illustration by Simon Pemberton

A Structural Revolution

A quarter century ago, structural biology was stalling. HHMI created a way to energize it.

Imagine there is a way to peer deep inside a cell, past the cytoskeleton and the organelles, beyond the large molecular complexes. A technology that reveals intimate details about a single protein’s structure, down to the location of its tiny carbon atoms. Now imagine that this method with the potential to unlock the secrets of biology is so obscure, expensive, and elaborate that only a handful of people can take advantage of it. This is, in essence, what structural biologists were up against in the mid-1980s.

"At that time, barely anybody could do structural biology because there wasn’t enough money to get all the necessary equipment,” says Thomas Steitz, an HHMI investigator at Yale University.

Steitz and his colleagues needed help, and assistance arrived in the form of an HHMI initiative. In 1986, the Institute created a program to fund structural biology research around the country. Over the next quarter century, the initiative produced three Nobel laureates, five high-powered x-ray beamlines, scores of innovations in microscopy, hundreds of protein structures, and answers to long-standing questions in biology.

A Trio of Accolades

Three HHMI structural biologists have won Nobel prizes, studying three very different structures.

“From the very beginning it was a very popular program with the Trustees and it was absolutely welcomed with great delight by the structural biology community,” says Purnell Choppin, who was then chief scientific officer of HHMI and became president in 1987. “Many people have told me that the Hughes program really transformed structural biology, not only in the United States but abroad as well.”

The Dawn of Structural Biology

Architects like to say that form follows function—a building’s shape should be based on its intended purpose. The same concept applies to the structure of biological molecules: their forms reflect their functions. Learning what a molecule such as a protein looks like can lead to ways to encourage or hinder its activity, which might be especially helpful if that protein lowers blood cholesterol levels, for example, or is part of a virus.

Unfortunately, protein molecules are much too small to be seen by light microscopes and even most electron microscopes. Structural biologists have developed technical workarounds, however. One of the earliest and most powerful techniques is x-ray crystallography, which involves the often arduous process of coaxing millions of copies of a molecule to organize themselves into a repeating three-dimensional pattern—a crystal. After working for weeks, even months, to grow a protein crystal, scientists then pelt it with intense beams of x-rays, thereby destroying their hard work but also obtaining valuable data.

Each atom in the crystal scatters the x-rays, producing what’s called a diffraction pattern. By rotating the crystal in the beam, scientists can gather diffraction data from many angles. With help from a high-powered computer, the data are translated into a three-dimensional map of the coordinates of each of the molecule’s atoms.

Linus Pauling and Robert Corey at the California Institute of Technology were the first scientists to use x-rays to probe the structures of amino acids—the building blocks of proteins. Combined with information from other groups, what they found was simple, yet profound: an elegant spiral of amino acids called an alpha-helix—one of the fundamental structures found in almost all proteins. They published their results in 1951.

Illustrating the Invisible

Over the years, scientists and artists have used an assortment of techniques to showcase molecular structure. Read More >>

Molecular Mimicry

See how the field of molecular illustration has evolved over the past 50 years.

Page 2 of 4

Less than a decade after Pauling and Corey’s remarkable discovery, Max Perutz and John Kendrew of Cambridge University went bigger. They solved the structures of the proteins hemoglobin and myoglobin with x-ray crystallography, a feat for which they were awarded the 1962 Nobel Prize in Chemistry.

“There were a number of rods in the original myoglobin structure and everyone believed those rods were alpha-helices,” recalls David Davies, a structural biologist at the National Institutes of Health who was at that time a visiting scientist in Kendrew’s lab. “Pauling had proposed the alpha-helix in 1951 but no one had actually seen one. So, in 1959 John and I [analyzed] a section through one of these rods in a higher resolution model of myoglobin and there was an alpha-helix. It was fantastic.”

In addition to publishing his work in the Proceedings of the Royal Society of London, Kendrew described the myoglobin structure in a 1961 Scientific American article. To help nonscientists understand this groundbreaking discovery, he enlisted the talents of scientific illustrator Irving Geis to create the first molecular illustration meant for a general audience.

