A New Way to 2,3,4-Trisubstituted Benzo[h]quinolines: Synthesis, Consecutive Reactions and Cellular Activities

The reaction of mercaptoacetic acid esters with pentachloro-2-nitro-1,3-butadiene provides the appropriate precursors for the synthesis of 2,3,4-trisubstituted benzo[h]quinolines. These heterocycles are easily accessible via a single-step reaction with naphthalen-1-amine or anthracen-1-amine as the precursor. Due to the steric bulk and high electron density ring, the ring closure of benzo[h]quinolines takes place exclusively. Such highly substituted annelated pyridine systems can be modified in subsequent, selective reactions to build up new N-heterocycles with promising microbiological properties. The antibacterial and antiproliferative assays against four mammalian cell lines demonstrate that some of the sulfur-substituted benzo[h]quinoline analogs display potent phenotypic bioactivities in the single-digit micromolar range.


Introduction
Previous articles in the field of polyhalogenated nitrobutadienes have already demonstrated the enormous potential of pentachloro-2-nitro-1,3-butadiene (1) as a precursor for the "click synthesis" of highly functionalized (hetero)cyclic, as well as acyclic, compounds [1,2]. The corresponding syntheses that we have developed to date always start with the attack of an appropriate nucleophile at the activated terminal carbon atom of the nitrodichlorovinyl group within 1 to undergo a vinylic substitution reaction. Thus, in the case of sulfur nucleophiles, the corresponding thioperchlorobutadiene derivatives are easily accessible [3,4]. In this paper, we describe the unexpected reaction of thioperchloronitrobutadienes to 2,3,4-trisubstituted benzo[h]quinolines, which are tricyclic azaheterocycles with an unusual substitution pattern. Benzo [h]quinolines are important natural products, which are isolated from the stem wood of Zanthoxylum nitidum [5] or from the roots of Zanthoxylum capense [6]. Alkaloid decarine ( Figure 1) shows high antibacterial activity against mycobacterial, Gram-positive, and Gram-negative bacteria, and low cytotoxicity toward human macrophages [7]. Sanguinarine (Figure 1) is one of the most examined members in the class of natural 3,4-disubstituted benzo[h]quinoline compounds and has multiple application possibilities due to its broad scale of bioactivities, such as antibacterial These selected applications, in combination with the unique substitution pattern of the novel benzo[h]chinolines, presented herein indicate that it is promising to continue this field of medicinal chemistry.
A conceivable mechanism for the cascade reaction leading to the benzo[h]quinoline 5 is presented in Scheme 3. The starting material 4 exists exclusively as an E-isomer, probably due to a strong hydrogen bond between the NH and NO 2 groups. Initially, (E,E)-imine A is formed from diene 4 upon enamine-imine tautomerization. Due to an almost free rotation around the central single bond of the side chain, an intramolecular electrocyclization of A to give the intermediate B as part of an equilibrium appears feasible. Subsequently, dihydropyridine B is aromatized by the elimination of HCl in the presence of triethylamine, giving benzo[h]quinoline 5. A Friedel-Crafts substitution or a Michael addition appears less likely under the applied reaction conditions. A conceivable mechanism for the cascade reaction leading to the benzo[h]quinoline 5 is presented in Scheme 3. The starting material 4 exists exclusively as an E-isomer, probably due to a strong hydrogen bond between the NH and NO2 groups. Initially, (E,E)imine A is formed from diene 4 upon enamine-imine tautomerization. Due to an almost free rotation around the central single bond of the side chain, an intramolecular electrocyclization of A to give the intermediate B as part of an equilibrium appears feasible. Subsequently, dihydropyridine B is aromatized by the elimination of HCl in the presence of triethylamine, giving benzo[h]quinoline 5. A Friedel-Crafts substitution or a Michael addition appears less likely under the applied reaction conditions. The structure of benzo[h]quinoline 5a was also confirmed by an X-ray single-crystal analysis ( Figure 3). The structure of benzo[h]quinoline 5a was also confirmed by an X-ray single-crystal analysis ( Figure 3).  