VIDEO: Dragomir Milovanovic on Condensate biology at the synapse

Dewpoint scientists hosted Dragomir Milovanovic, Principal Investigator – Molecular Neuroscience, German Center for Neurodegenerative Diseases in the next of our 2024 Kitchen Table Talk series. During his talk Dr Milovanovic discussed efforts to understand the cognizant biology of the synapse.

Dr Milovanovic published seminal work that helps us understand the molecular mechanisms that control local organization of biomolecules at the synapse and how this organization regulates neurotransmission. His work on how macromolecules and organs self-organize at the synapse using condensation has yielded important insights into how synaptic vesicles can self-organize into a liquid phase, how synaptic condensates act as a buffer for disordered proteins, such as alpha synuclein, and how condensates can actually harbor electric potential at the interface.

Click here to view the engaging discussion about condensate biology at the synapse below.

TRANSCRIPT

Diana Mitrea: It is a great pleasure to welcome everyone to our kitchen table talk number 37. I am Diana Mitrea, Head of Scientific and Corporate communications at Dewpoint Therapeutics, and I’ll be your host. Christiane Iserman, a senior scientist in the Biology team from Dewpoint, is joining us from Dresden, Germany, as your another co-host, Eric Martin Associate, principal computational scientist in the Data Science and Engineering team is joining us here in Boston. Thank you all for being here and thank you to Dragomir Milovanovic, today’s speaker for kindly accepting our invitation.

Diana Mitrea: Before we start with the awesome science that we’re going to hear today from Drago, I would like to start with a few housekeeping messages. This talk is a live event, which will be recorded. So feel free to turn on your camera. But keep your microphone on mute, and please be respectful – no on camera shenanigans. The event recording will be posted on condensates.com within a week or two, so you can refer to it at any time, we would prefer you keep your questions to the end. But Drago has agreed to answer any burning questions. If they arise in the middle of the talk.

Also feel free to type in your questions in the chat as they come to you, and at the end of the talk please wait for Chrissy, Eric, or I to call your name, we will ask you to unmute, and at that point you can ask your question if, for whatever reason you’re unable to speak. Let us know in the chat, and we’ll ask the question for you.

Chrissy Iserman:  I’m very happy to introduce Drago Milovanovic. Drago is originally from Serbia, and after his PhD at the MPI in Gotting and postdoc at Yale. He moved on to become a group leader at the German center for neurodegenerative diseases in Berlin. His work on how macromolecules and organs self-organize at the synapse using condensation has yielded important insights into how synaptic vesicles can self-organize into a liquid phase, how synaptic condensates act as a buffer for disordered proteins, such as alpha synuclein, and how condensates can actually harbor electric potential at the interface. I very much look forward to hearing more of this and more by Drago himself, and I’m sure we all have a fascinating talk ahead of us.

Dragomir Milovanovic: Thank you very much, Christiane and Diana, for the invitation. It is truly a great pleasure and honor, and, to be honest, I’m a little bit nervous to talking to what is the flagship industry, academic endeavor that you are doing at Dewpoint Therapeutics, and that we all, as a community benefit from. So today I would like to share with you our efforts to understand the cognizant biology of the synapse. Just to give you an overview: When we take neurons, primary hippocampal neurons from newborn pups and grow them in a culture, after about two weeks you get something which you’re seeing on the screen right now. So those are primary hippocampal neurons expressing synaptophysin, which is the bona fide synaptic vesicle protein.

So what you’re seeing here each of these dark spots in this network is basically one synapse and synapse are contact points between the neurons. And usually when we think about synapses, we always think about so-called synaptic vesicle cycle. So synaptic vesicles are the organelles, presumably the smallest organelles in mammalian cells, but only 50 nanometers in diameter that are filled with the neurotransmitters they need to dock at the very precise regions of the presynaptic plasma membrane, wait for the calcium signal to arrive, and then, within just a millisecond, undergo a complex event of membrane fusion to secrete those neurotransmitters.

Now, of course, you have compensatory steps where all of these excessive membrane has to be endocytosed and recycled back. And this is happening in every of these synaptic bhutans that you’re seeing. What is really striking is that these vesicles at the synaptic terminals are actually clustered, and they form this very beautiful presynaptic, three-dimensional domain. And what, in a certain way, is at the same time inspiring and puzzling, is that these vesicles that we are seeing on the EM are, in fact, surrounded by hundreds of thousands of proteins. So it’s a very dense environment. It’s a very densely packed environment. And this is not just a pretty picture. This is based on a complementary mass spectrometric, quantitative mass, spectrometry and super resolution stead imaging. So it is a very crowded environment. So for me, I’m trained basically as a chemist. And during my undergrads. And the question is really, why is this? How is this remaining soluble?

How are the vesicles able to undergo such a precise synaptic vesicle cycle within a spatial and temporal precision in such a densely crowded environment? And the mechanisms we try to understand this organization dynamics is basically try to use some of the principles of condensate biology and to look into this in details. So, already some years ago, when I was doing my postdoc in the Pietro de Camillis lab, we looked now at this catalog of proteins shown to be enriched the presynaptic terminals, and we recognize that there are some proteins, such as synapsin one or intersecting. So there are many proteins which have those general properties of proteins that often undergo phase separation. And this is in the case of synapsin. It has this very long proline, rich motif, very repetitive sequences and intersecting is just one example. There are many proteins with Sh 3 domains. So I purified these two components, put them together in physiological conditions and we saw this beautiful formation of these condensates. Now, as of course you know, and it’s one of the central themes in biology. And this is, how can we regulate the diversibility of these condensates? And this is really important for neurons, because you have these repeated cycles of neurotransmitter release.

