Dynamic, Morphotype-Specific Candida albicans β-Glucan Exposure during Infection and Drug Treatment

Candida albicans, a clinically important dimorphic fungal pathogen that can evade immune attack by masking its cell wall β-glucan from immune recognition, mutes protective host responses mediated by the Dectin-1 β-glucan receptor on innate immune cells. Although the ability of C. albicans to switch between a yeast- or hyphal-form is a key virulence determinant, the role of each morphotype in β-glucan masking during infection and treatment has not been addressed. Here, we show that during infection of mice, the C. albicans β-glucan is masked initially but becomes exposed later in several organs. At all measured stages of infection, there is no difference in β-glucan exposure between yeast-form and hyphal cells. We have previously shown that sub-inhibitory doses of the anti-fungal drug caspofungin can expose β-glucan in vitro, suggesting that the drug may enhance immune activity during therapy. This report shows that caspofungin also mediates β-glucan unmasking in vivo. Surprisingly, caspofungin preferentially unmasks filamentous cells, as opposed to yeast form cells, both in vivo and in vitro. The fungicidal activity of caspofungin in vitro is also filament-biased, as corroborated using yeast-locked and hyphal-locked mutants. The uncloaking of filaments is not a general effect of anti-fungal drugs, as another anti-fungal agent does not have this effect. These results highlight the advantage of studying host–pathogen interaction in vivo and suggest new avenues for drug development.


Introduction
Candida albicans, a common commensal organism of humans, has emerged as an important fungal pathogen clinically due to the immunocompromised status of many patients as well as the ineffectiveness of current anti-fungal drugs [1]. The immune system has many pathways for recognizing and responding to C. albicans, including innate immune Toll-like receptors, lectin receptors, antibody, complement, and mannose binding lectin [2]. A major innate immune receptor for C. albicans is Dectin-1, a lectin that can recognize b-glucan, a unique component of the fungal cell wall. In vitro analysis shows that Dectin-1 recognizes fungal b-glucan (comprising glucan polymers with mixed b1,3and b1,6linkages), and signals through unique pathways to induce phagocytosis, up-regulation of immune killing mechanisms, and production of pro-inflammatory cytokines [3].
The mechanism of b-glucan signaling through Dectin-1 is unclear because this signature molecule on Candida and other fungi is enveloped by a cell wall mannoprotein layer that masks almost all of the b-glucan from immune recognition and mutes the host immune response [4,5,6,7,8]. On one hand, the b-glucan masking observed in vitro runs contrary to an active role for Dectin-1 recognition of b-glucan during infection [8,9,10]. On the other hand, in vitro Dectin-1 is clearly able to mediate protective responses to fungi and zymosan, a treated fungal particle with exposed b-glucan [9,11,12,13,14]. In addition, work in knock-out mice suggests at least a conditional requirement for Dectin-1 in resistance to the SC5314 strain of C. albicans, which may be dependent on mouse strain and/or C. albicans strain [13,14].
An additional complexity at the Candida-host interface is that C. albicans has several developmental cell types including yeast, pseudohyphae, and hyphae, each with different modes of interaction with innate immune cells [15,16]. The ability of C. albicans to switch between the yeast and filamentous forms is strongly associated with virulence. Filaments are distinct from yeast-form cells in cell wall structure, cell wall proteins and transcriptional programs. In addition, the hyphal form of the fungus has been shown to cause more tissue damage than the yeast-form fungus in ex vivo models of candidiasis [17,18,19]. Furthermore, immune recognition of yeast provokes a different immune response compared to recognition of hyphae [15,20]. Intriguing recent work has suggested that only yeast-form cells possess exposed b-glucan at a few sites [9], but this has yet to be shown in vivo.
Although b-glucan can be an important signal for innate immune cells in vitro, and the b-glucan receptor Dectin-1 protects against fungal disease in mice, we still understand little of the dynamics of b-glucan-Dectin-1 interaction during infection or drug treatment. We have developed a new technique to directly measure morphotype-specific b-glucan exposure in mouse tissue during infection and anti-fungal treatment. We find unexpected dynamics of b-glucan exposure during infection and during treatment with the cell wall-directed drug caspofungin. Caspofungin treatment, either in vitro or in vivo, causes dramatic, hyphalbiased exposure of b-glucan. By contrast, although in the absence of the drug there is progressive unmasking of b-glucan during infection, there is no difference between yeast form and hyphal cells. Our results support a role for Dectin-1 in recognition of C. albicans during infection and drug therapy and demonstrate a novel morphotype bias in the action of the anti-fungal caspofungin, highlighting new avenues for drug development.

