# News this Week

Science  15 Jul 2005:
Vol. 309, Issue 5733, pp. 362
1. ENVIRONMENTAL HEALTH

# Lead Paint Experts Face a Barrage of Subpoenas

1. Jocelyn Kaiser

A high-stakes legal battle involving the health hazards of lead paint has turned ugly. A paint company sued by the state of Rhode Island has demanded raw and even unpublished data sets from several researchers so that the court can review scientific claims. Although a lawyer for the company calls the probe routine, one target says it feels like harassment.

Many states and legal experts are watching this trial because a victory by Rhode Island could spur lawsuits by other cities and states seeking compensation for cleaning up lead paint. In the meantime, some scientists whose research is involved say the dueling over data has become unreasonable. “I didn't realize what I was getting into,” says epidemiologist Kim Dietrich of the University of Cincinnati in Ohio, who has been asked to turn over 25 years' worth of data even though he is no longer an expert witness.

The company, Sherwin-Williams, is one of seven firms named in the lawsuit that once made or sold lead-based paint. Rhode Island claims that the companies knew of its risks and should pay to clean up a “public nuisance” they helped create. At issue is a long-established link between lead exposure and a drop in IQ, as well as recent studies suggesting that very low blood levels of lead pose a risk and that lead exposure can increase criminal behavior. Before a 2002 trial, Superior Court Judge Michael Silverstein denied the companies' request for raw data from the state's three expert witnesses. The trial ended with a hung jury.

But Silverstein changed his mind in advance of a new trial. Lawyers for Sherwin-Williams asked for data sets from cohort studies in four cities. They demanded even unpublished data, arguing that peer-reviewed studies might have omitted information, such as a father's IQ, that could invalidate the apparent links between lead and adverse health effects. In allowing their request last September, Silverstein cited a 1993 Supreme Court case, Daubert v. Merrell Dow, which called for pretrial hearings to evaluate scientific experts' opinions. “One cannot … blindfold the defendants and preclude the raw data,” he declared.

Two expert witnesses for the state have since turned over certain data sets from their published studies. They are psychiatrist Herbert Needleman of the University of Pittsburgh in Pennsylvania, who has battled data demands before, and environmental health researcher Bruce Lanphear of Cincinnati Children's Hospital Medical Center. In February and May, two more researchers—neurologist David Bellinger of Children's Hospital in Boston and Dietrich—received subpoenas.

For Bellinger, the request amounted to providing attorneys with a Boston data set he had given to Lanphear for a pooled analysis on IQ and blood lead level appearing this month in Environmental Health Perspectives. He initially vowed to resist but relented last month after the judge ruled this could disqualify Lanphear as a witness.

