The Catalytic Asymmetric Alkylation of Olefins with Alcohol Derivatives

Abstract

Both olefins and arenes are abundant and versatile feedstock chemicals. Although the electrophilic alkylation of arenes constitutes an advanced technology, the corresponding alkylation of olefins is rather undeveloped. Here, we report an asymmetric catalytic benzylation and allylation of unactivated olefins with acetates as electrophiles. The reaction is catalyzed by a single imidodiphosphorimidate organocatalyst operating as a silylium Lewis acid. Mechanistic experiments support an enantioconvergent S_N1 mechanism featuring stereoablative ion pair formation, resulting in a dynamic kinetic resolution. The enantioenriched hydrocarbon products are readily derivatized to odorants and nonsteroidal anti-inflammatory drugs in one or two steps. The developed method provides a metal-free alternative to cornerstone Tsuji–Trost-type allylations proceeding though π-allyl metal complexes.

Olefin alkylation reactions are of high relevance in terpene biosynthesis. The archetypical example is the evolutionarily conserved isoprenyl pyrophosphate synthase (IPPS)-catalyzed synthesis of geranyl pyrophosphate (Figure A). This reaction provides a building block for chain elongation toward polyprenyl pyrophosphates as well as cyclase-catalyzed reactions toward terpene natural products.

1.

Biocatalysis and small-molecule catalysis in olefin alkylation with alcohol derivatives. (A) Geranyl pyrophosphate biosynthesis. IPPS structure adapted from ref . Available under a CC-BY 4.0 license. Copyright 2011 Hsieh et al. (PDB ID: 3APZ, 2.60 Å resolution). , (B) Asymmetric olefin alkylation with small-molecule acid catalysts. (C) This work: realization of an asymmetric organocatalytic intermolecular olefin alkylation of unactivated olefins with acetates.

Small-molecule catalysis in olefin alkylation reactions remains underdeveloped (Figure B). Furthermore, metal-free catalytic asymmetric olefin alkylation reactions with alcohol derivatives are unprecedented. Unactivated olefins are challenging substrates in asymmetric transformations due to their weak nucleophilicity and lack of directing groups. Activated π-nucleophiles such as allylsilanes, − allylboranes, and silyl enol ethers have been used to overcome these hurdles, besides asymmetric Tsuji–Trost-type allylations. ,

In 2014, Carreira reported the first asymmetric allylation of olefins with alcohols in which iridium π-allyl complexes were leveraged to furnish 1,5-dienes with high regio- and enantiocontrol. Crucially, a strongly acidic sulfonimide cocatalyst was required for the activation of alcohol substrates bearing both an aryl- and vinyl substituent.

We envisioned our imidodiphosphorimidates (IDPis) as bifunctional catalysts to enable both substrate activation and enantioinduction. Confinement of the IDPi active site provides a well-defined stereochemical environment for enantioselective C–C bond formation through asymmetric counteranion-directed catalysis (ACDC). , Furthermore, our group recently reported on IDPi-catalyzed transformations of benzylic acetates. These developments motivated us to tackle the metal-free asymmetric alkylation of olefins with alcohol derivatives (Figure C).

By screening substrates and conditions (see Supporting Information), we identified a model reaction between 1a and 2a using IDPi 3a and N,O-bis­(trimethylsilyl)­trifluoroacetamide (BSTFA). Importantly, BSTFA acted as a silylium reservoir to deprotosilylate the IDPi, accessing a silylium Lewis acid regime. This suppressed olefin isomerization and provided a driving force for acetate activation by forming a strong Si–O bond.

Screening IDPi catalysts revealed the indispensability of electron-withdrawing 3,3′-“wings” (Figure A), as IDPi 3b favored elimination (see Supporting Information for further discussion of catalyst optimization). Tuning the wings and sulfonyl “core” (3c–3e) gave 3e as the optimal catalyst. Finally, decreasing the reaction temperature and concentration afforded 4a with 79% yield (75% isolated) and an e.r. of 96:4.

2.

Development of an asymmetric olefin alkylation. (A) Discovery and optimization. Reactions performed at 0.02 mmol scale. Yields determined by ¹H NMR with mesitylene internal standard. Yield in brackets: isolated yield at 0.2 mmol scale. (B) Arene scope and (C) olefin scope. Scope conditions: acetate 2 (0.2 mmol, 1.0 equiv), olefin 1 (8.0 equiv), IDPi 3e (2.5 mol %), BSTFA (2.5 equiv), CyMe (0.05 M), −60 °C, 3 d. Yields are isolated yields after column chromatography. Reactions performed at 0.2 mmol scale. ^a 3e (2.5 mol %), 1a (8.0 equiv), −60 °C, in CyMe (0.05 M), ^b–70 °C, and ^cthermal ellipsoids drawn at 50% probability level; two TFM molecules omitted for clarity. Color code: white, H; gray, C; green, Cl.

