News this Week

Science  24 Sep 1999:
Vol. 285, Issue 5436, pp. 2038
1. HUMAN GENOME

Team Wrapping Up Sequence of First Human Chromosome

1. Dennis Normile,
2. Elizabeth Pennisi

TOKYO—While the Human Genome Project races to finish a rough draft of the 3 billion bases in our DNA by next March, three sequencing teams are about to reach a different, potentially more significant milestone: a final draft of the first human chromosome. Sometime within the next week or two, the international consortium sequencing chromosome 22 will conclude that it has done everything possible to complete the sequence. As the team approaches that landmark, it is also setting precedents for those whose work on the rest of the human genome is scheduled to be finished by 2003, including a definition of what constitutes success.

The official announcement about chromosome 22 is not expected until early November, timing that coincides with both a conference in Tokyo and publication in a peer-reviewed journal. The scientists involved—from the Sanger Centre near Cambridge, United Kingdom; the University of Oklahoma, Norman; and Keio University School of Medicine in Tokyo—say theirs has been a model of international cooperation. “A lot of people have put a lot of blood, sweat, and tears into this,” says Oklahoma's Bruce Roe. “This has been international cooperation at its finest.”

Chromosome 22 will be the first across the finishing line primarily because, at roughly 53 megabases (53 million base pairs), it is the second-shortest chromosome. (Chromosome 21 is slightly shorter, but its sequencing is progressing more slowly.) Another factor has been the right mix of partners. The Sanger Centre, which can churn out a lot of sequence data in a short time, took on the seemingly straightforward half of the chromosome, while the two smaller groups tackled regions, such as those with numerous repeats and few genes, that could have slowed down the Sanger team.

Although part of the international Human Genome Project, the chromosome 22 consortium had enough of a head start over other groups that it was unaffected by the project's decision last year to concentrate on a rough draft first. Before human genome sequencing efforts were launched in earnest, chromosome 22 had already been the focus of a fair amount of mapping and even some sequencing work. Ian Dunham, who now heads Sanger's chromosome 22 group, was part of a team at Guy's Hospital in London that in the early 1990s used chromosome 22 to develop sequencing tools and techniques. By 1995 the team had moved to the Sanger Centre and completed a map of the chromosome.

At roughly the same time, a group at Keio led by Nobuyoshi Shimizu had completed what was then the largest complete human contig, a map depicting the relative order of a linked library of small overlapping clones. The contig covered a 1.02-megabase stretch of chromosome 22 containing the immunoglobulin-l gene cluster, which is involved in human immune response. In late 1995, Shimizu set up a sequencing effort for that region and for another region implicated in cat eye syndrome, which can result in congenital heart defects and mental retardation. And the Oklahoma lab got involved in sequencing chromosome 22 when one of Roe's graduate students, Stephanie Chissoe, sequenced a 152,000-base stretch of DNA that included the Bcr gene, which has been implicated in certain forms of leukemia.

From those humble beginnings, says Dunham, the chromosome 22 consortium “just sort of evolved,” with the trio finally deciding to tackle the entire chromosome. “We got together and divided it up into regions that we had reagents for,” Roe recalls. The effort also included other labs in Europe, North America, and Japan that helped map the chromosome and provided clones, reagents, and other material for the sequencers.

The sequence actually covers only the lower arm, the so-called q region, of the chromosome. It's roughly 32 megabases long and contains almost all of the chromosome's genes. The upper arm, called the p region, was ignored because it doesn't seem to code for proteins. The consortium also skipped the telomere—the tail end of the arm—and most of the centromere—the “waist” of the chromosome that separates the two arms. These two regions contain few genes and are very difficult to sequence.

The last bits of sequence were the most difficult. “We decided we were almost finished [last spring],” says Shimizu, “but then it took 6 months to actually finish.” For reasons that are not completely understood, the bacterial clones that researchers depend on to produce the DNA needed for actual sequencing don't retain certain human sequences. This led to an exhaustive and frustrating search through clone libraries in hope of finding a clone that would cover a particular gap.

The group succeeded in filling some of the gaps, but nine small gaps remain that Dunham says “seem to be unclonable.” “I'll be happy when we put the whole thing to bed,” says Roe, “but I wish we had [my] two gaps closed. We really want to get it done, done, done.”

The consortium's experience with chromosome 22 has helped the leaders of the Human Genome Project decide on a definition of “finished” that likely will be applied to sequencing of the remaining human chromosomes. Earlier this month, at a meeting in Cambridge of the international partners involved in this massive effort, the partners reached a consensus on what's needed. The three major criteria are: More than 95% of the chromosome must have been sequenced; the number, location, and size of remaining gaps must be pinned down; and individual gaps must be shorter than about 150,000 bases.

Dunham says the criteria may never be “written in stone,” but the chromosome 22 consortium is already well within these criteria if the sequence data “in the pipeline” count, along with what has been posted in databases. What remains, he says, is to “check everything” and make sure that the entire sequence is correctly labeled and deposited in a public database. Completion of that task will be a signal to pop the champagne corks. But the celebration will be brief. “Mapping and sequencing has already taught us a lot about the nature of the genome,” Shimizu says. The next step, he says, will be to clarify the biological significance of it all. And that work has barely begun.

2. SCIENCE PUBLISHING

Cell Editor Steps Down

1. Marcia Barinaga

Benjamin Lewin, the editor of Cell and its sister journal Molecular Cell, announced to his staff and editorial board last week that he plans to retire on 1 October. His sudden departure represents “a big loss for Cell,” says Cell biologist Tony Hunter of the Salk Institute for Biological Studies in La Jolla, California.

After Lewin founded Cell in 1974, it quickly became a premier journal of molecular and cell biology. Scientists attribute the journal's success largely to Lewin's depth of scientific knowledge and his hands-on management style. “It will be very different without Benjamin there,” says Hunter, who has been on the journal's editorial board since 1980. “He was always there to talk with you about your paper or someone else's. This was in contrast to most other journals.”

Lewin sold the journal, along with its three sister journals—Neuron, Immunity, and Molecular Cell—to Dutch science-publishing giant Elsevier Science in April, for an amount rumored to be close to $100 million. Insiders wondered how long Lewin would stay on, although Elsevier had announced that he would remain editor for 5 years. When reached by Science, Lewin declined to comment. Some close to the journal fear that Lewin's departure, combined with Elsevier's takeover, will trigger an exodus of staff. But Deputy Editor Vivian Siegel will stay on at the helm, and editorial board member Ira Herskowitz of the University of California, San Francisco, expresses “absolute faith” in her. Siegel, he says, shares Lewin's engaged management style, but he expects Cell to “evolve in some manner. She is not a clone of Ben.” 3. SCIENCE PUBLISHING NEJM Publisher Resigns 1. Constance Holden The publisher of The New England Journal of Medicine, Joel Baron, quit his job less than 2 months after former Editor-in-Chief Jerome Kassirer was forced out. In a 13 September letter to colleagues, Baron said that after an expansionary push in which the journal's owner, the Massachusetts Medical Society, launched several new publications, he was ready to move on. Under Baron's 2-year tenure, the society has started up new publications such as Heartwatch, a consumer newsletter, and acquired Hippocrates, a journal for physicians. It has also been looking into lucrative arrangements with commercial publishing enterprises. Now that things have quieted down, Baron, who calls himself a “strategist” rather than an “implementation” person, said in the letter, “I think I will be able to make a bigger contribution elsewhere.” (Baron couldn't be reached for comment.) Some observers suspect that in the turmoil following Kassirer's dismissal, Baron no longer had a free hand to do what he was hired to do. Kassirer was pushed out because of “differences of opinion” with the medical society over activities that he claimed would compromise the journal's good name (Science, 30 July, p. 648). “There's been enough concern expressed by editors of the journal and the academic community” over the new publications initiatives that management may have decided to shelve its plans for the present, says NEJM Associate Editor Morton Swartz, former chief of infectious disease at Massachusetts General Hospital in Boston. Taking it one departure at a time, the society last week announced the appointment of a search committee, headed by Harvard Medical School professor Ronald A. Arky, chair of the medical society's publications committee, to look for Kassirer's replacement. 4. AIDS THERAPY Ambitious Clinical Trial Stirs Debate 1. Cohen Jon The National Institute of Allergy and Infectious Diseases (NIAID) last week decided to fund what will likely be the largest and most expensive trial of an AIDS treatment the institute has ever backed. During the next 5 years, the$43 million study will follow 4000 HIV-infected people who are already taking anti-HIV drugs to see whether adding an immune-system messenger called interleukin-2 (IL-2) can help prevent disease and death. The study, which went through a stringent but unusual review process because NIAID director Anthony Fauci holds a patent on the treatment, will involve 210 sites in 18 countries, creating an enormous new clinical trials network that will include some of the world's leading AIDS clinicians. “It's tremendously ambitious,” acknowledges Jack Killen, head of NIAID's Division of AIDS. “But this is about as good a shot as we're going to get at answering a very important question.”

