Everhart Lecture: Illuminating the "Dark Genome" and Membraneless Organelles Dr. Prashant Bhat

hi everyone thank you for coming to our second Everhart lecture series Talk of the year my name is Rachel Tam I'm the academics chair of the graduate student council and also part of the Everhart lecture series committee along with Pond and Syria and so the Calta Everhart lecture series is sponsored by The Graduate Studies Office and organized with the graduate student council the everheart lecture series

is a forum for outstanding Caltech graduate students to present their path breaking interdisciplinary research to a broad scientific audience we're excited to continue the lecture series after about five years and so I will now like to ask Professor Gutman to introduce today's speaker Dr Prashant Bhatt [Applause] okay well um thank you all for coming out um it's a great pleasure and honor for me to get

to produce Prashant to all of you uh who many of you already know um our guest speaker who will deliver this Everhart distinguished graduate student lecture Prashant is currently an MD PhD student in our joint UCLA Caltech medical scientist training program which means he is going to be both a PhD and an MD when he finally finishes his training in the next few years when he

started his career he started at Berkeley as an undergraduate received a ba in molecular and cell biology where he worked with Jennifer doudna who was a Pioneer in the development of crispr in fact won the Nobel Prize recently for her pioneering work in this and even before crispr was a household name Prashant was working with her on some of the very early characterizations of many of

these different crispr enzymes in bacteria after uh finishing his undergraduate work he came to uh Caltech uh well to UCLA Caltech did some medical school work but we'll skip over that for a moment but as a PhD student here at Caltech he's been in our biology program uh and has I've had the privilege of working with him for the last six or so years which has

been really uh you know not to make this about me but really wonderful for me and been really wonderful opportunity but what he's done is really uh incredibly remarkable his work has really uncovered some fundamental new mechanisms of how cells work and how three-dimensional organization of molecules in the cell can can control fundamental information flow in the process of mRNA splicing I'm not going to tell

you about this because you'll learn all about this in today's lecture but what I want to highlight is is sort of this new and exciting new this this exciting new paradigm in cell biology that Prashant has uncovered which really required um as was alluded to about this black series required real interdisciplinary work in developing new Cutting Edge experimental approaches analytical methods and working with many different

colleagues across many different divisions here at Caltech in recognition of as many important contributions Prashant has been the recipient of several prestigious Awards including the de Carmen National Fellowship uh the Ruth kirchenstein National research Fellowship from NIH and not not least uh this everhard uh lecture selection for this lecture ship uh and amongst many others and uh certainly many many many more honors to follow uh

Prashant has also demonstrated an important commitment to diversity and inclusion uh he um has played an important leadership role in the Caltech prism community and as a recipient of the Caltech diversity and inclusion award so um please join me in welcoming Prashant and we're delighted to hear about all your amazing work thank you okay can we hear me in the back okay well uh thank you

so much Mitch for that introduction and thank you to the Everhart committee and The Graduate Studies Office for giving me the opportunity to present my work today so I have two parts of my talk the first part of my talk I'm going to be talking to you about how we developed a new method from biology to look at the highest resolution picture of how three-dimensional molecules

organize in space and we did this without ever looking at a microscope and the second part of my talk I'm going to then take these maps and we're going to ask the question of why molecules are organized in this way what makes it that these molecules are arranged in three-dimensional space what does that mean and so the structure of the talk is pretty simple so the

first part I'm going to be talking about structure so we're going to be making molecular Maps together and then finally and then we're going to go into the functional implications of this so like what do these Maps even mean and then finally the applications of these findings to clinical medicine and to get everyone oriented on the same page I want to introduce you to the biology

that I think of and how we think and so this is the fundamental principle in biological systems and that is the concept of structure and function so this is not only biology but in all aspects you know like when you're at a symphony there's structure function relationship how well when you're listening to like the wind the percussion the string instruments each and every single instrument that

you hear has a different shape and because of that it makes a unique sound or its particular function so in a similar way biological systems are no different so here we have different organs like there's tiny little alveolar air sacs in our lungs that help us breathe there's also in our colon little appendages that help us reabsorb water and then our stomach and its structure is

very important for the digestion of food so in all of these cases structure and function are very important so what happens if we go many layers deeper well humans and all biological organisms are made of cells humans are made of trillions of cells so your lungs your colon your stomach and everything in between made of cells so if you took a microscope and you peered down

under it and you took a cell and you looked at it it would look something like this this is a schematic of a cell and what I'm showing you here is that here in the cell there's a membrane the membrane is like the skin for the cell it keeps the outside out and the inside in now what happens if we go inside past the membrane into

the bells of the cell we're now entering the cytoplasm so cyto means cell plasm is gel so it's like a jelly-like substance that keeps a bunch of structures together so what are these structures well these are called membrane-bound organelles so these go by many names and I'll just highlight a couple examples one of them like the mitochondria and this is where the energy is produced in

