Expression cloning of the Candida albicans CSA1 gene encoding a mycelial surface antigen by sorting of Saccharomyces cerevisiae transformants with monoclonal antibody-coated magnetic

Expression cloning of the Candida albicans CSA1gene encoding a mycelial surface antigen by sortingof Saccharomyces cerevisiae transformants withmonoclonal antibody-coated magnetic beads

Claude Lamarre,1 NoeÈlla Deslauriers1,2 and

Yves Bourbonnais1,3*1DeÂpartement de biochimie et de microbiologie, Centres

de Recherche en 2EÂ cologie Buccale (GREB) et3sur la Structure, la Fonction et l'IngeÂnierie des

ProteÂ ines (CREFSIP) de l'UniversiteÂ Laval, QueÂbec,

Canada G1K 7P4.

Summary

The mycelial surface antigen recognized by monoclo-

nal antibody (mAb) 4E1 has previously been shown to

be present predominantly in the terminal third of the

hyphal structures in Candida albicans. We report here

the expression cloning of the corresponding gene

(CSA1 ) by mAb 4E1-coated magnetic beads sorting of

Saccharomyces cerevisiae transformants expressing a

C. albicans genomic library. The strategy is both highly

selective and highly sensitive and provides an addi-

tional genetic tool for the cloning and characteriza-

tion of C. albicans genes encoding surface proteins.

CSA1 is an intronless gene encoding a 1203-residue

protein composed of repetitive motifs and domains.

Northern analysis indicates that CSA1 is preferentially

expressed during the mycelial growth phase, although

a low level of CSA1 mRNA can be detected in the yeast

form. As evidenced by indirect immuno¯uorescence

microscopy with mAb 4E1, Csa1p is not randomly dis-

tributed over the surface of yeast cells, but localizes

predominantly in the growing buds. This suggests

that the distribution of Csa1p may be restricted to

sites of cell surface elongation. Both heterozygous

and homozygous C. albicans csa1D mutants are viable.

Upon induction of mycelial growth, the number and

size of hyphal structures derived from the mutants

are similar to those observed in the parental wild-

type strain. The physiological role of Csa1p has yet

to be determined. However, the presence in Csa1p

of repeated cysteine-rich hydrophobic domains with

signi®cant sequence similarity to motifs found in

surface proteins (Ag2 and Pth11) from two distantly

related fungal pathogens (Coccidioides immitis and

Magnaporthe grisea respectively) suggests a common

function in host interaction.

Introduction

Candida albicans is of major medical importance, being

the most commonly isolated fungal species from various

mucosal surfaces in healthy individuals and from infec-

tious sites in patients with candidiasis. Most frequently, it

causes super®cial, irritating infections of the oral and uro-

genital tracts. However, serious deep-seated or systemic

infections can develop, particularly in immunocompromised

subjects.

The performance of Candida albicans as an opportunis-

tic pathogen is associated with a number of factors, which

include the morphological and functional modi®cations

resulting from switching between the yeast and the hyphal

forms. Mycelium formation is believed to contribute to

fungal adhesion to host cell surfaces and to facilitate inva-

sion of a variety of host tissues through the expression of

specialized surface proteins and enzymes (Cutler, 1991;

Odds, 1994; Staab et al., 1999). On the basis that the

dimorphic process is likely to be associated with differential

expression of mycelial cell-speci®c molecules, biochemical

and immunological approaches have been used for their

identi®cation.

The success of immunological approaches largely

depends on the nature and speci®city of the antibody pre-

paration but, recently, the use of monoclonal antibodies

(mAbs) has proved invaluable in the screening of yeast

versus mycelial antigens. As summarized by Ponton et al.

(1993), different types of germ tube surface antigens

have been described, but true hyphal antigens (type I anti-

gens) appear to be scarce, because most hyphae-speci®c

mAbs also react with either dithiothreitol (DTT)-treated

(stripped) yeast cells (type II antigens) or both yeast cells

and germ tubes (type IV antigens).

We have recently produced an IgG mAb that reacts

strongly with the terminal third of the growing ®laments,

but not with the parent blastospore, in C. albicans mycelial

Molecular Microbiology (2000) 35(2), 444±453

Q 2000 Blackwell Science Ltd

Received 28 July, 1999; revised 7 October, 1999; accepted 11 Octo-ber, 1999. *For correspondence at Pavilion Marchand, UniversiteÂLaval, QueÂbec. E-mail [email protected]; Tel. (�1)418 656 2131, ext. 7069; Fax (�1) 418 656 7176.

cultures (Deslauriers et al., 1996). In immunoblots, the

mAb 4E1 detected two species of 117 kDa and 104 kDa

from DTT extracts prepared from the mycelial cells, but

not from yeast cells, suggesting that the 4E1 antigen

de®nes a true type I antigen.

To characterize this surface antigen further and also to

con®rm its differential expression in yeast versus hyphae,

we sought to clone the corresponding gene. We reasoned

that functional surface expression of this major mycelial

antigen in the yeast Saccharomyces cerevisiae might

provide a simple and rapid approach to clone the corre-

sponding gene using magnetic beads coated with mAb

4E1.

Results

Expression cloning of the C. albicans gene encoding

the 4E1 antigen

Preliminary experiments showed that S. cerevisiae cells

transformed with the yeast multicopy plasmid YEP24 did

not attach to either uncoated or mAb 4E1-coated magnetic

beads (Fig. 1A). This indicated that there is no non-speci®c

adherence of the cells to the beads and that the epitope

recognized by mAb 4E1 is either not expressed or not

exposed at the cell surface in this organism. S. cerevisiae

was thus transformed with a C. albicans genomic library,

and the pooled transformants were screened with the

magnetic beads coated with mAb 4E1 during a 16 h incu-

bation period at 48C. After extensive washes with PBS±

BSA, the free and cell-bound microspheres were ®nally

resuspended into the same buffer and spread onto selec-

tive agar plates (SC-ura). After an incubation period of 3

days at 308C, three colonies grew on this medium. The

plasmid p4E1, recovered from one of the colonies, con-

ferred the ability of freshly transformed S. cerevisiae cells

to be repeatedly sorted out by the coated beads. That

this sorting resulted from a speci®c interaction between

the candidal antigen exposed at the surface of S. cerevi-

siae transformants and mAb 4E1 was ®rst suggested by

the >1000-fold reduction observed in the sorting ef®ciency

when the mAb 4E1-coated microspheres were replaced

with either uncoated or mAb A2C7-coated (antienolase)

