# News this Week

Science  15 Sep 2000:
Vol. 289, Issue 5486, pp. 1850
1. FORENSIC EPIDEMIOLOGY

# Vaccine Theory of AIDS Origins Disputed at Royal Society

1. Jon Cohen

LONDON, ENGLAND—For 2 days this week, the staid Royal Society hosted a spirited, sometimes raucous, meeting on the origin of the AIDS epidemic, the first such gathering ever held. At center stage was a controversial theory that a contaminated polio vaccine tested in Africa more than 40 years ago sparked the epidemic. The theory took a hit when researchers revealed that tests of old samples of the vaccine provided no supporting evidence, and the main proponent of the theory, British writer Edward Hooper, endured a verbal battering himself from several prominent scientists. But Hooper, unbowed, got in plenty of jabs of his own.

Most AIDS researchers believe that HIV infected humans through the hunting and handling of chimpanzees, some of which harbor a closely related virus called SIVcpz. This “natural transfer” theory holds that a “cut hunter” was infected, and then urbanization, the use of dirty needles in medical campaigns, increased geographic mobility, and other effects of modernization in Africa caused the epidemic to explode.

Expanding on a theory that first received widespread attention in a 1992 article in Rolling Stone magazine, Hooper last year laid out a challenge to this conventional wisdom in a thick tome, The River, that pins the genesis of the AIDS epidemic on a long-forgotten oral polio vaccine (OPV) made by Hilary Koprowski and colleagues. Between 1957 and 1960, Koprowski, former head of the Wistar Institute in Philadelphia, Pennsylvania, tested his vaccine on a million people in what is now the Democratic Republic of Congo, Burundi, and Rwanda. Hooper posits that the vaccine became contaminated with SIVcpz because researchers used kidney cells from infected chimpanzees during the manufacturing process. As circumstantial evidence, Hooper contends that the earliest cases of AIDS closely match the sites of these vaccinations. And he argues in The River that the Wistar Institute could help settle the case by testing the few remaining samples of the vaccine—as recommended 8 years ago by an outside committee of experts that looked at the issue in the wake of the Rolling Stone article.

Now, three independent labs have done just that. At the Royal Society meeting, Claudio Basilico, the head of the expert committee and chair of microbiology at New York University Medical Center, described the results of tests on seven old vaccine samples found in Wistar's freezers. Koprowski has insisted that he used kidneys from monkeys, not chimpanzees, to make the vaccine, so one lab analyzed primate mitochondrial DNA in the samples. Another looked for SIV or HIV genetic material. And the third lab, headed by Simon Wain-Hobson of the Pasteur Institute—who has been sympathetic to Hooper's point of view and even helped do some research for him—ran tests for both virus and mitochondrial DNA. All the samples tested negative for simian and human viruses, and the mitochondrial DNA clearly came from monkey, not chimpanzee, cells. “The experiments were well done and the data were solid,” said Northwestern University's Steven Wolinksy, who conducts similar tests with HIV in his studies of viral evolution.

Hooper did not challenge the results; he simply dismissed them. “This means nothing at all for the polio vaccine theory,” said Hooper at a jam-packed press conference held a few minutes later. He noted that the samples didn't come from the exact lots of the polio vaccines tested in Africa. Indeed, Koprowski himself has acknowledged that no such samples still exist. Retrovirologist Robin Weiss of Chester Beatty Laboratories in London, who co-organized the meeting with Wain-Hobson, complimented Hooper for pushing the Wistar to do these tests. “I think it was worth doing,” said Weiss. But, he added, “I'm slightly surprised that Hooper pooh-poohs it now.”

Undaunted by the test results, Hooper asserted that his thesis is actually “looking considerably more impressive today.” He said he recently had discovered “two smoking guns”: Two people who worked in Africa on the project whose first-hand accounts support the idea that Koprowski or his collaborators used chimpanzee kidney cells to make the vaccine. But Stanley Plotkin, a professor emeritus at the University of Pennsylvania in Philadelphia who helped Koprowski make and test the vaccine, said he has spent the past year contacting former collaborators, and 16 scientists have testified, in writing, that they never worked with chimpanzee cells. “I'm sure Mr. Hooper will be disappointed by the results of this meeting,” said Plotkin. “There is no gun. There is no bullet. There is no shooter. There is no motive. There is only smoke created by Mr. Hooper.”

Several scientists who once worked on Koprowski's OPV continued in that vein. During one particularly heated session, they began attacking Hooper's conclusions and accusing him of misrepresenting their thoughts in his book. The ad hominem attacks from both the scientists and Hooper prompted a call to order from the session chairman Neal Nathanson, who last month retired from the top AIDS job at the National Institutes of Health. “I insist on some civility or we'll simply close the meeting right now,” said Nathanson.

Several other scientists challenged the OPV theory with data rather than rhetoric. Bette Korber, an evolutionary geneticist at the Los Alamos National Laboratory in New Mexico, added details to a recent paper she published in Science (9 June, p. 1789) that dated the origin of the main type of HIV now in humans to between 1915 and 1941. Korber's computer modeling of different HIV strains has now answered a question raised by those data: If the HIV epidemic began in the first half of the century, why did it take until the 1980s to surface? She concludes that the virus appears to have spread extremely slowly at first, gaining momentum only after infecting thousands of people. In two separate presentations using independent techniques, Anne-Mieke Vandamme of Belgium's Rega Institute and Paul Sharp of the University of Nottingham came up with timelines similar to Korber's.

Hooper has hypothesized that chimpanzees from a colony in eastern Congo, on which Koprowski tested his polio vaccine, may also have been the main source of kidneys used to make the vaccine. But Beatrice Hahn of the University of Alabama, Birmingham, has shown that all five of the SIVcpz strains found to date that closely resemble the HIVs in humans come from chimps in western Africa; the only other SIVcpz discovered so far, which she believes came from the region where Koprowski had his chimp colony, is very different. Hahn has since found no evidence of SIVcpz in urine and fecal samples from 24 wild chimps in Uganda and 28 others in Côte d'Ivoire—two regions outside the range of west African chimps. And she reported finding SIVcpz antibodies in urine samples from a chimp in an eastern African country—which Hahn said would be “irresponsible” to name at this point—that appears to resemble the odd sixth sample. “Every piece of evidence we currently have would support the cut hunter theory,” said Hahn. “That alone blows OPV out of the water.”

But supporters of Hooper's theory remained unconvinced. Brian Martin, a social scientist from Australia's University of Wollongong, argued that if people scrutinized the natural transfer theory as closely as they have examined Hooper's scenario, it would prove to be just as unsatisfying. “There is one thing I will predict as a social scientist,” said Martin. “Whatever happens at this conference, this controversy will continue.”

2. NATIONAL SECURITY

# Relief, Rebukes Follow Agreement on Lee

1. Andrew Lawler*
1. With reporting by David Malakoff.

What began as an explosive case of alleged nuclear espionage petered out in an Albuquerque, New Mexico, courtroom this week. Once last-minute legal wrangling is complete, physicist Wen Ho Lee is expected to be freed after 9 months in jail. The ignominious collapse of the government's case and Lee's release have embarrassed federal prosecutors. However, the news was a relief to Asian-American researchers and others who say Lee's status as a suspect had heightened racial tensions at the national labs.

Lee acknowledged in the draft agreement that he had mishandled classified government data while working at the Department of Energy's (DOE's) Los Alamos National Laboratory in New Mexico. But that single felony count is a far cry from the 59-count indictment brought last December, when government officials warned darkly that Lee had given secrets about the design of sophisticated nuclear weapons systems to China. The charges sparked an outcry from Asian Americans, who complained that Lee was singled out due to his ethnicity. His solitary confinement, the use of shackles during his jail term, and limitations on his family visits outraged both Asian Americans and many scientific organizations (Science, 8 September, p. 1669).

The government's case suffered a final blow last month during a bail hearing when an FBI agent admitted that he had been wrong to assert that Lee's behavior was deceptive. The two sides then moved to reach a plea agreement, which was to be finalized 13 September in Albuquerque's Federal District Court.

Secretary of Energy Bill Richardson, who took intense political heat during the controversy even though Lee's actions occurred before his appointment in 1998, said on 11 September that he remains concerned about the whereabouts of tapes Lee made containing nuclear weapons data. “The plea bargain enables us to get that information,” he said. And in Congress, the deal eased pressure on New Mexico lawmakers, who for months have struggled to defend the Los Alamos lab against harsh attacks.

“This is a good plea arrangement,” said Senator Pete Domenici (R-NM), a powerful lab ally. But the case “uncovered systemic and deep-rooted [security] problems at our labs and within the entire DOE management structure.” Domenici is a strong backer of the new National Nuclear Security Administration, a semiautonomous agency within DOE formed in response to a range of concerns about security at the department's weapons facilities.

Relief was also evident among Asian-American scientists. “We finally have come to something sensible,” says Bryan Kashiwa, a Los Alamos fluid dynamics researcher, about the plea agreement. “It's the best deal he could get,” added Cheuk-Yin Wong, a veteran physicist at Oak Ridge National Laboratory in Tennessee. But Wong warned that DOE must revise its security procedures to avoid future debacles at its network of national labs. Although Wong thinks Lee should be punished for mishandling the data, he says, “we still have a long way to go to eliminate injustice.”

That view is widely shared by other Asian Americans. “This is the end of a nightmare for the Lee family,” says Henry Tang, chair of the Committee of 100, a group of influential Asian Americans. “But we feel the issue of ethnic profiling at the national labs should clearly be investigated. As Americans—not Asian Americans—we are very concerned that what appears to be a procedural violation at a national lab could land you in prison for life.”

Such criticism prompted Richardson in January to name an Asian-American ombudsman, Jeremy Wu, to handle diversity issues for the department. DOE also released a report on racial profiling that found widespread concern among Asian Americans about “insensitive and offensive” accusations of spying aimed at ethnic Asians, whether foreign or U.S. nationals.

Meanwhile, Domenici said that he would like to see the government drop its investigation of the loss and recovery of two computer disks at Los Alamos holding classified weapons information (Science, 23 June, p. 2109). In a fiery statement during a debate over DOE's funding bill, Domenici challenged the FBI, saying, “If you can't prove there is spying or espionage, you ought to get off their backs.” The FBI hasn't responded.