Arene substitution gave m-alkyl-substituted 4b–4d with good yields and excellent enantioselectivities (Figure B). p-Chloro-substitution similarly yielded 4e, which formed a multicomponent crystal with 1,3,5,7-tetrakis­(2-fluoro-4-methoxyphenyl)-adamantane (TFM). Using anomalous dispersion in crystal structure determination, we found that (S,S)-3e afforded (S)-product. 4f and 4g bearing m-methoxy substituents were obtained with comparable enantioselectivities, alongside minor separable spirocyclic byproducts. Lastly, 4h–4j were obtained with very good to excellent enantioselectivities.

Investigation of olefins demonstrated that nucleophilic olefins expectedly formed 5b–5c with comparable enantioselectivities (Figure C). , E/Z-mixture 5d demonstrated near-identical enantioselectivities between diastereomers, indicating stepwise C–C bond formation and deprotonation.

Gratifyingly, tetramethylethylene underwent alkylation despite its low nucleophilicity, yielding product 5e with excellent enantioselectivity. Furthermore, formal vinylation product 5f was accessed from 1f.

Reactions with 1g–1i produced products with excellent enantioselectivities. The formation of E/Z mixture 5g indicated preferential deprotonation toward the least sterically hindered position, whereas 5h–5i favored the thermodynamic products. Although no byproducts were observed, incomplete conversion led to a moderate yield of 5h, possibly due to increased steric hindrance on 1h. Separation of the regioisomers of 5i was achieved on silica gel impregnated with AgNO₃ by virtue of their distinct coordination to silver­(I).

Finally, trisubstituted olefins were investigated. 1j preferentially yielded 1,1-disubstitued anti-5j. Interestingly, insertion of an additional methylene unit in 1k resulted in regioselective deprotonation toward the 1,1-disubstituted alkene. We hypothesize that a gradual increase in steric demand from the methyl- and ethyl substituents in 5j and 5k compared to 5i prevents the cationic intermediates from entering the thermodynamic “Zaitsev” deprotonation pathway. Slight diastereomeric enrichment was observed for 5k compared to 5j, a trend which carried over to 5l.

A number of mixed scope entries were explored, followed by transformations toward biologically active compounds (Figure A). First, olefins 1b and 1e were benzylated to give 6a–6d, demonstrating compatibility with p-bromo- and terminal alkyne groups and chemoselectivity toward terminal olefins and allyl benzyl ethers.

3.

Application of the olefin benzylation reaction. (A) Mixed scope and product derivatization. Yields are isolated yields after column chromatography. Reactions performed at 0.2 mmol scale. (B) Alkylation of allylsilane in (S)-florhydral synthesis. (C) Isomeric convergence of 5i. (D) One-pot acetylation and benzylation. Conditions A: TfOH (2.0 equiv), HFIP, 25 °C, 10 min. Conditions B: FeCl₃·6H₂O (5.0 mol %), Ac₂O (3.0 equiv), propylene carbonate, 80 °C, 2 d. Conditions C: RuCl₃·H₂O (2.0 mol %), NaIO₄ (4.1 equiv), MeCN/CCl₄/H₂O, 25 °C, 1 h. Conditions D: O₃, CH₂Cl₂/MeOH, Ph₃P (1.1 equiv), −78→25 °C, 1 h. ^a–70 °C; ^b–85 °C; ^callyltrimethylsilane (8.0 equiv) instead of BSTFA and olefin.

Preparation of m-methylated 6e from 1e enabled the two-step synthesis of a novel isomer of known musk odorant phantolide: (S)-isophantolide. The preparation of phantolide analogues is highly relevant because aromatic musk derivatives often possess desirable odorant properties and lower bioaccumulation propensity. −

Cyclizing 6e according to a modified procedure by Fu yielded 7 with excellent regioselectivity. The TfOH-mediated reaction occurred with near-complete stereoretention, contrary to an earlier report of a similar AlCl₃-catalyzed cyclization. Subsequent Friedel–Crafts acylation using conditions from Cherif delivered (S)-isophantolide in 32% overall yield.

