Thermodynamic Analysis for Binding of 4-O-β-tri-N-acetylchitotriosyl Moranoline, a Transition State Analogue Inhibitor for Hen Egg White Lysozyme

4-O-β-tri-N-acetylchitotriosyl moranoline (GN3M) is a transition-state analogue for hen egg white lysozyme (HEWL) and identified as the most potent inhibitor till date. Isothermal titration calorimetry experiments provided the thermodynamic parameters for binding of GN3M to HEWL and revealed that the binding is driven by a favorable enthalpy change (ΔH° = −11.0 kcal/mol) with an entropic penalty (−TΔS° = 2.6 kcal/mol), resulting in a free energy change (ΔG°) of −8.4 kcal/mol [Ogata et al. (2013) 288, 6,072–6,082]. Dissection of the entropic term showed that a favorable solvation entropy change (−TΔS solv° = −9.2 kcal/mol) is its sole contributor. The change in heat capacity (ΔC p°) for the binding of GN3M was determined to be −120.2 cal/K·mol. These results indicate that the bound water molecules play a crucial role in the tight interaction between GN3M and HEWL.


INTRODUCTION
Lysozyme (EC 3.2.1.17) is an enzyme that hydrolyzes the β-1,4-glycosidic bond between N-acetylmuramic acid and N-acetylglucosamine (GlcNAc) in peptidoglycan, a major structural component of the bacterial cell wall (Jollès and Jollès, 1984). Hen egg white lysozyme (HEWL) is the first enzyme to have its three-dimensional structure determined by X-ray diffraction (Blake et al., 1965), consequently its catalytic mechanism has been intensively studied. Based on the modeled structure of HEWL-chitohexaose, it was proposed that HEWL has six subsites for binding of sugar residues in the active site cleft, termed −4 to +2 (formerly A, B, C, D, E, and F), and the cleavage occurs between sugars located at subsites −1 and +1 through the cooperative action of Glu35 and Asp52 (Johnson and Phillips, 1965;Blake et al., 1967;Phillips, 1967). In the catalytic reaction, Glu35 is thought to act as a general acid catalyst to protonate the glycosidic oxygen, while Asp52 acts as a conjugate base and stabilizes the carbonium ion intermediate that adopts a half-chair conformation with C1 carbon displaying sp 2 hybridization. This is known as the Phillips mechanism, widely supported by a number of experimental observations including mutagenesis studies of these amino acids (Malcolm et al., 1989).
Meanwhile, Vocadlo et al. reported the crystal structure of HEWL, covalently bound to C1 carbon of the −1 sugar, which exhibits a chair conformation with C1 carbon in sp 3 hybridization (Vocadlo et al., 2001). Recently, we synthesized 4-O-β-tri-N-acetylchitotriosyl moranoline (GN 3 M) from chitoteratetraose (GlcNAc) 4 and moranoline with the aid of the lysozyme-catalyzed transglycosylation reaction and examined its inhibitory action toward HEWL ( Figure 1) (Ogata et al., 2013). GN 3 M with C1 carbon in sp 3 hybridization was found to be the most potent lysozyme inhibitor till date with an inhibition constant K i of 1.84 μM and bind tightly to HEWL (K d 0.76 μM at 25°C). Furthermore, in the HEWL-GN 3 M complex structure, GN 3 M was well superimposed on NAG2FGlcF ((GlcNAc) 2 -fluoro-glucosyl fluoride), covalently bound to HEWL mutant E35Q and the moranoline moiety bound to subsite −1 was in a chair conformation (without distortion) as the −1 sugar of the covalently bound NAG2FGlcF. From these results, we concluded to support the covalent glycosyl-enzyme intermediate formation in the reaction catalyzed by the wild-type HEWL. This is known as the Koshland mechanism, now more widely accepted by enzyme researchers (Koshland, 1953).
In this study, in order to elucidate the driving forces responsible for the tight binding of GN 3 M (a transition-state analogue inhibitor) to HEWL, we conducted a detailed thermodynamic analysis using isothermal titration calorimetry. The thermodynamic data obtained from our study would be useful to understand the substrate binding mechanism of HEWL and to design novel glycosidase inhibitors with moranoline moiety.

MATERIALS AND METHODS
Materials 4-O-β-tri-N-acetylchitotriosyl moranoline (GN 3 M) was prepared by lysozyme-mediated transglycosylation from the substrates tetra-N-acetylchitotetraose (GlcNAc) 4 , and moranoline (1deoxynojirimycin) as described previously (Ogata et al., 2013). HEWL, which was recrystallized six times, was purchased from Seikagaku Kogyo Co. (Tokyo, Japan). All other reagents were of the highest quality commercially available and were used without further purification.

