 You know he's an important personality in science, but I, where I stood, and I feel a form of euphoria, I'm here to do so. The only reason I decided to plan to use my gap after doing it yesterday, is because I didn't quite make it to the one place of the CV, so I'm going to complete the job today. Peter, yesterday, was researching as a phenomenon, and his lecture is a phenomenon, and I'm sure you'll be surprised that that wasn't all that he does. That's a subtraction. He's going to tell you a little bit about what he's done in science today. Peter started as another graduate in 1970. He finished another graduate at the Cal Tech in 1979, and now he's usually a bit derby, working on small molecules that are in DNA in 84, and then after that it's just incredible and hungry to just kind of see these. In two years he went from assistant professor to associate professor, two years to professor, a couple years after that he was a hard user, a few years after that, I mean two years after that. By the same time actually, by the age of 28 he was a member of the academy. A rare distinction is also a member of the Institute of Medicine at the National Academy, the undergraduate awards, graduate awards, postgraduate awards. I mean I don't know what's that, but you guys were competing for the award at the same time. The CERN scholarship gave us the latest loan, and it goes on later. Mending went of course, I mean the book, right? The most important lecture I think, and it's a city dance on this one, it says Harry Stimbal Pension, University of Wisconsin 2008, so that must be the drawing. And don't put it down, Harry Stimbal's lectures in the past years have gone on. If they haven't already won the prize, they've gone on to win one. Example, of course, says, good trick, we'll give this off and we can later follow the prize. So I think whether I'm sure or not. I can continue, I think I'll carry on. Here's a list of the awards. It goes on single line all the way up to here. But you know, stop here and let me give you the justification for all the tax dollars I've won. During today's 8th or 15th. Thanks I think. Anyway, it's really been fun being here. It's just perfectly talking to everybody. Faculty and grad students and both sides. It's really exciting. What's your brain going? What I'm going to do today is, I've got kind of a flip side of what I was talking about yesterday. As I said, we basically have hards and thiochemists. We make molecules. And our interest is in making molecules. Not with defined structures, but functions. Either chemical, biological or physical. And as I said, the chemists aren't very good at that. So in general, what we do is take lessons from nature. And use biological strategies and use any of the biological molecules. That help us to kind of create interesting molecular structures and assemblies. And yesterday I talked about complex molecular structures and assemblies in the context of living organisms altering their genetic code. Today I want to go to the other end of the spectrum and talk about small molecules. Okay, molecules are 500 lines of the way. And how do we find interesting biologically active small molecules that biologists might think are interesting? So that's where we're going to start today. I actually fell asleep before I finished the talk last night. So let's see how this goes. Okay, but so how do we find interesting small molecules in terms of molecules that do interesting things from biologists or something? Well, one way is to take a defined target and actually attempt to rationally design a molecule that binds and blocks it. And that's not so easy to do in the case of entire teams of personal accounts to do that. We take a different approach, and that's a biologically inspired approach. And that's really to make large collections of small molecules and simply look for those of the most interesting properties. And then work backwards. It's the same way the inusism works in terms of the right emission of them. You want to make an antibody that recognizes a leg and you make a billion antibodies. Okay, based on, you know, B, D, J rearrangements and recombination. And then you find one that works and fine tune it through affinity and attrition. Well, that's exactly what we're going to do. Instead of assembling molecules in terms of gene segments, variable, joint, and diversity, we break down small molecules in terms of building blocks. So if you take a benzodiazepine, you can break it down into three sets of building blocks. They're small molecular entities. And if you assemble a hundred of these, a hundred of these, and a hundred of these, you'll get a billion molecules. The same way the inusism works. Okay, so that's the strategy. It's called combinatorial chemistry. A lot of people practice it now. So we've actually developed it's scripts, Matt Gray's and Bing Xiu Wu and others have developed some really pretty simple chemistry for making very diverse chemical scaffolds where we actually use the heterocycles themselves that you see in broad light substances as pieces of the building blocks. So we can create huge libraries of compounds. And our constraints are these should actually look like molecules that are known to bind with the biological receptors. And they should be easy to make. Because once you find an interesting small molecule, that's not where the story ends. You have to improve its affinity and selectivity if you want to do anything with it down. So you really need straight forward synthetic routes. Otherwise every molecule you find you're going to need five times to optimize it. So we've built libraries around all kinds of molecules that actually are known to interact and modulate the activities of biological molecules. Many of which are in four or five states. And it's actually probably a collection of about 100,000 compounds. In GNF we've built a collection of two and a half million compounds. So we actually have relatively large collections. And the challenge then is when we realize we would take that, say two and a half million molecules, a big farm and if they're going to screen two and a half million molecules, they budget a million dollars to do the screen. So that presented a problem for us. Because we could have run two screens and got out of money. So what we did is we said, well if you really want to use this as a research tool, to find interesting new molecules, you