Thinking Big

"He's capable of tremendous intellectual and conceptual multitasking," notes J. Michael Bishop of UCSF. "There's no one who exceeds him in terms of the breadth of areas in which he has made truly seminal contributions."

Those seminal contributions include the discovery that the cell cycle is driven, in large part, by the degradation of cyclin proteins. In addition, Kirschner and colleagues identified the APC protein complex that tags these cyclins with ubiquitin, marking them for destruction. On the cytoskeletal side, Kirschner and associates discovered that microtubules exhibit dynamic instability, a property that allows them to form a variety of cellular structures, including neuronal axons or the spindles that capture chromosomes during cell division. "It's a pretty impressive set of scientific discoveries of the highest possible caliber," concludes Bruce Spiegelman of Harvard Medical School, one of Kirschner's early students.

Making of Microtubules

For someone with such far-ranging interests, Kirschner started off his scientific career much like everyone else: focused on a single, specific problem. As a graduate student at UC, Berkeley, in the late 1960s, Kirschner worked with Howard Schachman on aspartate transcarbamylase, an allosteric enzyme that regulates its own activity.

"It was a very successful PhD," notes Kirschner, whose work on the enzyme's activity-altering conformational changes spawned four papers. "But by the time I finished I was pretty well disenchanted with the physical chemistry of proteins," he says. "I felt it wasn't addressing an essential question in biology" - something like "how you go from a single cell to make all the complexity of the adult organism," says Kirschner. "I mean, I was looking for a really big problem."

That search - launched during a brief postdoc with John Gerhart at UC, Berkeley - drew Kirschner to his work on cell division and microtubule assembly.

"Microtubules were interesting because they seemed to be involved in so many different structures," he says. In addition to mitotic spindles, microtubules show up in nerve axons, cilia and flagella, "and in every cell, no matter what it looks like," says Kirschner. How could this one polymer - built from a protein called tubulin - do so many different things? As a newly appointed faculty member at Princeton University, he started reading the literature and testing researchers' hypotheses, "quickly to find out that their conclusions were mostly wrong."

For example, investigators were debating how microtubules were assembled. Many felt that the tubulin subunits joined together and wound their way around the filament, like the stripe on a barber pole. Others thought that tubulin might form little circles, which then stack into filaments like a pile of lifesavers. That model was supported by an observation that when microtubules fall apart, they leave behind little rings. Using high-resolution electron microscopy, Kirschner and his colleagues determined that the diameter of the discarded circles was larger than that of a mature microtubule cylinder. "So that ruled out the stacking model," he says. They also observed that de-polymerizing microtubules sometimes appeared frayed, which eliminated the winding helix model. Instead, their results suggested that microtubules are built from linear filaments of tubulin that then come together like a bundle of sticks.

'Mind' of Microtubules

Figuring out how microtubules assemble "was a nice accomplishment," Kirschner says. "But it didn't answer any interesting questions" - questions about how microtubules can "decide" whether to form a spindle, an axon, or some other structure.

That answer came a decade later, from a set of experiments designed to study the kinetics of microtubule assembly. By then Kirschner had moved to UCSF, and he and his student Tim Mitchison were watching microtubules polymerize off of a preparation of purified centrioles. "When we threw a lot of tubulin in there, the microtubules were all pretty much the same length," Kirschner says. "But when we tried low concentrations of tubulin, they were random sizes: some were longer, some were shorter." So they decided to do a control. They used high concentrations of tubulin to grow a set of microtubules to the same length. Then they diluted the sample. "To our surprise, some microtubules continued to grow. Others shrank." And they found that the process was dynamic. An individual filament would grow for awhile, then shrink, then do it all over again.

That behavior, says Kirschner, explains how microtubules can form specific structures. During mitosis, for example, chromosomes are strewn randomly around the cell interior. Microtubules have to find those chromosomes and connect them to the cell poles, so they can be segregated when the cell divides. The rapid growth and shrinkage allows microtubules to explore the cytoplasm until they hit a chromosome, which captures the filament like a piece of flypaper. Over time, the only microtubules that remain will be the ones attached to chromosomes. The others will shrink until they disappear. "So the target chooses the array," says Kirschner. "It's a profound idea that explains the kind of lifelike properties that excited me about the system in the first place."

Being able to extract such broad principles from a relatively simple observation is just one of Kirschner's skills. "For a lot of people, if they try to think big, they can't get anything meaningful done," says Douglas Koshland, an HHMI investigator at the Carnegie Institution, and another former postdoc. "Marc can think about the big picture and come up with insightful experiments to get at it."

He also enjoys a good story. The discovery that cyclins get ubiquitinated before they are degraded, for example, was made after student Michael Glotzer left a gel in the freezer while he went camping. He was monitoring the degradation of radiolabeled cyclin protein in frog extracts at different times during the cell cycle. Before the trip, Glotzer exposed the gel to a piece of film and confirmed that the band representing the intact cyclin protein was disappearing as the extracts progressed through the cell cycle. He then stuck gel and film back in the freezer and grabbed his gear. Two weeks later, looking at this longer exposure, Glotzer says, "you could see this faint ladder of bands - shifted up the gel - indicating that the cyclin was getting ubiquitinated before it was degraded."

Kirschner "loves to tell that story - although in his version I was off skiing," says Glotzer, now at the University of Chicago.

'Mind' of a Mentor

Kirschner is a big believer in giving his trainees the freedom and support to pursue their own interests. "Marc once told me that he sees his lab as sort of like an expat café in Paris, where artists lounge at tables, churning out their sketches, novels, and plays. He merely provides the venue - and the occasional coffee or pastry," says Deshaies.

And Kirschner allows his mentees to take those projects with them when they leave - a "win-win situation" that further entices the most independent investigators, says Glotzer. "If you look at the number of people he's trained who've gone on to successful independent careers, it's pretty remarkable. I can't think of too many people who have as large and successful a pedigree."

For Kirschner, giving away projects keeps him scientifically spry. "It puts pressure on me, a very good kind of pressure, to continually come up with new things for the lab to do," Kirschner says. "So I've gotten to enjoy that aspect of science." In 1993, he carried that adventurous spirit with him from UCSF to Harvard Medical School, where he created what Spiegelman calls "one of the best cell biology departments in the country." A decade later, he established the systems biology department, of which he is currently chairman.

Kirschner also led the charge to create interdisciplinary, cross-departmental graduate programs, both at UCSF and at Harvard, which allowed the schools to compete for the best students. "He's a forward-looking person in everything he does," says Bishop. "Not just in science, but in his approach to politics and to education."

In addition, writing two books on evolution with Gerhart doesn't leave much time for doing experiments with his own hands. "When I arrived in the lab in 1994, he still had a bench. But it didn't look like it had been used any time recently," says Jan-Michael Peters of the Research Institute of Molecular Pathology in Vienna, who identified APC while a postdoc in the Kirschner lab. "He had dissolved a peptide and started a notebook that had one page, noting the concentration. It became a running joke in lab: this is Marc's peptide that he dissolved."

Regardless of whether he ever did anything with that peptide, "Marc has created this place where you can develop as a scientist and do really cool, really great science," says Michael Rape of UC, Berkeley, another member of the Kirschner cabal. "When my friends ask me where they should do a postdoc, I tell them to go to Marc. What more can I say?"