Columbia University Medical Center

Drawing Back DNA “Curtains” on New Gene-Editing Method

DNA curtains – devised by P&S professor Eric Greene – provide a new way to study DNA-protein interactions. DNA strands are green; proteins are pink. Image: Eric Greene.

DNA curtains—devised by P&S professor Eric Greene—provide a new way to study DNA-protein interactions. DNA strands are green; proteins are pink. Image: Eric Greene.

Using a dazzling technology to watch proteins collide, clutch, and slide along strands of DNA, researchers at Columbia University College of Physicians & Surgeons and UC-Berkeley report online in Nature that they have uncovered some of the secrets behind a powerful new genetic engineering technique.

The CRISPR technique is a faster, cheaper, and more precise method of inserting new genes into DNA, and the new findings may allow researchers to expand and improve its use in areas such as gene therapy.

Greene_cover-natureLike many laboratory techniques used to cut and paste DNA, the CRISPR system was not invented by scientists but discovered in bacteria. It turns out that bacteria can steal small snippets of viral DNA and then retain these fragments like a library representing the complete history of viruses to which the bacteria have been exposed in the past.

If a virus ever tries to come back, the bacteria make RNA copies of these DNA sequences. In CRISPR, a protein called Cas9 then grabs the bits of RNA and uses them like a fingerprint to identify the viral invaders. The Cas9-RNA complexes then patrol the genome, sniff out invading viruses, and destroy the viral DNA.

The high specificity and programmability of Cas9 has allowed scientists to co-opt CRISPR systems as a powerful new tool for genome engineering to delete any gene they desire in any type of cell, as well as to turn genes on and off. Yet concerns remained that the system may have unwanted effects in other parts of the genome, essentially because many details of how CRISPR works are still unknown.

We’re watching proteins in real time as they’re doing their real jobs, instead of relying on more traditional techniques that may only indirectly suggest what’s going on.

To get a clearer picture of how Cas9 finds and removes viral DNA, Jennifer Doudna of UC-Berkeley and her graduate student Samuel Sternberg turned to Eric Greene, PhD, associate professor of biochemistry and molecular biophysics at P&S, and his graduate student Sy Redding.

In Eric Greene’s lab, biochemistry is not just done in the traditional way, with activity assays and gels, but also with nano-fabricated DNA “curtains” that let the scientists watch biochemistry take place before their own eyes.

“We call it visual biology, or visual biochemistry, because we can see at the molecular level what these things are doing,” Greene says. “We’re watching them in real time and seeing them as they’re doing their real jobs, instead of relying on more traditional techniques that may only indirectly suggest what’s going on.”

To imagine what a DNA curtain looks like, think no further than curtains made of strands of plastic beads. But instead, DNA curtains are made of strands of DNA hanging side-by-side from a glass slide.

In DNA curtains, DNA strands are labeled and proteins are pink. Image: Eric Greene.

In DNA curtains, DNA strands are green and proteins are pink. Image: Eric Greene.

Greene developed DNA curtains as an assistant professor at Columbia, while trying to watch proteins perform their jobs on a single strand of DNA. With only a single strand, he could spend a full day in the lab and end up with just three data points. But by hanging thousands of DNA strands next to each other from a glass slide, it became possible to watch thousands of separate protein-DNA interactions in just a few hours.

With the DNA curtains, one of the lab’s main interests is understanding how proteins search for and identify specific sites in DNA. In the CRISPR research, Greene and Redding simply adapted their DNA curtain technology to investigate the actions of Cas9.

The theory was that the Cas9 complex would grab onto the DNA and then slide along the strand, looking for its target. “But what we saw was pure, random three-dimensional diffusion,” Greene says. “Cas9 collides with the DNA randomly, but it is looking for something very specific during these collisions.”

(Watch an animation, created by Greene Lab Studios, of Cas9 searching DNA)

As Cas9 collides with the viral DNA, it searches for a specific 3-nucleotide sequence called a “PAM.” If the short PAM sequence isn’t present, Cas9 quickly detaches from the DNA strand and floats around until it crashes into a different spot, and the process is repeated.

Once Cas9 finds a PAM sequence, the complex latches onto the strand a little longer and opens the DNA a bit to see if it matches the RNA bound to the Cas9 protein. If it matches, Cas9 clamps down on the DNA and cuts the strand.

“It’s basically doing a two-part search, first looking for the short PAM sequence and then asking whether the adjacent DNA matches the virus that it’s trying to destroy. This two-part mechanism sets Cas9 apart from any other target search we’ve looked at,” Greene says.

You can understand the process by thinking about the way you find your favorite pair of red socks in the morning.

Though the sliding mechanism was believed to be the quickest-possible search mechanism, Redding explains that, “you can understand how a two-part search speeds the process by thinking about the way you find your favorite pair of red socks in the morning.”

“You don’t randomly open all of your dresser drawers until you happen to find your red socks. You just open your sock drawer and pick out your favorite pair of red socks. Cas9 is essentially doing the same thing: First it looks for the PAM (the ‘sock drawer’), then it asks whether the sequence next to the PAM is correct (the ‘red socks’). This lets Cas9 avoid wasting time searching through the entire genome by restricting its search to sites that have PAM sequences.”

For scientists using CRISPR for genetic engineering, the findings are good news, because they show that the system will target only genes next to the short PAM sequence. Even if the sequence targeted for removal by CRISPR is present in multiple places in the genome, Greene says, researchers need to consider only those lying next to the PAM sequences.


The study was supported by the National Science Foundation (MCB-1154511, MCB-1244557), National Defense Science & Engineering Graduate Research Fellowship program, NIH (GM074739), and the Howard Hughes Medical Institute.

Also on CUMC News: