Cloning and Characterization of a Gene Encoding an Immunoglobulin-Binding Receptor on the Cell Surface of Some Members of the Family Trypanosomatidae

Abstract

Several members of the Trypanosomatidae family, when freshly isolated from their mammalian hosts, have immunoglobulins adsorbed to their cell surfaces. However, a significant portion of these antibody molecules is not parasite specific, i.e., the immunoglobulins are bound to the parasite's cell surface molecules via noncognitive interactions. It has been proposed that this noncognitive adsorption of immunoglobulins to the parasite is mediated by an Fc-like receptor present in several members of the Trypanosomatidae family. However, the molecular identification of this receptor has never been defined. Here, we describe the cloning of a gene encoding a protein that might represent this molecule. The gene, named Lmsp1, was cloned by screening a Leishmania major cDNA expression library using a rabbit antiserum. Lmsp1 is present in both Leishmania and Trypanosoma and is expressed in all developmental stages of these parasites. The predicted protein has a molecular mass of 16.6 kDa and contains an RGD sequence starting at residue 104 and three cysteine residues at positions 55, 74, and 116. The purified recombinant protein strongly binds to normal immunoglobulins of various animal species (humans, rabbits, sheep, goats, guinea pigs, donkeys, rats, and mice) and the binding to human immunoglobulins appears to be immunoglobulin G (IgG) and IgM isotype specific. Moreover, Lmsp1 binds to both purified Fc and Fab fragments of IgG from both humans and rabbits. The mapping of the Lmsp1 epitopes that bind human IgG revealed that different sequences of the molecule bind to Fc or Fab. In addition, fluorescence-activated cell sorter analyses with a specific rabbit anti-Lmsp1 antiserum showed that Lmsp1 is associated with the parasite's cell surface. Finally, inhibition experiments point to an active role of this molecule in the immunoglobulin-mediated attachment and penetration of Trypanosoma cruzi in its macrophage host cells, thus suggesting that Lmsp1 is a putative Trypanosomatidae immunoglobulin receptor.

The mechanisms that intracellular parasites have developed to both interact with the host cells and escape immune surveillance are complex and intriguing. Cell surface ligands synthesized by the parasites themselves as well as molecules acquired from the host have been described as participating in the parasite's internalization in the host cells and in the escape from the host defense mechanisms. Examples of parasite-derived molecules are the cell surface mannan-fucose glycoproteins of various parasites and the leishmanial major surface protein gp63 (16, 38). These molecules facilitate the parasite's internalization in the host cells by binding to the mannose-fucose receptor and to complement receptor 3, in particular (30, 37, 39). Examples of host molecules involved in the parasite's escape mechanisms are the human blood group antigens, serum proteins, and major histocompatibility antigens (8, 15, 31, 34). In addition, most Trypanosomatidae that are pathogenic for mammals, when freshly isolated from their hosts, have immunoglobulin adsorbed to their cell surfaces (4, 9-12, 18, 19). Interestingly, a significant portion of these antibody molecules is apparently not parasite specific (4, 35); i.e., they are bound to the parasite's cell surface via the noncognitive regions of the antibody molecules. It is believed that these parasite ligands offer both an effective mechanism for antigen mimicry of the host antigens and an effective system for the internalization of the parasites in their target host cells. One possible parasite noncognitive ligand of immunoglobulins is an Fc-like receptor present on the cell surface of several members of the Trypanosomatidae (20, 25, 35). Indeed, immunoglobulins and purified Fc fragments of immunoglobulin G (IgG) have been shown to facilitate the internalization of Trypanosoma cruzi in their host cells and to consequently increase the infective capacity of these parasites (1, 22, 23, 35).

In summary, these results suggest that Trypanosomatidae organisms developed unique mechanisms to utilize the host immunoglobulins bound to their cell surface either via Fab or Fc fragments to facilitate the infection as well as to evade the lethal effects of the antibody-mediated immune response. Notwithstanding, up to now, the existence of a cell surface molecule in Trypanosomatidae that binds noncognitive regions of the immunoglobulin molecules has only been based on circumstantial or indirect evidences.

In the present study, we describe the cloning of the gene and the characterization of a Trypanosomatidae recombinant protein that binds, in a noncognitive manner, both Fc and Fab fragments of IgG. The gene is present and is expressed in both Leishmania and Trypanosoma, and its encoded protein is associated with the parasite's cell surface and is apparently involved in the IgG-mediated internalization of T. cruzi in its macrophage host cells.

MATERIALS AND METHODS

Microorganisms.

Leishmania major and Leishmania amazonensis were maintained in vivo in BALB/c mice. Mice were infected in the rear footpad with approximately 10⁴ amastigote forms of the parasites freshly obtained from the lesions of previously infected mice (7). Amastigotes were prepared (enriched) by differential centrifugation. Leishmania chagasi was maintained in vivo in golden hamsters. Hamsters were infected intracardially with approximately 10⁷ amastigote forms of the parasites freshly obtained from the spleens of previously infected hamsters (5). Amastigotes were prepared (enriched) by centrifugation of disrupted spleen cells over a Percoll gradient. Promastigote forms of all Leishmania species were obtained from cultures in Schneider's medium. T. cruzi (Y strain) epimastigote, amastigote, and trypomastigote forms were obtained as described previously (32).

Opsonization and phagocytosis of T. cruzi by murine macrophages.

