Distal γ-C(sp³)−H Olefination of Ketone Derivatives and Free Carboxylic Acids

Abstract

We herein report the distal γ-C(sp³)–H olefination of ketone derivatives and free carboxylic acids. Fine tuning of our previously reported imino-acid directing group and use of the ligand combination of mono-N-protected amino acid (MPAA) ligands and electron-deficient 2-pyridone ligands were critical for the γ-C(sp³)–H olefination of ketone substrates. In addition, MPAA ligands were found to enable the γ-C(sp³)–H olefination of free carboxylic acids to form diverse 6-membered lactones. Besides alkyl carboxylic acids, benzylic C(sp³)–H also could be functionalized to form 3,4-dihydroisocoumarin structures in a single step from 2-methyl benzoic acid derivatives. The utility of these protocols was demonstrated in large scale reactions and diversifications of the γ-C(sp³)–H olefinated products.

Entry for the Table of Contents

Distal γ-C(sp³)–H olefination reactions of ketone derivatives and free carboxylic acids were realized. For ketone substrates, fine tuning of the directing group and applying a mono-N-protected amino acid (MPAA) and 2-pyridone ligand combination were critical to achieve high reactivity. MPAA ligands also enabled free carboxylic acid directed γ-C(sp³)–H olefination.

Introduction

Carbonyl groups are common functional motifs in agrochemicals, pharmaceuticals and natural products. In recent years, palladium catalyzed C–H activation reactions directed by binding motifs have been proven to be useful synthetic strategies. Although ligand development has enabled carbonyl directed Pd-catalyzed C(sp²)–H activation reactions,^[1] the corresponding palladation of C(sp³)–H bonds remains challenging due to the weakly coordinating nature of the carbonyl oxygen to palladium (Scheme 1A). Thus, the development of synthetically useful directing groups (DGs) has been a major focus for the palladium catalyzed C(sp³)–H activation of ketones. The stoichiometric palladation of oximes^[2] was successfully exploited in pioneering studies towards developing catalytic reactions,^[3] albeit with limited substrate scope and transformations (Scheme 1B).

Scheme 1.

Pd(II)-Catalyzed C(sp³)–H Activation of Ketones and Free Carboxylic Acids

Recently, we have designed L,X-type covalent and transient directing groups to enable a wide range of C(sp³)–H activation reactions with both ketones and aldehydes.^[4] In particular, a new aminooxy acetic acid directing group enabled C(sp³)–H iodination^[5a] and (hetero)arylation^[5b-c] of ketones with a broad substrate scope and superior reactivity. However, most transformations with this DG target the β-C(sp³)–H bond, with only one example of γ-C(sp³)–H (hetero)arylation.^[5c] Because γ-functionalized carbonyls cannot be accessed through traditional carbonyl reactivity, such as enolate chemistry or conjugate addition, the selective functionalization of γ-C(sp³)–H bonds could be extremely useful.

To address this synthetic need, we aimed to further develop the γ-C(sp³)–H activation of ketones using our new imino-acid directing group. The γ-C(sp³)–H activation^[4e,6] is relatively uncommon and Pd-catalyzed C(sp³)–H activation of ketones has thus far been limited to Pd(II)/Pd(IV) catalytic cycles. To expand the scope of transformations, it is necessary to unlock the reactivity of Pd(II)/Pd(0) catalysis for ketones. Especially, C(sp³)–H olefination^[6b,6e,7], a formal alkyl-Heck reaction, poses its unique importance as Heck type reactions of alkyl halides are usually suffered by premature β-H elimination and slow oxidative addition.^[8] However, we also anticipated that the imino-acid bidentate directing group might prove challenging for the desired γ-C(sp³)-H olefination reaction, since the high stability of the resultant bicyclic palladation intermediate might hinder the necessary olefin migratory insertion step.

Herein we report the palladium catalyzed olefination of unactivated γ-C(sp³)–H bonds of ketone substrates through oxime-based directing group modification and use of a mono-N-protected amino acid (MPAA) and electron-deficient 2-pyridone ligand combination. Building on the observed ligand effect in γ-C(sp³)–H olefination of ketone derivatives, we also developed γ-C(sp³)–H olefination reaction of free carboxylic acids using MPAA ligands. Without installation of exogenous directing groups, both aliphatic carboxylic acids and 2-methyl benzoic acid derivatives were olefinated (Scheme 1C-D).

