3-Dimensional Model : Cold Virus: an Image of the Unseen

It began in the summer of 1981 with a chance conversation between two scientists at an international conference in Strasbourg, France.

Michael G. Rossmann, 55, a professor of biological sciences here at Purdue University, mentioned that he was looking for a new project and was thinking of tackling the polio virus.

Roland R. Rueckert, 52, a microbiologist at the University of Wisconsin in Madison, stiffened with alarm. At the time, he was helping a Wisconsin graduate student study just that subject. Rossmann would be tough competition.

Rueckert carefully suggested that Rossmann instead go after the common cold virus. Rueckert offered to supply samples of the cold virus he was then growing in his own lab. After five hours of talk, Rossmann agreed.

Scientific History

In this inadvertent fashion, the two became collaborators in an extraordinary endeavor that would make scientific history. Two months ago, they and their colleagues reported that they had created a three-dimensional model of a cold virus--the first such image of any virus that infects humans and animals.

Scientists marveled at the news. Accustomed to thinking theoretically about things they cannot see, they now had before them, in exquisitely fine, atom-by-atom detail, one of the world’s most resilient agents of disease.

They could see how it is structured and how its various parts fit together. They could see how it attacks the body and how the body reacts.

They could, tentatively, lay plans to map the structure and understand the workings of similar viruses that cause such diseases as multiple sclerosis, hepatitis, leukemia, diabetes, certain tumors, foot and mouth disease and AIDS.

Scientists could also wonder at the crafty purpose built into the cold virus’ design. Its structure showed clearly why all man’s knowledge has failed to prevent or cure colds.

Vague, Fuzzy Dots

Scientists who before saw the cold virus through electron microscopes only as a vague, fuzzy ball now found that it was composed of 20 triangle-shaped sides, marked on the surface by ridges and canyons.

They surmised that the part of the virus that grabs onto a human cell and causes infection lies deep within those canyons. As it happens, the body’s infection-fighting antibodies cannot fit into those narrow valleys.

A preventive vaccine, which works by generating antibodies, seemed more unlikely than ever.

“A clever little strategy,” Rossmann said admiringly. “Very pretty. It all makes sense. That’s what our work presupposes: When you get it right, it will make sense.”

If the pursuit of the cold virus unveiled a natural world of elegantly simple order, however, it revealed also an unruly scientific world full of complicated human emotions.

The drive for knowledge about the virus was also an aggressive contest in which ambition and personalities played major roles. Rueckert’s attempt to prevent competition in the end yielded much the opposite result.

The graduate student he tried to protect from Rossmann, James M. Hogle, soon after received an attractive offer from the Research Institute of the Scripps Clinic in La Jolla, Calif. In October of 1982, Hogle moved to California and set to work in his own lab, pursuing the structure of the polio virus.

Sent Cold Viruses

Rueckert, meanwhile, had kept his promise to send Rossmann cold viruses. He had also taught a visiting Purdue delegation how to grow and crystallize their own.

So, unintentionally, Rueckert became part of the team at Purdue pitted against Hogle in the race to generate the first animal virus model. It would prove a contest tinged with some bitterness.

“That’s just the way it is,” Rueckert said after it was over. “Science isn’t dispassionate. Human emotion is the engine that drives us all.”

Rueckert first learned how to grow polio viruses in large, pure amounts at the University of California, Berkeley, virology lab between 1962 and 1965. He found that they contained four protein chains, which was hard to figure. He had imagined there would only be one protein in something so small. How did they all fit?

“That was when I got interested in the structure of the virus,” Rueckert said.

He knew that a few proteins and plant viruses had been mapped in three dimensions by use of a method called crystallography. The proteins would be isolated, converted into crystals, then hit with X-rays for hours. Scientists would laboriously study and interpret the resulting photos for years before constructing a model.

Damage From X-rays

Animal viruses, however, seemed much too large and complex to tackle in this manner. They had to be X-rayed for so long that the crystals disintegrated before the job was completed. Even if scientists could collect the needed data, the task of making sense of it all would be overwhelmingly burdensome.

Rueckert knew, however, that new supercomputers were evolving that might be able to handle the computations. There were also special high-power machines that could X-ray virus crystals in minutes, rather than hours. Rueckert started to realize that, soon, scientists could do crystallography on the polio virus.

Rueckert’s own training was rooted in microbiology. He needed someone more familiar with the advanced X-ray techniques. That was when he met the graduate student, Hogle, who was then working in a crystallography lab on campus. Rueckert found Hogle to be bright and driven. Hogle grew intrigued with the older scientist’s ideas. Rueckert supplied him with polio virus.

