14 August 2017

Fragments distinguish allosteric from active site binders

As discussed last year, secondary binding sites on proteins appear to be quite common. Some of these sites have no functional relevance, but others are allosteric sites, which can modulate the activity of proteins. Allosteric ligands can be useful for several reasons. First, unlike molecules that bind at the active (that is, catalytic) site of an enzyme, which usually inhibit activity, allosteric site binders can increase activity. Second, allosteric sites are usually less conserved than active sites, allowing greater selectivity. Finally, combining an allosteric inhibitor with an active site inhibitor can lead to synergy as well as lower the incidence of resistance mutations for cancer and anti-infectives. In a recent ACS Med. Chem. Lett. paper, Lukasz Skora and Wolfgang Jahnke at Novartis describe a simple NMR approach to differentiate these two classes of ligands.

The researchers used 19F NMR to screen 540 fragments containing a CF3 group, each at 25 µM, in pools of 30 against the kinase ABL1 (at 4 µM); the BCR-ABL1 mutant form of this protein is a key driver for chronic myelogenous leukemia. Several approved drugs target the active site of ABL1, and Novartis researchers have recently launched clinical studies of a compound called ABL001, which binds to an allosteric pocket.

Fragments that bind to ABL1 showed a decreased 19F NMR signal due to line broadening. Adding ABL001 displaced fragments that bind to the allosteric site, thereby increasing their NMR signals, while adding the active-site binding drug imatinib displaced fragments that bind to the catalytic site. Follow-up experiments with individual fragments identified a selective catalytic-site binder (CAT-1) and a selective allosteric site binder (ALLO-1). Both fragments are commercially available and quite weak (Kd = 43 µM for ALLO-1 and IC50 = 380 µM for CAT-1), which in this case is a feature because they can easily be displaced.

Mixing these two fluorine-containing probes with ABL1, adding test compounds, and performing 19F NMR thus provides a simple means to determine whether a ligand binds to the allosteric site, the active site, or both sites. The researchers confirmed that the approved catalytic-site binding drugs nilotinib, dasatinib, and ponatinib displace CAT-1 but not ALLO-1, while allosteric-site binders such as ABL001 displaced ALLO-1 but not CAT-1.

Interestingly, a crystal structure of imatinib with the highly related protein ABL2 shows the compound binding to both the catalytic and allosteric sites, yet although imatinib clearly displaced CAT-1 it could not displace ALLO-1. This is a useful reminder that crystal structures say nothing about affinity.

The drug crizotinib, which binds to the active site of multiple kinases, has been reported by other researchers to bind to the allosteric pocket of BCR-ABL1, but this was not borne out in the competition assays. Similarly, the drug fingolimod has also been reported as an allosteric inhibitor of ABL1. This molecule did indeed displace ALLO-1, but only at concentrations so high as to be biologically irrelevant.

This is a nice paper, and a good reminder that fragments can make useful biophysical probes in and of themselves, even without the need for optimization.

07 August 2017

Assessing ligandability by thermal scanning

Ligandability refers to the ability to find small-molecule leads against a target. A protein might be ligandable but not druggable if, for example, potent inhibitors of the target do not affect a disease state. But knowing in advance whether a target is ligandable can be useful, both to decide whether to embark on a campaign and to plan the resources it will likely require. Fragment screens by NMR have been shown to be good predictors of ligandability, but not everyone has access to this technology. Computational methods (such as FTMap) are also useful, but require a structure of the target. In a recent paper in J. Med. Chem., Stefan Geschwindner and colleagues at AstraZeneca describe high-throughput thermal scanning (HTTS) for assessing ligandability.

Thermal scanning (alternately called, as the researchers note, thermal shift, differential scanning fluorimetry (DSF), or thermofluor) relies on the preferential binding of a fluorescent dye to protein that is heat-denatured. Since ligands generally stabilize a protein against denaturation, an increase in melting temperature (Tm) is taken as an indication of binding. The assays can be plate-based and thus very fast.

The researchers chose 16 diverse targets (mostly enzymes) and screened their 763-ligandability fragment set (described here) at 1 mM by HTTS. Hits were defined as compounds that increased  thermal stability at least 3-fold above the standard deviation of controls. Targets were then categorized as follows:

Low ligandability: hit rate < 1.5%
Medium ligandability: hit rate between 1.5 and 4.5%
High ligandability: hit rate > 4.5%

Nine targets ranked low, and all of these failed high throughput screening (HTS), while 5 out of the 7 targets ranked medium or high by HTTS yielded useful HTS hits. Of course, failure in an HTS does not preclude target advancement by other means – including FBLD. Ultimately all but three targets (including all of those ranked medium or high and 6 of 9 ranked low) went on to enter hit-to-lead optimization programs.

Encouragingly, HTTS and NMR agreed perfectly for low and high ligandability targets, but NMR assigned three targets as medium where HTTS assigned them as low. The researchers thus set out to increase the sensitivity of HTTS.

It turns out that entropically-driven binders tend to cause greater thermal shifts than enthalpically driven binders. The observation that most fragments bind largely enthalpically, and with low affinity too, makes them particularly challenging to detect. To try to shift the balance, the researchers repeated the HTTS assay for three of the low-scoring targets in D2O instead of H2O, which enhances entropic interactions at the expense of enthalpic interactions. Indeed, all three targets showed enhanced hit rates, and two moved from low to medium ligandability.

Another way to improve sensitivity of a thermal shift assay is to add urea, which destabilizes proteins by lowering the unfolding enthalpy. Adding non-denaturing amounts of urea (0.8 to 2.4 M concentration) to the three low-scoring targets above did indeed increase the hit rate for two of them.

One interesting tidbit is the observation that particularly stable targets, with unfolding temperatures >70 °C, tend to produce lower hit rates in HTTS than less stable targets. This could account for the very different experiences people have had with the technique.

This is a nice paper, and the approach may be worth implementing, as the researchers note has already happened at AstraZeneca. Although HTTS is unlikely to ever be as robust as SPR, NMR, or crystallography, it is hard to beat the low cost and high speed.

31 July 2017

Fragments in the clinic: PF-06650833

Of the more than 30 fragment-derived drugs that have entered clinical development, more than a third target kinases. While most of these are being developed against various types of cancer, a new paper in J. Med. Chem. by Katherine Lee, Stephen Wright, and their Pfizer colleagues describes the discovery of a compound that inhibits interleukin-1 receptor associated kinase 4 (IRAK4), a target for chronic autoimmune diseases. (Katherine also spoke about this project at FBLD 2016.) This details the earliest screens through development of the active clinical candidate.

