Effects of human lactoferrin on the cytoplasmic membrane of Candida albicans cells related with its candidacidal activity

Abstract

Human lactoferrin is an innate host defence protein with antimicrobial activity that exerts a candidacidal effect in a cation concentration-dependent manner. We investigated the ability of this cationic protein (with an isoelectric point of 8.7) to permeabilize the cytoplasmic membrane of Candida albicans cells. Despite minor K⁺-release in lactoferrin-treated C. albicans cells, the killing effect was not related to an extensive membrane permeabilization, as indicated by: (a) the non-release of macromolecular cytosolic constituents; (b) the non-permeabilization for extracellular propidium iodide nor for intracellular accumulated calcein; and (c) the inability to disrupt the phospholipid bilayer of 8-aminonaphthalene-1,3,6, trisulfonic acid/p-xylene-bis-pyridiniumbromide-loaded liposomes. These results suggest that lactoferrin exerts its candidacidal effect through a mechanism different from membrane permeabilization described for other cationic peptides.

Lactoferrin, Candida albicans, Candidacidal activity, Membrane permeabilization, Potassium release

1 Introduction

Candida albicans is a dimorphic fungal organism that is a commensal of the gastrointestinal and reproductive tracts as well as an opportunistic pathogen causing frequent mucosal and systemic infections [1,2]. Cell-mediated immunity and the innate antimicrobial proteins and peptides of mucosal fluids are involved in host defence against C. albicans infection [3,4].

Human lactoferrin (hLf) is an iron-binding protein (77 kDa) present in milk and mucosal secretions that possess in vitro antimicrobial activity by sequestering the free iron required for microbial growth and by a direct interaction with the bacterial surface [5]. In addition, hLf could release two N-terminal hLf-derived peptides with antimicrobial activity [6,7]. Although the candidacidal mechanism of lactoferrin is poorly understood, it has been related to the binding of lactoferrin with the C. albicans surface rather than to iron deprivation [8]. Furthermore, hLf-mediated cell wall damage with formation of bleb-like structures has been shown previously [9,10]. Recently, we found that the candidacidal activity exerted by hLf is dependent on the extracellular cation concentration and the cellular metabolic state, suggesting that the extracellular ionic medium may have significant effects on hLf candidacidal activity [11].

This study focuses on the initial events related to hLf candidacidal properties to determine its capability to permeabilize the cytoplasmic membrane of C. albicans as shown for other antimicrobial cationic peptides [12].

2 Materials and methods

2.1 Materials

Human lactoferrin (Lot # 98H3784, 18H3789) was obtained from Sigma (St. Louis, MO). Purity of the hLf was assessed as described [13] and the iron saturation (<0.03%) of the protein was determined by atomic absorption spectrometry. Gramicidin S and nystatin were from Sigma. 8-Aminonaphthalene-1,3,6, trisulfonic acid (ANTS), p-xylene-bis-pyridiniumbromide (DPX), calcein acetoxymethyl ester (calcein-AM) and propidium were purchased from Molecular Probes (Eugene, OR); phosphatidylcholine (PC) and phosphatidylserine (PS) were from Avanti Polar Lipids (Alabaster, AL); Sabouraud dextrose broth (SDB) or agarified (SDA) media were purchased from Difco (Detroit, MI).

2.2 Microorganisms and culture conditions

C. albicans ATCC 10231 was purchased from the American Type Culture Collection. C. albicans cells were grown at stationary phase in SDB and subcultured (1:400) in fresh SDB to the mid-logarithmic growth phase at 37 °C in a shaker. The cells remained in the yeast (blastoconidia) phase throughout the studies.

2.3 Antifungal activity assays

The anti-Candida effect of hLf was monitored using C. albicans cell suspensions (10⁵ cells/ml) in 5 mM potassium phosphate buffer (PPB; K₂HPO₄–KH₂PO₄, pH 7.4) incubated at 37 °C with 5 µM hLf. The number of viable cells was determined by counting colonies on SDA plates. Cell viability was expressed as percentage of control and the loss of viability was calculated as [1-(CFU from hLf-treated cells/CFU from control cells)] ×100.