And here, what was really exciting was to recognize that actually, within this proline-rich motif of synapsin there are certain target sites for certain kinases in particular, Kamk 2, which by its name tells you it’s a calcium dependent modelling kinase. And it has been known for decades that the camk 2 can interact with synapsin and that basically, it couples the influx of calcium, which will then activate the kinase with the synapsing phosphorylation and dispersion of vesicles from the nerve terminals.

So basically, we reconstituted this reaction. This was the time when actually Avina’s paper came out as well on the ATP as a hydrope. So we’ve now prepared everything, and we added, the ATPs, and the concentrations that we treat wouldn’t be a hydropes and the physiologically relevant concentrations. And we sort of you will see this really beautiful dispersion of these condensates, and that was a really important thing for us to see that it’s not only that you can form these condensates, but you can actually use the physiological stimulation to show that these structures are actually reversible.

So this was all of biochemistry and reconstitution biology, type of work. So the exciting thing was to see that this is conducive with what has been observed in living systems. So in this particular example, what you’re seeing is a giant lamprey synapse, and I’ll come to this in a few minutes. But what’s really important to observe here is that these particular animals in this particular synapse has actually a huge cluster of synaptic vesicles. So on average, mammalian synapses have around 400 synaptic vesiclesIn the prefrontal synaptic bhutans. In this case we talk about several 1,000 of synaptic vesicles clustered very tightly together. And it has been shown that when you inject the antibodies against synapsin, you basically disperse most of these vesicles, so by injecting the antibody, you do acute depletion of synapsin, and the only vesicles which remain are the vesicles which are docked to the Presynaptic plasma membrane. Through this active zone, which subsequently colleagues have shown to also be an example of a phase.

So when we published our in vitro work, these experiments have been reproduced now specifically with an antibody that targeted this IDR region of synapsin, and that was really striking because they could actually see a similar kind of dispersion, so disrupting the IDR of synapsin acutely in lamprey causes, results in a dispersion of synaptic vesicles from these condensates, more from these clusters at the time. So what’s emerging over the last, I would say 5, 6 years in the community is that many examples, both within neuronal soma, which is, of course, not just for neurons, but it’s also for many other cell types, but more excitingly within synapses, and both presynaptic and postsynaptic regions, most of these structures seem to represent some sort of seem to assemble through these principles of cognizant biology. Today I will focus quite a lot on the synaptic vesicle condensate. But again, the presynaptic active zone, the endocytic sites, the postsynaptic density, and even these adhesion zones between the pre and post synapse, seem to all assemble through these biomolecular condensates.

So the what is the question when you see such a picture that comes very, very clearly. The question that comes forward is, how are these bimolecular condensates interacting with the membrane-bound organelles? And again, in the context of a synapse where you have many synaptic vesicles, many clattering coated vesicles, endocytic intermediates, mitochondria er. So how these bimolecular condensates actually interact with the classical membrane bound organelles. And we thought of how to approach that problem and we coined the term dip in context to sort of explain the situation that we have in cell very often, and the sort of a question that we put to ourselves for understanding.

So that means that the initial contact between the membrane bound organelle, bimolecular, cognizant, will be particularly crucial in determining the fate of downstream processes. And there are in principle three potential opportunities or three potential scenarios that one can think of in the first scenario is that the interaction strand between the cognizant and the membrane is very weak. So basically, the cognizant will just not give a damn about the membrane. There will be no association.

Another very frequently observed scenario is where the condensates could actually interact with the surfaces of the membranes in this, with the principles of wetting. So where the adhesion and adsorption sort of collide together, they stabilize the surfaces of these, both you stabilize the surface of the membrane as well as the surface of the condensates. And this has been studied by many groups for decades. Now, trying to understand how are these principles, what drives those principles of vetting again the scenarios we see very often in the synapse as well, and I will also be discussing.

And then you have the scenario where the interaction between condensate the membrane bound organelle can be in this interaction. Strength can be very strong, leading to the full engulfment of the membrane bound organelles within these condensates. And again, synaptic vesicle cluster is one of prominent one prominent biological case where this is happening, another one. Another scenario, where this happens, is, for example, by the antibodies where you have mitochondria and other organisms really fully involved in the biomolecular condensates.

So what we try to understand is how the molecules and this is just a cartoon, so this is a sort of a hypothesis that we are trying to use to inspire our research. And this is how the organization of molecules and the interfaces of condensate. So this is in this case, it’s model synapsin. So how the synapsin orientation of the interfaces may interact with the protein lipid nanodomains that occur within these membrane bound organelles. And can these anchoring points be actually determining which of these downstream scenarios you have. So in the lab, we tried to look at interactions between condensates and membranes, and today I will share with you several, give you an overview of several stories we have to look at the function to characterize the interfaces and the specificity of these interfaces, how these condensates can act as reaction centers, and what this can mean for a disease.

So the first question I would like to tackle today is, can we scrutinize the face separation of synaptic vesicles in living neurons? So what I showed you in the beginning as an introduction was all done in vitro with recombinant purified proteins and liposomes. So basically, vesicle mimetics. And again, I will remind you about the situation we have here. This is a very simplistic video, but it should just emphasize to you that you have continuously these Ompa Sambotons. So along the entire axon that can be meters long, you have these clusters of synaptic vesicles, and this is really counterintuitive, because intuitively you would imagine the even distribution of vesicles. But what you’re having is not an even distribution. You basically have this local clustering of those of those components.

