Synthesis of trinorbornane

Lorenzo Delarue Bizzini a, Thomas Müntener a, Daniel Häussinger a, Markus Neuburger a and Marcel Mayor *abc
aDepartment of Chemistry, University of Basel, St. Johanns-Ring 19, 4056 Basel, Switzerland. E-mail: marcel.mayor@unibas.ch
bInstitute for Nanotechnology (INT), Karlsruhe Institute of Technology (KIT), P. O. Box 3640, 76021 Karlsruhe, Germany
cLehn Institute of Functional Materials (LIFM), Sun Yat Sen University (SYSU), XinGangXi Rd. 135, 510275 Guangzhou, P. R. China

Received 10th August 2017 , Accepted 26th September 2017

First published on 26th September 2017


Herein we report the synthesis and characterisation of the until recently unreported chiral C11 skeleton of tetracyclo[5.2.2.01,6.04,9]undecane (“trinorbornane”) which could be obtained in 7% overall yield in 9 steps. This new rigid structural type was found to be present in the computer generated Chemical Universe Data-base (GDB) and has until now no real-world counterpart.


The Chemical Universe Database GDB, which was introduced by the group of Reymond, is in its first iteration an exhaustive enumeration of all possible molecules up to eleven atoms (C, N, O and F) considering chemical stability and synthetic feasibility.1 Through employing the molecular quantum number (MQN) classification method, which is an equivalent of the periodic system of elements for molecules, the interactive search of the chemical space was rendered possible.2

An analysis of the enumerated database for structures with up to 11 carbon atoms revealed 124 polycyclic hydrocarbons without 3- and 4-membered rings. From these structures only three have no real world counterparts, neither as pure hydrocarbons nor as substructures containing heteroatoms or unsaturations.2 One of these structures is tetracyclo[5.2.2.01,6.04,9]undecane (1), which is a particularly appealing and eye catching example of a new structural type.1,2

Its scaffold consists of two norbornanes that share a pair of neighboring edges (red and blue in Fig. 1, left). Interestingly, this arrangement results in a third norbornane subunit (highlighted in yellow in Fig. 1, left). Furthermore, the structure is an example of axial chirality, and Fig. 1 shows the two enantiomers (1Sa,4S,6R,7S,9R)-tetracyclo-[5.2.2.01,6.04,9]undecane (left) and (1Ra,4R,6S,7R,9S)-tetracyclo[5.2.2.01,6.04,9]undecane (right), which we will refer to as 1-Sa and 1-Ra for convenience. We suggest for the scaffold of 1 consisting of three superposed norbornane units the trivial name trinorbornane. Fig. 1 also shows the mirror plane between both enantiomers and for 1-Ra the molecule's C2-axis. The numbering of the carbon atoms according to the IUPAC name is given for 1-Sa.


image file: c7cc06273g-f1.tif
Fig. 1 The two enantiomers (1Sa,4S,6R,7S,9R)tetracyclo[5.2.2.01,6.04,9]undecane (1-Sa) and (1Ra,4R,6S,7R,9S)-tetracyclo[5.2.2. 01,6.04,9]undecane (1-Ra) displaying the axial chirality. For 1-Sa, on the left side the three norbornane subunits of the structure are displayed in red, blue, and yellow together with the numbering of the eleven carbon atoms corresponding to the IUPAC name. The enantiomer 1-Ra is displayed together with the molecules on the C2-axis.

There is a long tradition in organic chemistry to explore the chemical space for complex, symmetrical and aesthetically appealing hydrocarbons by synthetic means.3 One paramount example is dodecahedrane (2), called “the Mount Everest of hydrocarbon chemistry”,4 which was first synthesised by Paquette in 1982.5 It is composed of twelve fused cyclopentane rings and possesses the highest known point group symmetry (Ih).4

Our search for compounds resembling the unstrained but highly rigid carbon framework of 1 guided us to the pentacyclic, chiral, and D3-symmetric trishomocubane 3 (pentacyclo-[6.3.0. 02,6.03,10.05,9]undecane). It is also composed of three partially superposed norbornyl units (Fig. 2) and was first synthesised by Underwood in 1970 via a sequence of brominations and eliminations from pentacyclo-[5.4.0.02,6.03,10.05,9]undecane-8,11-diol.6 Since then, several routes to 3 and its derivatives thereof have been developed.7


image file: c7cc06273g-f2.tif
Fig. 2 Examples of already known polycyclic hydrocarbons with rigid scaffolds: dodecahedrane (2), trishomocubane (3), brexane (4) and its isomer brendane (5).

