Synthesis of the Indole-Based Inhibitors of Bacterial Cystathionine γ-Lyase NL1-NL3

Bacterial cystathionine γ-lyase (bCSE) is the main producer of H2S in pathogenic bacteria such as Staphylococcus aureus, Pseudomonas aeruginosa, etc. The suppression of bCSE activity considerably enhances the sensitivity of bacteria to antibiotics. Convenient methods for the efficient synthesis of gram quantities of two selective indole-based bCSE inhibitors, namely (2-(6-bromo-1H-indol-1-yl)acetyl)glycine (NL1), 5-((6-bromo-1H-indol-1-yl)methyl)- 2-methylfuran-3-carboxylic acid (NL2), as well as a synthetic method for preparation 3-((6-(7-chlorobenzo[b]thiophen-2-yl)-1H-indol-1-yl)methyl)- 1H-pyrazole-5-carboxylic acid (NL3), have been developed. The syntheses are based on the use of 6-bromoindole as the main building block for all three inhibitors (NL1, NL2, and NL3), and the designed residues are assembled at the nitrogen atom of the 6-bromoindole core or by the substitution of the bromine atom in the case of NL3 using Pd-catalyzed cross-coupling. The developed and refined synthetic methods would be significant for the further biological screening of NL-series bCSE inhibitors and their derivatives.


Introduction
The majority of antibiotics used to treat infectious diseases are aimed at inhibiting essential bacterial proteins. However, the efficiency of traditional antibiotics decreases. At the same time, pathogenic bacteria develop resistance to their action, which makes it necessary to increase the concentrations of antibiotics used for therapy, combine them, and look for new antibacterial drugs. As an alternative approach, the effect of antibiotics that trigger damage in bacteria via oxidative stress can also be enhanced by blocking bacterial H 2 S-producing enzymes and, consequently, the production of H 2 S and glutathione. Quite recently, cystathionine γ-lyase (CSE) was suggested as an additional drug target, provided that selective inhibitors of this enzyme in bacterial pathogens (bCSE) and the low activity of these molecules against the human version of the enzyme (hCSE) are ensured [1]. In many pathogenic bacteria (e.g., Staphylococcus aureus or Pseudomonas aeruginosa), bCSE is the primary H 2 S producer and is usually involved in forming their resistance mechanisms relative to active antibiotics. In a recent study, Nudler's team performed large-scale in silico and in vitro screening. It succeeded in identifying three leaders among bCSE inhibitors that featured high activity, selectivity, and low toxicity [2]. These include the NL1 ((2-(6-bromo-1H-indol-1-yl)acetyl)glycine), NL2 (5-((6-bromo-1H-indol-1-yl)methyl)-2-methylfuran-3carboxylic acid), and NL3 (3-((6-(7-chlorobenzo[b]thiophen-2-yl)-1H-indol-1-yl)methyl)-1Hpyrazole-5-carboxylic acid) molecules, i.e., derivatives based on 6-bromoindole ( Figure 1).

Synthesis of 6-Bromoindole
Since the structures of all three target bCSE inhibitors, i.e., NL1, NL2, and NL3, are based on a fragment of 6-bromoindole or its derivative, the initial step was to choose a convenient synthesis of 6-bromoindole 4 as one of the main building blocks in amounts of at least several tens of grams. About a dozen of methods for its synthesis can be found in the literature [4][5][6]; a complete description of their drawbacks and advantages (along with a synthesis of another important natural derivative, 6-bromotryptamine) was summarized in the recent review article [7]. However, many reported methods possess problems with scalability, the availability of initial reagents, and low yields. From our point of view, the four-stage scheme based on the diazotization of para-aminotoluene followed by bromination [8] and the ring closure in two stages ( Figure 2) proved to be the most convenient approach [7].
This reaction sequence is characterized by moderate yields; it is easy to optimize and scale up to the required quantities, and it employs inexpensive starting reagents. Thus, the initial 6-bromoindole 4 was synthesized, and the target inhibitors were subsequently obtained.

