A Crack In Creation: A Nobel Prize Winner's Insight into the Future of Genetic Engineering

by Jennifer A. Doudna

In RNA interference, animal and plant cells form RNA-RNA double helixes to destroy invading viruses.

In much the same way, CRISPR RNA molecules might target phage RNA during an immune response by using RNA-RNA double helixes.

I was fascinated by the added possibility that, unlike in RNA interference, CRISPR RNAs might be able to recognize matching DNA too—a power that would enable this weapons sy...

Northwestern University, Luciano Marraffini and his mentor Erik Sontheimer, a colleague I knew from his student days at Yale, figured out that CRISPR RNA co...

elegant experiments to prove that CRISPR RNAs target the DNA of invading genetic parasites.

likely relied on base-pairing interactions—the

studies to a broad, unified theory about the inner workings of a microbial adaptive immune system.

intricate bacterial defense system.

a few researchers around the world weaving the fabric of what would eventually become the vast tapestry of the CRISPR field, with all of its applications and implications.

We were amazed to see just how diverse CRISPR was. In 2005, researchers had identified nine different types of CRISPR immune systems. By 2011, that number had decreased to three—but within these basic types there were thought to be ten subtypes. And by 2015, the classification would change yet again to include two broad classes comprising six types and nineteen subtypes.

In Type I systems like E. coli and P. aeruginosa, the Cas3 enzyme—that motorized hedge clipper—chewed the DNA to shreds. It wasn’t even possible to see the DNA destruction in action because this tiny machine mowed through it so rapidly; when we tried to observe the reaction in a test-tube experiment, all we could see was molecular chaos,

The Type II system found in S. thermophilus, by contrast, was more restrained and precise.

The more I read, the clearer it was that the Cas9 protein was likely to be a key player in the DNA destruction phase of the immune response in Type II CRISPR systems.

the Type I CRISPR systems we had been studying in the lab, the CRISPR RNA assembled with a multitude of Cas proteins to form a DNA-binding-and-cutting machine.

At fifty letters long, the DNA double helix would be just seventeen nanometers, or seventeen-billionths of a meter, long,

Looking at these results, we realized that we had defined the essential parts of the DNA-cutting machine, the mechanism that allowed S. pyogenes and S. thermophilus— and any other bacteria with a similar type of CRISPR system—to not only target specific phage DNA sequences, but also to destroy them. The crucial components for DNA cutting were the Cas9 enzyme, the CRISPR RNA, and the tracrRNA.

They found that Cas9 could latch onto a DNA double helix, pry open the two strands to form a new helix between the CRISPR RNA and one strand of DNA, and then use two nuclease modules to simultaneously slice through both strands of the DNA, creating a double-strand break. Depending on the sequence of its associated RNA molecule, Cas9 could target and cut virtually any matching DNA sequence. In effect, the CRISPR RNA molecule acted like a set of GPS coordinates, guiding Cas9 to a precise spot within the vast expanse of a long DNA molecule according to the matching letters in the CRISPR RNA and DNA.

For any twenty-letter sequence the guide RNA contained, Cas9 would find its matching counterpart in DNA and then cut.

In the warfare waged between bacteria and viruses, Cas9’s function made perfect sense. Armed with a cache of RNA molecules derived from the CRISPR array, where snippets of phage DNA had been stored, Cas9 could readily be programmed to slice up corresponding sites within viral genomes. It was the perfect bacterial weapon: a virus-seeking missile that could strike quickly and with incredible precision.

That single step—programming the CRISPR-Cas9 machine ourselves—actually consisted of two smaller steps: developing an idea, then performing an experiment.

the CRISPR RNA molecule that did the targeting, and the tracrRNA molecule that held it and Cas9 together—to

If we fused the tail of one to the head of the other, the resulting chimeric RNA, if functional, would simplify our programmable DNA-cutting machine; instead of having to combine Cas9 with both RNA molecules, the guide (CRISPR RNA) and the helper (tracrRNA), we’d be able to pair the enzyme with a single RNA molecule—a single-guide RNA—that did both jobs.

Out of this fifth bacterial weapons system, we had built the means to rewrite the code of life.

genome-editing applications.” On June 8, 2012, a sunny Friday afternoon, I clicked Confirm on my computer, formally submitting our paper for consideration to the journal Science. It would be published just twenty days later, on June 28, and nothing after that would ever be the same—not for me, not for my collaborators, and not for the field of biology.

In November 2013, we founded Editas Medicine with $43 million in financing from three venture capital firms. Just a half a year later, Emmanuelle co-founded another company, CRISPR Therapeutics, with an initial $25 million bankroll, and in November 2014, a third company, Intellia Therapeutics, joined the scene with $15 million in Series A funding.

