Mutually Isomeric 2- and 4-(3-Nitro-1,2,4-triazol-1-yl)pyrimidines Inspired by an Antimycobacterial Screening Hit: Synthesis and Biological Activity against the ESKAPE Panel of Pathogens

Starting from the structure of antimycobacterial screening hit OTB-021 which was devoid of activity against ESKAPE pathogens, we designed, synthesized and tested two mutually isomeric series of novel simplified analogs, 2- and 4-(3-nitro-1,2,4-triazol-1-yl)pyrimidines, bearing various amino side chains. These compounds demonstrated a reverse bioactivity profile being inactive against M. tuberculosis while inhibiting the growth of all ESKAPE pathogens (with variable potency patterns) except for Gram-negative P. aeruginosa. Reduction potentials (E1/2, V) measured for selected compounds by cyclic voltammetry were tightly grouped in the −1.3–−1.1 V range for a reversible single-electron reduction. No apparent correlation between the E1/2 values and the ESKAPE minimum inhibitory concentrations was established, suggesting possible significance of other factors, besides the compounds’ reduction potential, which determine the observed antibacterial activity. Generally, more negative E1/2 values were displayed by 2-(3-nitro-1,2,4-triazol-1-yl)pyrimidines, which is in line with the frequently observed activity loss on moving the 3-nitro-1,2,4-triazol-1-yl moiety from position 4 to position 2 of the pyrimidine nucleus.


Introduction
N-Aryl-C-nitroazoles represent a general class of heterocyclic compounds which have found utility as pesticides [1], herbicides, fungicides [2] and high-energy materials [3][4][5]. At the same time, compounds belonging to this broadly defined chemical class are noticeably underrepresented in the medicinal chemistry literature [6]. Possible reasons for this include the negative stigma associated with nitro heteroaromatic moieties in general which are redox-active moieties and can therefore exert non-specific toxicity and mutagenicity [7]. Although nowadays such moieties continue being avoided, there is a steadily growing sentiment (particularly in the antibacterial field [8][9][10]) that the toxic effects In our program aimed at discovering new efficacious antibacterial leads, we undertook screening of a set of diverse nitrogen heteroaromatic compounds bearing a bioreducible nitro group. From this effort, compound OTB-021 (5-methyl-7-(3-nitro-1,2,4-triazol-1-yl)-1,2,4-triazolo [1,5a]pyrimidine) surfaced as a moderately potent hit with specific activity against drug-sensitive H37Rv strain of Mycobacterium tuberculosis [14], while other Gram-positive (S. aureus and E. faecium) or Gramnegative (E. coli, P. aeruginosa, A. baumannii, K. pneumoniae) pathogens belonging to the so-called ESKAPE panel [15] and immortalized cancer cell lines (MCF-7, FS4-LTM, KB-3-1, L929) were not affected. We hypothesized that the 4-(3-nitro- [1,2,4]-triazol-1-yl)pyrimidine portion of OTB-021 (highlighted in red) was likely responsible for its antibacterial properties, and we sought to verify this premise by simplifying the structure of this hit molecule, making it more amenable to structural variations and investigating the structure-activity relationships (SAR). To this end, we designed two isomeric series, -2-and 4-(3-nitro-1,2,4-triazol-1-yl)pyrimidines 1 and 2, bearing diverse amino side chains in positions 4 and 2 of the pyrimidine ring, respectively. The principal idea behind such a design was to reduce the bicyclic aromatic nitrogen-rich core of OTB-021 to the more druglike pyrimidine as well as to provide sufficient room for structural diversity via the side chain variation ( Figure 2). Herein, we report the synthesis and comparative evaluation of the two novel series of compounds with respect to their ability to inhibit the growth of bacterial pathogens. In our program aimed at discovering new efficacious antibacterial leads, we undertook screening of a set of diverse nitrogen heteroaromatic compounds bearing a bioreducible nitro group. From this effort, compound OTB-021 (5-methyl-7-(3-nitro-1,2,4-triazol-1-yl)-1,2,4-triazolo [1,5-a]pyrimidine) surfaced as a moderately potent hit with specific activity against drug-sensitive H37Rv strain of Mycobacterium tuberculosis [14], while other Gram-positive (S. aureus and E. faecium) or Gram-negative (E. coli, P. aeruginosa, A. baumannii, K. pneumoniae) pathogens belonging to the so-called ESKAPE panel [15] and immortalized cancer cell lines (MCF-7, FS4-LTM, KB-3-1, L929) were not affected. We hypothesized that the 4-(3-nitro- [1,2,4]-triazol-1-yl)pyrimidine portion of OTB-021 (highlighted in red) was likely responsible for its antibacterial properties, and we sought to verify this premise by simplifying the structure of this hit molecule, making it more amenable to structural variations and investigating the structure-activity relationships (SAR). To this end, we designed two isomeric series, -2-and 4-(3-nitro-1,2,4-triazol-1-yl)pyrimidines 1 and 2, bearing diverse amino side chains in positions 4 and 2 of the pyrimidine ring, respectively. The principal idea behind such a design was to reduce the bicyclic aromatic nitrogen-rich core of OTB-021 to the more druglike pyrimidine as well as to provide sufficient room for structural diversity via the side chain variation ( Figure 2). Herein, we report the synthesis and comparative evaluation of the two novel series of compounds with respect to their ability to inhibit the growth of bacterial pathogens.
