(Not) getting misled by crystal structures part 6: low ligand occupancies

It’s been a while since our last “getting misled by crystal structures” post. That one described unrecognized conformational heterogeneity of ligands. A more basic issue is ligand occupancy. It’s normally assumed that every protein in the crystal lattice has a bound ligand. A new paper in Structure by Timothy Stachowski and Marcus Fischer at St. Jude Children’s Research Hospital reveals that this is not always the case.

 

The researchers selected roughly 10,000 protein-ligand structures from the Protein Data Bank (PDB) and did a simple re-refinement of the ligand occupancies and B-factors (measures of conformational heterogeneity and modeling errors). 10% of the structures already presumed ligand occupancies at or below 0.9, but re-refinement saw the fraction jump to 35%, more than three-fold higher. There were no overall differences between covalent and non-covalent ligands, but 37% of fragments (defined as having MW <300 Da) saw decreased occupancy in re-refinement compared to just 22% of larger ligands. A few structures even saw occupancy drop completely. The authors wrote that “manual reviewing these revealed that ligands were built into spurious electron-density.”

 

Crystallographers use several metrics to assess the quality of their structures. In addition to B-factors, unique to each residue or atom, real-space correlation coefficients (RSCC) and real-space R values are commonly used, and the researchers compared these parameters before and after re-refinement. In many cases the metrics improved with decreased occupancy, but not for all metrics, and not always meaningfully. This means that standard assessments do not always flag partially occupied ligands.

 

OK, so a ligand binds at only 80% occupancy rather than the 100% assumed: does this matter? The researchers describe three categories where the answer may be yes. In the first, correcting ligand occupancy reveals alternative conformations of protein side chains, which could be informative for understanding the mechanism of binding. In the second, correcting ligand occupancy can reveal water molecules that interact with the ligand and/or protein. As we noted more than a decade ago, water is an essential player in protein-ligand interactions, and a single water molecule can make the difference between a binder and a non-binder. Finally, correcting ligand occupancy can reveal alternative binding modes for the ligand, and even ligands binding to other sites.

 

Importantly, this analysis was done on individually refined structures in the PDB, and it seems likely that the issues would be even more severe for structures batch-refined in high-throughput crystallographic fragment screens. As we wrote last year, the community needs to figure out how to deal with the increasing number of these structures.

 

The fact that roughly a third of PDB structures have less than 100% ligand occupancy has implications for training AI models. It also has implications for individual targets. As the researchers note, “non-crystallographers who rely on the PDB often assume that deposition itself is an implicit stamp of approval. Structural biologists, however, know that this is not always the case.” Before using a structure, it would be wise to re-refine it yourself.

Surprise – a covalent histidine-targeting PDE3B inhibitor

Earlier this year I wrote about archiving crystallographic fragment data, and indeed a meeting is planned for early next year to establish guidelines. A new paper in J. Med. Chem. by Samuel Eaton and David Christianson at University of Pennsylvania illustrates why this is important.

 

The story starts with a paper published in 2024, also in J. Med. Chem., by Ann Rowley, Gang Yao, and collaborators at GSK and 23andMe. They were interested in finding inhibitors of PDE3B, a cyclic nucleotide phosphodiesterase that has been implicated in metabolic disease. However, this enzyme has a closely related counterpart significantly expressed in cardiac tissue: PDE3A, with 95% amino acid identity near the active site. So the researchers sought an inhibitor highly selective for PDE3B over PDE3A.

 

A DNA-encoded library (DEL) screen of 1.9 trillion(!) molecules was screened against both PDE3B and PDE3A. Hits were resynthesized without the DNA and tested in activity assays, leading to several chemical series, only one of which was selective for PDE3B. A key feature of this series was a boronic acid moiety, which was essential for activity. Optimization led to compounds such as GSK4394835A, with high nanomolar activity against PDE3B and >20-fold selectivity against PDE3A. The GSK researchers deposited a crystal structure of this molecule in the protein data bank (PDB), along with the structure factor amplitudes. It showed the boronic acid making non-covalent interactions with side-chain residues as well as the catalytic magnesium atoms and water molecules.

 

Further optimization at GSK led to compounds with as much as 300-fold selectivity for PDE3B, but like GSK4394835A, these were only high nanomolar inhibitors. The researchers could further improve potency, but this came at the expense of selectivity. Cell activity was modest at best, and the researchers noted that “the boronic acid is, in general, a challenge for development of an orally bioavailable drug.”

 

This is where the University of Pennsylvania researchers take up the story. As their paper points out, several drugs do contain boron, most notably bortezomib, which forms a covalent adduct with a threonine in the proteasome. When Eaton and Christianson took a closer look at the PDB entry showing GSK4394835A bound to PDE3B, they “noticed unusual features such as extra density around the boron atom of GSK4394835A, steric clashes between the boronic acid moiety and H737, and aberrant refinement statistics… from ideal bond lengths.” Upon re-refinement, they found that the boronic acid in fact makes a covalent bond with histidine 737. The structure explains why the boronic acid moiety was essential for activity, and the new paper suggest that other covalent warheads could potentially be used in place of the boronic acid. (Eaton and Christianson write that they contacted the GSK researchers in February of 2024, but it is not clear whether they heard back.)

 

This is a nice correction of the literature and a reminder not to take crystal structures at face value. The beauty of the PDB is that, with the experimental data deposited, the new researchers were enabled to re-refine the data even without input from the original authors.

 

As we’ve previously discussed, this example is not the only misleading crystal structure in the PDB. Many fragment structures have lower occupancy and more ambiguous electron density and would be even more prone to misinterpretation. As the community moves to establish guidelines for depositing fragment structures, it will be important to provide access to the raw data to facilitate this type of reanalysis.

Checking halogen bonds

Halogen bonds (or X-bonds) are one of the less appreciated protein-ligand interactions. As we discussed in 2022, the polarized nature of a carbon-halogen bond creates a partially-positively charged “σ-hold” at the bit of the halogen furthest from the carbon, and this can make favorable interactions with lone pairs on oxygen or sulfur atoms (or nitrogen, but in most proteins this is limited to histidine residues and is rare.) Halogens can also interact with aromatic π-systems such as the side chains of phenylalanine, tryptophan, histidine, and tryptophan. Since many fragments contain halogen atoms by design, halogen bonds may occur frequently. But how do you decide whether “a halogen in proximity of a possible acceptor” actually contributes to binding? In a new (open-access) paper in Protein Science, Ida de Vries, Robbie Joosten, and colleagues at Oncode Institute and The Netherlands Cancer Institute provide a new metric.