The next year, Hogle began postdoctoral work at the Harvard University lab of Steve Harrison, the first crystallographer to determine a plant virus’ structure. When Hogle was far enough along to apply for a National Institutes of Health grant, Rueckert decided it was time to bring him back to Wisconsin.

Persuading a faculty dean to hire Hogle, however, would prove more troublesome than understanding the nature of viruses. Rueckert lobbied hard, but, as time dragged on, the university made no offer.

It was then that Rueckert traveled to the conference in Strasbourg and persuaded Rossmann to tackle the cold rather than polio virus.

The two men who found themselves drawn together there could not have been more different in nature.

Protective Assistant

Rossmann seems to dart rather than walk across a room. He rushes back to his lab most nights after dinner, protected by a possessive assistant who blocks phone calls and budgets his time.

Rueckert answers his own phone and enjoys long walks in the woods near his home. People who speak fast or in a high-pitched tone tend to get on his nerves.

But their knowledge complemented each other’s.

Rueckert moved in a world filled with cells and amino acids and genes. Rossmann, his training rooted in mathematics and physics, was one of the deans in the rarefied field of crystallography.

He had been second only to Harrison at Harvard in determining the structure of a plant virus. As a younger man, he had worked in the Cambridge lab of Max Perutz, who won a Nobel Prize for solving the structure of the hemoglobin protein. Rossmann had always felt he had not received proper credit for his work there.

As Rueckert recalls it, he at first could not persuade Rossmann to switch his attention from the polio to the cold virus.

“Michael wanted the glamour of polio. The argument that turned Michael, I think, came when I emphasized that the polio project at Harvard was not Steve Harrison’s. Michael was determined to show he was better than Harrison. After I explained it was Jim’s, Mike agreed to study the cold virus.”

Among those who soon signed on at Rossmann’s lab were two young postdoctoral fellows. Edward Arnold, 28, had trained in organic chemistry and crystallography at Cornell. Gerrit Vriend, 29, had studied biochemistry and computer science in the Netherlands. Rossmann was famous, they both figured, and his project looked interesting. This could be the place to launch their careers.

The team first needed to grow and crystallize viruses. This was not an easy task. Rueckert showed the Purdue team what to do.

Viruses will not multiply by themselves. They are non-living objects that can reproduce only by seizing a host cell. Inside the hollow protein shell of a virus is a small amount of genetic material which, when released inside a cell, commandeers the cell’s replicating apparatus, directing it to make duplicate copies of the virus.

The Purdue scientists began with a sample of rhinovirus 14 from Rueckert’s lab, one of the more than 100 known types of cold viruses. They added these to human cells so they would replicate, then broke open the cells to get at the multiplied viruses. These they purified and combined with a chemical that would draw out all water.

If everything went well, virus crystals would begin to grow within 24 hours. Within four days, they would expand to three-tenths of a millimeter, about the dimension of a black pepper grain and the size needed for X-raying.

Many Ways to Go Wrong

Quite often, however, everything did not go well. Sometimes, the cells died. Other times, the process produced fine “crystal showers” too small to be X-rayed. Once, after a series of failures, they spent days rechecking everything from the virus to the machinery before learning that the problem was with the building’s water supply.

“It’s not so much fancy stuff here at this stage,” said virologist Marcia Kremer. “But there are so many places where simpler things go wrong.”

Once he had virus crystals, Rossmann did not want to expose them to the conventional X-ray generator he had in his lab. A crystal would have to be hit with such rays for more than 18 hours and, by then, he reasoned, it would be quite dead.

He thought about leasing time on Cornell University’s High Energy Synchrotron Source, essentially a particle accelerator which provides a thousand times stronger X-ray than conventional generators. Rossmann was not sure whether sensitive virus crystals would hold up under such intense rays, but, if they did, a crystal that normally had to be X-rayed for 18 hours could be shot in four minutes.

Rossmann liked those numbers.

He and Eddie Arnold first visited Cornell in 1983. The data they collected then was not useful, for the crystals they brought were a disordered, inconsistent lot. They did, however, establish that they could get high-resolution photos.

The next year, in May, 1984, Rossmann and a team of six returned to Cornell with 100 crystals carefully mounted for X-raying. To make best use of the limited and expensive synchrotron time, they worked around the clock for five days, with three-man teams alternating on 12-hour shifts. They repeated this process again in June and November.

Each photo produced by crystallography appears essentially as a circle containing many dots. Clear distinct dots are, to a crystallographer, a thing of beauty. By the end of the year, Rossmann had them.

