Correspondence: Maria Lucia Taylor, Laboratorio de Inmunología de Hongos, Departamento de Microbiología y Parasitología, Facultad de Medicina, Universidad Nacional Autónoma de México, Ciudad Universitaria 04510 México DF, Mexico. Tel./fax: + 52 55 5623 2462; e-mail: email@example.com
A nonimmune library, containing single chain variable fragments (scFv) of immunoglobulin human genes displayed on the surface of M13 filamentous phages, was used to recognize molecules exposed on Histoplasma capsulatum yeasts' surface, during their growth in synthetic medium. The scFv clones were checked in their consistency by Dot-ELISA using HRP/anti-M13 conjugate, and they were tested to recognize molecules on H. Capsulatum yeasts' surface by ELISA in plates. Three out of 80 scFv cones (C2, C6, and C52) reacted consistently with H. capsulatum molecules, and they recognized molecules from both H. capsulatum morphologic phases. However, C6 and C52 clones reacted better with molecules on the surface of whole yeasts, with molecules from the yeasts' cell-wall extract, and with molecules released to the supernatant of the yeast culture. Mycelial supernatants from other fungi, as well as from a Mycobacterium filtrate, were not recongnized by scFv phage monoclones. Monoclones C2, C6, and C52 recognized yeast molecules irrespective of the H. capsulatum strains used; the C6 clone revealed a specific immunohistochemistry reaction when tested against homologous and heterologous fungal infected tissues. The scFv clones isolated will be a useful toll to define the role of their target molecules in the host–parasite relationship of histoplasmosis.
Histoplasma capsulatum var. capsulatum, causative agent of ‘histoplasmosis capsulati’, has an intracellular yeast parasitic-phase in infected hosts (Tewari et al., 1998). The interaction of infective or parasitic fungal structures with host-cells could be mediated by several surface molecules from both the fungal and the host cells. Macrophages recognize, bind and ingest H. capsulatum that could be killed by intracellular antiparasitic mechanisms (Newman et al., 1990). However, ingested fungal cells can find access to a permissive intracellular environment for their survival and replication (Eissenberg & Goldman, 1994).
In regard to fungal components that could participate in the interaction with host molecules, a heat-shock protein of 60 kDa (hsp60) has been proposed as a probable fungal ligand that mediates binding to CD18, a 95-kDa molecule of the β2 family of adhesion glycoproteins from human leukocytes (Long et al., 2003). Interaction of carbohydrate-binding proteins (lectin-like molecules) from H. capsulatum with complementary glycosyl residues on the membrane surface of murine macrophages and mammal erythrocytes has been reported (Taylor et al., 1998, 2004; Mendes-Giannini et al., 2000; Duarte-Escalante et al., 2003), although a fungal lectin has not been isolated yet.
Keath et al. (Keath et al., 1989; Keath & Abidi, 1994) used a differential hybridization approach to identify H. capsulatum phase-specific genes. They isolated YPS (yeast phase-specific) genes, of which the most extensively studied is YPS-3. This gene encodes a small protein of 17.4 kDa (Weaver et al., 1996), but its function has not been determined yet. The yps-3 protein is produced only in the pathogenic yeast phase and is expressed differentially in H. capsulatum strains that differ in virulence. This protein is surface localized in virulent and thermotolerant strains (Weaver et al., 1996; Bohse & Woods, 2005). It has been found that yps-3, released from a G-217B culture, binds to the cell wall chitin of strains that do not express it (Bohse & Woods, 2005).
Our present attempt is aimed at detecting surface molecules that are constitutively expressed by H. capsulatum yeasts, hoping to define their role in the first contact with host phagocytes. To accomplish this, a nonimmune human antibody library, containing 1.2 × 109 single chain variable fragments (scFv) displayed on the surface of filamentous M13 phages (Griffin library), was used to recognize different molecules exposed on the yeast surface during its growth in synthetic culture medium.
Materials and methods
Histoplasma capsulatum yeast-phase cells of EH-361 and EH-375 strains from the Fungal Immunology Laboratory Culture Collection (School of Medicine, UNAM, Mexico), as well as two reference strains (Downs and G-217B), were used for the detection of constitutively expressed surface molecules. Yeasts from each −196°C strain stock were grown overnight in BHI-broth (Bioxón, Becton-Dickinson, Mexico City, MX) containing 0.1% cysteine and 1% glucose (Glc), at 37°C under orbital shaking (Lab-Line Environ Shaker, Lab-Line Instruments Inc., Melrose Park, IL) at 200 r.p.m. Harvested yeasts were washed with 150 mM PBS, pH 7.2, and transferred to RPMI-1640 with l-glutamine and 25 mM HEPES (Gibco Laboratories, Grand Island, NY), supplemented with 0.2% of sodium bicarbonate. After overnight incubation, cultured yeasts were centrifuged at 800 g for 10 min, washed with RPMI-1640, counted and adjusted to a desired number to be used in different assays.
