Disulphide-reduced psoriasin is a human apoptosis-inducing broad-spectrum fungicide

Significance

Fungi increasingly cause serious medical problems in immunocompromised populations. Antimicrobial peptides are primary effector molecules of innate immune systems. Antimicrobial peptides successfully protect healthy humans from bacterial infections. However, it is largely unknown how and why human body surfaces resist fungal infections. We identified the common epithelial protein, psoriasin (S100A7), in its disulphide-reduced form (redS100A7) as the principal antifungal factor of human body surfaces. redS100A7 kills several pathogenic fungi using a mechanism that differs from conventional antifungal agents. Thus, this study might contribute to a better understanding of human defense systems against fungal infection and the development of urgently needed novel antifungal therapeutics.

Abstract

The unexpected resistance of psoriasis lesions to fungal infections suggests local production of an antifungal factor. We purified Trichophyton rubrum-inhibiting activity from lesional psoriasis scale extracts and identified the Cys-reduced form of S100A7/psoriasin (redS100A7) as a principal antifungal factor. redS100A7 inhibits various filamentous fungi, including the mold Aspergillus fumigatus, but not Candida albicans. Antifungal activity was inhibited by Zn²⁺, suggesting that redS100A7 interferes with fungal zinc homeostasis. Because S100A7-mutants lacking a single cysteine are no longer antifungals, we hypothesized that redS100A7 is acting as a Zn²⁺-chelator. Immunogold electron microscopy studies revealed that it penetrates fungal cells, implicating possible intracellular actions. In support with our hypothesis, the cell-penetrating Zn²⁺-chelator TPEN was found to function as a broad-spectrum antifungal. Ultrastructural analyses of redS100A7-treated T. rubrum revealed marked signs of apoptosis, suggesting that its mode of action is induction of programmed cell death. TUNEL, SYTOX-green analyses, and caspase-inhibition studies supported this for both T. rubrum and A. fumigatus. Whereas redS100A7 can be generated from oxidized S100A7 by action of thioredoxin or glutathione, elevated redS100A7 levels in fungal skin infection indicate induction of both S100A7 and its reducing agent in vivo. To investigate whether redS100A7 and TPEN are antifungals in vivo, we used a guinea pig tinea pedes model for fungal skin infections and a lethal mouse Aspergillus infection model for lung infection and found antifungal activity in both in vivo animal systems. Thus, selective fungal cell-penetrating Zn²⁺-chelators could be useful as an urgently needed novel antifungal therapeutic, which induces programmed cell death in numerous fungi.

Fungi cause emerging infectious diseases that increasingly threaten human health (1–3). Superficial fungal infections affect 20–25% of the global population and opportunistic fungal infections cause serious medical problems, with high morbidity and mortality, particularly in immunocompromised patients (3, 4). Filamentous fungi are composed of a conidium, an asexual spore, and a hypha, a long, branching filamentous structure. Conidia, which are ubiquitously spread in soil and air, represent the primary infectious unit of fungi. Therefore, it is amazing that the permanent exposure and colonization of our body surfaces by various fungi (5) does not usually cause infections in healthy individuals. Surprisingly, it is largely unknown how and why human body surfaces resist fungal pathogens.

Healthy human lungs are highly efficient at clearing airborne fungal spores without causing lung inflammation, suggesting that innate defense strategies to control fungal pathogens do exist in the epithelium (6). Epithelial antimicrobial peptides are the candidate effector molecules that could play a role in defending the body against fungal infections. Although the disulphide-reduced form of human β-defensin-1 (hBD-1) shows—apart from its bactericidal activity—strong activity against Candida albicans (7), there is no systematic study investigating antifungals with human epithelial origin that might control the growth of filamentous fungi at body surfaces.

To address this important question, we analyzed lesional skin from patients with psoriasis—a skin disease with an unexpected resistance to fungal infections (8)—in an attempt to identify human antifungals (9, 10).

Results

Identification of Disulphide-Reduced Psoriasin as Antifungal Protein.

Human epithelia show a marked resistance toward bacterial and fungal infections, particularly skin lesions of psoriasis patients, which are unexpectedly rarely affected (8, 10). We hypothesized that antifungal activity is generated in psoriasis lesions. We analyzed psoriatic scale extracts and found Trichophyton rubrum growth inhibiting activity (Fig. 1A). The antifungal agent was purified by several HPLC steps to homogeneity (Fig. S1).

Fig. 1.

Identification of reduced S100A7 (redS100A7) as an antifungal protein. (A) Psoriatic scale extracts contain T. rubrum antifungal activity (n = 3 *P < 0.001). (B) SDS/PAGE and Western blot (WB)-analyses of the antifungal protein purified from the psoriasis scales. (C) redS100A7 is a potent antifungal protein compared with oxS100A7, reduced and alkylated S100A7 (red/alkS100A7), and S-sulfito-S100A7 (sulfS100A7). *P < 0.01, **P < 0.007 and ***P < 0.001. (D) redS100A7 decreases cell viability in T. rubrum dose-dependently. Viable fungi were detected by the MTT tetrazolium salt colorimetric assay. n = 3 independent experiments. Mean ± SD is shown. *P < 0.008. (E) Partial ascorbic acid (AA) sequences of redS100A7 and the three mutants. (F) Whereas antifungal activity is absent in all three Cys-Ala-mutants (*P < 0.002), antibacterial activity is still present. A minimum of three independent experiments was performed and the average ± SD is plotted.

Fig. S1.

Purification of the antifungal protein from psoriatic scales and properties. (A) Psoriatic scale extracts (molecular weight range: 3–50 kDa) contain T. rubrum antifungal activity (n = 3 independent experiments, *P < 0.0004). Psoriatic scale extracts (molecular weight range: 3–50 kDa) were first separated by heparin-affinity chromatography (B), followed by C₈ RP-HPLC of heparin-binding proteins (C). The fraction containing the highest antifungal activity was further purified by cation-exchange chromatography (D). As a final purification step, C₁₈ RP-HPLC was used (E). Extracts and HPLC fractions were tested in a microbroth dilution assay. (F) The molecular mass of the antifungal protein was analyzed by MALDI-MS using two internal calibration standards (IS). (G) The HPLC-purified protein fraction was incubated with a neutralizing monoclonal S100A7 antibody in phosphate buffer saline for 1 h. The mixture was tested against T. rubrum in a microbroth dilution assay. The S100A7 antibody (ab45091, Abcam) blocked antifungal activity of the purified protein preparation (n = 3 independent experiments, **P < 0.002). (H) ESI-MS analysis of Zn²⁺-bound redS100A7. Zinc-treated redS100A7 was analyzed by ESI-MS in ammonium formate buffer at pH 5.5. We found a peak at 22,958 Da, which indicates a dimer of red100A7 containing zinc and buffer ions and likely H₂O. (I) Antifungal activity of redS100A7 is pH-dependent. Note that redS100A7 works optimally between pH 4 and 7 and is sensitive to alkaline pH. n = 3 independent experiments, *P < 0.0001. Mean ± SD is shown. Statistical analysis was performed with two-tailed Student’s t test. Antifungal assays were performed in duplicate and repeated at least three times.

SDS/PAGE, Western blot analyses (Fig. 1B), protein-sequencing, and MALDI-MS analyses revealed that the antifungal protein was the reduced form of psoriasin, S100A7. Inhibition of antifungal activity by psoriasin antibodies (Fig. S1G) excluded a copurifying contaminant.

Two Free Cysteines of S100A7 Are Essential for Antifungal Activity.

Different bands upon SDS/PAGE analyses revealed easy air oxidation of S100A7 (Fig. 1B). Reduction of the single disulphide bridge in oxS100A7 results in two free-thiol groups of the cysteine residues, Cys46 and Cys95. To elucidate the involvement of these thiol groups in the antifungal effect, we analyzed the reduced and alkylated S100A7 (red/alkS100A7) and found very low antifungal activity (Fig. 1C), suggesting that free thiols are essential. Viability tests after redS100A7-treatment of T. rubrum indicated that it acts as a fungicidal agent (Fig. 1D). We then generated S100A7 mutants, in which either one or both cysteines were substituted with alanine (Fig. 1E). All three mutants, the Cys95Ala-S100A7 (mutant 1), Cys46Ala-S100A7 (mutant 2), and Cys46Ala, Cys95Ala-S100A7 (mutant 3) did not show any inhibition of T. rubrum growth (Fig. 1F). Interestingly, the growth of Escherichia coli was similarly inhibited, as seen with oxS100A7 (11) (Fig. 1F), corroborating that both free thiols are essential for antimycotic but not bactericidal activity.

redS100A7 Chelates Zn²⁺ Ions in Fungi.

oxS100A7 is highly abundant on healthy human skin and several mucosal surfaces (11–13). Its antibacterial activity is Zn²⁺-sensitive, acting via a histidine-coordinated low-affinity Zn²⁺-binding site (14). Correspondingly, we found that Zn²⁺ completely inhibited the antifungal activity of redS100A7 (Fig. 2A), whereas other divalent ions did not reduce antifungal activity.

