Crystal Structure of IMP-2 Metallo-β-lactamase from Acinetobacter spp.

Comparison of Active-Site Loop Structures between IMP-1 and IMP-2

Abstract

IMP-2, a subclass B1 metallo-β-lactamase (MBL), is a Zn(II)-containing hydrolase. This hydrolase, involved in antibiotic resistance, catalyzes the hydrolysis of the C–N bond of the β-lactam ring in β-lactam antibiotics such as benzylpenicillin and imipenem. The crystal structure of IMP-2 MBL from Acinetobacter spp. was determined at 2.3 Å resolution. This structure is analogous to that of subclass B1 MBLs such as IMP-1 and VIM-2. Comparison of the structures of IMP-1 and IMP-2, which have an 85% amino acid identity, suggests that the amino acid substitution at position 68 on a β-strand (β3) (Pro in IMP-1 versus Ser in IMP-2) may be a staple factor affecting the flexibility of loop 1 (comprising residues at positions 60–66; EVNGWGV). In the IMP-1 structure, loop 1 adopts an open, disordered conformation. On the other hand, loop 1 of IMP-2 forms a closed conformation in which the side chain of Trp64, involved in substrate binding, is oriented so as to cover the active site, even though there is an acetate ion in the active site of both IMP-1 and IMP-2. Loop 1 of IMP-2 has a more flexible structure in comparison to IMP-1 due to having a Ser residue instead of the Pro residue at position 68, indicating that this difference in sequence may be a trigger to induce a more flexible conformation in loop 1.

β-Lactamases catalyze the hydrolysis of β-lactams, opening the β-lactam ring and rendering the antibiotics inactive. β-Lactamases are classified into four classes, A–D^1,2): Classes A, C, and D are serine enzymes that use a serine residue as a nucleophile, whereas class B consists of metallo enzymes whose active sites contain one or two Zn(II) ion(s) and are referred to as metallo-β-lactamases (MBLs). MBLs are divided into three subclasses (B1, B2, B3) based on the sequence of the Zn(II) ligands.³⁾ MBLs hydrolyze most β-lactams used currently, such as cephems and carbapenems, but not monobactam such as aztreonam. MBLs are hardly blocked by the inhibitors for serine β-lactamases, including clavulanate, sulbactam and, tazobactam.

In 1994, IMP-1 MBL, belonging to subclass B1, was first identified from Serratia marcescens and Pseudomonas aeruginosa in Japan.^4,5) Its gene, bla_IMP, encodes the IMP-1 enzyme and is integrated as a gene cassette into integrons carried by transferable plasmids.⁶⁾ Therefore, the bla_IMP gene can spread among different nosocomial pathogens horizontally. To date, at least 48 variants of IMP-type MBLs have been deposited (http://www.lahey.org/Studies) by the end of July 2014.

In 1997, an IMP-2 MBL was identified from an Acinetobacter baumannii clinical isolate AC-54/97 in Italy,⁷⁾ followed by the isolation of IMP-2-producing A. baumannii, A. lowffii, and P. aeruginosa in Japan.⁸⁾ The IMP-2 gene (bla_IMP-2) is also carried as an integron-bone gene cassette, similar to the IMP-1 gene (bla_IMP).^6,7) IMP-2 possesses approximately an 85% amino acid identity with IMP-1, and differs in 36 amino acids from IMP-1: 10 amino acid residues are clustered within the signal peptide region and the remaining 26 amino acid residues are found in the mature protein⁷⁾ (Fig. 2C). The structure of IMP-1 suggests that 4 of 26 amino acid residues predicted to be involved in substrate recognition in IMP-2 (Ser68, Gln198, Asp227, and Ser261; the amino acid residues of IMP-1 and IMP-2 are designated by their BBL number³⁾) are located in the neighborhood at its active site within a distance of ca. 9 Å (Fig. 1). The remaining 22 amino acid residues are located at the protein surface or are far from the active site.

