Synthesis and antibacterial activity of Schiff bases and amines derived from alkyl 2-(2-formyl-4-nitrophenoxy)alkanoates

A series of novel Schiff bases and secondary amines were obtained in good yields, as a result of the reductive amination of alkyl 2-(2-formyl-4-nitrophenoxy)alkanoates with both aniline and 4-methoxyaniline under established mild reaction conditions. Sodium triacetoxyborohydride as well as hydrogen in the presence of palladium on carbon were used as efficient reducing agents of the Schiff bases, in both direct and stepwise reductive amination processes. The Schiff bases, amines, and amine hydrochlorides were designed as potential antibacterial agents, and structure–activity relationship could be established following in vitro assays against Gram-positive and Gram-negative bacteria. The minimal inhibitory concentration and zone of inhibition were also determined. In these tests, some of Schiff bases and secondary amine hydrochlorides showed moderate-to-good activity against Gram-positive bacteria, including S. aureus, M. luteus, and S. mutans.


Introduction
The major problem in the effective antibacterial treatment is increasing resistance of microorganisms to currently available antimicrobial drugs. Therefore, the development of novel antimicrobial drugs is an active area of research. Most of compounds bearing an azomethine group exhibit antimicrobial (da Silva et al., 2011;Mohini et al., 2013;Shi et al., 2007), antioxidant, and antiproliferative properties (Cheng et al., 2010). Schiff bases, such as nitrofurantoin or nifuroxazide, are commonly applied in medicine as antibacterial agents (Sztanke et al., 2013). Additionally, a variety of phenoxyalkanoic acid derivatives are also known to possess a wide range of bioactivities (Hullar and Failla, 1969;Pattan et al., 2009;Kumar and Kumaresan, 2012), and some of the Schiff bases derived from 2-formylphenoxyacetic acids exhibit antibacterial properties (Bala et al., 2010;Iqbal et al., 2007). Further, aromatic secondary amines, as well as their salts, are also known to possess antimicrobial activity (Kitahara et al., 2004;Singh et al., 2011). Finally, secondary amines containing an aromatic nitro group exhibit an arginase inhibitory effect on vascular smooth muscle cell proliferation (Curtis et al., 2013).
A widely useful method for the synthesis of amines is reductive amination, which involves the reaction of aldehydes and ketones with ammonia or primary/secondary amines in the presence of a selective reducing agent (Tarasevich and Kozlov, 1999;Gomez et al., 2002). This process is considered direct when a carbonyl compound Electronic supplementary material The online version of this article (doi:10.1007/s00044-015-1397-6) contains supplementary material, which is available to authorized users. and an amine are mixed together with a reducing agent in a single operation. On the other hand, a stepwise reductive amination involves the pre-formation of the intermediate imine, followed by reduction in a separate step (Abdel-Magid et al., 1996). A wide variety of reducing agents have been utilized for reductive amination; however, two methods have been used most commonly. The first method involves catalytic hydrogenation with platinum, palladium, ruthenium, cobalt or nickel catalysts (Klyuev and Khidekel, 1980;Petrisko and Krupka, 2005;Tripathi et al., 2008). The second method utilizes metal hydride reagents, mainly sodium borohydride (Panfilov et al., 2000), sodium triacetoxyborohydride (Abdel-Magid and Mehrman, 2006;Gribble, 2006), sodium or lithium cyanoborohydride (Borch et al., 1971;Grenga et al., 2009), and sodium borohydride modified with numerous polyvalent metal salts (Saxena et al., 2000;Saidi et al., 2007;Neidigh, et al., 1998) or activated by acids (Cho and Kang, 2005;Alinezhad et al., 2010).
Here, we report the synthesis of a series of Schiff bases and amines that were designed as potential antimicrobial agents. The synthesis involves the chemoselective reaction of primary amines with alkyl 2-(2-formyl-4-nitrophenoxy)alkanoates yielding Schiff bases bearing intact the ester group, as well as further reduction of the Schiff bases to the corresponding amino esters. Further, we have performed antibacterial screening of the obtained compounds against Gram-positive and Gram-negative bacteria and analyzed the influence of the electron-donating substituents such as methoxyl and amino groups in the phenyl rings, as well as length of hydrophobic side chain on the antibacterial activities.