 

The researchers examined structures of halogen-containing ligands bound to proteins in PDB-REDO, a database of carefully vetted and refined structures from the Protein Data Bank. They only included structures solved to better than 2.5 Å resolution and omitted structures where halogens had high B-factors, which may be the result of radiation damage. This led to 8423 structures in which a halogen possibly interacted with an oxygen or sulfur atom and 8096 potential halogen-π interactions, which were analyzed in detail.

 

A halogen bond to an oxygen or sulfur atom can be described by the interatom distance and two angles: θ1 (carbon-halogen-oxygen/sulfur) and θ2 (halogen-oxygen/sulfur-carbon). Halogen-π-system bonds can be defined by distance to the centroid of the π-system and θ1, the carbon-halogen-centroid angle. (The paper has a nice diagram.) These parameters were calculated and annotated for all the structures.

 

Median distances were 3.5 Å between halogen and oxygen/sulfur, regardless of the halogen. Median θ1 angles were smaller than the 150º-180º expected, particularly for fluorine atoms, while median θ2 angles were more consistent with theory, at 90º-120º.

 

For halogen-π-systems, median distances were 4.8 Å for all halogens except iodine, which came in slightly higher. But θ1 angles were still smaller than expected, mostly between 110º-140º.

 

Armed with this tranche of high-quality data, the researchers established a Halogen Bond Score, or HalBS. For any potential halogen bond in a new crystal structure or other structural model, the distance, θ1, and, if applicable, the θ2 values are calculated, and if any of these diverge too far from the median values, HalBS flags them. Importantly, the researchers acknowledge that “the current HalBS cannot be used as a direct validation metric but can provide an indication of genuine halogen bonds and ‘not so proper’ halogen bonds.”

 

With this caveat HalBS could be useful, and the researchers have made the source code available at https://github.com/PDB-REDO/HalBS (though the link doesn’t seem to work for me). As they note, more data, such as might be provided by widespread deposition of large crystallographic fragment screens, could further refine HalBS. Of course, the existence of a halogen bond exists says little about how much binding energy it contributes, but it’s a start.

How can we house our crystallographic data?

Three years ago we highlighted a growing debate about how and where to house crystallographic fragment data. With the recent surge in high-throughput crystallography, issues including access, accuracy, and capacity have only become more urgent. An open-access perspective in Nat. Comm. by Manfred Weiss (Helmholtz-Zentrum Berlin) and multiple coauthors, including yours truly, calls on the scientific community to make some difficult decisions. Indeed, a session at the 75^th annual meeting of the American Crystallographic Association going on today is devoted to the topic.

 

High-throughput crystallography can involve soaking more than 1000 crystals with fragments, sometimes yielding hundreds of protein-ligand structures. The paper tabulates a dozen synchrotrons around the world with current or planned high-throughput capabilities. We’ve written recently about the XChem facility at the Diamond Light Source, which is currently running about 80 fragment screens per year. Assuming similar productivity at the other synchrotrons, we might soon see 1000 fragment campaigns per year worldwide. If each of these involves 1000 crystals and we get 10% hit rates, that could mean 100,000 new fragment structures annually.

 

That is a big number. For reference, 10,000 new crystal structures are currently being released by the protein data bank (PDB) each year. (Director of the RCSB PDB Stephen Burley is one of the authors of the perspective.)

 

The problem is that, as we discussed in the 2022 blog post, most fragment structures from high-throughput screens are not refined to the level required for the PDB, a process which typically takes a day or two for the researcher and up to 3 hours by a biocurator at the PDB. Moreover, fragments are often identified using PanDDA (Pan-Dataset Density Analysis, which we wrote about here), a process which makes use of the many unbound structures obtained in a dataset. Ideally, these datasets should also be made available.

 

The challenge is balancing practicality with FAIR (Findable, Accessible, Interoperable, and Reusable) principles. The paper outlines four non-exclusive options. Very briefly, these are:

 

Option One: Fully refine and deposit all protein-fragment structures just as with other structures.

 

Option Two: Partially refine structures, and possibly flag or even segregate them from other structures in the PDB.

 

Option Three: Rather than treating each protein-ligand structure independently, treat each high-throughput screen as a single experiment, and archive all of the data in its entirety, including unbound structures. These data could be housed in the PDB or elsewhere.

 

Option Four: A hybrid approach, where fully refined structures would be deposited in the PDB and the rest of the data would be stored in a separate branch of the PDB or elsewhere entirely.

 

There are pros and cons for each option. At the extremes, the first option puts a tremendous burden on experimentalists and the PDB, and potentially valuable information regarding unbound structures is lost, while option three requires setting up new repositories to store vast quantities of data.

 

The paper intentionally avoids making a specific recommendation and instead calls for discussion within the scientific community. Personally, I favor some sort of hybrid approach such as option four. As the paper notes, no one could have foreseen AlphaFold2 when the PDB was launched in 1971. Over the next decade researchers around the world are likely to generate hundreds of thousands of protein-fragment structures. I don’t pretend to know what the artificial intelligence tools of the future will be able to make of such data, but I hope they will have access.

 

What do you think?

Novel Assay (SPR) Leads to Novel Allosteric Binding Site (nAChR)

I love this blog.  But sometimes it drags.  Finding new articles to blog about can be hard, there seems to be waves.  So, I love it when someone points one out to me.  This article was sent to me by someone associated with it.  We don't always blog those articles people point out to us, but this one has all the hallmarks: high visibility journal, good target, and "innovative" in the approach.  So, let's dive in and see what we have here.

Nicotinic receptors (nAChR) are pentameric, ligand-gated ion channels.  nAChR mutations are implicated in a wide range of neurological disorders and are the targets for a wide range of current drugs.  It is important to note that most of these drugs work through an allosteric site distant from the agonist binding site.  Most high resolution, structural data is from homologous molluscan Acetylcholine binding proteins.  The orthosteric site is located at an interface of a "principal" and "Complementary" subunit.  Ligand-binding induces conformational changes in the ligand binding site which are coupled to the ion opening.  There is little information on structural implications at allosteric sites.  The authors decided to address this unmet need with a chimeric human α7 ligand binding domain and AChBP which has 71% sequence similarity to the native protein, compared to 33% for the most commonly used target (Aplysia).  

The authors used SPR against a target with a blocked orthosteric site.  They did this one of two ways: 1. pre-incubation with a high affinity orthosteric ligand or 2. mixing each fragment with an orthosteric ligand of lower affinity.  This second approach (Figure 1) allows detection of fragments not competing for binding to the orthosteric site and were therefore potential allosteric ligands.  

Figure 1.  To distinguish allosteric binders from competitive binders using SPR spectroscopy we perfused each fragment alone (green triangle) or in combination with the competitive antagonist d-tubocurarine (black circle). In the case of an allosteric binder, the response units observed for the mixture of fragment + d-tubocurarine is close to the sum of fragment alone + d-tubocurarine alone (blue dashed line). No competition exists because the fragment and d-tubocurarine bind at distinct sites. (C) In the case of a competitive binder, the response units for the mixture of fragment + d-tubocurarine is lower than the sum of fragment alone + d-tubocurarine alone because both compounds compete for binding at the same site. (D) Example traces for fragment 4, which was identified as one of the allosteric binders in this study.

