A Sub-Micromolar MraYAA Inhibitor with an Aminoribosyl Uridine Structure and a (S,S)-Tartaric Diamide: Synthesis, Biological Evaluation and Molecular Modeling

New inhibitors of the bacterial tranferase MraY are described. Their structure is based on an aminoribosyl uridine scaffold, which is known to be important for the biological activity of natural MraY inhibitors. A decyl alkyl chain was introduced onto this scaffold through various linkers. The synthesized compounds were tested against the MraYAA transferase activity, and the most active compound with an original (S,S)-tartaric diamide linker inhibits MraY activity with an IC50 equal to 0.37 µM. Their antibacterial activity was also evaluated on a panel of Gram-positive and Gram-negative strains; however, the compounds showed no antibacterial activity. Docking and molecular dynamics studies revealed that this new linker established two stabilizing key interactions with N190 and H325, as observed for the highly potent inhibitors carbacaprazamycin, muraymycin D2 and tunicamycin.


Introduction
The continuous emergence of bacterial resistance to commonly used antibiotics is a major risk to global health [1][2][3]. It also has a significant societal impact, particularly in economic terms, due to the increased length of hospital stays because of nosocomial infections that considerably affects hospital costs [4][5][6][7]. Multidrug-resistant (MDR) bacterial strains are capable of developing a set of resistance mechanisms to circumvent the toxicity of antibacterial compounds, and the fight against antimicrobial resistance is a challenge of prime importance for the scientific community. Biological targets displaying a different mode of action than the one targeted by the approved antibiotics are particularly sought after in order to delay the emergence of resistance. Although peptidoglycan biosynthesis is a well-known target, it still involves enzymes specific to the bacterial world that are not targeted by existing drugs and therefore represent promising targets in the search of new antibiotics. Thus, the membrane [8] and intracytoplasmic [9] steps of peptidoglycan biosynthesis are underexploited and deserve special attention. Peptidoglycan biosynthesis is a complex process that takes place successively in the cytoplasm, the membrane and the periplasm (Figure 1). The MraY transferase catalyzes the first membrane-associated step of this biosynthesis, namely the transfer of the phospho-MurNAc-pentapeptide moiety of this biosynthesis, namely the transfer of the phospho-MurNAc-pentapeptide moiety from the cytoplasmic precursor (UDP-MurNAc-pentapeptide) to the membrane acceptor, undecaprenyl-phosphate (C55-P), yielding lipid I and releasing uridine monophosphate [10]. Further transfer of a GlcNAc moiety to lipid I catalyzed by the MurG transferase affords lipid II, which is flipped across the membrane to the periplasm. Then transglycosylation and transpeptidation steps, catalyzed by transglycosylases and transpeptidases, such as penicillin-binding proteins (PBPs), occur to give peptidoglycan, a giant macromolecule that protects the cell against internal osmotic pressure and maintains its shape. Since the enzymes involved in this biosynthesis are ubiquitous and essential among bacteria, they are pertinent targets for developing antibiotics such as fosfomycin-targeting MurA or penicillins and cephalosporins that covalently bind PBPs ( Figure 1). However, due to the mechanisms of resistance developed by the bacteria towards these compounds [11,12], it is interesting to focus on other targets, such as the MraY transferase. Several families of potent natural MraY inhibitors are known, such as liposidomycins [13][14][15], muraymycins [16] and caprazamycins [17,18], all of them sharing an aminoribosyluridine moiety as a common structural feature. Currently, there are no MraY-directed antibiotics in clinical use, which is a key advantage to delay the occurrence of resistance. The synthesis of analogs of these peptidonucleosidic compounds has been the matter of intensive synthetic work [19][20][21][22], especially regarding the goal of decreasing their complexity while maintaining potent inhibition of MraY enzymatic activity. This is particularly challenging, considering the hydrophilicity of these molecules, and interesting progress has been achieved towards this objective [23][24][25][26][27]. Based on the aminoribosyluridine skeleton known to be important for MraY inhibition, we developed the synthesis of several families of simplified analogs (Figure 2A), with the chemical diversity being introduced on this scaffold through a triazole [28] or a methylene triazole [29] linker. However, the docking of these triazole-containing inhibitors in either the 5CKR [30] or 6OYH [31] structural models revealed no significant interactions of the triazole with aminoacids of the MraY active site. We then focused on the synthesis of urea-containing compounds [32] with various substituents (Figure 2B), and the most active inhibitor (IC50 equal to 1.9 µM) was the one bearing a decyl chain. Molecular dynamics experiments showed that the urea linker could interact with H324 and H325 ( Figure 2B), but that this interaction was not stable [32]. This result prompted us to study the effect of other linkers ( Figure 2C) that are Several families of potent natural MraY inhibitors are known, such as liposidomycins [13][14][15], muraymycins [16] and caprazamycins [17,18], all of them sharing an aminoribosyluridine moiety as a common structural feature. Currently, there are no MraY-directed antibiotics in clinical use, which is a key advantage to delay the occurrence of resistance. The synthesis of analogs of these peptidonucleosidic compounds has been the matter of intensive synthetic work [19][20][21][22], especially regarding the goal of decreasing their complexity while maintaining potent inhibition of MraY enzymatic activity. This is particularly challenging, considering the hydrophilicity of these molecules, and interesting progress has been achieved towards this objective [23][24][25][26][27]. Based on the aminoribosyluridine skeleton known to be important for MraY inhibition, we developed the synthesis of several families of simplified analogs (Figure 2A), with the chemical diversity being introduced on this scaffold through a triazole [28] or a methylene triazole [29] linker. However, the docking of these triazole-containing inhibitors in either the 5CKR [30] or 6OYH [31] structural models revealed no significant interactions of the triazole with aminoacids of the MraY active site. We then focused on the synthesis of urea-containing compounds [32] with various substituents ( Figure 2B), and the most active inhibitor (IC 50 equal to 1.9 µM) was the one bearing a decyl chain. Molecular dynamics experiments showed that the urea linker could interact with H324 and H325 ( Figure 2B), but that this interaction was not stable [32]. This result prompted us to study the effect of other linkers ( Figure 2C) that are able to interact by hydrogen bonds with key residues of the MraY active site [31], notably H324 and H325, in order to improve their potency compared to that of the urea-containing compounds. The targeted molecules display various linkers such as squaramide, two isomers of tartaric diamide, amide and sulfonamide ( Figure 2C). Indeed, the squaramide ring in compound 2 could act as both a donor and an acceptor of hydrogen bonds [33]. Moreover, the diamide moiety in compounds (S,S)-3a and (R,R)-3b could promote the formation of Supplementary hydrogen bonds due to the presence of two hydroxyl groups [34][35][36]. The amide 4 should permit us to compare the activity of a single amide to that of the diamides (S,S)-3a and (R,R)-3b, whereas the tetrahedral geometry of the sulfur atom and the longer size of the S=O bond in compound 5 could generate hydrogen bonds different from those of the other linkers [37].
Molecules 2022, 27, x FOR PEER REVIEW 3 of 22 able to interact by hydrogen bonds with key residues of the MraY active site [31], notably H324 and H325, in order to improve their potency compared to that of the urea-containing compounds. The targeted molecules display various linkers such as squaramide, two isomers of tartaric diamide, amide and sulfonamide ( Figure 2C). Indeed, the squaramide ring in compound 2 could act as both a donor and an acceptor of hydrogen bonds [33]. Moreover, the diamide moiety in compounds (S,S)-3a and (R,R)-3b could promote the formation of Supplementary hydrogen bonds due to the presence of two hydroxyl groups [34][35][36]. The amide 4 should permit us to compare the activity of a single amide to that of the diamides (S,S)-3a and (R,R)-3b, whereas the tetrahedral geometry of the sulfur atom and the longer size of the S=O bond in compound 5 could generate hydrogen bonds different from those of the other linkers [37]. An identical C-10 alkyl chain was selected based on the best MraY inhibitory activity of the corresponding urea-containing inhibitor to allow for the accurate comparison of the inhibitory potency among the resulting inhibitors. The docking of these compounds in the structural models 5CKR and 6OYH was performed, and key interactions of the envisaged linkers with D196, N255, N190, L191, D265 and H325 of the MraY active site were highlighted. The stability of the best compounds was confirmed by molecular dynamics experiments. We were delighted to achieve a gain in activity of a factor of 5 with the (S,S) diamide inhibitor as compared to the urea reference compound, and we wish to present our results in detail.

