Clinical isolates of C. albicans were obtained: one commensal strain was isolated from the blood of a patient, and P. vulgaris and P. mirabilis were isolated from the urine of another patient. The identity of each microorganism was confirmed with the commercially-available identification systems (BioMeriéux, Marcy I'Etoile, France): API 32C for C. albicans and API 20E for P. vulgaris and P. mirabilis.

Prior to each experiment, C. albicans isolates were cultured at 30℃ for 48 hours on Sabouraud's dextrose agar (SDA, Difco™, Becton Dickinson, Spark, MD, USA), and one colony of yeast was inoculated into yeast nitrogen base (Difco™, Becton Dickinson) medium supplemented with 50 mM glucose. P. vulgaris and P. mirabilis were first subcultured at 37℃ for 18 hours on tryptic soy agar. One colony each of P. vulgaris and P. mirabilis was then inoculated into tryptic soy broth (Difco™, Becton Dickinson) and incubated at 37℃ for 18 hours. The experimental conditions were as follows: 1) the microorganism was cultured alone; 2) C. albicans was co-cultured with live P. vulgaris or P. mirabilis; 3) C. albicans was co-cultured with P. vulgaris or P. mirabilis killed at 100℃ for 30 minutes; or 4) C. albicans was treated with bacteria-cultured supernatants of P. vulgaris or P. mirabilis diluted four times, in which the bacteria were removed.

Biofilm formation was quantified using the method developed by Ramage, et al.23 Biofilms were formed on commercially available pre-sterilized, polystyrene, flat-bottomed, 96-well microtiter plates (Costar, Cambridge, MA, USA). Microorganisms were prepared for each condition and transferred to selected wells of a microtiter plate. The plate was incubated for 90 minutes at 37℃ in an orbital shaker at 75 rpm. After the initial adhesion phase, the cell suspensions were aspirated, and each well was washed twice with phosphate-buffered saline (PBS) to remove loose adherent cells. A volume of 200 µL of medium was added to each well, and the plate was then incubated for another 72 hours. After biofilm formation, the medium was aspirated, and non-adherent cells were removed by thoroughly washing the biofilm three times with PBS. A quantitative measure of biofilm formation was calculated using the XTT [2,3-bis(2-methyoxy-4-nitro-5-sulfo-phenyl)-2H-tetrazolium-5-carboxanilide] reduction assay. A 200-µL aliquot of XTT (1 mg/mL, Sigma, St. Louis, MO, USA) and menadione (0.4 mM, Sigma) solution was added to each well containing the prewashed biofilm and the control well. The plates were then incubated in the dark for up to 3 hours at 37℃. A colorimetric change resulting from XTT reduction was measured using a microtiter plate reader (EMax, Molecular Devices, Sunnyvale, CA, USA) at 490 nm.

After biofilm formation, the medium was aspirated, and non-adherent cells were removed by thoroughly washing the biofilm three times with PBS. Then, 1 mL of PBS was transferred to each well, and biomass was meticulously scraped off. The resultant solution containing the detached biofilm cells was gently vortexed for 1 minute to disrupt the aggregates and inoculated on an SDA plate. The colony forming units (CFUs) of C. albicans were quantified after 48 hours of incubation at 30℃.

We developed biofilms from single species, in addition to Candida biofilms that were co-cultured with bacteria on polystyrene coverslips as described. The coverslips were washed twice with PBS and placed in PBS with a fixative of 2.5% glutaraldehyde (Sigma) for 20 hours. Next, they were washed for 5 minutes in PBS and then placed in 1% osmium tetroxide for 30 minutes. After a series of alcohol washes, a final drying step was performed using the critical point drying method. Biofilms were then mounted and gold coated. Samples were imaged with a scanning electron microscope (TM-1000, Hitachi, Tokyo, Japan) in a high-vacuum mode at 15 kV.