An IgG fraction from human serum obtained from a patient with chronic Chagas disease was prepared by affinity chromatography with protein A Sepharose. T. cruzi trypomastigotes were incubated with 2 μg of purified IgG/ml in the presence and absence of the indicated concentrations of synthetic peptides for 30 min at room temperature. Thioglycolate-induced peritoneal exudate murine macrophages were obtained from BALB/c mice as described previously (2). Cells were plated at 2 × 10⁵ cells/well on eight-well chamber slides (Lab-Tek, Fort Washington, Pa.) and incubated for 2 h. Nonadherent cells were removed and counted to infer the number of adherent cells per well. Opsonized T. cruzi trypomastigotes, plus or minus peptides, at a multiplicity of infection of 1/1 were added to the cells and followed by subsequent incubation for 4 h. Noninternalized parasites were washed out. Slides were methanol fixed and Giemsa stained. Macrophages containing one or more parasites were scored out of 500 adherent cells.

Serological reagents.

Normal sera from rabbits, guinea pigs, donkeys, goats, sheep, rats, and mice were purchased from Sigma, St. Louis, Mo. Normal human gamma globulin was from Jackson ImmunoResearch Laboratories Inc., West Grove, Pa. Highly purified human and rabbit Fc and Fab fragments of IgG were obtained from ICN Pharmaceuticals, Inc., Aurora, Ohio. Purified human IgG, IgA, IgM, and IgE immunoglobulin isotypes were from Calbiochem, La Jolla, Calif. Peroxidase-labeled donkey F(ab′)₂ antisera specific for human, rabbit, guinea pig, goat, sheep, rat, and mouse immunoglobulins were from Zymed, South San Francisco, Calif. Rabbit anti-L. major secreted proteins and anti-purified leishmanial recombinant protein antisera were prepared as previously described (33). Briefly, 100 μg of either L. major secreted proteins or purified leishmanial recombinant protein were mixed with 100 μg of muramyl dipeptide, brought up to 1 ml with phosphate-buffered saline (PBS), and emulsified with 1 ml of incomplete Freünd's adjuvant. The emulsions were injected in multiple subcutaneous sites into female New Zealand rabbits (R&R Rabbitry, Stanwood, Wa.). The rabbits were given two subcutaneous boosters (100 μg of the corresponding antigen in incomplete Freünd's adjuvant) 6 weeks apart followed by one intravenous injection of 100 μg of antigen. One week after the final boost, the rabbits were sacrificed by exsanguinations and sera were collected and stored at −20°C. Purified IgG from both the preimmune and immune rabbit sera was prepared by using Staphylococcus aureus protein A columns (Montage antibody purification; Millipore Corporation, Bedford, Mass.) as instructed by the manufacturer.

Library generation and serological expression screening.

Anti-Escherichia coli antibody reactivity was removed from the rabbit serum by preadsorption on nitrocellulose filters containing lysed E. coli. The absorbed serum was initially evaluated by Western analysis on both promastigote lysate and the secretory shed antigens of L. major. The results showed that the rabbit serum was reactive with seven dominant antigens with molecular masses ranging from ∼10 to >200 kDa (data not shown). The serum was subsequently used to screen an L. major cDNA expression library. This library was made from L. major promastigote RNA by using the unidirectional Lambda ZAP (uni-ZAP) kit as suggested by the manufacturer (Stratagene). A total of 70,000 PFU of the amplified cDNA library was screened with the rabbit serum at a 1:250 dilution. Nineteen positive picks were confirmed in the tertiary screening. The phagemid was excised, and the DNA from each of the 19 clones was sequenced. The results from the sequencing revealed that 17 of the 19 clones represented sequences of one single gene, which was designated L. major secreted protein gene 1 (Lmsp1 gene).

Recombinant protein expression and affinity purification.

The full-length cDNA clone of Lmsp1 gene was PCR amplified by using as a template the phagemid and specific oligonucleotides for directional cloning into a protein expression vector. The sequences of the oligonucleotide primers were as follows: 5′ (5′-CAA TTA CAT ATG CAT CAC CAT CAC CAT CAC ATG TCC TCC GAG CGC ACC TTT ATT-3′) and 3′ (5′-GTA CGG ATA TCG TAA AAC GAC GGC CAG T-3′). The 5′ oligonucleotides contained an NdeI restriction site (underlined) preceding the ATG initiation codons followed by nucleotide sequences encoding six histidines and the sequence derived from the gene (in italics). The 3′ oligonucleotide contained the M13 −20 primer and an EcoRV restriction site. The resultant PCR product (∼1.0 Kbp) was digested with NdeI and EcoRV and subcloned into the pET17b vector similarly digested with NdeI and EcoRV for directional cloning. Ligation products were initially transformed into E. coli XL1-Blue competent cells (Stratagene) and were subsequently transformed into E. coli BL21(pLysE) host cells (Novagen, Madison, Wis.) for protein expression. The recombinant Lmsp1 protein was purified from isopropyl-β-d-thiogalactopyranoside (IPTG)-induced culture by using preparative sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) followed by excision of the respective band containing the overexpressed protein followed by electroelution as previously described (6). The purity of the recombinant proteins was assessed by SDS-PAGE and followed by Coomassie blue staining and N-terminal sequencing by Edman chemistry with a Procise 494 protein sequencer (Perkin Elmer-Applied Biosystems, Inc., Foster City, Calif.). The recombinant Lmsp1 was assayed for endotoxin contamination by using the Limulus amoebocyte assay (BioWhittaker, Walkersville, Md.) and was shown to contain <100 EU of protein/mg.

Immunoblot analysis.

