News this Week

Science  10 Apr 1998:
Vol. 280, Issue 5361, pp. 198

You are currently viewing the .

View Full Text

Log in to view the full text

Log in through your institution

Log in through your institution


    T Cells on the Mucosal Frontline

    1. Nigel Williams


    Over a decade ago, a long molecular hunt ended when immunologists finally bagged one of the field's most elusive molecules: the receptor that the T cells of the immune system use to recognize antigens, the distinctive molecules carried by pathogens and other foreign invaders. Until then, they had been stymied in their efforts to understand this key step in the initiation of many immune responses. But the discovery of the receptor almost immediately left researchers confronting another mystery.

    They found that T cells actually carry two different types of receptors. The great majority—some 90%—carry so-called αβ receptors and are the classic T cells that circulate around the body, performing such jobs as helping get rid of viruses and triggering antibody production. But a small percentage turned out to carry a different receptor variant—designated the γδ receptor—and no one knew what this maverick T cell population might do. A growing body of evidence now suggests, however, that these γδ cells have a more important role in the body than their relatively sparse numbers might suggest.

    Cross talk.

    Epithelial cells and γδ T cells may interact through growth factors and cytokines.

    Indeed, they may be the first line of defense against invading pathogens, especially those attempting to enter the body through mucosal surfaces such as the lining of the gut. “At these sites γδ cells represent a significant population, and interest in unraveling their role is now compelling,” says immunologist Adrian Hayday at Yale University. Already, researchers are learning that γδ cells act in ways that seem consistent with a role in mucosal immunity. “The development of responses is very clearly different in the mucosa compared with the periphery, and it is attracting increasing interest,” says immunologist Pierre Vassalli at the University of Geneva.

    Among other things, although γδ cells have much less variability in their antigen-recognition sites than αβ cells have, they nonetheless appear capable of dealing with the broad spectrum of antigens they are likely to encounter at mucosal surfaces. What's more, they may be able to go to work without needing to be activated by the same complex series of steps that αβ cells require—a big plus for cells that may need to act quickly on the front lines of pathogen defenses. Other findings are also raising researchers' estimation of γδ cells. Evidence suggests, for example, that they help the body damp down autoimmune reactions and repair tissues damaged by inflammation.

    The new respect for γδ cells is a big change from immunologists' view of these cells soon after they were discovered, when there was good reason to wonder whether the cells play any important role in the body. The researchers were intrigued by one finding: Although the cells are a relatively small T cell population—they constitute only about 5 % of circulating T cells—they are much more common in certain epithelial tissues. They make up about 50% of the lymphocytes in the intestinal mucosa of mice, for example, and as much as 40% of the lymphocyte population in human colon epithelium. That suggested that they might be frontline sentinels. As Hayday points out, “the γδ T cells are in a unique position to monitor epithelial cells.” Counterbalancing that, however, was the apparent lack of variability in what the cells could recognize.

    The receptors of both αβ and γδ cells consist of two similarly designed protein chains (designated α and β or γ and δ). But although the αβ receptors display an enormous range of structural variation, indicating that they can recognize an equally enormous range of different target antigens, the antigen specificities of the γδ receptors found in individuals appear to be much more limited.

    T cells assemble each of the genes for their receptor proteins by combining two to three separate segments of DNA, and the repertoire of these elements is much smaller for the γ and δ genes. In the most extreme case, essentially all the γδ cells in the epithelia of the mouse reproductive tract turned out to use the same genetic elements to construct the antigen-binding portion of their receptor. In the skin, anywhere from 60% to 99% of the T cells share a similar receptor construction. Although the receptor make-up does vary between locations, this lack of structural variation in the γδ receptors at any one site raised questions about whether they would be able to deal with the wide variety of pathogens mammals encounter.

    Indeed, there were indications that γδ cells might be nothing more than remnants of some evolutionarily primitive antigen recognition system in mucosa. Researchers found, for example, that γδ cells have a simpler, more direct way of recognizing antigens. The αβ cells can recognize antigens only after the target proteins have been “processed,” broken down into small peptides, which are then displayed on the surface of “antigen-presenting cells” in conjunction with proteins encoded by genes within the major histocompatibility complex (MHC).

    In contrast, γδ cells in the epithelia appear to recognize and respond to some MHC proteins directly without added peptides, eliminating the protein-processing The small γδ cell population in circulation also recognizes some bacterial antigens directly without any processing. In this way, γδ receptors seemed to behave more like antibodies than like αβ receptors, an idea that received additional support earlier this year. In an x-ray crystallography study that appeared in the 29 January issue of step. Nature, Roy Mariuzza's group at the University of Maryland Biotechnology Institute in Rockville found that the structure of the δchain's antigen-recognizing segment resembles that of antibody chains.

    Another indication that γδ cells are more primitive came from studies reported 7 years ago in mice, in which researchers showed that these cells appear in the fetus well before the first αβ cells. That could be of some benefit, Hayday says: “The greatest exposure to new antigens for a mammal occurs at birth and γδ T cells may be a vital defense at this time. There's good evidence the αβ cells are not yet working at full clip.” But whether they had a role in the adult was unclear, especially as Susumu Tonegawa of the Massachusetts Institute of Technology and his colleagues found no obvious immunological defects in mice that were unable to make γδ T cells because the researchers had knocked out the δ chain gene.

