21 May 2018

Fragments vs PKC-ι: 7-azaindole strikes again

A common question in library design concerns novelty: should you populate your library with custom-made, hitherto unseen molecules, or just buy off-the shelf compounds? While the first strategy might make it easier to get patentable leads, the second approach is faster and has a long history of success. Indeed, simple 7-azaindole served as a starting fragment for three clinical compounds: vemurafenib, PLX3397, and AZD5363. A new paper in J. Med. Chem. by Alvin Hung and colleagues at A*STAR illustrates just how versatile this scaffold can be.

The researchers were interested in protein kinase C iota (PKC- ι), one of a family of 10 kinases that has been implicated in cancer. A high concentration screen of 1700 fragments yielded 15 hits with measurable IC50 values, three of which were substituted 7-azaindoles. Compound 1, which has the highest ligand efficiency, was chosen to pursue.

Initial SAR quickly revealed that the bromine could be replaced with larger substituents, and a combinatorial library led to more potent molecules, such as compound 25. This was docked into a previously reported crystal structure of PKC- ι, which suggested the possibility of adding a positively charged moiety to interact with a couple aspartic acid residues. This strategy was successfully accomplished in compound 36, with low micromolar activity.

Adding a methoxy substituent to force a twist in the molecule led to an additional increase in potency, and rigidifying the amine led to compound 39, with mid-nanomolar activity. This was profiled against 101 kinases and found to be reasonably selective, though it did hit some other PKCs. The molecule was also not very permeable, and perhaps for this reason did now show good cellular activity.

To further optimize the series the researchers turned to group efficiency analysis, which revealed that the central benzimidazole element was the least efficient portion of the molecule. Earlier SAR and modeling had suggested that the unsubstituted nitrogen was making an important hydrogen bond to the protein, but “moving” the other nitrogen led to a more potent molecule. Further tweaking led to low nanomolar compound 49, which also had improved cellular activity.

Overall this is a nice example of advancing a generic, promiscuous fragment to a novel, potent, and selective lead – all without crystallographic support. Though further characterization of these molecules is not reported, the authors do mention optimization of a second series starting from a different fragment. Stay tuned!

14 May 2018

Fragments vs Gram-negative bacterial PPAT

Of the 30+ fragment-derived drugs that have entered the clinic, only one is an antibiotic. In part this reflects a shift away from this therapeutic area by many companies. Novartis, though, has continued to invest, as demonstrated by two consecutive papers in J. Med. Chem.

The researchers were interested in the enzyme phosphopantetheine adenylyltransferase (PPAT, or CoaD), which catalyzes the penultimate step of coenzyme A biosynthesis from ATP and 4'-phosphopantetheine. Although the enzyme is present in all organisms, the bacterial form is highly conserved across prokaryotes and significantly different than the human form. It is also essential for bacterial growth, thus making it an attractive target.

In the first paper, Robert Moreau and colleagues start big: a high-concentration screen (at 500 µM) of 25,000 fragments as well as NMR-based screens of their core 1408 fragment library. Triaging both hit sets led to a cornucopia of 39 crystal structures of bound fragments; the chemical structures of a dozen are provided in the paper, with IC50 values from 31 to >2500 µM. Perhaps surprisingly, all of these bound at the pantetheine binding site of the enzyme, suggesting that this is a “hotter” hot spot than the ATP-binding site.

Three of the fragments are described in more detail. The first was optimized from 273 µM to 4.3 µM, but subsequent advancement was unsuccessful. The second fragment, with an IC50 of 230 µM against E. coli PPAT, could be optimized to mid-nanomolar inhibitors; unfortunately these were much less active against PPAT from P. aeruginosa, so this series was also abandoned. But the third fragment discussed, compound 6, proved to be more tractable.

Initial optimization based on other hits led to compound 32, and addition of a methyl to the benzylic linker provided a satisfying 30-fold improvement in potency for compound 33. This “magic” methyl appeared to help desolvate the adjacent NH as well as pre-orient the molecule in the bound conformation. Further growing from this position led to compound 53, which provided a further 7-fold improvement in potency. Crystallography revealed a hydrogen bond between the nitrile nitrogen and a protein backbone amide. Unlike the previous series, this compound was active against PPAT from both E. coli and P. aeruginosa.

The second paper, by Colin Skepper and colleagues, describes further optimization of the molecules to picomolar binders. There’s a lot of lovely medicinal chemistry in both papers, but unfortunately all the molecules displayed at best only modest antibacterial activity. One problem is that Gram-negative bacteria have two membranes: an outer one which blocks lipophilic molecules and an inner one which blocks hydrophilic molecules. Compounds that can make it past these barriers also face an array of diverse efflux pumps, and these seemed to be the downfall of this project. The core of the molecule makes multiple hydrogen bonds to PPAT; about twenty different heterocycles were tested, but most of these had significantly lower potency, and the active ones were efflux pump substrates.

These difficulties in part explain why companies have been moving away from antibiotics. This was not a minor effort: each paper listed more than twenty authors. The second ends somewhat wistfully. “Although none of the series disclosed… yielded a clinical candidate, it is our hope that these studies will help pave the way toward the discovery of new Gram-negative antibacterial agents with novel modes of action.” It is a worthy – if arduous – quest.

07 May 2018

Fragment growing via virtual synthesis and screening

Practical Fragments has covered virtual screening for nearly ten years, and the tools continue to improve. More recently, researchers are using computational approaches not just to dock libraries of molecules, but to decide what compounds to make. The latest example, called AutoCouple, is described in an ACS Cent. Sci. paper by Cristina Nevado, Amedeo Caflisch, and colleagues at University of Zurich.

The researchers have a standing interest in bromodomains, epigenetic “reader” proteins that bind acetylated lysine residues. In particular, they were interested in CBP (cyclic-AMP response element binding protein). Previously the researchers had identified compound 1 through virtual screening, but although this compound had sub-micromolar affinity, it showed no cell-based activity, presumably due to the carboxylic acid, a moiety usually associated with poor cell permeability. Indeed, a CBP series we discussed earlier this year that also contained a carboxylic acid had no cellular activity.

To come up with better molecules the researchers used a program they developed and named AutoCouple because it virtually “grows” a fragment using common coupling reactions such as amide formation, Buchwald-Hartwig amination, and the Suzuki-Miyaura reaction. An initial set of 270,000 commercial compounds was computationally filtered to remove large molecules and those containing undesirable moieties. Potentially self-reactive building blocks were also removed. Ultimately 70,000 virtual compounds based on growing compound 2 (the key fragment of compound 1) were designed and docked into multiple crystal structures of CBP, and 53 were actually synthesized and tested.

