 Right, so go ahead and get started. So it's my great pleasure to introduce Sarah Hanson today for thesis defense. And so this is probably one most impressive audiences to assemble for a thesis defense here in a long time. Like a lot of splicing knowledge has space separated into this room. So Sarah joined my lab in 2012, coming from the University of St. Thomas through the IPID program. And my first impression of Sarah was that she's not as tan as someone as I thought who went to college in the Caribbean. But then I learned the University of St. Thomas is in Minnesota, so that helps with a lot of pieces of that together. So Sarah came into the lab with a strong interest in chemistry and chemical biology. And so she began working on an incredibly difficult project but it's one that I've been thinking about now for over 10 years. And the idea is that we've done a lot with single molecule work to study kind of the big picture events which are happening in splicing, but then we kind of need to take the next step and study things in a lot of detail to answer really important questions about how RNAs are recognized by the splicing machinery. And so this is an incredibly challenging project and Sarah has made absolutely terrific progress on this and hopefully the paper will be submitted quite soon on this. And it'll be I think something that will make waves within the splicing community. But she's also done quite a bit of other things in the lab. And so she also has a different project that is in press at cell chemical biology which is also I think going to be quite important for the splicing field and important for those who study splicing in vivo in yeast. And finally, Sarah's also mentored a large number of undergraduates and has been extensively involved in undergraduate teaching here at UW Madison both as a TA for my courses as well as volunteering to be TAs for several courses within the department. And she's won the Denton Award for undergraduate mentoring for all those accomplishments. And the number of undergraduates she's worked with has been really phenomenal and she's not even going to talk about that work but that's all been kind of slowly making progress on one of these splices homo helicases to understand it's basically it's biological function. And she has made absolutely terrific progress with all of them and taught them a lot about biochemistry and chemistry. So I'm extraordinarily proud to welcome Sarah to her thesis defense. And I wanna thank all of you for coming. I'm a little emotional because I see all my babies on the audience but I'll let Sarah take it away. Okay, thanks, Erin. That was very nice. So yeah, so thank you for coming here today. I'm really happy to be here at this point and to talk to you about this graduate research that I've done on intron recognition by the pre spliceosome. So an outline of what I need to talk about today, it is divided between two different research projects and they're united by their being about the pre spliceosome. So we'll start with some background on what the pre spliceosome is. And then the first story is about an in-depth single molecule kinetic analysis of five prime splice site selection by U1SNRP. And then the second story I'm gonna tell you is how we developed a new tool for studying the spliceosome, which targets a protein that's a component of U2SNRP. So to start with some background, so when the vast majority of our protein encoding genes are actually transcribed as precursor messenger RNA, which means that between the protein encoding exons, there's these non-coding intron regions. And these regions have to be identified and removed and the exons ligated together with nucleotide precision in order to get a correctly encoding messenger RNA. And that's the process of splicing. So how are these sequences recognized? Within the intron there are three conserved consensus sequences. There is at the five prime end of the intron where it meets the exon, there's the five prime splice site. At the junction between the intron and the downstream exon is the three prime splice site. And in between there's a branch point sequence. So this is what needs to get recognized. It's also involved in the chemistry that's going to result in these two exons ligated together. And the machine that is gonna do both the recognition and the chemistry of this process is called the spliceosome. It's extremely large. It's composed of five SNRNAs and dozens of proteins. They are organized into these complexes called the small nuclear ribonucleo protein complexes or the SNRPs. And they need to assemble on each intron in a stepwise fashion. And after they are all assembled they undergo dramatic conformational and compositional rearrangements in order to produce an active spliceosome. At which point this machine can undergo two steps of catalysis which produces an intron laryate product and the messenger RNA. After this process that messenger RNA is released and the components need to dissociate and be recycled so that this process can happen again on another intron. And what these projects are united by is that we're