Synthesis and Biological Evaluation of Lipophilic Nucleoside Analogues as Inhibitors of Aminoacyl-tRNA Synthetases

Emerging antibiotic resistance in pathogenic bacteria and reduction of compounds in the existing antibiotics discovery pipeline is the most critical concern for healthcare professionals. A potential solution aims to explore new or existing targets/compounds. Inhibition of bacterial aminoacyl-tRNA synthetase (aaRSs) could be one such target for the development of antibiotics. The aaRSs are a group of enzymes that catalyze the transfer of an amino acid to their cognate tRNA and therefore play a pivotal role in translation. Thus, selective inhibition of these enzymes could be detrimental to microbes. The 5′-O-(N-(L-aminoacyl)) sulfamoyladenosines (aaSAs) are potent inhibitors of the respective aaRSs, however due to their polarity and charged nature they cannot cross the bacterial membranes. In this work, we increased the lipophilicity of these existing aaSAs in an effort to promote their penetration through the bacterial membrane. Two strategies were followed, either attaching a (permanent) alkyl moiety at the adenine ring via alkylation of the N6-position or introducing a lipophilic biodegradable prodrug moiety at the alpha-terminal amine, totaling eight new aaSA analogues. All synthesized compounds were evaluated in vitro using either a purified Escherichia coli aaRS enzyme or in presence of total cellular extract obtained from E. coli. The prodrugs showed comparable inhibitory activity to the parent aaSA analogues, indicating metabolic activation in cellular extracts, but had little effect on bacteria. During evaluation of the N6-alkylated compounds against different microbes, the N6-octyl containing congener 6b showed minimum inhibitory concentration (MIC) of 12.5 µM against Sarcina lutea while the dodecyl analogue 6c displayed MIC of 6.25 µM against Candida albicans.


Introduction
The surge of antimicrobial-resistant strains of pathogenic bacteria is a significant concern to human health worldwide and has a profound impact on hospital born infections, chemotherapy, tuberculosis, and surgical procedures. This resistance crisis is estimated to cause 300 million cumulative premature deaths by 2050, with a loss of up to $100 trillion to the global economy [1]. A recent report from Public Health England stated that antibiotic resistance could make three million surgical procedures deadly in England alone [2]. The situation becomes more critical as there are very few antibiotics in the drug discovery pipeline [3]. To overcome antimicrobial resistance, various government organizations, The second approach deals with the attachment of a promoiety at the α-amine of aaSA ( Figure  1, structure b). Two carbamate promoieties which have been shown to be cleaved intracellularly following uptake, are the 4-nitrobenzyloxycarbonyl [25,26] and 4-acetoxybenzyloxycarbonyl carbamate [27,28]. The functional group at the para-position (either nitro or acetoxy here) on activation by nitroreductase or esterase respectively, should lead to cleavage of the prodrug and release of the active aaSA. The mentioned functionalities therefore are attached at the α-amine of leucylsulfamoyl adenosine (LSA) providing compounds 14 and 18 as the desired prodrugs. Following synthesis, their biological activity will be measured against different microorganisms while their intrinsic activity will be tested in a cellular extract using the aminoacylation assay. For this prodrug synthesis, we opted for leucylsulfamoyl adenosine, which already showed to be strongly active in vitro against leucine tRNA synthetase (LeuRS).
The first mentioned activation calls for the nitroreductase (NTR) [20], which are not found in human for activation of 14, and hence could provide a selectivity handle. In fact, the nitrobenzyl moiety has been used, as well as a selectivity handle, in cancer treatments in generating directed prodrugs with a metabolic trigger, but then requires the NTR to be introduced as well by transfection