the form of ATP so every muscle you um contract or every blink you take you expend energy in the form of ATP and it's made in the mitochondria additionally we have membrane-bound organelles like the nucleus and this is where our DNA is housed so you get one copy of chromosomes from your biological mom one from biological dad and it gets stuffed in this organelle called the

nucleus and encodes the genetic blueprint in order to for you to look how you look feel how you feel and think how you think now the title of my talk was about membrane-less organelles and so what do I mean by that and in order to explain that let me tell you what a membrane-bound organelle is so membrane-bound organelle so let's take the mitochondria for example just

like the cell which had a skin or a membrane over its entire surface mitochondria and other membrane-bound organelles do too so what this would look like and the names are not important but this is what it the membrane looks like it's a phospholipid bilayer and all that you have to realize is that it keeps the outside the cytoplasm out and the inside of the mitochondria in

and in this particular case it's important that these molecules are not moving out and it's known as restricted molecular exchange so molecules are not moving out of the mitochondria so what about the cell nucleus well the nucleus itself is a membrane-bound organelle but for the past Century there's actually a different type of organelle that has been really elusive to molecular biologists and that is there are

these membraneless organelles that decorate the entire nucleus of the cell so these are membraneless organelles in that they don't have any enclosing membranes but they still have an inside and an outside but in fact when you look there's like this invisible boundary and it's a question of like what are keeping these molecules together so these represent distinct entities with an inside and outside but in fact

there's free molecular exchange into and out of the cell and there are many such examples so let's make this a little bit concrete and provide some examples so if uh you're a biologist or had you know took a biology class in high school or I'll explain it to you right now there are many different uh membrane-less organelles and the most famous example of which is the

nucleolus pleural nucleoli and there are many others like Speckles which I'll be telling you about later in my second half of my talk kahal bodies histone Locus bodies are so many of them and the basic structure of these is that they're organizing molecules around shared molecular function but even though that we know that they have shared molecular function we actually still don't know why they exist

and that we don't know what the functions of these nuclear structures are so that's a little confusing so let me explain that once more and that is if I were to take the member so basically well I told you about the structure function relationships and so now you're now I'm saying that well maybe we don't understand the functions of these things and so you might remember

from high school biology that the nucleolus is the site of ribosome synthesis so the ribosome is the machine that actually makes the proteins like collagen and keratin and insulin and things like this it's made by the ribosome but the ribosome has to be made somewhere and it's made inside of the nucleolus but here's the catch if I were to take a mitochondria a membrane-bound organelle and

rip off its membrane and put its contents into a test tube it would not work so mitochondria would not be able to make energy it would not make ATP now if I were to do the same thing with the nucleolus where I took the contents of a nucleolus I put in a test tube I shook it up I would still get a ribosome so that's a

little confusing so now prashant's telling you the structure not equal function so this is the conundrum that we're in but of course all these molecules are organized Within These very beautiful structures and we really want to understand why and today I'm going to tell you why but before I tell you why I want to walk you through how we even got to the point of addressing

this question this was a question that has been throughout the past several decades of like why are molecules organized in this way and people have had hypotheses but experimentally proving it has been a little difficult and for that I want to take you a little bit back into the 60s and 70s when people were actually looking at where genes are being mapped and I have to

talk about this because modern technology has enabled us to make these technologies that enable us to understand why our membraneless organelles doing what they do so again let's make this concrete so we get one copy of chromosomes from our mom one from our dad and so we have 22 sets of chromosomes which is what you see here an X and a y y in males and

two x's if you're a female and so it was painstaking to map where genes were at the time and that is if you were interested in a gene like hemoglobin it carries oxygen in your blood so that seemed like a very important Gene to know where it is and so it was found okay it's on chromosome 11. and so you know people Brute Force went through

and then it was keratin you know keratin there's a bit on chromosome 12 a bit on chromosome 17 but of course we have thousands of genes that make up our human genome and so this is naturally a very difficult thing of mapping every single protein inside of the genome so how do we map all the genes on every single chromosome well this is where the Human

Genome Project came into being and so this was started in the Clinton Administration and completed in 2000 in which we sequenced the entire 3 billion nucleotides of the human genome so scientists took the nuclei of cells from humans they took out the DNA it's six feet long it gets packaged into this tiny little space of the nucleus and it's made up of these letters a t

g and c and so from that we got every single chromosome every single letter we can read it from end to end and I'm going to just highlight a few little nuggets of information that was learned from the Human Genome Project and that is one of the surprising things was that only one percent of the entire genome actually codes for protein and a very large proportion

