• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of prosciprotein sciencecshl presssubscriptionsetoc alertsthe protein societyjournal home
Protein Sci. Sep 2006; 15(9): 2071–2081.
PMCID: PMC2242601

Peptide deformylase is a potential target for anti-Helicobacter pylori drugs: Reverse docking, enzymatic assay, and X-ray crystallography validation


Colonization of human stomach by the bacterium Helicobacter pylori is a major causative factor for gastrointestinal illnesses and gastric cancer. However, the discovery of anti-H. pylori agents is a difficult task due to lack of mature protein targets. Therefore, identifying new molecular targets for developing new drugs against H. pylori is obviously necessary. In this study, the in-house potential drug target database (PDTD, http://www.dddc.ac.cn/tarfisdock/) was searched by the reverse docking approach using an active natural product (compound 1) discovered by anti-H. pylori screening as a probe. Homology search revealed that, among the 15 candidates discovered by reverse docking, only diaminopimelate decarboxylase (DC) and peptide deformylase (PDF) have homologous proteins in the genome of H. pylori. Enzymatic assay demonstrated compound 1 and its derivative compound 2 are the potent inhibitors against H. pylori PDF (HpPDF) with IC50 values of 10.8 and 1.25 μM, respectively. X-ray crystal structures of HpPDF and the complexes of HpPDF with 1 and 2 were determined for the first time, indicating that these two inhibitors bind well with HpPDF binding pocket. All these results indicate that HpPDF is a potential target for screening new anti-H. pylori agents. In addition, compounds 1 and 2 were predicted to bind to HpPDF with relatively high selectivity, suggesting they can be used as leads for developing new anti-H. pylori agents. The results demonstrated that our strategy, reverse docking in conjunction with bioassay and structural biology, is effective and can be used as a complementary approach of functional genomics and chemical biology in target identification.

Keywords: peptide deformylase, protein crystallization, reverse docking, enzyme inhibitor, Helicobacter pylori

Colonization of human stomach by the human pathogenic bacterium Helicobacter pylori is a major causative factor for gastrointestinal illnesses such as chronic gastritis and peptic ulcer (Tee et al. 1995). Infection of H. pylori is also associated with adenocarcinoma and stomach lymphoma, increasing the risk of gastric cancer (Cover and Blaser 1996). There is no effective therapy for eradicating H. pylori infection. Combination therapies employing one proton pump inhibitor (e.g., omeprazole) and two or three antibiotics (e.g., amoxicillin, clarithromycin, or tetracycline) have been used as preferred treatments (Ulmer et al. 2003). However, the multiple therapy regimens have not been very effective in a clinical setting, because H. pylori is likely to develop resistance (Cameron et al. 2004). Moreover, this treatment may disrupt the natural population of commensal microorganisms in the gastrointestinal tract, leading to undesired side effects such as diarrhea (Carcanague et al. 2002). Therefore, there are urgent needs for discovering novel anti-H. pylori agents. Nevertheless, all the current anti-H. pylori agents were almost discovered by random screening with the MIC (minimal inhibitory concentration) assays, such as broth dilution and agar dilution methods, because mature protein target for screening anti-H. pylori agents is destitute. Accordingly, identifying new molecular targets to develop new drugs against the pathogen of H. pylori is obviously necessary (Legrain and Strosberg 2002; Cremades et al. 2005).

Because small organic molecules can alter or perturb the functions of target proteins by inhibiting or activating their normal functions through binding, they have been widely used to illuminate the molecular mechanisms underlying biological processes. This approach is referred to as chemical biology (Stockwell 2004). Compounds with functions of activating or inhibiting cellular cycle should be likely probes to map the protein targets. To this end, proteomics may be an appropriate approach for identifying particular binding proteins of the small molecules by comparing the differences of protein expression profiles between pathological cells and cells treated by chemicals. However, this method is not very successful in target discovery because of its time consuming and slower rate of reproduction (Huang et al. 2004). An alternative approach that has been proved to be promising in recent years is to find the probable binding protein(s) for an active compound from the genomic or protein database by using computational methods, and then to validate the computational results by traditional molecular and/or cell biology methods (Rockey and Elcock 2005).

In the following, we report on the finding of peptide deformylase (PDF) as a potential target for anti-H. pylori agents. The result was discovered by using computational method and verified with bioassay and X-ray crystallography. Briefly, taking the natural product, N-trans-caffeoyltyramine (compound 1), discovered by anti-H. pylori assay as a probe, we searched the in-house potential drug target database (PDTD) by using a reverse docking method (http://www.dddc.ac.cn/tarfisdock/), TarFisDock (Li et al. 2006), and found that Escherichia coli PDF is a binding protein candidate. Sequence alignment indicated that H. pylori PDF (HpPDF) (GenBank NP_223447) contains 40% identity to E. coli PDF. Enzymatic assay demonstrated that compound 1 and its derivative compound 2 ((E)-N-phenethyl-3-(3,4-diacetoxyphenyl)acrylamide) are potent inhibitors of HpPDF. Finally, we determined the X-ray crystal structures of the apo-HpPDF and inhibitor-HpPDF complexes, which demonstrated that the two compounds tightly bind to the active site of HpPDF. Our results illuminate that the reverse docking method can facilitate the identification of target protein binding to active compounds.


