Investigation of Dipicolinic Acid Isosteres for the Inhibition of Metallo-β-Lactamases

Abstract

New Delhi Metallo-β-lactamase-1 (NDM-1) poses an immediate threat to our most effective and widely prescribed drugs, the β-lactam containing class of antibiotics. There are no clinically relevant inhibitors to combat NDM-1, despite significant efforts towards their development. Inhibitors that utilize a carboxylic acid motif for binding the Zn(II) ions in the active site of NDM-1 comprise of a large portion of the >500 inhibitors reported to date. New and structurally diverse scaffolds for inhibitor development are of urgent need. Herein, we report the isosteric replacement of one carboxylate group of dipicolinic acid (DPA) to obtain DPA isosteres with good inhibitory activity against NDM-1 (and related metallo-β-lactamases, IMP-1 and VIM-2). It was determined that the choice of carboxylate isostere influences both the potency of NDM-1 inhibition and the mechanism of action. Additionally, we show that an isostere with a metal-stripping mechanism can be re-engineered into an inhibitor that favors ternary complex formation. This work provides a roadmap for future isosteric replacement of routinely used metal binding motifs (i.e., carboxylic acids) for the generation of new entities in NDM-1 inhibitor design and development.

Graphical abstract

Dipicolinic acid isosteres were evaluated against metallo-β-lactamases. The concept of isosteres is used to show the choice of the carboxylate isostere not only impacts inhibition value, but also the mechanism of action. This study demonstrates the utility of isosteric replacement for routinely used metal binding motifs (e.g., carboxylic acids) in metalloenzyme inhibitor development.

Introduction

β-Lactam-containing antibiotics have been one of the most successful and popular class of antibiotics for combating a wide range of Gram-positive and -negative bacterial infections. Unsurprisingly, since the introduction of β-lactam containing antibiotics (beginning with penicillin in the 1940s), the widespread use of this class of antibiotics has led to the emergence of various resistance mechanisms. Resistance mechanisms that bacteria employ include mutation of penicillin binding proteins (PBPs), modification of outer membrane proteins, the production of efflux pumps, and the expression of β-lactamases.^1-2 β-Lactamases utilize either an active site serine residue (Ambler classes A, C, and D) or Zn(II) metal ion(s) (Ambler class B, also known as metallo-β-lactamases, MBLs) to hydrolyze the β-lactam ring of the target antibiotic and render the drug ineffective.^{1, 3} A dedicated β-lactamase database provides an up-to-date compilation of the biochemical and structural data of all MBLs (http://www.bldb.eu/).⁴ First observed in 1966 by Sabath and Abraham (merely two decades after the introduction of penicillin), there are now >80 reported unique MBL families.^{3, 5} MBLs have become one of the most problematic bacterial resistance mechanisms due to their wide substrate profile, with the ability to hydrolyze virtually all clinically used bicyclic β-lactam antibiotics.^6–7

MBLs are divided into B1, B2, and B3 subclasses depending on sequence identity and the number of Zn(II) ion(s) (either one or two) in the active site. Description of subclasses and their mechanism of action are reviewed elsewhere.^6–9 Commonly observed and clinically relevant members of the MBLs belong to subclass B1,¹⁰ of which New Delhi Metallo-β-lactamase (NDM) is a prominent representative. NDM bears a dinuclear Zn(II) active site, with Zn₁ ligated by H116, H118, H196, and a bridging hydroxide, and Zn₂ ligated by D120, C221, H263, the bridging hydroxide, and an apical H₂O (standard BBL numbering).¹¹ The active site is flanked between two flexible loops, allowing the protein to accommodate a wide range of antibiotic substrates.^12–13 Plasmids that carry the bla_NDM-gene to undergo horizontal gene transfer between different species of microorganisms, leading to an increase in the prevalence of bla_NDM-bearing pathogens.^14–15 Additionally, the threat of NDM is exacerbated (compared to the two other most prevalent members of the B1 MBLs, IMiPenemase, IMP, and Verona Integron-encoded Metallo-β-lactamase, VIM) by the ability of NDM to anchor to the cellular membrane, leading to higher protein stability and secretion.^16–17 Resistance to a broad spectrum of β-lactams and the high horizontal gene transfer ability have allowed for rapid propagation from nosocomial infections to infections within the general population.^18–20 Furthermore, bla_NDM is often carried on plasmids containing other genes that encode various resistance factors (including macrolides, aminoglycosides, rifampicin, and sulfamethoxazole),¹⁴ resulting in bacterial infections that are resistant to many different classes of antimicrobials.

It has been ten years since the first report of NDM-1, with >24 NDM variants currently identified.^{4, 14, 21} Some NDM variants exhibit an increase in thermal stability, Zn(II) affinity, and mononuclear Zn(II) activity that may provide an evolutionary advantage when Zn(II) availability is low.^22–24 Unfortunately, even with the rapid spread of the bla_NDM-1 gene and the evolution of NDM variants, little advancement has been made in inhibitor development. There are currently ~525 inhibitors reported in literature (representative structures shown in Figure 1),²⁵ but many inhibitors share similar structural features and none have progressed to clinical trials. The flexible active site of NDM-1, the diversity of related MBLs, a lack of understanding for inhibition mechanisms, and misinterpretation of structural data have all contributed to delayed inhibitor development.^26–27 A survey of current NDM-1 inhibitors reveals a large portion of inhibitors bear a carboxylic acid motif.^25–26 In line with this, our lab previously utilized a fragment-based drug discovery (FBDD) approach to identify dipicolinic acid (DPA) as a lead metal-binding pharmacophore (MBP) for NDM-1 inhibitor development.²⁸ Derivatization of DPA was performed to obtain 4-(3-aminophenyl)pyridine-2,6-dicarboxylic acid (Figure 1) as a highly selective inhibitor for NDM-1 (IMP-1 and VIM-2) that inhibits by forming a stable NDM-1:Zn(II):inhibitor ternary complex. In this study, we further investigate the DPA MBP as an inhibitor for NDM-1 utilizing the concept of isosteric replacement.

Figure 1.

Representative inhibitors of NDM-1.^28–36

Isosteres bear similar structural and physical properties to the parent functional group and are used as surrogates in an effort improve the overall drug-likeness (i.e., increased permeability, better pharmacokinetics, and/or reduced toxicity) of the original molecule. A prominent example includes substituting the carboxylic acid functional group with carboxylic acid isosteres.^37–38 There are an estimated >450 marketed drugs (including nonsteroidal anti-inflammatory drugs, antibiotics, anticoagulants, and cholesterol-lowering statins) that utilize a carboxylic acid motif.^37–39 The ionizable nature of the carboxylic acid under physiological conditions (pH 7.4) makes it a useful handle for generating strong inhibitor-target interactions (i.e., electrostatic and hydrogen bonds). In metalloenzyme inhibition, the carboxylic acid motif routinely participates in metal coordination, as seen in inhibitors for neutral endopeptidase (NEP), class II fructose-1,6-bisphosphate aldolase (FBP-aldolase), angiotensin-converting enzyme (ACE), and many others.⁴⁰ Despite the efficacy of this functional group, its usefulness can be limited in drug development due to poor cell permeability, metabolic instability, and off-target effects.^{37, 39} In an attempt to overcome such liabilities, chemists turn to the utilization of pro-drugs or isosteric replacement. Recent studies have begun to explore the effect of MBP isosteric replacement in metalloenzyme inhibition and pharmacokinetic profile.^41–42

In this study, we report the development and evaluation of 10 DPA isosteres (Figure 2) as inhibitors of NDM-1 and related MBLs. A selection of carboxylic acid isostere motifs was chosen to span a range of acidity (pK_a) values and metal coordination preferences. Varying the identity of the isostere impacts both inhibition value and inhibition mechanism. Additionally, we show the re-engineering of an isostere from a metal-stripping mechanism to one that favors the formation of a ternary complex. This study provides a roadmap for the isosteric replacement of current and future metal-binding motifs for the generation of new entities in NDM-1 inhibitor design and may be adopted for the inhibitor development of MBLs and other metalloenzymes.

Figure 2.

DPA isosteres reported here.

Results and Discussion

Isostere Syntheses.

The synthesis of isosteres (1 – 10) is presented in Scheme 1, Scheme 2, and Scheme 3. Synthesis of isosteres 1, 4, 9, and 10 are summarized in Scheme 1. Isostere 1 was obtained from a palladium-catalyzed Hirao cross-coupling reaction of commercially available methyl 6-bromopicolinate and diethylphoshate, followed by acid hydrolysis. Isostere 4 was achieved by conversion of the methyl 6-bromopicolinate to a nitrile via the Rosenmund-von Braun reaction, followed by an azide-nitrile cycloaddition. Isosteres 9 and 10 were synthesized via palladium-catalyzed Stille coupling of methyl 6-bromopicolinate with the corresponding organotin reagents, followed by saponification. Syntheses of isosteres 2, 3, 5, 6, and 8 are summarized in Scheme 2. Briefly, compound 11 was obtained from commercially available methyl 6-(hydroxymethyl)picolinate. Substitution of the alkyl bromide of compound 11 with various nucleophiles, followed by hydrolysis yielded isosteres 2, 3, 5, 6, and 8. Lastly, methyl 6-aminopicolinate was treated with methanesulfonyl chloride, followed by saponification to yield isostere 7.

