In this seminar, Nicole Martinez presents her laboratory’s research on the chemical modification pseudouridine (Ψ) in messenger RNAs. She demonstrates that Ψ is deposited co‑transcriptionally on pre‑mRNA by a subset of nuclear pseudouridine synthases (e.g., PUS1), using biochemical fractionation of HepG2 chromatin and high‑throughput CMC‑derivatization coupled with reverse‑transcriptase stop profiling to map Ψ sites. The data reveal that Ψ is enriched in introns and near splice sites, especially within alternatively spliced cassette exons, suggesting a role in splicing regulation. Functional assays—single‑site Ψ reporters in nuclear extracts and PUS1 knockout RNA‑seq—show that individual Ψ residues can enhance splicing and that loss of PUS1 leads to widespread alternative splicing changes. Integration with RNA‑binding protein eCLIP data indicates many Ψ sites overlap splicing factor binding sites, hinting at a mechanism where Ψ modulates protein‑RNA interactions. The talk also connects Ψ biology to human disease: loss‑of‑function mutations in PUS enzymes cause neurodevelopmental disorders, while over‑expression (e.g., PUS1) correlates with poor prognosis in cancers such as hepatocellular carcinoma, where PUS1 knockdown reduces tumor cell growth. Martinez discusses open questions about enzyme targeting specificity (sequence vs. structural motifs), evolutionary origins of Ψ enzymes, and broader implications, including the use of Ψ in mRNA vaccine stability and immune evasion. The Q&A addresses technical specificity of CMC labeling, the actual fraction of Ψ in mRNA, evolutionary context, occurrence at start/stop codons, and potential mechanisms of vaccine‑related stabilization.
Questions this video answers
CMC reacts with uridine and guanosine, but under alkaline conditions the adducts on U and G are reversed, while the adduct on pseudouridine remains, providing specificity.
Only a small fraction of uridines are pseudouridylated in mRNA—about 0.2–0.4% of Us by mass‑spec analysis.
Pseudouridine synthases are ancient, present across kingdoms, including bacteria that lack splicing, indicating the modification predates splicing. Some enzyme families expanded in humans, gaining additional functions.
There are few examples of pseudouridine at start codons; a few have been reported in stop codons and can cause translational read‑through. Some enzymes recognize a sequence motif, while others rely on RNA secondary structures such as stem‑loops; many lack a clear consensus.
Ψ can block cleavage by RNases and prevent activation of Toll‑like receptors, which contributes to the enhanced stability and reduced innate immune sensing of vaccine mRNAs. These mechanisms may be distinct from its role in splicing regulation.
Chapters
So if you thought this morning was good, just wait, because we have all kinds of amazing stuff coming up in our next session. And to kick things off, I'm going to pass the microphone to our own Nathaniel Gray, who will moderate the first talk.
Great.
Thanks, everybody. I think everyone's hopefully enjoying this as much as I am. Really had a wonderful time. I've also only been here about a year and a half, but really appreciating what a wonderful community we have here. So to start off our afternoon session, we're going to start with Nicole Martinez, who we've been really lucky to recruit to sort of three entities: Chem-H, the Department of Chemical Systems Biology, and also the Biology Department. And Nicole got her postdoctoral training at Yale with Wendy Gilbert, and her PhD at the University of Pennsylvania, where she studied alternative splicing with Kristen Lynch. So Nicole's lab is really doing fundamental work trying to understand how RNA is regulated and how those modifications affect stability and subsequent translation. I think you'll really see how her work is going to impact both our sort of understanding of physiological and pathological RNA regulation. So with that, I'll turn it over to Nicole.