really need to do something about the cost of doing these experiments as well. Although I just can't do the experiments. So at GNF we hired a group of engineers, actually my own college remunerated. Caltech was the person who built the first electronic transmission, which was in Saturn Core. And I decided that most things you build in biotech, you publish in science and then they break down. The comers have to drive 120,000 miles. So we hired automotive engineers and we really built high throughput cell based screening methodology. And this system, we can screen the entire collection in a day or two days. It costs pennies per molecule. The entire screen now instead of costing a million dollars, it costs 30,000 or 20,000 dollars. So almost a 50 fold reduction in cost. And the molecule collections we can buy might become a little molecule in the last ten years. So it becomes very simple to make quantities of molecules to build huge collections. And moreover we build imaging capabilities that allow us to screen in almost any format, including high content. So now that we have this collection of molecules that are diverse, but simple to make, and we have the ability to run screens in a research mode, we decided in my scratch lab, one of the things I'm talking about mostly is work on screens, is we've decided to run a whole series of screens to look for interesting molecules. And so for instance, you can run screens where you target a particular protein. So one of the proteins we were interested in, in fact work with a group of Gina, is a translocated kinase called NPAOK, which is one of these fusion events that leads to constitutive activity and a signaling kinase that leads to cancer. At first this translocation was identified by plastic large-cell lymphomas. Now all of a sudden it's found that one cancer is a barren cancer, so it's a huge population of potential patients. So in this case what we wanted to do was a cell-based screen to find inhibitors of this translocated kinase. And so you can do a very simple, we've generated a very simple construct where you turn this into a proliferative screen by taking these three cells and actually making their proliferation dependent on the activity of this kinase. And this you can do, it's quite commonly done now. And so then you can screen these collections of molecules and here we found a molecule that's active at about IC50, about a five-man long. And if you actually put it into a mouse model of NPAOK, you actually get remission, okay? I mean, this is actually really impressive. So 10 megs per K and 8th rodents, the timbers are belated. So actually this is probably going to go into people in the next 12 months, something like that. So you can do target-based screens. You can then begin to generalize these. So you can actually take tyrosine kinases and make them considerably active because it's known in certain fusions, a patient-derived fusion, you have the tel dimerization domain fused to the kinase domain that again leads to an activated translocated kinase. So what we said is why don't we just take John Melting, so why don't we just take this and begin to fuse it to every tyrosine kinase. And it turns out that worked. So he actually has made, we've made fusions of all, of roughly 80 human tyrosine kinases. And so we have all of these in cell-based screen mode and then we have a collection of our kinase selected in small uncle's capital. So the question was how do we screen all those? So Jeremy Caldwell on the GNF and the engineering group and Dan Sykes and others got together and actually built an automated system now with not only screens, it will take 200 cell lines and propagate them in an automated way. So it splits them, changes media, spins them down, reuses, spins them, plays them out, and then blow the compounds. And so on a single experiment, you can actually now, if you're looking for interesting kinases, in this first one, we took 42 tyrosine kinases in a single experiment, I'd say, I mean it's 5,000 compound, and generated about a million and a half data points. So the kinase of information you now get by doing all of the tyrosine kinases is now doing a large number of work in GPCRs and so forth the same way. But you know how you see these kinome branch charts? Well now you can actually take the kinome and begin the segment based on reactivities of kinases or inhibitory activities. The degree of inhibition of these kinases by specific molecules in your screening collection. So what you wind up doing is clustering these kinases based on their ability to be inhibited with specific chemical scaffolds. And then what you wind up seeing is, oh, if I had an inhibitor of I, it should inhibit IGFR, too. So oh, if I want to get my inhibitor, let's use an inhibitor molecule that hits AOK and so forth and so on. So you can just march through your collection of kinase inhibitors and go from one kinase to another to another and figure out start points. And then you can actually go on to the lecture and here's Gleevec, GNAF is funded by no borders. So you can see this is actually quite selective. And it's BCR-Able, OK, KIT, and PDF. The Bristol-Meyers-Squid competitor compound, OK, it's everything, OK. So this is actually BCR-Able in either, too. But the real point of this is actually, you can also use this molecule as a start point to inhibit when or after in B1 or whatever you want. So again, you can be able to go from singles specific bargain-based strains to families of proteins to generate small molecules that can be useful tools and potentially useful drugs. And then, the research part of this is really then going in and treating basically the cell as a black box and doing an unbiased cell-based screening. And you can screen for molecules that affect a heat drain to cellular processes or proliferation to differentiation apoptosis and so on and so on based on just the activity in the cell, the phenotype of the cell or a reporter-based assay. So we can look at morphology, proliferation. We can look at mitochondria. We can look at mitochondria segmenting. We can look at mitochondrial spindles. And you have the advantage here now as you're treating the cell with a small molecule and you're interrogating thousands of proteins and black acids at the top. So what you're now doing is not looking for your keys under the lamppost. You're actually interrogating the whole cell to try to