Dietrich is still in limbo. He was an expert witness in the first trial and initially in the second trial, until Sherwin-Williams requested raw data from his study, started in 1979, of some 300 Cincinnati children followed from birth. He asked to be reimbursed for the cost of reconstructing old data tapes and providing other materials—and stripping out identifying information—at about $125,000. He also refused to provide sensitive data on behavior, such as self-reports of criminal arrests. After those conditions were rejected, Dietrich stepped down as a witness in mid-2004. “I thought I was out of the picture,” he says. Then early on 6 May, “a gentleman barged into my office” and demanded materials, Dietrich says. He refused, then was served a subpoena later that day. “As far as I'm concerned, they have no right to the data,” he says. Dietrich has not heard from Rhode Island's lawyers since he withdrew from the case. University of Cincinnati lawyers have offered legal help, but it is unclear whether they will try to quash the subpoena, Dietrich says. Spokesperson Richard Puff said the university cannot comment. Whereas Dietrich sees it as an “extraordinary” case of harassment, Laura Ellsworth, a lawyer for Jones Day, which represents Sherwin-Williams, says “litigants are entitled to data on which the experts are relying,” even if the owners of the data aren't expert witnesses in the case. Such requests are rare but have sometimes been granted, for example in tobacco lawsuits, notes Joe Cecil of the Federal Judicial Center in Washington, D.C. One case last year involved documents from peer reviewers of a book on industrial pollution. The Rhode Island trial is set to begin on 7 September. For environmental health researchers, the case is a reminder of the risks that come with their work. “Regardless of what happens to me,” says Dietrich, “it's something scientists should be aware of.” 2. CLIMATE CHANGE # G8 Leaders Make a Promise to Do More 1. Eliot Marshall* 1. With reporting by Eli Kintisch. The U.K. government, champion of a global campaign to control greenhouse gases, let it be known in advance that the G8 meeting it hosted in Scotland wasn't likely to produce any miracles. It didn't. The G8 plan for mitigating global warming that came out on 8 July was heavy with proposals but light on commitments.* The heads of the eight leading industrial nations promised to boost energy-efficient technology; adopt low CO2-emitting energy sources (including possibly even nuclear power and hydrogen fuels); and back research collaborations, such as a huge monitoring network called the Global Earth Observation System of Systems (Science, 25 February, p. 1182). They pledged to ask the International Energy Agency to work up efficiency standards and the World Bank to boost technology investment. But they endorsed no new targets for reducing greenhouse gases. The plan's vagueness angered green groups that want action. “This is a very disappointing finale,” said Tony Juniper, leader of Friends of the Earth International in London. “The text conveys no sense of the scale or urgency of the challenge.” Others, such as the Natural Resources Defense Council in Washington, D.C., found a silver lining in the Bush Administration's failure to “block world action” on limits, “despite [its] intense lobbying.” U.K. Prime Minister Tony Blair acknowledged after the meeting that “we were never going to be able … to resolve the disagreement” over the Kyoto Protocol, the 1997 treaty that commits the 151 participants (the United States refused to sign on) to meet difficult goals for greenhouse-gas reduction by 2012. However, the G8 members should be “proud” of their solidarity, Blair insisted, because all attendees endorse the view “that climate change is a problem, that human activity is contributing to it, and that we have to tackle it.” In November, Blair is planning to hold “a new dialogue” on climate change in Britain before a meeting of the Kyoto partners later that month in Montreal. 3. ASTRONOMY # Improved X-ray Telescope Takes Flight 1. Dennis Normile TOKYO—Scientists hope that, as can happen with love, the Japanese-U.S. Astro-E2 x-ray satellite is better the second time around. Successfully launched on 10 July, Astro-E2 carries the same six instruments as the original mission that failed on launch 5 years ago. But their performance has been improved. And recent findings from NASA's Chandra X-ray Observatory and the European XMM-Newton Observatory, both launched in 1999, have enabled scientists to hone their list of observational targets. “Although it was a big catastrophe 5 years ago, the mission now is more timely because we know much more about the x-ray sky and have greater [observational] capabilities,” says Nicholas White, chief of the Laboratory for High Energy Astrophysics at NASA's Goddard Space Flight Center in Greenbelt, Maryland. The mission is a joint effort of NASA and the Japan Aerospace Exploration Agency (JAXA). Astro-E2's crown jewel is the X-ray Spectrometer (XRS), which “allows investigations that have never been possible before,” says Andrew Fabian, an astrophysicist at the University of Cambridge, U.K., and science adviser to the mission. Developed by JAXA's Institute for Space and Astronautical Science (ISAS) and NASA, XRS measures the energy of individual x-ray photons. The original XRS had higher spectral resolution, or a greater ability to distinguish energy levels, than any other x-ray instrument in space. By eliminating some sources of noise, scientists doubled that resolution. And improved cryogenics add half a year or more to the 2-year life of the instrument. These improved capabilities will be put to good use. Richard Kelley, NASA's principal investigator for XRS, explains that Chandra has seen evidence of blobs of material in the accretion disks surrounding certain black holes. Astro-E2 should be able to determine if these blobs are real and then to use them “as probes to tell whether the general picture of matter spiraling [into a black hole] makes sense,” he says. XRS should also be able to confirm previous glimpses of the telltale distortion of x-ray emissions expected when iron elements encounter a black hole's intense gravitational pull. XRS will also be looking at clusters of galaxies for clues to the role of dark matter in their evolution and dynamics. Other instruments are expected to keep working for 5 to 7 years, says Hideyo Kunieda, an astrophysicist at Nagoya University who is the principal mission scientist for ISAS. These include the hard x-ray detector, developed by ISAS and a group at the University of Tokyo, that looks at very high energy x-rays emanating from the most violent astrophysical phenomena. There is also a set of four x-ray charge-coupled device cameras covering a wide range of energies. All instruments will be observing the same objects simultaneously for a broad picture of emissions over the x-ray spectrum. The satellite was given a new name, “Suzaku,” after the launch from the Uchinoura Space Center in southwestern Japan. But mission scientists caution that it will take a few weeks to make sure everything is operating properly. “We can't be completely at ease yet,” says ISAS astrophysicist Noriko Yamasaki. If all goes well, scientific observations will start early next month. 4. GEOPHYSICS # Threshold Crossed on the Way to a Geodynamo in a Computer 1. Richard A. Kerr How do you tell whether you've got the right answer if you never get to look at the answer sheet? Geophysicists trying to conjure up the geodynamo that generates Earth's magnetic field have been grappling with that conundrum for years. Seemingly any computer model with the basic physics of the churning molten-iron core produces a reasonable-looking magnetic field as seen at Earth's surface. Yet the models produce those similar surface fields by creating entirely different geodynamos in their cores (Science, 10 January 1997, p. 160). Without access to the real core, how do you tell which one is right? On page 459, Japanese modelers report that one of the world's fastest computers has pushed their geodynamo model into a new regime, in which for the first time a crucial aspect of core behavior matches theoretical expectations. “It's the closest yet to describing what happens with the real Earth,” says planetary physicist David Stevenson of the California Institute of Technology in Pasadena. “This might actually be a realistic model.” If so, it might reveal whether Earth's magnetic field is beginning one of its once-in-100,000-years flip-flops, as geophysicists have recently wondered. Much as weather forecasting models use equations of motion to calculate the flow of the atmosphere, geodynamo models calculate the flow of the core's highly conductive molten iron. The models must also include calculations of the electric currents and magnetic forces generated when the conducting iron flows across magnetic fields. What gave the Japanese group an edge over previous modelers was the Earth Simulator, once the world's most powerful computer and now the fourth most powerful (Science, 1 March 2002, p. 1631). Using just a tenth of the Earth Simulator's 5120 processors during 6500 hours of computation, Futoshi Takahashi of the Japan Aerospace Exploration Agency in Kanagawa and Masaki Matsushima and Yoshimori Honkura of the Tokyo Institute of Technology achieved an unprecedented level of realism in their simulations. In particular, they were able to use the lowest-ever value of a key parameter, the Ekman number, which represents the relative importance of the fluid iron's viscosity to the rotation rate of the planet. In theory, a small enough Ekman number—approaching the real value in the core—would allow viscous drag forces to approach zero, but previous simulations couldn't calculate the flow of molten iron on a small enough scale to let that happen. The Earth Simulator run lowered the Ekman number by an order of magnitude. Cranking up all that computer power produced “a lot of Earth-like qualities in their maps” of changing magnetic fields, says modeler Andrew Jackson of the University of Leeds, U.K. But what impresses geodynamo researchers most is the way the core in the new simulations behaves. In previous simulations, the viscous drag of the fluid iron was on a par with the forces induced by flowing currents. Here, however, viscous effects have become negligible, just as theory says they should be. Although not all the input parameters are realistic yet, the Japanese modelers “seem to have reached a regime like Earth's,” says Stevenson. “It's a remarkable achievement.” A truly realistic simulation of geodynamo behavior could allow researchers to forecast the “weather” in Earth's core, with practical implications. Geophysicists have been worrying out loud lately that a 10% weakening of the magnetic field during the past 170 years may presage an overdue flipping of Earth's magnetic poles, a so-called magnetic reversal. It's been 780,000 years since the last reversal, although millennia-long reversals have occurred on average every 100,000 years or so. Aside from millennia of confused magnetic navigation, a reversal would greatly weaken Earth's magnetic shield that fends off cosmic radiation. In the Earth Simulator model runs, at least, reversals begin with distinctive pairs of magnetic flux patches, areas at the surface where magnetic fields of opposing polarity protrude from the interior. These patches first appear in the model at low to middle latitudes as a reversal begins and migrate poleward as the reversed field emerges from the outer parts of the core. Curiously, parts of the early stages of a simulated reversal “look a lot like the South Atlantic today,” says paleomagnetist Bradford Clement of Florida International University in Miami. Intriguing reversal simulations aside, “we've still got a ways to go,” says Gary Glatzmaier of the University of California, Santa Cruz, one of the first to run simulations of the dynamo 10 years ago. “They're getting closer [to realistic conditions], but you're still not close enough to be confident you've got something Earth-like.” Among other shortcomings, the model still doesn't have enough computing power to banish glitches such as too-frequent reversals. “It's going to take 5 to 10 years for computers to be fast enough,” says Glatzmaier. 5. AVIAN INFLUENZA # Chinese Ministry Questions Bird Flu Findings 1. Dennis Normile* 1. With reporting by Martin Enserink. The Chinese Ministry of Agriculture sharply criticized a paper published online by Nature last week suggesting that an outbreak of H5N1 avian influenza among wild birds in northwest China originated from apparently unreported outbreaks among poultry in southern China. The researchers defend their results and worry that the ministry's reaction may lead to overly strict enforcement of new guidelines on handling the H5N1 virus. The brief report by Honglin Chen of Shantou University Medical College and colleagues there and at the University of Hong Kong and other institutions, posted on 6 July, claims that the virus recovered from wild birds at Qinghai Lake is a genetic match to one found among poultry earlier this year at a live market in China's Guangdong Province. The following day, China's official Xinhua News Agency quoted Jia Youling, director general of the Agriculture Ministry's Veterinary Bureau, questioning the article's credibility because H5N1 has not been reported in Guangdong this year. He added that the Shantou lab doesn't have the proper biosafety features for handling the virus. Chen says the report's conclusions will be supported by additional data on H5N1 circulating in southern China in an upcoming publication. He adds that the research conducted in the Shantou lab, a biosafety level 2-plus lab, conforms “entirely with … World Health Organization guidelines.” And Chen denies news reports that the ministry is attempting to close their operations. “We are now establishing communications with the ministry to gain an understanding about our research,” he says. The Agriculture Ministry's guidelines for handling highly pathogenic viruses, adopted in late May, simply spell out in more detail previously established procedures, says George Gao, a virologist at the Chinese Academy of Sciences' Institute of Microbiology and the corresponding author of a paper on the Qinghai outbreak published online by Science on 7 July. (That paper focuses on the pathogenicity of the viral strain without tracing it to any previously reported outbreaks among poultry.) Roy Wadia, a spokesperson for the World Health Organization's Beijing office, says the agency hopes that the guidelines “will be interpreted in a way that will both ensure that research is carried out in a safe manner but also encourage, stimulate, and support research into this virus.” 6. CLIMATE FORECASTING # India Chucks Monsoon Model 1. Pallava Bagla NEW DELHI—The Indian government has decided that its controversial, homegrown computer model to predict the all-important monsoon is all wet. Turned off by the model's poor track record, officials at the Department of Science and Technology (DST) have teamed up with modelers at the U.S. National Center for Atmospheric Research (NCAR) in Boulder, Colorado, to use the center's new software for weather prediction and forecasting. The June-to-September monsoon is vital to Indian farmers and to the country's economy. But the complex atmospheric conditions make it extremely difficult to predict. A model developed in the 1980s by the India Meteorological Department (IMD) incorporated data from 16 factors in hopes of eliminating much of the uncertainty (Science, 23 August 2002, p. 1265). Instead, the model tended to predict normal levels of rainfall and missed extreme events like the massive flooding in 1994 and widespread droughts in 1987, 2002, and last year (see graph). IMD's poor performance goes back even further, say atmospheric scientists Sulochana Gadgil and Ravi Nanjundiah of the Indian Institute of Science in Bangalore and IMD's M. Rajeevan. Writing recently in India's Current Science, they argue that “the forecast skill has not improved over seven decades despite continued changes in the IMD operational models.” Although IMD officials continue to believe that the model has been a useful tool, DST Secretary Valangiman Subramanian Ramamurthy says he concluded late last month that it was time for a change. “If it's failing,” he says, “we can't continue with it.” Unfortunately, he says, India does not have the necessary skills or time to start from scratch. Thanks to the new Weather Research and Forecasting (WRF) model, it doesn't have to. Greg Holland, director of NCAR's mesoscale and microscale meteorology division, calls the model “a very advanced system, usable down to very fine resolutions of less than 1 kilometer.” Using globally available data as a starting point, scientists can plug in regional data to fine-tune the forecast. In addition to short-range weather prediction, the model is also capable of making the type of seasonal forecasts that Indian officials desire for the monsoon. Both sides see important benefits from the collaboration, which was worked out earlier this month when Ramamurthy visited NCAR. NCAR gets to expand the WRF model, already used in China, Taiwan, and South Korea, to the Indian subcontinent. The two countries agree to conduct joint research not only on the monsoon but also on a 2-year demonstration project to predict deadly tropical cyclones in the Bay of Bengal. And India gets the promise of a more accurate monsoon prediction. To make the most of the new software, Ramamurthy says DST is close to signing a$7 million deal to purchase a more powerful supercomputer. “A teraflop machine running WRF would definitely bring India to the leading edge of atmospheric modeling,” says Holland, as well as “substantially improving local weather forecasts.”

