01 March 2021

Fragments vs MEK1: allosteric binders

MEK1 is a central player in the MAP kinase signaling cascade, which is often dysregulated in cancer. As such the enzyme has been the focus of considerable research and the target of four approved drugs. Interestingly, these drugs bind not to the hinge region targeted by most kinase inhibitors but rather to an allosteric pocket adjacent to the ATP binding site. The drugs also look somewhat alike. Seeking something completely different, Paolo Di Fruscia, Fredrik Edfeldt, Helena Käck, and colleagues at AstraZeneca turned to fragments. They have recently published their results in ACS Med. Chem. Lett.
As we discussed in 2016, the AstraZeneca fragment library is quite large at 15,000 molecules. The researchers used a computational screen to narrow this down to a more manageable 1000 compounds for ligand-detected NMR screening. AMP-PNP, a nonhydrolyzable version of ATP, was included to block the hinge region, biasing the screen for fragments that bind the allosteric site. (See here for earlier work looking for ATP-competitive molecules.) A total of 142 fragments were identified and further characterized by SPR, and 46 showed dissociation constants better than 1 mM and similar affinities in both the presence and absence of AMP-PNP, suggesting they do indeed bind in the allosteric site.
Crystallography was attempted on all the fragments, but only two produced structures. Reassuringly, both bound in the allosteric site. But with only limited structural information, the researchers tested analogs of the fragment hits within their corporate collection. This identified compound 10, which is more potent than initial fragment 3. Moreover, compound 10 lends itself well to library synthesis.

All library members were initially made and tested as racemates. When the two enantiomers of the best hit were separated, compound 23 was found to be a sub-micromolar binder, roughly 100-fold better than the other enantiomer. At this point the researchers finally obtained a crystal structure of compound 23, confirming that it did bind in the allosteric pocket. Compound 23 is also still fragment-sized, just three heavy atoms larger than compound 3.
The astute reader will notice that the word “inhibitor” has not appeared until now, and indeed despite the encouraging affinity no mention is made in the paper of inhibition – a rather important feature! At a conference in 2019 Paolo did describe further optimization to a functional molecule, so hopefully we will see a second publication detailing this work.
Like the NPBWR1 story last month, this is another nice example of advancing fragments in the absence of structural information. It is also a good case study of fragments yielding completely different chemical matter in a crowded field.

22 February 2021

Antifreeze opens cryptic pockets, experimentally and computationally

Imagine an alien species sees a single photo of a human. They would have no idea how our arms and legs move, or that our mouths can open and close. So it is with protein crystal structures: even multiple static images often fail to show possible conformations. Pockets open and close in unexpected places, and these can be critical for drug discovery. But how do you find these “cryptic” pockets? Harsh Bansia, Suryanarayanarao Ramakumar, and collaborators at Indian Institute of Science, Bengaluru and Pennsylvania State University provide a new approach in J. Chem. Inf. Mod.
The researchers were studying a bacterial xylanase called RBSX and had mutated a tryptophan residue to an alanine. When they solved the crystal structure, they found that the mutation had created a surface pocket that was filled with a molecule of 1,2-ethanediol (EDO). EDO is an ingredient in antifreeze because of its ability to prevent ice formation, and this property also makes it a common cryoprotectant in crystallography. The EDO molecule was making both van der Waals contacts as well as a hydrogen bond with the protein. The researchers found similar results when they used propylene glycol. (See here for a related discussion of MiniFrags, the smallest of which are the size of propylene glycol.)
To see whether water could make these same interactions, the researchers determined another crystal structure in the absence of EDO. Surprisingly, a phenylalanine side chain rotated and closed the pocket. Had this been the only structure solved, the possibility of pocket formation would not have been suspected.
Next, the researchers conducted molecular dynamics simulations. Starting from the closed state, the pocket remained occluded by the phenylalanine, giving no hint of its potential presence. Starting from the open state and removing EDO, the pocket also rapidly closed. In other words, in the absence of a ligand, the pocket appears to collapse in upon itself.
Importantly, these observations are not limited to a mutant bacterial protein. The researchers looked at published crystal structures of four unrelated proteins with known cryptic pockets and found that EDO could bind in all of them. They also ran molecular dynamic simulations on two proteins in which EDO was included as a virtual cosolvent. For both NPC-2 and IL-2, addition of EDO was able to open up cryptic pockets that had been previously found using other molecules; we’ve discussed earlier computational work on IL-2 here.
This is a nice example of following up on an unexpected observation, and is well-suited for further study. For example, it would be interesting to do a systematic study of EDO and propylene glycol binding sites throughout the entire protein data bank. For those of you doing molecular dynamics or crystallography, it may be worth adding EDO – virtually or experimentally – to see if it reveals any surprises in your favorite proteins.

