ABSTRACT
Antifungal exposure can elicit immunological effects that contribute to activity in vivo, but this activity is rarely screened in vitro in a fashion analogous to MIC testing. We used RAW 264.7 murine macrophages that express a secreted embryonic alkaline phosphatase (SEAP) gene induced by transcriptional activation of NF-κB and activator protein 1 (AP-1) to develop a screen for immunopharmacological activity of cell wall-active antifungal agents. Isolates of Candida albicans and Aspergillus fumigatus that conditionally express genes involved in cell wall synthesis were also tested with the reporter macrophages. We found that growth of fungi in subinhibitory concentrations of glucan synthesis inhibitors (caspofungin and enfumafungin A) or repression of the β-glucan catalytic subunit of glucan synthase, FKS1, increased macrophage NF-κB/AP-1 activation in a dectin-1-dependent manner. This pattern of activation was also transiently observed with repression of chitin synthesis in C. albicans or when yeast cells were incubated in low concentrations of the chitin synthesis inhibitor nikkomycin Z.
INTRODUCTION
The host immune response plays a critical role in shaping the efficacy and safety of antifungal therapy in vivo. Some antifungals, such as amphotericin B, directly activate immune cells through microbial pattern recognition receptors (PRRs) (toll-like receptor 2 [TLR2] and CD14), causing the release of proinflammatory cytokines (tumor necrosis factor alpha [TNF-α], interleukin-6 [IL-6], IL-1Ra, and IL-1β), chemokines (IL-8, monocyte chemoattractant protein 1 [MCP-1], and MIP-1β), nitric oxide, and prostaglandins (1). This direct stimulation of immune cells underlies the frequent and sometimes severe infusion-related reactions (fever, chills, rigors, nausea, and tachypnea) observed in patients during amphotericin B infusions (2). Alternatively, antifungals can also indirectly stimulate host immune cells through their inhibitory or lethal effects in fungi. Exposure of Candida or Aspergillus species to echinocandin antifungals increases exposure of normally concealed branched glucan polymers, which elicit strong inflammatory responses from macrophages in vitro (3, 4) and in vivo (5). Although these mechanisms can enhance antifungal drug efficacy, immunopharmacological mechanisms are infrequently considered components of antifungal activity or spectrum in vivo.
Similarly, the immunopharmacological activity of antifungals is rarely screened for established or novel antifungal agents in a fashion similar to MIC assays. To that end, we sought to develop a facile assay for screening drug immunopharmacological activity that could complement other phenotypic endpoints used for screening of drug activity or conditional mutant libraries. We also reasoned that an immunopharmacological screen would be useful in the evaluation of resistance in clinical isolates in the research laboratory. For example, certain antifungal resistance mutations may also facilitate immunoevasion strategies in the host even in the absence of antifungal selective pressure, contributing to the persistence of resistant strains in vivo (6, 7).
Specifically, the purpose of our study was to develop a simple, cell-based reporter assay suitable for screening indirect immunopharmacological effects of antifungal agents in Candida albicans and Aspergillus fumigatus. During the development of this assay, we characterized the immunopharmacological profiles of caspofungin; a nonechinocandin glucan synthesis inhibitor, enfumafungin A; and the chitin synthase inhibitor nikkomycin Z over a wide range of clinically achieved concentrations. Additionally, we examined whether strains of C. albicans and A. fumigatus with conditional mutations in cell wall synthesis genes could be used to phenocopy the immunopharmacological effects of antifungals tested in the assay.
MATERIALS AND METHODS
Antifungal agents.Amphotericin B and tetracycline hydrochloride (Sigma, St. Louis, MO), fluconazole and voriconazole (Pfizer, New York, NY), and caspofungin and enfumafungin A (Merck Research Laboratories, Rahway, NJ) were prepared as stock solutions from analytical-grade powder, as described in the Clinical and Laboratory Standards Institute (CLSI) M27-A3 and M38-A2 protocols (8, 9). Amphotericin B lipid complex (ABLC; Sigma-Tau, Gaithersburg, MD) was obtained from the manufacturer and diluted in phosphate-buffered saline (PBS) prior to testing.
