# News this Week

Science  05 Sep 1997:
Vol. 277, Issue 5331, pp. 1432
1. GENOME SEQUENCING

# Laboratory Workhorse Decoded

1. Elizabeth Pennisi

The genome of Escherichia coli, a research favorite and common pathogen, has finally been finished—twice—allowing researchers to tour the organism's molecular workings

It's rare for Science to devote 10 pages to a single scientific achievement. It's even more extraordinary for the journal's editors to accept an article presenting data that had been released months or even years earlier. But when the article marks the end of a 15-year race to determine the sequence of all 4.6 million base pairs in the genome of the bacterium Escherichia coli, exceptions can be made.

The E. coli genome, described on page 1453 by geneticist Frederick Blattner of the University of Wisconsin, Madison, and his colleagues, is neither the first nor the largest to be sequenced. But because of the rich history of the organism itself, it stands out among the dozen or so genomes now in the public record. E. coli “is the organism we have the most information about genetically and metabolically,” says Francis Collins, director of the National Human Genome Research Institute (NHGRI) in Bethesda, Maryland.

For decades, microbiologists, biochemists, geneticists, and cell biologists have studied this easy-to-grow organism. More recently, a pathogenic strain in tainted meat has made E. coli a household word. Now, researchers have the full set of genetic instructions underlying the organism's structure and function. Already, partial E. coli sequences that Blattner and, independently, a Japanese team have been depositing in the public database have given researchers a taste of how valuable such data can be. Matches between new genes from other species, including mammals, and E. coli sequences previously filed have often provided researchers with a name and function for their discoveries.

Now that the genome is completed, there are more genes, known and unknown, to work with. E. coli has 4288 genes, about 40% of which are complete mysteries. With this genomic overview, researchers can begin to form a coherent picture encompassing all of E. coli's biology. “Having this complete set of instructions gets us one step closer to understanding how a free organism functions,” Collins points out. Moreover, comparing this genome to other microbial genomes that are rapidly being finished (see sidebar) should help explain how these organisms have evolved.

## Long haul

Blattner first thought about sequencing the E. coli genome in 1983 after tallying the genes that had already been sequenced. Together they totaled 2.3 million base pairs—half the length of E. coli's single chromosome. “I sort of convinced myself that [sequencing E. coli] was a feasible idea” and proposed doing it in a Science editorial, he recalls.

E. coli was the obvious choice for a sequencing effort, says Frederick Neidhardt, a microbiologist at the University of Michigan, Ann Arbor. “By the late '50s and '60s, so much information had been learned about metabolism by using E. coli that it meant there was an advantage for all kinds of people to study it,” he explains. “Your work had a better chance of being interpretable.” Figuring out the microbe's genetic code would help integrate all those years of study, Blattner reasoned.

Other efforts, including those that produced the complete sequences of the genomes of the yeast Saccharomyces cerevisiae and the bacterium Bacillus subtilis, have involved multiple labs on two or more continents. But Blattner decided to take on the E. coli genome on his own, a move that ultimately led to a race between his team and Japanese sequencers.

As a first step toward that goal, his lab began breaking the E. coli chromosome into small pieces that could be used to build a physical map of genetic landmarks along the DNA. He was about 85% done when a Japanese group beat him to the punch, however. In 1987, Yuji Kohara of Nagoya University and his colleagues published its complete set of clones, along with a detailed physical map. Disappointed, Blattner quickly finished off his own map, determined to beat the Japanese in sequencing the genome. The race had begun.

But the next few years were rough on both sides of the Pacific. “The E. coli genome project languished as much in Japan as it did in the United States, [held up] in part by the funding and in part by the technology,” notes Craig Venter, president of The Institute for Genomic Research (TIGR) in Rockville, Maryland. At the time, automated sequencing was in its infancy, and the existing methods meant sequencing was as much an art as a tested technology. “The equipment was so primitive that the speed was slower and the accuracy was lower than now,” says Takashi Horiuchi of the National Institute for Basic Biology in Okazaki, Japan.

Indeed, the first Japanese sequencing attempt resulted in 186 kilobases, but so much was inaccurate that eventually both the Japanese team and Blattner's would resequence that section. A second attempt by a different team did little better, and not until 1995 did Horiuchi and his colleagues come up with an effective sequencing strategy. By having each participating laboratory do a different step, “we improved our efficiency,” says Hirotada Mori, a molecular biologist at the Nara Institute of Science and Technology who compiled the incoming sequence data. That enabled the Japanese team to complete 2.6 megabases within a year.

Meanwhile, Blattner was having problems of his own. In 1990, he had received one of the first grants awarded by NHGRI's predecessor, the Human Genome Center. By the end of 1994, his team had completed 1.4 million bases—a significant chunk, but far short of the full genome he had promised to complete by then. As a result, his grant wasn't renewed (Science, 13 January 1995, p. 172). Only after the E. coli community protested (Science, 31 March 1995, p. 1899) and he reapplied and won new funding was he able to proceed.

Blattner was unaware of the progress the Japanese were making during this time until a September 1996 meeting in California, when both his team and theirs reported that they were about to make a final push to complete their work. The news precipitated “a hectic mad dash,” recalls Guy Plunkett III, who had been with the Blattner team almost from the beginning. “To not even be the first to finish E. coli would have been heartbreaking.” They did finish first, but not by much.