Fig. 1. Molecular Surface Representation of IMP-1

Zn(II) ions are shown as green spheres, and loop 1 and loop 2 are shown as yellow ribbon models. The acetate ion and the mutated amino acid residues as compared to the IMP-2 sequence are shown as sticks. The amino acid residues of IMP-1 are designated by their BBL number.³⁾

The kinetic parameters of the hydrolysis of several β-lactams by IMP-2 are overall similar to those by IMP-1, but the catalytic efficiency values of the two enzymes (k_cat/K_m) for ampicillin are different⁷⁾: the k_cat/K_m values are 4.8 µM⁻¹ s⁻¹ for IMP-1 and 0.21 µM⁻¹ s⁻¹ for IMP-2.⁷⁾ The k_cat/K_m value of IMP-1 to IMP-2 increases 23-fold, so IMP-1 hydrolyses ampicillin more efficiently than IMP-2. These differences in kinetic parameters might be related to the subtle structural changes arising from the different amino acid sequences of the enzymes, even though the 6 amino acid residues (His116, His118, Asp120, His196, Cys221, and His263) which construct the active site of the enzyme are conserved between IMP-1 and IMP-2.

Therefore, determination of the fine three-dimensional structure of IMP-2 would be very useful for elucidating the mechanism underlying the difference in the substrate specificity between IMP-1 and IMP-2 in order to develop inhibitors specific for MBLs. Here, we describe the crystal structure of IMP-2 MBL from Acinetobacter spp.

MATERIALS AND METHODS

Plasmid and Reagents

The pBC SK(+) plasmid vector was purchased from Agilent Technologies, Inc. (Santa Clara, CA, U.S.A.). Ampicillin and zinc(II) nitrate hexahydrate were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). 2-[4-(2-Hydroxyethyl)-1-piperazinyl]ethanesulfonic acid (HEPES) was purchased from Dojindo Laboratories (Kumamoto, Japan). Tris(hydroxymethyl)aminomethane (Tris) was purchased from Nacalai Tesque (Kyoto, Japan). Polyethylene glycol 4000 (PEG 4000) was purchased from Hampton Research (Aliso Viejo, CA, U.S.A.). All other reagents were of the highest grade commercially available.

Expression and Purification

The IMP-2 enzyme was expressed in Escherichia coli HB101 harboring pBC SK(+) vector carrying the bla_IMP-2 gene; pBC SK(+)/bla_IMP-2. The cells were cultured in 2 L of LB broth containing ampicillin (50 µg/mL) for 14 h at 37°C, then centrifuged at 6000×g for 15 min at 4°C. The pellet was resuspended in 30 mL of 50 mM sodium phosphate buffer (pH 7.0) containing 10 µM Zn(NO₃)₂. The cells were disrupted by sonication, then centrifuged at 105000×g for 75 min at 4°C. The supernatant was purified by column chromatography. Cation exchange chromatography was performed using a SP Sepharose Fast Flow column (ϕ26 mm×100 cm, GE Healthcare UK Ltd., Little Chalfont, U.K.) pre-equilibrated with 50 mM sodium phosphate buffer (pH 7.0) containing 10 µM Zn(NO₃)₂. Bound proteins were eluted with a linear gradient of 0 to 0.3 M NaCl in 50 mM sodium phosphate buffer (pH 7.0) containing 10 µM Zn(NO₃)₂. Fractions exhibiting β-lactamase activity were collected, pooled, and concentrated by ultrafiltration with an Amicon YM-10 (Merck KGaA, Darmstadt, Germany). Then, the sample buffer was exchanged with 50 mM Tris–HCl buffer (pH 7.4) containing 0.3 M NaCl, followed by concentration by ultrafiltration with a Centricon YM-10 (Merck KGaA) to 2 mL. The concentrated samples were applied to a gel filtration column (Sephacryl HR-100, ϕ16 mm×80 cm, GE Healthcare), pre-equilibrated with 50 mM Tris–HCl buffer (pH 7.4) containing 0.3 M NaCl. Fractions exhibiting β-lactamase activity were collected, pooled, concentrated by ultrafiltration using an Amicon YM-10 (Merck KGaA), and then stored at −80°C. The purity of the preparation was examined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE); the final preparation showed a single band using Comassie Brilliant Blue (CBB) dye, indicating more than 95% purity. For crystallization of the purified IMP-2 enzyme, the protein buffer was exchanged with 20 mM HEPES–NaOH (pH 7.5) using an Amicon Ultra (Merck KGaA).

