MM Study - American Chemical Society

indicate that the metal ion in Fe2+-PDF is always pentacoordinated during the reaction process, ... Bacterial peptide deformylase (PDF) represents a n...
0 downloads 0 Views 451KB Size
Download PDF
J. Phys. Chem. B 2007, 111, 6229-6235

6229

Catalytic Mechanism and Metal Specificity of Bacterial Peptide Deformylase: A Density Functional Theory QM/MM Study Chuanyun Xiao and Yingkai Zhang* Department of Chemistry, New York UniVersity, New York, New York 10003 ReceiVed: December 16, 2006; In Final Form: April 5, 2007

Bacterial peptide deformylase (PDF) represents a novel class of mononuclear iron peptidase, and has an intriguing metal preference different from most other metalloproteases. Using a hybrid density functional theory (B3LYP) QM/MM method, we have theoretically investigated its catalytic mechanism and metal specificity by studying both Fe2+-PDF and Zn2+-PDF. In both forms of PDF, the conserved Glu133 residue is protonated in the reactant complex, and acts as a general acid during the reaction. The initial reaction step is the nucleophilic attack of the metal-bound hydroxide on the carbonyl carbon of the substrate. Our calculations indicate that the metal ion in Fe2+-PDF is always pentacoordinated during the reaction process, while that in Zn2+-PDF is only tetrahedrally coordinated and not bound to the substrate in the reactant complex. This difference in their metal coordination is suggested to account for the lower activity of Zn2+-PDF in comparison with Fe2+-PDF.

1. Introduction Bacterial peptide deformylase (PDF) represents a novel class of mononuclear non-heme iron protein to catalyze amide hydrolysis,1,2 and is responsible for the removal of N-terminal formyl group (N-formyl) from nascent polypeptides generated in eubacterial protein synthesis:3,4

Figure 1. Structure of (PATH)M2+(O2CH).61

While most metalloproteases are zinc-dependent, PDF has been shown to be the first example of iron metallopeptidases which utilize a non-heme Fe2+ ion as its catalytic metal.1,2 Intriguingly, the replacement of Fe2+ with Zn2+ in bacterial peptide deformylase leads to the rate reduction by 2-3 orders of magnitude.5,6 Meanwhile, PDF has been an attractive target for the design of broad-spectrum antibacterial drugs.7-11 Structural studies by NMR spectroscopy12,13 and X-ray crystallography2,13-15 have revealed that the metal ion at the active site of PDF in E. coli is coordinated to a cysteine (Cys90) and two histidines (His132 and His136). The fourth ligand has been suggested to be either a water molecule or a hydroxide ion (OH-).2 The two histidine ligands reside in the conserved sequence motif of HEXXH, a characteristic motif found in many zinc metalloproteases represented by thermolysin.16,17 The Glutamate residue (Glu133) in the HEXXH motif is not bound to the metal ion but plays an essential catalytic role. The mutation of Glu133 to alanine abolishes the enzyme activity of PDF.5,18 On the basis of the extensive structural and biochemical studies of PDF and its structural similarity to thermolysin, * Address correspondence to this author. E-mail: [email protected].

Figure 2. The active site structures for the reactant of Fe2+-PDF and Zn2+-PDF determined with B3LYP (SDD, 6-31G*) QM/MM calculations.

several possible mechanisms for the PDF-catalyzed hydrolysis reaction have been proposed by Pei et al.7,14,18,19 and Becker et al.2 The unsettled fundamental mechanistic issues include the following: (I) Is the metal ion bound to a water molecule or a hydroxide ion? (II) Is the formyl group of the substrate directly bound to metal? (III) What is the coordination of the metal: tetrahedral or 5-fold? Does the coordination change during the reaction? (IV) Does the residue Glu133 directly participate in the reaction? What is its protonation state in the reactant? Does it act as the general acid or the proton shuttle during the reaction? Besides the reaction mechanism, a particularly interesting and fundamental question for bacterial peptide deformylase is its metal-dependent reactivity. In contrast to Ni2+ or Co2+, whose substitution for Fe2+ in PDF leads to enzymes with nearly

10.1021/jp068657f CCC: $37.00 © 2007 American Chemical Society Published on Web 05/16/2007

6230 J. Phys. Chem. B, Vol. 111, No. 22, 2007

Xiao and Zhang

Figure 3. Calculated minimum-energy path for Fe2+-PDF and Zn2+-PDF at the B3LYP (SDD, 6-31G*) QM/MM level. The energies at the stationary points are given in the figure. Regions with different reaction coordinates are separated by dashed lines: RC1 ) R(OW-CP); RC2 ) R(HW2-OW) - R(HW2-NP) - R(HW1-OE1); RC3 ) R(HW2-OE2) - R(HW2-NP) + R(HW1-OW) - R(HW1-OE1) + R(NP-CP); RC4 ) R(NP-CP).

