Time course global gene expression analysis of an in vivo Candida biofilm

Abstract

Candida device infections are common and invariably associated with biofilm growth. Exploratory microarray studies were undertaken to identify target genes associated with biofilm formation from an in vivo catheter model over time. We compared mRNA levels from C. albicans grown in an in vivo central venous catheter biofilm model at 12h (intermediate growth) and 24h (mature) to in vitro planktonic cells without a biofilm substrate using C. albicans oligo-arrays. 124 transcripts were similarly upregulated at the 12 and 24h timepoints. Ontology categories most highly represented included energy/metabolism (12%), carbohydrate (10%) and protein (13%) synthesis and modification, and transport (6%). Numerous genes were previously identified from in vitro biofilm studies. These genes included those associated with hyphal growth, amino acid metabolism, adherence, drug resistance, ergosterol biosynthesis, and β-glucan synthesis. In the current data set, adherence genes were unique to the earlier timepoint. Differences between the current in vivo biofilm expression data and that previously reported from in vitro models, including alterations in metabolism and carbohydrate processing, may be due to the continuous availability of nutrients from host serum and the incorporation of the host-pathogen interaction.

INTRODUCTION

Candida albicans is an important human pathogen causing a wide spectrum of diseases and is a leading cause of nosocomial bloodstream infection. Candidemia is frequently associated with implantantion of a medical device. When growing on the surface, such as a venous catheter or urinary catheter, Candida adapts to a biofilm lifestyle [1, 2]. Biofilm growth involves phenotypic changes distinct from planktonic growth [3-5]. Two such characteristics, drug-resistance and reduced susceptibility to the host immune system, contribute to the difficulty in treating Candida biofilm infections. Greater understanding of the gene expression patterns present in Candida biofilms is one method to begin to discern the pathways important for the biofilm phenotype and identify potential antifungal targets.

Many in vitro model systems have been developed to mimic the biofilm growth found on infected medical devices. These models have provided the foundation for investigation of biofilm composition, architecture, and mechanisms of drug-resistance [3, 5, 6]. More recently, transcription profiling experiments have identified potential biochemical pathways required for biofilm formation [7-9]. Although the substratum common to these clinical infection sites (i.e. catheter, denture material) can be used in vitro, the models lack other environmental conditions potentially important in human infections [10]. In device-associated infections, the substratum has been preconditioned with host proteins soon after implantation [1, 10]. This preconditioning film impacts biofilm formation and extracellular matrix production. In vitro models are not necessarily exposed to all the host proteins needed for the appropriate preconditioning film. Also, in many in vitro models, nutrients are depleted and waste products accumulate. However, in vivo biofilm cells are exposed to a continuous nutrient supply and waste products are eliminated while in the vasculature or other host tissue site [11]. Imaging of in vivo biofilms has also demonstrated incorporation of host cells, which may potentially impact biofilm formation. This host-pathogen interaction is not represented in these in vitro models. Finally, flow dynamics of host vasculature are difficult to reproduce in vitro.

In the current experiments, we utilize an in vivo rat catheter biofilm infection model [11]. This central venous catheter rat model mimics conditions encountered in human infections, including exposure to host proteins and immune components, host vasculature, and biofilm substratum. Inclusion of these variables in the model system allows us to account for the host-pathogen interaction and examine its impact on the transcriptional profile of Candida during biofilm development. The importance of the host-pathogen interaction, by use of in vivo model systems, has been previously demonstrated in study of both the host and pathogen transcriptomes [12]. The current study demonstrates the relevance and importance of incorporating the host infection site and immune system for the biofilm process.

METHODS

In vivo Venous Catheter Biofilm Formation

A C. albicans central venous catheter biofilm model was used for in vivo experiments as previously described [11, 13, 14]. Following catheter placement for 24h, a C. albicans SC5314 inoculum of 10⁶ cells/ml in 0.15M NaCl was instilled in the catheter. Catheters were removed for RNA isolation or imaging following a 12 or 24h incubation. To yield a sufficient amount of RNA for the microarray analysis, Candida cells from 2-5 catheters were pooled. Collected cells were flash frozen with liquid nitrogen in AE buffer (50 mM sodium acetate pH 5.2, 10 mM EDTA) prior to RNA processing as previously described. RNA was collected using the hot method as previously described [11]. The RNA integrity was assessed using Agilent Bioanalyzer 2100 with a RNA Nanochip before use in microarrays.

