Novel Antitubercular 6-Dialkylaminopyrimidine Carboxamides from Phenotypic Whole-Cell High Throughput Screening of a SoftFocus Library: Structure–Activity Relationship and Target Identification Studies

A BioFocus DPI SoftFocus library of ∼35 000 compounds was screened against Mycobacterium tuberculosis (Mtb) in order to identify novel hits with antitubercular activity. The hits were evaluated in biology triage assays to exclude compounds suggested to function via frequently encountered promiscuous mechanisms of action including inhibition of the QcrB subunit of the cytochrome bc1 complex, disruption of cell–wall homeostasis, and DNA damage. Among the hits that passed this screening cascade, a 6-dialkylaminopyrimidine carboxamide series was prioritized for hit to lead optimization. Compounds from this series were active against clinical Mtb strains, while no cross-resistance to conventional antituberculosis drugs was observed. This suggested a novel mechanism of action, which was confirmed by chemoproteomic analysis leading to the identification of BCG_3193 and BCG_3827 as putative targets of the series with unknown function. Initial structure–activity relationship studies have resulted in compounds with moderate to potent antitubercular activity and improved physicochemical properties.


■ INTRODUCTION
Tuberculosis (TB) is one of the world's most deadly infectious diseases, leading to 1.4 million deaths in 2015. 1 The disease is caused by Mycobacterium tuberculosis (Mtb) and primarily affects lower to middle income countries such as those in sub-Saharan Africa. A standard first-line four-drug regimen of rifampin, isoniazid, pyrazinamide, and ethambutol is used to treat drugsusceptible TB. 2 Although these drugs form the core of TB treatment regimens, they have become ineffective against resistant Mtb strains. TB that does not respond to at least isoniazid and rifampin, the two frontline drugs, is defined as multidrug resistant (MDR) TB. MDR TB that develops additional resistance to any fluoroquinolone and at least one of the three injectable second-line drugs (amikacin, kanamycin, or capreomycin) is referred to as extensively drug resistant (XDR) TB and is virtually untreatable. 3 One of the main drivers behind the rise of resistance is the 6-month-long treatment required for drugsensitive disease, resulting in poor patient adherence. Thus, the main goal of many TB drug discovery programs is the identification of compounds that will potentially contribute to treatment shortening, albeit there is no clear evidence as to how this can be accomplished. 4 One possible approach is to identify compounds that inhibit novel targets and do not show cross-resistance to currently used anti-TB drugs. Various novel drug combinations can then be evaluated for their potential toward treatment shortening while establishing clinical effectiveness. Identifying compounds that exert their anti-TB effect through novel targets is not a trivial process. This is exemplified by the approval in 2012 of bedaquiline, the first new drug approved for the treatment of TB in more than 40 years with a novel mechanism of action (MOA) that involves inhibition of the mycobacterial ATP synthase. 5−7 Two hit-generation approaches that have been traditionally used are target-based and whole-cell screening. 8,9 Historically, there has been extremely limited success with the target-based approach owing to the lack of translation from target activity to whole-cell activity. Nevertheless, efforts to discover and validate novel drug targets using various approaches must continue. This is exemplified by the recently reported validation of CoaBC as a bactericidal target using a chemical biology approach. 10 On the other hand, a more successful approach has been phenotypic whole-cell high-throughput screening (HTS) of chemical libraries to identify novel chemotypes whose target is then identified retrospectively. There are several compounds in the TB drug pipeline that have been identified through whole-cell HTS. Examples include Q203 (Figure 1), an inhibitor of the QcrB subunit of the cytochrome bc 1 complex which is involved in electron transport. 11 Another example is SQ109 (Figure 1) whose precise MOA remains to be elucidated but is believed to involve the mycobacterial membrane protein large 3, MmpL3, a mycolic acid transporter involved in cell wall synthesis,. 12,13 Similarly, some inhibitors of decaprenylphosphoryl-β-D-ribose 2′-epimerase (DprE1), e.g., pBTZ, 14 and azaindoles TBA7371 15,16 are poised to enter clinical trials (Figure 1). While these compounds show promise as new anti-TB agents, their targets are yet to be clinically validated. This necessitates the need to continuously identify new chemical matter with alternative novel MOAs toward populating the TB drug pipeline.
