 Just give us one minute. Welcome to today's open research webinar hosted by Elife. This series aims to give early career research as an online platform to continue to share their research as an alternative to in-person gathering. Anna Akmanova. I'm a deputy editor at Elife. And my own research is focused on the cell biology of the mammalian cytoskeleton. It's an exciting moment for me. It's the first time I am sharing a webinar. And I think it has a very important function to connect people in this very difficult time when we're all confined to our homes and trying to make the best out of it. And we are particularly excited here today at Elife to be able to give some of our young researchers who cannot travel to meetings anymore and opportunity to present their research and to connect to many of you. Today, we will hear talks from Sajal Davlon, who is a research associate at Advanced Science Research Center at the University of New York in the United States. James Thomas was Dr. Othello at St. Hutchinson's cancer research center, also in the United States. And Roger Amamoto is a post-doctoral fellow at the Harvard Medical School, also in the US. OK, so before we start with the talks, just a few explanations of how we're going to proceed. So each talk, which will take 10 minutes, we will have five minutes for questions for the speaker. Then we will move to the next talk. To ask a question, you can type it into the chat on Zoom or directly in the Google document linked here. The link is also present in the chat. We are joined today by Miran Den, Anya, and Naomi from Elife, who are working in the background to support you. They will help me line up your questions. I will read your questions out loud and include your name where possible. The open notes document is also a place for you to continue shared public notes. We welcome you to do so and to list yourself as a contributor in the list above the speakers for today's webinar. Thank you. Finally, I'd just like to let you know that we are also recording the webinar and live streaming it on YouTube. During this live webinar, we ask you, please, be respectful, honest, inclusive, accommodating, appreciative, and open to learning from everyone else. Do not attack the mean, disrupt, harass, or threaten others or encourage such behavior. If you feel uncomfortable or welcome on any of these webinars, please contact Elife by email. We have an inbox which is watched by Miranda Nye at Elife. Elife organizes a reserve right to ask anyone to leave and to deny access on the subsequent webinars on Zoom. If you need any help at any point, please send Miranda Nye or Naomi a chat message directly using Zoom. Okay, so we will now start with our first speaker, who is Sajal, and she will tell us all about it. The title is ANT1 Functions and Astrocytes to Regulate Sleep Homeostasis. Sajal, welcome to the webinar. Thank you, Anna. So I begin to share my screen. Just let me know if it works. Are we good? Can everyone see my screen? Hi, Sajal. It's Naomi. Hi. Do you hear me? Do you see my screen? Yeah, so we can't see your screen right now. If you try exiting and share screen again. Does it work? Not yet. Because I see it here. I did the share screen. If you close presentation, are you showing a normal PowerPoint? Yes. You're not presenting right now. PowerPoint window. How about now? Let's work out how to do this in the background. Should we go to James? James, would you be ready to go first? And then we'll come back to you, Sajal. In the background, we'll work this out. Yeah, I can present this one. Okay, excellent. If we can do that, let's do that. Sajal. Yeah. Okay, then we will first start with James. James, welcome and please share your screen. We'll see my screen here. Yeah, I see your screen. Great. Sorry, I was pulling up the chat to make sure I could see it. Okay, great. Yeah, thank you so much everyone for organizing this and having I'll start my video. Okay, perfect. So everybody's good. Okay, let me know if you can't see anything. Great. So yeah, thank you everyone for having us and yeah, participating in this during this difficult time. I hope everyone is doing well. Okay, so my name is James. I'm a postdoc at the Fred Hutchinson Cancer Research Center. And I'm going to talk to you today about RNA splicing and new tools that we've been developing to study RNA splicing. So here I'm showing you a very simplified gene model. So there's an Exxon here that has a stock code on and then Exxon downstream, which has a translation stock code on. And these two coding sequences are separated by an intron. And during the process of RNA splicing, the introns are removed and you can kind of think of the introns as quote unquote junk. So this junk gets removed. Now the exons are ligated together. And now this coding sequence can be translated to make protein. So you can sort of think of the coding regions as being the quote unquote important parts of the message. And the importance of these coding regions are consistent with the conservation we see. So here I'm showing you these blue regions here are regions of high sequence conservation. And if you just take a quick glance, it looks like the intron really has no conservation whatsoever or upstream. But really I've been hiding these regions of incredibly high sequence conservation across this gene. And these are actually regions known as ultra conserved