14093
2008, 112, 14093–14095 Published on Web 10/17/2008
Rationalization of the Difference in Lifetime of Two Covalent Sialosyl-Enzyme Intermediates of Trypanosoma rangeli Sialidase Laura L. Parker,†,§ Ying-Hua Chung,†,⊥ Claudio J. Margulis,*,† and Jan H. Jensen*,‡ Department of Chemistry, UniVersity of Iowa, Iowa City, Iowa 52242, and Department of Chemistry, UniVersity of Copenhagen, UniVersitetsparken 5, 2100 Copenhagen, Denmark ReceiVed: May 15, 2008; ReVised Manuscript ReceiVed: August 27, 2008
The difference in lifetime with respect to hydrolysis of two covalent syalosyl-enzyme intermediates of two difluorinated sialic acid analogues (1 and 2) bound to Trypanosoma rangeli sialidase is rationalized based on quantum mechanical calculations. The two intermediates differ only in a single functional group, acetamide in the sialidase-1 complex and hydroxyl in the sialidase-2 complex. It is shown that the acetamide group, which is also present in the natural substrate, increases the pKa of a catalytic base (Asp60) through electrostatic repulsion with the carbonyl oxygen on the ligand. This oxygen is absent in 2, resulting in a less basic Asp60 residue and, hence, a longer lifetime of the silaidase-2 complex. Presumably, the lifetime of a sialidase inhibitor complex could be increased further by substituents that stabilize the negative charge on (and lowers the pKa value of) Asp60 in T. rangeli sialidase. Introduction
SCHEME 1: Difluorinated Sialic Acid Analogue
Trypanosama are a genus of parasitic protozoa that cause diseases such as sleeping sickness and Chagas disease in humans. These parasites contain sialidases on their surface, which transfer the monosaccharide sialic acid from host cells to their own surface glycoproteins. This appears to further cell invasion and evasion of the immune system. Thus, inhibitors of sialidases present potential drugs in the fight again diseases such as Chagas diseases. Withers, Buschiazzo, and co-workers1 have recently developed a difluorinated sialic acid analogue (1, Scheme 1) that forms a covalent sialyl-enzyme intermediate with T. rangeli sialidase. However, the sialyl-enzyme bond is susceptible to hydrolysis, and the lifetime of the complex is too short (ca. 5 min) for 1 to be a viable drug candidate. The lifetime of the intermediate can be lengthened by replacing the acetamide group in 1 by a hydroxyl group (2). However, it is far from clear why a substitution at this remote position should affect the hydrolysis rate, which was therefore attributed to “a combination of several electronic, structural, and entropic factors, which are to be the subject of future research.”1 Here, we provide a simple structural rationale for the observed difference in the lifetimes of the two intermediates. Computational Methods The structural models used were derived from the 1.95 Å X-ray structure of Trypanosoma rangeli sialidase in complex with 1 (2A75) and protonated with PDB2PQR.2 The models shown in Figure 1 include the fluorosialic acid and the water * To whom correspondence should be addressed. E-amil:
[email protected] (C.J.M.);
[email protected] (J.H.J.). † University of Iowa. ‡ University of Copenhagen. § E-mail:
[email protected]. ⊥ E-mail:
[email protected].
10.1021/jp804314x CCC: $40.75
and surrounding side chains involved in the hydrolysis reaction. Nearby, charged side chains are also included. Only the position of the fluorosialic acid and the side chains involved in hydrolysis are energy-minimized (the R carbons of the side chains remain fixed; see Figure 1 for more details). Constrained geometry and single-point energy calculations were performed with the GAMESS program3 in the gas phase at the B3LYP/6-31G(d) level of theory. Results and Discussion The energy of the structural model of sialidase complexed with 1 in Figure 1a is 21 kcal/mol higher than that of the corresponding model of the hydrolyzed product in Figure 1b. The corresponding energy difference for the sialidase-2 complex is only 15 kcal/mol. On the basis of the linear freeenergy relationship4 (Figure 2), the activation energy for hydrolysis for the sialidase-1 complex should be ∼2-3 kcal/ mol lower than that for sialidase-2, which is consistent with the faster rate of hydrolysis observed for the sialidase-1 complex. The measured kcat for hydrolysis of sialidase-2 is 0.019 min-1, while the corresponding kcat for sialidase-1 is too fast to be measured since the sialidase-1 complex has regained 95% of its catalytic activity after only 5 min. However, the latter result indicates that the half-life of the sialidase-1 complex is 2008 American Chemical Society
14094 J. Phys. Chem. B, Vol. 112, No. 45, 2008
Figure 1. (a) Structural model of the sialidase-1 covalent complex. The dark bonds and atoms indicate regions of the model that are frozen during the energy minimization process. All but the catalytically important hydrogen atoms have been removed for clarity. (b) The corresponding structural model for the product of the hydrolysis reaction. The geometrical constraints are the same as those in (a) except that the oxygen from the catalytic water molecule is allowed to move. The distance between oxygen atoms in Asp60 and the acetamide group is indicated. The corresponding structures for sialidase-2 are constructed from these models and look virtually identical.
