Bacteria Modify Candida albicans Hypha Formation, Microcolony Properties, and Survival within Macrophages

Candida albicans is the predominant fungus colonizing the oral cavity that can have both synergistic and antagonistic interactions with other bacteria. Interkingdom polymicrobial associations modify fungal pathogenicity and are believed to increase microbial resistance to innate immunity. However, it is not known how these interactions alter fungal survival during phagocytic killing. We demonstrated that secreted molecules of S. gordonii and P. aeruginosa alter C. albicans survival within the phagosome of macrophages and alter fungal pathogenic phenotypes, including filamentation and microcolony formation. Moreover, we provide evidence for a dual interaction between S. gordonii and C. albicans such that S. gordonii signaling peptides can promote C. albicans commensalism by decreasing microcolony attachment while increasing invasion in epithelial cells. Our results identify bacterial diffusible factors as an attractive target to modify virulence of C. albicans in polymicrobial infections.

virulent form of fungal growth due to their extensive hyphae that invade epithelial cells, as well as their high expression of several virulence genes, including ECE1 (encoding candidalysin, a peptide toxin critical for mucosal infection), HYR1 (encoding a hyphal wall protein that modulates phagocytic killing activity), and HWP1 (encoding a hyphal wall protein that mediates tight binding to oral epithelial cells) (2).
In the oral environment, C. albicans typically resides in complex polymicrobial communities, along with multiple bacterial and other fungal species. Interactions between C. albicans and single bacterial species may be synergistic or antagonistic and have been well described in excellent reviews (3)(4)(5). C. albicans exhibits cooperative relationships with multiple oral streptococci but has a particularly high affinity for Streptococcus gordonii (6,7), a Gram-positive bacterium that is one of the first colonizers of the oral cavity (8). Coinfection of C. albicans and oral streptococci in a murine model of oral candidiasis increases the severity of fungal lesions (9), and C. albicans biofilms formed in vitro were synergistically increased by streptococci (7,10). In addition to physical interactions, both species respond to signaling metabolites, including quorumsensing molecules. For example, autoinducer 2 secreted by S. gordonii relieves farnesolinduced suppression of C. albicans hyphal formation within cospecies biofilms (7), so that C. albicans virulence and invasion into epithelial tissues is promoted. The role of S. gordonii in C. albicans hyphal morphogenesis and virulence is further supported by a transcriptional study by Dutton et al. showing that coculture of C. albicans with S. gordonii increases expression of genes required for morphogenesis (TEC1 and ALS1) and oxidative stress response (CAT1) (11).
In contrast with S. gordonii, interactions between Pseudomonas aeruginosa and C. albicans are mainly antagonistic (12,13). P. aeruginosa is commonly isolated as part of mixed infections with C. albicans from the lungs of cystic fibrosis patients (14) and as a part of the oral microbiota (15). P. aeruginosa, a Gram-negative opportunistic rod, is becoming increasingly important in clinical settings because of its antibiotic resistance (16). Clinical isolates of P. aeruginosa are able to reduce C. albicans survival (17) and kill germinated C. albicans cells (18,19) due to secretion of bacterial phenazines (20). Besides its candidacidal activity, P. aeruginosa influences C. albicans morphology by reducing filamentation (21), and can adhere to C. albicans hyphal cells in response to a secreted quorum-sensing molecule, 3-oxododecanoyl-L-homoserine lactone (22). This antagonistic interaction is bilateral, since C. albicans can produce farnesol that reduces P. aeruginosa secretion of pyocyanin (23) and motility (24). Importantly, prior colonization of lungs by C. albicans increased the clearance of subsequent P. aeruginosa infection via leukocyte recruitment (25).
Key components of innate immunity are phagocytic cells, including macrophages that recognize and engulf C. albicans and other microbes into phagosomes where they are killed (26)(27)(28). C. albicans, as well as bacteria, can survive phagocytic killing (29)(30)(31) by suppressing phagosomal maturation (31) or reactive oxygen species (ROS) production (32,33). Despite the well-known relationships of C. albicans with S. gordonii and P. aeruginosa, it is unclear whether simultaneous phagocyte interactions with yeast and bacteria affects the subsequent survival of C. albicans. Phagocytosis of S. gordonii alone by activated macrophages can permit phagosomal disruption on maturation, resulting in increased bacterial survival (34), and P. aeruginosa can evade phagocytosis by modifying its motility (35,36). It is likely that cospecies interactions within phagocytes will further modify their survival.
In this study, we investigated the effect of these two bacteria on C. albicans after phagocytosis by macrophages. We show for the first time that phagocytosis of C. albicans in the presence of selected strains of these bacteria altered the survival of C. albicans within macrophages that was associated with induction or repression of fungal hyphae. P. aeruginosa decreased fungal survival in macrophages and inhibited C. albicans filamentation and microcolony formation, while S. gordonii increased survival and C. albicans hyphal formation and their escape from macrophage. C. albicans microcolony density and biomass were also significantly altered by products of both bacteria. Phenazines produced by P. aeruginosa reduced microcolony biomass, while S. gordonii culture supernatants increased biomass, but simultaneously caused loss of adhesion of fungal microcolonies to the underlying substrate and reduced epithelial invasion. Coincubation with S. gordonii also led to altered expression of some C. albicans microcolony virulence genes, suggesting that S. gordonii can modify microcolony virulence.
although we could not ascertain their absolute proximity. At 3 h after coinfection, neither S. gordonii CH1 or S. gordonii DL1 had significantly altered C. albicans survival within phagosomes compared to C. albicans alone (C. albicans survival of 99.7%; Fig. 1, lower left panel). In contrast, coinfection with S. gordonii SK12 significantly increased survival (137 to 145%) of C. albicans after 3 h compared to the control. In contrast, cospecies infection of C. albicans with P. aeruginosa clinical isolates (0635 or PA14) resulted in a 23% decrease in C. albicans survival after 3 h compared to C. albicans alone, which was not observed with the laboratory strain of P. aeruginosa (PAO1).