Drug-induced b-glucan unmasking and fungicidal activities are filament biased
The anti-fungal drug caspofungin (CF) unmasks b-glucan at sub-inhibitory concentrations during growth of Candida albicans in culture media that promotes either exclusively yeast-form or hyphal-form growth [8]. During infection, C. albicans grows in conditions that promote both yeast and hyphal forms [21]. To study conditions more similar to in vivo growth, we grew C. albicans in sub-inhibitory concentrations of CF in RPMI media at a temperature that concurrently permit yeast-, pseudohyphal-, and hyphal-form growth. Growth at K the minimum inhibitory concentration (MIC) of CF causes striking b-glucan unmasking, as assayed with anti-b-glucan antibody specific for b1,3-glucan ( Figure 1A and 1B).
Consistent with our previous work [8], there is very little cell death during growth in K MIC of CF, as assayed by propidium iodide staining. This result was further confirmed by using the WT-GFP strain, engineered to constitutively express yEGFP in the SC5314 clinical isolate. WT-GFP fluoresces at similar intensity to CAI4-GFP [22] and grows at wildtype rates (data not shown). The cytoplasmic GFP fluorescence enables the identification of live cells, because live cells have a characteristic cytoplasmic pattern of green fluorescence (see Methods).
Unexpectedly, there is a striking bias in exposure of filaments as opposed to yeast-form cells-a high percentage of hyphal cells have exposed b-glucan, whereas only a few yeast-form cells have exposed b-glucan ( Figure 1E). The filament-specific exposure of bglucan was also found in several clinical isolates, suggesting this is a general feature of caspofungin ( Figure S1). To ensure that this differential effect on filaments was not due to persistence of the initial inoculum of yeast-form cells (which were pre-grown in the absence of CF), we followed the persistence of these original cells during CF treatment by marking them with the amine-reactive dye Alexa fluor 647-succinimidyl ester. As shown in Figure 1C, treatment of live cells with the dye homogeneously labels all cells in the inoculum. The label is retained on the surface of the original dye-tagged cells even when they are grown in hyphal-promoting conditions (37uC with serum), which results in no loss of label intensity on the mother cell and no transferral of label to growing hyphae in those cells ( Figure 1D). This suggests that the cell surface proteins labeled by the amine-reactive Alexa fluor 647 dye are tightly retained in the cell wall and provide a long-lasting, nondispersing mark for following the ancestry of C. albicans cells. We then used this to ask whether the yeast cells with no exposure are from the pre-culture or grew in the presence of caspofungin. As shown in Figure 1E, the majority of yeast-form cells in overnight cultures have no Alexa label and therefore were newly formed in the presence of CF. Thus, growth in pseudohyphal or hyphal form, as compared to yeast-form, predisposes C. albicans to bglucan unmasking by CF.
The above data suggest that CF specifically unmasks hyphal forms. To further test this we utilized C. albicans mutants that grow exclusively in hyphal-or yeast-form, independent of growth media. Due to the possibility that any given mutant might show altered sensitivity unrelated to its morphotypic defect, we chose a variety of strains. Specifically, we analyzed two mutants locked in filamentous growth form (tup1D/D [23] and nrg1D/D [24]), two mutants locked in yeast-form growth (edt1D/D and efg1D/D cph1D/D [25]), and several wildtype strains (SC5314, CAF2, and WT-GFP). The EDT1 gene, originally isolated based on its requirement for filamentation, results in constitutive yeast-form growth when deleted (Chen et al, unpubl. results). When these morphotype specific mutants grown at sub-inhibitory doses of CF are stained for b-glucan exposure, filament-locked mutants such as nrg1D/D have exposure similar to wildtype filamentous cells but yeast-locked cells such as edt1D/D show little exposure ( Figure 1F-1H). The finding that yeast-locked cells show no b-glucan unmasking at sub-inhibitory CF concentrations supports the conclusion that morphogenesis and sensitivity to unmasking are linked.
