From noncovalent fragment to covalent KRASG12C inhibitor

Last week we highlighted Steve Fesik’s presentation at the Discovery on Target meeting in which he discussed the discovery of a covalent inhibitor of the oncology target KRAS^G12C. The paper describing this work, by Joachim Bröker, Alex Waterson, and collaborators at Boehringer Ingelheim and Vanderbilt University, has just appeared (open access) in J. Med. Chem.

 

A decade ago we described how the Fesik lab reported finding millimolar fragments that bind to the so-called switch I/II pocket on KRAS. In collaboration with researchers at Boehringer Ingelheim, these were optimized to sub-micromolar ligands that block nucleotide exchange (see here). However, these molecules hit all RAS isoforms and show only modest cell activity. In contrast, the approved drug sotorasib binds in a different pocket, called switch II, and forms a covalent bond with an oncogenic cysteine mutation, G12C.

 

To find molecules that would bind in the switch II pocket, the researchers first needed to block the switch I/II pocket, which seems to be a hot spot for fragment binding. They did so by introducing a cysteine mutation near the pocket and linking this via a disulfide to a small fragment. All this was done in the context of KRAS^G12V, a mutant that is more common in cancer than KRAS^G12C. The modified protein was then screened using two-dimensional protein-observed (HSQC) NMR against 13,000 fragments. This process identified 20 fragments that bind outside of the switch I/II pocket, including compound 1, which bound to the modified protein with mid-micromolar affinity.

 

 

A combination of SAR-by-catalog and synthesis suggested the importance of both the amino group and the nitrile, and these observations were confirmed by a crystal structure of compound 1 bound to the protein deep in the switch II pocket, as predicted from the NMR data. The crystal structure also revealed a vector to grow the molecule, leading to compound 12. This molecule had sufficiently high affinity to bind to KRAS^G12V without the introduction of the switch I/II blocking fragment. Further growing to compound 19 and addition of a phenyl group (compound 20b) led to low micromolar binders. Installation of an acrylamide warhead and further decoration led to BI-0474, which rapidly reacted with the mutant cysteine in KRAS^G12C. Interestingly, the initial fragment is carried through unchanged.

 

In addition to potent biochemical activity, BI-0474 showed low nanomolar cell activity. The “bioavailability was not yet optimized,” but intraperitoneal administration led to anti-turmor activity in mouse xenograft models. The paper also notes that “a more advanced orally available analogue from this series has recently entered phase I clinical trials.” As we noted earlier this year, this is BI 1823911.

 

It is worth contrasting this work with the discovery of sotorasib, which we discussed in 2020. Sotorasib traces its origins to covalent fragment screens, and an electrophile was maintained throughout the optimization process. In contrast, the new paper starts with a non-covalent fragment that was optimized before an electrophilic warhead was introduced. This is probably more typical of how covalent drugs are discovered, as exemplified last year for the BTK inhibitor TAK-020. However, it is not necessarily easy; Steve mentioned in his presentation earlier this month that achieving the optimal configuration of the warhead took some effort.

 

KRAS has become a poster child for the power of fragment-based approaches to deliver drugs against previously intractable targets. The fact that the new molecules have good non-covalent affinity broadens the range of ligandable oncogenic mutants beyond KRAS^G12C. Indeed, the researchers end by noting that their approach has led to “molecules that are highly attractive for further use in the discovery of inhibitors against other KRAS mutants.”

 

Let’s hope they – and others – succeed.

Twentieth Annual Discovery on Target Meeting

Cambridge Healthtech Institute held its annual Discovery on Target meeting in Boston last week. Although it was technically a hybrid event, about 90% of attendees were physically present, so it really felt like a return to normalcy. Still, the online option was useful: just as with the spring DDC meeting, at least one speaker tested positive for COVID-19 and had to give his presentation remote. Also, with up to eight concurrent tracks, the fact that sessions were recorded for future viewing reduces FOMO, though likely at a cost of spontaneity (see our poll). Panel discussions were in-person only and not recorded, allowing for more candid conversations.

 

Fragments made appearances throughout the event. In a keynote talk, Steve Fesik (Vanderbilt) described work on several targets, most notably KRAS. Long-time readers will recall how NMR screens a decade ago identified molecules that bind to what has become known as the switch I/II pocket. Heroic efforts in collaboration with Boehringer Ingelheim have led to a chemical probe, but the biology around this particular site is complicated.

 

Also, the switch I/II pocket seems to be a magnet for fragments: all 25 crystallographically-characterized fragments from an NMR screen bound here. To look for new sites, the researchers introduced a cysteine residue to covalently block the switch I/II pocket with a known fragment, and then ran an NMR screen to find noncovalent binders at other sites. This identified fragments binding at the switch II pocket used by sotorasib. Extensive optimization and addition of a covalent warhead to target the G12C mutation led to clinical-stage BI 1823911. Steve emphasized the importance of diverse vectors for fragment growing and linking, and not being seduced by potency alone.

