Role of a Mutation at Position 167 of CTX-M-19 in Ceftazidime Hydrolysis

Abstract

CTX-M-19 is a recently identified ceftazidime-hydrolyzing extended-spectrum β-lactamase, which differs from the majority of CTX-M-type β-lactamases that preferentially hydrolyze cefotaxime but not ceftazidime. To elucidate the mechanism of ceftazidime hydrolysis by CTX-M-19, the β-lactam MICs of a CTX-M-19 producer, and the kinetic parameters of the enzyme were confirmed. We reconfirmed here that CTX-M-19 is also stable at a high enzyme concentration in the presence of bovine serum albumin (20 μg/ml). Under this condition, we obtained more accurate kinetic parameters and determined that cefotaxime (k_cat/K_m, 1.47 × 10⁶ s⁻¹ M⁻¹), cefoxitin (k_cat/K_m, 62.2 s⁻¹ M⁻¹), and aztreonam (k_cat/K_m, 1.34 × 10³ s⁻¹ M⁻¹) are good substrates and that imipenem (k₊₂/K, 1.20 × 10² s⁻¹ M⁻¹) is a poor substrate. However, CTX-M-18 and CTX-M-19 exhibited too high a K_m value (2.7 to 5.6 mM) against ceftazidime to obtain their catalytic activity (k_cat). Comparison of the MICs with the catalytic efficiency (k_cat/K_m) of these enzymes showed that some β-lactams, including cefotaxime, ceftazidime, and aztreonam showed a similar correlation. Using the previously reported crystal structure of the Toho-1 β-lactamase, which belongs to the CTX-M-type β-lactamase group, we have suggested characteristic interactions between the enzymes and the β-lactams ceftazidime, cefotaxime, and aztreonam by molecular modeling. Aminothiazole-bearing β-lactams require a displacement of the aminothiazole moiety due to a severe steric interaction with the hydroxyl group of Ser167 in CTX-M-19, and the displacement affects the interaction between Ser130 and the acidic group such as carboxylate and sulfonate of β-lactams.

As β-lactam antibiotics stable against β-lactamases have been developed and used clinically, extended-spectrum β-lactamases (ESBLs) have increased in prevalence, especially since the beginning of the 1990s (2). These enzymes cause the serious clinical problem of resistance to broad-spectrum cephalosporins such as cefotaxime, ceftazidime, or ceftriaxione. A precise elucidation of the mechanism of hydrolysis by ESBLs and an understanding of the requirements for the development of new antibiotics against ESBL producers are emerging from studies of ESBL kinetics and crystal structure. ESBLs can be classified into nine types (TEM, SHV, CTX-M, PER, VEB, GES, TLA, BES, and OXA) (6). TEM- and SHV-type ESBLs have been the most investigated. Point mutations in these ESBLs play an important role in the interactions between the enzymes and β-lactam antibiotics (3, 8, 14).

CTX-M-type β-lactamases have been more frequently found in Japan and have increased in recent years (23). Derivatives of CTX-M-type β-lactamases preferentially hydrolyze cefotaxime but not ceftazidime. However, Poirel et al. have recently reported that CTX-M-19, a Pro167Ser mutant, could hydrolyze ceftazidime efficiently (19). These authors reported a discrepancy between MIC results and the kinetic parameters of the enzyme against some β-lactams, such as aztreonam and ceftazidime, ascribed to a rapid loss of activity of the enzyme caused by instability of the protein.

To elucidate the mechanism of ceftazidime hydrolysis by CTX-M-type β-lactamases, we obtained the expected MICs of β-lactams, including cefotaxime, ceftazidime, and aztreonam. The kinetic parameters of these enzymes were then measured. The interactions between β-lactamases and β-lactams have been previously studied by computer modeling, an approach that may provide details of plausible interactions between enzyme and substrates (7, 13, 21). In the present study, the CTX-M-type β-lactamase substrate specificity was elucidated by modeling complex structures of β-lactamases with β-lactams (10, 20). Therefore, we constructed a model structure of CTX-M-18 and CTX-M-19.

MATERIALS AND METHODS

Construction of CTX-M-19 by site-directed mutagenesis.