TABLE 1: Comparison of the Experimental and Calculated Structures at the B3LYP Level with Three Basis Sets for (PATH)M2+(O2CH) with M ) Zn and Fea M ) Zn 6-311G* SDD LanL2DZ exptl M ) Fe 6-311G* SDD LanL2DZ

M-O1

M-O2

M-N1

M-N2

Zn-S

C-O1-Zn

C-O2-Zn

1.955 1.987 2.140 2.000

2.812 2.637 2.251 2.615

2.096 2.138 2.206 2.054

2.202 2.270 2.332 2.130

2.270 2.260 2.337 2.256

110.4 104.9 90.6 103.7

71.5 74.6 85.9 75.6

2.089 2.092 2.073

2.321 2.354 2.382

2.150 2.177 2.170

2.272 2.314 2.301

2.315 2.303 2.316

93.8 94.4 95.5

84.0 83.2 82.2

a The bond lengths and angles are in Å and deg, respectively. SDD and LanL2DZ denote respectively the SDD and LanL2DZ ECP basis set for the metal atom in combination with the 6-31G* basis set for all other atoms.

unchanged activity,2,6,13,18,20,21 the replacement of Fe2+ with Zn2+ reduces the activity of the enzyme by 2-3 orders of magnitude.1,5,6,20,22 Earlier structural studies revealed that the difference in the crystal structure of different metal forms of PDF is extremely small, with the only difference that Zn2+ might be bound more tightly to PDF.2,20 On the basis of this observation, it was suggested that the lower catalytic activity of the Zn2+ form could result from its difficulty in converting between tetrahedral and 5-fold coordination during the reaction.2 More recently, Jain et al.23 reanalyzed, with an enhanced resolution, the crystal structures of Fe, Co, and Zn forms of E. coli PDF in complex with its deformylation product (formate) they previously obtained, and found a clear difference in the binding mode between the formate and the metal ion. In both Fe2+-PDF and Co2+-PDF, the formate was found to bind to the metal in a bidendate fashion, with both oxygen atoms of the formate lying within the coordinating distance of the metal (2.44 and 2.30 ( 0.17 Å for Fe2+-PDF). In Zn2+-PDF, however, the formate is bound to Zn2+ in a monodendate fashion (2.09 and 2.88 ( 0.17 Å). On the basis of this finding, it has been hypothesized that the low activity of Zn2+-PDF may be due to the lack of activation of the formyl carbonyl by the Zn2+ ion in Zn2+-PDF.23 Subsequently, a combined experimental and theoretical study on the structure of two model complexes of PDF bound to formate suggested that an N,S-ligand Fe2+ ion prefers a coordination number >4 while a Zn2+ ion in the same environment does not show such a geometric preference.24 It should be noted that, in both studies,23,24 the bound ligand is the formate ion, which is the product, rather than the formyl group as in the PDF substrate. So it is unclear whether the metal coordination mode in Zn2+-PDF would also be different from that in Fe2+-PDF in the reactant complex, and how the coordination changes during the reaction. In contrast to the extensive experimental efforts, only a few theoretical studies have been performed on PDF. Madison et

al.25 explored the binding affinities and geometries of Fe2+PDF in complex with various inhibitors by a combined quantum mechanical/molecular mechanical (QM/MM) method. Very recently, density functional calculations on PDF model complexes26 have been carried out. In this model complex study, the size of the system is limited, and the enzyme environment cannot be simulated. The calculated results revealed no binding mode difference between Fe2+-PDF and Zn2+-PDF, in all stationary states including the product complex; and the reaction barrier for Zn2+-PDF is even a little lower than that of Fe2+PDF with use of the largest model complex.26 These findings26 are in disagreement with the available experimental results.1,5,6,20,22,23 Here we have carried out density functional theory QM/MM calculations to investigate the formyl-peptide hydrolysis reaction catalyzed by both Fe2+ and Zn2+ forms of PDF. With the QM/MM approach,27-36 the whole enzyme complex is simulated. The atoms in the enzyme active site, which consist of the metal ion, coordinating ligands, substrate, and residues directly participating in the reaction, are treated by density functional theory, while the rest of the enzyme environment including protein and water molecules is described by a molecular mechanical force field. Our calculations are based on a pseudobond ab initio QM/MM method,37-39 which has been demonstrated to be powerful in the study of several enzymes.40-46 2. Methods Preparation of the Enzyme-Substrate System. In most experimental works on PDF, formyl-Met-Ala-Ser (fMAS) or its shortened form formyl-Met-Ala (fMA) was used as the model substrate.2,5,6,12,20 Considering the fact that the PDF catalytic activity is relatively insensitive to the length of an N-formyl polypeptide, provided that it is composed of at least two residues,21 we have chosen the formyl-Met-Ala (fMA) as the