 

According to Christopher Davies of Genentech, one of the reasons KRAS has been so hard to drug is that it is very dynamic; in particular, the switch I and switch II loops can adopt multiple conformations. To constrain the protein, the researchers generated antibodies against the G12C mutant covalently bound to a small molecule inhibitor. One of these CLAMPs (Conformational Locking Antibody for Molecular Probe discovery) could stabilize the “open” form of the switch II pocket, thereby improving the affinity of ligands for this pocket and increasing the hit rate from an SPR-based fragment screen. (This work was published in Nat. Biotech. earlier this year.)

 

KRAS is a small GTPase. Samy Meroueh (Indiana University) discussed screening electrophilic fragments against Rgl2, which activates RAL, another small GTPase. We recently wrote about some of this work, and he mentioned that future publications are on the way.

 

Continuing the theme of difficult targets, Brad Shotwell described various hit-finding approaches used at AbbVie against the “cytokinome,” including IL-36γ, TNFα, and two sites on IL-17. We covered some of their TNFα work last year, and the IL-17 work will be the subject of a future post. In line with observations on other proteins, fragment hit rates predicted target ligandability.

 

The protein-protein interaction between NRF2 and KEAP1 is also a challenging target, and David Norton (Astex) discussed how a fragment-inspired virtual screen of the GlaxoSmithKline library ultimately led to low nanomolar inhibitors distinct from an earlier series. He emphasized the importance of growing fragments deliberately rather than attempting dramatic changes.

 

The pocket on KEAP1 is difficult because it is highly polar, but Marianne Schimpl (AstraZeneca) faced the opposite problem with the lipophilic allosteric site on MAT2a (work we highlighted last year). She mentioned the role of synthetic tractability: one fragment hit with higher LLE and Fsp³ was deprioritized in favor of a less shapely molecule that was more readily derivatized.

 

I spent much of the conference in PROTACs and molecular glues talks. Despite my arguments in 2018, FBLD is still not prominent here, but hopefully this will change. For PROTACs especially, which consist of two separate binding elements and a linker, minimizing the overall size is important. Indeed, Yue Xiong (Cullgen) described finding E3 ligands with molecular weights less than 300 Da. Despite only having micromolar affinity, they could be used to make highly effective PROTACs.

 

A broad view of drug discovery was provided by plenary keynote speaker Anabella Villalobos, who described the multiple therapeutic modalities used at Biogen to tackle neuroscience diseases. She mentioned that there are around 15 FDA-approved oligonucleotide-based drugs, but that this did not happen overnight: the first was approved more than a quarter century ago. This long “induction period” reminds me of my post last year comparing the rise of therapeutic antibodies with FBDD.

 

There is plenty more of interest; for those of you who attended, what talks would you recommend watching? And mark your calendars for September 25-28 next year, when DoT returns to Boston!

Inter-ligand STD NMR: Better than ILOE?

Although our poll in 2019 suggested that crystallography has surpassed NMR in FBLD, not all proteins can be crystallized. Ligand-detected methods such as saturation transfer difference (STD) NMR can be particularly useful for quickly identifying individual fragment binders and getting some sense of how they bind. A new variation published (open access) in Pharmaceuticals by Jesus Angulo and collaborators at University of East Anglia and Universidad de Sevilla provides information on the relative binding modes of two ligands.

 

Long-time readers may remember the inter-ligand NOE method (ILOE) we wrote about in 2010, in which proximity of two ligands is assessed by measuring NOE signals between them. However, despite being described more than 20 years ago, the technique seems to be rarely used, with fewer than a dozen papers in Pubmed, perhaps because ILOE requires large amounts of both protein and NMR time.

 

The new method is called inter-ligand STD NMR (IL-STD NMR), and it was discovered serendipitously while studying the binding of the drug naproxen to bovine serum albumin (BSA). As Teddy discussed several years ago, STD NMR normally involves irradiating specific protons in a protein (for example, the hydrogen atoms on buried methyl groups) and then measuring the “transfer” of this magnetization to bound ligands. When the researchers instead irradiated protons on one end of naproxen, they found that while the STD effect fell along the length of the molecule as expected based on distance, the signal suddenly increased at the other end of the molecule. Naproxen is known to bind to three sites on BSA, and this increased signal was attributed to the proximity of two adjacently bound naproxen molecules.

