Current Topics in Medicinal Chemistry Special FBDD Issue

This year ends with an entire issue of Current Topics in Medicinal Chemistry devoted to FBDD. Rob van Montfort and Ian Collins provide a brief editorial overview of the six papers. Collins and colleagues also describe their application of fragment-based methods to develop inhibitors of the anti-cancer target protein kinase B (AKT).

Half the papers cover various forms of fragment screening: the Medical Structural Genomics of Pathogenic Protozoa Consortium describes their crystallographic approach, Helena Danielson discusses the use of SPR, and Gregg Siegal and Johan Hollander present their Target Immobilized NMR Screening (TINS) methodology.

Charles Reynolds and colleagues review metrics, such as ligand efficiency and fit quality, for evaluating fragment hits.

Finally, in the longest article in the collection, Vicki Nienaber discusses how fragment-based methods may be particularly useful for discovering compounds that will cross the blood-brain-barrier to target the central nervous system.

As 2009 comes to a close, Practical Fragments would like to thank our readers, old and new. May you all have a happy and productive 2010!

Fragment-based conferences in 2010

There have been some great conferences this past year (see summaries of some of them here and here), and 2010 is shaping up nicely, particularly with the announcement of FBLD 2010 (below).

February 3-5: CHI’s Molecular Medicine Tri-Conference will be held in my beautiful city of San Francisco, with a program on medicinal chemistry that includes a fragment track, and a short course on “Fragment-Inspired Medicinal Chemistry” on February 2.

March 21-25: The spring ACS meeting will also be held in San Francisco. There will be a symposium on “Fragment Based Drug Design: Novel Approaches and Success Stories.”

April 20-25: The Keystone Symposium on computer-aided drug design will take place in Whistler, British Columbia. Although not exclusively devoted to fragments, the schedule shows plenty of talks on the topic.

April 27-28: Cambridge Healthtech Institute’s Fifth Annual Fragment-Based Drug Discovery will be held in San Diego. And if you missed the short course in San Francisco, you have another chance on April 26.

June 6-9: The 32nd National Medicinal Chemistry Symposium will be held in Minneapolis, Minnesota, and Dave Rees is organizing a session on fragments on June 9. Looks like a great lineup, with top speakers from Astex, Plexxikon, Novartis, Abbott, and UC Berkeley.

October 10-13: Finally, standing alone in the second half of the year, Fragment-based Lead Discovery 2010 will be held in Philadelphia, PA. This is the third in a popular series of conferences that started with FBLD 2008 in San Diego and continued this year in York, UK. An emphasis next year will be on biophysical methods - old and new - for fragment identification and characterization, as well as sessions on libraries, chemical strategies for fragment evolution, and success stories. A web site will be available shortly where further details will be posted. We anticipate a call for oral and poster contributions during the Spring of 2010. As far as we know this is the first major fragment event on the east coast of the US, so don't miss it!

Know of anything else? Organizing a fragment event? Let us know and we’ll get the word out.

Natural linking – though not of fragments

We’ve previously discussed the appeal and challenges of fragment linking. A new paper in Science describes how a naturally occurring antibiotic makes use of a linking strategy, albeit using rather Brobdingnagian fragments.

Simocyclinone D8 (SD8) is a dumbbell shaped molecule isolated several years ago from that ultimate micro-pharma, Streptomyces. Although SD8 blocks the action of bacterial DNA gyrase, which is also the target of fluoroquinolones such as ciprofloxacin and aminocoumarins such as novobiocin, it is mechanistically distinct from these older antibiotics. To understand why, Anthony Maxwell and colleagues at the John Innes Centre in Norwich, UK, solved the co-crystal structure of SD8 bound to GyrA. The structure reveals that the protein forms a dimer of dimers, with four molecules of SD8 bound to the four subunits of GyrA. Weirdly, each molecule of SD8 cross-links two separate subunits of GyrA, although mass-spectrometry, analytical ultracentrifugation, and modeling suggest that a single molecule of SD8 could also bind to a single GyrA subunit. The crystal structure shows that SD8 binds near – but not at – the fluoroquinolone binding site, blocking the DNA-binding portion of GyrA.