The filament bias of b-glucan exposure by CF raised the possibility that the drug's fungicidal activity may also act preferentially on filamentous cells. To test this, we probed the efficacy of CF against Candida mutants locked in either the yeast or filamentous form under conditions that normally permit growth of both filaments and yeast-form cells. Compared to wild type strains, mutant filamentous strains consistently show greater drug sensitivity and yeast-form mutants show greater resistance ( Figure 2). Because filamentation confers greater sensitivity to the fungicidal activity of CF, this suggests that both the b-glucan unmasking activity and fungicidal activity of CF preferentially affect filamentous cells in vitro.
Direct measurement of baseline b-glucan exposure during infection with ex vivo immunofluorescence Our in vitro results point to hyphal-biased activity of CF, but their relevance for understanding the drug's activity during therapy was unclear. We therefore developed a method that

Author Summary
Candida is a common human commensal but disseminated candidiasis is a serious clinical problem, especially among immunocompromised patients. The innate immune system controls Candida infection, in part through the germline-encoded b-glucan receptor Dectin-1. However, during in vitro growth, Candida albicans mutes Dectin-1 recognition by cloaking its b-glucan underneath a layer of mannan. Bridging these two seemingly contradictory observations, we demonstrate that C. albicans masks b-glucan early during infection, but it becomes exposed later, allowing Dectin-1 to recognize the fungi and mediate immunity. Remarkably, treatment of mice with subtherapeutic doses of the antifungal drug caspofungin causes exposure of b-glucan on C. albicans even when it would not be exposed naturally. We introduce a new technique for monitoring of epitope exposure during infection, which can be used to monitor the availability of any epitope for immune recognition. This technique allowed us to show that natural unmasking of b-glucan is not morphotype-specific, but drug-mediated unmasking is biased towards the invasive filamentous form of C. albicans. These results highlight the unexplored area of dynamic epitope exposure during infection and therapy, which might be targetable to enhance immune recognition and fungal clearance.
permitted us to isolate and assay the surfaces of live fungi for bglucan exposure directly from infected tissue without prior fixation or sectioning. The fixation and sectioning procedures normally used to assay cell surface molecules could be misleading as fixation can itself alter cell surface structure and exposure while sectioning artificially exposes underlying molecules. In this technique, called ex vivo fluorescence (EVF), tissue homogenates obtained from infected animals are washed and whole cells are stained directly without fixation in order to preserve the live cell wall organization. Use of WT-GFP C. albicans permitted both the unambiguous detection of fungi amidst non-fluorescent kidney tissue and the identification of live fungi based on characteristic fluorescence pattern (see Methods).
To provide a baseline against which to assay the affects of CF, the course of b-glucan exposure during disseminated infection in the absence of CF was measured in BALB/c mice infected intravenously with WT-GFP Candida. Fungal cells were subsequently identified in their kidney homogenates stained by EVF using a commercially available antibody directed against b-glucan. As shown in Figure 3A, EVF-stained C. albicans isolated from the kidney within 16 hours post-infection did not have appreciable bglucan exposure ( Figure 3A). Similar data were found in heart and brain homogenates (data not shown). The masking of b-glucan on Candida at early times post-infection in all cell types suggests that Dectin-1 plays a minimal role in recognition at this time during infection. Dectin-1 is reported to have a role in clearance of fungi late in infection [14], so b-glucan recognition by this receptor might not normally play a role until later during infection. To discover whether b-glucan is unmasked later during infection, we infected mice intravenously with lower doses of WT-GFP C. albicans, euthanized animals at 1, 3, 5, and 7 days post-infection, and examined b-glucan exposure by EVF. Remarkably, by five days post-infection a large percentage of cells had exposed b-glucan ( Figure 3B). The anti-b-glucan antibody is a mouse monoclonal IgG antibody; however, the control without primary antibody ( Figure 3C) was negative, showing that no detectable host IgG was bound to the fungus during growth in kidney tissue. In addition to this control, the robustness of our staining protocol was tested using different incubation times, omitting the blocking reagent (bovine serum albumin), and testing two different isotype control antibodies to ensure this assay does not introduce artifactual staining ( Figures S2, S3, and S4). To quantify the progression in bglucan exposure, cells were identified and cell wall-associated bglucan fluorescence was quantified as shown in Figure 3D using the CellProfiler software package [26]. As shown in Figure 3E, the percentage of cells with b-glucan exposure increases from approximately 20% to 80% over the time course, demonstrating time-dependent unmasking of b-glucan during infection.