Escherichia coli TG1 suppressor strain [K12, Δ(lac-pro), supE, thi, hsdΔ5/F′traD36, proA+B+, lacIq, lacZΔM15] was used as host for the bacteriophage VCS-M13 helper phage and the scFv phages that contain heavy and light immunoglobulin chains. All were a kind gift from Dr Greg Winter from the Medical Research Council Centre (Cambridge, UK).
Preparation of cell-wall extracts from H. capsulatum yeasts
Yeast cells from EH-361, EH-375, Downs and G-217B strains were obtained as described above and resuspended in 10 mL of 10 mM sodium phosphate buffer, pH 7.0, and broken with 212-μm diameter glass beads (Sigma Chemical Co, St Louis, MO) in a Braun MSK cell homogenizer (Braun, Melsungen, Germany) with six alternate periods of breakage and cooling (20 s/2 min). The homogenate was centrifuged at 1000 g for 10 min and the resulting supernatant containing the cell-free extract was saved. The cell walls' residue was resuspended in 3 mL of 50 mM sodium phosphate buffer, pH 7.0, adding 0.02% sodium azide and 0.5 mg lyticase (0.25 mg mL−1), incubated at 37°C for 3 h, and then centrifuged at 105 000 g for 10 min. Supernatants were freeze-dried and kept at −20°C until use.
scFv phage library
A nonimmune scFv library, displayed in pIII proteins on the surface of M13-filamentous phage (Griffin.1 library), was used to monitor molecules on the surface of H. capsulatum yeasts. The Griffin.1 library is a large naive human scFv phagemid library (total diversity 1.2 × 109) constructed from synthetic V-gene segments.
Phages from the original bacterial library stock or from bacteria infected by selected phage clones (see over leaf) were rescued and amplified by infection of phagemid-bearing TG1 cells with VCS-M13. Bacterial cultures in log-phase were infected with the helper phage at a 1:20 TG1:VCS-M13 ratio. Cultures were incubated for 30 min at 37°C without shaking. Transfected cells were harvested and resuspended in 2 × TY (16 g L−1 tryptone, 10 g L−1 yeast extract, and 5 g L−1 NaCl), supplemented with 100 μg mL−1 ampicillin (Amp) and 25 μg mL−1 kanamycin (Kan), and incubated overnight at 30°C under orbital shaking. Cultures were centrifuged and phages were concentrated from the supernatant by polyethylene glycol (PEG) precipitation. Resuspended phages in PBS were stored in aliquots at 4°C, until their use. Finally, for titration, serial dilutions of the phage preparations, either from library or single clones, were added to log-phase TG1 strain, incubated at 37°C for 30 min, and plated on 2 × TY containing Amp and 1% Glc.
scFv phage selection of constitutive molecules on the surface of cultured yeasts
EH-375 H. capsulatum yeasts (1 × 108) were blocked with 500 μL of 10% fish gelatin (Sigma) in PBS, 1 h at room temperature. Simultaneously, scFv phages (1 × 1012) were blocked with 500 μL of 5% fish gelatin in PBS, in the same conditions. Yeasts were washed in PBS and mixed with scFv phages overnight at 4°C. The mixture was centrifuged at 15 300 g for 10 min, washed in PBS/0.1% Tween-20 followed by PBS without Tween, 10 times each. Afterwards, scFv phages were dissociated from yeast cells with 500 μL 0.1 M glycine, pH 2.4, 5 min, at room temperature. Dissociated scFv phages, recovered in the supernatant after centrifuging at 15 300 g for 10 min, were mixed with TG1 and incubated 30 min without shaking, followed by plating on to TYE-Amp agar and incubated overnight at 30°C. Colonies carrying phagemids were harvested with 6 mL of 2 × TY; 100 μL were inoculated into 2 × TY-Amp-Glc and incubated at 37°C under shaking at 200 r.p.m. until 0.5 OD600 nm was reached. Then, VCS-M13 helper phage was added at a 1 : 20 ratio, and incubated at 37°C for 30 min, without shaking. The culture was centrifuged at 800 g for 30 min; the pellet was resupended in 50 mL 2 × TY supplemented with Amp and Kan, and incubated overnight at 30°C, under shaking at 200 r.p.m. Later, 40 mL were centrifuged at 3800 g for 10 min and the supernatant was treated with 10 mL PEG/NaCl (20%/2.5 M), 4 h at 4°C. The supernatant was centrifuged at 7800 g for 10 min and the pellet was recovered and resupended in 2 mL PBS. Finally, the suspension was centrifuged at 15 300 g for 10 min and the supernatant containing the selected scFv phages was stored at −20°C until required.
