A Kinetic Study of the Replacement by Site Saturation Mutagenesis of Residue 119 in NDM-1 Metallo-β-Lactamase

ABSTRACT

New Delhi metallo-β-lactamase 1 (NDM-1) is a subclass B1 metallo-β-lactamase that exhibits a broad spectrum of activity against β-lactam antibiotics. Here we report the kinetic study of 6 Q119X variants obtained by site-directed mutagenesis of NDM-1. All Q119X variants were able to hydrolyze carbapenems, penicillins and first-, second-, third-, and fourth-generation cephalosporins very efficiently. In particular, Q119E, Q119Y, Q119V, and Q119K mutants showed improvements in k_cat/K_m values for penicillins, compared with NDM-1. The catalytic efficiencies of the Q119K variant for benzylpenicillin and carbenicillin were about 65- and 70-fold higher, respectively, than those of NDM-1. The Q119K and Q119Y enzymes had k_cat/K_m values for ceftazidime about 25- and 89-fold higher, respectively, than that of NDM-1.

TEXT

New Delhi metallo-β-lactamase 1 (NDM-1) plays a key role in antibiotic-resistant bacterial infections. NDM-1-producing bacteria have been identified from both nosocomial and community-acquired infections, including urinary and pulmonary tract infections, peritonitis, septicemia, and osteomyelitis, throughout the world (1–4). The rapid emergence of this enzyme in multiple enteric bacteria is due to the favorable genetic characteristics of the producing plasmids. Multiple antibiotic resistance genes were found to be harbored by plasmids carrying the bla_NDM-1 determinant (5, 6). Therefore, NDM-1-producing strains showed high levels of resistance to a wide range of β-lactams, except monobactams, and also other antibiotic families (7). Structurally, the NDM-1 enzyme displays the typical αβ/βα sandwich architecture (8–11). The active site of NDM-1 is surrounded by several loops that are responsible for substrate binding and specificity and by two zinc ion binding sites, referred to as sites 1 and 2 (9). The two zinc ions promote substrate catalysis; Zn-1 is tetracoordinated by the imidazole groups of three histidine residues, H116, H118, and H196, and one water molecule, whereas Zn-2 is pentacoordinated by H263, D120, and C221 (BBL numbering [12]) and two water molecules. NDM-1 displays 32% sequence identity to the most common metallo-β-lactamases (MBLs), IMP-1 and VIM-2 (9). Several studies demonstrated that, in MBLs, D120 plays an important role in orienting the active site hydroxide ion, which serves as a general base (13). A recent study of mutagenesis at position 120 in NDM-1 validated the hypothesis that this residue acts as a strong Zn-2 ligand and is involved in the interaction with substrate (13).

