D.
Christopher Braddock
*,
James
Clarke
and
Henry S.
Rzepa
Department of Chemistry, Imperial College London, London, SW7 2AZ, UK. E-mail: c.braddock@imperial.ac.uk; Fax: +44 (0)2075945805; Tel: +44 (0)2075945772
First published on 23rd October 2013
DMDO epoxidation of bromoallenes gives directly α,β-unsaturated carboxylic acids under the reaction conditions. Calculated (ωB97XD/6-311G(d,p)/SCRF = acetone) potential energy surfaces and 2H- and 13C-labeling experiments are consistent with bromoallene oxide intermediates which spontaneously rearrange via a bromocyclopropanone in an intersecting bromoallene oxide – Favorskii manifold.
![]() | ||
Fig. 1 Metabolites 1–3 from Laurencia marilzae and proposed biogenesis via epoxidation events. |
While the epoxidation of allenes8,9 and vinyl bromides10 has been studied, the epoxidation of bromoallenes has not been reported.11 Herein, we report the hitherto unknown direct conversion of bromoallenes to α,β-unsaturated carboxylic acids via an initial epoxidation event and the presumed intermediacy of a bromoallene oxide. We also show by computational modeling and 2H- and 13C-labeling studies that the latter's spontaneous reorganization to an α,β-unsaturated carboxylic acid under the reaction conditions is consistent with a bromocyclopropanone intermediate in an intersecting allene oxide – Favorskii manifold.
Bromoallene 412 was selected as a suitable substrate for investigating epoxidation and was synthesized by a standard sequence from heptanal (ESI†).13 Much to our delight, epoxidation of bromoallene 4 using dimethyl dioxirane (DMDO), generated either in situ14 or as a solution (ESI†)15 (Scheme 1), gave a mixture of Z and E-α,β-unsaturated carboxylic acids 5 directly in low but reproducible yields (note §, ESI†). The low yields can be attributed to decomposition of DMDO16a under the reaction conditions to methyl radicals,16b and subsequent radical attack on either of the products or starting materials (note ¶, ESI†).
![]() | ||
Scheme 1 Epoxidation of bromoallene 4 using DMDO solution. |
Mechanistically, we invoke the following pathway for the formation of α,β-unsaturated carboxylic acids from DMDO mediated epoxidation of bromoallenes (Fig. 2). Initial epoxidation of the bromoallene would give bromoallene oxides of the type A and/or B (note ¥, ESI†). Spontaneous epoxide opening8cvia bromo oxyallyl cations C and D (note ††, ESI†) respectively converge on the same bromocyclopropanone E. This intermediate now intersects with the Favorskii rearrangement manifold of α,α- and α,α′-dibromoketones where the resulting bromocyclopropanones E are known to collapse after attack by water giving hydrate F to α,β-unsaturated carboxylic acids 5 (note **, ESI†).17,18 Evidently, there is sufficient water in the dioxirane solution to function as a nucleophile here (note ‡‡, ESI†). Interestingly, regardless of the initial site of epoxidation, this mechanism predicts that carbon atoms 1 and 2 in bromoallene 4 interchange positions in the α,β-unsaturated carboxylic acid products 5.
![]() | ||
Fig. 2 Mechanistic rationale for conversion of bromoallenes into α,β-unsaturated carboxylic acids, with the carbon atoms of the functional groups numbered 1–3 showing an interchange of carbon atoms 1 and 2 (see also interactive Fig. 2 in HTML version of this article). |
This mechanism can be subjected to scrutiny via density functional level (ωB97XD/6-311G(d,p)/SCRF = acetone)19 exploration of the potential energy surface (R = H, Me, presented as an interactive version of Fig. 2 (ref. 20) via a digital data repository21). Oxygen transfer from dimethyldioxirane to form both A and B (TS1) have thermally accessible free energy activation barriers (R = H, 26.8 for A, 27.3 for B; R = Me, 26.8 for A, 24.6 kcal mol−1 for B), followed by a second, lower energy dyotropic rearrangement (TS2) to give E. An intrinsic reaction coordinate (IRC) reveals that TS2 (R = H,Me) represents the concerted transformation of A or B to E, with C/D acting as “hidden intermediates” in the process.22 Such hidden intermediates can be potentially transformed to real ones by tuning the substituents, and in this instance changing R from H or Me to OMe is predicted to accomplish this by stabilization of C/D (see interactive Fig. 2). TS2 itself (R = Me) has some early character of C/D; the C–Br bond is calculated to initially contract in length due to a significant stabilising resonance contribution of Br lone pairs, from 1.924/1.896 Å (A and B respectively) via 1.840/1.885 (TS2), 1.856/1.868 (C/D acting as hidden intermediates) to 1.921/1.922 Å (E).23 Calculations having demonstrated the thermal accessibility of the epoxidation-bromocyclopropanone sequence, 2H- and 13C-labeling experiments were necessary to verify the overall reorganization (4 to A/B to E to F to 5, Fig. 2) of the carbon framework.24