70 to 90 percent of the genome the DNA is copied into its sister molecule or RNA and we're now just starting to understand what these RNA molecules do another surprising finding was that we share 99.9 percent of DNA so me and you share 99.9 of DNA you and Albert Einstein share 99.9 of DNA and everyone it's it's quite remarkable and that the variation that we have

is just the 0.1 percent that we see in DNA and then another reason why people were really interested in the Human Genome Project is if we understood the Book of Life the genetic code then when typos arise or mutations in our DNA we might be able to understand how to fix it and so it was found that there's mutations that really occur in both coding regions

that code for protein as well as non-coding regions that do not code for protein so this is all fine and dandy and so in the past 10 years people started taking a different approach and this is around the same time that I started my PhD fun fact this was when I was in like the fourth grade and so it's we've come a long way since then

and so now in the past 10 years around when I started when I moved to LA there was a Consortium that was started called the 40 nucleon project four-dimensional nucleum and so it was trying to take the genome in three dimensions so we had the string of letters but now how do we look at these string of letters in three dimensions and so this is what

they did and we only knew a little bit about how DNA organizes in three-dimensional space so for example there are some structures in DNA so like this is a representation of DNA in normal cells and it creates this Loop structure where you have these two different DNA sites they find each other There's A protein that can bind to these DNA sites and stabilize this Loop so

it kind of looks like a necklace and if you think about this necklace here is a pendant and this pendant is a gene and it's an oncogene so we want this to be turned off and in fact in normal cells right now as we're sitting and breathing there are these oncogenes that are just sitting there but they're insulated they're not being turned on but sometimes you

do get a mutation and in cases like glioblastoma brain cancer where you get a mutation this leads to disruption of the structure and loss of insulation and therefore this Gene turns on and you get cancer production not good so obviously we have to understand how exactly these structures form and actually what these structures even are because I can actually count on two hands the number of

examples like this that we know of many more are still to be discovered and in fact we discovered we developed a way that is going to be able to discover these structures so how do we map all the molecular interactions simultaneously I'm going to walk you through the solution that we developed but before that I'm going to do a little bit of chalkboarding of how or

whiteboarding whatever you want to call it um how we actually determined uh what strategy to go down so let's say I want to take a picture of the cell nucleus and I want to look at Region a b c and d so to give you examples so region a has these hemoglobin genes B has these genes that are encoding for keratin C is the other chromosome

that encodes for keratin and D is something else and so I want to take a picture of all these things and so what I could do is I could use a microscope and I could add a fluorescent fluorescent molecule on these different segments using technology I'm not going to really get into but basically you can label specific segments of DNA and say okay well these two

regions are in the same three-dimensional space well now what if I want to know where the Keratin genes are I want to know that they're far apart from each other and they're far apart from the hemoglobin genes well we can add a little like a couple more colors but now it gets a little tricky because the disadvantages are so at the time of when I started

this project in um 2016 um we had a finite number of colors you kind of only have the colors of the rainbow which are six or seven that you can choose from so you can really only image six or seven molecules at a time there's another limitation in that the resolution is limited to 0.2 microns since they give you a sense that is quite significant in

the context of a cell because if you have region a and region B and they're 0.2 microns or shorter distance near each other in reality they represent distinct areas of the cell but when you look under the microscope and get your data you're going to think that they were at the same place so also not great so what else can we do well we can use

this idea from sequencing and because I know the actual sequence of these DNA molecules from the Human Genome Project what I can do is I can isolate this complex I can put it in a single little tube I can sequence the DNA and get the exact string of letters and say okay well all of these molecules were interacting in the same space and I can do

the same for b c a and I can do this for thousands and thousands of different complexes but there's a problem I can't do a thousand experiments because that would take me many many years and I only have so much time here so we had to get creative and this is the solution we developed and this is a method called Sprite and it stands for split

pool recognition of interactions by tag extension doesn't matter what it stands for but I will walk you through How We Do It and so here we have the same nucleus with complexes a b c and d and we want to know that they are all interacting in the same space at the same time so the way that we do it is we freeze the cells with

a cross Linker called formaldehyde we then break up the complexes into smaller and smaller chunks then we pull everything into a single tube and then we split it 96 well 96 ways into individual Wells and so the chances that a molecule lands in any one of these Wells is 1 over 96. so we have right now we have complex A and B that ended up in

the same well complex C and D that ended up in the same well and therefore they share the same tag well this is good in that we can separate these complexes from these ones but right now we still think that A and B are together so that's not good enough so let's do it again so we can pool all the molecules in a single tube and

we can split again and once again there's a one out of 96 possibility that the molecules go into any of these Wells but the probability that any two complexes end up in the exact same well is now 1 over 96 squared so now A and B are in two separate well C and yarn two separate Wells all fine and dandy and now we've been able to

separate A and B but of course there's millions of molecules within the context of the nucleus so we're going to do this procedure not three times not four times but five and so after five rounds of barcoding we achieve 1 over 96 to the fifth power that's one in eight billion chances of two complexes being in the same exact well with the exact same barcodes it's

very very low and this is great because now we can use genome sequencing and we can identify these clusters so now we have the known barcode string and then we have the known genes that once we sequence we can align it to a reference genome and say oh these are hemoglobin genes and they work together and these were keratin genes and they were together and they

