Characterization of acquired β-lactamases in Pseudomonas aeruginosa and quantification of their contributions to resistance

ABSTRACT

Pseudomonas aeruginosa is a highly problematic opportunistic pathogen that causes a range of different infections. Infections are commonly treated with β-lactam antibiotics, including cephalosporins, monobactams, penicillins, and carbapenems, with carbapenems regarded as antibiotics of last resort. Isolates of P. aeruginosa can contain horizontally acquired bla genes encoding β-lactamase enzymes, but the extent to which these contribute to β-lactam resistance in this species has not been systematically quantified. The overall aim of this research was to address this knowledge gap by quantifying the frequency of β-lactamase-encoding genes in P. aeruginosa and by determining the effects of β-lactamases on susceptibility of P. aeruginosa to β-lactams. Genome analysis showed that β-lactamase-encoding genes are present in 3% of P. aeruginosa but are enriched in carbapenem-resistant isolates (35%). To determine the substrate antibiotics, 10 β-lactamases were expressed from an integrative plasmid in the chromosome of P. aeruginosa reference strain PAO1. The β-lactamases reduced susceptibility to a variety of clinically used antibiotics, including carbapenems (meropenem, imipenem), penicillins (ticarcillin, piperacillin), cephalosporins (ceftazidime, cefepime), and a monobactam (aztreonam). Different enzymes acted on different β-lactams. β-lactamases encoded by the genomes of P. aeruginosa clinical isolates had similar effects to the enzymes expressed in strain PAO1. Genome engineering was used to delete β-lactamase-encoding genes from three carbapenem-resistant clinical isolates and increased susceptibility to substrate β-lactams. Our findings demonstrate that acquired β-lactamases play an important role in β-lactam resistance in P. aeruginosa, identifying substrate antibiotics for a range of enzymes and quantifying their contributions to resistance.

IMPORTANCE

Pseudomonas aeruginosa is an extremely problematic pathogen, with isolates that are resistant to the carbapenem class of β-lactam antibiotics being in critical need of new therapies. Genes encoding β-lactamase enzymes that degrade β-lactam antibiotics can be present in P. aeruginosa, including carbapenem-resistant isolates. Here, we show that β-lactamase genes are over-represented in carbapenem-resistant isolates, indicating their key role in resistance. We also show that different β-lactamases alter susceptibility of P. aeruginosa to different β-lactam antibiotics and quantify the effects of selected enzymes on β-lactam susceptibility. This research significantly advances the understanding of the contributions of acquired β-lactamases to antibiotic resistance, including carbapenem resistance, in P. aeruginosa and by implication in other species. It has potential to expedite development of methods that use whole genome sequencing of infecting bacteria to inform antibiotic treatment, allowing more effective use of antibiotics, and facilitate the development of new antibiotics.

INTRODUCTION

The opportunistic pathogen Pseudomonas aeruginosa can cause severe and life-threatening infections in burn wounds, surgical sites, lungs of people with cystic fibrosis, and immunocompromised patients (1–7). β-lactam antibiotics play a key role in the treatment of P. aeruginosa infections (8, 9), with a large variety of β-lactams being used, including the penicillins ticarcillin and piperacillin, the cephalosporins ceftazidime and cefepime, the carbapenems imipenem and meropenem, and the monobactam aztreonam (3, 9, 10). Of particular importance is the carbapenem subclass as they are generally reserved as a last line of defense in the treatment of known or suspected multidrug-resistant bacteria due to their ability to resist hydrolysis by most β-lactamases (11–13). However, treatment is becoming more difficult because of the increasing frequency of β-lactam-resistant isolates (2, 14, 15). Indeed, the World Health Organization has classified carbapenem-resistant P. aeruginosa as a high-priority group for the development of new treatments (16). Novel β-lactams that inhibit the growth of P. aeruginosa have been developed, but resistance to these can arise in the same way as resistance to currently used antibiotics (17, 18).

There are many mechanisms by which β-lactam-susceptible strains of P. aeruginosa can become resistant. These include mutations that increase expression or alter substrate specificity of multidrug efflux pumps and the intrinsic low-activity β-lactamases AmpC and an OXA-50-like β-lactamase, resulting in increased export and degradation of β-lactams (19–23); mutations in porins responsible for uptake of β-lactams (24–26); mutations in penicillin-binding proteins that are the targets of β-lactam antibiotics; and acquisition of β-lactamases through horizontal gene transfer (27–32). These can be combinatorial, with resistant isolates often having several different mechanisms for reducing susceptibility to β-lactams.

