Interaction of triosephosphate isomerase from the cell surface of Staphylococcus aureus and α-(1→3)-mannooligosaccharides derived from glucuronoxylomannan of Cryptococcus neoformans

The glycolytic enzyme triosephosphate isomerase (TPI; EC 5.3.1.1) of Staphylococcus aureus is a candidate adhesion molecule for the interaction between the bacterium and the fungal pathogen Cryptococcus neoformans. TPI may recognize the mannan backbone of glucuronoxylomannan (GXM) of C. neoformans. We purified TPI from extracts of S. aureus surface proteins to investigate its binding by surface plasmon resonance analysis. The immobilized TPI reacted with GXM in a dose-dependent manner. Furthermore, the interactions between staphylococcal TPI and α-(1→3)-mannooligosaccharides derived from GXM were examined. The oligosaccharides exhibited binding with TPI; however, monomeric mannose did not. Differences in the slopes of the sensorgrams were observed between oligosaccharides with an even number of residues versus those with an odd number. A heterogeneous ligand-parallel reaction model revealed the existence of at least two binding sites on TPI. The enzymic activities of TPI were inhibited in a dose-dependent manner by α-(1→3)-mannooligosaccharides larger than triose. The binding of TPI and α-(1→3)-mannotriose near the substrate-binding site was predicted in silico (AutoDock 3.05). An oligosaccharide of size equal to or greater than triose could bind to the site, affecting enzymic activities. Moreover, affinities were indicated, especially for biose and tetraose, to another binding pocket, which would not affect enzymic activity. These data suggest a novel role for TPI, in addition to glycolysis, on the surface of S. aureus.


INTRODUCTION
The search for novel targets for the regulation of microbes has focused on interactions between micro-organisms. Diffusible signalling molecules that play essential roles in bacterial cellcell communication have been identified (Kaper & Sperandio, 2005;Ryan & Dow, 2008;von Bodman et al., 2008). Concentrating on the fungal2bacterial interaction, we previously reported contact-mediated killing of Cryptococcus neoformans by Staphylococcus aureus (Saito & Ikeda, 2005).
Cryptococcus neoformans, an encapsulated yeast pathogen that causes severe meningitis, is found predominantly in avian excreta, especially that of the pigeon. This fungus is characterized by a capsule composed mainly of the acidic heteropolysaccharide glucuronoxylomannan (GXM), which has a backbone consisting of a-1,3-mannan with a single branch of b-1,2-xylose or glucuronic acid per mannose residue. C. neoformans enters the human body via the respiratory pathway (Casadevall & Perfect, 1998). S. aureus is found in the nasal cavities of a large percentage of normal healthy subjects (Kluytmans et al., 1997). In S. aureus infection, nasal colonization is a risk factor for invasive infection (von Eiff et al., 2001;Foster, 2004;Park et al., 2008), and the elimination of S. aureus nasal carriage reduces the development of the infection. However, the presence of S. aureus in the nasal mucosa may also provide a defence against the entry of C. neoformans.
Previously, we found that cell-cell contact is required for the apoptosis-like cell death of C. neoformans (Ikeda & Sawamura, 2008). The cell surface adhesion molecules involved in this interaction were identified as triosephosphate isomerase (TPI) and a-(1A3)-mannooligosaccharides larger than triose in S. aureus and C. neoformans, respectively (Ikeda et al., 2007). In the present study, we purified staphylococcal TPI and examined its interactions with a-(1A3)-mannooligosaccharides derived from C. neoformans in order to characterize the binding affinities and predict binding models for TPI.
Preparation of surface proteins from S. aureus. Surface protein extracts from S. aureus were prepared as previously described (Ikeda et al., 2007;Komatsuzawa et al., 1997). Briefly, S. aureus cells that had been cultured in Trypticase soy broth (TSB; Becton Dickinson) at 37 uC for 6 h were suspended in 3 M LiCl. After mixing gently on ice for 15 min, the cell extract was dialysed against 10 mM phosphate buffer (pH 6.8) and the surface proteins were collected. TPI activity was not detected in the cellular debris remaining after surface protein isolation.
TPI activity assay. Enzymic activity was coupled to the oxidation of NADH by glycerol-3-phosphate dehydrogenase and measured as a change in absorbance at 340 nm, according to the method of Rozacky et al. (1971) with minor modifications. Briefly, S. aureus surface proteins were incubated with 0.45 mM DL-glyceraldehyde 3-phosphate as the substrate, 0.03 mg NADH and 3 mg glycerol-3-phosphate dehydrogenase in a final volume of 0.3 ml, and the A 340 of the supernatant was monitored.
