|Home | About | Journals | Submit | Contact Us | Français|
Until recently, NADPH oxidase (NOX) enzymes were thought to be a property of multicellularity, where the reactive oxygen species (ROS) produced by NOX acts in signaling processes or in attacking invading microbes through oxidative damage. We demonstrate here that the unicellular yeast and opportunistic fungal pathogen Candida albicans is capable of a ROS burst using a member of the NOX enzyme family, which we identify as Fre8. C. albicans can exist in either a unicellular yeast-like budding form or as filamentous multicellular hyphae or pseudohyphae, and the ROS burst of Fre8 begins as cells transition to the hyphal state. Fre8 is induced during hyphal morphogenesis and specifically produces ROS at the growing tip of the polarized cell. The superoxide dismutase Sod5 is co-induced with Fre8 and our findings are consistent with a model in which extracellular Sod5 acts as partner for Fre8, converting Fre8-derived superoxide to the diffusible H2O2 molecule. Mutants of fre8Δ/Δ exhibit a morphogenesis defect in vitro and are specifically impaired in development or maintenance of elongated hyphae, a defect that is rescued by exogenous sources of H2O2. A fre8Δ/Δ deficiency in hyphal development was similarly observed in vivo during C. albicans invasion of the kidney in a mouse model for disseminated candidiasis. Moreover C. albicans fre8Δ/Δ mutants showed defects in a rat catheter model for biofilms. Together these studies demonstrate that like multicellular organisms, C. albicans expresses NOX to produce ROS and this ROS helps drive fungal morphogenesis in the animal host.
We demonstrate here that the opportunistic human fungal pathogen Candida albicans uses a NADPH oxidase enzyme (NOX) and reactive oxygen species (ROS) to control morphogenesis in an animal host. C. albicans was not previously known to express NOX enzymes as these were thought to be a property of multicellular organisms, not unicellular yeasts. We describe here the identification of C. albicans Fre8 as the first NOX enzyme that can produce extracellular ROS in a unicellular yeast. C. albicans can exist as either a unicellular yeast or as multicellular elongated hyphae, and Fre8 is specially expressed during transition to the hyphal state where it works to produce ROS at the growing tip of the polarized cell. C. albicans cells lacking Fre8 exhibit a deficiency in elongated hyphae during fungal invasion of the kidney in a mouse model for systemic candidiasis. Moreover, Fre8 is required for fungal survival in a rodent model for catheter biofilms. These findings implicate a role for fungal derived ROS in controlling morphogenesis of this important fungal pathogen for public health.
Reactive oxygen species (ROS) including superoxide anion and hydrogen peroxide play diverse roles in biology. ROS can inflict severe oxidative damage to cellular components, but when carefully controlled, ROS can also be used to combat infection and act in cell signaling processes. A well-studied example of controlled ROS production involves NADPH oxidase (NOX) enzymes . These heme and flavin containing enzymes use electrons from NADPH to reduce molecular oxygen to superoxide . In macrophages and neutrophils, NOX enzymes generate bursts of superoxide in the extracellular milieu or phagolysosomal compartments to assault microbial pathogens. In non-immune cells, ROS from NOX enzymes are widely used in cell signaling pathways to promote growth, development and differentiation . As membrane proteins, NOX enzymes can vectorially release superoxide inside the cell or extracellularly and in either case, the superoxide can react with neighboring superoxide dismutase (SOD) enzymes that disproportionate superoxide to oxygen and hydrogen peroxide. In fact, NOX enzymes often partner with SODs in signaling processes, whereby SOD converts the cell impermeable superoxide to the diffusible hydrogen peroxide signaling molecule [1–5]. NOX-SOD interactions are also prevalent during infection where the microbial pathogen uses its arsenal of extracellular SODs to combat the oxidative burst of host NOX enzymes .
The opportunistic fungal pathogen Candida albicans has evolved with a family of three extracellular SOD enzymes (Sod4, Sod5, Sod6) believed to protect the fungus from the attack of host NOX-derived superoxide [7, 8]. We recently reported that these extracellular SODs represent a novel class of Cu-only SOD enzymes that are unique to the fungal kingdom and oomycetes [9, 10]. Much of what is known about fungal Cu-only SODs has emerged from studies on C. albicans Sod5. Sod5 can react with superoxide at rates limited only by diffusion [9, 10], and can effectively degrade superoxide radicals derived from macrophage and neutrophil NOX enzymes [11, 12].
Curiously C. albicans Sod5 appears specific to the filamentous form of the fungus [7, 13]. C. albicans is a polymorphic fungus that can transition from unicellular yeast-like form to pseudo hyphal and true hyphal filamentous states [14, 15]; Sod5 is evidently absent in the yeast-form of C. albicans. The rationale for selective expression of Sod5 during morphogenesis was not clear, as both yeast and hyphal forms exist in the animal host, are subject to immune surveillance and are essential for virulence [15–17]. Moreover, SOD5 is induced in filamentous C. albicans in the absence of any insult from the host [7, 13]. This raises the possibility that filamentous C. albicans witness a source of superoxide not seen in the yeast-like form.
Certain multicellular fungi are capable of generating superoxide themselves using fungal NOX enzymes as part of signaling during differentiation [18–20]. However, unicellular yeasts were believed to not express NOX, as NOX was characterized as a property of multicellular differentiation [21, 22]. This dogma of no NOX in unicellular fungi was recently challenged by the identification of Saccharomyces cerevisiae Yno1, a NOX that localizes to the endoplasmic reticulum and generates intracellular (not extracellular) superoxide . Other than Yno1, there has been no evidence for NOX enzymes in evolutionarily related yeasts including C. albicans.