2.4 Leakage of cellular constituents

Measurement of hLf-induced release of 260 nm-absorbing compounds of C. albicans cells was performed as described elsewhere [14], except that the determinations were made in 5 mM PPB, pH 7.4. Cell suspensions (10⁵ cells/ml) were incubated in the presence of 5 µM hLf at 37 °C. Cell samples (0.5 ml) were taken at different time points and centrifuged at 1500g for 15 min. The supernatant was used to determine the leakage of UV-absorbing materials from the cells at 260 nm (A₂₆₀). Measurements were adjusted for the presence of the hLf. The supernatant of the hLf-treated C. albicans cells was also analyzed by SDS–PAGE on 12% (w/v) gels and Coomassie blue staining.

2.5 Analysis of membrane permeabilization to propidium iodide and calcein

Cell permeability was monitored using the DNA-staining fluorescent probe propidium iodide (PI) as previously described [15]. Cell suspensions (10⁶ cells/ml) in 5 mM PPB, pH 7.4 were incubated with hLf for 2 h at 37 °C and subsequently treated with PI (9 µM, final concentration) for 5 min. Cell fluorescence was monitored using a flow cytometer (Cytoron Absolute, Ortho Diagnostics Systems Inc., Raritan, NJ).

Determination of calcein efflux from C. albicans cells was carried out as described previously [16]. Briefly, the cells (10⁷ cells/ml) were incubated with calcein-AM (6 µM, final concentration) for 2 h at 37 °C. Then, the calcein-loaded cells were washed four times with 5 mM PPB, pH 7.4 and diluted samples (10⁵ cells/ml) were incubated with or without hLf. The fluorescence intensity of calcein was measured every 15 min using a spectrofluorimeter Perkin–Elmer LS-50. The excitation and emission wavelengths were 485 and 530 nm, respectively. Calcein release was calculated according to the equation: % Release=[(F_f−F₀)/(F₁₀₀−F₀)]×100. F_f represents the equilibrium value of fluorescence after hLf addition, F₀ the initial fluorescence intensity of the cell suspension, and F₁₀₀ denotes the fluorescence value after boiling the calcein-loaded cells.

2.6 Preparation of liposomes and ANTS leakage assays

Liposomes were prepared as described [17] using phosphatidylcholine and phosphatidylserine (3:1). For each preparation, approximately 20 mg/ml of the lipid mixture was used. Dried phospholipids were obtained by evaporation of the solvents (CHCl₃–CH₃OH; 9:1) in a rotavapor. The dry lipid films were resuspended in Tris buffer (20 mM Tris–HCl, pH 7.4) containing 25 mM ANTS and 90 mM DPX. After a vigorous lipid dispersion the suspension was sonicated at intervals of 5–15 min until clear. The preparations were then centrifuged at 1000g for 10 min to remove particulate matter. Non-encapsulated ANTS and DPX were separated from the vesicle suspension through a Sephadex G-75 gel filtration column (Pharmacia, Uppsala, Sweden) eluted with Tris-buffer.

The leakage experiments were performed as described previously [18]. The fluorescence intensity was recorded in a Perkin–Elmer LS-50 spectrofluorimeter with excitation and emission wavelengths set at 350 and 510 nm, respectively. One hundred percent release of ANTS was achieved by adding Triton X-100 (0.1%, final concentration).

2.7 Measurement of cation concentrations

The potassium efflux was monitored by measuring the extracellular K⁺ concentration of C. albicans cell suspensions (10⁶ cells/ml) using an ion-selective electrode (Mettler Toledo), as described elsewhere [19]. The total K⁺ content (100% value) was determined by flame photometry in the supernatant of cellular suspensions previously treated with 0.5% (v/v) perchloric acid, heated at 95 °C, for 1 h and centrifuged to remove cell debris [11]. The percentage of K⁺ released from cell suspensions treated with nystatin (100 µg/ml) was 96±2% as determined by flame photometry.

