Two Trans-sialidase Forms with Different Sialic Acid Transfer and Sialidase Activities from Trypanosoma congolense*

Trypanosomes express an enzyme called trans-sialidase (TS), which enables the parasites to transfer sialic acids from the environment onto trypanosomal surface molecules. Here we describe the purification and characterization of two TS forms from the African trypanosome Trypanosoma congolense. The purification of the two TS forms using a combination of anion exchange chromatography, isoelectric focusing, gel filtration, and subsequently, antibody affinity chromatography resulted, in both cases, in the isolation of a 90-kDa monomer on SDS-PAGE, which was identified as trans-sialidase using micro-sequencing. Monoclonal antibody 7/23, which bound and partially inhibited TS activity, was found in both cases to bind to a 90-kDa protein. Both TS forms possessed sialidase and transfer activity, but markedly differed in their activity ratios. The TS form with a high transfer-to-sialidase activity ratio, referred to as TS-form 1, possessed a pI of pH 4–5 and a molecular mass of 350–600 kDa. In contrast, the form with a low transfer-to-sialidase activity ratio, referred to as TS-form 2, exhibited a pI of pH 5–6.5 and a molecular mass of 130–180 kDa. Both TS forms were not significantly inhibited by known sialidase inhibitors and revealed no significant differences in donor and acceptor substrate specificities; however, TS-form 1 utilized various acceptor substrates with a higher catalytic efficiency. Interestingly, glutamic acid-alanine-rich protein, the surface glycoprotein, was co-purified with TS-form 1 suggesting an association between both proteins.

TS was first described in the bloodstream form of the American trypanosome Trypanosoma cruzi (4), the pathogen of Chagas disease, afflicting millions of people in Latin America. TS has also been reported to occur in the procyclic insect forms of the African trypanosomes Trypanosoma brucei gambiense and Trypanosoma brucei rhodesiense (5), which are the cause of human sleeping sickness. Furthermore, TS has been found in procyclic forms of other African trypanosomes, such as Trypanosoma brucei brucei (6,7) and Trypanosoma congolense (5). These parasites are the agents of Nagana, the trypanosomiasis in African ruminants.
Trypanosomes are unable to synthesize sialic acids; instead they utilize TS to transfer sialic acid from the environment onto trypanosomal surface molecules. In the case of T. cruzi, TS is employed to acquire sialic acid from mammalian host glycoconjugates to sialylate mucin-like acceptor molecules in the parasite plasma membrane (8). Furthermore, TS sialylates host cell glycoconjugates to generate receptors, which are used for parasite adherence and subsequent entry into host cells (2). In the African species T. brucei brucei and T. congolense, where TS is only expressed in the procyclic insect stage, the enzyme is used to sialylate the major cell surface glycoprotein of the parasites (e.g. procyclic acidic repetitive protein and GARP) in the vector (tsetse fly). Thus, a negatively charged glycocalyx is formed, which is believed to protect the parasites from digestive conditions in the fly gut, or from immunocompetent substances present in the fly's blood meal and may enable them to interact with epithelial cells (6,9). Additionally, it has recently been reported that T. cruzi TS itself directs neuronal differentiation in PC12 cells (10), stimulates interleukin-6 secretion from normal human endothelial cells (11), and potentiates T cell activation through antigen-presenting cells (12).
Investigating TS has become increasingly attractive over the last years, not only because of its involvement in trypanosomal pathogenicity but also because of its biotechnological importance. That is, TS is a unique enzyme that, because of its ability to transfer Neu5Ac in a stereo-and regio-specific manner, can be utilized to synthesize a variety of biologically relevant structures of the type Neu5Ac␣2,3Gal␤1-R (13,14).
To date, only two trypanosomal TS have been studied in detail, the American T. cruzi TS (15)(16)(17) and the African T. brucei brucei TS (6,7,18), with different genes encoding T. cruzi TS (19 -21) and T. brucei brucei TS (22) being identified and analyzed. Furthermore, the SA expressed by Trypanosoma rangeli, a non-pathogenic relative of T. cruzi, has been isolated and characterized biochemically (23) and genetically (24). Although the crystal structure of T. cruzi trans-sialidase has recently been published (25), a number of questions concerning the exact TS transfer mechanism remain un-answered. Because native and recombinant enzymes can differ in their glycosylation, antibody specificity, and biochemical properties, it is important that the native enzyme be purified and characterized, with the subsequent aim of obtaining sequence information. This is especially important, because several genes encoding TS/SA enzymes, or even silent genes may exist in trypanosomes, as has been shown for T. cruzi (26). Here we describe the purification and characterization of two TS forms from the African trypanosome T. congolense and their identification using micro-sequencing. Moreover, we report on the production of monoclonal antibodies raised against both enzyme forms and their subsequent use in purification. Additionally, we present characterization studies that reveal significant differences between both TS forms concerning their transfer to SA ratios and catalytic efficiencies using various acceptor substrates.
Cultivation-Procyclic culture forms of T. congolense (STIB 249, kindly provided by Dr. Retro Brun from the Swiss Tropical Institute, Basel, Switzerland) were cultivated axenically in SM/SDM 79 medium (28), containing 10% fetal calf serum (FCS, PAA Laboratories, Austria) and 0.001% hemin. After 3-4 days of cultivation, the trypanosomes were transferred into new SM/SDM 79 medium without FCS and he-min. Following a further 3 days, the culture supernatant was harvested via centrifugation.