Four of the 33 amides synthesized were sub-micromolar, compound 5 being one example; another 17 were low micromolar. (Five of the 10 Suzuki-derived compounds were also sub-micromolar, as was at least one of the amines.) Compound 5 was improved by using information from one of the other tested molecules to generate compound 16, with low nanomolar affinity. Crystallography confirmed that this compound binds as the docking had predicted, in a similar manner to compound 1.

Happily, not only was compound 16 more potent than compound 1, it was also active in cells. Moreover, it showed reasonable selectivity against a dozen other bromodomains.

Overall AutoCouple looks like it could be a useful tool to design and prioritize compounds for synthesis. Moreover, like the growing via merging “PINGUI” approach we highlighted earlier this year, the Python scripts appear to be freely available. It would be fun to benchmark both methods on the same targets to see how they compare.

30 April 2018

Fragment linking vs IMPDH (with a little help from the literature)

Mycobacterium tuberculosis (Mtb), the cause of its eponymous disease, is making an unwelcome comeback. Current treatments are lengthy, and highly resistant strains are emerging – and spreading. Chris Abell’s lab at the University of Cambridge has been working on multiple tuberculosis targets, and his latest results – in collaboration with Tom Blundell, David Ascher, and collaborators at the University of Cape Town and the University of Melbourne – has just published open access in J. Med. Chem.

The researchers were interested in inosine-5’-monophosphate dehydrogenase, or IMPDH, which is important for the synthesis of guanine nucleotides and is essential for every pathogen examined. They started by screening 960 fragments at 1 mM in a biochemical (spectrophotometric) assay. The IMPDH enzyme from a related organism was used due to its better expression and higher diffracting crystals; selected compounds were cross-checked with the Mtb enzyme and showed similar behavior.

This screen identified 18 molecules that gave at least 50% inhibition. IC50 values were determined for the six most active, and these ranged from 325-675 µM. These molecules were soaked into crystals of IMPDH, but only compound 2 produced a structure. As if to compensate, two molecules of compound 2 bound in the active site. Moreover, these two fragments bound in the same region as a previously reported molecule, compound 1.

Initial attempts at fragment growing yielded only modest improvements in potency, so the researchers tried to link the two copies of compound 2. One linking attempt failed outright, a second gave a 58 µM inhibitor, but a breakthrough came when the linker taken from compound 1 was used. The (S) enantiomer of compound 31 is 2500 times more active than the starting fragment, and crystallography revealed that it binds as designed. Unfortunately, and in contrast to compound 1, compound 31 showed no activity against Mtb in culture. The researchers hope to figure out why.

This paper illustrates several points. First, fragment linking can be quite effective. Second, consistent with our poll from a few years ago, this is not necessarily easy. Indeed, given the reliance on the structure of compound 1, this study can be considered an example of fragment-assisted drug discovery as much as fragment linking. And finally, as we’ve said before, biochemical potency all too often does not translate to cell-based activity – let alone good pharmacokinetic properties. Potency is just the first step in the long march to a drug.

23 April 2018

Fragments in the clinic: eFT508

Multiple clinical candidates derived from fragments were described at the recent CHI FBDD Meeting. The story behind one of these has just appeared in J. Med. Chem. in a paper published by Siegfried Reich and colleagues at eFFECTOR Therapeutics.

The researchers were interested in mitogen-activated protein kinase interacting kinases 1 and 2 (MNK1/2), which appear to be important in tumorigenesis but not normal cells. As Paul Sprengeler described it, they started with a “library” of just six fragments – four from the literature and two designed. (The company began in a law office, so hands-on experiments were initially limited.) Some might ask whether this constitutes FBDD, but in the end it’s not the size of your library that counts, but what you do with it.

All the fragments had good affinity, and the researchers were able to obtain crystal structures of four of them bound to MNK2. Optimization proceeded on all six of the fragments, but compound 1 was considered particularly attractive due to its high ligand efficiency and multiple vectors for growing.

Initially the researchers deconstructed the bicyclic ring to compound 7, which led to a 10-fold loss in potency but reduced the molecular weight and lipophilicity. As they note, “loss of potency in exchange for improved physicochemical properties is an often overlooked yet powerful optimization strategy in medicinal chemistry.” Too often people focus on binding over drug-like properties, so it is refreshing to see smart tradeoffs explicitly acknowledged.

Next, the researchers cyclized the molecule to form a lactam and remove one hydrogen bond donor. This also improved the affinity (compound 10). Replacing the phenyl ring in compound 10 with a pyridone in compound 12 further reduced the lipophilicity and improved the selectivity due to non-covalent interactions between a non-conserved cysteine residue and the heterocyclic ring. More optimization led to eFT508.

In addition to low nanomolar potency against both MNK1 and MNK2 in biochemical and cell-based assays, eFT508 is metabolically stable, permeable, and orally bioavailable in mice, rats, dogs and monkeys. The molecule showed good activity in several mouse xenograft models, and tissue samples revealed reduced phosphorylation of the substrate protein eIF4E, as expected. Unlike the MTH1 story last week, in which a selective chemical probe devalidated the target, the results with eFT508 suggest that inhibiting MNK1 and MNK2 has merit, and the compound is currently in four clinical trials for both solid tumors and lymphoma.

Like the story last week, this program progressed rapidly: just 110 compounds, 30 crystal structures, and 1 year to the development candidate. This turned out to be somewhat lucky, as it took another two years to find an equivalently attractive backup candidate. It is also an excellent example of how a non-selective fragment (a purine, found in ATP!) can be turned into a selective molecule. And finally, this is another nice example of how even a public, generic fragment can lead to an attractive chemical series. There are myriad published fragments bound to legions of targets, and it’s worth keeping these in mind whether or not you have in-house biophysical screening capabilities.

16 April 2018

Fragments vs MTH1: a chemical probe

As mentioned last week, CHI’s FBDD Meeting was chock-full of success stories. Some of these have recently been published, including work in J. Med. Chem. by Jenny Viklund (Sprint Biosciences) and collaborators at Bayer, the University of Oxford, and the Structural Genomics Consortium.

The researchers were interested in the protein MutT Homologue 1 (MTH1), which helps clear the cell of oxidized nucleotide triphosphates. The enzyme is upregulated in several cancers, and previous research involving non-selective MTH1 inhibitors had implicated it in cancer cell survival. But other research suggested that the effects on cancer cells were due to off-target effects. Clearly what was needed was a high-quality chemical probe.

