Letter Cite This: Org. Lett. 2017, 19, 6252-6255
pubs.acs.org/OrgLett
Construction of 6,10-syn- and -anti-2,5-Dioxabicyclo[2.2.1]heptane Skeletons via Oxonium Ion Formation/Fragmentation: Prediction of Structure of (E)‑Ocellenyne by NMR Calculation Daeyeon Jeong,† Te-ik Sohn,† Jong Yup Kim,† Gyudong Kim,† Deukjoon Kim,*,† and Robert S. Paton‡ †
College of Pharmacy, Seoul National University, Seoul 08826, Korea Chemical Research Laboratory, Department of Chemistry, University of Oxford, Mansfield Road, Oxford OX1 3TA, U.K.
‡
S Supporting Information *
ABSTRACT: A highly efficient and stereoselective route to potential synthetic intermediates for ocellenyne and related C15 acetogenin natural products with 6,10-syn- and 6,10-anti-2,5-dioxabicyclo[2.2.1]heptane core structures has been developed by means of an iterative biogenesis-inspired oxonium ion formation/fragmentation sequence. In accordance with chemical transformations, the most likely stereostructure for (E)-ocellenyne on the basis of GIAO 13C NMR calculations possesses a 6,10anti-2,5-dioxabicyclo[2.2.1]heptane core, as predicted from a plausible biosynthetic pathway, instead of the spectroscopically proposed 6,10-syn-2,5-dioxabicyclo[2.2.1]heptane skeleton.
constant analysis based on comparison with the data for the nonideal systems available at the time could have led to an erroneous assignment (see below). The relative stereochemistry between the bicyclic skeleton and the syn-12,13-dibromo moiety in the C10 side chain could not be determined, so 1a and 1a′ in Figure 1 were proposed as possible ocellenyne structures. The biogenesis of (E)- and (Z)-ocellenynes proposed by Suzuki et al.2 is depicted in Scheme 1, elaborated with a more detailed stereochemical interpretation.3 In this scheme, (12E)(R,R)-laurediol 2 would be converted into prelaurefucin (3) via two consecutive intramolecular bromoetherification reactions by the action of bromide and a bromoperoxidase. Transannular displacement of bromide in prelaurefucin by the ring oxygen atom would generate tricyclic oxonium ion 4, where opening by chloride at C7 or H2O at C10 would give notoryne (5) or laurefucin (6), respectively. Meanwhile, attack of tricyclic oxonium ion 4 at C13 by bromide would lead to ocellenyne (1b). It is noteworthy that the stereochemical relationship of the C6and C10-alkyl substituents in the biogenesis-based ocellenyne structure (1b) is anti, in contrast to the syn relationship in the spectroscopy-based ocellenyne structures (1a and 1a′) proposed by Scheuer and co-workers. In addition, the biogenetic pathway predicts that the 12,13 stereochemical relationship in 1b would be (12S,13S).
(E)- and (Z)-Ocellenynes (1) were isolated by Scheuer and coworkers from the sea hare Aplysia oculifera collected on a reef flat near Pupukea, Oahu, in 1981 (Figure 1).1 These C15 non-
Figure 1. Proposed ocellenyne structures.
terpenoid acetogenins possess a novel 2,5-dioxabicyclo[2.2.1]heptane skeleton, and the structures of the two geometrical isomers were elucidated by chemical degradation and spectral analysis. Formation of a (Z)-olefin in a zinc debromination of the corresponding hexahydroocellenyne, which requires a trans−anti conformation of the bromine atoms, suggested the relative stereochemistry of the 12,13-dibromo moiety as syn, namely, either (12S,13S) or (12R,13R). Both alkyl side chains were assigned the exo configuration from the size of the 1H NMR coupling constants between the bridgehead (C7 and C9) and vicinal (C6 and C10) protons, both of which were zero or nearly so, in comparison with those of the related carbocyclic bicyclo[2.2.1]heptane systems as well as that of a related dioxabicyclo[3.2.1]octane. However, the vicinal 1H−1H coupling © 2017 American Chemical Society
Received: October 16, 2017 Published: November 7, 2017 6252
DOI: 10.1021/acs.orglett.7b03226 Org. Lett. 2017, 19, 6252−6255
Letter
Organic Letters Scheme 1. Proposed Biosynthesis of Ocellenynes