Deuterated bromoallene (1-2H)-4 was prepared by addition of ethynylmagnesium bromide to heptanal, in situ deprotonation of the propargylic alkoxide with n-butyllithium and quenching with MeOH-d4 to give labeled propargylic alcohol (1-2H)-6 (Scheme 2). Subsequent alcohol trisylation25 gave (1-2H)-7, and SN2′ displacement of the trisylate with bromide under the action of LiCuBr2 (ref. 26) provided bromoallene (1-2H)-4 with 70% deuterium incorporation at the 1-position.†
![]() | ||
Scheme 2 Synthesis of deuterated bromoallene (1-2H)-4. |
13C-labeled bromoallene (1-13C)-4 was similarly targeted, commencing with silyl enol ether 8 formation27 from octanal (Scheme 3). Oxidation using mCPBA gave interrupted Rubottom28 adduct 9, which could be acetylated to give acetate 10. Desilylation using buffered TBAF29 revealed protected α-hydroxyaldehyde 11, which we planned to use in a Wittig reaction with a suitably 13C-labeled phosphorous ylid. To the best of our knowledge, there is only a single report30 using methyltriphenylphosphonium iodide to generate the Stork–Wittig reagent31 using an in situ deprotonation–iodination–deprotonation procedure which we adapted using 13C-labeled salt 12 – available from relatively inexpensive 99% atom 13C-labeled methyl iodide – to give vinyl iodides Z-(1-13C)-13, E-(1-13C)-13 and diiodide (1-13C)-14.32 Acetate deprotection as a mixture gave the corresponding alcohols Z-(1-13C)-15, E-(1-13C)-15 and (1-13C)-16 all with 99% 13C at the alkene terminus.†
![]() | ||
Scheme 3 Synthesis of 13C-labeled bromoallene (1-13C)-4. |
Dehydrohalogenation of Z- and E-iodides (1-13C)-15 in the presence of inseparable diiodide (1-13C)-16 with LDA gave propargylic alcohol (1-13C)-7 in good overall yield, with the unprecedented observation that LDA converts vinyl 1,1-diiodides into terminal alkynes also (note §§, ESI†). Interestingly, 4% of the alkyne product was found to be the 2-13C isotopomer (ESI†), implicating a 1,1-elimination reaction pathway for diiodide 16 and competitive alkyl group migration from a vinylidene intermediate (note ¶¶, ESI†). Alcohol (1-13C)-7 was then converted to the desired bromoallene (1-13C)-7 (as 4% of its 2-13C-isotopomer, ESI†) as previously described (cf., Scheme 2).
With (1-2H)-4 and (1-13C)-4 in hand, epoxidation with DMDO was conducted. For deuterated (1-2H)-4, after the reaction was conducted in the usual manner (cf., Scheme 1), E-(2-2H)-5 and Z-(2-2H)-5 were isolated each showing 65% deuteration at the α-position only (note ‡, ¥¥, ESI†). Evidently, this result is consistent with the proposed mechanism (cf., Fig. 2) (note †††, ESI†). More compellingly, epoxidation of bromoallene (1-13C)-4 gave (E-2-13C)-533 and (Z-2-13C)-5 (28% isolated yield) where carbon atoms 1 and 2 from the bromoallene have entirely interchanged positions, giving also 4% of each of the (E-1-13C)-5 and (Z-1-13C)-5 isotopomers (ESI†). The expected 1JCH coupling constants experienced by the α-vinyl protons of the major isotopomers are clearly apparent in their 1H NMR spectra (Fig. 3).
![]() | ||
Fig. 3 1H NMR spectra of (a) (E-2-13C)-5 and (b) (Z-2-13C)-5 displaying the expected 1JCH values for the α-vinyl protons. |
In conclusion we have established that the hitherto unknown direct conversion of bromoallenes to α,β-unsaturated carboxylic acids using DMDO is consistent with an initial epoxidation event (note ***, ESI†) followed by a spontaneous reorganization via a bromocyclopropanone, a mechanism supported by calculations, in an intersecting bromoallene oxide – Favorskii manifold. These experiments support the proposed biogenesis of α,β-unsaturated carboxylate 3 from bromoallene 2 by epoxidation (note ‡‡‡, ESI†).
We thank the EPSRC for DTG funding (to J. C.).
Footnote |
† Electronic supplementary information (ESI) available: Notes §, ¶, ¥, ††, **, ‡‡, §§, ¶¶, ¥¥, †††, ***, ‡‡‡; general experimental; experimental details and characterising data for compounds leading to bromoallenes 4 (including ESI Scheme S1 for the synthesis of bromoallenes 4), (1-2H)-4 and (1-13C-4) and epoxidation thereof leading to E- and Z-5, (E-2-2H)- and (Z-2-2H)-5, and (E-2-13C)- and (Z-2-13C)-5; Copies of 1H and 13C spectra for all compounds showing 2H and 13C isotopic shifts and coupling constants where appropriate; ESI references. See DOI: 10.1039/c3cc46720a |
This journal is © The Royal Society of Chemistry 2013 |