Abstract

Natural product ovalicin and its synthetic derivative TNP-470 have been extensively studied for their antiangiogenic property, and the later reached phase 3 clinical trials. They covalently modify the conserved histidine in Type 2 methionine aminopeptidases (MetAPs) at nanomolar concentrations. Even though a similar mechanism is possible in Type 1 human MetAP, it is inhibited only at millimolar concentration. In this study, we have discovered two Type 1 wild-type MetAPs (Streptococcus pneumoniae and Enterococcus faecalis) that are inhibited at low micromolar to nanomolar concentrations and established the molecular mechanism. F309 in the active site of Type 1 human MetAP (HsMetAP1b) seems to be the key to the resistance, while newly identified ovalicin sensitive Type 1 MetAPs have a methionine or isoleucine at this position. Type 2 human MetAP (HsMetAP2) also has isoleucine (I338) in the analogous position. Ovalicin inhibited F309M and F309I mutants of human MetAP1b at low micromolar concentration. Molecular dynamics simulations suggest that ovalicin is not stably placed in the active site of wild-type MetAP1b before the covalent modification. In the case of F309M mutant and human Type 2 MetAP, molecule spends more time in the active site providing time for covalent modification.

Introduction

Fumagillin and ovalicin class of natural products have been explored as antiangiogenic agents [13]. TNP-470, a synthetic derivative reached phase 3 clinical trials, though did not go further due to neurotoxicity [47]. These molecules target Type 2 methionine aminopeptidase (MetAP2) [8,9]. Biochemical and structural studies suggest that they covalently modify a conserved histidine in the entrance of the active site through one of its two epoxide groups [9]. In spite of having the conserved histidine in Type 1 MetAP, these molecules inhibit only at a concentration above 2 mM, while it inhibits Type 2 MetAPs at low nanomolar concentrations [10,11]. There have been efforts to explain the molecular basis for selectivity [12,13]. Crystal structure of ovalicin in complex with Type 1 human MetAP did not provide satisfactory clues to the anomaly [13]. A threonine (T334) that is behind the active site was proposed to be responsible for resistance in the Type 1 human MetAP [12].

MetAPs are present in every living cell and remove initiator methionine from ∼60 to 70% of all new proteins synthesized [14]. Knockout or chemical inhibition of these enzymes is shown to be detrimental to both prokaryotes and eukaryotes [1517]. MetAPs are classified as Type 1 and Type 2. Type 1 is further sub-classified as Types 1a, 1b, 1c and 1d [14,18,19]. All prokaryotes contain at least one Type 1 enzyme while eukaryotes have three; Type 1b and Type 2 in cytosol, and Type 1d in mitochondria [2023]. Knockdown of either of the cytosolic proteins results in a slow growth phenotype of the yeast while results in a non-viable strain when both were knocked down [16].

Because of their importance in cell survival, several groups targeted MetAPs using small molecule inhibitors for antimicrobial and antiproliferative properties [2426]. Most of the reported studies have considered inhibitors against Type 1 versus Type 2. In the previous study, we have identified α-aminophosphonate derivatives as selective inhibitors against the Type 1a Streptococcus pneumoniae MetAP (SpMetAP1a) and Enterococcus faecalis MetAP (EfMetAP1a) compared with Type 1b human enzyme [24]. This was the first study dedicated to deriving sub-Type 1 selectivity. During this study, we realized that a single amino acid difference in the S1 pocket is responsible for selectivity. Phenylalanine (F309) in the human enzyme versus methionine (M205) in the SpMetAP1a and isoleucine in EfMetAP1a at the analogous position render them with different susceptibilities against the identified inhibitors. X-ray crystal structure analysis of various phosphonate-based inhibitors in complex with human Type 1b wild-type and F309M enzyme suggest that M309 is more flexible to accommodate the small molecules compared with F309. Hence, S. pneumoniae and E. faecalis proteins accommodate inhibitors of various sizes and shapes compared with their Type 1b human counterpart.

Crystal structure analysis of human MetAP1b–ovalicin complex (PDB: 2GZ5) suggests that isoprenyl side chain of the inhibitor is in very close proximity to the F309 [13]. Though this structure represents the final conformation of the molecules, we believe that F309 in HsMetAP1b resists the binding of the ovalicin by not allowing it to attain a suitable configuration in the active site for covalent reaction with the histidine. Type 2 human enzyme has isoleucine (I338) at the analogous position. Since SpMetAP1a has a methionine and an isoleucine in EfMetAP1a in the analogous position of F309 in human MetAP1b, we hypothesized that both these enzymes should be sensitive to ovalicin.

In this study, we aimed at understanding the inhibition capacity of ovalicin and its synthetic derivative TNP-470 (Figure 1A,B) on the SpMetAP1a and EfMetAP1a. In addition, we show that F309 in human MetAP1b is key for resistance against this class of inhibitors. We used site-directed mutagenesis (SDM), enzyme kinetics and protein crystallography to explain the molecular basis.

Chemical diagrams of inhibitors, and sequence and structural composition of inhibitor binding site of the enzyme.

Figure 1.
Chemical diagrams of inhibitors, and sequence and structural composition of inhibitor binding site of the enzyme.

Chemical structures of Ovalicin (A) and TNP470 (B). Partial sequence alignment of Type 1 MetAPs (C) and active site of HsMetAP1b (D). (C) Blue diamonds represent metal-binding amino acids and red stars represent substrate recognizing amino acids. Red circle represents the amino acid which shows large variation in the active site. An insert of 27 amino acids (103–129) in SpMetA1a was removed while aligning. (D) Active site of HsMetAP1b–Met complex (PDB: 4U6J) representing metal-binding and side chain recognition amino acids. Product methionine is shown in green.

Figure 1.
Chemical diagrams of inhibitors, and sequence and structural composition of inhibitor binding site of the enzyme.

Chemical structures of Ovalicin (A) and TNP470 (B). Partial sequence alignment of Type 1 MetAPs (C) and active site of HsMetAP1b (D). (C) Blue diamonds represent metal-binding amino acids and red stars represent substrate recognizing amino acids. Red circle represents the amino acid which shows large variation in the active site. An insert of 27 amino acids (103–129) in SpMetA1a was removed while aligning. (D) Active site of HsMetAP1b–Met complex (PDB: 4U6J) representing metal-binding and side chain recognition amino acids. Product methionine is shown in green.

Materials and methods

All chemicals used in this study were purchased from Sigma–Aldrich, U.S.A. and GenPro Biotech, India. Molecular biology enzymes and ladders were purchased from Thermo Fisher Scientific, New England Biolabs (NEB) and LAADH Biotech, India. Crystallization screens and plates were purchased from Hampton Research and Molecular Dimensions.

Bioinformatic analysis of various MetAPs

The amino acid sequences were retrieved from UniProt-KB [27]. Sequences were aligned using Clustal Omega (ClustalO) [28]. The alignment image was generated using ALINE software [29].

