Ceftazidime and Aztreonam Resistance in Providencia stuartii: Characterization of a Natural TEM-Derived ExtendedSpectrum β-Lactamase, TEM-60

Abstract

A plasmid-encoded β-lactamase produced from a clinical strain of Providencia stuartii has been purified and characterized. The gene coding for the β-lactamase was cloned and sequenced. It appears to be a new natural TEM-derived enzyme, named TEM-60. Point mutations (Q39K, L51P, E104K, and R164S) are present with respect to the TEM-1 enzyme; the mutation L51P has never been previously reported, with the exception of the chromosomally encoded extended-spectrum β-lactamase PER-1. Kinetic parameters relative to penicillins, cephalosporins, and monobactams other than mechanism-based inactivators were related to the in vitro susceptibility phenotype.

The use of stable β-lactams to treat bacterial infectious diseases in intensive care units or in immunocompromised patients has resulted in the large spread of resistant strains. These isolates are often characterized by the presence of mutant β-lactamases belonging to the TEM and/or SHV families, which are diffused worldwide in clinically relevant strains. The presence of these extended-spectrum β-lactamases (ESBL) usually confers a phenotype of resistance to ureido-penicillins and to newer cephalosporins. Occasionally these enzymes appear to be resistant to mechanism-based inactivators used in therapy, such as clavulanic acid and penicillanic acid sulphones (14).

Besides the presence of natural β-lactamase mutants, to better understand the molecular basis of the extension substrate specificity, a large number of laboratory mutants have been selected in vitro (18). A detailed knowledge of extended-spectrum activity allows us to anticipate the future; presumably, plasmid or chromosomally encoded β-lactamases, not necessarily belonging to the TEM or SHV families, could acquire favorable mutations to inactivate all antibiotics actually used in clinical therapy (11). The aim of the present study was to characterize a new TEM-derived enzyme showing an ESBL behavior found in a clinical isolate of Providencia stuartii. Usually, strains belonging to the Providencia genus are most frequently isolated from urinary tract infections in hospitalized patients (7). Moreover, P. stuartii may be the causative agent of fatal bacteriemia not related to the urinary tract. The increasing role of P. stuartii as a nosocomial pathogen in the dissemination of plasmid-mediated resistance is confirmed, although this species usually preserves its susceptibility to extended-spectrum cephalosporins such as cefotaxime and ceftazidime.

MATERIALS AND METHODS

Bacterial strain.

P. stuartii VR-1 was a clinical strain isolated in Milan, Italy, from the urinary tract of a hospitalized elderly patient. It was identified with API-20 E (bioMerieux, Marcy l’Etoile, France). On the basis of agar disk diffusion assay, the strain was resistant to extended-spectrum β-lactams other than aztreonam. P. stuartii VR-1 was also resistant to gentamicin, tobramycin, netilmycin, ciprofloxacin, and polymixin B.

Antibiotics and determination of MICs.

The following antibiotics were obtained from the respective manufacturers: piperacillin (Lederle Wyeth, Catania, Italy); cefotaxime and desacetyl-cefotaxime (Hoechst AG, Frankfurt, Germany); imipenem (Merck Sharp & Dohme, Pomezia, Italy); aztreonam (Squibb Institute for Medical Research, Princeton, N.J.); ceftazidime (Glaxo Wellcome, Verona, Italy); clavulanic acid (SmithKline Beecham Research Laboratories, Brentford, United Kingdom); 6-β-iodopenicillanic acid (β-IP) and sulbactam (Pfizer Central Research, Sandwich, United Kingdom); carumonam (Roche, Basel, Switzerland); and benzylpenicillin, ampicillin, cephalothin, and cephaloridine (Sigma Chemical Co., St. Louis, Mo.). Nitrocefin was purchased from Unipath (Milan, Italy). MICs were determined by the conventional macrodilution broth procedure, according to the National Committee for Clinical Laboratory Standards. Tazobactam and clavulanic acid were used at fixed concentrations of 8 and 4 μg/ml, respectively. Two different inocula (5 × 10⁴ and 5 × 10⁷ CFU) were used for testing.