The most complex task of all, however, lay in front of the team. They had to convert those dots back into the three-dimensional virus from which they originated.

“What we start with and what we want is the virus model,” Vriend said. “The process takes us to those dumb dots, unfortunately. So then we have to do all sorts of things to the dots. If you do it all correctly, you get what you started with. We are undoing what our own experimental process did to us.”

The scientists knew that those dots represented waves of light from X-rays that had scattered, or refracted, as they passed through the crystal. If they could properly measure those refracted waves, they could begin to figure out the atomic structure of the virus crystal that had scattered them.

Using an optical film scanner, they calculated the degree of brightness and darkness of the many dots. From this, they could determine the nature of the light waves.

No single shot, however, can yield a portrait of the virus. The film is flat and the crystal is three-dimensional. Moreover, a single shot can involve no more than 40,000 dots--otherwise the photo would simply be all black--while the team needed 580,000 distinct dots to generate their virus model.

So each shot represented only a slice of the virus, reflecting the angle and position of the crystal in relation to the X-ray. To form a complete picture, the scientists needed to combine the data from 86 separate shots.

First, they had to assemble as many waves as possible and place them in the proper position relative to one another. This difficult procedure, called phasing, confounded even some of the scientists. Those who knew the workings of the cell gave way to those who grasped the complexities of higher physics. Yet, phasing alone did not provide enough information for, at best, the scientists could match up less than one-third of the needed waves.

Many Doubted Theory

In the end, the fate of the Purdue team’s years of work hinged on a controversial theory that Rossmann had first advanced as a young man. His paper, published in January, 1962, in an arcane journal called the Acta Crystallographica, had attracted many doubters. Rossmann never forgot the reaction and never let go of his desire to prove them wrong.

If you know that you have a 20-faced symmetrical object such as a virus, Rossmann in essence had reasoned, you do not need all the data. You can extrapolate the information from a subsection to the rest of the object.

“Michael had predicted it would happen,” said Vriend. “Scientists simply lacked the computer power and courage to start it.”

The Purdue team had no alternative. In early March of this year, they turned to the university’s Cyber 205 supercomputer. In essence, they asked it to average what data they did have and apply that to the whole virus.

The result provided a few more matched waves and a clearer map, so they pressed on. Forty times they asked the computer to average and reconstruct the map.

What followed, for the scientists, was wondrous. On the screen, the computer began to build, atom by atom, a three-dimensional structure. Tubes, arms and hollows took shape.

Arnold could, in some early images, see the canyons and ridges of the virus. He could not yet see the protein chains fully resolved. For the scientists to be right, everything would have to match and fit together--structure, proteins, chemical formulas.

Arnold knew that their type of scientific pursuit often reached a dead end after years of labor. Still, he could not help but feel excited.

“It was like looking at a beautiful picture from far away,” he said, “and then from closer and closer.”

In Wisconsin, meanwhile, Rueckert and a graduate student, Barbara Sherry, were not sure they were heading anywhere with their part of the inquiry.

Since 1982, they had been working to identify where human infection-fighting antibodies attach themselves on the virus’ protein chain. Such sites are called antigenic binding sites.

They had observed that all the antibodies grouped at one of four places, so they knew there were four antigenic sites.

To learn precisely where those sites were situated on the virus protein chain, the scientists then looked at mutated viruses that were resistant to the antibodies.

Whatever had mutated in those viruses was preventing the antibodies from attaching and doing their job. So the scientists knew the mutations must be at the sites where the antibodies normally attach. If they could find those mutations on the protein chain, they would have found the four antigenic sites.

The only way to find those mutations was to compare the lengthy protein chains of normal viruses with those of the mutated, resistant viruses and look for the minute differences in their sequences. The limits of their science and their resources required that the Wisconsin team do this indirectly, by studying the nucleotide chain in the virus gene that codes for the protein. The laborious task took three years.

When they finished last January, Rueckert and Sherry had found the mutated sites.

Researchers Dismayed

What they came up with, however, did not make sense. Some mutations sat off by themselves, much farther down the protein chain than where the rest of the mutations were clustered, and where antibodies appeared to bind. Rueckert and Sherry were dismayed.

“They should have been together, where the antibodies indicated they were,” Sherry said. “We drafted a paper reporting this unsolved problem and were about to publish it. We never did, though.”

Before they could, the phone rang in their lab one morning. Rossmann was calling. He had his breakthrough.

From early March to early April, the Purdue team had harnessed 40% of the supercomputer’s capacity, feeding it 6 million distinct pieces of data. Without the supercomputer, the month of calculations would have taken 10 years.

