Substituted Azafluorenones: Access from Dihalogeno Diaryl Ketones by Palladium-Catalyzed Auto-Tandem Processes and Evaluation of their Antibacterial, Antifungal, Antimalarial and Antiproliferative Activities

from


Introduction
Many studies have been devoted to the access to azafluorenones, which are compounds of biological interest, for example in relation to their antifungal, 1 antimicrobial, 2 antimalarial, 2b,3 and cytotoxic 2b,4 activities, or else for their role in the treatment of neurodegenerative disorders. 5 Lithiations 6 and multicomponent reactions 2d,4a,7 can notably be cited among the modern synthetic methods by which to access them. In 2010, an approach was developed by Kraus and Kempema who used 2-bromoaryl 3pyridyl ketones, prepared by reaction of 3-pyridyllithiums with 2-bromobenzaldehydes followed by oxidation, in intramolecular Heck cyclization reactions. 2c In 2013, Ray and co-workers documented a two-step synthesis. Indeed, due to the facile oxidation of the corresponding α-aryl-2-bromo-3pyridylmethanols, they successfully performed both oxidation and cyclization reactions within one palladium-catalyzed step. 8 In the course of the development of 2,2,6,6-tetramethylpiperidido (TMP) bases for the deprotonative metalation of aromatic compounds, 9 we have developed the use of (TMP) 2 CuLi·LiCl, prepared in situ from CuCl and LiTMP (2 equiv). 10 Among the advantages of this lithiocuprate base, one can cite its possible use at room temperature and the possible trapping of the formed arylmetal species by aroyl chlorides to directly afford ketones. We have showed that, when applied to the synthesis of 2-chloro diaryl ketones, this method could be combined with direct arylation through C-H bond activation by intramolecular transition metal-catalysis, 11 to afford azafluorenones and related compounds in two steps. 12 Recently, we could achieve the synthesis of substituted azafluorenones by combining cyclization with cross-couplings such as Suzuki 13 and Heck 14 reactions in auto-tandem 15 processes starting from diaryl ketones suitably substituted by two halogeno groups. 16 Herein, the details of our synthetic investigations, including the testing of a large range of substrates, and the biological evaluation of the prepared compounds for their antibacterial, antifungal, antimalarial and cytotoxic activity are described.