Synthesis of 6-Bromoindole
Since the structures of all three target bCSE inhibitors, i.e., NL1, NL2, and NL3, are based on a fragment of 6-bromoindole or its derivative, the initial step was to choose a convenient synthesis of 6-bromoindole 4 as one of the main building blocks in amounts of at least several tens of grams. About a dozen of methods for its synthesis can be found in the literature [4][5][6]; a complete description of their drawbacks and advantages (along with a synthesis of another important natural derivative, 6-bromotryptamine) was summarized in the recent review article [7]. However, many reported methods possess problems with scalability, the availability of initial reagents, and low yields. From our point of view, the four-stage scheme based on the diazotization of para-aminotoluene followed by bromination [8] and the ring closure in two stages ( Figure 2) proved to be the most convenient approach [7]. NL2 (5-((6-bromo-1H-indol-1-yl)methyl)-2-methylfuran-3-carboxylic acid), and NL3 (3-((6-(7-chlorobenzo[b]thiophen-2-yl)-1H-indol-1-yl)methyl)-1H-pyrazole-5-carboxylic acid) molecules, i.e., derivatives based on 6-bromoindole ( Figure 1). It was shown that these compounds could be applied as potentiators for a considerable enhancement of the antibiotic effect on pathogenic bacterial microorganisms, including resistant strains. Thus, the further development of this approach requires efficient and readily available methods for synthesising NL1, NL2, and NL3 inhibitors, which have not been reported on a gram scale to date [3]. Our work aims to develop and optimize methods for synthesising NL1, NL2, and NL3 in near-gram quantities for biological tests with the careful verification of all the products using physicochemical methods.

Synthesis of 6-Bromoindole
Since the structures of all three target bCSE inhibitors, i.e., NL1, NL2, and NL3, are based on a fragment of 6-bromoindole or its derivative, the initial step was to choose a convenient synthesis of 6-bromoindole 4 as one of the main building blocks in amounts of at least several tens of grams. About a dozen of methods for its synthesis can be found in the literature [4][5][6]; a complete description of their drawbacks and advantages (along with a synthesis of another important natural derivative, 6-bromotryptamine) was summarized in the recent review article [7]. However, many reported methods possess problems with scalability, the availability of initial reagents, and low yields. From our point of view, the four-stage scheme based on the diazotization of para-aminotoluene followed by bromination [8] and the ring closure in two stages ( Figure 2) proved to be the most convenient approach [7].
This reaction sequence is characterized by moderate yields; it is easy to optimize and scale up to the required quantities, and it employs inexpensive starting reagents. Thus, the initial 6-bromoindole 4 was synthesized, and the target inhibitors were subsequently obtained. This reaction sequence is characterized by moderate yields; it is easy to optimize and scale up to the required quantities, and it employs inexpensive starting reagents.
Thus, the initial 6-bromoindole 4 was synthesized, and the target inhibitors were subsequently obtained.

Synthesis of NL1 Inhibitor
The NL1 molecule is the simplest one in terms of structure and synthesis [2]. NL1 is synthesized from 6-bromoindole in four steps by attaching the short peptide-like chain to the indole nitrogen atom (Figure 3). First, 6-bromoindole 4 is alkylated with bromoacetic Molecules 2023, 28, 3568 3 of 13 ester and then hydrolyzed to yield 2-(6-bromo-1H-indol-1-yl) acetic acid 6 [9,10]. The subsequent reaction of the latter with glycine ester hydrochloride under peptide synthesis conditions creates an amide bond [11].

Synthesis of NL1 Inhibitor
The NL1 molecule is the simplest one in terms of structure and synthesis [2]. NL1 is synthesized from 6-bromoindole in four steps by attaching the short peptide-like chain to the indole nitrogen atom ( Figure 3). First, 6-bromoindole 4 is alkylated with bromoacetic ester and then hydrolyzed to yield 2-(6-bromo-1H-indol-1-yl) acetic acid 6 [9,10]. The subsequent reaction of the latter with glycine ester hydrochloride under peptide synthesis conditions creates an amide bond [11]. DCC or EDC can be successfully used as a dehydrating agent. However, EDC shows the best results if the reagent ratio is excessive. The hydrolysis of the ester group in the resulting product 7 with lithium hydroxide provides the target compound: NL1 [12]. This synthetic route makes obtaining NL1 in near-gram quantities possible without special additional purification.

