12 August 2019

Achieving maximum diversity with minimum size

One theoretical advantage of fragment-based drug discovery is the ability to efficiently explore chemical space: there are vastly fewer possible fragment-sized molecules than lead-sized molecules. That said, even fragment space is daunting; the number of possible molecules with up to 17 non-hydrogen atoms is about three orders of magnitude larger than the largest computational screen. Maximizing diversity is thus a key goal in designing fragment libraries, but how do you actually do this? A new open-access paper in Molecules by Yun Shi and Mark von Itzstein at Griffith University provides a practical new approach.

As the researchers point out, diversity itself can be a slippery concept. Functional diversity (ie, what targets are bound) is important but hard-won knowledge. Physicochemical diversity is by definition limited for fragments. That leaves structural diversity, as defined by “molecular fingerprints.” These can be as simple as the presence or absence of a fluorine atom, or can require complicated calculations involving, say, the distance between a hydrogen bond donor and acceptor in the lowest energy conformation of a molecule. In their paper the researchers focus on “extended-connectivity” fingerprints, which take into consideration the physical connectivity between different types of atoms.

But how can you actually quantify structural diversity? One possibility is by comparing molecules to see how different they are, as used for example in Tanimoto similarity assessments. Each additional molecule would be chosen to be least similar to those in a library. Alternatively, one could consider “richness,” how much of chemical space is covered, by calculating how many unique structural features (such as specific bond connectivities) are represented. Each additional molecule would be chosen to provide as many new molecular fingerprints as possible. Shi and von Itzstein propose a third approach, “true diversity,” that considers the number of unique features as well as their proportional abundances. In other words, a library with a higher true diversity would have a “more even distribution of proportional abundances.” The researchers note that this approach has been used in ecology for decades.

To see how their approach performs, the researchers started with a set of 227,787 commercially available fragments, all of which were roughly rule-of-3-compliant and scrubbed of undesirable functionalities. They also considered a subset of 47,708 fluorine-containing fragments. For both sets, they then assessed structural diversity as a function of increasing fragment library size using Tanimoto similarity, richness, and true diversity, as well as random sampling.

Naturally, as the size of a fragment library rose, the diversity increased. As expected, applying Tanimoto similarity or richness led to greater diversity at a smaller library size than did random sampling. This was even more true for true diversity. Interestingly, true diversity reached a maximum at 8.8% or 15.7% (for the full and fluorinated libraries) and then began to decline. This conceptually makes sense because commercial compounds themselves are unlikely to be truly diverse.

More importantly, just 1% or 2.5% of fragments were sufficient to achieve the same true diversity as the full sets. This corresponds to 2052 fragments for the complete commercial set, the structures of which are provided in the supplementary material. As the researchers note, this is comparable to the size of many commonly used fragment libraries.

The method is computationally inexpensive (it runs on a desktop), and should be a useful tool for both building and curating fragment libraries, real and virtual. Of course, diversity is not everything, and it probably makes sense to include privileged pharmacophores even at the cost of lower diversity. But as Lord Kelvin said, “when you can measure what you are speaking about, and express it in numbers, you know something about it.” This paper provides a quantitative approach for measuring diversity.

05 August 2019

Fragments vs RAS family proteins: A chemical probe

RAS family proteins are considered a holy grail of oncology research. Way back in 2012 we discussed a couple papers disclosing low affinity fragments that bind in a small, shallow, polar pocket found in KRAS, NRAS, and HRAS. At the time we wondered “whether this is a ligandable site on the protein.” Last year we highlighted a paper proving that the site is, in fact, ligandable, as exemplified by the mid-nanomolar molecule Abd-7. A paper just published in Proc. Nat. Acad. Sci. USA by Darryl McConnell and collaborators from Boehringer Ingelheim and Vanderbilt University (including Steve Fesik, who published one of the 2012 reports) describes successful development of another ligand. (See here for a fun animated description set to music.)

Consistent with the “undruggable” reputation of RAS family proteins, a high-throughput screen of 1.7 million compounds failed to find anything useful. In contrast, a library of just 1800 fragments screened using STD NMR and MST identified 16 fragments that bind to an oncogenic mutant form of KRAS, as confirmed by 2-dimensional (HSQC) NMR. A separate HSQC NMR screen of 13,800 fragments identified several dozen more, though all the fragments from both screens have dissociation constants weaker than 1 mM. SAR by catalog led to amine-substituted indoles such as compound 11, which modeling suggested could form a salt bridge to an aspartic acid side chain.


The pocket in which all of these molecules bind, between the so-called switch I and switch II regions of KRAS, is much smaller than typical drug-binding sites, but modeling suggested that fragment growing could pick up an additional hydrogen bond, leading to compound 15. Crystallography confirmed the predicted binding mode of this molecule, and informed additional structure-based design, leading first to compound 18 and ultimately to BI-2852, with low or sub-micromolar affinity for wild-type and mutant KRAS, NRAS, and HRAS as assessed by ITC. The researchers also confirmed that the enantiomer is about 10-fold less potent, thereby providing a control compound. Commendably, the researchers have made BI-2852 and the enantiomer available (for free!) to the research community as a chemical probe.

A crystal structure of KRASG12D bound to BI-2852 (cyan) compared with Abd-7 (magenta) reveals how shallow the pocket is; both molecules are largely surface-exposed. The conformational flexibility of the protein is also interesting: Abd-7 would not be accommodated by the protein conformation bound by BI-2852.

The biology is also quite interesting – and complicated. RAS family proteins behave as molecular switches, cycling between the “on” (GTP-bound) state and the “off” (GDP-bound) state, with these transitions assisted by other proteins. On-state RAS drives cell-proliferation and survival. Molecules that bind at the switch I/II pocket block the transition from off to on, but they also block the transition from on to off. Thus, cellular effects are modest. Moreover, BI-2852 hits all RAS isoforms, which could lead to unacceptable toxicity in animals.

This is a lovely paper, but I do quibble that the promise of the title – “drugging an undruggable pocket on KRAS” – remains to be demonstrated. First, both the biochemical and cell-based potency need to be further improved. As the molecule is already large, gaining this needed potency could come at the cost of physicochemical properties. Indeed, the researchers do not discuss the pharmacokinetics of BI-2852. And finally, as the authors themselves note, they will probably need to improve selectivity to spare one or more wild-type RAS isoforms.