 In screening 3000 fragments, they found 300 putative allosteric binders.  Follow up, including dose-response, led to 24 fragments being selected for co-crystallization.  Crystal trials were set up using blocked nAChR and fragments that were soluble at 5-10mM; yielding in the end 5 crystal structures.  All five proved to be allosteric binders in three separate locations, including one never observed before (top pocket).  The most potent of these fragments had a IC50 of 34 uM, while the least potent was 400 uM. 

Figure 2.  Allosteric Binding Sites

This is a really nice piece of work.  I think the SPR assay is really clever and I would expect that many people will now be taking a similar approach to discovering allosteric sites in their targets.

Way Down in the X-Ray Weeds

So, what I know about the details of crystallography can fit on the head of a pin...a small pin.  You put pure protein in multiwell plates and then do a huge matrix of crystallization conditions until tiny little crystals form.  Big crystals are best, but you can use tiny crystals or seeds, or with recent advances in technology, to actually collect data.  Then, through some wizardry (some sort of inverse transform) you make spots go to electron density, then with will power and what used to be SGI machines, you thread your protein sequence in, et voila a model of the structure.  I typically don't go for methods papers in fields I have almost no clue in, but this one intrigued me.

This paper aims to increase the efficiency of soaking fragments into crystals to take advantage of 3rd generation synchrotrons. These machines/labs/setups/doohickeys use acoustic droplet injection (ADE), which many people may already be aware of.  In this approach, each fragment soaks into a protein crystal either directly on data collection media or on a moving conveyor belt which then delivers the crystals to the X-ray beam.  The source of inefficiency comes from the time required to soak the fragment in to the crystals (for those where the apparatus is inside the X-ray station. I have no idea what that means, but here is a google image search that might give you an idea.)  A second source is the limit of evaporative dehydration during the fragment soak.  

Using the model system lysozyme and thermolysin the identified factors which can increase efficiency: namely smaller crystals can be used to decrease the soak time.  By small crystals, they are talking things that are 100 microns or less.  The authors go on to state that:

These techniques efficiently use fragment chemicals (~2.5 nL per screened condition), protein (~25 nL per screened condition), space (1120 screened conditions per standard shipping Dewar; no limits using a conveyor belt), and synchrotron beam time (less than 1 second/screened condition).  Evaporative dehydration of the protein crystal limits these fragment screening applications to systems where the fragment soak time is not prohibitive. Slow-binding compounds can be screened (without time constraint) in trays using ADE, but will consume significantly more resources such as purified protein and chemical compounds (~1 µl per screened condition). Hence, it is desirable to identify promising cases where the cost-efficient on-micromesh or on-conveyor soaking methods are adequate.

So, what did I really get out of this paper.  I am amazed by the miniaturization and automation that exists in synchrotrons.  It is really amazing. Its good to get and read the literature for another field.  It can be enlightening.  If I am looking at this correctly, they can screen a 1000 fragments with ~1mg of protein.  With that said, how many people screen for fragments in this way?  It seems not to be resource intensive, if you are sitting at a synchrotron.  But, how much does it cost to sit at that synchrotron?  Are the problems called out here something people encounter every day, or is this a "First World problem" for those sitting at synchrotrons.

What really got me was that the majority of the authors are high school students and undergraduates. This emphasizes to me the commoditization of X-ray, and really all services.  There is a very high level of training that goes in to solving the structure; I get that.  But it seems that many of the steps are commmodities, if you will.  When I was at Merck, they had a directive called (in some form) I-C-E: Innovation-Commoditization-Experimentation.  The concept was the highly trained (and highly paid) scientists needed to focus on innovation.  Once innovation was achieved it led to experimentation (figuring out how to run it routinely).  After that it was a commodity and should be outsourced to enable those scientists to go back to innovating.  It makes sense from a business standpoint, but scary from the scientists standpoint.  I am all for full employment for scientists in industry (trust me on this), but outsourcing can co-exist in industry. Look at the growth in providers from 2011 to 2014. Not sure where I am going with this, but food for thought.

Benchmark Your Process

So, not everybody agrees with me on what a fragment is.  As has been pointed out years ago, FBDD can be a FADD.  In this paper, from earlier this year, a group from AZ discusses how FBDD was implemented within the infectious disease group. Of course, because of the journal, it emphasizes how computational data is used, but you skim over that and still enjoy the paper :-). They break their process into several steps.

Hot Spots: This is a subject of much work, particularly from the in silico side.  In short, a small number of target residues provide the majority of energy for interaction with ligands.  Identifying these, especially for non-active site targets (read PPI), is highly enabling, for both FBDD and SBDD. To this end, the authors discuss various in silico approches to screening fragments.  They admit they are not as robust as would be desired (putting it kindly).  As I am wont to say, your computation is only as good as your experimental follow up.  The authors indicate that the results of virtual screens must be experimentally tested.  YAY!  They also state that NMR is the preferred method; 1D NMR in particular being the AZ preferred method.  [This is something (NMR as the first choice for screening) that I think has become true only recently.  Its something I have been saying for more than a decade, but I guarantee my cheerleading is not why.] They do note that of the two main ligand-based experiments, STD is far less sensitive than WaterLOGSY.  There is no citation, so I would like to put it out there, is this the general consensus of the community?  Has anyone presented data to this effect?  Specifically, they screen fragments 5-10 per pool with WaterLOGSY and relaxation-edited techniques.  2D screening is only done for small proteins (this is in Infection) and where a gram or more of protein is available.

Biophysics:  They have SPR, ITC, EPIC, MS, and X-ray.  They mention that SPR and MS require high protein concentrations to detect weak binders and thus are prone to artifacts.  They single out the EPIC instrument as being the highest throughput.  [As an aside, I have heard a lot of complaints about the EPIC and wonder if this machine is still the frontline machine at AZ.]  60% of targets they tried to immobilize were successful.  They also use "Inverse" SPR, putting the compounds down; the same technology NovAliX has in their Chemical Microarray SPR.  In their experience, 25% of these "Target Definition Compounds" still bind to their targets. 

They utilize a fragment-based crystallography proof of principle (fxPOP).  Substrate-like fragments (kinda like this?) are screened in the HTS, hits [not defined] are then soaked into the crystal system, and at least one structure of a fragment is solved.  This fragment is then used for in silico screening, pharmacophore models, and the like.  So, this would seem to indicate that crystals are required before FBDD starts.  They cite the Astex Pyramid where fragments of diverse shape are screened and the approach used at JnJ where they screen similar shaped fragments and use the electron density to design a second library to screen.

