The enduring concept of isolobal relationships

A couple of weeks ago, I heard one of the most interesting lectures of the past year. Prof. Alcarazo, now at the University of Göttingen, was visiting our department as the external examiner of one of Doug Stephan’s PhD students. From his talk, I learned about the surprising reactivity of thiourea, known in all sorts of roles – from heterocycle synthesis to asymmetric catalysis. The novel reactivity shown in Figure has its origins in Roald Hoffman’s teachings on the power of isolobal relationships in chemistry. Alcarazo and his students have extended this enduring concept to the “sulfur version” of hypervalent iodine reagents. As it turns out, a lot of reactions known in organoiodine chemistry can now be carried out using significantly more user-friendly organosulfur compounds, many of which are accessible from thiourea derivatives using a couple of trivial transformations. By way of an example, I am showing the cyanation of N-methyl indole.

2.jpg

http://pubs.acs.org/doi/abs/10.1021/jacs.5b05287

Inspirational chemistry

Every once in a while, we all want to read something inspirational. Alas, we like different things and our choices reflect personal preferences, dogmas, and current fads. There are people who somehow get existential meaning out of “transition metal-free X”, with X being pretty much anything… Some other folks get a kick out of site-specific modification of amino acids in proteins… How about science coming out of places that have no business producing anything meaningful because they are entrenched in conflict and corruption? I like that. Below is a paper co-authored by a team of scientists from Enamine in Kiev, Ukraine. The fact that these individuals are able to produce science of this caliber under the conditions they are currently in, is quite admirable. The Org. Lett. paper describes a very counterintuitive participation of CF3 diazomethane in reactions with nucleophiles. Effectively, nucleophilic additions of certain nucleophiles appears to result in the attack at nitrogen, which goes counter to everything we know about diazomethane chemistry. As a result, a series of interesting transformations are enabled and I would call it a method par excellence for producing heterocycles of medicinal importance. I am at a loss as to why this transformation of diazo functionality has remained veiled through all these years…

2.jpg

http://pubs.acs.org/doi/abs/10.1021/acs.orglett.6b01565

At long last – an HDAC structure with a cyclic peptide inhibitor

Not too long ago, Christianson and colleagues published a notable paper in Nature Chemical Biology. It describes the molecular basis of catalysis and inhibition of histone deacetylase 6 (HDAC 6) and uses several small-to-medium sized probes to investigate this enzyme. Naturally, my attention was focused on the exciting co-crystal structure of HDAC 6 and HC toxin, which is well-known covalent cyclic peptide inhibitor. HDAC 6 comprises two tandem catalytic domains. One of them is specific for substrates bearing C-terminal acetyllysine residues. Now that we finally have a molecular-level view of a cyclic peptide inhibitor/HDAC interaction, this paper should encourage a new wave of attempts to design selective and potent inhibitors of HDACs. I am not sure I agree with the authors regarding their claim that the cis/trans/cis/trans geometry of the four amide bonds in the HDAC-bound HC toxin is particularly remarkable. I cannot think of anything else that is reasonable, particularly if proline is one of the residues. In fact, there are a number of crystallographically characterized cyclic tetrapeptides that feature exactly this arrangement. There is, nonetheless, an interesting clue regarding achieving selective HDAC inhibition using cyclic peptides: despite the presence of the strictly conserved cysteine 584 residue, HC toxin binding is dominated by zinc interacting with the gem-diol, leaving the epoxide intact. The thiol side chain of cysteine 584 is still well positioned for nucleophilic attack at one of the epoxide carbons, but the authors suggest that inhibitor binding to other HDACs would result in an even closer contact between the nucleophilic SH and the epoxide electrophile, leading to covalent bond formation. This offers an interesting bis(electrophile) selectivity filter.

4.jpg

http://www.nature.com/nchembio/journal/vaop/ncurrent/full/nchembio.2134.html?WT.feed_name=subjects_structure-determination

 

Reactivity and selectivity

On a recent trip to Munich, I got acquainted with the research of Professor Hendrik Zipse. His mechanistic understanding of catalysis of alcohol silylation is both deep and educational. It serves as an instructive reminder of the reactivity/selectivity principle, which is something chemists relate to on an intuitive level. Not too long ago, Zipse and co-workers published a series of papers aimed at understanding the fundamental underpinnings of Corey’s classical silylation of alcohols. Zipse reminds us that DMF was the prescribed solvent in the original system. In this process, imidazole was used to mop up HCl, making TBS transfer one of the most familiar processes in organic chemistry. The question is whether or not the role of all components is crystal clear. It is now, but only after Zipse’s kinetic analysis. In brief, DMF is not your innocent by-stander. Its role is to form the active silylating argent, which is the Lewis acid/base pair shown below. Due to the high activity of this adduct, reactions in DMF (the original solvent from Corey’s 1972 paper) do not show impressive selectivity among primary, secondary, and tertiary alcohols. This is a very important finding. In contrast, if one stays away from DMF/imidazole mixture and runs silylations in dichloromethane along with DMAP and triethylamine, the selectivity is excellent. Improved reaction profile correlates with lower activity of the DMAP-derived active silyl transfer agent.