And what we did here first. We now purified those synaptic vesicles from so native synaptic vesicles from red brains, and we incubate them with synapsin. And again we could see this beautiful formation of these colloids that over time relax into a round structure minimizing the surface tension. And this was a very powerful system in hand, because it allowed us to do exciting techniques such as single molecule, tracking. Single molecule tracking is a time resolved microscopic technique, which actually allows you to look precisely at the trajectories of molecules, at the different regions, both in a dilute and in a dense phase and the interfaces. And what I really find striking is that, for example, when you do classical, spatial, super resolution, microscopy like stead or pound storm, you can have a fabulous resolution. You can see those individual dots. But what’s really exciting when the time result microscopy, like single molecule tracking or FCS, is that you can actually now get a trajectory and see how, although the molecules can be very much confined in a condensate, they can be highly mobile.

And you can also see here how the presence of synaptic vesicles increases the density of these trajectories. And these are basically even. This is a single molecule followed over the course of even five seconds. So what we did now in the living system, we now basically overexpress two proteins. So halosynapsin one, which for us seems to be a driver of this content information, and synaptophysin. So again, this bona fide synaptic vesicle protein. And this is just one example of how those single molecule tracking movies look like. And that allows us now to look at these, to record these trajectories.

From those trajectories you can deduce different types of motility. For example, you can see whether it’s a Brownian motion confined motion, as you would expect in a condensate situation, or a directed motion, which basically would be a trafficking of organelles along the microtubules.

And what was really cool for us to see was that if you now look at the synapse into synaptic vesicle motion, you basically see that those vesicles are beautifully combined within these presynaptic terminals. What’s similarly exciting was to see that synapsin actually has two different types of trajectories. So just to give you a comparison if you would do frop and you would frop such a region, you would get an average ensemble measurement of a diffusion property, but with a single molecule tracking, we were able to dissect those confined motion of synapsing that you have precisely where synaptic vesicles are in the synaptic bhutans and see them distinctively from the actually directed motion of synapsin that you have between the neighboring terminals, because obviously these molecules communicate between these terminals. So of course, the question arises. Now, this is all in a wild type context in a wild type, neurons. But what would happen now could we genetically corroborate this data? And to do that we capitalize on a mouse model where synapses have been deleted, so synapses are called, are encoded by three genes. So this is why the synapse in triple lockout animal. And here what you can see the trajectories of synaptophysin in those synaptic vesicles. In those animals you can see that you completely lose the accumulation of vesicles inside these bhutans.

And, beautifully enough, when you do the rescues, of course, with the full length, you see a nice rescue, but also with just this Idr region. By the way, this proline reach motif, this is also at the same time intrinsically disordered region, which is highly basic. So just looking at this just overexpressing, this Idr region was sufficient to rescue also the clustering of, or confinement of, synaptic vesicles within these presynaptic terminals.

Now, of course, you may wonder the evolution and modern nature to come up with this evolutionary preserved feature of synaptic vesicle proteins just for diffusion, right? The question is is this any good for the neuronal function. And to do that, we went to a functional assay that relies on a Ph sensitive fluorophore, so fluorine. And in this particular case we tagged the fluorine probe to the luminal domain of the synaptic vesicle protein. So synaptic vesicles are loaded with neurotransmitters. And they’re very very acidic. So basically, what happens with this fluorophores, it’s quenched inside the vesicles. When you now stimulate your neurons, either electrically or chemically, you can now cause the secretion of these synaptic vesicles. So the exocytosis, which will increase the Ph cause the dequenching  and then you will see the increase in signal.

And this is one of those videos. So these are live neurons perfused with a high potassium chloride to trigger the release. And you will observe here the release, and at the same time in synapsin channel. Because phosphorylation occurs, you will observe it. It’s happening right now. You will observe the dispersion of synapses from these condensates. And when we now, of course, quantify this over many different synapses, we were able to really see that in the synapse in triple local animals we have the disruption in the secretion in the neurotransmitter release, which can be rescued not only by the full end protein, but also with just overexpressing. This intrinsically disordered region. So we really see that the ability of vesicles to form these condensates at the nerve terminals is really essential for neurotransmitter release.

So this is, I hope, that I could show you convincingly that the synaptic vesicles, not only in vitro but also in the neurons, can truly form these biomolecular condensates. But now the question is, there are many other proteins, and can these biomolecular condensates at the synapse be reaction centers for organizing some of these other molecules. And in particular, we focused on the acting cytoskeleton because actin cytoskeleton is important to mediate the membrane trafficking during this synaptic vesicle cycle.

So obviously, we have definitely not been the 1st one to think about this. So the role of the cytoskeleton, the presynapse, has been inspiring the neuroscience community for decades, and there is a lot of beautiful work showing or dealing with that problem. And here just took an Em image from 1989, which shows that. And this is the this very exciting active filaments within this presynaptic terminals. And what’s also very much cool to know is that when you do this fancy, quantitative super resolution, imaging, people are really visualizing many different morphologies of actin at the nerve terminal, so they can form these acting tracks. There are recent reports that actin can form corrals. So you have this variety of morphology and anatomy of actin cytoskeleton at this very, very localized region.