What caught my eye is the fact that both halves of the molecule are active by themselves, albeit with a loss in potency (see figure). The linker is over one nanometer long and doesn’t appear to make significant interactions with the protein; it would be fun to know how something like this evolved.

Antibiotics gleefully seem to ignore the Rule of 5, but it wouldn’t hurt to get a smaller, less complicated analog. Replacing either of the two ends with smaller fragments may be a productive approach, as would optimizing the individual “fragments” themselves.

Warren DeLano Memorial Award

We wrote last month about Warren DeLano, and in the December issue of Nature Structural and Molecular Biology Axel Brunger and Jim Wells have written a beautiful obituary.

Together with Warren’s family, Axel and Jim are also organizing a commemorative fund:

The Warren L. DeLano Memorial Award for Computational Biosciences

This award shall be given to a top computational bioscientist in recognition of the contributions made by Warren L. DeLano to creating powerful visualization tools for three dimensional structures and making them freely accessible. The award, accompanying lecture, and honorium will be given annually in the context of a national bioscience meeting or a Bay Area gathering of computational bioscientists at Stanford, UCSF or UC Berkeley. For the award special emphasis will be given for Open Source developments and service to the bioscience community.

For the award selection, a committee will be formed consisting of experts in the computational and biological sciences. Submission for nominations will be open to everybody.

Tax deductible donations can be made by check to the address below or by PayPal.

Silicon Valley Community Foundation
memo: Warren L. DeLano Memorial Fund
2440 West El Camino Real, Suite 300
Mountain View, CA 94040
tel: 650.450.5400

To endow this in perpetuity would require about $100,000, and the fund is off to a good start, with $23,000 contributed and another $30,000 pledged so far.

Enthalpy versus entropy

The earliest stages of lead discovery usually focus on obtaining a molecule with decent affinity for a given target. Affinity, or binding energy, can be dissected into two components: enthalpy and entropy. On a (very) simplistic level, enthalpic binding comes via specific molecular interactions, such as hydrogen bonds, while entropic binding results from nonspecific hydrophobic interactions. Optimizing enthalpy is usually more difficult than optimizing entropy: engineering a polar interaction requires more precision than adding a bit of grease. In a new paper in ChemMedChem, Andrew Scott and colleagues at Pfizer show how fragments that owe more of their binding affinity to enthalpy make better starting points for optimization than do fragments whose binding is more entropic, even if the entropic fragment is more potent.

The researchers used human carbonic anhydrase (hCA II), a venerable work-horse of biophysical studies. Benzenesulfonamide (compound 1, below) is a known binder, and the researchers studied the thermodynamics of 20 derivatives of this molecule using isothermal titration calorimetry (ITC), taking care to generate high-quality binding data. Adding a fluorine group to the 2-position of benzenesulfonamide (compound 2) improves the potency almost three-fold but lowers the ligand efficiency. In contrast, adding a fluorine to the 3-position (compound 3) improves the potency by seven-fold and also improves the ligand efficiency.



If you were choosing between these fragments solely on the basis of affinity or ligand efficiency, it would be reasonable to choose compound 3, and in fact a search of the literature turned up 15 carbonic anhydrase inhibitors that contained the 3-fluorobenzenesulfonamide substructure and none that contained the 2-flurobenzenesulfonamide substructure. However, a look at the thermodynamic parameters reveals that the affinity of compound 2 is driven by a sizable improvement in enthalpic binding, partially offset by lowered entropy. In contrast, compound 3 has a similar enthalpy of binding as compound 1 but increased entropy. What’s going on?

The researchers determined high-resolution crystal structures for all three of these molecules bound to hCA II. Interestingly, the structure of compound 2 shows a specific interaction between the fluorine atom and a main-chain NH of the protein. In compound 3, the fluorine points towards the hydrophobic wall of the protein.

Adding a 4-benzylamide substituent onto each of these molecules led to improvements in activity. However, this was a relatively modest boost for the more entropic compound 3 to compound 20, but considerably larger for the enthaplically-driven compound 2 to compound 19. Compound 19 shows a highly favorable binding enthalpy, and is the most potent and ligand-efficient of any of the three elaborated molecules.