The ’80s Tech Boom

Those first few discoveries made clear that x-ray crystallography would be a huge player in deciphering the nature and function of molecules. Although it took Kendrew more than 10 years to deduce the structure of myoglobin, subsequent technological advances sped the pace of discovery. “When I first started as a postdoc, if you could determine a structure in three to five years you were doing well,” recalls Brian Matthews, a biophysicist and HHMI alumnus at the University of Oregon. “By the time I came to Eugene in 1970 to start my own lab, the first structure we worked on took three of us a year. That was considered extraordinarily quick.”

“The 1980s were a time when a lot of the technologies that are now the backbone of structural biology and crystallography were introduced,” says Johann Deisenhofer, an HHMI alumnus at the University of Texas Southwestern Medical Center. By the middle of the decade, three developments had pushed crystallography into its heyday. The first was the recombinant DNA revolution. Genetic research had finally made it possible to clone DNA and make ample amounts of any protein. “It was a wonderful moment because we recognized that we were going to be liberated from the constraint of working on proteins that happened to be very abundant,” says Stephen Harrison, an HHMI investigator at Harvard Medical School.

The second advance was the availability of computers that could handle the complex algorithms that turned a diffraction pattern into a molecular map. It became possible to do scientific computations that were unthinkable in Kendrew’s day.

Third, and perhaps most significant, was the availability of a powerful new source of x-rays: the synchrotron. These massive machines fling subatomic particles faster and faster around a huge ring—about the size of a football field—until they approach the speed of light. The powerful radiation emitted by these flying bits of matter can produce x-rays about a thousand times stronger than the ones created in the average laboratory, allowing scientists to speed up their data collection by as much as 100-fold. “This was very important because it turned out in the long run that a lot of our laboratory-based x-ray facilities were not good enough for the job,” says Deisenhofer.

Breaking the Barrier

By 1985, nearly 200 protein structures had been solved, almost all of them by using crystallography. Despite this incredible progress, the field was stalling. The technology was there, but it was elaborate, expensive, hard to use, and often inaccessible.

In a 1985 report to the Board of Trustees, HHMI President Donald Fredrickson wrote, “Soon the access to [the technologies] and the paucity of persons trained to use them will be the critical barrier to continued progress in cell biology.”

Page 3 of 4

The situation prompted Fredrickson to assemble a committee to determine what the Institute could do to break through this barrier. Davies and seven other structural biology experts met in Boston on a Saturday in early March. They spent the day evaluating the state of structural biology and deliberating about how HHMI could support its development. The final verdict: The Institute should create several structural biology laboratories at research hospitals and medical schools around the United States, each associated with an existing HHMI “unit.” Each of the new laboratories would have 1 or 2 principal investigators and a team of 6 to 10 associates, all funded by HHMI. The cost of purchasing and maintaining all the necessary equipment—computers, microscopes, x-ray generators—would be covered. The intention was to make the resources available to HHMI investigators and other scientists at the universities as a way to bolster the field as a whole.

The Trustees supported the scientific leadership’s decision, allocating about $25 million initially and promising $60 million over the next five years. Structural biology became the fifth major area of research for HHMI, joining cell biology and regulation, genetics, immunology, and neuroscience. Despite the prevalence of x-ray crystallography, the new program also committed to supporting emerging technologies such as electron and optical microscopy, magnetic resonance imaging, and nuclear magnetic resonance (NMR).

“Crystallography wasn’t the only tool in the world, but it was the dominant tool,” says Purdue University’s Michael Rossmann, who was then a member of HHMI’s Scientific Review Board. “The labs that were funded were fairly solid crystallographic labs, but many of them have blossomed out to using other tools as they became available.”

Eight scientists at six institutions were selected for the program: David Agard and John Sedat at the University of California, San Francisco (UCSF); Stephen Harrison and the late Don Wiley at Harvard University; Wayne Hendrickson at Columbia University; Florante Quiocho at Baylor College of Medicine; Stephen Sprang at the University of Texas Southwestern Medical Center; and Thomas Steitz at Yale University.

HHMI also agreed to support the creation of a protein crystallography facility at the National Synchrotron Light Source at Brookhaven National Laboratory. More scientists would now have access to a high-intensity x-ray source.