Fig. 2.

redS100A7 sequesters Zn²⁺ ions in fungal cells. (A) redS100A7 antifungal activity is inhibited by addition of Zn²⁺ (*P < 0.05, **P < 0.001 vs. 0 μM Zn²⁺). (B) CD spectroscopy of redS100A7 at pH 5.4 in the presence or absence of Zn²⁺. Note changes in α-helical content characterized by a deep minimum between 209 and 222 nm. (C) In redS100A7-treated T. rubrum, intracellular liable Zn²⁺ are stained with a specific fluorescent probe, Zinquin. (Magnification: 300×.) (D) Antifungal activity (T. rubrum) of different Zn²⁺-chelators. TPEN is a cell-permeable chelator, and DTPA and EDTA are cell-impermeable chelators (*P < 0.002, **P < 0.001). A minimum of three independent experiments was performed and the average ± SD is plotted.

With few exceptions, protein zinc-binding sites are Cys, His, or Glu/Asp (15). To elucidate whether opening the disulphide bond of oxS100A7 causes secondary structure changes, we performed circular dichroism (CD) spectroscopy. Without exposure to Zn²⁺, redS100A7 displayed a CD spectrum of α-helical structural elements, which is similar to that of oxS100A7 (16) (Fig. 2B). However, Zn²⁺-exposure causes a shift of the CD spectrum (Fig. 2B), supporting the hypothesis that Zn²⁺ produces a conformation change. Because the low-affinity Zn²⁺-binding site of oxS100A7 is formed via four His of an oxS100A7 dimer (14), we analyzed Zn²⁺-treated redS100A7 through electrospray ionization-mass spectrometry (ESI-MS) and found a mass of 22,958 Da, corresponding to a redS100A7 dimer plus Zn²⁺ions (Fig. S1H). This finding supports the hypothesis that four thiols in two redS100A7 molecules bind Zn²⁺, suggesting formation of a new Zn²⁺-binding site in redS100A7. This hypothesis was corroborated by the finding that Ca²⁺ augmented the antifungal efficacy of redS100A7 (Fig. S2A), because Ca²⁺ increases the Zn²⁺-binding affinity of the S100A8/A9 complex calprotectin (17).

Fig. S2.

Calcium (Ca²⁺)-dependent antifungal activity of redS100A7. (A) T. rubrum conidia were incubated with redS100A7 in the presence or absence of a 10 mM Ca²⁺ supplement. Fungal growth was measured at the optical density of 595 nm. Sterile saline was used as control. Addition of Ca²⁺ increased the efficacy of redS100A7 antifungal activity (n = 3 independent experiments, *P < 0.002 vs. Ca ²⁺-free control). (B) Ca²⁺ alone showed no effect on the fungal growth up to 50 mM. CaCl₂ was used as Ca^2+-source. Mean ± SD is shown. Statistical analysis was performed with two-tailed Student’s t test. (C) 5 mM of AA blocked the antifungal activity of redS100A7 (**P < 0.001). (D) Fungal Cu/Zn-SOD enzyme activity was measured in the presence or absence of redS100A7. (ns, not significant). (E) redS100A7-induced antifungal activity was not inhibited by ergosterol, in contrast to the control antimycotic amphotericin B (Amp-B) at 500 μM ergosterol. (***P < 0.0007). Mean ± SD is shown. Statistical analysis was performed with two-tailed Student’s t test. Data are representative of three independent experiments.

To examine whether redS100A7 chelates Zn²⁺ in fungi, we incubated redS100A7-treated T. rubrum with Zinquin, a cell-permeable fluorescent probe specific for Zn²⁺ ions (18, 19), and observed an alteration in Zinquin staining pattern (Fig. 2C). This finding suggests that redS100A7 results in altered subcellular distribution of labile zinc. Together with the data presented in Fig. 2 A and B, this suggests that redS100A7 acts as a natural Zn²⁺-chelator. To test this hypothesis, we investigated synthetic Zn²⁺-chelators, and identified the cell-membrane–permeable Zn²⁺-chelator TPEN [N,N,N′,N′-Tetrakis(2-pyridylmethyl)- ethylenediamine] (20) as a potent antifungal (Fig. 2D).

redS1007 and TPEN Are Broad-Spectrum Fungicides.

Next we analyzed the effects of redS100A7 on other fungi. redS100A7 inhibited the growth of several filamentous fungi and yeasts, such as Aspergillus fumigatus, Malassezia furfur, Microsporum canis, Rhizopus oryzae, Saccharomyces cerevisiae, and Trichophyton mentagrophytes with a 90% minimal inhibitory concentration (MIC₉₀) of ∼2 μM (Fig. 3A). However, the yeast C. albicans was not affected at concentrations up to 20 μM, which is surprising because C. albicans is killed by calprotectin in a Zn²⁺-dependent fashion (21). We identified TPEN (20) as a broad-spectrum antifungal agent that kills a high variety of filamentous fungi and yeasts, including C. albicans, at a MIC₉₀ of ∼2 μM for all tested fungi (Fig. 3B), which supports our hypothesis that redS100A7 acts as a natural fungus cell-penetrating Zn²⁺-chelator.

Fig. 3.

Action mechanism of redS100A7. (A) Broad-spectrum antifungal activity of redS100A7 against various fungi. (B) Broad-spectrum antifungal activity of TPEN. (C) Intracellular ROS (determined with the fluorescent probe carboxy-H2DCFA) were dose-dependently raised in redA100A7-treated conidia. The ROS-inhibitor AA prevented the ROS-generation (4 h, *P < 0.0003, n = 3 independent experiments). (D) TUNEL assay (to observe DNA fragmentation) and SYTOX-green staining (to observe disintegrated plasma membrane) of redS100A7-treated (4 μM) T. rubrum. (Scale bars, 20 μm.) (E) TUNEL assay and SYTOX-green staining of redS100A7-treated (4 μM) A. fumigatus. (Scale bars, 20 μm.) (F) TEM images of redS100A7 (4 μM)-treated T. rubrum. Arrows indicate electron-dense depositions and arrowhead represents dilated nucleolemmal cisterns. (Scale bars, 0.2 μm.) (G and H) SEM analyses of redS100A7-treated (4 μM) A. fumigatus (G) and controls (H). Note destroyed fungal spores (conidia), premature conidia and conidia-producing phialides (circle). (Scale bars, 2 μm.) CM, cell membrane; CW, cell wall; M, mitochondria; N, nucleus; R, ribosomes; V, vacuole. (I) A cell-permeant pan caspase inhibitor blocks the antifungal activity of redS100A7 in a dose-dependent manner. (n = 3 independent experiments, *P < 0.02.)

redS100A7 and TPEN Kill Fungi by Inducing Apoptosis-Like Cell Death.

To analyze the mechanism underlying redS100A7 antimycotic activity, redS100A7-treated T. rubrum conidia were analyzed for intracellular reactive oxygen species (ROS) levels. Levels increased in a dose-dependent manner (Fig. 3C) and inhibition of the activity by a ROS-inhibitor ascorbic acid (Fig. S2C) suggest that antifungal activity is at least in part ROS-mediated. We excluded the fungal copper/zinc superoxide dismutase (Cu/Zn-SOD) (22), the SOD enzyme (Fig. S2D), and the fungal sterol biosynthesis (Fig. S2E), as primary targets of redS100A7 antifungal action.