Plasmid content and transformation.

Lysates of P. stuartii VR-1 cultures were prepared as reported by Kado and Liu (13). The plasmid content was analyzed running the DNA preparation on an 0.8% agarose gel electrophoresis (120 V, 2 h). An approximative plasmid size was established by comparison with a standard mixture of known DNA fragments (MW-XV Boehringer, Mannheim, Germany). Escherichia coli HB101 was used as the recipient for the plasmid. Transformation experiments were performed as reported by Sambrook et al. (20). Transformants were selected on Luria-Bertani (LB) agar plates containing ampicillin (100 μg/ml).

Recombinant DNA technology.

PCR was applied to amplify the gene coding the β-lactamase, by using the purified plasmid as DNA template. Two to 5 ng of DNA were subjected to 30 cycles of amplification by using 2 U of Taq polymerase (Perkin-Elmer, Roche Molecular Systems Inc., Branchburg, N.J.). Oligonucleotides designed on the basis of a published TEM-1 β-lactamase nucleotide sequence were used for the fragment amplification as reported also in Mabilat et al. (16). Direct sequencing of PCR amplimers was performed according to the dideoxy-chain termination method by using a Sequenase 2.0 DNA sequencing kit (USB, Amersham, Little Chalfont, United Kingdom), according to the manufacturer’s specifications (2).

β-Lactamase preparation and purification.

Two liters of Luria-Bertani medium was inoculated with 100 ml of an overnight preculture of P. stuartii VR-1 at 37°C in a rotary incubator (120 rpm). The bacterial suspension was recovered by centrifugation and the pellet resuspended in 50 mM sodium phosphate buffer (pH 7). Crude enzyme was obtained by sonication. The membrane debris was removed by high-speed centrifugation. Forty milliliters of the clarified supernatant containing the β-lactamase activity was extensively dialyzed against 50 mM sodium phosphate buffer (pH 7) and loaded onto a Sepharose Q FF column (2.5 by 30 cm; Pharmacia Biotech, Milan, Italy) equilibrated with the same buffer. The unbound material was eliminated from the column by washing with the same buffer (A₂₈₀ < 0.05). The enzyme was eluted by a linear gradient (0 to 1 M NaCl) in the same buffer. Five fractions containing the β-lactamase activity were pooled and dialyzed against 25 mM bis-tris buffer (pH 7.4). The protein was purified to homogeneity by the fast-chromatofocusing technique with a Mono P HR5/20 column equilibrated in the above buffer. The protein was eluted with 2.5% Ampholine solution (1:10 diluted) (pH 4). The pH of the single fractions was immediately adjusted to 7 by the addition of 1 M potassium phosphate buffer (pH 7) (1/10 of the fraction volume).

β-Lactamase activity.

The β-lactamase activity was measured in a spectrophotometer by following the hydrolysis of a 100 μM cephaloridine solution (Δɛ_M²⁶⁰ = −10.000 M⁻¹ cm⁻¹). One enzyme unit was defined as the amount of enzyme hydrolyzing 1 μmole of cephaloridine/min at maximum rate, at 30°C in 50 mM sodium phosphate buffer (pH 7).

M_r and isoelectric focusing analysis.

The relative β-lactamase molecular mass was estimated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis with 15% acrylamide (15). Proteins were boiled 5 min at 100°C in a solution containing 1% SDS and 4% mercaptoethanol before being loaded on the gel. The low-range marker proteins (Bio-Rad, Richmond, N.J.) were subjected to the same treatment and used as the standard reference.

Crude β-lactamase extract and purified enzyme were analyzed on precast polyacrylamide gel (5% wt/vol) containing Ampholines (pH range, 3.5 to 9.5) on a Multiphor II flatbed apparatus (Pharmacia Biotech) at 6°C. Enzyme activity was revealed by overlaying the gel with a paper strip soaked with 250 μM nitrocefin.

