Rahul
Sarkar
and
Santanu
Mukherjee
*
Department of Organic Chemistry, Indian Institute of Science, Bangalore 560 012, India. E-mail: sm@iisc.ac.in; Fax: +91-80-2360-0529; Tel: +91-80-2293-2850
First published on 15th January 2021
The first iridium-catalyzed enantioselective olefinic C(sp2)–H allylic alkylation is developed in cooperation with Lewis base catalysis. This reaction, catalyzed by cinchonidine and an in situ generated cyclometalated Ir(I)/phosphoramidite complex, makes use of the latent enolate character of an α,β-unsaturated carbonyl compound, namely coumalate ester, to introduce an allyl group at its α-position in a branched-selective manner in moderate to good yield with good to excellent enantioselectivities (up to 98:
2 er).
α,β-Unsaturated carbonyl compounds can be visualized as latent enolates and thereby nucleophilic at their α-position, especially under Lewis base (LB) activation (Scheme 1B).9 The reactions of such latent enolates with a wide range of electrophiles, even in enantioselective fashion, have been well documented and lead to an overall α-C(sp2)–H functionalization of α,β-unsaturated carbonyls.9 However, the combinations of these latent enolates with metal-activated electrophiles are rare,10 and an enantioselective variant is still unknown.
In 2003, Krische and co-workers reported an intramolecular α-C(sp2)–H allylic alkylation of α,β-unsaturated ketones through the merger of palladium catalysis and Lewis base activation (Scheme 1C).10c The same strategy was later on extended to an intermolecular variant of the same reaction by Huang and co-workers (Scheme 1C).10a Despite the exciting potential of these reactions, an enantioselective α-C(sp2)–H allylic alkylation of α,β-unsaturated carbonyl compounds under transition metal catalysis is yet to be accomplished.11
With our interest in catalytic asymmetric AA reactions6c,12 we sought to address this long-standing problem. Since C(sp2)–H AA reactions do not generate any stereocenter at the nucleophilic site, an enantioselective version of this reaction must necessarily arise from the prostereogenicity of the electrophilic partner. As iridium-catalyzed AAS reactions with unsymmetrical allylic electrophiles typically lead to the formation of branched products predominantly,3 we surmised that π-allyl-Ir would be a suitable electrophilic partner for the latent enolate generated from α,β-unsaturated carbonyl compounds.
While contemplating the choice of α,β-unsaturated carbonyl compounds, we were attracted to the report by Liu and Zu on the enantioselective cross-vinylogous Rauhut–Currier (RC) reaction13 of methyl coumalate with α,β-unsaturated aldehydes under iminium activation.14 Coumalates are an interesting class of electron-deficient heterocycles, which can act as synthetic precursors for various useful frameworks including electron-deficient arenes and heterocycles.15 Whereas coumalates are well-known for their use as electron-deficient dienes in various cycloaddition reactions,16 their applications as latent enolate are scarce. Besides Zu's report, the only other example of the usage of coumalate esters as latent enolate was recently documented in the form of a Morita–Baylis–Hillman (MBH) reaction by Thorimbert, Dechoux and co-worker.17
Coumalates (1) are known to be electrophilic at their C6 position18 and could suffer nucleophilic addition by a Lewis base (LB) to generate a dienolate A (Scheme 2).14,17 Although A is nucleophilic both at C3 (α-) and C5 (γ-), attack to the in situ generated π-allyl-Ir intermediate B was expected to take place from the sterically favored C3 of A. This enantiodetermining carbon–carbon bond formation would result in the formation of the intermediate C. Subsequent removal of the α-proton and elimination of LB from C would furnish the desired α-allylic alkylation product 3.
We envisioned that the enantioinduction in this C(sp2)–H AA reaction could be achieved either by using a chiral Lewis base (LB) or a chiral ligand (L*) on Ir(I) or both. Combining two different catalytic modes in cooperative fashion has emerged as a very effective strategy for reaction development during the past decade19 and has also been applied to Ir-catalyzed AA reactions with great effect.20
Entry | 1 | 2 | L | Lewis base | t [h] | 3 | Yieldb [%] | erc | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
a The catalyst was prepared via n-PrNH2 activation.
b Yields were determined by 1H NMR spectroscopy with mesitylene as internal standard. Isolated yields are given in the parentheses.
c Enantiomeric ratios (er) were determined by HPLC analysis on a chiral stationary phase.
