Medicinal chemists are interested in weird ring systems – I can’t deny it. We like them for several reasons, one of which is that we’re always looking for variations on structural themes that might affect the metabolism or absorption of our drug molecules. And the fact that these variations can lead to new patentable chemical space is not to be minimized, either. This is why you see so many cyclopropyls and cyclobutyls, azetidines and oxetanes in the med-chemi literature, and why new reactions to work with them get our attention.
So have a look at this new paper in Angewandte Chemie. They’re taking what is truly a small ring system – cubane, which is shaped exactly the way you’d figure – and substituting it into various biologically active structures. Cubane itself was first synthesized back in the 1960s by Philip Eaton, and it wasn’t a particularly easy task. Since then, I’d guess that the bulk of the literature on the ring system has come from two fields of chemistry: the theoreticians, looking at how different substituents affect the strained bonds in the ring, and the energetic materials folks, looking to pack that already high-energy structure with even more payload. Octanitrocubane is a fine example of that, densely functionalized with enough bang to the gram to send you briskly down the stairwell to avoid it.
But there are other reasons to work with this ring system. Eaton himself speculated in 1992 that cubane might serve as an isostere for a phenyl ring in many compounds, which at first sounds a bit odd. But keep in mind that aryl rings like that have a goodly amount of electron density above and below them – from that perspective, they’re chunkier than they look, and the spacing across opposite corners of a cubane molecule is about right as well. This latest paper puts that hypothesis to the test in what I think must be the most comprehensive way yet. (Examples have appeared in the literature from time to time, and there may well be some more out in patents, but I don’t recall a direct test like this one before). As the authors here say, “This apparent lack of interest in cubane presumably follows the incorrect assumption that such compounds are esoteric, unstable, or synthetically intractable“.
Well, they’re not easy, but they’re doable, for sure. In this case, of the molecules in which a cubane was worked in for a phenyl, two of them were more active (the derivatives of leteprinim and diflubenzuron), and two were equipotent (the cubane analogs of SAHA and benzocaine). Attempts to swap in a cube for one or both phenyls in benzyl benzoate (an acaricidal agent) let to lower potencies, though. These tests were done all the way to to in vivo (for example, in xenograft models with the SAHA analog), which makes the overall case pretty strong.
Overall, the cubane derivatives are slightly more lipophilic than the phenyls, which makes sense just by eyeballing them. Metabolically, the ring system appears very stable – those hindered tertiary carbons are not very reactive. Cubane itself (as mentioned above) is thermodynamically rather elevated, but it’s quite stable, because there are really no easy ways for the structure to rearrange and break down. That’s probably helping it out with the p450 enzymes as well. Some of the same CSIRO team from Australia reporting these results have also developed a large-scale synthesis of cubane 1,4-dicarboxyl systems, so the objection of “well, you can’t get any intermediates” may be disappearing as well.
A last note: the paper is dedicated to Philip Eaton, on his 80th birthday. I’m glad that he’s getting a chance to see his work getting some traction in medicinal chemistry!