Bmh1p (14-3-3) mediates pathways associated with virulence in Candida albicans

Abstract

The ability of the pathogenic fungus Candida albicans to cause disease requires rapid adaptation to changes in the host environment and to an evolving host immune response. The identification of ‘virulence factors’ using in vitro characterization of mutant strains has traditionally relied on a common set of phenotypic and biochemical assays (most often performed at 30 °C) and the subsequent correlation with their corresponding virulence in mouse models of disease. Utilizing a panel of isogenic mutants for the multifunctional signal-modulating 14-3-3 protein (Bmh1p), we have found that specific mutations affect a variety of different pathways currently associated with virulence, including those involved with the formation of filaments, as well as interaction with host immune cells. Surprisingly, our studies revealed that deficiencies in many of these pathways do not always correlate with virulence in a mouse model of disseminated infection. Mutations within the binding pocket of Bmh1p that affect the ability of the protein to efficiently bind ligand had varying effects on the results of a number of in vitro and in vivo assays. The capability, in vitro, to filament in embedment conditions, and to filament and form chlamydospores under microaerophilic conditions on cornmeal agar, does not correlate with virulence. It is likely that only a subset of hyphal signalling pathways is actually required for the establishment of infection in the disseminated mouse model. Most importantly, our results suggest that the delayed onset of lag-phase growth in vitro at 37 °C, and not at 30 °C, results in an inability of these mutants to rapidly adjust to environmental changes in vivo and may be responsible for their increased clearance and reduced virulence. It is critical, therefore, that future in vitro studies of putative virulence factors in C. albicans include careful characterization at physiological temperatures.

INTRODUCTION

Candida albicans is part of the normal microbial flora of humans that colonizes mucocutaneous surfaces. Both humoral and cell-mediated immune responses are evoked in the immunocompetent host and prevent this organism from causing disease. However, changes in the host environment can induce C. albicans to shift from a benign, commensal organism to an invasive pathogen, resulting in multiple clinical presentations from cutaneous to systemic candidiasis, depending on the patient's specific immune dysfunction (de Repentigny et al., 2004; Odds, 1988; Rogers & Balish, 1980). Transcriptome analyses of C. albicans in response to the host environment or after interaction with host cells have shown that several pathways are differentially induced, leading to upregulation of Candida genes involved in the stress response, antioxidant response, glyoxylate cycle, putative virulence attributes, and hyphal formation (Fradin et al., 2003; Lorenz & Fink, 2002; Lorenz et al., 2004; Nantel et al., 2002; Rubin-Bejerano et al., 2003). Interestingly, several of these pathways are also disregulated in Saccharomyces cerevisiae 14-3-3 mutants (Bruckmann et al., 2004; Ichimura et al., 2004).

The 14-3-3 family of proteins plays key functional roles in various pathways regulated by phosphorylation of serines and threonines, and is highly conserved in all eukaryotic species (Aitken, 2006; Yaffe, 2004). In yeast, two 14-3-3 isoforms are predominantly redundant and facilitate multiple cellular processes including mitotic checkpoint control and cellular differentiation in response to appropriate environmental signals (van Heusden & Steensma, 2006). C. albicans, unlike most other eukaryotes, encodes a single 14-3-3 gene (BMH1), which is essential. C. albicans strains expressing a single BMH1 allele demonstrate significant defects in two different filamentation assays compared with the wild-type. These include (1) a defect in filamentation under embedment conditions, and (2) an inability to filament or form chlamydospores under microaerophilic conditions on cornmeal agar (Table 1) (Cognetti et al., 2002; Palmer et al., 2004; Palmer & Sturtevant, 2004). Furthermore, with regard to these two filamentation responses, isogenic strains expressing mutant 14-3-3 isoforms demonstrate distinct aberrant phenotypes (Tables 1 and 2) (Palmer et al., 2004; Palmer & Sturtevant, 2004). For example, K51E and L231S mutants filament under chlamydospore-inducing conditions but do not form chlamydospores, while M125R and R142C mutants respond to embedment and chlamydospore conditions comparably to the parental strain. A conservative K51R substitution is less filamentous under embedment conditions than the other strains. A subset of bmh1 mutants also demonstrate reduced filamentation on solid agar media and a lag in the onset of filamentation in liquid media (Table 1).

Table 1.

Phenotypes of bmh1 mutants

In vitro filamentation phenotypes of bmh1 mutants have been previously published.

Strain	Growth*	Filamentation induction				Chlamy-dospores¶	Macrophage viability#	Virulence** (days)
		M199/FBS† (mm)	Embedment matrix‡	Liquid FBS (min)§	Chlamy-dospore||
6284	1.0	5–6	100 %	30	50–75 %	Yes	0–6	7
BH1P1	1.2	3–5	20–40 %††	30	0 %	No	nd	5.5
BMH	1.2	3–5	20–30 %††	30–60	0 %	No	0–5	7
K51E	3.5	0–1	5–9 %††	60–120	30–50 %	No	20–150	>31
K51R	1.3	0–2	0 %	30–60	0 %	No	0–6	12
L231S	3.3	0–1	19 %††	60–120	80–100 %	No	15–115	>31
M125R	1.2	3–6	100 %	30	40–60 %	Yes	4–16	24.5
R142C	1.0	5–6	100 %	30	40–75 %	Yes	0–7	7

*Lag time was calculated as time in hours to enter exponential growth compared with the WT control strain at 30 °C. Once cells entered exponential growth, generation times did not differ significantly (90–130 min, depending on the experiment).