To investigate whether redS100A7 causes apoptosis in fungal cells, DNA degradation was visualized using a TUNEL assay. After 4 h of exposure, redS100A7-treated T. rubrum and A. fumigatus showed TUNEL⁺ phenotypes representing apoptotic DNA breaks (Fig. 3D, Left). In T. rubrum cells, 24-h exposure to redS100A7 affected the plasma membrane integrity, as evidenced by a strong, intact plasma membrane-impermeable SYTOX-green nucleic acid staining in mycelia (Fig. 3D, Right). Similar findings were obtained when T. rubrum and A. fumigatus were treated with TPEN (Fig. 3 D and E).

redS100A7- and TPEN-Treated Fungi Show Signs of Apoptosis-Like Cell Death.

Further support of apoptosis induction in fungal cells by both redS100A7 and TPEN came from ultrastructural analyses by transmission electron microscopy (TEM). We found dark nuclei with dilated nucleolemmal cisterns, degraded mitochondria, electron-dense depositions, lipid droplets, and blebs in the plasma membrane in T. rubrum treated with redS100A7 for 24 h (Figs. 3F and 4C).

Fig. 4.

Target localization of redS100A7. (A) redS100A7-immunofluorescence visualization (arrows) on T. rubrum by confocal laser microscopy. Note redS100A7 accumulation on conidia (arrows). (Scale bars, 50 μm.) (B) Confocal laser microscopy of 2 h (Left) or 4 h (Center)-treated A. fumigatus. Arrows indicate the location of redS100A7 on conidia at 2 h and hyphae at 4 h. (Scale bars, 50 μm.) (C) Immunogold labeling electron microscopy (Upper and Lower). Arrows point toward gold particles that represent redS100A7 immune complexes. Arrowhead indicates gold particles accumulating within the conidia. (Scale bars, 0.2 μm.) C, conidium; CM, cell membrane; CW, cell wall; H, Hyphae; L, lipid droplet; M, mitochondria; N, nucleus; R, ribosomes; V, vacuole. Independent experiments have been repeated at least three times and showed similar results.

We then investigated the ultrastructure of redS100A7- and TPEN-treated A. fumigatus by scanning electron microscopy (SEM). redS100A7 and TPEN treatments cause shrunken filaments and damaged fungal spores (conidia), premature conidia, and open-ended amphora-like conidia-producing phialides (Fig. 3 G and H and Fig. S3).

Fig. S3.

SEM of A. fumigatus. A. fumigatus conidia were cultured on a Sabouraud dextrose agar for 7 d. The agar was cut into 25 mm² and incubated with redS100A7 (A) or TPEN (B) for 24 h. Sunken filaments, fungal cell debris, and damaged conidia were observed in both, redS100A7- and TPEN-treated fungi.

To confirm whether redS100A7 induces apoptotic cell death in fungi, we coincubated redS100A7 with a cell-permeant pan caspase inhibitor (Z-VAD-FMK) (23) and found dose-dependent inhibition of the redS100A7-induced cell death in T. rubrum, suggesting that redS100A7 kills fungi via a caspase-dependent apoptosis pathway (Fig. 3I).

Next, we determined the target location of redS100A7-treated T. rubrum and A. fumigatus. These fungi were incubated with S100A7 antibodies and then analyzed using immunofluorescence (IF) confocal microscopy and immunogold TEM. IF-S100A7 was seen to accumulate at the surface of and within conidia of T. rubrum (Fig. 4A), as well as in the conidia and along the hyphae of A. fumigatus (Fig. 4B). Ultrastructural analyses by immunogold TEM revealed redS100A7 along the outer surface of the cell wall, as well as penetrating the fungal cell near a conidium or within conidia (Fig. 4C and Fig. S4). Thus, it suggests that redS100A7 first accumulates at the conidium surface and then penetrates fungal membranes close to the conidium, corroborating that conidia are the primary target of redS100A7.

Fig. S4.

redS100A7 accumulates primarily at conidia. (A) Confocal laser microscopic immunofluorescence analyses with S100A7 antibodies reveal after 4-h incubation preferential accumulation of redS100A7 (arrows) at conidia surfaces of T. rubrum and A. fumigatus. (Scale bars, 20 μm.) (B) Immunogold TEM images of redS100A7-treated T. rubrum. The possible T. rubrum-entry site of redS100A7 after 4 h treatment. (Scale bar, 0.5 μm.) (C) A boxed area in B is enlarged in C. Gold particles (which represent redS100A7) inside the cells, close to an area missing the cell wall (arrow). (Scale bar, 0.5 μm.) (D and E) Gold particles (arrowheads) were seen along the outer cell wall. Intact cell wall and intracelluar organelles were only detected in a time period when gold particles were outside the fungal cell. (Scale bar, 0.5 μm.) (F) The Cys46Ala-S100A7 mutant (mutant) is able to penetrate the fungal cell membrane. T. rubrum, grown on an agar plate, was treated with the antimycotically inactive Cys46Ala-S100A7 for 24 h. After pretreatment of T. rubrum with the mutant psoriasin, fungi were treated with a polyclonal S100A7 antibody (which recognized the mutant S100A7, mutant1, upon immunodot analysis) and then analyzed by confocal laser-scanning microscopy for immunofluorescence. Note the focal accumulation of the psoriasin mutant (arrows) close to conidia. (Scale bars, 50 μm.) CM, cell wall; CW. cell membrane; M, mitochondria; N, nucleus; Nu, nucleolus; V: vacuole. (G) Quantitation of redS100A7 gold particles in the conidia and hyphae. (n = 3, *P < 0.001).

redS100A7 Is Generated During Dermatophyte Skin Infection.

To investigate whether redS100A7 acts in vivo, we investigated S100A7 levels in the stratum corneum of healthy and dermatophyte-infected skin. The amounts of S100A7 increased threefold in the dermatophyte-infected patients (Fig. S5A). To determine the redox status of S100A7, stratum corneum extracts from both groups were immediately alkylated and then analyzed by MALDI-MS. Markedly increased amounts of redS100A7, compared with healthy controls, were seen in lesional dermatophyte-infected skin (Fig. S5B), indicating an induction and local reduction of oxS100A7 to redS100A7 in response to fungal infections.

Fig. S5.

Disulphide reduction of S100A7 protein in vivo. (A) S100A7 levels in healthy human skin (24.2 ± 5.3 U, n = 4) and dermatophyte-infected skin (64.2 ± 5.6 U, n = 5, P < 0.0001 versus healthy; Student’s t test), determined by RP-HPLC. (B) Redox status of S100A7 in healthy and infected skin. The ratios of oxS100A7/ redS100A7 in healthy (n = 4) and dermatophyte-infected skin scales (n = 5) were 100/85.1 ± 5.7 and 100/154.0 ± 6.1, respectively (redS100A7: P < 0.0001 vs. healthy; Student’s t test). redS100A7α, mono–CAM-S100A7; redS100A7β, bi–CAM-S100A7. (C) The TRX-R system reduces the disulphide-bridge of oxS100A7. The reduction was analyzed by RP-HPLC. A minimum of three independent experiments was performed and the average ± SD is plotted.

Thioredoxin and Reduced Glutathione Generate redS100A7.

Next we analyzed the mechanism of redS100A7 generation. Epidermis has a very high antioxidant capacity because of high levels of reduced glutathione (GSH) (24) and thioredoxin (TRX) (7, 24). In gut mucosa and skin, the TRX-reductase (R) system regulates physiologically relevant oxidoreduction (25, 26). We found that the TRX-R system dose-dependently increased the amount of redS100A7 (Fig. S5C). Interestingly, GSH alone also generated redS100A7 (Fig. S6).

Fig. S6.

The GSH system reduces the oxS100A7. This experiment was performed by using HPLC. Results are representative of three independent experiments.

Sulfitation-Dependent Inactivation of Psoriasin Is Reversed by the TRX System.

Dermatophytes are specialized fungi that live in the stratum corneum, where they subsist on cysteine-rich proteins like keratins. These fungi secrete sulfite to make stratum corneum proteins digestible, forming S-sulfito-cysteines before proteolysis (27). As expected, we found the disulphide-bridge of oxS100A7 to be cleaved, forming the single thiol-group–containing S-sulfito-S100A7 during T. rubrum culture in vitro (Fig. S7). The S-sulfito-S100A7 showed low antifungal activity (Fig. 1C), which supports our finding that a single thiol-containing S100A7 mutant is inactive (Fig. 1F). These experiments indicate that T. rubrum and other sulfite-generating fungi can inactivate S100A7. However, S-sulfito-S100A7 was not detected in the lesional stratum corneum obtained from dermatophyte skin infections (Fig. S5), which suggests that in vivo the host-derived redox system converts it to redS100A7. To address this, we incubated S-sulfito-S100A7 with TRX-R and found its complete reduction toward redS100A7 (Fig. S7).

