trans-Sialidase from Trypanosoma cruzi binds host T-lymphocytes in a lectin manner.

Trypanosoma cruzi, the protozoan parasite responsible for Chagas' disease, expresses on its surface an uncommon membrane-bound sialidase, known as trans-sialidase. trans-Sialidase is the product of a multigene family encoding both active and inactive proteins. We report here that an inactive mutant of trans-sialidase physically interacts with CD4(+) T cells. Using a combination of flow cytometry and immunoprecipitation techniques, we identified the sialomucin CD43 as a counterreceptor for trans-sialidase on CD4(+) T cells. Using biochemical, immunological, and spectroscopic approaches, we demonstrated that the inactive trans-sialidase is a sialic acid-binding protein displaying the same specificity required by active trans-sialidase. Taken together, these results suggest that inactive members of the trans-sialidase family can physically interact with sialic acid-containing molecules on host cells and could play a role in host cell/T. cruzi interaction.

The surface of the protozoan parasite Trypanosoma cruzi, the causative agent of Chagas' disease (American Trypanosomiasis), displays a unique enzyme known as trans-sialidase (TS) 1 (1,2). TS is a modified sialidase (3) sharing the catalytic mechanism (4) and the active site architecture (5) with other known sialidases (6 -8). However, instead of releasing sialic acid, TS preferentially transfers sialic acid from ␤-galactopyranosyl (␤Galp)-containing exogenous donor molecules to terminal ␤Galp-containing acceptors, attaching it in an ␣2-3 linkage configuration (9). T. cruzi TS belongs to a large family of proteins (10,11), and several other members of this family, which lack enzymatic activity (11), are expressed. Comparison of the deduced amino acid sequences shows that enzymatic activity requires a Tyr at position 342, whereas inactive members contain a His at the same position (11). The Tyr 342 residue is involved in the stabilization of the transition-state sialyl carbocation formed during the hydrolysis reaction (4,5). Indirect evidence suggests that an enzymatically inactive recombinant TS acts as a lectin, agglutinating desialylated erythrocytes (12). However, no direct evidence for a ␤Galp binding site or for its role in the parasite host interaction has been established.
Trypomastigote-derived TS is anchored to the membrane through a glycosylphosphatidylinositol anchor and is released to the extracellular medium during acute T. cruzi infection in humans (13), thus acting distant from the parasite. Besides a role in mammalian cell invasion (14), the soluble TS functions as a virulence determinant molecule. Chuenkova and Pereira (15) demonstrated that in vivo injection of minute amounts of purified native TS increases subsequent parasitemia and mortality in T. cruzi-infected mice. As TS injection into severe combined immunodeficiency mice did not affect parasitemia or mortality, it was suggested that TS acts on the host adaptive immune response (15). It is well known that T lymphocytes bearing conventional ␣␤ T cell receptors are required for control of parasitemia and mortality in murine infection by T. cruzi (16). Host CD4 ϩ T cells are also involved in the immunopathology of T. cruzi infection, and their exacerbated function can lead to mortality in susceptible hosts (17). Recently, we demonstrated that both enzymatically active and inactive TS costimulated CD4 ϩ T cell activation in vitro and in vivo and blocked activation-induced cell death in CD4 ϩ T cells from T. cruzi-infected mice through CD43 engagement (18). In the present work, we extended our studies and, using CD43 Ϫ/Ϫ mice, we show that the major lymphocyte mucin CD43 is the counterreceptor for the inactive TS on host CD4 ϩ T cells. We also employed NMR spectroscopy and immunochemical approaches to investigate the nature of the CD43 epitope that functions as ligand for TS, and we demonstrated that inactive TS binds to sialic acid-containing molecules with the same specificity exhibited by active TS.