Obtaining thermodynamic parameters for small-molecule protein interactions has historically been challenging, but in recent years miniaturization and improvements in technology have brought ITC into more non-specialist labs. If you have the resources, it may be worthwhile characterizing the thermodynamic profiles of your fragment hits, and – perhaps – looking more closely at those that show enthalpically-driven binding.

Too many aromatics stink

A recent discussion centered on whether fragment libraries should be designed to include more “3-dimensional” molecules and reduce the number of flat, aromatic compounds. Two new papers suggest that doing so may improve pharmaceutical properties. What effect this would have on screening success is still unclear.

The first paper, published by Timothy Ritchie and Simon Macdonald of GlaxoSmithKline in this month’s Drug Discovery Today, correlates the number of aromatic rings with several metrics associated with success in drug development. For this analysis, each ring in a fused system is counted separately, so indole is counted as having two aromatic rings. The researchers conclude that more than three aromatic rings correlates with an increased risk of compound attrition during drug development:

The fewer the number of aromatic rings contained in an oral drug candidate, the more developable that candidate is likely to be.

This is not surprising, but with their access to a vast internal data set the researchers provide considerable supporting evidence. For example, the mean aromatic ring count declines from 3.3 to 2.3 as GSK compounds move from preclinical candidate selection to proof-of-concept in humans. Measured (kinetic) solubility decreases dramatically with increasing ring count: even two aromatic rings leads to many low solubility compounds, and with four aromatic rings the median solubility is only 0.012 mg/ml. Both c log P and log D increase with increasing ring count, as do serum albumin binding, P450 3A4 inhibition, and hERG inhibition – all factors one usually wants to decrease in drug development.

One caveat is that the authors do not control for size. As aromatic rings are added, molecular weight is likely to increase, and thus many of the properties could simply reflect the pharmaceutical liabilities of larger molecules. This is where the second paper comes in. In J. Med. Chem., Frank Lovering and colleagues at Pfizer (nee Wyeth) analyze the effect of aromaticity itself by defining a simple metric:

Fsp3 = number of sp3 hybridized carbons / total carbon count

The smaller the number, the more aromatic the compound; the larger the number, the less aromatic. Besides being a straightforward measure of saturation, the formula inherently controls for molecular size.

When the researchers examined published data sets, they found that the mean Fsp3 increases from 0.36 for 2.2 million molecules in discovery to 0.47 for 1179 approved drugs. They also investigated measured solubility and found a strong correlation: 104 molecules with a log S of -6 (quite insoluble) had an average Fsp3 of 0.31, while 194 molecules with a log S of 0 (very soluble) had an average Fsp3 of 0.56. The effect is even more striking with melting points, which negatively correlate with solubility: 1153 molecules with a melting point of 125 deg. C had an average Fsp3 of 0.31, while 375 molecules with a melting point of 275 deg. C had an average Fsp3 of 0.18.

OK, so let’s say we accept the premise that increasing aromatic character in a molecule leads to lower solubility and worse properties overall. The easiest solution might be to reduce the number of aromatics in a screening collection, but would this really be wise? Ritchie and MacDonald note that aromatics, with their rigid structures, are likely to have increased potency relative to unsaturated molecules. And particularly for fragment libraries, you want all the binding energy you can get.

An interesting study would be to correlate the hit rate for fragments with their aromatic character. Does the hit rate increase with decreasing Fsp3? These data must exist in companies that have been doing FBDD for years. Indeed, at FBLD 2009, Ijen Chen of Vernalis presented a nice analysis of hits against 12 targets, in which she noted that roughly 2/3 of the fragment library members didn’t hit any of the targets. I don’t think she mentioned aromaticity specifically, but she did note that the hits tended to be slightly more rigid and hydrophobic than the non-hits – just what you would expect for low-Fsp3 molecules.

So by all means avoid having too many aromatics, but don’t go to extremes: it’s finding the right balance of binding energy and pharmaceutical properties that makes drug discovery such a tricky business.