Results and discussion Chemistry
The desired alkyl 2-(2-formylphenoxy)alkanoates 1a-g were obtained in high yield by the condensation of adequately substituted 2-hydroxybenzaldehydes with alkyl 2-bromoalkanoates in the presence of potassium bicarbonate in dimethylformamide (Kwiecień, 2004). The Schiff bases 3a-l were prepared by the reaction of alkyl 2-(2-formylphenoxy)alkanoates 1a-g with aniline (2a) and 4-methoxyaniline (2b) (Scheme 1). To determine the optimal conditions of the process, a series of reactions of methyl 2-(2-formyl-4-nitrophenoxy)butanoate (1a) with aniline (2a) were carried out using different solvents, such as tetrahydrofuran, 1,2-dichloroethane, and methanol or without any solvent. The influence of the ratio of reagents, reaction time, and catalyst on the yield was also assessed, as indicated in Table 1. Based on these studies, we found that a nearly quantitative yield of the desired product was obtained when the reaction was carried out in methanol and in the presence of catalytic amount of the acetic acid, using equimolar ratio of the reactants (Entry 7, Table 1).
Under these established conditions, the reaction proceeded chemoselectively in the formyl group and leaving unchanged the ester group. Importantly, it was observed that the product 3a could be obtained in high yield by using an aprotic solvent, such as 1,2-dichloroethane in the presence of catalytic amounts of acetic acid. This is important for the direct reductive amination, which is more effective when it is carried out in 1,2-dichloroethane, as demonstrated later. The Schiff base was separated from the reaction mixture by dilution with water, followed by filtration of the precipitate.
Next, the Schiff bases 3b-l were readily prepared by reactions of formyl esters 1b-g with aniline (2a) and 4-methoxyaniline (2b), in the established reaction conditions (1:1 molar ratio of the formyl ester to amine, methanol as solvent, catalytic amount of acetic acid, room temperature, 3.5 h). The crude products (3b-l) were crystallized from methanol to yield stable crystals with high melting points (mp); yield: 72-95 %. Some of the Schiff bases exhibit a wide range of their melting points (see experimental section) that is probably caused by presence of (E-) and (Z-) diastereoisomers in the solid state of the products.
The structures of novel Schiff bases 3a-l were confirmed by gas chromatography mass spectrometry (GCMS), Fourier transform infrared spectroscopy (FTIR), and 1 H and 13 C nuclear magnetic resonance (NMR). The infrared spectra exhibited an intense absorption band in the range of 1623-1614 cm -1 , characteristic of the azomethine groups. Additionally, intense bands, originating from the valence vibrations of the ester carbonyl group, were observed in the range 1756-1733 cm -1 (C=O stretch) and 1208-1198 cm -1 (C-O stretch). Further, we observed a singlet of integration intensity equivalent to one hydrogen at 9.10-8.94 ppm in the 1 H NMR spectra of the Schiff bases, indicating the presence of the azomethine proton (-CH=N-).
Two methods were used for the reduction of Schiff bases. The first one was carried out using sodium triacetoxyborohydride (STAB) as a selective reducing agent of the imino group only and giving nitro amines 4af (Scheme 2). The second method consisted of catalytic reduction of both imino and nitro groups, resulting in the formation of diamine compounds 5a, c-f (Scheme 3).
To determine the optimal conditions for the reduction of Schiff bases with sodium triacetoxyborohydride to the nitro amines 4a-f, a series of reactions was carried out, starting from model compound 3a and using the following solvents: methanol, 1,2-dichloroethane, tetrahydrofuran, and N,Ndimethylformamide (Table 2). Utilizing methanol resulted in a low yield of amino ester, and neither extending the reaction time nor increasing the temperature improved the yield. Similarly, low yield was obtained using N,Ndimethylformamide in the presence of catalytic amounts of acetic acid. The best result was obtained using a 1:1.5 molar ratio of Schiff base/STAB in 1,2-dichloroethane and a catalytic amount of acetic acid. The reaction was carried out at room temperature for 4 h. Next, the reaction mixture was neutralized with 5 % aqueous solution of sodium bicarbonate, and the organic layer was separated and dried using magnesium sulfate. The product was isolated by solvent removal under reduced pressure and was recrystallized from methanol, to yield pure amino ester 4a. Under these optimal conditions, the synthesis of compounds 4bf was readily achieved with moderate-to-good yields (Scheme 2). All of the synthesized amino esters 4a-f are novel compounds. Their structures were established by spectroscopic methods: GC-MS, FTIR, 1 H, and 13 C NMR. The FTIR spectra exhibited an intense absorption band in the range of 3406-3385 cm -1 , characteristic of the amine groups. Additionally, intense bands, originating from the valence vibrations of the ester carbonyl group, were observed in the range 1751-1742 cm -1 (C=O stretch) and 1209-1200 cm -1 (C-O stretch). In the 1 H NMR spectra of the amino esters, the multiplet at 4.70-4.31 ppm is consist of two doublets and broad signal which are derived from two protons of CH 2 and a one proton of NH group, respectively. The deuterium exchange experiment with D 2 O was performed to confirm the presence of amine proton (see Supporting Information).