We isolated a plasmid from a CTX-M-18 producer provided by Ma et al. (15) and partially digested it with Sau3AI. The resulting digest containing bla_CTX-M-18 with promoter region was cloned into BamHI-digested pHSG396, yielding a recombination of plasmid pMTY135. To construct CTX-M-19, we used a QuikChange XL site-directed mutagenesis method (Stratagene, Inc.) according to the manufacturer's recommendations for the CTX-M-18 structural gene. Two degenerate complementary primers with the mutant sequences 5′-TGG ATC GCA CTG AAT CTA CGC TGA ATA CCG-3′ and 5′-CGG TAT TCA GCG TAG ATT CAG TGC GAT CCA-3′ containing the replaced base (underlined) were used for pMTY135. The resulting plasmid contained bla_CTX-M-19 and a promoter region yielding recombination plasmid pMTY136. Samples were prepared with ABI Prism BigDye Terminator cycle sequencing ready reaction kits (Applied Biosystems) and sequenced with the automatic sequencer ABI Prism 310 Genetic Analyzer (Applied Biosystems/Perkin-Elmer Biosystems). These plasmids were transformed into Escherichia coli MV1184 [ara Δ(lac-proAB) rpsL thi(φ80 lacZ ΔM15) Δ(srl-recA) 306::Tn10(Tet^r)/F′[traD36 proAB⁺ lacL^q lacZ ΔM15] (Takara Bio, Inc., Tokyo, Japan). The strains were grown in Luria-Bertani broth at 35°C.

Susceptibility testing.

MICs were determined by the National Committee for Clinical Laboratory Standards broth microdilution method with cation-adjusted Mueller-Hinton broth (Difco, Detroit, Mich.) (17). The antibacterial agents and their sources were as follows. Ampicillin and cephalothin were purchased from Sigma Co., Ltd. (Tokyo, Japan). Reference powders of the following antibiotics were kindly provided by their manufacturers: piperacillin (Toyama Chemical Co., Ltd., Tokyo, Japan), cefoxitin and imipenem (Banyu Pharmaceutical Co., Ltd., Tokyo, Japan), cefepime (Bristol Pharmaceutical Co., Tokyo, Japan), ceftizoxime (Fujisawa Pharmaceutical Co., Ltd., Tokyo, Japan), clavulanate and ceftazidime (Glaxo Smith Kline, Tokyo, Japan), cefotaxime (Aventis Pharma, Tokyo, Japan), aztreonam (Eisai Co., Ltd., Tokyo, Japan), and faropenem (Suntory, Ltd., Tokyo, Japan).

Purification of CTX-M-18 or CTX-M-19.

Purified CTX-M-19 was unstable under addition of bovine serum albumin (BSA) in the enzyme solution compared to CTX-M-18 (data not shown). This was consistent with the report of Poirel et al. of a discrepancy between MICs and kinetic parameters that was attributed to the instability of the enzyme. Hence, to obtain stabilized CTX-M-19, we immediately collected vast amounts of CTX-M-19 by using affinity chromatography. To fuse the His tag to the C-terminal amino acid sequence of CTX-M-19, the enzyme gene was amplified by using two primers containing a unique restriction site. The forward primer (CCT TGT TTG AGC ATA TGA CGA GTG CGG TGC AGC) contained the NdeI site (underlined). The reverse primer (ATT CTC GAG CAG CCC TTC GGC GAT GAT TCT) contained the XhoI site (underlined). The PCR conditions were as follows: 94°C for 30 s, 55°C for 30 s, and 72°C 1 min. After 25 cycles, the PCR product was purified with QIAquick PCR purification kit (Qiagen). These genes were digested with NdeI and XhoI for 2 h at 37°C and ligated into pET28b(+) by using a DNA ligation kit (Takara). The resulting plasmid pMTY135 was digested with EcoRI and NdeI (Takara Bio, Inc.) and cloned into the vector pET28a(+) (Novagen).