Recombinant TS-Recombinant active TS (rTS) and inactive TS (irTS), containing the C-terminal repeats, were obtained from Escherichia coli MC1061 electro-transformed with plasmids containing either the wild-type TS insert, TSREP (11), or the inactive mutant TS insert bearing a Tyr 342 3 His substitution, pTrcHisA (11). Bacteria were grown in supplemented Terrific broth in the presence of 100 g/ml ampicillin. When the culture reached an A 600 nm of 1.0, 30 mg/liter isopropyl-1-thio-␤-D-galactopyranoside was added, and incubation continued overnight. Bacteria were lysed at 4°C in 20 mM Tris-HCl containing 2.0 mg/ml lysozyme, 2% Triton X-100, 0.1 M phenylmethylsulfonyl fluoride, 5.0 g/ml leupeptin, 1.0 g/ml trypsin inhibitor, and 0.1 M iodoacetamide. Both rTS and irTS containing a poly-His tag were purified as described by Buschiazzo et al. (20) and modified by Todeschini et al. (4) using Ni 2ϩ -chelating chromatography on a HiTrap column and eluted with an imidazole gradient (0 -1 M). The eluates were dialyzed against 20 mM Tris-HCl, pH 7.6, further purified by ion exchange chromatography on Mono Q and Mono S columns, applying a linear NaCl gradient (0 -1 M), and stored in 20 mM Tris-HCl buffer, pH 7.6, at 4°C until used. The homogeneity of proteins was evaluated on 10% SDS-PAGE. For flow cytometry (FCM) and Western blotting analyses, irTS was biotin-conjugated as described previously (21).
rTS Activity Assay-Enzyme activity was assayed by incubating rTS preparations in 5 mM cacodylate buffer, pH 7.0, in the presence of 0.25 mol of ␣2-3-sialyllactose (␣2-3-SL) and 0.25 mol of [D-glucose-1- 14 C]lactose (400,000 cpm) (4). After incubation for 30 min at 37°C, the reaction mixture was diluted with 1 ml of water and applied to a column containing 1 ml of Dowex 2X8 (acetate form) equilibrated with water. To remove the excess of [D-glucose-1-14 C]lactose, the column was washed with 10 ml of water, and sialylated [D-glucose-1-14 C]lactose was eluted with 3 ml of 0.8 M ammonium acetate and quantitated by scintillation counting (Beckman LS 6500). One unit was defined as the amount of rTS required to catalyze the incorporation of 1 mol of sialic acid into lactose per minute.
Animals and T Lymphocytes-BALB/c mice (male, aging 4 -5 weeks) were obtained from Fundaçã o Oswaldo Cruz, Rio de Janeiro, Brasil; CD43 Ϫ/Ϫ and wild-type control mice (22) were from Universidade Federal de Sã o Paulo, Sã o Paulo, Brasil, animal facilities. All experiments were conducted according to protocols approved by the Committee on Ethics and Regulations of Animal Use of Instituto de Biofísica Carlos Chagas Filho, Universidade Federal do Rio de Janeiro, Brasil.
Primary T cell-enriched suspensions were obtained by nylon-wool filtration of fractionated splenocytes depleted of red cells by treatment with Tris-buffered ammonium chloride (23). Purified CD4 ϩ T cells were nylon-nonadherent cells treated with anti-CD8 mAb for 30 min at 4°C followed by anti-rat Ig chain mAb MAR 18.5 plus 10% rabbit complement for 45 min at 37°C.
Immunoprecipitation and Western Blotting-1 ϫ 10 7 CD4 ϩ T cells were lysed in PBS containing 1 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, 0.1 M iodoacetamide, 1 g/ml trypsin inhibitor, and 0.1% Triton X-100 for 1 h at 4°C under vigorous shaking. After pelleting (10,000 ϫ g for 30 min), the supernatant was precleared overnight at 4°C with 100 l of protein G-Sepharose beads and centrifuged, and the supernatant was incubated with 3 g of anti-CD43 mAb S7 for 2 h at 4°C, under shaking. Protein G-Sepharose (50 l) was added and incubated for 3 h. The beads were separated by centrifugation, washed three times, and incubated with SDS sample buffer for 5 min at 100°C. The supernatant was run on 10% SDS-PAGE and transferred to nitrocellulose membranes. Membranes were blocked overnight with 2% bovine serum albumin in Tris-buffered saline containing 0.2% Tween 20, incubated with anti-CD43 mAb S7 for 2 h followed by incubation with horseradish peroxidase-conjugated anti-mouse IgG. The reaction was detected using enhanced chemiluminescence (ECL) in Hyperfilm-ECL according to the manufacturer's instructions. For irTS immunoprecipitation, streptavidin beads and biotin-conjugated irTS were used.