Whether the so-called Esprit trial is likely to yield meaningful results is, however, being fiercely debated within the AIDS research community. Some researchers believe its flexible design and relatively healthy subjects may blur any results. And numerous other logistical, procedural, and ethical questions have also dogged this trial since it was first conceived 3 years ago, including whether the costly study is needed when smaller IL-2 trials are already planned in sicker subjects.

Small-scale studies have shown that genetically engineered IL-2 significantly boosts levels of CD4 cells in HIV-infected people. (CD4s are the main white blood cells that HIV selectively destroys.) “We have seen changes in CD4 counts, but we don't know what they mean clinically,” explains NIAID's clinical director Clifford Lane, who pioneered this treatment strategy and shares the patent with Fauci and NIAID's Joseph Kovacs. (The patent is assigned to the government, and Chiron, the maker of engineered IL-2, has a license; the researchers are entitled to a maximum of $150,000 of any payments each year, which Fauci donates to charity.) Specifically, none of the trials have yet shown that the CD4 increases result in longer, healthier lives, and the treatment does not seem to decrease the amount of HIV in a person's bloodstream. The trial aims to mimic the diverse ways IL-2 would be used in the real world. In the first 6 months, 2000 people already taking any combination of anti-HIV drugs will give themselves injections of IL-2 for 5 days every 8 weeks. After those three cycles, physicians will use their discretion to determine the frequency of subsequent cycles of IL-2 treatment, which can cause flulike symptoms. Another 2000 people who are taking only anti-HIV drugs will serve as the control group. “This will give some pretty clear information,” asserts Lane. Others aren't so sure—including the peer-review group that ultimately gave the trial a thumbs-up. “A lot of people are skeptical about whether it will be possible at the end of a large trial like this to sort out the cause and effect when people cycle through different treatments,” acknowledges Killen. And the link may be further blurred because Esprit will recruit people who have suffered relatively modest immune damage from HIV and thus are more likely to respond to the immune booster; to be eligible, HIV-infected people must have at least 300 CD4 cells per millimeter of blood at the trial's start. (The normal range is 600 to 1200.) As a result, it may take longer than 5 years to see enough AIDS-related disease and death to determine conclusively whether IL-2 helps. “You could be holding your breath a long time,” says Robert Schooley of the University of Colorado Health Sciences Center in Denver, who heads the AIDS Clinical Trials Group (ACTG), an NIAID-supported network that conducts most trials of AIDS drugs. The fact that ACTG will not be running this trial is another point of contention. James Neaton of the University of Minnesota, Minneapolis, a biostatistician who is Esprit's principal investigator, says he couldn't interest ACTG. “People I worked with [in the ACTG] wanted to participate, but they couldn't get approval from the executive committee,” says Neaton. Schooley explains that not only would the expense overwhelm the ACTG's budget, but ACTG is already conducting a smaller scale trial of IL-2 in sicker patients. “Our feeling is it really offers more to people with advanced disease,” he says. (Indeed, Chiron last month launched a large efficacy trial of the treatment in people with 50 to 300 CD4s.) Schooley also questions whether patients with relatively high CD4 counts will choose this toxic and expensive drug. “If you have 700 CD4s, you're going to do well for a long time,” says Schooley. Neaton also considered another NIAID-sponsored clinical trials network, the Community Programs for Clinical Research on AIDS (CPCRA), but it did not have enough sites to recruit the needed number of patients. So, on advice from NIAID, he turned to a mechanism that is rarely used to fund large clinical trials: He submitted an investigator-initiated, “R01” grant. To help avoid the perceived conflict-of-interest issues raised by Fauci's patent, the ad hoc “study section” of peers set up to evaluate the proposal was convened by the National Cancer Institute, not NIAID. “I can tell you for a fact that Tony had no influence whatsoever on the process of review or the decision about the funding of this,” says Killen. Lane, who did help design the trial, says National Institutes of Health lawyers gave him a waiver for that purpose. Unlike ACTG and CPCRA, study sections evaluate proposals behind closed doors. But interviews with members of the study section and documents provided by Neaton suggest that it received a rigorous review. When the study section first evaluated the proposal in June 1998, it gave it a score of 322, which put it in the unfundable 59.3 percentile. Among the many concerns listed by the study section was “whether the study as currently designed will result in interpretable data.” Another sensitive topic was that some testing would be done in poor countries that may not be able to afford IL-2 if it is found to work: Three treatment cycles cost at least$5000.

The same study section considered a revised proposal 8 months later, and although it still noted “several weaknesses,” it deemed the trial design “significantly improved.” This time around, it earned a priority score of 178, placing it in the 18.1 percentile, an excellent ranking. On 16 September, NIAID formally announced that it would fund the trial. Leading clinicians in the United States (including those with the CPCRA), Canada, Europe, Australia, Greece, Israel, Thailand, and Argentina will participate.

Even if the treatment itself doesn't pan out, Neaton thinks the trial could have an important benefit: to show how “large, simple trials” that mirror real-world situations can answer tough questions. “I'm hoping to make this a model for how other research is done,” he says.

5. IMMUNOLOGY

On the Way to a Better Immunosuppressant?

1. Michael Hagmann

“Kiwi” Clint Hallam, the world's first hand transplant recipient, celebrated the first anniversary of his new hand yesterday. He is living proof that organ transplantation has come a long way since the first patient received a donated kidney back in 1954. But in spite of milestones like Hallam's, swapping body parts between different individuals has a major downside. It challenges our immune system's very raison d'être—fighting off anything recognized as foreign, whether invading pathogens or potentially life-saving organ grafts—and requires that the patients be given powerful immunosuppressive drugs. Now, a molecular smart bomb that targets a key molecular event in graft rejection could lead to improved immunosuppressants with fewer side effects.

In the early 1990s, immunologists discovered that when a foreign antigen—on a graft, for example—triggers the immune system's T cells, an intracellular enzyme called calcineurin is activated. Calcineurin clips chemical tags called phosphate groups from other proteins, including the so-called NFATs. When that happens, these proteins switch on a variety of genes encoding proteins that rev up immune cell activities.

Current immunosuppressive drugs such as cyclosporin A (CsA) or FK506 work by blocking calcineurin, but they often have nasty side effects such as kidney failure, diabetes, and an increased risk of cancer, presumably because they inhibit calcineurin's ability to remove phosphates from other proteins besides the NFATs. The new compound, which molecular immunologist Anjana Rao and protein chemist Patrick Hogan of Harvard's Center for Blood Research in Boston and their colleagues describe on page 2129, inhibits calcineurin more selectively and may cause far less of this “collateral damage.”

The compound has only been tested in cultured cells and may not be suitable for clinical use, but if something like it eventually hits the pharmacy, it could be a boon to transplant recipients and other patients. “The pharmaceutical industry spent hundreds of millions of dollars to come up with a ‘better cyclosporin,’ but so far it all amounted to zero,” says Ernest Villafranca, a protein crystallographer at Agouron Pharmaceuticals in La Jolla, California. A more selective immunosuppressant, he adds, might be used to treat conditions caused by an overactive immune system, such as asthma or autoimmune disorders like multiple sclerosis, for which physicians hesitate to prescribe current drugs because of their dire side effects.

Rao, Hogan, and their colleague José Aramburu took a first step toward the design of a more selective calcineurin inhibitor last year. They identified the site on NFAT that docks with calcineurin and showed that a short protein snippet, or peptide, from within that region can block the docking and activation. The peptide presumably acts by binding to calcineurin and, in effect, clogging its NFAT binding site. In the current work, the researchers have come up with a more powerful inhibitor.

In collaboration with Lewis Cantley of the Harvard Institutes of Medicine they synthesized a “library” of around 1 billion peptides in which they varied several amino acids of the natural NFAT sequence and then selected the peptide that bound most tightly to calcineurin. This synthetic peptide, the researchers found, is 20 to 100 times more potent than its natural counterpart at blocking NFAT activation by calcineurin—and most important, because the peptide doesn't block the catalytic center, it does this without affecting calcineurin's action on other proteins. “This means that if you've got the right wedge you can pry the two proteins apart,” says Hogan.

Next, the team probed the effects of the peptide on gene activation in T cells. They found it was much more specific than CsA, turning off only genes thought to be NFAT-dependent, such as several genes for the immune system messengers called interleukins. Everyone agrees that much more work will be required to turn this promise into a drug, however. “There are still quite a few hurdles to overcome,” admits Hogan. For one, although the tests with cultured cells are encouraging, the researchers haven't yet demonstrated that the peptide is in fact immunosuppressive. They plan to address that question in animal studies.

But even if the new inhibitor passes that test, it still may not be suitable for use in therapy. “Peptides are really not practical to use clinically,” says Villafranca, because they are chopped up very fast in the body and often have problems getting inside the cell. But researchers might be able to find nonpeptide drugs with similar effects. “Pharmaceutical companies could simply take the assay with which we found the peptide and screen hundreds of thousands of compounds, for instance in bacterial broth,” says Hogan.

Stanford molecular biologist Gerald Crabtree considers that a reasonable approach. “The only good immunosuppressants so far, CsA and FK506, have both targeted calcineurin. … Maybe nature has already told us where the [immune system's] Achilles' heel is.”