SDM, expression and purification of various MetAPs

MtMetAP1c, EfMetAP1a, SpMetAP1a, HsMetAP1b and HsMetAP1b–F309M mutant enzymes were expressed and purified as reported earlier [8,18,24,30,31]. SDM of HsMetAP1b mutants was performed by amplifying the wild-type plasmid (pET15b containing HsMetAP1b gene) with overlapping primers to create F309L, F309Y, F309I, T334A and H212A mutants. Mutations in the gene sequences were confirmed by nucleotide sequencing (1st Base DNA sequencing services). The sequence of the oligonucleotide primers and SDM procedure are provided in the Supplementary material.

Enzyme kinetics

Concentration of each of the enzyme is adjusted such that they have similar reaction rates (HsMetAp1b, 1 µM; SpMetAP1a, 5 µM; EfMetAP1a, 1 µM; MtMetAP1c, 1 µM; F309M, 0.5 µM; F309L, 1 µM; F309Y, 4 µM; F309I, 13 µM; T334A, 2 µM; H212A, 2 µM). All assays were performed in 100 µl reaction containing 50 mM HEPES (pH 7.5 for SpMetAP1a and HsMetAP1b–T334A mutant; 8.0 for EfMetAP1a, MtMetAP1c, HsMetAP1b and all other mutants of HsMetAP1b), 150 mM NaCl and CoCl2 (three molar equivalents of corresponding enzymes). The reaction was started by adding 50 µM l-Methionine 7-amido-4-methylcoumarin (Met-AMC) and continuously monitored the fluorescence of released 7-amino-4-methyl coumarin (AMC) (380 nm excitation and 460 nm emission) in a micro-plate multimode reader (TECAN, Austria). For determining the substrate specificity of mutant enzymes, AMC derivatives of different amino acids were used at a concentration of 200 µM. Optimal pH of mutants was determined by testing the activity at different pH ranging from 4 to 10 (sodium acetate pH 4.0–5.0, MES pH 5.5–6.5, HEPES pH 7.0–8.0, Tris pH 8.5–9.0 and sodium carbonate/bicarbonate pH 9.5–10.0). For determining metal specificity, the activity of all MetAPs described in this study is tested for their activity at various concentrations of different metal ions (CoCl2, MnCl2, NiCl2 and ZnCl2). Michaelis–Menten constant and other kinetic parameters were determined by using different substrate concentrations ranging from 6.25 µM to 1.6 mM. Ki values were determined according to the Dixon method [32] using three different concentrations of substrates and four different concentrations of the inhibitor. All experiments were performed in triplicate. Sigmaplot®13 (Systat Software Inc., U.S.A.) was used to determine all the kinetic parameters.

Crystallization and structure determination of complexes with ovalicin and TNP470

A modified crystallization condition (0.1 M Bis–Tris pH 6.2, 19% PEG 3350 and 5% glycerol) was used to setup crystals of HsMetAP1b and its mutants (F309M and F309L) compared with that of the previous report [8]. For obtaining inhibitor-bound structures, 1 mM compound and 0.5 mM freshly prepared CoCl2 were added to the drop containing crystals and soaked for 24 h at 25°C. Well solution with additional 20% glycerol was used as a cryoprotectant. X-ray diffraction data were collected on home source diffractometer (Rigaku MicroMax™-007 HF rotating anode X-ray generator and R-AXIS IV++ detector). All images were processed using HKL2000 v716 [33]. Since the unit cell parameters were isomorphous with the ovalicin complex of human MetAP1b (PDB ID: 2GZ5), all structures were refined using these co-ordinates after removing the inhibitor and all water molecules [8,34]. Refinement and modeling were carried out using CCP4 suite [35,36] and Coot [37]. Using MolProbity [38], the quality of the model was verified. All the model graphics were generated using PyMOL [39].

Molecular dynamics simulation

Molecular dynamics (MD) simulation, the state-of-the-art technique was performed in GROMACS 2016.4 [40] using General Amber force field on workstation with LINUX platform. Three structures, HsMetAP1b, F309M mutant and HsMetAP2 in complex with ovalicin were simulated for understanding the dynamic stability of the inhibitor in the enzyme pocket. The crystal structures of respective complexes were used for MD simulation. In all crystal structures reported to date on fumagillin/ovalicin class of molecules, one of the conserved histidine residues flips out of the active site. For simulation, this histidine was flipped out in the starting model of the protein. Both the epoxides were intact and no covalent bond was defined between the ovalicin and the enzyme. The simulation parameters for ovalicin were prepared using AmberTools 16 [41,42]. The MetAP–ovalicin complexes were solvated using TIP3P water model in a cubic box of side 6.5, 6.5 and 7.2 nM (HsMetAP1b, F309M mutant and HsMetAP2, respectively) and energy minimized using steepest descent protocol. Equilibration of the system was done in two steps as NVT ensemble and NPT ensemble, for 200 ps. Production simulation was performed at 300 K for 20 ns and the co-ordinates, velocities and energies were saved at every 10 ps, using a leap-frog integrator. All bond constraints were implemented with LINCS algorithm [43]. Short-range electrostatic and short-range van der Waals cutoff was fixed at 1.0 nm by Verlet scheme and long-range electrostatics was treated with Particle Mesh Ewald (PME) scheme [44]. The thermostat and barostat used were V-rescale [45] and Parrinello-Rahaman [46] with 2.0 and 1.0 ps time constant, respectively.

The autocorrelation function (ACF) () for the rotation of molecules as a function of time was calculated by the following equation: 
formula
Here, is the discrete time intervals and N is the number of frames over which the ACF was performed. Change in the orientation of an orthogonal vector (on a plane defined by atoms of ovalicin) was considered for ACF calculation.

Mass spectrometry analysis of TNP-470

Intact mass of TNP-470 was determined on the sample dissolved in DMSO to a stock concentration of 20 mM using LC/MS (Agilent Triple quadrupole). To understand the effect of crystallization buffer on the stability of TNP-470, it was incubated with 1:1 mixture of crystallization solution without protein for 12 h (10 μl). Organic compounds were extracted with ethyl acetate and determined the mass of the components using LC/MS (Waters e2695).

Results and discussion

Discovery of natural product fumagillin as antiangiogenic compound and MetAP2 as the molecular target has accelerated one of its synthetic derivative TNP-470 to clinical trials [4,9,10]. However, due to unexpected toxic effects, the molecule was pulled back from phase 3 clinical trials [47]. Several attempts were made to improve the molecule by modifying it with PEG, etc., without success [47,48]. Meanwhile, there have been studies to understand the molecular basis to why fumagillin and its derivatives show more than a million-fold difference in selectivity towards Type 2 compared with Type 1 MetAP even though the mechanism of inhibition is same. In a yeast-based screen for HsMetAP2 mutations, it was demonstrated that a single amino acid change (A362T) could render HsMetAP2 resistant to ovalicin [12]. Similarly, T334A mutation in HsMetAP1b rendered it to be sensitive to ovalicin. However, the crystal structure of HsMetAP1b in complex with ovalicin did not indicate how this single amino acid difference manifests in such variation in the specificity [13]. Based on our recent study on the discovery of sub-Type 1-specific phosphonate inhibitors, we believe that phenylalanine (F309) in human MetAP1b in the S1 pocket, which is in direct contact with the substrate, could hold the key. We have studied the effect of ovalicin and TNP-470 on MtMetAP1c, SpMetAP1a and EfMetAP1a which have phenylalanine, methionine and isoleucine, respectively, at the F309 analogous position of HsMetAP1b (Figure 1C,D). Similar activity has been carried out on HsMetAP1b as well. It is important to note that the threonine is conserved in all these four enzymes in the analogous position of T334 of human MetAP1b.