Results and Discussion

Based on our previous success in the γ-C(sp³)–H arylation of ketone derivatives, γ-C(sp³)–H olefination was investigated with 4,4-dimethyl pentanone as a model substrate. Trace amount (less than 5%) of the desired product was observed with the optimal directing group (DG1) for γ-C(sp³)–H arylation. The yield was not further improved despite extensive screening of conditions with DG1. Therefore, we decided to tune the properties of the directing group. Gratifyingly, the yield increased to 22% when the electron deficient amide was introduced instead of the weakly coordinating carboxylic acid (DG2). Installation of an electron rich amide to the directing group afforded 2a-3 in 8% yield (DG3). Removal of the gem-dimethyl groups resulted in no product formation under the same conditions probably due to a diminished Thorpe-Ingold effect (DG4 and DG5).

Having identified DG2 as the optimal directing group, we turned our attention to exploring the ligand effects in γ-C(sp³)–H olefination. We first tried electron-deficient 2-pyridone ligand(L1) which was critical for the γ-C(sp³)–H arylation.^[5c] However, a noticeable increase in yield was not observed. The mono-N-protected amino acid (MPAA) Ac-Phe-OH (L2) ligand led to a dramatic increase in yield to 38%. Among various N-protecting groups, acetyl protection gave the best yield whereas carbamate based protecting groups resulted significant decrease in yields (L2-L4). Next, MPAA ligands bearing different side chains were also examined. Ac-Ala-OH (L5) provided the product in 45% yield. Further increasing the side chain sterics was detrimental to the reactivity (L6-L8). Ac-βAla-OH (L9) afforded a slightly lower yield than L5. Removing the methyl from L5 suppressed the reactivity (L10). The combination of a protected amino group and a free carboxylic acid group was essential for the higher reactivity for MPAAs in this reaction, since modification to either site led to no improvement in yields (L11-L13). Further screening of conditions revealed that the use of oxygen as a co-oxidant in addition to silver salts aided catalytic turnovers. Interestingly, adding the 2-pyridone ligand, which was not an effective ligand alone, in the presence of MPAA ligand promoted the reactivity synergistically. After screening of 2-pyridone ligands with various electronic properties, a ligand combination of L5 and L14 was found to be optimal for this reaction (See SI for detailed screening information).

The attempted β-C(sp³)–H olefination of substrate 1a’ (derived from pinacolone) under the same conditions afforded trace amount of product determined by ¹H NMR. Moreover, when 1a and 1a’ were reacted in the same pot, only 1a was functionalized in 70% yield whereas 1a’ provided trace amount of product (Scheme 2). These results were interesting considering that β-C(sp³)–H functionalization is typically favored over the γ-C(sp³)–H bond.^[9] This unusual preference for γ-C(sp³)–H functionalization is likely the result of the greater stability of the five-six bicyclic intermediate arising from 1a’ compared to the analogous six-six bicycle resulting from 1a; its greater stability thus renders it more inert to migratory insertion of the bound alkene. Similar trends favoring insertion to six-six palladacycles were observed for alkyne^[10a-b] and alkene^[10c] coupling partners when bidentate directing groups were applied.

Scheme 2.

Competition Experiment, conditions: 1a (0.05 mmol), 1a’ (0.05 mmol), Pd(OAc)₂ (10 mol%, 0.01 mmol), L5 (20 mol%, 0.02 mmol), L14 (20 mol%, 0.02 mmol), ethyl acrylate (3.0 equiv, 0.3 mmol), AgOAc (2.0 equiv, 0.2 mmol), Ag₂CO₃ (1.0 equiv, 0.1 mmol), HFIP (1 mL), 120 °C, 6 h, Combined yield of mono/di products determined by ¹H NMR analysis of the crude mixture using CH₂Br₂ as the internal standard. Yields were calculated based on the initial amount of substrates.

With the best conditions in hand, we next explored the scope of ketone derivatives. Oximes derived from methyl ketones with various alkyl chains were preferentially functionalized at the γ-position (2a-g). Substrates with linear alkyl chains afforded moderate to good yields (2a-d). Substrates with bulky β-groups such as isobutyl (2e) or isopropyl (2f) were also compatible. Substrates with cyclic substituents such as cycloheptyl (2h) and cyclohexyl (2i) provided the desired products. Functional groups such as tetrahydropyran ring (2j), methyl ester (2k) and protected piperidine (2l) were also tolerated. Aromatic substituents such as phenyl (2m) and 4-methoxyphenyl (2n) were compatible with the new conditions. In addition to methyl ketones, substrates bearing alkyl chains with different chain lengths (2o-q) were tested and exhibited similar yields to corresponding methyl ketone substrates (2h and 2j).