find some interesting thing. Now we'll see that it's actually hard to do this with small molecules. So if you want to find novel genes, you probably want to screen assay and CDNA life rates. Okay? It's far simpler than the deconvolution is far more straightforward. In the case of small molecules, the advantage there is that if we find a small molecule that does something interesting, we can take it into pharmacological models and give it on tests of sometimes more easily. Okay? So if you're doing this with small molecules, I think ultimately your motivation ought to be in needle models. So we're writing many cell-based screens and looking at various genetic and orphan diseases and neglected diseases and so on and so on. But I just want to tell you about a subset of those screens today. They came out of, as I was thinking, where should we go with this kind of methodology and where is big pharma not and so forth and so on. I was realizing I was getting older. So I thought, well, I ought to do something that's relevant to me as a person. So we started thinking about regeneration. Okay? And, you know, can we make body parts? Because I'm pretty seeming to need body parts. So, you know, and I always, you know, as a chemist, biology's kind of really fascinating. So if you cut, you know, the tail off a new, it grows back. If you cut your finger off, you have no finger. Okay? Why? Why are you going to grow organs and body parts and all that stuff? So we got really intrigued by this. Well, maybe we can take a small molecule approach towards this problem and learn something that's kind of a matter into what's going on in the biology world and some sort of biology world. And maybe we can find small molecules that affect these processes of differentiation. So if we don't want to reprogramming. And maybe we can test hypotheses and debate over these molecules. And maybe by learning how these molecules work, we'll actually gain some insight into the processes. So that was a motivation. And so we really got into this, not knowing anything about developmental biology. So we started out with more or less simpler systems. So the first thing we did was taking an advanced kind of cell screen and we took an MSc from my staff to ask whether we could actually find molecules. Normally, you all know these differentiations of the last kind of sites or different sites. And we said, could we find molecules that would control the selective differentiation of MSc's and say, hey, bum, selectively, or make hydrocytes. And so when you're screening large numbers of molecules, in general, the cost of the screen is very many. So you really want the first round of tasks to go from two and a half million molecules to 2,000 and then you can do almost any answer to 2,000. So if you want to look at osteoblasts, you'll look it up from phosphatase. If you want to look at kind of sites, the primary screen with an axiocene blue. And when we find the molecules that affect both of these processes, we'll select the legal osteoblasts and select the legal kind of sites. And we actually are looking at optimizing these tests and rolling the models with cartilage damage. But we did this. The first year grad student isolated one molecule we called chermorphamine. You can see we're somewhere between a chemical and a biological world. Because if we were biologists, we would have called this AKT or JAK or some three-letter acronym. But I don't know how anybody remembers these things, OK? If we were chemists, we would have called this the Schultz molecule, OK? It's kind of saying things after themselves. We actually called it a curing with a mean, but more, so chermorphamine, so we have another. And it actually goes different. Shade MSC is the BOM. And we confirmed this with a secondary reporter-based assay. And then we actually carried out assays of morphology and with other BOM-specific markers. And so the question was, how does this small molecule work? And so we began to carry out a lot of experiments, and this becomes a really challenging part when you have an act of small molecules, how do you figure out what it does? And in general, our approaches are either transfer-profiling, affinity-based approaches, which are the most successful, or CDNA-complication experiments, OK? In this case, the RNA expression analysis gave the clues when we actually did an analysis, a pathway analysis, we found that this molecule actually activated the hedgehog seedling. And we showed the molecules that intact on this hedgehog seedling, so just like both of them, it blocked the activity of the molecule. And in fact, Jen and Stanford subsequently showed in a competition experiment that this molecule actually binds and is at smoother naggings, OK? So we started out with not by a sulfate screen and actually hit on a major development, and agonists are made to develop an old pathway which said, well, maybe we will find interesting things here. So that kind of decided. And so we've subsequently gone in and set up screens to look for agonists and antagonists of hedgehog seedling, wet seedling, notch, and so forth and so on. And we have, again, we have some antagonists, you will in this team, but made antagonists of hedgehog seedling that will shortly go on to claim. But one thing we looked at was agonists and antagonists of wenced seedling and we found this molecule. We set up with Chi-shan and Shun-Ding this synergist screen, so looking for small molecules that would be synergistic with wen. And the top class reporter, I was saying, found this molecule and in the presence of 1-3-A, the report is activated roughly 200-fold. And it not only works in cells, but actually as a potent wen synergist and xenophilic you see increase in the partial axis duplication you see a weighty notochord phenotype and zebrafish. So it looks like it's actually a molecule that can be used in various simpler organic models. Another question is how does this molecule work? And so in this case, we may affinity reagents and we actually affinity purified this protein, that turned out to be RF gap. So it's kind of surprising because there are gap inhibitors described and is it an RF gap inhibitor? Well, RF, as you know, is a GTPase and a GTP-bound form is often GD-bound form is often and RF gap is a GTP-executated protein so it catalyzes a GTP-to-GDP hydrolysis reaction. So it turns out if you inhibit this with this compound, QS11 