In the present study, to better understand the function of residue 119 in the NDM-1 enzyme, site saturation mutagenesis was performed on the bla_NDM-1 gene (codons 247 to 249 [CAG]) to generate NDM-1 variants, named Q119X. The codon was randomized using overlap-extension PCR, as described previously (14), with the pFM-NDM-1 plasmid as the template (15). Mutations were introduced into a PCR amplicon using the mutagenic primers Q119X_for (5′-CTCACGCGCATNNNGACAAGATG) and Q119X_rev (5′-CATCTTGTCNNNATGCGCGTGAG), where N represents any of the four base pairs at the codon position to be randomized. The external primers NDM_for (5′-GGGGGGGTACCCATATGGGTGAAATCCGCCCGA) and NDM_rev (5′-GGGGGGAGCTCGAGTCAGCGCAGCTTGTCGGC) (KpnI/NdeI and SacI/XhoI restriction sites are shown in bold) were used to amplify the entire bla_NDM-1 gene. The amplification product, digested with KpnI and SacI, was ligated into the pBC-SK(+) Bluescript vector, and the ligation product was transformed into Escherichia coli XL-1 Bluescript cells. Functional random mutants were selected by growth in the presence of chloramphenicol (30 μg/ml), ampicillin (100 μg/ml), and meropenem (1 μg/ml) and then were sequenced (ABI Prism 3500 eight-capillary automated sequencer; Life Technologies, Monza, Italy) to identify the amino acid substitutions at the randomized position. Among functional clones, 11 NDM-1 mutants were identified, i.e., Q119I, Q119S, Q119L, Q119Y, Q119C, Q119T, Q119F, Q119K, Q119V, Q119E, and Q119M. The phenotypic profiles were characterized by the microdilution method using a bacterial inoculum of 5 × 10⁵ CFU/ml of E. coli XL-1/pBC-SK-Q119X, according to Clinical and Laboratory Standards Institute (CLSI) performance standards (16). The antimicrobial susceptibility to penicillins, cephalosporins, and carbapenems was tested for all recombinant strains, including wild-type E. coli XL-1/pBC-SK-NDM-1 and E. coli XL-1/pBC-SK (Table 1). Like NDM-1, all recombinant mutants exhibited resistance to penicillins, cefazolin, cefotaxime, ceftazidime, and cefepime. Only E. coli XL-1/pBC-SK-Q119E showed a MIC value of 8 μg/ml for cefepime. For cefoxitin, all recombinant strains showed MIC values higher than that of E. coli XL-1/pBC-SK-NDM-1. For all strains tested, the imipenem and meropenem MIC values ranged from 1 μg/ml to 16 μg/ml (Table 1). The mutants were categorized on the basis of MIC values and, in this study, only the following 6 NDM-1 variants were investigated on the basis of their phenotypic profiles: Q119C, Q119F, Q119E, Q119Y, Q119V, and Q119K. For overexpression of the 6 NDM-1 variants, the bla_Q119X mutant genes, without signal peptide, were subcloned into the pET-24a(+) vector and transformed into E. coli BL21(DE3). Purification of the different NDM-1 mutants was performed as described previously (15). From 400 ml of an overnight culture in tryptic soya broth (TSB), we obtained about 5 mg of each NDM-1 mutant. Kinetic experiments were performed by following the hydrolysis of each substrate at 25°C in 25 mM sodium phosphate buffer (pH 7). The data were collected with a Perkin-Elmer Lambda 25 spectrophotometer (Perkin-Elmer Italia, Monza, Italy). Steady-state kinetic analyses were performed under initial-rate conditions using the Hanes plot linearization method (16). For K_m values of ≤10 μM, the K_m values were determined as K_i values, using nitrocefin as the reporter substrate (17). Each kinetic value is the mean of five different measurements; errors were <5%. In Table 2, we show the K_m, k_cat, and k_cat/K_m values for Q119C, Q119F, Q119E, Q119Y, Q119V, and Q119K variants, compared to NDM-1 (data for NDM-1 were taken from reference 15), with some carbapenems, penicillins, and cephalosporins.

TABLE 1.

Pattern of β-lactam resistance mediated by NDM-1 and Q119X mutants in E. coli XL-1, in comparison with E. coli XL-1/pBC-SK

Antibiotic	MIC (μg/ml)
	NDM-1	Q119C/Q119S	Q119E	Q119K	Q119F/Q119T	Q119Y/Q119I/Q119L	Q119V	pBC-SK
Benzylpenicillin	512	64	64	>512	512	512	512	0.250
Piperacillin	256	>512	>512	>512	>512	>512	>512	0.500
Amoxicillin	128	256	256	256	256	256	256	0.500
Imipenem	1	1	1	8	16	4	8	0.125
Meropenem	2	16	4	16	16	16	16	<0.0625
Cefepime	32	32	8	>128	128	128	128	<0.0625
Cefoxitin	8	64	32	128	64	64	64	0.125
Cefazolin	128	256	256	>256	256	256	>256	<0.0625
Cefotaxime	128	128	256	256	256	256	256	<0.0625
Ceftazidime	256	128	128	>128	>128	>128	>128	<0.0625

TABLE 2.