In line with this, we initiated a HTS campaign using a diverse SoftFocus compound library 17 acquired from BioFocus and complemented this with a biology triage process that aimed specifically to exclude compounds acting on targets such as QcrB and the cell wall (e.g., MmpL3 and DprE1), as well as DNA damaging agents. The rationale was to exclude targets known for their promiscuity, with a large proportion of HTS hits having these as part of their MOA. 18,19 There is also the question around whether or not compounds targeting cell wall metabolism can contribute to the much sought after treatment shortening. 19 Compounds that induce DNA damage were also excluded in order to avoid general cytotoxicity, as well as targets such as DNA gyrase whose clinically used inhibitors are exemplified by the fluoroquinolones. Along with an aminopyrazolo [1,5-a]pyrimidine chemical series, 20 a 6-dialkylaminopyrimidine carboxamide scaffold was identified as a hit series with a potentially novel MOA, based on the initial biology triage process ( Figure 2). Herein, we describe the synthesis, structure−activity relationship (SAR), in vitro absorption, distribution, metabolism, and excretion (ADME), and in vivo pharmacokinetic (PK) and biological profiles of this chemical series.  serum albumin (BSA). The compound was slightly more potent in GAST-Fe (glycerol-alanine-salts) 20 minimal medium, likely driven by the absence of BSA which allowed for higher free concentration of the compound under the assay conditions. Hence, GAST-Fe was used alongside the 7H9/ADC medium during further SAR exploration of the series. Compound 1 was tested in the three biology triage assays in order to assess activity against promiscuous targets. The compound did not show hypersensitivity against a cytochrome bd oxidase knockout mutant strain (cydKO) 21,22 and also retained activity against a QcrB mutant (qcrB A317T ), thereby eliminating this as potential target. Compound 1 also did not elicit a positive response in two standard bioluminescence reporter assays: PiniB-LUX, 23 which is designed to detect compounds targeting Mtb cell wall biosynthesis, and PrecA-LUX, 23 which allows detection of genotoxic compounds (data in Supporting Information). Similar results were obtained with other analogues in the series obtained during SAR studies (data in Supporting Information). These results, along with chemoproteomics and cross-resistance studies (see biology results below), indicate that the 6-dialkylaminopyrimidine carboxamide series of compounds exemplified by compound 1 possess a novel MOA that is not shared with known TB drugs.
SAR Exploration Strategy. Compound 1 was identified as a suitable hit with a moderate minimum inhibitory concentration (MIC 99 ) of 20 μM in 7H9/ADC medium, low toxicity against the mammalian Vero cell line (IC 50 of 287 μM), yet poor kinetic solubility (KS, <5 μM). The poor solubility was assumed to be related to the high lipophilicity and flat aromatic character of the molecule. Since there were very few active near neighbors identified from the screen, very limited information was available on further optimization scope with the scaffold to improve anti-TB activity while addressing the poor physicochemical properties and potential safety issues that may be associated with the scaffold. A detailed SAR plan as summarized in Figure 3 was embarked upon. The molecule was divided into three parts for convenience: the central pyrimidine core, right-hand side (RHS) amide, and left-hand side (LHS) hydrophobic N-benzyl moiety ( Figure 3). The major SAR exploration focus was on understanding key hydrogen bond (HB) donor−acceptor interactions critical for activity, shape, and size of the central core, and scope for substitution/modification on either the RHS or LHS parts of the molecule to improve potency. The scope for addition of heteroatoms or polar groups into the molecule was also explored with the aim of reducing lipophilicity along with addressing related issues such as poor physicochemical properties and structure alerts such as the presence of potentially AMES positive anilines upon cleavage of the amide group.
