1-(1-Arylethylpiperidin-4-yl)thymine Analogs as Antimycobacterial TMPK Inhibitors

A series of Mycobacterium tuberculosis TMPK (MtbTMPK) inhibitors based on a reported compound 3 were synthesized and evaluated for their capacity to inhibit MtbTMPK catalytic activity and the growth of a virulent M. tuberculosis strain (H37Rv). Modifications of the scaffold of 3 failed to afford substantial improvements in MtbTMPK inhibitory activity and antimycobacterial activity. Optimization of the substitution pattern of the D ring of 3 resulted in compound 21j with improved MtbTMPK inhibitory potency (three-fold) and H37Rv growth inhibitory activity (two-fold). Moving the 3-chloro substituent of 21j to the para-position afforded isomer 21h, which, despite a 10-fold increase in IC50-value, displayed promising whole cell activity (minimum inhibitory concentration (MIC) = 12.5 μM).


Introduction
Tuberculosis (TB) is an airborne infectious disease caused by Mycobacterium tuberculosis (M. tuberculosis). Still belonging to the top ten causes of death worldwide, TB was responsible for claiming 1.5 million lives in 2018, thereby preceding AIDS [1]. The World Health Organization (WHO) launched the "END TB" strategy in 2014, aiming at reducing the incidence of TB by 90% and the number of deaths from TB by 95% by 2035 compared with 2015 levels [2]. Nevertheless, the progress towards this sustainable development goal is disappointing, and the continuing increase in drug-resistant TB cases makes the situation more challenging [1,3].
Patients with drug-sensitive TB are currently treated with a combination regimen, consisting of a two-month treatment with first-line agents rifampicin, isoniazid, pyrazinamide and ethambutol, followed by a four-month treatment with rifampicin and isoniazid. Although this schedule has decreased TB mortality, these gains are being threatened by the advent of coinfection with HIV/AIDS, poor patient adherence and a deficient health care system. Additionally, the emergence of multidrug-resistant TB (MDR-TB) and extensively drug-resistant TB (XDR-TB) further erodes the ambitions of the WHO program [1]. In the case of MDR/XDR TB, a wide palette of second-and third-line anti-TB drugs are used, e.g., fluoroquinolones, ethionamide, thioacetazone, clarithromycin and clofazimine [4,5]. However, this the WHO program [1]. In the case of MDR/XDR TB, a wide palette of second-and third-line anti-TB drugs are used, e.g., fluoroquinolones, ethionamide, thioacetazone, clarithromycin and clofazimine [4,5]. However, this regimen not only requires more toxic and costly medications; it also requires a longer treatment duration (up to 24 months), resulting in a poor outcome. Moreover, soon after their introduction, resistance has already developed to the newly approved agents bedaquiline [6] and delamanid [7], with the use of pretomanid restricted to a limited and specific population of patients [8]. Thus, novel anti-TB agents are needed to effectively shorten the treatment regime and cure MDR-/XDR-TB.
A previous work from our laboratory started from compound 1 (Figure 1), originally reported by AstraZeneca as an inhibitor of TMPK's of Gram-positive bacteria [22]. After we found that it also potently inhibited MtbTMPK, a SAR investigation demonstrated that it could be converted to the achiral inhibitor 2 [14]. Further modifications led to the identification of 1-(1-arylethylpiperidin-4yl)thymine analog 3, which, compared to 1, displayed a nine-fold lower minimum inhibitory concentration (MIC) value for H37Rv [14].  [14,20]. Compound 3 was resynthesized and re-evaluated herein as a reference. MIC: minimum inhibitory concentration.
In this manuscript, we describe our optimization efforts towards finding potent antimycobacterial agents based on 3 by introducing modifications on the A/B/C/D ring ( Figure 2) and by altering the linker part. An overview of the synthesized analogs is presented in Figure 2.  [14,20]. Compound 3 was resynthesized and re-evaluated herein as a reference. MIC: minimum inhibitory concentration.
In this manuscript, we describe our optimization efforts towards finding potent antimycobacterial agents based on 3 by introducing modifications on the A/B/C/D ring ( Figure 2) and by altering the linker part. An overview of the synthesized analogs is presented in Figure 2. (racemic compound 23) resulted in a substantial (> 10-fold) loss in the inhibitory potency. In-line with earlier observations, with a one-carbon linker [14,21], amide analog 26 displayed a weak enzyme inhibition. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity. (racemic compound 23) resulted in a substantial (> 10-fold) loss in the inhibitory potency. In-line with earlier observations, with a one-carbon linker [14,21], amide analog 26 displayed a weak enzyme inhibition. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity.  (racemic compound 23) resulted in a substantial (> 10-fold) loss in the inhibitory potency. In-line with earlier observations, with a one-carbon linker [14,21], amide analog 26 displayed a weak enzyme inhibition. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity.