Results and Discussion
The starting dihalogeno aroylpyridines were synthesized by deprotocupration-aroylation, as previously described. 10 Upon treatment by the lithiocuprate (TMP) 2 CuLi·LiCl, generated in situ from CuCl and LiTMP (2 equiv), in tetrahydrofuran (THF) containing N,N,N',N'tetramethylethylenediamine (TMEDA) at room temperature for 2 h, the different chloropyridines were regioselectively deprotocuprated; their subsequent trapping with aroyl chlorides in general afforded the expected ketones. In the case of 2-chloro-4-iodopyridine, the arylcuprate proved unstable before the conditions required to perform its trapping (Table 1).  Yield after purification by column chromatography. The rest is in general starting material. b Trapping step performed at 60 °C. c Estimated 7% yield by performing the trapping step at rt; degradation at higher temperature.
Pd catalysis C-X arylation C-H arylation Ar Scheme 1. Possible synthesis of 4-azafluorenones substituted on their phenyl ring using palladiumcatalyzed double arylation.
In our previous study aimed at cyclizing 3-benzoyl-2-chloropyridines to 4-azafluorenones, 12 we employed a protocol inspired by conditions described for the cyclization of 2-chloro diaryl amines to carbazoles. 17 This protocol used catalytic amounts of Pd(OAc) 2 as transition metal source, electron-rich and bulky trialkyl phosphine Cy 3 P (Cy = cyclohexyl) as ligand, and K 2 CO 3 as base, in DMF at 130 °C.
Starting from 3-(4-bromobenzoyl)-2-chloro-6-(trifluoromethyl)pyridine (4) and phenylboronic acid, the expected azafluorenone 8a was formed in 51% yield (entry 8). This moderate yield, when compared with that obtained using 1-Br, is at least partly due to the competitive formation (estimated 17% yield) of the bis-coupled product, 2-phenyl-3-(4-phenylbenzoyl)-6-(trifluoromethyl)pyridine. The trifluoromethyl group could be responsible, by its electron-withdrawing effect, for the competitive cross-coupling observed on the pyridine ring. The formation of the product resulting from a double Suzuki coupling suggests a Suzuki cross-coupling reaction easier than the C-H arylation giving the azafluorenone.
Our first goal was to identify a new method to synthesize onychine (R = 1-Me), an alkaloid endowed with anticandidal activity, 1a as well as analogs, in order to evaluate them for their biological activity.
Nevertheless, modulating the reactants nature showed that the reaction result was dependent on the nature of the arylboronic acids. By using 4-methoxyphenylboronic acid, the expected 1-substituted 4azafluorenone 12b was also isolated, but in a lower 52% yield ( Figure 2). This result could be due to a less efficient cyclization, uncyclized 3-benzoyl-2-chloro-4-(4-methoxyphenyl)pyridine (Suzuki coupling product, 12b') being also isolated in 24% yield (entry 2). In the case of 4-aminophenylboronic acid, no 1-substituted 4-azafluorenone was detected by GC analysis of the crude, but only 2,4-bis(4aminophenyl)-3-benzoylpyridine (resulting from a double Suzuki coupling) and 4-(4-aminophenyl)-3benzoyl-2-chloropyridine (corresponding to the uncyclized product) (entry 3). These results show that the presence of an electron-donating group on the phenyl substituent at the 4 position of the pyridine ring is not suitable for a subsequent palladium-catalyzed cyclization. Unfortunately, employing 3pyridylboronic acid furnished a complex mixture in which the expected 1-substituted 4-azafluorenone was not identified (entry 4). An azafluorenone was neither detected with 2-thienylboronic acid, but the corresponding uncyclized Suzuki coupling product 12c', which was isolated in 49% yield (entry 5). In the case of 2-benzo [b]thienylboronic acid, a complex mixture was produced, but from which the expected 1-substituted 4-azafluorenone 12d could be isolated in 10% yield (entry 6). By turning to methylboronic acid, it proved possible to generate onychine (1-methyl-4-azafluorenone, 12e), an alkaloid endowed with anticandidal activity. 1a The moderate 52% yield obtained is due to the competitive formation of the bis-coupled product (about 11% yield) and to a difficult palladiumcatalyzed cyclization, as demonstrated by the detection by GC of 3-benzoyl-2-chloro-4-methylpyridine (about 7% yield) (entry 7). That the nature of the substituents has an impact on the course of the reaction was also evidenced by performing the reaction from 4-bromo-2-chloro-3-(3methoxybenzoyl)pyridine (6); in the presence of methylboronic acid, the reaction did not provide any more an azafluorenone, but the uncyclized Suzuki coupling product 13e' (entry 8). Attempts to cyclize 2-chloro-3-(3-methoxybenzoyl)-4-methylpyridine (13e') in a separate reaction under the same reaction conditions neither proved successful.
Similarly, 2-phenyl-3-azafluorenone (18) was synthesized from 4-benzoyl-2,5-dichloropyridine (19)   Our experimental findings show that Suzuki coupling precedes intramolecular arylation during azafluorenone formation. To rationalize the results obtained for these cyclizations, we have performed quantum chemical calculations. One can predict that oxidative addition is quite related to the partial positive charge on the carbon bearing the chlorine, as it is for nucleophilic substitution. We thus accounted for this positive charge by linking it to the corresponding 1 H NMR chemical shift of the dechlorinated substrate, as shown by Handy and Zhang. 21 As far as the C-H activation step is concerned, it notably depends on the corresponding CH acidity. Thus, in addition to the electrophilicity 13 measure of the carbon bearing chlorine, we also calculated the acidity at the cyclization site of the phenyl ring for the expected Suzuki coupling products using a homodesmic reaction protocol effective for azines 10c,22 (Table 4). Table 4. Calculated values of pK a (THF) for selected substrates, and 1 H NMR chemical shift for their dechlorinated analogues.