Here we provide the first evidence for a NOX enzyme in the opportunistic fungal pathogen C. albicans. This NOX, known as Fre8, produces a burst of extracellular ROS in filamentous but not yeast-form cells. We demonstrate that Fre8-superoxide serves as substrate for Sod5, providing a rationale for inducing this extracellular SOD during morphogenesis. Strikingly, the ROS from Fre8 is concentrated at the growing tip of C. albicans hyphae and can promote formation and/or maintenance of elongated hyphae in vitro as well as in infected kidneys during a mouse model for disseminated candidiasis. Moreover, Fre8 enhances C. albicans survival in a rat venous catheter model of candidiasis. These studies show that host NOX is not the only source of ROS at the host-pathogen interface; C. albicans makes its own ROS for hyphal morphogenesis through Fre8.
Previously, Schroter et al. demonstrated that C. albicans produces ROS during the transition from the yeast form to the hyphal state . To probe this fungal ROS, we used luminol, a chemiluminescence probe typically used to measure ROS bursts in macrophages and neutrophils [11, 12, 25, 26]. Consistent with findings by Schroter et al , C. albicans cells induced to form hyphae by serum treatment exhibited a burst in luminol chemiluminesence not seen in yeast-form cells (Fig 1A). We observed that this ROS is not unique to serum stimulation but is also seen when morphogenesis is induced by elevated amino acid concentrations in IMDM medium and alkaline pH conditions (Fig 1A) [14, 27].
The luminol probe used in Fig 1A is not expected to penetrate the fungal cell wall, and should therefore only detect extracellular ROS. This notion of extracellular ROS was corroborated using the chemiluminescence probe lucigenin that cannot cross cell membranes, and is specific for superoxide compared to luminol which can detect both superoxide and hydrogen peroxide [25, 28]. As seen in Fig 1B, cells induced to form hyphae exhibited a defined lucigenin signal within 15 min of morphogenesis. However, in WT cells the lucigenin signal often declined at later times points (60 min, Fig 1B), and we tested whether this reflected induction of extracellular SOD enzymes. Indeed the superoxide signal from lucigenin was enhanced in sod4Δ/Δ sod5Δ/Δ sod6Δ/Δ cells lacking all three extracellular SODs (60 min, Fig 1B) and the same was true with luminol chemiluminescence (Fig 2B). Of the three extracellular SODs, deletion of SOD5 alone was sufficient to enhance ROS during morphogenesis stimulated by serum (Fig 2A) or IMDM (Fig 2B), consistent with the notion that Sod5 is the major extracellular SOD induced during hyphal formation . Compared to effects of sod5Δ/Δ mutations, there was no change in luminol ROS in sod1Δ/Δ mutants lacking the major intracellular Sod1 (Fig 2C). Together these studies demonstrate that cells undergoing morphogenesis produce a burst of extracellular ROS including superoxide that can serve as substrate for extracellular Sod5.
In multicellular organisms, extracellular ROS is derived from NOX enzymes, although C. albicans was not previously known to express NOX. To address whether the ROS burst of morphogenesis was derived from a NOX enzyme, we used DPI (diphenylene iodonium), a classical inhibitor of NOX enzymes [29–31]. As seen in Fig 2D, there was a dose response inhibition of the ROS burst of C. albicans using DPI, with full inhibition at 0.5 μM. DPI also eliminated the enhanced ROS of sod5Δ mutants (Fig 2C). These studies suggested that C. albicans expresses a NOX enzyme for extracellular ROS during morphogenesis. However, since DPI can also inhibit other flavin containing enzymes , we applied molecular genetic approaches to examine the source of the ROS burst.
In the fungal kingdom, NOX enzymes are part of an expanded family of NADPH oxidoreductases that use electrons from NADPH to either reduce oxygen to superoxide (NOX enzymes) or reduce ferric or cupric metal ions (FRE enzymes) . NOX and FREs are highly similar and it is difficult to predict functionality based on sequence analysis alone [23, 33, 34]. Yeasts are generally thought to only express FRE, not NOX [21, 22] although as mentioned above, this dogma was challenged by identification of S. cerevisiae Yno1 as an endoplasmic reticulum NOX . In C. albicans, there are at least 17 genes annotated as FREs [35, 36], three of which are known cupric or ferric reductases (Fre1, Fre7, Fre10 (S1 Fig, [37–39]); the remainder have uncharacterized functions. By qRT-PCR, we identified a number that are induced during early stages of hyphal morphogenesis coincident with the ROS burst (S1 Fig). The C. albicans orthologue to S. cerevisiaeYNO1  was not among the FREs induced with morphogenesis (S1 Fig). The most highly induced hyphal specific gene was FRE8, also known as CFL11 or C. albicans CR_06670W, orf19.701 (S1 Fig). FRE8 was additionally reported as the most abundantly induced FRE during C. albicans invasion of the kidney . We chose to focus on FRE8 as a potential NOX enzyme.
Recombinant versions of codon optimized FRE8 were expressed in Pichia pastoris under control of the methanol inducible AOX2 promoter. An identical procedure has been used to express and analyze activity of mammalian NOX enzyme complexes . In parallel, we expressed a second unknown member of the FRE family, namely FRP1, that is only moderately induced by hyphal stimulation (S1 Fig). P. pastoris cells expressing recombinant FRE8 and FRP1 were assessed for ferric reductase activity and ROS production. As seen in Fig 3A left, Pichia cells expressing FRP1 exhibited clear ferric reductase activity when FRP1 expression was induced with methanol. These cells however, exhibited no ROS that could be detected by luminol (Fig 3A right), indicating that FRP1 encodes a metalloreductase, and not a NOX enzyme. The opposite profile was obtained with Pichia cells expressing FRE8. These cells exhibited no methanol-inducible ferric reductase activity (Fig 3B left), but produced a strong methanol-inducible ROS signal (Fig 3B right). This ROS was eliminated by addition of 0.5 μM of the NOX inhibitor DPI or by addition of exogenous SOD enzyme (bovine SOD1). By contrast the ferric reductase activity of recombinant Frp1 was not altered by exogenous SOD or by DPI (Fig 3A left). These results indicated that FRP1 encodes a metalloreductase while FRE8 encodes a NOX.
