Harish K.
Potukuchi
,
Anatol P.
Spork
and
Timothy J.
Donohoe
*
Department of Chemistry, University of Oxford, Chemistry Research Laboratory, 12 Mansfield Road, Oxford OX1 3TA, UK. E-mail: timothy.donohoe@chem.ox.ac.uk
First published on 19th March 2015
Aromatic heterocycles are a very well represented motif in natural products and have found various applications in chemistry and material science, as well as being commonly found in pharmaceutical agents. Thus, new and efficient routes towards this class of compound are always desirable, particularly if they expand the scope of chemical methodology or facilitate more effective pathways to complex substitution patterns. This perspective covers recent developments in the de novo synthesis of aromatic heterocycles via palladium-catalysed α-arylation reactions of carbonyls, which is itself a powerful transformation that has undergone significant development in recent years.
Fig. 1 A selection of phosphine ligands and palladium-precatalysts used in the α-arylation reaction. |
However, despite its utility, the palladium-catalysed α-arylation reaction of carbonyls has perhaps not gained as widespread a recognition as other cross coupling reactions. In particular, we noticed that this reaction provides intermediates that would be useful in the synthesis of aromatic compounds via de novo routes. Synthetic sequences to produce aromatic compounds from non-aromatic precursors can provide valuable and complementary reactivity patterns to those obtained by starting from the parent arene itself.
Therefore, this perspective is intended to provide a specific overview of the de novo synthesis of aromatic heterocycles based on palladium-catalysed α-arylation reactions as a key transformation. Instead of supplying a broad introduction to the catalytic method, this discussion focuses on recent examples to highlight the general capability and the future potential of this approach in the synthesis of arenes.
In a related sequence involving imines rather than ketones, and while exploring the bidentate nature of the azaallylic anion, Barluenga and coworkers reported a cascade approach that involved a palladium-catalyzed imine α-arylation, followed by intramolecular C–N bond formation promoted by the same palladium catalyst.3 This was the first example of intermolecular imine arylation. Coupling of o-dihaloarenes 6 with imines 7 using a Pd(0) catalyst and the bulky, electron rich phosphine ligand X-Phos along with NaOtBu as base led to the formation of indoles 8 in a concise manner (Scheme 2).
In the case of differently substituted dihalides 6 (Scheme 2, X and Y), the regioselectivity of the process was governed by the relative oxidative addition of aryl halides to palladium complexes (I > Br > Cl).3b Thus, the initial imine α-arylation step determined the regioselectivity of the final product 8. The scope of this modular approach proved to be general, providing 2- and 2,3-disubstituted indoles with either aliphatic or aromatic substituents in the 1,2,3-positions of indole 8. However, the instability of N–H imines 7 did not allow the direct preparation of N–H indoles 8. After screening several protecting groups, imines derived from t-butylamine proved to be optimal for the indolization reaction yielding N-t-Bu indoles 8 in high yields. Deprotection of the t-Bu group was effected by using TFA or AlCl3 in refluxing dichloromethane. In order to overcome the relatively limited availability of aryl dihalides, the authors investigated the role of o-halosulfonates, which could be readily prepared from o-halophenols. When o-chlorotriflates 6 (X = OTf, Y = Cl) were employed, slow addition of the triflates was essential, presumably due to the sensitivity of the triflates to metal alkoxides. Additionally, an optimization of rate of addition for each individual substrate was required. Also, in certain cases, indoles 8 were obtained in low yields. However, the use of chlorononaflates 6 (X = ONf, Y = Cl), turned out to be advantageous providing a variety of structurally diverse indoles 8 in high yields.3b This methodology was then exemplified by a straightforward and regioselective synthesis of a 4,6-disubstituted indole, which was challenging to make by conventional methods.
During their studies on the palladium-catalyzed cyclization reactions of (2-iodoanilino) carbonyl compounds, Solé and coworkers observed the formation of indoles in several cases.4 Palladium-catalyzed intramolecular α-arylation of β-(2-iodoanilino)esters 9 resulted in the formation of indole-3-carboxylic acid derivatives 10 after column-chromatography. Presumably, the initially formed indolines were oxidised (by air?) to the corresponding indole derivatives (Scheme 3a). Note that the use of phenol additives in a polar solvent such as DMF afforded the indole products directly. In a similar manner, the Pd(0)-catalyzed α-arylation of β-(anilino) ketones/aldehydes/carboxamides 9 (i.e. variation of R2) resulted in formation of the respective indoles 10 in certain cases.
The same group also reported the synthesis of isoindole-1-carboxylic acid esters 12 from α-(2-iodobenzyl-amino) esters 11via a palladium-catalyzed cascade involving enolate-arylation and dehydrogenation of the initially formed isoindoline (Scheme 3b).4g In a related annulative approach amino-tethered 2- and 3-iodoindoles 13 were converted to pyrrolo[3,4-b]indoles 14 using enolate arylation (Scheme 3c).4h
Churruca et al. also reported a similar protocol for the synthesis of benzofurans,6 which were then converted into pentacyclic benzophenanthrofuran derivatives, through an intramolecular oxidative cyclization. The same group also developed a heterogeneous diarylbenzofuran synthesis by means of polymer anchored palladium catalyst FibreCat™ 1026 (not shown).
