N-Substituted 2-Isonicotinoylhydrazinecarboxamides — New Antimycobacterial Active Molecules

This report presents a new modification of the isoniazid (INH) structure linked with different anilines via a carbonyl group obtained by two synthetic procedures and with N-substituted 5-(pyridine-4-yl)-1,3,4-oxadiazole-2-amines prepared by their cyclisation. All synthesised derivatives were characterised by IR, NMR, MS and elemental analyses and were evaluated in vitro for their antimycobacterial activity against Mycobacterium tuberculosis H37Rv, Mycobacterium avium 330/88, Mycobacterium kansasii 235/80 and one clinical isolated strain of M. kansasii 6509/96. 2-Isonicotinoyl-N-(4-octylphenyl)hydrazinecarboxamide displayed an in vitro efficacy comparable to that of INH for M. tuberculosis with minimum inhibitory concentrations (MICs) of 1–2 μM. Among the halogenated derivatives, the best anti-tuberculosis activity was found for 2-isonicotinoyl-N-(2,4,6-trichlorophenyl)hydrazinecarboxamide (MIC = 4 μM). In silico modelling on the enoyl-acyl carrier protein reductase InhA confirmed that longer alkyl substituents are advantageous for the interactions and affinity to InhA. Most of the hydrazinecarboxamides, especially those derived from 4-alkylanilines, exhibited significant activity against INH-resistant nontuberculous mycobacteria.


Chemistry
The most straightforward method to obtain various N-substituted 2-isonicotinoylhydrazinecarboxamides 3 involves an addition of INH (1) to aryl isocyanates 2a-v. Commercially available isocyanates were used for the synthesis of the 4-methyl, 4-isopropyl, 4-tert-butyl, 4-n-butyl and 4-methoxy derivatives 3a-d, 3i and N-heptyl derivative 3v without a phenyl ring. To prepare the products 3e-h and 3k-u, the required aryl isocyanates 2e-h and 2k-u were generated in situ from the appropriate anilines 4 during treatment with triphosgene in the presence of triethylamine in dry dichloromethane (DCM; Scheme 1). These reaction intermediates were not isolated. The synthesis, subsequent isolation and purification gave yields within the range of 67%-97%. These described products 3 can serve as starting materials for the synthesis of 2,5-disubstituted 1,3,4-oxadiazoles. Namely, we previously developed an efficient procedure for the transformation of 1-acyl-4-substituted semicarbazides into the corresponding 1,3,4-oxadiazoles via in situ formed diazenes [14]. Some hydrazinecarboxamides 3, because they belong to the same type of compounds mentioned above, were treated with the mixture of triphenylphosphine and 1,2-dibromo-1,1,2,2tetrachloroethane in the presence of triethylamine in dry acetonitrile (MeCN) to generate the expected N-substituted-5-(pyridine-4-yl)-1,3,4-oxadiazol-2-amines 5, as depicted in Scheme 2, with yields within the range of 33%-81%.  Table 1).
When the N-(4-heptylphenyl) derivative 3g and the N-heptyl molecule 3v, which vary by the presence of a phenyl ring, were compared they exhibited identical MICs of 16/32 μM despite very different logP values. The influence of the 4-alkyl length on the activity is interesting: the expanding length from methyl 3a to n-butyl 3d increased the activity from 62.5 to 4/8 μM; pentyl 3e and hexyl 3f resulted in a milder activity; and heptyl 3g displayed lower MICs comparable to the isopropyl. Finally, the octyl substituted molecule 3h was evaluated and was determined to be the most potent compound in this series. One possible hypothesis that may explain this effect is the similarity of longer alkyl chain with fatty acids, which are structural fragments of biomembranes. The situation for the atypical mycobacteria was different. Eleven hydrazinecarboxamides showed significantly lower MICs for M. avium than INH: 3a-g, 3i, 3p, 3t and 3v (range, 16 to 125 μM) with the superiority of 3a and 3b (MICs of 16-32 μM) indicating that the best substitution pattern for the phenyl ring is a small 4-alkyl group (3a, 3b) followed by 4-methoxy (3i), and 4-heptyl groups (3g). Identical MIC values were obtained for 3v, the heptyl derivative without a phenyl ring. In general, N-substituted 2-isonicotinoyl hydrazinecarboxamides 3, incorporating either alkyl (3a-g, 3v) or methoxy (3i) groups, exhibited a significantly higher inhibition of M. avium growth than the halogenated derivatives 3j-u (MICs ≥ 125 μM).
For both strains of M. kansasii, the alkyl substituted hydrazinecarboxamides 3a-g and 3v showed a higher in vitro potency than the halogenphenyl derivatives 3j-u. Twelve compounds (3a-g, 3n, 3p, 3r, 3t and 3v) inhibited INH-resistant strain 235/80, with MICs ≤ 125 μM, but no molecule produced lower MIC values than INH for clinical strain 6509/96. Only the 4-isopropyl derivative 3b, the most active 2-isonicotinoylhydrazinecarboxamide 3 against all nontuberculous strains in this study (MICs 8-32 μM), displayed a similar in vitro efficacy. The most hydrophilic 4-methoxyphenyl 3i and, surprisingly, the most lipophilic derivative 4-octylphenyl 3h did not exhibit any significant activity. For the strain 6509/96, the presence of two or three chlorine atoms (3n, 3t, 3u) or a trifluoromethyl moiety at any position (3j, 3p-r) on the phenyl ring led to better activity among the halogenated compounds. In general, there is no general and unequivocal correlation of ClogP and antimycobacterial activity within this series. This finding supports the conclusions of Ventura and Martins [15].
According to the interpretation of Scior and Garcés-Eisele [6], 2-isonicotinoylhydrazinecarboxamides 3 act as INH (and then isonicotinic acid) prodrugs that must be activated. None of these compounds showed superiority over INH against drug-susceptible M. tuberculosis H 37 Rv. In contrast, a range of these compounds showed notable in vitro activity at micromolar concentrations against the INH-resistant strains of M. avium and M. kansasii. In the case of M. avium, due to the missing catalase/peroxidase enzyme, this observation can be explained on the basis of facilitated liberation of isonicotinoyl radicale from the less stable hydrazide derivative prodrugs [6].