Dot-enzyme linked immunosorbent assay (Dot-ELISA), for screening scFv phages able to recognize yeast fungal molecules
Firstly, consistency in scFv-phage detection was verified through Dot-ELISA. Briefly, 1 × 108 to 1 × 1010 scFv phage in 100 μL of PBS was applied in separate wells to a 0.45-μm Immobilon-P membrane (Millipore Corporation, MA) using a 96-well microfiltration Bio-Dot apparatus (Bio-Rad, Richmond, CA). Phages binding on to membrane were favored by the apparatus' vacuum production, for 15 min. The membrane was blocked with PBS/0.05% Tween-20 (PBST) pH 7.2, plus 10% fish gelatin (Sigma) for 1 h, at room temperature under gentle shaking, in order to prevent nonspecific binding. The membrane was washed thrice with PBST; HRP/anti-M13 monoclonal conjugate diluted 1:1000 in PBST-1% fish gelatin was applied to the membrane and incubated 18 h at 4°C. The membrane was washed again in PBST and then immersed in a fresh mixture of 25 mg of 3,3′-diaminobenzidine-4HCl (Sigma) prepared in 50 mL of 0.1 M Tris buffer, pH 7.5, plus 50 μL of 30% H2O2. After enzyme–substrate reaction, membranes were washed and dried. Dot-ELISA positive reactions were visible as brown spots.
Finally, Dot-ELISA was used in the initial screening of scFv-phage clones to allow for the recognition of H. capsulatum yeast molecules. The procedure was similar to that mentioned above, with one modification; each sample of yeast fungal cells (EH-375 strain) was applied in separate wells to the Immobilon-P membrane (Millipore) of a 96-well microfiltration Bio-Dot apparatus (Bio-Rad), followed by addition of tested scFv phage dilutions from different clones. Yeasts alone or yeasts incubated with HRP/anti-M13 conjugate were used as negative controls and M13-phages incubated with HRP/anti-M13 conjugate were used as positive control.
Screening of individual scFv clones that recognize constitutive yeast antigens
ELISA (Voller et al., 1979) was used to identify scFv-phage clones able to detect yeast surface molecules. Yeasts (100 μL), from the EH-375 strain, were mixed (v/v) with 50 mM carbonate buffer, pH 9.6, serially diluted in the same buffer into a 96-wells Nunc plate and incubated overnight at 4°C. The plate was then washed thrice with PBS/0.05% Tween-20 (PBST). To block any remaining reactive sites, PBST/3% albumin was added and the plate was incubated for 24 h; afterwards, three more PBST washes were performed and 100 μL scFv phages in PBST/3% albumin were added to the wells and incubated overnight at 4°C. Then, plates were washed twice in PBST, and 100 μL of 1:1000 HRP/anti-M13 monoclonal conjugate in PBST was added to each well and incubated for 2 h at 37°C. After PBST washes, 100 μL of the substrate solution (40 mg o-phenylenediamine in 100 mL of 0.05 M citrate-phosphate buffer, pH 5.0, with 40 μL of 30% hydrogen peroxide) was added. Incubation was performed under dark conditions at room temperature for 20 min. The reaction was stopped by adding 50 μL of 2.5 M sulphuric acid to each well, and OD values were read at 492 nm in a Multiskan MS Labsystems Apparatus (Helsinki, Finland). Assays were set up in duplicate, including yeasts alone and M13-phages as negative and positive antigen controls, respectively.