6. MUSEUM MANAGEMENT

Smithsonian Taps Banker as New Leader

1. David Malakoff

A banking executive with a flair for flamenco guitar and a passion for the arts will lead the world's largest museum and research complex. The Smithsonian Institution's Board of Regents announced last week that Lawrence M. Small, 58, second in command at the Fannie Mae mortgage corporation, will succeed retiring Secretary I. Michael Heyman in January.

Small's appointment breaks a 150-year tradition of naming a scientist or academic to lead the sprawling $570 million institution (see box). Some Smithsonian scholars are disappointed that they won't be working for one of their own, but many hope their new boss's management skills and boardroom connections will help inject cash into stagnating science budgets. “We're making a paradigm shift in leadership; it's clear the regents place a high value on corporate management skills,” says Chris Wemmer, who directs the Smithsonian's Conservation and Research Center in Front Royal, Virginia. The Smithsonian is best known as one of the United States' foremost tourist attractions, having lured 30 million visitors last year to its zoo and 16 museums, many of which line Washington, D.C.'s grassy mall. Less visible are the institution's more than 600 scientists, who toil amid world-class collections of everything from spiders to gems and conduct studies at seven far-flung research institutes, including the Harvard-Smithsonian Center for Astrophysics in Cambridge, Massachusetts, and a tropical research institute in Panama. View this table: Over the last decade, some Smithsonian scientists say their influence—and funding—has waned as the institution expanded and spent aggressively on new art galleries and flashier exhibits. “There has been a slow but inexorable shift away from scholarship and toward public entertainment,” believes spider expert Jonathan Coddington of the National Museum of Natural History. He and others worry that the perceived trend—which some have termed “Disneyfication” and others say is “dumbing down”—could undermine the institution's scientific prowess. Small says he has no desire to see research or educational outreach suffer under his tenure, which could last a decade or more. “I am more than aware that [Smithsonian founder] James Smithson's estate was bequeathed for the ‘increase and diffusion’ of knowledge—and that there was a point to putting ‘increase’ first,” he told Science. There is “no great educational institution where research is not also considered enormously important,” he says, adding that although “being a scientist is a wonderful way to spend your life, it doesn't necessarily prepare you for leading a large organization.” His own résumé includes 27 years at Citicorp, eight at Fannie Mae, and dozens of appointments to nonprofit boards. That background, observers say, should equip him well for the job of streamlining the Smithsonian's bureaucracy, a task Heyman began that earned him high marks from some researchers. Small will also have useful connections when it comes time to raise money, whether lobbying Congress for the 70% of the Smithsonian's budget that comes from taxpayers or seeking corporate gifts for a major upcoming capital campaign. Indeed, one of the new secretary's most valuable assets “may be his Rolodex,” jokes one researcher. But Small says he also has a lighter side, pointing to his year in Spain studying flamenco guitar and a number of other “scholarly passions.” Small will get plenty of suggestions on how to spend any new funds. Many researchers, for instance, would like to see more money for graduate stipends and postdoctoral fellowships, noting that those budgets have eroded dramatically since the 1980s. The number of fellowships available at the Natural History Museum has slumped from more than 20 a decade ago to about five, researchers note, while the astrophysics center now has just two junior slots available annually on a staff of 300 Ph.D.s. “We are also falling behind in our computing capability,” notes Irwin Shapiro, director of the astrophysics center. For the moment, however, Small is keeping quiet about any plans he has for the organization. Indeed, he notes that he will have his hands full in the short run just finishing off two new projects—an annex at Virginia's Dulles Airport for aerospace exhibits too large for the existing facility on the mall, and a new Native American museum—that were begun by his predecessor. 7. ASTRONOMY Italian Observatories to Form National Institute 1. Alexander Hellemans* 1. Alexander Hellemans writes from Naples, Italy. NAPLES, ITALY—Small organizations usually resist being subsumed into a larger one. But for astronomers at Italy's 12 observatories—all independent but funded directly by the government—their amalgamation into the new National Institute of Astrophysics (INAF) can't come soon enough. They argue that the current fragmentation of astronomy in Italy is hampering their ability to play in the international big leagues. “The foundation of this national institute has long been a wish of all Italian astronomers. They want to be able to compete at an international level, which was not possible with small entities like the observatories,” says Marcello Rodono, director of the Observatory of Catania. The 12 observatories, which employ about half of Italy's 700 astronomers, control their own budgets, choose their own scientific programs, and hire their own researchers. In addition, Italy boasts eight astronomical research institutes run by the National Research Council (CNR), and several universities have astronomical facilities. Without some form of central body to manage this scattered enterprise, the observatories have found it hard to work together on large national and international research projects. “It was rather difficult to start national projects that would imply big expenses,” says Rodono. “With a [national] institute, you can plan the extra money for the years to come and be sure that this will be allocated for the projects.” The government approved the new institute in July and is expected to appoint a president and two board members in November. Four more board members will be elected nationally, two from the observatories and two from universities. At some point, the research ministry may also transfer the CNR's eight astronomy institutes to INAF, says Rodono, who expects that INAF will become fully operational next summer. INAF headquarters will be in Rome, although technical facilities may be located on La Palma in the Canary Islands. The total budget of INAF will be at least$54 million, the sum of the budgets of the 12 observatories.

Massimo Capaccioli, director of the Capodimonte Observatory in Naples, says the new institute should improve the management of projects such as the recently completed Galileo Telescope, the national 3.5-meter telescope on La Palma; the Large Binocular Telescope, two joined 2.84-meter scopes now under construction at the University of Arizona, in which Italy has a 25% share; and a 2.65-meter survey telescope that the Capodimonte Observatory is building with the European Southern Observatory to aid its Very Large Telescope in Chile. INAF may also give a push to plans for Italy to join Spain in the construction of a replica of the Keck 10-meter telescope to be located on La Palma. The next few years will be a “very critical phase,” says Giancarlo Setti of the University of Bologna.

8. EARTHQUAKES

Prediction Claims Stir Greek Controversy

1. Richard A. Kerr

Bucking any scientific consensus can be rough, but insisting that you can predict earthquakes in a quake-prone country like Greece—when practically no one thinks it can be done anywhere—is sure to create a fuss. For almost 2 decades, a group of Greek scientists has claimed they could predict damaging earthquakes by monitoring electrical currents in the ground. Although many Greek colleagues have questioned the scientific rigor of the method (called VAN, after the initials of its inventors), a run of seeming successes in 1995 caught the attention of researchers outside of Greece (Science, 10 November 1995, p. 911). Now, after a lull, the VAN scientists are making new claims. They say the ground gave clear warning signs of the 7 September earthquake that struck near Athens, killing 67, and they think they may have picked up signs of another, perhaps larger, temblor in the offing.

These claims are meeting with scorn, especially in the Greek scientific community. “This has nothing to do with seismology or science,” says Leonidas Resvanis, director of the Physics Laboratory of the University of Athens. Adds Gerry Chouliaras, a seismologist at the National Observatory of Athens: “There's no scientific reason to make this alarm. I don't believe their ‘signals.’ I'm not going to believe anything.” This rancor has emerged over years of frustration with VAN, explains seismologist Robert Geller of Tokyo University. Outsiders must compare vague predictions made on the basis of ill-defined criteria against the earthquake record, he says, while being denied access to the full VAN observations.

The controversy began in the early 1980s with laboratory experiments conducted by solid state physicist Panayiotis Varotsos of the University of Athens and his colleagues. They found that rock squeezed in the lab produced a transient electrical current just before fracturing. Might it also give off electrical signals before fracturing under stress in Earth's crust, they wondered—that is, during an earthquake?

To find out, they set up what amount to giant voltmeters around Greece: up to several kilometers of wire connected to two electrodes stuck in the ground. Their equipment turned up signals aplenty, including extraneous currents such as radio broadcasts and industrial noise. But once Varotsos and his colleagues thought they could recognize and weed out noise, they identified “seismic electric signals,” or SESs, that seemed to precede quakes of all sizes in Greece.

Some seismologists were intrigued, but many objected that any apparent VAN successes were just dumb luck; by making enough predictions, the VAN group was sure to catch a few of the many quakes that strike Greece each year. Unfazed by such objections, Varotsos and his colleagues expanded their monitoring. On 1 and 2 September, a station near Lamia, about 150 kilometers northwest of Athens, recorded the first powerful signal in its 4 years of running, says Varotsos. He and his University of Athens colleagues, physicists Vassilios and Claire Hadjicontis, say they immediately recognized the signal as the SES of a significant forthcoming earthquake.

They concluded that the SES signaled a magnitude 5.5 quake that would strike something like 70 kilometers away sometime in the coming few weeks. The Athens quake came 5 days later, 140 kilometers away, with a magnitude of 5.9. “It is very impressive to see the signals and expect an event,” says Claire Hadjicontis. “I think it's very promising.”

