Akihiro
Yoshitani
a,
Tomonari
Sasage
a,
Ranmaru
Kato
a,
Takahiro
Suzuki
b,
Kazutada
Ikeuchi
*c and
Keiji
Tanino
*b
aGraduate School of Chemical Sciences and Engineering, Hokkaido University, Sapporo 060-0810, Japan
bDepartment of Chemistry, Faculty of Science, Hokkaido University, Sapporo 060-0810, Japan. E-mail: ktanio@sci.hokudai.ac.jp
cSchool of Pharmacy and Pharmaceutical Sciences, and Institute of Medicinal Chemistry, Hoshi University, Tokyo 142-8501, Japan. E-mail: ikeuchi.kazutada@hoshi.ac.jp
First published on 5th September 2025
We describe a halogen/chalcogen-mediated skeletal rearrangement of tricyclo[4.3.0.03,7]non-8-ene derivatives that feature a norbornene framework with two hydroxy/silyloxy groups substituted on the bridgehead carbons, yielding a tricyclo[4.2.1.03,7]nonane (brendane) skeleton. Four different types of consecutive Pd(II)-mediated skeletal rearrangement/C–C cross-coupling reactions are further demonstrated.
Bicyclo[2.2.1]heptane (norbornane) skeletons found in natural products are often fused to other carbocycles, such as tricyclo[4.3.0.03,7]nonane (brexane) and tricyclo[4.2.1.03,7]nonane (brendane) (Fig. 1a).4,5 The skeletal rearrangement of norbornane scaffolds contained in brexanes is a reasonable approach toward the synthesis of brendanes.† However, among studies adopting this approach, only two have involved cationic6 or anionic7 rearrangement from a simple substrate.
An electrophile-mediated rearrangement of norbornene to afford 1,3-difunctional norbornane is known,8 but such examples are limited owing to the competitive generation of 1,2-difunctional compounds (Scheme 1a). In 2003, the semi-pinacol rearrangement of 1-hydroxy-2,7,7-trimethyl-norborn-2-ene mediated by the electrophilic activation of the double bond assisted by the bridgehead hydroxy group was reported by the García Martínez and de la Moya Cerero group (Scheme 1b).9 However, these authors reported only three examples.
![]() | ||
Scheme 1 Transformation of a norbornene skeleton via (a) Wagner–Meerwein rearrangement, (b) a semi-pinacol rearrangement and (c) our strategy of skeletal rearrangement of the brexane skeleton (this work).† E = electrophile, Nu = nucleophile, L = ligand, Ph = phenyl. |
Although the Pd(II)-mediated activation of the norbornene double bond is known as the trigger of the Catellani reaction,10 to the best of our knowledge, there are no reports involving the Pd(II)-mediated skeletal rearrangement of norbornenes. This transformation is expected to enable cross-coupling at the C7 position of norbornanes. To date, reported examples of cross-coupling at this position have relied on a directing group substituted at the C2 pseudo-equatorial position.11
Herein, we demonstrate the electrophile-mediated semi-pinacol rearrangement of brex-2-ene derivative 1, whose synthesis we recently reported (Fig. 1c).12 This reaction proceeds chemoselectively and stereospecifically via the assistance of the hydroxy group at the C1 position to afford brendane 2.† When using a Pd(II) reagent as the electrophile, consecutive C(sp3)–C(sp3/sp2/sp) cross-couplings with various organometallic reagents are applicable. Synthesis of tricyclo[5.2.1.03,7]decanes (homobrendanes)13 is also reported.
To examine our remodelling strategy, we began with tricyclic compound 3 whose synthetic method is shown in SI-4. We first activated the olefin of 3 using halogenation reagents—N-bromosuccinimide (NBS), N-chlorosuccinimide, and N-iodosuccinimide—and the desired products 4a–c were obtained in high yields as single diastereomers (entries 1–3). These brendane structures were confirmed via the X-ray diffraction analysis of 4a. Oxidation of the double bond using an excess amount of m-chloroperbenzoic acid also induced the desired reaction, which afforded diol 4d in 40% yield (entry 4). Phenylsulfenyl chloride and phenylselenyl chloride easily produced 4e and 4f in high yields (entries 5 and 6). The use of mercury(II) acetate was also acceptable as the activator of 3 in the presence of catalytic scandium(III) triflate.14 The C–Hg bond of the resulting product was easily cleaved using NaBH4 under basic reaction conditions, which afforded 4g in 93% yield (entry 7). The same product was also obtained in 54% yield via the proton-catalysed transformation, although a high temperature and long reaction time were required (entry 8).
As shown in Fig. 2b, we applied the NBS-mediated skeletal rearrangement to various derivatives of 5. Notably, the skeletal rearrangement of bis trimethylsilyl (TMS) ether 5a also occurred by treatment with NBS to afford 6a in 85% yield. Compound 5b, wherein the methyl group of 3 was replaced with a benzyl (Bn) group, provided 6b in high yield. Attachment of a bromo group or an oxy-functional group at the C5 or C2′ position of 3 did not affect the reaction, affording 6c from 5c and 6d from 5d, in 95% and 93% yield, respectively. Although substrate 5e, which bears a cyano group at the C5 position in bis-TMS ether 5a,15 exhibited poor reactivity toward NBS under the tested conditions (rt, 4 d), resulting in 45% recovery of 5e and 30% yield of 6e, treatment with dibromoisocyanuric acid16 (rt, 23 h) dramatically facilitated the reaction to afford 6e in 81% yield. Homobrendanes 8a and 8b (dr = 7.2:
1) were also synthesized from 7a and 7b (dr = 7.1
:
1) in 83% and 74% yields, respectively.