Kinetic parameter determinations.

The complete time course of the substrate hydrolysis was recorded with a lambda2 spectrophotometer connected to an IBM-compatible microcomputer via an RS232 interface. The K_m and V_max values were derived as reported by De Meester et al. (9). When the K_m was quite low it was calculated as K_i by using 100 μM nitrocefin as a reporter substrate. The interaction with β-IP was studied with the substrate reporter method as described by De Meester et al. (8). The reported values are the means of three different experiments, with errors below 10%.

Determination of N-terminal amino acid sequence.

The mature TEM-60 was used to characterize its N-terminal amino acid sequence. N-terminal amino acid derivatives were analyzed with the help of a gas-phase Sequenator (Applied Biosystems, Foster City, Calif.).

Nucleotide sequence accession number.

The nucleotide sequence data appear in the GenBank database under accession no. AF 047171.

RESULTS

β-Lactamase purification.

A typical purification procedure is reported in Table 1. A 117-fold purification of the enzyme was obtained with a yield of 29%. The enzyme used for the kinetic measurements was 95% pure.

TABLE 1.

Purification of the TEM-60 β-lactamase produced by P. stuartii VR-1

Step	Total protein (mg)	Specific activity (U/mg)	Total activity (U)	Purification (fold)	Yield (%)
Crude lysatea	260	2.6	676	1	100
Sepharose Q-FF	8.5	52.0	442	20	65
Chromatofocusing	0.55	305.0	195	117	29

^aThe crude lysate was obtained from 2 liters of an overnight culture. 

Isolation of the P. stuartii VR-1 plasmid.

A plasmid (pVR-1) estimated to be 20 kb was isolated from the P. stuartii VR-1 strain. It was transferred into the E. coli HB101 strain by transformation (20). The size of the plasmid purified from the transformed E. coli HB101 was identical to that found in P. stuartii VR-1 and conferred the same phenotypic pattern.

MICs.

The reported MICs showed that the presence of the VR-1 β-lactamase in P. stuartii is able to confer resistance to a broad spectrum of β-lactam antibiotics, including monobactams. Susceptibility to penicillins is restored by the combination piperacillin-tazobactam at two inocula but not by amoxicillin-clavulanate or by ampicillin-sulbactam. However, the amoxicillin-clavulanate combination is able to lower the MIC (Table 2). The susceptibility data suggested the presence of an extended-spectrum β-lactamase responsible for the resistance phenotype in P. stuartii VR-1.

TABLE 2.

MICs (μg/ml) of P. stuartii VR-1 wild type and E. coli HB101 containing the pVR-1 plasmid

Antibiotica	E. coli HB101 5 × 107	P. stuartii 5 × 104	P. stuartii 5 × 107	E. coli pVR-1 5 × 104	E. coli pVR-1 5 × 107
AMP	4	>1024	>1024	>1024	>1024
AMP plus SULB	<0.5	>256	>256	>256	>256
PIP	<0.5	256	>256	256	>256
PIP plus TAZ	<0.5	4	4	2	1
AMX	4	>1024	>1024	>1024	>1024
AMX plus CLAV	<2	64	128	4	8
CAZ	<0.5	>256	>256	>256	>256
CTX	<0.5	8	16	4	8
IMP	<0.5	1	8	<0.5	4
CRM	<0.5	1	>64	8	>64
ATM	<0.5	64	>64	>64	>64

^aAMP, ampicillin; SULB, sulbactam; PIP, piperacillin; TAZ, tazobactam; AMX, amoxicillin; CLAV, clavulanic acid; CAZ, ceftazidime; CTX, cefotaxime; IMP, Imipenem; CRM, Carumonam; ATM, Aztreonam. 

Recombinant DNA technology and N-terminal amino acid sequence.