In order to test the specificity of individual scFv-phage clones to H. capsulatum yeast molecules, different ELISA assays were performed using crude mycelial (M) filtrate (histoplasmin) (Toriello et al., 1982, 1993) and yeast phase culture supernatant (YPCS) from the EH-375 strain of H. capsulatum. Additionally, a purified M-phase antigen (deproteinized polysaccharide-protein complex (D-PPC) (Toriello et al., 1982, 1993) and cell wall extracts from the four different H. capsulatum strains previously mentioned were also tested. Heterologous crude M-phase filtrates of Coccidioides immitis (Ci), Paracoccidioides brasiliensis (Pb), Blastomyces dermatitidis (Bd) (Toriello et al., 1982, 1993), as well as a Mycobacterium tuberculosis (Mtb) culture filtrate, a kind gift from Dr Iris Estrada from the Instituto Politécnico Nacional, Mexico, were also processed.
Each experiment was performed in triplicate and data were plotted after nonspecific OD values had been adjusted, taking into account the controls.
scFv phage interaction with surface molecules of fungal structures in infected tissues
Paraffin-embedded tissue sections of two human mycoses (histoplasmosis and blastomycosis) were dewaxed and dehydrated before retrieval antigenic sites using citrate buffer, pH 7.6, at 90°C for 10 min. Endogenous peroxidase was blocked and tissue sections were washed thrice in PBS and incubated in PBS/3% albumin, for 1 h at room temperature. Specimens were incubated overnight at 4°C with C6 scFv clone diluted 1:50 in PBS. As negative controls, tissue sections were incubated with M13-phages diluted 1:250 in PBS. Reactions were processed by HRP/anti-M13 monoclonal conjugate diluted 1:50 in PBS for 2 h at room temperature, and then revealed by 3,3′-diaminobenzidine/hydrogen peroxide solution. Before adding anti-M13 to each tested tissue section, three PBS washes were performed. Finally, tissue sections were counterstained with Mayer's hematoxylin. Additionally, some tissue sections were processed with PAS stain for morphological identification of fungi.
Data analysis and statistics
Taking into account the adjusted OD data from the different ELISA assays, the SD was calculated for each set of data, which were analyzed using an anova, considering α=0.05 (SPSS® and SYSTAT®, version 11.0, 2004). When a significant difference was found, multiple comparisons using Tukey HSD test (Montgomery, 1991) were applied.
Phage display was first engaged in human pathogenic fungi to identify Candida albicans surface antigens (Haidaris et al., 2001; Bliss et al., 2003), as well as to localize melanin in conidia of Alternaria alternata (Carzaniga et al., 2002). Phage display has also been used successfully to screen and identify binding peptides to the surface of Aspergillus fumigatus (Lionakis et al., 2005) and for the isolation of one human antimannan Fab fragment which, converted to full-length IgG1 antibody, confers resistance to a lethal hematogenous infection by C. albicans in mice (Zhang et al., 2006).
In this paper we selected a monovalent scFv-phage system to detect H. capsulatum yeast surface proteins constitutively expressed during fungal growth at 37°C.
To develop different assays, the Griffin.1 library of scFv phages, which consists of semi-synthetic variable heavy and light chain coding fragments clones into pHEN2 vector, was amplified and a 1 × 1014 CFU mL−1 titer of nonimmune human scFv phagemids was obtained; c. 30% of phagemids contained complete 1000-bp scFv fragments as revealed by PCR (data not shown).
First biopanning with H. capsulatum cultured yeasts produced 1 × 106 scFv phages. After two additional rounds of selection, 80 scFv-phage monoclones were recovered and mixed in eight pools, containing 10 clones each, in order to screen for pools that contained the best antibody fragments able to react with yeast surface molecules.
Dot-ELISA, using a M13-HRP/anti-M13 system, was first performed to check the sensitivity of the system, and then used to detect scFv-phage clones that recognize yeast surface molecules. Under these experimental conditions, a minimum of 1 × 108 phage particles were revealed by HRP/anti-M13 conjugate. Recognition of H. capsulatum yeast molecules by scFv-phage clones was also revealed by Dot-ELISA assays, and 1 × 107 yeast cells was found as the minimal antigen concentration necessary for the detection of scFv-phage clones (Fig. 1).