3. PLANETARY SCIENCE

# 'Spiders' Channel Mars Polar Ice Cap

1. Richard A. Lovett
1. Richard A. Lovett is a science writer based in Portland, Oregon.

REYKJAVIK, ICELAND—Scientists studying the latest high-resolution photos of the martian south polar ice cap think they may have found additional clues to its ebb and flow. These hints of the planet's bizarre atmosphere come from a new class of dramatic-looking terrain features whose dark, multilimbed, vaguely radial designs have earned them the moniker “black spiders,” and another group of dusky, spreading features called “dark fans.”

At a recent gathering of Mars researchers,* Hugh H. Kieffer, a planetary scientist at the U.S. Geological Survey in Flagstaff, Arizona, proposed that the spiders might be subsurface gas channels, visible through an unusually transparent section of the martian ice. Within the legs, he suggested, blow hurricane-speed jets of carbon dioxide generated as the spring sun vaporizes the CO2 ice deposited at the poles each winter. The jets may carry dust, he added, which spreads in fanlike shapes over the ice.

Steve Clifford, a planetary scientist at the Lunar and Planetary Institute in Houston, Texas, is excited by Kieffer's proposal, which he calls the first attempt to explain these features. Other scientists say the black spiders and the planet's other strange CO2 features are critical to understanding the martian atmosphere, one-third of which is deposited each winter as CO2 frost at the martian poles. Kieffer admits that his ideas are speculative but that the urge to interpret what he and others are seeing is irresistible. “I can make a wonderfully consistent story—which may or may not be what's going on,” he says.

During the spring, solar heating vaporizes up to 10 kilograms of CO2 per square meter per day, the equivalent of 1 cm of ice thickness. Kieffer proposes that black spiders, tens to hundreds of meters across, develop in regions where this vaporization happens not from the top down, but from the bottom up. The spider's legs collect gas from the transparent areas, conducting it beneath the surface to weak points, where it fountains free in roaring jets. Dust carried with the gas may then land atop the ice in spreading dark fans hundreds of meters in length. Although black spiders and dark fans have not yet been seen in tandem, Kieffer's theory suggests that they are linked, with the fans extending downwind from the vents of spiders that, for whatever reason, are not well enough defined to show up in satellite photos.

Kieffer's hypothesis requires the ice to be transparent, so that warm sunlight can penetrate deep enough for the resulting CO2 gas to have trouble breaking through to the surface. This would be no problem with pure CO2 ice, which is so clear that 75% of sunlight will penetrate at least a half-meter deep. But the CO2 ice that condenses out of the martian atmosphere isn't pure. Instead, a heavy peppering of dust makes it opaque.

Kieffer also has proposed a mechanism by which spring sunlight can purge dust from the ice. Sun-warmed dust motes, he says, should easily become hot enough to evaporate adjacent CO2. Near the surface, the vapor pressure may be enough to crack the ice, ejecting the dust in a puff of gas. Otherwise, gravity will cause the dust particles to tunnel ever deeper until they hit bottom, like a slow-falling rain. Because the martian polar atmosphere is predominantly CO2 and cools each winter to the freezing point of CO2, the microscopic tunnels are continuously reclosed with freshly deposited ice. The result is an increasingly transparent, self-sealing layer of CO2 ice.

Meanwhile, enough sunlight reaches the dark surface beneath the ice—perhaps a meter down—to evaporate more ice, allowing the CO2 gas to build up there. Kieffer predicts that small, branching gas channels will feed into larger ones, like merging raindrops on a windowpane. Where these big channels break through to the surface, he predicts that the gas could reach a velocity of 50 meters per second, fast enough to keep the jets open throughout the 150-day CO2 evaporating season. Clifford calls Kieffer's hypothesis “interesting” and “reasonably coherent” but notes that important details remain to be filled in. They include whether the proposed mechanism can produce gas channels large enough to be seen from space, and whether the martian CO2 ice is sufficiently clear for the bottom-up heating effect to occur. But he doubts that any of these will prove fatal to the basic theory.

Andrew Ingersoll, a planetary scientist at the California Institute of Technology in Pasadena, says black spiders and dark fans are just part of the “crazy stuff” that makes up the overall puzzle of the Red Planet's atmospheric dynamics.

• *The Second International Conference on Mars Polar Science and Exploration, 21–25 August, Reykjavik, Iceland.

4. BIOMEDICINE

# 'Glue Grant' Boosts Cell Signaling Consortium

1. Kathleen Fisher*
1. Kathleen Fisher is a writer in Alexandria, Virginia.

A “glue grant” may sound like financial support for a revolutionary adhesive, and in fact, that's not far off the mark. The National Institute of General Medical Sciences (NIGMS), in announcing its first glue grant last week, said it aims to use a novel funding approach to bind together researchers in cutting-edge fields at many institutions, allowing them to transcend their individual areas of expertise.

NIGMS has awarded $5 million a year for 5 years to a group of scientists studying cellular signaling. The project is headed by Alfred Gilman, chair of pharmacology at the University of Texas (UT) Southwestern Medical Center in Dallas and co-recipient of a Nobel Prize in 1994 for his work on “G proteins,” which act as gatekeepers for information entering cells. To speed their findings into the public domain and make them available for use in drug testing, Gilman and members of the project have agreed to post new results in a public database and forgo some patent and authorship claims. The group of 50 participating scientists at 20 universities, called the Alliance for Cellular Signaling (AFCS), expects to spend a total of$10 million a year on this work. They have raised half of this sum through pledges from individuals, institutions, and corporate backers, including Eli Lilly and Co., Johnson & Johnson, Merck, and Novartis, among others. The drug companies also have agreed to forfeit proprietary rights to alliance findings.

Researchers have identified thousands of cellular signaling molecules that carry information between and within cells. “The classic example is the fight-or-flight response,” said Gilman, in which signals from the brain trigger responses in the heart, blood vessels, lungs, and gastrointestinal system. The AFCS plans to chart interactions among these signaling molecules to produce a model of how mouse heart muscle cells (cardiac myocytes) and immune cells (B cells) respond to stimuli. Gilman foresees pharmaceutical companies using it to develop “a treasure chest of very specific drugs.”

Officials at NIGMS said that increased public funding, combined with Internet links that permit quick transfer of mammoth data sets, has made the time ripe for such huge cooperative studies. “This differs from anything else we're doing, or anything that we've done before, in that it doesn't supply underlying research support for investigators,” said Marvin Cassman, director of NIGMS. It will enable researchers “to reach a goal they can achieve only by working together.” Michael Rogers, director of NIGMS's Division of Pharmacology, Physiology, and Biological Chemistry, finds it “remarkable” that “traditionally independent-minded individuals” are now ready to join big collaborations.

Scientists participating in the alliance will be asked “to steer, to guide, to hypothesize, and to design models,” said Gilman. In addition, 250 “member” scientists around the world will host Web sites and augment the AFCS database. The glue grant will fund work in seven newly established labs. Stanford University will be in charge of microscopy; the California Institute of Technology, molecular biology. The San Francisco Veterans Affairs Medical Center will deal with signaling assays. The University of California's San Diego Supercomputer Center will be the hub of bioinformatics. UT Southwestern will have three new labs, one focusing on antibodies and one on each type of mouse cell under investigation.

Part of the glue that holds the alliance together is electronic. AFCS plans to communicate via Internet 2, a university-based system with enhanced bandwidth and speed. A virtual conferencing system will allow Gilman to meet simultaneously with 36 of his colleagues. Gilman said it was necessary to forfeit some intellectual property and first-time publication rights to allow “real-time” posting of group findings. “There will be publishing opportunities for the people employed in the labs,” he explained, “but it will be higher level interpretation, not conventional findings.” The community has “got to know that we're playing fair with them and that everybody's got an equal shot” at the alliance's data, he added.

Private donors may be sacrificing proprietary rights in the short run, but their investment may pay off later as their own researchers use the AFCS models to develop new drugs. Steven Paul, group vice president of Lilly Research Laboratories of Eli Lilly and Co., said his company made its $500,000-a-year pledge not out of philanthropy but because its scientists hope to glean immediately useful data. “This is a fantastic group of extraordinary scientists,” said Paul. “We feel strongly that there will be enormous public ramifications.” 5. U.S. POSTDOCS # Report Urges Better Treatment, Status 1. Jeffrey Mervis For years, U.S. postdocs have been complaining about paltry salaries, lack of benefits, and lowly status. This week, they won some high-level support. A committee of the National Academies of Sciences and Engineering and the Institute of Medicine has validated many of the complaints and lent its considerable weight to efforts to provide greater institutional support for postdocs. At the same time, however, the panel sidestepped two burning issues by explicitly declining to recommend a boost in postdoc salaries or take a position on whether to curtail the size of the postdoc workforce, which has more than doubled in the past 20 years to an estimated 52,000 (see graph). The recommendations are contained in a guide* issued this week by the academies' Committee on Science, Engineering, and Public Policy. The panel concludes that postdocs are “indispensable” to U.S. science but that low pay and uncertain job prospects have made them disgruntled. An electronic survey of leading research institutions conducted by the committee documents both the relative poverty and the precarious status of postdocs, including the fact that only about half of their academic employers provide them with vacation time and sick leave, and almost 60% give advisers complete control over the length of postdoctoral appointments. “Although many postdocs have stimulating and productive research experiences under the supervision of attentive, thoughtful mentors,” says panel chair Maxine Singer, president of the Carnegie Institution of Washington, many also receive “embarrassingly low pay and meager benefits.” The report says that the low salaries—averaging$28,000 for starting postdocs in 1998—are largely the result of a decision by universities not to supplement National Research Service Awards (NRSA), stipends provided by the National Institutes of Health (NIH) to cover training expenses. By not doing so, the panel notes, universities have made the NRSA levels the “de facto funding standard” for all academic-based postdocs.