Scheme 1.

Synthesis of isosteres 1, 4, 9, and 10. Reagents and conditions: (a) Pd₂(dba)₃, Pd(dppf)Cl₂, diethylphosphate, triethylamine, toluene, 90°C, 20 h; then 6 M HCl, 100°C, 20 h; two steps 19%; (b) CuCN, pyridine, 116°C, 4 h, 21%; (c) NaN₃, NH₄Cl, DMF, 130°C, 20 h; then 2 M HCl, 25°C, 1 h; two steps 98%; (d) 2-(tributylstannyl)thiazole/ 2-(tributylstannyl)oxazole, Pd(PPh₃)₂Cl₂, THF, 75°C, 18 h; then 3:1 1 M NaOH:THF, 70°C, 3 h; two steps 8-74%.

Scheme 2.

Synthesis of isosteres 2, 3, 5, 6, and 8. Reagents and conditions: (a) PBr₃, CHCl₃, 0°C, 3 h, 70%; (b) P(OEt)₃, toluene, 140°C, 2 h, 80%; (c) 6 M HCl, 100°C, 27 h, 98%; (d) KCN, THF, 50°C, 19 h, 49%; (e) 12 M HCl, 100°C, 12 h, then 1 M NaOH, 60°C, 6 h, 30%; (f) Na₂SO₃, H₂O, then 4 M HCl, 100°C, 16 h; two steps 41%; (g) NaSO₂CH₃, DMF, 120°C, 2 h, then 3:1 1M NaOH:THF, 70°C, 3 h; two steps 43%; (h) CH₃NHSO₂CH₃, K₂CO₃, ACN, 75°C, 25 h; then 1 M NaOH, 70°C, 3 h; two steps 55%.

Scheme 3.

Synthesis of isostere 7. Reagents and conditions: (a) Methanesulfonyl chloride, triethylamine, CH₂Cl₂, 0°C, 16 h; then 1 M NaOH: THF, r.t., 16 h; two steps 12.4%.

Inhibition Assay.

A preliminary screen of isosteres 1 – 10 was performed against NDM-1 utilizing meropenem as substrate (at a single concentration of 180 μM, Table 1). Meropenem was observed to have a K_M,NDM-1 = 80±7 μM and k_cat = 15.3±0.4 s⁻¹. Compounds 1 – 4 exhibited inhibition against NDM-1 with IC₅₀ values ranging from 0.13 – 7.7 μM. Compounds 1 and 2 (both bearing a phosphonic acid) exhibited lower IC₅₀ values (130±10 and 310±10 nM, respectively) compared to that of DPA (840±40 nM), while 3 and 4 revealed higher IC₅₀ values (7.7±0.6 and 7.0±0.5 μM, respectively). Interestingly, compound 5, which possesses a similar acidic motif as 1 and 2 (a sulfonic acid versus a phosphonic acid) showed no inhibition at concentrations up to 10 μM. Notably, all compounds where the substituted moiety was not acidic (6 – 10) showed no appreciable inhibition at concentrations up to 10 μM.

Table 1.

IC₅₀ values (μM) for DPA isostere inhibition of NDM-1 catalyzed meropenem hydrolysis.^[a]

Compound	IC50	Compound	IC50
DPA	0.84±0.04	6	>10
1	0.31±0.01	7	>10
2	0.13±0.01	8	>10
3	7.7±0.6	9	>10
4	7.0±0.5	10	>10
5	>10

^[a]Activity measurements were taken in triplicate using varying inhibitor concentrations that bracket each IC₅₀ value, with fitting errors listed above.

To determine the ability of isosteres to inhibit other B1 MBLs, the inhibition of 1 – 4 against NDM-1, IMP-1, and VIM-2 was investigated. An alternative substrate, Fluorocillin (K_{M, NDM-1} = 460 nM), was utilized to increase assay sensitivity.^{28, 43} Lower IC₅₀ values for Fluorocillin compared to meropenem were expected due to using a smaller [S]/K_M ratio in these experiments. IC₅₀ values determined for NDM-1 inhibition via Fluorocillin showed the same trends as with meropenem, with 1 and 2 showing the lowest IC₅₀ values, followed by the parent DPA fragment, with 3 and 4 yielding higher IC₅₀ values among these five compounds (Table 2). Similar trends, but with IC₅₀ values about an order of magnitude higher, were observed for IMP-1 (Table S1, Figure S1) and VIM-2 as well.

Table 2.

IC₅₀ values (μM) for isostere inhibition of B1 MBL catalyzed Fluorocillin hydrolysis.^[a]

Compound	NDM-1	IMP-1	VIM-2
DPA	0.33±0.04	2.54±0.04	2.34±0.04
1	0.064±0.001	0.85±0.06	0.85±0.02
2	0.068±0.002	0.63±0.01	0.62±0.01
3	2.05±0.09	59±5	35±1
4	2.19±0.13	22±1	27±2

^[a]Activity measurements were taken in triplicate using varying inhibitor concentrations that bracket each IC₅₀ value, with fitting errors listed above.

Equilibrium Dialysis.

Equilibrium dialysis experiments were performed to give insight into the inhibition mechanism (metal stripping versus ternary complex formation) of DPA and the active isosteres. After protein purification, NDM-1 was exchanged into ammonium acetate buffer to facilitate the use of ICP-AES in determining metal content. Metal analyses of the resulting protein showed that the enzyme binds ~1.7 equivalents of Zn(II) (Figure 3). While holoNDM-1 is expected to bind 2 equivalents of Zn(II), previous studies have shown that one of the Zn(II) ions binds more weakly than the other, resulting in protein that contain less than the full complement of two Zn(II) ions.^44–45 As previously reported, incubation of NDM-1 with L-captopril, a known competitive inhibitor (Figure 1),⁴⁶ did not show any evidence of Zn(II) removal. In contrast, the Zn(II) content of NDM-1 was significantly reduced when incubated with DPA.²⁸ Isosteres 1 – 4 exhibited different levels of Zn(II) removal from NDM-1. Compounds 3 and 4 behaved more like captopril, removing only small amounts of Zn(II) from NDM-1 even at concentrations up to 128 μM. Conversely, compounds 1 and 2 removed more Zn(II) from NDM-1 than the parent DPA.

Figure 3.

Zn(II) content of NDM-1 (8 μM) upon incubation with increasing concentrations of captopril, DPA, and 1 – 4 (16 μM – 128 μM).

UV−Visible Spectroscopy.

As previously reported, UV-vis spectrophotometry of Co(II)-substituted NDM-1 can be used to probe inhibitor binding to the enzyme.²⁸ The spectrum of CoCo-NDM-1 shows an intense peak between 320 – 350 nm, which has been assigned to a cysteine-to-Co(II) ligand-to-metal charge transfer (LMCT) band, and multiple peaks between 500 – 650 nm, which have been assigned to Co(II) ligand field bands (Figure 4).^{22, 47} The inclusion of captopril to the sample results in large changes in the ligand field transitions, and an increase in the LMCT band. These changes are consistent with the formation of a NDM-1:captopril ternary complex, with the captopril sulfur bridging the two Co(II) ions.⁴⁸ Incubation of CoCo-NDM-1 with EDTA resulted in significant reduction of the LMCT and ligand field peaks, indicating that EDTA strips the metal from NDM-1.²⁸ Incubation of CoCo-NDM-1 with compounds 1 – 4 and DPA resulted in a reduction of absorbance at 500 – 650 nm (Figure 4), which indicates that Co(II) is being removed from the active site to varying degrees. Compound 2 appears to remove the most Co(II), followed by compound 1, DPA, 3, and then 4. The findings with CoCo-NDM-1 are wholly consistent with the equilibrium dialysis results.

Figure 4.

UV-Vis spectrum of CoCo-NDM-1 with captopril and EDTA (top), and with DPA and 1 – 4 (bottom).

Derivatization of Isostere 2.

Although 2 was shown to remove Zn(II) from NDM-1, we sought to reduce the metal-removal behavior via elaboration using a fragment-growth strategy. We previously had success using this approach to optimize DPA.²⁸ Also during the time of this study, a structure of 2 in complex with IMP-1 (PDB 5HH4) was reported.⁴⁹ In this structure, the Zn₁ of IMP-1 is coordinated via residues H116, H118, H196, and a bridging hydroxide in a tetrahedral coordination geometry (Figure 5, standard BBL numbering). The bridging hydroxide also coordinates to Zn₂, which is further ligated by residues D120, C221, and H263, as well as by the pyridine nitrogen donor and carboxylic acid of 2 (in an overall octahedral coordination geometry). The phosphonic acid motif of 2 is shown to hydrogen bond with the bridging hydroxide ion and a nearby S80 residue. Guided by this crystallographic data and the homology of IMP-1 and NDM-1 enzymes, a small library of 2 derivatives (18a – 18m) were designed, synthesized, and tested for NDM-1 inhibition.

Figure 5.

6-(Phosphonomethyl)picolinic acid (2) was reported as a potent inhibitor of B1 and B3 MBLs, and able to restore β-lactam activity in MIC assays. Shown above is 2 complexed with IMP-1 (PDB 5HH4).⁴⁹ Zn(II) ions are shown in orange, coordinating ligands and protein ribbon are shown in grey, 2 is shown in green, and ligand-protein interactions are shown with a yellow-dashed line. Figure was rendered with Molecular Operating Environment (MOE).