The researchers started by screening their 2592-member Global Fragment Initiative library at 236 µM using STD NMR, resulting in 169 hits. A biochemical screen of the same library at 909 µM produced 160 hits, with 95 in common. Further triage using another assay along with modeling prioritized 15 fragments, of which 10 produced structures in co-crystallization trials. Fragment 51 was particularly interesting due to its impressive ligand efficiency and unusual binding mode to the hinge region of the kinase.

The crystal structure suggested that fragment growing could be productive, and indeed simply expanding the phenyl ring to a naphthyl improved the affinity to low micromolar for compound 10. Adding a nitrogen into the ring to lower lipophilicity while also adding a substituent to pick up additional interactions improved the affinity another order of magnitude (compound 14).


Guided by a co-crystal structure of compound 14 bound to IRAK4, the researchers used parallel chemistry to further improve the molecule, resulting in compound 20, which crystallography confirmed makes multiple interactions with the protein. Compound 20 also had promising selectivity and pharmacokinetic properties, but despite low nanomolar activity in a biochemical assay it had only high nanomolar potency in human peripheral blood mononuclear cells (PBMC).

At this point the medicinal chemistry began in earnest, again guided by structure and with a keen eye on maintaining good physicochemical properties. To a non-chemist the changes between compound 20 and PF-06650833 may appear subtle, but chemists will appreciate that you don’t introduce two new stereocenters without darn good reasons, which are discussed in depth in the paper. The results paid off, with the final molecule showing low nanomolar potency in the PBMC assay, excellent selectivity against a broad panel of kinases and other targets, and attractive ADME properties. It was also orally active in an acute rat inflammation model.

Sometimes publications only appear after a compound has dropped out of development, but that is not the case here. Indeed, after completing four phase 1 studies, PF-06650833 is currently being tested in a phase 2 trial for rheumatoid arthritis. Watch this space!

24 July 2017

Fragments vs Trypanosoma parasites

Last month we highlighted how fragments could be used to discover inhibitors of protein-protein interactions (PPIs). Today we continue the theme of fragments vs PPIs, in this case the interaction between PEX14 and PEX5, proteins which are important for glucose metabolism in disease-causing protists such as Trypanosoma.

The research, published recently in Science, was done by a large multinational team led by Grzegorz Popowicz, Michael Sattler (both at Helmholtz Zentrum München), and Ralf Erdmann (Ruhr University Bochum). They started by solving the NMR structure of the N-terminal domain of PEX14 from T. brucei, the organism that causes sleeping sickness. Previous work had shown that PEX5 binds to this domain, with two aromatic side chains of PEX5 binding in adjacent hydrophobic pockets. With this information in hand, the team performed a virtual screen of several million (non-fragment-sized) molecules. Eight of the best-scoring hits were tested, and four showed binding in an NMR assay, with compound 1 having the highest affinity.


Next, the researchers screened a library of 1500 fragments (each at 1 mM in pools of 5) using 1H, 15N HMQC NMR. This led to 12 hits with affinities better than 2 mM. Strikingly, all of these fragments contained fused bicylic aromatic ring systems, three of which were substituted naphthyls. Appending these onto compound 1 led to compound 4, with low micromolar affinity. Introducing an amine to interact with a glutamic acid residue in PEX14 led to compound 5, with high nanomolar affinity. This compound also showed activity against several species of pathogenic Trypanosoma. Further tweaking led to a molecule with activity in a mouse model of infection.

This example of fragment-assisted drug discovery (FADD) is reminiscent of other cases (described here, here, and here) in which fragments were used to replace elements of a previously identified molecule. While it is possible that traditional medicinal chemistry could have achieved the same result, fragments probably helped winnow down the number of molecules to be synthesized. It is also nice to see this technology applied to understudied diseases. 

17 July 2017

Native mass spectrometry revisited

Native electrospray ionization mass spectrometry (ESI-MS) is one of the less-commonly used fragment finding methods. The technique relies on gently ionizing a protein-fragment complex without causing denaturation; bound fragments reveal themselves as shifts in mass. The technique is truly label-free, and can use very small amounts of protein and fragments. In practice the technique can work really well, reasonably well, or quite poorly. Two new papers shed light on factors that influence success.

The first paper, by Kevin Pagel (Freie Universität Berlin), Benno Kuropka (Bayer), and collaborators, examines four different cancer-related proteins. Let me say up-front that that the paper is remiss in not disclosing the chemical structures of any of the fragments, so in a very real sense this work is not reproducible. It is a shame the editors of ChemMedChem were not more demanding. That said, there is some useful information here.

Most of the focus is on the protein MTH1, screened at 10 µM concentration with 100 µM of each fragment. This was not a naïve screen; the fragments were previously identified from a thermal shift assay (TSA): 24 stabilized the protein, 4 destabilized it, and 5 had no effect. Remarkably, all of the fragments showed complexes in ESI-MS ranging between 6 – 66%, even those that had no effect in the TSA! Choosing an (admittedly arbitrary) 20% cutoff weeded out most of the false positives: 16 of the 24 stabilizers passed, while none of the destabilizers or neutral molecules did.

The best hit by ESI-MS also gave the strongest thermal shift, and a titration curve revealed an impressive dissociation constant of 1.7 µM. However, even at high concentrations of fragment the amount of bound complex did not exceed 70%, meaning that interpretation of single-dose experiments (for example, from a primary screen) could be problematic.

The results were similar for the protein KDM5B: 8 of 9 stabilizing fragments were hits by ESI-MS, as were two of 7 destabilizing fragments. Note that fragments that destabilize proteins can still be tight binders, as illustrated here.

For two additional proteins, however, ESI-MS was disappointing. For BRPF1, ESI-MS didn’t find any of the 11 hits from TSA, while for UHRF1 it found only a single hit – though this hit was not one of the 10 stabilizers identified by TSA. One could argue that the TSA hits were false positives were it not for the fact that, in the case of BRPF1, 6 of them were confirmed by crystallography.

The second paper, in Angew. Chem., comes from Chris Abell and coworkers at the University of Cambridge, and focuses on the protein EthR, a potential target for tuberculosis that we’ve previously discussed.

EthR binds to DNA, so rather than look for direct binding of fragments to EthR the researchers instead looked for fragments that could disrupt the EthR-DNA complex. A small library of 73 fragments was tested (at 0.5 mM each, in 2% DMSO), yielding 8 hits. The same library was screened under the same conditions using differential scanning fluorimetry (DSF), yielding 7 hits, 4 of which had also been identified using ESI-MS. All 11 of these molecules were then tested under the same conditions in an SPR assay to see if they could disrupt the interaction between EthR and chip-bound DNA. The 7 best SPR hits were all fragments that had been identified by ESI-MS. Moreover, two fragments – one identified solely by ESI-MS and one identified by both ESI-MS and DSF – were characterized bound to EthR crystallographically, and these represent new chemotypes for this target.