Planktonic Culture Conditions

C. albicans SC5314 was propagated overnight in YPD at 37°C on orbital shaker at 250RPM and used for inoculation of planktonic cultures. Planktonic cultures were obtained in parallel with in vivo cultures by incubated the cells in 20 ml of RPMI in a glass flask at 37°C on orbital shaker at 250RPM. Log phase planktonic cells were harvested by centrifugation and the cells were flash frozen with liquid nitrogen in AE buffer until RNA processing. RNA was collected and integrity investigated as described above.

Biofilm Scanning Electron Microscopy (SEM)

Catheters were harvested at each timepoint for imaging as previously described [11]. Samples were imaged in a JEOL JSM-6100 in the high-vacuum mode at 10kV. The images were assembled using Adobe Photoshop 7.0.

Microarray and RT-PCR

C. albicans oligo microarray slides were purchased from the Biotechnology Research Institute, National Research Council, Montreal (Version Oligo Trial 2). Three biologic replicates and dye swaps were performed for each condition and timepoint as previously described [12, 15, 16]. The microarrays were scanned with an Agilent DNA microarray scanner (G2565BA). Lowess normalization and statistical analysis were performed using Genespring v.7 (Agilent Technologies, CA). Genes whose expression was either increased 1.5-fold (log2) or decreased 0.67-fold for C. albicans biofilm samples relative to planktonic samples were considered for further analysis. A two-sided Student’s t test was used determine the statistical significance of the log ratios.

Quantitative real-time reverse transcription-PCR (RT-PCR) was used to confirm mRNA abundances of a subset of upregulated transcripts from the microarray studies. The TaqMan probe and primer sets were designed using Primer Express (Applied Biosystems, Foster City, CA) (Table 1). The QuantiTect probe RT-PCR kit (Qiagen, Valencia, CA) was used in an ABI PRISM 7700 v1.7 sequence detection system (Applied Biosystems) as previously described. Reactions were performed in triplicate. The quantitative data analysis was completed using the C_t (2^−C_t) method [17]. The comparative expression method generated data as transcript fold-change normalized to a constitutive reference gene transcript (ACT1) and relative to a control or baseline condition.

Table 1.

Primers used for RT-PCR experiments

Gene	Primers	Probe
ALS1	Forward: 5′ -TTAGCATCTGCAAGTGAAGAAGACA Reverse: 5′ -ATGGAGCTTCTGTAGGACTGGTTG	5 ′-/56-FAM/AGCGGTTCTCATGAATCAGCATCCACAA/36-TAMSp
ARG1	Forward: 5′ -GTGAAGTTAGAGCCATCAGAGATCAA Reverse: 5′ -TGAACGAACGTATTCTCCTTCTGG	5′ -/56-FAM/TTGTGACACCAACTTGGGCCAAAATCTT/36-TAMSp
BGL2	Forward: 5′ –CTCGCAACTGTTCTTACTTCAGTTG Reverse: 5′ -TGACGTCTTTACAAGTACCGTCATC	5′ -/56-FAM/TGAAAGCCAAATCACCCATGGCG/36-TAMSp
CDR2	Forward: 5′ -CCACATGTCCGACATACCTGG Reverse: 5′ -GGAATCTGGGTCTAATTGTTCATGA	5′ -/56-FAM/CCATTCAACGGCAACAT/36-TAMSp
ECM3	Forward: 5′ -TCCACAGTGAAGGGTCAGCTG Reverse: 5′ -CGCAAATCGCTTGTCCTAGAA	5′ -/56-FAM/AAGACGCCGTACATCAAATGCATCTGAACT/36-TAMSp
ERG25	Forward: 5′ -TGGATTGGCAGCAGAATATGC Reverse: 5′ -GCAAGTTACCAGTGATAAGACACCA	5′ -/56-FAM/CAATCGGAATACCAACCGTACCCAATCCTA/36-TAMSp
FET3	Forward: 5′ -TCGGTGTGTCATGGGAAGGTA Reverse: 5′ -ACCTCTAGCAGTGAAACCAGTTGG	5′ -/56-FAM/TGCCGCCGCCAACAGTAAAGACTACTT/36-TAMSp
HGT2	Forward: 5′ -AGGTCAATGGGAAGCTGCTG Reverse: 5′ -TAGCGTATCCAATATGCTTGCTTG	5′ -/56-FAM/CACGATCACCATGAGCTTGAACCTTGG/36-TAMSp
INO1	Forward: 5′ -GTCTTACTGGCTCAAGGCTCCA Reverse: 5′ -CCAACACTGAAAGCAAGTTGACTAA	5′ -/56-FAM/AGCAAGACCAGGATTCAAACCTATCAACGG/36TAMSp
KRE1	Forward: 5′ -GGCAGATAAAACATCCAGTTCAGTATC Reverse: 5′ -CCCAATCCAATAGAACCAGACG	5′ -/56-FAM/TTTCCCATTGGCATCAGTACCTGTGACC/36-TAMSp
PCK1	Forward: 5′ -TACTCTGGTAACAAAACCGGTAGATC Reverse: 5′ -ACTTGTTTATTCACTGGACCCCA	5′ -/56-FAM/TGTCGACGAATCCACCTCATCCCAT/36-TAMSp
PHR1	Forward: 5′ -CTGCAAAGCTGTTAGTGGAGTAGC Reverse: 5′ -TGATGATCCAGAAGTAGATGCAGAG	5′ -/56-FAM/TCCTCTGGTGGAAGCTCCAAATCTGG/36-TAMSp
SAP5	Forward: 5′ -CCAGATGATCCAACTGTTGAAGC Reverse: 5′ -TCCACATCACCATCTCTACCAGTC	5′ -/56-FAM/AGTTTACTCCCTCAGAATTTCCCGTCGATG /36-TAMSp
ACT1	Forward: 5′ -AGCTTTGTTCAGACCAGCTGATT Reverse: 5′ -GGAGTTGAAAGTGGTTTGGTCAA	5′ - /56-FAM/CCAGCAGCTTCCAAACCT/36-TAMSp

