08 October 2008

Bacterial Paclitaxel from Fragments

The rise of new bacterial strains resistant to existing antibiotics has long been anticipated, but a lack of investor interest means few companies are doing anything about it. A nice exception is described in a recent paper in Science from UK-based Prolysis and collaborators at several other institutions. And they used fragment-based drug discovery.

The researchers started with 3-methoxybenzamide, which was shown several years ago to inhibit the bacterial protein FtsZ and inhibit cell division in Bacillus subtilis, leading ultimately to cell lysis. FtsZ is a distant relative of the mammalian protein beta tubulin, a target for such famous anticancer drugs as paclitaxel.

The paper does not discuss the details of the fragment-based methodology used; hopefully this will be published separately. What we do know is that after some 500 analogs the researchers arrived at PC190723, which is roughly two thousand-fold more effective against B. subtilis as well as a host of staphylococci, including such nasties as multi-drug-resistant Staphylococcus aureus (MDRSA). Molecular modeling and mutagenesis data suggest that the compound binds to the region of FtsZ corresponding to the taxane-binding region of tubulin. The compound was also able to completely protect mice from an otherwise lethal challenge of S. aureus.

This is a nice story on several levels, as befits its publication in Science. First, it describes a new lead series against deadly bacteria. Second, it validates a new drug target and provides a new tool compound for exploring microbiology. And finally, it again demonstrates the power of fragment-based drug discovery to generate useful, novel molecules. Indeed, perhaps the fact that the fragment-origins of PC190723 are merely a side note as opposed to a central focus of the publication shows that the technique has become main-stream.

1 comment:

Byron DeLaBarre said...

I noticed they went after a high resolution apo form of FtsZ but had to model in the fragment. I wonder if they got the structure of the fragment or final 1 uM inhibitor in and simply didn't want to publish that experimental result or was the liganded complex not amenable to crystallization? They must have tried to get the complexes - there were already multiple entries for FtsZ in the pdb.