S. gordonii promoted and P. aeruginosa repressed C. albicans hyphal formation within macrophages. Since C. albicans is known to initiate hyphal formation in macrophages in order to escape killing (37), we next examined whether the presence of bacteria altered C. albicans hypha formation within the lumens of phagosomes. To compare hyphal length outside macrophages to that inside phagosomes, we used a higher MOI (0.5 C. albicans alone or 0.5 C. albicans:5 bacteria) so that some C. albicans (and bacteria in coinfection experiments) would not be phagocytosed, allowing us to compare germination outside macrophages with that within phagosomes. Surprisingly, the presence of S. gordonii DL1 (SgDL1) or SgSK12 both resulted in a significant (P Ͻ 0.01) increase in phagosomal C. albicans hyphal formation (89 to 86% hyphal formation with bacteria compared to 75% germination with C. albicans monoinfection) after 1 h (Fig. 2, left upper panel). After 2 h, there was no statistical difference between number of germinated cells between monoinfection and coinfection. However, 2 h after coinfection with S. gordonii, most C. albicans within the phagosome formed very elongated hyphae (white arrows) compared to monoinfection in which many C. albicans had very short hyphae (black arrows, Fig. 2, upper panel). In comparison, extracellular nonphagocytosed C. albicans (Fig. 2, CaϩSg panel, purple stain) formed similar long hyphae as phagosome-localized C. albicans in the presence of S. gordonii. Also, at 2 h after coinfection with S. gordonii, many intracellular C. albicans hyphae could be seen penetrating the phagosome so that the distal tips of hyphae were outside the macrophage. In monoinfections with C. albicans alone, most hyphae did not have sufficient length after 2 h to penetrate macrophages.
In contrast, the presence of P. aeruginosa had no effect on phagosomal C. albicans germination at 1 h, but after 2 h P. aeruginosa 0635 and PA14 significantly (P Ͻ 0.05) reduced hypha formation compared to monoinfection of C. albicans (Fig. 2, lower panel). Microscopically, coinfection with P. aeruginosa reduced hyphal length of C. albicans within the phagosome compared to extracellular C. albicans; and few hyphae were seen to escape macrophages at 2 h. To better quantify these observations, we next tested the ability of C. albicans to escape from macrophages and form microcolonies after coinfection with either S. gordonii or P. aeruginosa.
Bacteria modify escape of C. albicans from macrophages and alter the size of resulting C. albicans microcolonies. In order to measure relative escape ratios, macrophages were infected at an MOI of 1 macrophage: 0.1 C. albicans: 0.1 bacteria (or 1 macrophage: 0.1 C. albicans for monoinfection) to ensure nearly complete phagocytosis of both yeast and bacteria, incubated together for 3 h, and then streptomycin and penicillin were added to suppress the growth of escaped or nonphagocytosed bacteria. Macrophage monolayers and fungal cells then were grown at 37°C in 5% CO 2 for 17 h to permit visualization of C. albicans able to form microcolonies following escape from macrophages. C. albicans cells from monoinfections formed clearly visible microcolonies that grew on top of macrophages without damaging the majority of underlying cells (Fig. 3A), suggesting that the growth of escaped fungal cells did not result in lysis of nearby macrophages. The mean escape ratio for C. albicans monoinfection was 76 microcolonies per 100 added C. albicans (Fig. 3B). C. albicans and SgDL1 coinfection resulted in the formation of microcolonies that were much larger and denser than for those of monoinfection alone and increased the escape ratio to 96/100. C. albicans and P. aeruginosa PA14 (PaPA14) coinfection led to a much lower escape ratio of 55/100, and the microcolonies formed were much smaller and less dense. Since phenazine production by P. aeruginosa is known to inhibit filamentation, we examined a P. aeruginosa phenazine mutant (⌬phz) expecting that these cells would more closely resemble monoinfection by C. albicans. Indeed, ⌬phz cells had a significantly higher escape ratio (70/100) and formed denser and larger microcolonies than did the parental P. aeruginosa strain, suggesting that phenazine production by P. aeruginosa is largely responsible for the reduction in C. albicans microcolony size and reduced escape from macrophages.