Design Rationale
In this work, we tried both the above-mentioned approaches to increase the lipophilicity of aaSAs, which should lead to an increased influx of compounds and antibacterial activity. The first approach involved the coupling of different lipophilic groups to the N 6 -amino group of aaSA ( Figure 1). As observed from the crystal structures of Thermus thermophilus (Tt) isoleucine tRNA synthetase (IleRS) in complex with isoleucylsulfamoyl adenosine (ISA) or with mupirocin (ileRS inhibitor; unpublished work), there is sufficient space available around the adenine base to accommodate some chemical modifications [22,23]. In addition, the N 6 -amine is only making one hydrogen bond interaction with the backbone of the protein chain. Likewise, the aliphatic chain of mupirocin is interacting with the adenine binding domain [24]. strengthening our choice for longer alkyl moieties. Therefore, we alkylated ISA at the N 6 -position with alkyl moieties of various chain length to analyze for the best fit within the cavity and to achieve the maximum inhibitory effect. Thus, various chain lengths from octyl, dodecyl to octadecyl, and a phenyl group were introduced at the N 6 -amino moiety of the adenosine part (6b-e). Additionally, the 6-NH-CH 3 (6a) and 6-O-methyl moieties (6f) were introduced to evaluate the effect of oxygen vs nitrogen regarding enzymatic affinity, as well as the orientation of the methyl substituent in the active site of ileRS.
The second approach deals with the attachment of a promoiety at the α-amine of aaSA ( Figure 1, structure b). Two carbamate promoieties which have been shown to be cleaved intracellularly following uptake, are the 4-nitrobenzyloxycarbonyl [25,26] and 4-acetoxybenzyloxycarbonyl carbamate [27,28]. The functional group at the para-position (either nitro or acetoxy here) on activation by nitroreductase or esterase respectively, should lead to cleavage of the prodrug and release of the active aaSA. The mentioned functionalities therefore are attached at the α-amine of leucylsulfamoyl adenosine (LSA) providing compounds 14 and 18 as the desired prodrugs. Following synthesis, their biological activity will be measured against different microorganisms while their intrinsic activity will be tested in a cellular extract using the aminoacylation assay. For this prodrug synthesis, we opted for leucylsulfamoyl adenosine, which already showed to be strongly active in vitro against leucine tRNA synthetase (LeuRS).
The first mentioned activation calls for the nitroreductase (NTR) [20], which are not found in human for activation of 14, and hence could provide a selectivity handle. In fact, the nitrobenzyl moiety has been used, as well as a selectivity handle, in cancer treatments in generating directed prodrugs with a metabolic trigger, but then requires the NTR to be introduced as well by transfection methodologies or other means [29]. The second principle makes use of various esterases to release LSA Antibiotics 2019, 8, 180 4 of 23 from 18. Both enzyme types are present in several bacteria such as E. coli. Following exposure to the enzyme, the inactive prodrug should be metabolized to produce the active compound. In a typical prodrug strategy, the 4-nitrobenzyl group is attached to a leaving moiety such as a phosphoramide or a carboxylate. We here opted for the 4-nitrobenzyloxycarbonyl which on reduction by NTR produces the p-hydroxylamino-benzyl carbamate. The latter is prone to hydrolytic cleavage (Scheme 1) as the transformation converts an electron-withdrawing nitro group into an electron donating group [25,26]. Likewise, esterases upon hydrolysis of 18 will provide the hydrolytically cleavable p-hydroxybenzylcarbamate leading to the active compound [27] (Scheme 1). methodologies or other means [29]. The second principle makes use of various esterases to release LSA from 18. Both enzyme types are present in several bacteria such as E. coli. Following exposure to the enzyme, the inactive prodrug should be metabolized to produce the active compound. In a typical prodrug strategy, the 4-nitrobenzyl group is attached to a leaving moiety such as a phosphoramide or a carboxylate. We here opted for the 4-nitrobenzyloxycarbonyl which on reduction by NTR produces the p-hydroxylamino-benzyl carbamate. The latter is prone to hydrolytic cleavage (Scheme 1) as the transformation converts an electron-withdrawing nitro group into an electron donating group [25,26]. Likewise, esterases upon hydrolysis of 18 will provide the hydrolytically cleavable phydroxybenzylcarbamate leading to the active compound [27] (Scheme 1).

Chemistry
For the synthesis of N 6 -alkylated compounds, initially, we attempted to introduce various N 6alkyl groups at the pre-final protected stage by nucleophilic aromatic substitution on the C 6 -position. However, we observed no reaction or degradation of the protected 6-chloropurine ISA analogue. Later we attempted to modify the 6-chloropurine analogue of sulfamoyl adenosine; however, when trying to substitute 6-Cl with different alkylamines on 3f, cleavage of the sulfamoyl group was observed, as of the presence of the base N,N-diisopropylethylamine (DIPEA) and quite harsh microwave conditions (Supplementary reaction scheme S1). We finally opted for introduction of the various alkylamines on the protected 6-chloropurine riboside (2), followed by sulfamoylation and amino acid coupling to obtain the respective coupled products 5a-f which on deprotection gave compounds 6a-f. 3.1.1. Synthesis of N 6 -alkylated Analogues of 5'-O-(N-(L-isoleucyl)) Sulfamoyl Adenosine As shown in Scheme 2, previously reported strategies were followed to obtain the desired N 6alkylated derivatives. Their synthesis started via acetonide protection of commercially available 6chloropurine riboside (1) utilizing dimethoxypropane using para-toluene sulfonic acid as the catalyst and acetone as solvent. The acetonide-protected 2 on microwave-assisted nucleophilic aromatic substitution with a series of alkylamines and aniline afforded compounds 3a-e [30].

