DNA array analysis of Candida albicans gene expression in response to adherence to polystyrene

Abstract

Candidiasis is often initiated by the colonization of inert surfaces. In order to elucidate the mechanisms involved in this adherence process, DNA macroarrays were used to analyze the transcriptome of Candida albicans, the main causative agent of this mycoses, in a simple adherence model using germ tubes produced in polystyrene Petri dishes. Non-adherent germ tubes produced on glass surface were used as a control. Analysis of gene expression displayed 77 genes identified as statistically overexpressed in adherent germ tubes. Among these genes, some encoded enzymes participating in metabolism of lipids (such as LIP6), of proteins (such as SAP1) or of carbohydrates (like PGI1, PMI40 and PSA1. Some of these genes have already been reported as playing a role in pathogenesis of C. albicans. However, functions were unknown for a large part (45.5%) of the overexpressed genes which will be analyzed further in order to define their relationship with adherence.

Candida albicans, Macroarrays, Adherence, Polystyrene

1 Introduction

Candida albicans which is the major fungal pathogen in humans, is a dimorphic fungus capable to cause superficial mucosal infections, as well as systemic infections in immunocompromised individuals [1]. During the past two decades, the prevalence of candidiasis has increased markedly and C. albicans has now become one of the most important cause of nosocomial infections [2]. The factors responsible for its pathogenesis are still not well understood, but several attributes related to the cell wall such as the ability to undergo dimorphic transition, as well as the production of extracellular proteinases and adherence, have been thought to contribute to its virulence [1,3,,5].

Adherence of the pathogens to the host cells is considered as an essential step in the establishment of infection. C. albicans has been shown to adhere to a large variety of epithelial cells, as well as to endothelial cells, but also to medical implants such as catheters and protheses. Multiple adherence mechanisms have been proposed for this yeast. The molecular interactions involved in adherence may be classified into non-specific (electrostatic or hydrophobic) interactions and specific ligand–receptor interactions [6,7]. These specific interactions, enhanced by morphogenetic transition from budding yeast to hyphae, are governed by multifunctional adhesins which are mannoproteins localized at the fungal surface [8]. Several studies have established the role of some fungal proteins in adherence to the host tissues. Among them, a hyphal wall protein (Hwp1p) and two agglutinin-like proteins (Als1p and Ala1p) have been suggested to be involved in adherence to epithelial and endothelial cells [9,10]. In addition, previous works from our group have been focused on the role of some cell wall mannoproteins of C. albicans in its interactions with plastic surfaces and with some plasma or matrix adhesive proteins. A set of three mannoproteins of 60, 68 and >200 kDa has been suggested to mediate the adherence of the yeast to inert surfaces, but also the binding of laminin and fibrinogen [11,,13]. For others, interactions with fibrinogen are mediated by a 58-kDa mannoprotein encoded by the fibrinogen binding protein 1 gene (FBP1) [14].

In order to get a better understanding of the adherence process, we have used an adherence model of C. albicans germ tubes to polystyrene surface. Gene expression was analyzed using DNA macroarrays containing 2002 Open Reading Frames (ORF) of the C. albicans genome [15], and non-adherent germ tubes produced on glass substrate as the control. These macroarrays enabled us to characterize a large part of the C. albicans transcriptome in response to adherence to plastic surfaces. Therefore, this study provides the first analysis of C. albicans gene regulation during the early steps of the development of candidiasis.

2 Materials and methods

2.1 Strain and culture conditions

C. albicans ATCC 66396 was used throughout. This strain was routinely maintained by biweekly passages on Sabouraud dextrose agar containing chloramphenicol 1 mg ml⁻¹. For germ tube production, 24-h-old blastoconidia were suspended in medium 199 pH 7 at a final concentration of 2 × 10⁶ cells ml⁻¹. The fungal suspension was then distributed in 140-mm neutral polystyrene Petri dishes (Fisher Labosi) or in 110-mm neutral glass Petri dishes (Duran^?– Fisher Labosi), at the rate of 50 ml or 31 ml per dish, respectively and incubated for 2 h 30 at 37 °C. Under these conditions, over 90% of the organisms produced germ tubes that adhered to the plastic surface [13].

