|Home | About | Journals | Submit | Contact Us | Français|
Recent studies have shown that the transcriptional landscape of the pleiomorphic fungus Candida albicans is highly dependent upon growth conditions. Here using a dual RNA-seq approach we identified 299 C. albicans and 72 Streptococcus gordonii genes that were either up- or down-regulated specifically as a result of co-culturing these human oral cavity microorganisms. Seventy five C. albicans genes involved in responses to chemical stimuli, regulation, homeostasis, protein modification and cell cycle were statistically (P ≤0.05) up-regulated, while 36 genes mainly involved in transport and translation were down-regulated. Up-regulation of filamentation-associated TEC1 and FGR42 genes, and of ALS1 adhesin gene, concurred with previous evidence that the C. albicans yeast to hypha transition is promoted by S. gordonii. Increased expression of genes required for arginine biosynthesis in C. albicans was potentially indicative of a novel oxidative stress response. The transcriptional response of S. gordonii to C. albicans was less dramatic, with only eight S. gordonii genes significantly (P ≤0.05) up-regulated ≥ twofold (glpK, rplO, celB, rplN, rplB, rpsE, ciaR, and gat). The expression patterns suggest that signals from S. gordonii cause a positive filamentation response in C. albicans, while S. gordonii appears to be transcriptionally less influenced by C. albicans.
Candida albicans is a commensal fungus and opportunistic pathogen found in the human gut, oral cavity and genital tract. It is present in 20-60% humans, depending upon the population studied (Martins et al., 1998). C. albicans can progress from commensal colonization to local invasion, according to subject susceptibility, and then to invasive candidiasis which is associated with high mortality rates (Pfaller et al., 2010; Eggimann et al., 2015). In the oral cavity, C. albicans tends to localize to mucosal surfaces and prostheses e.g. dentures (Zomorodian et al., 2011), but there is also evidence for association with dental caries (de Carvalho et al., 2006) and periodontal disease (Canabarro et al., 2013). C. albicans therefore often coexists with other microorganisms in polymicrobial biofilm communities. There is evidence that this influences the morphogenetic, pathogenic and antifungal susceptibility properties of C. albicans, and therefore such infections may be more difficult to control (Wright et al., 2013).
Biofilm formation in C. albicans is a multistage process (Douglas, 2003) and is dependent upon morphological transitions from yeast to pseudohyphal or hyphal forms. The formation of biofilms involves global changes in gene expression (Garcia-Sanchez et al., 2004) modulated by at least six transcription factors (Nobile et al., 2012). A set of eight genes has been found forming a core filamentation response (Martin et al., 2013). The genes include ALS3, ECE1, HGT2, IHD1 and RBT1, all of which encode cell wall proteins (CWPs).
Recently, RNA sequencing has been utilized to provide transcriptome maps of C. albicans under a range of growth conditions, both in vitro and in vivo (Bruno et al., 2010; Tierney et al., 2012; Grumaz et al., 2013). Such studies have revealed that relatively well established pathways may be affected unexpectedly by the prevailing growth conditions. For example, the arginine biosynthesis genes (e.g. ARG1, ARG3, ARG4) are induced under conditions of mild oxidative stress (0.5 mM hydrogen peroxide, H2O2) such as those which might occur within phagocytes (Jiménez-Lopéz et al., 2013).
In mixed species biofilms there are clearly many opportunities for trans-kingdom signalling to occur (Nobbs & Jenkinson, 2015) and it is therefore important to understand the nature of this communication. Quorum sensing molecules and other metabolites from Gram-negative bacteria such as P. aeruginosa (Hogan et al., 2004) and Burkholderia cenocepacia (Boon et al., 2008) have been shown to block hypha formation in C. albicans. Staphylococcus aureus appears to inhibit filamentation under some conditions (Fox et al., 2013), while the oral bacterium Streptococcus mutans inhibits hypha formation by production of trans-2-decenoic acid (Vilchez et al., 2010) and competence stimulating peptide (Jarosz et al., 2009).
Streptococcus gordonii is found associated with most surfaces in the human oral cavity (Aas et al., 2005) and is one of several oral streptococcal species that have been shown to co-aggregate with C. albicans (Jenkinson et al., 1990). It is suggested that these interactions with bacteria are crucial for C. albicans incorporation into oral cavity biofilms (Jenkinson, 2011), and for the development of polymicrobial communities. More recent studies have shown that one response of C. albicans to the presence of S. gordonii involves promotion of hyphal morphogenesis (Dutton et al., 2014). This could be influenced by cell-cell contact and modulated by secreted metabolites i.e. signalling molecules (Bamford et al., 2009; Jack et al., 2015). The aim of the work described in this paper was to outline the transcriptional responses of C. albicans and S. gordonii in the early stages of their interaction, mimicking the natural biofilm communication processes that occur in the oral cavity. The expression profiles obtained reveal how early recognition responses modulate downstream events involved in dual species, trans-kingdom biofilm development.