†Cells were plated on M199 or 10 % FBS solid media and length of filaments was assessed at 7 days. L231S and K51E were afilamentous on FBS media; K51R was only sparsely filamentous on FBS media.

‡Cells were mixed with molten YPS (yeast peptone, 2 % sucrose) and monitored for 6 days, and the percentage of filamenting colonies per 100 colonies was counted.

§Cells were incubated in 10 % FBS; the ability to filament was assessed by onset of filamentation (min). Extensions of K51E and L231S filaments were not as long as seen for other strains.

||Cells were grown on cornmeal agar under microaerophilic conditions for 9 days, and the number of filamentous colonies was assessed.

¶Cells were grown on cornmeal agar under microaerophilic conditions and monitored for production or no production of chlamydospores.

#Strains were incubated with macrophages for 24 h, and macrophage viability was assessed as the mean number of macrophages per field (×20); nd, not determined.

**Virulence was assessed in the murine disseminated mouse model as described above and presented as median survival.

††Colonies were heterogeneous and smaller in size than those of the WT.

Table 2.

bmh1 mutant strains and relevant genotypes

Strain	Parent	Relevant genotype	Reference
YJB6284	BWP17	ura3Δ/ura3Δ arg4Δ/arg4Δ : : ARG4URA3 his1Δ/his1Δ : : HIS1 BMH1/BMH1	Bensen et al. (2002)
BH1P1	BWP17	BMH1/bmh1Δ : : HIS1 ura3Δ/ura3Δ arg4Δ/arg4 his1Δ/his1Δ	Palmer et al. (2004)
BMH*	BH1	BMH1/bmh1Δ ura3Δ/ura3Δ arg4Δ/arg4Δ : : ARG4URA3 his1Δ/his1Δ : : HIS1	Palmer et al. (2004)
K51E	BH1	bmh1Δ : : ARG4/bmh1Δ : : HIS1 ura3Δ/ura3Δ : : URA3 : : BMH1 arg4Δ/arg4Δ his1Δ/his1Δ	Palmer et al. (2004)
K51R	BH1	bmh1Δ : : ARG4/bmh1Δ : : HIS1 ura3Δ/ura3Δ : : URA3 : : bmh1K51E arg4Δ/arg4Δ his1Δ/his1Δ	Palmer et al. (2004)
L231S	BH1	bmh1Δ : : ARG4/bmh1Δ : : HIS1 ura3Δ/ura3Δ : : URA3 : : bmh1K51R arg4Δ/arg4Δ his1Δ/his1Δ	Palmer & Sturtevant (2004)
M125R	BH1	bmh1Δ : : ARG4/bmh1Δ : : HIS1 ura3Δ/ura3Δ : : URA3 : : bmh1L231S arg4Δ/arg4Δ his1Δ/his1Δ	Palmer & Sturtevant (2004)
R142C	BH1	bmh1Δ : : ARG4/bmh1Δ : : HIS1 ura3Δ/ura3Δ : : URA3 : : bmh1M125R arg4Δ/arg4Δ his1Δ/his1Δ	Palmer & Sturtevant (2004)

*The bmh1 mutant strains with both endogenous BMH1 alleles removed were referred to as ‘Ud-’ in the original papers to differentiate the strains deleted in only one endogenous allele (‘U-’). Since only the Ud set of mutants were used in this study, ‘Ud-’ is omitted from the names of these strains.

The differential response of each 14-3-3 mutant to distinct hyphal-inducing stimuli strongly suggests a role for Bmh1p in multiple hyphal signalling pathways. Since filamentation has been reported to play a key role in virulence (Lo et al., 1997), we have utilized the Bmh1p mutant strains to identify fungal responses that are required for virulence. We examined the pathogenesis of these mutants using an in vitro macrophage model and in a murine model of disseminated candidiasis. We related our findings to our previous in vitro phenotypic analysis and this revealed that filamentation pathways involved in embedment in matrix and chlamydospore differentiation are surprisingly not required for virulence in the disseminated mouse model.

METHODS

Cell cultures.

The murine macrophage-like cell lines J774A.1, RAW 264.7 and RAW 264.7 gamma NO⁻ (American Type Culture Collection nos TIB-67, TIB-71 and CRL-2278, respectively) were grown in Dulbecco's Modified Eagle Medium (DMEM) with 4.5 g d-glucose l⁻¹ supplemented with 10 % heat-inactivated fetal bovine serum (FBS), 4 mM l-glutamine, and 50 U ml⁻¹ penicillin G sodium and 50 μg ml⁻¹ streptomycin sulfate at 37 °C under 5 % CO₂ in 75 cm³ flasks. For simplicity, these cell lines are termed ‘macrophages’.

Strains and growth conditions.