15 February 2021

Fragments vs Notum, three ways

The Wnt signaling pathway has been implicated in multiple diseases, from Alzheimer’s to cancer to osteoporosis. To be able to bind receptors, Wnt proteins must be post-translationally modified with a palmitoleate group. The carboxylesterase Notum removes this group, shutting of signaling. Thus, inhibitors of Notum could maintain Wnt activity. In two J. Med. Chem. papers, E. Yvonne Jones (University of Oxford), Paul Fish (University College London), and collaborators describe three series of inhibitors derived from fragments.
The palmitoleoyl group is highly lipophilic, and previous work with Notum had revealed a predilection for hydrophobic carboxylic acids. Thus, in the first paper, the researchers assembled a library of 250 diverse carboxylic acids, all rule-of-three compliant. Each was tested (10-point dose-response) in a biochemical screen up to 100 µM. Twenty compounds had IC50 < 25 µM, and all of these were soaked into crystals of Notum, resulting in 14 structures. Two series were pursued.
Compound 5 was one of three pyrroles with low micromolar potency. Fragment growing led to compound 20, and further SAR ultimately led to compound 20z, with high nanomolar activity. Unfortunately, this is a fairly lipophilic molecule, with clogP = 5.5. Indeed, despite best efforts, including paying close attention to lipophilic ligand efficiency (LLE), potency tracked closely with clogP for this series.
The second series, as represented by compound 8, was less potent but also less lipophilic. Walking various substituents around the phenyl increased both properties (compound 25n) and further tweaking led to compound 26. Although this molecule had promising in vitro ADME properties, it was deprioritized in favor of another series described in the second paper

In addition to the biochemical screen, the researchers also conducted a crystallographic fragment screen at the Diamond Light Source XChem platform. This yielded a whopping 60 hits of the 768 fragments screened, with compound 7 being notable for its high potency, ligand efficiency, and LLE. Iterative structure-based design led ultimately to low nanomolar compound 23dd.
Compound 23dd was active in cell-based assays and had acceptable pharmacokinetic properties in mice. The researchers were particularly interested in modulating Wnt signaling in the brain, and in vitro studies suggested that compound 23dd would have good blood-brain permeability. Unfortunately, this turned out not to be the case, for reasons that are still not clear. However, another molecule derived from compound 7 was superior – hopefully the subject of a future paper.

These are nice structurally-enabled fragment to lead stories, and the medicinal chemistry strategies are particularly well described. Notably, the researchers were able to optimize fragment hits to nanomolar binders while maintaining low molecular weights and (in the second and third cases) reasonable lipophilicity. In addition to clear examples of property-driven medicinal chemistry, these papers illustrate that fragment-based methods can yield a variety of starting points, which can be useful when one lead series runs into trouble.