So what are we to make of all this? In common with other techniques, ESI-MS works well for some targets and less well for others. The problem is that it is not clear what distinguishes the two classes of targets. If you have access to the equipment and expertise you might consider adding ESI-MS to your screening cascade. But if you can only afford to buy one instrument for fragment screening, you’d probably be better off investing in NMR or SPR.

10 July 2017

Reagents as covalent fragments

Covalent drugs are a thing these days: as long as you can get selectivity, there’s nothing like a covalent bond to juice up affinity for a target. Screening for covalent fragments is thus a reasonable approach, and multiple researchers have assembled libraries of fragments containing either irreversible or reversible covalent “warheads”. The latest example, by Marion Lanier, Mark Hixon, and collaborators at Takeda, appears in J. Med. Chem.

Boronic acids can form reversible covalent bonds with the side chains of serines or threonines in proteins, with a predilection for the highly reactive active-site residues found in hydrolytic enzymes. Indeed, three approved drugs – bortezomib, tavaborole, and ixazomib – contain boronic acids.

When medicinal chemists think of boronic acids, they probably think of them as reagents for the Suzuki coupling, a useful method for forming carbon-carbon bonds. Because the reaction is so popular, some 6000 boronic acids are commercially available, many of them fragment-sized. Thus, the researchers assembled a set of 650 into a boronic acid library (BAL).

To determine whether this BAL would be useful, the researchers screened it against autotaxin, a phospholipase with anti-cancer and anti-inflammatory potential. Fragments were tested at 100 µM in a functional assay, with hits retested in 11-point dose-response curves. This yielded a whopping 51 molecules with IC50 values better than 10 µM, some as good as 200 nM.

The researchers also screened autotaxin against a set of 1750 non-boronic acid containing fragments, this time at 500 µM. Not surprisingly, hits tended to be significantly weaker despite the similar sizes of the fragments. The BAL fragments had average ligand efficiencies of 0.61 kcal mol-1 per heavy atom, while the conventional fragments averaged a lower but still respectable 0.41. Some of the BAL fragments were also crystallized bound to autotaxin, revealing that they do in fact form bonds with the catalytic threonine. 

This is a nice paper, though I do wish that the researchers had tried to calculate the inherent reactivities of the boronic acids to determine how these differences affected their affinities for the protein, as has been done for other warheads such as aryl acrylamides. Also, it would be interesting to see how a docking program such as DOCKovalent performs against this target with the same set of fragments. Hopefully we’ll see these questions addressed in the future. In the meantime, expect to see commercial vendors start offering libraries of boronic acid fragments.

03 July 2017

Fragment events in 2017 and 2018

The year is halfway behind us, but there are still a couple upcoming fragment-based events, and next year is already taking shape.

2017

July 23-28: Australia is coming into its own as a destination for fragment experts, many of whom will be participating in the Royal Australian Chemical Institute's Centenary Congress in Melbourne. The entire event should be huge - think ACS with wombats - so if you've been looking for yet another reason to travel Down Under, this is it.

September 25-29: CHI's Discovery on Target in Boston includes an Inaugural Lead Generation Strategies track (September 26-27), and it looks like fragments will play a major role - as well they should!

2018

January 28 - February 1: The First Alpine Winter Conference on Medicinal and Synthetic Chemistry will take place in St. Anton am Alberg, Austria. This looks like a fun event and includes a section on FBDD.

April 2-6: CHI’s Thirteenth Annual Fragment-Based Drug Discovery, the longest-running fragment event, will be held in San Diego. You can read impressions of this year's meeting here, last year's 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.

June 6-9: Although not exclusively fragment-focused, the fifth NovAliX Conference on Biophysics in Drug Discovery will have lots of relevant talks, and will be held for the first time in Boston. You can read my impressions of the recent Strasbourg event here and Teddy's impressions of the 2013 event herehere, and here.

October 7-10: Finally, FBLD 2018 returns to San Diego, where it was born back in 2008. This will mark the seventh in an illustrious series of conferences organized by scientists for scientists. You can read impressions of FBLD 2016, FBLD 2014,  FBLD 2012FBLD 2010, and FBLD 2009.

Know of anything else? Add it to the comments or let us know!

26 June 2017

Fragments vs BCL6, two ways

Disrupting protein-protein interactions (PPIs) tends to be challenging: interfaces are often large and flat, with few deep pockets in which small molecules can bind. Also, much like unhappy families, PPIs are usually dissimilar, meaning that HTS collections yield fewer, less attractive hits. Both of these challenges are well-addressed by fragments, and indeed last year saw the approval of venetoclax, which targets a PPI. Two new papers in J. Med. Chem. report inhibitors of another PPI.

B-Cell lymphoma 6 (BCL6) binds to other proteins to regulate gene expression. As its name suggests, it was identified in diffuse large B-cell lymphoma (DLBCL), and is thus an interesting anticancer target. Also, structures of the protein in complex with a peptide suggested that it may not be impossible to find small molecule inhibitors. Yusuke Kamada and colleagues at Takeda set out to do just this.

An SPR-based screen of 1494 fragments, each at 1 mM, identified 64 hits which confirmed in dose-response titrations. However, only seven compounds confirmed in an STD-NMR experiment. Of these, compound 1 was characterized crystallographically bound in the peptide groove.

A few tweaks to the fragment led to compound 4, with improved potency. Meanwhile, an HTS screen had identified the very weak compound 5, and crystallography showed that it bound at the same site as compound 4. Merging the two molecules led to compound 7, with mid-nanomolar biochemical potency and low micromolar activity in a cell-based assay. This is a classic case of fragment-assisted drug discovery (FADD), and a good illustration of how FBLD and HTS can be complementary.

The second paper, by William McCoull and a large team of collaborators from AstraZeneca and Pharmaron, goes somewhat further. The researchers conducted an SPR-based screen of 3500 fragments along with a virtual screen. Both found hits, and led to molecules with the same core as compound A2. Crystallography revealed that these also bind in the peptide groove, and combining elements of both molecules while tweaking the properties led to compound A5, with improved affinity.