Fig. S7.

Reduction of S100A7 derivatives. (A) S-sulfito-S100A7 is produced by T. rubrum. oxS100A7 was incubated in T. rubrum culture medium for 24 h. The medium was collected and applied to a C₈ RP-HPLC column. The fraction containing S100A7 derivatives revealed only S-sulfito-S100A7 (11,445.60) as seen by MALDI-MS. (B) S-sulfito-S100A7 is converted toward redS1007 by the TRX-R system. The TRX-R system was able to reduce S-sulfito-S100A7, which shows a partial disproportionation toward oxS100A7 at experimental conditions, back to redS100A7 in a dose-dependent manner. (C) Oxidized GSH (GSSG) inhibits oxS100A7 antifungal activity (*P < 0.02; Student’s t test). Results are representative of three independent experiments.

Surprisingly oxS100A7 is also an antifungal that requires higher concentrations than redS100A7, however (Fig. 1C). One would have expected that complete absence of antifungal activity might be because of the lack of two free-thiol groups in oxS100A7. Therefore, intracellular oxS100A7 within the fungal cells and the endogenously produced intermediate S-sulfito-S100A7 should be reduced in situ toward redS100A7. IF-analyses with the inactive Cys46Ala-S100A7 mutant (which is recognized by the S100A7 antibody) indicate that oxS100A7 should also penetrate fungal cell membranes (Fig. S4F), which suggests that it is reduced within the cell. To test this theory, we added increasing amounts of oxidized GSH (GSSG), a competing disulphide-bond–containing compound, and found a dose-dependent decrease of oxS100A7 antifungal activity (Fig. S7).

redS100A7 and TPEN Are Protective in a T. mentagrophytes Guinea Pig Tinea Pedes Model.

To investigate whether redS100A7 and TPEN also act in vivo, we generated a guinea pig model of tinea pedis (28). An ablated guinea pig foot was treated with redS100A7, TPEN, or the vehicle. The foot was then infected with 1.5 × 10⁷ T. mentagrophytes conidia. After 3 d of infection, the infected areas were analyzed microscopically, and with Periodic acid Schiff (PAS) reagent and Fungiflora Y staining. Fungal invasion into the stratum corneum was recorded. Both redS100A7 and TPEN showed a significant protective effect in the guinea model of T. mentagrophytes infection (Fig. 5 A and B and Fig. S8).

Fig. 5.

In vivo antifungal activity of redS100A7 and TPEN. (A and B) After applied with vehicle, redS100A7 or TPEN, the plantar of guinea pigs was infected with T. mentagrophytes. Fungal invasion was determined histologically (A) and by scoring (B) (see SI Materials and Methods). (n = 4, *P < 0.005, **P < 0.0001). (C) Immunocompromised mice were infected with A. fumigatus and treated with redS100A7, TPEN, or the vehicle. The Kaplan–Meier survival curve comparing untreated control mice (black) with redS100A7-treated (red) and TPEN-treated mice (green) (n = 7 for each group, *P < 0.036 and **P < 0.003 vs. controls; log-rank test) is shown. (D) Histopathology and SEM analyses of mouse lungs infected with A. fumigatus. Arrowheads indicate A. fumigatus. Grocott’s Methenamine Silver (GMS), H&E. A, alveolus; Ad, alveolar duct; B, bronchiole; R, red blood cell; S, septum. (Magnification: D, GMS and H&E, 200×; SEM, 1,700×.)

Fig. S8.

Guinea pig model of tinea pedis. The foot of guinea pig was lightly abraded. Vehicle, redS100A7, and TPEN were applied to the skin and air-dried. The skin was then infected with 0.15 mL of 1 × 10⁸ conidia per milliliter of T. mentagrophytes and covered with a film. The infected sites were then covered with bandages. After 3 d of infection, the bandages were removed. The skin tissue was collected from the feet. The fungal tissues in the skin were observed microscopically after staining with PAS. (Magnification: 100×.)

redS100A7 and TPEN Prevent Death in a Lethal Aspergillus Infection Mouse Model.

Finally, we investigated the in vivo antifungal activity of redS100A7 in a mouse model of Aspergillus lung infection (29), in which immunocompromised mice were infected with 2 × 10⁷ A. fumigatus conidia for 2 consecutive days. Whereas all mice in the control group did not survive the infection, mice in the redS100A7- or TPEN-treatment survived the invasive fungal infection throughout the 7-d observation period (Fig. 5C). Numerous A. fumigatus conidia and hyphae, as well as massive neutrophil infiltrates, were observed in the untreated control lungs after 3 d (Fig. 5D and Fig. S9). In contrast, fungal burdens were hardly found in redS00A7- or TPEN-treated mice (Fig. 5D and Fig. S9). However, lung histology showed massive inflammatory infiltrates, possibly as a result of dead and degenerating hyphae (Fig. 5D and Fig. S9).

Fig. S9.

(A) At the end of 7-d observation period, redS100A7- and TPEN-treated mice were cured from A. fumigatus infection, compared with the vehicle-treated mice at 3-d after infection. Histopathology of mouse lung after A. fumigatus infection and treatment with redS100A7 or TPEN. (B, Top) Untreated, redS100A7-treated and TPEN-treated mice after intranasal challenge with 2 × 10⁷ A. fumigatus conidia for 2 consecutive days (n = 7 mice for each group). Mouse lungs were stained with Grocott's Methenamine Silver (GMS), PAS, and H&E stain. In untreated control lungs, numerous A. fumigatus conidia and hyphae (arrows) were observed. Neutrophils and monocytes were the dominating leukocytes in infected tissue sites. In redS100A7- and TPEN-treated lungs, less fungi were observed (mainly conidia). After a 7-d observation period, redS100A7- and TPEN-treated lungs were cured from A. fumigatus infection. Monocytes and lymphocytes were the dominating leukocytes in the treated lungs.

Discussion

Lesional psoriatic skin contains the Cys-reduced form of psoriasin (S100A7) as a principal antifungal component, which has antifungal activity against various pathogenic fungi. Although the Cys-oxidized form also shows antifungal activity, it is markedly less potent (Fig. 1), a fact that would explain our failure to purify the oxygen-sensitive Cys-reduced form of psoriasin. The finding that Cys-thiol–lacking S100A7 derivatives show less antifungal activity and S100A7 mutants lacking Cys are not fungicidal suggests that free-thiol groups are essential for the antifungal activity of S100A7 and points toward a unique mode of action of redS100A7. Antibacterial activity of oxS100A7 is mainly restricted toward E. coli; far lower activity was seen against Pseudomonas aeruginosa and Staphylococcus aureus (11), and in the oxS100A7 it is inhibited by Zn²⁺ at pH 7.4 (11). This finding suggests that the His-based Zn²⁺-binding sites in the oxS100A7 (14) and in the Cys-Ala- S100A7 mutants are involved in E. coli-cidal activity at pH 7.4. However, at pH 5.5, the normal skin pH, where His-based Zn²⁺-binding sites are inactive, oxS100A7 is bactericidal with a different mode of action, targeting the bacterial membrane by forming pores (16). redS100A7 is a potent antifungal at pH 5.5, but sensitive toward an alkaline pH (Fig. S1I). A plausible explanation for the antifungal activity of redS100A7 at an acidic pH, low stability at alkaline pH and Zn²⁺-sensitivity would be the involvement of a Cys-thiol–based Zn²⁺-binding site. This, however, would need the opening of the single disulphide group in S100A7, causing structural changes upon Zn²⁺-binding (Fig. 2B). Our finding that both Cys-thiols in S100A7 are necessary for activity (Fig. 1) and redS100A7 forms Zn²⁺-binding dimers at acidic pHs (Fig. S1) supports the idea that the Zn²⁺-binding site could be of the (Cys)₄-type (15). Indeed, a previous study revealed inhibition of fungal growth by synthetic zinc chelators and calprotectin (S100A8/A9) (30). Therefore, a synthetic zinc-chelator should have antifungal activity. We tested three zinc chelators and identified the cell-penetrating zinc-chelator TPEN as the most potent antifungal with an ED₅₀ at ∼1 µM, similar as that seen for redS100A7 (Figs. 2 and 3). Because noncell-penetrating zinc chelators are the least-potent antifungals, this would indicate that intracellular zinc-chelation could be the mechanism behind redS100A7-dependent antifungal activity. To test this hypothesis, we investigated redS100A7-treated fungi with the fluorescent zinc sensor Zinquin, which allows monitoring of loosely bound, labile intracellular Zn²⁺ and which does not mobilize tightly bound Zn²⁺ from enzymes (18). We saw a stronger vesicle-associated fluorescence (Fig. 2C), a phenomenon also seen in apoptotic lymphocytes (19), but barely—if any—increase in fluorescence. Perhaps this unexpected finding is caused by the presence of two zinc binding sites with different affinities in redS100A7. In such a case Zinquin fluorescence analyses reveal complicated results, possibly because of the analyses’ inability to report on two K_d values that differ by several orders-of-magnitude (31). One may speculate that redS100A7 enters the fungal cell, sequesters zinc from cytosolic sources, and then migrates to the endosomal compartments where zinc is released from redS100A7 (because of the local endosomal environment), resulting in increased labile zinc, which increases Zinquin fluorescence. The absence of similar fluorescence signals in TPEN-treated fungi is a result of TPEN’s higher affinity toward Zn²⁺, which causes quenching of the Zinquin-fluorescence (18).