Apart from pay, the panel urges institutions to adopt a common definition for postdocs and policies for their appointment, training, compensation, evaluation, and career guidance. It also recommends that universities set up a central office to handle postdoc affairs and emphasizes that faculty members should view postdocs as “apprentices” who require mentoring rather than as a “pair of hands” to carry out research at the bench. “Everybody has to ante up,” says Singer about the issues facing the scientific community. “If everybody points to somebody else, then nothing will happen.”

Several academic administrators and science managers give the report high marks. “I think the scientific community would be well advised to take these recommendations very, very seriously,” says Michael Teitelbaum, program director for the Alfred P. Sloan Foundation, which helped pay for the study and which is supporting the creation of a national postdoc network (nextwave.sciencemag.org). Joel Oppenheim, an associate dean at the New York University School of Medicine, also welcomes the panel's advice but adds, “it's just a report. The real power to change things lies with the funding agencies, in particular NIH and the National Science Foundation.” Walter Schaffer, research training officer at NIH, which helped fund the study, says that “I think they did a heck of a job. Most of what they are saying is right on.”

Some observers claim, however, that the academies' panel downplayed what they see as the “exploitation” of postdocs by institutions that depend on them to get the work done. “They don't want to alienate the university faculty, who would have to pay higher salaries out of their grants,” says Letitia Yao, a former chemistry postdoc and current staff member at the University of Minnesota, Minneapolis, who helped form one of the first postdoc associations at the University of California, San Francisco. “It all comes down to money: If institutions were paying postdocs 45 or 50 thousand [dollars], they'd also treat them right. You wouldn't even need a guide.”

Jack Bennink, a section chief at the National Institute of Allergy and Infectious Diseases, believes that the status of postdocs is a moral as well as an economic issue. “In many cases their treatment borders on abuse and exploitation,” he says. At the same time, Bennink agrees with Schaffer and others that the best course is “to make small fixes on a problem that is really, really complex.”

With three-quarters of the postdocs working in the life sciences, many officials look to NIH for answers. And they see its growing budget as a painless way to boost salaries without trimming the number of postdoc slots and disrupting research. Schaffer agrees that the report puts pressure on NIH to raise its NRSA stipends from the current $26,916 starting point. But he says that it's not clear what the standard of comparison should be. “We need to figure out what's reasonable,” he says, “and it should probably be on a cost-shared basis with universities.” 6. ASTROPHYSICS # Neutron Stars Linked to Celestial Runaway 1. Govert Schilling* 1. Govert Schilling is an astronomy writer in Utrecht, the Netherlands. A million years ago, in the constellation Scorpius, one of a pair of binary stars erupted into a supernova. Its nonexploding partner shot off into space and is now Zeta Ophiuchi, a bright, giant “runaway star” racing through the neighboring constellation Ophiuchus. The supernova has been harder to trace. Astrophysicists know it must have collapsed into a neutron star, but where it wound up has been anyone's guess. Now, however, astronomers are fingering two candidates, one of which is the closest known neutron star to Earth. The discovery, if confirmed, will give astronomers a better understanding of the dynamics of supernova explosions in binary systems and the origin of runaway stars. The nearby neutron star, known as RX J1856.5-3754, was first detected by the German ROSAT x-ray satellite, and it was spotted in visible light in 1997 by Fred Walter and Lynn Matthews of the State University of New York, Stony Brook, using the Hubble Space Telescope. New Hubble images, taken in March and September 1999, enabled Walter to calculate both the faint star's distance from Earth—a mere 200 light-years—and its proper motion, or apparent path across the sky. Traced backward, Walter says, “the proper motion brings the neutron star from the general vicinity of the Upper Scorpius association”—the group of bright, young stars in which Zeta Ophiuchi was born. In a paper scheduled for publication in the 10 January 2001 issue of The Astrophysical Journal, Walter theorizes that RX J1856.5-3754 is the collapsed core of the supernova that flung Zeta Ophiuchi into space. Because the neutron star's radial velocity (its movement toward or away from Earth) is unknown, astronomers can't tell whether the path leads into the association or passes in front of it. But Walter calculates that if the neutron star came from the association, it did so 1.15 million years ago—just when Zeta Ophiuchi made its own explosive exit. That timing matches evidence of the star's youth, Walter says. Lone neutron stars are expected to cool down over time, yet RX J1856.5-3754′s x-ray brightness indicates that its surface is blazing away at more than 500,000 K. Such a hot neutron star, he concludes, must be young. Some other astronomers, however, believe RX J1856.5-3754 may be far too old to have been Zeta Ophiuchi's companion. Marten van Kerkwijk of Utrecht University in the Netherlands and Shri Kulkarni of the California Institute of Technology in Pasadena announced this week that observations with the European Southern Observatory's Very Large Telescope in Chile show that RX J1856.5-3754 is trailing a small, faint cone of glowing hydrogen gas, supposedly heated by the star's intense x-rays. From the brightness of the glow, the scientists calculate that the gas near RX J1856.5-3754 must be about 100 times as dense as the galactic average. The gas would slam onto the compact star at half the speed of light, heating the surface and making it appear younger than it actually is. With so much matter available to heat it, Van Kerkwijk says, RX J1856.5-3754 could well be billions of years old—thousands of times too ancient to have been born in Upper Scorpius. But Walter argues that if the star is that old, it would have attracted enough interstellar hydrogen to make its visible light hundreds of times brighter than astronomers observe. Van Kerkwijk thinks a much more likely candidate for Zeta Ophiuchi's erstwhile partner is a radio pulsar known as PSR J1932+1059. Pulsars are spinning neutron stars that emit radio pulses. From the rate at which PSR J1932+1059′s rotation is slowing down, Van Kerkwijk calculates that the pulsar is at most a few million years old. In another Astrophysical Journal paper, scheduled to appear in October, Ronnie Hoogerwerf, Jos de Bruijne, and Tim de Zeeuw of Leiden Observatory in the Netherlands show that PSR J1932+1059 also left the Upper Scorpius association 1 million years ago, assuming that its radial velocity—also unknown—is about 200 kilometers per second. Walter agrees that PSR J1932+1059 could be the former binary companion of Zeta Ophiuchi. But he believes RX J1856.5-3754 is young enough to be a candidate as well. In any case, Walter says, it's perfectly possible that the hot neutron star and the young pulsar flared into being in the same part of the sky at about the same time. “There was at least one supernova in Upper Scorpius about 1 million years ago. Why not two?” 7. APPOINTMENTS # Salk Institute Goes North for New CEO 1. Robert F. Service* 1. With reporting by Wayne Kondro in Ottawa. The revolving door at the top of the Salk Institute for Biological Studies in La Jolla, California, took another spin last week with the appointment of neuroscientist Richard Murphy as president and CEO. Murphy, director of the Montreal Neurological Institute (MNI), will become Salk's fourth chief executive in 4 years when he takes up the reins on 1 October. Murphy, 56, says his main job will be to raise enough money to keep the endowment-poor research institute in the scientific big leagues. Founded in 1960 by polio vaccine developer Jonas Salk, the institute has become a dominant player in molecular biology and genetics. But Salk executives have had a difficult time translating scientific success into long-term financial health. Part of the problem is that Salk, dedicated exclusively to research, does not graduate students and is therefore not blessed with generous alumni. More recently, however, the institute's frequent changes in leadership have created another problem—a suggestion that the institute lacks a clear sense of direction. Born and trained in the United States, Murphy says that he doesn't plan to run a research group at Salk and that “this will be a full-time job with all the fund raising involved.” At MNI he raised more than$25 million during his 8-year tenure, allowing the institute to hire 20 new faculty members and carry out needed renovations. “He rejuvenated that institute,” says Salk molecular biologist Tony Hunter.

The need to raise money isn't a new task for the head of Salk, of course. Murphy's immediate predecessors, Salk structural biologist Thomas Pollard and current CEO Frederick Rentschler, managed to boost Salk's endowment from under $50 million in 1996 to$120 million today, although Pollard stepped down as CEO a year ago to focus on his research. Even with that success, Salk's endowment is only one-tenth the size of those of similar places like the Rockefeller University in New York City, notes Charles Stevens, a Salk neuroscientist and member of the search committee that tapped Murphy.

Large endowments allow top-tier research institutes to pay for expensive equipment such as gene chip arrays for genetics and nuclear magnetic resonance machines for structural biology, says Hunter. Not having that pot of money, he adds, “makes it harder for us to compete.” It also prevents Salk from providing much salary support for its 54 faculty members and puts it at a disadvantage in recruiting, says Stevens.

Hunter, Pollard, and others are optimistic that Murphy can keep Salk in the race. “It's a terrific selection,” says Pollard. And although Salk's scientific luster doesn't need much polishing, Hunter hopes that Murphy's arrival will also stop the revolving door: “We would love to have someone who sticks around for a while.”

8. MOLECULAR BIOLOGY

# Cancer Fighter's Modus Operandi Revealed

1. Jean Marx

Researchers have deciphered how a promising cancer drug acts like a smart bomb, homing in on only a very narrow range of its potential targets in the cell. The compound, known as STI-571, has shown remarkable success in early clinical trials on patients with chronic myelogenous leukemia (CML). Now, in work reported on page 1938, John Kuriyan and Thomas Schindler of the Rockefeller University in New York City and their colleagues reveal just how the compound works—information that could aid in the design of similar cancer therapies. “It's a very neat story,” says cell biologist Tony Hunter of the Salk Institute for Biological Studies in La Jolla, California.

Scientists already knew that STI-571 blocks the enzyme produced by the abl oncogene, whose activation has been linked to the massive proliferation of leukemia cells in CML patients. They have been hard put to explain, however, why the compound doesn't also block closely related enzymes. The abl oncogene is one of hundreds of genes identified over the past 25 years that can, when abnormally active, lead to cancer. The hope is that the proteins made by these oncogenes will provide good drug targets. But many oncogenic proteins, including the Abl protein, belong to one of the largest enzyme families in the cell—the protein kinases, which transfer a phosphate group from ATP to proteins. Thus any effort to snuff out an aberrant kinase could easily produce a great deal of collateral damage and unacceptable side effects for cancer patients.