The VAN group never made this prediction public, according to Varotsos, because of an understanding with the Greek government that they would only announce predictions of imminent quakes of magnitude 6.0 or larger. But Varotsos soon thought he had another prediction, which did fit the bill. The signal of 1 to 2 September, he had noticed, changed polarity before disappearing, something that had happened before when a single station had picked up merged SESs from two impending quakes. Then, on the 13th, the Lamia station picked up another SES of the same polarity as the end of the earlier signal—seemingly a continuation of the first.

“This strengthened our interpretation that the last part of the signal should correspond to future activity,” Varotsos told Science on the 14th. On the 16th, VAN group member Kostas Eftaxias went public on national TV with both their “postdiction” of the 7 September quake and suggestions of another impending temblor somewhere around Lamia, this time with a magnitude of about 6.0.

Chouliaras is not impressed. “It is ridiculous to continue this debate,” he says. In recent published papers, he says, he and colleagues have shown that the SES-like signals they recorded independently at the VAN station in western Greece are radio and phone transmissions, not crustal signals. Resvanis also remains to be convinced. “If they did predict [the 7 September quake], it would be random coincidence,” he says. Adds Geller: “His ‘predictions’ are on the same level as those of the oracle at Delphi.” To be taken seriously, he says, the group needs to change its ways. “Varotsos is simply not carrying out scientific research as it is understood by scientists. … None of the necessary conditions—free availability of continuous raw data, publication of the prediction algorithm—are satisfied.”

Even those who have offered some support in the past are being cautious. Stephen Park of the University of California, Riverside, says he “would back off and take a little more conservative view than in '95,” when his analysis suggested VAN was doing better than chance at predicting quakes. With a longer VAN record to work from, Park now finds that any claims of real success “could be questioned by statisticians.”

Varotsos has answers to all these criticisms. For example, he acknowledges that he and his colleagues “record a lot of noise, but we apply certain criteria and immediately classify noises versus signal,” and he can point to a published algorithm. But he gets the feeling that his critics are actually sending a broader message: “The problem [of earthquake prediction] is very difficult, and therefore no one should try.” Varotsos insists he must, although he now faces both the mysteries of earthquakes and the deep skepticism of his colleagues.

9. NASA

Space Science Feels Budget Ax in Senate

1. Andrew Lawler*
1. With reporting by Jeffrey Mervis.

Sighs of relief resounded everywhere at NASA last week, with Hurricane Floyd blowing past the Kennedy Space Center without damaging the shuttered shuttles, and a Senate panel granting the agency its full $13.6 billion request for 2000. Everywhere except Ed Weiler's office, that is. “This is bizarro-land,” the space science chief complained after hearing that his division was the only one at NASA to get clobbered. “What have we done to deserve this?” What distressed Weiler was the Senate Appropriations Committee's bottom line for NASA space science:$2.08 billion in 2000, $43 million less than this year's budget and a hefty$120 million shy of his request. The cut was especially painful because the committee granted NASA and the National Science Foundation (NSF) the overall amounts the White House asked for—thanks to a critical decision by Senate Republican leaders to break strict budget caps. NSF scored a 7.9% boost for research, putting it just over the $3 billion level, an outcome a relieved NSF chief Rita Colwell calls “wonderful.” The proposed budgets are in stark contrast to the House plan to stick with the spending caps and chop funding for NASA and NSF (Science, 17 September, p. 1827). The full Senate is expected to vote this week, and the two chambers will meet in coming weeks to hammer out a final plan that will go to President Bill Clinton for approval. A White House official told Science that the Administration will fight to restore space science funding. Although the Senate panel bit half as deeply into the space science budget as did the House, which had slashed$240 million from NASA's request, Weiler warns that even the more modest cut could cripple programs ranging from the Hubble Space Telescope to comet and planetary missions. “The irony is that this is the heyday of space science,” he says. “We had eight successes out of 10 launches this year,” including the Chandra x-ray telescope, which began sending back images this month.

But Senate members take a dimmer view of NASA's record. Appropriations Committee documents refer to “mixed successes and some outright failures,” including the loss in space of the Wide Field Infrared Explorer mission in March and the unexpected need for an expensive mission to repair the Hubble's gyros. The panel also notes that NASA may be shortchanging data analysis—research and analysis account for one in four space science dollars—and asks the White House and NASA to consider developing a data warehouse. Weiler does not quibble with the need for better dissemination and analysis of the flood of data streaming back from probes circling Mars and Jupiter and from observatories like Hubble and Chandra, although other NASA officials note that much of the data is available on the Internet. But he warns that the proposed Senate cuts would mean less money for everything, including analysis.

The Senate panel finds more to praise in projects that will benefit particular states. For example, Senate Majority Leader Trent Lott's (R-MS) desire for more spending on space transportation—specifically, engine testing at Stennis Space Center in his home state—took precedence over space science, according to NASA officials. The Senate plan includes $100 million above the$1.1 billion requested for aerospace technologies. The bill also includes a host of pork projects having nothing to do with space, ranging from $1 million for a museum on “the underground adventure” of soil ecosystems to$14 million for a life sciences upgrade at the University of Missouri, Columbia. That's the home state of Senator Kit Bond (R-MO), who chairs the panel that appropriates NASA funding. The pork projects—some of which would have to be paid for out of Weiler's budget—put even more pressure on space science funding.

With space science likely destined for a big cut, finger-pointing has begun in earnest. Weiler worries that most scientists don't understand the extent of the threat and adds that congressional staffers have told him that only a handful of researchers have complained about the proposed cuts. But Kevin Marvel, who heads public policy for the American Astronomical Society, says the problem is the larger politics of budget caps and the jockeying for funds inside NASA. “Blaming the community for a battle being lost internally is the wrong road,” he says.

10. BIOMEDICAL RESEARCH

AIDS Researchers Blast NIH Peer Review Plan

1. Eliot Marshall

A scheme to overhaul peer review at the National Institutes of Health (NIH) is drawing intense fire from the AIDS community. Complaints from patient activists and scientists have been piling up for the past 2 weeks at NIH's Center for Scientific Review (CSR), which is considering recommendations from a panel headed by Bruce Alberts, president of the National Academy of Sciences, to reshuffle the groups that rank grant applications (Science, 30 July, p. 666).

The Alberts committee suggested grouping peer review panels under broad areas of science rather than specific disease categories or research methods, as many are grouped now. For example, the panel proposed doing away with the category “AIDS and AIDS related research” and moving the seven study sections grouped under this heading into new, more general science categories (such as immunology). The scheme allows reviewers to be moved readily from one panel to another within each grouping. But one critic, Mario Stevenson, a virologist at the University of Massachusetts, Worcester, says: “The logic [of the new proposal] isn't apparent to me. … I think reviews in the AIDS area are working very well.”

Stevenson is part of a group of scientists who endorsed a protest letter circulated by AIDS researcher Ron Desrosiers of Harvard University's New England Primate Research Center in Southborough, Massachusetts. They argue that eliminating the AIDS-specific category would dilute expertise and lower the quality of peer review. In addition, Charles Carpenter of Brown University, chair of the council that advises the NIH Office of AIDS Research, has sent CSR a letter on behalf of council members warning that the proposed reform could “cause irreparable harm” by exposing grant proposals “to review by investigators lacking the appropriate knowledge of AIDS research.” Neal Nathanson, director of NIH's Office of AIDS Research, has also expressed his concerns about the plan in an informal e-mail to Elvera Ehrenfeld, director of CSR. Nathanson was unavailable for comment.

Ehrenfeld says she was surprised by the angry response from the AIDS community. She thinks AIDS researchers may be confused by “an unfortunate misunderstanding” that existing study sections would disappear. The AIDS panels would simply be placed in new groupings, she says. For example, the panel reviewing AIDS-related behavior research might be grouped with behavioral research, and other AIDS study sections might be grouped with virology or immunology. Furthermore, none of this is set in stone. Some of the criticism “may be valid,” Ehrenfeld says, “and that's why we asked for comments.” Alberts also wants to dispel concern: “Clearly we need to explore with the AIDS researchers exactly what it is that bothers them and why. After this detailed discussion, the committee will decide, based on science, how to modify our report.”

Comments are due by 15 October. In early November, a CSR advisory council will discuss the next step.

11. UNIVERSITY OF ILLINOIS

Chancellor Quits After Research Shutdown

1. Constance Holden

Already reeling from a federal suspension of its clinical research, the University of Illinois, Chicago (UIC), got another jolt this month: the sudden resignation of its chancellor, David Broski, on 9 September. Broski appears to be the third—and highest ranking—school official to fall in the course of a simmering 2-year conflict involving the university's Institutional Review Boards (IRBs), which review research proposals that involve human subjects.

On 27 August, the National Institutes of Health's Office for Protection from Research Risks (OPRR), a watchdog agency that monitors compliance with federal rules on human subjects research, suspended about 1000 NIH-funded projects at UIC. This is the latest in a wave of OPRR crackdowns across the country, including one last fall at Rush-Presbyterian-St. Luke's Medical Center in Chicago. The office acted after determining that some UIC research had been conducted without IRB review, and in some instances it found that investigators had failed to obtain informed consent. Senior officials, OPRR said, “knew, or should have known, about these deficiencies.”