beads respectively (Fig. 1A). This was con®rmed by com-

petition experiments in which an excess of mAb 4E1 was

incubated with exponentially growing p4E1 transformants

before sorting with the coated microspheres (Fig. 1B). In

two parallel experiments, one conducted with 105 cells

(Expt 1) and the other with 106 cells (Expt 2), competition

with mAb 4E1 led to 90% and 82% inhibition of sorting

respectively. We therefore concluded that plasmid p4E1

carries the C. albicans gene coding for the 4E1 surface

antigen. However, the low percentage of sorting observed

with a homogeneous culture of p4E1-transformed yeast by

mAb 4E1-coated beads (<2 ´ 103/105 cells or <2%;

Fig. 1A) suggested a much reduced expression or surface

exposure of this mycelial antigen in S. cerevisiae yeast

cells compared with mycelial cultures of C. albicans. Micro-

scopic examination of the mAb 4E1-sorted cells revealed

that the microspheres were not randomly distributed over

the cell surface, but preferentially attached to the growing

bud or at the mother±daughter neck junction (Fig. 1C).

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 35, 444±453

Fig. 1. Sorting of S. cerevisiae transformants by uncoated, control mAb A2C7- and mAb 4E1-coated magnetic beads.A. Exponentially growing cultures of S. cerevisiae transformed with the indicated plasmids were sorted out with uncoated, control mAbA2C7- and mAb 4E1-coated microspheres as described in Experimental procedures. The results are expressed as the average number ofcfu recovered per 105 cells 6 SD from an experiment carried out in triplicate.B. Either 105 cells (Expt 1) or 106 cells (Expt 2) from an exponentially growing culture of S. cerevisiae transformed with p4E1 were ®rstincubated with either mAb 4E1 or PBS±BSA before sorting with the mAb 4E1-coated microspheres. The results are expressed as thepercentage of inhibition of the sorting after a preincubation with mAb 4E1, as calculated from the number of cfu recovered with the coatedbeads.C. Representative differential interference contrast micrograph (100 ´) of the sorted S. cerevisiae p4E1 transformant decorated with the mAb4E1-coated magnetic beads.

Cloning of the C. albicans CSA1 gene 445

The deduced amino acid sequence of CSA1 reveals

the presence of repeated domains with sequence

similarity to the C. immitis antigen 2 and M. grisea

Pth11 protein

Partial restriction mapping of plasmid p4E1 indicated that

it carries a 7.4 kb genomic fragment (Fig. 2A). Subcloning

experiments and immunocapture of the corresponding

transformants showed that the gene encoding the 4E1

surface antigen (referred to hereafter as Candida surface

antigen 1; CSA1 ) lies on the 4.2 kb Sal I±XbaI fragment

(Fig. 2A). Sorting of yeast expressing this shorter DNA

fragment (p4E1DS) was in fact signi®cantly better than

that observed with the p4E1 transformants, suggesting

that deletion of 58 ¯anking sequences resulted in derepres-

sion of CSA1 gene expression (Fig. 2A). However, this pos-

sibility was not investigated further. The 4.2 kb Sal I±XbaI

genomic fragment was entirely sequenced on both strands

and found to contain a single, uninterrupted, open reading

frame (ORF) of 3609 nucleotides. The deduced amino

acid sequence of the ORF (1203 residues) revealed sev-

eral important features (Fig. 2B). First, both N- and C-ter-

mini contain a core of hydrophobic residues, which may

function as a signal sequence and a GPI-anchoring deter-

minant respectively. Anchoring to membranes through a

GPI moiety is a common feature of many cell wall-associ-

ated proteins in fungi, including C. albicans and S. cerevi-

siae (Kapteyn et al., 1994; Hoyer et al., 1995; Bailey et al.,

1996; Caro et al., 1997). That this putative GPI-anchoring

determinant is important for the correct assembly of the

4E1 antigen into the cell wall is suggested by the observa-

tion that yeast expressing a C-terminal 164-amino-acid

truncated version of the protein cannot be sorted by the

coated microspheres (Fig. 2A). A striking feature of the

protein is a 102-residue cysteine-rich hydrophobic domain

(CH domain) that is repeated ®ve times in the sequence.

The sequence identity between each domain exceeds

95%, except for the central repeat (amino acids 403±

504 in Csa1p), which diverges slightly from the other

repeats (84% sequence identity, <94% sequence similar-

ity). These CH domains are interspersed by segments of

variable length (60±89 amino acids), almost exclusively

composed (89%) of the residues P, E, T, S, A and Q,

with an overall net charge of ÿ54. Within these acidic,

proline-rich domains, at least two motifs, TSAP and P(A/

S/V)ETSS(E/Q), can be distinguished. A copy of the

TSAP motif is always found upstream (15±16 amino

acids) of a CH domain, whereas the longer motif is

repeated several times within the segments separating

the CH domains. Finally, the 333-amino-acid domain

located between the last CH domain and the putative

GPI-anchoring determinant (amino acids 852±1184) is

also enriched in the residues P, E, T, S, A and Q (66%)

with a net negative charge of ÿ13. This domain contains

all the putative N-glycosylation sites (10).

A search for sequence similarity in the S. cerevisiae

protein database revealed that there is no homologue of

Csa1p in this organism. However, a signi®cant similarity

was noticed between the CH domains of Csa1p, the

immunoreactive antigen 2 (Ag2) of Coccidioides immitis,

and the Pth11 protein of Magnaporthe grisea, two other

fungal pathogens (Fig. 3). Amino acids 19±88 of Ag2

and 34±104 of Pth11p show 35% and 29% sequence


Fig. 2. Deduced amino acid sequence of the C. albicans CSA1gene.A. Schematic representation of the genomic DNA fragments carriedby the indicated plasmids and the ability of the corresponding yeasttransformants to be sorted out by mAb 4E1-coated magneticbeads. The hatched box represents the CSA1 coding region andthe arrow the direction of transcription. The dashed linecorresponds to DNA sequence derived from YEP24. The averagenumber of cfu 6 SD per 105 cells from an experiment carried out intriplicate was 6680 6 580 (��), 715 6 156 (�) and 60 6 33 (ÿ).B. Deduced amino acid sequence of CSA1 using the single lettercode. The predicted signal peptide (amino acids 1±17) and thehydrophobic stretch predicted to serve as a GPI-anchoringdeterminant (amino acids 1184±1203) are in bold italics. Therepeated hydrophilic sequences TSAP and P(A/S/V)ETSS(E/Q)are underlined and twice underlined respectively. The ®ve CHdomains are black boxed. The putative N-glycosylation siteslocated in the C-terminus of Csa1p are denoted by asterisks.