So in the very early beautiful biochemistry work, Paul Grignard and his team have shown, or had tried to reconstitute and understand the role and the interplay between acting and synapsing. So what they did, they isolated the acting from bovine, and then sort of reconstituted it. Put it under em, and you can see this beautiful acting network. And then they spiked synapsin to this acting filaments, and they saw this fantastic bundling of acting that was phosphorylation dependent, because if they would add phosphosynapsin, they would not see that bundling effect and the take home from series of work that was published in in several years in the late eighties and early nineties, sort of put forward this idea that the synapsin is a cross-linker, that bundles acting filaments.

So reading this is a great biochemistry done beautifully well, but the question that, or simply the absence of some information led to a misleading conclusions. So, first of all, in this particular reconstitution, they were spiking synapsin to actin. And that’s very intuitive. If you think what’s the most? What’s more abundant in a cell, and you don’t know the numbers you would think. Yeah, actin is probably super highly abundant, and you probably have just a few synapsin molecules. So actually, this is not the case. So synapsin is more abundant at the nerve terminals than actin. So synapsin is present in around 120 micromolar local concentration at these presynaptic terminals.

And then, of course, as I showed you in the last few minutes now, we have quite convincing data that synapsin can form these condensates. So we wanted to revisit this idea of what’s the role of synapsin and actin, knowing the more precise concentrations and more precise conditions.

So what we did here and this is phenomenal work from Akshita and Linda, two students in the lab. They reconstituted the synapsin condensates, together with the actin. They saw that the synapsin condensates can actually enrich actin, and after some time you can see the beautiful appearance of these. So you can see that the actin starts polymerizing these beautiful aster structures coming out of these condensates. And that was really striking. Again, this is in the absence, of course, Atp is present, and I think she’s necessary for actin to be functional. But there is no nucleation point at this. There’s no nucleation factor. There’s no Arp 2, 3. So this is just synapsin condensates and actin. And if you do turf imaging, which was what Christian has done here, you can really see that some of these are really having this. This is just the acting channel that I’m showing you. You can see these actin enrichments at the periphery.

So actually synapsing condiments in the absence of any nucleation factors are fully sufficient to trigger the actin polymerization. And this is just one of these 3D reconstructions of those in vitro generated acting polymers. And what was really important for us now was to recognize that the polymerized actin, first of all this aster shape which is really striking and phenomenal. And then these asters can actually be connected with the very, very long distances. So the acting fibers can spend quite, quite some ranges of space with tens of micrometers that then led us to see okay, or to examine can those hubs of synapsin condensates and actin also, then sequester synaptic vesicles. And this was indeed the case. So here you have synapsing condensates and after half an hour of incubation with the actin. You can again see these beautiful actin esters, which are zoned here, and if you now add the natively isolated synaptic vesicles from urine brain, which are doped with the lipid dye, with the FM dye so you can see them. You can see that the vesicles are beautifully enriched in these condensates, and they actually also interact with these actin fibers. And this is exactly what has been seen that the vesicles can be on these actin fibers already decades ago.

So the question that clearly comes when you do these assays is that this is all in vitro. Could this make any sense in a living synapse? And can we really see such a complex acting architecture also ourselves in the in vivo system And to do that, we teamed up with my dear colleague, Jennifer Morgan, and over the course of three years I have a summer lab in Woods Hall. I’m a Whitman fellow, and a lot of my lab members actually, every year join me in Woods Hall, where, in this particular case, we work with our oldest vertebrae predecessor. So what you’re seeing here is a lamprey. It’s maybe not the most particularly beautiful organism, but it’s a really cool one. It still lives in the sand. It’s very easy to maintain, at least at the larva stage.

And what’s valuable for us by a chemist is that it has this super giant reticular spinal axon. So you can basically do the dissection of the spinal cord, and you can actually inject into these axons, which is really striking. So you can purify your proteins of interest. You can prepare your liposomes, and you can now inject those directly and do the acute imaging.

So I will now come back to the image I showed you, just to emphasize that, in fact, the function of synapsin has been discovered in a lamprey. So this image that I showed you from 1995, now almost 30 years ago. So the role of synapses has been defined through these lamprey experiments. Of course, with the development of mouse genetics, we are all now moving towards stringent animals. But just the value of these of these non-standard model system is remarkable.

So in our experiments, what we did, we now injected the lamprey with phalloidin-488. And what you can see here, this is the injection imaging in a ringer solution immediately after injection. So this is the still living axon, and you can see that the phalloidin stains is beautiful. So you can see this beautiful tori. And then here you’ll have now the zoom in again. This is the Coinja in this particular case, on the right side it’s a co-injection of phalloidin FM dye. So this is the live axon, imaging, if you want to say it like that, acutely in a lamprey and we were able to again see this beautiful acting tori around the cluster of synaptic vesicles, because the FM. Dye will stain all the membranes. You’ll see a lot of. So this will probably be some of the other organelles, perhaps mitochondria. But this very bright signal comes because in a cluster of vesicles you have a lot of lipids, so the FM dye gives you a very bright signal.

So actin forms those rings around lipid vesicles reminiscent of our in vitro reconstitutions also in the living synapse. And here what was really these are now the fixed samples. So we can now actually look in more detail with the confocal or super resolution microscopes, you can choose different setups, and you can really try to dissect. Now the morphologies that you are seeing in vivos and not just in vitro. And here I’m just showing you multiple staining. So in this particular one we have the synapsin, the Sv. 2. Another synaptic vesicle, bona fide, synaptic vesicle, protein, phalloidine.

And you can see this beautiful actin and wrapping, and if you try, some of the convolution software is available, you can see. Also, I can start seeing a little bit of these fibers coming out. So it’s really exciting, because all of the complex morphology we could see in vitro. We were also able to at least the main properties of this actin accumulation tori formation we were able to see in a living synapse.