you actually see the GTP-bound form of RF6 and RF1. You can measure a direct interaction on the order of hundreds of animal or by vehicle or if you ever express RF gap 1 you block the synergistic effect of the molecule and the fel name blocks the effect of the molecule. So this is really how the molecule is functioning. So what we think we're doing is using RF1 which is involved in protein trafficking and trafficking in cell and that's activating acting as a wind-center just by modulating either an endocytosis of E. cadherin and beta-catenin or it could be affecting other proteins in the wind-signaling pathway for instance trafficking of wind-less outside of the cell. And so we're not, we know it's working at this level and it's modulating the pathway but the exact nature of this we don't know yet although we now have assays in-house to look at trafficking of these proteins. We then decided to begin to look at other cell types so we got together with Rusty Gage and so we'll actually begin to do screen-solving neural stem cells so we actually carried out a screen in neural stem cells isolated from adult rat in the campus and the screen was related to liquor molecules so they'd selectively induce mirrored genesis whereas if you'd read about acid you'd induce relatively unspecific differentiation so we found this molecule page that's 101 which is pretty active in cell-based assays it's about 5 micromolar in this case we optimized the properties of this molecule so that they would pass the blood-brand barrier and then we put these into rats and we actually are beginning to see when we treat rats and do look at neural genesis and proliferation we're actually beginning to see the case as 101, 6 makes for a K actually induces neural genesis and the rat-ventate gyros so it looks like now we're actually how small a molecule it appears to have in the activity in vivo which is pretty neat we don't know how this molecule acts we've done affinity-based experiments and we've got four proteins we've isolated that we're characterizing now making assays for each one of those we know that it does inhibit BMP4 induced astrocyte formation we know it affects M4 signaling but that's really all we know so once we identify the molecular interactors I think it would give more insight hypo, who's done that work is also beginning to look at glioblastoma in the right of stem cells so he's isolated these cells and proliferated them and they're actually quite tumorogenic and so he's carried out screens to look for molecules that will selectively induce apoptosis or differentiation of these glioblastoma in the right of stem cells and he's found this class of molecules and you can see this class of molecules induces apoptosis apparently selectively of these glioblastoma in the right of stem cells relative to other neuro-progenitor cells so sour scorn kills both relatively not specifically as a control and here you see at 6.2 micromolar levels we're actually sparing neural stem cells including glioblastoma in the right of stem cells so we're starting to move into this area now to look for things that might be useful to test the cancer stem cell hypothesis in vivo so then we looked at other we're really kind of, as you can tell, inserting those chemists to see what interesting things we can find and then collaborate so we then looked at hematophox stem cells working tummy boy Tom did this a post-hoc in my lab working with Mike Cook and so the idea was to take HSC purified population of human HSCs and then look at what we can control differentiation to so look at megacaryocytes, erythrocytes, macrophages and so forth and also look at cell renewal and in this case we actually found a molecule that induced differentiation of HSCs in the megacaryocytes and you can see these molecules are pretty potent one micromolar they made megacaryocytes in a dose-dependent fashion and consistent with that is to do a polygene analysis you see what we're making again megacaryocytes of the small molecule roughly the same concentrations and so the question was how does this molecule work and in this case we actually happened when we ran through our bank of kinase assays we wound up having a molecule that inhibited PDGFR and so that's the activity of the molecule it's a pretty potent inhibitor and likewise we took anti-PDGF antibodies anti-PDGFC antibody and you can see it has the same activity as a small molecule inducing megacaryocyte formation so we can induce over this a specific ligand specific antibody anisolongylapyloxacetate so it turns out PDGFR alpha is highly expressed in CMPs and if you look PDGF A and B are made by megacaryocytes and PDGFC by HSCs so our current model is that you're actually there's a negative feedback loop to PDGF being released by the HSCs and the megacaryocytes that actually regulates the differentiation of CMPs in our PDGF alpha our receptor alpha inhibitors actually blocking activity with these molecules and the negative growth regulation so the neat thing here is these molecules should actually be synaptistic with TPO so they actually might be useful clinically for making pills so now we've turned our attention to embryonic stem cells and we didn't know how to grow embryonic stem cells so we started out with very embryonic stem cells and then did this collaborating with Shenzhen who's a graduate student sorry this was done by Shu Wu in collaboration with Shenzhen so what we did is we took in this case for our height through the quick screen we used mal-CC cells and we used a reporter we were asking whether we could differentiate nearing the HSCs and the cardiomyocytes and here we generated a reporter in the Sifera base based on A and a half the Sifera's reporter pass right we went through the collection and found this molecule which again we characterize in a second series of secondary assays based on markers and transcription factors and expression and so forth interestingly if you take this molecule and show you our one ESC you do get a beating cardiomyocytes so it's kind of an interesting molecule again there have been reports that this actually has activity in vivo but we certainly haven't recreated those reports so we then began to look at Shaotan began to look at differentiation of other cell types so we really like to go from ES cells to endoderm to say the hepatocytes or the beta cells and so we carried out the first step of that where