Expanding the Knowledge Around Antitubercular 5-(2-aminothiazol-4-yl)isoxazole-3-carboxamides: Hit–To–Lead Optimization and Release of a Novel Antitubercular Chemotype via Scaffold Derivatization

Abstract

Tuberculosis is one of the deadliest infectious diseases in the world, and the increased number of multidrug-resistant and extensively drug-resistant strains is a reason for concern. We have previously reported a series of substituted 5-(2-aminothiazol-4-yl)isoxazole-3-carboxamides with growth inhibitory activity against Mycobacterium tuberculosis strains and low propensity to be substrate of efflux pumps. Encouraged by these preliminary results, we have undertaken a medicinal chemistry campaign to determine the metabolic fate of these compounds and to delineate a reliable body of Structure–Activity Relationships. Keeping intact the (thiazol-4-yl)isoxazole-3-carboxamide core, as it is deemed to be the pharmacophore of the molecule, we have extensively explored the structural modifications able to confer good activity and avoid rapid clearance. Also, a small set of analogues based on isostere manipulation of the 2-aminothiazole were prepared and tested, with the aim to disclose novel antitubercular chemotypes. These studies, combined, were instrumental in designing improved compounds such as 42g and 42l, escaping metabolic degradation by human liver microsomes and, at the same time, maintaining good antitubercular activity against both drug-susceptible and drug-resistant strains.

Graphical Abstract

Introduction

Mycobacterial infections are a group of multisystem diseases caused by microorganisms of the genus Mycobacterium which includes strictly aerobic, non-motile, acid-fast and rod-shaped bacilli with fastidious growth requirements.[1] There are over 190 species recognized within this genus, with very few not reported as pathogenic in humans or animals.[2] Mycobacteria are generally classified into three groups: (i) Mycobacterium tuberculosis (Mtb) complex (MTBC), including the causative agent of tuberculosis (TB); (ii) Mycobacterium leprae, the causative agent of leprosy and (iii) Nontuberculous mycobacteria (NTM).[3–5] In addition to Mtb, MTBC comprises Mycobacterium bovis, including the M. bovis bacillum Calmette-Guérin (BCG), Mycobacterium africanum, Mycobacterium caprae, Mycobacterium canetti and Mycobacterium microti, among the others. Although they share a large number of essential genes, they are widely different in terms of their host tropisms, phenotypes, and pathogenicity.[6–8] Mtb, first described by Robert Koch in 1882,[9] represents the most common etiological agent of TB, a highly contagious airborne disease. Nowadays TB is one of the top ten causes of death worldwide and the leading cause from a single infectious agent (above HIV/AIDS). It was estimated that, in 2020, TB caused about 1.3 million deaths among HIV-negative people and additional 214,000 deaths among HIV-positive people.[10,11] It is also estimated that one third of the world population is infected with latent TB, and thus at risk of developing active TB disease during their lifetime.[12] The situation is worsened by the increasing emergence of drug-resistance, which is becoming an even more severe public health issue.[13] In 2018, 51% of people with bacteriologically confirmed TB were tested positive for rifampicin resistance (RR). A total of 186,772 cases of MDR/RR-TB were detected and notified in 2018, and 156,071 cases were enrolled in treatment.[14–16] RR-TB/MDR-TB are 3.5% of new TB cases and 18% of previously treated cases had RR/MDR-TB, and the numbers are increasing. These resistant phenotypes represent an important issue as they are more difficult to eradicate and require longer therapeutic regimens and poorly tolerated drugs for their treatment.[17] The morbidity and mortality associated with TB has made its control a top priority for the World Health Organization, which in 2015 started the “END TB Strategy”,[18] a global program to fight TB worldwide. In recent years, the incidence of NTM lung diseases, mostly caused by the Mycobacterium avium and M. abscessus complexes, are on the rise, mainly in regions with low prevalence of TB, such as Europe and US.[19–21] Unlike Mtb infections, public authorities are not usually made aware of NTM diseases, thus their epidemiology is difficult to define. Furthermore, their intrinsic resistance to many available antibiotics complicates therapeutic regimens, reinforcing the need to identify novel small molecule inhibitors with activity against NTM.[22]

Finally, the COVID-19 pandemic has posed a severe menace to TB care, setting back progress in the fight against TB by several years.[23]

Considering that the antitubercular first-line drugs were developed more than half a century ago the recent approval of bedaquiline has represented an important achievement.[24] In addition, few years after bedaquiline, two compounds of the nitroimidazole class, delamanid and pretomanid were also introduced in the clinic.[25,26] Bedaquiline is an ATP synthase inhibitor, which acts by binding to the subunits c and ε of the mycobacterial F-ATP synthase enzyme, thus blocking its activity.[27,28] Moreover, when treated with bedaquiline, patients may show side effects including nausea, hepatitis and prolongation of the QT interval which is enhanced by the association with other drugs, like clofazimine or moxifloxacin. Delamanid (deltyba^®, formerly OPC-67683)[29] and pretomanid (formerly PA-824)[30] are bicyclic nitroimidazoles, originally investigated as radiosensitizers for cancer chemotherapy, but later found to show potent bactericidal activity against multidrug resistant Mtb strains. Upon reduction, they inhibit methoxy-mycolic and keto-mycolic acid synthesis via the formation of a radical intermediate between delamanid and desnitro-imidazooxazole. Severe side effects were observed, including cardiac arrhythmia and general central nervous system toxicity, especially when used in combination with isoniazid or fluoroquinolones.

In the last years, our efforts led to the discovery of several 2-aminothiazoles as effective antimycobacterial compounds through iterative medicinal chemistry efforts.[31–33] Recently, we have described the preliminary biological evaluation of a series of 2-aminothiazolyl isoxazoles as valuable antitubercular chemotypes (Figure 1).[34–36] The assays carried out on the hit compounds demonstrated that the strong bactericidal activity is accompanied by low cytotoxicity toward eukaryotic cells and high selectivity toward other microorganisms such as fungi, Gram-positive and Gram-negative bacteria. More interestingly, these compounds were found to be not affected by the action of mycobacterial efflux pumps, unlike many molecules that are currently in clinical and preclinical evaluation.[33] Although this characteristic was considered particularly interesting as it anticipates low propensity to generate resistant phenotypes, these compounds suffered from sub-optimal metabolic stability in an in vitro assay, with a half-life in human liver microsomes (HLM) of 16.1±0.6 and 35±0.8 min for compounds 3 and 4, respectively. These premises prompted us to start a hit-to-lead optimization process via a convergent Structure-Activity and Structure-Metabolism relationships (SAR and SMR) approach to identify structurally related analogues containing either the 2-aminothiazole feature or its bioisosteres.

Figure 1.

Antitubercular 2-aminothiazoles and planned modifications to explore the SAR

Results and Discussion

In vitro metabolism of compounds 3 and 4.

Compounds 3 and 4, previously reported by us, were considered valuable antitubercular chemotypes not only by virtue of their potent bactericidal activity and low cytotoxicity, but especially for their improved stability to hepatic metabolism over compound 2 (Figure 1).[33]

The highly labile ester moiety of compound 2 was indeed cleaved after only one minute in HLM, to give the corresponding acid derivative that was found to be inactive. The replacement of the ester moiety with a substituted arylamide led to more stable hit compounds, although the half-life was still considered suboptimal. This prompted us to further investigate the metabolism of compounds 3 and 4 in HLM. Three metabolites were detected for each compound and characterized by High Resolution Mass Spectrometry (HR-MS) and tandem mass spectrometry (MS/MS) analysis. Two of them, M1 and M2, showed molecular mass M = M+16 [M+O] with respect to the parent compound and they were indicated, in descending order of abundance, M1_cpd3, M2_cpd3 if derived from compound 3 and M1_cpd4, M2_cpd4 if derived from compound 4 (Table 1).

Table 1.

Metabolism of compounds 3 and 4 in HLM

Cpd	RT (min)	Formula	Ion	Experimental mass (m/z)	Calculated mass (m/z)	Mass Error (ppm)	Major MS/MS fragments	Structure Assigned
3	28.17	C20H16N4O2S	[M+H]+	377.1075	377.1067	2.30	258.1, 217.2
4	26.81	C19H15N5O2S	[M+H]+	378.1026	378.1019	1.87	360.1, 284.0, 258.1, 217.2
M1cpd3	22.95	C20H16N4O3S	[M+H]+	393.1023	393.1016	1.71	375.1, 256.1
M2cpd3	25.44	C20H16N4O3S	[M+H]+	393.1024	393.1016	2.19	274.1, 251.1, 232.9
M3cpd3	23.65	C20H14N4O4S	[M+H]+	407.0815	407.0809	1.69	389.1, 288.1, 270.0, 247.1	[Mcpd3+2O-2H]
M1cpd4	21.01	C19H15N5O3S	[M+H]+	394.0975	394.0968	1.68	376.2, 256.2
M2cpd4	23.73	C19H15N5O3S	[M+H]+	394.0975	394.0968	1.68	376.1, 300.1, 274.1
M3cpd4	21.97	C19H13N5O4S	[M+H]+	408.0769	408.0761	1.96	390.1, 288.1, 270.0, 121.1	[Mcpd4+2O-2H]

Furthermore, a third metabolite with a molecular mass M = M+30 [M+2O−2H] was detected for each parent compound and designated M3_cpd3 or M3_cpd4. After this preliminary analysis, hydroxylation seemed to be the main phase I metabolic modification, and therefore M1 and M2 for both series were characterized. Amide hydrolysis was ruled out, confirming the value of replacement of the ester moiety. Although the nature of the metabolic modification to compounds 3 and 4 could be easily identified by HR-MS, the exact site of hydroxylation could not be predicted. Therefore, the identification of the metabolites was accomplished via the synthesis of a panel of plausible metabolites of the parent compounds, followed by the comparison of the features of the synthesized molecules with those of the corresponding metabolites in the HLM incubates. The comparative identification was carried out analyzing the following features: a) High Resolution (HR) mass value, as derived from HR-MS analysis; b) MS/MS spectrum; c) Retention Time (RT) in HPLC runs; d) Co-elution of the standard compounds with the putative metabolites in the HLM incubates. Based on the analysis of the MS/MS spectra, and considering that the methyl group of toluene might be hydroxylated during phase I metabolism, the synthesis of the potential metabolites was mainly focused on structures bearing the hydroxyl group at the aromatic ring attached to the aminothiazole (42–43e, 42–43f, 42–43n). However, in the case of compound 3, since the analysis of MS/MS spectra could not rule out hydroxylation of the amide portion, also compounds 20a, 21a, and 22a were included in the panel of potential metabolites. It should be considered that the synthesis of these derivatives was important not only to elucidate the metabolic fate of compounds 3 and 4, but also to extend the knowledge around the SAR of these 5-(2-aminothiazol-4-yl)isoxazole-3-carboxamides, as it will be described in the following paragraph. Nine hydroxylated derivatives were synthesized and evaluated as described above (Table 1).