Chemistry
For the synthesis of N 6 -alkylated compounds, initially, we attempted to introduce various N 6 -alkyl groups at the pre-final protected stage by nucleophilic aromatic substitution on the C 6 -position. However, we observed no reaction or degradation of the protected 6-chloropurine ISA analogue. Later we attempted to modify the 6-chloropurine analogue of sulfamoyl adenosine; however, when trying to substitute 6-Cl with different alkylamines on 3f, cleavage of the sulfamoyl group was observed, as of the presence of the base N,N-diisopropylethylamine (DIPEA) and quite harsh microwave conditions (Supplementary reaction Scheme S1). We finally opted for introduction of the various alkylamines on the protected 6-chloropurine riboside (2), followed by sulfamoylation and amino acid coupling to obtain the respective coupled products 5a-f which on deprotection gave compounds 6a-f. 3.1.1. Synthesis of N 6 -alkylated Analogues of 5 -O-(N-(L-isoleucyl)) Sulfamoyl Adenosine As shown in Scheme 2, previously reported strategies were followed to obtain the desired N 6 -alkylated derivatives. Their synthesis started via acetonide protection of commercially available 6-chloropurine riboside (1) utilizing dimethoxypropane using para-toluene sulfonic acid as the catalyst and acetone as solvent. The acetonide-protected 2 on microwave-assisted nucleophilic aromatic substitution with a series of alkylamines and aniline afforded compounds 3a-e [30]. In the next reaction, these compounds 3(a-e) were sulfamoylated at the 5'-O-position using in situ prepared sulfamoyl chloride, which is synthesized by reacting formic acid with chlorosulfonylisocyanate [20], to obtain 4(a-e). These were then coupled with the Nhydroxysuccinimide active ester of Boc-Ile (Boc-Ile-OSu) in the presence of DBU, leading to the formation of the coupled products 5(a-e). Further deprotection of boc and acetonide functionality by the action of 60% TFA:water mixture lead to the formation of the desired compounds 6(a-e). For the synthesis of the O 6 -methylated analogue, 5'-O sulfamoylation of acetonide-protected 6-chloropurine riboside 2, was performed generating compound 3f. Next, this compound on nucleophilic aromatic substitution by sodium methoxide in methanol [31]. yielded compound 4f, which on further coupling with Boc-Ile-OSu and deprotection using 60% TFA: H2O mixture led to the desired 6f.

Synthesis of α-amine Promoiety Analogues
The synthesis of the intended prodrugs was carried out based on reported methodologies [30] and are described in Scheme 3 and Scheme 4. Synthesis started from the commercially available adenosine 7, which was first persilylated to generate compound 8 after which the 5'-position was liberated using a mixture of TFA:THF:H2O. The obtained 9 on reaction with the in situ generated sulfamoyl chloride led to the formation of 10. The sulfamoylated adenosine 10 was coupled with Boc-Leu-OSu in the presence of DBU to yield 11. The coupled product 11 was then treated with a 50:50 mixture of TFA and DCM to yield 12. The 4-nitrobenzyl chloroformate was reacted by using DIPEA as base and DMF as the solvent. The promoiety coupled compound 13 is finally treated with HF in TEA to cleave the silyl protection which lead to the desired prodrug 14. In the next reaction, these compounds 3(a-e) were sulfamoylated at the 5 -O-position using in situ prepared sulfamoyl chloride, which is synthesized by reacting formic acid with chlorosulfonylisocyanate [20], to obtain 4(a-e). These were then coupled with the N-hydroxysuccinimide active ester of Boc-Ile (Boc-Ile-OSu) in the presence of DBU, leading to the formation of the coupled products 5(a-e). Further deprotection of boc and acetonide functionality by the action of 60% TFA:water mixture lead to the formation of the desired compounds 6(a-e). For the synthesis of the O 6 -methylated analogue, 5 -O sulfamoylation of acetonide-protected 6-chloropurine riboside 2, was performed generating compound 3f. Next, this compound on nucleophilic aromatic substitution by sodium methoxide in methanol [31]. yielded compound 4f, which on further coupling with Boc-Ile-OSu and deprotection using 60% TFA: H 2 O mixture led to the desired 6f.

