Dectin-2-Targeted Antifungal Liposomes Exhibit Enhanced Efficacy

Invasive fungal diseases caused by Candida albicans, Cryptococcus neoformans, and Aspergillus fumigatus have mortality rates ranging from 10 to 95%. Individual patient costs may exceed $100,000 in the United States. All antifungals in current use have serious limitations due to host toxicity and/or insufficient fungal cell killing that results in recurrent infections. Few new antifungal drugs have been introduced in the last 2 decades. Hence, there is a critical need for improved antifungal therapeutics. By targeting antifungal-loaded liposomes to α-mannans in the extracellular matrices secreted by these fungi, we dramatically reduced the effective dose of drug. Dectin-2-coated liposomes loaded with amphotericin B bound 50- to 150-fold more strongly to C. albicans, C. neoformans, and A. fumigatus than untargeted liposomes and killed these fungi more than an order of magnitude more efficiently. Targeting drug-loaded liposomes specifically to fungal cells has the potential to greatly enhance the efficacy of most antifungal drugs.


RESULTS
Preparation of fungicide-loaded sDectin-2-coated fluorescent liposomes. A model of the fungicide-loaded, sDectin-2-coated liposomes constructed herein is shown in Fig. S1 in the supplemental material. Liposome construction methods and liposome composition paralleled closely those of sDectin-1-coated liposomes described recently (51). We remotely loaded 11 mol% amphotericin B (AmB) relative to moles of liposomal lipids into the membrane of pegylated liposomes to make AmB-loaded liposomes (AmB-LLs) (51). For reference, the widely used commercial AmB-loaded untargeted liposomal product AmBisome contains 10.6 mol% AmB relative to moles of liposomal lipid (57-59) (see Table S1 in the supplemental material). We designed the murine sDectin-2 sequence to contain a small lysine tag on its amino-terminal end (see Fig. S2 in the supplemental material), expressed it in Escherichia coli (see Fig. S3 in the supplemental material), and conjugated the purified sDectin-2 protein via this lysine tag to NHS-PEG-DSPE (N-hydroxysuccinimide plus polyethylene glycol plus 1,2distearoyl-sn-glycero-3-phosphoethanolamine), making DEC2-PEG-DSPE. We then incorporated DEC2-PEG-DSPE via its DSPE lipid moiety into AmB-LLs at 1 mol% protein molecules relative to moles of liposomal lipid (1,500 molecules of sDectin-2 per liposome) to make DEC2-AmB-LLs. Similarly, and as a protein-coated liposomal control, 0.33 mol% bovine serum albumin (BSA) was incorporated via a lipid carrier into AmB-LLs to make BSA-AmB-LLs. This resulted in equivalent microgram amounts of 22-kDa sDectin-2 and 66-kDa BSA proteins on the surface of these two sets of liposomes. Uncoated AmB-LLs, which closely resemble the commercial product AmBisome, also served as a liposomal control. Two moles percent DHPE (1,2dihexadecanoyl-sn-glycero-3-phosphoethanolamine, triethylammonium salt)-rhodamine was also incorporated into the liposomal membrane of all three liposome preparations. Hence, all three sets of liposomes contained the same 11 mol% AmB and 2 mol% rhodamine. The composition of DEC2-AmB-LLs is compared to that of BSA-AmB-LLs, AmB-LLs, AmBisome, and AmB/micelles (a deoxycholate [DOC] micelle suspension delivering AmB) in Table S1.
sDectin-2-coated DEC2-AmB-LLs bound more strongly to diverse fungal species than control liposomes. sDectin-2-coated, red fluorescent, DEC2-AmB-LLs bound strongly to C. albicans yeast cells, pseudohyphae, and hyphae (Fig. 1). The vast majority of sDectin-2-coated liposomes bound in large clusters to the extracellular polysaccharide matrix associated with these cells. Furthermore, DEC2-AmB-LLs bound to a large subset of the extracellular matrix (Ex ϩ ) surrounding these cells, while some regions of the matrix clearly did not bind sDectin-2-coated liposomes (Ex Ϫ ) (Fig. 1E). Although 100-nm liposomes are too small to be resolved by light microscopy, the presence of an estimated 3,000 rhodamine molecules per liposome (Fig. S1) allows the fluorescent signal from individual liposomes to be visualized (51). We only rarely saw individual DEC2-AmB-LLs (yellow arrows in Fig. 1A) or liposome clusters binding directly to C. albicans cell walls. In contrast, we previously showed individual sDectin-1-coated liposomes bound frequently to the cell wall of A. fumigatus cells (51).