446 C. Lamarre, N. Deslauriers and Y. Bourbonnais

identity (56% and 50% sequence similarity), respectively,

with a 66-amino-acid motif internal to the Csa1p CH

domains. Remarkably, all the cysteine residues with an

insertion of a single amino acid in Ag2 align with those

found in the Csa1p CH domains. In addition to the similar-

ity in the primary amino acid sequence, the content of

hydrophobic residues within this motif is similar (<26%)

in all three proteins.

The 4E1 surface antigen can be detected in the

growing buds of C. albicans yeast cells

We have shown previously that the 4E1 surface antigen is

exposed on the hyphal extensions in the mycelial form of

C. albicans (Deslauriers et al., 1996). The antigen could

not be detected either in the parent blastospore from myce-

lial cells or in the yeast form of C. albicans by immunohisto-

chemistry. Northern analysis con®rmed the presence of

an abundant CSA1 transcript (<4.0 kb) with total RNA

extracted from C. albicans mycelial cultures (Fig. 4). A

low level of mRNA could also be detected in the RNA

sample prepared from C. albicans blastospores harvested

during the early exponential growth phase. This, and the

low level of expression of the candidal surface antigen in

S. cerevisiae yeast cells, therefore prompted us to re-

examine the presence of Csa1p by indirect immuno¯uor-

escence microscopy in C. albicans yeast cells (Fig. 5).

As for the Northern analysis, the culture was grown to

early exponential phase to increase the proportion of bud-

ding yeast. Under these conditions, mAb 4E1 reacted with

a fraction of the cell population. As observed previously

with the S. cerevisiae transformants, the antigen was

detected predominantly, if not exclusively, in the growing

buds. A control experiment, in which only the secondary

¯uorescein-conjugated antibody was added, con®rmed


Fig. 3. Alignment of the Csa1p CH domain with peptide sequences derived from the C. immitis Ag2 and M. grisea Pth11 proteins. Alignmentwas performed with the CLUSTALW software available on the ExPASy molecular biology WWW server of the Swiss Institute of Bioinformatics.Sequence of Ag2 and Pth11p were taken from Dugger et al. (1996) and DeZwaan et al. (1999) respectively. Black characters indicateidentical residues, whereas grey characters correspond to conservative changes in two of the three sequences. Dashed lines represent sitesof insertion, and the dots indicate the amino acid positions not conserved.

Fig. 4. Northern hybridization of total RNA extracted from theyeast and mycelial forms of C. albicans with probes derived fromCSA1 and the S. cerevisiae ACT1 gene. Total RNA was preparedfrom C. albicans cultures growing in IMDM medium either at 258C(predominantly yeast cells; Y) or at 378C (mycelial cells; M).Identical amounts of total RNA (20 mg) were fractionated byagarose-formaldehyde gel and transferred onto nylon membrane.The RNA blot was then hybridized with the CSA1 3.9 kb Hpa Ifragment and a probe derived from the S. cerevisiae ACT1 gene(actin).Top. Autoradiogram of the Northern hybridization.Bottom. Ethidium bromide staining of the agarose-formaldehydegel.

Fig. 5. Indirect immuno¯uorescence microscopy of C. albicansyeast cells with mAb 4E1.A. Phase-contrast micrograph (40 ´) of C. albicans yeast cellsincubated with mAb 4E1 and the ¯uorescein-conjugated anti-mouseIgG secondary antibody.B. Epi¯uorescence micrograph of (A).Arrows point to the growingbuds.


the speci®city of the immuno¯uorescence pro®le (not

shown). Hence, either a subpopulation of C. albicans

yeast cells expresses Csa1p or the blastospores may

transiently express the surface antigen during the budding

process.

Disruption of CSA1 indicates that this gene is not

essential for viability in C. albicans

As a ®rst step towards assessing the functional role of

CSA1, we sought to construct a strain with disrupted alleles

of this gene. As shown by Southern analysis (Fig. 6), the

parental strain (CAI4) used for the targeted disruption

carries shorter alleles of CSA1 than that observed in the

ATCC 32354 strain. The HpaI fragment, which contains

most of the CSA1 coding region and part of the 58 ¯anking

sequence, is <3.4 kb long in CAI4 compared with 3.9 kb in

ATCC 32354. As predicted from the sequence of the gene,

®ve fragments, including a doublet of <0.5 kb, should

hybridize when the genomic DNA digested with BamHI

is probed with the HpaI fragment (Fig. 6, top). Fragments

of identical sizes were observed in genomic DNA prepared

from both strains. However, the signal intensity ratio of the

0.5 kb over the 1.0 kb fragment was signi®cantly reduced

for strain CAI4 compared with ATCC 32354. This strongly

suggests that the alleles of CSA1 from CAI4 are missing

one of the two 0.5 kb BamHI fragments. To con®rm this

and to locate the deletion site more precisely, genomic

DNA was digested with Pst I and again probed with the

HpaI fragment. In contrast to the <1.8 kb predicted from

the sequence, this revealed that the internal Pst I fragment

of CSA1 is <1.3 kb in size in strain CAI4. Collectively there-

fore, these results indicate that the alleles of CSA1 found

in strain CAI4 lack the <0.5 kb BamHI fragment located

towards the 38 end of the gene in strain ATCC 32354.

Based on the nucleotide sequence, deletion of this BamHI

fragment predicts a protein composed of four instead of ®ve

CH domains.