Germ tubes were removed from the plastic surface with a rubber policeman and resuspended in culture supernatant. The obtained suspension as well as germ tube suspension obtained in glass Petri dishes were separately isolated from culture medium by filtration through 1.2-μm-pore size filters (Fisher Labosi). Cells were then washed in sterile distilled water, and finally resuspended in extraction buffer (Tris–HCl 100 mM pH 7.5 containing 100 mM LiCl and 1 mM EDTA) to reach a cell density of 5 × 10⁷ cells ml⁻¹. After centrifugation at 2350g for 10 min at 4 °C, pellets were stored at −80 °C until use.

2.2 RNA extraction and purification

Extraction of total RNA from germ tube pellets was performed with an equal volume of extraction buffer and a mix of phenol pH 5–chloroform–isoamyl alcohol (25:24:1 v/v; Sigma– Aldrich). SDS 20% (5 μl) and ice-cold glass beads (mix 1:1 of 0.25 and 1-mm diameter glass beads; Braun) were then added to the samples which were vortexed vigorously for 8 × 30 s with periodic chilling of 30 s on ice. After centrifugation at 2350g for 5 min at 4 °C, the supernatants were extracted twice with 0.5 ml phenol–chloroform–isoamyl alcohol. Total RNA was precipitated from supernatants with 0.3 M sodium acetate (Sigma–Aldrich) and 2.5 volumes absolute ethanol, then washed with 80% ethanol, dried and dissolved in 20 μl sterile distilled water. Total RNA concentration was calculated from the absorbance at 260 nm.

2.3 cDNA labelling and hybridization to DNA arrays

cDNA probes were synthesized and labelled according to the Yeast GeneFilters^? Microarrays protocol (http://www.invitrogen.com/content/sfs/manuals/genefilteryeast.pdf).

To do this, 1 μg of total RNA was mixed with 2 μl of oligo dT (Invitrogen Life Technologies) and heated at 70 °C for 10 min. After chilling on ice, 1.5 μl of dNTP mixture containing dATP, dGTP and dTTP (Amersham Pharmacia Biotech, 20 mM each), 1 μl DTT 0.1 M (Invitrogen Life Technologies), 6 μl of ³³P dCTP (10 mCi ml⁻¹ with a specific activity of 3000 Ci mmol⁻¹) (Perkin–Elmer NEN Life Sciences) and 1.5 μl reverse transcriptase (Superscript™ II RNase Transcriptase; Invitrogen Life Technologies) were added to the solution which was then incubated at 37 °C for 90 min. The volume was adjusted to 100 μl with standard sodium citrate buffer 1 × [(SSC) 0.15 M NaCl, 0.015 M citrate tri-sodique] and the free nucleotides were removed by gel filtration through a Sephadex Bio-Spin^? P-6 column (Bio-Rad).

DNA arrays used for hybridization experiments were purchased from Eurogentec. They consisted of 7 × 11 cm nylon membranes on which 2002 PCR-amplified open reading frames (ORFs) of the C. albicans genome were printed in duplicate. These ORFs were a random representation of the whole genome of C. albicans. Negative controls, consisting of PCR amplified DNA of Bacillus subtilis genome, were deposited on the arrays as well as positive controls which corresponded to dilutions of C. albicans genomic DNA.

The membranes were washed for 5 min in boiling 0.5% SDS and prehybridized for 2 h at 42 °C using roller bottles (Thermo Hybaid) with 5 ml MicroHyb solution (Invitrogen Life Technologies) containing 5 μl of Poly A (Invitrogen Life Technologies). The labelled cDNA probes were denatured for 10 min at 70 °C and added to the prehybridization mixture. Hybridization was carried out for 16–18 h at 42 °C. Then, each membrane was washed twice for 20 min at 50 °C with 30 ml of 2 × SSC supplemented with 1% SDS and once for 15 min at room temperature with 0.5 × SSC containing 1% SDS. The membranes were finally wrapped in plastic bags (Kapak) and exposed to a PhosphorImager screen (MP; Perkin–Elmer Life Sciences) for 24 h. For subsequent utilization, arrays were stripped with 1 l of boiling dehybridation solution (0.5% SDS) for 1 h.

2.4 Experimental design, image acquisition and data analysis

In our experimental design, macroarray hybridizations were performed nine times, using for each experiment, new cDNA probes derived from independent cell preparations. All filters were used up to three times.