RESULTS

In vivo Candida Biofilm Model

Images of the intermediate biofilm (12h) demonstrated yeast and hyphal cells attached to the substratum. At 24h, the mature biofilm consisted of an extensive network of yeast and hyphae embedded in strands of heterogeneous matrix material (Figure 1). Host cells, including red and white blood cells, were also embedded in the Candida biofilm.

Figure 1.

Scanning electron micrograph of C. albicans in vivo biofilm. C. albicans SC5314 was inoculated into a rat venous catheter and allowed to dwell for 24h. Catheters were placed in a fixative, treated with osmium tetroxide and dehydrated through a series of ethanol washes and critical point drying.

Transcript Profiles of Intermediate and Mature In vivo Candida Biofilms

The genome-wide changes in gene expression in intermediate and mature Candida biofilms are shown in Supplemental Tables 1 and 2. From the intermediate biofilm and planktonic comparison, we identified 545 transcripts of the 6,354 ORFs (8.6%) as differentially regulated at least 1.5 fold. Within this group, 457 transcripts were upregulated during biofilm formation (biofilm/planktonic ratio greater than 1.5) and 88 transcripts were downregulated (biofilm/planktonic ratio less than 0.67). In the comparison of mature (24h) biofilm cells and planktonic cells, 1034 transcripts (16.3%) were differentially regulated. 523 transcripts were upregulated during mature biofilm formation and 511 transcripts were downregulated.

Of the genes differentially expressed in intermediate biofilm formation, multiple functional categories were represented (Table 2). During intermediate biofilm development, transcripts involved in protein synthesis (17%), transport (5%), stress (4%), amino acid metabolism (4%), cell wall metabolism (3%), energy and general metabolism (9%), and carbohydrate processing (7%) were abundant. The functions of 44% of the genes with transcriptional upregulation in intermediate biofilm formation have not yet been identified. Transcripts reduced during intermediate biofilm formation were primarily involved with DNA processing and cell cycle (19%), although changes in transcription and protein synthesis (10%), transport (5%), cytoskeleton (5%), and carbohydrate processing (6%) were also evident.

Table 2.