Probably the structurally most similar known compound is brexane 4 (tricyclo[4.3.0.03,7]nonane), which lacks only one interconnecting ethylene bridge compared to 1 and was synthesised by Nickon in 1964 in 9 steps (Fig. 2).8 Later its synthesis has been improved by Brieger profiting from an intramolecular Diels–Alder reaction as the key step.9 In similarity to 1, this tricyclic structure is also C2-symmetric but is composed of two superposed norbornyl units instead of three. Interestingly, the tricyclic scaffold has been found in the natural products pallambins A and B, whose total syntheses were recently achieved.10

Our approach to construct the tetracyclic scaffold of 1 (Scheme 1) was based on the tricyclic brexene derivative 6, which might form the desired target structure 1 in a 6-endo-trig/5-exo-trig radical cyclisation mechanism. The compound 6 should be formed in an intramolecular Diels–Alder reaction9 from the cyclopentadiene 7, which is envisioned to be accessible from the literature known compound 8via a Michael addition to phenyl vinyl sulfone and a series of reduction and elimination reactions.


image file: c7cc06273g-s1.tif
Scheme 1 Retrosynthetic analysis for racemic trinorbornane 1 (to improve the visibility, only the 1-Ra enantiomer is displayed).

The synthesis of the target structure trinorbornane 1 as a racemate is displayed in Scheme 2 and starts with the assembly of the 3-ethyloxycyclopent-2-enone derivative 8. The assembly of 11 starting from methyl (E)-4-chloro-3-methoxy-2-butenoate (9) is straightforward and even suitable for industrial production.11 The addition of 9 to a solution of dimethyl malonate (10) in DMF under basic conditions at room temperature resulted in trimethyl (E)-3-methoxy-3-butene-1,1,4-tricarboxylate (11) in a very good yield of 85%. The triester 11 was subsequently cyclised in a solution of sodium methoxide in methanol at 60 °C providing methyl 4-methoxy-2-oxo-3-cyclopentenecarboxylate (12) in 86% isolated yield.


image file: c7cc06273g-s2.tif
Scheme 2 Syntheses of the polycyclic hydrocarbons 1 (for clarity only the 1-Sa enantiomer is displayed), 18, 19, and A as the not (yet) synthesised constitutional isomer of 1. Reagents and conditions: (a) NaOMe, DMF, r.t., 2 h, 85%; (b) NaOMe, MeOH, 60 °C, 3 h, 86%; (c) KH, DMF, 13, r.t., 4 h, 76%; (d) KOH, EtOH, reflux, 16 h, 76%; (e) (1) LDA, THF, −78 °C, (2) DMI, 0 °C, (3) 15, 0 °C to r.t., 1 h, 82%; (f) DIBAL-H, toluene, r.t., 30 min, then 2 M HCl (aq), r.t., 20 min 86%; (g) (1) NaBH4, CeCl3·7H2O, MeOH, 0 °C, 1 h, (2) cat. p-TsOH, toluene, reflux, 32 h, 83%; (h) SmI2, THF, DMPU, 80 °C, 2 h, 67% determined by GC-MS, 28% isolated. (Alternative: LDMAN, THF, −78 °C, 30 min, 1(≈2%), 18(≈1%), 19(13%) determined by GC-MS.)

