Joshua R. Sanes, Professor of Molecular and Cellular Biology, Paul J. Finnegan Family Director, Center for Brain Science, Harvard University PORTER GIFFORD

I did my thesis research during the Watergate hearings,” says Josh Sanes, who was studying the development of moth sensory neurons at Harvard in the early 1970s. “I sat in the lab cutting 1-micron sections for about a year,” he says. “I did not miss a single word.” The experience inspired him to spend some time on Capitol Hill. “I wanted to see the characters I’d been listening to on the radio,” he says.

The gig at the Office of Technology Assessment was “a gas,” says Sanes. “I got to brief Teddy Kennedy on one of my reports. That was a big thrill. I walked him all the way from the Congressional Office Building across the lawn to the hearing room.” But Sanes decided that being a...

Sanes’ subsequent accomplishments certainly support that conclusion. Over the past 30-odd years, Sanes has amassed an impressive body of basic research on the molecules and mechanisms that direct the formation of synapses—at the neuromuscular junction and in the retina. Here he shares his thoughts on Ramón y Cajal, Larry Summers, and the meaning of life.


Playing with fire. As an undergraduate at Yale University in the late 1960s, Sanes did a research project with future Nobelist Paul Greengard. “He was just a house on fire,” says Sanes. “He was groping toward the kind of biology that can get to behavior—and he had set his sights on cyclic AMP. He came to realize that cyclic AMP worked by activating enzymes, and my senior thesis involved purifying cyclic nucleotide–activated enzymes. Paul is an extraordinary person. He’s turning 87, but still in the lab—and still 100 percent passionate about what he’s doing. Far more than I am! I admire that.”

If at first you don’t succeed.  In 1970, Sanes applied to join Harvard’s department of neurobiology as a grad student. He didn’t get in. “It was a small, elite department, and they only took two or three students a year,” says Sanes. So he enrolled in a different department. Then he got lucky. “The students they accepted my year bombed out. So they were stuck, and took me in my second year.” The first thesis project he attempted was a flop. “Turned out that all the hypotheses we’d made, in the end, were right. But for technical reasons, and because of my incompetence, we couldn’t prove it.” Sanes considered dropping out, but settled on switching labs to work with John Hildebrand on moth neurons.

“The NIH is giving more and more money to translational research. It’s moving toward doing a second-rate job of what biotech does well.”

Seeking communion. As a postdoctoral fellow, Sanes launched a search for the molecules that mark the spot where a nerve cozies up to a muscle. A severed motor neuron, when it regenerates, will return to the same location on a muscle fiber. How does it know where to go? Sanes and his advisors—Zach Hall at UCSF and Jack McMahan, who started at Harvard and moved to Stanford—suspected that the secret lay in the basement membrane: a thin but tough sheath of extracellular matrix that surrounds the muscle fiber. “We came up with the hypothesis that there may be molecules on the basement membrane that are recognized by the axon,” he says. To test that hypothesis, Sanes prepared ‘ghosts’—basement membranes stripped of both muscle and nerve—and asked whether axons would be able to find the original synaptic sites. They did. The next step was to identify the axon-attracting entities. He isolated antibodies that selectively recognized the basement membrane at the synaptic site but ignored the rest of the muscle fiber surface. And when he joined the faculty at Washington University in Saint Louis in 1980, Sanes discovered that the first antibody he’d isolated as a postdoc recognized a particular isoform of a laminin protein, called laminin 2. This protein, produced by muscle, hooks up with a particular calcium channel in nerve terminals. His postdoctoral mentor MacMahan discovered that nerve terminals produce a protein, called agrin, which recognizes receptors present on muscle cells. “So the basement membrane serves as kind of a bulletin board where the muscle and nerve post messages for each other,” says Sanes.

To confirm the importance of these interactions, Sanes and his long-time Wash U. collaborator John Merlie engineered mice that were missing the laminin 2 gene or the agrin gene. “One postdoc in each of our labs spent years making these knockouts,” says Sanes. “It was much harder then than it is now. But we found out that agrin and laminin mutants both had devastating neuromuscular junction defects. So it was quite a satisfying story.” But also a sad one. “John and I saw the very first agrin mutants together two days before he died of a massive heart attack at the age of 49.”

Act locally. In the membrane of a muscle fiber, receptors for the neurotransmitter acetylcholine cluster at the neuromuscular junctions. They’re drawn there, in part, by the agrin released from the nerve terminal. But Sanes and Merlie thought that something more was involved. “In a muscle fiber, there are hundreds of nuclei,” he says. “What if the few that are right at the synapse selectively transcribe the acetylcholine receptor genes? That would make sense, because if a more distant nucleus does the transcribing, the protein might get degraded before it gets all the way to the synapse.” The first experiment Merlie and Sanes did together showed that, sure enough, the mRNA encoding the receptor is localized to the synapse.” This is a very sweet memory for me because John and I actually did the experiment with our own hands, dissecting an obscene number of muscles into tiny little pieces,” Sanes recalls. “Now it would probably be a rotation project you could do with a fraction of one muscle and a PCR machine. But back then, it was quite painful.” The results, which appeared in Nature in 1985, set off a flurry of additional studies of synaptic transcription. Sanes and Merlie went on to demonstrate that synaptic nuclei were indeed responsible. “It’s another way to think about putting together a synapse,” says Sanes: “by local synthesis of the pieces.”