Crystallization

Initial screening of IMP-2 crystallization conditions was performed using the hanging drop method at 293 K by referring to the IMP-1 crystallization conditions.⁹⁾ Drops prepared by mixing 3 µL of protein solution (5 mg/mL) with 3 µL of reservoir solution, and were equilibrated against 350 µL of reservoir solution in the well. Crystals of IMP-2 were appeared after one month using a reservoir solution consisting of 30% (w/v) PEG 4000, 0.1 M citric acid/sodium citrate buffer containing 0.2 M sodium acetate (pH 6.0).

Data Collection and Refinement

X-Ray diffraction data were collected in house X-ray diffraction system. CuKα X-ray radiation from a rotating-anode X-ray generator (Rigaku Micro Max007, Rigaku Corporation, Tokyo, Japan) and an imaging-plate detector (Rigaku R-AXIS VII) were used. Crystals could be flash-cooled at 100 K in a stream of cold nitrogen without cryoprotectant to avoid crystal cracking. Diffraction data from IMP-2 crystals were collected to 2.30 Å resolution. The diffraction data sets were processed using program CrystalClear (Rigaku Corporation). The crystallographic statistics of the collected data are summarized in Table 1.

Table 1. Crystallographic Data Collection and Refinement Statistics for IMP-2

Data collection
Resolution (Å)	44.2–2.30 (2.38–2.30)a)
Wavelength (Å)	1.5418
Cell dimensions
a, b, and c (Å)	37.9, 68.5, 88.3
α, β, and γ (°)	90.0, 90.0, 90.0
Space group	P212121
Redundancy	6.81 (6.62)
Completeness (%)	99.8 (100.0)
Rmergeb)	0.095 (0.255)
No. of observed reflections	73284 (7065)
No. of unique reflections	10767 (1068)
〈I/(σ)〉	6.6 (2.3)
Refinement statistics
σ Cutoff	None
Resolution (Å)	44.2–2.30 (2.63–2.30)
No. of reflections used	9994 (751)
B factors (Å2)
Average	31.9
Protein	31.8
Ligand	26.6
Water	34.1
No. of non-H atomsc)
Protein	1717
Ligand	5
Water	113
R.m.s.d deviation from ideald)
Bond lengths (Å)	0.009
Angles (deg.)	1.17
Rworkinge)	0.232 (0.276)
Rfreef)	0.299 (0.265)

a) Values in parentheses are for the highest resolution shell. b) R_merge=∑_hkl ∑_i |I_i(hkl)−〈I(hkl)〉|/∑_hkl ∑_iI_i(hkl), where I_i(hkl) is the observed intensity for reflection hkl and 〈I(hkl)〉 is the average intensity calculated for reflection j from replicate data. c) Per asymmetric unit. d) R.m.s.d: root-mean-square-deviation. e) R_working=∑_hkl ||F_o|−|F_c||/∑_hkl |F_o|, where F_o and F_c are the observed and calculated structure factors, respectively. f) R_free=∑_hkl ||F_o|−|F_c||/∑_hkl |F_o| for 5% of the data not used at any stage of structural refinement.

The structure of IMP-2 was solved by molecular replacement using the program Morlep¹⁰⁾ of the CCP4 suite ver. 6.3¹¹⁾ using the structure of IMP-1 from P. aeruginosa (PDB code: 1DD6) as a search model. The initial model was refined with REFMAC 5.5¹²⁾ in the CCP4 suite¹¹⁾ using resolution limits of 44.2–2.30 Å. Water molecules were added using Coot 0.7¹³⁾ selected from peaks in the 2|F_o|−|F_c| difference density map (σ=1.8). The final model had an R_working factor of 23.2% and an R_free factor of 29.9%. The quality of the final model was checked with RAMPAGE (http://mordred.bioc.cam.ac.uk/~rapper/rampage.php).¹⁴⁾ The Ramachandran plot showed 95.8% (207 residues) of the residues in the favoured region and 3.7% (8 residues) of the residues in the allowed region, while 0.5% (1 residue; Asp84) was in the outlier region.