Antibiotics 2020, 9, x FOR PEER REVIEW 2 of 21 non-specific toxicity and mutagenicity [7]. Although nowadays such moieties continue being avoided, there is a steadily growing sentiment (particularly in the antibacterial field [8][9][10]) that the toxic effects of nitro heteroaromatics to the human host can be alleviated-and detrimental effects to the pathogen retained or even increased-by careful optimization of the molecular periphery around the nitro heteroaromatic warhead. The feasibility of such an approach has been demonstrated by the recent approval of antitubercular nitroimidazole drugs delamanid [11] and pretomanid [12] ( Figure  1) which act as metabolically activated prodrugs. The progress in this field as well as the pros and cons of incorporating nitro (hetero)aromatic groups in drug candidate molecules have been comprehensively summarized in a recent review [13]. In our program aimed at discovering new efficacious antibacterial leads, we undertook screening of a set of diverse nitrogen heteroaromatic compounds bearing a bioreducible nitro group. From this effort, compound OTB-021 (5-methyl-7-(3-nitro-1,2,4-triazol-1-yl)-1,2,4-triazolo [1,5a]pyrimidine) surfaced as a moderately potent hit with specific activity against drug-sensitive H37Rv strain of Mycobacterium tuberculosis [14], while other Gram-positive (S. aureus and E. faecium) or Gramnegative (E. coli, P. aeruginosa, A. baumannii, K. pneumoniae) pathogens belonging to the so-called ESKAPE panel [15] and immortalized cancer cell lines (MCF-7, FS4-LTM, KB-3-1, L929) were not affected. We hypothesized that the 4-(3-nitro- [1,2,4]-triazol-1-yl)pyrimidine portion of OTB-021 (highlighted in red) was likely responsible for its antibacterial properties, and we sought to verify this premise by simplifying the structure of this hit molecule, making it more amenable to structural variations and investigating the structure-activity relationships (SAR). To this end, we designed two isomeric series, -2-and 4-(3-nitro-1,2,4-triazol-1-yl)pyrimidines 1 and 2, bearing diverse amino side chains in positions 4 and 2 of the pyrimidine ring, respectively. The principal idea behind such a design was to reduce the bicyclic aromatic nitrogen-rich core of OTB-021 to the more druglike pyrimidine as well as to provide sufficient room for structural diversity via the side chain variation ( Figure 2). Herein, we report the synthesis and comparative evaluation of the two novel series of compounds with respect to their ability to inhibit the growth of bacterial pathogens.  Initially, our synthetic efforts focused on the installation of the 3-nitro-1,2,4-triazol-1-yl unit on the pyrimidine nucleus using commercially available 5-nitro-1H-1,2,4-triazole and 2,4-dichloropyrimidine. Firstly, it was promptly established that the reaction did not require the use of a metal-based catalyst (Pd 0 or Cu I ) and proceeded as direct nucleophilic aromatic (S N Ar) substitution. Secondly, performing the reaction in the presence of even a slight excess of the base diminished the yield as it led to the degradation of the 2,4-dichloropyrimidine starting material. Hence, 5-nitro-1H-1,2,4-triazole potassium salt (3) was obtained [16] in a separate step and used in the chloride displacement reactions. Finally, we discovered that achieving a good yield of the monosubstitution product was not straightforward as even at low conversions of the monosubstitution, the product formed reacted with the second equivalent of 3, even in the presence of unreacted 2,4-dichloropyrimidine. Considering that the 3-nitro-1,2,4-triazol-1-yl moiety itself can served as a good leaving group in S N Ar reactions [17], we obtained disubstituted product 4 in excellent yield (Scheme 1) and subsequently employed it in S N Ar reactions with amines. Initially, our synthetic efforts focused on the installation of the 3-nitro-1,2,4-triazol-1-yl unit on the pyrimidine nucleus using commercially available 5-nitro-1H-1,2,4-triazole and 2,4dichloropyrimidine. Firstly, it was promptly established that the reaction did not require the use of a metal-based catalyst (Pd 0 or Cu I ) and proceeded as direct nucleophilic aromatic (SNAr) substitution. Secondly, performing the reaction in the presence of even a slight excess of the base diminished the yield as it led to the degradation of the 2,4-dichloropyrimidine starting material. Hence, 5-nitro-1H-1,2,4-triazole potassium salt (3) was obtained [16] in a separate step and used in the chloride displacement reactions. Finally, we discovered that achieving a good yield of the monosubstitution product was not straightforward as even at low conversions of the monosubstitution, the product formed reacted with the second equivalent of 3, even in the presence of unreacted 2,4dichloropyrimidine. Considering that the 3-nitro-1,2,4-triazol-1-yl moiety itself can served as a good leaving group in SNAr reactions [17], we obtained disubstituted product 4 in excellent yield (Scheme 1) and subsequently employed it in SNAr reactions with amines. Scheme 1. Preparation of 2,4-bis(3-nitro-1H-1,2,4-triazol-1-yl)pyrimidine (4).
The SNAr reactions of 4 with aliphatic amines indeed proceeded rather smoothly at room temperature in acetonitrile, while more forcing conditions (DMSO, 100 °C, 24-48 h) had to be applied with aromatic amines. The reactions displayed a pronounced selectivity toward the displacement of the 3-nitro-1H-1,2,4-triazol-1-yl leaving group in position 4 of the pyrimidine nucleus and allowed obtaining satisfactory yields of compounds 1a-s (Table 1). Chromatographic isolation of sufficiently pure (>90% of purity according to 1 H NMR) regioisomeric products 2 from these reactions was not feasible; hence, 4-(3-nitro-1,2,4-triazol-1-yl)pyrimidines were accessed via a different strategy. The S N Ar reactions of 4 with aliphatic amines indeed proceeded rather smoothly at room temperature in acetonitrile, while more forcing conditions (DMSO, 100 • C, 24-48 h) had to be applied with aromatic amines. The reactions displayed a pronounced selectivity toward the displacement of the 3-nitro-1H-1,2,4-triazol-1-yl leaving group in position 4 of the pyrimidine nucleus and allowed obtaining satisfactory yields of compounds 1a-s (Table 1). Chromatographic isolation of sufficiently pure (>90% of purity according to 1 H NMR) regioisomeric products 2 from these reactions was not feasible; hence, 4-(3-nitro-1,2,4-triazol-1-yl)pyrimidines were accessed via a different strategy. Initially, our synthetic efforts focused on the installation of the 3-nitro-1,2,4-triazol-1-yl unit on the pyrimidine nucleus using commercially available 5-nitro-1H-1,2,4-triazole and 2,4dichloropyrimidine. Firstly, it was promptly established that the reaction did not require the use of a metal-based catalyst (Pd 0 or Cu I ) and proceeded as direct nucleophilic aromatic (SNAr) substitution. Secondly, performing the reaction in the presence of even a slight excess of the base diminished the yield as it led to the degradation of the 2,4-dichloropyrimidine starting material. Hence, 5-nitro-1H-1,2,4-triazole potassium salt (3) was obtained [16] in a separate step and used in the chloride displacement reactions. Finally, we discovered that achieving a good yield of the monosubstitution product was not straightforward as even at low conversions of the monosubstitution, the product formed reacted with the second equivalent of 3, even in the presence of unreacted 2,4dichloropyrimidine. Considering that the 3-nitro-1,2,4-triazol-1-yl moiety itself can served as a good leaving group in SNAr reactions [17], we obtained disubstituted product 4 in excellent yield (Scheme 1) and subsequently employed it in SNAr reactions with amines. Scheme 1. Preparation of 2,4-bis(3-nitro-1H-1,2,4-triazol-1-yl)pyrimidine (4).