PCR experiments performed on the purified plasmid gave a single 1,100-bp fragment including an open reading frame of 865 bp potentially encoding a 286-amino-acid polypeptide. The nucleotide sequence determined on both DNA strands showed the main characteristics of a class A β-lactamase, such as the four conserved elements involved in the active site delimitation, i.e., S*XXK (where S* is the active-site Ser residue), SDN, EXXLN, and KSG. Sequence comparison with other TEM enzymes revealed the highest sequence homology with TEM-8 and TEM-7 enzymes, even if other amino acid substitutions related to TEM-2, TEM-3, and TEM-18 were present (Table 3).

TABLE 3.

Comparative analysis of VR-1 residues with some key amino acid residues in the TEM β-lactamases family

	pI	39	51	104	164	238
TEM-1	5.4	Q	L	E	R	G
TEM-2	5.6	K
TEM-3	6.3	K		K		S
TEM-7	5.4	K			S
TEM-8	5.9	K		K	S	S
TEM-18	6.3	K		K
TEM-60a	6.4	K	P	K	S

^aTEM-60, β-lactamase from P. stuartii VR-1. 

The N-terminal amino acid sequence (HPETLVKVKDAEDKL…) determined by microsequencing is in complete agreement with that deduced from the gene sequence. By comparison with TEM-1 enzyme, four mutated residues were found (Table 3).

M_r and isoelectric point.

The SDS-polyacrylamide gel electrophoresis analysis showed the purified enzyme to have a relative mass of 27,500 Da (data not shown). An isoelectric point value of 6.4 was also determined.

Kinetic parameters.

The values of K_m and k_cat determined for the purified enzyme are reported in Table 4 in comparison with the values calculated for TEM-1 (19). Penicillins were the best substrates; moreover, cephalosporins had relatively high k_cat values compared to penicillins. Surprisingly, aztreonam is one of the best substrates, showing a catalytic efficiency higher than that of most of the cephalosporins tested in the present study. A comparative analysis of catalytic efficiency between P. stuartii VR-1 β-lactamase, TEM-1, TEM-3, and TEM-18 is reported in Table 5.

TABLE 4.

Kinetic parameters determined for the TEM-60 β-lactamase purified from P. stuartii VR-1 compared to TEM-1 enzyme^a

β-lactam	Km (μM)		kcat (s−1)		kcat/Km (μM−1 s−1)
	TEM-60	TEM-1	TEM-60	TEM-1	TEM-60	TEM-1
Benzylpenicillin	9	25	26	1,600	2.8	64
Carbenicillin	20	n.d.	10	n.d.	0.5	n.d.
Ampicillin	9	31.8	18	1050	0.9	33
Piperacillin	10	n.d.	8.8	n.d.	0.88	n.d.
Amoxicillin	11	n.d.	14	n.d.	12	n.d.
Cephaloridine	28	681	15	1,500	0.53	2.2
Cephalothin	54	246	3	160	0.06	0.65
Cefuroxime	31	1,000	3.7	6	0.12	0.006
Cefotaxime	31	6,000	1.5	9	0.05	0.0015
Desacetyl cefotaxime	29	n.d.	1.1	n.d.	0.04	n.d.
Ceftazidime	59	4,298	12	0.3	0.2	0.0007
Nitrocefin	20	54.7	5	930	0.85	17
Aztreonam	55	1,430	9	1	0.16	0.007
Carumonam	17	n.d.	0.2	n.d.	0.01	n.d.

^aData from reference 18; n.d., not determined. 

TABLE 5.