7. ENERGY SCIENCE

# Tight Budgets Force Lab Layoffs

1. Eli Kintisch

Shifting priorities and the expectation of worsening budgets have triggered layoffs at two U.S. synchrotron facilities in California. Officials at the Department of Energy's (DOE's) Advanced Light Source (ALS) at Lawrence Berkeley National Laboratory and the Stanford Synchrotron Radiation Laboratory (SSRL) say the cutbacks, which include shutting down one beamline at each facility, are necessary to free up money for new projects, such as brighter beams. The cuts may not be reversed, managers say, even if Congress beefs up the lean 2006 DOE budget proposed by the Bush Administration.

“I had to make some tough choices,” says SSRL director Keith Hodgson, who in April let go eight scientists and technicians and stopped scheduling users on one beamline doing materials research. “I could have stopped advanced beamline development and saved jobs” in response to the president's proposal to cut the lab's 2006 budget by 8%. Instead, Hodgson says, he chose to protect his most promising programs in the face of uncertainty.

The two labs are funded by DOE's $1.1 billion Office of Basic Energy Sciences (BES), which received a$41 million boost in the president's request for 2006. But that 4% increase includes $151 million in new funds to expand nanoscale research facilities and begin work on the Linac Coherent Light Source at Stanford and a high-flux neutron beam under construction in Oak Ridge, Tennessee, the pair of which should benefit physicists, doctors, and chemists. Taking into account$52 million in construction costs that DOE will save in 2006, the result is a proposed $58 million belt tightening within existing BES programs, including 7% at ALS. Pat Dehmer, DOE associate director of science for BES, says funds for new projects have led to similar tradeoffs elsewhere in the DOE budget. “As a result of those high priorities, cuts had to be made virtually across the Office of Science,” she says. Two other DOE-funded synchrotron sources, at the Argonne (Illinois) and Brookhaven (New York) national labs, have so far avoided layoffs. But there are hiring cutbacks at each facility that affect about a dozen positions now vacant. In addition to financial pressures, the cutbacks also reflect shifting science trends on a local scale. This spring, ALS acting Director Janos Kirz shuttered an x-ray spectroscopy line used for diagnosing silicon wafers and removed from the Berkeley lab's payroll 16 scientists and support staff. New analysis methods have been shown to be “simpler or better,” he said. At the same time, Stanford's Hodgson stopped scheduling users for an undersubscribed topography imaging beamline used to characterize stress in metallic thin films. Some users acknowledge the need for change, despite the disruption. “Given that nanotechnology is moving toward more biomaterials, I would support [Hodgson],” says materials scientist John Bilello of the University of Michigan, Ann Arbor, who used the SSRL beamline and now eyes the CERN high-energy physics lab near Geneva, Switzerland, as an alternative. Congress seems eager to help. In separate bills that must be reconciled, the Senate and House have added back$95 million and $27 million, respectively, to the BES budget. “Existing capabilities cannot be sacrificed to purchase new facilities,” Senate appropriators wrote in a committee report that observers see as a vote of confidence in current work. But even with a restored 2006 budget, Hodgson and Kirz say it may be difficult to rehire staff or reopen beamlines. 8. NONLINEAR OPTICS # To Physicists' Surprise, a Light Touch Sets Tiny Objects Aquiver 1. Adrian Cho Much as a child might make a soda bottle shake by blowing across its top and filling it with sound waves, physicists have set a tiny disk of glass vibrating by “whistling” light through it. The effect could lead to optically controlled micromachines but might also limit the sensitivity of giant gravitational-wave detectors. “I'm deeply impressed,” says Dirk Bouwmeester, a physicist at the University of California, Santa Barbara. “The findings add a completely new tool to the fields of optical interferometry and information processing.” The fat-rimmed disk of silica used by physicist Kerry Vahala and colleagues at the California Institute of Technology (Caltech) in Pasadena is an “optical microcavity” that “rings” with light of distinct frequencies, just as a soda bottle whistles at specific pitches. Optical microcavities control lasers in CD and DVD players, and higher-quality cavities that can hold more light might help shuttle photons through “photonic” circuits. This week, the Caltech researchers report online in the journal Optics Express that light coursing through a microcavity can set the thing in motion. In the experiment, light from a nearby optical fiber bled into the disk and raced around its rim. Pressure from the circulating light set the disk vibrating. The vibrations stretched the disk and altered the frequency of the light in telltale ways, the researchers found. Theorists had predicted that light pressure might cause an optical cavity to vibrate, but the rattling caught the researchers off-guard. “We were studying the nonlinear optical properties of these cavities,” Vahala says. “This one really came out of the blue.” The effect could prove useful, says Ming Wu, an electrical engineer at the University of California, Berkeley. For example, researchers are already developing micrometer-sized mechanical oscillators that interact with microwaves. So the light-to-vibrations connection might make it possible to control microwaves with light, Wu says. On the other hand, the vibrations could prove a nuisance for researchers working on the Laser Interferometer Gravitational-Wave Observatory (LIGO). With installations near Livingston, Louisiana, and in Hanford, Washington, LIGO relies on high-power, 4-kilometer-long optical cavities to search for gravitational waves, which would stretch the cavities. Vibrations caused by light pressure might limit LIGO's ultimate sensitivity, Bouwmeester says, and that's no small shakes. 9. RESEARCH MANAGEMENT # NSF Looks Inward for Geoscience Head 1. Jeffrey Mervis It's been 16 years since the National Science Foundation (NSF) has hired a senior research manager from within the ranks. But that streak seems about to end. Unlike other federal science agencies, NSF has relied on so-called rotators—academic scientists chosen after a national search and on temporary leave from their institutions—to head each of its seven research directorates as assistant NSF directors (ADs). The theory is that the best people to allocate the agency's$5 billion research and education portfolio are those working at the frontiers of science—and who want to go back. Indeed, the succession of outsider appointments since Mary Clutter was promoted to lead the biology directorate in 1989 has created the impression that such a policy was carved in stone.