1f was applied to the synthesis of nonsteroidal anti-inflammatory drugs (NSAIDs). Benzylation products 6f–6g were transformed into (R)-flurbiprofen and (R)-ibuprofen with near-complete stereoretention according to a Djerassi–Rylander oxidation under Sharpless conditions reported by Bäckvall and others. ,

In lieu of α-olefins undergoing alkylation (see Supporting Information), allyltrimethylsilane was explored as a C₃-nucleophile to obtain 6h with excellent enantioselectivity (Figure B). Ozonolysis then gave (S)-florhydral, the more potent enantiomer of this synthetic muguet odorant.

We were eager to isomerically enrich olefinic mixture 5i. Screening a variety of catalysts (see Supporting Information), we found that subjecting 5i to catalytic p-TsOH significantly enriched the thermodynamic product without racemization or cyclization side-reactivity (Figure C).

During reaction optimization, the acetate provided a reactivity handle at low temperatures. We envisioned that telescoping acid-catalyzed acetylation and olefin alkylation in one pot would improve step economy. Consequently, we sought to employ the IDPi in both the Brønsted- and Lewis acidic regimes (Figure D). Screening of reagents and conditions (see Supporting Information) revealed efficient acetylation of 1-phenylethanol (8) using Ac₂O and 3e. At gram scale, acetylation was followed by olefin alkylation according to the optimized conditions, affording 4a with comparable yield and enantioselectivity.

Revisiting our inspiration of geranyl pyrophosphate biosynthesis, we explored olefin allylation. By regulating reaction temperature, we were able to achieve allylation of 1a with (E)-pent-3-en-2-yl acetate, affording 10a with moderate yield and an e.r. of 94.5:5.5 (Figure ). Encouraged by this result, we explored 3-substitution, affording products 10b–10d with superior yields and enantioselectivities. The reaction toward 10d produced a number of byproducts related to cationic cyclization, which were removed using AgNO₃-silica. Surprisingly, halogenated substrates were smoothly converted to 10e–10f. Multicomponent crystallization of 10f with TFM allowed us to unambiguously assign the (S)-configuration using single-crystal X-ray diffraction (SC-XRD).

4.

Olefin alkylation with allylic acetates. Yields are isolated yields after column chromatography. Reactions performed at 0.2 mmol scale. All products obtained with shown double bond geometry in ratios >20:1. ^a–90 °C, ^b–50 °C, and ^c–60 °C. ^dThermal ellipsoids drawn at 50% probability level; two TFM molecules and one 10f molecule omitted for clarity. Color code: white, H; gray, C; brown, Br.

Next, allylic acetates possessing nonequivalent electrophilic sites were explored for the regioselective allylation of 1a. Substrates 9g–9i were cleanly converted to 10g–10i with very good yields and enantioselectivities. Nucleophilic attack occurred almost exclusively at the least sterically hindered site, providing complementary selectivity to Carreira’s synthesis of branched 1,5-dienes.

Allylation with linear 9j yielded 10j with excellent regioselectivity, highlighting the confined catalyst’s ability to distinguish between methyl- and ethyl substituents. Moreover, reactions with allylic transposition isomers 9k–9k′ yielded near-identical results for 10k–10k′, suggesting an allylic cation intermediate. Methylene-extended acetates 9l–9m demonstrated high regioselectivities. However, additional conformational freedom in 10k–10m may cause steric clash with the catalyst, decreasing enantioselectivity.

Reaction monitoring was performed to probe the reaction mechanism (Figure A). At high conversion, a starkly decreased reaction rate was observed. Furthermore, 2a became significantly enantioenriched early on, plateauing at higher conversions. Variable time normalization analysis (VTNA) of reactions including TMSOAc (0.0–1.0 equiv) revealed that this byproduct acts as a competitive inhibitor with an estimated partial order of −0.5.

5.

Mechanistic investigations. Yields determined by ¹H NMR using mesitylene or 1,2-dichloroethane internal standard. (A) Reaction monitoring and TMSOAc inhibition. (B) Leaving group scrambling. (C) Intermolecular competition KIE. (D) Reactions with enantiopure starting materials. (E) Solid-state structure of 3e. Thermal ellipsoids drawn at 50% probability level. Disordered parts (CF₃-fluorine atoms) and apolar hydrogens omitted for clarity. Intramolecular H-bond highlighted in light blue. Color code: white, H; gray, C; light blue, N; red, O; light green, F; orange, P; yellow, S. (F) Proposed mechanism.