Purification of TPI from S. aureus. To purify TPI from the sample, ammonium sulfate at 50 % saturation was added followed by centrifugation to remove the precipitates, and the supernatant was dialysed against 50 mM phosphate buffer (pH 6.8). Additional ammonium sulfate (2.5 M) was added and the sample was applied to a column containing about 0.5 ml phenyl-Sepharose CL-4B. The protein was then eluted with a step gradient of 2.520 M ammonium sulfate in 50 mM phosphate buffer (pH 6.8). The fraction with TPI activity was dialysed against 10 mM phosphate buffer (pH 6.8) and the sample was applied to a column containing about 0.5 ml DEAE-Toyopearl 650M (Tosoh Bioscience). The protein was eluted with a step gradient of 021.0 M NaCl in 10 mM phosphate buffer (pH 6.8). The fraction with TPI activity was dialysed against PBS (pH 7.2) and the sample was applied to a Bio-Gel P-100 column (1.56100 cm; Bio-Rad) for final purification.
Molecular mass of TPI. SDS-PAGE was performed at a constant current of 20 mA using a Multi Gel II mini 12.5 system (Cosmo Bio). After electrophoresis, the bands were silver stained. Briefly, the gel was fixed in 25 % (v/v) ethanol and 5 % (v/v) acetic acid for 3 h or overnight, washed twice for 10 min each with deionized water, incubated in sodium thiosulfate (40 mg in 200 ml) for 1 min, and washed twice for 1 min each with deionized water. The gel was then incubated with silver nitrite solution (200 mg in 200 ml) for 30 min, washed once for 30 s with deionized water, and developed for about 10 min with 2 % (w/v) sodium carbonate and 0.015 % (v/v) formaldehyde solution. The reaction was quenched with acetic acid, and the gel was washed three times for 5 min each with deionized water. The molecular mass of the purified protein was determined based on a set of protein markers (broad-range; ColourPlus Prestained Protein Marker, BioLabs).
Preparation of oligosaccharides. Oligosaccharides were prepared as previously described (Ikeda et al., 2007). Briefly, the controlled Smith degradation product was hydrolysed with 0.4 M H 2 SO 4 at 100 uC for 1 h. After neutralization with BaCO 3 , the product was applied to a Bio-Gel P-2 column (2.56120 cm; Bio-Rad) and eluted with water (10 ml h 21 , 2 ml fractions). Carbohydrate concentrations were determined by the phenol/sulfuric acid method. To identify mannooligosaccharides, the NMR spectra of the oligosaccharide fractions were recorded (Ichikawa et al., 2001;Ikeda & Maeda, 2004).
Surface plasmon resonance (SPR) analysis of the interaction between cryptococcal GXM and staphylococcal TPI. The interaction of S. aureus TPI and C. neoformans GXM was analysed using a Biacore 3000 biosensor system (GE Healthcare). The purified TPI, confirmed using SDS-PAGE followed by silver staining (Fig. 1), was diluted with 10 mM sodium acetate buffer (pH 3.48) and immobilized on a standard sensor chip CM 5 (GE Healthcare) using an amine coupling kit, according to the manufacturer's instructions. The analyte, GXM, was diluted with HBS-EP running buffer [10 mM HEPES (pH 7.4), 150 mM NaCl, 3 mM EDTA and 0.005 % (w/v) surfactant P20] and injected into the flow cell. The buffer flow rate was maintained at 10 ml min 21 for the immobilization and at 20 ml min 21 for the analysis.
SPR analysis of the interaction between a-(1A3)-mannooligosaccharides and TPI. The interactions between TPI and a-(1A3)mannooligosaccharides derived from C. neoformans were determined as described in the preceding section, except that a-(1A3)mannooligosaccharides were used as analytes. Each analyte was diluted with HBS-EP buffer and the buffer flow rate was maintained at 20 ml min 21 .
Docking simulation for mannotriose from C. neoformans and TPI from S. aureus. The docking simulation of mannotriose and S. aureus TPI was performed using AutoDock 3.05 (Morris et al., 1998). The PDB file of mannotriose was obtained from the Sweet website (http://www.dkfz-heidelberg.de/spec/) (Bohne et al., 1999). The PDB file of TPI from S. aureus was obtained from the SWISS-MODEL website (Guex & Peitsch, 1997;Schwede et al., 2003).