C. albicans wild-type strain SC5314 was grown aerobically for 16 h in YPD medium (2% yeast extract, 1% mycological peptone, 2% dextrose) at 37°C, with shaking at 220-rpm. Cells were then harvested by centrifugation (5000 × g for 5 min), washed twice in YPT medium (1 × Difco yeast nitrogen base, 20 mM phosphate buffer pH 7.1, 0.1% Bacto tryptone) by alternate centrifugation (5000 × g for 5 min) and suspension, and finally suspended at optical density 600 nm (OD600) = 1.0 (approximately 1 × 107 cells ml−1) in YPT medium. Aliquots (10 ml; 1 × 107 cells ml−1) were transferred into conical flasks containing YPT medium (90 ml) supplemented with 0.4% glucose (YPT-Glc). The cultures were then incubated at 37°C for 2 h with shaking at 50-rpm to induce hypha formation (Dutton et al., 2014). S. gordonii cells were grown anaerobically for 16 h in 10 ml BHY medium (per litre: 37g Brain Heart infusion broth, 5 g yeast extract) and then harvested by centrifugation (5000 × g for 7 min). The bacterial cells were washed twice with YPT (no glucose) and finally suspended at OD600 = 0.5 (2 × 108 cells ml−1) in YPT-Glc medium.
Several combinations of C. albicans, S. gordonii and growth medium were designed to specifically identify changes in gene expression as a result of co-incubation (Table 1). For the dual-species cultures of S. gordonii and C. albicans, S. gordonii cell suspensions (50 ml; 2 × 108 cells ml−1) were added at 2 h, while for the C. albicans monospecies culture, YPT-Glc medium alone (50 ml) was added. S. gordonii cell suspension (50 ml) was added to pre-warmed (37°C) YPT-Glc medium for 1 h for the S. gordonii monoculture. To prepare C. albicans spent medium, C. albicans cells were removed from the culture medium by centrifugation (5000 × g, 5 min), and the supernatant was vacuum filtered through a 0.45 μm nitrocellulose membrane. The filtered medium was transferred to a sterile glass bottle and warmed to 37°C before S. gordonii suspension (50 ml; 2 × 108 cells ml−1) was added. All cultures were incubated at 37°C with shaking (50-rpm) for a further 1 h.
Cells were harvested by centrifugation (5000 × g, 10 min) in 50 ml-Falcon tubes and all but 5 ml supernatant was aspirated. The cell pellet was suspended in the remaining supernatant, transferred to sterile 15 ml-Falcon tubes and harvested by centrifugation (5000 × g, 5 min). The supernatant was aspirated until only 0.5 ml remained, and this was used to suspend the cell pellet. The cell suspension was frozen into small balls by dropping portions (200 μl) into liquid nitrogen. The balls were stored at −70°C prior to RNA extraction.
Frozen microbial cell balls were thawed on ice and suspended in ice-cold RLT buffer (Qiagen Ltd., Manchester, UK) containing 2-mercaptoethanol and transferred to a sterile screw cap microfuge tube containing acid-washed Biospec glass beads (0.6 ml). The suspension was mixed with the glass beads and the fungal and bacterial cells were disrupted by alternating shaking (30 s) using a Fast-prep 25 bead beater (MP Biomedicals, Santa Ana, CA) and incubating 1 min on ice (repeated 3 times). The beads were allowed to settle and the supernatant was transferred to a sterile microfuge tube. The disrupted cells were centrifuged (13000 × g, 2 min) and the supernatant transferred to a sterile microfuge tube. An equal volume of 70% ethanol was added and the RNA was extracted and purified using an RNeasy Mini Kit (Qiagen) with the use of an on-column DNAse digestion (Qiagen). The quality of the RNA was checked by formaldehyde agarose-gel electrophoresis. The RNA concentration of each sample was measured spectrophotometrically (Nanodrop 1000, Thermo Scientific, Fisher Scientific UK Ltd, Loughborough, Leics., UK) and stored at −20°C.
ERCC RNA Spike-In Control Mix (Ambion, Foster City, CA) was mixed with 2.5 μg RNA. Ribosomal RNA was depleted with a RiboZero Magnetic Gold Kit (Epicentre) and lllumina sequencing libraries were prepared using ScriptSeq v2 (Epicentre, Illumina Inc., Madison, WI) with 10 cycles of PCR amplification. The quality and quantity of each library was determined using a Bioanalyzer and the average (modal) insert size of samples was 400 bp with the spread ranging between 200 bp and 1,000 bp. An equimolar library pool was denatured, diluted to 6.5 pM and clustered on a cBot (Illumina) to create clonal clusters from single molecule DNA templates. One hundred base pair paired-end sequencing was undertaken using HiSeq2500 (Illumina) in high output mode with Truseq v3 reagents. The resulting FASTQ data were then filtered using the fastq-mcf command from the EA-Utils suite to remove adaptor sequences and low quality bases (Aronesty, 2011). The filtered data were then aligned against the reference ERCC transcripts using Bowtie v1.0.0 using the -X 600 flag. SAMtools v0.1.19 was utilized to convert the resulting SAM formatted-file to BAM (Li et al., 2009). The number of reads mapping to each transcript was extracted using the SAMtools idxstats command and used to calculate RPKM values for each ERCC transcript. Log2 values of observed RPKM were then plotted against log2 expected RPKM values and inspected to establish lower limits of detection.
