ARTICLE pubs.acs.org/JPCB
Theoretical Investigation on Reaction of Sulbactam with Wild-Type SHV-1 β-Lactamase: Acylation, Tautomerization, and Deacylation Rui Li,† Jun-Min Liao,† Chi-Ruei Gu,† Yeng-Tseng Wang,†,‡ and Cheng-Lung Chen*,† † ‡
The Department of Chemistry, National Sun Yat-Sen University, Kaohsiung 80424, Taiwan National Center for High-Performance Computing, Hsin-Shi, Tainan County, Taiwan ABSTRACT:
Molecular dynamics (MD) simulation and quantum mechanical (QM) calculations were used to investigate the reaction mechanism of sulbactam with class A wild-type SHV-1 β-lactamase including acylation, tautomerization, and deacylation. Five different sulbactam enzyme configurations were investigated by MD simulations. In the acylation step, we found that Glu166 cannot activate Ser70 directly for attacking on the carbonyl carbon, and Lys73 would participate in the reaction acting as a relay. Additionally, we found that sulbactam carboxyl can also act as a general base. QM calculations were performed on the formation mechanism of linear intermediates. We suggest that both imine and trans-enamine intermediates can be obtained in the opening of a five-membered thiazolidine ring. By MD simulation, we found that imine intermediate can exist in two conformations, which can generate subsequent trans- and cis-enamine intermediates, respectively. The QM calculations revealed that trans-enamine intermediate is much more stable than other intermediates. The deacylation mechanism of three linear intermediates (imine, trans-enamine, cis-enamine) was investigated separately. It is remarkably noted that, in cis-enamine intermediate, Glu166 cannot activate water for attacking on the carbonyl carbon directly. This leads to a decreasing of the deacylation rate of cis-enamine. These findings will be potentially useful in the development of new inhibitors.
1. INTRODUCTION For decades now, β-lactam antibiotics have been one of most effective weapons against bacterial infections.1,2 However, using β-lactams in the clinic treatment is constantly resisted by the appearance of β-lactamases.3,4 These β-lactamases can be divided into four classes (A D) on the basis of the activity site differences.5 Those of class B β-lactamases are metalloenzymes, whereas classes A, C, and D are serine enzymes. The most commonly encountered β-lactamases in Escherichia coli and Klebsiella pneumoniae (TEM and SHV, respectively) are class A enzymes.6,7 In order to overcome class A β-lactamase-mediated resistance and preserve the efficacy of antibiotics in medical treatment, β-lactamase inhibitors (clavulanate, sulbactam, and tazobactam) were used. The mechanism of class A β-lactamase inactivation by inhibitors has been studied extensively by means of enzyme kinetics, r 2011 American Chemical Society
X-ray, and Raman crystallography.8 17 These studies suggested a common mechanism as follows. In the first step, the catalytic serine (Ser70) attacks the carbonyl carbon of the β-lactam ring, forming an acyl-enzyme intermediate (AEI) with the opening of the β-lactam ring. In the second step, AEI undergoes further reaction to generate a linear imine intermediate with five-membered ring-opening. Lastly, a series of proton transfers can occur in which the imine AEI tautomerizes to yield a more stable cis- or trans-enamine intermediate and thus providing transient inhibition. The acylation of class A β-lactamases with β-lactam compound such as benzylpenicillin have been studied intensively by Received: December 6, 2010 Revised: July 23, 2011 Published: July 28, 2011 10298
dx.doi.org/10.1021/jp111572v | J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B
ARTICLE
Figure 1. The reaction path and MD models.
experimental and theoretical methods.18 26 The active-site residues such as Ser70, Glu166, Lys73, and Ser130 have been found to play important roles in the acylation reaction. However, except for the nucleophilic reaction of hydroxyl group on Ser70, the exact mechanisms of other active-site residues are still unclear. On the other hand, the mechanism of AEI deacylation is well understood, and experiments have confirmed that Glu166 acts as general base, which is essential for efficient deacylation.27 30 Experimental studies have utilized the deacylation deficient enzyme E166A to trap its AEIs for crystallographic analysis.14 17 In E166A, the active-site Glu166 was replaced by Ala166 and thus prevented regeneration of active enzyme. The accumulation of AEIs facilitates structure detection. Experimental evidence showed that all three inhibitors formed mainly trans-enamine intermediates in E166A β-lactamases.14 17 In addition to E166A acyl-enzyme, Kalp et al. also analyzed the intermediate of wild-type SHV-1 acyl-enzyme. Their study of sulbactam interacted with wild-type SHV-1 showed larger population of cis-enamine intermediate than that of E166A variant.31 This result suggested that laboratory mutants, such as E166A, may provide information on intermediates that is different from wild-type enzyme, and therefore may not be a reliable basis for structure-based drug design. In the present work, molecular modeling methodologies were employed to study the entire reaction of sulbactam (Sul) with wildtype SHV-1 β-lactamase, which includes three steps: acylation, tautomerization, and deacylation. Initially five sulbactam-enzyme models (1 5) were constructed, and subsequent molecular dynamics (MD) simulations were carried out. Figure 1 shows five AEI
models: preacylation form, five-membered ring form, imine form, trans-enamine form, and cis-enamine form, respectively. The structural variations of active sites are analyzed base on MD trajectories. In addition to MD simulations, quantum mechanical (QM) density functional theory (DFT) calculations and ONIOM method were performed for the mechanistic study on the formation of linear intermediates and their stabilities. For three different linear intermediates, subsequent deacylation reaction was investigated, and different mechanisms between cis-enamine and other intermediates were proposed. All of these theoretical results have provided further insights into the reaction processes of sulbactam with SHV-1 β-lactamases and will be potentially useful in the development of new inhibitors.