The C. albicans strains used in this study are listed in Table 2. Strain BWP17 was provided by Dr A. Mitchell (Columbia University). The prototrophic derivative of BWP17, YJB6284, was provided by Dr J. Berman (University of Minnesota) (Bensen et al., 2002). In our studies, the YJB6284 strain will be designated ‘wild-type’ (WT). The strains constructed for this study were derivatives of BWP17 (Palmer et al., 2004; Palmer & Sturtevant, 2004). Strains were grown in yeast extract-peptone-dextrose (YPD) broth at 37 °C with shaking for 18 h.

Macrophage interactions and viability assays.

C. albicans strains were incubated with the murine macrophage line J774A.1 as previously reported (Palmer et al., 2005). Briefly, J774A.1 cells were seeded overnight in 12-well plates for a final count of 2×10⁵ cells per well onto 18 mm coverslips and incubated at 37 °C under 5 % CO₂. C. albicans strains were incubated with J774A.1 cells at an m.o.i. of 2 : 1 for 1.5, 5 and 24 h, and coverslips were observed using a light microscope. Macrophage viability was determined after 24 h, as previously described, using 1 μM calcein AM and 1.25 μM Calcofluor White (final concentrations) (Palmer et al., 2005). Macrophage survival was quantified by counting four fields per coverslip; three wells were counted per strain per experiment. The experiment was repeated at least three times per strain. For pre-germination experiments, C. albicans strains were suspended in DMEM supplemented with 10 % FBS and incubated at 37 °C until the length of the extension was equal to WT germ tube formation at 1 h. Macrophages were also incubated with heat-killed germ tubes and remained viable after incubation with all strains after 24 h (data not shown).

Attachment/ingestion.

J774A.1 cells were seeded overnight in 12-well plates at 2×10⁵ cells per well onto 18 mm coverslips and incubated at 37 °C under 5 % CO₂. For staining of C. albicans, strains were grown overnight at 30 °C, washed twice in PBS, resuspended at 10⁷ cells ml⁻¹ in PBS containing 1 μM Cell Trace Oregon Green 488 (Molecular Probes), and incubated at 30 °C for 1 h (Fernandez-Arenas et al., 2007). Staining efficiency was confirmed by microscopy. Cells were washed twice with PBS prior to incubation with J774A.1 cells at an m.o.i. of 2 : 1 in 12-well plates for 1.5 h in fresh media. Wells were washed twice with PBS, and fixed with 4 % paraformaldehyde in PBS for 10 min at 37 °C. External (non-ingested) C. albicans were stained with 1.25 μM Calcofluor White for 10 min at room temperature (RT) (Supplementary Fig. S1). Coverslips were washed three times with PBS and inverted onto slides with Fluoromount G mounting fluid (Southern Biotech). Micrographs were acquired using an Axio Observer Inverted High-End Microscope (Carl Zeiss) equipped with DAPI and FITC filters. Differential interference contrast (DIC) and false-coloured fluorescent images were merged using AxioVision Software (Zeiss). Two coverslips for each strain were counted and the number of ingested C. albicans (bright-green cells) versus attached only (blue-green cells) were calculated for 100 macrophages. When determining external cells, only those that were attached to a macrophage were included. Results are presented as the average number of C. albicans cells ingested or attached per macrophage.

C. albicans survival.

J774A.1 cells and C. albicans strains were co-incubated as described above. In order to avoid overgrowth of non-attached/ingested yeast, supernatants were removed after 1 h of incubation and replaced with fresh medium. After 24 h, samples were treated with 100 μl 1 % SDS (total concentration 0.05 %) for 10 min at RT, and the entire contents of each well were recovered by scraping the bottom of the well using a cell scraper. Samples were centrifuged and washed with sterile deionized water, and serial dilutions were plated onto YPD agar. Controls were wells with C. albicans but no macrophages. Viability was assessed by comparing each strain with its own control. The percentage growth inhibition was calculated by dividing the number of colonies in the presence of macrophages by the number of colonies in the absence of macrophages. Experiments were repeated at least three times per strain. Survival of C. albicans was also determined by the end point dilution assay at lower m.o.i. values, as previously described (Palmer et al., 2005).

Tumour necrosis factor-α (TNF-α) expression assay.

J774A.1 macrophages (5×10⁴ per well) were seeded into 96-well plates in a total volume of 150 μl medium per well. Overnight C. albicans cultures were washed twice in glucose–HEPES buffer (GH; 25 mM HEPES, 2 % glucose, 150 mM NaCl) and suspended in complete DMEM. Medium was then removed from macrophages by inverting and blotting plates onto clean paper towels. C. albicans strains were added to wells at a 2 : 1 effector : target (E : T) ratio in a total volume of 50 μl per well and incubated for 3 h. LPS (50 μl of a 10 ng ml⁻¹ stock solution) was added to macrophages as a positive control. TNF-α expression in supernatants was measured using the Quantikine mouse TNF-α ELISA kit according to the manufacturer's instructions (R & D Systems). Optical density was measured using a microplate reader. Concentrations of TNF-α were determined using the Logistic 5PL (Cook) Curve Fit of the standard curve. Experiments were repeated three times with n=3 per experiment.