But STI-571 is a notable exception. It was identified in the early 1990s by scientists at the pharmaceutical company Novartis, who found that it inhibits the kinase that acts as the receptor for platelet-derived growth factor. Subsequent tests by Brian Druker's group at the Oregon Health Sciences University in Portland showed that the compound, a small 2-phenylaminopyridine, also inhibits the oncogenic form of Abl and the c-kit kinase, but none of the 50 or so other kinases screened.

The discovery that STI-571 blocks Abl activity raised the possibility that it might be used to treat CML. Bone marrow transplants offer a potential cure, but suitable donors can be found for only one-third of patients. And current drug therapies, usually with interferon _, only control the disease for a few years before it eventually progresses to the acute—and fatal—stage. One of the diagnostic hallmarks of this cancer is an abnormal chromosome—the so-called Philadelphia chromosome—formed when a portion of chromosome 22 fuses with the chromosome 9 segment bearing the abl gene. As a result, a portion of the bcr gene becomes attached to abl, and for reasons that are poorly understood, the hybrid protein produced by this fusion gene shows increased kinase activity. What's more, work in animals has shown that the Bcr-Abl protein is all that it takes to cause leukemia.

In the past 2 years, Druker has coordinated a series of trials to determine whether STI-571 does in fact help CML patients, and so far the work looks very promising. For example, at last year's meeting of the American Society of Hematology, he presented the results of a phase I clinical trial—which is primarily aimed at determining tolerable doses—involving 54 patients who no longer responded to interferon _. The cancer cells seemed to disappear from the blood of all 31 of the patients who got doses of 300 milligrams and above, Druker says. What's more, 30 of those patients have remained in remission for about a year of follow-up, and they have reported only mild side effects including some nausea, diarrhea, and muscle cramping.

Druker and his colleagues have since concluded enrollment in a phase II trial to test STI-571's efficacy and have embarked on a phase III trial pitting the drug against standard CML treatments. Researchers familiar with the trials are already enthusiastic. “I can certainly say the results are terrific. It's a home run,” says Owen Witte of the University of California, Los Angeles, School of Medicine, whose own work was instrumental in showing Abl's importance in CML. James Griffin, a leukemia expert at Harvard Medical School in Boston, agrees, predicting that “this is going to change the way we treat CML.”

The Rockefeller team's work now explains how STI-571 homes in on the Abl kinase. The researchers crystallized the catalytic region of the human Abl protein together with an STI-571 variant. They then used x-ray crystallography to determine the three-dimensional structure of the drug-protein complex. Abl, like many other kinases, has an “activation loop” that has to have a phosphate group added before the enzyme can add phosphate groups to other proteins. This alters the shape, or conformation, of the enzyme, opening it up so that the kinase can bind ATP and its target proteins. What the crystal structure reveals, Kuriyan says, is that STI-571 binds to the inactive conformation of Abl.

This presumably prevents it from acquiring the activating phosphate, thus locking Abl in its inactive conformation, a mode of action that Witte says “makes good sense with how [STI-571] works in the leukemia.” Targeting the inactive conformation also explains why so few other kinases are inhibited by the drug. “When the kinases are active, they look very similar,” Kuriyan says. “But when they are turned off, they can look very different from one another.” He also notes that drug developers tend to think in terms of designing enzyme inhibitors that bind to the active enzyme, but for the kinases the inactive forms may make better targets.

A good many questions remain to be answered about STI-571. Clinicians will want to know, for example, how long its effects can be maintained in patients and whether it will work against solid tumors as well as CML. There are hints that it might. For example, Griffin's team has shown that it inhibits the growth in culture of small lung cancer cells, a cancer in which the kit oncogene kinase is often activated. But researchers are already pleased that what they have learned about the gene changes leading to cancer is beginning to pay off. “It's total vindication of the need to do basic science on the mechanisms of cancer,” Witte says.

9. EUROPEAN SCIENCE

# Call to Arms for Life Scientists

1. Robert Koenig

GENEVA—Seeking to create “a force for change in European research,” nearly 2000 European life scientists gathered here last week at their first congress to promote solidarity, forge collaborations, and air complaints about how European managers dole out science funding.

“There is an enormous amount for [us] to do, when we compare the European situation to that in the United States,” says cell biologist Kai Simons, president of the new European Life Scientist Organization (www.elso.org). Modeled in part on the American Society for Cell Biology, ELSO sprang from the brows of Simons and several other alumni of the European Molecular Biology Laboratory (EMBL), who saw the need for an organization to help unite Europe's molecular life scientists and to lobby Brussels to improve European Union (E.U.) research policies.

Simons—who recently became head of the Max Planck Society's new Institute for Molecular Cell Biology and Genetics in Dresden—and many others are dissatisfied with the E.U.'s flagship research effort, Framework 5. The 5-year, $17 billion program restricts much of its spending to priority areas. Innovative research tends to slip through the cracks, Simons contends. Another common complaint, he says, is that Framework mostly ignores young researchers, leaving it to the national programs to shoulder much of the support for grad students and postdocs. Simons and other ELSO council members are bringing their guns to bear on the architects of Framework 6, a 5-year portfolio to start in early 2003. The organization's key aims are for the new Framework to offer more grants for postdocs and far more “generic” funds for research that does not fit into Framework spending categories. ELSO's leaders have already presented Brussels with a list of priorities for Framework 6, and the organization will now push to add a new grant category to allow more independence for young European researchers. “A lot of European scientists complain about the Framework Program, but we don't do anything about it. This is going to change,” says the Finnish-born Simons. The group also intends to fight for better job opportunities for female scientists and to stimulate mobility and collaborative research among European life scientists. The meeting featured presentations by four Nobel laureates and symposia on topics ranging from cell death to trends in mammalian genetics. Although most conferees hailed from Germany, France, and Switzerland, more than two dozen came from Eastern Europe, where Framework is just taking root. ELSO “can help draw good scientists in Eastern Europe into the mainstream, and to lobby Brussels on the importance of expanding research programs there,” says ELSO council member Maya Simionescu, who directs Romania's Institute of Cell Biology and Pathology. Although there are several Europe-wide life sciences organizations, ELSO organizers say their group is different because it does not restrict who can join, and it is not a federation of national societies. “It's important to have a bottom-up organization that allows life scientists from different fields to develop some priorities and to lobby,” explains ELSO council member Denis Duboule, a University of Geneva biologist. All meeting attendees became enrolled in ELSO; European newcomers are also welcome. Now funded mainly by corporate grants, ELSO won't charge dues until it has “proven its worth” through lobbying, communications, and conferences, Simons says. To keep up team spirit between annual gatherings, ELSO has launched a free bimonthly magazine, The ELSO Gazette (www.the-elso-gazette.org). It will feature research reviews by “up and coming young European scientists,” along with news articles about European research, job listings, and editorials, says editor Carol Featherstone, a former EMBL cell biologist. Although attendance in Geneva reached only half the original goal, Simons says he is pleased. “Up until this conference, ELSO was a virtual organization,” he says. “Now it is real.” 10. X-RAY SCIENCE # French 'Sun' to Rise at Site Near Paris 1. Michael Balter PARIS—The sun is shining on French science this week with the selection of a site outside Paris as the home of the country's first “third-generation” x-ray source. On Monday research minister Roger-Gérard Schwartzenberg announced that the machine, called SOLEIL, or “sun,” would be built near Saclay, about 20 kilometers southwest of Paris. The project had been cancelled by his predecessor, Claude Allègre, who feared that its$200 million price tag for construction and 8 years of operation would pinch other research budgets. But Schwartzenberg said that the national government's share should not exceed 20%, with regional and local authorities paying 75% and the rest coming from the United Kingdom, Spain, Belgium, and Portugal. Paris regional authorities have said they hope the new machine will attract companies to the area, which already boasts several universities and research institutions. SOLEIL's construction will begin in fall 2001, and it should go online in 2005, Schwartzenberg said.

“France needs to have a third-generation synchrotron on its own soil,” Schwartzenberg said, adding that several other European countries have their own state-of-the-art machines. High-powered x-rays will allow researchers to probe the atomic structures of biological molecules and industrial materials at resolutions of just a few angstroms. The accelerator will have an energy of 2.5 to 2.75 giga electron volts and 24 beamlines for experiments.

Schwartzenberg said that Saclay beat out a site near Lille as the home for SOLEIL. France has already agreed to support the planned Diamond synchrotron in Britain (Science, 6 August 1999, p. 819).

11. ECOLOGY

# Louisiana's Vanishing Wetlands: Going, Going ...

1. Joel Bourne*
1. Joel Bourne is a writer in Silver Spring, MD.

An ambitious $14 billion plan aims to restore Louisiana's wetlands, which are disappearing at record pace. But the scientific and political hurdles are huge What do you do if you are one of the poorest states in the nation, have a projected budget shortfall of more than$200 million, and are losing Rhode Island-sized swaths of land every 50 years? If you are Louisiana, you propose the largest, most expensive environmental restoration project on the planet —and hope the federal government picks up the tab.

The $14 billion plan known as Coast 2050 attempts to protect more than 10,000 square kilometers of marsh, swamp, and barrier islands, an area nearly twice as large as Florida's Everglades—the nation's most beloved swamp and the former record holder for an environmental salvage project. Although the plan has yet to be funded by Congress, current legislation sponsored by Senator Mary Landrieu (D-LA) would tap royalties from federal oil and gas leases on the outer continental shelf to help pay for more than 500 projects ranging from replanting marsh grasses on barrier islands to building locks on the Mississippi River. A$6 million feasibility study of wetland restoration projects designed for Barataria Basin, just east of the Mississippi River, is already under way; state leaders hope it will spur on the larger project.

“This is the major delta system on the North American continent, and it is dying,” says biologist Bill Good, director of the Coastal Restoration Division of Louisiana's Department of Natural Resources (DNR) and one of the driving forces behind Coast 2050. “It has a tremendous impact on the overall ecology of the Gulf of Mexico. I mean, God bless the Everglades, but God bless the Mississippi Delta, too.”

