Although Serge Zaretsky defended his PhD thesis and left my lab almost 10 months ago, we still continue our fruitful cooperation. During his graduate stay, Serge was working in the area of macrocyclization reactions driven by amphoteric aziridine aldehydes. Mechanistically, this turned out to be a difficult problem. Right now, this process is straightforward and we make macrocycles rather efficiently, but the reaction itself is still an enigma. My other former PhD student, Ben Rotstein, once likened this macrocyclization to an onion. There are just so many layers and you don’t know where to stop. While we all know that at some point we must stop (hoping all of the findings are consistent with some logic), there is one extra experiment that throws it all down the drain. Convergence is not something we are inherently good at in science. And this is for the better.

In Serge’s case, we got to the bottom of some key selectivity controls. The imidoanhydride species you see below is the most important constituent of this process. We think that this intermediate accounts for the reactivity trends we observe. A paper describing this work just appeared in Chemical Science. I am providing a link as well as a photo of Serge (he’s wearing a light-green shirt) having Jägerbombs after his graduation. We sat down in our group room, raised our drinks, and sincerely thanked Serge for his contributions.

http://pubs.rsc.org/en/content/articlelanding/2015/sc/c5sc01958c#!divAbstract

I just came from a short vacation to Bermuda. Speaking of vacations, what would we do without them? I find it exceptionally annoying when graduate students or postdocs refuse to take them, which might sound strange to some of you, especially if you think that professors are all about driving people mad in the lab. If you subscribe to this view, I point you to the following paper (this work should be read by everyone):

http://www.ecologyandsociety.org/vol20/iss2/art3/

The authors make a point about two distinct processes that lead to creative thinking. There is this fast system-I, which produces intuition. There is also the more deliberate system-II, which is more analytical and leads to what we call “reasoning”. The point of this paper by Scheffer and colleagues is that the two systems are stimulated under different conditions.

System-I works unconsciously and is based on instantaneous associations, whereas system-II keeps it all in check and modifies the results coming from system-I. Here is the problem: traditional work expectations overemphasize system-II, which makes some sense if you think that decision making is all about rationality. This is wrong.

Unstructured associative thinking is critical to creativity because our web of associations is extremely complex. The authors provide a great example: “banana” is linked to “fruit”, to “yellow”, to taste, and so on. Once you have triggered one idea, others are activated and cause ripples through your mind web. It turns out that the same person would be better or worse at generating remote associations based on his/her mood and state of mind. A cheerful state tends to result in novel associations. Because intuitions arising from system-I are often wrong, there certainly needs to be a substantial contribution from system-II. But not too much… Apparently, Feynman deliberately tried to stay away from knowing previous explanations for various phenomena he had been interested in.

In closing, there is scientific evidence that suggests that creativity is directly related to our ability to trigger associations between unrelated items (and we may not even anticipate in advance which ones they happen to be). For that one needs to be relaxed. Going for a nice walk in the woods could be way better than staying in office (or lab).

I am always on the lookout for ways to convert one class of heterocycle into another. I think this is easier than building rings de novo. One can reasonably expect scaffold hopping approaches in fragment-based discovery to be greatly simplified with such metamorphosis-like pathways. Today I want to discuss one particular type of molecule that is as suited for this general idea, or is better than, any other heterocycle. I refer to indole, whose versatility is astounding. One recent sequence that attracted my attention is the one reported by Guan and colleagues. Here we have a Baeyer-Villiger-like oxidation of indole to give benzoxazinone nucleus.

http://pubs.rsc.org/en/Content/ArticleLanding/2013/CC/c3cc44215b

This reaction was accomplished using oxone, which ruptured the indole moiety. There was another recent report, by Cui and co-workers, who accomplished a fairly similar ring-expansive goal, except they used indoles and amines under copper catalysis and ended up with arylquinazolinones (http://pubs.acs.org/doi/pdf/10.1021/acs.joc.5b00957 ).