The reads which did not map to the ERCC transcripts were then aligned to a merged FASTA file containing both the Ca21_C_albicans_SC5314 genome (Version 21 from www.candidagenome.org) and the NC_009785 S. gordonii CH1 genome. The Tophat2 v2 2.0.8b program was used with the following parameters: -G -library-type fr-secondstrand -I 10000 -r 50, --mate-std-dev 100 -p 8. The -G parameter was followed by the combined gff file containing annotation for both organisms (Kim et al., 2013). An additional analysis was carried out with the DESeq analysis tool to calculate differential gene expression (Anders & Huber, 2010). The output of Tophat2 was processed using the Bedtools multicov command (Dale et al., 2011). This produces numbers of reads mapping to each annotated gff feature. The gene/read count files were then processed to separate the eukaryote and prokaryote gene features. Each was then analyzed separately using the DESeq package v1.12.1. Default parameters were utilized as outlined in the DESeq manual. DESeq uses a negative binomial model to account for the dispersion of the reads and the variation between replicates (n = 3), and uses a general linear model for comparisons (Love et al., 2014). P-values were calculated from DESeq, and Benjamini-Hochberg adjusted P-values ≤ 0.05 were deemed significant. All transcriptional data have been submitted to the GEO repository and assigned the GEO accession number GSE68477.
Confocal scanning laser microscopy of C. albicans (stained with Calcofluor white) with or without S. gordonii (stained with fluorescein isothiocyanate, FITC) was performed as previously described (Dutton et al., 2014). Transmission electron microscopy (TEM) was performed as follows: cell suspensions were centrifuged (5000 × g, 10 min) and to the pellets was added TEM fixative (4% paraformaldehye, 5% glutaraldehyde, 0. 1M sodium cacodylate buffer pH 7.2, 0.05% Tween 20). Tubes were shaken gently and incubated for 15 min at 22°C. The samples were centrifuged (3000 × g, 2 min), the supernatant carefully removed, and the pellet suspended in TEM fixative (2% paraformaldehye, 2.5% glutaraldehyde; 2 ml) for 16 h at 4°C. The pellets were then washed 3 times in 0.1 M sodium cacodylate buffer (pH 7.2), and set into low melting point 2% agarose. The agarose pellets were cut into ~3 mm sections and incubated for 1 h at 22°C in 1% osmium tetroxide solution. The samples were rinsed in sterile H2O, dehydrated at room temperature using sequential incubations in ethanol (30%, 50%, 75%, 90% and 100%) then propylene oxide, and embedded in Spurr resin. The resin was cut by microtome (RMC Powertome PC) into sections of 80-100 nm using a diamond knife, and sections were collected on copper grids and imaged at 80 kV by TEM.
Scanning Probe Microscopy (SPM) was performed on C. albicans-S. gordonii cultures deposited onto glass cover slips. Cover slips were attached to metal pucks with double sided tape and mounted on a Multimode AFM fitted with a Nanoscope IIIa controller (Veeco, Santa Barbara, CA). Cells were imaged in contact mode using triangular silicon nitride (Si3N4) tips with a nominal spring constant of ~0.06 N m−1 (Veeco). Images were obtained at typical scan rates of 25 μm s−1 and processed using Nanoscope 8 software (Veeco).
Coaggregation of C. albicans and S. gordonii cells occurs rapidly after mixing of the two cell types (Jenkinson et al., 1990). In order to visualize these interactive events, S. gordonii cells, fluorescently labelled with FITC, were mixed with hyphae-forming C. albicans cells in YPT-Glc medium (see Methods) and incubated for 1 h at 37°C. The associations between the bacteria and fungi were then viewed by confocal scanning laser microscopy (CSLM). After 3 h growth in monoculture, ~50% C. albicans cells had formed hyphae of 20-50 μm in length (Fig. 1A). In the presence of S. gordonii, bacterial cells bound along the lengths of the hyphal filaments (Fig. 1B) and could also be seen to aggregate, forming microcolonies at discrete attachment sites (Fig. 1B). These aggregates must form mainly by recruitment of other streptococci since extensive bacterial cell division does not occur in YPT-Glc medium within the experimental time-frame of 1 h.