± 2 21a
Molecules 2020, 25, 2805 6 of 22 (racemic compound 23) resulted in a substantial (> 10-fold) loss in the inhibitory potency. In-line with earlier observations, with a one-carbon linker [14,21], amide analog 26 displayed a weak enzyme inhibition. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity. (racemic compound 23) resulted in a substantial (> 10-fold) loss in the inhibitory potency. In-line with earlier observations, with a one-carbon linker [14,21], amide analog 26 displayed a weak enzyme inhibition. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity.

NI b 21b
Molecules 2020, 25, 2805 6 of 22 (racemic compound 23) resulted in a substantial (> 10-fold) loss in the inhibitory potency. In-line with earlier observations, with a one-carbon linker [14,21], amide analog 26 displayed a weak enzyme inhibition. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity. (racemic compound 23) resulted in a substantial (> 10-fold) loss in the inhibitory potency. In-line with earlier observations, with a one-carbon linker [14,21], amide analog 26 displayed a weak enzyme inhibition. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity.

± 53 21n
Molecules 2020, 25, 2805 6 of 22 (racemic compound 23) resulted in a substantial (> 10-fold) loss in the inhibitory potency. In-line with earlier observations, with a one-carbon linker [14,21], amide analog 26 displayed a weak enzyme inhibition. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity. (racemic compound 23) resulted in a substantial (> 10-fold) loss in the inhibitory potency. In-line with earlier observations, with a one-carbon linker [14,21], amide analog 26 displayed a weak enzyme inhibition. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity.