The next step was growing compound A5 to try to make a hydrogen bond interaction with the protein. That led to compound A8, with submicromolar affinity. A crystal structure of a related compound bound to BCL6 revealed that two chemically distant portions of the molecule were in close proximity, suggesting a macrocyclization strategy similar to what we described last week. This proved highly successful, improving the affinity by more than two orders of magnitude for compound A11. NMR studies of the linear and cyclized compounds revealed that the conformation of the latter was indeed closer to that seen in the crystal structure.

The medicinal chemistry continues extensively from here. In particular, compound A11 showed some activity against the kinase CK2, but this could be engineered out. Multiple additional changes were explored, with many compounds showing low nanomolar activity in a biochemical assay and some showing high nanomolar activity in a cell-based assay. Unfortunately, none were very effective at inhibiting the proliferation of lymphoma cell lines. The authors state, “we conclude that the BCL6 hypothesis as a means of treatment for DLBCL is still unproven and we have elected not to progress this series of BCL6 inhibitors further into development.”

This makes sense, though I wouldn’t abandon all hope. For two other PPIs, BCL2 and MCL1, robust cell activity required picomolar affinity in a biochemical assay. Whether this level of potency is achievable for BCL6 remains an open question.

19 June 2017

Fragments vs BRD4, two ways

Bromodomains, epigenetic targets that recognize acetylated lysine residues, have received considerable attention from the fragment community. (I devoted all of last July to the topic, and covered it more recently here.) Of the dozens of bromodomain-containing proteins, the four BET-family members have been highly studied, and the second bromodomain of BRD4 (BRD4-BRDII) in particular has been implicated in cancer and inflammation. In two new papers, researchers from AbbVie describe inhibitors of this target.

In the first Bioorg. Med. Chem. Lett. paper, George Sheppard and colleagues briefly describe a protein-detected (13C-HSQC) NMR screen of BRD4 in which the methyl groups of isoleucine, leucine, valine, and methionine were 13C-labeled. About 18,000 fragments were screened in pools of 30, and hits were then tested individually in NMR and time-resolved fluorescence resonance energy transfer (TR-FRET) assays. Despite extensive work on this target by multiple groups, these screens were able to identify several new fragments, such as the related compounds 1 and 2.


Crystallography of each fragment bound to BRD4 revealed that they bind in the acetyl lysine recognition site and make contacts with the conserved asparagine residue as well as a nearby water molecule. Merging the fragments led to compound 5, with a slight increase in affinity.

Comparison with other BRD4 inhibitors suggested a growth strategy, leading to compound 15, with nanomolar activity in the TR-FRET assay and two cell-based assays. The compound was orally bioavailable but had relatively high clearance, so further medicinal chemistry focused on changing the original core. This ultimately led to compound 38, with improved oral bioavailability, lower clearance, good selectivity against non-BET bromodomains, and activity in a mouse xenograft assay.

The second paper, by Le Wang, John Pratt, and colleagues in J. Med. Chem., starts with a different fragment from the original screen, compound A1. This molecule was even weaker than the fragments described above, but crystallography confirmed that it binds in the same acetyl lysine binding pocket.


Again, comparison with known inhibitors provided ideas for fragment growing, rapidly leading to compound A11. Further medicinal chemistry – which is extensively described in the paper – led to compound A30a, which bears considerable resemblance to the series reported in the previous paper.

Crystal structures of compounds bound to the protein suggested that it might be possible to make a macrocycle, which would in theory increase the affinity by locking the molecule in a low energy conformation. This proved to be synthetically challenging but ultimately worthwhile in the form of compound A74b. (Incidentally, this is the first case I can recall where a fragment led to a macrocycle. It won’t be the last.) Not only was this molecule more potent than the open form, it also showed excellent oral bioavailability and pharmacokinetics, good selectivity against non-BET bromodomains, and even better activity in a mouse xenograft model. 

One lesson from these papers is that fragments can generate new ideas even for heavily pursued targets. A second is that, as we saw in the recent poll, crystallographic information can be critical for advancing fragments to leads. The discovery of new moieties along with clear data on their binding modes can be a powerful combination for creative medicinal chemists.

12 June 2017

Fourth NovAliX Biophysics in Drug Discovery Conference

NovAliX held its fourth Biophysics in Drug Discovery meeting last week in the beautiful city of Strasbourg. This was my first time attending, and although Teddy’s recounting of the first meeting (here, here and here) had given me high expectations they were easily exceeded. With 134 participants from 13 countries, most of whom stayed for the full time, the event felt like a Gordon Research Conference, with lots of lively discussions over excellent food and drink. Rather than trying to summarize all 30+ talks and nearly as many posters, I’ll just provide a few impressions. Conference Chairman Jean-Paul Renaud posted a 2 minute video overview here.

Factors that drive success in FBLD were a theme of Jenny Sandmark (AstraZeneca). She discussed several projects in which fragments were able to generate useful chemical matter, including one (Complement Factor B) where HTS and DNA-encoded libraries both came up short. In the case of neutrophil elastase, SAR on HTS hits was making rapid progress, but chemistry to explore the S1 pocket was difficult. A directed NMR screen of 800 fragments did not yield anything better, but it did save considerable synthetic effort and was thus judged a success.

Tweaking experimental conditions was often essential to get informative results: soaking crystals of neutrophil elastase with fragments dissolved in DMSO produced no hits, while dissolving the fragments in water did. With crystals of Factor XIa, glycerol – not fragments – bound in the active site. Fortunately the researchers recognized this and were able to use a different cryoprotectant.

Glyn Williams (Astex) has been doing FBLD since the earliest days, and discussed some of the lessons learned. Although the current Astex fragment library consists of about 1800 compounds, some 8000 fragments have been evaluated over the years. Small fragments in particular can be quite volatile, so Astex stores its library at -80 °C under nitrogen. Researchers also do rigorous quality control (QC) and maintain “fragment CVs”, which summarize analytical and screening data for each molecule. Astex is heavily invested in synthesizing novel fragments to explore specific regions of chemical space, and this means being constantly on guard for risks such as redox cyclers.

Understanding the enemy is always useful, so Martin Redhead (UCB) has constructed a library of bad actors – including PAINS, chelators, and metals – to stress-test his assays. But even robust assays show plenty of false positives. A screen of 20,000 molecules against a protein-protein interaction revealed 105 stabilizers, only 3 of which turned out to be real, but four times as many inhibitors, none of which were legitimate. Some of the stabilizers could subsequently be modified to inhibitors, and Martin suggested that screening for stabilizers initially could be a general way of improving signal to noise.