And what’s another cool thing is that now we went back to our transgenic animals, and here we were able to do a more systematic analysis. So we were able to now, together with Yaco Branch and Helga Everest, we were able to look at those actin enrichments in the presynapse. And here what you’re seeing, and this is the 2 color Stat. So you see these actin rings that Xiaowei Zhang described now about a decade ago, and those actin rings are between the synaptic buttons, and you can see wherever you have the presynaptic buttons. So wherever you have a vesicle signal there, you don’t have the rings, but you actually have the accumulation of the actin. And this is fully disrupted in animals that lack synapses because there are no synaptic vesicle clusters. So there is no local actin accumulation, although the actin rings actually remain fully unaltered. And we were able to basically conclude now not only upon the acute injections in lamprey, but also in this mouse model, that the synapsin-driven s 3 condensates, or the absence of this synapsin-driven synaptic vesicle condensates, also leads to the disruption of the local accumulation of actin at the presynaptic terminal. These are just the quantifications to show you. So the actin rings remain unaltered, but the accumulation of actin within synaptic bhutans drops in triple knockout animals, and can be rescued with the full length, and the Idr protein.

So these condensates can be reaction centers. But at the same time the question is, now, once you have those condensates, you can have recruitment of many molecules. You also generate new properties inside the cytosol, and you generate those new surfaces, these new interfaces. And the question is, how can we think now about these interfaces between condensates, between neighboring condensates, or between condensates and the surrounding membrane-bound organelles?

And here what inspired our work quite a lot is the formation of those preformed fibrils. So what Christian did here, he constituted preform fibril of Alpha-synuclein. This is known as a pathological hallmark in many of the synuclinopathies, many of the diseases, and also very often use experimental tool because you generate these preformed fibrils, and then you can add them. You can inject them into animals to sort of trigger something like a Parkinson’s disease like phenotype, and by because Alpha synapse in one interacts with Alpha-synuclein by adding the synapsin one condensates. We’re able to see those equidistant formation of these sort of hotspots or sort of beads and a string of beads, type of behavior. And, in fact, this is one example where recreation captured the landing of one of those condensates to the interface of the Preform Fibril. And as we were doing these experiments. A paper came out from Howard Stone Petrie Lab. And they basically focused on the other type of interface in a cell. In this case the microtubules, and the Tpx. 2, which is a microtubule branching protein, and they really showed that these local hotspots along the microtubules can actually act as a sites for indicating where the microtubule branching will occur. So that really then stimulated us to think systematically. How can we look at interaction between synapsing, condensate the interface of synapsing condensates and surrounding structures.

And to do that we turned on to a model system using giant unilameter vesicles. So this system is based on the preparation of the membrane bilayers, which are tens of micrometers in diameter. So there really can be almost like a cell size structures. And for the condensate, if you think about it. There is no curvature in this in this. It’s almost like working with the Planar layer.

And this is really cool, because now you, this allows you to change the interface, because you can decide what lipid composition you will use in those reconstitutions. And in this particular case, what you’re seeing are the giant vesicles made with the neutral lipids, and we incubate the synapsin condensates with these, and you can see this beautiful wetting of synapsin condensates how they’re sort of landing in the interface and wetting the interface of these neutrally charged surfaces.

Now, if we switch the charge, so, instead of the neutrally charged surface you now make a negatively charged surface by introducing the negatively charged phospholipid. You see a fundamentally different behavior where now you’re instead of this individual condensate being in the surface of these giant unilamar vesicles. Now you’re seeing this beautifully formed complete vetting almost like a sunny egg, sunny side up, egg type of structure vetting the interface, and you can quantify. I mean you, can we quantify the contact index. But you can also do the contact angles different ways to quantify this. But also I think it’s very much visually clear.

So we could really see that the changing the nature of the interface, you can change the extent of condensate to membrane interaction. And the question was, here we see this charge-dependent behavior. But how can we quantify it? And we spend a lot of time thinking about this. And here I would really like to shout out to my very, very dear friend and colleague, Alexander Matkovich, who is a physicist, and who works in Austria and in his work at the University of Leban they’re characterizing different kinds of, it’s a coal industry center in Austria, and they’re characterizing different kind of minerals and inorganic compounds. And we were discussing, can we use some of these sensors that they have to quantify the biological systems? And after quite some thinking how to do this in their physiologically relevant conditions, right? Not to be in vacuum at 0 Kelvins. We decided to try out the graphenes based sensors. So graphene is a very unique material. It’s a very thin one atomic layer area. And actually, this is a photograph of the actual sensor that Alex made. And what’s very great with these transistors, with these graphene based transistors is that they are ambivalent to electrons, so they neither have the affinity nor repulsion from electrons which basically allows you to be. You can put it in a circuit, you can apply a current and then you can measure the changes which are happening when you put a different solution to those interfaces.

And in this particular experiment, what we are doing, basically we are adding now

a solution containing our synapse in proteins. This is just a synapse in a normal buffer, 150 millimolar sodium chloride, so billions of electrons, endless presence of electrons in a solution. And what you’re seeing here is that this graphene sensor is measuring the actual electron accumulation of the interface.

And you see this initial bump within 1020 seconds, and then a saturation of the surface. And for all of us who do cell biology, this is a very common phenomenon, because this is what we do very often record. The cover slips right. You add the pll or pdl, because you want to change the surface of the cover. So you want to make it, for example, hydrophilic.