We used the Ura-blaster technique to disrupt the CSA1

gene (see Experimental procedures ). The coding region

of CSA1 internal to the BamHI sites was replaced by the

disrupting cassette composed of the CAT and URA3 sequ-

ences (Fig. 7, top). The DNA was then restricted with

HpaI, and the linear DNA fragment was used to transform

strain CAI4. Integration of the disrupting cassette at the

CSA1 locus was monitored after each round of transforma-

tion by Southern analysis of the genomic DNA prepared

from the Ura� transformants and digested with EcoRV

and XbaI (Fig. 7, bottom). After the second round of trans-

formation, three bands of 6.0 kb, 3.9 kb and 3.2 kb, corre-

sponding to csa1::CAT-URA3-CAT, CSA1 and csa1::CAT,

respectively, were revealed by Southern analysis (Fig. 7,

lane 4). Hence, strain CAI4 carries three alleles of CSA1,

and a third round of transformation was required to obtain

a strain lacking all functional alleles of this gene (Fig. 7,

lane 6).

The construction and selection of the csa1 mutants was

carried out with the yeast form of C. albicans CAI4. All the

mutant strains (single, double and triple deletions) showed

the same viability, indicating that CSA1 is not an essential

gene. Furthermore, these strains were grown under condi-

tions that elicit the transition from the yeast to the mycelial

form, and both the number and the size of the hyphal

extensions were similar in all three cultures (Fig. 8A±C).

Hence, CSA1 is not essential for cell growth in either mor-

phological phase, and its absence does not preclude the

emergence and elongation of the hyphal structures. Indirect

immuno¯uorescence microscopy performed on the myce-

lial cells with mAb 4E1 indicated that the relative abundance


Fig. 6. Southern analysis of the genomic DNA extracted from theC. albicans strains ATCC 32354 and CAI4.Top. Schematic representation of the CSA1 locus from strainsATCC 32354 and CAI4. Grey and black boxes correspond,respectively, to the N- and C-terminal sequences of the CSA1coding region. The hatched and open boxes represent the repeatedCH domains and the hydrophilic sequences respectively. Therelevant restriction sites are shown: B, BamHI; H, Hpa I; P, Pst I.Bottom. The genomic DNA (10 mg) prepared from strains ATCC32354 (A) and CAI4 (C) was digested with the indicated restrictionenzymes, fractionated on agarose gel, transferred onto nylonmembranes and probed with the CSA1 3.9 kb Hpa I fragment. TheDNA size markers (1 kb ladder; Gibco BRL) are indicated on theleft.


of the surface antigen was subjected to a gene dosage

effect, as evidenced by the decreased ¯uorescence inten-

sity between the single and double mutant (Fig. 8D and E).

Most importantly, the antigen could not be detected in the

triple mutant, indicating that Csa1p is the only C. albicans

surface antigen containing the 4E1 epitope (Fig. 8F).

Discussion

The composition of the Candida albicans cell wall has

been thoroughly studied over the past decade, and anti-

genic variations in cell wall mannoproteins as a function

of dimorphism have been investigated with polyclonal

and monoclonal antibodies (Smail and Jones, 1984; Sund-

strom and Kenny, 1984; Brawner and Cutler, 1986; Chardes

et al., 1986; Hopwood et al., 1986; Chaf®n et al., 1988;

Sundstrom et al., 1988; Casanova et al., 1989; Garzon

et al., 1989; Poulain et al., 1989; Ollert and Calderone,

1990; Torosantucci et al., 1990; Li and Cutler, 1991; Ponton

et al., 1993). These antibodies were most frequently direc-

ted towards carbohydrates carried by the cell wall proteins.

Recently, we described the production of mAbs directed

against C. albicans cell wall proteins and showed that

mAb 4E1 recognizes proteinaceous antigens on the sur-

face of mycelial cells (Deslauriers et al., 1996). The aim

of the present study was to clone the corresponding

gene as a ®rst step towards understanding the role of

this mycelial cell wall protein.

A number of expression cloning systems have been

developed to isolate cDNA clones corresponding to cell sur-

face molecules, and immunological screening by antibody


Fig. 7. Disruption of the CSA1 gene from C. albicans.Top. Schematic representation of the wild-type, CAT-URA3-CAT-and CAT-disrupted alleles of CSA1 from strain CAI4. The greyboxes correspond to the CSA1 coding region. The size of thecorresponding EcoRV±XbaI fragment is indicated on the right. Therelevant restriction sites are shown: B, BamHI; E. EcoRV; H, Hpa I;X, XbaI.Bottom. Genomic DNA (10 mg) prepared from the parental strainCAI4 (lane 1), a ®rst-round Ura3� transformant (lane 2), a ®rst-round FOA-resistant segregant (lane 3), a second-round Ura3�

transformant (lane 4), a second-round FOA-resistant segregant(lane 5) and a third-round Ura3� transformant (lane 6) wasdigested with EcoRV and XbaI, and the resulting Southern blot wasprobed with the CSA1 Hpa I±BamHI fragments. The DNA sizemarkers (1 kb ladder; Gibco BRL) are indicated on the left.

Fig. 8. Indirect immuno¯uorescencemicroscopy of the C. albicans csa1D mutantstrains with mAb 4E1. Differential interferencecontrast (A±C) and epi¯uorescence (D±F)micrographs (40 ´) of the single (A and D),double (B and E) and triple (C and F) csa1Ddeletant constructed in strain CAI4. Themycelial form was induced as described inthe legend to Fig. 4, and the indirectimmuno¯uorescence with mAb 4E1 wasperformed as indicated in the legend to Fig. 5.Note that the epi¯uorescence micrographspresented in (D) and (E) were exposed for10 s compared with 40 s for that shown in (F).


capture after panning, ¯uorescence-activated cell sorting

(FACS) or magnetic bead sorting has been used success-

fully in mammalian cells. Here, we describe an adaptation

of this approach for isolating C. albicans genes encoding

cell surface molecules. From <1.5 ´ 107 yeast cells

derived from 8 ´ 103 independent transformants, three

were sorted out by the mAb 4E1-coated magnetic

beads, and one carried the C. albicans gene CSA1 (p4E1

transformant), illustrating the exquisite selectivity of the

technique. As the Csa1p surface antigen is only weakly

expressed in budding yeast (see below), the method is

also very sensitive. During the course of this study, Gaur

and Klotz (1997) reported the cloning of the C. albicans

ALA1 gene through sorting of S. cerevisiae transformants

with magnetic beads coated with extracellular matrix pro-

teins. Immunocapture of S. cerevisiae transformants thus

offers an attractive alternative for the identi®cation of C.

albicans genes encoding surface antigens. In both C. albi-

cans and S. cerevisiae yeast cells, Csa1p was not randomly

distributed over the cell surface, but localized predomi-

nantly in the growing buds. The strong induction of the

CSA1 transcript observed upon transition from the yeast

to the mycelial form and the absence of Csa1p from the

parent blastospore (Deslauriers et al., 1996; Fig. 8) there-

fore suggest that the distribution of the antigen may be

restricted to sites of cell surface elongation.