Test isolates.Table 1 lists the Candida albicans (n = 5) and Aspergillus fumigatus (n = 3) isolates used for all experiments (kindly provided by Merck Research Laboratories). Candida albicans strains consisted of the parent wild-type (WT) strain, C. albicans CAI4, and four strains constructed in the CAI4 background with conditional heterozygous or homozygous loss-of-function mutations in chitin synthase (CHS3) or glucan synthase (FKS1) under the control of a tetracycline promoter (10). A. fumigatus isolates consisted of an A. fumigatus CEA17 uracil/uridine auxotroph parent strain and strains with conditional loss-of-function mutations in chitin synthase (CHS2) and glucan synthase (FKS1) under the control of the pNiiA nitrogen-regulated promoter (11). Candida MICs were determined according to CLSI M27-A3 (9) methods in RPMI 1640 plus 0.165 M MOPS [3-(n-morpholino)propanesulfonic acid], pH 7.0, without tetracycline. Minimal effective plasma concentrations (MECs) for Aspergillus were determined using a modification of the CLSI M38-A2 method (8) in Aspergillus complete medium (ACM) with uridine/uracil supplementation (12).
Isolates used for testingb
Macrophage reporter cells.We used a commercially available reporter macrophage cell line (RAW-Blue cells; InvivoGen, San Diego, CA). Cells were derived from murine RAW 264.7 macrophages with chromosomal integration of a secreted embryonic alkaline phosphatase (SEAP) reporter construct, induced by NF-κB and activator protein 1 (AP-1) transcriptional activation. An NF-κB reporter was selected because of its rapid posttranslational activation in response to activation of microbial pattern recognition receptors (PRRs). The macrophage reporter cells express PRRs, including toll-like receptors (TLRs) and NOD-like, RIG-I-like, and C-type lectin receptors, including the PRR for fungal 1,3-β-glucan, dectin-1 (13). Activation of these PRRs by fungal organisms induces signaling pathways leading to the activation of NF-κB and AP-1 and the subsequent production of SEAP. Concentrations of SEAP are then measured by spectrophotometry to characterize the magnitude of NF-κB and AP-1 transcriptional activation using a colorimetric enzyme assay (Quanti-Blu; InvivoGen).
Macrophages were propagated, banked, and prepared for testing daily using protocols developed by the manufacturer. Cells were grown to 80% confluence in T-25 culture flasks in Dulbecco's modified Eagle's medium (DMEM) containing 4.5 g/liter glucose, heat-inactivated 10% fetal bovine serum, 2 mM l-glutamine, and 200 μg/ml Zeocin antibiotic (InvivoGen) at 37°C in 5% CO2. To minimize experimental variability, only cells with fewer than 10 passages were used. On each experimental day, cells were harvested by removing growth medium, followed by washing with sterile phosphate-buffered saline (PBS), trypsinization, and resuspension of the cells in fresh DMEM growth medium. Cell viability was confirmed to be >95% by trypan blue counting using a Countess cell imager (Life Technologies, Grand Island, NY). The final test concentration of reporter macrophages used for all experiments was 5 × 105 cells/well.
Candida albicans drug analysis.We used a modification of methods described by Wheeler and Fink (3) for evaluating macrophage response to inactivated C. albicans. Yeasts were grown for 18 h in a shaking incubator at 35°C in 50-ml polypropylene tubes containing RPMI growth medium alone or in the presence of increasing serial concentrations (0 to 200 ng/ml) of antifungal agents. Additional replicate tubes were included for higher antifungal concentrations that produced near-fungicidal effects to achieve the required fungal inoculum required for testing. On the morning of the experiment, yeast cells were pelleted by centrifugation (5,000 rpm), washed twice with PBS, and then resuspended in 5 ml of PBS. An aliquot (1 ml) of the suspension was extracted using the Y-PER yeast protein extraction kit (Thermo Scientific, Rockford, IL) and analyzed for protein concentration using a modified Lowry assay (Bio-Rad detergent compatible protein kit; Bio-Rad, Hercules, CA) using bovine serum albumin standards as recommended by the manufacturer. Once the protein concentration was determined, the inoculum was standardized to dispense 20 μg of C. albicans protein equivalent into each well of a 96-well, flat-bottom microtiter tray. This protein concentration was roughly equivalent to 5 × 106 C. albicans CAI cells per well, confirmed by hemocytometer counts.