 

Inspired by this observation, the researchers developed IL-STD NMR. The experiment requires two samples, with two NMR experiments on each. One sample contains the protein and ligand of interest, while the other sample also contains a “reporter ligand” with a known binding mode. For each sample, one NMR experiment is run with selective irradiation of protons on the protein, while the other is run using irradiation of the reporter ligand. Comparison of the spectra reveals which regions of the ligand of interest are near the reporter ligand. The researchers demonstrated that the method works using a model system they had previously studied, the cholera toxin subunit B (CTB), which binds two ligands at nearby sites.

 

Importantly, the time and amount of protein is considerably less than required for ILOE: in this case 2 hours of NMR time and 0.3 mg of protein compared with 88 hours (!) and 1.8 mg. Moreover, the experiment could be run on a 500 MHz NMR, which is a relatively common instrument.

 

IL-STD NMR does have limitations. The researchers note that it is important to avoid irradiating the ligand of interest while irradiating the reporter ligand. Also, the approach obviously only applies to proteins with two nearby pockets (or one larger pocket). Still, it does look interesting, and I could imagine it being used as part of a screening cascade to find candidate fragments for merging or linking. What do the NMR aficionados think?

Crystallographic covalent fragment screening – but why?

Once considered a time-intensive method justified only by the most-promising molecules, crystallography is fast becoming a routine technique for general fragment screening. Just last month we highlighted an example where crystallographic hits were successfully used in virtual screens. In a new J. Med. Chem. paper, Jeffrey St. Denis, Benjamin Cons, and colleagues at Astex describe a crystallographic screen with a covalent fragment library.

 

Driven in part by the rapid development of KRAS^G12C inhibitors such as sotorasib, covalent fragments are continuing to increase in popularity. To screen, you first need a library, and here the researchers assembled a set of 114 molecules, most of which were commercially available from Enamine. The majority (101) contained an acrylamide moiety, the most common warhead in approved covalent drugs. The library is rule of three compliant and fairly shapely, as assessed by deviation from planarity. Laudably, the structures of all 114 library members are disclosed in the supporting information.

 

In addition to physicochemical properties, the library was also characterized experimentally. Importantly, it seemed to be quite stable when stored frozen at -80 °C in deuterated DMSO for up to a year. Most (83%) of the compounds were soluble to at least 3 mM in aqueous buffer. And just over half of the molecules withstood a test for reactivity with 5 mM glutathione, the main nucleophile found in cells, assessed as >90% remaining after 30 minutes at 37 °C.

 

As a test case, the researchers screened the kinase ERK2, an oncology target for which Astex previously developed the non-covalent clinical compound ASTX029. Soaking ERK2 crystals in 50 mM of each fragment led to 29 bound ligands, a whopping 25% hit rate. Of these, 16 compounds bound to C166, a cysteine near the hinge region in the active site. Unsurprisingly, less reactive fragments made hydrogen bonding interactions with the protein, while the more reactive fragments did not and also had poor electron density, suggesting non-specific alkylation. As we noted in 2014, distinguishing specific from non-specific binding is a general challenge when screening irreversible fragments.

 

Among fragments making specific interactions, compound 9 was chosen for further study due to its low inherent reactivity and interesting structure. The molecule showed no inhibition of the protein after 1 hour at 1 mM, but scaffold hopping and fragment growing led to compound 22, with low micromolar inhibition after 1 hour. (A back of the envelope calculation suggests k_inact/K_i ~ 25 M^-1s^-1, though I wish the researchers had reported this.)

 

This is an interesting paper, and it is useful to get structural information up-front, but I can’t help thinking that it would be more efficient to use an alternative method for the initial screen. As the researchers acknowledge, many covalent fragment screens use intact-protein mass spectrometry. For example, in a paper we highlighted in 2019, nearly 1000 fragments were screened against ten different proteins. Moreover, as we noted last year, the ability of crystallography to find such very weak hits can be a distraction as much as a blessing.

 

The current paper promises to “report on the further developments and observations pertaining to our electrophilic library in due course.” I look forward to seeing more.

Metal-binding fragments vs glutaminyl cyclases

Metal-binding fragments have a long history in FBLD; the first mention on Practical Fragments was back in 2010. The idea is to use the strong interaction between a fragment and a protein-bound metal as an affinity anchor for further optimization. The latest example, by Jie-Young Song, Soosung Kang and collaborators at Korea Institute of Radiological & Medical Sciences, Ewha Womans University, and elsewhere was published in ACS Med. Chem. Lett.

 

Glutaminyl cyclases such as glutaminyl-peptide cyclotransferase (QC) and glutaminyl-peptide cyclotransferase-like protein (isoQC) convert N-terminal glutamine or glutamate residues on proteins to pyroglutamates. This modification tends to stabilize proteins, and it has been implicated in several diseases. In particular, modification of CD47 by isoQC seems to be important for the ability of cancer cells to evade the immune system.

