Paul Nurse is one of the most celebrated biologists of his generation. His work on the behaviour of cells is at the heart of cancer research, yet he thinks that a cure for Britain's No 1 killer is still 30 years away. Jon Turney reports

Your cells got you where you are today. Once there was just one. By now, there are trillions of them, each making two new cells just often enough to keep you going. But any one of them could yet be the death of you. Any one can break ranks, start reproducing out of turn. Some of its descendants shake off other restraints and seek nutritious sites to build their new object in life: a malignant tumour.

This is why understanding cell behaviour is at the heart of cancer research. And why a man who, 25 years ago, set out to study cell division as a problem in pure science is today head of the Imperial Cancer Research Fund, the country's largest cancer charity. Paul Nurse's unravelling of his chosen problem shows the prescience, clear thinking and luck that a successful scientist needs, in a combination that has made him one of the most celebrated biologists of his generation. Yet it nearly did not happen.

In the early 1970s, Nurse was half-way through a not particularly satisfactory PhD at the University of East Anglia. He had been drawn to science since his schooldays, but the move from learning to doing was proving trying. "As an undergraduate, you only read about the beautiful experiments that all work," Nurse says. "When you are a postgraduate and you do them yourself, they are not such beautiful experiments and most of the time they do not work." Sitting up nights, keeping watch over a temperamental amino acid analyser, he wondered if the experimental life was really for him, and he toyed with the idea of giving it up for philosophy.

³ÉÈËVRÊÓƵ

If not philosophy, what? He had turned away from his interest in ecology and natural history in favour of the controlled experiments of the biology lab. But he was still interested in whole organisms, while the molecule-by-molecule analysis that dominated the laboratory invariably meant ripping such organisms apart. What might give him a way forward in science that combined a concern with systems that were still living with the control of variables he craved? And what was important enough to induce him to carry on in the face of all those failures, to deal with "the cold water of having to do experimental manipulations all the time"? The answer to both questions was the same: the cell, the smallest piece of anything that is still alive.

Biology textbooks told of the life cycle of a single cell, which microscopic observation had recorded in great detail. It was a complex affair, with several different phases and wondrously ordered preparations for cell division involving chromosome copying, separation and repackaging of the chromosomes into two cell nuclei where once only one existed. The scheme presented was in essence descriptive: what might control when and how often a cell divides was unknown. Nurse decided to find out.

³ÉÈËVRÊÓƵ

It must have required unusual self-confidence, though looking back now as an affable near 50-year-old at the peak of his profession, Nurse makes it sound obvious. "The problem in biology is how you make an organism, how organisms change in space and time. Seeing a cell develop was that in its simplest possible way, because a cell just grows, doubles its size, reproduces everything in itself, and divides. I decided that the cell cycle had all the characteristics of a problem that defined the interesting features of life." Moreover, it should be open to study with the genetic techniques then coming into use.

Nurse had never worked on the cell cycle, nor did he know much about genetics. But he persuaded a more experienced researcher, Murdoch Mitchison in Edinburgh, to take him on. He flashes a smile and says: "I just got into the problem with a running jump."

The best-studied cells were in yeast, and that was where he began. An American, Lee Hartwell, had paved the way with studies of budding yeast. Nurse would work with so-called fission yeast - the names indicate how the types reproduce. Before settling in Edinburgh, he spent half a year in Berne learning yeast genetics. Within a couple of years, he reported finding a yeast strain that split into two when it was only half the normal size for cell division. There followed a host of other mutant strains, about 50 were known by 1990, with alterations in their genes that affected the cell cycle in various ways.

The most useful strains were temperature-sensitive, permitting normal reproduction but getting stuck somewhere in the cycle if they warmed up. To find out what the products of all these genes do meant shifting from overall genetic mapping to the molecular level. After moving to Sussex, Nurse set up a group to do this, building up libraries of normal yeast DNA and inserting pieces into mutant strains one at a time, to find out which one would kick-start a temperature-sensitive cell.

The most important early finding was a gene that woke up cells whose cycles halted after they made two sets of chromosomes but before they began to move apart. This gene enables the cell to make a protein that, when bound to another key player called cyclin, works on several other proteins to allow the cell to move on to division.

