Expedient one-pot synthesis of indolo[3,2-c]isoquinolines via a base-promoted N–alkylation/tandem cyclization

Abstract

A transition metal-free, one-pot protocol has been developed for the synthesis of 11H-indolo[3,2-c]isoquinolin-5-amines via the atom economical annulation of ethyl (2-cyanophenyl)carbamates and 2-cyanobenzyl bromides. This method proceeds via sequential N–alkylation and base-promoted cyclization. Optimization data, substrate scope, mechanistic insights, and photoluminescence properties are discussed.

Graphical Abstract

1. Introduction

Indoles and quinolines are ubiquitous heterocyclic scaffolds appearing in numerous natural products and synthetic pharmaceuticals. The family of indoloquinoline alkaloids depicted in Figure 1 has been isolated from the root bark extracts of Cryptolepis sanguinolenta (Asclepiadaceae), a Ghanaian medicinal shrub native to West Africa.¹ Well-recognized as components in traditional folk remedies for malaria, these alkaloids display interesting pharmacology – for example, expressing hypotensive, antibacterial, anti-fungal, anti-plasmodial, antipyretic, and anti-inflammatory activities.² Further, recent studies have shown that cryptolepine is a potent cytotoxic agent, which binds to DNA in a base-stacking intercalation mode.³

Figure 1.

Bioactive examples of indoloquinoline alkaloids.

Because of their remarkable therapeutic potentials, the synthesis of indoloquinoline derivatives has attracted considerable attention from the synthetic community.⁴ In this report, we were particularly interested in 11H-indolo[3,2-c]isoquinoline, a close relative of cryptolepine and isocryptolepine. Our literature survey revealed only a few synthetic approaches to this unique motif, and further functionalized derivatives are rarely reported.⁵ These existing methods generally require multiple prefunctionalizations of requisite indole or isoquinoline starting materials (Scheme 1). For example, Black and co-workers demonstrated a flexible route to 1 via the acid-catalyzed cyclization of 3-amido-2-phenylindoles^5a and Timari presented a concise synthesis of 2 by thermal cyclization of 2-azidophenylisoquinoline.^5b Both are effective, but utilize moderately harsh conditions and involve multi-step precursor preparations.

Scheme 1.

Reported indolo[3,2-c]isoquinoline and benzofuro[3,2-c]isoquinolin-5-amine synthetic methods.

With the ascendency of organometallic catalysis in modern organic synthesis, metal-catalyzed cross-coupling C–N/C–C bond formations have become a powerful tool for constructing heterocyclic structures.⁶ In that context, the Maes group described an interesting assembly of 2 in two steps via a Buchwald–Hartwig amination followed by a Pd-catalyzed intramolecular arylation.^5c Importantly, the Kalugin group reported the transition metal-free two-step synthesis of benzofuro[3,2-c]isoquinolin-5-amine 3 from 2-cyanophenol.^5d Building on that work, we believed that an appropriate method could be developed for the facile construction of indolo[3,2-c]isoquinoline from readily available reagents.

Recently, our group has exploited the atom and step-economy of tandem/domino and multicomponent reactions to assemble complex heterocyclic skeletons.⁷ These strategies provide a rapid means to introduce molecular complexity by enabling multiple bond forming events to occur in one simple operation – thus avoiding the inconvenience of intermediate purifications.⁸ Herein, we report the base-promoted tandem annulation of ethyl (2-cyanophenyl)carbamate derivatives with 2-(bromomethyl)benzonitriles as a one-pot route to amine-functionalized indoloisoquinolines (Scheme 2a). To the best of our knowledge, this is the first report on the synthesis of 11H-indolo[3,2-c]isoquinolin-5-amines.

Scheme 2.

(a) Overall transformation; (b) initial approach; and (c) preparation of ethyl (2-cyanobenzyl)(2-cyanophenyl)carbamate.