Active natural product identification

In a random screening of a diverse small molecules and herbal extracts library by using agar dilution method to identify active components against H. pylori, compound 1 isolated from Ceratostigma willmottianum, a folk medicine used to remedy rheumatism, traumatic injury, and parotitis (Yue et al. 1997), showed inhibitory activity against H. pylori with MIC value of 180 μg/mL. Chemical modification on compound 1 afforded a number of analogs, and compound 2 is the most active one with the improved MIC value of 100 μg/mL against H. pylori. The chemical structures of compounds 1 and 2 were shown in Figure 1.

Figure 1.
The chemical structures of compounds 1 and 2.

Potential target identification

Potential binding proteins of compound 1 were screened from our in-house potential drug target database (PDTD; http://www.dddc.ac.cn/tarfisdock/) using the reverse docking approach TarFisDock (Li et al. 2006). Fifteen protein targets (see Supplemental Table 1) with interaction energies to compound 1 under −35.0 kcal/mol were identified. Homology search indicated that only two proteins of the 15 candidates, diaminopimelate decarboxylase (DC) and PDF, have homologous proteins in the genome of H. pylori. DC encoded by the lysA gene converts meso-DAP to L-lysine, which is the final step in the bacterial lysine biosynthetic pathway. Lysine is essential for bacterial viability and development. Unlike bacteria, humans lack the lysine biosynthetic pathway and obtain lysine from dietary sources. Thus, DC could be a potential anti-bacterial drug target. In prokaryotes and eukaryotic organelles (e.g., mitochondria and chloroplast), protein synthesis is initiated with N-formylmethionyl-tRNAi, resulting in N-terminal formylation of all nascent polypeptides. During elongation of polypeptide chain, PDF catalyzes the removal of a formyl group at the N-terminal. PDF is essential for bacterial cell growth (Mazel et al. 1994; Meinnel and Blanquet 1994), and it has been proved that the inactivation of the def gene encoding PDF in Streptococcus pneumoniae cannot be achieved (Margolis et al. 2001). Although PDF has also been discovered in human beings (Lee et al. 2003; Serero et al. 2003), PDF takes no effect on cytoplasmic protein synthesis in mammalian cells. Therefore, PDF is also an attractive target for discovering novel antibiotics (Nguyen et al. 2003). Thus, we chose DC and PDF as probable binding proteins of compound 1 for target validation.

Figure 2 represents the sequence alignment of HpPDF with the other PDFs, indicating that PDFs are very conservative and they share high percentage of homology; the similarities of PDF sequences range from 50% to 65%. The metallic ions (e.g., Co2+, Fe2+, or Zn2+) in the active site of PDF were predicted to be coordinated with two histidines from the conserved motif HEXXH, a cystine from the conserved motif EGCLS, and a water molecule. Bacterial PDF was first characterized as a Fe2+-containing enzyme that is very labile because of its conversion to Fe3+ by molecular oxygen and H2O2, resulting in inactivation (Rajagopalan and Pei 1998). The Fe2+ can be replaced by other metal ions such as Ni2+, Zn2+, and Co2+. The catalytic activities of different metal forms of PDF have been shown to be of great diversity (Ragusa et al. 1998). Based on the phylogenetic tree analysis and systematic sequence alignment, PDFs can be classified into two types, i.e., type-I (Gram-negative type) and type-II (Gram-positive type) (Guilloteau et al. 2002). HpPDF belongs to the type-I family.

Figure 2.
Sequence alignment of HpPDF with several other PDFs, including type-I (E. coli, P. aeruginosa, Plasmodium falciparum, and Leptospira interrogans) and type-II (Staphylococcus aureus, Bacillus stearothermophilus, and S. pneumoniae) PDFs. The sequence alignments ...

Enzymatic validation

The inhibitory activities of compounds 1 and 2 to HpPDF were measured by using the formate dehydrogenase (FDH)-coupled assay (Lazennec and Meinnel 1997). Figure 3 depicts the dose-dependent inhibitions of HpPDF by these two compounds. The IC50 values of compounds 1 and 2 inhibiting the catalytic activity of Co-HpPDF were estimated as 10.8 and 1.25 μM, respectively, which are in agreement with the MIC values. This result indicates that HpPDF is really a target of compound 1 and its analog compound 2. In parallel, the inhibitory activities of compounds 1 and 2 to H. pylori DC (HpDC) were also measured by using the double-enzyme coupled assay (Scriven et al. 1988; Ray et al. 2002). However, compounds 1 and 2 have no inhibitory effect on HpDC, indicating that HpDC is not a target of compounds 1 and 2.

Figure 3.
Dose-response curves of HpPDF enzyme inhibition by compounds 1 and 2.