All structural figures were prepared using the program PyMOL v0.99rc6.¹⁵⁾

PDB Accession Code

Coordinates and structural factors have been deposited in the PDB under the accession code: 4UBQ

RESULTS AND DISCUSSION

Overall Structure of IMP-2

The final refined model of IMP-2 per asymmetric unit included one IMP-2 molecule consisting of residues Leu39–Lys298, two Zn(II) ions, 113 water molecules, and one acetate ion. The overall structure of IMP-2 adopts an αβ/βα sandwich structure, with an interface comprising two central antiparallel β-strands surrounded by two α-helices (Fig. 2A), similar to the structural fold in other subclass B1 MBLs such as IMP-1,⁹⁾ CcrA,¹⁶⁾ IND-7,¹⁷⁾ and VIM-2.¹⁸⁾ The N-terminal domain consists of four α-helices (α1–α4) and six antiparallel β-strands (β1–β6), whereas the C-terminal domain is formed by two α-helices (α5 and α6) and five antiparallel β-strands (β7–β11). The active site of IMP-2 contains two Zn(II) ions (Zn1, Zn2) separated by 3.2 Å and is located at the bottom of a wide, shallow cleft enclosed by two extended loops (loop 1, loop 2, Fig. 2A). Loop 1, a β-turn connected by two antiparallel β-strands (β2, β3), comprises residues 60–66 (EVNGWGV) (Fig. 2C). Loop 1 is likely involved in the binding of substrates or inhibitors.^19,20) Loop 2, which connects a strand (β10) and a helix (α5), is composed of residues 224–240 (Fig. 2) and is located on approximately the opposite side of loop 1 centered around the Zn(II) ion-binding site. Lys224 and Asn233 on loop 2 participated in substrate and inhibitor binding.^9,21,22)

Fig. 2. A) Overall Structure of IMP-2 from Acinetobacter spp.

α-Helices, β-strands, loops, and Zn(II) ions are shown in red, green, yellow, and orange, respectively. B) Superposition of IMP-1 (orange) and IMP-2 (beige) structures. Considerable differences are observed in the loop 1 motifs in both IMP-1 and IMP-2 structures. C) Sequence alignment and secondary structures of IMP-2 from Acinetobacter spp. with that of IMP-1 from Serratia marcescens using the PDB file, 1DDK (IMP-1), and the structure from this study (IMP-2). References for each sequence are as follows: IMP-1 (EMBL/GenBank/DDBJ accession number: IMP-1 (S71932)) and IMP-2 (AB182996). The figure was produced using the ESPript 3.0 program (http://espript.ibcp.fr).³⁰⁾ The BBL number is indicated above the sequences.³⁾ The dashed lines indicate the signal peptide sequences. Invariant residues are shown in red columns and conserved residues are shown in boxes. The arrows indicate β-sheets, the coils indicate α-helices, TT indicates β turns, and η indicates 3₁₀ helices. The loop 1 and loop 2 regions in IMP-1 and IMP-2 are underlined in blue.

Asp84 in IMP-2 is the outlier in the Ramachandran plot and has a sterically strained main chain conformation, with ϕ and ψ angles of 59° and 150°, respectively. The carboxylate oxygen, OD1, is hydrogen bonded to Ser115OG (3.0 Å) and Ser115N (2.8 Å), whereas OD2 is hydrogen bonded to Lys69NZ (2.7 Å), Ser115OG (3.1 Å), and Ser121OG (2.7 Å). In IMP-1, Asp84 also has a sterically strained main chain conformation in both the native and in the inhibitor complex, 2-[5-(1-tetrazoylmethyl)thien-3-yl]-N-[2-(mercaptomethyl)-4-(phenylbutyrylglycine)] with mean ϕ and ψ angles of 81° and 148°, respectively.⁹⁾ The carboxylate oxygen atoms of Asp84 in IMP-1 form hydrogen bonds to Lys69NZ (2.8 Å), Ser115N (2.8 Å), Ser115OG (2.8 Å), and Ser121OG (2.7 Å).⁹⁾ Asp84 has a common strained conformation not only in IMP-1 and IMP-2, but also in other subclass B1 MBLs.^9,16,23) Therefore, Asp84 likely plays an important role in the folding of MBLs.

Structural Comparison with IMP-1

The overall structure of IMP-2 superposed on IMP-1 (PDB code: 1DDK the structure discussed here) with a root-mean-square deviation (rmsd) of 0.55 Å (for the Cα atoms of Leu39–Gly293, Fig. 2B). Significant differences were located in loop 1 of the IMP-1 and IMP-2 structures. Different conformations for the loop 1 were observed between the two enzymes, even though there was an acetate ion in the active site of both enzymes (see discussion below).