L. major promastigote-secreted proteins, whole leishmanial cell extracts (promastigote and amastigotes forms of L. major, L. amazonensis, and L. chagasi), whole T. cruzi cells (epimastigotes, trypomastigotes, and amastigotes), and the purified recombinant Lmsp1 were separated by electrophoresis on SDS-16% PAGE gels and transferred to nitrocellulose with a semidry transfer apparatus (Bio-Rad Laboratories, Hercules, Calif.). Blots, in duplicate, were blocked for a minimum of 1 h with PBS-1% Tween 20 and probed with either rabbit anti-Lmsp1 antiserum or with the preimmune serum from the same rabbit. Reactivity was assessed by using ¹²⁵I-labeled protein A followed by autoradiography.

Cytofluorometric analysis.

The expression of Lmsp1 on the parasite's cell surface was examined by a standard procedure on a FACSCalibur (Becton Dickinson) as described previously (6). Briefly, L. major promastigotes were washed in Hanks balanced salt solution (Gibco, Grand Island, N.Y.) and suspended to a density of 5 × 10⁶ parasites per ml followed by 1 h of incubation at 4°C with Hanks or various dilutions (in Hanks) of either a rabbit anti-Lmsp1 antiserum or a normal rabbit serum (preimmune serum). Parasites were washed and then incubated for 30 min at 4°C with fluorescein isothiocyanate-labeled donkey F(ab′)₂ anti-rabbit IgG. Stained cells were washed twice and analyzed by fluorescence-activated cell sorter (FACS).

Binding of purified leishmanial recombinant protein to normal immunoglobulins.

Binding of the recombinant leishmanial protein to whole immunoglobulins and immunoglobulin fragments was carried out primarily by enzyme-linked immunosorbent assay (ELISA). Ninety-six-well Nunc-Immuno ELISA plates, MaxiSorp surface (Nalge Nunc International, Roskilde, Denmark), were coated at room temperature for 2 h with 200 ng of the recombinant protein/well. Free reactive groups on the plates were quenched by using PBS containing 0.05% Tween and 2% bovine serum albumin. Plates were then incubated for 2 h with various concentrations of either normal sera or purified normal immunoglobulin diluted in quenching buffer. Wells were washed five times with quenching buffer followed by incubation with peroxidase-labeled specific anti-immunoglobulin antibody. Wells were washed, and reactions were developed by using the substrate H₂O₂ and the chromogen 3,3′,5,5′-tetramethylbenzidine (TMB) and read at 450 nm.

Peptides.

Overlapping peptides containing 20 amino acids covering the entire Lmsp1 molecule were synthesized on a Rainin Symphony multiple-peptide synthesizer by using the O-benzotriazole-N,N,N′,N′-tetramethyluronium hexafluorophosphate activation system. Cleavage of the peptides from the solid support was carried out with the following cleavage mixture: trifluoroacetic acid-ethanedithiol-thioanisole-water-phenol (40:1:2:2:3). After cleaving for 2 h, the peptides were precipitated in cold ether. The peptide pellets were then dissolved in 10% (vol/vol) acetic acid and lyophilized prior to purification by C₁₈ reverse-phase high-performance liquid chromatography. A gradient of 0 to 60% acetonitrile in water (containing 0.1% trifluoroacetic acid) was used to elute the peptides. The purity of the peptides was verified by high-performance liquid chromatography and mass spectrometry. Each peptide was purified to >95% before use. The peptides contain an overlapping sequence of 10 amino acids with each other. A total of 14 peptides were synthesized.

RESULTS

Serological expression cloning of Lmsp1.

It was recently shown that culture supernatants of either L. major or L. chagasi contain antigens (secreted antigens) that can protect BALB/c mice against challenge with L. major (36). To search for such antigens, a rabbit antiserum was raised against the supernatant of L. major incubated in protein-free medium for 2 days. This antiserum was then used to screen an L. major cDNA expression library. One clone in particular, which was designated Lmsp1 (for L. major-secreted protein) was selected because it was repeatedly and preferentially identified by the rabbit antiserum. Figure 1A shows the full-length and predicted amino acid sequence of the Lmsp1 cDNA. The EcoRI/XhoI insert is 1,019 bp long containing the following features: (i) the last 17 nucleotides of the spliced leader sequence characteristic of all Trypanosomatidae nuclear mRNA, (ii) 39 nucleotides of 5′ untranslated sequence, (iii) a 453-bp open reading frame coding for a 151-amino-acid sequence, and (iv) a stop codon followed by 471 bp of 3′ untranslated sequence terminating with a poly(A) tail. The full-length open reading frame encodes a protein with a predicted molecular mass of 16.64 kDa, an isoelectric point of 7.86, and a net charge of 1.33 at pH 7.0. The predicted protein contains three potential phosphorylation sites (S/T-X-R/K) at amino acid residues 3, 85, and 102. In addition, the predicted protein sequence of Lmsp1 has an RGD sequence starting at residue 104. This tripeptide sequence is known to bind to several proteins, particularly those encoded by genes of the integrin family (28). Moreover, amino acid sequence comparison analyses revealed that Lmsp1 shares 65 to 70% homology with a eukaryotic nucleoside diphosphate kinase protein. Unexpectedly, no obvious leader sequence at the amino-terminal portion of the protein could be identified. Alignment of this sequence with those of T. cruzi and Trypanosoma rucei revealed high homology spanning throughout the molecule, with diversion occurring only at the very end of the C terminus portion of the T. cruzi homologue (Fig. 1B). Moreover, Southern blot analyses revealed that the Lmsp1 gene, in addition to L. major, is present in the following species of the Trypanosomatidae family thus far tested: Leishmania tropica, L. amazonensis, Leishmania brasiliensis, L. chagasi, T. cruzi, and T. brucei (data not shown).

FIG. 1.