So I'm really excited to be here celebrating with you the anniversary of Sarafan-Kim H, which is one of the scientific homes of my lab. And I've been very fortunate to start my lab in this community with amazing mentors and colleagues. So my lab is interested in chemical modifications to RNA and how they rewrite the message in health and disease. Okay, so pre-messenger RNA is transcribed from DNA, and then it is processed by splicing to remove introns and join exons. It's then cleaved and polyadenylated to make mature messenger RNA that can then be translated into proteins to achieve cellular functions. And all of these steps are really highly subject to regulation to control gene expression. And what I want to tell you today is that over the past decade, it has become apparent that messenger RNAs are chemically modified to make non-canonical nucleosides. And these non-canonical nucleosides change or rewrite the information that's encoded in our genome. And if these modifications are added early to the pre-messenger RNA, these chemical modifications have the potential to impact every step shown here to then regulate gene expression in a variety of cellular conditions. I have been studying one of these modifications, pseudouridine. When I started my postdoc, we knew that messenger RNAs contained some of these chemical modifications like pseudouridine. But what we didn't know is when these were added to the messenger RNA in their life cycle. And we also didn't know what they were doing there. Why do these modifications get added to the messenger RNA? Okay. So this is that modification, pseudouridine. It's an isomer of the canonical base uridine, and it is catalyzed by these pseudouridine synthases that carry out a base rotation of uridine to pseudouridine. And there's 13 of these enzymes in human cells, and the vast majority of them have been associated with a wide range of human diseases. And these range from neurodevelopmental disorders to cancer. So this really highlights the need to understand where these modifications are in messenger RNAs and what they do so that we can begin to understand how they contribute to human health and disease. Okay. So two of these enzymes that we knew about were nuclear or chromatin-associated. So we hypothesized that this modification might be added to the nascent pre-messenger RNA. RNA, again, where it could potentially have functions in the nucleus in early steps of pre-mRNA processing, or then go on to impact other stages of the mRNA life cycle. Okay, so this is the first question that we wanted to address when I started this work is, do these modifications get added to the pre-messenger RNA? And in order to ask that question, we took a biochemical approach and fractionated HepG2 cells and isolated RNA from chromatin. The idea being that we could use this to enrich for unspliced pre-messenger RNA that's still tethered to the polymerase. And we then performed high-throughput pseudouridine profiling to figure out where these modifications and whether they were added to the pre-messenger RNA. And what I'll tell you is that these modifications, pseudouridines, aren't on the Watson-Crick face of the base. So they don't normally cause any perturbations to reverse transcriptase. So in conventional RNA sequencing, you can't detect this modified base. So we need ways to specifically label the base and be able to detect it by high-throughput sequencing. So we've been using approaches to derivatize the base, such as this chemical or this carbodiimide called CMC. And this reacts with the pseudouridine base and it creates a bulky covalent adduct that forms a robust roadblock to reverse transcriptase. So then we can sequence those truncated cDNAs and determine the location of the modification based on these reverse transcriptase stops. So we did this and we found that pseudouridines are added to the nascent RNA. There's many of them that are added to nuclear non-coding RNAs, and the majority are added to pre-messenger RNAs. And these are both within exons shown in purple as well as in introns shown in red. And these ones in introns were particularly interesting because they're there when the pre-messenger RNA is unspliced, but the pre-mRNAs then get spliced and the mature messenger RNA no longer has that modification. So we thought that if these had a function, it had to be in the nucleus or in pre-mRNA processing. Okay, so then we asked, can these modifications potentially impact pre-mRNA processing steps such as splicing? And what we found is that these modifications are enriched around alternatively spliced regions. Alternative splicing is the differential inclusion of a portion of the pre-mRNA into the final mature messenger RNA. And in the most prevalent example, in the very top of cassette exons, these are exons that can either be included or skipped from the mRNA. So you can generate distinct mRNA isoforms that can encode even distinct proteins with different functions in some cases. And what we found was that the pseudouridines, which are shown in the filled bars compared to the uridines in the line bars, were enriched in the introns of regulated regions of these cassette exons and then also other types of alternative splicing. And we also found that these tend to happen very close to splice sites or within splice sites. Where regulatory elements for splicing are found. So again, this suggested that these modifications could be playing a role in splicing regulation. So we wanted to ask this: do these modifications affect pre-mRNA splicing? We use splicing reporters that contain two exons and introns, and we either left this unmodified or we site-specifically introduced a single pseudouridine, and we added these to nuclear