The 5-membered ring of 12 is already a subunit of the tetracyclic structure 1 and in order to form the next two rings in a similar manner to the strategy reported for brexane9via an intramolecular Diels–Alder reaction, a but-3-ene chain had to be introduced. Following a reported procedure,12 compound 12 was deprotonated with potassium hydride in DMF at room temperature and alkylated by the addition of 4-bromo-1-butene (13), providing 1-(but-3-enyl)-4-methoxy-2-oxo-cyclopent-3-enecarboxylic acid methyl ester (14) in 76% isolated yield. In order to form the quaternary centre, decarboxylation of 14 was required. Following a reported protocol,1214 was treated in refluxing ethanol with potassium hydroxide for 16 hours. Under the applied reaction conditions, not only the compound was decarboxylated, but also the complete substitution of the methoxy group by an ethoxy group was observed, which does not handicap the further proceeding along the reaction sequence. Thus, 5-(but-3-enyl)-3-ethoxycyclopent-2-enone (8) was obtained in 76% isolated yield. A key step of the synthesis was the formation of the quaternary centre, which was achieved by modifying the procedure reported by Cory and Reeneboog13 using phenyl vinyl sulfone (15) as the Michael acceptor. After the formation of the kinetic enolate of 8 with LDA in dry THF under cryogenic conditions, the solution was allowed to warm up to 0 °C and 1,3-dimethyl-2-imidazolidinone (DMI) was added. A solution of phenyl vinyl sulfone (15) in dry THF was added at 0 °C and allowed to warm up to room temperature. After about 1 hour the reaction mixture was worked up and 5-(but-3-enyl)-3-ethoxy-5-(2-(phenylsulfonyl)ethyl)cyclopent-2-enone (16) was isolated in good yields of 74–82% using column chromatography (CC). Water exclusion turned out to be crucial for the success of this reaction as in the presence of traces of water mainly polymeric material was formed.

First attempts to reduce 16 with either DIBAL-H or LiAlH4 in THF and subsequent treatment with hydrochloric acid gave, in agreement with the literature,14 mixtures of 1,2-, 1,4- and over-reduced species. Changing the solvent to toluene and using DIBAL-H as the reducing agent provided selectively 4-(but-3-en-1-yl)-4-(2-(phenylsulfonyl)ethyl)cyclopent-2-enone (17), which was, after stirring in 2 M aqueous hydrochloric acid for 20 minutes, isolated in excellent 86% yield. The cyclopent-2-enone fragment of 17 was reduced to the cyclopent-2-enol with the typical conditions for a Luche reduction.15 Thus, 17 was treated with cerium(III) chloride and sodium borohydride in 0 °C cold methanol. The crude reaction product comprising 4-(but-3-en-1-yl)-4-(2-(phenylsulfonyl)ethyl)cyclopent-2-enol as a mixture of stereoisomers was not purified but instead directly dehydrated in toluene with catalytic amounts of p-TsOH at 110 °C. Under these conditions, the in situ formed cyclopentadienyl derivative 7 reacts with the but-3-ene chain in an intramolecular Diels–Alder reaction9 providing 1-(2-(phenylsulfonyl)-ethyl)tri-cyclo [4.3.0.04,9]non-7-ene (6), which was isolated in very good 83% yield over three steps.

The second key step of the synthesis was the reduction of the sulfone moiety of 6 with single-electron reducing agents in order to form the 2-ethyl radical which was expected to cyclise to the target structure 1 by reacting intramolecularly with the olefin of the brexene scaffold. In first attempts, 6 was treated with large excesses of lithium 1-N,N-dimethylaminonaphthalenide (LDMAN)16 as the single-electron reducing agent at −78 °C in THF. To our delight, we were able to observe up to 30% conversion and further analysis of the reaction mixture using GC-MS unravelled a 5[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio in favour of the reduced brexene derivative 19 compared to the cyclised target structure 1. In addition, also the formation of the radical dimerisation product 18 was observed. The mixture of compounds was subjected to CC using silica gel as the solid phase and pentane as the eluent at −20 °C. The target structure 1 was isolated together with the dimer 18 in a combined yield of about 3%, while the open brexene derivative 19 was isolated in 13% yield. The dimer 18 was separated from the mixed fraction by crystallisation in a mixture of diethyl ether and acetonitrile and was isolated in 1% yield.

