were collected for staining at 30-min intervals following the induction of germ tube formation. Decreasing fluorescence was observed with germ tube elongation for both small and large yeasts, with the same final result of nonfluorescent mother yeast. However, a longer time was required to effect equivalent results from small mother yeasts than from larger yeast cells (data not shown). Results from the indirect immunofluorescence experiments indicated that Als proteins were localized on the C. albicans cell surface and raised interest in the distribution of Als proteins during the yeast-to-hypha conversion.

Variability in ALS gene size and expression pattern.

It is well documented that the sizes of ALS genes in any C. albicans strain are highly variable (25, 26). In certain strains, alleles of a given ALS gene produce different-sized proteins due to variation in the numbers of tandem-repeat copies present in each allele (Table 1). Also, certain Als proteins are likely to be larger than others, with Als1p in one strain, for example, being twice the size of Als1p in another strain (26).

Variability also exists within the ALS family with respect to patterns of gene expression. Previous work demonstrated the differential expression of ALS1 (26) and the hypha-specific expression of ALS3 (25). Northern analysis established that ALS4 expression was correlated with the growth phase of a C. albicans culture. This effect was first noted in a pilot experiment in which C. albicans cells from an overnight YPD culture were subcultured into fresh media and incubated at 30°C with shaking at 200 rpm. RNA was analyzed at 3-h time points by Northern blotting with an ALS4-specific probe. ALS4-specific message was present except when cells were in early- to mid-log phase (data not shown). Cultures were followed for 33 h, at which time ALS4 message was abundant (data not shown). To more carefully analyze the time period in which ALS4-specific message was absent, the experiment was repeated with 1-h time intervals (Fig. 6). A Northern blot hybridized to detect ALS4-specific messages was deliberately overexposed to identify lanes in which ALS4-specific message was absent and to pinpoint the time when the synthesis of ALS4-specific message began (Fig. 6). Because the half-life of the ALS4-specific RNA was not known, it was unclear whether signals present at the 0-, 1- and 2-h time points were due to new synthesis or to dilution and decay of message present in the stationary-phase cells used to inoculate the culture. Synthesis of ALS4-specific message began during the fifth hour, when cells reached mid-log phase, and increased as the culture reached stationary phase. In previous experiments, the increase in ALS4-specific signal continued as the culture reached stationary phase (data not shown). These experiments were also done with C. albicans 3153A, and the same pattern of expression was noted (data not shown).

FIG. 6.

ALS4 expression in YPD-grown C. albicans cells. Strain SC5314 was grown overnight in YPD; cells from this culture were used to inoculate fresh YPD medium. Immediately after inoculation (0 h) and at each hour for the next 8 h, an aliquot of cells was removed. Cells were counted to generate a growth curve (right panel) and harvested for RNA extraction and Northern blotting with an ALS4-specific probe (left panel). A fragment of the C. albicans TEF1 gene was used as a control for equal loading of RNA.

In contrast to the definable pattern of ALS4 expression, ALS2-specific message was not detected in cultures grown under a wide variety of in vitro conditions, including all growth stages in YPD and YND (neopeptone substituted for peptone); RPMI 1640-induced germ tubes and hyphae; Lee (33) and Soll medium (supplemented Lee medium [3]) at pHs 4.5, 5.5, 6.5, and 7.5; Emmons-modified Sabouraud medium (32) with various carbon sources, including dextrose, galactose, maltose, and sucrose; and hyphal cells induced by adding 10% serum to YPD, 10 mM imidazole buffer (49), and PBS. Cells in these media were grown at various temperatures, including 25, 30, and 37°C. C. albicans strains in these studies included B311, B792, SC5314, 1177, 3153A, and WO-1. The lack of detection of an ALS2-specific message under so many in vitro conditions suggested a number of possibilities, including the possibility that ALS2 was a pseudogene and the possibility that ALS2 required in vivo signals for expression.