2. METHODS 2.1. Construction of the Sulbactam Enzyme Model. The initial protein structure of the β-lactamase was adopted from the crystal structural data 1SHV32 of the Protein Data Bank (PDB). To construct the structure of enzyme-drug complex from 1SHV, molecule docking was performed. LibDock Module of Discovery Studio 2.5 was employed, and the active site side chain residues (Ser70, Lys73, Ser130, Asn132, Glu166, Gly236, and Ala237) were defined as the binding site sphere. The ligands with optimized structures were docked in the active site of the enzyme. The generated conformations were manually analyzed, and the one corresponding to the lowest energy was chosen for the initial preacylation structure of the subsequent MD run. We found that the distance 10299
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B SulO8—Ala237N was 2.93 Å, and SulO8—Ser70N was 2.91 Å in the complex. This indicates that two hydrogen bonds were generated, and SulO8 was trapped by the “oxyanion hole”.33 2.2. MD Simulation of Five Models. According to the reaction sequence, MD simulations were performed in three stages. On the basis of the docking result, an MD simulation of 1600 ps was performed first on the unbonded prereactive Michaelis complex (1). After this, complex 2 was constructed, in which the C in the CdO of β-lactam ring was bonded with the Oγ in Ser70. An MD simulation of 1600 ps was then performed on the model 2 system. Then, three linear AEI intermediates (3, 4, and 5), which were derived from the structure of 2 by breaking the C S bond of the five-membered ring, were constructed. An MD simulation of 1600 ps was performed for each of the systems with different linear intermediates. In each of the computed system, the complex was surrounded by a periodic box containing TIP3P water molecules that were extended 8 Å from the protein. Two Na+ counterions were placed in the simulation box by the LEaP module of the AMBER 10 computational program package.34 Atomic charges for sulbactam and substrate-bound Ser70 were determined using RESP module. The electrostatic potential at points selected according to the Merz Shigh Kollman scheme for use in the RESP module was calculated at the RHF/6-31G** level by the Gaussian 03 software program.35 Force field parameters for the protein were assigned from the “parm99” set of parameters, while the parameters of the substrate were obtained from the “gaff” parameters within AMBER 10. At the start of each of the MD run, the energy minimization was carried out first. The steepest descent method followed by the conjugate gradient method was performed for the minimization. After that, MD simulation with position-restrain, which restrained the atomic positions of the macromolecule while allowing the solvent to move, was carried out for 20 ps. Finally, a 1.6 ns MD production run was performed for the system. In the MD process, the SHAKE procedure was applied to constrain all bonds involving hydrogen atoms. The Langevin dynamics36 was used to control the temperature at 300 K using a collision frequency of 1.0 ps 1. Isotropic position scaling was used to maintain the pressure at 1 atm and a relaxation time of 2 ps was used. Periodic boundary conditions were used with a particle mesh Ewald (PME)37 implementation of the Ewald sum for the description of long-range electrostatic interactions. A cutoff of 10 Å was used for other nonbonded interactions. 2.3. Quantum Chemical Calculation. In order to study the proton transfer process, DFT quantum chemical calculations were performed. The geometries of reactants, intermediates, transition states, and products were fully optimized at the B3LYP/6-31G** level.38,39 All energies were calculated including zero-point vibrational energy (ZPE). Intrinsic reaction coordinate (IRC)40calculations were carried out to confirm the reaction paths on the potential energy surface connecting the different reactants, transition states, and products. Considering the effect of the protein environment, two-layer ONIOM calculations41 were performed for a series of intermediates, transition states, and product complexes along the reaction pathways. In the ONIOM method, the system is divided into an inner and an outer layer. The entire system is called “real” and is treated with the low level of theory. The inner layer is termed “model” and is treated with both the low and high levels of theory. The total ONIOM energy E(ONIOM)42 is given by E(ONIOM) = E(high, model) + E(low, real) E(low, model). In our calculations, the degraded sulbactam was treated on the inner layer using a high level of theory: B3LYP/6-31G**. The outer layer using the semiempirical PM3 method included eight residues (Ser70, Lys73,
