Sideromimic Modification of Lactivicin Dramatically Increases Potency against Extensively Drug-Resistant Stenotrophomonas maltophilia Clinical Isolates

Abstract

Acetamido derivatives of the naturally antibacterial non-β-lactam lactivicin (LTV) have improved activity against their penicillin binding protein targets and reduced hydrolysis by β-lactamases, but penetration into Gram-negative bacteria is still relatively poor. Here we report that modification of the LTV lactone with a catechol-type siderophore increases potency 1,000-fold against Stenotrophomonas maltophilia, a species renowned for its insusceptibility to antimicrobials. The MIC₉₀ of modified lactone compound 17 (LTV17) against a global collection of extensively drug-resistant clinical S. maltophilia isolates was 0.063 μg · ml⁻¹. Sideromimic modification does not reduce the ability of LTVs to induce production of the L1 and L2 β-lactamases in S. maltophilia and does not reduce the rate at which LTVs are hydrolyzed by L1 or L2. We conclude, therefore, that lactivicin modification with a siderophore known to be preferentially used by S. maltophilia substantially increases penetration via siderophore uptake. LTV17 has the potential to be developed as a novel antimicrobial for treatment of infections by S. maltophilia. More generally, our work shows that sideromimic modification in a species-targeted manner might prove useful for the development of narrow-spectrum antimicrobials that have reduced collateral effects.

INTRODUCTION

Lactivicin (LTV) is highly unusual in that it is the only non-β-lactam natural product known to target penicillin binding proteins (PBPs). Unlike the β-lactams, which remain the most important antimicrobial class, lactivicin contains cycloserine and γ-lactam motifs, but like the β-lactams, lactivicin reacts covalently with PBPs to form a stable acyl-enzyme complex (1) (Fig. 1A and B). However, lactivicin has poor penetration into Gram-negative bacteria and is susceptible to at least some β-lactamase enzymes (2–4). A deeper understanding of the interactions between lactivicin and its derivatives and their various enzyme targets has led to the rational design of synthetic derivatives with higher potency against bacteria and reduced susceptibility to β-lactamases (1, 5, 6), including LTV13, which has the “ATMO”-type side chain (6) (Fig. 1C). Recently, it has been shown that sideromimic modification of the LTV γ-lactone (4) results in more-favorable 50% inhibitory concentrations (IC₅₀s) against Pseudomonas aeruginosa PBPs and improved penetration into P. aeruginosa strain PAO1 via interaction with the siderophore receptors and uptake systems of this strain. One of these lactivicin derivatives is phthalimide-conjugated compound 17 (here referred to as LTV17) (Fig. 1C) (6).

FIG 1.

Structures and mechanism of action of β- and γ-lactams. (A) Structures of β-lactam antibiotics and the cycloserine derivative lactivicin. (B) Outline of mechanism of action of LTV derivatives against penicillin binding proteins. (C) Structures of LTV13 and LTV17 used in the current study.

Here we describe the activity of lactivicin derivatives against Stenotrophomonas maltophilia, which is an important nosocomial pathogen, primarily causing bloodstream and respiratory tract infections in severely debilitated patients. S. maltophilia also causes sporadic urinary tract and ocular infections and is a colonizer of the lungs of a significant proportion of adult patients with cystic fibrosis. Clinical isolates of S. maltophilia are notoriously resistant to antimicrobial drugs, with resistance to most β-lactams, quinolones, and aminoglycosides being an intrinsic property of the species (7, 8). Intrinsic resistance mechanisms in S. maltophilia include the expression of antibiotic-modifying enzymes, e.g., two β-lactamases, L1 and L2, which together can hydrolyze all known β-lactams (9–11), and of multidrug efflux pumps, most notably SmeDEF (12, 13), SmeVWX (14, 15), and SmeYZ and SmeIJK (16).

MATERIALS AND METHODS

Bacterial isolates and materials.