β-lactamases hydrolyze the β-lactam ring of substrate compounds, inactivating the antibiotics (33). These enzymes are categorized into four classes (A to D) based on sequence homology. Enzymes in classes A, C, and D have a catalytic serine for substrate hydrolysis, whereas class B are metallo-enzymes that catalyze the hydrolysis of β-lactam rings using a metal ion, most commonly a zinc ion (34, 35). Collectively, serine β-lactamases can hydrolyze all β-lactams in clinical use, and metallo-β-lactamases can hydrolyze all clinically used β-lactams except monobactams, although most individual enzymes are only active against a subset of compounds (32, 36, 37). β-lactamases capable of degrading carbapenems (carbapenemases) are especially problematic due to the last-resort role of carbapenems in managing severe infections by multidrug-resistant P. aeruginosa. Carbapenemases can be present in carbapenem-resistant isolates of P. aeruginosa that are often a cause of severe infections (27, 38–41). β-lactamase inhibitors, such as tazobactam, clavulanate, avibactam, vaborbactam, and relebactam, that inhibit some serine β-lactamases can enhance the activity of β-lactam antibiotics by reducing the rate of antibiotic hydrolysis and are often co-administered with β -lactams (42–47).

P. aeruginosa isolates can acquire bla genes that encode β-lactamases through horizontal gene transfer (28–32). The bla genes are typically carried on mobile elements, predominantly chromosomally located composite transposons, although they are on plasmids in some cases (48, 49). Horizontal gene transfer can occur between different species of bacteria, providing the potential for P. aeruginosa to acquire β-lactamases from unrelated species, such as from a variety of Enterobacteriaceae species (50).

Although the substrate antibiotics of some β-lactamases have been determined (31, 51, 52), new enzyme variants are regularly discovered (53–55), and the substrates of most of these are unknown. Consequently, in many cases, it is not known which antibiotics will be effective against a β-lactamase-containing isolate. Furthermore, although many studies have reported the presence of horizontally acquired β-lactamases in P. aeruginosa, there is currently very limited quantitative information on the extent to which these enzymes reduce β-lactam susceptibility in this species. The overall goal of this research was to quantify the effects of acquired β-lactamases on resistance of P. aeruginosa to different β-lactams.

RESULTS

Prevalence of acquired β-lactamases in P. aeruginosa

Determining the frequency of acquired bla genes within the P. aeruginosa population is a first step towards quantifying their contribution to resistance in this species. The genomes of 858 P. aeruginosa isolates with unknown antibiotic resistance phenotypes from the International Pseudomonas Consortium Database (56, 57) were interrogated for the presence of bla genes. The isolates (Table S1) were of worldwide origin (five continents, multiple different countries and centers) and were from a range of different infections and from environmental and veterinary sources. Of the 858 isolates, 26 (3%) had acquired bla genes. These isolates harbored a wide range of acquired genes (Table 1; Table S1). Genes encoding class D enzymes were the most frequent, with OXA-10 being the most prevalent of these (39% of class D enzymes). Class A enzymes were the next most common, with CARB-2 being the most prevalent of these (57%). Class B enzymes were the least frequent, with VIM-2 being the most prevalent (50%). No acquired Class C β-lactamases were found.

TABLE 1.

β-lactamases in uncharacterized and meropenem-resistant P. aeruginosa isolates

Class A	Number of genomes	Class B	Number of genomes	Class D	Number of genomes
β-lactamases encoded by 858 uncharacterized P. aeruginosa genomes
CARB-2	8	IMP-14	1	OXA-2	1
VEB-1	4	IMP-54	1	OXA-4	1
VEB-2	3	SPM-1	2	OXA-10	8
		VIM-2	4	OXA-21	1
				OXA-56	6
				OXA-129	1
β-lactamases encoded by genomes of 238 meropenem-resistant P. aeruginosa
CARB-2	4	IMP-1	1	LCR-1	2
GES-1	5	IMP-7	1	OXA-1	1
GES-5	6	IMP-31	2	OXA-2	13
GES-9	1	VIM-2	18	OXA-4	3
GES-19	4	VIM-4	3	OXA-10	8
GES-20	3	VIM-47	4	OXA-17	1
KPC-2	1			OXA-35	2
PER-1	1			OXA-36	1
SHV-12	2			OXA-74	1
VEB-1	4			OXA-233	1

We next investigated whether bla genes are enriched in isolates of P. aeruginosa that are resistant to meropenem, a carbapenem antibiotic. The frequency of acquired β-lactamases was determined in 238 independent meropenem-resistant P. aeruginosa isolates in our collection (clinical isolates of worldwide origin) and from a previous study (clinical isolates from hospitals in Europe) (58) (Table 1; Table S2). Sixty-one of these isolates (35%) had acquired one bla gene and 23 had acquired multiple genes (Table S2), indicating a strong selection for β-lactamases in carbapenem-resistant isolates. The most common β-lactamases found in carbapenem-resistant isolates for each class were GES-5 for class A, VIM-2 for class B, and OXA-2 for class D, all of which have activity against carbapenems, including meropenem. Meropenem resistance is multi-factorial, and other factors, such as absence of OprD or increased expression of efflux pumps, may also contribute to the meropenem-resistant phenotype of these isolates.