ARTICLE
Table 1. Important Average Distance (in Å) for All Models system
1
2
3
4
5
Asn170Nδ—Glu166Oε1
2.87
2.83
2.93
2.92
3.93
Glu166Oε2—Lys73Nζ
2.75
2.80
2.77
2.74
2.75
Lys73Nζ—Ser130O
2.87
2.82
2.84
2.82
2.81
Ser70Oγ—Lys73Nζ
2.97
3.87
3.49
3.57
5.47 2.90
Lys73Nζ—Asn132Oδ
2.90
2.85
2.86
2.89
Asn132Nδ—Asp104O
2.86
2.93
2.88
2.89
2.91
Ser70N SulO8
2.98
3.11
3.01
3.07
3.03
Ala237 N SulO8
2.89
3.06
2.93
2.89
3.14
ASN170, SER130, Glu166, Ala237, Asn132, Asp104) and eight water molecules around the active site. Hydrogen atoms were added to the truncated residues. In calculations of water-assisted reactions, the catalytic water molecule was included in the inner layer. The initial coordinates of these complexes were obtained from the average structure of MD simulations. In the optimization procedure, the outer layer residues are fixed at their initial positions, because optimization of the outer layer might lead to an unrealistic structure. All calculations were carried out using the Gaussian 03 package of programs.35
3. RESULTS AND DISCUSSION 3.1. Hydrogen Bond Network in Active Site. On the basis of the MD trajectories, we found that several amino acid residues in the active site formed hydrogen bonds with sulbactam or with residues of each other. The catalytically and structurally important interatomic average distances of five models are listed in Table1. In model 1, The substrate sulbactam was surrounded by several conserved residues including Ser70, Ala237, Lys73, Ser130, Glu166, Asn170, Asn132, and Asp104 (Figure 2). Glu166 is known to be the catalytically important residue. The average distance of Asn170Nδ— Glu166Oε1 was 2.87 Å, and the average distance of Glu166Oε2— Lys73Nζ was 2.75 Å. These distances indicate that Oε1 of the Glu166 makes two rigid hydrogen bonds with the side chains of Asn170 and Lys73. In the present study, Lys73 wass located at the center of the hydrogen bond network, and it formed four hydrogen bonds with Glu166, Ser70, Ser130, and Asn132, respectively. This is consistent with previous reported experimental results.26,43 The Asn132 also made contact with the backbone carbonyl group of Asp104, and the average distance of Asn132Nδ—Asp104O was 2.86 Å. The entire hydrogen bond network represents a crisscross as shown in Figure 2. In models 2, 3, and 4, the conformations of sulbactam were different in different systems. Compared to 1, the average distance of Ser70Oγ—Lys73Nζ was increased by about 0.5 0.9 Å in these three systems. However, the other hydrogen bonds were preserved at a fixed distance. Interestingly, there are significant differences between 5 and the other systems. In model 5, the average distance of Asn170Nδ—Glu166Oε1 was 3.93 Å, and the average distance of Ser70Oγ—Lys73Nζ was 5.47 Å. This indicates that two hydrogen bonds were broken in 5. However, other hydrogen bonds were not affected, and represented stable contacts. MD results showed that the carboxyl oxygen of the sulbactam (SulO8) was trapped by an “oxyanion hole” that consists of two NH groups of the main chains in Ser70 and Ala237. The average distance of the two hydrogen bonds was about 3 Å in all five trajectories. This indicates that the two hydrogen bonds are very 10300
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B
ARTICLE
Figure 2. The hydrogen bond network around the active site. 10301
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B
ARTICLE
Figure 3. Change in distances of 1_Wat_1O Glu166Oε2 and 1_Wat_1O Ser70Oγ during MD simulation.