NMR vs other methods

Fragment-based lead discovery owes much of its popularity to NMR: the SAR by NMR papers published by Abbott in the mid 1990s demonstrated both the power and the practicality of the approach. Recently SPR has also come into its own as a means for screening fragments, and in a paper in this month’s issue of Drug Discovery Today Claudio Dalvit of Novartis and the Italian Institute of Technology compares these two techniques, along with fluorescence spectroscopy. Not surprisingly given the author’s longstanding research interest, NMR comes out favorably, though with the recommendation that the techniques are complementary, so researchers should combine techniques rather than simply selecting one over another.

As I read the paper, I wondered why fluorine-labeled fragments are not used more widely; Dalvit’s group published another paper about this approach recently in JACS. Fluorine has a strong NMR signal and is very sensitive to the local environment, so when a fluorine-containing fragment binds to a protein this can be easily detected. In fact, the dynamic range for this type of assay is so great that fragment binding can be detected at concentrations several orders of magnitude lower than their dissociation binding constants.

This seems like a very powerful approach, but I haven’t seen many other people using it. Are folks concerned about the need for fluorine in every fragment (although many are commercially available) or is there something else I’m missing?

A tale of two deals

Two deals involving companies in the fragment space were announced today. I don’t have inside information on either of these, but superficially they are strikingly different.

In the first, Australia’s Biota has agreed to acquire UK-based Prolysis Limited, which previously published some nice work on using FBLD to discover new antibiotic leads (see here and here). The price? Just $10.8 million. However, Biota did say it plans to invest up to $25 million over the next three years on programs Prolysis started.

At the same time, UK-based Astex Therapeutics announced a new partnership with GlaxoSmithKline. The deal is for multiple targets in multiple therapeutic areas, with Astex focused on fragment screening and lead discovery and GSK focused on optimization of the resulting leads as well as preclinical and clinical development. The price? $33 million in up-front cash and equity, with a total potential of more than $500 million (BioBucks).

Astex of course is one of the few intact survivors of the first wave of fragment-based companies and has put several compounds into the clinic, including AT9283, AT7519, and others. It’s encouraging to see that deals of this size are still being done for what look to be fairly early stage collaborations.

Remembering Warren Delano

For those of you who haven’t heard, Warren Delano, author of PyMOL, died earlier this week. His sister Jen has established a blog for people to record their memories.

Many of you have used PyMOL for visualizing crystallographic and NMR structures. Even if you haven’t used the program directly, hardly a week goes by without papers appearing in Science, Nature, and other high-profile journals adorned with beautiful illustrations created with his software. Every protein-fragment structure I’ve looked at in detail has been through the lens of PyMOL.

Warren and I overlapped in Jim Wells’ group at Genentech, and Warren was one of the first people to join Sunesis, where he stayed until PyMOL became so successful that he left to devote himself to it full time. He was a brilliant programmer, a gifted scientist, and a valued friend.

Warren’s insights and creations illuminated macromolecular structures with the clarity of science and the grace of art. The world is darker without him.

To grow or to link: why not both?

What can you do with fragments? The idea of linking a couple together, while successfully demonstrated in the first SAR by NMR paper, generally seems to be more difficult than gradually growing one fragment. Now a new paper in Angewandte Chemie from Chris Abell and colleagues at the University of Cambridge presents a lovely comparison of these two strategies applied to a single target.

The researchers were interested in the M. tuberculosis enzyme pantothenate synthetase (PS) as a potential therapy for TB. Using a number of biophysical techniques including thermal shifts, NMR, and isothermal titration calorimetry, Abell and colleagues identified indole fragment 1 as a low-affinity binder from a library of about 1300 fragments (see figure below). X-ray crystallography revealed that the fragment binds in the ATP-binding site. An attempt to partially mimic the triphosphate by introducing negatively charged moieties led to modest improvements in potency (compounds 1a, 1b, and 2). Compound 2 bound in a similar position as compound 1, with the advantage that the methyl group off the sulfonamide is nicely positioned for further growing the molecule. Replacing this methyl group with a methylpyridine produced compound 4, increasing the affinity by about two orders of magnitude while maintaining ligand efficiency, and crystallography revealed that this moiety binds in the P2 pocket. Thus, the fragment growing approach began with an indole of low millimolar affinity and produced a molecule with low micromolar affinity after several iterations.