The SNAr reactions of 4 with aliphatic amines indeed proceeded rather smoothly at room temperature in acetonitrile, while more forcing conditions (DMSO, 100 °C, 24-48 h) had to be applied with aromatic amines. The reactions displayed a pronounced selectivity toward the displacement of the 3-nitro-1H-1,2,4-triazol-1-yl leaving group in position 4 of the pyrimidine nucleus and allowed obtaining satisfactory yields of compounds 1a-s (Table 1). Chromatographic isolation of sufficiently pure (>90% of purity according to 1 H NMR) regioisomeric products 2 from these reactions was not feasible; hence, 4-(3-nitro-1,2,4-triazol-1-yl)pyrimidines were accessed via a different strategy. Initially, our synthetic efforts focused on the installation of the 3-nitro-1,2,4-triazol-1-yl unit on the pyrimidine nucleus using commercially available 5-nitro-1H-1,2,4-triazole and 2,4dichloropyrimidine. Firstly, it was promptly established that the reaction did not require the use of a metal-based catalyst (Pd 0 or Cu I ) and proceeded as direct nucleophilic aromatic (SNAr) substitution. Secondly, performing the reaction in the presence of even a slight excess of the base diminished the yield as it led to the degradation of the 2,4-dichloropyrimidine starting material. Hence, 5-nitro-1H-1,2,4-triazole potassium salt (3) was obtained [16] in a separate step and used in the chloride displacement reactions. Finally, we discovered that achieving a good yield of the monosubstitution product was not straightforward as even at low conversions of the monosubstitution, the product formed reacted with the second equivalent of 3, even in the presence of unreacted 2,4dichloropyrimidine. Considering that the 3-nitro-1,2,4-triazol-1-yl moiety itself can served as a good leaving group in SNAr reactions [17], we obtained disubstituted product 4 in excellent yield (Scheme 1) and subsequently employed it in SNAr reactions with amines. Scheme 1. Preparation of 2,4-bis(3-nitro-1H-1,2,4-triazol-1-yl)pyrimidine (4).
The SNAr reactions of 4 with aliphatic amines indeed proceeded rather smoothly at room temperature in acetonitrile, while more forcing conditions (DMSO, 100 °C, 24-48 h) had to be applied with aromatic amines. The reactions displayed a pronounced selectivity toward the displacement of the 3-nitro-1H-1,2,4-triazol-1-yl leaving group in position 4 of the pyrimidine nucleus and allowed obtaining satisfactory yields of compounds 1a-s (Table 1). Chromatographic isolation of sufficiently pure (>90% of purity according to 1 H NMR) regioisomeric products 2 from these reactions was not feasible; hence, 4-(3-nitro-1,2,4-triazol-1-yl)pyrimidines were accessed via a different strategy. The regiochemistry of series 1 was unequivocally confirmed by the single-crystal X-ray structure obtained for compound 1e ( Figure 3, Table S1). We reasoned that the seemingly unavoidable formation of disubstituted product 4 in reactions of 3 with 2,4-dichloropyrimidine could be circumvented if excess of the latter was used while the concentration of 3 was maintained low over the course of the reaction. After substantial experimentation, we established that with 0.5 equiv. of 3 added as a DMF solution to a solution of 2,4-dichloropyrimidine in DMF using a syringe pump over 8 h (at a rate of about 0.82 mmol in 2.5 mL per hour) at 80 °C, The regiochemistry of series 1 was unequivocally confirmed by the single-crystal X-ray structure obtained for compound 1e ( Figure 3, Table S1). We reasoned that the seemingly unavoidable formation of disubstituted product 4 in reactions of 3 with 2,4-dichloropyrimidine could be circumvented if excess of the latter was used while the concentration of 3 was maintained low over the course of the reaction. After substantial experimentation, we established that with 0.5 equiv. of 3 added as a DMF solution to a solution of 2,4-dichloropyrimidine in DMF using a syringe pump over 8 h (at a rate of about 0.82 mmol in 2.5 The regiochemistry of series 1 was unequivocally confirmed by the single-crystal X-ray structure obtained for compound 1e ( Figure 3, Table S1). We reasoned that the seemingly unavoidable formation of disubstituted product 4 in reactions of 3 with 2,4-dichloropyrimidine could be circumvented if excess of the latter was used while the concentration of 3 was maintained low over the course of the reaction. After substantial experimentation, we established that with 0.5 equiv. of 3 added as a DMF solution to a solution of 2,4-dichloropyrimidine in DMF using a syringe pump over 8 h (at a rate of about 0.82 mmol in 2.5 The regiochemistry of series 1 was unequivocally confirmed by the single-crystal X-ray structure obtained for compound 1e ( Figure 3, Table S1). We reasoned that the seemingly unavoidable formation of disubstituted product 4 in reactions of 3 with 2,4-dichloropyrimidine could be circumvented if excess of the latter was used while the concentration of 3 was maintained low over the course of the reaction. After substantial experimentation, we established that with 0.5 equiv. of 3 added as a DMF solution to a solution of 2,4-dichloropyrimidine in DMF using a syringe pump over 8 h (at a rate of about 0.82 mmol in 2.5 The regiochemistry of series 1 was unequivocally confirmed by the single-crystal X-ray structure obtained for compound 1e ( Figure 3, Table S1).
We reasoned that the seemingly unavoidable formation of disubstituted product 4 in reactions of 3 with 2,4-dichloropyrimidine could be circumvented if excess of the latter was used while the concentration of 3 was maintained low over the course of the reaction. After substantial experimentation, we established that with 0.5 equiv. of 3 added as a DMF solution to a solution of  The regiochemistry of series 1 was unequivocally confirmed by the single-crystal X-ray structure obtained for compound 1e ( Figure 3, Table S1).