Chemistry
The targeted compounds were synthesized from the 1′',5′'-dideoxy-2′',3′'-O-isopentylidene-5′'-azido-1′'-[2′,3′-O-isopropylidene-5′(S)-aminomethyl-uridinyl]-β-D-ribofuranose 1 (Figure 3), as we previously described [32]. This common intermediate resulted from the glycosylation of the 5′(S)-phthalimidoalcohol B [32] with the 5-azidoribosyl fluoride A [38] readily obtained in a few steps from uridine and D-ribose, respectively, followed by hydrazinolysis of the phthalimide (Figure 3). An identical C-10 alkyl chain was selected based on the best MraY inhibitory activity of the corresponding urea-containing inhibitor to allow for the accurate comparison of the inhibitory potency among the resulting inhibitors. The docking of these compounds in the structural models 5CKR and 6OYH was performed, and key interactions of the envisaged linkers with D196, N255, N190, L191, D265 and H325 of the MraY active site were highlighted. The stability of the best compounds was confirmed by molecular dynamics experiments. We were delighted to achieve a gain in activity of a factor of 5 with the (S,S) diamide inhibitor as compared to the urea reference compound, and we wish to present our results in detail.

Chemistry
The targeted compounds were synthesized from the 1 ,5 -dideoxy-2 ,3 -O-isopentylidene-5 -azido-1 -[2 ,3 -O-isopropylidene-5 (S)-aminomethyl-uridinyl]-β-D-ribofuranose 1 (Figure 3), as we previously described [32]. This common intermediate resulted from the glycosylation of the 5 (S)-phthalimidoalcohol B [32] with the 5-azidoribosyl fluoride A [38] readily obtained in a few steps from uridine and D-ribose, respectively, followed by hydrazinolysis of the phthalimide (Figure 3). The retrosynthesis towards the targeted compounds is straightforward, the chemical diversity being introduced in the last key step of the synthetic pathway. We first turned to the preparation of the required building blocks. The synthesis of squarate 6 was performed from decylamine and diethyl squarate in ethanol in the presence of triethylamine (Scheme 1). After 2 hours of reaction, the unsymmetrical squarate 6 was isolated in a good 95% yield after purification. For the synthesis of the diamide compounds, both (2S,3S) and (2R,3R) isomers of dimethyl tartrate were chosen. Indeed, due to the C2 axis of symmetry within these compounds, each of the acid functions is equivalent, and the first reaction on one of these acids, which will desymmetrize tartaric acid, can only lead to the formation of a single diastereomer. Both (2S,3S) and (2R,3R) isomers were submitted to a selective monosaponification in the presence of potassium hydroxide in methanol to afford the corresponding acids (Scheme 1). After activation of these intermediates with dicyclohexyl carbodiimide (DCC) and N-hydroxysuccinimide (NHS) in DMF for 3 h, peptidic coupling with decylamine for an additional 24 h afforded the crude (S,S)-7a and (R,R)-7b. Flash chromatographic purification of these compounds led to the pure (S,S)-7a compound in 66% yield over two steps and to the pure (R,R)-7b compound in 60% yield over two steps. With these building blocks in hand, we next turned to the synthesis of compounds 8-11 (Scheme 2). The synthesis of the squaramide derivative 8 was carried out by stirring the squarate 6 and the amine 1 in the presence of triethylamine in ethanol for 30 minutes at r.t., which led to the squaramide 8 in 42% yield. The diamide compounds (S,S)-9a and (R,R)-9b resulted from the saponification of the tartrate derivatives (S,S)-7a and (R,R)-7b in the presence of sodium hydroxide in methanol, giving the corresponding acids, followed by their peptidic coupling with the amine 1 that was activated in the presence of DCC and N-hydroxysuccinimide in DMF. Compounds (S,S)-9a or (R,R)-9b were isolated in 47 and 48% yield, respectively, from 1. The condensation of amine 1 with undecanoyl The retrosynthesis towards the targeted compounds is straightforward, the chemical diversity being introduced in the last key step of the synthetic pathway. We first turned to the preparation of the required building blocks. The synthesis of squarate 6 was performed from decylamine and diethyl squarate in ethanol in the presence of triethylamine (Scheme 1). After 2 hours of reaction, the unsymmetrical squarate 6 was isolated in a good 95% yield after purification. For the synthesis of the diamide compounds, both (2S,3S) and (2R,3R) isomers of dimethyl tartrate were chosen. Indeed, due to the C 2 axis of symmetry within these compounds, each of the acid functions is equivalent, and the first reaction on one of these acids, which will desymmetrize tartaric acid, can only lead to the formation of a single diastereomer. Both (2S,3S) and (2R,3R) isomers were submitted to a selective monosaponification in the presence of potassium hydroxide in methanol to afford the corresponding acids (Scheme 1). After activation of these intermediates with dicyclohexyl carbodiimide (DCC) and N-hydroxysuccinimide (NHS) in DMF for 3 h, peptidic coupling with decylamine for an additional 24 h afforded the crude (S,S)-7a and (R,R)-7b. Flash chromatographic purification of these compounds led to the pure (S,S)-7a compound in 66% yield over two steps and to the pure (R,R)-7b compound in 60% yield over two steps.  