DEC2-AmB-LLs bound strongly to C. neoformans yeast cells (Fig. 2). Monoclonal antibody 18B7 is specific for the glucuronoxylomannan (GXM) found in the capsule and exopolysaccharide of C. neoformans (60,61). We found antibody 18B7 stained most ( Fig. 2B) but not all of the cell capsules and most but not all of the exopolysaccharide that were visible in bright-field images (Ex ϩ in Fig. 2A). DEC2-AmB-LLs costained strongly most of the 18B7-stained GXM regions in the exopolysaccharide matrix, but did not stain the 187B-labeled GXM of the capsule ( Fig. 2C and D). Furthermore, there were some regions of extracellular matrix that did not stain with either 18B7 or DEC2-AmB-LLs (Ex Ϫ/Ϫ in Fig. 2A).
DEC2-AmB-LLs also bound in large clusters to the exopolysaccharide matrices produced by A. fumigatus germinated conidia and hyphae ( Fig. 2E to G). Again, little if any binding was associated with the cell wall itself. Also, there appear to be areas where the exopolysaccharide matrix is not stained or poorly stained (Ex Ϫ in Fig. 2E and F).
Because the DEC2-AmB-LLs bound poorly or not at all to mannans within the tightly cross-linked polysaccharide of cells walls of all three fungal species examined, we considered that the 100-nm-diameter size of our liposomes restricted their access or that sDectin-2 was somehow restricted from full activity when presented in a liposomal membrane. The rotational diameter of DEC2-AmB-LLs in solution would be even larger than their physical size estimate, due to their coating with sDectin-2 protein and associated water molecules. sDectin-2 itself has an atomic weight of 22 kDa and hence a rotational diameter in solution that may be estimated at ϳ4 nm (62). We prepared rhodamine-coupled sDectin-2, DEC2-Rhod. The atomic weight of rhodamine (0.48 kDa) and one or two rhodamine-coupled molecules will have little effect on this size estimate for DEC2-Rhod. Red fluorescent DEC2-Rhod bound strongly to most of the exopolysaccharide matrix surrounding A. fumigatus hyphal cells (see Fig. S4A and B in the supplemental material). The pattern of binding and intensity of binding was indistinguishable from that of DEC2-AmB-LLs ( Fig. S4C and D).
We quantified the efficiency of DEC2-AmB-LLs binding to fungal cells compared to control uncoated AmB-LLs and BSA-coated liposomes (BSA-AmB-LLs). The area of fluorescent liposome signal was measured from multiple red fluorescent photographic images taken of liposome-stained cultures that were nearly confluent with fungal cells (Fig. 3). DEC2-AmB-LLs bound to C. albicans pseudohyphae and hyphae, C. neoformans yeast cells, and A. fumigatus hyphae 50-to 150-fold more efficiently than AmB-LLs or BSA-AmB-LLs (Fig. 3A, D, and G). Examples of the photographic images of fluorescent liposomes quantified to make these measurements are presented adjacent to each bar graph (Fig. 3B, C, E, F, H, and I).
