News this Week

Science  08 Aug 1997:
Vol. 277, Issue 5327, pp. 763

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


    What Makes Fruit Flies Roam?

    1. Elizabeth Pennisi


    In 1976, Marla Sokolowski, then an undergraduate at the University of Toronto, began noticing that some of the fruit flies she was studying seemed a bit lazy. Some feeding fruit fly larvae would sup a little yeast paste, then wiggle in ever wider circles in search of more food, covering about 5 centimeters in a 10-minute period. But other, seemingly lazy, larvae did little roaming. Through breeding experiments, Sokolowski determined that these “sitter” variants, as she called them, and their more energetic “rover” cousins had different versions of a gene that she called foraging (for). Now, after 2 decades of searching, Sokolowski and her colleagues at York University in Toronto have finally uncovered the for gene.

    On page 834, they report that they have tracked down its location and found that it is actually a gene called dg2, which codes for a protein that helps to relay chemical messages inside cells. Mutations leading to lower-than-normal levels of this protein, one of the enzymes known as cyclic GMP-dependent protein kinases (PKGs), cause the sitter behavior.

    And these mutations aren't just a laboratory curiosity. Shortly after Sokolowski induced the sitter mutation into her lab fly larvae, she found that adult flies, including those in the wild, showed just as pronounced differences in their food-foraging behaviors. That she has now been able to use observed differences in a real-life behavior to track down the molecular underpinnings of that activity “is a first,” comments Bambos Kyriacou, a behavioral geneticist at Leicester University in the United Kingdom.

    The discovery that this “foraging” gene codes for a generic protein also points to a larger lesson. Researchers should not expect to find unique genes for behaviors, says neurogeneticist Ralph Greenspan, now at the Neurosciences Institute in San Diego. After all, a different set of generic messenger proteins, the cyclic AMP-dependent protein kinases, has turned out to play a role in learning and memory in both fruit flies and mammals. Behaviors, too, are likely to be shaped by genes that code for proteins with broad activities, affecting multiple traits.

    “The organism [would be] crazy to use a gene for just one behavior,” agrees Jeff Hall, a behavioral geneticist at Brandeis University in Waltham, Massachusetts. It's much more cost effective, so to speak, to use the same gene in different cells, at different times, and in combination with other genes to generate an organism's behavioral repertoire.

    Because PKG proteins are widely distributed across species, similar genes are likely to be important in shaping behavior in other organisms, perhaps even in people. “This is not something that's going to just apply to insects,” predicts James Truman, a neurobiologist at the University of Washington, Seattle.

    Sokolowski had been slowly homing in on the for gene since 1981. She decided it was worth pursuing, she recalls, after finding that the behavior was linked to one gene and that 70% of the fruit flies she collected from local Toronto orchards were rovers, while the rest qualified as sitters. The persistence of both behaviors in natural populations suggested that each confers some benefit, and although Sokolowski did not know at the time what those benefits might be, she did know that she had a rare opportunity to explore the role of this gene in a behavior critical to the fly's survival.

    As a first step toward finding for, Sokolowski did a series of breeding experiments that tracked the gene to the left arm of fruit fly chromosome 2. Then, to further narrow down for's location, Sokolowski and her colleagues developed a way to introduce an easy-to-trace mutation into the fly's chromosomes, one that was lethal once the fly larvae became pupae. By watching how often the lethal mutation was inherited together with sitter behavior, the group pinned down its quarry in a particular 150-kilobase section of the chromosome, in or near the location where other researchers had found the dg2 gene. But Sokolowski still needed to prove that for and dg2 are the same.

    Help came in 1995. At the time, Kim Kaiser of the University of Glasgow in Scotland had been making mutant fruit flies by allowing a movable piece of DNA called a transposable element to insert randomly into their genomes. In one case, the transposable element had sneaked into the region that Sokolowski was studying. Kaiser offered to send the mutant strain to Sokolowski, who found that they acted like sitters. By using restriction enzymes to dissect that part of the chromosome, she was able to see that the insertion was right in the middle of the dg2 gene, suggesting that for and dg2 are one and the same. Further evidence for that came when Sokolowski's team removed the extra DNA from the gene. The fruit flies “reverted back to rovers,” she says. That showed that “[Sokolowski] had the right gene,” Kyriacou notes.