The key step of the reaction path to form benzo[h]quinolines 5 is a cyclization reaction. There are two possible ways to obtain product 5. The first one is a one-pot reaction of mercaptoacetate 3, naphthalene-1-amine, and triethylamine as a base in THF with up to 33% yield (path A, Scheme 2). Following the synthetic route A, it also proves to be possible to synthesize naphtho[h]quinoline 6 at a 54% yield. To improve the yield, we varied the base (triethylamine, N,N-dimethylaniline, and pyridine, without a base), solvent (THF, methanol, and chloroform), reaction temperature, and time. Benzo[h]quinolines were formed at moderate yields when triethylamine or N,N-dimethylaniline was used as a base. Without a base, the intermediate product 4 could be isolated at yields up to 89%. The interaction of isolated 4 in a subsequent reaction with triethylamine (path B) The key step of the reaction path to form benzo[h]quinolines 5 is a cyclization reaction. There are two possible ways to obtain product 5. The first one is a one-pot reaction of mercaptoacetate 3, naphthalene-1-amine, and triethylamine as a base in THF with up to 33% yield (path A, Scheme 2). Following the synthetic route A, it also proves to be possible to synthesize naphtho[h]quinoline 6 at a 54% yield. To improve the yield, we varied the base (triethylamine, N,N-dimethylaniline, and pyridine, without a base), solvent (THF, methanol, and chloroform), reaction temperature, and time. Benzo[h]quinolines were formed at moderate yields when triethylamine or N,N-dimethylaniline was used as a base. Without a base, the intermediate product 4 could be isolated at yields up to 89%. The interaction of isolated 4 in a subsequent reaction with triethylamine (path B) led to the formation of 5 at yields up to 93%. Therefore, it appeared advantageous to carry out the reaction in a two-step mode (Scheme 2).
Since path B had been found to be the more efficient one, an attempt was made to synthesize more intermediate products under these reaction conditions. A variety of aromatic amines were selected (Scheme 4).
The key step of the reaction path to form benzo[h]quinolines 5 is a cyclization reaction. There are two possible ways to obtain product 5. The first one is a one-pot reaction of mercaptoacetate 3, naphthalene-1-amine, and triethylamine as a base in THF with up to 33% yield (path A, Scheme 2). Following the synthetic route A, it also proves to be possible to synthesize naphtho[h]quinoline 6 at a 54% yield. To improve the yield, we varied the base (triethylamine, N,N-dimethylaniline, and pyridine, without a base), solvent (THF, methanol, and chloroform), reaction temperature, and time. Benzo[h]quinolines were formed at moderate yields when triethylamine or N,N-dimethylaniline was used as a base. Without a base, the intermediate product 4 could be isolated at yields up to 89%. The interaction of isolated 4 in a subsequent reaction with triethylamine (path B) led to the formation of 5 at yields up to 93%. Therefore, it appeared advantageous to carry out the reaction in a two-step mode (Scheme 2).
Since path B had been found to be the more efficient one, an attempt was made to synthesize more intermediate products under these reaction conditions. A variety of aromatic amines were selected (Scheme 4). Interestingly, conversion of 3 with a one-ring system, such as 3,5-dimethylaniline or 4-(benzyloxy)aniline in methanol, afforded (Z)-thiazolidinones 7a and 7b at a yield of 76% and 89%, respectively, which is similar to the literature [3], and not the open-chain product 4. Quinoline-8-amine produced (Z)-thiazolidinone 7c as two atropisomers (1:0.75, 1 H-NMR) at a 70% yield. This phenomenon is due to a less hindered rotation of the C-N bond of the former quinoline-8-amine unit. Unfortunately, there were no conversions to the desired product in the case of 5-aminonaphthalene-2-sulfonic acid, naphthalene-1,8-diamine, naphthalene-1,5-diamine, and anthracen-1-amine.