The synthesis of 18a – 18m is presented in Scheme 4 and the IC₅₀ values (measured for NDM-1 catalyzed hydrolysis of meropenem) are reported in Table 3. This library consisted of simple aryl derivatives substituted at the 4-position of the pyridine ring. The aryl ring bears substituents (methoxy, hydroxyl, amine, and chorine) in the para-, ortho-, and meta- position of the ring in attempt to generate interactions with nearby residue side chains. To prepare these compounds, 4-hydroxypyridine-2,6-dicarboxylic acid was esterified to 12 with MeOH and catalytic H₂SO₄. Next, 12 was converted to 13 with tetrabutylammonium bromide (TBAB) and phosphorus pentoxide (P₄O₁₀). Compound 13 was treated with sodium borohydride (NaBH₄) at 0 °C to obtain 14, which was further converted to intermediate 15 via phosphorous tribromide (PBr₃). Compound 16 was obtained by heating 15 and triethyl phosphite in toluene at 140 °C for 22h. Intermediates 17a – 17m were obtained via Suzuki cross-coupling procedures utilizing the corresponding boronic acid, (Pd(PPh₃)₄), and K₃PO₄. Compounds 18a – 18m were obtained via hydrolysis.

Scheme 4.

Synthesis of derivatives of isostere 2 (compounds 18a – 18m). Reagents and conditions: (a) MeOH, H₂SO₄ (cat.), 70°C, overnight, 60%; (b) P₂O₅, TBAB, toluene, 100°C, 3 h, 75%; (c) NaBH₄, MeOH/CH₂Cl₂, 0°C, 1 h, 80%; (d) PBr₃, CHCl₃, 0°C, 1 h, 68%; (e) P(OEt)₃, toluene, 140°C, 22 h, 92%; (f) boronic acid, K₃O₄P, Pd(PPh₃)₄, 1,4-dioxanes, 80°C, 18 h, 20 – 90%; (g) 6 M HCl, 100 °C, 24 h, 29 – 97%.

Table 3.

IC₅₀ values (μM) for compounds 18a – 18m against NDM-1 catalyzed meropenem hydrolysis. R refers to substituent in Scheme 4.^[a]

R =		para-		meta-		ortho-
	18a	0.41±0.02	–		–
	18b	0.20±0.01	18c	0.42±0.02	18d	0.30±0.01
	18e	0.48±0.01	18f	0.28±0.01	18g	0.28±0.01
	18h	0.30±0.01	18i	0.40±0.02	18j	0.24±0.01
	18k	0.40±0.01	18l	1.29±0.07	18m	0.95±0.05

^[a]Activity measurements were taken in triplicate using varying inhibitor concentrations that bracket each IC₅₀ value, with fitting errors listed above.

Compounds 18a – 18m did not yield greater inhibitory activity than that of 2 (IC₅₀ = 0.13±0.01 μM). A simple aryl ring (18a), or an aryl ring with methoxy (18b – 18d) or hydroxyl substituent (18e – 18g) were generally well-tolerated, with IC₅₀ = 0.20±0.01 – 0.48±0.01 μM, but did not lead to a marked decrease IC₅₀ value compared to that of 2. Given the structural similarity of 18h – 18j (bearing the aniline substituent) to the previous reported inhibitor, (4-(3-aminophenyl)pyridine-2,6-dicarboxylic acid) (Figure 1), no improvement in inhibition was observed compared to that of 2 (IC₅₀ = 0.24±0.01 – 0.40±0.02 μM). Notably, a chlorine substituent in the ortho- and meta- position of the aryl ring yield a 2- to 3-fold loss in potency. Due to the similarities between the observed inhibition values, no structure-activity relationship (SAR) could be obtained. Although isosteric replacement of the carboxylate group maintains the ability to bind metal ions, the relative positioning of attached substituents is likely altered, possibly due to changes in metal coordination geometries. In the case of DPA and NDM-1, this prevents a simple group-swapping approach, and may necessitate re-optimization starting with the new isostere fragment.

Although lower IC₅₀ values for compounds 18a – 18m (compared to that of isostere 2) were not observed, previous evaluation of DPA isosteres 1 – 4 has shown that an increase in IC₅₀ value may be due to a shift in inhibitor mechanism (from metal stripping to ternary complex formation). To evaluate if this was the case for 18a – 18m, two candidates (18b and 18i) were selected for further inhibition mechanism analysis. Indeed, dialysis experiment revealed 18b and 18i did not perturb the Zn(II) content of NDM-1 at concentrations as high as 16 μM, with only slight Zn(II) removal at concentration of 32 μM (Figure S2). In addition, one equivalent of Zn(II) was retained at the highest tested concentration (128 μM) for 18b and 18i, while less than one equivalent was retained when performing the same experiment with compound 2. This further supports our previous findings and shows the possibility to re-engineer a compound which removes metal from the active site of NDM-1 to one which favors ternary complex formation.

Herein, we investigate the isosteric replacement of a carboxylic acid group of DPA with various surrogate structures. Compounds 1 – 10 yield a range of chemical diversity (various acidic or metal-binding groups) and show that the choice of the carboxylate isostere not only impacts the IC₅₀ value, but also the propensity for metal removal. Replacement of one carboxylic acid with a phosphonic substituent (2) results in a MBP with an extraordinarily low IC₅₀ value, but also with a greater tendency for NDM-1 Zn(II) removal. Replacement of the carboxylic acid with a tetrazole motif (4) yield a higher IC₅₀ value, but reduces NDM-1 Zn(II) removal. Equilibrium dialysis data, in conjunction with UV-vis spectroscopy of CoCoNDM-1, reveals differences in inhibition mechanism: inhibition by 2 occurs primarily via Zn(II) removal, likely forming the inactive mono-Zn(II) NDM-1, while inhibition by 4 is mainly via formation of a ternary complex at the dinuclear Zn(II) site. Because 4 forms a ternary complex at the active site, we can use IC₅₀ and the Cheng-Prusoff relationship for competitive inhibitors to calculate a K_i of 2.2±0.5 μM.⁵⁰ It is important to note that calculating K_i values for metal-stripping inhibitors (such as 1 or 2) is not appropriate because changes in IC₅₀ may not reflect changes in binding affinity. Available data suggesting potential varying inhibition mechanisms between homologous MBLs and 2 (ternary complex formation with IMP-1,⁴⁹ but metal-stripping with NDM-1) was unexpected. This difference may be due to a lower binding affinity for Zn(II) in the Zn₂ site of NDM-1 (K_d = 2 μM)⁴⁵ compared to that of IMP-1 (K_d = 0.3 μM),^51–52 resulting in more facile metal removal. Future studies are required to elucidate and conclude the observed mechanistic differences of 2 against IMP-1 and NDM-1.

In an attempt to reduce the propensity to strip metal from the NDM-1 active site, a small library of derivatives of compound 2 were synthesized (compounds 18a – 18m). It was predicted that 18i, bearing the same amine backbone as a reported inhibitor that uses DPA as a metal-binding group, would result in a decrease in the observed IC₅₀ value. Unfortunately, 18i (and other derivatives) did not yield a lower IC₅₀ value than that of 2, suggesting that the isosteres are not exactly equivalent substitutions of their parent fragment. Isosteric replacement may alter the relative positioning of distant substituents and would necessitate further optimization. Additionally, compounds 18a – 18m were screened against IMP-1 (Supporting Information); however, no structure activity relationship was observed, with relatively flat IC₅₀ values spanning 1.20±0.05 – 4.1±0.2 μM. Dialysis experiment of derivatives 18b and 18i against NDM-1 reveal a reduced tendency to remove Zn(II) from the NDM-1 active site, albeit no reduction in the IC₅₀ value, compared to that of isostere 2 (Figure S2).

Despite the understanding that compounds can inhibit MBLs via a number different mechanisms (covalent inhibition, metal-chelation, reversible competitive inhibition, allosteric inhibition, etc.), the mechanism of action of most reported MBL inhibitors is not well defined.²⁶ Many current studies focus on achieving low IC₅₀ values, with little consideration regarding the mechanism of action. Here, we clearly demonstrate how the mechanism of inhibition can significantly differ between similar isosteres and homologous enzymes and illustrate challenges to optimizing inhibitors with metal-stripping mechanisms. In the design of future MBL inhibitors, a loss of potency may be preferred when selecting compounds with a desired mechanism of action for further optimization (i.e., selection of 4 over 2). This kind of early stage evaluation of how the inhibitor interacts with the active site will crucial for accelerating NDM (and other MBLs) inhibitor development.

Conclusions

In summary, we report the synthesis and inhibitory activity of ten DPA isosteres and investigated the mechanism of action of the most active compounds. We show that the exchange of the carboxylic acid with other acidic isosteres not only drastically affects the IC₅₀ value but also the mechanism of action. The compounds which showed double-digit nanomolar IC₅₀ is predicted to act via a metal-stripping mechanism, while compounds which displayed single-digit micromolar IC₅₀ is predicted to form a ternary complex with the enzyme. Preliminary data suggests differences in inhibition mechanism between homologous NDM-1 and IMP-1 and isostere 2. Additionally, we demonstrate the potential to morph metal-stripping compounds into inhibitors which favor ternary complex formation. This study demonstrates the importance of considering the inhibition mechanism, and the utility of bioisosteric replacement for routinely-used metal binding motifs (i.e., carboxylic acid) in NDM-1 inhibitor development.