Fre8 contains all the features predicted for a member of the NADPH family of oxidoreductases, including seven transmembrane domains and sequences for binding heme, FAD and NADPH (Fig 3C). To assess the function of Fre8 in vivo, we deleted both copies of genomic FRE8 in C. albicans and tested effects on metals and ROS formation. We observed no impact of fre8Δ/Δ mutations on accumulation of copper or iron (Fig 3D) or on whole cell ferric reductase activity (Fig 3E). However, the fre8Δ/Δ strain was completely defective in generating ROS, and this result held true regardless of the stimuli for morphogenesis including serum (Fig 4A), alkaline medium, IMDM and spider medium (Fig 4B). The heterozygous fre8Δ/+ strain retaining a single genomic copy of FRE8 exhibited roughly a 50% reduction in ROS generation and the same haploinsufficiency was seen with the fre8Δ/Δ strain complemented with a single copy of FRE8 (Fig 4A). As previously mentioned (Fig 2), the ROS burst of hyphal cells is enhanced in a sod5Δ/Δ mutant, and as shown in Fig 4C, this elevated ROS is Fre8-mediated, as no ROS is detected in a double sod5Δ/Δ fre8Δ/Δ mutant. The ROS burst emitted by C. albicans during the morphogenic switch is clearly Fre8-dependent.
We observed that FRE8 mRNA is induced within 1 hour of serum treatment (Fig 5A), congruent with the ROS burst (Fig 1A) and the induction of SOD5 (Fig 5A). In C. albicans, morphogenesis involves complex signaling pathways that converge on the Efg1 and Cph1 transcription factors, and efg1Δ/Δ cph1Δ/Δ null cells are incapable of forming hyphae [14, 42]. We observed that efg1Δ/Δ cph1Δ/Δ mutations block induction of FRE8 and SOD5 by serum (Fig 5A) and accordingly, the ROS burst is also eliminated (Fig 5B). SOD5 has previously been shown to fall under control of Efg1 . SOD5 and FRE8 were not identified by ChIP as direct targets of Efg1 or Cph1  ; both are subject to chromatin remodeling control by Hir1 that works in concert with Efg1 to control genes for morphogenesis .
Is hyphal formation required for ROS production? We uncoupled FRE8 expression from morphogenesis through ectopic expression of FRE8 under control of the repressible MET3 promoter in cph1Δ/Δ efg1Δ/Δ cells, fre8Δ/Δ cells and WT C. albicans. Strains were grown under conditions favoring yeast-only growth, and then FRE8 expression was de-repressed by removal of methionine. As seen in Fig 5C–5E there was a burst of ROS from all cells in which FRE8 expression was de-repressed. Importantly, these cells remained in the budding yeast-form, with no evidence of hyphal forms or germ tubes (Fig 5C–5E bottom). Thus the ROS burst is not dependent on morphogenesis, but rather the expression of FRE8, regardless of the morphogenic state. This study also demonstrated that Fre8-ROS is not sufficient to induce hyphal formation.
We next examined localization of Fre8-dependent ROS using nitrobluetetrazolium (NBT) which has been used to localize NOX superoxide in multicellular fungi [46, 47]. NBT is reduced by superoxide, forming a purple formazan precipitate. As seen in Fig 6, intense NBT staining was observed at the tip of elongating germ tubes in WT cells but not fre8Δ/Δ mutants (Fig 6A). This staining was discernable within 30 min of stimulating morphogenesis (Fig 6B). Thus, Fre8-dependent ROS is specific to the growing tip of the polarized cell.
Why would C. albicans produce ROS during hyphal formation? We considered whether this ROS might act as a signal to modulate morphogenesis, as has been shown for ROS modulating development in differentiating fungi [18–20]. Although no detectable change in morphology could be noted in the NBT experiment of Fig 6, these experiments were conducted at early time points when cells were initially forming germ tubes. However, at later time points when cell formed elongated hyphae, a fre8Δ/Δ deficiency could be observed (Fig 7A). This fre8Δ/Δ defect was complemented in the FRE8 re-integrant (Fig 8A top) and appeared completely specific to later stages of hyphal development. After 3–4 hours when WT cells were assembling into elongated hyphae, fre8Δ/Δ cells accumulated abundant yeast-forms (Fig 7C).
The experiments of Fig 7 employed cultures seeded at 6 x 107 cells/ml, a relatively high density where quorum sensing is prominent . However, quorum sensing cannot explain the fre8Δ/Δ defect, as these mutants showed no increased sensitivity to the quorum sensing molecule farnesol (S2 Fig parts A,B), and high density cultures of fre8Δ/Δ cells exhibited WT-like quorum sensing properties  (S2 Fig part C and legend). Most importantly, the fre8Δ/Δ defect in hyphal development can also be seen in low density cultures not subject to quorum sensing, e.g., when 4 x 106 cells/ml are stimulated with serum at 34°C (Fig 8B–8D). It is important to note that the requirement for Fre8 in hyphal formation is not absolute, and can be bypassed by potent stimuli for morphogenesis, e.g., high levels of serum (Fig 8A bottom) or when low density cells are shifted to temperatures ≥37°C (S2 Fig Part C). We conclude that Fre8 can act as a modifier of hyphal development, but is not unconditionally essential for the process.
