Vast fields of biologically active chemical space

Exactly 17 years ago, I highlighted a paper from Brian Shoichet’s lab in which he hypothesized that only tiny regions of chemical space are biologically relevant. If so, Practical Fragments quoted, “a major reason why the screening of synthetic compounds ever finds notable hits is that our libraries are biased toward the sort of molecules that proteins have evolved to recognize.” The best scientists continually question their assumptions, and this is exactly what Brian, Bryan Roth, and collaborators at University of California San Francisco and University of North Carolina at Chapel Hill have done in a new open-access J. Med. Chem. paper.

 

The idea that molecules more closely resembling metabolites and natural products would be more likely to be biologically active is reasonable given the promiscuity of many naturally occurring molecules. As the researchers point out, dopamine signals through 5 receptors, while serotonin binds to 14. Nature is the ultimate recycler, constantly reusing chemical motifs.

 

But since 2009 there has been a massive growth of make-on-demand libraries, and these have increasingly diverged from known molecules. Brian has been among the most prolific explorers of this new chemical space, and as he noted in his keynote at DDC 2026, these gargantuan libraries are yielding more hits against more targets.

 

Are these hits more selective? In other words, are molecules that are less “bio-like” likely to bind to fewer targets? This is the question the new paper addresses, focusing on the 5-HT_2A serotonin receptor (5-HT_2AR), the target for a variety of drugs including psychedelics. The researchers have extensively studied this GPCR; we wrote about a successful computational screen here. In the new paper, the researchers compared hits from a previous “make-on-demand” (MoD) library screen of 1.6 billion molecules with a new screen of 3.5 million “in-stock” molecules. As predicted, the in-stock compounds were much more bio-like than the MoD compounds as assessed by several metrics (Tanimoto similarity, Avalon fingerprints, and RDK7 fingerprints, for the cheminformatics aficionados among you).

 

Both libraries were computationally screened using DOCK3.8 against 5-HT_2AR. Of the top scoring hits, 85 molecules from the in-stock library and 384 molecules from the MoD library were tested in a radio-ligand displacement assay. Hit rates were nearly identical at around 24% each. Functional assays revealed the hits to be active, with some acting as agonists while others were antagonists.

 

When tested against related GPCRs, specifically 5-HT_2BR and 5-HT_2CR, the in-stock molecules were more promiscuous than the MoD molecules. But when tested against a panel of 318 GPCRs there were no differences: the in-stock molecules bound on average 1.7% of the receptors while the MoD molecules bound on average 2.1%. The similar promiscuity persisted even when focusing on the 55 aminergic GPCRs (5-HT_2AR is an aminergic GPCR). In fact, the most promiscuous compound of all came from the MoD set and is distinctly not bio-like.

 

Of course, this is just one study, but I suspect it will generalize. In 2009 I questioned whether biologically relevant chemical space is really so sparse. As I wrote then, “consider a vast field of some crop that can only be harvested by night. There are lights scattered haphazardly throughout the field. One might expect that the crops immediately under the lamp posts would be harvested more intensively than crops in darker parts of the field, even if other areas are equally productive. In this scenario, the lamp posts reveal natural products and similar molecules, but much – or even most – of (unlit) chemical space may also be biologically active, it just hasn’t been sampled yet.”

 

There is grandeur in this view of chemical space, which is supported by the new paper. Imagine yourself in the midst of the field on a moonless night. Turn on your headlamp and step forward in any direction. Drug leads glimmer as far as your light can reach.

An unexpected lysine-reactive covalent fragment against a Pleckstrin Homology domain

Last week we highlighted a study in which an attempt to optimize a covalent fragment led to a non-covalent fragment that bound in a different location. A new open-access paper in J. Med. Chem. by David Spring, Marko Hyvönen and colleagues at University of Cambridge is almost the inverse – a surprisingly covalent binder from a conventional fragment screen.

 

The researchers were interested in Pleckstrin Homology (PH) domains. The 250 or so PH domains in humans bind to the intracellular side of the plasma membrane by interacting with phosphoinositides such as PIP3. Blocking membrane recruitment could be useful for multiple diseases, but the highly charged nature of the interaction (the “P3” in PIP3 means three phosphate groups) makes drugging PH domains difficult, a perfect challenge for fragments.