Metabolite M1_cpd4 was eluted at 21.01 min in the ESI⁺ trace (Fig. 2B) with a [M+H]⁺ ion at m/z =394.0975, corresponding to the elemental composition reported in Table 2 with a mass error of 1.68 ppm (see also Figure S6A, Supporting Information). The characteristic product ions in the ESI⁺ MS/MS spectrum (Table 2 and Figure S6B) were at m/z = 376.2 (elimination of water) and m/z = 256.2 due to the cleavage of pyridin-2-yl-aminocarbonyl fragment followed by elimination of water. M1_cpd4 was synthesized following the hypothesis that it corresponded to the hydroxylation product on the methyl group of toluene ring. MS/MS spectrum of M1_cpd4 matched that of the obtained chemical standard (i.e. compound 43n, Figure S6C). Identity of metabolite M1_cpd4 was further confirmed, verifying the co-elution with 43n in LC-MS/MS runs by spiking experiments in HLM incubates (Figure S6D).

Figure 2.

A. LC-HR-MS Extracted Ion Chromatograms (XIC) in ESI⁺ of compound 3 (RT = 28.17 min) in HLM at t=3h showing the peaks corresponding to metabolites M1_cpd3 (RT = 22.95 min), M2_cpd3 (RT = 25.44 min) and M3_cpd3 (RT = 23.65 min). B. XIC of compound 4 (RT = 26.81 min) in HLM at t=3h showing the peaks corresponding to metabolites M1_cpd4 (RT = 21.01 min), M2_cpd4 (RT = 23.73 min) and M3_cpd4 (RT = 21.96 min).

Table 2.

Antitubercular activity, cytotoxicity and metabolic stability of compounds synthesized

Cmpd	Ring A	R	R1	X	MIC90a μg/mL (μM) MtbH37Rv	MIC90b μg/mL (μM) M. Marinum		IC50c μg/mL (μM)	HLM stabilityd (min)
3	Phenyl	H	4-CH3	S	1.0(2.7)	0.5(1.3)	88(234)		t1/2 = 35 ± 1
4	2-Pyridine	H	4-CH3	S	0.5(1.3)	0.5(1.3)	22.1(59)		t1/2 = 16 ± 1
9a	Phenyl	4-F	4-CH3	S	1.0(2.5)	nde	46.8(118)		nd
10a	Phenyl	3-F	4-CH3	S	2.5(6.3)	nd	nd		nd
11a	Phenyl	2-F	4-CH3	S	1.0(2.5)	nd	33.9(86)		nd
12a	Phenyl	4-Cl	4-CH3	S	>40	nd	nd		nd
13a	Phenyl	3-CF3	4-CH3	S	>20	nd	nd		nd
14a	Phenyl	4-CH3	4-CH3	S	1.0(2.6)	nd	47.5(122)		nd
15a	Phenyl	3-CH3	4-CH3	S	0.5(1.3)	nd	10.7(27)		nd
16a	Phenyl	2-CH3	4-CH3	S	1.0(2.5)	nd	63.7(163)		nd
17a	Phenyl	3,5-CH3	4-CH3	S	>20	nd	nd		nd
18a	Phenyl	2,4-CH3	4-CH3	S	>20	nd	nd		nd
19a	Phenyl	3-OEt	4-CH3	S	>20	nd	nd		nd
20a	Phenyl	4-OH	4-CH3	S	0.5(1.3)	1(2.6)	79.1(202)		nd
21a	Phenyl	3-OH	4-CH3	S	0.5(1.3)	nd	48.7(124)		nd
22a	Phenyl	2-OH	4-CH3	S	0.5(1.3)	nd	53.1(135)		nd
23a	4-Pyridine	H	4-CH3	S	1.0(2.6)	nd	29.9(79)		nd
24a	3-Pyridine	H	4-CH3	S	10(26)	nd	nd		nd
25a	2-Pyridine	3-CH3	4-CH3	S	>20	nd	nd		nd
26a	2-Pyridine	6-CH3	4-CH3	S	0.5(1.3)	1.0(2.6)	66.8(171)		nd
27a	2-Pyridine	4-CH3	4-CH3	S	1.0(2.6)	nd	86.3(221)		nd
28a	Phenyl	2-Cl, 4-CH3	4-CH3	S	>20	nd	nd		nd
29a	4-Pyridazine	H	4-CH3	S	>20	nd	nd		nd
30a	2-Thiazole	H	4-CH3	S	>40	nd	nd		nd
31a	2-Thiazole	5-CH3	4-CH3	S	>20	nd	nd		nd
32a	3-Thiophene	H	4-CH3	S	1.0(2.6)	1.0(2.6)	>100		nd
33a	(Thiophen-2-yl)methyl	H	4-CH3	S	>20	nd	nd		nd
35b f	Phenyl	H	3,5-Cl	S	>20	nd	nd		nd
36b f	2-Pyridine	H	3,5-Cl	S	>20	nd	nd		nd
35c	Phenyl	H	3,5-F	S	>20	nd	nd		nd
36c	2-Pyridine	H	3,5-F	S	>20	nd	nd		nd
35d	Phenyl	H	3,5-CH3	S	1.0(2.5)	nd	70.9(175)		nd
36d	2-Pyridine	H	3,5-CH3	S	>20	nd	nd		nd
20e	Phenyl	4-OH	3-OH,4-CH3	S	5.0(12.2)	nd	nd		t1/2 = 19 ± 1
42e	Phenyl	H	3-OH,4-CH3	S	2.5(6.3)	nd	>100		nd
43e	2-Pyridine	H	3-OH,4-CH3	S	5(12.6)	nd	>100		nd
42f	Phenyl	H	2-OH,4-CH3	S	nd	nd	nd		nd
43f	2-Pyridine	H	2-OH,4-CH3	S	>20	nd	nd		nd
42g	Phenyl	H	H	S	1.25(3.4)	nd	11.9(33)		t1/2 = 43 ± 3
43g	2-Pyridine	H	H	S	>20	nd	nd		t1/2 = 66 ± 9
42h	Phenyl	H	4-CF3	S	>20	nd	nd		71 ± 7 %
43h	2-Pyridine	H	4-CF3	S	>20	nd	nd		85 ± 6 %
42j	Phenyl	H	3,4,5-Cl	S	>20	nd	nd		101 ± 1 %
43j	2-Pyridine	H	3,4,5-Cl	S	>20	nd	nd		102 ± 6 %
42l	Phenyl	H	4-Cl	S	1.25(3.2)	2(6.0)	>100		t1/2 = 59 ± 9
42m	Phenyl	H	3-OH	S	5.0(13.2)	nd	nd		t1/2 = 34 ± 1
42n	Phenyl	H	4-CH2OH	S	>20	nd	nd		nd
43n	2-Pyridine	H	4-CH2OH	S	>20	nd	nd		nd
47	Phenyl	H	4-CH3	O	1.0(2.8)	nd	>100		nd
48 g	2-Pyridine	H	4-CH3	O	2.5(6.9)	nd	70.9(197)		nd
34a					>20	nd	>100		nd
42i					>20	nd	nd		97 ± 12 %
43i					>20	nd	nd		96 ± 14 %
53					1.25(3.8)	nd	>100		nd
SM h					0.25	nd	nd		nd
INH i					0.025	nd	nd		nd

^aMinimum Inhibitory Concentration, determined by REMA method toward M. tuberculosis H37Rv;[69]

^bMinimum Inhibitory Concentration, determined toward M. marinum ATCC BAA-535;

^cMTT assay has been used to assess the cytotoxicity of synthetized compounds toward Human monocytes THP-I cells;

^din vitro metabolic stability in HLM was alternatively reported as half-life (t_1/2) in min or as residual fraction (%) at 60 min if % compound at t=60 min was > 50%. Reported are Means ± S.E.M. (n=3).

^eNot Determined;

^fActivity for these compounds has already been reported;³²

^gActivity for this compound has already been reported;⁴⁹

^hStreptomycin;

ⁱIsoniazide.

Metabolite M2_cpd4 was eluted at 23.73 min in the ESI⁺ trace (Fig. 2B) with a [M+H]⁺ ion at m/z = 394.0975, corresponding to the elemental composition reported in Table 2 with a mass error of 1.68 ppm (See also Figure S7A, Supporting Information). The characteristic product ions in the ESI⁺ MS/MS spectrum (Figure S7B) were at m/z = 376.1 (elimination of water) and at m/z = 300.1 and m/z = 274.1 due to the subsequent cleavage of the 2-aminopyridyl and carbonyl fragments from M2_cpd4 structure. Two standards, 43e and 43f, were synthesized, bearing the hydroxyl group at the ortho and meta positions with respect to the methyl group of the toluene ring and their MS/MS spectra and chromatographic retention times were compared to those of M2_cpd4. While comparison of MS/MS spectra was not conclusive, since both 43e and 43f returned the same signals (Figure S7C–D), a 120 min LC-MS gradient (see experimental part for details) ultimately enabled the identification of M2_cpd4 as 43e, via the spiking of the chemical standard in the HLM incubate and subsequent co-elution with M2_cpd4 (Figure S7E–F). Metabolite M3_cpd4 was eluted at 21.96 min in the ESI⁺ trace (Fig. 2B) with a [M+H]⁺ ion at m/z = 408.0769, corresponding to the elemental composition reported in Table 2 with a mass error of 1.96 ppm and a Δm = +30 with respect to parent compound 4 (See also Figure S8A, Supporting Information). This mass shift is compatible with the complete oxidation of the methyl group of the toluene ring to carboxylic acid, as this would be consistent with the mass, with the position of the peak in the LC traces and with the scientific literature reporting this biotransformation. The characteristic product ions in the ESI⁺ MS/MS spectrum (Figure S8B) were at m/z = 390.1 and m/z = 270.0 (elimination of water and subsequent cleavage of the pyridin-2-yl-carboxyamide fragment), at m/z = 288.0 (cleavage of pyridin-2-yl-aminocarbonyl fragment from M3_cpd4 structure) and m/z = 121.1 due to the ionization of the pyridin-2-yl-aminocarbonyl fragment.

