Total Synthesis of (±)-Integerrines B–F via Bismuth-Catalyzed Oxidative Homo- and Cross-Couplings of 3‑Hydroxycarbazoles

Abstract

The total synthesis of (±)-integerrine B and the first total synthesis of (±)-integerrine C–F were realized, and 30–65% yields were achieved over 1–6 step(s) from the corresponding 3-hydroxycarbazole monomers. The key step involved the chemo- and regioselective bismuth-catalyzed oxidative homo- and cross-couplings of 3-hydroxycarbazoles with m-chloroperbenzoic acid. This strategy not only provides access to oxygen-sensitive biaryl alkaloids but also establishes a foundation for future asymmetric syntheses and broader applications in biaryl natural products.

In nature, 3-hydroxycarbazoles 1 and their (bis)­carbazoles 2 and 3, which are homo- and cross-coupled at the 4-position of 1, are widely distributed alkaloid compounds (Scheme ). − These compounds exhibit diverse biological activities, including cytotoxicity, antibacterial properties, and free radical-scavenging capabilities, and have therefore garnered significant research interest. Recently, (±)-integerrines A–E (2a, 2b, 3b–3d) and (R)-integerrine F (2c), which exhibited cytotoxicity, were isolated from Micromelum integerrimum.

1. Natural Products of 3-Hydroxy­(bis)­carbazoles.

Biscarbazoles 2 and 3 can be synthesized via the oxidative homocoupling (OHC) and oxidative cross-coupling (OCC) of 1. OHC and OCC reactions are among the most important reactions in organic synthesis and have gained significant attention in green chemistry owing to their atom-economical nature and efficiency. − In 2023, Knölker et al. reported a synthesis of C ₂-symmetric carbazole integerrine B (2a) via OHC at room temperature using an iron catalyst in an oxygen atmosphere (Scheme A). In addition, a similar natural product, sorazolon E2 (2d), was synthesized via iron and vanadium catalysis (Schemes A and B). Recently, Kozlowski reported the first total synthesis of C ₁-symmetric carbazole clausenawalline E (3a) via vanadium-catalyzed OCC of 1d and 2-hydroxycarbazole 4 (Scheme C). Although vanadium-catalyzed OCCs have been reported, − their application is hindered by the intrinsic toxicity of vanadium. In 2023, we reported bismuth-catalyzed OCC of 1 with 2-naphthols 5, affording C ₁-symmetric carbazoles 3′ (Scheme D). This method exhibited high chemo- and regioselectivity under mild conditions, comparable to metal-catalyzed reactions. Moreover, it has a significant advantage over reactions involving other metal catalysts because bismuth is a nontoxic main-group element, despite its location among toxic heavy metals. To our knowledge, however, neither bismuth-catalyzed OHC of 3-hydroxycarbazoles nor OCC between differently substituted 3-hydroxycarbazoles has been reported. In this study, we report novel bismuth-catalyzed OHC and OCC reactions for biscarbazoles and their applications to the total synthesis of (±)-integerrines B–F and sorazolon E2 (Scheme E).

2. Previous OHCs over [A] an Iron Catalyst and [B] a Vanadium Catalyst, OCCs over [C] a Vanadium Catalyst and [D] a Bismuth Catalyst, and [E] This work.

Initially, we investigated the OHC of unprotected monomer 1a with our previously reported Bi­(OTf)₃–O₂ system to synthesize 2a (Table , entries 1–2). However, under standard conditions, the reaction proceeded very slow and afforded a low yield (entry 1). Replacing MeCN with MeOH improved the reaction time; however, the yield remained low because of the lack of stability of 2a in an oxygen atmosphere (entry 2).

