Mapping the protein world

In the early days of X-ray crystallography obtaining a three-dimensional model of a protein required wire models, screws, bolts and years of tedious calculations by hand. Today macromolecular models are built by computers – thanks to sophisticated software and in particular a package called ARP/wARP. Developed by Victor Lamzin at the Hamburg Outstation of the European Molecular Biology Laboratory (EMBL) and Anastassis Perrakis at the Netherlands Cancer Institute (NKI) in Amsterdam, ARP/wARP is currently used by over 2,000 researchers throughout the world. The capabilities of this software will now expand even further - thanks to a grant of over 800,000 US Dollars from the U.S. National Institutes of Health (NIH).

The grant, which will run over four years, comes at a perfect time. "More than 1,000 research laboratories from over 50 countries are holding ARP/wARP licenses and by June this year our paper that described the key innovative feature of ARP/wARP in Nature Structural Biology in 1999 has reached the magic number of 1,000 citations in the scientific literature. This has created an incredible drive for further scientific development," Lamzin says. "The new funding gives us a push to advance the software's ability to recognise and distinguish different types of macromolecular objects, for example DNA, and to improve the automated generation of structural models. ARP/wARP has made the life of structural biologists worldwide a lot easier and will do even more so once the new features planned under the NIH grant have been implemented."

ARP/wARP transforms 'electron density maps', produced in experiments that bombard protein crystals with X-rays, into 3-dimensional structures. "X-ray experiments result in 'diffraction patterns' that can't be interpreted using our eyes," Lamzin says. "These have to be reconstructed into a three-dimensional image through mathematics and models. This was a very tedious, time-consuming, and subjective process."

ARP/wARP was the first, and for a while the only, software that could generate models to fit experimental data automatically and very accurately. It has cut down the time necessary to create structural models from weeks to minutes.

This grant will allow the scientists to explore new concepts of model-building and enlarge the scope of data that the software can handle. ARP/wARP deals very well with high-resolution data that allows to distinguish individual atoms, but much of the data that scientists have to deal with is of lower quality. The software has steadily been improved to work with lower-resolution data, and Lamzin and Perrakis know how to stretch it even further.

"The high-throughput revolution in Structural Biology allows us to work on more and more complex problems relevant to human health," Perrakis says. "Knowing the structures of molecules that play crucial roles in cancer, cardiovascular and neurodegenerative diseases and molecules from pathogenic bacteria or viruses will contribute to design new revolutionary therapeutic strategies."

To meet this objective the scientists intend to study crystals of proteins bound to diverse drug candidates or containing different types of large molecules.

"ARP/wARP needs to meet a two fold challenge: firstly, it needs to be able to work with structural information at lower resolution, within the range of 3.0 to 3.5 Ångstroms, and secondly, the models produced have to be complete and validated. The new NIH grant will help us to approach these aims. In the future researchers will be able to focus on structure analysis rather than just building the structure and, who knows, by combining ARP/wARP with new cell imaging techniques we might be able to model the molecules of a complete cell," Perrakis concludes.

Source : European Molecular Biology Laboratory