extracts and quantified percent spliced to unspliced and found that a single pseudouridine could enhance splicing in nuclear extracts from different cell types. And we were quite excited about this because in these wild-type cells, the only difference is the presence and absence of just one pseudouridine. So then we wanted to ask, well, what about in cells? Do these modifications impact splicing in cells? And we used a high-throughput assay to assign the direct targets of these different pseudouridine synthases that I've been telling you about.. And we tested a variety of these in human cells and found a number that modify pre-mRNA sequences. And we focused on the ones that modified the most pre-mRNA sequences to ask what happens when you lose these enzymes and modification to splicing. All right. So then we made PUS1 knockout cells and we isolated RNA from the knockout and wild-type cells, performed RNA sequencing, and looked for evidence of differential splicing. And we found examples like the one shown here where in this cassette exon it's included when PUS1 is there and when you have pseudouridine. And in the absence of PUS1, you get skipping of the exon. And that's quantified in the graph below. And what I'm showing right below the cartoon schematic is data showing that there is a pseudouridine in the intron downstream of this exon. And you can see that by that RT stop that's seen Cmc-dependent and it's there when PUS1 is there and not when you don't have PUS1. Okay. So, then we looked genome-wide and we found that there was widespread alternative splicing upon PUS1 knockout, the majority in cassette exons but also other types of alternative splicing. And when we knocked down the two other enzymes that had the most targets in pre-messenger RNAs, we also saw widespread alternative splicing showing that these enzymes control gene expression programs at the level of alternative splicing, and those programs are really largely distinct between these different enzymes, and they recognize different substrates as well. Then our next question was, we see these effects on alternative splicing, what could be potential mechanisms that mediate these changes in splicing? And one of those is the possibility that pseudouridines impact the binding of RNA binding proteins. Or in the case of splicing, specific splicing factors that bind RNA. And these modifications could, in theory, change the affinity of the splicing factor for the RNA and therefore change splicing outcome. So we collaborated with Dean Yeo's lab, and they have mapped the binding sites of hundreds of RNA-binding proteins in high throughput in the same cell line where we found these pseudouridine sites. And what we found was that there's a number of RNA-binding proteins that are enriched for having pseudouridines in their binding site. And the majority of these highlighted in blue are splicing factors, suggesting a potential mechanism for changes that we see in alternative splicing. And we find that the large fraction of all of the pseudouridine sites that we found in pre-messenger RNAs are actually present in the binding sites of RNA-binding proteins. So now, we don't know of any endogenous readers of these modifications that can detect whether the modification is there or not, so we're following up on these candidate factors as potential readers of this modification. So, what I've told you so far is that these enzymes modify pre-messenger RNA co-transcriptionally, that they regulate gene expression at the level of alternative pre-mRNA processing, and that individual modifications can directly affect splicing. Now, in my lab, we're really interested in a number of questions related to these modifications and their role in broadly regulating gene expression. And one of these is how do specific sites select— get selected for modification and what directs the enzymes to specific locations? And since we see it's co-transcriptional, we're also wondering if there's coupling between transcription and deposition of these modifications. And my student, Becca, is working on that. We're also interested in what I told you earlier, that these enzymes are dysregulated in disease. Loss of function mutations cause neurodevelopmental disorders in one of the enzymes that I told you about, PUS7. And Wendy in my lab is studying that. And we also know that there's high expression of these enzymes and that high expression is correlated with poor outcome in a variety of cancers. So two people in my lab, Christina and Matthias, are working on from that. And this is one piece of data from Matthias, a postdoc in the lab, showing that in hepatocellular carcinoma, where high expression of PUS1 is correlated with poor outcome, we see that knockdown of PUS1 leads to a reduction in hepatocellular carcinoma cell growth. So this enzyme is promoting cancer cell growth. So we're now really interested in what are the mechanisms behind this. And can we tie some of the RNA targets of this enzyme to these phenotypes? And lastly, this is all— what I've told you today is about one modification, pseudouridine, but we're interested in how other modifications might impact gene expression and be used to control gene expression in different conditions. So Nico, a student in the lab, is trying to exploit the chemistry of distinct nucleosides to develop new methods to detect and study the function of other modifications. And I hope you saw some of their posters. I realize the poster session has happened, but if not, please catch up with them. And with that, I'd like to thank my postdoc advisor, Wendy Gilbert, for her support where some of this work started, our collaborators, and the members of my new lab, and the fantastic ChemH community. Thank you.
