VIDEO: Liam Holt on Crowding and Compression

Liam Holt (00:03:13):
And if you just look at this movie right here, even a 20% volume fraction of these sticky particles will start to jam. And the cell is at a volume fraction, so the amount of volume taken up by stuff in the cell is on the order of 40%. And so we think that the system is really on the edge of what’s physically possible in terms of having efficient molecular motion. That’s one thing. And so let’s take a look at a consequence of that in terms of what happens when you start to increase the volume fraction, increase the crowding in the cell.

Liam Holt (00:03:58):
One way that we’ve been doing this is by compressing cells. Physical compression is something that has to be dealt with by all kinds of organisms, from microorganisms through to our tissues, our cells and our tissues. And this is a microfluidics approach that was developed by Morgan Delarue, who was in my lab and now has his own group in Toulouse. What we’re looking at here is a PDMS chamber. And these are yeast cells, and they’re loaded in through this large loading channel–well it’s big enough for cell to get in. And the chamber here is fed constantly by these nutrient channels. The nutrients that the media is exchanged every second. And what I’m going to show is the cells growing and dividing. And as they do so, pay attention to these little pincers here and the shape of the chunk chamber in general, which is going to start to get distorted as the cells grow and divide.

Liam Holt (00:05:00):
This is PDMS, it’s elastic. And you see the pincers now close off this loading channel, you can get completely confined and this allows the system to build up compressive stress. And by looking at how much the chamber is distorted, we know the Young’s modulus the stiffness of this material. From looking at this distortion, we can calculate how much compression, how much pressure is in the system. And we can look at molecules inside the cell as we do this. This is an RNA molecule. And if we look at a couple of different pressures, what we see is that at low, relatively low pressure, this RNA molecule moves around relatively freely. And as we get up to higher pressures, you start to see the molecules moving much more slowly. And we think this is a consequence of this kind of jamming effect.

Liam Holt (00:05:58):
And if you look at what happens to the growth rate of cells as a function of this growth-induced pressure, you see this exponential decrease in growth rate. And again, we don’t 100% understand why this is, but this is a universal phenomenon. You see this in mammalian cells and yeast cells and bacterial cells; always this exponential dependence of growth rate on pressure. And we again, we think that this has an underpinning in the physics of jamming. That’s something that Morgan and I are continuing to study. On the other hand, it turns out that this crowded environment is also important, we think, for biology to go ahead efficiently.

Liam Holt (00:06:51):
And here’s a collaboration that we had recently with Angelika Amon, looking at what happens when you arrest cells in the cell cycle, and therefore uncouple the nucleus–the chromosome content, essentially, from the cell size. What happens, I won’t go into the details, but what happens is that at a certain critical size of the cell, it seems that the cytoplasm becomes too dilute. And then when you try and release the cells from this arrest, they can’t; they can’t go, that’s an senescent. The idea is that some degree of crowding, and in this case also some concentrations on critical concentration of certain molecules, are necessary for biology to be able to continue.

Liam Holt (00:07:40):
Another example of this is, the example of the xenopus extract, one of the reasons that you can do amazing things with the xenopus extract, get the entire cell cycle to go, get the mitotic spindle to build and divide, is probably that you’re able to isolate cytoplasm as a very high crowding, high concentration. And if you dilute that just a little bit too much, more than about 10%, the entire thing just stops working. That’s a little bit speculative, but why might this be? Why is it that having enough crowding is also important? And so one simple way of thinking about this is that crowding also helps to favor molecular assembly.

Liam Holt (00:08:28):
If you consider this simple binding reaction, if you add a crowder to the system, then what happens is that usually for the bound state, there’s this entropic cost. But if you add the crowder, now when you have the bound state here, although the entropy of these two molecules has decreased, there’s more space opened up for the rest of the crowded molecules. And so the entropy of the system is maybe even increased or certainly not decreased as much. And so this is called the depletion attraction effect, and it can help to drive reactions forward. And a good example of this would be, for those of us that remember doing DNA ligation reactions, before Gibson cloning and all that.

Liam Holt (00:09:26):
We had a regular DNA ligase kit and we had our rapid ligase kit. The reason that the rapid ligase kit goes in five minutes instead of an hour is that crowding agent was added to it. You can really drive reactions a lot more quickly by adding crowder to the system. On the other hand, if you go too far as I presented earlier, then you start to jam the system up and these molecules can’t find each other anymore. And what this means is that if you look at the rate constant of some reaction, that depends on this kind of binding effect as a function of crowding agent, you’ll get this kind of biphasic behavior with an optimum. And presumably the cell has mechanisms to maintain its internal crowding at or around this optimum value.

Liam Holt (00:10:16):
And I’ll point out that this optimum might not be the same for every reaction, so that’s an interesting potential source of sensing and regulation on the cell. For example, just say dilution is not good and too much compression is not good in terms of growth rate, but there are some more subtle things that can happen. For example, if you slightly compressed the cell, you can get these dramatic phenotypic changes to occur. And this has been observed for a while, but here’s a really nice example of that from Kevin Alessandri. What he’s doing here is growing just a spheroid of mammalian cancer cells. And what you see on the left here, the spheroid of cells, it kind of just grows and it stays fairly cohesive. Nothing particularly dramatic here.