2. Results and Discussion

We commenced initial screening by examining the feasibility of the reaction between 2-aminobenzonitrile and 2-(bromomethyl)benzonitrile in the presence of strong base (KO^tBu in DMF; Scheme 2b) – conditions adapted from the work of Li et al.⁹ Based on Kalugin's work,^5d we envisioned that S_N2 displacement would lead to intermediate 4, which would be subsequently deprotonated in situ by KO^tBu. The resultant benzylic anion would then attack the aniline ring nitrile and trigger the annulation process in a cascade fashion. Unfortunately, this N–alkylation/tandem cyclization furnished the expected tetracyclic core 5 in low yield and as a complicated reaction mixture. LCMS analysis revealed a significant amount of aniline remained post-reaction, suggesting that weak nucleophilicity of the 2-aminobenzonitrile results in an incomplete S_N2 reaction. Presumably the NH moiety of 4, exhibiting a lower pK_a than benzylic protons, might undergo deprotonation in this basic environment forming the amide anion, which could then attack electrophilic sites to form side products. In order to circumvent these hypothesized difficulties, we decided to modify the NH₂ moiety on the starting aniline with an appropriate protecting group (PG) – one easy to install, compatible with the desired transformations, and readily removed after cyclization. Among N–alkyl, N–acyl, and N–carbamoyl PGs, we found N–ethylcarbamoyl was the most satisfactory as treating ethyl (2-cyanophenyl)carbamate and the benzylic bromide with stoichiometric KO^tBu or NaH in DMF at 0 °C cleanly delivered 8a in quantitative yield within 1.5 h (Scheme 2c). With this result in hand, we turned our attention to an investigation of optimal conditions for the annulation using 8a as the model substrate (Table 1).

Table 1.

Optimization studies: base-promoted heterocyclization of 8a to indoloisoquinolines 9a^a

Entry	Base	Solvent	t (°C)	Time (h)	9a yieldb (%)	9a/9a’c
1	KOtBu	DMF	50	1	27	–
2	NaH	DMF	0	1	49	90:10
3	NaH	DMF	rt	24	26	20:80
4	NaH	DMF	50	0.5	70	90:10
5	NaH	THF	80	2	trace	–
6	NaH	MeCN	80	2	trace	–
7	LiHMDS	DMF	rt	24	20	–
8	K2CO3	DMF	80	2	NR	–
9	DBU	DMF	rt	24	trace	–

^aReactions were performed using 8a (0.5 mmol), base (0.55 mmol), and solvent (5.0 mL) at different temperatures and reaction time

^bIsolated yield.

^cDetermined by HPLC analysis of the crude reaction mixture.

DBU = 1,8-diazabicyclo[5.4.0]-undec-7-ene.

KO^tBu achieved the desired transformation, but afforded 9a in only 27% yield (entry 1). Examination of other bases revealed that NaH gave the best efficiency in terms of yield and clean reaction mixture. Cyclization was sluggish with LiHMDS and ineffective with K₂CO₃ and DBU, resulting in low conversion (entries 7–9). Among the solvents tested, DMF was superior to THF and MeCN (entries 5,6). Notably, stoichiometric NaH was essential for complete conversion. Also, prolonged stirring has a detrimental effect on the overall yield (entry 3) with 9a’, the N-deprotected indoloisoquinolines, being the major product. When the reaction was conducted with NaH/DMF at 50 °C for 30 min, the best conversion was achieved with minimal formation of 9a’ (entry 4); therefore, these conditions were found optimal. The structure of tetracyclic product 9a (Table 1) was unambiguously established by X-ray crystallographic analysis.¹⁰

This mild and efficient annulation (8a → 9a) encouraged us to explore combining the robust S_N2 reaction with the subsequent cyclization steps to give a one-pot cascade synthesis of 9. With effective sodium hydride mediation of both steps established (see Scheme 2 and Table 1), we set out to explore the scope and generality of this method for transforming appropriate carbamates 6 and bromides 7 into substituted 11H-indolo[3,2-c]isoquinolin-5-amines 9.

