Bad Behaviour

In my lab we do a lot of peptide work with a particular emphasis on structure determination. This is a multistage and complex process that requires our utmost attention to detail, and is subject to many stumbling blocks of highly technical nature. The often capricious and solvent-dependent solubility and aggregation of peptides are enough to make me cringe. I can recall a number of occasions where we fell victims to this trap: we would change one amino acid in a sequence and record some utterly unanticipated behavior. For instance, one sequence would be soluble and well-behaved, whereas its close relative would be a complete “dud”. Welcome to the world of peptides! No wonder there is American Peptide Society out there… Have you heard of American Alkene Society? Or American Bromine Society? Nope, me neither. It tells you that peptide chemistry, as a field, is in a class of its own.

But let’s go back to some tangible ways of tackling peptide behavior, in particular peptide aggregation and solubility. I was reading a 5-year old paper by Geyer and colleagues in OBC, and came across the use of SDS in order to prevent self-aggregation of peptides. One of the sequences reported in this manuscript corresponded to a ten amino acid peptide with a high antiparallel amphipathic beta-sheet content. This peptide was inclined to self-aggregate in aqueous solution. Fortunately, this undesired behavior was effectively countered by adding perdeuterated sodium dodecyl sulfate to the aqueous phosphate buffer. Under these conditions, the precipitation was prevented, allowing the authors to study their molecules using NMR. I am going to discuss this property with my lab and see how SDS might help with some of our intractable cyclic peptide NMRs that occasionally raise their ugly heads. I think we need to compile a list of the magical additives that help avoid the pitfalls of peptide self-association. If you are aware of tricks along these lines, please let me know.

Imagehttp://pubs.rsc.org/en/content/articlehtml/2010/ob/b917549k

Science blogging, the way I see it

Hello everyone. Many of my friends, students, and colleagues have been asking me about how I feel about this whole blogging thing. I like it, that’s for sure. A more detailed set of reasons why I think it is worth my time is outlined in the following article that I put together. It just came out yesterday on the Chemical and Engineering News website (thanks a lot, Carmen Drahl, for making it happen!):

http://cenblog.org/grand-central/2014/03/guest-post-why-i-am-blogging-on-amphoteros-com-by-andrei-yudin/

Have a great weekend.

In praise of C1 stitching units

When it comes to the so-called “C1 building blocks” (in other words, building blocks that lead to straightforward transfer of one carbon unit), formaldehyde is on the very top of my list. This simple chemical acts as a direct precursor to the methylamine fragment that can be put together using a range of condensation reactions. Formaldehyde-derived iminium ions are fantastic starting materials for the Mannich reaction and are known to partake in cascade processes driven by the incipient primary amine units. In this regard, I enjoyed reading the total synthesis of lyconadins A and C by the Dai group of Purdue University. The cornerstone of their synthesis (shown below) demonstrates how formaldehyde takes on a scaffold containing a ketone and an alpha,beta unsaturated ketone, and stitches a complex multi-ring system in one pot. There are three distinct mechanisms one can propose for this process. One of them – the aza-Michael/Mannich sequence – can be safely ruled out because the control experiment using the secondary amine intermediate shown below, failed. This leaves two possibilities, neither of which can be ruled out at this point of time. Overall, this gram-scale route opens doors towards libraries of lyconadin derivatives and their evaluation as anti-neurodegenerative agents.

Image

http://onlinelibrary.wiley.com/doi/10.1002/anie.201400416/abstract

Get your rings in order

The title of the paper by Richard Taylor and colleagues speaks for itself: “Rings in Drugs”. Below you can see the top 6 rings from marketed pharmaceuticals. I am also showing a morpholine (number 29) and a beta-lactone (the last ring on the list of top 100). Macrocycles were not considered in this study as maximum single ring size was set at 9. There are several interesting “ring” anecdotes in this paper. For instance, approximately six new ring systems enter drug space annually. Here is a somewhat surprising fact for those of us who are interested in asymmetric synthesis: 40% of drugs do not contain any sp3 carbons in a ring system.

Image

http://pubs.acs.org/doi/abs/10.1021/jm4017625

In my view, this paper provides a compelling rationale for going after novel heterocycles. In this regard, I recalled an interesting study published by Erick Carreira’s lab several years ago. In it, the authors made a point about a novel spirocyclic ring that had subtly perturbed the established electronic structure of morpholine. In the graphic below, you can see the relative orientation of the oxygen lone pairs that is clearly different in the two structures. Due to the fact that biological activity of a small molecule can often be traced back to the vectorial relationship of electron pairs on heteroatoms at key positions, the oxetane unit has been employed to expand the chemical space around morpholine. Because morpholine is common (number 29 in the Taylor study), its oxetanyl analogues have found wide use. And so shall many other, yet to be identified, heterocycles that are “vectorially challenged”!