NMR is uniquely capable of providing extensive QC information about both proteins and ligands, and Alvar Gossert (ETH) discussed how a “validation cross” confirming both integrity and binding can improve confidence. Alvar also discussed using dynamic nuclear polarization to reduce the number of spectra required to assign a protein from five collected over two weeks to one obtained in under three days. This requires a specialized setup involving a second magnet, but it does appear powerful. Another unusual NMR configuration was described by Ad Bax (NIH), who is studying protein folding by rapidly (< 1 msec) increasing the pressure of a sample to 2500 bar – the equivalent energy of “discharging a small handgun into your NMR.”

Plenty of more accessible technologies were also described, including some interesting new commercial offerings. Matyas Vegh (Creoptix) discussed using waveguide interferometry to study protein-ligand interactions. This is similar to surface plasmon resonance (SPR) but with higher sensitivity and lower bulk effect. Their new four-channel instrument can be temperature controlled from 4-45 °C and is capable of accurately measuring kinetics even for weak binders, such as the 2.79 s-1 off-rate for the binding of methylsulfonamide (MW = 95 Da, Kd = 419 µM) to carbonic anhydrase II. And Sven Malik (Sierra Sensors) described a sensitive new SPR instrument with 8 channels, each with 4 spots, capable of running 384 samples in under 3 hours.

Several talks or posters highlighted an instrument from Biodesy which is capable of studying sub-Ångström conformational changes in dye-labeled, surface-immobilized proteins using second harmonic generation. Their plate-based instrument can measure 20,000 samples per week. Elizabeth Vo (UCSF) is using this to study the protein K-Ras, and has identified a number of active fragments which are being further characterized using orthogonal methods. It will be fun to see how these compare with previously reported K-Ras binders.

Finally, the keynote lecture was delivered by Jean-Pierre Changeux (Collège de France and Institut Pasteur), who described some of the highlights – many of them quite recent – from a career that spans nearly six decades. Jean-Pierre literally invented the model for allostery at a time when the three dimensional structures of only two proteins were known; today the Protein Data Bank contains more than 130,000 structures, and at least 90 marketed drugs are known to work through allosteric mechanisms. It is humbling to be reminded of how far we've come, yet how little we still know.

Strasbourg is a wonderful city, but with today’s travel budgets it can be difficult to access for some researchers, so next year the conference will head to Boston (June 12-15), returning to Strasbourg in 2019 before moving on to Kyoto in 2020. Hope to see you there!

05 June 2017

Poll results: what structural information is needed to optimize fragments

Our latest poll asked “how much structural information do you need to begin optimizing a fragment?” Over the past few weeks we received 143 responses, and the results are shown here.
Just over a third of you said that you wouldn’t work on a project without a crystal structure, while nearly a quarter said you’d settle for an NMR-based model. In other words, more than half of you demanded fairly detailed structural data to embark on a fragment optimization campaign.

But consistent with continuing improvements in modeling, almost a fifth said that a computational model would be just fine.

And perhaps most surprisingly, fully one quarter of you said that SAR alone would be sufficient to begin optimizing a fragment. Presumably this is driven at least somewhat by internal successes, and I look forward to seeing these disclosed in meetings and publications.

All of these approaches are rapidly developing, and it will be fun to revisit this poll in a few years to see whether crystallography maintains its lead, or whether lower resolution methods gain dominance.

In the meantime, are there other topics you’d like to see polled?

29 May 2017

Fragment hot spots revisited: a public validation set and method

This is the last week for our poll on how much structural information you need to begin optimizing a fragment – please vote on the right-hand side of the page if you haven’t already done so. We’ve recently discussed crystallography and NMR, so this post is focused on computation.

Predicting hot spots – regions on proteins where fragments are likely to bind – is becoming something of a cottage industry (see for example here and here). These can provide some indication as to whether or not a protein is ligandable, and ideally even provide starting points for a lead discovery program. But how should one searching for promising hot spots and binders choose a method, or evaluate a new one? In a recent paper in J. Med. Chem., Marcel Verdonk and colleagues at Astex provide a method as well as a validation set, both of which are freely available.

The validation set consists of 52 high-quality crystal structures pulled from the Protein Data Bank (PDB). These were chosen to be maximally diverse in terms of fragments (41 of them) and proteins (45). The fragments were not taken in isolation; rather, fragments of larger molecules were considered if they bound in the same region of the protein when presented from at least three different ligands. For example, the researchers note that there are no structures in the PDB of resorcinol bound to HSP90A, even though this is a privileged fragment that usually binds in a conserved fashion at the ATP-binding site in the context of a larger molecule.

Fragments chosen for the validation set have at most one rotatable bond and are quite small, just 5 to 12 non-hydrogen atoms. However, as they are culled from larger molecules, some (such as adamantane) are more lipophilic than standard “rule of three” guidelines.

The 52 examples in the test set were divided into 40 hot spots and 12 “warm” spots, depending on the occupancy of the binding site in the protein across the PDB. For example, the canonical purine binding site of kinases is a hot spot, while the nearby chlorophenyl-binding site of the PKA-Akt chimeric kinase is classified as warm.

With this validation set in hand, the researchers tested an in-house developed fragment mapping method called PLImap (which relies on the previously published Protein-Ligand Informatics force field, PLIff) to see how well they could reproduce the bound conformations of the fragments. The results were quite favorable in comparison with other docking methods tested. That’s exciting, and since PLImap is free to download and use (here), it should be a useful tool for modelers everywhere.

But of course PLImap also made mistakes, and some of these were “wrong in an interesting way.” Water often plays a critical role in protein-ligand interactions, but water was not included in the docking. Several cases where PLImap did not choose the experimentally observed conformation of the fragment involved water molecules. For example, PLImap placed the bromodomain-privileged 3,5-dimethoxylisoxazole fragment in the right location but in a flipped orientation, because the highly conserved water was not present.

Perhaps more interestingly, in some cases warm spots were ignored in favor of hot spots. For example, in the case of the PKA-Akt chimeric kinase mentioned above, the chlorophenyl fragment bound not at the chlorophenyl-binding subsite, where it sits in the context of larger ligands, but rather in the “hotter” purine binding subsite. This phenomenon was observed experimentally several years ago by Isabelle Krimm and colleagues; large BCL-xL ligands that were deconstructed into component fragments bound mostly at a single site, rather than the two sites occupied by the larger molecules. It would be fascinating to test this same set of molecules using PLImap.

All of which is to say that, while computational methods continue to make impressive strides, we are still (happily!) some way from getting rid of the experimentalists.

22 May 2017

Extracting more information from crystallographic data

The current poll on how much structural information is needed for fragment optimization is still open - if you haven't done so already, please vote on the right hand side. Last week we discussed new developments in NMR. This week we turn to crystallography.