We reasoned that the seemingly unavoidable formation of disubstituted product 4 in reactions of 3 with 2,4-dichloropyrimidine could be circumvented if excess of the latter was used while the concentration of 3 was maintained low over the course of the reaction. After substantial experimentation, we established that with 0.5 equiv. of 3 added as a DMF solution to a solution of a Reactions were performed at 100 • C in DMSO over 24-48 h. b No regioisomer was formed.
The regiochemistry of series 1 was unequivocally confirmed by the single-crystal X-ray structure obtained for compound 1e ( Figure 3, Table S1). The regiochemistry of series 1 was unequivocally confirmed by the single-crystal X-ray structure obtained for compound 1e ( Figure 3, Table S1).
We reasoned that the seemingly unavoidable formation of disubstituted product 4 in reactions of 3 with 2,4-dichloropyrimidine could be circumvented if excess of the latter was used while the concentration of 3 was maintained low over the course of the reaction. After substantial experimentation, we established that with 0.5 equiv. of 3 added as a DMF solution to a solution of We reasoned that the seemingly unavoidable formation of disubstituted product 4 in reactions of 3 with 2,4-dichloropyrimidine could be circumvented if excess of the latter was used while the concentration of 3 was maintained low over the course of the reaction. After substantial experimentation, we established that with 0.5 equiv. of 3 added as a DMF solution to a solution of 2,4-dichloropyrimidine in DMF using a syringe pump over 8 h (at a rate of about 0.82 mmol in 2.5 mL per hour) at 80 • C, monosubstituted 2-chloro-4-(3-nitro-1H-1,2,4-triazol-1-yl)pyrimidine (5) could be obtained in 43% yield (from 3), while unreacted 2,4-dichloropyrimidine could be isolated and utilized again. The substitution pattern of 5 was confirmed by the single-crystal X-ray analysis (Scheme 2, Table S1).
With sufficient amounts of 5 in hand, we proceeded in preparing compounds belonging to series 2 with the same set of amines as was used to prepare series 1 (to allow for direct comparison of the biological effects exerted by the two series). The substitution of the chlorine atom in position 2 with aliphatic amines generally gave moderate to high yields in acetonitrile at room temperature, while anilines, again, required heating at 100 °C in DMSO for the reaction to proceed (Table 2). Notably, in this case, no substitution of the 3-nitro-1,2,4-triazol-1-yl group in position 4 of the pyrimidine nucleus was observed. With sufficient amounts of 5 in hand, we proceeded in preparing compounds belonging to series 2 with the same set of amines as was used to prepare series 1 (to allow for direct comparison of the biological effects exerted by the two series). The substitution of the chlorine atom in position 2 with aliphatic amines generally gave moderate to high yields in acetonitrile at room temperature, while anilines, again, required heating at 100 • C in DMSO for the reaction to proceed (Table 2). Notably, in this case, no substitution of the 3-nitro-1,2,4-triazol-1-yl group in position 4 of the pyrimidine nucleus was observed.  (5) and its single-crystal X-ray structure (CCDC 1582455, thermal ellipsoids are shown at 50% probability).
With sufficient amounts of 5 in hand, we proceeded in preparing compounds belonging to series 2 with the same set of amines as was used to prepare series 1 (to allow for direct comparison of the biological effects exerted by the two series). The substitution of the chlorine atom in position 2 with aliphatic amines generally gave moderate to high yields in acetonitrile at room temperature, while anilines, again, required heating at 100 °C in DMSO for the reaction to proceed (Table 2). Notably, in this case, no substitution of the 3-nitro-1,2,4-triazol-1-yl group in position 4 of the pyrimidine nucleus was observed.
With sufficient amounts of 5 in hand, we proceeded in preparing compounds belonging to series 2 with the same set of amines as was used to prepare series 1 (to allow for direct comparison of the biological effects exerted by the two series). The substitution of the chlorine atom in position 2 with aliphatic amines generally gave moderate to high yields in acetonitrile at room temperature, while anilines, again, required heating at 100 °C in DMSO for the reaction to proceed (Table 2). Notably, in this case, no substitution of the 3-nitro-1,2,4-triazol-1-yl group in position 4 of the pyrimidine nucleus was observed.
With sufficient amounts of 5 in hand, we proceeded in preparing compounds belonging to series 2 with the same set of amines as was used to prepare series 1 (to allow for direct comparison of the biological effects exerted by the two series). The substitution of the chlorine atom in position 2 with aliphatic amines generally gave moderate to high yields in acetonitrile at room temperature, while anilines, again, required heating at 100 °C in DMSO for the reaction to proceed (Table 2). Notably, in this case, no substitution of the 3-nitro-1,2,4-triazol-1-yl group in position 4 of the pyrimidine nucleus was observed.
With sufficient amounts of 5 in hand, we proceeded in preparing compounds belonging to series 2 with the same set of amines as was used to prepare series 1 (to allow for direct comparison of the biological effects exerted by the two series). The substitution of the chlorine atom in position 2 with aliphatic amines generally gave moderate to high yields in acetonitrile at room temperature, while anilines, again, required heating at 100 °C in DMSO for the reaction to proceed (Table 2). Notably, in this case, no substitution of the 3-nitro-1,2,4-triazol-1-yl group in position 4 of the pyrimidine nucleus was observed.    The regiochemistry of series 2 was unequivocally confirmed by the single-crystal X-ray structures obtained for compounds 2h, 2k, 2n and 2o (Figure 4, Tables S2 and S3). The regiochemistry of series 2 was unequivocally confirmed by the single-crystal X-ray structures obtained for compounds 2h, 2k, 2n and 2o (Figure 4, Tables S2 and S3). The regiochemistry of series 2 was unequivocally confirmed by the single-crystal X-ray structures obtained for compounds 2h, 2k, 2n and 2o (Figure 4, Tables S2 and S3).