Since Sod5 reacts with Fre8 superoxide to produce H2O2, we tested whether sod5Δ/Δ mutations likewise affect hyphal morphogenesis. As seen in Fig 9A and 9B, sod5Δ/Δ mutants exhibited a defect in hyphal development that was less pronounced than that of fre8Δ/Δ cells. Such an intermediate effect could be expected since Fre8 superoxide may also be converted to H2O2 through spontaneous disproportionation  or through extracellular Sod4 and Sod6. Even so, the parallel trends seen with sod5 and fre8 mutations would imply that H2O2 (and not superoxide) underlies the Fre8-defect. To more definitely test this, we addressed whether exogenous H2O2 could bypass the requirement for Fre8 in hyphal development. Previous studies have shown that mM concentrations of H2O2 can induce hyperpolarized buds in C. albicans [50, 51], or pseudohyphae [52, 53], neither of which resemble true C. albicans hyphae. Rather than using a single bolus of H2O2 as was previously done, we used glucose oxidase (GO) to continuously generate exogenous H2O2. As seen in Fig 9C, as little as 0.1 mU of GO (generates 100 pmoles H2O2/min) was able to restore hyphal development to fre8Δ/Δ cells while heat-inactivated GO was without effect. It is important to note that the hyphae formed with GO treated fre8Δ/Δ cells were indistinguishable from that of WT cells (Fig 9D), unlike the elongated buds and pseudohyphae reported for C. albicans treated with mM H2O2 [50–53]. Together, our in vitro studies of Figs Figs77–9 support a model in which the H2O2 produced by Fre8 can act as a modifier of hyphal morphogenesis.
It was important to examine the impact of Fre8 in vivo, as the animal host is the only natural environment for C. albicans. One model examined was the mouse model for disseminated candidiasis where kidney is the target organ. In late stages of infection, kidneys were harvested from mice infected with WT versus fre8Δ/Δ strains and subjected to histological analysis of invading fungi by PAS staining. As seen in Fig 10A top, WT C. albicans predominantly showed elongated hyphal filaments in the infected kidney. By comparison, the fre8Δ/Δ mutant from 4 independent mice produced a mixture of morphological forms with a much larger proportion of shorter filaments or yeast form cells in fre8Δ/Δ mutants compared to WT C. albicans (Fig 10A and 10B). These findings are similar to fre8Δ/Δ effects on morphology in vitro (Figs (Figs77–9). Yet in spite of the changes in morphology observed in vivo, there was no overall impact on pathogenesis with fre8Δ/Δ mutants. fre8Δ/Δ mutants showed no deficiency in virulence (Fig 10C), and markers of host inflammation [54, 55] were similar between SC5314 and fre8Δ/Δ infected kidneys (S3 Fig part A). Colony forming units (CFUs) in the kidney 48 hours post infection were modestly (2–3 fold) lower in fre8Δ/Δ mutations relative to WT (S3 Fig part B). All morphological forms are thought to contribute to infection and invasion , and our data with fre8Δ/Δ mutants supports this view.
We additionally tested the fre8Δ/Δ mutant in a rodent model of catheter biofilms. C. albicans is capable of forming surface adherent aggregates of biofilms on either biological surfaces (e.g., epithelial cells) or on medical implant devices, and such dense fungal communities are highly tolerant to antifungals . One of the most common clinical biofilm infections involves venous catheter implants and a rodent vascular catheter model faithfully recapitulates the human disease . In this model, WT SC5314 C. albicans forms robust biofilms within 24 hours of injection into the animal catheter (Fig 11A–11C). Over three independent trials, fre8Δ/Δ cells exhibited deficiencies whereby the biofilms were either sparse in number (Fig 11A) or undetected (Fig 11C), and when present, biofilms were often attenuated with few elongated hyphae (Fig 11B). This defect was partially reversed with the Fre8 re-integrant (Fig 11C), consistent with the haploinsufficiency and partial restoration of ROS formation in vitro (Fig 4A).
The fre8Δ/Δ defect in biofilms in vivo may very well reflect changes in morphogenesis similar to what we observed in vitro (Figs (Figs77–9). Yet host factors may also contribute. Neutrophils represent the primary leukocytes of Candida biofilms in catheters  and we tested whether fre8Δ/Δ cells were more sensitive to neutrophil killing. As seen in Fig 12A, the total mass of C. albicans biofilms in vitro in the absence of any host cells was unchanged in fre8Δ/Δ cells, indicating that there is no primary defect in biofilm formation including adherence. By comparison, fre8Δ/Δ biofilms exhibited a consistent increase in killing by neutrophils (Fig 12B). Thus, the fre8Δ/Δ defect with in vivo biofilms may not reflect a deficit in biofilm formation but rather increased clearance by host immune mechanisms including neutrophils.
We tested whether this increased killing was unique to neutrophils or could be extended to other phagocytes, e.g., macrophages. Our control for macrophage studies was the sod5Δ/Δ C. albicans strain that has been previously shown to be sensitive to macrophage killing due to an inability to degrade host superoxide . We observe that both sod5Δ/Δ and fre8Δ/Δ mutants show statistically significant increases in killing by bone marrow derived macrophages (BMDM) (Fig 12C). The impact of Fre8 derived ROS appears to extend beyond the hyphal morphology effects seen in fungal-only cultures and interactions with host cells are also important.
NOX enzymes have evolved to intentionally produce ROS, and until recently, were believed to be a property of multicellular differentiation [21, 22]. The discovery of Yno1 in S. cerevisiae demonstrated that a unicellular fungi can produce ROS through NOX, although the ROS in this case was found to be intracellular [23, 34]. Here we describe C. albicans Fre8 as the second example of NOX in an organism that can grow as a unicellular yeast and the first for this opportunistic fungal pathogen. Moreover, unlike S. cerevisiae Yno1, C. albicans Fre8 is capable of producing extracellular ROS, akin to NOX enzymes in multicellular organisms . In animal cells, NOX enzymes can partner with extracellular SODs that convert the extracellular superoxide free radical to the diffusible H2O2 molecule [1–5]. Likewise C. albicans Fre8 appears to partner with extracellular Sod5, providing a rationale for expressing Sod5 only in hyphal cells . These studies also provide a new twist to Sod5 function at the host-pathogen interface. While Sod5 can clearly react with superoxide from macrophage and neutrophil NOX enzymes [11, 12], our studies here with C. albicans Fre8 indicate that the superoxide for Sod5 is not just coming from the host. It is conceivable that when hyphal cells interact with neutrophils or macrophages, that a “superoxide superstorm” ensues with ROS coming from both the sides of the host-pathogen axis, and with Sod5 operating in the middle.