Apparently not. This month, NSF Director Arden Bement took a chisel to that philosophy by laying the groundwork for the incumbent, paleoceanographer Margaret Leinen, to remain as head of the geosciences directorate. A job notice posted 5 July says that candidates must be federal employees and that the successful applicant will be hired as a career appointment. They also have a scant 2 weeks to apply. That combination has raised eyebrows among those within the foundation who feel that NSF should conduct a national search before filling such an important job. But the rules reflect Bement's willingness to consider alternatives to the rotator model as well as his high regard for Leinen.

“She's eminently qualified, and we'd like to have her around for a long time,” says Bement. “She's truly outstanding, and she has great experience.” Bement also says he's “much more open” to the idea of having a career AD than some of his predecessors.

Like a typical rotator, Leinen, 58, took leave from her post as a top administrator at the University of Rhode Island (URI) to come to NSF in January 2000. Three years later, then-NSF Director Rita Colwell announced a national search for the geosciences post that culminated in extending Leinen's stay at NSF until January 2007, in a temporary category called a limited-term employee. If Leinen is hired under the new job posting, she would need to sever her ties to URI and become a career federal worker.

“I have not yet applied for the job,” Leinen said last week, “and whether or not I do is my business.” She declined further comment.

Scientists who have seen Leinen in action give her high marks. “She's good at articulating environmental issues and programs in ways that excite people and at getting the geosciences involved in interdisciplinary efforts,” says Susan Brantley, head of the Earth and Environment Institute at Pennsylvania State University, University Park, and a member of NSF's geosciences advisory committee. Another advisory panelist, Robert Harriss of the National Center for Atmospheric Research in Boulder, Colorado, which receives more than $60 million a year from NSF, says approvingly that “she likes to think big.” Bement acknowledges that NSF's “routine approach” to finding an AD is through a national search. But Leinen rose to the top during two such previous searches, he notes, adding that he sees no reason “to stretch things out for 6 months” if there's a person “who's fully qualified.” 10. AVIAN INFLUENZA # Vietnam Battles Bird Flu ... and Critics 1. Dennis Normile While international experts warn of a potential human pandemic, scientists in Vietnam are striving to do their share to bring the H5N1 avian influenza virus under control HANOI AND HO CHI MINH CITY—Ever since the H5N1 avian influenza virus started its devastating sweep through Southeast Asia, international health experts and Western scientists have vacillated between praising and criticizing Vietnam's efforts and willingness to cooperate. The World Health Organization (WHO) has repeatedly complained that Vietnamese authorities have been slow to report new human cases and relatively tight with epidemiological data. But the government's own epidemiologists did sound an alarm this spring, alerting international authorities after they had spotted apparent changes in the virus's behavior. Vietnam asked for help in interpreting data but also seems intent on going its own way in vaccine development, to the consternation of some public health experts. Research institutes here are looking for collaborations, but they have subtly let it be known they will pick and choose among proposals to ensure that they are mutually beneficial. This leaves some prominent virologists begging for samples and information. A particularly sore point among researchers trying to track the outbreak in poultry has been the paucity of relevant data from Vietnam. But Vietnamese animal health authorities readily admit that their surveillance system is rudimentary, and this is the area that's received the least international assistance. Despite the system's shortcomings, scientists working with counterparts in Vietnam say that the country deserves praise for doing so much with so little. “We have been very impressed with the work that the Vietnamese are doing,” says Tim Booth, director of the Viral Diseases Division at Canada's National Microbiology Laboratory in Winnipeg, which is collaborating with Vietnamese colleagues on serum surveys that might uncover subclinical cases. Adds Masato Tashiro, who has been on several missions to Vietnam as director of WHO's Collaborative Center for Influenza Surveillance and Research at Japan's National Institute of Infectious Diseases (NIID) in Tokyo: “The surveillance and research effort by Vietnam, I think, is highly respected in the community.” Vietnamese scientists are both stung and genuinely puzzled by claims that they aren't cooperating. At the National Institute of Hygiene and Epidemiology (NIHE) in Hanoi, Vice Director Pham Ngoc Dinh lists the foreign labs the institute has worked with: NIID, the U.S. Centers for Disease Control and Prevention (CDC), Hong Kong's Department of Public Health, and Canada's National Microbiology Laboratory. At the Pasteur Institute in Ho Chi Minh City, virologist Phan Van Tu mentions many of the same partners and adds a group in Israel. Still, Tu emphasizes that as long as the outbreak continues, there will be additional opportunities for international support. “We want to do more [to combat the outbreak] on our own,” Tu says. “But we need collaborations with labs in developed countries for transfers of technology.” ## Starting from scratch Vietnam has every incentive to take H5N1 seriously. Not only has it been hit with the largest number of human cases (87) and deaths (38), but losing 46 million chickens and ducks to disease and culling has also had a cruel economic impact. The government estimates that in the first half of 2004 alone, direct and indirect costs ran to$190 million and shaved 0.5% off the nation's gross domestic product.

But in confronting this challenge, Vietnam's public health system was starting almost from scratch. “Before the outbreak, few scientists in Vietnam were interested in influenza because very few people die of ordinary flu here,” says Tu. Despite the country's experience with severe acute respiratory syndrome in 2003, its labs weren't equipped to deal with infectious pathogens. “We didn't even have personal protection equipment” such as masks and gloves, says Le Thi Quynh Mai, an NIHE virologist. (Both NIHE and the Pasteur Institute are now Vietnamese governmental institutions.)