Nucleotide sequence and deduced amino acid sequence of the Lmsp1 gene (A) and alignment with T. cruzi and T. brucei homologues (B). In panel A, numbers on the right refer to the nucleotide positions. The spliced leader sequence, and the 5′ untranslated (5′-UT) sequence are indicated immediately before the putative translation initiation codon (ATG). An RGD sequence is indicated in the protein sequence in the boxed residues. The stop codon followed by the 3′ untranslated (3′-UT) sequence and the poly(A) tail are also indicated. In panel B, numbers on the right refer to the amino acid positions. The nonshaded lower line highlights the degree of homology between the genes.

The open reading frame of the full-length gene was amplified by PCR with 5′- and 3′-end-specific oligonucleotides and cloned into the pET17b expression vector. Unfortunately, the recombinant protein could not be obtained by affinity chromatography because of insolubility problems. However, the purification of the recombinant Lmsp1 was achieved by using preparative SDS-PAGE followed by excision of the respective band and electroelution of the protein. Approximately 500 μg of purified Lmsp1 was obtained from 1 liter of E. coli culture. Figure 2 shows a Coomassie blue-stained SDS-PAGE gel of the E. coli cultures before and after induction as well as the purified recombinant antigen.

FIG. 2.

Expression of the Lmsp1 gene as a recombinant protein. E. coli (BL21/pLysE) transformed with the expression vector pET containing the Lmsp1 gene was grown and induced with IPTG. Cells were lysed, and the Lmsp1 protein was purified by using preparative SDS-PAGE followed by excision and electroelution of the respective band containing the overexpressed protein. Expression and purification were evaluated by SDS-PAGE (10%) under reducing conditions, and the gel was stained with Coomassie blue. Lane 1, noninduced E. coli lysate; lane 2, induced E. coli lysate; lane 3, purified recombinant Lmsp1 protein. Numbers on the left side indicate the molecular masses of the markers. The arrow points to a major band of an ∼16-kDa protein present in both lanes 2 and 3 and absent in lane 1.

Expression of native Lmsp1 in Trypanosomatidae organisms.

To investigate the expression of Lmsp1 in Trypanosomatidae organisms, the purified recombinant protein was used to raise a rabbit anti-Lmsp1 antiserum. The serum was initially adsorbed with E. coli and then tested against the purified recombinant molecule and against a crude lysate of promastigote and amastigote forms of L. major, L. chagasi, and L. amazonensis as well as crude lysates of epimastigote, trypomastigote, and amastigote forms of T. cruzi. Figure 3A shows that the preabsorbed antiserum at a 1/500 dilution reacted with the recombinant Lmsp1 as well as with a 16-kDa single molecule present in the crude antigenic preparations of all developmental forms of all Trypanosomatidae family members tested under reduction conditions. However, when both the purified recombinant Lmsp1 and a crude antigenic preparation of L. major promastigotes were separated under nonreduction conditions, several bands of increased molecular masses were detected (Fig. 3B), indicating that both the recombinant Lmsp1 and the native molecule aggregate or form several homopolymers. No reaction to either the recombinant antigen or crude lysate was observed with the preimmune rabbit serum diluted at 1/500 (data not shown). However, at lower dilutions (≤1/100) this serum reacted with several proteins, including a 16-kDa molecule. This band was particularly strong with the crude lysates of both promastigote and amastigote forms of L. chagasi and with both trypomastigote and amastigote forms of T. cruzi (data not shown), thus suggesting that Lmsp1 also binds normal immunoglobulins. However, because the normal serum reacted with multiple bands, it is not possible to conclude that this 16-kDa molecule is indeed Lmsp1.

FIG. 3.

Western blot analysis of Lmsp1. Two micrograms of either purified recombinant Lmsp1 (rAg), crude secreted proteins of L. major (CSP) and crude antigenic preparation of the different forms of the indicated members of the Trypanosomatidae family were submitted to electrophoresis under reducing conditions (A) and nonreducing conditions (B) in 16% gels and transferred to nitrocellulose membranes. Proteins were identified by using a rabbit anti-Lmsp1 recombinant protein. Reactivity was detected with ¹²⁵I-labeled protein A followed by autoradiography. Lc, L. chagasi; Lm, L. major; La, L. amazonensis; Tc, T. cruzi; p, promastigote; a, amastigote; e, epimastigote; t, trypomastigote; Lm_p, L. major promastigotes. Numbers on the left side indicate the molecular masses of the markers. The arrow points to a major band of an ∼16-kDa protein present in all forms and species of the Trypanosomatidae family.

Cell surface expression of Lmsp1 on L. major.

To investigate a possible presence of cell surface expression of Lmsp1 in Trypanosomatidae parasites, FACS analysis was performed with purified rabbit IgG anti-recombinant Lmsp1 antibody. Promastigote forms of L. major were washed and incubated with medium or with various concentrations of either purified normal rabbit IgG or purified rabbit IgG anti-Lmsp1 antibody. The IgG molecules were purified from serum samples obtained from separated bleedings of the same rabbit (before and after immunization, respectively). Binding of the rabbit IgG to the promastigotes was revealed by using fluorescein-labeled goat F(ab′)₂ anti-rabbit immunoglobulins. Figure 4 shows the results and points to an abundant binding of normal rabbit IgG (preimmune) to the parasite's surface. However, further binding, as indicated by marked increase in fluorescence intensity, was observed with all concentrations of the rabbit IgG anti-Lmsp1. Indeed, the fluorescence intensity given by the lower concentrations of the specific IgG (50 to 12.5 μg/ml) was clearly superior to that given by the highest concentration of the normal IgG (100 μg/ml) purified from the rabbit serum before immunization with Lmsp1. These results support previous observations, which proposed the presence of a noncognitive immunoglobulin-binding ligand on the cell surface of Trypanosomatidae, and indicate that Lmsp1 is present on the parasite's cell surface.