Synthesis. Compounds 1−20 were synthesized to explore the SAR around the N-methyl-1-phenylmethanamine motif (SAR on LHS, Table 1). Compounds 21−30 were synthesized to explore the central pyrimidine core of the scaffold (SAR on central core, Table 2), and compounds 31−54 were synthesized to explore the SAR around the amide portion of the scaffold (SAR on RHS, Table 3). In general, target compounds were synthesized using either one of two routes. Scheme 1 was used to synthesize an advanced intermediate with a specific amide in place allowing for the exploration of the LHS of the molecule. The same route was also used to synthesize several of the core modifications. The route involved a straightforward two-step procedure starting from a chloro-heterocyclic-carboxylic acid, where coupling with corresponding amine using 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophosphate (HATU) was first performed, followed by nucleophilic displacement of the relevant chloro group by an appropriate amine. Thus, synthesis of compounds 1−20 and 39 (Table 1) was achieved by coupling of 6-chloropyrimidine-4-carboxylic acid with aniline, 4-methoxyaniline, and 4-aminopyridine to give intermediates 55, 56, and 57, respectively, followed by displacement of 6-chloro group with various amines. Compound 21 was synthesized by coupling of 4-chloropicolinic acid with aniline to form intermediate 58 followed by reaction with 1-(4-fluorophenyl)-Nmethylmethanamine.
Synthesis of compounds 22−26 (Table 2) with changes in the central core ring are summarized in Scheme 2. Compound 22 was synthesized by nucleophilic displacement of 59 with 1-(4fluorophenyl)-N-methylmethanamine followed by carbamoylation according to the procedure described by Ren et al. 24 Compound 23 was synthesized in a similar manner except the initial nucleophilic displacement step of 60 resulted in a mixture of the intended 4-chloro-N-(4-fluorobenzyl)-N-methylpyrimidin-2-amine

Journal of Medicinal Chemistry
Article (61) as well as the 2-chloro regioisomer in a 1:2 ratio. The pyridazine core replacement analogue was synthesized by heating 6-oxo-1,6-dihydropyridazine-4-carboxylic acid with POCl 3 followed by quenching of the 6-chloropyridazine-4carbonyl chloride intermediate with aniline to afford 63. Intermediate 63 was subsequently subjected to nucleophilic displacement with 1-(4-fluorophenyl)-N-methylmethanamine under microwave heating to deliver the final compound 24. The pyridone analogue (25) was synthesized from the commercially available precursor, methyl coumalate (64), which was reacted with 4-fluorobenzylamine at room temperature (25°C) to give the N-substituted pyridinone ester (65). The ester was then hydrolyzed using LiOH followed by an amide coupling to afford 25. The thiazole analogue 26 was synthesized by reductive amination of methyl 2-aminothiazole-4-carboxylate 66 with 4-fluorobenzaldehyde and NaBH 4 , followed by methylation under basic conditions to afford 67. Intermediate 67 was then subjected to the same hydrolysis and amide coupling conditions as 65 to afford 26.

Article
The 2-and 5-methylpyrimidine analogues 27 and 28 (Table 2) were synthesized in a similar fashion by condensation of the appropriate 1,4-diethyl ester and imidamide 68 and 69 to form the corresponding dihydropyrimidine carboxylic ester and acid intermediates 70 and 71, respectively (Scheme 3). Intermediate 70 was chlorinated using SOCl 2 followed by nucleophilic displacement of the 6-chloro with 1-(4-fluorophenyl)-N-methylmethanamine delivered 72, which was then hydrolyzed under standard basic conditions followed by an amide coupling with HATU to produce 27. Intermediate 71 was subjected to chlorination of both the hydroxyl and carboxylic acid groups using oxalyl chloride. The reaction was then quenched with aniline to afford 73 which was subsequently aminated with 1-(4-chlorophenyl)-N-methylmethanamine to afford 28.
In order to access an advanced intermediate that would allow for exploration of the RHS of the molecule (31, 33−38, 40−52), another relatively straightforward approach (Scheme 4), involving four steps from the commercially available methyl 2,4-dichloropyrimidine-6-carboxylate, was used. Briefly, quantitative amination by selective displacement of the 6-chloro group using an appropriate N-methyl-1-phenylmethanamine at 0°C in THF followed by hydrogenation with 10% Pd/C and subsequent hydrolysis using LiOH in methanol afforded the carboxylic acid (78−80). Lastly, amide coupling with appropriate amines using  The reverse amide 32 was synthesized by reacting 6-chloropyrimidin-4-amine 82 with benzoyl chloride in the presence of a base followed by nucleophilic substitution of the 6-chloro with 1-(4-fluorophenyl)-N-methylmethanamine (Scheme 5). The trifluoromethyl ketone, 83, was synthesized from the free acid, 78, via formation of a Weinreb amide followed by trifluoromethylation using TMS-CF 3 . 25 Reductive amination using NaCNBH 3 in the presence of TiCl 4 then afforded compound 53. Finally, triazolyl compound 54 was synthesized by heating the hydrazide 84 with benzonitrile in refluxing butanol in the presence of potassium carbonate.