As I have always said, there are non-X-ray methods to obtain structural information.  AZ notes that SOS-NMR, INPHARMA, and iLOE are three ways.  These are three of the most resource intensive methods: SOS-NMR requires labeled protein (and not of the 15N kind), INPHARMA requires NOEs between weakly competitive ligands (and a boatload of computation), while iLOE requires NOEs of simultaneously binding ligands.  I think there are far better methods, read as requiring fewer resources, to give structural information more quickly (albeit at lower resolution).

The Library:  The describe in detail how they generated their fragment libraries.  They have a 20,000 fragment HCS library.  The only hard filter is to restrict HA less than 18.  I fully support that.  They also generated a 1200 fragment NMR library biased towards infection targets.

The Process:   The authors list three ways to tie these methods together:

1. Chemical Biology: Exploration of binding sites/development of pharmacophores.  I would add that this is also for target validation.  As shown by Hajduk et al. and Edfeldt et al., fragment binding is highly correlated to advancement of the project. 
2. Complementary to HTS.  At the conference I am at today, one speaker (from Pfizer) said that HTS was for selectivity, FBDD was for efficiency (or Lord, here comes Pete with that one).  I really like that approach.
3. Lastly, stand alone hit generation.  

I think this paper is a nice reference for those looking to see how one company put their FBDD process in place. Not every company will do it the same, nor should they.  But there is a FBDD process for every company.

It's A Start

As the readers of this blog know, I tend to be harsh on academic "Drug discovery" papers.  Sometimes, there is a really worthwhile academic paper, but by and large I find that they tend to publish things that are barely "drug discovery" and more the For Dummies...of what they think drug discovery is.  Which way will I swing on this paper from researchers Down Under?  

The bacterial Sliding Clamp, aka polymerase 3beta, is a key player in bacterial replication and is an "emerging" target. It interacts with other proteins via LM (Linear Motifs): 4-10 amino acid disordered regions.  This is typically a weak interaction (1-100 uM). These LM exist at termini, but sometimes in loops.  A consensus sequence for the LM that interacts with the Sliding Clamp has been identified: QLx1Lx2F/L (S/D preferred at x1; x2 may be absent).  Two classes of compounds have been identified previously but with >10 uM affinity and no -cidal activity. 

So, these authors went after this target using X-ray as the primary screen.  The Zenobia Fragment Library was used (352 molecules) to soak into crystal in pools of 4 fragments.  Four fragments (below) were found to bind to Subsite I on Chain A.  However, no changes in the main chain density were observed.

They also found several other fragments with weak density, and several that were deemed crystallographic artifacts.  None of these compounds showed significant activity below 1 mM in their competition assay.  So, the story then continues that they "sought to improve binding affinity by identifying fragments that could more completely occupy" the binding site.  

[An aside:  To me, this brings up an important point about the choice of fragment collection.  Fragments that are designed for X-ray soaking tend to be small (10-12 HAC).  Just from a theoretical standpoint, those fragment would have to have an affinity in the 250 uM range (LEAN 0.3).   This was covered in a poll and most most people are happy going < 10 HAC.  My question is how often is a very small fragment found as an active?]

To do this, they noted that the fluoro-phenyl group in 1 was previously reported, leading to investigations with compound 5. It was found to fully occupy the binding site.

They searched ZINC for compounds similar to 1-5.  Their initial purchases failed to find any compounds with activity < 1mM.  Eventually, they landed on the hypothesis that chlorocarbazoles were "promising", leading to compound 6.  At this point, I think Dan's head exploded, Scanners-style.  Yes, that is an epoxide.  The co-crystal structure showed that it was binding in the active, albeit with weak electron density.  Their SAR, wisely, did not include the N-alkyl epoxide. 

Both 7 and 8 show good LE and LLEAT.  Only the R enantiomer of 8 caused movement in the main chain.  It also was the most potent in the replication inhibition assay (64 uM).  It was also the most potent in terms of -cidal activity against both Gram positive and negative microbes. 

So, is this a good or bad paper?  I would say it is a start, but if I had been a reviewer I would have made them change the title "Discovery of Lead Compounds Targeting the Bacterial Sliding Clamp
Using a Fragment-Based Approach" to "Discovery of ACTIVE Compounds Targeting the Bacterial Sliding Clamp Using a Fragment-Based Approach".

Review of 2013 reviews

The year is coming to an end, and as we did last year, Practical Fragments is looking back at notable events as well as reviews that we haven’t previously highlighted.

The fragment calendar started in March in Oxfordshire, at the RSC Fragments 2013 conference, closely followed in April by CHI’s FBDD meeting in San Diego (here and here). Closing out the year for conferences that Teddy or I attended was the Novalix conference on Biophysics in Drug Discovery in Strasbourg (here, here, and here).

There weren’t any new books published (though the special issue of Aus. J. Chem. practically counts as one), but there were several notable reviews.

Stephen Fesik and colleagues at Vanderbilt University published “Fragment-based drug discovery using NMR spectroscopy” in J. Biomol. NMR. This is an excellent overview that covers library design, NMR screening methodologies, and compound optimization. The researchers make an interesting case for including multiple similar compounds and allowing for larger, more lipophilic fragments, while always being careful to avoid “bad actors”. They also do a good job of summarizing the various NMR techniques, including their strengths and limitations, in language accessible to a non-spectroscopist. Finally, the section on fragment linking discusses the theoretical gains in affinity, the practical challenges to achieving these, and strategies to overcome them.

Turning to the other high-resolution structural technique, Rocco Caliandro and colleagues at the CNR-Istituto di Cristallografia in Italy published “Protein crystallography and fragment-based drug design” in Future Med. Chem. This provides a fairly technical description of X-ray crystallography and its role in FBDD, along with a table summarizing around 30 examples, five of which are discussed in some detail.

Of course, it’s always best to use multiple techniques for finding fragments, so it’s well worth perusing “A three-stage biophysical screening cascade for fragment-based drug discovery,” published in Nature Protocols by Chris Abell and colleagues at the University of Cambridge. This expands on a gauntlet of biophysical assays (involving differential scanning fluorimetry (DSF), NMR, crystallography, and isothermal titration calorimetry (ITC)) that we discussed earlier this year. Nature Protocols are highly detailed, with lots of troubleshooting tips, so this is a great resource if you’re exploring any of these techniques.

Finally, Christopher Wilson and Michelle Arkin at the University of California San Francisco published “Probing structural adaptivity at PPI interfaces with small molecules” in Drug Discovery Today: Technologies. Protein-protein interactions are frequent targets for FBLD: see for example here, here, here, here, and here – and that’s just for 2013! The current review gives a nice overview of the technology called Tethering, focusing on the cytokine IL2 and an allosteric site on the kinase PDK1.