Early in the outbreak, Tu recalls, Pasteur drew up a wish list of needed equipment; WHO arranged for roughly $60,000 worth of polymerase chain reaction (PCR) test gear, safety cabinets, centrifuges, incubators, and freezers. And NIHE got a similar package of goodies. One important item that didn't come through, however, was a biosafety level 3 (BSL-3) lab—ruling out local work on some important aspects of the contagious H5N1 virus. Vietnamese epidemiologists and virologists have done well with the resources at hand, says Jeremy Farrar, director of the Oxford University Clinical Research Unit at the Hospital for Tropical Diseases in Ho Chi Minh City, who points to two major achievements. The first was simply catching the few human H5N1 cases among the hundreds of patients with respiratory illnesses that turn up each week. As soon as H5N1 was reported among poultry in the region in late 2003, NIHE's Mai readied reverse transcription-PCR (RT-PCR) assays, which probe for the RNA of a replicating virus, as a precaution. And in January 2004, her lab detected the H5N1 virus in a flu victim, confirming the first human case of the outbreak in Southeast Asia. “It was one of the biggest achievements of our institute,” NIHE's Dinh says. NIHE and Pasteur immediately started collecting throat swabs and serum samples from family members and contacts of victims, as well as from random poultry workers. Through the first months of 2004, NIHE collected several hundred samples in northern Vietnam; Pasteur got several dozen more in the south. In addition to patients, their contacts and poultry workers were tested using the RT-PCR assay; the results were overwhelmingly negative. The two institutes were unable to check for antibodies to the virus in blood samples, a sign of past infection, because the most sensitive procedure, the microneutralization assay, requires a BSL-3 lab. Instead, they shipped the samples to CDC in Atlanta, Georgia, where tests confirmed the negative findings. The second achievement was detecting a change in the behavior of the virus in early 2005, when epidemiologists traced contacts in northern Vietnam and turned up clusters of cases, including asymptomatic and mild cases. This raised fears that the virus was becoming less lethal but more infectious. At Vietnam's request, WHO sent in a fact-finding team of virologists and public health experts and organized expert meetings to review the data. (In late June, WHO issued a brief report concluding that the virus was not undergoing a major change.) Vietnamese scientists seem miffed that even achievements such as these don't always earn credit. For instance, although scientists from both NIHE and Pasteur presented findings from their 2004 surveys at avian influenza meetings in Asia last year, as late as spring of 2005, members of the global influenza community were publicly urging Vietnam to undertake these kinds of surveys (Science, 25 March, p. 1865). And when the asymptomatic cases surfaced this spring, many Western scientists—instead of praising the subtle finding—openly worried that Vietnam's epidemiologists were missing a hidden iceberg. Yet later, WHO concluded it was unlikely that large numbers of cases were going undetected. This gap in communication is due partly to the fact that Vietnamese researchers are only now forming ties to the tight-knit global influenza community, and partly to differing attitudes toward publishing. Peter Horby, an epidemiologist in WHO's Hanoi office, explains that Western researchers rush results into high-profile journals. But that tradition “is much less strong here,” he says, adding that Vietnamese scientists feel their first responsibility is to provide scientific data to guide government policy. ## Respect the dead Although Vietnam's epidemiological work has earned respect, international health officials and scientists are frustrated with some aspects of the clinical research and are questioning the country's vaccine efforts. In the 18 months since the outbreak hit Southeast Asia, only four or five autopsies have been performed on victims—all of them in Thailand. Nguyen Hong Ha, an infectious disease specialist at the Institute for Clinical Research in Tropical Medicine in Hanoi, says scientists there understand the importance of autopsies in elucidating how the virus attacks the body. “But by Vietnamese law, the consent of the family is required to conduct any postmortem research, and families have not given that consent,” he says. Farrar adds that a national association of physicians has urged the Vietnamese government either to modify the law or to exercise the right to mandate autopsies under exceptional circumstances. But authorities hesitate to go against a deeply rooted cultural bias. WHO is also concerned about—and very interested in—NIHE's effort to develop an H5N1 vaccine for humans. David Wood, coordinator of WHO's Quality Assurance and Safety of Biologicals Team, says the organization applauds Vietnam's initiative but worries that the development process used for one of its vaccines “goes into uncharted territory,” raising questions about efficacy and contamination. NIHE's researchers derived a live but modified H5N1 virus in so-called 293 cells and are culturing the virus in primary monkey kidney cells for use in vaccines. This process takes some unprecedented steps, the first being the use of 293 cells, which in some cases have produced tumors when injected into mice with deficient immune systems. Wood says that he's seen little data on the specific batch of 293 cells used by NIHE. “Consequently, we do not know the risks posed by any 293 cell residuals in the vaccine,” he says. Wood also expresses concerns about potential contaminants in the primary monkey kidney cells: “There are tests to exclude such risks, but we do not know how they are being implemented for the flu vaccine production [in Vietnam].” WHO has provided an alternative vaccine seed stock that meets international quality standards. But there is a catch. It was produced using a patented reverse-genetics process in which the viral genes are cloned individually and assembled into a “safe” strain. Hoping to avoid fees that could make the vaccine unaffordable for Vietnam and other developing countries, WHO is negotiating deals with MedImmune Inc. in Gaithersburg, Maryland, which holds the patent for the reverse-genetics process, and with those who hold patents covering other aspects of the seed stock. Meanwhile, NIHE's Dinh says the vaccine program is continuing, without confirming current plans. “Our institute has a lot of valuable experience producing vaccines, and we strongly believe we can reduce the number of human infections with our own H5N1 vaccine,” he says. ## It's about the birds The biggest hole in Vietnam's effort to contain H5N1 may be its spread in poultry—which many experts believe must be controlled to keep it out of humans. Hoang Van Nam, deputy director of animal health for the Ministry of Agriculture and Rural Development, admits that the animal health infrastructure is woefully incomplete. Many local animal health officers aren't on the government payroll. “They earn money by treating sick animals, and they don't have any responsibility to report disease outbreaks,” he says. Faced with the ongoing outbreak, many commercial poultry farms have adopted effective biosafety practices, so the brunt of the animal epidemic is now falling on rural households that keep small flocks for their own consumption. In such settings, sick birds are even harder to spot. When chickens die, Nam says, “they don't think to report it; they just discard or burn the dead chickens.” To reduce the threat, the Ministry of Agriculture is starting a pilot poultry vaccination program in two provinces this summer; it hopes to extend it to all affected provinces before the winter flu season. For now, says Nam, the ministry plans to mandate vaccination only at farms with 200 or more birds. Extending the campaign to small holders—those with one or two dozen birds—would require additional governmental or international support. Indeed, there is no shortage of needs. Asked if her hospital could handle H5N1 patients turning up by the dozens instead of in ones and twos, Nguyen Thi Dung, a physician handling human avian flu cases at the Hospital for Tropical Diseases, just winces. For one, they don't have a proper isolation ward. “We have a plan to build one but no funding yet,” she says. Pasteur's Tu says that to extend the search for H5N1 cases, his institute and NIHE are setting up a network of six sentinel hospitals around the country; each week, each hospital will screen throat swabs and serum samples from 10 patients with respiratory problems to see if any cases of avian flu are going undetected. Again, Tu says Pasteur would like to expand the network to more regional hospitals—a move recommended by many epidemiologists—“but we'd need more [outside] support.” Help might just be on the way. At a conference on avian influenza in Kuala Lumpur last week, representatives of WHO, the U.N. Food and Agriculture Organization, and the World Organisation for Animal Health renewed appeals to the international community to support the fight against bird flu. On the animal side, the organizations emphasize the need to bring vaccination campaigns and biosafe farming practices to small-scale and backyard poultry operations. For public health, they want to strengthen the capabilities of affected countries in laboratory diagnosis, vaccine development, surveillance, and public education. They are seeking donations of$100 million for the poultry farming plans and $150 million for human health efforts. Vietnam is ramping up its own funding for H5N1 efforts. The government, for instance, is providing$5 million each to Pasteur and NIHE for BSL-3 labs. But NIHE's Dinh says they will continue to request further support from developed countries. “Any pandemic would not only affect Vietnam but the world, and we have to share responsibility for averting it,” he says.

11. AVIAN INFLUENZA

# Pandemic Influenza: Global Update

1. Martin Enserink

An explosion of the H5N1 strain in Asia is fueling worries that a lethal new flu virus may soon sweep the world. Around the globe, other avian influenza strains have occasionally infected humans in recent years as well, as shown on this map.