Murine model of disseminated candidiasis.

Four to five-week-old female BALB/c mice were obtained from Charles River laboratories. Animals were housed under approved conditions at the vivarium at The Research Institute at Children's Hospital (New Orleans, LA). Age-matched mice were challenged with C. albicans WT, and bmh1 mutant strains intravenously via lateral tail vein injection with 5×10⁵ yeast cells per mouse in 100 μl 0.15 M sterile saline. Cohorts of five mice per C. albicans strain were inoculated per survival study. Survival studies were carried out for 30 days and mice were observed daily for signs of morbidity (weight loss, ruffled fur, hunched appearance, and decreased activity). Moribund animals were sacrificed via CO₂ fixation. For tissue c.f.u. and cytokine analysis, three mice were infected per C. albicans strain per experiment. Mice were sacrificed 48 h post-infection (p.i.) and organs were removed aseptically for subsequent analyses. Experiments were repeated twice. Survival curves and statistical differences between strains were analysed by Prism software using the logrank test.

Tissue cytokine analysis.

Spleens and left and right kidneys were removed from naïve and infected mice, and immediately submerged into 500 μl RNA Later (Ambion). Samples were stored at −20 °C. RNA was extracted from tissues using the RiboPure RNA isolation kit (Ambion) following the manufacturer's protocol. The RNA samples were stored at −80 °C. Primers and probes for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and TNF-α cDNA were constructed, based on those reported in the literature (Giulietti et al., 2001; Overbergh et al., 1999). Comparative quantitative reverse transcriptase PCR (QRT-PCR) was performed using the Mx3000P QPCR System (Stratagene). Fluorogenic probes were dual-labelled with the reporter dye 6-carboxyfluorescein (FAM) covalently attached at the 5′ end and the quencher dye Black Hole Quencher-1 (BHQ-1) covalently attached at the 3′ end (Integrated DNA Technologies). To determine optimal analysis conditions, a genomic murine kidney DNA sample was tested with multiple combinations of primers (100–300 nM) and probes (100–200 nM) using TaqMan Universal Master Mix, which contains the internal reference dye 6-carboxyl-X-rhodamine (ROX) (Applied Biosystems). The optimal primer/probe concentrations for GAPDH and TNF-α were 100 nM/100 nM and 300 nM/100 nM, respectively. Reaction mixtures for comparative real-time analysis were set up using 1 μl tissue cDNA, optimal primer/probe concentrations, and TaqMan mastermix in a total volume of 25 μl per well in MicroAmp optical 96-well plates (Applied Biosystems). The concentration of each template cDNA was normalized against GAPDH, and the relative expression compared with naïve samples was calculated using the ΔΔCt method (Livak & Schmittgen, 2001). These experiments were repeated twice, with n=6 per experiment. Statistical differences between strains were measured using the unpaired Student's t test. P<0.05 was considered significant.

Histology.

Naïve and infected mice were sacrificed 48 h p.i. Spleens and left and right kidneys were weighed and tissues sliced longitudinally using a clean razor blade. Half the tissue was processed for c.f.u. and the other half placed in 10 % buffered formalin (3.7 % formaldehyde in PBS) for subsequent histological analysis. Organs were also collected during a survival study. Mice that were moribund were sacrificed and their kidneys treated as described above. Tissue samples were processed and embedded in paraffin, sectioned in 10 μm slices, and stained with haematoxylin and eosin or Grocott methenamine silver (AML Laboratories). Stained sections were analysed by a pathologist, Dr Timothy Morgan, DVM, Louisiana State University School of Veterinary Medicine. Changes in the tubular and interstitial regions in each kidney section were graded in a numeric form from 1 to 4. Inflammatory changes and distribution were also noted.

Statistical analysis.

Unless otherwise noted, an unpaired, two-tailed Student's t test was used to establish significant differences in experiments. A P value of 0.05 or less was considered significant.

RESULTS

C. albicans bmh1 mutants show attenuated killing of J774A.1 cells

In order to determine whether Bmh1p-regulated pathways play a role in C. albicans host cell interactions, we first studied the initial interactions between murine macrophages and a panel of C. albicans bmh1 mutants by light microscopy. The WT strain YJB6284 was readily ingested, germinated in the macrophage and within 5 h escaped from the host cell [Fig. 1a (i, ii)]. The isogenic control and three bmh1 mutant strains (R142C, M125R and K51R) were similar to the WT (results not shown). The ligand-binding domain mutants L231S and K51E similarly associated with the macrophages. Although L231S and K51E germinated and extended the macrophage membrane by 5 h, it was to a much lesser degree than for the control strains [Fig. 1a (v, vi)]. Viable J774A.1 cells were quantified using calcein AM after 24 h of co-culture [Fig. 1a (iv, viii)]. At this time point, macrophages were seldom visible in wells incubated with the WT, the isogenic control, and K51R, M125R and R142C mutant strains [Fig. 1a (iii, iv), and data not shown]. Macrophages showed significantly higher viability counts when incubated with the bmh1 ligand-binding domain mutants L231S and K51E compared with the isogenic control and other bmh1 mutants (P<0.001) [Fig. 1a (vii, viii), b, and data not shown]. Interestingly, an intermediate number of macrophages remained viable after co-culture with M125R, although this number was not statistically significantly different from the isogenic control.