1. Screening Reaction Conditions .

Entry	Substrate	Bi cat.	Oxidant	Time	Yield
1	1a	Bi(OTf)3	O2 instead of N2	30 d	10%
2	1a	Bi(OTf)3	O2 instead of N2	3 h	19%
3 c	1e	Bi(OTf)3	O2 instead of N2	24 h	6%
4	1e	Bi(OTf)3	O2 instead of N2	24 h	38–58%
5	1e	Bi(NO3)3·5H2O	O2 instead of N2	24 h	9%
6	1e	Bi(OAc)3	O2 instead of N2	24 h	trace
7	1e	Bi(OTf)3	mCPBA	0.5 h	54%
8	1e	BiCl3	mCPBA	1.5 h	39%
9	1e	BiPh3	mCPBA	1.5 h	2%
10	1e	thiabismin-Ph	mCPBA	1.5 h	5%
11	1e	thiabismin-OTs	mCPBA	1.5 h	69–72%
12	1e	thiabismin-OTs	CHP	24 h	45%
13	1e	thiabismin-OTs	TBHP	24 h	19%
14	1e	thiabismin-OTs	H2O2 (5 equiv)	0.5 h	32%
			AcOH (2 equiv)
15	1e	none	mCPBA	1.5 h	n.d.

^aReaction conditions: 1e (0.2 mmol), Bi catalyst (5 mol %), oxidant (1.1 equiv), MeOH (2 mL), and N₂ at 30 °C for 1.5 h. mCPBA = meta-chloroperbenzoic acid. CHP = cumene hydroperoxide. TBHP = tert-butyl hydroperoxide.

^bYields were estimated by ¹H NMR spectroscopy using tetrachloroethane as an internal standard.

^cMeCN instead of MeOH.

^d1.5 equiv of mCPBA was used.

^eIsolated yield. n.d. = not detected.

Therefore, we proceeded to optimize the OHC using N-Bn-protected 6-Me-3-hydroxycarbazole (1e) to synthesize C ₂-symmetric biscarbazole 2e (entries 3–15). The Bi­(OTf)₃–O₂ system promoted regioselective coupling, but resulted in a low yield (6%) after 24 h (entry 3). Using MeOH as the solvent improved the yield (38%–58%). However, the same problem occurred in entry 2, where the decomposition of 2e occurred in an oxygen atmosphere, causing a lack of reproducibility (entry 4). Other bismuth catalysts did not lead to improved results (entries 5 and 6). To address this problem, we used m-chloroperbenzoic acid (mCPBA), a well-known oxidant for bismuth, and a controllable reagent with respect to equivalents, providing 2e in a moderate yield (54%) without the recovery of 1e (entry 7). Subsequent screening of bismuth catalysts indicated that the Lewis acidity or electron deficiency of the Bi atom was critical, with BiCl₃ and BiPh₃ exhibiting lower Lewis acidity than Bi­(OTf)₃, which resulted in reduced yields of 39% and 2%, respectively (entries 8 and 9). As other organo-bismuth catalysts, we focused on thiabismin derivatives. These have been used for stoichiometric arylation − and oxytosylation, as well as for catalytic cross-couplings of boronic aryls, − aryl silicon, and allylic C–H functionalization. Unfortunately, thiabismin-Ph was ineffective (entry 10). However, the use of thiabismin-OTs significantly improved the yield to 69%–72% (entry 11) with good reproducibility. To our knowledge, organo-bismuth-catalyzed OHC reactions have not been reported thus far; therefore, we performed further optimization (entries 12–14 and Table S1–S3). Although several solvents were investigated, none provided superior results to those achieved with MeOH. Other oxidants, such as cumene hydroperoxide (CHP), gave a moderate yield (45%) with a recovery of 1e (51%); however, mCPBA proved to be optimal. In the absence of a bismuth catalyst, the reaction scarcely proceeded, and most of the starting material remained unreacted (65%, entry 15).

With the optimized conditions established, we investigated the substrate scope and limitations of the bismuth-catalyzed OHC (Scheme ). Substrates bearing weak electron-withdrawing groups (methyl ester 1f and bromo 1g) and electron-donating groups (methoxy 1h, alkyl ether 1i, or aryl 1j) reacted efficiently to afford the corresponding biscarbazoles (2f–2j) in excellent yields (67%–80%). However, this reaction has certain limitations. The yield of unsubstituted biscarbazole 2k was modest (38%) due to decomposition of both the starting carbazole and the product, as observed by TLC analysis. Furthermore, the substrate containing a strong electron-withdrawing formyl group (1l) failed to react, and the acetal-protected compound (1m) was unsuccessful. Deprotection of the acetal group occurred during the reaction because of the Lewis acidity of thiabismin-OTs, a monocationic bismuth species that serves as a strong and soft Lewis acid. Sterically hindered substrate 1n was also unreactive. These results indicate that this reaction is sensitive to electronic and steric factors. Investigation of the N-substituent revealed that while an N-Me group 1o afforded a moderate yield (2o, 50%), N-aryl substituents 1p and 1q were well tolerated, providing 2p and 2q in high yields (69% and 70%, respectively).

