Synthesis and Structure–Activity Relationship Studies of Pyrido [1,2-e]Purine-2,4(1H,3H)-Dione Derivatives Targeting Flavin-Dependent Thymidylate Synthase in Mycobacterium tuberculosis

In 2002, a new class of thymidylate synthase (TS) involved in the de novo synthesis of dTMP named Flavin-Dependent Thymidylate Synthase (FDTS) encoded by the thyX gene was discovered; FDTS is present only in 30% of prokaryote pathogens and not in human pathogens, which makes it an attractive target for the development of new antibacterial agents, especially against multi-resistant pathogens. We report herein the synthesis and structure-activity relationship of a novel series of hitherto unknown pyrido[1,2-e]purine-2,4(1H,3H)-dione analogues. Several synthetics efforts were done to optimize regioselective N1-alkylation through organopalladium cross-coupling. Modelling of potential hits were performed to generate a model of interaction into the active pocket of FDTS to understand and guide further synthetic modification. All those compounds were evaluated on an in-house in vitro NADPH oxidase assays screening as well as against Mycobacterium tuberculosis ThyX. The highest inhibition was obtained for compound 23a with 84.3% at 200 µM without significant cytotoxicity (CC50 > 100 μM) on PBM cells.


Introduction
The excessive use of antibiotics in humans and animals has led to the appearance of multi-resistant bacteria (BMRs) and as a consequence to an increased mortality [1,2]. In 2014, the World Health Organization warned of a risk of antibiotic shortages by 2050 if nothing was done. Today, E. coli, K. pneumoniae and S. aureus are resistant to more than 50% of the main antibacterial drugs. From a simple natural genetic evolution, the situation has become a global public health problem, promoting the discovery of new therapeutic targets in order to develop new antibacterial substances. In most eubacteria, plants and eukaryotic cells, thymidylate synthases (TS or ThyA) [3] provide the only de novo source of 2 -deoxythymidine-5 -monophosphate (dTMP) required for DNA synthesis. The activity of these enzymes is pivotal for bacterial DNA replication and repair. Reductive methylation of 2 -deoxyuridine-5 -monophosphate (dUMP) to 2 -deoxythymidine-5 -monophosphate (dTMP) was catalyzed by three coupling enzymes of folate metabolism [4]. TS catalyzes methylation by using (6R)-N 5 ,N 10 -methylene-5,6,7,8-tetrahydrofolate (CH 2 THF) as a methylene and hydride donor, which results in the formation of dTMP and 7,8-dihydrofolate (DHF). Dihydrofolate reductase (DHFR) catalyzes reduction of DHF to THF and the serinehydroxymethyl transferase (SHMT) catalyzes the serine-glycine conversion which is concomitant to the a methylene and hydride donor, which results in the formation of dTMP and 7,8-dihydrofolate (DHF). Dihydrofolate reductase (DHFR) catalyzes reduction of DHF to THF and the serinehydroxymethyl transferase (SHMT) catalyzes the serine-glycine conversion which is concomitant to the conversion of THF to CH2THF as cofactors. TS inhibitors were used as cytotoxic agents, but the lack of selective TS bacterial inhibitors over human has hampered their application. More recently, a new class of TS was discovered [5]. This enzyme is encoded by the ThyX gene (formerly Thy1) and is absent in the vast majority of eukaryotic cells and only present in approximately 30% of gram-positive or negative pathogenic prokaryotes [6]. It is a flavin-dependent thymidylate synthase (FDTS), which has a unique mechanism, structure and gene sequence, making it an attractive new therapeutic target for the development of new selective bioactive compounds. FDTS uses a FAD as methylene career intermediate but also as hydride donor through the reduced form FADH2 obtained from NADPH oxidase activity. Unlike human TS, FDTS produces tetrahydrofolate (H4folate) indeed of dihydrofolate (H2folate) [7]. Thus, FDTS can catalyze multiple biotransformation reactions in comparison to the classical TS. So far, few ThyX inhibitors have been reported like the 5-fluoro-2′-deoxyuridine-5′-monophosphate (5-FdUMP) and Raltitrexed ® , but their poor selectivity between ThyA and ThyX has hampered their development as new drugs [8,9]. The thiazolidine analog 1 (Figure 1), reported by Myllykalio et al. [10], exhibited an IC50 of 0.057 µM against ThyX. The same group reported a second series of FDTS inhibitors discovered from high-throughput screening (HTS) of natural and synthetic compounds [11]. After synthetic modifications and biological evaluation, the benzoquinone analogue 2 was reported to have an inhibitory Ki value of 28 nM against ThyX. Those molecules did not show any mitochondrial toxicity. However, the major matter of discussion is linked to the quinone properties in medicinal chemistry that allow strong redox stress and Michaël acceptor which could conduct to cellular damage and cell protein alkylation, [12]. C 5 -modifed dUMP analogs were described by Herdewijn et al. [13]; among them, compound 3 exhibited an IC50 value of 0.91 µM. More recently, new series of inhibitors bearing benzo [b][1,4]oxazin-3(4H)-one scaffold was reported by highthroughput screening of commercially available compounds. Structure activity was performed and led to compound B1-PP146 (4) with an IC50 value of 0.69 µM against ThyX [14]. FDTS is thus an attractive antibacterial target to the development of new and specific drugs to overcome bacterial resistance. By screening of our in house-synthesized compound libraries, two new pyridopurines analogs 5 and 6 were found to exhibit 23.1% and FDTS is thus an attractive antibacterial target to the development of new and specific drugs to overcome bacterial resistance. By screening of our in house-synthesized compound libraries, two new pyridopurines analogs 5 and 6 were found to exhibit 23.1% and 33.2% ThyX inhibitory effect at 200 µM, respectively. Based on those results, we decided to use pyrido [1,2-e]purine-2,4(1H,3H)-dione as scaffold to new and more active compounds through diversity-oriented synthesis. The current study aimed at identifying substituents at N 1 and/or N 3 positions, which could increase ThyX inhibitory activity. In this manuscript, the FDTS enzyme from M. tuberculosis (MtbThyX) was chosen for biological and docking assays due the emergence of multidrug-resistant strains of Mycobacterium tuberculosis (MDR-TB) [15].