Competitive nucleophilicity toward the cationic intermediate could underlie inhibition by TMSOAc, causing reversible ion-pair formation. Leaving-group scrambling studies were performed to investigate this hypothesis (Figure B). 2a was subjected to reaction conditions lacking an olefin and equimolar TMSOAc-d ₃. Monitoring 2a/2a-d ₃ revealed convergence toward a 1:1 mixture, while decomposition remained minimal. These findings confirmed the reversibility of leaving group activation, clarifying byproduct inhibition. Additionally, leaving-group scrambling confirms that 2a undergoes in situ racemization, which becomes dominant at higher concentrations of TMSOAc, causing a dynamic kinetic resolution (DKR).

A large normal secondary kinetic isotope effect (KIE) of 1.30 corroborated significant sp³→sp² rehybridization (Figure C) and suggested that ionization could be rate-limiting in the early phases since there is no upstream sp³→sp² rehybridization and catalyst silylation is fast.

An S_N2 pathway was ruled out using enantiopure starting materials (Figure D), which afforded the same enantiomer of 4a with comparable enantioselectivities. Moreover, higher conversion for (R)-2a is incompatible with an S_N2 pathway, leading to (S)-4a considering Walden inversion. Recovered starting material was enriched for (S)-2a in all cases (see Supporting Information), substantiating in situ racemization.

We obtained the solid-state structure of IDPi 3e using SC-XRD (Figure E), which revealed that fluorine atoms F6 and F28 flank the acidic core of the catalyst. Nonclassical C–H···F interactions between these fluorine atoms and the substrate could explain the high enantioinduction observed during reaction optimization. Furthermore, a N2–O7 distance of 2.710 Å and a slightly elongated P1–N2 bond (1.653(3) Å) indicate an intramolecular hydrogen bond.

These mechanistic experiments allowed us to propose a mechanistic hypothesis (Figure F). Deprotosilylation generates fast-equilibrating silylium-IDPi diastereomers as the resting state and facilitates a self-drying cycle. , Acetate activation then furnishes an ion pair that can undergo forward reaction with the olefin in the enantiodetermining C–C bond-forming step, giving the product after deprotonation. The competing backward reaction with TMSOAc causes racemization of the acetate, prompting a DKR.

This study presents the first organocatalytic intermolecular asymmetric olefin alkylation reaction with alcohol derivatives. Key to the discovery of the method was the development of a highly acidic IDPi catalyst that allowed for the activation of unbiased acetate electrophiles. The use of the IDPi as a silylium Lewis acid proved crucial to suppress olefin isomerization, introduce a self-drying and repair cycle, and provide a driving force for electrophile activation by forming a strong Si–O bond. Using the developed method, a number of enantioenriched odorants and pharmaceuticals were accessed. Experimental mechanistic studies evidenced reversible ion pair formation and an enantioconvergent cationic DKR. Notably, this method encompasses the first organocatalytic regio- and enantioselective olefin allylation, providing a complementary approach to Tsuji–Trost-type allylations proceeding through π-allyl metal complexes.

Glossary

Abbreviations

ACDC

asymmetric counteranion-directed catalysis

BSTFA

N,O-bis­(trimethylsilyl)­trifluoroacetamide

CyMe

methylcyclohexane

DKR

dynamic kinetic resolution

HFIP

1,1,1,3,3,3-hexafluoropropan-2-ol

HOPP

pyrophosphoric acid

IDPi

imidodiphosphorimidate

IPPS

isoprenyl pyrophosphate synthase

KIE

kinetic isotope effect

NSAID

nonsteroidal anti-inflammatory drugs

p-TsOH

p-toluenesulfonic acid

SC-XRD

single-crystal X-ray diffraction

TMS

trimethylsilyl

TMSOAc

trimethylsilyl acetate

TFM

1,3,5,7-tetrakis­(2-fluoro-4-methoxyphenyl)-adamantane

TfOH

trifluoromethanesulfonic acid or triflic acid

VTNA

variable time normalization analysis

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.6c12914.

• Experimental procedures, characterization data for new compounds, detailed discussion of reaction optimization, limitations of the method, and mechanistic investigations (PDF)

Deutsche Forschungsgemeinschaft (DFG, German Research Foundation under Germany’s Excellence Strategy, EXC 2033-390677874-325 RESOLV); National Institutes of Health grant R01AB123456 (PV, GS); European Research Council (ERC, European Union’s Horizon 2020 research and innovation program “Early Stage Organocatalysis, ESO” Advanced Grant Agreement No. 101055472); German Research Foundation grant AB 1234/1-1. Open access funded by Max Planck Society.

The authors declare the following competing financial interest(s): Benjamin List is listed as an inventor on patent WO 2017/037141 filed by the Max-Planck-Institut für Kohlenforschung covering the IDPi catalyst class and its applications in asymmetric synthesis. The other authors declare no competing financial interest.