SAR on the LHS. After ascertaining that the 4-methoxy substituent of the phenyl carboxamide was not crucial for activity (Table 1; compound 2, R 1 = H) the LHS (Table 1) of the scaffold was varied to identify the key pharmacophore needed for activity while keeping the central pyrimidine core unchanged and the RHS constant as 4-methoxyphenyl or phenyl.
Replacement of the aryl ring with a saturated cyclohexyl ring (compound 3) led to considerable loss of activity, suggesting essentiality of an aromatic ring on the LHS. Similarly, replacement of the N-CH 3 linker with "O" or "NH" led to inactive compounds 4 and 5, respectively, suggesting the critical importance of a tertiary nitrogen at the linker position or intolerance toward more polar "O" or "NH" groups. Furthermore, linker modifications such as addition of a methylene group (N-methylphenethyl, 6) or removal of CH 2 (N-methylaniline, 7), cyclization of the N onto the phenyl ring as tetrahydroisoquinoline moiety (8) or extension of N-methyl to N-ethyl (9) all led to loss or deterioration of MICs. This all indicated stringent shape requirements at this presumably hydrophobic part of the molecule.

Article
Realizing the essential nature of the aryl ring and the linker, we focused efforts on exploring substitution on the phenyl ring to improve potency and introduce polarity. Para-halogenated compounds 10 (4-F), 11 (4-Cl), and 12 (4-CF 3 ) were at least ≥4-fold and ≥2-fold more potent in 7H9/ADC and GAST/Fe, respectively, compared to 1. This was encouraging and the first sign of a SAR trend emerging and leading to the conclusion that a parahalogenated phenyl ring gave the best activity with 4-F phenyl compound 10 showing the most promise. Compounds with 3-F (13) and 3-Cl (14) were less potent compared to corresponding 4-halogenated analogues, whereas the 2-Cl analogue (15) was inactive up to the highest concentration tested, once again suggesting stringent shape requirements on the LHS. Addition of more polar groups such as 4-CN (16) and 4-SO 2 Me (17) led to deterioration or loss of MICs.
Up to this point, all compounds tested had poor aqueous solubility of ≤5 μM. In an attempt to improve solubility by introducing polarity, the phenyl ring was replaced with a pyridyl ring. Compound 18 with a 2-pyridyl ring was found to be inactive, whereas compound 19 with a 3-pyridyl ring was equipotent to compound 1 albeit with a much improved aqueous solubility of 150 μM at pH 7.4. Further substitution of the 3-pyridyl ring as in compound 20 with a 3-pyridyl-4-trifluoromethyl group retained similar activity but now with loss of solubility, albeit this substitution improved the in vivo pharmacokinetic properties as will be discussed in the DMPK section.
SAR at the Central Core. Exploration of the SAR around the central core of the scaffold was aimed at understanding whether or not the 4,6-disubstituted pyrimidine was essential for activity while retaining a 4-halo group on the LHS phenyl ring and a 4-methoxy, halogen, or hydrogen on the RHS phenyl ring. Initial efforts looked at shuffling the nitrogen atoms around the core ring, including the pyridine (21), pyrazine (22), and pyrimidine (23), with one of the nitrogen atoms shifted between the aminobenzyl and carboxamide groups, and the pyridazine (24) analogues. Only the pyridine retained partial activity at 10 μM in the GAST/Fe medium, which was not sufficient to warrant a scaffold hop. Compounds based on the pyrimidinone isostere (25) and

Journal of Medicinal Chemistry
Article thiazole (26) all lost activity. Lastly 5-and 2-substitutions (27,28,29, and 30) on the pyrimidine core were evaluated and found to be similarly inactive. It was thus concluded that the original pyrimidine core was optimal for activity and could not be replaced with other diazine and heteroaromatic rings.
