Histidine-rich glycoprotein protects from systemic Candida infection

Abstract

Fungi, such as Candida spp., are commonly found on the skin and at mucosal surfaces. Yet, they rarely cause invasive infections in immunocompetent individuals, an observation reflecting the ability of our innate immune system to control potentially invasive microbes found at biological boundaries. Antimicrobial proteins and peptides are becoming increasingly recognized as important effectors of innate immunity. This is illustrated further by the present investigation, demonstrating a novel antifungal role of histidine-rich glycoprotein (HRG), an abundant and multimodular plasma protein. HRG bound to Candida cells, and induced breaks in the cell walls of the organisms. Correspondingly, HRG preferentially lysed ergosterol-containing liposomes but not cholesterol-containing ones, indicating a specificity for fungal versus other types of eukaryotic membranes. Both antifungal and membrane-rupturing activities of HRG were enhanced at low pH, and mapped to the histidine-rich region of the protein. Ex vivo, HRG-containing plasma as well as fibrin clots exerted antifungal effects. In vivo, Hrg(-/-) mice were susceptible to infection by C. albicans, in contrast to wild-type mice, which were highly resistant to infection. The results demonstrate a key and previously unknown antifungal role of HRG in innate immunity.

Figure 1. Antifungal activity and binding of HRG to Candida.

(A) Antifungal activity of HRG. C. parapsilosis ATCC 90018 (1×10⁵ cfu) was incubated with purified human HRG at concentrations ranging from 0.03 to 6 µM for 2 hours in 10 mM Tris, pH 7.4 (•) or 10 mM MES, pH 5.5 (○), plated and the number of cfu determined (n = 6). (B) Antifungal effects of HRG in salt. C. parapsilosis ATCC 90018 (1×10⁵ cfu) was incubated with 6 µM HRG for 2 hours in 10 mM MES, pH 5.5 containing 0, 25, 50, 100 or 150 mM NaCl, plated and the number of cfu was determined. (C) Killing kinetics. 0.3 or 3 µM HRG were incubated with 1×10⁵ cfu C. parapsilosis ATCC 90018 for 0, 5, 15, 30, 60 or 120 minutes in 10 mM Tris, pH 7.4 (•) or 10 mM MES, pH 5.5 (○), plated and the number of cfu determined. (D) Antifungal activity of HRG against different strains of Candida. 3 µM HRG were incubated with 1×10⁵ cfu C. parapsilosis ATCC 90018 or BD 17837, C. albicans ATCC 90028 or BD 1060, C. glabrata ATCC 90030 or C. krusei ATCC 6258 in 10 mM Tris, pH 7.4 (black bars) or 10 mM MES, pH 5.5 (white bars) for 2 hours, plated and number of cfu determined (n = 6). (E) Binding of HRG to fungi. C. parapsilosis (1×10⁵ cfu) was incubated with HRG (0.6 µM) in 10 mM MES, pH 5.5. For inhibition studies, heparin (50 µg/ml) was added. Samples were centrifuged and the pellet and supernatants were extracted and run on 8% SDS-PAGE under reducing conditions. HRG was detected by western and immunoblotting using polyclonal antibodies against GHH20. Purified HRG was used as a positive control (labeled C). (F) Flow cytometry analysis of binding of HRG to fungal membranes. C. parapsilosis (5×10⁷ cfu) were incubated with FITC-labeled HRG in 10 mM Tris pH 7.4 or 10 mM MES pH 5.5.

Figure 2. HRG induces membrane permeabilisation of Candida cells as well as liposomes.

(A) Negative staining and electron microscopy analysis of C. parapsilosis exposed to HRG. C. parapsilosis ATCC 90018 were incubated in the absence of HRG in 10 mM Tris, pH 7.4 or 10 mM MES, pH 5.5. These fungi did not exhibit signs of membrane perturbations. In contrast, when treated with 10 µM HRG in 10 mM Tris, pH 7.4 or 10 mM MES, pH 5.5 membrane damage, blebbing and ejection of cytoplasmic components was observed. Fungi treated with 10 µM LL-37 was used as a positive control for membrane damage. The scale bar corresponds to 2 µm. (B) Fungal viability after incubation with HRG and LL-37. C. albicans ATCC 90028 were incubated with 10 µM HRG or LL-37 in either 10 mM Tris, pH 7.4 (left panel) or 10 mM MES, pH 5.5 (right panel). The left images in each row are Nomarski Differential Interference Contrast images, whereas the right images show FITC fluorescence of fungi. (C) Effects of 1 µM HRG on liposome permeability. Left panel. Increase in HRG-induced permeabilization of ergosterol-containing liposomes is detected at pH 6.0. Right panel. Increased HRG-induced lysis of (at pH 7.4) ergosterol containing liposomes. (D) CD spectroscopy of HRG under different conditions. CD spectra for 0.25 µM HRG in buffer and in presence of S. cerevisiae mannan are presented.