The retrosynthesis towards the targeted compounds is straightforward, the chemical diversity being introduced in the last key step of the synthetic pathway. We first turned to the preparation of the required building blocks. The synthesis of squarate 6 was performed from decylamine and diethyl squarate in ethanol in the presence of triethylamine (Scheme 1). After 2 hours of reaction, the unsymmetrical squarate 6 was isolated in a good 95% yield after purification. For the synthesis of the diamide compounds, both (2S,3S) and (2R,3R) isomers of dimethyl tartrate were chosen. Indeed, due to the C2 axis of symmetry within these compounds, each of the acid functions is equivalent, and the first reaction on one of these acids, which will desymmetrize tartaric acid, can only lead to the formation of a single diastereomer. Both (2S,3S) and (2R,3R) isomers were submitted to a selective monosaponification in the presence of potassium hydroxide in methanol to afford the corresponding acids (Scheme 1). After activation of these intermediates with dicyclohexyl carbodiimide (DCC) and N-hydroxysuccinimide (NHS) in DMF for 3 h, peptidic coupling with decylamine for an additional 24 h afforded the crude (S,S)-7a and (R,R)-7b. Flash chromatographic purification of these compounds led to the pure (S,S)-7a compound in 66% yield over two steps and to the pure (R,R)-7b compound in 60% yield over two steps. With these building blocks in hand, we next turned to the synthesis of compounds 8-11 (Scheme 2). The synthesis of the squaramide derivative 8 was carried out by stirring the squarate 6 and the amine 1 in the presence of triethylamine in ethanol for 30 minutes at r.t., which led to the squaramide 8 in 42% yield. The diamide compounds (S,S)-9a and (R,R)-9b resulted from the saponification of the tartrate derivatives (S,S)-7a and (R,R)-7b in the presence of sodium hydroxide in methanol, giving the corresponding acids, followed by their peptidic coupling with the amine 1 that was activated in the presence of DCC and N-hydroxysuccinimide in DMF. Compounds (S,S)-9a or (R,R)-9b were isolated in 47 and 48% yield, respectively, from 1. The condensation of amine 1 with undecanoyl With these building blocks in hand, we next turned to the synthesis of compounds 8-11 (Scheme 2). The synthesis of the squaramide derivative 8 was carried out by stirring the squarate 6 and the amine 1 in the presence of triethylamine in ethanol for 30 minutes at r.t., which led to the squaramide 8 in 42% yield. The diamide compounds (S,S)-9a and (R,R)-9b resulted from the saponification of the tartrate derivatives (S,S)-7a and (R,R)-7b in the presence of sodium hydroxide in methanol, giving the corresponding acids, followed by their peptidic coupling with the amine 1 that was activated in the presence of DCC and N-hydroxysuccinimide in DMF. Compounds (S,S)-9a or (R,R)-9b were isolated in 47 and 48% yield, respectively, from 1. The condensation of amine 1 with undecanoyl chloride or decanesulfonyl chloride in the presence of triethylamine and 4-dimethylaminopyridine in dichloromethane led to the amide 10 in a satisfactory 72% yield or to the sulfonamide 11 in a good 87% yield. Finally, the reduction of compounds 8-11 was performed in an 85/15 THF/H2O mixture under Staudinger conditions, using polymer-supported triphenylphosphine to optimize the removal of supported triphenylphosphine oxide by filtration through a celite pad. Then the acidic hydrolysis of the alcohol protective groups was carried out in a cold 4/1 mixture of trifluoroacetic acid/water. The targeted compounds 2, (S,S)-3a, (R,R)-3b, 4 and 5 were isolated as their free amine in 42 to 73% yield after flash chromatographic purification on silica gel (Scheme 3). Finally, the reduction of compounds 8-11 was performed in an 85/15 THF/H 2 O mixture under Staudinger conditions, using polymer-supported triphenylphosphine to optimize the removal of supported triphenylphosphine oxide by filtration through a celite pad. Then the acidic hydrolysis of the alcohol protective groups was carried out in a cold 4/1 mixture of trifluoroacetic acid/water. The targeted compounds 2, (S,S)-3a, (R,R)-3b, 4 and 5 were isolated as their free amine in 42 to 73% yield after flash chromatographic purification on silica gel (Scheme 3).