have a unique and distinct barcode from the ones that were in cluster a and so on for C and D so this is all hypothetical but what did we find what are what's the data set that we um uncovered well as you can tell if you didn't completely understand it that's fine you can ask me later it's obviously a very like elegant sophisticated way but what

the data we got was very elegantly simplistic in that the genome actually arranges in two distinct hubs and so here is like the blue Hub which is um so this is a spider plot showing different chromosomes so chromosome one all the way to 19 and then chromosome X and so here what we found is that there's regions of chromosomes 18 and 19 coming together chromosomes two

and four and five coming together but the red and the blue interaction hubs never interact they're always separate so what are they upon further inspection what we found is that this blue one represents the inactive Hub in that they are enriched for genes that are always turned off or for the most part are turned off versus these regions that are in the red correspond to what

we're calling the active Hub and these are typically enriched for genes that are turned on but what are some of the other molecular hubs that we found so this is just in the context of DNA but because I told you that DNA is copied into RNA and because we could extend our Sprite procedure to include RNA molecules we started to understand how RNA RNA hubs actually

engage with each other and so that's what I'm going to show you next and that we revealed not only these two DNA interactions but many other RNA hubs so this is a heat map that shows um orange are molecules that are interacting close with one another and blue means that they're not interacting very close with one another and so I'm just going to highlight a couple

examples one of them is the nucleolus so we actually pulled down the entire nucleolus and when we sequence the RNA we find that a lot of the rnas that are within the nucleolus correspond to locations that are very close to one another as you might expect and all these other different hubs that I'm not going to get into but suffice it to say that we found

these multi-way hubs of RNA and DNA locations throughout the nucleus and just to give you a picture of what this looks like we had a pretty good idea of specific molecules that were in the nucleolus or the nuclear speckle the kahal bodies but we didn't have the full picture of every single molecule that exists Within These compartments and because of this unbiased sequencing approach we were

able to discover dozens of unidentified interactions between RNA and DNA Within These different membraneless bodies so this was really exciting because now we have a list of RNA and DNA molecules that exist Within These different compartments and we can start to test hypotheses of what happens when we look at those locations or what happens when we perturb it and so I'm going to focus on one

example and this is the spicesomel Hub which I will show you corresponds to the nuclear speckle and that leads me to the second part of my talk which is on function so why do these Maps matter why is it that the genome has to be organized in this way and for that I'm going to give you a brief introduction into how information flows through the cell

in terms of genetic information and that is the concept of the central dogma molecular biology so this is the nucleus of a cell with all of our DNA it gets made into a copy of RNA this RNA that engages with a ribosome the thing that makes the protein and now you have a protein that's made from the message that's contained within messenger RNA the protein can

do all the things that it needs to do for you to function normally but there's one additional step that is critical in the production of mRNA and that is a process called mRNA splicing so before you have a happy Jolly mRNA that's going to produce a protein it exists as this long ribbon of nucleotides as a pre-messenger RNA so it has these exons which are the

meaty parts that we want and these introns that are kind of the junk that we want excise and removed and so in proper splicing you have the exons that are stitched together into a mature form and then this goes out into the cytoplasm and gets made into a functional protein but what happens when this goes wrong well when this goes wrong what you can have is

improper splicing where the intron didn't get cut out so now this is a problem for the cell and it executes different things so one way that it could react to this improper splicing is that it degrades the MRNA altogether and you don't get the protein that you need to make not good and so this is the basis of some diseases and then you can also have

a defective protein If the message actually gets translated and the ribosomes like I don't understand this but I'm still going to make a protein that happens too and so these have been associated with a whole host of different uh diseases including the ones you see here neurodegeneration cancer developmental disorders immune dysregulation they are often marked by mutations and spliceosomes splicing factors Etc therefore for this part

it's very important for us to ask the question how does the cell ensure rapid and efficient splicing because if the cell isn't able to do it can't make the proteins it needs to make and it'll die so this is not a New Concept this is something that has been thought about for several decades now since the discovery of splicing which was 35 40 years ago and

so I'm going to walk you through what people have gathered about what is the way that mRNA splicing is efficient and so what people have thought is okay so here we have a piece of DNA and then we have this enzyme called polymerase that finds this location on DNA and it starts synthesizing this long ribbon of nucleotides this pre-messenger RNA now where does this enzyme the

spliceosome come into being well there's this region on the polymerase called the c-terminal domain name not important but what is important is that it now acts as a landing pad for the spliceosome the enzyme that executes the splicing reaction so now you have the spliceosome and this ribbon of nucleotides directly in con like near each other such that the spliceosome can now deposit on the pre-messenger