What this work does establish indisputably is that the switch I/II pocket is ligandable, though not without effort, as indicated by the 42 authors. Whether or not the site is actually druggable may require another seven years to determine.

29 July 2019

SAR by WaterLOGSY?

Among ligand-based NMR methods, WaterLOGSY is nearly as popular as STD NMR. Normally the information obtained is limited: does a given small molecule bind to a protein or not? In a new paper in J. Enzyme Inhib. Med. Chem., Isabelle Krimm and collaborators at the Université de Lyon and University of York try to wring more data from this common experiment.

In WaterLOGSY, magnetization is transferred from water, to protein, and then to bound ligand. This can happen through multiple mechanisms, and even talented NMR spectroscopists have told me they have trouble understanding exactly what is going on. In short, the WaterLOGSY spectra of molecules bound to proteins show a change in sign compared to molecules that don’t bind. Examining ligands in the presence and absence of protein can thus provide evidence for whether a ligand binds.

The researchers go beyond this simple qualitative approach and look at changes in peaks corresponding to specific hydrogen atoms in each ligand. They define a “WLOGSY factor,” which shows an inverse correlation to solvent exposure. In other words, a smaller WLOGSY factor means that a given hydrogen atom in a ligand is more exposed to water, and thus less exposed to protein. If all the hydrogen atoms in a bound ligand have the same WLOGSY factor, this suggests either multiple binding modes, or that the ligand is completely enclosed by the protein. If, on the other hand, different hydrogen atoms in a bound ligand have different WLOGSY factors, this could provide information on the binding mode. This analysis is conceptually similar to the STD epitope mapping the Krimm lab described several years ago, and STD experiments were also run on the proteins here for comparison.

To validate the approach, the researchers tested six proteins (with molecular weights ranging from 22 to 180 kDa) for which fragment ligands had been previously identified with affinities from 50 µM to worse than 1 mM. Screens were done using 400 µM fragment and 5 to 20 µM protein. (NMR aficionados, please see the paper for details on the effects of mixing times and ligand exchangeable protons.)

The results look pretty impressive: for PRDX5, HSP90, Bcl-xL, Mcl-1, and glycogen phosphorylase, the ligand hydrogen atoms previously shown to be solvent exposed from crystallographic or two-dimensional NMR structures do in fact show reduced WLOGSY factors. In the case of human serum albumin, a ligand showed uniform WLOGSY factors, suggesting multiple binding modes, as expected given the multiple promiscuous binding sites on this protein.

To a non-NMR spectroscopist such as myself, this seems like a useful approach for obtaining binding information in the absence of crystallographic data. It also seems easier to run than the LOGSY titration we highlighted a couple years ago. But the first word of this blog is “Practical.” We recently discussed work demonstrating that STD NMR data is perhaps not as easily interpretable as many assume. Have you tried anything like this yourself, and if so how well does it actually work?

22 July 2019

Fragments vs the PWWP1 domain of NSD3: a chemical probe

Epigenetics is a topic we’ve covered frequently at Practical Fragments. Much attention has been focused on bromodomains, which recognize acetylated lysine residues. However, lysine side chains are also methylated to affect gene expression. The PWWP1 domain of the protein NSD3 (NSD3-PWWP1) recognizes these modified lysines. This protein is amplified in several tumour types, and so makes an intriguing cancer target. At the CHI DDC conference last year Jark Böttcher presented how Boehringer Ingelheim and a large multinational group of collaborators developed a chemical probe for NSD3. The story now appears in Nat. Chem. Biol. (and see here for a fun animated short set to music).

The researchers started by screening a library of 1899 fragments against NSD3-PWWP1. STD NMR (at 0.25 mM of each fragment, in pools of four) as well as differential scanning fluorimetry (at 0.5 mM of each fragment) resulted in 285 and 20 hits, respectively. Two-dimensional NMR was used to confirm hits. Interestingly, only three fragments were identified from both STD-NMR and DSF, and these did not confirm – a cautionary reminder that screening orthogonal methods is not necessarily the best path.

Fortunately, 15 fragments not only confirmed, but also caused the same changes to the 2D-NMR spectra as a histone-derived peptide containing a dimethyl-lysine residue, suggesting that the fragments bind at the recognition site for modified lysines. Those fragments with dissociation constants better than 2 mM were pursued crystallographically, and some of the successes included compound 4. This molecule was used in a virtual SAR-by-catalog screen of internal compounds. Of the 601 fragments experimentally tested, compound 8 was the most potent. Crystallography confirmed that the compound binds in the expected site, and further structure-based design ultimately led to BI-9321.


BI-9321 was put through a battery of tests. Affinity was confirmed in biochemical, SPR, and ITC assays, and crystallography revealed the binding mode to be similar to the initial fragment. BI-9321 was selective for NSD3-PWWP1 when tested against 14 other PWWP domains, and showed no activity against 35 protein methyltransferases, 31 kinases, and 48 bromodomains. Solubility, in vitro metabolic stability, permeability, and plasma protein binding all look good.

Multiple assays also demonstrated selective target engagement in cells at a concentration of around 1 µM. BI-9321 showed downregulation of MYC mRNA levels, though the effect was both modest and transient. Antiproliferative activity was also observed in cells, and the effects were synergistic with a bromodomain inhibitor. Moreover, these effects were only seen in NSD3-dependent cells, suggesting that the activity is on-target and that the compound is not generally cytotoxic.

All of this makes BI-9321 an attractive chemical probe, at least for cell-based assays. More work will need to be done to improve potency and further understand the biology. Laudably, to this end, the researchers have made the molecule publicly available.

15 July 2019

Fragments vs viral protein EBNA1

Epstein-Barr virus (EBV) infects more than 90% of adults. In most cases it remains latent, but even then it expresses genes that cause cellular proliferation, which can lead to cancer. In fact, up to 2% of human cancers are caused by the virus. In a recent paper in Sci. Transl. Med., Troy Messick, Paul Lieberman (both at the Wistar Institute) and a large group of collaborators (including Teddy Zartler) take aim at this pathogen.