At the same time, the researchers also identified benzofuran fragment 5 (see figure below) and discovered that it binds in the P1 pocket some distance from the indole fragment 1, suggesting the two could be linked. In fact, a crystal structure revealed that the two fragments are able to bind to PS simultaneously. Linking these together through the acylsulfonamide linker employed above led to compound 8, with a potency similar to that obtained from fragment growing. Compounds 4 and 8 structurally resemble each other, but although the indole fragment of each binds in the same location, the terminal fragments (the methylpyridine in compound 4 and the benzofuran fragment in compound 8) bind in different locations, the former in the P2 pocket with the later in the P1 pocket. However, the benzofuran is somewhat twisted relative to the binding mode it adopts as a free fragment.



As the researchers observe, the ligand efficiency of compound 8 derived from fragment linking is lower than those derived from fragment growing, though even the molecules developed from growing have lower ligand efficiencies than the initial fragments.

The researchers conclude:

The two strategies resulted in similar compounds with similar potencies. This outcome obscures the fact that although the linking strategy appears more elegant, the limited repertoire of linkers is likely to compromise the binding of the original fragments. In comparison, the fragment-growing strategy provides more freedom for development at each stage and allows more room for further optimization.

True. But, the fragment linking strategy does provide a clear starting point for further optimization. The researchers did not describe how they selected the methylpyridyl fragment in compound 4 or how many other moieties they tested; 5-methylpyridine-2-sulfonamide does not seem like the first reagent one would grab from the shelf. However, the methylpyridine fragment is not dissimilar to the benzofuran fragment: swap the (hydrogen-bond accepting) oxygen for the (hydrogen-bond accepting) nitrogen, and the methyl would sit in a similar position as the phenyl ring (see figure above). In other words, medicinal chemistry on compound 8 could lead quite naturally to compound 4.

Infarmatik In-3D Library

In what we hope is a new series bringing the latest in Fragment Science up for discussion, we present today to you a discussion of Infarmatik's In-3D Library. We look forward to this discussion, and hopefully, many more.

Fragment based drug discovery has been shown to provide a rapid means for transforming low affinity “hits” to optimized leads. However, most currently available fragment libraries are limited in usefulness, mainly because over 90% of the molecules are planar and thus do not fit well into 3-dimensional receptor protein binding sites. InFarmatik realized [Ed: and others] that “real 3-D” structures offer a better fit within the uneven binding surfaces of protein hot-spots (business sites) than do planar compounds. To address this issue, we have developed a specific series of novel and diverse 3-D fragments, which are not available from any other commercial sources. The structure types of the first release contain 2,3, and 4 member non-aromatic ring systems, with various attachment points, including spiro and 1,2 anellation, 10 electron systems connected to saturated ring systems, saturated bis-heterocyclics and rod shaped compounds. We believe these compounds will exhibit the ability to bind to a wide array of protein targets. In addition, we can offer another 435 structures from existing stock, which conform to Ro3 and are quite “fragment-like”.

Most of the compounds have soft scaffold structures: meaning they were designed to have low reactivity centers to avoid non-specific binding, while preserving the ease of chemically coupling them to each other or to other fragments. The attachment points in the molecules in many cases are useful for regiospecific reactions.

Here are the relevant properties of the 3-D fragment library:
Size: 119 3-D Fragments
Average MW=230 Da
average logP value (calculated) =1.88
confirmed minimum water solubility of at least 0.1% in 2% aqueous DMSO.
Solubility data available for all compounds
Highly diverse, as shown by 3-D Diversity Analysis using ChemAxon supplied tools
Here are the relevant properties of the new standard fragment set
· Size: 435 Fragments
· Average MW=237.8 Da
· Average LogP value (calculated) =2.27

Genentech’s affinity for Graffinity

Heidelberg-based Graffinity today announced that they would be collaborating with the Genentech division of Roche. Graffinity will apply its surface plasmon resonance (SPR) fragment-based technology to several Genentech targets. Financial details and specific targets have not been released, though Graffinity CEO Kristina Schmidt is quoted as saying that they plan “to explore drug targets that would remain white spaces on the map of drug discovery” with conventional high-throughput screening.

