By Jonathan Rabinovitz
Illustration by Calef Brown
Photograph by Trujillo-Paumier
Time is running out. Jen Liou, nine months pregnant, guesses how many hours she has until delivery as she rushes to compile records needed to document her research. Over the previous year, she has endured physical pain, juggled dozens of tasks and made meticulous plans to get to this point.
The delivery in question is not the infant girl in her womb. Yes, the baby could arrive at any moment, but first Liou needs to submit the last bit of material required for publication of a paper that could resolve a fundamental question in biology: How does a cell signal its outer membrane to let in more calcium when it has depleted its store of the element?
“Baby, don’t come now,” Liou prays. “Mommy’s not ready. Just give me time to do this last thing.” She confides to a labmate that a trip to the maternity ward may be imminent: She is bleeding. “What are you doing?” he pleads. “You need to get out of here.”
Liou does not budge from her stool, fearful that any movement could trigger the birth of her second child. She asks her colleague to get her a sandwich, recruits another to gather records, then gets down to work.
This day — May 24, 2005 — could be the culmination of several years of Liou’s work as a postdoctoral scholar in the laboratory of Tobias Meyer, PhD, at the Stanford School of Medicine, and she has staked her career on it. It isn’t exactly what Liou, who has a PhD in immunology from UCSF, had imagined working on when she arrived at Stanford in 2001. She did not consider herself “a calcium person.” She was not part of the tight-knit circle of scientists, with their own journals and their own conferences, who had been hammering away for the last 20 years to figure out the mechanism behind what some refer to as CRAC (calcium release-activated calcium) and others SOC (store-operated calcium) influx.
The project that Meyer launched — and Liou is hoping to complete — showcases what it takes to achieve a breakthrough in the basic sciences. It requires a vexing, long-standing problem. It demands feats of innovative wizardry. And it often must inspire a heroic and selfless effort from researchers. The bottom line: It is a tremendous gamble with many losers and only the remote possibility of a jackpot.
Figuring out the mechanism behind calcium influx would certainly have huge repercussions. The flow of calcium allows neurons to fire and muscle cells to contract. It activates lymphocytes to fight off diseases. Its failure may be the culprit in a host of autoimmune disorders. Calcium’s role extends far beyond building bones and teeth: Ultimately its flow regulates nearly every aspect of human physiology.
Meyer, the Mrs. George A. Winzer Professor in Chemistry, helped make some of the early breakthroughs in calcium research working in the lab of Stanford’s Lubert Stryer, MD, in the late 1980s, but his interests extended in the ensuing decades far beyond that. “The ultimate systems biology problem,” is how he describes his lab’s work. “How can a cell’s entire control system be quantitatively modeled?” His team studies how cells integrate inputs from multiple receptors, how they compute this information and how they then make decisions about such functions as migration, synapse formation and differentiation. He has published findings ranging from the mechanism of the herpes virus to the anti-cancer activity of bryostatins. While doing that, he has developed new biosensors and pioneered new microscopy techniques.
In late 2002, Meyer had a wild idea for a new technology: He wanted to build the first-ever comprehensive library of RNA interference, or RNAi, a phenomenon discovered four years earlier. Although RNA molecules normally enable the cell to build proteins needed for a range of cellular functions, two scientists in 1998 showed that these molecules sometimes do just the opposite: They knock out genes responsible for making specific proteins. (The discovery of RNAi resulted in a 2006 Nobel Prize for Craig Mello, PhD, at the University of Massachusetts, and Andrew Fire, PhD, who is now at Stanford.)
Scientists had started to use RNAi in experiments to silence genes in a cell so they could test how a gene’s absence would affect different tasks. The work was conducted in fruit flies and worms, but had not been used much in mammalian cells, in which the long strands of RNA initially used for RNAi were toxic. Meyer hoped to apply a novel technique for making short snippets of RNA interference, or siRNA, created in a neighboring lab, that would not trigger such adverse reactions in human cells, yet would still be precise and inexpensive. He wanted to run screens with hundreds of specially selected siRNAs to test how using them to silence genes one by one would affect a variety of cell signals — including what causes a cell to signal that its calcium store had been depleted.
The project would require an unusual partnership between labs and a larger team than the typical study in cell biology. It would involve entering a race with many competing to be the first to make the discovery. When Meyer asked his lab members to volunteer, many declined, saying it was too dicey. Liou was an exception. “She was very quickly excited,” says Meyer. “She was ready to take on something that hundreds had not succeeded in doing.” Four others also joined in from his and the neighboring lab, under James Ferrell, chair of the Department of Chemical and Systems Biology, who also agreed to help.