Scheme 2. Retrosynthetic Plan for 6,10-syn- and 6,10-anti-2,5Dioxabicyclo[2.2.1]heptane Cores 7a and 7b
protocol6 to known oxocene amide 106a using 3-benzyloxypropylmagnesium bromide. Subsequent L-Selectride reduction4−7 of the resultant ketone 11 gave Felkin−Anh selectivity to provide the requisite (6R)-oxocene alcohol 9b in 74% overall yield (Scheme 3). A straightforward Mitsunobu inversion−saponifiScheme 3. Synthesis of Rearrangement Substrates 9a and 9b We have been involved in the development of an efficient synthetic method for generating the pivotal dioxatricyclic oxonium core in 4, which led to two organoselenium-mediated oxonium ion formation/fragmentation protocols, the second of which involves a SiO2-promoted fragmentation. Application of the former methodology culminated in a successful biomimetic asymmetric total synthesis of a series of 2,8-dioxabicyclo[5.2.1]decane derivatives (i.e., laurefucin, (E)-6) by regioselective fragmentation at C10.4 Meanwhile, the latter protocol with a SiO2-promoted fragmentation was applied to the total syntheses of natural products with a bis(2,2′)-tetrahydrofuran core (i.e., notoryne, (Z)-5) by C7 fragmentation.5 However, we have been unable to achieve the fragmentation of the tricyclic oxonium ion at C13, which would provide access to the remaining 2,5dioxabicyclo[2.2.1]heptane skeleton such as that found in ocellenyne. Herein we describe our realization of the desired remaining C13 fragmentation, which allows us optional access to any of the three dioxabicyclic structures in the biogenetic pathway. The preliminary discovery5b that C13-unsubstituted oxonium ions undergo preferential fragmentation at C13 seemed to present a fortuitous opportunity to steer the SiO2-promoted fragmentation to C13. Thus, we formulated a concise and highly stereoselective strategy for the construction of both 6,10-synand 6,10-anti-2,5-dioxabicyclo[2.2.1]heptanes 7a and 7b, respectively, via the selenium-mediated oxonium ion formation/SiO2-promoted fragmentation protocol,5 as shown in our retrosynthetic plan (Scheme 2). We envisioned that these targets 7a and 7b could be constructed from readily accessible (6S)- and (6R)-γ,δ-unsaturated oxocene alcohols 9a and 9b, respectively, via attack of chloride ion at C13 of oxonium ions 8a and 8b. Advantageously, these key intermediates possess two differentiated functional group handles for elaboration of C6 and C10 side-chain appendages. Our preparation of the rearrangement substrates 9a and 9b began with the application of our direct ketone synthesis
cation sequence8 on 9b furnished the corresponding (6S)oxocene alcohol 9a in 78% overall yield for the two steps, setting the stage for the pivotal organoselenium-mediated oxonium ion formation/SiO2-promoted fragmentation.5 With the key rearrangement substrates in hand, we first proceeded to address the construction of 6,10-anti-2,5dioxabicyclo[2.2.1]heptane 7b, a potential synthetic intermediate for ocellenynes predicted from biogenesis and related C15 acetogenin natural products,9 via the one-pot tandem organoselenium-mediated oxonium ion formation/SiO2-promoted fragmentation of (6R)-γ,δ-unsaturated oxocene alcohol 9b. Unfortunately, in contrast to the experience with α,α′disubstituted oxocene alcohols,5 the action of PhSeCl on oxocene alcohol 9b in the presence of activated silica gel was far less effective.10 Fortunately, we could overcome this unforeseen obstacle by developing an iterative oxonium ion formation/fragmentation sequence that was far more efficacious. To that end, the first iteration involved our biogenesis-inspired 6253
DOI: 10.1021/acs.orglett.7b03226 Org. Lett. 2017, 19, 6252−6255
Letter
Organic Letters hydroxyetherification protocol,4a in which subjection of oxocene alcohol 9b to organoselenium-mediated oxonium ion formation/ fragmentation afforded the desired hydroxyether 12b in excellent yield (96%), as shown Scheme 4.