The presence of different bla genes in different isolates implies multiple independent acquisitions of bla genes into P. aeruginosa, although the presence of identical bla genes in some cases suggested the possible occurrence of multiple independent isolates of a clonal strain. A phylogenetic tree was therefore constructed with genomes of 64 representative P. aeruginosa isolates from five different continents and including 20 carbapenem-resistant isolates, 21 susceptible isolates, and 23 isolates of unknown phenotype (Fig. 1). An extended phylogenetic tree with all isolates containing acquired β-lactamase was also constructed (Fig. S1).

Fig 1.

Phylogenetic tree of representative β-lactamase containing P. aeruginosa isolates. The tree comprises 37 isolates with acquired β-lactamases and 27 isolates with no acquired β-lactamases (Table S3). The identities of the β-lactamases are shown. Meropenem-resistant isolates are shown in red, meropenem-susceptible isolates in green, and isolates with unknown phenotype in black. Five P. aeruginosa reference strains (in blue) are also included (59). Sequence types (ST) are shown.

The genomes (Fig. 1; Fig. S1) have a similar distribution to a larger tree that represents the breadth of the P. aeruginosa species (60), showing that it encompasses most of the genetic diversity of the P. aeruginosa species, except for the absence of isolates related to the PA7 subgroup. Acquired β-lactamases were present in a wide distribution of carbapenem-resistant isolates and isolates of unknown phenotype (Fig. 1; Fig. S1). The P. aeruginosa isolates have a few groups of closely related strains harboring the same β-lactamases (Fig. 1; Fig. S1), indicating that these clones are highly successful. Notably, there are multiple isolates of the highly virulent clone ST235 (61–63), and these isolates have acquired a range of β-lactamases. There are also multiple isolates of other high-risk clones, including ST111 (61, 63) and ST277 (61, 64–66).

Substrate specificities of β-lactamases

Substrate antibiotics were determined for a range of β-lactamases to better understand the contributions of different enzymes to resistance. Genes encoding β-lactamases from P. aeruginosa isolates were cloned and expressed in P. aeruginosa PAO1, and the effects on MICs of representative antibiotics from all β-lactam classes were measured (Table 2). The β-lactamases chosen that were prevalent in clinical isolates (VIM-2, OXA-10) (Table 1) had not been investigated previously (OXA-21, OXA-129, and the OXA-2-related OXA-226) and/or in β-lactamase families associated with carbapenem resistance (KPC-2, GES-19, GES-20, VIM-2, SPM-1. and IMP-54) (13, 42, 67). Most bla genes are present in the chromosome or on a low copy number of plasmids (27), and so a chromosome-integrating vector (pSW196) was used to represent the normal situation. MIC testing was performed on P. aeruginosa PAO1 expressing the β-lactamases (Table 2).

TABLE 2.

MICs of P. aeruginosa PAO1 expressing bla genes^c

Expressed β-lactamase	Antibiotic MICs mg/La
	Carbapenems		Penicillins			Cephalosporins		Monobactam
	MEM	IMI	CARB	PIP	TIC	CTZ	CPM	AZT
Noneb	0.5	2	16	2	8	0.5	0.5	2
Empty vector	0.5	2	16	2	8	0.5	0.5	2
Class A
KPC-2	8	4	512	16	128	4	16	32
GES-19	2	2	512	16	256	512	64	64
GES-20	16	2	128	16	32	8	2	2
Class B
VIM-2	16	16	>1024	16	512	16	8	2
SPM-1	32	8	128	8	32	256	128	2
IMP-54	32	4	32	2	16	64	64	2
Class D
OXA-226	0.5	2	32	2	32	16	1	8
OXA-10	0.5	1	128	8	64	0.5	2	4
OXA-129	0.5	1	32	4	32	0.5	1	2
OXA-21	1	1	256	8	128	2	1	2

^a Median of three replicates, with all replicate values in Table S4.

^b None, strain PAO1; empty vector, PAO1 containing pSW196 vector with no cloned bla gene.

^c MEM = Meropenem, IMI = Imipenem, PIP = Piperacillin, TIC = Ticarcillin, CTZ = ceftazidime, CPM = cefepime, AZT =Aztreonam, CARB = Carbenicillin.

The 10 β-lactamases all increased the MIC for antibiotics in at least one class of β-lactams, with the spectrum of substrates and the extent of the increase differing between enzymes. The three class A β-lactamases (KPC-2, GES-19, and GES-20) and three class B enzymes (VIM-2, SPM-1, and IMP-54) all increased the MICs for at least six of eight β-lactams tested, with KPC-2 increasing the MIC of all β-lactams. Class B β-lactamases did not increase aztreonam MIC, but some enzymes in classes A and D did. Classes A and B enzymes had carbapenemase activity, increasing the MIC for at least one carbapenem, whereas the four class D β-lactamases had little or no carbapenemase activity.

MICs for β-lactamase-containing isolates

The MICs of P. aeruginosa containing a variety of acquired bla genes were determined for comparison with strain PAO1 expressing the same genes (Table 3).

TABLE 3.