Microtiter trays were then subjected to 4 cycles of UV irradiation (120,000 μJ/cm2) (Stratalinker UV cross-linker 1800; Stratagene, La Jolla, CA) to inactivate cells with shaking between each cycle. Cells were inactivated to reduce the confounding effects of growth in the cell culture medium, which could alter the cell wall surface. Each microplate was centrifuged (3,500 rpm) to pellet cells, and PBS was carefully aspirated using a micropipette, avoiding disruption of the fungal pellet.
Freshly harvested and washed macrophages (5 × 105) were then added (180 μl) to each well of a microplate containing the fungal inoculum (fungal cell/macrophage ratio, 10) and mixed by micropipette. A negative control (endotoxin-free PBS) and a positive control (10 μg zymosan, a TLR2/dectin-1 agonist) were included in each experimental run, as well as macrophages preincubated for 30 min with 10 μg/ml anti-dectin-1 monoclonal antibody (MAb) (InvivoGen) to assess the specific contribution of β-1,3-glucan/dectin-1-mediated activation. After 6 h of incubation at 37°C in 5% CO2 with intermittent shaking, the microtiter tray was centrifuged to pellet macrophages and fungal material, and 50 μl of the cell culture supernatant and negative and positive controls was transferred for analysis of SEAP concentrations using methods described by the manufacturer (Quanti-Blu assay; InvivoGen). SEAP absorbance was read at an optical density at 655 nm (OD655) on a microplate spectrophotometer at 6, 12, and 24 h. Data reported in this publication are for 24-h readings.
Candida albicans conditional mutants.Isolates with heterozygous or homozygous conditional mutations in FKS1 and CHS3 were grown in nonrepressing (RPMI alone) or repressing (RPMI plus 100 μg/ml tetracycline) liquid medium for 18 h in a shaking incubator at 35°C (10). Fungal cells were then collected and processed for testing in a fashion similar to that of antifungal-exposed cells. We also examined how macrophage inflammatory responses changed with different “simulated antifungal doses” by growing the homozygous FKS1 mutant in different (repressing) tetracycline concentrations.
Aspergillus fumigatus drug analysis.A modification of methods described by Hohl et al. (4) was used to test macrophage responses against drug-exposed A. fumigatus. A. fumigatus CEA17 conidia (5 × 106) suspended in 200 μl of ACM plus uridine/uracil alone or ACM plus uridine/uracil with antifungal agents (0 to 64 μg/ml) were prepared and dispensed into a 96-well flat-bottom microtiter tray. After 18 h of incubation at 35°C, microplates were centrifuged (3,500 rpm) and medium was carefully aspirated by micropipette using an inverted microscope to leave a hyphal mat on the bottom of each well. Hyphae were then gently washed twice with PBS and subjected to 4 cycles of UV irradiation (120,000 μJ/cm2) with mixing between each cycle, and the remaining PBS was carefully aspirated. No protein standardization of inoculum was performed, as the antifungal agents had minimal effects on total fungal biomass (<15%) at the tested concentrations (14). Freshly harvested macrophages (5 × 105) or macrophages preincubated with anti-dectin-1 MAb were then added to each well with PBS (negative control) or zymosan (positive control) in a fashion similar to that in Candida studies. After 6 h of coincubation, the culture medium was collected and analyzed for SEAP concentrations as previously described.
Aspergillus fumigatus conditional mutants.A. fumigatus CHS2 and FKS1 conditional mutants, controlled by the nitrogen-responsive pNiiA promoter, were grown overnight (18 h) from 5 × 106 conidia per well in a microplate containing inducing (Aspergillus minimal medium [AMM] plus nitrate) or repressing (AMM plus ammonium tartrate) medium as described by Hu et al. (11). Microplates were then centrifuged to pellet the hyphal mat, washed, and UV inactivated as previously described before addition of macrophages (5 × 105), anti-dectin-1-blocked macrophages, PBS (negative control), or zymosan (positive control). Plates were incubated for 6 h before medium was collected for analysis of SEAP concentrations.
Statistical analysis.Experiments were performed in duplicate on two separate days. Mean SEAP concentrations ± standard deviations were compared by Kruskal-Wallis test, with Dunn's post hoc test for individual comparisons. Differences in SEAP concentrations associated with a P value of <0.05 were used to define significant differences in macrophage NF-κB/AP-1 activation status.