 

QC and isoQC are closely related enzymes with a zinc-containing active site. Capitalizing on this, the researchers tested a library of 36 potential metal-binding fragments in a functional assay against QC. Most of the compounds tested were inactive, though 11 had IC₅₀ values less than 0.8 mM. A few of these, including compound ab, were used to generate a second library of just half a dozen larger fragments, and compound 9 turned out to quite potent.

 

The researchers recognized that compound 9 has two potential zinc-binding moieties, and docking suggested the newly added amino-thiadiazole was likely responsible for the increased activity. Structure-based design ultimately led to compound 22b, with low nanomolar activity against QC and isoQC. The molecule did not seem to be generally cytotoxic, but it did increase phagocytosis of cancer cells in vitro, consistent with an effect on the “don’t eat me” function of CD47.

 

Unfortunately, no information is provided on the selectivity of compound 22b against other zinc-dependent enzymes. Moreover, unlike an earlier example of starting with metallophilic fragments, no ADME data are provided. But whether or not this particular series advances, it is nice to see metallophilic fragments being explored.

FBLD meets DEL part two: let there be light

DNA-encoded libraries (DEL) are collections of peptides or small molecules attached to DNA tags. In a typical application, libraries are mixed with a protein of interest, non-binders are washed away and those that remain are identified by using PCR to amplify the DNA tags. Two years ago we highlighted an article in which previously identified fragments were merged with molecules identified from DEL. However, because fragments typically have low affinities, screening fragments directly by DEL would seem to be difficult. In a new open-access RSC Medicinal Chemistry paper, Rod Hubbard and collaborators at Vernalis and HitGen describe how to do so. (Rod presented some of this work in April at the CHI DDC meeting.)

 

To identify weak binders, the researchers turned to photoactivatable fragments that – in the presence of UV light – would bind irreversibly to a nearby protein. Specifically, they used the diazirine tag, which has proven useful in both cell-based screening as well as screens of isolated proteins. Here, the researchers generated two libraries of fragments bound to DNA, with each library member also containing a diazirine tag. The libraries were built using different chemistries and consisted of 15,804 and 23,905 members, small by DEL standards (which often range in the millions) but large by fragment standards.

 

The PAC-FragmentDEL libraries were incubated against two proteins: the kinase PAK4 and the bacterial enzyme 2-epimerase. Each protein was incubated with both libraries for one hour at room temperature and then treated with ultraviolet light for 10 minutes on ice. Next, the proteins were captured on an affinity resin and washed extensively under denaturing conditions to remove any non-covalently bound library members. Finally, the DNA was amplified by PCR and quantified; any library members that bind to the protein stand out over background.

 

Of course, there is plenty of opportunity for non-specific binding, so the researchers incorporated several controls, such as omitting the UV-crosslinking step or protein. Moreover, they repeated the experiment in the presence of known high-affinity binders and looked for fragments that were competed.

 

In the case of PAK4, the researchers identified 301 fragments that could be competed. Eleven of these were further examined (without the DNA tags), all of which demonstrated binding by ligand-observed NMR, and ten of them yielded crystal structures bound to the protein. The examples shown in the paper occupy the hinge-binding site, which the researchers acknowledge is a low bar for fragment screens.

 

The second target, 2-epimerase, has a more challenging active site, and indeed the hit rate was lower: just 21 competitive fragments were found. But all 9 of those selected for further testing confirmed by ligand-observed NMR, and 5 of them yielded crystal structures.

 

This paper demonstrates that DEL can be used to identify fragment hits with a fairly low false-positive rate. But do we need yet another fragment-finding method? The researchers point out that PAC-FragmentDEL is fast, with screening and sequencing analysis taking just a few weeks. This also means that fragment libraries can be much larger than for most techniques. Protein requirements are also modest, at around 250 pmol (12.5 mg for a 50 kD protein). They also note that – because of the DNA tag – less intrinsically soluble fragments can be screened, increasing chemical diversity, though one might counter that this could lead to problems down the road.

 

On the downside, it is not clear whether affinity information can be obtained from the primary screen. Also, the need for a competitive tool molecule could limit choice of targets, as some of the most interesting targets lack any chemical probes. Still, as the researchers note, the competitor could be a peptide or protein, and in a pinch the site of interest could be mutated.

 

In summary, this looks to be an interesting approach, and I look forward to seeing more applications.

Crystallography first, then virtual screening: application to PKA

Fragment-based screening is often funnel-shaped: a virtual screen might identify dozens or hundreds of potential hits that are tested in various assays, eventually leading to a few chosen for crystallography. But a paper we highlighted back in 2016 argued that many assays miss genuine hits, and crystallography should be moved to the front of the line. A paper just published in J. Med. Chem. by Serghei Glinca and collaborators at CrystalsFirst, BioSolveIT, Enamine, and elsewhere provides a proof of concept.