Isothermal Titration Calorimetry (ITC) Experiments
The HEWL solution (45 μM) in 20 mM phosphate buffer (pH 7.0) was degassed and its concentration was determined by measuring the absorbance of ultraviolet light at 280 nm. GN 3 M was dissolved (0.5 mM) in 20 mM phosphate buffer (pH 7.0), degassed, and loaded into a syringe, whereas the HEWL solution (0.2028 ml) was loaded into the sample cell. Calorimetric titration was performed with an iTC 200 system (Microcal Northampton, MA, United States). For the titrations, 2.5 μl of a ligand was injected into the sample cell at an interval of 180 s with a stirring speed of 1,000 rpm. The heat of dilution caused by an injection of GN 3 M was measured under identical buffer, injection, and temperature conditions but by adding ligand to a sample solution that lacked protein. The heat of dilution was subtracted from the heat change that occurred in presence of the protein.
Origin ® software was used to analyze the ITC data. Using the single-site binding model, individual datasets obtained from the titration experiments fitted well to the theoretical curves, providing stoichiometries (n), equilibrium binding association constants (K a ), and enthalpy changes (ΔH°) of the protein-ligand interactions. The value of n was found to be between 1.02 and 1.22 for all titrations. The binding free energy change (ΔG°) and the entropy change (ΔS°) were calculated from the relationship described in Eq. 1.
To examine temperature dependence, ITC measurements were performed at pH 7.0 and the temperature was varied in fivedegree increments from 15 to 35°C. Methods for obtaining the heat capacity change (ΔC p +) and parameterizing the entropic term have been described previously (Ohnuma et al., 2011). Briefly, the binding heat capacity change (ΔC p°) was obtained from the ITC titrations over temperature range. Measurements of temperature dependence of ΔH°for GN 3 M binding to HEWL in