According to Stanley Schade, professor of hematology and oncology and former chair of the biomedical IRB, the problems began about 2 years ago when a woman complained that her confidentiality had been violated after she was given brain scans while going through an induced episode of a dissociative disorder. When it was discovered that this study had never been submitted to the IRB, the university set up a task force to determine how to tighten up procedures. Despite these efforts, sources say, there was continuing friction between the IRB and its support staff, and university officials anxious to keep the stream of grants coming in. For example, Lynda Brodsky, the former staff chief, says one department would get preliminary NIH approval for a study and then “pressure the IRB to rubber-stamp it.” Brodsky says she was removed from her job in July 1998. Then last January, 10 of the 12 members of the biomedical IRB resigned to protest, among other things, Brodsky's departure and staff shortages.

In March, after receiving a whistleblower's complaint, OPRR started investigating the university's human subjects research procedures. Three months later the vice chancellor for research, Mi Ja Kim, resigned. But if the move was meant to appease OPRR, it was either too little or too late: In addition to suspending UIC's human subjects research, the office said that staffing and technical support for the university's three IRBs was “markedly insufficient” to the point that it “undermined the mission of the IRB.” On 30 August, immediately after the suspension, Broski told an open meeting at the university that “the buck stops at my desk, and I take responsibility for the findings.” University president James Stukel would say only that Broski left for “personal reasons.”

The university has since issued a mea culpa. In a statement, Eric Gislason, interim vice chancellor for research, admitted that “our office did not keep up with” the rapid growth in UIC's research program over the past 4 years, a period in which total funding from the Department of Health and Human Services doubled to its current level of \$80 million. UIC has followed OPRR's orders, says UIC spokesperson Bill Burton. It has put relevant people through educational programs on research ethics and is revising its ethics procedures, which it plans to submit to OPRR before the end of the month. On 1 October, IRBs will start what is likely to be a yearlong process of re-reviewing all active projects. (See www.uic.edu/depts/ovcr/oprr/index.html for documentation of university actions.)

12. STRUCTURAL BIOLOGY

The Race to the Ribosome Structure

1. Elizabeth Pennisi

Large and complex, the ribosome has resisted efforts to decipher its structure for 4 decades, but now four groups have it in their sights, to the excitement of everyone in the field

Every so often, an actress who has struggled through years of bit parts and waitressing gigs suddenly finds that she's an overnight success. The change comes when she lands that key role that finally shows off her talents. In the molecular world, a similar happy fate has now befallen the ribosome, the tiny particle in the cell that translates the genome's messages into all the proteins needed for life. After 4 decades in which researchers' best efforts to introduce the ribosome to a wider audience by laying bare its complete molecular structure have been frustrated, that structure is now having its debut.

The ribosome itself is to blame for its long period of obscurity. It's a tangle of 54 proteins and three RNA strands, which complicated every step needed to work out its structure by x-ray crystallography. But last month, in the 26 August issue of Nature, Venki Ramakrishnan's team at the University of Utah, Salt Lake City, published the structure of the smaller of the two subparticles that together make up a complete ribosome, while the team of Peter Moore and Thomas Steitz at Yale University described the structure of the larger subunit (Science, 27 August, p. 1343). Ada Yonath and her colleagues at the Weizmann Institute of Science in Rehovot, Israel, and the Max Planck Research Unit in Hamburg, Germany, say they, too, have solved the structure of the smaller subunit, although their results are not yet published. And on page 2095 of this issue, a fourth team, led by Harry Noller of the University of California, Santa Cruz (UCSC), reports the structure of the complete ribosome.

The structure's debut took only months, but behind the scenes the story unfolded over several decades, during which time Yonath doggedly worked to make progress. Only recently have others joined the fray, sparking what will be an intense sprint to the first truly high-resolution image of the ribosome and its subunits. The images are not quite there yet.

The resolution of the individual subunit structures, at between 4.5 and 5.5 angstroms, reveals the overall arrangement of the ribosomal proteins and RNAs, but does not show individual atoms. Noller's structure of the complete ribosome has a lower resolution, 7.8 angstroms, but does give the details of how the two subunits interact. What's more, the UCSC team also has three-dimensional images of the ribosome bound to messenger RNA (mRNA), which carries the genetic information needed for protein synthesis, and bound to transfer RNA (tRNA), the kind of molecule that supplies the ribosome with amino acids, the building blocks of proteins.

“There's no question that it's more interesting to see the two subunits together,” says Moore. “If you want to see how protein synthesis occurs, the low-resolution map [of the whole ribosome] will be very significant.” As Anders Liljas, a crystallographer at the University of Lund in Sweden, points out in his commentary on page 2077, ribosome researchers can now see the functional centers of the ribosome clearly. Thus, they can better check how their ideas on the mechanism of protein synthesis stack up against the real thing.

Breaking ground

Although three groups have gotten into print with their ribosome structures before Yonath, they and others credit her with paving the way. “She proved that it could be done,” says structural biologist Wayne Hendrickson of Columbia University in New York City—something that many crystallographers doubted. When Yonath began her studies in the late 1970s, ribosomes seemed too big and too variable to crystallize and study by x-ray diffraction.

It took decades of persistence for her to overcome that belief. Yonath started by determining the structure of an “initiation factor,” a protein that helps jump-start protein production by binding temporarily to the ribosome. But even that single protein proved difficult to purify in quantities sufficient for structural studies. And in 1978, Yonath suffered a further setback.

By then she had arranged to work at the Max Planck Institute for Molecular Genetics in Berlin, Germany, which had undertaken a large research effort on ribosomes and which had the protein-purification equipment she lacked at her lab in Israel. But a bicycle accident sidelined her for several months. Once she arrived in Berlin as a visiting professor, however, her focus—and her luck—changed. There she discovered a crystallographer's gold mine in the lab's refrigerators: lots of highly purified ribosomes left over from previous experiments. Yonath decided to ask the institute's director, Heinz Günter Wittmann, if she could try to crystallize them. “He said this was the dream of his life and gave me everything I needed,” Yonath recalls.

Still, with the techniques then available, it took Yonath months of trying different solutions and crystallization procedures to get tiny crystals of the larger, or 50S, subunit of the ribosome from a Bacillus bacterium, and more than a year to get the first very fuzzy x-ray crystallographic images. But when she showed colleagues her results at an August 1980 meeting, “everyone laughed at me,” Yonath recalls. A few key people kept the faith, however.

One was Wittmann; another was Sir John Kendrew, a Nobel Prize-winning x-ray crystallographer who then was director of the European Molecular Biology Laboratory in Heidelberg. He helped ensure that Yonath was able to continue to get beamtime for her x-ray diffraction studies, despite the high risk of failure. Then in 1981, Yonath produced crystals of the Bacillus 50S subunit perfect enough that they yielded diffraction patterns in which she could distinguish atoms down to 3 angstroms apart from one another—an accomplishment that Hendrickson describes as “pivotal,” because that resolution was about what would be needed to discern all the atoms in the structure.

Even so, getting to that resolution still proved difficult, because she needed stabler crystals that would last long enough in the x-rays for her to collect enough data to solve the ribosome's structure. Over the next several years she made a number of improvements. For one, ribosomes from the Dead Sea microbe Haloarcula marismortui proved stabler and better suited to x-ray studies than those from Bacillus. Together with Hakon Hope, a crystallographer at the University of California, Davis, Yonath made another key advance by adapting Hope's supercooling technique to the ribosome crystals so they would last longer in the beam of x-rays. “She was one of the first people to recognize that you had to cryocool [freeze] the crystals to get data,” says Utah's Ramakrishnan. “She should get a lot of credit,” now that the technique is universal in macromolecular crystallography.

With stabler crystals in hand, the next step was to figure out how to create specific landmarks for phasing in the diffraction patterns, a first step toward making sense of those patterns. Crystallographers typically do this by doping their crystals with heavy atoms, which contain so many electrons that they stand out like beacons on the electron-density maps. But because the ribosome is so big and electron-filled, Yonath needed larger concentrations of electrons than single atoms could provide and so decided to use clusters of heavy atoms instead. Even with these advances, by 1995 when the international ribosome community met in Victoria, British Columbia, her report on the 50S subunit was disappointing. And other ribosome researchers were beginning to champ at the bit.

Among crystallography's purists, the tradition has been that once a researcher crystallizes a challenging molecule, he or she is given the latitude to see that molecule through to its atomic resolution. After all those years, Yonath “had convinced everyone that you could do structural analysis [on the ribosome] by x-ray crystallography,” recalls Roger Garrett, a molecular biologist at Copenhagen University in Denmark. But she hadn't really figured out how to solve the structure, and he adds, “People started getting impatient.”