And with that, Practical Fragments thanks you for reading and says goodbye to 2013. May your 2014 be happy and fulfilling!

Fragment linking on RPA: another protein-protein interaction inhibitor

Protein-protein interactions have often been targeted by fragment efforts, partly I think out of desperation when all else fails. That said, there have been notable successes. Earlier this year we highlighted one example from Stephen Fesik’s group at Vanderbilt University. In a recent paper in J. Med. Chem., the same lab now reports progress on a different target.

Replication protein A (RPA) is important for DNA replication and repair, and is thus an intriguing anti-cancer target. RPA binds to single-stranded DNA as well as to various other proteins involved in the DNA-damage response, such as ATRIP. The site targeted here is the “basic cleft” of RPA that binds ATRIP.

The researchers used the venerable SAR by NMR approach, screening a library of 14,976 fragments against ¹⁵N-labeled protein using HMQC and looking for changes in chemical shifts. A total of 149 fragments produced significant and specific chemical shift differences at 0.8 mM concentration. One of the nice features of SAR by NMR is that not only do you get hits, you find out where they bind. In this case, most of the hits bind in the basic cleft. This region has two sub-sites; some fragments bind to one or the other, while many bind to both. Not surprisingly for a basic binding site, most of the fragments identified are negatively charged.

Although all the hits are relatively weak (the best have dissociation constants around 0.5 mM), some could be improved through various strategies to low micromolar inhibitors, the subject of a paper earlier this year.

In the current paper, the researchers used crystallography to further define the binding modes of select fragments. They found that fragment 2 and fragment 4 could bind to both subsites of the basic cleft, but that when co-crystallized together fragment 2 binds to one subsite while fragment 4 binds to the other. The two fragments come within a few Ångstroms of one another, suggesting that they could be linked.

Fragment linking doesn't always work as well as one might hope, and although the initial linked compound 7 is nearly 30-fold more potent than fragment 4, its ligand efficiency drops considerably. However, structure-based optimization to compound 8 was able to improve the affinity by another two orders of magnitude.

Teddy has argued that SAR by NMR is dangerous because of its reliance on labeled protein and because the initial application involved fragment linking, leading people to believe that these are necessary requirements for successful prosecution of fragments. There are now plenty of examples of using other methods to find and advance fragments, but this paper illustrates that SAR by NMR can still be incredibly powerful.

Of course, the final molecule reported here has warts, notably a thioamide, two carboxylic acids, a molecular weight over 600, and a ClogP>7. Indeed, the absence of reported cellular data is perhaps telling. And yet, Bcl inhibitors are also superficially unattractive but are in the clinic. Clearly more medicinal chemistry needs to be done on these molecules, if nothing else to improve potency, but that’s not to say there isn’t a path forward. It will be fun to watch this story progress.

Fragments vs CHK2: high-concentration screening comes through

Checkpoint Kinase 2 (CHK2) is an oncology target that has been kicking around for years. Its relevance is still debated, so having more small molecule inhibitors would go a long way toward assessing its therapeutic potential. In a recent paper in PLoS One, Rob van Montfort and colleagues at The Institute of Cancer Research (UK) present their fragment-based efforts on CHK2.

The researchers describe the design of their screening library in some detail, starting with a series of typical computational filters on commercially available molecules. Although most of the molecules had MW < 300, molecular weights up to 320 Da were allowed for fragments containing F, Cl, or SO₂ moieties. Also, all fragments were required to have at least 10 heavy atoms, which is on the high-side for a minimum. A total of 1869 fragments were purchased. All of these were analyzed for solubility and purity (by nephelometry and LC-MS, respectively), though unfortunately the researchers do not provide pass rates.

Having assembled the library, the researchers then screened each fragment at 300 micromolar against CHK2 in a biochemical assay (AlphaScreen). This led to 45 hits, but 25 of these showed some interference with the AlphaScreen assay itself. However, the remaining 20 all showed dose-response curves in a different assay format, giving IC₅₀ values from 2.7 to 944 micromolar.

In parallel, the researchers screened CHK2 using a thermal shift assay, with each fragment present at 2 mM. Perhaps not surprisingly given the higher concentrations used, this led to 63 hits.

Where things got interesting – and encouraging – was when the researchers compared hits identified using the two methods. In contrast to others' experiences, there was reasonable overlap; of the 14 hits from both assays, 12 yielded measurable IC₅₀ values when assessed using a microfluidic functional assay. Most of the AlphaScreen hits that didn’t produce thermal shifts came from the set of 25 that had previously been flagged as interfering with the AlphaScreen assay itself, and several were also insoluble. Of the 49 thermal shift hits that did not show up in the AlphaScreen assay, 13 were insoluble. Regarding the remaining 36, the researchers propose that they may bind to CHK2 outside its active site and thus don’t inhibit enzymatic activity.

Next, the researchers attempted to characterize the binding modes of the fragments crystallographically. Of the nine fragments that produced structures, eight came from the set of fragments confirmed using both AlphaScreen and thermal shift. Significantly, the only fragment to yield a structure that was identified solely from the thermal shift assay also produced the worst IC₅₀ value (228 micromolar) and the lowest ligand efficiency. All nine fragments bind to the so-called hinge region of the kinase.

One interesting observation was that, although the library did contain larger molecules, 6 of the 9 fragments characterized crystallographically had MW < 200, and the other 3 were well under 300 Da. This is exactly what you would expect according to the concept of molecular complexity, and suggests that adding larger fragments to your library may actually lower your hit rate (though admittedly it may be a stretch to conclude too much from this one study).

Another interesting note is that co-crystallization was used in all cases. Folks sometimes believe that you need to grow vats of crystals for fragment soaking experiments, but co-crystallizing worked fine here, and in some cases may allow the protein to adopt different conformations than if grown in the absence of small molecules.

Overall, the paper presents some nice starting points against CHK2. Perhaps more important, this is a thorough, well-written, and open access account of fragment screening that is well worth perusing by anyone embarking on a fragment campaign.

Fragments vs MDM2 – retrospectively

Protein-protein interactions (PPIs) are some of the tougher targets in drug discovery: they often have relatively flat interfaces that lack obvious small-molecule binding sites. But despite large surface areas, these interfaces often have smaller hot spots, and fragment-based approaches have succeeded against targets where traditional screening approaches have failed. In a paper in this month's issue of ACS Med. Chem. Lett., David Fry and colleagues at Roche describe a retrospective fragment study against a classic PPI (this story was also presented at the CHI conference earlier this year).