The researchers started with a thermal shift assay of just 723 fragments screened at 1 mM, of which 166 increased the melting temperature by at least 1°C – a remarkably high hit rate suggesting good ligandability. Of the 49 fragments tested in full dose response thermal shift assays, 48 showed dose dependence. Compound 1 was not the most potent or ligand efficient, but it was synthetically tractable and different from other reported MTH1 inhibitors. Isothermal titration calorimetry revealed a dissociation constant of 49.5 µM, and the compound was also active in an enzymatic assay.

A crystal structure of compound 1 bound to MTH1 guided the selection of similar molecules from an in-house collection, such as compound 3. The structure also revealed a small pocket near the 2-position of the azaindole ring, and compound 5 – also available from the in-house collection – gave a nice pop in potency. Synthesis of a few analogs quickly led to compound 7, with mid-nanomolar activity. Crystallography revealed that the molecule bound mostly as expected. But because an asparagine side chain shifted to accommodate it, standard rigid-protein computational techniques would likely not have predicted its binding.

Further optimization for both potency and DMPK properties ultimately led to BAY-707, which is orally bioavailable in mice. In the interest of space I won’t go into details, but the paper is worth reading for a lovely, well-written account of lead optimization. Astute readers will recognize that all these molecules contain a 7-azaindole core, which is the same moiety that led to three clinical kinase inhibitors. The researchers tested representative molecules against a large panel of kinases as well as other ATPases and determined that the series is quite selective.

With probe in hand, the researchers set off to test whether inhibiting MTH1 would be useful for treating cancer. Unfortunately, as reported in another paper, the results actually “devalidate” the target. Despite potently inhibiting enzymatic activity in cells, BAY-707 showed no growth inhibition on several cancer cell lines, nor did it show activity in mouse xenograft models. While certainly disappointing, the results with this selective inhibitor at least provide a better understanding of biology.

This is also an example of just how quickly FBLD can yield results: at the CHI meeting Jenny said that it took 3.5 FTEs just 14 months from the start of synthesis to discover BAY-707, and the paper says this required only 35 compounds. A nice counterexample the next time someone says fragment approaches take too long.

09 April 2018

Thirteenth Annual Fragment-based Drug Discovery Meeting

CHI’s Drug Discovery Chemistry (DDC) meeting was held last week in San Diego. The event continues to grow, and this year hosted some 800 attendees, three quarters from the US and two thirds from biotech or pharma. The first DDC meeting in 2006 had just four tracks, of which FBDD is the only one that remains. The current event had nine tracks and three one-day symposia. There was always something interesting happening, and usually several – at one point three talks involving fragments were going simultaneously. Like last year, I’ll just try to give a few impressions.

What struck me most was the number of success stories, several involving clinical compounds. Last year we highlighted Pfizer’s discovery of a chemical probe against ketohexokinase (KHK); Kim Huard described how this was optimized to PF-06835919, the first and only KHK inhibitor to enter the clinic, which is now in phase 2 trials for NAFLD.

Another phase 2 compound was described by Paul Sprengeler (eFFECTOR Therapeutics). A handful of fragments designed from published work were characterized crystallographically bound to the kinase MNK1, and careful structure-based design resulted in eFT508, an MNK1/2 inhibitor which is being tested against various cancers.

A few years back we highlighted Genentech’s work on the kinase ERK2. In a lovely example of fragment-assisted drug discovery, Huifen Chen told “the convoluted journey of an ERK2 fragment series (with an HTS detour)”. SAR from the fragment series was used to inform the optimization of an HTS series originating from partner Array BioPharma, and was particularly useful for fixing some pharmacokinetic liabilities. Huifen emphasized the importance of using information from multiple strategies, ultimately leading to GDC-0994, which entered phase 1 trials for cancer.

Rounding up the list of clinical compounds, I heard through the grapevine that AbbVie’s dasabuvir, approved for hepatitis C, had fragments in its ancestry. I’d be interested to know more; though since success usually has many fathers, precise parentage can be tricky to ascertain.

Earlier stage success stories included the discovery of BI-9321, a highly selective inhibitor of NSD3-PWWP-1, which binds to methylated lysine residues in proteins. Jark Böttcher described how a collaboration between Boehringer Ingelheim and the Structural Genomics Consortium started with NMR and DSF-based screens of 1899 fragments to identify the cell-active chemical probe.

Jenny Viklund (Sprint Bioscience) described the discovery of potent, selective inhibitors of MTH1, a potential anti-cancer target. The project was successful, but unfortunately the molecules did not have the desired effect in cancer cell lines; this and other evidence helped to devalidate the target. Although undoubtedly disappointing, knowing what not to pursue is still important, and who knows – perhaps the target will turn out to be important in the future.

Finally, Steve Fesik (Vanderbilt) described a number of success stories against the KRAS protein, one of the holy grails of oncology. He also described how a fragment screen against a similarly hot target, the transcription factor MYC, failed utterly – the numerous compounds reported in the literature turned out to be artifacts or DNA intercalators. However, colleague Bill Tansey found that MYC interacts with the protein WDR5, and this protein-protein interaction turned out to be tractable, ultimately yielding potent inhibitors. This is a useful reminder that even if your target is not directly ligandable, biology is complicated enough that you may be able to modulate it through one of its partners.

Success sometimes requires breaking rules, as illustrated by the rule-of-5-defying drug venetoclax. Indeed, as noted by AbbVie’s Phil Cox, 18 of the 76 oral drugs approved since 2014 are bRo5s (beyond rule of 5). But if you’re going to break rules you should expect a harder path, and Phil described factors that correlate with success. Pete Kenny will be delighted to know that this has resulted in a new metric, AB-MPS, which is defined as the sum of the number of rotatable bonds, aromatic rings, and the difference of the ClogD from 3; values less than 12 are correlated with a higher probability of being orally bioavailable among AbbVie’s bRo5s.

Former guest blogger Brian Stockman described NMR-based functional screens he is doing with undergraduates at Adelphi University. Library acquisition can be challenging for a small organization, but happily Dean Brown at AstraZeneca has established an Open Innovation program for neglected diseases – if you’re interested and eligible you can receive a high-quality 1963-fragment library plated and ready for screening.

Of course there was plenty to learn about fragment-finding methods too, both in talks and in a discussion session led by Rod Hubbard (University of York and Vernalis). Microscale thermophoresis (MST) continues to be controversial, with researchers from a couple companies commenting that it’s fantastic the 20% of the time it works, while another company had success rates of ~95%. Thermal shift assays were also contentious, though Fredrik Edfeldt’s (AstraZeneca) method of adding urea or D2O (see here) to improve the sensitivity created significant buzz.