 

The study focuses on the PH domain from Bruton’s Tyrosine Kinase (BTK), an important oncology and immunology target with several approved drugs, all of which bind to BTK’s kinase domain. Its PH domain was screened against 720 fragments using differential scanning fluorimetry (DSF). Seven fragments stabilized the melting temperature by a whopping 5 degrees C or more. Crystallography was successful for compound 1, which surprisingly revealed that the ketone reacts with the terminal amine of lysine 12, which normally makes electrostatic interactions with two phosphates in phosphoinositides.

 

The crystal structure showed unoccupied space nearby; subsequent fragment growing and optimization led to compound 24, with low micromolar affinity as assessed by DSF. The researchers obtained some two dozen crystal structures, which they were able to correlate with structure-activity relationships (SAR). All of the molecules formed the imine with K12; reducing the ketone to an alcohol abolished stabilization in DSF.

 

Intact protein mass spectrometry was used for assessing SAR and confirming that the covalent bond formation is reversible: adding two fragments with different affinities led to the same distribution of products regardless of the order of addition of the fragments. Not surprisingly, the reaction was faster at higher pH, but nearly complete covalent modification still took place within an hour at pH 7.4.

 

Even at high compound to protein ratios the molecules largely gave single protein modification, and two separate computational studies of the pK_a values of the 15 lysine residues in the BTK PH domain revealed that K12 is an outlier, with a calculated pK_a of 7.1 or 8.8 compared to an average of 10.5 or 10.8 for the others. This means that K12 is largely unprotonated at physiological pH, increasing its reactivity. A similar analysis of four related PH domains suggested that the equivalent lysine residues also have anomalously low pK_a values.

 

The fact that K12 is conserved across related PH domains does raise the question as to whether selectivity will be possible with this type of ligand. Also, no in vitro ADME properties are provided, so it is not clear whether this warhead is advanceable. I wish the researchers had explored other heterocycle alternatives to the furan moiety to assess the tunability of the reactivity; perhaps those will come later. Overall though, this paper is a nice case study where following up on unexpected observations identified a new approach for covalently targeting lysine residues. As Isaac Asimov famously said, “The most exciting phrase to hear in science … is not ‘Eureka!’ but ‘That’s funny.’”

From covalent to noncovalent 14-3-3 modulator, unintentionally

The seven members of the 14-3-3 family act as “hub proteins” that bind to other proteins, bringing them together or affecting their subcellular localization. Some of the client proteins, such as estrogen receptor alpha (ERα), are associated with diseases such as cancer, and stabilizing the interaction with 14-3-3 could thus be useful therapeutically. The idea is that a small “molecular glue” could bind at the interface between 14-3-3 and a client protein, enhancing the interaction. We’ve written here, here, and here about successful examples, both covalent and noncovalent. An open-access paper in ACS Med. Chem. Lett. earlier this year by Richard Doveston and colleagues at University of Leiscester starts with a similar strategy, but ends up somewhere else entirely.

 

The researchers had previously found that WR-1065, the active metabolite of the approved drug amifostine, can covalently bind to C38 of 14-3-3σ. (This is the same cysteine residue that had been targeted by Michelle Arkin and coworkers in 2019 using tethering.) Disulfide bond formation between WR-1065 and 14-3-3σ enhances the affinity of ERα for 14-3-3σ by a modest 2.8-fold.

 

In the new paper, the Doveston group made a small set of analogs in which the thiol was replaced by other warheads. Compound 7, containing an acrylamide, improved the affinity of ERα for 14-3-3σ by nearly two orders of magnitude, from 206 to 2.8 nM. But surprisingly, mass spectrometry revealed no covalent modification. Indeed, compound 7 was also active at low micromolar concentrations against the C38A mutant form of 14-3-3σ, while WR-1065 was inactive.

An examination of the structure-activity relationships (SAR) revealed that the acrylamide was essential; reducing the double bond abolished activity, while replacing the amide nitrogen with an oxygen made it significantly less active. Removing the primary amine or replacing it with a hydroxyl or carboxylic acid was also not tolerated.

 

If compound 7 does not bind covalently to C38, how does it work? Surprisingly, the molecule does not even bind at the interface between 14-3-3σ and ERα. Indeed, it shows additivity with a well-characterized molecular glue (fusicoccin A) known to bind at the interface. Where exactly compound 7 does bind is uncertain. A series of biophysical experiments suggests that it may stabilize the dimeric form of 14-3-3σ, though the mechanism remains to be determined.

 

This paper is a useful reminder that, as the authors conclude, “even rational design approaches can lead to unexpected outcomes.” Sometimes a warhead does not behave as one.