Molecules that disrupt the interaction between p53 and MDM2 have been sought since the 1990s as potential anti-cancer agents. The first such inhibitor to enter the clinic was RG7112, one of the so-called nutlins developed at the Roche Nutley site, which is sadly in the long process of closing. Although this series of compounds originated from a high-throughput screen, the researchers were interested to know whether they could have been discovered from fragments. To do so, they deconstructed RG7112 into a series of smaller pieces.

RG7112 is a “tripod-like” molecule with three hydrophobic moieties (blue, red, and green, below) that bind to subpockets on MDM2, as well as a polar “cap” (pink) that projects into solvent but is nonetheless important for activity. The researchers were unable to detect the binding of any of these individual fragments by either SPR or two-dimensional (HSQC) NMR, suggesting that each moiety alone was too weak, so the researchers tried lopping off one or two pieces from RG7112. In all cases, pared-down molecules containing three out of the four moieties led to molecules with affinities in the 14 to 1000 micromolar range. The tipping-point seemed to be when they kept just two of the four component fragments: in two cases the resulting molecules showed no detectable binding, while in two other cases they did. Compound 5, in fact, was a decent hit by any measure, and crystallography revealed that this molecule binds in the same manner as RG7112.

Interestingly, compound 5 makes interactions with two adjacent subpockets, including a deep hole that seems to be critical for all MDM2 binders; p53 inserts the indole moiety of a tryptophan into this site.

However, compound 5 does teeter on the edge of what could be called a fragment: with a molecular weight of 305 it falls just outside the Rule of Three (sorry Teddy!), and with 20 non-hydrogen atoms it sits right on the border of what most people seem to include in their libraries. The researchers suggest that larger fragments may be necessary for tackling PPIs, though they recognize that doing so will increase the number of fragments necessary to sample chemical space.

This is a beautiful, thorough study, but I do worry about super-sizing fragments. There are now multiple success stories of finding and advancing fragments against PPIs, including such “teflon-targets” as Mcl-1. Perhaps the nutlins would not have been discovered starting from small fragments, but with hundreds of billions of possibilities there are certainly other, even more ligand-efficient leads out there waiting to be discovered.

Eight Annual Fragment-Based Drug Discovery Meeting (part 2)

The last major fragment event of 2013, CHI’s FBDD, wrapped up earlier this week in San Diego. As with last year this summary is not meant to be comprehensive (and you can also read Teddy’s impressions here.)

The FBDD track is just one of six within the CHI Drug Discovery Chemistry Conference. One indication of the success of the field is its appearance in several of the other tracks: attendees were likely to hear about fragments without going to the FBDD track at all.

In the GPCR track, Robert Cooke from Heptares discussed the application of fragments to the β1-adrenergic receptor (see also here). In the kinase track, Hongtao Zhao presented in silico fragment work (see also here). And in the protein-protein interaction track, David Fry from Roche described a deconstruction of the p53-HDM2 inhibitor RG7112 into its component fragments to see whether the molecule could have been identified from FBLD. RG7112 consists of a central core with four appendages, and although the mono-substituted core was too weak to detect, some of the cores with two substituents could be identified and bound to the protein in the same manner as the parent compound. However, these did tend to be super-sized fragments, with molecular weights in the 300-350 Da range.

Protein-protein interactions were also a theme of Richard Taylor, from the company UCB. They built a sizable fragment library of about 23,000 (mostly commercial) compounds designed to cover molecular frameworks found in known drugs. UCB has invested heavily in SPR technology, and with a stable of four Biacore 3000 instruments could rapidly screen this entire library against a dozen protein-protein interaction targets. Not surprisingly, given the difficulty of this target class, the hit rate was much lower than in conventional fragment screens, averaging just about 1%. What was interesting is that only 964 fragments hit any target – at less than 5%, this is much lower than the roughly 33% hit rate seen in other fragment libraries. Most of these fragments were reasonably specific, though; 908 hit ≤ 8 targets. It will be interesting to see whether anything can be learned about “privileged” protein-protein interaction fragments from this set.

Of course, extracting general trends from collections of fragments is not necessarily straightforward. Teddy has already brought up the difficulties of describing molecular shape; as he pointed out in his presentation, Fsp³ is not the best measure of “three-dimensionality” for several reasons. For example, even toluene has an Fsp³ = 0.14, and while Pete Kenny correctly points out that aromatic molecules do have volume, most chemists would think of this as a very “flat” compound. Principal moment of inertia (PMI) is better, but is harder to calculate. Happily, as Justin Bower described in his presentation, the Beatson Institute is allowing other researchers to use their 3DFIT software to calculate PMI and other properties.

One of the criticisms of 3D fragments is that, as Rod Hubbard pointed out, they can be a “pain in the neck” for chemistry. One solution that researchers at Vernalis took was to do analog work on a simpler molecule, then scaffold-hop back to the original fragment once the SAR was sufficiently understood to justify investment in more challenging chemistry.

Finally, the question of what to do with fragment hits that don’t reproduce in different assays was the topic of at least one breakout discussion and was also extensively discussed by Peter Kutchukian, who presented an analysis of 134 fragment screens using a variety of techniques against 34 different targets at Novartis. Some of this was presented at FBLD 2012, but one interesting finding was that hits from biophysical screens (such as SPR, NMR, or DSF) tended to cluster separately from hits in biochemical assays. Given the number of ways to find fragments, pursuing hits that confirm in both a biochemical and a biophysical method may help to weed out artifacts, though at the risk of increasing false negatives.

Feel free to chime in with your thoughts and impressions, whether or not you were there. And if you are kicking yourself for not attending, next year’s meeting is scheduled to return to San Diego from April 22-25.

Seventh Annual Fragment-Based Drug Discovery Meeting

CHI’s annual FBDD meeting took place in San Diego this week, and since this was the first time in a while both Teddy and I have been in the same place we’ve decided to make this a joint post. As with last year this does not aim to be comprehensive.

One of the highlights of the conference was a set of three talks on BACE1 inhibitors from Amgen (Ted Judd), Lilly (David Timm), and Pfizer (Ivan Efremov), the first two of which have been discussed here and here. It’s nice to see fragments playing a pivotal role in delivering advanced leads and – at least in the case of Lilly and Merck – clinical candidates against what has been one of the most difficult drug targets in industry.

Speaking of difficult targets, Till Maurer of Genentech gave a lovely presentation on using NMR-based fragment screening to discover inhibitors of the holy grail of oncology, Ras. They’ve recently published some of this story, which we’ll highlight in an upcoming post.

A common question is ‘how often do you find the same fragment using different methods?’ (see here for an ongoing discussion on LinkedIn). Cynthia Shuman from GE gave a nice case study in which she screened the protein PARP15 against 987 fragments using a Biacore T200. Of the 15 fragments with shapely, well-behaved sensorgrams, 14 were confirmed by NMR. On the other hand, only one of these hits was detected in a differential scanning fluorimetry (thermal melt) assay.