interested in how is an intron identified and selected so that it can be brought into this process. So if we look in closely at the process of assembly we're gonna look at this stage which is called the pre-spliceosome. So we have a cryo-EM structure as of this last year and what is happening in the pre-spliceosome there's U1 snrp with its SNRNA base paired to the five prime splice site. However you can see it's a very large complex because it has many proteins as part of its composition. And similarly the U2 snrp has its SNRNA base paired to the branch site sequence. And so overall it makes this very large pre-spliceosome structure. And what we're gonna ask in these two projects are how are splice sites identified? And it's a really important question to understand at a basic level because we know that about 14% of all disease related point mutations can be found at splice sites. And furthermore about 50% of all disease related point mutations change splicing in some way. So we wanna understand how this works in order to go forward and understand what happens when something goes wrong. So with that I wanna tell you the first story about U1 snrp. So U1 is responsible for five prime splice site recognition during spliceosome assembly. In yeast it's a really large complex composed of a single 568 nucleotide SNRNA and 17 proteins. And it's gonna associate with the five prime splice site sequence by a combination of SNRNA, RNA base pairing, protein RNA interactions, and then protein-protein interactions with other components of the spliceosome. And there's a really interesting thing that's going on with this association. U1 snrp has a single SNRNA and it's five prime and needs to recognize many five prime splice site variants. So in our genome we have thousands of variations of the five prime splice site sequence which is represented in the sequence logo. And these nucleotides all interact with the same 10 nucleotides at the five prime end of U1 SNRNA. This sequence is so important that it's perfectly conserved between yeast and humans including two post-transcriptional pseudo-uridine modifications. So how is it possible that it's so important that it's perfectly conserved and yet it has to be flexible enough to interact with all of this variation with the exception of two positions? So to keep in mind, G plus one and U plus two are nearly invariant because they're necessary for the catalytic steps of splicing. So our approach to answering this question is to use single molecule microscopy. We use a method called co-localization single molecule spectroscopy or COSMOS. And so this is a picture of the microscope that Josh Larson who's in the audience and with help from Mad City Labs, Eric's in the audience as well, help build. And what you do in this technique is we fluorescently label biomolecules and watch their interactions one molecule at a time. So in this example, there's an RNA that's immobilized to the slide surface and it has a blue fluorophore on it. And so because it's immobilized it's gonna be there constantly through the experiment. We can put in solution other biomolecules that are able to bind and they're fluorescently labeled. And by looking at the signal from those fluorophores separately we can see when binding events occur. Now to do this sort of experiment we have to first fluorescently label what we want to study. And I wanna study U1 snrp and I wanna study it in a purified system. To do that I worked with a yeast strain where U1 protein, SNU71 is fused to a tap tag and that allows me to use affinity chromatography to purify the yeast U1 snrp complex. In that same strain there's another U1 protein SNP1 that's fused to the snap tag. And the snap tag is used to form a covalent attachment with a fluorescent small molecule. In this case I have a small molecule that's both contains a fluorophore and a biotin moiety. And that allows me to immobilize U1 to the slide surface using stripped avidin. So this process requires me to grow about 12 liters of yeast and in the end what I get is a fluorescent purified complex of about 15 picomoles of U1 snrp, which is not a lot. It's enough to do a lot of single molecule experiments or very, very few biochemistry experiments. And we did do some biochemistry because we wanted to show that we had successfully purified an active U1 snrp. So to show its activity we did in vitro splicing assays. If you take yeast whole cell extract and incubate it with a radio labeled pre-MRNA you'll get the products of splicing as you can see here. And I removed the splicing ability in that yeast whole cell extract by ablating the five prime end of the endogenous U1. And you can see most of the splicing activity is lost. And in that background if I add back purified U1 you can see a return of most of the splicing activity showing that this complex is active. We also collaborated with the Smith Lab here at UW to do some mass spectrometry on what I purified. And we can see that all 17 proteins that are part of U1 are present. And in general you see a trend line where the larger the protein is the more peptides we detect by mass spec. Now