Assays-For all enzyme assays the formation of product was linear with respect to time and protein amount. In all activity tests, controls were performed in the absence of enzyme sample or using heat-inactivated enzyme. For fluorescence detection a 96-well-plate fluorometer (Fluorolite 1000, Dynatech Laboratories) was used.
SA activity was routinely tested in the presence of 1 mM MUNeu5Ac in 20 mM Bis/Tris buffer, pH 7.0 (29). The reaction mixture was incubated for 120 min at 37°C in black 96-well-plates (Microfluor, Dynex). By the addition of 0.08 M glycine/NaOH buffer, pH 10, the reaction was terminated, and the fluorescence of MU released measured immediately at an excitation and emission wavelength of 365 and 450 nm, respectively. The instrument was calibrated with MU standard solutions. One unit of SA activity equals 1 mol of MU released per minute, which is equivalent to 1 mol of sialic acid released per minute.
TS activity was routinely tested using the non-radioactive assay described by Schrader et al. 2 Briefly, TS activity was monitored by incubating 25 l of enzyme solution in 50 mM Bis/Tris buffer, pH 7.0, containing 1 mM ␣2,3-SL as the donor and 0.5 mM MUGal as the acceptor in a final volume of 50 l at 37°C for 2 h. The reaction was terminated by the addition of ice-cold water and, subsequently, applied to mini-columns containing Q-Sepharose FF. After washing, the sialylated product was eluted with 1 M HCl, hydrolyzed at 95°C for 45 min, and cooled on ice. The samples were neutralized, adjusted to pH 10, and MU released was measured as stated above. One unit of TS activity equals 1 mol of MU released per minute, which is equivalent to 1 mol of sialic acid transferred per minute. During the course of this study the TS test described above was modified by applying the assay principle to a 96-well plate format. Because of its enhanced throughput, all TS tests for mAb screening, as well as kinetic experiments were performed using the 96-well-plate assay (Schrader et al.). 2 Protein concentration was determined using either the BCA protein assay kit from Pierce (Cologne, Germany) or the Bio-Rad protein assay (30) from Bio-Rad (Munich, Germany), as described by each manufacturer. All assays were performed in 96-well plates employing bovine serum albumin as the standard, and photometric determinations were performed using a 96-well plate photometer (Tecan Sunrise, Tecan Deutschland GmbH). Total amounts of bound sialic acid were measured by the micro-adaption of the orcinol/Fe 3ϩ /HCl reaction (31).
Separation and Purification of the Two TS Forms-The crude culture supernatant was filtered (1.2-m membrane, Millipore GmbH, Schwalbach, Germany) and concentrated in an Amicon ultrafiltration device (molecular mass cut off 20 kDa, Sartorius, Göttingen, Germany) prior to undergoing chromatography. Following all purification steps, fractions were concentrated with the aid of the following devices depending on the volume: Centrex UF-2 (molecular mass cut off 30 kDa, Schleicher & Schuell, Dassel, Germany), Centricon Plus-20 (molecular mass cut off 30 kDa, Millipore, Eschborn, Germany), or an Amicon ultrafiltration device (molecular mass cut off 20 kDa). Unless otherwise stated all purification experiments were performed at 4°C.
The separation of two major TS activity peaks was provided by chromatography on Q-Sepharose FF. The concentrated culture supernatant was applied to a column (2 ϫ 20 cm) of Q-Sepharose FF, equilibrated with 20 mM Bis/Tris buffer, pH 7.0, at a flow rate of 0.6 ml/min. Following extensive washing bound TS activities were eluted using a 600-ml continuous NaCl gradient (0 -0.8 M) in 20 mM Bis/Tris buffer, pH 7.0. Fractions of 6 ml were collected and analyzed for transfer and SA activity. A larger Q-Sepharose column could not be employed due to poor separation of the two TS forms. Therefore, several Q-Sepharose runs were performed using the column size stated above, with separated TS-form 1 and TS-form 2 following each run being combined and further purified individually by isoelectric focusing (IEF).
Isoelectric focusing was carried out in a 16-ml Rotor cell (Rotofor Preparative Isoelectric Focusing Cell, Bio-Rad) using ampholytes that provided a pH range between pH 4 and 6 (Biolyte pH 4 -6, Bio-Rad). The buffer contained in the collected fractions was immediately exchanged, and fractions were concentrated and activity determined. Active fractions were pooled and further purified by gel filtration chromatography.
Each individual TS form was applied to a column (1 ϫ 90 cm) of Sephadex-G150 SF equilibrated with 20 mM Bis/Tris buffer, pH 7.0, containing 100 mM NaCl at a flow rate of 0.125 ml/min, which had been calibrated using the high molecular weight calibration kit (Amersham Biosciences, Freiburg, Germany) as described by the manufacturer. Fractions of 500 l were collected and analyzed for activity. Active fractions were pooled, concentrated, and analyzed by SDS-PAGE.
T. congolense TS Antibody Production, Detection, Isolation, and Isotyping-BALB/c (H-2 d ) mice, obtained from Harlan/Winkelmann (Borchen, Germany) and reared under conventional conditions, were used for the production of monoclonal antibodies (mAb). Female BALB/c mice 6 weeks of age were injected three times intraperitoneally with 25 g of the partially purified TS forms adsorbed to 2 mg of Al(OH) 3 (Imject Alum, Pierce, Rockford, IL). Three days after the last injection spleen cells of one mouse was fused with non-secretor Ag8.653 myeloma cells (32) by the conventional polyethylene glycol-mediated fusion technique (33). After fusion, cells were plated in 288 wells of 24-well hybridoma plates (Greiner, Nü rtingen, Germany) in RPMI 1640 supplemented with 10% FCS, 100 units/ml penicillin, 100 g/ml streptomycin, 100 M hypoxanthine, 0.4 M aminopterin, and 16 M thymidine. The medium was further supplemented with 10% conditioned medium from the J774 cell line. Wells containing antigen-specific IgGsecreting hybridomas were identified via ELISA using mouse-IgG-specific antiserum and an enzyme immunoassay (TS activity binding assay). Clones in positive wells were subcloned and reanalyzed.