The determination of the metabolic sites of the parent compound 3 confirmed the hypothesis that the toluene moiety was the region affected by the oxidative phase I metabolism. Following the same approach, M1_cpd3 and M2_cpd3 were identified and, analogously to the results obtained for compound 4, the major metabolite M1_cpd3 was found to be the hydroxylated derivative on the methyl group (42n) and the minor metabolite M2_cpd3 corresponded to the hydroxylated derivative on the ortho position with respect to the methyl group of the toluene (42e). The complete characterization of metabolites M1_cpd3 and M2_cpd3 in terms of accurate mass, MS/MS fragmentation schemes and related comparisons with MS/MS and chromatographic features of synthetic standards 42n and 42e are reported in the Supporting Information (Figures S3A–D and Figures S4A–D for M1_cpd3-42n and M2_cpd3-42e, respectively). Metabolites M1_cpd3, M2_cpd3, M1_cpd4 and M2_cpd4, corresponding to compounds 42n, 42e, 43n, and 43e, respectively, were tested for their antitubercular activity against wild type Mtb H37Rv, and although the most abundant metabolites 42n and 43n displayed no whole cell activity, the less abundant metabolites 42e and 43e still retained a certain activity (see the next paragraph for comments on the SAR). Since phase I metabolism has a significant impact on the activity of this series of derivatives, hampering their future development, a panel of derivatives with predicted improved metabolic stability was designed and synthesized, according to two different strategies: (i) removing the site of metabolism, (ii) capping the metabolic soft-spots with halogens. A total of 8 compounds were synthesized and their in vitro metabolic stability was measured in HLM (Table 1). Interestingly, all the novel molecules synthesized showed a significant improvement in metabolic stability as compared to the parent compounds 3 and 4 and, for the majority of derivatives, the residual fraction at the last time point of HLM incubation (i.e. 60 min) was above 50%.

More in detail, the removal of methyl group of toluene moiety on both series led to a half-life (t_1/2) in HLM of 42g, 43g which was relatively longer than that of corresponding parent compounds 3 and 4. Substitution of methyl with a trifluoromethyl group (42h, 43h), a well-known strategy of metabolic protection, also significantly improved metabolic stability with percentages of compound recovered at t=60 min equal to 71% and 85%, respectively. The most stable derivatives were those in which either the toluene moiety was completely removed (42i, 43i) or the three metabolic spots on the toluene ring were capped with chlorine atoms (42j, 43j). In both reported modifications and for both tested series, all the novel metabolically protected compounds did not show any significant biotransformation in HLM for the tested period.

In conclusion, the analysis of the in vitro metabolism for compounds 3 and 4 enabled the identification of two metabolic soft-spots and revealed that hydroxylation was the major phase I metabolic transformation. For both parent compounds, the major metabolic transformation, i.e hydroxylation of the methyl group of the toluene ring, resulted in a loss of antitubercular activity (42n, 43n, MIC₉₀ >20 μg/mL), whereas, encouragingly, the minor metabolites 42e and 43e retained moderate whole cell activity, displaying MIC₉₀ values of 2.5 μg/mL and 5 μg/mL, respectively. To further corroborate our findings, a small panel of derivatives was designed and synthesized, and all of them showed significantly improved metabolic stability profiles, despite failing to show measurable MICs, as reported in the next paragraph.

Investigation of the Structure-Activity Relationships.

A total of 63 compounds were prepared and evaluated to determine the Minimum Inhibitory Concentration required to inhibit the growth of 90% of Mtb_H37Rv strains (MIC₉₀, Table 1 and 2). To eliminate the possibility that the antitubercular activity arises from general toxicity, THP-derived macrophages were used for an in vitro cytotoxicity evaluation. Based on the limited set of information around the hit compounds 3 and 4,[33] we decided to undertake 3 different design strategies to investigate the SAR. From one side, we focused on the central 5-(2-aminothiazol-4-yl)isoxazole-3-carboxamide scaffold, that was kept intact throughout the chemical manipulation as considered a plausible pharmacophore. 2-aminothiazoles have been considered privileged scaffolds in the field of medicinal chemistry since a long time, as they are widely distributed both in natural and synthetic compounds and are thought to confer many therapeutic and pharmacological effects. For instance, the 2-aminothiazole core has been used to adorn cefdinir, a semi-synthetic third generation cephalosporin used for treatment of pneumonia, among the others.[37] The 2-aminothiazole ring not only affects the pharmacodynamic properties of cefdinir, but it can also influence its kinetics, as it has been speculated that in the digestive tract iron(II) chelates with the thiazole ring and restricts gastrointestinal absorption. Therefore, it is reasonable to have such a dynamic feature in the architecture of a biologically active molecule. In addition, both the 2-aminothiazole and the isoxazole-3-carboxylate scaffold have demonstrated to be privileged structures in the search of potent antitubercular compounds.[38–43]

The second approach consisted in changing the structural arrangement of atoms of hit compounds 3 and 4. In this case, the thiazole ring is broken, and the sulfur atom connected to the benzene ring to give a 2-aminobenzothiazole ring (Figure 1), variously substituted, on turn connected through ad amide bond to the C5 of the isoxazole-3-carboxylate ring, which maintains its pattern of substituents as in 3 and 4.

As the last approach, a small set of 2-aminothiazole bioisosteres was synthesized. Hinsberg has extended the concept of isosterism and bioisosterism to heterocycles according to the “ring equivalence” theory, based on which benzene, thiophene, pyridine and furan can be considered ring equivalents.[44,45] Therefore, compounds 47 and 48, where the sulfur atom of the thiazole ring is substituted by oxygen, and compound 53, where the whole thiazole ring is substituted by a pyridine, were synthesized and tested,

Considering the first approach, most of the modifications were made either at the amide side or at the aniline side, where variously substituted aromatic and heteroaromatic rings were investigated to evaluate their impact on the antitubercular activity (Figure 1). This led to the implementation of two different synthetic approaches, in order to optimize the synthesis of a consistent number of molecules starting from one common synthon. Concerning the amide side, variously substituted anilines (9–22a, 28a) and aminopyridines (23–27a) were coupled to the isoxazole-3-carboxylic moiety to give the corresponding secondary amides. In addition, the effect of substitution with different aromatic rings, such as thiazole (30a), 5-methylthiazole (31a), thiophene (32a, 33a), and pyridazine (29a) was also evaluated, as some of them are isosteres of benzene and pyridine as well as privileged scaffolds in medicinal chemistry. In one case, a tertiary amide was prepared by coupling the carboxylic moiety with pyrrolidine (34a). Anilines and aminopyridines were adorned with small functional groups chosen based on properties such as lipophilicity, electronic character, and steric hindrance. In general, lipophilic substituents were preferred over hydrophilic ones, as a certain degree of lipophilicity has proven to be beneficial for the antitubercular activity.[46,47] In addition, also the electronic effect of the substituents was taken into account, exploring both electron-donor (methyl, dimethyl, ethoxy, hydroxyl) and electron-withdrawing (fluorine, chlorine) groups (EDGs and EWGs). In a similar vein, modifications at the aniline side were investigated as well, although to a lesser extent.

When the phenyl or the pyridine ring of the hit compounds 3 and 4 is substituted with other five- and six-membered heterocycles, loss of the antimycobacterial activity was noticed (29a, 31a, 33a, MIC₉₀ >20 μg/mL and 30a, MIC₉₀ >40 μg/mL), as well as in the case of the tertiary amide 34a (MIC₉₀ >20 μg/mL). On the other hand, substitution with thiophene (32a, MIC₉₀ = 1 μg/mL), a benzene bioisostere, was found not only to maintain a remarkable antitubercular activity, but also to improve the cytotoxic profile (IC₅₀ >100 μg/mL). Small EWGs such as fluorine, attached either at the ortho, meta or para position of the benzene of the amide side, were able to maintain good activity toward Mtb (9a, 11a, MIC₉₀ = 1 μg/mL, 10a MIC₉₀ = 2.5 μg/mL). Other EWGs such as p-chlorine (12a, MIC₉₀ >40 μg/mL) and m-trifluoromethyl (13a, MIC₉₀ >20 μg/mL) led to a sharp decrease of the mycobactericidal activity. It can be speculated that EWGs are well tolerated as substituents at this position, although size constraints must be taken into consideration. Standard EDGs such as the methyl group facilitated maintenance of potent antimycobacterial activity (14a, 16a, MIC₉₀ = 1 μg/mL, 15a, MIC₉₀ = 0.5 μg/mL), with compound 15a resulting one of the most active of the series, although also one of the most toxic to THP-derived macrophages (IC₅₀ = 10.7 μg/mL). Attachment of the bulkier ethoxy substituent to the phenyl ring (19a, MIC₉₀ >20 μg/mL) resulted in a loss of activity, as did di-substitution with methyl (compounds 17a and 18a, MIC₉₀ >20 μg/mL) or with a combination of EWG (chlorine) and EDG (methyl) (28a, MIC₉₀ >20 μg/mL). Altogether, these findings led to the speculation that the hindrance of the substituents at the benzene of the amide side is more relevant than their electronic nature (EWGs and EDGs). Since small EDGs were well tolerated, hydroxyl functional groups were introduced at the ortho, meta and para position of the benzene. It must be noted that, in general, the introduction of polar groups that reduce the polar surface area are not advised in the design of antitubercular agents, as they must penetrate the thick and waxy Mtb cell wall.[46,47] On the other hand, substitution with a hydroxyl group could improve the overall drug-likeness of the molecules, and the role of its hydrogen-bond donor/acceptor ability in the interaction with the active site of the target, which is as yet unknown. Encouragingly, all of the compounds bearing a hydroxyl group were able to maintain the antitubercular activity, resulting in the most active compounds of the series (20–22a, MIC₉₀ = 0.5 μg/mL). The activity of these compounds might be rationalized by the existence of a hydrophobic pocket in the target binding site, located within the vicinity of the amide moiety, however this hypothesis can only be confirmed after the identification of the protein inhibited by this class of compounds. Regarding the pyridine ring, the first attempt was to move the pyridine nitrogen at different positions of the ring, specifically at position 3 and 4. Surprisingly, a substantial difference in activity could be noticed depending on the position of the nitrogen atom within the ring. When it is placed at position 2 or 4, (4 = 0.5 μg/mL, 23a, MIC₉₀ = 1 μg/mL), the activity is maintained. However, when the nitrogen atom is at position 3, a sharp drop in the activity is noticed (24a, MIC₉₀ = 10 μg/mL). It is not simple to assess the collocation of this result within the SAR, however, it seems that the electronic effect given by the nitrogen atom plays a role in the antitubercular potency. Next, two advantageous characteristics such as the methyl group and the nitrogen at position 2 were investigated. Compound 26a (MIC₉₀ = 0.5 μg/mL) was one of the most active of the series, and its regioisomer 27a (MIC₉₀ = 1 μg/mL), where the methyl group is moved from position C3 to C4, retained considerable activity, supporting the rationale of the design. Surprisingly, compound 25a (MIC₉₀ >20 μg/mL) was devoid of appreciable activity.