Entry
Substrate δ (C 2 , X = H) pK a (C 2' , C 6' , X = Cl) 14 Because of the strict steric requirements for cyclization, we only considered here the CH acidities of the suspicious sites, omitting the others. It is obvious that some of these compounds exist in form of several rotamers, and the data from Table 4 refer to the most stable ones (see Supplementary data). In general, these polycycles demonstrate a trend, which is present in phenyl rings (rather than pyridyl) to be in conjugation with a carbonyl moiety. Unfortunately, in our set of compounds, the reaction center is quite remote from the substituent that is changed. This leads to their effect attenuation, and quantities discussed (δ, pK a ) are varying in a narrow range. Nevertheless, their change is quite logical with the electron-withdrawing or -donating properties (in particular, see entries 7-10).
We have previously demonstrated that CH acidities 10c,22 are an essential factor in deprotometallation of azines including biaryls. In the tandem process under consideration, one can see that the pK a`s contribution is not so important, but the electron-withdrawing groups still favor the reaction (e.g. see Table 2, entries 2-5; Table 3, entries 1,2).
When compared with 3-benzoyl-2-chloropyridine (Table 4, entry 1), 12b the phenyl-and 2benzo [b]thienyl-substituted derivatives coming from 1-Br (entries 2 and 3), which have nearly the same partial positive charges at C2 and CH acidities at C2', are converted into the tricycles 7a and 7d in similar (good) yields. The reason why the furyl-substituted derivative coming from 1-Br (entry 4) only furnished the dechlorinated Suzuki coupling product 3-(4-(2-furyl)benzoyl)pyridine does not seem to be in relation with these quantities.
To rationalize the lower reactivity of the phenyl-substituted derivative coming from 2 (entry 5), one can advance a lower partial positive charge at C2 when estimated from the 1 H NMR chemical shift of the dechlorinated substrate. On the same basis, a high partial positive charge at C2 is evidenced from the phenyl-substituted derivative coming from 3-Br (entry 6). This could be at the origin of the formation of the bis-coupled product 2-phenyl-3-(2-phenylbenzoyl)pyridine. The latter appears as a more likely precursor of 10a' if we consider the low CH acidity at C2" of the calculated substrate (entry 6).
Compared with 3-benzoyl-2-chloropyridine (entry 1), the corresponding 4-phenyl (entry 7), 4-(2benzo [b]thienyl) (entry 8) and 4-methyl (entry 9) derivatives show lower partial positive charges at C2 when estimated from the 1 H NMR chemical shift of the dechlorinated substrates. In the case of 4phenyl (entry 7) and 4-methyl (entry 9), it is combined with lower CH acidities at C2' but with nearly no alteration of the cyclization yield. Thus, the complex mixture obtained from the 4-(2benzo [b]thienyl) (entry 8) has to be explained differently, and could result from possible coordination of palladium by the sulfur-containing ring. The reluctance to react of 2-chloro-3-(3-methoxybenzoyl)-4methylpyridine (entry 10) could result from both a sterically congested position at C2' and a lower CH acidity at C6'.

Biological evaluation
Because of structural similarity with natural azafluorenone antimicrobial agents such as onychine, 1a,2b the synthesized compounds 7b-d, 9a', 9b, 10a', 12a,b, 17 and 18 were screened for their antibacterial activity against Gram-negative (Escherichia coli) and Gram-positive (Staphylococcus aureus) bacteria, and for their antifungal activity against Candida albicans (Table 5). It was found that all the tested compounds have high potent antifungal activity against Candida albicans, more than the reference drug (nystatin). For the antibacterial activity, all the compounds except 9b and 18 have high to moderate effect against Gram-positive bacteria (Staphylococcus aureus) and, in case of Gram-negative bacteria (Escherichia coli), the compounds 9b, 12a,b and 18 exhibit high activities and the compounds 9a', 10a' and 17 moderate activities whereas the compounds 7b-d were found to have no effect.  23 The compound 18, bearing no cytotoxicity at 10 μM and a quite interesting antiplasmodial activity (4.8 μM -1.2 μg/mL), meets this criteria and could constitute an interesting antimalarial lead for further chemical modifications. Finally, a study has been carried out to investigate the cytotoxic potential of the derivatives 7b-d, 9b, and 12a,b (Table 7). The antiproliferative activity of the derivatives was determined using breast cancer cell line MCF-7, which is an invasive differentiated mammary epithelial breast cancer cell line used worldwide to screen and compare the antiproliferative activity of new molecules vs standard anticancer compounds. Among the molecules tested, the compounds 7c,d containing a thienyl ring showed a strong antiproliferative activity, higher to that exhibited by doxorubicin, whereas the compound 12b showed a similar activity.