On 16 January, Blattner's group made its final deposit of sequence into GenBank, having completed 2.6 megabases in the previous 12 months. The Japanese team submitted its fully assembled genome on 23 January. It was the seventh genome to be made public and the largest to date for a prokaryote.

Although both genomes come from the same strain, K-12, of E. coli, they are not quite identical because the Japanese and U.S. groups used two different versions, or isolates. For example, there's a 496-kilobase-long stretch that's inverted in the Japanese isolate. Also, the completed genome from Japan includes sequences submitted by other groups, including Blattner's, and therefore it is a mosaic of isolates. Blattner, known for his meticulousness, had insisted that his team do the whole genome themselves, even to the point of redoing E. coli DNA already represented in GenBank. “What we wanted to present was a sequence from one particular isolate,” says Plunkett. In this way, they were assured that all the genes in this genome actually work together to create a functioning organism.

## Genomic revelations

By providing a panorama of the microbe's genome, the sequence is allowing researchers to spot new details. Indeed, describing those features became an important component of his effort, says Blattner. With computer programs, the sequencers first identified the main landscape feature—the open reading frames, which are long stretches of bases that code for amino acids and likely represent genes. This revealed that E. coli has some 4288 potential genes, encompassing about 88% of its DNA. A search through GenBank showed that 1827 of E. coli's genes had already been characterized.

To piece together the identities of the rest, Blattner, Monica Riley of the Marine Biological Laboratory in Woods Hole, Massachusetts, Julio Collado-Vides of the University of Mexico in Cuernavaca, and their colleagues compared the E. coli sequences to those of known genes from other organisms or identified genes based on their neighbors. For example, biochemists had pinned down just two of the six genes needed to make a set of enzymes that E. coli uses to break down aromatic compounds. Because the genes needed for a particular metabolic pathway in microbes are often located together, his team found the rest by looking next to one of those genes. Those types of analyses enabled the Wisconsin team to identify hundreds of additional genes. “We were discovering new things every day,” says Blattner.

For each gene newly deciphered, however, there remains another with an unknown function—some 40% of the total. “It shows how little we know, even with something like E. coli, about the biology [of these organisms],” says Venter. Adds Plunkett: “There are all these things waiting to be discovered.

## Global resource

Genome enthusiasts expect those discoveries to start pouring in, in large part because of the experiments made possible by knowing E. coli's genetic code. “There are a lot of experiments you can do that you couldn't do before,” says Richard Roberts, a molecular biologist at New England Biolabs in Boston.

By examining the genome sequence, for example, Kenneth Rudd, a bacterial geneticist at the University of Miami School of Medicine, can determine whether the small proteins that he finds in E. coli are true proteins with possible functions of their own, or merely the breakdown products of larger proteins. “Without the complete genome sequence, we would just be waving our hands,” he notes.

Other researchers are using the genome to guide them to proteins crucial to E. coli metabolism. Microbiologist Tyrrell Conway of Ohio State University in Columbus says that searching for genes known to be involved in sugar acid metabolism now takes “a matter of days” instead of months, enabling him to focus his efforts on assessing the function of that gene's protein product. In addition, when examining a cluster of E. coli genes known to control sugar metabolism, he homed in on another sugar acid pathway. Among those genes, he noticed an extra gene that had no known function. It turned out to code for an enzyme that initiates an alternative pathway that researchers had never known about before. “[The genome] will be a tool to really accelerate our understanding of how this organism makes a living,” he predicts.

By comparing the E. coli genes to those of other organisms, the Blattner group also spotted clues to how particular genes originated. In some cases, the results were expected: Genes for the proteins that make up the whiplike flagella that microbes use for locomotion are alike in both E. coli and its close relative, Salmonella, suggesting they evolved long ago in a common ancestor. Other comparisons turned up surprises. For example, E. coli turns out to have not one, but two sets of enzymes for breaking down aromatic compounds, and the genes for the newly found second set are almost identical to the equivalent set in the soil microbe Pseudomonas, which is not closely related to E. coli. The question remains whether the genes sneaked into E. coli's genome somehow from that soil bacterium.

Databases and computer programs already being developed to store and analyze the E. coli data and compare its genome with those of other organisms will speed such efforts. Rudd, notes, however, that researchers will have to follow up on the computer analyses in the laboratory, to verify that a gene does what the analyses say it does. He also worries that the E. coli community is not as well positioned to complete this next phase of the work as are the biologists interested in B. subtilis and yeast. They learned how to cooperate and coordinate efforts while sequencing the genomes of their organisms and can now put this lesson to use as they work to identify all the proteins encoded by those genomes.

Because the E. coli genome was sequenced by two competing groups, no such infrastructure exists for completing the next step. “It shouldn't be that everyone is knocking out the [same] 50 most exciting genes” to find out what they do, Rudd says.

Efforts are under way, however, to improve communications among E. coli researchers, if not coordinate their efforts. The half dozen or so E. coli-related sites on the World Wide Web link to one another, and Blattner's E. coli Web page, once completed, will likely be a cyberspace meeting place for researchers studying the microbe. And even if a full-blown, coordinated attack on the inner workings of E. coli never develops, the genome will still hold its value, says TIGR's Venter. “It's scientific information that will be used for centuries.”