Questions?
Carolyn? This is so exciting, such a cool wild area of RNA metabolism. And I have two questions. The first is really probably a dumb question, which is what makes the CMC chemical modification specific for pseudouridine and not uridine? And is there a selectivity issue there?
Yeah, so it actually reacts with uridine and Gs. So it reacts with Us and Gs. But under alkaline conditions, it reverses off of U's and G's and stays on pseudouridines. So that's how the specificity achieve— is achieved, not on the modification step, but really the resistance to reversal for the pseudouridine.
Oh, that's cool. Okay, that's a great like mechanistic problem for students in a class somewhere, I'm sure. And then the next question was, you showed in the bar graph where you were quantitating the percent of pseudouridines in different types of RNA. And I think it was in the pre-messenger RNA. It was like more than 50% of the Us are actually pseudoUs. Is that— did I read that correctly?
No, no. So it's a small fraction. It's something like 0.2% or 0.4% of Us is pseudouridine in messenger RNAs by mass spec.
So maybe I just didn't understand the bar graph then. I wish I could point to it. I don't remember which slide it was, but it was the y-axis was— oops, sorry. It was somewhere where the red bar was really big. No?
I saw that too, and I think it was normalized for total.
Oh, is that what it was?
Okay.
Okay.
Then I missed it. Oh, yeah, this one.
Yeah, that's percent of total.
So, right, that's the percentage of pseudouridines in that kind of alternative splicing.
I see.
Okay. But still, it's pretty dramatic.
Thank you.
Hi, Nicole.
I was wondering if you could comment on sort of the evolution of these chemical modifications and splicing. And are these sort of like co-evolved processes, or did splicing come first and the enzymes that put on these modifications later? And how do you think about that?
Yeah, so the enzymes are really ancient. They're present across all kingdoms of life. So the enzymes are there in bacteria where there's no splicing. There's also been expansion of several of the families. So there's some families that are only present in human, for example.
So the modifications came first, and they have—
some of them have non-coding RNA substrates like tRNAs. So they likely evolved additional functions. Most of them have additional domains and different features which might suggest additional targets and ways of recognizing mRNA, for example. But it's a really interesting question.
Question there. Yeah, fascinating. You ever see the start codon modified? And then is there any consensus site or sequence that makes a uridine more susceptible to modification versus not?
Yeah, that's a good question. So for the first, there aren't many examples of pseudouridines in translation start sites. There's now a few examples of some in stop codons and work from other groups. Yi-Tao Yu's lab, for example, has shown that if you artificially put a pseudouridine in a stop codon, it leads to read-through, translational read-through. So there is the potential for these modifications to be regulating multiple steps of mRNA processing throughout the life cycle of the mRNA. And related to your question, some of the enzymes recognize a sequence motif But most of the others do not. And for some, we've shown that there's a structural motif, that is that the RNA folds into a specific stem loop structure that can be recognized by the enzymes. But that is a question we're really interested in, is what directs the specificity of these enzymes, especially for the ones that don't have a sequence motif associated with their targets.
All right, we have one more last question.
That was so interesting. So pseudouridine got famous in the last couple years, right, for its stabilization of the mRNA vaccine and also the avoidant— the lack of binding to, I guess, its toll receptors. Does any of that, especially the stabilization, have to do with its function here, or is the structural change that prevents degradation sort of the same as is affecting splicing?
Yeah, no, I think that's a really interesting point. We know that the modification can prevent cleavage by some RNases, So those could be readers of the modification, but this has all been sort of artificial. The mechanisms there aren't really known, but if you add a pseudouridine and then try to digest with RNase L or E, then that blocks it. So there is some potential mechanism there. And the other thing with the vaccines is that they prevent recognition by toll-like receptors, as you suggest, in activation of foreign nucleic acid centers or innate immunity. And that's something we've gotten really interested in lately, and that is the role of these marks in marking host mRNAs as self as a potential additional function. So that is something we're interested in and potentially distinct functions.