Coast 2050 is the latest and by far the most ambitious of a series of coastal management and restoration plans that date to the early 1970s, when scientists first measured dramatic rates of land loss in Louisiana's coastal zone—roughly all lands lying south of Interstate 10, which stretches from Slidell to Lake Charles. Between 1956 and 1990, some 3460 square kilometers of coastal wetlands reverted to open water. The state continues to lose between 65 and 91 square kilometers each year, one of the highest rates of land loss in the world. Louisiana alone contains 40% of the wetlands in the contiguous United States, yet accounts for 80% of wetland loss in the Lower 48 states. At this rate, even with current conservation projects under way, another 1800 to 4500 square kilometers will vanish under the Gulf in the next 50 years—and that area could conceivably include the city of New Orleans. The loss of public resources, including fisheries, wildlife habitat, navigation, flood control, and hurricane protection, has been estimated at more than $37 billion. Previous attempts to help the state's wetlands led to passage of the 1990 Coastal Wetlands Planning, Protection and Restoration Act, which was sponsored by Senator John Breaux (D-LA). The “Breaux Act” currently funnels about$40 million to $50 million annually into the state for wetlands restoration projects. Yet wetlands advocates contend that the program, which has funded mostly small projects equally distributed throughout the state, has served as little more than the proverbial finger in the dike. In the mid-1990s the Coalition to Restore Coastal Louisiana, an influential advocacy group, pushed for a single, overarching restoration and management plan that would meld federal, state, and local efforts with public participation to create a focused blueprint for saving the state's beleaguered wetlands. With the support of Louisiana Governor Mike Foster and Secretary Jack Caldwell of the DNR, Good led dozens of state, federal, and university scientists in an 18-month effort that selected projects from previous plans and also included new techniques. Those include methods for predicting future land loss, the first coastwide assessment of subsidence rates and patterns, and the first comprehensive look at changes in fish and wildlife populations. Public hearings were held in the 20 coastal Louisiana parishes, all of which eventually passed resolutions supporting the plan, which was released in December 1998. “We have a chance to plan and design a major natural system—a world-class system—that produces renewable resources and that can sustain itself and sustain human activity,” says coastal geologist Sherwood Gagliano of Coastal Environments Inc., one of Coast 2050's architects and the researcher who first raised the alarm about coastal land loss in the early 1970s. “I'm not sure that's been done anywhere, not on this scale.” Other wetlands experts aren't as optimistic about the ability of the U.S. Army Corps of Engineers to correct the damage they largely helped create. “The Army Corps' reputation is not good,” says ecologist Stuart Pimm of Columbia University. “They like highly managed systems. My rule of thumb when it comes to ecosystems is larger is better than smaller, connected is better than fragmented, and natural is better than managed.” ## Sinking and slumping To succeed, coastal scientists have to overcome three major obstacles that are fueling wetland demise, according to Gagliano. The first is called concentrated margin gravity slumping, or fault-induced sinking. Geologists working for the oil and gas industry have long known that the entire delta region is chopped up into thousands of pieces by subsurface fault zones that crisscross the delta from east to west. The seaward block, consisting of about 4100 square kilometers, has been slipping southward since the turn of the century. As it slips, the Gulf of Mexico invades. Gagliano estimates that such slumping may have caused as much as 60% of the land loss that has occurred since the 1890s. The sinking rate began to accelerate in the 1960s, largely because of the second major obstacle: the levying of the Mississippi River. Unlike the crisis in the Everglades in which a naturally low-energy freshwater system is suffering from nutrient overload, Louisiana's wetlands are starving for nutrients, sediments, and water. Big Muddy was once the greatest land builder in the world, dumping an estimated 400 million tons of sediment a year on the delta. It created six different delta lobes during the last 7000 years, jumping its bed every millennium or so in its rush to find the path of least resistance to the Gulf. As the old deltas withered, leaving fringing barrier islands, new ones formed, nourished by spring floods. The natural overflow of fresh water, sediments, and nutrients kept the wetlands fresh and aggressive, according to Gagliano. Marshes were able to overcome their natural sinking and dewatering by constantly increasing their vegetative surface. The devastating flood of 1927 that drove nearly 1 million people from their homes and inundated more than 70,000 square kilometers led to the Depression-era construction of high concrete levies that now line the banks for nearly 2000 kilometers. The levies did end the spring floods, but with a cost. Meanwhile, four major upstream dams built on the Missouri River trapped more than half of the river's sediment load. Most of the remaining 150 million tons of clay, silt, and sand that could sustain the delta now scour out the river's ship channel before flowing into a deep offshore trench in the continental shelf. Navigation channels and canals comprise the third major problem. More than 13,000 kilometers of canals lace Louisiana's marsh, along with nine major shipping lanes that have made the state the nation's leader in shipping tonnage. The canals—built primarily by the oil and gas industry—and shipping lanes are directly responsible for up to 30% of the state's current wetland loss. They are indirectly responsible for even more, because they have drastically changed the hydrology of the marsh. Driven by winds and tides, salty Gulf water that was once relegated to the coastal fringe now flows up the humanmade waterways far inland, causing massive “brownouts” or marsh die-offs along the way. “Essentially, you are bringing salt water and tidal action into a freshwater regime,” says Good. “We're trying to reduce that tidal energy and marine influence in those fresh and brackish water systems.” One of the worst offenders has been the Mississippi River Gulf Outlet (MRGO). Built in the 1960s to give large container ships access to New Orleans, the channel is now lined with forests and marshes killed by saltwater intrusion. Moreover, ship wakes are eroding a 60-kilometer stretch of the northern bank at an average of 5 meters every year. ## Heavy lifting To combat these problems, Coast 2050 would plug some canals and close MRGO as soon as another container facility is completed farther downstream. More controversial, however, are plans to relocate the Mississippi River Navigation Channel from the lower river to capture the bed load currently lost to the deep waters of the Gulf. This would entail building a new channel to the east or west of the current river entrance with at least two locks—a plan that is anathema to powerful shipping interests along the river because of the potential delays and the possible shoaling of the main navigation channel. In the more immediate future, however, Coast 2050 depends largely on water diversions, both large and small structures, that scientists hope will recreate at least part of the river's former natural functions by spilling water back into the marshes. The prototype for such efforts is the Caernarvon Freshwater Diversion, a$26 million opening in the levy built by the Army Corps about 16 kilometers south of New Orleans. Authorized by Congress in 1965 but not funded until the mid-1980s, the gated concrete culvert can divert up to 240 cubic meters of river water per second into a canal that feeds into the marshes behind Breton Sound, an area that had been losing up to 400 hectares a year. After 3 years of operation, Good's staff measured a 6% annual increase in marsh within the study area and a sevenfold rise in the population of freshwater marsh plants. Oyster numbers in the public seed grounds of Breton Sound jumped by more than three orders of magnitude, while waterfowl, alligators, and muskrats flourished in the rejuvenated marsh. An even larger diversion known as Davis Pond is scheduled to come online this year. Coastal planners estimate that the two separately funded diversions, combined with current Breaux Act projects, will prevent up to 22% of the projected loss by 2050.

“From what we know about how the marshes are naturally formed and from research that's been done at Caernarvon and other areas, the use of sediment diversions appears to provide a feasible means of large-scale marsh restoration,” says Irving Mendelssohn, a coastal ecologist at Louisiana State University (LSU) in Baton Rouge who has studied the region's wetlands problems for more than 20 years.

Louisiana's barrier islands, the state's first line of defense against hurricane storm surge, are another matter. After comparing 130 years of data depicting shoreline positions, geologist Randolph McBride of George Mason University in Fairfax, Virginia (formerly of LSU), found that some islands are eroding at up to 20 meters per year. Others have lost half of their land area over the last century. According to McBride, several islands, including Isles Dernieres, Timbalier Island, and the Grand Terre islands, will disappear within the next 20 to 50 years if left to their fate. This would leave large swaths of wetlands exposed to the direct erosive force of the Gulf. Although the exact buffering effect of the barrier islands on storm surge is unknown, Joseph Suhayda of LSU's Department of Civil and Environmental Engineering, using computer modeling, estimated in 1997 that certain configurations of islands and inlets along the coast could reduce storm surges inland by more than 1 meter.

Under Coast 2050, Louisiana's barrier islands would be restored or maintained using the most cost-effective means. This would most likely include beach nourishment with dredged material combined with marsh creation projects on the bay side of the islands, although hard structures such as sea walls and groins are also being considered. Such hardening of the shoreline worries Orrin Pilkey, a coastal geologist at Duke University, who specializes in barrier island migration.

“Two sea walls on East Timbalier Island are already out to sea,” says Pilkey. “In the context of rising sea levels, I don't see how they'll be able to stop the islands from disappearing. We have to look at this realistically and recognize that we're going to lose a lot of stuff.” Pilkey thinks the best chance of success lies with a plan to replant marsh grasses on the inland side of the islands and then pump dredged material behind them to give them a platform on which to migrate.

## Competing interests

A number of unanswered questions swirl around the plan. Scientists still don't know how much fresh water, sediment, or nutrients a marsh needs to thrive or when it needs them. Nor do they know if diversions such as Caernarvon and Davis Pond will funnel too many nutrients into the marshes, leading to eutrophication and algal blooms. And the jury is still out on the exact rates of subsidence and sea level rise.

“Sometimes you can't wait until all the answers are in,” says Mendelssohn. “You have to start the management process, but you have to be able to change that management approach and modify those procedures when new information becomes available.” Such “adaptive management” and monitoring components have been built in to many Coast 2050 projects.

Nonetheless, Mendelssohn believes nature has provided scientists with a good model of how large diversions will work: the Atchafalaya River. Capturing some 30% of the Mississippi River's flow at the Old River Control Structure near Simmesport, the Atchafalaya River is producing the healthiest marshes in the state as well as a small but growing delta that appeared after the 1973 flood. It's one of the few places in Louisiana where marshes are actually expanding.

“We've been studying these wetlands for 30 years,” says James Johnston, chief of the Spatial Analysis Branch at the U.S. Geological Survey's National Wetlands Research Center in Lafayette and co-chair of the Coast 2050 monitoring program. “A lot of what needs to be done is just common sense.”

