Friday, February 14, 2014

Structural Reassignment by Cool Science

Today, Vinogradova, Muller, and Buchwald have an excellent piece of investigative science in Angew. Chem. Int. Ed. that I think is great, just great. And honestly, I'm a bit jaded and easily unimpressed (especially when the work isn't directly applicable to me finishing my Ph.D.). In fact, I liked this paper enough to write about it, and not just tweet it on the sidebar <---over there.

Lets start with a summary of why I think its great.
1. They decided to take a closer look at a "known" system - and they learned something new.
2. They crystallized a liquid using the work by Fujita et al.

...And onward with a more in depth look...

So it seems that Buchwald et al. were interested in aryl isocyanates, specifically group transfer of isocyanates and isothiocyanates utilizing hypervalent iodine reagents. Upon preparing compound 3 and treating with silver isocyanate or silver isothiocyanate they observed ring-closed I-bound cyanates (4) and ring-opened O-bound isothiocyanates (5) [see Figure 2 below, taken without permission from the paper].

Which brings us to the thiotrifluoromethylating reagent (1) reported by Shen et al. back in 2013. Following all literature precedent, they assigned compound 1 as the ring-closed benziodoxole, and since the characterization data matched their desired compound there was no reason to worry. There is a decent body of literature to support this conclusion.

However, when the Buchwald group compared the NMR spectra of compounds 1-6, something seemed strangely amiss. Notably, proton Hb fell into one of two categories: ring-open benzyl alcohols (1, 2, 5, 6) around 7 ppm, or ring-closed benziodoxoles around 7.5 ppm (3, 4). Now here is reason #1 I like this paper - they decided to take a closer look at what was going on. I personally think these details would have been easy to overlook... after all it didn't affect the reactivity.

First, they utilized some 13C NMR work by the late Alan Katritzky, which corroborated their thoughts. They also prepared a crystalline variant of compound 1 (not shown), which confirmed an O-bound thiotrifluoromethyl group.

But it was time to go deeper.

And this is the second reason I like this paper, and actually the reason I read it in the first place... remember this Nature paper on crystallizing non-crystalline compounds that everyone was stoked on? And then how everyone was less than stoked on? Well, the Buchwald lab stuck with it, and even showed me this Nature Protocols paper by Fujita (that I missed) which details a better protocol for their "crystalline sponge" method. And, it worked. Cool! I suppose the full value of this method will be realized in time, but this a great (first, to the best of my knowledge) use of Fujita's method outside their own lab. Thus, Buchwald confidently revised the structure of compound 1 to the benzyl alcohol shown (which they refer to as a "thioperoxide," this is probably IUPAC, but I don't like it... peroxide to me is O-O... thioperoxide sounds like S-S, but that's disulfide. Now that I think about it, I kind of like dioxide to talk about O-O. I digress. Sorry.).

As Buchwald et al. point out, these new observations don't affect the reactivity, although they may open for discussion new mechanistic pathways. And most interestingly (to me), is that this is the first use of the still promising "crystal sponge" method by the Fujita group.


  1. I also noted this Angewandte paper and its use of the Fujita methodology, and I think you are correct in assuming that it's the first time it has ben picked up by someone else.
    Think of the applications of the soak for x-ray! The mind boggles. I hope the Nat. Prod. isolation people use it.

  2. Yes, this is indeed the first time this method has been used successfully outside of Fujita's lab.

    1. You sound so calm... AM I THE ONLY ONE REALLY STOKED ON THIS!?

    2. No, I am quite thrilled. Really. Imagine me jumping up and down while I write this.

      Fujita's method works, but it needs much love and care and patience. Katya Vinogradowa very patiently and persistently did all the preparative work (she grew the crystals, she soaked them in cyclohexane and she soaked them in the target compound), and, as you may imagine, it did not work the first time, nor the second time. This method is great, but it is not easy to use and the structure determination of such guest-host structures is nothing for inexperienced crystallographers. There is too much room for error and also ample opportunity to make a structure look nicer than it is (which means that the crystallographer may face a few ethical challenges when refining one of those Fujita structures).

    3. Thanks for the insight! I can believe that it didn't work at first, does anything ever? ;) Kudos for sticking with it.

      If I recall, when the first Fujita paper came out, we didn't have much luck... But I need to go question my labmates about which step didn't work, I suspect with the new Nature Protocols paper we might have more luck.

    4. An important (and sometimes overlooked) detail is the selection of the crystals for soaking. Many people make the mistake to select large, good-looking crystals for soaking, but one wants crystals with a large surface-to-volume ratio, so that smaller, needle shaped crystals are actually better than big blocks. Katya and I spent much time at the microscope selecting crystals before soaking them in cyclohexane and later-on in the target compound. I think this was time spent well.

    5. Peter, from the way you described it, it almost sounds like the way protein crystallographers/structural biologists would do their crystallography work...

    6. Interesting you'd say that. It did not occur to me, but you are correct. In an earlier incarnation I spent three years as a protein crystallographer (my postdoc time...) and I have always thought that both groups, small molecule and protein crystallographers, can learn a lot from one another.