    But although most of these early findings hinted that γδ cells are a minor player in immunity, at least in adults, more recent results are reviving interest in the cells. Driving much of this revival is evidence that the cells may have a role in mucosal immunity, a critical component of the body's defenses because most invading pathogens have to cross a mucosal surface to enter the body. One indication of that is, of course, their relative abundance in those surfaces.

    More recent work by Tom Spies and his colleagues at the Fred Hutchinson Cancer Research Center in Seattle suggests how γδ cells overcome the apparent limitation of their sparse receptor variability. They may respond to two cell surface molecules that are related to the main family of MHC molecules, known as MICA and MICB, that epithelial cells may display when stressed—say, by an infection (Science, 13 March, p. 1737). In experiments performed on epithelial cells in lab culture, the Spies team found that γδ T cells can recognize these molecules and kill the cells bearing them. “There is a real possibility these γδ T cells respond to these molecules in the body,” says Vassalli. “The data are very compelling.”

    If so, these findings would help explain how γδ cells could be effective at immune surveillance despite the narrow repertoire of their receptors. If the same molecules, such as MICA, are expressed in response to a range of pathogen infections or other sorts of stress, then this mechanism could be highly effective at eliminating infected or damaged cells. “If γδT cells worked in this way, it would make sense,” says immunologist Delphine Guy-Grand at the Necker Hospital in Paris.

    The idea is also consistent with a recent finding by Hayday and his colleagues. They looked at the γδ receptors in knockout mice lacking a gene segment commonly used in the construction of a typical receptor in one particular tissue, the skin. They found that the mice were able to circumvent the loss of this particular gene by co-opting another gene segment to create a receptor of similar structure and specificity (Science, 13 March, p. 1729). This suggests, Hayday says, that the final γδ receptor repertoire, however narrow, is driven by the need to recognize specific molecules in the part of the body where the T cells reside. If cells in which one γ or δ gene segment is knocked out can use another to create a receptor with the same specificity, the Hayday team's result may also help explain why these knockouts don't show obvious effects.

    In addition to recognizing MICA and MICB on epithelial cell surfaces, γδ cells apparently receive signals of another kind from epithelial cells—signals that may help maintain γδ cell populations. They have receptors for growth factors, including stem cell factor and interleukin-7 (IL-7), that epithelial cells secrete. Evidence that these receptors are functionally important comes from knockout studies in which researchers have found that mice lacking the gene for the IL-7 receptor also lack γδ T cells in the gut epithelium, whereas the αβ T cell population is only slightly reduced.

    Researchers have also found that γδcells produce growth factors that may play a role in healing epithelia damaged by infection or inflammation by promoting cell growth there. “These are two complex players, and it's important to understand how they communicate with each other,” says mucosal immunologist John Klein at the University of Tulsa.

    Beside killing infected cells directly, γδ T cells may also help protect the mucosa by drawing other immune cells in to help combat an infection. Work by several researchers has shown, for example, that the cells release immune cell messengers called cytokines that may activate αβ cells and attract inflammatory cells useful in cleaning up damaged cells.

    In some cases, though, γδ T cells may help damp down immune responses that might damage epithelia. Immunologist Martin Kagnoff at the University of California, San Diego, says that areas of the bowel affected by celiac disease, an autoimmune condition resulting from an adverse reaction to cereal proteins, contain higher than normal numbers of the cells. Although their role in the pathology of the disease is not known, the numbers of γδT cells are high during “silent” periods of the disease, when the pathology is mildest, suggesting that the cells may help suppress the autoimmune reactions.

    Similarly, other studies found that γδ knockout mice develop a more serious disease, with more damage to the intestinal epithelium, than controls do when infected with the common protozoan Eimeria. Because animals lacking normal T cells don't develop this damage, the finding suggests that γδ T cells damp down the αβ T cell response to the parasite.

    With evidence for their importance now accumulating, γδ T cells are no longer a scorned minority, and the number of researchers looking at ways of modulating immune responses via the mucosal route is growing sharply. Far from being a backwater of the immune system, “everyone is agreed mucosal immunity is an important field,” says Vassalli.

    For more information on recent immunology research, please see the special section, “Turning the Immune System Off,” which begins on page 237.


    Pinpointing the Source of Intestinal T Cells

    1. Nigel Williams

    Studies of γδ T cells are not only toppling narrow notions about how the immune system defends the body (see main text); they are also challenging the classical view that T cells must pass through the thymus gland to develop properly. Over the past few years, many studies on mice have suggested that some T cells found in the gut, including γδ T cells, can develop even when the mice lack a functional thymus. Just where these cells develop if not in the thymus has been much less clear, however. But a team of Japanese researchers led by Hiromichi Ishikawa of the Keio University School of Medicine in Tokyo may have solved the problem.

    T cell source?

    Intestinal T cells may mature in cryptopatches (light areas).

    SAITO et al.

    Although the gut T cells, like those found elsewhere in the body are born in the bone marrow, the Keio team's work, which is described on page 275, indicates that the cells mature in the cryptopatches, clusters of cells located just under the layer of epithelial cells lining the gut. If so, cryptopatches would give the intestinal lining its own homegrown supply of T cells to combat the large array of pathogens and antigens that an animal constantly ingests.