The primary amino acid sequence of Csa1p reveals the

presence of repeated, nearly identical, cysteine-rich hydro-

phobic domains that are separated by acidic, proline-rich

hydrophilic domains composed of two repetitive units,

TSAP and P(S/A/V)ETSS(E/Q). Interestingly, the number

of repeats within CSA1 was found to be different from

strains CAI4 and ATCC 32354 (Fig. 7), as well as from

various laboratory strains and clinical isolates (C. Lamarre

et al., unpublished data). Different tandem repeats have

been found in surface proteins from a variety of organ-

isms, including the C. albicans Als1, Hwp1 and Hyr1 cell

wall proteins (Hoyer et al., 1995; Bailey et al., 1996; Staab

et al., 1996), and these are frequently involved in host cell

attachment, evasion of phagocytosis, invasion of host

cells or act as neutralization epitopes (Klein et al., 1993;

Watari et al., 1994; Hogan et al., 1995; Madoff et al.,

1996; Jenkinson and Lamont, 1997; Ramasamy, 1998;

Staab et al., 1999). The role of these repeats in Csa1p is

currently unknown. However, the presence of domains

with sequence similarity to the Csa1p CH domains in sur-

face proteins from two distantly related pathogenic fungi,

C. immitis and M. grisea, suggests a common function.

C. immitis is an important fungal human pathogen,

whereas M. grisea is a plant fungal pathogen responsible

for the infection of many grass species, including rice. The

C. immitis Ag2 is a 194-amino-acid protein that is expressed

in the mycelium- and spherule-phase cell walls (Cox, 1989;

Galgiani et al., 1992; Dugger et al., 1996; Zhu et al.,

1996a,b). Pth11p is a 628-amino-acid protein that acts

as an upstream component of pathogenicity signalling in

M. grisea (DeZwaan et al., 1999).

Cell surface hydrophobicity (CSH) in C. albicans and

M. grisea has been linked to a plethora of host interactions

and fungal functions, and several surface proteins are

thought to be involved in CSH in C. albicans (Chaf®n et

al., 1998). Given the hydrophobic character of the CH

domain and the preferential expression of Csa1p during

the mycelial growth phase, its potential function may be

to increase the overall hydrophobicity of the fungal cell

wall associated with the transition from the yeast to the

mycelial form (Chaf®n et al., 1998). In support of this

hypothesis, the CH domains of Csa1p and the analogous

domains found in Ag2 and Pth11p show some similarity to

a class of small secreted fungal proteins (96±125 amino

acids) called hydrophobins. Hydrophobins from different

species have now been identi®ed (Kershaw and Talbot,

1998). They present a weak sequence identity (<4%),

but they all possess eight cysteine residues dispersed

throughout a sequence rich in proline and hydrophobic

amino acids. Despite their weak sequence identity, they

appear to be functionally interchangeable (Kershaw et al.,

1998). In response to environmental stimuli, these mole-

cules self-assemble into polymeric structures to form a

coat that dramatically increases the hydrophobic character

of the fungal cell walls.

It is obvious that morphogenesis in C. albicans is more

than a change in cell shape and entails the expression of

physiological attributes linked to its performance as a suc-

cessful commensal and opportunistic pathogen. Until now,

molecular genetic approaches have identi®ed a few genes

encoding hyphae-speci®c surface proteins that may con-

tribute to differences in cell wall structure and functions

(Hoyer et al., 1995; Bailey et al., 1996; Staab et al., 1996;

Fu et al., 1998). In a recent study, Staab et al. (1999)

showed that Hwp1p is involved in the stabilized adhesion

of C. albicans to buccal epithelial cells. Homozygous

hwp1D mutants, while still capable of forming hyphal exten-

sions, enabled mice to survive intravenous administrations

of normally lethal doses of C. albicans. We report here that

C. albicans Csa1p is a non-essential protein differentially

expressed in blastospores and mycelia. Its accessibility

to external ligands, its dynamic expression and hydrophobic

character, together with its sequence similarity to domains

found in the immunogenic Ag2 protein (C. immitis ) and

Pth11p (M. grisea ), suggest that this protein may be

involved in surface interactions with its changing molecular

and cellular environment. However, its physiological sig-

ni®cance and function remains to be elucidated. Towards

this goal, we are currently investigating whether any

phenotypic variation can be linked to the loss of CSA1

in C. albicans and/or its functional expression in S.

cerevisiae.



Experimental procedures

Strains and media

Candida albicans ATCC 32354 was used for the production ofmonoclonal antibodies (Deslauriers et al., 1996) and through-out this study. Yeast cells were cultured in Iscove's modi®edDulbecco medium at 258C, and mycelium formation wasinduced at 378C, as described previously. Disruption of CSA1was performed in the C. albicans strain CAI4 (Fonzi andIrwin, 1993). The S. cerevisiae strain used was W303-1b(Thomas and Rothstein, 1989) (MATa ade2-1 can1-100his3-11,15 leu2-3112 trp1-1 ura3-1 ), which was grown at308C in either YPD or SC-ura broth as described previously(Kaiser et al., 1994) for untransformed or transformed yeastrespectively. The E. coli strain used for plasmid puri®cationand subcloning experiments was MC1061, which was culturedin 2 ´ YT medium supplemented with 1% glucose and50 mg mlÿ1 ampicillin (Sambrook et al., 1989).