Biological Studies
The inhibitory activity of the synthesized compounds 2, (S,S)-3a, (R,R)-3b, 4 and 5 was evaluated on MraY transferase purified from Aquifex aeolicus (MraYAA), which was prepared as previously described by Chung et al. [39]. Their activity was compared to the inhibitory activity of the unprotected amino precursor 12, urea-containing inhibitor 13

Biological Studies
The inhibitory activity of the synthesized compounds 2, (S,S)-3a, (R,R)-3b, 4 and 5 was evaluated on MraY transferase purified from Aquifex aeolicus (MraY AA ), which was prepared as previously described by Chung et al. [39]. Their activity was compared to the inhibitory activity of the unprotected amino precursor 12, urea-containing inhibitor 13 and Cand N-triazole-containing inhibitors 14 and 15 ( Figure 4) that we previously synthesized [29,32] (Table 1). Commercially available tunicamycin from Streptomyces sp. was used as a positive control in the test.

Biological Studies
The inhibitory activity of the synthesized compounds 2, (S,S)-3a, (R,R)-3b, 4 and 5 was evaluated on MraY transferase purified from Aquifex aeolicus (MraYAA), which was prepared as previously described by Chung et al. [39]. Their activity was compared to the inhibitory activity of the unprotected amino precursor 12, urea-containing inhibitor 13 and C-and N-triazole-containing inhibitors 14 and 15 ( Figure 4) that we previously synthesized [29,32] (Table 1). Commercially available tunicamycin from Streptomyces sp. was used as a positive control in the test.  a Experiments were performed in triplicate, and each experiment was repeated at least twice, except for tunicamycin as a control that was tested twice. The background color corresponds to reference compounds already known while the lines with no background correspond to the new compounds described in the manuscript.  a Experiments were performed in triplicate, and each experiment was repeated at least twice, except for tunicamycin as a control that was tested twice. The background color corresponds to reference compounds already known while the lines with no background correspond to the new compounds described in the manuscript.
As shown in Table 1, the five tested compounds are relevant inhibitors of the enzymatic activity catalyzed by the transferase MraY AA , with IC 50 ranging from 0.37 to 17.97 µM. All the synthesized compounds revealed better activity than that of the reference amine 12, which displayed an IC 50 value around 50 µM, confirming that substitution of the amine at C6 is beneficial to the inhibitory effect. Among the tested compounds, the (S,S)-3a diamide displays the strongest inhibition with an IC 50 equal to 0.37 µM, representing a gain in activity of a factor of 5 as compared to the reference compound 13 with an urea linker. It is noteworthy that its diastereoisomer (R,R)-3b is less active with an IC 50 equal to 1.38 µM. Interestingly, the simple amide 4 displays an IC 50 equal to 6.49 µM, thus demonstrating that the presence of the hydroxyl groups in (S,S)-3a and (R,R)-3b, which are capable of creating additional hydrogen bonds with amino acids of the MraY catalytic site, plays a prominent role in enhancing the inhibitory potency of the corresponding inhibitors. The sulfonamide 5 retains good inhibitory activity in the same range as C-triazole 14 but is a bit less active than the urea 13 or the C-triazole 15, while the squaramide 2 is probably too rigid for an optimal positioning in the MraY active site, resulting in a significant loss of activity (IC 50 equal to 17.97 µM).
Even if our goal was to improve the MraY inhibitory activity of the reference urea 13, and that was successfully achieved, the antibacterial activity of MraY inhibitors 2-5 was also evaluated against several bacterial strains. Gram-negative (E. coli ATCC 8730, C. freundii ATCC8090 and P. aeruginosa ATCC 27853) and Gram-positive pathogenic bacterial strains (S. aureus ATCC 25923 and E. faecium ATCC 19434), including a methicillin resistant strain (S. aureus MRSA ATCC 43300), were selected as representative of pathogen bacterial diversity. Piperacillin and vancomycin were used as positive controls in the tests. As we observed for the urea 13 [32], and as one could expect, none of the tested MraY inhibitors showed antibacterial activity (MICs values were higher than or equal to 128 µg/mL (see Supplementary Materials Table S1) against the six selected bacterial species. Indeed, we previously reported that the urea 13 with a C10 alkyl chain does not display antibacterial activity, since it is probably not able to cross the cytoplasmic membrane to reach its target. Our previous data on differently substituted urea [32] suggested that a linear chain of at least 12 carbon atoms, or a branched substituent, is required to show antibacterial activity. Therefore, further optimization of the lipophilic side chain is still required to increase the antibacterial activity by modulating the lipophilicity of the molecule.