Next, we examined the scope of partner olefins. A range of acrylates with various alkyl chains (3a-3i) and olefins with electron withdrawing groups were compatible with the optimized conditions. The reaction was mono selective when diethyl phophonate was applied as the coupling partner (3j). Phenyl vinyl sulfone also gave the olefinated product albeit in low yield (3k). Complex acrylic acid derivatives could be used for this reaction (3l-3n). Acrylic esters derived from (−)-menthol (3l), tetrahydrogeraniol (3m) and cholesterol (3n) were also successfully employed in the reaction.

C(sp³)–H activation^[11] reactions of carboxylic acids employing various nitrogen-based auxiliaries have been extensively studied.^[12] The apparent disadvantages of installing and removing auxiliaries prompted us and others to use free carboxylates as directing groups through ligand development. A series of β-C(sp³)–H functionalization reactions of free carboxylic acids have been achieved including olefination,^[7f] oxidation,^[13a-b] arylation,^[13c-e] and cross-coupling.^[13f-g] However, it is rather challenging to perform γ-C(sp³)–H activation with weakly coordinating carboxylates because the formed 6-membered palladacycles are highly unstable. Indeed, only γ-C(sp³)–H arylations are reported for free carboxylic acids so far.^[14]

Given the fact that MPAA ligands were able to promote γ-C(sp³)–H olefination reactions of ketone substrates, we decided to revisit the γ-C(sp³)–H olefination of free carboxylic acids which previously required nitrogenous directing groups. With the model substrate 4,4-dimethyl butanoic acid (4a), the ligand effect on γ-C(sp³)–H olefination of free carboxylic acids was explored (Table 5). No product was observed without ligand. Gratifyingly, mono-N-protected amino acid ligand Ac-Phe-OH (L2) provided 55% yield of a cyclized lactone product generated via olefination followed by intramolecular conjugate addition of the carboxylate. Replacing the acetyl N-protecting group with different groups such as Boc or Fmoc was detrimental to the reactivity (L2-L4). L5, which was the best MPAA ligand for the olefination of ketone derivatives gave a lower yield compared to L2. Changing the side chain sterics or chain length decreased the reactivity (L7, L9 and L10). 2-Pyridone ligand (L1) was not as effective as MPAA ligands. Acetyl-protected aminoethyl phenyl thioether ligand (L15), the critical component for the successful β-C(sp³)–H olefination of free carboxylic acids, provided just 20% yield in γ-C(sp³)–H olefination. Other monodentate ligands previously used to promote C(sp³)–H olefination (L16 and L17)^[6b,7g] showed poor to almost no reactivity in this case. Fine tuning of the reaction conditions revealed that increasing the loading of silver carbonate to 2 equiv. is beneficial to the reactivity.

Table 5.

Ligand Evaluation for Free Carboxylic Acid γ-C(sp³)–H Olefination^[a,b]

^[a]Reaction conditions: 4a (0.10 mmol, 1.0 equiv), Benzyl acrylate (2.0 equiv), Pd(OAc)₂ (10 mol%), ligand (10 mol%), Na₂HPO₄·7H₂O (1.0 equiv), Ag₂CO₃ (1.0 equiv), HFIP (1.0 mL), 120 °C, under air, 18 h.

^[b]Yield determined by ¹H NMR analysis of crude mixture using CH₂Br₂ as internal standard.

^[c]2.0 equiv of Ag₂CO₃ instead of 1.0 equiv.

With the optimized conditions in hand, a series of free carboxylic acid substrates were tested for newly discovered reactivity (Table 6). Substrates bearing various alkyl chains were functionalized in moderate to good yields (5a-5d). A β-dialkyl substituted carboxylic acid was also compatible with this conditions (5e). Substrates containing methoxy or phenoxy group provided the desired products albeit in moderate yields (5f and 5g). Substrates with phenyl and 4-methoxyphenyl groups were also compatible substrates (5h and 5i).

Table 6.

Scope of Free Carboxylic Acids^[a,b]

^[a]Reaction conditions: 4 (0.10 mmol, 1.0 equiv), benzyl acrylate (2.0 equiv), Pd(OAc)₂ (10 mol%), L2 (10 mol%), Na₂HPO₄·7H₂O (1.0 equiv), Ag₂CO₃ (2.0 equiv), HFIP (1.0 mL), 120 °C, under air, 18 h.