DNA manipulations and transformations

All DNA manipulations were carried out according to standardprocedures (Sambrook et al., 1989). All restriction enzymesand other DNA modifying enzymes were purchased fromNew England Biolabs. Plasmid p4E1DS was constructed bydigesting p4E1 with Sal I followed by self-ligation of the result-ing plasmid. To construct p4E1DS-Stop, the 4.2 kb Sal I±XbaIfragment of p4E1DS was ®rst subcloned into pTZ18R, digestedat the unique Spe I site, followed by ®lling in with the Klenowfragment of E. coli DNA polymerase I and self-ligation of theresulting plasmid. The 4.2 kb Sal I±XbaI fragment was thensubcloned back to the YEP24 yeast expression plasmid. Puri-®cation of total RNA was performed according to the proceduredescribed by Kohrer and Domdey (1991). For the constructionof the library and for Southern analysis, puri®cation of genomicDNA from 40 ml of yeast culture was carried out as describedby Kaiser et al. (1994). Standard procedures were used forSouthern and Northern analyses (Sambrook et al., 1989).Nucleic acids were transferred onto nylon membranes in allcases (Hybond-N; Amersham Life Science), and the radio-active probe was prepared with the Rediprime kit (AmershamLife Science) using [32P]-dCTP (ICN) according to the manu-facturer's instructions. Automatic DNA sequencing reactionswere performed by the dye terminator cycle protocol withdouble-stranded DNA on a GeneAmp polymerase chain reac-tion (PCR) system 9600 and a 373 DNA sequencer (AppliedBiosystems, Perkin-Elmer). Complete sequencing of CSA1using universal primers was achieved by the constructionof overlapping subclones in pTZ18R (Pharmacia Biotech).The CSA1 nucleotide sequence has been deposited in theGenBank database under the accession number AF080221.

Transformation of yeast using lithium acetate salt was per-formed according to the rapid procedure described by Kaiseret al. (1994). The CaCl2 protocol for transformation of E. coliwas used for all subcloning experiments and recovery of thep4E1 plasmid from yeast, whereas electroporation was usedto construct the genomic library (Sambrook et al., 1989).

Construction of the C. albicans genomic library

High-molecular-weight genomic DNA prepared as described

above was partially digested with Sau 3AI and size-selectedon a discontinuous potassium acetate gradient (5±25%) accord-ing to the method described by Aruffo and Seed (1987). Frac-tions containing DNA fragments > 5 kb were pooled and ligatedto the BamHI-cut, dephosphorylated yeast 2 m plasmid YEP24.The ligation mixtures were then used to transform E. coli byelectroporation. Approximately 15 000 total independent trans-formants were obtained, pooled into two aliquots of <8000and 7000 clones, and restriction digests performed on ran-domly selected plasmid DNA clones indicated that <90%had an insert of an average size of 8 kb. Plasmid DNA pre-pared from pool I (<8000 clones) was introduced into S. cere-visiae cells, and the resulting transformants were selectedonto SC-ura plates (<15). A total of <30 000 colonies werescraped from the selective solid medium, pooled into 10 mlof SC-ura and 1 ml aliquots were frozen at ÿ808C.

Screening of S. cerevisiae transformants with

mAb-coated microspheres

Immunocapture of yeast transformants was performed withsheep anti-mouse IgG-linked magnetic microspheres (Dyna-beads M-280; Dynal) coated with mAb 4E1 (Deslauriers etal., 1996) and a particle concentrator (Dynal). Coating of themicrospheres (1 mg) with mAb 4E1 (1 ml of hybridoma super-natant; <50 mg mlÿ1 IgG) was done by a 2 h incubation atroom temperature in a rotary shaker. Coated beads werethen washed four times in phosphate-buffered saline supple-mented with 0.1% bovine serum albumin (PBS±BSA) for10 min each on a rotary shaker. As a negative control, themagnetic beads, either uncoated or coated with an irrelevantantibody, were incubated under the same conditions. Controlantibodies were monoclonal IgG1 anti-Candida enolase anti-bodies produced from hybridoma A2C7 (Strockbine et al.,1984) obtained from the American Type Culture Collection.For immunoscreening, yeast transformants growing exponen-tially into SC-ura medium were harvested by low-speed cen-trifugation and suspended into 1 ml of PBS±BSA at a ®nalconcentration of 1.5 OD600 mlÿ1. Coated beads (100 ml; 6 ´106 beads) were then added, and the suspension was rotatedfor 16 h at 48C. At the end of the incubation, the microsphereswere maintained on the wall of the tube by a lateral magnet,and the supernatant was discarded. The beads were washedfour times with PBS±BSA as described above. Free and cell-bound beads were ®nally recovered into 110 ml of PBS±BSAfor plating onto SC-ura plates (100 ml) and microscopic exami-nation (10 ml).

Magnetic bead sorting from homogeneous cultures of S.cerevisiae carrying either YEP24 or p4E1 was performedessentially as described above with the following modi®ca-tions. Exponentially growing cultures (OD600 mlÿ1 between 1and 1.5) were adjusted to a ®nal cell density of 105 cells mlÿ1

in PBS±BSA before the addition of the coated (mAb 4E1) oruncoated beads (106 mlÿ1). The incubation was also done atroom temperature for 2 h rather than overnight at 48C. Finally,serial dilutions (10-fold) of the free and cell-bound beads wereplated onto selective SC-ura solid medium.

Competition experiments were conducted by ®rst incu-bating the yeast transformant (105±106 cells) harvestedfrom exponential cultures and washed once with PBS±BSAwith either 1 ml of PBS±BSA or the mAb 4E1 supernatant



for 1 h at room temperature before sorting with the coatedmicrospheres.

Indirect immuno¯uorescence microscopy

Indirect immuno¯uorescence microscopy was carried outessentially as described by Pringle et al. (1991) using undilutedmAb 4E1 hybridoma supernatant as primary antibody and¯uorescein-conjugated goat anti-mouse IgG antibodies (Bio/Can Scienti®c) at a 1:500 ®nal dilution.