3 Results and discussion

3.1 Lactoferrin was unable to permeabilize the cytoplasmic membrane

Some antimicrobial peptides exhibit antibiotic activity via the pore formation on the target cell membrane causing the leakage of essential metabolites and leading to cell lysis. The disruption of the membrane structure is the major killing mechanism for antibacterial and antifungal cationic proteins [12]. We have examined whether hLf would cause leakage of cellular components. The leakage of cytosolic constituents was determined at different time points (0, 30, 45, 60, and 90 min) using the supernatant from the cell suspensions, which were pre-incubated with a candidacidal concentration of hLf (5 µM). Neither the absorbance at 260 nm nor the SDS–PAGE analysis from the supernatants revealed a significant release of intracellular constituents or proteins in comparison with the controls (data not shown). This indicated that the cell membrane damage sustained by C. albicans cells, although lethal, did not result in a release of cytoplasmic constituents.

The ability of hLf to permeabilize the cytoplasmic membrane of C. albicans was investigated using the fluorescent probes propidium iodide and calcein. PI is a membrane impermeable DNA stain that can diffuse into cells only when the plasma membrane is permeabilized. The results of intracellular PI measurements were analyzed by flow cytometry and are displayed in Fig. 1(a). They indicate the inability of hLf (5 µM) to cause PI-permeabilization in C. albicans cells after 2 h incubation. In contrast, cells incubated with the membrane-perturbing antifungal drug amphotericin B [20], used as positive control, showed an increase in fluorescence due to an intracellular PI-accumulation indicative of membrane permeabilization.

Figure 1

Effect of human lactoferrin on cytoplasmic membrane of C. albicans cells and liposomes. (a) Fluorescence profiles of C. albicans cell suspensions treated for 2 h with hLf, amphotericin B (AmB), and non-treated (C, control) stained with propidium iodide and analyzed by flux cytometry. (b) C. albicans cells preloaded with calcein-AM and then exposed to hLf (▀) or non-exposed (□). The fluorescence intensity of the calcein released was recorded by spectrofluorimetry at excitation and emission wavelengths of 485 and 530 nm, respectively. In parallel, the hLf-killing effect was monitored (inset). (c) Liposomes containing ANTS/DPX were exposed to 5 µM (▀) or 10 µM of human lactoferrin (▲) or 40 µM gramicidin S (□). Arrows indicate the addition of gramicidin S (GS), human lactoferrin (hLf) or Triton X-100 (TX). Leakage is marked by an increase in fluorescence intensity expressed in arbitrary units (a.u.).

Calcein-AM is a polar molecule that is accumulated in the cytoplasm and subsequently hydrolyzed by intracellular esterases. The non-permeable calcein is only released by membrane permeabilizing agents and the leakage is detected by increased fluorescence. The exposure of calcein-preloaded cells to candidacidal concentrations of hLf (5 µM) showed a progressive release of non-significant amounts of calcein as compared to control cells (Fig. 1(b)). The maximal calcein efflux (13±3%; n=3) from hLf-treated cells was observed after 90 min. The release of calcein occurred also in non-treated cells and was interpreted as a spontaneous leakage from the cells, an event that has been described previously [16].

Determination of cell viability indicated a non-correlation between calcein efflux and loss of viability (i.e. 13% of calcein efflux vs. 76±5% of viability loss). Cell counting revealed that cell death induced by hLf was not due to a cytolytic effect.

