 I think we'll get started. The first one will be Dr. Deere-Cold. And he's going to share with us some of his research he's been doing in Dr. Umbody's lab. And his title is Identification of Brave or Two as a Novel Regulator of Escalation. Thanks, Leah. Thanks, everybody, for coming. I'm happy to have a chance to talk a little bit about some of the research I've been able to do with Dr. Umbody. I'm sure that everyone here will agree that my eye is the most amazing organ in the body. And part of the elegant structure of the eye involves proper vascular demarcation. And there are two important and striking examples illustrated here of that ocular vascular demarcation. The first is the boundary between the avascular cornea and the vascularized conjunctiva, which we call the corneal limbis. The second example is the avascular outer retina. So this is some immunofluorescence of the retina where this perlican stains blood vessels red and dappie stains nuclei blue. And you can see that the outer retina is avascular and that's separated, the avascular outer retina is separated from the highly vascularized coroid by the RPE Brooks membrane complex. So anytime, so the umbody lab is very interested in blood vessels, we're interested in blood vessels, we're interested in vasculature in the eye, things that make blood vessels grow in the eye and molecular factors that prevent blood vessels from growing places where they're not supposed to be. And when you're talking about regulation of blood vessel growth, one of the most important factors of course is VEGF. And this year marks 40 years since the original identification of a diffusable factor that regulates blood vessel growth. So in 1971, Judith Folkman identified through painstaking biochemistry work from actually multiple different tumor sources. The isolated biochemical fractions that had biological activity that in a rat assay of neovascularization led to blood vessel growth when you added some of these biochemical fractions. And so this is the model slide from Folkman's original paper where they postulated that this diffusable factor that they called tumor angiogenesis factor was secreted by tumors and led to growth of blood vessels into the tumor to enable the tumor to get larger. And 40 years later, we know that the situation is actually much more complex. We call this factor now VEGF, but in fact there are multiple different genes within the VEGF family. And many of those genes within the VEGF family actually exist in multiple different splicing isoforms. And in addition to that, different members of the VEGF family bind to multiple different VEGF receptors. And I have focused my research on VEGF receptor one. And scientists like physicians like to give multiple names the same thing. So we also call VEGF R1 flint one, and that's what I'll call it for the rest of the talk. Now the situation is even a little more complex than what is diagrammed here because these receptors actually exist in different isoforms as well. And those different isoforms of the flint one gene are generated by alternative RNA processing. So to give you just a context of what I mean by alternative RNA processing, first I'll briefly review some fundamental molecular biology about normal sort of canonical RNA processing. So the central dogma molecular biology is that DNA makes RNA and RNA makes protein. And the DNA to RNA step actually involves multiple different processing steps. That includes, so if we look at the genomic DNA that's transcribed into a pre-MRNA that undergoes a formation of a five-prime cap splicing out of introns and three-prime cleavage and polyadenylation. So in the case of the flint one gene, the pre-MRNA can undergo RNA processing along two different pathways. The first we can consider is kind of the more canonical situation where the flint one gene contains 30 exons. In this case, all the introns are spliced out and the 30 exons are spliced together and the cleavage and polyadenylation occur downstream of exon 30. And that results in formation of a full-length receptor that has a ligand binding domain, a transmembrane domain, and an intracellular signaling domain like the diagrams in the previous slide. However, if this pre-MRNA instead undergoes cleavage and polyadenylation within intron 13, that results in production of a truncated RNA that is stable and produces an isoform that contains the ligand binding domain only and not a transmembrane domain, so therefore it's soluble. This full-length protein binds to VEGF and elicits VEGF signaling within the cell, so it acts as a VEGF agonist. This soluble protein combined VEGF, but it doesn't elicit any signals and so it acts as a soluble decoy and actually is a VEGF antagonist. So we have been very interested in studying the biology of this soluble isoform S-flit and a lot of work, some of which I've reviewed in previous talks, but none of which I will review in detail today, has shown that S-flit is a key regulator of the vascular demarcations within the eye. So these are the two of the important examples of vascular