Disruption of CSA1

The 3.5 kb CAT-URA3-CAT cassette from plasmid pCUC(Fonzi and Irwin, 1993) was isolated by digestion withBamHI and inserted at the BamHI site of plasmid p4E1DSto create p4E1DS::CUC. The 4.6 kb fragment released fromp4E1DS::CUC by digestion with Hpa I was used to transformC. albicans CAI4. Early logarithmic cells (OD600 0.3) weretransformed with <5±10 mg of DNA. Cells were plated ontoSC-ura medium. Approximately three transformants per mgof DNA were visible after 3 days of incubation at 308C. Primarytransformants were replated onto SC-ura medium containinguridine (50 mg mlÿ1) and 58-¯uoorotic acid (FOA) (1 mg mlÿ1),and FOA-resistant colonies were subjected to further roundsof transformation. At each stage of this process, integrationof the disrupting cassette at the CSA1 locus was con®rmedby Southern analysis.

Acknowledgements

This study was supported in part by grants from the NaturalSciences and Engineering Research Council (Y.B.), CanadianMedical Research Council (N.D.) and the Fonds de laRecherche en SanteÂ du QueÂbec (N.D. and Y.B.). The authorswould like to thank Jessy Tremblay for the construction of theCandida albicans genomic library, Annie Roy for assistance inindirect immuno¯uorescence microscopy, and Brigitte CarreÂ

and Chantal Larouche for excellent technical assistance inthe early phase of this work. We are very grateful to JeanRenaud for help in the automated DNA sequencing reactions.We thank Dominick Pallotta and Michel Frenette for helpfulcomments on the manuscript.

References

Aruffo, A., and Seed, B. (1987) Molecular cloning of a CD28cDNA by a high-ef®ciency COS cell expression system.Proc Natl Acad Sci USA 84: 8573±8577.

Bailey, D.A., Feldmann, P.J., Bovey, M., Gow, N.A., andBrown, A.J. (1996) The Candida albicans HYR1 gene,which is activated in response to hyphal development,belongs to a gene family encoding yeast cell wall proteins.J Bacteriol 178: 5353±5360.

Brawner, D.L., and Cutler, J.E. (1986) Ultrastructural and bio-chemical studies of two dynamically expressed cell surfacedeterminants on Candida albicans. Infect Immun 51: 327±336.

Caro, L.H., Tettelin, H., Vossen, J.H., Ram, A.F., vanden Ende, H., and Klis, F.M. (1997) In silico identi®cation

of glycosyl-phosphatidylinositol-anchored plasma-membraneand cell wall proteins of Saccharomyces cerevisiae. Yeast13: 1477±1489.

Casanova, M., Gil, M.L., Cardenoso, L., Martinez, J.P., andSentandreu, R. (1989) Identi®cation of wall-speci®c anti-gens synthesized during germ tube formation by Candidaalbicans. Infect Immun 57: 262±271.

Chaf®n, W.L., Skudlarek, J., and Morrow, K.J. (1988) Vari-able expression of a surface determinant during prolifera-tion of Candida albicans. Infect Immun 56: 302±309.

Chaf®n, W.L., Lopez-Ribot, J.L., Casanova, M., Gozalbo, D.,and Martinez, J.P. (1998) Cell wall and secreted proteins ofCandida albicans : identi®cation, function, and expression.Microbiol Mol Biol Rev 62: 130±180.

Chardes, T., Piechaczyk, M., Cavailles, V., Salhi, S.L., Pau,B., and Bastide, J.M. (1986) Production and partial charac-terization of anti-Candida monoclonal antibodies. Ann InstPasteur Immunol 137C: 117±125.

Cox, R.A. (1989) Antigenic structure of Coccidioides immitis.Immunol Ser 47: 133±170.

Cutler, J.E. (1991) Putative virulence factors of Candida albi-cans. Annu Rev Microbiol 45: 187±218.

Deslauriers, N., Michaud, J., Carre, B., and Leveillee, C. (1996)Dynamic expression of cell-surface antigens probed withCandida albicans-speci®c monoclonal antibodies. Micro-biology 142: 1239±1248.

DeZwaan, T.M., Carroll, A.M., Valent, B., and Sweigart, J.A.(1999) Magnaporthe grisea pth11p is a novel plasmamembrane protein that mediates appressorium differentia-tion in response to inductive substrate cues. Plant Cell 11:2013±2030.

Dugger, K.O., Villareal, K.M., Ngyuen, A., Zimmermann,C.R., Law, J.H., and Galgiani, J.N. (1996) Cloning andsequence analysis of the cDNA for a protein from Cocci-dioides immitis with immunogenic potential. Biochem Bio-phys Res Commun 218: 485±489.

Fonzi, W.A., and Irwin, M.Y. (1993) Isogenic strain construc-tion and gene mapping in Candida albicans. Genetics 134:717±728.

Fu, Y., Rieg, G., Fonzi, W.A., Belanger, P.H., Edwards, Jr,J.E., and Filler, S.G. (1998) Expression of the Candidaalbicans gene ALS1 in Saccharomyces cerevisiae inducesadherence to endothelial and epithelial cells. Infect Immun66: 1783±1786.

Galgiani, J.N., Sun, S.H., Dugger, K.O., Ampel, N.M., Grace,G.G., Harrison, J., et al. (1992) An arthroconidial-spheruleantigen of Coccidioides immitis : differential expressionduring in vitro fungal development and evidence forhumoral response in humans after infection or vaccination.Infect Immun 60: 2627±2635.

Garzon, S., Marquis, G., Montplaisir, S., Kurstak, E., and Ben-hamou, N. (1989) Antigenic structure of Candida albicans.Electron microscopic localization of polysaccharide andimmunodeterminants in the cell wall. Immunol Ser 47: 3±36.

Gaur, N.K., and Klotz, S.A. (1997) Expression, cloning, andcharacterization of a Candida albicans gene, ALA1, thatconfers adherence properties upon Saccharomyces cere-visiae for extracellular matrix proteins. Infect Immun 65:5289±5294.

Hogan, L.H., Josvai, S., and Klein, B.S. (1995) Genomiccloning, characterization, and functional analysis of the



major surface adhesin WI-1 on Blastomyces dermatitidisyeasts. J Biol Chem 270: 30725±30732.

