Resurrecting Old β-Lactams: Potent Inhibitory Activity of Temocillin against Multidrug-Resistant Burkholderia Species Isolates from the United States

Burkholderia spp. are opportunistic human pathogens that infect persons with cystic fibrosis and the immunocompromised.

ABSTRACT

Burkholderia spp. are opportunistic human pathogens that infect persons with cystic fibrosis and the immunocompromised. Burkholderia spp. express class A and C β-lactamases, which are transcriptionally regulated by PenR_A through linkage to cell wall metabolism and β-lactam exposure. The potency of temocillin, a 6-methoxy-β-lactam, was tested against a panel of multidrug-resistant (MDR) Burkholderia spp. In addition, the mechanistic basis of temocillin activity was assessed and compared to that of ticarcillin. Susceptibility testing with temocillin and ticarcillin was conducted, as was biochemical analysis of the PenA1 class A β-lactamase and AmpC1 class C β-lactamase. Molecular dynamics simulations (MDS) were performed using PenA1 with temocillin and ticarcillin. The majority (86.7%) of 150 MDR Burkholderia strains were susceptible to temocillin, while only 4% of the strains were susceptible to ticarcillin. Neither temocillin nor ticarcillin induced bla expression. Ticarcillin was hydrolyzed by PenA1 (k_cat/K_m = 1.7 ± 0.2 μM⁻¹ s⁻¹), while temocillin was slow to form a favorable complex (apparent K_i [K_i app] = ∼2 mM). Ticarcillin and temocillin were both potent inhibitors of AmpC1, with K_i app values of 4.9 ± 1.0 μM and 4.3 ± 0.4 μM, respectively. MDS of PenA revealed that ticarcillin is in an advantageous position for acylation and deacylation. Conversely, with temocillin, active-site residues K73 and S130 are rotated and the catalytic water molecule is displaced, thereby slowing acylation and allowing the 6-methoxy of temocillin to block deacylation. Temocillin is a β-lactam with potent activity against Burkholderia spp., as it does not induce bla expression and is poorly hydrolyzed by endogenous β-lactamases.

INTRODUCTION

In the 1980s, temocillin, a 6-methoxy-carboxypenicillin, was developed and approved for use in Europe. Temocillin demonstrates significant stability against many β-lactamases, including class A extended-spectrum β-lactamases (ESBLs) as well as AmpC (1–4). However, due to a lack of activity of temocillin against Gram-positive pathogens, anaerobes, and Pseudomonas aeruginosa, temocillin never reached its potential in the clinicians’ toolbox (1, 5, 6). In 2009, Livermore and Tulkens published an elegant review on the capabilities of temocillin and argued why this old β-lactam should be revisited (6). Indeed, likely due to our dwindling antibiotic arsenal, interest in temocillin as an alternative monotherapy or as part of combination therapy has increased over the last decade. Several studies have demonstrated the potent in vitro activity of temocillin against multidrug-resistant (MDR) Enterobacteriaceae producing ESBLs, AmpCs, and even KPCs (7–18). In addition, temocillin demonstrated efficacy in in vivo infection models as well as in patients (19–21).

Temocillin was previously shown to possess antimicrobial activity against members of the Burkholderia cepacia complex (Bcc), a group of Gram-negative pathogens that cause infections in susceptible persons, such as those with cystic fibrosis (CF) (22–26). β-Lactam antibiotics (e.g., ceftazidime or meropenem) are often recommended as treatment options for infections due to Bcc (27). A study in France found that 82% of Bcc isolates (n = 66) obtained from individuals with CF were susceptible to temocillin, whereas only 42% were susceptible to ceftazidime; temocillin was the most effective agent in this study (28). Another study evaluated 37 isolates belonging to 17 different Bcc species and found that 76% were susceptible to temocillin (29). A more recent study from Belgium using 91 Bcc isolates revealed that 67% were susceptible to temocillin (30).

Importantly, the in vivo efficacy of temocillin for individuals with Bcc infections was also previously reported. In a retrospective study conducted in the United Kingdom, 18 out of 32 courses of intravenous temocillin resulted in improvement for 20 CF patients with a Bcc infection (31). In addition, temocillin was used as salvage therapy to successfully treat a Bcc infection in an individual with cervical osteomyelitis (32).