The researchers focused on the viral protein EBNA1, a DNA-binding protein essential for viral replication as well as host cell survival. They started with a virtual screen of 1500 fragments from Maybridge, and then did fragment soaking of the top 100 hits. Happily, this resulted in structures of 20 fragments in four separate sites on the protein. Less happily for modelers, none of the fragments bound as predicted – more grist for the “crystallography first” argument.

A dozen fragments bound in a deep hydrophobic pocket, and most of them contained an acidic moiety that made hydrogen bond contacts to conserved threonine and asparagine residues that normally contact the DNA backbone. Merging two of these fragments led to VK-0497, which disrupts binding of EBNA1 to DNA at sub-micromolar concentrations. Crystallography confirmed that it bound as expected. The molecule contains a potentially unstable pyrrole, and replacing this with an indole and growing led to molecules such as VK-1248, with high nanomolar activity in the DNA-binding assay. Additional biophysical techniques including SPR, ITC, and 2-dimensional (HSQC) NMR confirmed binding for this and related compounds.


The carboxylic acid moiety likely reduces permeability across cell membranes, and indeed the compounds showed no activity in cells. However, methyl esters were found to be rapidly cleaved by intracellular esterases, and these prodrugs were tested in a variety of assays.

The prodrugs inhibited proliferation of EBV-positive human cells but had no effect on non-infected cells. The prodrugs also reduced expression of both viral and host proteins. More importantly, they inhibited tumor growth in xenograft models using four different cancer cell lines, two of which were patient-derived. Prolonged dosing over as long as eight weeks showed a sustained effect, which is reassuring in terms of drug resistance. Finally, the molecules could effectively be combined with existing drugs or radiation.

There is still much work to be done, not just in terms of potency but also further pharmacokinetics, pharmacology and toxicology. And as the researchers acknowledge, xenograft models are regrettably poor surrogates for humans. Still, this is an interesting approach, and hopefully further work will be done on this series, or at least the target.

08 July 2019

Stabilizing apolipoprotein E4 with fragments

Among fragment-derived drugs that have entered the clinic, BACE1 inhibitors are well-represented. Sadly, multiple drugs targeting this protein have failed to show efficacy against Alzheimer’s disease. That said, every drug that has been thrown at Alzheimer’s has failed to slow the disease, so perhaps we need to think more boldly. An example was published recently in J. Med. Chem. by Andrew Petros, Eric Mohler, and colleagues at AbbVie.

The researchers were interested in apolipoprotein E4 (apoE4), one of three isoforms found in humans. Folks who have two apoE4 alleles are at increased risk for Alzheimer’s, suggesting that the protein might make a good drug target. Unfortunately, although it is known to be a lipid carrier, its precise function is unclear. What is known is that apoE4 is less stable to thermal denaturation than apoE2 or apoE3, so the team set out to find molecules that would stabilize the protein. This being AbbVie, they used two-dimensional NMR to find fragments.

The methyl groups of all the isoleucine, leucine, methionine, and valine residues in apoE4 were 13C labeled, and the researchers looked for changes in the 13C-HSQC spectra upon addition of fragments; just over 4000 were screened in pools of 12, each at 1 mM. Of the dozen or so hits, compound 1 was among the best.


NMR titration studies revealed an affinity just under 1 mM, while SPR suggested slightly stronger binding. As hoped, compound 1 raised the melting temperature of apoE4. Adding the fragment also altered the kinetics of liposome breakdown, causing the protein to behave more like apoE2 and apoE3. Although this assay isn’t necessarily physiologically relevant, the reasoning is that causing apoE4 to behave more like the other isoforms may be useful.

A crystal structure of compound 1 bound to apoE4 revealed the fragment to be binding in a small pocket, and growing led to compound 2, with a slightly improved affinity. Introduction of polar substituents to interact with a nearby aspartic acid side chain led to compound 8, with low micromolar affinity (assessed by NMR). This molecule also stabilized apoE4 with respect to thermal denaturation.

As noted above, it is not entirely clear why apoE4 is associated with Alzheimer’s, but researchers had previously found that overexpression in a neuronal cell line caused release of the inflammatory cytokines IL-6 and IL-8. When human induced pluripotent stem cell (iPSC)-derived astrocytes carrying two copies of apoE4 were treated with compound 8, release of IL-6 and IL-8 cytokines was reduced to levels similar to those from iPSC-derived astrocytes carrying two copies of apoE3. The compound also showed no toxicity, even at relatively high concentrations (100 µM).

There is still a tremendous amount to do: affinity needs to be improved considerably, and permeability is also mentioned as an issue. Moreover, the highly polar nature of compound 8 will likely make transport across the blood-brain barrier challenging. Optimizing activity against a target whose function is poorly understood will present a host of problems. But if it were easy, Alzheimer’s disease would not be the scourge that it is. Practical Fragments salutes thinking outside the box, and wishes those involved the best of luck.

01 July 2019

Fragment events in 2019 and 2020

We're already halfway through 2019, but there are still some excellent upcoming events, and 2020 is taking shape, so start planning your calendar!

2019
September 1-4: BrazMedChem2019 will be held in the Brazilian holiday destination of Pirinopolis, and will include a section on FBLD.

September 17-19: CHI's Seventeenth Annual Discovery on Target takes place in Boston. Multiple biological targets are covered, and there are also more general talks on a variety of topics of interest to readers, including two sessions on fragments. You can read my impressions of last year's event here.

November 12-15: FBDD Down Under 2019 will take place in beautiful Melbourne. This is the third major FBDD event in Australia, and given the success of the first, I expect it to be excellent.

November 13-15: Although not exclusively fragment-focused, the Seventh NovAliX Conference on Biophysics in Drug Discovery will have several relevant talks, and will be held in the lovely city of Kyoto. You can read my impressions of the 2018 Boston event here, the 2017 Strasbourg event here, and Teddy's impressions of the 2013 event herehere, and here.

2020
April 13-17CHI’s Fifteenth Annual Fragment-Based Drug Discovery, the longest-running fragment event, will be held in San Diego. You can read impressions of 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 20-23: FBLD 2020 will be held for the first time in the original Cambridge (UK). This will mark the eighth in an illustrious series of conferences organized by scientists for scientists. You can read impressions of FBLD 2018FBLD 2016FBLD 2014,  FBLD 2012FBLD 2010, and FBLD 2009.