    But Sokolowski herself was not convinced until her team inserted four extra copies of dg2 into the nuclei of eggs from a sitter fruit fly. The larvae that developed from these eggs began moving around as if they were rovers, the group reports, suggesting that the level of gene expression is what influences the behavior.

    Confirmation came from experiments in which the group teamed up with Greenspan and his colleagues at New York University to analyze the PKG activity in rover and sitter fruit flies and in the larvae with the extra genes. The activity levels were higher in the rovers than sitters, and the transgenic animals had levels similar to that of the rovers. “There was a nice correlation between behavior and the amount of PKG [activity],” says Sokolowski.

    The activity levels didn't differ all that much, however. The sitters still retained about 75% of the PKG activity found in the rover fruit flies. “The enzymatic abnormalities are incredibly mild,” Brandeis's Hall notes. But Greenspan and others say that's what you would expect for an important enzyme like PKG—flies with a serious deficit of the protein are unlikely to survive.

    The researchers do not yet know exactly how the varying levels of PKG activity change foraging behavior, but other work suggests that the PKG signaling pathway affects the excitability of nerve cells. Higher levels of the enzyme could thus cause nerve cells involved in sensing food or in controlling foraging movements to fire more readily, increasing foraging behavior.

    Meanwhile, other recent work by Sokolowski and her colleagues has pointed to why the sitter variant exists in wild populations. In her early experiments she had shown that sitters aren't really lazy. When there is no food, sitter larvae and adults wander just as far as rovers do to look for sustenance. The different behaviors come into play only after the flies have eaten, and results reported by the Sokolowski team in the 8 July issue of the Proceedings of the National Academy of Sciences help explain why.

    Starting with wild populations of fruit flies, Sokolowski and her colleagues put mixtures of rovers and sitters into jars that contained either 50 or 800 to 1200 individuals. The researchers removed extra flies to keep those numbers steady for 73 generations. In that time, the ratio of rovers to sitters increased in the crowded jars and decreased in the uncrowded ones. Those changes make sense, says Sokolowski. Moving takes a lot of energy, so larvae that do not have fierce competition for food do better to move less, while those in crowded conditions gain an edge if they go out searching for food ahead of their jarmates. “It seems empirically that [the sitter behavior] has an adaptive significance,” says Hall.

    However, Sokolowski says that many questions remain to be answered. She still has to pin down the specific cells that control foraging behavior. In addition, the dg2 gene produces three different versions of the enzyme, and it's not yet known whether just one or all three forms of the kinase are involved. Most likely, other, as yet unidentified genes will prove important to foraging behavior as well. But even with these unknowns, Sokolowski's colleagues salute her persistence. “She had the tenacity and whiz to stick with it,” says Hall. “She's to be congratulated.”


    Once, Maybe Still, an Ocean on Europa

    1. Richard A. Kerr

    BOSTON—Does the mightiest ocean in the solar system lie beneath the ice-sheathed surface of Jupiter's moon Europa? Planetary scientists are piling up the evidence, and even the skeptics are now finding it hard to resist. Images of the moon's tortured surface returned by the Galileo spacecraft had persuaded some, but not all, researchers that an ocean tens of kilometers deep lies beneath the ice (Science, 18 April, p. 355). But the evidence was indirect, and skeptics argued that a solid layer of ice slightly warmed by Europa's internal heat could mimic the effects of a deep ocean by slowly churning, reshaping the surface without ever melting.

    Telltale blemish.

    The debris-ringed 140-kilometer impact scar Tyre (in false color) is almost perfectly flat, suggesting it formed on a thin skin of ice floating on water.


    At last week's annual meeting here of the Division for Planetary Sciences of the American Astronomical Society, however, researchers presented new analyses of Europa's surface scars that had even the skeptics agreeing that, either now or in the recent past, an ocean stirred just beneath the icy surface. “It's likely that at the time that the surface we see formed, there was an ocean,” says Robert Pappalardo of Brown University, a Galileo team member and self-described skeptic. That may not have been so long ago, he says, judging from the fresh appearance of the surface.