Fig. 1.

Altered interactions between bmh1 mutant strains and cultured murine macrophages. (a) Candida was incubated with J774A.1 cells. WT (i–iv) and K51E (v–viii) were incubated for 1.5 h (i, v), 5 h (ii, vi) or 24 h (iii, iv, vii, viii). To assess macrophage viability, viable cells were stained with calcein AM (green) and Candida was counterstained with Calcofluor White (blue) (iv, viii). (b) Macrophage viability after 24 h co-incubation with bmh1 mutant yeast cells was assessed as described in Methods. Data are presented as the mean±sd number of viable macrophages per field. *Significantly different (P<0.0001) from isogenic control. The difference between M125R and BMH is not significant (P<0.06). (c) Macrophage viability after 24 h co-incubation with pre-germinated isogenic control (BMH) and the K51 bmh1 mutants was assessed as described in Methods. The experiment was repeated three times and the results combined. Data are presented as the mean±sem number of viable macrophages per field. Macrophage survival after incubation with K51E is significantly different from that for BMH (P≤0.002). (d) BMH and pre-germinated K51E cells were co-incubated with J774A.1 cells for 24 h and viable macrophages stained as in (a).

Previous phenotypic analyses revealed that both L231S and K51E demonstrate a lag in germination in liquid media. In order to test the significance of this lag with respect to macrophage viability, the mutants were pre-germinated prior to co-incubation with J774A.1 cells and subsequent viability assessment. In a representative experiment we compared strains BMH, K51R and K51E (Fig. 1c, d). Pre-germinated K51E readily associated with the macrophages, but after 24 h of co-culture, J774A.1 cells remained viable. Therefore, pre-germination of K51E was not sufficient to restore macrophage killing.

Initially, J774A.1 cells were chosen as a representative phagocyte due to the simplicity and reproducibility of an established model system. However, these cells are reported to lack the mannose receptor and have a weak oxidative burst response (Stahl & Gordon, 1982). Experiments were repeated in mannose receptor-positive murine macrophage-like cell lines: RAW 264.7 and the nitric oxide-deficient RAW 264.7 gamma cells. Similar to J774A.1 cells, survival of both RAW 264.7 cell lines was significantly higher after incubation with K51E and L231S compared with the control and the other bmh1 mutants (data not shown).

bmh1 mutants are unaffected in attachment to and ingestion by J774A.1 cells

Attachment and ingestion were assessed to further characterize the C. albicans–host cell interaction. C. albicans strains were pre-labelled with Oregon green (Fernandez-Arenas et al., 2007). After co-incubation with the macrophages, C. albicans were counterstained with Calcofluor White, which cannot permeate the macrophage. Under this protocol, ingested C. albicans will appear green, while external (non-ingested) C. albicans will be dually stained and fluoresce blue-green (Supplementary Fig. S1). After 1.5 h incubation, almost all C. albicans cells were ingested; there were very few fields with external C. albicans. There was no significant difference in attachment and ingestion of any of the strains. A comparison between WT and K51E is shown in Fig. 2(a). The experiment was repeated at least three times for each strain, and the average number of C. albicans cells ingested per macrophage was consistently between 1 and 3, while those attached ranged from 0.2 to 0.5 (Fig. 2b, and data not shown).

Fig. 2.

Attachment to and ingestion by J774A.1 cells of bmh1 mutants is similar to that of the WT. (a) Candida was pre-labelled with Oregon green, incubated for 1.5 h under cell culture conditions, fixed, and counterstained with Calcofluor White. Ingested Candida are stained green, while external Candida appear blue-green. A representative micrograph of external labelled Candida can be seen in Supplementary Fig. S1. (b) The numbers of attached and ingested Candida were assessed per 100 macrophages in two separate wells. A representative graph is depicted.

bmh1 mutants survival is similar to that of the WT following J774A.1 ingestion

The increased macrophage survival seen with the K51E and L231S strains may be because the mutants are more susceptible to the candidacidal activity of the phagocyte. Due to clumping of hyphae after 24 h it was difficult to accurately quantify survival. Therefore, we performed two different assays and obtained similar results. C. albicans survival in the presence of macrophages was assessed by c.f.u. after 24 h. Survival percentages ranged between 36.6 and 48.7 % for all strains at an m.o.i. of 2 (data not shown). C. albicans survival with the RAW cell lines was also assessed using the c.f.u. method. In agreement with the results from the J774A.1 cells, survival of K51E and L231S mutant cells was similar to that of cells of the control strains (data not shown). It is possible that an E : T ratio of 2 : 1 overshadows differences between strains. To determine whether reducing the E : T ratio would result in different survival rates between strains, we used an end point dilution assay (Marcil et al., 2002; Palmer et al., 2005). At E : T ratios between 2 : 1 (∼50 %) and 0.125 : 1 (∼30 %), all strains survived equally well in the presence of the phagocytes. Thus, while the K51E and L231S mutants are inefficient at killing J774A.1 cells, they are able to survive within them following ingestion.