Using the same quantitative assay of the area of fluorescent liposome binding to C. albicans pseudohyphae and hyphae, we characterized the specificity, stability, and rate of DEC2-AmB-LL binding (Fig. 4). DEC2-AmB-LL labeling was 75% inhibited by the inclusion of solubilized yeast mannans during the binding assay, but not by the same concentrations of the soluble ␤-glucan laminarin or of the glucose-fructose-containing disaccharide sucrose ( Fig. 4A to C). This result confirmed that sDectin-2-targeted liposome binding to the extracellular matrix was mannan specific, in agreement with the published carbohydrate specificity of sDectin-2 (42)(43)(44)(45)(46). We examined the stability of binding of DEC2-AmB-LLs to C. albicans cells by taking the DEC2-AmB-LL-stained preparations examined in Fig. 3A and storing them in the dark in phosphate-buffered saline (PBS) at 4°C. After 2 months, these cells were rephotographed and liposomal staining quantified. The fluorescent intensity of DEC2-AmB-LLs bound to cells remained strong and was estimated to be 50-fold stronger than the nonspecific binding of AmB-LLs ( Fig. 4D to F). This result suggests that the DEC2-AmB-LLs themselves are relatively stable and that their binding to cells is also relatively stable. The rate of DEC2-AmB-LL binding was estimated by exposing dense fields of C. albicans pseudo- hyphal and hyphal cells to liposomes for periods ranging from 10 s to 90 min before washing off unbound liposomes ( Fig. 4G and H). The area of DEC2-AmB-LL labeling increased rapidly and exponentially for the first 15 min (Fig. 4G) and then slowed, but labeling did not appear to be complete after 90 min (Fig. 4H).
In summary, Dectin-2-coated AmB-loaded liposomes (DEC2-AmB-LLs) bound specifically, stably, and rapidly to fungal mannans present in the extracellular matrix of C. albicans grown in vitro, while little nonspecific binding of control liposomes was observed. There were only trace amounts of DEC2-AmB-LL binding to the cell wall. DEC2-AmB-LLs also bound efficiently to portions of the extracellular matrices surrounding C. neoformans and A. fumigatus cells.
Growth inhibition and killing by DEC2-AmB-LLs. We performed various fungal cell growth and viability assays after treating actively growing cultures of C. albicans, C. neoformans, and A. fumigatus with sDectin-2-coated and control liposomes delivering AmB concentrations near the MICs for this fungicide (Fig. 5). Depending upon the assay conditions and delivery method for AmB, the MICs for AmB estimated for these species range from 0.06 to 1.3 M (63-67). It seemed reasonable to consider that at reported AmB concentrations near the MICs for these species, we might best resolve the improved performance of targeted AmB-loaded liposomes relative to untargeted drug.
We inoculated 4,000 C. albicans yeast cells into individual wells of 96-well microtiter plates, grew them for 6 h to the pseudohyphal and early hyphal stages, and treated the cells with drug-loaded liposomes. After a 30-min incubation, liposomes were washed out and cells were grown for an additional 16 h. Figure 5A shows that targeted DEC2-AmB-LLs delivering from 1 M down to 0.125 M AmB killed or inhibited of C. albicans cells from 90-fold to 3-fold more efficiently than uncoated AmB-LLs or BSA-AmB-LLs delivering the same concentrations of AmB. The difference is remarkable given that in these experiments, cells were exposed to liposomal drug for only 30 min. These data were obtained using CellTiter-Blue (CTB) reagent to assess cytoplasmic reductase activity as a proxy for cell integrity and viability. Dead cells or metabolically inactive cells do not reduce the resazurin substrate to fluorescent resorufin product. Treatment of C. albicans cells continuously with liposomes for the entire 16-h time period or with higher drug concentrations resulted in too much cell death for all three drug-loaded liposome samples to clearly resolve differences among the different liposome preparations. Six-month-old preparations of DEC2-AmB-LLs retained their full antifungal activity, as long as they were freshly reduced (Materials and Methods). We also assayed viable cell numbers after liposome treatment. C. albicans yeast cells growing in rich medium in liquid culture were incubated with liposomes delivering 2 M AmB. The liposomes were washed out after 30 min. After an additional 6 h of growth, the cultures were diluted and assayed for CFU on agar plates with rich medium. Based on CFU, DEC2-AmB-LLs were 3-fold more effective at inhibiting or killing C. albicans yeast cells in liquid than AmB-LLs or BSA-AmB-LLs (Fig. S2A).
