B.R.S.M. Home of Woodward Wednesdays


Unnatural Products 3: Pentaprismane

For the last planned post in my Unnatural Products series, I’m going to write about Eaton’s 1981 synthesis of pentaprismane.[1] At the time, unnatural hydrocarbons were hot targets, and as the next largest prismane on the list this target was the subject of much research by groups around the world. Perhaps Eaton's biggest rivals were the groups of Paquette and Petit, and in fact all three had, at various times, synthesised hypostrophene as an intended precursor to the target.

Unfortunately, the ‘obvious’ [2 + 2] disconnection from pentaprismane turned out to be a dead end and the photochemical ring closure was unsuccessful. The 1970s and early 1980s saw the publication of a number of other similarly creative, but sadly ill-fated, approaches based on various ring contractions, and the compound gained a well-earned reputation for extraordinary synthetic inaccessibility.

Eaton’s route began, as with the cubane and dodecahedrane syntheses previously covered in this series, with a Diels-Alder reaction. The diene used was the known tetrachlorocyclopentadienone acetal shown that upon heating neat with benzoquinone produced the endo adduct shown in excellent yield. Next, an even higher yielding photochemical [2 + 2] reaction was used to close the cage-like structure by cyclobutane formation. Treatment with lithium in liquid ammonia simultaneously reduced both ketones and removed all four chlorine atoms. The resulting diol was converted to the ditosylate, which, under carefully controlled conditions with sodium iodide in HMPA, underwent a mono-Finkelstein reaction to give the iodotosylate shown. When this was treated with t-BuLi halogen-lithium exchange, followed by an extraordinary fragmentation, gave a diene reminiscent of hypostrophene shown above. However, the extra carbon atom in the skeleton made all the difference, and unlike the parent compound, this did undergo a [2 + 2] cycloaddition when exposed to UV light. Finally, acetal hydrolysis gave homopentaprismane in 34% yield from benzoquinone, putting the group a single ring contraction from victory.[2]

With significant amounts of homopentaprismanone in hand, the group now intended to employ the transformation that had been the cornerstone of their cubane synthesis – the Favorskii rearrangement. Unfortunately, this required the introduction of a leaving group in the ketone α-position, a transformation made incredibly difficult due to the strained system and Bredt’s rule, which prevented enolisation.[3] Eventually a six-step sequence (!) to introduce a tosyloxy group was devised, beginning with a Baeyer-Villiger reaction using m-CPBA. A remarkable CH oxidation with RuO4, generated in situ, then gave the hydroxylactone. Treatment of this with diazomethane gave the corresponding δ-ketoester in almost quantitative yield. The group then reformed the starting norbornane-like bridge through use of an unusual acyloin type reaction effected by treatment with sodium in liquid ammonia. Finally, oxidation of the secondary alcohol and tosylation gave the Favorskii precursor, apparently preparable in muti-gram quantities.

Treatment with aqueous potassium hydroxide solution effected Favorskii rearrangement in excellent yield, especially considering that this was the first time the elusive pentaprimane ring system had been prepared. Finally, Eaton used the three-step decarboxylation he had developed for cubane to remove the extraneous acid and give pentaprismane in 18 steps. Awesome.[4]

References and suchlike

  1. J. Am. Chem. Soc., 1981, 103, 2134. Much like Eaton’s seminal cubane paper, the title is a single word, ‘Pentaprismane’. I love the lack of hype.
  2. Although Petit had prepared this compound a full decade earlier, his approach relied on a cycloaddition of the difficult to prepare cyclobutadieneiron tricarbonyl with the acetal of tropone, and proved difficult to scale  up. In fact, in his own paper Eaton rather directly described it as ‘conceptually fascinating [but] useless synthetically’.
  3. Eaton uses the phrase ‘invasion at the bridgehead’, which I find delightfully evocative. Makes it sound like a second world war campaign. Apparently the group initially planned, in spite of Bredt’s rule, to deprotonate the bridgehead position, relying on inductive stabilisation of the anion rather than enolate formation, but were unable to do so.
  4. Pentaprismane is the most recent of three prismanes synthesised to date, the other two being cubane, and triprismane. Although I think triprismane looks quite silly, it was actually synthesised some 8 years previouslyby T. J. Katz in far fewer steps. Go figure.



Comments (6) Trackbacks (0)
  1. Another excellent post! Interestingly, hexaprismane has not yet been synthesized despite considerable effort. This molecule is quite interesting, as it is a dimer of benzene. Keep up the great work!

  2. That’s homopentaprismanONe. (Anyone prepare pentahomopentaprismane? That looks rather interesting.)
    A couple other typos – norboRnane-like & 2,4,6-trIisopropylnitrobenzene.
    How important are those chlorine atoms in the starting material, & why?
    How about a post on triprismane? Bullvalene is another interesting one.

    • Thanks spotting the typos, some of those are quite silly. I’m not really sure why the chlorine atoms are there – maybe to discourage the irritating dimerisation that cyclopentadienes undergo? A similar tactic is used in Prinzbach’s pagodane synthesis (it’s covered in a great presentation from the Trauner group I found this morning: http://www.cup.uni-muenchen.de/oc/trauner/synstrat/Syntheseplanung09.pdf) where an odd looking tetrachlorothiophene S,S-dioxide is used as a diene then all the chlorines are removed.

      I would like to cover at least triprismane, propellane and pagodane (the Prinzbach route is great) if I revisit this series. There’s also more Woodward Wednesdays and some newer stuff planned, I’m just so damn busy right now.

  3. Idle thought – instead of the extra CO, could you use sulfur? I have the impression that desulfurization is rather easy.

    • Yes, I would have though that would work quite well – using something like thiophene S,S-dioxide in place of cyclopentadiene and then later doing an alpha-halogenation (in place of that rather lengthy alpha-tosylation sequence), followed by a Ramberg-Bäcklund Reaction in place of the Favorskii.

Leave a comment

nine × = 54

No trackbacks yet.