Liam Holt (00:11:06):
On the right it’s exactly the same cells, but what they developed was a really nice alginate encapsulation technology. They essentially injected the same cells into a thin gel capsule. And the capsule is designed such that the cells could build up compressive stress and pressure, but and then eventually rupture the capsule. Just pay attention as it ruptures to what happens to the cells upon release from this compressive environment. And you see that there’s this dramatic change in their mortality, their coherence and so forth.

Liam Holt (00:11:43):
And this kind of [EMT inaudible 00:11:44] type behavior driven by compression has been observed in multiple places. There’s a sense that these compressive events can drive these large regulatory changes. And of course there are active signaling processes involved here, things like Piezo channels and various signaling pathways. But it’s interesting to consider that some of these general physicochemical, environmental changes inside the cell may also play a role. In fact, they’re certainly going to play a role in all biology.

Liam Holt (00:12:24):
The cell has this crazy crowded environment. And of course this audience realizes that this particular picture from David Goodsell, it’s based on a prokaryotic system. In fact, it’s far more messy than that in eukaryotic cells where there’s a huge fraction of disorder in this proteome and in terms of the polymers that are around in the cell. This is actually what got me into this whole mess in the first place was thinking about disorder.

Liam Holt (00:12:56):
This thing is, the cell is also, it’s not a simple frictionless sphere system where all the spheres being the same size, it’s tremendously polydisperse. You have molecules that range over orders of magnitude of length scale and then you get up into organelles and so forth. And a particularly poorly understood length scale is what we call the mesoscale, which is about the range of between 10 and a couple of 100 nanometers of diameter.

Liam Holt (00:13:27):
And that’s where a lot of really interesting biology is going on. For example, all of the transcriptional machinery or the translational machinery, proteasomes, et cetera, et cetera. The assemblies of proteins and nucleic acids that are controlling a lot of the reactions and the regulation of the cell. Getting back to this disorder, as I was saying, the reason that I got into this whole mess is that, back as a graduate student and then the year right after that I was looking at the cyclin dependent kinase. And we did a quantitative mass spectrometry screen to figure out where a bunch of it’s, well what a bunch of its substrates were, where those substrates were being phosphorylated in the protein. And this is just a very simple diagram to show that, minimize this, that… What I’m showing here is the fraction of the proteome that’s predicted to be in helix, sheet or loop with one prediction algorithm or over here, disorder versus domain.

Liam Holt (00:14:34):
And the gray is the fraction of the proteome in general. And we just focus over here in this sort of. So first of all, 30 to 40% of a typical eukaryotic proteome is predicted to be disordered, which was kind of shocking to me at the time. And now I think is becoming more and more, there’s more and more awareness of this now. But this is where all of the regulation is happening in terms of phosphoregulation and lots of other post-translational modifications. Really the reason I started thinking about all of this physicochemical environment stuff is, I was trying to figure out, well how on earth do you derive regulation from phosphorylating and all of the disordered protein. It didn’t really fit with this Lehninger textbook model of phosphates coordinating a precise change in the conformation of a protein.

Liam Holt (00:15:27):
While of course that does happen, it’s perhaps a minority of cases. The questions that we were asking how is the environment, this crowded crazy environment regulated in the cell? How does this crowding impact physiology? And finally, does this crowding change in disease conditions? In particular, we’ve been thinking about what happens when we start to build up compressive stress or changes in the mechanical environment. Okay, the first story I’ll tell you about is driven by Greg Brittingham, a grad student in my lab, and Morgan Delarue who now has his own lab Toulouse. And it’s about the control of physical properties.

Liam Holt (00:16:15):
I should say that I’ll pause in a couple of places explicitly for questions, but you can also jump in at any point with a question if you have one. The way that we’ve been studying the physical properties, the crowding and the cell is through this technique of passive microrheology, which is simply the idea that if you watch the motion of some kind of passive tracer inside of a material, you can infer things about the properties of that material. Here’s a very simple kind of illustration of that. If you have, imagine you can’t see the environment, but you can see this green bead. If you watch the bead, you can guess that this bead that just jiggles a little bit is in a far more crowded environment than this bead over here that is moving more freely. That’s just a very gross simplification of the field of microrheology.

Liam Holt (00:17:13):
Now the way that people have studied this in the past has been to microinject inert nanoparticles or to get them to get into cells by other techniques, such as pinocytosis. But these kinds of techniques have been quite labor intensive and typically very low throughput. They tend to perturb cells, for example, by diluting of cytoplasm and disrupting membranes. And very importantly, they’ve been completely impossible in any organism that has a cell wall, for example, cerevisiae which is one of our work horses for figuring out the initial genetics of systems.

Liam Holt (00:17:54):
To overcome these limitations, what we decided to do was develop these genetically encoded multimeric nanoparticles or GEMs. What these are is a gene that encodes a protein that we’d got from hyperthermaphilic archaea, and these proteins are encapsulants. Prokaryotes don’t have membrane bound organelles, but they do almost universally make these protein cages. And what’s cool about the encapsulants is that it’s just a single protein, a single monomer, that very robustly self assembles, in this case, 120 copies of this monomer to form this precisely defined nanoparticle cage, which is about 49 meters in diameter. And all we had to do was to tag this thing with green fluorescent protein. And the C-terminus is projected into space. And so what you wind up with is this nanoparticle with perfectly defined shape and size that is coated with 120 GFP molecules. And so it’s nice and bright and we can use this as a genetically encoded nanoparticle. Any cell that has this gene in its genome will constantly produce nanoparticles that we can then look at and track and do physical analysis for it.