    A deep pool of seawater on another world is sure to nurture speculations about the possibility of life there. But Pappalardo stresses that Europa's deep waters could have frozen up since they left their traces on the surface: “I'm still a skeptic [about a present-day ocean]; we have to prove it by some direct technique.” That may be more than Galileo can deliver, even when it shifts its orbit about Jupiter late this year to begin an extended mission devoted to Europa.

    Making even the skeptics edge toward tentative belief in an ocean are geologic features such as fractures pulled open in the ice, stress cracks formed from the weight of ridges, and disrupted crustal blocks. The sizes and distributions of these features imply that when the features formed, the brittle outer crust of ice was less than 6 kilometers thick—and perhaps as thin as a few hundred meters, says Pappalardo. Below this thin, brittle crust was presumably a layer of warmer, and therefore ductile, churning ice, he says, which implies that temperatures were rising so sharply from the surface into the interior that liquid water would have been found not much farther down.

    Supporting the idea that the ductile ice beneath the surface gave way in turn to liquid water is a different set of features: 7- to 15-kilometer-wide pits, domes, and spots. James Head of Brown, Pappalardo, and their Galileo teammates argue that heat-driven convection in the ductile ice beneath the surface is responsible for these blemishes. Plumes of warm, rising ice hit the underside of the cold, brittle surface layer, they say, pushing the surface upward to form a dome. Plumes that carried enough heat could have caused some of the surface ice to sublimate away or melt, forming pits and spots.

    If all these features do share a common origin, their spacing—every 5 to 20 kilometers, on average—says something about the thickness of the ductile ice layer. Convection tends to space its rising or falling plumes about as far apart as the depth of the convecting layer, the researchers note. The 20 kilometers of ice inferred from the spacing would leave plenty of room for liquid water, because the subtle gravity variations Galileo detected during its Europa flybys imply that the water or ice layer sheathing the moon's rock core is some 150 kilometers thick.

    Places where large impacts have battered Europa's icy shell also point to a deep ocean beneath. At the meeting, Jeffrey Moore of NASA's Ames Research Center in Mountain View, California, and his colleagues argued that some of the so-called maculae of Europa—huge dark splotches 60 to 140 kilometers across—are debris-encircled craters produced by asteroid or comet impacts. Unlike the high-rimmed, bowl-shaped impact craters that form in solid rock, maculae are flat—the 45-kilometer crater Callanish, for example, is roughened by just 100-meter variations in the height of the surface. The lack of relief implies that Europa's icy surface was weak, and therefore thin, when the meteorite hit; Moore estimates that the ice floated on liquid water just 10 kilometers below.

    All these features look very fresh to geologists, so they have little trouble imagining that the ocean they imply is still there today. But they are reserving final judgment about the age of Europa's surface to a small subgroup of their colleagues—the crater counters. These specialists count the number of impact craters in a given area, estimate the rate at which asteroids and comets have been bombarding the surface—a far more contentious question—and from those two quantities calculate how long it has been since geologic processes last wiped the surface clean of craters.

    So far, the cratering ages have been in conflict. At the meeting, Clark Chapman and his colleagues at the Southwest Research Institute in Boulder, Colorado, offered an age of 1 million to 10 million years, a geologic blink of an eye, for one area that for all the world looks like Arctic ice floes caught in a freshly frozen sea. Such recent resurfacing would persuade even the most skeptical geologist that an ocean is reshaping the surface even now. But Gerhard Neukum of the German Aerospace Research Organization (DLR) in Berlin, like Chapman a Galileo team member, has been advocating a truly ancient age for parts of Europa: 3 billion years. Few would agree with Neukum's pivotal assumption—that a swarm of asteroid impacts billions of years ago, rather than the steadier rain of comets assumed by Chapman, dominates Europa's cratering record—but most geologists take the controversy as a sign that the final age for the surface of Europa is not in.