Frequency distribution of differentially expressed genes among functional categories

Description	No. of genes in functional category (%)
	Biofilm Upregulated 12 h	Biofilm Downregulated 12 h	Biofilm Upregulated 24 h	Biofilm Downregulated 24 h
Amino Acid	18 (4)	2 (2)	21 (4)	15 (3)
Carbohydrate	30 (7)	5 (6)	27 (5)	17 (3)
Cell Wall	15 (3)	3 (3)	22 (4)	18 (4)
Cytoskeleton	2 (0)	4 (5)	5 (1)	15 (3)
DNA/Cell Cycle	10 (2)	17 (19)	24 (5)	38 (7)
Energy/Metabolism	41 (9)	1 (1)	40 (8)	28 (5)
Fatty Acid	6 (1)	0 (0)	3 (1)	0 (0)
Heat Shock Protein	1 (0)	0 (0)	3 (1)	0 (0)
Hyphal	4 (1)	3 (3)	8 (2)	6 (1)
Iron	9 (2)	0 (0)	9 (2)	4 (1)
Plasma Membrane	2 (0)	3 (3)	15 (3)	13 (3)
Resistance/Stress	20 (4)	2 (2)	9 (2)	10 (2)
Transcription/Protein	76 (17)	9 (10)	91 (17)	67 (13)
Transport	24 (5)	3 (5)	28 (5)	37 (7)
Vacuole	0 (0)	0 (0)	2 (0)	12 (2)
Unknown	199 (44)	36 (41)	216 (41)	231 (45)
Total	457	88	523	511

Functional categories are based on the Candida Genome Database http://www.candidagenome.org

In the mature biofilm, we noted similar changes in gene regulation (Table 2). Again, multiple functional categories were represented. Transcription of genes associated with protein processing (17%), energy and metabolism (8%), transport (5%), DNA/cell cycle (5%), and carbohydrate processing (5%) was upregulated in the mature biofilm cells. Similar to the intermediate biofilm, mature biofilm formation was associated with reduction of transcripts involved in DNA processing and cell cycle (7%), transcription and protein synthesis (13%), and transport (7%). The functions of near 40% of the differentially regulated genes are not yet known.

Next, we identified transcript abundance during both the intermediate biofilm and mature biofilm cells as a marker for relevance throughout the majority of biofilm formation. Transcript abundance of 124 genes was found at both the 12h and 24h timepoints, when compared to cells growing in planktonic conditions (Table 3). Genes involved in carbohydrate synthesis and processing (10%), transcription and protein synthesis (13%), and energy and metabolism (12%) were among those highly expressed at both timepoints. Transcripts involved in cell wall metabolism (4%), iron metabolism (3%) and amino acid metabolism (5%) were also abundant. Transcription of only 27 genes was downregulated in both the intermediate and mature biofilms (Table 4). Genes involved in DNA processing (30%) were most represented among the reducted transcripts.

Table 3.

Transcripts upregulated in both intermediate and mature biofilms

Category	Genes (Candida Genome Database)	No. genes in functional category
Amino acid	AGP2, ARG1, CAN1, CDG1, CPA1, CPA2, SMM1	6 (5)
Carbohydrate	CAT8, CIT1, DLD3, FBP1, GAL1, GAL10, HGT1, HGT2, HGT14, KGD1, MLS1, PAM16, PCK1	13 (10)
Cell wall	BGL2, CYC3, ECE1, FMP45, orf19.6489	5 (4)
Cytoskeleton	SDA1	1 (1)
DNA/Cell Cycle	AAH1, orf19.1005, orf19.5845	3 (2)
Energy/Metabolism	ALD6, AOX2, ARG8, COX11, COX 17, CTN1, GUT2, ICL1, INO1, NCE103, NDE1, OYE22, OYE23, YAT2, orf19.361	15 (12)
Fatty Acid	FAA2, orf19.732	2 (2)
Heat Shock Protein	HSP30,	1 (1)
Hyphal	RLP1	1 (1)
Iron	CFL2, FRE10, ISU1, SIT1	4 (3)
Plasma Membrane	ERG25, PDR16	2 (2)
Resistance/Stress	AQY1, EBP1, MDR1, SOD6, YHB3	5 (4)
Transcription/Protein	CIC1, DBP1, DBP7, MRPL33, MRT4, MSS116, MTO1, PEX4, REX2, RLP7, RPC19 RRN11, SAP5, SOF1, UGA32, orf19.124	16 (13)
Transport	GNP1, JEN2, LEU5, OAC1, SFC1, TIM9, TNA1	7 (6)
Unknown		43 (35)
Total		124

Table 4.