X-ray crystal structure of HpPDF

To verify that HpPDF is a binding protein of compounds 1 and 2 at the atomic level, the X-ray structures of both HpPDF and HpPDF-inhibitor complexes were determined by using the crystallographic method. Analysis of the diffraction data indicated that the crystal structure of HpPDF belongs to space group P212121 and contains one HpPDF molecule per asymmetric unit. The statistics of the diffraction data and structure refinement are shown in Table 1. All the X-ray crystal structures were refined to resolution of 2.2 Å.

Table 1.
Statistics of X-ray data collection and structure refinement

Phases of all three sets of diffraction data were obtained by molecular replacement method using the program CNS (Brunger et al. 1998). A homology model was built using the structure of PDF from Pseudomonas aeruginosa (PDB code 1IX1; Yoon et al. 2004) as the template, and the homology-built structure was taken as the initial search model. Remarkably, >20 nonconserved C-terminal residues were excluded from the search model. During the later refinement of HpPDF crystal structure, we can find that some of these residues constitute an α-helix possessing a unique spatial orientation compared with the other PDFs. The subsequent refinements were preformed with the program CNS, including rigid body refinement, energy minimization, and B factor refinement. Manual building of the model was carried out with the program O (Jones et al. 1991) according to the 2Fo-Fc map. After several cycles of model rebuilding and refinement, most of the residues were clearly positioned in the 2Fo-Fc map. The cobalt ion in the active site was unambiguously distinguished by direct examination of the Fo-Fc map. Water molecules were added automatically with CNS using 3δ peaks in the Fo-Fc map and hydrogen bonding requirements, and the uncertain water molecules were removed manually by inspecting the electron density map. After incorporating the cobalt ion and water molecules, some previously unrecognizable regions of the electron density map were greatly improved (e.g., the flexible loop of residues 66–71), but the last 10 residues of the C terminus (residues 165–174) were still disordered in the map. The final R-factor is 20.9% (free R-factor 24.4%), and statistics show that the final model of native HpPDF is a well-refined structure (Table 1). The positions of inhibitors were determined based on the Fo-Fc map of the complexes, using the solved structure of HpPDF as the model. Similar to the structure of apo-HpPDF, the last seven C-terminal residues (residues 168–174) of the complexes were not found in the electron density maps. The R-factors of HpPDF-1 complex and HpPDF-2 complex are 20.4% and 20.1%, respectively. The crystal structures are shown in Figures 4 and and55.

Figure 4.
Comparison of the X-ray structure of HpPDF with that of other PDFs. (A) The stereo view of the overall structure of HpPDF. (B) Structure alignment of the Cα atoms of HpPDF with that of currently available PDF structures in PDB. Blue, EcPDF (E. ...
Figure 5.
Stereo images of SIGMAA-weighted 2Fo-Fc maps (1σ contoured level) of compound 1 (A) and compound 2 (B). The final coordinates of the structures of compounds 1 and 2 are shown as stick models. The figures were generated by using PyMOL (DeLano 2002 ...

Similar to the other PDFs, the overall structure of HpPDF contains three α-helices, seven β-strands, and four 310 helices (Fig. 4A). A cobalt ion is located at the active site of the enzyme, which is tetrahedrally coordinated with His138 and His142 from the motif HEXXH, Cys96 from the motif EGCLS, and a water molecule (Figs. 2, ,4A).4A). Detailed structural comparison of HpPDF with several currently determined X-ray crystal structures of PDFs (Hao et al. 1999; Guilloteau et al. 2002; Kumar et al. 2002; Yoon et al. 2004; Zhou et al. 2004) reveals some distinct structural differences (Fig. (Fig.4B).4B). One distinctive difference comes from the CD loop composed of Asn62–Cys75 between the βC and βD, which is structurally different from any of the other PDFs (Fig. (Fig.4C).4C). This is consistent with the phenomenon that the CD loops of PDFs are conformationally flexible (Fig. (Fig.4B4B,,C).C). Another difference comes from the C terminus. Leu152 and Ser153 before the C-terminal helix fold into a conformation different from either that of type-II PDF or that of type-I PDF (Fig. (Fig.4B4B).

Inhibitor-HpPDF interactions

Like the binding pockets of the other PDFs, the substrate-binding pocket of HpPDF consists of S1′, S2′, and S3′ subsites (see Supplemental Fig. 1). S1′ and S3′ are two cavities, and S2′ is a “saddle” linking S2′ and S3′. We made a comparison of the binding pockets of different PDFs. The surfaces of the pockets are also shown in Supplemental Figure 1. The shapes of the binding pockets of PDFs are diverse, reflecting the selectivity and specificity for substrate or inhibitor binding.