rigid, and play an important role in stabilizing the negative charge on the SulO8 atom. 3.2. The Acylation Mechanism of SHV Enzyme with Sulbactam. Most experimental and theoretical studies have been performed to investigate the acylation mechanism of Class A βlactamase, and a common mechanism has been proposed: A proton is abstracted from Ser70, and subsequently the deprotonated Ser70 approaches the carbonyl group of the β-lactam ring to form a tetrahedral intermediate. However, the proton abstraction mechanism is still unclear, and a number of proposed mechanisms have been reported. A widely accepted hypothesis suggests that the conserved Glu166 residue is the general base that accepts a proton from Ser70.18,29,44 We have checked the MD trajectory of model 1, and
found that the average distance of Glu166Oε2—Ser70Oγ was about 6 Å, so that a direct abstraction of the proton might be impossible. Previous experimental and theoretical studies on TEM β-lactamase22,45 suggested a mechanism that Glu166 abstracts a proton from Ser70Oγ via a water molecule. In our simulation on SHV-1 β-lactamase, we found that there was only one conserved water 1_Wat_1 anchored by Glu166 (Figure 3), and the average distance of 1_Wat_1O Glu166Oε2 was about 2.6 Å. However, the average distance of 1_Wat_1O Ser70Oγ was about 6.5 Å. This indicated that 1_Wat_1 was not located between Glu166 and Ser70, so it is impossible for this water molecule to mediate proton transfer. It was also proposed that Lys73 acts as the general base.46,47 However, some theoretical and experimental investigations 10302
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B
ARTICLE
Figure 4. Suggested reaction path for the formation of linear intermediates.
indicate that Lys73 is more likely to be protonated under physiological conditions,30,48 which rules out the possibility that Lys73 acts as the general base. In the present study, the H-bond network of Glu166 Lys73 Ser70 was quite stable from our MD simulation (Figure 2). The average distance of Glu166Oε2 Lys73Nζ and Ser70Oγ Lys73Nζ in model 1 indicates that a couple of proton transfers from Ser70 to Lys73 and from Lys73 to Glu166 are fairly possible. Therefore, we suggested that Lys73 can act as a relay station, and Glu166 can abstract a proton from Ser70 via the amino group of Lys73. Experiments also show that Glu166 and Lys73 are essential for efficient acylation. Lietz et al.49 have reported that mutation of Lys73 to alanine in β-lactamase resulted in a substantial reduction in catalytic efficiency. For E166 mutants, significant decreases in the acylation rate were also observed.28,50,51 Díaz et al. studied benzylpenicillin hydrolysis catalyzed by the TEM1 β-lactamase.21 They proposed a substrate-assisted mechanism in which benzylpenicillin carboxyl oxygen abstracts a proton from Ser70 via a hydroxyl group in the Ser130 side chain. In our MD simulation of model 1, the average distance of Ser130Oγ SulO9 was 2.78 Å, which indicates a hydrogen bond was formed. On the other hand, the average distance of Ser70Oγ Ser130Oγ was 3.65 Å, which is beyond the hydrogen bond range. Therefore, the Ser70Oγ Ser130Oγ proton transfer is hard to occur because of the far distance. However, the average distance of Ser70Oγ SulO9 was 2.79 Å. This result suggests that sulbactam carboxyl can abstract a proton from Ser70Oγ easily, viz., it can act as general base directly.
3.3. The Formation of Linear Intermediates. After the ser70 hydroxyl oxygen bounded to SulC7, the β-lactam ring was opened, and a five-membered ring intermediate was formed. This intermediate is thought to be the initial intermediate of the multistep inhibitory process employed by mechanism-based inhibitors. Previous studies suggested five-membered ring intermediate can undergo further reaction to generate a linear imine species by opening the sulfone ring in which the S1 moved away from the C5.14 17 This imine species may rearrange to form cisor trans-enamine. The possible reaction route is the five-membered ring intermediate undergoing an endocyclic cleavage of the S1 C5 bond, accompanied by an intramolecular proton transfer. We propose here that the five-membered ring intermediate can generate imine by proton transfer from N4 to O11, as well as generate trans-enamine by proton transfer from C6 to O11. Then the imine can tautomerize to form trans- and cis-enamine via proton transfer from N4 to C6 (Figure 4). Some significant structural features of model 2 have been observed from MD simulation. Compared to 1, the H-bond network in 2 was preserved except for Ser130Oγ SulO9. In Figure 5, the MD trajectories of model 2 showed that the distance between Ser130Oγ and SulO9 was suddenly increased in the time range 550 600 ps, while the distance between Ser130Oγ and SulO11 was reduced to 2.8 Å. This indicates that Ser130Oγ made a new hydrogen bond with sulfone oxygen O11. Figure 5 also shows that the dihedral angle N4 C5 C6 C7 changed 10303
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B
ARTICLE
Figure 5. Changes in distances of Ser130Oγ SulO9 and Ser130Oγ SulO11 and dihedral angle N4 C5 C6 C7 during MD simulation.