DNA elements, which are defined as regions greater than 200 base figures in length that are perfectly conserved between human mouse and rat. And also very highly conserved and even more distant organisms such as the puffer fish. Surprisingly, most ultra conserved elements are in non-coding regions of the genome. And they're often located within or adjacent to essential genes. So DMRT3 is a gene that's known to be important for viability in mice. So if you knock out this gene, there's phenotypes you see in a mouse model. So you might think, great, let's delete one of these ultra conserved elements, since they have all these remarkable features, certainly the animals would have viability defects. But when this was done, using a single ultra conserved element knockout, the mice are viable. And there was no postnatal phenotypes. So this was very surprising. It turns out actually more recent studies have shown that combinatorial deletion between two ultra conserved elements actually begin to unmask some phenotypes. And this suggests that these ultra conserved elements could be functionally redundant. So here I'm going to show you another example of some ultra conserved elements in other surprising regions of the genome. So here I'm showing you the gene structure for SRSF3, which is an RNA binding protein. And again, the coding sequences are very highly conserved. But again, I'm hiding something. And that's this other exon that has incredible sequence conservation across it, but also very high conservation within the flanking introns. One other thing you might notice about this exon, which might be a little surprising, is it contains a premature termination codon. And we call this class of exons poison exons. And we call them poison exons because when an exon gets spliced in and introduces a premature termination codon, this is a signal for that transcript to be degraded by a process called nonsense mediative decay. So essentially poison exons poison the host transcript. So you get the exon in and the transcript gets degraded. These poison exons tend to be enriched in many RNA splicing factors, such as SRSF3. And they are speculated to provide, speculated by also some direct evidence showing that these provide a mechanism of auto regulatory feedback. For example, SRSF3 encodes an RNA binding protein, which can regulate the inclusion of its own poison exon. So this sort of provides the cell a feedback loop to maintain the levels of SRSF3 and other RNA binding proteins. A lot of these poison exons, we can actually see them in even more distant organisms, such as certain molds. Again, really highlighting the high conservation of these elements. So again, intriguing and almost obvious hypothesis is that these poison exons must be important for organism viability. The challenge though with studying poison exons and really splicing in general is that there aren't until, you know, we hope recently, you know, great tools to study these, especially in a high throughput manner. So again, this is an isoform that would contain a poison exon, and this would be degraded by nonsense, maybe 2k, and the isoform on the right would make a functional protein. So we want to ask the question, what happens when we just delete this isoform? Unfortunately, we can't use RNA to do this. You can envision designing an SI RNA to target the poison exon, but that just copies what the poison exon does in the first place, which is knock down the message. And really we want the message to go up. Another strategy would be to use an antisense oligonucleotide to block RNA binding proteins that regulate the inclusion of that poison exon, but ASOs are low throughput, and it can often be challenging to predict which ones will be effective at blocking a certain exon's inclusion. So we decided to take advantage of CRISPR-Cas9 to delete these poison exons, and more specifically use pairs of guide RNAs in a Cas9-expressing cell to delete the exon. And this was inspired by this paper and others showing that delivery of pairs of guide RNAs can lead to deletion of large genomic segments. So here we can get up to kilobase size deletions. So the model that we started with was to delete the exon, so remove that region of the genome. So now you take away the cell's ability to even make that isoform, because the exon important for that isoform simply doesn't exist. So again, GDNA editing can enforce the production of a specific isoform. And now we can ask what happens when we just lose this isoform. So we're calling this technique pickfarm. This thing is for paired guide RNAs for alternative exon removal. And I'll show you how we wanted to adapt this to our studies. So the first thing was we wanted to show that we can use paired guides to deletion exon. We then wanted to use this for high throughput screens. And then we wanted to apply this into an in vivo model, hopefully to study some biologically relevant or clinically relevant systems. So to get pickfarm working, we first looked at an ultra conserved exon in an RNA binding protein called MDNL. I want to note that this isn't a poisoned exon. This exon is actually an ultra conserved exon that contains the nuclear localization signal for MDNL. So just keep that in mind. So we designed a variety of strategies to deliver paired guides to delete this exon. We can use Sanger sequencing to show when we turn on Cas9, we get the expected