ABOVE: A ribosome (purple) translating mRNA (orange) into protein (red) © ISTOCK.COM, SELVANEGRA

When it comes to controlling the amount of protein made based on a gene’s code, cells have a variety of tools at their disposal: attaching epigenetic marks to a gene can make it less accessible to be copied into mRNA, for example, or cells can make more or fewer of the transcription factors that get the process started. Now, a paper published on January 17 in Cell Systems uses a novel method to zero in on a lesser-studied control mechanism: small changes in an mRNA’s start region that affect how efficiently the transcript attracts ribosomes, and thus how much protein is churned out. 

“This study uses an ingenious way of being able to monitor ribosome binding” to the start regions, Maria Barna, a geneticist at Stanford University who was not involved in the study, tells The Scientist. Using this method, the researchers “identified a whole slew of potential cis-regulatory elements, which diversify gene expression very profoundly, some over 1,000-fold. This was very exciting and unexpected.” 

Previous studies had established that translational activity can vary between mRNAs and that changes in the start area of these transcripts, known as 5-prime untranslated regions (5'UTRs), may be associated with diseases. Wendy Gilbert, a molecular biochemist at Yale University and lead author of the study, had a “long-standing interest in how the regulatory information at the five-prime end of a messenger RNA controls gene expression at the level of translation,” she tells The Scientist. Her lab had tried several methods to try to get at this question, culminating in the newly published method, dubbed direct analysis of ribosome targeting (DART). 

For DART, numerous synthetic full-length 5'UTRs are incubated in cell extracts, which contain the necessary factors for translation to start, including ribosomes. After the 5'UTRs and ribosomes are left to mingle for 30 minutes, the mixture is centrifuged, causing 5'UTRs that have bound to a ribosome to sink, while lighter unbound 5'UTRs float. The researchers then sequence the RNA from both fractions, and finally compare how many 5'UTRs with a given sequence managed to recruit a ribosome versus how many didn’t. Using this method, Gilbert and her team found up to 1,000-fold differences in ribosome recruitment between 5'UTRs, which, she says, “is what we expected from the work that we had been doing, very painstakingly, over the years, testing five 5'UTRs at a time—but now we are able to test 10,000.” 

They then used this high-throughput approach to investigate how secondary structure affects translation initiation. Inserting stable stem-loop structures at varying positions of different 5'UTRs, they observed that ribosome recruitment was inhibited differently depending on where the stem-loops were placed. “There were cases where moving the insertion site by three nucleotides totally changed the effect: You put it in one place and it squashed translation, which is what we expected. And then you move it just a little bit, and now it doesn’t inhibit translation at all,” Gilbert recalls. In future research, Gilbert plans to use DART to further investigates the rules of translation initiation. 

Gilbert and members of her team, including postdoc Rachel Niederer, also used DART to identify other features in the 5'UTR that influence how strongly the region recruits ribosomes. One such motif were sequences containing long stretches of the nucleic acid uridine, which stimulated ribosome recruitment. They also found that 5'UTRs with higher ribosome recruitment tended to have AU-rich sequences, while those with lower ribosome recruitment had more C-rich sequences. 

“The sequences they’ve identified […] will be very dogma-changing in our ability to understand the specificity from translational control,” Barna says. “It raises our attention to these translational regulatory elements that are very underexplored.” By demonstrating that these elements “can have such profound effects on gene expression” the study reveals “a new layer of control to how gene products are being turned on and off,” she says. 

Shu-Bing Qian, a biologist at Cornell University who was not involved in the study, agrees that “using a pool of synthetic sequences to monitor ribosome loading is a clever approach.” However, he writes in an email to The Scientist, the cell-free system is a limitation as “the ribosome loading in the lysates does not necessarily reflect the situation inside cells,” and the rate at which full-length mRNA transcripts decay inside cells would also affect their translation rate. 

However, for Gilbert, it was a deliberate decision to establish DART using extracts instead of cells, as she and her team sought to “separate out effects of mRNA sequence on multiple steps of protein production.” As regulatory regions can affect RNA stability, RNA localization, and ribosome recruitment, they wanted to establish a system in which they could measure just one step “and be confident that the differences that we were measuring were coming from one thing, ribosome recruitment.” 

All experts who spoke with The Scientist about the study point out the potential for this technology in finding ways to optimize mRNA therapeutics, such as mRNA vaccines, and improve translation output. 

“There’s a lot of interest in finding these regulatory elements also from the point of view of RNA therapeutics,” Barna says.