The scientific questions may be the easiest to resolve. Currently, the state is involved in five lawsuits related to reduced oyster harvests allegedly caused by the Caernarvon Diversion, which has pushed the freshwater gradient farther toward the Gulf. Political opposition has kept Caernarvon running at roughly half of its capacity. Shrimpers, farmers, the petroleum industry, shipping interests, and low-lying municipalities are heavily vested in the status quo, which could change dramatically if Coast 2050 accomplishes its goals. Some towns, for instance, may be forced to relocate to avoid the risk of flooding. Nonetheless, President Clinton recently threw his support behind the Landrieu bill, known as the Conservation and Reinvestment Act, which seems to be gathering momentum in Congress. The bill would return most of the federal royalties on outer continental shelf oil and gas leases to states where offshore drilling is allowed. Coastal states such as Alaska, Mississippi, and Louisiana would receive the lion's share of the money, although other states that are landlocked would benefit as well. The House of Representatives approved the bill in May with a two-thirds majority, but it faces a stiff floor fight in the Senate.

Coast 2050's staggering price tag alone is enough to raise a red flag for ecologist Pimm, who has been a vocal critic of the Everglades restoration. “Water projects in the U.S. have a reputation for embodying the worst excesses of pork-barrel spending,” says Pimm, and Coast 2050 may be no exception. “What we'd really like for the Mississippi is to take out all the dams and levies and diversions and let the damned thing flow free.”

Even if the big money doesn't arrive, the delta's defenders say they will not be deterred. They contend that the government will either spend billions now to save the marshes or many more billions later to bail out New Orleans, half of which already lies below sea level. “We'll do whatever we can,” says DNR's Good. “Even if we have to fill sand bags and throw them into the breach.”

12. COMPUTER SCIENCE

# Flushing Out Nasty Viruses in the Balkans

1. Robert Koenig

Bulgaria created a pioneering center to tackle the threat that its homegrown viruses pose to the world; now the lab is struggling to stave off obsolescence

SOFIA—Shortly before its corrupt Communist regime toppled a decade ago, Bulgaria was overrun by a spate of devastating infectious agents that went by names such as Dark Avenger, Anthrax, and Evil. Cooked up by shadowy figures in the besieged country, these plagues were not the errant concoctions of Soviet-era bioweapons labs. They were the scribbles of computer geeks.

How this mostly rural Balkan country with a frayed telecommunications infrastructure became a “virus factory” is a tale wrapped in Cold War intrigue. Just as compelling is how the Bulgarian Academy of Sciences responded to the crisis. The academy established a National Laboratory of Computer Virology, where antivirus hunters match wits with unseen virus creators. The laboratory has been a major force in reining in Bulgaria's viral threat, says Lars-Olof Stromberg, a virus expert with the Royal Institute of Technology in Stockholm.

While Bulgaria's significance as a computer virus reservoir has waned, the virology lab continues to serve as a training ground for antivirus experts, including a few who made major contributions during recent efforts to disarm high-profile scourges such as the ILOVEYOU virus. Despite a ludicrous budget—the government gives the lab roughly half of what a single Western whiz kid fresh out of college might earn in a year—the facility remains a force to be reckoned with. “They are certainly good at looking at emerging viruses and analyzing them quickly,” says Fred Cohen, a researcher at Sandia National Laboratories in Livermore, California, who in 1983 coined the term “computer virus.”

## A Balkan Silicon Valley

No one knows where the first viral-type programs came from; they may have originated in instructions given to early computers to fill memory space by copying bits. In the late 1970s, some malignant codes infected Russia's Rijad mainframe computers. Then a few years later a virus known as the “Xerox worm” (so named because it made copies of itself) blighted a U.S. computer network.

It was not until the late 1980s that the public at large became acquainted with the destructive powers of viruses. That's when a rapidly proliferating virus from Israel dubbed Jerusalem bogged down computers around the world, and Brain—a “boot-sector infector” from Pakistan that hit MS-DOS systems—went on a rampage. Bulgaria then put itself on the map, unleashing a slew of globe-girdling viruses in the late ‘80s and early ‘90s. In the years since, the rogues' gallery has swelled to nearly 50,000 known viruses worldwide, with 10 to 15 new ones popping up every day. This proliferation has turned antivirus and computer security R&D into a multibillion-dollar business (see table).

Bulgaria's contributions to this global scourge trace their roots to the 1970s, when Soviet planners designated the loyal satellite as the lead East Bloc country for producing computers and software. Implementing that decision, Bulgarian Communist Party leader Todor Zhivkov decreed that personal computers would be manufactured in his hometown of Pravetz, which lent its name to two generations of PCs. Unable to import user-friendly Western software, a cadre of computer-literate Bulgarians soon became intimately familiar with the code that runs computer programs—knowledge that helped them understand how viruses operate, says Klaus Brunnstein, who founded the University of Hamburg's pioneering “Virus Test Center” in 1988.

Several theories have been put forward to finger who among the Bulgarian population turned their hands to virus writing. Some say the culprits were a handful of talented but frustrated young computer experts who were looking for an escape from the economic and social turmoil of the late 1980s and early ‘90s. Others see more insidious forces at play. Stromberg, for one, asserts that Bulgaria was a center of the once top-secret “InfoWar” effort, initiated by the KGB in the late 1970s to develop software and viruses that could be sicced upon the West. After the Soviet Union dissolved, he contends, some of those highly trained experts left the government “to freelance and to hack Western computer systems”—sometimes for organized crime rings.

Bulgarian experts contacted by Science demur, insisting that viruses sprang from the brows of amateurs, not spies. Thousands of computer-savvy teens cut their teeth in neighborhood “computer clubs” set up by the government in the 1980s. Although most of their activities were benign, some learned how to pirate Western software, and a few dallied in viruses. “The first viruses created here in Bulgaria were innocent things—just kids who experimented,” claims computer scientist Eugene Nickolov, who runs the Sofia lab.

If that's true, it didn't take long for some computer jocks to lose their innocence. By the late 1980s, Bulgarians “certainly displayed more aggressive virus examples than many of their contemporaries” in viral hot zones elsewhere, Cohen says. Notorious Bulgarian viruses included Yankee Doodle and Dark Avenger—perhaps the first “fast infector” virus, it ripped through the industrialized world. Viruses like Dark Avenger infect files not only when the user executes them, but also any time the computer accesses them for other reasons. “This was a novel idea that permitted the virus to spread very fast,” says Vesselin Bontchev, the founding director of Sofia's virology lab and author of the first analysis of the Bulgarian virus factory. Also around that time, the first highly publicized “virus exchange”—a venue in which hackers swap viruses—began operating in Sofia, a development that put the Bulgarian threat under a harsher international spotlight.

…The academy founded the virology lab in late 1990 to combat the country's mushrooming virus problem. To lead the facility, the academy's president tapped Bontchev, then a young hot shot at the Institute of Industrial Cybernetics and Robotics. He had made a mark for antivirus research done in his spare time, including devising a way to neutralize “Vienna 648”—an early virus that damaged *.com files—by writing a program to locate the infected files and purge the virus.

“I became director of the laboratory before I even had an office, equipment, or personnel,” recalls Bontchev, who soon left the lab for Hamburg to write his dissertation under Brunnstein. Bontchev is now an antivirus expert at FRISK Software International in Iceland. “One of my strongest motivations then and now,” he says, “is to counter the public image of Bulgaria as a place that makes viruses.”

View this table:

## The antiviral diaspora

Virus creators and antivirus researchers are constantly at war, with both sides in an arms race to launch increasingly sophisticated attacks and counterattacks. Although some viruses evade detection long enough to inflict global economic losses, Nickolov argues that he and his colleagues benefit from the thrust and parry. “From this race emerge better operating systems, applications, and technologies” that are more resistant to viral infection, he says, thanks to the development of “bloodhound” technologies capable of lightning-fast virus detection.

Devising such technologies is a lucrative pursuit, although you wouldn't know it from touring the Sofia lab. On a summer day the facility, tucked in a corner of the academy's Institute of Mathematics and Informatics, is hot and cramped, a far cry from sleek, air-conditioned Western centers. All the Sofia lab has to work with are 15 personal computers, upgraded occasionally thanks to European Union grants. The government gives Nickolov, who took over from Bontchev shortly after the lab was up and running, a budget of less than \$50,000 a year. Most of the lab's 10 computer scientists must take second jobs, says Nickolov, who himself moonlights as a lecturer at three universities.

Unable to pay competitive wages, Nickolov says it's “virtually impossible” to hold onto young experts. Although the lab has produced some world-class antivirus researchers, most of Bulgaria's domestic computer industry disintegrated after Communism fell, and the Bulgarian software companies that remain can't match the salaries paid by their Western counterparts. Bulgaria's loss is the world's gain, as the lab's far-flung progeny have established themselves as leaders in antivirus efforts. Bulgarian luminaries include Bontchev and Katrin Tocheva, a senior researcher at F-Secure Corp. in Finland. Tocheva, who got hooked on computer programming when she was 15, joined the virology lab in 1991 and stayed for 6 years until taking a higher paying job at F-Secure.

One of Tocheva's crowning moments, she says, was fighting the ILOVEYOU virus, which inflicted billions of dollars of damage on computer and information systems last spring. “When a copy of ILOVEYOU arrived from one of our distributors, I realized that the case was really big,” she says. “I informed all my colleagues, called Vesselin [Bontchev], and dragged him out of bed.”

After taking a few minutes to figure out how the virus functions—it's a “worm” that infects computer networks via Microsoft Outlook and uses a separate code to steal passwords—Tocheva set to work analyzing the script that forms the main body of the virus, while a colleague tackled the password-grabbing code. “When there is a new virus, the first and main part of the job is to analyze it. The next step is to give users a solution,” she says. F-Secure was among the first security firms to alert the computer world about the virus, and it and other companies soon ginned up antidotes.

Bontchev, who has disarmed hundreds of viruses, along with Brunnstein is a founding member of the world's most exclusive club of antivirus researchers, the Computer Anti-Virus Researchers' Organization. The group's 25 members, most of whom work for competing computer security companies, ship viruses to each other for analysis. “We often exchange information,” Bontchev says. One major concern among today's virus fighters is that as more communications functions are merged—such as WAP cell phones that can access the Internet, or cable television systems that double as Internet service providers—clever virus programs will propagate faster and become more destructive. This phenomenon, called “convergence,” is “a major potential threat,” notes Stromberg.