Desialylation and Resialylation of T Lymphocytes-1 ϫ 10 7 T cells were resuspended in serum-free Dulbecco's modified Eagle's medium and incubated for 30 min at 37°C with 0.05 units of Clostridium perfringens (Type X) sialidase in a total volume of 500 l. Sialidase was removed by washing three times with serum-free Dulbecco's modified Eagle's medium. Desialylated T cells (5 ϫ 10 6 ) were resialylated by incubation with 0.05 units of recombinant rTS in the presence of 1 mM ␣2-3-SL for 30 min at room temperature and washed as described above. Desialylated and resialylated T lymphocytes were resuspended in sorting buffer (10 mM PBS, pH 7.4, containing 2% bovine serum albumin, 0.02% NaN 3 , and 0.1 M lactose) incubated with biotin-conjugated irTS for 30 min at 4°C, washed, stained with PE-conjugated streptavidin, and analyzed by FCM as described below.
FCM Analysis-T cell-enriched suspension was incubated with 10 g/ml Fc block for 5 min at 4°C followed by addition of 10 g/ml fluorescein isothiocyanate-labeled anti-CD43 mAb S7 or 10 g/ml biotin-conjugated irTS for 30 min at 4°C. The T cells were then washed in sorting buffer, incubated for 30 min with PE-conjugated streptavidin at 4°C, washed again, and resuspended in 0.4 ml of sorting buffer plus 2% paraformaldehyde. T lymphocytes were gated by forward scatter and side scatter parameters, and 10,000 cells were analyzed on a fluorescence-activated cell sorter Xcalibur system using Cell Quest software. For irTS inhibition assay, biotin-conjugated irTS was preincubated with ␣2-3or ␣2-6-SL in a range of 0 -1 mM for 30 min at 4°C.
ELISA Analysis of irTS Binding to Sialic Acid-containing Molecules-Analysis of irTS binding to sialic acid-containing molecules was done by ELISA (24). Wells in microtiter plates were coated overnight at 4°C with a monoclonal antibody against the TS repeats (25) (500 ng/well) in 50 mM carbonate/bicarbonate buffer, pH 9.5. The plates were washed with ELISA buffer (3% bovine serum albumin in PBS, pH 7.4) and incubated overnight at 4°C with irTS (500 ng/well) in the same buffer. Plates were blocked with ELISA buffer containing Triton X-100 1% for 1 h at room temperature and subsequently washed. Wells were then incubated with biotin-conjugated PAA probes substituted with sialylated glycans at a final concentration of 5.0 g/well or a range between 0.25 and 10.0 g/well for 2 h at room temperature. Following washing, wells were incubated for 1 h at room temperature with peroxidase-conjugated streptavidin, diluted (1:500) in blocking buffer, and developed with the 2,2Ј-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) diamonium substrate system (100 l/well). The plates were read at 405 nm in an automatic microplate reader (Bio-Tek Instruments).  (24). The products were diluted 5ϫ with ELISA buffer and used in the assay. For sham treatment, 2 mM NaIO 4 and 10 mM NaBH 4 were mixed for 1 h at 4°C and diluted with ELISA buffer, and sialylated PAA probes were added to this mixture just before use in the assay.
Carboxyl Reduction of Sialylated PAA Probes by Iodoethano and Sodium Borohydride Treatment-Sialylated PAA probes were carboxylreduced with sodium borohydride treatment after esterification with iodoethane (24). Briefly, 100 g of lyophilized sialylated PAA probes was solubilized in 350 l of dimethyl sulfoxide, esterified with 35 l of CH 3 CH 2 I for 1 h at room temperature, and then reduced by addition of 1.115 ml of PBS buffer containing 10 mM NaBH 4 . The reaction products were diluted 4ϫ with ELISA buffer and directly used in the assay. For sham treatment, the same procedures were performed without adding CH 3 CH 2 I.
Analysis of Relative Biotinylation Level of Sialylated PAA Probes-Microtiter plates were coated with biotinylated PAA probes (200 ng/ well) overnight at 4°C in 50 mM carbonate/bicarbonate buffer, pH 9.5. After blocking with ELISA buffer and subsequent washing, the wells were incubated for 1 h at room temperature with peroxidase-conjugated streptavidin diluted 1:500 times and developed with 100 l/well 2,2Јazino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diamonium substrate system. The plates were read at 405 nm in an automated microplate reader.