Exposed PhosphorImager screens were scanned on a PhosphorImager (Cyclone^? Storage Phosphor System; Packard) to obtain a digital image of the membranes, and analyzed by the ArrayVision™ 6.0 commercial software (Imaging Research, Inc.). To determine induction or repression of gene expression, all spots intensities were compared between the filters corresponding to both culture conditions (cultures on plastic or glass surfaces). Data were analyzed after normalization of each spot. The normalized value of a spot was calculated from the spot signal minus the background level of hybridization (determined from the intensity of the signal surrounding the entire filter), divided by the median value of all the signals detected on the filter.

Statistical evaluation of the data was performed on SPSS software version 10.1 using the Wilcoxon non-parametric test, based on the ranks of the paired differences between the two samples (cultures on plastic or glass surfaces). A p value <0.05 was considered significant.

2.5 Reverse transcription and quantitative PCR

The differential gene expression revealed by macroarray hybridization was confirmed by reverse transcription and quantitative PCR (RT-PCR) for some of the genes.

In this aim, 2 μg of DNAse-treated total RNA was mixed with 1 μl (2.65 μg μl⁻¹) of random hexamers (Amersham Pharmacia biotech) and completed to 15 μl with distilled water. After heating at 70 °C for 5 min and chilling on ice, 10 μl of 5X first strand buffer (Promega) were added to the mix, together with 2 μl of 10 mM dNTPs (Sigma–Aldrich), 1 μl of 40 U μl⁻¹ recombinant RNasin ribonuclease inhibitor (Promega), 1 μl of 200 U μl⁻¹ M-MLV reverse transcriptase (Promega) and 21 μl of distilled water. The samples were incubated at 37 °C for 1 h. cDNAs were purified with the QIAquick PCR purification kit (Qiagen) and stored at −20 °C until use.

Quantitative PCR reactions were carried out in a M × 4000 sequence detection system (Stratagene) with Brilliant^? SYBR^? Green QPCR Master Mix (Stratagene), in a 25-μl final volume. Each reaction contained 150 nM of each upstream and downstream primers and 2.5 mM MgCl₂. PCR was optimized by following the recommendations of the manufacturer (Stratagene). PCR cycling conditions were as follows: initial denaturation at 95 °C for 10 min, followed by 40 cycles of 30 s at 95 °C, 1 min at a temperature adapted for each primer set used for annealing, and 30 s at 72 °C for extension. The identity of the PCR products was checked by melting curve analysis on the Mx4000.

Transcript levels of a housekeeping gene (ACT1) were evaluated to normalize the values, and data were analyzed by using the C_T comparative method.

3 Results and discussion

Macroarrays with the C. albicans cDNA probes allowed us to investigate global changes in gene expression associated with adherence to polystyrene. Results were not strictly identical between the nine different biological experiments, but normalization of each spot and then statistical analysis using the Wilcoxon non-parametric test allowed comparisons between the experiments. In addition, no differences in gene expression between the two culture conditions (adherent germ tubes produced on plastic surface compared to non-adherent organisms obtained on glass surface) were seen for negative controls (Bacillus subtilis DNA) nor for positive controls (C. albicans genomic DNA) in the nine experiments.

3.1 Gene expression on plastic surface versus glass surface

The Wilcoxon T test displayed a total of 77 C. albicans genes statistically overexpressed in adherent germ tubes versus non-adherent cells, and 40 genes were identified as statistically underexpressed. These genes, presented in Table 1, were assigned to functional categories on the basis of homology with S. cerevisiae genes (http://www.pasteur.fr/recherche/unites/RIF//transcriptdata/index.html).