Cryo-electron microscopy continues to make rapid strides for structurally characterizing difficult targets, such as membrane proteins. Christopher Arthur (Genentech) did not downplay the many technical hurdles, particularly in sample preparation, but he thought that 2 Å resolution structures would be routine within the next decade. Although they have yet to analyze fragment binding, this is only a matter of time.

Ben Cravatt (Scripps) discussed ligand discovery on a proteome-wide scale using electrophilic fragments. His group has currently discovered more than 2000 ligandable cysteine residues in human cells – an exciting if daunting number of potential new targets.

And in the category of now for something completely different, Josh Wand (University of Pennsylvania) described nanoscale encapsulation – in which individual proteins are confined in reverse micelles suspended in liquid pentane; the low viscosity increases tumbling time and thus resolution for NMR, while the miniscule volume increases the concentration of protein and any accompanying fragments. This allows detection of extraordinarily weak interactions (dissociation constants of several hundred millimolar or worse). The technique is limited to very polar fragments because less polar ones would diffuse into pentane, but it would be interesting to see if a fluorocarbon replacement for the hydrocarbon allowed a wider range of fragments to be tested.

I could keep writing but I’ll stop here, hopefully before you stop reading; please leave comments. There are still several good events coming up this year, and mark your calendar for next year, when DDC returns to San Diego April 8-12, 2019!

01 April 2018

Universal crystallography

More mature readers may remember a column by Daedalus, aka David E. H. Jones, which used to run in Nature. Sadly he passed away last year, but his company, DREADCO, is still going strong. They have just launched a new product that should be of wide interest.

Our poll last year found that nearly a third of respondents would not begin fragment optimization without a crystal structure. Although there are successful counterexamples, it is fair to say that just about everyone would like a crystal structure if possible. Thus DREADCO has launched UniC, their Universal Crystallography platform.

The idea is based on previous work in which “crystalline sponges” can be used to absorb small molecules. X-ray data are collected on the sponge-molecule complex, and since the sponge structure is already known, the small molecule structure can be readily determined (see here for a nice summary by Derek Lowe). This is a powerful approach for small molecules, but the metal-organic frameworks used for the crystalline sponges are too small for proteins.

DREADCO researchers have solved this problem by using DNA origami to construct a cage-like structure that contains large pores yet is incredibly rigid, and therefore diffracts to high resolution. They have also inserted binding sites for a variety of DNA-binding proteins. All you need to do is generate a fusion between your protein and a DNA-binding protein and soak this into the crystallized DNA cages. Then soak in your fragment, and collect diffraction data to your heart’s content.

UniC is similar to the well-established method of tackling difficult-to-crystallize proteins by generating fusion proteins with antibodies or maltose-binding protein, but there you still need to find and optimize crystallization conditions for the construct. Here, since crystals of the DNA cage can be pre-grown, the time from construct generation to structure determination is dramatically shortened. Whatever the specifics of your protein of interest, all the world’s a cage.

A few years ago Teddy wrote, “The age of the medchemist is over; now is the time of the biophysicist.” Could the same be true for structural biologists who aren’t crystallographers? I hope Teddy puts in a good word for me when he reaches Valinor.

26 March 2018

Acceptable tradeoffs: From fragment hit to fragment lead against mGluR2, without structures

Membrane-bound proteins such as GPCRs are often ignored by practitioners of FBLD in part because – Heptares notwithstanding – they are usually difficult to characterize structurally. This seems like a missed opportunity. A large fraction of drugs target GPCRs, and the vast majority of these were developed without crystallographic information, so why is the fragment community so fixated on structure? A paper just published in J. Med. Chem. by György Szabó, György Keserű and colleagues at Gedeon Richter, the Hungarian Academy of Sciences, and Mitsubishisi Tanabe shows how much can be done without strcutures.

The researchers were interested in metabotropic glutamate receptor 2 (mGluR2), a popular target for schizophrenia. In particular, they sought positive allosteric modulators (PAMs), which act outside the main ligand binding site to enhance signaling. A functional screen yielded compound 4 as a fairly potent fragment-sized hit. Comparison with other larger reported inhibitors suggested growing could be productive, leading to molecules such as compound 5, with sub-micromolar activity. Further optimization for potency and ADME properties led to compound 29, with low nanomolar potency.

Unfortunately, this molecule is very lipophilic (cLogP > 5), resulting in poor solubility, high plasma protein binding, and thus limited efficacy in a mouse pharmacodynamic model. All attempts to reduce lipophilicity came at the cost of potency.

To determine which elements of compound 29 were most important for binding, the researchers turned to group efficiency analyses; that is, they systematically removed different chemical groups and weighed the loss in binding energy versus the reduction in size. Even though they could not visualize precisely how each group interacted with mGluR2, the researchers could measure it. This effort revealed that the biaryl moiety was not particularly efficient, and although trimming it came at a cost in potency, this was compensated for by improved ligand efficiency. Substitution at another position off the initial fragment led to a satisfying boost in activity (compound 30). Further optimization for pharmacokinetic properties led to the fragment-sized compound 60, which is considerably less potent in vitro than compound 29 but which has better brain penetration and also better efficacy in two mouse models.

Several lessons can be drawn from this story. First, as Mike Hann warned seven years ago, molecular properties should not be ignored in the push for potency. Indeed, despite the 25-fold decrease in potency for compound 60 compared with compound 29, the smaller molecule is more effective in vivo. This is reminiscent of the Merck verubecestat story, which also involved optimization of a fragment hit to a potent but lipophilic lead that was ultimately abandoned in favor of an initially less active but more ligand-efficient series. The second lesson is that in vitro models can only take you so far. And finally, creative chemists are able to advance fragments even in the absence of structural information. Hopefully more of them will give it a try.

19 March 2018

Industrializing native MS: hundreds of fragments against dozens of targets

Native mass spectrometry (MS) is a direct binding assay in which fragment binding to a target is detected when the complex is ionized and “weighed” in high vacuum. The technique is less commonly used than others, and there is some debate as to how well it works. A paper just published in ACS Infect. Dis. by Ronald Quinn and collaborators at Griffith University, the University of Washington, and the University of Toronto provides some encouraging data.

To demonstrate just how high-throughput native MS could be, the researchers started with 79 different proteins. These were all from Plasmodium falciparum, one of the main organisms that causes malaria. The proteins were chosen based on their size (< 50 kDa, for easier MS analysis) and likely importance for the parasite. Of these, 62 gave a good signal-to-noise ratio by native MS and were screened.