NMR Experiments-␣2-3-SL, ␣2-6-SL, SLeX, or methyl ␣-mannoside was dissolved in deuterated PBS, pH 7.6 (not corrected for isotope effects). irTS solution in 20 mM Tris-HCl was exchanged with deuterated PBS by gel filtration on a G25 column. 20 l of a stock solution containing 10 mg/ml of irTS was added to a solution of sialyl glycoside (2 mM final concentration), and the total volume was adjusted to 500 l. NMR spectra were obtained at a probe temperature of 20°C on a Bruker DMX 600 equipped with a 5-mm self-shielded gradient triple resonance probe or on a Bruker DRX 600 with a 5-mm triple resonance probe.
One-dimensional Saturation Transfer Difference (One-dimensional STD)-One-dimensional STD experiments were performed by low power presaturation of the methyl region of the protein during the 2-s relaxation delay. The pulse scheme was as follows: relaxation delay with or without presaturation of the protein resonances, 90 degrees pulse, and acquisition of 256 scans (16,000 points for 10 ppm of spectral width). The data were obtained with interleaved acquisition of onresonanse and control spectra to minimize the effects of temperature and magnet instability.
Saturation Transfer Difference-Total Correlation Spectroscopy (STD-TOCSY) spectra were recorded with a mixing time of 66 ms, 32 scans per t 1 increment. 200 t 1 increments were collected in an interlaced mode with presaturation on or off for 2 s. Prior to subtraction, both spectra were processed and phased identically. The acquisition time for the two-dimensional experiments were typically 16 h. The spectra were multiplied with a square cosine bell function in both dimensions and zero-filled two times. All spectra were referenced relative to trimethylsilyl-2,2Ј,3,3Ј-d 4 -propionic acid-sodium salt (␦ ϭ 0.0 ppm).

RESULTS
irTS Binds CD43 on T Cells-We recently demonstrated that irTS from T. cruzi binds CD4 ϩ T cells and that this binding is abrogated by prior treatment with anti-CD43 S7 mAb (18). To prove that CD43 is the counterreceptor for TS on CD4 ϩ T cells, we investigated the interaction between irTS and the leukosialin (CD43), using splenic CD4 ϩ T cells from CD43-deficient mice (CD43 Ϫ/Ϫ ). As compared with wild type, CD4 ϩ T cells from CD43 Ϫ/Ϫ mice failed to bind either anti-CD43 mAb S7 (Fig. 1A) or biotin-conjugated irTS (Fig. 1B). To confirm that CD43 is the counterreceptor for irTS, CD4 ϩ T cell extracts from wild-type and CD43 Ϫ/Ϫ mice were immunoprecipitated with biotin-conjugated irTS or with anti-CD43 mAb S7. Precipitates were immunoblotted and revealed with anti-CD43 mAb S7 (Fig. 1C). Remarkably, both irTS and anti-CD43 mAb S7 immunoprecipitated the same 115-kDa protein band expected for CD43 (19). However, this protein was absent when CD4 ϩ T cells from CD43 Ϫ/Ϫ mice were submitted to the same treatment (Fig. 1C). These results indicate that CD43 expression is required for irTS binding on T cells, showing that soluble irTS physically interacts with CD43.