In Vitro Mycobacterial Thyx Inhibition Assay for Structure-Activity Relationship Studies
To investigate Structure-Activity Relationship (SAR) in a systematic way, we used a NADPH oxidase spectrophotometric assay, adapted from Basta et al. [16], to test the in vitro inhibitory activities of synthesized compounds on Mtb ThyX, and discuss the influence of substituents. The most active of our compounds was used as reference scaffold for future optimization. Assay reactions, with a final volume of 100 µl, consisted of 750 µM NADPH, 100 µM dUMP, 2 mM MgCl 2 , 1% glycerol, 50 µM FAD and 10 µM of ThyX. The reactions were initiated by injection of NADPH (15 µL of a 5 mM solution) to each well of the microtiter 96-well clear flat-bottom plate, followed by rapid shaking of the microplate. ThyX activity was determined by following the decrease of absorbance at λ 340 (due to oxidation of NADPH). All assays used a kinetic mode of a multilabel microplate reader with an injector. The primary screen was performed with molecules dissolved in DMSO, including DMSO alone as low-activity control. B1-PP146 (4), with a 1,4-benzoxazine moiety and described as tight-binding inhibitors, was used as reference compound [17]. All screening reactions were performed in duplicates at 200 µM concentration.

33.2%
ThyX inhibitory effect at 200 µM, respectively. Based on those results, we decided to use pyrido [1,2-e]purine-2,4(1H,3H)-dione as scaffold to new and more active compounds through diversity-oriented synthesis. The current study aimed at identifying substituents at N 1 and/or N 3 positions, which could increase ThyX inhibitory activity. In this manuscript, the FDTS enzyme from M. tuberculosis (MtbThyX) was chosen for biological and docking assays due the emergence of multidrug-resistant strains of Mycobacterium tuberculosis (MDR-TB) [15].