December 15-20: The second Pacifichem Symposium devoted to fragments will be held in Honolulu, Hawaii. The Pacifichem conferences are 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? Add it to the comments or let us know!

23 June 2019

New chemistries for covalent fragments

Good things come in threes, and since our last two posts have covered covalent fragments, we thought we’d continue the theme with two more papers on the topic.

The first, in MedChemComm by György Keserű and colleagues at the Hungarian Academy of Sciences, is actually a companion to a paper we covered at the end of last year. The researchers were interested in electrophilic heterocycles, and assembled a library of 84, of which 57 were commercial and the rest were synthesized. In their previous paper, the researchers focused on reactivity against proteins. This paper is focused more on aqueous stability and intrinsic reactivity with glutathione, a biologically important thiol. The paper includes handy figures and tables summarizing reactivity rate constants and half-lives (at pH 7.4). These should be useful for selecting warheads that are sufficiently reactive as to be able to label proteins, but not so reactive as to label many proteins nonselectively.

The researchers also did computational studies to try to understand different trends in reactivity. And if you’re interested in testing the compounds yourself, they note that the “library is available for screening against relevant targets upon request from the authors.”

The second paper, by Philippe Roche, Xavier Morelli, and collaborators at Aix-Marseille University, was published in J. Chem. Inf. Model. Last year we highlighted their diversity-oriented target-focused synthesis (DOTS) approach, which combines virtual screening with automated synthesis to rapidly generate new compounds for testing. They have now expanded this approach (called CovaDOTS) to focus on covalent modifiers.

Conceptually, CovaDOTS is akin to fragment linking, in which one of the “fragments” is the nucleophilic residue in the target protein. The process starts with a known noncovalent ligand which is computationally grown by attaching it to commercially available building blocks that contain reactive warheads. These new molecules are then linked with the side chain of an amino acid (cysteine and serine in the paper), and the assemblies are docked against the protein to find molecules that fit well. Ultimately, the best would be resynthesized and tested.

The researchers applied CovaDOTS to three proteins for which covalent and non-covalent ligands had been previously characterized crystallographically. In the case of two kinases, EGFR and ERK2, the program performed well, with the “correct” (i.e., published) ligand observed in the top 14% and 4.4% of hits. For the serine peptidase PREP, the published ligand was the top hit of 303 molecules scored. In all three cases, the predicted binding mode also closely resembled the experimentally determined structure.

One limitation of CovaDOTS is that, as currently implemented, it considers only commercially available building blocks that also “possess both a warhead and an activated function compatible with the selected chemical reaction(s).” It would be interesting to combine the approach with the “make-on-demand” molecules we discussed a few months ago. And of course, it will be interesting to see real-world examples of how the program performs, in addition to these retrospective case studies.

This ends the June 2019 covalent fragment trilogy, but I think it’s fair to say that covalent fragments have a bright future. Look forward to many sequels!

17 June 2019

Screening irreversible covalent fragments, computationally and experimentally

Last week we discussed a paper that characterized commercially available irreversible fragments and screened them against ten proteins. The “warheads” used were either acrylamides or chloroacetamides. This week we’re continuing the theme of irreversible fragments with two papers, each using different warheads.

The first paper, published in Bioorg. Med. Chem. Lett. by Alexander Statsyuk and collaborators at University of Houston, Schrödinger, and Northwestern University, uses a computational approach.

Research we previously highlighted from the Statsyuk lab found that methyl vinylsulfones have a narrower range of reactivities than – for example – acrylamides. This property is important  to ensure that differences in binding to a target are caused by (specific) noncovalent interactions rather than mere differences in warhead reactivity. For the current campaign, the researchers constructed a virtual library in which 1648 commercially available carboxylic acids were coupled to H2N-CH2-CH=CH-SO2-CH3.

The researchers used the CovDock program from Schrödinger to dock the fragments. (Another computational tool for doing so is DOCKovalent, which we described back in 2014.) The specific target chosen was cathepsin L, a cysteine protease which has been implicated in a variety of diseases from cancer to osteoporosis. The virtual screen yielded 33 high-scoring compounds, five of which were synthesized based on price and diversity. Unfortunately, most of these had “solubility issues,” but compound 11 did show time-dependent inhibition of cathepsin L. The researchers also found that the racemic methyl substituent could be removed (compound 13), suggesting that compound growing might be productive. Compound 13 was also selective against three other cysteine proteases.

The second paper, by David House, Katrin Rittinger and collaborators at GlaxoSmithKline, the Francis Crick Institute, and Cellzome, was published in J. Am. Chem. Soc. The researchers were interested in protein ubiquitination, in which the small protein ubiquitin is conjugated to various other proteins to cause a variety of effects depending on the context. The biology is fiendishly complex, but the final step is done by an E3 ubiquitin ligase, of which there are more than 600 in human cells. Needless to say, selective chemical probes would be useful.

The researchers were specifically interested in LUBAC, an RBR E3 ubiquitin ligase which conjugates ubiquitin to proteins with an N-terminal methionine to modulate cellular pathways important in cancer and inflammation. The ligase itself actually consists of three protein subunits, with HOIP containing the catalytic cysteine residue. Although a couple inhibitors had been previously reported, the researchers found these to be nonspecific. Thus, they built and screened their own fragment library. For a warhead, they chose the 4-aminobut-2-enoate methyl ester, which Statsyuk had previously shown has a narrow range of reactivity and is about 10-fold less reactive than the vinylsulfones discussed above. The researchers constructed a small set of 104 fragments, grouped them into pools of 4 or 5, and screened these against 2 µM HOIP at 20 µM each using intact protein mass spectrometry. Compound 5 was one of the best hits.

Compound 5 functionally inhibited HOIP and was selective against about a dozen other cysteine-containing enzymes. The researchers obtained a crystal structure of the molecule bound to HOIP, which confirmed covalent binding to the active site cysteine. Limited SAR studies led to a slightly more potent analog (containing a six-membered ring instead of a five-membered ring), and this molecule showed pathway inhibition in a cell-based assay with EC50  = 37 µM. Activity-based profiling in two cell lines revealed only 8 or 11 proteins that were significantly modified by the compound in addition to HOIP.

