Synthesis, Antifungal Evaluation and In Silico Study of N-(4-Halobenzyl)amides

A collection of 32 structurally related N-(4-halobenzyl)amides were synthesized from cinnamic and benzoic acids through coupling reactions with 4-halobenzylamines, using (benzotriazol-1-yloxy)tris(dimethylamino)phosphonium hexafluorophosphate (BOP) as a coupling agent. The compounds were identified by spectroscopic methods such as infrared, 1H- and 13C- Nuclear Magnetic Resonance (NMR) and high-resolution mass spectrometry. The compounds were then submitted to antimicrobial tests by the minimum inhibitory concentration method (MIC) and nystatin was used as a control in the antifungal assays. The purpose of the tests was to evaluate the influence of structural changes in the cinnamic and benzoic acid substructures on the inhibitory activity against strains of Candida albicans, Candida tropicalis, and Candida krusei. A quantitative structure-activity relationship (QSAR) study with KNIME v. 3.1.0 and Volsurf v. 1.0.7 softwares were realized, showing that descriptors DRDRDR, DRDRAC, L4LgS, IW4 and DD2 influence the antifungal activity of the haloamides. In general, 10 benzamides revealed fungal sensitivity, especially a vanillic amide which enjoyed the lowest MIC. The results demonstrate that a hydroxyl group in the para position, and a methoxyl at the meta position enhance antifungal activity for the amide skeletal structure. In addition, the double bond as a spacer group appears to be important for the activity of amide structures.


Introduction
The genus Candida includes over 200 species of human pathogens. Among the most important of these are Candida albicans, Candida tropicalis and Candida krusei. Changes in host defense mechanisms, invasive medical procedures, and anatomical barrier failures (through burns) are all factors that favor infection with these micro-organisms [1]. However, there are compounds in Nature that can inhibit such invasions. Benzoic and cinnamic acid derivatives exhibit pharmacological versatility with antitumor, anti-inflammatory, anti-microbial and immunostimulatory activities [2,3]. In the literature, cinnamic amides inhibit the growth of fungi such as Phytophthora infestans [4], Aspergillus niger, Candida albicans, and bacteria such as Escherichia coli, Bacillus subtilis and Staphylococcus aureus [5]. Cinnamic amides with cinnamic, caffeic, ferulic, sinapic, p-coumaric and 3,4,5-trimethoxycinnamic cores display proven antimicrobial activity [5]. Benzoic acid acts as an unspecific antimicrobial, derivatives such as protocatequetic, gentisic, vanillic, and p-hydroxybenzoic acids exhibit antimicrobial properties against various bacterial and fungal strains [7].
The literature reports that halogenated aromatic rings in cinnamic amides potentiate the biological activity, such as EGFR kinase inhibition [8], pesticidal effect [9,10] and microbial growth inhibition [11]. Among the halogenated amides studied, chlorinated cinnamic analogues showed higher microbial inhibition in bacterial and fungal strains than their fluorinated and brominated analogs. Furthermore, certain studies show that salicylanilides with ortho and para hydroxyls, and meta methoxyl on the aromatic ring all display increased microbial inhibitory activity [12]. Some studies also show the importance of substituents in the aromatic ring such as nitro groups, methyls and sterically bulky groups [13]. In this present study, we prepared a collection of structurally related cinnamic and benzoic acid amides with halogenated substituents in the para position, illustrated in Figure 1, in a coupling reaction using benzotriazol-1-yloxy-tris(dimethylamino) phosphonium hexafluorophosphate (BOP) as the coupling agent [14,15]. It is expected that the study will provide more information about structure-activity relationships (SAR) in this group of amides. Chemical parameters, such as lipophilicity, electronic effects, and hydrogen bonds caused by the substituents R, including the presence of the spacer (n = 1) between the carbonyl group and the aromatic ring, were analyzed in the SAR.

Antimicrobial Activity
In the antifungal activity study, the N-(4-halobenzyl)amides 1-32 were evaluated against Candida strains. The technique used was the broth microdilution method according to published protocols [27,28] using seven strains of Candida: C. albicans ATCC 76645, C. albicans LM-106; LM-23 C. albicans, C. tropicalis ATCC 13803, C. tropicalis, LM-36, LM-13 C. krusei, and C. krusei LM-656. The control medium result showed no fungal growth, while growth of fungi in the medium without any added drug (sterile control) was detected.

