Getting the right fold

For almost two decades Arthur Horwich has been unraveling a basic biological mystery: how proteins achieve their native shapes.

By Cathy Shufro
Photograph by Terry Dagradi

Christian B. Anfinsen, Ph.D., a biochemist at the National Institutes of Health (NIH), in the early 1960s answered one of biology’s fundamental questions. He explained where proteins get the marching orders that direct their metamorphosis from amino acid chains into active, functional three-dimensional sheets and helices. Function follows form, and those distinctive shapes dictate how the proteins work. Shape allows the lock-like structure of an antibody to trap the key-shaped antigen to fight infection; it allows an enzyme to work on a specific substrate to speed up a chemical reaction, such as digesting food; and it permits a hemoglobin protein to bind to oxygen and carry it in the blood. But when a protein is improperly folded, its function is lost, and that can lead to disease.

Scientists already understood that DNA embodies a code that is transcribed onto RNA, which in turn directs the cell to produce the 30,000 vital proteins that are the body’s building blocks. Scientists recognized that these polypeptide chains—composed of the 20 amino acids—don’t function as proteins until they have folded, but they did not know, before Anfinsen, where proteins got the instructions to fold.

Anfinsen shared a Nobel Prize in chemistry in 1972 for a simple experiment that yielded remarkable results. He unfolded an enzyme to see if it could find its way back to its active, folded form. It did, showing that the sequence of amino acids itself contains the information that tells a protein how to fold, and that the protein folds spontaneously.
That seemed to be the final answer when Arthur L. Horwich, M.D., HS ’78, was studying biochemistry as a freshman at Brown University in 1969. But 18 years later, Horwich, by then an assistant professor of genetics at Yale, began to think there might be more to protein folding. While studying the passage of unfolded proteins through membranes into mitochondria, the cell’s power source, he asked whether a specialized protein might be assisting the folding process.

Anfinsen had correctly discerned that amino acid sequences do provide the information needed for folding, but Horwich found evidence that many amino acid chains need help: he discovered “folding machines” that provide that help.

Now the Eugene Higgins Professor of Genetics and Pediatrics and a Howard Hughes Medical Institute (HHMI) investigator, Horwich has spent 17 years studying those machines.

“This work is as basic to biology as understanding the nature of genes and how genes are expressed and translated into proteins,” says colleague Richard P. Lifton, M.D., Ph.D., the chair and Sterling Professor of Genetics at Yale and an HHMI investigator. “The part that is really breathtaking about Art’s work is that he went all the way from basic discovery of the protein-folding machine to understanding how it works on an atomic level. … It’s just one of the nicest pieces of work one will ever see in biology.”

Seeking what’s never been seen
Horwich rarely sets foot in his office at the Boyer Center for Molecular Medicine. He’d rather be in the lab.

“I just have never matured beyond postdoc,” jokes Horwich, a slim man with unruly brown hair, glasses and a droopy, graying moustache. “I still work at the bench every day. I still like to do my own experiments. I like to be able to live with and suffer through the problems of understanding how things work side by side with my own people, and I always have one or two things for myself that I consider my own laboratory struggle.”

When Horwich must be away, he calls daily to check on the progress made by his 10-member team. “I drive them nuts.”

“He’s a very, very intense person,” says Krystyna J. Furtak, M.S., an HHMI research technician who has worked with Horwich for 20 years.

Dressed in faded corduroys, winter and summer, Horwich sets up his experiments beside the makeshift office he has created at the end of his bench, where piles of journals and papers rise from the floor and windowsill. The lab draws Horwich like a magnet, because there each day holds out the chance, however small, “to see something that’s not been seen before.”

Horwich grew up in River Forest, a Chicago suburb where he enjoyed snowball fights that endangered priceless stained-glass windows in the neighborhood’s Frank Lloyd Wright homes. His father, Walter, a Chicago businessman, shares Horwich’s interest in science.