1

Genes of C. albicans differentially expressed (p < 0.05) in germ tubes produced on plastic surfaces compared to glass surfaces

Gene/IPF*	Description of gene product	p value
Amino acid and protein metabolisms
Overexpressed
	CaYML6	Mitochondrial ribosomal protein	0.038
	CaEFT3	Translation elongation factor	0.021
	CaMRPL19	Ribosomal protein	0.038
	CaFMT1	Methionyl-tRNA transformylase	0.036
	CaTRM1	N2,N2-dimethylguanine tRNA methyltransferase	0.021
	IPF 56 homolog of ScNAS6	Nas6p subunit of 26S proteasome	0.038
CaSAP1	Secreted aspartyl protease	0.038
CaHYU1	Hydantoin utilization protein A	0.038
Underexpressed
	CaODC1	Ornithine decarboxylase	0.021
IPF 8591 homolog of ScCAR1	Putative arginase family member	0.011
1036IPF 96 homolog of ScITT1	C3HC4type zinc finger protein	0.011
Lipid, fatty acid metabolism
Overexpressed
	CaLIP6	Secretory lipase 6	0.028
	CaACH1	Acetyl-CoA hydrolase	0.038
CaCOQ5	C-methyltransferase	0.011
Homolog of ScSPO14	Phospholipase D	0.021
Underexpressed
CaGUT1	Glycerol kinase	0.038
Carbohydrate metabolism
Overexpressed
	CaPMI40	Phosphomannose isomerase	0.038
	CaPSA1	GDP-mannose pyrophosphorylase	0.021
	CaPGI1	Glucose-6-phosphate isomerase	0.011
CaCDC19	Pyruvate kinase	0.015
CaARA1	d-arabinose dehydrogenase	0.021
Underexpressed
CaMDH1	Mitochondrial malate dehydrogenase	0.038
Cellular organization and transport
Overexpressed
	CaATS1	α-Tubulin suppressor	0.038
	CaARP1	Centractin	0.035
	CaTUB4	γ-Tubulin	0.028
	CaHTB1	Histone H2B	0.028
	CaCDC28	Cell-division control protein 28-protein kinase	0.021
	IPF 3355 homolog of ScCDC1	Cell division control protein	0.011
	IPF 11212 homolog of ScBUB2	Cell cycle arrest protein	0.028
	CaTOP1	Canal DNA topoisomerase I	0.021
	IPF 4085 homolog of ScAPG1P	Essential for autophagocytosis	0.028
	CaSSC1	Mitochondrial heat shock protein 70-related protein	0.011
	CaVPS35	Protein-sorting protein, vacuolar	0.011
	CaVPS1	Member of the dynamin family of GTPases	0.021
	CaMIR1	Phosphate transport protein, mitochondrial (MCF)	0.011
	CaZRT1	High affinity zinc transport protein	0.021
	CaDJP1	DnaJ-like protein involved in peroxisome biogenesis	0.038
IPF 13042 homolog of ScIMH1	Vesicular-mediated transport	0.011
IPF 11873 homolog of ScSWA2	Swa2p clathrin binding protein	0.008
Underexpressed
	CaMYO5	Myosin I	0.011
	CaVRP1	Verprolin	0.011
CaGTT1	Glutathione S-transferase	0.038
CaSAC7	GAP for RHO1	0.011
Homolog of ScSIM11	Cell cycle regulation	0.038
CaPHR2	PH-regulated protein 2	0.028
CaSEC4	GTP-binding protein	0.008
CaVPS24	Endosomal Vps protein complex subunit	0.028
CaVPS8	Vacuolar sorting protein	0.043
IPF 4537 homolog of ScAVT7	V1 small nuclear ribonucleoprotein	0.011
Transcription
Overexpressed
	CaSMD3	Core snRNP protein	0.008
	CaPRP5	Pre-mRNA processing RNA-helicase	0.017
CaPRP22	RNA-dependent ATPase	0.028
CaRPA12	DNA-directed RNA polymerase I	0.011
Underexpressed
	CaCTA4	Probable transcription factor	0.043
CaSPT20	Transcription factor, member of the histone acetyltransferase	0.011
CaBDF1	Transcriptor regulator activity	0.008
Energy production
Overexpressed
	CaABC1	Ubiquinol-cytochrome-c reductase	0.021
	CaCYC3	Cytochrome c heme lyase	0.038
Others
Overexpressed
	CaCBP1	Corticosteroid binding protein	0.008
	CaHOG1	Ser/thr protein kinase of MAP kinase family	0.038
	Underexpressed
CaSBP1		RNA binding protein like	0.021
CaMAK32		Sugar kinase	0.028
CaCTA1	Catalase A, peroxisomal	0.015
IPF 8044 homolog of ScSNP1	Similar to ScSnp1p	0.028

Among the genes differentially expressed, 53 (35 overexpressed and 18 underexpressed) were of unknown function.

