Cu-Catalyzed Chan-Evans-Lam Coupling reactions of 2-Nitroimidazole with Aryl boronic acids: An effort toward new bioactive agents against S. pneumoniae

Abstract

The coupling of phenylboronic acids with poorly-activated imidazoles is studied as a model system to explore the use of copper-catalyzed Chan-Evans-Lam (CEL) coupling for targeted C—N bond forming reactions. Optimized CEL reaction conditions are reported for four phenanthroline-based ligand systems, where the ligand 4,5-diazafluoren-9-one (dafo, L2) with 1 molar equivalent of potassium carbonate yielded the highest reactivity. The substrate 2-nitroimidazole (also known as azomycin) has documented antimicrobial activity against a range of microbes. Here N-arylation of 2-nitroimidazole with a range of aryl boronic acids has been successfully developed by copper(II)-catalyzed CEL reactions. Azomycin and a range of newly arylated azomycin derivatives were screened against S. pneumoniae, where 1-(4-(benzyloxy)phenyl)-2-nitro-1H-imidazole (3d) was demonstrated to have a minimal inhibition concentration value of 3.3 μg/mL.

Graphical Abstract

The coupling of phenylboronic acids with poorly-activated imidazoles is studied as a model system to explore the use of copper-catalyzed Chan-Evans-Lam (CEL) coupling for targeted C—N bond forming reactions. The N-arylated 2-nitroimidazoles (also known as azomycin) has documented antimicrobial activity against a range of microbes, here azomycin and arylated derivatives synthesized here are screened against S. pneumoniae.

Introduction

Imidazoles, nitroimidazoles, and benzimidazoles are important precursors for the generation of key heterocyclic aromatic compounds. They are also found in a wide-range of applications in coordination and materials chemistry.^[1–4] Five-membered heterocycles, like imidazole, are commonly observed in pharmaceuticals and natural products, which show antifungal, antitumor, antibacterial, and anti-inflammatory activity.^[5–7] More specifically, a plethora of pharmaceuticals contains N-phenylimidazole moieties.^[8,9] Several FDA approved therapeutics that contain this or related architectures are shown in Figure 1; they include clonidine (sedative and antihypertensive drug), alpidem (anxiolytic drug), zolimidine (gastroprotective drug), olprinone (cardiotonic agent), megazol-CL 64855 (protozoan infections), metronidazole (antibiotics-vaginal infections), ronidazole (antiprotozoal- histomoniasis and swine dysentery), satranidazole (antibiotic), benznidazole (antiparasitic-Chagas disease), fexinidazole (African trypanosomiasis-Trypanosoma brucei gambiense), misonidazole (radiosensitizer), and azathioprine (immunosuppressive drug-rheumatoid arthritis).^[10–17] One of the simplest and most efficient methods of synthesizing N-phenylimidazole is coupling functionalized arenes directly to imidazole or related derivatives, which is a cost-effective method of generating C—C and C—heteroatom bonds using first-row transition metals and metal complexes catalysts.^[18–22]

Figure 1.

Representative natural products and therapeutical molecules containing the imidazole heterocycles.

One method of coupling heteroatoms to arenes is Chan-Evans-Lam (CEL) coupling, which traditionally uses copper(II) salts as a catalyst under mild conditions.^[23–25] More recent literature has also shown a variety of bidentate ligands can be employed to improve yields as well.^[26] Similar to the Pd-catalyzed Suzuki coupling,^[27] the CEL coupling typically utilizes a phenylboronic acid and oxidatively couples it to an N—H or O—H functionality. Chan,^[23] Evans,^[24] and Lam^[25] were the first to independently develop this copper-catalyzed coupling reaction in the late 1990s. The CEL reaction provides an air-stable, alternative pathway to C—N bond formation compared to the Ullmann coupling reaction^[28] and the palladium-catalyzed Buchwald-Hartwig coupling.^[29,30] Similar to the Ullmann coupling, it was thought that the stability of the copper(I) species in solution is the main contributor to reactivity, but this is made difficult in part by the instability of copper(I) under typical atmospheric conditions.^[30] Copper(II)-catalyzed or copper(II)/cobalt(II) co-catalyzed C—N cross-coupling of 1H-pyrrole and phenylboronic acid have been reported by Batey,^[31] Ghanbari,^[32] Aberi,^[33] and Allahresani^[34] as shown in Scheme 1a.