We first examined the effects of a variety of R¹ substituents on the one-pot process (Scheme 3). Ethyl (2-cyanophenyl)carbamates possessing either e-donating or e-withdrawing substituents are well-tolerated, affording the desired products in moderate to good yields (9a–h). Aryl halides (F / Cl / Br) are accommodated by the transformation and provide the possibility of further functionalization via various coupling reactions. It was found, in general, that e-rich anilines displayed retarded reactivity, resulting in lower yields than e-deficient substrates (cf., 9g in 45% vs 9d in 57%). A cyclic-β-enaminonitrile reacted smoothly to deliver 9i in 66%, and a pyridyl substrate was converted to 11H-pyrido[3',2':4,5]pyrrolo-[3,2-c]isoquinolin-5-amine (9n) in 68% yield. A limitation of this protocol was revealed with a thiophenyl substrate, which failed to undergo cyclization (→ 9o) resulting in recovery of the S_N2 intermediate. We next surveyed benzyl bromide R² substituents. The presence of an e-withdrawing fluorine substituent slightly lowers yields (9j-l) and, in the case of an e-donating methoxy, the S_N2 intermediate leads to 9m in only a trace amount.¹¹

Scheme 3.

One-pot substrate scope: synthesis of 11H-indolo[3,2-c]isoquinolin-5-amines. ^aSee SI for experimental procedures. ^bIsolated yield. ^cThree equivalents of NaH were used. ^dn.d. = not detected.

With various analogs of indoloisoquinolines 9 in hand, we set out to establish a practical N–deprotection method since most indoles require the free N–H to be biologically active. Fortunately, removal of the N–ethylcarbamate group can be achieved by simply treating a solution of 9 with LiOH at ambient temperature, which affords the free indolyl nitrogen in quantitative yield (Scheme 4). These conditions were found to be applicable to substrates with various substituents.

Scheme 4.

N–Deprotection of indoloisoquinolines 9.

A plausible mechanism for the base-promoted tandem cyclization is depicted in Scheme 5. Initial abstraction of a benzylic hydrogen by NaH generates benzylic anion intermediate A, which subsequently attacks the aniline-derived nitrile (→ B). Tautomerization furnishes 3-aminoindole C through which intramolecular nucleophilic attack onto the benzylic bromide-derived nitrile delivers D. Finally, tautomerization (→ E) and protonation deliver 9a.

Scheme 5.

Proposed mechanistic pathway.

Finally, we examined the photophysical properties of these novel indoloisoquinolines (UV−vis and fluorescence spectral data for 9a and 10a in DCM are illustrated in Figure 2). Compound 9a has a maximum absorption at 341 nm, which can be ascribed to an aromatic π–π* transition. As expected, removal of the carbamate moiety increases the availability of π-electrons on the indolyl nitrogen for easier conjugation with the isoquinoline, thus possibly explaining the bathochromic shift [9a (341 nm) → 10a (367 nm)]. Both compounds displayed strong fluorescence in the blue-green region (400-500 nm) and a red-shift trend was observed for fluorescence (9a → 10a ∥ 425 nm → 452 nm).

Figure 2.

Absorption (dashed lines) and emission (solid lines) spectra of 9a and 10a (c = 10 μM) in DCM (left). Fluorescence of 9a and 10a under UV illumination (365 nm) (right).

3. Conclusion

In conclusion, a transition metal-free and operationally simple one-pot protocol for the synthesis of 11H-indolo[3,2-c]isoquinolin-5-amines has been developed starting from readily available ethyl (2-cyanophenyl)carbamates and 2-cyanobenzyl bromides. This mild and efficient method proceeds via base-promoted N–alkylation and cyclization steps to forge novel indoloisoquinolines via a domino process that features atom and step economy as well as broad substrate scope. Photophysical studies demonstrate that this class of compounds affords potentially useful fluorophore properties.