Figure 3. The antifungal activity of the histidine-rich domain of HRG is significantly increased at low pH.

(A) Sequence of HRG and synthetic peptides used in this study are indicated. (B) Screening of antifungal epitopes of HRG. 20-mer peptides spanning the whole sequence of HRG (for sequences see Table S1) were used in radial diffusion assays against C. parapsilosis ATCC 90018 and C. albicans ATCC 90028 in 10 mM Tris, pH 7.4 or in 10 mM MES, pH 5.5. A 4 mm diameter well was loaded with 6 µl of 100 µM peptide. The clearance zones (mm) were measured after an overnight incubation at 27°C (n = 6). (C) Comparison of the antifungal activity of rHRG and the truncated version rHRG1-240. C. parapsilosis ATCC 90018 (1×10⁵ cfu) was incubated with 0.6 µM rHRG or rHRG1-240 in 10 mM Tris, pH 7.4 or in 10 mM MES, pH 5.5. Samples were plated and the number of cfu was determined. Significance was determined using Kruskall-Wallis one-way ANOVA analysis (*** p<0.001, n = 6).

Figure 4. Antifungal activity and fungal binding of GHH20 peptide.

(A) Antifungal activity of GHH20. C. parapsilosis ATCC 90018 or C. albicans ATCC 90028 (1×10⁵ cfu) were incubated with GHH20 peptide (0.03 to 6 µM) for 2 hours in 10 mM Tris, pH 7.4 (•) or 10 mM MES, pH 5.5 (○), plated and the number of cfu determined. A representative experiment (of three) is shown. (B) Flow cytometry analysis of binding of GHH20 to fungal membranes. C. parapsilosis (5×10⁷ cfu) were incubated with 20 µg TAMRA-labeled GHH20 in 10 mM Tris pH 7.4 or 10 mM MES pH 5.5. (C) Binding of 2 µg TAMRA-labeled GHH20 peptide to C. parapsilosis ATCC 90018 and inhibition by an excess of heparin. C. parapsilosis were incubated with TAMRA-labeled GHH20 in 10 mM MES (panel 3), pH 5.5 or the same buffer supplemented with heparin (50 µg/ml) (panel 4). The left panel shows Nomarski images (1 and 3), whereas the right panel shows red fluorescence of peptide bound to fungi. (D) The GHH20 peptide permeabilizes ergosterol-containing liposomes preferably at pH 6.0. 1 µM GHH20 was used (n = 6).

Figure 5. Localization and activities of HRG.

(A) Analysis of HRG in biological fluids. The indicated biological materials were electrophoresed on a 8% gel (Tris-Glycine, non-reducing conditions) (left panel) or on a 16.5% Tris-Tricine gel under reducing conditions (right panel) and transferred to a nitrocellulose membrane. Western blot was performed using polyclonal antibodies directed against the GHH20 epitope of HRG. (B) Localization of HRG in fibrin clots. Human control plasma (panel 4) or plasma depleted of HRG (panel 2) were incubated with FITC-labeled HRG and clots were generated overnight after addition of 10 mM Ca²⁺ at 37°C. The clots were mounted on slides and visualized by fluorescence microscopy. The left side shows Nomarski images (1 and 3), whereas the right part shows fluorescence of HRG associated with the clots. (C) Candida growth in plasma. C. parapsilosis (2×10⁷ cfu) was inoculated in human plasma (•) or human HRG-deficient plasma (○) and incubated at 27°C for 0, 4 ,8 or 18 hours and the number of cfu was determined (n = 6). (D) Antifungal activity of HRG ex vivo. Mouse control plasma or plasma of Hrg^−/− mice were used to form fibrin clots in the presence of 10 mM Ca²⁺. Clots were incubated with C. parapsilosis ATCC 90018 (1×10⁵ cfu) in 10 mM MES, pH 5.5 for 2 hours, plated and the number of cfu determined (p = 0.043, n = 6).

Figure 6. Candida produces severe infection in Hrg^−/− mice.

(A) Weight loss in mice after infection with C.albicans. C57BL/6 (solid line) and C57BL/6 Hrg^−/− (dotted line) mice were infected (i.p.) with 1×10⁹ cfu of C. albicans, and the body weight was followed from day 0 to day 3 (n = 6,4 and 2, respectively) (p = 0.02). (B) Fungal dissemination to the bloodstream. C57BL/6 (•) (n = 4) and C57BL/6 Hrg^−/− (○) (n = 5) mice were infected as above and the animals were sacrificed on day 2 and number of cfu in blood was determined (p = 0.032). (C) HRG suppresses fungal dissemination to the spleen and kidney. C57BL/6 (•) and C57BL/6 Hrg^−/− (○) mice were infected as above and the cfu of C. parapsilosis in spleen and kidney was determined (p = 0.009, n = 10) (D) In two animals kidneys were collected on day 3, fixed, sectioned and then stained with hematoxylin and eosin (HE) (left panel) or PAS (right panel) (upper section; magnification ×10 and lower section; magnification ×30).