Experimental Section

All reagents and solvents were obtained from commercial sources and used without further purification. Corning UV-transparent 96-well microplates (3635), Corning black polystyrene round-bottom 96-well microplates (3792), 3-((3-cholamidopropyl)-dimethylammonio)-1-propanesulfonate (CHAPS), 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), dimethyl sulfoxide (DMSO), L-captopril and Fluorocillin green 495/525 β-lactamase substrate, soluble product, were purchased from Thermo Fisher Scientific Inc. (Fair Lawn, NJ). All other reagents were purchased from Sigma-Aldrich Inc. (St. Louis, MO). Screening assays were performed on a PerkinElmer Victor3 V 1420 multilabel counter plate reader. Fluorescent assays were performed on a PerkinElmer Victor3 V fluorescent plate reader. Column chromatography was performed using a Teledyne ISCO CombiFlash Rf system with prepacked silica cartridges. All ¹H NMR and ¹³C spectra were recorded at either ambient temperature or at 35 °C using Varian 400 Mercury Plus or Varian VX500 instrument located in the Department of Chemistry and Biochemistry at the University of California, San Diego. Mass spectrometry data were obtained from the University of California San Diego Chemistry and Biochemistry Mass Spectrometry Facility (MMSF). The purity of all compounds used for screening were determined to be ≥95% by high performance liquid chromatography (HPLC).

6-Phosphonopicolinic acid (1).

A solution of methyl 6-bromopicolinate (300 mg, 1.39 mmol), diethylphosphate (172.0 μL, 1.39 mmol), Pd₂(dba)₃ (67 mg, 0.69 mmol), Pd(dppf)Cl₂ (77 mg, 0.14 mmol), and trimethylamine (387.0 μL, 2.78 mmol) were dissolved in toluene (10 mL), and heated to 90°C for 20 h. Ethyl acetate (20 mL) was added to the reaction mixture, and the solution was filtered through a pad of celite. The collected organic layers were concentrated in vacuo. The crude mixture was purified via flash column chromatography, with the intermediate eluting at 50% ethyl acetate in hexanes. The intermediate was heated under reflux conditions with 6 M HCl (3 mL) for 20 h. The excess HCl was removed in vacuo, and co-evaporated with copious amounts of methanol and water until a precipitate was observed. The product was collected via vacuum filtration as a white solid in 19% yield (54 mg, 0.27 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ = 8.26 (d, J = 7.7 Hz, 1H), 8.19 – 8.00 (m, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.4, 156.2, 149.0, 138.0, 129.4, 126.1. ESI-MS(−) calculated for [C₆H₅NO₅P]⁻ m/z 202.10, found m/z 201.98 [M-H]⁻.

6-(Phosphonomethyl)picolinic acid (2).

A solution of 11 (200 mg, 0.87 mmol) and P(OEt)₃ (1.6 g, 9.56 mmol) were heated in toluene (20 mL) at 140°C for 2 h to give a clear liquid solution. Excess P(OEt)₃ and toluene were removed in vacuo, and the intermediate was purified via flash column chromatography eluting at 100% ethyl acetate in hexanes as a clear oil in 80% yield (200 mg, 0.70 mmol). ¹H NMR (400 MHz, CDCl₃-d): δ 8.02 (d, J = 8.0 Hz, 1H), 7.80 (td, J = 7.8, 2.1 Hz, 1H), 7.62 (d, J = 8.0 Hz, 1H), 4.13 – 4.06 (m, 4H), 3.99 (d, J = 2.0 Hz, 3H), 3.53 (d, J = 22.0 Hz, 2H), 1.31 – 1.25 (m, 6H). ESI-MS(+) calculated for [C₁₂H₁₉NO₅P]⁺ m/z 288.10, found m/z 288.29 [M+H]⁺.

The intermediate, methyl 6-((diethoxyphosphoryl)methyl)picolinate (200 mg, 0.70 mmol) was refluxed in 6 M HCl (5 mL) for 27 h. The excess HCl was removed in vacuo and co-evaporate with copious amounts of methanol and water until a white precipitate was observed. Compound 2 was collected via vacuum filtration as a white solid in 86% yield (130 mg, 0.60 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.02 (t, J = 7.7 Hz, 1H), 7.95 (d, J = 7.5 Hz, 1H), 7.67 (d, J = 7.6 Hz, 1H), 3.38 (s, 1H), 3.32 (s, 1H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.6, 155.4, 148.0, 137.9, 127.8, 122.8, 38.3. ESI-MS(−) calculated for [C₇H₈NO₅P]⁻ m/z 217.01, found m/z 216.00 [M-H]⁻.

6-(Carboxymethyl)picolinic acid (3).

To a solution of 11 (2.5 g, 10.87 mmol) dissolved in THF (50 mL) was added a pre-dissolved solution of KCN (1.1 g, 16.30 mmol) in water (9 mL). All supplies in contact with KCN were quenched with 1 M sodium thiolsuflate prior to disposal. The mixture was stirred at 50°C for 19 h and quenched with 1 M sodium thiosulfate solution. The mixture was extracted with CH₂Cl₂ (25 mL X5). The combined organic layers were dried over MgSO₄, filtered, and concentrated in vacuo. Intermediate methyl 6-(cyanomethyl)picolinate was purified via flash column chromatography, eluting at 58% ethyl acetate in hexanes to yield yellow crystals in 49% yield (941 mg, 5.34 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.09 – 7.95 (m, 2H), 7.68 (t, J = 5.5 Hz, 1H), 4.30 (s, 2H), 3.88 (s, 3H). ESI-MS(+) calculated for [C₉H₉N₂O₂]⁺ m/z 177.06, found m/z 177.18 [M+H]⁺.

The intermediate, methyl 6-(cyanomethyl)picolinate was dissolved in 12 M HCl and heated to 100 °C for 12 h. The solvent was removed in vacuo and the crude was hydrolyzed in 3 mL of 1 M NaOH at 60 °C for 6 h. The solution was the acidified with 4 M HCl to pH 4, and product 3 was collected via vacuum filtration as a white solid in 30% yield (37 mg, 0.20 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 7.98 – 7.86 (m, 2H), 7.56 (d, J = 6.3 Hz, 1H), 3.83 (s, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 172.1, 166.6, 155.8, 148.3, 138.3, 128.0, 123.4, 43.6. ESI-MS(+) calculated for [C₈H₈NO₄]⁺ m/z 182.04, found m/z 182.25 [M+H]⁺.

6-(1H-Tetrazol-5-yl)picolinic acid (4).

Methyl 6-bromopicolinate (3 g, 13.886 mmol) and copper(I) cyanide (2.49 g, 27.77 mmol) were dissolved in pyridine (120 mL) and heated to 116°C for 4 h. The reaction was monitored via TLC. Upon completion of the reaction, the mixture was filtered, and the filtrate was concentrated in vacuo. Aqueous NaHCO₃ (50 mL) and CH₂Cl₂ (50 mL) were added to the crude mixture. The aqueous layer was extracted with CH₂Cl₂ (50 mL X 3), and the combined organic layers were dried over MgSO₄, concentrated in vacuo, and purified with flash chromatography. Intermediate methyl 6-cyanopicolinate eluted 40% ethyl acetate in hexanes as a yellow solid in 20% yield (470 mg, 2.90 mmol). ¹H NMR (400 MHz, CDCl₃-d): δ 8.34 (d, J = 7.9 Hz, 1H), 8.04 (t, J = 7.9 Hz, 1H), 7.88 (d, J = 7.8 Hz, 1H), 4.04 (s, 3H). ESI-MS(+) calculated for [C₈H₇N₂O₂]⁺ m/z 163.05, found m/z 163.10 [M+H]⁺.

In a round bottom flask, intermediate methyl 6-cyanopicolinate (200 mg, 1.23 mmol), sodium azide (408 mg, 6.29 mmol), and ammonia hydrochloride (336 mg, 6.29 mmol) were dissolved in anhydrous DMF (10 mL). The reaction was heated under N₂ at 130°C for 20 h. After cooling, the inorganic salts were removed via vacuum filtration, and washed with hot DMF. The organic filtrate was concentrated in vacuo to yield a yellow solid. The solid was suspended in a solution of 2 M HCl (3 mL), and stirred at 25°C for 1 h. The precipitate was collected via vacuum filtration and washed with copious amounts of cold water to afford product 4 as a white solid in 98% yield (232 mg, 1.21 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.37 (d, J = 7.5 Hz, 1H), 8.26 – 8.13 (m, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 165.2, 155.1, 149.1, 144.3, 140.3, 126.8, 126.2. ESI-MS(−) calculated for [C₇H₄N₅O₂]⁻ m/z 190.04, found m/z 189.25 [M-H]⁻.

6-(Sulfomethyl)picolinic acid (5).