Marcel Verdonk at Astex described general trends from mining in-house and published data. After looking at 43 in-house targets, he found that 8000 compounds had been tested against 2 or more proteins, and after plotting by molecular weight found that, consistent with the original Hannian model, larger compounds are more selective. In a separate analysis of 53 fragments that had been advanced to leads, he found that in most cases the initial fragment maintained roughly the same position and orientation from start to finish.

Rod Hubbard, Teddy and I all ran round-table discussions, but only Teddy kept notes, which are summarized here.

The topic started as a discussion of 2D vs. 3D fragment libraries. In the recent Pfizer fragment build, a group of diverse chemists eyed every compound, and at least five had to agree to each molecule before it went into the library. 

The discussion went briefly to the old fight: Are nitros masked amines or noxious moieties?  The table ended up agreeing that if you would remove the nitro group or change it anyways, why put it in the first place?

We then dove right back into the 2D vs. 3D debate. Kinases seem to love 2D fragments, while other classes of targets seem to NEED 3D fragments. One idea discussed was 3D fragments as complements to 2D fragments. It was mentioned that 3D fragment libraries would need to be MUCH larger to cover equivalent chemical space. I thought the idea that 3D fragments would be exploring “vector” space rather than “chemistry” space would mean that you could go with a much smaller library, if you want to use it for vector space searching. It was also proposed that 2D fragments tend to be much smaller (~150-180Da-ish) and 3D fragments would be, by necessity, bigger (~250 Da).

The topic then changed to SBDD (structure-based drug discovery) as part of FBDD. Most people at the table were of the opinion that they wouldn’t use FBDD (on a normal priority target) without SBDD. And with SBDD, you don’t need 3D fragments to explore vector space since you will have the X-ray to guide you. 

The point was made that 3D fragments also HAVE to be scaffolds which would end up in the final compound. If you are simply using a scaffold to explore vector space without any hope of it ending up in the final product, why bother. [TZ Note: I think this is very shortsighted and people do not understand that targets will most likely NOT have SBDD to guide them, at least in the fragment-based lead generation stage.] 

The question was asked to the table: is FBDD a valid approach in the absence of structure? Three people said yes, the rest (~8) said no. All three who said yes were small company/ CRO people. Everyone agrees if you do go with FBDD without structure you need to have a HEAVY investment in biophysics to characterize the protein and the hits. 

One person asked about low confirmation hits following a fragment screen: the table agreed that this is most likely the result of the library being “wrong”. We finished the discussion by asking if 2D fragments are more pan-class (not targeted to a specific class of targets) and 3D fragments may be more target-focused. The resounding answer was “Who knows, but why would they be?”

As this post is already getting long we’ll stop here, but for those of you who were also at the conference please add your comments. And if you missed this one, there are still several exciting upcoming events this year!

Fragment Based Lead Discovery San Diego 2011

Zenobia Therapeutics hosted a one-day FBLD meeting in San Diego last Friday. It was a very full day of seven talks and plenty of informal discussion, with both lunch and happy-hour on a rooftop terrace overlooking the ocean. I’ll try to give some flavor of the event below – those of you who were there, please chime in with your own impressions.

In the first talk of the day, Michael Recht of Xerox’s PARC discussed the construction and use of array-based nanocalorimetry devices, which analyze samples as small as a couple hundred nanoliters. Although conceptually related to isothermal titration calorimetry, the scale of nanocalorimetry makes direct measurement of binding enthalpy difficult. However, the devices are well-suited to measuring the heat output from enzymatic reactions, thus allowing detailed study of enzyme kinetics and inhibitor characterization (see here for a full description and here for a review, both open-access).

Maurizio Pellecchia of the Sanford-Burnham Medical Research Institute gave a wide-ranging talk with the common theme of NMR as an enabling method. In particular he noted that, due to its low protein requirements, ILOE (which we’ve written about previously) could be considered the “homeopathy of NMR;” a recent paper in J. Med. Chem. describes its application to generate hits against Bcl-family proteins. He used a different NMR technique to discover a potent and selective JNK inhibitor that binds in both the ATP and substrate binding sites. On a completely different topic, Maurizio noted that roughly one-third of drug targets contain metals, prompting the construction of a fragment-library designed with metal-chelation in mind, a strategy we’ve written about previously (see also here). Many of these techniques and libraries have now been licensed into a new company, AnCoreX Therapeutics, founded by Maurizio.

Schrödinger was a major sponsor of the meeting, and company scientist Kathryn Loving gave a good overview of the potential of their suite of software tools for fragment-based lead discovery. After lunch I presented a PDK1 story (see here) as well as applications of Carmot’s Chemotype Evolution, and Michelle Arkin from UCSF discussed some of the capabilities of the university’s Small Molecule Discovery Center, with special focus on protein activators and protein-protein disruptors.

Tony Giannetti of Genentech gave a nice presentation on caspase-6 inhibitors and also discussed the power of SPR, a topic on which he’s written extensively. He noted that as long as one is careful about conducting the experiment, the number of promiscuous or non-specific binders found by SPR tends to be less than the number of true hits – he described the former as “a managed headache.”

Tony is in the enviable position of having screened numerous targets using the same platform, allowing him to do some very interesting data mining. One phenomenon that we’ve previously commented on is the fact that, across multiple companies, only about a third of fragments hit in one or more screens, and Roche is no exception (36% across 13 screens). With the merger of Genentech and Roche now complete, Tony has been able to compare the fragment libraries assembled separately at each institution. Interestingly, although they contain < 5% common fragments and < 10% similar fragments, the libraries are, by broad measurements such as size or lipophilicity, statistically indistinguishable. However, the hits from the Roche collection seem to have higher ligand efficiencies. Also, the most potent binders tend to have 13-17 atoms, with the most ligand-efficient molecules generally having a molecular weight < 200. Finally, there has been a lot of recent discussion on making fragments less flat (aromatic) and more three-dimensional (often aliphatic). However, as I predicted 2 years ago, one needs to be cautious: it turns out that hits were actually enriched for flatter fragments, although Tony did mention that most of the screens were against kinases, which may be particularly “flatophilic”.

Vicki Nienaber, founder and president of Zenobia, closed the conference by discussing the potential for fragment-based approaches to tackle diseases of the central nervous system (CNS). The rationale is that, since CNS drugs need to be small to cross the blood-brain-barrier, it pays to start with fragments. Indeed, Zenobia’s fragment library has a molecular weight averaging just 150-175, with an upper limit of 225, which should allow them to avoid molecular obesity. Vicki described the discovery of nanomolar leads against LRRK2, a Parkinson’s disease target, and against PDE10, a Huntington’s disease target.