in a few replicates I could detect non-U1 splicing proteins but you can see that despite how large some of these proteins are they're not detected frequently they're not detected in all replicates. And so we conclude that they're not there to an appreciable degree. We have purified an active U1 snrp. And so with that we were ready to go forward and do single molecule experiments. And this is what it looks like. So in an experiment I have U1 snrp immobilized to the slide surface and its fluorophore is excited by a red laser. And so you can see in this field of view a bunch of fluorescent spots and each one of those is an individual molecule of U1. In these experiments in solution there is a 29 nucleotide long oligomer which contains the five prime splice site and is labeled with a Psi3 fluorophore. So it is excited with a green laser. And it is only going to be excited when it's bound to U1 because only then is it held close enough to the slide surface and is no longer diffusing so rapidly that it can't be seen. So when we look in that field of view you can see where an RNA has bound to U1 by the co-localization of the two fluorescent signals. And we look at what happens at each U1 molecule over time and that generates a trace which is the fluorescence intensity from the Psi3 fluorophore over time. And when an RNA is bound to U1 you can see that appearance of that fluorophore in the video and a signal in the trace. And when that RNA dissociates you go back to no spot being present and back to background intensity. And the information we're getting from these videos and we're getting hundreds of these data points is the time it takes for something to bind and the dwell time of that interaction. So I'll show that what it looks like in a video form. In the case of this experiment the RNA that's in solution contains the yeast consensus five prime splice site sequence. You can see how it base pairs to U1 up here. And in yeast the consensus sequence actually includes an internal mismatch. So here's U1 on the slide surface. Here's a sped up video of those RNAs interacting with U1. And just qualitatively what you can notice is some of these are very rapid. They're only there for a frame or two and they're gone. But there are other interactions that are really sticking around for a long time. And when I actually analyze the distribution of hundreds of dwell times what I see is that that distribution is best described by a double exponential decay function which is also shown above. And what that means is that within the data there are two populations. There's a short lived population and a long lived population. They each have an associated tau and amplitude. So the short lived population has a characteristic dwell time of about 12 seconds and a long lived of about 120 seconds. This was a really exciting result for me to get when I first got these experiments to work because Josh Larson had been working on experiments where he was looking at how U1 interacted with full length pre-M RNA and mobilized the slide surface. And he was doing his experiments in yeast whole cell extract. So he can ablate some of these interactions by not having ATP in his experimental conditions by mutating the branch site, by having a cap analog that's gonna inhibit this complex binding. And still in these various conditions what he sees is a dwell times that are best described by a double exponential decay function. And my kinetics are also similar to those in yeast whole cell extract. And so what this told us is that there is this behavior that's intrinsic to U1 because I could see it in a purified system and what it's consistent with is a two step binding mechanism for how U1 interacts with the five prime supply site. So when U1 first binds to a five prime supply site it's able to dissociate rapidly and that's what gives rise to the short lift population. So we call that the weak complex. Now this weakly bound complex can transition into a more stably bound state which is slower to dissociate and that's what gives rise to the long lift population. And what was really nice about being able to see this in a purified system is I was able to go forward and interrogate how the sequence of that RNA, how changing the sequence of the five prime supply site is gonna affect various steps of this mechanism. So that's what I did. I designed a library of RNAs that I wanted to test to see how sequence affects this behavior. So the first thing I did was look at how changes to complementarity would affect it. So the five prime end of U1, the single stranded region is 10 nucleotides long so we could have an RNA that's perfectly complementary to the sequence and then just remove one base pair at a time to see where the mechanism falls apart. Now as I mentioned earlier that consensus sequence includes an internal mismatch in the helix and so we also looked at how position of the mismatch might affect the mechanism. And there was one more thing we wanted to look at. So this is with a purified U1 snrp but it has all of these protein RNA interactions that are contributing to the mechanism in some way. So I