Taking into account the previous Dot-ELISA data, a more quantitative method to detect the reaction between constitutive H. capsulatum yeast molecules and scFv clones was developed using ELISA in plates. Analysis of results from Fig. 2a by anova showed significant differences (P<0.0001) among the eight pools of scFv-phage clones tested. Two out of eight pools of scFv-phage clones, P-I and P-V, reacted preferentially with molecules on the yeast surface, using different amounts of yeast. P-I and P-V pools were further tested in three different dilutions, which confirmed their consistency to detect specific interactions (Fig. 2b). In contrast, P-VIII, included as a weakly reacting control, was only reactive with the highest phage concentration. Comparisons among P-I, P-V and P-VIII scFv-phage pools and the positive control C (+) showed again a significant difference (P<0.0001). In consequence, individual clones from pools P-I and P-V were selected for further analyses. Four clones from P-I (C1, C2, C4, and C6) and two clones from P-V (C52 and C54) reacted with epitopes present in cultured yeasts (Fig. 3a and b). anova revealed significant differences (P<0.0001) among all individual scFv-phage clones depicted in Fig. 3a and b. After several assays using different concentrations of yeasts and phage particles, from the six clones selected only three were found to be stable (C2, C6, and C52); for this reason they were tested in the following assays.
Individual scFv clones were tested in their specificity towards antigens from homologous (histoplasmin, M-phase D-PPC and Y-phase culture supernatant) and heterologous mycelial supernatants of C. immitis, P. brasiliensis and B. dermatitidis, as well as from M. tuberculosis filtrate. Recognition of homologous antigens was detected in the three clones tested (Fig. 4a–c). Clones C6 and C52 reacted mainly with Y-phase antigen released in the culture supernatant (Fig. 4b and c), whereas clone C2 did not react with this antigen, although it was able to recognize molecules on the surface of whole yeasts (Fig. 3). Moreover, clone C2 reacted better with M-phase histoplasmin than clones C6 and C52 (Fig. 4a). This last result establishes the difference between molecules recognized by C2 and/or C6 and C52 clones. None of the three clones reacted with the purified D-PPC H. capsulatum antigen (Fig. 4a–c), suggesting that this fraction no longer contains the epitope recognized by the three scFv clones in the histoplasmin. All tested M-phase heterologous antigens showed the lowest OD values, supporting the clones' specificity (Fig. 4a–c).
In general, OD values using scFv clones were rather low when compared to those obtained with sera from infected patients. This could be explained by taking into account that scFv antibody fragments were obtained from a nonimmune library and, therefore, their affinities were similar to those derived from a primary immune response. Affinity could be further improved by means of directed mutagenesis. Besides, serum antibodies are polyclonal, in contrast to phage-displayed antibodies which are monoclonal.
Selected clones were tested in their ability to recognize constitutive yeast molecules of different H. capsulatum strains. Figure 5 shows that selected scFv clones interacted with cell-wall fractions of the four strains used, emphasizing that C6 presented less reaction with H. capsulatum Downs strain components, even though no significant difference (P>0.05) was detected among the fractions of the four strains tested. These findings also suggest that antigens recognized on the four strains tested are surface-conserved molecules of H. capsulatum and point out a distinction between clones C6 and C52.
It is probable that fungal molecules involved in the scFv-phage clone's recognition are constitutive surface components, probably associated with the fungal cell wall. According to the present results it is possible to suggest that the ligand for the C6 scFv clone is a released molecule just like yps-3 protein (Bohse & Woods, 2005), except that this molecule is also found in the M-phase of H. capsulatum.
PAS staining demonstrated that tested tissue sections corresponded to histoplasmosis (Fig. 6a) and blastomycosis (Fig. 6d). Only H. capsulatum yeasts were recognized by the C6 scFv clone, which had previously shown the best reaction with H. capsulatum yeast surface molecules, depicting positive peroxidase immunohistochemistry reactions on yeasts' surface (Fig. 6b). In contrast, B. dermatitidis' (Fig. 6e) fungal structures were always negative, as were the control tissue sections incubated with M13 phages (Fig. 6c and f). Several dilutions of C6 scFv clone and M13 phage (negative control) were tested ranging from 1:25 to 1:500, and the best dilution to reveal C6 scFv positive reaction was 1:50. All M13 phage dilutions tested were systematically negative.
The phage-display methodology represents an advantage over other methods that reveal gene transcription, as it permits also to isolate genes and detect their cognate proteins. The isolated scFv-fragments allow for the possibility of developing further studies to define the role of these molecules in the host-parasite relationship of histoplasmosis.
This research was partially supported by CONACYT-Mexico (Ref: 43944-M). The authors thank Ingrid Mascher for editorial assistance, as well as Martha Ustarroz Cano for technical support with the immunohistochemistry. E. Curiel-Quesada is a COFAA-IPN fellow.