Synthesis of α-amine Promoiety Analogues
The synthesis of the intended prodrugs was carried out based on reported methodologies [30] and are described in Schemes 3 and 4. Synthesis started from the commercially available adenosine 7, which was first persilylated to generate compound 8 after which the 5 -position was liberated using a mixture of TFA:THF:H 2 O. The obtained 9 on reaction with the in situ generated sulfamoyl chloride led to the formation of 10. The sulfamoylated adenosine 10 was coupled with Boc-Leu-OSu in the presence of DBU to yield 11. The coupled product 11 was then treated with a 50:50 mixture of TFA and DCM to yield 12. The 4-nitrobenzyl chloroformate was reacted by using DIPEA as base and DMF as the For the synthesis of 4-acetoxybenzyloxycarbonyl protected α-amine, the 4-acetoxybenzyl chloroformate 16 was generated in situ by reacting triphosgene with 4-acetoxybenzyl alcohol 15 in the presence of DIPEA [32]. The 4-acetoxybenzyl chloroformate is used as such for further coupling with compound 12 which led to the silyl protected intermediate 17 that on further treatment with HF in TEA yielded the desired compound 18. For the synthesis of 4-acetoxybenzyloxycarbonyl protected α-amine, the 4-acetoxybenzyl chloroformate 16 was generated in situ by reacting triphosgene with 4-acetoxybenzyl alcohol 15 in the presence of DIPEA [32]. The 4-acetoxybenzyl chloroformate is used as such for further coupling with compound 12 which led to the silyl protected intermediate 17 that on further treatment with HF in TEA yielded the desired compound 18. For the synthesis of 4-acetoxybenzyloxycarbonyl protected α-amine, the 4-acetoxybenzyl chloroformate 16 was generated in situ by reacting triphosgene with 4-acetoxybenzyl alcohol 15 in the presence of DIPEA [32]. The 4-acetoxybenzyl chloroformate is used as such for further coupling with compound 12 which led to the silyl protected intermediate 17 that on further treatment with HF in TEA yielded the desired compound 18.

Biology
All synthesized compounds were evaluated in vitro against either purified enzyme (IleRS) in a buffer or in a cellular extract (S30), followed by determination of MIC values against different microbes.

Measurement of in vitro Inhibitory Activity with Purified E. coli IleRS
Following successful synthesis of compounds (6a-f), inhibitory activity was determined using radiolabel transfer assay. IleRS and total tRNA isolated from E. coli were used. In this assay the quantity of C 14 labelled isoleucine transferred to tRNA Ile was determined by precipitating the [C 14 ] Ile-tRNA Ile complex using a 10% TCA solution. Out of six tested derivatives, four showed IC 50 in the nanomolar range in the radiolabel transfer assay (6a,b and 6e,f; Figure 2A and Table 1). However, for all four a 2-or 3-fold decrease in inhibitory activity was observed versus ISA in analogy with which was reported in the past [20,33]. A general trend of decrease in inhibitory activity was observed with increasing alkyl chain length.

Biology
All synthesized compounds were evaluated in vitro against either purified enzyme (IleRS) in a buffer or in a cellular extract (S30), followed by determination of MIC values against different microbes.

Measurement of in vitro Inhibitory Activity with Purified E. coli IleRS
Following successful synthesis of compounds (6a-f), inhibitory activity was determined using radiolabel transfer assay. IleRS and total tRNA isolated from E. coli were used. In this assay the quantity of C 14 labelled isoleucine transferred to tRNA Ile was determined by precipitating the [C 14 ] Ile-tRNA Ile complex using a 10% TCA solution. Out of six tested derivatives, four showed IC50 in the nanomolar range in the radiolabel transfer assay (6a,b and 6e,f; Figure 2A and Table 1). However, for all four a 2-or 3-fold decrease in inhibitory activity was observed versus ISA in analogy with which was reported in the past [20,33]. A general trend of decrease in inhibitory activity was observed with increasing alkyl chain length. The activity of the enzyme is reported as a percentage value relative to that measured in the absence of inhibitor. The presented fit of the measured points was calculated using the Sigmoidal dose-response (variable slope) [34]. (B). Relative inhibitory activity of the dodecyl and octadecyl derivatives at higher concentration against purified E. coli IleRS. The relative activity was determined by comparing to values measured in the absence of an inhibitor and assuming its 100% enzyme activity. Average of three experiments with SD error bar. Indeed, the octadecyl compound 6d was found to be active only in the micromolar range, while the dodecyl derivative 6c was active in submicromolar range ( Figure 2B); in view of their lower inhibitory activity, their dose-response curves were not determined. Remarkably, the O 6 -methyl derivative 6f showed about 3-fold better inhibitory activity compared to the N 6 -methylated ISA 6a and matched the activity of ISA (Table 1), while the phenyl substituted ISA (6e) showed activity similar to methylamine substituted ISA (6a). The activity of the enzyme is reported as a percentage value relative to that measured in the absence of inhibitor. The presented fit of the measured points was calculated using the Sigmoidal dose-response (variable slope) [34]. (B). Relative inhibitory activity of the dodecyl and octadecyl derivatives at higher concentration against purified E. coli IleRS. The relative activity was determined by comparing to values measured in the absence of an inhibitor and assuming its 100% enzyme activity. Average of three experiments with SD error bar. Indeed, the octadecyl compound 6d was found to be active only in the micromolar range, while the dodecyl derivative 6c was active in submicromolar range ( Figure 2B); in view of their lower inhibitory activity, their dose-response curves were not determined. Remarkably, the O 6 -methyl derivative 6f showed about 3-fold better inhibitory activity compared to the N 6 -methylated ISA 6a and matched the activity of ISA (Table 1), while the phenyl substituted ISA (6e) showed activity similar to methylamine substituted ISA (6a).

