The Code Breaker, page 17
In his January 2012 grant application, Zhang did not describe the full role of the tracrRNA. Likewise, in his notebook pages and declaration describing the work he had done before June 2012, there is no evidence that he appreciated the role the tracrRNA plays in cleaving the targeted DNA. One of the relevant pages, Carroll says, “includes a rather detailed recipe of the components included, and that list does not have anything that suggests that a tracrRNA was included.” Zhang’s lack of understanding of the role of the tracrRNA, Doudna and her supporters would later say, was the main reason that his experiments were not working well before June 2012.17
Zhang himself, in the paper that he and his colleagues eventually published in January 2013, seemed to acknowledge that a full understanding of the role of the tracrRNA did not come until he saw what Doudna and Charpentier had published. He noted that it had “previously been shown” that the tracrRNA was needed to cleave DNA, and he footnoted the Doudna-Charpentier paper at that point. “The reason that Feng knew those two RNAs were required was based on reading our paper,” Doudna says. “If you look at Feng’s 2013 article, we are cited and we’re cited for that reason.”
When I ask Zhang about this, he says that he included the footnote as a standard practice, because the Doudna-Charpentier paper was the first to publish on the full role of tracrRNA. But he and the Broad Institute say that he was already experimenting with systems that linked the tracrRNA to the crRNA.18
These are murky claims to sort out. For what it’s worth, my own assessment is that Zhang was working on using CRISPR for human gene editing beginning in 2011, and by mid-2012 he was focusing on the Cas9 system and showing some success, but not a lot, in getting it to work. However, there is no clear evidence, and certainly no published evidence, that he had fully sorted out the precise components that were essential or that he appreciated the ongoing role of the tracrRNA until after reading the Doudna-Charpentier paper of June 2012.
* * *
Zhang was open about one thing he learned from the Doudna-Charpentier paper: the possibility of fusing the crRNA and the tracrRNA into a single-guide RNA that could be programmed to target a desired DNA sequence. “We adapted a chimeric crRNA-tracrRNA hybrid design recently validated in vitro,” he later wrote, with a footnote citing the Doudna-Charpentier paper. Marraffini, who was still working with Zhang in June 2012, agrees: “Feng and I began using a single-guide RNA only after we saw Jennifer’s paper.”
As Zhang points out, the creation of the single guide was a useful but not totally essential invention. The CRISPR-Cas9 system can work with the tracrRNA and crRNA remaining separate rather than fused into a simpler molecule, as Doudna and Charpentier’s team had done. The single guide simplifies the system and allows it to be delivered more easily into human cells, but it’s not what enables the system to work.19
CHAPTER 25 Doudna Joins the Race
“We were not genome editors”
It was surprising that Jennifer Doudna was even a contender in the race to make CRISPR-Cas9 work in humans. She had never experimented with human cells, nor had she ever engineered gene-editing tools such as TALENs. That was also true of her primary researcher, Martin Jinek. “I had a lab full of biochemists and people doing crystallography and that sort of thing,” she says. “Whether it was creating cultured human cells or even those of nematode worms, that was not the kind of science my lab was expert in.” So it was a testament to her willingness to take risks that she jumped into what she knew would be a crowded race to take their discoveries about CRISPR-Cas9 and turn it into a tool that would work in human cells.
Doudna realized, correctly, that using CRISPR to edit human genes was the next breakthrough waiting to happen. She assumed that other researchers, including Eric Sontheimer and probably people at the Broad, were racing to do it, and she felt a competitive urgency. “After our June paper, I knew we had to speed up, and it wasn’t clear that our collaborators had the same commitment,” she recalls. “That was a frustration to me. I’m competitive.” So she pushed Jinek to work more aggressively. “You need to make this your absolute priority,” she repeatedly told him, “because if Cas9 is a robust technology for human genome editing then the world changes.” Jinek worried that it would be difficult. “We were not genome editors, unlike some of the labs that pioneered the method,” he says, “so we had to reinvent what others had already done.”1
Alexandra East
At first, Doudna later admitted, she suffered “many frustrations” in her quest to make CRISPR-Cas9 work in human cells.2 But as the fall semester of 2012 began—and Zhang was racing to finish his own experiments—she got a lucky break. A new graduate student named Alexandra East, who had experience working with human cells, joined the lab. What made her arrival especially interesting was where she came from: she had received her training and honed her gene-editing skills as a technician at the Broad Institute, working with Feng Zhang and others.