Bacterial Peptide Deformylase

J. Phys. Chem. B, Vol. 111, No. 22, 2007 6231

Figure 4. Calculated reaction mechanism for Fe2+-PDF.

substrate in our theoretical treatment. The initial structure of M2+-PDF/fMA (M ) Fe, Zn) was prepared on the basis of the crystal structure of the Ni2+-PDF in complex with the product MAS (PDB code 1BS6).2 Since there are three almost identical protein complexes in the pdb file, we have removed two of them by deleting the B, C, E, and F chains. The substrate fMA was constructed by removing the Ser residue from MAS and adding a formyl group. The Ni2+ ion was replaced by an M2+ ion to form M2+-PDF (M ) Fe, Zn). The fourth ligand of the metal center was initially considered to be a water molecule and Glu133 is chosen to be in the deprotonated form. Since the pKa for cysteine in water is 8.3 and its coordination to the metal ion would lower the pKa value, the metal-coordinated Cys90 is assumed to be depronated. Hydrogen atoms were added to heavy atoms by the Tinker program.47 There are four histidine residues (7, 54, 132, and 136) in PDF. According to their local hydrogen network, His7 and His54 are assigned to be doubly protonated since they form salt bridges with Asp162 and Asp52, respectively, while His132 and His136 are assumed to be neutral with Nδ protonated. This enzyme-substrate complex was then

solvated by using a sphere of TIP3P water48 of radius 28 Å centered at the metal center. Water molecules overlapping with the protein atoms are removed from the sphere. The total number of atoms in our enzyme-substrate system is 8182. DFT QM/MM Calculations. The prepared M2+-PDF/fMA (M ) Fe, Zn) enzyme-substrate model described above is partitioned into a QM subsystem and an MM subsystem. The QM subsystem comprises the metal ion, the metal-bound water molecule, the side chains of Cys90, His132, His136, and Glu133, and all atoms of the substrate fMA. The MM subsystem consists of the rest of the protein and water molecules. The boundary between the QM and MM subsystems was treated by the pseudobond approach.37 The total number of QM and pseudobond atoms is 74. With this prepared QM/MM system, first the MM subsystem is equilibrated with minimizations interspersed by short MD simulations, and then an iterative optimization procedure38 is applied to the system with B3LYP QM/MM calculations, leading to an optimized structure for the reactant. A spin singlet is assumed for Zn2+-PDF, while the singlet, triplet, and quintet are considered for Fe2+-PDF. The

6232 J. Phys. Chem. B, Vol. 111, No. 22, 2007

Xiao and Zhang

Figure 5. Optimized structures for the stationary points of Fe2+-PDF around the active site at the B3LYP (SDD, 6-31G*) QM/MM level. Important interatomic distances (in Å) are indicated.

Figure 6. Optimized structures for the stationary points of Zn2+-PDF around the active site at the B3LYP (SDD, 6-31G*) QM/MM level. Important interatomic distances (in Å) are indicated.