    Two years ago, the Keio group found that the patches consist primarily of cells that could be immature T cells, a few hundred per patch. The team found, for example, that the cryptopatch cells carry on their surfaces two proteins, one called c-kit and the other the receptor for the immune signaling molecule interleukin-7, that are hallmarks of immature T cells. That suggested that the cryptopatches might be the key site in which stem cells from the bone marrow develop to provide mature intestinal T cells.

    In the current work, the team set out to see whether that is in fact the case. They did this by removing tiny fragments of gut containing the cryptopatches and transferring cells from the gut fragments into the blood system of immunodeficient mice lacking T cells. As predicted, the recipient animals developed mature T cells in their gut. In contrast, when the team injected cells from pieces of lymph nodes removed from the gut region, no T cells developed at epithelial sites in the gut. “It's strong evidence pointing to a role for the cryptopatches,” says immunologist John Klein at the University of Tulsa.


    The Subtle Flirtation of Ultracold Atoms

    1. James Glanz

    If high-energy accelerators make the rap music of physics—with their whirling particles and rapid-fire smashups—then collisions between ultracold atoms are its Wagnerian opera. They are a spectacle of slow atomic duets in which subtle forces and rare internal states hardly ever seen in the everyday world can emerge. Trapped in cages of light and magnetic fields and cooled almost to absolute zero, these atoms can form fragile molecules far larger and more tenuously bound than could survive in our room-temperature world. Their interactions can be fine-tuned with lasers and magnetic fields. And rare, spontaneous changes of the atoms' internal states can take place as a collision unfolds, transforming a gentle encounter into a rollicking escape from the cage.

    In the past few months, physicists have also discovered new links between the cold, slow-motion world of these collisions and other, better known phenomena. At even lower temperatures, clouds of atoms can form a collective state—in effect, a giant atom—called a Bose-Einstein condensate or BEC (Science, 14 July 1995, p. 152; 25 August 1995, p. 1047). The stability of a BEC depends almost entirely on how the atoms making it up interact one-on-one, and ultracold collisions provide a glimpse of these interactions. And in a first-ever detection, a group led by Pierre Pillet of the Laboratoire Aimé Cotton at the Université Paris-Sud has made new strides by getting cesium atoms to form ordinary molecules chilled to 300 millionths of a kelvin—something only seen before with atoms. That breakthrough could be the first step toward studying complex materials in the world of the ultracold.

    “This ultracold collision business is a very exciting thing,” says William Phillips of the National Institute of Standards and Technology (NIST) in Gaithersburg, Maryland, who won the Nobel Prize last year for his part in the development of laser cooling and trapping. “The more people work on it, the more astounding things we're going to learn.”

    Magneto-optical traps slow atoms, cooling them, by bathing them in laser light. The cooling works because photons of light carry momentum, and atoms can absorb photons that have specific frequencies. If a laser's frequency is tuned just below one of those frequencies, an atom will “feel” the light only when it is moving into the beam, as the Doppler or train-whistle effect raises the light's apparent frequency. By combining laser cooling with magnetic fields, which can cause light to trap certain atoms once they are moving slowly enough, researchers can capture cold atoms for many minutes at a stretch.

    These caged atoms can interact in ways never seen in our hotter, faster paced world. Extending ideas first put forth by William Stwalley of the University of Connecticut, Storrs, NIST's Paul Julienne, and others, Phillip Gould, a physicist who is also at Connecticut, focused last year on the attractions that the trap laser can create by distorting the normally symmetric electron clouds around atoms. The excitation creates a charge pattern in one atom that causes it to attract another one over huge distances. This “photoassociation” can begin dragging two atoms slowly together over distances of 100 nanometers—forming, in effect, a nascent diatomic molecule several hundred times larger than could exist at room temperature.

    This acceleration is so gradual, however, that the asymmetry decays in midcollision, after about 30 nanoseconds, causing the excited atom to emit a photon and return to the lower energy, symmetric state (see graphic). “The atoms are still moving toward each other,” says Gould, although now that the attraction has been turned off, they are coasting. Still, that coasting speed can be enough to “spit atoms out of the trap,” ejecting them before the lasers can turn them back, says Stwalley. He adds that this phenomenon “is something you'd like to completely understand” to minimize the leakage of atoms from optical traps.

    Excitation and attraction.

    One laser pulse briefly excites an atom, causing it to be attracted to another atom. A second, probe laser clocks how long it takes for the atoms to converge.


    Gould and Connecticut colleague Steven Gensemer have been measuring the strength of this photoassociative force by first turning on the trap laser to get the atoms moving toward each other, then hitting them with a second, probe laser, which revives the attraction. If the atoms are close together at that point, the result is a collision violent enough to throw the atoms out of the trap immediately. By varying the interval between the pump and the probe lasers and counting the leaked atoms, Gould can gauge how fast the photoassociated atoms were moving together—and how strong and sustained was the original force pulling them together.

    The work, which Gould and Gensemer describe in the 2 February issue of Physical Review Letters (PRL), might help atom trappers turn off the attraction or avoid it. But Pillet is courting it—and creating ultracold ordinary molecules in the process.