Image

http://onlinelibrary.wiley.com/doi/10.1002/anie.200800450/abstract

Electrophilic fragments: small but mean

The idea of ligand efficiency as a metric for comparing new heterocyclic motifs has always intrigued me. Things get particularly interesting when heterocyles are combined with the elements of reversible covalent inhibition. In his 2013 paper in JACS, Taunton and colleagues described a series of reversible covalent inhibitors of MSK/RSK-family kinases that contain noncatalytic cysteine residue close to the active site. In fact, C436 is found in only 11 of 518 human kinases (so there are reasons to go after this cysteine). The acrylamide fragments described in the Taunton paper were later used to develop potent kinase inhibitors, underscoring the fact that ligand efficiency of reversible covalent fragments was sufficient for further elaboration. Below you see a view I created using PyMol, showing how the indazole scaffold of the acrylamide inhibitor forms the expected hydrogen bond pattern with the hinge region. The authors point out that the indazole core does not extend beyond the gatekeeper T493.  Interestingly, unfolding of the indazole fragment/RSK2 adduct with guanidinium-HCl resulted in quantitative recovery of the fragment, indicating that the covalent bond was formed reversibly. It will be interesting to see if good levels of kinetic discrimination can be achieved with reversible covalent inhibitors. From my point of view, one of the most interesting lessons offered by this study is that potency is not solely driven by the free energy of covalent bond formation.

Image

http://pubs.acs.org/doi/abs/10.1021/ja401221b

Poking at the Diels-Alder reaction

I am sitting on my Oakville-bound train that is about to depart Union Station in Toronto. This winter has been super cold in our part of the country. It’s kind of funny because I have been hearing way less about global warming on the news. Weird, eh? I guess we’ll wait for a couple of months for people to start complaining. OK, I am being a troll.

Tonight I want to talk about the venerable Diels-Alder reaction. There is no need to praise it beyond the superstar status it already has. Instead of empty accolades, I will pay a facts-based tribute to this process and, in the spirit of recent discussions, try to poke at it.

Are there 6-membered rings that cannot be made using this reaction? I can’t think of too many. The classroom value of Diels-Alder reaction is also undisputed: we beat this reaction to death when we teach frontier molecular orbitals (FMO) method, to the extent that some students leave our classes with an impression that this cycloaddition (and perhaps some electrocyclizations and sigmatropic reactions) defines the FMO theory in its entirety, which is not true at all. Nonetheless, this is still one of the most awe-inspiring reactions out there. To challenge its bulletproof status, one might want to subject Diels-Alder reaction to the limits of angular strain, hoping that that the cycloaddition might “crack”. Time and again, though, this resilient reaction has surprised us in most admirable ways. Take a look at one of my favorite papers on this subject. This is a study published by Dirk Trauner and Ken Houk some years ago. I would not have expected that imposing such severe strain on the 6-membered transition state would deliver any reasonable outcome. But the reaction works at 30 oC. There are some other interesting insights offered by this paper, so check it out. I think the lesson here might be that no matter how strain-crazy your idea might be, you should just give it a shot if it involves the magical 6p-electron transition state.

Image

http://pubs.acs.org/doi/abs/10.1021/ja050135a

Now… Keeping up with the flavor of my recent post on heteroatom-heteroatom bonds, here is another viewpoint from my “vault of near-impossibles.” While there are countless examples of C-C and C-heteroatom bond formations using Diels-Alder reaction, it is interesting to note that heteroatom-heteroatom connections aren’t really made using this process. If you have a good example – please let me know. There is probably a fairly decent energetic argument against the transition state that produces a link between two heteroatoms (I should ask Ken Houk about this). Overall, I feel a bit better about showing that not everything is hunky-dory in the Diels-Alder bag of tricks. I will feel this way until you guys show me that I am wrong (or will you?).

Stitching with radicals

Tandem cyclizations that result in rapid formation of complex molecular skeletons have always attracted my attention, especially when fairly unusual intermediates are implicated in the corresponding reactions. Below you see a really cool sequence recently reported by Jennifer Stockdill of Wayne State University. The reaction targets the tricyclic system of daphniyunnine (what a mouthful…). The first step is N-chlorination, which is achieved by the use of NCS (N-chlorosuccinimide). N-chloroamines are interesting synthetic intermediates that can get tantalizingly close to losing the elements of HCl upon treatment with base, yet are often surprisingly stable and isolable. Upon alcohol oxidation in the example below, the tricyclic system is set up by way of a tandem radical cyclization, which starts off the aminyl radical. The authors highlight the neutral nature of the aminyl radical undergoing 6-exo cyclization in their sequence. It will be interesting to see a completed synthesis (hopefully some time soon).

Image

http://pubs.acs.org/doi/abs/10.1021/ol4034868