Fragment screening by crystallography is a little like finding needles in haystacks. Typically, dozens or hundreds of crystals are individually soaked with one or more fragments. Diffraction data gathered from each crystal are used to generate electron density maps, which are iteratively refined by tweaking the conformation of side chains and adding water molecules. In theory, any unexplained electron density that remains after refinement should correspond to bound fragments.

In practice, the process of manually inspecting so many data sets can be both tedious and subjective. Although a narrow focus on the active site reduces the amount of work, doing so risks missing the many fragments that bind at interesting secondary sites. Also, because fragments have low affinities, they may only bind to a fraction of protein molecules; this "partial occupancy" lowers the signal to noise ratio. And fragments sometimes bind in more than one conformation, thereby smearing out the electron density and further reducing the signal.

Of course, even though crystallographic fragment screening can give very high hit rates, most crystals will not have bound fragments. In a new paper in Nat. Comm., Frank von Delf at the Structural Genomics Consortium and collaborators at several institutions describe how these "empty" structures can be turned from lemons into lemonade.

The method, called Pan-Dataset Density Analysis (PanDDA), is essentially a form of background correction. Dozens of datasets from empty crystals are averaged and computationally subtracted from a dataset of interest. This averaging gives much cleaner maps, allowing fragments to be more rapidly and easily detected. It’s almost as if you could subtract all the hay from a haystack to reveal any needles.

The researchers present four case studies of crystallographic fragment screens, each with more than 100 datasets, and the results are stunning: in one case manual inspection revealed just 2 fragment hits, both at a single site, while PanDDA revealed 24 fragments at 5 different sites!

One limitation of PanDDA is that it does require dozens of empty datasets – ideally more than 30. In a new paper in Acta Crystallogr. D Struct. Biol., Dorothee Liebschner at the Lawrence Berkeley National Laboratory and collaborators at other institutions describe an alternative approach suitable for lower throughput applications.

One common tool in crystallography is the OMIT map. Atoms in question (such as from a ligand) are omitted from the model, and the calculated electron density is then compared with the observed electron density; if the density remains, this suggests that the atoms really belong. Of course, there is no truly empty space in a crystal – solvent fills any space not occupied by protein or ligands. Typically this is accounted for by treating “bulk solvent” (ie, water molecules not making specific interactions) as being present at a constant level of background electron density. The problem is that when calculating an OMIT map, this bulk solvent could obscure weak but real electron density.

To address this challenge, the researchers develop “polder OMIT maps,” named after land that is kept dry despite being below the surrounding water level. Essentially, the bulk solvent is not allowed into polder OMIT maps when they are generated, thus enhancing any actual density and allowing low-occupancy ligands to be observed. Several lovely figures in the paper illustrate that the process works well.

It is nice to see that, despite its long history, crystallography continues to make practical and creative advances.

15 May 2017

NMR structures without protein assignments

Our latest poll asks how much structural information you need to advance a fragment (please vote on the right hand side of the page). On this subject, a recent paper by Marielle Wälti, Roland Riek, and Julien Orts in Angew Chem. demonstrates a new NMR method.

Researchers typically begin an NMR structure campaign by examining the chemical shift perturbations (CSPs) of proton-nitrogen or proton-carbon crosspeaks from an isotopically labeled protein in the presence and absence of a ligand. If you know which crosspeaks correspond to which specific atoms in an amino acid residue, you can deduce the ligand binding site by looking for the residues with the largest CSPs. Next comes the measurement of nuclear Overhauser effects (NOEs) between atoms in the ligand and atoms in the protein; these are exquisitely dependent on distance, so if you have enough measurements you can use these to accurately dock your ligand into the binding site of your protein.

This is how SAR by NMR was done more than twenty years ago, and it still works well today, but it is neither fast nor easy. In particular, the initial step of assigning the hundreds of protons, nitrogens, and carbons in a typical protein can be daunting.

To streamline the process, the researchers developed NMR molecular replacement (NMR2), first published last year (here) and presented by Julien at FBLD 2016. Rather than requiring knowledge of which peaks correspond to which specific protein atoms, NMR2 relies on the increasing power of computers to run large numbers of complex calculations. Various docking poses will generate different NOEs, so exhaustively and iteratively examining these possibilities and comparing them with the experimental data should generate an optimal model. (NMR2 does require that the structure of the protein is known, so you know the residues surrounding a ligand-binding pocket, even if you don’t know their chemical shifts. Also, the protons of the ligand are assigned, and in fact the intramolecular NOEs of the ligand itself are an important input.)

In the new paper the researchers apply NMR2 to a complex between the onocology target MDMX and a previously disclosed high nanomolar binder and find good agreement (1.35 Å RMSD) with the crystal structure.

The researchers then turn to “ligand #845,” which binds with millimolar affinity to the oncology target HDM2. A total of 33 intramolecular NOEs (from ligand #845) and 21 intermolecular NOEs (between ligand #845 and HDM2) were fed into NMR2 and used to crank through 54,000 structure calculations in a few hours to produce a binding model. No crystal structure was available, but conventional NMR methods support the model.

This seems like a rapid and powerful approach, but readers of this blog are probably wondering how well it will apply to fragments. Clearly NMR2 is sufficiently sensitive to weak binders. However, with 24 non-hydrogen atoms and a molecular weight of 354 Da, ligand #845 is too large to be called a fragment. Smaller molecules will have fewer hydrogen atoms and thus fewer intramolecular and intermolecular NOEs, decreasing the information content of the model. It will be fun to plumb the lower ligand size limits for this technique – leave a comment if you’ve done so!

08 May 2017

Poll: structural information needed for fragment optimization

As mentioned last week, advancing fragments in the absence of structure is a major challenge. But how much of a barrier is it really?

I know some researchers who would not consider moving forward with a fragment in the absence of a crystal structure. As crystallography continues to advance, more targets will be available, but many will remain out of reach for the foreseeable future.

Of course, the first SAR by NMR paper used NMR rather than crystallography, and the early work that ultimately led to venetoclax relied only on NMR-derived structures. Similarly, crystallography was initially unsuccessful against MCL-1, but NMR-based models allowed effective fragment advancement.

When crystallography and NMR both fail, there is in silico modeling, which continues to improve. Last year we highlighted how modeling succeeded in merging fragments to a nanomolar binder.

But the real challenge is advancing fragments with no structural information whatsoever. There are a few published examples (such as this and this). And it’s worth remembering that optimization in the absence of structure was how drug discovery was done decades ago, before the rise of biophysics. Indeed, until recently most GPCR-based drug discovery was done without the benefit of structural information.

So, in the poll to the right please choose the minimum level of structural information you would need to embark on a fragment to lead program. Happy voting!