One characteristic of C. albicans hypha formation is that the hyphal filaments clump together. As expected therefore, transmission electron micrographs showed C. albicans hyphal filaments in close contact with each other (Fig. 2A), seemingly connected by networks of fibrillar material. S. gordonii cells were often seen intimately interacting with hyphal filaments, with the cell surfaces in direct contact along a length of about 100 nm (Fig. 2B). There were also streptococci present that were not interacting with hyphae (Fig. 2B). The closeness of the bacterium-fungus interaction was further shown by scanning probe microscopy. In scans of wet mounts, streptococcal and C. albicans cell surfaces are apparently coalesced (Fig. 3A), while under dried conditions there appears to be a structural difference on the hyphal cell surface at the point of contact with bacteria (Fig. 3B) which could suggest some form of fungal cell wall remodelling. It is these close associations between the two cell types that led us to hypothesize that recognition signals (contact or diffusible) could be relayed through cell wall sensors to modulate gene expression in response to the other microorganism.
Following suspension of C. albicans cells in YPT-Glc medium, and incubation for 2 h at 37°C, fungal cells were undergoing early-stage hyphal morphogenesis (see above). At this point the C. albicans cells were allowed to proceed for a further 1 h, either in the presence or absence of S. gordonii, or in the presence of spent C. albicans culture medium (to correct for metabolic effects), and RNA was then prepared. Transcriptomic data therefore represent the response of C. albicans to the presence of S. gordonii for 1 h, and they take into account also the transcriptional effects of nutritional shift-down and C. albicans culture medium on S. gordonii cells (see Table 1). Under these conditions, physical trans-kingdom interactions occur, and chemical signals would carry on being exchanged between the two microorganisms since both C. albicans and S. gordonii continue to metabolize within this medium (Bamford et al., 2009).
Illumina sequencing from the combined C. albicans and S. gordonii samples yielded 258,347,928 raw reads (Table S1). A total of 6,653 open reading frames (features) from all the samples were functionally annotated to the haploid assembly 21 of the C. albicans genome (van het Hoog et al., 2007) using the Candida Genome Database (CGD) (www.candidagenome.org), while 2,051 open reading frames were matched to the S. gordonii database for annotation, visualization and integrated discovery (DAVID version 6.7 - http://david.abcc.ncifcrf.gov) (Huang et al., 2007; Sherman et al., 2007). The total numbers of candidal and streptococcal reads for all three sample replicates were calculated, and the % distribution of candidal and bacterial reads were calculated for the three combined replicates for each sample (Table S2).
When C. albicans and S. gordonii were co-cultured in YPT-Glc medium, the overall ratio of reads was 37.7% C. albicans to 62.7% S. gordonii. This showed there was a suitable distribution of C. albicans and S. gordonii reads with no undue bias towards one single organism. It should be noted that features can include any sequence belonging to the genome, including reads that are not translated into proteins. This explains why there were a small number of reads (0.09% and 0.03%) from the S. gordonii samples that matched to the C. albicans genome. These relate to regions of the genome with some homology that are similar in the two organisms e.g. tRNA or mitochondrial RNA.
One technical consideration was if there might be a bias towards long or short transcripts. To check this, the gene (orf) coordinates from the 6,653 orfs in the CGD were used to prepare a dataset of the lengths of every orf. The normalized expression reads (tag counts) for every gene were then plotted against gene length for the entire C. albicans genome. The plot was compared with a corresponding graph of normalized expression reads versus gene length for the up- and down-regulated C. albicans genes in the presence of S. gordonii. These data, presented in Fig.S1, show that there was no shift in length distribution between total C. albicans orfs and differentially regulated genes, and no obvious change in range of distribution. Therefore it was concluded that the expression data were not biased by gene length.
The Illumina sequencing data from co-incubated cultures of C. albicans and S. gordonii were compared to the data obtained from C. albicans grown with only the addition of growth medium (YPT-Glc) minus S. gordonii. This was to rule out the likelihood that any changes in gene expression were caused by the addition of the extra growth medium after 2 h rather than an effect caused by S. gordonii cells. Volcano plots of P-value vs. mean fold change in gene expression (Fig. 4) derived from analysis of the Illumina data (by the statistical DESeq package v1.12.1 with default parameters) showed that when C. albicans and S. gordonii were co-cultured in YPT-Glc medium the expression levels of a large number of C. albicans genes were significantly (P ≤0.05) up- or down-regulated by at least a twofold change (Fig. 4A, Table 2). On the other hand, only one S. gordonii gene was significantly (P ≤0.05) decreased in expression when the bacteria were incubated with C. albicans. S. gordonii gene expression significantly increased only slightly overall, with the majority of increases around twofold or less (Fig. 4B). Statistical analysis using DESeq v1.12.1 also showed the total number of genes with altered expression (either up- or down-regulated) was much higher in C. albicans (299 genes) compared with S. gordonii (72 genes).