X-ray data for trigonal crystals of HpPDF soaked with compounds 1 or 2 were also collected and refined to 2.2 Å resolutions (Table 1). After several cycles of refinements, the structure of HpPDF in complex with compound 2 could be fitted unambiguously in the electron density map (Fig. 5B), whereas the structure of HpPDF in complex with compound 1 is partially disordered with an average B-factor of 59.8 Å2 and has a weak broken electron density (Fig. 5A). This suggests that the occupation of compound 1 at the binding pocket of HpPDF is not 100%, which is consistent with its weak binding affinity to HpPDF. Structure superposition shows that both compounds 1 and 2 adopt a similar orientation and conformation in the substrate-binding pocket. The heads for both compounds (phenyl group for compound 1 and phenol group for compound 2) enter into the S1′ subsite, and the tails extend along the pocket entrance (Fig. 6A). For either compound 1 or compound 2, the acylamide oxygen atom forms two hydrogen bonds to the main-chain nitrogen atoms of Ile45 and Gly46, respectively, and the acylamide nitrogen atom hydrogen bonds to the carbonyl oxygen atom of Gly95 (Fig. 6B,,C).C). The tail ring (3,4-diacetoxyphenyl) of compound 2 flips ~180° in comparison with that of compound 1 (3,4-dihydroxyphenyl) to avoid the steric hindrance between the 3-acetoxy and the enzyme (Fig. 6A). This led to different hydrogen bonding models for the tails of compounds 1 and 2 to HpPDF. The oxygen atom of 3-hydroxyl of compound 1 forms hydrogen bonds to the main-chain nitrogen atom of Gly101, the main-chain nitrogen atom of Cys96, and the main-chain oxygen atom of Phe102 through a water molecule. The tail end of compound 2 only forms one hydrogen bond to HpPDF through the 4-acetoxy carbonyl oxygen atom with the hydroxyl of Tyr103. Therefore, the hydrogen bonding of compound 1 to HpPDF might be stronger than that of compound 2 to HpPDF, which is not in agreement with the inhibitory activities (compound 2 is more potent than compound 1). Further analysis reveals that the binding pocket of HpPDF, especially S1′ subsite (Fig. 6A; Supplemental Fig. 1), is hydrophobic, and compound 2 is structurally more hydrophobic (XlogP = 2.87) than compound 1 (XlogP = 2.78). Accordingly, compound 2 may form a more favorable hydrophobic interaction than compound 1 does. This conclusion is confirmed by the component analysis for the interaction energies of compounds 1 and 2 with HpPDF (see Supplemental Table 2). The result demonstrates that van der Waals interaction dominates the binding of the two compounds with HpPDF, suggesting that the interactions of the compounds with HpPDF are driven by hydrophobic force. In addition, compared with the structure of HpPDF, Tyr92 in HpPDF-1 complex moves toward solvent due to the repulsive interaction between the hydroxyl group in the head of compound 1 and the hydroxyl of Tyr92 (Fig. 7). In the HpPDF-2 complex, Tyr92 and the phenyl group of compound 2 close up together due to the hydrophobic interaction. The movement of Tyr92 makes room for a water molecule, which forms a hydrogen bond to the hydroxyl group of Tyr92 (Fig. 7). This hydrogen bond may stabilize the protein structure. Therefore, we can conclude that the hydrophobic interaction dominates the activity difference between 1 and 2.

Figure 6.
Bindings of compounds 1 and 2 to HpPDF. (A) Comparison of the conformations of compounds 1 and 2 in the binding pocket of HpPDF. The carbon atoms of compounds 1 and 2 are shown in cyan. (B) Hydrogen bonding between compound 1 and HpPDF. (C) Hydrogen bonding ...
Figure 7.
Structural superposition of the residues composed of S1′ subsites of apo-HpPDF (magenta), HpPDF-1 complex (green), and HpPDF-2 complex (yellow).

Binding predication of compounds 1 and 2 to the other PDFs

To test the selectivity of compounds 1 and 2 to HpPDF, we docked these two compounds into the binding pockets of several other PDFs, and the binding free energies of the compounds to PDFs (including HpPDF) were calculated using the scoring function of AutoDock 3.1 (Morris et al. 1998). The results are listed in Supplemental Table 3. The predicted binding free energies indicate that, in general, these two inhibitors may bind to the type-I PDFs but may not bind to the type-II PDFs, suggesting that these two compounds are selective inhibitors of type-I PDFs. In addition, the binding affinities of these two compounds to HpPDF are higher than those of them to the other type-I PDFs. To verify this prediction, we determined the inhibitory activities of compounds 1 and 2 against E. coli PDF (EcPDF). The result indicates that these two compounds do not show inhibitory effect on EcPDF, suggesting that compounds 1 and 2 may be specific inhibitors to HpPDF and it is possible to obtain HpPDF-selective inhibitors by modifying the structures of these two compounds.


We described here that PDF is a potential target protein for anti-H. pylori agents. Compound 1 was first found as an active compound to inhibit the growth of H. pylori by screening a series of natural products and herbs using the MIC approach. Afterward, a computational method called “reverse docking” was used to search the possible binding proteins of compound 1 from the potential drug target database (PDTD). Totally, 15 candidate proteins were found, and homology search revealed that there are only DC and PDF homologous proteins in the genome of H. pylori.

We verified that HpPDF is a target of compound 1 by enzymatic assay and crystal structure determination. Before that, compound 2, one analog of compound 1, was found as a more potent inhibitor of H. pylori. The FDH-coupled assay indicated that compounds 1 and 2 are exact inhibitors of HpPDF, and the IC50 values of these two compounds are, respectively, 10.8 and 1.25 μM, which are in agreement with the MIC values. The X-ray crystal structures indicated that compounds 1 and 2 fit well into the binding pocket of HpPDF, demonstrating at the atomic level that PDF is the binding protein of these two compounds. However, enzymatic assay indicated that DC is not a real target protein for compounds 1 and 2.

