HYPOTHESES

REDUCTIONISM is one of those important ideas that some people treat as if it had bad breath and were to be avoided. Halitosis or not, I am an unrepentant reductionist biologist. Reductionism's apparent bad odour comes from its perception by some as an attempt to take the "soul" out of life, to reduce human behaviour to just the complicated dance of molecules. This is an old issue and was partly the inspiration for Mary Shelley's Frankenstein; she was influenced by "blasphemous" materialist views suggesting the brain alone was the basis of mind. This is a view modern molecular biology has more than vindicated; the life of the cell, the fundamental unit of life, is just complicated chemistry. Vitalism, the idea that there is some indefinable something distinguishing animate from inanimat, is dead - though there are those who would insert a vitalist element into consciousness.

The development of human beings, flies, elephants and plants from a single cell, the fertilised egg, ranks as one of the most beautiful and wonderful processes in biology. Only the evolution of the cell itself - the true "miracle" of evolution - and the workings of the brain can compete . For 20 years we have been working on one aspect of embryonic growth, that of the wing of the chicken. It is similar to the human arm, though it has only three fingers, but the advantage of studying the development of the chicken wing is that one can observe the process by opening eggs.

How should we think about such a marvellous process as the generation of the arm and the five fingers of the hand? The early limb bud grows out as a paddle-shaped protrusion which is made up of an outer sheet of cells, containing and restraining a mass of dividing cells on the inside - cells that will give rise to all the cartilage, bone, muscles and tendons of the limb. How do the cells know to make these structures? There is no doubt the process is controlled by genes, but how?

Elegance matters in science. I have always found it hard to believe that evolution would not have found an elegant solution for translating genetic information in the eggs into patterns and forms during embryonic development. Moreover, I have the strong suspicion that evolution is rather lazy, and that having found ways of doing it, will use the same tools again and again. This implies that there ought to be universal mechanisms for patterning the embryo, and this turns out to be clearly the case. Our own development is more like that of flies than even I could have hoped for.

A way of thinking about the limb and patterning of cells generally is in terms of the cells acquiring positonal information - so that they know, as it were, where they are in the embryo. Then if every cell has the same set of instructions as to what to do in every position, the embryo could generate an enormous number of different patterns. It is rather like the patterns people make at the start of, for example, the Olympic Games by holding up coloured cards. This seems to be a mechanism used quite widely by different embryos. The way the cells learn their position requires signalling from the cells to which they relate their position. One way of doing this is for the signalling cells to make a molecule that diffuses away so that other cells know that, the lower the concentration, the further they are away from the signallers.

Some 30 years ago a group of cells at the posterior margin of the bud (the little finger side) were identified which provide a signal that can tell the cells just how far they are from the posterior margin, and so which digit to make. Close to the signalling region a little finger would develop, far away a thumb. The problem, then, was the nature of the signal.

We thought we had the answer when we discovered that retinoic acid, a derivative of vitamin A, could mimic the natural signal in the limb. We made this discovery on the lottery ticket principle - if anyone had a "feeling" as to what the signal was they should take a chance and try the experiment. We were a bit lucky. But then came more experiments which cast doubt on retinoic acid being the signal.

The animal whose development we understand best is the fruit fly Drosophila, and the 1995 Nobel Prize in Medicine was awarded for work on identifying the genes that control its development. It was with enormous excitement that we learned that similar genes were active in similar ways in vertebrates, including humans.

Somewhat neglected was the "hedgehog" gene which is involved in patterning insect segments by coding for a protein that provides a key signal; flies with a mutant gene have an abnormal hedgehog-like pattern of denticles on the surface of the larva. Over the last year, hedgehog has become much loved. For it turns out that the hedgehog-like vertebrate gene, playfully called "sonic hedgehog", provides the key positional signal for digit identity. It also provides the signal for patterning parts of the early embryo, the neurons in the spinal cord and the brain, and probably for the embryo being able to tell its left from its right. It also plays many roles in the developed fruit fly, where it patterns both legs and wings. When evolution finds a positional signal, it sticks to it.

The language of cells is molecules. Embryonic development can be understood largely in terms of signals between cells switching on genes with consequent changes in cell behaviour. This occurs because genes control the production of proteins which determine cell behaviour. There are no mysterious forces. Cell behaviour is governed simply by chemical interactions between its components. We understand the principles and, as the molecular details are filled in, all will be revealed. Reductionism works and has a delightfully sweet smell.