also designed an SN RNA MIMIC which has the same single stranded region at its five prime end to interact with RNA and solution and so we can look at these interactions when it's just two pieces of RNA binding. And I will say first of all when I put that SN RNA MIMIC down on the slide surface so same distribution of single molecules and put this RNA in solution I get what are the saddest single molecule videos you can possibly create. This isn't a video but it might as well be. There is no co-localization observed above background levels and that's true even when I was doing these experiments with an RNA that's fully complementary. So that wasn't great but we can interpret something from that. We can interpret that U1 snrp with all of its proteins is accelerating the on rate of this interaction. And so beyond that there was something interesting that we saw. The U1 oligo association is approximately the same across many diverse five prime splice site sequences. So something that's fully complementary to something that has very minimal complementarity has a very similar apparent on rate all within one order of magnitude. So that was an interesting thing to see. It says that the association is to a limited degree somewhat sequence independent. You have some minimum sequence complementarity that's required but above that it doesn't increase the on rate. Now that doesn't mean that the protein free experiments didn't bring us any information because what we can do is anneal all of the RNA pieces together in a high salt buffer immobilize the three pieces together onto the slide surface and then do a buffer exchange and monitor the rate of dissociation. So we can see the off rate in this way. And when I did these experiments I found that the RNA only interactions are governed by thermodynamic stability. So the off rate is a single step dissociates in a single step and there is approximately an exponential relationship between how stable that duplex is and what the rate constant is. Now when I do these experiments with U1 and I'm looking at the off rate for the stable complex what I saw was that once again this rate is approximately the same across many diverse five prime supply sites. And furthermore they all fall within the range that was observed in yeast whole cell extract. And so what that's telling us is that U1 is able to modify the stability of these interactions and this is in some cases a very large change. So for instance with the 10 base paired duplex the mean lifetime when it's a protein free interaction is more than 1800 seconds but the stable complex lifetime is only about 140 seconds it's making this huge change. And beyond being able to just modify these interactions it's doing so in opposing directions depending on the identity of the RNA. So it's stabilizing weak duplexes and destabilizing these strong duplexes relative to RNA only interactions. And the question that I had proposed at the start of this section was how does one U1 SN RNA interact with so many variants of the five prime supply site sequence. And so we think of it like the Goldilocks principle it's taking things that are too hot and things that are too cold and it can make everything just right. One thing that doesn't remain the same across a diverse five prime supply site sequences though is the amplitude of the long with population. And the amplitude represents the efficiency at which U1 transitions from a weakly bound complex to a stably bound complex. And we started to look at this in terms of end to end duplex length. So where you take complementarity that is for instance seven base pairs, contiguous base pairing and you treat it the same as a duplex that has six Watson Creek base pairs but also contains an internal mismatch. And if you think of duplexes this way and you look at the data what we found is that in general duplex length increases the efficiency at which U1 transitions to the stably bound complex but there is a sharp cutoff so that when the SN RNA duplex is less than seven base pairs long the stable complex is not observed. So this indicates that U1 has the ability to sense the length of the duplex that it's interacting with. And that is an interesting concept that is very much different from how we typically consider a five prime splice site. So this is a sequence logo for all five prime splice sites found in yeast. So if I show you this and I ask you what is a five prime splice site you would look at it and you would say it's G-U-A-U-G-U it's five base pairs and it has certainly not meeting this seven base pair length. So what we did is we actually went back to all of the annotated five prime splice sites that are in yeast and we looked at them not for their individual sequence but how long of a duplex would be formed with U1 SN RNA if you allow single nucleotide mismatches. And if you look at it this way you see that 85% of introns actually have a five prime splice site that would form a long enough duplex that we would predict they're competent for stable U1 binding. So this really showed us that sometimes you can't look at the aggregate of the data you have to