In Vitro Biological Evaluation
Surprisingly, when tested against the drug-sensitive H37Rv strain of Mycobacterium tuberculosis, none of the 38 compounds, 1a-s and 2a-s, displayed any appreciable activity. Gratifyingly, when screened against the ESKAPE panel of six bacterial pathogens commonly found to carry antimicrobial resistance genes [15], both series displayed a promising antibacterial profile, which is summarized in Table 3 ((bis(3-nitro

In Vitro Biological Evaluation
Surprisingly, when tested against the drug-sensitive H37Rv strain of Mycobacterium tuberculosis, none of the 38 compounds, 1a-s and 2a-s, displayed any appreciable activity. Gratifyingly, when screened against the ESKAPE panel of six bacterial pathogens commonly found to carry antimicrobial resistance genes [15], both series displayed a promising antibacterial profile, which is summarized in Table 3 ((bis(3-nitro-1,2,4-triazol-1-yl) compound 4 is shown for comparison and ciprofloxacin was employed as a positive control for all six microorganisms).   Several important observations emerge from the data presented in Table 3. Firstly, (bis(3-nitro-1,2,4-triazol-1-yl) compound 4 was virtually inactive (except for marginal activity on E. faecium). This clearly demonstrates that the combination of only one 3-nitro-1,2,4-triazol-1-yl moiety with an electron-donating amino substituent is the correct definition of the pharmacophore of both series 1 and 2. Secondly, there appears no apparent biologic activity dependence on the relative position of the two groups around the pyrimidine nucleus. Quite frequently, the high antibacterial potency of series 2 compounds was lost on switching to series 1 (cf. 2a →1a, 2b→1b, 2f→1f vs. K. pneumonia; 2e→1e, 2k→1k, 2s→1s vs. A. baumannii; 2e→1e vs. S. aureus; 2n→1n vs. E. faecium; 2b→1b vs. E. aerogenes). Less frequently, the opposite was true and series 1 was highly active, while series 2 was not (cf. 1i→2i and 1m→2m vs. S. aureus; 1j→2j and 1r→2r vs. A. baumannii; 1o→2o vs. S. aureus and E. faecium). It is an accepted view that the antibacterial activity of bioreducible nitro heteroaromatic compounds with respect to a particular bacterial species, among other factors, will depend on their ability to be metabolically activated by the membrane-bound nitro reductase enzyme of that species [18] as well as on the ability of the resulting reactive chemical entity to cross the bacterial membrane and damage the pathogen's DNA [19]. Considering such a multifactorial nature of the observed inhibitory effects on bacterial growth, it is unsurprising that no definitive bioactivity pattern between series 1 and 2 emerged.
Other notable features of the bioactivity profile of the compounds 1a-s and 2a-s include the complete absence of activity against P. aeruginosa and the distinct susceptibility of A. baumannii to many compounds in the investigated set. In fact, some of the compounds (cf. 1b, 1c, 1j, 1l, 1m, 1q, 2b,  2e, 2k, 2s) displayed MIC values comparable or even lower than those displayed by ciprofloxacin towards this particular pathogen. Another pathogen that demonstrated susceptibility to a number of compounds tested is E. faecium. However, the best MIC values achieved in this case (2 µ g/mL) are six times lower than the respective value for ciprofloxacin. At the same time, the activity of compounds 1i and 1o is only three times lower with respect to S. aureus than that of the comparator drug. Compound 2b certainly leads the way in terms of the single-digit µ g/mL activity displayed across the panel, strongly inhibiting the growth of E. faecium, K. pneumonia, A. baumannii and E. aerogenes. In contrast, compounds 2c, 2k, 1l and 2o appear to be distinctly selective towards A. baumannii, which is a characteristic tendency of the entire set. The complete absence of activity across the ESKAPE panel displayed by compound 1n is somewhat surprising and, again, attests to the multifactorial nature of the bioactivity patterns observed. Several important observations emerge from the data presented in Table 3. Firstly, (bis(3-nitro-1,2,4-triazol-1-yl) compound 4 was virtually inactive (except for marginal activity on E. faecium). This clearly demonstrates that the combination of only one 3-nitro-1,2,4-triazol-1-yl moiety with an electron-donating amino substituent is the correct definition of the pharmacophore of both series 1 and 2. Secondly, there appears no apparent biologic activity dependence on the relative position of the two groups around the pyrimidine nucleus. Quite frequently, the high antibacterial potency of series 2 compounds was lost on switching to series 1 (cf. 2a →1a, 2b→1b, 2f→1f vs. K. pneumonia; 2e→1e, 2k→1k, 2s→1s vs. A. baumannii; 2e→1e vs. S. aureus; 2n→1n vs. E. faecium; 2b→1b vs. E. aerogenes). Less frequently, the opposite was true and series 1 was highly active, while series 2 was not ( cf. 1i→2i and 1m→2m vs. S. aureus; 1j→2j and 1r→2r vs. A. baumannii; 1o→2o vs. S. aureus and E. faecium). It is an accepted view that the antibacterial activity of bioreducible nitro heteroaromatic compounds with respect to a particular bacterial species, among other factors, will depend on their ability to be metabolically activated by the membrane-bound nitro reductase enzyme of that species [18] as well as on the ability of the resulting reactive chemical entity to cross the bacterial membrane and damage the pathogen's DNA [19]. Considering such a multifactorial nature of the observed inhibitory effects on bacterial growth, it is unsurprising that no definitive bioactivity pattern between series 1 and 2 emerged.

Electrochemical Behavior
Other notable features of the bioactivity profile of the compounds 1a-s and 2a-s include the complete absence of activity against P. aeruginosa and the distinct susceptibility of A. baumannii to many compounds in the investigated set. In fact, some of the compounds (cf. 1b, 1c, 1j, 1l, 1m, 1q, 2b,  2e, 2k, 2s) displayed MIC values comparable or even lower than those displayed by ciprofloxacin towards this particular pathogen. Another pathogen that demonstrated susceptibility to a number of compounds tested is E. faecium. However, the best MIC values achieved in this case (2 µ g/mL) are six times lower than the respective value for ciprofloxacin. At the same time, the activity of compounds 1i and 1o is only three times lower with respect to S. aureus than that of the comparator drug. Compound 2b certainly leads the way in terms of the single-digit µ g/mL activity displayed across the panel, strongly inhibiting the growth of E. faecium, K. pneumonia, A. baumannii and E. aerogenes. In  contrast, compounds 2c, 2k, 1l and 2o  Several important observations emerge from the data presented in Table 3. Firstly, (bis(3-nitro-1,2,4-triazol-1-yl) compound 4 was virtually inactive (except for marginal activity on E. faecium). This clearly demonstrates that the combination of only one 3-nitro-1,2,4-triazol-1-yl moiety with an electron-donating amino substituent is the correct definition of the pharmacophore of both series 1 and 2. Secondly, there appears no apparent biologic activity dependence on the relative position of the two groups around the pyrimidine nucleus. Quite frequently, the high antibacterial potency of series 2 compounds was lost on switching to series 1 (cf. 2a →1a, 2b→1b, 2f→1f vs. K. pneumonia; 2e→1e, 2k→1k, 2s→1s vs. A. baumannii; 2e→1e vs. S. aureus; 2n→1n vs. E. faecium; 2b→1b vs. E. aerogenes). Less frequently, the opposite was true and series 1 was highly active, while series 2 was not (cf. 1i→2i and 1m→2m vs. S. aureus; 1j→2j and 1r→2r vs. A. baumannii; 1o→2o vs. S. aureus and E. faecium). It is an accepted view that the antibacterial activity of bioreducible nitro heteroaromatic compounds with respect to a particular bacterial species, among other factors, will depend on their ability to be metabolically activated by the membrane-bound nitro reductase enzyme of that species [18] as well as on the ability of the resulting reactive chemical entity to cross the bacterial membrane and damage the pathogen's DNA [19]. Considering such a multifactorial nature of the observed inhibitory effects on bacterial growth, it is unsurprising that no definitive bioactivity pattern between series 1 and 2 emerged.