we actually carried out a screen again it's for Psi 17 positive Psi 7 negative cells and again these using these markers so the actual technical detail is running the screen across the fact that fashion has become important but we actually identified this molecule that selectively induces differentiation in the definitive endoderm ok it looks like a star of sworn animal so we thought this was completely non-specific but it turned out when you run this through there's a human piracy kind of collection that hits no kinase so it's not acting as a kinase inhibitor again we just carried out affinity purification experiments and we're trying to deduce proteins and bonds and we shouldn't have those seen but you are able to take this molecule and make the definitive endoderm and then differentiate both cells into albumin express and hepatocytes and this molecule does work in human embryonic sense all these one cells as well so here we have an example of a molecule that crosses between human and mouse and again it's useful for selective differentiation of ESC so then so we're beginning to find molecules that induce selective differentiation and we'd like to take some of those molecules into appropriate animal models for regeneration and we next turned our attention to asking whether we could find self-remolving molecules and so again because of a lack of extortates in the air we started out, Suzanna Tent started out with murineous cells and with molecules it would simply allow us to flow for ADFCs in a differentiated state and so the screen is really simple to use and I have four GFP, the core would be for grain and we found this molecule SC1 that actually allows you it's CC50, it's about one micromole and it allows you to actually expand ESCs in the absence of lip for feeder cells or what have you, that's how the screen was carried out without a feeder or lip and so you can actually pass its murine in brain stem cells serially with this molecule they maintain expression of ES markers and I got socks and they retain the morphology of ES cells and you can then differentiate these into the multiple sulfides and they can trigger the germline so this works well for mouse and brain stem cells it doesn't work for human and brain stem cells we just done the first round screen there and it actually isolated the number of small molecules now it appeared to work for human ES cells so how does this molecule work so this turned out to be kind of interesting and so again we carried out a affinity experiment so we make an active molecule and we make a dead molecule and we do differential affinity experiments and we found that the active molecule bound ERP1 and RASCAP so you can purify RASCAP and ERP1 and measure direct binding and it's on the order of 100 gallons or so so you can measure the direct interaction you can over express ERP1 and RASCAP and you advocate the activity of the small molecules ok and then you can go in independently accomplish the same thing so we can take another small molecule on part Davis compound that hits ERP and we can actually take an SI against RASCAP and in the presence of both of these you recapitulate the cell phenotype so in this case we serendipitously have two targets to the small molecule and it appears that both are required and so here's what we have we're blocking RASCAP again this is kind of surprising because there hasn't been reports of this and we're blocking ERP1 so we may be blocking differentiation signals but at the same time it's being telling cells to proliferate ok that's a simple model so then we said well this is an interesting approach if we can do self-renewal let's go back to adults themselves and ask whether there's an adult stem cell type that we really want to be able to propagate therapeutically so we went back to this hematopoietic stem cell screen and we asked whether we could find molecules that promote HFC self-renewal so we can expand HFCs in an undifferentiated state and do a bone marrow transplant and they would actually differentiate in the blood lineage so this screening demo is pretty simple a CD34, CD133 positive screen we carried out a screen went through the whole collection and found molecules that actually expanded HFCs in an undifferentiated state they work far better than HAC computers you can actually expand these for 21 days ok we then went back and took those expanded HFCs and put them in a non-skid mice and actually you can repopulate non-skid mice this is actually one of the first experiments we tended to say cells stem and stem cells you can repopulate non-skids pretty well and we've done howling forming assays howling forming assays and we do generate all predictors so this was interesting so then we went and talked to people who do bone marrow transplants and said what would really change the way to do bone marrow transplants and they said what you really want to do is be able to expand cord blood HFC so we just went in and isolated telling isolated HFCs from cord blood and now we can expand these so you see we're out for 35 days expanding these and they expand very very efficiently and again they differentiate into blood languages so we're now actually asking whether we can take cord blood HFCs HFC of spanna and use them therapeutically which would be a way to really quickly generate large numbers of HFCs you can use the bone marrow transplants and a large number of diseases where you don't have a problem with rejection that you do so that was interesting and we said okay so we're expanding we can expand embryonic stem cells and adult stem cells while we just take thermally differentiated cells and ask whether we can just proliferate those so not even worry about controlling the differentiation process just take the cells that you'd like to expand whether they're dopamine or do neurons and modern neurons or whatever and just make more and so we set up a screen to do that and at the time I was called by the GERF for some reason they wanted me as an advisory group on the GERF which I offered to do but I started learning a little about type 1 diabetes so we said well we just do a beta cell proliferation and the real problem here is when we looked into it we couldn't get our hands on a beta cell to do a screen you know you talk about a billion cells 10 billion cells to screen 2 million counterpounds okay you're at 500,000 cells for a while so the trick we did is we