look at things one at a time. And so what about though the 14% of introns that appear to be too short to be competent for stable binding? Well we can't say what's going on in each individual case but what we do notice as a trend amongst these genes is many are known to have low U1 occupancy and there's a significant portion of these shorter duplexes that have either unusual intron features or they're being regulated by the cell. So that's could be some of the things that are going on in the cell that explain why there's this division between longer and shorter duplexes. And overall we think that this agrees with what we are observing in single molecule experiments. So, you know, we had to concede this project and do almost all of this work without any structures for the yeast U1 snrp. We had no cryoEM data until 2018. And when finally we've got a structure of the yeast presplicis zone we had this really great observation that we could make. This is the presplicis zone structure where we're only looking at the duplex between the five prime splice site and U1 snRNA. And there are two proteins that form extensive contacts with this duplex. The first is U1c. It's a conserved U1 snrp protein. And the other is LUX7, which is a U1 protein in yeast but an alternative splicing factor in humans. And the footprint of these contacts between these two proteins in the duplex is seven base pairs long, the minimum length requirement. So this might be the mechanism by which it can sense the length of the duplex. The other interesting thing we see from the structure is that the five prime splice site in the structure contains mismatches to U1 snRNA or non-canonical base pairs. And yet it's in this extended helix. So it's possible that the U1 proteins are in these contacts can compensate for mismatches. And we wanted to know is the long-lived complex sensitive to exactly where the position of that mismatch is located. To test that, I designed RNA oligomers, which are complementary to U1 in order to form a seven base pair long duplex with the same start and end positions but with a different position of the internal mismatch. Now, if that internal mismatch is at the plus four, plus five, plus three, plus four, or plus five position of the five prime splice site, the kinetics are all approximately the same. However, if that mismatch is at the plus one or plus two position, the long-lived population is completely lost. And to be clear, U1 can still bind these RNAs but it dissociates with a single step and it's very rapid and that's consistent with only the weekly bound complex being formed. And what I want you to remember from earlier in this presentation is that these two positions are nearly invariant and it's true for both yeast and humans. And that's because they're necessary for the catalytic steps of splicing. So what our data is telling us is that U1 has evolved a kinetic selection for the conserved positions of the five prime splice site while retaining the ability to be flexible at those other positions. So all together, we think we've put together some rules for five prime splice site selection when U1 is not interacting with other proteins. We need to have at least five Watson Creek base pairs in order to observe binding at all. And then transitions into the stably bound complex are stimulated by additional complementarity. You cannot do this transition if the duplex length is less than seven base pairs or if there are mismatches at the plus one or the plus two position. So we conclude that what's going on in vivo when you have weak splice sites is that they would be unlikely to transition to the stable complex and thus would require other splicing factors to stabilize U1 to the sequence. And this is a starting point, a foundation for a lot of future study about U1. We can ask a lot of interesting questions having done this work. So for instance, which of these steps are dependent on U1 proteins? We can, for instance, delete a non-essential protein such as an AM8 or we can make mutations in the zinc fingers of U1C in LUC7, which contact that duplex and see if they affect certain steps or certain positions of complementarity. Another interesting question is, is there a role for these U1 proteins in the destabilization of some of the stronger duplexes that we tested? And Josh, in his work, actually found evidence in yeast host cell extract that U1C may have a role in this function. This would also be very interesting to go forward and study this mechanism, but with human U1 snub. So do the kinetics differ between a core complex and one that contains alternative splicing factors, such as LUC7, which is a part of U1 in yeast? Okay, so that is the end of one project. I'm gonna take a quick sip of water. And then my second project is how we developed a tool to inhibit the splice's own by targeting a protein called HSH-155. So when we study the central dogma in some of our model organisms, one really useful tool is to inhibit the process. And we have useful tools for small molecule tools for inhibiting transcription and translation, but we did not have a tool for inhibiting