The molecules in both of these papers still require considerable work to become chemical probes, let alone development candidates. Nonetheless, they are useful starting points, and together demonstrate the increasing interest and utility of irreversible covalent fragments.

10 June 2019

Characterizing and screening commercially available irreversible covalent fragments

A few years ago we highlighted the utility of irreversible fragments. Because these molecules form covalent bonds with their targets, they can be more effective than similarly sized noncovalent molecules at inhibiting proteins. However, compared with conventional fragments, the quality and quantity of commercial irreversible fragments is limited. This is changing, as described (open access!) by Nir London (Weizmann Institute of Science) and a large, multinational group of collaborators in J. Am. Chem. Soc.

The researchers assembled a collection of 993 fragments from Enamine, all of which contained a cysteine-reactive warhead, either a chloroacetamide (76%) or an acrylamide (24%). The molecules were largely rule of three compliant, even with the warhead included.

A major concern with screening irreversible fragments is that binding to the target protein can be dominated by the inherent reactivity of the warheads rather than non-covalent (and presumably target-specific) interactions from the fragment. Indeed, a previous study found that the reactivities of acrylamides ranged over more than three orders of magnitude. To assess fragments for this, the researchers developed a rapid, plate-based spectrophotometric assay based on labeling the reduced form of Ellman’s reagent. Not surprisingly, the chloroacetamides tended to be more reactive than the acrylamides, but overall the reactivity range across both classes was a relatively modest ~100-fold.

Next, the researchers screened their library against ten cysteine-containing proteins. Fragments were screened in pools of five (200 µM each) with 2 – 10 µM protein for 24 hours at 4 °C. As with Tethering, intact protein mass spectrometry was used to identify hits, which were found for seven of the ten proteins. Hit rates ranged from 0.2 to 4%.

Not surprisingly for fragments, some hits were promiscuous: they strongly labeled two or more proteins. However, these represented less than 3% of the library. Surprisingly, promiscuity did not correlate with reactivity, and in fact some of the most reactive fragments did not label any of the proteins. This suggests that non-covalent interactions are playing a role in promiscuity, and indeed many of the frequent hitters were aminothiazoles – which have previously been found to be promiscuous.

The researchers also screened their fragments (at 10 µM) against three cell lines, and here they did see a correlation with reactivity, with the most reactive fragments tending to be more toxic.

Next, the researchers began optimizing hits against two targets. The first, OTUB2, is a deubiquitinase (DUB) implicated in diverse diseases from amyotrophic lateral sclerosis to diabetes to cancer. The primary screen yielded 47 hits which labeled at least 50%, of which 37 were quite selective. Co-crystal structures were solved for 15 fragment-protein complexes, and two shared a hydrazide moiety (as in PCM-0102954) which made multiple hydrogen bonds with the protein. Two rounds of SAR-by-catalog eventually led to OTUB2-COV-1, which inhibited the enzyme with a respectable kcat/KI = 3.75 M-1 s-1. Despite containing a chloroacetamide, the molecule labeled just 26 of 2998 cysteines in proteins detected in a cell-based proteomic assay.

The researchers also found 36 fragment hits against NUDT7, a protein potentially associated with diabetes, and many of these stabilized the protein in a differential scanning fluorimetry (DSF) assay. Crystal structures were obtained for several, and compound PCM-0102716 showed an overlap with the non-covalent molecule NUDT7-REV-1 derived from a previous crystallographic fragment screen. When the researchers merged these, the resulting NUDT7-COV-1 showed low micromolar inhibition and rapid labeling (kcat/KI = 757 M-1 s-1). This is all the more impressive given that the original noncovalent hit showed no activity. NUTDT7-COV-1 also showed target engagement in a cell assay, and hit only 37 of 2025 detected cysteine residues in a proteomics screen.


This is a nice, thorough paper, though I suspect people in industry will be wary of the chloroacetamides that form the bulk of the library. Nonetheless, chemical structures and reactivity data for all the fragments are reported in the supporting information, making this a useful resource for anyone wishing to dip their toes into covalent fragment screening.

03 June 2019

Is thermodynamic data useful for drug discovery?

Just over a decade ago Ernesto Freire suggested that small molecules whose binding energy is dominated by the enthalpic – rather than the entropic – term make superior drugs. He also suggested that such molecules may be more selective for their target. But the backlash came quickly, and a couple years ago we wrote that focusing on thermodynamics probably isn’t particularly practical. A new perspective in Drug Disc. Today by Gerhard Klebe (Philipps-University Marburg) revisits this topic.

Klebe suggests that enthalpy was initially embraced “because readily accessible and easily recordable parameters are much sought after for the support of the nontrivial decision over which molecules to take to the next level of development.” (I would be interested to know whether sales of isothermal titration calorimetry (ITC) instruments spiked around 2010.) Unfortunately, both theoretical and practical reasons make thermodynamic measurements less useful than hoped.

First, and as we noted previously, “in an ITC experiment… the balance sheet of the entire process is measured.” In particular, water molecules – which make up the bulk of the solution – can affect both enthalpic and entropic terms. Klebe describes an example in which the most flexible of a series of ligands binds with the most favorable entropy to the target protein; this is counterintuitive because the ligand adopts a more ordered state once bound to the protein. It turned out that in solution the ligand traps a water molecule that is released when the ligand binds to the protein, thus accounting for the favorable entropy.

Indeed, water turns out to be a major confounding factor. We’ve previously written about “high-energy” water; Klebe notes that an individual water molecule can easily contribute more than 2 kcal/mol to the overall thermodynamic signature. And of course, proteins in solution are literally bathed in water. The structure of this water network, which may change upon ligand binding, is rarely known experimentally, but optimizing for it can improve affinity of a ligand by as much as 50-fold. Conversely, attaching a polar substituent to a solvent-exposed portion of a molecule to improve solubility sometimes causes a loss in affinity, and Klebe suggests this can be due to disruption of the water sheath.

Beyond these theoretical considerations, experimental problems abound. We’ve previously discussed how spurious results can be obtained when testing mixtures of ligands in an ITC experiment, but even with single protein-ligand complexes things can get complicated. Klebe shows examples where the relative enthalpic and entropic components to free energy change dramatically simply because of changes in buffer or temperature. This means that the growing body of published thermodynamic data needs to be treated cautiously.