FIG. 4.

Surface expression of Lmsp1 on L. major. Indirect immunofluorescence (FACS analysis) was performed with live L. major promastigotes. Parasites were incubated with either medium (dotted line), preimmune IgG purified from rabbit serum before immunization with recombinant Lmsp1 (solid line), or IgG anti-Lmsp1 purified from rabbit serum after immunization with Lmsp1 (shaded histogram). Panels A, B, C, and D show the concentrations of both preimmune IgG and anti-Lmsp1 IgG (100, 50, 25, and 25.5 μg/ml, respectively) used to stain the parasites. After reaction with the immunoglobulins, parasites were washed and incubated with purified fluorescein isothiocyanate-conjugated goat F(ab′)₂ anti-rabbit immunoglobulins (2 μg/ml). The fluorescence intensity was analyzed by FACScan.

The strikingly high background observed with normal immunoglobulins in both Western blot and FACS analyses was also seen when ELISA was used for the titration of the rabbit anti-leishmanial Lmsp1 antibody (data not shown). Together, these observations suggested that the recombinant protein either is recognized by natural antibodies or binds to noncognitive domains of the immunoglobulin molecules.

Binding of Lmsp1 to normal immunoglobulins of various mammals.

To begin to characterize the nature of the Lmsp1 binding to immunoglobulins of nonimmunized animals, a standard sandwich ELISA was used. Plates (96 wells/plate) were coated with 200 ng of Lmsp1 followed by incubation with a 1/20 dilution of normal serum obtained from various mammals (rabbits, humans, guinea pigs, donkeys, goats, sheep, rats, and mice). Peroxidase-labeled species-specific anti-immunoglobulin was used to detect the binding of the various immunoglobulins to Lmsp1. Figure 5 shows the results and indicates that Lmsp1 binds to the immunoglobulins of all tested animals. When a serial dilution analysis was performed, a positive ELISA signal (arbitrarily defined as an optical density at 450 nm of 0.2 above background signal) was observed up to a 1/320 dilution for the sera from most species. However, no ELISA signal was observed at dilutions above 1/80 with sera from donkeys, mice, and rats (data not shown).

FIG. 5.

Binding of Lmsp1 to immunoglobulins of various mammals. ELISA plates were coated with 200 ng of Lmsp1. Plates were washed and blocked. Normal sera (1/20 dilution) from various species were added to individual wells (in triplicate for each serum) and incubated for 1 h at room temperature. Wells were washed with blocking buffer and incubated for 45 min with species-specific horseradish peroxidase-labeled IgG anti-immunoglobulin antibodies. The washing of the plates was followed by the addition of substrate (H₂O₂) and chromogen (TMB). The reaction (OD) was read at 450 nm. The standard error of the triplicate OD readings was <5%. This is a representative experiment of five separate experiments with essentially the same results.

To investigate further the binding of Lmsp1 to immunoglobulins, ELISA was performed with isotype-specific anti-human IgG, IgA, IgM, and IgE antibodies. ELISA plates were coated with Lmsp1 (200 ng/well), blocked, and incubated with 10 μg of purified human IgG, IgM, IgA, and IgE/ml. Detection of the isotypes bound to Lmsp1 was carried out by using isotype-specific peroxidase-labeled donkey F(ab′)₂ antibody. Results are shown in Fig. 6 and indicate that Lmsp1 strongly binds to both IgG and IgM. In contrast, very weak or no binding was observed with IgA and IgE.

FIG. 6.

Selective binding of Lmsp1 to human immunoglobulin isotypes. ELISA plates were coated with Lmsp1 (200 ng/well), blocked, and incubated in triplicate with 10 μg of purified human IgG, IgM, IgA, and IgE/ml. Detection of the isotypes bound to Lmsp1 was carried out with isotype-specific peroxidase-labeled donkey F(ab′)₂ antibody. The washing of the plates was followed by the addition of substrate (H₂O₂) and chromogen (TMB). The reaction (OD) was read at 450 nm. The standard error of the triplicate OD readings was <5%. This is a representative experiment of two separate experiments with essentially the same results.

Because of the broad reactivity with immunoglobulins of various mammals and with both human IgG and IgM immunoglobulin isotypes, these results point to a binding of Lmsp1 to domains other than the antigen-binding site of the antibody molecules, i.e., to the noncognitive constant regions of the heavy or light chains of the immunoglobulins.

Characterization of the IgG fragments that bind Lmsp1.

Purified Fc and Fab fragments of both human and rabbit origins were used to define the regions of IgG that bind to Lmsp1. Various concentrations of Fc and Fab were tested against fixed amounts of Lmsp1 in an ELISA format with species-specific peroxidase-labeled donkey F(ab′)₂ anti-Fc or Fab antibody. Figure 7 shows that regardless of the immunoglobulin origin (human or rabbit), both fragments of IgG bind to Lmsp1. No inhibition of the binding of Fc to Lmsp1 was observed with the addition of an excess of Fab. Likewise, an excess of Fc did not inhibit the binding of Fab to Lmsp1, thus excluding cross contamination in both IgG fragments (data not shown).

FIG. 7.

Binding of Lmsp1 to purified human and rabbit Fc and Fab. ELISA plates were coated with 200 ng of Lmsp1. Plates were washed and blocked. Several concentrations of purified human (A) and rabbit IgG (B) Fc and Fab fragments were added to the individual wells (in triplicate) and incubated for an hour at room temperature. Wells were washed with blocking buffer and incubated for 45 min with horseradish peroxidase-labeled donkey F(ab′)₂ anti-human or -rabbit Fc or Fab antibodies. The washing of the plates was followed by the addition of substrate (H₂O₂) and chromogen (TMB). The reaction (OD) was read at 450 nm. The standard error of the triplicate OD readings was <5%. This is a representative experiment of three separate experiments with essentially the same results.