demarcation that are both critical for vision that we discussed earlier and we now know that S-flit is critical for corneal avascularity. Loss of S-flit leads to growth of blood vessels into the cornea and it's also critical for the avascularity of the outer retina is loss of S-flit leads to coroidal neovascularization. So these are some FA and OCT images of mouse models where a flit one has been knocked out or versus control. So if we're going to understand the biology of S-flit better, the production of S-flit hinges on this critical RNA processing event. So what is known at this point about this RNA processing event? Well, it was known for some time that there were a few other genes that underwent the same sort of alternative processing. It wasn't just the soluble vagus receptor one, soluble vagus receptor two, the neuropilin receptor, or a couple of examples, but recently a group from New York actually found that this was a much more widespread mechanism of alternative processing than was previously appreciated. And in fact, it seems to be a property of receptor tyrosine kinases in general. So this is kind of a busy slide, but you can see that there are multiple receptor tyrosine kinase families here. And in fact, members of all these families were found to generate both soluble and full length isoforms. You can see some of those isoforms identified here. So it's kind of a general property of receptor tyrosine kinases. This group also went on to show that in fact the U1SN RNA, which is involved in splicing, is also involved in inhibiting intronic polydenylation. So in that way, the U1SN RNA promotes formation of the full length protein. And what they were able to show is that by blocking U1 function, they were able to induce production of the soluble isoform. So what we know now is that U1SN RNA through inhibition of intronic polydenylation leads to formation of the full length isoform. But we've been interested in some time in trying to identify factors that promote formation of S-flit. And when we were thinking about this project, we thought, well, how would we go about identifying these factors? And we came up with a strategy that involves several steps. So first we thought, well, if there's a gene that promotes production of S-flit, then tissues that express high levels of S-flit should probably express high levels of this protein, a vice versa. So can we identify a set of genes whose expression mirrors that of S-flit because we might be able to identify a candidate within that set? If we can identify a set of genes that has the correct expression pattern, then can we find within that set a gene that looks like it could be involved in RNA processing? So by inspecting the protein domain structure and known function of the gene, is there one that would make sense in terms of a possible role in RNA processing? If we could find that, then could we do some functional experiments and show that if we change the expression of gene X, we also change the expression of S-flit. If it is a regulator of S-flit, then that would be critical to be able to demonstrate that. We'd also like to be able to place that functional effect within a relevant biological context. So if we modify expression of gene X, do we get a phenotype that is relevant and that makes sense in terms of regulation of S-flit? If we can show both of these functional things to be true, then we would like to be able to localize this particular gene to the RNA of interest. If it's involved in alternative RNA processing, then we'd like to do some experiments that directly look at association of the factor in question with the F-flit1 RNA. And if we can demonstrate all of these, then we'd like to think about ways that we can get at the molecular mechanism of action. How could it actually be involved in RNA processing? So first let's look at our efforts to identify factors whose expression mirrored that of S-flit. This data has been presented before, so I'm going to go over it very quickly. What we did was identified a relevant model of corneal neovascularization. We identified two mouse strains, the MRL mice, also called the healer mice, that are resistant to corneal neovascularization. Paxix heterozygous gnome mice that undergo spontaneous corneal neovascularization. Both these mice are derived from the same wild type background, which has an intermediate susceptibility to corneal neovascularization. And as you would suspect, S-flit is the key regulator of corneal avascularity. It's expressed at high levels in the MRL corneas, at low levels in the S-flit corneas. And so what we did was, well, let's look at the genome-wide expression. The expression of all genes in the corneas of these mice. See, we can identify factors that have the same pattern of expression. So we did that, and this is the set of genes that we pulled out. This is the microarray data represented in heat map form. Each row here represents a probe on the array. Each column represents a different mouse. And it's color coded according to expression levels, so the bright colors means