In a round bottom flask, 11 (200 mg, 0.87 mmol) and Na₂SO₃ (110 mg, 0.87 mmol) were dissolved in H₂O (10 mL), and heated to 100°C for 16 h. The reaction was cooled to room temperature, and H₂O was removed in vacuo until it reached one-third of the original volume. The white precipitate was collected via vacuum filtration as the ester intermediate. The intermediate was hydrolyzed via stirring in 4 M HCl at 100 °C for 16 h. The excess HCl was removed in vacuo, and the product was co-evaporated with copious amounts of MeOH and H₂O until white crystals were observed. The precipitate was collected via vacuum filtration to yield compound 5 as a white crystal in 41% yield (77.0 mg, 0.35 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.08 – 7.88 (m, 2H), 7.79 (d, J = 6.0 Hz, 1H), 4.06 (s, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 164.7, 155.0, 145.2, 140.7, 129.9, 124.1, 57.6. ESI-MS(−) calculated for [C₇H₆NO₅S]⁻ m/z 216.10, found m/z 216.08 [M-H]⁻.

6-((Methylsulfonyl)methyl)picolinic acid (6).

In a round bottom flask, 11 (300 mg, 1.30 mmol) and sodium methanesulfinate (266 mg, 2.61 mmol) were dissolved in DMF (15 mL). The solution was heated to 120°C for 2 h. DMF was removed in vacuo, and the crude product was purified via flash column chromatography eluting at 90% ethyl acetate in hexanes. The intermediate was hydrolyzed by stirring in 3:1 1M NaOH:THF at 70°C for 3 h. The THF was removed in vacuo, and the solution was acidified with 4 M HCl until pH 4. The aqueous layer was extracted with ethyl acetate (20 mL X 3) to afford 6 a white solid in 43% yield over two steps (120 mg, 0.56 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.03 (s, 2H), 7.70 (t, J = 4.5 Hz, 1H), 4.75 (s, 2H), 3.08 (s, 3H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.3, 150.7, 149.2, 138.9, 129.3, 124.5, 61.5, 41.1. ESI-MS(−) calculated for [C₈H₈NO₄S]⁻ m/z 214.02, found m/z 214.16 [M-H]⁻.

6-(Methylsulfonamido)picolinic acid (7).

Commercially available methyl 6-aminopicolinate (300 mg, 1.97 mmol) was dissolved in a solution of triethylamine (1 mL):CH₂Cl₂ (5 mL) and cooled to 0°C. Next, methanesulfonyl chloride (168.0 μL, 2.17 mmol) was added drop-wise to the mixture. The reaction was stirred at room temperature for 16 h. The solvent was removed in vacuo, and the crude mixture was purified via flash column chromatography. The ester intermediate eluted at 62% ethyl acetate in hexanes as a white solid. The intermediate was hydrolyzed via stirring at room temperature for 16 h in 3:1 1M NaOH:THF (8 mL). The excess THF was removed in vacuo, and the solution was acidified with 4 M HCl to pH 4. The aqueous layer was extracted with ethyl acetate (10 mL X 3), and the combined organic layers were dried over MgSO₄, filtered, and concentrated in vacuo to afford 7 as a white solid. ¹H NMR (400 MHz, MeOD-d₄): δ 7.89 (t, J = 7.9 Hz, 1H), 7.79 (d, J = 7.5 Hz, 1H), 7.28 (d, J = 8.3 Hz, 1H), 3.33 (s, 3H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.2, 152.4, 147.1, 140.1, 119.5, 115.9, 42.5. ESI-MS(−) calculated for [C₇H₇N₂O₄S]⁻ m/z 215.10, found m/z 215.05 [M-H]⁻.

6-((N-Methylmethylsulfonamido)methyl)picolinic acid (8).

In a round bottom flask, 11 (158 mg, 0.69 mmol), N-methylmethanesulfonamide (74 mg, 0.69 mmol), and potassium carbonate (48 mg, 0.34 mmol) were dissolved in acetonitrile (10 mL) and heated to 75°C for 20 h. The solvent was removed in vacuo and the crude mixture was purified via flash column chromatography. The ester intermediate eluted at 3% MeOH in CH₂Cl₂ as a yellow oil. The intermediate was hydrolyzed by stirring in 1 M NaOH at 70°C for 3 h. The solution was the acidified with 4 M HCl to pH 4, and the aqueous layer was extracted with ethyl acetate (20 mL X 4). The combined organic layers were dried over MgSO₄, filtered, and concentrated in vacuo to afford 8 a pale solid in 55% yield over two steps (92 mg, 0.38 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.04 – 7.90 (m, 2H), 7.61 (d, J = 7.3 Hz, 1H), 4.46 (s, 2H), 3.08 (s, 3H), 2.75 (s, 3H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.5, 157.4, 148.5, 138.9, 125.7, 124.1, 54.9, 36.7, 35.2. ESI-MS(−) calculated for [C₉H₁₁N₂O₄S]⁻ m/z 243.04, found m/z 243.31 [M-H]⁻.

6-(Thiazol-2-yl)picolinic acid (9).

In a round bottom flask, methyl 6-bromopicolinate (300 mg, 1.39 mmol) and 2-(tributylstannyl)thiazole (624 mg, 1.67 mmol) were dissolved in anhydrous THF (15 mL). The solution was purged with N₂ for 15 min, followed by the addition of bis(triphenylphosphine)palladium(II) dichloride (97 mg, 0.14 mmol). The reaction was heated to 75°C for 18 h. Upon completion, water (20 mL) and ethyl acetate (20 mL) were added to the brown slurry. The aqueous layer was extracted with ethyl acetate (20 mL X 3), and the combined organic layers were dried over MgSO₄, filtered, and concentrated in vacuo. The ester intermediate was purified via column chromatography, eluting at 45% ethyl acetate in hexanes as a yellow solid. The intermediate was hydrolyzed in 3:1 1 M NaOH:THF at 70 °C for 3 h. THF was removed in vacuo, and the aqueous layer was acidified with 4 M HCl to pH 4. The precipitate was collected via vacuum filtration to afford compound 9 as a tan solid in 74% yield (213 mg, 1.03 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 9.18 (s, 1H), 8.68 (s, 1H), 8.22 (d, J = 8.0 Hz, 1H), 8.05 (t, J = 7.8 Hz, 1H), 7.94 (d, J = 7.7 Hz, 1H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.1, 157.0, 150.5, 149.0, 142.3, 139.7, 139.1, 124.2, 123.4. ESI-MS(−) calculated for [C₉H₅N₂O₂S]⁻ m/z 205.01, found m/z 205.18 [M-H]⁻.

6-(Oxazol-2-yl)picolinic acid (10).

In a round bottom flask, methyl 6-bromopicolinate (300 mg, 1.39 mmol) and 2-(tributylstannyl)oxazole (596 mg, 1.67 mmol) were dissolved in anhyrdous THF (15 mL). The solution was purged with N₂ for 15 min, followed by the addition of bis(triphenylphosphine)palladium(II) dichloride (79 mg, 0.14 mmol). The solution was heated to 75°C for 18 h. Upon completion, water (20 mL) and ethyl acetate (20 mL) were added to the slurry. The aqueous layer was extracted with ethyl acetate (20 mL X 3), and the combined organic layers were dried over MgSO₄, filtered, and concentrated in vacuo. The ester intermediate was purified via column chromatography, eluting at 60% ethyl acetate in hexanes as a yellow solid. The intermediate was hydrolyzed in 3:1 1 M NaOH:THF at 70°C for 3 h. THF was removed in vacuo, and the aqueous layer was acidified with 4 M HCl until pH 4. The aqueous layer was extracted with ethyl acetate (20 mL X 3), and the combined organic layers were dried over MgSO₄, filtered, and concentrated in vacuo to yield 10 as a white solid in 8% yield (20 mg, 0.12 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.35 (s, 1H), 8.28 (dd, J = 6.4, 2.6 Hz, 1H), 8.16 – 8.10 (m, 2H), 7.49 (s, 1H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.2, 160.0, 149.4, 145.9, 142.0, 139.4, 129.4, 126.2, 125.3. ESI-MS(−) calculated for [C₉H₅N₂O₃]⁻ m/z 189.03, found m/z 189.20 [M-H]⁻.

Methyl 6-(bromomethyl)picolinate (11).

To a solution of methyl 6-(hydroxymethyl)picolinate (1.0 g, 5.98 mmol) in chloroform (25 mL) at 0°C was added PBr₃ (1.9 g, 7.18 mmol) dropwise over the course of 15 min. The reaction was stirred at room temperature for 3 h. The reaction mixture was quenched with saturated sodium carbonate in water and extracted with chloroform (20 mL X 4). The combined organic layers were dried over MgSO₄, filtered, and concentrated in vacuo. The product was purified via flash column chromatography, eluting at 45% ethyl acetate in hexanes to afford 11 as a white crystal in 70% yield (95 mg, 4.14 mmol). ¹H NMR (400 MHz, CDCl₃-d): δ 8.06 (d, J = 7.8 Hz, 1 H), 7.86 (td, J = 7.8, 1.4 Hz), 7.68 (d, J = 7.8 Hz, 1 H), 4.64 (s, 2H), 4.01 (s, 3H). ESI-MS(+) calculated for [C₈H₉BrNO₂]⁺ m/z 229.98, found m/z 230.24 [M+H]⁺.