The original substituents of benzo[h]quinolines 5 allow their selective transformation into other functional groups, too. For example, the sulfoxide group of 8 is a good leaving group and can therefore be used in SN reactions, probably proceeding via an additionelimination mechanism. This way, both S-(sulfide 10a-d) and N-substituents (amines 11a,b) can be introduced. Access to sulfoxide 8 is gained by the selective oxidation of 5a Scheme 4. Synthetic pathways to aminothiodienes 4a,b and thiazolidinones 7a-c.
The original substituents of benzo[h]quinolines 5 allow their selective transformation into other functional groups, too. For example, the sulfoxide group of 8 is a good leaving group and can therefore be used in S N reactions, probably proceeding via an additionelimination mechanism. This way, both S-(sulfide 10a-d) and N-substituents (amines 11a,b) can be introduced. Access to sulfoxide 8 is gained by the selective oxidation of 5a with 35% of hydrogen peroxide, and this subsequently leads to sulfone 9 in very good yields. The reaction steps are temperature controlled (Scheme 5).
Molecules 2023, 28, x FOR PEER REVIEW 6 of 21 with 35% of hydrogen peroxide, and this subsequently leads to sulfone 9 in very good yields. The reaction steps are temperature controlled (Scheme 5).
The selective substitution reactions of the sulfoxide group of 8 proceeded well with different thiophenolates and morpholine (Scheme 6). The nucleophilicity of alkoholates proved to be too low for an SN reaction. The selective substitution reactions of the sulfoxide group of 8 proceeded well with different thiophenolates and morpholine (Scheme 6). The nucleophilicity of alkoholates proved to be too low for an S N reaction. The selective substitution reactions of the sulfoxide group of 8 proceeded well with different thiophenolates and morpholine (Scheme 6). The nucleophilicity of alkoholates proved to be too low for an SN reaction.

Scheme 6. Substitution reactions of sulfoxide 8 with S-and N-nucleophiles.
In the case of quinoline 11b, we were able to carry out an X-ray single-crystal analysis to prove the structural conclusions that we had drawn from the NMR spectra (see Figure  4). In the case of quinoline 11b, we were able to carry out an X-ray single-crystal analysis to prove the structural conclusions that we had drawn from the NMR spectra (see Figure 4).  Another method to modify compound 5a is the almost quantitative acidic cleavage of its ester group to 2-((4-(dichloromethyl)-3-nitrobenzo[h]quinolin-2-yl)thio)acetic acid (13). This molecule deserves interest because of its high activity against methicillin-resistant Staphylococcus aureus (MRSA) in microbiological tests. The oxidation of acid 13 with hydrogen peroxide furnished lactam 14. The expected sulfone E (Scheme 7), in contrast to sulfone 9 (Scheme 5), was not isolable. We assume that the primarily formed E was hydrolyzed instead under the reaction conditions to give quinolin-2-ol F. The tautomerization of the intermediate F led to lactam 14 at a 70% yield. The esterification of 14 with benzoyl chlorides gave 15a and 15b at a 77% and 78% yield, respectively.  Another method to modify compound 5a is the almost quantitative acidic cleavage of its ester group to 2-((4-(dichloromethyl)-3-nitrobenzo[h]quinolin-2-yl)thio)acetic acid (13). This molecule deserves interest because of its high activity against methicillin-resistant Staphylococcus aureus (MRSA) in microbiological tests. The oxidation of acid 13 with hydrogen peroxide furnished lactam 14. The expected sulfone E (Scheme 7), in contrast to sulfone 9 (Scheme 5), was not isolable. We assume that the primarily formed E was hydrolyzed instead under the reaction conditions to give quinolin-2-ol F. The tautomerization of the intermediate F led to lactam 14 at a 70% yield. The esterification of 14 with benzoyl chlorides gave 15a and 15b at a 77% and 78% yield, respectively. (13). This molecule deserves interest because of its high activity against methicillin-resistant Staphylococcus aureus (MRSA) in microbiological tests. The oxidation of acid 13 with hydrogen peroxide furnished lactam 