E. coli BL21(DE3)pLysS [F⁻ ompT hsdS_B(r_B⁻ m_B⁻)gal dcm (DE3)pLysS; Novagen] was used to produce the enzyme as a soluble protein. CTX-M-18- or CTX-M-19-producing strains were grown in 2 liters of Super Broth on a rotating shaker at 25° or 37°C, respectively (16). When the optical density of the culture at 660 nm reached an absorbance of 0.5, IPTG (isopropyl-β-d-thiogalactopyranoside; Sigma) at a concentration of 400 or 100 μM gave an appropriate level of expression of CTX-M-18 or CTX-M-19, respectively. The maximum activity of CTX-M-18 or CTX-M-19 was reached 7 or 4 h after induction, respectively. Crude extracts of these enzymes were harvested by the method of Alba et al. (1). Purification of CTX-M-18 from cell extract was done by cation-exchange chromatography by using MonoS HR5/5 column (Amersham Pharmacia Biotech, Uppsala, Sweden) and MonoQ HR5/5 column (Amersham Pharmacia Biotech). All of the purification process was done with AKTA purifier (Amersham Pharmacia Biotech). Purification of CTX-M-19 was done by affinity chromatography with QIAexpressionist (Qiagen). The purity of these β-lactamase preparations was controlled by using sodium dodecyl sulfate-12% polyacrylamide gel electrophoresis gels stained by Coomassie brilliant blue.

Determination of steady-state kinetic parameters.

Hydrolysis of β-lactam antibiotics was detected by monitoring the variation in the absorbance of β-lactam solution (4 to 9,000 μM) in 50 mM phosphate buffer (pH 7.0). Hydrolysis rates were measured by use of spectrophotometer Shimadzu UV-2550 (Shimadzu, Kyoto, Japan), connected to a personal computer. Steady-state kinetic parameters were determined for the β-lactam compounds reported by Ibuka et al. (11). For the dilution of the enzyme (0.01 to 9,170 nM), BSA was added to the buffer at a final concentration of 20 μg/ml to prevent denaturation of the enzyme. Six different substrate concentrations were used to determine the kinetic parameters for each substrate. In addition, each reported parameter is an average of three independent measurements. All kinetic parameters were determined by measuring the initial hydrolysis rate of the selected antibiotic and by using Hanes-Wolf linearization of the Michaelis-Menten equation. The K_m of poor substrates was determined as the competitive inhibition constant K_i by the competition experiment between the tested β-lactam antibiotic and a good substrate used as a reporter substrate. Nitrocefin at 100 μM was used as the reporter substrate. In that case, the catalytic constants (k_cat) for all substrates were determined by hydrolyzing substrate at a concentration ≥10 times higher than the K_m (K_i) value.

Modeling of complex structures of enzymes and β-lactams.

The three-dimensional crystal structure of Toho-1 was used for modeling of CTX-M-18 and CTX-M-19. The substrates (Fig. 1) were placed at the binding site, dictating the binding mode of acyl moiety of cefotaxime in Toho-1. Molecular dynamics calculations were performed at a time step of 1 fs for 100 ps within the substrate-binding site (residues within 9 Å from the chromophore) with the capped water molecules (20 Å thick) by using Discover 3 (version 98; Molecular Simulations, Inc., San Diego, Calif.). Structural minimization of the whole structure was carried out until the final root-mean square deviation became less than 0.1 kcal/mol/Å.

FIG. 1.

Chemical structure of β-lactam antibiotics.

RESULTS

Antibiotic susceptibility test of transformant with CTX-M-18 or CTX-M-19.

In order to compare the MICs of the CTX-M-18- or CTX-M-19-producing strain against β-lactams, we used transformants of E. coli MV1184 with plasmid pMTY135 (including bla_CTX-M-18) or pMTY136 (including bla_CTX-M-19), which includes the same promoter region reported by Ma et al. (15). Both CTX-M-18- and CTX-M-19-producing strains exhibited high levels of resistance to ampicillin (MIC, 128 μg/ml) and cephalothin (MIC, >64 μg/ml) (Table 1). Both strains were susceptible to cefoxitin (MIC, 2 μg/ml), cefepime (MIC, <0.5 μg/ml), ceftizoxime (MIC, ≤0.25 μg/ml), faropenem (MIC, <8 μg/ml), and imipenem (MIC, <2 μg/ml). However, MICs for CTX-M-18-producing E. coli MV1184(pMTY135) relative to those of CTX-M-19-producing E. coli MV1184(pMTY136) were eightfold or at least fourfold higher against cefotaxime or aztreonam, respectively, whereas CTX-M-18-producing E. coli MV1184(pMTY135) relative to those of E. coli MV1184(pMTY136) were at least sixteenfold lower against ceftazidime. Clavulanate potentiated the activity of ceftazidime (in the presence of CTX-M-19) and cefotaxime (in the presence of either CTX-M-18 and CTX-M-19).

