Antibacterial Activity of Salicylanilide 4-(Trifluoromethyl)benzoates

The development of novel antimicrobial agents represents a timely research topic. Eighteen salicylanilide 4-(trifluoromethyl)benzoates were evaluated against Mycobacterium tuberculosis, M. avium and M. kansasii, eight bacterial strains including methicillin-resistant Staphylococcus aureus (MRSA) and for the inhibition of mycobacterial isocitrate lyase. Some compounds were further screened against drug-resistant M. tuberculosis and for their cytotoxicity. Minimum inhibitory concentrations (MICs) for all mycobacterial strains were within 0.5–32 μmol/L, with 4-chloro-2-[4-(trifluoromethyl)phenylcarbamoyl]phenyl 4-(trifluoromethyl)benzoate superiority. Gram-positive bacteria including MRSA were inhibited with MICs ≥ 0.49 μmol/L, while Gram-negative ones were much less susceptible. Salicylanilide 4-(trifluoromethyl)benzoates showed significant antibacterial properties, for many strains being comparable to standard drugs (isoniazid, benzylpenicillin) with no cross-resistance. All esters showed mild inhibition of mycobacterial isocitrate lyase and four compounds were comparable to 3-nitropropionic acid without a direct correlation between in vitro MICs and enzyme inhibition.


Introduction
The inappropriate use or misuse of established antimicrobial drugs has led to the increasing emergence of microbial resistance. The possible loss of the global control of infectious diseases represents a serious threat for public health, since resistance has been described for many important human pathogens.
Tuberculosis (TB) is responsible for millions of human deaths and about 1.4 million people died from TB in 2010 [1]. The drug-resistant TB forms, especially multidrug-resistant and extensively drugresistant tuberculosis (MDR-and XDR-TB), bring grave problems. MDR-TB consists in a resistance to at least isoniazid (INH) and rifampicin (RIF), the most effective first-line oral agents, while XDR-TB is MDR-TB in combination with both resistance to any fluoroquinolone and at least one secondline injectable drug (kanamycin, amikacin, capreomycin) [2]. Latent TB and also coincidence of TB with HIV-infection signify other principal impetuses for novel antimycobacterial drugs.
Contemporary research trends are focused, i.e., on the identification of unique mycobacterial pathways. Isocitrate lyase (ICL), an essential enzyme for the metabolism of fatty acids which splits isocitrate to succinate and glyoxylate, is one of two glyoxylate shunt enzymes. This pathway replenishes tricarboxylic acids cycle intermediates during growth on C 2 substrates and is lacking in vertebrates. Disruption of icl gene attenuated bacterial persistence and virulence without affecting the bacterial growth during the acute phase of infection. ICL might contribute to mycobacterial adaptation to hypoxia. Conventional anti-tuberculosis drugs exhibit mostly an insufficient activity against slowor non-growing (latent) mycobacteria. This is thought to be an important reason for the length of anti-tuberculosis therapy [3]. Novel ICL inhibitors may bring the shortening of TB treatment course and targeting of ICL represents one of the up-to-date research topics [4,5].
Nosocomial infections are a major challenge because of the high rates of morbidity and mortality. Worldwide, the state of resistance increased considerably in both Gram-positive and Gram-negative pathogens. Methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant Enterococcus spp., Klebsiella pneumoniae, Escherichia coli, Pseudomonas spp. or Acinetobacter baumannii producing extended spectrum β-lactamases (ESBL) represent such a most common and problematic drug-resistant bacteria, which have been rapidly spread throughout the world [6].
Consequently, novel drugs and strategies to control microbial spread are of critical importance, especially molecules with unique mechanism of action and no structural similarity. They can replace or combine with conventional antibiotics to combat problematic strains. The question if a new drug should be broad-spectral rather than have a narrow-spectrum of antimicrobial activity is controversial [7].
Salicylanilides derived from 4-(trifluoromethyl)aniline have exhibited predominantly a high antimycobacterial activity; in general, the introduction of CF 3 moiety into the salicylanilide scaffold enhanced antimicrobial properties [4,[9][10][11][12][13]. Salicylanilide benzoates revealed a notable activity especially toward MDR-TB [13] as well as antibacterial one [11]. Aromatic esters of 4-(trifluoromethyl)benzoic acid have been reported as a mild antibacterial agents against Pseudomonas fluorescens and Bacillus subtilis, since inactive for S. aureus and E. coli [21], and some salicylanilide 4-(trifluoromethyl)benzoates exhibited antifungal activity, especially against moulds with MICs  0.49 μmol/L [22]. Additionally, an increased lipophilicity may ameliorate passing through cell walls and biomembranes, that is why we selected highly lipophilic salicylanilide 4-(trifluoromethyl)benzoates for the searching of a potential antibacterial and antimycobacterial agents with expected lower MIC.