12. AVIAN INFLUENZA

# Who Controls the Samples?

1. Dennis Normile

In early 2004, University of Hong Kong virologist Guan Yi was tracking the evolution of the H5N1 virus that had infected poultry in Japan that winter when he found a source of samples. But he also ran into a Catch-22. Concerned about bioterrorism, both Japan and Hong Kong had put H5N1 on the list of infectious agents requiring special handling. Guan's lab needed an import permit, and Japan's National Institute of Infectious Diseases (NIID) needed an export permit. In theory, the transfer was doable, but Hong Kong's import permit was valid for only 1 week, and NIID couldn't get export approval that fast.

“We tried time after time, but we were never able to export the material,” says NIID virologist Masato Tashiro.

At a time when averting a global influenza pandemic may depend on the rapid sharing of samples and information, researchers in developed and developing countries alike are running into roadblocks. The bioterror concerns are being added to the usual difficulties of sharing coveted samples that may have commercial value or may give researchers an advantage in scientific prestige and funding. The need for speed is imperative. Like all flu viruses, H5N1 is continually changing. Researchers want to track its genes to see if the virus is becoming more easily transmissible among humans, which could trigger a pandemic. This also means that vaccines and reagents used in diagnostic kits must be updated regularly to be effective.

“For this kind of pandemic preparedness, we need samples almost in real time,” says Guan. The problem is finding the best way to share them.

Despite a consensus that samples of H5N1 strains should be sent to international reference labs as soon as they appear, no existing international agreement requires labs or countries to do so. “You have to ask” individual labs and governmental authorities for samples and information, says Astrid Tripodi, former avian influenza coordinator for the United Nations Food and Agriculture Organization (FAO) in Hanoi.

And those requests have to get to the right people. Although Vietnam was faulted for delays in sharing viral samples earlier this year (see main text), Phan Van Tu, head of the Department of Microbiology at the Pasteur Institute in Ho Chi Minh City, says no one requested the serum samples he collected over the winter and spring. So he sent one bulk shipment of samples to the U.S. Centers for Disease Control and Prevention (CDC) in Atlanta, Georgia, in mid-June, just as he did last year.

Even sending samples to CDC and other World Health Organization (WHO) collaborating centers may not mean that samples and results get to all who want them. The collaborating centers must first report the results of any studies to the country or lab that supplied the samples; they also need the source's permission to pass samples on to third parties.

Bioterror precautions impede sharing, even with highly regarded labs. In early 2004, Canada's National Microbiology Laboratory (NML) in Winnipeg asked CDC for recent H5N1 samples to update diagnostic kits. “If an outbreak hit Canada, local hospitals would expect us to provide the proper reagents for diagnostics,” says NML virologist Yan Li. Even so, he says, approval took 4 or 5 months.

David Daigle, a spokesperson for CDC's National Center for Infectious Diseases, confirms that an initial clearance for exporting a “select agent” can take several months, although the agency is trying to speed the process.

Approvals generally go more smoothly when the receiving lab is affiliated with WHO or FAO, says Tashiro, who heads WHO's Collaborative Center for Reference and Research on Influenza at Japan's NIID. But he would like to send samples more easily to nonnetwork labs. He would also like to see the WHO network database, which has extensive information on sequences and viral genetics, opened to outside researchers.

Margaret Chan, WHO's new director of Communicable Disease Surveillance and Response, agrees with the need to disseminate specimens and data more widely. She says expanding the WHO network is a possibility, but new labs “would have to satisfy biosafety requirements.”

But that won't solve the problem for university researchers, who feel left out. Albert Osterhaus, a virologist at Erasmus University Medical Center in Rotterdam, the Netherlands, thinks WHO and FAO should assemble a permanent H5N1 task force, modeled on the one that tackled the severe acute respiratory syndrome virus. He would like to see government, university, and even private sector labs share samples and research results rapidly and freely. “This goes way beyond scientific cooperation,” he admits, but should be done to safeguard public health.

Tim Booth, director of the Viral Diseases Division of Canada's NML, among others, remains unconvinced of the need to abandon traditional approaches to collaboration just yet. “I don't think rigid regulations on [sharing samples and information] would be useful,” he says, adding, “it is better to build collaborations by developing a mutual understanding of the roles and responsibilities of all parties involved.” The question is which model of cooperation will best suit the challenges posed by a potentially pandemic virus.

13. EVOLUTION 2005 MEETING

# Fungal Trees Grow Faster With Computer Help

1. Elizabeth Pennisi

FAIRBANKS, ALASKA—At Evolution 2005, from 10-14 June, evolutionary biologists, natural historians, and systematists shared results about fungi, mice, yeasts, and other organisms.

Researchers trying to determine the relatedness of organisms are finding it hard to keep up with the torrent of DNA sequence data gushing from biology's spigots. Now, two new computer programs are coming to the rescue, at least for biologists constructing the fungal family tree. One program, created by Frank Kauff of Duke University in Durham, North Carolina, and his colleagues, helps validate, assemble, and keep track of raw data from fungal DNA sequencing efforts. The other, developed by David Hibbett of Clark University in Worcester, Massachusetts, automatically retrieves fungal DNA sequences from the public archives and incorporates the data into an ever-improving phylogeny of this diverse group of microorganisms. Both efforts are part of the “Assembling the Fungal Tree of Life” project begun in 2003 and may be bellwethers of taxonomy's future. “It's great to have this all automated,” says Michael Donoghue, a botanist at Yale University. “It means that progress can be made while we sleep.”

Molecular studies now dominate fungal systematics, but the plethora of data they provide has not necessarily brought clarity. There are hundreds of published family trees for the fungi or their various branches, and many conflict with one another. Yet no one has really tried to piece together where the discrepancies lie. That's where automated computer analyses will help, says Hibbett, a fungal systematist.

Among other fungal projects, Hibbett's lab focuses on mushroom-forming varieties, which make up an estimated 20,000 of the more than 70,000 known fungal species. To deal with the ever-growing number of DNA sequences for this group, Hibbett's program, which he dubbed mor, sifts through GenBank for newly deposited data on a single gene, called nuc-lsu rDNA, in mushrooms. If a researcher has deposited a new sequence of this gene for a species, the computer program compares it with other deposited copies of the gene for that species, weeding out any redundancies. It then compares the best version with the sequence of the gene in other species and uses the differences to adjust the branches of the fungal family tree. It even assigns names to new subgroups as needed. So far, mor has 2401 sequences representing 1899 mushroom species in 562 genera, Hibbett reported in Fairbanks, Alaska.

“It's one of the first attempts to automate large-scale phylogenetic analysis,” says Roderic Page, a systematist at the University of Glasgow, United Kingdom.

Although fungal experts may need that help more than most—these organisms are among the most diverse and the most difficult to sort out—Hibbett's approach should also be portable. “It's easy to see how it could be expanded to fit other organisms,” says David Baum, a botanist at the University of Wisconsin, Madison. Adds Donoghue, “I'd love to have something like this for plants.”

Kauff's program, dubbed WASABI for Web Accessible Sequence Analysis for Biological Inference, comes into play before fungal family trees are created. In essence, it ensures that such trees sprout from good seeds. The consortium working on the fungal tree of life project is sequencing eight genes in 1500 different fungi, and WASABI rates the accuracy of each newly submitted DNA sequence. The program also pieces together short fungal DNA sequences into ever longer ones and compares these so-called contigs with existing sequence information. This all happens automatically, providing researchers with one place to find refined data that originated from various consortium members. Finally, WASABI archives its manipulations and analyses of the raw information. “WASABI considerably reduces the time users would otherwise have to spend,” verifying and piecing together sequences, says Kauff. “The speedup is many orders of magnitude.”

