CRISPR-based tool solves genetic mystery 80 million years in the making

We don’t have much in common with mice. After 80 million years of diverging evolution, seeing any similarities in our DNA takes some squinting. So it seems obvious that any bits of DNA that have resisted the forces of evolution and remained identical between mice and humans (and rats and pufferfish) must be critical. Probably even essential. Known as ultraconserved elements, these sections of DNA drew immediate scientific attention when the first human genome sequence was released in 2003. And while scientists have discovered some molecular functions for these elements, they have been failing to show their essentiality ever since.

Until now.

In a new study published today in Nature Genetics, scientists at Fred Hutchinson Cancer Research Center showed that yes, these ultra-conserved DNA elements are essential. Looking at a subclass of ultraconserved elements known as poison exons, the investigators found that certain poison exons were essential for cell growth while others acted to suppress the growth of lung tumor cells in mice. By showing that ultraconserved elements operate on a cellular level, these discoveries help shed light on why they have remained unchanged over millions of years.

This is the “first study finding large-scale importance of these highly conserved [DNA] elements,” said Dr. Rob Bradley, a computational biologist and the study’s senior author.

Ultra-conserved, but function unclear

The 481 ultra-conserved DNA elements shared between mice, rats and humans immediately jumped out once these three genomes were sequenced and compared.

The finding was “really odd, really interesting,” Bradley said. “Everybody assumed these pieces [of DNA] were super important.”

But initial attempts to reveal their obviously oh-so-crucial function came up empty. Scientists removed ultra-conserved elements from mouse genomes — and produced perfectly healthy mice.

“But they don't seem to matter, even though they're the most conserved things in our genome,” Bradley said. “It was really surprising. These previous studies were extremely good.”

Over the years, researchers have revealed some modest roles for ultra-conserved elements, but nothing like the dramatic effects they’d initially expected.

As it happens, there are different kinds of ultra-conserved elements. The type that most scientists had focused on were called enhancers. These are stretches of DNA that help turn genes on or off. Researchers had also noticed that ultra-conserved DNA elements often overlapped with stretches of DNA that encode so-called “poison” exons.

Genes are like recipes for proteins. Between DNA and protein is an intermediary molecule, RNA, that carries a gene’s protein-making instructions to the cell’s protein-making factories. But like a pie that doesn’t need cinnamon, or would be just as tasty with less sugar, proteins can be tweaked. The recipe “steps” that can be swapped in or out are called exons. Cells can skip exons by snipping them out of an RNA molecule and then pasting (also known as splicing) that gap back together.

Poison exons act as a kill switch for protein production. It’s as though, instead of directing the cook to add the cinnamon right after the sugar, the pie recipe read, “Add sugar. Now throw everything away.”

Poison exons are helpful. Sometimes certain proteins shouldn’t be made, so these exons poison the protein for the greater good of the cell.

But it seems counterintuitive that poison exons would be so important that they’d be able to resist millions of years of evolution. It’s tricky to estimate exactly how much a given exon should have changed as mice and humans diverged (time is only one factor that shapes evolution), but a rough estimate would be about 80 changes to an exon that’s 200 DNA “letters” in length, Bradley said.

RNA splicing, which can dramatically alter how proteins are made — and which proteins are made — is a major focus of Bradley’s research group. Many different proteins regulate RNA splicing, but all 12 genes of a specific family of RNA-splicing-regulating proteins known as SR proteins contain ultra-conserved poison exons. Though previous studies had revealed that poison exons act on a molecular level to help regulate amounts of these proteins, Bradley suspected ultra-conserved poison exons in SR protein genes and other genes functioned on a cellular level. No one had yet tested whether these DNA elements were necessary for cell survival.

Adapting CRISPR for RNA splicing in high throughput

Prior studies of ultra-conserved DNA elements, including those looking at enhancers, had taken an old-school tack: choose one and go deep. Bradley and Dr. James Thomas, the postdoctoral fellow in Bradley’s lab who spearheaded the work, decided to take advantage of the emerging field of CRISPR-based genome editing to cast a wide net. They developed a tool that allowed them to examine the effect of removing hundreds of poison exons at the same time, a type of experiment known as a high-throughput screen.

The tool uses two molecules known as guide RNAs to direct the CRISPR-based genome editor to the right stretch of DNA that needs snipping. Thomas and Bradley designed these molecules to direct CRISPR to remove poison exons. They dubbed the approach paired guide RNAs for alternative exon removal, or pgFARM (pronounced pig farm).

PgFARM builds off prior approaches, Bradley noted.

“What’s unique about what James did is not the approach per se, but that he’s doing it in high throughput,” he said.

Important roles for poison exons

Thomas and Bradley focused on 465 highly conserved poison exons and compared them to 91 poison exons that are not highly conserved. Using pgFARM, Thomas first tested the sequences’ importance in cells growing in lab dishes.

“When you remove certain [ultraconserved] poison exons, the cells end up dying,” Thomas said.

This was exactly what researchers originally investigating ultra-conserved enhancers had expected — but failed — to see, making it a both expected and unexpected finding.

It was the high-throughput nature of his approach that made it clear that he was on to something when he first assembled the data, Thomas said.

“In the past, people have really just focused on one [ultraconserved element]. If you just look at one point, you say maybe there’s an artifact [in the experiment],” Thomas said. But he was able to see that many ultraconserved poison exons had the same effect on cells — the screen gave him the big picture that a one-by-one approach could not.

Thomas suspected that removing poison exons from cells living in an animal, not just a lab dish, would allow him to tease out different, and possibly stronger, effects.

Bradley’s Hutch colleague, lung cancer researcher Dr. Alice Berger, helped the team test the role of poison exons in lung tumor cells. The team used mice with lung adenocarcinoma to screen the same poison exons and found that a subset of them helped promote tumor growth.

Many of the anti-tumor poison exons lie within the genes for RNA splicing factors, some of which are produced at higher-than-normal levels in cancer. The findings suggest that some poison exons help rein in RNA-splicing factors that could otherwise promote tumor growth.

Future directions

Many questions about highly conserved poison exons remain. Thomas is planning to further test whether any are important for early embryonic development, the ultimate test of a DNA element’s “essentiality.”

He’s also excited about the potential for poison exon-targeted anti-cancer therapeutics. RNA mis-splicing is at the root of diseases including several muscular dystrophies, certain neurological diseases and many cancers — and small-molecule drugs to counteract mis-splicing are already in clinical trials. Success in these trials would be good news for anyone hoping to target a similar problem in cancer.

“Once you find a target, it's easy to build upon the previous technology that exists for therapeutics,” Thomas said.

The team also believes that pgFARM is just the beginning of genetic screens for investigations into RNA splicing. The method could easily be adapted to study the many different types of RNA splicing that occurs in cells, as well as specific diseases and different biological systems, Thomas said.

“I can't possibly envision [pgFARM] being more flexible, easy, cheap … and just robust,” Thomas said.