Molecular Modelling Studies
The enoyl-acyl carrier protein reductase InhA, the target of INH, was further investigated as a possible target of INH derivatives. Molecular modelling studies were performed to suggest possible conformations of the novel compounds in the active site of the enzyme and to search for possible interactions between the ligand and InhA.
In previous studies [20,21], the crystal structure of InhA and its active site were described in detail. InhA belongs to the family of NADH -dependent dehydrogenase enzymes. Participating in fatty acid biosynthesis, the structure of InhA is adapted to accommodate its natural substrates, namely, long chained fatty acids. Adjacent to the NADH binding site is a substrate binding loop (residues 196-219) forming an oval-shaped cavity with plenty of hydrophobic amino acid residues. One side of the cavity is widely accessible to solvent. A crystallographic study of a ternary complex of InhA and C16 fatty acyl substrates revealed that the substrate occupies the substrate binding cavity in a U-shaped conformation, which is stabilised by surrounding hydrophobic side chains of amino acid residues at one side of the cavity and by hydrogen bonds with NADH at the other [22]. Molecular modelling studies of N-(4-alkylphenyl) substituted hydrazinecarboxamides (3a-h) indicate a similar binding of the derivatives to the active site of InhA, as in the previously described study of Rozwarski et al. [22]. The pyridine moiety of the ligands is positioned near the entrance to the cavity, and the hydrazinecarboxamide connecting linker is on the top of NADH within hydrogen bonding distance. Some of the molecules also display one direct hydrogen bond to the protein between the carboxamide nitrogen and the hydroxyl group of Tyr158. The lipophilic portion of the molecule point deep into the cavity and is surrounded by the hydrophobic side chains of Ile202, Leu218, Pro193, Met199, Met161, Met103, Phe149 and Ile215. The longer the alkyl substituent is, the more interactions are available in the area of the substrate cavity. One of the highest affinities towards the enzyme was observed in the case of 4-octylphenyl derivative 3h (docking score of −8.6 kJ/mol; Figure 1).
Concerning the docking of variously halogenated derivatives, the conformation of these compounds was similar to that observed in the 4-alkylphenyl group with a pyridine moiety pointing towards the entrance of the cavity and a phenyl moiety pointing into the cavity. However, these compounds did not fill the substrate binding cavity as did the compounds with the longer alkyl substituents. This could result in the formation of fewer interactions with the hydrophobic residues and therefore lead to a weaker affinity towards the enzyme. The derivatives containing an oxadiazole moiety displayed similar conformation with hydrogen bonding to NADH and few hydrophobic interactions with Met103, Met199, Met161 and Phe149.
According to the observed conformations displayed by various derivatives, longer alkyl substituents are advantageous in the formation of interactions and therefore in the affinity to InhA. This phenomenon can be assumed from the structure of the natural substrates -long chain fatty acids. These results indicate that presented compounds 3 may affect InhA directly without previous necessary activation to form isonicotinoyl radical.

General Information
The starting materials for the synthesis of the examined compounds were used as obtained from commercial sources (Sigma-Aldrich, Prague, Czech Republic, Fluka, Prague, Czech Republic, Penta Chemicals, Prague, Czech Republic, Fluka Alfa Aesar, Ljubljana, Slovenia, Maybridge Chemical Company Ltd, Ljubljana, Slovenia). The melting points were determined on a Kofler micro hot stage or on a Büchi Melting Point machine B-540 apparatus using open capillaries and were uncorrected. The NMR spectra were recorded at 29 °C on a Varian Mercury-Vxbb 300 (300 MHz for 1 H and 75.5 MHz for 13 C; Varian, Inc., Palo Alto, CA, USA) or with a Bruker Avance III 500 spectrometer ( 1 H-NMR spectra at 500 MHz; 13 C-NMR spectra at 125.8 MHz) using deuterated dimethyl sulfoxide (DMSO-d 6 ) solutions of the samples. Proton spectra are referenced to TMS as the internal standard; the carbon shifts are given against the central line of the solvent signal (DMSO-d 6 at δ = 39.5 ppm; CDCl 3 at δ = 77.1 ppm). The coupling constants (J) are reported in Hz. IR spectra were obtained with a Perkin-Elmer Spectrum 100, equipped with a Specac Golden Gate Diamond ATR as a solid sample support, or they were recorded on a Nicolet 6700 FT-IR spectrometer in the range of 400-4,000 cm −1 (ATR). MS spectra were recorded with an Agilent 6224 Accurate Mass TOF LC/MS spectrometer. Elemental analyses (C, H, N) were performed with a Perkin Elmer 2400 Series II CHNS/O Analyzer. TLC was performed on Fluka silica-gel TLC-cards or on plates coated with 0.2 mm of Merck 60 F254 silica gel, and the chromatograms were visualised by UV irradiation (254 nm). The purity of synthesised compounds was checked by TLC, NMR, elemental analysis and for some compounds by HRMS. The calculated logP values (ClogP), which are the logarithms of the partition coefficients for octan-1-ol/water, were determined using the program CS ChemOffice Ultra version 12.0 (CambridgeSoft, Cambridge, MA, USA).

Synthesis of Isonicotinoylhydrazinecarboxamides 3
Method A   Isoniazid (1, 137.1 mg, 1.0 mmol) was suspended in anhydrous dichloromethane (5 mL), followed by the addition of the appropriate aryl isocyanate (2, 1.0 mmol) in one volume. The reaction mixture was stirred at room temperature for 30 min. A solid material was filtered off and recrystallised from boiling ethyl acetate. The reaction was monitored by TLC using the mixture n-hexane/ethyl acetate 3:1 as an eluent.

Conclusions
The synthesis of new INH derivatives, 2-isonicotinoyl-N-(substituted)hydrazinecarboxamides, was performed from commercially available or in situ synthesised isocyanates with good yields, and their structures were confirmed by physico-chemical analyses. The biological test data for Mycobacterium tuberculosis showed that none of the synthesised compounds exhibited higher activity than INH, but the N-(4-octylphenyl) derivative 3h reached almost the same in vitro efficacy with MICs 1-2 μM. Among the halogenated molecules, the best activity was found for the 2,4,6-trichloro derivative 3u, with a MIC of 4 μM. An incorporation of the alkyl substituent on the carboxamide nitrogen of 2-isonicotinoylhydrazinecarboxamides led to a significantly higher inhibition of the growth of M. avium and both M. kansasii strains when compared to the halogenated derivatives and INH. Molecular modelling studies suggested a possible conformation of these novel compounds in the active site of the enzyme InhA. The 4-alkylphenyl group, the lipophilic portion, may be located in the cavity surrounded by hydrophobic side chains of amino acids; thus, the longer the alkyl substituent is, the more interactions are available in this area of the substrate cavity. Based on this hypothesis, derivatives bearing long alkyl or alkylphenyl connected with isoniazid moiety may bring an additional benefit.
Cyclisation of the selected compounds 3 to the corresponding 1,3,4-oxadiazol-2-amines 5 resulted in the loss of antimycobacterial efficacy. Although this observation is in contrast with previously published results, similar compounds of type 5 were not further investigated.