^[b]Isolated yields.

^[c]Ethyl acrylate (2.0 equiv) was used instead of benzyl acrylate.

^[d]KOAc (1.0 equiv) was used instead of Na₂HPO₄·7H₂O (1.0 equiv).

To further demonstrate the generality of this free carboxylic acid directed olefination reaction, we tried a different class of substrate. To our delight, not only open chain aliphatic carboxylic acids, but also 2-methyl benzoic acid derivatives with benzylic γ-C(sp³)–H bonds could be functionalized under the conditions. Benzylic γ-C(sp³)–H substrates afforded cyclized 3,4-dihydroisocoumarin structures, which are commonly found in key intermediates for the synthesis of bioactive molecules.^[15] Further optimization revealed that simply changing the base to potassium acetate (KOAc) could increase the yield to 72% for 2,6-dimethyl benzoic acid substrate (5j). These new conditions for benzylic γ-C(sp³)–H were then applied to various 2-methylbenzoic acid derivatives. With the methyl group at the meta position, a lower yield of 58% was obtained (5k). Substrates containing electron-withdrawing groups (nitro, fluoro and trifluoromethyl) afforded the products in good to excellent yields (5l-5n). An electron-donating ortho-methoxy substituent was also tolerated (5o). 2,6-Dimethyl benzoic acids with different substituents were successfully converted into the corresponding 3,4-dihydroisocoumarin derivatives (5p and 5q). 2-Methyl-1-naphthoic acid underwent olefination in 40% yield (5r). Interestingly, a 5-bromo substituted substrate delivered a mixture of mono- and di- olefinated product (5s and 5s’).

A range of olefins were shown to be compatible with the new conditions (Table 7). Various acrylate derivatives afforded the desired δ-lactone products in good yields (6a-6c). Other α,β-unsaturated activated olefins including Weinreb amide derived acrylamide, methyl vinyl ketone and phenyl vinyl sulfone gave the desired product in moderate to good yields (6d-6f). Notably, five-membered lactone product was isolated as the major product when pentafluorostyrene was applied (6g). This product is probably generated through sequential γ-C(sp³)–H olefination followed by allylic C–H functionalization.

Table 7.

Scope of Olefin Partners for Free Carboxylic Acid^[a,b]

^[a]Reaction conditions: 4a (0.10 mmol, 1.0 equiv), olefin (2.0 equiv), Pd(OAc)₂ (10 mol%), L2 (10 mol%), Na₂HPO₄·7H₂O (1.0 equiv), Ag₂CO₃ (2.0 equiv), HFIP (1.0 mL), 120 °C, under air, 18 h.

^[b]Isolated yields.

^[c]KOAc (1.0 equiv) instead of Na₂HPO₄·7H₂O (1.0 equiv).

^[d]KHCO₃ (1.0 equiv) was used instead of Na₂HPO₄·7H₂O (1.0 equiv).

To showcase the synthetic utility of these reactions, the olefinations of the ketone substrate 1a and carboxylic acid substrate 4a were conducted on larger scales (2.0 mmol) and afforded 73% and 79% of isolated yields respectively confirming the robustness of these conditions. These olefinated products were successfully derivatized. The ketone auxiliary was easily removed by acidic hydrolysis followed by esterification.^[16] Hydrogenation of 7 using Pd/C catalyst enabled the reduction of the double bond (8a) in 98% yield. Moreover, the ketone carbonyl group was selectively reduced to secondary alcohol without affecting other functional groups of the molecule (e.g. ester, 8b). Treatment with 1 equiv. of LiHMDS base at −78°C selectively deprotonated the α-C–H of the methyl group and provided the cyclized product through an intramolecular Michael addition. (8c) The uncyclized product of the free carboxylic acid could be obtained when 5a was treated with 1.2 equiv. of LiHMDS (9a). Global reduction of 5a with LiAlH₄ afforded the 1,3,7-triol (9b). Selective hydrolysis of the benzyl ester was accomplished under acidic conditions and afforded δ-lactone with a carboxylic acid group which could be used as a handle for further functionalization of the product (9c) (Scheme 3).

Scheme 3.