3. Scope and Limitations of OHC .

^* Isolated yields are shown. ^a 1 (0.2 mmol), thiabismin-OTs (5 mol %), mCPBA (1.1 equiv), MeOH (0.1 M), 30 °C, N₂. ^b mCPBA (1.5 equiv). ^c 0 °C. n.d. = not detected.

The N-H substrate 1r, bearing a Weinreb amide at the 6-position, gave 2r in 70% yield at 0 °C. In this study, we developed an organo-bismuth-catalyzed OHC reaction for oxygen-sensitive biscarbazole compounds.

With the developed bismuth-catalyzed OHC established, we proceeded to the total synthesis of integerrines B, D, F and sorazolon E2 (Scheme ). Initially, integerrines B and F were synthesized via hydrogenolysis of their N-Bn-protected precursors (2e and 2f) in high yields (87% and 89%, respectively, Scheme A). However, from the perspective of atom economy, we attempted to use unprotected 1a and 1b under the optimized conditions. Notably, the direct OHC of 1a was successful and afforded 2a in a slightly higher yield (38%) than that reported in the literature (Scheme A, 35%). We also applied this method to synthesize 2c, obtaining it in 65% yield. In addition, we obtained a similar natural product, sorazolon E2 (2d), from sorazolon E (1c); however, the yield of 2d is lower than that reported previously (Scheme B). Although the yield could likely be improved through further optimization of conditions, this represents a future challenge for this methodology.

4. Total Synthesis of Integerrine B, D, and F and Sorazolon E2.

To access formyl group-containing target 2b, a Weinreb amide served as an aldehyde synthon. The Bicatalyzed OHC of 1r affords 2r in 70% yield (Scheme C). However, our initial approach involving the subsequent diisobutylaluminum hydride (DIBAL-H) reduction of 2r gave a complex mixture, and 60% of 2r was recovered because of its low solubility and over-reduction of 2r. To circumvent this problem, we implemented a protective strategy. The protection of the phenolic hydroxy groups of diester 2c as tert-butyldimethylsilyl (TBS) ethers, followed by DIBAL-H reduction and Dess–Martin periodinane (DMP) oxidation, smoothly yielded diformyl group-containing compound. Final deprotection by tetrabutylammonium fluoride (TBAF) with AcOH afforded 2b in an excellent overall yield of 63% over the five steps from the corresponding monomer.

Next, we proceeded to synthesize the C ₁-symmetric biscarbazoles, integerrine C (3c) and integerrine E (3d). To achieve this, we investigated OCC between two different 3-hydroxycarbazoles (Scheme ). Since both 3c and 3d contain a 6-formylcarbazole moiety, we first examined the reactions of the N-H derivative of 1l with 1a and 1b, respectively, under the OHC conditions. In both reactions, the desired cross-coupled products were not detected; instead, only decomposition of the starting materials was observed. Next, we conducted the reaction using esters 1b and 1f instead of the formyl compound. When 1b was treated with 1e using mCPBA, fortunately, the OCC proceeded with excellent regioselectivity to afford the desired cross-coupling product 3e in 42% yield, along with 41% of the homocoupled product 2e. Using a less reactive oxidant, CHP, surprisingly, gave 48% of 3e without the formation of 2e, although the reason remains unclear. The OCC between the 6-bromo derivative 1g and 1b afforded the cross-coupled biscarbazole 3f in 44% yield with CHP. In the presence of CHP, 1e also reacted with the N-Bn derivative 1f to give the cross-coupled product 3g; however, it was obtained in a low yield of 18% due to the diminished reactivity of the combination between 1e and 1f. Furthermore, 1e could also be used for OCC with 1g to produce 3h in 31% yield with mCPBA and 21% yield with CHP, respectively. The OCC between 1b and Weinreb amide 1r in the presence of CHP proceeded to afford the desired 3i in 46% yield.