high level of expression, and dark color means a low level of expression. And you see that just for this whole table of about 80 genes or so, it's bright on the left and dark on the right. So they're expressed at high levels in MRL, lower levels in PAC-6. So we were able to do the first step, and find a set of genes whose expression mirrored that of S-flit. The next step was to be to look through this set and see if we can find a gene that looks interesting. And so we did that, and we found a gene, Raver-2. Obviously the title of the talk is the fact that this is the one that we've been studying. We're interested in. And the first thing we did was confirmed using a different method that, yes, Raver-2 is a real-time PCR of MRNA expression data in the cornea that shows like S-flit, which is expressed at high levels in MRL and lower levels in PAC-6. Raver-2 also is expressed at higher levels in MRL and lower levels in PAC-6. What is known about Raver-2? Well, it's part of this small Raver family of genes. There's a Raver-1 and a Raver-2. Each of them have three RRM domains. It stands for an RNA recognition motif. They have nuclear localization signals and loosing rich region. So just the domain structure, the fact that they have these RNA recognition motifs, suggests that they could be involved in RNA processing, potentially. What do we know about the biology of these two factors? Well, both co-localization and structural studies, this is some NMR spectroscopy data, show that Raver-1 can interact with PTB, the polyperimidine tract binding protein, which is a well-studied RNA processing factor. We'll talk more about PTB a little later. So Raver-1 is known to interact with a factor that regulates RNA processing and experiments in a splicing reporter system show that Raver-2 regulates splicing. Or Raver-1, I'm sorry. Raver-2, a little less is known about it, but co-localization and co-immunoprecipitation studies also show that Raver-2 can interact with PTB. So this looks like a factor that could be involved in regulation of RNA processing. So we found a candidate that makes some sense in terms of a possible regulator. Now we'd like to look if modifying expression of Raver-2 changes S-flit expression. So it's a simple idea. If Raver-2 promotes production of S-flit, if we decrease the levels of Raver-2, we should block this and decrease the levels of S-flit. So how will we go about doing this? Well, we'd like to use an S-I-R-N-A mediated system. S-I-R-N-As are small RNAs that are targeted to... can be targeted in a sequence-specific manner to a specific RNA that, through this risk complex, leads to some slicing and downregulation at the end of the target RNA. So we thought let's perform S-I-R-N-A mediated Raver-2 knockdown in both an in vivo system by injecting plasmids that bear Raver-2 S-I-R-N-As into mouse cornea and an in vitro system, one that's relevant to blood vessels, HUVAC, this is a human umbilical vein endothelial cells, endothelial cells, transfect these S-I-R-N-As into this cell culture. So when we did the injections into mouse cornea, we saw, first of all, that injection of control plasmid did not result in significant change in Raver-2 expression, but injection of two different S-I-R-N-As targeting Raver-2 resulted in about a 30% knockdown of Raver-2 expression that was reproducible and statistically significant. So that's great, we're able to knock it down in the cornea, so what happens to S-flit? Well, again, injection of the control S-I-R-N-A did not change S-flit levels significantly, but injection of either of the Raver-2 S-I-R-N-As resulted in decreased expression of S-flit that was both reproducible and statistically significant. The levels of the membrane-bound isoform did not change significantly, and the levels of the control gene, and the H, did not change significantly. We did a similar experiment in Hubek cell culture and saw similar results, that is, reproducible and significant knockdown of Raver-2 and decreased expression levels of S-flit without a change in the membrane-bound isoform. So this was great, we're able to show that if you decrease Raver-2 expression, you knock down the S-flit RNA. We also wanted to show that this resulted in changes in S-flit protein expression, because it's really the protein that is functional. So those mouse corneas are really, really small. It's kind of hard to harvest enough of those and be able to do protein expression assays. So we decided to do this in the Hubek cells. So this is a soluble protein, it's secreted into the culture media. We can look at expression levels of the protein in the culture media. So we did that by Western blotting. You can see that a mock transfection versus control transfection, we compare these two. There's no significant change by Western blotting. We can see decreased intensity of the band with treatment of either of the Raver-2 siRNAs. We can do this a little more quantitatively using ELISA. And again, we see reproducible and significant knockdown of S-flit protein levels within the culture media following treatment