GDNA deletion. This is an RNA RT-PCR gel where we can show that at time zero, you get about 50-50 ratio of the two isoforms. But at day seven, you get a great reduction in the exon-5-containing isoform. And then most compelling to me, we can do a Western blot for MDNL. And without Cas9, you can see that both MDNL isoforms are produced. But when we use paired guide RNAs to remove that exon, we get specific knockout of the exon-5-containing isoform. And then again, I mentioned that this contains the nuclear localization signal for MDNL. And when we do immunofluorescence, you can see in non-targeting treated cells in here in orange, MDNL, accumulates in the nucleus of the cells. But we can see a loss of MDNL and many of the cells in our paired guide RNA treated cells. Great. So now that we knew the technology work and especially that it worked in a polyclonal setting, we can now apply it to high throughput screens. So for this, we can package a variety of paired guide RNAs into lentivirus, add the cells, add the lentivirus to cells and then let the cells grow over many population doublings and you then would count the guide RNAs before and after. So the blue guide RNA must have killed the cells and maybe the red one would help the cells to grow even better. We targeted 500 poison exons, 500 coating exons, and we typically target the three prime supply site of the exons. I'm skipping many details here, but the technology reveals coating exons that are essential for cell viability and also poison exons that are essential for cell viability. Regardless of what cell line we do it in, we often see qualitatively similar results. We can confirm that for specific substrates, both deleting a coating exon and the poison exon decreases cell viability. And then just in my last two slides, I'll just mention we also got curious in these exons which seem to have an anti-proliferative effect. But for this, we really switched the model. We added paired guides to cells, put them in a mouse, sequenced them, the guides from the tumors and found enriched exons. And then when we pick one of those exons, we can show that when we delete that exon, it actually increases cell growth in vivo. So here's the summary. Pick Farm is a tool to study isoforms. And I just want to highlight our lab and say thank you to everyone who contributed and our collaborators. And then I'll just put this up while I take questions. Great. Let me turn off. Great, I can hear now. Sorry, sorry I rushed through at the end, but I see this question stopping up. Thank you. Very exciting story about poison exons. I'm sure some of you have never even heard about them. So the first question comes from Valentina Snitkova. Wonderful talk, James. Are stop codons more conserved than surrounding nucleotides and poison exons? Yeah, that's a really good question. I would say that for the classically defined ultra conserved elements, these are defined as regions where all along a stretch of 200 nucleotides with perfect sequence identity. So in this instance, at least between human mouse and rat, it would be perfectly conserved. So we want to see enriched conservation for the three nucleotides itself. I think for other highly conserved exons, that might be a possibility, but I haven't listed that specifically. Okay, thank you. Then we have a question from Ashwariya Sigur, sorry if I mispronounce your name. Is the ultra conserved region in the coding region specific to start and stop codons? Yeah, that's another great question. It's not. So you might remember I showed the ultra conservation around the poison exon of SRSF3. And you can see there that the conservation extends throughout the exon itself, but even into the flanking and tronic regions, suggesting that there's other important regions beyond just the stop codon. One of the sort of models for why that's the case is again a lot of these poison exons can have auto regulatory feedback mechanisms. So SRSF3 would bind to its own poison exon to regulate its inclusion. So you might envision that the binding motifs for SRSF3 might need to be conserved in order to maintain the regulation. Thank you. Then we have a question from Jacob, our Colte. Do you see off target effects of your PGRNA class nine exon skipping method? Yes, we look at this a lot. And we look at it from a variety of ways. So both at the DNA level and RNA level. So first I'll speak to the DNA level. We did a lot of work using Sanger sequencing of topoclone fragments, I think, for maybe 20 different exons. And we can see that we get the expected on target effects of the pair of dives. So we at least we know that we're hitting the right spot. Most importantly for us though is the RNA level. And we want to make sure that we're just disrupting the inclusion of that single exon and we're not just messing up splicing for the entire gene. So you might envision if Cas9 is doing crazy things to the gene structure, you might get the accumulation of all these strange cryptic splice isoforms that would have undesirable functions in the cells. So to address that, we did a lot of RT-PCRs where we took primers and moved them back across the entire message. And we can show that we just get skipping of the intended exon. And we also did a variety of RNA-seq studies to look for low abundance splice functions. And for all the exons we've looked at currently, it seems to be highly specific. Do