The TS activity-binding assay was performed using Dynabeads M-450 goat anti-mouse IgG (Dynal, Hamburg, Germany) and a magnetic particle concentrator for microcentrifuge tubes (Dynal MPC-S, Dynal, Hamburg, Germany). Briefly, 200 l of beads were washed twice with phosphate-buffered saline (PBS) as described by the manufacturer. Following incubation with putative anti-T. congolense TS mAb at room temperature for 1 h the beads were washed again five times with 900 l of PBS buffer and further incubated with 200 l of TS-containing solution at 4°C for 1 h. The incubation was terminated by transferring the supernatant to a new cap, and the beads were subsequently washed five times with 900 l of PBS buffer. In the supernatant, as well as on the beads, TS activity was determined and compared with a control provided by binding non-TS-specific IgG2b antibodies to the Dynabeads. The reduction of TS activity in the TS-containing solution in comparison to the control, as well as the activity detected on the beads, enabled the determination of a clone producing anti-T. congolense TSspecific mAb.
Purification of the anti-T. congolense TS mAb from hybridoma supernatant was performed by affinity chromatography using rProtein A-Sepharose FF (Amersham Biosciences, Freiburg, Germany) according to the manufacturer. The antibody concentration of the eluted preparation was determined with an enzyme immunoassay for the quantitative determination of mouse IgG (Roche Diagnostics GmbH, Mannheim, Germany). Immunoglobulin subclass determination was performed with the Hybridoma Subtyping kit (Roche Diagnostics GmbH).
Immunoaffinity Chromatography-Purified anti-T. congolense TS mAb (7/23, 24 mg) were incubated for 2 h with rProtein A-Sepharose FF (5 ml) and equilibrated with binding buffer (20 mM Na 2 HPO 4 , NaH 2 PO 4 , pH 7.0). Following washing with 70 ml of binding buffer, the matrix was further washed with cross-linker buffer (0.2 M triethylamine, pH 8.5) and, subsequently, incubated with 10 ml of cross-linker reagent (20 mM dimethyl pimelimidate in cross-linker buffer) at room temperature for 1 h. The cross-linking reaction was terminated by washing the column with 70 ml of 0.2 M ethanolamine, pH 9, followed by 70 ml of binding buffer and 70 ml of sodium citrate buffer, pH 3.0. A flow rate of 1-1.25 ml/min was used at all stages of matrix preparation. The column was washed with binding buffer prior to immunoaffinity chromatography.
The immunoaffinity matrix equilibrated in binding buffer was incubated with the partially purified TS forms overnight at 4°C. Unbound protein was removed by washing with binding buffer. TS activity was eluted stepwise with 70 ml of 100 mM Bis/Tris, pH 7.0; 100 mM Bis/Tris, pH 7.0, containing 1 M NaCl; 20 mM sodium citrate, pH 4.5; and 20 mM sodium citrate, pH 3.0, at a flow rate of 1-1.25 ml/min. The last fraction was immediately neutralized prior to TS activity determination.
Micro-sequencing-In-gel trypsin digestion and mass spectrometric analysis of peptides were performed by WITA GmbH (Berlin, Germany).
Kinetic Studies-Kinetic data for both enzyme forms were obtained by making suitable modifications to the standard SA and TS assay. To measure transfer activity various concentrations of the acceptor substrates MUGal and MULac (0 -2 mM) were used at a constant donor substrate (␣2,3-SL) concentration of 1 mM. The kinetic parameters for ␣2,3-SL were obtained by varying ␣2,3-SL concentrations (0 -3 mM) at a constant concentration of MUGal (0.5 mM). Various concentrations of MUNeu5Ac (0 -0.2 mM) were used to obtain kinetic data for SA activity.
Apparent V max and apparent K m values were determined by non-linear regression using the computer program Enzfitter from Elsevier Biosoft. Additionally, the temperature and pH optima of both forms were investigated in the range of 5-55°C and pH 4.5-10.5, respectively.
Donor and Acceptor Substrate Specificities and Inhibitor Studies-A number of glycoconjugates, as well as mono-and oligosaccharides (Table IV) were assayed as potential donors using the TS assay described. Known viral and bacterial SA inhibitors, as well as salts (NaCl and KCl), cations (20 mM Ca 2ϩ , Mg 2ϩ , Mn 2ϩ ; 5 mM Cu 2ϩ , Zn 2ϩ , Fe 2ϩ , Co 2ϩ ), and other compounds, including anti-T. congolense TS mAb (0 -20 g/ ml), were assayed for their ability to inhibit TS activity using essentially the standard TS assay described, except additives were preincubated in the assay mixture for 30 min at room temperature prior to starting the reaction. Potential TS acceptors (Table IV) were assayed in a similar manner to that described for potential inhibitors.