08 February 2021

Fragments in the clinic: PF-06835919

A year into the COVID-19 pandemic more than 2.3 million people have died, with deaths in the US approaching 500,000. These are staggering numbers, and the scientific community has rapidly responded. Amidst this disaster, it is easy to lose sight of longstanding, even more deadly threats, such as heart disease.
A leading cause of metabolic disease is overconsumption of fructose. Because it is sweeter than other natural sugars and cheap to produce, fructose is widely used in processed foods. Fructose is not subject to the negative feedback regulation of other sugars, and overconsumption has been linked to nonalcoholic fatty liver disease (NAFLD), insulin resistance, and cardiovascular disease. The first step in fructose metabolism is mediated by the enzyme ketohexokinase (KHK), so blocking it seems like a reasonable approach.
More than three years ago we highlighted a paper from Pfizer describing the fragment-based effort which led to compound 1, an inhibitor of KHK. That post ended by noting that there was “still some way to go” to reach a drug. A paper published late last year in J. Med. Chem. by Kentaro Futatsugi and colleagues from Pfizer describes the journey to the clinic.

Compound 1 was well-suited to SAR by parallel synthesis, and a variety of replacements for the methylpyrrolidine (on top) led to compound 3. Although this molecule had similar affinity as compound 1, a crystal structure revealed that it had shifted its binding mode such that the other pyrrolidine ring was pointing towards an important arginine residue. Exploring a diverse range of replacements led to compound 4, with improved affinity driven in part due to interactions between the hydroxyl and the arginine side chain. Replacing this hydroxyl with a carboxylic acid led at last to a low nanomolar lead.
Compound 6 was unstable when incubated with human hepatocytes, and various studies revealed that glucuronidation at the remaining hydroxyl was responsible. Removing the hydroxyl and lowering lipophilicity by removing the nitrile ultimately led to PF-06835919. This compound is potent, orally bioavailable, and clean in a variety of off-target assays.
This is a beautiful example of lead optimization guided by structure with a keen focus on molecular and pharmaceutical properties. The initial fragments are difficult to discern in the final molecule, which is not a bad thing: the whole point of fragment-based discovery is giving multiple options for creative medicinal chemistry. In contrast to last week’s post, crystallography was essential for the program; it also benefited from the applied serendipity of parallel synthesis.
Often these sorts of publications are the valediction of a halted program, but not here: PF-06835919 is moving forward in three clinical trials, including a phase 2 trial for NAFLD. Interestingly, the compound was first dosed in humans in 2016 – a year before the initial paper. This gap between clinical efforts and publications is a reminder that our list of fragment-derived clinical compounds will always be incomplete. I look forward to watching PF-06835919 advance.

01 February 2021

Advancing fragments without structures: NPBWR1

Last week’s post highlighted how biophysical methods, and in particular structural insights, can be critical for advancing fragments to leads. But while everyone likes a structure, one quarter of respondents to our 2017 poll said they were comfortable optimizing fragments on the basis of SAR alone. (See also a recent review.) A new example of structure-free optimization has been published in Bioorg. Med. Chem. Lett. by Remond Moningka and colleagues at Merck.
The researchers were interested in the GPCR neuropeptide B/W receptor subtype 1 (NPBWR1, also known as GPR7), a potential target for obesity. Although impressive advances have been made towards obtaining structural information on membrane-bound proteins such as GPCRs, especially using cryo-EM, routine structure-based design is generally not an option.
The researchers started with a 30,000 member library of fragments between 200-350 Da. Both the size of the library and the size of the fragments are on the large side compared to what is typically used. A cell-based screen (cAMP assay) at 100 µM yielded 500 hits that inhibited at least 30%. Counter-screening against an unrelated GPCR whittled down the number to 20, of which just 3 provided dose-responses. The low confirmed hit rate illustrates both the utility of a larger library as well as the number of false positives likely to arise in a cell assay.
SAR by catalog on compound 1 led to compound 2, and further SAR led to compound 3c, with low micromolar activity and good ligand efficiency. Replacing the nitro group with a more pharmaceutically acceptable trifluoromethyl group produced compound 10. It is worth noting that compound 10 is still fragment-sized yet is >300-fold more active than the initial hit. This is a useful reminder that one can often make significant improvements even before fragment growing. Finally, extensive SAR studies around the phenyl ring ultimately led to compound 21a, with low nanomolar activity.