␣2-3-linked Sialic Acid Is the Epitope for irTS on T Cells-CD43 is the most abundant glycoprotein bearing ␣2-3and ␣2-6-linked sialic acids expressed on the surface of T cells (26). To investigate whether sialic acid is the epitope for irTS, T cells were treated with C. perfringens sialidase, and the binding of biotinylated irTS was tested. Fig. 2A (a) shows that irTS binds to T cells, and this binding is abrogated by sialidase treatment ( Fig. 2A (b)). irTS binding was reconstituted with high intensity by resialylation of T cells with rTS in the presence of ␣2-3-SL as donor substrate ( Fig. 2A (c)). Since TS catalyzes the transfer of sialic acid residues from NeuAc␣2-3Galp␤1-x-containing donors and attaches them in an ␣2-3 linkage to terminal ␤Galp-containing molecules (9), our results clearly show that irTS binds ␣2-3-linked sialic acid. The specificity of the interaction of irTS with ␣2-3-linked sialic acid was confirmed by inhibition studies using ␣2-3-( Fig. 2A (d)) or ␣2-6-SL ( Fig.  2A (e)). As shown in Fig. 2A (d), binding of irTS to T cells was abrogated by the previous incubation of irTS with ␣2-3-SL but not with ␣2-6-SL ( Fig. 2A (e)). These results are consistent with the irTS bright pattern of staining observed after resialylation of T cells by rTS ( Fig. 2A (c)) since these cells now must bear almost exclusively ␣2-3-linked sialic acid. In addition, previous treatment of irTS with ␣2-3-SL completely disrupt irTS binding to resialylated T cells ( Fig. 2A (f)). This inhibition was dose-dependent and had an IC 50 of 0.49 mM (Fig. 2B).
Binding Preferences of irTS for Sialic Acid-containing Molecules-The irTS preference for ␣2-3-linked sialic acid was further investigated using sialylated PAA probes. Fig. 3 shows that irTS strongly binds to ␣2-3-linked sialic acid. Binding was dependent on the concentration of ␣2-3-SL-PAA, being maximal at 10 g/ml, and relies on the carboxylate group of sialic acid since irTS binding is abrogated after carboxyl reduction (Fig. 3). Furthermore, when ␣2-6-SL-or SLeX-PAA probes were used as ligands, low level binding or no binding to irTS was observed, respectively (Fig. 4).
Periodate treatment of sialylated PAA probes did not reduce irTS binding (Fig. 4), demonstrating that the sialic acid side chain is not required for irTS recognition. Taken together, these data suggest that irTS displays a binding site that recognizes ␣2-3-linked sialic acid and its 7-carbon analog and that this binding can be abolished by either fucosylation or carboxyl reduction. Interactions of irTS with all sialylated PAA probes were inhibited in the presence of a polyclonal antibody directed against epitopes present in the catalytic domain of the active TS ( Fig. 4) (25).
NMR Study of irTS Binding to Sialic Acid-containing Molecules-To better understand the binding specificity observed for irTS, NMR spectroscopy studies were employed. Saturation transfer difference (STD) experiments were used to verify soluble ␣2-3-SL interaction with irTS. The binding assay was done in the presence of the methyl ␣-mannoside, used as negative control. Fig. 5 shows the one-dimensional STD experiment of irTS in the presence of ␣2-3-SL and methyl ␣-mannoside in D 2 O PBS, pH 7.6, at 20°C. Binding of ␣2-3-SL can be followed by measuring the signal at 2.050 ppm assigned to the 5NAc protons of N-acetylneuraminic acid (NeuAc) (27) and by the NeuAc H3 equatorial (eq) and axial (ax) resonances at 2.727 and 1.812 ppm, respectively. Binding of methy ␣-mannoside would be detected by observation of the H1 and methyl protons at 4.775 and 3.405 ppm, respectively, in the STD spectrum. The subtraction of the spectrum in which protein resonance was presaturated (0.5 ppm) (Fig. 5B) from the reference spectrum without protein saturation (Fig. 5A) reveals the resonances at 2.727, 2.050, and 1.812 ppm (Fig. 5C) of NeuAc from ␣2-3-SL ligand involved in the binding to the protein. No signals arising from methyl ␣-mannoside were observed, showing that ␣2-3-SL binds to irTS in a specific interaction.