Why does C. albicans produce ROS during hyphal morphogenesis? Lessons may be taken from multicellular fungi or fruiting body fungi that use NOX derived ROS for morphogenesis and differentiation [18–20]. For example, ROS from these fungal NOX have been implicated in calcium signaling , MAP kinase and Rac1 GTPase signaling [46, 59] and cell re-modeling involving cytoskeleton effects [23, 60]. We show that Fre8 derived H2O2 can modulate morphogenesis. Work in other systems has shown that NOX derived H2O2 targets multiple redox sensitive molecules, including protein tyrosine phosphatases, casein kinases with peroxide sensitive degrons; even actin itself can be modulated by oxidation [1, 5, 61, 62]. Since Fre8 ROS is specifically seen at the growing tip of developing hyphae, the H2O2 produced may act locally on one or more redox sensitive targets that promote polarized growth. The precise mechanism of Fre8 control of hyphal biology is the subject of ongoing investigations.
In addition to the morphology defects of fre8Δ/Δ mutants in vitro and in vivo, we observed that these cells are more sensitive to killing by neutrophils and macrophages in vitro. It is possible that the morphological changes in fre8Δ/Δ may somehow render these fungal cells more susceptible to attack by phagocytes. As an alternative possibility, the ROS from Fre8 may help condition cells for the oxidative attack by immune cells. It has been proposed that low dose exposures of C. albicans to H2O2 or to ROS from macrophages can induce adaptive mechanisms to guard against subsequent oxidative insults [63–66]. FRE8 ROS may promote such adaptation against neutrophil and macrophage attack. In future studies, it will be important to determine the impact of fre8Δ/Δ mutations in immunocompromised settings.
Is Fre8 the only NOX of C. albicans? This organism has evolved with a very large family of 17 NOX/FRE enzymes, four of which are metalloreductases (including Frp1 characterized here), leaving 12 with unknown functions [35–39]. The extracellular ROS burst studied here is completely eliminated in fre8Δ/Δ cells suggesting that Fre8 is the only NOX for extracellular ROS in C. albicans at least under these in vitro conditions. During C. albicans invasion of the kidney, Fre8 is the most abundantly expressed member of the FRE/NOX family . Even so, it is possible that other members of this family induced during fungal infection and hyphal morphogenesis such as Fre2 (S1 Fig) may similarly function in a NOX capacity, perhaps secondary to Fre8 . C. albicans may also express NOX enzymes for intracellular ROS analogous to Yno1 of S. cerevisiae. With such a large family of NOX/FRE enzymes, we speculate that additional NOX enzymes will come to light as mediators of ROS signaling in C. albicans. Regardless, it will be of interest to integrate Fre8-ROS signaling into known pathways of hyphal regulation.
Cultures of Candida albicans cells were typically maintained at 30°C in a yeast extract, peptone based medium (YPD) with 2% (wt/vol) glucose, conditions which support the budding yeast-form of the fungus. The C. albicans fre8Δ/Δ and sod5Δ/Δ strains used in these studies grew identical to WT SC5314 in the yeast-form (S4 Fig). To induce hyphal morphogenesis, yeast-form cells were harvested, starved for 30 min in sterile H2O at 30°C, followed by harvesting and induction of hyphal formation by incubating at 37°C (or 34°C, see below) in various media known to stimulate hyphal formation, including Iscove’s Modified Dulbecco’s Medium (IMDM; Gibco), alkaline YPD (50 mM glycine, pH 9.5), spider medium (1% nutrient broth, 1% mannitol, 11.5 mM potassium phosphate, pH 7.2) or YPD with 5–20% fetal bovine serum (heat inactivated, Corning/Cellgro). Hyphal morphogenesis was stimulated in either “low density” (optical density, OD600 = 0.1–0.2) or “high density” (OD600 = 3.0) conditions. In studies of hyphal morphology, yeast-form cells were cultured to OD600 ≈ 8.0, conditions where all cells grew identically (S4 Fig), and were stimulated to form hyphae with YPD-serum. The level of serum used to investigate the fre8Δ/Δ defect ranged from 5–15% depending on the lot of serum, with low density cultures typically requiring less serum and temperatures of 34°C to demonstrate a dependence on serum for hyphal morphogenesis. Where indicated, cultures were supplemented with 0.1–10 mUnits glucose oxidase (Type II Sigma#G6125) to bypass the fre8Δ/Δ defect in hyphal development. Experiments involving yeast-form cells expressing FRE8 under the MET3 promoter used a synthetic complete (SC) based medium containing 0.67% yeast nitrogen base lacking cysteine and either containing or lacking 85.6 mg/L methionine. Cell were seeded at OD600 = 0.1 and grown for 1–7 hours.