Antimicrobial Activity
In the antifungal activity study, the N-(4-halobenzyl)amides 1-32 were evaluated against Candida strains. The technique used was the broth microdilution method according to published protocols [27,28] using seven strains of Candida: C. albicans ATCC 76645, C. albicans LM-106; LM-23 C. albicans, C. tropicalis ATCC 13803, C. tropicalis, LM-36, LM-13 C. krusei, and C. krusei LM-656. The control medium result showed no fungal growth, while growth of fungi in the medium without any added drug (sterile control) was detected. Scheme 1. General procedure for synthesis of halogenated amides.

Scheme 2.
General procedure for preparation of esters and ethers 23-32 derived from vanillic amide (15). R = substituents in reactions of esters and ethers.

Scheme 2.
General procedure for preparation of esters and ethers 23-32 derived from vanillic amide (15). R = substituents in reactions of esters and ethers.

Scheme 2.
General procedure for preparation of esters and ethers 23-32 derived from vanillic amide (15). R = substituents in reactions of esters and ethers.       The results are shown in Table 3; the minimum inhibitory concentrations (MICs) of the compounds were significantly different, ranging from 256 to more than 1024 µg·mL −1 . The antimicrobial activity of the products was interpreted and considered active or not, according to the following criteria: 50-500 µg·mL −1 = strong/optimum activity; 600-1500 µg·mL −1 = moderate activity; Above 1500 µg·mL −1 = weak activity or inactive product [29,30]. The results of the control culture medium show no microbial growth occurred while control was positive in the yeast viability.

Qualitative-Structure Activity Relantioship (QSAR)
Three-dimensional structures (3D) of amides were used as input data to generate 128 descriptors together with the dependent variable (binary classification), which describes the compound as active (A) or inactive (I). The data were used as input to the KNIME v. 3.1.0 software [31]. Importantly, the generation of descriptors is relatively fast for all 32 amides which the training data sets comprised generating all 128 descriptors for Volsurf +, it took less than 1 min, using a computer equipped with an i7, running at 3.4 GHz and equipped with 12 GB of RAM.
The KNIME software inserts the data of active and inactive compounds in mathematical algorithms that try to find the descriptors that explains the influence of the structure on microbial activity. A match is given when the software can separate the truly active and truly inactive compounds. Table 4 summarizes the statistical indices of the match model for training and cross-validation in all antifungal tests. For training set the decision tree generated high rates of correct answers for inactive compounds, up to 83.3%, and lower rates for the active compounds, 22.2%. For cross-validation, the model performed similar to the training set. The specificity (true negative) was greater than the sensitivity (true positive). Overall this means that, there was a lower false positive percentage, if it was compared with true positive prediction, which shows that the method is suitable only to screen active compounds and detect physico-chemical properties of inactive compounds. Table 4. Summary of training, internal cross-validation, test results and corresponding match results which were obtained using a leave-one out validation method in KNIME of the total set of 32 haloamides subjected to antifungal tests. According to the decision tree (an organization chart that compares the descriptors, to predict the biological activity), the prediction of actives and inactives was based on two Volsurf descriptors [32] of N-(4-halobenzyl)amides: DRDRDR and DRDRAC, which are pharmacophore descriptors forming the maximum triangular area DRY-DRY-DRY (three hydrophobic regions), and DRY-DRY-Acceptor (two hydrophobic regions and one H-bond acceptor region).

C. albicans
It was possible to establish regions for analogs that produced fungal growth sensitivity. Figure 2 shows the molecular interaction of the grid fields (Molecule Interaction Fields-MIF) around the most active compounds 14 and 15, the weakly active compound 17 and inactive compound 12. The DRY probe is shown in all active and weakly active structures. Looking at the N1 probe (dark blue), regions in the aromatic ring which show an outstanding hydrogen accepting character could be observed. This N1 probe is seen more present near the gallic (14), vanillic (15), and 4-hydroxybenzoic (17) amide hydroxyls. This study did not have a high accuracy rate for matches, decreasing the specificity of software to find descriptors related to biological activity of the evaluated amides.  (14), vanillic amide (15), 4-hydroxybenzoic amide (17) and benzoic amide (12) interaction fields: energy level −0.6 kcal/mol DRY, and −3 kcal/mol N1 in VolSurf.