S. maltophilia clinical isolates used in this study either originated from the SENTRY antimicrobial resistance survey, as previously described (17), or were isolated from patients being treated at the Bristol Oncology Centre (18). Isolates of other species were obtained from type strain collections or were clinical isolates collected by SENTRY and provided as gifts by Mark Toleman, Cardiff University, or have previously been described (19). Growth media were from Oxoid. Chemicals were from Sigma, unless otherwise stated. LTV13 was synthesized according to the literature protocol (6). LTV17 was kindly supplied by Pfizer.

β-Lactamase induction and measurement of β-lactamase activity in cell extracts.

MICs were determined using CLSI broth microtiter assays (20) and interpreted using published breakpoints (21). For β-lactamase induction assays, an overnight culture of bacteria was diluted to an optical density at 600 nm (OD₆₀₀) of 0.1 in nutrient broth and grown at 37°C until the OD₆₀₀ was 0.4. Inducer (100 μg · ml⁻¹ cefoxitin, 10 μg · ml⁻¹ imipenem, 50 μg · ml⁻¹ LTV13, and 0.35 μg · ml⁻¹ LTV17) was then added and incubation continued for 2 h before cell extracts were prepared and levels of β-lactamase activity in cell extracts measured as described previously (22) using 100 μM meropenem as a substrate. Protein concentrations were determined using Bio-Rad protein assay dye reagent concentrate, and an extinction coefficient at 299 nm of 9,600 absorbance units (AU)/M/cm for meropenem was used to calculate the specific meropenem hydrolyzing activity in each cell extract.

Expression and purification of L1 and L2.

Recombinant L1 protein was produced in Escherichia coli and purified as previously described (23). For L2 protein production, the putative signal sequence (residues 1 to 27) was “removed” by amplifying positions 82 to 912 from the L2 gene of S. maltophilia K279a genomic DNA with forward and reverse primers 5′-AAGTTCTGTTTCAGGGCCCGGCGGGCAAGGCCAC-3′ and 5′-ATGGTCTAGAAAGCTTTATCCGATCAACCGGTCGGC-3′. Primer sequences included extensions which allowed recombination into pOPINF vector (24), resulting in a construct encoding L2 with an N-terminal hexa-His tag, with the tag being cleavable with rhinovirus 3C protease. The resultant plasmid was designated pOPINF-L2Δ27. For protein overproduction, E. coli SoluBL21 (DE3) cells (Genlantis) bearing pOPINF-L2Δ27 were grown in 2× tryptone-yeast extract (TY) medium containing ampicillin (50 μg · ml⁻¹) to an OD₆₀₀ of 0.9 at 37°C. Protein production was induced with 1 mM isopropyl-β-D-1-thiogalactopyranoside (IPTG) at 18°C for 16 h. All subsequent purification steps were performed at 4°C. Cells were harvested by centrifugation (6,500 × g, 10 min) and resuspended in 50 mM Tris (pH 7.5)–150 mM NaCl–1 mM Tris(2-carboxyethyl)phosphine (TCEP) supplemented with EDTA-free protease inhibitor (Roche). Cells were lysed by two 30,000-lb/in² passages through a cell disruptor. After centrifugation at 100,000 × g for 45 min, the supernatant was incubated for 1.5 h with nickel-nitrilotriacetic acid (Ni-NTA) resin (Qiagen). The resin was washed with buffer A (50 mM Tris [pH 7.5], 400 mM NaCl, 1 mM TCEP) plus 10 mM imidazole, then with buffer A plus 0.1% (vol/vol) Triton X-100, and then with buffer A plus 20 mM imidazole. Protein was eluted with 50 mM Tris (pH 7.5)–200 mM NaCl–400 mM imidazole–1 mM TCEP. Imidazole and NaCl concentrations were reduced to 5 mM and 150 mM, respectively, in an Amicon 10-molecular-weight-cutoff (mwco) concentrator, and the His tag was then removed by overnight incubation with His-tagged 3C protease. The digestion mixture was then incubated with Ni-NTA resin for 30 min, and the flowthrough containing purified L2 was collected and concentrated to 30 mg · ml⁻¹ using a 10-mwco Amicon concentrator.