± 63 21c
Molecules 2020, 25, 2805 6 of 22 (racemic compound 23) resulted in a substantial (> 10-fold) loss in the inhibitory potency. In-line with earlier observations, with a one-carbon linker [14,21], amide analog 26 displayed a weak enzyme inhibition. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity. (racemic compound 23) resulted in a substantial (> 10-fold) loss in the inhibitory potency. In-line with earlier observations, with a one-carbon linker [14,21], amide analog 26 displayed a weak enzyme inhibition. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity. Having established that 1-(1-arylethylpiperidin-4-yl)thymine is preferable for inhibitory potency, our efforts were then directed toward the exploration of the biphenyl ether tail. Deletion of the terminal phenoxy group (analog 21a) resulted in a 40-fold decrease in inhibitory potency. Docking studies based on the X-ray co-crystal structure of MtbTMPK with our previously reported 1-(piperidin-4-yl)thymine inhibitor (PDB 5NR7) [14] were performed to rationalize the observed SAR. Docking indicated that the loss of a hydrophobic interaction with Tyr39 accounts for the drop in inhibitory potency ( Figure 3A). The addition of a methylene moiety between the terminal phenyl ring and C-phenoxy group (analog 21n) caused a five-fold decrease in the inhibitory potency. The docking pose of compound 21n in MtbTMPK ( Figure 3B) showed that the elongated benzyloxy phenyl resulted in a weaker hydrophobic interaction with Tyr39 than the phenoxy moiety of 3. Omitting the oxygen atom between the two phenyl rings led to a complete loss of enzyme inhibitory potency (compound 21b). Moreover, moving the terminal phenoxy ring of 3 to the ortho or para-position of ring C (compounds 21c/21e) had a negative effect on the inhibitory activity.  [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency.   [14] active site. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Based on these results, further optimization efforts focused on the substitution pattern of the terminal phenyl ring of 3 and 21n. As shown in Table 2, most of the substituted analogs exhibited a small decrease in inhibitory activity compared to 3. Of note, the introduction of sterically demanding electron withdrawing substituents (21k/21l/21m) resulted in a significant drop in MtbTMPK inhibitory potency. Interestingly, the introduction of a 3-chloro (21j) but not a 4-chloro substituent (21h) afforded a significant improvement in the inhibitory potency due to an edge-to-face π-stacking interaction between the 3-chlorobenzene ring and Tyr39, as found for compound 3 (Figure 4). Introduction of a 3-chloro substituent in the benzyl analog 21n also had a beneficial impact on the inhibitory activity. Interestingly, the introduction of a 3-chloro (21j) but not a 4-chloro substituent (21h) afforded a significant improvement in the inhibitory potency due to an edge-to-face π-stacking interaction between the 3-chlorobenzene ring and Tyr39, as found for compound 3 (Figure 4). Introduction of a 3-chloro substituent in the benzyl analog 21n also had a beneficial impact on the inhibitory activity. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Finally, all compounds were evaluated for their in vitro antimycobacterial activity (Table 3). Consistent with the observations on the enzyme inhibitory activities, modifications of the scaffold of 3 did not yield analogs with superior antimycobacterial activity (21a-21c, 21e, 21n, 23 and 26). Remarkably, analog 27, in which the thymine moiety is replaced by a phenyl ring, showed potent antimycobacterial activity but lacked selectivity, as evidenced by its equipotent cytotoxicity. The introduction of substituents on the distal phenyl ring of 3 afforded several analogs with improved antimycobacterial activity (21f-21j). Substitution of the D-ring of the benzyloxy analog 21n also contributed to the growth inhibitory activity. However, the selectivity vs. MRC-5 fibroblasts was modest. Table 3. Antimycobacterial activity against H37Rv and cytotoxicity against MRC-5 fibroblasts of the compounds in this study. All residues interacting with the inhibitors, including the hydrophobic contact (gray wire) and hydrogen-bonding interaction (residues in orange wire, hydrogen bonds indicated in magenta), were calculated using LigPlus [31]. Illustration was created using Chimera [32].
Finally, all compounds were evaluated for their in vitro antimycobacterial activity (Table 3). Consistent with the observations on the enzyme inhibitory activities, modifications of the scaffold of 3 did not yield analogs with superior antimycobacterial activity (21a-21c, 21e, 21n, 23 and 26). Remarkably, analog 27, in which the thymine moiety is replaced by a phenyl ring, showed potent antimycobacterial activity but lacked selectivity, as evidenced by its equipotent cytotoxicity. The introduction of substituents on the distal phenyl ring of 3 afforded several analogs with improved antimycobacterial activity (21f-21j). Substitution of the D-ring of the benzyloxy analog 21n also contributed to the growth inhibitory activity. However, the selectivity vs. MRC-5 fibroblasts was modest.

Computational Studies
For the molecular modeling, X-ray structure of the MtbTMPK (PDB entry 5NR7 [14]) was analyzed using AutoDock vina and AutodockTools-1.5.6. [35]. In ChemDraw 3D 16.0, the PDB files of all ligands were generated after the energy was minimized (minimum RMS gradient: 0.001). The PDBQT file of the ligands and receptors were prepared by AutodockTools-1.5.6, including atom types, atomic partial charges and the information on the ligand torsional degrees. Using a grid spacing of 0.375 and 60 × 60 × 60 numbers of grid points, the prepared PDBQT files of ligands and receptors were docked (centered on the MtbTMPK active site PHE70 CE2, the coordinates x, y and z were −0.997, 26.240 and −4.528, correspondingly) through the Lamarckian 4.2 method. Each ligand was docked in Autodock vina 3 times, with each time generating 20 possible conformations. Chimera in combination with LigPlus were used to analyze the results.