Scheme 1.

CEL reactions from previously reported studies and our work herein.

Collman and co-workers have shown that some copper(II) catalysts require an aerobic atmosphere to form these C—N bonds,^[35] and while a myriad of different copper-catalyzed reactions have been employed to produce the N-arylated or O-arylated products. Arvidsson et al. reported the Cu(OAc)₂-catalyzed CEL coupling reaction of sulfonimidaimdes with aryl boronic acid to obtain N and Ń-arylated products.^[36] In recent years, many research groups, including Schaper,^[37,38] Emerson,^[39,40] and Park^[41,42] have reported the copper(II)-catalyzed C—N cross-coupling reactions between imidazole and aryl boronic acids (Scheme 1b). Using this scheme, the mechanism of CEL coupling remains a topic of ongoing research. Recently, we have reported several copper(II) complexes supported by phen and phen-derivatives shown in Scheme 1d, which supported CEL coupling to generate new C—N bonds.^[39] Suwinski et al. developed the method for the formation of N-aryl-C-nitroazoles using C-nitroazoles with aryl boronic acids. However, this method provides a single example of arylated-2-nitroimidazole with low yields of 40%.^[43] Most of the reported methods are plagued with harsh reaction conditions, poor selectivity, and poor yield. Therefore, the development of a new synthetic protocols is required, which can be applied to a broad range of nitroimidazoles. Here we report our utilization and optimization of the CEL reaction with these ligand systems toward an unactivated aromatic heterocycle like the imidazole ring in 2-nitroimidazole, which is known as the antimicrobial agent azomycin (in Scheme 1c). Azomycin is a prodrug, where it is thought to be reduced in vivo leading to covalent DNA modification and broad biological effects,^[42,44,45] where it has been used for the treatment of community-acquired pneumonia (CAP), skin and soft tissue infection, urethritis, pharyngitis, bacterial sinusitis, bacterial bronchitis, and acute otitis.^[42,46] Here we report our efforts to generate a series of arylated 2-nitroimidazole derivatives through the CEL reaction, where these new compounds are screened against Streptococcus pneumoniae as potential new antimicrobial framework.

Results and Discussion

Previously, we have reported and characterized four different, 1,10-phenathroline-derived ligand systems which supported some level of reactivity in CEL coupling.^[39] While our previous work focused on generating and characterizing our pre-catalyst for CEL coupling by coordinating two of our ligands to copper triflate, we have found that generating the catalyst in situ by charging a flask with a copper salt and molar equivalence of ligand does not have an effect on the observed reactivity of the system and provides a more simple route for catalysis. Here we use L1 - L4, along with a few other classical bidentate ligand systems to support copper(II) for the coupling of phenylboronic acid with imidazole (Table 1).

Table 1.

Optimization of CEL reactions using different supporting ligands and bases.

Entry	Ligand (equiv)	Base (equiv)	% Yielda
1	L1 (1)	-	<5
2	-	K2CO3 (1)	10
3	L1 (1)	K2CO3 (1)	>95
4	L1 (2)	K2CO3 (1)	52
5	L1 (3)	K2CO3 (1)	24
6	Bipy (1)	K2CO3 (1)	>95
7	L1:Bipy (1:1)	K2CO3 (1)	54
8	L1:Bipy (1:2)	K2CO3 (1)	50
9	EDA (1)	K2CO3 (1)	ND
10	L2 (1)	K2CO3 (1)	>95
11	L2 (1)	-	50
12	L3 (1)	K2CO3 (1)	>95
13	L3 (1)	-	<5
14b	L4 (1)	K2CO3 (1)	>95
15	L4 (1)	-	21
16	L2 (1)	NaOAc(1)	<5
17	L2 (1)	KOH(1)	<5
18	L2 (1)	Pyridine(1)	4
19	L2 (1)	TEA(1)	36

Reaction Conditions: 0.4 mmol phenylboronic acid, 0.2 mmol imidazole, 0.016 mmol copper(II) triflate, 0.2 mmol K₂CO₃, 1 mL MeOH.