To assess the cell-type specificity of natural b-glucan unmasking in vivo, we compared exposure between the yeast-form and hyphalform cell wall at early and late points during infection. As shown in Figure 3F, there was no overall difference in unmasking between cell forms. Overall, there is very little b-glucan available for immune recognition in either cell type early during infection but there is significant exposure of both cell types as infection progresses. These data demonstrate that, during infection, C. albicans b-glucan is initially masked from immune recognition, but, over time, becomes unmasked in a morphotype-independent fashion.

Caspofungin treatment during C. albicans systemic infection causes b-glucan unmasking
To determine if treatment of disseminated candidiasis with caspofungin (CF) could artificially increase b-glucan exposure, we used different levels of CF to identify a sub-therapeutic dose (30 mg/kg) that would allow isolation of live fungi that had grown in the presence of drug ( Figure 4A). These concentrations are in line with those observed to be sub-inhibitory by other investigators (C. Douglas, pers. comm.) We then infected mice with WT-GFP Candida, isolated kidneys at 16 hr post-infection when all bglucan should be masked, and stained fungi with EVF to look at bglucan exposure. Fungi isolated from mice treated with intermediate doses of CF had very high levels of b-glucan exposure ( Figure 4D), in contrast to C. albicans isolated from mice treated with vehicle ( Figure 4B). The no-primary-antibody controls ( Figure 4C and 4E) show that this increase is not due to bound host IgG.
To test if this increase in b-glucan exposure has biological relevance for immune recognition, we probed with the purified carbohydrate recognition domain of the b-glucan receptor Dectin-1 (Dectin1-CRD). As shown in Figure 5, the increased b-glucan exposure is accompanied by increased availability of b-glucan for immune recognition ( Figure 5). In each of the other types of experiments presented, Dectin-CRD binding was assayed and found to mirror antibody binding (data not shown). Thus, treatment with CF results in exposure of the pro-inflammatory b-glucan epitope and increased binding to the innate immune receptor Dectin-1.
To determine whether b-glucan unmasking is a consequence of treatment with another known anti-fungal compound, we treated WT-GFP-infected mice with sub-therapeutic doses of either fluconazole, an anti-fungal drug that targets sterol biosynthesis, or CF, and then examined b-glucan exposure 16 hours postinfection ( Figure 6). In this experiment, both drug treatments are active and result in 10-fold decreased fungal burden ( Figure 6A) but only CF treatment causes b-glucan exposure ( Figure 6B-6D). Thus, b-glucan unmasking is specifically associated with CF treatment and is not a default stage in killing of C. albicans.
b-glucan exposure in vivo preferentially affects the filamentous cells The potent b-glucan unmasking activity of caspofungin in vivo provided a foundation for looking at the morphotype dependence of b-glucan unmasking during treatment of disseminated infection. Examination of individual cells for b-glucan exposure due to CF showed that filament cell walls are exposed, but yeast cell walls are not ( Figure 7A). Quantification with CellProfiler demonstrates the robustness of this phenomenon, in which filamentous cells stain with anti-b-glucan whereas yeast-form cells show undetectable levels of b-glucan exposure ( Figure 7B). For Candida isolated from vehicle-treated mice, neither morphotype showed detectable bglucan exposure at these times after infection. Dectin-CRD binding closely paralleled anti-b-glucan staining for both caspofungin and vehicle treatment (data not shown).