14. The expected sulfone E (Scheme 7), in contrast to sulfone 9 (Scheme 5), was not isolable. We assume that the primarily formed E was hydrolyzed instead under the reaction conditions to give quinolin-2-ol F. The tautomerization of the intermediate F led to lactam 14 at a 70% yield. The esterification of 14 with benzoyl chlorides gave 15a and 15b at a 77% and 78% yield, respectively. Additionally, a convenient way to synthesize 2-substituted azolylbenzoquinolines in three steps, starting from nitrodiene 1, had been developed. At first, diene 1 was reacted with azoles, such as benzotriazole, 1,2,4-triazole, and pyrazole, leading to the formation of 1,1-diazolyl-2-nitrotrichlorobutadienes 16a-c in good yields (76-92%). The transamination of compound 16 with an equimolar amount of naphthalen-1-amine ran smoothly in ether at −30 °C and then r.t., furnishing 1-azolyl-1-(naphthylamino)diene 17 (86-90%). The treatment of butadienes 17 with triethylamine as a base in CHCl3 again led to the formation of 2,3,4-trisubstituted benzo[h]quinolines 18, at 36-87% yield (Scheme 8). A change of solvent (CH2Cl2, DMSO, MeOH, or Et2O) and base (NaOH or NaHCO3) decreased the yields of quinoline 18a to 9-55%. Additionally, a convenient way to synthesize 2-substituted azolylbenzoquinolines in three steps, starting from nitrodiene 1, had been developed. At first, diene 1 was reacted with azoles, such as benzotriazole, 1,2,4-triazole, and pyrazole, leading to the formation of 1,1-diazolyl-2-nitrotrichlorobutadienes 16a-c in good yields (76-92%). The transamination of compound 16 with an equimolar amount of naphthalen-1-amine ran smoothly in ether at −30 • C and then r.t., furnishing 1-azolyl-1-(naphthylamino)diene 17 (86-90%). The treatment of butadienes 17 with triethylamine as a base in CHCl 3 again led to the formation of 2,3,4-trisubstituted benzo[h]quinolines 18, at 36-87% yield (Scheme 8). A change of solvent (CH 2 Cl 2 , DMSO, MeOH, or Et 2 O) and base (NaOH or NaHCO 3 ) decreased the yields of quinoline 18a to 9-55%. Because of the diverse bioactivities reported for benzo[h]quinolines before, a selection of the newly prepared analogs shown above was made to represent all key structural variants. These selected analogs were characterized using phenotypic cellular assays. Their antibiotic activities were measured by using growth inhibition assays against one Gram-positive pathogen, i.e., methicillin-resistant Staphylococcus aureus (MRSA), and two Gram-negative pathogens, i.e., Escherichia coli and Pseudomonas aeruginosa. None of the compounds could inhibit the Gram-negative strains at a concentration of 50 µM, but some analogs were active against MRSA (Table 1). Notably, compounds 8, 12 and 13 prevented the growth of MRSA with the minimal inhibitory concentrations (MICs) of 17.5, 1.2, and 8.4 µM (corresponding to 7.7, 0.5, and 3.3 µg/mL), respectively. Thus, 12 was the most potent antibiotic among all tested analogs. Interestingly, the formal reduction of sulfoxide in 12 to a sulfide, as in 10d, led to a complete loss of activity against MRSA. The importance of the naphth-1-yl residue in 8 was illustrated by the fact that its replacement by an anthracen-1-yl moiety, as present in 6, also led to an inactive compound. Additionally, the sulfur substituent was crucial for activity because nitrogen-substituted benzo[h]quinolines, such as the amines 11a and 11b, or the triazols and pyrazol 18a-18c, were inactive. Additionally, the tested non-cyclized precursors, such as 16c and 17a-17c, had no activity. Because of the diverse bioactivities reported for benzo[h]quinolines before, a selection of the newly prepared analogs shown above was made to represent all key structural variants. These selected analogs were characterized using phenotypic cellular assays. Their antibiotic activities were measured by using growth inhibition assays against one Gram-positive pathogen, i.e., methicillin-resistant Staphylococcus aureus (MRSA), and two Gram-negative pathogens, i.e., Escherichia coli and Pseudomonas aeruginosa. None of the compounds could inhibit the Gram-negative strains at a concentration of 50 