[ Winter 2013 / Vol. 26 No. 1 ]

For a cell, the past informs the present. We humans have search engines like Google and Yahoo to sift through the Internet’s gobs of historical information and learn from others’ mistakes and successes. In some cells of the worm Caenorhabditis elegans, it turns out, a type of RNA, called Piwi-interacting RNA (piRNA), and its partner, an Argonaute protein called Piwi, run a similar search. The piRNA and protein continuously peruse the cell’s library of data and detect how it previously dealt with a particular molecule—whether an invading virus or a cell’s own genetic material.

That historical review determines the cell’s next step, according to research by HHMI investigator Craig Mello of the University of Massachusetts Medical School. piRNA—short stretches of 21 to 36 nucleotides—exists in tens of thousands of different sequences in germ cells (those that become eggs and sperm) of many animals including humans, but scientists have struggled to understand its purpose. Mello’s lab has now uncovered the reason the molecules are so ubiquitous and exist in so many forms in C. elegans: so they can pair with essentially any genetic sequence they encounter during their endless scanning.

“The Internet is full of information, yet we can navigate it pretty efficiently because we have search engines,” says Mello. “Argonaute proteins are like cellular search engines, and their small RNA cofactors are like the short search queries we type into Google.” The piRNAs have a huge capacity for scanning, explains Mello, because they allow imperfect pairing—like when you misspell a search query, but still find what you’re after. In short, Mello has discovered, the piRNA-based search engine can find everything out there, like running many millions of two- or three-word searches in Google to assemble every Web page on the Internet.

Mello’s findings suggest that to make sense of this massive search process, the piRNA system interfaces with two other pathways that serve as cellular memories of “self” and “non-self” RNA. Sequences that were seen before in a previous generation (“self” RNAs) are thought to be protected from piRNA silencing by a pathway involving an Argonaute protein called CSR-1. Sequences not seen before lack this pathway’s protection and so they are recognized by the Piwi/piRNA complex. In this case, the complex recruits a different Argonaute, dubbed WAGO, to create a permanent memory of the “non-self” RNA sequence.

The Mello lab stumbled upon this system while attempting to introduce foreign genes into C. elegans. When scientists insert a new gene into the worm’s germline—the genetic material passed down to future generations via germ cells—the gene is sometimes expressed and sometimes silent. Mello’s lab group wanted to know why expression is so unpredictable, even when genes are inserted into exactly the same corresponding spot in the genome.

To follow expression, the team attached coding sequences for jellyfish green fluorescent protein (GFP) to the foreign genes. By identifying worms whose germlines glowed green, they could easily see whether the associated gene was silenced. Then, they began crossing worms—those that had silenced the gene and those that hadn’t—with each other. In the resulting offspring, they expected to see half the brightness of the green fluorescence—the gene inherited from one parent would be on, the second gene off. But the germlines of cross progeny were dark.

“You get this transfer of silencing from one copy of the gene to another,” says Mello. “And this was permanent, very stable silencing.” Even in future generations, they found, GFP was always turned off.

However, this is only half of the story, Mello says. Equally remarkable was the observation that, in another line of worms, active versions of the engineered genes had become resistant to the transfer of silencing over time and instead activated the silent genes. These observations suggest that in some cell lines, silencing trumps, while in other cases, the on-switch prevails. Importantly, the researchers found, once an on-or-off decision was made, it held true for every descendant of that animal for generations. “The animal is actually remembering which genes are supposed to be on and which genes it wants off,” says Mello.

The connection with piRNA came when Mello’s team repeated the experiments on cells that lacked Piwi, the protein that binds to piRNA molecules, and found that genes that had been silenced in other iterations of the tests were now always turned on. Through a series of experiments, the scientists provided evidence that piRNAs are forever scanning every bit of free RNA in germ cells. They also showed that molecular memories were maintained through two groups of molecules: one that signals activation, another silencing. Both rely on different Argonaute proteins to establish the memories.

When a piRNA successfully binds to RNA, the attached Piwi protein kicks into action: If the RNA sequence has not been seen before, the Piwi turns on a molecular “non-self” pathway that enforces silencing. If the RNA sequence has been seen before, it is recognized by the “self” Argonaute pathway, and Piwi allows the cell to express the gene. The findings were published in two papers in the July 6, 2012, issue of Cell.

While an animal’s immune system is built to recognize foreign particles or cells, the piRNA system offers a second level of protection at the genetic level. If a virus, for example, gets past a worm’s immune system, it can insert its genes into the worm’s genome. The piRNA system can ensure that those viral genes remain turned off—it’s a second line of defense. “Instead of recognizing a structural feature as foreign, the animal is looking at thesequence information itself,” says Mello.

But the system—which Mello has dubbed RNAe, for RNA-induced epigenetic silencing—could also be a way that organisms generate heritable diversity that can be acted on by natural selection. Epigenetic silencing is any form of inherited genetic regulation that allows two organisms with the same set of genes to express those genes differently. Most previously known epigenetic mechanisms are based on protein modifications or chemical tags on chromatin or DNA, but RNAe is based on an inherited RNA signal and is the first epigenetic mechanism discovered to scan new genes by comparing them with a memory of self RNA expression.

“The bottom line,” Mello says, “is that cells appear to have a previously unappreciated level of information technology sophistication, including both an actual memory of every gene that’s been expressed and a constant surveying of information to keep track of what’s new.”