For the first time, we determined the crystal structure of HpPDF. The overall structure of HpPDF folded in a similar way to the other PDFs, and a cobalt ion tetrahedrally coordinated with two histidines, one cystine, and a water molecule. The CD loop of HpPDF adopted a conformation different from any other PDF, and the C-terminal helix shifted away from its original position as in the other type-I PDFs. The binding pocket of HpPDF is different from those of the other PDFs, implying that the selective HpPDF inhibitors could be designed. As a matter of fact, compounds 1 and 2 are the selective inhibitors of the type-I PDFs. Moreover, these two compounds, especially compound 1, may be modified as selective HpPDF inhibitors, because the predicted binding affinities of these two compounds to HpPDF are higher than those of them to the other PDFs, and enzymatic assay also demonstrated that these two compounds cannot inhibit EcPDF. In addition, different from the other inhibitors of PDF (e.g., actinonin), compounds 1 and 2 bind to HpPDF noncovalently, and these two compounds have novel and simple chemical scaffolds. Therefore, these two compounds can be used as leads for developing new anti-H. pylori agents.

Target identification and validation is the first key stage in the drug discovery pipeline. Numerous technologies for addressing targets have been developed recently (Wang et al. 2004). Genomics and proteomics approaches, including bioinformatics analysis, are the major tools for target identification. Chemical biology, which generally uses small molecules as probes to map the genomic functions, is an emerging tool for target identification (Stockwell 2004). In this study, we provided an alternative approach for target identification, i.e., discovering the potential binding protein candidates of active compounds (natural products or existing drugs) from the protein databases (e.g., PDTD or PDB) by using the reverse docking method TarFisDock, a computational method we developed. The computational clues were verified by enzymatic assay and even X-ray crystallography determination. The result of the present study demonstrated that this approach is effective and can be used as a complementary approach of genomics and chemical biology in target identification for other systems. However, the reverse docking approach TarFisDock still has certain limitations (Li et al. 2006). The major one is that the protein entries are not enough to cover all the protein information of disease-related genomes. The second one is that TarFisDock has not considered the flexibility of proteins during docking simulation. These two aspects will produce a false negative. Another limitation is that the scoring function for reverse docking is not accurate enough, which will produce a false positive.

Materials and methods


H. pylori strain SS1 was obtained from Shanghai Institute of Digestive Disease. E. coli host strain BL21(DE3) was purchased from Stratagene. Crystallization screen kits were purchased from Hampton Research. All chemicals were of reagent grade or ultrapure quality, and commercially available.

In vitro anti-H. pylori determination

The in vitro anti-H. pylori activities of compounds or crude extracts of herbs were estimated by determining the MIC with agar dilution method. A series of agar plates were prepared with the base of Campylobacter selective agar (Merck) containing 5% of fetal bovine serum. Then, various concentrations of twofold diluted test compounds or crude extracts of herbs were dispersed into the prepared agar plates. Cells of H. pylori strain SS1, suspended in saline at the density of 108 cfu/mL, were added to the well-prepared agar plates and were incubated at 37°C for 96 h under an atmosphere of 5% O2, 10% CO2, and 85% N2. Blank controls and positive controls were performed using the same conditions as described above, except that in the case of blank controls no compound or crude extract was dispersed into the agar plates, while in the case of positive controls various concentrations of twofold diluted metronidazole were dispersed. The MIC value was defined as the lowest concentration of the compound or crude extract for inhibiting the visible growth.

Reverse molecular docking

Taking compound 1 as a probe, we searched our in-house potential drug target database (PDTD). PDTD contains 698 protein structures isolated from PDB. Missing residues and atoms of each protein structure were repaired using the Biopolymer module of Sybyl 6.8 (Tripos Associates), and Kollman (Weiner et al. 1986; Cornell et al. 1995) charges were assigned to the protein. A grid of each protein binding pocket (site) was constructed by using the grid module of DOCK4.0 (Kuntz 1992; Ewing and Kuntz 1997), which was mapped onto the original protein structure deposited in PDTD. Afterward, compound 1 was docked into the binding site of each protein using DOCK4.0, and interaction energies between compound 1 and the proteins were calculated using the scoring function of DOCK4.0. Proteins with interaction energies to compound 1 < −35.0 kcal/mol were selected for further analysis. Recently, based on this computational method, we developed a Web-based tool, TarFisDock (Target Fishing Dock), for searching the probable binding proteins for active small molecules (Li et al. 2006), which can be accessed through http://www.dddc.ac.cn/tarfisdock/.

Homologous proteins search

Reverse docking produced a series of binding protein candidates for compound 1. Homology search was performed to identify the homologous proteins of these candidates from H. pylori genome. The candidates were selected as queries for searching the GenBank database (http://www.ncbi.nlm.nih.gov/).