µM, but some analogs were active against MRSA (Table 1). Notably, compounds 8, 12 and 13 prevented the growth of MRSA with the minimal inhibitory concentrations (MICs) of 17.5, 1.2, and 8.4 µM (corresponding to 7.7, 0.5, and 3.3 µg/mL), respectively. Thus, 12 was the most potent antibiotic among all tested analogs. Interestingly, the formal reduction of sulfoxide in 12 to a sulfide, as in 10d, led to a complete loss of activity against MRSA. The importance of the naphth-1-yl residue in 8 was illustrated by the fact that its replacement by an anthracen-1-yl moiety, as present in 6, also led to an inactive compound. Additionally, the sulfur substituent was crucial for activity because nitrogen-substituted benzo[h]quinolines, such as the amines 11a and 11b, or the triazols and pyrazol 18a-18c, were inactive. Additionally, the tested non-cyclized precursors, such as 16c and 17a-17c, had no activity. In addition to probing effects on bacterial pathogens, the compounds' ability to interfere with the proliferation and/or viability of four mammalian cell lines was assessed by quantifying their mitochondrial dehydrogenase activity in a colorimetric assay with the tetrazolium dye WST-1. For this purpose, the cells were exposed to the compounds at varying concentrations for a time period of 5 days (for L929, KB-3-1, and MCF-7 cells) or 24 h (for FS4-LTM cells). For 8 and 13, similar activity trends were observed against mammalian cell lines: The sulfoxide 8 was most potent and inhibited the proliferation of all four cell lines with EC 50 s of 2.5-3.8 µM. Compound 13, which differs from 8 by the hydrolysis of ester to a carboxylic acid and the oxidation of sulfide to a sulfoxide, had weaker activities with EC 50 s of 6.9-45 µM. In contrast, the aliphatic substituent of sulfoxide in 8 could be replaced by an aromatic p-fluorophenyl group, because 12 inhibited the proliferation of L929, KB-3-1, and FS4-LTM cells with EC 50 s of 3.3, 3.1, and 4.4 µM, respectively. Compound 10d, which differs from 12 by the oxidation state at sulfur, displayed significantly reduced activities in L929 and FS4-LTM cells, but it was more potent in KB-3-1 cells.
In summary, the phenotypic cellular assays demonstrate that a distinct subset of the benzo[h]quinolines synthesized in this study possess high bioactivities against mammalian cell lines and the bacterial pathogen MRSA.

General Remarks
The solvents and reagents were used as received from commercial sources without further purification. The TLC was performed with Merck aluminum-backed TLC plates using silica gel 60, F 254 . A flash column chromatography was performed using Macherey-Nagel silica gel 60 M (0.040-0.063 mm) with appropriate mixtures of petroleum ether (PE, boiling range 60-70 • C) and ethyl acetate as the eluents. The melting points were determined in capillary tubes with a Büchi B-520. The FTIR spectra were recorded with a Bruker "Alpha-T" spectrometer (Bruker, Ettlingen, Germany) with the solid compounds measured as KBr pellets. The ATR-IR spectra were measured on the same instrument with a Bruker "Alpha Platinum ATR" single-reflection diamond ATR module. The 1 H NMR and 13 C NMR spectra at 600 and 150 MHz, respectively, were recorded with an "Avance III" 600 MHz FTNMR spectrometer (Bruker, Rheinstetten, Germany). The 1 H NMR and 13 C NMR spectra at 400 and 100 MHz, respectively, were recorded with an "Avance" 400 MHz FT-NMR spectrometer (also Bruker). The 1 H and 13 C NMR spectra were examined with reference to the residual solvent peak: CDCl 3 at δ 7.26 ( 1 H) and 77.0 ppm ( 13 C), and DMSO-d 6 at δ 2.50 ( 1 H) and 39.7 ppm ( 13 C). The NMR spectra and HR-MS data of the newly synthesized compounds are available in the Supplementary Materials. The mass spectra were obtained with a Hewlett-Packard MS 5989B spectrometer (HP Inc., Palmer, MA, USA), usually in a direct mode with the electron impact (70 eV). For the chlorinated and brominated compounds, all peak values of molecular ions and fragments refer to the isotope 35 Cl. High-resolution mass spectra were recorded with the "Impact II" from BRUKER (Bruker Daltonik GmbH, Bremen, Germany).