We then investigated the mechanism underlying how redS100A7 kills fungi. redS100A7 has been shown to have fungicidal activity (Fig. 1D). Polyene antifungals (e.g., amphotericin B) and azole antifungals (e.g., fluconazole) are commonly used antifungal agents. Polyene antifungals bind to ergosterol in the fungal cell membrane, forming transmembrane channels that lead to monovalent ion leakage (32). Azole antifungals inhibit the fungal cytochrome P450 enzyme, preventing the formation of essential ergosterol (32). We observed that redS100A7 neither interacted with ergosterol nor inactivated fungal Cu/Zn-superoxide dismutases for its antifungal action (Fig. S2). Instead, ultrastructural analyses (Fig. 3F) of redS100A7-treated fungi suggest that it induces apoptosis-like cell death in fungi. Indeed, TUNEL analyses, SYTOX-green analyses, and inhibition of fungicidal activity by a caspase inhibitor (Fig. 3I), as well as the Zinquin-experiments (Fig. 2) corroborated the hypothesis that it induces apoptosis-like cell death in fungi in a zinc-dependent manner. Zinc plays an important role in the regulation of apoptosis. It has been shown that the addition of Zn²⁺-chelators leads to apoptosis-induction in various hepatocytes, as well as in leukemia cells (33, 34). Sequestering inhibitory Zn²⁺ ions in inactive procaspase-3 generates active caspase-3, which subsequently triggers the downstream effector caspase-6 and induces apoptosis (33). We also demonstrated that redS100A7-induced fungal cell death is reversible through a pan caspase inhibitor (Fig. 3I). Therefore, in contrast to conventional antifungal agents, our data suggest that redS100A7 induces apoptosis-like fungal cell death, possibly via activation of the metacaspase system (i.e., the caspase system of fungi, plants, and protozoa) by sequestering zinc ions. However, further analyses are needed to prove this hypothesis.

Ultrastructural analyses (Fig. 3 F–H) suggest that the main target of redS100A7 could be conidia, the asexual nonmotile spores of a fungus. This finding was supported by both immunofluorescence-staining and immunogold detection of S100A7, accumulating at or close to conidia in T. rubrum and A. fumigatus (Fig. 4). Kinetic analyses revealed that redS100A7 is penetrating the fungal cell close to a conidium, then diffusing within the cell and eventually accumulating within the conidium (Fig. 4). It is tempting to speculate that the highly hydrophobic S100A7 accumulates at conidia of filamentous fungi because of the presence of hydrophobins (35, 36). The absence of hydrophobins in C. albicans (36) would explain the failure of redS100A7 to kill this yeast. We investigated the sensitivity of a hydrophobin mutant of A. fumigatus (37) but found no difference in sensitivity (Fig. S10), which suggests that either hydrophobin is not the target of S100A7 or other hydrophobins are still present.

Fig. S10.

redS100A7 inhibits the growth of hydrophobin-mutant A. fumigatus. A. fumigatus ΔKU80 (parent strain) (final 1 × 10⁵/mL) and hydrophobin-mutant A. fumigatus ΔKU80 ΔrodA (final 1 × 10⁵/mL) were cultured in the 96-well plate. Serial dilutions of redS100A7 were added to the assay medium. Fungal growth was measured at the absorbance of 595 nm. n = 3 independent experiments. Mean ± SD is shown. Both A. fumigatus ΔKU80 (parent strain) and A. fumigatus ΔKU80 ΔrodA were kindly provided by J. P. Latgé, the Aspergillus Unit at the Pasteur Institute in Paris, France.

If redS100A7 is an important antifungal defense component of human skin, one would expect it to form during fungal infection. Indeed, stratum corneum of patients with tinea pedes, a dermatophyte skin infection of the feet, revealed increased amounts during infection (Fig. S5). This finding was unexpected because dermatophytes are cleaving disulphide-bridges with sulfite, forming sulfitocysteines (27), as shown in oxS100A7 (Fig. S7). Skin has a negative redox potential (24), containing high amounts of reduced GSH (24) and TRX (7, 25). Because both redox-compounds are able to convert oxS100A7 and sulfito-S100A7 into antifungal redS100A7 (Figs. S5–S7), skin has an efficient counter-strategy to compensate dermatophyte’s inactivation of S100A7. Thus, the presence of the TRX-R system, reduced hBD-1 (7), and GSH (24) in healthy skin suggests that S100A7 is also present in its reduced state.

S100A7 Cys-Ala mutants and other derivatives indicate that free Cys thiols are essential for antifungal activity (Fig. 1). Therefore, it is surprising that oxS100A7, which lacks free-thiol groups, is an antifungal. A possible explanation would be its disulphide-bridge reduction within the fungal cell by a fungal oxidoreductase. Although direct proof has yet to be collected, this hypothesis is supported by a dose-dependent inhibition of antifungal activity in the presence of oxidized GSH as a competing, disulphide-containing substrate (Fig. S7C).

redS100A7 and TPEN were identified as broad-spectrum antifungals in vitro (Fig. 1). To investigate whether both are antifungals in vivo, we developed a guinea pig tinea pedes skin model system, allowing us to treat an experimental superficial dermatophyte infection with our agents of interest. Our results show that topical administration of either redS100A7 or TPEN prevents fungal infection in vivo (Fig. 5 A and B). In addition, in a mouse model of Aspergillus lung infection, in which immunocompromised mice were infected with a lethal dose of A. fumigatus conidia, redS100A7 or TPEN treatment prevented death from invasive fungal infection (Fig. 5 C and D). Therefore, redS100A7 and fungus cell-penetrating zinc chelators might have a therapeutical potential as novel antifungals. Lethal fungal infections are a growing threat to humans because of the increasing number of immunocompromised individuals, such as those with HIV, cancer, and transplant recipients (1–4, 38). Many antifungal agents are limited by a narrow spectrum, as well as by the development of drug resistance and toxicity, engendering an urgent need for innovative antifungal agents (1, 38).

In summary, our data represent a previously undescribed mechanism of action for an antimicrobial protein. We propose that the Cysteine-reduced form of the ubiquitous epithelial protein psoriasin (S100A7) acts as a principal human antifungal protein that induces apoptosis in fungi by penetrating the fungal cell membrane and sequestrating Zn²⁺ from an intracellular target via a newly formed thiol-based metal-binding site, which is similar to that seen with the antimicrobial peptide human defensin 5, hD5 (31). We therefore suggest that in general, fungus-cell–penetrating Zn²⁺-chelators, like redS100A7 and TPEN, could be useful as an important new therapeutic for opportunistic, superficial, or invasive fungal infections.

Materials and Methods

Collection of human skin material was approved by the ethical committee of the Shimane University. Written informed consent was obtained from each human subject involved in this study.