To ensure that this differential effect on filaments was not due to persistence of the initial inoculum of yeast-form cells, we Alexa fluor-labeled fungi pre-infection and followed the appearance of unlabeled yeast-form cells (Figure 7C and 7D). As shown above in Figure 1, this covalent labeling technique tags cell surfaces with a long-lasting, non-dispersing mark for following the ancestry of cells. By three hours post-infection, there was already significant appearance of unlabeled yeast-form cells, which grew exclusively post-labeling and therefore post-infection. By six hours, there were

Discussion
A critical goal in understanding both host-pathogen interaction and drug action is to probe their dynamics during infection. In vivo experiments show that cell wall b-glucan is a key fungal signature molecule used by the innate immune system to clear fungal infection, yet in vitro experiments suggest that Candida albicans masks b-glucan from immune recognition. To resolve this apparent paradox, we developed and exploited an ex vivo fluorescence (EVF) technology to measure the changes in exposure of b-glucan in live cells during infection. This procedure showed that in the normal course of infection b-glucan is initially masked but subsequently exposed on the surface of C. albicans. This reconciles in vitro data that C. albicans b-glucan is masked by mannoprotein [8,9] with in vivo infection data suggesting a potential role for Dectin-1 in immunity to C. albicans. In vivo data with knockout mice is inconsistent-clearance of the SC5314 strain of C. albicans, used in this study, from mice in the mixed BL/66129/Sv background, is dependent on Dectin-1 [14], whereas clearance of a different strain of C. albicans from mice in the BL/6 background does not require Dectin-1 for efficient clearance [13]. Taken together, these data suggest that the Dectin-1 b-glucan receptor plays a role in resistance to C. albicans in conditions most similar to those used in our study [14]. Although it is not known how this unmasking occurs during the course of the infection, it is possible that, with time, immune cells accumulate in sufficient numbers to directly damage the cell wall and expose b-glucan. Infiltrating immune cells may also create an environment that induces physiological defenses in the fungus, resulting in altered cell wall signaling and increased exposure of b-glucan. Innate immune cells with the capacity to damage Candida are recruited to the site of infection in the first several days of infection, favoring these models [21,27,28]. Alternatively, there may be a novel pathogen-intrinsic pathway late in infection which is programmed to alter architectural changes independent of immune function.
EVF also permits unambiguous identification of cellular morphotype, allowing a rigorous analysis of b-glucan exposure in yeast vs. hyphal cells. In the absence of drug there is no difference in exposure of b-glucan between yeast-form and hyphal cells: both are effectively masked from the immune system early during infection and both are unmasked later. This provides evidence that during systemic infection of tissues, filamentous growth does not allow the fungus to evade recognition by the Dectin-1 b-glucan receptor. Previous work with in vitro-grown C. albicans using different media to promote morphotype-specific growth had suggested that hyphal cells are uniquely immune from Dectin-1-mediated recognition [9]. However, our in vivo results show clearly that during infection there is no morphotype-specific masking or unmasking of b-glucan. This contrast highlights the critical advantage of looking at the host-pathogen interaction during infection rather than in vitro.
Treatment of mice with CF during disseminated infection unmasks C. albicans b-glucan and makes it available for immune recognition by the Dectin-1 b-glucan receptor. Although unmasking during the normal course of infection shows no morphotype specificity, CF-mediated unmasking is filament-biased and both bglucan unmasking and fungicidal activities of CF are directed more at filaments than yeast-form cells. The bias may be due to intrinsic compositional differences between hyphal and yeast wall [29,30,31,32] or due to differences in transcription of cell wall regulators sensitive to caspofungin [33]. Filament bias may contribute to the clinical efficacy of CF by selecting for yeastform growth which is less invasive, more easily cleared and stimulates a more protective immune response [15,20]. The finding that a particularly effective anti-fungal drug has morphotype bias opens the possibility of new anti-fungals designed for increased effectiveness against the filamentous, invasive form of C. albicans. The morphotype bias also highlights the potential shortcomings of using a single growth condition to determine in vitro drug effectiveness, which can favor one morphotype over another and fail to reveal such biases.     Our findings draw attention to a relatively unexplored avenue for drug development-the identification of drugs that do not affect the growth or viability of the pathogen per se but render the pathogen more susceptible to the host immune system. It is likely that clinical use of CF also causes increased recognition through the b-glucan receptor, making hyphal cells more susceptible to immune attack and contributing to the success of the drug. Recent work shows that caspofungin causes unmasking in other fungal species and makes them more susceptible to immune attack, underlining the broad-spectrum robustness of this drug's activity to expose b-glucan even in the absence of classical fungicidal activity [4,6]. Development of new anti-fungal drugs that specifically cause b-glucan unmasking may reveal novel classes of drugs that target a non-essential process, provide broad-spectrum clinical protection, and do not select for resistance in the absence of immune pressure.