TABLE 1.

Antibiotic MICs for E. coli MV1184 cells harboring CTX-M-19 or CTX-M-18

Antibiotic	MIC (μg/ml)
	E. coli MV1184/ pMTY136a (CTX-M-19)	E. coli MV1184/ pMTY135b (CTX-M-18)	E. coli MV1184/ pHSG396
Ampicillin	128	128	2
Cephalothin	64	128	2
Cefoxitin	2	2	2
Cefepime	≤0.25	0.5	2
Ceftizoxime	≤0.25	≤0.25	≤0.25
Cefotaxime	1	8	2
Cefotaxime-clavulanatec	≤0.25	≤0.25	≤0.25
Ceftazidime	4	≤0.25	≤0.25
Ceftazidime-clavulanate	≤0.25	≤0.25	≤0.25
Aztreonam	≤0.25	1	≤0.25
Faropenem	4	8	8
Imipenem	1	2	1

^apMTY136 was constructed from pMTY135 by using site-directed mutagenesis method.

^bStructure gene of CTX-M-18 was inserted into pHSG396.

^cClavulanate was used at a final concentration of 4 μg/ml.

Kinetic parameters of CTX-M-18 and CTX-M-19.

We measured kinetic parameters of both purified enzymes against selected β-lactams. In the case of CTX-M-19, which was produced with the overproduction system, the enzyme was purified by using histidine-tagged chromatography (Table 2). Since the kinetic parameters of histidine-tagged CTX-M-19 against penicillin G or nitrocefin were identical to those of native CTX-M-19, we used histidine-tagged CTX-M-19 in the following experiments. The purified enzymes gave a single band on SDS-PAGE. As shown in Tables 2 and 3, the purified CTX-M-18 and CTX-M-19 showed high catalytic activities toward penicillin G, cephalothin, and nitrocefin. Inhibition by ceftriaxone, cefoxitin, imipenem, faropenem, or clavulanate was monitored with 100 μM nitrocefin as a reporter substrate. On the other hand, CTX-M-19 and CTX-M-18 showed different behaviors in either k_cat, K_m, or k_cat/K_m against cefotaxime, ceftazidime, or aztreonam. CTX-M-18 and CTX-M-19 had a similar catalytic activity (k_cat) with cefotaxime, whereas CTX-M-18 had 35-fold-higher enzymatic saturation (K_m) than CTX-M-19. This difference in K_m affects the catalytic efficiency (k_cat/K_m) in a similar manner. In contrast, CTX-M-18 and CTX-M-19 had similar K_m values for aztreonam, CTX-M-19 had a 44-fold-lower k_cat than CTX-M-18. This difference in catalytic activity was also reflected in the catalytic efficiency. For ceftazidime, the catalytic activities of CTX-M-18 and CTX-M-19 were not determined, since both enzymes exhibit very high K_i values (2.7 to 5.6 mM) against ceftazidime.

TABLE 2.