In Vitro Mycobacterial Thyx Inhibition Assay for Structure-Activity Relationship Studies
To investigate Structure-Activity Relationship (SAR) in a systematic way, we used a NADPH oxidase spectrophotometric assay, adapted from Basta et al. [16], to test the in vitro inhibitory activities of synthesized compounds on Mtb ThyX, and discuss the influence of substituents. The most active of our compounds was used as reference scaffold for future optimization. Assay reactions, with a final volume of 100 µl, consisted of 750 µM NADPH, 100 µM dUMP, 2 mM MgCl2, 1% glycerol, 50 µM FAD and 10 µM of ThyX. The reactions were initiated by injection of NADPH (15 µL of a 5 mM solution) to each well of the microtiter 96-well clear flat-bottom plate, followed by rapid shaking of the microplate. ThyX activity was determined by following the decrease of absorbance at λ340 (due to oxidation of NADPH). All assays used a kinetic mode of a multilabel microplate reader with an injector. The primary screen was performed with molecules dissolved in DMSO, including DMSO alone as low-activity control. B1-PP146 (4), with a 1,4-benzoxazine moiety and described as tight-binding inhibitors, was used as reference compound [17]. All screening reactions were performed in duplicates at 200 µM concentration.
The synthesized compounds, pyrido [1,2-e]purine-2,4(1H,3H)-dione 9a-j and thio derivative 11 substituted at N 3 -position by phenyl, were evaluated for their abilities to inhibit the ThyX enzyme (Table 1)   To probe into the optimal scaffold of central heterocycle, the phenyl para substitution was studied. With small electron-withdrawing groups, such as fluorine (9a), chlorine (9b) and bromine (9c), a 49-59% inhibition was obtained with the best value for the fluorine derivative. By increasing hydrophobic and electron-withdrawing effects (9e), less inhibitory activity was observed (31.7%). Substitution at para position with electro-donating group (-OCH3) 9f decreases the activity to 8.1% whereas the methyl group (9d) led to 55% inhibition. On the other hand, fluorine analogs of 9a, by substitution at meta position (9g) or a second fluorine substitution at ortho/para (9h) or meta/para (9i) positions, results in loss of ThyX inhibition. Polysubstitution with electron-donating group such as methyl groups (9j) also display low enzyme inhibition. We investigate the substitution of oxygen by sulfur atom at 2 position, and we observe that compound 11 decreases drastically the inhibitory effect from 59.4% to 19.8% to compare with 9a. We also looked at the influence of flexibility induced by benzylic substitution (Table 2). Both benzyl analogs 10a-d and 12 showed a significant loss of activity (<25%) in comparison with the aryl derivatives 9. To probe into the optimal scaffold of central heterocycle, the phenyl para substitution was studied. With small electron-withdrawing groups, such as fluorine (9a), chlorine (9b) and bromine (9c), a 49-59% inhibition was obtained with the best value for the fluorine derivative. By increasing hydrophobic and electron-withdrawing effects (9e), less inhibitory activity was observed (31.7%). Substitution at para position with electro-donating group (-OCH 3 ) 9f decreases the activity to 8.1% whereas the methyl group (9d) led to 55% inhibition. On the other hand, fluorine analogs of 9a, by substitution at meta position (9g) or a second fluorine substitution at ortho/para (9h) or meta/para (9i) positions, results in loss of ThyX inhibition. Polysubstitution with electron-donating group such as methyl groups (9j) also display low enzyme inhibition. We investigate the substitution of oxygen by sulfur atom at 2 position, and we observe that compound 11 decreases drastically the inhibitory effect from 59.4% to 19.8% to compare with 9a. We also looked at the influence of flexibility induced by benzylic substitution (Table 2). Both benzyl analogs 10a-d and 12 showed a significant loss of activity (<25%) in comparison with the aryl derivatives 9.  To probe into the optimal scaffold of central heterocycle, the phenyl para substitution was studied. With small electron-withdrawing groups, such as fluorine (9a), chlorine (9b) and bromine (9c), a 49-59% inhibition was obtained with the best value for the fluorine derivative. By increasing hydrophobic and electron-withdrawing effects (9e), less inhibitory activity was observed (31.7%). Substitution at para position with electro-donating group (-OCH3) 9f decreases the activity to 8.1% whereas the methyl group (9d) led to 55% inhibition. On the other hand, fluorine analogs of 9a, by substitution at meta position (9g) or a second fluorine substitution at ortho/para (9h) or meta/para (9i) positions, results in loss of ThyX inhibition. Polysubstitution with electron-donating group such as methyl groups (9j) also display low enzyme inhibition. We investigate the substitution of oxygen by sulfur atom at 2 position, and we observe that compound 11 decreases drastically the inhibitory effect from 59.4% to 19.8% to compare with 9a. We also looked at the influence of flexibility induced by benzylic substitution (Table 2). Both benzyl analogs 10a-d and 12 showed a significant loss of activity (<25%) in comparison with the aryl derivatives 9. tration, through tritium release assay (Table 3). Without FAD adding, only 1.1-1.5 of 4 active sites were occupied by this natural cofactor. Without excess of FAD, we can detect a competition effect between FAD and potential inhibitors. By decreasing inhibitors concentration to 50 µM, we can study the true potential of our molecules. The tritium release assay permitted us to study inhibition on the second mechanism part (methylene transfer). Compound 9a showed better activity, by increasing inhibition to 76.1% at 200 µM in absence of FAD, which was not the case for compounds 9b and 9d, for which the inhibition effect had decreased. With a 50 µM inhibitor concentration, only compound 9a provided an equal inhibitory activity to the reference compounds 5 and 6, and no inhibition was observed for 9b and 9d. Through a tritium release assay, compounds 9b and 9d exhibited increased inhibitory activities on ThyX with 77.1% and 69.2% at 200 µM, respectively, which suggested that these molecules would be more active in the methylene transfer mechanism. On the other hand, compound 9a showed similar inhibition on Mtb ThyX (59.5% at 200 µM, Table 3) to the NADPH oxidase assay (59.5% at 200 µM, Table 1). Because most of the active compounds on the first part of the mechanism were also active on the second part (tritium release assay), NADPH oxidase assay was taken as a reference test to determine inhibitory activities and compound 9a as the scaffold to perform other modifications in search of better activity.