All C. albicans strains used in this study were isogenic to SC5314 or its derivative CA-IF100 (arg4Δ/arg4Δ, leu2Δ/leu2Δ::cmLEU2, his1Δ/his1Δ::cdHIS1, URA3/ura3Δ). The sod1Δ/Δ, sod5Δ/Δ and sod4Δ/Δ sod5Δ/Δ sod6Δ/Δ strains derived from CA-IF100 were kind gifts of Karl Kuchler as previously described . The cph1Δ/Δ efg1Δ/Δ strain derived from SC5314 was a gift from Gerald Fink . Mutations in FRE8 and SOD5 were introduced in SC5314 using the SAT1-flipper cassette method . Deletion in a single FRE8 allele was achieved using plasmid pJGFRE8LKO, in which FRE8 regions -926 to -581 and +2402 to +2802 were inserted into the Kpn1 and XhoI and the NotI and SacI sites respectively of pSFS2 . Following liberation of the cassette by KpnI and SacI digestion and transformation of SC5314 by electroporation, accurate deletion of a single FRE8 allele was verified by PCR, generating the fre8Δ/+ mutant strain CA-JG201. The second FRE8 allele was deleted similarly using a pSFS2 construct, pJGFRE8SKO, containing FRE8 sequences -587 to -3 and +2075 to +2427, creating the fre8Δ/fre8Δ strain CA-JG211. Homozygous sod5Δ/Δ mutations were introduced in either SC5314 (generating strain CA-JG201) or CA-JG211 (generating CA-JG221) using a construct containing SOD5–492 to +53 and +808 to +1253 inserted into the Kpn1 and XhoI and the Not I and SacI sites respectively of pSFS2 . Deletion of both copies of sod5Δ/Δ in strain CA-JG201 and CAJG-221 was verified by PCR. A single copy of FRE8 was introduced into the fre8Δ/Δ strain CA-JG211 as follows: FRE8 sequences -926–+2802 were inserted into the KpnI and XhoI sites of pJGFRE8LKO. Integration into the FRE8 locus at position -926 to +2802 was achieved by transformation of the cassette liberated by digestion with KpnI and SacI generating the fre8Δ/Δ:FRE8 re-integrant strain CA-JG231. To create the construct for expressing FRE8 under control of the MET3 repressible promoter, C. albicans MET3 sequences -1643 to -1 were inserted into Sph1 and Nhe1 sites engineered at FRE8 position -1 in the pJGFRE8LKO re-integrant plasmid described above. Following digestion with Kpn1 and Sac1, the MET3-FRE8 containing cassette was used to transform the cph1Δ/Δ efg1Δ/Δ strain, the fre8Δ/Δ strain CA-JG211 and SC5314 by electroporation. Accurate integration at the FRE8 locus -926 to +2802 was verified by PCR.
Expression of recombinant FRE8 and FRP1 in Pichia pastoris used the PichiaPink Strain 1: ade2 (Thermo Fisher Scientific). P. pastoris cells were maintained in YP-Gal medium (1% yeast extract, 2% peptone, 2% galactose). Protein expression experiments used a buffered YP medium (1% yeast extract, 2% peptone, 100 mM potassium phosphate, pH 6.0, 1.34% yeast nitrogen base, 0.00004% biotin) that was supplemented with either 0.5% methanol to induce protein expression under the AOX2 promoter or with 1% glycerol for non-inducing conditions. The plasmid for expressing FRE8 or FRP1 under the P. pastoris AOX2 promoter represented a modified version of pPINK α-HC (Thermo Fisher Scientific) in which a 10X HIS tag was introduced downstream of the α-factor pre-sequence (plasmid pRPp718, kind gift of Ryan Peterson). Following the insertion of a Afe1 site downstream of the HIS tag, FRP1 sequences +1 to +1665 and FRE8 +1 to +2220 were inserted into the Afe1 and Fse1 sites of this expression plasmid, creating in-frame fusions to the N-terminus secretion sequence and HIS tag. The CTG codons from both genes were altered to TCG for optimal expression in P. pastoris; plasmids were linearized by digestion with Spe1 and integrated into the TRP2 locus by transformation.
For luminol and lucigenin measurements of ROS, either yeast-form cells (grown in YPD to OD600 of 1.0–2.0) or cells induced to form hyphae as described above were used. Cells were harvested, washed, and suspended in a final OD600 of 0.2 in Hanks buffered saline solution (HBSS) containing 0.2 mM luminol (Cayman chemicals) and 0.5 units/ml horseradish peroxidase. In studies with lucigenin, cells were first washed in 25 mM glycine pH 9.5, 0.5% glucose prior to resuspending in 200 μl of the same alkaline buffer containing 5 μM lucigenin. Samples were analyzed for luminol or lucigenin chemiluminescence in 96 well plates using a BioTek Synergy HT plate reader. Analysis was carried out over 1.5 hours at 37°C with a gain setting at 120–135 and integration time of 1.0 second. Results were plotted according to relative luminescence units (RLU) per 0.04 OD600 units of cells. With experiments involving DPI, 5 μl of DMSO containing the indicated amount of DPI (or no DPI as control) was added to the reaction at time zero.
For qRT-PCR analysis of fungal-only cultures, 50 ml cultures of cells were induced to form hyphae for 1 hr by growth in YPD-10% FBS (as described above); these early hyphal cells or the control yeast-form were harvested, washed and RNA prepared by the hot acid phenol method . cDNA was prepared using the Maxima H Minus First Strand cDNA Synthesis Kit (ThermoFisher Scientific) and qRT PCR carried out using iTaq Universal SYBR Green Supermix (Bio-Rad). Values were normalized to TUB2 and graphed according to the fold change in FRE8 and SOD5 expression in early hyphal versus yeast-form cells. Amplicons of ≈150 residues were prepared using primers as described in S1 Table.
For analysis of inflammatory mRNA markers (TNF-α, IL-17a, IL-6), RNA from whole kidneys was extracted as previously described . cDNA prepared from 5.0 μg of RNA was diluted 1:50 prior to PCR analysis as above. Values were normalized to ActB and graphed according to the fold change over uninfected controls. Primers for host mRNA analyses are listed in S1 Table.
For ferric reductase and NOX activity analyses in P. pastori transformants, cells were grown overnight in 10 ml YP-Gal, washed twice in either glycerol or methanol containing buffered media (described above) and resuspended at an OD600 of 0.1 in 15 mls of the same medium. Following growth at 30°C for 6 hrs, cells were harvested and washed in either HBSS for the luminol assay or 50 mM citrate, pH 6.6, 5% glucose for the ferric reductase assay. Cells were subjected to luminol chemilumiscence precisely as described above for C. albicans. Compared to C. albicans assays, the luminol substrate appears rapidly depleted in P. pastoris expressing high levels of FRE8. For the ferric reductase assay, cells at a OD600 of 0.5 were incubated in 200 μl of a reaction containing 1 mM FeCl3 and 1 mM bathophenanthrolinedisulfonic acid (BPS) in 50 mM citrate, pH 6.6, 5% glucose. Absorbance at 515 or 520 nm was read in 96 well plates on a BioTek Synergy HT plate reader over 1.5 hours at 30°C. Where designated, 0.1 U of bovine Cu/Zn SOD1 or 50 nM DPI were added to the luminol or ferric reductase assay at t = 0. Ferric reductase measurements in C. albicans cells was conducted similarly, using cells induced to form hyphae in IMDM for 1 hr as described above and assayed for ferric reductase using the same conditions described for P. pastoris except C. albicans cells were assayed at OD600 of 0.1.