But now, if, instead of the single state solution, we now add a solution that contains condensates of synapsin. What you’re seeing is this continuous accumulation of electrons. This is over the course of about 2 min. So basically, what’s happening is suddenly this condensate seemed to start generating this interfacial potential. And this is how we visualize this. Of course this is an animation, but the idea here was to this is the alpha fold, different structures of synapsing from the alpha fold, and the idea is that this polybasic tail can actually orient itself at the interface of the condensate and many labs, for example, rockets lab has shown theoretically, this can happen with many condensates, that you can have this preferential orientation of molecules which can then also, because in this case it’s a positively charged IDR can actually also drag the electrons and generate

relevant and measurable potential differences. And I think this is really striking, because it shows us that at least for synapsin condensates it seems to be emerging in the field as a general property that the condensates can harbor the electric potential of their interfaces, and in a certain way can act as mesoscale capacitors inside the cell that can suddenly store that can store charge. And I think this is a very important concept to think when we think about these interfaces. And what other reason is that these differential interfaces can play in the subcell organization of the synapse and beyond is that they can drive the specificity

So inspired by these observations, we wanted to now understand what drives the specificity of in our particular case synapsin condensates as a model system. So here I just picture for you several key proteins that are associated that are integral to synaptic vesicles, or associated to synaptic vesicles. And many of these proteins have very distinct physical chemical features To understand this, and some of these, as you know, from the community. This has been shown by many labs in many different biological contexts. These chemical properties can be oftentimes complementary and helping and accelerating the formation of condensates. So we teamed up with Rohit and a very talented senior scientist, Kirsten Raf and his team to sort of systematically analyze the properties of these intrinsically disordered regions within synaptic or of synaptic proteins. Synaptic vesicle associated proteins, and indeed, many of these proteins have very distinct properties. And in our particular case we took one example to see, to validate whether this is truly happening in a living system. So we took synapsin one and we co-express synapsin one with synaptophysin. This is in hex cells, so just cell in a test tube. And you’re basically seeing these beautiful condensates appearing. So now, what’s interesting with synapsin, if you look at the chart, is that it has the two very unique properties. It has a very high abundance of proline, polar residue blockiness within its idr, and it has a lot of Arginines.

So in the first mutant suggested by Kirsten secretion clone, now they scramble this IDR, so no deletion. So the full protein is there. All the amino acids are there, and by scrambling the IDR by ruining this polar proline blockiness, you have this very exciting gain of function property where now, suddenly, these new constructs basically become a cytoskeleton binding construct. They’re not anymore forming condensates.

And what for me was really striking. When we did this arginine to lysine initially, when we were discussing over the Zoom, I thought like, but why don’t we do just arginine to alanine, and they’re like, no, no, do arginine to lysine. Of course you know from work, and the TDP 43 and many other fields that these arginine have very different chemical properties. Let’s try to see whether this will have an effect. And indeed, mutating arginine to lysine fully abolish the ability of synapsin to form condensates.

And this is all in cells. We were able to validate this in a very, very systematic fashion, using the microfluidic devices by Tomas Nova’s lab, and then really see that there is a different boundary, different phase separation boundaries between wild type and these different mutants, and basically that this change of arginine to lysine, and this ruining of this partial pi property of the arginine side chain really disrupts fully the formation of condensates. And what was really exciting is that in Rohit’s lab Min and Christian were able to look at the clustering of these proteins in not in a phase separating conditions, but in those subsaturating conditions, and really see that the cluster size and density of the synapsing wild type sequence, or the scramble sequence, is fundamentally different than the subclustering of synapses which have the arginines mutated to lysines. So I think this very nicely showed that indeed there is a presence of these different molecular grammars, so to speak, within these synapsing condensates or synaptic condensates in general. And now we are full force, interested and excited about really understanding how these different complementary grammars between different proteins come together, and a lot of work is currently put into that into that direction.

So when we saw this one thing which was really important for us was, okay so we have these different constructs. But ultimately, under given conditions, you can push all of these proteins to form condensates. So we wanted to seem, can we decouple the grammar that would drive the condensate formation from the properties that those condensates have once they are formed. Once you choose a condition that all proteins were formed condensates. And the particular unit of property we wanted to analyze was pH. Why pH? Because you need a pH gradient in synaptic vesicles and in the condensate to accelerate and to help the loading of the neurotransmitters.

So here, what Christian did he basically capitalized on the C-snarf ratio metric dies and basically incubated the synapsing condensates with the snarf dye which you can excite and then you can record the different wavelengths. And that allows you, for example, before you have your sample with condensate, you can do this in a solution of a very precise pH, and then you can generate those standard calibration curve which now, if you now from the standard calibration you compare now the ratio of the images that you get once you have the once you form the condensate, you can get the relationship between the ratio of the fluorescence, intensity, and the actual pH. And what was really important and striking was that once you generate these condensates, you actually have a very consistent acidic environment, where basically for 0 point 3 pH units for about 40 nanomolar of the protons more inside the condensate inside the dense phase versus a dilute phase, and I think this really shows that once you can form those condensates, whatever the conditions. And of course the evolution in different species will have different properties within these sequences. Once you form these condensate, you can have some of these features and properties nicely so nicely preserved

And just at the end, why, all of this is important for biology of the disease. Just one quick teaser slide to put forward that many, many examples of formation of aberrant inclusions that are classically perceived as a protein only inclusions actually contain membrane bound organelles. And you can imagine if I go back to this idea of dipping context, right? So you can imagine there is an initial interaction between condensates and membranes where some membranes will be selectively included. They’ll be wetting the condensates or accumulating at the interface.