Seeing specificity. “If you want to study synaptic specificity, you have to study the central nervous system, because there isn’t much specificity at the neuromuscular junction: once an axon gets out to the muscle, it doesn’t much care which muscle fiber it innervates,” Sanes explains. “So in the mid-1990s, we decided to study the specificity of synapse formation in the visual system. The retina is quite accessible and has a defined number of cell types. So this is our new playground.” In teasing apart the mechanisms that direct the intricate wiring of the chick retina, Sanes and his team—now at Harvard—discovered a key set of related proteins called Sidekicks  and Dscams. “I think they’re the first bona fide synapse-specificity molecules to be discovered in vertebrates,” says Sanes. “We can show by loss- and gain-of-function experiments that these molecules are responsible for making a subset of synapses in the right place in the retina.”

Cajal in color. Since 1990, Sanes and his colleagues have been generating mice with color-coded neurons. The first line, made by postdoctoral fellow Guoping Feng (now a professor at MIT), painted subsets of neurons with GFP. Subsequently, Sanes and his team—in collaboration with Jean Livet and Jeff Lichtman—developed Brainbow mice: animals whose neurons are labeled with a stunning variety of brightly colored fluorescent proteins. How are such tools useful? “They’re really good for getting pictures on the covers of journals and books,” jokes Sanes. “Seriously, they allow you to distinguish individual neurons out of a morass. Just like what Cajal did with his Golgi stain. If you stain a single neuron in its entirety you can learn about its shape and its connections.” But why label multiple neurons at once? “If you label 100 neurons in a single mouse, you learn much more than if you study 100 mice with one neuron labeled in each. Cajal was an incredible genius in that he could look at one neuron in each of 100 mice and then go home and draw a picture that synthesized all of that information. And almost always he got that right. Anyone else would have gotten it wrong. Because you can’t register all those variations against each other unless you’re supernaturally brilliant like Cajal.” Although the Brainbow team continues to improve on the original design, Sanes says Feng’s original “single-color” transgenic mice continue to be used by labs around the world. “It’s kind of wacky that there haven’t been better ones made in the past decade. But Guoping’s lines were just el primo.”


Best intentions. “Larry Summers was a forceful advocate for science at Harvard. My feeling is that he made a really big mistake in what he said, but that it wasn’t a hanging offense. Everyone makes mistakes. Some people make stupid mistakes. Some people make stupid mistakes repeatedly. But I do feel that this university was a place with high aspirations under Summers, and I miss his ambition.”

The end of basic research? “The NIH is giving more and more money to translational research, like this misguided new center that Collins is starting. It’s moving toward doing a second-rate job of what biotech does well. What’s worse is that the money NIH claims is going to basic research is increasingly judged by translational criteria. Us old guys will no doubt do fine. But this could encourage researchers who are not yet established to disguise their real motivations—or, worse, propose fundable, disease-oriented projects, rather than the great stuff they really want to do. That’s going to cut the guts out of the whole enterprise.”

“Now the deciding votes on any job search in this country are cast by the editors of Cell, Science, and Nature.”

Slippery slope. “It’s gotten to the point that study sections don’t want to fund research that isn’t almost certain to succeed. The best way to guarantee that is to fund research that’s already been done. So you have to submit a grant in which you pretend you haven’t done the work, but let them know that you really have. That’s devastating to young investigators who want to do something that isn’t exactly what they did as a postdoc.” As a result, researchers are not landing their first grants until they’re close to their scientific peak, says Sanes. “So NIH is funding their long slide downward.”

Pyramid schemes. “The hierarchy of journals has gotten steeper. Now the deciding votes on any job search in this country are cast by the editors of Cell, Science, and Nature. If you publish a paper in one these journals you’ve got a chance. If you haven’t, you’re screwed. I had a postdoc some years ago who had done very well. He had two first-author papers in the Journal of Cell Biology and a few others. When he applied for jobs, he was getting nothing. Then one day, he e-mails the places he’d applied to and hadn’t yet been rejected by, and says that one of the papers he’d listed as ‘submitted’ on his CV had just gotten accepted in Nature. He had five interviews within that week. So there’s this incredible nonlinearity. It’s not like a Nature paper is worth a Journal of Cell Biology paper. Or two. It may be worth five or ten. It’s very distorting.”

Alternative history. “If I were coming into the field today I might say, ‘I like doing research, but if I can’t do basic research, then the hell with it, I’ll work in a company and make some money.’ It never occurred to me to go to a company or to start a company. I was just pleased as punch to be able to do what I like. But if I felt I weren’t going to have the freedom to pursue curiosity-driven research, I would ask myself, ‘Should I take a different course?’ I don’t know what I’d answer myself. I really don’t.”

Not-to-do list. “It takes just as much time to do a dumb experiment—or an unpublishable one, or one that’s publishable but totally unimportant—as it does to do a good one. A big part of my job is helping students figure out which experiment not to do.”


What’s a weekend? “I think I’ve been a bit of failure at living a balanced life. I’ve never managed to work 6 days a week and take one day off. It’s more like I work six years and then go on sabbatical for a year. It amounts to the same proportions. But it’s not very balanced.”

The meaning of life. “Like many people, when I was in college, I thought a lot about philosophy and whether there’s more to mind than the brain. I don’t know whether it’s age or achievement orientation or practicality, but I have to admit I think less about that stuff now. There were better drugs then, too.”

No Plan B. “There are people who will tell you they just can’t imagine doing anything else. I can easily imagine not being a neurobiologist. But when I stop to think of what else I could do, I never come up with anything that would be nearly as much fun.”

Exit strategy. “I am not going to be doing this when I’m 80. Absolutely not. But there are no great passions I’ve suppressed. So when I retire, I plan to just hang out and read novels, travel. It will be like having a very long vacation.”


Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?