Kinetic parameters of CTX-M-19 or CTX-M-18 against good substrates

Compound	Enzyme	Concn range (μM)	Enzyme concn (nM)	Km or Ki (M)	kcat (s−1)	kcat/Km (s−1 M−1)
Penicillin G	CTX-M-18	10-100	0.26	(28.6 ± 0.1) × 10−6	37.4 ± 4.4	1.31 × 106
	CTX-M-19	10-100	0.46	(25.0 ± 0.1) × 10−6	22.3 ± 4.4	0.89 × 106
	CTX-M-19a	10-100	0.46	(24.9 ± 0.1) × 10−6	22.3 ± 0.4	0.89 × 106
Nitrocefin	CTX-M-18	20-100	0.86	(27.0 ± 0.0) × 10−6	633 ± 0	23.4 × 106
	CTX-M-19	10-100	0.76	(18.3 ± 0.0) × 10−6	397 ± 0	21.7 × 106
	CTX-M-19a	10-100	0.76	(18.3 ± 0.0) × 10−6	595 ± 0	32.5 × 106
Cephalothin	CTX-M-18	10-100	0.26	(50.9 ± 0.0) × 10−6	1,190 ± 0	23.4 × 106
	CTX-M-19a	10-80	0.23	(54.8 ± 0.0) × 10−6	882 ± 1	16.1 × 106
Cefotaxime	CTX-M-18	40-120	5.18	(1.68 ± 0.0) × 10−6	74.5 ± 2.5	44.3 × 106
	CTX-M-19a	60-140	13.8	(58.0 ± 0.0) × 10−6	84.8 ± 0.6	1.47 × 106
Ceftazidime	CTX-M-18	1,800-9,000	0.86	(5,610 ± 5) × 10−6b	NDc	ND
	CTX-M-19a	1,800-9,000	1.53	(2,720 ± 4) × 10−6b	ND	ND
Aztreonam	CTX-M-18	50-300	0.01	(278 ± 0) × 10−6	18.5 ± 1.7	6.65 × 104
	CTX-M-19a	360-9,000	1.15	(313 ± 0) × 10−6b	0.420 ± 0.080	1.34 × 103
Cefoxitin	CTX-M-18	10-50	2.20 × 103	(36.0 ± 0.0) × 10−6b	(1.92 ± 0.60) × 10−3	53.3
	CTX-M-19a	10-80	9.17 × 103	(23.9 ± 0.0) × 10−6b	(1.48 ± 0.19) × 10−3	62.2

^aThis enzyme was modified with a His₆ tag.

^bNitrocefin at 100 μM was used as a reporter substrate to obtain the kinetic parameters of these substrates.

^cND, not determined.

TABLE 3.

Kinetic parameters of CTX-M-19 or CTX-M-18 against poor substrates^a

Compound	Enzyme	Concn range (μM)	Enzyme concn (nM)	Kcal (M)b	k+3 (s−1)	k+2/K (s−1 M−1)
Imipenem	CTX-M-18	10-80	0.22	4.38 × 10−5	(4.60 ± 1.00) × 10−3	1.05 × 102
	CTX-M-19	20-100	0.76	6.25 × 10−5	(7.50 ± 0.30) × 10−3	1.20 × 102
Faropenem	CTX-M-18	10-60	0.22	5.58 × 10−6	(5.30 ± 0.42) × 10−3	9.50 × 102
	CTX-M-19	10-60	0.76	1.34 × 10−5	(6.35 ± 0.42) × 10−3	4.73 × 102
Clavulanate	CTX-M-18	4-10	0.22	8.55 × 10−7	(1.00 ± 0.02) × 10−3	1.17 × 103
	CTX-M-19	4-8	0.76	1.07 × 10−6	(1.79 ± 0.04) × 10−3	1.68 × 103

^aNitrocefin at 100 μM was used as a reporter substrate to obtain the kinetic parameters of these substrates.

^bKcal, K calculated from measured values.

Model structures of CTX-M-18 and CTX-M-19.

A model structure of CTX-M-18 was built with the homology modeling by using the crystal structure of Toho-1. Pro167 was then mutated to a Ser residue to provide a model structure of CTX-M-19. The change of the cyclic structure of the proline residue of Pro167 to the serine residue allowed the hydroxyl group of the serine residue protruding from the omega loop into the aminothiazole (AT) moiety binding site. This caused a severe steric interaction with the AT moiety when the AT moiety was bound to the binding cleft.

Model structures of aztreonam-bound CTX-M-18 and CTX-M-19.

