Biochemical Characterization of the Subclass B3 Metallo-β-Lactamase PJM-1 from Pseudoxanthomonas japonensis

ABSTRACT

Biochemical properties of the novel subclass B3 metallo-β-lactamase (MBL) PJM-1 expressed in Pseudoxanthomonas japonensis, which is often isolated from the environment, were determined. The 906-bp bla_PJM-1 gene in P. japonensis is a species-specific MBL gene, and PJM, with 301 predicted amino acids, has 81.8% amino acid identity with AIM-1. In this study, PJM-1 was recombinantly expressed and purified. PJM-1 showed a low catalytic activity against ceftazidime and cefepime, and it was strongly inhibited by EDTA.

INTRODUCTION

Metallo-β-lactamases (MBLs) have high diversity and are divided into three subclasses: B1, B2, and B3 (1, 2). Subclass B1 MBLs (e.g., IMP, VIM, and NDM) are found on mobile elements in several species (2). In contrast, subclass B2 and B3 MBL genes are frequently present in nonglucose fermenting Gram-negative organisms with a few exceptions, including several rare subclass B3 enzymes such as SMB-1 and AIM-1 (3–8).

Pseudoxanthomonas, a Gram-negative bacterium, was first reported by Finkmann et al. in 2000 (9). Pseudoxanthomonas japonensis was originally isolated from polluted urban soil in Tokyo, Japan and was reported as a novel species by Thierry et al. in 2004 (10). Although Pseudoxanthomonas species are widely distributed in the environment, they rarely infect humans. Although they harbor MBL genes, the biochemical characteristics of the MBL enzymes have not been evaluated.

Recently, we isolated P. japonensis TUM20250 strain from the water of a hospital sink for screening multidrug-resistant bacteria. Species identification of TUM20250 was performed using MALDI-TOF MS spectrometry on a MALDI Biotyper (Bruker Daltonics, Hamburg, Germany). TUM20250 was identified as P. japonensis with a high spectral score (>2.0). MICs for TUM20250 were determined using the broth microdilution method, according to the Clinical and Laboratory Standards Institute (CLSI) (Table 1) (11). TUM20250 had a high imipenem MIC (4 mg/L), and MBL production in TUM20250 was confirmed. In contrast, the MICs of ceftazidime and cefepime were lower than those of cephalothin and cefoxitin, respectively (Table 1).

TABLE 1.

MICs (mg/L) for TUM20250 strain, BL21, and their transformants

Antibiotic	Pseudoxanthomonas japonensis TUM20250	Escherichia coli BL21
		pET9a (–)	pET9a (+)a
			Vector only	bla PJM	blaPJM +EDTAb
Ampicillin	4	2	2	>128	128
Piperacillin	2	0.5	0.5	>128	32
Carbenicillin	2	2	2	>128	2
Cephalothin	16	1	1	>128	32
Cefoxitin	32	1	1	>128	8
Cefotaxime	4	<0.125	<0.125	32	0.125
Ceftazidime	1	0.25	0.25	2	0.125
Ceftazidime/clavulanate	1	0.25	0.25	2	0.125
Cefepime	0.5	<0.125	<0.125	<0.125	<0.125
Imipenem	4	0.25	0.25	128	2
Meropenem	1	<0.125	<0.125	64	4
Ertapenem	2	<0.125	<0.125	128	1
Aztreonam	16	<0.125	<0.125	<0.125	<0.125

^aThe MICs remained changed under induced and noninduced conditions of 100 μM IPTG.

^bContains EDTA at a final concentration of 100 μM.

We further analyzed TUM20250 using the draft whole-genome sequence (WGS) obtained with MiSeq (Illumina, San Diego, CA, USA) (GenBank accession no. NZ_BOUK00000000) (12). The WGS of TUM20250 shared high concordance with that of P. japonensis type strain DSM 17109 based on the average nucleotide identity (95.8%) (13). Furthermore, we identified an MBL gene, bla_PJM-1 (PJM, Pseudoxanthomonas japonensis metallo-β-lactamase-1), consisting of a 906-bp nucleotide sequence with 69.8% GC content and a predicted 301 amino acid sequence. Phylogenetic analysis based on amino acid sequences showed that PJM-1 belongs to the MBL subclass B3 (Fig. 1) based on the BBL nomenclature, displaying a conserved zinc-binding region H-X-H-X-D-H at positions 116 to 121, 196H, and 263H (Fig. 2) (14). PJM-1 shared 81.8% identity (contained signal peptide) with AIM-1, originally identified from Pseudomonas aeruginosa. In contrast, low identities were observed in other subclass B3 MBLs, such as SMB-1 (45.0%) from Serratia marcescens and THIN-B (44.6%) from Janthinobacterium lividum (7, 15). P. mexicana harbors bla_AIM-1-like (97.7% identity versus bla_AIM-1), suggesting that bla_AIM-1-like from P. mexicana is the origin of bla_AIM-1 (16). bla_AIM-1 is bracketed by two copies of ISCR15 in P. aeruginosa chromosome, as an acquired MBL gene (8). A previous study was unable to reproduce the mobilization of bla_AIM-1-like via ISCR15, suggesting that it may not be easy to mobilize it to other bacterial species (16). Meanwhile, Pseudoxanthomonas species harbor MBL genes with high diversity (~59% to 81%). In addition, their comprehensive genomes surrounding the MBL genes are well conserved and there are no transfer elements (Fig. 3). Although these MBL genes may not be easily transferred to other Gram-negative bacteria, they may have become an antimicrobial-resistance pool in the environment.