SDS-PAGE and Immunoblot Analyses-SDS-PAGE was performed according to Laemmli (34) in a Mini Protean II Cell (Bio-Rad, Munich, Germany) in the presence of a reducing agent (dithiothreitol). Polyacrylamide gels usually consisted of 8% resolving and 4% stacking gel, with the exception of gels used for immunoblot analyses using anti-T. congolense GARP mAb, where the resolving gel was 12%. As molecular weight markers, pre-stained SDS-PAGE standards from Bio-Rad (for immunoblotting) or SDS-PAGE Marker High Range from Sigma were applied (for staining). Gels were subsequently stained with either silver (35) or Coomassie Brilliant Blue R-250 (36). For immunoblot analyses, after SDS-PAGE, proteins were transferred onto a nitrocellulose membrane (Schleicher and Schuell, Dassel, Germany) using a Mini-V 8 -10 Blot Module (Life Technologies, Eggenstein, Germany) as described by the manufacturer.
For immunodetection, blots were blocked overnight at 4°C in TBS (Tris-buffered saline) buffer containing 0.05% Tween 20 (TBST) and 5% skim milk (blocking buffer), washed six times with TBST for 5 min, and then incubated for either 24 h at 4°C (antiserum to T. cruzi TS, dilution 1:5000) or 1 h at room temperature (anti-T. congolense TS mAb, dilution 1:3000 or anti-T. congolense procyclin (GARP) mAb, dilution 1:1000) in blocking buffer solution containing the appropriate primary antibody. Following incubation, the blots were washed again six times with TBST buffer and then incubated for 1 h at room temperature with horseradish peroxidase-conjugated anti-rabbit or anti-mouse IgG (1:10000). After washing six times for 10 min with TBST buffer, bands were visualized using the ECL immunoblotting detection reagent kit (Amersham Biosciences, Braunschweig, Germany) as described by the manufacturer.
Immunoprecipitation-Immunoprecipitation was performed in a similar manner to that described for the TS activity-binding assay. Briefly, 30 l of Dynabeads M-450 goat anti-mouse IgG were washed twice with PBS buffer and incubated with either 30 l of PBS buffer (control) or 30 l of anti-T. congolense TS mAb (mAb7/23) at room temperature for 1 h. Following incubation, the beads were washed again five times with 900 l of PBS buffer and further incubated at 4°C for 1 h with 200 l of TS containing concentrated culture supernatant. The beads were washed five times with 900 l of PBS buffer, subsequently boiled in SDS-PAGE sample buffer (containing SDS and dithiothreitol) at 95°C for 5 min, analyzed by SDS-PAGE, and immunoblotting as described above. The detection of T. congolense TS was achieved as described above. The detection of GARP was performed using anti-T. congolense GARP mAb as the primary (dilution 1:1000) and biotinylated anti-mouse IgG3 mAb (dilution 1:4000) as the secondary antibody to avoid cross-reactivity with the light chain of anti-T. congolense TS mAb. Following washing with TBST buffer, blots were incubated with streptavidin (dilution 1:5000) for 45 min at room temperature. After washing again with TBST buffer, GARP was visualized using the ECL system as stated above.

RESULTS
Cultivation-In axenic cultures the trypanosomes grow as flagellate forms called epimastigotes, which correspond to the forms found in the guts of blood-sucking vectors (procyclic or insect forms) (37). In contrast to the African trypanosome T. brucei brucei, where TS activity was found to be only membrane-bound (6, 7), TS activity in procyclic forms of T. congolense was detected both in the culture supernatant and membrane-bound (5). During the cultivation of procyclic forms of T. congolense in FCS/hemin-containing media the cell number increased during 3-4 days from 1 ϫ 10 6 cells/ml to 7 ϫ 10 6 cells/ml, whereas the parasites died within 3-4 days when cultivated in FCS/hemin-free media. However, in both cases the parasites did produce enzymatic activity during cultivation. When cultivating the parasites in FCS/hemin-containing media 84% of transfer activity and 97% of SA activity were measured in the culture supernatant. In contrast, cultivation of the trypanosomes in FCS/hemin-free media resulted in only 25% of transfer activity and 60% of SA activity in the culture supernatant. In both cases the residual activity was bound to the cell pellet. Additionally, transfer and SA activity measured in the FCS/hemin-containing and FCS/hemin-free media were found to possess similar characteristics (data not shown). However, in the culture supernatant derived from the cultivation without FCS/hemin the specific enzymatic activity was four times higher than that seen in the culture supernatant obtained from FCS/hemin-containing media. Therefore, and because of the quenching effect of hemin on enzyme activity assays, parasites were cultivated in FCS/hemin-free media.
Separation and Purification of Two TS Forms from T. congolense-The separation and partial purification of the two TS forms, summarized in Table I, was achieved by employing a combination of anion exchange chromatography, isoelectric focusing, gel filtration, and, subsequently, immunoaffinity chromatography. Ion exchange chromatography was chosen as a first step in the purification cascade, mainly because of its high capacity, but also due to its ability to sufficiently separate the two enzyme forms (Fig. 1A). Activity-positive fractions eluting at a salt concentration higher than 0.3 M were combined and referred to as TS-form 1, whereas TS-active fractions eluting at a salt concentration below 0.3 M were combined and called TS-form 2. As can be seen in Table I, following Q-Sepharose, a difference in the transfer to SA activity ratios for both TS forms could already be observed. That is, TS-form 1 and TS-form 2 exhibited a transfer to SA activity ratio of 17 and 2.4, respectively. Even though no significant enrichment of transfer and SA activity was obtained, this first chromatography step provided effective separation of TS-form 1 and TS-form 2.