2. GENOME SEQUENCING

# Microbial Genomes Come Tumbling In

1. Elizabeth Pennisi

It's an avalanche,” says Daniel Drell, project manager for microbial genomes at the U.S. Department of Energy in Germantown, Maryland. Barely 26 months ago, researchers at The Institute for Genomic Research (TIGR) in Rockville, Maryland, submitted the first complete genomic sequence of a bacterium, Haemophilus influenzae, to GenBank, a public database run by the National Library of Medicine. Since then, 10 more bacterial genomes, plus that of the eukaryote yeast, have been completed, and at least a half dozen more should be finished by year's end. And the avalanche is just beginning. All told, some 50 microbial genome projects are under way, with more in the planning stages.

These sequencing efforts are already changing the way microbiologists do their work. By comparing the genes of these many microbes, researchers can gain insights into how they evolved and which genes are likely to be essential for particular functions. “Having genomes to compare makes it easier to draw conclusions about how [a microbe] lives,” explains Anthony Kerlavage, a computational biologist at TIGR. That information in turn can help pharmaceutical companies search for new antimicrobial medicines.

Indeed, the surge in interest in microbial genomes stems from the realization of just how much they have to offer. In the past, “the human genome was seen as the big prize,” explains sequencer Bart Barrell of the Sanger Centre in Hinxton, United Kingdom. “But once TIGR had sequenced Haemophilus, people realized it was possible and that immense amounts of information could be gotten quickly [from microbial genomes].”

Some of the sequencing efforts have been aimed at laboratory workhorses, such as Escherichia coli, whose genome appears in this issue (see main text), and Bacillus subtilis, the complete genome of which was reported at a July meeting. Because these organisms have already been so well studied, knowledge of their complete genome sequences will be particularly useful to researchers trying to link genes to function. The other genomes finished to date provide a diverse representation of the microbial world, including archaea that live in extreme environments. Together these sequences will enable researchers to sort out evolutionary relationships.

View this table:

But the potential usefulness of their genome sequences has won pathogens the lion's share of attention. Haemophilus causes ear infections and meningitis, for example. And just last month, TIGR's Jean-François Tomb and colleagues described the genome of Helicobacter pylori, a bacterium responsible for peptic ulcers and linked to stomach cancer. (The results appeared in the 7 August issue of Nature.) Because H. pylori is specialized for living in the stomach's acid environment, “the standard [antibiotic] treatments don't work too well,” notes molecular geneticist Antoine Danchin of the Pasteur Institute in France. Consequently, he says, having its genome is critical for finding new surface proteins that might be vaccine candidates.

Also completed recently is the genome of Borrelia burgdorferi, the spirochete responsible for Lyme disease, which TIGR's Claire Fraser put up on the TIGR World Wide Web site on 18 July. And the list will not stop there. This year alone, TIGR expects to sequence as much microbial DNA as is in the entire E. coli genome, which took 6 years to finish. That DNA includes partial genomes of Mycobacterium tuberculosis, which causes tuberculosis, of the malarial parasite Plasmodium falciparum, and of Salmonella and other important human pathogens.