RNA cut it up take out the intron and then execute the splicing reaction just fine so now you have a messenger RNA that's fine and ready to be made into protein and so in this nice way you've coupled the production of the RNA with the processing of RNA so it's like an assembly line while you're making it you're processing it and then the cell can do

what it needs to do and this led to some questions about well where are these spliceosomes where are these splicing factors in the nucleus aha while it turns out that a lot of these spliceosomes were nuclear Speckles so uh this is just a cell nucleus here in blue the red is the nucleolus this is the kaha body in white and then there's these 20 to 25

tiny teeny little dots uh known as nuclear Speckles and these actually um have been known for several decades to include splicing proteins and the same non-coding rnas that we found in our Sprite maps and so this was really interesting and what people thought when Speckles were first discovered is that because they contain all these splicing proteins they represent the site of splicing so they thought okay

this membraneless organelle represents the side of splicing because they have spliceosomes which is a fair assumption but instead what was found is that spliced some exit the speckle and they go to sites of active transcription the synthesis of RNA where they execute the splicing reaction in other words speckles are warehouses of inactive splicing factors and splicing doesn't happen at the speckle so that is a view

that's held until today but um oh and then so Speckles have long been thought about as RNA and protein but I just told you that we have this entire new atlas of genomic DNA and how it's organized and so this is the view that I showed you where we have these active Hub DNA regions and now I'm going to tell you what this active Hub is

and what we found is that it's actually the DNA that's organized around nuclear Speckles so let me walk you through what I mean by that let's take this long chromosome and so there's these two genes Gene one gene 2 and they're separated by a very long genomic distance 10 000 kilobases as a reference point just an average Gene length is like just 10 kilobases so this

is very very far apart they don't really see each other uh if we can humanize them a bit and so what we found is that if we look at these speckles this DNA actually folds in such a way that these two regions that were really far in linear distance actually three-dimensionally come around the same nuclear speckle so which are these active genes that contact nuclear Speckles

and for this I'm going to go through the next slide is highly technical so if you don't follow that's okay but it's for any of the biology aficionados for me to Showcase how we can use our Sprite technology and how we analyze it so that we can get to these conclusions and so the way that we visualize our Sprite data is by plotting a histogram of

what we call our specklehub contact frequency so this is how frequently does each genomic Region contact a nuclear speckle so this is chromosome 10 41 megabases that's 41 million bases of DNA and then these are two regions that are on that are zoomed in and so a higher value on this histogram means that it's speckled close a lower value means it's speckled far so as an

example this is a speckle closed region because it has a high Sparkle contact frequency and this is a speckle far region because it has a low special contact frequency well we can look at these same regions by Imaging and we can image by microscopy because we have methods to do this we can look at the speckle protein this is just one of the sparkle protein names

and then we can also add a fluorescence probe to the DNA sites that correspond to these two regions and what we find again this is a nucleus of a cell and then in white are the different proteins and then blue which is this tcf3 close to the speckle this Gene far from the speckle so in essence it's basically showing us what we're seeing in Sprite and

when we look across the entire genome we can plot on the y-axis the distance of each and every Gene from the nearest speckle and then on the y-axis we can plot our Sprite Sparkle contact frequency this is all to say that we can actually see that there's a very high anti-correlation where the closer you are to the speckle and Sprite distance the closer you are also

in this microscopy-based data the bottom line here is that Sprite is accurately measuring our distance to the nuclear speckle so that's really cool and so now the question is well I told you that these active genes are around the nuclear speckle the question now is is it just simply that a gene it being turned on does that mean that it goes to the speckle is it

all active genes there's a lot of active genes well the short answer is no not every active Gene goes to the speckle so what's the flavor of a gene that goes to a speckle well here I'm showing you again low contact frequency High contact frequency so this is the zoom in of equivalently length bins for for chromosome seven and then now I'm going to show you

um these two regions and how much they're transcribed or how much RNA they're producing and what we find is that okay here there is a gene that's on but it's not close to the speckle so then what's unique about this site that actually does make it close to the speckle it turns out if we look at how transcription is occurring there there is a lot of

transcription that's occurring across the entire block and so there's an increased density of transcription or increased number of genes that are turned on in the similar length scale and these genes are the ones that tend to be associated with nuclear speckles and this is just to show you that this is correlated very well genome-wide where we look at the density of transcription versus the speckle contact

frequency again correlated and these are not just cherry-picked examples so what this might look like in the context of this um of Speckles is that if we zoom in at a speckle so here is our speckle here so regions that have very little amount of RNA they still are being transcribed but they're just not going to the speckle but regions that have a lot of transcription

a lot of RNA tend to be more associated with the speckle so what I've showed you so far is that there are these actively transcribed genes and they're around the speckle and there's also active transcription that happens away from the speckle is just less all right so I started this part by telling you that speckles are these warehouses of inactive splicing factors but now I've also

told you that there's these increased number of genes that are at the nuclear speckle and so this gave us a really interesting idea of now we have a list of genes that actually contact the speckle so can we look at the genes that contact the speckle inner Sprite data and ask are the splicing of those genes in any way affected so does organization increase splicing efficiency