bmh1 mutants induce an altered cytokine profile in murine phagocyte cell lines

C. albicans strains were incubated with J774A.1 (Fig. 3a) and RAW 264.7 (Fig. 3b) cells for 3 h at an E : T ratio of 2 : 1, and TNF-α expression was measured in the supernatant by ELISA. At this time point, all C. albicans cells had been ingested, but viability of macrophages (as determined by calcein AM) had not yet been affected. A significantly higher expression of TNF-α was seen in strains K51E and L231S relative to the other bmh1 mutants, the WT and the isogenic control (Fig. 3). Although RAW 264.7 cells secreted significantly higher levels of TNF-α than J774A.1 cells, the relative differences between bmh1 mutant strains were the same for both cell lines.

Fig. 3.

bmh1 mutants induce an altered cytokine response in murine macrophage cell lines. (a) J774A.1 cells were incubated for 3 h with WT, isogenic control and the bmh1 mutants, and supernatants were measured for expression using a TNF-α ELISA kit. LPS at 5 ng ml⁻¹ was used as a positive control. Results are presented as the mean±sd concentration in pg ml⁻¹. *Significantly different from WT and BMH; P<0.05 was considered significant. Experiments were set up in triplicate and repeated three times, with similar results. (b) RAW 264.7 cells were incubated with Candida strains as in (a). Supernatants were diluted 1 : 10 and evaluated for TNF-α expression.

bmh1 mutants are affected in virulence in a mouse model of disseminated disease

The bmh1 mutants were tested in a murine model of systemic candidiasis. Mice infected with the control strains had median survival rates of 6–7 days (Fig. 4). Interestingly, the mutant strain M125R displayed attenuated virulence, with a median survival of 24.5 days. In three separate experiments, 40–60 % of mice infected with M125R survived for the duration of the experiment. This was significantly different from both the control and avirulent strains. The bmh1 mutant K51R, which was more efficient at killing macrophages than M125R, was more virulent in vivo, having a median survival time of 12 days, which was significantly different from that for control strains, although to a much lesser degree than strains K51E, L231S and M125R (P=0.0075 versus P<0.0001). Mice infected with K51E and L231S showed 100 % survival after 30 days. As shown above, these two mutant strains also failed to kill macrophages and induced an increased TNF-α response. These results demonstrate that a subset of bmh1 mutants were less virulent in a disseminated mouse model. Since colonization is a critical step in establishing infection, c.f.u. of murine tissues was assessed at 48 h p.i. All strains were able to colonize the mice by 48 h (Table 3). However, infection of kidneys by L231S, K51E and M125R was significantly less than that by the WT, isogenic control or the other bmh1 mutants. There was no significant difference in colonization of any of the spleen samples.

Fig. 4.

A subset of bmh1 mutants are attenuated for virulence in the mouse model of systemic candidiasis. Mice were inoculated via tail-vein injection with 5×10⁵ yeast cells of the WT, BMH1 control and mutant strains, and their disease progression was assessed for 30 days. Survival curves and statistical differences between strains were analysed by Prism software using the logrank test. K51E and L231S are superimposable on the graph.

Table 3.

Tissue burden of murine tissues 48 h p.i

Strain	Median survival*	Right kidney [log10(c.f.u. g−1)], mean±sem	Left kidney [log10(c.f.u. g−1)], mean±sem	Spleen [log10(c.f.u. g−1)], mean±sem
WT	7	5.3±0.7	4.9±1.0	4.0±0.4
BH1P1	5.5	4.8±0.8	4.9±0.8	4.2±0.2
BMH	7	4.8±0.4	4.6±0.2	3.9±0.4
K51E	31	3.1±0.3†‡	2.9±0.6†‡	3.2±0.7
K51R	12	4.1±0.2	4.2±0.2	2.4±0.7
L231S	31	2.6±0.8†‡	2.1±1.3†‡	2.9±1.7
M125R	24.5	2.8±1.4†	2.7±1.2†‡	3.4±1.7
R142C	7	4.9±0.8	4.8±0.7	3.6±1.0

*Median survival of mice infected with control and mutant strains is shown for comparison. Results are presented as the mean±sem of two separate experiments with n=3 mice per experiment. Student's t test was used to determine differences between mean values.

†Significantly different from WT.

‡Significantly different from isogenic control (BMH). P<0.05 was considered significant.