As far as I know this was the last face-to-face fragment event this year, but there are still several webinars this month and next, and it’s not too soon to start planning for 2012; please see here for upcoming events.

How effectively can fragments sample chemical space?

One of the key advantages of fragment-based drug discovery is that, since there are fewer fragments than lead-sized or drug-sized molecules, it is possible to sample chemical space far more efficiently with fragments than with larger molecules. At least, that’s the theory, but does is it hold true in the real world?

To put it another way, do fragments sample all of the space in which drugs are found? And what kinds of fragments are best for this sampling? In the most recent issue of J. Med. Chem., Stephen Roughley and Rod Hubbard of Vernalis address such questions.

The system they investigate, heat shock protein 90 (Hsp90), is an ideal model system: it is both a popular anti-cancer target as well as structurally tractable, and is thus arguably the most heavily explored single target in terms of fragment-based lead discovery. At least 8 antagonists have entered the clinic, of which at least 2 have come from fragments (see the posts on AT13387, NVP-BEP800/VER-82576, and posts on Evotec compounds discovered by fragment growing or linking.)

Vernalis has had a long-running fragment-based program targeting Hsp90, which has resulted in numerous fragments whose binding modes have been determined by X-ray crystallography. Roughley and Hubbard analyzed these fragments and compared them to published inhibitors. Just 5 distinct fragments can be mapped onto all of the clinical compounds: a handful of fragments effectively samples relevant chemical space. As the authors put it:

For Hsp90 at least, the fragments do cover an appropriate chemical space; what is then important is the imagination of the chemist in evolving the fragments into potent inhibitors.

The second point – about the imagination of the chemist – is critical. Mapping fragments onto elaborated molecules is easier to do retrospectively than prospectively; a cynic could argue that methane is a fragment of just about any drug out there. However, Roughly and Hubbard also point out that, particularly in cases such as this where there are many co-crystal structures, fragments can help identify bioisosteres, including cryptic ones that would not be obvious purely from studying functional SAR.

The paper also addresses the issue of optimal library design, in particular the dilemma of size. Although all five representative fragments were found in an initial set of just 719 fragments, subtle changes can dramatically change the binding mode, an issue we’ve touched on previously. It may not be practical to have multiple similar fragments present in a primary screening library, but testing close analogs after identifying initial fragment hits is likely to be worthwhile.

Finally, one of the concerns about fragment-based approaches is that, if everyone is buying the same set of fragments from the same suppliers and screening them against the same targets, they will end up in the same place – and stumbling over each others’ intellectual property. Reassuringly, this turns out not to be the case:

Even though the various companies discovered rather similar compounds from a fragment screen, exploiting similar binding motifs, there were no exact matches. [Also], the subsequent evolution of the fragments sometimes took very different paths and produced mostly very different chemical leads and candidates.

If this holds true for such heavily mined targets as Hsp90 (and kinases, as discussed previously) it should be even more true for newer classes of targets.

There is a wealth of information in this paper, and it is worth perusing, especially if you find yourself longing for some science over the long holiday weekend.

Sixth Annual Fragment-Based Drug Discovery

The only US-based conference completely devoted to fragment-based drug discovery ended in San Diego this week. As with last year, I won’t attempt to summarize all of the talks – there was far more information presented than I have time to write (or that you probably have patience to read!) For those of you who were there, please feel free to mention some of the things I missed.

One of the points that Don Huddler (GlaxoSmithKline) and I (Carmot) made in the pre-conference short-course is that finding fragments is a solved problem. As Rod Hubbard (Vernalis, University of York) noted in his opening presentation, “it’s pretty simple to find fragments that bind; a graduate student can do it in a couple months.” Even membrane proteins are starting to yield to fragment-based screening, as Gregg Siegal (ZoBio, Leiden University) discussed in his closing session (see also here).

That’s not to say that new methods for finding fragments aren’t useful, particularly if they open new target space, are faster or more reliable, or provide new information. An example of the latter was the presentation by Denis Zeyer (NovAliX) on native mass-spectrometry (see also here). Because hydrophobic interactions are weaker in the gas phase than in water, it should be possible to select for molecules that bind predominantly through polar interactions. In fact, by gradually increasing the voltage in their MS instrument, Zeyer and colleagues generated “VC50” curves, the voltage at which half the compound dissociates from the protein. At least in one case, a higher VC50 correlated with the presence of an additional hydrogen bond to the protein compared with related molecules.

Polar contacts are generally associated with enthalpic rather than entropic interactions, and whether such fragments are preferable was the subject of some discussion, particularly at a breakfast round-table discussion. In contrast to a meeting just last year, several participants were actively collecting thermodynamic data, though there was some uncertainty as to what to do with it. This is a controversial subject; one person suggested that enthalpic binders are likely to be more hydrophilic than entropic binders, so just keeping an eye on lipophilicity is likely to be just as useful and far easier than actually measuring thermodynamic parameters. Charles Reynolds (Ansaris) provided an analysis that illustrates some of the difficulties in using thermodynamic data – I’ll follow up on this in a later post.

The shape of fragments has been previously discussed, and Ivan Efremov (Pfizer) gave a nice case study of a strikingly three-dimensional fragment: an X-ray screen of 340 molecules against BACE resulted in a single hit, a spirocyclic pyrrolidine. The electron density of this was so clear that it didn’t even need to be deconvoluted from the other three compounds in the pool, and medicinal chemistry ultimately led to low micromolar inhibitors.

There was general consensus that ligand efficiency (and various lipophilicity adjusted versions) is a helpful metric. One practitioner said that his company had sometimes pursued more chemically tractable but less ligand efficient fragments and generally came to regret those decisions. But a fragment with lower ligand efficiency could still be interesting: with fragments, even small changes could have huge effects on binding (see for example AT13387, which was discussed by Chris Murray of Astex). Thus, a bit of initial fragment optimization could be a good investment before pursuing more intensive chemistry, particularly if commercial or in-house analogs are available. Interestingly, I couldn’t find anyone who uses either fit quality or %LE.

In the early days of fragment-based lead discovery a common selling point was that it sped up drug discovery, but a theme in this meeting was that it is not necessarily faster but can provide leads against more difficult targets or better leads against “normal” targets. Of course, one has to be wary of taking a good fragment, slapping a bunch of grease on it, and turning it into a lipophilic monster.

Indeed, an analysis of fragment-derived leads published a couple years ago was not flattering. Taking up the thrown gauntlet on behalf of fragments, Chris Murray presented a retrospective analysis of all 42 fragment to lead programs at Astex (including 21 kinases and 9 proteases). The average parameters of these leads were considerably more attractive in terms of both molecular weight and ClogP that the published values of the HTS hits. At least according to this analysis, fragment approaches have the potential to deliver superior molecules, as long as one is disciplined and creative in how these approaches are applied.