SAR on the RHS. An extensive SAR exploration on the RHS was undertaken in order to evaluate the scope for optimization of potency and physicochemical properties ( Table 3). The free amide NH was deemed critical for potency as N-methylation (31) led to a complete loss in activity. Similarly, reversal of the amide functionality led to inactive compound 32. For further exploration of substitution on the aryl ring of the amide, a 4-halophenyl or 4-CF 3 -3-pyridyl was fixed on the LHS of the scaffold. Replacement of the phenyl ring with an aliphatic chain such as isopropyl (33) was detrimental to activity, while replacement with a saturated cyclohexyl ring (34) resulted in retention of activity in the GAST/Fe medium albeit it was completely inactive in 7H9/ADC. The overall trend suggested a preference for an aryl ring on the RHS. Since the original hit contained a methoxy group, additional ethers were explored in the para position, including isopropoxy (36), 2-methoxyethoxy (37), and a N,N-dimethyl-2-ethoxy ether (38). Encouragingly, compounds 36 and 37 were some of the most active compounds synthesized up to this point but were poorly soluble (solubility pH 7.4 ≤ 5 μM). Compound 38 with the basic dimethylaminoethoxy side chain (predicted pK a of 8.77) was less active (7H9/ADC MIC of 25 μM) even though it showed considerably improved aqueous solubility of 155 μM. Nonetheless, these results suggested scope for substitutions on the aryl ring to modulate physicochemical properties while retaining potency. Replacement of the phenyl ring with a 4-pyridyl ring, as in compound 39, was tolerated indicating a wider scope for SAR exploration at this position with heterocycles and weakly basic substituents. Addition of a less basic N-methylpiperazine ring at the 4-position of the phenyl ring (predicted pK a of 7.85) led to potent compound 40 which was poorly soluble at pH 7.4 but sufficiently soluble at a slightly acidic pH of 6.5 (refer to DMPK section). Solubility could be improved further by combining N-methylpiperazinylphenylamide on the RHS with a 4-CF 3 -3-pyridyl group on the LHS as in compound 41 with some loss in potency (MIC = 12.5 μM, solubility at pH 7.4 of 165 μM). Moving the N-methylpiperazine ring to the 3-position of the phenyl ring (42) was detrimental to activity with MIC values dropping by 10-to 30-fold. Similarly, 3-pyridyl-4-(N-methylpiperazinyl) analogue 43 was found to be much less potent (7H9/ADC MIC > 50 μM; GAST/Fe MIC = 19 μM). Compounds with neutral substituents at the 4-position of the phenyl ring like N-acetylpiperazinyl (44) or morpholino (compound 45) retained good MICs (<5 μM) but were much less soluble. Interestingly, the 4-morpholino-2-pyridyl analogue (46) retained good MIC activity (2−3 μM) indicating scope for the introduction of a nitrogen into the phenyl ring to address potential reactivity issues of electron-rich anilines during further optimization.
A small set of amides with five-membered heterocycles was synthesized to evaluate the scope for further SAR exploration at this position. Compounds with N-methylpyrazole (47), N-methyltriazolyl (48), and isoxazolyl (49) substituents were either inactive or weakly active. Compound 50 containing a 4-methylthiazolyl group showed a GAST MIC equivalent to its phenyl counterpart (5 μM). However, its MIC in 7H9/ADC medium (37 μM) was weaker and had a poor solubility (<5 μM). Similarly, compounds containing 1,3,4-thiadiazole moieties (51 and 52) showed reasonable MIC in GAST/Fe medium (∼5 μM) but had weaker potency in 7H9/ADC medium (12.5 and 50 μM, respectively). Unfortunately compounds with these moieties were cytotoxic in vitro and were also found to be toxic at higher doses in animal studies. Hence, these modifications were abandoned even though more lead-like DMPK properties were achieved with these compounds.