The Sanger Centre, meanwhile, expects to finish the sequence of a laboratory strain of M. tuberculosis by the end of the year, and it is also sequencing the pathogenic bacteria that cause leprosy and meningitis. In addition, Sanger just expanded a pilot project aimed at sequencing the malarial parasite to a full-fledged, $12.8 million effort, in which the center will complete 15 megabases—one-half the malarial genome—in the next 3 years, says Barrell. And just last week, the U.K. Biotechnology and Biological Sciences Research Council announced that it will support the sequencing by Sanger of the genome of a Streptomyces bacterium, which is not a pathogen but an important source of antibiotics. “The goal is to establish a pathogen-sequencing facility that would be able to sequence 25 megabases a year,” says Barrell. That's good news for microbiologists wanting to compare genomes. “Anything we can find out about one genome starts feeding back to the other organisms,” says Guy Plunket III, a bacteriologist at the University of Wisconsin, Madison. Soon, for example, he and his colleagues will be able to compare the virulent mycobacterium strain being sequenced by TIGR with the lab strain being sequenced by Sanger—to find out what makes one so much more deadly. TIGR's Fraser can't wait to have the genomes of several human pathogens in hand to look for unusual genes common to them all. “If we get a big enough set, we can begin to see if there are sets of genes responsible for the [pathogens'] specificity for the human host,” she notes. The protein products of such genes might make particularly good vaccine or drug targets. All of which puts a heavy burden on bioinformatics specialists who are developing the tools for making useful comparisons. So far, says Drell, “nobody is equipped to handle all the knowledge that's coming out of these sequencing projects.” But Kerlavage says he welcomes the challenge, because “we're not working in a vacuum anymore.” 3. DEVELOPMENTAL BIOLOGY # Haeckel's Embryos: Fraud Rediscovered 1. Elizabeth Pennisi Generations of biology students may have been misled by a famous set of drawings of embryos published 123 years ago by the German biologist Ernst Haeckel. They show vertebrate embryos of different animals passing through identical stages of development. But the impression they give, that the embryos are exactly alike, is wrong, says Michael Richardson, an embryologist at St. George's Hospital Medical School in London. He hopes once and for all to discredit Haeckel's work, first found to be flawed more than a century ago. Richardson had long held doubts about Haeckel's drawings because they didn't square with his understanding of the rates at which fish, reptiles, birds, and mammals develop their distinctive features. So he and his colleagues did their own comparative study, reexamining and photographing embryos roughly matched by species and age with those Haeckel drew. Lo and behold, the embryos “often looked surprisingly different,” Richardson reports in the August issue of Anatomy and Embryology. One striking deviation from reality, Richardson says, appears in Haeckel's drawings of embryos in the “tail bud” stage, which he depicted as identical for different species. While real embryos do share many features at this stage, such as a tail and identifiable body segments, they also have key differences. Human embryos, for example, have tiny protrusions called limb buds, says Richardson, particularly if they have developed to the point of having as many body segments as Haeckel gives them. But Haeckel did not include limb buds. And in his drawings, the chick embryo eye is blackened, like a mammal's, “but it wouldn't be pigmented this early,” Richardson says. He adds that Haeckel has given the bird embryo a curl in the tail that resembles a human's. Not only did Haeckel add or omit features, Richardson and his colleagues report, but he also fudged the scale to exaggerate similarities among species, even when there were 10-fold differences in size. Haeckel further blurred differences by neglecting to name the species in most cases, as if one representative was accurate for an entire group of animals. In reality, Richardson and his colleagues note, even closely related embryos such as those of fish vary quite a bit in their appearance and developmental pathway. “It looks like it's turning out to be one of the most famous fakes in biology,” Richardson concludes. This news might not have been so shocking to Haeckel's peers in Germany a century ago: They got Haeckel to admit that he relied on memory and used artistic license in preparing his drawings, says Scott Gilbert, a developmental biologist at Swarthmore College in Pennsylvania. But Haeckel's confession got lost after his drawings were subsequently used in a 1901 book called Darwin and After Darwin and reproduced widely in English-language biology texts. The flaws in Haeckel's work have resurfaced now in part because recent discoveries showing that many species share developmental genes have renewed interest in comparative developmental biology. And while some researchers—following Haeckel's lead—like to emphasize the similarities among species, Richardson thinks studying the contrasts may be more interesting. Gilbert agrees: “There is more variation [in vertebrate embryos] than had been assumed.” For that reason, he adds, “the Richardson paper does a great service to developmental biology.” 4. GEOSCIENCE # Ring Laser Senses Earth's Spin 1. Alexander Hellemans 1. Alexander Hellemans is a writer in Naples, Italy. The sun will surely rise tomorrow, but it may not rise exactly on schedule. Because tides, atmospheric changes, and perhaps movements in Earth's core are constantly shifting the planet's mass, the time it takes to make a full turn fluctuates by a few milliseconds per day. And milliseconds matter, at least to researchers trying to understand the fluctuations and people wanting to get a fix on their exact position by comparing precise time signals transmitted from global-positioning system satellites. Now a square block of glass 1.2 meters across in a cave in New Zealand may help keep track of Earth's vagaries. The device is the world's largest ring laser gyroscope, similar to the laser gyros used in the navigation systems of aircraft but very much more accurate. “We believe it is the most stable gyro that has ever been produced,” says theoretical physicist Geoffrey Stedman of the University of Canterbury in Christchurch, New Zealand. Built by scientists from the Federal Office for Cartography and Geodesy in Frankfurt, Germany, the Technical University of Munich, and the University of Canterbury, the$1 million gyro, dubbed C-II, should be up and running by late fall. When it reaches its peak sensitivity, it should be able to detect changes as tiny as a few milliseconds per day.

Earth scientists can already detect such minute fluctuations using very long baseline interferometry, in which two or more radio telescopes separated by thousands of kilometers monitor slight changes in the apparent position of beacons in the distant universe, called quasars. But analyzing the measurements involves sending tapes to a processing center where readings from the telescopes are compared on a single computer. A faster alternative is satellite tracking, in which the movement of a satellite relative to the spinning Earth is tracked by reflecting laser beams off it.

The laser gyro promises to be faster still, and just as accurate. The gyro consists of a glass block 1.2 meters on a side, with a square channel a meter on a side bored through it. The channel is filled with helium and neon gas. Researchers excite the gas with radio waves, and the whole device acts as the cavity of a helium-neon gas laser, producing two coherent beams, traveling in opposite directions. Mirrors at each corner guide the beam around the ring.

The two counterrotating beams interfere with each other to produce a so-called standing wave—a fixed pattern of several million bright and dark spots, called nodes and valleys, crammed into the ring. “This standing wave is standing still in absolute space,” says Hans Bilger of Oklahoma State University in Stillwater, a ring laser expert who helped design C-II, but the wave moves relative to the gyro. “If you rotate the physical housing of this standing wave, then this housing rotates around the standing wave, and if you look through a mirror at the standing wave, then you simply see valleys and nodes walking by you,” he says.