The separation of both activity peaks following chromatography on Q-Sepharose was also observed when culture supernatant derived from FCS/hemin-containing media was used. This indicates that the described TS forms exist under different cultivation conditions and, therefore, represent major transsialidase forms.
Following ion exchange chromatography each form was treated independently. Further purification of TS-form 2 was achieved by IEF (Table I), however, IEF was not particularly effective at further enriching TS-form 1, instead only leading to a loss of TS activity, and was therefore not used in the purification of TS-form 1. The isoelectric points obtained for TS-forms 1 and 2 were found to be pH 4 -5 and pH 5-6.5, respectively (Fig. 1C).
The two TS forms were further purified independently using gel filtration on Sephadex-G150 SF. As shown in Fig. 1 (B and  D), the molecular mass of TS-form 1 and 2 were found to be 350 -600 and 130 -180 kDa, respectively. The transfer activities of TS-form 1 and form 2 were purified by 30-and 150-fold, respectively, but interestingly in the case of TS-form 1 only very low amounts of SA activity could be detected (Table I). This shows that after complete separation and partial purification of two TS forms both possessed very different transfer to SA activity ratios (8000 for TS-form 1; 45 for TS-form 2). The SDS-gel in Fig. 2 depicts the protein pattern during the various stages of purification. For both TS forms a clear enrichment of a 90-kDa band on SDS-PAGE under reduced conditions was observed (Fig. 2, lanes 1-3 and 5-7). However, a number of contaminating proteins still remained. Therefore a further specific purification step was used.
Production of Anti-T. congolense TS Monoclonal Antibodies and Immunoblot Analyses-Both partially purified TS forms were used for mAb production, with a combination of ELISA and the TS binding test, as described under "Experimental Procedures," being employed for the detection of TS-specific antibodies. Clone 7/23 was found, using the TS binding test, to reduce TS activity in a TS-containing sample by 75%. TS activity present on the Dynabeads was also detected for clone 7/23. Monoclonal Ab 7/23 was found to belong to the subclass IgG2b. Additionally, the V region of mAb 7/23 sequenced was analyzed (V L sequence: AY198310; V H sequence: AY198309). After resolving the concentrated supernatant, as well as the two separated TS forms on SDS-PAGE, immunoblotting with anti-T. congolense TS mAb (mAb 7/23) led to the staining of a single protein band at about 90 kDa, which had been shown to be enriched after purification of the two TS forms (Figs. 2 and  3A). In contrast, anti-T. congolense TS mAb did not cross-react with the 70-kDa band representing recombinant T. cruzi TS and the 80-kDa band representing T. brucei brucei TS (Fig. 3B). Furthermore, immunoblotting analysis revealed that by employing anti-T. cruzi TS antiserum a 70-kDa protein band representing recombinant T. cruzi TS was detected, whereas the 90-kDa protein band of the two T. congolense TS forms was not (Fig. 3C). These results are very similar to that seen for antiserum and mAb raised against T. cruzi TS, which showed no cross-reactivity with T. brucei brucei TS (6).
Immunoaffinity Chromatography-Immunoaffinity chromatography employing the anti-T. congolense TS mAb 7/23 was used for further purification of the two TS forms. For both forms the majority of TS activity was eluted using 20 mM sodium citrate buffer, pH 3.0. The transfer activities of TS-form 1 and TS-form 2 were enriched by 1071-and 4200-fold, respectively, whereas the SA activity, which could only be measured for TS-form 2, was enriched by ϳ28-fold. The two purified TS forms migrated with an apparent molecular mass of 90 kDa on SDS-PAGE (Fig. 2, lanes 4 and 8). However, TS-form 1, which was purified to apparent homogeneity, provided a far greater recovery and was therefore used for micro-sequencing. Micro-sequencing-Following the purification scheme outlined in Table I the 90-kDa protein of TS-form 1 (Fig. 2, indicated with an arrow) was excised following SDS-PAGE, and peptides were analyzed after in-gel trypsin digestion. Subsequently, a peptide with the amino acid sequence VVDPTVVAK, in which the mass data showed a high similarity with the mass data of a peptide from T. brucei brucei TS, was obtained. Additionally, a BLAST data base search revealed that the observed peptide showed 100% sequence identity with a peptide from T. brucei brucei TS (amino acids 193-201), as well as with a peptide from one of two T. congolense TS gene sequences (T. congolense TS1) obtained using a PCR-based cloning approach 3 (Table II). With that, the 90-kDa protein of TS-form 1, as well as indirectly through immunoblot analysis the 90-kDa protein of TS-form 2, were identified as a TS. In contrast, the analyzed peptide did not show 100% sequence identity with peptides from T. rangeli SA, T. cruzi TS, and a second T. congolense TS sequence 3 (T. congolense TS2) (Table II).
Immunoblotting with Anti-T. congolense GARP Monoclonal Antibodies-GARP is the major surface glycoprotein of procyclic forms of T. congolense, and it is bound to the parasite membrane by a glycosylphosphatidylinositol (GPI) anchor (38). It has been shown that GARP acts as an excellent acceptor molecule for T. congolense TS and is therefore believed to be the major natural sialic acid acceptor on the surface of procyclic T. congolense (5). Therefore, and because of the differences in the molecular weight of both TS forms, the possibility that GARP might interact with TS was investigated.