C. neoformans yeast cells growing in liquid were treated for 4 h with liposomes delivering 0.4 M AmB and overnight with liposomes delivering 0.4, 0.2, and 0.1 M AmB (Fig. 5B). At the end of each treatment, cells were diluted and CFU were assayed on agar plates with rich medium. DEC2-AmB-LLs were 2.5-to 11-fold more effective at killing C. neoformans than AmB-LLs or BSA-AmB-LLs, with the optimum treatment being 0.2 M AmB overnight. As an alternative assay, C. neoformans yeast cells growing on  Cells were diluted and plated on YPD medium, and CFU were counted from multiple plates. (C) A. fumigatus with DEC2-AmB-LLs. Conidia were germinated for 9 h in VMM plus glucose plus 0.5% BSA in 96-well polystyrene microtiter plates, treated for 2 h with liposomes delivering 0.5 and 0.25 M AmB to the medium as indicated, washed twice with medium, grown overnight, and then assayed for metabolic activity using CTB reagent in RPMI lacking phenol red indicator plus 0.5% BSA. The control wells were overgrown with hyphae protruding out the medium. (D) A. fumigatus with DEC1-AmB-LLs. Assay conditions were similar to the assay in panel C, except liposomes were first diluted into LDB1 buffer (PBS plus 0.5%BSA plus 1 mM BME) prior to dilution into growth medium. For CTB assays in panels A, C, and D, the fluorescence background from medium incubated with CTB reagent was subtracted. Standard errors are shown for all values, and fold differences and P values were estimated comparing the performance of AmB-LLs to DEC2-AmB-LLs. Two or more biological replicates gave similar results. microtiter plates were treated for 5 h with liposomes delivering 1 M AmB. Cells were immediately assayed for cell death by incubating them with propidium iodide. Propidium iodide enters dead but not live cells and fluoresces red when it intercalates into double-stranded DNA or short regions of double-stranded RNA. Propidium iodide assays showed that under these treatment conditions, DEC2-AmB-LLs were 5-fold more efficient at killing C. neoformans cells than uncoated AmB-LLs (Fig. S2B, C, and D).
A. fumigatus was treated with liposomes delivering AmB concentrations near and below the MIC estimated for AmBisome: 0.5 M (67). Conidia were germinated and grown until the very early germling stage, when the germ tube first began to emerge from 95% of conidia. Cells were then treated for 2 h with AmB-containing liposomes or liposomal dilution buffer, and unbound liposomes were washed off with growth medium. Cells were grown for an additional 19 h and assayed for viability and metabolic activity with CellTiter-Blue reagent. DEC2-AmB-LLs delivering 0.5 and 0.25 M AmB killed or inhibited the growth of A. fumigatus 20-fold and 36-fold more efficiently than AmB-LLs, respectively (Fig. 5C). It should be noted that the dilution buffer control-treated A. fumigatus cells overgrow and generate a thick hyphal mat in the microtiter wells during this assay. Hence, the metabolic activity and CellTiter Blue signal from these control cells was low.
Dectin-1-targeted AmB-loaded liposomes also bind efficiently to A. fumigatus swollen conidia, germlings, and hyphal cells and inhibit and kill these cells (51), although they are binding ␤-glucans and not ␣-mannans. In this previous study, cells were incubated continuously with liposomes during the entire assay and not washed out. For a more direct comparison of the drug-targeting efficiency of the two Dectins, we examined DEC1-AmB-LLs using the same assay design used herein for DEC2-AmB-LLs, washing out the liposomes after 2 h, except that liposomes were initially diluted into LDB1 and the cells were grown out for 16 h (see Materials and Methods). DEC1-AmB-LLs delivering 0.5 and 0.25 M AmB killed or inhibited the growth of A. fumigatus 28-fold and 5-fold more efficiently than AmB-LLs, respectively (Fig. 5D). Using this assay condition, the results from AmB-loaded liposomes targeted by Dectin-1 and Dectin-2 are very similar.