Maturation of C. albicans-containing phagosomes and ROS production was not altered by S. gordonii and P. aeruginosa. To test whether changes in killing and escape of C. albicans from macrophages after coinfection with bacteria may be a result of changes in phagosomal maturation or ROS production, we measured the maturation kinetics of C. albicans-containing phagosomes by the loss of EEA1 (a marker of early maturation) and the acquisition of LAMP1 or lysosome-chased dextran (markers of phagosomal maturation) (34,38). Coinfection of macrophages with C. albicans and either S. gordonii or P. aeruginosa did not change the rate of phagosomal maturation, as measured by the loss of EEA1 or the increase of dextran. Acquisition of LAMP1 differed only for one strain of S. gordonii (SK12) (see Fig. S2), in total showing that C. albicans survival and escape from macrophages in the presence of bacteria was not likely to be altered by phagosomal maturation. Since production of ROS by phagosomes is an important killing mechanism (39), we also examined the total release of FIG 2 C. albicans hypha formation within macrophages is altered in the presence of bacteria. C. albicans CAF2-dTomato was added in an MOI of 0.5 to RAW 264.7 macrophages in the presence of S. gordonii (upper panels) or P. aeruginosa strains (lower panels). Macrophages and C. albicans were incubated without bacteria as a control. Nonphagocytosed C. albicans were stained with calcofluor white (blue), and macrophages were stained with phalloidin (green). At least 100 phagocytosed C. albicans (red) were counted and classified morphologically as yeast or hyphae microscopically. The percentage of hyphal cells was calculated by obtaining the ratio of total C. albicans with hyphae/total number of C. albicans counted ϫ 100. C. albicans long hyphae (white arrows) or short hyphae (black arrows) within macrophages are shown with or without bacteria (right panel). Experiments were carried out in duplicate, and graphs represent the means Ϯ the SD of three independent experiments. Significance was obtained by one-way ANOVA with post ad hoc Dunnett's multiple-comparison test (*, P Ͻ 0.05; **, P Ͻ 0.01). Scale bar, 10 m. ROS over 2.5 h from macrophages incubated with C. albicans in combination with S. gordonii or P. aeruginosa. There were no significant differences in ROS in macrophages coinfected with C. albicans and either S. gordonii or P. aeruginosa strains compared to monoinfection with C. albicans alone (Fig. S2). Thus, altered C. albicans survival in macrophages in the presence of bacteria did not appear to be a result of altered ROS production or delayed phagosomal maturation within the initial 90 to 180 min after coinfection.
C. albicans microcolonies are modified by secreted molecules of P. aeruginosa and S. gordonii. Since these data suggested a role for secreted bacterial products in altering C. albicans filamentation within macrophages, we tested whether either bacteria themselves or secreted products modified microcolony formation outside macrophages. For these experiments, C. albicans microcolonies were grown on solid surfaces in the presence of bacteria or bacterial supernatants to allow quantification of growth. Similar to what we observed within macrophages, C. albicans microcolonies formed in the presence of three P. aeruginosa strains (PAO1, Pa0635, and PA14) were smaller and had fewer hyphal projections (Fig. 4A), and addition of cell-free P. aeruginosa culture supernatants (sPa) resulted in nearly 50% reduction in microcolony density for PAO1 or PA14 strains (Fig. 4B). Quorum-sensing molecules secreted by P. aeruginosa play an important role during polymicrobial interactions (22,24) and, among them, phenazines , lower panels) for 17 h. The cells were allowed to phagocytose for 3 h. Bacterial growth was controlled by adding 1ϫ antibiotic solution. After incubation for 24 h, the microcolonies were imaged using bright-field microscopy. (B) The total number of microcolonies was counted for each well. Each experiment was carried out in duplicate, and graphs represent the means Ϯ the SD of three independent experiments. Significance was obtained by one-way ANOVA with post ad hoc Dunnett's multiplecomparison test (**, P Ͻ 0.01; ***, P Ͻ 0.001). Scale bar, 100 m.
are known to suppress C. albicans hyphal morphogenesis (19,20). We therefore decided to further examine the role of phenazines by using a P. aeruginosa phenazine knockout (⌬phz) mutant, along with synthetic the methylphenazinium analogs phenazine methosulfate (PMS) and pyocyanin (PYO) that suppress hypha morphogenesis in agar (40). The addition of ⌬phz cells still resulted in some reduction in C. albicans microcolony FIG 4 C. albicans microcolony density is decreased by P. aeruginosa and phenazines. (A) C. albicans microcolonies were cultured using RPMI 1640 medium at 37°C in 5% CO 2 , along with P. aeruginosa strains POA1, 0635, or PA14 or of fresh-filtered P. aeruginosa culture supernatant (10%; sPa), phenazine methosulfate (PMS; 5 M), or pyocyanin (PYO; 20 M) for 17 h. C. albicans incubated alone was used as control. Images were obtained using bright-field microscopy, and the microcolony density was calculated using ImageJ. (B) The microcolony density per square micron was obtained by using ImageJ. The addition of three P. aeruginosa strains or sPa each significantly decreased microcolony density compared to C. albicans grown in culture media alone. Experiments were carried out in duplicate, and the means Ϯ the SD of four independent experiments are shown. Significance was obtained by Student t test or one-way ANOVA with post ad hoc Dunnett's multiple-comparison test (*, P Ͻ 0.05; **, P Ͻ 0.01; ****, P Ͻ 0.0001). Scale bar, 100 m.