“I thought it was absolutely timely and an important thing to try, but I had no idea whether we could pull it off,” says Ferrell. Liou was also aware of the risk. She could invest months in the project and have nothing to show for it. “I thought we could wind up producing something that was absolutely junk,” she says.
Breakthroughs involve hard work and genius, but they often turn on serendipity. In this case, both Meyer and Ferrell trace it to a chance encounter at the airport in Denver in 1999. Joshua Jones was waiting in line for his flight out to Stanford for graduate school interviews. He noticed the guy next to him was looking at the same agenda. “We started talking and didn’t stop,” says Jones. “We hit it off.”
As luck would have it, Jones and Jason Myers were assigned the same room for that weekend, and they talked through the nights. They chose to be roommates in their first year of graduate school. The two kept talking in the coming years about music, sports and dating — but mostly about research. Jones was in Tobias Meyer’s lab, Jason Myers in Ferrell’s, but they bounced between their stations.
Jason Myers proposed to investigate for his dissertation a new way to make siRNA for mammalian cells. A recent study had shown that an enzyme, Dicer, could cleave RNA in fruit flies into siRNA. He had an idea to make Dicer so that it could be used in vitro to make siRNAs for use in human cells. He thought his method could cheaply and easily produce vast quantities of siRNAs.
“We started talking and didn’t stop. We hit it off.”
Myers knew his proposal did not fit neatly into the work being done in Ferrell’s lab, but Ferrell encouraged him to pursue his passion. Along with Ferrell, Tobias Meyer was on his thesis committee. “I don’t think I was initially super-supportive,” says Tobias Meyer. Still, Jason Myers talked it through with his friend, Jones, who helped on the experiments. Myers ultimately showed proof of concept in a few cells. When he submitted his findings for publication in fall 2002, the second and third co-authors were, respectively, Jones and Tobias Meyer, with Ferrell as senior author. The following year, when Jones published a paper further confirming the validity of the method — it shows how mitosis speeds up when a particular gene is knocked out — Myers and Ferrell were listed as co-authors, with Tobias Meyer the senior author.
“It is two people from two labs who set this in motion,” says Meyer of his decision to try to build an siRNA library. “Without their getting along personally and supporting each other, it wouldn’t have happened.” Meyer saw that it worked in a few instances, so why not try more?
Still, imagine making a layer cake and then being told you had to make another 2,000 as quickly as possible — and, by the way, each one had to be a different flavor.
Building the library would demand creative thinking, engineering savvy — and work that is monotonous yet requires precision and consistency. Meyer found five scientists — postdocs Liou and Won Do Heo, PhD; graduate students Jones and Myers; and research assistant Man Lyang Kim — as well as Ferrell and himself. That’s bigger than most studies in cell biology, but it provided the needed range of expertise. “Tobias is a great mathematician, and Jim has an amazing understanding of biochemistry,” says Jason Myers. “Josh and I know how to create siRNA.” Kim has critical skills in cloning proteins and creating fluorescent tags. Heo has experience in acquiring reagents that would be needed to complete the work.
And then there was Liou.
As a young girl growing up in Taipei, Liou’s ambition was to become a housewife. She excelled in cooking, gardening and embroidery. A colleague once wondered if, with her gracious, soft-spoken manner, she was “too nice.” But Liou has a quiet confidence. At her all-girls middle and high schools, she excelled in science, and it became her passion in college. She had no problem competing in what often are stereotyped as “male” disciplines. After graduating high school, for instance, she voluntarily did two weeks of military training on a small, remote island off Taiwan, learning how to fire an M-16 and to work with her unit to scale a 3-meter wall. “I have this desire to test my limits,” she says. Adds Jason Myers, “Jen would do whatever it took to get the job done.”
In the spring of 2003, Tobias Meyer designed an algorithm to search the database of the roughly 25,000 genes in the human genome, plucking out those with signaling domains. He then followed that with text searches of signaling-related terms. He had compiled a list of some 3,000, but wanted it more focused. At that point, he spent two days reviewing them manually one by one, looking at the repeating sequences of more than 1,000 characters. “I just wanted to get a feeling for what was in there,” he said. He instinctively eliminated ones that he did not think would be involved in cell signaling.
The final list contained 2,304 siRNAs that he hoped could be used to answer a variety of questions about cell signaling.
