In a research career that lasted more than 60 years from 1945 until his death on 23 July, Dan Koshland used his skills as a chemist and highly creative thinker to advance our understanding of proteins and cells. An energetic and caring man, passionate about all aspects of science, his persuasive and insightful leadership greatly strengthened every institution of which he was part.

Koshland decided to pursue science after reading, as a young teenager, Paul de Kruif's collection of biographical essays, Microbe Hunters, and Arrowsmith, Sinclair Lewis's novel of an idealistic scientist and doctor. After receiving a bachelor's degree in chemistry from the University of California, Berkeley, in 1941, Koshland was recruited to the Manhattan Project, and became a group leader under Glenn Seaborg in the team at the University of Chicago that successfully purified plutonium. When the war ended, he entered graduate school in Chicago under another giant of chemistry, Frank Westheimer, then a young assistant professor. Koshland attributed much of his success to the mentoring and inspiration of these two men.

After a brief stint at Harvard, Koshland moved in 1951 to Brookhaven National Laboratory on Long Island, New York, where he focused on understanding through physical organic chemistry how enzymes catalyse biological reactions. He became fascinated by the puzzle of enzyme specificity: how does an enzyme select only one of many closely related molecules as its substrate?

At the time, the experts believed that Emil Fischer's 'key–lock' or 'template' theory, proposed in the late nineteenth century, had solved the problem. But Koshland realized that this model of a rigid enzyme could not explain why an enzyme catalyst is much less effective when a distant part of the substrate molecule is removed. The correct answer, published by Koshland in 1958 only after numerous rejections, is that enzymes are activated by changing their shape when they bind their substrates, and that process is influenced by the distant part of the substrate. This principle, called 'induced fit', was eventually confirmed fully through detailed structures of enzyme–substrate complexes produced by X-ray crystallography. Induced shape change has proved to be a ubiquitous phenomenon, applying not only to substrate recognition, but also to much of the regulation of enzyme activity in biological systems.

In 1965, Koshland moved back to Berkeley, to the Department of Biochemistry. A year later, he published a classic paper, 'Comparison of experimental binding data and theoretical models in proteins containing subunits', now recognized to provide the most general model for cooperative (and anti-cooperative) responses in proteins. In the 1970s, he and his lab shifted to focus on understanding how cells receive and interpret signals, wisely selecting bacterial chemotaxis — the movement of bacteria in response to changes in the concentrations of certain chemicals — as a system that could be explored in exquisite detail. By rapidly changing the concentrations of the molecules that bacteria detect, Koshland demonstrated that bacterial cells have a 'memory' of their environment. In later years, members of his lab worked out much of the biochemistry of the receptor-protein methylation that underlies this process.

In 1976, Koshland showed that differences between bacterial daughter cells can be caused by stochastic fluctuations inherent in systems driven by small numbers of molecules. How biology deals with such fluctuations is now a focus of research in many laboratories. In 1984, together with Albert Goldbeter, Koshland introduced the concept of 'ultrasensitivity', revealing how large differences can originate from a cascade of enzymatic reactions downstream from an initial signalling event. All this work has changed the way we think about biology; Koshland's mentorship has spawned many leaders in these fields.

While maintaining an active research laboratory, Koshland became chair of the Berkeley biochemistry department in 1973 and began a series of institution-building activities across the university. In the 1980s, he drove a revolutionary reorganization of Berkeley's 200-strong biology faculty from more than ten different small departments into three large ones. This was a seemingly impossible task to those versed in university politics, but its beneficial effect on the quality of both faculty and student recruitments was immediately apparent. The recent development of biology-focused interdisciplinary science and engineering efforts at Berkeley also owes much to Koshland.

In 1985, Koshland became editor-in-chief of Science, a pivotal position for world science that he occupied, as a half-time post, for ten years. He transformed the journal, not only vastly improving the scientific research it published by appointing a cadre of talented full-time editors with PhD degrees, but also greatly increasing the quality and quantity of the sections dealing with science news and science policy. He wrote more than 200 editorials, many in the form of dialogues with 'Dr Noitall' that were testaments to both his originality and his remarkable wit. These duties did not dent his formidable scientific productivity: during his editorship, he authored more than 100 scientific articles stemming from work at his beloved Berkeley laboratory.

Dan Koshland was both a model scientist and a model human being. He exemplified all the values that the idealists among us associate with science. He was completely open and honest, with a generous, happy spirit that made him a pleasure to be around. He derived great pleasure from being helpful to everyone, and he went out of his way to support young scientists. I was amazed to discover late in his life that he, a modest man who always flew in economy class and drove a dented car, was very wealthy, heir to part of the Levi Strauss fortune. In fact, Dan was a great philanthropist for science, helping to finance new buildings at Berkeley and at Haverford College in Pennsylvania, and both designing and endowing a new science museum at the National Academy of Sciences in Washington DC. I had the privilege of working closely with him in designing that museum. It is named after his first wife, Marian (Bunny) Koshland, an outstanding scientist in her own right, to whom he was married for 52 years until her death in 1997.

Dan is survived by five children from that first marriage, and by his second wife, Yvonne, a former college friend whom he married in 2000. Family, friends and colleagues will come together to celebrate the breadth of his contributions to science in a memorial service and symposium at Berkeley on 16 September.