In the IMP-1 structure, Gly63–Trp64–Gly65 (the GWG portion) located near the apex of loop 1, are disordered, and Trp64 is positioned away from the active site groove, towards the solvent.⁹⁾ NMR studies on CcrA by Scrofani et al. suggest that Trp64 of IMP-1 plays a role in recruiting and stabilizing the substrate ligand.¹⁹⁾ The conformational flexibility of the GWG portion likely creates an open cavity in the active site, allowing the accommodation of a variety of bulky substrates. In contrast, judging from the 2|F_o|−|F_c| electron density map, the backbone of the GWG portion in IMP-2 is in a single conformation with a well-defined electron density (Fig. 3A). Two antiparallel β-strands (β2, β3) in IMP-2 extend perpendicularly to the active site cleft, where the indole ring of Trp64 is situated, thus covering the active site from the upper part (Fig. 3A). Residues 60–66 in loop 1 of IMP-2 are transformed from an open conformation, as seen in the IMP-1 structure (Fig. 3B),⁹⁾ to a closed conformation (Fig. 3A), resulting in a tunnel-shaped cavity in the active site.

Fig. 3. Comparison of the IMP-2 Structure and the IMP-1 Structure

α-Helices, β-strands, and loops are shown in red, green, and yellow, respectively. Zn(II) ions are shown as orange spheres. Trp64, His116, His118, Asp120, His196, Cys221, and His263 residues and an acetate ion are represented as sticks (carbon, gray; nitrogen, blue; oxygen, red; and sulfur, light green). A) Structure of the active site in IMP-2. The electron density map (cyan mesh) is shown contoured at the 1.0σ level in the 2|F_o|−|F_c| map. B) Structure of the active site in IMP-1.

Interestingly, this closed conformation of loop 1 in IMP-2 is similar to those found in the crystal structures of IMP-1 complexed with inhibitors.^9,24,25) However, the active site cleft showed no major difference between IMP-1 and IMP-2 (Fig. 2B). One structural factor that may be triggering the conformational change of loop 1 may be the nature of the residue at position 68, located between Val67 and Lys69 on a β-strand (β3) that creates part of the hydrophobic pocket for the substrate of loop 1. Position 68 in IMP-1 is Pro, which is conformationally rigid, whereas that of IMP-2 is a Ser residue. Substitution of the residue at position 68 led to changes in the dihedral angle of the adjacent Val67 (for the Cα atom of Val67: ϕ −147°, φ 122°, and ω 180° for IMP-1 and ϕ −95°, φ 149°, and ω 174° for IMP-2) and to changes in hydrogen bond formation of loop 1 between IMP-1 and IMP-2. Rotational transfer of Vla67 (ϕ: −147°; IMP-1 to −95°; IMP-2) may influence interaction with substrates. Palzkill et al. analyze the residues in or near the active site of IMP-1 by codon randomization and selection experiments^22,26,27) and suggest that Val67 is essential for ampicillin hydrolysis.^26,27)

Loop 1 of IMP-2 seems more flexible due to the lack of steric hindrance with the cyclic side chain of Pro, compared with IMP-1. Borra et al. pointed out that loop 1 of VIM-7 MBL with Ser at position 68 is more flexible than that of VIM-2, with Pro at position 68.²⁸⁾ The crystal structures of IMP-1 with and without a mercaptocarboxylate inhibitor indicate that IMP-1 takes an open conformation without an inhibitor and converts to a closed conformation upon binding of the inhibitor to the active site.⁹⁾ Such an observation is found in X-ray crystal structures of unliganded MBL from Bacteroides fragilis (CcrA) and its 4-morpholinoethanesulfonic acid (MES) complex.²⁹⁾

From the results of the IMP-2 structure although there is only one case, it is thought that IMP-2 can take a closed conformation, even when a substrate or an inhibitor is not present in the active site, because of the conformational flexibility.