the splice's own in yeast using a small molecule. And that is what this project is about. And to do that, we targeted a protein called HSH-155. So this protein is a component of the U2 snub, and it is a homologue of the branch point finding protein SF3B1 in humans. So both of these proteins contact the duplex that's formed between U2 SNRNA and the branch site sequence. And the reason that we are looking to this protein is because SF3B1 is a hotspot for mutations found in several types of cancer, and thus it's been the target for several promising anti-tubor drugs. And we want to adapt this system in the use of these drugs to be used in yeast. Now, these proteins are very highly conserved, and the largest domain in HSH-155 is the heat repeat domain. It's composed of 20 of these two alpha helices motifs called the heat repeat motif. And this domain overall forms a clamp-like structure that binds to the U2 branch site duplex. Now with this duplex, the branch point adenosine actually gets flipped out into a binding pocket that's formed by heat repeats 15 and 16. This binding pocket, which you can see here in a structure from HSH-155, is the same site where these anti-tumor drugs bind for SF3B1. So you can see pladenalide B occupying the same binding pocket form between heat repeats 15 and 16. So that tells us that these SF3B inhibitors work in this way. When they occupy that branch point A binding pocket, this protein cannot form a stable complex with this duplex. There's just one problem. This inhibition is specific to SF3B1. It does not work in yeast. So I picked up this project from another graduate student named Tucker Karocchi. And what he had shown when he was a graduate student is that you can make chimeric constructs of HSH-155. You can replace heat repeats of this yeast protein with the human sequence and yeast are viable. Furthermore, he showed that you can take these yeast and make whole cell extracts and look at the splicing activity. And if you have replaced heat repeats five through 16 with the human sequence, that yeast whole cell extract is able to splice, but it's sensitive to the drug pladenalide B, whereas the wild type HSH-155 sequence is not sensitive to the drug. However, there is no effect in vivo. So if you put this chimeric construct into yeast and you put pladenalide B in the growth medium, you don't see an inhibition of growth. And that's where I picked up this project from because we wanted to develop a method to inhibit splicing in vivo in this important model organism. And the way we were able to overcome this problem, we had hypothesized that it was due to drug efflux. So the drug is getting pumped out too rapidly to observe inhibition. And so we started working in a transporter deletion strain where several multi-drug transporters were deleted and we put into these strains the chimeric constructs that Tucker had developed. So as I showed you before, if HSH-155 is wild type, it's not sensitive to pladenalide B in the growth medium. But now in this background, if you put in that humanized construct, you can see robust inhibition. And so we have developed a way to inhibit in vivo. And furthermore, this yeast strain is sensitive to inhibition by structurally diverse SF3B1 inhibitors. So in the future, this could be used to screen for drugs, to test the optimization of lead compounds. And we also saw that you can test specific alleles of HSH, or SF3B1. So when I put in a humanized construct, but it contained a point mutation that conveys drug resistance in humans, we see that drug resistance in our assay. Going forward, I wanted to look at the binding pocket at what is actually involved in this sensitivity. And so what we had done is overlay the structure of the two binding pockets, yeast and human, and align the sequences of these heat repeats. And for most of these residues, especially the ones that actually line the binding pocket, they are almost all identical. And there's two exceptions to that. There is, N747 is availing in humans and L777 is disparaging in humans. And so we wondered if these two residues could be responsible for the difference between HSH-155 and SF3B1 in binding these drugs. So I made, or I used single point mutations as well as a double point mutation construct and tested them using our assay. So if HSH-155 includes the L777N point mutation, there's no sensitivity. However, if it contains the N747V point mutation, you now see drug sensitivity. So a single point mutation is all that's necessary to see inhibition by SF3B1 drugs and yeast. Now, if you put those both mutations into HSH-155, you see an enhancement of that sensitivity. So L777N, while not being sensitive on its own, increases the sensitivity to this drug. I showed that this arrest of growth is due to the inhibition of splicing by using primer extension. So I put in a reporter pre-MRNA into the yeast strain and grew them in the presence of the drug. Now, if the HSH-155 is wild type, you see robust splicing of that reporter pre-MRNA with or without the drug present. But if the yeast contains the double mutation, the splicing of the reporter pre-MRNA