Other notable features of the bioactivity profile of the compounds 1a-s and 2a-s include the complete absence of activity against P. aeruginosa and the distinct susceptibility of A. baumannii to many compounds in the investigated set. In fact, some of the compounds (cf. 1b, 1c, 1j, 1l, 1m, 1q, 2b,  2e, 2k, 2s) displayed MIC values comparable or even lower than those displayed by ciprofloxacin towards this particular pathogen. Another pathogen that demonstrated susceptibility to a number of compounds tested is E. faecium. However, the best MIC values achieved in this case (2 µ g/mL) are six times lower than the respective value for ciprofloxacin. At the same time, the activity of compounds 1i and 1o is only three times lower with respect to S. aureus than that of the comparator drug. Compound 2b certainly leads the way in terms of the single-digit µ g/mL activity displayed across the panel, strongly inhibiting the growth of E. faecium, K. pneumonia, A. baumannii and E. aerogenes. In contrast, compounds 2c, 2k, 1l and 2o  Several important observations emerge from the data presented in Table 3. Firstly, (bis(3-nitro-1,2,4-triazol-1-yl) compound 4 was virtually inactive (except for marginal activity on E. faecium). This clearly demonstrates that the combination of only one 3-nitro-1,2,4-triazol-1-yl moiety with an electron-donating amino substituent is the correct definition of the pharmacophore of both series 1 and 2. Secondly, there appears no apparent biologic activity dependence on the relative position of the two groups around the pyrimidine nucleus. Quite frequently, the high antibacterial potency of series 2 compounds was lost on switching to series 1 (cf. 2a →1a, 2b→1b, 2f→1f vs. K. pneumonia; 2e→1e, 2k→1k, 2s→1s vs. A. baumannii; 2e→1e vs. S. aureus; 2n→1n vs. E. faecium; 2b→1b vs. E. aerogenes). Less frequently, the opposite was true and series 1 was highly active, while series 2 was not (cf. 1i→2i and 1m→2m vs. S. aureus; 1j→2j and 1r→2r vs. A. baumannii; 1o→2o vs. S. aureus and E. faecium). It is an accepted view that the antibacterial activity of bioreducible nitro heteroaromatic compounds with respect to a particular bacterial species, among other factors, will depend on their ability to be metabolically activated by the membrane-bound nitro reductase enzyme of that species [18] as well as on the ability of the resulting reactive chemical entity to cross the bacterial membrane and damage the pathogen's DNA [19]. Considering such a multifactorial nature of the observed inhibitory effects on bacterial growth, it is unsurprising that no definitive bioactivity pattern between series 1 and 2 emerged.
Other notable features of the bioactivity profile of the compounds 1a-s and 2a-s include the complete absence of activity against P. aeruginosa and the distinct susceptibility of A. baumannii to many compounds in the investigated set. In fact, some of the compounds (cf. 1b, 1c, 1j, 1l, 1m, 1q, 2b,  2e, 2k, 2s) displayed MIC values comparable or even lower than those displayed by ciprofloxacin towards this particular pathogen. Another pathogen that demonstrated susceptibility to a number of compounds tested is E. faecium. However, the best MIC values achieved in this case (2 µ g/mL) are six times lower than the respective value for ciprofloxacin. At the same time, the activity of compounds 1i and 1o is only three times lower with respect to S. aureus than that of the comparator drug. Compound 2b certainly leads the way in terms of the single-digit µ g/mL activity displayed across the panel, strongly inhibiting the growth of E. faecium, K. pneumonia, A. baumannii and E. aerogenes. In contrast, compounds 2c, 2k, 1l and 2o appear to be distinctly selective towards A. baumannii, which is a characteristic tendency of the entire set. The complete absence of activity across the ESKAPE panel displayed by compound 1n is somewhat surprising and, again, attests to the multifactorial nature of the bioactivity patterns observed. Several important observations emerge from the data presented in Table 3. Firstly, (bis(3-nitro-1,2,4-triazol-1-yl) compound 4 was virtually inactive (except for marginal activity on E. faecium). This clearly demonstrates that the combination of only one 3-nitro-1,2,4-triazol-1-yl moiety with an electron-donating amino substituent is the correct definition of the pharmacophore of both series 1 and 2. Secondly, there appears no apparent biologic activity dependence on the relative position of the two groups around the pyrimidine nucleus. Quite frequently, the high antibacterial potency of series 2 compounds was lost on switching to series 1 (cf. 2a →1a, 2b→1b, 2f→1f vs. K. pneumonia; 2e→1e, 2k→1k, 2s→1s vs. A. baumannii; 2e→1e vs. S. aureus; 2n→1n vs. E. faecium; 2b→1b vs. E. aerogenes). Less frequently, the opposite was true and series 1 was highly active, while series 2 was not (cf. 1i→2i and 1m→2m vs. S. aureus; 1j→2j and 1r→2r vs. A. baumannii; 1o→2o vs. S. aureus and E. faecium). It is an accepted view that the antibacterial activity of bioreducible nitro heteroaromatic compounds with respect to a particular bacterial species, among other factors, will depend on their ability to be metabolically activated by the membrane-bound nitro reductase enzyme of that species [18] as well as on the ability of the resulting reactive chemical entity to cross the bacterial membrane and damage the pathogen's DNA [19]. Considering such a multifactorial nature of the observed inhibitory effects on bacterial growth, it is unsurprising that no definitive bioactivity pattern between series 1 and 2 emerged.