Table 2 shows all genes identified as significantly (P ≤0.05) up- or down-regulated by DESeq in C. albicans when hypha-forming cells were incubated with S. gordonii for 1 h. Eighteen out of 75 genes up-regulated ≥ twofold are annotated as being associated with stress responses (core, oxidative, acid, macrophage up-regulated), while 15 genes are annotated as being up-regulated in biofilm formation (Table 2). Only one up-regulated gene (TSA1) was associated with both stress and biofilm formation, suggesting that at least two response pathways were being activated. These results, together with the volcano plots, indicated that when C. albicans and S. gordonii were co-incubated there was a much larger overall effect on C. albicans compared to a relatively small effect on S. gordonii.
Within the Gene Ontology (GO) category for Biological Processes, from 152 C. albicans up-regulated genes (Table 2), 75 were significant (P ≤0.05), while 36 of the 147 down-regulated C. albicans genes were found to be significantly (P ≤0.05) down-regulated. The genes with significant differential gene expression were assigned to 41 GO Slim categories (Fig. 5). The numbers of genes assigned to biological process, other and chemical stimulus (the most abundant) were all found to be significantly (P ≤0.05) affected (Fig. 5). From the C. albicans down-regulated genes the terms biological process and transport were the most abundant (25%) and the terms other and translation were the next highest in abundance (13.9%). The genes assigned to transport were found to be significantly (P ≤0.05) affected (Fig. 5).
Cellular Components describe locations, at the levels of subcellular and macromolecular complexes. For GO category Cellular Components, 75 up-regulated C. albicans genes were assigned to 27 Slim categories (Fig. 6). From the up-regulated C. albicans genes the term cytoplasm (48.6%) was the most abundant with the terms cellular component (37.8%) and mitochondrion (28.4%) showing the next highest abundance. Gene ontology terms associated with the nucleus, mitochondrion, and cell wall were all found to be significantly (P ≤0.05) enriched in the up-regulated gene sets (Fig. 6). For the down-regulated C. albicans genes, GO Slim terms assigned to cytoplasm and cellular components were found to be significantly (P ≤0.05) affected (Fig. 6).
Molecular Functions describe activities, such as catalytic or binding activities that occur at the molecular level. For the GO category Molecular Function, the 75 significantly up-regulated C. albicans genes (from a total of 152 up-regulated genes) (Fig. 7) and 36 significantly down-regulated C. albicans genes were assigned to 26 GO Slim categories. From the up-regulated C. albicans genes, the term molecular function (36.5%) was the most abundant, with the terms oxidoreductase activity (25.7%) and transferase activity (14.9%) showing the next highest abundance. The number of genes assigned to oxidoreductase activity and hydrolase activity were found to be significantly (P ≤0.05) affected (Fig. 7). From the C. albicans down-regulated genes the term molecular function (27.8%) was the most abundant, with oxidoreductase activity (16.0%) and transporter activity (16.7%) significantly (P≤0.05) affected.
From 211 genes associated with pathogenesis by CGD, only 10 genes were shown to be up- or down-regulated by ≥ twofold (Table 3). The genes ALS1, CAT1 and TEC1 are strongly associated with the transition from yeast to hyphae during filamentous growth as well as pathogenesis. The up-regulation of TEC1 in this study tends to suggest that S. gordonii stimulation of C. albicans hyphal growth might occur through the Cyr1/cAMP signalling pathway (Sudbery, 2011).
For successful colonization and pathogenesis, once C. albicans blastospores adhere to host surfaces they rapidly begin to produce hyphal filaments. Hyphae physically penetrate host endothelial and epithelial cells. Aided by a large array of hyphal surface proteins they also create stronger attachments to the host surfaces. Overall there were 18 genes associated with filamentous growth whose expression was affected (six significantly) when C. albicans was challenged with S. gordonii (Table 3). FRG42, ALS1, CAT1, and TEC1 each showed a mean > four-fold increase. ALS1 and TSA1 transcripts were of highest abundance (Table 3). Tec1 and Fgr42 proteins are both filamentous growth regulators (Sudbery, 2011), so are involved in the switch from yeast form to hyphal form, suggesting that the gene expression network associated with the yeast to hypha transition was stimulated by the presence of S. gordonii in the culture. This is supported by evidence for growth synergy of streptococci and C. albicans in biofilms (Bamford et al., 2009) and morphological observations of more extensive hyphal filament formation with S. gordonii present (Dutton et al., 2014).
Up-regulation of CAT1 (catalase) during the yeast-hypha transition might be consistent with reports that reactive oxygen species (ROS) levels are increased under hypha-forming conditions (Schroter et al., 2000). During the transition of yeast to hyphal filament, the cell wall must go through a considerable conformational change, involving restructuring, which ultimately requires a greater energy input from the cell. The energy comes from oxidative phosphorylation in mitochondria, which generate ROS as a by-product. The ROS will attack and inhibit the functions of proteins, lipids and DNA if not decomposed by Cat1 and other anti-oxidant enzymes. So an increased energy output during hypha formation will generate a need for an increase in anti-oxidants.