01 May 2017

Twelfth Annual Fragment-based Drug Discovery Meeting

CHI’s Drug Discovery Chemistry meeting took place over four days last week in San Diego. This was easily the largest one yet, with eight tracks, two one-day symposia, and nearly 700 attendees; the fragment track alone had around 140 registrants. On the plus side, there was always at least one talk of interest at any time. On the minus side, there were often two or more going simultaneously, necessitating tough choices. As in previous years I won’t attempt to be comprehensive but will instead cover some broad themes in the order they might be encountered in a drug discovery program.

You need good chemical matter to start a fragment screen, and there were several nice talks on library design. Jonathan Baell (Monash University) gave a plenary keynote on the always entertaining topic of PAINS. Although there are some 480 PAINS subtypes, 16 of these accounted for 58% of the hits in the original paper, suggesting that these are the ones to particularly avoid. But it is always important to be evidenced-based: some of the rarer PAINS filters may tag innocent compounds, while other bad actors won’t be picked up. As Jonathan wrote at the top of several slides, “don’t turn your brain off.”

Ashley Adams described the reconstruction of AbbVie's fragment libraries. AbbVie was early to the field, and Ashley described how they incorporated lessons learned over the past two decades. This included adding more compounds with mid-range Fsp3 values, which, perhaps surprisingly, seemed to give more potent compounds. A 1000-member library of very small (MW < 200) compounds was also constructed for more sensitive but lower throughput biophysical screens. One interesting design factor was to consider whether fragments had potential sites for selective C-H activation to facilitate fragment-to-lead chemistry.

Tim Schuhmann (Novartis) described an even more “three-dimensional” library based on natural products and fragments. Thus far the library is just 330 compounds and has produced a very low hit rate – just 12 hits across 9 targets – but even a single good hit can be enough to start a program.

Many talks focused on fragment-finding methods, old and new. We’ve written previously about the increasingly popular technique of microscale thermophoresis (MST), and Tom Mander (Domainex) described a success story on the lysine methyltransferase G9a. When pressed, however, he said it did not work as well on other targets, and several attendees said they had success in only a quarter to a third of targets. MST appears to be very sensitive to protein quality and post-translational modifications, but it can rapidly weed out aggregators. (On the subject of aggregators, Jon Blevitt (Janssen) described a molecule that formed aggregates even in the presence of 0.01% Triton X-100.)

Another controversial fragment-finding technique is the thermal shift assay, but Mary Harner gave a robust defense of the method and said that it is routinely used at BMS. She has seen a good correlation between thermal shift and biochemical assays, and indeed sometimes outliers were traced to problems with the biochemical assay. The method was even used in a mechanistic study to characterize a compound that could bind to a protein in the presence of substrate but not in the presence of a substrate analog found in a disease state. Compounds that stabilized a protein could often be crystallized, while destabilizers usually could not, and in one project several strongly destabilizing compounds turned out to be contaminated with zinc.

Crystallography continues to advance, due in part to improvements in automation described by Anthony Bradley (Diamond Light Source and the University of Oxford): their high-throughput crystallography platform has generated about 1000 fragment hits on more than 30 targets. Very high concentrations of fragments are useful; Diamond routinely uses 500 mM with up to 50% DMSO, though this obviously requires robust crystals.

Among newer methods, Chris Parker (Scripps) discussed fragment screening in cells, while Joshua Wand (U. Penn) described nanoscale encapsulated proteins, in which single protein molecules could be captured in reverse micelles, thereby increasing the sensitivity in NMR assays and allowing normally aggregation-prone proteins to be studied. And Jaime Arenas (Nanotech Biomachines) described a graphene-based electronic sensor to detect ligand interactions with unlabeled GPCRs in native cell membranes. Unlike SPR the technique is mass-independent, and although current throughput is low, it will be fun to watch this develop.

We recently discussed the impracticality of using enthalpy measurements in drug discovery, and this was driven home by Ying Wang (AbbVie). Isothermal titration calorimetry (ITC) measurements suggested low micromolar binding affinity for a mixture of four diastereomers that, when tested in a displacement (TR-FRET) assay, showed low nanomolar activity. Once the mixture was resolved into pure compounds the values agreed, highlighting how sensitive ITC is to sample purity.

If thermodynamics is proving to be less useful for lead optimization, kinetics appears to be more so. Pelin Ayaz (D.E. Shaw) described two Bayer CDK kinase inhibitors having either a bromine or trifluoromethyl substitution. They had similar biochemical affinities and the bromine-containing molecule had better pharmacokinetics, yet the trifluoromethyl-containing molecule performed better in xenograft studies. This was ultimately traced to a slower off-rate for the triflouromethyl-substituted compound.

The conference was not lacking for success stories, including MetAP2 and MKK3 (both described by Derek Cole, Takeda), LigA (Dominic Tisi, Astex), RNA-dependent RNA polymerase from influenza (Seth Cohen, UCSD), and KDM4C (Magdalena Korczynska, UCSF). Several new disclosures will be covered at Practical Fragments once they are published.

But these successes should not breed complacency: at a round table chaired by Rod Hubbard (Vernalis and University of York) the topic turned to remaining challenges (or opportunities). Chief among these was advancing fragments in the absence of structure. Multiprotein complexes came up, as did costs in terms of time and resources that can be required even for conventional targets. Results from different screening methods often conflict, and choosing the best fragments both in a library and among hits is not always obvious. Finally, chemically modifying fragments can be surprisingly difficult, despite their small size.

I could go on much longer but in the interest of space I’ll stop here. Please add your thoughts, and mark your calendars for next year, when DDC returns to San Diego from April 2-6!

24 April 2017

Fragment optimization without purification

Compound purification can be a major hassle: separating the desired product from starting materials, reagents, and byproducts often takes far longer than making the compound in the first place. As we’ve previously noted, this is especially true for small, polar fragments – which are particularly attractive for drugs. Two new papers address this challenge. (Shameless plug: my company Carmot Therapeutics also has a solution to this problem.)

In J. Med. Chem., Paul Brough and Vernalis colleagues describe their discovery of inhibitors of all four isoforms of pyruvate dehydrogenase kinase (PDHK), potential targets for diabetes and oncology. The ATP-binding site of these four enzymes is similar to that of oncology target HSP90, in which Vernalis has a long-standing interest.