MICs of isolates with acquired β-lactamases

Isolate	Acquired β-lactamase	Antibiotic MICs mg/La
		Carbapenems		Penicillins			Cephalosporins		Monobactam
		MEM	IMI	CARB	PIP	TIC	CTZ	CPM	AZT
PA01	None	0.5	2	16	2	8	0.5	0.5	2
Class A
1255885	KPC-2	1024	256	>2048	256	2048	64	256	512
Class D
1257147	OXA-1/17/129	8	1	1024	128	512	4	64	16
1260715	OXA-10	16	8	256	16	128	1	2	2
1586981	OXA-2	32	16	1024	128	256	64	16	32
1607533	OXA-2	8	16	128	16	32	8	8	8
AUS106	OXA-21	16	8	512	64	256	8	8	8
Classes A and D
1275655	OXA-2, GES-19/19/20	128	32	>2048	128	1024	2048	256	256
1324459	OXA-10/233 VEB-1	64	16	2048	32	1024	1024	256	1024
1420275	OXA-2, GES-19/20	256	34	512	128	256	2048	256	128
1690076	OXA-2, GES-19/20	16	8	1024	32	256	256	64	32
Classes B and D
1260990	OXA-10, VIM-2	128	128	>2048	64	2048	16	16	8
172BR	SPM-1, OXA-56	256	128	1024	32	128	256	128	16
4068	OXA-10, IMP-54	>2048	256	>2048	128	>2048	512	2048	128

^a Median of three replicates, with all replicate values shown in Table S5. Values at or above the threshold for resistance (68) are shown in bold font.

Most bla-containing isolates had higher MICs than reference strain P. aeruginosa PAO1 for all β-lactams tested (Tables 2 and 3). However, the results were consistent with the effects of cloned bla genes expressed in strain PAO1. All P. aeruginosa strains with a β-lactamase that increased carbapenem MIC in strain PAO1 (Table 2) were resistant to both carbapenems meropenem and imipenem (Table 3). Class B and D enzymes had little or no effect on the MIC for aztreonam in strain PAO1, and four isolates that were susceptible to aztreonam had only class B and/or D β-lactamases. All the clinical isolates that were susceptible to cephalosporins had only class D β-lactamases, which showed little to no activity against cephalosporins in strain PAO1 (Table 2).

Deletion of carbapenemase genes from clinical isolates

In order to quantify the contributions of carbapenemases to resistance, genes encoding three carbapenemases VIM-2, GES-20 and IMP-54 were deleted from the genomes of P. aeruginosa isolates. These genes were selected for deletion because they conferred the highest increases in meropenem MIC when expressed in P. aeruginosa PAO1 (Table 2), because carbapenems are “last-resort” antibiotics in treatment of P. aeruginosa infections, and because complete genome sequences necessary for engineering the deletions were available. The genes encoding GES-19 and GES-20 are adjacent to each other and flanked by repetitive DNA sequences, making it necessary for technical reasons to delete blaGES19 as well as blaGES20. MIC testing was performed on the engineered deletion-containing mutants (Table 4).

TABLE 4.

MICs of β-lactams for P. aeruginosa isolates with deletions of acquired carbapenemase genes

Isolate	Antibiotic MICs mg/La
	Carbapenems		Penicillins			Cephalosporins		Monobactam
	MEM	IMI	CARB	PIP	TIC	CTZ	CPM	AZT
PA01	0.5	2	16	2	8	0.5	0.5	2
1260990	128	128	2048	64	1024	16	16	8
1260990 ΔVIM-2	16	16	512	32	256	2	8	8
1420275	256	32	>2048	128	1024	2048	256	128
1420275 ΔGES-19/20	16	8	256	8	64	4	8	32
4068	>2048	256	>2048	128	>2048	512	2048	128
4068ΔIMP-54	256	8	>2048	128	2048	4	64	64

^a Median of three replicates, all replicates are in Table S6. Values at or above the threshold for resistance (68) are shown in bold font.

Deletion of the blaVIM-2 gene from the genome of isolate 1260990 reduced the MIC for all antibiotics except aztreonam, consistent with the effect of this gene on strain PAO1 (Table 2), although the bacteria remained resistant to carbapenems and ticarcillin. Deletion of the blaGES-19 and blaGES-20 genes from the genome of isolate 1420275 reduced the MIC of all antibiotics tested, making the bacteria susceptible to penicillins, cephalosporins, and aztreonam and consistent with the activity of class A enzymes against all classes of β-lactams. Deletion of the blaIMP-54 gene from isolate 172BR also reduced the MICs of all β-lactams although only rendered the isolate susceptible to ceftazidime. Overall, none of the gene deletions were sufficient to render the isolates carbapenem-susceptible, but all three deletions reduced carbapenem MIC. The decreases in MICs caused by deleting bla genes are consistent with the increases in MICs that the β-lactamases provide when expressed in P. aeruginosa PAO1 (Table 2). All three of these isolates had deletions or other mutations that inactivate OprD, the porin that is responsible for uptake of carbapenems, and the absence of functional OprD may explain their resistance to carbapenems despite the absence of the bla genes (69, 70).