Fungi were cultured in a Sabouraud liquid medium for assaying antifungal proteins. Fungal growth was photometrically measured at 595 nm. Purification of the antifungal protein from lesional psoriatic scale extracts was performed by HPLC (39, 40). Its structural identification as reduced psoriasin (redS100A7) was performed using MALDI-MS, ESI-MS, amino acid sequencing, and Western blot analyses. Psoriasin mutants were generated as SUMO-fusion proteins, which were cleaved by SUMO-protease and further purified by HPLC. Morphological studies of redS100A7- and TPEN-treated fungi were performed with TEM and SEM. For immunogold TEM, S100A7 locations in treated T. rubrum were identified with antibodies. Apoptosis induction was tested with the TUNEL-assay and SYTOX-green staining. In vivo activities of redS100A7 and TPEN as antifungal agents were investigated in guinea pig tinea pedis (28) and mouse Aspergillus lung infection models. All animal manipulations were approved by the ethical committee of Shimane University, Faculty of Medicine (approval no. 459). See SI Materials and Methods for full details.

SI Materials and Methods

Human Material Approval.

The collection and use of skin stratum corneum specimens from patients with psoriasis, fungal infection, and healthy persons was approved by the ethical committee of the Shimane University, Faculty of Medicine (approval no. 459). Before its initiation, this study was fully explained, together with the methodology to be used, to each subject and written informed consent was obtained.

Fungal Strains and Isolates.

Aspergillus fumigatus (IFO: 4400), Candida albicans (NBRC: 1385), Malassezia furfur (NBRC: 101594), Microsporum canis (IFM: 54199-L1), Rhizopus oryzae (NBRC: 5780), Trichophyton mentagrophytes (IFM: 46027-L4), and Trichophyton rubrum (IFM: 46157; IFO: 5467; IFM: 57436-L1; NBRC 5807) were obtained from the Japan Society of Culture Collection. Fungi were cultured on potato dextrose agars containing chloramphenicol and cycloheximide at 28 °C. After 7 d of culture conidia were collected and stored at 5 × 10⁶ conidia mL⁻¹ in water containing 40% (vol/vol) glycerol at −70 °C. Malassezia species were cultured on Sabouraud dextrose agar containing chloramphenicol and cycloheximide with olive oil overlay at 32 °C.

Microbroth Dilution Assay.

The test was performed in flat-bottom 96-well plates. Each well was filled with 180 μL of Sabouraud liquid medium containing fungal conidia at the final concentration of 1x 10⁵ conidia mL⁻¹. Chloramphenicol and cycloheximide were added to prevent bacterial contaminants. Sterile PBS and fluconazole (Pfizer) or amphotericin-B (Wako) served as negative and positive controls, respectively. Microplates were incubated at 28 °C and fungal growth was monitored by microscopy and absorbance at 595 nm from culture day 0–14. For Malassezia species, a microbial viable assay kit (Dojindo) was used.

Purification of Reduced S100A7 from Psoriatic Scales.

One gram of lesional psoriatic scales or healthy persons’ heel stratum corneum was homogenized in 0.1 M citric acid/ethanol [50:50 (vol/vol)] buffer using a Polytron (Kinematica), similar to as described previously (39, 40). The extracts were cut off 3–50 kDa by diafiltration (Amicon ultra, Millipore, MA), and applied to a heparin affinity column (HiTrap 5 mL, GE Healthcare). After washing the column with equilibration buffer (0.1 M tris-citrate, pH 7.5), bound proteins were eluted with 2 M NaCl in 0.1 M tris-citrate buffer. The eluted fractions were tested for antifungal assay against T. rubrum in a microbroth dilution assay system. The heparin-bound material, that had active antifungal activity, was adjusted to pH 3.0 by adding trifluroacetic acid (TFA) and loaded onto a preparative C8 RP-HPLC column (YMC-Pack C8, 150 × 10 mm i.d., S-5 μm, 30 nm; YMC Co. Ltd., Japan). The column was previously equilibrated with 0.1% TFA. Proteins were eluted with a gradient of increasing concentrations of acetonitrile containing 0.1% TFA. The eluted fractions were lyophilized, dissolved in sterile distilled water, and tested for antifungal assay in a microbroth dilution system.

The fraction with strong antifungal activity against T. rubrum was purified with cation exchange HPLC column (YMC-BioPro SP-F, 100 × 4.6 mm i.d., S-5 μm; YMC Co. Ltd., Japan) that was previously equilibrated with 20 mM ammonium formate buffer. The proteins were eluted with 1M NaCl in equilibration buffer containing 5% (vol/vol) acetonitrile. Each eluted fraction was desalted with C8 RP-HPLC column and tested for antifungal assay. The fraction containing antifungal activity against T. rubrum was finally purified with C18 RP-HPLC column (YMC-Triart C18, 150 × 4.6 mm i.d., S-5 μm, 12 nm; YMC Co. Ltd., Japan).

Fungicidal Assay.

The in vitro fungicidal activity of redS100A7 was performed by a modification of the method of Espinel-Ingroff et al. (41). T. rubrum spores (1 × 10⁵ conidia mL⁻¹) were incubated with various concentrations of redS100A7 in a 96-well plate for 72 h. After incubation, 20 μL was subcultured from the redS100A7-treated wells and from the growth control wells (drug-free medium) into 180 μL of Sabouraud dextrose medium in a 96-well plate: the specified volume was pipetted without agitation of the well. The plate was incubated at 28 °C for 22 h. Ten microliters of coloring reagent (Dojindo Microbial Viability Assay Kit-WST: the electron mediator in the kit receives electrons from viable cells and transfers the electrons to WST, the water-soluble tetrazolium salt) was added to the wells, and the plate was incubated at 28 °C for 2 h. Viable T. rubrum was determined by monitoring the color intensity of WST formazan dye (wavelength at 450 nm). Fungicial activity was calculated by comparing the color intensity of the treated wells with that of the drug-free controls.

SDS/PAGE and Immunoblot Analysis.

SDS/PAGE analyses were performed at nonreducing conditions, similar to as described previously (39, 40). Western blotting was performed with a mouse monoclonal S100A7 antibody (ab13680, Abcam), as described previously (39, 40).

Tryptic Digestion and Mass Spectrometry.

After in-gel trypsin-treatment, protein fragments were analyzed by an AB SCIEX TOF/TOF 5800 system (Applied Biosystems). The resulting mass spectra were processed by Data Explorer 4.0 software (Applied Biosystems). Both PMF and MS/MS database searches were performed on a Mascot Server (Matrix Science).

Amino Acid Sequencing.

Sequences of PVDF membrane-transferred proteins were determined with a model PPSQ-33A protein sequencer (Shimadzu), as recommended by the manufacturer.

MALDI-MS Analysis.

Molecular masses of the proteins were measured using an AB SCIEX TOF/TOF 5800 system (Applied Biosystems). All samples were cleaned up with ZipTip C₁₈ cartridges (Millipore) and analyzed at standard conditions. The molecular mass was estimated in the positive ion linear mode with cytochrome c and apomyoglobin (Sigma-Aldrich) as internal standards. The processing of the resulting mass spectra was performed by Data Explorer 4.0 software (Applied Biosystems).

Reduction and Alkylation of S100A7.

One-hundred micrograms of oxidized psoriasin (oxS100A7) was dissolved in 500 μL of 0.1 M NaHCO₃ containing 4 M urea and adjusted to pH 8.3 with 0.1 M HCl. Then 50 μL of 0.4 M DTT in water was added under nitrogen atmosphere, stored for 4 h, and followed by addition of 100 μL 1 M iodoacetamide (Wako). After incubation in the dark for 1 h at room temperature, the reaction was stopped by adjusting to pH 3.0 with TFA. The reduced and alkylated S100A7 was purified by C₁₈ RP-HPLC and stored lyophilized.

Generation of Recombinant S100A7 Variants.