Lastly, two amide isosteres CF 3 -amine (53) 26 and triazole (54) 27 were synthesized in an attempt to explore the scope for variation and in line with metabolite identification studies (Supporting Information Table S1). Metabolite identification, performed on 12 and 51, indicated cleavage of the amide bound on the RHS as a primary route of metabolism compared to N-demethylation on the LHS. Even though both these modifications resulted in compounds showing some activity, there was no improvement in other properties like solubility and hence were not explored further. Overall, as summarized in Figure 4, the SAR around the LHS and the central core was narrow with

Journal of Medicinal Chemistry
Article fewer modifications tolerated; SAR around the RHS was more dynamic, allowing scope for further optimization. From this, several compounds were identified for further profiling both in vitro and in vivo. Biology Profiling. A spontaneous resistant mutant of Mtb H37Rv was successfully raised against compound 13. Most of the analogues from the series were cross-resistant against this mutant except the thiadiazole analogues 50 and 51, indicating additional targets and/or a different MOA (Table 4). Additionally, these mutants were not cross-resistant to any known TB drugs, suggesting a potential novel MOA for the series. Unfortunately, whole genome sequencing of the mutant did not reveal any genetic polymorphism(s) suggestive of the target or MOA. Hence, further studies using different target identification approaches were explored in order to identify the molecular target of these compounds. Representative compounds were screened against drugsensitive clinical isolates and were found to retain MICs similar to those in H37Rv (Table 5).
Chemoproteomics. In order to supplement our MOA identification efforts, we used a chemoproteomics approach to identify the potential target proteins of the 6-dialkylaminopyrimidine carboxamides. This affinity-based approach was recently successfully employed for the direct target identification of antitubercular compounds. 28−30 This approach is based on the modification of the compound 40 with a linker for covalent immobilization on sepharose-beads by N-hydroxysuccinamide (NHS) coupling. The modified compound 85 ( Figure 5, synthesis Scheme S1) retained antimycobacterial activity (MIC 90 = 18.6 μM as determined against M. bovis BCG) suggesting that the derivatization with the linker moiety did not interfere with target binding. The 85 beads were incubated with Mycobacterium bovis BCG extract, and proteins captured by the beads were digested with trypsin, labeled with isobaric mass tags (TMT 10plex), and quantitatively identified by liquid chromatography−tandem mass spectrometry (LC−MS/MS). In order to distinguish true targets from nonspecific background binding, the bacterial extracts were incubated prior to the bead-incubation step with the active test compound 40, the inactive test compound 25, or a vehicle (DMSO only) control. Only the active compound should bind to its target proteins in the lysate and thus reduce the binding of these proteins to the beads, but not the inactive compound. Indeed, two proteins, BCG_3193 (Rv3169) and BCG_3827 (Rv3768), showed reduced binding to the compound beads in the presence of competing "free" 40 ( Figure 5A) but not with inactive 25, which suggests that these proteins are the potential targets of 40. In order to estimate inhibitor

Journal of Medicinal Chemistry
Article potency, we performed the affinity step in the presence of different concentrations of active compound 40 in the protein extract, which allowed the determination of an apparent dissociation constant K d app ( Figure 5C). The K d app takes into consideration the affinity of the target for the bead-immobilized ligand. The latter effect can be deduced by measuring the depletion of the target protein by the beads. 31 The apparent dissociation constant of compound 40 for BCG_3193 was determined in two separate experiments as 0.61 μM and 0.69 μM and for BCG_3827 as 2.1 μM and 3.8 μM ( Figure 5C). Both potential novel target proteins, BCG_3193/Rv3169 and BCG_3827/Rv3768 (contains a polyketide cyclase SnoaL-like domain, IPR009959), are conserved hypothetical proteins of unknown function. Both proteins are nonessential but notably found using Triton X-114 detergent phase separation for extraction of lipophilic proteins, suggesting they are associated with the cell envelope despite lacking detectable transmembrane helices. 32 Physicochemical, DMPK, and hERG Profiles. Representative compounds were profiled for physicochemical properties and in vitro DMPK properties which included kinetic solubility, human plasma protein binding (PPB), microsomal metabolic stability (MS), and permeability, as shown in Table 6.