Modification at the N 1 Position of Compound 9 Taken as Scaffold and SAR Studies
Starting from 9a as scaffold, a library of molecules with structural modifications through N 1 -alkylation with various benzyl groups was obtained (13a-f) (Scheme 2). In the first attempt, in the presence of K 2 CO 3 , in DMF at room temperature during 12 h a competition between N 1 -and O 2 -alkylation was observed (by HMBC-NMR), which decreased the yield and deteriorated the purification. In this case, we investigated the solvent effects (DMF, THF), base influence (K 2 CO 3 , Cs 2 CO 3 , LiH, NaH, etc.,) and activation mode (microwave, sonication) to enhance the N 1 -regioisomers. We observed that only the N 1 -alkylated regioisomer was obtained under microwave activation at 120 • C during 20 min with potassium carbonate and molecular sieves in presence of various bromobenzyl analogs (Scheme 2). Compound 9a showed better activity, by increasing inhibition to absence of FAD, which was not the case for compounds 9b and 9d, fo tion effect had decreased. With a 50 µM inhibitor concentration, only vided an equal inhibitory activity to the reference compounds 5 and 6 was observed for 9b and 9d. Through a tritium release assay, compo hibited increased inhibitory activities on ThyX with 77.1% and 69.2% tively, which suggested that these molecules would be more active in t fer mechanism. On the other hand, compound 9a showed similar inhi (59.5% at 200 µM, Table 3) to the NADPH oxidase assay (59.5% at 20 cause most of the active compounds on the first part of the mechanis on the second part (tritium release assay), NADPH oxidase assay was test to determine inhibitory activities and compound 9a as the scaffo modifications in search of better activity.

Modification at the N 1 Position of Compound 9 Taken as Scaffold and SA
Starting from 9a as scaffold, a library of molecules with struc through N 1 -alkylation with various benzyl groups was obtained (13afirst attempt, in the presence of K2CO3, in DMF at room temperature petition between N 1 -and O 2 -alkylation was observed (by HMBC-NMR the yield and deteriorated the purification. In this case, we investigate (DMF, THF), base influence (K2CO3, Cs2CO3, LiH, NaH, etc.,) and activ wave, sonication) to enhance the N 1 -regioisomers. We observed that on regioisomer was obtained under microwave activation at 120 °C duri tassium carbonate and molecular sieves in presence of various bro (Scheme 2). The obtained benzylic derivatives 13a-f were then evaluated against FDTS (Table 4). The obtained benzylic derivatives 13a-f were then evaluated against FDTS (Table 4). We observed that the presence of benzyl group (13a), and substituted derivatives at the para position by methyl (13b) or electron-withdrawing group (13c-f) result in lack of inhibition activity compared to the reference 9a (59.4%). From this study, further modifications at C7 and C8 moiety of compound 9a with free N 1 -position were thus performed.

Modification at the C 7 and C 8 position of Compound 9a and SAR Studies
After modulation on the pyrimidine-2,4-dione ring, we investigated the influence of various substituents on the aromatic ring at C 7 and C 8 positions. We first synthesized under microwave activation two series of substituted molecules by methyl group (18a,b) or bromine (19a,b) (Scheme 3). Halogen derivatives 19a,b were used as starting material to develop a library of 24 molecules through palladium cross-coupling reactions (Scheme 4).