Infection with attenuated bmh1 mutant strains results in a decreased proinflammatory response

To determine whether infection with Bmh1p mutants resulted in an altered proinflammatory response in the host, relative TNF-α expression levels in kidneys and spleens of infected mice were measured using QRT-PCR. At 48 h p.i., TNF-α expression was significantly less in the kidneys of mice infected with the attenuated mutants L231S, K51E and M125R (Fig. 5a). Results from both kidneys are pooled in Fig. 5(a). There was no significant variation in TNF-α expression in spleens of infected mice (Fig. 5b). To determine whether the immune cell infiltrate and pathology of infected tissues correlated with reduced virulence in K51E, L231S and M125R, histology was performed on kidney and spleen samples at 48 h. The control strain and a majority of Bmh1p mutants displayed inflammatory changes of a suppurative nature (+1 to +3), predominantly in the intra-tubular regions of the kidneys. Inflamed tubules appeared dilated and filled with large numbers of neutrophils and sloughed epithelial cells (see Supplementary Fig. S2). On the other hand, no inflammatory lesions were present in the kidneys or spleens of K51E-infected mice (Supplementary Fig. S2) and none, or mildly suppurative lesions (+1), were observed in mice infected with L231S. Kidneys from M125R-infected mice revealed similar kidney pathology (+1 to +2) and neutrophilic index (2) to the virulent strains. When mice infected with control strains were close to death (6–8 days p.i.), inflammation was still of the suppurative type, with moderate to severe neutrophil indices in the kidneys. This was often accompanied by large mats of fungi in the renal pelvis, tubular abscesses, interstitial fibrosis, and fibrin in the lymphatics. The hyperfilamentous mutant R142C caused similar pathology to the WT. In contrast, kidney sections from K51E-, L231S- and M125R-infected mice sacrificed at day 30 displayed no inflammatory lesions and had a neutrophil influx of 0.

Fig. 5.

Attenuated bmh1 mutants induce an altered cytokine response in vivo. Organs were removed from mice infected with Candida control and bmh1 mutant strains 48 h p.i. Expression of TNF-α was determined by QRT-PCR from kidney (a) and spleen (b). Gene expression was normalized against GAPDH expression levels, and the results expressed as fold induction compared with untreated controls. *Significantly different (P<0.05) from WT and BMH. All samples were analysed in triplicate.

DISCUSSION

Results with the bmh1 mutant strains in the disseminated mouse model have begun to delineate the pathways that are required for C. albicans survival and virulence. Multiple nutrient and/or environmental conditions differentially induce C. albicans filamentation pathways; this flexibility may be responsible and is likely required for its ability to adapt and survive over the course of a developing infection within the host (Giusani et al., 2002). Historically, the ability to filament has been reported to be required for virulence; however, the bmh1 mutant strains demonstrated varied defects in filamentation which did not fully correlate with pathogenesis. These findings emphasize the importance of the identification of the different cellular processes that result in hyphal formation. Therefore, activation of appropriate hyphal signalling pathways may be essential for different stages of infection in the disseminated murine model.

It has been suggested that C. albicans growth under the microaerophilic conditions of the matrix assay mimics growth in tissue (Giusani et al., 2002). We found, however, that the ability of C. albicans to filament under these conditions does not always correlate with virulence. The BMH1 heterozygotes, BH1P1 and BMH, are defective in filamentation in matrix, both in onset and colony morphology, yet are equally as virulent as the WT. Furthermore, K51R shows the slowest and most altered response to embedment in matrix, yet remains virulent. These results strongly suggest that the ability to filament is regulated by pathways that are condition-specific and that filamentation under all conditions does not necessarily correlate with virulence. Under chlamydospore-inducing conditions, the virulent strains BH1P1, BMH and K51R fail to filament, form suspensor cells or produce chlamydospores, while the less-virulent mutant M125R filaments, form suspensor cells and produces chlamydospores (although to a lesser extent than the WT). These results, taken together with the matrix assay findings, clearly indicate that these assays are not accurate predictors of C. albicans virulence in the disseminated mouse model.

A subset of bmh1 mutants demonstrated altered host–fungus interactions and virulence. The delay in growth of K51E and L231S most likely contributes to the reduced colonization by these strains of the mouse. Although these strains do not grow as vigorously as the control strains, it is important to note that they (1) do grow at 37 °C, (2) can survive within a hostile environment as well as control strains (macrophage cell lines, spleen), and (3) are not sensitive to nitrogen starvation (data not shown). M125R was also attenuated in virulence, although it exhibited a slightly longer lag phase than the WT (Table 1); this same delay in the onset of growth was also observed in the virulent strains BMH and K51R. Previously, the effect of temperature on growth was assessed by plating serial dilutions and incubating at 30 and 37 °C. Overnight cultures demonstrated healthy growth of all mutants, and colony size was larger at 37 °C than at 30 °C. However, liquid growth curves were performed only at 30 °C. Growth curves were repeated at 37 °C (see Supplementary Fig S3). All isogenic strains demonstrated similar log doubling times, and by 24 h attained equivalent cell densities. However, K51E, L231S and, to a lesser extent, M125R had extended lag periods compared with the other strains (2-, 1.7- and 1.4-fold longer than the WT, respectively). At 37 °C, BMH and K51R strains no longer demonstrated the extended lag periods seen at 30 °C. This implies that K51E, L231S and M125R have defects in the signalling or metabolic pathways that respond to environmental changes. This reduced adaptive capacity may contribute to their attenuated virulence. The identification of lag period differences among cells grown at 37 and 30 °C is a significant finding, as the vast majority of laboratories routinely characterize mutant strains at 30 °C. Additionally, caution should be taken when strains are tested in insect models with incubation temperatures other than 37 °C. Identification of the altered pathway(s) that cause this lag in growth will be instrumental in identifying factors that allow C. albicans to survive in the host.