Transcripts downregulated in both intermediate and mature biofilms

Category	Genes (Candida Genome Database)	No. genes in functional category
Amino acid	GLT1	1 (4)
Carbohydrate	UGP1, PFK2	2 (7)
Cell wall	CWH43, SCW11	2 (7)
Cytoskeleton	DYN1, IQG1	2 (7)
DNA/Cell Cycle	HHF1, HHF22, HTA2, PCL2, POL30, RNR3, TMP1, orf19.7634	8 (30)
Plasma Membrane	ERG7, PLB1	2 (7)
Unknown		10 (37)
Total		27

Validation of Microarray Data by Real-time RT-PCR

Thirteen genes of interest were chosen to include several functional categories, including those involved in adherence, drug resistance, ergosterol synthesis, transport, amino acid metabolism, cell wall synthesis, and iron metabolism. We found that the expression direction for each gene chosen was consistent with the array data (Table 5).

Table 5.

Comparison of RT-PCR and microarray results for genes of interest

Gene	Microarray Fold Change (Biofilm/Planktonic)	RT-PCR Fold Change (Biofilm/Planktonic)
ALS1	11.8	1370
ARG1	3.3	154.3
BGL2	1.5	3.1
CDR2	1.5	2.6
ECM3	1.8	2.3
ERG25	1.8	21.1
FET3	2.1	19.8
HGT2	52.4	1060
INO1	2.9	13.1
KRE1	1.9	32.7
PCK1	2.2	66.3
PHR1	2.4	24.2
SAP5	2.0	29.5

DISCUSSION

Nearly all device-associated infections are associated with organisms growing as biofilms [1, 2, 18]. Most of our knowledge about Candida biofilms is based on in vitro studies, but such systems cannot completely simulate the environment of the infection site and host immune system. The in vivo central venous catheter model was developed to incorporate these experimental variables and has been shown useful for the investigation of Candida biofilm development, diagnosis, drug treatment and resistance, and study of the role of specific individual gene products [11, 13, 19, 20]. The current investigation expands the examination of the molecular basis for biofilm formation by assessing the temporal global transcript profile during C. albicans biofilm development.

These experiments identified differential regulation of many genes previously shown important in biofilm formation in vitro [4, 21-23]. Also, a subset of differentially expressed transcripts in the current in vivo study has been previously identified in genomic investigation of the planktonic Candida response to the mammalian host and specific host cells [12, 24]. Among the similarities to prior biofilm study was the upregulation of genes involved in adherence (ALS1 and ALS4), cell wall metabolism (ECE1, SAP5), carbohydrate and general metabolism (CGT, ICL1, MLS1, PCK1, PDK1), and hyphal formation (HWP2).

Noteworthy, in vivo biofilm formation involved differential regulation of the glyoxylate cycle, which allows organisms to use two carbon molecules as an energy source. Transcripts for two enzymes of the glyoxylate cycle, MSL1 and ICL1, encoding malate synthase and isocitrate lyase, were 2.4 and 2.3-fold more abundant in the mature biofilm. Increased expression of these genes has been described upon inoculation and infection of animal models in vivo and by phagocytosis [24, 25]. Mature biofilms are heterogeneous with cells presumably exposed to gradients of nutrients. Biofilm cells closest to the catheter surface may experience a decreased supply of glucose. Increased expression of the glyoxylate cycle potentially allows the cells to utilize additional carbon sources.

The current analysis identified several expression differences between the 12 and 24h biofilm timepoints. These differences may be accounted for by the stepwise morphologic and architectural changes during biofilm development or perhaps due to a change in metabolic state or a quorum sensing process. Initial biofilm formation requires the attachment of yeast cells to a substratum [26, 27]. As expected, expression of a greater number and intensity of adherence genes was observed at the earlier 12h timepoint. ALS (agglutinin-like sequence) genes are a family of adhesins recognized to play a role in adherence and early biofilm formation [21, 28, 29]. In the current studies, transcript abundance of ALS1 (12-fold) and ALS2 (4-fold) was present only at the earlier timepoint. The current study did not observe a transcript abundance of ALS3. This result supports the hypothesis that adhesins may have overlapping functions in biofilm formation in vivo. Another, possibility is that ALS3 may have been upregulated at a timepoint other than those examined in the current microarray analysis.