The researchers used an existing fragment library of 643 natural products; we highlighted an earlier version of this library in 2013. Of these, 602 molecules met the strict criteria defined in that design, with MW < 250 Da but with other properties more relaxed than rule of three guidelines. The library also contained significantly fewer aromatic rings than conventional fragment libraries and was more “three dimensional,” as assessed both by PMI and Fsp3.

Fragments were screened in pools of 8 at 5-400 µM each, with protein present at 1-20 µM; final ratios were 5:1 to 20:1. Hits were judged qualitatively as strong, medium, or weak, and the researchers estimate that strong and medium binders have dissociation constants < 100 µM.

Just over half of the proteins (32) had at least one hit, and a total of 96 fragments came up as hits. Importantly, many of these were selective: 48 fragments bound just one target, while another 18 bound just two (fragments that hit more than 6 proteins were considered promiscuous and excluded from further analysis).

Similarly to what has been done with NMR and thermal-shift assays, the researchers suggest that native MS can be used to assess ligandability. This is an appealing suggestion, though the researchers do not correlate MS-assessed ligandability with other methods such as SPR or high-throughput screens.

Conventionally, the next step would be to confirm binding with orthogonal techniques. Instead, the researchers took the rather bold move of testing fragment hits against the parasite directly. Remarkably, 79 of the fragments were active at 100 µM, with 13 having IC50 values < 45 µM.

A major strength of this paper is the disclosure of all the hits against all the targets. Not only does this allow others to confirm the results, it also provides starting points for further studies. So what do the fragments look like? Many of them are somewhat PAINful – we previously mentioned the promiscuity of one of their compounds, securinine. Although this molecule only hits two proteins in their panel, previous research has found that native MS can give high false-negative rates. Moreover, even if a molecule is truly inactive against a few dozen proteins, that doesn't mean it won’t hit many of the thousands of other proteins in a live protozoan.

Ultimately I would take any of these molecules with a huge dose of caution. That said, there are lots of interesting molecular structures in here, so if you’re looking to jump-start a program against malaria while exploring new chemistry, it may be worth digging into the data.

12 March 2018

Fragments vs PDE10A: Astellas’ turn

The 11 members of the phosphodiesterase (PDE) family cleave cyclic nucleotides such as cAMP and cGMP to regulate cell signaling. These enzymes are established drug targets – sildenefil inhibits PDE5, for example. PDE10A inhibitors have been heavily investigated for a variety of neurological disorders, and fragments have played a role in several efforts: we’ve highlighted work from Merck, AstraZeneca, and Zenobia/PARC on this target. A new paper in Chem. Pharm. Bull. by Ayaka Chino and colleagues describes work from Astellas.

A previous HTS screen at the company had led to a series of low nanomolar inhibitors, but these had metabolic liabilities and also inhibited CYP3A4. Thus, the researchers turned to fragments. No details are given as to library size, screening method, or hit rate, though it is worth noting that Astellas has previously reported fragment screening by crystallography. Compound 2 turned out to be a hit, and examination of the crystallographically determined binding mode proved quite useful. (Astute readers will also note the similarity of compound 2 to one of the Merck fragments.)

Because the chlorophenyl moiety was pointing towards solvent, the researchers decided to lop this off  to lower both lipophilicity and molecular weight. Previous publications had also revealed the presence of a “selectivity pocket”, and the researchers therefore grew towards this pocket, yielding molecules such as compound 7. Further tweaking led to compound 13, with low nanomolar potency. In contrast to the HTS-derived lead, this molecule was metabolically stable in vitro and showed negligible inhibition against a panel of 13 CYP enzymes.

This is a nice – albeit brief – example of how fragments can generate new chemical matter even against an extensively explored class of enzymes. Plenty of questions remain around pharmacokinetics, selectivity, and brain penetration, but the paper does end by promising that more will be revealed.

05 March 2018

Fragments deliver (another) inhibitor for CBP and EP300

In 2016 we highlighted a chemical probe that binds two closely related bromodomains, CBP (cyclic-AMP response element binding protein) and EP300 (adenoviral E1A binding protein of 300 kDa). These proteins bind to acetylated lysine residues in various nuclear receptors and are implicated in several types of cancer. Multiple chemical probes are always nice to have, and in a new paper in Eur. J. Med. Chem., Yong Xu and collaborators at Guangzhou Medical University, the University of Chinese Academy of Sciences, Jilin University, the University of Hong Kong, and the University of Auckland go some way towards this goal.

The researchers started with a virtual screen of 272,741 fragments (MW < 300 Da) docked against CBP. The top 5000 were clustered into related subsets and analyzed manually. Of thirteen fragments purchased and tested in an AlphaScreen assay, two had IC50 values better than 40 µM. Compound 6 was slightly less potent, but showed good selectivity against three other bromodomains.

The docking model of compound 6 suggested that more bulk between the indole and the carboxylic acid could be beneficial. Several molecules were made and tested, with compound 25e being the most potent. A related molecule was characterized crystallographically bound to CBP; this suppored the predicted binding mode.

Next, various small lipophilic elements were added to try to pick up additional interactions, ultimately leading to compound 32h, with low nanomolar affinity. This compound, which is equally active against EP300, also showed promising selectivity: it had no activity in a panel of six other bromodomains, including BRD9, which is inhibited by the chemical probe (CPI-637) mentioned above. Unfortunately compound 32h has no activity in cells, which the researchers speculate is due to the carboxylic acid. Masking this moiety with a tert-butyl ester causes a modest reduction in the biochemical activity but does lead to low micromolar activity in several cell assays.

Although much remains to be done, this is a nice example of advancing a computationally-derived fragment with limited structural information. I suspect we’ll see more of these, particularly for well-understood target families.

26 February 2018

Computationally-enabled fragment growing without a structure

Advancing fragments without high-resolution structural information remains a challenge scientists often choose not to take on, according to our poll last year. But for many appealing targets, such as membrane proteins, structural information is difficult to obtain. In a new paper in J. Med. Chem., Peter Kolb and collaborators at Philipps-University Marburg and Vrije Universiteit Brussel describe a computational strategy.

The approach, called “growing via merging”, starts with a core fragment that binds to a target, in this case the β2-adrenergic receptor (β2AR). Ideally this interaction is structurally characterized, but if not a model can suffice. Here, the researchers started with five fragments they had previously discovered. All of these had in common a lipophilic core with a primary or secondary amine appendage; this is a known pharmacophore for β2AR, so modeling could be used to orient the fragments.