Peptide mapping of the Lmsp1 binding sites to Fc and Fab.

Once it was determined that both Fc and Fab bind to Lmsp1, it became interesting to map the peptide sequences of Lmsp1 that bind to these two distinct fragments of IgG. To achieve this, peptides containing 20 amino acids covering the entire Lmsp1 molecule were synthesized. In addition the peptides contained an overlapping sequence of 10 amino acids with each other. A total of 14 peptides were obtained and tested by ELISA for the ability to bind to purified Fc and Fab fragments of IgG. ELISA plates were coated with the individual peptides (200 ng/well), blocked, and incubated with 10 μg of either Fc or Fab/ml. After the proper washes, reactions were detected with peroxidase-labeled donkey F(ab′)₂-specific antibodies. Figure 8 shows the results and indicates that when the wells of the ELISA plates were coated with peptide 1 (amino acids [aa] 1 to 20), peptide 4 (aa 31 to 50), peptide 5 (aa 41 to 60), peptide 6 (aa 51 to 70), peptide 12 (aa 111 to 130), peptide 13 (aa 121 to 140), and peptide 14 (aa 131 to 151), no signal above the background could be detected for wells incubated with either Fc or Fab molecules. In contrast, the overlapping peptides 2 to 3 (aa 11 to 30 and 21 to 40, respectively) and peptides 7 to 8 (aa 61 to 80 and 71 to 90, respectively) reacted primarily with and Fab. Conversely, overlapping peptides 10 to 11 (aa 91 to 110 and 101 to 120, respectively) reacted preferentially with Fc. The overlapping peptides 2 and 3 appear to react equally to Fab (same optical density). In contrast, the signals given by peptide 7 for both immunoglobulin fragments were much stronger than the signals given by its overlapping peptide 8. Similarly, peptide 11 reacted more strongly with Fc and Fab than did its overlapping peptide 10. It is interesting that both peptides 10 and 11 include the tripeptide RGD, an amino acid sequence that binds several proteins encoded by genes of the integrin family (28).

FIG. 8.

Peptide mapping of Lmsp1 binding site to human IgG Fc and Fab. Wells of ELISA plates were coated with 200 ng (per well) of overlapping peptides (20mers) covering the entire Lmsp1 molecule. A total of 14 peptides were used (top panel). The RGD motif is underlined in peptides 10 and 11. ELISA plates were blocked, and wells (in triplicate for each peptide) were incubated in a triplicate manner with 10 μg of either Fc or Fab/ml. Wells were washed with blocking buffer and incubated for 45 min with horseradish peroxidase-labeled donkey F(ab′)₂ anti-human Fc or Fab antibodies. The washing of the plates was followed by the addition of the substrate (H₂O₂) and chromogen (TMB). The reaction (OD) was read at 450 nm (lower panel). The standard error of the triplicate OD readings was <5%. This is a representative experiment of three separate experiments with essentially the same results.

These results indicate that the binding of Lmsp1 to Fab and Fc fragments of IgG is mediated by distinct amino acid sequences of Lmsp1 and further confirm, at the amino acid level, that this leishmanial protein binds to noncognitive domains of IgG.

Inhibition of IgG-mediated opsonization of T. cruzi by Lmsp1 IgG-binding peptides.

Among the several Trypanosomatidae parasites, T. cruzi is perhaps the species that most clearly utilizes the immunoglobulin bound to the cell surface to facilitate interaction with the host cells (1, 22, 23). Therefore, these parasites, instead of L. major, were used in inhibition experiments designed to investigate the participation of Lmsp1 in this step of parasite biology. Moreover, as suggested by Western blot analyses (Fig. 3), an Lmsp1-like protein is also produced by T. cruzi. To perform these experiments, culture forms of T. cruzi trypomastigotes were opsonized with specific human anti-T. cruzi IgG in the presence or absence of either the Fab-binding peptide 7 or the Fc-binding peptide 11. Parasites were then exposed to freshly obtained peritoneal macrophages from BALB/c mice. Inhibition of internalization of parasites was then analyzed (scored) by direct microscopy in Giemsa-stained preparations. Figure 9A shows the results and indicates that peptides 7 and 11 used individually at 50 μg/ml greatly inhibited the penetration of opsonized parasites into the macrophages (66.3 and 28.7% inhibition, respectively). Moreover, when both peptides were used at the same time, a much lower concentration (4 μg/ml) caused greater inhibition (>95%) of the T. cruzi internalization in the macrophages (Fig. 9B). To exclude possible harmful effects of the peptides on the trypomastigotes, the parasites were incubated for 6 h at 37°C with medium alone or with various concentrations of a mixture of peptides 7 and 11 (100, 50, 20, 4, and 0.8 μg/ml). The viability of the parasites was assessed by direct-phase microscopic observation (parasite integrity and motility). At no concentration did the peptides interfere with the trypomastigote numbers, morphology, and motility compared to parasites incubated with medium alone. Moreover, regardless of the presence of inhibitory peptides, inoculation of the trypomastigote-infected macrophages in liver infusion-tryptone medium 48 h after the exposure of the parasites to the phagocytic cells resulted in the growth of epimastigote forms of T. cruzi (data not shown).

FIG. 9.