East was able to grow the necessary human cells and then began testing ways to get Cas9 into the nucleus. When she started getting the data from her experiments, she was not sure that they showed evidence of gene editing. Sometimes biology experiments do not have clear results. But Doudna, who had a far better eye for assessing results, saw the experiments as successful. “When she showed me the data, it was immediately clear to me that she had beautiful evidence of genome editing by Cas9 in the human cells,” Doudna says. “This is a classic difference between a student who is in training and someone like me who’s been doing this for a while. I knew what I was looking for, and when I saw the data she had, it just clicked and I thought, ‘Yes, she’s got it.’ Whereas she was unsure and thought she might have to do the experiments again, I was saying, ‘Oh my gosh, this is huge! This is so exciting!’ ”3
To Doudna, this was evidence that getting CRISPR-Cas9 to edit in a human cell was not a difficult leap or a major new invention: “It was very well known how you could tag proteins with nuclear localization signals to get them to go into the nucleus, which is what we did with Cas9. It was also well known how to change the codon usage in a gene so it would be expressed well in mammalian cells versus bacteria, and we did that as well.” So she did not feel that it was a great inventive step, even though she was racing to be the first to do it. It merely required adapting methods that others had used in the past, such as with TALENs, to get enzymes into the nucleus of a cell. East had been able to do it in a few months. “It was easy once you knew the components,” Doudna says. “A first-year grad student was able to do it.”
Doudna felt it was important to publish something as soon as possible. She realized—correctly, as it turned out—that if other labs became the first to show that CRISPR-Cas9 could be ported to human cells, they would claim that to be a major discovery. So she pushed East to firm up her data through repeated experiments. In the meantime, Jinek worked on ways to turn the single-guide RNA that they had devised in test tubes into a guide that could get Cas9 to the right target in a human cell. It was not easy. The single-guide RNA that he had engineered was not, it turned out, quite long enough to work most efficiently on human DNA.
CHAPTER 26 Photo Finish
Zhang’s final lap
When Feng Zhang began to test the idea of using a single-guide RNA, he discovered that the version described in the Doudna-Charpentier paper of June 2012 worked poorly in human cells. So he made a longer version of the single-guide RNA that included a hairpin turn. That made the single guide more efficient.1
Zhang’s modification showed one difference between doing something in a test tube, like Doudna’s team, and doing it in human cells. “Jennifer was probably convinced by the biochemical results that the RNA didn’t need that extra chunk,” he says. “She thought the short single guide that Jinek had engineered was sufficient, because it worked in a test tube. I knew that biochemistry does not always predict what will actually happen in living cells.”
Zhang also did other things to improve the CRISPR-Cas9 system and optimize it so that it would work in human cells. It’s sometimes hard to get a large molecule through the membrane surrounding a cell nucleus. Zhang used a technique that involved tagging the Cas9 enzyme with a nuclear localization sequence, which grants a protein access to the otherwise impenetrable cell nucleus.
In addition, he used a well-known technique called “codon optimization” to make the CRISPR-Cas9 system work in human cells. Codons are the three-letter snippets of DNA that provide instructions for the specific arrangement of amino acids, which are the building blocks used to make proteins. A variety of codons can code for the same amino acid. In different organisms, one or another of these alternative codons may work more efficiently. When trying to move a gene-expression system from one organism to another, such as from bacteria to a human, codon optimization switches the codon sequence to the one that works best.
On October 5, 2012, Zhang sent his paper to the editors of Science, who accepted it on December 12. Among the authors were Shuailiang Lin, the postdoc who said that Zhang was making little progress until after the Doudna-Charpentier paper appeared, and Luciano Marraffini, who had helped Zhang focus on Cas9 but would later be dropped from his main patent application. After describing their experiments and results, their paper concluded with one of those significant final sentences: “The ability to carry out multiplex genome editing in mammalian cells enables powerful applications across basic science, biotechnology, and medicine.”2
Zhang vs. Church
For twenty-five years, George Church had been working on various methods to engineer genes. He had trained Feng Zhang and was still nominally the academic advisor of Zhang’s lead coauthor, Le Cong. But until the late fall of 2012, he hadn’t been told—or thought he hadn’t been told—by either of them that they had been working for more than a year on turning CRISPR into a human gene–editing tool.
It was not until November of that year, when Church went to the Broad Institute to give a talk, that he found out that Zhang had submitted a paper to Science on using CRISPR-Cas9 in human cells. That was a shock, because Church had just submitted a paper to the same journal on the same topic. He was furious and felt betrayed. He had previously published papers on gene editing with Zhang, and he didn’t realize that his former student now considered him a rival rather than a collaborator. “I guess Feng didn’t get the full culture of my lab,” Church says. “Or maybe he just felt the stakes were so high so he didn’t tell me.” Although Le Cong had moved to the Broad to work with Zhang, he was still a graduate student at Harvard and Church was still officially his advisor. “It was upsetting and seemed to me a breach of protocol that my own student was doing something he knew would interest me but he kept from me,” Church says.
Church raised the issue with the Harvard Medical School’s dean for graduate studies, who agreed that it was improper. Eric Lander then accused Church of bullying Le Cong. “I didn’t want to make a federal case out of it,” Church says. “I didn’t think I was bullying him, but Eric did. So I backed off.” 3
In order to sort this out, I shuttled back and forth between the various contending parties, finding myself constantly reminded that memory can be an unreliable guide to history. Zhang insists that he did, in fact, tell Church that he was working on CRISPR in August 2012, when they drove together to the San Francisco airport from a cutting-edge conference, known as Science Foo Camp, held on the Google campus an hour away. Church has narcolepsy, and he admits he could have dropped off to sleep while Zhang was talking. But even if that happened, it does not, at least in Church’s opinion, get Zhang off the hook for failing to communicate his plans, since he surely would have noticed that he was getting no response from Church.