Kinetic parameters for Q119X mutants, compared with NDM-1, with some β-lactams

Substrate and parametera	NDM-1b	Q119C	Q119F	Q119E	Q119Y	Q119V	Q119K
Imipenem
kcat (s−1)	64	69	336	41	79	229	393
Km (μM)	35	50	157	40	30	65	90
kcat/Km (μM−1 s−1)	1.83	1.38	2.14	1.02	2.63	3.52	4.37
Meropenem
kcat (s−1)	75	252	1500	82	561	269	357
Km (μM)	80	100	252	146	105	20	37
kcat/Km (μM−1 s−1)	0.94	2.52	5.95	0.56	5.34	13.45	9.65
Benzylpenicillin
kcat (s−1)	105	100	166	100	34	231	414
Km (μM)	250	71	89	5c	8c	45	15
kcat/Km (μM−1 s−1)	0.42	1.41	1.86	20.0	4.25	5.10	27.6
Carbenicillin
kcat (s−1)	108	52	158	234	216	267	929
Km (μM)	285	59	131	18	31	32	35
kcat/Km (μM−1 s−1)	0.38	0.88	1.21	13.00	6.97	8.34	26.54
Cefazolin
kcat (s−1)	42	48	169	63	561	269	125
Km (μM)	20	11	27	4c	105	20	32
kcat/Km (μM−1 s−1)	2.10	4.36	6.26	15.75	5.34	13.45	3.90
Cefoxitin
kcat (s−1)	23	99	152	64	85	193	58
Km (μM)	26	23	21	23	17	18	49
kcat/Km (μM−1 s−1)	0.88	4.30	7.24	2.78	5.00	10.72	1.18
Cefotaxime
kcat (s−1)	20	43	59	4	7	21	89
Km (μM)	14	7c	6c	2c	2c	10c	3c
kcat/Km (μM−1 s−1)	1.43	6.14	9.83	2.00	3.50	2.00	29.67
Ceftazidime
kcat (s−1)	18.5	74	257	14	68	7	106
Km (μM)	50	163	403	140	2c	3c	11
kcat/Km (μM−1 s−1)	0.37	0.45	0.64	0.10	34.00	2.33	9.64
Cefepime
kcat (s−1)	13	22	67	18	30	71	60
Km (μM)	35	30	16	40	4c	10c	18
kcat/Km (μM−1 s−1)	0.37	0.73	4.19	0.45	7.50	7.10	3.33

^aEach kinetic value is the mean of five different measurements; the error was <5%.

^bKinetic data for NDM-1 were from reference 15.

^cK_m values were determined as K_i values, using nitrocefin as the reporter substrate (19).

Imipenem and meropenem were good substrates for all Q119X variants analyzed, with k_cat values higher than that calculated for NDM-1 (except for Q119E with imipenem). The Q119F variant showed a k_cat value for meropenem of 1,500 s⁻¹, about 20-fold higher than the NDM-1 value. Regarding penicillins, the Q119X variants showed lower K_m values toward both benzylpenicillin and carbenicillin, with values ranging from 5- to 50-fold less than the NDM-1 values. The k_cat values for benzylpenicillin were quite similar to the NDM-1 value except for Q119Y, with a k_cat value about 3-fold lower than the NDM-1 value, and Q119K, with a k_cat value 4-fold higher than the NDM-1 value. Concerning carbenicillin, the Q119X variants (not including Q119C) showed higher k_cat values.