Toxicity of DEC2-AmB-LLs to animal cells. Rapidly grown cultures of human HEK-293 and human HT-29 cells were treated overnight with various liposomes and a deoxycholate micelle suspension each delivering 15 M AmB (AmB/micelles). Based on CellTiter-Blue assay of metabolic activity and viability, DEC2-AmB-LLs were 10% to 20% more toxic than AmB-LLs or BSA-AmB-LLs and 2-to 5-fold less toxic than AmB deoxycholate (AmB/DOC) micelles (Fig. S3). When these cells were treated to receive lower AmB concentrations (for example, 3 M AmB), only AmB/DOC micelles showed measurable toxicity. None of the three liposomal preparations appeared to have any particular affinity for either of these cell lines when examined by fluorescence microscopy.

DISCUSSION
We coupled the N-terminal domain of Dectin-2, sDectin-2, to a lipid carrier and inserted this conjugate into liposomes such that the C-terminal carbohydrate recognition domain (CRD) of each monomer faced outward from the liposomal membrane. Considering the efficient binding we obsered for these DEC2-AmB-LLs to the extracellular matrices of three diverse human fungal pathogens, the sDectin-2 monomers must have been conformationally free to form the functional dimers necessary for efficient mannan binding (48,68). C-type lectin receptors often bind some of their substrates weakly. The published estimated effective concentrations for 50% of sDectin-2 binding to mannan-related polysaccharides (EC 50 s) range from approximately 20 mM for mannose and 2 mM for mannan-␣-1-2-mannan (68) down to 150 M for mannoglycan (44,68). This is indeed weak binding compared to Dectin-2's closest paralog, Dectin-1, which has EC 50 s for binding various ␤-glucans ranging from 2 mM down to 2.2 pM (69). It is likely that the greater avidity created by having approximately 1,500 sDectin-2 monomers on each liposome resulted in the rapid, strong, and stable binding we oberved for DEC2-AmB-LL binding to fungal cells.
Most pathogenic fungal species, including Candida, Cryptococcus, and Aspergillus, live within polysaccharide matrices that promote adherence to host cells, tissues, and implanted biomedical devices, and which partially protect them from host antifungal ativities (70)(71)(72)(73). The exopolysacchride matrix of C. albicans contains approximately 13% ␤-1,6-glucan and 87% ␣-1,6-mannan (74). C. neoformans produces a glucuronoxylomannan-and galactoxylomannan-rich capsule (75) that is shed and contributes to formation of the extracellular matrix. The matrix is composed of 20% mannan and 23% glucan residues (76). The exopolysaccharide matrix of A. fumigatus contains approximately 14% galactomannan and 13% ␤-glucans (77). Our data on the efficient binding of DEC2-AmB-LLs to the extracellular matrix of all three species is consistent with the mannan-rich content of their exopolysaccharides. However, some regions of the C. albicans, C. neoformans, and A. fumigatus matrices did not stain or stained poorly with sDectin-2-coated liposomes, suggesting that either the distribution of mannans in the matrix is heterogeneous or the mannans in these regions are masked from exposure to liposomal sDectin-2. This distinction between mannan exposure or expression within their exopolysaccharide matrices remains to be resolved. We showed that the binding of DEC2-AmB-LLs to exopolysaccharide mannans was no more restricted than the binding of the much smaller DEC2-Rhod, suggesting size restriction may not be the primary limitation.
We did not find convincing evidence that DEC2-AmB-LLs bound more than at trace levels to the cell walls of any of these fungal species. Although the cell wall mannan content varies widely based on growth media and methods of chemical analysis, the estimated cell wall mannan polysaccharide content of C. neoformans is 22% (78), that of C. albicans is 40% (79), and that of A. fumigatus is 15% to 41% (73,79). Although rhodamine-labeled sDectin-2 (DEC2-Rhod) bound with similar intensity and pattern to the exopolysaccharide matrix of A. fumigatus as DEC2-AmB-LLs, DEC2-Rhod did not stain the cell wall. Hence, again a size barrier for the much larger DEC2-AmB-LLs does not appear to explain their lack of cell wall binding. Hence, our findings suggest the cell wall mannans may be chemically masked from sDectin-2-mediated liposomal binding. This result parallels the masking of C. albicans cell wall ␤-glucans from sDectin-1 binding and from DEC1-AmB-LL binding (52,(54)(55)(56).