ground state for Fe2+-PDF corresponds to the quintet, which is lower than the singlet and triplet by 42 and 29 kcal/mol, respectively, at the B3LYP (SDD, 6-31G*) QM/MM level. Since the energy difference between the quintet and the other two spin states is considerably larger than the energy barrier of the PDF-catalyzed hydrolysis reaction process, the spin cross is not expected to occur, and thus Fe2+-PDF should be in the quintet spin state during the entire reaction process. With the determined reactant structures, we employed the reaction coordinate driving method38,49 to determine the minimum energy path along a proposed reaction coordinate. Provided that the obtained minimum energy path is smooth and continuous, it has been shown that the energy maximum along the path is a good approximation for the transition state.50,51 The nature of the stationary points on the reaction path was confirmed by frequency calculations. The energy maxima with one and only

one imaginary frequency are the transition states (TS), while the energy minima along the path with no imaginary frequencies are characterized as the reactant, the intermediates (INT), or product. All calculations were carried out with a modified version of Gaussian0352 and TINKER programs.47 The QM subsystem is treated by the hybrid B3LYP functional,53-55 while the AMBER 95 all-atom force field56 and the TIP3P water model48 were employed for describing the MM subsystem. The basis set used for geometry optimizations and the frequency calculation of all stationary points along the reaction path is a SDD effective core potential (ECP) basis set for the metal57 in combination with the 6-31G* basis set for all nonmetal QM atoms.58 We denote this computational scheme by B3LYP (SDD, 6-31G*). Pseudobonds were treated with the 6-31G* basis set and its corresponding ECP parameters.37 For the QM subsystem, the

Bacterial Peptide Deformylase convergence criteria for geometry optimizations follow Gaussian03 defaults. For the MM subsystem, the convergence criterion is that the root-mean-square (rms) energy gradient should be less than 0.1 kcal‚mol-1‚Å-1. In the MM minimizations, only atoms within 20 Å of the metal center were allowed to move. No cutoff for nonbonded interactions was used in the QM/MM calculations and the MM minimizations. Choice of the Effective Core Potential and Basis Set for the Metal Atom. In comparison with previous theoretical studies of Fe2+-PDF inhibitor binding25 and PDF model complexes,26 we have employed the same hybrid functional B3LYP and the same basis set 6-31G* for all nonmetal QM atoms. However, for the metal atom, we have employed the SDD ECP basis set57 for the iron and zinc ions instead of the LanL2DZ ECP basis set,59,60 which has been employed previously.25,26 Our choice is based on the test calculations on the structures and metal coordination pattern of (PATH)M2+(O2CH) (M ) Fe, Zn), which is a model complexes of M2+-PDF in complex with formate61 (shown in Figure 1). We have compared our results with both the experimentally determined structure for (PATH)Zn2+(O2CH)61 and the previous B3LYP calculation results with the large 6-311G* basis set for all atoms.24 From Table 1, we can see that for (PATH)Zn2+(O2CH), the calculated geometric parameters using the B3LYP (SDD, 6-31G*) scheme agree very well with the experimental as well as B3LYP (6-311G*) results, while the B3LYP (LanL2DZ,6-31G*) scheme yields a much shorter distance between Zn and the O2 atom of the formate, leading to a bidendate binding mode between Zn and the formate that is inconsistent with the observations in (PATH)Zn2+(O2CH)24,61 and Zn2+-PDF bound to formate.23 For (PATH)Fe2+(O2CH), there are no experimental data for comparison, but the results obtained with both SDD and LanL2DZ agree well, and are very close to those with B3LYP (6-311G*) calculations, which prefers a bidentate binding mode between Fe and the formate.24 These test results also explain why the previous DFT calculations on PDF model complexes26 did not observe the coordination difference between Fe2+-PDF and Zn2+-PDF. 3. Results and Discussion The QM/MM optimized structures for the reactants of Fe2+PDF and Zn2+-PDF are shown in Figure 2. Two important features are worth noticing. First, in both optimized structures, the metal ion is bound to a hydroxide ion (OH-) and the Glu133 residue is protonated. It should be noted that in our prepared initial structures, it is a water molecule that is bound to the metal ion and the Glu133 is assumed to be depronated. This indicates that the metal ion has significantly lowered the pKa of the metal-bound water and thus facilitated the proton transfer from the water to Glu133. Second, Fe2+ and Zn2+ ions have different binding modes with their ligands. In addition to the coordination to the three conserved residues (His132, His136, Cys90), the Fe2+ ion is bound to both the hydroxide ion and the carbonyl group of the substrate formyl-peptide. The bond lengths between Fe2+ and the nearest atoms of its five ligands range from 1.92 to 2.45 Å, which are all within the coordinating distance of the iron ion. Therefore, the Fe2+ ion is pentacoordinated in the reactant. In contrast, the Zn2+ ion is bound only to the three conserved residues and the hydroxide ion, while the distance between Zn and the OP atom of the carbonyl group is as long as 3.45 Å. Therefore, the Zn2+ ion is only tetrahedrally coordinated in the reactant. With the optimized reactant structures of Fe2+-PDF and Zn2+PDF, the hydrolysis reaction mechanisms for both enzymes have