Enzyme inhibition assay

Recombinant HpPDF and EcPDF were overexpressed and purified from E. coli strain BL21(DE3) as described previously (Rajagopalan et al. 1997; Han et al. 2004). A FDH-coupled assay was used to determine the inhibitory activity of the compounds (Lazennec and Meinnel 1997). The formate generated by PDF from its substrate N-formyl-Met-Ala-Ser is oxidized by the enzyme FDH, reducing NAD+ to NADH, which causes specific absorption at 340 nm (epsilonM = 6300 M−1cm−1). All assays were conducted at 37°C in a 96-well plate system (Tecan GENios reader) by measuring the increase in absorbance at 340 nm. The reaction mixture contained 50 mM HEPES (pH 7.5), 10 mM NaCl, 0.2 mg/mL BSA (Roche), 8 mM NAD+ (Roche), 0.5 U/mL FDH (Fluka), and 2 mM N-formyl-Met-Ala-Ser (Sangon). To determine IC50 (the concentration needed to inhibit 50% of enzyme activity) of a compound, PDF activity was measured in the presence of increasing concentrations of the compound in the presence of an f-MAS concentration corresponding to Km value (Han et al. 2004). Compounds were added to assay mixtures from concentrated stocks dissolved in dimethyl sulfoxide (Me2SO). The final Me2SO concentration in all assays was 0.1% (v/v). IC50 value was obtained by fitting the data to a sigmoid dose-response equation using the Origin software (OriginLab). The reaction was initiated by the addition of the diluted HpPDF (or EcPDF) enzyme.

Recombinant HpDC was expressed and purified in E. coli system. A double-enzyme coupled assay was used to determine the inhibitory effect of compounds 1 and 2 (Scriven et al. 1988; Ray et al. 2002). In this assay, HpDC activity was estimated by monitoring the decrease of specific absorption at 340 nm caused by NADH. The degassed assay buffer contained 100 mM Tris (pH 8.0), 10 mM MgCl2, 2 mM phosphoenolpyruvate (Fluka), 2 mM 2,6-diaminopimelic acid (Fluka), 0.7 mM NADH (Fluka), 0.2 U/mL phosphoenolpyruvate carboxylase, and 1.25 U/mL malate dehydrogenase. Compounds dissolved in Me2SO were added to assay mixtures. The final Me2SO concentration in all assays was 0.1% (v/v). All assays were conducted at 37°C in a 96-well plate system. The reaction was initiated by the addition of the diluted HpDC enzyme.

Crystallization and X-ray structure determination

The purified HpPDF protein was prepared in buffer A (10 mM Tris-HCl at pH 8.0, 0.1 M NaCl, 1 mM DTT). Prior to crystallization, the protein was concentrated to 20–30 mg/mL and stored at 4°C. Initial screening was performed at 4°C by the hanging-drop vapor-diffusion method. Drops were prepared by mixing 1 μL of protein solution with 1 μL of reservoir solution, and were equilibrated against 500 μL of reservoir solution. The conditions yielding small crystals were further optimized by variation of the buffer pH and precipitant concentration. The best crystals were grown at a reservoir solution of 60%–70% MPD in 0.1 M HEPES (pH 7.8). The crystals of HpPDF in complex with compounds 1 and 2 were obtained by soaking method.

All diffraction data were collected in-house on a Rigaku rotating-anode X-ray generator operated at 50 kV and 100 mA (λ = 1.5418 Å). Diffraction images were recorded by a Rigaku R-AXIS IV++ imaging-plate detector with an oscillation step of 1°. The crystal-to-detector distance was set to 15 cm. The crystals were picked up with a nylon loop and flash-cooled in liquid nitrogen. Data collection was performed at 100 K using the original reservoir solution as cryoprotectant. All data were processed and scaled using the CrystalClear program (Pflugrath 1999).

The structures of both apo-enzyme and the inhibitor-enzyme complexes were solved by molecular replacement method using the program CNS (Brunger et al. 1998), taking the X-ray crystal structure of P. aeruginosa PDF (PDB code 1IX1; Yoon et al. 2004) as the initial model. Refinement was performed by using CNS, and the atomic models were built by using the computer graphics program O (Jones et al. 1991). The Ramachandran statistics of the final models were listed in Table 2. The resolved structures of both apo-enzyme and the inhibitor-enzyme complexes have been deposited in the PDB database (codes 2EW5, 2EW6, and 2EW7).

Table 2.
Ramachandran statistics of the final models

Electronic supplemental material

The Supplemental Material contains one figure and three tables: Supplemental Table 1, possible binding protein candidates of compound 1 searched out from the PDTD by using reverse docking; Supplemental Table 2, component analysis for the interaction energies of compounds 1 and 2 with HpPDF; Supplemental Table 3, binding free energies of different PDFs to compounds 1 and 2; and Supplemental Figure 1, the surface expressions of the pockets of different PDFs. The chemical identities and purities of compounds 1 and 2 can be found in the Supplemental Materials.


This work was supported by the State Key Program of Basic Research of China (grants 2002CB512802, 2002CB512807, and 2004CB518905), the National Natural Science Foundation of China (grants 30525024 and 20372069).