Effectiveness of tazobactam on P. aeruginosa containing acquired β-lactamase genes

The β-lactamase inhibitor tazobactam is commonly co-administered with β-lactams to restore their function against β-lactamase-containing isolates (71, 72). To investigate the effectiveness of tazobactam against P. aeruginosa containing acquired β-lactamases, antibiotic MICs were determined for a P. aeruginosa PAO1 mutant that has a deletion of ampC. This mutant was used to avoid any confounding effects of the intrinsic β-lactamase AmpC, as inhibition of AmpC by tazobactam might affect the MICs. β-lactams that had a higher MIC in the presence of one or more β-lactamases (Table 2) were tested (Table 5).

TABLE 5.

Effect of tazobactam on P. aeruginosa PAO1 ampC expressing acquired β-lactamases

Antibiotic MICs mg/La
β-lactamase	MEM	MEM+TZb	β-lactamase	CPM	CPM + TZ	β-lactamase	TIC	TIC+TZ	β-lactamase	CTZ	CTZ+TZ
Nonec	0.5	0.5	None	1	1	NA	8	8	None	0.5	0.5
GES-20	16	16	KPC-2	16	16	OXA-10	64	64	OXA-226	16	8
VIM-2	16	16	GES-19	64	64	OXA-129	32	32
			SPM-1	128	128	OXA-21	128	16
			IMP-54	64	64	OXA-226	32	16

^a Median of three replicates, all replicates are in Table S7. MICs that were lower in the presence of tazobactam are indicated in bold font.

^b TZ = Tazobactam.

^c None, PAO1ΔampC.

Tazobactam had activity against two class D β-lactamases, OXA-21 and to a lesser extent OXA-226. Class A enzymes (KPC-2 and GES19/20) were not inhibited by tazobactam, consistent with previous findings for KPC-2 (47, 73). Class B β-lactamases (metallo-β-lactamase, VIM-2, IMP-54, and SPM-1) were also not inhibited by tazobactam, which is known to lack activity against class B enzymes (33).

bla genes are parts of complex mobile genetic elements

Mobile genetic elements allow for the spread of β-lactamase-encoding genes and a large range of other antibiotic-resistance genes (27, 74). The mobile elements that contained β-lactamases were investigated for carbapenem-resistant isolates for which complete (fully assembled) genome sequences were available to determine how the β-lactamases were likely acquired and spread. Genes encoding VIM-2 and OXA-10 (Fig. S2a), and GES-19, GES-20, and OXA-2 (Fig. S2b) are on transposons in the chromosomes of the isolates 1260990 and 1420275, respectively. The blaKPC-2 gene present in isolate 1255885 is encoded by a large plasmid (44 kb), named in this study as pPAE1255885 (Fig. S2c).

The transposon-located resistance genes were predominantly found in non-Pseudomonas species, in particular members of the Enterobacteriaceae. These included members of the Aeromonas, Citrobacter, Enterobacter, Klebsiella, and Raoultella genera, indicating that these transposons or the genes on them have likely been acquired by P. aeruginosa from other species. The transposons contained multiple genes for resistance to other compounds, notably aminoglycoside antibiotics and also mercury and sulfonamide (Fig. S2).

The blaKPC-2 gene is part of a plasmid pPAE1255885 (Fig. S2c) that is very similar to plasmids pPA2047 and pPae-13 (75, 76), as well as plasmids pPA-1 and pPA-2 (BioSamples; SAMN28550649 and SAMN28550875, respectively). The main differences are multiple rearrangement/duplication events, although pPae-13 lacks a transposon present in the other plasmids. Plasmid pPAE1255885 was transformed into P. aeruginosa PAO1, and MIC testing was carried out (Table 6). The MIC for P. aeruginosa PAO1 containing the KPC-2-encoding pPAE1255885 was significantly higher for all antibiotics tested than for P. aeruginosa expressing KPC-2 from the chromosomally integrated vector. The difference in MIC may be due to an uncharacterized antibiotic resistance mechanism on pPAE1255885, or to the multi-copy nature of pPAE1255885, compared with the chromosomally expressed KPC-2 gene. Isolate 1255885 that contains pPAE1255885 had similar MIC values to strain PAO1 harboring the same plasmid (Table 6). This indicates that the majority of the β-lactam resistance in 1255885 is due to the plasmid.

TABLE 6.

Effect of plasmid pPAE1255885 on the MICs of P. aeruginosa PAO1

Strain	Antibiotic MICs mg/La
	Carbapenems		Penicillins			Cephalosporins		Monobactam
	MEM	IMI	CARB	PIP	TIC	CTZ	CPM	AZT
PAO1	0.5	2	16	2	8	0.5	0.5	2
PAO1 +KPC-2	8	4	512	16	128	4	16	32
PAO1 +pPAE1255885	256	128	>2048	256	2048	32	128	512
1255885	1024	256	>2048	256	2048	64	256	512

^a Median of three replicates, all replicate values are in Table S8.