of those genes that's what we wanted to test next and to introduce that I want us to take a step back because I know I've shown a lot of data and I want to introduce about just chemical reactions and what they depend on so now we're going to just briefly illustrate okay so you have a reactant that goes to a product and there's an enzyme in

between that catalyzes this reaction and the other thing that reactions depend on are pH in temperature which are relatively constant in a Cell so that I'm going to remove from the equation so in the context of splicing our reactant is our pre-messenger RNA our product is our mature messenger RNA and our enzymes are the spliceosomes and so what I just showed you is that instead of

these genes being all throughout the nucleus there is an enrichment of what we're calling these so these are the reactions of pre-mrnas and they actually are localized at the speckle so that means that they have a high concentration of reactants at the speckle so now the next two questions are for us to actually complete this model we need to demonstrate that the spliceosomes are concentrated also

at these genes and that the product is also increased so that's what we're going to go next so this result might seem um a bit trivial but it was an important one which I can't get into the specific details but what we can do is we can ask what is the concentration of spliceosomes at these specific regions and so the way we do this is we

look in our Sprite data and I told you that we use this cross-linking agent called formaldehyde it creates these crosslinks between all these different spliceosomes as well as the pre-mrna and the DNA so we have this entire molecular Hub that we've captured and Frozen and so now I can measure how many spliceosomes are engaging with this pre-mrna so I can look at a speckle close and

a speckle far region and on the x-axis I can map basically the spliceosome concentration which I'm calling snrna density and what we find these are four different rnas that are found within the spliceosome and they are all enriched at the speckle and not enriched when there are four ad genes away from the speckle so what this looks like is I began by telling you that there's

active genes and these pre-mrna reactant molecules around the nuclear speckle and now I'm showing you that the splicing protein concentration is also highest near the speckle and it decreases as you get further and further away what this looks like at the level of an RNA is that if you have five RNA molecules far away from the speckle there's less spliceosomes available so you're going to have

what the hypothesis you'll have less splicing versus up here you have high concentration of spliceosomes same number of molecules but they're all engaging with spliceosomes so that's basically what I showed you here and so now we have a increased number of enzymes great so now this is a very interesting biochemical problem which is if we concentrate reactants and enzymes does this in fact enhance splicing the

way we did this is by isolating RNA molecules as they're associated with the speckle through a method called nascent RNA sequencing I'm not going to get too much into the nitty-gritty of how we did it other than to say that we have to do a very stringent biochemical purification method in order to strip away all other rnas the other 99 of rnas in the cell and

we just pick out the rnas as they're close to the speckle and so we then can do next Generation sequencing and sequence the RNA molecules and ask what is their splicing State as in how many of these molecules are spliced and how many of these molecules are in spliced as in they have introns unsplaced and this will give us a measure of how effectively this process

is happening of splicing and the data will look something like this where if you have this is the intron region of a gene and this is the Exon so a region that has low splicing efficiency would mean that there's a high number of introns versus a gene that has high splicing efficiency would have a low number of introns so pretty simple well how does the data

look now here I'm showing you the Sprite specklehub contact frequency this is just to show you that this is a far region and this is a close region because it has a high value so how does the splicing look here well when we look at this nascent RNA this region has a lot of introns meaning it's low percent spliced and in other words it's splicing that's

not happening very efficiently versus this Gene that is located within the speckle has very low number of introns reflected by a higher splicing efficiency and therefore proper splicing and again to show you that these are not cherry-picked examples when we look across speckle close and speckle far regions and on the x-axis we're plotting the percent splicing speckle closed regions on average have two-fold higher levels of

splicing than do specklefar regions and when we look across the Continuum of sparklehub contact frequency there is a nice correlation between specklehub contact frequency as well as splicing efficiency and this is all to show you that genes are exhibiting different splicing efficiencies based on Sparkle proximity great so what I showed you is high concentration of spliceosomes and reactant molecules at the speckle more engagement at pre-mrnas

the splicing reaction the details here do not matter but basically you're getting spliced mRNA at both locations but what that means for the cell is that it's low efficiency far away and high efficiency close to the speckle so indeed we're finding that there's increased amount of mRNA product when you get to the close to the speckle so we've kind of completed this entire picture of more

reactant more enzyme more product the rich get richer is kind of the motto here unfortunately well we wanted to take it one step further and that was we wanted to decouple genomic position and speckle proximity can we get cause and effect and the way that we wanted to address this is can I produce the RNA anywhere within the cell and drive it to a speckle to

drive it to get spliced that would demonstrate that speckles are actually explicitly coordinating splicing so how did we do this well the way that we did this is by recruiting different art like one specific RNA to the nuclear speckle so this is a speckle protein called srm1 and refuse to it a fluorescent tag M Cherry uh which fluoresces red and so we can visualize it by