Time-dependent in vitro Inhibitory Activity with E. coli Cellular Extract
The compounds with either a 4-nitrobenzyloxycarbonyl (14) or 4-acetoxybenzyloxycarbonyl moiety (18) attached to the alpha-amino group of LSA were tested in E. coli S30 cellular extract to determine the time required for metabolic activation of the mentioned promoieties. Therefore, the compounds were incubated with the cellular extract at 37 • C for different time periods and inhibitory activities were measured in comparison with LSA. The cellular lysate of E. coli appeared to be enriched with the nitroreductases and esterases responsible for activation to the parent compound. Both prodrugs showed equal inhibitory effect on LeuRS compared to LSA, irrespective of the time of incubation. Hence, in view of using only 250 nM of prodrug equivalent to the concentration of the parent inhibitor LSA, the (partial) early release of the LSA warhead must be concluded (Figure 3). The compounds with either a 4-nitrobenzyloxycarbonyl (14) or 4-acetoxybenzyloxycarbonyl moiety (18) attached to the alpha-amino group of LSA were tested in E. coli S30 cellular extract to determine the time required for metabolic activation of the mentioned promoieties. Therefore, the compounds were incubated with the cellular extract at 37 °C for different time periods and inhibitory activities were measured in comparison with LSA. The cellular lysate of E. coli appeared to be enriched with the nitroreductases and esterases responsible for activation to the parent compound. Both prodrugs showed equal inhibitory effect on LeuRS compared to LSA, irrespective of the time of incubation. Hence, in view of using only 250 nM of prodrug equivalent to the concentration of the parent inhibitor LSA, the (partial) early release of the LSA warhead must be concluded. The activity was determined by measuring the transfer of the appropriate C 14 -labeled amino acid to tRNA in the presence of 250 nM of each compound. The stated prodrugs were incubated at 37 °C in the cell extract for either 2, 15, 60, and 120 minutes, after which aliquots were taken, and the aminoacylation activity was measured. The relative activity was determined by comparing to the values of the extract measured in the absence of an inhibitor and assuming 100% enzyme activity. The results correspond to an average of three experiments, with SD presented as error bars.

Antimicrobial Assay
All the synthesized analogues were tested against six different microbes to cover the spectrum of activity against gram-positive-Staphylococcus aureus ATCC 6538P, Staphylococcus epidermidis RP62A, Sarcina lutea ATCC9341; gram-negative-Escherichia coli NCIB 8743, P. aeruginosa PAO1; and fungi-Candida albicans CO11. The results are described in Table 2. The antimicrobial activities were obtained by measuring the optical density at 600 nM of the individual wells of a microtiter plate reached by the microbial culture in the presence of different concentrations of respective inhibitors.  The activity was determined by measuring the transfer of the appropriate C 14 -labeled amino acid to tRNA in the presence of 250 nM of each compound. The stated prodrugs were incubated at 37 • C in the cell extract for either 2, 15, 60, and 120 min, after which aliquots were taken, and the aminoacylation activity was measured. The relative activity was determined by comparing to the values of the extract measured in the absence of an inhibitor and assuming 100% enzyme activity. The results correspond to an average of three experiments, with SD presented as error bars.

Antimicrobial Assay
All the synthesized analogues were tested against six different microbes to cover the spectrum of activity against gram-positive-Staphylococcus aureus ATCC 6538P, Staphylococcus epidermidis RP62A, Sarcina lutea ATCC9341; gram-negative-Escherichia coli NCIB 8743, P. aeruginosa PAO1; and fungi-Candida albicans CO11. The results are described in Table 2. The antimicrobial activities were obtained by measuring the optical density at 600 nM of the individual wells of a microtiter plate reached by the microbial culture in the presence of different concentrations of respective inhibitors. The aaSA derivatives having a dodecyl or octyl substitution at the N 6 -position proved to be active against some microbes. The octyl derivative (6b) showed the best MIC against S. lutea of 12.5 µM and similarly, dodecyl derivative (6c) showed the best MIC of 6.25 µM against C. albicans. This proves an amelioration of the uptake properties via increase of hydrophobic-hydrophilic balance, but less than was hoped for. Potentially, the distance between hydrophobic and hydrophilic portions of the molecule or so called "amphiphilic moment", likewise contributes to improved uptake properties [35]. Unexpectedly, both compounds with a prodrug moiety were found to be inactive under the applied assay conditions. Possibly the promoieties are released too rapidly to secure sufficient uptake of these prodrugs. The compound with octadecyl (6d) substitution on the other hand, might be too lipophilic to dissolve thoroughly in the culture media (providing aggregates) or proved unable to cross the bacterial membrane to show its activity.