While the reaction sequence with its poor conversion, low yield of the desired compound, and complex product mixture was not particularly appealing, it shows that the target structure 1 can indeed be formed via a radical cyclisation reaction. Furthermore, the dimeric side product 18 was crystalline in the solid state and indeed, single crystals suitable for X-ray analysis were obtained by sublimation in a sealed capillary at 130 °C. While the dimer was formed as a mixture of diastereomers, the single crystal consisted exclusively of the meso-form combining a 1-Ra subunit with a 1-Sa subunit (left and right parts respectively in Fig. 3). The solid state structure of 18 is displayed in Fig. 3. 18 crystallises in the monoclinic space group P21/n with two formula units per unit cell. Only one half of the dimeric molecule is present in the asymmetric unit. The complete molecule is generated by an inversion centre of the space group P21/n.


image file: c7cc06273g-f3.tif
Fig. 3 Solid state structure of the dimer 18 corroborating the existence of the tetracyclo[5.2.2.01,6.04,9]undecane system as it consists of an interlinked 1-Ra subunit (left) and a 1-Sa subunit (right). Hydrogen atoms are omitted for clarity and rotational ellipsoids are displayed at the 50% probability level.

As an alternative, SmI2 in THF with 1,3-dimethyltetrahydropyrimidin-2(1H)-one (DMPU) as the cosolvent is reported to be a strong enough single-electron reducing agent for sulfones.17 To our delight, employing 10 equivalents of SmI2 in THF with 8 equivalents of DMPU with respect to SmI2 at 80 °C for 2 h afforded exclusively cyclised 1 with 67% conversion determined by GC-MS. After the usual work up, the desired tetracyclic target structure 1 was isolated in 28% yield as a colourless crystalline solid by filtering off the remaining starting material and polar residues through a silica pad using pentane as the eluent. The low isolated yield compared to the conversion determined by GC-MS is due to the high volatility of the compound resulting in losses during solvent evaporation. Pentane was distilled off through a Vigreux column and the residual solvent was removed under vacuum while the flask was placed in an ice bath. It is noteworthy that the ring closing radical reaction seems to occur in a regioselective manner, as exclusively the formation of the trinorbornane scaffold is observed and the also not yet known C11-analogue (A in Scheme 2), which should be obtained when the other olefinic carbon atom would be attacked as well, could not be detected with any of the investigated reaction conditions.

The racemic mixture of 1 was fully characterised by 1H- and 13C-NMR, and high resolution mass spectrometry. Despite the apparent simplicity of the molecular architecture of 1 with only six carbon and eight proton resonances, it turned out exceedingly difficult to unambiguously confirm the suggested structure by standard HSQC/HMBC/NOESY NMR experiments due to the high order nature of the 1H NMR spectrum and multiple pathways for long-range H–C couplings in the rigid tetracyclic framework. We therefore resorted to the very sensitive 1,1-ADEQUATE experiment18 using chirp pulses19 that selectively delivers correlations between the protons and carbon atoms that are separated by two bonds, thus allowing us to assign the topology of 1 without any doubt (Fig. 4). With the proton and carbon assignment at hand, it was possible to distinguish the diastereotopic protons in the CH2-groups by 2D-NOESY NMR using an unusually long mixing time of 3 seconds (Fig. S1 shows the key NOE contacts, ESI). The crucial NOE contact between symmetry related protons 5b and 8b could be obtained from a pseudo 3D HSQC-NOESY experiment without carbon decoupling during the 1H acquisition which splits the diagonal peak into a 130 Hz doublet allowing for the detection of a weak but significant NOE between the protons 5b and 8b (data not shown).


image file: c7cc06273g-f4.tif
Fig. 4 1,1 ADEQUATE spectrum of trinorbornane 1 recorded in C6D6 with assignment.

While all the spectra unambiguously prove the identity of 1, all attempts to analyse its solid state structure using X-ray analysis failed. Despite the fact that 1 is a crystalline solid at room temperature, the compound crystallises upon sublimation in thin needles, which are notoriously too thin to allow for analysing the solid state structure with the available X-ray sources.