Over dinner one night, I ask Lander his view of the dispute. Church’s narcolepsy issue is “nonsense,” he insists, and he accuses Church of starting his own work on CRISPR only after Zhang told him he was embarked on that task. When I ask Church about this, I think that I can detect his placid face tightening beneath his beard. “That is absurd,” he replies. “If my students had told me that they wanted to establish their own name in this, I would have backed off. I had a lot else I could have done.”
The quarrel so unsettled Le Cong, who is shy and polite, that he subsequently avoided doing much more work in the CRISPR field. When I tracked him down at Stanford Medical School, where he is focusing his research on immunology and neuroscience, he had just returned from his honeymoon. He told me that he thought he had behaved properly when he withheld from Church the details of what he was doing in Zhang’s lab. “The two labs were independent research groups at two institutions,” he says. “The principal investigators [Zhang and Church] were responsible for sharing information or materials. This is what we were taught as entering PhD students in our Responsible Conduct of Research class.”4
When I tell him Cong’s version of the story, Church chuckles. He teaches an ethics course at Harvard, and he agrees that the behavior of Zhang and Cong was not unethical. “It was within the norms of science.” It did, however, violate the norms he tried to cultivate in his own lab. History would have been a little different, he says, if Zhang and Cong had stayed working for him rather than moving to the Broad. “If they had stayed in my lab, where there was a culture of open behavior, I would have made sure that their relationship with Jennifer was much more collaborative, and there wouldn’t have been all the patent battles.”
Ingrained in Church’s character are instincts that promote reconciliation. Zhang, likewise, avoids conflict. He uses his disarming smile as an effective shield to avoid confrontation. “When one of our grandchildren was born, Feng sent us a colorful play-mat with the alphabet on it,” Church says. “He also invites me to his workshops each year. We all move on.” Zhang feels likewise. “We hug when we see each other.”5
Church succeeds
Church and Zhang ended up in a virtual tie in showing how CRISPR-Cas9 could be engineered for use in human cells. Church submitted his paper to Science on October 26, three weeks after Zhang sent his. After dealing with referee comments, they were both accepted by the editors on the same date, December 12, and were published online simultaneously on January 3, 2013.
Like Zhang, Church created a version of Cas9 that was codon-optimized and had a nuclear localization sequence. Drawing on (and crediting more generously than Zhang did) the Doudna-Charpentier paper of June 2012, Church also synthesized a single-guide RNA. His version was longer than the one Zhang devised and ended up working even better. In addition, Church provided templates for the homology-directed repair of the DNA after CRISPR-Cas9 had made its double-strand break.
Though their papers differ somewhat, they both came to the same historic conclusion. “Our results establish an RNA-guided editing tool,” Church’s paper declared.6
The editor at Science was surprised, and a bit suspicious, that the journal had received two papers on the same topic from researchers who were supposed to be colleagues and collaborators. Was he being gamed? “The editor felt as if Feng and I were doing some kind of double dipping, doing two papers when we should have submitted one,” Church recalls. “He required a letter from me saying these papers were actually done without knowledge of each other.”
Martin Jinek
CHAPTER 27 Doudna’s Final Sprint
In November 2012, Doudna and her team were pushing hard to pin down the results of their experiments so they could win the race to publish on the use of CRISPR-Cas9 in humans. She didn’t know that Church had just submitted a paper to Science, and she had barely heard of Feng Zhang, who also had. Then she got a phone call from a colleague. “I hope you’re sitting down,” the caller said. “CRISPR is turning out to be absolutely spectacular in George Church’s hands.”1
Doudna already knew from Church’s email that he was working on CRISPR, and when she heard about his progress in making it work in humans, she gave him a call. He was gracious and explained the experiments he had done and the paper he had submitted. By then, Church had learned about Zhang’s work, and he told Doudna that it was slated for publication as well.
Church agreed to send Doudna a copy of his manuscript as soon as the editors at Science accepted it. When she received it in early December, she was deflated. Jinek was still doing experiments in her lab, and the data they had were not as extensive as those of Church.
“Should I still go ahead and try to publish my work anyway?” she asked Church. He said yes. “He was very supportive of our work and of us publishing,” she says. “I thought he behaved as a great colleague.” Whatever experimental data she produced, Church told Doudna, would add to the accumulation of evidence, especially on how best to tailor the RNA guide.
“I felt it was important to keep pushing with our experiments, even if others were already doing the same work,” Doudna later told me, “because that would show how easy it was to use Cas9 for human genome editing. It showed that you didn’t have to have special expertise to use the technology, and I felt that that was important for people to know.” Publishing their work would also help her stake a claim that she had demonstrated CRISPR-Cas9 could work in human cells at approximately the same time as competing labs had.
That meant she needed to get her paper published quickly. So she called a colleague at Berkeley who had recently started an open-access electronic journal, eLife, that published papers after less review time than traditional journals such as Science and Nature. “I talked to him, described the data, and sent him a title,” Doudna says. “He said it sounded interesting and he would get it reviewed quickly.”