Concentrated culture supernatant, and TS forms at various stages of purification, were analyzed by immunoblotting with anti-T. congolense GARP mAb. Under reducing conditions GARP migrates as a 28-to 32-kDa protein band on SDS-PAGE (39), and as can be seen in Fig. 4 a protein band at about 30 kDa was detected. Surprisingly, GARP was only detected in the concentrated culture supernatant and in two purification stages of the higher molecular weight TS form, TS-form 1, with the intensity of the signal increasing proportionally with specific TS activity (Fig. 4). This points toward an association between TS-form 1 and GARP under the mild purification conditions used in Q-Sepharose FF and Sephadex-G150 SF chromatography. However, immunoblot analysis of the immunoaffinity-purified TS-form 1 was unable to detect GARP (Fig.  4), indicating that the final purification step disrupted the possible association between GARP and TS-form 1. This may have been due to the intensive washing steps (1 M NaCl and 20 mM sodium citrate, pH 4.5) used during immunoaffinity chromatography. Interestingly, no bands reacting with anti-T. congolense GARP mAb were observed in all TS-form 2 samples investigated (Fig. 4). Nevertheless, the association of TS-form 1 with GARP must be specific and strong, because anti-T. congolense TS mAb (mAb7/23) was able to co-immunoprecipitate TS and GARP from a concentrated culture supernatant sample (Fig. 5). The 90-kDa protein band visualized following immunoprecipitation represents TS (Fig. 5A, lane 2), and the 32-kDa protein band represents GARP (Fig. 5B, lane 2). The protein band at 55 kDa represents the heavy chain of the T. congolense TS mAb that reacted with the secondary antibody, horseradish peroxidase-conjugated anti-mouse IgG (Fig. 5A, lane 2).
Kinetic Studies-Kinetic studies were performed on the two TS forms following their separation and partial purification using ion exchange chromatography and gel filtration. Various donor and acceptor substrates, generally employed to determine SA and transfer activities, were utilized. As can be seen in Table III, when using MUGal as the acceptor, TS-form 1 and 2 bind the donor substrate ␣2,3-SL with very similar affinities. The K m values for ␣2,3-SL of 0.3 mM and 0.2 mM for TS-form 1 and 2, respectively, are in good agreement with that reported for native T. brucei brucei TS (K m , 0.5 mM) (7). Both TS forms were found to prefer MULac over MUGal as the acceptor, however, the catalytic efficiency (expressed as approximate V max /K m ) is three to four times higher for both acceptors in the case of TS-form 1. TS-form 2 was also found to bind MUNeu5Ac, the donor substrate routinely used to measure the hydrolyzing reaction, with high affinity. A K m value of 0.09 mM is very similar to that reported for the native T. brucei brucei TS (K m , 0.16 mM) (7).
Kinetic parameters for the hydrolyzing reaction could not be determined in the case of TS-form 1, because of insufficient SA activity. Interestingly, in comparison to TS-form 1, which possessed predominately transfer activity, TS-form 2 was found to hydrolyze sialic acid from MUNeu5Ac five times more efficiently than transferring sialic acid from the donor ␣2,3-SL to GalMU (Table III). These results suggest that not only do the two TS forms consist of different transfer to SA activity ratios but, in addition, that TS-form 2 hydrolyzes sialic acid more efficiently than TS-form 1. On the other hand, TS-form 1, which is more efficient in transferring sialic acid, behaves similar to the previously reported native and recombinant T. cruzi TS and T. brucei brucei TS (7, 17, 22, 40). Moreover, both TS-forms exhibited no differences in their pH and temperature optima,  The italicized sequences show 100% identity with the found peptide. The GenBank accession numbers of the sequences are stated in parentheses.

Enzymes
Peptide sequence with a pH optimum of 7 and a temperature optimum of 37-40°C being determined, similar to those reported for T. brucei brucei TS (6, 7).
Donor and Acceptor Substrate Specificities-A number of sialoglycoconjugates were tested as potential sialic acid donors for TS-form 1 and 2 from T. congolense (Table IVA). Both TS forms revealed no differences in their donor substrate specificities, with the exception of apolactoferrin. As has been previously observed for T. cruzi TS (17,41), T. brucei brucei TS (7), and crude T. congolense TS (5), both isolated T. congolense TS forms preferably catalyze the transfer of ␣2,3-linked sialic acid (␣2,3-SL and Neu5Ac␣2,3-N-acetyl-lactosamine). The finding, that sialic acids in ␣2,6-linkage also serve as reasonable donor substrates (Table IVA), differs from the findings reported for T. cruzi TS (41). However, the results are similar to those reported for T. brucei brucei TS, which can also utilize ␣2,6linked sialic acid but at a lower rate in comparison to T. congolense TS (7). However, as previously reported for the crude T. congolense TS (5), fetuin served as a good donor substrate with a high transfer rate being observed for both T. congolense TS forms. The presence of a fucose near the terminal galactose residue (sialyl-Lewis x ) resulted in a decrease in T. congolense TS activity. This has also been shown for T. cruzi TS (17) and T. brucei brucei TS (6). Moreover, sialyloligosaccharides from bovine and human milk, as well as the -casein glycomacropeptide, known to inhibit bacterial and viral adhesion (42), proved to be excellent sialic acid donors for both TS forms (Table IVA).