Earth's rotation has exactly that effect, causing 79 nodes of the standing wave to pass by any point on the gyro every second. To detect changes in Earth's rate of rotation, the researchers must measure variations in this node rate to an accuracy of one in 10 million. “We are getting toward parts per million, and I believe we shall be there in a few weeks' time,” says Stedman. Ultimately, he says, “we hope it will be able to pick up something like lunar tides.” That would require a sensitivity of four parts in 100 million, he says, which he calls “quite a tall demand.”

For a start, any variations in temperature would prove disastrous because thermal expansion in the glass could change the laser's path length and thus its frequency. So the researchers installed the instrument in an artificial cave 30 meters underground, where the temperature is stable to a few hundredths of a degree. They are also stabilizing the size of the ring using “adaptive optics” techniques. The researchers constantly compare the ring laser frequency to an extremely stable reference laser, and correct for any discrepancies by bending one of the mirrors with an actuator.

Already, though, the ring is picking up the rotational component from the seismic waves of nearby earthquakes. And even if C-II does not reach the sensitivities needed to explore the mechanisms affecting Earth's rotation, its successor might. The team is seeking funds for a larger, 4-meter ring laser that they hope to install in a cave to be built in Wettzell, Germany, around 2002.

5. EVOLUTION

# 'Living Fossil' Fish Is Dethroned

About 370 million years ago, a restless faction of the fishes traded in their fins for feet and set out to colonize land. Biologists have debated for decades exactly which members of the fish family made this bold move—and therefore which of their descendants are our closest living gilled relatives. Now it seems that the popular favorite, the “living fossil” known as the coelacanth, is out of the running.

That's the tentative conclusion reached by two researchers who have completed the most comprehensive survey to date of coelacanth mitochondrial DNA (mtDNA). Mitochondria, organelles that serve as power plants in all higher cells, carry their own small complement of genes that mutates over evolutionary time, enabling scientists to infer how long any two species have been diverging by comparing their mtDNA. In this month's issue of the German journal Naturwissenschaften, geneticists Axel Meyer of the University of Konstanz in Germany and Rafael Zardoya of the Museo Nacional de Ciencias Naturales in Madrid, Spain, report that the mtDNA of lungfish—an ancient class of air-breathing fish found in Africa, Australia, and South America—is closer than that of the coelacanth to the mtDNA of land animals such as frogs.

That's “an interesting piece of information,” says S. Blair Hedges, an evolutionary biologist at Pennsylvania State University in University Park. He explains that knowing which extant fish is closest to the first terrestrial tetrapods, or four-legged creatures, might tell biologists which key anatomical innovations enabled our fishlike ancestors to conquer the land. “It helps reconstruct what the organisms looked like at that time, and maybe what environmental factors may have been involved,” says Hedges, who published a study in 1993—based on several mtDNA sequences—that also pointed toward the lungfish.

Paleontologists of the 19th and early 20th centuries knew coelacanths only from the fossil record, but that was enough to convince them that the unattractive creatures, with lobed fins that resemble primitive tetrapod limbs, were close relatives of the first land animals. Then, in 1938, anglers off the Comoro Islands in the Indian Ocean stunned the scientific world by catching a live coelacanth, the first of many. The discovery caused such a sensation, says Meyer, that the coelacanth-tetrapod connection “is still the predominant textbook dogma. It has to do to some degree with the romance of it.”

In the 1980s, paleontologists began finding hints that the dogma might be wrong. For one thing, features of fossil and living lungfish such as their external nasal openings—important for any animal that needs to breathe and chew at the same time—pointed to lungfish, not coelacanths, as the closest sister group to the tetrapods. At the same time, molecular biologists such as the late Allan Wilson at the University of California, Berkeley, had begun to examine the evolutionary relationships of species by comparing similar fragments of their mitochondrial genes, which are often simpler and easier to analyze than nuclear genes. That allowed Wilson and Meyer to announce in a 1990 paper that tetrapods arose from the branch of the evolutionary tree leading to the lungfish, not the coelacanth. Later, Hedges and two colleagues reported similar findings.

As researchers sequenced more coelacanth mtDNA, however, the creature edged back into contention. In the July issue of Genetics, for example, Meyer and Zardoya reported that a statistical comparison using the complete coelacanth mtDNA sequence didn't point unambiguously to either lungfish or coelacanths as the tetrapods' closest sister group. As Hedges points out, however, mtDNA may mutate at different rates in different lineages, sometimes resulting in phylogenetic trees that contain “highly significant but wrong” groupings. Indeed, when Meyer and Zardoya reanalyzed their data for their latest study, they concluded that they could “clearly reject” the possibility that coelacanths are the closest sister group to tetrapods. (The possibility that coelacanths and lungfish are equally close relations of tetrapods, although unlikely, could not be formally ruled out.)

That means that traits seen in the lungfish, such as external nostrils and modifications in the circulatory system and blood chemistry, may well provide the best clues to what the earliest land animals looked like. But to settle the issue once and for all, says Meyer, biologists will need to examine the more complex nuclear genes of coelacanths and lungfish. “It's an important question,” Meyer says, “and of course I would like to be the one to answer it.”