Inhibition of IgG-mediated opsonization of T. cruzi trypomastigotes by Lmsp1 peptides 7 and 11. T. cruzi trypomastigotes were incubated for 30 min with purified human IgG anti-T. cruzi antibody (2 μg/ml) in the absence or in the presence of 50 μg of peptides 7 and 11/ml (A) or with various concentrations of a mixture of these two peptides (B) followed by incubation with peritoneal exudate murine macrophages at a multiplicity of infection of 1/1. Noninternalized parasites were washed out 2 h later, and cells were incubated for an additional 4 h, fixed with methanol, and stained with Giemsa. Infected cells containing one or more parasites were scored out of 500 cells. The percentage of inhibition was calculated with the following formula: (number of infected macrophages in the presence of IgG − number of infected macrophages in the presence of IgG plus peptides × 100)/number of infected macrophages in presence of IgG. The number of infected macrophages in the absence of IgG was subtracted from the number of infected macrophages in the presence of IgG plus or minus peptides. PEC, peritoneal exudate cells.

Together, these results suggest that an Lmsp1-like molecule of T. cruzi is involved in the IgG-mediated attachment and internalization of these parasites in their macrophage host cells.

DISCUSSION

The presence of normal immunoglobulins on the cell surface of Trypanosomatidae isolated from their mammallian hosts has long been documented (4, 9-12, 18, 19). More recently, experimental evidences indicated that intracellular Trypanosomatidae such as and T. cruzi and Leishmania mexicana ingeniously utilize antibody molecules to facilitate their internalization in the host cells, hence using one of the most abundant host mediators of immunity as an important asset to survive within the host and to perpetuate the infectious process (17, 21, 25). Despite the fact that these observations are undisputable, the presence of a putative immunoglobulin ligand on the parasite's surface has not yet been demonstrated. However, the protein Lmsp1 described herein may represent one such ligand.

The Lmsp1 gene is present in both Leishmania and Trypanosoma genera of the Trypanosomatidae family, and the protein is expressed in all developmental forms of Leishmania (promastigotes and amastigotes) and T. cruzi (epimastigotes, trypomastigotes, and amastigotes). Two lines of evidence support the suggestion that this protein may represent a putative immunoglobulin ligand. First, Lmsp1 strongly binds immunoglobulins in a noncognitive manner. Second, Lmsp1 is associated with the parasite's cell surface.

The demonstration that Lmsp1 is an immunoglobulin ligand was initially suggested because of the ubiquitous reactivity of this molecule with normal immunoglobulins of various animal species (humans, rabbits, sheep, goats, guinea pigs, donkeys, rats, and mice) and with two different isotypes of human immunoglobulins (IgG and IgM). These results pointed to the binding of the recombinant protein with the antibody molecules to noncognitive regions of the immunoglobulins, i.e., to the constant (C) rather than the variable (V) regions of either light or heavy chains of these molecules. Subsequently and more importantly, the experiments described here demonstrated that Lmsp1 interacts with both Fc and Fab fragments of the immunoglobulin molecules. At this point, we still cannot categorically affirm that the interaction of Lmsp1 with the Fab fragment is via noncognitive regions of this molecule because this fragment contains both C and V regions of the antibody molecule. However, the binding of Lmsp1 to Fc unquestionably indicates that Lmsp1 interacts with a noncognitive region of the antibody molecule and therefore can be a putative immunoglobulin ligand present on the parasite's surface. In reality, such types of interactions have been described for a variety of microbial proteins, with the most commonly known being protein A from the S. aureus Cowan strain and protein G from Streptococcus pyogenes (3, 26). Interestingly, competition experiments suggested that Lmsp1 might bind to the same epitopes that are recognized by protein A on the IgG molecule. This suggestion is based on the fact that we observed that Lmsp1 substantially inhibits the binding of peroxidase-labeled protein A to both human and rabbit IgG (data not shown). It is also interesting that both protein A and protein G, although binding preferentially to Fc, and similarly to Lmsp1, also bind the Fab fragment of the immunoglobulin molecule (14, 27).

The evidences pointing to a cell surface localization of Lmsp1 were provided by FACS analyses with purified rabbit IgG specific for the recombinant protein. Because Lmsp1 binds immunoglobulins, the use of a quantitative FACS analysis was necessary to convincingly demonstrate that the anti-Lmsp1 antibody specifically recognized the antigen Lmsp1 on the parasite's cell surface. These experiments clearly showed that the fluorescence intensity of leishmania parasites stained with purified rabbit IgG anti-Lmsp1 antibody was much greater than that observed with the same concentration of normal IgG. In realty, approximately 10 times more normal IgG was required to provide the same levels of fluorescence intensity given by the IgG anti-Lmsp1 antibody. Indeed, the binding of normal immunoglobulin to an immunoglobulin ligand on the cell surface (Fc-like receptor) of members of the family Trypanosomatidae was proposed many years ago (20) and has been confirmed (25, 35). However, the results presented here with purified rabbit IgG anti-Lmsp1 antibody clearly showed that the fluorescence intensity observed with this purified antibody molecule was markedly higher than the fluorescence intensity given by immunoglobulins present in the normal rabbit serum. Therefore, this result clearly points to the association of a defined IgG-binding molecule (Lmsp1) with the parasite's cell surface. Surprisingly though, no obvious signal peptide-encoding sequence is associated with the Lmsp1 gene. Therefore, additional experiments with classical subcellular fractionation and immunoblot assays should help to demonstrate the localization of Lmsp1 in the parasites and to determine whether this protein is actively secreted or is carried to the cell membrane and extracellular milieu complexed with other proteins.