    Variations in Europa's surface features also make some geologists hesitate to endorse a present-day ocean. While the maculae suggest a thin ice layer, noted Moore, the large crater Pwyll looks more like a classic crater, suggesting the ice there was far thicker at the time of impact. Pappalardo also notes that the fractures and ridges, apparently formed when the ice was thin, seem to have preceded the pits, domes, and spots that formed when a thicker layer of warm ice was convecting. That would suggest that at least in some places, the ice has been thickening in geologically recent times. Europa's interior may be cooling because the store of heat from its formation is dwindling or because of variations in another source of heat, the tidal massaging of the moon by Jupiter's gravity.

    Proving that a global ocean or even local lakes still lie below the Europan ice could take a while. Galileo will take an even closer look during its extended mission starting in December, improving the resolution of its best images from 70 meters to 10 meters. If an even sharper look at Europa fails to convince everyone, the job will probably fall to a future geophysics mission in which a spacecraft would orbit the moon and probe it with ice-penetrating radar, finally plumbing the depths beneath its tantalizing exterior.

    Additional Reading

    1. 218.

    Flickers From Far-Off Planets

    1. James Glanz

    The telescopes of the South African Astronomical Observatory (SAAO) are thousands of kilometers from their hemispheric counterparts in Australia and South America. But two and a half years ago, SAAO's John Menzies joined a project that turned this geographic isolation to advantage. Called PLANET—for Probing Lensing Anomalies Network—the project monitors stars in the Milky Way's crowded central bulge for apparent brightenings that could indicate another, dimmer star with a retinue of planets drifting across the line of sight. Like a set of lenses passing in front of a candle, the system's complicated gravitational field would bend light rays from the bulge star and make it seem to flash and dim.

    Tracking these fluctuations, which would take place over a matter of hours, requires round-the-clock monitoring of the southern sky, where the galactic bulge can be seen. So SAAO's perch between collaborators in Chile and Australia was crucial to the project. And now the lonely watching may have paid off. In recent months, this unusual collaboration has uncovered two strong candidates for companions circling stars thousands of light-years from Earth. The most spectacularly fluctuating event, sketchily described in Internet alerts by PLANET team members as they observed it through June and July, is “exactly what one would expect for a Jupiter-mass planet orbiting around a solar-mass star,” says Abraham Loeb, a theorist at the Harvard-Smithsonian Center for Astrophysics (CfA) in Cambridge, Massachusetts.

    The team still has to complete mathematical modeling of the event to nail down the arrangement of bodies that produced it, says Penny Sackett of the Kapteyn Astronomical Institute in the Netherlands, who, along with Kailash Sahu of the Space Telescope Science Institute in Baltimore, founded PLANET in early 1995. She and Loeb, who is not a member of PLANET, both caution that the system responsible for the flicker could turn out to be, say, a binary star rather than a single star with a planet. But the data themselves are likely to be solid, as another international collaboration called GMAN (Science, 7 March, p. 1416) monitored the same star and saw a similar pattern of brightening and dimming, according to team member David Bennett of the University of Notre Dame in Indiana. If the planet discoveries are confirmed, the technique could take its place as astronomers' most sensitive means of finding planets around other stars.

    The idea of searching for planets by watching for gravitational lensing can be traced to Shude Mao of CfA and Bohdan Paczyński of Princeton University. Paczyński had realized that by scanning large chunks of the sky for stars that gradually brighten over weeks or months, astronomers could say something about how often unseen stars or stellar cinders drift across the lines of sight to those stars. The symmetric gravity of each passing object would act like a single magnifying glass slowly passing in front of a distant streetlight. That insight spawned collaborations including MACHO (Massive Compact Halo Object) and OGLE (Optical Gravitational Lensing Experiment), which are scanning the skies for the gradual “Paczyński curves” in order to estimate how much unseen matter might be swarming through the galaxy in the form of such objects. While such lensing events are rare toward the galaxy's thinly populated outskirts, they are plentiful toward the bulge.