Total cellular accumulation of copper and iron was measured by inductively coupled plasma mass spectrometry (ICP-MS) using C. albicans cells induced to form hyphae for 1 hr in 10% FBS as described above. Cells were washed twice with 10 mM Tris, 1 mM EDTA, pH 8 and twice with MiliQ deionized water. Cell pellets containing 10.0 OD600 units of cells were resuspended in 500 μl of 20% nitric acid and digested by incubation at 90°C overnight. Samples were diluted 10-fold in MiliQ deionized water and subjected to elemental analysis on a Agilent 7700x ICP-MS instrument.
In vitro biofilms were grown in the wells of 96-well microtiter plates, as previously described . Briefly, C. albicans resuspended in RPMI-MOPS at 1.5 x 106 cells/ml (200μL/well) was added, and incubated for 24 hours at 37°C with 5% CO2. To assess biofilm burden an XTT (2,3-Bis-(2-Methoxy-4-Nitro-5-Sulfophenyl)-2H-Tetrazolium-5-Carboxanilide) assay was performed as an estimate of viable burden, as previously described .
For assays involving neutrophils, human neutrophils were collected as follows: Blood was obtained from volunteer donors with written informed consent through a protocol approved by the University of Wisconsin Internal Review Board (IRB). Primary human neutrophils were purified by negative antibody selection using the MACSxpress Neutrophil Isolation and MACSxpress Erythrocyte Depletion kits (Miltenyi Biotec Inc., Auburn, CA), as previously described . Experiments with neutrophils were performed in RPMI 1640 (without phenol red) supplemented with 2% heat-inactivated fetal bovine serum (FBS) and glutamine (0.3 mg/ml). Incubations were at 37°C with 5% CO2. An adaptation of the XTT metabolic assay was used to estimate C. albicans viability following co-culture with neutrophils . Following a 24 h incubation period, biofilms were washed with DPBS and neutrophils were added at 1.5 x 106 cells/ml, which represented an effector:target of 1:2. Following a 4 h incubation, 90 μL of 9:1 XTT working solution (0.75 mg/ml XTT in DPBS with 2% glucose: phenazine methosulfate 0.32 mg/ml in ddH2O) was added to each well. After a 25 minutes incubation, samples were transferred to a Falcon 96 well U bottom plate and centrifuged at 1,200×g for three minutes to pellet cells. Supernatants (110 μl) were then transferred to a 96 well flat bottom plate for absorption reading at 492 nm. A neutrophil only control was used to subtract their contribution to the XTT values. To determine percent killing, values were compared to wells without neutrophils after subtraction of the baseline absorbance.
Macrophage infection assays used bone-marrow derived macrophages (BMDM) isolated from the marrow of hind leg bones of 5- to 8-wk-old C57BL-6 female mice. For differentiation, cells were seeded in 100 mm treated cell culture dishes (Corning, Corning, NY) in Dulbecco’s Modified Eagle medium (DMEM; Corning) with 20% L-929 cell-conditioned medium, 10% FBS (Atlanta Biologicals, Flowery Branch, GA), 2mM Glutamax (Gibco, Gaithersburg MD), 1% nonessential amino acids (Cellgro, Manassas, VA), 1% HEPES buffer, 1% penicillin-streptomycin and 0.1% 2-mercaptoethanol for 6–7 days at 37°C with 9.5% CO2. 105 BMDM were seeded on 96 well plates and activated by incubating overnight using 100 U/ml of IFN-γ (Roche, Indianapolis, IN). C. albicans obtained from overnight cultures in YPD (OD600 = 8.0) and starved in water as for hyphal morphogenesis studies (see above) were washed twice with PBS and incubated for 30 min at 37°C with Guinea pig complement (MP biomedicals, LLC, OH) for opsonization. The fungus was then added to macrophages at a multiplicity of infection (MOI) ratio of 1:10 for 4 hours. After incubation, the media was removed and macrophages lysed in water. Fungal viability was assessed by the XTT assay according to Pierce et al . The same XTT assay was used to determined fungal viability following farnesol treatment.
For the murine model of disseminated candidiasis, ten male BALB/c mice (10 weeks old) per strain were inoculated with 2x105 C. albicans cells of WT SC5314, the fre8Δ/Δ strain or the fre8Δ/Δ strain complemented by FRE8 by lateral tail vein injection. Moribund mice were sacrificed by CO2 asphyxiation and immediately dissected for harvesting kidneys for histology (see below). Fungal burden and host inflammatory markers were analyzed following 48 hours of infection. The spleen and one kidney was processed for CFUs as previously described . The other kidney was placed in 500 μL Trizol and frozen at -80°C for subsequent RNA analyses (see above). Mouse survival was plotted using a log rank test (Mantel Cox) to query any statistical difference.
A jugular vein rat central venous catheter biofilm infection model was used as previously described . Briefly, 24 h following surgical implantation of a jugular venous catheter, C. albicans at 106 cells/ml was instilled in the catheter lumen and flushed after 6 h. After 24 h biofilm growth period, catheters were harvested and fixed overnight (4% formaldehyde, 1% glutaraldehyde, in PBS). They were then washed with PBS, treated with 1% osmium tetroxide, and washed again. Samples were dehydrated through series of ethanol washes followed by critical point drying and mounted on aluminum stubs. Following sputter coating with platinum, samples were imaged in a scanning electron microscope (LEO 1530) at 3kV.