Various substrates were tested for their ability to act as sialic acid acceptors for T. congolense TS (Table IVB). Both TS forms were found to possess a similar acceptor preference, however, some acceptor substrate specificity differences were observed between the two forms. In agreement with that reported for T. cruzi TS (41), ␤1,4-linked galactose-containing substrates were better acceptors for both T. congolense TS forms than ␤1,3-linked galactose-containing substrates (Table  IVB). Lactose and its derivatives serve as good acceptor substrates. In agreement with earlier studies on T. brucei brucei TS (5, 7), monosaccharides did not act as sialic acid acceptors for the T. congolense TS forms, apart from a slight effect using galactose (Table IVB). Parodi et al. (20) reported that maltose and cellobiose can be sialylated by T. cruzi TS, however, they did not serve as sialic acid acceptors for both T. congolense TS forms, even at a concentration of 5 mM. Moreover, the lipooligosaccharides and lacto-N-tetraose and lacto-N-neotetraose, which prevent the adherence of bacteria to epithelial cells, showed good acceptor properties for both TS forms (Table IVB). This is important because the sialylation of these biologically relevant structures increases their survival in blood serum (43).
However, for all acceptor substrates tested, TS-form 1 was found to utilize the various acceptors at a rate that was about 2-fold higher. This mirrors the kinetic results summarized in Table III, where it is shown that TS-form 1 utilized the various acceptors with a greater efficiency in comparison to TS-form 2.
Inhibitor Studies-Several known viral and bacterial SA inhibitors (44) were tested for their ability to inhibit the SA and transfer activity of both T. congolense TS forms. None of the compounds tested showed any significant inhibitory effect. Neu5Ac, Neu2en5Ac (a natural inhibitor of SA), 4-amino-Neu2en5Ac, as well as the potent inhibitor of influenza virus SA (45), 4-guanidino-Neu2en5Ac, exhibited no more than 20% inhibition of transfer activity for both TS forms at a concentration of 2 mM. Interestingly, the SA activity of TS-form 2 was not inhibited by these compounds at all concentrations tested.
In contrast, the synthetic SA inhibitor N-(4-nitrophenyl)oxamic acid inhibited SA activity by 25% at a concentration of 2 mM but the transfer activity by only 5 and 10% for TS-form 1 and 2, respectively. Furthermore, the anti-malaria drug suramin, which has previously been shown to be a strong inhibitor of ganglioside SA from human brain tissue (IC 50 , 7 g/ml) (46), exhibited a slight (17%) inhibitory effect on the SA activity of TS-form 2 at a concentration of 25 g/ml, whereas the transfer activity of the two TS forms was not effected.
Anti-T. congolense TS mAb was found to only slightly inhibit (20% at 100 g/ml) the transfer activity of both enzyme forms, and it had no effect on the SA activity of TS-form 2. Glycosaminoglycans, like heparan sulfate and chondroitin sulfate A, as well as dextran sulfate are known inhibitors of mammalian sialidases (47), but they neither inhibited the transfer activity of TS-form 1 and 2 nor the SA activity of TS-form 2 at a concentration of 25 mg/ml.
The addition of 1 M NaCl or 1 M KCl resulted in the reduction of SA and transfer activities by greater than 50%; however, full activity could be restored after desalting. Moreover, the addition of 5 mM Co 2ϩ , Zn 2ϩ , and Fe 2ϩ inhibited SA and transfer activities of both TS forms by 20 -40%, whereas the addition of 20 mM Ca 2ϩ , Mg 2ϩ , and Mn 2ϩ had no effect on SA and transfer activities of both TS forms. This confirms earlier findings that TSs are not activated by Ca 2ϩ (5,18) as opposed to most viral and bacterial SA that require Ca 2ϩ for full activity (48). DISCUSSION Apart from TS expressed in trypanosomes, TS activity has also been reported in Endotrypanum species (49), in Corynebacterium diphtheriae (50), and in human plasma (51). However, only trypanosomal TS from the American trypanosome T. cruzi and the African trypanosome T. brucei brucei, have been studied in detail. To expand our knowledge concerning this peculiar enzyme, we investigated the TS expressed from the African trypanosome T. congolense. The fact that T. congolense produces a soluble TS simplified the purification, because the usage of detergents and other substances or enzymes enhancing solubilization, which often decrease enzyme activity, could be avoided. Nevertheless, different cultivation conditions were tried to reduce the content of contaminating protein in the culture supernatant.
Cultivation in FCS-containing media resulted in the majority of the enzyme activity being detected in the culture supernatant, whereas when cultivated in FCS-free media T. congolense TS activity was found to be mainly membrane-associated. It is unclear if T. congolense produces a membrane-bound TS, which is GPI-anchored and released into the medium due to the action of the proteases and phospholipases of the parasites, or if two different T. congolense TS species exist, one soluble and one membrane-bound, which are expressed depending on cultivation conditions. Because the GPI anchor has no influence on the enzymatic activity (52) and soluble proteins are easier to purify, we isolated T. congolense TS from concentrated culture supernatant using FCS-free media, because the specific activity of TS was markedly increased.
Employing different purification techniques, two major peaks of TS activity were detected, both possessing SA and transfer activity but differing in their transfer to SA activity ratio, as well as molecular weights and isoelectric points. Following SDS-PAGE, both purified T. congolense TS forms ap-peared as a single 90-kDa protein band, indicating that they may be aggregates of the same monomer. TS-form 2, as observed by gel filtration, seems to form homodimers (ϳ180 kDa), whereas TS-form 1 probably exists as oligomers (tetramer or higher), resulting in the high molecular mass observed by gel filtration (ϳ350 -600 kDa).