Parallel pursuits

One rival effort had begun behind the Iron Curtain during the 1980s. Meeting as graduate students in Alexander Spirin's lab at the Protein Research Institute in Pushchino, Russia, Gulnara Zh. Yusupova and Marat Yusupov soon married and later embarked on their own pursuit of ribosome crystals at the institute in 1983. By 1987, they had produced crystals of the ribosome and of the smaller 30S subunit from the bacterium Thermus thermophilus. But Yusupov recalls, “It was impossible to solve the [structure] in Russia,” because biologists there lacked access to the x-ray beamlines.

After the dissolution of the Soviet Union in 1989, however, the Yusupovs began collaborating with Dino Moras at the University of Strasbourg in France. There they produced a new crystal that diffracted better than the original one, but the work bogged down in the early 1990s because of shortages of funding. So in 1996, the Yusupovs packed up their crystals and moved to California to work with Noller at UCSC. “It was the right decision,” Yusupov says.

Noller already had a 20-year interest in the ribosome, one that began when he sequenced and characterized its various RNA components. Because directly determining the entire structure seemed impossible at the time, he and his team developed biochemical approaches to figuring out where the RNAs link with the individual ribosomal proteins. But by the mid-1990s, Noller was ready to try the impossible. He recruited crystallographer Jamie Cate, who had just finished solving a much simpler RNA structure. “In terms of structural biology, this was the biggest challenge I could pick,” Cate recalls.

At Santa Cruz, the Yusupovs improved their purification procedures, and Yusupova began working out ways to get crystals of ribosomes attached to pieces of either tRNA or mRNA. Gradually the crystals got better, and complexes of the ribosome with the other RNAs often proved stabler than the ribosome alone. Even as the crystals improved, the group was stumped about the best way to study them. At first it seemed “we just had to roll up our sleeves and fight with them,” says Cate. Daily he and Yusupov would sit down for coffee to hash out ideas. Cate found Yusupov's methodical approach a good counter to his seat-of-the-pants, “let's do an experiment and work out the details later” impulsiveness, and the two were able to make progress on all fronts.

Noller's group also took an incremental approach to making sense of the diffraction patterns obtained from the team's crystals. As a first step, they used lower resolution structures reconstructed from images obtained by cryo-electron microscopy, in which specimens are frozen in vitreous ice before they are imaged (see sidebar). His team then refined those maps by diffracting x-rays of different wavelengths through ribosome crystals containing clusters of heavy atoms, a technique called multiwavelength anomalous dispersion. These clusters provided improved phasing information, which the researchers used to locate single heavy atoms inserted into another set of crystals.

This approach, aided by improvements in both the crystals and the synchrotron facility that provided the x-rays, helped the researchers surpass their expectations. “My feeling was that if [the crystals] only went to 12 to 15 angstroms, it would still be important,” says Thomas Earnest, Noller's collaborator at Lawrence Berkeley National Laboratory in California. But so far the team has resolved the structure to 7.8 angstroms.

And they're off

Like Noller, Yale's Moore had been in the ribosome field for many years. With Yale colleague Donald Engelman, he had spent a decade working out the relative positions of all the proteins in the 30S subunit by scattering neutrons through crystals of the subunit, a structural technique that gives less precise information than x-ray crystallography. Still, says Noller, it was “a monumental piece of work.” But as it began winding down in the early 1990s, Moore began to think seriously about the next step. “I got interested in higher resolution information,” he recalls. For years, he and Yale structural biologist Steitz, who had distinguished himself by solving the structure of several proteins, had talked about tackling the ribosome.

In 1994, Steitz brought in Nenad Ban, who had worked on crystallizing viruses. Although Yonath's and Yusupov's work had shown that ribosomes can produce high-quality crystals, “we had to stretch every available method to its limits,” Ban says. The team decided to follow Yonath's lead, producing 50S crystals by using the methods that she reported as giving the best crystals. But like Yonath, they stumbled. “It was no wonder that she had trouble,” says Moore. “These [crystals] are really hard.”

The Yale team had particular trouble with “twinning,” in which two mirror-image crystal structures form within what seems to be one crystal, resulting in confusing diffraction patterns. It took 2 years for the researchers to work out the problems, but once they were solved, “we were able to proceed very fast,” Moore says. In the 26 June 1998 issue of Cell, they described a structure with 9-angstrom resolution, demonstrating that heavy atoms could be used to solve the ribosome structure. They then continued to improve on their data and analytical techniques, reaching 5-angstrom resolution by the spring of this year.

Meanwhile, Ramakrishnan, another longtime ribosome researcher, was edging toward the starting line as well. He had worked with Moore on the neutron scattering experiments from 1978 to 1982 and, while working first at Brookhaven National Laboratory on Long Island and then at the University of Utah, Salt Lake City, had helped solve the structures of about a half-dozen individual ribosomal proteins. He had become ever more eager to solve the structure of the 30S subunit, which plays the crucial role of accepting or rejecting the tRNAs that carry successive amino acids to be added to the protein.

He had also deduced from Yonath's presentation in Victoria that no one had really pursued getting good crystals of the 30S subunit, and he decided to try. His team managed to make steady progress, even though he moved across the Atlantic from Salt Lake City to the Medical Research Council Laboratory of Molecular Biology in Cambridge, U.K., earlier this year. During that time, though, Ramakrishnan never really let on that he was making this attempt. “He was the dark horse,” says Joachim Frank, a physicist at the New York State Department of Health Wadsworth Center in Albany.

But when Ramakrishnan, Moore and Steitz, and Yonath gave consecutive talks about the structures they were working on at the opening session of an international ribosome meeting in Helsingør, Denmark, this June, Frank was most impressed by how far this dark horse had come. At the meeting, Ramakrishnan explained that turning his postdoctoral fellow Brian Wimberly loose on solving the structure “was like handing the keys of a Ferrari to a teenager.” Wimberly not only sped picking out key landmarks in the resulting electron-density map, but also was able to trace parts of the RNA's path through the subunit.

Yonath also described her 30S research at that meeting, but her lower resolution results paled in comparison with those of Ramakrishnan and the others. Some of her colleagues, particularly her rivals, couldn't help but wonder if she had been left behind in the field she started. “She has invested most of her career in this problem,” says Moore. “If she had been the first person to get the structure, there's no question that she would be given the Nobel Prize. Now, the final answer is unlikely to come from her.”

While many in the audience were pleased to see new talent attacking the problem, a few crystallographers think that Moore and Steitz and Ramakrishnan should have held off. “She's done most of the backbreaking work,” says Kenneth Holmes, a structural biologist at the Max Planck Institute for Medical Research in Heidelberg. “The others have jumped the gun. They should have left Ada in peace.” But Yonath herself says she doesn't mind. “After being alone for so long, this gave the field a big push and gave me much satisfaction,” she says. And colleagues who have seen her latest work, which she did not present at the June meeting, say she should not be counted out.

Indeed, everyone agrees that the race to the ribosome structure is not over. At the current level of resolution, says Steitz, “the [structure's] impact on biochemistry is small compared to what we want it to be.” The ultimate goal is atomic resolution that reveals the exact interactions between the ribosome's components and thereby provides an understanding of how the structure makes protein synthesis possible. But like Yonath, he expects rapid progress. “We've defined for ourselves what's necessary to solve the structure,” he says, adding that he expects the solution in a matter of months.

Yonath, for example, has already taken a step in that direction. In August at the XVIII Congress of the International Union of Crystallographers in Glasgow, U.K., she presented data on a 4.5-angstrom structure of the 30S subunit with, she says, “markers that mimic the functional sites.” Liljas describes the structure as “very exciting,” because the markers reveal where the mRNA hooks up with the subunit, presenting an even clearer picture of the ribosome than those now published.

But the crystallographers point out that even the expected atomic resolution structure won't reveal everything about the workings of the ribosome during protein synthesis. The images from structural biologists “are snapshots,” says Noller. The ribosome is really a machine that moves along the messenger RNA, all the while transferring amino acids from incoming tRNA and forging peptide bonds in the growing protein. To visualize all of this, says Noller, “what we really want is the movie.”

That means capturing multiple images of the ribosome, recording each step. Making crystals suitable for x-ray diffraction studies of all those ribosome forms would be extremely difficult at best. It may be possible, however, with cryo-electron microscopy, which requires only that the various forms be properly frozen, although it's not clear that such studies will provide atomic resolution in three dimensions the way x-ray crystallography could.

But even if the “movie” turns out to be far in the future, researchers still expect benefits from the atomic resolution structure, and even from the less detailed structures like those that have now appeared. “There's a huge amount of biochemical evidence that can be put into perspective,” says Ban. “It's going to be fantastic for everyone who is working with the ribosome.”

13. STRUCTURAL BIOLOGY

Challenge From Electron Microscopy

1. Elizabeth Pennisi

Physicist Joachim Frank may well feel a twinge of envy when he turns to pages 2095 and 2133 and sees the x-ray crystallographic images of the complete ribosome made by Harry Noller's team at the University of California, Santa Cruz. For 2 decades, both Frank's team and a rival group in Europe led by Marin van Heel have been pushing the limits of another imaging technology, called cryo-electron microscopy (cyro-EM), to achieve the same goal: seeing the ribosome—the cell's protein factory—in all its molecular detail. And 4 years ago, when both teams published structures resolved to about 25 angstroms, it seemed they might succeed ahead of the crystallographers. Now, crystallographers have taken the lead, but for understanding the dynamic nature of this molecular complex, cryo-EM is still very much in the running.