fluorescence microscopy and then we also added within the same cells this mRNA reporter which contains a stem Loop structure called an MS2 and this all you have to know is that when you put them in the same cell this mCP protein actually binds with high Affinity to this MS2 so it's just a way for us to bring the RNA to the speckle and to show you

that this works when we do imaging for this RNA we show that the the speckle protein co-localizes with the RNA very well and this is an important check because many times in biology and anyone who's in science here will know that sometimes things don't work so you have to make sure that everything is looking and span which is why we have controls and the control here

efficiency so when we put all these molecules together in the cell only if this intron gets cut out spliced out well we have this green fluorescent protein be made so now we can measure green fluorescent protein as a proxy for splicing levels how does the data look well we didn't only make just the nuclear periphery and the speckle protein we actually made different classes of proteins

that the RNA will get recruited to and that also includes splicing proteins like these ones SRS F3 and 9. these localize in a diffuse pattern throughout the nucleus of a cell and there's another protein srsf1 which exists in a dual state where it's both compartmentalized and it's also diffusely scattered throughout the nucleus so it'll be important in a second and so we basically generated two versions

of this one where we can actually recruit the RNA to The Domain and one where the protein doesn't have the recruitment domain and then we can measure the change in gfp so basically the y-axis is the splicing efficiency when we recruit to all these proteins so just to orient you when we look at the nuclear periphery for all concentration regime of lbr we see that there's

no difference in splicing when we recruit to it not surprising because lbr is not involved in splicing but what happens when we recruit to these other splicing proteins but they're not at the speckle again there's not really any difference in splicing efficiency next when we recruit to this Buckle protein SRS F1 we find that there is a market increase in the levels of splicing and then

finally we recruit to srrm1 which again is localized exclusively within the speckle and when we recruit to srm1 we see a over 30-fold increase in the levels of splicing so that's pretty remarkable in that pre-mrna localization to a speckle is sufficient to increase splicing efficiency and this suggests that there is a causal rule in organization around the speckle and splicing occurring and that wraps up that

part of uh that talk and I just want to uh impress upon you that I began my talk by introducing all these different membrane-less bodies within the nucleus of a cell and although we've very comprehensively identified how the nuclear speckle functions and how the structure is important for its function we can now use this as a paradigm to understand whether the efficiency of the processes that

occur within these other membrane-less compartments is also the same Paradigm that we've discovered for the speckle and finally I'll just provide one short slide on what applications I'm thinking about based on these results and so I showed you this slide earlier on where if you have proper splicing you make functional protein if you have improper splicing you can have various negative outcomes and you might imagine

that there are instances where making more functional protein might be very advantageous to either um someone with not enough protein or even just in like vaccine development where you need to make copious amounts of protein and decreasing improper splicing would also be a very important thing to do so one idea would be to take any Gene that may be not being as well expressed and you

could recruit it to the nuclear speckle just so that you can increase the amount of protein and so there's quite a few diseases in which you have what's called haplo insufficiency where you're not making enough protein and by recruiting to the speckle we may be able to start to understand how we may be able to modulate splicing around Speckles as like a new therapy and that

completes all three of those parts and so I showed you how we developed Sprite to map DNA and RNA organization and we made some of the highest resolution maps of a 3D organization in the cell we then talked about how DNA organization around nuclear Speckles increases splicing efficiency and then finally just very briefly I just mentioned some implications how we can think about this for enhanced

protein synthesis lots of people to thank I just want to specifically mention a few people who have been instrumental for these projects of course my advisor Mitch Gutman he has been an incredible source of support and ideas he's just a very like awesome guy open door policy can always go to him for anything and it's been a pleasure to work with him Amy Chow Ben emmer

Olivia etland have helped a lot with the ms2mcp assay you saw and Sophie kinados for the first half of my talk it was her PhD project to um that she started with to generate this technology called Sprite and so a lot of the structural work was done by her and everyone else that's been on the project enough for the beautiful illustrations you saw in our talk

my funding sources and all of you for your attention thank you so much and do we have time for questions I have a quick question about the top which was we don't know how after 32 sub nuclear components some of your organelles you would get a function and so I wonder if some of your further Studies have added more clarity of that question for the other

bodies or something together is that that are any more details yeah so the question was uh previous like in the beginning of my talk I was talking about how you can kind of execute a lot of the reactions that are within these compartments inside of a test tube and that um here we demonstrated that organizing around these compartments is important for their functional efficiency and so

uh one of the things that I've thought about in the context of that is when we look at these older papers so for example for the nuclear speckle if you take the nuclear special components out and you put in a test tube you can still get splicing inside of a test tube but what people weren't looking at was the efficiency they were just looking on like

a um like an agarose gel like basically do it was a binary like do I splice or do I not splice it wasn't what percentage of splicing do I get and it was in a very synthetic system so I think a lot of those experiments were important for understanding the biochemistry of splicing but not really important for understanding um like what it means in the context

