08 October 2014

PPI Fragment Libraries...what do YOU think they should look like?

Dan and I are at the CHI Discovery on Target meeting in Boston.  We taught our award-winning course on FBDD yesterday, but with an emphasis on PPI.  And just in time, Andrew Turnbull, Susan Boyd, and Bjorn Walse  published a paper on PPIs and FBDD (it's open access[Ed:Link fixed]).  The paper is a nice review of PPIs in general (which have already been covered here): structural characteristics, the "hot spot", and computational approaches.  What I thought interesting was their discussion of the physico-chemical properties of PPI fragments.  This is an area were this is a lot of common knowledge, but nothing rigorously studied.  So, into the breach.

They discuss vendor supplied PPI libraries: Asinex, Otava (which does not seem to be a fragment library with a mean MW of~500),  and Life Chemicals.  PPI compounds are thought to need to be more 3D and obey the rule of 4: PPI compounds will tend to be larger and more lipophilic.  Does ontogeny recapitulate philogeny?  To explore this, they looked at 100 fragments orthosterically active against PPI targets (unpublished data) and compared to 100 fragments active against non-PPI targets. 
It appears that PPI fragments are a little larger and more lipophilic than "standard" fragments, but NOT any more 3D.  It should also be noted that the PPI fragments had double the acid and base containing fragments than "standard" fragments.  The authors agree that their dataset is small, and other groups are looking at larger datasets.  But, the conclusion they draw is that PPI fragments should be larger, more lipophilic, and contain at least one polar moiety. 


Anonymous said...

Hi Edward,
I would like to notice you that "PPIs and FBDD" link to a paper is pointing to your FBDD course. And also I would like to ask you to relink it.
Thank you in advance


P.S. thank you (and Dan) for blogging.

Anonymous said...

That should be the correct link:


Dr. Teddy Z said...

Link fixed. Thank you.

Dan Erlanson said...

The researchers suspected that PPI-binding fragments would be larger and more lipophilic. They examined 100 PPI fragments and 100 fragments that didn't hit PPIs, and that's exactly what they found. However, it is not clear how these fragments were chosen, so I worry about the potential for selection bias.

Also, with many fragments in the 350-400 amu range and clogP > 3, they are departing somewhat from what most people consider fragments. Indeed, I don't believe any of the fragments shown in their table approach this size, which again makes me concerned about how they chose their 100 PPI fragments.

Rod Hubbard presented a larger, unbiased analysis of Vernalis fragment hits earlier this year in which he showed that those that hit PPIs did seem to be slightly larger and more lipophilic, but hardly to the extremes reported here, and still well within standard "fragment space."

Angelo said...

We also were curious about the property-space of the PPI fragments in the paper. We decided to have a look at two fragments sets we have: one is a non-specific set (e.g. not PPI-focused hits) and the other one was extracted from the TIMBAL database which lists “all” known PPI inhibitors. To define the fragment-space we filtered both sets using only Heavy Atom Count of 8 to 20 and our SMARTS set for unwanted functionalities. We ended up with two “med-chem friendly” sets (one generic, one PPI focused) of 318 fragments hits each (by chance). Then, we analysed the property space of the two sets.
PPI fragments hits seemed to be larger and more lipophilic but 189 out of 318 fragments in our PPI-focused set are hits from just 3 proteins targets and I am not sure how the set in the paper could be that different from ours in terms of number of targets covered. It would be nice to know. It is smaller, 100 vs 318. I still think that there are not statically relevant data at the moment to clearly define a property space for fragments targeting PPIs.

Dr. Teddy Z said...

This is terrific. Thanks for sharing.


Angelo said...

Thanks Teddy,

When I said "We" I meant Drug Discovery at the CRUK Beatson Institute in Glasgow (UK).

I point I would like to make about 3D fragments. When we decided to increase the 3D content of our fragment library (almost 4 years ago now) the main reason was to add another layer of diversity on top of the “canonical” ones (e.g. chemical, pharmacophore,….) which, in our view, could not do any harm and could possibly improve the library properties (for instance, the number of aromatic rings seems to be inversely correlated with drug developability). Then, if these more “shapy” fragments also prove more useful for probing new/challenging targets (they do explore a larger interaction space presenting different vectors) it will be a plus but at the moment it is too early to say so.

Chris Murray said...

My concern is that a lot of what are purported to be PPI fragments are just too big. If PPI targets are hard, why do we think it's a viable strategy to screen "lead-like" molecules as if they're fragments. Surely we expect that we are not going to be able to effectively sample chemical space with a small number of fragments of this size?

It's interesting that the data in this paper suggests that there is little evidence that more 3-dimensionality is required in fragment libraries targeted at PPI targets. This ties in with our experience at Astex (which has been discussed in recent FBLD talks and also reviewed recently in Progress in BioPhys and Molecular Biology). We too find little evidence that current fragment libraries have an inappropriate balance of 2D versus3D fragments. We have also found that 3D fragments give lower hit rates than 2D fragments and have suggested this is because they are more complex (in the sense that increasing complexity implies the possession of more features that can produce mismatches with a protein binding site). The data suggest that generally 3D fragments need to be smaller than 2D fragments (by one or two heavy atoms) if they are to retain the same sampling advantage. There is surely a need for new 2D and 3D fragments to be synthesised but I'd like library suppliers and academics to focus on fragments that are small enough to sample chemical space effectively.


Angelo said...

We measured the complexity of our fragments and we found no clear correlation between complexity and 3dimensionality. We have our own method to measure complexity which is independent from molecular weight (most of the other methods are) and is conceptually closer to the interaction complexity introduced by Hann. So, yes we did monitor it.

Angelo P