we have time for one more question? Yes. Alison Anders would like to know, do the ultra conserved regions contain conserved RNA structures? Yeah, that might be the case. It's not something I've looked at. I mean, I think that there's algorithms available to predict RNA structures. So I think an analysis like that is relatively straightforward and something we can do. I personally haven't looked at it. In terms of structures that I do know about, there's work, and I'm liking on the group that did this, but there's work showing that ultra conserved elements are enriched in certain domains of chromatin or kind of DNA-DNA interactions. But I don't believe, I can't remember if that study looked at only ultra conserved enhancer elements or all ultra conserved regions, but there's also some research showing that these could be enriched in certain DNA elements. So, yeah. Okay. Thank you, James. Time is up. We'll move on to the next speaker, but thanks a lot for the very exciting talk. We'll move to Rogi, who will share with us his insights on transcriptional profiling of specific cell types. Welcome, Rogi. Thank you. Let me share my screen. Can people see me? Okay. Thank you, Eli, for giving me this opportunity during these really difficult times. My name's Rogi, and I'm a postdoc at Harvard Medical School at the Kani Sapkola. And today I'm very excited to share with you our new method called probe seek, which allows transcriptional profiling of specific cell types from heterogeneous tissue using fish. As many of you probably know, you've worked with organs or different tissues, and they're usually composed of many, many different cell types. Deficted here is just one example. It's the retina, which allows us to see. And you can see I'm just depicting the seven major classes here, the rods, the cones, bipolar cells, horizontal cells, amocrine, RGC, and mule glia. But recent single self-strategies have identified more than 100 different subtypes in the retina. And this is not unique just to the retina. It's also happening in all the other different organs, different tissues, and different organisms. So when you're thinking about doing an RNA-seq experiment to, for example, let's see, let's say you want to understand what's going on in one of your cell types of interest during disease, then you have to really think about what kind of experiment you're going to run. So, you know, there are different common transcriptional profiling approaches that are already available. And I'm going to just illustrate this by using a group basket as an analogy. So imagine this group basket is your favorite tissue or organ. They just dissected it. Just like tissue, this group basket has different kinds of fruits, and there are some more common fruits like apples here, and some rare ones like kiwis. Now, if you mash all of this, all the fruits up, and you make a smoothie, then what you're going to do is you're going to be tasting what's the most common fruit, like the apple here, but you're not going to really taste, you know, the kiwis. And this is similar to a whole tissue RNA-seq. In which, you know, you get the very common abundant cell types and the transcriptional profiles associated with it, but not really the rare cell types. More recently, we have this fruit salad method where you can take all the different fruits, basically all of them, and then cut them up, and then you can taste each different ones. But to get to the kiwi, you're going to have to go through all the different apple pieces. And this is something like single cell RNA-seq, which is great, and it's very transformative in biology. But, you know, for you to get the 1% of the cell types that you're interested in, you might have to go through 99% of unwanted cell types. So, in the middle is something where I'd like to think of it as single fruit bowl. Let's say you're interested in the very rare kiwis, then you can sort them out, you can pick them out, cut them up, and then you can enjoy just the kiwis. And that's something that's like cell-type specific sequencing. So, there are, you know, there are different transcriptional profiling strategies for better cell-type specific already. For example, if you're working with a genetically tractable organism like mice or drosophila, you can, maybe you might have a promoter or an enhancer for your cell type of interest, and you can drive the expression of a fluorescent protein like GFP, and then sort them out. Or if you have an antibody, you're lucky, you know, you'd be able to take them out using anti-bodies against cell-type specific antigen, and again, back sort them. But oftentimes, even if you're working with mice or human or drosophila, you're not lucky and you don't have these approaches. And it's also a huge problem if you're working with any of these different organisms, these understudied organisms, and, you know, antibodies or promoters are just not that available for these organisms. So with probe-seq, we had one goal in mind. It was to obtain deep transcriptional profile from any cell type, from any organism, potentially. And this is the workflow for probe-seq. And as a proof of concept, we decided to go with a fresh mouse retina where we know all the different cell types that are involved, and we decided to take out a very specific subtype of bipolar cells in this layer here, and it's about 2% of all the cells in the retina. So