NEW YORK (GenomeWeb) – A University of Pennsylvania group has put its novel sequencing strategy for profiling RNA-protein interactions to task in a study of RNA-binding proteins and RNA structure in the cell nuclei of Arabidopsis seedlings.
Taking this unbiased or global look at RNA-protein interaction specifically in the nucleus, the study, published this week in Molecular Cell, offers a snapshot of regulatory processes that occur before RNA moves out of the nucleus to a cell's cytoplasm.
Brian Gregory, a Penn assistant professor and senior author of the study, told GenomeWeb that techniques abound for sequencing or otherwise measuring RNA-protein interactions. Many have been focused on detecting the binding of specific proteins to RNA. Others, while allowing a broad look at RNA-protein interactions, have not been able to also capture the overall structure of these complexes.
Gregory and his team's method, protein interaction profile sequencing, or PIP-seq, is an RNA-centric rather than protein-centric approach, Gregory said, that can capture both the pattern of protein-RNA binding sites and the larger structure of nuclear or other transcripts.
Earlier this year, Gregory and his team published an initial study debuting PIP-seq in Genome Biology. The researchers describe PIP-seq as a "universal, high-throughput, ribonuclease-mediated protein footprint sequencing approach" that reveals RNA-protein interaction sites throughout a transcriptome of interest.
"We use structure-specific nucleases, meaning they only cleave after a base that is in a specific structural conformation — either single-stranded or double-stranded. We treat the RNA till completion, and if it didn’t get degraded by the single-stranded RNAse it's considered double-stranded and vice versa. Then we put that information back together," Gregory explained.
In their initial study, Gregory and his colleagues demonstrated how the technique could be used to catalogue RNA-protein binding in human cell lines. In the new study published this week, the team then set out to try to catalogue the landscape of RNA-binding proteins, or RBPs, as well as the secondary structure of RNA transcripts in the nuclei of Arabidopsis seedlings.
"We definitely wanted to show the full power of the method, and look at both aspects of information you can get out of this approach, both transcriptome-wide RBP binding and transcriptome-wide structure, which we didn't do in our first paper," Gregory said.
"We also wanted to go a step further because most previous studies have focused on taking a cell and bursting it open and looking at the total population of RNA. We wanted to focus down on what RNA looks like in the nucleus before it moves out. We wanted to see if we could capture more RNAs in their immature state, as they are being spliced or before."
In the study, the researchers performed PIP-seq on the nuclei of 10-day-old Arabidopsis seedlings. In cataloguing nuclear RBPs, they made several different discoveries. For one, they achieved their goal of measuring RBP binding patterns in RNA in its pre-spliced or immature state. "In the paper we reported that about half of our reads come from these pre-mRNAs or mRNAs in the process of splicing," Gregory said.
Moreover, the group was able to measure patterns of structure and of protein binding that mark what look to be important aspects of the mRNA maturation process, he explained. The group saw that RBP-binding and certain patterns of secondary structure were much more common at sites where alternative splicing and alternative polyadenylation occur.
Interestingly, Gregory said, the team also found that RBPs bind frequently near the start codons of nuclear RNA. Heightened RBP binding near start codons has been known from studies of full cells, but was thought to be the result of ribosomes attaching to RNA in the cytoplasm at the start of translation.
"There should be no known functional ribosomes [in the nucleus]," he said. "But we saw that a large fraction of mRNAs actually have their start codon bound by some RBPs even before they come out to the cytoplasm."
The researchers also observed that the sections of RNA where RBPs bind appear to have been conserved over evolutionary time, suggesting they are likely to play an important role in the gene expression and regulatory processes
In addition, the group found a strong inverse relationship between patterns of RBP binding and the secondary structure of RNA molecules. Where structure is less complex there tended to be more RBP binding, and vice versa.
The group's PIP-seq analysis also revealed that some collections of transcripts were linked by the sequence near their RBP binding sites into distinct groups or "motifs," potentially indicating a functional relationship between these molecules.
Tracking one of these motifs, the group identified an RBP, CP29A, as being active in the nucleus. This protein was previously known to operate only in the chloroplast.
Moving forward, Gregory said he and his team are now expanding their work to examine RBP binding and structural patterns earlier in Arabidopsis development. Specifically, the team is hoping to find a stronger connection between RBP binding and alternative splicing, and to investigate its connection to the differentiation of root hair cells from their precursors.
"All the cells start the same," Gregory said. "But at one point in development every other cell becomes a root hair cell: you have non-hair, hair, non-hair, hair — alternating all the way down the same cell file. So we are comparing alternative splicing events in these two types to see if there are structural and RBP binding differences in a developmental switch type of context."
The researchers also plan to use PIP-seq to try to understand some of the same processes in human cells.