2.jpg

http://pubs.acs.org/doi/abs/10.1021/acs.orglett.5b01536

Protecting groups in drug structures

Are you aware of any therapeutic agent whose structure contains protecting group(s)? It is as if an intermediate en route to some target molecule was put through screening, revealing that it might be wise to leave the protecting group intact. Talk about the irony of chemical synthesis… I make this point in some of my classes and particularly like to mention the anti-epileptic drug topiramate as an example. There are also some really nice cases among recently approved peptide drugs. I think there is a molecule with a Cbz group at nitrogen, but I can’t recall its name at the moment.

2.jpg

http://pubs.acs.org/doi/abs/10.1021/acsmedchemlett.6b00176

Coming back to the “acetonide-rich” framework of topiramate, I want to mention an outstanding retrospective recently published by Dr. Bruce Maryanoff. He traces the discovery of topiramate to its origins and uses his narrative as an opportunity to remind us that the so-called phenotypic screens (rather popular these days) are nothing new. In fact, they represent the stomping grounds of drug discovery, a once widely accepted practice that eventually gave way to target-based approaches (rather unfortunately). My favorite passage in this paper is when Bruce recalls his fruitless attempts to convince upper management of Johnson&Johnson of the virtues of phenotypic screening. He says: “At the beginning of the 21st century, I felt like a lumbering dinosaur in the halls of the company…”.

The power of low-tech methods

In my wildest dreams I would never have predicted that, one day, PAMPA (parallel artificial membrane permeability assay) would be so meaningful in my lab’s research.  I think it is a superb way to evaluate the capacity of complex structures to undergo conformational rearrangements that result if modulation of polar surface area (PSA). The reason these measurements are all the rage nowadays is that PAMPA data correlates with cellular permeability, which is paramount in chemical biology and drug discovery. Briefly, the idea is to set up an experiment that is sketched below. You can see a small apparatus that contains two compartments separated by a hydrophobic membrane. Once injected into the donor well, a compound will eventually equilibrate and, depending on its balance of properties, will be more or less membrane-permeable. The ratio of concentrations is eventually converted to a number on a log scale. The lower the number, the better is the chance that the molecule of interest would be capable of traversing hydrophobic cellular membranes. You might ask: why use this contraption in favor of the “real world”? After all, a myriad different cell lines are available, of which Caco-2 is most commonly used in drug discovery. The answer lies in the presence of transporters in all those cells. Transporter proteins work as chaperones, pumping molecules in and out of cells. As a result, if one relies on cellular assays, too many false negatives would be generated. Interesting molecules might then be discarded too early. This is why researchers have embraced PAMPA as the go-to tool that is not confounded by nature’s tricks of active transport.

Synthetic chemists love to think in terms of intricate three-dimensional arrangements of atoms in their molecules, so how can PAMPA data relate to structure?  This is exactly why I used to look down on seemingly “low-tech” methods such as PAMPA, but I have changed my mind. We recently published a paper where we were able to show how our macrocycles exercise control over the formation of hydrogen bond networks. You can see a comparison between one of our molecules on the left and the corresponding homodetic control. The linker region is our patented trick to modulate PSA. Dr. Jen Hickey of Encycle Therapeutics deserves a ton of credit for bringing this study to fruition.

2.jpg

http://pubs.acs.org/doi/abs/10.1021/acs.jmedchem.6b00222

Cannabinoids

A couple of weeks ago I gave a talk at the Gordon Conference on Heterocycles (https://www.grc.org/programs.aspx?id=11391) and got to see my old friend Professor Erick Carreira of ETH (and, of course, a ton of other cool people). Erick gave a really interesting lecture, which was largely dedicated to his lab’s ongoing interests in developing synthetic routes to all possible diastereomers of a given chiral molecule. As you all know, it is the enantioselectivity that is easy: if you need the opposite enantiomer of a given molecule for which an asymmetric route exists, you choose the corresponding enantiomer of your catalyst. In contrast, the development of synthetic strategies that provide controlled access to any given stereoisomer of a target with multiple stereocenters, is in its infancy.

In his talk, Erick described an innovative strategy to access all stereoisomers of cannabinoid receptor activators. The route involves simultaneous use of two chiral catalysts, each of which possessing full control over the configuration of one of the product stereocenters. For their purposes, the Carrera team used a combination of iridium and enamine catalysis. Asymmetric matters aside, I truly enjoyed the way the core of the target was constructed. You can see a brief summary of the route below. There are countless other examples of cannabinoid total synthesis and they look nothing like this original contribution. I particularly like the last step. Stars indicate the fact that all stereoisomers can be prepared at will by choosing the right combination of catalysts.

2.jpg

http://onlinelibrary.wiley.com/doi/10.1002/anie.201408380/abstract