There is a recent report that documents dose-dependent disruption of amyloid fibryls by tabersonine, a terpene indole alkaloid. The paper might be interesting for a couple of reasons. It is rare to see molecular-level evidence for an interaction between alkaloids and peptides. It is also interesting to trace back the origins of the material used by the authors. Here are the initial stages of how Zhou and co-workers got their stuff: “Voacanga africana beans (100 g) were ground and crushed into powders, which were added into 300 mL of 95% ethanol and soaked at 65 °C for 1 h…”

Below I am showing a scheme that involves the monomer of A-beta peptide interacting with tabersonine. The evidence in the paper points towards the preferred interaction between tabersonine and the aggregated form of peptide, but it is easier to draw the monomer.

http://pubs.acs.org/doi/pdf/10.1021/acschemneuro.5b00015

While there is no real chemistry in this paper (and why should there be if you can get your Voacanga africana beans?), I did recall one of my favorite syntheses of tabersonine. It was published by my good old friend S. Kozmin (now at the University of Chicago) when he was a graduate student in Viresh Rawal’s lab. The paper features a Diels-Alder approach to the central ring and makes use of the amine variant of the Danishefsky’s diene, developed by Rawal and Kozmin at the end of 1990’s.

Coming back to the disruption of amyloids, I am not sure tabersonine will see any sort of medicinal use, although it is encouraging to keep in mind that alkaloids are often very blood/brain barrier-permeable. Maybe nature is indeed trying to tell us something here?

http://pubs.acs.org/doi/pdf/10.1021/ja983198k

It is great to witness your students succeed, especially when you meet them at conferences some years after graduation and learn about their notable accomplishments. As Ved Srivastava and I were running the American Peptide Symposium in Florida last week, I ran across Naila Assem and heard about her recent work. Naila recently accepted a job at Novozymes in San Francisco (http://www.novozymes.com/en/Pages/default.aspx), which is excellent. I asked her to do a guest post on her recent work with Phil Dawson at Scripps. Here it is, verbatim:

After working on peptidomimetic ligation in the Yudin lab (http://www.nature.com/nprot/journal/v7/n7/abs/nprot.2012.066.html), pursuing my post-doctorate with a pioneer in the peptide ligation field, such as Dr. Philip Dawson, seemed like the logical next step for me. Upon completing my Ph.D., I packed my bags and joined the Dawson lab at the Scripps Research Institute. One of the things I have greatly enjoyed about working in the Dawson lab is how multidisciplinary the projects have been. While working on synthesizing small peptide mimics of a highly conserved epitope on the Hepatitis C virus, I came across dichloroacetone (DCA), a cross-linking reagent. On today’s post I would like to highlight the utility of DCA, which is the subject of a recently published paper in Angewandte Chemie (http://onlinelibrary.wiley.com/doi/10.1002/anie.201502607/abstract).

Since short peptides do not typically form stable secondary structures, the Dawson lab utilizes synthetic techniques to display small peptides in a biologically relevant manner. In our work with Hepatitis C we found that DCA was effective at cyclizing two cysteine side chains. From this observation we wondered if we could use DCA to stabilize helical secondary structure. What further attracted us to DCA was the ketone within the akyl chain that could potentially be used as a site for functionalization. We first tested to see if dichloroacetone could be used to cross-link two cysteine side chains, in the i and i+4 positions, to induce helical formation. Although the cross-linking reaction between two cysteine side chains proved to be selective and high yielding we found that it did not give enough slack for proper helix formation. On the other hand, cross-linking between two homo-cysteines gave proper helix-stabilization.

Once we were able to establish helicity we explored the ketone’s ability to undergo oxime ligation. We first tested typical aniline catalyzed oxime ligation conditions and found full conversion to the desired product after 16 hours. We were able to demonstrate the ability to tag the cross-linker with a diverse set of labels including fluorophores such a Alexafluor 647 and Alexafluor 488. We were also able to biotynilate or add peptide tags such as a poly-arginine or FLAG-tag. Additionally, we could dimerize the helical peptide with a bis-aminooxy linker or simply add an aminooxy functionality that can be used to conjugate on to the surface of carrier proteins or virus like particles.