Reduction of both nitro and azomethine groups of 3a and 3c-f was carried out in mild conditions: in methanol with the addition of dimethoxyethane (DME), using a 1:0.1 weight ratio of Schiff base to catalyst, 10 % Pd/C (Scheme 3). The reaction was completed after 7 h, and the Scheme 3 Catalytic reduction of Schiff bases 5a and 5c-f products 5a and 5c-f were obtained after removing of the solvent under vacuum. Diamines 5a, c-f were obtained as brown semi-solids, in good yield (71-86 %, Table 3). The FTIR spectra of 5a, c-f exhibit an intensive absorption band in the range of 3410-3374 cm -1 , characteristic of the amine groups. Additionally, intensive bands, originating from the valence vibrations of the ester carbonyl group, were observed in the range 1735-1737 cm -1 . Finally, the signal at 9.10-8.94 ppm in the 1 H NMR spectra, associated with azomethine group, disappeared and instead signals at ranges 3.90-3.53 and 1.51-0.78 ppm were observed, indicating the presence of the amine protons.
Subsequently, reductive amination of methyl 2-(2formylphenoxy)alkanoate 1a-f was investigated as a onestep process. Synthesis of amines via direct reductive amination is very useful, because it does not require isolating the intermediate Schiff bases. This greatly speeds up the process of synthesis and limits losses associated with isolation of the intermediates.
To determine optimal conditions for the direct reductive amination, the model formyl ester 1a was reacted with aniline (2a) in the presence of sodium triacetoxyborohydride as the reducing agent. The reactions were conducted at ambient temperature, using different solvents, and changing the molar ratio of the reactants and reaction times ( Table 4).
The highest yield of the desired product was achieved when the reaction was performed in 1,2-dichloroethane with a catalytic amount of acetic acid for 4 h, with an equimolar ratio of formyl ester and aniline and 1.5 mol excess of the catalyst.
The direct reductive amination of 1b-f was carried out using the same conditions as for 1a, resulting in good yields (71-85 %) of amines 4b-f (Scheme 4). Finally, amino esters 4a-f were converted into their hydrochloride salts.

Microbiology
All of the synthesized compounds were screened for antibacterial activity against selected clinically important Gram-positive (S. aureus, M. luteus, S. mutans, E. faecalis) and Gram-negative (E. coli, P. aeruginosa, A. baumannii) bacteria by the disk diffusion method. Acetone was used as solvent for Schiff bases, and DMSO as solvent for amines and hydrochloride salts of amines. The antibiotic ciprofloxacin (5 mg/mL) was used as a positive control. The results of antibacterial screening indicate that four of eleven tested Schiff bases 3a, 3c-e exhibit varied activity against Gram-positive bacteria, including S. aureus, and S. mutans. The bacterial inhibition zone values of the Schiff bases are summarized in Table 5. Schiff base 3d exhibited antibacterial activity only against S. aureus strains and caused the strongest inhibition of the growth of methicillin-resistant S. aureus. Schiff base 3a, on the other hand, caused the strongest inhibition of the growth of S. mutans. It should also be mentioned that the synthesized Schiff bases did not inhibit the growth of E. faecalis and M. luteus. Further, introducing the methoxyl group into Schiff base structure has a significant impact on antibacterial activity. No antibacterial activity was observed for the Schiff bases containing methoxyl in N-substituted phenyl ring (3h-l, R=OMe). Likewise, the presence of methoxyl group at 2-position in 3d structure abolishes activity against S. mutans, in contrast to S. aureus, where activity increases against MRSA. Increasing alkyl, hydrophobic chain length results in decreased antibacterial activity, except for the influence of 3e on S. mutans.
Followed the screening, the minimum inhibitory concentration (MIC) was determined for the Schiff bases that demonstrated activity against specific species of bacteria. The results are presented in Table 6.