Modifications by Sonogashira Cross-Coupling Reaction
Several substituents were introduced at C 7 and C 8 positions of 19a,b under different palladium cross-coupling reactions. Sonogashira conditions [22] were investigated under different activation modes (thermic [23], ultrasonication [24], and microwaves [25]). Under microwave irradiations in DMF with triethylamine, in presence of CuI and Pd(PPh3)4, We observed that the presence of benzyl group (13a), and substituted derivatives at the para position by methyl (13b) or electron-withdrawing group (13c-f) result in lack of inhibition activity compared to the reference 9a (59.4%). From this study, further modifications at C 7 and C 8 moiety of compound 9a with free N 1 -position were thus performed.

Modification at the C 7 and C 8 position of Compound 9a and SAR Studies
After modulation on the pyrimidine-2,4-dione ring, we investigated the influence of various substituents on the aromatic ring at C 7 and C 8 positions. We first synthesized under microwave activation two series of substituted molecules by methyl group (18a,b) or bromine (19a,b) (Scheme 3). Halogen derivatives 19a,b were used as starting material to develop a library of 24 molecules through palladium cross-coupling reactions (Scheme 4).
The obtained benzylic derivatives 13a-f were then evaluated against FDTS (Table 4). We observed that the presence of benzyl group (13a), and substituted derivatives at the para position by methyl (13b) or electron-withdrawing group (13c-f) result in lack of inhibition activity compared to the reference 9a (59.4%). From this study, further modifications at C7 and C8 moiety of compound 9a with free N 1 -position were thus performed.

Modification at the C 7 and C 8 position of Compound 9a and SAR Studies
After modulation on the pyrimidine-2,4-dione ring, we investigated the influence of various substituents on the aromatic ring at C 7 and C 8 positions. We first synthesized under microwave activation two series of substituted molecules by methyl group (18a,b) or bromine (19a,b) (Scheme 3). Halogen derivatives 19a,b were used as starting material to develop a library of 24 molecules through palladium cross-coupling reactions (Scheme 4).

Modifications by Sonogashira Cross-Coupling Reaction
Several substituents were introduced at C 7 and C 8 positions of 19a,b under different palladium cross-coupling reactions. Sonogashira conditions [22] were investigated under different activation modes (thermic [23], ultrasonication [24], and microwaves [25]). Under microwave irradiations in DMF with triethylamine, in presence of CuI and Pd(PPh3)4,  All those compounds were then evaluated (Table 5). Table 5. SAR of the C 7 and C 8 moiety by Sonogashira cross-coupling and Mtb ThyX inhibitio µM by the NADPH oxidase assay.
All those compounds were then evaluated (Table 5). the aromatic analogues 20 and 21a,b and aliphatic alkynes 20 and 21c-g derivatives were obtained in moderate to good yield (40 to 92%) (Scheme 4). All those compounds were then evaluated (Table 5). All those compounds were then evaluated (Table 5). The different substituents on C 7 and C 8 position were chosen for their abilities to generate flexibility, to create a hydrogen bond, hydrophobic interactions and π-stacking interactions with the active site of the enzyme. Modifications at the C 7 -position led to derivatives with aromatic alkynes, which induce flexibility, the presence of aliphatic chain at para position (compounds 20a and 20b respectively) or aliphatic chain with 5,6 and 7 (compounds 20c, 20d and 20e respectively) which have shown no to low inhibition still below derivatives with aromatic alkynes, which induce flexibility, the presence of aliphatic chain at para position (compounds 20a and 20b respectively) or aliphatic chain with 5, 6 and 7 (compounds 20c, 20d and 20e respectively) which have shown no to low inhibition still below that of the 9a activity. Compounds 20f and 20g, substituted by side-chain presenting hydrogen-bound site with amide and urea moiety (respectively) and hydrophobic chain, did not show more efficiency against ThyX. The same substituents were introduced at C 8 -position of 9a. Even if most of the molecules do not present an inhibitory activity higher than 33%, this position is however more favorable than that in C 7 . In fact, we observe for the aromatic group derivatives 21a and 21b, higher inhibitions than the position C 7 , and especially for 21b inhibitory activity was higher (69.1% at 200 µM) than 9a (59.4%). Overall, this SAR study reveals that the aromatic ring at C 8 -position increases inhibitory activities as well as maybe some flexibility. π-stacking interactions seem to be predominant in order to increase the activity, which is why we have synthesized a new series of molecules bearing aromatic groups by the Suzuki-Miyaura cross-coupling reaction. Expected molecules were planar and could be intercalated between the different co-factors into the enzyme pocket.
para position (compounds 20a and 20b respectively) or aliphatic chain with 5 pounds 20c, 20d and 20e respectively) which have shown no to low inhibit that of the 9a activity. Compounds 20f and 20g, substituted by side-chain drogen-bound site with amide and urea moiety (respectively) and hydroph not show more efficiency against ThyX. The same substituents were introd sition of 9a. Even if most of the molecules do not present an inhibitory activ 33%, this position is however more favorable than that in C7. In fact, we o aromatic group derivatives 21a and 21b, higher inhibitions than the positi pecially for 21b inhibitory activity was higher (69.1% at 200 µM) than 9a (5 this SAR study reveals that the aromatic ring at C 8 -position increases inhib as well as maybe some flexibility. π-stacking interactions seem to be predom to increase the activity, which is why we have synthesized a new series of m ing aromatic groups by the Suzuki-Miyaura cross-coupling reaction. Expec were planar and could be intercalated between the different co-factors in pocket.