Inhibition of TPI activity by mannooligosaccharides. We investigated the possible inhibition of TPI activity by the mannooligosaccharides mannose, mannobiose, mannotriose, mannotetraose and mannopentaose. TPI activity in the presence of each mannooligosaccharide at 0, 10 and 100 mM was determined using the TPI assay described above.

RESULTS
Purification of TPI from S. aureus As summarized in Table 1, TPI was purified from S. aureus surface proteins by salting out followed by hydrophobic, anion-exchange, and gel-filtration column chromatography. The purified TPI had a molecular mass of 28.5 kDa, as determined by SDS-PAGE (Fig. 1). The yield of 1.7 mg TPI reflected a 2400-fold purification.
Interaction between GXM from C. neoformans and TPI from S. aureus Fig. 2 shows the SPR sensorgrams for the interaction between GXM and immobilized TPI. The SPR response occurred in a dose-dependent manner.

saccharides and TPI
The SPR analyses revealed that the a-(1A3)-mannooligosaccharides derived from C. neoformans interacted with TPI from S. aureus in a dose-dependent manner (Fig. 3).
The rate at which the response increased (expressed in response units) over the injection time (0-60 s) for each mannooligosaccharide decreased in the following order: mannopentaose, mannotetraose, mannotriose and mannobiose; mannose did not display interaction with TPI. However, from 60 to 600 s, the rate of increase in the response was greater for mannotetraose and mannobiose than for mannopentaose and mannotriose. The sensorgrams of the TPI interactions showed steeper slopes for oligosaccharides containing an even number of residues than for those with an odd number of residues, suggesting a difference in the mode of binding between the two types of oligosaccharides.

Analysis and evaluation of the interactions between TPI and a-(1A3)-mannooligosaccharides
The association and dissociation kinetics of each mannooligosaccharide were evaluated using BIAevaluation version 4.1. Based on the sensorgrams, a 1 : 1 Langmuir binding model was not considered for the interaction between TPI and a-(1A3)-mannooligosaccharides. Instead, we used a heterogeneous ligand-parallel reaction model (Smith et al., 2003;Yaqub et al., 2003;Mader et al., 2004) and a two-state reaction model to study the binding kinetics. As shown in Table 2, the heterogeneous ligandparallel reaction model indicated two binding sites. Similar affinity constants (K d1 ) were calculated for the a-(1A3)mannooligosaccharides at the first binding site, although the order depended on size (mannopentaose,mannotetraose,mannotriose,mannobiose); however, the second parallel reaction affinity (K d2 ) clearly increased according to the order mannotetraose, mannobiose, mannopentaose and mannotriose. The differences among these affinities might have contributed to the variation in the sensorgrams. In contrast, the two-state reaction model gave similar affinity constants (K d ) among the oligosaccharides, with values and an order similar to those of K d1 .
Docking simulation for mannotriose from C. neoformans and TPI from S. aureus We performed docking simulation analyses for S. aureus TPI (253 amino acids) with AutoDock 3.05 using a PDB file for TPI from S. aureus obtained by homology modelling. Several possible conformations were obtained, and the interaction between TPI and mannotriose near the substrate binding site was suggested (Fig. 4).

DISCUSSION
Glycolytic enzymes have been found on the cell surfaces of certain bacteria and fungi. In addition to their roles in glycolysis, some glycolytic enzymes, such as D-glyceraldehyde-3-phosphate dehydrogenase, that occur both on the cell surface and in a secreted form have been shown to possess multi-functional properties, including adhesion and nutrient uptake (Modun & Williams, 1999;Bergmann et al., 2004;Gatlin et al., 2006;Terao et al., 2006;Antikainen et al., 2007;Egea et al., 2007). Furthermore, interactions have been reported between glycolytic enzymes and human bioactive proteins, such as fibrinogen, plasminogen and fibronectin, and glycolytic enzymes expressed on microbial cell surfaces have become targets for the regulation of bacteria and fungi (Pancholi & Fischetti, 1992;Gil-Navarro et al., 1997;Pancholi & Fischetti, 1998;Bergmann et al., 2001).  Previously, we found that the bacterium S. aureus adheres to and kills the fungal pathogen C. neoformans (Saito & Ikeda, 2005). We identified TPI as a candidate adherence molecule on S. aureus that recognizes the mannan backbone in GXM, a major component of the capsule of C. neoformans (Ikeda et al., 2007). The glycolytic enzyme TPI, which catalyses the reversible reaction between Dglyceraldehyde 3-phosphate and dihydroxyacetone phosphate, is located on the cell surface, although it is not as abundant as the cell surface glyceraldehyde-3-phosphate dehydrogenases. TPI is also thought to play a role in adhesion with the fungal pathogen Paracoccidioides brasiliensis (Pereira et al., 2007).