Accelerating Discovery

HHMI invested in a synchrotron beamline to speed protein crystallography.

A Torrent of Findings

The initiative worked and had a cumulative effect.

“Yale already had a center for structural biology that included five senior investigators studying diverse problems,” says Steitz. “HHMI provided technical support, technicians, equipment, and soon there were about a hundred postdocs and students who were using the Yale facility.”

Collaborations, and even new facilities with new tools, took hold. “Because I had one foot in the medical faculty, I started interacting collaboratively with a variety of scientists around the medical school,” says Harrison. “The dean asked me to help lead a small center for structural biology in the medical school to try to enhance its presence. I decided that one of the things this new center should do was spearhead a modest initiative in cryoelectron microscopy.”

It wasn’t long before, as Davies recalls, “the findings started coming out in torrents.”

In 1987, just one year into the initiative, Don Wiley at Harvard University answered a central question in immunology. For decades, scientists had wondered how the immune system could tell the difference between normal healthy cells and infected cells. They knew that a molecule called the major histocompatibility complex (MHC) played a role in tagging the unhealthy cells. But how did MHC flag down the passing T cells that would trigger an immune response? Using x-ray crystallography, Wiley and his student Pamela Björkman, who later became an HHMI investigator, showed that MHC contains a deep groove—perfect for cradling a short piece of protein, much like a hotdog in a bun. He surmised that MHC uses the groove to present foreign peptides to the T cells, which recognize the non-native bits of protein as a signal to act.

X-ray crystallography also proved to be an extremely powerful tool for obtaining information that could be used in fighting viruses. “There was an intense focus on HIV at the time,” says Harrison. “This was only shortly after the discovery of the virus that causes AIDS; there were no adequate drugs, and there was still a scramble to understand as much as possible about the virus and its properties to assist in thinking about therapeutics.” Wiley and Harrison decided to use the flexibility provided by HHMI funding to devote some of their joint effort to work on HIV.

Viruses usurp normal cellular processes to slip inside a cell and hijack its molecular machinery. One of the weapons that allows HIV to gain entry into a host cell is a molecular complex called gp120/gp41. The virus uses the complex to clamp onto a cell surface molecule called CD4. Wiley and Harrison focused their efforts on this entry process. Their work revealed some of the radical changes in shape that gp120 and gp41 undergo when the complex binds to CD4. Conformational changes in gp120 alert gp41 that it is time to go through its own transformation, which, in turn, launches a series of events that lead to the membrane fusion necessary for the virus to enter the host cell. Other HHMI investigators contributed crucial components of this picture, notably Wayne Hendrickson and his student Peter Kwong. According to Harrison, these early HHMI-based structural studies have become one of the foundations underlying current work on HIV vaccines.

Another of HIV’s armaments is reverse transcriptase. This enzyme allows the virus to merge its genome with that of its host, tricking the unsuspecting cell into producing new virus particles to invade neighboring cells. Fortunately, reverse transcriptase has a fatal flaw—it isn’t normally found in human cells, making it an ideal target for drugs. But the enzyme mutates very rapidly—another strategy that results in a very elusive quarry.

In the late 1980s, Steitz began work on solving the structure of reverse transcriptase bound to the experimental drug nevirapine. When nevirapine was discovered, it fit snugly into a pocket on reverse transcriptase where it easily blocked the enzyme’s activity. It wasn’t long, however, before reverse transcriptase mutated, making the drug useless. Steitz’s crystal structure uncovered the reason why. Several parts of the pocket had changed shape, and the drug could no longer bind and inhibit the enzyme’s activity. Luckily, several other areas of the pocket were unaltered, and they were used to create a new version of nevirapine.

Page 4 of 4

But How Does It Fold?

A phenomenon that puzzled structural biologists for decades is how a string of amino acids can fold into an orderly three-dimensional shape like Pauling and Corey’s alpha-helix. Brian Matthews, who became an HHMI investigator at the University of Oregon three years after the program started, set about answering this question. Using a protein called phage T4 lysozyme, he methodically substituted each of its amino acids with a different amino acid and then compared the resulting three-dimensional structures using x-ray crystallography.

“Right now, we are at the point where the techniques have become almost perfect.”