Assay of lactivicin hydrolysis by NMR.

¹H nuclear magnetic resonance (NMR) spectra were acquired at 298 K on a Bruker AVIII 700 spectrometer with a ¹H/¹³C/¹⁵N TCI CryoProbe. The data were recorded employing a pulse sequence with water suppression (excitation sculpting with gradients using perfect echo) (25, 26). Spectra were collected with a sweep width of 11,161 Hz, a relaxation delay of 2 s, 65,536 data points, and 16 scans. For data processing, line broadening of 0.3 Hz was applied. The NMR samples were prepared in 50 mM Tris-d₁₁ buffer (pH 7.5), supplemented with 10% D₂O. LTV13 was prepared as a 100 mM stock mixed in H₂O and LTV17 as a 25 mM stock mixed in 50% H₂O–50% dimethyl sulfoxide (DMSO) and diluted to a final concentration of 0.4 mM. The final concentrations of L1 and L2 were 150 nM.

RESULTS AND DISCUSSION

MICs of lactivicin derivatives against S. maltophilia clinical isolates.

MICs for LTV13 and LTV17 were determined against a selection of clinical isolates representing key Gram-negative species whose multidrug resistance is frequently a problem. These data confirmed the antibacterial potential of LTV17; its MIC against multidrug-resistant S. maltophilia bloodstream isolate K279a (18) is remarkably low, and it is more than 1,000-fold more potent than LTV13, an improvement in potency greater than that against any other test species (Table 1).

TABLE 1.

MICs of lactivicin derivatives against clinical Gram-negative isolates

Isolate	MIC (μg · ml−1)
	LTV13	LTV17
Klebsiella pneumoniae NCTC5055	2	0.5
Citrobacter freundii D571	512	16
Enterobacter cloacae 30950	512	16
Enterobacter aerogenes 15-8358A	64	4
Serratia marcescens NCTC1041	4	4
Pseudomonas aeruginosa NCTC10662	256	1
Stenotrophomonas maltophilia K279a	64	≤0.0625
Acinetobacter baumannii 39-4034C	256	8

According to published CLSI susceptibility testing performance standards, there are only six antimicrobials that have potential for treatment of S. maltophilia and for which resistance breakpoints have been defined (21). The current drug of choice is trimethoprim-sulfamethoxazole (SXT), and there are five alternatives: ceftazidime, ticarcillin-clavulanate, minocycline, levofloxacin, and chloramphenicol. We used the CLSI performance standards to define resistance phenotypes for a collection of 50 clinical S. maltophilia isolates from around the world (11, 13, 17). Initially, the isolates were divided into two groups: 23/50 that are STX resistant and 27/50 that are STX sensitive. The two groups were then subdivided based on how many alternative antimicrobials they remain sensitive to. Finally, the MICs of LTV13 and LTV17 were determined against the 50 isolates using standard CLSI broth microdilution methodology (Table 2). Clearly, LTV17 is very potent against all of these S. maltophilia clinical isolates, including extensively drug-resistant strains. The highest MIC seen was 0.25 μg · ml⁻¹, and the MIC₉₀ for the 50 isolates was 0.063 μg · ml⁻¹.

TABLE 2.

MICs of lactivicin derivatives against S. maltophilia clinical isolates having different profiles of resistance to last-resort antibiotics

Resistance phenotype	No. of isolates with indicated MIC (μg · ml−1)
	LTV13					LTV17
	16	32	64	128	256	≤0.031	0.063	0.125	0.25
SXTr; sensitive to 0 to 2 alternatives	0	0	3	5	1	5	3	0	1
SXTr; sensitive to 3 to 5 alternatives	1	0	7	6	0	13	1	0	0
SXTs; sensitive to 1 to 3 alternatives	0	0	4	9	1	8	3	3	0
SXTs; sensitive to 4 or 5 alternatives	0	5	3	5	0	12	1	0	0

Induction of β-lactamase production in S. maltophilia by lactivicin derivatives.