 

The researchers started with a set of 19 crystal structures of fragments bound to Protein Kinase A (PKA) from a campaign we wrote about in 2020. Four diverse fragments were chosen for further study. Importantly, the affinities of these fragments had not been measured; selection was based on the diversity of chemical structures and binding modes.

 

Next, the crystal structures of the four fragments were used as starting points for four virtual screens using 208,293 Enamine REAL Space fragments (see here for more on these). These were docked using BioSolveIT’s FlexX algorithm, and 50 from each of the four screens were then computationally grown. Just over half a million of these elaborated molecules were then docked, and after clustering, triaging, and visual selection, 106 were chosen for synthesis, of which 93 were delivered and 75 were soluble at 200 mM in DMSO.

 

The soluble fragments were tested in a functional assay, and 30 of these showed inhibition. Most were weak (double digit micromolar or higher) but fragment EN093 (derived from Frag2) was a low micromolar inhibitor. All of the initial fragments were very weak inhibitors, with at best millimolar activity.

The 75 soluble compounds were also tested in a thermal shift assay (each at 2.5 mM), revealing 29 hits, of which 19 were also active in the functional assay. These included EN093. Interestingly, only one of the initial fragments (not Frag2) showed any activity in the thermal shift assay.

 

To assess how well the docking performed, 13 of the most active compounds were tested in co-crystallization experiments, yielding 6 high-quality bound structures. These confirmed the virtual screens, with the rmsd for EN093 being 0.74 Å.

 

Impressively, the whole study, including compound synthesis and crystallographic validation, took just 9 weeks.

 

This “Crystal Structure First” is conceptually similar to the V-SYNTHES approach we discussed earlier this year, with the difference being that while V-SYNTHES is entirely virtual, Crystal Structure First starts with an actual structure. As the researchers state, “using crystallographically validated fragments and bound ligands for template-based docking can be thought of as introducing a ‘magnet’ to help find the needle in an ever-growing haystack in a more targeted way.”

 

This is a nice case study, and intuitively it makes sense to start with an experimentally determined structure. Indeed, the increasing number of publicly available fragment structures should be a boon for this approach. That said, it is interesting that most of the molecules made and tested are quite weak, and only two have ligand efficiencies equal to or greater than 0.3 kcal/mol per heavy atom. As we suggested earlier this year, crystallography may find ligands that are just too weak to be useful. Perhaps adding a functional screen before computational elaboration could lead to even more and better binders.

Growing fragments in silico with FastGrow

Growing fragments is probably the most common approach to improving affinity, and it is immeasurably faster to do this virtually than experimentally. But as anyone who has ever tried can attest, this is often easier said than done. In a new open-access J. Comput. Aided Mol. Des. paper, Matthias Rarey and collaborators at Universität Hamburg, Servier, and BioSolveIT describe a free tool to help.

 

The application is called FastGrow, and it can be accessed through this web server or the SeeSAR 3D software package. It relies on the “Ray Volume Matrix (RVM) shape descriptor,” which simplifies chemical fragments and protein binding pockets into three-dimensional shapes. This allows extremely rapid assessments of whether a given fragment can fit into a binding pocket. A scoring function called JAMDA assesses interactions beyond simple shapes, such as hydrogen bonds and hydrophobic contacts, and also allows fragments to shift slightly to optimize complementarity with the protein.

 

One nice feature of FastGrow is that users can input fragments into multiple binding sites with different amino acid conformations, allowing for protein flexibility. You can also specify an important interaction, such as a critical hydrogen-bond, that you prefer to maintain.

 

To validate the approach, the researchers turned to the database PDBbind and looked for examples in which two ligands with identical cores but different substituents bound to the same protein. They chopped off the substituents from the first ligand and used the resulting fragment as a starting point to try to grow the second ligand. Running 425 of these took just 3 and a half hours and successfully recapitulated the binding mode 71% of the time. This was higher than the popular program DOCK (version 6.9), which seemed to be a pleasant surprise. They attribute the difference to a higher clash tolerance for FastGrow in the initial stages.

 

For additional validation, the researchers turned to real-world examples of fragment-growing for the kinases DYRK1A/B, which we highlighted last year (here and here). Here too FastGrow outperformed DOCK and was also about five-fold faster when using JAMDA (and 600-times faster without JAMDA, though at some cost in performance).

 

FastGrow looks to be a valuable tool, and indeed the researchers note that it is currently in use at Servier. There is a lot more detail in the paper and supplementary materials, including the full code for the FastGrow web server and all the underlying data. It would be interesting to compare its performance to the V-SYNTHES approach we highlighted earlier this year.