It was a stunning illustration of the power of genetic techniques to unravel complex phenomena. No wonder Nurse is irked by some portrayals of genetics. As he told a meeting on reductionism in London last year: "I'm a geneticist, and I get sick of the fact that we are always being pilloried as Neanderthal reductionists, one step away from a Nazi concentration camp warden. Genetics looks at the whole organism."

Further work on the gene Nurse had found showed that this controller is controlled, in turn, by other proteins. The details rapidly become too complex to relate. Read the papers and you understand the force of Nurse's comment that one limit to a reductionist programme to enumerate every molecule in a cell is boredom. But the basic principle is simple. There is a fundamental mechanism of switches in cells - used to control progress through the cell cycle, among other things - that depends on a relatively simple chemical change. The switches themselves can be changed in the same way, and they can often sense the presence of other chemicals. Combining such elements means that you can build complex networks of signals whose components interact with each other and with other parts of the cell. A whole set of sensors and switches that automatically responds to the levels of particular molecules present at any particular time can coordinate the movement of one cell towards becoming two.

³ÉÈËVRÊÓƵ

By now, the relevance of all this to cancer was clear, and Nurse moved to the Imperial Cancer Research Fund's laboratories in London's Lincoln's Inn Fields, leaving for a stint as professor at Oxford before returning to the ICRF as director of research, then director-general. Now he has to help fashion scientific strategies for hundreds of scientists in dozens of laboratories, but he is still mindful that the work can grind you down and that the really satisfying moments are few and far between. The bottom line is that the problems must be worthwhile, he says, because if the stakes are high, "it's worth the pain".

³ÉÈËVRÊÓƵ

No one could argue that the stakes are not high in cancer, which has just overtaken heart disease as Britain's number one killer. An organisation with a mission to tackle cancer must work with treatments for patients and schemes for prevention. But Nurse the biologist is equally committed to longer term programmes - he speaks of part of the institute's portfolio being "risk-invested". His own career is a good example of the way such investments may pay off - we now know that the proteins that regulate the cell cycle in yeast do the same job in humans.

Nurse's work still focuses on the simpler organism. At the ICRF he has managed to do what most scientists-turned-administrators rarely achieve: keeping in touch with the lab. He says he spends half his time running his research team and half running the organisation, but he admits that others' idea of half time in the lab may be different from his.

Aside from refining details of the cell cycle, his team is working on another problem - in essence, how cells know which way is up. They have found a protein that locates end-markers in yeast cells. Again, the possible connections with cancer are there to be explored, for cancer cells change shape when they migrate. Finding out how a cell knows its shape should help explain such changes.

Together, the two investigations bracket properties of cells that Nurse emphasises cannot be studied one molecule at a time: organisation in time and organisation in space. If all of the control elements that regulate the cell cycle can be identified, for example, what will the list tell us? Even if their connections are drawn in a pretty diagram on paper, it is likely to be a pale reflection of what is actually going on inside the rich little world of the simplest living unit. Nurse the scientific strategist wants to move beyond the naming of parts: "The methodologies we are good at tend to lose information." Understanding the cell cycle, he suggests, will mean "actually describing in real cells in real time and space how these signalling pathways work".

Suppose, for example, that what matters is not whether a particular signal is activated but how long it is activated. Suppose in fact that cells have some kind of internal Morse code. Then consider the distances messages written in such a code must travel, which are small by our standards but very large compared with the size of a protein molecule. So how can cells monitor their internal states as three-dimensional empires with outlying provinces of proteins as well as nucleic acid heartlands?

"We may not find it easy to think about these sorts of problems," Nurse reckons, but he is looking to control theory and techniques to produce new ways of describing the emergent complexities of the cellular interior. In the end, though, however much this language speaks to the limits of one kind of reductionism, practical approaches to cancer will demand going down to the next level. As he said last year, "when you are interested in engineering and manipulating a system, you have to go down to the molecular level because that is the only way you are going to be able to operate on it".

He still thinks that major improvements in cancer treatment - the hope that secures the Pounds 1 million a week the ICRF needs to keep going - are 30 years away. That could at least mean that the scientists who deliver such improvements have already started work, even if they are still at the stage Paul Nurse was when he first fixed on the cell cycle as the part of life he really wanted to understand. And when they come, the treatments will almost certainly have one thing in common, that they eliminate the cells that cut loose from the controls on cell division that Nurse has helped to make visible.

Jon Turney teaches science communication at University College London.

³ÉÈËVRÊÓƵ