In conclusion, we report the first synthesis of trinorbornane, a tetracyclic saturated hydrocarbon with the elemental formula C11H16. While theoretical considerations predicted the stability and accessibility of the trinorbornane scaffold, it has neither been synthesised nor been found in nature so far. The tetracyclic scaffold has been assembled in a linear sequence of 9 steps with a total isolated yield of 7%. The key steps of the synthesis are the introduction of a quaternary carbon atom in an early stage providing the precursor 16, already comprising not only the first 5-membered ring of the target structure, but also all the required eleven carbon atoms. An intramolecular Diels–Alder reaction resulted in the next two ring structures of the framework and the last cyclisation step was achieved via an intramolecular radical reaction.

We are currently exploring functionalised trinorbornane structures as well as synthetic routes towards enantiomerically pure samples of this tetracyclic C11-hydrocarbon.

The manuscript is dedicated to Prof. Jean-Louis Reymond as the inspiring muse provoking the endeavor. The authors gratefully acknowledge financial support for this work from the Swiss National Science Foundation (SNF, grant number 200020-159730) and the University of Basel.

Conflicts of interest

There are no conflicts to declare.

Notes and references

  1. T. Fink and J.-L. Reymond, J. Chem. Inf. Model., 2007, 47, 342–353 Search PubMed .
  2. J.-L. Reymond, L. C. Blum and R. van Deursen, Chimia, 2011, 65, 863–867 Search PubMed .
  3. H. Hopf, Classics in hydrocarbon chemistry: syntheses, concepts, perspectives, Wiley-VCH, Weinheim, 2000 Search PubMed .
  4. Nachr. Chem., Tech. Lab., 1977, 25, 59–70 Search PubMed.
  5. R. J. Ternansky, D. W. Balogh and L. A. Paquette, J. Am. Chem. Soc., 1982, 104, 4503–4504 Search PubMed .
  6. G. R. Underwood and B. Ramamoorthy, Tetrahedron Lett., 1970, 11, 4125–4127 Search PubMed .
  7. I. A. Levandovsky, D. I. Sharapa, O. A. Cherenkova, A. V. Gaidai and T. E. Shubina, Russ. Chem. Rev., 2010, 79, 1005 Search PubMed .
  8. A. Nickon, H. Kwasnik, T. Swartz, R. O. Williams and J. B. DiGiorgio, J. Am. Chem. Soc., 1965, 87, 1613–1615 Search PubMed .
  9. G. Brieger and D. R. Anderson, J. Org. Chem., 1971, 36, 243–244 Search PubMed .
  10. C. Ebner and E. M. Carreira, Angew. Chem., Int. Ed., 2015, 54, 11227–11230 Search PubMed .
  11. R. Fuchs and J. F. Mcgarrity, Synthesis, 1992, 373–374 Search PubMed .
  12. B. Guay and P. Deslongchamps, J. Org. Chem., 2003, 68, 6140–6148 Search PubMed .
  13. R. M. Cory and R. M. Renneboog, J. Org. Chem., 1984, 49, 3898–3904 Search PubMed .
  14. R. L. Funk and K. P. C. Vollhardt, Synthesis, 1980, 118–119 Search PubMed .
  15. A. L. Gemal and J. L. Luche, J. Am. Chem. Soc., 1981, 103, 5454–5459 Search PubMed .
  16. R. Ivanov, I. Marek and T. Cohen, Tetrahedron Lett., 2010, 51, 174–176 Search PubMed .
  17. G. E. Keck, K. A. Savin and M. A. Weglarz, J. Org. Chem., 1995, 60, 3194–3204 Search PubMed .
  18. B. Reif, M. Kock, R. Kerssebaum, J. Schleucher and C. Griesinger, J. Magn. Reson., Ser. B, 1996, 112, 295–301 Search PubMed .
  19. M. Köck, R. Kerssebaum and W. Bermel, Magn. Reson. Chem., 2003, 41, 65–69 Search PubMed .

Footnote

Electronic supplementary information (ESI) available: Experimental details, analytical data and crystallographic data of 18 (CIF). CCDC 1556107. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c7cc06273g

This journal is © The Royal Society of Chemistry 2017