    Until now, says Randall Hulet of Rice University in Houston, “nobody's been able to detect ultracold molecules.” Directly cooling and trapping ordinary molecules doesn't work, as the complex internal structure of molecules makes them difficult to cool with lasers. In work now in press at PRL, Pillet and his colleagues Andrea Fioretti, Daniel Comparat, and others succeeded by first photoassociating pairs of ultracold cesium atoms in a trap. Photoassociated molecules usually don't survive long, because the internal energy left over after the atoms come together rips the pair apart. But Pillet says that the cesium molecules rapidly shed this extra energy by emitting a photon, which—a lucky peculiarity of cesium—stabilizes them. At that point, they are close enough to be held by the usual valence bonds, and the union is essentially forever. “That's beautiful,” says Hulet of the experiment. “The ultimate goal would be not only to make ultracold molecules but to trap them” by some other means, he says. That could lead to studies of novel chemical reactions, extremely precise molecular spectroscopy, or even Bose-Einstein condensation of molecules.

    As those fresh directions open up, other researchers have been studying the gentle collisions that are the first step toward a BEC, the quantum-mechanical fluid that forms when the right kinds of atoms are cooled in a sufficiently dense cloud. “It's essential to understand these cold collision processes if you're going to do Bose-Einstein condensation,” says Phillips. He explains that a single number describing the trapped atoms, their “scattering length,” determines much of the behavior of the resulting BEC. The scattering length gauges how two passing atoms interact as they briefly linger in each other's vicinity; the length depends sensitively on the very weakest binding state the atoms could have and still stick together.

    That number—although often virtually impossible to calculate—can be measured by photoassociating a pair of ultracold atoms and then forcing them apart with another laser pulse, as Hulet and many others have shown. Others have recently confirmed theoretical predictions that the length is not immutably fixed for a given atom. A group led by Wolfgang Ketterle of the Massachusetts Institute of Technology reported in the 12 March issue of Nature that they had observed a “Feshbach resonance,” in which applied magnetic fields could make the weakest bound state weaker or stronger, totally altering the character of a sodium BEC. Although that work was done with the BEC fluid itself, Daniel Heinzen of the University of Texas, Austin, and several colleagues have seen a Feshbach resonance more recently in ultracold collisions of rubidium-85 atoms.

    For devotees of this kind of stately atomic opera, the season has just opened. “I suspect that what everybody is going to do is keep their eyes open for really juicy things to measure,” says Phillips.


    Flying by the Seat of Their Halteres

    1. Elizabeth Pennisi

    The fly is the Ferrari of the insect world. It specializes in high-speed maneuvers as it pursues mates or dodges fly swatters, darting around obstacles 10 times faster than a blink of an eye. Now researchers may have uncovered the guidance system responsible for these acrobatics: a pair of Tootsie Pop-shaped organs on the fly's back called halteres.

    The fly responds so rapidly to what it sees that researchers once thought the visual system in its brain must connect directly to its flight muscles. But the new results suggest that this neural hot line may run instead through the halteres, vestigial wings that until now were supposed to function only as gyroscopes that help stabilize the insect in flight. The findings, reported on page 289 by neuroethologist Michael Dickinson of the University of California, Berkeley, and his colleagues suggest that halteres both stabilize the flight and guide its twists and turns. If so, “it's going to [lead] to a rewriting of how flight control works,” says Dickinson.

    This new picture of fly flight control is “interesting and exciting,” says R. Meldrum Robertson, a neurobiologist at Queen's University in Kingston, Ontario, Canada. But beyond that, he adds, the work also provides intriguing insights into how this modern insect's sophisticated flight system evolved from the neural connections in the fly's four-winged ancestor.

    Rear steering.

    Flight control in both tethered (right) and free-flying blowflies may work through the halteres (arrow).


    Researchers have known for decades that the fly steers based on what it sees, rather than, say, by what it smells. To find out more about how it does that, the Dickinson team, building on work done in Germany, studied how visual cues affect wing movements. The researchers tethered flies to a thin rod and surrounded them with a screen displaying moving patterns of dark and light stripes. The flies took off, flying as if moving through a real landscape.

    The team found that the various wing steering muscles respond to the stripes, exhibiting one pattern of activity when the stripes shift sideways and another when they move up and down. Even though this meant, Dickinson says, that “there had to be a way for visual information to affect these steering muscles,” years of electrophysiological studies found no evidence of a direct, functional connection between the visual system and the flight muscles. “It was so disturbing,” he recalls. “I was racking my head to understand what was going on.”

    Then, in about 1994, he came across a paper from 1947 that described in great detail a small set of muscles at the base of the halteres. These muscles are equivalent to the wing's steering muscles, Dickinson says, but they had no known function. Out of curiosity about what they might do, his lab team monitored their electrical activity while moving striped patterns in front of the fly.

    To their surprise, the researchers found that specific muscles stiffened, depending on whether the stripes were moving vertically, sideways, or diagonally. “Very robust visual information [is] going to the halteres,” says Dickinson. To him, this result indicates that the vibrating halteres could be acting as way stations, receiving visual cues and then relaying signals to the wing steering muscles.

    He notes that the halteres do have the necessary neural connections, as researchers had shown in studying their role as gyroscopes. If the fly rotates or starts to tumble while flying, the halteres are deflected by the changes in the forces acting on them. This activates some 300 sensory neurons at the haltere base, which send a strong signal directly to the nerves controlling the wings' steering muscles. The muscles in turn stiffen and alter the wing beat to prevent rolling, pitching, or yawing.