So what is to be done? First, thermodynamic data should always be treated relatively: “we should avoid classifying ligands as enthalpy- or entropy-driven binders; in fact, we can only differentiate them as enthalpically or entropically more favored binders relative to one another.”

Klebe argues that collecting data on a variety of ligands for a given target under carefully controlled conditions will be useful for developing computational binding models. This is important, but not the kind of work for which people usually win grants, let alone venture funding.

He also suggests that, by collecting thermodynamic data across a series of ligands, unexpected changes in thermodynamic profiles might reveal “changes in binding modes, protonation states, or water-mediated interactions.” Maybe. But it takes serious effort to collect high-quality ITC data. Are there examples where you’ve found it to be worthwhile?

27 May 2019

Fragments vs PKC-ι: A*STAR’s second series

Just over a year ago we highlighted work out of A*STAR describing a series of inhibitors for the cancer target protein kinase C iota (PKC-ι). We ended by mentioning that the group had a second undisclosed series. This has now been described in ACS Med. Chem. Lett. by Jacek Kwiatkowski, Alvin Hung, and colleagues.

Compound 1 was among the fragment hits from the high-concentration biochemical screen previously mentioned. Although the researchers did not have a crystal structure, they assumed that the aminopyridine moiety was acting as a hinge binder, which helped them produce a computational model. A simple replacement of the phenyl ring with a pyridyl ring led to compound 2, with a satisfying improvement in potency and ligand efficiency.


As it turned out lots of diverse moieties could be substituted in place of the phenyl, including indoles and phenols. This promiscuity led the researchers to propose that the added heteroatom was making a water-mediated hydrogen bond to the protein; the water could rotate to either accept or donate a hydrogen bond to the ligand. Unfortunately, further growing from this ring did not improve potency.

Returning to their model, the researchers sought to grow from the aminopyridine ring towards a hydrophobic region of the protein. Adding a phenyl group (compound 16) was tolerated, though did not improve the affinity. However, the model suggested that an aspartic acid might be accessible from the phenyl ring, and indeed adding a positively charged piperazine as in compound 19 led to a nearly 100-fold boost in affinity. Unfortunately, the compound’s permeability (measured in a Caco-2 assay) was low, and perhaps because of this it showed only weak antiproliferative activity against hepatocellular carcinoma cells.

Ultimately the researchers were able to solve the cocrystal structure of compound 19 with PKC-ι, which mostly confirmed the model: the aminopyridine interacts with the hinge region, and the second pyridyl moiety likely makes a water-mediated hydrogen-bond with the protein, although the low resolution of the structure makes this somewhat ambiguous. The added piperazine appears to interact with a different aspartic acid than the one targeted.

Although there is more work to be done, it is notable that the researchers were able to optimize a fairly weak fragment to a sub-micromolar compound in the absence of experimental structural information. As they note, “while the empirical SAR remained our ultimate guide in fragment optimization, the model aided the successful design of potent inhibitors.” This is another nice example supporting our 2017 poll results, and recent review, that drug hunters can successfully advance fragments without NMR or crystallography.

20 May 2019

SAR by STD: NOT

As noted last week, Practical Fragments has been on something of a crystallography binge. But according to polling, NMR is the most common fragment-finding method. And, according to a different poll, saturation transfer difference (STD) is the most popular NMR technique. Familiarity breeds complacency, and widespread assumptions go untested. A new paper in Front. Chem. by Jonas Aretz and Christoph Rademacher (Max Planck Institute and Freie Universität Berlin) suggests that this is a mistake.

In STD NMR, a protein is saturated by specific electromagnetic pulses, and the resulting magnetization transfers to bound ligands. Assuming that the bound ligands are in rapid equilibrium with ligands free in solution, this “saturation transfer” results in a reduction of NMR signal for the small molecule in the presence of protein compared to no protein. High affinity ligands will remain bound to the protein and thus be missed by STD NMR, but this is usually not relevant in FBLD, where most fragments bind with dissociation constants weaker than 10 µM.

A common assumption with STD NMR is that the strength of an STD signal increases with the affinity of the ligand (again, in affinity ranges between about 10 µM and 10 mM). Indeed, when STD NMR is used as part of a screening cascade, molecules showing the strongest effect are generally prioritized as hits. But is this assumption correct?

To find out, the researchers retrospectively analyzed a fragment screen against langerin, a carbohydrate-binding protein we discussed last year. When they plotted the STD amplification factor against the affinity (measured by SPR) for several dozen fragments, the resulting scatter plot showed no correlation.

Recognizing that experimental errors could obscure a true correlation, the researchers ran virtual STD experiments using COmplete Relaxation and Conformational Exchange MAtrix (CORCEMA) theory. They used well-characterized fragments with published crystal structures and affinities for some dozen diverse proteins. As they conclude, “varying saturation time, receptor size, binding kinetics, and interaction site… there were no conditions in which the STD NMR amplification factor correlated unambiguously with affinity.”

But it gets worse. When the researchers explored the effects of binding kinetics, they found that ligands with slower on-rates or off-rates also had lower STD signals. Several groups have advocated prioritizing compounds with slower-off rates, yet these are the very compounds STD is most likely to miss.

All in all this paper could go some way toward explaining the sometimes poor correlation between different fragment-finding methods.

That said, I’m no NMR spectroscopist, so I’m certainly not as qualified to comment on the importance of this paper as someone like Teddy, who co-wrote this how-to guide for STD NMR. I’d be interested to hear what NMR folks think, and whether we should rethink use of STD. In any case, this work is a useful reminder that skepticism is a scientific virtue.

13 May 2019

Crystallographic vs computational fragment screening

Several recent Practical Fragments posts have touched on crystallographic screening: from ultra-high concentration screening of “MiniFrags,” to an extensive analysis of fragment structures in the protein data bank, to an open-source effort to develop new antibiotics. A new paper in Phil. Trans. R. Soc. A by Tom Blundell and collaborators at University of Cambridge, the Diamond Light Source, University of Oxford, and several other institutes provides a useful synthesis and an interesting comparison with computational approaches.

