Nature achieves many highly selective and specific transformations at surprisingly mild conditions just by  controlling the supramolecular environment within the enzymatic cavities. Impressive examples thereof are terpene cyclases which promote polyene cyclizations, where highly complex carbon units with multiple stereocenters are created by the controlled ring closure of linear molecules. Enzymatic reactions have inspired many fields of (organic) chemistry thus far and led to a significant extension of the synthetic toolbox.

An interesting approach envisioned by our groups was the employment of fluorinated alcohol environments thus harnessing their unique ability to promote extended H-bonding networks either with themselves and/or Lewis bases.  Our initial studies begun shortly after disclosing the constructive effect that hexafluoro-2-propanol (HFIP) had on haliranium-induced cyclization of linear polyenes in 2018 (Scheme 1).¹

During the extensive screening of suitable Lewis bases and, for ease of handling, their respective salts, we observed some proton-cyclized materials as – in this case – undesired side products. These results were later revisited with the aim to allow for the proton-induced cyclization to occur in the most selective and controlled manner possible. Again, extensive optimizations were undertaken, with the first step being the screening of various protonated Lewis bases in HFIP. It was already clear at this time, that both the employed Lewis base and the counterion have a decisive role on the selectivity as well as on the speed of the reaction. The most optimal conditions were found to produce the desired product in a maximum yield of ca. 80% at this time with HFIP being the solvent and pyridinium bromide as the most effective catalyst.

It was only after more than 100 separate optimization reactions – a big shout out to our gas chromatograph at this point, which made analysis of conversions and yields very convenient – that we performed a final solvent screening of non-standard fluorinated alcohols, for example 1,1,1-trifluoro-2-propanol, trifluoroethanol or perfluoro-tert-butanol (PFTB). At the one-hour mark, no major surprises were observed, but after analyzing the mixtures again after 24 hours it was evident that using the latter solvent the reaction proceeded in overwhelming selectivity and a GC yield of 96%.

At this time, the real work began by identifying the underlying mode of action, both supported by wet chemistry (i.e. NMR titrations, kinetic studies, and deuteration experiments), and computational methods.  

We wanted to investigate the influence of the solvent PFTB on the coiling behavior of GerBn by molecular dynamics (MD) simulations, for two reasons:

First, it was observed that the reaction is occurring specifically well in this solvent. Second, the aggregation behavior in PFTB is of interest due to the fact that fluoro-alcohols (more specifically HFIP) have shown micro-structuring before (e.g. by Hollóczki et. al²).

This investigation led to a big surprise: The solvent by itself did not significantly stabilize the coiled form compared to other solvents. In the presence of the investigated ions however, percolating structures were observed in the simulations and a stabilization of the coiled form of GerBn (by the comparison of potentials of mean force) was observed.

This means the apparent solvent effect on the ring-closing reaction is likely due to this supramolecular aggregation which explicitly depends on a combination of both the solvent and the catalytic amount of the salts used in the reaction.

This is indeed a fascinating finding and reminiscent of environmental control, although pre-organizations are not achieved by the use of spatial confinement or specific binding – rather by supramolecular chemistry!

This is obviously not restricted to just the model system GerBn. It was possible to show in the experiments that it is possible to transform a broad variety of structurally diverse alkenes to the corresponding ring-closed products in up to quantitative yields with astounding functional group tolerance efficiencies while retaining excellent stereocontrol (diastereoselectivity up to d.r >95:5).

For practical applications, it is important to mention that the method itself only requires cheap, off-the-shelf components and is done under mild, and easily achievable reaction conditions. Also, it is now clear that the previously rarely used PTFB might be as interesting in syntheses as HFIP, because it is an outstanding reaction medium for proton-induced polyene cyclization and can furthermore be used for the stabilization of cationic species in a dipolar medium.

If you are interested in the detailed substrate scope, and/or the mechanistic investigation, feel free to take a look at the details in our recent paper by Gulder et. al..

1. Arnold, A. M., Pöthig, A., Drees, M. & Gulder, T. NXS, Morpholine, and HFIP: The Ideal Combination for Biomimetic Haliranium-Induced Polyene Cyclizations. J. Am. Chem. Soc. 140, 4344–4353 (2018).
2. Hollóczki, O. et al. The Catalytic Effect of Fluoroalcohol Mixtures Depends on Domain Formation. ACS Catal. 7, 1846–1852 (2017).