    In a 1991 paper, Mao and Paczyński expanded on the idea. They pointed out that adding a planet around such a star would be akin to spattering water on the magnifying glass, embellishing the more gradual curve with rapid spikes in brightness. Sahu and Sackett realized that well-spaced telescopes of moderate size—a meter or less—could sample lensing events detected by MACHO and OGLE on much finer time scales, searching for planetary anomalies. So they formed PLANET, which now includes telescopes operated by SAAO, Perth Observatory and the University of Tasmania in Australia, and the European Southern Observatory and Cerro Tololo Inter-American Observatory in Chile.

    Each night, PLANET keeps an eye on a handful of events reported by MACHO, immediately stepping up the sampling rate if any of them shows anomalous variations. “When anomalies are detected … e-mail and phone services run very hot,” says John Greenhill of the University of Tasmania.

    The glimmer that galvanized PLANET and GMAN began with a sharp spike around 19 June, followed by a slow rise like a Paczyński curve, and—by some accounts—a strange, double-humped peak around 24 July before a final downturn. Follow-up observations and analysis that could eliminate an ordinary binary-star system as the cause should be completed in a few months, says Sackett.

    The candidate planet circling around this putative dim star is a massive one, like the other extrasolar planets detected so far. But the search method could be sensitive enough to detect planets as small as Earth, unlike techniques that rely on finding a “wobble” in the parent star as a giant planet whirls around it (Science, 30 May, p. 1336). On the downside, it gives just a brief glimpse of the planet, and because of the very gradual convergence of light bent by gravity, any planets it reveals are so far out in the galactic blackness that they can't be studied by any other method. Says Jean Schneider of the Observatoire de Paris-Meudon in France, “I would be frustrated by the impossibility of investigating the planet any further.”

    For now, PLANET observers like SAAO's Menzies are simply enjoying the chase. “It is interesting to speculate,” he says, “that one may be the only person on Earth to be aware, while watching the light curve unfold, of the existence of this other possible world.”


    New Respect for Metal's Role in Ancient Arctic Cultures

    1. Heather Pringle
    1. Heather Pringle is a science writer in Vancouver, British Columbia.

    When English naval officer Sir William Parry was searching for the elusive Northwest Passage in the Canadian High Arctic in 1821, he got a vivid glimpse of the Inuit passion for iron tools. Wherever Parry's party went, they encountered aboriginal groups eager to barter their most cherished possessions for iron nails and hoops; on one occasion, he wrote in his journal, he even witnessed an Inuit woman offer a 4-year-old child in exchange for a metal knife. His experience was shared by others who ventured into Arctic regions in the early 19th century. When British explorer Robert M'Clure's ship Investigator became jammed in the ice off northern Banks Island in 1853, for example, aboriginal families flocked from hundreds of kilometers away to salvage metal for making tools.

    Metal traders.

    Remains of Thule Eskimo winter house, consisting of whale bones and boulders, on Bathurst Island, central Canadian Arctic.


    This ardor for iron, it turns out, was not a new phenomenon sparked by the metal's novelty. Archaeologists have long known that prehistoric Arctic cultures possessed tools crafted of iron scavenged from meteorites and fashioned from native copper. But only recently have they come to realize how widely dispersed and relatively abundant metal objects were in the ancient Arctic. By employing metal detectors in their excavations, searching for rust stains on wooden and bone handles, and studying slots that might have held metal blades, four independent research teams have recently recovered and identified troves of metal artifacts and other clues to its use at five sites in the Canadian Arctic and Greenland.

    Trade routes.

    Metal sources for ancient Arctic cultures: 1. Asian trade metal; 2. Coronation Gulf-Coppermine River native copper; 3. Cape York meteoritic iron; 4. Disko Bay telluric iron; 5. and 6. Norse settlement trade iron.


    The discoveries are giving new insight into the complexity of ancient Arctic society. Metal objects were common at sites hundreds of kilometers from the few known northern sources of copper and iron, implying the existence of elaborate trade networks. And at some sites, the possession of metal objects seems to reflect patterns of social ranking. The desire for metal implements, says Peter Whitridge, an archaeologist at the University of Northern British Columbia in Prince George, both “united distant communities and divided them internally.”