did this reversible immortalization with large tannins and so we took beta cells reversibly expanded those in the presence of doxcycline okay you get an extended population you take dox away and now you have your screenings now they may not be actually beta cells but they're pretty good primary screen so we did a massive screen and found molecules that would proliferate these and you can make sure that they're not doxcycline thematic simply by reversing with the dox offers and dox on screen and so we began to find these small molecules and along the whole we found you know four-ball, you can specify four-ball it's proliferated cancer-causing molecules we found a novel went agonist so when agonists were at the same time that they would came out right after we found this it said bio proliferated beta cells and we found a lot of other interesting small molecules that we done on the mechanism of action interestingly we found a cell type calcium channel agonist it turns out one of these had been in people 20 years ago for psychiatric disease so these had actually been in people so the idea that you're going to use a proliferated molecule to expand beta cells in vivo and give somebody cancer is for sure that's the menace because it's been in people and so we actually did took our first shot at an STZ model or we would bled the beta cells we probably did too high a level of STZ we bladed 90 to 95% of the beta cells but you can see as we began to dose with an alloy of this molecule glucose had started to come down so we got pretty excited about this and so we actually did histology on these mice and we actually see proliferated beta cell population so in the first experiment this looks actually really interesting we had to go back and run better models and we're collaborating with the Baldor's lab and using his model and other models as well but I think this approach could be interesting and maybe I've put a goal we'd love to do taking the population of motor neurons or dopaminergic neurons to actually ask whether we can direct these scan nodes as well or hepatocytes or perinealic sites so this could be an interesting approach for the regenerative therapy so now a few years back probably 10 or 15 years ago we got interested in the idea of how we can actually carry out the program and we thought about this and finally Shundang and Subhan Shah and I were on the screen we actually didn't know quite how to do this at the time so we did with a phenotypic screen or we took a mild blast which normally in the presence of DNA will differentiate the these conditions in the mine tubes and so we took this molecule and screened a collection of molecules of the idea that we really didn't know what the marker here would be but if we did go back if we did reprogram then we could differentiate gaseous labs or dipocytes under osteogenesis or dipogenesis conditions and we found a molecule and the reagents and in fact we found a molecule we called reversum that will allow you to go through your reprogram and then make osteoblasts or make adipocytes and these work at a clone level so we're not just selectively killing cells and again in this case the affinity experiments confirm my big course say we're binding two things and go to their sides and I say to our names that hit both parties I've got one small molecule so this molecule is kind of our first step into this so all we really know now is we're modulating activities of these molecules we're not sure whether there's cell cycle effects clearly what else is being affected whether the gang annihilation status is being affected and so forth we're looking at but we've begun to ask whether we can generalize this so we went into another system where there was no example of plasticity by condom and all in and that's in the case of the ligand endocyte precursors cells OPCs and here we set up a screen to ask whether we could take OPCs and actually reprogram them for a few months instead of a ligand endocyte and so we set up a screen here and the screen wants to go backwards to a neural stem cell neural stem-like cells using Cyc-Studio P-reporter and we carry out the screen and then we add neuron-inducing conditions and so here's the screen we can carry out train-style molecules they actually could up-regulate the Cyc-Studio P-reporter and then you can actually deform neurons and all eight molecules we identified were H-stagnant inhibitors so that may not be surprising so here they are H-stagnant inhibitors you can see when you treat the OPCs with these inhibitors you actually get down-regulation of OPC lineage markers and up-regulation of NSC lineage markers and if you do a chip analysis you actually see changes in the R1 focus of the Cyc-Studio P-reporter that are consistent with this reprogramming so now what we're really going is to run many massive screens looking for molecules that reprogram HSCs and various so many just derived from HSCs a pretty client system and we also set up a collaboration with the English lab where we're taking maps with a male on Luciferase reporters and then we introduce I-4 and then we introduce pairwise combinations of I-4 and Cyc-2 or I-4 and K-4 and look for replacements so one strategy is to march backwards where you take all four factors and then form an I-CSO and the lead one or two at a time to see what you can make up for those of the small molecule and so we've done screens of a hundred to two of a million to two million molecules with probably five combinations right now we're working through the data but here's one where we took I-4 and Cyc-2 just these two and added these two glasses of molecules and you can see you're actually forming the right cells so here's a mess a negative control area of C and so these are what these two molecules do so we may be able to now have a molecule that replaces K-4 and MEC and as I said we're taking all fossil combinations, singles, doubles and triples so this is the direction we're headed in in this case so I think maybe ultimately we can find chemical factors and do this so you can also have fun with this so I was sitting around talking to a colleague of mine and Arden is one of the hot and cold receptors so the cold receptors are these trig channels and so he has channels that are separate to high sensitive to various temperatures characterized by a lot of characterization so Arden and I said you know what we added a screen for