Selective inhibition of SpMetAP1a and EfMetAP1a by ovalicin and TNP-470

HsMetAP1b, MtMetAP1c, SpMetAP1a and EfMetAP1a were purified to homogeneity (Supplementary Figure S1A). Enzyme activity was optimized using the Met-AMC and reaction kinetics were determined and described in Table 1 and Supplementary Figure S2. All enzymes displayed processivity of the substrate within 4-fold difference. Inhibition of all four wild-type enzymes was tested using the ovalicin and TNP-470. As expected, HsMetAP1b and MtMetAP1c were not inhibited at the tested highest concentration (50 μM). However, these molecules inhibited SpMetAP1a and EfMetAP1a at 1.5 and 3.5 μM, respectively. This is the first observation that a native Type 1 MetAP is inhibited by ovalicin or TNP-470 at this low concentration. Current data suggest that threonine (T334 in HsMetAP1b), which is thought to be responsible for the resistance against ovalicin and TNP-470 may not directly influence the binding of the inhibitor [12]. Because enzymes without phenylalanine in the analogous position of F309 in HsMetAP1b are sensitive to ovalicin and TNP470, we proceeded to explore the molecular mechanism.

Table 1
Kinetic parameters of wild-type MetAPs. Kinetics parameters were determined for each enzyme using CoCl2 as a metal cofactor and Met-AMC as a substrate
 KM (µM) kcat (s−1kcat/KM (mM−1s−1Ki ovalicin (µM) Ki TNP470 (µM) 
EfMetAP1a 175.22 ± 10.17 0.69 3.92 1.49 ± 0.10 3.45 ± 0.25 
SpMetAP1a 269.23 ± 7.64 0.43 1.61 1.49 ± 0.10 2.01 ± 0.15 
HsMetAP1b 465.90 ± 20.21 4.04 8.66 ND ND 
MtMetAP1c 335.88 ± 24.12 3.00 8.92 ND ND 
 KM (µM) kcat (s−1kcat/KM (mM−1s−1Ki ovalicin (µM) Ki TNP470 (µM) 
EfMetAP1a 175.22 ± 10.17 0.69 3.92 1.49 ± 0.10 3.45 ± 0.25 
SpMetAP1a 269.23 ± 7.64 0.43 1.61 1.49 ± 0.10 2.01 ± 0.15 
HsMetAP1b 465.90 ± 20.21 4.04 8.66 ND ND 
MtMetAP1c 335.88 ± 24.12 3.00 8.92 ND ND 

Abbreviation: ND: not determined.

Inhibition of HsMetAP1b–F309 mutants by ovalicin and TNP-470

F309M, F309L, F309I, F309Y point mutations were created in the HsMetAP1b. To understand the role of T334 of HsMetAP1b to the inhibitor sensitivity, T334A mutant was made. In addition, histidine (H212) that is responsible for binding of ovalicin is mutated to alanine to confirm its role in covalent binding of the inhibitor. All mutants were purified to homogeneity (Supplementary Figure S1B) and biochemical parameters were determined (Table 2; Supplementary Figure S3). All mutants displayed ∼±2-fold activity except that F309I that has ∼9-fold lower activity towards Met-AMC as substrate compared with the wild-type enzyme. To check for the effect on cofactor specificity, in addition to CoCl2, NiCl2, MnCl2 and ZnCl2 were tested against the mutant and the wild-type enzymes (Figure 2A). While none of the enzymes were active in the presence of ZnCl2, the human wild type and the F309L lost more than 50% of the activity when the cofactor is changed from CoCl2 to MnCl2. On the other hand, F309I, F309M and H212A gained activity by almost two folds. F309Y had similar activity in the presence of either CoCl2 or MnCl2. Surprisingly, T334A has displayed better activity in the presence of NiCl2 compared with all other metal ions tested. After establishing the metal specificity, substrate specificity has been tested (Figure 3) in the presence of respective enzymes. All mutants preferred to hydrolyze only Met-AMC, though H212A displayed additional, but very mild activity towards Leu-AMC (Figure 3). To check if the mutants have any effect of pH on the optimum activity, enzyme reactions were carried out for each mutant between 4 and 10 pH range in the presence of respective metal cofactor and Met-AMC substrate. While all enzymes retained the highest activity at pH 8.0, T334A displayed better activity at pH 7.5 (Figure 2B; Supplementary Figure S4). Inhibition of all mutants of the human enzyme was tested with ovalicin and TNP470. As predicted, except for the F309Y, ovalicin and TNP-470 inhibited all other F309 mutants at low micromolar to nanomolar concentrations (Table 2). T334A mutant, in fact, was inhibited better compared with wild type as demonstrated earlier, but not as good as in the case of F309 mutants [12]. As expected, H212A was not inhibited by either of the inhibitors.

Metal and pH preferences of HsMetAP1b and its mutants.

Figure 2.
Metal and pH preferences of HsMetAP1b and its mutants.

(A) While the wild-type enzyme (A(i)) and F309L mutant (A(iii)) are activated mainly by CoCl2, F309I (A(ii)), F309M (A(iv)) and H212A (A(vii)) mutants have displayed maximal activity with MnCl2. F309Y (A(v)) is equally activated by both CoCl2 and MnCl2. In contrast, T334A mutant (A(vi)) shows maximum activity with NiCl2. ns: not significant; ****P < 0.0001 (one-way ANOVA performed with CoCl2 as control group). (B) Activity graphs of HsMetAP1b and its mutants at different pH points (complete pH range from 4 to 10 image is provided in Supplementary Figure S4). All mutants except for T334A display maximum activity at pH 8.0 like the wild-type enzyme. T334A mutant shows maximum activity at pH 7.5. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (one-way ANOVA performed with pH 8.0 as control group).

Figure 2.
Metal and pH preferences of HsMetAP1b and its mutants.

(A) While the wild-type enzyme (A(i)) and F309L mutant (A(iii)) are activated mainly by CoCl2, F309I (A(ii)), F309M (A(iv)) and H212A (A(vii)) mutants have displayed maximal activity with MnCl2. F309Y (A(v)) is equally activated by both CoCl2 and MnCl2. In contrast, T334A mutant (A(vi)) shows maximum activity with NiCl2. ns: not significant; ****P < 0.0001 (one-way ANOVA performed with CoCl2 as control group). (B) Activity graphs of HsMetAP1b and its mutants at different pH points (complete pH range from 4 to 10 image is provided in Supplementary Figure S4). All mutants except for T334A display maximum activity at pH 8.0 like the wild-type enzyme. T334A mutant shows maximum activity at pH 7.5. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (one-way ANOVA performed with pH 8.0 as control group).

Substrate preferences of HsMetAP1b and its mutants.

Figure 3.
Substrate preferences of HsMetAP1b and its mutants.