As discussed previously, most of the compounds based on phenyl amides had poor to moderate solubility at pH 7.4 except those with basic solubilizing side chains (e.g., compounds 38 and 41). These compounds showed considerably improved solubility in slightly acidic Fassif medium (fasted-state simulated intestinal fluid, pH 6.5), 33 e.g., compound 40 (kinetic solubility at pH 7.4 of <5 μM; kinetic solubility in Fassif of 150 μM). This may be relevant for oral absorption of these compounds. Compounds containing heterocyclic amides like thiadiazole compounds 51 and 52 had good solubility of >100 μM in both media. In general, compounds in the series had 5−10% free fraction in human plasma as measured in a PPB assay while some more polar compounds like 52 had up to 30% free fraction in human plasma. The ability of compounds to cross biological membranes by passive diffusion was assessed using the PAMPA assay. 34 All compounds tested showed good to moderate permeability with values ranging between −3.8 and −4.9 log P app presumably due to their lipophilic character . Caseum binding was also assessed for certain compounds in order to give some insight into the ability of the compounds to permeate the caseous necrotic regions of granulomas to access the bacteria. 35 Compound 52 had the highest free fraction (3.6% unbound) from the compounds assessed with phenyl compounds generally being highly bound (0.06−0.1% unbound). Further safety profiling was conducted against the human ether-a-go-go-related gene (hERG) potassium channel, and compounds were found to be low micromolar inhibitors. While hERG activity is a well-established safety concern, 36 compounds were not optimized to remove hERG activity due to the initial focus being on increasing potency and solubility.
In general, compounds showed moderate metabolism in human, mouse, and rat liver microsomes except compound 40, which showed rapid in vitro clearance. The compounds were also tested in male C57/BL6 mice for pharmacokinetic parameters at an intravenous dose of 2 mg/kg and an oral dose of 20 mg/kg ( Table 7). Most of the compounds showed low in vivo clearance in mice in accordance with their in vitro microsomal metabolic stability. Compound 40 showed low clearance iv clearance despite its poor microsomal stability, probably due to extensive tissue distribution as indicated by high volume of distribution (Vd = 21.8 L/kg). Most of the compounds had moderate to good oral bioavailability. Moderate permeability was sufficient to give good oral absorption as well as moderate bioavailability for compounds with low solubility (12 and 20). Compound, 52 with

Journal of Medicinal Chemistry
Article a thiadiazole group at LHS showed improved physicochemical and ADME parameters such as good solubility, improved microsomal stability leading to higher oral absorption, higher AUC and better bioavailability of 67%. Unfortunately, this compound was found to be toxic to mice when dosed orally above 50 mg/kg and hence was not progressed any further.

■ CONCLUSION
Screening of a Biofocus library led to the identification of 6-dialkylaminopyrimidine carboxamides with potent antitubercular properties and a novel mode of action. A biology triage cascade allowed deprioritization of compounds potentially acting through inhibition of QcrB, cell-wall targets and DNA damage. Cross-resistance studies of TB drugs against the resistant mutant raised against this series indicate a potentially novel MOA for the series, and this was reinforced by chemoproteomic studies albeit the exact target (and MOA) remains unknown. Detailed SAR studies have demonstrated limitations as well as scope for further optimization of these series to deliver lead-like compounds with potent anti-TB properties and oral bioavailability. Chemoproteomics. Chemoproteomics experiments were performed as previously described. 31 Briefly, sepharose beads were derivatized with compound 85 at a concentration of 0.04 mM and subsequently washed and equilibrated in lysis buffer (50 mM Tris-HCl, pH 7.4, 0.4% Igepal-CA630, 1.5 mM MgCl 2 , 5% glycerol, 150 mM NaCl, 25 mM NaF, 1 mM Na 3 VO 4 , 1 mM DTT, and one Complete EDTAfree protease inhibitor tablet (Roche) per 25 mL). The 85 beads were incubated at 4°C for 1 h with 0.1 mL (0.25 mg) M. bovis BCG extract, which was preincubated with compound or DMSO (vehicle control). Our experimental design employed isobaric tandem mass tags (TMTs) 37 which allowed us to analyze 10 samples together by LC− MS/MS enabling the relative quantification of proteins across these samples. The experiments were configured