6. INFECTIOUS DISEASES

# Malaria Fighters Gather at Site of Early Victory

1. Pallava Bagla
1. Pallava Bagla is a science writer in New Delhi.

HYDERABAD, INDIA—One hundred years ago, Britain's Sir Ronald Ross discovered that the malaria parasite is transmitted by the Anopheles mosquito. To honor that historic work, done in a lab outside this city, and to assess the state of the war against the disease, some 700 scientists from 30 countries gathered here on 18 to 22 August for the Second Global Meeting on Parasitic Diseases.

## Malaria and AIDS

In addition to its staggering political and economic problems, sub-Saharan Africa is also home to two of the world's most dreaded infectious diseases, AIDS and malaria. Now researchers are gathering evidence in both field and lab studies suggesting that those two killers may be stalking the continent in tandem.

In vitro studies conducted at the U.S. Centers for Disease Control and Prevention (CDC) in Atlanta show that malaria infection, by stimulating the white blood cells that are targeted by HIV, can speed up replication of the virus. Recent epidemiological studies in Africa also hint that malaria may be a deadly cofactor for HIV, hastening the death of those who carry both infections. But more field data are needed to draw a complete picture of the interaction. “In vitro evidence suggests a link between malaria and HIV, and now we need to find evidence for it in vivo,” says Bernard Nahleen, a medical epidemiologist and director of the CDC's station in Kenya.

The laboratory evidence comes from studies by Altaf Lal, chief of CDC's molecular vaccine section. He found that when malaria antigens activate CD4 T lymphocytes—the white blood cells HIV infects—and stimulate the production of the immune system messengers called cytokines, HIV replication increases sharply. The result was a 30- to 100-fold increase in HIV load in activated CD4 cells, he reported at the meeting.

That kind of synergy, if it takes place in infected people, could have deadly effects in Africa. Recent estimates from the World Health Organization, for example, note that about 60% of all cases of HIV—some 13.3 million—occur in sub-Saharan Africa, which is also home to almost 90% of the nearly half-billion cases of malaria reported globally each year. In addition, epidemiological data suggest that the HIV-malaria connection may be real.

Studies in Malawi and Kenya, for example, show that infants born to HIV-positive and malaria-positive mothers have four times the mortality rate of those whose mothers suffer from a single infection. That's because the maternal antigens that protect them against malaria in the first year also serve as cofactors for the HIV. “From a clinical point of view, HIV-infected individuals deteriorate faster with every bout of malaria,” says Subhash Hira, a professor of infectious diseases in the Health Sciences Center at the University of Texas, Houston. What's more, HIV infection in sub-Saharan Africa progresses to full-blown AIDS in roughly 5 years—about half the time it takes in the West. “This sudden acceleration of the disease could be due to the rampant presence of malaria in the region,” speculates Lal, who hopes to do longitudinal studies on humans at the CDC's Kenya station.

“The implications of the HIV-malaria connection are very serious,” says John LaMontagne of the U.S. National Institute of Allergy and Infectious Diseases, in Bethesda, Maryland. In an ironic side effect, a demonstrated link may also boost the willingness of organizations to raise their investment in malaria research, says Louis Molineaux, former chief of operational research for the WHO's tropical disease control division. “AIDS scares the rich and malaria does not,” says Molineaux. “So if there is a connection, more money could come in for research on malaria.”

## Itching for a Solution to Drug Resistance

A one-time magic bullet against malaria may be ready to make a comeback, if some suggestive results in Africa hold up. Field trials by researchers at the University of Ibadan in Nigeria suggest that the combination of chloroquine and various anti-itching drugs could make the former wonder treatment work again.

Chloroquine is the preferred drug for treating malaria because of its low cost and easy availability. But parasites resistant to the drug have spread across Africa in recent years and reduced its effectiveness as both a treatment and a prophylactic, forcing doctors to rely on more expensive compounds.

A widely reported side effect of chloroquine in Africa, however, yielded a hint about how to restore its potency. Many patients complained of intense itching; to control it, health workers routinely prescribed various antihistamines. A few years ago, Ayodae M. J. Oduola, a physician at the University of Ibadan, found that chloroquine seemed to be working in areas where it should have failed when it was prescribed with chlorpheniramine, a cheap, easily available antihistamine.

Since then, Oduola and A. Sowunmi of Ibadan's Malaria Research Group have done four trials involving 400 patients with acute uncomplicated malaria; half were treated with the combination drugs. In an area with 45% chloroquine resistance, the WHO-defined success rate for the combination treatment was 81% for those with chloroquine-resistant malaria and 96% for those with acute, uncomplicated infections, said co-investigator Colonel Wil Milhous, director of experimental therapeutics at the Walter Reed Army Institute of Research in Washington, D.C. In an earlier study with Aotus monkeys, Oduola found similar results with a combination of chloroquine and promethazine, another antihistamine that appears to help reverse resistance. A larger study is under way to explore the mechanism of action, said Milhous, who presented highlights of the in-press findings.

Milhous cautions that “this may only be a stopgap arrangement, as the half-lives of these antihistamines are very short and you need to have large dosages of them in the body.” Moreover, not everybody is convinced that the two-drug approach will actually make a dent in the toll from malaria. “The people who need treatment for malaria are so poor that they can't even buy one drug,” says Win L. Kilama, director-general of the National Institute for Medical Research in Dar es Salaam, Tanzania. “Why would you now expect them to buy two drugs?”