SPR is rapidly becoming a workhorse in the stable of FBDD techniques. Although it provides less information than NMR or X-ray approaches, SPR is faster, and can rapidly distinguish true hits from bad-acting artifacts. Typically a protein is immobilized on a gold surface, and fragments are allowed to flow past to detect those that bind. Graffinity reverses this process: they have a collection of about 110,000 small molecules, just over a fifth of which are fragments, immobilized in microarrays which can be screened against proteins (see here for full description).

Genentech is no stranger to SPR; one of the highlights of the recent FBLD 2009 meeting was a talk by Tony Giannetti on the use of this technology at Genentech against roughly 40 target proteins. The collaboration further validates the use of SPR for FBDD, and suggests that Graffinity has an interesting – and useful – angle.

Fragment-based events in 2009 and 2010 (and calls for abstracts)

We’re in the last quarter of 2009, and I know of just one more event this year involving fragments:

October 13: The Life Science Regional Technology Symposium will be held in Somerset, NJ, and Dr. Teddy Z. will be one of several excellent speakers.

2010

Next year is starting to take shape nicely, and two events have put out calls for abstracts, so if you have something interesting to present, now’s your chance!

February 3-5: Cambridge Healthtech Institute’s 17th International Molecular Medicine Tri-Conference will be held in my beautiful city of San Francisco, with a track on medicinal chemistry that will have some fragment talks, and a short course on “Fragment-Inspired Medicinal Chemistry” on February 2.

March 21-25: The spring ACS meeting will also be held in San Francisco. There will be a symposium on “Fragment Based Drug Design: Novel Approaches and Success Stories,” and Rachelle Bienstock at the FBDD LinkedIn site has put out a call for abstracts, due October 19.

April 20-25: The Keystone Symposium on computer-aided drug design will take place in brisk Whistler, British Columbia. Although not exclusively devoted to fragments, the schedule shows several talks on the topic.

April 27-28: Cambridge Healthtech Institute’s Fifth Annual Fragment-Based Drug Discovery will be held in summery San Diego. This conference has also put out a call for speakers, with a deadline of October 16.

Know of anything else? Organizing a fragment event? Let us know and we’ll get the word out.

Looks can be deceiving: Getting misled by crystal structures - part 2

Last year we highlighted a paper that touched on some of the ways crystal structures can mislead, and a theme of FBLD 2009 was how dubious data can derail modeling efforts. Now, Jens Erik Nielsen and colleagues at University College Dublin add to the discussion by showing how the crystal lattice can potentially distort protein-ligand interactions. Their paper in J. Med. Chem. provides an analysis of the prevalence of two common structural artifacts, plus a practical tool for detecting them.

The first problem the authors consider is that some ligands make “crystal contacts.” Because a crystal is made up of a three-dimensional lattice of proteins packed together, a ligand bound near the surface of one protein may be in close contact with another protein in the crystal (a nonbiological “symmetry mate”); this contact occurs only in the context of a crystal and could distort how the ligand binds to its (true) partner protein.

The second, related problem is that water molecules that appear in the crystal structure can form bridges between a ligand and its nonbiological symmetry mate.

The authors examined a set of 1300 protein-ligand crystal structures with noncovalently bound ligands and experimentally measured binding affinities (PDBbind Database). Of these, 36% of ligands showed crystal contacts, and a similar number (37%) had crystal-related water bridges.

This doesn’t mean that all of these structures are misleading: the researchers note that “it is entirely possible that crystal contacts in some cases do not perturb the geometry of a protein-ligand complex whatsoever.” However, removing these structures before running docking experiments did improve the results.

The tricky thing about these structural artifacts is that they are often invisible, even when suspected. Most non-crystallographers focus on just on a single protein-ligand complex and don’t consider the crystal lattice when examining a crystal structure. Happily, Nielsen and colleagues have constructed a simple online tool (LIGCRYST) that can evaluate structures from the pdb to search for these types of problems. Although I’m not a crystallographer, I found it quite easy to use.

Hopefully modelers will increasingly take crystal contacts into account, and the next time you examine a structure from the pdb, you may want to give it a quick run through LIGCRYST.

