By Kevin Davies
August 18, 2004 | According to a 1996 back-of-the-envelope calculation by computational chemist Regine Bohacek and colleagues, the total universe of small-molecule compounds — 30 atoms or less of carbon, oxygen, nitrogen, and sulfur — is estimated to be a staggering 1060. The latest version of the Chemical Abstracts Service registry lists more than 23 million organic and inorganic compounds in existence. The Beilstein database, a compendium of organic chemicals going back to 1771, includes records of some 9.3 million compounds.
Why then, given such seemingly vast chemical diversity, do drug companies so often struggle to find just one molecule to specifically bind a given target (of which there may be only a few thousand of therapeutic relevance)? When it comes to screening chemical compounds, quantity is a poor substitute for quality. (Of course, there are exceptions: At Oxford University, Graham Richards and his colleagues' highly publicized distributed computing projects produced a virtual library of 3.5 billion molecules by concocting 100 modifications for each of 35 million compounds.) Problems include the instability of many aging compound collections, the mediocre quality of some commercial offerings, and the inherent lack of diversity in most collections.
At the 2004 Beilstein Bozen Workshop*, held in the Dolomites of northern Italy, University of Cambridge chemistry professor Steven Ley derided as "absolutely pathetic" the standard practice of building small-molecule libraries merely by tinkering with a fixed chemical scaffold. Why not change the core itself? he asked. He also characterized as "rubbish" the fact that most of the major drugs on the market require up to 12 synthetic steps. Such convoluted syntheses are required in part because of patent restrictions, which prevent chemists from using many classic synthetic reactions.
Taking a dynamic approach to chemical synthesis, sometimes called target-oriented synthesis, Ley's group has created the equivalent of an Aldrich catalog of compounds in a matter of months, using suites of immobilized reagents — even reagent stirrer bars! A growing trend is the automated procedures to assess the literature when planning routes to synthesize a particular compound. The impact of these changes, Ley said, is that what might have been a Ph.D. project 10 years ago can now be completed in six weeks.
Friedrich Beilstein: encyclopedist extraordinaire
A different solution to this problem is championed by Harvard University's Stuart Schreiber, who advocates diversity-oriented synthesis, or DOS (see Feb. 2003 Bio·IT World
, page 48). In a paper last year, he wrote that DOS is about "populating chemical space efficiently with small molecules that have complex and diverse molecular skeletons." His first effort produced 1,200 compounds with diverse skeletons synthesized in a combinatorial process.Better Targeting Through Chemistry
Back at the Beilstein workshop, the view from pharma was no different: Keith Russell, AstraZeneca's head of CNS chemistry, noted that "since 1980, new medical entities (NMEs) are flat-lining despite increased R&D costs." Only 1 in 100 projects results in an NME — roughly 80 percent of projects focus on the wrong target, whereas 95 percent look at the wrong compound.
To improve small-molecule libraries, Russell's team turns to computational programs such as the SynGen program for route generation, developed by James Hendrickson at Brandeis University, which generates the most economic route for a particular organic synthesis project, often devising novel chemical transformations. The workstation where all of this advance planning takes place happens to be "the only Mac in AstraZeneca," Russell joked.
The biennial Beilstein workshop takes an admirably interdisciplinary approach to "address contemporary issues in the chemical and related sciences." This intimate workshop tackled many other related topics, including chemical biology, enzymology, molecular simulation, and cheminformatics. (The proceedings of the 2002 workshop are available at www.beilstein-institut.de/bozen2002/proceedings.) Given the surging interest in this field, there will be much to discuss when the Beilstein Institute hosts its next gathering in 2006.
* Beilstein Bozen Workshop: "The Chemical Theatre of Biological Systems"; Bozen, Italy; May 24-28, 2004.
PHOTO of Beilstein by Science Photo Library