5. Scope and Limitations of OCC .

^* Isolated yields are shown. ^a 1x (0.1 mmol), 1y (1.0 equiv), thiabismin-OTs (10 mol %), mCPBA (1.1 equiv), MeOH (0.1 M), 30 °C, N₂. ^b CHP (1.5 equiv). ^c CHP (2.0 equiv).

With biscarbazoles 3e and 3i in hand, the total synthesis of 3c and 3d was investigated (Scheme ). Application of our established five-step sequence (hydrogenolysis, TBS protection, reduction, oxidation, and deprotection) converted 3e into 3c in 30% overall yield from 1b. For the synthesis of 3d, TBS protection of 3i followed by DIBAL-H reduction afforded 3j (67%) and the over-reduced product 3k (30%). Notably, 3k was found to be stable and could be efficiently reoxidized to 3j in 92% yield via DMP oxidation. Finally, quantitative deprotection completed the total synthesis of 3d (41% overall yield over five steps from 1b).

6. Total Synthesis of Integerrine C and E.

Lastly, we performed several experiments to elucidate the reaction mechanism. In our previously reported Bi­(OTf)₃/O₂ system, we observed that the addition of radical scavengers, such as 3,5-di-tert-butyl-4-hydroxytoluene (BHT) or 2,2,6,6-tetramethylpiperidin-1-oxyl (TEMPO), significantly decreased the yield of the coupled product under the optimized conditions. These results indicated that the reaction with Bi­(OTf)₃/O₂ likely proceeds through a radical mechanism, which is consistent with the literature. In contrast, the addition of BHT had only a minor effect on the OHC of 1e with thiabismin-OTs and mCPBA, affording 2e in 56% yield (compared to 72% without BHT, entry 9, Table ) with 9% of 1e remaining (Scheme A). This result suggests that a two-electron mechanism may be more reasonable than a radical one. In the biaryl coupling reported by Ball et al. using thiabismin compounds, they clarified the generation of pentavalent thiabismin species in situ by oxidation of trivalent Bi with mCPBA. Furthermore, we found that treatment of 1e with stoichiometric pentavalent Bi reagents, such as BiPh₃(CO₃) or NaBiO₃, gave the coupling product 2e (Scheme B). These results suggest that the coupling reactions with mCPBA may proceed through pentavalent Bi intermediates formed in situ. Therefore, we next examined the reaction of thiabismin-OTs and 1.1 equiv of mCPBA in the absence of carbazole. We found that they formed a complex Bi-1 which was detected by HRMS-ESI (Scheme C). Based on these findings, we propose a plausible mechanism shown in Scheme . Thiabismin-OTs reacts with mCPBA to either Bi-1, followed by ligand exchange with 1 to form pentavalent Bi A. A second carbazole 1 couples with oxidized carbazole B to give the product, and trivalent bismuth C reacts with mCPBA again. In the OCC reaction, the more reactive (electron-rich) carbazole might react with Bi complex to form a similar intermediate of A which then couple with the less reactive one. In cases involving highly reactive carbazole and/or potent oxidants such as mCPBA, both oxidation and coupling steps would proceed rapidly. Therefore, the formation of homocoupled product is preferred. Additionally, competitive oxidative decomposition of both the substrates and products would also be accelerated. However, the mechanism remains elusive, and further investigation is needed.

7. Mechanistic Studies.

8. Plausible Reaction Mechanism.

In conclusion, we completed the total synthesis of (±)-integerrine B and (±)-sorazolon E2 and achieved the first total synthesis of (±)-integerrines C, D, E, and F. To this end, we developed mild oxidative homo- and cross-couplings of 3-hydroxycarbazoles to afford the chemo- and regioselective coupled C ₁- and C ₂-symmetric biscarbazoles using the bismuth catalyst with mCPBA. This method demonstrated over 15 examples of OHCs and OCCs, serving as an effective tool for accessing these oxygen-sensitive biscarbazole natural products. Further asymmetric syntheses of these alkaloids are currently ongoing in our laboratory.

The data underlying this study are available in the published article and its Supporting Information.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.orglett.6c01363.

• Detailed experimental procedures, characterization data of all of the new compounds and copies of NMR spectra of the products (PDF)

K.K. designed the project. K.K. and T.Y. wrote the manuscript. H.D. and K.U. performed all experiments. All authors analyzed the results and approved the final version of the manuscript.

The authors declare no competing financial interest.