DISCUSSION

The overall aim of this research was to quantify the contribution of acquired β-lactamases to the resistance of P. aeruginosa to β-lactam antibiotics, with a particular focus on carbapenems and carbapenemases. Our findings show that acquired bla genes are enriched in carbapenem-resistant isolates. Classes A and B enzymes, although not class D, had carbapenemase activity in P. aeruginosa. Deletion of bla genes from clinical isolates of P. aeruginosa increased susceptibility to multiple β-lactams, including carbapenems and allowed quantification of the effects of the corresponding β-lactamases on MIC. Tazobactam had limited activity against the enzymes characterized in this study.

Quantification of the effects of acquired bla genes can contribute to ongoing efforts to predict antibiotic susceptibility of P. aeruginosa from genome sequences (77, 78). Knowledge of the substrate specificities of different β-lactamases will also facilitate the application of artificial intelligence for development of new antibiotics and inhibitors (79, 80).

Although the effects of some of the enzymes tested here (KPC-2, GES-19, and VIM-2) had been assessed previously in E. coli for a subset of the antibiotics used here (51, 81–83), only SPM-1 and OXA-10 had previously been investigated for their effects in a P. aeruginosa reference strain (84, 85). So far, the remaining enzymes GES-20, IMP-54, OXA-21, OXA-129, and OXA-226, had not been characterized for any species. Substrates of β-lactamases were identified by expressing bla genes from an integrative vector that mimics the acquisition of genes into the chromosome of P. aeruginosa by integrative mobile genetic elements. This approach enabled an efficient survey of the effects of acquired β-lactamases of different classes on P. aeruginosa. Class A β-lactamases collectively increased the MICs of all tested β-lactams, although only KPC-2 increased MICs towards all β-lactams (Table 2). KPC-2 increases the MIC of imipenem, piperacillin, and ceftazidime when expressed in E. coli (82, 83), consistent with the results of this study. Although GES-19 and GES-20 are only two amino acid residues different in sequence, they had significantly different effects on β-lactam susceptibility, with GES-19 increasing the MIC of penicillins, cephalosporins, and monobactams, and GES-20, that of meropenem (Table 2). Consistent with these results, deletion of both GES-19 and GES-20 genes from a clinical isolate reduced the MIC of all β-lactams tested (Table 4). Intriguingly, deletion of both GES-19 and GES-20 reduced the MIC for imipenem, whereas expression of the individual genes did not increase the MIC of reference strain PAO1 (Table 2). The effect of co-expression of both GES-19 and GES-20 genes in strain PAO1 remains to be determined.

Class B β-lactamases increased the MIC of all β-lactams tested except for the monobactam aztreonam (Tables 2 and 4). These findings are consistent with previous findings in E. coli (32, 34, 37), as well as the known lack of activity of class B β-lactamases against monobactams (32), and suggest that aztreonam could be used to treat clinical isolates with only class B β-lactamases. The activity of Class D β-lactamases was largely limited to penicillins and cephalosporins, except for OXA-226 that increased the MICs for ceftazidime and aztreonam MICs. For example, OXA-10 increased piperacillin, ticarcillin, cefepime, and aztreonam MIC when expressed in P. aeruginosa PAO1 (Table 2), consistent with an earlier study (85).

The β-lactamase inhibitor tested here, tazobactam, is commonly co-administered with β-lactams and has been reported to inhibit many class A and some class C and D β-lactamases, although not KPC β-lactamases (47, 73, 86–88). Tazobactam inhibits GES-2 that has a very similar sequence (two or three amino acids different) to both GES-19 and GES-20 (88) but did not affect the MIC of P. aeruginosa PAO1 expressing GES-19 or GES-20 (Table 5) indicating that the differences between enzymes are sufficient to make GES-19 and GES-20 resistant to tazobactam. Instead, tazobactam only showed activity against the class D β-lactamases OXA-21 and OXA-226 (Table 5). Tazobactam also has activity against OXA-2 (47) that is very similar to OXA-226 and relatively closely related to OXA-21. These findings indicate that tazobactam likely has activity against other class D β-lactamases closely related to OXA-2 but has variable effects on class A enzymes. Other β-lactamase inhibitors, such as avibactam and relebactam, are effective against class A β-lactamases (89, 90). It will be of importance to investigate the activities of a range of β-lactamase inhibitors against enzymes expressed in strain PAO1 to enable comparison with direct enzymatic assays (86, 91). Relating this information to the frequencies at which different β-lactamases are present in P. aeruginosa may allow fine-tuning of antibiotic/inhibitor combinations in treating P. aeruginosa infections.

Although acquired bla genes are at a low but significant level (3%) in the P. aeruginosa population, they are enriched in carbapenem-resistant isolates (35% of isolates), indicating their importance to resistance. It should be noted that most currently available bacterial genome sequences, including almost all of those analyxed here, are assembled from short read-data that leads to poor assembly of mobile elements (92, 93), and so the frequency of bla genes could be significantly underestimated in both carbapenem-resistant and carbapenem-susceptible sub-populations of P. aeruginosa. Knowing the distribution and frequency of acquired β-lactamases in the P. aeruginosa population, and in other bacteria, can contribute to development of approaches to manage the spread of the threat.