3.2 Lactoferrin was unable to induce leakage from phospholipid vesicles

The results from permeabilization assays indicated that the integrity of C. albicans cytoplasmic membrane did not have significant permeability changes during the hLf-killing time. Given that hLf is a positively charged molecule (isoelectric point (pI) of 8.7), it could be expected to be sequestered by anionic molecules (i.e. mannoproteins) present in the cell wall. In this case, a diminished hLf-membrane interaction could explain the non-disruption of the cell membrane. Consequently, we tested the ability of hLf to permeabilize negatively charged phospholipid vesicles (liposomes) containing the fluorophore ANTS and the quencher DPX, which results in a low fluorescence emission when both are encapsulated in vesicles. Breakdown of the vesicle membrane causes leakage of the contents and dilution of ANTS and DPX, resulting in an increment of ANTS-fluorescence. The spectrofluorometry data showed the inability of hLf (5 and 10 µM) to induce leakage of the fluorophore during the 20 min incubation whereas the permeabilizing peptide gramicidin S (40 µM), used as positive control [21], caused a rapid (approximately 1 min) increase in ANTS fluorescence intensity (Fig. 1(c)). Adding Triton X-100 to the cuvette achieved total fluorophore release. Although a direct interaction of hLf with liposomes PC/PS (3:1) has been observed (Dr. H. Ostolaza, Department of Biochemistry, University of Basque Country, Spain; personal communication), the absence of ANTS leakage suggests the inability of hLf to cause an extensive disorganization of the phospholipid bilayer architecture.

3.3 Kinetic of K+-release induced by lactoferrin

We have previously reported a potassium efflux from cells exposed to hLf [7,11]. To determine the relationship between K⁺-release and loss of cell viability, we performed an accurate determination of both kinetic K⁺-release and the number of viable cells in the presence of hLf. Fig. 2(a) shows a representative assay where cells have been exposed to candidacidal concentrations of hLf (5 µM) or nystatin (100 µg/ml), the latter being an antifungal agent that causes leakage of monovalent cations by increased fluidity or disruption of the cell membrane [22]. Fig. 2(a) shows that hLf caused a maximal release (26±7%) of total K⁺ content after 17 min, whereas a maximal K⁺-release induced by nystatin (96±2%) was achieved after 5 min. Higher hLf concentrations (up to 20 µM) were unable to increase the percentage of released potassium. However, the loss of cell viability was not significantly increased when a maximal K⁺-release (26±7%) was reached (Fig. 2(b)), indicating that in C. albicans K⁺-release does not correlate with the onset of cell death induced by lactoferrin. Unlike other antibacterial and antifungal peptides [12,23], human lactoferrin did not cause an immediate and massive K⁺-release upon addition to the cells, indicating that membrane permeabilization due to the formation of pore-like structures was not involved. There are other antimicrobial peptides whose antifungal activity is exerted by unusual mechanisms [24,25]. For instance, recent studies on the salivary peptide histatin 5 suggest that this candidacidal peptide could cause cell cycle arrest due to a disruption of the cell volume regulatory mechanisms [26]. Since both histatin 5 and hLf induce a non-cytolytic K⁺-efflux not related to cell viability loss and are dependent on the extracellular salt concentration and cell metabolic state [11,25,27], it could be speculated that they exert their antifungal activity by a similar mechanism. The amount of K⁺-released (26±7%) from hLf-treated cells in addition to the previously observed concomitant depolarization of the plasma membrane of C. albicans cells [11] suggest that hLf induced a K⁺-efflux. However, further work is needed to determine if the non-lytic K⁺-release induced by hLf is an ion homeostatic response, irrelevant for the cell death process, or the first event of the hLf-candidacidal effect, as proposed for histatin 5 [26].

Figure 2

Kinetics of potassium-efflux induced by lactoferrin on C. albicans cells. (a) Potassium efflux from C. albicans cells exposed to hLf and nystatin (Nys) was monitored with an ion-selective potassium electrode. The K⁺ release induced by hLf and Nys is represented as the percentage relative to the total K⁺ released by cell suspensions treated with perchloric acid (0.5%) considered as 100%. (b) Candidacidal effect of human lactoferrin. Cell viability was determined by plating aliquots of the cell suspensions. Each value shown is the mean±standard deviations from duplicates from at least three separate experiments.

In summary, these results indicate that human lactoferrin cannot be considered as a classical pore former as described for many other antimicrobial cationic peptides, thus suggesting that other intracellular or cell membrane component could be the target for this innate defence protein.

Acknowledgements

We thank Dr. A.J. Miranda (Departamento de Química y Física Analítica) for assistance with potassium measurements. This work was supported by University of Oviedo (CN-96–133-B1/Laboratorio de Microbiología Oral).