Comparison of the Active Site Structure between IMP-1 and IMP-2

Zn1 in IMP-2 showed a very clear electron density and was coordinated by three His residues (His116, His118, and His196) and one acetate ion. The average bond distance between Zn1–His and the average angle for His–Zn1–His were 2.3 Å and 105°, respectively, which are almost identical to those found in IMP-1 (2.3 Å, 94°). An acetate ion in the active site of IMP-2 exhibited two alternate conformations with half-occupancy (ACT A, ACTB), with one of the two oxygen atoms in ACT A located 2.7 Å from Zn1 (IMP-1; 2.9 Å). No apparent electron density for a bridging water molecule/hydroxide ion in between Zn1 and Zn2 was observed, in contrast with the majority of other MBL structures. The coordination environment around Zn1 can be described as a distorted tetrahedral geometry, as can be seen in the Zn1 site of IMP-1.

The coordination geometry of Zn2 in IMP-2 is different from that of IMP-1. Unlike Zn1, the 2|F_o|−|F_c| electron density map at Zn2 showed the existence of partially dissociated Zn(II) ion from the active site. The occupancies for Zn1 and Zn2 were set to 1.0 and 0.3, respectively, for subsequent refinement. As a result, the final B-factors approached 33.6 Ų for Zn1 and 38.3 Ų for Zn2 (B-factor average: 35.8 Ų). This result indicates that the Zn(II) binding affinity of the Zn2 site is lower than that of the Zn1 site. Moreover, the side chain of Cys221 adopted alternate conformations, where the occupancy of Cys221A was refined by 0.3, and that of Cys221B was refined by 0.7. The former conformer was the Zn2-bound form, whereas the latter was the Zn2-unbound form. The Zn2–Cys221A and Zn2–His263 bond distances were 2.3 Å and 3.0 Å, respectively. Thus, the Zn2–Cys221A bond distance was similar to that of IMP-1 but the Zn2–His263 bond distance in IMP-2 was much longer by 0.6 Å than that of IMP-1. The side chain of Asp120 in IMP-2, the Zn2 ligand, displayed a well-defined single conformation and the Zn2–Asp120 bond distance in IMP-2 is 2.6 Å, very similar to that of IMP-1 (2.6 Å). One of the two oxygen atoms in ACTB is located 3.1 Å from Zn2, which is the same position as the apical water of plane in IMP-1. The IMP-2 ligand–Zn2–ligand bond angle of 74–104° is close to the optimal tetrahedral angles, although those of IMP-1 are 64–88°. Thus, the coordination environment around Zn2 can be described as a distorted tetrahedral geometry. In the IMP-1 structure, Zn2 is coordinated with Asp120, His196, Cys221, and one water molecule, and a bridging water/hydroxide ion (but not seen due to a low resolution), forming a trigonal bipyramidal geometry.⁹⁾ In addition, an acetate ion in IMP-1 is positioned 2.8 Å from Zn2. Thus, there is a considerable difference in the coordination geometry of the Zn2 sites between IMP-1 and IMP-2.

Another interesting difference in and near the active site is the portion of residues 261–263: IMP-2 harbors two contiguous (Ser261–Ser262) residues adjacent to the Zn2 ligand His263, whereas IMP-1 harbors Pro261–Ser262 adjacent to His263. In the crystal structure of IMP-2, the hydroxyl oxygen atom of Ser261 is hydrogen bonded to the main chain carbonyl of Ser264 (3.0 Å), indicating that the conformational freedom of this portion of the protein by the participation of this hydorgen bond is decreased relative to IMP-1. In addition, His263Nδ1 in IMP-1 and IMP-2 is hydrogen bonded to the main chain carbonyl of the residue at position 68. From these findings, we propose that the conformational flexibility of residues 261–263 may well also influence the position, mobility, or affinity of Zn2.

CONCLUSION

In conclusion, we have determined the crystal structure of a subclass B1 MBL, IMP-2. Comparison of the structures of IMP-1 and IMP-2 revealed that the substitution of the amino acid residue at position 68 (Pro in IMP-1, Ser in IMP-2) causes conformational flexibility of loop 1 (comprising residues at positions 60–66) in IMP-2 that may be responsible for substrate binding. Our data will help elucidate the correlation between substrate specificity and structural polymorphism among MBLs belonging to the IMP family. Crystallographic studies of IMP-2 complexed with the hydrolyzed product of ampicillin are in progress in order to quantitatively analyze the structure–activity relationship of IMP-2.