This was not the science of old. Scientists not that long ago used little more than logic and chemical formulas to imagine what a gene or protein looked like. Then they built oversized structures with twists of wire and plastic, much as a youth assembles a model airplane, studying and adjusting the parts until the triumphant moment when everything seemed to fit.

Now, the supercomputer did much of that work. As monumental a task as any for the scientists was to design the programs that guided the Cyber 205.

On Saturday night, April 6, Arnold asked the computer to plot out the map as it then stood. This the computer did in a series of separate frames, for the image was too complex to integrate into one snapshot.

On Sunday, Arnold printed out the series of separate frames and converted them into transparencies. He then stacked these transparencies one on top of another.

If they contained a complete and accurate map of the virus, they would all match. As the scientists looked down through the pile of transparencies, they would see a single unified image, with the structural borders and protein strings forming continuing lines rather than overlapping. The protein would be discretely resolved, and the mathematical and chemical formulas would fit with the structural shape.

Everything would make sense.

Just as likely, though, they would see nothing. For most of the team, that would mean they had wasted three or four years. For Rossmann, it would mean that for 20 years his vision of the biological world had been fundamentally wrong.

When the team gathered here in their basement lab in Purdue’s molecular biology lab at 4 p.m. Sunday, Rossmann said later, “I frankly was very skeptical. There had been so much doubt about my molecular replacement idea I thought maybe they were right. This was just supposed to be a diagnostic test to make sure we were not up a dead end. I was just seeking encouragement to continue. In fact, we had plans to go back to Cornell the next Thursday to shoot more crystals.”

Vriend was the first to look down at the stacked transparencies. He did not know what to think. “I had never seen such a map before. I thought, well, that looks nice.”

Then Rossmann looked. He said nothing at first, preferring to study what was before him. After four years, a few more minutes did not seem to him a long time. He knew, though, what he was looking at, for he had worked many times with such transparency maps. He had it. The job was done.

“This is the most beautiful map possible,” he wrote on the blackboard.

If there was any doubt, the team had only to look at Rossmann. He was dancing on the conference table. Soon after, team members and their spouses were down the street at the China Palace restaurant, pouring champagne.

One last critical step remained before the four-year pursuit of the cold virus could be considered completely successful. The Purdue team’s model had to match with Rueckert’s and Sherry’s information about antigenic sites, or something was fundamentally wrong.

Both teams in essence had been working only with parts of the whole. The Purdue team had a structure that looked good but existed in a vacuum, as an object to marvel at. The Wisconsin team had biological insights that tapped particular secrets of the virus but could not explain it as a complete entity.

On May 1, Rueckert and Sherry arrived at the Purdue lab to compare findings.

At the start of that morning, they knew only the locations of their antigenic sites on a linear chemical protein chain. They had no idea where they were in the virus’ three-dimensional structure. Most important, they had no explanation for their oddly situated mutations.

“They hadn’t sent us data because they didn’t trust what they had,” Rossmann said. “I expected their data not to fit. There was tension as we looked at each piece of data. I didn’t want to see, but I had to.”

Point by point, they slowly checked what they had.

When Rueckert’s and Sherry’s four antigenic sites were matched with Rossmann’s structure, the sites turned out to be just where they should be--on the elevated ridges above the canyons. The scientists could see where antibodies bind to the virus.

Tension gave way to excited cheers. “It was like when you laugh so much your sides hurt,” Rossmann said. “I just felt I couldn’t take any more happiness. I had to leave and walk around.”

As they looked longer, it became clear why Rueckert and Sherry found mutations far from where they should be on the protein chain. In the three-dimensional structure, they saw that the protein chain was folded and curled over itself in such a way that the two distant spots actually were positioned together.

This time, it was Barbara Sherry who danced on the conference table.

“It was not just a confirmation, it was the ground-breaking next step,” said Arnold. “Without this, we would see a structure but not know what to make of it. We would just know it looked sort of like a plant virus. Our work was a monument. Their work made the rhinovirus 14 come to life.”

Rueckert considered the day one of the most exciting moments of his life. “When you get a breakthrough like that, all sorts of things make sense. The feeling that you can explain everything is very appealing. You feel like you can control the universe.”

They could not, however, control the common cold virus.

As they gazed at their three-dimensional model, they could do no more but begin to understand why the cold virus has been such a hearty survivor.

Scientists had hoped for years that they might find a structural feature common to all of the more than 100 types of cold viruses. If they did, they might be able to develop a vaccine that could produce antibodies to attack that common feature.