Docking Studies
To predict the binding mode of this series of compounds, we performed docking studies starting from the X-ray crystal structures of MraY AA in complex with muraymycin D2 (MurD2, PDB: 5CKR) [30] and carbacaprazamycin (PDB: 6OYH) [31], using the CDOCKER docking program [40] implemented in the Biovia Discovery Studio docking package 2016. Preferred poses were selected based on the docking score (-CDOCKER interaction energy) and favorable ligand/protein interactions.

Docking Score
CDOCKER interaction energy included the energy of non-bonded interaction between the protein and the ligand [40]. Higher negative energy values for the docking score indicated stronger binding between MraY AA and the compounds. Tartaric diamides (S,S)-3a, (R,R)-3b performed well in terms of docking scores in comparison to known inhibitors (MurD2, carbacaprazamycin) and urea 13, both in 5CKR and 6OYH. Lower scores were obtained for hits 2, 4 and 5 (Table 2).

Binding Mode of the Hits
Hits bind to hotspots (HSs) of MraY that were identified by Chung et al. [31] by following two modes (one in 5CKR and one in 6OYH), as previously observed for the urea 13 [32].

•
Binding mode in 5CKR: The uracil moiety of all compounds occupied the uridine pocket establishing H-bond interactions with K70, N255 and D196 residues ( Figure 5). A strong H-bond interaction (d1 < 2.1 Å) involving the key D196 residue [30,42] was retrieved for all compounds, as observed with the urea 13 (Supplementary Materials Table S2) and MurD2 in the crystallographic structure 5CKR (d~1.5 Å). All compounds, except for 4, formed additional H-bond interactions with both residues K70 and N255, as suggested by distance measurements d2 and d3 (Supplementary Materials Table S2).

•
Binding mode in 5CKR: The uracil moiety of all compounds occupied the uridine pocket establishing H-bond interactions with K70, N255 and D196 residues ( Figure 5). A strong H-bond interaction (d1 < 2.1 Å) involving the key D196 residue [30,42] was retrieved for all compounds, as observed with the urea 13 (Supplementary Materials Table S2) and MurD2 in the crystallographic structure 5CKR (d ~ 1.5 Å). All compounds, except for 4, formed additional Hbond interactions with both residues K70 and N255, as suggested by distance measurements d2 and d3 (Supplementary Materials Table S2).  Subtle variations in the binding pattern were observed for the 5-aminoribosyl and linker moieties of new compounds (Table 3). Amide 4 binds to three hotspots (HS1, HS3 and HS5), as previously seen for the urea 13 through the residues K121, N190, D193 and D265. In addition to binding HS1 and HS5, squaramide 2 formed H-bond interactions with H324 and H325 in HS2, while sulfonamide 5 targeted HS1, HS2 and HS3. Focusing on diamide compounds (S,S)-3a, (R,R)-3b, we see that the results analysis revealed that both diastereoisomers have an interesting binding profile combining four hotspots (HS1, HS2, HS5 and HS6, Table 3; see Figure 5). Our binding mode analysis suggests that the extensive network of interactions formed by diamide (S,S)-3a within the MraY binding site may stabilize this compound in a more active conformation. Interestingly, the moderately active compound 2 adopts a strained conformation with spatial reorientation of the aminoribose moiety to minimize steric clashes with H324 and H325, leading to the loss of stabilizing interactions. Taken together, these results suggest that the structural geometry of the linker might guide the positioning of the hit in the binding site. In addition, all compounds oriented their aliphatic tail in the HS2 area, including hydrophobic residues V302, I306 and A321, as seen for the urea 13 in 5CKR model. Binding site hotspots were colored as follows: HS1 (green), HS2 (purple), HS3 (pink), HS5 (orange) and HS6 (gray). A detailed view of the interactions between MraY AA and the hits is available in Supplementary Materials Figure S1.
• Binding mode in 6OYH: New hits adopted a second binding mode similar to that observed for our reference compound 13 and carbacaprazamycin in 6OYH model ( Figure 6). Compared to the 5CKR model, the major difference is the aliphatic tail shifting from HS2 binding site to the hydrophobic groove (HS4). The anchoring of the uracil moiety involved the same residues, namely K70, D196 and N255, as seen in 5CKR. The strong H-bonding interaction with the D196 residue (d1 ≤ 2Å; see Supplementary Materials Table S2) was preserved for all compounds, and contact with N255 was maintained (d2 ≤ 2.5Å; see Supplementary Materials Table S2). Moreover, a strong H-bond with K70 (d3 ≤ 1.8Å) was also predicted for diamides (S,S)-3a and (R,R)-3b. All ligands, except for 5, were also stabilized by hydrophobic interactions with residues F262 and G194 (Supplementary Materials data). Furthermore, an in-depth analysis of the binding profile for the aminoribosyl and linker moieties revealed that hits only have one common HS (HS1, Table 4). In addition to HS1, compound 2 bound HS2 and HS3, while compound 4 interacted with HS5 and HS6. More interestingly, interaction with H325 in HS2 was retrieved for both tartrates (S,S)-3a and (R,R)-3b, as previously observed in 5CKR (Table 4). These results suggest that residues D196, N255, K70 (uridine pocket), T75, N190, L191, D193, G264 (HS1), H325 (HS2) and D265 (HS5) may play a key role in ligand binding.

