A Chemo- and Regioselective Tandem [3 + 2]Heteroannulation Strategy for Carbazole Synthesis: Combining Two Mechanistically Distinct Bond-Forming Processes

A modular approach to prepare tri- and tetracyclic carbazoles by a sequential [3 + 2]heteroannulation is described. First, optimization of Pd-catalyzed Buchwald–Hartwig amination followed by C/N-arylation in a one-pot process is established. Second, mechanistic analyses identified the origins of chemo- and regioselective sequential control of both bond-forming steps. Finally, the substrate scope is demonstrated by the preparation of a range of tri- and tetracyclic carbazoles, including expedient access to several natural products and anti-cancer agents.


■ INTRODUCTION
Carbazoles are ubiquitous N-heterocycles used throughout medicinal chemistry and material sciences. 1−3 From a pharmaceutical perspective, the carbazole core features extensively in drugs and natural products, many of which exhibit potent anti-proliferative activities (Figure 1a). 4−9 The breadth of applications has inspired the development of numerous methodologies for their preparation. 2,9,10 Pd-catalyzed pro-cesses in particular enable a [3 + 2]heteroannulation via Pd(0)catalyzed Buchwald−Hartwig amination 11 followed by Pd(II)catalyzed C-arylation at a late stage in a synthetic workflow. 12−25 However, a generalized set of guidelines that control the chemoselectivity of each C−N/C−C bond-forming reaction, and the factors that influence Pd(0) 23 versus Pd(II) 17 catalysis in a one-pot process have not been established. Furthermore, a mechanistic understanding of any regiocontrol underpinning the second C−H activation step for the formation of tetracyclic carbazoles is unknown.
In this manuscript, we establish reaction guidelines to prepare tri-and tetracyclic carbazoles by controlling the chemoselectivity of the first Pd(0)-catalyzed C−N bond-forming step and the regioselectivity of the second Pd(II)-catalyzed C/ N-arylation (Figure 1b). Our rationale was to use o-chloroanilines (1) to define the A-ring of a carbazole core and heteroaryl bromides to form the C/D-rings of tricyclic and tetracyclic products. An initial version of this work was deposited in ChemRxiv on the 24th of November 2021. 26

