Synthesis of 2-BMIDA Indoles via Heteroannulation: Applications in Drug Scaffold and Natural Product Synthesis.

A Pd-catalyzed heteroannulation approach for the synthesis of C2 borylated indoles is reported. The process allows access to highly functionalized 2-borylated indole scaffolds with complete control of regioselectivity. The utility of the process is demonstrated in the synthesis of borylated sulfa drugs and in the concise synthesis of the Aspidosperma alkaloid Goniomitine.

Azaheterocycles are prolific in agrochemicals, pharmaceuticals, and natural products. Among the variety of classes, indoles remain a template of enduring prominence.[1] The academic and industrial utility of this scaffold has inspired the development of numerous methodologies for its construction and functionalization.[2] Selective functionalization of the indole scaffold has been integral to the development of bioactive compounds (e.g., Scheme a), and strategies that allow selective and/or late-stage modification remain a target for methodological development.[3] On the basis of their wide scope of potential applications and familiarity of use, methods to install boron functional groups have been a particular target for development. These methods include classical strategies based on stoichiometric metalation and reaction with, for example, B(OMe)3,[4] and extend to contemporary approaches using C–H activation[3,5] and direct borylation with borenium cations (Scheme b).[6]

Scheme 1

Accessing Borylated Indoles

Cat. = catalyst, MIDA = N-methyliminodiacetoxy, Pin = pinacolato, TM = transition metal.

These methods rely on borylation of an established indole scaffold and are necessarily constrained by available functionality. Regioselectivity is a key consideration, and examples of these methodologies have demonstrated exquisite selectivity, with others exhibiting lower levels of regiocontrol.

Here we report an alternative approach to the regioselective synthesis of C2-borylated indoles. A Larock-type annulation[7−9] allows regioselective synthesis of functionalized 2-borylated indoles under mild conditions (Scheme c).[10,11] The process avoids the need for protecting groups on the indole nitrogen and avoids the restrictions imposed by using commercial indole scaffolds. The utility of this approach is demonstrated in the synthesis of drug scaffolds and alkaloid natural products.

Exploration of the annulation began with an initial survey of reaction conditions using 2-iodoaniline (1a) and propynyl BMIDA[12,13] (2a) as a benchmark system (Table ). Optimization provided a system that delivered 2-BMIDA-3-methylindole 3 in good yield (entry 1; for full details, see Supporting Information (SI)). These conditions were equivalent to more standard Larock-type conditions (entry 2); however, the chloride effect[7−9] could be replicated by the catalytic chloride available from the Pd catalyst, which delivered a small practical advantage. In the absence of chloride, reaction efficiency was ca. 20% lower (entry 3). The main issue that required navigation was compatibility of the reaction conditions with the BMIDA unit. For example, stronger bases led to MIDA hydrolysis[14] and subsequent protodeboronation[15−17] lowering the yield of 3 (entry 4).

Table 1

Reaction Developmenta

entry	components	deviation from “standard condtions”	yield (%),b product
1	1a/2a		84,c 3
2	1a/2a	Pd(OAc)2, LiCl (1 equiv)	83, 3
3	1a/2a	Pd(OAc)2	66, 3
4	1a/2a	replace NaOAc with K2CO3 or K3PO4	<20%, 3
5	1a/2b		16, 4
6	1b/2b	Pd(OAc)2 (10 mol %), LiCl (2 equiv), DMF, 65 °C	60,c 4

Reactions performed on 0.2 mmol scale.

Determined by 1H NMR using an internal standard as an average of 2 runs.

Isolated yield.

Moving from propynyl BMIDA 2a to phenylacetylenyl BMIDA 2b was less straightforward than expected. The optimal conditions for 2a delivered only 16% of 2-BMIDA-3-phenylindole (4) when using 2b (entry 5), and an independent optimization was necessary (see SI). Ultimately, this required the use of N-acyl 2-iodoaniline (1b) under the more classical Larock conditions for this aryl-substituted alkyne, giving 4 in good yield (entry 6). Acetate was the optimal N-protecting group (see SI). The origin of this difference in reactivity is uncertain, but the increased steric bulk of alkyne 2b is very likely to dominate,[7−9,18−20] with electronic effects also a minor contributor.[21,22] These conditions were subsequently assessed for generality across a series of annulations (Scheme ).

Scheme 2

Example Scope of the Annulation Process

Determined by 1H NMR assay using 1,4-dinitrobenzene as an internal standard.

Using 10 mol % Pd(dppf)Cl2.

Using Pd(OAc)2 (5 mol %), NaOAc (2.5 equiv), LiCl (2 equiv), DMF, 65 °C.

A series of alkyl alkynes were successfully accommodated to generate a small library of 2-BMIDA-3-alkyl indole products in good to excellent yield (Scheme a). The alkyl-substituted BMIDA alkyne progenitors were accessed via a metalation/borylation sequence using the requisite alkyl alkyne (see SI).[23] Compound 19 was delivered in low yields under the PdCl2(dppf) general conditions; however, this was found to improve when using the ligand-free conditions developed for the aryl/alkenyl alkynes. The origin of this subtle substrate divergence remains unclear.

Aryl- and alkenyl-substituted alkynes were also broadly compatible with the annulation, giving a similar series of products; however, a general lower efficiency was noted for aryl-substituted alkynes, consistent with previous observations with bulky alkynes in this area.[7−9,18−20]

The aryl-substituted BMIDA alkyne components can also be accessed via metalation/borylation or via a simpler Sonogashira coupling of the commercially available acetylene BMIDA (see SI).[24,25]

The reaction is completely regioselective. Regioselectivity was unequivocally established by X-ray crystallography (3 and 4, Scheme ) and NMR, showing that the BMIDA occupies the 2-position consistent with a larger steric footprint of this unit in comparison to the alkyl/aryl groups.[26]

A demonstration of the utility of the 2-BMIDA indole products is shown in Scheme . The sulfa drugs are a particularly important class of antibiotics.[27] The developed methodology enables the rapid, regioselective synthesis of the sulfa drug chemotype, included marketed compound 37(28) via annulation and subsequent Suzuki–Miyaura cross-coupling (Scheme a). Importantly, while a Larock approach to 38 could be envisaged via direct heteroannulation using the appropriate diaryl alkyne, steric issues lead to low yields for diaryl alkynes, and in addition, the subtle electronic differences lead to regioisomeric mixtures in these diaryl systems.[7−9,18−22]

Scheme 3

Utility of the Process in Drug and Natural Product Synthesis

Finally, this annulation/cross-coupling approach can be deployed to enable the modular synthesis of the Aspidosperma alkaloid goniomitine (Scheme b).[29] Annulation using alkyne 38 delivers indole 39 on a multigram scale. Cross-coupling with lactam fragment 40 provided 41, establishing the full carbon framework needed for the natural product. Hydrogenation, cyclization, and deprotection rapidly provided goniomitine (42). Importantly, intermediate 41 can be potentially diverted to other members of the Aspidosperma family following established approaches.[30]

In summary, a Larock-type annulation has been developed for the synthesis of 2-BMIDA indoles, allowing access to readily modifiable borylated heterocyclic scaffolds. The process accommodates a range of functionalized alkyne and aryl iodide coupling partners and delivers the products in good to excellent yield. The utility of the products has been highlighted in the rapid synthesis of drug and natural product scaffolds.[31]