Together with other consortium members, Hibbett and Kauff have already published one 588-species fungal tree, with all the major branches, such as the mushrooms, represented. The goal is to have the 1500 under study linked up in the proper relationships by 2006. Says Baum, “Fungal systematists are really leading the pack in terms of their critical use of cutting-edge analytical tools.”

14. EVOLUTION 2005 MEETING

# Color Genes Help Mice and Lizards

1. Elizabeth Pennisi

FAIRBANKS, ALASKA—At Evolution 2005, from 10-14 June, evolutionary biologists, natural historians, and systematists shared results about fungi, mice, yeasts, and other organisms.

The light-skinned deer mice (Peromyscus polionotus) found along Florida's shoreline didn't always have such a bleached-out look. It took the beach rodents less than 5000 years to go from brown to blond; the darker look may have provided camouflage in the dense fields in which they used to dwell, but on the white sand, it would have made the mice a conspicuous meal for predators. At the meeting, a research team described how a key gene aided the animal's colorful transformation. And another group reported that changes in the same gene helped lizards evolve a similar adaptation.

Researchers have studied the genetics of color in lab mice for decades, implicating more than 100 genes, half of which are now sequenced. But Hopi Hoekstra, an evolutionary biologist at the University of California, San Diego, says she “wanted to see what kinds of genes are involved” in shaping color patterns in nature.

In the southeastern United States, deer mice living in forests and dense fields have brown backs and light gray underbellies. But their cousins living on the vegetation-sparse white dunes on islands along the Gulf Coast have lost most of the brown on their backs, and their bellies look bleached. The beach mice have also dropped a characteristic dark stripe running down their face for a more muted look that helps camouflage the animals in their burrows.

To get at the genetics behind such adaptations, Hoekstra and her colleagues bred male beachcombers with female forest mice and vice versa. They now have 600 second-generation mice. “We see a lot of variation in pigmentation” among the animals, says Hoekstra, estimating that about a dozen genes control the pattern of colors distributed across the rodent's flanks, faces, tails, and other body parts. With these crossbred mice, she began testing whether various genes shown to have roles in coloration in lab mice are involved in the beach mouse's new look. “Hoekstra can ask where in the pathway natural selection is working,” notes Johanna Schmitt, an evolutionary biologist at Brown University in Providence, Rhode Island. By happenstance, Hoekstra and her colleagues scored a hit with Mc1R, a gene involved in the switch between light and dark pigments. A single base change in the gene resulted in the Mc1R protein having abnormally low activity, causing less melanin to be made in the beach mice and resulting in whiter fur. In fact, the change in just this one gene accounts for 34% of the color variation in beach mice, Hoekstra reported. Hoekstra's postdoc Cynthia Steiner subsequently showed that a second gene called agouti is more significant for patterning than overall color.

Further analyses indicate that the two genes influence each other, a process called epistasis, in defining the overall patterns of body coloration. “It's the interaction that explains the variation” in color from body part to body part, Hoekstra notes.

Lizards from White Sands, New Mexico, also seem to have exploited changes in Mc1R to transform themselves from dark brown to light-colored, Erica Rosenblum of the University of California, Berkeley, reported. She studied three distantly related lizard species that have moved into the dunes in the past 600 years. Rosenblum found that all three had mutations in the gene, dramatically reducing their colors. “What is most striking is the repeating pattern as different species converge on the same phenotype,” says Hoekstra.

Lizards and mice are far apart on the tree of life, and scales and fur bear little resemblance, but the metabolic pathways to produce melanin pigment in both animals are very similar. As a result, “it may be evolutionarily 'easy' to evolve color and color pattern differences” by means of the Mc1R gene, says Rosenblum.

15. EVOLUTION 2005 MEETING

# Wine Yeast's Surprising Diversity

1. Elizabeth Pennisi

FAIRBANKS, ALASKA—At Evolution 2005, from 10-14 June, evolutionary biologists, natural historians, and systematists shared results about fungi, mice, yeasts, and other organisms.

Since the days of the pharaohs, the yeast Saccharomyces cerevisiae has enabled us to make bread, as well as wine, beer, and other alcoholic beverages. More recently, it has become a model organism for cell and molecular biologists. Yet it has barely been studied outside the lab. Now, a research team has begun to trace the genetic diversity of this simple eukaryote in the wild.

Evolutionary biologist Jeffrey Townsend of the University of Connecticut, Storrs, and his colleagues have identified several distinct S. cerevisiae strains from forests and vineyards in Italy and the United States. Different strains found on grapes from different vineyards “may in part be responsible for the distinctive tastes of naturally fermented wines,” Townsend speculates.

Until recently, yeast researchers paid little mind to grapes, thinking that any yeasts on the grapevines were escapees from the nearby vats, where the microbes are often added for the fermentation process. That thinking came into question, however, in 2004, when Paul Sniegowski of the University of Pennsylvania in Philadelphia discovered S. cerevisiae just below the bark of oak trees and in the soil around the base of these trees, establishing that this organism had a broader distribution beyond rotting fruit and vineyards. He “demonstrated that there are isolated, variant populations of S. cerevisiae,” says Townsend.

Sniegowski's finding led researchers to wonder how many yeast strains there are in the wild, how the oak strains are related to those in vineyards, and whether one is derived from the other. While working in John Taylor's lab at the University of California, Berkeley, Townsend and graduate student Erlend Aa of the University of Tromsø in Norway compared DNA of 15 S. cerevisiae strains from Italian vineyards—primarily from grapes used in Chianti wine—with two lab samples and a strain from crushed grapes used to make wine. They also analyzed yeast strains provided by Sniegowski that were found on and near oak trees.

Aa sequenced four genes from each yeast and found 78 single-base differences in these genes among the strains. Various combinations of these altered genes established distinguishable genotypes for each sample. Aa and Townsend demonstrated that the yeast found on grapes were not that similar to the yeast recovered from the wine must in fermentation vats. Instead, yeast from wine vineyards around the world include many wild strains and greater genetic diversity than that of yeast from the must. “The wine yeast does not represent a [global] population of domesticated strains as has been suggested,” notes Christian Landry of Harvard University in Cambridge, Massachusetts. The vineyard yeast were also quite different than the yeast recovered from oaks.

Two samples from Italy's Elba Island also hinted that the yeast found on grapes may differ significantly from vineyard to vineyard within a region. Townsend discovered that yeast from the Elba samples resembled mainland strains but also contained genotypes unique to the island. He plans to expand the study to determine whether other places have distinctive yeast populations and, perhaps as a result, distinctive wines.

Two of the four yeast genes studied by Townsend and Aa had telling changes that may explain some of the vineyard-to- vineyard strain variation. One, the SSU1 gene, is involved in transporting sulfite—a toxin—out of the yeast cell. The second is a gene whose protein regulates SSU1's activity. The more active SSU1 is, the more resistant the yeast is to this toxin. The SSU1 regulatory gene showed the greatest number of differences from strain to strain, which translated into slightly different proteins and indicated that it had evolved the fastest of the four genes studied. Viniculture practices could explain this rapid change, says Townsend. In the vineyard, grapes are treated with sulfite and sulfite-containing compounds that destroy mold and other microbes, presumably killing all but those yeast with high SSU1 activity. Also, winemakers add sulfite to sterilize fermentation vats, again presumably killing all but the most tolerant yeast.

Townsend notes that with such treatments, winemakers end up with ever more useful strains. The more resistant a S. cerevisiae strain is to sulfur-based chemicals, the longer the yeast cells will survive in vats treated with sulfite, and the more alcohol they make. “[Wild] wine yeast has inadvertently been domesticated,” concludes Townsend. That's worth a celebratory drink.