The SAR at the amide side indicates that, in general, five- or six-membered heterocycles other than benzene or pyridine are somewhat detrimental to the activity. EDGs and EWGs are well tolerated at any position of the ring, provided that they are sufficiently small in size. This is likely why di-substitution leads to loss of activity. Surprisingly, polar groups such as the hydroxyl were found to confer high antitubercular activity, coupled with a net improvement of the drug-likeness of the molecule. It is likely that the additional pattern of interactions established by this group might counterbalance the decrease in lipophilicity, a characteristic that accompanies the design of many antitubercular agents.[48,49]

Concerning the substituents at the aniline side, we were surprised to find out that the majority of the modifications investigated led to a sharp decrease in the activity, either when at the amide side the phenyl or the 2-pyridinyl appendage are present. Removal of the p-methyl (43g, MIC₉₀ >20 μg/mL) or its substitution either with a trifluoromethyl (42h, 43h MIC₉₀ >20 μg/mL) or a hydroxymethyl moiety (42n, 43n, MIC₉₀ >20 μg/mL) led to loss of activity. Surprisingly, compound 42g was the only exception with an encouraging MIC₉₀ of 1.25 μg/mL, although coupled to a relatively high toxicity (IC₅₀ = 11.9 μg/mL). Removal of the p-methyl and mono-, di- and tri-substitution patterns with EWGs such as the fluorine and the chlorine (35b, 36b, 35c, 36c, 42j, 43j, MIC₉₀ >20 μg/mL), led to loss of activity. However, when a chlorine atom is inserted at the C3 position (42l, MIC₉₀ = 1.25 μg/mL), high activity is obtained coupled to irrelevant cytotoxicity. This result was particularly important, especially considering the higher metabolic stability of 42l as compared to the corresponding hit compound 3. Concerning the substitution with EDGs, hydroxylation at position C3 is generally well tolerated (42e, MIC₉₀ = 2.5 μg/mL 43e, MIC₉₀ = 5 μg/mL, 42m, MIC₉₀ = 5 μg/mL), with and without the presence of the p-methyl group. Moreover, this modification seems also to confer an improved safety profile (42e, 43e, IC₅₀ >100 μg/mL), and this result is even more significant taking into account that 42e and 43e are among the metabolites of hit compounds 3 and 4. However, when the hydroxylation occurred at position C2, activity was lost (43f, MIC₉₀ >20 μg/mL). Surprisingly, when two methyl groups were introduced at positions C3 and C5 of the benzene, the activity was maintained only if the phenyl (35d, MIC₉₀ = 1 μg/mL), but not the 2-pyridinyl (36d, MIC₉₀ >20 μg/mL), was present at the amide side. As expected, removal of the phenyl ring at the nitrogen of the 2-aminothiazole led to loss of the antitubercular potency (42i, 43i, MIC₉₀ >20 μg/mL). Interestingly, we observed that the activity of compounds 47 and 48, characterized by the 2-aminooxazole heterocycle in place of the 2-aminothiazole, was comparable to that of the parent compounds (47, MIC₉₀ = 1.0 μg/mL and 48, MIC₉₀ = 2.5 μg/mL). This might be considered a further confirmation of the ring equivalence between the 2-aminothiazole and the 2-aminooxazole. In addition, these derivatives were found to possess a decreased cytotoxic activity, confirming our previous findings about 2-aminooxazole as a novel privileged scaffold in medicinal chemistry.[50] In a similar vein, compound 53, where the thiazole ring is substituted with a pyridine and the same substitution pattern of parent compound 2 is maintained, (Figure 1) was synthesized and tested. We were pleased to notice that compound 53 had an extremely encouraging antitubercular activity (MIC₉₀ 1.25 μg/mL), and its SAR evaluation and biological characterization is currently under consideration and will be reported in a separated publication. Overall, the design strategy pivoting on the exploitation of ring bioisosterism led to the disclosure of 5-(pyridin-2-yl)isoxazole-3-carboxylate as a novel antitubercular chemotype.

Merging the set of information above reported, it was possible to plan the design and synthesis of compound 20e, where the introduction of two hydroxyl moieties might be beneficial for the general drug-likeness of the molecule. Indeed, compound 20e maintained good activity (20e, MIC₉₀ = 5 μg/mL), confirming the rationale of the design. However, the presence of the hydroxyl group at the C3 did not prevent the metabolic oxidation of the vicinal methyl group, resulting again in a suboptimal predicted half-life. Altogether, these results indicate that, despite the metabolic liability, the p-methyl at the aniline side is important to confer acceptable antitubercular potency. However, in one case, its removal and the presence of a chlorine atom at the position C3 led to a novel lead compound (42l) with good activity, absent toxicity and improved metabolic stability, suggesting how this portion of the molecule is worthy of further investigation for its role in modulating potency and pharmacological characteristics.

Quite surprisingly, all the molecules where the thiazole ring is constrained in a benzothiazole structure (58-63a, 58-61b and 60c), were found to be inactive (Table 2). Since the biological target of these molecules is still unknown, it might be only speculated that enlarging the size of planar surface of the 2-aminothiazole moiety has a detrimental impact on the recognition of the molecule for the target binding site.

In order to gain insights into the mechanism of action/resistance of this class of compounds, one of the most promising compounds (26a, MIC₉₀ = 0.5 μg/mL, IC₅₀ = 66.8 μg/mL) was tested against a panel of 10 Mtb mutants harboring different mutations in genes encoding for drug targets (DprE1, MmpL3, CoaA, and PyrG),[51–54] activators (EthA and Rv2466c),[55] inactivator (Rv3405c)[56] and mechanism of drug resistance (Rv0579).[57] Furthermore, the activity of this compound against two Mtb MDR clinical isolates (IC1 and IC2) was evaluated.[58]

Encouragingly, 26a was found to be very active against the two MDR clinical strains as well as against all the mutants tested, suggesting a novel mechanism of action with respect to most of the drugs used in therapy or under evaluation in the clinical studies (Table 3). Given the propensity for novel small molecule inhibitors with antitubercular activity to translate into activity against NTM[22], selected compounds (9a, 10a, 15a, 16a, 20a, 21a, 23a, 27a, 35d, 43e, 43g, 47) were evaluated for their ability to inhibit growth of clinically relevant NTM species, namely M. abscessus Bamboo[59] and M. avium 11.[60] None of the compounds tested displayed significant activity against either species at up to 100 μM. However, 27a was identified as the only compound displaying growth inhibitory activity approaching 50% (MIC₅₀) against both at the highest concentrations tested (Figure S9 and S10), whereas compounds 42g and 43g, although weak, displayed a concentration-dependent inhibitory activity especially against M. abscessus Bamboo.

Table 3.

Antitubercular activity, cytotoxicity and metabolic stability of compounds synthesized

Cmpd	R	X	R1	R2	MIC90a (μg/mL) MtbH37Rv
58a	H	C	H	H	>40
59a	H	N	H	H	>40
60a	3-CH3	C	H	H	>40
61a	4-OH	C	H	H	>40
62a	3-F	C	H	H	>40
63a	3-OH	C	H	H	>40
58b	H	N	H	CH3	>40
59b	H	C	H	CH3	>40
60b	3-CH3	C	H	CH3	>40
61b	4-OH	C	H	CH3	>40
60c	3-CH3	C	4-OCH3	H	>40
SM b					0.25

^aMinimum Inhibitory Concentration, determined by REMA method toward M. tuberculosis H37Rv;[69]

^bStreptomycin.

In order to assess in vivo efficacy and to consolidate the finding of improved metabolic stability, most promising compounds were assessed in the zebrafish larvae model of M. marinum infection. We initially evaluated the MIC of most interesting compounds (3, 4, 20a, 26a, 32a, 42l) against M. marinum strain M (ATCC BAA-535) and M. smegmatis mc2155 and we also assessed potential toxicity issues in vivo (zebrafish embryos). Encouragingly, the tested compounds displayed a pronounced inhibitory activity against the multi-drug resistant M. marinum strain (table 1) whereas many compounds active toward Mtb fail in this task. Compounds were not active up to >64 μg/mL toward M. smegmatis mc2155 (data not shown).

We then assessed the developmental and acute toxicity of the compounds on zebrafish embryos 1 day post-fertilization and treated them for a period of 4 days. We identified three compounds (3, 26a and 32a), for which the maximum tolerated concentrations (MTC) appeared to be in a favourable range and the embryos tolerated concentrations of 25 and 50 μg/mL in the surrounding medium. However, it must be noticed that MTC could not be uniquely determined as all compounds partly precipitated when added to the egg water (Figure S11). Based on the poor solubility of the compounds we tried to administer them via microinjection into the caudal vein of M. marinum infected zebrafish larvae (Figure S12). However, also in this setup compounds precipitated and clogged the microneedle. Nevertheless, it was possible to treat the infected larvae with compound 3 at low concentration (ca. 6–8 ng), observing that – by tendency – granuloma formation was slightly less pronounced than in the untreated control group (Figure S12). Due to the technical difficulties related to the poor aqueous solubility we were not able to generate a valid and statistically relevant data set for the compound series in the zebrafish larvae model of M. marinum infection; however these preliminary data encourage further investigation with optimized delivery media.

Despite the possibility that 2-aminothiazoles may cause a false positive interacting with a large number of proteins, and taking into consideration the call made by some editors of the ACS journals with regard to the early identification of PAINS,[61,62] we would like to point out that the data described herein rule out the possibility of an unselective and unspecific mechanism of action. Indeed, the compounds in this study were tested toward different cellular lines, and also in the presence of liver subcellular fractions (microsomes), yielding a wide range of activities and allowing an analysis of their SAR. Moreover, compounds 3 and 4 have already been tested for their reactivity to glutathione, ruling out the possibility of indiscriminate reactivity toward biological nucleophiles.[50] Finally, some of these derivatives were not recognized as PAINS using the software “False Positive Remover” (http://www.cbligand.org/PAINS/) nor as aggregator according to the software “Aggregator Advisor” (http://advisor.bkslab.org/). All of the above findings demonstrate how this series of antitubercular compounds show biological properties that can be considered particularly appealing, especially when compared to the current therapeutic arsenal. In particular, the activity against single drug-resistant, multidrug resistant and extensively drug resistant mycobacterial strains, the high selectivity, and the low propensity to be substrate of efflux pumps make these derivatives particularly suitable for additional investigation.

Chemistry.