is inhibited by about 90% in the presence of this drug or box of dying. So this together shows that we've developed a system that is a way to potently inhibit splicing in vivo and could be an important tool for the splicing community. We also looked at how these changes to the active site might affect its function in binding the U2 branch site duplex. And I did this with an undergraduate in the lab named Brandon Nikolai, and what we did is we put in an Act 1 Cup 1 reporter into these strains. So this reporter pre-MRNA, if it is properly spliced, you get the Cup 1 gene in frame and that encodes for copper tolerance. So the more efficiently this pre-MRNA is spliced, the better that yeast can grow on higher concentrations of copper in the plate. So if that pre-MRNA contains a consensus branch site sequence that's perfectly complementary to U2, we see no difference in growth between yeast that contains wild type HSH155 or the double mutant. But when we start making point mutations in the sequence so that there are now mismatches between the branch site and U2 SN RNA, we start to see a difference. If there are mismatches at the one or two positions upstream of that branch point A, yeast that contain the double mutant HSH155 actually grow better. So what we see is that the same residues which convey sensitivity to SF3B1 inhibitors are improving the splicing of branch sites that contain mismatches with U2 SN RNA. So the conclusions that we drew from this work. So SF3B1 inhibitors can arrest splicing and yeast growth when the binding pocket of HSH155 is humanized by the replacement of 12 heat repeats or by a single point mutation in that binding pocket. Mutations in that binding pocket convey inhibitor sensitivity and increase the use of branch sites that have mismatches with U2 SN RNA. So we believe this will be a powerful tool for the splicing community that wants to study what happens when you inhibit the splice's own in a living cell. And some of those things that we can, or that people will be able to test. So one question we're interested in is using RNA-seq, are certain genes or certain branch site sequences affected earlier or more strongly by this inhibition? Another interesting question that I think a lot of labs would be interested in using this technique for is what happens when you decouple splicing from other processes that are occurring in the cell, such as the maintenance of chromatin state or the RNA polymerase, which we know is coupled to splicing? One question that I would be very interested in is what happens to U1 occupancy when pre-splice's own formation is inhibited? So, yes, with that, that's all the research I have to present to you today. So first and foremost, thank you, Erin, for your help during this project. It was a difficult one, and I feel like we got through it together at difficult at times, great at the end. And thank you for your mentorship, and I want to thank the whole lab for the support that I've received while I was here. I have some great lab members, some of which helped very directly on this project. So Brandon Nicolai is an undergrad in the lab who helped a lot with the HSH-155 project. I stood on Tucker's shoulders to get that HSH-155 project. So thank you for all the work you started towards it. For the U1 project, we collaborated with Floyd Smith Lab and Mark Scalf to do the mass spectrometry. So I've been supported by the integrated program in biochemistry, and I've received a lot of support from the chemistry biology interface training program as well as funding from the NIH. Here's our happy lab. We do a lot of fun stuff together as a lab. We have a lot of fun lab activities. And just a lot of opportunities to travel and go to really fun places and talk about research and hear great research. And I think that the lab, I chose the reason I got to have all of these opportunities. And furthermore, we didn't just interact in the lab space. This was a great group of people that I met in the lab or friends that I made or brought with me to graduate school and that friendship blowing off steam when it was necessary was a really important part of getting to where I am today. So I mean, I had people who flew in from Yale, Johns Hopkins and University of Washington to come in to see me defend. And I appreciate that so much to have fun and do some of this one more time. But I think the most support that you receive is from your home life. So I think Vlad has literally carried me at times through the last six years. So I think, I don't know why you were surprised when you saw the dedication page and you're like, you dedicated your thesis to me. And I was like, yeah, I didn't get here except for you. And so, yeah, thank you, Vlad, first most. And to my committee, just so you know, you might want to take this into consideration. Vlad and I did say that once one of us successfully defend, our cats get honorary PhDs. I'll just say in their defense, that one can open doors and that one can scam out a second dinner out of a second person. So they're kind of smart. So, but thank you for coming here today. I really appreciate getting