■ RESULTS AND DISCUSSION
The motivation for this work was to develop a robust, one-pot synthetic framework that can access both tri-and tetracyclic carbazoles. The synthesis of tetracyclic carbazoles using a heteroaryl bromide substrate such as 3 has potentially two competing C−H activation sites. If the C-arylation proceeded via a Pd(II)-catalyzed intermediate, then we surmised that the more nucleophilic 5-position of an isoquinoline will result in preferential C−H bond activation at this site. 27 However, the mechanistic determinants that control the regioselectivity of such Pd(II)-catalyzed C-arylation are not known with respect to a convergent [3 + 2]heteroannulation approach. 28−31 Our studies commenced with optimizing reaction conditions for the Pd(0)-catalyzed C−N bond-forming step using 4methoxy-2-chloroaniline (1a) and 6-bromoisoquinoline (6a) as exemplar substrates (Scheme 1). An extensive screen (Table S1) identified DavePhos and K 3 PO 4 as the optimal ligand/base pairing, which formed 7a in 87% yield.
With conditions for the Pd(0)-catalyzed C−N bond-forming step established, the optimization of the one-pot process was explored (Table S2). The optimal ligand/base pairing of HPCy 3 BF 4 with K 3 PO 4 was identified, which formed 9a from 1a and 6a in 82% yield (Scheme 2a). This highlights that the second C−H activation step is regioselective for the 7-position of 6a. To further understand how the nature of the chloroaniline and heteroaryl bromide substrates influenced the chemo-and regioselectivity of both bond-forming steps, a series of test reactions was undertaken. Exchanging 6-bromoisoquinoline (6a) for isoquinoline (10) formed dimethoxyphenazine (11) and the tertiary aniline (12) in 35 and 25% isolated yields, respectively (Scheme 2b). No dihydrophenazine was isolated from the reaction, which suggests that an in situ oxidation occurred. 33 No reaction occurred when isoquinoline (10) was the coupling partner with 1-chloro-3-methoxybenzene (13, Scheme 2c). Only secondary aniline (15) was isolated when para-anisidine (14) was reacted with 6-bromoisoquinoline (6a, Scheme 2d). Taken collectively, these test reactions highlighted the following requirements for the preparation of the tetracyclic core: (i) aryl bromide is essential and prevents dimerization of the o-chloroaniline, (ii) whilst the absence of a bromo substituent in the heteroaryl substrate results in C−H activation at the same site, there is little regiocontrol, (iii) a chloro substituent is essential for C-arylation.
The influence of the electron-donating aniline lone pair in the direct C-arylation step was then explored (Scheme 2e). We surmised that the acetylated substrate (16) would deactivate the C-ring and render the C-arylation less efficient. This indeed occurred as highlighted by the formation of deacetylated Carylated regioisomers, 9a and 17, in 27% total yield (1.2:1.0, 9a:17). We assume that deacetylation occurs in situ because of the high temperatures and the presence of a base in the reaction mixture. An unexpected result was the formation of the linear tetracyclic carbazole 17, which arises from C−H activation of the isoquinoline 7-position of 16. The formation of 17 suggests that C−H activation of the 7-position is favored if the acetyl group is present prior to C-arylation. In contrast, if 7a is present, presumably formed by deactylation of 16, C−H activation at the 5-position occurs. DFT calculations confirmed that the 5position is indeed the more electron-rich site (Scheme 2e and the Supporting Information Section 6). We speculate that the acetyl group (i.e., 16) directs C−H activation at this site via coordination of a Pd(II) species through the amide carbonyl. 34−36 This series of reactions has guided us to propose that oxidative addition of the C−Cl bond in 7a occurs first and proceeds via a Pd(0) species to form 18 (Scheme 2f). Pd(II)-catalyzed Carylation forms the palladacycle (20), which could proceed through a concerted metalation−deprotonation pathway via the formation of 19a. 12 Alternatively, since the efficiency of the reaction is higher with the lone pair of aniline (18) donating into the ring, a Friedel− Crafts-like electrophilic aromatic substitution mechanism proceeding via imine (19b) followed by tautomerization might also be possible to form 20. 37 Finally, reductive elimination of 20 produces the C-arylated product (9a). With mechanistic knowledge of the second C−C bond-forming step   ). The presence of a nitro group resulted in only trace amounts of 9d formed, with the secondary aniline (7e) isolated in 79% yield. The reaction conditions also tolerated heteroatom changes and saturation in the D-ring of the heteroaryl bromide (9e−j). 30 The [3 + 2]heteroannulation strategy was also compatible with the formation of carbazole-1,4-quinones (21a−f). 28,38 Access to N-arylated products is also possible, forming a mixture of fused imidazoles (9k−l) via an Narylation step, alongside tertiary anilines (22−23). Steric bulk ortho to the aniline substrate is also tolerated, forming 9m in 62%. However, when 3-methyl-2-chloroaniline is used as a substrate, a mixture of regioisomers was formed as an inseparable mixture (see Supporting Information).
The modularity of this strategy is also exemplified by the preparation of natural products glycozoline (9n), 39 harmane (9o), 40 and murrayafoline A (9p). 41,42 In addition, preparation of 9q demonstrates that the reaction conditions tolerate functional groups bearing potential Pd-chelating sites and bulky substituents ortho to the corresponding aryl bromide.
Our [3 + 2]heteroannulation strategy was extended to the targeted synthesis of biologically active tetracyclic carbazoles.
Carbazole-1,4-quinones (e.g., 21a−f) have established anticancer activity via topoisomerase inhibition or by the production of reactive oxygen species. 38,43 We used 21b as a key intermediate for the targeted synthesis of a deaza analogue of the natural product 9-methoxyellipticine (Scheme 4a), producing 26 in three steps and in an overall yield of 20%. 44 Further exemplification of our strategy was demonstrated by the preparation of alkylated 7H-pyridocarbazoles (e.g., 9a−e), which have well-established anti-cancer activity. 45,46 Previous preparative methods of this series of compounds have involved a six-step linear synthesis affording compound 9e in 28% overall yield. 47 Our two-step convergent strategy accessed the 7Hpyrido [4,3-c]carbazole core 9e in a single step (83%), followed by alkylation (30−32) to produce mono-N-alkylated examples (27−28) and the potent anti-cancer agent ditercalinium dichloride (29, Scheme 4b). 48

■ CONCLUSIONS
In summary, we have established a mechanistic framework for the preparation of fused tetracyclic carbazoles. The key to the modularity of this [3 + 2]heteroannulation approach is knowledge of the molecular hallmarks that underpin both the chemo-and regioselectivity of the process. The strategy is amenable for the diversification of tri-and tetracyclic carbazoles and is a scalable method for target-focused synthesis of tetracyclic carbazoles. We envisage that this convergent approach could find application in medicinal chemistry for structure−activity profiling and in broader synthetic applications that require step-efficient access to carbazole scaffolds.