Supplemental material: see www.proteinscience.org

Reprint requests to: Xu Shen, Hualiang Jiang, or Jianmin Yue, Drug Discovery and Design Center, State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Graduate School of Chinese Academy of Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai 201203, China; e-mail: xshen/at/mail.shcnc.ac.cn, hljiang/at/mail.shcnc.ac.cn, or jmyue/at/mail.shcnc.ac.cn; fax: 86-21-50807088.

Article published online ahead of print. Article and publication date are at http://www.proteinscience.org/cgi/doi/10.1110/ps.062238406.

Abbreviations: MIC, minimal inhibitory concentration; PDF, peptide deformylase; PDTD, potential drug target database; DC, diaminopimelate decarboxylase; FDH, formate dehydrogenase.


  • Brünger A.T., Adams P.D., Clore G.M., DeLano W.L., Gros P., Grosse-Kunstleve R.W., Jiang J.S., Kuszewski J., Nilges M., Pannu N.S. et al. 1998. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54: 905–921. [PubMed]
  • Cameron E.A., Powell K.U., Baldwin L., Jones P., Bell G.D., Williams S.G. 2004. Antibiotic resistance and eradication rates in Suffolk, UK, 1991–2001. J. Med. Microbiol. 53: 535–538. [PubMed]
  • Carcanague D., Shue Y.K., Wuonola M.A., Uria-Nickelsen M., Joubran C., Abedi J.K., Jones J., Kuhler T.C. 2002. Novel structures derived from 2-[[(2-pyridyl)methyl]thio]-1H-benzimidazole as anti-Helicobacter pylori agents, Part 2. J. Med. Chem. 45: 4300–4309. [PubMed]
  • 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. 1995. A second generation force field for the simulation of proteins, nucleic acids, and organic molecules. J. Am. Chem. Soc. 117: 5179–5197.
  • Cover T.L. and Blaser M.J. 1996. Helicobacter pylori infection, a paradigm for chronic mucosal inflammation: Pathogenesis and implications for eradication and prevention. Adv. Intern. Med. 41: 85–117. [PubMed]
  • Cremades N., Bueno M., Toja M., Sancho J. 2005. Towards a new therapeutic target: Helicobacter pylori flavodoxin. Biophys. Chem. 115: 267–276. [PubMed]
  • DeLano W.L. 2002. The PyMOL molecular graphics system. DeLano Scientific, San Carlos, CA.
  • Ewing T.J.A. and Kuntz I.D. 1997. Critical evaluation of search algorithms for automated molecular docking and database screening. J. Comput. Chem. 18: 1175–1189.
  • Guilloteau J.P., Mathieu M., Giglione C., Blanc V., Dupuy A., Chevrier M., Gil P., Famechon A., Meinnel T., Mikol V. 2002. The crystal structures of four peptide deformylases bound to the antibiotic actinonin reveal two distinct types: A platform for the structure-based design of antibacterial agents. J. Mol. Biol. 320: 951–962. [PubMed]
  • Han C., Wang Q., Dong L., Sun H., Peng S., Chen J., Yang Y., Yue J., Shen X., Jiang H. 2004. Molecular cloning and characterization of a new peptide deformylase from human pathogenic bacterium Helicobacter pylori. Biochem. Biophys. Res. Commun. 319: 1292–1298. [PubMed]
  • Hao B., Gong W., Rajagopalan P.T., Zhou Y., Pei D., Chan M.K. 1999. Structural basis for the design of antibiotics targeting peptide deformylase. Biochemistry 38: 4712–4719. [PubMed]
  • Huang C.M., Elmets C.A., Tang D.C., Li F., Yusuf N. 2004. Proteomics reveals that proteins expressed during the early stage of Bacillus anthracis infection are potential targets for the development of vaccines and drugs. Genomics Proteomics Bioinformatics 2: 143–151. [PubMed]
  • Jones T.A., Zou J.Y., Cowan S.W., Kjeldgaard M. 1991. Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A 47: 110–119. [PubMed]
  • Kumar A., Nguyen K.T., Srivathsan S., Ornstein B., Turley S., Hirsh I., Pei D., Hol W.G. 2002. Crystals of peptide deformylase from Plasmodium falciparum reveal critical characteristics of the active site for drug design. Structure 10: 357–367. [PubMed]
  • Kuntz I.D. 1992. Structure-based strategies for drug design and discovery. Science 257: 1078–1082. [PubMed]
  • Lazennec C. and Meinnel T. 1997. Formate dehydrogenase-coupled spectrophotometric assay of peptide deformylase. Anal. Biochem. 244: 180–182. [PubMed]
  • Lee M.D., Antczak C., Li Y., Sirotnak F.M., Bornmann W.G., Scheinberg D.A. 2003. A new human peptide deformylase inhibitable by actinonin. Biochem. Biophys. Res. Commun. 312: 309–315. [PubMed]