C. albicans strains and growth conditions
C. albicans strain information is summarized in Table 1. C. albicans was routinely grown at 37uC in YPD media (DIFCO; 20 g/L peptone, 10 g/L yeast extract) containing 2% glucose. For infections, C. albicans was washed, counted by hemacytometer, and diluted to 1*10 7 /ml. For MIC experiments, C. albicans was pregrown in YPD at 30uC overnight, washed, sonicated briefly with a Branson sonicator (3610 seconds at setting of 1.5), and diluted 1:100 to a concentration of 2.5*10 6 /ml (for WT and yeast-locked strains) or 1:20 (for hyphal-locked strains that grow poorly) into RPMI-PS, RPMI (Invitrogen) containing phenol red, sodium pyruvate, HEPES pH 7.2, with added penicillin-streptomycin. Caspofungin (Cancidas intravenous formulation, Merck & Co., Inc.) was made fresh from lyophilized stock and diluted into RPMI-PS for in vitro studies or into PBS for intraperitoneal (i.p.) injection into mice (see below).

Alexa fluor labeling of live C. albicans
For Alexa fluor 647 labeling of C. albicans, cells pre-grown at 37uC overnight in YPD were washed three times in PBS and resuspended in PBS at a concentration of 1*10 9 /ml in 100 ul. Alexa fluor 647-succinimidyl ester (Molecular Probes) was dissolved and stored in DMSO at a concentration of 10 mg/ml. Just prior to labeling, 11 ml of 1 M NaCO 3 pH 10 was added to the Candida and then 2 ml of Alexa fluor 647 was added and mixed briefly by vortexing. Cells were labeled for 1 hour at room temp, then washed five times with PBS, counted by hemacytometer, and used for infection or in vitro experiments. This resulted in no loss of viability, as assayed by broth dilution and plating to YPD.

Ex vivo fluorescent staining (EVF)
Dounce-homogenized kidney tissue was washed three times with PBS, then mammalian cells were lysed with two washes of distilled water. For washes, cells were spun at 14,0006g for 30 seconds and pellets were vigorously resuspended with vortexing. Homogenates were washed again three times with PBS and then blocked with 2%BSA in PBS for 1 hour at room temp. Homogenates were then stained with anti-b-glucan antibody, specific for b1,3-glucan (Biosupplies, Inc., Australia) diluted 1:200 in 2%BSA/PBS overnight at 4uC on a rotator and washed three times with PBS. Then homogenates were stained for 1 hour at room temp with goat anti-mouse Cy3 antibody (Jackson Immunoresearch) diluted 1:100 into 2%BSA/PBS with or without Alexa647-labeled Dectin-CRD (diluted 1:40 in 2%BSA/PBS). The carbohydrate recognition domain of mouse Dectin-1 (Dectin-CRD) was purified from E. coli and labeled as described previously [8]. After three washes, cells were visualized by fluorescence microscopy. Live cells were identified based on characteristic GFP fluorescence, with cytoplasmic fluorescence excluded only from vacuoles. Co-staining with propidium iodide shows that the small proportion of PI-positive cells (,1%) do not have characteristic GFP fluorescence, and vice versa (data not shown).

Plasmid and C. albicans strain construction
Construction of the pENO1-yEGFP3-NAT plasmid was performed as follows. pENO1GFP3 [22] was digested with KpnI. The ENO1 locus was amplified from SC5314 genomic DNA with forward primer (aattggccctgcagatgtcttacgccactaaaatccac) and reverse primer (aattggccggtaccccagcgtagatagcttcagaacct), which introduce KpnI and PstI sites, respectively. The NAT1 gene was digested from pJK795 [34] with KpnI and PstI. A triple ligation of these two fragments with the KpnI-digested pENO1GFP3 resulted in pENO1-yEGFP3-NAT. EcoRV-digested plasmid was transformed into strain SC5314 by lithium acetate. Integrants were selected on 100 mg/ml nourseothricin (Werner Bioagents, Jena, Germany) and verified by PCR. This resulted in the WT-GFP strain, and transformation of the homozygous deletion strain edt1::hisG/edt1::hisG-URA3-hisG (a derivative of CAI4) yielded edt1-GFP.