The majority of the compounds (9–34a, 35–36b, 35–36c, 35–36d and 20e) were obtained following an established protocol based on the Hantzsch thiazole synthesis. Commercially available isoxazole 5 was brominated to obtain the α-bromoketone 6 which undergoes Hantzsch condensation with the appropriate thioureas, either commercially available or synthesized in one step from the corresponding anilines. The key intermediates 7a-e obtained were hydrolyzed with LiOH in THF/MeOH/H₂O at room temperature to give the corresponding carboxylic acids, which were then reacted with the suitable amine, generally using TBTU and EDC as the coupling reagents.[63] Compounds 42e-m and 43e-k were synthesized with the same pattern of chemical reactions, but in a different order compared to 9–34a, 35–36b, 35–36c, 35–36d and 20e previously described. This allowed to obtain key intermediates 40 and 41, that could be condensed with various thioureas thus expanding the series for SAR recognition. In this regard, condensation of intermediate 40 with 1-(4-(hydroxymethyl)phenyl)thiourea n in EtOH at reflux failed to give the desired compound 42n, as ethyl etherification of the hydroxyl group occurred as side reaction (data not shown). Likely, the high temperature, the large excess of EtOH as nucleophile, and the acidic conditions generated by the release of HBr favored the formation of the ethyl ether. To avoid this, the thiourea n was first protected on the hydroxyl group with tert-butyl(chloro)diphenylsilane to give 1-(4-(((tert-butyldiphenylsilyl)oxy)methyl)phenyl)thiourea k, that smoothly reacted with α-bromoketones 40 and 41 to give thiazoles 42–43k. Deprotection of the hydroxyls with TBAF on silica gel in dry DMF at 0 °C allowed to obtain the title compounds 42–43n in good yields. Because of the different reactivity of the oxygen atom with respect to sulfur, the compounds 47 and 48 could not be obtained according standard Hantzsch condensation protocol with substituted ureas. Thus, the synthesis of 47 and 48 was carried out using a synthetic protocol previously developed and optimized in our research group.[50] Briefly, α-bromoketone 6 was condensed with urea and the resulting 2-aminothiazole 44 underwent Buchwald-Hartwig coupling reaction to give the key ester 45 in moderate yields. Ester hydrolysis with LiOH in THF/MeOH/H₂O at room temperature and reaction with suitable amine in the conditions described above led to the desired compounds 47 and 48. Compound 53 was prepared in part following a reported protocol,[64] starting from 6-bromopyridin-2-amine and ethynyltrimethylsilane to give intermediate 51 (Scheme 3). Hydrolysis of the trimethyl silane and condensation with ethyl-chloro-(2-hydroxyimino)-acetate led to intermediate 52, which is reacted with 4-bromotoluene according to a Buchwald-Hartwig protocol, as reported earlier in our group,[50] to give the desired final compound 53 in low yield. Finally, for the synthesis of compounds 58–63a, 58–61b and 58c, the reaction between propiolic acid and ethyl-chloro-(2-hydroxyimino)-acetate at reflux gave the key isoxazole intermediate 54, that is reacted with the suitable benzo[d]thiazol-2-amines 55a-c to give esters 56a-c. When not commercially available, intermediates 55a-c were prepared by intramolecular cyclization of the proper phenylthiourea, synthesized in situ with ammonium thiocyanate, in presence of bromine. The ethyl esters were then hydrolyzed with lithium hydroxide, leading to the corresponding acids 57a-c that underwent coupling reaction promoted by various coupling agents to give final compounds 58–63a, 58–61b and 58c in moderate yields (Scheme 4).

Scheme 3^a.

^aReagents and conditions: a) Ethynyltrimethylsilane, TEA, Pd(PPh₃)₂Cl₂, CuI, THF, 12h, (53%) b) KOH, MeOH, 10 min, quantitative; c) triethylamine, DMF, 90°C, 17h, (82%); d) 4-bromotoluene, tBuONa, Tris(dibenzylideneacetone)dipalladium, Dave-Phos, t-BuOH, toluene, 12h, 140°C (22%).

Scheme 4^a,b.

^aReagents and conditions: a) propiolic acid, triethylamine, THF, 90°C, 17h, (28–65%); b) R-NH2, ammonium thiocyanate, bromine, glacial acetic acid, 20h, (40–70%); c) R-NH₂, HATU, TEA, DMF, 2–17h (25–48%) d) LiOH, THF/H₂O/MeOH 3:1:1, rt, 3 h, quantitative; e) R-NH₂, HATU or COMU or TBTU/EDC-HCl, TEA, DMF, 2–17h (30–50%). ^bFor complete structures, see Table 3.

Conclusion

We have herein reported the development of a series of 2-aminothiazoles highly active against susceptible and resistant Mtb, with improved metabolic stability compared to the parent compounds 3 and 4. Compounds 3 and 4 were previously reported by us as promising antitubercular chemotypes by virtue of their strong antitubercular activity coupled to irrelevant cytotoxicity toward eukaryotic cells, high selectivity toward other microorganisms, and little tendency to be substrate of Mtb efflux pumps. Therefore, in this work, we have undertaken a hit-to-lead optimization based on the convergent analysis of the Structure-Activity Relationships (SAR) and Structure-Metabolism Relationships (SMR). Since the preliminary phases of the study, metabolic stability was investigated only in a subcellular model, so as to avoid time-consuming assays involving animals, and allowed to obtain interesting information. The hydroxylation resulted to be the primary phase I transformation and the toluene ring attached to the 2-aminothiazole core was found to be the main substrate of the metabolism. The two metabolic soft-spots, both for compounds 3 and 4, are represented by the methyl group of the toluene attached to the 2-aminothiazole and the ortho position to the methyl. Interesting, while the major metabolite resulted inactive, the minor metabolite (compounds 42e and 43e) was found to be moderately active and less toxic to mammal cells compared to the parent compounds. These findings proved to be important in the design and synthesis of derivatives with ameliorated stability toward HLM degradation, that was successfully achieved, and, in the case of compounds 42g and 42l, also coupled with the maintenance of a good antitubercular activity. Despite the metabolic liability, most of the compounds prepared in this work had a 4-methyl substitution at the right wing to allow a comprehensive SAR analysis of the left wing of the molecule. SAR analysis revealed that substitutions at the aromatic ring attached to the isoxazole-3-carboxyamide are more tolerated than those at the toluene attached to the 2-aminothiazole, and this position offers a wider scope for modification in order to modulate the antitubercular activity of the compounds and, in some cases, also their cytotoxicity. Despite the molecular target of this class of compounds is still unknown, it seems that only small functional groups can be attached to the arylamide portion, regardless their position around the ring (ortho, meta, para) and their electron-withdrawing or electron-donor nature. Moreover, the presence of a polar pocket[65] within the target binding site is suggested by the remarkable activity of hydroxyl derivatives 20–22a. This can be considered an exception in the field of antitubercular compounds as hydrophilic substituents generally have a detrimental effect on antimycobacterial activity, probably because they hamper cell wall permeability. On the other hand, this modification is important in the improvement of the general drug-likeness of the molecule, especially concerning its solubility and formulability. Finally, compounds belonging to this series, despite being devoid of activity toward NTM, were found to be highly active against a large panel of resistant Mtb phenotypes.

This study was accompanied by the design of rearranged versions of the 5-(2-aminothiazol-4-yl)isoxazole moiety, characterized either by thiazole bioisosteres or by constrained structures such as benzothiazoles. Whereas benzothiazoles were devoid of any antitubercular activity, compound 53, embodying a 5-(pyridin-2-yl)isoxazole-3-carboxylate core, was found to be the prototype of a novel chemical class of antitubercular compounds, whose deeper investigation is currently ongoing in our laboratories.

Experimental section

Chemistry

General information.

All the reagents were purchased from Sigma-Aldrich and Alfa-Aesar at reagent purity and, unless otherwise noted, were used without any further purification. Dry solvents used in the reactions were obtained by distillation of technical grade materials over appropriate dehydrating agents. Thioureas a, b, c, d, g, i, l, m were purchased while thioureas e, f, h, j, k, n were prepared as previously described, and the analytical data matched with those already reported (scheme 1).[42,66,67] Reactions were monitored by thin layer chromatography on silica gel-coated aluminium foils (silica gel on Al foils, SUPELCO Analytical, Sigma-Aldrich) at both 254 and 365 nm wavelengths. Where indicated, intermediates and final products were purified through silica gel flash chromatography (silica gel, 0.040–0.063 mm), using appropriate solvent mixtures.

Scheme 1^a,b.

^aReagents and conditions: a) Br₂, chloroform, AcOH, 50°C, 1h (70–81%); b) suitable thioureas, absolute EtOH, 80 °C, yield: 17–70% c) LiOH, THF/H₂O/MeOH 3:1:1, r.t., 1h, yield: quantitative; d) for compounds 9–19a, 23–34a, 35–37b, 35–37c, 35–37d and 20e: TBTU, EDC∙HCl, N₂ atm, dry DMF, 10 min; then proper amine, TEA, r.t., 4h, yield: 17–79%; for compounds 20–22a: the appropriate aminophenol, TEA, COMU, N₂ atm, dry DMF, r.t., 4h, yield: 10–28%; e) TBAF on silica gel in dry DMF, 0 °C to r.t., yield: 32–48%. ^bFor complete structures, see Table 1.

¹H-NMR and ¹³C-NMR spectra were recorded on a BRUKER AVANCE spectrometer at 400 and 100 MHz respectively, with TMS as internal standard. ¹H-NMR spectra are reported in this order: multiplicity and number of protons. Standard abbreviation indicating the multiplicity was used as follows: s = singlet, d = doublet, dd = doublet of doublets, t = triplet, q = quadruplet, m = multiplet and br = broad signal. HPLC/MS experiments were performed with HPLC: Agilent 1100 series, equipped with a Waters Symmetry C18, 3.5 μm, 4.6 mm × 75 mm column and MS: Applied Biosystem/MDS SCIEX, with API 150EX ion source. HRMS experiments were performed with LTQ ORBITRAP XL THERMO.

Microwave reactions were conducted using a CEM Discover Synthesis Unit (CEM Corp., Matthews, NC). The machine consists of a continuous focused microwave power delivery system with an operator-selectable power output from 0 to 300 W. The temperature inside the reaction vessel was monitored using a calibrated infrared temperature control mounted under the reaction vessel. All experiments were performed using a stirring option whereby the reaction mixtures were stirred by means of a rotating magnetic plate located below the floor of the microwave cavity and a Teflon-coated magnetic stir bar in the vessel.

All compounds were tested as 95–100% purity samples (by HPLC/MS).

General procedure for α-bromination of the ketones.

The appropriate ketone (1 eq) was dissolved in dioxane (2 mL per mmol) and a catalytic amount of acetic acid (0.1 mL per mmol) was added. The mixture was heated up to 40°C and bromine (1 eq) was added dropwise. Reaction mixture was stirred at 70°C for 2 h. Then, the mixture was carefully washed with aq. sat. NaHCO₃ and extracted with EtOAc (3 × 20 mL). Organic layers were collected, dried over Na₂SO₄ and purified by Combiflash^® eluting from 10% to 30% v/v EtOAc in petroleum ether. Yields and analytical data are reported in the Supporting Information.

General procedure for Hantzsch synthesis: method A.

The suitable α-bromoketone (1 equiv) and the proper thiourea (1 equiv) were solubilized in anhydrous ethanol (20 mL/mmol) and reacted at 70° C until consumption of the starting materials as indicated by TLC. After cooling, the desired compound was collected by filtration after precipitating in the reaction mixture. When the precipitation of the desired compound was not observed, reaction mixture was evaporated under reduced pressure, partitioned between water (5 mL) and EtOAc (5 mL), extracted with EtOAc (20 mL x3) and purified by Combiflash^®. Purification conditions, yields and analytical data are reported in the Supporting Information.

General procedure for microwave-assisted synthesis: method B.

The α-bromo ketone (1 eq) and the appropriate thiourea (1 eq) were dissolved in EtOH (4 mL per mmol) under nitrogen atmosphere and the mixture was heated up to 60°C using microwave irradiation (power: 200 W) for 3 minutes. Purification conditions, yields and analytical data are reported in the Supporting Information.