to talk to you about my work and I'd love to take questions about this. Thank you. Yeah, Heidi, I saw your hand go up first. I don't know if you need to replace 12 or 12 heat repletes. I think there is work coming out of the areas lab too where they just replaced heat repeats 15 and 16. And I think there's drug inhibition in vitro as well. So Tucker was just looking in his previous paper at what chimeric constructs were viable. And he found that this particular chimeric construct is both viable and sensitive to in vitro inhibition. So that is why it became our starting point going forward. I think there could also be a role for the U1 SNRNA structure in limiting this influx of light. Because I recall it's the five prime supply side binding that is very close to a four helix junction or an extensive secondary structure in the RNA which I'm wondering if that might help prevent recognizing too far into the exosail of light through the light forward because you would be further away. Yes, I think that's very possible as being a limitation for how far base pairing can extend in this region. So one thing that we are testing going forward before we publish the paper is shifting the seven base paired duplex to see what happens. So for instance, the big question there is, is it any seven base pairs of complementarity or does it need to be precisely positioned? Because I agree there could be constraints that come from having helix one start just a few base pairs upstream of the five prime supply side. Yes, hi. Yeah. The connect. Mm-hmm. Josh's work did show that the branch side is contributing to the association of U1 with the five prime supply side. If he deletes both the branch side and the pseudo branch side, it significantly reduces the long lift population. And so there is, when I say the kinetics are approximately the same, that's what we're kind of getting at is that they're not different by an order of magnitude. They're two-fold different and that purified U1 keeps it at that same level across it. So it definitely is shorter in yeast-hole cell extract if you don't have interactions with the branch point binding protein. No, I haven't. I think that's very interesting, but we haven't looked at that yet. I think we have all the data at hand. We just haven't analyzed it in that way yet. Yeah, we got, we share the. Are there any other questions? I'll ask it. Yeah, hey. Yes, in yeast. Yes. Oh, you're saying where that base pairing is of the short duplex? That's possible. I don't know for sure about those short duplexes. One of the unusual features we see is actually a long five prime exon, short one, but that might be a mechanism, both the length of the exon and the weaker five prime splice site might be a mechanism of controlling that particular transcript, but I can't say anything about the particular genes. Into the closed session, but I think that's an interesting question. So normally, like when you have something that's conserved, conserved's a particular sequence for a particular job, but here, it conserves spliceisome. It's conserved, but it seems like from your talk, it's conserved in order to allow variability. Yeah. So that's kind of interesting. It's like a little backwards from normal. And I wondered what the selective advantage is of nature doing it that way instead of having nature just make all the splice site sequences the same, which would have been the more obvious thing to do. I think the advantage of this is that it allows for other steps of regulation where you won is not the ultimate regulator of how well a particular intron is spliced. It's saying this has a GU and it has the minimum amount of complementarity that it could very well be a five prime splice site, but it also gets defined by other features like the cap binding complex, branch point binding, and then before catalysis by the spliceisome, it's exchanged with U6. So I think that why wouldn't there just be one sequence that's always used? I think it allows for transcriptional control. So some of these introns that only, you only want to be spliced and expressed at levels during meiosis, for instance. You can control that by maybe having a low level of affinity baseline for you one, but then you express other factors during meiosis that will promote that interaction. So you have all these protein-protein interactions that can be dependent on cellular context and control. Because it gives you flexibility in changing the splicing when you want to by having other factors get kicked in. Yeah, so you can be inefficient under some contexts and efficiently splice that transcript under other contexts. Yes? That affects splicing. And so if you go back and sort of map the splice sites, are there any patterns that jump out that? I'm not sure about that. And that 50% figure means mutations that will put a point mutation into a splicing factor, like a mutation that'll make HSH-155, a cancer associated mutation. So, yeah. You know, I can't think of any off the top of my head. And I don't know if there's, I don't know if there's a very particular pattern. Thanks, Sarah, for the wonderful talk.