Results and Discussion
Salicylanilide 4-(trifluoromethyl)benzoates 1 were evaluated for their in vitro antimycobacterial and antibacterial activity, cytotoxicity and inhibition of mycobacterial isocitrate lyase.

Chemistry
The synthesis and characterization of the salicylanilide 4-(trifluoromethyl)benzoates (Scheme 1) were published recently [22]. Yield of esters synthesized via N,N'-dicyclohexylcarbodiimide coupling ranged from 49% up to 86%.  Table 1 summarizes the antimycobacterial activity of esters 1, isoniazid and para-aminosalicylic acid, two comparative drugs. All of the 4-(trifluoromethyl)benzoates displayed a significant antimycobacterial activity, with MICs of 0.  In general, the most active derivatives share a trifluoromethyl moiety (compounds 1o-r) or two chlorines (compounds 1e, 1f) on the aniline ring; on the other side, fluorine atom provided only a minimal benefit. Esters of 5-chlorosalicylanilides showed a higher antimycobacterial activity than those derived from 4-chlorosalicylanilides, with fluorinated molecules 1k-n being an exception.

Praha 1
Mtb. Importantly, all evaluated esters showed a significant activity towards MDR-and XDR-TB strains independently of the particular resistances. Interestingly, not only esters with the best activity towards M. tuberculosis (MICs ≤ 1 μmol/L) were assayed, as in previous studies [9,13,24], as even additional derivatives with MICs in the range of 2-4 μmol/L, were also powerful. All of these derivatives affected the growth of drug-resistant M. tuberculosis at equal or almost equal concentrations required for the inhibition of drug-sensitive one; drug-resistant strains were more susceptible to esters 1j, 1o and 1r than M. tuberculosis H 37 Rv. These findings indicate that salicylanilide 4-(trifluoromethyl)benzoates do not share any cross-resistance with conventionally used drugs (INH, rifamycines, EMB, STM, OFX, clofazimine, aminoglycosides) and they may be prospective and perspective agents for combating drug-resistant TB. 4-(Trifluoromethyl)benzoates 1 possess potent in vitro antimycobacterial properties; although sharing a high lipophilicity, their efficacy did not surpassed esters with N-acetyl-L-phenylalanine [9] with MDR-and XDR-TB being an exception, but esters 1 exhibited lower MICs than, e.g., salicylanilide benzenesulfonates [10] or esters with N-benzyloxycarbonyl amino acids [25]. Carbamates showed better in vitro activity against M. tuberculosis including drug-resistant strains, while atypical mycobacteria were inhibited at similar concentrations [24]. When compared to closely related salicylanilide benzoates [13], the introduction of trifluoromethyl moiety into a benzoic acid ring did not improved activity, although the presence of this substituent in parent salicylanilides resulted unambiguously in the excellent antimycobacterial properties.
The additional ICL inhibition assay was performed with the derivatives at the concentration of 100 µmol/L; nevertheless, some of the compounds precipitated in the medium after a very short period during the investigation. This effect at the higher concentration may be render due to a high lipophilicity.
Obviously, even though salicylanilide 4-(trifluoromethyl)benzoates 1 exhibited a significant ICL-1 inhibition, their concentration values are higher (IC 50 > 10 µmol/L) in comparison with MICs necessary for the mycobacterial growth arrest. Importantly, the inhibition of isocitrate lyase does not facilitate the elimination of mycobacteria during the acute phase of infection -in contrast to the chronic phase [3]. Therefore there is not a direct relationship of in vitro MICs of salicylanilides towards growing mycobacteria and the ICL inhibition, especially when the suppression of ICL may affect the slowly-or non-growing mycobacterial subpopulations existing on the nutrition poor substrates. Thus, the salicylanilide derivatives share a complex and still not fully elucidated mechanism of action for actively growing mycobacteria, since more cellular targets have been identified for this chemical group (e.g., [4,5,15]). Further experiments dealing with the mechanism(s) of action are appropriate. Table 4 overviews the antibacterial activity. For the remaining ten esters (1a, 1c, 1d, 1f, 1g, 1i, 1l,  1n, 1q, 1r), whose MICs are not reported, it was not possible to determine them. Similarly as in previously reported antifungal assay [22], they could not be dissolved in the testing medium well or they precipitated rapidly in them due to an escalated lipophilicity. 4-(Trifluoromethyl)benzoic acid showed no antibacterial action under our conditions (MIC > 500 µmol/L).