^[a]Yield by GC-MS,

^[b]Ligand was coordinated to copper before reaction rather than in situ catalyst formation.

The CEL reaction is highly sensitive to the solvent, where previous studies determined that methanol is a preferred solvent for the CEL reaction and supports good yields.^[47] No product was observed in N,N´-dimethyl formamide, dimethyl sulfoxide, water, or acetonitrile.^[39] Next simple copper salts such as CuCl, CuBr, CuI, CuCl₂, CuBr₂ and Cu(OAc)₂ were surveyed for CEL reactivity. All trials resulted in poor or undetectable yields (data not shown). Cu(OTf)₂, however, supports moderate yields indicating the counter ion plays a role in catalysis. Initially we investigated the C—N cross-coupling of imidazole 1 with phenylboronic acid 2 under the influence of Cu(OTf)₂ and ligand (L1) in MeOH, giving the corresponding N-arylated product in low yield (Table 1, entry 1). We found that the addition of base (K₂CO₃) to Cu(OTf)₂ only slightly improves conversion (Table 1, entry 2). The CEL coupling process was optimized using varying amounts of ligand (L1-L4, bipy) in different molar equivalents compared to Cu. There is an inverse relationship between L1 equivalents and product yield, where the highest yield was observed with a 1:1 ration of L1 to Cu (Table 1, entries 3–5). Notably, the use of greater amounts of bipy did not show a reduction in yield (c.f. Table S1). Interestingly, by adding varying ratios of L1:bipy, the reaction yields were also impacted. These competing ligand systems followed the same inverse relationship, where increased molar equivalents of L1 resulted in a reduction of product formation. Conversely increasing the amount of bipy seems to increase yields (Table 1, entries 6–7 and Table S1). Trials including ethylenediamine (EDA) showed no catalytic activity (Table 1, entry 9). Notably, L2 showed extremely high yields regardless of the amount of ligand present. 1:1, 1:2, and 1:3 ratios of copper(II):L2 gave high yields (Table S1, entries 17–19). L2 supported moderate yield even without added base (Table 1, entry 10–11). L3 and L4 followed a similar trend as L1, where higher molar amounts reduced yields and some base was required to achieve good turnover (Table 1, entry 12–15). Other bases were screened, including NaOAc, KOH, pyridine, and triethylamine (TEA), but reaction yields of 5%, 5%, 4%, and 36%, respectively were measured (Table 1, entry 16–19). An interpretation of this data is that at high ligand concentrations shifts the complex ion equilibria toward copper complexation, where coordinatively saturated copper(II) complexes are unlikely to efficiently support catalysis. Based upon these trials, the dafo ligand (L2) performed the best. While the other bidentate ligands support good yields under very specific conditions, L2 performed well regardless of the amount added to the reaction and functioned moderately when no base was added. L2 has a slightly larger bite angle compared to the other phen-based ligands which was originally identified by Stahl and co-workers.^[48] We believe that the perturbed binding angle leads to increased lability allowing copper-complexes of dafo to support the formation of reactive or intermediate species more readily than other phen derivatives. One equivalent of L2 with one equivalent of K₂CO₃ supported the best yields in Table 1. The copper(II)/phen complex equilibria have been studied previously,^[49] where approximately 90% of the copper(II) in solution will be in a [Cu(phen)]²⁺ form using a 1:1 phen/Cu²⁺ solution. This fact is consistent with notion that L2 is highly labile, where [Cu(L2)]²⁺ is likely the most active catalytic species in solution in this study.