Most clinical isolates harboring bla genes had higher MICs than P. aeruginosa PAO1 expressing individual genes (Tables 2 and 3). For example, unlike strain PAO1 all five isolates with a class A β-lactamase were clinically resistant towards all β-lactams (Table 3), except that one isolate had only intermediate resistance towards piperacillin. The higher MICs relative to strain PAO1 expressing individual bla genes could be due in part to the presence of multiple bla genes in many of the isolates. About a quarter (24.5%) of isolates that have class B β-lactamases also have either a class A or D β-lactamase, and over half (54%) of isolates with a class D enzyme also have other acquired β-lactamases (Table 3; Tables S1 and S2). However, the higher MICs of the clinical isolates could also be due to the presence of mutations affecting antibiotic efflux or uptake, or mutations in penicillin-binding proteins. For example, mutations inactivating OprD explain the finding that deletion of bla genes encoding enzymes with carbapenemase activity (GES-20, VIM-2 and IMP-54) reduced MICs of carbapenems and other β-lactams, it did not render the bacteria carbapenem-susceptible (Table 4). The relative sensitivity of strain PAO1 to β-lactams means that MICs of this strain expressing individual β-lactamases can likely be used to indicate the minimum MICs of different antibiotics for clinical isolates with the same β-lactamases. bla-containing isolates and carbapenem-resistant isolates encompass a large range of the genetic diversity of the P. aeruginosa species. This indicates that bla genes have been acquired by P. aeruginosa on multiple occasions, and that most if not all P. aeruginosa isolates are capable of acquiring bla genes and becoming carbapenem-resistant. Notably, high-risk clones, such as ST235, that are associated with poorer patient outcomes following infection (61–63) have acquired a range of different bla genes. The genes encoding VIM-2 and OXA-10 are co-located on a transposon in the chromosome, as well as those encoding GES-19, GES-20, and OXA-2. Almost identical transposons were identified in a wide range of Gram-negative genera, indicating that the transposons have likely been acquired by P. aeruginosa from other species. The presence of bla genes on mobile genetic elements that frequently contain multiple resistance genes provides opportunities to acquire multiple bla genes simultaneously. A significant proportion of bla-containing isolates (28%) had multiple bla genes. The GES-19 and GES-20 genes are adjacent to one another in all three clinical isolates that contain them (Table 3; Table S2; Fig. S1b) and are often reported together in literature (94, 95). The presence of multiple different bla genes on the same element indicates that combinations are selected most likely due to additive or synergistic effects of different enzymes in conferring resistance to a wider range of antibiotics. The transposons also carry genes for resistance to aminoglycoside antibiotics, emphasizing how multiple antibiotic resistance genes can be acquired in a single genetic event.

The blaKPC-2 gene is located on a plasmid, pPAE1255885 (Fig. S2). Strain PAO1 containing this plasmid had very similar MIC values to the clinical isolate containing the same plasmid (Table 6), indicating that the plasmid may be responsible for the majority of β-lactam resistance in this isolate. The plasmid resulted in significantly higher MICs for strain PAO1 than KPC-2 expressed alone (Tables 2 and 6), although KPC-2 is the only gene on pPAE1255885 known to contribute directly to antibiotic resistance. The difference may be due to increased expression of KPC-2 from the multi-copy plasmid or may indicate the presence of an uncharacterized mechanism of β-lactam resistance on pPAE1255885. Collectively, these examples emphasize how mobile elements can carry a diverse range of β-lactamases and other antibiotic resistance genes. Many mobile elements are poorly characterized, and their prevalence in bacterial populations is not well understood. However, understanding and reducing the spread of mobile elements will likely be important in combating antibiotic resistance.

In conclusion, acquired β-lactamases contribute towards resistance to a large range of β-lactams in P. aeruginosa. Different β-lactamases of both the same and different classes have differences in their target β-lactams and also the extent of their effect on MIC. β-lactamases likely play a significant role in carbapenem resistance as they are highly enriched in carbapenem-resistant P. aeruginosa clinical isolates. Because many of the β-lactamases studied here are present in a wide range of bacterial species (54), determining their activities has potential to inform decisions on antibiotic treatment of infections by other bacterial species nd P. aeruginosa.