 

If you have experience using FastGrow, please leave a comment!

Is phenotypic fragment screening worthwhile?

Fragment-based drug discovery is almost always target-based. Indeed, not until the development of powerful biophysical techniques such as protein-labeled NMR did FBLD really began in earnest. Phenotypic fragment screens against cells, tissues, or animals are uncommon. In an open-access Front. Pharmacol. paper, Chris Lipinski and Andrew Reaume (Melior Discovery) argue that they should be used more often.

 

The researchers analyzed all 184,139,678 compounds in the CAS registry with molecular weights between 100 and 999 Da. These were divided into 18 bins (100-149 Da, 150-199 Da, etc.) Next, they calculated the percentage of molecules within each bin with any biological data as evidenced by the “biological study” tag in SciFinder-n.

 

In terms of raw numbers, fragments are well-represented, with the 250-299 Da bin containing close to 40 million molecules. However, only about 4% of these had any biological data. Molecules with molecular weights between 300 and 549 were abundant and also had considerably more biological data – up to roughly 50% of compounds in the 500-549 Da bin. In other words, people don’t seem to be screening lower molecular weight compounds in biological assays as often as they are screening larger molecules.

 

The assumption may be that small fragments are not biologically active, but the researchers revisit a classic In the Pipeline post in which Derek Lowe lists 56 drugs with molecular weights equal to or less than that of aspirin (180 Da). Most of these are old drugs, with all but three first reported in the chemical literature before 1980.

 

The researchers suggest that more effort should go into exploring the biology of smaller molecules, particularly those for which some activity is already reported. They also draw an interesting distinction between two uses of the word pleiotropic. People often say that a drug has pleiotropic effects if it acts on multiple targets; a classic example is imatinib, which hits several kinases in addition to the target BCR-ABL. However, the term pleiotropic originates in genetics and initially referred to one gene having multiple effects. Thus, a drug that acts on a single protein can have multiple effects, as in the case of the PDE5 inhibitor sildenafil.

 

As an example of a pleiotropic fragment, the researchers discuss MLR-1023, a fragment-sized molecule first discovered in a phenotypic screen at Pfizer in the 1970s. The molecule has shown promise in disease models ranging from atherosclerosis to myeloproliferative neoplasms and was taken into the clinic by Melior in 2014 as an anti-diabetic agent. All of these varied effects seem to stem from the ability of the compound to act as an activator of Lyn kinase. With just 15 non-hydrogen atoms and a molecular weight of 202 Da MLR-1023 is comfortably within rule of three space. Despite its small size, the molecule is a potent activator of Lyn, with an EC₅₀ around 50 nM, giving it a ligand efficiency of 0.66 kcal/mol per heavy atom.

 

Is MLR-1023 an outlier or an example of an underexplored pool of pharmacological riches? My suspicion is the former. It is rare to find fragments with EC₅₀s < 1 µM, let alone < 100 nM. Moreover, I suspect that many proteins are so difficult to drug that a molecule will need to be well beyond fragment-space – and even rule-of-five space – to have an effect. The protein-protein interaction targeted by venetoclax (MW = 868 Da) immediately comes to mind.

 

That said, the idea that a large group of tiny molecules is underexploited is worth exploring. For some types of drugs perhaps we don’t need extreme potency: Mike Hann noted a decade ago that the EC₅₀ values of approved drugs average 20-200 nM and cautioned against an “addiction to potency.” And because fragments are likely to have low affinities towards most proteins, they may even be more specific than larger drugs. It will be fun to discover how much room there really is at the bottom.

Diverse function – not structure – in fragment libraries

Successful fragment-based lead discovery typically starts with a good library. But what is “good”? Given that most fragment libraries are small, diversity is generally prized. The idea is to cover as much chemical space as possible with the fewest molecules. When most chemists hear the word diversity they think of structural diversity; tetrahydrofuran looks quite different from pyridine, for example. Functionally though, both contain a hydrogen bond acceptor. In a paper recently published (open access) in J. Med. Chem., Charlotte Deane and collaborators at University of Oxford and Diamond Light Source argue that functional diversity is more important.

 

Frank von Delft and his XChem colleagues at the Diamond Light Source have been screening dozens of targets crystallographically, many of them using the DSI-poised library, designed to enable rapid elaboration of hits. (We described it here). For the present analysis, the researchers considered ten diverse proteins (maximum pairwise sequence identity of 27%) that had all been screened against 520 fragments. Of these, 225 bound to at least one target.

 

The researchers considered what types of interactions the bound fragments made with the protein at either the residue or atomic level. For example, a fragment might serve as a hydrogen bond acceptor to the hydroxyl group of a serine residue. These interaction fingerprints, or IFPs, were calculated and compared.