J. Phys. Chem. B, Vol. 111, No. 22, 2007 6233 been characterized. The determined minimum-energy reaction paths are presented in Figure 3. For both paths, we can see that the first reaction step is rate determining, which has a barrier of 14.3 and 17.4 kcal/mol for Fe2+-PDF and Zn2+-PDF, respectively, while the other reaction steps are relatively quite flat or downhill to the product state. The two intermediates (INT1 and INT2) are only metastable. The calculated overall reaction barriers are 15.2 and 17.4 kcal/mol for Fe2+-PDF and Zn2+-PDF, respectively. Our results are consistent with the activation barrier of 13.0 kcal/mol for Fe2+-PDF and 16.416.7 kcal/mol for Zn2+-PDF estimated from the experimental value of kcat6,20 by the simple transition-state theory: kcat ) κ(kBT/h)e-∆G/RT. It should be noted that the calculated energy barriers can be quite dependent on a number of factors, including the initial structure, the QM/MM system partition, the QM method and the basis set, the MM force field, and so on. Thus the calculated 2.2 kcal/mol energy barrier difference between Fe-PDF and Zn-PDF, which seems to be in good agreement with the experimental results, is likely to be fortuitous. In the following, we will first describe the hydrolysis reaction mechanism for Fe2+-PDF, and then make a comparison between Zn2+-PDF and Fe2+-PDF with respect to their reaction mechanisms and structural features. Mechanism of Fe2+-PDF. A scheme of the characterized mechanism for Fe2+-PDF is depicted in Figure 4 and the key structures are presented in Figure 5. The overall reaction mainly consists of three steps. The initial step is the nucleophilic attack of the OW atom of Fe-bound hydroxide toward the CP atom of the carbonyl group. The OW-CP bond is partially formed in the transition state TS1 and fully formed in the intermediate INT1, with the bond length decreasing from 2.61 Å in the reactant to 1.80 Å in TS1 and 1.51 Å in INT1 (Figure 5). Due to the formation of the OW-CP bond, the CP-OP bond is weakened and undergoes a transition from the double bond toward the single bond, with the bond length elongated from 1.25 Å in the reactant to 1.31 Å in TS1 and 1.36 Å in INT1. The resulting oxyanion is stabilized by hydrogen bonds with Gln50 and Leu91 as well as the interaction with the metal ion. The second step of the reaction mainly involves the formation and breaking of the hydrogen bonds, with a quite flat reaction energy curve (Figure 3). In the third step of the reaction, the proton HW2 transfers from the Glu133 to the amide NP atom, and subsequently the amide bond is spontaneously cleaved, leading to the product. Throughout the reaction process, we can see that the Fe2+ ion is always pentacoordinated. In the optimized product structure as shown in Figure 5, the Fe2+ ion is bound to both oxygens of the formate, with the determined Fe-OW and Fe-OP bond lengths being 2.37 and 2.06 Å, respectively, which are in agreement with the recent experimental values of 2.44 ( 0.17 and 2.30 ( 0.17 Å for Fe2+-PDF bound to formate in the crystal structure of the Fe2+-PDF product complex.23 Zn2+-PDF vs Fe2+-PDF. With the same choice of reaction coordinates as Fe2+-PDF, the hydrolysis reaction path for Zn2+PDF is also determined. The optimized structures for the stationary points of Zn2+-PDF around the active site are presented in Figure 6. Although the overall hydrolysis reaction catalyzed by Zn2+-PDF involves three very similar steps, the metal coordination is very different from that of Fe2+-PDF throughout the reaction process. For Zn2+-PDF, the coordination of the Zn2+ ion to the ligands is flexible, and it is always less than five. As described above, the Zn2+ ion is tetrahedrally coordinated to His132, His136, Cys90, and the hydroxide ion, but is not bound to the OP of the carbonyl group in the reactant.