Bacteria Modify C. albicans Microcolony Formation size, but the P. aeruginosa ⌬phz culture supernatants did not inhibit microcolony formation consistent with the loss of secreted phenazines (Fig. 4). The addition of purified phenazines PMS (5 M) or PYO (20 M) almost completely blocked microcolony formation (Fig. 4A), showing that phenazines suppress C. albicans filamentation and repress microcolony formation.
Since we observed promotion of germination by S. gordonii cells in macrophages, we expected a similar effect on microcolony formation. Indeed, the addition of S. gordonii CH1, DL1, and SK12 cells, cell supernatants, and heat-fixed (HF) supernatants all increased the size of C. albicans microcolonies by increasing hyphal length and number of radiating hyphae (Fig. 5A). Quantification of microcolony density showed a significant increase in density between 25 and 50% among S. gordonii strains and with only fresh culture supernatants compared to C. albicans alone (Fig. 5B, left). However, we also noticed that microcolonies formed in the presence of S. gordonii lost their ability to adhere to the substrate, a phenomenon we did not find with P. aeruginosa albicans microcolonies were cultured using RPMI 1640 medium at 37°C in 5% CO 2 , along with S. gordonii strains CH1, DL1, and SK12; fresh-filtered S. gordonii culture supernatant (10%; sSg); or heat-fixed (HF sSg) S. gordonii culture supernatant (10%) for 17 h. Images were obtained using bright-field microscopy, and the microcolony density was calculated using ImageJ. The addition of S. gordonii or sSg each significantly increased microcolony density compared to C. albicans grown in culture media alone (B, right panel). Microcolony biomass (as determined with crystal violet staining) when grown with HF sSg were significantly increased with S. gordonii strains CH1, DL1, and SK12 (B, right panel). Experiments were carried out in duplicate, and the means Ϯ the SD of four independent experiments are shown. Significance was obtained by one-way ANOVA with post ad hoc Dunnett's multiple-comparison test (*, P Ͻ 0.05; ***, P Ͻ 0.001; ****, P Ͻ 0.0001). Scale bar, 100 m. experiments. Incubation with HF supernatants from each of the three S. gordonii strains with C. albicans that resulted in microcolonies that were so loosely adherent that they could not be quantitated by ImageJ. Instead, the total biomass of microcolonies was measured by crystal violet (CV) staining. The HF supernatants of all S. gordonii strains significantly increased C. albicans microcolony biomass by 3-fold compared to untreated C. albicans cells (Fig. 5B, right panel).
S. gordonii Eep contributes to C. albicans hypha formation and microcolony biomass. Many streptococcal signaling molecules known to affect C. albicans biofilm formation are small peptides and competence factors. Therefore, we first treated S. gordonii supernatants with trypsin (200 g/ml) or DNase (500 g/ml) before heat inactivation. Trypsin treatment significantly reduced microcolony biomass observed by 40%, while DNase treatment had no effect on microcolony formation compared to Todd-Hewitt (TH) medium only (Fig. 6), suggesting that peptides but not competence factors are likely to be responsible for this affect. To further validate these findings, we examined knockout mutants of S. gordonii strains defective in the production of secreted heat-stable competence factors (S. gordonii CH9278) or pheromone secretion (S. gordonii CH1Δeep). Among S. gordonii competence factors, the comCDE system encodes a sensor-regulator system (ComD ComE), which is activated by the comC gene product (CSP, competence stimulating peptide) and regulates the competence response. Deletion of the comCDE system has been shown to alter biofilm biomass by increasing extracellular DNA content in C. albicans and S. gordonii mixed biofilms (41). The total fungal biomass of microcolonies grown with HF supernatant of S. gordonii CH9278 cells (deficient in comCDE) was not different than microcolonies formed with the parental strain (S. gordonii CH1) (Fig. 6). However, the microcolony biomass was significantly decreased by 40% when incubated with HF supernatants of CH1Δeep that are deficient in the zinc metalloprotease Eep required for the processing of pheromones from lipoproteins (42). Furthermore, this effect was specific to Eep, as shown by the restored biomass with the complemented strain CH1Δeep/pSgEep1 (Fig. 6). These albicans microcolonies were cultured using RPMI 1640 medium at 37°C in 5% CO 2 , along with heat-fixed supernatants from S. gordonii (HF sSg) strains (10%) for 17 h, and the biomass was measured by crystal violet straining. Trypsin treatment of HF sSg CH1 significantly reduced C. albicans microcolony biomass, while DNase treatment had no effect. HF sSg of CH9278 missing comCDE competence regulatory system had no effect on C. albicans microcolony biomass. In contrast, the CH1 Δeep strain missing the zinc metalloprotease Eep significantly reduced C. albicans microcolony biomass, and this effect was abolished in the CH1Δeep/pSgEep1 complemented strain. C. albicans microcolony biomass was not altered by Todd-Hewitt (TH; 10%) medium alone. Experiments were carried out in duplicates, and the means Ϯ the SD of three independent experiments are shown. Significance was calculated by one-way ANOVA with post ad hoc Dunnett's multiple-comparison test (**, P Ͻ 0.01). results show that an S. gordonii heat-resistant, trypsin-sensitive pheromone processed by Eep promotes C. albicans hyphal microcolony formation.