Next, this core fragment is derivatized in silico with other fragments using a selection of 58 common reactions. Since all five fragments contained an amine, reductive amination was used here. A set of nearly 19,000 fragment-sized aldehydes and ketones was extracted from the ZINC database and computationally transformed into amines – as if they were reacted with one of the core fragments. These were then docked into the receptor, and those that did not overlap with the core fragments and also placed the amine near the amine of the core fragment were kept for further analysis.

The top 500-scoring fragments were then “reacted” – again in silico – with the core fragments and again docked. Eight of these were actually synthesized and tested for binding, of which four had higher affinity than the initial fragments. The best, compound 11, showed a 40-fold boost in affinity over its starting fragment.

This is an appealing approach, and it will be interesting to see how generalizable it proves. The β2AR is a somewhat forgiving test case due to prior work on the target and the fact that the ligand’s amine interaction with a critical asparate residue helps to orient the core fragment. Laudably though, the computational toolbox (called PINGUI, for Pyton in silico de novo growing utilities) is open access. Please leave a comment and share your experiences if you’ve tried it.

19 February 2018

More hits from a complex library?

One of the cornerstones underpinning fragment-based lead discovery is molecular complexity: fragments are less complex than larger molecules, and are thus likely to bind to more sites on more proteins. In theory, then, you want relatively simple fragments, and in fact Astex has actually formalized this with the concept of the “minimal pharmacophore”, in which each fragment contains a single pharmacophore (such as a hydrogen bond donor next to a hydrogen bond acceptor). But this is not the only way to build a fragment library; in 2016 we noted a paper out of the University of Dundee describing fragment libraries built with “caps” for easy derivatization. In a new paper in ChemMedChem, Paul Wyatt, Peter Ray, and collaborators at the University of Dundee and GlaxoSmithKline describe a screen with this “functional group complexity” (FGC) library.

The researchers were interested in the protein InhA, a drug target for Mycobacterium tuberculosis, the organism causing the eponymous disease. A relatively small library of 1360 fragments was assembled from six different sources, loosely defined by the authors:
  • 573 commercial fragments
  • 170 “3D” fragments from the 3DFrag consortium
  • 326 of the designed FGC fragments
  • 46 commercial fragments chosen based on known InhA inhibitors
  • 124 “inventory” fragments
  • 121 “project” fragments
These were screened against InhA in pools of 8, with each fragment present at 0.5 mM, using STD NMR, resulting in a fairly high hit rate of 11% (149 fragments). The commercial fragments and FGC fragments both gave a marginally higher hit rate (12.6%, 72 fragments and 13.2%, or 46 fragments respectively) while the 3D fragments gave a considerably lower hit rate (5.9%, or 10 fragments).

Previous work had suggested that more potent molecules seemed to reduce the STD signals for the NADH cofactor, so these molecules (32 fragments) were prioritized. The 13 FGC fragments represented a hit rate of 4%, nearly double the 2.4% for the library as a whole.

All 149 of the initial fragments were tested in a biochemical assay at 0.5 mM, but only 4 gave measurable inhibition – too few to draw conclusions. Five compounds were characterized crystallographically bound to InhA, including two of the FGC fragments. This information was used to merge two fragments, compound 24 (an FGC fragment) and compound 12 (a commercial fragment), yielding a mid-micromolar inhibitor. Adding a “magic methyl” gave a satisfactory ten-fold boost in potency. Fragment 24 was also merged with a previously reported molecule, compound 3a, to produce compound 42.

These results suggest that more heavily functionalized fragments don’t necessarily have a lower hit rate, albeit for a small library and a single target. And as we noted last year, molecular complexity is difficult to define; it is not immediately obvious that FGC fragment 24 is actually more complex than commercial fragment 12. The old cliché still holds: more data are needed.

12 February 2018

Fragments in the clinic: ABBV-075 / Mivebresib

Bromodomains bind to acetylated lysine residues in proteins to control gene transcription. These epigenetic regulators have received considerable attention as drug targets, particularly for oncology. Last year we highlighted work out of AbbVie in which fragments found in an NMR screen were advanced to two series of molecules that potently inhibit the four members of the BET family of bromodomains. A more recent publication in J. Med. Chem. by Keith McDaniel and his colleagues at the company describes how one of the fragments was transformed into the clinical compound ABBV-075, or mivebresib.

Compound 6 was not the most potent fragment identified, but crystallography confirmed that it binds in the acetyl lysine binding pocket. The earlier work described how the pyridazinone moiety was replaced with a pyridone and another phenyl ring was added to make molecules such as compound 9, with sub-micromolar activity.

Further modification of the pyridone led to compound 19, with a nearly 20-fold boost in affinity. Crystallography revealed that the pyrrolopyridone makes a bidentate interaction with a critical asparagine residue in BRD4, and also displaces a “high-energy” water molecule.

Next, the researchers sought to pick up additional interactions, and it turned out that introducing a nitrogen off the central ring was synthetically straightforward and would point substituents towards a pocket in the protein. This led to low nanomolar inhibitors such as compound 25, and crystallography revealed that one of the sulfonamide oxygen atoms makes a hydrogen bond with a backbone amide. Happily, the improvement in potency was also accompanied by an improvement in stability in liver microsome assays.

Unfortunately, although the pharmacokinetics in mice were reasonable, these compounds showed high clearance in rats. Analysis of the metabolites revealed that this was largely due to oxidation of the unsubstituted phenyl ring, so the researchers took the classic route of introducing halogen atoms to both deactivate the ring and block metabolism sites. This ultimately led to ABBV-075.

In addition to excellent potency in biochemical, biophysical, and cell-based assays, ABBV-075 showed excellent antitumor effects in a mouse xenograft assay when dosed orally at the low concentration of just 1 mg/kg. In addition to BRD4, the compound binds tightly to the other BET family members but is selective against most of the other bromodomains. It also demonstrates good pharmacokinetic properties in mice, rats, dogs, monkeys, and humans. ClinicalTrials.gov lists a Phase 1 study currently recruiting.

This is a lovely, textbook example of how structurally-enabled fragment growing combined with careful pharmacokinetic-based optimization can lead to a clinical candidate. Obviously there is a long and uncertain road ahead for the molecule prior to approval, but getting this far is a victory in itself.