Johann Deisenhofer

His conclusion? Proteins are very tolerant. “Many people had the idea that a protein structure is very complicated—that there’s very fine balance between the unfolded and the folded protein,” says Matthews. “People believed that if you just randomly made substitutions here and there you would probably prevent folding or at least seriously compromise the protein.” Matthews found the contrary: He made multiple substitutions at many sites and the three-dimensional structure was not changed at all. The amino acids on the surface of the protein were especially forgiving—Matthews changed many of them and the protein remained active.

David Agard, an HHMI investigator at UCSF, tackled the protein-folding problem with a different technique—NMR spectroscopy. This powerful method is useful for looking at smaller proteins in buffer solutions. The molecules are exposed to a giant magnet that causes their nuclei to absorb and re-emit electromagnetic radiation. This emitted energy gives clues about each atom’s orientation and location in the protein.

Agard chose a small bacterial enzyme, called alpha-lytic protease, to look at the different shapes a protein assumes as it folds into its final active form. Normally, these intermediate structures are difficult to examine because they’re so unstable and transient. But Agard discovered that if he cut off a piece of the enzyme called the proregion, the protease would be frozen in an intermediate form, which could easily be visualized with NMR spectroscopy. His experiments revealed a surprise: The proregion helps alpha-lytic protease fold into a very unstable, high-energy form before assuming its final active shape.

Not ones to shy away from new technology or big projects, Agard and his UCSF colleague and then HHMI investigator John Sedat also developed several microscopy techniques to help them look at protein machinery at work inside cells.

“Our HHMI support allowed us to make leaps in technology that wouldn’t have been possible any other way,” says Agard. “At the time, the electron microscopists generally weren’t doing big complex cellular things. So we had to combine methodologies and formulate new strategies for collecting and processing data in three dimensions.”

One of the techniques they pioneered was cryoelectron tomography, which involves flash-freezing a cell, photographing it from different angles, and combining the photos to create a three-dimensional model of the cell’s contents. Agard and Sedat used this technique to examine a molecular complex called the centrosome, which is responsible for ensuring that equal numbers of chromosomes are distributed to the mother and daughter cells during cell division. Their images revealed just how the centrosome goes about organizing microtubule fibers in the cell and how the chromosomes then line up on the fibers and move to their respective ends of the dividing cells.

The Shape of Things to Come

Today, there are 36 investigators at 24 institutions in the HHMI structural biology program. More than 86,000 protein structures have been solved and submitted to the Protein Data Bank, an international repository for structural data. “Right now, we are at the point where the techniques have become almost perfect,” says Deisenhofer. “As soon as you have a crystal you can almost certainly determine the structure in a relatively short amount of time. It’s become a standard technique in many laboratories.”

In the coming years, emerging technologies are likely to increase the number and type of molecules that can be studied. For example, a team of scientists led by HHMI investigator Axel Brunger at Stanford University is using a free-electron laser that shoots x-rays at very small crystals of proteins that are hard to crystallize (membrane proteins, for example), opening up a world of structures that had been off-limits for x-ray analysis.

Then there’s the new trend of combining techniques to look at larger assemblies, like Agard and Sedat did. For example, Steitz, who shared the 2009 Nobel Prize in Chemistry for solving the structure of the ribosome—a huge complex of RNA and protein—is blending cryoelectron microscopy and crystallography to capture snapshots of the ribosome as it goes about its job transcribing proteins.

“Between cryoelectron microscopy and the free-electron laser, I imagine we won’t need crystals at all,” says Agard. “The combination of these tools will make it so that we understand the structure of all the molecular complexes at high resolution in isolation and at moderate resolution in the context of the cell. The ability to look at how these complexes are interacting, how they interchange, and what their dynamics are, that’s where I imagine things going.”

Part 1 of 2. In the next issue, our series on HHMI’s structural biology program continues with a look at some of the research that is coming out of the program.

  • 1
  • 2
  • 3
  • 4

Scientist Profile

Yale University
Harvard Medical School
Structural Biology, Biochemistry
California Institute of Technology
University of California, San Francisco
Stanford University
Structural Biology, Neuroscience
Show More