*IPF : Individual Protein File.

Among genes identified as statistically overexpressed on plastic surface, 22% played a role in cellular organization and intracellular transport, 10.4% were involved in amino acid and protein metabolism, 6.5% in carbohydrate metabolism, 5.2% in lipid and fatty acid metabolism, 5.2% in transcription, and functions were unknown for 45.5% of these genes (Fig. 1). Differential expression of genes LIP6, SAP1, PMI40 and of two genes with unknown function (IPF17021 and IPF2147) was confirmed by quantitative RT-PCR (Table 2). The agreement between data from RT-PCR and the macroarray transcription profile underlined the relevance of the macroarray data in these experimental conditions. The distribution of the 40 genes underexpressed on plastic surface versus glass surface among the different functional categories was equivalent to that observed for the overexpressed genes (data not shown).

1

Pie chart presenting the genes overexpressed on plastic surface versus glass surface grouped in functional classes (n= 77); functions of 45.5% of the overexpressed genes are unknown.

2

Confirmation by quantitative RT-PCR of the differential expression of some genes identified by macroarray hybridization

Gene	2-(ΔΔCt)a	p valueb
SAP1	5.8	0.038
PSA1	1.23	0.021
LIP6	1.35	0.028
IPF 17021	2.68	0.028
IPF 2147	6.10	0.028

^aΔΔCt was obtained by substracting the average Ct value of the housekeeping gene ACT1 from the average Ct value of the gene studied and ΔΔCt was the difference between ΔCt obtained for C. albicans germ tubes produced in adherence condition (polystyrene surface) and ΔCt obtained for C. albicans germ tubes obtained in non-adherence condition (glass surface).

^bData were calculated from the results obtained by DNA macroarray analysis.

Thus these results indicate a differential expression of some genes associated with adherence to polystyrene. The main changes concerned genes involved in cellular organization and transport since 27 genes (17 overexpressed and 10 underexpressed) were found to be differentially expressed. Among them, genes involved in cytoskeletal modification such as ATS1, ARP1, TUB4 were overexpressed in our study while MYO5 was underexpressed. Other overexpressed genes like HTB1, CDC28, IPF 3355 and IPF 11212 play a role in cell division or cell regulation, but a gene homolog of S. cerevisiae SIM11 was underexpressed. Four VPS genes encoding proteins required for the sorting of vacuolar proteins presented a differential gene expression. VPS1 and VPS35 were overexpressed while VPS8 and VPS24 appeared underexpressed. However, all the VPS gene products do not play the same role in sorting and/or in delivery of proteins to the vacuole [16,,18]. For example, trafficking in the endosomal system was analyzed by monitoring the movement of a plasma membrane ATPase in several vps mutants defective in vacuolar protein sorting. These experiments showed that the biosynthetic membrane traffic follows different routes in vps8 and vps1 mutants. For instance, in vps8, in contrast to vps1, the plasma membrane ATPase moves to the surface via endosomal intermediates, implicating an endosome-to-surface traffic pathway [18]. Thus one may speculate that some Vps proteins could be involved in the transfer of some mannoproteins mediating the adherence to inert surfaces or synthesized in response to adherence.

In addition, some of the overexpressed genes encode general components of the transcriptional machinery, particularly components involved in pre-mRNA splicing (SMD3, PRP22, PRP5) [19,,,22]. Likewise, other overexpressed genes encode enzymes participating in lipid, protein or carbohydrate metabolisms and some of them have already been described as playing a role in pathogenesis of C. albicans. For example, expression of the gene LIP6 which encodes one of C. albicans lipases [23], was enhanced in germ tubes adhering to polystyrene. It has been shown that the lipolytic activity of these enzymes could increase hydrophobic interactions through the release of fatty acids [22,24,25]. Therefore, one may speculate that these enzymes play a role in the colonization of prosthetic devices [23].

Genes encoding enzymes participating in protein metabolism may also be involved in pathogenesis. For example, the relationship between the production of the secreted aspartic proteinase Sap1p and invasiveness of C. albicans is well known [26,27]. Here we showed that the expression of SAP1 which is regulated during the yeast to mycelium transition, is also enhanced during the adherence process. In contrast, the expression of SAP2 and SAP3 which have also been suggested to be involved in virulence [26], was not modified between the two culture conditions, suggesting that the corresponding enzymes do not play any role in adherence to inert surfaces.