05 February 2018

Pointless stereochemistry

Designing fragments to be more “three dimensional” than the flatter aromatic molecules that dominate most libraries is a topic often discussed in fragment library design. One way to make fragments more shapely is to introduce a stereocenter, but doing so often complicates the synthesis. In fact, new methods for efficient enantioselective synthesis constitute a major theme of organic chemistry research. In a recent paper in Angew. Chem. Int. Ed., Niklaas Buurma (Cardiff University), Andrew Leach (Liverpool John Moores University) and collaborators at Hawler Medical University Erbil and AstraZeneca demonstrate that the effort is sometimes not worthwhile.

Because proteins are chiral, different enantiomers can have profoundly different activities. The classic case is thalidomide, the racemic mixture of which was sold as a sedative in the 1950s, leading to the birth of thousands of babies with profound birth defects. Only one enantiomer appears to be responsible for the teratogenic effects, and many people are taught that had the manufacturer sold just one enantiomer, the disaster would have been averted. Unfortunately, biology is not so simple: the hydrogen atom attached to the chiral center is slightly acidic, and thalidomide rapidly racemizes at physiological pH.

Such racemization is more common than generally appreciated. The researchers experimentally measured the racemization of a couple dozen compounds using either circular dichroism (CD) spectroscopy or NMR (in the latter case, this involved dissolving the molecule in deuterated buffers and measuring the rate of deuterium incorporation, which occurs through an achiral intermediate).

The experimental results were then compared with those obtained through computational methods. Initially these were intensive quantum mechanical calculations, but the researchers also developed a rapid and effective approach by considering each of the attached substituents around the stereocenter independently. Importantly, the details for doing this are provided in the supporting information.

How much of a problem is this? The researchers provide four examples of what they call “potentially pointless stereoselective syntheses,” all published in high profile journals in 2016 (interestingly, three are fragment sized).

According to calculations, all of these molecules would undergo 19 to 70% racemization in 24 hours under physiological conditions.

So before embarking on any onerous stereoselective synthesis, it would be worth running a quick calculation. If the molecule goes forward you’ll still need experimental evidence for stability, but at least you’re less likely to be unpleasantly surprised by the answer.

28 January 2018

FragNet: The next generation

The first fragment event of 2018 was held in Barcelona last week. This was part of FragNet, established “to train a new generation of researchers in all aspects of FBLD.” Fifteen graduate students from 13 European countries are participating over the course of three years. This meeting marked the midway point for them. I was privileged to serve as a scientific advisor, and was impressed at how much they’ve been able to accomplish in just 16 months. They’ll be on the market next year, so you’ll definitely want to prioritize them if they apply to your institution.

One interesting feature of the program is that, in addition to their primary research, each student completes two “secondments” in other labs – one in academia and one in industry. This is unusual (in the US), and gives them a much broader range of experiences than is typical in graduate school.

The projects themselves are diverse, ranging from synthetic chemistry through computational approaches and biophysics. Fragment library design is a major theme: David Hamilton is building substituted cyclobutanes, Hanna Klein is focusing on pyrrolidines and piperidines, and Aaron Keely is exploring covalent fragments. Darius Vagrys, Sebastien Keiffer, Edward FitzGerald, Pierre Boronat, Lorena Zara, Eleni Makraki, Bas Lamoree, and Lena Muenzker are applying multiple (mostly) biophysical techniques against a variety of different targets. Andrea Scarpino, Moira Rachman, and Maciej Majewski are focusing on computational approaches. Finally, Angelo Romasanta is exploring the diffusion of FBLD techniques through industry. Often multiple students work on one problem from different angles: for example, Andrea is using modeling to explain some of the experimental results produced by Aaron. Plenty of interesting data are being generated in the projects, and I look forward to seeing the eventual publications.

In addition to the student presentations, there was a one day workshop open to the public, with a strong focus on computational approaches. Chris Murray (Astex) discussed how these play a role in all aspects of FBLD, from library design to finding related compounds using the Fragment Network (discussed here). Having a good set of validated experimental data is essential for benchmarking computational methods, and Astex has contributed one of these. But not every computational approach is applicable to every problem. Free energy perturbation (FEP), a rigorous method for predicting SAR, worked well retrospectively for the target XIAP but was not useful prospectively for a target in which the researchers were trying to find a less lipophilic replacement for a phenyl ring. Chris also pointed out that computational methods have a high hurdle – not just to make predictions but to do so better than experienced scientists.

Jenny Sandmark (AstraZeneca) discussed structure-guided design, with a heavy focus on crystallography. She emphasized the importance of quality control: resolution better than ~2.4 Å, with good electron density and low B factors. (Computation can help: Maciej gave an example where dynamic undocking was able to clarify an ambiguous crystal structure.) Jenny also highlighted a set of 52 crystal structures of fragments bound to the capacious binding site of soluble epoxide hydrolase that has been made publicly available for the benefit of modelers.

Chun-wa Chung (GlaxoSmithKline) discussed the importance of understanding your screening technologies and all their limitations. How to establish a cascade assay depends on the needs: if crystallography is challenging, you may want to limit the hits to those that confirm in multiple methods, as these are more likely to confirm crystallographically. If, on the other hand, you have the capacity to do lots of structures you should examine hits from all screens, as those that don’t repeat may be false negatives. Chun-wa also discussed the importance of biophysics for HTS (though this may require different protein constructs for different methods). An HTS screen of 1.7 million molecules against ATAD2 produced a 1% hit rate, of which 444 were studied using a variety of methods including fluorescence polarization, SPR, and NMR. Ultimately only 16 compounds turned out to be useful – all in a single series. (See here for their fragment efforts.)

John Overington (Medicines Discovery Catapult) gave an overview of the open-access database ChEMBL, which holds data from publications and patents on more than 11,000 targets and 14.5 million molecules, including 13,000 clinical candidates and 1500 drugs. Of course, the entries are only as good as the underlying publications: biochemical assays can vary by about 10-fold, cell-based assays can differ by about 100-fold, and in vivo results can vary by 1000-fold. Still, studying these data can produce interesting insights. For example, the observation that antibacterial compounds tend to be larger and more polar appears to be due to the fact that many antibiotics bind to bacterial RNA – those that just bind to bacterial proteins have more standard properties.

Finally, Anthony Bradley described the computational resources at XChem. We’ve recently discussed some of these, including their open-access version of Fragment Network for analog searching. XChem uses extremely high concentrations of fragments for soaking – DMSO stocks are 500 mM and are soaked at 30-50%, so the final concentration can be as high as 250 mM! This often results in multiple fragments binding to a crystal, many of which are of uncertain functional relevance; Anthony used the term “putosteric” for putative allosteric site. Achieving functional activity can be challenging, but it is encouraging that of 16 targets initiated in the past 12 months, 7 have produced compounds with IC50 values better than 100 µM.