small molecules that activated cold okay and so we just started screening a million molecules and the idea is, think about it we got a molecule where you want something that turns on cold that's absolutely odorless and tasteless okay but it's not really uh, systemically there's no systemic exposure swollen because then you can go out on your boat or go to the ballgame and take a can of beer and just have a little bit of this molecule like a spark tank and you'll have a cold drink okay, so it's pretty nice you know and maybe in Wisconsin you can actually do the same thing with a hot receptor and put it in your clothes so we're doing things like this which could be kind of fun at all we're also, as I said, starting to use these tools now big farm and farms act up with major diseases but a lot of orphan diseases have not been really looked at from our research to the cancellation effort so we began to set out we're actually setting up an institute to look at orphan and neglected diseases and so we really applied these tools and we're learning a lot about the genetics RET and SNA and so forth and so on but here with Elizabeth Wensler we did a screen of malaria so we just took malaria and we did a red cell and adhesion screen and blood cells we actually did a few million molecules and we actually went through and we have a database now at GNF where we have 300 cell based screens across 2 million molecules so if you get a hit you can go through here and here are all the screens and here are the hits these molecules look at this molecule well, it does nothing but it hits a malaria it hits Jack too, just a little so all of a sudden we have a huge collection of molecules that we can just go through in the end of the day so we have molecules that just hit malaria and so we put together a collection of 5000 or 6000 confirmed cell based hits that we're proposing now the Gates Foundation distributed to the whole academic community and we've done the same thing with the merculoses and the Schmeine-Ian other diseases we've also gotten to look for endogenous small molecules that affect developmental pathways so instead of looking for synthetics let's just go in and look at endogenous small molecules so we have a corner around the field pig market in Indiana at the back and we're grinding up field pig organs and spraying those for molecules that will modulate these processes, either differentiation cell from home and so we found now molecules that actually will activate we found both agonists and antagonists for a hedgehog signaling that are in 5 minutes and now we're trying to sort out their structures we think one of these is a monogamous steroid family plan we've also in GMF now kind of had the idea let's go back and we built a lot of tools let's just go back and look at proteins and peptides that affect developmental pathways and other pathways so people have looked at secreted peptides and proteins before there are about 4000 and only about 15% have been characterized so we actually said can we actually go into this set of 4000 express and purify all 4000 and take purified proteins and then run them through cell based assets now with our automated cell based proponic system we can run about 100 to 200 assets at a time so we thought that would work and so we're also doing this with conditioned media but so we took 700 genes that we want to do this year and we didn't know you know how to express cloning expressed purifying assets 700 proteins in the amount of cells so again we took the engineering group and GMF together with Scott Leslie and others we actually built this automated pure protein expression purification system where this system will automatically grow up in the million or insect cells up to 400 cells at a time distinct expression constructs at a time we can do this transient or viral expression we grow up we isolate roughly a milligram of protein per expression construct and we automatically lice these are the static centrifugal lice if any purify run these through a series of biophysical and biochemical assets to live for integrity, homogeneity and so forth and so on so we built this system it's actually impressive in person and we took a set of 48 proteins okay half chosen random and half is kind of controls and in the 48 the first pass through we made both N and C terminal fusions the various fusion constructs CIS tags or FC fusions we ran these through 75% actually gave purifying isolated protein and you could isolate and was active in the first round pass 75% so we were we thought it might be 10% so we're pretty excited so if you take that number and apply it to 4,000 to 2,000 so and if you run these through a panel of assays we have now 100 assays that we're assaying these proteins again 0 well here's glucagon okay there's PCSK9 okay there's ILA and they're all coming up exactly where you would expect them to come up so now what we want to do is take this collection now and also look at reprogramming with this collection of proteins to see whether we can get a combination of sure how the proteins that we can do reprogramming or anything so for now now small molecules are neat and if you find something you can test it and be well but it's the hard way to find model genes so if you're just interested in finding model genes as much as people talk about chemical genetics chemical genetics is harder than hell compared to genetics so what you should do is just take us to RNA libraries, lentiviral libraries transfection libraries, cNA libraries and screen needs we think this is a lot easier so again we built these large collections Tony Morris and Jeremy Caldwell and others built these huge collections of array SI and cNA by the way we had these collections available to academic collaborators and also small molecule collections and so this is where we're going to reprogramming and differentiation where we're screening these collections of SI and cNA so here we're looking at methylation status both from a cancer and reprogramming perspective so that we built p16 and oct4 reporter based screens looking at the methylation and so the readout is a leciferase readout and we're doing these screens now with small molecules, sRNAs and cDNAs and so here's the first round of the cDNA library so here's the control GAD45a and these are the first seven genes that have come out including sRNA4 so I would like the interesting in the sense