It was a leap of faith that Meyer had chosen the right ones, but Liou and her colleagues didn’t have room for doubts. They set about turning the abstract list into a physical reality. They began the assembly line that would produce multiple sets of the 24 trays — each about the size of a Pop-Tart — with 96 wells, one for each siRNA. (The math: 24 times 96 equals 2,304.)
The team repeated ad nauseum the steps needed to transform raw DNA into the desired siRNA. They did at least 12,000 squirts from a 12-pronged pipette of a specific solution for each well, needed to do the polymerase chain reactions to produce small strands of DNA. When they discovered that too much water was evaporating from the wells on the rim of the tray, they devised a new way to wrap it in an aluminum-foil-like cover; they then rubbed the rows of the trays roughly 5,000 times to ensure the wells were snugly covered during the reactions. “Doing it once is easy,” Liou says, “but when you do it dozens of times a day — well, it gets brutal.” A lump developed at the end of her thumb.
The work then involved in vitro transcription to turn the DNA into RNA, being careful to add just the right amount of reaction component in each well, and incubating it for three hours at 37 degrees Celsius. The next-to-last step was to add the Dicer, which they had made themselves earlier in the year following Myers’ specifications, and incubate it overnight, for another 14 to 16 hours, making sure that one of them would be back first thing in the morning. The final step was to run the product twice through a 96-well filter devised by Myers, washing the product after each time to remove unwanted chemicals and any undiced long RNA.
Over the entire process, Liou and Kim cooked up more than 100 batches of gel, with water and salt, which then needed to be injected with DNA or RNA samples and bathed in a highly toxic stain, so that they could then take ultraviolet photos to confirm that the “product” was acceptable.
By December 2003, they had finished. It had taken six months of work by all five. The result was in the freezer: 10 sets of 24 trays.
During much of the time that she toiled on the library, Liou was going over in her mind how she would conduct her screens. She was determined to use the library to address a daunting problem that has foiled some of the world’s greatest scientists.
The challenge could be traced back to 1986 when James Putney, PhD, now the principal investigator at the Calcium Regulation Group at the National Institutes of Health, proposed a theory for how a cell sensed that it was depleting its calcium and then opened channels in its plasma membrane to take in more. In an article in the journal Cell Calcium, Putney recounted how the cell has a store of calcium in its endoplasmic reticulum, a spider-weblike organelle comprised of tubules, vesicles and cisternae that stretches across the cell’s cytoplasm from the nucleus to its plasma membrane. In response to a variety of stimuli, he observed, there is an initial transient release of sequestered calcium from the endoplasmic reticulum, which is then followed by a longer-lasting entry of calcium from outside the cell. Putney ran through several theories — such as the calcium level being mediated by phosphatidic acid in the plasma membrane or a trigger inhibiting pumps that typically expel calcium from the calcium membrane — before proposing his own: The emptying of the ER prompts the signal.
In a recent e-mail, Putney explains that shortly after proposing the idea of such a store-regulated channel, much of his work was devoted “to trying to obtain solid evidence to convince the skeptics that such a mechanism even existed.” Gradually, research mounted supporting his hypothesis, including some critical work by Stanford’s Richard Lewis, PhD, who was the first to observe the CRAC channel in T cells, but the puzzle was incomplete. “More laboratories began to tackle the key questions: What was the signal from the ER to the plasma membrane, and what was the identity of the channel?” he says.
Instead of adding genes back, they knock them out.
When Meyer asked Liou in late 2002 to work on the project, she was 32 years old, married, with a 2-month-old son and in the midst of studying an entirely different signaling pathway. But her graduate studies in the lab of UCSF’s Arthur Weiss, MD, PhD, made her an ideal candidate. Weiss does research on the molecular mechanisms that activate lymphocytes, so Liou appreciated how the trigger for SOC influx is a critical missing piece to this puzzle, among others.
But perhaps most important, Liou learned how Weiss had discovered genes involved in regulating other calcium-related processes. In addition to using biochemical purification to search for key molecules, Weiss pioneered a genetic approach, starting by zapping lymphocytes with mutation-inducing radiation and the collecting cells with damaged calcium processing as a result. Then he added back in one gene at a time to see which one would restore the function, ultimately identifying several.