Modifications by the Suzuki-Miyaura Cross-Coupling Reaction
Modifications under the Suzuki-Miyaura cross-coupling conditions performed under microwave irradiations [27,28] in presence of Cs2CO3, P various boronic acid derivatives in DMF at 120 °C for 40 min. A library substituted by aromatic derivatives was isolated (22,23a-e) in 11 to 50% yie All those compounds were then evaluated (Table 6). All those compounds were then evaluated ( Table 6). The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure and π-stacking interactions looked to be the best characteristic to develop potential Mycobacterium tuberculosis ThyX inhibitors.  The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure  The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure  The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure  The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure  The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure  The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure  The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure  The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure  The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure 35  The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure 29.4 23e H  The aromatic substitution at C 7 -position-bearing with hydrogen or methyl (compound 22a and 22b respectively) showed low (22a: 20.7%) to moderate (22b: 45.7%) inhibition. With electron-withdrawing phenylic substitution such as methoxy, trifluoromethyl ether or fluorine (22c, 22d and 22e respectively) inhibitory activities do not exceed 28%. At C 8 position, the same substitutions were realized with electron-donating substituents at phenylic para position (compounds 23a and 23b). With 23b equal inhibition was observed as the C 7 position. For 23a a promising inhibition was observed with 84.3% at 200 µM. For electron-withdrawing substituents at phenylic para position, higher inhibitory activities were shown with methoxy (35.4%) and fluorine (43.1%) group (compound 23c and 23e respectively). In the case of trifluoromethyl ether 23d, equal inhibition was observed. Overall, this SAR study revealed that small electron-withdrawing and donating groups showed only less to reasonable inhibitions. One of those compounds with phenyl substitution on the C 8 showed the best inhibition of this study at 200 µM. Planar structure 43.1

B1-PP146 a
95.1 b a B1-PP146, a compound with a 1,4 benzoxazine ring, used as positive control; b value at 50 µM. The standard deviations from two independent experiments is between 3 and 9, as a function of the molecules.

Mycobacterial ThyX Docking Studies
dUMP was docked into the binding site of ThyX from the Mycobacterium tuberculosis complex ( Figure 2) with FAD, dUMP (PDB: 3GWC) as the starting structure to perform an initial molecular study. dUMP binds in the FDTS pocket by π-stacking with FAD, but also by several hydrogen bonds [29]. On the pyrimidine-2,4-dione ring with four hydrogen bonds: two between C 2 =O and Arg199; one between N 3 and Arg199; and finally one between C 4 =O and Arg107. On the 2 -deoxyribosyl part, only two hydrogen bonds were created between 3 -OH, Arg95 and Gln103. The last hydrogen bonds were localized on the phosphate moiety of dUMP. This part of the natural subtract acted like an anchor into active site; six hydrogen bonds were reported between oxygens of phosphonic group and Arg87 (2 bonds), Gln106 (1 bond), Arg172 (2 bonds) and Arg107 (1 bond).
also by several hydrogen bonds [29]. On the pyrimidine-2,4-dione ring with four hydrogen bonds: two between C 2 =O and Arg199; one between N 3 and Arg199; and finally one between C 4 =O and Arg107. On the 2′-deoxyribosyl part, only two hydrogen bonds were created between 3′-OH, Arg95 and Gln103. The last hydrogen bonds were localized on the phosphate moiety of dUMP. This part of the natural subtract acted like an anchor into active site; six hydrogen bonds were reported between oxygens of phosphonic group and Arg87 (2 bonds), Gln106 (1 bond), Arg172 (2 bonds) and Arg107 (1 bond).

