The Code Breaker, page 38
Fyodor Urnov’s reflection as he photographs the note
Janice Chen and Lucas Harrington
Feng Zhang with Patrick Hsu
CHAPTER 51 Mammoth and Sherlock
CRISPR as a detection tool
At the March 13 meeting that Doudna convened to address the coronavirus, she decided that a top priority was to create a high-speed conventional PCR testing lab. But during the discussion, Fyodor Urnov suggested that they also consider a more innovative idea: using CRISPR to detect the RNA of the coronavirus, similar to how bacteria use CRISPR to detect attacking viruses.
“There’s a paper that just came out on that,” a participant interjected.
Urnov showed a slight flash of impatience and interrupted, for he knew the paper well. “Yes, from Janice Chen, formerly of the Doudna Lab.”
There were actually two similar papers that had just come out. One was from former members of the Doudna Lab who had formed a company to use CRISPR as a detection tool. The other, not surprisingly, sprang from Feng Zhang of the Broad Institute. Once again, the two realms were competing. This time, however, it was not a race to patent methods for editing human genes. In this new race, the goal was to help save humanity from the novel coronavirus, and their discoveries were being shared for free.
Cas12 and Mammoth
Back in 2017, Janice Chen and Lucas Harrington were doctoral students working in Doudna’s lab exploring newly discovered CRISPR-associated enzymes. Specifically, they were analyzing one that became known as Cas12a, which had a special property. It could be targeted, like Cas9, to find and cut a specified sequence of DNA. But it didn’t stop there. Once it cleaved the double-stranded DNA target, it went into an indiscriminate cutting frenzy, chopping up any single-stranded DNA that was nearby. “We started to see this very weird behavior,” Harrington says.1
Over breakfast one day, Doudna’s husband, Jamie Cate, suggested that this property could be harnessed to create a diagnostic tool. Chen and Harrington had the same idea. They combined a CRISPR-Cas12 system with a “reporter” molecule, which was a fluorescent signal connected to a bit of DNA. When the CRISPR-Cas12 system found a targeted sequence of DNA, it would also chop up the reporter molecules and cause a glowing signal. The result was a diagnostic tool that could detect whether the patient had a particular virus or bacteria or cancer. Chen and Harrington dubbed it the “DNA endonuclease targeted CRISPR trans reporter,” a very clunky phrase that was crafted in order to create the CRISPR-like acronym DETECTR.
When Chen, Harrington, and Doudna submitted their findings in an article to Science in November 2017, the editors requested that they write more about how to turn the discovery into a diagnostic test. Even the traditional scientific journals were now showing greater interest in connecting basic science to potential applications. “If a journal tells you to do something like that,” Harrington says, “you start working on it very hard.” So over Christmas break of 2017, he and Chen collaborated with a researcher at UC San Francisco to show how their CRISPR-Cas12 tool could detect human papillomavirus (HPV), a sexually transmitted infection. “We were going back and forth with a giant piece of lab equipment in an Uber, testing different patient samples,” he says.
Doudna prodded Science to expedite publication as part of its fast-track program. They resubmitted their article in January 2018 with the data the editors had requested showing that DETECTR detected HPV infections; it was accepted, and a version went online in February.
Ever since Watson and Crick ended their famous DNA paper by saying, “It has not escaped our notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material,” it has become standard to end journal papers with an understated but important forward-looking sentence. Chen, Harrington, and Doudna ended their paper by saying that the CRISPR-Cas12 system “offers a new strategy to improve the speed, sensitivity and specificity of nucleic acid detection for point-of-care diagnostic applications.” In other words, it might be used to create a simple test to detect virus infections quickly, at home or in a hospital.2
Even though Harrington and Chen had not yet gotten their doctorates, Doudna encouraged them to form a company. She was now a strong believer that basic research should be combined with translational research, moving discoveries from bench to bedside. “A lot of other technologies that we had discovered were bought as a defensive strategy by big companies that then didn’t develop them,” Harrington says. “So that motivated us to start our own company.” Mammoth Biosciences launched officially in April 2018 with Doudna as chair of its scientific advisory board.