All the mutants are specific towards methionine as is the case with wild-type enzyme except for H212A with additional but mild activity towards Leu-AMC. ****P < 0.0001 (two-way ANOVA performed with Met-AMC as control group).

Figure 3.
Substrate preferences of HsMetAP1b and its mutants.

All the mutants are specific towards methionine as is the case with wild-type enzyme except for H212A with additional but mild activity towards Leu-AMC. ****P < 0.0001 (two-way ANOVA performed with Met-AMC as control group).

Table 2
Kinetic parameters of HsMetAP1b mutants. For determining kinetic parameters, Met-AMC was used as the substrate. CoCl2 is used as a metal cofactor for F309L and F309M while MnCl2 is used as a metal cofactor for F309M, F309I and H212A, and NiCl2 in the case of T334A
 KM (µM) kcat (s−1kcat/KM (mM−1s−1Ki ovalicin (µM) Ki TNP470 (µM) 
F309M 343.34 ± 13.64 5.76 16.78 0.09 ± 0.01 0.46 ± 0.04 
F309L 191.57 ± 8.22 0.62 3.23 0.18 ± 0.022 4.57 ± 0.34 
F309I 408.27 ± 15.35 0.39 0.95 7.66 ± 0.90 7.17 ± 0.73 
F309Y 391.58 ± 27.99 1.49 3.80 ND ND 
T334A 86.90 ± 2.88 0.85 9.79 51.62 ± 6.07 35.17 ± 3.91 
H212A 663.74 ± 52.77 7.85 11.83 ND ND 
 KM (µM) kcat (s−1kcat/KM (mM−1s−1Ki ovalicin (µM) Ki TNP470 (µM) 
F309M 343.34 ± 13.64 5.76 16.78 0.09 ± 0.01 0.46 ± 0.04 
F309L 191.57 ± 8.22 0.62 3.23 0.18 ± 0.022 4.57 ± 0.34 
F309I 408.27 ± 15.35 0.39 0.95 7.66 ± 0.90 7.17 ± 0.73 
F309Y 391.58 ± 27.99 1.49 3.80 ND ND 
T334A 86.90 ± 2.88 0.85 9.79 51.62 ± 6.07 35.17 ± 3.91 
H212A 663.74 ± 52.77 7.85 11.83 ND ND 

Abbreviation: ND: not determined.

Crystal structure analysis of F309M and F309L in complex with ovalicin and TNP-470

Crystallization trials were carried out on all mutants, SpMetAP1a and EfMetAP1a. However, only F309M and F309L mutants provided X-ray diffraction quality crystals in complex with ovalicin and TNP-470. X-ray diffraction data were collected on inhibitor-soaked crystals using the home source diffractometer at near-atomic resolution (Table 3). All structures were refined using the co-ordinates from ovalicin complex of wild-type human MetAP1b (PDB: 2GZ5) [13]. As expected, both the inhibitors are covalently attached to H212. All four structures were similar to the wild-type HsMetAP1b–ovalicin complex (F309M–Ova (RMSD of 0.19 Å), F309M–TNP (0.2 Å), F309L–Ova (0.19 Å), F309L–TNP (0.17 Å)) and no major structural variations were observed. Clear electron density is observed for mutations and bound inhibitors in all four structures (Figure 4A–D). Occupancy of TNP-470 in both bound structures is ∼0.8. Also in both the TNP-470 complexes, density for chloroacetyl group in the side chain of the inhibitor is not visible (Figures 1 and 4B,D). Like in all fumagillin, ovalicin and TNP-470 bound MetAP structures irrespective of Type 1 or Type 2, a conserved histidine (H310) flips away from the active site in all complexes reported here (Figure 4A–E). Only observable difference between the F309M holo form (PDB: 4U76) and ovalicin/TNP-470 bound structures is slight outward movement (0.8–1.1 Å) of the side chain of Y195 which is in close contact with the isoprenyl side chain of the inhibitor (Figure 4E). Due to the movement of Y195, the side chain of Y196 is pushed further out (0.6–1.6 Å). Since side chains of Y195 and Y196 experience small changes, surface residues that are in contact with these two residues also undergo some changes in their side chain conformations. A holo form of the F309L could not be determined. There are no major differences in the conformations of F309M holo and complex forms suggesting the structural role of this mutant is more prominent during the entry and covalent reaction of the inhibitor rather than after the complex is formed.

X-ray crystallography structures of F309 mutants.

Figure 4.
X-ray crystallography structures of F309 mutants.

2Fo − Fc maps covering respective mutations are shown in blue density and omit maps were shown in green. (A) F309M mutant in complex with ovalicin (cyan). Omit map representing the density of ovalicin is contoured at 2.0σ. (B) F309M mutant in complex with TNP470 (peach) (Omit map at 1.0σ). (C) F309L mutant in complex with ovalicin (cyan) (Omit map at 2.0σ). (D) F309L mutant in complex with TNP470 (peach) (Omit map at 1.0σ). (E) Alignment image of F309M-apo (light green), F309M–Ova (yellow) complex, F309M–TNP470 (lavender) complex, F309L–Ova (dark green) complex and F309L–TNP470 (pink) complex showing their structural differences among them. (F) Alignment of all available structures of F309M mutant (lavender) (4U6Z, 4U70, 4U71, 4U73, 4U75, 4U76, 5YKP and 5YR4) with equivalent number of wild-type HsMetAP1b structures (gray) (2B3K, 2B3L, 2GZ5, 4U1B, 4U6C, 4U6E, 4U6J and 4U69). The ovalicin complexes of F309M mutant (green) and wild-type HsMetAP1b (pink) are highlighted.

Figure 4.
X-ray crystallography structures of F309 mutants.

2Fo − Fc maps covering respective mutations are shown in blue density and omit maps were shown in green. (A) F309M mutant in complex with ovalicin (cyan). Omit map representing the density of ovalicin is contoured at 2.0σ. (B) F309M mutant in complex with TNP470 (peach) (Omit map at 1.0σ). (C) F309L mutant in complex with ovalicin (cyan) (Omit map at 2.0σ). (D) F309L mutant in complex with TNP470 (peach) (Omit map at 1.0σ). (E) Alignment image of F309M-apo (light green), F309M–Ova (yellow) complex, F309M–TNP470 (lavender) complex, F309L–Ova (dark green) complex and F309L–TNP470 (pink) complex showing their structural differences among them. (F) Alignment of all available structures of F309M mutant (lavender) (4U6Z, 4U70, 4U71, 4U73, 4U75, 4U76, 5YKP and 5YR4) with equivalent number of wild-type HsMetAP1b structures (gray) (2B3K, 2B3L, 2GZ5, 4U1B, 4U6C, 4U6E, 4U6J and 4U69). The ovalicin complexes of F309M mutant (green) and wild-type HsMetAP1b (pink) are highlighted.