Yet this and biochemical methods had not pinpointed the molecules behind SOC influx. The search for the signaling protein was painfully slow, because researchers had to make an educated guess as to which one it would be and then run experiments to test it. “RNAi really changes the game,” says Stefan Feske, MD, an assistant professor at New York University, who has been doing research in this area for more than a decade. It essentially lets scientists reverse the genetic approach; instead of adding genes back, they knock them out. “An siRNA library focusing on 2,300 signaling genes in mammalian cells is a novel technology in 2004, and it takes a leap of faith to go ahead with it.”
Jen Liou now was among the first scientists to have such a tool at her disposal, though she knew that other groups were trying different approaches. The library was promising, but it had only a slice of the genome. And it could test for only one siRNA at a time. If two similar proteins happened to be involved in the process, knocking out just one might not work.
Nonetheless, Liou forged ahead, deciding to run her first experiment on HeLa cells, a line of cervical cancer cells widely used in labs because of their robustness. She would place them in each of the 2,304 wells holding a unique siRNA and then inject “agonists” that would promote the depletion of the calcium stores and the opening of the calcium channels. The idea was to see which cells didn’t get the second wave of calcium and then identify the protein that was knocked out in that particular well.
Liou selected a microplate reader in the basement of Stanford’s Center for Clinical Science Research that could measure the change in the amount of calcium in the cell over a four-minute period. She programmed it to chart the calcium as it first spiked with the depletion of the store in the endoplasmic reticulum, and then continued on a plateau as channels formed to let in calcium from outside the cell — or, if she were lucky, would drop precipitously indicating that the signal had not been sent. She chose two common compounds to prompt such a response, histamine and thapsigargin, both of which were relatively inexpensive. “It takes some imagination,” she says. “There isn’t an obvious control for SOC influx.”
“At that point, I was very excited and happy.”
Liou was able to sign up to use the microplate reader every week on Thursdays and Fridays, starting on April 22. She would sit there all day, breaking sometimes only to buy her regular vegetarian meatball sub on a Dutch-crunch roll from the cafe upstairs, which she would bring back to the desk so that she would not waste a moment of her time with the machine. She didn’t know what would be unusual, so occasionally she circled in her logbook the wells that had interesting graphs. It was on the sixth day of screening, on May 7, at 2:14 p.m., that she saw a particularly steep drop from one of the wells. It seemed to stand out from the others, but she wasn’t quite sure. She put a star next to it on her list, noticed its name — stromal interaction molecule 1, or STIM1 — and kept screening. She did not tell anyone. Several weeks later, on June 24, she noticed another strong response and was struck when she saw the source was a sister molecule: STIM2.
Still, Liou had circled roughly 100 possible candidates. She needed to narrow it down. She repeated the screening using a smaller concentration of the selected siRNA. This time STIM1 and STIM2 remained the top hits, while many of the others did not. She did a literature search on STIM proteins and turned up a paper about STIM as a possible tumor suppressor. It mentioned that it has an “EF hand” — a loop of amino acids at one end that binds to calcium ions. “It was like, Ooooooh!” she says, explaining that it suggested a mechanism by which STIM worked: Its EF hand would be in the endoplasmic reticulum and sense when the calcium had been depleted.
“At that point, I was very excited and happy,” she recalls. She had been putting in long hours, working all day, coming home to make dinner and put her son to bed, then returning to the lab to put in a few more hours. She finally felt free to take a breather. “That’s probably how I became pregnant!” she says.
Meyer was intrigued when she presented the results, but still skeptical. He wanted visual evidence. “I want to see it,” he told her. How did the STIMs lead to the opening of the channels in the plasma membrane? The next step, she and Meyer agreed, was to use his spinning-disk confocal microscope, which is not standard equipment in most labs but allows for greater depth perception. She began to work with research assistant Man Lyang Kim to place fluorescent tags on STIM1 so she could see how the glowing molecules responded to the calcium depletion. She was working at an intense pace, while dropping off and picking up her toddler son every day from day care and sharing chores with her husband. After dinner and the bedtime story, it was back to the lab.
On a Saturday morning in the first week of November, she was forced to take a break. She started bleeding in the lab and had to go to the emergency room. She learned that she had been carrying twins, and that one no longer had a heartbeat. She was devastated: She took a day to rest, but the doctor said there was nothing else to be done, and no benefit from staying in bed. She spoke with her mother, who encouraged her to concentrate on the baby she was still carrying and to go back to living her life. Her sadness endured, but work was a respite from the grief.
Liou began observing what happened to HeLa cells, transfected with fluorescent-tagged STIM1, when histamine and thapsigargin were added. The results were striking. While STIM1 was distributed throughout the endoplasmic reticulum at the start of the experiment, it quickly moved to spots on the periphery of the endoplasmic reticulum, adjacent to the plasma membrane. The dance took less than a minute. It was clear evidence of how the signal reached the cell’s periphery.