In Vitro Antituberculosis Assay
The MIC values of all compounds were determined as previously described [36]. In brief, M. tuberculosis H37Rv (ATCC 27294) was grown to optical density at 650 nm wavelength (OD 650nm ) 0.2 in Middlebrook 7H9 medium supplemented with 0.4% glucose, 0.03% Bacto casitone and 0.05% Tyloxapol (7H9/glucose/casitone/Tyloxapol) prior to further 1000-fold dilution in fresh medium. Drugs were 2-fold serially diluted in duplicate in 7H9/glucose/casitone/Tyloxapol (50 µL/well) in a concentration range spanning 100-0.049 µM in sterile 96-well U-bottom clear polystyrene microtiter plates. Isoniazid and DMSO as positive and negative controls, respectively. An equal volume (50 µL) of diluted cells was added to the plates with the serial drug dilution. Plates were sealed in Ziplock bags and incubated at 37 • C. After 7-14 days, plates were read with an enlarging inverted mirror plate reader. The MIC was recorded as the concentration that fully inhibited all visible growth.

In Vitro Cytotoxicity Assay
The cytotoxicity of compounds on MRC-5 fibroblasts was performed exactly as previously reported [14].

Chemistry
All reagents and solvents were purchased from standard commercial sources and were of analytical grade. All synthetic compounds described in this study were checked with analytical TLC (Macherey−Nagel precoated F254 aluminum plates, Düren, Germany), visualized under UV light at 254 nm and purified by column chromatography (CC) on a Reveleris X2 (Grace, BÜCHI, Flawil, Switzerland) automated flash unit. All final compounds and some intermediates were measured with Varian Mercury 300/75 MHz (Palo Alto, CA, USA) or a Bruker AVANCE (Fällanden, Zürich, Switzerland) Neo ® 400/100 MHz spectrometer at 298.15 K using tetramethylsilane (TMS) as an internal standard. The analysis and confirmation of the final compounds were conducted with 1 H, 13 C, HSQC and HMBC NMR spectral data (Supplementary Materials). High-resolution mass spectrometry was performed on a Waters LCT Premier XE TM (Waters, Zellik, Belgium) time-of-flight (TOF) mass spectrometer equipped with a standard electrospray ionization (ESI) and modular LockSpray TM interface (Waters, Zellik, Belgium). The purity of the tested compounds was determined by LC-MS analysis using a Waters AutoPurification system equipped with a Waters Cortecs C18 column (2.7 µm, 100 × 4.6 mm), as was a gradient system of formic acid in H 2 O (0.2%, v/v)/MeCN with a gradient of 95:5 to 0:100 in 6.5 min at a flow rate of 1.44 mL/min. General procedure A: Synthesis of biphenyl ether aldehyde building blocks. According to a literature report [26], hydroxyphenylacetic ester derivatives (1.0 eq), phenylboronic acid derivatives (3.0 eq.), Cu(OAc) 2 (2.0 eq.), 4Å molecular sieves (0.18 g/mmol ester) and pyridine (3.0 eq.) in 1,2-dichloroethane (6.0 mL/mmol ester) afforded the biphenyletheracetic ester intermediates. To a solution of the biphenyletheracetic ester intermediates (1.0 eq.) in dry tetrahydrofuran (THF) (6.0 mL/mmol ester intermediate) was added LiAlH 4 (2.0 eq.) at 0 • C under N 2 atmosphere, and the resulting mixture was then stirred at room temperature for 1 h [23]. After complete consumption of the starting material, the reaction mixture was quenched with aq. Na + /K + tartrate solution (5.0 mL/mmol LiAlH 4 ), and the mixture was stirred at room temperature overnight and then filtered. The collected filtrate was dried and concentrated to afford a crude alcohol intermediate, which was oxidized by PCC (2.0 eq.) for 2 h in dichloromethane (DCM) (5.0 mL/mmol PCC) [24,25]. The reaction mixture was filtered through a short silica column. The collected filtrate was concentrated in vacuo and used in the next step without any additional purification.