The researchers were interested in the bacterial protein PurC, also known as SAICAR synthetase, which is essential for purine biosynthesis and is sufficiently different from its human orthologue to be an attractive antimicrobial target. The protein has an extended binding site that can accommodate ATP as well as its substrate CAIR and an aspartic acid. Using a traditional screening cascade, 960 fragments were screened at 5 mM in a thermal shift assay, resulting in 43 hits. Each hit was then soaked at 10 mM into crystals of PurC, resulting in 8 bound structures, all of which occupy the ATP-binding pocket. Isothermal titration calorimetry revealed dissociation constants as good as 178 µM, with a ligand efficiency of 0.39 kcal/mol/atom.

Next, the researchers ran a computational screen, Fragment Hotspot Maps. This confirmed the main fragment-binding site. Indeed, the crystallographically-identified fragments even make the hydrogen-bonding interactions predicted by the model. However, the computational approach also identified three other hot spots, two in the active cleft and one on the rear of the protein. There was also a “warm spot” next to the ATP-binding site. Are these real, or computational artifacts?

To address this question, the researchers screened fragments at a much higher concentration at XChem, and processed the data using the PanDDA software we’ve previously described. They screened two libraries of fragments at 30-50 mM: 125 “shapely” fragments and 768 “poised” fragments designed for rapid follow-up chemistry. The 8 hits from the first crystallographic fragment screen were also included. This exercise yielded structures for 35 fragments, 60% of which bound in the ATP-binding site, including all 8 of the previously identified ones. Most of the other fragments bound in shallow pockets or near crystallographic interfaces; only one of the other hot spots predicted computationally had a bound fragment, and that was present at low occupancy. Some hits made new interactions around the ATP-binding site, but none bound in the predicted warm spot. Unfortunately, the proportions of fragment hits coming from the two libraries are not broken out.

So in summary, both computational and crystallographic screening correctly identified the “hottest” hot spot, but each approach also identified additional sites that were not confirmed by the other. The researchers ask, “are these sites truly hot spots… or are they weak binding sites routinely seen in crystals?”

This is indeed the key question, and it would be interesting to see whether other computational approaches – such as FTMap or SWISH – are able to shed light on the matter.

06 May 2019

Fragments in the clinic: AZD5991

Venetoclax, the second fragment-based drug to reach the market, binds to and blocks the activity of the anti-apoptotic protein Bcl-2, allowing cancer cells to undergo programmed cell death. The drug is effective in certain cancers such as chronic lymphocytic leukemia and small lymphocytic lymphoma. However, a related protein called Mcl-1 is more important in other types of cancers. Like Bcl-2, it binds and blocks the activity of pro-apoptotic proteins, allowing cancer cells to survive even when Bcl-2 is inactivated. A paper in Nat. Comm. by Alexander Hird and a large group of collaborators (mostly at AstraZeneca) describes a successful effort to target Mcl-1.

Given that the researchers were targeting a protein-protein interaction, they took multiple approaches, including their own fragment-based efforts. They also characterized previously reported molecules, such as those the Fesik group identified using SAR by NMR (which we wrote about in 2013). A crystal structure of one of these revealed a surprise: two copies of compound 1 bound to Mcl-1, which had undergone conformational changes to accommodate the second molecule in an enlarged hydrophobic pocket.


Recognizing the potential synergies of linking these together, the researchers prepared a dimer of a related molecule, but unfortunately the affinity of this much larger molecule was actually worse. However, they wisely isolated and tested a side product, compound 4, and found that this had improved potency. A crystal structure of this molecule bound to Mcl-1 revealed that the pocket had expanded to accommodate the added pyrazole moiety. Since compound 4 adopted a “U-shaped” conformation, the researchers decided to try a macrocyclization strategy to lock this conformation and reduce the entropic penalty of binding. This produced compound 5, and adding a couple more judiciously placed atoms led to AZD5991, with a nearly 300-fold improved affinity. The molecule binds rapidly to Mcl-1 and has a relatively long residence time of about 30 minutes. A crystal structure reveals a close overlay with the initial compound 1 (in cyan).

In addition to picomolar affinity, AZD5991 showed excellent activity in a variety of cancer cell lines dependent on Mcl-1. The compound was tested in mouse and rat xenograft models of multiple myeloma and acute myeloid leukemia and showed complete tumor regression after a single dose. This is all the more remarkable given that AZD5991 is about 25-fold less potent against the mouse version of Mcl-1 than the human version. The molecule was also effective in cell lines resistant to venetoclax, and combining the two molecules caused rapid apoptosis in resistant cell lines. AZD5991 is currently being tested in a phase 1 clinical trial.

This paper holds several lessons. First, the researchers did extensive mechanistic work (beyond the scope of this post to describe) to demonstrate on-target activity. Second, although the initial dimerization strategy was unsuccessful, the researchers turned lemons into lemonade by pursuing a byproduct; we’ve written previously about how even synthetic intermediates are worth testing. Third, the macrocyclization and subsequent optimization is a lovely example of structure-based design and medicinal chemistry. And finally, the fact that the researchers started with a fragment-derived molecule reported by a different group is a testimony to the community nature of science. Last week we highlighted the Open Source Antibiotics initiative, which is actively encouraging others to participate in advancing their early discoveries. Good ideas can come from anywhere, and it takes a lot of them to make a drug.

29 April 2019

Help develop new antibiotics from fragments!

The state of antibiotic drug discovery is – to put it mildly – dangerously poor. Not only do you have all the challenges inherent to drug discovery, you’re dealing with organisms that can mutate more rapidly than even the craftiest cancer cells. And then there’s the commercial challenge: earlier this month the biotech company Achaogen filed for Chapter 11 bankruptcy, less than a year after winning approval for a new antibiotic.

As Douglas Adams’s Golgafrincham learned, complacency about microbial threats is suicidal. But what can any one of us do? Chris Swain, whom we’ve previously highlighted on Practical Fragments, is involved with a consortium of researchers called Open Source Antibiotics. Their mission: “to discover and develop new, inexpensive medicines for bacterial infections.” And they are asking for our help. More on that below.

The researchers initially chose to focus on two essential enzymes necessary for cell wall biosynthesis, MurD and MurE, both of which are highly conserved across bacteria and absent in humans. They conducted a crystallographic fragment screen of both enzymes at XChem, soaking 768 fragments individually at 500 mM concentration. As we’ve written previously, you’ll almost always get hits if you screen crystallographically at a high enough concentration.