DISCUSSION

PCR screening of a C. albicans fosmid library with primers based on the consensus tandem-repeat sequence of ALS1 yielded fosmids encoding ALS3 (25), ALS2, and ALS4. These genes account for the ALS1-tandem-repeat-hybridizing fragments detected in high-stringency Southern blots of C. albicans genomic DNA (26). An additional gene described in the literature, ALA1 (18), also belongs in the ALS family and is designated ALS5. Although ALS5 encodes tandem repeats similar to those in ALS1, there is sufficient nucleotide sequence divergence between the two consensus repeat sequences that they detect different genomic fragments on high-stringency Southern blots.

Sequences N-terminal of the tandem repeats are highly conserved in the five aligned Als proteins; however, sequences C-terminal of the tandem repeats are divergent. N-terminal hydrophobic sequences function as a signal peptide which is cleaved following Ala17, a site conserved in each Als protein. C-terminal conserved hydrophobic sequences within the last 50 amino acids of each predicted Als protein have characteristics of the site for GPI addition. These observations are consistent with cell-surface localization of Als proteins, a feature demonstrated by indirect immunofluorescence with an anti-Als serum. The anti-Als serum stains both mother yeasts and germ tubes, with the intensity of staining of the mother yeast diminishing as the germ tube elongates.

Analysis of expression patterns of the two newly characterized ALS genes indicates that the expression of ALS4 is correlated with the growth phase of the culture. ALS2 message was not detected in vitro despite the fact that a wide variety of growth conditions were tested. These data support previous evidence that genes in the ALS family are differentially regulated.

Als protein profile on the C. albicans cell surface.

The profile of Als proteins on the C. albicans cell surface is highly variable and depends upon several factors, including growth conditions and strain (Table 1) (25, 26). Indirect immunofluorescence experiments demonstrated that the Als protein profile on the C. albicans surface is dynamic. Both mother yeasts and germ tubes stain with an anti-Als serum, and staining of the mother yeast diminishes with germ tube elongation. Possible explanations for the diminishing fluorescence of the mother yeast include migration of the Als proteins from the mother yeast to the growing germ tube, masking of the Als protein antigens on the mother yeast by synthesis of new cell wall material, and shedding of the Als protein antigens into the culture medium. Migration of cell wall material from mother yeast to germ tube is unlikely in light of data published by Staebell and Soll (51), which demonstrated growth of a germ tube predominantly by apical expansion. Masking of cell-surface proteins by carbohydrate or by glycosylation of other proteins has been discussed in the context of C. albicans cell-surface hydrophobicity (20, 41). Because the N-terminal epitope recognized by the anti-Als antiserum is predicted to be glycosylation free, it is unlikely that the epitope is masked by the direct addition of carbohydrate. However, modification of preexisting glycosylated cell-surface proteins or the production of new ones might explain the diminished fluorescence of the mother yeast. Another possible reason for the diminishing fluorescence of the mother yeast is the shedding of antigens into the culture medium, a phenomenon that is well documented (reviewed in reference 40). High-molecular-weight mannoproteins within the range of sizes predicted for Als proteins are among those shed (1, 40). Previous work by Brawner and Cutler described a C. albicans cell-surface antigen with the same pattern of expression we observed with our anti-Als serum (4–6). Immunogold electron microscopy with a monoclonal IgM indicated that the antigen studied by Brawner and Cutler was associated with the outer flocculent layer of the cell surface (4–6); they discussed shedding of the antigen into the growth medium as an explanation for its diminished intensity on the mother yeast during germ tube elongation.

Nature of the putative binding domain.