with Raver-2 siRNAs. Whereas the total level of protein in the culture media was unchanged. So that's great. We were able to show a functional effect in terms of S-flit expression that by knocking down Raver-2, we decreased the expression of S-flit. Next, can we put this in a biological context and show a phenotypic effect that makes sense in a relevant model? And so we chose to look in the cornea again and see if we could induce corneal neovascularization. So again, these are some representative photographs. So following injection of buffer alone, we see a cornea that is beautifully clear with no neovascularization. Injection of the control siRNA also results in nice clear cornea. There's no change there. But injection of the Raver-2 siRNA results in marked corneal neovascularization, which I wasn't sure how the picture would show up. We can highlight a little bit. You can see the blood vessels there, obviously. And the same effect was seen with injection of the other Raver-2 siRNA. All right, so this is great. So we see a phenotypic effect that makes sense in terms of regulation of S-flit. All right, so the next step would be is to show that Raver-2 acts at the level of the flit1 pre-mRNA. So can we localize it to the flit1 pre-mRNA? And how would we go about doing that? Well, we can use a technique called RNA immunoprecipitation or RIP. And in this technique, we harvest first a cell lysate that contains a mixture of RNA, some of which are bound by RNA-associated proteins. We can identify generically as an RNP here. So we harvest this cell lysate. We subject it to an immunoprecipitation with an antibody specific to our RNA-binding factor of interest or a control antibody. And that will enrich for RNAs that are bound by the protein of interest. Then we can isolate the RNA and analyze it any way we like. People do arrays, people do sequencing. If you know the gene you're interested in, the best way to look at it is just gene-specific PCR. So we did this experiment in Hubex cells that express a flag-tagged version of Raver-2. One reason for that is because the flag antibody is just a beautiful antibody. Immunoprecipitation is very well, it's very clean. So we went ahead and did this experiment, and this is an atherium-stained agrose gel showing the results from... This is the flag immunoprecipitation, a control immunoprecipitation, IP with a Raver-2 polyclonal antibody. So these two antibodies will both recognize Raver-2 in different ways. And then a control antibody for the polyclonal experiment. And then we did PCR for our gene of interest, and we saw a pull-down of the flip-1 RNA with both the flag monoclonal, and this is a weak band, but it's there and it's reproducible with the Raver-2 polyclonal. So this is great. We were able to localize Raver-2 to the flip-1 RNA. These other bands here just represent primer dimer from the PCR. So now we've been able to show that Raver-2 has the correct expression for an S-flit regulator. It is functionally important in terms of regulating S-flit expression levels, and that it localizes to the flip-1 RNA. Now let's try to think about what it might be doing there. And again, we know that, at least in the other studies we're done in HeLa cells, that Raver-2 and its homolog Raver-1 can associate with PTB. So the first thing was to show that this association between Raver-2 and PTB exists within our relevant system, like endothelial cells. So we did an immunoprecipitation experiment in Hubex, followed by immunoprecipitation of Raver-2, followed by Western blotting for PTB. So this is a co-immunoprecipitation experiment. And so we can see that this is our input lane, and then we pull down from the same input with a control antibody and with the Raver-2 polyclonal antibody and then do a Western blot for PTB. And we can see significant enrichment of PTB in the Raver-2 IP relative to control IP. So this indicates that Raver-2 associates with PTB in our endothelial cells. So what do we know about PTB? Well, it binds to this region called a polyprimiting tract, which is found in introns, and it is well-known, PTB is well-known to repress splicing. However, PTB has a lot of other functions, too. It is known in some cases to activate splicing. It can activate or repress cleavage and polydentalation. It can regulate mRNA stability, mRNA localization. It can even regulate translation by regulating internal ribosomal intracytes. In terms of these different functions, the two that would make the most sense for us in terms of regulating the alternative processing of S-flit would be either repression of splicing or activation of cleavage and polydentalation. Either one of those could lead to formation of the truncated isoform. So if PTB is associating with Raver-2 and playing a role in flit-1 processing, the next thing to be able to show is that PTB also localizes to the flit-1 RNA. So we can do the same experiment, a RIP experiment for PTB, and so that was done. We can see the PTB immunoprecipitation and the control immunoprecipitation, and again PCR for flit-1, and we can see