The Regulation Game: The key to controlling cellular microRNA levels

What can viruses, culprits of the flu, or even worse pathogens teach us about our biochemistry? For decades, viruses have brought scientists from a wide swath of specialties together and proven to be a hotbed of discovery because their relatively tiny genome can only encode the most crucial functions—such as the regulation of gene expression.

Due to their genetic efficiency, viruses are ideal for the study of non-coding RNA, a powerful multi-purpose tool. As the result of various structural features, RNA serves the unique dual role of both carrying genetic messages and regulating them. A form of small RNA called microRNA (miRNA) plays a key role in regulating the levels of messenger RNA (mRNA). But in the two decades since the discovery of miRNA, the question has remained: what regulates the regulator? This question so tantalizing that it united RNA biochemists, crystallographers, and a tenacious virologist with a common goal: to reveal the mechanism for regulating the levels of miRNA.

From both the Steitz lab in Yale’s Molecular Biophysics and Biochemistry Department and the MacRae lab at The Scripps Research Institute, researchers recently found the key mechanism for target-directed miRNA degradation (TDMD), in which the RNA target of miRNA bites back, causing the degradation of its would-be “assassin.” Through a series of RNA and protein mutation experiments, Yale postdoctoral associate Paulina Pawlica and her collaborators pinpointed sequence and structure requirements for miRNA degradation. Their findings reveal the molecular mechanism of how cells regulate their own genetic regulators.

A Fine-Tuned Regulator

As genetic information is transmitted from DNA through RNA to protein, the greatest regulation occurs at the level of mRNA. By dynamically controlling the levels, localization, and even the mRNA code of proteins, a cell can quickly fine-tune the amounts of different proteins in response to its changing needs. The best molecule for regulating RNA is, in fact, itself. The mRNA of most mammalian genes are regulated by miRNA in complex with an Argonaut (Ago) protein. Ago exposes the miRNA’s “seed sequence,” which binds to a complementary sequence on the target mRNA. Once a target is bound, the Ago protein may recruit RNA-degrading or suppressing enzymes, or it may cleave the RNA itself. The outcome depends on which Ago protein is involved and, for reasons yet unknown, on the strength of complementarity between the miRNA and its mRNA target.

Among RNAs, miRNA enjoys unusual longevity. Typically, an mRNA has an extended protective tail that eventually gets chewed away, exposing the mRNA to degradative enzymes. In contrast, the tail of miRNA is sheltered in binding pocket domains of its partner Ago protein, such as the domain known as PAZ. This means that a specific regulation mechanism must be in place to degrade miRNA and inversely control the level of its target mRNA. In other words, killing the assassin allows more targets to live. Viruses were the gateway to discovering this mechanism. “Viruses end up being sort of a gold mine for giving you an entrée into what cells are doing because they’re completely dependent on their host cells,” Steitz said. If a virus is hijacking host cell proteins to do something, the host is probably doing it too.

An Unexpected Journey

Pawlica never expected to work on RNA biochemistry. Her passion is viruses, and she joined Steitz’s lab to investigate an intriguing viral non-coding RNA process. Usually, when a miRNA/Ago complex binds to an RNA target, the target gets degraded or suppressed. However, with one viral RNA known as HSUR1, the opposite occurred: instead of HSUR1 being degraded, its miRNA assassin was itself degraded. HSUR1 seemed to turn cellular RNA regulation against itself. More “backfiring” miRNA targets were soon discovered, including in mammals, suggesting that target-directed miRNA degradation (TDMD) was widespread.

Pawlica first used a bioinformatic approach to decipher the TDMD mechanism. By mutating HSUR1, Pawlica found that the target RNA had to fulfill three requirements: it had to be complementary to the miRNA’s seed sequence (as with all other miRNA-mRNA pairs); it had to have a short middle stretch of non-complementarity; and, unlike typical targets, it had to be complementary to the tail of the miRNA. Using these criteria, Pawlica then screened a set of viral genomes to identify other “anti-targets”. However, she got a bewildering number of matches, most of which were not successful TDMD targets. “In one virus I had hits for almost half of its micro-RNA’s…without additional information like structures you can’t go too far,” Pawlica said. Then, a structural gift dropped from the heavens.

“I learned of Paulina’s project after running into [Steitz] at an RNA conference in Heidelberg,” said Ian MacRae, a structural biochemist at Scripps, whose lab specializes in X-ray crystallography of miRNA bound to Ago protein. While Pawlica puzzled over the breadth of her results, Jessica Sheu-Gruttadauria—then a graduate student in MacRae’s lab—was pondering an unusual miRNA-target-protein structure. “TDMD was on the list of possibilities, but seeing Paulina’s data made it very clear,” MacRae said. Sheu-Gruttadauria’s structure showed a striking feature: the tail of the miRNA, rather than being safely tucked in its Ago partner’s PAZ domain, was lured out of hiding to pair with the TDMD RNA.

Small Mutation, Big Results

If miRNA stability is determined by whether its tail is in the Ago PAZ domain, then altering the PAZ domain to release the tail should also lead to miRNA degradation—and that is exactly what happened. “That was very satisfying, sort of the nail in the coffin,” Steitz said. Sheu-Gruttadauria then worked to obtain crystal structures of Pawlica’s HSUR1 constructs. The resulting structures provided mechanistic links between the sequence criteria Pawlica uncovered and the process of TDMD.

The miRNA seed sequence, the primary requirement for target binding, formed strong interactions with both the target and the Ago2 protein. In contrast, the sequence from TDMD RNA complementary to the miRNA tail destabilized the interaction between the miRNA and the Ago2 PAZ domain, pulling the tail out of the protective region. Hypothesizing that miRNA-regulating enzymes could bind to this exposed tail, MacRae’s lab decided to test whether this were spatially possible with computer modeling. This binding appeared to be feasible.

The non-complementary region, by bulging out the paired RNAs in the central cleft of Ago2, seemed to be protecting the TDMD RNA from being sliced—a molecular shield. The unstructured region was also found to be crucial to the tail-exposing mechanism. However, it is still unclear how the three regions work together to induce TDMD: whether miRNA tail release causes Ago’s central cleft to open, or vice versa. “I suspect that the two processes go hand-in-hand,” MaRae said. Knowing how the mechanisms are connected would require a new set of experiments. “Crystal structures tell you the starting and ending states, but not how you got from one to the other,” Steitz said.

Future Directions

With a tentative mechanism in hand, Pawlica hopes to find further distinguishing features of TDMD RNA’s. “This simple complementarity…would occur much too often to be biologically relevant,” Pawlica said. She also hopes to learn more about the proteins involved in TDMD—some of which regulate miRNA by modifying rather than degrading it, depending on complementarity of the target RNA. With a growing list of TDMD targets being discovered, the implications of the study may eventually lead to treatments for the many diseases governed by miRNA levels.

This project has been a great undertaking, from transitioning from viruses to RNA to trying to narrow down a massive bioinformatic screen to uniting distinct scientific perspectives to tell a coherent story. “It hasn’t been that easy, actually,” Pawlica said. But having two fronts of evidence for a new mechanism and reaping the fruits of this collaboration seem worth the effort. “Isn’t it beautiful the way the structure really tells you what’s happening?” Steitz said. “Then you can go back and make mutations to prove it…and it all becomes clear how it works.”