16. NANOMATERIALS

# 'Smart Coatings' Research Shows The Virtues of Superficiality

1. John Bohannon*
1. John Bohannon is a writer in Berlin, Germany.

Thin, shallow, and out to strike it rich—high-tech protective paints and varnishes look poised to become the first “killer apps” for nanotechnology

BERLIN—Clothing with computers woven into the fabric. Microscopic robots that make repairs with tools the size of a virus. No question about it: Nanotechnology, the applied science of the very small, has generated its share of megahype. For companies researching nanomaterials, however, profitability is the priority—and not in the dreamy future but now. Many are concluding that the beauty of the technology is literally skin deep.

At a recent meeting here,* researchers from around the world swapped news about efforts to spin nanotech into products based on surfaces with novel properties. “Coatings applications are among the first true everyday uses of nanotechnology,” says Dirk Meine, a chemist who organized the conference for Vincentz Network, a coatings industry media group. Examples include nanoparticle-laden varnishes that combine the scratch resistance of an inorganic crystal with the versatility of an organic plastic. (Super-scratch-resistant coatings are already on the market.) Researchers offered a glimpse of what may be the next wave of nano applications to enter daily life.

## Combating corrosion

The biggest task in the coatings industry is to slow down corrosion. Pipes rust, bricks crumble, and timbers rot, calling for repairs that add up to 4% of the gross national product of Western countries, according to Ubbo Gramberg, a corrosion chemist at Bayer in Leverkusen, Germany. “Not all these corrosion problems can be solved by coatings, but a considerable percentage can,” says Michael Rohwerder, a physicist at the Max Planck Institute for Iron Research in Düsseldorf, Germany.

Top prize will go to a coating that prevents the corrosion of steel. Today, even the best protective coatings allow oxygen to diffuse slowly through to the metal surface. Corrosion kicks into overdrive when coatings begin to peel off, a process called delamination.

The trouble starts at microscopic nicks or pits on the surface introduced during manufacturing or through wear and tear. These defects form miniature circuits in which electrons flow through the metal in one direction while positive ions such as sodium flow back along the metal surface, leaving a degraded metal-coating interface in their wake. The coating becomes separated from the metal and flakes away, exposing fresh metal and accelerating the process.

That is where nanotechnology could come to the rescue. Rohwerder's group is working on coatings that allow a corroding metal surface to “self-heal.” The oxidative attack at the site of a defect triggers nanoparticles to release corrosion-inhibiting ions—in this case, negatively charged molybdate ions—that stand in for the metal and form a protective oxide skin. Once the defect is sealed, the coating stops releasing ions until the next attack.

But there's a catch. Because these coatings sense corrosion with innately conductive polymers (ICPs)—carbon chains that allow charge to flow along their length like the semiconductors in microchips—they actually promote corrosion except under controlled conditions. Designing “smart” ICP coatings that remain protective in unpredictable environments requires a “balancing act,” says Sze Yang, a chemical engineer at the University of Rhode Island in Kingston. Another problem is that most ICPs are difficult to work into standard coating solutions, a drawback that could make them a commercial nonstarter despite their excellent anticorrosion properties.

Yang says he and colleagues have discovered an elegant solution to some of these problems. They found that synthesizing an ICP called polyaniline into a DNA-like double helix makes it far less corrosion-prone. The helix form is also easier to integrate into several coating mixtures. The researchers hope to find a replacement for chromates, the nearly universal additive to metal coatings that protects against corrosion but is a potent toxin, causing environmental havoc when it leeches out. It's too soon to say whether the badly needed successor to chromate will be Yang's double-helix ICP, but most coating experts agree that whatever it is, it will likely come from nanotechnology.

## Fighting fire with nanoparticles

Nanomaterials may also help hold at bay rust's dramatic cousin, fire. Flame-retardant coatings have been widely used since the 1970s, but they have a serious drawback. According to Stefan Sepeur, a chemist at technology company Nano-X in Saarbrücken, Germany, more than 90% of fire-related deaths “are not caused by the flames but by the emission of toxic and corrosive gases”—many of which come from the fire-retardant coatings themselves. So finding alternatives for these formulas, which include toxic epoxy and acrylates, would save lives.

At Inomat, a coating research company in Bexbach, Germany, engineers have developed a way of coating surfaces with nanoparticles of flame-retardant oxides of aluminum or silicon. Because the particles are so small, they can be incorporated into a water-based solution, sidestepping the toxic organic compounds that make up standard formulas. One problem still to be overcome is that it requires a temperature of 100°C during application, limiting its use to steel and aluminum rather than the inside walls of houses where it is most needed.

In a very different approach, researchers at the Fraunhofer Institute for Wood Research in Braunschweig, Germany, are trying to endow surfaces with their own fire extinguishers. What looks and behaves like a normal paint or varnish at room temperature suddenly erupts in a layer of carbon foam in the presence of flame. The foam, composed of so-called ceramizing elastomers, was developed 35 years ago to insulate the combustion chambers of rockets, says Sebastian Simon, a chemical engineer on the project. The first challenge, he says, was to engineer the unexpanded polymers into a heat-sensitive coating that could pass muster as a household varnish. To test it, Simon and colleagues coated a wooden staircase and roasted it at 900°C for half an hour. Each stair could still bear a 100-kilogram weight after the ordeal.

War and peace

Other researchers are pitting smart coatings against even worse worst-case scenarios. At the University of Pittsburgh, Pennsylvania, molecular biologists Richard Koepsel and Alan Russell are working on a coating that protects against attacks with biological or chemical weapons. With funding from the U.S. military, they are developing a “bioreactive plastic” embedded with antibodies and enzymes that decontaminate surfaces as soon as pathogens or toxins arrive.

The biological principles are simple, Koepsel says. The coating contains enzymes for breaking down various poisons into harmless smaller molecules. And for each of the prime pathogenic suspects, such as anthrax or smallpox, a specific antibody lies in wait to grab it with enzymes such as cell-popping lysozymes nearby.

One problem is that all these proteins evolved to function in the wet, salty environment within organisms, not outdoors. But after much tinkering, Koepsel and Russell found mixtures of water-retaining materials such as polyurethane in which enzymes remain 60% active after more than 20 weeks. Other chemical tricks helped keep the proteins at the outermost surface of the coating where the action is, instead of trapped and useless within the interior.

Koepsel says the coating has done well against simulated attacks of E. coli bacteria and harmless molecules, and the same principles should apply for the real deal. A self-decontaminating surface alone, he acknowledges, won't keep people inside a building or vehicle perfectly safe. But it should provide at least “a moderate level of protection for occupants of unsealed buildings and vehicles” and could make it easier to clean up after an attack. He says he is also pondering ways to equip the coating with an alarm system to alert people that an invisible attack is taking place.

Antimicrobial coatings could also soon find niches in operating rooms and in medical devices such as catheters that must remain inside the body for days at a time. When it comes to fighting infection, “nano is a natural given the size of bacteria,” says Alexander Klibanov, a materials scientist at the Massachusetts Institute of Technology in Cambridge. The main obstacle right now, Klibanov says, is that antibacterial nanocoatings are expensive. He expects that the coatings won't make it onto the consumer market until they've been adopted by the “price-insensitive” hospital and homeland security areas. But once they have become cheap and proven effective, he predicts, they're bound to become as common as a coat of paint.

• * Fourth Annual Smart Coatings Conference, 9-10 June.