 

Interestingly, there was no correlation between fragments that made similar IFPs and their structural similarity. In other words, “structurally dissimilar compounds can exploit the same interactions.” Moreover, many different fragments made similar or identical interactions: “structurally diverse fragments can be described as functionally redundant.”

 

In fact, just 135 fragments could make all the interactions observed for the 225 fragments. Some made more novel interactions than others, with “promiscuous” fragments that bound to multiple targets tending to be more informative.

 

The top 100 of these 135 functionally diverse fragments tended to have molecular weights between 175 and 240 Da and 12 to 16 non-hydrogen atoms, putting them comfortably within rule of three space. Interestingly, fragments that never hit any target skewed smaller, with many having molecular weights less than 175 Da and fewer than 12 non-hydrogen atoms; this is slightly at odds with work from Astex which found many tiny fragment hits.

 

The researchers considered sub-libraries consisting of either these functionally diverse fragments, randomly selected fragments, or structurally diverse fragments. The number of interactions discovered was significantly higher for the functionally diverse sets of fragments than for the other sets.

 

On one level the findings are not surprising: the whole concept of bioisosterism relies on the fact that different functional groups can make the same interactions, meaning that structurally disparate fragments can be functionally redundant. This suggests that libraries could be optimized to capture more information with fewer molecules. How to do so prospectively is not entirely clear, but laudably the researchers have provided chemical structures for all the fragment hits in the Supporting Information. It may be worth adding some of the functionally diverse fragments to your library; perhaps some enterprising vendor will start selling the top 100 as a set.

Fragments vs human Adensoine 2a Receptor using SPR

Last week we highlighted the use of surface plasmon resonance (SPR) to find ligands against RNA. Although RNA is not a typical protein target, it is at least normally free in solution. Targets such as GPCRs are more technically challenging because they are bound within membranes. Challenging, but not impossible, as illustrated by this post from 2012. A new ACS Med. Chem. Lett. paper by Reid Olsen, Iva Navratilova, and colleagues at Exscientia, University of Dundee, and AstraZeneca provides the latest example.

 

Navratilova and colleagues previously described using SPR to screen the β2 adrenergic receptor. In the new paper, the researchers studied the human adenosine 2a receptor (hA_2AR), a “rheostat for energy homeostasis” that also plays a role in cancer immunotherapy. hA_2AR is one member of a small family of adenosine receptors, and the researchers expressed all four of them, each with a polyhistidine tag that could be captured in the SPR instrument using a nickel-NTA sensor chip. Other labs (such as Heptares) have used mutant, stabilized forms of GPCRs, but here the researchers used native proteins and stabilized them by crosslinking them to the surface of the chip. They confirmed that these GPCRs bound known ligands with similar affinities to those reported in the literature.

 

Next the researchers screened a library of 656 fragments, each at 50 µM, against hA_2AR. This led to 72 potential hits taken into dose-response experiments, of which 17 confirmed with affinities ranging from 1.1 to 410 µM. All the sensorgrams are shown, as are the structures of the fragment hits. These confirmed hits were also screened against A₁, A_2B, and A₃; most of the fragments bound to all the receptors, though two were selective for hA_2AR.

 

To assess where the fragments bind, the researchers added a known high-affinity ligand; ten of the fragments could be competed, while seven showed less or no competition, suggesting that they may bind to an allosteric site.

 

GPCRs biology is complicated, and just because a ligand binds does not mean it will have any effect on signaling. In cell experiments, none of the fragments behaved as agonists, but five fragments could act as antagonists of a known agonist. Another fragment seemed to increase the signal, suggesting it is an allosteric modulator. As the researchers conclude, “while SPR can screen fragment-like molecules that allow for extrapolation of extremely large and diverse chemical spaces, it cannot predict the biological activity of these binders."

 

Nonetheless, this paper provides a nice guide on how to use SPR, with its low protein requirements, to screen GPCRs. And the fragments disclosed could be interesting starting points for medicinal chemistry.

Fragments vs RNA with SPR: A guide

Fragment-based lead discovery on RNA has a long history: the first mention on Practical Fragments was in 2009. Most often, various NMR methods have been used (see this example from last year), though isothermal titration calorimetry (ITC) is also effective. However, both of these techniques generally require considerable amounts of RNA. In a recent Biochemistry paper, J. Winston Arney and Kevin Weeks describe using SPR, which could increase the speed and ease of screening RNA.

 

Non-specific binding is a significant problem in characterizing RNA ligands. RNA is negatively charged, and many ligands are positively charged, leading to non-specific interactions. In a typical SPR experiment, the target is bound to a surface and the analyte is allowed to flow over the immobilized target; binding causes a change in refractive index that can be detected. However, if the analyte interacts non-specifically with the target, this will also be detected. For high affinity ligands the non-specific interactions may be minimal at low concentrations, but for low-affinity ligands such as fragments, it can be difficult to differentiate specific from non-specific binding.