FBLD 2009

Fragment-based Lead Discovery Conference 2009 just concluded in York, UK; it was the second in what will hopefully be a continuing series. With more than two dozen talks and as many posters spread over three days, most of them very high quality, it is impossible to summarize even the highlights (and I don’t want to scoop pending publications). Instead I’ll just jot down a few impressions.

On the broad topic of why FBLD is useful, an interesting shift in emphasis seems to have occurred. A few years ago a key argument in favor of fragments was getting compounds to the clinic faster, but there is now a greater focus on quality over speed. In summarizing over a decade of fragment work at Abbott, Phil Hajduk noted that FBLD hits consistently bind more efficiently than those from HTS. Similarly, Chris Murray of Astex noted that, among their five clinical candidates (four of which target kinases), the average ClogP was 1.7 (vs 4.1 for a set of 45 reported orally active kinase inhibitors), while the average molecular weight was 390 (vs 457).

One theme that differentiated this meeting from others was a strong focus on modeling: an entire day was devoted to sessions on “fragments, scoring functions and docking” and “design from fragments.” This concluded with a lively round table discussion, chaired by Vernalis’ James Davidson, titled “Chemistry challenging modeling.” But challenges didn’t only come from chemists: one prominent modeler noted that there have been no fundamentally new approaches to modeling in the past two decades; another asked why, despite the number of interesting new chemistries out there, so many modelers restrict themselves to the same old standbys such as amide bonds.

Part of the problem with modeling, of course, is separating hits from noise: true hits often show up near – but not at – the top of a ranked list, so how does one decide what is worth pursuing? Phil Hajduk discussed the use of “Belief Theory”, in which the similarity of an unknown molecule to a known active is used to evaluate the unknown.

Another problem is the quality of primary data: As Hajduk noted, “no one takes experimental error into account” when predicting ligand binding, and a recent analysis suggests that over-fitting data is a substantial problem with many computational approaches. This is all the more problematic when the data are not just noisy but spurious; Practical Fragments has noted the problem of aggregation, and UCSF’s Brian Shoichet emphasized this point, noting that 85-95% of hits from a high-throughput screen could be artifacts, while 85-100% of what remains could also be bogus. He did note, though, that fragments are less problematic in this regard than larger molecules. And Genentech’s Tony Giannetti, Vernalis’ James Murray, and others illustrated how surface plasmon resonance is effective at weeding out bad actors.

Getting better data will clearly be essential to getting better models, but one essential category, the forces involved in protein-small molecule interactions, is still poorly understood. Gerhard Klebe of the University of Marburg presented a detailed and elegant set of experiments exploring the effects of chemical structure on the enthalpy and entropy of binding to the protein thrombin. He emphasized that desolvation of fragments from water is critical, and only possible if compensated by strong interactions with the protein. This also implies that you want fragments that have low desolvation penalties as well as high solubilities – a tricky balancing act.

FBLD 2009 was held barely six months after Fragments 2009, and it is a testament to the vibrancy of the field that both conferences managed to be so successful and exciting while sharing very few speakers in common.

For the other two hundred plus attendees at the conference, what were some of your impressions?

Upcoming Fragment Talks

There is an upcoming conference with an extraordinary FBDD lineup. :-)
Don Huddler from GSK will be talking about SPR in fragment screening.
Bill Metzler from BMS will be talking about the uses of biophysical methods and structural information for hit prioritization.
I will be talking about how to put together an integrated FBDD paradigm.
There is one more talk of the TBD variety, but I think it will be a very nice complement to these other three.
Please come out and see what the state of the art is.

Who’s doing FBDD?

Lots of companies are using FBDD, but aside from big pharma it’s not always easy to find them. As a public service we have started a running list with live links. This first installment is taken largely from a nice review by Wendy Warr in the JCAMD special issue we highlighted; we’ve removed companies that have been bought or ceased working in FBDD.

Astex Therapeutics
Beactica
BioLeap
BioSolveIT
Carmot Therapeutics
Crystax Pharmaceuticals
deCODE Chemistry and Biostructures
Evotec
Graffinity Pharmaceuticals
IOTA Pharmaceuticals
Locus Pharmaceuticals
MEDIT
Plexxikon
Proteros Fragments
Pyxis Discovery
Structure Based Design
Vernalis
Zenobia Therapeutics
ZoBio

I’m sure there are plenty of omissions; put them in the comments and we’ll add them in the next update.

