Mild and Practical Indole C2 Allylation by Allylboration of in situ Generated 3‐Chloroindolenines

Abstract

C2 allylation of indole derivatives is a challenging but important transformation given the biological relevance of the products. Herein we report a selective C2 allylation strategy that proceeds via allylboration of in situ‐generated 3‐chloroindolenines. The reaction is mild, practical, and compatible with a wide range of C3‐substituted indoles. As allylboronates are readily accessible from commercial precursors, various substituted allyl moieties can be introduced using the same protocol. To showcase the utility of this method we applied it to the synthesis of the natural product, tryprostatin B.

ALL(yl) IN(dole)!: Allylboration of in situ generated 3‐chloroindolenines allows the mild and operationally simple C2 allylation of 3‐substituted indoles. The reaction has a broad functional group tolerance and is compatible with various substituted allylboronates. The utility of this method is demonstrated by the total synthesis of tryprostatin B.

Introduction

Indoles have been an important target in organic synthesis ever since Baeyer reported the first synthesis of the parent heterocycle in 1866.1 In these early days of organic chemistry, many classical methods (e.g. the Fischer, Bischler, Reissert, and Madelung indole syntheses), were developed, several of which are still widely used today.2 The interest in this “privileged scaffold” originates in part from the numerous naturally occurring bioactive indole alkaloids (Figure 1).3

Figure 1.

Selected naturally occurring indole alkaloids bearing an allyl moiety.

Unfortunately, the state of the art in indole synthesis does not always allow construction of the indole ring system with the desired substitution pattern. Consequently, research in this area has shifted focus to selective functionalization of readily available indoles.4 While functionalization at the C3 and N1 positions is typically straightforward owing to their nucleophilic properties (in the latter case after deprotonation), and substituents at the benzenoid ring are mostly introduced during construction of the indole core, selective functionalization of the C2 position is more challenging. Radicals have been reported to react preferentially at the indole C2 position.5 Recently, Bach et al. reported a convenient C–H activation strategy that allows alkylation and arylation at the indole C2 position.6

During our studies on the reactivity of indole‐functionalized isocyanides,7 we became interested in the C2 allylation of indoles en route to natural product scaffolds. The importance of this transformation is underlined by the numerous examples in the literature.8 Most strategies involve either directed lithiation (either by lithium–halogen exchange or by deprotonation) of the C2 position or C–H activation by a directing group (Scheme 1).

Scheme 1.

C2 allylation strategies of indoles. (a) lithiation of the C2 position via either lithium–halogen exchange or deprotonation. (b) transition metal catalyzed C–H activation by use of a directing group. (c) two step chlorination of indoles followed by a allylation–rearomatization sequence.

Although these strategies have great potential, they generally require a directing group at N1 which needs to be removed afterwards. In our pursuit of a more general and practical procedure for this selective conversion, we came across a method developed by Danishefsky et al. involving nucleophilic additions on in situ‐generated 3‐chloroindolenines.9 After addition of a nucleophile to the imine, rearomatization by elimination of HCl gives back the indole. In our opinion, this method represents the most convenient strategy for the direct C2‐allylation of indoles to date. However, the authors did not show the generality of the reaction with respect to allylating reagents and C3‐substituted indoles. Moreover, the addition of nucleophiles (mostly toxic organotin reagents) required activation by BF₃ ·Et₂O, thus limiting the functional group tolerance. In this communication, we present a milder and more general procedure for allylboration of 3‐chloroindolenines.

Results and Discussion

In search of a mild and readily available nucleophilic allylation reagent, we arrived at allylboronic acid pinacol ester (allylBpin, 8a). AllylBpin and related reagents effectively react with aldehydes and imines, and several convenient ways to synthesize substituted allylBpin derivatives from allylic alcohols and halides, 1,3‐dienes, and allenes have been reported.10 The reaction of 3‐methylindole (5a) with NCS as the electrophilic chlorine reagent in the presence of triethylamine at 0 °C selectively generated 3‐chloroindolenine 6a in situ. After subsequent addition of 8a and stirring this crude mixture for 1 h, we conveniently obtained 2‐allylindole 7aa in 44 % yield (Table 1, entry 1). We then showed that no reaction occurs in the absence of either Et₃N or NCS (entries 2 and 3).11 Optimization of reagent stoichiometry (entries 4–7) allowed us to identify conditions affording 7aa in 75 % isolated yield. Finally, we found that our initial choice for CH₂Cl₂ as the solvent was ideal for this reaction.