Heat-fixed supernatant of S. gordonii increases the expression of C. albicans genes required for hyphal morphogenesis while decreasing the expression of adhesion genes. Since we observed that incubation with S. gordonii HF supernatant increased C. albicans microcolony size but decreased yeast adhesion to the substrate, we next measured the expression of genes required for C. albicans filamentation (EFG and HGC1), virulence (ECE1), and adhesion (HYR1, EAP1, HWP1, ALS3, and HWP2) by quantitative reverse transcription-PCR (RT-qPCR). Consistent with increased filamentation and microcolony size induced by S. gordonii (Fig. 5), C. albicans expression levels of EFG1 and HGC1 genes associated with filamentation were significantly increased (2to 7-fold) in the presence of S. gordonii HF supernatant (Fig. 7, upper panel). Intriguingly, we found that the expression of ECE1, required for virulence through candidalysin production, was significantly reduced in the presence of S. gordonii HF supernatant (Fig. 7, lower panel). Importantly, the expression of HYR1, EAP1, and HWP2 adhesins that are involved in initial attachment to surfaces was significantly decreased in the presence of all HF supernatants (Fig. 7, lower panel) and may partially account for the loss of adhesion of microcolonies to the surface that we observed. However, genes involved in some cell-cell and cell-substrate adhesion, including ALS3 and the hyphal wall protein HWP1, were upregulated, consistent with their hypha-specific expression. Together, these data confirm the increase in filamentation induced by S. gordonii, as shown by increased hypha-specific gene expression, along with altered expression levels among genes involved in adhesion.

FIG 7 S. gordonii heat-fixed supernatant increases the expression of C. albicans filamentation genes
while it decreases the expression of virulence and adhesion genes. C. albicans microcolonies were cultured using RPMI 1640 medium at 37°C in 5% CO 2 , along with heat-fixed supernatant from S. gordonii (HF sSg) strains CH1, DL1, and SK12 for 17 h. Total RNA from microcolonies was isolated, and the relative gene expression obtained by RT-qPCR was normalized to actin and GAPDH. C. albicans microcolonies grown without HF sSg were used as control, and the expression levels of each gene were set as 1. Microcolonies formed in the presence of HF sSg increased the expression of genes required for filamentation (EFG1 and HGC1), as well as cell-cell and cell-substrate adhesion (ALS3 and HWP1), but decreased the expression of other adhesins involved in surface attachment (HYR1, EAP1, and HWP2). Experiments were carried out in duplicate, and graphs indicate the means Ϯ the SD of four independent experiments. Significance was calculated one-way ANOVA with post ad hoc Dunnett's multiplecomparison test (*, P Ͻ 0.05; **, P Ͻ 0.01; ***, P Ͻ 0.001).
Heat-fixed supernatants of S. gordonii decreased C. albicans microcolony adherence to oral epithelial cells while increasing microcolony size. Since adhesion and virulence genes were altered in microcolonies in the presence of S. gordonii supernatants, we examined whether the presence of S. gordonii would modify C. albicans microcolony adherence and invasion with oral epithelial cells. We first grew microcolonies with 1.25 to 20% HF S. gordonii culture supernatants (HF sSg) for 17 h and then gently washed each well to remove nonadherent cells before visualizing adherent microcolonies (Fig. 8A). Even addition of 1.25% of HF sSg resulted in some loss of adhesion compared to microcolonies grown without HF sSg that remained firmly attached and evenly distributed over epithelial monolayers (Fig. 8A). The addition of increasing concentrations of HF sSg resulted in greater detachment of microcolonies until 70% of microcolonies grown with 20% HF sSg were removed by gentle washing. We also observed that among the approximately 30% microcolonies remaining attached with 20% HF sSg, many were much smaller. To determine whether invasion into epithelial monolayers might account for these differences in adhesion, hyphal invasion of microcolonies was assessed (Fig. 8B). Untreated C. albicans microcolonies showed a typical phenotype consisting of a large mass of noninvasive hyphae (Fig. 8B, left panel, shown in green) surrounding a central region of hyphae invading epithelium (Fig. 8B, left panel, gray regions indicated by a white dotted line). In contrast, unwashed floating microcolonies grown with 20% HF sSg were much larger and denser than microcolonies grown without HF sSg, and we were unable to detect invasive hyphae within these floating microcolonies. However, among the roughly 30% of small microcolonies grown with HF sSg that remained after washing, nearly 90% of the total area of the microcolony was found to be invading (Fig. 8B, lower right panel, white dotted line). Thus, while the majority of microcolonies grown with HF sSg on epithelium were similar to those formed on a solid substrate exhibiting a large dense floating phenotype, we discovered a subpopulation of microcolonies with a smaller but highly invasive phenotype.