Total RNA was isolated from cultured foreskin-derived keratinocytes using the RNeasy Mini kit (Qiagen). A total of 2 mg of total RNA was reverse-transcribed with an oligo(dT)₁₈ primer and SuperScript II RNaseH⁻ RT (Invitrogen). All primers were designed from the S100A7/psoriasin gene (GenBank accession no. NM_002963.3). To retrieve a general human psoriasin cDNA template, PCR was performed with the primer pair psoriasin-f and psoriasin-r and Phusion Hot Start II DNA Polymerase (Finnzymes) under the following conditions: 98 °C for 34 s; 30 cycles (98 °C for 5 s, 72 °C for 10 s); and finally 72 °C for 1 min. 95 °C for 2 min; 32 cycles at 95 °C for 20 s, 64 °C for 30 s, 72 °C for 30 s; and finally 72 °C for 10 min. The PCR product was cloned and confirmed by sequencing before its plasmid DNA was used as the template for the generation of different psoriasin-variant templates.

To generate the three different psoriasin structural variants, PCR was performed with Taq DNA polymerase (Invitrogen) and one pair of primers containing appropriate mutations (shown below) under the following conditions: 95 °C for 2 min; 5 cycles (95 °C for 20 s, 58 °C for 30 s, 72 °C for 45 s); 25 cycles (95 °C for 20 s, 72 °C for 45 s); and finally 72 °C for 10 min. To generate inserts, the PCR product was double-digested with BsaI and BamHI for subcloning into the pE-Sumo3 vector (Lifesensors). The expression construct was amplified in Escherichia coli Top10 (Invitrogen) and confirmed by sequencing.

For recombinant protein expression, the constructed plasmid was transformed into BL21(DE3)pLysS cells (Novagen), selected on LB agar plates containing kanamycin (50 µg/mL⁻¹) and chloramphenicol (34 µg/mL⁻¹). Transformants were grown at 37 °C and 225 rpm in LB medium containing appropriate antibiotics to an OD₆₀₀ of 0.6. Protein expression was induced with 1 mM IPTG (isopropyl-beta-d-thio-galactopyranoside) at 30 °C for 3 h. After incubation, cells were harvested by centrifugation and resuspended in phosphate buffer (20 mM NaH₂PO₄, 20 mM Na₂HPO₄, 10 mM imidazole pH 7.6). Resuspended cells were subjected to one cycle of freeze-thawing and sonicated on ice. After centrifugation at 13,500 × g for 30 min, the clarified supernatant was applied to a HisTrap HP column (GE Healthcare). Separation of the polyhistidine-tagged protein was carried out using Äkta FPLC (GE Healthcare) with a linear imidazole gradient elution buffer (20 mM NaH₂PO₄, 20 mM Na₂HPO₄, 500 mM NaCl, 250 mM imidazole, pH 7.6). The fusion protein was further purified by RP-HPLC using preparative wide-pore C8 RP-HPLC with a column (SP250/10 Nucleosil 300-7 C8; Macherey-Nagel) that was previously equilibrated with 0.1% (vol/vol) TFA in HPLC-grade water containing 10% (vol/vol) acetonitrile. Proteins were eluted with a gradient of increasing concentrations of acetonitrile containing 0.1% (vol/vol) TFA (flow rate, 3 mL/min⁻¹). Fractions containing UV (215 nm)-absorbing material were collected, analyzed by ESI-QTOF-MS (Micromass), and lyophilized.

To remove the N-terminal tag sequences, fusion proteins were dissolved in PBS containing 3 mM DTT and digested with SUMO-Protease 2 (Lifesensors). The recombinant protein was purified by RP-HPLC using a Jupiter-5µ-C18-300A (150 × 2.0 mm) HPLC column (Phenomenex) equilibrated with 0.05% (vol/vol) TFA and 10% acetonitrile in water. Proteins were eluted with a gradient of increasing concentrations of acetonitrile containing 0.05% TFA (flow rate, 0.5 mL/min⁻¹). Fractions were collected, analyzed by ESI-QTOF-MS, and vacuum-dried.

CD Spectroscopy.

CD measurements were carried out as described previously (16). Proteins were dissolved to a concentration of 10 μM in 50 mM sodium phosphate buffer, pH 5.4 or pH 7.4. CD spectra in the absence and presence of zinc chloride (100 μM) have been recorded in 50 mM Tris, pH 7.4 and 50 mM sodium acetate, pH 5.4. The spectral bandwidth was 2 nm. All measurements were performed at 24 °C.

Thioredoxin and Glutathion Reduction Assays.

TRX reduction assays were performed as described previously (7). Briefly, for RP-HPLC analysis oxidized psoriasin (8.8 µM) or S-sulfito-psoriasin (8 µM) was incubated with 0.8 mM NADPH (Biomol), 100 nM rat TRX reductase (IMCO), and 0–2 µM human TRX (Sigma-Aldrich) for 30 min at 37 °C in 0.1 M potassium phosphate-2 mM EDTA, pH 7.0 buffer. For GSH reduction assays, oxidized psoriasin was incubated with 0–10 mM reduced GSH (Sigma-Aldrich) in 0.1 M potassium phosphate-2 mM EDTA, pH 7.0 buffer.

Incubation mixtures were acidified with TFA and analyzed with RP-HPLC on a Vydac 218TP-C18 column (250 × 4.6 mm, 5 µm; Grace). Conversion from oxidized into reduced S100A7 was followed by retention time.

Intracellular ROS Accumulation.

Intracellular ROS production in T. rubrum was measured using a fluorescent dye 2′,7′-dichlorofluorescin diacetate (DCFHDA; Molecular Probes, Invitrogen) and a fluorescent spectrometer. T. rubrum conidia were grown for 12 h and incubated with 2 or 8 μM of redS100A7 in the presence or absence of 5 mM of the ROS-inhibitor ascorbic acid in the liquid medium for 2 h at 28 °C. After washing, the cells with PBS and incubating with DCFHDA for 1.5 h at 28 °C; fluorescence intensity was determined by using a fluorescence microplate reader.

SOD Inhibition Assay.

To extract SOD enzymes, filamentous fungi (T. rubrum) were cultured for 7 d as described above, collected, and washed with PBS. Approximately 20 mg of wet fungus were homogenized with glass beads in 400-μL ice-cooled PBS by using a Qiagen TissueLyser (Qiagen). Homogenized samples were then centrifuged and supernatants containing the cytoplasmic contents were collected and passed through a 50-kDa cut-off filter device (Amicon ultra, Millipore). Extracts were then monitored by SDS/PAGE electrophoresis for the presence of copper/zinc-SOD (Cu/Zn-SOD) enzyme and stored at −70 °C. Ethanol/chloroform 62.5/37.5 (vol/vol) was added to the T. rubrum cytoplasmic extracts to inactivate manganese-SOD (Mn-SOD) and iron-SOD (Fe-SOD). The antioxidant action of T. rubrum Cu/Zn-SOD was measured in the presence or absence of redS100A7 using a SOD assay kit (Dojindo), as described by the manufacturer.

TUNEL Assay.

DNA strand breaks were demonstrated by the TdT-mediated dUTP nick end-labeling TUNEL method using an In Situ Cell Death Detection kit, fluorescein (Roche). A. fumigatus and T. rubrum were cultured on Falcon four-well culture slides (BD Biosciences) for 36–48 h. Then the fungi, treated with 4 μM of red S100A7 for 4 h in Sabouraud liquid medium, were fixed with 3.7% (vol/vol) formaldehyde and 5% (vol/vol) dimethyl sulfoxide in PEM [50 mM piperazine-N,N’-bis(2-ethanesulfonic acid) pH 6.7, 25 mM ethylene glycol tetra-acetic acid pH 7.0, 5 mM magnesium sulfate] for 30 min at room temperature. Fungi were washed with PEM and digested for 30 min at 37 °C with 200 μL of 10 mg/mL⁻¹ Zymolyase-20T (Seikagaku). The slides were rinsed with PEM and incubated for 5 min at room temperature with permeabilization solution (100 mM Pipes pH 6.7, 25 mM EGTA pH 7.0, 0.01% Igepal CA-630). The slides were rinsed three times with PEM and incubated with 50 μL of TUNEL reaction mixture (Roche) in the dark for 1 h at 37 °C, rinsed with PEM, mounted with a drop of Fluorescence Mounting Medium (Dako), and then examined by a fluorescence microscope.

SYTOX-Green Uptake Assay.

Fungi, grown as described above, were incubated for 24 h with 4 μM of redS100A7 or PBS in Sabouraud liquid medium. SYTOX-green (Molecular Probes-Invitrogen) (final concentration: 1 μM) was added. After 10-min incubation with shaking in the dark, fluorescence was viewed and captured with an Olympus BX51 fluorescence microscope and an Olympus DP 21 digital camera system.