MATERIALS AND METHODS

Bacteria and growth conditions

Bacterial strains and plasmids used in this study are listed in Table S9. Bacteria were grown at 37 ◦C in Lysogeny broth (LB) with shaking (200 rpm) or on Luria agar (LA) unless stated otherwise, with additional chemicals added as required. Gentamicin (E. coli, 20 µg/mL; P. aeruginosa, 30 mg/L) was used to select for bacteria containing plasmid pJN105 tetracycline (E. coli, 12.5 mg/L; P. aeruginosa, 25 mg/L) for bacteria containing plasmid pEX18Tc or pSW196 and meropenem (2 mg/L) for P. aeruginosa strain PAO1 containing plasmid pPAE1255885. For conjugation, ST18 was used as a donor strain and grown in media supplemented with δ-aminolaevulinic acid (ALA) (50 mg/L), and P. aeruginosa recipients were grown at 42°C in LB with 0.4% KNO₃ as described previously (96–98). For expression of cloned genes from plasmid pSW196, arabinose (5 mg/mL) was included in the growth medium.

Molecular biology methods

Genomic DNA was purified using the UltraClean Microbial Kit (Qiagen, Hilden, Germany), plasmids were purified using the NucleoSpin Plasmid kit (Macherey Nagel, Dueren, Germany), and PCR products were purified and extracted from agarose gels using the NucleoSpin Gel and PCR Clean-up kit (Macherey Nagel, Dueren, Germany) according to the manufacturers’ instructions. For DNA cloning, PCR amplification was carried out with genomic DNA from isolates listed in Table S9 as template and with appropriate primers (Table S10). Amplicons for expression of the bla genes included the predicted ribosome-binding sites. Amplicons were cloned into pSW196 (99) for expression of cloned genes in P. aeruginosa. DNA cloning was carried out using restriction enzymes and T4 DNA ligase (New England Biolabs, Ipswich, MA, USA). DNA sequencing (Genetics Analysis Service, University of Otago, Dunedin, New Zealand) was carried out with appropriate primers on all plasmid constructs to confirm that the intended plasmids had been obtained with no mutations.

Genetic manipulation

bla genes that had been cloned into pSW196 were transferred into the chromosome of the β-lactam-susceptible reference strain PAO1 by conjugation (97). Plasmid pPAE1255885 was transformed into P. aeruginosa PAO1 that had been made competent by the sucrose washing method (100). bla and ampC genes were deleted from the genomes of P. aeruginosa isolates using two-step allelic exchange as previously described (96–98). In brief, PCR amplification with appropriate primers (Table S10) was used to amplify ∼1-kb fragments flanking the gene(s) to be deleted, using genomic DNA from the relevant isolate as template. The PCR fragments were cloned adjacently into pEX18Tc (101). Plasmid pEX18Tc containing the two cloned fragments was transferred into the chromosome of recipient strains by conjugation from E. coli ST18, followed by homologous recombination. Mutants in which the engineered deletion alleles had replaced the wild-type DNA were identified by sucrose selection, followed by PCR amplification and sequencing using primers diagnostic for the presence of the deletion.

Minimum inhibitory concentration (MIC) testing

Overnight cultures of P. aeruginosa were diluted to 1.5 × 10⁶ CFU/mL and two 5-µL aliquots were spotted onto LA plates supplemented with doubling concentrations of antibiotic (102, 103), and, if required, arabinose to induce expression of cloned genes. Tazobactam was used at a set concentration of 4 µg/mL. Agar plates were incubated for 24 h, and the antibiotic MIC was defined as the lowest concentration that inhibited growth. CLSI breakpoints were utilized to classify resistant and susceptible phenotypes (68). Meropenem and imipenem have resistance breakpoints of 8 mg/L, aztreonam, ceftazidime, and cefepime of 32 mg/L, piperacillin of 64 mg/L, and ticarcillin of 128 mg/L. CLSI does not have thresholds for carbenicillin, which is not used clinically.

Bioinformatic analysis

Genome sequences were determined by the International Pseudomonas Consortium Database (IPCD) (56, 57), by Khaledi et al. (58), and for isolates in our laboratory collection (96). Genome sequences were from isolates from five continents and included isolates from a range of human infections and from environmental and animal sources (Tables S1 to S3). ResFinder V 4.0 was used to identify acquired antibacterial resistance genes from draft genome sequences (104). The identified acquired resistance genes were verified using the CARD BLAST (https://card.mcmaster.ca/) (105). For genomes with complete fully assembled sequences, mobile genetic elements, the genes present on them and the boundaries of transposons were identified with TnCentral BLAST (106) and NCBI BLAST (107). For phylogenetic analysis, ParSNP version 1.2 from the Harvest suite 1.1.2 was used to construct a core genome alignment of isolates with P. aeruginosa PAO1 as the reference strain (108, 109). The phylogenetic trees were visualized on TreeGraph 2 (110). The MLST type was determined using FastMLST (https://github.com/EnzoAndree/FastMLST) (111).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/spectrum.00694-24.

Figures S1 and S2. spectrum.00694-24-s0001.pdf.

Figure S1: Extended phylogenetic tree. Figure S2: Maps of acquired genetic elements.

Tables S1 to S10. spectrum.00694-24-s0002.xlsx.

Supplemental tables detailing genomes analyzed, MIC replicate data, and methods-related materials.

OPEN PEER REVIEW. reviewer-comments.pdf.

An accounting of the reviewer comments and feedback.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.