The pharmacology around GPCRs can be complicated, and compound 21a turned out not to be a simple competitive (orthosteric) antagonist of NPBWR1. Rather, it seems to act as a negative allosteric modulator: it reduces the affinity of the natural ligand.

This is a concise success story of advancing a fragment in the absence of structural information. Does this mean we should not strive for structures? Heck no! Not only would structures likely facilitate faster and further improvements, they might explain the mechanism of action of the compounds. I, for one, would love to know where and how they bind.
But this paper is another reminder that you do not always need crystallography - or even a model -  to take a fragment to a lead.

25 January 2021

Fragments vs TNFα advanced with biophysics, linking, and growing

The cytokine tumor necrosis factor α (TNFα) is a key mediator of inflammation and has long been a target for rheumatoid arthritis, Crohn’s disease, psoriasis, and a host of other inflammatory diseases. Several biologic drugs, such as adalimumab, are approved but these monoclonal antibodies and fusion proteins can be immunogenic or induce neutralizing antibodies. Small molecules could avoid these pitfalls and also reach organs, such as the brain, less accessible to biologic agents. Impressive efforts towards this goal have just been reported in J. Med. Chem. by Justin Dietrich, Chaohong Sun, and colleagues at AbbVie. (Andrew Petros presented this work at the CHI DoT meeting last September.)
The researchers began by screening 18,000 fragments using two-dimensional (13C-HSQC) NMR against TNFα in which the methyl groups of isoleucine, valine, leucine, and methionine were isotopically labeled. Only 11 fragments caused significant perturbations, an 0.06% hit rate reflecting the difficulty of finding hits against this target. All the fragments were characterized by SPR, and compound 1 turned out to have reasonable affinity and ligand efficiency. Synthesis of a few dozen analogs led to compound 2, with improved activity.

TNFα forms a homotrimer, and a crystal structure of compound 2 bound to TNFα revealed that two copies of the fragment bind within a large hydrophobic cavity at the interface of the three protein monomers. Not present in the apo-form of the protein, this central pocket is formed by the movement of tyrosine side chains, causing desymmetrization of the protein trimer. The researchers linked the two nearby fragments to produce compound 3 with improved affinity but decreased ligand efficiency. Further optimization led to compound 4, which was active in cells. But perhaps not surprisingly given its size and lipophilicity, this molecule had high clearance and poor oral bioavailability in mice. 
A second fragment, compound 6, had lower affinity than fragment 1, and parallel chemistry efforts generated only flat SAR. A crystal structure revealed that compound 6 bound in a similar manner as compound 1, with two copies in the central cavity. Surprisingly, a crystal structure of compound 8, which differs from compound 6 only by a single methyl group and actually has slightly lower affinity, revealed a singly copy bound in the central void. Scaffold hopping led to compound 9, which was ultimately optimized through structure-based design and careful attention to drug-like properties to compound 12. This molecule is orally bioavailable and showed activity in a mouse arthritis model.
This is a lovely paper that illustrates several important lessons. First, as the researchers note, “we have learned from multiple programs, including this one, aimed at developing small-molecule inhibitors of protein-protein interactions, that biophysical methods, when used to drive a fragment-based approach, offer the greatest chance of success.” NMR was essential for finding the initial fragments, SPR provided necessary thermodynamic and kinetic information, and crystallography led to the breakthrough discovery of the binding mode of compound 8.
Second, although the initial series generated by fragment linking ultimately did not advance, it proved critical for developing chemical tools, validating assays, and providing structural insights.
And finally, this paper is a paean to persistence for difficult targets. As the researchers note, scientists have been seeking small molecule inhibitors of TNFα for decades, and although compounds were reported as early as 2005, most of these have had poor physicochemical properties. Seemingly undruggable targets can sometimes be unlocked. But it usually takes time.