Rossmann, studying the model, came to believe there indeed was a common feature--the apparatus with which the virus grabs onto a host cell. However, because this apparatus was at the bottom of the viruses’ canyons, where the antibodies cannot fit, nothing could be done with the common feature.

Antibodies instead attach to the ridges atop the canyons. Those ridges differ for each cold virus and regularly change by mutation, so no one antibody can neutralize all of them. In fact, the body needs a different antibody for each of the more than 100 cold viruses.

This makes the cold virus a much tougher opponent than the deadly polio virus, which does not often successfully mutate.

“The cold virus won’t let the antibody into the canyon, so, instead, the antibody binds to the ridge,” Rossmann said. “Then the cold snarls and says, ‘OK, you beat me’--and changes its ridge . . . . Cold viruses have taken a long time to fine-tune themselves. It’s interesting how random selection creates something of a higher order.”

Although they did not see possibilities for vaccines, the scientists, looking at their model, did see the chance for drugs to fight colds.

A drug might be able to cover the site on the human host cell to which the virus attaches, preventing infection.

The virus has to shed its protein shell when it enters the human cell. Then, in order to leave a cell and infect another one, the virus must assemble a new protein coat. Understanding the structural changes involved might suggest a way to block those processes.

Models of Other Viruses

Scientists believe also that determining the cold virus’ structure will affect their pursuit of other diseases. The success of Rossmann’s molecular replacement theory could rapidly enable scientists to generate similar three-dimensional models of other viruses and begin to understand their workings.

“The method is now known to work, so we can apply it to other groups of viruses,” Rossmann said.

However, when the Purdue and Wisconsin teams published a paper reporting their breakthrough in the Sept. 12 edition of the magazine Nature, not everyone felt an unqualified thrill.

The drive for knowledge is often as much a contest of man against man as it is scientist against a problem. The search for personal recognition at times yields knowledge as a side product. However, scientists’ ideas feed and fuel off one another in a way that makes assigning credit a confused matter involving vague codes.

Jim Hogle, working with polio in La Jolla, had expected to be the first to produce a three-dimensional map of an animal virus. Such a feat by a young man of 34 would have startled the scientific community.

Lacking Purdue’s more high-powered resources, he had been using a conventional X-ray generator and a relatively commonplace Digital minicomputer. His team consisted of one other person, David Filman, 36. But they had taken pride in creating special calculations that enabled them to sort their data efficiently and use one-sixth of the information processed by Rossmann’s team. “When you have a supercomputer, you don’t have to think of a new idea,” Filman said.

Hogle could not help but feel some dismay at Rossmann’s triumph. “I did feel some anxiety. I care more about science than glory, but the glory gets you a bigger lab and more funding and attracts good people. So it does matter.”

Rossmann had sent Hogle an advance copy of his article, an act that, intentionally or not, served to lay first claim. There was talk of announcing both projects jointly, but Rossmann rejected the idea.

When Hogle completed his model of the polio virus in mid-July and published his report in the Sept. 27 edition of Science--just two weeks after Rossmann’s paper appeared--he did not credit the Purdue scientist’s molecular replacement theory. Rossmann felt considerable disappointment. Hogle thought his work’s lineage instead traced back to his Harvard mentor--and Rossmann’s competitor--Steve Harrison. “You have to look back to where the ideas began,” he said. “There are different ways of looking at the history of it all.”

At a conference that Rossmann and Hogle attended recently in Europe, the feelings of resentment were palpable.

This troubled turn of events left Rueckert with a feeling of ironic detachment. He had launched both men on their way but was now watching from the sidelines as they strove for recognition. He saw no simple or orderly resolution to the squabble, such as those that gave him comfort in the natural world. Nor was he inclined to choose sides. He was of two minds about this dimension of the team’s four-year-long project.

“Some people think I’m a fool, the way I help others and don’t maneuver for credit,” he said recently, sitting in his lab. “It’s just that I don’t want to be famous. I have my woods to walk in. But, on the other hand, science depends on driven people feeding on each other’s ideas. That’s what it’s all about. That’s what makes science work.”

Rueckert was even more struck by the general public’s apparently unbounded interest in the cold virus. He felt bemused, knowing how inadvertently his team chose that target. Such a common tormentor had more glamour than the polio virus after all.

“I’ve gotten letters from all over. Everyone wants to talk about the cold. There was a man who wrote to say he fell asleep in a very chilly and humid chicken coop. When he woke, the bad cold he had had for six months was gone. He wanted to know how I can explain that by the virus’ structure.”

Rueckert laughed and shrugged.

“I had no idea, of course. This is a major fundamental achievement, but it doesn’t cure the cold. Biology is much more complicated than people think.”