Resistance to β-lactams is common among Bcc strains due to the expression of β-lactamases (33, 34). Most Bcc strains express two β-lactamases, a Pen-like class A β-lactamase and a class C β-lactamase (AmpC); some Bcc strains also produce a class D oxacillinase (35). We previously characterized the PenA1 and AmpC1 β-lactamases present in Burkholderia multivorans, a prevalent species within the Bcc. PenA1 is a carbapenemase with a substrate profile similar to that of KPC-2 (34). Conversely, the AmpC1 of B. multivorans possesses limited β-lactamase activity; thus, data to date suggest that this AmpC1 may play a role in peptidoglycan metabolism (36). Expression of these bla genes is regulated by PenR_A through a system analogous to the AmpC/AmpR regulatory pathways present in Enterobacteriaceae (33, 37).

The goal of this study was to assess the activity of temocillin against MDR Burkholderia spp. (i.e., Bcc and Burkholderia gladioli) strains from the United States and to determine the mechanistic basis for the potency of temocillin. Using a panel of 150 Burkholderia species isolates, we examined the antimicrobial properties of temocillin compared to those of ticarcillin, which is identical in structure to temocillin but lacks the C-6 methoxy. We further determined temocillin’s and ticarcillin’s effect on bla expression in B. multivorans and characterized the biochemical activity of both agents against purified PenA1 and AmpC1 β-lactamases. Using molecular modeling, we further explored the mechanistic impact of the 6-methoxy side chain present on temocillin by directly comparing it to ticarcillin docked into the PenA1 active site.

RESULTS AND DISCUSSION

Temocillin demonstrates potent activity against Bcc.

Susceptibility testing with temocillin was conducted using a previously characterized panel of 150 MDR clinical strains, including 140 Bcc isolates (14 different species) and 10 B. gladioli isolates obtained from the sputum of persons with CF (38). As the breakpoints for temocillin against Burkholderia spp. are not well defined, previously recommended interpretative indices (≥32 μg/ml for resistant and ≤8 or 16 μg/ml for susceptible) were used (39–41). The site of the infection is a critical determinant for efficacy. For example, the lower breakpoint seems more likely to be achieved in systemic infections than in urinary tract infections, where larger amounts of drug are observed (11).

At a breakpoint of 16 μg/ml, out of 150 isolates, 87% were susceptible to temocillin and only 13% were resistant (Table 1). Twenty isolates were not susceptible to temocillin, including B. cenocepacia (3 of 33), B. cepacia (2 of 15), B. contaminans (1 of 7), B. dolosa (1 of 4), B. gladioli (2 of 10), and B. multivorans (11 of 50). For comparison, 63%, 64%, and 93% of these 150 isolates previously tested susceptible to trimethoprim-sulfamethoxazole, ceftazidime, and ceftazidime-avibactam, respectively (38). Using a breakpoint of 8 μg/ml for temocillin, 66% of the isolates were susceptible. Thus, temocillin compared similarly to the recommended first-line agents, trimethoprim-sulfamethoxazole and ceftazidime.

TABLE 1.

Distribution of MICs for temocillin and ticarcillin against Burkholderia spp. clinical isolates^a