BioLeap leaps into collaborations

Pennsylvania-based BioLeap, which uses computational FBDD, has just signed a deal with GlaxoSmithKline to work on “difficult” targets. The announcement came September 8, just a month after BioLeap started a collaboration with Lycera on autoimmune disorders. I haven’t personally seen any talks or papers out of BioLeap, but there have certainly been plenty of improvements in computational chemistry applied to FBDD recently (see here, here, and here), and given the lag between discovery and disclosure there are likely many new developments.

This is also the second fragment deal that GSK has done in the past month; we already noted their collaboration with Vernalis.

What do you think? Does this flurry of new deals signify increasing use of FBDD?

Journal of Computer-Aided Molecular Design Special FBDD Issue

Our friends over at FBDD-Literature have already highlighted this, but it bears repeating that the entire August issue of J. Comp. Aid. Mol. Des. is devoted to FBDD. For aficionados of all things silicon, there are articles on computational chemistry applied to FBDD generally as well as on more specific topics such as MCSS, NovoBench, FTMap, and two papers on Glide (here and here).

But don’t be put off by the name of the journal: with 14 articles covering close to 200 pages, there is something here for almost everyone, even for those whose interest in computers ends at using them to read this blog! A brief editorial outlines the challenges of FBDD, and a longer introductory piece gives an overview of the field. Several articles focus largely on specific targets such as p38alpha, heparanase, and Eg5, while one is devoted to assessing druggability.

Finally, two articles address the important topic of designing fragment libraries, one from the perspective of big pharma (nicely summarized here), the other from biotech.

Destructible ligands

Crystallography-based methods of fragment screening often rely on growing many crystals of a protein and soaking these in fragment-containing buffers. But how do you get biologically relevant crystals in the first place? Many proteins adopt a variety of different conformations in solution, and their freedom of movement is constrained once they are forced into a crystal lattice. Crystallizing the protein in a state that is relevant for binding ligands often means co-crystallizing them in the presence of a known ligand. In fact, some proteins are so disordered on their own that the only way you can get them to crystallize at all is by adding a small molecule. In many cases, these “co-crystals” can then be soaked in a solution containing new ligands; the existing ligands will diffuse out of the crystal, making room for new ligands. Unfortunately, in some cases the original molecule binds so tightly that it can’t be forced out. Two recent papers in J. Am. Chem. Soc. provide a clever solution.

Both papers focus on the major histocompatibility complex (MHC) Class I proteins. These proteins bind 8-11 amino acid intracellular peptides and present them on the cell surface, allowing passing T cells to survey the contents of cells for viruses, bacteria, or other nasties and, when appropriate, eliminate the infected cells. As might be expected given their function, the MHC proteins are quite promiscuous in which peptides they bind to, frustrating a general understanding of the molecular recognition. Moreover, crystallography is complicated by the fact that MHC class I proteins do not crystallize in the absence of a bound ligand.

In the first paper, Anastassis Perrakis, Ton Schumacher, and colleagues at the Netherlands Cancer Institute designed a 9-amino acid "conditional" peptide ligand for MHC that contains two internal photosensitive nitrophenyl substituents. They were able to crystallize this in complex with MHC and solve the structure. When they exposed these crystals to UV-light, the nitrophenyl groups caused the peptide to break apart into into three pieces. Interestingly, structural characterization after this exposure revealed that while the central portion of the peptide was gone, the two end bits were still bound to MHC. However, the researchers were able to successfully replace these remnants with new, full length peptides derived from HIV and avian flu proteins by soaking the crystals for just a few hours in buffer containing the new peptides. The resulting structures were identical with previously determined structures, even revealing some side-chain movement. A second paper from Ton Schumacher, Huib Ovaa, and colleagues reports a similar strategy, this time using diol-containing peptides and mild chemical cleavage with sodium periodate rather than UV-light, although in this case the reaction is done in solution rather than in crystals.

This seems like an interesting approach for tackling peptide-binding proteins, and possibly even small-molecule binding proteins, though this would require more effort to design destructible ligands.