that we're able to pull down the flit-1 RNA in the PTB, but not in the control. The PTB is better studied, so we know some genes that should be targeted by PTB and other genes that should not be targeted by PTB. It's actually been studied on the genome-wide scale. So this is a nice control for our RIP experiments to be able to show, pull down of a known PTB-regulated gene and to show that we don't pull down a gene that is known to not be bound by PTB. So the next step that we were interested in was seeing what happens, is there a relationship between binding a PTB to the flit-1 RNA and RAVR2? So we decided to do a sequential experiment where we knocked down RAVR2 first and then did RNAIP to see if that affected occupancy of PTB at the flit-1 RNA. And so this is the data for that experiment. We can see that the Hubex cells were treated with either RAVR2-SIRNAs or control treatments, either a control SIRNA or a mock SIRNA treatment. And then immunoprecipitation was performed either with a PTB monoclonal antibody or control antibody. And then PCR was performed for flit-1. And what we can see is that in the control situation, we pulled down PTB, as we saw before. But when we knocked down RAVR2, we pulled down significantly less PTB. So this suggests that PTB association with flit-1 is at least in part RAVR2 dependent. That if you decrease RAVR2 levels, you decrease occupancy of PTB on the flit-1 RNA. So we've been able to start to put together some of the other processing factors that might be involved in a molecular mechanism of action for RAVR2. So again, we return back to this critical processing event. And what we now know is that U1 through intronic polyane inhibition promotes formation of the full-length protein. The data I've showed you today supports a model wherein RAVR2, likely in association with PTB, promotes production of the soluble isoform. And given what we know about PTB, we're favoring a model in which it either does this by intronic poly-A activation or by splicing repression. So we've generated a lot of new questions now with this data and a lot of future directions. Some of the things that we're considering is I showed you that lots of receptor tyrosine kinases are regulated by this kind of alternative RNA processing. So does RAVR2 and PTB regulate the expression of other receptor tyrosine kinase genes? What other complexes are involved? Can we show that the cleavage-poly vanillation factors versus splicing factors are involved? We'd like to be able to localize regulatory sequence elements within intrum-13 that are important for this processing event. We have binding sites for RAVR2 and PTB within intrum-13. And of course, look at what happens when we knock down PTB. What happens to expression of S-flit and M-flit and what happens to RAVR2 occupancy. Then finally, of course, I'd like to acknowledge Dr. Ambadi for his wonderful mentorship and for allowing me to do this work in his lab. There are lots of people in the Ambadi lab who have been instrumental and helpful in this work. I'd like to thank Dr. Olson and Dr. Mifflin for giving me the opportunity. I have some time in the lab. And of course, I'd like to thank my wonderful family and thank all of you for listening. These are my big boys here. We went for a little hike up Little Cottonwood Canyon to Little Waterfall a few weeks ago, and then we went to the new Museum of Natural History, which was really fun for a while ago. I'd be happy to answer any questions that you might have. Well, I think we have a pretty good story right now, but the main thing that we'd like to figure out now is whether or not it's acting primarily through a cleavage implied inhalation pathway or splicing pathway. And those two are interrelated, of course, but there are some experiments that we're trying to set up now to try and address that. So I think that's the first thing. I think that what makes the most sense is that Raver 2's function actually is to target PTB there, and then PTB helps to either inhibit splicing and make the intron available or activate the intronic polydentalation through interaction with the cleavage and polydentalation machinery. So we're trying to figure out, first of all, which pathway in general is involved, and then we can try to figure out exactly which cleavage and polydentalation factor is being recruited by the complex or how it may be interrupting splicing. No, the crystal structure... I don't think the crystal structure of Raver 1 or Raver 2 is known, but it's certainly Raver 2 is not known. There's really only two papers that have ever been published on Raver 2. It would be useful to map the regions of Raver 2 that are important for interaction with PTB. And some of those experiments have been done, at least I know for sure in the setting of Raver 1, so we could get some idea from Somology to Raver 1 in terms of the domains that could be important. And those could be important mutants to use in the experiment. There's lots of ways that we could think of using those. Not that I know of, but I'd like to look at that. Thank you.