Antibiotic and strain (total no. of isolates)	No. of isolates with the following MIC (μg/ml):									No. of isolates:
	1	2	4	8	16	32	64	128	>128	R	I	S
Temocillin
B. ambifaria (4)	1	1	1	1	0	0	0	0	0	0		4
B. arboris (2)	0	0	2	0	0	0	0	0	0	0		2
B. cenocepacia (33)	1	3	14	7	5	1	1	1	0	3		30
B. cepacia (15)	0	4	9	0	0	2	0	0	0	2		13
B. contaminans (7)	0	3	1	1	1	1	0	0	0	1		6
B. diffusa (1)	0	0	1	0	0	0	0	0	0	0		1
B. dolosa (4)	0	0	0	2	1	0	1	0	0	1		3
B. gladioli (10)	0	0	3	2	3	0	2	0	0	2		8
B. multivorans (50)	0	0	0	24	15	7	3	0	1	11		39
B. pseudomultivorans (1)	0	0	0	1	0	0	0	0	0	0		1
B. pyrrocinia (2)	1	0	1	0	0	0	0	0	0	0		2
B. seminalis (1)	0	0	1	0	0	0	0	0	0	0		1
B. stabilis (2)	0	0	0	1	1	0	0	0	0	0		2
B. ubonensis (1)	0	0	1	0	0	0	0	0	0	0		1
B. vietnamiensis (9)	2	2	3	1	1	0	0	0	0	0		9
Bcc indeterminate (8)	1	0	5	1	1	0	0	0	0	0		8
Total (of 150) with MIC of 16 μg/ml	6	13	42	41	28	11	7	1	1	20		130
Total (of 150) with MIC of 8 μg/ml	6	13	42	41	28	11	7	1	1	48		102
Ticarcillin
B. ambifaria (1)	0	0	0	0	0	0	0	0	1	1	0	0
B. arboris (1)	0	0	0	0	0	0	0	0	1	1	0	0
B. cenocepacia (3)	0	0	0	0	0	0	0	0	3	3	0	0
B. cepacia (2)	0	0	0	0	0	0	0	0	2	2	0	0
B. contaminans (2)	0	0	0	0	0	0	0	0	2	2	0	0
B. diffusa (1)	0	0	0	0	0	0	0	0	1	1	0	0
B. dolosa (2)	0	0	0	0	0	0	0	0	2	2	0	0
B. gladioli (2)	0	0	0	0	0	1	1	0	0	0	2	0
B. multivorans (3)	0	0	0	0	0	0	0	0	3	3	0	0
B. pseudomultivorans (1)	0	0	0	0	0	0	0	0	1	1	0	0
B. pyrrocinia (1)	0	0	0	0	0	0	0	0	1	1	0	0
B. seminalis (1)	0	0	0	0	0	0	0	0	1	1	0	0
B. stabilis (2)	0	0	0	0	0	0	0	0	2	2	0	0
B. ubonensis (1)	0	0	0	0	0	0	0	0	1	1	0	0
B. vietnamiensis (2)	0	1	0	0	0	0	0	0	1	1	0	1
Total (25)	0	1	0	0	0	1	1	0	22	22	2	1

^aFor temocillin, resistance versus susceptibility is presented for breakpoints of 8 μg/ml and 16 μg/ml for all of the Bcc tested; however, for each species, the 16-μg/ml breakpoint was used to assign resistance versus susceptibility. CLSI interpretative indices are not available for Burkholderia spp. for temocillin or ticarcillin. Thus, for temocillin, previously published breakpoints were used. For temocillin, resistant was an MIC of ≥32 μg/ml and susceptible was an MIC of either ≤ 16 μg/ml (39) or ≤ 8 μg/ml (40, 41). For ticarcillin, the ticarcillin-clavulanic acid CLSI breakpoints for Burkholderia spp. were used. For ticarcillin, resistant was an MIC of ≥128 μg/ml, intermediate was an MIC of 32 to 64 μg/ml, and susceptible was an MIC of ≤16 μg/ml. R, resistant; I, intermediate; S, susceptible.

A select panel of 25 Burkholderia species clinical isolates representing different species and strains with susceptible and resistant MICs toward temocillin were tested against ticarcillin (see Table S1 in the supplemental material). Only 4% were susceptible to ticarcillin (Table 1).

Based on the aggregate temocillin susceptibility profiles, 13% of 150 isolates were resistant to temocillin. The lingering factors contributing to resistance to temocillin in these isolates may be efflux, changes to the outer membrane of the bacterium that affect the penetration of temocillin, changes to the target penicillin binding protein (PBP) that results in diminished inhibition, and/or the presence of a class D oxacillinase. In P. aeruginosa, resistance to temocillin is largely associated with mutations in the genes that encode the efflux pump MexAB-OprM (42, 43). To assess the contribution of efflux toward temocillin resistance in the 20 resistant Burkholderia spp., temocillin was combined with two different efflux pump inhibitors, Phe-Arg β-naphthylamide dihydrochloride (PA-β-N) or carbonyl cyanide m-chlorophenylhydrazone (CCCP). PA-β-N or CCCP can demonstrate efflux pump inhibition; however, CCCP is more potent and consistent as an efflux pump inhibitor than PA-β-N against Burkholderia spp. (44–46). Neither PA-β-N nor CCCP significantly improved the MICs of resistant isolates, with the exception of B. cenocepacia AU14381, which demonstrated a MIC of ≤0.06 μg/ml for the temocillin-CCCP combination (Table 2).