Letters
Figure 3. Simplified structural models of the sialidase-1 covalent complex used for analysis of the energy differences. The coordinates are taken from the large models such the ones shown in Figure 2 and are not reminimized.
TABLE 1: Difference in Energy between the Sialidase-Inhibitor Complex and the System after Hydrolysis of the Sialidase-Inhibitor Bond structure
R
∆E (kcal/mol)
1 2 3 4 5
NHac OH CH3 NH2 NH3+
-20.9 -15.3 -14.9 -13.9 -7.9a
a Since geometry optimization with the NH3+ substituent leads to proton transfer to Asp97, the energy difference for NH3+ is calculated from a single-point energy calculation on optimized NH2 structure 4 with R ) NH3+ (the NH3+ structure used from the CH3NH3+ optimization) replacing R ) NH2.
Figure 2. Schematic representaition of the linear free-energy relationship, which states that the difference in activation energy (∆∆E‡) is a linear function of the difference in exothermicity (∆∆E0), ∆∆E‡ ≈ k∆∆E0, where 0 e k e 1 is the progress of the reaction. Since both reactions are exothermic, it is reasonable to assume that k e 0.5 and that a reasonable estimate of ∆E‡ is 2-3 kcal/mol.
no longer than 1 min, suggesting a kcat of at least 0.7 min-1 and a decrease in the activation free energy of at least 2 kcal/ mol (using transition-state theory) relative to those of the silalidase-2 complex, which is consistent with our estimate. Thus, the experimental observation can be explained by differences in the potential energy surfaces, indicating that entropic effects are not the main source of the difference in the estimated barrier height. To determine the origin of the difference in the estimated barrier height, we recalculated the difference in the relative energy of the reactants and products using two simpler structural models. For the large model shown in Figure 1, the difference in exothermicity is 6 kcal/mol. When this model is reduced to one that contains only the sugar, catalytic water, Tyr343, and Asp60 (Figure 3a), the energy difference is 9.8 kcal/mol, indicating that the source of the energy difference is retained in this simpler model. Subsequent removal of Asp60 (Figure 3b) reduces the energy difference to 0.5 kcal/mol, indicating that the source of the energy difference is not a difference in inductive (or electronic) effects due to the different ligands. Rather, the energy difference must come from a difference in interaction between the ligands and Asp60. In the sialidase-1 X-ray structure (2A75), the distance between one of the carboxyl oxygen atoms in Asp60 and the
carbonyl oxygen of the acetamide group is 3.6 Å, while the corresponding distance to the hydroxide oxygen in the sialidase-2 structure (2AGS) is 4.9 Å. The corresponding distances in the predicted models used in this study are 3.4 and 3.9 Å, respectively, and the corresponding distances in the product structures are 3.6 and 4.4 Å. Presumably, there is a repulsive electrostatic interaction between one of the carboxyl oxygen atoms in Asp60 and the carbonyl oxygen of the acetamide group in 1 that is absent in 2. Indeed, the interaction energy between acetamide and ethanoate arranged in the same orientation as the sugar substituent and Asp60, respectively, in Figure 1a is +3.2 kcal/mol. The corresponding value for acetamide and ethanoic acid is -0.6 kcal/mol, indicating that this repulsive interaction is decreased as the reaction proceeds primarily because the negative charge of Asp60 is neutralized and secondarily because the inter-oxygen distance increases slightly. In effect, the acetamide group, which is also present in the natural substrate, increases the pKa of the Asp60 and thus contributes to catalysis. Presumably, the lifetime of a sialidase inhibitor complex could be increased further by substituents that stabilize the negative charge on (and lowers the pKa value of) Asp60 in T. rangeli sialidase. This hypothesis is supported with three additional calculations. The difference in energy between the sialidase-inhibitor complex and the corresponding hydrolyzed product is reduced further (and the activation energy to hydrolysis is presumably increased) when the acetamide group (1) is replaced with functional groups having neutral or positive electrostatic interactions with Asp60 (3, 4, and 5). Table 1 shows that the most positive group, NH3+, exhibits the smallest energy difference between the bound inhibitor and its hydrolyzed product compared to NH2 and CH3.
Letters Acknowledgment. J.H.J. acknowledges support from the NSF (MCB 209941) and the Danish Research Agency (through a Skou Fellowship). C.J.M. acknowledges support from the Roy J. Carver Charitable Trust (05-2182). References and Notes (1) Watts, A. G.; Oppezzo, P.; Withers, S. G.; Alzari, P. M.; Buschiazzo, A. J. Biol. Chem. 2006, 281, 4149–4155.
J. Phys. Chem. B, Vol. 112, No. 45, 2008 14095 (2) Dolinsky, T. J.; Nielsen, J. E.; McCammon, J. A.; Baker, N. A. Nucleic Acids Res. 2004, 32, W665-W667. (3) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S.; Windus, T. L.; Dupuis, M.; Montgomery, J. A., Jr. J. Comput. Chem. 1993, 14, 1347–1363. (4) Jencks, W. P. Catalysis in Chemistry and Biology; Dover Publications, Inc.: New York, 1987.
JP804314X