Table 3
Crystallographic data collection and refinement statistics
 F309M–Ova F309M–TNP F309L–Ova F309L–TNP 
Cell parameters 
 Space group P21 P21 P21 P21 
a (Å) 47.71 47.21 47.39 47.44 
b (Å) 77.40 77.24 77.34 77.33 
c (Å) 47.49 47.17 47.46 47.76 
 β (°) 91.62 92.03 87.95 91.7 
 Mol/ASUa 
Data collection 
 Resolution (Å)b 23.73–1.67 (1.73–1.67) 23.57–1.82 (1.89–1.82) 31.23–1.60 (1.657–1.60) 25.60–1.75 (1.81–1.75) 
 Reflections 141471 110139 161299 110325 
 Uniqueb 39512 (3648) 30217 (2948) 45153 (4513) 34885 (3311) 
 Completeness (%)b 99.21 (92.72) 99.73 (97.81) 99.92 (99.62) 99.40 (95.39) 
 Redundancy 3.6 3.6 3.6 3.2 
 Mean I/σ (I)b 8.23 (2.08) 10.70 (2.16) 9.35 (2.70) 8.64 (2.15) 
Rmerge (%)b 7.02 (50.98) 5.86 (56.43) 5.62 (43.10) 6.05 (50.74) 
Rpim (%)b 4.29 (32.96) 3.57 (35.70) 3.41 (0.28) 3.85 (34.19) 
 CC1/2b 0.997 (0.803) 0.998 (0.787) 0.998 (0.844) 0.998 (0.942) 
 Wilson B-factor (Å227.04 29.41 24.24 29.4 
Refinement statistics 
Rwork (%)b 17.84 (31.94) 17.86 (30.75) 17.69 (27.55) 18.38 (33.85) 
Rfree (%)b 21.33 (32.47) 23.16 (35.61) 20.66 (31.13) 21.29 (41.20) 
No. of atoms 2691 2623 2709 2625 
 Protein 2464 2425 2479 2431 
 Water 212 164 205 168 
RMS deviationsc 
 Bond lengths (Å) 0.015 0.013 0.014 0.016 
 Bond angles (°) 1.85 1.55 1.74 1.56 
Ramachandran analysis 
 Favored (%) 98.02 97.35 98.01 98.01 
 Outliers (%) 0.33 0.33 0.33 0.33 
 Average B-factor all atoms (Å230.01 32.67 27.14 33.30 
PDB ID 5YKP 5YR4 5YR5 5YR6 
 F309M–Ova F309M–TNP F309L–Ova F309L–TNP 
Cell parameters 
 Space group P21 P21 P21 P21 
a (Å) 47.71 47.21 47.39 47.44 
b (Å) 77.40 77.24 77.34 77.33 
c (Å) 47.49 47.17 47.46 47.76 
 β (°) 91.62 92.03 87.95 91.7 
 Mol/ASUa 
Data collection 
 Resolution (Å)b 23.73–1.67 (1.73–1.67) 23.57–1.82 (1.89–1.82) 31.23–1.60 (1.657–1.60) 25.60–1.75 (1.81–1.75) 
 Reflections 141471 110139 161299 110325 
 Uniqueb 39512 (3648) 30217 (2948) 45153 (4513) 34885 (3311) 
 Completeness (%)b 99.21 (92.72) 99.73 (97.81) 99.92 (99.62) 99.40 (95.39) 
 Redundancy 3.6 3.6 3.6 3.2 
 Mean I/σ (I)b 8.23 (2.08) 10.70 (2.16) 9.35 (2.70) 8.64 (2.15) 
Rmerge (%)b 7.02 (50.98) 5.86 (56.43) 5.62 (43.10) 6.05 (50.74) 
Rpim (%)b 4.29 (32.96) 3.57 (35.70) 3.41 (0.28) 3.85 (34.19) 
 CC1/2b 0.997 (0.803) 0.998 (0.787) 0.998 (0.844) 0.998 (0.942) 
 Wilson B-factor (Å227.04 29.41 24.24 29.4 
Refinement statistics 
Rwork (%)b 17.84 (31.94) 17.86 (30.75) 17.69 (27.55) 18.38 (33.85) 
Rfree (%)b 21.33 (32.47) 23.16 (35.61) 20.66 (31.13) 21.29 (41.20) 
No. of atoms 2691 2623 2709 2625 
 Protein 2464 2425 2479 2431 
 Water 212 164 205 168 
RMS deviationsc 
 Bond lengths (Å) 0.015 0.013 0.014 0.016 
 Bond angles (°) 1.85 1.55 1.74 1.56 
Ramachandran analysis 
 Favored (%) 98.02 97.35 98.01 98.01 
 Outliers (%) 0.33 0.33 0.33 0.33 
 Average B-factor all atoms (Å230.01 32.67 27.14 33.30 
PDB ID 5YKP 5YR4 5YR5 5YR6 
a

Number of molecules per asymmetric unit.

b

Values in parentheses refer to the highest resolution shell.

c

Root-mean-square deviations from ideal values.

To understand the possible role of methionine in allowing the ovalicin in the active site of MetAP1, we compared six random structures each of wild-type human MetAP1b and F309M which includes ovalicin complexes (Figure 4F). The aromatic rings of F309 in different structures have clustered together with the maximum displacement of 0.2 Å while the Cε atom of methionine has a maximum displacement of 2.5 Å. This suggests that while ovalicin enters the active site methionine could adjust for the demands of the inhibitor, which is not possible in the case of phenylalanine. In addition, we have analyzed role of all active site residues by understanding their interactions with ovalicin or TNP-470 in complex with wild-type enzyme of F309M using LigPlot+ (Supplementary Figure S5) [49]. Except for differences at F309(M), no changes were noticed with other residues further confirming the importance of this residue in the binding of ovalicin and its derivative TNP-470.

MD simulation

Crystal structures of the MetAP and ovalicin/TNP-470 represent the snapshot of the complex formed after the covalent reaction. These structures do not provide a clue to the status of the active site residues while ovalicin/TNP-470 enters the active site and takes the position for covalent reaction with H212. To understand this situation, we have carried out unconstrained MD simulations of the wild-type human MetAP1b, F309 mutant and human MetAP2 in complex with ovalicin. The inhibitor was placed in the active site without the covalent bond and epoxide intact. The visual inspection (Supplementary Video v01, Video v02 and Video v03) of the trajectory suggests that the HsMetAP1b–ovalicin complex is not stable and the ovalicin tends to move away from the active site. The F309M–ovalicin complex is relatively stable than the wild-type ovalicin complex. As expected, HsMetAP2–ovalicin complex is comparatively stable. This result is quantitatively understood by plotting the rotational ACF (ACFrot) () of the ovalicin during simulation (Figure 5A). The ACFrot decay of ovalicin in the HsMetAP1b–ovalicin complex is the fastest whereas the decay is similar in F309M–ovalicin and HsMetAP2–ovalicin complex (Figure 5B–D). However, it was also noted that the RMSD (root-mean-square deviation) (Supplementary Figure S6) of Cα of trajectories plateaued ∼2 Å. The RMSD data suggest that the degree of freedom of rotational motion of ovalicin, as represented by ACFrot, is less affected by the thermal fluctuation in protein.

MD simulations of the wild-type human MetAP1b, F309 mutant and human MetAP2 in complex with ovalicin.

Figure 5.
MD simulations of the wild-type human MetAP1b, F309 mutant and human MetAP2 in complex with ovalicin.