I could already see a new era of genetic command and control on the horizon—an era in which CRISPR would transform biologists’ shared toolkit by endowing them with the power to rewrite the genome virtually any way they desired.

the genome would become as malleable as a piece of literary prose at the mercy of an editor’s red pen.

in our 2012 Science article, Martin and Krzysztof had demonstrated something groundbreaking: that a CRISPR-associated protein called Cas9, isolated from flesh-eating bacteria, worked with two molecules of RNA to target matching twenty-letter DNA sequences and cut them apart. The RNA acted like a guide, dictating the GPS coordinates of the attack, and Cas9 acted like the weapon to eliminate the target.

this CRISPR machine was mobilized to slice up and destroy specific DNA molecules from the virus as part of an adaptive immune response.

Like us, they found that Cas9 cut apart DNA sequences that matched the letters of the CRISPR RNA.

We had also proposed that this defense system could be repurposed for a different function inside cells, not to destroy viral DNA, but to precisely edit the cell’s DNA.

If we changed the twenty-letter RNA code to match the sequence of a specific human gene and then transplanted Cas9 and the new guide RNA into human cells, CRISPR would make a surgical cut in the targeted gene, marking that site for repair.

Our own research had revealed that the Cas9 protein and its guide RNA were very picky about their partners and stuck together tightly, indicating they should have no problem finding each other inside a human cell.

for sending them into the cell’s nucleus, where the DNA is located, we could simply provide a chemical zip code that would let the cell do the work for us.

Martin began by transferring the bacterial DNA encoding Cas9 and the CRISPR-derived RNA into two plasmids, little ringlets of DNA that act like artificial mini-chromosomes. The first plasmid contained genetic instructions for the guide RNA as well as separate instructions that directed human cells to produce gobs of it. The second plasmid contained the cas9 gene, but it had been “humanized” so that it could be interpreted by protein-synthesizing factories inside human cells.

Martin also fused the cas9 gene to two genes, routinely used by biologists, that coded for other proteins: a tiny one, called a nuclear localization signal, which directed a protein to the cell’s nucleus, and the green fluorescent protein that would cause any human cells successfully producing Cas9 to fluoresce green when exposed to ultraviolet light.

combining all these molecular parts, Martin and I intended to convert human cells into CRISPR-producing factories that unwittingly churned out molecules progr...

Thus, in the most basic scenario, if CRISPR succeeded in cutting a gene, the cell would respond by simply gluing the DNA back together, much like welding two pieces of metal pipe together.

Scientists refer to this process as nonhomologous end joining, since, unlike homologous recombination, the mending does not involve a matching repair template. (Homologous derives from the Greek homologos, meaning “agreeing.”) A key property of this repair process is its inherent sloppiness. Just as a welder needs to be sure that the two pipes have clean edges before he or she joins them, the cell needs to ensure that the broken pieces of DNA have clean ends before putting them back together. Generating clean ends sometimes involves deleting or inserting a few letters of DNA, which results in permanent genetic changes after the repair process is completed. This meant that a gene would most likely be altered after being targeted by CRISPR, sliced apart, and repaired by the cell. This messy, error-prone repair would give Martin and me a simple way to detect successful gene editing. By targeting a specific gene and analyzing its DNA sequence letter by letter before and after CRISPR treatment, we would find any signs of sloppy repair, thereby proving that CRISPR had located and cut its target.

the real reason that CRISPR exploded onto the biotech scene with such force and vitality was its low cost and ease of use. CRISPR finally made gene editing available to all scientists. Previous tools—primarily ZFNs and TALENs—were difficult to design and prohibitively expensive. For this reason, many labs, including my own, were unwilling to take on the challenges of research using gene editing. With CRISPR, however, scientists can easily design a version to target their gene or genes of interest, prepare the requisite Cas9 protein and guide RNA, and execute the experiments themselves using standard techniques, all within mere days and without requiring any outside assistance.

The only thing necessary to get started is a copy of the basic CRISPR-containing artificial chromosome, or plasmid. This need has been conveniently met on a massive scale by the nonprofit organization Addgene, a highly successful and ever-expanding plasmid repository and plasmid-distribution service.

Addgene keeps careful track of the plasmids it has on file, advertises the plasmids and their exact specifications on a website, and generates thousands of duplicate copies that can be distributed to eager customers. The cost to academic laboratories: in 2016, just sixty-five dollars per plasmid.

In 2015 alone, Addgene shipped some sixty thousand CRISPR-related plasmids to researchers in over eighty different countries.

empirical data from the scientific literature on what kinds of targeting sequences work better than others, various software packages offer researchers an automated, one-step method to build the best version of CRISPR to edit a given gene.

some of the most complex and sophisticated gene-editing experiments to date: the design and execution of genome-wide screens, in which CRISPR is exploited to edit every single gene in the genome.

Some experts have suggested that, with today’s tools, anyone can set up a CRISPR lab for just $2,000.

CRISPR was even the star of a crowdfunded venture that raised well over fifty thousand dollars to generate and distribute DIY gene-editing kits. For $130, donors received “everything you need to make precision genome edits in bacteria at home.”

The CRISPR Menagerie

TOMATOES THAT CAN SIT in the pantry slowly ripening for months without rotting. Plants that can better weather climate change. Mosquitoes that are unable to transmit malaria. Ultra-muscular dogs that make fearsome partners for police and soldiers. Cows that no longer grow horns.

they already exist, thanks to gene editing.