Comparative analysis of catalytic efficiency between TEM-60 from P. stuartii VR-1 and some related TEM enzymes

Antibiotics	kcat/Km (μM−1 s−1)
	TEM-60	TEM-3a	TEM-18a	TEM-1a
Benzylpenicillin	2.8	9.5	69	84
Carbenicillin	0.5	n.d.b	n.d.	n.d.
Ampicillin	0.9	3.8	26	33
Piperacillin	0.9	n.d.	n.d.	n.d.
Amoxicillin	12	n.d.	n.d.	n.d.
Cephaloridine	0.53	1.1	2.1	2.2
Cephalotin	0.06	1.9	1.2	0.65
Cefuroxime	0.12	0.15	0.02	0.006
Cefotaxime	0.05	1.0	0.08	0.0015
Desacetyl cefotaxime	0.04	n.d.	n.d.	n.d.
Ceftazidime	0.2	0.04	0.008	0.0007
Nitrocefin	0.85	7.3	38	17
Aztreonam	0.16	0.2	0.014	0.007
Carumonam	0.01	n.d.	n.d.	n.d.

^aData from reference 18. 

^bn.d., not determined. 

Interaction with β-iodopenicillanate, clavulanic acid, tazobactam, and sulbactam.

The P. stuartii VR-1 β-lactamase was rapidly inactivated after incubation with β-IP. The complete inactivation of the enzyme was reached at a 1:1 ratio with the inhibitor. No reactivation was observed by dilution of the complex in the presence of a good substrate such as 100 μM nitrocefin at a pH of 7. The inactivation rate constants were computed according to the following scheme: K k₊₂ k₊₃E + I ⇄ EC → EC* → E + P ⇕ k₊₄EC_i

where EC* is the acylenzyme and EC_i is the irreversibly inactivated complex.

The variation of the pseudo-first order inactivation rate constant (k_i) versus the inhibitor concentration allowed the calculation of the acylation constant (k₊₂ = 0.021 s⁻¹), the dissociation constant (K = 0.08 μM), and the acylation efficiency (k₊₂/K = 260,000 M⁻¹ s⁻¹) of the process.

After incubating clavulanic acid (0.48 to 1.6 μM) as a competitive inhibitor with 100 μM nitrocefin and measuring the initial reaction rate, a K_i = 0.7 μM was computed. The values of the first-order rate constant were also estimated by longer incubation and by monitoring the relative hydrolysis of the same substrate as reported for β-IP interaction. The deacylation process was measured by the direct hydrolysis of clavulanic acid at 265 nm (Δɛ_M²⁶⁵ = + 2,000 M⁻¹ cm⁻¹) and a k₊₃ of 8.7 10⁻³ s⁻¹ was calculated. The acylation efficiency reported as k₊₂/K was estimated to be 12,000 M⁻¹ s⁻¹. These values were estimated as reported by Galleni et al. (10).

Tazobactam (0.1 to 5 μM) was used as a competitive inhibitor to calculate a K_i of 0.018 μM. The deacylation process of the enzyme-inhibitor complex was estimated by incubating the enzyme and the inhibitor (1:5,000 ratio) at 30°C. After 10 min, a small aliquot of the complex (5 μl) was removed and added to 500 μl of 100 μM nitrocefin. The hydrolysis of nitrocefin was continuously monitored at 30°C until the rate remained constant. The k₊₃ computed for the deacylation process was 4.2 10⁻⁴ s⁻¹ and a k₊₂/K of 21,000 M⁻¹ s⁻¹.

No direct hydrolysis was detected with sulbactam by using a large quantity of enzyme (5 to 10 μg). A K_i value of 0.075 μM was estimated in competitive inhibition assays when a range of 0.5 to 5 μM inhibitor was used with 100 μM nitrocefin as the reporter substrate. The results of interactions with inhibitors are summarized in Table 6.

TABLE 6.

Kinetic parameters for the inactivation of the TEM-60 β-lactamase

β-lactam	K (μM)	k+2 (s−1)	k+3 (s−1)	k+2/K (M−1 s−1)
β-IP	8 · 10−2	2.1 · 10−2	n.d.b	2.6 · 105
Sulbactam	7 · 10−2	n.d.	n.d.	n.d.
Tazobactam	2 · 10−2	n.d.	4.2 · 10−4a	2.1 · 104
Clavulanate	0.7	n.d.	8.7 · 10−3	1.2 · 104

^aMeasured from the deacylation of the complex using 100 μM nitrocefin as substrate. 