    Dickinson's new scenario proposes that visual information—another fly moving through the air, for example—triggers the same pathway by causing the haltere steering muscles to stiffen. The stiffening activates the sensors at the base of each haltere, and they in turn send a revised message to the wing's steering muscles, one that might translate into “turn left,” as well as “rotate the body 5 degrees.” “It's a different way to think about halteres,” comments Cole Gilbert, a neuroethologist at Cornell University in Ithaca, New York.

    Gilbert's team has additional results supporting this new view. The position of the head relative to the rest of the body also influences the flight path, and his group recently found that sensors in the neck that monitor head position relative to the thorax in the fly relay that information not just to the wings but also to the halteres.

    A dual role for the halteres makes sense from a functional standpoint, Dickinson says. Instead of having to override the stabilizer system in the halteres, which tends to keep a fly flying straight, signals from the brain telling the fly to change course may work through the stabilizer system. As a result, the fly can change course in less than 30 milliseconds, without disabling its gyroscope.

    This setup would also be consistent with evolution's penchant for tinkering with preexisting circuitry, Robertson adds. Most researchers think that the fly's ancestors had two pairs of wings and that during evolution the rear pair degenerated into halteres. This improved the fly's maneuverability by helping it keep from tumbling out of control as it zips about. But the neural connections that originally allowed the two pairs of wings to coordinate their activities may have been retained and refined as this new function evolved. “You can really see how evolution has exploited circuitry in the brain that was there in a much more primitive system to create a system that looks very, very different,” says neurobiologist Heinrich Reichert at the University of Basel in Switzerland.

    Still, Dickinson has yet to prove that the haltere is relaying flight commands to the wing. For one, despite Dickinson's failure to detect input from the fly visual system to the wing steering muscle, neurobiologist Nicholas Strausfeld at the University of Arizona in Tucson has described a physical connection between the visual parts of the brain and the wing steering muscle. The role of that connection is unclear, but its existence means that direct input to the wing muscles can't be ruled out.

    To try to confirm his new ideas, Dickinson hopes to show that the haltere steering muscles do influence the wing steering muscles. And he may not be alone in his efforts. “It's such an interesting idea,” Robertson says, “that now people will go out and try to find evidence for or against it.”


    Ozone Loss, Greenhouse Gases Linked

    1. Richard A. Kerr

    Buffeted by scores of news accounts, the public has sometimes mixed up global warming, which is fueled by heat-trapping gases in the lower atmosphere, and stratospheric ozone destruction, spurred by voracious chlorine compounds like chlorofluorocarbons. Now a computer model suggests there may be a real connection between the two processes. For the first time, a single model has combined greenhouse warming and ozone depletion in a long-term simulation, and the results are sobering: Greenhouse gases and chlorofluorocarbons together may be ganging up to destroy ozone.

    In the wake of global controls on ozone-destroying compounds, most observers expected that the annual Antarctic ozone hole would fade, and the more modest Arctic ozone losses diminish, as atmospheric chlorine declines. But in this week's issue of Nature, a group from NASA's Goddard Institute for Space Studies (GISS) in New York City reports that their model indicates that during the next few decades, greenhouse gases will trigger a springtime ozone hole over the Arctic, much like the one now over the Antarctic.

    Atmospheric researchers are taking these results seriously, although they caution that the model is rather crude. “It's a bold calculation and an important result,” says atmospheric physicist Paul Newman of NASA's Goddard Space Flight Center in Greenbelt, Maryland. “But it's a really soft result,” because the model is relatively simple. Still, the recent behavior of Arctic ozone tends to support the result. Says ozone researcher Ross Salawitch of the Jet Propulsion Laboratory in Pasadena, California: “I think they may be on to something.”

    The link between ozone destruction and greenhouse gases involves temperature, but cooling rather than warming. Stratospheric chlorine-ozone reactions in polar regions are catalyzed by icy cloud crystals that form only in the most extreme cold. The colder it gets and the longer that cold persists into the spring—when the other key ingredient, sunshine, appears—the more ozone will be destroyed. Greenhouse gases warm the lower atmosphere, but they can also cool the polar stratosphere by radiating heat to space and by changing atmospheric heat transport.

    When GISS modelers Drew Shindell, David Rind, and Patrick Lonergan ran their global ocean-atmosphere climate model with ozone chemistry included, they found that projected increases in greenhouse gases progressively chilled the model's wintertime stratospheric temperatures over the poles by 8° to 10°C. Arctic ozone losses, rather than declining in step with decreasing chlorine, worsened until the 2010s and then slowly recovered. During the worst years, up to 65% of Arctic ozone was destroyed, a larger proportion than in the current Antarctic ozone hole. The same process would delay healing of the Antarctic ozone hole, but it would cause minimal additional ozone loss because the hole there is already so severe.

    Researchers are intrigued by the model results but note that the only way a modeler can find enough computer time to project this far into the future “is to use a rather crude model,” as Newman puts it. One concern, for example, is that some other models wouldn't produce as much greenhouse-induced reduction in the heat delivered to the polar stratosphere. Still, even though the GISS model “is fudged up and has a lot of patches,” says stratospheric modeler Jerry D. Mahlman of the Geophysical Fluid Dynamics Laboratory in Princeton, New Jersey, “their punch lines are all plausible.”