C. albicans reversibly switches among unicellular ovoid yeast and multicellular ellipsoid pseudohyphal and elongated hyphal morphologies (80). All three stages are capable of producing an extracellular matrix and adhering to host tissues. The matrices of all three bound DEC2-AmB-LLs, suggesting fungicide-loaded Dectin-2-coated liposomal therapeutics have the potential to reduce the virulence of C. albicans.
DEC2-AmB-LLs inhibited and killed C. albicans, C. neoformans, and A. fumigatus far more efficiently than plain uncoated AmB-LLs or BSA-AmB-LLs delivering the same concentrations of AmB. A combination of metabolic activity assays based on CellTiter-Blue (CTB) reagent, cell growth assays based on CFU, and propidium iodide staining of dead cells confirmed that both inhibition and killing of cells occurred. Incubation with DEC2-AmB-LLs for as little as 30 min to a few hours resulted in significant killing. DEC2-AmB-LLs were 3-to 90-fold more efficient at inhibiting or killing fungal cells, when delivering AmB concentrations near or below the MIC values reported for AmB, concentrations at which our AmBisome equivalent uncoated AmB-LLs had little or no impact on cell inhibition or survival. The DEC2-AmB-LLs bound all three species 50-to 150-fold better than AmB-LLs and, under some conditions we tested, inhibited and killed them 11-to 94-fold more efficiently than AmB-LLs. We believe the differences in the optimal results for the binding and killing assays and the variability within the killing assays are due primarily to the fact that inhibition and killing assays are much more complex and difficult to optimize. They are influenced by many variables, including (i) the amount of AmB being delivered, (ii) the cells proceeding through more than one developmental stage during the assay, (iii) the time of liposome exposure, (iv) the length of the growth period after liposome exposure, and (v) the mechanics of the assay being used. The binding assays are influenced primarily by the concentration of sDectin-2 and the extent of washing.
A significant reduction in the MIC for AmB or other liposomal packaged therapeutic fungicides should result in reducing fungicide doses and the frequency of drug administration, which, in turn, should reduce host toxicity. We showed that DEC2-AmB-LLs were not particularly toxic to animal cells when delivering 15 M AmB, which was 15-to 150-fold higher than the AmB concentrations used here to kill fungal cells. However, there are number of important factors that we have not yet explored in vitro, which may alter the amount of exopolysaccharide matrix and its composition and may influence the effectiveness of Dectin-2-targeted antifungals in vivo. These include growth substrate, temperature, carbon source, serum, and the particular antifungals being delivered.
Candida, Cryptococcus, and Aspergillus species belong to three evolutionarily disparate classes of fungi (81), the Saccharomycetes (phylum Ascomycota), Tremellomycetes (phylum Basidiomycota), and Eurotiomycetes (phylum Ascomycota), respectively. It is estimated that they separated from common ancestry relatively early in the evolution of the fungal kingdom, 0.8 to 1.3 billion years ago (82). DEC2-AmB-LLs bound specifically to the extracellular matrix of all three. This suggests the mannans found in the extracellular matrices of most pathogenic fungi will be conserved enough in structure to bind sDectin-2-targeted liposomes. However, mannan expression and masking from host detection may vary widely among pathogenic fungal species or in different host niches. Consistent with the masking of mannans to sDectin-2 binding, our assays showed very inefficient binding to the mannan-rich cell walls of all three species.
This preliminary study was performed with the mouse sDectin-2 protein sequence to avoid problems of sDectin-2 immunogenicity during the future testing of sDectin-2-targeted antifungals in mouse models of candidiasis, aspergillosis, and cryptococcosis. The human sDectin-2 protein sequence is 72% identical to the mouse protein and is only 2 amino acids (aa) shorter. Therefore, we do not expect problems manipulating the human protein to target fungicide-loaded therapeutic liposomes for use in clinical studies.