Liam Holt (00:19:30):
And this size here, is at that sweet spot, that length scale, the mesascale, where we think that a lot of interesting biology is happening. Here’s what it looks like in a mammalian cell. This is a pancreatic cell and this is a yeast cell. Here we have the GEM nanoparticles, and here we have the nucleus, actin. This is a realtime movie. So the movie is looping, and this is four seconds of acquisition. We’re acquiring a frame every 10 milliseconds to be able to do accurate particle tracking on these nanoparticles. And we can get thousands of tracks from a single cell in this four second experiment. We can really brute force through lots and lots of experiments using this kind of approach. Here over, this is cerevisiae, no one had ever been able to do rhealogy on a yeast cell before.

Liam Holt (00:20:28):
But now we can get these nanoparticles into these cells that have cell walls, and you immediately start to see some really interesting things just by looking at this cells. For example, in a single cell right here, you have some nanoparticles that are barely moving and other ones that are moving quite freely. There’s a, even within a single cell, there’s quite a bit of heterogeneity you would guess in the intracellular physical properties. And my guess is that this kind of a low impeded movement right here might be a reflection of some kind of local jamming.

Liam Holt (00:21:08):
As soon as we started looking at these cells with the GEM nanoparticles, honestly someone designed this technology to ask a different set of questions. But what we immediately observed was that even on a global scale, the physical properties of the cell were not constant. And from day to day and experiment to experiment, it seemed like there was quite a bit of variation and the diffusion coefficients that we would get from tracking these particles. We decided to try and figure this out and one thing that we figured out was that the cells are way more sensitive to the precise growth conditions and nutrition conditions than people often take care with in yeast experiments. And you don’t need to have very much perturbation at all before you start to see these big physical changes. And we tracked it down to amino acid starvation.

Liam Holt (00:22:08):
This is the effective diffusion coefficient normalized to control here. When you start to run out of amino acids, you start to see this big bump up in the diffusion coefficient of these nanoparticles, suggesting that the crowding inside the cell is decreasing. And since this is amino acids, we immediately thought about a potential regulator of the system, which is TORC 1, which is known to be the one of the main sensors of amino acids. TORC kinase, it’s essential regulator of growth and metabolism, it integrates a lot of information about, in mammalian cell’s growth factors in all cells, amino acids, carbon source and stress. And it controls pretty much everything.

Liam Holt (00:22:53):
And so the reason it’s called target of rapamycin is because of this guy. This is one of those heads from Easter Island. Easter Island also known as Rapa Nui, which is where this small molecule was discovered that turns out to very, very specifically and potently inhibit this kinase. We can simply use this ancient technology from Rapa Nui, rapamycin, to ask the question: is TORC kinase the thing that is controlling the physical properties of the cell? Here’s the experiment, it’s the same movie I showed you before. Again, look at these jammed nanoparticles. After treatment with rapamycin, and this takes an hour to manifest this effect, so it’s not an immediate kinase inhibition effect. What you see is that after treatment with rapamycin, I think even by eye you can get a sense that the nanoparticles are moving more rapidly and you don’t really see these jammed nanoparticles as much anymore. And if we quantify that, so this is the value of the effective diffusion coefficient of these nanoparticles in the control, and treatment with rapamycin increases this diffusion coefficient, again suggesting a decrease in crowding in the cell.

Liam Holt (00:24:15):
How is TORC impacting these physical properties of the cytoplasm? And this was a question that we would never have answered if we didn’t have this technology because like I said, TORC controls everything. And all of our great ideas and hypotheses that we had for like the first couple of years as we’ve tried to like slowly work our way through our wrong ideas, they were incorrect. And things like cytoskeleton, cell size, and all of this stuff didn’t really explain what TORC was doing. But because we were in yeast and because each experiment we could have been done in a few seconds and generating a mutant and yeast only takes a couple of days, we could go through hundreds of mutants and finally find the thing that was controlling the physical properties here.

Liam Holt (00:25:08):
What we’re looking for is this. In wild type cell, I’m showing you the same data again. When you treat with rapamycin, the diffusion coefficient increases. We’re looking for mutants where this effect of rapamycin goes away, and eventually we found such mutants. And here’s an example, so this is SFP 1. The control and the rapamycin-treated cells look the same. There’s no further effect of inhibiting TORC. And actually, maybe unexpectedly the control condition now is already de-crowded. It looks like the rapamycin treatment before doing anything.

Liam Holt (00:25:50):
What is SFP 1? It turns out that SFP 1 is the major transcription factor that regulates the rate of ribosome biogenesis. And as we look through the various mutants we found half a dozen or so that had this kind of epistasis with rapamycin effect. We found two classes that were involved either in ribosome biogenesis or in autophagy. And autophagy is the only thing that really degrades ribosomes. Together these two processes are tuning ribosome concentration in the cell. We sent that off and got rejected. And so then I collaborated with my friend Ben Engel, who we were in grad school together and he was in Wolfgang Baumeister Institute in Munich and they’re sort of raison d’etre is doing this amazing Cryo-FIB milling EM tomography. You basically take a focused ion beam and you mill out a thin lamella. Inside this droplet here you have cells that have been flash frozen in liquid helium, I think it is.