One of the reasons for the initial failure of lactivicin and its early derivatives as antibiotics was susceptibility to β-lactamases, coupled with an ability to induce β-lactamase production in bacteria where inducible enzymes exist (2–4). S. maltophilia has two inducible β-lactamases, L1 and L2, which are coordinately controlled by an AmpR-type transcriptional regulator (27). Genetic disruption of one of the main targets of lactivicin, PBP1A, has been shown to constitutively activate β-lactamase production in S. maltophilia (22, 28). As expected, we found that treatment of two well-characterized S. maltophilia clinical isolates, K279a and N531 (18), with LTV13 or LTV17 induced β-lactamase production to a similar extent as treatment with the β-lactam antibiotics cefoxitin and imipenem when the antibiotics were added to growing cells at concentrations proportionate to the compounds' relative MICs (Fig. 2).

FIG 2.

Effect of β-lactams and lactivicin derivatives on the production of β-lactamase by S. maltophilia clinical isolates N531 (upper panel) and K279a (lower panel). Bacteria were incubated in the presence of inducer (100 μg · ml⁻¹ cefoxitin, 10 μg · ml⁻¹ imipenem, 50 μg · ml⁻¹ LTV13, and 0.35 μg · ml⁻¹ LTV17), cell extracts were prepared, and β-lactamase activity in cell extracts was measured as described previously (22) using 100 μM meropenem as the substrate. Protein concentrations were determined using Bio-Rad protein assay dye reagent concentrate, and an extinction coefficient at 299 nm of 9,600 AU/M/cm for meropenem was used to calculate the specific meropenem hydrolyzing activity in each cell extract. Data presented are means ± standard errors of the means; n = 3.

Breakdown of lactivicin derivatives by S. maltophilia β-lactamases.

Since LTV13 and LTV17 both strongly induce L1 and L2 β-lactamase production, one explanation for the increased potency of LTV17 versus LTV13 is that LTV17 is a poorer substrate than LTV13 for the S. maltophilia β-lactamases. To test this hypothesis, NMR spectroscopy was used to evaluate the time-dependent hydrolysis of LTV13 and LTV17 by purified recombinant L1 and L2 β-lactamases. L2 was able to totally hydrolyze 400 μM ampicillin in less than 5 min (data not shown), while ∼95% of the LTV13 and LTV17 remained intact. Longer incubation times confirmed that LTV13 and LTV17 are not substrates for L2. They were both found to be substrates for L1 β-lactamase, however, and underwent enzyme-catalyzed hydrolysis (Fig. 3). LTV17 was broken down faster than LTV13, so reduced susceptibility to the L1 β-lactamase does not explain the increased potency of LTV17 against S. maltophilia.

FIG 3.

L1 catalyzed turnover of lactivicin derivatives as observed by ¹H NMR spectrum analysis. (A) Overlay of ¹H NMR spectra of LTV13 before addition of L1 and after 1.5 h of incubation with L1. (B) Overlay of ¹H NMR spectra of LTV17 before addition of L1 and after 1.5 h of incubation with L1. (C) Time course of L1-mediated hydrolysis of LTV13 and LTV17.

Protection of S. maltophilia from lactivicin derivatives by the activity of β-lactamases.