RESULTS
Macrophage reporter cell assay.A number of test variables were examined during assay development, including fungal cell culture conditions, the optimal target inoculum/macrophage ratio, and the potential for direct macrophage activation of the reagents and antifungals. Among antifungals initially screened for direct immunostimulatory activity, only amphotericin B was found to activate NF-κB/AP-1 transcription at concentrations above 1 μg/ml (see Fig. S1 in the supplemental material). Exposure to caspofungin, enfumafungin A, or nikkomycin Z did not significantly increase macrophage transcriptional activation of NF-κB/AP-1 up to concentrations of 10 μg/ml. The final test conditions described in Materials and Methods produced consistent results that were in agreement with previous work using RAW 264.7 macrophages or primary cell lines (3, 4).
Glucan and chitin synthesis inhibitors enhance macrophage inflammatory responses to C. albicans.Macrophages exposed to C. albicans CAI4 grown without antifungals elicited marginal, nonsignificant increases in NF-κB/AP-1 macrophage activation over that in the negative controls. However, growth of CAI4 in the presence of subinhibitory (0.125 to 0.5× MIC) concentrations of caspofungin or enfumafungin elicited 2- to 4-fold-higher rates of NF-κB/AP-1 transcriptional activation in macrophages (P < 0.05) compared to non-drug-exposed fungi, which were blocked by the addition of anti-dectin-1 MAb (Fig. 1). Notably, macrophage activation diminished with higher drug concentrations (200 ng/ml), particularly for enfumafungin A. Similarly, CAI4 grown in the presence of low (0.1- to 1-ng/ml) but not higher (10- to 500-ng/ml) concentrations of nikkomycin Z induced 2-fold-higher rates (P < 0.05) of NF-κB/AP-1 transcriptional activation in macrophages, which was attenuated by anti-dectin-1 MAb, suggesting that chitin synthesis inhibitors may also induce increased surface exposure of β-1,3-glucan in C. albicans.
Exposure of C. albicans to subinhibitory concentrations of caspofungin, enfumafungin A, or nikkomycin Z increases activation of macrophage NF-κB/AP-1 through dectin-1. (a) Caspofungin (CAS); (b) enfumafungin A (ENF-A); (c) nikkomycin Z (NKZ). Anti-dectin-1 antibody, 10 μl. Z, zymosan, 10 μg/ml (TLR2/dectin-1-positive control). Each experiment was performed in triplicate. *, P ≤ 0.05 versus control, Kruskal-Wallis test with Dunn's post hoc test; **, P ≤ 0.05 versus non-anti-dectin-1-exposed macrophages.
Conditional cell wall mutants of C. albicans phenocopy immunopharmacological activity of glucan and chitin synthesis inhibitors.Transcriptional repression of echinocandin (FKS1) or chitin synthesis (CHS3) drug targets should elicit patterns of dectin-1-dependent macrophage activation similar to those of drug-exposed fungi. Macrophages exposed to homozygous FKS1 and CHS3 conditional mutants grown under repressing conditions (100 μg/ml tetracycline) induced patterns of dectin-1-dependent macrophage activation similar to those of cells exposed to glucan synthesis inhibitors (Fig. 2). Similar experiments performed with heterozygous mutants did not elicit increased macrophage activation, confirming that the homozygous deletion background was required to phenocopy immunopharmacological effects of exposure to glucan synthesis inhibitors or nikkomycin Z in the assay. These results suggested that the immunopharmacological effects of the antifungals result from direct target inhibition, rather than pleiotropic nonspecific mechanisms on the fungal cells.
Tetracycline repression of homozygous CHS3 and FKS1 in C. albicans phenocopies growth in the presence of nikkomycin Z and caspofungin. Tetracycline concentration, 100 μg/ml; anti-dectin-1 antibody, 10 μg/ml; HET, heterozygous conditional mutant; HOM, homozygous conditional mutant; WT, wild type; CHS3, chitin synthase 3; FKS1, β-1,3-glucan synthase subunit; Z, zymosan, 10 μg/ml (TLR2/dectin-1-positive control). Experiments were performed with four replicates. *, P ≤ 0.05 versus control, Kruskal-Wallis test with Dunn's post hoc test; **, P ≤ 0.05 versus non-anti-dectin-1-exposed macrophages.