“We have to get this published,” Meyer said upon viewing the images in mid-November. Liou redoubled her efforts, working even later into the night. “I was so excited, I was only sleeping from 3 a.m. to 7 a.m.,” she says, noting that the prospect of finishing had given her such a surge of energy that it was difficult to go to bed at all. She sent her latest draft to Meyer when she left the lab, and his comments were waiting when she arrived early the next morning. The two of them worked through the Thanksgiving break, and then sent the paper to Nature on Nov. 27. About a week later, they heard back; the reviews were enthusiastic, though they asked for more experiments. Liou scrambled to get it done. She sent another draft to Meyer on Christmas Eve. A few hours later he responded:
“Dear Jen, happy Christmas. Please check the final version if you can. I could submit it later tonight.” At 4:39 a.m. on Dec. 26, 2004, it was done. “Jen, I submitted it,” he e-mailed. “Best wishes and happy holidays.”
Sadly, the road to publication was not smooth. On April 25, 2005, Nature informed them that it wished to get new reviews. Liou and Meyer decided it was time to try another journal. Four days later, they submitted it to Current Biology. The delay cost them the chance to be first with the news about STIM1: On May 2, a competing group of scientists published a paper that identified STIM1 as an essential component of store-operated calcium channel function in The Journal of Cell Biology.
Liou was frustrated but all she could do was wait. On May 23, after a few exchanges between the editors, her and Meyer, Current Biology accepted the paper. Still, the editors needed her to provide high-resolution images, a paper copy of the data and other documentation within the week for it to be published.
So that is why on that morning of May 24, Liou rushed to meet their requests, fearing that the labor pains could prevent her from finishing. The research assistant, Kim, dropped what he was doing on another project to help her out. It was done by 7 p.m. Her baby girl was born two days later.
Although Liou’s paper was technically not “first,” it made a huge impression when it was posted online a week later. “Given the checkered past of studies trying to solve this problem, it was important that two independent and well-respected labs came to essentially identical conclusions,” says the NIH’s Putney. What’s more, he adds, there were important details in the Liou-Meyer paper that were not in the other. “Most significant is the first report that STIM proteins move and redistribute in response to store depletion,” he says. Feske, the NYU researcher, agrees, describing as “heroic” the work that Liou and her colleagues did. “The whole field had been something of a mess,” he explains. “They took a high-risk approach that could have left them with nothing, but that’s how the big advances in science are made.”
Liou became something of a rock star in the world of calcium studies. The Stanford team’s discovery was selected as a “Signaling Breakthrough of 2005” by Science Signaling. It was named a “Hot Paper” by ISI Essential Science Indicator of Thomson Scientific.
More important, the discovery sparked a renaissance in calcium research. “I went to a conference in 2006, and about half of the posters were on the newly discovered calcium sensor STIM,” Liou says. “It’s an explosion.” That same year, three independent groups, including NYU’s Feske, identified ORAI-1, another protein, as the second essential regulator of SOC influx. STIM initiates signaling from the ER to the plasma membrane, then ORAI constitutes the calcium channel. “STIM is the beginning of the pathway and ORAI is the end,” Liou says.
Patents have been filed on the mechanisms involving STIM and ORAI, and pharmaceutical companies are interested in them. Liou has continued to do research on the store-operated calcium influx, publishing with Meyer several more papers. She plans to launch her own lab in September as an assistant professor and endowed scholar at the University of Texas Southwestern Medical Center in Dallas. Her daughter, now 4, inspires her to keep investigating immune responses — the girl has severe allergies. Someday her daughter might even be taking a drug that targets STIM or ORAI.
In the meantime, siRNA libraries have become commonplace.
On a recent afternoon, Tobias Meyer recalls how “most people didn’t want in” when asked to join in the effort. He acknowledges it wound up taking longer than he had expected, as they needed to improvise and “tweak” things to adjust to the larger scale. With the sun streaming into his office and his bicycle propped against the wall, he gestures at nine plastic crates on the shelf above his desk. Several are labeled “Primers: Human siRNA.” They’re kits for constructing the exact same library Liou used to discover the STIM proteins. Labs around the world can buy similar ones for a few hundred dollars. A single scientist can build it in a few days. “Once you have written out protocols,” he observes, “it all becomes routine.”