FIG 1.

Phylogenetic tree of metallo-β-lactamase based on amino acid sequence, constructed using MEGAX with maximum-likelihood and 1,000 replication bootstrapping with Jones-Taylor-Thornton model.

FIG 2.

Amino acid alignment of PJM-1 and other closely related subclass B3 metallo-β-lactamases. The red box indicates zinc-binding regions.

FIG 3.

Comprehensive similarities in the genome sequences of surrounding metallo-β-lactamase genes in MBL-producing Pseudoxanthomonas species. The block arrows indicate putative open reading frames (ORFs) and their orientation. The black and orange arrows indicate genes encoding MBL and other protein coding sequences, respectively.

To characterize the kinetic properties of PJM-1, a recombinant PJM-1 protein was purified from Escherichia coli BL21(DE3). Briefly, cloning was performed using the In-Fusion HD Cloning kit (TaKaRa Bio, Shiga, Japan) after PCR amplification of bla_PJM-1 and the linear pET9a vector using the following primers with a 15-bp overlapping sequence (underlined): pET9a_F_Infusion, 5′-TAACAAAGCCCGAAAGGAAGC-3′; pET9a_R_Infusion, 5′-ATGTATATCTCCTTCTTAAAGTTA-3′; F-PJM_infusion, 5′-GAAGGAGATATACATATGACGCTCCGCTCCACC-3′; and R-PJM_infusion1, 5′-TTTCGGGCTTTGTTATCAAGGCGCAGGCACGGG-3′. The antimicrobial susceptibility of the transformed E. coli BL21(DE3) recombinant strain and the wild type was determined according to CLSI (Table 1). After growth at 35°C in LB broth containing 50 μg/mL kanamycin, induction with 1 mM isopropyl β-thiogalactopyranoside was performed. After 4 h of induction, the cell pellet was collected by centrifugation, resuspended in 20 mM HEPES buffer, and sonicated before ultracentrifugation at 50,000 rpm for 30 min. The supernatant was loaded onto a HiTrap Q column (Cytiva, MA, USA) and recombinant PJM-1 was eluted using a linear NaCl gradient (0 to 200 mM) in 20 mM HEPES buffer. The purified enzyme was dialyzed using 50 mM HEPES buffer (pH 7.5) with 50 μM ZnSO₄ and analyzed using SDS-PAGE. Recombinant PJM-1 was obtained at >95% purity, which fits with the consensus on β-lactamase nomenclature (17). The molecular mass and predicted isoelectric point of PJM-1 were estimated to be 29,800 kDa without the predicted signal peptide sequence and 4.99, respectively, using ExPASy (https://web.expasy.org/compute_pi/).

The kinetic parameters of purified PJM-1 were determined as previously described (18, 19). The K_i values of ampicillin, piperacillin, and imipenem were determined using nitrocefin as the reporter substrate. The IC₅₀ values of EDTA (Nacalai Tesque, Kyoto, Japan) and 2,6-pyridinedicarboxylic acid (DPA Sigma-Aldrich, St. Louis, MO, USA) were determined using cephalothin as a reporter substrate. Three individual assays were performed and the mean ± standard deviation for K_m, k_cat, and K_i was calculated (Table 2).

TABLE 2.