Antibacterial Evaluation
All soluble derivatives showed a very good activity against Gram-positive cocci, in some cases (1e,  1j, 1p, particularly 1b, 1h and 1k) even excellent MICs lower than 1 μmol/L, despite the methicillinresistance. Based on MIC values, these esters act likely mainly as bactericidal agents. MICs against S. aureus were comparable to benzylpenicillin and, moreover, 4-(trifluoromethyl)benzoates 1 exceeded its activity for MRSA, S. epidermidis and partly for Enterococcus (1e, 1j, 1p). Among Gram-negative bacteria, only E. coli possessed a partial susceptibility from 31.25 μmol/L (1p), while the growth of P. aeruginosa and both strains of K. pneumoniae was not reduced up to the concentration of 500 μmol/L.
Due to a limited number of assayed esters, it is difficult to discuss the structure-activity relationship. In the pair of 1o and 1p, the 5-chloro derivative (1p) showed lower MICs and when compared 1k vs. 1m, 3'-F moiety resulted in a higher activity.
IC 50 of 4-(trifluoromethyl)benzoates 1 were within the range 1.33-4.82 μmol/L and IC 50 of parent salicylanilides of 0.36-1.98 μmol/L. For all pairs salicylanilide and its 4-(trifluoromethyl)benzoate, the esterification alleviated mildly the strong cytotoxicity of the parent salicylanilides containing free phenolic group, which is probably mainly responsible for the cytotoxicity. Its temporary masking leads to the derivatives with somewhat higher IC 50 values -from 1.2-fold (1r) up to 3.7-fold for 1e and 1p, but still in similar concentration range. 4-(Trifluoromethyl)benzoic acid was many time less toxic with IC 50 of 563.7 µmol/L. Selectivity indexes (SI) ranges from 0.1 up to 1.67 for M. tuberculosis for parent salicylanilides, while esters with a significant activity against MDR-TB showed SI of 0.89-4.46 for drug-sensitive M. tuberculosis and of 0.78-4.46 for drug-resistant strains. SI values for S. aureus were within 0.17-9.84; SI of 1j almost reached the value of 10, which is generally considered being borderline. With respect to the purposed cytotoxicity reduction of the parent salicylanilides, the results are not satisfactory.  When compared to other salicylanilide esters, 4-(trifluoromethyl)benzoates showed about ten times higher cytotoxicity than salicylanilide carbamates [24], esters with N-acetyl-L-phenylalanine [9] or even more than esters with N-benzyloxycarbonyl amino acids [25], while it was approximately comparable with salicylanilide benzenesulfonates and benzoates [4].
However, salicylanilide esterification may represent one successful way to reduce their cytotoxicity for human cells as it was demonstrated especially for esters with N-protected -amino acids [9,25] and carbamates [24]; it is necessary to search new type(s) of acid(s), because 4-(trifluoromethyl)benzoic acid brought a very mild and still insufficient benefit.
The mechanism of cytotoxicity seems to be multiple -besides non-specific damage of cellular structures and function by the in vivo released phenolic group [8], it has been described a mild inhibition of human methionine aminopeptidase [4], inhibition of EGFR protein tyrosine kinases [17,19,20,28,29] or other human enzymes [18,19].