to generate values for the affinity of the beads to the bound proteins ("depletion" values, four samples) and to allow determination of IC 50 values (six samples) in a single experiment. Samples 1 and 2 were vehicle control duplicates, samples 3 and 4 were duplicates of the "rebinding" experiment, and samples 5−10 served to generate IC 50 values by adding compound over a range of 5 concentrations (highest concentration of 30 μM, then four dilution steps of 1:4 each). In the "rebinding" experiment, the nonbound fraction from the first bead incubation step was incubated again with "fresh" beads, allowing the determination of target depletion by the beads. Apparent dissociation constants were determined by taking into account the protein depletion by the beads. 31 A repeat experiment was performed with vehicle control versus 20 μM and 5 μM compound. The beads were transferred to filter plates (Durapore PVDF membrane, Merck Millipore), washed extensively with lysis buffer, and eluted with SDS sample buffer. Proteins were alkylated, separated on 4−12% Bis-Tris NuPAGE (Life Technologies), and stained with colloidal Coomassie. Gel lanes were cut into three slices and subjected to in-gel digest using LysC for 2 h and trypsin overnight. Digestion, labeling with TMT isobaric mass tags, peptide fractionation, and mass spectrometric analyses were performed. 31 Proteins were quantified by isobaric mass tagging and LC−MS/MS. The proteins.fasta file for M. bovis BCG was downloaded on May 11, 2011, and supplemented with the sequences of bovine serum albumin, porcine trypsin, and mouse, rat, sheep, and dog keratins. Decoy versions of all proteins were created and added. The search database contained a total of 11 492 protein sequences, 50% forward and 50% reverse. Protein identification and quantification were performed. 37 Proteins identified with >1 unique peptide matches were considered for further data analysis.
All commercial reagents were purchased from Sigma-Aldrich, Combiblocks, Waterstone, or Fluorochem and were used without further purification. Solvents were used as received unless otherwise stated. Analytical thin-layer chromatography (TLC) was performed on SiO 2 plates on aluminum backing. Visualization was accomplished by UV irradiation at 254 and 220 nm. Flash column chromatography was performed using a Teledyne ISCO flash purification system with SiO 2 60 (particle size 0.040−0.055 mm, 230−400 mesh). Purity of all final derivatives for biological testing was confirmed to be >95% as determined using an Agilent 1260 Infinity binary pump, Agilent 1260 Infinity diode array detector (DAD), Agilent 1290 Infinity column compartment, Agilent 1260 Infinity standard autosampler, and a Agilent 6120 quadrupole (single) mass spectrometer, equipped with APCI and ESI multimode ionization source. Using a Kinetex Core C18 2.6 μm column (50 mm × 3 mm); mobile phase B of 0.4% acetic acid, 10 mM ammonium acetate in a 9:1 ratio of HPLC grade methanol and type 1 water, mobile phase A of 0.4% acetic acid in 10 mM ammonium acetate in HPLC grade (type 1) water, with flow rate of 0.9 mL/min, detector diode array (DAD). Or an Agilent UPLC−MS was used: Agilent Technologies 6150 quadrupole, ES ionization, coupled with an Agilent Technologies 1290 Infinity II series UPLC system Agilent 1290 series HPLC at two wavelengths 254 and 290 nm using the following conditions: Kinetex 1.7 μm Evo C18 100A, LC column 50 mm × 2.1 mm, solvent A of 0.1% (formic acid) water, solvent B of 0.1% (formic acid) acetonitrile. The structures of the intermediates and end products were confirmed by 1 H NMR and mass spectrometry. Proton magnetic resonance spectra were determined in an appropriate deuterated solvent on a Varian Mercury spectrometer at 300 MHz or a Varian Unity spectrometer at 400 MHz.
General Protocol for Synthesis of Compounds 1−21 and 28. The appropriate 6-chloro-N-phenylpyrimidine-4-carboxamide (1 equiv), N-methylbenzylamine (1.2 equiv), and DIPEA (2 equiv) were added to a 50 mL round-bottomed flask containing i-PrOH (10 mL) and heated to reflux overnight. Compounds that precipitated upon cooling were filtered and washed with cold EtOH. For those that did not precipitate, the i-PrOH was removed under vacuum and the residue redissolved in EtOAc (20 mL) and washed with a saturated aqueous solution of NaHCO 3 , the organic layer dried and removed under reduced pressure. The resulting solid or residue was either recrystallized from EtOH or purified by flash column chromatography to afford the appropriate 6-(benzyl(methyl)amino)-N-phenylpyrimidine-4-carboxamide.