    Researchers have been slow to recognize this brisk commerce in metal and the material's importance in prehistoric Arctic cultures largely because metal objects were so precious that they were rarely left behind for archaeologists to find. The Thule Inuit, who inhabited the Canadian Arctic and Greenland from A.D. 1000 to historic times, and their predecessors, the Dorset people, recycled broken pieces over and over again. While many tool handles bear rust stains, for example, few are found today with metal blades in place, suggesting that their owners thriftily removed the precious iron when an implement cracked or broke. “Metal is a material that you can keep reusing until it is dust practically,” notes veteran Arctic researcher Allen McCartney, a professor of anthropology at the University of Arkansas, Fayetteville. “So it's been hard to find, because if it was big enough for someone to have seen it, they walked off with it.” This realization led McCartney, in an influential paper published 6 years ago, to urge his colleagues to intensify the search for metal with electronic metal detection and other recovery techniques.

    Several groups of researchers took this advice to heart. In the 1994 field season, for example, a team led by James Helmer and Genevieve LeMoine of the University of Calgary in Alberta recorded and confirmed 288 pieces of copper and iron with the help of a metal detector in two Dorset villages on Little Cornwallis Island in the Canadian High Arctic. Of 98 pieces ultimately collected—blades, points, needles, fasteners, and debris unearthed in dwellings dating from A.D. 450 to 1250–53 were made of copper and 45 of iron.

    There's no obvious source of metal anywhere near Little Cornwallis Island, and ancient Arctic peoples did not smelt metal ores, as they lacked both trees for fuel and a tradition of pyrotechnology. But in a paper published in the 1980s, Vagn Buchwald, a chemical engineer at the Technical University of Denmark, gathered data on three major sources of metallic iron that prehistoric Arctic cultures could have hammered into tools: telluric iron that occurs as pea-sized inclusions in iron basalt at Disko Bay in western Greenland, wrought iron from the two Norse colonies founded in Greenland during the late 10th century A.D., and iron from fragments of the massive Cape York meteorite in northwestern Greenland. Other research teams have identified several other possible sources of Arctic metal, including important drift-copper deposits in the Coronation Gulf-Coppermine River region of Canada, and Asiatic metal traded across the Bering Strait.

    Recent analyses tie the newly excavated artifacts to several distant sources. Trace element analysis on 10 of the iron pieces recovered by Helmer's team, for example, showed nickel levels between 6.8% and 7.9%—the same levels found in iron from the Cape York meteorite shower in Greenland, more than 600 kilometers away. Trace element analysis on the copper pieces proved disappointing, however. “The copper contains almost no trace elements,” says Helmer, “so we can't say for a certainty where it comes from.” But Little Cornwallis Island lies more than 800 kilometers from the nearest known copper source. Although little research has yet been done on trade routes in Dorset times, Helmer is intrigued by the findings. “I think the most interesting thing to come from this is the fact that the world view of these people was a lot broader than we often give them credit [for],” he says.

    The metal could have been easily transported by sled and water craft in Thule times, from society to society across the Arctic. Moreover, the relative abundance of metal, says McCartney, dramatically illustrates how closely linked these far-flung societies were. “I think we've got to see the Thule now as major wheelers and dealers in metal,” says McCartney. “I envision a latticework of trade networking that goes on from village to village, group to group, annual fair to annual fair. This stuff is just constantly on the move.”

    The distribution of metal within some individual Arctic settlements also sheds light on ancient social organization. For years, explains Northern British Columbia's Whitridge, researchers relied heavily on ethnographic records of the historic Canadian Inuit for their ideas about Thule social organization and economy. Because the historic Canadian Inuit lived in a small-scale egalitarian band society, many researchers assumed the earlier Thule did, too.

    Whitridge, who has been studying Qariaraqyuk, a large Thule whaling village on Somerset Island in the Canadian Arctic that was inhabited from A.D. 1100 to 1450, now has evidence that this assumption is wrong. He carefully collected all visible traces of metal—including splashes of turquoise color in the excavation units that chemical analysis later proved to be copper—during digs from 1992 to 1994. But the sample sizes were too small to reveal statistically meaningful differences in the abundance of metal in the excavated houses. So Whitridge turned to another indicator of metal use: the widths of blade slots in 194 excavated slotted tools such as knife handles. McCartney had previously pointed out that the ground slate blades favored by the Thule are generally 2 millimeters or greater in thickness, while their metal blades are approximately 1 mm in thickness.