  • Legrain P. and Strosberg D. 2002. Protein interaction domain mapping for the selection of validated targets and lead compounds in the anti-infectious area. Curr. Pharm. Des. 8: 1189–1198. [PubMed]
  • Li H., Gao Z., Kang L., Zhang H., Yang K., Yu K., Luo X., Zhu W., Chen K., Shen J. et al. 2006. TarFisDock: A web server for identifying drug targets with docking approach. Nucleic Acids Res. (in press). [PMC free article] [PubMed]
  • Margolis P., Hackbarth C., Lopez S., Maniar M., Wang W., Yuan Z., White R., Trias J. 2001. Resistance of Streptococcus pneumoniae to deformylase inhibitors is due to mutations in. defB. Antimicrob. Agents Chemother. 45: 2432–2435. [PMC free article] [PubMed]
  • Mazel D., Pochet S., Marliere P. 1994. Genetic characterization of polypeptide deformylase, a distinctive enzyme of eubacterial translation. EMBO J. 13: 914–923. [PMC free article] [PubMed]
  • Meinnel T. and Blanquet S. 1994. Characterization of the Thermus thermophilus locus encoding peptide deformylase and methionyl-tRNA(fMet) formyltransferase. J. Bacteriol. 176: 7387–7390. [PMC free article] [PubMed]
  • Morris G.M., Goodsell D.S., Halliday R.S., Huey R., Hart W.E., Belew R.K., Olson A.J. 1998. Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. J. Comput. Chem. 19: 1639–1662.
  • Nguyen K.T., Hu X., Colton C., Chakrabarti R., Zhu M.X., Pei D. 2003. Characterization of a human peptide deformylase: Implications for antibacterial drug design. Biochemistry 42: 9952–9958. [PubMed]
  • Pflugrath J.W. 1999. The finer things in X-ray diffraction data collection. Acta Crystallogr. D Biol. Crystallogr. 55: 1718–1725. [PubMed]
  • Ragusa S., Blanquet S., Meinnel T. 1998. Control of peptide deformylase activity by metal cations. J. Mol. Biol. 280: 515–523. [PubMed]
  • Rajagopalan P.T. and Pei D. 1998. Oxygen-mediated inactivation of peptide deformylase. J. Biol. Chem. 273: 22305–22310. [PubMed]
  • Rajagopalan P.T., Datta A., Pei D. 1997. Purification, characterization, and inhibition of peptide deformylase from Escherichia coli. Biochemistry 36: 13910–13918. [PubMed]
  • Ray S.S., Bonanno J.B., Rajashankar K.R., Pinho M.G., He G., De Lencastre H., Tomasz A., Burley S.K. 2002. Cocrystal structures of diaminopimelate decarboxylase: Mechanism, evolution, and inhibition of an antibiotic resistance accessory factor. Structure 10: 1499–1508. [PubMed]
  • Rockey W.M. and Elcock A.H. 2005. Rapid computational identification of the targets of protein kinase inhibitors. J. Med. Chem. 48: 4138–4152. [PubMed]
  • Scriven F., Wlasichuk K.B., Palcic M.M. 1988. A continual spectrophotometric assay for amino acid decarboxylases. Anal. Biochem. 170: 367–371. [PubMed]
  • Serero A., Giglione C., Sardini A., Martinez-Sanz J., Meinnel T. 2003. An unusual peptide deformylase features in the human mitochondrial N-terminal methionine excision pathway. J. Biol. Chem. 278: 52953–52963. [PubMed]
  • Stockwell B.R. 2004. Exploring biology with small organic molecules. Nature 432: 846–854. [PMC free article] [PubMed]
  • Tee W., Lambert J.R., Dwyer B. 1995. Cytotoxin production by Helicobacter pylori from patients with upper gastrointestinal tract diseases. J. Clin. Microbiol. 33: 1203–1205. [PMC free article] [PubMed]
  • Ulmer H.J., Beckerling A., Gatz G. 2003. Recent use of proton pump inhibitor-based triple therapies for the eradication of H. pylori: A broad data review. Helicobacter 8: 95–104. [PubMed]
  • Wang S., Sim T.B., Kim Y.S., Chang Y.T. 2004. Tools for target identification and validation. Curr. Opin. Chem. Biol. 8: 371–377. [PubMed]
  • Weiner S.J., Kollman P.A., Nguyen D.T., Case D.A. 1986. An all atom forcefield for simulations of proteins and nucleic acid. J. Comput. Chem. 7: 230–252.
  • Yoon H.J., Kim H.L., Lee S.K., Kim H.W., Kim H.W., Lee J.Y., Mikami B., Suh S.W. 2004. Crystal structure of peptide deformylase from Staphylococcus aureus in complex with actinonin, a naturally occurring antibacterial agent. Proteins 57: 639–642. [PubMed]
  • Yue J., Xu J., Zhao Y., Sun H., Lin Z. 1997. Chemical components from Ceratostigma willmottianum. J. Nat. Prod. 60: 1031–1033.
  • Zhou Z., Song X., Li Y., Gong W. 2004. Unique structural characteristics of peptide deformylase from pathogenic bacterium Leptospira interrogans. J. Mol. Biol. 339: 207–215. [PubMed]

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...