Starting from compound 6, compounds 7a (Ethyl 5-(2-(p-tolylamino)thiazol-4-yl)isoxazole-3-carboxylate) and 7b (Ethyl 5-(2-(3,5-dichlorophenyl)thiazol-4-yl)isoxazole-3-carboxylate) were synthesized as previously reported.[33] Analytical data matched those reported in literature.

General procedure for the hydrolysis of the esters.

The appropriate ester (1 equiv) and LiOH·H₂O (4 equiv) were dissolved in a solution of THF/MeOH/H₂O (3:1:1, 1 mL/mmol) and stirred at room temperature until consumption of the starting material as indicated by TLC (7:3 Petroleum ether/Ethyl Acetate, then 9:1 Dichloromethane/methanol). The reaction mixture is then evaporated under reduced pressure, and the crude obtained is taken up with H₂O, acidified with 2N HCl and extracted with ethyl acetate (3×10 mL). After evaporation of the solvent, the product is used for the next reaction step without further purification. Purification conditions, yields and analytical data are reported in the Supporting Information.

Starting from compound 5, 7a,7b, 45, 56a-c the corresponding acids 37, 8a, 8b, 46, and 57a-c were obtained as previously reported.[33] Analytical data matched those reported in literature.

General procedure for the synthesis of amides: method C.

O-(Benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium tetrafluoroborate (TBTU, 1 equiv) and N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC·HCl, 1 equiv) were added to a solution of the proper carboxylic acid (1 eq) in dry DMF (4 mL/mmol). The reaction mixture was stirred at room temperature under N₂ for 15 minutes, then triethylamine (1.5 equiv) and the suitable amine (1 equiv) were added to the mixture that was stirred at the same temperature until the complete consumption of the starting material, as indicated by TLC (usually 7:3 Petroleum ether/Ethyl Acetate). Water (10 mL) was added and the mixture extracted with ethyl acetate (3×10 mL). The organic layers were treated with water, washed with brine and dried over Na₂SO₄. After filtration, the solvent was removed in vacuo and the crude material was purified by flash column chromatography to give the title compounds. Compounds and purification conditions, yields and analytical data are reported in the Supporting Information.

General procedure for the synthesis of amides: method D.

1-[(1-(Cyano-2-ethoxy-2-oxoethylideneaminooxy) dimethylaminomorpholino)] uronium hexafluorophosphate (COMU) (1 eq) was added to a solution of the carboxylic acid (1 eq), the appropriate amine (1 eq) and triethylamine (1.5 eq) in dry DMF (4 mL/mmol) under N₂ atmosphere. After stirring at room temperature for 4h, brine (5 mL) was added and the mixture was extracted with EtOAc (3 × 10 mL). The organic layers were collected, washed with brine (2 × 5 mL) and dried over Na₂SO₄. After filtration, the volatiles were removed under vacuum and the crude material was purified by flash column chromatography. Compounds and purification conditions, yields and analytical data are reported in the Supporting Information.

5-(2-((4-(hydroxymethyl)phenyl)amino)thiazol-4-yl)-N-phenylisoxazole-3-carboxamide (42n).

The crude 42k was dissolved in DMF (0,5 mL) at 0 °C under nitrogen atmosphere. Then TBAF on silica gel (2 eq) was added and the mixture was left to stir for 3h. After this time, the reaction was quenched with water and the mixture extracted with EtOAc (3 × 10 mL). Organic layers were collected and dried over Na₂SO₄ and purified by flash column chromatography eluting from 20% to 30 % EtOAc in petroleum ether. The title compound was obtained as a pale-yellow powder (yield: 9 mg, 32%). ¹H-NMR (300 MHz, DMSO-d₆) δ: 10.76 (1H, s), 10.47 (1H, s), 7.81 (2H, d, J=7.9 Hz), 7.69 (1H, s), 7.64 (2H, d, J=8.5 Hz), 7.39 (2H, t, J=7.9 Hz), 7.31 (2H, d, J=8.5 Hz), 7.21 (1H, s), 7.19–7.14 (1H, m), 5.10 (1H, t, J=5.7 Hz), 4.46 (2H, d, J=5.7 Hz); ¹³C-NMR (75 MHz, DMSO-d₆) δ: 166.9, 165.0, 160.1, 157.6, 139.9, 138.5, 138.3, 136.5, 129.3, 128.0, 125.0, 121.1, 117.6, 110.3, 100.9, 63.2. HR-MS analysis: calculated for C₂₀H₁₆N₄O₃S: 393.10; found: 393.10241. Mp = 194.4

Using a similar procedure, but starting from compound 43k, compound 43n was obtained.

5-(2-((4-(hydroxymethyl)phenyl)amino)thiazol-4-yl)-N-(pyridin-2-yl)isoxazole-3-carboxamide (43n).

The title compound was allowed to precipitate overnight at −20°C (yield: 48%, pale yellow powder). ¹H-NMR (600 MHz, DMSO-d₆) δ: 10.78 (1H, s), 10.41 (1H, s), 8.39 (1H, ddd, J= 4.8, 1.9, 0.9 Hz,), 8.10 (1H, d, J=8.3 Hz,), 7.86 (1H, ddd, J=8.3, 7.4, 1.9 Hz,), 7.64 (1H, s), 7.61 (2H, d, J=8.3 Hz,), 7.30 – 7.24 (3H, m), 7.20 (1H, ddd, J=7.4, 4.8, 0.9 Hz), 5.05 (1H, t, J=5.7 Hz), 4.41 (2H, d, J=5.7 Hz). ¹³C-NMR (150 MHz, DMSO-d₆) δ: 167.0, 164.7, 159.5, 158.2, 151.4, 148.8, 139.7, 139.0, 138.4, 136.5, 128.0, 127.9, 117.5, 115.3, 110.4, 100.9, 63.2. HR-MS analysis: calculated for: C₁₉H₁₅N₅O₃S: 393.09; found: 394.09684. Mp = 195.8

Ethyl 5-(2-aminooxazol-4-yl)isoxazole-3-carboxylate (44)

In a 10 mL test tube for microwave, urea (228 mg, 3.8 mmol) was added to a solution of α-bromoketones 6 (100 mg, 0.38 mmol) in dry DMF (3 mL/mmol), and the reaction mixture was irradiated in a microwave reactor after setting the parameters as follows: temperature = 120°C, time = 3 minutes, pressure = 250 psi, power = 300 W, power max = off. After the complete consumption of the starting material, as revealed by TLC (95:5 dichloromethane/methanol), water was added, and the mixture extracted with ethyl acetate (3 × 10 mL). The organic layers were washed with water, brine and dried over Na₂SO₄. After filtration, the solvent was removed in vacuo and the crude material was purified by flash column chromatography eluting petroleum ether/ethyl acetate 1:1. 42 mg, 49% yield. Analytical data matched those already reported.[33]

Ethyl 5-(2-(p-tolylamino)oxazol-4-yl)isoxazole-3-carboxylate (45)

In a 10 mL test tube for microwave, a solution of 2-aminooxazole 44 (50 mg, 0.22 mmol), p-bromotoluene (19 mg, 0.11 mmol), sodium t-butoxide (21 mg, 0.22 mmol) in a mixture of anhydrous toluene (2.5 mL/mmol) and t-butanol (0.5 mL/mmol) was stirred under argon flux for 15 minutes. After this time, X-phos Pd G2 (17.3 mg, 0.022 mmol) was added and the reaction mixture was irradiated in a microwave reactor after setting the parameters as follows: temperature = 130°C, time = 15 minutes, pressure = 250 psi, power = 300 W, power max = off. The reaction was monitored by TLC. After the complete consumption of the starting material, as revealed by TLC (95:5 dichloromethane/methanol), water was added, and the mixture extracted with ethyl acetate (3 × 10 mL). The organic layers were washed with water, brine and dried over Na₂SO₄. After filtration, the solvent was removed in vacuo and the crude material was purified by flash column chromatography, eluting petroleum ether/ethyl acetate 9/1. 18 mg, 16% yield. Analytical data matched those already reported.[33]

6-((trimethylsilyl)ethynyl)pyridin-2-amine (50)

To a solution of 2-amino-6-bromopyridine (2 g, 11.56 mmol) in anhydrous THF, Pd(PPh₃)₂Cl₂ (243 mg, 0.03 mmol) and CuI (110 mg, 0.05 mmol) were added. After 10 minutes, anhydrous triethylamine (40.5 mL, 80.9 mmol) was added dropwise followed by ethynyltrimethylsilane (1.703 g, 17.34 mmol). The mixture was allowed to react overnight at room temperature. After the complete consumption of the starting material, as revealed by TLC (95:5 dichloromethane/methanol), the mixture was filtered through an alumina pad, water was added, and the mixture extracted with ethyl acetate (3 × 10 mL). The organic layers were washed with water, brine and dried over Na₂SO₄. After filtration, the solvent was removed in vacuo and the crude material was purified by flash column chromatography, eluting petroleum ether/ethyl acetate 8/2. 1.460 g, 66% yield. Analytical data matched those already reported.[64]

6-Ethynylpyridin-2-amine (51)

To a solution of 20% KOH in MeOH (13 mL), 6-((trimethylsilyl)ethynyl)pyridin-2-amine (190 mg, 1 mmol) is added and the reaction is stirred at room temperature for 1h. To obtain higher yield, MeOH is not evaporated, rather, water was added, and the mixture extracted with ethyl acetate (3 × 10 mL). The organic layers were washed with water, brine and dried over Na₂SO₄. After filtration, the solvent was removed in vacuo and the crude material was used in the next reaction step without further purification. Analytical data matched those already reported.[64]

Ethyl 5-(6-aminopyridin-2-yl)isoxazole-3-carboxylate (52)

Triethylamine (255 mg, 2.53 mmol) was added dropwise to a solution of ethyl-chloro-(2-hydroxyimino)-acetate (128 mg, 0.843 mmol) and 6-ethynylpyridin-2-amine (150 mg, 1.26 mmol) in anhydrous DMF (1.8 mL/mmol) at 0 °C. The reaction mixture was stirred at 60 °C for 1 hour. After this period, 1N HCl (5 mL/mmol) was added, and the mixture extracted with ethyl acetate (3 × 10 mL). The organic layers were washed with water, brine and dried over Na₂SO₄. After filtration, the solvent was removed in vacuo and the crude material was purified by flash column chromatography eluting dichloromethane/methanol 96/4. 107 mg, 21% yield. ¹H-NMR (400 MHz, DMSO-d₆) δ: 7.55 (t, J=8.0 Hz, 1H), 7.22 (s, 1H), 7.18 (d, J=8.0 Hz, 1H), 6.58 (d, J=8.0 Hz, 1H), 6.35 (s, 1H), 4.39 (q, J=8.0 Hz, 2H) 1.34 (t, J=8.0 Hz, 3H)

With this procedure, starting from propiolic acid, the isoxazole derivative 3-(ethoxycarbonyl)isoxazole-5-carboxylic acid 54 was synthesized. 38% yield. ¹H-NMR (400 MHz, DMSO-d₆): δ 9.73 (s, 1H), 7.45 (s, 1H), 4.42 (q, J = 8 Hz, 2H), 1.33 (t, J = 8 Hz, 3H).