■ EXPERIMENTAL SECTION
General Information. All reagents and solvents were obtained from commercial suppliers and were used without further purification unless otherwise stated. Purification was carried out according to standard laboratory methods. Starting materials were purchased from commercial suppliers and used without further purification unless otherwise stated. Dry solvents for reactions were purchased from Sigma-Aldrich and stored under nitrogen. Dichloromethane, chloroform, methanol, ethyl acetate, and petroleum ether (40−60°C) for purification purposes were used as obtained from suppliers, without further purification. Reactions were carried out using conventional glassware for the preparation of starting materials. Microwave reactions were carried out in capped 2−5 mL microwave vials purchased from Biotage. Microwave reactions were carried out at elevated temperatures using a Biotage Initiator+ equipped with a Robot Eight microwave system. Thin-layer chromatography was carried out using Merck silica plates coated with a fluorescent indicator UV254, and they were analyzed under both 254 and 375 nm UV light or developed using potassium permanganate solution. Normal phase flash chromatography was carried out using 60 Å, 40−63 μm silica gel from Fluorochem. Semi-preparative reversed-phase HPLC purification was carried out on a Kinetex 5 μm, 150 × 21.2 mm XB C18 using a DIONEX 3000 series HPLC system equipped with a VWD3400 variable wavelength detector. Preparative purifications of small molecules were performed using a 30−90% gradient B (solvent A: 0.1% TFA in water, solvent B: 0.1% TFA in acetonitrile), with a flow rate of 12.0 mL/min. The absorbance of the UV-active material was detected at 254 nm. Analytical reversed-phase HPLC (RP-HPLC) was carried out on a Shimadzu Prominence instrument equipped with a PDA detector scanning from 190 to 600 nm using a Thermofisher Hypersil GOLD column 100 × 4.6 mm, with a particle size of 5 μm. The Fourier transform infrared (FTIR) spectra were obtained on a Shimadzu IR Affinity-1 instrument. Only major absorbance bands are reported. The 1 H NMR, 13 C NMR, and 19 F NMR spectra were obtained on a Bruker AV 400 at 400, 101, and 376 MHz, respectively. Chemical shifts are reported in parts per million (ppm), and coupling constants are reported in hertz with DMSO-d 6 referenced at 2.50 ( 1 H) and 39.52 ppm ( 13 C) and MeOD-d 4 referenced at 3.31 ( 1 H) and 49.0 ppm ( 13 C). Assignment of 13 C NMR signals is based on HSQC and HMBC experiments. The COSY and NOESY spectra were used to assign unequivocally atom connectivities. The high-resolution mass spectra were recorded on a Bruker microTOF II mass spectrometer at the SIRCAMs facility at the University of Edinburgh or on an LTQ Orbitrap xL at the EPSRC National Facility in Swansea.
General Experimental Procedure for Microwave-Assisted Buchwald−Hartwig Amination. Aryl bromide (0.50 mmol, 1.00 equiv), chloroaniline (0.60 mmol, 1.2 equiv), Pd(OAc) 2 (5 mol %), DavePhos (10 mol %), and K 3 PO 4 (1.50 mmol, 3 equiv) were added to a microwave vial (2−5 mL). 1,4-Dioxane (5 mL, 0.1 M) was added and the vial was capped, evacuated and purged with argon three times, and then heated at 120°C for 30 min under microwave irradiation in a Biotage microwave. The reaction was allowed to cool to rt, diluted with ethyl acetate (50 mL), and the solid was filtered under vacuum. The organic phase was washed with water and brine, dried with Na 2 SO 4 , filtered, and the solvent was removed in vacuo. The crude compound was purified by silica column chromatography.