Hopwood, V., Poulain, D., Fortier, B., Evans, G., and Vernes,A. (1986) A monoclonal antibody to a cell wall componentof Candida albicans. Infect Immun 54: 222±227.

Hoyer, L.L., Scherer, S., Shatzman, A.R., and Livi, G.P. (1995)Candida albicans ALS1 : domains related to a Saccharo-myces cerevisiae sexual agglutinin separated by a repeat-ing motif. Mol Microbiol 15: 39±54.

Jenkinson, H.F., and Lamont, R.J. (1997) Streptococcal adhe-sion and colonization. Crit Rev Oral Biol Med 8: 175±200.

Kaiser, C., Michaelis, S., and Mitchell, A. (1994) LaboratoryCourse Manual for Methods in Yeast Genetics. Cold SpringHarbor, NY: Cold Spring Harbor Laboratory Press.

Kapteyn, J.C., Montijn, R.C., Dijkgraaf, G.J., and Klis, F.M.(1994) Identi®cation of beta-1,6-glucosylated cell wall pro-teins in yeast and hyphal forms of Candida albicans. EurJ Cell Biol 65: 402±407.

Kershaw, M.J., and Talbot, N.J. (1998) Hydrophobins andrepellents: proteins with fundamental roles in fungal mor-phogenesis. Fungal Genet Biol 23: 18±33.

Kershaw, M.J., Wakley, G., and Talbot, N.J. (1998) Comple-mentation of the mpg1 mutant phenotype in Magnaporthegrisea reveals functional relationships between fungalhydrophobins. EMBO J 17: 3838±3849.

Klein, B.S., Hogan, L.H., and Jones, J.M. (1993) Immuno-logic recognition of a 25-amino acid repeat arrayed in tan-dem on a major antigen of Blastomyces dermatitidis. J ClinInvest 92: 330±337.

Kohrer, K., and Domdey, H. (1991) Preparation of high mole-cular weight RNA. Methods Enzymol 194: 398±405.

Li, R.K., and Cutler, J.E. (1991) A cell surface/plasma mem-brane antigen of Candida albicans. J Gen Microbiol 137:455±464.

Madoff, L.C., Michel, J.L., Gong, E.W., Kling, D.E., andKasper, D.L. (1996) Group B streptococci escape hostimmunity by deletion of tandem repeat elements of thealpha C protein. Proc Natl Acad Sci USA 93: 4131±4136.

Odds, F.C. (1994) Pathogenesis of Candida infections. J AmAcad Dermatol 31: S2±S5.

Ollert, M.W., and Calderone, R.A. (1990) A monoclonal anti-body that de®nes a surface antigen on Candida albicanshyphae cross-reacts with yeast cell protoplasts. InfectImmun 58: 625±631.

Ponton, J., Marot-Leblond, A., Ezkurra, P.A., Barturen, B.,Robert, R., and Senet, J.M. (1993) Characterization ofCandida albicans cell wall antigens with monoclonal anti-bodies. Infect Immun 61: 4842±4847.

Poulain, D., Cailliez, J.C., and Dubremetz, J.F. (1989) Secre-tion of glycoproteins through the cell wall of Candidaalbicans. Eur J Cell Biol 50: 94±99.

Pringle, J.R., Adams, A.E., Drubin, D.G., and Haarer, B.K.(1991) Immuno¯uorescence methods for yeast. MethodsEnzymol 194: 565±602.

Ramasamy, R. (1998) Molecular basis for evasion of hostimmunity and pathogenesis in malaria. Biochim BiophysActa 1406: 10±27.

Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989) MolecularCloning: a Laboratory Manual. Cold Spring Harbor, NY:Cold Spring Harbor Laboratory Press.

Smail, E.H., and Jones, J.M. (1984) Demonstration andsolubilization of antigens expressed primarily on the sur-faces of Candida albicans germ tubes. Infect Immun 45:74±81.

Staab, J.F., Ferrer, C.A., and Sundstrom, P. (1996) Develop-mental expression of a tandemly repeated, proline- andglutamine-rich amino acid motif on hyphal surfaces onCandida albicans. J Biol Chem 271: 6298±6305.

Staab, J.F., Bradway, S.D., Fidel, P.L., and Sundstrom, P.(1999) Adhesive and mammalian transglutaminase sub-strate properties of Candida albicans Hwp1. Science 283:1535±1538.

Strockbine, N.A., Largen, M.T., and Buckley, H.R. (1984)Production and characterization of three monoclonal anti-bodies to Candida albicans proteins. Infect Immun 43:1012±1018.

Sundstrom, P.M., and Kenny, G.E. (1984) Characterizationof antigens speci®c to the surface of germ tubes of Candidaalbicans by immuno¯uorescence. Infect Immun 43: 850±855.

Sundstrom, P.M., Tam, M.R., Nichols, E.J., and Kenny, G.E.(1988) Antigenic differences in the surface mannoproteinsof Candida albicans as revealed by monoclonal antibodies.Infect Immun 56: 601±606.

Thomas, B.J., and Rothstein, R. (1989) Elevated recombi-nation rates in transcriptionally active DNA. Cell 56: 619±630.

Torosantucci, A., Palma, C., Boccanera, M., Ausiello, C.M.,Spagnoli, G.C., and Cassone, A. (1990) Lymphoprolifera-tive and cytotoxic responses of human peripheral bloodmononuclear cells to mannoprotein constituents of Candidaalbicans. J Gen Microbiol 136: 2155±2163.

Watari, J., Takata, Y., Ogawa, M., Sahara, H., Koshino,S., Onnela, M.L., et al. (1994) Molecular cloning and ana-lysis of the yeast ¯occulation gene FLO1. Yeast 10: 211±225.

Zhu, Y., Yang, C., Magee, D.M., and Cox, R.A. (1996a)Coccidioides immitis antigen 2: analysis of gene and pro-tein. Gene 181: 121±125.

Zhu, Y., Yang, C., Magee, D.M., and Cox, R.A. (1996b) Mole-cular cloning and characterization of Coccidioides immitisantigen 2 cDNA. Infect Immun 64: 2695±2699.



Documents

Expression cloning of the Candida albicans CSA1 gene encoding a mycelial surface antigen by sorting of Saccharomyces cerevisiae transformants with monoclonal antibody-coated magnetic