When aztreonam was docked into the active site of the enzymes, the thiazole moiety was too close to the hydroxyl group of Ser167 of CTX-M-19 to maintain the original position observed in the cefotaxime-bound Toho-1 (20). Thus, the structure optimization gave a binding mode distinct from that of aztreonam complexed in CTX-M-18. The structure of aztreonam in CTX-M-19 clearly showed a displacement of the AT moiety to escape the steric interaction with the Ser167 hydroxyl group. The dislocation of the AT moiety caused a hydrogen bond network distinct from that of CTX-M-18, in which the carboxylate of the oxime side chain forms hydrogen bonds with Ser237 (Fig. 2A). Consequently, the sulfonate group preferentially formed a hydrogen bond with Ser130. In the CTX-M-19 model, the carboxylate group was located too close to form hydrogen bonds with Ser237 and thus had a different conformation (Fig. 2B). The side chain hydroxyl group of Ser237 also had a different conformation to preferentially form a hydrogen bond with the sulfonate group of aztreonam, and this interaction interfered with the interaction between the sulfonate group and Ser130. Since the electronic interaction of the Ser130 hydroxyl group a the negative group such as sulfonate group in aztreonam would facilitate the enzyme reaction (12), the preferential interaction between the sulfonate group and Ser130 of CTX-M-18 would lead to effective hydrolysis of aztreonam, as observed in the hydrolysis of aztreonam by CTX-M-18.

FIG. 2.

The active site of CTX-M-type β-lactamase with aztreonam (A and B), ceftazidime (C and D), or cefotaxime (E and F). (A, C, and E) CTX-M-18; (B, D, and F) CTX-M-19. Red broken lines indicate hydrogen bonds. Black thin double arrow indicates an interatomic distance longer than hydrogen bond. Blue broken double arrow indicates a steric interaction with the hydroxyl group of S167 and the thiazole ring.

Model structure of ceftazidime-bound CTX-M-18 and CTX-M-19.

The same steric interaction between the AT moiety and the Ser167 residue of CTX-M-19 drove the Ser237 residue forming a hydrogen bond with the carboxylate group on C-4 of ceftazidime (Fig. 2-D). Unlike the sulfonate group of aztreonam, this hydrogen bond appears to assist the carboxylate group to further form a hydrogen bond with Ser130, whereas the carboxylate on C-4 did not form an effective hydrogen bond with Ser130 of CTX-M-18 (Fig. 2C). These complex structures of ceftazidime in CTX-M-18 and CTX-M-19 were consistent with the results that ceftazidime can be efficiently hydrolyzed by CTX-M-19 but not by CTX-M-18.

Model structure of cefotaxime-bound CTX-M-18 and CTX-M-19.

Although the oxime side chain of cefotaxime does not have a carboxylate group to form hydrogen bond with Ser237, Ser237 kept an original conformation without a steric interaction with the methoxime in CTX-M-18 (Fig. 2E). The C-4 carboxylate formed hydrogen bonds with Ser130 and Arg276. On the other hand, the steric interaction between the AT moiety and Ser167 in CTX-M-19 invoked a conformational change of Ser237 to form a hydrogen bond with the C-4 carboxylate (Fig. 2F). Thus, the C-4 carboxylate formed the hydrogen bond with Ser130 in both CTX-M-18 and CTX-M-19. This may be compatible with the finding that both enzymes have a similar affinity for cefotaxime. Further inspection of the complex models, Ser130 in the CTX-M-18 model had a hydrogen bond with Lys73, which is assumed to be a general base for activating the hydroxyl group of Ser70 for acylation (12), whereas Ser130 in the CTX-M-19 model was not located in a hydrogen bond distance to Lys73.

DISCUSSION

It has been reported that CTX-M-type β-lactamases do not hydrolyze ceftazidime. Recently, however, several researchers have reported ceftazidime-hydrolyzing CTX-M-type β-lactamases. In 2001, CTX-M-19 was first reported as a ceftazidime-hydrolyzing CTX-M-type β-lactamase (19). The bacteria producing CTX-M-15, which is a single amino acid mutant (Asp240Gly) of CTX-M-3, were then shown to be resistant to both cefotaxime and ceftazidime (18). In addition, Ho et al. reported that the class A BPS β-lactamase from Burkholderia pseudomallei hydrolyzes ceftazidime. This enzyme has a high degree of amino acid homology to the CTX-M-type β-lactamases (9). These results suggest that CTX-M-type β-lactamases can acquire the ability to hydrolyze ceftazidime through mutation.