Liam Holt (00:27:08):
And as you zap a thin lamella out, the cell will be revealed into a 500 nanometer piece of cytoplasm or nucleus or what have you. You can then take this lamella, you can stick it into a direct detector EM beam and rotate the sample and do tomography. And so this is our experiment. So here are organelles and so forth. And as we pan back, you can fill in the electron densities with known structures. And one of the things that dominates is the cyan thing and this is ribosomes. And you can see just how crazy-packed the cytoplasm is with ribosomes. These orange things are our nanoparticles. And so for example, you can imagine that right here in this area, you might have a local jamming, while over here, it might be a little bit more free.

Liam Holt (00:28:06):
I mean that’s speculation. But if you inhibit TORC kinase, what you immediately see by eye, is that this concentration of ribosomes has dropped by almost twofold, and a lot more space has opened up. Very clearly, TORC is controlling the concentration of ribosomes in the cytoplasm. And we were able to quantify in a whole bunch of conditions and mutants and treatments what the relationship was between the diffusion coefficient of nanoparticles versus the concentration of ribosomes. And we were able to completely parameterize this Doolittle equation that predicts how crowder should impact diffusion.

Liam Holt (00:28:54):
And this blue line is a prediction; it’s not fit to these data points at all. It’s just what we predict should happen. And the data points fall pretty much on the line, which doesn’t say the model is true, but it definitely says that our model is not crazy. This is what we think is going on with TORC kinase and the regulation of the physical properties of the cytoplasm at this length scale. What might consequences of these crowding changes be? Well, an obvious one that we’ve already seen is that the rate at which particles move in the cell will change. But beyond that, as I introduced earlier you can also have dramatic effects potentially on the interaction of molecules. And this is where we started getting back into this idea of phase separation. Here’s the famous vinaigrette Google image search and then on the right we have Cliff’s worm from Woods Hole. And as we all know in this call this is the spontaneous de-mixing of macromolecules.

Liam Holt (00:30:07):
And while I introduced earlier that a simple bimolecular reaction, the binding of this bimolecular interaction can be favored by crowder. This is perhaps even more true of phase separation. And if you look at all of the phase separation papers and the methods of all of the in vitro work, the vast majority of them are adding Ficoll or PEG or what have you. And every single crystal structure that was ever made screened through these crowding conditions to try and get phase separation to work. And the reason is similar, that as you add crowder, you offset the entropic cost of this condensation reaction, of these networks of molecules forming, and you tend to drive that phase separation event.

Liam Holt (00:30:54):
But is the change that we see in the concentration of ribosomes in the cell sufficient to make a difference to any kind of phase separation that might be biologically relevant? And it’s not clear that that would be the case. I mean, we’re seeing a twofold change in ribosome concentration. We decided to explicitly ask this question. And what we used as a model system, so we were working in the Summer Institute at Woods Hole along with Mike Rosen and a bunch of other people, Tony was there as well. And so we were using Mike Rosen’s SUMO-SIM system, where you have this poly-SUMO multivalent molecule interacting with the SUMO interaction motif.

Liam Holt (00:31:37):
And what we see here is a graph showing partition coefficient as a function of ribosome concentration. And these ribosome concentrations are the same ribosome concentrations that you find in the cell. In a happily growing crowded cell, you have about, well you have exactly 23 micromolar ribosome concentration. And in this condition you have a partition coefficient above three and that’s shown in this micrograph here. This is a completely in vitro system again, this is just ribosomes and then the poly-SUMO and the poly-SIM.

Liam Holt (00:32:15):
And then this is the concentration of ribosomes once you’ve inhibited TORC kinase with rapamycin, so you go down to about 13 micromolar. And you can see that the degree of phase separation, the partition coefficient, has gone down. You also see things like the amount of wettening is higher, so the surface tension is lower I guess. Very clearly in vitro, just as very, very simple system, you can certainly tune phase separation. But how about in the cell? Well the cool thing about the SUMO-SIM system is that you can do the same reaction, the same condensation reaction in a cell. In this case we use a SUMO(10)-SIM(6) single polypeptide fusion to make sure that the stoichiometry of these things are always the same, fused to mCherry. And again, this is a plasma that we designed based on Mike’s work. And so, here we have cerevisiae cells and human cells.

Liam Holt (00:33:20):
Everything we did and cerevisiae we also did in mammalian cells. And the two things track, both in terms of the earlier work with genetics and also this work with phase separation. What you can see is that with high ribosome concentration, you have a large amount of phase separation, and with lower ribosome concentration after rapamycin treatment, the number and size of these droplets goes down. And that’s quantified here. After rapamycin you have less total droplets. And very importantly, we have this control where we use sorbitol to compress cells and therefore increase crowding again. And we can figure out how to perfectly increase crowding back to what it was prior to rapamycin treatment just by looking at the GEMs. We put in enough sorbitol to restore the gem motion to pre-rapamycin conditions.

Liam Holt (00:34:12):
This turns out to be a pretty heavy compression for cerevisiae, much less so for mammalian cells. And what you see as you recover phase separation. Not 100%, and this is probably because we’re losing a little bit of protein as we inhibit translation with TORC kinase inhibition as well. But you suddenly recover most of the phase separation. This paper was published a little while ago now, like a year ago. And just in summary, with these nanoparticles we can see that the concentration of ribosomes modulates the effective viscosity, the crowding in the cell, and then also at the same time tunes phase separation. I want to pause here and see if anyone has any questions that they’d like to raise.