While they are clearly substrates, the rate of LTV17 and LTV13 hydrolysis by L1 was very low compared with that of meropenem (L1 was able to totally hydrolyze 400 μM meropenem in less than 5 min using similar assay conditions when ∼95% and 90% of LTV13 and 17 were still intact—data not shown), so we hypothesized that LTV13 and LTV17 actually kill S. maltophilia, even though they induce β-lactamase production, simply because cellular β-lactamase hydrolysis is too slow to protect the cells. To test this hypothesis, we incubated LTV17 with or without purified L1 and spotted the two mixtures onto a lawn of S. maltophilia K279a. In parallel, we used meropenem as a control (Fig. 4). Incubation with L1 does not significantly reduce the ability of LTV17 (or of LTV13—data not shown) to kill S. maltophilia K279a, whereas meropenem is rendered totally ineffective by L1. Preincubation of LTV17 with L1 for 1 h prior to spotting onto K279a did reduce the zone of clearing, suggestive of modest destruction of LTV17 as shown in the NMR experiments. Importantly, the inhibition zone diameter for LTV17 (and LTV13—data not shown) is the same for K279a as it is for the ampR frameshift mutant (FS) derivative K279a::ampR^FS, which cannot induce β-lactamase production (27), even though the latter is far more sensitive to meropenem (Fig. 4). These results imply that even though L1 can break down LTV17 at a relatively modest rate (Fig. 3), its induction by LTV17 is not sufficient to protect the cell. This also explains why S. maltophilia clinical isolates in our worldwide collection that are known to express β-lactamase constitutively at high levels (11, 22) are no less susceptible to the lactivicin derivatives than are isolates with normally inducible β-lactamases (Table 2). Indeed, to confirm this, we tested four L1/L2-hyperproducing mutants previously derived from S. maltophilia K279a (22, 27, 29) and found that the MICs of both lactivicin derivatives against K279a and these mutants are the same (32 and ≤0.031 μg · ml⁻¹, respectively, for LTV13 and LTV17).

FIG 4.

Inhibition zone produced by LTV17 and meropenem with and without L1 β-lactamase. A lawn of S. maltophilia K279a (left) or the ampR frameshift mutant (ampR^FS) (right) was spread as if for CLSI disc susceptibility testing (using a 10⁸ CFU/ml suspension of bacteria). In all cases, 2 μl of reaction mixture (50 mM Tris [pH 7.5]) with meropenem (MER, 12 mM) or LTV17 (0.4 mM) mixed (+L1) or not mixed (−L1) with purified L1 β-lactamase (1 μM final) was immediately spotted onto the lawn (t = 0 h) (top panels) or was spotted following 1 h of preincubation at 37°C (t = 1 h) (bottom panels). The plate was then incubated for 18 h at 37°C. This figure is representative of the results determined for three biological replicates.

Conclusions.

The reason for the dramatically increased potency of LTV17 versus LTV13 against S. maltophilia is not its relatively limited ability to induce L1/L2 β-lactamase production or its relatively slow hydrolysis by either of these β-lactamases. Both LTV17 and LTV13 are hydrolyzed by L1 β-lactamase slowly and not at a level that is detectable by L2, so β-lactamase production by S. maltophilia is not actually protective against either lactivicin derivative. Accordingly, while we have not excluded the possibility that there is some increased affinity for its PBP target(s), the 1,000-fold increased potency of LTV17 over that of LTV13 is most likely due to an increased rate of entry into S. maltophilia. The major difference between LTV13 and LTV17 is the presence of a catechol-type siderophore on the lactone ring of LTV17 (6). Notably, all S. maltophilia clinical isolates previously tested, including the K279a isolate used here, exclusively produced catechol-type siderophores (30). Thus, it is reasonable to infer that they preferentially take up this type of siderophore and the antibiotics conjugated to them. Accordingly, it would appear that the siderophore used for LTV17 particularly favors uptake by S. maltophilia, explaining its remarkable potency against this otherwise extensively drug-resistant bacterium. This observation is important because it implies that side-chain modification of the core fused bicyclic non β-lactam ring system of the lactivicins has the potential to improve activity in the same way as it has done for the β-lactams, e.g., for BAL30072, which is in the early phase of clinical development. This mechanism can operate by increasing potency versus PBPs and/or reducing β-lactamase susceptibility and, in addition, by improving uptake (31). Moreover, the results presented in Table 1 suggest that siderophore-mediated uptake is not a general effect, equally seen in all species. It may be that the apparent species specificity of the effect seen is dependent on the conjugation of a particular siderophore preferentially used by a particular species. In an era of improved diagnostics for infection (32), the routine use of narrow-spectrum antimicrobials is becoming a realistic proposition, and the benefit would be reduced collateral damage to the host microbiome and cross-selection for the acquisition of resistant isolates of other species of bacteria.