RESULTS AND DISCUSSION
In addition to chitin oligosaccharides (GlcNAc) n (n 2-3), we previously investigated the inhibitory effects of 4-O-β-tri- Figure 1). GN 3 M was found to be the most potent inhibitor (K i 1.84 μM). Based on the thermodynamic parameters provided by ITC experiments for HEWL binding of these compounds, GN 3 M was found to be a tight binding inhibitor (ΔG° −8.4 kcal/mol; K d 0.76 M) (Ogata et al., 2013). Figure 2A shows a typical ITC thermogram and theoretical fit to the experimental data for binding of GN 3 M at 20°C. Figure 3A shows superimposed structures of (GlcNAc) 3 -liganded HEWL (PDB code 1lzb) and GN 3 M-liganded HEWL (PDB code 4hp0) complexes (Maenaka et al., 1995;Ogata et al., 2013). The main chain of GN 3 M-liganded HEWL overlapped well with that of (GlcNAc) 3 -liganded HEWL with a RMS deviation of 0.235 Å in the superimposition of the corresponding 128 Cα-atoms. The (GlcNAc) 3 moiety of GN 3 M bound to HEWL also overlapped with (GlcNAc) 3 bound to HEWL. All saccharide rings in the complex structures were in the 4 C 1 chair conformation with small variations. Thus, the potent inhibitory activity and tight binding ability of GN 3 M (a (GlcNAc) 3 -moranoline conjugate) towards HEWL, in comparison to other compounds may be attributed to the moranoline moiety of GN 3 M. The ΔG°value for (GlcNAc) 3 binding to HEWL was found to be −6.9 kcal/mol. In this case, (GlcNAc) 3 binds to the subsites −4, −3, and −2 of HEWL, as shown in the crystal structure of the HEWL-(GlcNAc) 3 complex ( Figure 3B, left). Therefore, it appears that moranoline residue attached to the reducing end of (GlcNAc) 3 contributes to the binding free energy of −2.5 kcal/mol (ΔΔG° −8.4 − (−6.9) kcal/mol), resulting in a 12.2-fold enhancement in the binding affinity (K d of 0.76 μM for GN 3 M and 9.3 μM for (GlcNAc) 3 ). Williams et al. synthesized the xylobio-deoxynojirimycin analogue (a xylanase inhibitor and xylose-moranoline conjugate) and demonstrated that it binds to the retaining family of 10 xylanase Cex from Cellulomonas fimi approximately 830-fold more tightly than xylobiose (K i of 5.8 μM for xylobio-deoxynojirimycin analogue and 4,800 μM for xylobiose) . Arai et al. synthesized (glucose) n -deoxynojirimycin (n 1-8) conjugates by the transglucosylation action of bacterial saccharifying amylases and showed that these compounds have inhibitory activity against αand β-amylases of various origins (Arai et al., 1986). Therefore, conjugation of moranoline to the reducing end site of mono-or oligosaccharides is an excellent strategy to design potent inhibitors of the target glycosidases. Since GN 3 M lacks an acetamido group at the C2 of the moranoline moiety bound to the −1 subsite of HEWL ( Figure 3B, right), conjugation of N-acetylmoranoline to the reducing end of (GlcNAc) 3 could improve the inhibitory activity of (GlcNAc) 3 (de la Fuentea et al., 2016).
Determination of temperature dependence of ΔH°for binding of GN 3 M at pH 7.0 in the temperature range of 15-35°C, yields a change in the reaction heat capacity (ΔC p°) ( Figure 2B). The ΔC p°v alue for GN 3 M binding to HEWL was found to be −120.2 cal/K·mol (Table 1). Negative heat capacity changes are often attributed to the release of well-ordered water molecules from the interface between the protein and the ligand into a bulk solvent by forming hydrophobic contacts. Therefore, we decomposed the entropic terms for binding of GN 3 M to HEWL as follows. By recognizing that the entropy of solvation is close to zero for proteins near 385 K, ΔC p°c an be related to the solvation entropy change (ΔS solv°) of the binding reaction at t GN 3 M -120.2 2.6 2.4 -9.2 9.3 a Data are derived from the temperature dependence of ΔH°. b ΔS mix° Rln (1/55.5) −8 cal/K mol (Baker and Murphy, 1997). c ΔS solv° ΔC p ln (T 298 K /T 385 K ) (Baldwin, 1986;Murphy et al., 1990;Baker and Murphy, 1997 25°C as described in Eq. 2 (Baldwin, 1986;Murphy et al., 1990;Baker and Murphy, 1997).
Furthermore, the mixing entropy change (ΔS mix°) is a statistical correction that reflects mixing of solute and solvent molecules and the changes in translational/rotational degrees of freedom as described in Eq. 3 (Baker and Murphy, 1997).
The conformational entropy change (ΔS conf°) details the change in side-chain and backbone conformational entropy associated with binding. The reaction entropy change (ΔS°), which is derived from the ITC experiment, can be viewed as the sum of ΔS solv°, ΔS mix°a nd ΔS conf° ( Baker and Murphy, 1997). The results summarized in Table 1 show that at pH 7.0 − TΔS solv°i s equal to −9.2 kcal/mol (ΔS solv° 30.8 cal/K·mol), −TΔS mix° 2.4 kcal/mol (ΔS mix° −8.0 cal/K·mol) and −TΔS conf°i s equal to 9.3 kcal/mol (ΔS conf° −31.1 cal/K·mol). As shown in Table 1, while the −TΔS solv°v alue for binding was negative (−9.2 kcal/ mol), the −TΔS conf°v alue was positive (9.3 kcal/mol), resulting in a net entropic penalty of 2.6 kcal/mol. These results indicate that the favorable solvation entropy change is the only contributor in the entropic term. A negative value of −TΔS solv°( positive value of ΔS solv°) implies water molecules are expulsed upon ligand binding due to hydrophobic interactions. Although the precise information regarding the solvation state of HEWL before and after binding to GN 3 M is not available, upon GN 3 M binding to HEWL, the apolar solvent accessible surface area (ASA apolar ) was reduced with over 7%. On the other hand, the same difference in ASA apolar between (GlcNAc) 3 -bond and ligand free HEWL was negligibly small (calculated using GetArea 1.1, TX, United States, based on the crystal structures to ligand-bound and ligand-free HEWL) ( Table 2) (Fraczkiewicz and Braun, 1998). Figure 3B shows the binding modes of (GlcNAc) 3 and GN 3 M to HEWL. Compared to the HEWL-(GlcNAc) 3 complex structure, more water molecules were observed at HEWL-GN 3 M interfaces. However, the solvation entropy change upon GN 3 M binding to HEWL is favorable (positive) ( Table 1). Therefore, it is likely that dehydration for GN 3 M binding involves the ligand, the active site, or, in case of conformational changes upon binding, the bulk of protein.
In this study, the thermodynamic analysis for binding of GN 3 M yielded valuable information regarding the driving forces behind the tight binding of this inhibitor. In addition to the enthalpic contribution, the favorable solvation entropy change is the only contributor in the entropic term. Since they mimic the transition states of the hydrolytic reactions, iminosugar-conjugated glycosidase inhibitors containing moranoline residues have been designed and synthesized (Arai et al., 1986;Notenboom et al., 2000;Williams et al., 2000;Kato et al., 2011). They are known to have potent inhibitory activity against the target glycosidases. Therefore, it would be relevant to consider their thermodynamic properties for binding to the target enzymes, especially the solvation entropy change, in the rational design of novel inhibitors with improved properties.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
MO, TF, and TO designed research; MO performed experiments; MO, TF, and TO analyzed the data; and TO wrote the paper.

FUNDING
This study was supported in part by a grant from Agricultural Technology and Innovation Research Institute (ATIRI), Kindai University. GetArea 1.1 software was used for calculating the ASAs from the crystal structures of HEWL (PDB ID: 1lzd), GN 3 M-liganded HEWL (PDB ID: 4hp0) and (GlcNAc) 3 -liganded HEWL (PDB ID: 1lzb). This program separates the solvent accessible surface area (ASA total ) into the apolar (ASA apolar ) and polar (ASA polar ) components.