Dimethyl 4-hydroxypyridine-2,6-dicarboxylate (12).

Commercially-available 4-hydroxypyridine-2,6-dicarboxylic acid (10.0 g, 55 mmol) was dissolved in MeOH (500 mL) and H₂SO₄ (cat.) was added. The reaction was stirred at 70°C for 16 h. The solvent was removed in vacuo, and the crude mixture was purified via column chromatography, eluting at 1% MeOH in CH₂Cl₂ as a yellow solid in 60% yield (7.0 g, 30 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 7.59 (s, 2H), 3.96 (s, 6H). ESI-MS(+) calculated for [C₉H₁₀NO₅]⁺ m/z 212.06, found m/z 212.07 [M+H]⁺.

Dimethyl 4-bromopyridine-2,6-dicarboxylate (13).

Tetrabutylammonium bromide (28.623 g, 88.789 mmol) and phosphorus(V) oxide (12.6 g, 89 mmol) were dissolved in dry toluene (50 mL), and heated to 100°C for 30 min under N₂. Compound 12 (7.5 g, 36 mmol) was added to the solution and the mixture was heated at 100 °C for 3 h. The toluene layer was decanted, and an additional 50 mL of toluene was added to the residual brown oil. The mixture was heated to 100°C for an additional 30 min, and the toluene layer was decanted again. Addition and removal of toluene was repeated three times. The combined toluene layers were dried in vacuo, and the crude mixture was purified via column chromatography. Compound 13 eluted at 45% ethyl acetate in hexanes as yellow needles in 75% yield (7.3 g, 27 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.42 (s, 2H), 3.91 (s, 6H). ESI-MS(+) calculated for [C₉H₉BrNO₄]⁺ m/z 273.97, found m/z 274.26 and 275.25 [M+H]⁺.

Methyl 4-bromo-6-(hydroxymethyl)picolinate (14).

In a round bottom flask, compound 13 (1.0 g, 3.65 mmol) was dissolved in a solution of MeOH:CH₂Cl₂ (16 mL:4 mL) and cooled to 0°C. NaBH₄ (138 mg, 3.65 mmol) was added to the mixture and stirred at 0°C for 1 h. The reaction was quenched with aqueous NaHCO₃ (10 mL). The mixture was extracted with CH₂Cl₂ (10 mL X 3), and the combined organic layers were dried in vacuo. The crude mixture was purified via column chromatography, with 14 eluting at 70% ethyl acetate in hexanes as a white crystal in 80% yield (689 mg, 2.80 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.17 (s, 1H), 7.97 (s, 1H), 4.73 (s, 2H), 3.97 (s, 3H). ESI-MS(+) calculated for [C₈H₉BrNO₃]⁺ m/z 245.98, found m/z 246.16 and 248.09 [M+H]⁺.

Methyl 4-bromo-6-(bromomethyl)picolinate (15).

Compound 14 (400 mg, 1.63 mmol) was dissolved in CHCl₃ (4 mL) and cooled to 0°C. Next, PBr₃ (528 mg, 1.95 mmol) was added dropwise, and the reaction was stirred at 0°C for 1 h. The reaction was tracked via TLC. Upon completion, the reaction mixture was quenched with aqueous Na₂CO₃ (40 mL), and the aqueous layer was extracted with chloroform (40 mL X 4). The combined organic layers were dried over MgSO₄, filtered, and concentrated in vacuo to yield a yellow oil with white precipitate. The crude product was purified via flash column chromatography, with compound 15 eluting at 42% ethyl acetate in hexanes as a white solid in 68% yield (342 mg, 1.11 mmol). ¹H NMR (400 MHz, CDCl₃-d): δ 8.21 (s, 1H), 7.86 (s, 1H), 4.59 (s, 2H), 4.02 (s, 3H). ESI-MS(+) calculated for [C₈H₈Br₂NO₂]⁺ m/z 307.89, found m/z 308.07 and 310.00 [M+H]⁺.

Methyl 4-bromo-6-((diethoxyphosphoryl)methyl)picolinate (16).

In a round bottom flask, 15 (3.5 g, 11.39 mmol) and triethyl phosphite (5.7 g, 34.16 mmol) were dissolved in toluene (30 mL). The reaction was heated at 140°C for 22 h. Excess triethyl phosphite and toluene were removed in vacuo, and the crude product was purified via flash column chromatography. Compound 16 eluted at 70% ethyl acetate in hexanes as a clear oil in 92% yield (3.9 g, 10.52 mmol). ¹H NMR (400 MHz, CDCl₃-d): δ 8.17 (s, 1H), 7.78 (s, 1H), 4.12 (dq, J = 10.2, 3.6 Hz, 4H), 3.99 (s, 3H), 3.50 (d, J = 22.0 Hz, 2H), 1.29 (t, J = 7.0 Hz, 6H). ESI-MS(+) calculated for [C₁₂H₁₈BrNO₅P]⁺ m/z 366.01, found m/z 366.03 [M+H]⁺.

General procedures for the synthesis of compounds 17a – 17m.

In a round bottom flask, compound 16 (1 eq.), the corresponding boronic acid (1.2 eq.) and K₃O₄P (2 eq.) were dissolved in 1,4-dioxanes (5 mL). The solution was purged under N₂ for 20 min, followed by the addition of tetrakis(triphenylphosphine)palladium(0) (0.1 eq.). The reaction was heated to 80°C for 18 h under N₂. The crude mixture was filtered through a pad of celite, washed with ethyl acetate, and the combined organic layers were concentrated in vacuo. The crude products were purified via flash column chromatography to afford derivatives 17a – 17m.

Methyl 6-((diethoxyphosphoryl)methyl)-4-phenylpicolinate (17a).

Yield: 74% (59 mg, 0.16 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.26 (s, 1H), 7.90 (d, J = 2.1 Hz, 1H), 7.77 (d, J = 6.5 Hz, 2H), 7.65 – 7.31 (m, 3H), 4.28 – 4.07 (m, 4H), 3.65 (d, J = 22.2 Hz, 2H), 1.28 (t, J = 7.0 Hz, 6H). ESI-MS(+) calculated for [C₁₈H₂₃NO₅P]⁺ m/z 364.13, found m/z 364.11 [M+H]⁺.

Methyl 6-((diethoxyphosphoryl)methyl)-4-(4-methoxyphenyl)picolinate (17b).

Yield: 79% (120 mg, 0.31 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.14 (s, 1H), 7.87 (s, 1H), 7.79 (d, J = 7.3 Hz, 2H), 7.10 (d, J = 7.7 Hz, 2H), 4.11 – 3.94 (m, 4H), 3.89 (s, 3H), 3.82 (s, 3H), 3.56 (d, J = 21.7 Hz, 2H), 1.18 (t, J = 6.2 Hz, 6H). ESI-MS(+) calculated for [C₁₉H₂₅NO₆P]⁺ m/z 394.14, found m/z 394.28 [M+H]⁺.

Methyl 6-((diethoxyphosphoryl)methyl)-4-(3-methoxyphenyl)picolinate (17c).

Yield: 58% (86 mg, 0.22 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.26 (s, 1H), 7.90 (s, 1H), 7.45 (t, J = 8.0 Hz, 1H), 7.34 (d, J = 7.8 Hz, 1H), 7.30 (s, 1H), 7.07 (d, J = 8.2 Hz, 1H), 4.21 – 4.07 (m, 4H), 4.00 (s, 3H), 3.88 (s, 3H), 3.66 (d, J = 22.2 Hz, 2H), 1.29 (t, J = 7.0 Hz, 6H). ESI-MS(+) calculated for [C₁₉H₂₅NO₆P]⁺ m/z 394.14, found m/z 416.13 [M+Na]⁺.

Methyl 6-((diethoxyphosphoryl)methyl)-4-(2-methoxyphenyl)picolinate (17d).

Yield: 87% (125 mg, 0.32 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.20 (s, 1H), 7.81 (s, 1H), 7.50 – 7.36 (m, 2H), 7.15 (d, J = 8.4 Hz, 1H), 7.09 (t, J = 7.5 Hz, 1H), 4.24 – 4.06 (m, 4H), 3.98 (s, 3H), 3.86 (s, 3H), 3.63 (d, J = 22.2 Hz, 2H), 1.28 (t, J = 7.0 Hz, 6H). ESI-MS(+) calculated for [C₁₉H₂₅NO₆P]⁺ m/z 394.14, found m/z 416.13 [M+Na]⁺.

Methyl 6-((diethoxyphosphoryl)methyl)-4-(4-hydroxyphenyl)picolinate (17e).

Yield: 20% (27 mg, 0.07 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.21 (s, 1H), 7.83 (s, 1H), 7.65 (d, J = 6.4 Hz, 2H), 6.92 (d, J = 6.7 Hz, 2H), 4.31 – 4.04 (m, 4H), 3.99 (s, 3H), 3.61 (d, J = 23.0 Hz, 2H), 1.28 (t, J = 7.0 Hz, 7H). ESI-MS(+) calculated for [C₁₈H₂₃NO₆P]⁺ m/z 380.12, found m/z 402.13 [M+Na]⁺.