Molecular Dynamics
To provide insights on the stability of tartaric diamides (S,S)-3a and (R,R)-3b in the binding site of MraY AA , we conducted Molecular Dynamics (MD) simulations using NAMD with the CHARMM36m force field [43] implemented in Biovia DS 2021. The experiments were performed starting from the best docking solutions of ligands obtained from CDOCKER and using the procedure previously reported for compound 13 and carbacaprazamycin. Root Mean Standard Deviation (RMSD) analysis of MraY AA and MraY AA /ligand complexes measured over protein backbone atoms revealed that all systems are relatively stable along the 50 ns MD runs with no major structural variations ( Figure 7A). Diastereoisomer (S,S)-3a displayed a stable RMSD plot similar to that observed for carbacaprazamycin with marginal elevation between 25-30 ns indicating that the ligand binding mode was preserved during the timescale of the simulation ( Figure 7B). Diastereoisomer (R,R)-3b seems to be less stable as suggested by a significant elevation of 0.9 Å between 25-50 ns in the deviation profile ( Figure 7B). an in-depth analysis of the binding profile for the aminoribosyl and linker moieties revealed that hits only have one common HS (HS1, Table 4). In addition to HS1, compound 2 bound HS2 and HS3, while compound 4 interacted with HS5 and HS6. More interestingly, interaction with H325 in HS2 was retrieved for both tartrates (S,S)-3a and (R,R)-3b, as previously observed in 5CKR (Table 4). These results suggest that residues D196, N255, K70 (uridine pocket), T75, N190, L191, D193, G264 (HS1), H325 (HS2) and D265 (HS5) may play a key role in ligand binding. , and π-π and amide-π interactions (magenta). The hydrogen atoms of apolar groups were omitted for clarity.  , and π-π and amide-π interactions (magenta). The hydrogen atoms of apolar groups were omitted for clarity.   Figure S3B) suggesting better uracil stability of this compound than the one observed for (R,R)-3b. In addition, the residues F262 and G194 contributed to the binding through hydrophobic π-π interactions (data not shown).
periments were performed starting from the best docking solutions of ligands obtained from CDOCKER and using the procedure previously reported for compound 13 and carbacaprazamycin. Root Mean Standard Deviation (RMSD) analysis of MraYAA and MraYAA /ligand complexes measured over protein backbone atoms revealed that all systems are relatively stable along the 50 ns MD runs with no major structural variations ( Figure 7A). Diastereoisomer (S,S)-3a displayed a stable RMSD plot similar to that observed for carbacaprazamycin with marginal elevation between 25-30 ns indicating that the ligand binding mode was preserved during the timescale of the simulation ( Figure 7B). Diastereoisomer (R,R)-3b seems to be less stable as suggested by a significant elevation of 0.9 Å between 25-50 ns in the deviation profile ( Figure 7B).  Figure S3B) suggesting better uracil stability of this compound than the one observed for (R,R)-3b. In addition, the residues F262 and G194 contributed to the binding through hydrophobic π-π interactions (data not shown). More variations in the distance profile were retrieved for the aminoribose and linker moieties of both tartaric diamides between 25 and 50 ns. Compound (S,S)-3a established an electrostatic interaction with D265 (d4 ≤ 3 Å for 26% of conformations; see Supplementary Materials Figure S3C). Persistent H-bond interactions were also retrieved with N190 through the amino group (d5~2.6 Å; see Figure 9B) and the H97 hydrogen atom of the linker (d6~1.9 Å; see Figure 9C). In addition, the O45 atom of the tartaric moiety formed a stable H-bond with H325 (d7~2.8 Å; see Figure 9D). Compared to (S,S)-3a, (R,R)-3b lost H-bond interactions with N190 after 25 ns (d6 > 4 Å; see Figure 9C) and H325 contributed weaker to ligand binding (d7~3.4 Å, Figure 9D) than was observed for the urea linker (d~3.0 Å; data not shown). Interestingly, the position of aliphatic chain for both tartaric diamides remained stable in the hydrophobic groove (HS4) (Figure 8) establishing the same hydrophobic interactions with V302 and I303 as seen for carbacaprazamycin (data not shown), while previous results have revealed that alkyl chain of urea 13 was flipped from HS4 to HS2 after 50 ns MD simulations. Altogether, these results confirmed that More variations in the distance profile were retrieved for the aminoribose and linker moieties of both tartaric diamides between 25 and 50 ns. Compound (S,S)-3a established an electrostatic interaction with D265 (d4 ≤ 3 Å for 26% of conformations; see Supplementary Materials Figure S3C). Persistent H-bond interactions were also retrieved with N190 through the amino group (d5~2.6 Å; see Figure 9B) and the H97 hydrogen atom of the linker (d6~1.9 Å; see Figure 9C). In addition, the O45 atom of the tartaric moiety formed a stable H-bond with H325 (d7~2.8 Å; see Figure 9D). Compared to (S,S)-3a, (R,R)-3b lost H-bond interactions with N190 after 25 ns (d6 > 4 Å; see Figure 9C) and H325 contributed weaker to ligand binding (d7~3.4 Å, Figure 9D) than was observed for the urea linker (d~3.0 Å; data not shown). Interestingly, the position of aliphatic chain for both tartaric diamides remained stable in the hydrophobic groove (HS4) (Figure 8) establishing the same hydrophobic interactions with V302 and I303 as seen for carbacaprazamycin (data not shown), while previous results have revealed that alkyl chain of urea 13 was flipped from HS4 to HS2 after 50 ns MD simulations. Altogether, these results confirmed that (S,S)-3a adopts a more stable conformation than the (R,R)-3b diastereoisomer within the MraY binding site.

