Eric and Kyle’s paper “Distributed biotin-streptavidin transcription roadblocks for mapping cotranscriptional RNA folding” is now online at Nucleic Acids Research! This work makes cotranscriptional SHAPE-Seq even easier and more accessible by developing a “one-pot” synthesis for roadblocking DNA templates and uncovers important details about how collision with different transcription roadblocks can affect the experiment.
This work is our immediate follow up to our creation of cotranscriptional SHAPE-Seq that can map the structures of nascent RNAs at nucleotide resolution as they exit RNA polymerase in vitro. A core element of cotranscriptional SHAPE-Seq is the ability to “roadblock” RNA polymerase at every position across a DNA template so that we can characterize the structure of all intermediate RNAs while they’re still part of an elongation complex. For our first paper we used a catalytically dead EcoRI mutant called Gln111 that has been used as a transcription roadblock for several decades. The major drawback to this strategy is that because roadblock binding is dependent on recognition of an EcoRI site, we had to make a unique DNA template for every roadblock site – that’s A LOT of primers and PCR! While it ended up working great, we quickly realized that this would prohibit the adoption of cotranscriptional SHAPE-Seq unless we can make it easier.
We decided that an easier way to do this would be to use randomly biotinylated DNA templates so that we’d only need two primers and one PCR reaction and the result was fantastic! Cotranscriptional SHAPE-Seq analysis of the Bacillus cereus fluoride riboswitch using biotin-streptavidin is nearly identical to the analysis we performed with Gln111 (see figure below).
Interestingly there was one distinction – some of the RNA folding transitions we observed with biotin-streptavidn were shifted relative to what we saw with Gln111. Eric thought that this might be due to RNA polymerase sliding backwards after it collided with a flexible biotin-streptavidin roadblock – a well-known process called backtracking. To determine whether this hypothesis was correct, we collaborated with Irina Artsimovitch and Yuri Nedialkov from The Ohio State University. They were able to show that while collision of RNAP and Gln111 doesn’t cause significant backtracking, collision with biotin-streptavidin roadblocks can. What this means is that while biotin-streptavidin is less expensive and easier, it does lose some precision, and while Gln111 is more expensive and cumbersome it is also higher precision. Because the strengths of these two roadblocking strategies are so complementary, we suggested that the best workflow involves using biotin-streptavidin to get a broad view of the RNA folding landscape and then to focus in on important sections of the pathway using Gln111 to get a high-precision view.
We hope that our new streamlined strategy for roadblocking DNA template synthesis will encourage others to try out cotranscriptional SHAPE-Seq! We also think that the characterization of transcription roadblock properties in this paper will be instructional not only for those who want to use cotranscriptional SHAPE-Seq, but for anyone who needs to use transcription roadblocks! Stay tuned to see how we’ve been using these methods to understand how RNAs work their magic!