irTS interaction with ␣2-3-SL was further verified by twodimensional NMR experiments, which show in detail the key structural elements of ␣2-3-SL involved in binding to irTS. A reference TOCSY spectrum of irTS in the presence of ␣2-3-SL and methyl ␣-mannoside was recorded without protein pre-saturation and with protein presaturation (0.5 ppm) (Fig. 6A). In the STD-TOCSY obtained (Fig. 6B), the on-diagonal crosspeaks identify hydrogens in close proximity with the protein in the complex. It is evident that ␣2-3-SL binds to irTS. As the relative signal intensities on the STD spectrum correlate with the proximity to the protein, we can conclude that NAc and the H3ax protons from the NeuAc are in close contact with irTS. In the STD-TOCSY spectrum, off-diagonal cross-peaks help identify key hydrogens involved in binding. It is clear that the H3eq, H4, and H5 from the NeuAc residue and the H1, H3 and H4 from the ␤Galp ring also interact with the irTS as they receive saturation from the protein (Fig. 6B). The H1 and H2 peaks from the ␣-anomer of Glcp are low in intensity, indicating a loose contact with the irTS binding site. Cross-peaks arising from ␣-Glcp or from methyl ␣-mannoside were not detected (Fig. 6B). Binding of irTS to ␣2-6-linked sialic acid was also investigated using STD. The STD-TOCSY spectrum obtained from ␣2-6-SL in the presence of irTS (Fig. 7) shows cross-peaks in which only NAc (2.039 ppm), H3ax (1.755 ppm), and H3eq (2.721 ppm) from NeuAc residue receive saturation from the protein. No signals emerging from ␤Galp or ␣Glcp are visible. These results indicate that irTS binds only to the sialic acid ring of ␣2-6-SL, consistent with an improper positioning of the entire molecule in the irTS binding site. Furthermore, in agreement with the experiments using the SLeX-PAA probe (Fig. 4), STD-NMR does not show any binding between irTS and soluble SLeX (data not shown). T. cruzi presents in its genome hundreds of genes encoding a family of TS molecules and sialic acid acceptor glycoproteins. Combined, the molecules expressed by both sets of genes are likely to cover most of the parasite surface (13), warranting a parasite-host interface. Around 140 genes encode for the TS family, comprising enzymatically active as well as inactive members (10,28,29). Although no function has been assigned to the inactive TS, genes coding for inactive members are present in a similar number in the parasite genome as in those encoding active TS (12). In the present work, we show for the first time that irTS can interact with host T cells. Using wildtype and CD43 Ϫ/Ϫ mice and a combination of FCM and immu-noprecipitation techniques, we identified the sialomucin CD43 as the counterreceptor for irTS on the T cell surface. Furthermore, using biochemical, immunological, and spectroscopic approaches, we directly demonstrate that irTS binds to ␣2-3linked sialic acid.
CD43, or leukosialin, is an abundant mucin expressed on T lymphocytes and other bone marrow-derived cells with ubiquitous physiological roles in cell-to-cell interactions. It is highly glycosylated with 80 -90 O-linked glycosylation sites (26,30). High surface density, pronounced length of protruding molecules, and abundance of sialic acid residues (26) make CD43 a candidate receptor for T. cruzi TS on CD4 ϩ T cells. Natural ligands for CD43 include ICAM-1 (31), galectin-1 (32), class I major histocompatibility complex (33), E-selectin (34), and a macrophage sialoadhesin (35). Our studies indicate that TS is a parasite ligand for CD43 on CD4 ϩ T cells. Furthermore, soluble active and inactive TS costimulate CD4 ϩ T cell activation through CD43 engagement, being a candidate molecule for induction of immunopathology in T. cruzi infection (18). CD43 is a known receptor for other pathogen-associated molecules. CD43 is a neutrophil receptor for influenza A hemagglutinin (36), responsible in part for neutrophil deactivation observed during infection (37). Macrophage invasion by Mycobacterium also requires the extracellular domain of CD43 (38). Since the immunopathology of T. cruzi infection involves multiple host cell types (16), binding of irTS to other myeloid cells expressing CD43, such as macrophages, neutrophils, and dendritric cells, should be investigated.
CD43 expressed on resting T lymphocytes is highly sialylated and carries the tetrasaccharide core NeuAc␣2-3Gal␤1-3(NeuAc)GalNAc␣1-Ser/Thr (39). We therefore, tested the hypothesis that sialic acid is the epitope for irTS binding on CD43. Our results demonstrated that ␣2-3-linked sialic acid is the epitope for irTS binding on CD43 and that irTS recognizes its ligand with similar specificity to that described for active TS (9). Thus, irTS showed a preferential binding for ␣2-3-linked sialic acid and for its C7 derivative obtained by truncation of the sialic acid side chain after mild periodate oxidation (1,9). Another characteristic shared by both proteins is that active and inactive TS are unable to recognize ␣2-3-sialyllactosamine bearing a fucose residue, forming the SLeX epitope (9). We converted the carboxyl group of sialic acid into an alcohol on the ␣2-3-SL-PAA probe to explore its role on irTS binding. Our results show that the sialic acid carboxyl group is essential for irTS binding. It is known that the catalytic domain of TS contains three arginine residues that bind the carboxyl group of sialic acid (3,5). Therefore, the sialic acid binding site found in irTS might be the same present in active TS. Consistent with these findings, an antibody directed against an epitope present in the catalytic domain of active TS (25) was able to inhibit irTS binding to sialic acid-containing PAA probes.
