about the lab
A Brief History of the Miura lab
“Progress in science depends on new techniques, new discoveries and new ideas, probably in that order”- Sydney Brenner
As my lab approaches 10 yrs of age, I’ve found that our progress has mirrored this saying quite closely. Our progress has been dependent on the maturation of two groundbreaking technologies– RNA-Seq and CRISPR gene editing. These enabled our new discoveries. As for “new ideas”, they haven’t been at the forefront except as a starting motivation to try something out. One thing missing from the Brenner aphorism is “people”– scientific progress in our lab has only been possible because of the great trainees we’ve been lucky to have in the group. I attribute the bulk of our progress to the hard work and passion of the trainees.
My lab started up in 2014 at the University of Nevada, Reno, in the Department of Biology. To get a job running an independent research lab, one lays out very specific research plans in order to impress on others that success will be highly likely. But the reality is that the practice of science is mostly missteps. We took many steps in many directions, whereas other steps led to productive and interesting paths. If we’ve kept consistent with anything it is our long standing interest in alternative polyadenylation and circRNAs. We’ve used flies, worms, cultured cells, and mice to investigate questions in these two general topics. I’ve found most people in science to be quite nice and collaborative. Or at least I’ve tried to be around people in science that are nice and collaborative. Our projects routinely go outside of whatever expertise the group might have and are aided by collaborations with experts in their respective fields (that are nice and collaborative).
circRNAs
During my postdoc in Eric Lai’s lab at MSKCC, we found that circRNAs increased in expression at a global level during aging in the heads of Drosophila melanogaster (Westholm et al. 2014). We weren’t looking specifically for this– a fellow postdoc, Jakub Westholm, found the trend in modEncode RNA-seq data (day 1, day 5, day 20 flies). Brent Graveley’s group at UConn School of Medicine had sequenced these samples using a new ribo-depleted total RNA-seq technique which allowed circRNA detection. Wow– this new and strange class of RNAs were age-regulated! This is an example of a new technique leading to new discovery. There was no brilliant “new idea” that enabled the discovery.
In my new lab, I wondered if this aging expression pattern also occurred in other animals during aging, and whether it was unique to the nervous system. We obtained some pilot funding to profile circRNAs in aging mice. We got the aged mice from NIA. It turns out that NIA only gives you those mice if you have active funding that specifically mentions the mice. I didn’t know that at the time, so I got lucky. Otherwise, we would have been aging mice for many months before getting results. Hannah Gruner and Mariela Cortez-Lopez in the lab led a project profiling circRNAs using RNA-Seq in various tissues of old and young mice. They found that circRNAs were globally increased in brain tissues during aging. This led to the first independent publication from the lab (Gruner et al. 2016). I am very proud of this paper– including the author list, which consists of a postdoc (Daphne), a PhD student (Hannah), and undergrad (Mariela), and high school student (Matt Bauer).
At UNR, we developed close relationships with neighboring labs in the Biology department. We started a collaboration with Sander van der Linden’s lab to study circRNAs in the aging worm. Why worms? They aged faster than flies and mice for one. When we profiled their global expression patterns during aging, we found most circRNAs to increase with age, and for some these increases were very dramatic(Cortés-López et al. 2018). Since an adult worm consists of mostly post-mitotic cells that do not turnover during aging, and since circRNAs are very stable, we interpreted the dramatic increase of circs during aging to be due to accumulation and not alterations in alternative splicing(Knupp and Miura 2018).
We wanted to test if individual circRNAs had functions related to aging. We had no experience in the aging field. The known functions of circRNAs then (and still now) really had no generalizable principles– our question was really basic– are circRNAs “good” or “bad” for aging? This was difficult to assess because circRNAs are usually generated from protein-coding genes. So it is hard to create a genetic mutant that would specifically remove a circRNA and not mess around with the proteins generated from the host gene. Earlier work had shown that sequences in introns that flank circularizing exons play a role in back-splicing by base pairing with one another. In worms, we found an abundantly expressed circRNA exon flanked by short introns that paired together really well. We used CRISPR to delete paired intronic regions, and this led to specific loss of the circRNA (circ-crh-1). These worms turned out to have an increased median lifespan(Knupp et al. 2022). So this circRNA seems to be detrimental to the aging process. How it exactly works? We really aren’t sure yet! Most paths one takes in science are hard to predict– and I think this work on circRNAs in worms was a surprising and fruitful endeavor filled with great collaborators and trainees.
In addition to circRNAs generally increasing during aging, they exhibit enhanced expression in brain versus other tissues, and in neurons in particular. We thought that RNA-binding proteins that are expressed highly in neurons could control expression of some circRNAs. David Knupp, a PhD student in the lab, together with postdoc Daphne Cooper profiled circRNAs from public datasets of mouse brain samples lacking the RNA-binding protein Nova2. They found evidence that Nova2 promotes circRNA biogenesis. David was able to uncover intronic sequences required for Nova2 to regulate circRNA biogenesis(Knupp et al. 2021). Importantly, much of this evidence for Nova2 regulation was obtained from the mouse, in vivo. I think that the specific expression patterns of circRNAs are highly dependent on cell type and we will be teasing out the details of how Nova2 controls regulation of circRNAs using cultured neurons.