    What's more, the GISS model simulation of the 1990s resembles what actually happened—the wintertime Arctic stratosphere has gotten progressively colder and more depleted in ozone. In the spring of 1997, ozone loss was already half as severe as the gloomiest model projection, for around 2015. But as if to show how hard it will be to sort out the effect of greenhouse gases in the real world, Arctic ozone losses this spring dropped to near zero, presumably because of natural climate variability. Given all these uncertainties, the best result of the GISS study, says Newman, may be the enhanced awareness that two great human alterations of the atmosphere—greenhouse warming and ozone depletion—are indeed interdependent.


    Jumbo Gene Offers Clue to Parkinson's

    1. Mutsumi Stone
    1. Mutsumi Stone is a correspondent for Newton Magazine in Washington, D.C.

    People with Parkinson's disease often can't keep their hands from trembling or relax their face to form a smile. For years researchers have known that these hallmark symptoms arise from the slow death of neurons in the substantia nigra, a brain region that pumps out the neurotransmitter dopamine. Now a Japanese research team appears to have uncovered a huge clue to what triggers this neuronal death.

    The team, led by Nobuyoshi Shimizu of Keio University School of Medicine and Yoshikuni Mizuno of Juntendo University School of Medicine, both in Tokyo, reports in the 9 April issue of Nature that it has discovered a massive gene, dubbed parkin, whose mutated version causes a rare inherited form of parkinsonism. Although the finding won't resolve a long-running debate over whether a genetic flaw or some environmental insult is the primary culprit behind most Parkinson's cases, neurobiologists hope that it will offer a glimpse at the shadowy players involved in the molecular wreckage of a patient's brain.

    The Parkin protein's structure links it to protein-degrading machinery in the cell, prompting speculations that a cell with abnormal Parkin may accumulate toxic levels of proteins. “This research is very important,” says Robert Nussbaum of the National Human Genome Research Institute in Bethesda, Maryland, “because we'll want to see if a deficiency of Parkin might be occurring in nonhereditary Parkinson's disease as well,” perhaps as a secondary effect of underlying disease processes.

    To hunt down the new gene, Shimizu, Mizuno, and their colleagues studied Japanese families afflicted with a rare disease called autosomal recessive juvenile parkinsonism (AR-JP), the symptoms of which sometimes start in teenagers. In contrast, in common nonhereditary Parkinson's, symptoms start in people over 40.

    A few years ago, the Japanese group tracked the AR-JP gene to the long arm of chromosome 6. In their latest work, they used positional cloning—a technique in which ever-finer maps of inherited mutations narrow a gene's possible location—to nab parkin. It turned out to be a monster gene. It has more than 500,000 nucleotides, making it the second largest known disease gene after dystrophin, involved in muscular dystrophy. Consistent with parkin's role in AR-JP, the gene is highly active in the substantia nigra. But it may not be functional in patients with AR-JP, because large swathes are missing.

    The team soon discovered that Parkin is “one of the most unique proteins ever known,” says Shimizu. One portion resembles ubiquitin, a protein that, like Charon—the mythical boatman of the river Styx—ferries defective or spent proteins to proteosomes that chop them up. Defects in Parkin's ubiquitin-like section may subvert protein degradation and lead to toxic protein buildup. Parkin also has a zinc finger, a motif often found in proteins that help regulate gene expression.

    The big question now is whether the discovery of parkin will help explain the more common nonhereditary form of Parkinson's. Researchers note that while substantia nigra cells in patients with nonhereditary Parkinson's are littered with aggregated proteins called Lewy bodies, AR-JP patients don't have Lewy bodies. Still, the dissimilarities between the diseases don't rule out an underlying connection. In nonhereditary Parkinson's, Shimizu says, “some environmental factor may be altering the form of Parkin over time.”


    El Niño Brings Winter of Discontent

    1. Melissa Blouin
    1. Melissa Blouin is a science writer in Arkansas.

    At Kitt Peak National Observatory in Arizona, workers had to remove six trash bags full of snow from the top of the 4-meter telescope's dome in December. In Chile, heavy rains recently washed out the road to Cerro Tololo Interamerican Observatory, stranding astronomers on the mountaintop. At Apache Point Observatory in New Mexico, heavy snowfall kept domes closed for days. As El Niño watchers everywhere know, it has been no ordinary winter, and astronomers, who need clear skies and access to mountaintops, felt it more than most scientists.

    True to form, El Niño—a warming of the eastern tropical Pacific—brought cloudy, wet weather to the Southwestern United States and Chile, exactly where astronomical observatories are concentrated. “Southern California, Arizona, New Mexico—these people feel it strongly,” says Nate Mantua, an atmospheric scientist at the Joint Institute for the Study of Atmosphere and Ocean at the University of Washington in Seattle. “If you look at the storm tracks, they all hit here,” adds astronomer Judy Cohen of the California Institute of Technology in Pasadena, which runs Mount Palomar Observatory in southern California. “We're getting dumped on.”

    Some observers at Palomar and Lick Observatories in California say they thought the weather wasn't much worse than in an average winter. But at Cerro Tololo in Chile, which has a reputation for clear skies, observers got to peek at the sky only 11 nights in February—half the usual number. At Kitt Peak, bad weather wiped out 40% of observing time in January, says Phillip Massey, an astronomer there. Because astronomers have to apply for telescope time 6 to 10 months in advance, it isn't always possible to reschedule a night of observing, he adds. And a makeup night may not help. “If your objects are up once a year, it can mean a year's delay.”