18 January 2021

Does configurational entropy explain why fragment linking is so hard?

Linking two weak fragments to get a potent binder is something many of us hope for. Unfortunately, as a poll taken a few years back indicates, it often doesn’t work. But why? This is the question tackled by Lingle Wang and collaborators at Schrödinger and D. E. Shaw in a recent J. Chem. Theory Comput. paper.
When a ligand binds to a protein it pays a thermodynamic cost in terms of lost translational and orientational entropy. By linking two fragments, this cost is paid only once instead of twice. In theory this should lead to an improvement of 3.5-4.8 kcal/mol in binding energy, resulting in a 400-3000-fold improvement in affinity over what would be expected from simple additivity. As we noted here, this is possible, though rare. Linker strain often takes the blame as a primary villain. But is there more to the story?
The researchers computationally examined published examples of fragment linking (most of which we’ve covered on Practical Fragments) using free energy perturbation (FEP) to try to understand why the linked molecules bound more or less tightly than expected. Impressively, they were able to computationally reproduce experimentally derived numbers, and by building a thermodynamic cycle they could extract the various components of the “connection Gibbs free energy.” These included changes in binding mode or tautomerization, linker strain or linker interactions with the protein, and the previously mentioned entropic benefits of fragment linking.
The analysis also identified two additional components. If two fragments favorably interact with each other, covalently linking them may not give as much of a boost. This concept had been considered decades ago, though the current work provides a more general understanding.
The more important factor appears to be what the researchers refer to as “configurational entropy.” The notion is that even when a fragment is bound to a protein, both the ligand and protein retain considerable flexibility, which is entropically favorable. Linking two fragments reduces the configurational entropy of each component fragment, and the linked molecule binds less tightly than would be expected. The researchers argue that this previously unrecognized “unfavorable change in the relative configurational entropy of two fragments in the protein pocket upon linkage is the primary reason most fragment linking strategies fail.” They advise that maintaining a bit of flexibility in the linker can help, as has been previously suggested.
This is an interesting analysis, and explicitly considering configurational entropy is likely to improve our understanding of molecular interactions. But is it really the main barrier to successful fragment linking? The researchers explore only nine different protein-ligand systems, though they did consider multiple linked molecules for three of these (pantothenate synthetase, RPA, and LDHA). Still, these represent just a fraction of the 45 examples collected in a recent review, and they only considered one somewhat contrived case (avidin) in which especially strong superadditivity was observed. It will be interesting to see whether the analysis holds true for more examples of fragment linking.