Nickolov looks forward to the possibility that his lab, freshly infused with grant money from the Swedish government, will be able to help defang any convergence threat. But he also casts a wary eye toward the past. In a hallway near the virology lab sits a display case containing the dusty metal skeletons of antiquated Bulgarian and Russian computers. “They are relics now,” says Nickolov, running his fingers along a 1960s Bulgarian keyboard that resembles an antique cash register. Unless Nickolov can somehow lure new equipment and a decent budget, his lab could become just as obsolete.

13. VIROLOGY

# Evolution on Life's Fringes

1. Michael Balter

Fresh evidence that viruses have existed for billions of years has scientists wondering what role these stripped-down microbes played in evolution

TOURTOUR, FRANCE—In the beginning, there was the land, the sky, and a sterile sea. After hundreds of millions of years in that primordial soup, molecules with carbon backbones began replicating and eventually begat cells. The cells were fruitful and multiplied, giving rise to life's three domains: Bacteria, Archaea, and Eukarya.

This creation story, or variations on the theme, is an article of faith to most scientists. But there is one embarrassing omission. Where did viruses come from? Most scientists agree that viruses are life-forms. They are not cells, but they have their own DNA or RNA genomes and can reproduce with the unwilling help of a cellular host. Yet although viruses are able to hijack organisms of all sorts, they have long been consigned to a taxonomic ghetto that had little to do with the origins of the three domains.

Earlier this summer, two dozen scientists gathered here in southern France's Var region, at the forested estate of the Les Treilles Foundation—a family-run organization that promotes scientific exchange—to take another crack at the question of viral origins and evolution.* They had some powerful new findings to charge the discussion. Despite the difficulties of reconstructing events that may have taken place more than 3 billion years ago, recent work using the structure of viral genes and proteins to infer relationships between organisms has sparked some provocative ideas. Among them is the notion that viruses arose very early—perhaps before the three domains diverged—and the hypothesis that viruses, rather than being an aberrant branch on the tree of life, have played a major role in the evolution of multicellular organisms. “I hope we will dismiss the notion that they are somehow alien, like the Andromeda Strain,” virologist Stephen Morse of Columbia University in New York City told the meeting attendees.

Even before the first electron microscopes in the 1930s opened a window onto the shadowy world of viruses—which scientists in the late 19th century had recognized only as infectious particles that could pass through filters fine enough to ensnare bacteria—scientists were speculating on how these mysterious microbes originated. In 1924, French-Canadian microbiologist Félix d'Herelle proposed that viruses were the ancestors of cells. Echoing that idea, Nobel Prize-winning biologist Salvador Luria suggested in the 1960s that modern viruses are relics of the precellular primordial soup. Others have argued that viruses came into being more recently as the vestiges of bacteria that lost their cell walls and degenerated from a free-living to a parasitic existence. Currently, the most widely held view, inspired by research in the 1970s by Nobel laureate Howard Temin and others, is that viruses trace their heritage to cellular genes that somehow broke free and spun a protein sheath that enabled them to survive briefly outside a cell's protective environs.

It is only in the last few years that researchers have been able to put these ideas to the test. The enormous genetic diversity of viruses—whether a result of being around for billions of years, from mutating rapidly, or both—makes it difficult to trace evolutionary lineages by comparing gene sequences from one species or strain to another. Researchers find many nearly identical genes in organisms as diverse as fruit flies, mice, and humans. But bacteriophages—viruses that infect bacteria—seldom bear more than a trifling resemblance when their genomes are lined up side by side, even when outwardly they appear very similar.

So instead of simply lining up genomes, virologists have been hunting for other measures of evolutionary relatedness—and they have come up with some surprising pairings. Two unlikely cousins are the herpes virus HSV-1, which infects people, and the bacteriophage T4. The two viruses are remarkably similar in the shapes of their outer coats, the complicated steps they take to assemble their coat proteins, and how their DNA molecules are packaged inside—even though they are very dissimilar genetically. At the meeting, molecular biologist Dennis Bamford of the University of Helsinki described an even more compelling parallel between the coat protein structure of the bacteriophage PRD1—which was discovered nearly 30 years ago in the sewers of Kalamazoo, Michigan—and that of human adenovirus, blamed for many chronic respiratory and intestinal infections.

Bamford's group, in collaboration with structural biologist Roger Burnett's team at the Wistar Institute in Philadelphia, worked out the x-ray crystal structure of P3, the main protein of PRD1's outer coat. Like many viral coat proteins, P3 features a compact “jelly roll,” made up of eight parallel polypeptide strands (see diagram). A comparison of P3 with the coat protein of human adenovirus type 2, called hexon, revealed that the configurations of each virus's jelly rolls are remarkably similar. Moreover, both PRD1 and adenovirus arrange their protein subunits in the same way during viral assembly, in a fashion atypical of other viruses. “The parallels are striking,” says Stanford University geneticist Allan Campbell, a pioneer in bacteriophage research, because P3 and hexon lack any detectable similarity in their amino acid sequences. Campbell and others agree that the structures must have been conserved even though the amino acid combinations changed radically during the course of evolution.

Indeed, meeting participants interpreted these results—as well as other fresh data showing that genetic similarities can sometimes be found, for example, between viruses that infect bacteria and those that infect archaea—as evidence that viruses arose before life's three domains diverged. “It is difficult to imagine this could be convergent evolution,” says Campbell. Adds molecular biologist Roger Hendrix of the University of Pittsburgh, “The similarities are great enough that it is unlikely they rose independently.” But because the molecular clock used to unravel genetic relationships of other organisms doesn't work well for viruses, researchers admit that it's impossible for now to trace the viral family tree back to its roots. The best guess, says Hendrix, is that “their divergence must have been really far back in time.” If so, that would be gratifying to some researchers. “There must have been something that existed before cellular life, assembling subunits but having no membrane,” says molecular biologist Henry Krisch of the Laboratory of Microbiology and Molecular Genetics in Toulouse, France. “What else would you call it other than a virus?”

## A “moronic” hypothesis

Another theory of viral origins, the “moron accretion hypothesis,” was put forward at the meeting by Hendrix. He and Pittsburgh colleagues have been sifting the genomes of bacteriophages for clues to how they evolved. Earlier this year, they reported the complete DNA sequences of two bacteriophages that infect Escherichia coli, HK97 and HK022. When the sequences were compared to each other as well as to those of two other outwardly similar bacteriophages, it became clear that all four have some DNA stretches in common. But these stretches are interspersed with longer sequences that vary greatly from one bacteriophage to the next. Such genetic mosaicism has often been taken as evidence that viral strains swap genes.

But the Pittsburgh group also found extra genes, apparently capable of acting independently of the rest of the viral genome—because they carry their own instructions for the initiation and termination of gene expression—inserted into the genomes of some bacteriophages. The function of these extra genes is unclear, but they do not appear to act in concert with adjacent genes. The researchers whimsically dubbed these genes “morons,” says Hendrix, because they represent “more DNA.” Some morons are similar to known bacterial genes.

Extrapolating from these observations, Hendrix laid out an intriguing, albeit assumption-laden, scenario. If the DNA of early primitive cells was not yet organized into chromosomes but floating free, and if a gene mutation led to a protein that self-assembled into the kind of icosahedrally shaped shell typical of many virus coats (a penchant of a number of proteins under certain conditions, such as the bacterial enzyme lumazine synthase), the result could have been a DNA segment trapped in protein—a “primitive virus,” as Hendrix says. Over time, the viral genome could have grown moron by moron, particularly if the addition of a new gene made it more likely to survive in its environment. “This would allow you to build a virus from the ground up,” Hendrix says. “The morons are simple, but if you put them together over time you might get something that is pretty smart genetically.”

Hendrix's proposal generated debate at the meeting. “I'm not convinced that viral genomes have been built up by starting with one gene and then adding genes until we get where we are now,” says Campbell. And while agreeing that the wholesale defection of genes or groups of genes from early cells may well have given rise to viruses, Krisch argues that the moron accretion idea would require the viral coat to get bigger and bigger as new genes are added, to accommodate the growing genome. “That part of the theory leaves me a little ill at ease,” he says.

## The whale and the virus

Although the origins of viruses remain obscure, their potential role as forces shaping the early evolution of life is itself taking life. Underscoring the fact that viruses are much more than a sideshow in the biosphere, marine microbiologist Curtis Suttle of the University of British Columbia in Vancouver offered data on the influence of viruses on the carbon cycle in the oceans. A typical milliliter of seawater contains on average some 10 million virus particles; extrapolating from that figure, seagoing viruses lock up as much as 270 million metric tons of carbon, more than 20 times the top estimate for the amount stored in all the whales on Earth. “With all those numbers of [viruses], it is hard to imagine that they were not a significant force in evolution,” says Krisch.

In an essay in the June 1998 issue of the Proceedings of the National Academy of Sciences, microbiologist Carl Woese—who is credited with discovering the Archaea—suggested that early life was a hotbed of “lateral gene transfer” between cells, and that this gene swapping was the key driver of evolution. Only after the domains of life diverged, he argued, did the barriers to such promiscuity grow formidable. Woese didn't touch upon the question of how viruses originated. But building on this idea, molecular biologist Patrick Forterre of the University of Paris's Orsay campus suggested at the meeting that viruses might have been important early vehicles for such gene exchange. If so, Forterre argues, it might explain puzzling relationships between the proteins involved in DNA replication in each domain. Many such proteins in the bacteriophage T4, for example, are more similar to their eukaryotic counterparts than to those in bacteria. Indeed, many viral proteins are more closely related to viral or cellular proteins from organisms in domains other than that of their own host.

If viruses did play an active gene-shuffling role in life's early days, then their reputation as being somehow alien to other life-forms would be way off the mark. “These viruses have been having a global orgy of gene swapping for 3.5 billion years,” Hendrix says. Indeed, Krisch submits, “if there hadn't been all these things around exchanging genetic information with their hosts, we wouldn't be what we are today.”

• *The Origin and Evolution of Viruses, 25 to 29 July.