Figure 2. Hydrogen bond and hydrophobic interactions between the binding site of ThyX from
Mycobacterium tuberculosis complex and natural substrate dUMP. The green dotted lines depict the hydrogen bonds whereas the other residues represent the hydrophobic interactions with the respective compounds.
Following this molecular modeling, compounds 9a ( Figure 3) and 23a (Figure 4) with the highest inhibitions and results at NADPH oxidase assay and tritium release assay, were docked into the FAD and dUMP pocket with 13 amino acid residues. Compound 9a (Figure 3) was the first docked into Mycobacterium tuberculosis pocket. It showed reasonable inhibition at 200 µM (59.4%), a reduced activity at 50 µM (18.4%). By diminution of [FAD] it showed higher inhibition (76.1%) that suggested possible competition with the natural co-factor. Compound 9a took place closure than the FAD, its fluorophenyl group near the central ring of FAD, which let assumed π-stacking interactions. On the other hand, 2D representation into the pocket showed several hydrogen bond interactions with amino acid residues. Compound 9a created five interactions with amino acid responsible of hydrogen bond formation with dUMP. The pyrimidine ring created three interactions with Leu104 and Gln103 on its N 1 position. One bond was created with the C 4 =O and Arg172. Following this molecular modeling, compounds 9a ( Figure 3) and 23a (Figure 4) with the highest inhibitions and results at NADPH oxidase assay and tritium release assay, were docked into the FAD and dUMP pocket with 13 amino acid residues. Compound 9a (Figure 3) was the first docked into Mycobacterium tuberculosis pocket. It showed reasonable inhibition at 200 µM (59.4%), a reduced activity at 50 µM (18.4%). By diminution of [FAD] it showed higher inhibition (76.1%) that suggested possible competition with the natural co-factor. Compound 9a took place closure than the FAD, its fluorophenyl group near the central ring of FAD, which let assumed π-stacking interactions. On the other hand, 2D representation into the pocket showed several hydrogen bond interactions with amino acid residues. Compound 9a created five interactions with amino acid responsible of hydrogen bond formation with dUMP. The pyrimidine ring created three interactions with Leu104 and Gln103 on its N 1 position. One bond was created with the C 4 =O and Arg172. The last two hydrogen bonds were placed on the N 5 with Arg172 and on F with Arg199. The loss of inhibition could be explained by the large size of the S atom of compound 11, which then changes the positioning of the substrate in the active site of the protein. Delocalization 9g and polysubstitution 9h, 9i and 9j changed the capacity to create the hydrogen bond, but also the ring aromaticity. When fluorine was substituted by trifluorimethyl ether 9e, the capacity to create bond hydrogen was replaced by a hydrophobic interaction, leading to a loss of inhibitory activity. With a smaller group like methyl 9d, inhibition stays equal to compound 9a. Hydrophobic interaction with a small group was equal to hydrogen bond interaction in this case. With methoxy substitution 9f, the loss of inhibition could result from the change of bond angle. Insertion of carbon between the pyrimidinedione ring and phenyl 10a-d, 12 increase flexibility of the structure and broke the planar scaffold. The loss of inhibition observed on N1-alkylated compound 13a-h could be explained by the loss of the two hydrogen interactions from this position. Compound 23a (Figure 4) was docked into Mycobacterium tuberculosis ThyX pocket. This compound presented the highest inhibition at 200 µM.
hydrogen bond interaction in this case. With methoxy substitution 9f, the loss of inhibition could result from the change of bond angle. Insertion of carbon between the pyrimidinedione ring and phenyl 10a-d, 12 increase flexibility of the structure and broke the planar scaffold. The loss of inhibition observed on N1-alkylated compound 13a-h could be explained by the loss of the two hydrogen interactions from this position. Compound 23a (Figure 4) was docked into Mycobacterium tuberculosis ThyX pocket. This compound presented the highest inhibition at 200 µM.