Computational Analysis of Molecular Properties
For understanding the change in physicochemical properties of the synthesized compounds with respect to the parent analogues, computational prediction of molecular properties was attempted. The different parameters were determined using the online physicochemical predictor toolkit (www. molinspiration.com/cgi-bin/properties). Both modification strategies as expected appeared to be increasing the compounds' partition coefficient (logP). The analogues carrying either a phenyl, octyl, or dodecyl substitution at the N 6 -position and both synthesized prodrugs more or less satisfied the Lipinski rules (Table 3). Obviously, there is no change in the number of hydrogen bond donors or acceptors and in total polar surface area for C6-modified compounds. However, the number of hydrogen bond acceptors and polar surface area slightly increases for the synthesized prodrugs in comparison to LSA (not included in Table 3).

Discussion
The primary reason for the failure of antibacterial lead molecules discovered after a rational design approach by SAR, is their limited permeability through the bacterial membrane. In the past, trying to overcome the permeability issue of aaSA and their derivatives, our group already synthesized several conjugates comprising of aaSA analogues coupled with various transporter peptide [17][18][19][20][21] or siderophores [21]. However, we achieved limited success due to tedious synthesis, purification, and stability of peptide and siderophore coupled compounds.
There is widespread belief that the Lipinski "rule-of-five" guidelines for oral uptake of classical drugs should not be extrapolated to antibiotics and indeed many (natural) antibiotics do not comply with this guideline. A more recent report however, made a clear distinction between compounds targeting riboproteins or those targeting regular bacterial protein targets, where the latter upon analysis mostly comply with the Lipinski rules [36]. In addition, it has been observed that increase in lipophilicity of sulfamoyladenosine derivatives leads to increased permeability of the latter through the bacterial membrane [12]. Following our abovementioned approach, we were able to synthesize six compounds having lopP values from one to five obeying the Lipinski rule (Table 3), except for the compounds having either a methoxy or octadecylamine substitution at the C6-position resulting in a too polar or too lipophilic compound, respectively. Therefore, in aim to improve the in vivo efficacy of aaSAs, we used 2 different strategies utilizing either one of two amino functionalities present in aaSAs. As one can observe from crystal structures that there is considerable space around the adenine binding region [24], we decided to attach lipophilic moieties to the adenine heterocycle. Thus, we synthesized six lipophilic N 6 -modified analogues of ISA with the intention to find a compound which can cross the bacterial membrane while retaining the inhibitory activity. As expected, four compounds (6a, 6b, 6e, and 6f) out of six were found to have quite similar inhibitory activity against IleRS as compared to the parent analogue ISA. The gradually decreasing inhibitory activity with increasing chain length (6a-d) could be due to increasing entropic losses of the longer alkyl moieties, insufficiently compensated by increased hydrophobic interaction. Compounds 6b and 6c showed some improved antimicrobial activity against different microbes, but less than was hoped for. The longer octadecyl chain presumably leads to aggregates, leading again to reduced uptake. These results also show that our approach of substituting the N 6 -position was the correct one, as only a limited reduction of the enzyme inhibitory activity was observed. This is in stark contrast to our previous efforts of methylating the alpha-amine, which was accompanied with a dramatic loss in inhibitory activity [37]. The adenine moiety seems not to be essential to generate high affinity molecules against class I aaRS, as it can be readily substituted with other nucleobases or even tetrazole moieties while still showing good inhibitory activity [11,38].
In the second approach, we used prodrug moieties to enhance the permeability and also increase the bacterial selectivity towards these compounds. Therefore, we opted for 4-acetoxybenzyloxycarbonyl and 4-nitrobenzyloxycarbonyl moieties, which can be cleaved by esterases or nitroreductases, respectively. Following coupling of the promoieties to the α-amine of the LSA, the compounds were evaluated in a cellular lysate. We hypothesized that compounds would take some time for activation and therefore planned a time-point study from 2 to 120 min. However, the compounds appeared to be sufficiently metabolized already within 2 min to retain the full activity of the parent molecules. On further testing of these prodrugs against different microbes, no significant inhibitory activity could be observed, which might hint to preliminary degradation in the LB media, or alternatively, still no uptake in bacterial cells is accomplished. Nitrobenzyl carbamates (NBC) of a variety of cytotoxic amines are metabolized efficiently by nitroreductases to the hydroxylamines, which fragment to release the amines [29,39,40]. Preliminary chemical hydrolysis of our prodrugs is unlikely at the physiological pH buffer conditions used (see experimental). This type of prodrug has been used before at many occasions, as a means of directed prodrugs for cancer treatment. NBC derivatives of doxorubicine with the carbamate attached to the aminated sugar daunosamine [29], resemble the closest our alpha-amine carbamate as in 14, and still show 36% remaining carbamate following 24 h incubation in minimum essential medium (Eagle) supplemented with 5% fetal calf serum. Obviously however, selectivity in inhibiting the bacterial aaRS is another issue in development of the aaSA compounds, only dealt with in this work in using the nitrobenzylated prodrug. But the feasibility of selective targeting has been shown previously by the availability of two marketed drugs inhibiting an aaRS, with mupirocin (likewise equipped with a long aliphatic tail) and tavaborole, and in using heterocyclic sulfonamide scaffolds [41,42]. Our extensive 3D structural work [33] on various aaRS in complex with inhibitory ligands (including unpublished work) will further pave the way for improved selectivity.