Ethyl 5-(6-(p-tolylamino)pyridin-2-yl)isoxazole-3-carboxylate (53)

To solution of ethyl 5-(6-aminopyridin-2-yl)isoxazole-3-carboxylate (35 mg, 0.15 mmol), 4-bromotoluene (25 mg, 0.15 mmol), [Pd₂(dba)₃] (7 mg, 5 mol %), and Dave Phos (6 mg, 10 mol%) in anhydrous toluene (2.5 mL/mmol), sodium tert-butoxide (22 ng, 0.225 mmol) was added portionwise and the reaction was stirred at 140°C overnight. The reaction was monitored by TLC. After the complete consumption of the starting material, as revealed by TLC (95:5 dichloromethane/methanol), the suspension is filtered through a thin celite pad, then water was added, and the mixture extracted with ethyl acetate (3 × 10 mL). The organic layers were washed with water, brine and dried over Na₂SO₄. After filtration, the solvent was removed in vacuo and the crude material was purified by flash column chromatography eluting petroleum ether/ethyl acetate 85/15. A yellow solid is obtained (0.012 g, 43% yield). ¹H-NMR (400 MHz, DMSO-d₆): δ 9.27 (s, 1H), 7.73 (t, J = 8 Hz, 1H), 7.64 (d, J = 8 Hz, 2H), 7.40 (d, J = 8 Hz, 1H), 7.30 (s, 1H), 7.14 (d, J = 8 Hz, 2H), 6.93 (d, J = 8 Hz, 1H), 4.42 (q, J = 8 Hz, 2H), 2.22 (s, 1H), 1.37 (t, J = 8 Hz, 3H). ¹³C-NMR (150 MHz, DMSO-d₆) δ: 168.0, 156.5, 155.1, 154.7, 150.0, 139.5, 137.9, 131.2, 129.8, 120.3, 113.6, 104.8, 100.5, 60.9, 21.3, 14.1. HR-MS analysis: calculated for: C₁₈H₁₇N₃O₃: 324.1279; found: 324.1298. Mp = 162.6

General procedure for the synthesis of 2-aminobenzothiazoles (55a-c)

To a solution of the properly substituted amine (1 equiv) in glacial acetic acid, ammonium thiocyanate (2 equiv) was added. The reaction was stirred for 10 minutes on ice bath. Bromine (1.1 equiv) was then added dropwise: the reaction was stirred on ice bath for two hours and then stirred at room temperature overnight. A solution of ammonium hydroxide was added to reach pH 9 keeping temperature below 10 °C. The product was then extracted three times with ethyl acetate. The organic layers were dried over Na₂SO₄ and the crude material was purified by silica gel flash chromatography. Purification conditions, yields and analytical data are reported in the Supporting Information.

Analytical chemistry

Chemicals and reagents.

Human liver microsomes (HLM, pooled fraction derived from 200 male and female donors) were purchased from Xenotech, LLC (Cambridge, Kansas City, USA). 85% v/v formic acid was provided by ACEF Spa (Piacenza, Italy); HPLC-grade acetonitrile (ACN) and dimethylsulfoxide (DMSO) were supplied by Sigma Aldrich (Milan, Italy) and VWR Chemicals (Radnor, Pennsylvania, USA), respectively. Ultra-pure Millipore water (Darmstadt, Germany) was employed for HPLC mobile phase and sample preparations.

Metabolic Stability in Human Liver Microsomes (HLM).

Metabolic stability assays of compounds 3, 4, 20e, 42g-m and 43g-j were conducted in the presence of pooled human liver microsomes (HLM). Briefly, a solution of enzymatic co-factors containing glucose-6-phosphate (10 mM), NADP⁺ (2 mM), MgCl₂ (5 mM) and glucose-6-phosphate-dehydrogenase (1.5 U/mL) was prepared in 100 mM phosphate buffered saline (PBS) pH 7.4. HLM suspension (i.e. 15 μL; final protein concentration = 1 mg/mL) was added to 60 μL of co-factors mix and 222 μL of PBS buffer. Samples were pre-incubated under continuous stirring at 37 °C for 5 min, then 3 μL of DMSO stock solution of each test compound (final compound concentration = 1 μM; final DMSO concentration: 1% v/v) were added to start the reaction. Sample aliquots were collected at different time points (t = 0, 15, 30, 45, 60 min), enzymatic reactions were quenched by the addition of a double volume of acetonitrile (ACN) containing a structural analogue as internal standard (IS) (100 nM), samples were centrifuged (16000 g, 10 min, 4 °C) and the supernatant was directly analysed by HPLC-Multiple Reaction Monitoring (MRM)-MS for the percentage of test compound remaining over time. In vitro half-life (t_1/2) was calculated by the following equations: Elimination rate constant (k) = -slope; Half-life (t_1/2) min = ln2/k.[68] Microsoft Excel (Microsoft Corp., 2010, USA) was employed for data analysis.

Phase I metabolite ID profiling.

For profiling phase I metabolites in HLM, initial concentration of test compounds 3 and 4 in HLM incubates was increased to 100 μM. Aliquots of the incubates were collected at the beginning of incubation time and after 180 min, quenched by ACN addition, centrifuged and analysed by HR-MS by means of a LTQ-Orbitrap mass spectrometer (Thermo, USA).

In HLM spiking experiments, in which the HPLC retention times (RT) of hydroxylated metabolites of 3 and 4 were compared to those of corresponding synthetic standards, 10 μL of HLM incubates were diluted with either 90 μL of ACN (vehicle) or with a 250 nM solution of each synthetic standard dissolved in the same vehicle. After centrifugation, samples were analyzed by HPLC-MS/MS.

HPLC-MRM-MS analysis.

For the analysis of metabolic stability of 3, 4, and metabolically protected derivatives 20e, 42g-m and 43g-j in HLM, HPLC-MRM-MS traces were acquired both in positive (ESI⁺) and in negative (ESI⁻) ion mode. An Accela UHPLC system (Thermo, USA), equipped with a Synergy Fusion C18 80 Å RP-column (2.0 × 100 mm, 4 μm; Phenomenex, USA) was employed for HPLC-MRM-MS analysis. Mobile phases were: A: ACN and B: ultra-pure water respectively, both containing 0.1% v/v formic acid. The following general gradient was applied for the elution: 5%A between 0 and 1 min; linear gradient to 100%A from 1 to 6 min; 100%A between 6 and 10 min returning to 5%A in 2 min with a 3-min reconditioning time. Total run time: 15 min. The flow rate was maintained at 0.35 mL/min and injection volume was 10 μL. A TSQ Quantum Access Max Triple Quadrupole mass spectrometer (Thermo, USA), equipped with a heated electrospray (H-ESI) ion source, was employed for compound detection in stability assays. All analyses were performed by setting ion source voltage at 3500 V, and capillary temperature at 270 °C. Nitrogen served both as sheath and auxiliary gas at 15 psi and 35 psi, respectively; argon with a pressure of 1.5 mtorr was employed as collision gas. Xcalibur software version 2.2 (Thermo, USA) was employed for both data acquisition and processing. Tube lens voltages (TL) and collision energies (CE) for each parent-product ion transition were optimized by Flow Injection Analysis (FIA) of 10 μM solutions of each compound dissolved in MeOH. 3: m/z = 377.09 [M+H]⁺→ m/z = 235.00, 258.06, 189.04 (TL = 104 V; CE = 13, 15, 32 eV, respectively); 4: m/z = 378.10 [M+H]⁺→ m/z = 257.97, 234.97, 121.03 (TL = 74 V; CE = 17, 18, 24 eV); 20e: m/z = 409.06 [M+H]+→ m/z = 273.99, 251.01, 233.00 (TL = 103 V; CE = 16, 17, 27 eV); 42g: m/z = 364.09 [M+H]⁺→ m/z = 221.02, 244.02, 121.10 (TL = 72 V; CE = 15, 18, 23 eV); 42h: m/z = 432.09 [M+H]⁺→ m/z = 412.10, 291.96, 121.09 (TL = 94 V; CE = 19, 24, 28 eV); 42i: m/z = 288.06 [M+H]⁺→ m/z = 168.04, 121.11, 78.19 (TL = 65 V; CE = 16, 20, 35 eV); 42j: m/z = 465.98 [M+H]⁺→ m/z = 322.89, 121.06, 345.87 (TL = 83 V; CE = 13, 21, 23 eV); 42l: m/z = 397.05 [M+H]⁺→ m/z = 277.97, 256.96, 236.95 (TL = 101 V; CE = 14, 16, 27 eV); 42m: m/z = 379.10 [M+H]⁺→ m/z = 260.01, 236.99, 218.99 (TL = 108 V; CE = 15, 15, 25 eV); 43g: m/z = 363.09 [M+H]⁺→ m/z = 220.99, 244.02, 175.03 (TL = 86 V; CE = 12, 14, 32 eV); 43h: m/z = 429.03 [M-H]⁻→ m/z = 309.93, 282.96, 184.98 (TL = 81 V; CE = 22, 26, 37 eV); 43i: m/z = 287.1 [M+H]⁺ → m/z = 145.05, 168.01, 127.05 (TL = 73 V; CE = 11, 11, 22 eV); 43j: m/z = 462.93 [M-H]⁻→ m/z = 343.88, 316.82, 218.84 (TL = 78 V; CE = 22, 25, 39 eV).

In HLM spiking experiments, a longer HPLC gradient was employed to discriminate between the Retention Times of the hydroxylated derivatives in ortho and meta-positions to the methyl group of the toluene ring. Chosen mobile phases were: A: ACN and B: ultra-pure water respectively, both containing 0.1% v/v formic acid. HPLC gradient conditions were: 5%A between 0 and 1 min; linear gradient to 100%A from 1 to 120 min, returning to 5%A in 2 min with a 3-min reconditioning time. Total run time: 125 min. The flow rate was kept at 0.35 mL/min and injection volume was 10 μL. TSQ Quantum Access Max mass spectrometer acquired both in MRM and in product ion mode. For MRM acquisition mode, the following parent-product ion transitions were optimized by FIA and employed for M1 and M2 metabolites derived from 3 and 4: M1_cpd3: m/z = 393.09 [M+H]⁺→ m/z = 256.01, 233.00, 116.10 (TL = 124 V; CE = 16, 20, 39 eV). M2 _cpd3: m/z = 393.06 [M+H]⁺→ m/z = 250.98, 273.96, 204.95 (TL = 106 V; CE = 11, 15, 32 eV); M1_cpd4: m/z = 394.08 [M+H]⁺→ m/z = 376.07, 255.97, 121.02 (TL = 90 V; CE = 15, 20, 23 eV); M2_cpd4: m/z = 394.07 [M+H]⁺→ m/z = 250.97, 274.00, 121.07 (TL = 90 V; CE = 18, 19, 23 eV); 43f: m/z = 394.075 [M+H]⁺→ m/z = 274.01, 121.13, 205.00 (TL = 77 V; CE = 19, 24, 28 eV).