6234 J. Phys. Chem. B, Vol. 111, No. 22, 2007 With the formation of the OW-CP bond in TS1, the Zn2+ ion is brought to coordinate with the OP, but its coordination to His136 is almost lost. In INT1, the distances from the Zn2+ ion to the OW atom (2.56 Å) and to the His136 residue (2.63 Å) are both quite long. In TS2, INT2, and TS3, the Zn2+ ion is bound to the OW atom again, but the Zn2+-His136 distance remains quite long. In the product complex, the Zn2+ ion loses its binding with the OW atom, but it is bound to His136, resulting in a coordination number of four. The determined distances between the Zn2+ ion and the two formate oxygen atoms are 2.61 and 2.00 Å, which are in good agreement with the corresponding distances of 2.88 ( 0.17 and 2.09 ( 0.17 Å in the crystal structure.23 The different coordination modes of Fe2+ and Zn2+ to their ligands in PDF imply different catalytic roles in the hydrolysis reaction. It has been suggested18 that the metal ion in PDF can either act to lower the pKa of its bound water so that the water is ionized or act as a Lewis acid to polarize the carbonyl group so that it becomes more susceptible to nucleophilic attack, or act as both. In addition, the metal ion may play the role of stabilizing the oxyanion that is formed in the transition states and intermediates. In our characterized mechanism, the Fe2+ ion is always bound to five ligands (His132, His136, Cys90, the hydroxide ion, and the carbonyl group) throughout the reaction process and thus it should play three catalytic roles in the Fe2+-PDF catalyzed hydrolysis reaction: ionizing the water molecule by lowering its pKa, activating the carbonyl group in the substrate, and stabilizing the oxyanion in the TS1 and subsequent states. In contrast, the Zn2+ ion is only tetrahedrally coordinated to the first four ligands but not bound to the carbonyl group in the reactant. Thus the Zn2+ ion plays only twofold catalytic roles: ionizing the conserved water molecule and stabilizing the oxyanion. The lack in the activation of the carbonyl group by the Zn2+ can be suggested to be a key reason for the dramatically reduced activity of Zn2+-PDF in comparison with Fe2+-PDF. Our results support the previous hypothesis that the metal coordination difference may be responsible for the PDF’s reactivity difference based on the finding of the different formate binding modes with Fe2+ or Zn2+ in the structures of the peptide hydrolysis product.23 The protonation state and catalytic function of the Glu133 residue in PDF has been examined extensively by experimental studies.5,18,19,62 It was suggested that Glu133 could serve as a general acid (proton donor)18 or proton shuttle19 during the reaction. Our calculated result indicates that Glu133 is protonated in the reactant of both Fe2+-PDF and Zn2+-PDF, and acts as a general acid by donating its proton to the amide NH group of the substrate to facilitate the cleavage of the amide bond during the reaction. 4. Conclusions In this paper, the catalytic mechanism and metal specificity of peptide deformylase in the peptide hydrolysis reaction has been studied by a hybrid density functional theory (B3LYP) QM/MM method. The SDD ECP basis set for the metal atom in combination with the 6-31G* basis set for all other atoms is employed in our present calculations, and has been shown to provide a reasonable description of both the Fe2+ and Zn2+ coordination in an N,S-ligand environment. The structures of the reactant, transition states, intermediates, and product for Fe2+-PDF and Zn2+-PDF are optimized, and the reaction mechanism and the minimum-energy reaction paths for both enzymes are determined. The metal-bound water is found to be ionized in the reactant of both enzymes, resulting in a