(A) The rotational ACF of ovalicin in complex with HsMetAP1b (green), F309M mutant (violet) and HsMetAP2 (red) is depicted. The faster the decay, the lower is the stability of the complex. (B) Position of Ovalicin (Magenta) in HsMetAP1b after 3 ns of simulation. In this model, molecules move out of the active site shown in gray sphere. (C) Position of Ovalicin (Magenta) in F309M after 3 ns of simulation. (D) Position of Ovalicin (Magenta) in HsMetAP2 after 3 ns of simulation. The histidine that gets alkylated is marked in cyan. The phenylalanine (F309) of HsMetAP1b (B) and its equivalent amino acids in F309M (M309) mutant (C) and HsMetAP2 (I338) (D) are marked in orange. HsMetAP1b–Ova complex is most unstable, hence, ovalicin moves out of active site earlier than the other two enzymes.

Figure 5.
MD simulations of the wild-type human MetAP1b, F309 mutant and human MetAP2 in complex with ovalicin.

(A) The rotational ACF of ovalicin in complex with HsMetAP1b (green), F309M mutant (violet) and HsMetAP2 (red) is depicted. The faster the decay, the lower is the stability of the complex. (B) Position of Ovalicin (Magenta) in HsMetAP1b after 3 ns of simulation. In this model, molecules move out of the active site shown in gray sphere. (C) Position of Ovalicin (Magenta) in F309M after 3 ns of simulation. (D) Position of Ovalicin (Magenta) in HsMetAP2 after 3 ns of simulation. The histidine that gets alkylated is marked in cyan. The phenylalanine (F309) of HsMetAP1b (B) and its equivalent amino acids in F309M (M309) mutant (C) and HsMetAP2 (I338) (D) are marked in orange. HsMetAP1b–Ova complex is most unstable, hence, ovalicin moves out of active site earlier than the other two enzymes.

The MD simulation results corroborate well with the experimental results. The highly unstable complex of HsMetAP1b–ovalicin non-covalent complex explains the least sensitivity of this enzyme towards the inhibitor. Relative stabilities of F309M–ovalicin and HsMetAP2–ovalicin complexes confirm their sensitivities. It could be inferred from the simulation data that a stable MetAP–ovalicin complex formation is required before the covalent bond could be formed. The probability of effective collision for the product formation is high if the ovalicin stays in the active site. A close look at F309 in the wild-type HsMetAP1b–ovalicin MD simulation, in the wild-type enzyme, phenylalanine side chain moves the least while the methionine in F309M–ovalicin and isoleucine (I338) in HsMetAP2–ovalicin undergo large movements to accommodate the inhibitor (Supplementary Video v01, Video v02 and Video v03). Although it could be a concerted movement of several residues in the active site, F309 or its analogous residue in other enzymes seems to be critical for the stability of the enzyme–inhibitor complex which further manifests into a covalent complex.

TNP-470 and its metabolism to AGM-1883

In both the structures of TNP-470, there is no electron density for 2-chloroacetyl group in the TNP-470 (Figure 4B,D). This molecule is identical with AGM-1883, which is a major metabolite of TNP-470 identified during the clinical trials [4]. TNP-470 metabolism to AGM-1883 was noticed within 20 min of i.v. infusion. Typical pH of the human blood is 7.3–7.4. Crystals of the human mutant MetAP were grown in slightly acidic conditions (0.1 M Bis-Tris pH 6.2, 19% PEG 3350 and 5% glycerol). AGM-1883 was noticed after the incubation of TNP-470 for 12 h in the crystallization solution without the enzyme (Supplementary Figure S7A). Together, these data suggest that chloroacetylcarbamoyl is unstable and undergoes hydrolysis even under slightly acidic conditions. Mass spectral analysis of the inhibitor suggests that this group was intact before soaking (Supplementary Figure S7B). Given the sensitivity of the TNP-470, we re-looked at the crystal structures of TNP-470 in complex with human MetAP-2 (PDB: 1B6A) [9] and Encephalitozoon cuniculi MetAP2 (PDB: 3FMR) [50]. A close look at the densities suggests that in both the cases chloroacetyl group is partially hydrolyzed (Supplementary Figure S8).

Conclusions

In this study, we have discovered new sub-Type 1 MetAP enzymes that are sensitive to ovalicin and TNP-470. Phenylalanine in the active site of the Type 1 MetAP like human MetAP1b seems to block the proper orientation of the inhibitor for covalent modification of the active site histidine. We have identified SpMetAP1a and EfMetAP1a enzymes where the phenylalanine is replaced by a methionine or isoleucine to be sensitive for ovalicin and TNP-470. Mutation of F309 in human MetAP1b to a methionine, leucine or isoleucine renders it to be as sensitive as natural enzymes with this variation. Crystal structures and MD studies have provided a clue to the sensitivity. Hence, we conclude that F309 in the human MetAP1b holds the key for the resistance of this enzyme to the ovalicin and TNP-470.

Abbreviations

     
  • ACF

    autocorrelation function

  •  
  • EfMetAP1a

    Enterococcus faecalis MetAP1a

  •  
  • HsMetAP1b

    Type 1 human MetAP

  •  
  • HsMetAP2

    Type 2 human MetAP

  •  
  • MD

    molecular dynamics

  •  
  • Met-AMC

    l-Methionine 7-amido-4-methylcoumarin

  •  
  • MetAP2

    Type 2 methionine aminopeptidase

  •  
  • SDM

    site-directed mutagenesis

Author Contribution

V.P., T.A. and A.A. designed the experiments. V.P. crystallized, collected and solved the structures of F309L mutant and carried out all biochemical assays. T.A. crystallized, collected and solved the structures of F309M mutant. N.H. carried out MD simulations. S.C.B. expressed and purified H212A mutant. A.K.M. assisted in X-ray data collection. A.A. supervised all experimental work. V.P. and A.A. wrote the manuscript.

Funding

The current study is supported by research grant from Department of Science & Technology (DST), New Delhi, India (EMR/2015/000461).