Antimycobacterial Evaluation
All compounds were tested for their in vitro against Mycobacterium tuberculosis 331/88 (H 37 Rv; dilution of the strain was 10

Isocitrate Lyase Inhibition Assay (ICL1)
The Mycobacterium tubeculosis H 37 Rv genomic DNA was used as a template. The icl gene (Rv0467) of 1.28 kb was amplified using PCR. The amplified DNA was cloned into the pET-28b(+) plasmid vector Novagen (Merck KGaA, Darmstadt, Germany) using NdeI and HindIII restriction sites. The recombinant plasmid was transferred into E. coli HB101. DNA sequencing was employed to confirm that the inserted coding sequence had no mutations. For bacterial expression, 25 mL culture volumes were inoculated with BL21(DE3) cells containing the recombinant plasmid and allowed to grow until an optical density of OD 595 = 0.6 was achieved. Then, the culture was induced with 1 mmol/L isopropyl-β-D-thiogalactopyranoside solution and incubated at 30 °C for additional 4 h. The cells were harvested at 4 °C by centrifugation at 6,000 g for 10 min. The resulting pellet was resuspended in BugBuster Protein Extraction Reagent Novagen (Merck KGaA, Darmstadt, Germany). The cell debris was removed by centrifugation, and the histidine-tagged protein was purified using an Äkta purifier (Amersham Biosciences, Valley Stream, NY, USA). The purity of the protein was confirmed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) followed by Coomassie blue staining of the gel. The protein concentration was determined by the Bradford method [30].
The isocitrate lyase activity was assayed according to the protocol reported by Dixon and Kornberg [31]. The enzyme assay was optimised in the final volume of 100 μL using 96-well plates (NUNC, Schoeller, Prague, Czech Republic) and the final concentration of tested compound in the reaction mixture 10 μmol/L. The reaction buffer contained 50 mmol/L of KH 2 PO 4 , 4 mmol/L of MgCl 2 ·6H 2 O, 4 mmol/L of phenylhydrazine hydrochloride, 12 mmol/L of L-cysteine, H 2 O and KOH to pH 7.0. The isocitrate cleavage was measured by the change in the absorbance at 324 nm, which is associated with the formation of glyoxylate phenyl hydrazone. Each tested compound was dissolved in DMSO to prepare a 1 mmol/L solution, and 1 μL of this solution was added to 93.9 μL of reaction buffer. Then, it was added 0.1124 μL of the recombinant enzyme in phosphate buffer and glycerol solution with concentration of 0.58 mg/mL (Bradford). Finally the reaction was started by the addition of 0.2 μmol of (+)-potassium D s -threo-isocitrate in solution.
Isoniazid was employed as a negative control (inhibition 0%), and 3-nitropropionic acid (3-NP) served as the positive control. All of the control compounds were added to the reaction mixture the same way like tested compounds and also their concentration in the reaction mixture was 10 μmol/L. The inhibitory activity of DMSO alone (1 μL) was subtracted from the activities of the evaluated compounds.
The microdilution broth method modified according to standard M07-A07 in Mueller-Hinton broth (HiMedia Laboratories, Mumbai, India) adjusted to pH 7.4 (±0.2) was used. The tested compounds were dissolved in DMSO to the final concentrations ranging from 500 to 0.49 mmol/L. Benzylpenicillin (penicillin G) was used as a comparative drug. Bacterial inoculum in sterile water was prepared to match 0.5 McFarland scale (1.5 ± 10 8 CFU/mL). The minimum inhibitory concentrations were assayed as 90% (IC 90 ) or higher reduction of growth when compared to the control. The determination of results was performed visually and spectrophotometrically (at 540 nm). The values of MICs were determined after 24 and 48 h of incubation in the darkness at 35 °C (±0.1) in a humid atmosphere.