“He was a sort of frustrated scientist,” Horwich says. “He let me be a ham [radio] operator and he was just as interested as me. He was up on the roof with me putting up antenna systems. He had to draw the line when I wanted to bounce the signal off the moon. He wouldn’t let me put a dish up on the roof.”

Horwich graduated in 1975 from Brown Medical School, where he was class valedictorian, a fact he mentions only to tell a story on himself. “I gave this speech saying Brown should train great primary care physicians. So I promptly went off and did research.” He jokes that his parents took the speech to heart. “I think the folks are still waiting for me to practice medicine.”

During his pediatrics residency at Yale he felt the lure of genetics. After his residency, he spent three years at the Salk Institute in California with tumor biologists Walter Eckhart, Ph.D., and Tony Hunter, Ph.D.

“I watched Tony make what is probably one of the most important discoveries of the latter part of the 20th century,” Horwich recalls. In 1979 Hunter discovered tyrosine phosphorylation, a process that plays a role in normal cell signaling and growth and that also contributes to the formation of cancerous tumors if it goes awry: of 300 genes implicated in cancer, 90 encode tyrosine kinases.

“Art still works in the lab, which amazes me,” says Hunter, still at the Salk Institute. “He seems to have retained the enthusiasm for science and the day-to-day excitement that most well-established scientists running big labs seem to lose.”

Horwich returned to Yale in 1981 to work with human geneticist Leon E. Rosenberg, M.D., HS ’63, who later became dean of the School of Medicine. “I worked with a large number of postdoctoral fellows, but none had Art’s combination of brilliance, a love of experimentation and a fearlessness to learn what he needed to learn to go further,” says Rosenberg, now a professor at Princeton.

A late-night discussion and a mutant yeast
Horwich’s first significant discovery grew out of a late-night conversation in 1987 with then-graduate student Ming Cheng, M.D., Ph.D., about whether unfolded proteins crossing membranes into mitochondria might get help folding into active forms. “Maybe there’s such a thing as a folding machine,” Horwich proposed. Using a library of mutant yeast strains, they looked for one in which a protein that had clearly entered the mitochondria had reached its mature size, yet showed no sign of enzyme activity. That would suggest that it had not folded. They found such a mutant within days. They could hardly believe what they were seeing.

They traced the mutation that disabled folding to a defective gene for mitochondrial heat shock protein (called Hsp60). Heat shock proteins, scientists believed, played a role in protecting already folded proteins. When Horwich and Cheng inserted a normal gene for Hsp60 into the mutant yeast, protein folding resumed.

“It didn’t just hold the proteins,” says Horwich, who still sounds excited 17 years later. “It folded them. It was a folding machine.”

But Horwich’s team was not ready to announce the discovery. “We thought there must be something else to explain it. … We were terrified of being wrong.” They worked with another group in Germany to confirm their findings. That group was led by Franz-Ulrich Hartl, M.D., Ph.D., a director of the Max Planck Institute of Biochemistry, who along with Horwich and R. John Ellis, Ph.D., of the University of Warwick in the U.K., won the 2004 International Award from the nonprofit Gairdner Foundation, which recognizes outstanding achievements in biomedical research. The three were honored for having “revolutionized our understanding about basic cellular functions.”

In 1989, after a year of testing and retesting of a variety of proteins in mutant yeasts, Horwich and Hartl published their findings in Nature. Working together, the scientists and their teams went on to show that specialized proteins mediate protein folding in archaebacteria, prokaryotes and eukaryotes.

In the years that followed their discovery of the role of Hsp60, Horwich and his colleagues turned to understanding the inner workings of a closely related folding machine in bacteria called GroEL, using genetic, biochemical and structural methods. Like Hsp60, GroEL is a chaperonin, a double-ringed molecular “machine” that resembles two stacked doughnuts.