Synthetic Applications: a) 1a (2.0 mmol, 1.0 equiv), ethyl acrylate (3.0 equiv), Pd(OAc)₂ (10 mol%), L5 (20 mol%), L14 (20 mol%), AgOAc (2.0 equiv), Ag₂CO₃ (1.0 equiv), HFIP (20.0 mL), 120 °C, oxygen atmosphere, 40 h. b) 4a (2.0 mmol, 1.0 equiv), benzyl acrylate (2.0 equiv), Pd(OAc)₂ (10 mol%), L2 (10 mol%), Na₂HPO₄•7H₂O (1.0 equiv), Ag₂CO₃ (2.0 equiv), HFIP (20.0 mL), 120 °C, under air, 24 h. c) H₂, Pd/C, MeOH, RT, 15 h. d) NaBH₄ (1.0 equiv), THF, 0 °C, 4 h. e) LiHMDS (1.0 equiv), THF, −78 °C, 3 h, f) LiHMDS (1.2 equiv), THF, −78 °C, 3 h. g) LiAlH₄ (4.0 equiv), THF, 0 °C to RT, 10 h. h) 6 M HCl (aq.), 80 °C, 12 h.

Conclusion

In summary, the distal γ-C(sp³)–H olefination reactions of ketone derivatives and free carboxylic acids were developed. For ketone substrates, using modified imino-amide L,X-type directing group and applying both MPAA ligand and 2-pyridone ligand were critical to achieve a high reactivity. MPAA ligands were found to be efficient ligands for promoting the free carboxylic acid directed γ-C(sp³)–H olefination reactions. Under similar conditions, both open chain alkyl carboxylic acids and 2-methyl benzoic acid derivatives were functionalized in good yields. Considering the difficulties inherent in the alkyl variant of Heck reaction, these methods provide new synthetic disconnections. Gram-scale reactions and derivatizations of the olefinated products show the potential applicability of these reactions.

Table 1.

Directing Group Evaluation for γ-C(sp³)–H Olefination of Ketone Derivatives^[a,b]

^[a]Reaction conditions: 1a-x (0.05 mmol, 1.0 equiv), ethyl acrylate (2.0 equiv), Pd(OAc)₂ (10 mol%), AgOAc (2.0 equiv), HFIP (0.5 mL), 110 °C, under air, 15 h.

^[b]Yield determined by ¹H NMR analysis of the crude mixture using CH₂Br₂ as the internal standard. HFIP = 1,1,1,3,3,3-Hexafluoro-2-propanol.

Table 2.

Ligand Evaluation for γ-C(sp³)–H Olefination of Compound 1a^[a,b]

^[a]Reaction conditions: 1a (0.05 mmol, 1.0 equiv), ethyl acrylate (2.0 equiv), Pd(OAc)₂ (10 mol%), ligand (20 mol%), AgOAc (2.0 equiv), HFIP (0.5 mL), 110 °C, under air, 15 h.

^[b]Combined yield of mono/di products determined by ¹H NMR analysis of the crude mixture using CH₂Br₂ as the internal standard. Product observed as E/Z mixture.

^[c]Two ligands were applied 20 mol% each, oxygen atmosphere.

^[d]AgOAc (2.0 equiv) and Ag₂CO₃ (1.0 equiv) were used instead of AgOAc (2.0 equiv)

Table 3.

Scope of Ketone Derivatives ^[a,b,c]

^[a]Reaction conditions: 1 (0.1 mmol, 1.0 equiv), ethyl acrylate (3.0 equiv), Pd(OAc)₂ (10 mol%), L5 (20 mol%), L14 (20 mol%), AgOAc (2.0 equiv), Ag₂CO₃ (1.0 equiv), HFIP (1.0 mL), 120 °C, oxygen atmosphere, 15 h.

^[b]Isolated yields; products isolated as E/Z mixture.

^[c]Numbers in parentheses indicate the ratio of mono:di.

^[d]L9 (20 mol%) instead of L5.

Table 4.

Scope of Olefin Partners^[a,b,c]

^[a]Reaction conditions: 1a (0.1 mmol, 1.0 equiv), olefin (3.0 equiv), Pd(OAc)₂ (10 mol%), L5 (20 mol%), L14 (20 mol%), AgOAc (2.0 equiv), Ag₂CO₃ (1.0 equiv), HFIP (1.0 mL), 120 °C, oxygen atmosphere, 15 h.

^[b]Combined isolated yields of mono and di products; products isolated as E/Z mixture.

^[c]Numbers in parentheses indicate the ratio of mono:di.