Diana Mitrea (00:35:08):
Hi, this is the Diana. That was beautiful and that’s very elegant. I just have a question. What do you think happens to the ribosomes in that time span of an hour? Given that the lifetime of a ribosome is five to 10 days? Do they dissociate to decrease the crowding?

Liam Holt (00:35:32):
We know that autophagy is really important. If we inhibit autophagy, we don’t get these same changes. What is probably happening is that the, as you inhibit TORC kinase, you turn off production, but you also turn up autophagy and that’s probably degrading a lot of the ribosomes. Cell volume is increasing a little bit as well, not a ton; that may help dilute a little bit. That would be my answer.

Diana (00:35:58):
Okay. And also talking about timelines, can you, what would be the timeline of reversing the process? So if you wash out rapamycin how quickly do you? [crosstalk 00:36:10]

Liam Holt (00:36:10):
Good question. Yeah, we haven’t done that experiment, honestly. It’s a little tricky to wash out rapamycin. It tends to stick around. But we honestly haven’t tried. It’s a really good question. I’m sure we can figure out how to do it with a different molecule, like a rapalog or something.

Diana (00:36:31):
Okay. Thank you so much.

Liam Holt (00:36:33):
Yeah, thanks. Any more thoughts or questions?

Mark (00:36:39):
On that last point, I think the rapalog idea is a good one because as I’m sure you know, some of the other rapalogs that actually made it all the way to market or into the clinic had very different physical properties. And so their behavior, their physical behavior in your system could be quite different. That’s probably how I would think about it. Try to find the most structurally diverse, the most polar rapalogs that are still highly potent.

Liam Holt (00:37:05):
Yeah. I think that would be very interesting to see if they could be distinguished based on this physical effect as well, if that tracks with any clinical applications.

Mark (00:37:16):
Yeah, that’d be fantastic.

Liam Holt (00:37:18):
Yeah. Other thoughts? Okay, I’ll keep going. If you have any other ideas, feel free to interject. Okay. Just a quick a little side note here. We’ve been studying the nucleus as well. These nanoparticles we started out in the cytoplasm, but it turns out we can also send them to the nucleus just by putting a nuclear localization signal on them. It turns out that the way that we have to do that is by putting the NLS on the N-terminus. Because if we put the NLS on the C-terminus, you get these very strong interactions with the nuclear periphery, probably there’s a lot of interactions with the nuclear pore. But if we put it on the N-terminus, what happens is we think, in fact, we pretty much know, that the monomers go into the nucleus and then the particles assemble inside the nucleus.

Liam Holt (00:38:20):
And they’re big enough that they, we don’t think that they fall out of the nucleus anymore. These are nuclear GEMs here, and this is a human cell, a pancreatic cell. And this is more or less where the nucleuses is. And I just told you that we don’t think that the nanoparticles can really get out through the nuclear pore, but there are nanoparticles out here in cytoplasm in a mammalian cell, so we wondered about that. And it turns out that they’re falling out during mitosis. Here’s a much longer timescale movie and the particles start in the nucleus but then upon mitosis, you’ll see that all the nanoparticles, they not only fall out, its almost like they are forcibly ejected from the nucleus. You wind up with these two daughter cells that have nanoparticles in the cytoplasm and not in the nucleus.

Liam Holt (00:39:12):
And if you wait for a while, you’ll start to see that the particles form again in the nucleus and you eventually go back to this kind of state where you have particles in both compartments, which is fine. We can, stay in the nucleus and we can analyze each compartment individually. We thought it was kind of interesting. And one of the reasons this was really interesting, I’ll just show that in a sensible organism, like yeast, that like most biology does closed mitosis and not this crazy open mitosis the mammalian cells do. Here are nuclei and the cell outline in cerevisiae cell and here are our nanoparticles. They are completely in the nucleus. Here’s a time projection. And basically you never see nanoparticles in the cytoplasm in yeast. And again, yeast don’t have an open mitosis, so there’s no opportunity for these nanoparticles to get out.

Liam Holt (00:40:05):
Now the reason this was a bit confusing at first, was that, oops let’s go back. Sorry. If we look at these movies again, so here’s the nuclear GEMs and the nanoparticles in the cytoplasm and in the nucleus. But our original GEMs that we looked at, they were never in the nucleus. What’s going on? Because this is a stable cell line. This has gone through dozens and dozens and dozens of divisions. And yet there are no nanoparticles at all in the nucleus. We sort of talked with Daniel Gerlich about this at Woods Hole, and he’s been very interested in how the chromatin and the nucleus are reorganized during mitosis and then the exit from mitosis.

Liam Holt (00:40:54):
I just want to share this one result because it’s really cool. Hopefully this will come out soon. This is Daniel’s work, Sarah Cuylen and Mina Petrovic in his lab. He took advantage of our GEMs and here we have chromosomes that are in, they’re actually arrested in nocodazole in mitosis. And you can see that actually when you have the chromosomes condensed, there’s an intermingling of the nanoparticles with the DNA. You can see it down here with the GEMs alone here. If we watch a movie, he’s adding, reversing, so he’s driving exit from mitosis. And what you see is that as the cell reassembles its nucleus, the nanoparticles get forced out. We can look and see, just stills here. This is before, and this is after reassembly of the nucleus. The chromosomes cluster together, the nuclear envelope forms around those chromosomes, and the GEMs have been completely pushed out of the nucleus. And this is a very efficient process.