In this study, we purified TPI from S. aureus and characterized its interaction with a series of a-(1A3)mannooligosaccharides, with various degrees of polymerization, obtained from C. neoformans. TPI was purified from S. aureus surface proteins by salting out and hydrophobic, anion-exchange and gel-filtration column chromatography. The purified TPI yielded a single band in SDS-PAGE. SPR analysis of the interaction between TPI from S. aureus and GXM from C. neoformans showed that purified TPI interacted with GXM in a dose-dependent manner, providing evidence that staphylococcal TPI and cryptococcal GXM participate in the adherence between these two organisms. Subsequently, we examined the interactions between purified TPI and a-(1A3)-mannooligosaccharides derived from the backbone of GXM. Mannose did not display specific binding. During the first 60 s after injection, the SPR responses of mannobiose and mannotetraose were lower than those of mannotriose and mannopentaose, respectively. However, between 60 and 600 s, the responses of mannobiose and mannotetraose were greater than those of mannotriose and mannopentaose, respectively. For the kinetic analyses, two fitting models were selected: the heterogeneous ligand-parallel reaction model and the two-state reaction model. Two binding affinities, K d1 and K d2 , were calculated from the former model; the K d1 values were similar among the oligosaccharides, whereas the second reaction appeared to occur with significantly different affinities for the oligosaccharides. This suggests that TPI has at least two binding sites with different affinities. In the two-state reaction model, binding is followed by a conformational change (A+B«AB«AB x ). The affinity increased slightly with increasing number of mannose residues. The sensorgrams showed an increasing response according to the order mannotetraose, mannobiose, mannopentaose and mannotriose. We considered the heterogeneous ligand-parallel reaction model to be a better fit for the interaction of a-(1A3)-mannooligosaccharides, especially even-numbered oligosaccharides, and TPI from S. aureus.
The values for K d2 were higher for odd-numbered compared to even-numbered oligosaccharides. Based on the K d2 and the shape of the sensorgrams, the participation of the second affinity may be less significant for the oddnumbered oligosaccharides. Previously, we proposed threedimensional models of the oligosaccharides (Ikeda et al., 2007). The characteristic structure with a curve appeared in oligosaccharides larger than mannotriose. From the curved chemical structures, we inferred that common binding sites within oligosaccharides of a size equal to or greater than that of triose could be postulated. In the docking simulation, a conformation was suggested that included the formation of hydrogen bonds to Lys 11 , Asn 13 , Gly 175 , Ser 215 and Lys 217 . Furthermore, other binding pockets, which would not affect enzyme activity, could be predicted for biose and tetraose. In particular, mannobiose may bind TPI in a multiple-binding manner; therefore, the affinity of mannobiose was higher than that of mannopentaose or   mannotriose. Comparing mannobiose and mannotetraose, mannotetraose displayed a higher affinity (K d2 1.46610 28 ) than mannobiose (K d2 5.37610 26 ). Overall, mannotetraose showed the highest affinity among the oligosaccharides tested.
The results of computational docking and inhibition assays for TPI activity revealed the possibility that mannooligosaccharides larger than mannotriose may bind to S. aureus TPI near the substrate-binding site. The structural model of TPI suggests that this enzyme may have multiple binding pockets and that they bind to mannose residues.
It is important to characterize the binding of TPI proteins to C. neoformans cells, including their localization. For this purpose, it is necessary to prepare anti-TPI antibodies. The preparation of sufficient TPI, including recombinant TPI, should be helpful.
Since mannose plays an important role in recognizing molecules involved in infection and innate immunity (van de Wetering et al., 2004;Dommett et al., 2006), the presence of unknown substances that interact with TPI is expected. Therefore, TPI may be a multifunctional protein that binds with eukaryotic glycoproteins. The use of carbohydrates as anti-adhesion drugs for the regulation of pathogens has also been reported (Liu et al., 2000;Sharon, 2006). The elucidation of the role of TPI as a virulence factor of S. aureus may prove valuable for screening novel anti-infective agents.