that we actually isolated things that appear to be even more interesting than GAD45a you can also use the sample products to look for non-coating RNAs that modulate processes so we went in and with John Hoganich we made a collection of sRNAs that targeted 512 conserved non-coating RNAs between them and now these are larger non-coating RNAs so you can just make these sI libraries and screen them and here we found molecules that affect head-talk signaling and also non-coating RNAs affected infant signaling and the guy who was in the state of calling these non-coating RNA repressors of infant signaling is the N-line but this is a real result so it turns out this non-coating RNA appears to be a repressor of infant signaling and if you look at N-line expression it's in lymphoid cells if you look at the expression tissue expression tables we actually did pull-down non-coating RNA and put it on a finished support and we actually pulled-down important beta inside the protein that showed actually that N-line actually interacts with this important beta and protects it in an RNA protection experiment and the knock-down of this actually increases infant and infant levels so the current model and specific current facts so the current model here is as we think this non-coating RNA is a modulator of trafficking and perhaps for specificity and so the final as we went further I said, oh guys, you know why don't we just randomly take mice and look for interesting candidates and so Chris Goodinale at the time was doing this and others so Steve Kay and myself and Colin Fletcher and others got together and we said, could we do this and so we did it on a pretty massive scale at G&N and this is a recessive screen so we breed out the DNA-immuted synthesis, the G3's and then when we ran these things there was a huge number of asses so we have mice that have no T cells G cells, you know, altered LDL levels and so on and so on we have a mouse that I come up with these weird screens because as a chemist you can call them weird stuff because a lot of just don't think you're a fool they just think you're naive so it's nice, you can have a really silly thing so I said let's feed the mice McDonnell's french fries and look for mice a bit thinner and we found the mouse it gains muscle mass and loses weight, we call it army I said, oh, we're mapping those genes but I went over to Rusty and he said, well, how do we screen these mice for regenerative phenotype and he said, oh, I'll do a spinal cord in any section and I'm like, Rusty we're doing 200 mice a week, what are you talking about? okay, so I said come on, I said hey Rusty how about if we just give them urine holes okay, and he said that'll work so what we did is we punched holes and these all mice are 200 per week, we just looked at a one young phenotype and we actually have this mouse that has a one young phenotype but there's actually more than one young one there's Carleage there are blood vessels there are hair follicles so it almost looks like a regenerative phenotype and so we've actually sequenced the mutation and matched the TGF, they are one and it's not a constituent of active or inhibited it's kind of a kinetic mutation and a signaling pathway and we've actually mapped the mutation so we're actually characterizing these mice now in more detail, looking at counter-genesis of these mice looking at immune system function and so forth and so on and looking for other phenotypes for instance, heart regeneration and so on and so forth and then finally you can go all the way with this idea so you know, biology solves problems by making many possible solutions and simply selecting the best one so one day I was sitting in a meeting with a bunch of physicists and he said, how long are you going to be working on white people? I know a lot of physics community are working on these things and the answer was is they didn't know how to make anything more interesting so that is the only game in town so and the reason is is if you're going to look at the materials of these really complex materials with all these elements five at a time in theory it is absolutely useless and advanced materials problems so I said wow, what are you going to find interesting things by just doing what the immune system does make huge libraries work for the immune system it works for small molecules and it will work for elements so I was watching my son at the time crayon, have you got any stencils here? do you have a stencil? or these are classic and they're half blocked here instead of blue crayon and he paints a little blue here if you rotate the stencil you get a yellow crayon you get yellow, you get orange, you get red so then if you take all the stencils you get all combinations of the colors so all we did was replace the stencils with metal mats and the crayons with laser-bladed L-oxides and you can make this library thousands of them and you want a blue phosphor here's a new blue phosphor right there so all of a sudden you're going to make any kinds of materials this way Henry Weinberg in his group really took this to the town they were able to make all kinds of new things so now we're playing this back and actually getting back to the materials and we're trying to make hydrogen storage materials and lithium silicon 10 negative electron batteries for lithium ion batteries to kind of solve the energy problem for these, you know, pretty shy so that's the story of the last two days is what chemists do better than anybody is make interesting new structures and what we really like to do when chemists are 10 or 100 years is make really interesting new properties and do so in a rational way and so our kind of approach and philosophy is really to look how nature made these interesting molecules and the strategies nature used combine that with chemistry and apply it to a huge number of interesting problems in chemistry, biology and physical sciences and again I tried to point out the people that did the work and the work was done by this is my current group and these are the collaborators especially I like to point out Shu Wu Charles Cho who right now we're doing really extensive collaborations with American T&F and have also appointments at TSRI and again all the work was done but I wanted to really throw it at the people that I really improved this to work with and again I have a terrific time here and thank you for your attention