Methyl 6-((diethoxyphosphoryl)methyl)-4-(3-hydroxyphenyl)picolinate (17f).

Yield: 90% (135 mg, 0.40 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.23 (s, 1H), 7.86 (s, 1H), 7.35 (t, J = 7.9 Hz, 1H), 7.22 (d, J = 7.9 Hz, 1H), 7.16 (s, 1H), 6.92 (d, J = 8.0 Hz, 1H), 4.21 – 4.06 (m, 4H), 4.00 (s, 3H), 3.65 (d, J = 20.8 Hz, 2H), 1.29 (t, J = 6.3 Hz, 6H). ESI-MS(+) calculated for [C₁₈H₂₃NO₆P]⁺ m/z 380.12, found m/z 402.13 [M+Na]⁺.

Methyl 6-((diethoxyphosphoryl)methyl)-4-(2-hydroxyphenyl)picolinate (17g).

Yield: 63% (91 mg, 0.24 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.31 (s, 1H), 7.91 (s, 1H), 7.41 (d, J = 7.4 Hz, 1H), 7.28 (t, J = 7.8 Hz, 1H), 7.02 – 6.80 (m, 2H), 4.21 – 4.05 (m, 4H), 3.98 (s, 3H), 3.63 (d, J = 22.1 Hz, 2H), 1.28 (t, J = 7.0 Hz, 6H). ESI-MS(+) calculated for [C₁₈H₂₃NO₆P]⁺ m/z 380.12, found m/z 402.13 [M+Na]⁺.

Methyl 4-(4-acetamidophenyl)-6-((diethoxyphosphoryl)methyl)picolinate (17h)

Yield: 43% (72 mg, 0.170mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.36 (s, 1H), 8.02 (s, 1H), 7.87 – 7.72 (m, 4H), 4.23 – 4.09 (m, 4H), 4.03 (s, 3H), 3.72 (d, J = 22.3 Hz, 2H), 2.16 (s, 3H), 1.29 (t, J = 7.0 Hz, 6H). ESI-MS(+) calculated for [C₂₀H₂₆N₂O₆P]⁺ m/z 421.15, found m/z 421.20 [M+H]⁺.

Methyl 4-(3-acetamidophenyl)-6-((diethoxyphosphoryl)methyl)picolinate (17i).

Yield: 39% (60 mg, 0.14 mmol). ¹H NMR (400 MHz, CDCl₃-d): δ 8.22 (s, 1H), 8.16 (s, 1H), 7.85 (d, J = 6.8 Hz, 2H), 7.68 (d, J = 7.0 Hz, 1H), 7.43 – 7.32 (m, 2H), 4.28 – 4.05 (m, 4H), 4.00 (s, 3H), 3.67 (d, J = 21.9 Hz, 2H), 2.22 (s, 3H), 1.30 (t, J = 7.1 Hz, 6H). ESI-MS(+) calculated for [C₂₀H₂₆N₂O₆P]⁺ m/z 421.15, found m/z 421.19 [M+H]⁺.

Methyl 4-(2-acetamidophenyl)-6-((diethoxyphosphoryl)methyl)picolinate (17j).

Yield: 84% (136 mg, 0.32 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.17 (s, 1H), 7.80 (s, 1H), 7.64 – 7.29 (m, 4H), 4.32 – 4.08 (m, 4H), 4.02 (s, 3H), 3.74 (d, J = 22.3 Hz, 2H), 1.99 (s, 3H), 1.31 (t, J = 7.0 Hz, 6H). ESI-MS(+) calculated for [C₂₀H₂₆N₂O₆P]⁺ m/z 421.15, found m/z 421.21 [M+H]⁺.

Methyl 4-(4-chlorophenyl)-6-((diethoxyphosphoryl)methyl)picolinate (17k).

Yield: 60% (91 mg, 0.23 mmol). ¹H NMR (400 MHz, MeOD-d₄₎ : δ 8.28 (s, 1H), 7.91 (s, 1H), 7.80 (d, J = 8.7 Hz, 2H), 7.55 (d, J = 8.6 Hz, 2H), 4.23 – 4.08 (m, 5H), 4.00 (s, 3H), 3.66 (d, J = 22.2 Hz, 2H), 1.28 (t, J = 7.1 Hz, 6H). ESI-MS(+) calculated for [C₁₈H₂₂ClNO₅P]⁺ m/z 398.09, found m/z 398.24 [M+H]⁺.

Methyl 4-(3-chlorophenyl)-6-((diethoxyphosphoryl)methyl)picolinate (17l).

Yield: 42% (63 mg, 0.16 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.28 (s, 1H), 7.93 (s, 1H), 7.84 (s, 1H), 7.78 – 7.68 (m, 1H), 7.61 – 7.47 (m, 2H), 4.24 – 4.08 (m, 4H), 4.01 (s, 3H), 3.67 (d, J = 22.0 Hz, 2H), 1.29 (t, J = 6.8 Hz, 6H). ESI-MS(+) calculated for [C₁₈H₂₂ClNO₅P]⁺ m/z 398.09, found m/z 398.15 [M+H]⁺.

Methyl 4-(2-chlorophenyl)-6-((diethoxyphosphoryl)methyl)picolinate (17m).

Yield: 71% (114 mg, 0.29 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.11 (s, 1H), 7.74 (s, 1H), 7.61 – 7.40 (m, 4H), 4.23 – 4.07 (m, 4H), 3.99 (s, 3H), 3.67 (d, J = 22.2 Hz, 2H), 1.28 (t, J = 7.1 Hz, 6H). ESI-MS(+) calculated for [C₁₈H₂₂ClNO₅P]⁺ m/z 398.09, found m/z 398.13 [M+H]⁺.

General procedures for the synthesis of 18a – 18m.

Compounds 17a – 17m were dissolved the in a solution of 6 M HCl and heated to 100°C for 24 h. Excess HCl was removed in vacuo followed by co-evaporation with copious amounts of water and MeOH until precipitate was observed. The precipitate was collected via vacuum filtration and washed with cold water to afford products 18a – 18m.

4-Phenyl-6-(phosphonomethyl)picolinic acid (18a).

Yield: 29% (14 mg, 0.05 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.65 (s, 1H), 8.39 (s, 1H), 8.14 – 7.57 (m, 5H), 3.82 (d, J = 22.3 Hz, 2H). ESI-MS(−) calculated for [C₁₃H₁₁NO₅P]⁻ m/z 292.03, found m/z 292.04 [M-H]⁻.

4-(4-Methoxyphenyl)-6-(phosphonomethyl)picolinic acid (18b).

Yield: 61% (48 mg, 0.15 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.08 (s, 1H), 7.82 (s, 1H), 7.75 (d, J = 8.6 Hz, 2H), 7.09 (d, J = 8.5 Hz, 2H), 3.81 (s, 3H), 3.30 (d, J = 21.4 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.7, 161.0, 156.2, 148.9, 148.3, 129.2, 128.7, 124.3, 119.6, 115.2, 55.8, 38.3. ESI-MS(−) calculated for [C₁₄H₁₃NO₆P]⁻ m/z 322.05, found m/z 322.10 [M-H]⁻.

4-(3-Methoxyphenyl)-6-(phosphonomethyl)picolinic acid (18c).

Yield: 71% (50 mg, 0.15 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.33 (s, 1H), 8.01 (s, 1H), 7.53 – 7.31 (m, 3H), 7.09 (d, J = 7.4 Hz, 1H), 3.89 (s, 3H), 3.56 (d, J = 21.8 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.7, 160.4, 156.2, 149.4, 148.6, 138.8, 131.0, 125.3, 120.4, 119.6, 115.7, 112.7, 55.8, 38.4. ESI-MS(−) calculated for [C₁₄H₁₃NO₆P]⁻ m/z 322.05, found m/z 322.02 [M-H]⁻.

4-(2-Methoxyphenyl)-6-(phosphonomethyl)picolinic acid (18d).

Yield: 67% (67 mg, 0.21 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.32 (s, 1H), 7.96 (s, 1H), 7.59 – 7.38 (m, 2H), 7.17 (d, J = 8.2 Hz, 1H), 7.11 (t, J = 7.4 Hz, 1H), 3.88 (s, 3H), 3.55 (d, J = 21.7 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.7, 156.7, 155.5, 147.9, 147.2, 131.2, 130.6, 127.7, 126.6, 123.3, 121.5, 112.5, 56.1, 38.3. ESI-MS(−) calculated for [C₁₄H₁₃NO₆P]⁻ m/z 322.05, found m/z 322.04 [M-H]⁻.

4-(4-Hydroxyphenyl)-6-(phosphonomethyl)picolinic acid (18e).

Yield: 82% (16 mg, 0.06 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.38 (s, 1H), 8.08 (s, 1H), 7.81 (d, J = 8.2 Hz, 2H), 6.96 (d, J = 8.0 Hz, 2H), 3.56 (d, J = 21.4 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.7, 159.5, 155.9, 148.8, 148.7, 128.7, 127.5, 124.0, 119.4, 116.6, 38.2. ESI-MS(−) calculated for [C₁₃H₁₁NO₆P]⁻ m/z 308.03, found m/z 308.02 [M-H]⁻.

4-(3-Hydroxyphenyl)-6-(phosphonomethyl)picolinic acid (18f).