Acknowledgements

V.P. and T.A. thank University Grants Commission, New Delhi, India for fellowship. N.H. thanks Department of Science & Technology (DST) for postdoctoral fellowship. S.C.B. thanks Department of Biotechnology, New Delhi, India for fellowship. A.K.M. thanks Council for Scientific and Industrial Research, New Delhi, India for fellowship.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Ingber
,
D.
,
Fujita
,
T.
,
Kishimoto
,
S.
,
Sudo
,
K.
,
Kanamaru
,
T.
,
Brem
,
H.
et al.  (
1990
)
Synthetic analogues of fumagillin that inhibit angiogenesis and suppress tumour growth
.
Nature
348
,
555
557
2
Kusaka
,
M.
,
Sudo
,
K.
,
Matsutani
,
E.
,
Kozai
,
Y.
,
Marui
,
S.
,
Fujita
,
T.
et al.  (
1994
)
Cytostatic inhibition of endothelial cell growth by the angiogenesis inhibitor TNP-470 (AGM-1470)
.
Br. J. Cancer
69
,
212
216
3
Yamamoto
,
T.
,
Sudo
,
K.
and
Fujita
,
T.
(
1994
)
Significant inhibition of endothelial cell growth in tumor vasculature by an angiogenesis inhibitor, TNP-470 (AGM-1470)
.
Anticancer Res.
14
,
1
3
PMID:
[PubMed]
4
Bhargava
,
P.
,
Marshall
,
J.L.
,
Rizvi
,
N.
,
Dahut
,
W.
,
Yoe
,
J.
,
Figuera
,
M.
et al.  (
1999
)
A phase I and pharmacokinetic study of TNP-470 administered weekly to patients with advanced cancer
.
Clin. Cancer Res.
5
,
1989
1995
PMID:
[PubMed]
5
Stadler
,
W.M.
,
Kuzel
,
T.
,
Shapiro
,
C.
,
Sosman
,
J.
,
Clark
,
J.
and
Vogelzang
,
N.J.
(
1999
)
Multi-institutional study of the angiogenesis inhibitor TNP-470 in metastatic renal carcinoma
.
J. Clin. Oncol.
17
,
2541
2545
6
Herbst
,
R.S.
,
Madden
,
T.L.
,
Tran
,
H.T.
,
Blumenschein
, Jr,
G.
,
Meyers
,
C.A.
,
Seabrooke
,
L.F.
et al.  (
2002
)
Safety and pharmacokinetic effects of TNP-470, an angiogenesis inhibitor, combined with paclitaxel in patients with solid tumors: evidence for activity in non-small-cell lung cancer
.
J. Clin. Oncol.
20
,
4440
4447
7
Logothetis
,
C.J.
,
Wu
,
K.K.
,
Finn
,
L.D.
,
Daliani
,
D.
,
Figg
,
W.
,
Ghaddar
,
H.
et al.  (
2001
)
Phase I trial of the angiogenesis inhibitor TNP-470 for progressive androgen-independent prostate cancer
.
Clin. Cancer Res.
7
,
1198
1203
PMID:
[PubMed]
8
Addlagatta
,
A.
,
Hu
,
X.Y.
,
Liu
,
J.O.
and
Matthews
,
B.W.
(
2005
)
Structural basis for the functional differences between type I and type II human methionine aminopeptidases
.
Biochemistry
44
,
14741
14749
9
Liu
,
S.
,
Widom
,
J.
,
Kemp
,
C.W.
,
Crews
,
C.M.
and
Clardy
,
J.
(
1998
)
Structure of human methionine aminopeptidase-2 complexed with fumagillin
.
Science
282
,
1324
1327
10
Griffith
,
E.C.
,
Su
,
Z.
,
Niwayama
,
S.
,
Ramsay
,
C.A.
,
Chang
,
Y.H.
and
Liu
,
J.O.
(
1998
)
Molecular recognition of angiogenesis inhibitors fumagillin and ovalicin by methionine aminopeptidase 2
.
Proc. Natl Acad. Sci. U.S.A.
95
,
15183
15188
11
Lowther
,
W.T.
,
McMillen
,
D.A.
,
Orville
,
A.M.
and
Matthews
,
B.W.
(
1998
)
The anti-angiogenic agent fumagillin covalently modifies a conserved active-site histidine in the Escherichia coli methionine aminopeptidase
.
Proc. Natl Acad. Sci. U.S.A.
95
,
12153
12157
12
Brdlik
,
C.M.
and
Crews
,
C.M.
(
2004
)
A single amino acid residue defines the difference in ovalicin sensitivity between type I and II methionine aminopeptidases
.
J. Biol. Chem.
279
,
9475
9480
13
Addlagatta
,
A.
and
Matthews
,
B.W.
(
2006
)
Structure of the angiogenesis inhibitor ovalicin bound to its noncognate target, human type 1 methionine aminopeptidase
.
Protein Sci.
15
,
1842
1848
14
Bradshaw
,
R.A.
,
Brickey
,
W.W.
and
Walker
,
K.W.
(
1998
)
N-terminal processing: the methionine aminopeptidase and N alpha-acetyl transferase families
.
Trends Biochem. Sci.
23
,
263
267
15
Chang
,
S.Y.
,
McGary
,
E.C.
and
Chang
,
S.
(
1989
)
Methionine aminopeptidase gene of Escherichia coli is essential for cell growth
.
J. Bacteriol.
171
,
4071
4072
16
Li
,
X.
and
Chang
,
Y.H.
(
1995
)
Amino-terminal protein processing in Saccharomyces cerevisiae is an essential function that requires two distinct methionine aminopeptidases
.
Proc. Natl Acad. Sci. U.S.A.
92
,
12357
12361
17
Bernier
,
S.G.
,
Taghizadeh
,
N.
,
Thompson
,
C.D.
,
Westlin
,
W.F.
and
Hannig
,
G.
(
2005
)
Methionine aminopeptidases type I and type II are essential to control cell proliferation
.
J. Cell. Biochem.
95
,
1191
1203
18
Arya
,
T.
,
Kishor
,
C.
,
Saddanapu
,
V.
,
Reddi
,
R.
and
Addlagatta
,
A.
(
2013
)
Discovery of a new genetic variant of methionine aminopeptidase from Streptococci with possible post-translational modifications: biochemical and structural characterization
.
PLoS ONE
8
,
e75207
19
Giglione
,
C.
,
Fieulaine
,
S.
and
Meinnel
,
T.
(
2015
)
N-terminal protein modifications: bringing back into play the ribosome
.
Biochimie
114
,
134
146
20
Arfin
,
S.M.
,
Kendall
,
R.L.
,
Hall
,
L.
,
Weaver
,
L.H.
,
Stewart
,
A.E.
,
Matthews
,
B.W.
et al.  (
1995
)
Eukaryotic methionyl aminopeptidases: two classes of cobalt-dependent enzymes
.
Proc. Natl Acad. Sci. U.S.A.
92
,
7714
7718
21
Giglione
,
C.
,
Serero
,
A.
,
Pierre
,
M.
,
Boisson
,
B.
and
Meinnel
,
T.
(
2000
)
Identification of eukaryotic peptide deformylases reveals universality of N-terminal protein processing mechanisms
.
EMBO J.
19
,
5916
5929
22
Serero
,
A.
,
Giglione
,
C.
,
Sardini
,
A.
,
Martinez-Sanz
,
J.
and
Meinnel
,
T.
(
2003
)
An unusual peptide deformylase features in the human mitochondrial N-terminal methionine excision pathway
.
J. Biol. Chem.
278
,
52953
52963
23