    Atsuko Nitta, a graduate student at the University of Texas, Austin, who is studying a variable star, had one clear night in seven when she observed in Chile last July, just after the current El Niño started. She had three clear nights out of eight in March. Because her research requires continuous observations of the star from sites around the globe, “when you don't get data you really suffer,” she says. “We were counting on the data from Chile.”

    Winter is over and the eastern Pacific is cooling, but astronomers aren't yet in the clear. The Climate Prediction Center of the U.S. National Oceanic and Atmospheric Administration predicts unusually high rainfall through April or May in the Southwest. Chile may continue to have wetter than usual weather through July.

    The clouds do have a silver lining: The same weather patterns that dumped clouds and moisture on the continental Americas bypassed Hawaii. Astronomers at the Keck Observatory, which has the largest telescopes in the world, have reported unusually dry, clear weather, with plenty of starry nights.


    Viruses Have Many Ways to Be Unwelcome Guests

    1. Michael Balter

    Lyons, France—It's not easy being a virus. Unwanted and uninvited, these tiny microbes are totally dependent on their hosts for survival. They usually show up with little baggage of their own, often wearing just a thin protein coat wrapped around a small cluster of genes. But if it seems that these lowly visitors hardly merit respect, they are nonetheless taken very seriously for the disease and misery they can cause. And despite their apparent molecular simplicity, viruses have evolved sophisticated strategies to survive and propagate within their hosts.

    At a recent meeting* in this city straddling the Rhône and Saône rivers, some 150 virologists gathered to review current research on the lifestyles of more than 40 viruses that infect organisms from bacteria to humans. “This is the first time in my memory that there has been a meeting that set out to look at as many viruses as possible,” said Timothy Greenland of the Louis Pradel Hospital in Lyons. During the 3-day meeting, researchers reviewed recent work on the many different schemes viruses use to pursue their unwelcome cohabitation with host cells, including slavish dependency, blatant molecular thievery, and occasionally haughty independence.

    Borrowers and lenders. Although viruses form very intimate bonds with their hosts, many contribute only the bare minimum to the relationship. This is particularly true of the RNA-containing retroviruses, which reverse the normal flow of genetic information by using an enzyme called reverse transcriptase to transcribe their single-stranded RNA into double-stranded DNA. Retroviruses, whose genomes are rarely longer than 10,000 nucleotides—the building blocks of RNA and DNA—can produce very few proteins of their own and must beg, borrow, and steal whatever host proteins they need to carry out their life cycle. Stephen Goff of Columbia University in New York City reviewed work that he and his collaborators, including Jeremy Luban of Columbia, have done over the past few years on the hijacking of host proteins by HIV-1, the retrovirus that causes most cases of AIDS.

    Goff's group hunted for specific interactions between virus and host proteins using a technique called the yeast two-hybrid system, in which “bait” and “prey” proteins are fused to separate genetic factors that turn on detectable “reporter” genes when the proteins bind. The team demonstrated that the human enzyme cyclophilin A—which is thought to help cellular proteins fold properly—binds tightly to a specific site on an HIV-1 structural protein called Gag. Goff and other researchers went on to show that the enzyme actually becomes incorporated into the virus structure and that each virus particle contains some 200 molecules of cyclophilin A. Just why HIV-1 needs this host protein is still unclear, although some researchers have speculated that cyclophilin A may help the virus assemble itself properly. But Luban and other investigators have shown that without cyclophilin A, HIV-1 cannot infect its target cells, making the protein's interaction with Gag a possible target for therapies.

    Rather than just borrowing host proteins, as retroviruses do, many larger viruses have pirated host genes and made them a permanent part of their own genetic repertoire, thus acquiring the ability to produce their own versions of proteins that originally came from cells. “Viruses with large genomes have the capacity to incorporate host genes,” says Robin Weiss of the Institute for Cancer Research in London. “They already have 50 or 60 genes, so they can borrow a few more.” One such pilferer, described by Weiss in a talk at the meeting, is Kaposi's sarcoma-associated herpesvirus (KSHV), which was identified 3 years ago by scientists at Columbia University and which many researchers believe is responsible for a type of skin tumor often found in older European men and HIV-infected patients. KSHV's DNA-based genome, which contains at least 140,000 nucleotide base pairs, includes genes coding for myriad humanlike proteins, among them homologs of molecules that regulate cell proliferation, intercellular signaling, and immune functions.

    Exactly what use the virus makes of these proteins is still under investigation, although Weiss points out that another member of the herpesvirus family, cytomegalovirus, is known to produce its own versions of cellular proteins that modulate the immune response (see Article on p. 248). “We are sure these genes do something for the survival of the virus, although they may not be necessary for basic replication.” Weiss adds that various combinations of these pirated genes “may be necessary for viral growth in particular cell types under certain conditions.”

    Another intriguing case of gene piracy was reported at the meeting by Michael Skinner of the Institute for Animal Health in Compton, U.K. In as-yet-unpublished work, Skinner's group has identified several genes in the fowlpox virus, which infects chickens, that are homologous with genes found in a wide variety of other organisms, including yeast, roundworms, and mammals. The pirated genes include some that resemble genes coding for so-called SNAP proteins in mammals and squid, proteins involved in shuttling large molecules around within the cell in bodies called vesicles. Others look like genes that code for proteins of unknown function in the roundworms Caenorhabditis elegans and Trichinella spiralis.