With the optimized reaction conditions explored, we examined the scope and generality of copper(II)-catalyzed C—N cross-coupling product of 2-nitroimidazole (1) and a range of other nitrogen containing heterocycles with the aryl boronic acids 2a-2l (Scheme 3). Initial concerns were raised that the nitro substitution on the imidazole ring would deactivate 1 toward CEL reactions. However, the C—N cross-coupling of 1 and aryl boronic acids with para-substituted electron-donating groups (EDG’s), such as methyl (Me), methoxy (OMe), and benzyloxy (OBn) produced N-arylated-2-nitroimidazoles good yields (85–77%, 3b-3d). Further, the substrates with electron-withdrawing groups (EWG’s) such as fluoro (F), chloro (Cl), bromo (Br), cyano (CN), and trifluoromethoxy (OCF₃) groups were well tolerated in the C—N cross-coupling reaction, which offered the desired products 3e-3i in good to moderate yields (49–84%). Next, the coupling of 1-napthyl boronic acid (2j) and 3,4-dimethoxyphenylboronic acid (2k) with 1 afforded the desired products in good yields (82%−69%, 3j-3k), whereas benzo[b]thiophene-2-boronic acid (2l) and benzo[b]furan-2-boronic acid (2m) showed no C–N coupled product formation in the reaction. Unfortunately, boronic acid 2l and 2m did not produce the targeted product, where the self-coupling product was prominently observed. In short, we have efficiently synthesized a range of C—N cross-coupling products of 1 in air, using aryl boronic acid, base, and a copper(II) catalyst.

Scheme 3.

Substrate Scope study of C-N coupling products. Reaction Conditions: 0.4 mmol phenylboronic acid, 0.2 mmol imidazole, 0.016 mmol [Cu(L2)₂](OTf)₂, 0.2 mmol K₂CO₃, 1 mL MeOH.

To further demonstrate the scope of the N-arylated substrates, various N-nucleophiles were applied as cross-coupling partners in the reaction with phenylboronic acid (Scheme 3). The cross-coupling of 9H-carbazole (1b) with phenylboronic acid 2a gave the N-arylated product in moderated yields (3n, 45%). When benzimidazole 1c, and 2-phenylbenzimidazole 1d were employed as N-nucleophilic substrate, the yields of the N-arylated product were produced 3o–3p in good yields, 85% and 72%, respectively. Unfortunately, we failed to obtain the N-arylated products of 3q–3s when phthalimide 1e, phenothiazine 1f and benzotriazole 1g were involved to our optimized conditions. Based on the observations and the above experiment results, we concluded that Cu(OTf)₂ with ligand (L2) is a key role in the C—N cross-coupling reaction to obtain the N-arylated products.

Moving beyond the N-arylation, these 2-nitroimidazole derivatives produced in this study can be easily reduced to generate the corresponding 2-amino derivatives for further synthetic procedures. For example, 2-nitro-1-phenyl-1H-imidazole (3a) was transformed to 2-amino-1-phenyl-1H-imidazole by reduction using zinc-acetic acid in THF, yielding 4 (56%) as shown in Scheme 4. The 2-amino functionality provides a second site for potential derivatization such as further arylation or alkylation. The 2-aminoimidazole moiety has been found in the framework for potential antimalarials,^[50] biofilms,^[51] and Na⁺/K⁺ ATPase inhibitors.^[52]

Scheme 4.

Functionalization of 3a.

Although 1 has shown antimicrobial activity toward a range of organisms, to our knowledge it has not been applied toward Streptococcus pneumoniae growth.^[53,54] S. pneumoniae is a common opportunistic pathogenic microbe that is the leading cause of community acquired bacterial pneumonia. Bacterial pneumonia unequally targets people with compromised immune systems including children and the elderly. Pneumonia leads to an estimated $10B per year in medical costs,^[55] and even more when linked to other diseases like COVID 19 and those targeting the cardiovascular system.^[56] Additionally, multidrug resistant strains of S. pneumoniae are becoming alarmingly prevalent, serotype replacement is undermining effectiveness of current capsule-based vaccines, and new treatments are required for pneumococcal infections. Here we report an initial effort to screen 1 and its arylated derivatives generated in this study (3a-3k) against S. pneumoniae to showcase how CEL cross-coupling can be used to generate libraries of potential antimicrobial agents.