HSP21, one of the up-regulated genes in Table 3, encodes a multifunctional heat shock protein, with a role in both stress adaptation and virulence in C. albicans (Mayer et al., 2012). Hsp21 modulates thermal stress by fine tuning homeostasis of compatible solutes and activation of the Cek1 pathway, which has previously been shown to be influenced by S. gordonii (Bamford et al., 2009). Hsp21 also mediates adaptation to oxidative stress, while an hsp21Δ/Δ mutant forms shorter hyphae than the wild-type and is strongly attenuated in virulence in vivo (Mayer et al., 2012).
Numerous C. albicans anti-oxidant genes in addition to CAT1 responded to the presence of S. gordonii including ORF19.3537 a putative sulfiredoxin, two oxidoreductases (ORF19.2262 and CIP1), cytochrome c peroxidase (CCP1), glutathione reductase (GLR1), glutathione-S-transferase (GTT11), glutathione peroxidase (GPX2), thioredoxin reductase (TRR1) and the thiol-specific antioxidant protein (TSA1). TSA1, TRR1, CAT1, GLR1 and CPA2 transcripts were of high abundance (Table 2). However, not all known oxidative stress response genes were up-regulated e.g. TRX1, SOD1, GRX2, GPX31-33 (da Silva Dantas et al., 2015) suggesting a more specific type of response to S. gordonii. Many of the antioxidant genes are reported to be regulated by Cap1 (Wang et al., 2006) and their products are up-regulated upon H2O2 treatment of C. albicans (Kusch et al., 2007). Overall, protection against H2O2 potentially confers added resistance to macrophages and cells of the innate immune system that utilize oxidative stress to kill C. albicans.
ARG1, ARG3, ARG4, ARG5,6, CPA1, and CPA2 are all involved with arginine biosynthesis are also greatly up-regulated in C. albicans with S. gordonii present. Conversely CAR1, an arginase, is strongly down-regulated suggesting that increased arginine production is important for the response of C. albicans to S. gordonii. Since the ARG genes are up-regulated in the presence of H2O2 (Jiménez-Lopéz et al., 2013) these effects could be in direct response to the production of H2O2 by S. gordonii (Liu et al., 2011). The importance of arginine biosynthesis is further strengthened by the up-regulation of PUT1, a putative proline oxidase that is expected to convert L-proline into Δ1-pyrroline-5-carboxylate and glutamate-γ-semialdehyde that act as the main precursors for arginine production and polyamine production. In Saccharomyces cerevisiae, exposure to H2O2 and freeze-thaw stress also leads to an accumulation of arginine (Almeida et al., 2007; Momose et al., 2010), with supplementary arginine conferring resistance to oxidative and thermal stress (Nishimura et al., 2010).
The C. albicans cell wall consists of an internal scaffold of (β-1,3)- and (β-1,6)-linked glucan and chitin, to which an outer protein coat is attached (Klis et al., 2001; Ruiz-Herrera et al., 2006). The protein coat is thought to contain about 20 different polypeptides attached to the cell wall by covalent bonds linking proteins to the inner glucan/chitin skeleton (Klis et al., 2009). These cell wall proteins (CWPs) have been associated with many functions including host cell adhesion and invasion, biofilm formation, cell-cell and intergeneric aggregation, and enzymatic functions such as superoxide dismutases and yapsin-like aspartic proteases (Monod et al., 1998; Martchenko et al., 2004; Krysan et al., 2005). Because of the diversity of their roles, it is not surprising that expression of CWP-encoding genes can vary enormously, not only with mode of growth, but also with environmental signals and input from distinct signalling pathways, triggered by changing environmental conditions (e.g. temperature, pH, N-acetyl-D-glucosamine etc.). This analysis investigated 36 experimentally-validated covalently-linked CWPs of C. albicans (Klis et al., 2009) and how their gene expression profiles changed when C. albicans cells were co-incubated with S. gordonii (Table S3).
Thirteen out of 18 C. albicans CWP genes showed ≥ twofold up-regulation of gene expression, while five were ≥ twofold down-regulated (Table 4). Chi-squared tests showed that cell wall protein genes PGA57, ALS1, PGA34, PHA36 (IHD1), PGA61, and ORF19.4653 were all significantly up-regulated (P ≤0.05). PGA10 was up-regulated > four-fold, though this was not considered significant at P ≤0.05. Pga10 (Rbt51) belongs to a sub-set of fungal proteins with an eight cysteine residues domain CFEM (Common in several Fungal Extracellular Membrane proteins). Pga10/Rbt51 along with other proteins Rbt5 and Wap1/Csa1, each contain CFEM domains, which play key roles during biofilm formation (Perez et al., 2006).