The utility of DCA in a protein/peptide interaction setting was demonstrated when we synthesized an S-peptide analog and bound it to RNAse S. The DCA cross-linked peptide analog (Ac-KETAAhCKFEhCQHMDS-NH₂) was successfully co-crystallized at 2.2 Å. The cross-linked peptide structure was found to be highly conserved when compared to the natural S-peptide in complex to RNAse S. Additionally, we were able to show that DCA can be used the make peptide cycles and bicycles.

My lab has been interested in new ways to access uncommon substitution motifs in aromatic heterocycles. Here is our rationale: while there are countless examples of imaginative approaches to borylate an existing ring using directed ortho-metallation or C-H activation, the corresponding reactions are governed by the innate reactivity of a given heterocycle. The selectivity also depends on the reagent that effects the site-selective transformation.

We have been thinking about ways to assemble structures using different rules. As part of this undertaking, we placed our bets on amphoteric molecules and explored ways of controlling the selectivity of boron group transfer using simple condensation reactions.

In grey rectangles below you see some of the amphoteric species described by our lab in the past. In green you see a new and surprisingly stable class of compounds described by my postdoctoral fellow Piera Trinchera and graduate student Victoria Corless in a recent Angewandte paper. The dicarbonyl scaffold, prepared using photocatalysis and organocatalysis (special thanks to MacMillan, Stephenson, and Yoon for trailblazing this area) can be transformed into several types of heterocycles with unusual regiochemistry of the C-B bond. I think medicinal chemists might be interested in this technique as there are many molecules one can imagine making using dicarbonyls. I applaud Piera and Victoria’s efforts. I was also glad to see Chemical and Engineering News comment on their research. The next issue of Angewandte will have the cover designed by Piera and Victoria, which is a nice touch (although I am left with a bill for 2100 Euros for their beautiful artwork…).

http://onlinelibrary.wiley.com/doi/10.1002/anie.201504271/abstract

http://cen.acs.org/articles/93/web/2015/06/New-Route-Rare-Heterocycles.html

I would end by pondering over a curious finding that most isomers accessible through our condensative approach are not easily accessible by alternative metal-based techniques. You might ask “Is there a reason for that?”. I would say – there ain’t one, although it is tempting to speculate on the innate selectivity of condensation reactions.

I want to talk about the so-called “rabbit ears”, which is one of the most enduring concepts in organic chemistry. There have been several exciting discussions related to this topic at the American Peptide Society meeting this week.

By way of background, Professors Ron Raines and Dek Woolfson published a thought-provoking paper back in 2010, in which they analyzed the pdb (protein database) and came up with a conclusion that nearby amino acid residues in proteins display n-pi* interaction. I discussed this back in 2013 (https://amphoteros.com/2013/11/27/neighbourly-ties/). According to the Raines/Woolfson analysis, the molecular orbital-based view of the amide oxygen is more adequately described not as A (below), but as B, in which there is a p-like molecular orbital perpendicular to the C-O vector. The authors stated: “Contrary to the expectations of valence shell electron pair repulsion (VSEPR) theory, the two lone pairs of divalent oxygen do not occupy equivalent orbitals that resemble rabbit ears”.

While “B” adequately explains the interaction between adjacent amides seen by Raines and Woolfson, it does not (in my view) provide grounds for saying that the good old rabbit ears concept has no merit. However, this is exactly what some people tend to conclude from their work. What might be done in defense of the poor old rabbit ears? I suppose if there were a study clearly demonstrating the 120^o (or so) angle formed upon Lewis acid coordination to oxygen, it would be convincing enough. There is indeed a paper that shows that: Corey’s 1992 report in Tetrahedron Letters exemplifies the 120^o B-O-C angle both in X-ray and in solution. I am somewhat relieved that the good old VSEPR model is alive and well.

http://www.sciencedirect.com/science/article/pii/S0040403900609024