Figure 3. Hydrogen bond and hydrophobic interactions between the binding site of ThyX from
Mycobacterium tuberculosis complex and 9a. The green dotted lines depict the hydrogen bonds whereas the other residues represent the hydrophobic interactions with the respective compounds.  Compared to 9a, compound 23a looked to be less stabilized into the pocket because of three hydrogen bonds interactions' loss. It still conserved two major hydrogen bond interactions with Arg95 (hydrogen bond interaction with 3′OH of dUMP) with C 4 =O and Arg172 (hydrogen bond with phosphonic acid part of dUMP) with C 2 =O. By looking at 3D modeling, we observed three new interactions which could explain biological results. Compared to 9a, compound 23a looked to be less stabilized into the pocket because of three hydrogen bonds interactions' loss. It still conserved two major hydrogen bond interactions with Arg95 (hydrogen bond interaction with 3 OH of dUMP) with C 4 =O and Arg172 (hydrogen bond with phosphonic acid part of dUMP) with C 2 =O. By looking at 3D modeling, we observed three new interactions which could explain biological results. Two π-stacking interactions from the phenyl with Tyr108 and His203. Another π-stacking could be observed on the p-fluorophenyl part with His91. The loss of inhibition by decreasing FAD concentration could be explained by the possible π-stacking interaction between isoalloxazine and central ring of pyrido [1,2-e]purine-2,4(1H,3H)-dione. However, one of the biggest difficulties with molecular docking of weak inhibition was the position into the pocket. Some of them were reproducible but a small modification could change everything; the large flexible pocket of FDTS was responsible for this issue.

Chemistry General Section
Commercially available chemicals were provided as reagent grade and used as received. Some reactions requiring anhydrous conditions were carried out using oven-dried glassware and under an atmosphere of dry Argon. All anhydrous solvents were provided from commercial sources as very dry reagents. The reactions were monitored by thin layer chromatography (TLC) analysis using silica gel precoated plates (Kieselgel 60F254, E. Merck). Compounds were visualized by UV irradiation and/or spraying with sulfuric acid (H 2 SO 4 5% in ethanol) stain followed by charring at average 150 • C. Flash column chromatography was performed on Silica Gel 60 M (0.040-0.063 mm, E. Merck). The infrared spectra were measured with Perkin-Elmer Spectrometer. The 1 H and 13 C NMR spectra were recorded on BrukerAvance DPX 250 or BrukerAvance 400 Spectrometers. Chemical shifts are given in ppm and are referenced to the deuterated solvent signal or to TMS as internal standard and multiplicities are reported as s (singlet), d (doublet), t (triplet), q (quartet) and m (multiplet). Carbon multiplicities were assigned by distortion less enhancement by polarization transfer (DEPT) experiments. 1 H and 13 C signals were attributed on the basis of H-H and H-C correlations. High Resolution Mass spectra were performed on a Bruker Q-TOF MaXis spectrometer by the "Fédération de Recherche" ICOA/CBM (FR2708) platform. LC-MS data were acquired on a Thermo-Fisher UHPLC-MSQ system equipped with an electron spray ionization source (ESI). The temperature of the source was maintained at 350 • C. Initially, the cone voltage was set at 35 V and after 5 min was increased to 75V. In full scan mode, data were acquired between 100 and 1000 m/z in the positive mode with a 1.00 S scan time. In addition, a UV detection was performed with a Diode array detector at three wavelengths 273, 254 and 290 nm, respectively. A water/methanol (70%/30%) solution mixture with 0.1% formic acid was used as mobile phase. The composition of the mobile phase was increased to 100% methanol with 0.1% formic acid with a 7% ramp. The flow rate was set at 0.300 mL.min −1 . Samples diluted in the mobile phase were injected (3 µL) on a C18 column (X-terra, Waters), 2.1 mm internal diameter, and 100 mm length placed into an oven at 40 • C. Electronic extraction of ions was performed and the subsequent areas under the corresponding chromatographic peaks determined.

General Synthetic Procedure 1 for Strecker-Ugi Cyclization
A mixture of amino pyridine derivatives (200 mg) and ethyl glyoxalate (50% solution in toluene) (1 eq.) was stirred at 25 • C for 2 min. THF (4.5 mL) and 1,4-diazabicyclo [2.2.2] octane (1 eq.) were subsequently added. The reaction mixture was cooled to 0-5 • C and cyanotrimethylsilane (1 eq.) was added. The mixture was heated under microwave irradiation at 120 • C for 15 min. The solvent was evaporated under reduced pressure. The crude product was dissolved with EtOAc, washed with K 2 CO 3 , dried over MgSO 4 , filtrated and concentrated under reduced pressure.

General Synthetic Procedure 3 for the N 1 -Alkylation
To a solution of fluorobenzyl compound (150 mg) in anhydrous DMF (4 mL), were added subsequently potassium carbonate (1.5 eq.) and bromide derivative (1.5 eq.) under inert atmosphere. The reaction mixture was heated at 120 • C under microwave irradiation for 20 min. The mixture was dissolved with EtOAc, washed twice with saturated NH 4 Cl, dried over MgSO 4 , filtrated and concentrated under vacuum. Pure compound was