TEM.

Controls and redS100A7-treated fungi grown on an agar plate were immediately fixed with 2% (vol/vol) paraformaldehyde and 2% (vol/vol) glutaraldehyde in 0.1 M phosphate buffer (PB; pH 7.3) overnight at 4 °C. Samples were cut into 100-μm-thick sections. For conventional electron microscopy, sections were postfixed with 1% osmium tetroxide in PB for 1 h at room temperature. Samples were stained en block with 0.5% uranyl acetate in distilled water overnight at 4 °C. Dehydration was achieved in a graded series of ethanol. Dehydrated samples were cleared in QY-1, flat-embedded in Spurr’s resin, and ultrathin-sectioned at 80 nm on an ultramicrotome. Sections were collected on one-slot grid coated with collodion, and stained with lead citrate. For pre-embedding immunogold labeling, vibratome sections were incubated in PB containing 0.04% Triton X-100, 1.5% normal goat serum and rabbit polyclonal anti-S100A7 antibody (ab83534, Abcam) overnight at 4 °C, followed by washing with PB and incubation with PB containing 0.04% Triton X-100, 1.5% normal goat serum and 0.8-nm gold-conjugated goat anti-rabbit IgG (Aurion, 1:100) for 8 h at room temperature. After treatment with 1% glutaraldehyde in 0.1 M PB for 10 min, samples were rinsed with PB followed by 0.2 M sodium citrate buffer (pH 7.4). To enhance immunogold particles, samples were then treated with a HQ Silver enhancement kit (Nanoprobes) for 15 min. After silver-gold intensification, samples were processed for electron microscopy as described above. Specimens were monitored with a TEM (JEM1200EX, JEOL).

SEM.

Controls, redS100A7-, and TPEN-treated fungi on an agar plate were fixed with 2.5% (vol/vol) glutaraldehyde for 4 h and postfixed with 1% (vol/vol) osmium tetroxide in PB for 1 h at room temperature. Samples were dehydrated in a graded ethanol series, critical-point dried and mounted with a silver conducive adhesive (Nisshin EM). Specimens were then sputter-coated with gold and monitored under a SEM (JSM-6510, JEOL).

Immunostaining and Confocal Laser-Scanning Microscopy.

Controls and redS100A7-treated fungi were immediately fixed with 2% (vol/vol) paraformaldehyde and 2% (vol/vol) glutaraldehyde in PB overnight at 4 °C. Samples were cut into 100-μm-thick sections and mounted on glass slides. Both control and treated samples were immersed in blocking solution [3% (wt/vol) BSA and 0.1% Triton-X100 in PBS] for 30 min at room temperature. Samples were incubated for 2 h at room temperature with rabbit polyclonal anti-S100A7 antibody (ab83534, 1:250, Abcam) in blocking solution. After washing three times with PBS, samples were incubated for 1 h at room temperature with goat anti-rabbit Ig G (H+L) 488 (Invitrogen) in PBS containing 0.1% (vol/vol) Triton-X100 at 1:500 dilution. After washing three times with PBS, glass slides were mounted with a drop of Fluorescence Mounting Medium (Dako) and then examined by a confocal laser-scanning microscope (FV300, Olympus).

Analysis of the in Vivo S100A7 Redox Status.

Stratum corneum was collected by using a scalpel from T. rubrum-infected skin lesions on the foot (n = 4) and from healthy volunteers skin at the same location (n = 4). Immediately after collecting, equal amounts of stratum corneum from healthy volunteers and patients (100 mg) were homogenized in ice-cold 0.1 M citric acid/ethanol [50:50 (vol/vol)] buffer. Homogenized samples were adjusted to pH 8.3, and then iodoacetamide (final concentration: 150 mM) was added to prevent artificial air-oxidation of redS100A7. S100A7 derivatives were purified as described above. The ratio of secreted psoriasin in T. rubrum-infected skin and healthy skin was analyzed by UV (215 nm) absorbance integration of C₈RP-HPLC-peaks. Peak area (U) = Peak height (mAU) × length of peak base (min) ± SD. The redS100A7-content was analyzed as the sum of intensity percentage of mono-CAM (carboxamidomethylated)-psoriasin (redS100A7α) and bi-CAM-psoriasin (redS100A7β) ± SD by MALDI-MS.

Animal Model of Pulmonary Aspergillosis.

Female BALB/c mice, 20–23 g, were purchased (CLEA Japan) and housed at specific-pathogen-free conditions. Mice were kept in sterile cages and fed with sterile food and water. All procedures were reviewed and approved by the Animal Research Ethics Board at Shimane University Faculty of Medicine.

Chemotherapy-induced immunosuppression was achieved by vinblastine sulfate (Nippon Kayaku). The dose of 5 mg/kg per mouse was given intravenously 2 d before infection. A. fumigatus (IFO: 4400) was grown on a potato dextrose agar at 30 °C for 7 d. Conidia were harvested as described above. Under light anesthesia with inhaled isoflurane, mice were given an intranasal instillation of 2 × 10⁷ conidia in 40 μL of sterile PBS using a sterile gel-loading tip (29). Noses were decontaminated with 70% (vol/vol) ethanol. This procedure was repeated for 2 consecutive days.

Bulk amounts of oxS100A7 were purified from psoriatic scales (39, 40) and reduced to redS100A7 as described above. redS100A7 and TPEN (Sigma-Aldrich) were dissolved in sterile PBS to the desired concentrations and to each group (seven mice per group), simultaneously with A. fumigatus application, one of these regimens was administered under light anesthesia with inhaled isoflurane: redS100A7 [5 mg per kg, intranasally, once daily (OD), 2 d], TPEN (2.5 mg/kg, intranasally, OD, 2 d) or vehicle only (sterile PBS, intranasally, OD, 2 d). Mice were observed twice daily for survival up to 7 d.

Animal Model of Fungal Skin Infection.

T. mentagrophytes TIMM2789, which was obtained from the Teikyo University Institute of Medical Mycology (Tokyo, Japan), was precultured on Sabouraud dextrose agar (Becton, Dickinson) slants at 27 °C for 13 d. Sterilized saline containing 0.1% (vol/vol) Tween 80 was added to the slants, and the conidia were collected as suspension by gentle rubbing the surface of them with a loop. The suspension was filtered through sterilized gauze to remove the hyphal fragments. The number of conidia in the filtrate was measured using a Thoma hemacytometer, and the concentration was adjusted to 1.0 × 10⁸ conidia mL⁻¹, using sterile saline containing 0.1% (vol/vol) Tween 80.

Inhibition assay for tinea pedis-animal model was performed as described previously (28). The foot of guinea pig was cleaned with a cotton swab moistened with sterile saline, and then the surface of the skin was lightly abraded with sandpaper. Test drugs prepared above (0.05 mL per site) were applied to the feet and air-dried. A sterile adhesive bandage (Band-Aid; Johnson & Johnson) soaked with 0.15 mL of the T. mentagrophytes inoculum was applied to the planta and covered with a film (Saran Wrap; Asahi Kasei Corporation). A form pad (Reston Self-Adhering Foam Pads 1560M; 3M Japan Limited) was placed on the site with adhesive elastic tape (Tensoplast; BSN medical) so as to avoid excessive pressure to the infected site. The guinea pigs were killed after 3 d of inoculation, and the bandages were removed.

Skin-tissue specimens, including epidermis and dermis, were collected from the feet by use of blades. The fungal tissues in the skin were observed microscopically after staining with PAS and Fungiflora Y (Trust Medical). The scoring system for T. mentagrophytes invasion into stratum corneum is as follows: score 0 = no signs of invasion; score 1 = superficial invasion of a small number of fungi into stratum corneum; score 2 = deep invasion of a small number of fungi into stratum corneum; score 3 = deep invasion spreading up to one-third portion of stratum corneum; score 4 = deep invasion spreading from one-third to one-half portion of stratum corneum; score 5 = deep invasion spreading whole area of stratum corneum.

Statistics.

All data represented as means ± SD, unless otherwise specified. For Fig. 5A, data were represented as means ± SEM. All P values were calculated using the two-samples, independent Student’s t test, with the exception of the data from Fig. 5C, in which the logrank test (Mantel–Cox) was used.