RNA was isolated from C. albicans cells using the MasterPure Yeast RNA Extraction kit (Epicentre Biotechnologies, Madison, WI, USA). RNA was treated with amplification grade DNase I (Epicentre Biotechnologies) and used for cDNA synthesis with random hexamer primer (Invitrogen Life Technologies, Carlsbad, CA, USA) using Superscript II reverse transcriptase reagents (Invitrogen Life Technologies). Each reaction contained 1 µg of total RNA, 1 µL of 50 µM hexamer, and 1 µL of 10 mM dNTP in a final volume of 10 µL. Reactions were incubated at 65℃ for 5 minutes and cooled on ice. To each reaction tube, 10 µL of the following mixture was added: 4 µL of 5x First-Strand Buffer, 2 µL of 10 mM MgCl₂, 2 µL of 0.1 M DTT, 1.4 µL of RNase inhibitor, and 1 µL of Superscript II. Reactions were incubated at 42℃ for 50 minutes and then at 70℃ for 15 minutes. Real-time polymerase chain reaction (PCR) contained 10 µL of Power SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA), as well as forward and reverse primers (1 µL of each) (Table 1)22 and sterile water, at a final volume of 20 µL. The PCR was run on MicroAmp® Optical 384-well reaction plates in an ABI 7900 Real-Time PCR system (Applied Biosystems). Real-time PCR reactions were performed at 95℃ for 5 minutes, followed by 40 cycles of 15 seconds at 95℃ and 1 minute at 60℃. Dissociation curves were analyzed for all reactions to verify single peaks/products. Expression levels were analyzed using ABI 7900 System SDS software (Applied Biosystems). Real-time PCR data were normalized with the geometric mean of two reference genes. The ACT1 and PMA1 genes were used for this purpose.

All experiments were performed in triplicate on three different occasions. All data are expressed as mean values with corresponding standard deviations (SDs). Student's t-tests and Mann-Whitney U-tests were used to compare the differences between Candida only and Candida co-cultured with P. vulgaris or P. mirabilis. All p-values <0.05 were considered statistically significant.

The biofilm value generated when each type of microorganism was incubated separately was 0.07±0.007 for P. vulgaris, 0.120±0.004 for P. mirabilis, and 1.403±0.103 for C. albicans. When culturing C. albicans and P. vulgaris or P. mirabilis together, biofilm formation was reduced by more than 80%, compared to the C. albicans culture alone. To assess the mechanism by which the bacteria impeded Candida biofilm formation, bacteria were initially treated for 30 minutes at 100℃ to eliminate biological activity, followed by co-culture with C. albicans. Biofilm formation of C. albicans cultured together with heat-killed P. vulgaris or P. mirabilis elicited a 70% reduction, compared to the control group. These data indicate that dead bacteria interfere with the structural formation of C. albicans biofilms (Fig. 1A).

When C. albicans and P. vulgaris or P. mirabilis were co-cultured, biofilm formation of C. albicans was significantly reduced. In order to determine if this reduction was due to the effect of the bacteria or of secretory products when cultured, P. vulgaris and P. mirabilis were cultured for 72 hours, and remaining bacteria were removed by filtration. As a result of treatment with P. vulgaris and P. mirabilis culture supernatants, C. albicans biofilm formation was reduced by 60–70%, compared to C. albicans cultured alone (Fig. 1B). To determine whether this effect was due to a depletion of nutrients in the medium or to the secretory products of the bacteria, we diluted the cultured supernatants and tested the concentration effect on C. albicans biofilm formation. C. albicans biofilm formation decreased in proportion to the concentration of the bacterial-cultured supernatants (data not shown). To determine whether P. vulgaris and P. mirabilis inhibited the growth of C. albicans, C. albicans and P. vulgaris or P. mirabilis or bacterial-cultured supernatants were cultured together for 72 hours, and C. albicans CFUs were calculated. When C. albicans was cultured alone, the count was 2.85×10⁸ CFU/mL; when cultured together with P. vulgaris, the count of C. albicans was reduced to 9×10⁶ CFU/mL; and when C. albicans was cultured together with P. mirabilis, the count of C. albicans dropped to 2.4×10⁷ CFU/mL (Fig.1C). It seems that P. vulgaris and P. mirabilis inhibit C. albicans biofilm formation and also interfere with its growth. As a result, we investigated the C. albicans CFUs after treating C. albicans with bacterial-cultured supernatants and culturing for 72 hours. In the case of culturing C. albicans alone, the count was 2.85×10⁸ CFU/mL. Meanwhile, the count from the treated culture supernatants of P. vulgaris was 3.6×10⁷ CFU/mL, and the count from the treated culture supernatants of P. mirabilis was reduced to 2.0×10⁷ CFU/mL (Fig. 1C).