For MurD, four hits were found, all of which bind in the same pocket (in separate structures). Interestingly, this pocket is not the active site, but adjacent to it. The binding modes of the fragments are described in detail here, and the researchers suggest that growing the fragments could lead to competitive inhibitors. The fragments also bind near a loop that has been proposed as a target for allosteric inhibitors, so growing towards this region of the protein would also be an interesting strategy.

MurE was even more productive, with fragments bound at 12 separate sites. (Though impressive, that falls short of the record.) Some of these sites are likely artifacts of crystal packing, or so remote from the active site of the enzyme that they are unlikely to have any functional effects. However, some fragments bind more closely to the active site, and would be good candidates for fragment growing.

If this were a typical publication one might say "cool," and hope that someone picks up on the work sometime in the future. But this, dear reader, is different.

The researchers are actively seeking suggestions for how to advance the hits. Perhaps you want to try running some of these fragments through the Fragment Network? Or do you have a platform, such as “growing via merging,” AutoCouple, or this one, that suggests (and perhaps even synthesizes) new molecules? Perhaps you want to use some of the fragments to work out new chemistry? The consortium has a budget to purchase commercial compounds, and will also accept custom-made molecules. In addition to crystallography, they have enzymatic assays, and are building additional downstream capabilities.

The Centers for Disease Control identifies antibiotic resistance as one of the most serious worldwide health threats. Some have called for a global consortium—modeled after the International Panel for Climate Change—to tackle the problem. But in the meantime, you can play a role yourself. If you would like to participate, you can do so here. The bugs are not waiting for us – and they are already ahead.

22 April 2019

A bestiary of fragment hits

What do fragment hits look like, and how do they bind? Fabrizio Giordanetto, David Shaw, and colleagues at D. E. Shaw Research were interested in these questions, and their answers are provided in a recent J. Med. Chem. paper (open access, and also covered well by Derek Lowe).

The researchers started by searching the protein data bank (PDB) for the word “fragment” and selecting higher resolution structures (at least 2.5 Å) with a ligand containing 20 or fewer non-hydrogen atoms. Those of you who have done bibliometric searchers will appreciate that a lot of manual curation is required, and the initial list of 5115 complexes was ultimately winnowed down to 489, with 462 unique fragments and 126 unique proteins, about two-thirds solved to ≤2.0 Å resolution.

In contrast to a previous study, only a minority (18%) of proteins contained more than one binding site, suggesting that secondary (possibly allosteric) sites may be less common than hoped.

As to the fragments themselves, 21 bound in more than one pocket (not necessarily on the same protein), including the universal fragment 4-bromopyrazole. The fragments ranged in size between 6 and 20 non-hydrogen atoms, with 81% having 10 to 16, consistent with our poll last year. Given these sizes, it is perhaps not surprising that the vast majority of fragments conformed to the rule of three.

Roughly two thirds of the fragments were uncharged, while 22% contained a negative formal charge (usually a carboxylic acid) and 11% contained a positive formal charge such as an aliphatic amine. Interestingly, more than 90% of the fragment hits were achiral. Since our 2017 poll found that most fragment libraries contain chiral compounds, these results might suggest lower hit rates for these compounds. Fragment hits also tended to have lower Fsp3 scores than those in a popular commercial library, which is consistent with the observation that “less shapely” fragments give higher hit rates.

Digging more deeply into the chemical structures themselves, nearly a third of the fragments contained a phenyl ring, while 6% contained a pyridine and 5% contained a pyrazole. Thiophenes, indoles, indazoles, piperidines, furans, and pyrrolidines were present in 2-3% of fragment hits. But there was also plenty of diversity: more than half of fragments contained a unique ring system.

So that’s what fragment hits look like. How do they bind? Deeply, for starters: about three quarters of the fragment hits buried more than 80% of their solvent-accessible surface area, and 21 fragments were completely engulfed within a protein.

Not surprisingly, more than 90% of complexes showed at least one polar interaction, such as a hydrogen bond or a coordination bond to a metal ion. Many complexes contained more than one, and one had seven! Interestingly, these polar interactions also tended to be buried. Nitrogen and oxygen atoms from the fragments were equally likely to form hydrogen bonds. Interactions with bound water molecules were considered for high-resolution (≤1.5 Å) structures, and nearly half of these contained a water molecule with at least two hydrogen bonds to the protein and one to a fragment.

Beyond these conventional sorts of polar interactions, there were also less traditional interactions such as arene-mediated contacts, which occurred in nearly half of cases. As we’ve noted, these are often under-appreciated but can be useful for improving affinity. The subject of halogen bonds came up recently, but these turned out to be quite rare, appearing in just 3% of cases. Sulfur-mediated contacts and carbon hydrogen bonds were more common, appearing in 11-12% of complexes, respectively.

All of this has important implications for fragment library design. As the researchers note, this set of 462 fragments could be used as the basis for a library, and laudably all the structures are provided in the supporting information. Generalizing beyond these specific molecules, roughly a quarter of the atoms in the fragment hits are polar (nitrogen or oxygen) and thus more likely to form classic hydrogen bonds. The researchers “strongly suggest” maintaining this ratio in designing new fragments.

The researchers also suggest presenting “a minimum set of individual polar pharmacophoric elements, as opposed to distributing several pharmacophores on a given fragment,” which is essentially the minimal pharmacophore strategy described here.

The one category of data I would have liked to see was affinity. Many binding measurements were probably not reported, and experimental error can be particularly confounding for weaker interactions, but even a subset of the data should allow some conclusions about the strength of various molecular interactions. Hopefully this will be the basis of a follow-up publication.

16 April 2019

Third fragment-based drug approved!

Last Friday the US FDA approved erdafitinib (Balversa) for certain bladder cancers with FGFR2 or FGFR3 mutations. Although the fragment-to-lead story has yet to be published, those of you who were fortunate enough to attend Fragments 2019 last month heard some of it from Harren Jhoti.

Congratulations to the folks at Astex and J&J for a new tool in the campaign against cancer!

And earlier in the pipeline, several more drugs have entered the clinic starting from fragments, taking the number above 45.