Antibacterial Assays
Overnight cultures of the bacteria were grown aerobically at 37 • C in Müller-Hinton broth with added 1% glucose and a pH of 7.2 for Gram-negative strains, or with a Trypticase soy yeast extract medium (TSY-30 g/L Trypticase soy broth, 3 g/L yeast extract, pH 7.2) for Gram-positive strains. The cultures were adjusted to an OD600 nm of 0.001, which resulted in a final start OD600 nm of 0.0005 in the test. A total of 25 µL of test culture was added to 25 µL of a serial dilution of the test compounds in the appropriate medium for the different strains in accordance with standardized procedures in 384-well plates. For screening purposes, the residual absorbance in % was tested at the compound concentrations of 5 and 50 µM. For the selected compounds, the concentration-dependent growth inhibition curves were recorded from the stock solutions in DMSO at the final concentrations of 100, 50, 25, 12.5, 6,25, 3.125, 1.56, 0.78, 0.39, and 0.2 µM. As the positive control compounds, Linezolid (both MRSA strains) Ciprofloxacin (E. faecium, E. coli), and Amikacin (P. aeruginosa) were applied. The highest DMSO concentration in the assay was 1%, which had no apparent effect on the growth of the bacteria. After an incubation time of 18 h at 37 • C under moist conditions, the optical density at 600 nm was measured with a Fusion Universal Microplate Analyzer (Perkin-Elmer, Waltham, MA, USA). The lowest concentration that completely suppressed growth defined the MIC values. The following bacterial strains were used: Gram-negative: Escherichia coli (DSM 1116) and Pseudomonas aeruginosa PA7 (DSM 24068), and Gram-positive: Staphylococcus aureus MRSA (clinical isolate, RKI 11-02670) and Staphylococcus aureus MRSA (DSM 11822). The MIC values were determined by curve fitting with Sigma Plot.

Antiproliferative Assays
The effects of the compounds on cell viability were probed with a WST-1 test using the procedure of Ishiyama et al. [24] as modified by Sasse et al. [25]. The following cell lines were used: mouse fibroblast cell line L929 (DSM ACC 2), human cervical carcinoma cell line KB-3-1 (DSM ACC 158), and human breast cancer cell line MCF-7 (DSM ACC 115). In addition, the conditional immortalized human fibroblast cell line FS4-LTM (InScreenex, Braunschweig, Germany) was used without doxycyclin to induce primary cell-like behavior

Conclusions
Starting from the versatile building block, pentachloro-2-nitro-1,3-butadiene (1) The phenotypic cellular assays demonstrated that some of the synthesized benzo[h]quinolines possess high bioactivities against mammalian cell lines and the bacterial pathogen MRSA. Although the underlying molecular mechanism or target is unknown so far, the fact that the activity is not ubiquitous across the whole series, but depends on the distinct substitution patterns found in some analogs, suggests that it is not due to unspecific effects. In addition, the overlapping, but non-parallel effects against bacteria and mammalian cells imply that it might be possible to find compounds that selectively target bacteria vs. eucaryotic cells. These findings encourage further exploration of benzo[h]quinolines as scaffolds in compound collections assembled for bioactivity testing. The synthetic procedure reported in this study significantly facilitates the generation of such collections.