Experimental Section
4.1. General. All reactions were performed in Schlenk tubes under an argon atmosphere. THF was distilled over sodium/benzophenone. DMF was dried over CaH 2 and distilled before use. Liquid chromatography separations were achieved on silica gel Merck-Geduran Si 60 (63-200 μm). Nuclear Magnetic Resonance spectra were acquired using Bruker AC-300 spectrometer (300 MHz and 75 MHz for 1 H and 13 C respectively). 1 H chemical shifts (δ) are given in ppm relative to the residual solvent peak, and 13 C chemical shifts relative to the central peak of the solvent signal. 24  X-ray Crystallography. The samples were studied with graphite monochromatized Mo-K radiation ( = 0.71073 Å). The X-ray diffraction data were collected using APEXII, Bruker-AXS diffractometer at T = 150(2) K. All structures were solved by direct methods using the SIR97 program, 25 and then refined with full-matrix least-square methods based on F 2 (SHELX-97) 26 with the aid of the WINGX program. 27 All non-hydrogen atoms were refined with anisotropic atomic displacement parameters. H atoms were finally included in their calculated positions. Molecular diagrams were generated by ORTEP-3 (version 2.02). 27 4.2. General procedure 1: Deprotonation using the lithium-copper base prepared from CuCl (1 equiv) and LiTMP (2 equiv) before trapping with an aroyl chloride. A stirred cooled (0 °C) solution of LiTMP prepared at 0 °C in THF (6 mL) from 2,2,6,6-tetramethylpiperidine (1.7 mL, 10 mmol) and BuLi (1.6 M hexanes solution, 10 mmol) was treated with TMEDA (0.77 mL, 5.0 mmol) and CuCl (495 mg, 5.0 mmol). The mixture was stirred for 15 min at 0 °C before introduction of the required substrate (5 mmol). After 2 h at rt, a solution of the required aroyl chloride (10 mmol) in THF (3 mL) was added. The mixture was stirred at rt or 60 °C overnight before addition of a 1M aqueous solution of NaOH (20 mL) and extraction with Et 2 O (2 x 20 mL). After washing the organic phase with an aqueous saturated solution of NH 4 Cl (10 mL) and drying over anhydrous Na 2 SO 4 , the solvent was evaporated under reduced pressure, and the product was isolated after purification by flash chromatography on silica gel (the eluent is given in the product description).

In vitro antimicrobial and antifungal assays.
Applying the agar plate diffusion technique, 28 the compounds were screened in vitro for their bactericidal activity against Gram-positive bacteria (Staphylococcus aureus) and Gram-negative bacteria (Escherichia Coli), and for their fungicidal activity against Candida albicans. In this method, a standard 5 mm diameter sterilized filter paper disc impregnated with the compound (0.3 mg/0.1 ml of DMF) was placed on an agar plate seeded with the test organism. The plates were incubated for 24 h at 37 °C for bacteria and 28 °C for fungi. The zone of inhibition of bacterial and fungi growth around the disc was observed.
4.6. In vitro antiplasmodial assay. Activity against P. falciparum chloroquine-sensitive 3D7 strains was assessed by following the procedure already described by Frederich et al. 29 The parasites were 24 obtained from Prof. Grellier (Museum d'Histoire Naturelle, Paris, France). Each alkaloid and extract was applied in a series of eight three-fold dilutions (final concentrations ranging from 0.09 to 200 µg/mL for an extract and from 0.02 to 50 µg/mL for a pure substance) on two rows of a 96-well microplate and were tested in duplicate (n = 2) or triplicate (n = 3). Parasite growth was estimated by determination of lactate dehydrogenase activity as described previously by Jonville et al. 30 Artemisinin (98%, Sigma-Aldrich) was used as positive control (IC 50 = 36 ±14 nM). All compounds were tested 3 times (n = 3).

4.7.
In vitro cytotoxicity. The compounds were tested against breast carcinoma cell line MCF-7. The method applied is similar to that reported by Skehan et al. 31 using 20 sulfo-rhodamine-B stain (SRB).
Cells were plated in 96-multiwell plate (104 cells/well) for 24 h before treatment with the test compound to allow attachment of cell to the wall of the plate. Different concentrations of the compound under test (0, 1.0, 2.5, 5.0, and 10 mg/mL) were added to the cell monolayer in triplicate wells in individual dose, and monolayer cells were incubated with the compounds for 48 h at 37 °C and in atmosphere of 5% CO 2 . After 48 h, cells were fixed, washed and stained with SRB stain, excess stain was washed with acetic acid, and attached stain was recovered with Tris-EDTA buffer. Color intensity was measured in an ELISA reader, and the relation between surviving fraction and drug concentration is plotted to get the survival curve of the tumor cell line after the specified compound and the IC 50 was calculated.

Theoretical calculations.
All the calculations were performed at DFT B3LYP level of theory.
The geometries were fully optimized using the 6-31G(d) basis set. In order to perform stationary point characterization and to calculate zero-point vibrational energies and thermal corrections, vibrational frequencies were calculated at the same level of theory. The single point energies were obtained using the 6-311+G(d,p) basis set and tight convergence criteria. 1 H NMR chemical shifts were calculated with Gauge-Independent Atomic Orbital (GIAO) method 32 using corresponding TMS shielding 25 calculated at the same level of theory as the reference. The solvent effect was simulated within the polarized continuum model (PCM) with the default parameters for THF.