Before tackling the ribosome, Frank had spent years working out a computer program that could build two- and three-dimensional models of molecules by averaging and aligning electron micrographs of those molecules taken from different angles. He used the technique in 1981 to create a 2D reconstruction of the ribosome, but rapid progress didn't become possible until the late 1980s. By then, Jacques Dubochet of the European Molecular Biology Laboratory in Heidelberg, Germany, had developed a way of rapidly freezing molecules in water so that they end up encased in glassy vitreous ice, which lacks the crystals that would otherwise distort the structure.

Prepared this way, the ribosome could be seen much more clearly, and both Frank and van Heel, who was then at the Fritz Haber Institute of the Max Planck Society in Berlin, quickly applied cryo-EM to this structure. In 1991, Frank published the first 3D reconstruction of ribosomes based on cryo-EM. The resolution was only about 45 angstroms. But, Frank says, “we saw the two [ribosome] subunits and the triangular shape of the intrasubunit space.”

Meanwhile, van Heel, who is now at Imperial College in London, was working out his own method for making 3D reconstructions. In 1995, at a Gordon Conference in New Hampshire, his group and Frank's “had two posters right next to each other,” with models of the ribosome resolved to about 25 angstroms, van Heel recalls. For the first time, the molecule's surface topography, with all of its cavities and bulges, came into view. “All of a sudden, the ribosome had a spatial reality,” Frank says. The researchers could see, for example, how the messenger RNA, which specifies a protein's composition, might thread itself through the ribosome. The images also helped crystallographers interpret their x-ray diffraction data. “Crystallographers wouldn't have gotten anywhere without the cryo-electron microscopy maps,” says Roger Garrett, a molecular biologist at Copenhagen University in Denmark.

Since then, both cryo-EM teams have been getting structures with progressively higher resolutions. Frank now has some 70,000 projections of the entire ribosome with an 11.5-angstrom resolution, and until recently, he says, “we thought we could win the race against Noller.”

Van Heel still thinks he can win the next phase of the race: the quest for images with 4-angstrom resolution, almost good enough to reveal the positions of individual atoms. At the international ribosome meeting in Finland in June, he showed data on ribosome reconstructions with a resolution as fine as 8 angstroms. Frank isn't convinced that the resolution is that good, as he thinks the images didn't have the kind of detail he would have expected. But van Heel is bullish. “We're expecting a new microscope, and I'm expecting it to get below four [angstroms],” he says.

Even if the crystallographers also beat Frank and van Heel to the next milestone, cryo-EM has a unique advantage over crystallography. With cryo-EM, freezing a complex of molecules takes less than a second, instead of the weeks it takes to grow a crystal. So Frank and van Heel can easily catch the ribosome at various stages of operation, revealing how the molecules needed for protein synthesis move during the process. And that, notes Albert Dahlberg, a molecular biologist at Brown University in Providence, Rhode Island, “is going to be the next step, exploring the dynamic nature” of this protein factory.

14. PALEONTOLOGY

T. rex Was Fierce, Yes, But Feathered, Too

1. Tim Appenzeller

A new “feathered” fossil, this one a close relative of Velociraptor, adds to evidence that many predatory dinosaurs had some kind of plumage

Dinosaurs live on as sedate green lizards on the Sinclair gas station signs, but for the real beasts, that look was passé long ago. As anyone familiar with Jurassic Park can tell you, dinosaurs were brightly colored and awesomely athletic. Now, paleontologists say it's time for yet another makeover, at least for some of Jurassic Park's most menacing characters.

Picture this: Adult Velociraptors, savage man-sized hunters with slashing claws, may have been covered in downy feathers, like newly hatched chicks. The same goes for the young of Tyrannosaurus rex, and even the full-grown monster may have had a tuft or two.

That's the implication of the stunningly well-preserved fossils unearthed over the last several years from spectacular fossil beds in China: a total of five small dinosaurs apparently clad in some kind of feathers, ranging from fibrous down to true feathers as unmistakable as a pigeon's. The most recent find, which a Chinese team reported last week in Nature, is a small, birdlike animal called a dromaeosaur. The 125-million-year-old fossil is an undersized ancestor of Velociraptor, which makes it almost certain that the later, larger creatures had similar plumage, says dinosaur paleontologist Mark Norell of the American Museum of Natural History (AMNH) in New York City. “We have as much evidence that Velociraptors had feathers as we do that Neandertals had hair,” he says.

From the widely scattered perches of this and the other feathered fossils on the dinosaur family tree, paleontologists are realizing that many other dinosaurs on intermediate branches probably also had some kind of feathers, which may have served as insulation. “Feathers are marching their way down” the tree, says Paul Sereno of the University of Chicago—and settling on such unlikely creatures as T. rex and its relatives. It's a proliferation that, to some, dramatically underscores the proposed evolutionary link between dinosaurs and birds. But even for ardent proponents of that link, it takes some strenuous mental adjustment to picture such a wide range of dinosaurs in their new clothes.

And a small, energetic band of dissenters is not at all ready to make this kind of adjustment. The halos of fibers found around several of the crucial fossils, including the latest one, are far more likely to be some kind of internal connective-tissue fibers left behind when the flesh decayed than anything related to feathers, according to John Ruben of Oregon State University in Corvallis. “We think these are just collagen,” he says.

This controversy first erupted when a “feathered dinosaur,” Sinosauropteryx, emerged 3 years ago from the fossil deposits in China's Liaoning province, where volcanic eruptions some 125 million years ago in the early Cretaceous period entombed a menagerie of ancient animals (Science, 1 November 1996, p. 720, and 13 March 1998, p. 1626). The fine-grained sediments had preserved a fibrous mane along the creature's neck and back—“protofeathers” to many paleontologists, internal fibers to some others.

No one could deny the reality of the feathers on two fossils found later in the same deposit, creatures called Caudipteryx and Protoarchaeopteryx (Nature, 25 June 1998, p. 753, and Science, 26 June 1998, p. 2051). Caudipteryx, the better preserved of the two, had unmistakable feathers on its forelimbs and tail, although it was evidently flightless. But that did not win over the doubters, who think birds originated from reptiles that lived well before the dinosaurs. They agreed that the feathers were real enough but said that based on other features, both creatures were not dinosaurs at all but unusual flightless birds. Caudipteryx, says Ruben, is “a Cretaceous turkey, so to speak.”

Many paleontologists had a very different view, considering Caudipteryx and Protoarchaeopteryx as feathered dinosaurs all right, but ones that were somewhat removed from the main line of descent to birds. A better candidate for close cousins of the first birds seemed to be the dromaeosaurs, a group of small predators with a massive claw on each hindlimb and shoulder joints that gave the forelimbs a wide range of motion, a prerequisite for the evolution of flight. But paleontologists had few good fossils of these animals, let alone a view of one with feathers.

Now, Xiao-Chun Wu and his colleagues at the Institute of Vertebrate Paleontology and Paleoanthropology in Beijing have filled that gap with yet another fossil from the Liaoning deposits. The fossil they describe in Nature, a 40-centimeter-tall, downy dromaeosaur called Sinornithosaurus, “has more bird features than any of the other feathered dinosaurs,” says James Clark of George Washington University.

“Everything about the fossil is skewed toward similarity to birds,” adds Wu, citing the mobility of the shoulders, the shape of the breastbone, and the proportions of its limbs. Unlike other meat-eating dinosaurs, it has forelimbs nearly as long as its hindlimbs, which may have originally served in “a form of predation associated with grabbing and clutching,” says Thomas Holtz of the University of Maryland, College Park, and could easily have evolved into wings. And the downy fringe that appears to cover much of the new dromaeosaur is “a nice confirmation of the existence of feathers—or whatever we want to call them—on dromaeosaurs,” notes Holtz.

What Sinornithosaurus lacks are complex feathers like those of Caudipteryx, which troubles advocates of the bird-dino connection. If Caudipteryx—which they view as less birdlike—had true feathers, so should this missing link. This conundrum leads Wu to suggest that the dromaeosaur did have true feathers that somehow weren't preserved, perhaps because the wind blew them away, leaving only the down, before the animal was buried and fossilized. To doubters such as Larry Martin of the University of Kansas, Lawrence, however, the absence of true feathers is telling. “Flight feathers are bigger and have more structure than these fibers,” so they should have been preserved along with any down, he argues.

Ruben adds that if this dromaeosaur—supposedly so close to birds—didn't have true feathers, the “down” is probably a chimera as well. He thinks the same is true for Sinosauropteryx, the first “feathered” dinosaur, as well as another downy creature, a so-called therizinosaur, reported in the 27 May issue of Nature. As Ruben puts it, “All of these things are in all likelihood something like collagen connective fibers.”