The up-regulation of expression of HYR1 is most likely to be associated with an increase in the extent or rate of hypha formation. Studies have shown HYR1 is induced specifically in response to hyphal development when morphogenesis is stimulated by growth conditions such as serum, temperature elevation, pH and the addition of N-acetyl-D-glucosamine (Bailey et al., 1996). The findings suggest that the just over twofold increase in expression of HYR1 could be associated with yeast to hypha transition, as shown by Spiering et al. (2010), stimulated by the addition of S. gordonii to the culture.
Glycophosphatidylinositol (GPI)-modified proteins all share conserved features of an N-terminal signal sequence and C-terminus tethered to the cell wall or cell membrane by a preformed GPI anchor. In silico predictions suggest that there are 115 genes encoding GPI-modified proteins in C. albicans (Richard & Plaine, 2007). Several lists of C. albicans GPI-modified proteins (GpiPs) published in the literature (De Groot et al., 2003; Garcera et al., 2003; Eisenhaber et al., 2004) have been amalgamated and refined (Richard & Plaine, 2007) to avoid duplications and allow for differences in the algorithms used to define GpiPs.
The functions of the majority of these GpiPs (66%) still remain unknown. The others can be assigned to functions related to cell wall biosynthesis or remodelling, and cell-cell adhesion and interactions. Since the compilation of this list of GpiPs, 45 knock-out mutants of computer predicted GpiPs have been screened for their roles in cell wall structure (Plaine et al., 2008). Deletion mutants that result in cell wall modifications and reduced caspofungin sensitivity included DFG5, PHR1, PGA4 and PGA62.
In our data, from the 115 predicted GpiPs, only 16 had a change in expression > twofold in the presence of S. gordonii (Table S4). A small number of genes (five) were down-regulated while 11 genes showed up-regulation (Table S4), six of which were significant (P <0.05). These were ALS1, ORF19.4653, PGA34, PGA36, PGA57 and PGA61 (Table 4). The CGD provides information on PGA34 (role in host infection) and PGA36 (induced during hyphal development), while ORF19.4653, PGA57 and PGA61 are currently uncharacterized but clearly respond transcriptionally to S. gordonii.
C. albicans cells within the oral cavity must avoid being washed away by the continuous flow of saliva, therefore adhesion to a multitude of host surfaces including epithelial cells, teeth, and dentures is of paramount importance. C. albicans possesses numerous potential adhesins (Zordan and Cormack, 2012) and this part of the study investigated 23 proteins which have been linked to cell adhesion in the CGD.
When C. albicans was co-incubated with S. gordonii, 19 out of the 23 C. albicans cell wall protein genes associated with adhesion showed up-regulation of gene expression, while four were seen to be down-regulated (these were all less than twofold changes). Both ALS1 and TEC1 showed the highest mean fold change in up-regulated adhesion-associated genes with significant four-fold changes (P ≤0.05) (Table 4). ALS1 expression is known to be associated with hypha formation and, on the emergence of the germ tube, Als1 is the first member of the Als family to be expressed. Interestingly, the expressed protein is localized at the neck of the growing hypha (Fu et al., 2002). Differential expression patterns of the ALS genes have shown that there is a major spike in ALS1 expression when cells are inoculated into fresh medium, a procedure which also triggers hyphal growth, indicating that Als1 might have a regulatory role as well as an adhesin function. The up-regulation of TEC1, a TEA/ATTS transcription factor, is consistent with its reported role in regulating hypha formation and virulence (Schweizer et al., 2000).
EAP1 (2.6-fold increase), HIS4 (three-fold increase) and HYR1 (2.1-fold increase) were also seen to be up-regulated. Eap1 is a glucan-cross-linked cell wall-localized protein that has been reported to be required for C. albicans to form robust biofilms on polystyrene surfaces (Nobbs et al., 2010) and in central venous catheters (Li et al., 2007) under shear flow in vitro and in vivo. Although expressed in both yeast and hyphal cells, it has been suggested that Eap1 protein expression is not directly associated with hypha formation (Li and Palecek, 2003). However, there are contradicting reports where transcriptional profiling studies of the yeast to hypha transition reveal a twofold increase in Eap1 at 6 h (Nantel et al., 2002) similar to the approximately three-fold up-regulation seen here. Interestingly, studies of S. cerevisiae strains expressing Eap1 have confirmed that Eap1 is able to bind S. gordonii in planktonic culture (Grubb et al., 2009). This indicates that Eap1, along with a number of other adhesins such as Als3 and Hwp1, promotes trans-kingdom interactions with other microorganisms to aid successful colonization and the formation of polymicrobial communities (Nobbs et al., 2010; Xu et al., 2014). Although EAP1 expression was up-regulated, expression of the genes encoding other C. albicans adhesins ALS3 and HWP1 was not over the time-course of the experiment.