Synthesis of NL2 Inhibitor
NL2 is one of the most promising bCSE inhibitors [2]. Its structure is based on the 2methylfuran-3-carboxylic acid derivative in combination with 6-bromoindole linked via a methylene bridge. NL2 is synthesized by coupling these two fragments, followed by the hydrolysis of the ester group in the last two steps ( Figure 4). Methyl 5-(chloromethyl)-2methylfuran-3-carboxylate 12 is obtained in three stages from methyl 2-methylfuran-3carboxylate 9 [13][14][15] and is used as the furan building block for coupling with 6bromoindole. Methyl 2-methylfuran-3-carboxylate 9 can be obtained in one step by the rhodium-catalyzed cyclization of vinyl acetate with ethyl diazoacetate [16]. Next, methyl 2-methylfuran-3-carboxylate is introduced into the formylation reaction, followed by the reduction of the aldehyde group with sodium borohydride [13] and the replacement of the hydroxyl by chloride by treatment with mesyl chloride to produce the essential intermediate compound, methyl 5-(chloromethyl)-2-methylfuran-3-carboxylate 12 [15]. This synthetic pathway was significantly optimized during this work to obtain multigram quantities in high yields, which allowed us to obtain the NL2 molecule in near-gram quantities. DCC or EDC can be successfully used as a dehydrating agent. However, EDC shows the best results if the reagent ratio is excessive. The hydrolysis of the ester group in the resulting product 7 with lithium hydroxide provides the target compound: NL1 [12]. This synthetic route makes obtaining NL1 in near-gram quantities possible without special additional purification.

Synthesis of NL2 Inhibitor
NL2 is one of the most promising bCSE inhibitors [2]. Its structure is based on the 2-methylfuran-3-carboxylic acid derivative in combination with 6-bromoindole linked via a methylene bridge. NL2 is synthesized by coupling these two fragments, followed by the hydrolysis of the ester group in the last two steps ( Figure 4). Methyl 5-(chloromethyl)-2-methylfuran-3-carboxylate 12 is obtained in three stages from methyl 2-methylfuran-3-carboxylate 9 [13][14][15] and is used as the furan building block for coupling with 6bromoindole. Methyl 2-methylfuran-3-carboxylate 9 can be obtained in one step by the rhodium-catalyzed cyclization of vinyl acetate with ethyl diazoacetate [16]. Next, methyl 2-methylfuran-3-carboxylate is introduced into the formylation reaction, followed by the reduction of the aldehyde group with sodium borohydride [13] and the replacement of the hydroxyl by chloride by treatment with mesyl chloride to produce the essential intermediate compound, methyl 5-(chloromethyl)-2-methylfuran-3-carboxylate 12 [15]. This synthetic pathway was significantly optimized during this work to obtain multigram quantities in high yields, which allowed us to obtain the NL2 molecule in near-gram quantities.

Synthesis of NL3 Inhibitor
has the most complex structure of the antibiotic potentiators above. NL3 inhibits bCSE more efficiently than NL2 but is inferior to the latter in selectivity toward hCSE [2]. In contrast to NL1 and NL2, the bromine atom in the indole fragment of NL3 is replaced by 7-chlorobenzo[b]thiophene, which can practicably be performed by Pdcatalyzed cross-coupling based on the 6-bromoindole building block that is already developed for NL1 and NL2. The attachment of the heterocyclic fragment by crosscoupling requires that (7-chlorobenzo[b]thiophen-2-yl) boronic acid 17 be obtained first. The latter was synthesized from 2,3-dichlorobenzaldehyde in three steps ( Figure 5), including the cyclization of the benzothiophene ring, decarboxylation, and subsequent borylation [17][18][19]. The cross-coupling step was performed at the very end of the synthetic sequence after the pyrazole heterocycle had been assembled. Thus, with 17 as the key reagent, the main part of the NL3 synthetic sequence consists of four steps ( Figure 6). 6-Bromoindole 4 is first alkylated with propargyl bromide in the presence of sodium hydride to produce propargylindole 18 [20], which is then introduced into the [3+2]-cycloaddition reaction with ethyl diazoacetate at the triple bond to form the pyrazole ring in product 19 [4].