DISCUSSION
C. albicans and P. aeruginosa are found together in the sputum and lungs of cystic fibrosis patients (14), while S. gordonii and C. albicans are commonly isolated together in various oral biofilms, including prosthetic surfaces in denture stomatitis (10,43). Macrophages act as immune sentinels performing a variety of functions (44), so that macrophage depletion increased the risk for systemic candidiasis and streptococcal bacterial load, both leading to increased mortality in mice (27,45). Furthermore, tissue-resident macrophages are important for triggering immune responses in the lung during P. aeruginosa pneumonia (46). Although phagocytic cells are highly effective at controlling infection, pathogenic bacteria and fungi have evolved multiple mechanisms to escape killing. C. albicans hypha formation within macrophages is an escape mechanism to subvert phagosomal clearance (47). Similarly, P. aeruginosa and certain strains of S. gordonii can also resist phagosomal killing in macrophages by promoting autophagy or suppressing phagosomal maturation (34,48). Our results indicate that C. albicans survival in macrophages in the presence of S. gordonii or P. aeruginosa are not dependent upon changes in macrophage phagocytic uptake, phagosome maturation, or ROS production. Instead, we found that germination and hyphal production induced by coinfecting bacteria were most predictive of fungal survival within macrophages (Fig. 2), so that hyphal expansion results in phagosome membrane damage (49) and eventual escape. Thus, we found that C. albicans escape and subsequent microcolony formation after 17 h was proportional to the degree of hyphal promotion or suppression induced earlier (2 to 3 h) by coinfecting bacteria.
To understand how S. gordonii and P. aeruginosa alter C. albicans filamentation, we examined bacterial effects on microcolony formation outside macrophages and found that the addition of heat-fixed supernatants of S. gordonii significantly altered fungal biomass. Although previous work found that S. gordonii cells promote fungal hyphal morphogenesis via direct contact (50), we found that S. gordonii supernatant alone was effective in promoting microcolony formation. Thus, while physical interaction between fungal and bacterial cells may provide synergy, S. gordonii binding to C. albicans is not required for this effect on filamentation. We also found that S. gordonii competence factors are not involved. Instead, a heat-stable, trypsin-sensitive small peptide processed by the Eep metalloprotease appears to be crucial for hyphal elongation in C. albicans microcolony formation. Although the S. gordonii pheromone autoinducer 2 has been suggested as one diffusible molecule affecting filamentation of C. albicans (7), to our knowledge this is the first instance of an Eep processed pheromone signal peptide affecting fungal growth. In addition to potential lipoprotein signal sequences known to be cleaved by Eep, any protein signal peptide encoded in the S. gordonii genome could potentially be cleaved by Eep, including a large number of unannotated small hydrophobic peptides. Thus, the range of possible targets of Eep processing and the numbers of released peptides are numerous. Eep has similarities to proteases processing certain pheromone precursors in Enterococcus faecalis (42,51); thus, it is likely that Eap may generate related small peptides from other bacteria that may similarly influence C. albicans filamentation. Further studies are needed to determine with more detail the range of bacteria able to influence microcolony formation.
The Dongari-Bagtzoglou group previously showed that the presence of oral streptococci increases the ability of C. albicans to invade organotypic models of oral mucosa (9) and that Streptococcus oralis can activate C. albicans EFG1 gene expression (a key inducer of hyphae and microcolonies) that increases polymicrobial biofilm (52). In addition, coinfections of S. oralis and C. albicans increase mucosal fungal invasion (53), and coculture of C. albicans with S. gordonii induces the expression of fungal filamentation genes and hyphal adhesins, including Als1, Hyr1, and Eap1 (11), some of which we showed are important for microcolony adhesion (54). However, we found that S. gordonii supernatants induced a dual phenotype, with the majority of C. albicans microcolonies becoming enlarged and detached from an underlying epithelium, while some microcolonies (30%) had less biomass but were highly invasive. This subpopu-lation of invasive cells may be a result of a microenvironment in which fungal cells have low exposure to secreted bacterial products that cause detachment or, alternatively, some fungal cells might become hyperinvasive in response to S. gordonii when in close contact with epithelium. Further work is needed to understand this phenotype. We found that the floating dense microcolonies had increased in the expression of ALS3 and HWP1, hypha-specific cell adhesion proteins. It is possible that in the context of these S. gordonii modified microcolonies, Als3 and Hwp1 function in cell-to-cell adhesion rather than cell-to-substrate adhesion, resulting in the phenotype of a denser and larger mass of hyphae that are not attached to the underlying surface, potentially increasing dissemination of C. albicans to distant sites. However, S. gordonii might also promote C. albicans commensalism by repressing expression of virulence genes involved in epithelial invasion (ECE1) and decreasing expression of genes required for adhesion (HYR1, EAP1, and HWP2). What differences exist in gene expression between these two microcolony phenotypes induced by S. gordonii remains an important question.
In contrast, P. aeruginosa induced a striking reduction of C. albicans microcolony size that was replicated when purified phenazines were added, suggesting that these secreted compounds are responsible for the repression of C. albicans hypha formation and possibly reduced survival within macrophages. Nonlethal concentration of phenazines secreted by P. aeruginosa are known to suppress C. albicans respiration and to acidify the extracellular pH, resulting in suppression of filamentation (40). Important differences have been found among P. aeruginosa clinical isolates in their ability to secrete phenazines and their quorum-sensing inducers (24,55). The production of phenazines is promoted in environments of low oxygen and nutrient deprivation (56) and by the presence of C. albicans (57), suggesting that these molecules may be more functional in confined spaces, such as the lumens of macrophages while phagocytosed together with C. albicans, compared to the more open environment of respiratory mucosa where phenazines are easily inactivated (58).
Interkingdom signaling is a major contributor in development of the microbiome. The identification of C. albicans genes modified by S. gordonii diffusible factors represents an attractive target to modify virulence of polymicrobial biofilms. Also, potential identification of new fungal pathways affected by S. gordonii or P. aeruginosa may suggest a broad basis for the regulation of fungal biofilms by cocolonizing bacteria.