In contrast with the original description of an immunoglobulin ligand (Fc-like receptor) on the cell surface of members of the family Trypanosomatidae (20), the mapping of the binding sites of Lmsp1 on the antibody molecules revealed that both Fc and Fab fragments of these molecules strongly interact with the recombinant protein. However, in the experiments describing the Fc receptor on members of the family Trypanosomatidae, the exclusion of a Fab-binding ligand on the parasite's surface was based on inhibition of rosette formation between sheep red blood cells (SRBC) coated with specific antibody and protozoan cells (Leishmania or T. cruzi). It is possible that the Fab molecules did not inhibit the rosette formation because this fragment of the antibody molecules is by definition interacting with the SRBC membrane and therefore not accessible to interact with a Fab ligand (possible Lmsp1) present on the parasite's surface. In contrast, the Fc fragment of the anti-SRBC antibody is readily exposed and available for the interaction with the corresponding receptor on the parasite's surface and is therefore susceptible to inhibition with soluble Fc. Alternatively, it is possible that Lmsp1 and the proposed Fc-like receptor are distinct molecular entities.

One intriguing aspect of the molecular structure of Lmsp1 is the presence of the tripeptide RGD motif starting at the position 104 of the protein sequence. The RGD motif is one of the key ligand epitopes of fibronectin, a major extracellular adhesive protein found in all vertebrates. Fibronectin has approximately 15 repeated domains that contain RGD sequences, which are ultimately the major epitopes recognized by several members of the integrin family of cell surface proteins (28). For example, the complement receptor 3 (CR3), an important cell surface receptor that facilitates the internalization of microbial organisms into the phagocytic cells is a well-known member of the integrin family. It has been proposed that adherence and subsequent internalization of several intracellular parasites, including Leishmania, is mediated by the RGD motif of fibronectin-like molecules present on the parasite's surface. The interaction of these molecules with members of the integrin family, such as the CR3 on the host cells, would ultimately lead to the internalization of the parasite. And internalization of leishmania parasites by macrophages can be inhibited by anti-RGD monoclonal antibodies (24, 29). It is tempting, therefore, to postulate that that Lmsp1 could be a fibronectin-like candidate molecule present on the surface of leishmania parasites. Lmsp1 has, in addition to RGD, three cysteine residues at positions 55, 74, and 116, which can lead to polymerization of the molecule and generation of a fibronectin-like structure (13). Indeed, Western blot analyses performed under nonreducing conditions showed that both recombinant and native Lmsp1 migrate as several polymers of a single polypeptide chain. This multimeric structure of the RGD containing Lmsp1 could consequently confer to this molecule a fibronectin-like activity on the parasite's surface.

The experiments designed to map the Lmsp1 epitopes that bind to the immunoglobulin molecule revealed that different sequences of the molecule preferentially bind to Fc or Fab. These experiments also determined that the epitope size is smaller than 20 amino acids but longer than 10 amino acids. This conclusion is based on the fact that the peptides used in these experiments were 20-mers, with 10 amino acids overlapping the contiguous peptides. Thus, peptides 7 and 11 strongly bound Fab and Fc, respectively, and their overlapping peptides 8 and 10, respectively, although to a much lesser extent, also bound to these immunoglobulin fragments. However, a more systematic evaluation, with truncated peptides, needs to be performed to finely define the precise binding site of Lmsp1 to the immunoglobulin fragments. This type of approach will also clarify whether the RGD sequence participates or not in the interaction of Lmsp1 with the immunoglobulin molecules. It seems, though, that this motif is not involved in the binding of Lmsp1 to either immunoglobulin fragment because the binding of both Fc and Fab to the RGD-containing peptide 10 was only marginally detected.

The possible involvement of Lmsp1 in the biology of the Trypanosomatidae family parasitism was supported by inhibition experiments with peptides 7 and 11. Despite the fact the Lmsp1 was cloned from L. major, these experiments were carried out with T. cruzi simply because the latter, but not the former, organisms clearly utilize the IgG molecules for their internalization in the macrophages (1, 22, 23). Moreover, an Lmsp1-like molecule is also expressed in T. cruzi (as indicated by Western blot analyses), and more importantly, the sequences of both peptides are highly conserved among these two species of parasites (Fig. 1B). Significant inhibition of the IgG-mediated internalization of T. cruzi in macrophages could be observed with the individual peptides (∼63% for peptide 7 and ∼28% for peptide 11). Moreover, a mixture of these two peptides caused >95% inhibition of T. cruzi internalization in the macrophages. Control peptides (peptides that do not bind Fc or Fab) had no effect on the internalization of the parasite (not shown). These results suggest that these two Lmsp1 peptides can block antibody-mediated parasite invasion of the host cells, probably by binding and inhibiting the activities of the osponizing immunoglobulin molecules. Interestingly, the fact that inhibition of the IgG opsonization was much stronger when both Fc-binding and Fab-binding peptides were used together suggests that this immunoglobulin-binding protein on the parasite's cell surface binds both the specific antibody molecules and the non-T. cruzi-specific antibodies present in the human IgG preparation used in these studies. Unfortunately these studies could not be performed with the whole recombinant molecule because of solubility problems. Thus far, we have not been able to produce soluble Lmsp1 in the absence of strong detergents like SDS, which obviously lends a high level of toxicity to the protein preparation.

Finally, the results presented here point to an important finding, i.e., the presence of a cell surface molecule on members of the family Trypanosomatidae that binds noncognitively to both Fc and Fab fragments of the antibody molecule. Such a ligand may represent a unique evolutionary acquisition of the intracellular Trypanosomatidae pathogens. This ligand can in principle be used by the parasites either as an escape mechanism (e.g., Leishmania organisms) or as an efficient mechanism for them (e.g., T. cruzi) to reach their final intracellular localization in both nonimmune and immune hosts, i.e., by binding the Fab fragment of the antibody molecule, either in a noncognitive or in a cognitive manner, and utilizing the Fc fragment of this molecule for attachment and internalization in the host cells.