Electrochemical Behavior
Other notable features of the bioactivity profile of the compounds 1a-s and 2a-s include the complete absence of activity against P. aeruginosa and the distinct susceptibility of A. baumannii to many compounds in the investigated set. In fact, some of the compounds (cf. 1b, 1c, 1j, 1l, 1m, 1q, 2b,  2e, 2k, 2s) displayed MIC values comparable or even lower than those displayed by ciprofloxacin towards this particular pathogen. Another pathogen that demonstrated susceptibility to a number of compounds tested is E. faecium. However, the best MIC values achieved in this case (2 µ g/mL) are six times lower than the respective value for ciprofloxacin. At the same time, the activity of compounds 1i and 1o is only three times lower with respect to S. aureus than that of the comparator drug. Compound 2b certainly leads the way in terms of the single-digit µ g/mL activity displayed across the panel, strongly inhibiting the growth of E. faecium, K. pneumonia, A. baumannii and E. aerogenes. In contrast, compounds 2c, 2k, 1l and 2o appear to be distinctly selective towards A. baumannii, which is a characteristic tendency of the entire set. The complete absence of activity across the ESKAPE panel displayed by compound 1n is somewhat surprising and, again, attests to the multifactorial nature of the bioactivity patterns observed. Several important observations emerge from the data presented in Table 3. Firstly, (bis(3-nitro-1,2, 4-triazol-1-yl) compound 4 was virtually inactive (except for marginal activity on E. faecium). This clearly demonstrates that the combination of only one 3-nitro-1,2,4-triazol-1-yl moiety with an electron-donating amino substituent is the correct definition of the pharmacophore of both series 1 and 2. Secondly, there appears no apparent biologic activity dependence on the relative position of the two groups around the pyrimidine nucleus. Quite frequently, the high antibacterial potency of series 2 compounds was lost on switching to series 1 (cf. 2a →1a, 2b→1b, 2f→1f vs. K. pneumonia; 2e→1e, 2k→1k, 2s→1s vs. A. baumannii; 2e→1e vs. S. aureus; 2n→1n vs. E. faecium; 2b→1b vs. E. aerogenes). Less frequently, the opposite was true and series 1 was highly active, while series 2 was not (cf. 1i→2i and 1m→2m vs. S. aureus; 1j→2j and 1r→2r vs. A. baumannii; 1o→2o vs. S. aureus and E. faecium). It is an accepted view that the antibacterial activity of bioreducible nitro heteroaromatic compounds with respect to a particular bacterial species, among other factors, will depend on their ability to be metabolically activated by the membrane-bound nitro reductase enzyme of that species [18] as well as on the ability of the resulting reactive chemical entity to cross the bacterial membrane and damage the pathogen's DNA [19]. Considering such a multifactorial nature of the observed inhibitory effects on bacterial growth, it is unsurprising that no definitive bioactivity pattern between series 1 and 2 emerged.
Other notable features of the bioactivity profile of the compounds 1a-s and 2a-s include the complete absence of activity against P. aeruginosa and the distinct susceptibility of A. baumannii to many compounds in the investigated set. In fact, some of the compounds (cf. 1b, 1c, 1j, 1l, 1m, 1q, 2b, 2e, 2k, 2s) displayed MIC values comparable or even lower than those displayed by ciprofloxacin towards this particular pathogen. Another pathogen that demonstrated susceptibility to a number of compounds tested is E. faecium. However, the best MIC values achieved in this case (2 µg/mL) are six times lower than the respective value for ciprofloxacin. At the same time, the activity of compounds 1i and 1o is only three times lower with respect to S. aureus than that of the comparator drug. Compound 2b certainly leads the way in terms of the single-digit µg/mL activity displayed across the panel, strongly inhibiting the growth of E. faecium, K. pneumonia, A. baumannii and E. aerogenes. In contrast, compounds 2c, 2k, 1l and 2o appear to be distinctly selective towards A. baumannii, which is a characteristic tendency of the entire set. The complete absence of activity across the ESKAPE panel displayed by compound 1n is somewhat surprising and, again, attests to the multifactorial nature of the bioactivity patterns observed.

Electrochemical Behavior
Since it was hypothesized that, like other nitro (hetero)aromatic drugs [20], (3-nitro-1,2,4-triazol-1-yl)pyrimidines 1 and 2 investigated in this work are activated via the nitro group reduction by the bacterial nitroreductase, we investigated the electrochemical behavior of selected compounds from both series using cyclic voltammetry. The aim of this effort was to establish if there is an apparent correlation of the antibacterial properties against any bacterial species (as was reported for some antimicrobial nitroaromatic compounds [21] as well as metal chelate complexes [22]) and the compounds' reduction potential. Additionally, we thought it interesting to obtain a pairwise comparison of the reduction potential of series 1 and 2 compounds which might shed light on some of the bioactivity trends noted above.