Acknowledgment

This work was supported by H24-Shinkou-Ippan-010 from the Ministry of Health, Labour and Welfare of Japan, and in part by JSPS KAKENHI Grant Numbers 24659059 and 10363524.

Conflict of Interest

The authors declare no conflict of interest.

REFERENCES

• 1) Ambler RP. The structure of β-lactamases. Philos. Trans. R. Soc. Lond. B Biol. Sci., 289, 321–331 (1980).
• 2) Bush K, Jacoby GA, Medeiros AA. A functional classification scheme for β-lactamases and its correlation with molecular structure. Antimicrob. Agents Chemother., 39, 1211–1233 (1995).
• 3) Galleni M, Lamotte-Brasseur J, Rossolini GM, Spencer J, Dideberg O, Frère J-M, The metallo-β-lactamase working group. Standard numbering scheme for class B β-lactamases. Antimicrob. Agents Chemother., 45, 660–663 (2001).
• 4) Watanabe M, Iyobe S, Inoue M, Mitsuhashi S. Transferable imipenem resistance in Pseudomonas aeruginosa. Antimicrob. Agents Chemother., 35, 147–151 (1991).
• 5) Osano E, Arakawa Y, Wacharotayankun R, Ohta M, Horii T, Ito H, Yoshimura F, Kato N. Molecular characterization of an enterobacterial metallo β-lactamase found in a clinical isolate of Serratia marcescens that shows imipenem resistance. Antimicrob. Agents Chemother., 38, 71–78 (1994).
• 6) Arakawa Y, Murakami M, Suzuki K, Ito H, Wacharotayankun R, Ohsuka S, Kato N, Ohta M. A novel integron-like element carrying the metallo-β-lactamase gene bla_IMP. Antimicrob. Agents Chemother., 39, 1612–1615 (1995).
• 7) Riccio ML, Franceschini N, Boschi L, Caravelli B, Cornaglia G, Fontana R, Amicosante G, Rossolini GM. Characterization of the metallo-β-lactamase determinant of Acinetobacter baumannii AC-54/97 reveals the existence of bla_IMP allelic variants carried by gene cassettes of different phylogeny. Antimicrob. Agents Chemother., 44, 1229–1235 (2000).
• 8) Shibata N, Doi Y, Yamane K, Yagi T, Kurokawa H, Shibayama K, Kato H, Kai K, Arakawa Y. PCR typing of genetic determinants for metallo-β-lactamases and integrases carried by Gram-negative bacteria isolated in Japan, with focus on the class 3 integron. J. Clin. Microbiol., 41, 5407–5413 (2003).
• 9) Concha NO, Janson CA, Rowling P, Pearson S, Cheever CA, Clarke BP, Lewis C, Galleni M, Frère J-M, Payne DJ, Bateson JH, Abdel-Meguid SS. Crystal structure of the IMP-1 metallo β-lactamase from Pseudomonas aeruginosa and its complex with a mercaptocarboxylate inhibitor: binding determinants of a potent, broad-spectrum inhibitor. Biochemistry, 39, 4288–4298 (2000).
• 10) Vagin A, Teplyakov A. MOLREP: an automated program for molecular replacement. J. Appl. Cryst., 30, 1022–1025 (1997).
• 11) Collaborative Computational Project, Number 4. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D Biol. Crystallogr., 50, 760–763 (1994).
• 12) Murshudov GN, Skubák P, Lebedev AA, Pannu NS, Steiner RA, Nicholls RA, Winn MD, Long F, Vagin AA. REFMAC5 for the refinement of macromolecular crystal structures. Acta Crystallogr. D Biol. Crystallogr., 67, 355–367 (2011).
• 13) Emsley P, Lohkamp B, Scott WG, Cowtan K. Features and Development of Coot. Acta Crystallogr. D Biol. Crystallogr., 66, 486–501 (2010).
• 14) Lovell SC, Davis IW, Arendall WB 3rd, de Bakker PIW, Word JM, Prisant MG, Richardson JS, Richardson DC. Structure validation by Cα geometry: ϕ, ψ and Cβ deviation. Proteins, 50, 437–450 (2003).
• 15) DeLano WL. The PyMOL Molecular Graphics System. DeLano Scientific, San Carlos, California, U.S.A., (2002).