Upon initial cell adherence to a solid surface and the instigation of biofilm formation, genes involved in amino acid biosynthesis have been shown to be differentially regulated in vitro [7, 30]. Pathways involved in sulfur metabolism and regulation of methionine and cysteine biosynthesis are among those upregulated as early as 30 min following adhesion. In the current study, increased expression of MET3 (1.6-fold), MET10 (1.7-fold), CYS3 (3-fold), and CYS4 (1.7-fold) was detected in the intermediate biofilm. The specific role of these pathways in early biofilm formation has not yet been elucidated. MET3 is a primary activator of sulfur assimilation in planktonic cells and is generally repressed in the presence of extracellular cysteine and methionine. However, transcript abundance of MET3 has been identified in biofilm systems containing methionine and cysteine concentrations above that normally needed for repression [30]. Although it is possible that amino acids are limited within microenvironments in a heterogeneous biofilm, the swift upregulation of these pathways upon adherence to a solid surface suggests a role in sensing and responding to the surface [30]. In contrast to the intermediate biofilm, changes in expression of genes involved in methionine and cysteine biosynthesis were not predominant in the mature phase of biofilm growth. Instead, abundance of transcripts encoding amino acid permeases (DAO2, DIP5, GAP6, and GNP1) was noted [31, 32].

Mature Candida biofilms are marked by a basal layer of yeast cells with subsequent layering of filamentous morphotypes and extensive matrix production [33]. Transition to the hyphal morphology intuitively requires differential regulation of hyphal-associated genes [20, 34]. The current studies identified expression of a number of hyphal-associated genes in both the intermediate (15 genes) and the mature biofilms (40 genes). Furthermore, transcription of TEC1, a transcription factor involved in hyphal morphogenesis, was upregulated (2.4-fold) in the mature biofilm [35]. Of note, we also identified downregulation (2.5-fold) of TUP1, a negative regulator of hyphal formation [9, 36].

Recent studies have identified quorum sensing molecules in Candida similar to those described in bacterial biofilm systems [37, 38]. Farnesol acts to inhibit formation of biofilm, while tyrosol induces replication and hyphal morphogenesis at low cell densities [39, 40]. CHK1, a histidine kinase gene, appears to be a component of the farnesol quorum sensing pathway because the chk1/chk1 mutant is unresponsive to the action of farnesol [41]. In the current study, the CHK1 transcript was abundant (1.8-fold) in the mature biofilm, but not the intermediate biofilm. This timing is consistent with CHK1 involvement in the regulation of mature biofilms and perhaps not surprising given the much greater burden of organisms in the milieu at this later phase of development. A recent microarray analysis compared the transcriptome of a 24h mature biofilm exposed to farnesol to an unexposed biofilm and identified 274 differentially regulated transcripts [9]. Genes involved in hyphal morphology, drug resistance, and cell wall maintenance were highly represented. Sixty-nine of these farnesol-responsive genes were differentially regulated in the mature biofilm of the current study (Supplemental Table 3) [9].

We also identified altered regulation of ergosterol β-glucan pathways associated during in vivo biofilm growth. Similar changes in gene regulation have been recently described in the most basal layer of in vitro biofilm cells, blastospores [23]. When compared to the current studies, both identified increased transcripts of ERG25 and β-1,6 glucan synthesis genes KRE1 and SKN1 (25). ERG25 is putative C-4 methyl sterol oxidase with a role in C4-demethylation of ergosterol biosynthesis intermediates. It has been proposed that this upregulation may allow for increased conversion of lanosterol to nonergosterol intermediates, including eburicol and 14-methyl fecosterol, at the expense of conversion to ergosterol [23, 42].

Additionally, altered expression of the β-1,3 glucan synthesis and modification pathways (FKS1, BGL2, XOG1, PHR2, FEN1, GDB1, and SGA1) was prominent in the current study [43, 44].. It has been hypothesized that the glucan pathway may restructure the biofilm cell wall and contribute to the drug resistant phenotype [23]. The β-glucan pathway may also play a role in biofilm matrix production. Formation of a mature biofilm requires production of an extracellular polymeric matrix, which is composed primarily of carbohydrate, glucose, and protein with smaller amounts of hexosamine and phosphorus [3, 18]. The presence of glucan molecules in secreted Candida biofilm material and their possible role in biofilm resistance has recently been reported [45]. A number of transcripts involved in carbohydrate processing and synthesis (44) and cell wall metabolism (40) were differentially expressed in the mature biofilm. These gene products may participate in production of matrix material. Interestingly, transcripts involved in the β-1,3 glucan degradation pathway (ENG1 and SWC1) were reduced at the 24h timepoint. Both gene products, ENG1 and SWC1, have β-1,3 glucosidase activity. Disruption of CaENG1 results in decreased β-1,3 glucanase activity 8]. It is possible that downregulation of these β-1,3 glucosidases and altered regulation of the β-1,3 glucan pathway may serve to conserve glucans for construction of a mature biofilm matrix.