Cytotoxicity Evaluation (HepG2 Cells)
Seven esters and their parent salicylanilides with the most robust antimicrobial activity were tested for cytotoxicity in the human hepatocellular liver carcinoma cell line HepG2 (passages 21-24; ECACC, Salisbury, UK) using a standard colorimetric method measuring a tetrazolium salt reduction (CellTiter(R) 96 AQueous One Solution Assay, Promega G3580, Madison, WI, USA).
The cells were routinely cultured in Minimum Essentials Eagle Medium (Sigma-Aldrich, Darmstadt, Germany) supplemented with 10% foetal bovine serum (PAA; Biotech, Prague, Czech Republic), 1% of L-glutamine solution (Sigma-Aldrich), and non-essential amino acids solution (Sigma-Aldrich) in a humidified atmosphere containing 5% CO 2 at 37 °C. For subculturing, the cells were harvested after trypsin/EDTA (Sigma-Aldrich) treatment at 37 °C. The cells treated with the tested substances were used as experimental groups. Untreated HepG2 cells were used as control groups.
The cells were seeded in density 1 × 10 4 cells per well in a 96-well plate with microscopic control. Next day, the cells were treated with each of the tested substances. Because of the low solubility of some investigated compounds in water they were dissolved in a very small amount of DMSO and in a small volume added to cell culture; a final solution contained less than 1% DMSO. The most of the tested compounds were prepared in incubation concentrations of 0.01-20 µM in triplicates. 4-(Trifluoromethyl)benzoic acid was applied in the concentrations of 10-4,000 µM in triplicates. The following types of controls were included: determination of 100% viability and 0% viability (the cells treated by 10% DMSO), no cell control and vehiculum controls. These checking samples were prepared in triplicates.
The treated cells were incubated together with controls at 37 °C for 24 h in 5% CO 2 atmosphere. After this time reagent from the kit CellTiter 96 AQueous One Solution Cell Proliferation Assay was added. The CellTiter 96 assay is based on the reduction of tetrazolium salt MTS [3-(4,5dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] to water-soluble formazan dye by metabolically active cells. The reduction of the reagent is attributed to availability of NADH or NADPH. The decline in levels of these metabolically important compounds in the cell causes that the production of formazan is reduced. The tested plate was incubated for 2 h at 37 °C and 5% CO 2 and after this time, the absorbance was recorded at 490 nm using a 96-well plate reader (TECAN, Infinite M200, Grödig, Austria).
The results were expressed as inhibitory concentration which is necessary to inhibit cell viability to 50% from the maximal (control) viability (IC 50 ). A standard toxicological parameter IC 50 was calculated in each of the tested substances using GraphPad Prism software (version 5.02; GraphPad Software Inc., San Diego, CA, USA).

Conclusions
Eighteen salicylanilide 4-(trifluoromethyl)benzoates with a known antifungal activity underwent additional biological screening for the antimycobacterial and antibacterial activity, cytotoxicity and inhibition of mycobacterial isocitrate lyase. All compounds showed a significant antimycobacterial activity including against drug-resistant strains; in some cases comparable or even better than isoniazid. 3,4-Dichloro-and trifluoromethyl moieties represent the optimal substitution of aniline ring. Esters caused consistently a mild inhibition of isocitrate lyase; moreover, some derivatives are comparable to 3-nitropropionic acid. The substitution of aniline ring at the position 3 enhances this action. Although salicylanilide 4-(trifluoromethyl)benzoates failed with respect to our intention of finding the most antimycobacterial active salicylanilide esters, they offer very low MICs against Gram-positive strains and IC 50 values for HepG2 cells.
In contrast to the aniline ring, the introduction of a 4-trifluoromethyl moiety into an acyl part of the salicylanilide esters did not provide sharply improved antimycobacterial activity, when compared to salicylanilide benzoates. The most potent esters are still considerably cytotoxic in micromolar values, although less than parent salicylanilides. Their cytotoxicity was mostly a slightly higher than for corresponding benzoates. In summary, salicylanilides and their 4-(trifluoromethyl)benzoates may be promising initial step compounds for anticancer research with advantageous broad-spectrum antimicrobial properties.