C. albicans morphology changes to the hyphae form from the yeast form as biofilm formation progresses, and the expression pattern of genes related to morphology at these stages was identified. Previous studies have shown that expression patterns of Candida species-related genes are significantly increased during hyphae-specific gene expression over the duration of biofilm formation. On the contrary, the expression of genes suppressing the formation of yeast-specific genes and filament are decreased.22 These results suggest that Candida present in yeast form adhere to a surface to increase the expression of hyphae-related genes and promote the formation of biofilms by decreasing the expression of genes that inhibit the yeast form and filament formation. To clarify the effect of co-cultures on biofilm formation, we analyzed changes in C. albicans gene expression levels in biofilms co-cultured with P. vulgaris or P. mirabilis. In contrast to C. albicans cultured alone, als3 and hwp1 showed a reduction in expression by 80% and ece1 and sap5 by 90% (Fig. 2A). When C. albicans was cultured together with P. vulgaris or P. mirabilis, the expression of tup1 and nrg1, which are genes known to suppress filament formation, increased by about two-fold, compared to when C. albicans was cultured alone (Fig. 2B). Regarding the expression of yeast-specific rhd1 and rbe1, the expression levels of rhd1 increased by 2.5-fold and rbe1 by about more than 3-fold (Fig. 2C).

When C. albicans was co-cultured with heat-killed P. vulgaris and P. mirabilis, biofilm formation was decreased. We confirmed that these results were due to changes in the expression of biofilm-related genes in C. albicans. First, ece1, hwp1, and sap5 gene levels, which are associated with the formation of hyphae, did not increase. While als3 level was slightly increased, the change was not significant (Fig. 3A). There were also no significant changes in tup1 and nrg1, which are filament formation inhibiting genes (Fig. 3B). Yeast-specific genes also showed no difference in expression levels (Fig. 3C). The reduction in C. albicans biofilm formation in co-cultures with killed bacteria is considered to be caused not by changes in gene expression, but by the interference of killed bacteria acting as small particles that fit between C. albicans cells and, thus, suppress the structural formation of biofilm.

As described above, treatment of P. vulgaris and P. mirabilis cultured supernatants inhibited C. albicans biofilm formation. We examined how this treatment affected the expression of various genes involved in biofilm formation. The hyphae-specific genes als3, ece1, hwp1, and sap5 all showed a significant reduction in expression in contrast to the cultured supernatants in which C. albicans was cultured alone (Fig. 4A). In contrast, the filament formation inhibiting genes tup1 and nrg1 slightly increased (Fig. 4B). The yeast-specific gene rhd1 increased by 2-fold, and rbe1 slightly increased (Fig. 4C). This suggests that secretory products that are formed and released with growth of P. vulgaris and P. mirabilis inhibit the growth of C. albicans and regulate the expression of biofilm-related genes, thereby inhibiting biofilm formation in C. albicans.

It is known that biofilms are not formed with a simple structure and that Candida exist in the yeast form in the basal layer and in the hyphae and pseudohyphae form in the layer above. These layers form a network with a three-dimensional structure. The structural difference between the biofilm when C. albicans is cultivated separately and when co-cultured with P. vulgaris or P. mirabilis was investigated via scanning electron microscopy (Fig. 5). The biofilm of C. albicans when cultured alone was high in density and was a multi-layer solid (Fig. 5A and B). The biofilm formed when C. albicans and P. vulgaris or P. mirabilis were cultured together showed that the P. vulgaris or P. mirabilis were attached to the hyphae of C. albicans and inserted between the C. albicans cells. The thickness of the biofilm in these cases also appeared to be thin with a low density (Fig. 5C and E). The structure of biofilms formed after culturing heat-killed bacteria and C. albicans together was thinner and showed lower density than biofilms formed after separate culture of C. albicans (Fig. 5D and F). However, the heat-killed bacteria biofilm was thicker than the biofilm formed after C. albicans had been cultured together with live bacteria, and more of the C. albicans was in the yeast form. The above results agree with the outcome of XTT reduction assays, which quantitatively confirmed biofilm formation (Fig. 1A).