Alternative 3’UTRs
During my PhD I mostly studied how a gene called Utrophin A is regulated at the translational level via its 5’UTR in skeletal muscle in Bernard Jasmin’s group at the University of Ottawa. Toward the end of my PhD, microRNAs became a thing and I got really interested because you could go on a website and predict which microRNAs might regulate which 3’UTRs. Predicting how proteins might bind and regulate UTRs was much more difficult and messy. Other people in the lab were studying BDNF, which they found to be downregulated during myogenic differentiation. Putzing around on Targetscan, I found that BDNF had 3 target sites for the muscle-specific microRNA miR-206. Bernard was so good at letting me try whatever I wanted, and so I started studying this as a side project. Turned out that BDNF made two 3’UTRs, one with all three of the miR-206 target sites. We found that the longer 3’UTR was regulated by miR-206 in muscle cells(Miura et al. 2012). So I wanted to work on 3’UTRs because they had this predicable regulation by microRNAs.
In Eric Lai’s lab, we found that there was a really wide scope of alternative polyadenylation leading to longer 3’UTRs in the brain of flies, mice, and humans(Smibert et al. 2012; Miura et al. 2013). CRISPR started to become very hot toward the end of my postdoc. At one point it felt like the whole building was trying out CRISPR. So in my own lab, I planned to make CRISPR deletions to remove long 3’UTRs, leaving only the short one remaining. This had been done before (actually for BDNF) by conventional techniques(An et al. 2008). I was excited by how many we could assess with this revolutionary approach.
We managed to generate “long 3’UTR deletion mutants” in Drosophila, mouse ESCs and mice. Bong Min Bae and Hannah Gruner worked on the Calm1 long 3’UTR deletion mouse and found that it was required for proper neural development of the DRG, and that its loss impacted hippocampal neuronal activation(Bae et al. 2020). This project started out so quickly since the gene editing by Wei Yan’s group was completed in months. But the project went on for 6 years and was really difficult to publish! We had help from expert collaborators including Grant Mastick and Simon Pierault in the Biology department. Although we found that the long 3’UTR of Calm1 was important for development, we really couldn’t get a handle on what it was doing to Calm1 regulation. Kind of like how we found that circ-crh-1 was important for worm aging, but we couldn’t find out the “how”. At least not yet. But I feel this work was important because it added to the few studies that could convice neurobiologists that long 3’UTRs mattered for the nervous system. Establishing that 3’UTR lengthening can be recapitulated in mouse ES derived neurons and that showing that CRISPR deletions 3’UTR work well in these cells gives us a cultured cell platform to tease out the mechanisms involved(Bae and Miura 2021).
The prediction for these 3’UTR CRISPR deletion projects was that long 3’UTR loss would mean that the gene would lose the ability to be regulated at the post-transcriptional level via the binding of microRNAs and proteins. In Drosophila, we made a bunch of long 3’UTR deletion mutants. Many of them led to no phenotype that we could discern (sigh!). This was a huge bummer. One deletion that turned out to have a phenotype was for the long 3’UTR of Dscam1. These flies died early and had impaired locomotion. But when we examined the protein levels of Dscam1 in the mutants (or after knockdown), they surprisingly didn’t show any change. So how could long 3’UTR loss be causing a phenotype? I used my one skill to browse IGV tracks of Drosophila RNA-seq tracks and found that there were cassette exon alternative splicing events that seemed to be regulated during the same conditions as the long 3’UTR was regulated (This was my minimal contribution to the work!) Through the work of Zhiping, and collaborative contributions of real Drosophila neurobiologists in the department (Tom Kidd, Jung Hwan Kim, Yong Zhang), we found that most of the phenotype resulting from long 3’UTR loss was due to the loss of protein coding exons upstream of the long 3’UTR(Zhang et al. 2019).
This really surprising discovery led to an NIH R35 grant and a main focus of the lab today– how alternative splicing and alternative polyadenylation are coordinated. We have most recently developed a new approach for quantifying how alternative splicing and APA events come together, called PL-Seq. We have found that other genes show very tight connectivity between selection of alternative polyadenylation decisions and distant upstream alternative splicing events.
Looking ahead
Approaching 10 years as a lab, I think we’ve definitely exceeded my expectations! In an academic science career, the future seems to be always blurry. Things could have turned out very differently. The basis of the long 3’UTR and circRNA work started during my postdoc was enabled by Brent Graveley’s RNA-Seq datasets of fly tissues. Now, I’m a faculty in the department that Brent chairs in the same building where those samples were prepared! Funny how things turn out. It’s exciting to start a new adventure for the lab at UConn Health in Connecticut.
I’m very proud of how well so many of our trainees have done, and all the cool places they went after their time in the lab. This is mostly due to how excellent they are, but I’m very happy to think that the opportunities provided by our lab environment helped a bit. I’m really lucky to have worked with them and it has been fun, challenging, and exciting. I can’t predict the future from here on, but I know that the discoveries will depend on recruiting more fantastic trainees!