    Skinner told the meeting attendees that because these homologous genes are found in all strains of fowlpox so far studied, they must play an important role in the virus's interactions with its host. For example, he speculated that the viral SNAP proteins might subvert normal cellular transport processes to allow viral assembly. Alternatively, he said, they could interfere with antigen presentation, the process by which infected cells ferry proteins from a foreign invader to the cell surface to alert the immune system.

    But how did all these diverse genes find their way into the fowlpox virus? Skinner hypothesized that retroviruses may have served as go-betweens. At some point in evolution, he suggested, the reverse transcriptase of retroviruses may have copied messenger RNA—an intermediary molecule between DNA and proteins—from a host organism back into the complementary DNA form. This DNA could then have been picked up by the fowlpox virus and incorporated into its genome.

    Going it alone. Most viruses depend heavily on the host cell machinery to get their genomes replicated. Retroviruses can use cellular replication enzymes once they have created DNA copies of their genomes via reverse transcriptase, and DNA-containing viruses such as herpesviruses can use the cellular machine directly. But one group of viruses, those containing double-stranded RNA molecules, have to be more self-sufficient, bringing most of their own replication enzymes into the cell with them. The reason is simple: Host cells do not have the enzymes necessary to replicate double-stranded RNA.

    A talk by B. V. Venkataram Prasad of Baylor College of Medicine in Houston illustrated the lengths to which these viruses must go to reproduce. Over the past few years, Prasad's group has been exploring the replication machinery of the rotavirus, one of the most important causes of severe diarrhea in children, causing more than a million deaths around the world each year. Earlier work by Prasad's team and other researchers had shown that rotaviruses wear three protein coats: an outer garment that is shed when they enter their target cells in the intestinal lining, and two inner layers that shelter 11 separate segments of double-stranded RNA.

    Last year, using a technique called electron cryomicroscopy, which gives a high-resolution picture of frozen, well-preserved viral particles, Prasad and his co-workers demonstrated that the RNA segments are actually embedded in the innermost protein layer, which also contains the enzymes necessary for replication. Unlike most viruses, which come apart after infecting a cell and then go in search of molecular building blocks, double-stranded RNA viruses “stay intact and suck in the nutrients they need to replicate their RNA,” says David Bishop of the Pasteur Institute in Paris. And remarkably, once rotavirus has replicated its double-stranded RNA, Prasad saw single-stranded messenger RNA molecules exiting from the virus via narrow channels through the protein layers. They were presumably on their way to the cell's protein-making machinery, where they would direct the production of proteins needed by new virus particles.

    Given the wide range of strategies viruses have evolved to adapt to life within their hosts, it is no wonder researchers have faced such daunting challenges in coming up with therapies against these unwelcome guests. Says Bishop: “The meeting gave lots of illustrations that, as far as our health and welfare are concerned, viruses are a moving target.”

    • *Strategies in Virus-Host Relationships, Lyons, France, 16–18 February.


    Einstein's Theory Rings True

    1. Andrew Watson
    1. Andrew Watson is a science writer in Norwich, U.K.

    Tshe ancients knew that glass could bend light, but it took the genius of Albert Einstein centuries later to realize that gravity can perform the same trick. In much the same way as a glass lens can focus light, a sufficiently huge mass—say a galaxy—can focus the light from some source far beyond it. “Gravitational lensing” usually takes the form of multiple images of a single distant source. But a team of U.S. and European astronomers has now used three telescopes to pin down an “Einstein ring,” the complete circular image formed when source, gravitational lens, and telescope are in perfect alignment.

    The finding, announced at the U.K. National Astronomy Meeting in St. Andrews last week and published in the 1 April Monthly Notices of the Royal Astronomical Society, “is a clear, textbook example of gravitational lensing at work,” says team member Roger Blandford of the California Institute of Technology in Pasadena. Although Einstein rings have been seen in radio observations, this is “the first time a really complete, unambiguous Einstein ring has been seen in the optical and infrared wavebands,” says team member Neal Jackson, of Britain's Jodrell Bank radio telescope, near Manchester.

    Full circle.

    Gravity bends radio waves from a distant galaxy into an arc, infrared light into a complete ring.


    The group, which also includes Dutch and French astronomers, has been finding and counting gravitational lenses as a way of gauging the “geometry” of space, which depends on its density of mass and background energy (Science, 21 November 1997, p. 1402, and 13 February 1998, p. 981). As a first pass in their lens survey, the team uses the Very Large Array in New Mexico, a system of linked radio telescopes, to spot distant radio-emitting objects that seem to be more than simple bright spots. MERLIN, a six-telescope network centered on Jodrell Bank and spanning 250 kilometers across England to Cambridge, then zooms in for a closer look.

    At radio wavelengths, the new system, dubbed B1938+666, looked like an arc rather than a complete ring, because the radio emissions come from two off-center regions in the source galaxy. But when the researchers took another look with an infrared camera aboard the Hubble Space Telescope, the complete ring was revealed—a dazzling demonstration of Einstein's theory at work.

Stay Connected to Science