d EtOH was used as the solvent.
e With 1.5![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1d | 1a | 2a | L1 | DABCO | 40 | 3aa | 11 | 88![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
2 | 1a | 2a | L1 | DABCO | 72 | 3aa | 10 | 92.5![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
3 | 1a | 2b | L1 | DABCO | 48 | 3aa | 16 | 95![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
4 | 1a | 2b | L1 | B1 | 48 | 3aa | 14 | 97![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
5 | 1a | 2c | L1 | B1 | 60 | 3aa | 32 | 97![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
6 | 1a | 2c | L1 | B2 | 72 | 3aa | 21 | 97![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
7 | 1a | 2c | L1 | B3 | 72 | 3aa | 42 | 98![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
8 | 1a | 2c | L1 | B4 | 72 | 3aa | 38 | 97![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
9 | 1a | 2c | L1 | B5 | 72 | 3aa | 25 | 97![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
10 | 1a | 2c | L1 | — | 48 | 3aa | <5 | — | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
11 | 1b | 2c | L1 | B3 | 72 | 3bc | 44 | 98![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
12 | 1b | 2c | L2 | B3 | 72 | 3bc | 42 | 98![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
13 | 1b | 2c | L3 | B3 | 72 | 3bc | 40 | 3![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
14e | 1b | 2c | L1 | B3 | 72 | 3bc | 51 | 97.5![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
15f | 1b | 2c | L1 | B3 | 72 | 3bc | 54 | 97![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
16g | 1b | 2c | L1 | B3 | 72 | 3bc | 53(51) | 97![]() ![]() |
Cinchona alkaloids and their derivatives as Lewis base are known to promote MBH and RC reactions.9,23 This fact inspired us to use quinine B1 as a Lewis base catalyst. As anticipated, B1 facilitated the reaction to generate 3aa with improved er (entry 4). We realized that the low yield of the reaction, among other reasons, is due to the decomposition of 2b to cinnamyl alcohol under the reaction conditions. With 2,2,2-trichloroethyl cinnamyl carbonate 2c as an allylic electrophile, this decomposition was suppressed, and the product was obtained in 32% yield with the same level of enantioselectivity (entry 5).
A Lewis base screening at this point revealed cinchonidine B3 as the optimal both in terms of the reaction efficiency and enantioselectivity (entries 6–9). Please note that the same sense of stereochemical outcome as B1 and B3 was observed with pseudoenantiomeric quinidine (B4) and cinchonine (B5), which shows that the phosphoramidite ligand on Ir(I) is primarily responsible for enantioinduction. However, the presence of Lewis base is essential as no product formation was observed in the absence of any Lewis base (entry 10), thereby indicating the cooperative interplay between iridium and Lewis base. The use of tert-butyl coumalate 1b instead of methyl coumalate 1a, along with B3, resulted in a much cleaner reaction, delivering the product 3bc in 44% yield (entry 11). Efforts to ameliorate the reaction outcome using other phosphoramidite ligands (L2 and L3) met with failure (entries 12–13). Increasing the amount of 1b turned out to be beneficial as excess of allylic electrophile (2) led to considerable side reactions. A 2:
1 ratio of 1b and 2c provided the best results, generating 3bc in 54% yield and with 97
:
3 er (entry 15). A 1
:
1 mixture of 2-methoxyethanol and THF as the solvent was proven to be equally effective as 1
:
1 EtOH/THF mixture and afforded the product 3bc with 51% isolated yield and 97
:
3 er (entry 16).
After optimizing the ligand, Lewis base and other parameters for the reaction between 1b and 2c (Table 1, entry 15), we chose to test the generality of our protocol for other substrate combinations. The Ir(I)/L1 catalyst system in cooperation with the Lewis base B3 was found to be rather general and catalyzes the enantioselective α-C(sp2)–H allylic alkylation of coumalates 1 with a large variety of allylic carbonates 2. As illustrated in Table 2A, tert-butyl coumalate 1b smoothly underwent allylic alkylation with cinnamyl carbonates (2c–o) bearing either electron-donating or electron-withdrawing substituent at various positions of the aryl ring. While the enantioselectivity of the reaction was found to be independent of the electronic nature of the substituents, highly electron-deficient aryl groups (e.g.2g–h) adversely affected the yield of the reaction. In the case of p-methoxyphenyl substituted allylic carbonate 2f, the product 3bf was obtained as an inseparable mixture with unreacted 1b.