Cas13 and SHERLOCK
As was often the case, Doudna and her team were in a competition with her cross-country rival, the Broad Institute’s Feng Zhang. Working with the CRISPR pioneer Eugene Koonin of the NIH, Zhang had used computational biology to sort through the genomes of thousands of microbes, and in October 2015 they reported on their discovery of many new CRISPR-associated enzymes. In addition to the previously known Cas9 and Cas12 enzymes that target DNA, Zhang and Koonin found a class of enzymes that target RNA.3 They became known as Cas13.
Cas13 had the same odd trait as Cas12: when it found its target, it went into a cutting frenzy. The Cas13 not only cut its targeted RNA, it then proceeded to cut up any other nearby RNA.
At first Zhang assumed this was a mistake. “We thought that Cas13 would cleave the RNA just the way that Cas9 cleaved DNA,” he says. “But whenever we did a reaction with Cas13, the RNA got shredded in many different places.” He asked his lab team whether they were sure they had been purifying the enzyme correctly; maybe it was contaminated. They painstakingly eliminated all possible sources of contamination, but the indiscriminate cleavage kept happening. Zhang speculated that it was an evolutionary method to have the cell commit suicide if it got too infected by an invading virus, thus preventing the virus from spreading as fast.4
Doudna’s lab then contributed to the study of precisely how Cas13 works. In a paper in October 2016, Doudna and her coauthors—including her husband, Jamie Cate, and Alexandra East-Seletsky, the graduate student who had done some of the key 2012 experiments on CRISPR in human cells—explained the different functions that Cas13 performs, including being able to indiscriminately chop up thousands of other nearby RNAs once it reaches its target. This promiscuous chopping makes it possible to use Cas13 with fluorescent reporters (as was done with Cas12) to be a detection tool for a specified RNA sequence, such as that of a coronavirus.5
Zhang and his colleagues at the Broad were able to create such a detection tool in April 2017, which they named “specific high sensitivity enzymatic reporter unlocking,” which was reverse-engineered (though not very well) to produce the acronym SHERLOCK. The game was afoot! They showed that SHERLOCK could detect specific strains of Zika and Dengue viruses.6 Over the next year, they made a version that combined Cas13 and Cas12 to detect multiple targets in one reaction. Then they were able to simplify the system and make it possible for the detection to be reported on paper lateral flow strips, similar to pregnancy tests.7
* * *
Zhang decided to start a diagnostics company to commercialize SHERLOCK, just like Chen and Harrington had launched Mammoth. Zhang’s cofounders included the two graduate students who were the lead authors on many of the papers from his lab describing CRISPR-Cas13: Omar Abudayyeh and Jonathan Gootenberg. Gootenberg recalls that they almost decided not to publish a paper when they first discovered the tendency of Cas13 to go into a frenzy of indiscriminate RNA cutting. It seemed like a useless quirk of nature. But once Zhang figured out how to harness that quirk to create a virus-detection technology, Gootenberg realized how discoveries in basic science can turn out to have unexpected real-world applications. “You know, nature’s got a ton of amazing secrets in it,” he says.8
It took a while to get Sherlock Biosciences funded and launched because Zhang and his two graduate students did not want profit to be the main goal of the company. They wanted the technologies to be affordable in the developing world. So the company was structured in a way that allowed it to profit on its innovations while still taking a nonprofit approach in places where there was great need.
Unlike the Doudna-Zhang competition for the patents, the one involving diagnostic companies was not very contentious. Both sides knew that the technologies had enormous potential to do good. Whenever there was a new epidemic, Mammoth and Sherlock could quickly reprogram their diagnostic tools to target the novel virus and produce testing kits. The Broad team, for example, sent a team with SHERLOCK to Nigeria in 2019 to help detect victims of an outbreak of Lassa fever, a virus in the same family as Ebola.9
At the time, using CRISPR as a diagnostic tool seemed to be a worthy endeavor, though not a particularly exciting one. It did not get as much buzz as using CRISPR to treat diseases or edit human genes. But then, at the beginning of 2020, the world suddenly changed. The ability to quickly detect an attacking virus became critical. And the best way to do it faster and cheaper than the conventional PCR tests, which required a lot of mixing steps and temperature cycles, was to deploy RNA-guided enzymes that had been programmed to detect the genetic material of the virus—in other words, adapt the CRISPR system that bacteria had been deploying for millions of years.