MATERIALS AND METHODS
C. albicans and bacterial cultures. C. albicans and bacteria strains are listed in Table 1. C. albicans CAI-4 or SC5314 wild-type (WT) were cultured overnight in yeast extract-peptone-dextrose (YPD; BD Difco) broth supplemented with 50 g/ml of uridine (Sigma-Aldrich) at 30°C in an orbital shaker at 220 rpm. Cultures were diluted to an optical density at 600 nm (OD 600 ) of 0.3 to 0.4 in fresh YPD medium and allowed to reach an OD 600 of 0.6 to 0.7. The cells were then washed twice by centrifugation at 2,800 ϫ g for 5 min, with phosphate-buffered saline (PBS; pH 7.4; Corning), counted in a hemocytometer, CH9278 and CH1Δeep strains were maintained in TH agar supplemented with 250 g/ml spectinomycin (Sigma-Aldrich), and S. gordonii complement CH1Δeep/pSgEep1 was kept in TH agar supplemented with 250 g/ml spectinomycin and 5 g/ml erythromycin (Sigma-Aldrich). All three mutant strains were grown overnight in TH broth without antibiotics prior to use. P. aeruginosa strains were kept in Luria-Bertani (LB) agar (BD Biosciences) and cultured in broth in an orbital shaker at 220 rpm at 37°C. To evaluate the role of phenazines, P. aeruginosa Δphz strains, characterized by the deletion of the two redundant 7-gene operons phzA1-phzG1 (phz1) and phzA2-phzG2 (phz2) encoding the biosynthetic enzymes responsible for phenazine production, were compared to the PA14 parental strain. Each bacterial culture was grown for 10 h, diluted in fresh medium, and allowed to reach mid-log phase (OD 600 Ϸ 1) prior to use. The cultures were then spun down at 2,500 ϫ g for 3 min and washed two times in 1ϫ PBS and counted using a hemocytometer. S. gordonii cells were briefly sonicated for 30 s on ice to break bacterial chains into individual bacteria prior to counting. Fresh-filtered S. gordonii (sSg) or P. aeruginosa (sPa) culture supernatants were collected by centrifugation after log-phase growth and filtered using a 0.20-m syringe filter (Corning, Inc.). Heat-fixed S. gordonii culture supernatants (HF sSg) were obtained by boiling fresh-filtered supernatant of S. gordonii strains for 15 min and then stored at 4°C until use. When indicated, fresh-filtered S. gordonii supernatant was treated with DNase (500 g/ml; Sigma-Aldrich) or trypsin (200 g/ml; Sigma-Aldrich) for 10 min at 37°C with gentle shaking before heat treatment. Phagosomal survival and escape of C. albicans during coinfection with bacteria. Survival of C. albicans in macrophage phagosomes was performed as described previously (59) with modifications. Murine RAW 264.7 macrophages from the American Type Culture Collection (ATCC TIB-71) were seeded (5 ϫ 10 5 cells/ml) in 24-well plates (Corning, Inc.) with RPMI 1640 supplemented with L-glutamine (Corning) and 10% fetal bovine serum (FBS; Seradigm). The cells were activated using 10 ng/ml gamma interferon (BioLegend) for 12 h prior to experiments. C. albicans and bacterial cultures were each grown to mid-log phase as described above, added simultaneously at ratios of 10 macrophages:1 C. albicans:1 bacteria (MOI of 0.1:0.1) or 10 macrophages:1 C. albicans:10 bacteria (MOI of 0.1:1), and then incubated for 3 h at 37°C and 5% CO 2 . For survival assays, 0.25% SDS (Thermo Fisher Scientific) and sterile water were added to lyse macrophages and release phagocytosed C. albicans and bacteria. Lysates were serially diluted and cultured for 24 h at 30°C on yeast-dextrose-peptone agar supplemented with 50 g/ml streptomycin and 50 U/ml penicillin (Sigma-Aldrich) to remove bacteria and obtain C. albicans CFU. The percentage of survival was determined as follows: (recovered C. albicans CFU after phagocytic cell lysis/total number of phagocytosed C. albicans) ϫ 100. For escape assays, 10 macrophages:1 C. albicans:1 bacteria (MOI of 0.1:0.1) were incubated together for 3 h at 37°C and 5% CO 2 to allow for phagocytosis of yeast and bacterial cells, and then 50 g/ml streptomycin and 50 U/ml penicillin (Sigma-Aldrich) were added to each well to suppress the growth of nonphagocytosed bacteria or bacteria released by the lysis of macrophages. Macrophage monolayers were incubated at 37°C in 5% CO 2 for 17 h to permit visualization of C. albicans able to form microcolonies. Microcolonies on the surface of macrophage monolayers were imaged using a Zeiss Axio microscope, and the total number of microcolonies per well was counted. Assays were performed in duplicates, and results are representative of at least three independent experiments.