In summary, there is a dire need for new antifungal therapeutics. DEC2-AmB-LLs bound efficiently to the extracellular matrices produced by diverse cellular stages of C. albicans, C. neoformans, and A. fumigatus. DEC2-AmB-LLs delivering AmB concentrations near or below AmB's MICs for growth inhibition and killing of all three species showed sDectin-2 targeting of the liposome-packaged drug improved the fungicidal effect by an order of magnitude or more over untargeted liposomal AmB. It is reasonable to propose that, drug-loaded liposomes targeted to fungal cells have significant potential as pan-antifungal therapeutics with a wide range of applications. The next step in our research will be to examine the efficacy of targeted liposomal antifungal drugs in mouse models of invasive fungal diseases.

MATERIALS AND METHODS
Cell culture. C. albicans CAI4 expressing green fluorescent protein (GFP) under the control of the ADH1 promoter (83), A. fumigatus A1163 (51), and wild-type C. neoformans H99-alpha (84) were grown in Vogel's minimal medium (VMM) plus 1% glucose (85) plus 0.5% BSA or RPMI 1640 medium with no red indicator dye (Thermo Fisher SKU-11835-030) plus 0.5% BSA or YPD (1% yeast extract, 2% peptone, 2% dextrose) in liquid with shaking or on 24-or 96-well polystyrene microtiter plates or on glass microscope chamber slides and incubated at 37°C for 3 to 36 h. Plates were precoated with poly-L-lysine for A. fumigatus, and glass microscope slides were precoated with for all three species. All fungal cell growth was carried out in a biosafety level 2 (BSL2) laboratory. Before cells were treated with fluorescent liposomes for microscopic analysis of binding, fungal cells were washed three times with PBS, fixed in 4% formaldehyde in PBS for 60 min, washed three times, and stored at 4°C in PBS.
The human colorectal adenocarcinoma cell line HT-29 (ATCC HTB-38) and human embryonic kidney cell line HEK-293 (ATCC CRL-1573) were grown in 96-well microtiter plates in RPMI medium lacking red indicator dye plus 10% fetal calf serum in a 37°C incubator in air supplemented with 5% CO 2 . Their viability and metabolic activity after overnight antifungal treatment were assayed using CTB reagent diluted 1:10 into the medium and with incubation for 60 to 90 min at 37°C with 8 wells per treatment.
Production and chemical modification of sDectin-2. The carboxy-terminal end of Dectin-2 contains its mannan recognition domain, sDectin-2. The sequence of the codon-optimized E. coli expression construct with MmsDectin-2lyshis synthesized and cloned into pET-45B by GenScript is shown in Fig. S2.
The DEC2-PEG-DSPEand BSA-PEG-DSPE conjugates in RN#5 buffer and PBS, respectively, were integrated via their lipid DSPE moiety into the phospholipid bilayer membrane of AmB-LLs by 30 min of incubation at 60°C to make DEC2-AmB-LLs and BSA-AmB-LLs, parallel to the protocol used previously (51). During these same 60°C incubations, 2 mol% of the red fluorescent tag Lissamine rhodamine B-DHPE (Invitrogen, no. L1392) was also incorporated into sDectin-2-and BSA-coated liposomes and AmB-LLs (51). A duplicate sample of the DEC2-AmB-LLs was subjected to gel exclusion chromatography over 0.5 M Bio-Gel A resin (Bio-Rad, no. 151-0140). Fluorescent liposomes were efficiently excluded from the resin. Because no sDectin-2 (optical density at 280 nm [OD 280 ] of 2.2/mg/ml) or rhodamine was detected in the low-molecular-weight fractions included in the gel, we concluded that both were efficiently loaded into liposomes. Fresh 2 mM BME was added to DEC2-AmB-LLs before each use in binding or killing assays. DEC2-AmB-LLs stored at 4°C in RN#5 appeared to retain full fungal cell binding specificity and killing activity for at least 12 months.