At 500 μM, all compounds (1, 3a-3k) showed an impact on growth of S. pneumoniae cultures. All cultures observed showed an increase rate of autolysis limiting the length of the stationary phase (c.f. Figure S22.) For example, 3a had no impact on the culture through the exponential phase, but then showcases a truncated stationary phase, indicating limited viability of this culture. Other compounds like 3i showed both a retardation of the exponential phase and a rapid decline in cell population. Zero growth was observed for cultures with 3d or 3j at 500 μM. At 100 μM, most of these complexes showed little impact on the viability of S. pneumoniae cultures. Azomycin (1) and 3j impacted the initiation of the exponential phase of S. pneumoniae compared to the control and showed subtle effects to the stability of the stationary phase of these cultures. Compound 3d, however, showed no growth over a 20-hour period. Further efforts were made to measure a MIC value for 3d, which is determined to be 3.3 μg/mL for S. pneumoniae over a 10-hour period. Although this MIC value indicates that 3d is several orders of magnitude more effective than azomycin, it will require continued improvement to become a competitive antimicrobial agent against S. pneumoniae.^[57] The specific activity and biological target of these agents toward S. pneumoniae remain under investigation by our research team. However, this study demonstrates the power of C—N cross-coupling in therapeutic development and represents a good step forward in developing new templates for the development of antimicrobial agents targeting S. pneumoniae and possibility related bacteria.

Conclusion

Copper-based catalytic systems exhibit high levels of reactivity for the CEL coupling of unactivated imidazoles with phenylboronic acids under ambient atmosphere. While the 4,5-diazafluoren-9-one (L2) clearly exhibits the highest level of reactivity at a wider range of molar equivalents, we have shown that any of the ligand systems described, except EDA, will support high yields of CEL coupled products if 1 molar equivalent is used with respect to copper(II). Additionally, 1.0–1.25 molar equivalents potassium carbonate with respect to the imidazole substrate, is also needed to support high reactivity. The small library of arylated azomycins were screened against cultures of S. pneumoniae, where 3d showed promising activity (MIC ~ 3.3 μg/mL) as a starting framework for the development of alternative therapeutics.

Experimental Section

General procedure for copper(II) catalyzed CEL coupling.

A 20 mL vial was charged with a magnetic stir bar, 0.4 mmol of phenylboronic acid, 0.2 mmol of 2-nitroimidazole, 0.1 mmol of potassium carbonate (1 eq.), 0.016 mmol of copper(II) triflate (8 mol%) and the desired molar equivalence of ligand. This mixture was solvated in 1 mL of methanol and the reaction was left to stir for approximately 24 hours exposed to ambient atmosphere at room temperature in a capped vial. Upon completion the reaction was filtered through celite. The volume of this sample was reduced and then reconstituted in dichloromethane (DCM) and applied to a flash thin-layer column chromatography. Pure C—N coupled product was eluted with a different combination of hexanes/ethyl acetate mobile phase, which was evaporated to yield pure product. Characterization data for all purified complexes is available in the supporting information.

General procedure for S. pneumoniae growth studies.

S. pneumoniae serotype 4 strain TIGR4 were grown liquid culture in C+Y medium. Growth curves were performed by diluting bacteria to final concentration of 1 × 105 CFU/mL into various media conditions, 200 μL were inoculated into 96-well-plates in triplicate. Plates were then read at OD₆₀₀ every 30 min for 24 hrs. All growth curves were performed at twice and figures display a representative curve for each strain. Inhibitors were added to media using concentrated DMSO solutions of 1 and 3a-k to generate media with 100 to 500 μM inhibitor. Minimal inhibition concentrations values that limit growth of S. pneumoniae over an overnight growth (10 hrs) was measured using various concentrations of 3d. Further characterization of the cytotoxicity toward higher organisms was not investigated.

Figure 2.

Representative growth curves of S. pneumoniae TIGR4 challenged by different N-arylated complexes in the absence of antimicrobial agent (black circles), and in the presence of 100 μM 1 (open circles), 3a (red squares), 3d (blue triangles), 3i (green squares), and 3j (pink X symbols).

Scheme 2.

Substrate Scope study of C—N coupling products. Reaction Conditions: a 0.4 mmol phenylboronic acid, 0.2 mmol imidazole, 0.016 mmol [Cu(L2)₂](OTf)₂, 0.2 mmol K₂CO₃, 1 mL MeOH. b 8.0 mmol phenylboronic acid, 4.0 mmol imidazole, 0.32 mmol [Cu(L2)₂](OTf)₂, 4.0 mmol K₂CO₃, 20.0 mL MeOH.