Over the past several days, I have been suspiciously absent from the blogosphere. I was on a lecture trip to California, during which I visited the University of California, Santa Cruz and the University of California, San Francisco. The usual thing: you arrive in the morning, you give your talk, and you get exposed to the latest developments in other labs. Nothing can beat this way to spend time. I am very grateful to my hosts – Professor Scott Lokey in Santa Cruz (http://www.chemistry.ucsc.edu/faculty/singleton.php?&singleton=true&cruz_id=slokey) and Professor Jack Taunton (http://cmp.ucsf.edu/faculty/jack-taunton/) in San Francisco. Tonight I will talk about Scott. He and I share a common passion for macrocycles. I think you are all familiar with the byzantine difficulties in forcing cyclic peptide molecules to “behave”. I refer to the chasm that exists between small molecules and cyclic peptides when it comes to drug-like properties. Scott’s lab is well known for its findings that have been reverberating through the scientific community. In particular, I refer to his ongoing research that shows how important hydrogen bonds are in maintaining the conformations of complex macrocycles. Here is a cool example from the Lokey lab that tells you that we are still far from understanding this class of compounds. What you see is that the serine derivative has 96 mL min-1 kg-1 RLM clearance (RLM: rat liver microsomes), whereas the threonine-containing congener is substantially less stable (44 mL min-1 kg-1 RLM clearance). Oddly enough, the serine-containing peptide actually has an oral bioavailability of only 2% compared to the threonine-containing peptide which is 23.8% orally bioavailable. The difference between these two molecules is just one methyl group…
Click to access c2md20203d.pdf
This example underscores the highly empirical nature of efforts to identify orally bioavailable macrocycles and suggests that finding a correlation between oral bioavailability and scaffold design is likely to be challenging. I think it will take us all a long time to understand the intricate factors that turn cyclic peptides into drug-like molecules. I have no doubt that it will eventually happen and, when it does happen, we will probably collectively quote Winston Churchill, who famously said: “We always come to the right decision, having tried everything else first”.
On this Halloween night, let’s talk about something that has bugged me for a while. I recently realized that I tend to make contradictory remarks in my discussions with faculty colleagues and students. Most of these polarizing statements come out unintentionally. For example, I often talk about my lab’s interest in chirality. In doing so, I naturally imply and state the importance of asymmetry in drug design. Interestingly, as I switch to discussing some elements of my lab’s joint work with the Structural Genomics Consortium (SGC), I am forced to remember about our recent findings that chiral fragments are underperforming in our search for protein binders. Just to remind you about what we do: we run soaking experiments that are aimed at identifying small molecule fragments that bind to proteins. We literally take cocktails of small molecules, soak protein crystals in them, and occasionally get co-crystals. Peter Brown’s group at SGC is doing some really nice work in this regard. As I already mentioned, we have had comparatively little luck with the so-called “3D fragments”, or molecules that are more complex by virtue of having chiral centers. So tell me why should I, in a scientific discourse, continue to overstate the importance of chiral compounds? I have a contradiction here, ladies and gentlemen.
Molecular complexity is important, but mainly in process research, at a stage when one needs to prepare large amounts of a known (potentially complex) target. The corresponding molecule has likely emerged from iterative rounds of optimization that have inevitably led to increased molecular weight and structural complexity. On the other hand, the track record of chiral molecules at the discovery stage is not too impressive. What I just mentioned extends beyond fragment screening. In fact, if we go back 12 years, we find an interesting report by Hann and colleagues that suggests that collections enriched in very complex molecules generally have a low chance of individual molecules binding to protein targets. The authors suggest that it is far better to start with less complex molecules and increase the potency by increasing the complexity. These findings are correlated with experimental observations we have recently made in fragment screening (and may publish at an opportune time). For me, the implication is clear: avoid chiral centers and complex structures early on. Those who think that complexity favors discovery are profoundly misled. Here is that thought-provoking Hann paper:
There is an interesting consortium in the UK, called 3Dfrag (http://www.3dfrag.org), whose stated objective is to exploit complex chiral structures in fragment screening. I would be very interested in seeing publications that are hopefully going to come out of their work in the future. For now, I am not convinced that there is definitive data suggesting that chiral molecules enable discovery. So, let’s turn it down a notch with overzealous statements about asymmetric catalysis. Prove me wrong, though, by all means.