Conclusions and Future Perspective
A challenging and lengthy synthesis of C 6 -purine substituted analogues of ISA and of two LSA based prodrugs was performed to increase the lipophilicity of their parent compounds. The C 6 -purine substituted compounds showed potent inhibitory activity versus purified IleRS in a radiolabel transfer assay, albeit at a slightly higher concentration than the parent compound. The synthesized prodrugs appeared very effective against LeuRS in an E. coli cellular extract, showing rapid conversion to the parent compound without apparent loss in efficacy. The compounds 6b and 6c proved most effective displaying MIC in the micromolar range against few microbes indicating a positive effect of octyl and dodecyl substitution on permeation through the bacterial membrane and subsequent IleRS inhibitory activity. In general, although we met some difficulties, we still believe these novel approaches for altering the physicochemical properties of potent but too polar lead molecules could be utilized for their further development towards a novel antibiotic targeting an aminoacyl-tRNA synthetase. Detailed crystallographic studies on the interaction of the N 6 -modified compound with IleRS will further guide the rational design of future compounds against aaRSs.

Materials and Methods
Reagents and solvents were purchased from commercial suppliers and used as provided unless indicated otherwise. DMF and THF were of analytical grade and were stored over 4 Å molecular sieves. All other solvents used for reactions were analytical grade and used as provided. Reactions were carried out in oven-dried glassware under a nitrogen atmosphere with stirring at room temperature unless indicated otherwise. All microwave irradiation experiments were carried out in a dedicated CEM-Discover mono-mode microwave apparatus. C 14 -radiolabeled amino acids and scintillation liquid were purchased from Perkin Elmer. 1 H and 13 C NMR spectra of the compounds dissolved in CDCl 3 , CD 3   As reported in literature [43], compound 1 (6-chloropurine riboside, 8 g, 0.029 mol) was stirred with a mixture of dimethoxypropane (DMP) (34.31 mL, 0.29 mol) and paratoluenesulfonic acid (PTSA) (2.66 g, 0.014 mol) in dry acetone (80 mL) at room temperature for overnight. Thin layer chromatography TLC (developed at 10% methanol in dichloromethane (DCM)) was used to monitor the reaction. Saturated sodium bicarbonate was added to quench the reaction. Afterwards, the solvent was evaporated under reduced pressure. The crude product was dissolved in DCM, and the organic layer was washed two times with saturated sodium bicarbonate and one time with brine. Column chromatography was performed with a gradient of 5% methanol in DCM to obtain the desired compound 2 at 86% yield.   Formic acid (1.73 mL, 46 mmol) was added dropwise to CSI (4.0 mL, 46 mmol) in an ice bath for 10 min. After several minutes, the formation of a white solid was observed. Afterwards, acetonitrile (20 mL) was added and the solution was stirred for four hours at room temperature. Following stirring, the solution was added to compound 2 (5 g, 15 mmol) in DMA (20 mL) and was reacted overnight at room temperature. Column chromatography was performed with a gradient of 15-25% acetone in hexane to obtain desired compound at 73% (4.53 g) yield. 1  Formic acid (0.62 mL) was added dropwise to CSI (1.42 mL) in an RBF kept in an ice bath for 10 min. After a few minutes, the formation of a white solid was observed. Afterwards, acetonitrile (20 mL) was added, and the solution was stirred for four hours at room temperature. The obtained solution of sulfamoyl chloride is used in the next reaction as such. Sulfamoyl chloride solution (2.65 mL, 1.95 mmol) was added to a solution of compound 3a (208 mg, 0.65 mmol) in DMA (10 mL) and was stirred overnight. TLC (50:50 acetone/hexane) was used to monitor the reaction. After overnight stirring, the solvent was evaporated under reduced pressure. Column chromatography was performed with a gradient of 40-50% acetone in hexane to obtain 4a. Yield: 25% (0.065 g). 