TABLE 2.

MICs for temocillin paired with various partners tested against temocillin-resistant Burkholderia species clinical isolates^a

Strain	MIC (μg/ml)
	Temocillin	Temocillin-avibactam	Temocillin–PA-β-N	Temocillin-CCCP
B. cenocepacia AU10321	128	128	128	128
B. cenocepacia AU141094	64	64	64	64
B. cenocepacia AU14381	32	32	32	≤0.06
B. cepacia AU0329	32	16	16	32
B. cepacia AU13354	32	16	32	16
B. contaminans AU20979	32	64	64	64
B. dolosa AU29985	64	128	128	128
B. gladioli AU29541	64	64	64	64
B. gladioli AU16341	64	128	128	128
B. multivorans AU28442	>256	>128	>128	>128
B. multivorans AU14786	32	64	32	64
B. multivorans AU19659	32	32	32	64
B. multivorans AU23919	32	32	32	32
B. multivorans AU10047	32	32	32	32
B. multivorans AU17534	32	128	16	16
B. multivorans AU11772	64	64	64	64
B. multivorans AU10086	64	32	32	32
B. multivorans AU29198	32	64	32	64
B. multivorans AU15814	32	32	32	32
B. multivorans AU25626	64	32	32	64

^aAvibactam was held constant at 4 μg/ml, Phe-Arg β-naphthylamide dihydrochloride (PA-β-N) was held constant at 50 μg/ml, and carbonyl cyanide m-chlorophenylhydrazone (CCCP) was held constant at 12.5 μM. Interpretative indices are listed in footnote a of Table 1.

To assess if susceptibility to temocillin could be restored in the resistant isolates by partnering temocillin with a β-lactamase inhibitor, avibactam, a diazabicyclooctane non-β-lactam β-lactamase inhibitor was tested. Avibactam was previously shown to inhibit PenA1, the major β-lactamase in B. multivorans (38). The addition of avibactam had little effect on the temocillin-resistant isolates (Table 2).

The whole-genome sequences (WGS) for the 50 B. multivorans isolates used in this study are available at the GenBank WGS repository (BioProject accession number PRJNA434393), and the 11 temocillin-resistant isolates do not possess OXA β-lactamases. The remaining mechanisms of temocillin resistance in these isolates (i.e., permeability and/or amino acid substitutions in PBPs) are under investigation.

Temocillin does not induce expression of AmpC1 or PenA1 in B. multivorans.

To assess the impact of temocillin and ticarcillin on bla induction, analytic isoelectric focusing was conducted using B. multivorans ATCC 17616. B. multivorans ATCC 17616 was chosen because the members of the Bcc remain largely uncharacterized, including their β-lactamases. Imipenem robustly induced the production of PenA1 and AmpC1, as evidenced by the nitrocefinase activity of the β-lactamases (Fig. 1). Conversely, temocillin and ticarcillin did not affect bla expression, as indicated by the lack of hydrolytic activity.

FIG 1.

Isoelectric focusing gel of B. multivorans ATCC 17616 grown in LB alone, LB with induction with 1 mg/liter of imipenem, LB with induction with 0.5 mg/liter of temocillin, and LB with induction with 1 mg/liter of ticarcillin. The concentrations of β-lactams are at sub-MIC levels. Purified PenA1 and AmpC1 were used as controls.

Temocillin is a potent inactivator of AmpC1 but not PenA1.

The contribution of PenA1 and AmpC1 to the antimicrobial activity of temocillin and ticarcillin was determined by expressing the β-lactamases in an isogenic Escherichia coli strain background. Temocillin MICs were all in the susceptible range, while the presence of PenA resulted in an MIC of >512 μg/ml for ticarcillin (Table 3). In these cellular assays, temocillin was effective against the single β-lactamases.

TABLE 3.