Synthesis of NL3 Inhibitor
NL3, i.e., 3-((6-(7-chlorobenzo[b]thiophen-2-yl)-1H-indol-1-yl) methyl)-1H-pyrazole-5-carboxylic acid, has the most complex structure of the antibiotic potentiators above. NL3 inhibits bCSE more efficiently than NL2 but is inferior to the latter in selectivity toward hCSE [2]. In contrast to NL1 and NL2, the bromine atom in the indole fragment of NL3 is replaced by 7-chlorobenzo[b]thiophene, which can practicably be performed by Pdcatalyzed cross-coupling based on the 6-bromoindole building block that is already developed for NL1 and NL2. The attachment of the heterocyclic fragment by crosscoupling requires that (7-chlorobenzo[b]thiophen-2-yl) boronic acid 17 be obtained first. The latter was synthesized from 2,3-dichlorobenzaldehyde in three steps ( Figure 5), including the cyclization of the benzothiophene ring, decarboxylation, and subsequent borylation [17][18][19]. The cross-coupling step was performed at the very end of the synthetic sequence after the pyrazole heterocycle had been assembled. Thus, with 17 as the key reagent, the main part of the NL3 synthetic sequence consists of four steps ( Figure 6). 6-Bromoindole 4 is first alkylated with propargyl bromide in the presence of sodium hydride to produce propargylindole 18 [20], which is then introduced into the [3+2]-cycloaddition reaction with ethyl diazoacetate at the triple bond to form the pyrazole ring in product 19 [4]. Thus, with 17 as the key reagent, the main part of the NL3 synthetic sequence consists of four steps ( Figure 6). 6-Bromoindole 4 is first alkylated with propargyl bromide in the presence of sodium hydride to produce propargylindole 18 [20], which is then introduced into the [3+2]-cycloaddition reaction with ethyl diazoacetate at the triple bond to form the pyrazole ring in product 19 [4].
In the alkylation of bromoindole 4 with propargyl bromide, the excess base reacts with the resulting alkyne 18. It causes its isomerization to produce the allene as a side product, which is hard to separate from the target propargylindole. To reduce the formation of the undesirable isomer, the base excess should be minimized, and a temperature increase should be avoided; presumably, the yield of the target alkyne can also be increased by using weaker bases that preferably promote NH-activation, such as DBU [21].
The resulting product 19 is introduced into a Pd(dppf)Cl 2 -catalyzed cross-coupling reaction with the previously synthesized (7-chlorobenzo[b]thiophen-2-yl) boronic acid 17 under thoroughly selected conditions. The target product NL3 is then easily obtained by the hydrolysis of the ester group, yet the cross-coupling step limits scaling due to the complexity of this stage and the low reactivity of the building blocks being coupled. In the alkylation of bromoindole 4 with propargyl bromide, the excess base reacts with the resulting alkyne 18. It causes its isomerization to produce the allene as a side product, which is hard to separate from the target propargylindole. To reduce the formation of the undesirable isomer, the base excess should be minimized, and a temperature increase should be avoided; presumably, the yield of the target alkyne can also be increased by using weaker bases that preferably promote NH-activation, such as DBU [21].
The resulting product 19 is introduced into a Pd(dppf)Cl2-catalyzed cross-coupling reaction with the previously synthesized (7-chlorobenzo[b]thiophen-2-yl) boronic acid 17 under thoroughly selected conditions. The target product NL3 is then easily obtained by the hydrolysis of the ester group, yet the cross-coupling step limits scaling due to the complexity of this stage and the low reactivity of the building blocks being coupled.
We tested several phosphine ligands (XPhos, SPhos, RuPhos, DavePhos, MePhos, CyJohnPhos, PhJohnPhos, XanthPhos, CyXanthPhos, PPh3, dppf, DTBPF, and t Bu-XPhos), bases (Na2CO3, K2CO3, Cs2CO3, Na3PO4, CsF, and t BuONa) and solvents (dioxane/water, toluene, and DMF) [22,23] to improve the cross-coupling step. However, the formation of target product 21 was not observed in any alternative variants. In fact, with weak bases and at room temperature, only the side dimerization of thiopheneboronic acid to produce the dimeric adduct was observed ( Figure 6) [24,25]. On the other hand, the reaction using aqueous carbonate bases (Na2CO3, K2CO3, and Cs2CO3) at elevated temperatures mainly resulted in the hydrolysis of 19 relative to the corresponding acid. In contrast, the side reaction of 17 dimerization still occurred. Thus, to date, using Pd(dppf)Cl2 as the catalyst We tested several phosphine ligands (XPhos, SPhos, RuPhos, DavePhos, MePhos, CyJohnPhos, PhJohnPhos, XanthPhos, CyXanthPhos, PPh 3 , dppf, DTBPF, and t Bu-XPhos), bases (Na 2 CO 3 , K 2 CO 3 , Cs 2 CO 3 , Na 3 PO 4 , CsF, and t BuONa) and solvents (dioxane/water, toluene, and DMF) [22,23] to improve the cross-coupling step. However, the formation of target product 21 was not observed in any alternative variants. In fact, with weak bases and at room temperature, only the side dimerization of thiopheneboronic acid to produce the dimeric adduct was observed ( Figure 6) [24,25]. On the other hand, the reaction using aqueous carbonate bases (Na 2 CO 3 , K 2 CO 3 , and Cs 2 CO 3 ) at elevated temperatures mainly resulted in the hydrolysis of 19 relative to the corresponding acid. In contrast, the side reaction of 17 dimerization still occurred. Thus, to date, using Pd(dppf)Cl 2 as the catalyst and Na 2 CO 3 as the base is the only option for synthesising compound 20 according to the sequence described above. Moreover, it is important to perform the reaction in a waterdioxane medium at a slow rate by stirring Na 2 CO 3 at the bottom with a magnetic stirrer. Cross-coupling product 20 is not separable from the original bromoindole 19 by column chromatography. As a result, a mixture of 19 and 20 were used for the hydrolysis, followed by the separation of products by preparative HPLC on the reverse phase C18.
Thus, using the cross-coupling method at the final step of NL3 assembly is the main complicating factor in the synthetic route and requires further studies and optimization, for example, by applying the benzothiophene ring assembly from 6-indole acetic ester under transition-metal-free conditions [26].