• 16) Concha NO, Rasmussen BA, Bush K, Herzberg O. Crystal structure of the wide-spectrum binuclear zinc β-lactamase from Bacteroides fragilis. Structure, 4, 823–836 (1996).
• 17) Yamaguchi Y, Takashio N, Wachino J, Yamagata Y, Arakawa Y, Matsuda K, Kurosaki H. Structure of metallo-β-lactamase IND-7 from a Chryseobacterium indologenes clinical isolate at 1.65-Å resolution. J. Biochem., 147, 905–915 (2010).
• 18) Garcia-Saez I, Docquier J-D, Rossolini GM, Dideberg O. The three-dimensional structure of VIM-2, a Zn-β-lactamase from Pseudomonas aeruginosa in its reduced and oxidised form. J. Mol. Biol., 375, 604–611 (2008).
• 19) Scrofani SDB, Chung J, Huntley JJA, Benkovic SJ, Wright PE, Dyson HJ. NMR characterization of the metallo-β-lactamase from Bacteroides fragilis and its interaction with a tight-binding inhibitor: role of an active-site loop. Biochemistry, 38, 14507–14514 (1999).
• 20) Moali C, Anne C, Lamotte-Brasseur J, Groslambert S, Devreese B, Van Beeumen J, Galleni M, Frère J-M. Analysis of the importance of the metallo-β-lactamase active site loop in substrate binding and catalysis. Chem. Biol., 10, 319–329 (2003).
• 21) Yamaguchi Y, Jin W, Matsunaga K, Ikemizu S, Yamagata Y, Wachino J, Shibata N, Arakawa Y, Kurosaki H. Crystallographic investigation of the inhibition mode of a VIM-2 metallo-β-lactamase from Pseudomonas aeruginosa by a mercaptocarboxylate inhibitor. J. Med. Chem., 50, 6647–6653 (2007).
• 22) Brown NG, Horton LB, Huang W, Vongpunsawad S, Palzkill T. Analysis of the functional contributions of Asn233 in metallo-β-lactamase IMP-1. Antimicrob. Agents Chemother., 55, 5696–5702 (2011).
• 23) Carfi A, Pares S, Duée E, Galleni M, Duez C, Frère J-M, Dideberg O. The 3-D structure of a zinc metallo-β-lactamase from Bacillus cereus reveals a new type of protein fold. EMBO J., 14, 4914–4921 (1995).
• 24) Kurosaki H, Yamaguchi Y, Yasuzawa H, Jin W, Yamagata Y, Arakawa Y. Probing, inhibition, and crystallographic characterization of metallo-β-lactamase (IMP-1) with fluorescent agents containing dansyl and thiol groups. ChemMedChem, 1, 969–972 (2006).
• 25) Toney JH, Hammond GG, Fitzgerald PMD, Sharma N, Balkovec JM, Rouen GP, Olson SH, Hammond ML, Greenlee ML, Gao Y-D. Succinic acids as potent inhibitors of plasmid-borne IMP-1 metallo-β-lactamase. J. Biol. Chem., 276, 31913–31918 (2001).
• 26) Materon IC, Beharry Z, Huang W, Perez C, Palzkill T. Analysis of the context dependent sequence requirements of active site residues in the metallo-β-lactamase IMP-1. J. Mol. Biol., 344, 653–663 (2004).
• 27) Materon IC, Palzkill T. Identification of residues critical for metallo-β-lactamase function by codon randomization and selection. Protein Sci., 10, 2556–2565 (2001).
• 28) Borra PS, Leiros H-KS, Ahmad R, Spencer J, Leiros I, Walsh TR, Sundsfjord A, Samuelsen Ø. Structural and computational investigations of VIM-7: insights into the substrate specificity of VIM metallo-β-lactamases. J. Mol. Biol., 411, 174–189 (2011).
• 29) Fitzgerald PMD, Wu JK, Toney JH. Unanticipated inhibition of the metallo-β-lactamase from Bacteroides fragilis by 4-morpholineethanesulfonic acid (MES): a crystallographic study at 1.85-Å resolution. Biochemistry, 37, 6791–6800 (1998).
• 30) Corpet F. Multiple sequence alignment with hierarchical clustering. Nucleic Acids Res., 16, 10881–10890 (1988).