NMR spectroscopy was used to investigate a physical interaction between irTS and sialic acid-containing molecules. NMR spectroscopy has been a valuable tool in studying interactions between ligands and receptors and provides detailed information about binding site and binding conformations in a noninvasive manner (40). The STD NMR technique has been used to distinguish those molecules that bind to a macromolecule from nonbinding compounds (41)(42)(43)(44). This technique relies on selective saturation of resonances arising from the receptor protein that spin-diffuses efficiently over the entire protein and consequently is transferred to the bound ligand. This disturbance on spin states in the population is carried by the ligand when released and detected as a reduction in the intensity of resonances in the free ligand. The relative intensity of the signals present in the STD spectrum identifies those nuclei closest to the protein in the binding site, permitting a detailed mapping of the interactions involved (42,44,45). Using one-dimensional and two-dimensional STD experiments, we demonstrated that ␣2-3-SL ligand binds to irTS. From our results, it is possible to map the structural regions of ␣2-3-SL most involved in binding to irTS. The 5NAc, H3eq, H3ax, H4, and H5 of the NeuAc residue and the H1, H3, and H4 from ␤Galp are in contact with irTS, whereas the ␣Glcp residue is not involved in the binding to the protein. No interaction between the glycerol side chain and irTS was found, reinforcing our finding that mild periodate treatment does not impair irTS recognition. Furthermore, only NAc, H3ax, and H3eq from the NeuAc residue of ␣2-6-SL receive saturation from irTS, indicating that the loose binding of ␣2-6-SL to irTS is due to an incorrect positioning of the entire molecule in the TS binding pocket. Recently, the involvement of the Trp 312 in the T. cruzi TS activity was demonstrated (46). The single mutation Trp 312 3 Ala rendered the TS capable of hydrolyzing both ␣2-3and ␣2-6-linked sialylmolecules. Here we verified that the ␣2-6-SL does not bind correctly to the irTS pocket, suggesting that Trp 312 prevents ␣2-6-linked sialic acid binding and that a proper positioning of the sialoside is necessary in active TS for the transfer reaction to take place. This is the first time that binding of sialic acid-containing molecules by a catalytically inactive member of the T. cruzi TS family has been unambiguously proven. Our results show that irTS conserves its binding site for sialic acid despite its lack of enzymatic activity. It has been suggested that the Tyr 342 3 His mutation abrogates the catalytic activity due to the role of Tyr 342 in stabilizing the carbonium ion transition state intermediate (4,5). We now demonstrate that this site mutation does not abolish binding to sialic acid donor molecules and confirm the prediction that loss of the enzymatic activity might generate lectin-like molecules from TS (12). Molecules having mutually exclusive glycosidase or lectin activities have been described in mammalian cells, in a family of chitinases. Although some of their members are glycosylhydrolases, others are chitin-specific lectins that lack enzymatic activity (47).
Another example is a disulfide-bonded dimer of the Golgigalactoside ␣2-6-sialyltransferase that is catalytically inactive and retains the ability to bind galactose (48). In addition, site-directed mutagenesis of Newcastle disease virus hemagglutinin-neuraminidase found functional and topological relationships between the neuraminidase and receptor binding activities of hemagglutinin (49).
In summary, we demonstrated that CD43 is the counterreceptor for irTS on CD4 ϩ T cells. Using different approaches, we directly demonstrated that irTS is a sialic acid-binding protein.
Taken together, our results suggest that inactive molecules of the trans-sialidase family still retain sialic acid binding activity and could behave as lectins that bind and transmit signals to cells of the host immune system during T. cruzi infection.