what you do is you can dissociate a fresh mouse retina and then you can dissociate and then fix in 4% BFA, and you get this fixed single-cell succession, and you can use gene-specific probe sets for your markers of interest. Here we use VSX2 and GRIK1, and VSX2 labels all the different bipolar cells and muller glia, and GRIK1, which labels a subset of those cells. And these probes are what we use for stable fish as well, and you can do this in tissue. And this was developed, at least in part, by our lab, the SEPL lab. And once you do this overnight incubation with these probes, you can put these green or red fluorescent oligos onto these concatomor tails, and you can use a different combination. And you can fact sort them and then put them into different tubes. Here we have the VSX2 positive, GRIK1 positive population, and VSX2 positive, GRIK1 negative population, and VSX2 negative population. Then you can de-crosslink and perform RNA sequencing. Now if you do this in tissue, this is stable fish, using the same probe, you can see that VSX2 is confined to the INL layer. These are all the bipolar cells in Heliglia, and GRIK1, which labels a subset of them. And if you use this on fixed single cells and go through facts, you can see that there are different peaks. This is debris peak, and then you can see this is a single cell peak, which is what you want, and other contaminants. And if you take just the single cells, and then you can see this negative population, and there's this bump that is for VSX2. And if you get just that, now you can see there's a GRIK1 negative population and a GRIK1 positive population. That is very prominent here. And if you look at them under the scope, well, the double negatives, you have the host staining, but you don't have the punctile of VSX2 or a GRIK1. Now the VSX2 positive, GRIK1 negative cells, they have the punctile just for VSX2, but not GRIK1, and you have the double positives. And based on the markers we use, GRIK1 and VSX2, we expected bipolar cell subtype 2, 3A, 3B, and 4 to be in the GRIK1 positive population, and other bipolar cells and mule glia in the GRIK1 negative, and all other cell types in the VSX2 negative population. If you do RNA sequencing and you look at different markers, which is in these different rows, you can see that in here on the right side are the GRIK1 positive population. You can see that BC2, 3A, 3B, and 4 markers are enriched specifically, but in the GRIK1 negative, all the other markers are enriched. So it seems like probe seek is profiling the correct cell types. And the question we wanted to know was, okay, well, if people want to use this, how good is probe seek compared to normal bulk RNA sequencing? So this is kind of the gold standard. You have a promoter. We use GRIK1 promoter, driving GFP. So here we electroporated it into the retina. And if you do that, about 72% of GFP positive cells are GRIK1. And then from that retina, it's from the unlabeled, unelectroporated area, we did probe seek for GRIK1 positive cells and perform RNA sequencing. And in the electroporated area, we fact purified live GRIK1 GFP positive cells using a very standard trizol approach and performed RNA sequencing. And the results were very good. We thought it was very good. If you compare the transcriptome of the live GRIK1 positive cells and the transcriptome of probe seek GRIK1 positive cells, we saw on an average a correlation of 0.78 indicating that high quality, you know, transcriptome can be obtained by probe seek. And if you're interested in just, not just using two probes, but if you're interested in four or five different probes because you're looking at very specific subtypes, then we thought we can do a serially multiplex probe seek where we had GAD1, GSX2, GRM6. We had an overnight incubation of all three of them, but we put fluorescent probes only on two of them, GAD1 and GSX2. And you can see this population of GSX2 positive cells and GAD1 positive cells. And then you can very quickly strip these fluorescent oligonucleotides, but not the probes themselves. And that only takes like five, ten minutes. And you can see that the labeling is gone. And then you can rehybridize the fluorescent oligonucleotides onto GRM6 here. And you can see a new population appear. So I don't have time to go into this, but we also did this with frozen human retina nuclei as well as Drosophila mig gut and fresh chick retina. And they all worked great. So that shows that maybe this can be possible with any organism. And in the future, I think that probe seek can be combined with a tactic to look at open chromatin in the specific cell type or do single cell RNA sequencing after probe seek just to get a more finer, finer view of your cell type of interest. And with that, I'd like to thank everybody in the SEPCO lab for being very supportive, especially Connie for being a great mentor and the Taven and Demecki lab biopolymer facility for the sequencing, immunology facts for and the imaging and the funding. Thank you. Thank you very much, Roger. Thank you for sharing with us the smoothies, the TVs, the fruit salad, and particularly the probe set. So I would like to start with the first question, which comes away from Chantane. Is this method also working for plant cells? So that's interesting. If there is a way to dissociate plant cells, well, I have no experience with plant cells, but if you