Regarding genes involved in carbohydrate metabolism, three of the overexpressed genes encode enzymes directly involved in the GDP-mannose synthesis pathway: the gene PGI1 encoding the phosphoglucose isomerase involved in the synthesis of fructose-6-phosphate; the gene PMI40 encoding the phosphomanose isomerase, an enzyme involved in mannose catabolism, but also in the production of the GDP-d-mannose which is necessary for protein mannosylation; and the gene PSA1 encoding a GDP-mannose pyrophosphorylase involved in the synthesis of GDP-mannose which is essential for the addition of mannose residues. [28,,,31]. These results suggest that adherence to polystyrene is associated with an increased synthesis of GDP-mannose, leading to an increased synthesis of mannoproteins or to an increased mannosylation. Interestingly, previous studies from our laboratory [13] have demonstrated at the surface of germ tubes adherent to polystyrene, the presence of an outermost fibrillar layer, mannoproteinic in nature and allowing the interconnections germ tubes–substrate. In contrast, this network was not detected at the surface of non-adherent germ tubes produced on glass surface.

3.2 Analysis of genes involved in adherence phenomenons

In the present study, a simple experimental model was used to analyze the expression of genes associated with adherence to inert surfaces. In this model, adherent germ tubes of C. albicans were produced in neutral polystyrene Petri dishes and compared to non-adherent cells produced on glass surface. This enabled us to analyze gene expression during the early steps of infection contrary to previously published works studying later states in the development of candidiasis, for example the gene expression pattern in C. albicans biofilms obtained after a 48–72-h incubation [32]. In addition, due to the use of non-adherent germ tubes as the control, this study was focused on the expression of genes associated with adherence, independently of hyphal development. Besides, genes involved in filamentation such as ALS1, CPH1 or EFG1[33] were not overexpressed in our experimental model.

The macroarrays used here were the only tools available at the beginning of this study. However, these arrays covered only about one third of the 6000 genes expected in the whole C. albicans genome. Thus, some genes suggested to be involved in adherence such as FBP1, and MNT1 which encodes an α-1-2 mannosyltransferase adding a second mannose to singly mannosylated hydroxy amino acids, were not present on the array filters [14,34]. In contrast, other genes involved in adherence and deposited on the array filters were not found to be overexpressed on plastic versus glass surfaces. For example, the expression of HWP1 and ALS1 was clearly unchanged in our adherence model to inert surface [34,35]. If it is not surprising that HWP1 was not among the overexpressed genes, since it is likely to be involved in covalent binding to epithelial cells [36], an overexpression of ALS genes has been reported in biofilm populations [32,37]. One may speculate that the expression of ALS1 was not altered in our adherence model due to the short cultivation time whereas fungal biofilm formation requires a longer incubation of 48–72-h.

In conclusion, this study provides a better knowledge of the molecular mechanisms associated with the colonization of inert surfaces which may have important clinical involvements. Indeed, after adherence to medical devices, the yeast may contribute to deterioration of the devices and/or initiate an acute disseminated infection. Organisms adhering to plastic may also be less susceptible to antifungal drugs. In addition, the apparent lack of overexpression of known adhesin genes, joint to the overexpression of LIP6 and genes involved in mannosylation of proteins further support the importance of hydrophobic interactions in the adherence of C. albicans to inert surfaces. Our objective is now to study the genes with unknown function which were found to be overexpressed on plastic. The deduced amino acid sequences of these genes will be analyzed by using a bioinformatic approach to select putative cell-surface proteins. Potentially interesting genes will be disrupted and the obtained mutants will be studied in vitro in our adherence model.

Acknowledgements

The authors wish to thank Jean-Michel Camadro (UMR 7592 CNRS, Institut Jacques Monod, Paris, France), Claude Gaillardin (CNRS URA 1925 INRA UR 216, Thiverval-Grignon), and Bruno Vielle (CHU, Angers, France) for their expert assistance. This work was supported by the Ministère de la Recherche et de la Technologie (Réseau Infections fongiques).

Author notes

Véronique Marchais and Marie Kempf contributed equally to this work