Evaluation of phagosome maturation in mouse macrophages. Macrophages were seeded in 12-well culture plates with cover glass (Azer Scientific) and activated as described above. Phagosomal maturation was measured (34) with some modifications. C. albicans and bacterial cells were counted, and added to macrophages at an MOIs of 1 and 10, respectively. Nonphagocytosed C. albicans cells were stained with calcofluor white (CW; Sigma-Aldrich) for 2 min on ice to differentiate them from phagocytosed cells. Cover glasses were washed, returned to fresh RPMI 1640 medium, and maintained at 37°C. Cells were sampled at indicated time points and immunostained for lysosomal associated membrane protein 1 (LAMP1) (clone 1D4B; Developmental Studies Hybridoma Bank, University of Iowa) or early endosomal antigen 1 (EEA1; Cell Signaling Technology) and Alexa 594-conjugated donkey anti-rat or anti-rabbit secondary antibody (Jackson Immunoresearch). For dextran labeling, macrophages were incubated with Alexa 594-dextran (Invitrogen) with a chase of 2 h to ensure all dextran was in lysosomes prior to performing phagocytosis experiments as described above. In all cases, macrophages were costained with Alexa Fluor 488-conjugated phalloidin (Invitrogen). At least 50 phagosomes were observed per time point using a Zeiss Axio Observer Z1 inverted fluorescence microscope (Carl Zeiss, Germany) and ZEN 2011 (blue edition) software. Phagosomes of internalized C. albicans (no CW stain) with total surrounding staining were counted as positive, and phagosomes with incomplete or nonexistent staining were counted as negative. The percentage of label-positive phagosomes was calculated as the number of positive phagosomes/total number of phagosomes counted ϫ 100.
Evaluation of hypha formation within macrophages. Fluorescent C. albicans CAF2-dTomato and unlabeled bacterial strains were added to macrophages at MOIs of 0.5 and 5, respectively, spun to allow contact, and incubated at 37°C and 5% CO 2 for 2 h. These MOIs were used so that some yeast and bacterial cells would not be phagocytosed and the formation of hyphae extracellularly could be compared to that of phagocytosed yeast. The phagocytic index for C. albicans cells was evaluated as previously described (59) with or without added bacteria to ensure that bacteria did not alter the phagocytic indices of yeast cells. Extracellular C. albicans was stained with CW, cells were fixed with 4% paraformaldehyde, and then macrophages were permeabilized and stained with phalloidin. A minimum of 100 phagocytosed Candida cells was counted using a Zeiss Axio microscope and classified as yeast or hyphae. The percentage of hyphal cells was calculated by obtaining the ratio of total hyphal form C. albicans/total number of Candida counted ϫ 100.
C. albicans microcolony formation. C. albicans microcolonies were formed by seeding 100 cells in 12-well plates using RPMI 1640. Whole S. gordonii or P. aeruginosa bacterial strains (1,000 cells per well), sSg or sPa (10% final volume), or heat-fixed (HF) sSg (1.25 to 20% final volume) were added to wells, followed by incubation at 37°C and 5% CO 2 for 17 h to form microcolonies. For experiments with phenazines, purified pyocyanin (30 mM) (PYO; Cayman Chemical) and phenazine methosulfate (25 mM; PMS; Acros Organic) were diluted in H 2 O to reach a final concentration after addition to wells (PYO [20 M] and PMS [5 M]). Microcolonies were imaged, and density was measured with ImageJ using inverse gray values/m 2 (2). The biomass of microcolonies grown with HF sSg was determined by crystal violet (CV) staining as described previously (60). The total absorbance (A 595 ) was obtained by subtracting negative controls (no cells) from experimental samples of destained solutions using a FlexStation 3 multimode microplate reader (Molecular Devices).
Microcolony invasion on oral epithelial cells. TR146 epithelial cells, a buccal epithelial squamous carcinoma cell line from the European Collection of Authenticated Cell Cultures (ECACC), were grown on glass coverslips to confluence in 1:1 Dulbecco modified Eagle medium (DMEM)/F-12 medium supplemented with 10% FBS. Invasion of epithelial cells by C. albicans microcolonies was performed as described previously (2) with or without 10% HF sSg. After 17 h of incubation, microcolonies were photographed with white light using an InGenius imaging system (Syngene). The medium covering the cells was then aspirated, and monolayers were washed and stained as described previously (2).
Microcolony RNA isolation and qRT-PCR. C. albicans microcolonies formed in the presence of HF sSg were collected and pelleted by centrifugation at 10,000 ϫ g for 5 min to isolate RNA as previously described (2). Total RNA was further purified by using an RNeasy minikit (Qiagen, Hilden, Germany) and quantified with a NanoDrop One (Thermo Scientific). Total RNA was used to quantitate HWP1, HWP2, ECE1, HYR1, EAP1, ALS3, EFG1, and HGC1 gene expression using primers listed in Supplemental Table S1. Total cDNA was synthesized using iScript cDNA synthesis kit (Bio-Rad). All samples were prepared with SsoAdvanced Universal SYBR Green Supermix (Bio-Rad) and cycled using a CFX Connect Real time system (Bio-Rad). Data were analyzed with CFX Maestro software (Bio-Rad). The relative fold changes in gene expression were calculated using both C. albicans actin and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) as controls.
Statistical analysis. Data were analyzed by a Student t test or one-way analysis of variance (ANOVA) with a post ad hoc Dunnett's multiple-comparison test using Prism v7 (GraphPad Software, La Jolla, CA) at a significance level of P Ͻ 0.05 for all experiments.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.