Susceptibility testing results^a

Strain	Susceptibility testing result (MIC, μg/ml)
	Temocillin	Ticarcillin
B. multivorans ATCC 17616	4	32
E. coli DH10B/pBC SK(+)	16	4
E. coli DH10B/pBC SK(+) blapenA1	16	>512
E. coli DH10B/pBC SK(+) blaampC1	16	4

^aInterpretative indices are listed in footnote a of Table 1.

The PenA1 and AmpC1 β-lactamases were purified to assess their biochemical activity against temocillin versus that against ticarcillin. Progress curves revealed that temocillin was a poor substrate for PenA1, with limited hydrolysis being observed with 1 μM PenA1 during a 1-h incubation (Fig. 2A, red line). Conversely, 10 nM PenA1 robustly hydrolyzed ticarcillin within 500 s (Fig. 2B, pink line). AmpC1 was unable to hydrolyze either substrate (Fig. 2A and B, green lines).

FIG 2.

(A) Temocillin was spectroscopically monitored for 1 h alone or with the addition of sodium hydroxide (NaOH), purified PenA1, or purified AmpC1. (B) Ticarcillin was spectroscopically monitored for 25 min alone or with the addition of sodium hydroxide (NaOH), purified PenA1 (100 nM or 10 nM), or purified AmpC1.

A more in-depth kinetic analysis revealed that PenA1 was also not inhibited by temocillin (apparent K_i [K_i app] value, 1,820 ± 200 μM) (Table 4). PenA1 demonstrated detectable levels of ticarcillin hydrolysis with a k_cat value of 55 s⁻¹. The purified AmpC1 β-lactamase was inhibited by both temocillin and ticarcillin, with K_i app values of <5 μM for both β-lactams (Table 4).

TABLE 4.

Steady-state-kinetic parameters for PenA1 and AmpC1 with temocillin and ticarcillin^a

β-Lactam	β-Lactamase	Ki app (μM)	Km (μM)	kcat (s−1)	kcat/Km (μM−1 s−1)
Temocillin	PenA1	1,820 ± 200	NM	NM	NM
	AmpC1	4.3 ± 0.4	NM	NM	NM
Ticarcillin	PenA1	NM	33 ± 3	55 ± 5	1.7 ± 0.2
	AmpC1	4.9 ± 1.0	NM	NM	NM

^aThe error values represent the standard deviation from the mean. NM, not measurable.

Time-based electrospray ionization mass spectrometry (ESI-MS) was performed to obtain a greater understanding of the interactions of PenA1 and AmpC1 with temocillin. The molecular weights of the apo-enzyme forms of PenA1 (29,419 Da) and AmpC1 (39,750 Da) were near their theoretical values (PenA1, 29,418.27 Da; AmpC1, 39,748.61) (Fig. 3). The PenA1 and AmpC1 enzymes were incubated with temocillin at a 1:1 molar ratio, and after 15 min or 24 h of incubation, the reactions were terminated and mass spectra were determined. Consistent with the kinetic observations, apo-PenA1 was mostly observed after 15 min of incubation with a 1:1 molar ratio of temocillin. By 24 h with the slow hydrolysis of temocillin, the acyl-enzyme peak (29,833 Da) disappeared (Fig. 3). Increasing the ratio of temocillin to PenA1 by 10-fold resulted in only an acyl-enzyme peak, observed at 15 min, which suggests slow acylation (Fig. 3). Moreover, at 24 h, the acyl-peak was predominant, revealing that temocillin is also slow to deacylate from PenA1. At 15 min and 24 h, AmpC1 remained bound to temocillin as an acyl-enzyme (40,165 Da); the apo-enzyme was not detected at either time point (Fig. 3). Temocillin potently inhibits AmpC1 by formation of a highly stable acyl-enzyme complex.

FIG 3.

Time-based ESI-MS was conducted on PenA1 and AmpC1 alone and with temocillin incubated for 15 min or 24 h at a 1:1 ratio (gray box) or a 1:10 ratio. A summary of the masses identified is presented in the lower right.

Ticarcillin is favored for hydrolysis by PenA1, but temocillin is not.