Table 1.

Optimization of reaction conditions

Entry	Solvent	Et3N	NCS	8a	Yield
		[equiv]	[equiv]	[equiv]	[%]a
1	CH2Cl2	1.5	1.1	1.5	43
2	CH2Cl2	0.0	1.3	1.5	0
3	CH2Cl2	1.5	0.0	1.5	0
4	CH2Cl2	1.5	1.3	1.1	72
5	CH2Cl2	1.5	1.3	1.3	73
6	CH2Cl2	1.5	1.3	1.5	75b
7	THF	1.5	1.3	1.5	0c
8	Toluene	1.5	1.3	1.5	15 %
9	MeCN	1.5	1.3	1.5	12 %

^a) Yield determined by ¹H NMR analysis with 2,5‐dimethylfuran as internal standard.

^bIsolated yield on 1.0 mmol scale.

^cThe 3‐chloroindolenine intermediate was not formed.

With the optimized conditions in hand, we started to evaluate the scope with regard to indoles 5. We initially checked if unsubstituted indole would provide the desired allylated species. Unfortunately, the 3‐chloroindolenine intermediate rapidly rearomatized to form 3‐chloroindole. The reaction of 1,3‐dimethylindole only led to decomposition under the reaction conditions. We therefore limited our selection to 3‐substituted indoles (Scheme 2). As hypothesized, the mild reaction conditions tolerated a variety of functional groups. For example, the reactions of ester‐functionalized substrates 5b and 5c afforded the corresponding allylation products 7ba and 7ca in very good yield. In contrast, the reaction of methyl indole‐3‐carboxylate (5d) gave product 7da in lower efficiency. This can be rationalized by the instability of the 3‐chloroindolenine intermediate, which is destabilized by the electron‐withdrawing properties of the ester. On the other hand, cyclohexyl‐ and phenyl‐substituted products 7ea and 7fa were obtained in good yield. We were pleased to see that also N‐Boc‐tryptamine and Boc‐Trp‐OMe (5g and 5h) were well accepted. Finally, we successfully achieved double allylation with bis(indolyl)methane (5i) which was converted to bisallyl species 7ia in high yield (83 %). Monoallylation of 5i was also possible, however, this gave a statistical mixture of starting material, monoallylation and bisallylation products.

Scheme 2.

Scope of indoles 5. Reaction conditions: indole 5 (1 mmol), Et₃N (1.5 mmol), NCS (1.3 mmol) in CH₂Cl₂ (4 mL, 0.25 M) at 0 °C; then 8a (1.5 mmol) at r.t. [a] reagent stoichiometry adapted to bisindole 5i.

Encouraged by the high functional group tolerance of our method in comparison to other literature precedents for this transformation, in particular by the compatibility of Boc‐Trp‐OMe (to give 7ha), we wondered whether we could even use an N‐protected amino acid. This would be highly interesting for peptide chemistry, as the resulting C2‐allylated tryptophan could be readily incorporated in a peptide sequence by standard solid‐phase peptide synthesis. To our delight, the reaction of Fmoc‐Trp‐OH (5j) afforded 7ja in 76 % yield (Scheme 3).

Scheme 3.

C2 allylation of Fmoc‐Trp‐OH (5j).

Allylboronates are generally readily accessible from simple starting materials in one or two reaction steps. In addition, some simple allylboronates are commercially available. Moreover, they are non‐toxic and easy to handle (i.e., air and temperature stable). We sought to exploit these advantages by evaluating the compatibility of a set of readily available allylboronates with our reaction manifold (Scheme 4). The reaction of 5a with commercially available trans‐crotylboronic acid pinacol ester (8b) cleanly afforded reverse crotylation product 7ab in 69 % yield, nicely comparable to 7aa. Encouraged by this result, we then tested prenylBpin (8c), however, only traces of reverse prenylation product 7ac were obtained.12 Possibly, the sterically encumbered γ‐position is not sufficiently nucleophilic to attack the in situ‐generated 3‐chloroindolenine under these conditions. In contrast, C2 prenylation using 8d proved highly compatible with our method, as we could obtain product 7ad in 83 % yield. Unlike all other 2‐allylindoles, 7ad needed to be handled with care, as the product readily decomposed during chromatography if the silica gel was not pretreated with a base.13 Next, we could even demonstrate the possibility to introduce a propargyl substituent at the C2 position by using commercially available allenylBpin (8e).14 This significantly broadens the scope for post‐modification as alkynes are versatile reagents in robust chemistry (e.g. azide‐alkyne cycloaddition, Sonagashira coupling, Favorskii reaction, A3 coupling). We then successfully performed the allylation with 1,2‐diboryl reagent 8f which was conveniently prepared by diborylation of phenylallene. (β‐Boryl)cinnamyl indole 7af is again interesting for post‐modification as vinylboronates are suitable reactants for Suzuki cross‐coupling, Petasis reaction, or oxidation to the corresponding ketone.