Similar findings have been reported for T. cruzi and T. brucei brucei TS. That is, TS from T. cruzi trypomastigotes generates multimeric aggregates (15,16), the monomers varying from 120 to 180 kDa in the cell-derived, bloodstream forms and 90 kDa in the insect forms (53). In T. brucei brucei the TS was also found to be multimeric, with two major broad peaks of ϳ180 and ϳ660 kDa. SDS-PAGE under reduced conditions of those peaks revealed major bands between 60 and 80 kDa (6). Other studies reported the purification of a 67-kDa monomeric surface TS from T. brucei brucei (7,18). However, so far, the reason for the generation of multimeric aggregates of TS, as well as their composition, has not been studied.
In the case of T. congolense we were able to prove, using micro-sequencing, that the isolated 90-kDa protein of TS-form 2 is indeed a TS. Interestingly, in both T. congolense TS forms  a 90-kDa protein band was detected using the anti-T. congolense TS mAb 7/23, indicating that TS-form 2, which was not micro-sequenced, could also be identified as TS. However, further investigations are necessary to determine whether the 90-kDa proteins from both forms represent monomers with the same primary structure and perhaps different folding or are products of different TS genes, because the anti-T. congolense TS mAb 7/23 might recognize a common epitope or tertiary structure of both TS forms. The finding, that under mild conditions GARP is co-purified and co-immunoprecipitated with TS-form 1, seems plausible for a number of reasons. First, GARP is the natural substrate of T. congolense TS. Second, protein complexes concerning various SA (54) and protein-protein interactions involving other glutamic acid-rich proteins, also referred to as GARPs, have been reported (55,56). Moreover, the possibility that the trypanosomal GARP can mediate or facilitate the formation of oligomers of TS-form 1 is supported by the finding that the interaction between the cGMP-gated channel and peripherin-2 proteins of the rod photoreceptor outer segment of vertebrates are mediated by a glutamic acid-rich protein (56). At this stage, it is unclear whether an association between trypanosomal GARP and TS-form 1 would have a stabilizing effect on TSform 1 that could account for not only the differences in transfer and SA activities but also the differences in kinetic properties observed for TS-form 1 and 2. However, Poetsch et al. (56) also observed an interaction between mammalian phosphodiesterase and mammalian GARP, which led them to speculate that this association may play a role in modulating phosphodiesterase activity.
The fact that only TS-form 1 is co-purified and co-immunoprecipitated with GARP may be due to the possibility that only one of the two TS forms, TS-form 1, can interact with GARP. Considering that TS-form 1 has been shown to utilize various acceptor substrates with a higher catalytic efficiency, it is possible that GARP stabilizes TS-form 1 or that the 90-kDa protein bands of TS-forms 1 and 2 actually represent monomers of identical molecular weight but are encoded by two different genes, from which one codes for a protein with a slightly enhanced acceptor binding capacity (TS-form 1). This concept is sustained by recent findings identifying two different T. congolense TS gene sequences, which share only 50% identity with each other. 3 The isolated T. congolense TS forms were found to have the same donor and acceptor substrate preferences; however, interestingly, TS-form 1 utilized the acceptor substrates more efficiently than TS-form 2. On the other hand, the various donor substrates tested were utilized with similar efficiencies by both TS forms. Furthermore, SA activity was predominately found in TS-form 2, whereas TS-form 1 possesses significantly less SA activity and higher transfer activity. Taken together, these results suggest that the activity associated with TS-form 2 mainly represents SA, in which transfer activity is decreased, possibly due to reduced acceptor binding capacity. This is further substantiated by the finding that some known SA inhibitors, such as suramin inhibited SA activity of T. congolense TS, whereas the transfer activity was unchanged. In contrast, Neu2en5Ac, a known SA inhibitor, inhibited transfer activity but not SA activity.
At this stage a potent TS inhibitor is not available. This may reflect the complexity of the SA and transfer mechanism of TS. It might be possible that donor substrate analogues as TS inhibitors may only be effective in the presence of an appropriate acceptor substrate analogue. That is, the synthesis of multivalent inhibitors possessing both a donor and acceptor sub-strate analogue may provide very specific TS inhibitors for combating trypanosomiasis.
In conclusion, using micro-sequencing we have identified a single protein of ϳ90 kDa as T. congolense TS. We were able to produce anti-T. congolense TS mAb (mAb 7/23), which could, because of its specificity, be a valuable diagnostic tool for distinguishing between procyclic forms of T. brucei brucei TS and T. congolense TS. We also show for the first time that TS forms exist that differ remarkably in their transfer to SA activity ratios, with these differences possibly being due to different acceptor substrate binding capacities. Moreover, we demonstrated another intriguing difference between both TS forms, the ability of TS-form 1 to associate with GARP. Therefore one can speculate that not only do active and inactive TS forms exist but also TS forms with different transfer efficiencies, perhaps due to variations in their primary structure and/or protein folding or due to the possible stabilization of TS through an interaction with GARP.
Studies on the enzymatic transfer of sialic acid catalyzed by TS, in particular substrate specificities, will enhance its biotechnological applications. For example, TS could potentially be utilized for the sialylation of the T-and T N -antigens, as has been shown using bacterial SA (57) and human and mouse recombinant sialyltransferases (58). The sialylation of recombinant glycoproteins, as well as human milk compounds like lacto-N-neotetraose and lacto-N-tetraose, could also conceivably be carried out utilizing TS.
Given that trypanosomiasis has reached epidemic magnitude in some countries, one should consider methods to control not only the disease but also its transmission stage inside the vector (59). Efforts have been made to establish bush-free belts to reduce the spread of the tsetse fly. Taking this into consideration, the development of various TS inhibitors could not only serve in combating trypanosomes inside the host, in the case of T. cruzi, but also inside the vector, in the case of T. brucei brucei and T. congolense.