Although most scientists deferred comment until the results are published, several pointed to the potential payoff from the dual-drug approach. “Chloroquine was such a marvelous drug, and everyone was desperate when it was lost,” says Wallace Peters, head of malaria chemotherapy at the International Institute of Parasitology in Herts, United Kingdom. “If this combination drug is safe and it works, it would have greatly restored the value of chloroquine.”

## Malaria-Free Anopheles Mosquitoes

Even better than an effective treatment for malaria, or even a vaccine, would be eliminating the vector itself. But because eradicating mosquitoes is a Sisyphean challenge, the next best thing may be to genetically engineer the Anopheles mosquito so that it can't carry the malarial parasite. Scientists have now taken the first step toward that goal, by successfully transferring foreign genes into a mosquito.

These particular genes don't affect the mosquito's ability to transmit disease, and the work was done in Aedes aegypti, a mosquito that transmits yellow fever and other diseases but not malaria. Still, says Vinod Prakash Sharma, director of the Malaria Research Center in New Delhi, “this is an important milestone in our understanding of the basic research behind inserting alien genes into mosquitoes.”

Medical entomologist Frank H. Collins of the University of Notre Dame in Indiana and Anthony James, a vector biologist at the University of California, Irvine, used genetically modified transposable elements—stretches of DNA that readily move around in the genome—to introduce foreign eye-color genes into mosquito embryos. The result: eight stably transformed lines of mosquitoes sporting altered eye colors. “Until now people always thought this could be done, but we have provided the proof to the principle,” says James, who presented the work at the conference.

The next step is to identify and sequence the major genes that make certain mosquitoes resistant to carrying malaria. Once that is done, a similar technique could be used to breed a mutant of the Anopheles mosquito that would not transmit the parasite.

To be an effective deterrent, of course, the modified mosquitoes must also be able to establish themselves in the general population. Then there are the environmental safety issues that would surround what would be the first release of a transgenic animal species into the wild. “It's a one-way street,” says Louis Miller, chief of the Laboratory of Parasitic Diseases at the U.S. National Institute of Allergy and Infectious Diseases, in Bethesda, Maryland. “You may not be able to pull the genes back once they have been pushed into the wild.”

Even so, Miller and others see genetically altered mosquitoes as a promising approach in their age-old battle against the disease. “It gives you a quantum leap in technology and, at a very low cost, the hope of eliminating malaria,” says Miller.

7. SOLAR PHYSICS

# SOHO Traces the Sun's Hot Currents

1. David Ehrenstein

The mostly ionized gases of the sun's outer layers, at temperatures of hundreds of thousands of degrees, have little in common with Earth's chilly upper atmosphere. But last week researchers announced that data from the Solar and Heliospheric Observatory (SOHO) spacecraft, which watches the sun continuously from a vantage point 1.5 million kilometers sunward of Earth, have revealed rivers of solar material flowing beneath the surface of the sun near its north and south poles. The rivers look somewhat like the jet streams—narrow, high-speed air flows that encircle Earth high in the atmosphere.

Reported last week at a NASA press conference, the solar jet streams are just one of several glimpses of large-scale circulation in the sun described by the SOHO team. The findings, says Douglas Gough, a Cambridge University astrophysicist, “herald what I believe will be a new era of solar meteorology.” SOHO also detected other slower bands of gas flow encircling the sun, which Gough and other researchers—extending the meteorological metaphor—compared to Earth's trade winds. These complex circulation patterns, the solar researchers hope, may prove key to unraveling the mystery of the sun's 11-year cycle of magnetic activity.

The results come from SOHO's Michelson Doppler Imager (MDI), one of 12 instruments on board the European Space Agency craft. The MDI makes sensitive measurements of the undulations on the sun's surface at about 700,000 points at once. These undulations show where acoustic waves traveling through the sun reach the surface, and their frequencies and patterns hold clues to conditions in the subsurface regions probed by the waves. Although Earth-based instruments can also study these oscillations, “we would never have seen [the rivers]” from Earth, says Gough, because of the high sensitivity required to detect them.

These rivers are about 30,000 kilometers across, flowing 40,000 kilometers beneath the sun's visible surface at 140 kilometers an hour, 10% faster than the surrounding gases. The trade-wind-like bands circulate more slowly, at 30 km/hour, at depths ranging from at least 18,000 km all the way to the surface. Astronomer John Harvey of the National Solar Observatory in Tucson, Arizona, says the bands had been glimpsed from Earth as patterns on the sun's visible surface, but the SOHO data confirm “that the surface effect actually does go deeper into the sun; it's not some skin effect.”

The SOHO results also confirm that the bands coincide with collections of sunspots, which are a manifestation of the sun's magnetic activity. That raises researchers' hopes that the newly discovered circulation patterns will prove key to understanding the sun's magnetic cycle. Says Gough, “We're getting these fantastic measurements, and if things go the way they went with terrestrial meteorology, this is going to enable us to have a much deeper understanding of the dynamics of the sun.”