However, switching the stoichiometry of 1b and 2f from 2:
1 to 1
:
1.2 facilitated the isolation of 3bf. For a particular substituent, very similar level of yield and enantioselectivity was observed, irrespective of its position (meta vs. para). Along the same line of observation, 3,4-dichlorocinnamyl carbonate 2o afforded the product 3bo with 97.5
:
2.5 er.
Apart from simple aryls, pharmaceutically relevant heterocycles can also be incorporated into the products. For example, dioxolane, furan and thiophene containing allylic carbonates (2p–r) reacted to give the corresponding products (3bp–br) with good to high enantioselectivity (Table 2B). In addition, alkenyl substituted allylic carbonates (2s–t) were well tolerated under the optimum reaction conditions to provide the products (3bs–bt) as a single regioisomer with high er (Table 2C).
The effect of the ester substituent of coumalate was next examined (Table 3). The bulky tert-butyl group can be replaced with simple methyl (1a), isopropyl (1c) and benzyl (1d) without affecting the enantioselectivity, even though the yield of these reactions remained relatively low compared to the tert-butyl coumalate 1b.
Entry | R (1) | t [h] | 3 | Yielda [%] | erb | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
a Isolated yield after chromatographic purification.
b Er was determined by HPLC analysis on a chiral stationary phase.
c Reaction in MeOCH2CH2OH/THF (1![]() ![]() ![]() ![]() ![]() ![]() |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1c | Me (1a) | 96 | 3aa | 37 | 97![]() ![]() |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
2c,d | t-Bu (1b) | 72 | 3bc | 51 | 97![]() ![]() |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
3 | i-Pr (1c) | 72 | 3cc | 44 | 97![]() ![]() |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
4 | Bn (1d) | 88 | 3dc | 40 | 97![]() ![]() |
The scalability of our α-C(sp2)–H allylic alkylation protocol was showcased by performing the reaction between 1b and 2c on a 1.0 mmol scale (Scheme 3A). Under the optimum reaction conditions, product 3bc was isolated in a slightly improved yield with the same level of enantiopurity as the smaller scale reaction.
![]() | ||
Scheme 3 (A) Scale-up of α-C(sp2)–H allylic alkylation and (B) synthetic elaborations of α-allyl-coumalates. |
Although further investigation is required, the generally low yield of these C(sp2)–H allylic alkylation reactions may be attributed to the possible oligomerization of coumalates under Lewis basic conditions. In addition, 2H-pyran-2-ones are known to undergo ring-opening and decarboxylation.24
The densely functionalized enantioenriched α-allylic coumalates can serve as synthetically important building blocks for electron-deficient arenes and heterocycles. In addition, the newly installed allyl unit can be transformed into other useful functionalities. For example, treatment of 3bc with benzylamine in MeOH furnished 3-allyl-substituted 2-pyridone derivative 4 in 67% yield (Scheme 3B). The regioselective Wacker oxidation25 of 3aa and 3bc provided the corresponding aldehydes 5 and 6, respectively. The absolute configuration of 5 was previously established by Zu et al.14 The stereochemistry of 6 and the other allylated products 3, shown in Tables 2 and 3 were inferred in analogy with 5.21
Ir-catalyzed hydroboration26 of the terminal double bond of 3bc resulted in the formation of alkyl boronate 7 in 63% yield. The 2H-pyran-2-one core of the coumalate esters is known to participate in [4+2]-cycloaddition reactions.27 The 2H-pyran-2-one moiety present in compound 3bc was found to undergo [4+2]-cycloaddition/decarboxylative retro-cycloaddition cascade reaction when treated with phenylacetylene to give 3-allyl tert-butyl benzoate 8 as a single regioisomer in 42% yield (Scheme 3B). Direct synthesis of 8 would require an enantioselective allylic alkylation of an electron-deficient arene and would be electronically disfavoured. This strategy, therefore, compliments the usual Friedel–Crafts type C(sp2)–H allylic alkylation, which are generally favoured with electron-rich arenes.7 Enantiopurity of 3 was preserved during all these synthetic elaborations.
Footnote |
† Electronic supplementary information (ESI) available: Experimental details, characterization and analytical data. See DOI: 10.1039/d0sc06208a |
This journal is © The Royal Society of Chemistry 2021 |