All NDM-1 variants hydrolyzed first-, second-, third-, and fourth-generation cephalosporins very efficiently (Table 2). Indeed, the k_cat values calculated for cefazolin, cefoxitin, and cefepime were higher than the NDM-1 values. Differences in the k_cat values determined with cefotaxime were observed among the Q119X variants. The Q119V variant exhibited the same k_cat value as NDM-1, the Q119E and Q119Y variants showed lower k_cat values, and the Q119C, Q119F, and Q119K variants showed higher k_cat values. Regarding ceftazidime, we observed different behavior among the Q119X variants. For example, the Q119Y showed a k_cat of 68 s⁻¹, a K_m of 2 μM, and a k_cat/K_m value about 10-fold higher than the NDM-1 value. Cefepime, a fourth-generation cephalosporin, was more stable than cefazolin, cefoxitin, cefotaxime, and ceftazidime with all NDM-1 variants. For all Q119X variants (with rare exceptions), increases in the catalytic efficiency (k_cat/K_m), compared with NDM-1, were detected with all β-lactams examined (Table 3). In particular, the Q119E, Q119Y, Q119V, and Q119K variants showed significant improvements in k_cat/K_m for penicillins, due to lower K_m values. The catalytic efficiencies of the Q119K variant for benzylpenicillin and carbenicillin were about 65- and 70-fold higher, respectively, than those of NDM-1. The Q119K and Q119Y enzymes had k_cat/K_m values for ceftazidime about 25- and 89-fold higher, respectively, than that of NDM-1. The Q119E variant showed high catalytic efficiencies with penicillins but lower efficiencies with carbapenems and ceftazidime. The replacement of glutamine (Q) with cysteine (C), phenylalanine (F), glutamic acid (E), tyrosine (Y), valine (V), or lysine (K) considerably increased the activity of NDM-1 for some β-lactams, such as benzylpenicillin, carbenicillin, meropenem (except Q119E), ceftazidime (particularly true for Q119Y and Q119K), and cefepime (particularly true for Q119Y and Q119V). However, no clear correlation between the MIC values and the kinetic data could be observed, mainly because the MIC values for the majority of substrates did not differ substantially between wild-type NDM-1 and the mutants. For cefazolin, cefoxitin, and meropenem, the Q119V variant was always the most efficient enzyme (up to 10-fold greater than NDM-1), followed by Q119F and Q119E (about 2- to 5-fold greater than NDM-1). For benzylpenicillin, carbenicillin, cefotaxime, ceftazidime, and meropenem, Q119K was the most efficient enzyme, with catalytic efficiencies about 66-, 69-, 20-, 25-, and 10-fold higher, respectively, than those of NDM-1.

TABLE 3.

Ratios of k_cat/K_m values for NDM-1 and mutants

Substrate	kcat/Km for Q119X variant/kcat/Km for NDM-1
	Q119C	Q119F	Q119E	Q119Y	Q119V	Q119K
Benzylpenicillin	3.4	4.4	47.6	10.1	12.1	65.7
Carbenicillin	2.3	3.2	34.2	18.3	21.9	69.8
Imipenem	0.7	1.2	0.6	1.4	1.9	2.4
Meropenem	2.7	6.3	0.6	5.7	14.3	10.3
Cefazolin	2.1	2.9	7.5	2.5	6.4	1.9
Cefoxitin	4.9	8.2	3.2	5.7	12.2	1.3
Cefotaxime	4.3	6.9	1.4	2.4	1.4	20.7
Ceftazidime	1.9	1.7	0.3	89.5	6.1	25.4
Cefepime	2.0	11.3	1.2	20.3	19.2	9.0

Residue 119 is positioned in the L5 loop, which includes important catalytic residues such as H116, H118, and D120. Q119 is adjacent to D120 and we presume that it could play an important role in the activities of NDM-1. Structural studies of NDM-1 (18) suggested that Q119 maintains the correct orientation of H118 and D120 in the Zn-1 coordination and therefore plays a major role in the catalytic properties of NDM-1. The study by Chiou et al. demonstrated that Q119D and Q119A mutants showed reduced carbapenem, ampicillin, and cefepime MIC values and Q119D showed reductions in k_cat values for the same β-lactams (18); this seems to be a discrepancy with respect to the results obtained with our mutants, for which substitutions at position 119 led to enzymes with major activity, compared to the wild-type NDM-1 enzyme. The results obtained in the present study suggest that the involvement of residue 119 in NDM-1 activity is related to the amino acid structure. Our results highlight this position as a good candidate for the evolution of the NDM MBLs.