    In research yet to be published, Whitridge found that blade slots at the site fell into two groups, with peaks centered at 1.9 mm and 1.1 mm. Taking 1.5 mm as the dividing point between metal and stone blades, he found that 34% of the slotted tools at the site had once boasted metal blades. Intrigued, he then looked to see if some houses had more of these desirable tools than others. The results revealed some unexpected disparities. In one large dwelling, replete with whaling gear and other exotic goods such as amber, 46.2% of the slotted tools had had metal blades; in a smaller house where most of the gear was for terrestrial hunting, the figure was just 9.6% of the slotted tools.

    Such findings, says Whitridge, suggest that hunters playing key roles in whaling had far greater access to trade metal than did others in the community. “These were people who were appropriating more of the fruits of the whaling harvest [such as whale oil]. And they were converting that into other kinds of wealth to maintain their position in the community,” he says. Indeed, McCartney likens these Thule leaders to the umialiks or whaling captains of the historic North Alaska Eskimos.

    For McCartney, who almost single-handedly sparked this new archaeological interest in Arctic metal, the findings are gratifying. But many questions remain, he notes. One is how the far-flung trade in meteoritic iron got started. Current research suggests that prehistoric Arctic dwellers began hammering the metal into tools about 1200 years ago. McCartney initially wondered whether this expansive trade began soon after the Cape York meteorite shower —estimated at a total of some 200 tons—crashed to Earth.

    Meteorite experts, however, believe the impact date was much earlier. Vagn Buchwald, an authority on the Cape York meteorite shower, says that both the degree of corrosion on the fragments and the absence of a crater suggest that the shower fell more than 2000 years ago, when the region was covered by ice. Indeed, he favors an impact date as early as 10,000 years ago. To McCartney, this raises a tantalizing puzzle: If Arctic hunters first colonized Greenland some 4000 years ago, why did it take them so long to capitalize on such a desirable supply of metal? “That's not a meteorite problem, but it is an archaeological problem,” he adds. Certainly, it's a question for a new generation of Arctic archaeologists.


    How TRAIL Kills Cancer Cells, But Not Normal Cells

    1. Trisha Gura

    Even when battered by heavy-duty chemotherapy and radiation, cancer cells all too often don't give up. In about half of all tumors, this tenacity stems in part from mutations in the cells' p53 tumor suppressor gene, which render it incapable of performing one of its key functions: activating an internal suicide program after the DNA has been damaged by drugs or x-rays.

    Decoying death.

    The decoy receptor DcR1 can't trigger apoptosis on its own and may also interfere with signaling by TRAIL receptors DR4 and DR5.


    About a year and a half ago, however, researchers made a discovery that raised hopes that they might still get such mutated cancer cells to kill themselves: Tumor cells are far more susceptible than healthy ones to another suicide signal, a protein called TRAIL, which acts independently of p53. At the time, though, this hypersensitivity to TRAIL was a mystery. Now three groups, two of them writing in this issue, may have solved the puzzle by showing that normal cells, but not cancerous ones, have a “decoy” protein that can throw TRAIL off the scent.

    On pages 815 and 818, Avi Ashkenazi's group at the biotech firm Genentech Inc., in South San Francisco, and Vishva Dixit of the University of Michigan Medical School in Ann Arbor and his colleagues report the discovery of two new receptor proteins for TRAIL. One transmits TRAIL's message into the cell interior. But the other, although it can bind the death-inducing protein, is incapable of transmitting its signals. A team at Immunex Corp. in Seattle, led by Ray Goodwin and Craig Smith, has come up with similar results that are in press in the Journal of Experimental Medicine. The researchers find this decoy receptor, called either DcR1 (for decoy receptor 1) or TRID (TRAIL receptor without an intracellular domain), in many types of normal human tissues, but not in most cancer cell lines.