Considering the apparent difference in the activity trends against different pathogens of the ESKAPE panel, we focused on the activity against one specific pathogen, E. faecium, in nominating compounds for cyclic voltammetry experiments. To this end, we selected highly potent (1i, 1j, 2i), moderately potent (1m, 2b, 2e, 2r) and inactive (1n) compounds with respect to this pathogen. Cyclic voltammograms obtained for these compounds relative to the ferrocene external standard demonstrate a one-electron reduction (HetArNO 2 →HetArNO 2 − ) and a reversible oxidation (HetArNO 2 − →HetArNO 2 ) wave with reduction potentials (E 1/2 ) tightly grouped in the −1.3-−1.1 V range ( Figure 5). As it follows from the data collated in Table 4, there appears no apparent correlation between the activity displayed by the nine compounds investigated against E. faecium and their reduction potential. A brief glance at the activity data against the other five pathogens revealed that there is no correlation with the reduction potential either. This strongly suggests that, as noted previously, there are likely other factors at play, besides the compounds' reduction potential, which determine the observed antibacterial activity. This being said, however, one can note the generally more negative E 1/2 values observed for the series 1 compounds. All other factors being equal, this (i.e., the lower tendency of compounds 1 to undergo a one-electron reduction compared to compounds 2) could be the likely reason for the frequently observed activity loss on switching from series 2 to series 1.  As it follows from the data collated in Table 4, there appears no apparent correlation between the activity displayed by the nine compounds investigated against E. faecium and their reduction potential. A brief glance at the activity data against the other five pathogens revealed that there is no correlation with the reduction potential either. This strongly suggests that, as noted previously, there are likely other factors at play, besides the compounds' reduction potential, which determine the observed antibacterial activity. This being said, however, one can note the generally more negative E1/2 values observed for the series 1 compounds. All other factors being equal, this (i.e., the lower tendency of compounds 1 to undergo a one-electron reduction compared to compounds 2) could be the likely reason for the frequently observed activity loss on switching from series 2 to series 1. Table 4. Reduction potentials (E1/2, V) of selected compounds 1 and 2 and their MIC values (µg/mL) observed against E. faecium.

Conclusions
Starting from the structure of antimycobacterial screening hit OTB-021 which was devoid of activity against ESKAPE pathogens, we designed, synthesized and tested two mutually isomeric series of novel simplified analogs, 2-and 4-(3-nitro-1,2,4-triazol-1-yl)pyrimidines (series 1 and 2), bearing various amino side chains. These compounds demonstrated a reverse bioactivity profile being inactive against M. tuberculosis while inhibiting the growth of all ESKAPE pathogens (with variable potency patterns) except for Gram-negative P. aeruginosa. The observed inhibitory patterns allowed drawing some generalizations. In particular, frequent loss of activity on switching from series 2 to series 1 with the same substituents was noted (although, less frequently, the opposite trend was observed). Measurement of the reduction potentials (E 1/2 ) by cyclic voltammetry for compounds selected based on their activity against E. faecium revealed that all compounds investigated displayed a reversible single-electron reduction with the E 1/2 values tightly grouped in the −1.3-−1.1 V range.
No apparent correlation between the E 1/2 values and the ESKAPE minimum inhibitory concentrations was established, suggesting possible significance of other factors, besides the compounds' reduction potential, which determine the observed antibacterial activity. However, the frequent SAR trend noted above (the absence of activity for series 1 analogs while series 2 counterparts are active) correlates with the generally more negative E 1/2 values displayed by series 1. Collectively, these findings fortify the position of bioreducible nitro heteroaromatic chemotypes as antibacterial leads.

General Experimental
All commercial reagents and solvents were used without further purification, unless otherwise noted. Analytical thin-layer chromatography was carried out on UV-254 silica gel plates using appropriate eluents. Compounds were visualized with short-wavelength UV light. NMR spectroscopic data were recorded with a Bruker Avance 400 spectrometer (400.13 MHz for 1 H and 100.61 MHz for 13 C) DMSO-d 6 and were referenced to the residual solvent proton signal (2.51 ppm,) and solvent carbon signal (39.5 ppm). Melting points were determined with a Stuart SMP50 instrument in open capillary tubes. Mass spectra were recorded with a Bruker Maxis HRMS-ESI-qTOF spectrometer (electrospray ionization mode). Electrochemical measurements were performed with the Autolab PGSTAT30 (EcoChemie, The Netherlands) potentiostat/galvanostat.

Preparation of 2,4-bis(3-nitro-1H-1,2,4-triazol-1-yl)pyrimidine (4)
A round-bottom flask equipped with a magnetic stir bar was charged with 2,4-dichloropyrimidine (3 g, 20.14 mmol), potassium 3-nitro-1,2,4-triazol-1-ide (3, 6.13 g, 40.28 mmol) and dry DMF (40 mL). The reaction mixture was stirred at 80 • C for 24 h, cooled to room temperature and concentrated under reduced pressure. The residue was taken up in distilled water (100 mL) and the resulting suspension was filtered. The solid was washed with another distilled water (100 mL) and air-dried to afford the title compound A glass screw-capped vial containing a magnetic stir bar was charged with 2,4-bis(3-nitro-1H-1,2,4-triazol-1-yl)pyrimidine (4) (0.98 mmol, 300 mg), amine (1.97 mmol) and acetonitrile (3 mL). The suspension was stirred at room temperature. After completion of the reaction (TLC analysis), the solvent was removed in vacuo. The resulting solid was suspended in water (5 mL) and the suspension was placed in a fridge. After 2 h, the resulting thick precipitate was filtered off, dissolved in ethyl acetate, absorbed on silica gel (ca 0.3 g) and loaded on a silica gel chromatographic column. Elution with ethyl acetate-n-hexane (1:2) afforded an analytically pure product. (1 mg/10 mL) were prepared and diluted to a volume of 1 mL with deionized water. The resulting solutions aliquots (5 mL) were added to a Petri dish containing Muller-Hilton agar inoculated with a bacterial suspension (McFarland OD 1 /4 0.5). After drying of the compound solution, the Petri dish was incubated at 37 • C for 18 h. By measuring the bacterial growth inhibition zone diameter around the disc with ciprofloxacin or the compounds' dried solution circular spot, the susceptibility to a drug was assessed. Additionally, minimum inhibitory concentrations (MIC, µg/mL) were determined using serial broth dilutions [26].

Cyclic Voltammetry
Cyclic voltammetry studies were performed with the potential scan rate of 100 mV·s −1 in a sealed three-electrode cell at 25 • C under argon atmosphere. A flow of argon (extra-purity grade, 99.998%) was bubbled through the solutions to remove dissolved oxygen. Electrochemical measurements were performed in 10 −3 mol·dm −3 solutions of the tested compounds in 0.1 M TBAClO 4 /DMF. A typical cell used consisted of a working electrode (glassy carbon disk, 0.07 cm 2 ), a counter electrode (platinum plate, 1 cm 2 ) and a reference electrode (BAS MF-2062 Ag/0.1 M AgNO 3 solution in CH 3 CN, calibrated by 10 −3 mol ·dm −3 ferrocene external standard as a pseudo-reference electrode to comprise −188 mV referred to Fc/Fc + ). All potentials are quoted versus the above reference electrode.