Xiao and Zhang hydroxide ion as the fourth ligand of the metal ion. The conserved Glu133 residue is in a protonated state in the reactant, and acts as a general acid during the reaction. In spite of such similarities, different binding modes are revealed in the two enzymes between the metal ion and its ligands: the Fe2+ ion is pentacoordinated throughout the reaction process, while the Zn2+ ion is only tetrahedrally coordinated and it does not bind to the carbonyl group of the substrate formyl-peptide in the reactant. The distinct binding modes between the metal ion and the formate we found in the products of Fe2+-PDF and Zn2+-PDF agree with the recent experimental findings of Jain et al.23 in corresponding systems. This difference in their metal coordination implies different roles of the metal in this hydrolysis reaction, and is suggested to account for lower activity of Zn2+PDF in comparison with Fe2+-PDF. One main limitation of the current study is that only one initial structure has been employed in the DFT QM/MM study for each form of peptide deformylase. To quantitatively investigate the product specificity of peptide deformylase, more advanced simulations need to be carried out, such as DFT QM/MM studies with multiple initial structures from a MD trajectory42 or ab initio QM/MM molecular dynamics simulations.63 Thus the results presented here should be considered as suggestive rather than conclusive. Acknowledgment. This work has been supported by grants from the National Science Foundation (CHE-CAREER-0448156 and CHE-MRI-0420870). Y.Z. also thanks NYSTAR for a James D. Watson Young Investigator Award and NYU for a Whitehead fellowship. References and Notes (1) Rajagopalan, P. T. R.; Yu, X. C.; Pei, D. H. J. Am. Chem. Soc. 1997, 119, 12418-12419. (2) Becker, A.; Schlichting, I.; Kabsch, W.; Groche, D.; Schultz, S.; Wagner, A. F. V. Nat. Struct. Biol. 1998, 5, 1053-1058. (3) Mazel, D.; Pochet, S.; Marliere, P. EMBO J 1994, 13, 914923. (4) Meinnel, T.; Blanquet, S. J. Bacteriol. 1994, 176, 7387-7390. (5) Meinnel, T.; Lazennec, C.; Blanquet, S. J. Mol. Biol. 1995, 254, 175-183. (6) Ragusa, S.; Blanquet, S.; Meinnel, T. J. Mol. Biol. 1998, 280, 515523. (7) Hao, B.; Gong, W. M.; Rajagopalan, P. T. R.; Zhou, Y.; Pei, D. H.; Chan, M. K. Biochemistry 1999, 38, 4712-4719. (8) Giglione, C.; Pierre, M.; Meinnel, T. Mol. Microbiol. 2000, 36, 1197-1205. (9) Yuan, Z. Y.; Trias, J.; White, R. J. Drug DiscoVery Today 2001, 6, 954-961. (10) Nguyen, K. T.; Hu, X. B.; Colton, C.; Chakrabarti, R.; Zhu, M. X.; Pei, D. H. Biochemistry 2003, 42, 9952-9958. (11) Jain, R.; Chen, D.; White, R. J.; Patel, D. V.; Yuan, Z. Curr. Med. Chem. 2005, 12, 1607-1621. (12) Meinnel, T.; Blanquet, S.; Dardel, F. J. Mol. Biol. 1996, 262, 375386. (13) Dardel, F.; Ragusa, S.; Lazennec, C.; Blanquet, S.; Meinnel, T. J. Mol. Biol. 1998, 280, 501-513. (14) Chan, M. K.; Gong, W. M.; Rajagopalan, P. T. R.; Hao, B.; Tsai, C. M.; Pei, D. H. Biochemistry 1997, 36, 13904-13909. (15) Becker, A.; Schlichting, I.; Kabsch, W.; Schultz, S.; Wagner, A. F. V. J. Biol. Chem. 1998, 273, 11413-11416. (16) Lipscomb, W. N.; Strater, N. Chem. ReV. 1996, 96, 2375-2433. (17) Matthews, B. W. Acc. Chem. Res. 1988, 21, 333-340. (18) Rajagopalan, P. T. R.; Grimme, S.; Pei, D. H. Biochemistry 2000, 39, 779-790. (19) Deng, H.; Callender, R.; Zhu, J. G.; Nguyen, K. T.; Pei, D. H. Biochemistry 2002, 41, 10563-10569. (20) Groche, D.; Becker, A.; Schlichting, I.; Kabsch, W.; Schultz, S.; Wagner, A. F. V. Biochem. Biophys. Res. Commun. 1998, 246, 342-346. (21) Ragusa, S.; Mouchet, P.; Lazennec, C.; Dive, V.; Meinnel, T. J. Mol. Biol. 1999, 289, 1445-1457. (22) Rajagopalan, P. T. R.; Datta, A.; Pei, D. H. Biochemistry 1997, 36, 13910-13918.