14. CLIMATE

# Ice, Mud Point to CO2 Role in Glacial Cycle

1. Richard A. Kerr

Antarctic ice and deep-sea records suggest that orbital variations work through carbon dioxide, not ice sheets, to drive the ice ages

Every 100,000 years or so for the past million years, kilometers-high sheets of ice have ground southward from the Arctic. The vast glaciers scour landscapes and drive all life before them. Yet if the effects of the ice ages are far from subtle, their exact causes have left researchers scratching their heads for more than a century. Somehow, the rhythmic stretching of Earth's orbit seems to drive glacial cycles, but how this feeble “orbital variation” could cascade through the climate system of air, land, water, and ice to produce such monstrous climate shifts has remained a mystery. On page 1897 of this issue of Science, paleoceanographer Nicholas Shackleton of the University of Cambridge finds a likely strongman to transmit and enforce the orbital variations' demands: carbon dioxide. Comparing records preserved in deep-sea muds with those in antarctic ice, he finds that orbital variations may muster carbon dioxide into and out of the atmosphere, and the resulting waxing and waning of greenhouse warming may drive the glacial cycle.

“It's quite an exciting development,” says paleoceanographer John Imbrie, professor emeritus at Brown University. “He's made a major step forward.” Imbrie, Shackleton, and James Hays of Lamont-Doherty Earth Observatory in Palisades, New York, co-authored the 1977 paper in Science that made orbital variations the leading candidate for the ultimate driver of the ice ages. Imbrie went on to suggest that the ice sheets themselves amplify the weak orbital signal, but Shackleton concludes that ice is farther down the chain of command; carbon dioxide is the executive officer who hands Earth its climatic marching orders.

Shackleton reached his conclusion by precisely comparing geologic records in two different media, a method that enabled him to tease out intimately entangled climatic influences with unprecedented accuracy. First he consulted the paleoceanographers' mainstay, bottom muds containing fossil skeletons of microscopic marine animals called foraminifera. The mix of heavy and light isotopes of oxygen in the forams' carbonate skeletons depends partly on the mix of isotopes in the seawater the forams lived in and partly on the water's temperature. The composition of the seawater, in turn, depends on how much ice is trapped on land at any time—a factor long considered the dominant influence on such isotopic records. The oscillating oxygen isotopes of forams seemed to show that ice volume varied in step with the 100,000-year variation in the shape or eccentricity of Earth's orbit, although the correlation was not impressive.

To get a cleaner identification of changing ice volume, Shackleton turned to bubbles in the 400,000-year-long ice core retrieved from the antarctic ice sheet by Soviet drillers at Vostok station and analyzed by a team of French glaciologists headed by J. R. Petit of the Laboratory of Glaciology and Geophysics in St. Martin d'Heres. Because bubbles in glacial ice come from the atmosphere, their oxygen-isotope composition—unlike that of foram shells—does not respond to ocean temperatures. But the composition is sensitive to the volume of ice in the world, because isotopic changes in seawater oxygen are passed on to atmospheric oxygen through marine photosynthesis. Thanks to that difference in response, Shackleton could separate ocean-temperature changes from ice-volume changes.

To Shackleton's surprise, deep-sea temperature accounted for more variation of oxygen isotopes than ice volume did. And orbital eccentricity, deep-sea temperature, and atmospheric carbon dioxide as recorded in Vostok gas bubbles all varied in step, on the same 100,000-year cycle. Ice volume, however, lagged behind.

Shackleton sees the lockstep of eccentricity, greenhouse gas, and temperature as a sign of cause and effect. In his view, at the beginning of glaciation changes in eccentricity—presumably by shifting the distribution of sunlight across the globe—could have decreased atmospheric carbon dioxide, weakening the greenhouse and cooling the ocean and atmosphere. At the end of an ice age, the changes are in the opposite sense. Those changes are relatively rapid and would appear to coincide; the sluggish ice volume would lag behind. That delay rules out ice as a prime mover, Shackleton says; it's only a follower.

Shackleton's results impress many researchers who specialize in sorting out the cause of the ice ages. The paper “is really well argued,” says geophysicist Richard Peltier of the University of Toronto. “It has inevitable drawbacks because of the short record,” but it would appear that “carbon dioxide is a primary driver, not just a weak feedback.”

Peltier says his own computer models reinforce that conclusion. Hoping to show that ice sheets themselves were crucial to glacial cycles, he developed a model that included not only orbital forcing and carbon dioxide variations but also the way ice sheets grow and decay. The ice turned out to be essential for the distinctively abrupt end of an ice age, but not crucial the way carbon dioxide seems to be. “The model only works because it includes the forcing from carbon dioxide,” Peltier says. “If we exclude that, we get no glacial cycle. The ice dynamics just doesn't do it for you at all without carbon dioxide.”

Even if Shackleton's elimination of ice as a primary factor and support for carbon dioxide stand up, “there's a lot of questions that remain,” says geochemist Daniel Schrag of Harvard University. In particular, he says, no one knows how orbital variations would send the carbon dioxide into and out of the atmosphere.

And there are likely to be other significant factors besides carbon dioxide, notes climate modeler André Berger of the Catholic University of Louvain in Belgium. “I have quit looking for one cause of the glacial-interglacial cycle,” he says. “When you look into the climate system response, you see a lot of back-and-forth interactions; you can get lost.” Schrag agrees. “It's a complicated problem,” he says. “We still don't know the cause of the ice ages.”

15. BIOCHEMISTRY

# High-Tech Lures Hook Into New Marine Microbes

1. Elizabeth Pennisi

Two groups of researchers have identified large populations of bacteria that convert sunlight hitting the sea surface into energy

Two scientific teams have found an entirely unexpected phenomenon: new kinds of ocean bacteria that convert light to energy. What's more, combined, these two previously undetected microbial groups may account for a whopping 20% of bacterial life on the ocean surface, says Edward DeLong, a marine microbiologist at the Monterey Bay Aquarium Research Institute in Moss Landing, California, and head of one of the teams who report their findings this week in Science (p. 1902) and Nature. The other team was led by biological oceanographer Paul Falkowski of Rutgers University in New Brunswick, New Jersey. “These two papers are going to have a major impact on the way oceanographers think about the ecology of the ocean surface,” comments Stephen Giovannoni, a microbiologist at Oregon State University in Corvallis.

These two groups of bacteria, which harness light to move electrons and power cellular processes, likely help bring energy into the food chain. And that could help to explain a puzzle that has confounded researchers trying to understand marine ecosystems: how so many bacteria can survive in the open ocean, where there seems to be relatively little for them to eat.

The bacteria DeLong discovered have light-harnessing abilities previously known to exist in a fungus and in nonbacterial microbes called archaea that hang out in the most hostile salty environments, such as salt ponds, where sustenance is scarce. They are equipped with a protein, known as bacteriorhodopsin, that enables them to thrive by harnessing light to generate ATP—the source of energy needed to power cellular processes. The bacteria Falkowski found use a type of bacterial chlorophyll that, until now, no one had found in bacteria in the open ocean.

DeLong and Falkowski both discovered their bacteria somewhat serendipitously, using novel and distinct techniques for tracking down new species. Traditionally, biologists have assessed microbial diversity by growing environmental samples in the laboratory. But researchers have come to realize that the technique misses many organisms that don't survive long enough to be counted, so they are trying new approaches. “DeLong went fishing using a molecular biology tool; we went fishing with a biophysical technique,” notes Falkowski.

DeLong's group collects DNA directly from seawater samples, without first isolating the individual organisms. In one 130-kilobase DNA fragment they fished out this way, the team found a gene that encoded the small ribosomal subunit RNA belonging to a group of marine bacteria called SAR86, discovered by Giovannoni about a decade ago. No one had ever been able to grow the bacteria in the lab, so it was known only by this one ribosomal gene.

But the DeLong team had that gene and more in their fragment. When they compared the rest of the sequence—which included some 120 genes—to other genes on file in the public archives, most looked unexceptional for this type of bacterium. But one stood out: a gene for bacteriorhodopsin, a rhodopsin protein that, contrary to what the name implies, had never before been found in bacteria. “It's a mechanism of photosynthesis that's totally unknown in bacteria,” says Norman Pace, a microbial biologist at the University of Colorado, Boulder.

DeLong's team dubbed this new protein proteorhodopsin and tested its function by putting it into Escherichia coli. As they hoped, the modified bacteria not only made the protein, but they reacted to light by moving protons out of the cell and into the surrounding medium. That's characteristic of bacteriorhodopsin, whose proton-pumping activity helps set up a gradient whereby energy is generated by protons flowing back into the cell. “It's a really strong finding,” says Kenneth Nealson, a microbiologist at the California Institute of Technology in Pasadena. “And I don't think you could have ever discovered this without genomics.”

Falkowski's team, in contrast, was not after genes. They were searching for new types of photosynthesizing organisms that might use dim infrared light emitted from deep-sea vents. For almost a decade, researchers had postulated that such microbes might exist in those extreme environments, “but there was no solid proof,” says Falkowski. So he and his Rutgers colleague Zbigniew Kolber built a fluorometer designed to detect any changes in the electromagnetic radiation that would be caused by microbes using light energy to move electrons.

They found no organisms by the deep-sea vents. But, to their surprise, when they scanned surface waters they picked up many positive signals in samples taken along a 1000-km line in the eastern tropical Pacific and later off the coast of New Jersey. Because only bacterial—as opposed to plant—chlorophyll makes use of light at the wavelengths detected by the fluorometer, the team is certain that they have found a new home for phototrophic bacteria—the sea surface. Based on the strength of the signals, they estimate that these aerobic phototrophic bacteria represent about 1% of the surface phytoplankton. These bacteria and DeLong's bacteria use different mechanisms to harness light, but they both use that light to generate ATP. This gives them a leg up in environments where food is sparse, says DeLong, and they in turn become nourishment for organisms higher up the food chain. Neither Falkowski nor DeLong knows whether their light-loving organisms also use light to fix carbon the way plants do; if they do, they would enrich the ocean surface further.

Few microbiologists have imagined that aerobic phototrophic bacteria could live on the ocean's surface, because they usually thrive where the water's oxygen content is low. And even fewer suspected that some bacteria might have a bacteriorhodopsin. Says Giovannoni: “Most people are going to do a double take when they encounter these papers.”