Liam Holt (00:42:05):
And it turns out that there’s this chromosome clustering mechanism that’s super important. We can look at ribosomes as well, so ribosomes are similar size to gems. This is a GFP tag ribosome sub unit. And similar things, so cells forced out of mitosis and what you see is that the ribosomes all get pushed out. If we use a mutant that’s deficient in clustering the chromosomes, what you see is that upon forced exit from mitosis, you don’t get this perfect exclusion of ribosomes anymore. And we can show that the exclusion of ribosomes doesn’t really depend on nuclear export pathways because we can inhibit this typical pathway that would drive most nuclear export with Leptomycin B and you still see in this case, very nice physical exclusion of ribosomes from the newly forming nucleus. And that’s quantified here. The control and Leptomycin B, they both exclude ribosomes, but the clustering deficient cells don’t really do it so well.

Liam Holt (00:43:18):
The point is that the cell can do amazing things in terms of self-organization using these principles of density and differential crowding. And then, you can, the chromatin of a cell can really redefine this length scale, the entire nucleoplasm versus cytoplasm before the nuclear envelope is even formed. And this is really important. One might imagine for all of these molecular complexes that are too big to get out through a nuclear pore. I thought this was really cool. Again, Daniel’s work. It was really fun that he was able to follow up on that observation.

Liam Holt (00:44:04):
If we look at a nucleus and interface now, so here are our nuclear nanoparticles and here we’ve stained the DNA with sirDNA. And we assume that the brighted stain here corresponds to most of the heterochromatic regions. And if you just look at the nanoparticles moving around, you can kind of see that the nanoparticles aren’t going into the heterochromatin, they’re corralled and moving around inside this, what’s probably more like the euchromatic compartment. This physical control is powerful enough to completely define the nucleus versus the cytoplasm. And then once you have a nucleus and it starts to selectively recondense, one might imagine that there’s a tremendous regulatory potential here.

Liam Holt (00:44:51):
We’re just starting, I mean, we’ve been working on this for a while, but it’s really kind of a challenging thing to figure out how to really study this. This will be a fun thing to maybe chat with you about. Beyond the nucleus and in terms of physical organization of space, we’ve also targeted the nanoparticles to the endoplasmic reticulum. This was with Joe Chambers and Cambridge. And what you see here, if we focus on the merge is that the nanoparticles tend to be, this is actually a 20 nanometer nanoparticle, the smaller one.

Liam Holt (00:45:27):
They tend to be in the tubules. And the sheet’s maybe too thin for the nanoparticles to get into. And this is something that’s been known about the ER and sorting in the secretary pathway is that there’s the size-dependent sorting that happens, but we can really see it here. How might this thing be impacted by physical changes like compression? Okay. we have these big questions. We’re interested in what controls crowding and the physical properties of a nucleus. And we’re really interested in whether these physical properties, how and when do they impact transcription?

Liam Holt (00:46:04):
This is a huge set of questions and excited to think about how we might address them and whether other people want to work with us. Okay. That was a sort of a little bit of an aside. Now I’ll take a moment with this wonderful gif of a rubber band ball being squished, to transition into the work that we’re doing on compression and crowding. And I’ll, sort of zip through this really quickly. Everything’s a work in progress. Greg has been really spearheading this. Tamas has been sort of working more on the nuclear side of things. And we continue to collaborate always with Morgan.

Liam Holt (00:46:47):
Again, here’s our yeast-, Morgan’s yeast torture device that we use for compressing yeast cells. And what we see as the cells build up this one mega Pascal of pressure, which is 150 PSI, like double the pressure you would put in racing road bike tire. What we did is we looked at different particles in the cell, different particles in different compartments as well, and quantified how the diffusion coefficient on the Y axis normalized to the condition with no pressure. How does this diffusion coefficient scale with pressure? This is a semi log plot. The straight line indicates that this is exponential dependence, just as the growth rate has an exponential dependence.

Liam Holt (00:47:35):
And what you can see is that different sized particles. If you look in the cytoplasm at this 20 nanometer particle versus an RNA, which is more like 100 nanometers, there’s this different size scaling in terms of diffusion versus pressure. What was even more interesting to me is that the same exact nanoparticle, 40 nanometer nanoparticle, in the cytoplasm versus the nucleus, scales differently. The nucleus feels compression differently than the cytoplasm feels compression. And we don’t know why this is, but of course the crowder here is ribosomes here, we don’t know probably some combination of chromatin and RNA. But it’s actually kind of hard to nail it down right now. These different systems are behaving differently. That’s interesting.

Liam Holt (00:48:30):
Like I said, we’d been thinking about this in terms of disease models. And the two ones that we’ve been thinking about are cancer and we’re starting very, very recently to think about neurodegenerative disease. But let’s start with cancer. Tumors build up a really tremendous amount of compressive, solid compressive stress. Here’s a very evocative image, slightly gruesome from [inaudible 00:48:55] lab. What they did is they took, this is a pancreatic tumor and they took a razor blade essentially and did a very well-defined slice and watched how this things pops open. And the reason it pops open is that it has built up solid compressive stress inside.