Many factors have been proposed to contribute to Candida biofilm drug resistance, including upregulation of efflux pumps, decreased perfusion of antimicrobials through the matrix, slow growth, and alterations in plasma membrane ergosterol content [46, 47]. Early biofilm resistance coincides with increased transcript levels of efflux pumps MDR1 and CDR1 [48]. The current studies identified transcript upregulation of CDR2 at 12h (1.5-fold) and MDR1 at both 12h (2.1-fold) and 24h (1.9-fold). PDR16 transcript abundance was also noted in both the intermediate (2.2-fold) and mature biofilms (3.7-fold). PDR16 encodes a phosphatidylinositol transfer protein of the Sec14p family and is upregulated in fluconazole-resistant cells overexpressing CDR1 and CDR2 [49].

Candida biofilm resistance is likely multifactorial, with contributions from efflux pumps, ergosterol changes, matrix and other, not yet identified, processes. In more mature biofilms, changes in the ergosterol content of the plasma membrane have been hypothesized to contribute to biofilm resistance [4]. The ergosterol content of cells in a mature biofilm is 50% less than that found in an early-phase biofilm. The current studies observed altered regulation of the ergosterol pathway, as noted above.

Three prior studies have examined the global transcriptional response during Candida biofilm development [7, 9, 30]. The report with the most expression similarities to the current study was that which utilized a perfused model with continued nutrient repletion [7]. Forty-nine genes were differentially regulated in both the perfused model and current study (Supplemental Table 3). Similar to the perfused model, the current study found differential regulation of numerous amino acid synthesis genes in both the intermediate biofilms (18 genes) and mature biofilms (21 genes). The similarities of these models may, in part, be due to the unlimited supply of nutrients.

Similarities were also noted in comparison to an in vitro biofilm transcriptional profiling study utilizing a polystyrene petri dish as a substrate in a static condition with regard to nutrient supply [30]. This study focused on the early phase of Candida biofilm formation and might best be compared to the intermediate (12h) biofilm from the current study. Transcripts involved in sulfur metabolism (MET3, MET10) were abundant in both studies (Supplemental Table 3). Also, transcription of NCE103, a carbonic anhydrase-like gene, was upregulated in both biofilm models and may represent response to the metabolic state of the cells growing in a biofilm.

Analysis of the in vivo data set also identified differences from the in vitro transcriptional studies [7, 30]. For example, transcripts of genes encoding glucose transporters were among those most abundant at both timepoints. , Increased expression of both HGT1 (5.4-fold) and HGT2 (52-fold) was observed in intermediate biofilms. Mature biofilm formation was associated with transcriptional abundance of HGT1 (4.7-fold), HGT2 (7.6 fold), HGT14 (2.2-fold), HGT15 (5-fold), and HGT19 (4.5-fold). One possible explanation for this expression is the need for scavenging glucose for nutrient-starved biofilm cells. Another possibility is a glucose requirement for production of the carbohydrate matrix.

A sizable subset of genes was differentially regulated in both the intermediate (8.5%) and mature (16%) biofilms. Genes involved in amino acid metabolism (AGP2, ARG1, CAN1, CDG1, CPA1, CPA2, and SMM1), cell wall metabolism (BGL2 and ECE1), iron metabolism (CFL2, FRE10, ISU1, and SIT1), and drug resistance (ERG25, PDR16 and MDR1) were among those abundant during the time periods examined (Table 3). We suspect these gene products may play a role in the biofilm phenotypes exhibited throughout the biofilm lifestyle, including drug resistance, cell wall changes, altered metabolism, and maintaining cell attachment.

Of note, the study did not identify differential expression of several genes of demonstrated importance in biofilm formation (ADH1, BCR1, CPH1, EGF1, MKC1, and YWP1) [19, 21, 34, 50]. The timing of expression of these genes may not correlate with the timepoints examined in our study. For example, BCR1 and ALS3 may play a role in earlier biofilm formation. Also, the magnitude of the differences in transcript level may be below that detectable by our microarray analysis or regulation may occur at the post-transcriptional level. The absence of these genes in the current data set points to the importance of time course experimentation.

The identification of transcripts unique to the current study demonstrates the value of in vivo models that more closely mimics the disease state. The data set provides a resource for laboratories to guide directed biofilm investigations and target specific gene products.