Scheme 4.

Scope with regard to allylic boronates 8. Reaction conditions: indole 5a (1 mmol), Et₃N (1.5 mmol), NCS (1.3 mmol) in CH₂Cl₂ (4 mL, 0.25 M) at 0 °C; then 8a (1.5 mmol) at r.t. [a] 2.5 equiv of 8d was added. [b] reaction was performed on 0.2 mmol scale.

To demonstrate the versatility of the 2‐allyl moiety we performed some follow‐up transformations with 7aa (Scheme 5). As expected, catalytic hydrogenation of the alkene readily furnished 9 in 82 % yield. Hydroboration with 9‐BBN followed by H₂O₂ oxidation afforded anti‐Markovnikov hydration product 10, albeit in a rather modest yield. Finally, cross metathesis of 7aa and ethyl acrylate in the presence of Grubbs' 2^nd generation catalyst gave 11 in 42 % yield as a 3:1 E/Z mixture.

Scheme 5.

Chemical transformations of 7aa.

Having established the scope of this C2 allylation procedure of indoles, we wanted to demonstrate its utility in the synthesis of a natural product. We envisioned that we could efficiently access tryprostatin B (4) via this synthetic strategy. Danishefsky et al. already showed that their method could be applied to a five‐step synthesis of this natural product starting from N‐phthalamide protected 12 (Scheme 6).9 Our initial plan was to simply prenylate brevianamide F (16) to obtain 4 in a single step. Unfortunately, this route proved unsuccessful, not even producing a trace of 4. We hypothesized that the diketopiperazine was not stable under the reaction conditions, resulting in a mixture of unidentified products. However, since we had shown that Boc‐Trp‐OMe was well tolerated in the allylation reaction, we anticipated that Boc‐Pro‐Trp‐OMe (14) would be a suitable substrate for prenylation with boronate 8d. To our delight, dipeptide 14 was smoothly converted to prenylated product 15 in 62 % yield without losing optical purity. Then, Boc deprotection by treatment with TMSI followed by base‐mediated cyclization (NH₃/MeOH) as described previously by Danishefsky completed the total synthesis of tryprostatin B.

Scheme 6.

Total synthesis of tryprostatin B by Danishefsky and our method.

Conclusions

In conclusion, we report a mild and practical C2 allylation strategy of 3‐substituted indoles via allylboration of in situ‐generated 3‐chloroindolenines. The reaction is compatible with a range of functionalized indoles, providing the products in a short amount of time. We also demonstrated the compatibility with various allylic boronates, thus further expanding the range of accessible products. To show the potential of this method we synthesized tryprostatin B in a very efficient three‐step sequence starting from dipeptide 13. We believe that this method is the most efficient C2 allylation of 3‐substituted indoles in terms of scope and practical use, which can be employed as either an early or a late stage modification strategy to generate valuable building blocks.

Experimental Section

To a solution of an indole 5 (1.0 equiv) in CH₂Cl₂ (0.25 M) was added triethylamine (1.5 equiv) and N‐chlorosuccinimide (1.3 equiv) at 0 °C. After stirring for 15 minutes at this temperature allylboronate 8 (1.5 equiv) was added, followed by stirring for an additional hour at room temperature. The reaction was quenched by the addition of aq. NaOH solution (0.125 M) and stirred for an additional two hours. Then, the reaction mixture was extracted with EtOAc (3 ×), washed with brine, dried with Na₂SO₄, and concentrated in vacuo. The crude product was purified by silica gel column chromatography.

Supporting information

Supporting Information