Liam Holt (00:49:16):
And so we’ve been focusing on pancreatic cancer, because it is the cancer that builds up the most compressive stress of them all. They all build up quite a lot, but pancreatic especially so, because it builds up all of this change and the ECM. All of these tumor associated macrophages, this tissue becomes very stiff. And then as the tumor grows within this stiff tissue, you get this compressive stress. Pancreatic cancer is obviously its huge need. It’s becoming one of the most deadly cancers that really. We don’t have good treatments.

Liam Holt (00:49:52):
And one thing that as a geneticist was appealing to me is that, it’s always driven by KRAS. And this is crazy, like more than 95% of the time these tumors, these ductal adenocarcinomas are initiated by a KRAS oncogene. We decided to check the idea that maybe KRAS would have a impact on crowding in the cell and the physical properties of cells and the ability of cells to respond to compressive stress. And of course KRAS is related to mTORC in its signaling.

Liam Holt (00:50:32):
And so we did this, preliminary experiment. Just we took MEFs, and we either took wild type or KRAS, a single copy of KRAS oncogenic activation mutation K-12 B, this is the one that we find in pancreatic cancer. And we just osmotically compressed cells just as a very quick and easy way of asking: did the physical properties get modulated? And what you see is that in wild type cells, as you get to these 200, 300 millimolar sorbital conditions, they all start to die off, but the KRAS cells do better.

Liam Holt (00:51:05):
And so this is enough to motivate us to continue on. And so we decided to make some models for human pancreatic cancer. We started with these human pancreatic nestin expressing cells. These are fairly well and used model for relatively normal pancreatic cells. We knocked out p53 with CRISPR and then, and the reason we did this is because it’s difficult to introduce KRAS oncogenic mutations before you knock out p53. This is not what happens in disease, so that’s a caveat, but it allows us to do the comparison to this control, p53 knockout, versus a mutant where we’ve introduced the KRAS oncogenic mutation. And again, this human disease is pretty much always iniatiated by KRAS and then you typically get loss of a bunch of tumor suppresses including p53.

Liam Holt (00:52:03):
What we immediately see is that there is a substantial change in crowding, well, the physical properties of the cell. I’ve been saying in the yeast part of the talk that this change in diffusion coefficient, apparent diffusion coefficient, is due to crowding. Here in the case of these very much more cytoskeletaly active systems, we see an increase in the diffusion of the nanoparticles. This could be a change in crowding. It could also be that the actin/miosin cytoskeleton for example, is getting way more contractile and this could have an impact on mobility in the cell. Nevertheless, there is a big change. And so we thought that was interesting and we decided to look and see how the diffusion coefficient of nanoparticles changes in cells as we do this sorbitol shock condition.

Liam Holt (00:53:03):
Here we have a pancreatic cell again. In control conditions, we have this free diffusion, or this rapid diffusion. If we do a 300 millimolar sorbitol shock, we got this. This is a movie right here, you can’t really tell except for the fact that it’s photobleaching the nanoparticles, they really almost completely stopped moving. We’ve really jammed up the system at this length scale. Now of course, smaller things like individual proteins can still move around in these osmotically compressed conditions. And that’s kind of interesting to think about.

Liam Holt (00:53:45):
Now, 24 hours later, if you come back to these cells, they really haven’t started recovering at all. And these cells will ultimately go on to die. If you look at the KRAS mutant cells on the other hand, which start with a higher effective diffusion coefficient. If we shock them, the diffusion slows down, but then 24 hours later they’ve completely recovered. And we don’t really understand how this is working yet. But this is something that we were trying to follow up on.

Liam Holt (00:54:17):
But that’s osmotic compressive stress. And while that is relevant to the tumor microenvironment, you have fairly substantial changes and the osmotic strength inside tumors, we’re really also interested in mechanical compressive stress. Greg has come up with this very simple system for looking at how compression of these HPNE cells impacts their physiology. He 3D-printed this wafer and surrounds it in agarose pads. And then he uses these washers he got from a hardware store to, sticks these down on cells on an imaging well. Depending on whether he puts washers on there or not, you get compression. The cool thing about the washers and this wafer is that we can image through the hole in the middle. So toroids are useful, not only for donuts.

Mark (00:55:09):
That’s just brilliant, brilliant idea.

Liam Holt (00:55:12):
That was all Greg, so I was very impressed. Okay, this is what we see, is that it’s different than the osmotic compression. After adding the washer you get this slow down, and this is probably around a 300 Pascal pressure, less than you see in a tumor. And then the KRAS, you also see the slow down. But the difference here is that the KRAS doesn’t really recover. There’s a big difference between osmotic and mechanical compressive stress. These cells, they can actually survive this condition. They’re not quite so perturbed as the osmotically compressed cells.

Liam Holt (00:55:53):
They do continue to grow and they continue to divide. And we need to take out this experiment to longer times to see if the cells eventually recover a bit more. But the cells are managing to continue their biology with this very perturbed crowding environment. And one of the consequences of that is that when you look at mitosis, cells going through mitosis, here we have control cells, p53-deleted cells. And we’re looking at the fraction of cells that go through mitosis that have an error. And what you see is that in the control cells, compressing them doesn’t really make a big difference.