Structure Elucidation of NL1, NL2, and NL3
All compounds synthesized were converted to their lyophilized states, and their structures were confirmed using the standard methods of one-and two-dimensional NMR spectroscopy, including 15  Cross-coupling product 20 is not separable from the original bromoindole 19 by column chromatography. As a result, a mixture of 19 and 20 were used for the hydrolysis, followed by the separation of products by preparative HPLC on the reverse phase C18.
Thus, using the cross-coupling method at the final step of NL3 assembly is the main complicating factor in the synthetic route and requires further studies and optimization, for example, by applying the benzothiophene ring assembly from 6-indole acetic ester under transition-metal-free conditions [26].

Structure Elucidation of NL1, NL2, and NL3
All compounds synthesized were converted to their lyophilized states, and their structures were confirmed using the standard methods of one-and two-dimensional NMR spectroscopy, including 15   The compounds thus synthesized (NL1, NL2, and NL3) were passed for biological tests.
All NMR spectra are represented in Supplementary Material.

Conclusions
As a result, we have developed and presented convenient methods for synthesising two efficient bCSE inhibitors, NL1 and NL2, that make it possible to obtain them in gram quantities with a high degree of purity for further biological tests. Unfortunately, during the synthesis of NL3, due to the extremely low yields in the cross-combination reaction, the target substance was obtained only in tens of milligrams. The main synthetic scheme for all inhibitors (NL1, NL2, and NL3) uses 6-bromoindole as the main building block. The rest of the heterocyclic system (pyrazoline, furan, or peptide-like chain) is assembled at the nitrogen atom or with the replacement of the bromine atom by the Pd-catalyzed cross-coupling reaction in the case of NL3. bCSE inhibitors are promising compounds for potentiating antimicrobial therapy and circumventing bacterial resistance. The development of available methods for synthesising these inhibitors would allow their in vivo application modes to be perfected.

General Experimental Details
All reagents and catalysts were purchased from Sigma-Aldrich, Acros, J&K Scientific and TCI Europe and used without further purification unless otherwise mentioned. TLC analysis was performed on Silufol chromatographic plates. For preparative chromatography, silica gel 60 (0.040-0.063 mm) was used. 1 H, 13 C NMR spectra were recorded on a Bruker AVANCE II 300 MHz (300.1, 75.5 MHz and 282.4 MHz, respectively) and a Bruker AMX III 400 MHz (400.1, 100.6 MHz and 376.5 MHz, respectively) spectrometers in CDCl3, containing 0.05% Me4Si as the internal standard. Determination The compounds thus synthesized (NL1, NL2, and NL3) were passed for biological tests.
All NMR spectra are represented in Supplementary Material.

Conclusions
As a result, we have developed and presented convenient methods for synthesising two efficient bCSE inhibitors, NL1 and NL2, that make it possible to obtain them in gram quantities with a high degree of purity for further biological tests. Unfortunately, during the synthesis of NL3, due to the extremely low yields in the cross-combination reaction, the target substance was obtained only in tens of milligrams. The main synthetic scheme for all inhibitors (NL1, NL2, and NL3) uses 6-bromoindole as the main building block. The rest of the heterocyclic system (pyrazoline, furan, or peptide-like chain) is assembled at the nitrogen atom or with the replacement of the bromine atom by the Pd-catalyzed cross-coupling reaction in the case of NL3. bCSE inhibitors are promising compounds for potentiating antimicrobial therapy and circumventing bacterial resistance. The development of available methods for synthesising these inhibitors would allow their in vivo application modes to be perfected.