Next, it was observed that amino esters 4a-f did not show any antimicrobial activity, while their hydrochlorides showed good inhibition of the Gram-positive bacteria ( Table 7). The most active were those without the methoxyl groups, but no influence of hydrophobic side chain on the Scheme 4 Direct reductive amination of formyl esters 1af to 4a-f  antibacterial activity was observed. The lack of inhibition for amino esters might be caused by the presence of an intramolecular hydrogen bond that can be formed between amino proton and carbonyl group in the amino esters. The broadest spectrum of antibacterial activity was noted for 4cÁHCl. This compound inhibited the growth of all Gram-positive bacteria and was the only compound synthesized in these studies that inhibited the growth of the Gram-negative A. baumannii. The MIC of hydrochloride salts of 4a-f is given in Table 8.
Three other hydrochloride salts of amino esters, 4aÁHCl, 4bÁHCl, 4fÁHCl, inhibited the growth of all Gram-positive bacteria, having MICs of 0.50-1.00, 0.05-1.0, and 0.01-0.25 mg/mL, respectively, for each bacterial strain. Moderate antibacterial activity was also recorded for 4eÁHCl, which inhibited the growth of E. faecalis and M. luteus. The most encouraging results against Gram-positive bacteria were obtained for compounds 4cÁHCl and 4fÁHCl having MICs of 0.05-0.50 and 0.01-0.25 mg/mL, respectively.
The structure-activity relationships of the tested compounds can be summarized as follows: (1) in series of the Schiff bases, the presence of methoxyl group at the Nphenyl ring has a significant negative impact on their antibacterial activity; (2) in the same series, elongation of hydrophobic, alkyl chain causes the decrease in antibacterial activity against S. aureus MSSA and MRSA; (3) reduction of azomethine group to amine causes loss of activity against all of the tested microorganism; (4) in series of amine hydrochlorides, the presence of methoxyl group at the N-phenyl ring has the same negative impact on their antibacterial activity as in the case of Schiff bases; opposite effect is observed for elongation of hydrophobic, alkyl chain that causes the increase in antibacterial activity.
Lack of activity of the tested substances against Gramnegative bacteria could be explained by the differences in the structure of the cell walls of Gram-positive and Gramnegative microorganisms. In most Gram-positive bacteria, the cell wall consists of many layers of peptidoglycan, forming a thick, rigid structure. The cell walls of Gramnegative bacteria consist of one or a very few layers of peptidoglycan and a lipid-rich outer membrane (Beveridge, 1999). However, the mechanism responsible for the antibacterial activity of examined compounds is not known at the moment; work is in progress to clarify in detail the mechanism of antibacterial action, as well as the design of more effective compounds.

Conclusion
We demonstrate simple and efficient methods for both stepwise and direct reductive amination of 2-(2-formyl-4nitrophenoxy)alkanoic acid derivatives, yielding secondary N-arylated amines via Schiff bases under mild conditions. The reduction step was performed utilizing sodium triacetoxyborohydride, as well as catalytic hydrogenation using palladium(0) catalyst. Our antibacterial screening assay indicates that some of Schiff bases and secondary amine hydrochlorides possess moderate-to-good activity against Gram-positive bacteria, including S. aureus, M. luteus, and S. mutans. In a series of Schiff bases, we observe some of the influence of chain length and presence of methoxyl group on the antibacterial activity. Further modification of the selected compounds based on the information obtained from these results, as well as molecular modeling and structure-activity relationship studies, are in progress.
General procedure for the synthesis of methyl 2-(4amino-2-((phenylamino)methyl)phenoxy)alkanoate (5a and 5c-f) A mixture of methanol (10 mL) and 10 % Pd/C catalyst (0.033 g) was stirred magnetically under a slow stream of hydrogen (1 bubble per second) at room temperature for 30 min. Then, a solution of 3a or 3c-f (0.91 mmol) in methanol (10 mL) and DME (10 mL) was added, and reduction was carried out until the conversion of 3a or 3cf was completed (7 h), which was determined by gas chromatography. The mixture was left overnight, the catalyst was separated on the next day, the solvent was evaporated, and the residue was obtained to give 5a and 5c-f. added to each Petri plate and incubation was continued for 3 h at the same culture condition. In the last step, the areas of growth inhibition were measured for each tested compound and controls and the average diameters of the zone of inhibition (in mm) were calculated. Tests were performed in triplicate.