Kinetic parameter of PJM-1 and other class B3 metallo-β-lactamases

Substrate	PJM-1			AIM-1c			kcat/Km (M−1 s−1)
	kcat (s−1)	Km (μM)	kcat/Km (M−1 s−1)	kcat (s−1)	Km (μM)	kcat/Km (M−1 s−1)	L1c	SMB-1c	THIN-Bc
Ampicillin	370 ± 3.1	51 ± 0.55b	7.2 × 106	594	41	1.4 × 106	4.4 × 106	2.4 × 106	3.7 × 105
Piperacillin	150 ± 4.4	14 ± 0.44b	1.1 × 107	337	17	2 × 107	7.0 × 106	1.8 × 105	2.0 × 105
Cephalothin	38 ± 0.36	13 ± 0.19	2.9 × 106	529	38	1.4 × 107	-	1.9 × 106	-
Cefuroxime	86 ± 1.8	120 ± 3.8	6.9 × 105	292	29	9.9 × 106	2.7 × 106	1.4 × 106	2.8 × 106
Cefoxitin	31 ± 0.50	26 ± 0.98	1.2 × 106	145	26	5.7 × 106	5.5 × 105	-	-
Cefotaxime	90 ± 1.9	130 ± 7.0	7.1 × 105	609	49	1.2 × 107	2.6 × 106	8.9 × 105	2.0 × 106
Ceftazidime	0.091 ± 0.00036	150 ± 3.7	6.2 × 102	7	148	4.9 × 104	-	7.7 × 104	1.4 × 105
Cefepime	1.0 ± 0.027	230 ± 6.1	4.3 × 103	93	594	1.6 × 105	1.9 × 104	3.6 × 103	7.9 × 103
Imipenem	66 ± 2.1	36 ± 1.4b	1.8 × 106	1,700	97	1.7 × 107	7.5 × 105	3.9 × 106	1.5 × 106
Meropenem	110 ± 3.8	34 ± 2.3	3.3 × 106	1,000	163	6.8 × 106	4.5 × 106	4.2 × 106	5.0 × 106
Ertapenem	190 ± 3.4	57 ± 1.3	3.3 × 106	-f	-	-	-	-	-
Aztreonama	NHd	>1,000b	NDe	NH	NH	ND	ND	ND	ND
Nitrocefin	79 ± 1.5	2.6 ± 0.11	3.1 × 107	-	-	-	2.9 × 106	-	-

^aThe K_i value for aztreonam exceeds the measurable range.

^bK_i value.

^cThe references values of AIM-1, L1, SMB-1, and THIN-B are 8, 3, 7, and 15, respectively.

^dNH, no hydrolysis.

^eND, not determined.

^f-, no data available.

The MICs of PJM-1 for E. coli BL21 were significantly higher than those for nonrecombinant E. coli BL21 for almost all β-lactams; however, the MICs of ceftazidime and cefepime remained low (Table 1). In addition, in the presence of 100 μM EDTA, the MICs were reduced by 4- to 8-fold, indicating that EDTA strongly inhibits the catalytic activity of PJM-1.

PJM-1 showed a high catalytic activity against penicillin antibiotics, cephalothin, cefoxitin, and carbapenems (k_cat/K_m > 10⁶ M⁻¹s⁻¹), which were better substrates than broad-spectrum cephalosporins. In particular, the catalytic activity of PJM-1 against ceftazidime was lower than that of AIM-1 (~80-fold), SMB-1 (~120-fold), and THIN-B (~230-fold) (7, 8, 15). The overall kinetic properties (k_cat/K_m) of purified PJM-1 were found to be similar to those of L1 on the tested substrates (3). Although the amino acid sequence of AIM-1 is the most similar to that of PJM-1, the catalytic activity of AIM-1 was 2- to 80-fold higher than that of PJM-1, except for ampicillin, which may be a result of assay conditions such as temperature and buffer (8). The kinetic parameters could be influenced by various factors such as enzyme stability, folding, and presence of tags in addition to the assay conditions. Thus, comparison of kinetic parameters should be carefully performed, and unification of experimental methods is also needed.

Overall, the kinetic properties of the purified PJM-1 were in agreement with the antimicrobial susceptibility of PJM-1-producing recombinant E. coli. For instance, high catalytic activities (k_cat/K_m > 10⁶ M⁻¹s⁻¹) were observed against ampicillin, piperacillin, cephalothin, cefoxitin, and carbapenems, and the MICs were ≥64 mg/L. In contrast, low catalytic activities (k_cat/K_m ≤10⁻³ M⁻¹s⁻¹) were observed against ceftazidime and cefepime, and the MICs were low (ceftazidime, 2 mg/L; cefepime, <0.125 mg/L). The IC₅₀ values of cephalothin for EDTA and DPA were 35 and 100 μM, respectively, suggesting that PJM is more inhibited by EDTA than by DPA.

In conclusion, PJM-1 is a novel subclass B3 MBL that exhibits very low catalytic activity against ceftazidime and cefepime and is strongly inhibited by EDTA.