from ca. 30° to ∼120° within 550 600 ps. This dihedral angle change corresponds to the rotation of the five-membered ring around the C5 C6 bond. The rotation changed the cisN4 C5 C6 C7 to the trans form. Thus, a trans-enamine can be obtained by proton transfer from C6 to O11. To further verify this mechanism, we performed QM calculation on this reaction (Figure 4). The initial structure of 2_qm was derived from MD simulation, and the truncated residue Ser70 was replaced by a hydrogen atom. The transition state of the trans-enamine formation reaction (ts24) was obtained, and the calculated energy barrier was 14.68 kcal/mol. The corresponding reaction barrier was calculated to be 11.76 kcal/mol via the ONIOM method. We also calculated the transition state (ts23) of the imine formation reaction involving N4 O11 proton transfer, and the energy barrier was 13.26 kcal/mol. The corresponding reaction barrier was calculated to be 10.50 kcal/mol via the ONIOM method. Compared to the trans-enamine formation, the imine formation is slightly favorable from an energy point of
view. However, the small difference between the two calculated energy barriers also indicates that forming both imine and transenamine is possible in the opening of the five-membered thiazolidine ring. Three conformations of linear intermediates were analyzed from MD trajectories. We found that the dihedral angle N4 C5 C6 C7 was stable in models 4 and 5, which indicates that both trans- and cis-enamine conformations were preserved along the simulations. In 3, the dihedral angle varied from ∼30° to ca. 150° (Figure 6), which indicates the imine conformation rotating around the C5 C6 bond during the simulation. The result indicates that if the imine represents the trans conformation (dihedral angle N4 C5 C6-C7 180°), trans-enamine can be generated by proton transfer from C5 to N4. We performed QM calculations to check this proton transfer reaction. The imine reactant was 3_trans_qm (Figure 4). The transition state ts34 was obtained by geometry optimization, and the calculated energy barrier was 41.20 kcal/mol. On the other hand, if the 10304
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B
ARTICLE
Figure 6. The distribution of dihedral angle N4 C5 C6 C7 along MD simulations of model (A) 3, (B) 4, and (C) 5.
imine represents cis conformation 3_cis_qm (dihedral angle N4 C5 C6 C7 = 30°), a cis-enamine 5_qm can be generated by proton transfer from C5 to N4. For this reaction, the calculated transition state was ts35, and the calculated energy barrier was 42.24 kcal/mol. In the ONIOM calculations, the
calculated energy barriers were 37.26 and 38.87 kcal/mol respectively. Compared to DFT calculations, the ONIOM energy barriers are lower. However, the calculated energy barriers for both tautomerization reactions are so high that the reactions are difficult to proceed. In the real case, there are water molecules 10305
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B
ARTICLE
around the substrate, which may assist the proton transfer. Therefore, we investigated these possibilities by introducing a water molecule in the neighborhood of the imine intermediate. The determined transition states ts34w and ts35w are shown in Figure 4. The calculated energy barriers for the water-assisted tautomerizations are 22.95 and 25.81 kcal/mol, respectively. In Table 2. Energy of the Linear Intermediates from Quantum Chemical Calculationa B3lyp/6-31 g** structure
a
E+ZPE
rel energy
ONIOM E+ZPE
rel energy
3_trans_qm
1216.16893
0
1217.61210
0
3_cis_qm
1216.16762
0.82
1217.61013
1.24
4_qm
1216.17811
5.76
1217.63612
15.07
5_qm
1216.17504
3.83
1217.62314
6.92
E and ZPE are in hartrees and relative energies are in kcal/mol.
the ONIOM calculations, the calculated energy barriers are 18.44 and 21.57 kcal/mol, respectively. This indicates that the tautomerization reactions may occur due to the presence of water as a catalyst. We also calculated the tautomerization from trans-imine (3_trans_qm) to cis-imine (3_cis_qm). The process involved in the C5 C6 bond rotation and the transition state was ts3. The energy barriers for this process are very low, being 2.32 and 3.68 kcal/mol in DFT and ONIOM calculations, respectively. In previous experimental studies,14 17 E166A mutant (deacylation deficient) β-lactamase was used to monitor the formation of AEIs. In E166A β-lactamase, glutamic acid residue is replaced by alanine, which cannot activate a water molecule for nucleophilic attacking on the acyl-enzyme. Therefore, the subsequent deacylation is not activated, and AEIs can be detected easily by spectroscopic or crystallographic methods. These experimental data showed that the trans-enamine was the dominant intermediate, and suggested that the imine and cis-enamine acyl-enzyme were less stable and thus tautomerized to more stable trans-enamine. In our MD simulation, we observed that the sulfone group forms an
Figure 7. Distribution of water between Glu166Oε2 and SulC7 and changes in distances during model 3 MD simulation. 10306
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B
ARTICLE
Figure 8. Distribution of water between Glu166Oε2 and SulC7 and changes in distances during model 4 MD simulation.