Willis and coworkers synthesised several benzofurans 23 and benzothiophenes 24via a palladium catalysed intramolecular enolate O-arylation and thio-enolate S-arylation sequence (Scheme 5).7 While Cs2CO3 was sufficient to achieve excellent yield of benzofurans 23 from bromoarenes 21, variation of the base was important in achieving optimal yields with chloroarenes 21. The enolate and thio-enolate starting materials 21 and 22 respectively, were in turn obtained by palladium-catalyzed α-arylation of ketones 19 and thio-ketones 20 with dihaloarenes 18. In an attempt to develop a one-pot cascade process, after screening of ligands, cyclohexanone was coupled with 2-bromoiodobenzene to yield cyclohexane-fused benzofuran 23 in 91% yield. Note that these conditions required some optimisation for individual substrates.
Burch and co-workers reported a one-pot synthesis of benzofurans 23 from o-bromophenols 25 and ketones 19 using Pd(OAc)2 and a binaphthyl phosphine ligand 1 (Scheme 6).8 The use of sodium tert-butoxide was essential as other bases did not promote the coupling reaction. Treatment of intermediate 26 with a 1:1 mixture of CH2Cl2–TFA cleanly afforded the benzofurans 23. The use of microwave irradiation shortened reaction times for the arylation reaction to 30 min without any significant change in isolated yields. The utility of this method was then demonstrated by the synthesis of eupomatenoid, a natural product, in three steps.
The substituents at the C3 and C4 positions of the isoquinoline products were limited by the availability of the requisite ketone coupling partners. In order to circumvent this limitation and to broaden the scope of viable ketones a C4 functionalization reaction was envisaged after the α-arylation reaction but prior to the deprotection/cyclisation/aromatisation sequence. The coupling product 30 from the α-arylation of a methyl ketone 3 with an aryl bromide 27 was a prime candidate for selective manipulation (Scheme 8).10 Since the arylated ketone (31) is more acidic than the starting ketone 3 at least 2 equivalents of base are required to guarantee full conversion of 3 and which effectively yields the enolate 30 as the initial product of α-arylation. The α-carbon atom of the resulting carbonyl compound which would eventually become C4 on the final isoquinoline can be effectively functionalised by reaction of various electrophiles E+ with enolate 30in situ.1l,11 This feature allowed incorporation of the enolate functionalization step in the one-pot protocol developed earlier, with the final conversion of intermediate 31 into isoquinoline 32.
Scheme 8 Modular synthesis via in situ functionalization; yields reported for both steps. aElectrophile E+. |
In general this α-arylation-based methodology was limited to ketones with only one enolisable α-position or at least a very strong preference for one α-carbon atom to be arylated over the other to ensure a regioselective product formation. However, the in situ enolate functionalization strategy also facilitates the regioselective formation of products directly from a methyl ketone, and these would not be accessible selectively by direct α-arylation of the corresponding more functionalised ketones (Scheme 8).
The scope of this approach was further broadened by introducing an additional aryl moiety at the carbon atom of the intermediate coupling product that would eventually become C4 on the isoquinoline. Thus, the intermediate enolate (30) generated by α-arylation of a methyl ketone 3 with an aryl bromide 27 was coupled in situ with a second aryl bromide ((Het)Ar–Br) without further addition of catalyst providing α,α heterodiarylated compounds 33 (Scheme 9). All three steps including the final deprotection/cyclisation/aromatisation sequence were again conducted in one-pot. Since no diarylation was observed in the coupling of ketone 3 with bromide 27 the second aryl bromide species ((Het)Ar–Br) was restricted to aromatic systems with less steric demand than the first. Despite this limitation both the first as well as the second α-arylation reaction tolerate a large variety of electron-deficient and electron-rich aryl bromides furnishing the corresponding C-4 arylated isoquinoline products 34 in good to excellent yields.
By employing nitrile compounds instead of ketone derivatives as coupling partners in the α-arylation it was also possible to generate 3-amino isoquinolines, thereby enabling the direct synthesis of products at a higher oxidation level (not shown).10
Since the α-arylation conditions actually yield an enolate as the primary product a supplementary one-pot functionalization could be accomplished by adding a suitable electrophile after the coupling reaction was complete, vide supra.10 This approach allowed for the introduction of an additional substituent at the carbon atom which will eventually become C13 on the final protoberberine skeleton. In order to prove the validity of this concept both the C13-unsubstituted as well as the C13-functionalised natural products palmatine and dehydrocorydaline were synthesized (Scheme 11).12 The α-arylation reaction between aryl bromide 27a and ketone 3b provided the desired coupling products 37 and 38 either lacking or including supplementary functionalization by the addition of MeI as an electrophile. While treatment of 37 with NH4Cl at elevated temperature facilitated direct conversion into palmatine, the corresponding reaction with 38 effected only acetal cleavage and aromatization but not pivaloate displacement. This finding was rationalized by invoking restricted rotation about the isoquinoline–aryl bond, induced by the additional methyl group and thus disfavouring the near planar conformation required for the cyclisation. The lack of reactivity was overcome by introduction of a more effective leaving group via pivaloyl ester removal and chloride formation to yield the desired dehydrocorydaline (Scheme 11).
Since the majority of synthetic approaches in this field of isoquinoline synthesis have relied on electrophilic aromatic substitution to form the heterocyclic ring, the α-arylation-based strategy enlarges the scope of accessible core structures by facilitating the formation of (carbocyclic) electron-poor derivatives. Both the modular character and the extended scope of the method were illustrated by the synthesis of the naturally occuring pseudocoptisine as well as an unnatural fluorinated analogue (not shown).12
This journal is © The Royal Society of Chemistry 2015 |