General Information
Purification of the compounds was performed by column chromatography on silica gel 60 (ART. 7734 Merck, Saint Louis, MO, USA) using a Hex:EtOAc solvent gradient and confirmed by analytical thin layer chromatography on silica gel 60 F254, using ultraviolet light at two wavelengths (254 and 366 nm) from a Mineralight apparatus (UVP, Upland, CA, USA) or H2SO4 in 5% ethanol for detection. FTIR spectra were recorded in a Prestige-21 FTIR spectrometer (Shimadzu, Kyoto, Japan) using KBr pellets. 1 H-and 13 C-NMR spectra were obtained on a MERCURY machine (200 and 50 MHz for 1 H and 13 C, respectively. Varian (Palo Alto, CA, USA) in deuterated solvents (CDCl3, MeOD or DMSO-d6) and tetramethylsilane (TMS) was used for the internal standard. Chemical shifts were measured in parts per million (ppm) and coupling constants (J) in Hz. Measurements of atomic mass for the compounds was carried out using an Ultraflex II TOF/TOF mass spectrometer (Bruker Daltonik GmbH, Bremen, Germany) equipped with a high performance solid state laser (λ = 355 nm), and reflector. The system was operated by the Bruker Daltonik FlexControl 2.4 software package (Bruker, Bremen, Germany). Infrared spectrum, 1 H-NMR spectrum, Expansion of the spectrum of 1 H-NMR, 13 C-APT NMR spectrum, expansion of the 13

General Preparation of N-(4-Halobenzyl)amides
Procedure 1: In a 100 mL flask equipped with magnetic stirring, the organic acid (1.35 mmol, 200 mg) was dissolved in dimethylformamide (DMF, 2.7 mL) and trimethylamine (0.14 mL, 1.35 mmol). The solution was cooled in an ice bath (0 °C). Then, 4-chlorobenzylamine (1.35 mmol) was added. Soon after a 1.35 mmol solution of BOP in CH2Cl2 (10 mL) was added to the flask. The reaction was stirred at 0 °C for 30 min, and then for an additional period, at room temperature for 2 h. After the reaction, the CH2Cl2 was removed under reduced pressure and the solution was poured into a separatory funnel containing water (10 mL) and EtOAc (10 mL). The product was extracted with EtOAc (3 × 10 mL). The organic phase was washed sequentially with 1 N HCl, water, 1 M NaHCO3 and water (10 mL of each); dried with Na2SO4, filtered and concentrated in a rotavapor. The amide was purified by gel chromatography on a silica gel column using as the mobile phase an EtOAc:Hex mixture gradient of increasing polarity [11]. The following compounds were prepared by this procedure:   (14), vanillic amide (15), 4-hydroxybenzoic amide (17) and benzoic amide (12) interaction fields: energy level −0.6 kcal/mol DRY, and −3 kcal/mol N1 in VolSurf.

General Information
Purification of the compounds was performed by column chromatography on silica gel 60 (ART. 7734 Merck, Saint Louis, MO, USA) using a Hex:EtOAc solvent gradient and confirmed by analytical thin layer chromatography on silica gel 60 F 254 , using ultraviolet light at two wavelengths (254 and 366 nm) from a Mineralight apparatus (UVP, Upland, CA, USA) or H 2 SO 4 in 5% ethanol for detection. FTIR spectra were recorded in a Prestige-21 FTIR spectrometer (Shimadzu, Kyoto, Japan) using KBr pellets. 1 H-and 13 C-NMR spectra were obtained on a MERCURY machine (200 and 50 MHz for 1 H and 13 C, respectively. Varian (Palo Alto, CA, USA) in deuterated solvents (CDCl 3 , MeOD or DMSO-d 6 ) and tetramethylsilane (TMS) was used for the internal standard. Chemical shifts were measured in parts per million (ppm) and coupling constants (J) in Hz. Measurements of atomic mass for the compounds was carried out using an Ultraflex II TOF/TOF mass spectrometer (Bruker Daltonik GmbH, Bremen, Germany) equipped with a high performance solid state laser (λ = 355 nm), and reflector. The system was operated by the Bruker Daltonik FlexControl 2.4 software package (Bruker, Bremen, Germany). Infrared spectrum, 1 H-NMR spectrum, Expansion of the spectrum of 1 H-NMR, 13 C-APT NMR spectrum, expansion of the 13