But if Ruben and his fellow skeptics are wrong about the nature of these fibers, the images of many familiar dinosaurs should be softened with a coating of down. In the family tree of small, meat-eating dinosaurs, Sinosauropteryx lies well away from the dromaeosaurs and the putative ancestors of birds. “It's an ordinary ground-running dinosaur,” in Holtz's view. So creatures that perch in between could well have been downy too. Says the AMNH's Norell, “If you're willing to consider the fluffy stuff that's covering the body of Sinosauropteryx as feathers, you've got to contemplate T. rex, ornithomimids [a group of long-legged, ostrich-sized dinosaurs], and many others as having this kind of plumage at one time in their lives.”

Don't start describing T. rex to your 4-year-old as a toothy version of Big Bird, though. “Whether an adult T. rex had full plumage—well, there's no direct evidence for it, and it might not have been great to have a lot of insulation when you weighed 5 or 6 tons and lived in an environment like Louisiana,” says Holtz. “I wouldn't be at all surprised if adult T. rex had lost its plumage, although it may have had feathers here and there.”

15. AMERICAN CHEMICAL SOCIETY

Raising a Glass to Health and Nanotubes

1. Robert F. Service

NEW ORLEANS, LOUISIANA—For the 12,000 researchers who sweated out the dog days of summer at the American Chemical Society (ACS) meeting here from 22 to 26 August, the hot papers included a novel explanation for the healthy effects of moderate alcohol consumption, and both triumph and trials on the road to electronic devices based on carbon nanotubes.

The French seem to have it all. They eat an exquisite diet full of high-fat foods such as cheese and meats washed down with fine wines, and yet they suffer from only one-third as much heart disease as do inhabitants of the United States. One explanation for what health experts have dubbed the “French paradox” is that antioxidant compounds in red wine prevent fats from being oxidized into forms that tend to build up in coronary arteries, among other places. Yet simply eating grapes that harbor the same compounds as wine doesn't seem to confer the same benefits. “So something else must be going on,” says Yousef Al-Abed, an organic chemist at the Picower Institute for Medical Research in Manhasset, New York. At the meeting, Al-Abed reported that rat studies suggest a new possibility: that an alcohol metabolite prevents the formation of harmful compounds called advanced glycation end products (AGEs), which are thought to initiate the potentially deadly plaque buildup in coronary arteries.

“I feel quite excited about it,” says Helen Vlassara, an AGE expert at the Mount Sinai School of Medicine in New York City. Although the notion that alcohol suppresses AGEs has been around for a while, “this was the first time it was solidified [experimentally],” she says. In addition to helping explain the French paradox, the new work may also lead to the development of novel drugs that combat heart disease by targeting AGEs—work that biochemist Richard Bucala, the senior scientist on the Picower team, says is already under way.

AGEs begin to take form when common proteins circulating in the blood pick up a sugar group or two in their wanderings, which bind to the proteins to create key AGE intermediates called Amadori products (APs). The sugar groups on APs can adopt either a looped structure that behaves like a molecular Dr. Jekyll or a Hyde-like linear form. The linear form is harmful because the glucose remains reactive and can act as a molecular coupler, linking proteins together to form AGEs. The AGEs in turn can aggregate to form a thick, cross-linked web of proteins, which is thought to play a role in everything from atherosclerotic plaques to the loss of tissue flexibility with age. AGEs also bind low-density lipoprotein, the so-called bad cholesterol, and slow the rate at which it is cleared from the blood, thus increasing a person's cholesterol level and overall risk for heart trouble.

Amadori products with looped sugars are less reactive, but the rings can still open up into the Hyde-like chains—except on certain APs containing a pair of rings that seem frozen in their safe, unreactive form. Researchers weren't sure what causes the freezing, but Al-Abed thought some ring-forming compounds called aldehydes could be responsible. So he decided to test one called acetaldehyde, a byproduct of alcohol's chemical breakdown, to see if it could transform the ring-shaped glucose molecules on APs into a stable double ring, thereby preventing the APs from going on to create AGEs.

In the test tube, acetaldehyde did prove capable of turning the single-ringed glucose into its double-ringed cousin. To see if the same thing happens in the body, Al-Abed and his colleagues turned to rats that had been bred to serve as animal models for diabetes. Like human diabetics, these rats tend to have high levels of glucose in their blood, which the researchers thought might enable them to see sugar-protein reactions more quickly than in animals with normal blood sugar levels.

For their study, they fed one group of diabetic rats a normal diet supplemented with a modest amount of alcohol. The other group received just water and a bit more food, so that their overall calorie intake was the same. After 4 weeks the researchers measured the AGE levels in the animals and found that rats receiving the alcohol had only about half the AGE levels of their counterparts. The researchers also isolated an AP called HbA1c from the alcohol-fed rats and showed that some of it had the same stabilized double-ringed glucose structure that forms in the test tube.

The new work “gives us a real molecular mechanism” that may help explain alcohol's protective effect, says Bucala. As an encore to their current work, the Picower researchers are currently searching for potential drugs that may stabilize the Jekyll-like glycoproteins even better than acetaldehyde does. Such a drug could benefit diabetics, who are more likely to form AGEs because of their higher levels of blood glucose, and elderly patients at risk of heart disease. For the rest of us, the new work offers yet another line of evidence that drinking alcohol in moderation may produce more than just bonhomie.

Nanotubes Show the Way

Carbon nanotubes are the materials science counterpart of the high school student judged “most likely to succeed.” Made of a hexagonal lattice of carbon atoms, nanotubes are strong, light, flexible structures that conduct heat well, and depending on their precise arrangement of carbon atoms, can conduct electricity freely like a metal or reluctantly like a semiconductor. Last year, researchers built the first nanotube-based transistor, which sparked hopes for creating computers based on these molecular circuits. At the ACS meeting, Alex Zettl, a physicist with the University of California, Berkeley, and the Lawrence Berkeley National Laboratory, offered more evidence of nanotubes' promise—and also a major caution.

Zettl initially fanned hopes by reporting that his group had built a nanotube version of a diode, a standard electronic device that routes an electric current in a preferred direction. But then just minutes later he described other experiments showing that nanotube electronics suffer greatly from electronic “noise,” a background hiss that interferes with a device's ability to send and receive signals.

Zettl's new work was “one of the highlights” of a nanotube symposium that drew many of the top names in the field, says Robert Haddon, a physicist and nanotube expert at the University of Kentucky, Lexington. Haddon conceded that the tubes' noisiness—likely due at least in part to atomic contaminants on their outer surface—could spell trouble for their use in electronics, although he thinks it could be overcome by controlling contamination, as conventional chipmakers do. The new work, he says, “is a heads-up that eventually we'll have to work on the same problem.”

For now, wiring nanotubes into electronic devices is challenge enough. The standard printing techniques that pattern conventional silicon circuitry can't come close to creating anything as small as nanotubes. So, Zettl, Berkeley physicist Paul McEuen, postdoc Michael Fuhrer, and their colleagues had to take advantage of a little randomness to make their devices. First, they dusted a silicon-dioxide surface with a small collection of nanotubes. They then scanned the surface with the fine probe of an atomic force microscope to find a pair of tubes that formed a cross and didn't touch any neighbors. Once they had tracked one down, they used a conventional technique involving an electron beam microscope to position tiny electrical contacts made of gold at each of the four ends of the cross.

After hooking these contacts to a power supply and current meter, the researchers turned on the current through one tube and watched as it jumped to its neighbor. A conventional diode routes current in a preferred direction by juxtaposing a metal and a semiconductor, as electrons can flow easily from the metal to the semiconductor but hit an electronic wall at the junction when they try to go the other way. Zettl and his colleagues reasoned that any crosses that happened to combine a semiconducting and a metallic nanotube would work the same way. Current measurements showed that the new nanotube-based devices “look as good as any diode you can buy,” says Zettl. “One can envision using these devices in certain kinds of [electronic] architectures.”

All of which caused Zettl to wonder just how good wires made of nanotubes would be. So he and another associate, Philip Collins, decided to see if nanotubes suffer much from noise—fluctuations in the voltage measured when current is passed through a conductor. They wired up gold contacts to opposite ends of individual tubes, tube bundles, and even thick mats. After running current through the tube or tubes, they simply measured the noise with a standard device known as a spectrum analyzer. To their surprise they found that it didn't matter if they looked at individual tubes or bundles. All were noisy, as much as 10 orders of magnitude noisier than conventional metal conductors.

“These nanotubes are some of the noisiest conductors in electronics,” says Zettl. Part of the problem could be that they are long and thin, says Haddon. Whereas the noise in bulk metallic conductors increases linearly as they get longer, in molecule-based systems such as nanotubes, it increases exponentially.

Equally important, Zettl suggests, may well be oxygen and other electron-hungry contaminants that scurry up and down the nanotubes, attracted by whirling electrons that dance between the tubes' carbon atoms. Those contaminants could be deflecting some of the conducting electrons. “The message is that chemistry obviously influences the electronics,” says Zettl. But that is not necessarily a bad thing, he adds, suggesting that the effects could be tapped to produce an extremely sensitive oxygen sensor—yet another potential talent of these well-rounded materials.