Chaperonins begin their work by binding an unfolded or misfolded polypeptide molecule inside one of the rings (see last diagram on left). The working principle, Horwich says, is that unfolded or misfolded polypeptide chains expose hydrophobic (“greasy”) surfaces that become buried to the interior in the fully folded, functional (“native”) form. In the nonnative state, the exposed surface of one protein can stick to that of another, setting off a process of multimolecular aggregation. This results in an inactive protein that cannot perform its intended function; in addition, aggregates can harm the cell. Binding in a chaperonin ring serves to mask such surfaces of the nonnative protein, because the chaperonin itself has a cavity lining that is a hydrophobic surface, which interacts with that of the nonnative protein. It is thus prevented from aggregating.

The machine then switches to a folding-active state by binding ATP, a unit of energy currency, and a lid structure, a co-chaperonin called GroES. The chaperonin ring that binds GroES undergoes large, structural changes that cause its hydrophobic lining to turn away from the bound polypeptide and interact directly with GroES. In the process, the poly-peptide is released into the now-encapsulated chamber, where it folds to its native form. Proper folding is favored in this chamber, because the polypeptide is in solitary confinement with nothing else to aggregate with, and because it is now next to watery walls that encourage the burial of its hydrophobic surfaces and exposure of its watery ones, properties of the final native state. Finally, after 10 seconds, a step of conversion of ATP to ADP causes the machine to release the lid structure, and the polypeptide leaves the cavity. As the cavity of the ring that had been occupied becomes emptied, the other ring swings into action to perform the same function with another polypeptide chain.

But each protein does not always fold properly on the first try, and Horwich says the protein-folding mechanism reveals something fundamental about where a living cell directs its energy. The machine will try again and again to fold the protein correctly. Each cycle consumes seven ATPs (energy-storing molecules), and for some proteins, up to 20 trials may be required. “Think of the cost,” says Horwich. “It’s a very expensive process. The cell has decided it will spend a lot of ATP to try and correctly edit the protein-folding process.”

The consequences of misfolded proteins can be dire. Misfolded proteins are linked to hundreds of devastating diseases—including amyloid diseases such as Alzheimer’s, Parkinson’s, Huntington’s, ALS (also known as Lou Gehrig’s disease) and spongiform encephalopathies like mad cow disease. In the amyloid disorders, misfolded proteins stick together to form fibrils, or plaques, in the brain. Scientists are not yet sure how these diseases occur. For example, they don’t know whether the fibrils themselves injure the brain or whether the misfolded proteins cause damage on their way to forming fibrils.

Looking outside the cell
After two decades of studying the genesis of proteins within the cell, Horwich is interested in what happens to proteins in the extracellular spaces of the brain. During a break from the lab on a hot afternoon last summer, he mentions being both troubled and intrigued by a recent article in Nature examining the aging of the human brain. The article, “Gene Regulation and DNA Damage in the Ageing Human Brain,” reports that brain function begins to deteriorate when we get “old”—age 40. “That’s the shocker,” says Horwich. “All the good things are going down—vesicular trafficking, synaptic plasticity. The DNA is damaged increasingly. … Yet all these machines that help us to repair damage, including the heat shock proteins, are getting induced.”

He sees some good news in this observation: ramped-up defenses suggest avenues for treating disease. Horwich envisions two approaches. “If you want to think about what to do about neurological diseases that involve misfolded proteins, you could shut down production of the proteins that get misfolded. Or you could upregulate this line of defense [provided by molecules like chaperonins].”

He wonders whether the body has other machines similar to chaperonins that might prevent the misfolding of proteins that causes disease. “What other machine is out there that we’ve really missed, that’s outside the cell? Every-one’s been focused inside the cell.”

One step is to learn about the structure of the fibrils formed by misfolded proteins in the brain. Horwich and his team, in collaboration with Wayne Hubbell, Ph.D., at UCLA, have been using a spectroscopy approach to study how the normal structure of a serum protein, transthyretin (pre-albumin) gets converted into an amyloid in a disease known as familial amyloid polyneuropathy.