irreversible SiO2-promoted fragmentation to provide the desired 6,10-anti-2,5-dioxabicyclo[2.2.1]heptane 7b and erythro-bis(2,2′)-THF 14b via nucleophilic attack by chloride at C13 and C7 in oxonium ion 8b, respectively.13 With an efficient synthesis of 6,10-anti-isomer 7b in hand, we next turned our attention to the construction of 6,10-syn-2,5dioxabicyclo[2.2.1]heptane 7a, which could serve as an intermediate for the synthesis of ocellenynes, with the structure predicted from spectroscopy. To this end, 7a was synthesized from (6S)-γ,δ-unsaturated oxocene alcohol 9a in comparable yield and selectivity by a route parallel to that employed for (6R)γ,δ-unsaturated oxocene alcohol 9b, as depicted in Scheme 5. It is
Scheme 4. Synthesis of 6,10-anti-2,5Dioxabicyclo[2.2.1]heptane 7b via Iterative Oxonium Ion Formation/Fragmentation
Scheme 5. Synthesis of 6,10-syn-2,5Dioxabicyclo[2.2.1]heptane 7a
interesting to note that a small amount (7%) of bis(tetrahydrofuranyl)tetrahydropyran 14a′ was also produced in addition to erythro-bis(2,2′)-THF 14a (3%) by intramolecular attack at C7 of oxonium ion 8a by the benzyl ether functionality. Even as a minor isomer, the formation of 14a′ strongly supports the exo configuration of the C6 side chain. Considering that the spectroscopic techniques available at the time (1981) were relatively unsophisticatednotably, the 2D NOESY technique was not availablewe embarked on predicting the stereostructure of (E)-ocellenyne by comparison of the 13C NMR chemical shifts of the natural (E)-ocellenyne with the Boltzmann-weighted GIAO 13C NMR chemical shifts calculated using DFT methods.5c,14,15 First, the performance of the calculations against a pair of known 6,10-syn- and 6,10-anti-2,5-dioxabicyclo[2.2.1]heptanes 7a and 7b (modeling the protecting groups as Me groups) was evaluated. The biggest deviation between theory and experiment was 1.19 ppm, while the root-mean-square deviation was 0.68/ 0.50 ppm depending on the epimer (see the Supporting Information (SI)). This was a useful benchmark: in the identification of (E)-ocellenyne, which has a similar core structure, deviations between theory and experiment that are much greater than these values would suggest an incorrect assignment. As summarized in Figure 2, the C5 and C8 chemical shifts of the four candidate structures for (E)-ocellenyne were most diagnostic: syn diastereomers showed large deviations from the natural product shifts, strongly suggesting that the actual 6,10-
The second iteration entailed organophosphorous-mediated oxonium ion formation followed by SiO2-promoted fragmentation. After some experimentation, we were pleased to find that exposure of hydroxyether 12b to (n-Oct)3P and CCl4 under the Hooz chlorination conditions11 in the presence of activated SiO2 produced a 6.4:1 mixture of the desired 6,10-anti-2,5dioxabicyclo[2.2.1]heptane 7b (82% isolated yield) and (6R)erythro-bis(2,2′)-THF 14b in excellent total yield (95%). The C6/C10-anti relative stereochemistry in 7b was established by spectroscopic studies, with the diagnostic NOE interactions being observed between H6 and H8, H5 and H10, and H8 and H11. Furthermore, the coupling constants between the bridgehead protons of anti isomer 7b (J6,7 and J9,10 ≈ 0) agree well with those of (E)-ocellenyne (J6,7 and J9,10 ≈ 0),1 which constitutes strong support that the natural product possesses a 6,10-anti-2,5dioxabicyclo[2.2.1]heptane skeleton; the corresponding coupling constants in syn isomer 7a are close to 2 (see below). We reasoned that hydroxyether 12b would be converted into the corresponding oxophosphonium intermediate upon exposure to (n-Oct)3P and CCl4, which in turn would generate dioxatricyclic oxonium ion 8b via transannular participation by the ring oxygen atom. Nucleophilic attack by chloride at C10 in oxonium ion 8b could then produce chloroether 13b with double inversion of configuration.12 Chloroether 13b would be in equilibrium with oxonium ion 8b, which would undergo 6254
DOI: 10.1021/acs.orglett.7b03226 Org. Lett. 2017, 19, 6252−6255
Letter
Organic Letters ORCID
Deukjoon Kim: 0000-0003-4079-8734 Robert S. Paton: 0000-0002-0104-4166 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the Government of Korea (MSIP, NRF Grant 2007-0056817), the Oxford University Press John Fell Fund, and the Royal Society (RG110617).
■ ■
DEDICATION In memory of the late Professor Gilbert Stork.
Figure 2. Values of Δδ for four diastereomeric candidate structures for (E)-ocellenyne.
relative stereochemistry is anti. The relative stereochemistry between the bicyclic skeleton and the syn-12,13-dibromo moiety in the C10 side chain was, as expected, much harder to assign on the basis of the calculations, as there were no large errors seen for (S,S)-anti or (R,R)-anti diastereomers relative to the experimental spectrum. The (R,R)-anti assignment agreed closest with experiment: it has the lowest root-mean-square error (1.1 ppm) and is the most probable assignment on the basis of a 85% DP4 metric (vs 15% for (S,S)-anti and
