In the preivious post I gave the Tm data for homo-adenine and homo-thymine Oxepane Nucelic Acids (ONA) and it was quite clear that the low melting temperatures renders ONA useless from a pharmaceutical point of view. However, as I said it has a high stability in serum and activates RNase H. To date only very few oligonucleotide analogues have activated RNase H. The authors seem to be of the opinion that only four RNase H activating oligonucleotide (ON) analogues have been reported to date. However, they seem to forget the very first and most famous analogue - Phosphorus monothioates (PS). PS have been well known since the early 80s and ISIS Pharmaceuticals tried deveoloping drugs based on this class of analogues for a loooong time (they may still be doing so for all that I know). Anyway, let's have a look at the chemistry. They choose a somewhat surprising starting material and do a very funky Vorbrüggen coupling. I've never seen anything quite like it (notice the counterintuitive stereochemical outcome). Very impressive although the yields are poor. They get a fair bit of the diene oxepan product and some of the alfa-anomer too. Nevertheless a nice piece of work. Apparently an adaption of some work by Hoberg (JOC, 1997, 62, p. 6615) that I haven't checked out. People who haven't worked with nucleosides probably don't realise how difficult even the simplest transformations can be a times. Thymine and Adenine are by far the two "easiest" nucleobases to deal with. Cytosin is bad news and Guanine can be an outright nightmare. This is probably the main reason why everyone tests T and A first. Anyway, after taking the protection group off they proceed to hydrogenate the olefin which goes well. However, from here on it's a bit nasty. Firstly, they decide to use monomethoxy trityl (MMT) rather than dimethoxy trityl (DMT) because they get better yields this way. Now different molecules behave differently but I have a hard time accepting that you can't get DMT on in a higher yield than they report for MMT protection. Very odd! I wish they informed what happens instead. Maybe di-protection is a problem for some reason? Furthermore, they proceed to synthesise the phosphoramidite using the classic phosphoramidochlorodite (PCl) reagent. I see this all the time. Virtually, everyone in the area is using this phosphitylation reagent instead of the superior phosphordiamidite (PN2) reagent (See below). Back in the Jurassic when I used to make phosphoramidites I always used PN2. As a consequence no chromatography was necessary and I got >95% yield. We published a paper on how to make LNA (pioneered by Wengel) phosphoramidites using this reagent some years ago (Synthesis, 2002, 6, p. 802). Anyone interested in improving their phosphoramidite synthesis can request a copy (email@example.com). So later on they proceed to synthesise their oligonucleotides and only achieve coupling efficiencies of 98-99%. As always it is hard to know why the yield is reduced. I wonder if the use of MMT protection has anything to do with it. One final comment regarding their conclusion. They seem to spot a connection between RNase H activity of ON analogues and sugar conformer flexibility along the ON strand. However, they have just mentioned that alpha-L-LNA is a known ON analogue that activates RNase H. This analogue contains a highly constrained bicyclic sugar (See Figure) and hence doesn't support their conclusion. They could conceivably be right. Maybe some other mechanisms are at work with alpha-L-LNA but I think they should at least have mentioned this. Anyway, overall a good paper from this Canadian research group. Keep them coming. D!