Yield: 82% (77 mg, 0.30 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.47 (s, 1H), 8.21 (s, 1H), 7.48 – 7.33 (m, 2H), 7.29 (s, 1H), 7.01 (d, J = 7.4 Hz, 1H), 3.73 (d, J = 22.1 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.6, 158.6, 156.2, 148.9, 148.9, 138.5, 131.0, 125.0, 120.2, 118.0, 117.0, 113.9, 38.3. ESI-MS(−) calculated for [C₁₃H₁₁NO₆P]⁻ m/z 308.03, found m/z 308.05 [M-H]⁻.

4-(2-Hydroxyphenyl)-6-(phosphonomethyl)picolinic acid (18g).

Yield: 76% (56 mg, 0.18 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.87 (s, 1H), 8.48 (s, 1H), 7.68 (d, J = 6.7 Hz, 1H), 7.45 (t, J = 7.1 Hz, 1H), 7.13 – 7.01 (m, 1H), 3.86 (d, J = 22.4 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.8, 155.5, 155.3, 147.8, 147.6, 130.9, 130.5, 127.2, 124.5, 123.1, 120.2, 116.8, 38.4. ESI-MS(−) calculated for [C₁₃H₁₁NO₆P]⁻ m/z 308.03, found m/z 308.09 [M-H]⁻.

4-(4-Aminophenyl)-6-(phosphonomethyl)picolinic acid (18h).

Yield: 56% (14 mg, 0.05 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.68 (s, 1H), 8.43 (s, 1H), 8.14 (d, J = 8.2 Hz, 2H), 7.56 (d, J = 8.2 Hz, 2H), 3.86 (d, J = 22.4 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.6, 155.5, 151.1, 149.2, 148.6, 128.2, 123.1, 123.0, 118.5, 114.6, 38.0. ESI-MS(−) calculated for [C₁₃H₁₂N₂O₅P]⁻ m/z 307.05, found m/z 307.09 [M-H]⁻.

4-(3-Aminophenyl)-6-(phosphonomethyl)picolinic acid (18i).

Yield: 96% (48 mg, 0.16 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.35 (s, 1H), 8.01 (s, 1H), 7.96 – 7.47 (m, 4H), 3.59 (d, J = 21.9 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.4, 156.3, 148.9, 148.2, 138.3, 137.9, 131.1, 125.1, 123.6, 122.4, 120.3, 119.4, 38.3. ESI-MS(−) calculated for [C₁₃H₁₂N₂O₅P]⁻ m/z 307.05, found m/z 307.09 [M-H]⁻.

4-(2-Aminophenyl)-6-(phosphonomethyl)picolinic acid (18j).

yield: 60% (60 mg, 0.19 mmol). ¹H NMR (400 MHz, MeOD-d₄): δ 8.24 (s, 1H), 7.87 (s, 1H), 7.54 – 7.32 (m, 2H), 7.25 – 7.05 (m, 2H), 3.57 (d, J = 21.7 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.7, 155.8, 155.7, 148.9, 148.7, 145.9, 130.3, 127.2, 122.6, 122.1, 117.3, 116.3, 38.3. ESI-MS(−) calculated for [C₁₃H₁₂N₂O₅P]⁻ m/z 307.05, found m/z 307.06 [M-H]⁻.

4-(4-Chlorophenyl)-6-(phosphonomethyl)picolinic acid (18k).

Yield: 97% (60 mg, 0.18 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.13 (s, 1H), 7.86 (s, 2H), 7.83 (d, J = 8.1 Hz, 3H), 7.62 (d, J = 8.1 Hz, 2H), 3.34 (d, J = 21.5 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.6, 156.3, 149.1, 147.5, 136.0, 135.0, 129.8, 129.2, 125.0, 120.2, 38.3. ESI-MS(−) calculated for [C₁₃H₁₀ClNO₅P]⁻ m/z 326.00, found m/z 326.01 [M-H]⁻.

4-(3-Chlorophenyl)-6-(phosphonomethyl)picolinic acid (18l).

Yield: 90% (47 mg, 0.14 mmol). ¹H NMR (400 MHz, DMSO-d₆): δ 8.14 (s, 1H), 7.95 – 7.35 (m, 5H), 3.34 (d, J = 21.5 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.6, 156.4, 149.1, 147.2, 139.4, 134.6, 131.7, 129.8, 127.2, 126.2, 125.2, 120.4, 38.4. ESI-MS(−) calculated for [C₁₃H₁₀ClNO₅P]⁻ m/z 326.00, found m/z 325.97 [M-H]⁻.

4-(2-Chlorophenyl)-6-(phosphonomethyl)picolinic acid (18m).

Yield: 72% (67 mg, 0.20 mmol). ¹H NMR (400 MHz, MeOD-d₄) δ 8.57 (s, 1H), 8.30 (s, 1H), 7.77 – 7.48 (m, 4H), 3.89 (d, J = 22.5 Hz, 2H). ¹³C NMR (126 MHz, DMSO-d₆): δ 166.5, 155.8, 148.2, 147.5, 137.0, 131.7, 131.4, 131.2, 130.7, 128.4, 127.9, 123.2, 38.3. ESI-MS(−) calculated for [C₁₃H₁₀ClNO₅P]⁻ m/z 326.00, found m/z 325.98 [M-H]⁻.

Determination of IC₅₀ values for MBL inhibition

MBL metalloforms NDM-1, IMP-1, and VIM-2 were over-expressed and purified as previously described.²⁸ The IC₅₀ values for which meropenem was used as a substrate is described. The decrease in absorption of meropenem at 300 nm (buffer: 50 mM HEPES, 2mM CHAPS, pH 7) was monitored in UV-transparent 96 well plates (Corning product #3635).⁴⁵ Briefly, 20 μL of each compound at various concentrations (final concentration 0 – 10 μM) was added to each well, followed by addition of 50 μL NDM-1 (final concentration of either 5 or 10 nM). The wells were incubated at 25 °C for 20 min. Next, 30 μL of meropenem (final concentration 180 μM) was added to each well to initiate the reaction. The plate was then centrifuged at 600 rpm for 30 s to eliminate air bubbles, and placed into a Synergy H4 plate reader (BioTek). The absorbance was monitored at 300 nm over 5 min with 15 s intervals. The ΔAbs_300nm/min was calculated from each slope as described below, fixing the uninhibited value at 100%, and solving for both the IC₅₀ and Hill coefficient values.²⁸

The IC₅₀ values for which Fluorocillin was used as a substrate were determined as described previously,²⁸ with minor modifications made to the volume of reagents used. Briefly, 20 μL of each compound at various concentrations (final concentration 0.001 – 10 μM) was added to each well, followed by addition of 50 μL enzyme (final concentration, NDM-1, 0.2 nM; VIM-2, 2 nM; IMP-1, 0.06 nM). Each plate was incubated for 20 min at 25 °C, followed by the addition of 30 μL Fluorocillin (final concentration 87 nM). The hydrolysis of Fluorocillin is monitored at λex/em of 495/525 nm. Plates were prepared in parallel to minimize any variability due to compound storage and handling. The rates of fluorescence increase are determined and IC₅₀ values determined as described above for the chromogenic assay.

Equilibrium dialysis

ZnZn-NDM-1 (final concentration 8 μM) in 5 mL of 100 mM ammonium acetate, pH 7.5, was mixed with the compounds 1 − 4 at concentrations of 0 − 128 μM. After incubation for 1 h, the solutions were dialyzed versus 500 mL of metal-free ammonium acetate, pH 7.5, overnight (dialysis tubing MWCO 6000−8000, Fisherbrand). The Zn(II) content in the resulting NDM-1 samples was determined using inductively coupled plasma with atomic emission spectroscopy (ICP-AES, Perkin Elmer Optima 7300DV). The emission wavelength was set to 213.856 nm, as previously described.²⁸

UV-Visible spectroscopy

To prepare Co(II)-substituted NDM-1, NDM-1 (150 μM) was dialyzed twice against 2 mM EDTA, 50 mM HEPES, pH 6.8, containing 150 mM NaCl, and 2 mM EDTA, followed by three separate dialysis steps against 50 mM HEPES, pH 6.8, containing 150 mM NaCl, and 0.5 g/L Chelex resin. Buffers were exchanged at approximately 12 h intervals. Metal-free NDM-1 was diluted to 300 μM with 50 mM HEPES, pH 6.8, containing 150 mM NaCI, 10% glycerol, and 2 mM TCEP (tris(2-carboxyethyl)phosphine). CoCl₂ (100 mM stock in water) was added to result in a protein with 2 molar equivalents of Co(II). The resulting CoCo-NDM-1 enzyme was separated into 500 μL aliquots, and captopril, DPA, and compounds 1 – 4 (stock solutions of 50 mM in DMSO) were added to result in samples with 2 molar equivalents of compound. The EDTA stock was dissolved in water. The samples were then incubated on ice for 5 min. The samples were added to a 500 μL quartz cuvette, and UV−vis spectra were collected on a PerkinElmer Lambda 750 UV/vis/NIR spectrometer measuring absorbance between 300 and 800 nm at 25 °C. A blank spectrum of apo-NDM-1 (300 μM) was used to generate difference spectra. All data were normalized at 800 nm.