    Researchers have long known that viruses help protect themselves against immune attack by cranking out dummy receptors for the molecules that recruit dangerous immune cells. But the TRAIL finding is the first evidence that the body's own cell-bound proteins can serve as decoys against death-inducing proteins. “What is pretty neat is this shows that the body can do it, too,” says Craig Thompson, an apoptosis researcher at the University of Chicago.

    The decoy protein's function is unknown, as is the function of TRAIL itself. But Thompson and others think that the new findings could aid efforts to exploit the TRAIL pathway to kill cancer cells selectively. “This work suggests there might be specific ways to intervene in the survival regulation of cells,” he says.

    Researchers studying programmed cell death, or apoptosis, have been trying to solve the mystery of why cancer cells are so susceptible to TRAIL almost from the time of its discovery in 1995 by Goodwin's team at Immunex and by Ashkenazi and his colleagues. At first, they reasoned that tumor cells are preferentially killed by TRAIL because the cells exclusively express an unknown TRAIL receptor that normal cells lack.

    But when Dixit and his colleagues at Michigan and Human Genome Sciences in Rockville, Maryland, identified the first TRAIL receptor, a protein called DR4 (for death receptor 4), in April of this year, the discovery created more questions than it answered (Science, 4 April, p. 111). Both normal and tumor cells appeared to produce similar amounts of DR4. That sent researchers to the DNA databases looking for sequences indicating other TRAIL receptors.

    Those sequences had to fit one main criterion: They had to resemble a DNA segment that encodes the so-called “death domain,” a stretch of 60 to 80 amino acids found in receptors for other apoptosis-inducing molecules that is essential for their activity. Working independently with two separate databases, Ashkenazi's and Dixit's groups both pulled out a TRAIL receptor sequence, called DR5. Meanwhile, the Immunex team purified the DR5 protein from membrane extracts of TRAIL-sensitive cells. But to everyone's dismay, both normal and cancer cells expressed DR5 messenger RNA, implying that both kinds of cells produced this receptor as well.

    The researchers made another visit to the databases. Instead of searching for the death domain, they simply looked for sequences that resemble the putative TRAIL-binding region of DR4. Almost immediately, the investigators pulled out another gene that turned out to encode the so-called decoy receptor.

    This sequence encodes a protein that contains the external TRAIL-binding region, as well as a stretch of amino acids that anchors the receptor to cell membranes. But the new receptor, DcR1, lacks the intracellular segment needed to spark the cell-death pathway. Most importantly, the receptor is expressed almost exclusively by normal cells.

    Further evidence that DcR1 acts a decoy came when the researchers introduced the gene into tumor cell lines normally killed by TRAIL and found that it dramatically reduced the cells' susceptibility to apoptosis. “It appears that we found a decoy receptor that is highly expressed in normal cells and protects them from the death pathway,” concludes Dixit. However, he and others are at a loss to explain why normal cells carry such a receptor, while tumor cells do not. “It doesn't make sense,” Thompson says. “Tumor cells have evolved to not want to die, so this is really puzzling.”

    That isn't stopping researchers from pondering how the discovery might be parlayed into novel cancer therapies. Researchers have been looking for cell suicide signals that don't involve p53 that might be wielded against cancer cells. One such signal, the apoptosis-inducing protein known as tumor necrosis factor (TNF), triggers severe inflammatory reactions, but TRAIL seems to lack this drawback. “This may be the long-sought-after means to obtain p53-independent cell death using a molecule that doesn't possess the toxicity of previously tried agents like TNF,” says Dixit.

    Researchers at Genentech and Immunex have already begun testing TRAIL in rodents with cancer to see whether it, either alone or in combination with traditional therapies, might thwart tumor growth. So far there have been no signs of toxicity. But animal studies have just begun, and it's too early to tell whether the novel treatment will work or exactly how the new decoy findings will help.

    Indeed, researchers caution that a great deal needs to be learned about TRAIL. “We don't have a clue yet of what TRAIL does biologically,” says Thompson. Still, he adds, “just the sheer complexity of it all says that this is really important. After all, death is an important decision that a cell cannot make twice.”

Stay Connected to Science