The aim of the present study was to confirm ceftazidime hydrolysis by the CTX-M-19 β-lactamase and to elucidate the hydrolytic mechanism involved. First, in order to examine the ability of a CTX-M-type β-lactamase mutant for the hydrolysis of ceftazidime, we determined the MICs of CTX-M-18- and CTX-M-19-producing transformants. For an accurate comparison of MICs, we did not use the clinically isolated strains but rather the transformants, which have an activity identical to that of the promoter located upstream of the structural gene. MICs measured in our experiments were similar to MICs reported by Poirel et al. for clinical isolates (Table 1). Therefore, we used these strains in the following experiments. We compared β-lactam MICs and kinetic parameters of purified enzymes because Poirel et al. reported a discrepancy between MICs and kinetic parameters of CTX-M-19 against ceftazidime and aztreonam (19). These authors suggested that the change of a single amino acid Pro to Ser at 167 causes an instability of the enzyme that leads to the rapid loss of its activity. The instability of the enzyme may be due to the cis peptide bond at 167 on the omega loop, since the cis peptide bond formed by proline is more stable than that formed by serine (22). The purified enzyme was unstable, in particular at a low protein concentration (data not shown). However, we found that the enzyme becomes more stable at a high enzyme concentration with BSA. Under this condition, we obtained more accurate kinetic parameters. We obtained detailed kinetic parameters for CTX-M-18 and CTX-M-19 against cefoxitin and imipenem and for aztreonam against CTX-M-18, in contrast to the more limited findings of Poirel et al. These results suggest that using high concentrations of enzyme with BSA makes the enzyme more stable, thus leading to more accurate kinetic parameters than those obtained by Poirel et al. Comparison of the MICs versus the catalytic efficiency (k_cat/K_m) of these enzymes demonstrated a similar correlation for good substrates such as nitrocefin and cephalothin and for poor substrates such as imipenem and faropenem, as well as for the β-lactams bearing the AT moiety, cefotaxime, ceftazidime, or aztreonam.

In CTX-M-type β-lactamases, the crystal structure of the Toho-1 mutant and the acyl enzyme structures of the Toho-1 and β-lactam antibiotics have been reported (10, 20). The cefotaxime-bound crystal structure of Toho-1 indicates that Pro167 of CTX-M-type β-lactamases has good contact with the AT moiety. The modeling study of CTX-M-18 and CTX-M-19 and the complex model of AT-bearing β-lactams such as aztreonam, ceftazidime, and cefotaxime suggests that the binding of the AT moiety in the CTX-M-19 enzyme requires a displacement of the AT moiety due to a severe steric interaction with the hydroxyl group of Ser167. This dislocation of the AT moiety affects the binding of β-lactams and the interaction of the C-4 carboxylate or sulfonate group with the Ser130 residue. The influence on the interaction would invoke the distinct ability to hydrolyze the AT-bearing antibiotics. Moreover, the hydrogen bond of the C-4 carboxylate of ceftazidime with Ser130 of CTX-M-19 may well correlate with the strong ability of CTX-M-19 to hydrolyze ceftazidime. The ability of CTX-M-18 to hydrolyze aztreonam may also correlate with the hydrogen bond between the sulfonate and Ser130. Since cefotaxime has a smaller substituent at the oxime oxygen, the steric interaction of the AT moiety with Ser167 of CTX-M-19 may not influence the binding affinity with both of the enzymes but cause the reorientation of Ser237 and Ser130 to interact with the C-4 carboxylate. The reorientation of Ser130 results in a lack of the hydrogen bond with Lys73, leading to a reduced ability of CTX-M-19 to hydrolyze cefotaxime.

Poirel et al. suggested that Ser167 in CTX-M-19 enlarges the catalytic site of the enzyme for ceftazidime. However, our results from modeling studies suggest that CTX-M-18 has a larger cavity for binding the AT moiety than CTX-M-19. As a result of this small but significant difference of the binding cavity within these two enzymes, the recognition mode of the enzymes against β-lactams may differ for different β-lactams. Thus, the mutation at the omega loop site may cause a change of the interaction modes of β-lactams with Ser237 on the β3 strand and Ser130 on the SXN loop sites, which are fairly distant from the mutation site. This would correlate with the observation that a Ser237Ala or Arg276Asn mutation of CTX-M-4 decreases its hydrolytic efficiency against cefotaxime (4, 5).