Microscopy of liposomes and DEC2-Rhod bound to fungal cells. Formalin-fixed fungal cells were incubated with liposomes at 23°C in liposome dilution buffer LDB2 (20 mM HEPES, 10 mM triethanolamine, 150 mM NaCl, 10 mM CaCl 2 , 1 mM BME, 5% BSA, pH 8.0), wherein the BME was added fresh. Liposomal stocks were diluted 1:200 before incubation with cells such that the sDectin-2 protein concentration was 0.5 g/100 l. After incubations of 15 min, 1-h, or longer, unbound liposomes washed out with 4 changes of LDB2. Merged images of rhodamine red fluorescent liposomes, green fluorescent cells, and differential interference contrast (DIC) illuminated cells were taken of cells grown on microscope slides under oil immersion at 63ϫ on a Leica DM6000B automated microscope. The DEC2-Rhod stock was also diluted 1:200 before incubation with cells such that the sDectin-2 protein concentration was also 0.5 g/100 l. Bright-field, DIC, and red (excitation of 560 and emission of 645 [Ex560/Em645]) and green (Ex500/Em535) fluorescent images were taken of cells on microtiter plates at 20ϫ on an Olympus IX70 inverted microscope using an Olympus PEN E-PL7 digital camera, and the bright-field and/or fluorescent-colored layers were merged in Photoshop. The area of liposome binding at ϫ20 magnification was quantified by taking an 8-bit grayscale copy of the unmodified red fluorescent TIF image into Image J (imagej.nih.gov/ij), using ImageϾAdjustϾThresholdϾApply to capture just the red fluorescent areas illuminated by liposomes, and using AnalyzeϾMeasure to place the area data for each image in a file. Six to 10 photographic images were generally analyzed and averaged for most area estimates. However, because the staining intensities varied widely among photographic fields of C. neoformans cells, 90 images were analyzed for each treatment. Bright-field, DIC, and fluorescent images were also taken of cells grown in microsope chamber slides at 20ϫ or under oil immersion at 63ϫ on a Leica DM6000B automated microscope.
Growth inhibition and viability assays following liposome treatment. Liposomal stocks were stored at 615 to 800 M AmB and typically diluted first 30-to 600-fold into liposome dilution buffer, LDB2, and then diluted 1:11 into growth medium to achieve the indicated final fungicide concentrations ranging from 2 M down to 0.1 M. Control cells received an equivalent amount of LDB2. CellTiter-Blue (CTB) cell viability and metabolic activity assays of C. albicans and A. fumigatus were conducted as we recently described for A. fumigatus, by incubating cells with CTB reagent for 3 to 4 h (51) and analyzing 96-well plates in a Bio-Tek Synergy HT fluorescent microtiter plate reader. The fluorescent background from control wells was subtracted from experimental wells. Data from eight wells were averaged for each msphere.asm.org 13 data point. These assays had a lot of background and were less sensitive if the cells were assayed in VMM. We were unable to detect any fluorescent signal from CTB reagent when performing parallel cell viability assays on C. neoformans using this protocol or another CTB protocol published for this species (86). As an alternative measure of C. neoformans and C. albicans cell viability, assays were conducted by growing 1 ml of cells in YPD, adding drug-loaded liposomes for the indicated times of growth, diluting the cells, plating on YPD, and counting CFU. The fraction of dead C. neoformans cells among cells growing in YPD after treatment with fungicide-loaded liposomes was assayed by adding propidium iodide to the medium at 50 g/ml and incubating for 60 min at 37°C. The medium was removed and replaced with PBS for fluorescence microscopy using the red fluorescent protein channel (Ex560/Em595) and scoring the percentage of dead stained cells relative to the total number of stained and unstained cells. In experiments with DEC1-AmB-LLs, all liposomes were diluted with LDB1 (PBS plus 5% BSA plus 1 mM BME) (51) instead of LDB2.