There’s some other membrane-bound organelles may be just passing by, and this can be disrupted in aging and disease. And what I’m showing you here is a human postmortem tissue showing the Alpha-synuclein inclusion, and you can really appreciate this accumulation of, in this case, the aberrant accumulation of mitochondria, the interfaces, and the lysosomes inside these structures. And in the lab, we have a massive effort. These type of inclusions occur in different disorders. It includes in Huntington’s and Alzheimer’s, but also in not only neurodegenerative, they occur also in diabetes and in cancer. And in the lab, we put a lot of effort trying to develop the system, to topically generate the lewy, body-like structures. And we were really able to see that actually, when we generate those aberrant Alpha-synuclein conclusions that we start seeing the accumulation of the mitochondria, this aberrant mitochondria. But I’m happy to discuss this during the Q. and A.

And with this I would like to conclude, basically today, what I showed you is that an example of synapsin and synaptic vesicle condensate. You can generate the 3D territory inside the cell, inside the cytosol, which can be very distinct where the molecules can be accumulated, yet very mobile, very dynamic to allow for the functional in this case, for the functional release of neurotransmitters. Synapsin and synaptic vesicle condensates are also reaction centers for sequestering soluble proteins, such as actin and allowing for the local acting organization of the synapse.

Synapsing condensates, we see harbor electric potential in their interfaces, and you come with this oxymoron right? That membraneless organelles, as these structures are oftentimes called, can have a membrane potential, because in the neuroscience community, every time you talk about the electric potential, you think immediately about the membrane potential. So actually, indeed, it seems that we can have the generation of those membrane potentials or electric potential, also at the interfaces of the condensates. And condensation of synaptic vesicles is driven by this. It’s not a random stochastic event. It’s driven by sequencing code, molecular grammars of the key synaptic proteins. I showed you today the case of synapsin and now we are examining the contribution of other proteins. And of course, there are many interesting questions ahead of us. We focus here on a few proteins. But again, this picture, from beginning, where you have hundreds of thousands of different molecules. So what determines the specificity of these synaptic condensates? How is the coupling between synapsin condensates and synaptic vesicle organization achieved in context of the neuronal activity. And of course, this is what we all and I’m sure, also at Dewpoint, you want to develop the therapeutics and tools that will allow us to functionally manipulate these condensates both in health and especially in disease.

And with this I would like to say a huge thank you to my lab members. I just took some photos from winter times and summer times, like many collaborators I mentioned here and along to talk the collaborators for the topics I covered today. Of course, we have several other topics which I didn’t cover this particular occasion, but I’m happy to talk about offline or in the future. And with this I would like to thank you for your attention. I’m very much looking forward to the questions and discussion. Thank you.

Diana Mitrea: Thank you so much. Drago. Chrissy, do you want to take the questions from chat?

Chrissy Iserman: So thanks so much also from our side here. That was a fascinating talk, and we already have one question here from the chat which is from Cardarelli. Do you want to ask the question yourself, Ross?

Ross Cardarelli: Sure. Hi, Dragomir! Is there any role of physiologically relevant metabotropic receptor signaling in dispersing or regulating the formation. I’m thinking of, like dopamine. D 2 autoreceptors or things like that, and how they modulate, transmitter, release.

Dragomir Milovanovic: Yeah, absolutely. That’s a great question. So what’s happening was that I focus here on just one kinase for the purpose of our in vitro reconstitutions is the Camp K, 2. But actually, synapsin has been isolated from cow brain in 1970s and named the initial name was Phosphoprotein one, because it’s the most highly abundant phosphoprotein. So it’s a hub for many kinases, including some of the map, kinase and terkinases, which are indeed stimulated by the metabotropic receptors, and of course, the strengthening between a pre and post synapse. And you, of course, have this long-term potentiation, long-term depression which will favor some of those bhutans while deforing the others, and we have a full topic in the lab. So Gerard is doing his PhD. Trying to understand and dissect the contribution of these different pathways so different kind of because you ultimately what you generate, you generate different forms of phosphosynapsin, because you don’t phosphorylate just within this Idr, but you can phosphorylate with tyhere is a dimerization region. There is a vesicle binding region. And yeah, we start, we are starting to decouple those hierarchy of these different effects. So the short answer is, yes, there is a definite effect of a different type of receptor activation pathways.

Ross Cardarelli: Awesome. Thank you.

Chrissy Iserman:: So are there any other questions from Centrosome or from the Kitchen Table?

Diana Mitrea: We have Avi, who has a question.

Avinash Patel: Yeah. Thank you Drago, fantastic talk. Right? So now there is, you know, this growing plethora of many proteins. That kind of form condensates or undergo, you know, phase separation at the membranes. Right? So what basically makes synapsin different. So why doesn’t a different protein just do the same job as synapsin does for neurons, because, you know, there are junctional proteins and all those things. So what makes synapsin so specific for his job and nothing else?