In product ion mode, MS/MS spectra were acquired in the m/z = 100–400 amu range and CE was set at 30 V. Parent mass was set at m/z = 377.1 (3), 393.1 (M1-M2), and at m/z = 378.1 (4) and 394.1 (M1-M2).

HR-MS analysis.

For the determination of high-resolution mass values of parent compounds 3 and 4 and of corresponding synthetic metabolite standards M1 and M2, 10 μM solutions of test compounds in MeOH were directly infused into a LTQ-Orbitrap high-resolution mass spectrometer (Thermo, USA), interfaced with a heated electrospray (H-ESI) ion source. Analysis was performed in Full Scan mode (m/z = 150–1500 amu) and in positive electrospray (ESI+). Capillary temperature was set at 275 °C, sheath gas flow rate at 8.03 (arbitrary unit). Source parameters were as follows: source voltage: 3.51 kV, capillary voltage: 13 V and tube lens voltage: 100 V with an ion injection time of 250 ms. For tandem mass analysis, Collision-induced dissociation (CID) value was set at 30 or 35 V and an isolation width of 1 m/z and an ion injection time of 1000 ms were employed.

For the HR-MS analysis of HLM incubates of 3 and 4, a Dionex HPLC (Thermo, USA) was coupled with the LTQ-Orbitrap mass analyzer (Thermo, USA). An AERIS Peptide 3.6um XB-C18 150×2.10mm column equipped with a SecurityGuard ULTRA cartridge (Phenomenex, USA), thermostated at 35°C, was employed for compound separation. Mobile phases A and B were ACN and ultra-pure water both added with 0.2% v/v formic acid. The following elution gradient was applied: 10%A between 0 and 5 min; to 95%A between 5 and 36 min; to 96%A between 36 and 40 min; back to 10%A at 41 min with a 9-min reconditioning time. Total run time was 50 min. Flow rate was set at 0.2 mL/min and injection volume was 10 μL. Analysis was first performed in Full Scan mode (m/z = 150–1500 amu), both in ESI⁺ and ESI⁻. Capillary temperature was set at 275 °C, sheath, auxiliary and sweep gas pressures were 20, 5 and 5 psi, respectively. Source parameters in ESI⁺ employed a source voltage of 3000 V, a capillary voltage of 13 V and a tube lens voltage of 85 V. For ESI⁻ they were 2500 V, 35 V and −110 V, respectively. Resolution was set at 15000 (at m/z = 400). For tandem mass analysis, the first and second most intense ions of the parent list were acquired if having a minimum signal of 10000 in data dependent scan, an isolation window of 1 m/z in collision-induced dissociation (CID) and a collision energy of 30–35 V. Xcalibur software version 2.07 (Thermo, USA) was employed for both data acquisition and processing.

Biology

Determination of the MIC for the SAR and SMR studies.

The drug susceptibility of M. tuberculosis H37Rv strain was determined using the resazurin microtiter assay (REMA), as previously described.[69] Briefly, log-phase bacterial cultures were diluted to a theoretical OD₆₀₀=0.0005 and grown in a 96-well black plate (Fluoronunc, ThermoFisher) in the presence of serial compound dilution. A growth control containing no compound and a sterile control without inoculum were also included. After 7 days of incubation at 37°C, 10 μL of resazurin (0.025% w/v) were added and fluorescence was measured using a Fluoroskan^™ Microplate Fluorometer (Thermo Fisher Scientific; excitation=544 nm, emission=590 nm). Bacterial viability was calculated as a percentage of resazurin turnover in the absence of compound. The activity of 26a against Mtb mutants and MDR clinical isolates was determined using the micro-broth dilution method. The Mtb strains were grown at 37 °C in Middlebrook 7H9 broth (Becton Dickinson), supplemented with 0.05% Tween 80, or on solid Middlebrook 7H11 medium (Difco) supplemented with oleic acid-albumin-dextrose-catalase (OADC). Dilutions of M. tuberculosis cultures (about 10⁵-10⁶ cfu/mL) were streaked onto 7H11 solid medium containing a range of drug concentrations. Plates were incubated at 37 °C for about three weeks and the growth was visually evaluated. The lowest drug dilution at which visible growth failed to occur was taken as the MIC value.

Determination of the MIC for M. abscessus Bamboo and M. avium 11.

The minimum inhibitory concentrations (MIC) of the compounds were determined by broth microdilution, as previously described.[70] Briefly, 2-fold serial dilutions of each compound were inoculated with a suspension of M. abscessus Bamboo or M. avium 11 at a cell density (OD₆₀₀) of 0.05 in a flat-bottom 96-well microtiter plate. Following incubation at 37 °C with orbital shaking at 110 rpm for 3 and 4 days for M. abscessus Bamboo and M. avium 11, respectively, the MIC₅₀ was scored as the concentration at which growth was inhibited by 50% relative to the untreated control. Clarithromycin was included as a positive control in all assays, and all data are representative of two independent replicates.

Cytotoxicity assay and IC₅₀ determination. Cell culture.

Human monocytes THP-I cells were grown in suspension in RPMI 1640 (Euroclone, Italy) supplemented with 1% Sodium Pyruvate 100 mM (Life Technologies, USA), 0.5% Gentamicin 10mg/mL (Sigma Aldrich, USA), 0.1% 2-Mercaptoethanol 50mM (Life Technologies, USA), Glucose (0,25 g/mL) and 10% Fetal Bovine Serum (FBS, Euroclone, Italy).

THP-1 cells were seeded in 24-well plates (Sarstedt, Germany) in the presence of 100ng/mL of phorbol 12-myristate 13-acetate (PMA; Sigma-Aldrich, USA) for 72 hours to allow the cells to differentiate into macrophages.

Cytotoxicity assay (MTT assay).

MTT assay has been used to assess the cytotoxicity of synthesized compounds in THP-derived macrophages by evaluating the ability of mitochondrial succinate dehydrogenase to catalyze the enzymatic reduction of yellow water-soluble 3-[4,5-dimethylthiazole-2-yl]-2,5-diphenyltetrazolium bromide (MTT; Sigma Aldrich, USA) to insoluble purple formazan, index of cell viability. Differentiated macrophages were treated with synthesized compounds at increasing concentrations (1 μM, 10 μM, 50 μM and 100 μM; all compounds were tested twice) for 72 hours. Afterwards, cells were incubated with a solution of 1 mg/mL of MTT dissolved in RPMI 1640, supplemented with 5% FBS, at 37°C, 5% CO₂ for 2 hours in dark. The solution was then removed and resulting formazan crystals were dissolved in 0,2 mL DMSO under shaking for 10 minutes. Finally, absorbance was measured on an aliquot of 0.15 mL at a wavelength of 570 nm using an absorbance microplate reader (Spark^® Tecan, Switzerland).[71]

Statistical analysis.

Statistical analyses were performed using GraphPad Prism software version 7.0 (GraphPad Software Inc., La Jolla, CA). Results were expressed as the percentage between the viability of cells treated with the compounds and that of control cells. IC₅₀ values for each tested compound were calculated by interpolation through a non-linear regression analysis.

Experiments on Zebrafish larvae:

[72,73] Single-cell preparation. M. marinum M/mCherry cultures were grown in 7H9 OADC medium to mid-log phase and then repeatedly syringed through an 18-gauge needle followed by passaging through a 5-μm filter to generate single-cell mycobacteria and to get rid of bacterial clumps as previously described by Takaki et al. Caudal vein injections. Larvae were injected with around 100–150 CFU in the posterior (caudal) cardinal vein at 48–52 hours post fertilization (hpf). For each injection, embryos were first anesthetized in tricaine (Sigma) and the injection performed by positioning the larva so that the caudal vein region is directly below the tip of the needle. Bacterial cultures were injected upon piercing of the caudal vein region. The injected larvae were incubated at 28 °C and the same caudal vein injection procedure was performed 24 hours post infection (hpi) with the desired compounds.

Scheme 2^a,b.

^aReagents and conditions: a) urea, DMF, MW (120 °C, 300W), 3 min, 49%; b) 4-bromotoluene, tBuONa, X-Phos Pd G2, tBuOH, toluene, MW (130°C, 300W), 15 min, 16%; c) LiOH, THF/H₂O/MeOH 3:1:1, rt, 3 h, quantitative; d) 2-aminopyridine or anyline, TBTU, EDC, TEA, anhydrous DMF, rt, 5 h, 17–57%. ^bFor complete structures, see Table 2.

Table 4.

Activity of 26a against a panel of Mtb mutants and against MDR clinical isolates.

Strain	MIC90a (μg/mL)
	26a	Isoniazid
H37Rv	0.5	0.025
53.3 (Rv2466c, W28S)[55]	0.5	0.025
53.8 (Rv0579, L240V)[57]	0.5	0.025
NTB1 (dprE1, G387S)[54]	0.5	0.025
DR1 (mmpl3, V681I)[51]	0.5	0.025
Ty1 (Rv3405c, c190t)[56]	0.5	0.025
88.1 (coaA, Q207R)[52]	0.5	0.025
88.7 (pyrG, V186G)[53]	0.5	0.025
81.10 (ethA, Δ1109–37)[53]	0.5	0.025
IC1 (res. to STR, INH, RIF, EMB, ETH)[58]	0.5	>0.2
IC2 (res. to STR, INH, RIF, EMB, PYR, ETH, CM)[58]	0.5	>0.2

^aMinimum Inhibitory Concentration, determined by micro-broth dilution method.

Abbreviations^a

DMF

N,N-dimethyl formamide

DCM

dichloromethane

DMSO

dimethylsulfoxide

EDC

1-ethyl-3-(3-dimethylaminopropyl)carbodiimide

EDG

electron-donor group

EWG

electron-withdrawing group

EI

efflux inhibitor

HLM

Human Liver Microsomes

MIC

minimum inhibitory concentration

MDR-TB

multidrug-resistant tuberculosis

Mtb

Mycobacterium tuberculosis

MW

microwave

PAINS

pan-assay interfering compounds

RIF

rifampin

SAR

Structure–Activity Relationships

SMR

Structure–Metabolism Relationships

SM

streptomycin

TB

tuberculosis

TBTU

2-(1H-Benzotriazole-1-yl)-1,1,3,3-tetramethyluronium tetrafluoroborate

TDR

totally drug resistant

TEA

triethylamine

THF

tetrahydrofuran

TLC

thin-layer chromatography

XDR-TB

extensively drug-resistant tuberculosis