A total of 72 S. gordonii genes were either up- or down-regulated specifically in response to co-culturing with C. albicans (Table 5). Eighteen S. gordonii genes were significantly up-regulated including ≥ twofold increases in glpK, rplO, celB, rplN, rplB, rpsE, ciaR, and gat (Table 5). Glycerol kinase (GlpK) allows glycerol to be utilized as a carbon source. Notably, a highly up-regulated gene in C. albicans biofilms was RHR2 encoding the glycerol biosynthetic enzyme glycerol-3-phosphatase. Glycerol is five times more abundant in C. albicans biofilm cells (Desai et al., 2013) so it seems possible that S. gordonii may respond to glycerol production or secretion by C. albicans hypha-forming cells by utilizing this as an alternate carbon and energy source.
CiaR is a response regulator of competence for DNA uptake (Mascher et al., 2003) and biofilm formation (Blanchette-Cain et al., 2013) in Streptococcus pneumoniae. We have shown recently that S. gordonii competence-development is involved in formation of dual species biofilms with C. albicans (Jack et al., 2015), and so CiaR may be a factor in this process. cel genes annotated as encoding components of cellobiose metabolism were up-regulated, perhaps indicating that S. gordonii is responding to the presence of C. albicans cell wall glucans. Up-regulation of rplO, rplN, rplB, and rpsE, all involved in translation, may be an artefact of ribosomal depletion. Only glgP-2 (maltodextran phosphorylase) was significantly down-regulated (Table 5). Regions of the S. gordonii genome containing ORFs SGO_1105 to SGO_1108 showed coordinated down-regulation, and are all implicated in pyrimidine biosynthesis. Overall, these results suggest that S. gordonii was transcriptionally much less reactive to the presence of C. albicans for 1 h in mixed culture, while the effects of CiaR up-regulation in S. gordonii are a topic of future investigation.
This work describes the responses of C. albicans and S. gordonii to each other at the transcriptional level. The experiments were undertaken under growth conditions that induce filamentation in C. albicans. The results presented suggest that S. gordonii has a range of significant effects on the biological processes occurring within C. albicans during the early phase of co-culture, with a large number of genes affected in C. albicans in comparison to a smaller number in S. gordonii. Genes involved in responses to chemical stimuli, regulation, homeostasis, protein modification and cell cycle were up-regulated, while genes involved in transport and translation were down-regulated. These patterns suggest that C. albicans was responding positively to signals produced by S. gordonii. Mitochondrial genes were up-regulated together with genes encoding cell wall proteins, suggesting triggering of metabolic functions.
On the other hand, down-regulation of triplet codon-amino acid adaptor and transporter activities suggest modulation of the rate of protein synthesis. Oxidoreductase activities were up- and down-regulated, while hydrolase activity genes were up-regulated, suggesting perhaps that new macromolecular substrates e.g. polysaccharide, peptidoglycan etc. were now available for metabolism by C. albicans. Overall the data are consistent with C. albicans not being growth-inhibited in the presence of S. gordonii, unlike the inhibitory effects of some other bacteria (e.g. P. aeruginosa) on growth and hyphal development (Hogan & Kolter, 2004; Fox et al., 2013). There was some evidence of up-regulation of genes that might be linked with growth-stimulatory effects, supporting previous observations from biofilm experiments (Bamford et al., 2009; Dutton et al., 2014; Xu et al., 2014).
The major changes in expression of morphogenesis-related genes in response to S. gordonii were upregulation of TEC1, ALS1, and CAT1. Notably, Tec1 is a hyphal-development activator (Nobile & Mitchell, 2005) that regulates expression of cell wall protein genes, but probably not ALS1 (Nobile & Mitchell, 2006). ALS1 also appeared in the up-regulated genes encoding covalently-linked CWPs and GPI-modified proteins. Clearly TEC1, ALS1 and anti-oxidant genes seem to be major players in the response of C. albicans to S. gordonii. The Als1 protein is not thought to be a component of the hyphal cell wall, but appears at the initial site of hyphal filament growth from the mother cell (Coleman et al., 2012). ALS1 has been shown to be one of the genes that is first up-regulated following adhesion to a surface (Garcia-Sanchez et al., 2004) that then leads onto biofilm formation. Further work is required to establish the factors affecting expression of ALS1 and if it is involved in regulation as well as adhesion. Of the GPI-modified proteins, six genes (including ALS1) encoding these were significantly (P ≤0.05) up-regulated in the presence of S. gordonii. The functions of the other five PGA (Protein with Glycosylphosphatidylinositol Anchor) genes are unknown. Further studies could help to identify the functions of the proteins encoded by these genes and they may reveal new factors for adhesion, biofilm formation and pathogenesis. In summary, these transcriptional data findings seem to correlate with the biological data indicating that S. gordonii promotes hyphal development and grows synergistically with C. albicans. They also identify a number of new target genes for further study of their roles in development of interkingdom biofilm communities.
We thank Massimo Micaroni for help with electron microscopy and Jane Brittan for technical assistance. The support of NIH (NIDCR), Bethesda, USA to HFJ and RJL is gratefully acknowledged (R01DE016690).