Chemical Synthesis
When needed, reactions were carried out under an argon atmosphere. They were monitored by thin-layer chromatography with precoated silica on aluminum foil. Flash chromatography was performed with silica gel 60 (40-63 µm); the solvent systems are given in v/v. Spectroscopic 1 H and 13 C NMR, MS and/or analytical data were obtained by using chromatographically homogeneous samples. 1 H NMR (500 MHz) and 13 C NMR (125 MHz) spectra were recorded in CDCl3 unless otherwise indicated. Chemical shifts (δ) are reported in ppm, and coupling constants are given in Hz. For each compound, detailed peak assignments were made according to COSY, HSQC and HMBC experiments. The numbering of molecules is indicated in the Supplementary Materials. Optical rotations were measured with a sodium (589 nm) lamp at 20 °C. IR spectra were recorded on an FTIR spectrophotometer, and the wavelengths are reported in cm −1 . High-resolution mass spectra (HRMS) were recorded with a TOF mass analyzer under electrospray ionization (ESI) in positive ionization mode detection, atmospheric pressure chemical ionization or atmospheric pressure photoionization (APPI).

Chemical Synthesis
When needed, reactions were carried out under an argon atmosphere. They were monitored by thin-layer chromatography with precoated silica on aluminum foil. Flash chromatography was performed with silica gel 60 (40-63 µm); the solvent systems are given in v/v. Spectroscopic 1 H and 13 C NMR, MS and/or analytical data were obtained by using chromatographically homogeneous samples. 1 H NMR (500 MHz) and 13 C NMR (125 MHz) spectra were recorded in CDCl 3 unless otherwise indicated. Chemical shifts (δ) are reported in ppm, and coupling constants are given in Hz. For each compound, detailed peak assignments were made according to COSY, HSQC and HMBC experiments. The numbering of molecules is indicated in the Supplementary Materials. Optical rotations were measured with a sodium (589 nm) lamp at 20 • C. IR spectra were recorded on an FTIR spectrophotometer, and the wavelengths are reported in cm −1 . High-resolution mass spectra (HRMS) were recorded with a TOF mass analyzer under electrospray ionization (ESI) in positive ionization mode detection, atmospheric pressure chemical ionization or atmospheric pressure photoionization (APPI). KOH (59 mg, 1.04 mmol, 1 equiv.) was dissolved in MeOH (1.5 mL) and added dropwise to a solution of S,S-protected tartaric acid (228 mg, 1.04 mmol, 1 equiv.) in MeOH (3 mL). The reaction mixture was stirred at 25 • C for 16 h, and then solvents were removed in vacuo. The resulting crude oil was then dissolved in water (15 mL) and washed 3 times with DCM (3 × 15 mL). The aqueous phase was then acidified with 1M HCl until pH 2-3. After evaporation of the water, the compound was dissolved in DCM, and the remaining salts were filtered out. Solvents were removed in vacuo, and the resulting acid was engaged in the next step without further purification. The acid (167 mg, 818 µmol, 1 equiv.) was dissolved in dry DMF (7 mL KOH (57 mg, 1.02 mmol, 1 equiv.) was dissolved in MeOH (1.5 mL) and added dropwise to a solution of R,R-protected tartaric acid (224 mg, 1.02 mmol, 1 equiv.) in MeOH (3 mL). The reaction mixture was stirred at 25 • C for 16 h, and then solvents were removed in vacuo. The resulting crude oil was then dissolved in water (15 mL) and washed 3 times with DCM (3 × 15 mL). The aqueous phase was then acidified with 1M HCl until pH 2-3. After evaporation of the water, the compound was dissolved in DCM, and the remaining salts were filtered out. Solvent was removed in vacuo, and the resulting acid was engaged in the next step without further purification. The acid (185 mg, 906 µmol, 1 equiv.) was dissolved in dry DMF (7 mL) under argon. DCC (243 mg, 1.18 mmol, 1.3 equiv.) and N-hydroxysuccinimide (109 mg, 951 µmol, 1.05 equiv.) were successively added, and the reaction mixture was stirred at r.t. for 3 h. Decylamine (190 µL, 951 mmol, 1.05 equiv.) was then added at r.t. After 24 h, solvents were removed in vacuo, and the crude mixture was dissolved in EtOAc and filtered through a celite pad. After removal of the solvent, flash NaOH (7.22 mg, 180 µmol, 1 equiv.) was dissolved in MeOH (0.5 mL) and added dropwise to a solution of (S,S)-7a (62 mg, 180 µmol, 1 equiv.) in MeOH (1 mL). The reaction mixture was stirred at 25 • C for 24 h, and then solvents were removed in vacuo. The resulting crude oil was then dissolved in water (10 mL) and DCM (10 mL). The medium was then acidified with 1M HCl until pH 2-3. Aqueous phase was then extracted three times with DCM. Combined organic phases were washed with brine, dried over Na 2 SO 4 , filtered and concentrated in vacuo. The resulting acid was engaged in the next step without further purification. The resulting acid (23 mg, 69 µmol, 1 equiv.) was dissolved in dry DMF (525 µL) under argon. DCC (18.45 mg, 89 µmol, 1.3 equiv.) and N-hydroxysuccinimide (8.31 mg, 72.2 µmol, 1.05 equiv.) were successively added to the reaction mixture and stirred at r.t. for 3 h. Amine 1 (50 mg, 69 µmol, 1 equiv.) was then added at r.t. After 48 h, solvents were removed in vacuo, and the crude mixture was dissolved in EtOAc and filtered through a celite pad. After removal of the solvent, flash chromatography of the residue (Cyclohexane/EtOAc = 6/4) afforded (S,S)-diamide 9a as a colorless oil (33.7 mg, 47% yield from 1): R f 0.55 (Cyclohexane/EtOAc = 5/5); [α]   (C 2 ), 85.0 (C 4 ), 84.9 (C 4 ), 81.7 (C 3 ), 80.2 (C 5 ), 75.2 (C 2 ), 71.6 (C 3 ), 53.8 (C 5 ), 41.9 (C 6 ), 36.9 (C 1× ), 32.0 (C 3× ), 29.7 (C 7 a ), 29.6 (C 7 b ), 29 To a solution of 1 (45 mg, 62 µmol, 1 equiv.) in dry DCM (4 mL) were successively added, at 0 • C, DMAP (9 mg, 69 µmol, 1.2 equiv.) and undecanoylsulfonyl chloride (18 µL, 74 µmol, 1.2 equiv.). The reaction mixture was stirred at r.t. for 16