Feng Zhang (top left) with Omar Abudayyeh (top right) and Jonathan Gootenberg (middle right) at a Zoom meeting on COVID detection
CHAPTER 52 Coronavirus Tests
Feng Zhang
In early January 2020, Feng Zhang started getting emails about coronavirus written in Chinese. Some were from Chinese academics he had met, but he also got an unexpected one from the science officer at China’s consulate in New York City. “Even though you are American and not living in China,” it said, “this is really a problem that’s important for humanity.” It quoted an old Chinese saying: When one place is in trouble, assistance comes from all quarters. “So we hope that you can think about it and see what you can do,” the email urged.1
Zhang knew little about the novel coronavirus other than what he had read in a New York Times article describing the situation in Wuhan, but the emails “gave me a sense of urgency about the situation,” he says. This was especially true of the exchange he had with the Chinese consulate. “I usually don’t have any interactions from them,” says Zhang, who had immigrated to Iowa with his parents when he was eleven.
I asked him whether Chinese authorities think of him as a Chinese scientist. “Yeah, probably,” he says after a pause. “I think they probably think of all Chinese people as Chinese. But that’s irrelevant because the world is so connected now, especially in a pandemic.”
Zhang decided to reconfigure the SHERLOCK detection tool so that it could test for the new coronavirus. Unfortunately, he didn’t have anyone in his lab to handle the necessary experiments. So he resolved to go to his bench and do the experiments himself. He also enlisted his two former graduate students, Omar Abudayyeh and Jonathan Gootenberg. They had moved on to open their own lab at MIT’s McGovern Institute, a block from the Broad, and they agreed to collaborate with him again.
Zhang did not initially have access to samples of the coronavirus from human patients, so he made a synthetic version of it. Using the SHERLOCK process, he and his team devised a detection test that took only three steps and could be done in an hour without fancy equipment. All it required was a small device to keep the temperature constant while the genetic material from the samples was amplified through a chemical process that was simpler than PCR. The results could be read using a paper dipstick.
On February 14, well before most of the U.S. had focused on the novel coronavirus, Zhang’s lab posted a white paper describing the test and inviting any lab to use or adapt the process freely. “Today we are sharing a research protocol for SHERLOCK-based COVID-19 #coronavirus detection, and hope it will help others who are working to combat the outbreak,” Zhang tweeted. “We will continue to update this as we make further progress.”2
The company he had founded, Sherlock Biosciences, quickly began work on turning the process into a commercial testing device that could be used in hospitals and doctors’ offices. When the CEO, Rahul Dhanda, told his team that he wanted the company to focus on COVID, the researchers literally swung their chairs back to their workbenches to take on the mission. “When we say a pivot, there was a literal pivot of chairs at the same time there was a pivot of the company towards a new goal,” he says. By the end of 2020, the company was working with manufacturing partners to turn out small machines that could be used to get results in less than an hour.3
Chen and Harrington
Around the time that Zhang began working on his coronavirus test, Janice Chen got a call from a researcher on the scientific advisory board of the company she had founded with Doudna and Lucas Harrington, Mammoth Biosciences. “What do you think about developing a CRISPR-based diagnostic to detect the SARS-CoV-2 virus?” he asked. She agreed that they should try. As a result, she and Harrington became part of yet another cross-country competition between Doudna’s circle and Zhang’s.4
Within two weeks, the Mammoth team was able to reconfigure its CRISPR-based DETECTR tool so that it would detect SARS-CoV-2. One benefit of collaborating with UC San Francisco, which has its own hospital, was that they could test on real human samples, drawn from thirty-six COVID patients, unlike the Broad, which initially had to use synthetic viruses.