Fluorescence microscopy and Cellprofiler quantification
Samples were visualized with a Nikon TE2000-S microscope (Nikon, Tokyo, Japan) equipped with Spot RT camera (Diagnostic Instruments, Sterling Heights, MI, USA) and processed in Photoshop (Adobe Systems, Palo Alto, CA, USA). In Cellprofiler version 1.0.4884 (www.cellprofiler.org) 12-bit TIFF images of GFP fluorescence were used to manually define the edges of individual cell segments and 12-bit b-glucan fluorescence images without saturate pixels were used to quantify the average fluorescence intensity for the whole cell segment. Yeast form cells had a length-width ratio of ,1.5 and characteristic shape and size. Average fluorescence values were analyzed using JMP 5 statistical software package. For fluorescence microscopy in supplementary figures, cells were imaged on an Olympus IX81 inverted microscope with 406 PlanApo objective using an Orca-ER cooled CCD camera (Hamamatsu Photonics) and using the IPLab v. 4.04 software package. Figure S1 Several clinical isolates also display filament-specific exposure of b-glucan. Wildtype SC5314 (A-B) and clinical isolates CAN89 (C-D), CAN91 (E-F) and CAN92 (G-H) were grown overnight in YPD, washed, and diluted to 4*10 6 cells/ml in fresh RPMI-PS with different concentrations of caspofungin or vehicle. After growth overnight at 30uC, cells grown in vehicle (B, D, F, and H) and at 1/2 MIC 50 caspofungin (A, C, E, and G) were harvested and stained with anti-b-glucan antibody and Cy3labeled secondary antibody. Cells were imaged by fluorescence microscopy and representative images were cropped and processed in Photoshop. Scale bar in (A) is 10 microns long and applies to all images. Found at: doi:10.1371/journal.ppat.1000227.s001 (2.48 MB TIF) Figure S2 Altering incubation times during staining procedure does not alter b-glucan exposure. WT-GFP C. albicans was pregrown overnight in YPD, then diluted to 4*10 6 cells/ml and grown overnight in RPMI-PS media at 30uC with either subinhibitory doses of caspofungin (Caspofungin; A-B, E-F, and I-J) or vehicle (Vehicle; C-D, G-H, and K-L), then washed with PBS. Cells were stained with normal length incubations (1 hour block, overnight primary antibody, 1 hour secondary antibody; A-D), short incubations (15 minutes block, 15 minutes primary antibody, 15 minutes secondary antibody; E-H), or extra-long incubations (overnight block, 24 hours primary antibody, 1 hour secondary antibody; I-L). Samples without primary antibody were used to identify non-specific staining. Cells were imaged by fluorescence microscopy and representative images were cropped and processed in Photoshop. Scale bar in (A) is 10 microns long and applies to all images. Found at: doi:10.1371/journal.ppat.1000227.s002 (2.67 MB TIF) Figure S3 Caspofungin exposes b-glucan but does not cause increased non-specific binding of mouse IgG antibody to C. albicans. WT-GFP C. albicans was grown overnight in YPD, washed, and diluted to 4*10 6 cells/ml in fresh RPMI-PS with different concentrations of caspofungin or vehicle. After growth overnight at 30uC, cells grown in vehicle (E-H) and at 1/2 MIC 50 caspofungin (A-D) were harvested and stained with anti-b-glucan antibody (Biosupplies, Inc; A and E), isotype control mouse IgG (Becton Dickinson; B and F), whole mouse IgG control (Santa Cruz Biotech; C and G) or no primary antibody (D and H) and Cy3-labeled goat anti-mouse IgG secondary antibody (Jackson Immunoresearch). Cells were imaged by fluorescence microscopy and representative images were cropped and processed in Photoshop. Scale bar in (A) is 10 microns long and applies to all images. Found at: doi:10.1371/journal.ppat.1000227.s003 (4.56 MB TIF) Figure S4 BSA blocking prevents non-specific antibody binding and doesn't cause extra b-glucan exposure. WT-GFP C. albicans was pre-grown overnight in YPD, then diluted to 4*10 6 cells/ml and grown overnight in RPMI-PS media at 30uC with either sub-inhibitory doses of caspofungin (Caspofungin; A, C, and E) or vehicle (Vehicle; B, D, and F), then washed with PBS. (A-B) Cells were stained with a shortened protocol (30 minutes block, 30 minutes with primary, 15 minutes with secondary) using buffers including 2% BSA. (C-F) Washed cells were incubated for 24 hours at 4uC with 2% BSA to determine if BSA can expose b-glucan. Then, cells were washed and stained using the shortened protocol with (C-D) or without (E-F) BSA. Samples without primary antibody were used to identify non-specific staining. Cells were imaged by fluorescence microscopy and representative images were cropped and processed in Photoshop. Scale bar in (A) is 10 microns long and applies to all images. Found at: doi:10.1371/journal.ppat.1000227.s004 (4.10 MB TIF)