The degree of macrophage NF-κB/AP-1 transcriptional activation by homozygous mutants correlated with degree of tetracycline-mediated repression in the medium (Fig. 3). Graded increases in tetracycline proportionally increased NF-κB/AP-1 activation. No evidence of a paradoxical decrease in inflammatory responses (as was observed with higher concentrations in drug experiments) was evident over the range of tested tetracycline concentrations.
Graded NF-κB/AP-1 transcriptional activation against CAI4-GRACE-FKS1 grown in increasing concentrations of tetracycline in culture medium. Experiments were performed in four replicates. Z, zymosan, 10 μg/ml (TLR2/dectin-1-positive control).
Glucan and chitin synthesis inhibitors enhance macrophage inflammatory responses to A. fumigatus.A. fumigatus hyphae grown in the absence of antifungal agents caused minimal increases in macrophage activation (OD655 of <0.2 for SEAP) relative to negative controls. In contrast, hyphae grown in the presence of caspofungin or enfumafungin (≥0.5× MEC) elicited a 4- to 5-fold increase in NF-κB/AP-1 transcriptional activation in macrophages (P < 0.05) relative to that of non-drug-exposed fungi (Fig. 4). This activation was blocked by anti-dectin-1 antibodies. Similar to studies with C. albicans, the magnitude of dectin-1-dependent macrophage activation decreased with higher caspofungin and enfumafungin A concentrations. In contrast to studies with CAI4, no increase in macrophage transcriptional activity was observed when A. fumigatus CEA17 was grown in the presence of nikkomyzin Z.
Exposure of A. fumigatus to increasing concentrations of caspofungin or enfumafungin A, but not nikkomycin Z, increases activation of macrophage NF-κB/AP-1 through dectin-1. (a) Caspofungin (CAS); (b) enfumafungin A (ENF-A); (c) nikkomycin Z (NKZ). Anti-dectin-1 antibody, 10 μl; Z, zymosan, 10 μg/ml (TLR2/dectin-1-positive control).*, P ≤ 0.05 versus control, Kruskal-Wallis test with Dunn's post hoc test; **, P ≤ 0.05 versus non-anti-dectin-1-exposed macrophages.
The conditional glucan synthase mutant of A. fumigatus phenocopies immunopharmacological activity of glucan synthesis inhibitors.Transcriptional repression of FKS1 caused by ammonium tartrate in growth medium induced patterns of dectin-1-dependent macrophage activation similar to those of hyphae grown in the presence of caspofungin or enfumafungin A (Fig. 5). Similar to the studies with nikkomycin Z-exposed hyphae, conditional repression of CHS2 was not associated with increased macrophage transcriptional NF-κB/AP-1 activation.
Nitrogen repression of FKS1 in A. fumigatus phenocopies growth in the presence of caspofungin or enfumafungin. Repressing conditions were induced by addition of 10 mM ammonium tartrate to Aspergillus minimal medium. Anti-dectin-1 antibody, 10 μg/ml; WT, wild type; CHS2, chitin synthase isoenzyme 2; FKS1, β-1,3-glucan synthase subunit; Z, zymosan, 10 μg/ml (TLR2/dectin-1-positive control). Experiments were performed in four replicates. *, P ≤ 0.05 versus control, Kruskal-Wallis test with Dunn's post hoc test; **, P ≤ 0.05 versus non-anti-dectin-1-exposed macrophages exposed to isolates grown under repressing conditions.
DISCUSSION
Studies in antimicrobial pharmacology typically focus on the ability of the drug to inhibit a specific target and direct phenotypic effects of the target inhibition (15). While these biochemical concepts are central to understanding the intrinsic activity of antifungal agents, direct or indirect interactions with the host immune response also shape the efficacy and safety of the antifungals in vivo (16). Indirect immunological effects may be especially important for cell wall-active antifungals, such as the echinocandins. For example, despite their lack of appreciable in vitro activity against a number of clinically important molds (Mucorales, Fusarium, and Scedosporium), echinocandins have been found to be modestly effective in vivo, and sometimes highly effective in combination with other agents, in experimental models of mucormycosis (17–19), fusariosis (20), and scedosporiosis (21). Counterintuitively, lower doses of echinocandins were generally more effective than higher doses in these animal infection models. This unusual dose-response pattern may be a signature of indirect immunopharmacological effects in vivo and/or limited activation of fungal compensatory responses to echinocandin exposure at lower doses that have been shown (at least with Candida albicans) to be associated with diminished activation of host immune cells (22–24).