can dissociate them and if you can get these things into the cell wall by permeabilization, I'm not quite sure which permeabilization works for plant cells, but plant cells have RNA and if the Nc2 hybridization works for plant cells in tissue, then it should work for plant cells. Okay. The second question is, what are some challenges that you would encounter if you combine probe set with single cell RNA sequencing? Yeah, I get this question a lot. So I think, so right now, you have to fix the cells and if you want to do a high throughput droplet-based method, then all of the end of de-crosslinking at the end for probe seek occurs with protease. So if you wanted to combine single cell RNA seek with the droplet method, then somehow you have to de-crosslink beforehand and then put them into the droplet, or you might want to use like a methanol fixation and that's something that we can try. But I think, at least right now, it's probably already possible to do a plate-based single cell RNA seek method. I see. Another question, what are the, like, what is your pipeline for introducing fluorescent nucleotide into the cell? Is it enough to permeabilize the cells after fixing or do you have some other tricks? So the probe has to, probe can go away and only once it's fixed and hybridized. So there's a tritonex step and if it's not permeabilized, I don't know if it can go into the cell. I see. Then we have a question from Julia Piazza, who found this talk very cool. What's your pipeline for... So, actually I think it was the same question, sorry. Can you comment maybe a little bit more about the tricks that you use because I mean you must have been optimizing this a lot. It's interesting because so I had a previous paper where we were using antibodies and basically if you, once other lab members in our lab got the favor of fish to work in tissue, a lot of the same things that they were doing in tissue, if you applied them to single cells, they worked almost perfectly. The only optimization that we had to do was they were using tween in their, basically all the wash steps and then I found that if you use tween, they clump up a lot. So we just have to get rid of that. But it wasn't that much optimization that has to go into it. Okay. Then we have a question from David Sun. The abundance of these gene markers, is it important, does the marker has to be very abundant for the method to work? Yeah, so all of the markers that we used in the paper, they were very abundant genes. The two things that we found that were important were abundance of the genes and the length of the genes because these probes tile the RNA of interest. So the longer the transcript, you have many more probes that can bind to it. So we found that both of those parameters are important. The good thing is that now that everybody's doing single cell RNA sequencing on basically everything, we have 10 to 100 of different markers that we can use. Okay. Pretty cheap. I think if I understand correctly, we have to wrap up. Is it correct Naomi? So thank you very much, Roger. We now finally move to Sejal. Welcome Sejal and thanks Roger. And so the last talk will be about this mysterious long abbreviation and its function and astrocytes to regulate sleep home spaces. Can you see my screen? Hey Naomi. Is it good? Yes. Okay, perfect. I can see that. Okay. Hi everyone. My name is Sejal and I'm a postdoctoral researcher at the City University in New York with the work that I'm going to talk to you about today is a work I did as a PhD student in John Van Maal's laboratory at McGill University. So as most of you know that the most abandoned cell types in the mammalian nervous system are astrocytes. So here is a picture of this single astrocyte in the cortex of a rat. As you can appreciate here that these cells have very profuse remarkable morphologies. And because of the structures they are key for neural circuit assembly and functions. So for example astrocytes, they release different molecules during development that are important for synaptogenesis. Because of their morphology where they sit very close to synapses they have important functions in neurotransmitter uptake and processing. For example, in expressing neurotransmitter receptors or enzymes. And because of this molecular feature they can respond to neuronal activity as well as they release different neuromodulatory factors in response to this neuronal activity. And because of all these key functions that astrocytes serve in the neural systems they are very important in different thesis mechanisms. So overall the field is interested in how do astrocytes differentiate and acquire specific molecular profiles? How do they influence CNS function and behavior? And perhaps maybe we can use this knowledge and exploit astrocyte to better understand and develop therapies for human diseases. So I study astrocytes in drosophila because they have cellular and molecular features that are conserved. And drosophila it's a beautiful genetic model system that is easy for in vivo manipulation and you can readily observe phenotype. So if you were to take a cross section for example here it's the larva it's a brain of first and start drosophila larva. If you were to look at the cross section you see two histological segments. So here it's called the cortex which is shown here in green and it covers all the neuronal cell bodies. And you have this