11 January 2021

Hundreds of fragments hits for the SARS-CoV-2 Nsp3 Macrodomain

COVID-19 will be with us for some time. Despite the unprecedented speed of vaccine development, it is worth remembering that humanity has only truly eradicated two widespread viral diseases, smallpox and rinderpest. Thus, the long march of small molecule drug discovery against SARS-CoV-2 is justified. In a paper recently posted on bioRxiv, Ivan Ahel and more than 50 multinational collaborators take the first steps.
Last year we highlighted two independent crystallographic screens against the main protease of SARS-CoV-2. Another potential viral target is the macrodomain (Mac1) portion of non-structural protein 3 (Nsp3), an enzyme which clips ADP-ribose from modified proteins, thus helping the virus evade the immune response.
The researchers soaked crystals of Mac1 against a total of 2683 fragments curated from several collections. This yielded 214 hits, and most of the structures were solved at high resolution (better than 1.35 Å). About 80% of the fragments bound in the active site, with many binding in the adenosine sub-pocket. Two different crystal forms were used for soaking, and one set of 320 fragments was soaked against both. Interestingly, this yielded a hit rate of 21% for one crystal form and just 1.3% for the other. Even more surprising, of the five hits found in both crystal forms, only two bound in the same manner in both. This is a clear demonstration that it is worth investing up-front effort to develop a suitable crystal form of a protein before rushing into soaking experiments.
Independently, the researchers computationally screened more than 20 million fragments (mostly from ZINC15) against the protein using DOCK3.7, a process which took just under 5 hours on a 500-core computer cluster. Of 60 top hits chosen for crystallographic soaking, 20 yielded structures, all at high resolution (0.94-1.01 Å). The ultra-high resolution structures revealed that four fragments had misassigned structures (wrong isomers), which long-time readers may not find surprising. Importantly, most of the 20 experimentally determined structures confirmed the docking predictions.
A strength and weakness of crystallographic screening is that it can find extraordinarily weak binders, which may be difficult to optimize. To see whether they could independently verify binding, the researchers tested 54 of the docking hits in a differential scanning fluorimetry (DSF) assay. Ten increased thermal stability, and all of these had yielded crystal structures. Only four of 19 fragments tested yielded reliable data in isothermal titration calorimetry (ITC) assays, but encouragingly these four also gave among the most significant thermal shifts in the DSF assay. Finally, 57 of the docking hits and 18 of the crystallographic hits were tested in a homogenous time-resolved fluorescence (HTRF) based peptide-displacement assay, yielding 8 and 3 hits respectively, the best with an IC50 of 180 µM.
This paper is a tour de force, and may represent the largest collection of high-resolution crystallographic fragment hits against any target. Laudably, all 234 of the crystal structures have been released in the public domain, and the researchers have already suggested ideas for merging and linking. As they point out, many of the fragments bind in the adenine pocket, so selectivity will be an issue not just against human macrodomains but also against kinases and other ATP-dependent enzymes. Still, as the dozens of approved kinases inhibitors demonstrate, achieving selectivity is possible. 
From a technology perspective, this publication affirms the rising power of both crystallographic and computational screening. Indeed, the hundreds of crystal structures will themselves be useful input for training new computational methods. And from a drug discovery perspective, each of these fragments represents a potential starting point for SARS-CoV-2 leads.
Let’s get busy!

04 January 2021

Fragment events in 2021

Gotten vaccinated yet? Don't worry - the first few conferences of the year will be virtual, but hopefully we'll be meeting in person later this year.

March 9-12:  While not exclusively fragment-focused, the Second NovAliX Virtual Conference on Biophysics in Drug Discovery will have several relevant talks. You can read my impressions of the 2018 event here, the 2017 Strasbourg event here, and Teddy's impressions of the 2013 event herehere, and here.
May 18-19: CHI’s Sixteenth Annual Fragment-Based Drug Discovery, the longest-running fragment event, will again be held virtually. This is part of the larger Drug Discovery Chemistry meeting, running May 18-20. You can read impressions of the 2020 virtual meeting here, the 2019 meeting here, the 2018 meeting here, the 2017 meeting here, the 2016 meeting here; the 2015 meeting herehere, and here; the 2014 meeting here and here; the 2013 meeting here and here; the 2012 meeting here; the 2011 meeting here; and 2010 here.
September 27-30: CHI’s Nineteenth Annual Discovery on Target returns to the real world - or at least Boston. As the name implies this event is more target-focused than chemistry-focused, but there are always plenty of FBDD-related talks. You can read my impressions of the 2020 virtual event here, the 2019 event here, and the 2018 event here.

December 16-21: What better place to say goodbye to COVID than Hawaii? Postponed from last year, the second Pacifichem Symposium devoted to fragments will be held in Honolulu. Pacifichem conferences are normally held every 5 years and are designed to bring together scientists from Pacific Rim countries including Australia, Canada, China, Japan, Korea, New Zealand, and the US. Here are my impressions of the 2015 event.
Know of anything else? Please leave a comment or drop me a note!