1 H NMR (300 MHz, CDCl 3 ) δ 1.3 (s, 3H, C-CH 3 ), 1.6 (s, 3H, C-CH 3  Compound 10 (300 mg, 0.52 mmol), Boc-Leu-OSu (1.2 equivalent, 205.85 mg, 0.63 mmol), DBU (1 equivalent, 0.08 mL, 0.52 mmol) were added together using DMF (10 mL) as solvent. The reaction, which after a short period of time turned pink, was let to react at room temperature and overnight. A small sample was work-upped with EtOAc and 10% KHSO 4 for TLC analysis which was later developed with 1% MeOH in EtOAc and sprayed with ammonium molybdate (R f = 0.73). The solvents were evaporated after completion of the reaction and the obtained wine-red dense liquid was partitioned between water and EtOAc. A small amount of 10% KHSO 4 was added in the first wash to assure that the pH from the aqueous layer had pH 5-6. The organic layers were pyrophosphatase. After 10 min, pre-warmed ATP was added to the mixture at a final concentration of 500 µM. The reaction was quenched by the addition of 4 µL of quenching buffer containing 0.2 M sodium acetate pH 4, 0.1% N-lauroylsarcosine and 5 mM unlabeled isoleucine. 20 µL was spotted on 3MM Whatman paper. After thorough washing with cold 10% TCA, the filters were washed twice with acetone and air dried. Addition of scintillation liquid was followed by measurement of the radioactivity using the scintillation counter. The linear zone of enzyme activity was determined for each aaRS. The quenching time was picked within this zone at which approximately 50% of total RNA is aminoacylated. The quenching time of six minutes was used.

Time-Dependent in vitro Inhibitory Activity with E. coli Cellular Extract
To determine the time-dependent inhibitory activity of prodrugs (compound 14 and 18), a mixture of inhibitor (at a stock concentration of 5 µM): S30 extract (1:4) was incubated at 37 • C for the specified period of time. The S30 extract was prepared as disclosed before 20 . The addition of inhibitor to cellular extract was done at time point zero and after 2, 15, 60 and 120 min, respectively, 5 µL of this mixture was added to 15 µL of the aminoacylation mixture which was kept at 37 • C and which contains phosphate (50 mM, pH 7.5), DTT (1 mM), E. coli MRE 600 tRNA (5 g/L purchased from Sigma), ATP (3 mM), magnesium acetate (10mM), potassium acetate (100mM), and 28.6 µM of 14C-radiolabeled leucine. The aminoacylation reaction was quenched after one minute by addition of 4 µL mixture of 0.2 M sodium acetate pH 4, 0.1% N-lauroylsarcosine, and 5 mM leucine. Then 10 µL of the reaction mixture was spotted on 3MM Whatman paper and this was transferred to 10% cold TCA solution. The papers were washed thoroughly with 10% cold TCA (twice), then the papers were washed twice with acetone and later dried in air. Dried papers were transferred to scintillation vial followed by the addition of scintillation liquid (12 mL), the amount of radionuclide incorporation was determined using a Tri-card 2300 TR liquid scintillation counter.

Antimicrobial Testing
The respective microbes were inoculated overnight in LB medium (5 mL) and cultured again in the next morning in fresh LB medium (5 mL) till it reached the OD 600 of approximately 0.9. An amount of 10 µL of compounds made up in 50:50 DMSO:H 2 O solution was used for testing. Compounds were serially diluted using 50:50 DMSO:H 2 O mixture in a 96-well plate and 50:50 DMSO:H 2 O solution was used as control. Next, 90 mL of bacterial cell culture grown to a OD600 of 0.05 was added. The cultures were next placed in an incubator at 37 • C, and subsequently, the OD600 was determined after 20-24 h. Bacterial strains and fungi used for the evaluations are S. aureus ATCC 6538P, S. epidermidis RP62A, E. coli NCIB 8743, P. aeruginosa PAO1, S. lutea ATCC9341 and C. albicans CO11. All experiments were performed in triplicate.