To obtain a mechanistic insight into the differences between temocillin and ticarcillin, both compounds were docked into the PenA1 active site as Michaelis-Menten and acyl-enzyme complexes. In the Michaelis-Menten complex with ticarcillin and PenA1, a catalytic water molecule is favorably positioned near K73, S130, and S70 for rapid acylation (Fig. 4A). Conversely, in the temocillin model, K73 and S130 are rotated and the 6-methoxy of temocillin displaces the catalytic water molecule; thus, temocillin is less likely to acylate PenA1 efficiently (Fig. 4B).

FIG 4.

(A) The Michaelis-Menten complex of ticarcillin (cyan) in the PenA1 active site reveals that the catalytic water molecule (WAT; blue sphere) is primed for acylation. (B) The Michaelis-Menten complex of temocillin (magenta) in the PenA1 active site shows that the water molecules are not positioned favorably for acylation. (C) The acyl-enzyme complex of ticarcillin (cyan) in the PenA1 active site displays the catalytic water molecule in an ideal position for deacylation. (D) The acyl-enzyme complex of temocillin (magenta) in the PenA1 active site discloses clashes between residues that will impede deacylation. Dotted green, orange, and gray lines indicate potential hydrogen bonding interactions. Dotted blue lines are used to show potential hydrogen bonding interactions with water molecules. Chemical structures of ticarcillin (left) and temocillin (right; cyan, 6-α-methoxy moiety) are presented. (E, F) The graphs display the hydrogen bonding interaction patterns (red line, a hydrogen bond is present) of active-site water molecules during the course of MDS (y axis) with the acyl-enzyme complexes with ticarcillin (E) and temocillin (F).

In the acyl-enzyme model with ticarcillin, the deacylation water is oriented for deacylation to occur, while with temocillin, the 6-methoxy group again obstructs the catalytic water molecules (Fig. 4C and D). The methoxy group is added to temocillin to block the deacylation water molecules in β-lactamases; undoubtedly, similar mechanistic observations were made with other β-lactams possessing a methoxy moiety at C-6, including cefoxitin and moxalactam (47). Perusing the placement of the water molecules over the course of the molecular dynamics simulations (MDS) in both acyl-enzyme complexes reveals that fewer interactions with waters are formed with temocillin than with ticarcillin (Fig. 4E and F). The molecular modeling supports the biochemical and antimicrobial susceptibility testing data, as ticarcillin is a good substrate, while temocillin is a poor substrate.

Conclusions.

In this study, we found that temocillin is as effective in vitro as the currently recommended agents used to treat infections caused by Burkholderia spp. However, ceftazidime-avibactam remains superior against this panel of isolates, likely due to avibactam’s robust inactivation of PenA1 (38). Temocillin is a potent β-lactam against Burkholderia spp. because it is not hydrolyzed by the endogenous β-lactamases. The lack of hydrolytic turnover by PenA1 is due to the strategically placed 6-methoxy of temocillin. Based on the data presented herein, temocillin may offer some utility for the treatment of infections caused by Burkholderia spp.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The Burkholderia species clinical isolates used in this study were obtained from the Burkholderia cepacia Research Laboratory and Repository (BcRLR) strain collection; the 50 B. multivorans isolates were previously characterized (38). These 150 isolates were recovered from cultures of respiratory specimens from 150 individuals with CF receiving care in 68 cities in 36 states in the United States. Strain genotyping was performed using a repetitive element PCR assay as previously described (48). The construction of pBC SK(+) bla_penA1 and bla_ampC1, expressed in E. coli DH10B, and pGEX-6p2 bla_penA1 and bla_ampC1, expressed in E. coli Origami 2 DE3, was previously described; the genes were cloned from B. multivorans ATCC 17616 (34, 36).

Antibiotic susceptibility.

Mueller-Hinton (M-H) agar dilution MICs, determined according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI) (49), were used to phenotypically characterize the strains as previously described (50). Temocillin was obtained from BOC Sciences and Alfa Chemistry. Ticarcillin, Phe-Arg β-naphthylamide dihydrochloride (PA-β-N), and carbonyl cyanide m-chlorophenylhydrazone (CCCP) were obtained from Sigma. Avibactam was procured from Advanced ChemBlocks. The susceptibility breakpoints utilized are outlined in footnote a of Table 1. The concentrations of avibactam, PA-β-N, and CCCP used are indicated in footnote a of Table 2.

