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.