The laboratory based C. albicans SC5314 was used in this study (O'Donnell et al., 2016). Yeast cells were cultivated as working stocks on fresh Sabouraud agar (Sigma-Aldrich, UK) for 48 h at 30°C and maintained at 4°C. One unique colony was used to grow the cells in yeast-peptone-dextrose (YPD) medium (Sigma-Aldrich) for 18 h at 30°C and 150 rpm orbital shaker. The cells were washed twice by centrifuging in sterile phosphate buffered saline (PBS; Sigma-Aldrich, UK) and standardized using a Neubauer haemocytometer.

Planktonic and sessile cells were first investigated for their susceptibility to the polyphenol CUR (HPLC grade, Acros Organics, Belgium). Stock CUR was prepared immediately preceding the experiment using an non-antimicrobial concentration of dimethyl sulfoxide (DMSO) as a solvent and adjusted to < 5% v/v in RPMI-1640 medium (Sigma-Aldrich, UK; Shahzad et al., 2015). Standardized CLSI M-27A broth microdilution methodology was initially undertaken for planktonic yeast cultures in 96 well round bottomed microtitre plates (CLSI-M27-A, 2008). Clear wells with no visible growth were considered as the minimum inhibitory concentration (MIC). For sessile susceptibility testing, pre-formed 24 h biofilms were challenged with CUR using standardized sessile antifungal testing (Ramage et al., 2001; Pierce et al., 2008). Reduction of tetrazolium to formazan through an XTT assay was used, and the optical densities quantified at 492 nm using a microtitre plate reader (FluoStar Omega, BMG Labtech, UK). Negative and positive controls were included, and the experimental wells data were compared to the positive control data to reveal the SMIC₈₀, where the optical density is reduced more than 80% in comparison to the positive control optical density, reflecting significant bioactivity against the biofilm. These procedures were repeated in three independent occasions where three replicates have been considered.

The potential capability of CUR to be adsorbed to denture material was investigated. Heat cure poly methyl methacrylate (PMMA) denture base material (Chaperlin and Jacobs Ltd, Surrey, UK) was used to fabricate 12 mm diameter discs using the dental compression molding technique. These discs were immersed in ddH₂O for 7 days to ensure excess toxic monomers were removed. CUR was diluted in RPMI-1640 medium to 200, 400, and 800 μg/ml. Discs were distributed in 24 wells plates (Costar, Corning Incorporated, USA) and 1 ml of the CUR suspension was added. The plates were incubated at room temperature for a series of time points (1, 5, 10, 30, 60, 120, 240, and 1,440 min). Next, discs were transferred to fresh wells and washed with 1 ml of distilled water and then 1 ml of DMSO was added to dissolve the adsorbed CUR. To quantify the released CUR, a standard curve (0.39–100 μg/ml serially double diluted) was developed and measured at 436 nm using the spectrophotometer. Based on these data, PMMA discs were immersed in 1 ml of 800 μg/ml of CUR for 10 min (equivalent to 50 μg/ml) then washed with PBS to remove the unabsorbed molecules. These treated discs were inoculated with 1 ml of 5 × 10⁵ CFU of C. albicans SC5314 cells and incubated for 30 min at 37°C. Following the initial adhesion, cells were washed in PBS and the adherent cells removed by sonication at 35 kHz for 10 min (Ultrasonic bath, Fisher scientific, UK), and enumerated using the Miles and Misra plate counting method (Miles et al., 1938). The final cell number was expressed per cm² PMMA, which was compared to a CUR negative control. All experiments were performed with three independent sections on three independent occasions. Scanning electron microscopy (SEM) was also performed using the same experimental parameters, then processed and imaged, as described previously by our group. Briefly, biofilms were grown on Thermanox™ coverslips or hydrogel cellulose matrix and treated, as previously described. Biofilms were washed twice with PBS, before being fixed in 2% para-formaldehyde, 2% glutaraldehyde, 0.15M sodium cacodylate, and 0.15% w/v alcian blue, at pH 7.4, and prepared for SEM as previously described (Erlandsen et al., 2004). The specimens were sputter-coated with gold and viewed under a JEOL JSM-6400 scanning electron microscope.