Dragomir Milovanovic: So what we’re seeing, I mean, there are multiple things which put synapsin into a very unique position. And one thing is that it has this very specific synaptic vesicle binding interface. So the n-terminal part of synapsin, which is also a target of a phosphokinase, a pka target binds in a phosphorylation dependent manner to the synaptic vesicles, the negatively charged phospholipids. Then you have different regions of synapsin that, for example, recognize a highly curved membrane. So this is the there is a particular region between the main B and C. This out motif. Then you have this intrinsically disordered region of synapsing which can engage in the multivial interactions with other synaptic vesicle proteins. So it is, in a way, a combinatorial effect of these multiple properties that synapsin has that puts it in a unique position to bind synaptic vesicles with the 50 nanometer diameter are the smallest organelles, meaning the highest curvature. They have a negatively charged phosphatidylserine, just as the plasma membrane has, but these are small and curved and synaptic vesicles oftentimes contain transmit domain proteins with the cytosolic tails, which are disordered. So all of these things, individually, I think, contribute to specificity. So our data also, so far with this molecular grammar project starts really suggesting that. But we are currently quite invested in understanding the contribution of lipids, the contribution of these multivalent interactions. And I hope in a few years we’ll have a more systematic matrix. But so far all the data indicate that there is a combinatorial effect of recognizing negatively charged membranes, curved membranes having the IDR that can interact with the resident IDR, and that all of this is necessary because there are many other membranes. There’s a mitochondria. There’s er, there is plasma membrane, but those members will have different transform domain proteins. They will have different lipid interfaces. They will have different charge density and packing density.

Avinash Patel: Okay. Thank you.

Dragomir Milovanovic: Thank you.

Chrissy Iserman: Thanks very much. Any more questions? Otherwise I have one. So I was wondering. I find it really cool. You’re finding about the electric potential of the synapsin condensate. So I was wondering if you see any potential in using this knowledge for possibly drug targeting, for instance, or also to develop something out of it right? But to develop a mini battery or something like this, so have you had any thoughts in this regard.

Dragomir Milovanovic: Yes. So what we’re doing now we’re trying to, we continue working with Alex, and we’re trying to develop a device that will allow us to measure the electric potential and do sort of a simple microscopy at the same time, like a combined potential measurement with microscopy measurements. And we think that will help us, because you can now think about many ideas of adding things that will act as potential surfactants, change the surfaces of these condensates. But the question is, what is going to be your quantitative readout? So we think that we still have to work on a methodology to be able to visualize, not only to measure the potential, but to visualize what’s happening in the interfaces before we start looking at the potential molecules like surfactants that you could add, and you could then play around and tweak and change these potentials.

Chrissy Iserman: Very cool. I look forward to those publications. Any other questions? I cannot see the kitchen table. Is there anything there in Boston?

Diana Mitrea: I have a question. That was a brilliant talk. Thank you so much. You mentioned that there’s three different synapsins in the cell. Can you comment a little bit about the conservation of the IDR?

Dragomir Milovanovic: Yeah, that is such a relevant question, because there are two aspects to that question. The first aspect is that the consideration of let’s say, just synapse one IDR among the species, and while you may not see when you do, the classical cluster W, or cluster omega sequence, alignment, while the precise sequences or residues may not be conserved. The properties are remarkably well conserved, and this is also in this pre-print that we did, together with Rohit and Kirsten really could see, even as far back as lamprey, which is presumably the first organism, the oldest living vertebrate the property of synapse IDR, synapsin one IDR seems to be conserved, even though the sequence is completely different. So I think this is incredibly exciting, that evolutionary preservation of physical chemical properties.

The second aspect of that question is much more nuanced, and I guess much more important also for the thinking about the drug and therapies. And this is the different neuronal types and different synapse types may express different synapse in isoforms. So synapses one and 2 both have the IDRs, and both can readily undergo phase separation. We don’t see this with synapsin 3 and the synaps in one and 2 seem to be expressed in most neurons. But there are some neuronal types where one synapse is expressed preferentially over another one, and that can also then contribute to this specific vulnerability of some of the synapses, or like inhibitory versus excitatory, or some of the neuronal types. I would just say we are on it. But it’s as you can imagine. This is very complex because you would need to really specifically dissect that question in the given neuron type to see the individual contributions.

Diana Mitrea: Thank you. That’s awesome. And it also makes you wonder how you know the local pH and the emergent properties can be tuned that way. Are you looking at that?

Dragomir Milovanovic: Absolutely so. This is one of the future directions we would like to explore, and this is whether the rate of acidification, the rate of the neurotransmitter loading will be different different synapses. Because if you think about loading the dopamine into a synapse versus loading glutamate, it may require quite different microenvironment. And yeah, this is something very, very exciting.

Similar question. I mean the problem there is again similar to what Christiane was mentioning earlier with the potential measurements is that one has to really develop the methodologies to be able to measure this at the individual vesicle level, because the question is very clear that you’re asking. Right? So would, let’s say, dopaminergic synapse that has dopamine as a neurotransmitter, and the synapse that has the glutamate as a neurotransmitter, will the kinetics or loading be different? And will the condensate change this in the same or in a differential manner? But to actually quantitatively measure that we have to work on the methods for those assays.

Chrissy Iserman: Very cool, and I’m realizing we’re running out of time. So it’s already 4 o’clock here. And whatever in Boston the time difference? It’s too complicated. So yeah, so I think that fit perfectly with the amount of questions. And yeah, I want to thank you again. That was really cool, very exciting.

Dragomir Milovanovic: Thank you all very much. That was such a pleasure.

Diana Mitrea: Yeah, thank you so much, Drago. And if any other questions come up in the meantime feel free to email them to condesates.com, and we’ll share them with Drago.

Diana Mitrea: And with that, thank you so much for being here. Thank you so much, Drago, for a wonderful talk and we’ll let you know when the recording is live. See you at the next Kitchen Table Talk. Have a great day everyone.

Dragomir Milovanovic: Thank you all. Goodbye. Thank you.