intramolecular hydrogen bond with its N4 atom in enamine-enzyme intermediates, and it did not appear in imine-enzyme. This result is in agreement with the previous experimental study of Padayatti et al.16 Additionally, we also observed that Ser130 formed a hydrogen bond with the sulfone oxygen in trans-enamine-enzyme intermediates. This interaction was observed in 2 and 4 but not in 3 and 5. Therefore we concluded this hydrogen bond formation can stabilize the trans-enamine intermediate. Table 2 shows the calculated relative energies of the four linear intermediates (3_trans_qm, 3_cis_qm, 4_qm, and 5_qm). The results indicated that the transenamine is the most stable conformation among the linear intermediates. 3.4. Deacylation Mechanism of AEIs. Reported experimental and theoretical studies have indicated the Glu166 plays an important role in the deacylation process in the wild-type SHV-1 βlactamase. A widely accepted mechanism has been proposed: Glu166 activates a water molecule to attack the carbonyl carbon
and forms a tetrahedral intermediate. After that, the acyl-enzyme bond is broken, and the leaving Ser70 is protonated by means of proton transfer from Glu166. Herein, we investigated the deacylation mechanism of three linear intermediates by analyzing the MD trajectories. In 3, the MD trajectories showed that two water molecules were within a distance of 5 Å to the atom of Glu166 (Figure 7). One of water molecules 3_WAT_1 was fixed at a position between Glu166 and SulC7 during the simulation, and the average distances of 3_WAT_1O Glu166Oε2 and 3_WAT_1O SulC7 were 3.26 Å and 3.66 Å, respectively. This suggests that Glu166 can abstract a proton from 3_WAT_1, and the attack of this activated water molecule on the carbonyl carbon atom of the β-lactam is highly possible. In 4, we observed there were three water molecules within a distance of 5 Å to the atom of Glu166, and these water molecules were also in close distance to the C7 atom (Figure 8). However, only 4_WAT_1 was fixed at a position between Glu166 10307
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B
ARTICLE
Figure 9. Distribution of water between Glu166Oε2 and SulC7 and changes in distances during model 5 MD simulation.
and C7 after 100 ps simulation, and the average distances of 4_WAT_1O Glu166Oε2 and 4_WAT_1O SulC7 were 3.10 Å and 3.91 Å, respectively. The result suggests that water molecule 4_WAT_1 is activated by Glu166 to hydrolyze the acylenzyme intermediate. In both 3 and 4, the distances of Glu166 Ser70 were very large, which indicates that direct proton transfer from Glu166 to Ser70 in the second step of deaclaytion is impossible. However, in both 3 and 4, the residue Lys73 was located in the neighborhood between Glu166 and Ser70, so we conclude that Lys73 can participate in the deaclyation reaction, which acts as bridge of proton transfer. In 5, we also observed that there were three water molecules that have ever appeared within 5 Å of both the Glu166 atom and the SulC7 atom, but only one water molecule 5_WAT_1 was preserved after 100 ps in the simulation (Figure 9). The average distances of 5_WAT_1O Glu166Oε2 and 5_WAT_1O SulC7 were 3.02 Å
and 4.94 Å, respectively. It is hard for 5_WAT_1 to attack the SulC7 atom directly after being activated by Glu166 because of the long distance, and the second water molecule is required to participate in the reaction. In our MD simulation, we observed that two water molecules, 5_WAT_1 and 5_WAT_2, appeared between Glu166 and SulC7 at the same time (about 800 1050 ps) (Figure 9). It is considered that Glu166 activates 5_WAT_2 for attack on the carbonyl carbon via 5_WAT_1, which acts as a relay of a concerted double proton transfer. However, this water dimer between Glu166 and SulC7 was not stable in our simulation, so it could affect the deacylation rate of the cis-enamine intermediate. Recently, Kalp et al. examined the reaction of sulbactam with wild-type and E166A SHV-1 β-lactamase, respectively.31 Their data show that a significant difference exists between the wild-type and SHV E166A acyl-enzyme populations. In E166A SHV-1, the transenamine population appears to predominate, and cis-enamine forms 10308
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B a small population. However, in the wild-type SHV-1, the population of cis-enamine was increased. Herein, we attempt to elucidate this based on our simulation. In E166A enzyme, all three linear AEIs hardly undergo deacylation because catalytic Glu166 is absent, and the imine and cis-enamine intermediates tend to tautomerize to the more stable trans-enamine intermediates. In wild-type SHV-1 βlactamase, Glu166 can catalyze all AEI deacylation. However, the active site structure of cis-enamine is different from others, in which Glu166 cannot activate water for attack on the carbonyl carbon directly. Thus the deacylation rate of the cis-enamine intermediate is decreased, and it facilitates the accumulation of cis-enamine. Therefore, the population of cis-enamine in wild-type SHV-1 is increased.