Given the potential importance of these immunological effects, we sought to develop a simple assay for screening antifungal-mediated modulation of inflammatory responses that could provide information complementary to conventional assays typically used to screen antifungal activity. To this end, we developed a cell-based assay with a commercially available mouse macrophage reporter cell where transcriptional activation of NF-κB/AP-1 induces production of SEAP, an easily measured reporter frequently employed in high-throughput assays.
In agreement with previous studies (3, 13), we found that echinocandin-treated C. albicans and A. fumigatus increased dectin-1-mediated inflammatory responses in macrophages. This enhanced inflammatory response with caspofungin occurred over a range of drug concentrations that have been observed in vivo with 50- to 70-mg doses of caspofungin (100 to 10,000 ng/ml) (25). New findings were that we also observed similar patterns of dectin-1-dependent macrophage activation in C. albicans and A. fumigatus cells treated with enfumafungin A, a semisynthetic, nonechinocandin glucan synthesis inhibitor synthesized from an antimicrobial peptide produced by the fungal endophyte Hormonema species (26). We also found that C. albicans treated with low concentrations of nikkomycin Z, a specific inhibitor of chitin synthase (CHS3) with relatively poor inhibitory activity against Candida albicans (27), elicited a modest but significant increase in dectin-1 macrophage activation in C. albicans in a fashion similar to that of glucan synthase inhibitors. It is possible that exposure to higher concentrations of nikkomycin Z induces activation of other chitin synthase isozymes in Candida (CHS2 and CHS1), reversing the inhibitor's effects at higher concentrations (27). Macrophage inflammatory responses were largely unchanged in A. fumigatus exposed to nikkomycin Z.
Notably, C. albicans and A. fumigatus isolates with conditional mutations in glucan (FKS1) and chitin (CHS3) synthase elicited macrophage inflammatory responses in a fashion similar to that of drug treatment when the mutants were grown under repressing conditions. These results suggest that a macrophage reporter cell-based assay could be used as an immunopharmacological screen to complement typical static or fungicidal phenotypic endpoints of candidate drug molecules or libraries of isolates with different mutations affecting cell wall synthesis.
Although the assay is relatively simple and can be performed in any laboratory with cell culture facilities, a couple of limitations should be highlighted. Testing of Candida cells was complicated by the dimorphism of the fungus, which elicits different macrophage responses depending on cell morphology and in the presence of fungicidal drugs. Therefore, it was necessary to standardize the inoculum by protein concentration and include replicate tubes to account for lower fungal inoculum at test concentrations near the MIC. Similar protein or biomass standardization approaches would be required for testing fungicidal agents against Aspergillus. Culture medium composition (i.e., nutrient-rich versus -poor medium) can also profoundly influence the dynamic and adaptable composition of Candida and Aspergillus cell walls (28, 29), which also alters macrophage responses. Consequently, test medium needs to be carefully controlled and individualized to the pathogens of interest and the study objective. Finally, although transcriptional activation of NF-κB/AP-1 was simple to monitor in this assay and often parallels cytokine production, it does not provide information regarding specific immune effector mechanisms such as the generation of reactive oxygen species.
In summary, we describe the development of a simple, cell-based reporter assay suitable for screening indirect immunopharmacological effects of antifungal agents in Candida albicans and Aspergillus fumigatus. This assay has provided interesting insight into how FKS1 mutations associated with echinocandin resistance in A. fumigatus (30) and C. albicans (6) in vivo may impact treatment outcome and organism persistence and fitness in vivo. Therefore, this assay could be a useful complementary screen in the evaluation of new antifungal agents or mutation libraries or studies of intrinsically resistant or adaptively resistant pathogens in vitro and in vivo.
ACKNOWLEDGMENTS
This project was funded by Merck & Co., Inc. K.Y. and C.D. are employees of Merck & Co., Inc. D.P.K. acknowledges the Francis King Black Endowment.
FOOTNOTES
- Received 7 November 2013.
- Returned for modification 29 November 2013.
- Accepted 28 December 2013.
- Accepted manuscript posted online 6 January 2014.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/AAC.02408-13.
- Copyright © 2014, American Society for Microbiology. All Rights Reserved.