“I really want to focus on these things. I don’t have an unlimited amount of time. After reading the Nature paper about the brain, I’m not too optimistic,” says Horwich with a wry smile. He’s 53.

Horwich is not wedded to the idea of curing disease, however. “Of course it would be wonderful,” he says perfunctorily. And he does not seem interested in fame. Although he was recently elected to the National Academy of Sciences, and although as a Gairdner winner he has, statistically, nearly a one-in-four chance of going on to win a Nobel, Horwich shrugs when he is described as a distinguished scientist. It’s the science itself that drives him. “I’m particularly interested in seeing the beauty of how Mother Nature has handled the problems of protein folding and this last step of information transfer. … Mother Nature decided it’s not going to leave even the step of protein folding to chance.”

Outside the lab, fishing and tennis
Although Horwich spends long hours in the lab, that’s not to say he works nonstop. On weekends, he plays tennis, sometimes with his 24-year-old daughter Annie, a photographer. A few years ago, 26-year-old Michael, an M.D./Ph.D. student at the University of Massachusetts, introduced his father to fly-fishing. When it comes to fishing nowadays, his youngest child, sixth-grader Dave, is “a complete fanatic.”

Horwich is happy to get away occasionally, taking a few days off to relax on Block Island or backpack in the Adirondacks, “because you think about things on a different plane when you’re not right up against it. I’m OK with being away from it for a while, but there’s a feeling that it’s time to get back and touch it again.”

His wife, Martina Brueckner, M.D., FW ’90, understands the pull of the laboratory. An associate professor of pediatrics (cardiology), she studies the development of left-right asymmetry in the embryo, which affects the placement of the heart and viscera. They both spend long hours in the lab.

What keeps him in the lab is not the anxiety of falling behind, says Horwich. “It’s just: can we find the next really big thrill? There’s no substitute for the thrills we’ve had. Most people are lucky if they have one. I’ve had two.”

The first came after he and Cheng found the yeast mutant that failed to fold proteins. The second came when Horwich saw for the first time the structure of the bacterial chaperonin. For four challenging years in the 1990s, Horwich and his team had collaborated with the group led by the late Yale structural biologist Paul B. Sigler, M.D., Ph.D., trying to get an image of the protein molecule. They grew hundreds of crystal forms of GroEL, altering each slightly in hopes of finding one that would create a clear image by means of X-ray crystallography. Then, at last, one of Horwich’s graduate students, Kerstin Braig, hit upon a crystal that worked. It was a molecule that had an accidental mutation that made it diffract well. Horwich calls the discovery pure luck.

Just a few days later, Horwich’s and Sigler’s groups drove to the synchrotron at Cornell in Ithaca, N.Y., where they could use its high-intensity X-ray source to collect more data about the molecule. Assembling the data, or “phasing the crystal,” required several trips over several months, and Zbyszek Otwinowski, Ph.D., a member of Sigler’s lab at the time, performed what Horwich describes as “an amazing feat.” In just two days, Otwinowski “solved the structure” of GroEL and created a three-dimensional model of the protein.

“We worked for four years trying to get the crystalline structure of this, and there it was, at atomic-level resolution. That was a religious moment.”

Having experienced these highs creates a certain tension. “It’s a wonderful experience, but you always wonder, are you ever going to have another really fabulous revelation? You try to set things up so you will. You try to see beyond where things are, to see a way to open up a system so you can see how things work in one fell swoop. Science is usually more incremental, and that’s OK. It’s where you get to see a whole vista at once—that is really special stuff.

“It’s like pulling Venus from the waves. You have no idea what you’re going to see until you see it. It’s just unbelievable.” YM

Cathy Shufro is a contributing editor of Yale Medicine.


			Originally published in Yale Medicine, Fall/Winter 2004. Copyright © 2004 Yale University School of Medicine. All rights reserved.