All in all a great start to the year – and lots of good events ahead – hope to see you at some!

21 January 2018

Linking fragments on DNA

DNA-encoded chemical libraries are one of the sexier new approaches for lead discovery. Typically, small molecules are synthesized while covalently linked to DNA and then screened for binding to a target. The structure of the molecule is encoded in the sequence of the DNA, and since incredibly tiny amounts of DNA can be sequenced (wooly mammoth genome, anyone?) you can fit massive libraries into a single Eppendorf tube. Indeed, some companies boast 100-billion compound libraries, nearly three orders of magnitude more than the number of molecules indexed by Chemical Abstract Service.

One might think this has no relevance for fragments. Indeed, the only mention of DNA-encoded libraries I recall on Practical Fragments was a comment by Teddy back in 2012 that the approach is “as opposite from FBDD as you can go”. A recent paper by Dario Neri, Filippo Sladojevich, and their collaborators at the ETH Zürich and Philochem in ChemMedChem suggests otherwise.

The researchers have developed an approach called DNA-encoded self-assembling chemical (ESAC) libraries (see also their earlier paper in Nat. Chem.). Rather than synthesizing a single molecule on each strand of DNA, this approach involves assembling two separate sub-libraries of DNA-linked molecules, one attached to the 5’-end and the other attached to the 3’-end. These are then mixed together, allowed to hybridize in a combinatorial mixture, and screened against the target; if a specific combination of fragments is identified (through elegant PCR experiments), this indicates that the two fragments bind to the target in close proximity.

The researchers have focused on the protein alpha-1-acid glycoprotein (AGP), a prominent plasma protein whose function is poorly understood. In their Nat. Chem. paper, a library of 111,100 members (550 x 202 fragments) identified fragments A-117 and B-113. Neither of these fragments showed any binding themselves, but when linked together the resulting compound 1 bound with low micromolar affinity as assessed by isothermal titration calorimetry (ITC).

The linker connecting the two fragments is long, flexible, and not particularly drug-like; its improvement is the focus of the ChemMedChem paper. The researchers increased the size of their second fragment library from 202 to 428 elements, and an ESAC screen revealed that the pair of fragments A-117 and B-217 – both still attached to DNA – had a dissociation constant of 110 nM; B-217 itself (attached to DNA) was around 9900 nM.

To find out how these fragments could be productively linked, the researchers coupled them to 11 different scaffolds, each of which was attached to DNA. All of these bound to AGP, with dissociation constants ranging from 9.9 to 1300 nM. The moment of truth came when the researchers resynthesized some of the molecules no longer attached to DNA. Compound A117-L1-B217 bound with a Kd of 76 nM as assessed by SPR, while the weakest on-DNA binder (Kd = 1300 nM) showed no binding by itself. Although no explanation is provided for this discrepancy, it could be due to low solubility.

This is an interesting approach, though the molecules reported do tend towards molecular obesity (A117-L1-B217 weighs 765 Da and has a ClogP approaching 8). Indeed, this may be an inherent liability – the minimum allowable distance between two fragments that are each attached to DNA may be larger than desirable for most targets. Still, it will be fun to watch this develop.

15 January 2018

Fragments vs USP7, two ways, both allosteric

Proteins in cells are constantly synthesized and degraded in a complex, highly regulated manner managed in part by the ubiquitin proteasome system. Simplistically, a ubiquitous small protein called ubiquitin is conjugated to other proteins, targeting them for destruction, and some of the proteins thus targeted control the stability of still other proteins. But ubiquitination is not destiny: ubiquitin can be removed by more than 100 deubiquitinating enzymes, or DUBs.

As I said, this is complex. But complexity has never stopped folks from pursuing drug targets, and multiple groups are interested in a particular DUB called USP7, which is implicated in cancer and other indications. USP7 is one of more than 50 members of a subfamily of DUBs that use cysteine as a catalytic residue. Selectivity is an obvious challenge, and since cysteine is chemically reactive, any screening result carries a high risk of being an artifact. Two recent papers describe how fragment-based approaches led to potent, selective inhibitors.

The first, published in J. Med. Chem. by Paola Di Lello, Vicki Tsui, and coworkers at Genentech, started with an NMR fragment screen. This identified molecules such as compound 1, which NMR data suggested bound near the active-site cysteine. This and other fragments were used to conduct virtual screens of the much larger Genentech library, and 21 of these were then tested experimentally. Most of these either didn’t bind, bound to multiple sites, or caused protein aggregation, but four of them, including compound 2, showed clear binding to a specific site on USP7 and also inhibited the enzyme in a biochemical assay.

Surprisingly, protein-detected NMR suggested that these four molecules did not bind in the active site as expected but rather in an adjacent “palm site”, a hypothesis that was confirmed by a crystal structure of compound 2 bound to USP7. This led the researchers to reexamine other hits from the original NMR screen, where they identified several aminopyridinephenols, such as compound 13.

Meanwhile, a biochemical HTS against USP7 had identified 76 hits, but most of these turned out to be artifacts, and none of them yielded co-crystal structures with the enzyme. The fragment findings led the researchers to revisit some of the weaker hits that had been overlooked, such as compound 15. This led to a crystal structure showing binding in the palm site, and further medicinal chemistry ultimately led to molecules such as compound 28 (GNE-6640), with nanomolar activity in both biochemical and cell-based assays. A separate paper in Nature characterizes the biology in more detail, revealing that molecules in this series interfere with ubiquitin binding and are highly selective for USP7.

Another fragment effort on this target was reported by Timothy Harrison and collaborators at Almac and Queen’s University, Belfast in Nat. Chem. Biol. An SPR screen of 1946 fragments against the catalytic domain of USP7 led to compounds such as fragment B. This was combined with molecules from other groups that had been reported in the literature, leading to compound 1. Subsequent medicinal chemistry, informed by crystallography, led to compound 4, with low nanomolar biochemical and cell-based activity and excellent selectivity. The enantiomer is much less active, and compound 4 should be a useful chemical probe to further understand the biology of USP7.

Remarkably, not only do the two series of molecules bind some distance away from the active site cysteine (yellow, upper right), they bind in completely different, non-overlapping sites!

These papers illustrate the importance of allosteric sites for tackling specific members of large protein families. They are also both cases of “fragment-assisted drug discovery.” Unlike many success stories we’ve highlighted, it is difficult or impossible to find the initial fragment in the final molecules. Heck, Genentech’s best molecules bind in a completely different site from where the first fragment hits bound. Being open to such possibilities, and using all available data from every possible source, are keys to success.