8. MOLECULAR ENDOCRINOLOGY

# Differing Roles Found for Estrogen's Two Receptors

1. Elizabeth Pennisi

To the public, estrogen is one of the most familiar of all hormones. But to scientists, it is still among the most mysterious. Estrogen affects many tissues, including some, such as the ovaries and bladder, that until as recently as last year appeared to lack the receptors needed for estrogen responses. And estrogen-like anticancer drugs, which block the hormone's effects by binding to the same receptor molecules, can check tumor growth in some tissues while stimulating it in others. “There had been mysteries of estrogen action that had been intractable,” says molecular biologist Peter Kushner of the University of California, San Francisco (UCSF).

On page 1508, chemist Thomas Scanlan, Kushner, and their UCSF colleagues, working with a group at the Karolinska Institute and the company Karo Bio in Huddinge, Sweden, now report a major step toward dispelling those mysteries. Their work follows up on a discovery made last year, when the Karolinska group, led by biochemist Jan-Åke Gustafsson, found a second estrogen receptor located on some tissues—including the ovaries and the prostate—that don't contain the first. Now the two teams have joined forces to show that although the new receptor, ERβ, looks like its older sibling, ERα, the two molecules can act quite differently, depending on the particular substance, or ligand, binding to them. For example, estrogen bound to ERα turns on certain genes, whereas in combination with ERβ, it has no effect.

Until now, “it's been hard to find major differences between the receptors,” comments Donald McDonnell, a pharmacologist at Duke University in Durham, North Carolina. “But it makes sense that at least in some circumstances, they'd have opposite effects.” Such differences in activity, together with the different distributions of the two receptors in the body, may help explain estrogen's broad spectrum of activity and some of the paradoxical effects of estrogenlike drugs.

The results also suggest that it may be possible to develop “designer estrogens” that would have more specific effects than the drugs currently in use because they target one receptor and not the other. The goal would be, for example, an estrogen that could protect postmenopausal women against cardiovascular problems, Alzheimer's disease, and osteoporosis without raising their risk of developing breast or uterine cancers, as current supplements do, or a version of tamoxifen that can counter estrogen's cancer-promoting effects in the breast without increasing the risk of uterine cancer.

To tease apart the differences between the two receptors, UCSF graduate student Kolja Paech first put the gene for either ERα or ERβ into cells growing in lab culture that have no estrogen receptors of their own. Then he and his colleagues added a so-called reporter gene linked to either of two different estrogen “response elements”—DNA sequences to which an estrogen-receptor complex has to bind to regulate the expression of other genes. If the reporter gene was activated by that binding, it then caused the cell to light up.

The researchers could detect no differences when they tested cells carrying one of the response elements. No matter which receptor they carried, they lit up in response to all the compounds used, including estrogen itself, the estrogenic compound DES, and three substances that are considered “antiestrogens” because they block some of the hormone's effects—tamoxifen, raloxifene, and one called Imperial Chemical Industries 164384. They got the same results with cells carrying the second element, called AP1, and ERα. But when the cells contained AP1 and ERβ, the researchers found that while the antiestrogens again increased the activity of the gene, estrogen and DES had no effect. “[The differences] were really startling,” Kushner notes. “Up till now, it [had] looked like the ERα and ERβ worked the same.”

In addition, the researchers picked up a hint of why tamoxifen sometimes ceases to be effective in stemming breast cancer growth and even seems to promote it. When they repeated the experiments using breast cancer and uterine cancer cell lines, they found that gene activation by the antiestrogens was particularly strong in one of the breast cancer cell lines containing ERβ. The UCSF team plans to look at whether tamoxifen might somehow induce breast cancer cells to make more of this receptor, and thus contribute to the drug's loss of effectiveness.

Although the genetically altered cells used for these experiments constitute a somewhat artificial experimental system, the researchers suspect that ERα and ERβ will behave similarly in the body, producing different effects depending on the type of cell where they are located and the identity of the ligand that binds to them. The presence of ERβ in the ovaries and urogenital tract, for example, may explain how these tissues are influenced by estrogen even though no ERα had been found there.

Sorting out the effects of the two receptors will not be easy, however. For one thing, there appears to be more than one form of ERβ. Sietse Mosselman, Rein Dijkema, and their colleagues at the Dutch-based drug company, N. V. Organon, have evidence that this receptor may also exist in a longer form than has been reported thus far. And while Gustafsson and his collaborators found that the receptor form they used is by far the most common in the tissues studied, they agree it will be necessary to determine the roles of these other forms.

In addition, just a few weeks ago, molecular biologist Paul Meltzer of the National Human Genome Research Institute in Bethesda, Maryland, and his colleagues described a gene called amplified in breast cancer-1 (AIB1) that makes a protein that works in the nucleus with other proteins to help receptors like ERα and ERβ turn on genes (Science, 15 August, p. 965). “It's a complicated story with multiple proteins vying for interactions with the receptor and with other transcription factors,” says Meltzer.

All these complexities could make it harder to translate the new information into better synthetic estrogen compounds. Still, Gustafsson says, these are exciting times for estrogen researchers. After years of struggling with the hormone's mysteries, Kushner adds, “this [result] provides a potential explanation.”