For NBT staining and microscopic analyses of cell morphology, C. albicans cells were induced to form hyphae as described above using YPD containing 10% FBS (in the case of NBT staining). Cells were harvested, washed once with HBSS and resuspended in 1 ml HBSS containing 0.05% nitroblue tetrazolium (NBT). Following an incubation for 30 min in the dark at 37°C, cells were washed 1X with HBSS, 1X with 70% Ethanol and resuspended in 200 μl 50% Glycerol/HBSS. Cells were visualized by light microscopy at 100x magnification on a Zeiss Axio ImagerA2 microscope. For analysis of hyphal morphogenesis, dark field microscopy of live C. albicans cells was accomplished using a Nikon Infinity 1 microscope at 40x magnification. Where indicated, enumeration of cells was carried with culture aliquots first passed through 26 gauge and 31 gauge needles to help break up dense aggregates and enhance visualization of individual cells. Passage through these needles did not affect integrity of the individual cells.
To analyze C. albicans morphology in infected kidneys, freshly harvested kidneys from infected mice were flash frozen in Tissue Tek O.C.T. compound in a dry ice/ethanol bath and were sectioned to 20 μM thickness by cryotome. Tissue slices were adhered to Superfrost Plus Microscope Slides (Fisherbrand Cat. No. 12-550-15) and subjected to Periodic Acid Schiff (PAS) staining by treatment with 0.5% periodic acid (Sigma) for 5 minutes, rinsing briefly with distilled water, then staining 5 minutes with Schiff’s Reagent (Sigma Aldrich). Following a 5 min rinse with water, the mounted tissue was dehydrated using successive 2 min treatments with 50%, 70%, 80%, and twice 95% and 100% ethanol, followed by three 2 min treatment with xylene isomer mixture (Sigma Aldrich) to remove residual ethanol. Cover slips were then mounted with Permount (Fisher) and slides then imaged on a microscope at 40X magnification.
All experiments involving animals were approved by the Johns Hopkins University (protocols # MO16M168 and MO15H134) and University of Wisconsin (protocol # DA0031, MV1947) Institutional Animal Care and Use Committees according to guidelines established by the Animal Welfare Act, The Institute of Laboratory Animal Resources Guide for the Care and Use of Laboratory Animals, and the Public Health Service Policy. Experiments involving neutrophils were approved by IRB (protocol #2013 1758) and involved cells isolated from healthy human adult donors in which written informed consent was obtained at the time of blood draw, following the guidelines and approval of the University of Wisconsin-Madison Center for Health Sciences Human Subjects Committee.
Expression of the various members of the FRE family listed by ORF designation were examined by qRT-PCR as described in Materials and Methods. Shown is the fold change in expression after 1 hour stimulation of hyphal morphogenesis by IMDM compared to yeast-form cells (cultured as in Fig 1). Results represent the averages of triplicate cultures. Genes that were previously characterized as cupric or ferric metalloreductases [37–39] or genes induced during fungal invasion of the kidney  are indicated by + marks. The C. albicans orthologue to S. cerevisiae Yno1  is indicated.
(A,B) WT SC5314 and isogenic fre8Δ/Δ cells were induced to form hyphae by culturing cells seeded at 4 x 106 cells/ml at 37°C with 5% serum in the presence of the indicated levels of farnesol or methanol vehicle. Following four hours, cells were either (A) photographed or (B) assayed for viability by XTT as described in Materials and Methods where results represent the averages of biological triplicates. The decrease in cell mass/viability with 300 and 400 μM farnesol is statistically significant as determined by ANOVA with Tukey post-test; ****p<0.0001. (C) SC5314 cells seeded at 4 x 106 cells/ml (low density) were cultured for four hours at 37°C in YPD supplemented with or without 5% serum or with conditioned 5% serum media derived from high density WT or fre8Δ/Δ cultures (6 x 107 cells/ml). The conditioned medium was obtained by removing cells from the high density cultures through centrifugation. Results show that low density SC5314 forms hyphae at 37°C even in the absence of serum (-serum), but hyphal formation is blocked by conditioned medium from high density WT and fre8Δ/Δ cultures, indicative of quorum sensing . Photographs are representative of 5–10 images over 2 experimental trials.
Mice were infected with either C. albicans WT SC5314 or the isogenic fre8Δ/Δ or the FRE8 complemented fre8Δ/Δ (fre8Δ/Δ Res) strain by lateral tail vein injection. Following 48 hours of infection, kidney and spleen were harvested and examined for (A) RNA markers of inflammation in the kidney by qRT-PCR as described in Materials and Methods, and (B) CFUs. Results are from 7–8 mice from each group. (A) The mRNA levels of the indicated inflammatory markers is shown as a fold change over uninfected controls. In all three infected strains, the increases in TNF- α, IL-6 and IL-17A are statistically significant compared to uninfected controls (****p<0.0001; ***p<0.0007). There is no statistically significant difference between WT and fre8Δ/Δ for any samples as determined by a one-way ANOVA with a Tukey post-test. There was a small (<2 fold) increase in expression of IL-17 and IL6 in the fre8Δ/Δ RES compared to WT, but the significance of this small variation is uncertain. (B) CFUs are shown as a function of tissue wet weight. The difference between fre8Δ/Δ and the fre8Δ/Δ strain complemented with FRE8 (fre8Δ/Δ RES) is significant (*p = 0.039). There is no statistically significant difference in CFUs obtained from spleen.
The indicated yeast strains were seeded at OD600 = 0.001 and grown at 30°C in YPD where growth by OD600 was either monitored continuously (TOP) or following a 16 hour period (BOTTOM). Results represent the averages of triplicate cultures (TOP) or of two to five experimental trials of hyphal morphogenesis (BOTTOM).
We thank Dr. Angelique Besold for informative discussions and for assistance with statistical analyses. We are also grateful to Dr. Ryan Peterson for plasmids.
This work was supported by National Institutes of Health Grants RO1 GM 50016 (to VCC), RO1 AI 119949 (to VCC), RO1 AI 073289 (to DRA), RO1 HL 059842 (to CC), KO8 AI108727 (to JEN), and F31 DK111114 (to EMC). Funding was also provided by a grant from Burroughs Wellcome Fund 1012299 (to JEN). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
All relevant data are within the paper and its Supporting Information files.