neuropile region that has axon student rights and synapses but they do not have any neuronal cell bodies. Now at the intersection of these two histological segment you have the astrocyte cell bodies that infiltrate that beautiful membranous branches and connect with synapses and serve and functions. And here is if you were to label astrocytes in vivo with GFP that's what it would look like. So again the question that I'm addressing is how do astrocytes in drosophila influence CNS function and behavior? And I'll give you a small example today that's my PhD work where I studied this gene called A1 and how it modulates sleep in astrocytes. So before I jump into the results I just want to give you a brief introduction on sleep. That sleep is modulated by two regulatory mechanisms. So if you look here you have sleep traces of drosophila where on the x-axis 0 to 12 it's a light phase and 12 to 24 it's a dark phase and you have the sleep profile of individual flies either in black or in red and it resembles some of our sleep pattern now that we don't have to go out to work. So this sleep is more or less governed by circadian rhythms something we call a baseline sleep. However you can imagine a scenario here that if a fly is sleep deprived overnight compared to this control fly that is not sleep deprived the sleep deprived fly exhibits higher amount of sleep the next day which is something we call a recovery or a rebound sleep and this is governed by the homeostatic mechanisms. So one of the key questions in the sleep field overall is try to understand the mechanism that governs circadian or the homeostatic rhythms and there is one example that we know of this mutant which is in a gene called A nut 1 these mutants they do not have any defects in their daily sleep baseline rhythms which is the circadian sleep however if you were to look at it in the bottom right corner you have the wild side flies when they are sleep deprived in this black bar they have a higher amount of rebound sleep the A nut 1 mutant flies they exhibit even higher amount of recovery sleep compared to the controls so how does A nut 1 participate in these sleep functions so first of all what this A nut 1 gene is it's an enzyme called aryl alkylamine anacetyltransferase that acetylates an inactive monamine such as serotonin dopamine and octopamine so the key question is which cell types in the fly brain express this enzyme and in order to probe that I developed an antibody against this exposed region and we tested this antibody so what do you see here in the control animal that the antibody beautifully lights up different cells in the neurosystem of drosophila larvae however the animal that has a complete denlation of A nut 1 gene does not have any staining suggesting that this is A nut 1 specific antibody so I looked at the A nut 1 positive cells in the drosophila adult drosophila brain what you see here in the first panel in green you have the A nut 1 positive cells and magenta ages the antibody to reveal the brain morphology I further did the marker analysis using a panneuronal marker LAV and a panglial marker called repo which suggests that both neurons and glial cells express this enzyme based on the morphology of glial cells I can already pinpoint that these cells are astrocytes which I further confirm using a marker analysis where alarm which is a promoter which only expresses in astrocytes if I were to label astrocytes with nuclear RFP all astrocytes in the fly brain are positive for A nut 1 and this is sort of a snapshot summary of how we characterized other cell types where among all glial cells astrocytes are the key cell type that expresses A nut 1 and there are different subset of neurons that also expresses this enzyme so here is the model until now we have a very good idea about the circuit and how they generate behavioral output through different receptors how do astrocytes participate and here is the one example where I found the expression of A nut 1 in astrocytes so the question was do astrocytes alter the level of monoamines via A nut 1 and affect the behavior in this instance sleep and then the second question would be what monoamines are processed by A nut 1 so the first example, the first experiment I did was what happens if you were to knock down A nut 1 from astrocytes only and this was achieved with GAL4US system and RNAI so here is an example if I were to knock down A nut 1 only from astrocytes you can only see the neuron specific expression but all the astrocytes expression is gone the number of astrocytes remain unchanged in this manipulation here is the sleep data that was done in collaboration with Amita Segel's lab so in black and blue you have the control traces of flies as I showed earlier and in green it's the experimental animals so what we see is during daytime and nighttime the amount of sleep remains unchanged so the circadian sleep component remains unchanged when we knock down A nut 1 from astrocytes however in these flies if we were to do overnight sleep deprivation as shown here you can appreciate that the green bars here are the experimental animal and this is the recovery sleep 6 hours after the sleep deprivation which is significantly increased compared to control GAL4 and UAS flies and this is a very robust phenotype now this phenotype essentially is what was reported before for this A nut 1 low mutant phenotype so what