Analytical isoelectric focusing.

B. multivorans ATCC 17616 was grown in lysogeny broth (LB) to an optical density at 600 nm (OD₆₀₀) of 0.6. Either imipenem, temocillin, or ticarcillin was added to the cultures to induce the expression of β-lactamases, and the cultures were grown for 1 h. Cells were pelleted by centrifugation at 12,000 rpm for 10 min. Crude extracts were prepared and loaded onto a FocusGel 3-10 24S gel (Serva Electrophoresis GmbH) with a pH gradient of from 3 to 10 and electrophoresed using a Multiphor II apparatus. The gels were focused at 4°C at 8 W for 120 min. The detection of β-lactamases was performed by the overlay of 2 mM nitrocefin. Purified PenA1 or AmpC1 β-lactamases were used as controls, as was B. multivorans ATCC 17616 grown in LB without β-lactams.

Purification of PenA1 and AmpC1 and steady-state kinetics.

PenA1 and AmpC1 were purified as previously described (34, 36), and theoretical molecular weight values were determined using the ProtParam tool available through the ExPASy bioinformatics resource portal (51). Steady-state kinetic parameters were determined using an Agilent 8453 diode array spectrophotometer (Santa Clara, CA), as previously described (50, 52).

(i) Progress curves. Temocillin and ticarcillin were monitored at an absorbance (λ) of 230 nm. Ten millimolar sodium hydroxide was used as a positive control for hydrolysis of temocillin and ticarcillin; also, the substrates were monitored alone to control for spontaneous breakdown. Temocillin hydrolysis with the addition of 1 μM PenA1 or AmpC1 was monitored for 1 h. Ticarcillin hydrolysis with the addition of 100 nM PenA1 or AmpC1 as well as 10 nM PenA was monitored for 25 min.

(ii) Obtaining K_m and k_cat. Origin (version 8.0) software was used to fit the initial velocity data to the Henri-Michaelis-Menten equation and obtain the kinetic parameters V_max and K_m for ticarcillin, as previously described (50). The extinction coefficient (−490 M⁻¹ cm⁻¹) for ticarcillin was determined at an absorbance (λ) of 230 nm.

(iii) Determination of K_i app. Using a direct competition assay under steady-state conditions, the K_i app of temocillin was determined with AmpC1 and PenA1. A direct competition assay was performed to estimate the apparent dissociation constant for the Michaelis-Menten complex, K_i app, of the inhibitor. The data were analyzed according to equation 1 to account for the affinity of nitrocefin (ncf) for the β-lactamase.

$$K_{i\text{\hspace{0.17em}app}}\left(\text{corrected}\right)=K_{i\text{\hspace{0.17em}app}}\left(\text{observed}\right)/(1+[S]/K_m\text{ncf})$$ (1)

where [S] is the concentration of nitrocefin.

Time-based ESI-MS.

Electrospray ionization mass spectrometry (ESI-MS) was conducted on a Waters Synapt G2-Si mass spectrometer as previously described (38). Purified AmpC1 and PenA1 were incubated with temocillin at a 1:1 enzyme/inhibitor ratio for 15 min or 24 h. In addition, purified PenA1 was also incubated with temocillin at a 1:10 ratio. Reactions were terminated by the addition of 0.1% formic acid and acetonitrile.

Molecular modeling.

The crystal coordinates of PenA1 (PDB accession number 3W4Q) were used to construct the Michaelis-Menten and acyl complexes with ticarcillin and temocillin as previously described using the BIOVIA Discovery Studio 2017 (DS2017; Accelrys, Inc., San Diego, CA) molecular modeling software (52). The crystallographic water molecules were maintained during modeling. The PenA1 apo-enzyme was solvated and minimized to a root mean square deviation of 0.05 Å using the conjugate gradient method. Temocillin and ticarcillin were constructed using the Fragment Builder tools and minimized using the Standard Dynamics Cascade protocol of DS2017. The intact and acylated β-lactams were automatically docked into the active site of PenA1 using the LibDock module of DS2017. The generated conformations were further analyzed. To check the stability and to look for possible conformational changes, molecular dynamics simulation (MDS) was conducted for 60 ps, as previously described (52).