The result of the Hg(II)-mediated skeletal rearrangement encouraged us to investigate the sequential Pd(II)-mediated skeletal rearrangement/cross-coupling. The use of palladium(II) reagent would induce skeletal rearrangement, and the resulting alkyl-Pd species 4h may undergo an insert reaction with allyl chloride (Scheme 2a).17 Then, β-Cl elimination of the resulting intermediate 9 would give 7-allyl brendane 10a accompanied by a palladium(II) reagent, leading to a catalytic allylation reaction of 3.
![]() | ||
Scheme 2 (a) Pd(II)-catalysed skeletal rearrangement of 3 along with sequential cross-coupling with allylchloride. (b) Other examples under our catalytic reaction conditions. |
Screening of the reaction conditions (see Table S1 in SI) revealed that treatment of 3 and an excess amount of allyl chloride with 10 mol% of PdCl2(MeCN)2 and an excess amount of LiCl in MeCN at 120 °C induced the desired reaction to afford 10a in 83% yield. When using 2,6-di-tbutylpyridine as the additive, chloride 10b that was generated via the β-H elimination of 9 was observed.
This catalytic reaction occurred when using 2-methylallylchloride to produce 10c in 61% yield. Preparation of homobrendane 11a was also demonstrated through the reaction of 7a with allylchloride.
The successful result for the transformation of 3 into 10a prompted us to explore cross-couplings with organometal reagents. We first investigated the Stille coupling of allyltributyltin and 4h, which was prepared by treatment of 3 with 1.2 equiv. of PdCl2(MeCN)2 in MeCN at 70 °C. Screening of the reaction conditions (see Table S2 in SI) revealed that the presence of fumaronitrile (12),18 an electron-deficient ligand,19 was essential for the reaction progress, and the reaction of 4h with 1.2 equiv. of allylbutyltin, and 1.2 equiv. of 12 in MeCN at 70 °C for 3 h afforded 10a in 72% yield (Scheme 3a, method A). This method enabled the use of other organotin reagents, resulting in the introduction of a vinyl (10d), 4-(trimethylsilyl)-2E-butenyl (10e), methyl (10f), phenyl (10g), phenylethynyl (10h), and 1-propynyl (10i) group at the C7 position. Interestingly, the use of vinyltributyltin proceeded smoothly at 50 °C in the absence of 12.20 We also synthesized homobrendane 11a and 11b from 7a and 7b, respectively, via method A using allyltributyltin.
![]() | ||
Scheme 3 Examples of various cross-couplings. a![]() ![]() ![]() |
We next investigated the Suzuki–Miyaura reaction using 4h and phenyl boronic acid. Screening results (see Table S3 in SI) revealed that the reaction in the presence of 12 (2.4 equiv.) and CsF (3.0 equiv.) in 1,2-dimethoxyethane (DME) at 80 °C provided 10g in 68% yield (method B). The use of 4-methoxycarbonyl- and 4-methoxyphenylboronic acids was acceptable to afford 10j and 10k in 75% and 20% yields, respectively. In the reaction with 2-methoxycarbonylpheylboronic acid, lactonization occurred spontaneously because of the proximity of the bridgehead hydroxy group and the methyl ester moiety, to isolate 10l in 40% yield. The application of method B using PhB(OH)2 into the alkyl-palladium intermediate derived from 5d was acceptable to isolate the desired product as acetylation form 13 in 42% yield as a single isomer. The use of allylboronic acid pinacol ester (allyl-Bpin) also provided 10a in 58% yield. In addition, the reaction of 4h with (E)-9-styryl-9-borabicyclo[3.3.1]nonane [(E)-9-styryl-9-BBN] and bis(pinacolato)diboron (pinBBpin) afforded 10m and 10n in 39% and 27% yields, respectively. Development of the carbonylation of 4h was also achieved by stirring in MeCN/MeOH under 1 atm of CO (method C), to give methyl ester 10o in 69% yield.
In summary, we established a facile method for synthesizing divergent brendane derivatives via the remodelling of various brex-2-enes with two bridgehead hydroxy groups. The rearrangement trigger is the electrophilic activation of the double bond, and the reaction proceeds both chemoselectively and stereospecifically. We further developed consecutive skeletal rearrangement and cross-coupling with allyl chloride by using 10 mol% of PdCl2(MeCN)2. The Stille and Suzuki–Miyaura cross-coupling and carbonylation of the Pd(II)-substituted brendane species were also achieved. Further investigation toward establishment of general catalytic reaction conditions is currently underway in our laboratory.
This research was supported in part by JSPS KAKENHI (grant numbers JP20K05485, JP21H01923, JP21K14616, JP23K04737, and JP24K01477), JST SPRING (Grant number JPMJSP2119), and the Photo-excitonix Project of Hokkaido University.
K. I. and K. T. conceived the research theme, and designed the experiments. K. I., A. Y., T. S., and R. K. performed the experiments and analysed the data. K. I. wrote the manuscript. T. S. and K. T. assisted in writing and editing the manuscript. All authors contributed to the discussions.
CCDC 2408963 (for 4a), 2408964 (for 6a), 2408965 (for 6d), 2408966 (for 10d), and 2408967 (for 10g) contain the supplementary crystallographic data for this paper.21a–e
Footnote |
† Because this paper focuses on the remodelling of brexane to brendane, carbon numbering of both skeletons according to the IUPAC rules hampers comprehension. Therefore, we adopted the original numbering of both skeletons, as shown in Fig. 1. To clarify which carbons of brendane correspond to those of brexane, both sets of carbon numbering are further indicated in the structure of 2, as shown in Scheme 1. |
This journal is © The Royal Society of Chemistry 2025 |