Als proteins were named because of the similarities between predicted Als sequences and the sequence of α-agglutinin of S. cerevisiae. Information about the well-characterized α-agglutinin has provided numerous clues about the localization and function of Als proteins (36). The most-significant sequence identity between Als proteins and α-agglutinin exists between the first 300 amino acids of each protein (26). Within this region, α-agglutinin has an immunoglobulin fold structure that is characteristic of many different cell adhesion molecules (10, 13, 35, 56). Because the Als proteins have significant sequence similarity to Agα1p across domain-sized blocks, they are likely to have a three-dimensional structure similar to that of α-agglutinin (34). This observation implicates the first 300 amino acids of the Als protein N terminus as a putative binding domain. Examination of Cys residues in the two sequences indicates that the six Cys residues in the first 300 amino acids of α-agglutinin are conserved in the first 300 amino acids of all Als proteins; however, in this region, each Als protein contains two additional Cys residues that are not found in α-agglutinin. Conservation of the eight Cys residues in the first 300 amino acids of each mature Als protein strongly suggests structural similarity in this portion of each molecule. Additional studies are planned to predict the structure of the putative binding domain of Als proteins and to evaluate its relatedness to the immunoglobulin fold.

The tandem-repeat region and C-terminal domain are predicted to be highly N and O glycosylated. Because O glycosylation can confer a rigid structure on a given protein (27), the tandem repeats and C-terminal sequences are likely the means by which the putative binding domain is displayed on the cell surface. Estimates of protein length demonstrate that most Als proteins could be localized in the cell membrane or cell wall and still display this binding domain on the cell surface; shorter proteins, such as Als3p (25), may require cell wall localization to expose the putative binding domain on the cell surface. While lack of the dibasic motif suggests that Als proteins are localized in the cell wall, membrane localization cannot be completely ruled out, because the GPI addition site of C. albicans Phr1p, which is localized in the cell membrane, also lacks the dibasic motif (47, 53). It is also possible that not all Als proteins have the same subcellular localization.

Als protein function.

The growing size of the ALS family prompts questions about the function of the encoded proteins and conservation of function between proteins in the family. Als protein function was addressed by two recent studies in which S. cerevisiae was transformed with a C. albicans library seeking C. albicans genes that could confer an adhesive phenotype on the nonadherent S. cerevisiae. In the first study, expression of ALA1/ALS5 in S. cerevisiae conferred adherence to fibronectin, laminin, type IV collagen, and buccal epithelial cells (18). In the second study, the expression of ALS1 in S. cerevisiae conferred adherence to endothelial cells and to cells from an esophageal epithelial line (16). Conclusions drawn in both these studies are consistent with our ideas about the function of Als proteins. The nature of the adhesive interaction observed in these studies, however, needs further clarification to determine whether adhesion is due to the putative N-terminal binding domain or to another portion of the molecule. One portion of the molecule with the potential to mediate an adhesive effect is the region predicted to be highly glycosylated. The effect of protein must be separated from that of the likely carbohydrate in order to define the nature of the adhesive interaction.

The availability of amino acid sequences for several Als proteins allows initial speculation about the conservation of function among proteins in the family. Presumed adhesive function due to extensive glycosylation suggests that Als proteins require only a sequence that serves as the scaffold for the addition of carbohydrate; this is provided by the tandem-repeat and C-terminal domains which are rich in consensus N glycosylation sites and in serine and threonine, which are potential sites for the addition of O-linked carbohydrate (2, 27). The amino acid sequence of the C-terminal domain is highly variable between Als proteins, but each protein has features conducive to abundant carbohydrate addition. Adhesive function due to an N-terminal binding domain requires conservation of sequence determining similar structural features. This is observed for each mature Als protein, which has eight conserved Cys residues, a similar N-terminal domain length, and a high degree of sequence identity. Of the five Als protein sequences, Als4p is least like the others in regard to its N-terminal domain sequence, suggesting that it is the least likely to exhibit conserved function. More-meaningful information about conservation of function will be gained from structural predictions with even-less-similar sequences, such as that of Als7p, which is only about 50% identical to Als1p in the N-terminal domain (24). Differential glycosylation at the N terminus, which is possible for Als2p (Table 1), could also lead to alterations in the three-dimensional structure of the region and to corresponding variations in function. Whether adhesive function is due to protein, carbohydrate, or both, proteins encoded by the differentially regulated, multigene ALS family have the potential to explain much of the strain- and growth-medium-dependent differences in adhesion commonly observed for C. albicans.