 

SPR experiments generally use a reference cell, in which the analyte is allowed to flow over the surface in the absence of target; this signal is then subtracted from the target channel. Arney and Weeks decided to use a reference cell containing mutant RNA not expected to bind to the ligand.

 

The researchers developed their approach using two different riboswitches, each with known high-nanomolar ligands. Immobilizing the riboswitches to the chip and flowing ligand led to non-specific binding at concentrations of 100 µM or so. However, when the reference cell contained a mutant riboswitch designed not to bind to the ligands, this non-specific binding could easily be subtracted, leading to simple single-site binding models.

 

Of course, creating a mutant RNA assumes you already know where your ligand binds, which is not true if you are looking for ligands to a new target. To increase the generality of their approach, the researchers used a different riboswitch or a completely arbitrary RNA for the reference. These also worked, though not quite as well as the targeted mutants.

 

Finally, the researchers tested a dozen RNA-ligand pairs that had previously been rigorously characterized. Importantly, these varied considerably in affinity, from 8 nM to 2 mM. Most of them were also fragment-sized, with molecular weights as low as 119 Da. The correlation between SPR dissociation constants and those reported in the literature was excellent.

 

The technique does have limitations. First, the RNA-bound surfaces do seem somewhat unstable over a period of days. Also, larger RNAs present technical challenges, though the researchers do state that they have been able to examine molecules as large as 300 nucleotides. Overall this looks like a nice approach for measuring RNA-ligand affinities.

Solving structures with selective labeling and NMR2

Protein-detected NMR first enabled fragment-based lead discovery way back in 1996, but improvements in crystallography have now allowed synchrotrons to surpass big magnets as preeminent tools to determine how fragments bind to proteins. One of the major challenges in NMR is assigning the chemical shift values of atoms in all the individual amino acid residues. A technique called NMR Molecular Replacement (NMR²) sidesteps the need for this tedious, time-consuming process. A refinement to this technique, making it more broadly applicable, has just been published (open-access) in Sci. Reports by Julien Orts (University of Vienna), Martin Scanlon (Monash University) and collaborators.

 

As we discussed previously, NMR² relies on intensive calculations using experimental intermolecular NOEs between a protein and a ligand to generate a model. Although the method does not require assignment of backbone or side chain chemical shifts, it does require high-quality spectra. For example, if the spectra of several amino acid residues overlap it is impossible to distinguish them (this applies to conventional NMR methods too). The researchers realized that one way to simplify the spectra is through selective labeling, in which the methyl groups of the amino acid residues alanine, isoleucine, leucine, valine, and threonine are isotopically labeled with ¹³C. Going one step further, the entire protein can be deuterated (rendering most of the protein invisible to NMR), while these methyl groups retain ordinary hydrogen atoms.

 

For the present study, the researchers focused on the protein EcDsbA, an antibacterial target we’ve written about previously. They selectively labeled methyl groups so that, in isoleucine, leucine, and valine, only one of the two methyl groups was labeled. That reduced the total number of protons to just 6% of the unlabeled protein.

 

The researchers then solved the structure of EcDsbA with a previously identified ligand. At 23 heavy atoms the ligand is on the large side, though with an affinity of just 0.9 mM it presents a difficult test case. A total of twelve intermolecular NOEs were used in NMR² to build a model of the complex. One challenge with NMR² is that there may not be a single solution. For example, if two methionine methyl groups are both near a ligand, it may be impossible to determine a unique binding mode. This turned out to be the case, and the top two structures had different positions for a carboxylic acid group and a phenyl in the ligand.

 

To benchmark NMR², the protein-ligand complex was also determined using conventional two-dimensional techniques (HADDOCK and CYANA, which made use of assigned chemical shifts) as well as X-ray crystallography. These all agreed with the NMR² model in placing a phenylpropyl moiety from the ligand in a hydrophobic groove, but they differed in the placement of the carboxylic acid and the other phenyl moiety: the top scoring NMR² model agreed with the crystal structure and the CYANA NMR structure but differed from the HADDOCK structure, which was similar to the second-best NMR² model. Before assuming that the crystallographic structure is correct, though, it is worth noting that the ligand makes crystal contacts with a neighboring protein, and the electron density around the ambiguous phenyl is weak.

 

This is a nice demonstration of the utility of NMR². It seems to provide similar information as classic NMR methods, but the time taken is “orders of magnitude” less. And selective labeling should make NMR² applicable to even larger proteins. I look forward to seeing more people use this strategy.