Discs were adsorbed with CUR as described above, then the discs (untreated and treated) were inoculated with either 1 ml of 5 × 10⁵ CFU + 3 min CUR (50 μg/ml) or PBS (negative control) treated C. albicans SC5314 cells, and incubated for 30 min at 37°C. Adhesion of C. albicans was then assessed and quantified as described in the previous section, with the levels of adhesion expressed as a proportion of the negative control (PMMA-/C. albicans-). In parallel we then assessed whether longer CUR exposure time negatively impacted adhesion by treating cells for 3, 30, and 90 min, and the levels of adhesion to PMMA quantified. Finally, we assessed whether or not CUR exposure to C. albicans cells at different growth phases played a role, with the hypothesis that there may be differences in how yeast (Y), germlings (G), or hyphae (H) responded to this molecule. Briefly, cells were grown overnight, standardized to 1 × 10⁶ CFU in RPMI and inoculated into a 96-well microtitre plate. Cells were then exposed to 50, 100, or 200 μg/ml CUR at either 0 h (Y), 2 h (G), or 4 h (H) post-inoculation and incubated for a further 24 h at 37°C. Thereafter, the cells were washed in PBS and the resultant biofilm quantified using an XTT metabolic reduction assay. All experiments were performed in duplicate on three independent occasions.

In order to assess whether CUR has additional effects on the physicality of these yeasts, we assessed its impact on aggregation. C. albicans SC5314 cells (Y) were standardized (1 × 10⁶ cells) in PBS and exposed ± to subinhibitory concentrations of CUR (50 μg/ml) for 90 min at 37°C under constant agitation (200 rpm). Following incubation the cells were serially 10 diluted in PBS diluent and plated onto Sabouraud agar using the Miles and Misra methodology. The plates were then incubated overnight at 37°C and the colonies enumerated. In parallel, cells were examined under a light microscope to evaluate aggregation visually. All experiments were performed in triplicate on three independent occasions.

Preparation of Y and H cells was performed using an initial inoculum of 1 × 10⁸ cells and 5 × 10⁵ cells of C. albicans SC5314 in RPMI, respectively. For H cells these were incubated on PMMA sections within 24 wells plates for 4 h. Both Y and H cells were then treated ± CUR (50 μg/ml) in RPMI for 3, 30, and 90 min, after which they were prepared for RNA extraction. Cells were either centrifuged or sonicated in a 35 kHz for 10 min (Ultrasonic bath, Fisher scientific, UK) to harvest the cells. These were then washed by centrifugation prior to RNA extraction using a combined mechanical disruption (0.5 mm glass beads) and chemical TRIzol™ method (Invitrogen, Paisley, UK). After DNase treatment (Qiagen, Crawley, UK) and purification (RNeasy MinElute clean up kit, Qiagen, Crawley, UK), cDNA was synthesized using a High Capacity RNA to cDNA kit (Life Technologies, Paisley, UK), and quantitative PCR performed using a SYBR® GreenER™ assay (Life Technologies Ltd, Paisley, UK). The primers used for quantitative PCR were ALS1, ALS3, ALS5 (agglutanin-like sequence 1, 3, and 5), EAP1 (epithelial adhesion protein 1), and AAF1 (adhesion and aggregation factor 1). Each parameter was analyzed in duplicate using MxProP Quantitative PCR machine and MxProP 3000 software (Stratagene, Amsterdam, Netherlands). Gene expression was normalized to the housekeeping gene ACT1 according to 2^−ΔΔCT method (Livak and Schmittgen, 2001). Table 1 summarizes all the primer details used in this study. A heatmap was created for the differential expression of genes (log₂) over the period of 3–90 min from the untreated control compared to CUR exposed cells. Maps and clusters were generated in R with the use of heatmap 0.2 function from the gplots package. All experimnets were performed in triplicate on three independent occasions.

As we were unable to ascertain that the data conformed to a Gaussian distribution data analysis was performed on non-parametric data using either a Mann–Whitney test or a Kruskal–Wallis test with Dunn's multiple comparison post-test. All independant data points are presented, with error bars representing the median with interquartile range. Where proportional data is presented, analysis was performed on the original data sets. All statistics and figures were produced using GraphPad Prism v.5 (GraphPad Software Inc., La Jolla, CA).

First, we tested the potential of CUR to inhibit and kill planktonic and biofilm cells to establish biologically active working concentrations suitable for use in downstream analyses. The planktonic MIC (PMIC) was shown to be 100 μg/ml of CUR for SC5314 and two other clinical strains tested (data not shown), whereas the sessile (biofilm) MIC that caused an 80% reduced metabolic activity (SMIC₈₀) was ≥ 200 μg/ml, demonstrating that the biofilm's activity and/or viabilty are significantly reduced.