4. CONCLUSION MD and QM calculations were performed to clarify the reaction mechanism of sulbactam with wild-type SHV-1 βlactamase including acylation, tautomerization, and deacylation. A new mechanism has been proposed: five-membered ring AEI can first generate imine by proton transfer from N4 to O11, as well as trans-enamine by proton transfer from C6 to O11, and the two reactions are competitive. The MD simulation showed that the imine AEI can exist in two conformations. On this basis, we suggest that trans-enamine and cis-enamine can be obtained by two different imine conformations. By analyzing our MD trajectories, we observed that there were eight residues constructing a cross-shaped hydrogen bond network. Lys73 was located at the center and made a fixed salt bridge with catalytic residue Glu166. Lys73 can play the role of a relay not only in acylation but also in deacylation of AEI. Additionally, sulbactam carboxyl can also participate in acylation reaction acting as a general base. Among the three linear intermediates, trans-enamine is the most stable form. These linear intermediates all can undergo deacylation under Glu166 catalysis. However, the deacylation rate of cisenamine intermediate is decreased because Glu166 cannot activate water molecules for attack on the carbonyl carbon directly and needs the assistance of a water dimer, which was not stable around the active site. This could be responsible for the increased population of cis-enamine in wild-type SHV-1. ’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ REFERENCES (1) Frere, J. M.; Joris, B. C. Rev. Microbiol. 1985, 11, 299. (2) Frere, J. M. Mol. Microbiol. 1995, 16, 385. (3) Davies, J. Science 1994, 264, 375. (4) Page, M. I.; Laws, A. P. Chem. Commun 1998, 1609. (5) Bush, K.; Jacoby, G.; Medeiros, A. Antimicrob. Agents Chemother. 1995, 39, 1211. (6) Knox, J. R. Antimicrob. Agents Chemother. 1995, 39, 2593. (7) Bush, K.; Jacoby, G. J. Antimicrob. Chemother. 1997, 39, 1. (8) Chen, C. C.; Herzberg, O. J. Mol. Biol. 1992, 224, 1103. (9) Brown, R. P. A.; Aplin, R. T.; Schofield, C. J. Biochemistry 1996, 35, 12421. (10) Swaren, P.; Golemi, D.; Cabantous, S.; Bulychev, A.; Maveyraud, L.; Mobashery, S.; Samama, J.-P. Biochemistry 1999, 38, 9570. (11) Sulton, D.; Pagan-Rodriguez, D.; Zhou, X.; Liu, Y.; Hujer, A. M.; Bethel, C. R.; Helfand, M. S.; Thomson, J. M.; Anderson, V. E.; Buynak, J. D.; Ng, L. M.; Bonomo, R. A. J. Biol. Chem. 2005, 280, 35528. (12) Bush, K.; Macalintal, C.; Rasmussen, B. A.; Lee, V. J.; Yang, Y. Antimicrob. Agents Chemother. 1993, 37, 851.
ARTICLE
(13) Imtiaz, U.; Billings, E. M.; Knox, J. R.; Manavathu, E. K.; Lerner, S. A.; Mobashery, S. J. Am. Chem. Soc. 1993, 115, 4435. (14) Helfand, M. S.; Totir, M. A.; Carey, M. P.; Hujer, A. M.; Bonomo, R. A.; Carey, P. R. Biochemistry 2003, 42, 13386. (15) Padayatti, P. S.; Helfand, M. S.; Totir, M. A.; Carey, M. P.; Carey, P. R.; Bonomo, R. A.; van den Akker, F. J. Biol. Chem. 2005, 280, 34900. (16) Padayatti, P. S.; Helfand, M. S.; Totir, M. A.; Carey, M. P.; Hujer, A. M.; Carey, P. R.; Bonomo, R. A.; van den Akker, F. Biochemistry 2004, 43, 843. (17) Totir, M. A.; Helfand, M. S.; Carey, M. P.; Sheri, A.; Buynak, J. D.; Bonomo, R. A.; Carey, P. R. Biochemistry 2007, 46, 8980. (18) Hermann, J. C.; Hensen, C.; Ridder, L.; Mulholland, A. J.; H€ oltje, H.-D. J. Am. Chem. Soc. 2005, 127, 4454. (19) Hermann, J. C.; Ridder, L.; Mulholland, A. J.; H€oltje, H.