The Mammoth test relied on the CRISPR-associated enzyme that Chen and Harrington had studied in Doudna’s lab, Cas12, which targets DNA. That would seem to make it less suited than SHERLOCK’s Cas13, which targets RNA, the genetic material of the coronavirus. However, both detection techniques need to convert the RNA of the coronavirus into DNA in order for it to be amplified. In the SHERLOCK test, it has to be transcribed back into RNA to be detected, thus adding a small step to the process.
Chen and Harrington rushed to get a white paper online with the details of their Mammoth test. In many ways it was similar to the SHERLOCK process. All that was necessary was a heating block, the reagents, and paper flow strips to give a readout of the results. Like Zhang, the Mammoth team decided to put what they had devised into the public domain, to be shared freely.
On February 14, while they were preparing to put their white paper online, Chen and Harrington saw a message pop up on the Slack channel they were using. Someone posted the tweet that Zhang had just sent out announcing that he had just published his white paper on how to use the SHERLOCK protocol for detecting the coronavirus. “We were like, ‘Oh, shoot’ ” Chen recalls of that Friday afternoon. But after a few minutes, they realized that having both papers appear was a good thing. They appended a postscript to the paper they were just about to post. “While we were preparing this whitepaper, another protocol for SARS-CoV-2 detection using CRISPR diagnostics (SHERLOCK, v.20200214) was published,” it said. They then included a useful chart comparing the workflows of the two techniques.5
Zhang was gracious, though it was easy for him to be since he had beaten the Mammoth team by a day. “Check out the resource provided by Mammoth,” he tweeted, including a link to its white paper. “Glad that scientists are working together and sharing openly. #coronavirus.”
That tweet reflected a welcome new trend in the CRISPR world. The passionate competition for patents and prizes had led to secrecy about research and the formation of competing CRISPR companies. But the urgency that Doudna and Zhang and their colleagues felt about defeating the coronavirus pushed them to be more open and willing to share their work. Competition was still an important, and useful, part of the equation. There continued to be a race between Doudna’s world and Zhang’s to publish papers and make advances on the new COVID tests. “I’m not going to sugarcoat it,” Doudna says. “There’s definitely competition going on. It makes people feel an urgency to move ahead or, if they don’t, other people are going to get to something first.” But coronavirus made the rivalry less cutthroat, because patents were not a paramount concern. “The awesomely good thing about this terrible situation is that all the intellectual property questions have been put aside, and everyone’s really intent on just finding solutions,” says Chen. “People are focused on getting something out there that works, rather than on the business aspect of it.”
At-home tests
The CRISPR-based tests developed by Mammoth and Sherlock are cheaper and faster than conventional PCR tests. They also have an advantage over antigen tests, such as the one developed by Abbott Labs that was approved in August of the plague year. The CRISPR-based tests can detect the presence of the RNA of a virus as soon as a person has been infected. But the antigen tests, which detect the presence of proteins that exist on the surface of the virus, are most accurate only after a patient has become highly infectious to others.
The ultimate goal for all of these methods was to create a CRISPR-based coronavirus test that would be like a home pregnancy test: cheap, disposable, fast, and simple, which you could buy at the corner drugstore and use in the privacy of your bathroom.
Harrington and Chen of the Mammoth team unveiled their concept for such a device in May 2020 and announced a partnership with the London-based multinational pharmaceutical company GlaxoSmithKline (maker of Excedrin and Tums) to manufacture it. It would provide accurate results in twenty minutes and require no special equipment.
Likewise, Zhang’s lab that same month developed a way to simplify the SHERLOCK detection system, which originally required two steps, into a process that required just a single-step reaction. The only equipment necessary was a pot to keep the system heated at a steady 140 degrees Fahrenheit. Zhang named it STOP, for SHERLOCK Testing in One Pot.6 “Let me show you what it will look like,” Zhang says to me with his boyish enthusiasm as he shares slides and renderings on a Zoom call. “You just put a nasal or saliva sample into this cartridge, slide it into the device, break one blister to release a solution that will extract the virus RNA, and then break another blister that will release some freeze-dried CRISPR for a reaction in the amplification chamber.”