A screen of 1063 fragments (each at 0.5 mM) against PDHK-2 yielded 78 hits that were positive in three different NMR-based assays and also ATP-competitive. These yielded a whopping 43 structures when soaked into crystals of the related isoform PDHK-3. Compound 6 was one, and the binding mode was very similar to that previously seen for the same fragment with HSP90. Fragment growing rapidly led to molecules such as compound 8, with low micromolar potency. This compound was almost equipotent against HSP90, but modeling suggested that it might be possible to further grow this molecule in a direction that would be accommodated in the PDHKs but not in HSP90.

The next step was to make a bunch of analogs, and here's where avoiding purification becomes advantageous. Specifically, the researchers turned to off-rate screening (ORS), which entails making compounds and then testing the impure mixtures using surface plasmon resonance (SPR) to look for those which dissociate more slowly. Since off-rate is not dependent on the concentration of ligand, a low yield shouldn’t change the results of the assay.


An initial library of 56 compounds led to the discovery of compound 18, and subsequent libraries and medicinal chemistry ultimately yielded VER-246608, which is a potent pan-PDHK inhibitor. As designed, it is also completely inactive against HSP90. The molecule is described more thoroughly in this Oncotarget paper, which reveals that despite activity against PDHKs in cells, VER-246608 is not particularly effective at slowing the proliferation of cancer cells. Still, it does appear to be a useful chemical probe for further exploring the biology of the PDHKs.

Shifting methods but staying with the theme of assaying impure compounds brings us to a paper in SLAS Discovery by Sten Ohlson, Brian Dymock, and colleagues at Nanyang Technological University and the National University of Singapore. The protein tested was HSP90, and the method used was weak affinity chromatography, or WAC (see here, here, and here).

Like SPR, WAC also uses an immobilized protein. However, whereas SPR provides the (kinetic) off-rate, WAC provides the (thermodynamic) dissociation constant, which is calculated from the change in retention time of the molecule as it passes through a column containing protein-bound resin. In this case the researchers synthesized a mixture of five different compounds which varied from 7-24% of the mixture. This crude sample was analyzed by WAC, and the resulting dissociation constants, ranging from 48-147 µM, were satisfactorily similar to the values obtained using pure compounds.

Both of these approaches should accelerate screening and facilitate the analysis of complicated mixtures, such as natural product extracts. It will be fun to watch for more examples.

17 April 2017

Fragments vs PRC2 revisited: a chemical probe

Earlier this year we highlighted a paper from Novartis in which ligand deconstruction was used to deconstruct an HTS hit against the epigenetic target methyltransferase polychrome repressive complex 2 (PRC2). In a new J. Med. Chem. paper, Ying Huang and Novartis colleagues report a similar approach on a different HTS hit, ultimately yielding a promising chemical probe.

Compound 7 was identified as one of about 1400 hits from a high-throughput biochemical screen of 1.4 million molecules (described here at PLOS ONE). Crystallographic studies revealed that, like the previous molecule, this one also binds in the site on the EED subunit of PRC2 that normally recognizes trimethylated lysine 27 on histone H3.


Crystallography also suggested that the left half of the molecule didn’t seem to be making productive interactions with the protein. Lopping this off actually increased the activity and dramatically improved the ligand efficiency. Fragment-sized compound 8 was then subjected to extensive medicinal chemistry, ultimately resulting in EED226. Crystallography revealed that the binding modes of the initial hit and the final molecule are quite similar.

In addition to good biochemical activity, EED226 also shows good cell potency and impressive selectivity against other histone methyltransferases, kinases, and unrelated safety targets. It also shows excellent oral bioavailability in mice and acceptable pharmacokinetic properties. The compound caused complete tumor regression in a mouse xenograft model.

A separate paper in Nat. Chem. Biol. further characterizes EED226. A chemoproteomics study of a labeled version of EED226 revealed that it is remarkably selective for the PRC2 complex in human cell lysates. Also, EED226 is active against mutant cell lines that are resistant to other PRC2 inhibitors currently in the clinic, which are competitive with the cofactor rather than the trimethylated lysine residue. In fact, EED226 can bind to the PRC2 complex simultaneously with these other inhibitors, so dosing both together could give improved efficacy and slow the emergence of resistance.

As with the previous post on this target, the discovery of EED226 is a nice example of fragment-assisted drug discovery (FADD). Unlike that case, in which fragmentation led to an initial loss in potency, here trimming back the molecule paid immediate dividends.

Artists often talk about finding a sculpture within a stone by cutting away excess material. It is rewarding to see that chemists can use the same strategy.

10 April 2017

Unexplored but promising fragments

Sir James Black famously said that the best way to find a new drug is to start with an existing one. A drug not only has to bind to a target with reasonable affinity, it also has to survive an onslaught of metabolic insults – and avoid doing too much collateral damage. Compound libraries are often populated with derivatives of known drugs, but as Richard Taylor and collaborators at UCB and Bohicket Pharma Consulting show in a recent J. Med. Chem. paper, there is plenty of untapped chemical real estate out there.

The researchers started by deconstructing all FDA-approved drugs into component rings. As they’ve previously shown (and presented), this gives a surprisingly small set: just 95 monocyclic rings (such as benzene and succinimide), 124 bicycles (purine and quinazoline), and 58 tricycles.

Next, they computationally combined these rings with one another in various ways, focusing on monocycles and bicycles to maintain low molecular weights. For example, one set contained all combinations of drug-derived monocycles connected either to another monocycle or to a bicycle by linkers containing up to four bonds. That provides about 14.4 million possibilities. Among commercially available molecules, about 1.6 million are monocycles connected to another monocycle or a bicycle by up to four bonds, but many of these monocycles and bicycles have never appeared in a drug. Remarkably, the overlap among the computed and commercial sets is less than 58,000 compounds: only about 3% of relevant commercial compounds contain two rings which have both appeared in a drug.

Of course, chemical space is large; how do things fare among fragments? The researchers examined a subset of theoretical molecules having two monocycles or a monocycle and bicycle connected by just two bonds and with molecular weights less than 280 Da. They also allowed “decoration” with a fluorine atom or a methyl, amino, or hydroxyl group. This provided 421,929 molecules – a sizable number but, as the researchers note, a small enough set to be tractable with computational docking approaches.

Even with this fragment set the commercial availability is less than 1%. In fact, less than half of the decorated monocycles and less than 40% of the decorated bicycles are for sale. This seems like a ripe business opportunity for enterprising vendors of fragments. Unfortunately the researchers do not provide a comprehensive list of structures, but the analysis would be relatively straightforward to repeat.

This paper draws similar conclusions to one we highlighted a couple years ago focused on kinase inhibitors. Some chemists enjoy the challenge of making entirely novel molecules, but it may be worth taking another look at more conventional pharmacophores, particularly when they are connected in new ways.