Bacterial Peptide Deformylase (23) Jain, R. K.; Hao, B.; Liu, R. P.; Chan, M. K. J. Am. Chem. Soc. 2005, 127, 4558-4559. (24) Karambelkar, V. V.; Xiao, C. Y.; Zhang, Y. K.; Sarjeant, A. A. N.; Goldberg, D. P. Inorg. Chem. 2006, 45, 1409-1411. (25) Madison, V.; Duca, J.; Bennett, F.; Bohanon, S.; Cooper, A.; Chu, M.; Desai, J.; Girijavallabhan, V.; Hare, R.; Hruza, A.; Hendrata, S.; Huang, Y.; Kravec, C.; Malcolm, B.; Mccormick, J.; Miesel, L.; Ramanathan, L.; Reichert, R.; Saksena, A.; Wang, J.; Weber, P. C.; Zhu, H.; Fischmann, T. Biophys. Chem. 2002, 101, 239-247. (26) Leopoldini, M.; Russo, N.; Toscano, M. J. Phys. Chem. B 2006, 110, 1063-1072. (27) Warshel, A.; Levitt, M. J. Mol. Biol. 1976, 103, 227. (28) Singh, U. C.; Kollman, P. J. Comput. Chem. 1986, 7, 718-730. (29) Field, M. J.; Bash, P. A.; Karplus, M. J. Comput. Chem. 1990, 11, 700-733. (30) Gao, J. In ReView in Computational Chemistry; VCH: New York, 1995; Vol. 7, pp 119-185. (31) Gogonea, V.; Suarez, D.; van der Vaart, A.; Merz, K. W. Curr. Opin. Struct. Biol. 2001, 11, 217-223. (32) Field, M. J. J. Comput. Chem. 2002, 23, 48-58. (33) Gao, J.; Truhlar, D. G. Annu. ReV. Phys. Chem. 2002, 53, 467505. (34) Cui, Q.; Karplus, M. AdV. Protein Chem. 2003, 66, 315-372. (35) Friesner, R. A.; Guallar, V. Annu. ReV. Phys. Chem. 2005, 56, 389427. (36) Zhang, Y. Theor. Chem. Acc. 2006, 116, 43-50. (37) Zhang, Y.; Lee, T. S.; Yang, W. J. Chem. Phys. 1999, 110, 4654. (38) Zhang, Y.; Liu, H.; Yang, W. J. Chem. Phys. 2000, 112, 34833492. (39) Zhang, Y. J. Chem. Phys. 2005, 122, 024114. (40) Liu, H.; Zhang, Y.; Yang, W. J. Am. Chem. Soc. 2000, 122, 65606570. (41) Zhang, Y.; Kua, J.; McCammon, J. A. J. Am. Chem. Soc. 2002, 124, 10572-10577. (42) Zhang, Y.; Kua, J.; McCammon, J. A. J. Phys. Chem. B 2003, 107, 4459-4463. (43) Cisneros, G. A.; Liu, H.; Zhang, Y.; Yang, W. J. Am. Chem. Soc. 2003, 125, 10384-10393. (44) Cheng, Y.; Zhang, Y.; McCammon, J. A. J. Am. Chem. Soc. 2005, 127, 1553-1562. (45) Hu, P.; Zhang, Y. J. Am. Chem. Soc. 2006, 128, 1272-1278. (46) Corminboeuf, C.; Hu, P.; Tuckerman, M. E.; Zhang, Y. J. Am. Chem. Soc. 2006, 128, 4530-4531.

J. Phys. Chem. B, Vol. 111, No. 22, 2007 6235 (47) Ponder, J. W. TINKER, Software Tools for Molecular Design, Version 3.6. The most updated version for the TINKER program can be obtained from J. W. Ponder’s World Wide Web site at http://dasher.wustl.edu/tinker; February 1998. (48) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. J. Chem. Phys. 1983, 79, 926-935. (49) Williams, I. H.; Maggiora, G. M. THEOCHEM 1982, 6, 365-378. (50) Rothman, M. J.; Lohr, L. L. Chem. Phys. Lett. 1980, 70, 405. (51) Scharfenberg, P. Chem. Phys. Lett. 1981, 79, 115-117. (52) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R., 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.; 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, Revision B.05; Gaussian, Inc.: Pittsburgh, PA, 2003. (53) Becke, A. D. J. Chem. Phys. 1988, 88 (4), 2547-2553. (54) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (55) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (56) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.; Kollman, P. A. J. Am. Chem. Soc. 1995, 117, 5179-5197. (57) Dolg, M.; Wedig, U.; Stoll, H.; Preuss, H. J. Chem. Phys. 1987, 86, 866-872. (58) Hehre, W.; Radom, L.; Schleyer, P.; Pople, J. Ab Initio Molecular Orbital Theory; John Wiley & Sons: New York, 1986. (59) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270-283. (60) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299-310. (61) Chang, S. C.; Sommer, R. D.; Rheingold, A. L.; Goldberg, D. P. Chem. Commun. 2001, pp 2396-2397. (62) Meinnel, T.; Lazennec, C.; Villoing, S.; Blanquet, S. J. Mol. Biol. 1997, 267, 749-761. (63) Wang, S.; Hu, P.; Zhang, Y. J. Phys. Chem. B 2007, 111, 37583764.