of like and situ in the cell Ellen know there are a number of messenger are supposed to be spliced altering where sometimes the preference for one form versus another is variable I was just curious if you looked at any of those loci to see bring them up closely to the cell type itself so um yeah so the question uh is that right now so I was

showing um that we were looking at a very specific type of splicing and that is intron retention where whether or not an intron is retained but there's other types of splicing reactions that can occur and one of them is alternative slicing so you can have three exons and one and three get splice one and two get splice and so they can do different combinations and so

the question I think is have we looked at whether or not that might be influenced by proximity to the nuclear speckle it is an amazing question that I uh for the um for our Sprite data because of our our resolution our read lengths we can't really answer a lot of the um questions about whether alternative splicing happens like close to the speckle versus not but we

can look at some of our nascent RNA sequencing data and we are looking right now to see um if there are differences in in alternative splicing yeah yeah Lambda efficiency yeah so that's an interesting question so like um uh Lando's question was that I showed that you can recruit to Speckles as like a clinical strategy in order to enhance protein production but can we do the

converse of moving it away from the speckle to decrease protein production when you don't want it um yeah so I think the my immediate gut reaction is that does make sense but one thing is that not all genes are located at the speckle so we would first have to identify that the gene that you're interested in um that is expressing at very high levels is indeed

at the speckle and if it is then you could see that as a viable Target for disrupting that interaction so um there are some genes as I mentioned that are far from the speckle that are still on and they're expressed uh and not all genes are out the speckle so that that would just be the one caveat but yeah it's a great idea and this is

consistent between cells and then the second question is um fan of your top suggested that recruitment of rnas to speculates could increase glycine efficiencies just wondering if there's alternative ways to increase yeah um so for the first um question um your advisor Michael ellowitz also asked me this question um which is a very good question um and uh are basically are all Speckles the same and

do they vary by cell type and the short answer is we don't really know um and we would love to like test that question um and so like Speckles of course are like demarcated by many different proteins and so right now our Sprite technology is able to measure DNA and RNA but not really great at measuring protein but there's other methods like seek fish that yodaitake

has developed where we can start to look at DNA RNA organization around different speckle proteins and different Cellular Systems and start to answer that question but that experiment hasn't been done yet and the second question you had was regarding other alternative mechanisms like increasing splicing protein production as a mechanism to increase splicing efficiency so there are diseases in which you have up regulation of um like

different splicing factors specifically in the context of cancer and you have rampant splicing and I don't know if there's a cause and effect here like I don't know if it's like a consequence of disease but I my my hunch is if you now increase the global concentration of spliceosomes you've now increased the level of spliceosomes at regions that are far from the speckle and you're going

to aberrantly splice things that shouldn't be spliced so you need to have a tightly controlled buffer system where you're controlling splicing where it should happen and when it should happen and not um and because biochemical reactions depend on the enzyme and substrate concentration by globally increasing you're going to increase everything not specifically so that would be a little scary chemical reaction that attracts like you mentioned

you know 20 megabases capital is there some specificity for Gene low science people and what what is the biochemistry of ten genes to yeah so um the the question was like what are the contents of a speckle like what determines whether two genes like end up at the same speckle so I didn't get a chance uh to show this data but uh what we as well

as uh yodai and like other folks have shown is that most speckle contacts between different cell types do not change so the same genes are in most contexts contacting in a um in a conserved way but the things that are really important are that genes that tend to be really over highly expressed in different cellular contexts so for example in a muscle cell you make a

lot of myod1 that is far away from the speckle in an embryonic stem cell but then it moves closer to a speckle in the context of a muscle cell so that's data that we have that we did with Barbara wolves lab and so basically what we're seeing is that it's transcription of the gene that's toggling a gene either to or from the speckle um but yeah

about like 80 percent of speckle contexts are conserved between cell types yes he was thank you [Music] it'll take me a bit to get there but [Music] why is oh yeah so uh the um yeah so the spider plot that I showed the circle with the two different hubs was the blue and the red and so there were some chromosomes that didn't have any contacts so

that's a very astute observation so again it was a little bit of a simplistic so when we when we um I'm going to answer this like um technically because I know that you you're familiar with it but basically what we did is we have this inter-chromosomal contact score where we know that these three different trios of chromosomes interact with one another and if they interact in

a trio which means a higher order more than two then we call it as an active Hub contact but not every chromosome has that and so uh what we then do for the specklehub contact frequency which is a separate metric we divide the entire genome into equivalently spaced bins and we ask how does this bin contact one of these regions that I showed you on the

spider plot and then that's how we get a distance of speckle so the regions that were like not um in the spider plot it's not that they're not near the speckle it's just that they don't participate in a like multi-way contact at least that we can detect by Sprite all right let us thank uh [Applause] before we conclude the session I would like to make an

announcement that uh there will be a third talk in the lecture series that's going to be happening over the summer in August uh the date uh is still to be decided so please watch out for for our email and thank you everyone for coming out today uh to to listen to the to the uh wonderful talk [Applause]