I'm showing you here is here is the gene A nut 1 that do not participate in the circadian mechanism of sleep which is important for the homeostatic sleep and just to end here if I have time so what happens to the level of monoamines so when we look at the control verses the A nut 1 low mutant flies we do not see any difference in serotonin and dopamine levels under baseline sleep conditions however if that window is in the sleep deprivation there is a robust increase in both serotonin and the dopamine levels and if I were to repeat that experiment just by knocking down A nut 1 in astrocytes we do see a trend in these green bars where there is an increase in serotonin and dopamine so just to wrap up here what I showed you today we found this enzyme A nut 1 that expresses in astrocytes this enzyme is important to process monoamines this enzyme is important to drive homeostatic sleep in astrocytes I did not show you this data but it is important to drive baseline sleep and then this enzyme controls the levels of dopamine and serotonin under sleep deprivation conditions and with that I would like to thank my lab and our collaborators and thank you for your time and I'll be happy to take your questions thank you very much Sejal it was fascinating I never have been thinking about sleep deprivation and flies in my life but it's absolutely fascinating first question from Ariel how can the understanding of sleep and drosophila be applied more widely to humans or other vertebrates absolutely that's a fantastic question and you know just a bit of a historical perspective so you know in 2017 the Nobel Prize went through the discovery of clock genes in drosophila and turns out it's the same mechanism that operates and flies in circadian rhythms and sleep it's the same mechanisms that we humans have so for example the molecular clock is incredibly conserved just talking about the a01 which I investigated in drosophila astrocytes so a01 it's also present in the human brain especially in the SCN which makes melatonin the enzyme that precedes the first step in the melatonin synthesis so there's definitely conserved elements in terms of a circuit mechanism the monoamine circuit is also very well conserved so the kind of work we do in drosophila to understand sleep it's definitely very informative for humans sleep let me see we actually had a very related question are there any human diseases associated with mutations in a01 or any disorders has it been looked at genetically? absolutely that's another very nice question so we looked at the mutation data and there is actually a syndrome in human which is in the a01 related gene which is called the SNAT and these people have something we call delayed sleep onset phase so they are sort of like in a way nocturnal, their sleep phase runs from late at night and they have definitely problem waking up early in the morning so they start that day a little bit late so that is the one definite example we have where a01 is related to human sleep disorder we have a question from Catherine can you see the same effects in serotonin or dopamine mutant background do they have disrupted sleep overall? yes so there is like a long list of mutations in genes that are important for serotonin and dopamine pathway for example transporters or receptors and they do exhibit a lot of sleep defects what was really fascinating from that literature was most of them have baseline sleep deficits and you know when you if you think about it if you already have a baseline sleep defects and on top of that if you were to do sleep deprivation experiment the results are a little bit confounded so this is the only molecule that had that clean separation that you do not have a baseline sleep defect but you only have the homostatic sleep defect I think we have to wrap up is that correct Naomi? yes so unfortunately our time is up we had one more question left and we will have to stop now so first of all I would like to thank three speakers Sejal James and Roger for sharing their work with us today which was completely fascinating and very diverse and I would also like to thank all of you for listening and for asking questions our online research talk series will continue next week with three neuroscience talks on Tuesday at 9 a.m. British time in partnership with Neuromatch conference for online unconference for those of you who don't know what unconference is conference run bottom up and not top down for computational neuroscience starting on Monday you can check out their agenda online even if neuroscience is not your area of interest we hope that you will come along to support our fellow early career researchers and on Thursday we will be back with talks from Catherine, Madan and Elka so for updates about this series please follow us at elive community and you can find this schedule and the registration information online thank you very much thanks for joining and we hope to that you will join again thanks very much for sharing Anna thank you very much thank you very much thanks everyone for coming along join us on Tuesday yes absolutely have a nice day everyone thank you very much Naomi Miranda Anya for organizing this so well it was fantastic it was exciting and thank you very much all the speakers for sharing your science and starting this series and some of them these difficult times I agree thank you everyone take care bye bye