General Preparation of N-(4-Halobenzyl)amides
Procedure 1: In a 100 mL flask equipped with magnetic stirring, the organic acid (1.35 mmol, 200 mg) was dissolved in dimethylformamide (DMF, 2.7 mL) and trimethylamine (0.14 mL, 1.35 mmol). The solution was cooled in an ice bath (0 • C). Then, 4-chlorobenzylamine (1.35 mmol) was added. Soon after a 1.35 mmol solution of BOP in CH 2 Cl 2 (10 mL) was added to the flask. The reaction was stirred at 0 • C for 30 min, and then for an additional period, at room temperature for 2 h. After the reaction, the CH 2 Cl 2 was removed under reduced pressure and the solution was poured into a separatory funnel containing water (10 mL) and EtOAc (10 mL). The product was extracted with EtOAc (3 × 10 mL). The organic phase was washed sequentially with 1 N HCl, water, 1 M NaHCO 3 and water (10 mL of each); dried with Na 2 SO 4 , filtered and concentrated in a rotavapor. The amide was purified by gel chromatography on a silica gel column using as the mobile phase an EtOAc:Hex mixture gradient of increasing polarity [11]. The following compounds were prepared by this procedure:  (27). For the acetylation of 15, to a 50 mL flask, equipped with magnetic stirrer was added the chlorinated vanillic amide 15 (0.1000 g, 0.034 mmol), pyridine (0.13 mL, 0.15924 mmol) and acetic anhydride (0.08 mL, 0.8111 mmol). The reaction mixture was subjected to constant magnetic stirring for 24 h. The first step of the reaction product extraction was then carried out by pouring into ice water (30 mL) in a separation funnel using ethyl acetate as extractor solvent (3 × 10 mL). The organic phase was treated with saturated copper sulfate solution (3 × 20 mL). The ethyl acetate phase was washed with water (3 × 30 mL) and dried with anhydrous sodium sulfate (Na 2 SO 4 ). Subsequently, the organic phase was filtered and concentrated by rotary evaporation [22]. The product was purified by silica gel column chromatography using EtOAc:Hex (65:35) as the mobile phase system to give a crystalline solid; yield: 98% (159 mg), m.p. lowest concentration capable of inhibiting fungal growth in the wells as visually observed compared with the control. All tests were performed in duplicate and the results were expressed as a geometric mean of the MIC values obtained in both tests [34].
3.4. Qualitative-Structure Activity Relationship of Antimicrobial Activity in Amides

Volsurf Descriptors
The molecular structures in three dimensions (3D) were used as input data to the Volsurf + v program. 1.0.7, and subjected to molecular interaction fields (MIC) to generate descriptors using the following probes: N1 (amide hydrogen-nitrogen bond donor probe), O (hydrogen-oxygen carbonyl bond acceptor probe), OH2 (probe water), and DRY (hydrophobic probe). Descriptors not using probes were generated to create a total of 128 descriptors [35].
3.4.2. Models for the Study of Qualitative-Structure Activity Relationship KNIME 3.1.0 software (3.1.0 KNIME from Konstanz Information Miner, copyright 2003-2014, www.knime.org) [31] was used to perform all of the analyses. For internal validation, cross validation was employed using the "leave-one-out" method. Descriptors were selected, and a model was generated using the training and cross-validation set, using the WEKA nodes [35]. The internal performances of the selected models were analyzed for sensitivity (true positive rate), specificity (true negative rate), and accuracy (overall predictability).

Conclusions
A number of N-(4-halobenzyl)amides 1-32 were prepared and their antifungal potential was evaluated in screening experiments carried out with seven fungal strains. Compounds 2, 3, 5, 6, 10, 14, and 15 (MICs = 256 and 512 µg·mL −1 ) showed considerable antifungal activity against all tested strains of the genus Candida. According to the SAR, it was observed that disubstituted amides, such as metaand para-hydroxylated or meta-methoxylated/para-hydroxylated or 3,4,5-trihydroxylated N-(4-halobenzyl)amides have better antifungal activity against Candida strains. In addition, the presence of the nitro group in an ortho position in the benzene ring contributes to the antifungal activity. This study with haloamides could help in the development of future therapeutic approaches to the growing problem of microbial pathogens via the discovery of novel antifungal agents.