As 14-3-3 proteins are multifunctional proteins with hundreds of binding partners, there is great potential for involvement in numerous cellular processes. K51E and L231S substitutions may affect multiple pathways, resulting in compound effects that contribute to their attenuated virulence. These mutants are not altered in their ability to express secreted aspartic proteases (SAPs) or in their SAP and phospholipase proteolytic activities in vitro (data not shown). Therefore, these characteristics are unlikely to contribute to the reduced virulence of K51E and L231S mutants. In contrast, the differential abilities of these mutants to filament under specific conditions may well play significant roles in reduced virulence. Under standard in vitro conditions, macrophages readily ingest C. albicans yeast cells, which then filament and extend the membranes of the host cell, ultimately resulting in lysis of the macrophage (Lo et al., 1997; Lorenz et al., 2004; Rocha et al., 2001). Membrane extension per se is not sufficient to cause macrophage death, since extended cells are viable [Fig. 1a (vii, viii)] (Marcil et al., 2002). It may be the degree to which the macrophage is extended that is critical. Pre-germinated K51E and L231S mutants continued to demonstrate attenuated killing. This implies that the onset of germination itself does not significantly contribute to macrophage death. However, it is possible that the slower rate of germination, along with the truncated hyphal extension of K51E and L231S, accounts for increased macrophage survival. Although the onset of germination does not contribute to macrophage viability, it could influence the response of the macrophage to the invading pathogen. It has been reported that TLR4 mediates proinflammatory cytokine induction after C. albicans stimulation, and that these signals are lost upon germination (van der Graaf et al., 2005). Therefore, the reduced germination response seen in K51E and L231S may induce the TLR4 (TNF-α) response within the macrophage and allow it to escape C. albicans-induced lysis. This may account for the more rapid clearance of these strains in the host. The lag in growth and hyphal extension of K51E and L231S strains may also impose less stress on the macrophage, which could result in increased macrophage survival and increased induction of TNF-α. In contrast to the in vitro results, expression of TNF-α was inversely correlated with protection during disseminated candidiasis. As others have reported, the significant proinflammatory cytokine response correlated with significantly higher tissue burden and high mortality rates (Mullik et al., 2004). In contrast, mice infected with attenuated K51E, L231S and M125R mutant strains had very low induction of TNF-α in the kidneys. This correlated with decreased tissue burden, inflammatory cell influx and tissue damage. The increased clearance of these strains could be due to inherent defects in L231S and K51E, or it could be due to an initial increase in TNF-α production as observed in vitro at 3 h. The earlier induction of TNF-α by the mutants would result in a more rapid initial clearance. By 48 h, the fungal burden would be lower and thus result in a decreased TNF-α response.

K51E, and to a lesser extent L231S, are affected in cell surface properties (J. E. Sturtevant, unpublished data). Altered cell surface would affect the cell's interaction with environmental cues, with correspondingly differential activation of intracellular signalling pathways. This could in turn account for the disregulation of filamentation and growth by C. albicans and differential recognition by host cells. Our results support the suggestion that Bmh1p alters both initiation of germination and hyphal extension rates. Defects in filamentation occur under distinct inducing conditions for each mutant. However, all mutants are competent to produce hyphal forms. These defects are in signal transduction response pathways rather than a structural impediment. This is not surprising given the roles of 14-3-3 in other organisms and its role in binding phosphoproteins. K51, L231 and M125 all lie within the binding pocket, and thus will be differentially affected in their ability to bind ligands.

During the interaction between C. albicans and the host, a restructuring of a wide range of metabolic and non-metabolic pathways occurs in both fungus and host (Fradin et al., 2003; Lorenz & Fink, 2002; Lorenz et al., 2004; Rubin-Bejerano et al., 2003). This is a dynamic relationship, in that C. albicans must not only first adapt to the different host environment but also then respond to challenge by the host anti-C. albicans response. This requires the co-regulation of multiple signalling pathways. While not all of these pathways may be required for C. albicans to survive and establish infection in the host, it is very likely that specific pathways are required during the different stages of infection as the fungus adapts to variations in both the environment and the host immune response. Indeed, we have shown that the in vitro ability to filament after embedment in matrix or produce chlamydospores does not correlate with the ability of C. albicans to successfully initiate invasion of the host. We have begun to dissect different pathways by using a panel of isogenic mutants of the multifunctional signalling modulatory protein Bmh1p. K51E, L231S and M125R mutant strains show an increased lag in the onset of exponential phase growth at 37 °C, which suggests an impaired ability to adapt and readjust to the environment. The pathway(s) that regulates this characteristic may play a pivotal role in virulence in the disseminated mouse model. These studies establish that the bmh1 mutants differentially regulate important cellular growth/survival pathways which affect interactions with the host.

Abbreviations

• GAPDH, glyceraldehyde-3-phosphate dehydrogenase

• p.i., post-infection

• QRT-PCR, quantitative reverse transcriptase PCR

• TNF, tumour necrosis factor

• WT, wild-type