Tiny big problems

FOR Most of us, an antibiotic is a drug used to fight infections. But very few of us know about the hard work that has gone into making these little pills that boost out immune system and help fight germs and viruses. Antibiotics are essentially very complex molecules. Sir Alexander Flemming is credited with formulating the very first antibiotic, penicillin, in 1928. His discovery was a result of a lucky accident. Flemming noticed that bacteria simply would not grow on a petridish contaminated with Penicillium notatum, a mould. A few more experiments and penicillin was born. Our present-day scientists, however, cannot leave anything to chance and they, unlike Flemming, cannot wait for such lucky accidents to happen.

The biggest obstacle these scientists face is determining the three-dimensional structure of these large and complex molecules. The 'ball-and- stick' models we saw in our chemistry textbooks are okay for simpler molecules like water and carbon dioxide. But for molecules with 1,000 or more atoms, the number of arrangements possible is astronomical and it could take months before the researchers, frantically examining the reaction pathways and chemical properties of these molecules, hit upon the precise structure.

This is where the recent work of Nobel laureate H Hauptman and his team come in. They have devised a new algorithm called 'Shake- and-Bake' which claims to be able to determine the structure of molecules with a thousand atoms in a fraction of Lbe time taken by direct methods.

Hauptman, a mathematician who shared the Nobel Prize in Chemistry in 1986, is the president of a medical research institute at Buffalo, New York. The essence of the new algorithm is very much like the direct method of guessing the structure by using an assumed arrangements of atoms. But here, an arrangement is tried and if the phases seem close to the observed pattern, that guessed arrangement is frozen-in or 'baked'. This alternate shaking and baking leads them quickly to the true structure. Exactly why the method works so well is not clear even to its originators. But it certainly works and several researchers have used it to determine the structure of several proteins and antibiotics: feats which would not have been possible before. The only limitation in using the method is that it requires extremely powerful x-ray sources. This is because the pictures of the pattern should be of very high quality. Another problem is that the method can only work with compounds which can be easily made into high quality crystal. This leaves several key organic molecules.

Nevertheless, researchers are confident that even with the limitations, they will be able to map the 3-o structure of many biological molecules whose structure have not been determined (Science Vol 282, No 2136).

Other popular methods of determining molecular structures include x-ray crystallography which gives accurate structural information for molecules that can be obtained in solid, crystalline form. The basic principle of the method is simple. Energetic x-rays are used to take 'snapshots' of a crystal. The atoms in the crystal scatter the X-rays, producing a pattern of dots on photographic film. The intensity and location of these dots give us information regarding the arrangement of the atoms in the molecule.

Rebuilding the molecule form the obtained pattern, however, is a long and tedious process because the pattern lacks a very important ingredient: phase information. To the uninitiated, phase information is the relative positions of the crests and troughs of the x-rays. To get this information, heavy atoms are implanted into the crystal, changing the pattern of the dots. Then, using these atoms as reference points scientists can deduce the structure of the crystal. The only weakness of this method is that it is extremely time-consuming.

Another technique used to determine crystal structure is basically a computer-intensive process where different arrangements of atoms are tried and the results of an x-ray diffraction simulated on fast computers. Then the arrangement which matches the observed pattern is chosen as the one which is closest to the actual crystal structure.