^bn.d., not determined. 

DISCUSSION

The contribution of TEM-derived or chromosome-encoded ESBLs in the resistance mechanism to β-lactam-stable antibiotics has been reported worldwide (4, 5, 12). Nevertheless P. stuartii spp. usually have not shown a loss of sensitivity to extended-spectrum cephalosporins. In this study a P. stuartii β-lactamase-producing strain has been isolated during a survey from institutionalized elderly patients with urinary tract infections, whose phenotypic resistance pattern toward expanded-spectrum cephalosporins was evident (7). The β-lactamase encoded by a 20-kb plasmid had a molecular mass and reactivity parameters relative to the interaction with β-IP consistent with a class A β-lactamase as proposed by Ambler (1). This finding was confirmed by gene sequencing, which revealed the typical conserved motifs of class A β-lactamases (1). Moreover, our data confirmed that the P. stuartii VR-1 enzyme is a natural TEM-1-derived β-lactamase with the following mutations: G39K, L51P, E104K, and R164S. The mutation L51P is a unique characteristic of the P. stuartii VR-1 enzyme compared to TEM-2, TEM-3, TEM-7, and TEM-18 enzymes that have the other mutations. This mutation is also present on the chromosomally encoded β-lactamase PER-1 (17); but at the present its role is not yet explored.

The Leu-51 residue lies on the beta sheet S1, just a few residues before beta sheet S2. The L51 side chain is in the vicinity of R191, L194, and T195 belonging to helix H8 and I260 (beta sheet S5). Thus a substitution of P for L51 could contribute to the β-lactamase hydrolytic properties. Site-directed mutagenesis could help clarify how this residue influences β-lactamase activity.

Comparison of the catalytic efficiency with TEM-1 β-lactamase showed a marked diminution of the k_cat/K_m ratio relatively to some oldest β-lactams, while the same ratio was relatively higher for the P. stuartii enzyme, especially for cefuroxime, cefotaxime, ceftazidime and aztreonam. The higher ratio of k_cat/K_m is due to the higher affinity for cefotaxime and cefuroxime, and the same behavior was found for aztreonam and ceftazidime. The high hydrolytic activity against ceftazidime and a greater degree of resistance to ceftazidime than cefotaxime confirm that, as reported for TEM-8 (21), the mutation R164S leads to an increase of resistance to ceftazidime. Hydrolysis of desacetyl-cefotaxime, a cephalosporin lacking a good leaving group in the C-3 position, was no different from that for cefotaxime with TEM-60. A completely different behavior was found for carumonam; for this monobactam a 35-fold reduction of the k_cat was measured with respect to the aztreonam. This finding could prompt the investigation of structural analogs of monobactams that could be used with ESBL producing strains.

The comparative analysis of mechanism-based serine-active β-lactamase inhibitors confirmed the partial sensitivity of this β-lactamase to these compounds. β-IP leads to the complete inactivation of the enzyme; no enzyme activity recovery occurred after dilution of the complex in a good substrate such as nitrocefin. The high acylation efficiency value (k₊₂/K = 260,000 M⁻¹ s⁻¹) was similar to that reported for other class A β-lactamases (8). Clavulanic acid, although showing a high k₊₂/K value and a low K_i value (0.8 μM), was hydrolyzed by the enzyme. This could explain the MICs reported for the combination ampicillin-clavulanic acid; clavulanic acid reduces at least 10-fold the MIC but is not able to completely restore susceptibility of the strain. Tazobactam was able to restore piperacillin sensitivity in the P. stuartii VR-1 wild type and also in E. coli HB101 harboring the pVR-1 plasmid. The P. stuartii TEM-60 β-lactamase, as suggested by Bush and Jacoby (6), confirms that additional combinations of mutations could play a key role in the future dissemination of antibiotic resistance.