General Procedure for Compounds Deprotection
To a solution of the protected compounds 8-11 (1 equiv.) in dry THF (2 mL) was added Polymer-supported triphenylphosphine (3 mmol/g; 6 equiv.) and pure water (0.7 mL). The reaction mixture was carefully stirred at r.t. for 48 h. The reaction was then filtered through a celite pad, carefully rinsed with THF and concentrated in vacuo to afford the crude amine. To the crude residue was added pure H 2 O, and the resulting suspension was stirred at 0 • C. Then, at 0 • C, TFA (300 equiv.) was added dropwise. The orange resulting solution was stirred at 0 • C for 10 min and then at r.t. for 18 h. After concentration in vacuo, flash chromatography of the residue (DCM/MeOH/NH4OH 14% 80/18/2) afforded the fully deprotected compounds (2)(3)(4)(5) in 42 to 73% yield over two steps 3.1.9. Squaramide 2 Compound 2 was prepared according to the general procedure for compounds deprotection from the squaramide 2 (16 mg, 17 µmol, 1 equiv.) and was obtained as a colorless oil

Docking and MD Simulations
Ligands were docked into the crystal structures of the MraY AA in complex with muraymycin D2 (PDB: 5CKR) and carbacaprazamycin (PDB: 6OYH) from CDOCKER program implemented in Biovia Discovery Studio 2016 [41] and according to the protocole described previously [32]. MD simulations were carried out on the 6OYH model using NAMD program [43] in Biovia Discovery studio 2021 as reported previously [32].

Conclusions
In this paper, we reported the synthesis of five new inhibitors of the bacterial MraY transferase displaying an aminoribosyl uridine scaffold substituted in 5 position by various linkers (squaramide, diamides, amide or sulfonamide) bearing an identical decyl chain. Their straightforward synthesis involved classical transformations, such as amidation, peptidic coupling or sulfonylation reactions from a conveniently protected azidoribosyl uridine with an aminomethyl group at C-5 . Their biological activity was evaluated in vitro on purified MraY AA and compared to that of related reference compounds with a urea, a Cor an N-triazole linker. All compounds revealed MraY inhibition, with IC 50 ranging from 0.37 to 17.97 µM. The binding mode of the inhibitors was studied by docking experiments. Molecular dynamics studies revealed the crucial role of the diamide linker to stabilize the more active compound (S,S)-3a through interactions with N190 and H325, which are two invariant residues in the active site of the MraY enzymes [46]. These interactions promote an optimal orientation of the key fragments (aminoribosyl uridine and alkyl chain) within the hotspots of MraY AA catalytic site and explain the improved activity of this compound compared to the reference urea 13. The in cellulo evaluation of the synthetized inhibitors on different Gram-positive and Gram-negative bacterial strains showed no significant activity that can be explained by the too-short length of their alkyl chain. Further optimization of the lipophilic side chain is still required to increase the antibacterial activity.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules27061769/s1. 1H and 13C spectra of all compounds. Table S1: Antibacterial activity of compounds 2-13 and reference compounds. Table S2: Distance measurement (Å) between hits and interacting residues of MraYAA binding site from docking experiments in model 5CKR and 6OYH. Figure S1: 2D diagram of ligands interactions in MraY AA from docking experiments (PDB: 5CKR). Figure S2: 2D diagram of ligands interactions in MraY AA from docking experiments (PDB: 6OYH). Figure S3: Time evolution of the distances between binding site residues of MraYAA and ligands atoms. Figure S4: 2D diagram of ligands interactions in MraY AA from 50 ns MD simulations (PDB: 6OYH).
the Macromolecular Modelling Platform and the NMR platform core facilities of the BioTechMed facilities INSERM US36|CNRS UMS2009|Université de Paris for docking and MD simulation and NMR experiments, respectively. H.W. thanks the Chinese Scholarship Council for the financial support of her PhD thesis.