-D. J. Am. Chem. Soc. 2003, 125, 9590. (20) Fujii, Y.; Okimoto, N.; Hata, M.; Narumi, T.; Yasuoka, K.; Susukita, R.; Suenaga, A.; Futatsugi, N.; Koishi, T.; Furusawa, H.; Kawai, A.; Ebisuzaki, T.; Neya, S.; Hoshino, T. J. Phys. Chem. B 2003, 107, 10274. (21) Díaz, N.; Suarez, D.; Sordo, T. L.; Kenneth M. Merz, J. J. Phys. Chem. B 2001, 105, 11302. (22) Díaz, N.; Sordo, T. L.; Kenneth M. Merz, J.; Suarez, D. J. Am. Chem. Soc. 2003, 125, 672. (23) Herzberg, O. J. Mol. Biol. 1991, 217, 701. (24) Maveyraud, L.; Pratt, R. F.; Samama, J. P. Biochemistry 1998, 37, 2622. (25) Adediran, S. A.; Deraniyagala, S. A.; Xu, Y.; Pratt, R. F. Biochemistry 1996, 35, 3604. (26) Matagne, A.; Lamotte-Brasseur, J.; Frere, J. M. Biochem. J. 1998, 330, 581. (27) Knox, J. R.; Moews, P. C.; Escobar, W. A.; Fink, A. L. Protein Eng. 1993, 6, 11. (28) Escobar, W. A.; Tan, A. K.; Lewis, E. R.; Fink, A. L. Biochemistry 1994, 33, 7619. (29) Vijayakumar, S.; Ravishanker, G.; Pratt, R. F.; Beveridge, D. L. J. Am. Chem. Soc. 1995, 117, 1722. (30) Lamotte-Brasseur, J.; Lounnas, V.; Raquet, X.; Wade, R. C. Protein Sci. 1999, 8, 404. (31) Kalp, M.; Totir, M. A.; Buynak, J. D.; Carey, P. R. J. Am. Chem. Soc. 2009, 131, 2338. (32) Kuzin, A. P.; Nukaga, M.; Nukaga, Y.; Hujer, A. M.; Bonomo, R. A.; Knox, J. R. Biochemistry 1999, 38, 5720. (33) Strydnadka, N. C. J.; Adachi, H.; Jensen, S. E.; Johns, K.; Sielecki, A.; Betzel, C.; Sutoh, K.; James, M. N. J. Nature 1992, 359, 700. (34) Case, D. A.; Darden, T. A.; T.E. Cheatham, I.; Simmerling, C. L.; Wang, J.; Duke, R. E.; Luo, R.; ; Crowley, M.; R. C..Walker; Zhang, W.; Merz, K. M.; B.Wang; Hayik, S.; Roitberg, A.; Seabra, G.; Kolossvary, I.; K. F..Wong; Paesani, F.; Vanicek, J.; X.Wu; Brozell, S. R.; Steinbrecher, T.; Gohlke, H.; Yang, L.; Tan, C.; Mongan, J.; Hornak, V.; Cui, G.; Mathews, D. H.; Seetin, M. G.; Sagui, C.; Babin, V.; Kollman, P. A. AMBER 10; University of California: San Francisco, 2008. (35) Frisch, M. J.; Trucks, G. W., ; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr. , ; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G. N.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03; Gaussian, Inc.: Pittsburgh, PA, 2003. 10309
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
The Journal of Physical Chemistry B
ARTICLE
(36) Pastor, R. W.; Brooks, B. R.; Szabo, A. Mol. Phys. 1988, 65, 1409. (37) Essman, V.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. J. Chem. Phys. 1995, 103, 8577. (38) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (39) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (40) Gonzalez, C.; Schlegel, H. B. J. Phys. Chem. B 1990, 94, 5523. (41) Vreven, T.; Morokuma, K. J. Comput. Chem. 2000, 21, 1419. (42) Humbel, S.; Sieber, S.; Morokuma, K. J. Chem. Phys. 1996, 105, 1959. (43) Jelsch, C.; Mourey, L.; Masson, J. M.; Samama, J. P. Proteins 1993, 16, 364. (44) Gibson, R. M.; Christensen, H.; Waley, S. G. Biochem. J. 1990, 272, 613. (45) Minasov, G.; Wang, X.; Shoichet, B. K. J. Am. Chem. Soc. 2002, 124, 5333. (46) Oefner, C.; Winkler, F. K. Nature 1990, 343, 284. (47) Wladkowski, B. D.; Chenoweth, S. A.; Sanders, J. N.; Krauss, M.; Stevens, W. J. J. Am. Chem. Soc. 1997, 119, 6423. (48) Raquet, X.; Lounnas, V.; Lamotte-Brasseur, J.; Frere, J. M.; Wade, R. Biophys. J. 1997, 73, 2416. (49) Lietz, E. J.; Truher, H.; Kahn, D.; Hokenson, M. J.; Fink, A. L. Biochemistry 2000, 39, 4971. (50) Guillaume, G.; Vanhove, M.; Lamotte-Brasseur, J.; Ledent, P.; Jamin, M.; Joris, B.; Frere, J. M. J. Biol. Chem. 1997, 272, 5438. (51) Escobar, W. A.; Tan, A. K.; Fink, A. L. Biochemistry 1991, 30, 10783.
10310
dx.doi.org/10.1021/jp111572v |J. Phys. Chem. B 2011, 115, 10298–10310