Giglione
,
C.
,
Boularot
,
A.
and
Meinnel
,
T.
(
2004
)
Protein N-terminal methionine excision
.
Cell. Mol. Life Sci.
61
,
1455
1474
24
Arya
,
T.
,
Reddi
,
R.
,
Kishor
,
C.
,
Ganji
,
R.J.
,
Bhukya
,
S.
,
Gumpena
,
R.
et al.  (
2015
)
Identification of the molecular basis of inhibitor selectivity between the human and streptococcal type I methionine aminopeptidases
.
J. Med. Chem.
58
,
2350
2357
25
Kishor
,
C.
,
Arya
,
T.
,
Reddi
,
R.
,
Chen
,
X.
,
Saddanapu
,
V.
,
Marapaka
,
A.K.
et al.  (
2013
)
Identification, biochemical and structural evaluation of species-specific inhibitors against type I methionine aminopeptidases
.
J. Med. Chem.
56
,
5295
5305
26
Frottin
,
F.
,
Bienvenut
,
W.V.
,
Bignon
,
J.
,
Jacquet
,
E.
,
Vaca Jacome
,
A.S.
,
Van Dorsselaer
,
A.
et al.  (
2016
)
MetAP1 and MetAP2 drive cell selectivity for a potent anti-cancer agent in synergy, by controlling glutathione redox state
.
Oncotarget
7
,
63306
63323
27
Pundir
,
S.
,
Martin
,
M.J.
and
O'Donovan
,
C.
(
2017
)
Uniprot protein knowledgebase
.
Methods Mol. Biol.
1558
,
41
55
28
Sievers
,
F.
,
Wilm
,
A.
,
Dineen
,
D.
,
Gibson
,
T.J.
,
Karplus
,
K.
,
Li
,
W.
et al.  (
2011
)
Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega
.
Mol. Syst. Biol.
7
,
539
29
Bond
,
C.S.
and
Schüttelkopf
,
A.W.
(
2009
)
ALINE: a WYSIWYG protein-sequence alignment editor for publication-quality alignments
.
Acta Crystallogr. D Biol. Crystallogr.
65
,
510
512
30
Addlagatta
,
A.
,
Quillin
,
M.L.
,
Omotoso
,
O.
,
Liu
,
J.O.
and
Matthews
,
B.W.
(
2005
)
Identification of an SH3-binding motif in a new class of methionine aminopeptidases from Mycobacterium tuberculosis suggests a mode of interaction with the ribosome
.
Biochemistry
44
,
7166
7174
31
Kishor
,
C.
,
Gumpena
,
R.
,
Reddi
,
R.
and
Addlagatta
,
A.
(
2012
)
Structural studies of Enterococcus faecalis methionine aminopeptidase and design of microbe specific 2,2′-bipyridine based inhibitors
.
MedChemComm
3
,
1406
1412
32
Dixon
,
M.
(
1953
)
The determination of enzyme inhibitor constants
.
Biochem. J.
55
,
170
171
33
Otwinowski
,
Z.
and
Minor
,
W.
(
1997
)
Processing of X-ray diffraction data collected in oscillation mode
.
Methods Enzymol.
276
,
307
326
34
Berman
,
H.M.
,
Westbrook
,
J.
,
Feng
,
Z.
,
Gilliland
,
G.
,
Bhat
,
T.N.
,
Weissig
,
H.
et al.  (
2000
)
The protein data bank
.
Nucleic Acids Res.
28
,
235
242
35
Winn
,
M.D.
,
Ballard
,
C.C.
,
Cowtan
,
K.D.
,
Dodson
,
E.J.
,
Emsley
,
P.
,
Evans
,
P.R.
et al.  (
2011
)
Overview of the CCP4 suite and current developments
.
Acta Crystallogr. D Biol. Crystallogr.
67
,
235
242
36
Murshudov
,
G.N.
,
Vagin
,
A.A.
and
Dodson
,
E.J.
(
1997
)
Refinement of macromolecular structures by the maximum-likelihood method
.
Acta Crystallogr. D Biol. Crystallogr.
53
,
240
255
37
Emsley
,
P.
,
Lohkamp
,
B.
,
Scott
,
W.G.
and
Cowtan
,
K.
(
2010
)
Features and development of Coot
.
Acta Crystallogr. D Biol. Crystallogr.
66
,
486
501
38
Chen
,
V.B.
,
Arendall
, III,
W.B.
,
Headd
,
J.J.
,
Keedy
,
D.A.
,
Immormino
,
R.M.
,
Kapral
,
G.J.
et al.  (
2010
)
MolProbity: all-atom structure validation for macromolecular crystallography
.
Acta Crystallogr. D Biol. Crystallogr.
66
,
12
21
39
Schrodinger, LLC. The PyMOL Molecular Graphics System, Version 1.8. 2015
40
Spoel D
,
V.D.
,
Lindahl
,
E.
,
Hess
,
B.
,
Groenhof
,
G.
,
Mark
,
A.E.
and
Berendsen
,
H.J.
(
2005
)
GROMACS: fast, flexible, and free
.
J. Comput. Chem.
26
,
1701
1718
41
Wang
,
J.
,
Wang
,
W.
,
Kollman
,
P.A.
and
Case
,
D.A.
(
2006
)
Automatic atom type and bond type perception in molecular mechanical calculations
.
J. Mol. Graph. Model.
25
,
247
260
42
Wang
,
J.
,
Wolf
,
R.M.
,
Caldwell
,
J.W.
,
Kollman
,
P.A.
and
Case
,
D.A.
(
2004
)
Development and testing of a general amber force field
.
J. Comput. Chem.
25
,
1157
1174
43
Hess
,
B.
(
2008
)
P-LINCS: a parallel linear constraint solver for molecular simulation
.
J. Chem. Theory Comput.
4
,
116
122
44
Essmann
,
U.
,
Perera
,
L.
,
Berkowitz
,
M.L.
,
Darden
,
T.
,
Lee
,
H.
and
Pedersen
,
L.G.
(
1995
)
A smooth particle mesh Ewald method
.
J. Chem. Phys.
103
,
8577
8593
45
Bussi
,
G.
,
Donadio
,
D.
and
Parrinello
,
M.
(
2007
)
Canonical sampling through velocity rescaling
.
J. Chem. Phys.
126
,
014101
46
Parrinello
,
M.
and
Rahman
,
A.
(
1981
)
Polymorphic transitions in single crystals: a new molecular dynamics method
.
J. Appl. Phys.
52
,
7182
7190
47
Satchi-Fainaro
,
R.
,
Puder
,
M.
,
Davies
,
J.W.
,
Tran
,
H.T.
,
Sampson
,
D.A.
,
Greene
,
A.K.
et al.  (
2004
)
Targeting angiogenesis with a conjugate of HPMA copolymer and TNP-470
.
Nat. Med.
10
,
255
261
48
Benny
,
O.
,
Fainaru
,
O.
,
Adini
,
A.
,
Cassiola
,
F.
,
Bazinet
,
L.
,
Adini
,
I.
et al.  (
2008
)
An orally delivered small-molecule formulation with antiangiogenic and anticancer activity
.
Nat. Biotechnol.
26
,
799
807
49
Laskowski
,
R.A.
and
Swindells
,
M.B.
(
2011
)
LigPlot+: multiple ligand-protein interaction diagrams for drug discovery
.
J. Chem. Inf. Model.
51
,
2778
2786
50
Alvarado
,
J.J.
,
Nemkal
,
A.
,
Sauder
,
J.M.
,
Russell
,
M.
,
Akiyoshi
,
D.E.
,
Shi
,
W.
et al.  (
2009
)
Structure of a microsporidian methionine aminopeptidase type 2 complexed with fumagillin and TNP-470
.
Mol. Biochem. Parasitol.
168
,
158
167

Author notes

*

Present address: Department of Biochemistry and Molecular Medicine, Université de Montréal, Montréal, Canada

These authors contributed equally to this work.