Based on these data we wanted to evaluate whether these levels of CUR could be adsorbed to PMMA material to prevent C. albicans adhesion. We therefore adsorbed 200, 400, and 800 μg/ml of CUR to PMMA sections over different time periods and quantified these adsorbed concentrations using an elution method alongside an optimized standard curve. Figure 1A illustrates the kinetics of adsorption for each concentration. It was shown that 800 μg/ml CUR was required to achieve concentrations with anti-biofilm activity (200 μg/ml), though 90 min adsorption was required to achieve this. Nevertheless, after 10 and 30 min adsorption, 50 and 100 μg/ml concentrations were achieved from this initial concentration, respectively. The lower concentration of 400 μg/ml was able to achieve PMIC levels, though this took ~90 min adsorption. Finally, 200 μg/ml was not able to achieve any antimicrobial level concentrations, even after 24 h adsorption. Based on these data we focussed on adsorption of 800 μg/ml for 10 min, which was able to achieve 50 μg/ml on the surface of the PMMA for downstream analysis. Next, we evaluated the capacity of C. albicans to attach to PMMA for 30 min adsorbed with CUR (50 μg/ml) and compared a control (Figure 1B). Here we showed a significant three-fold reduction in adhesion of C. albicans was observed (p < 0.004). When analyzed by SEM the visible reduction of yeasts cells can be shown on the PMMA surfaces (Figure 1C).

Given that CUR was shown to elicit antimicrobial effects, both planktonically and to a lesser extent against sessile biofilms cells, and that adsorption appeared to influence adhesive capacity, we decided to investigate whether combining an early direct effect on C. albicans with that of adsorption to PMMA would synergize CUR activity in prevention of adhesion and colonization. C. albicans was treated for 3 min ± CUR 50 μg/ml (CA + CUR), which was compared to both PMMA adsorbed with CUR (PMMA + CUR), or a combination of CUR treated C. albicans and CUR adsorbed PMMA (PMMA CA + CUR). Figure 1D demonstrates that the direct treatment of C. albicans alone only inhibited 30 min adhesion by 27%, which was significantly lower than CUR adsorption alone (70%, p < 0.05). However, combining the effect on C. albicans with adsorbed CUR resulted in a significant reduction than adsorption alone of 93% (p < 0.001), thus improving the anti-adhesive capacity of C. albicans.

Our data above showed a positive anti-candidal effect with respect to surface adsorption, but also showed that a brief CUR pre-exposure (3 min) resulted in reduces adhesion of C. albicans, suggesting sub-inhibitory concentrations elicited some biological activity. We therefore sought to further investigate this effect on C. albicans by extending the CUR pre-exposure time. We were able to show that extending the time from 3 to 30 and 90 min significantly enhanced anti-adhesion (p > 0.01). Though, between 30 and 90 min there were no significant improvements in anti-adhesion properties (Figure 2A).

In order to understand the longer term effects of CUR treatment on C. albicans and how this could impact biofilm formation, we prepared yeast cells (Y = 0 min), germlings (G = 120 min) and hyphae (H = 240 min) prior to CUR exposure at 50, 100, and 200 μg/ml, which were then allowed to develop for biofilm for 24 h (Figure 2B). The resultant data showed that at lower sub-inhibitory concentrations (50 μg/ml) and PMIC levels (100 μg/ml) the anti-biofilm effects moderately impacted the overall biofilm metabolism, though at SMIC levels (200 μg/ml) a significant reduction in biofilm formation was observed for Y and G to approximately >90% of the control (p < 0.01), though H cells were least impacted (p < 0.05).

Next, we sought to determine whether CUR exhibited any additional effects on C. albicans. The premise of the experiment was to assess whether CUR induced aggregation through alteration of the cell wall surface. We reasoned that using a plate counting based approach of cells treated at sub-inhibitory concentrations, then if aggregation occurred then the CFU would be lower than the respective control as each aggregate would result in only one CFU due to a heterogeneous population of cells. Indeed, we demonstrated a significant reduction in cell counts in the CUR group (p < 0.01), which was further confirmed through light microscopy observations (Figure 2C).

Our data above suggest that CUR elicits biological effects on C. albicans, most significantly on preventing biofilm formation and impacting aggregation. We therefore aimed to assess a panel of associated genes through transcriptional analysis. To do this we focussed on Y and H cells at 3, 30, and 90 min post CUR exposure. Visual representation of these patterns was illustrated using heat map analysis and hierarchical clustering (Figure 3; raw C_T data profiles are presented in Supplementary Figure 1). We demonstrated that Y cells treated with CUR showed temporal changes in gene expression, most notably the down-regulation of the adhesin ALS3, and minimal impact on its related ALS1. Whereas, the clustered aggregative and flocculation genes AAF1, EAP1, and ALS5 transcripts were all up-regulated in a time dependant manner. Similar patterns and clustering of expression were also observed for the H cells, with AAF1 showing the highest levels of expression at 30 min compared to the control, and reciprocally ALS3 being the most down-regulated. Overall though the levels of differential expression were consistently lower in the H cells (Figure 3B) than Y cells (Figure 3A).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.