Liam Holt (00:56:35):
They don’t really have very high rate of mitotic catastrophe. But if you have compression and you have KRAS, so KRAS alone, they are dividing just fine. But if you combine these two things, it’s a complete just disaster. You have this huge rate of mitotic error. And so here is what a controlled cell should look like: This is just a DNA stain, it’s false colored for the Z stack, so that’s why it’s rainbow coloured is just DNA. And this is anaphase in the cells, merrily dividing in the control cell, this is with no compression. But if you take KRAS mutant cells under compression and you watch the kind of thing that happens, this is the same timescale. This is now these cells are staying in anaphase for hours and hours and hours. You see this, I mean, I wouldn’t even call this a bridge, this is just like a failure of segregation of certain chromosomes.

Liam Holt (00:57:37):
We suspect maybe the larger chromosomes are particularly susceptible to this, like chromosomes one, two, three. And you get this just disaster. And this is not the only thing you see. You also see multipolar divisions, and you see more subtle mitotic errors. And we’ve been spending a lot of time acquiring high content movie video of microscopy and then figuring out how to get computers to track cells and classify divisions, which has been kind of fun. Very challenging, especially, like tracking divisions when they are so messed up is a real challenge.

Liam Holt (00:58:19):
Why are these cells having such a hard time? We went back to again, the nuclear nanoparticles and the cytoplasmic nanoparticles to see what happens when we mechanically compress cells. What I’m showing you here is on the Y axis is the percentage decrease in diffusion upon compression. At zero that would mean that if we compress the cells, there was no change in diffusion. And as we go down, there’s more and more decrease in diffusion when we compress the cell. If we look at the control cell versus the KRAS mutant, and look at cytoplasmic nanoparticles here, we see that they both slow down in this particular condition about 20%. But if we look at the nuclear nanoparticles, what we see is that in the control cells, actually the nuclear nanoparticles slow down less than the cytoplasmic nanoparticles.

Liam Holt (00:59:16):
And perhaps this is in contrast to what I showed you earlier with yeast, where we’re compressing everything isotropically. In mammalian cells, we’re applying the compression from above, and the nucleus is actually the most mechanically stiff part of the cell. And so I think the nucleus is resisting the compressive stress more effectively than the cytoplasm in the controlled cell, so you’ve got less slow down. But in the KRAS mutant you actually see, way more decrease in diffusion than in the p53 control cell. There’s a bigger perturbation there and we kind of guessed that there might be less mechanical stability in a KRAS mutant cell.

Liam Holt (01:00:00):
We also looked in cells that were expressing a GFP that’s targeted to the nucleus as a way of checking whether there is nuclear rupture that’s happening just during growth. And so what you see, if we focus on this cell right here, every time the GFP blinks out and leaves the nucleus, that’s because the nuclear membrane ruptures. These cells are growing, these are KRAS cells growing under compression and you see that there’s this frequent nuclear rupture happening. And here’s another example down here. And you see that the structure of a cell and you can even see micro nuclei being ejected out.

Liam Holt (01:00:38):
And so we think that both in interface and in mitosis, and this might lead to all kinds of DNA damage that there might be resolved or lead to problems in anaphase. Again, we’re trying to figure all this out. But one thing that’s very interesting about all of this is that we know that in recent sequencing studies, this idea of this slow linear accumulation of mutations in tumors has kind of been, it’s fallen out of favor in many examples.

Liam Holt (01:01:16):
And in fact what you tend to see is a pattern of mutation that’s more consistent with giant catastrophes happening, where multiple genes all get chopped up at the same time through things like chromothripsis and massive aneuploidies. And then subsequent selection upon that new genome. And so this is from the Gallinger lab. This is what they kind of found to be happening all the time in pancreatic cancer as well. And so what’s not known is what is the driving force for these huge genomic catastrophes. We suggest that perhaps what we’re seeing with this combination of the KRAS mutation, which again is always present, and compressive stress, which again is always present, might be part of this story. We’ve been doing evolution experiments and we’re doing genome sequencing and we’re seeing interesting things in terms of chromosome amplifications and losses and so forth.

Liam Holt (01:02:24):
It’s on pause for now as everything is, but I’m excited to see how that all plays out. Okay, that’s kind of what I wanted to tell you for the most part. Keeping crowding perfect homeostatic level is crucial for life. I’ve alluded to this a number of times, but we’re very interested in looking at neurons as well. Here’s an image of a neuron from Ramony Cajal and then here’s an Actin mRNA that’s trying to make its way up and down an exon. And this is a tremendously crowded environment.

Liam Holt (01:03:05):
It’s a huge problem to move these RNPs up and down these meter-long axons. And as I was talking to Mark about earlier, these connections between these RNA granules, which themselves are condensates, and things that they track with like lysosomes that they’re all, speaking to things like TORC kinase. They’re all going to be dependent on things like the concentration of ribosomes, concentration of microtubules. And they’re also going to be impacted, I’m sure, by changes in the mechanical environment that may impact the amount of space available.

Liam Holt (01:03:46):
It was a very nice paper from Kevin Chalut, earlier last year, at the end of last year showing that the mechanical properties of bringing change vary substantially as a function of age. And so we’re interested in working with Kevin and other people, [inaudible 01:04:05]here in New York, and starting to build models. So we have nanoparticles, we can put them in neurons. We want to start looking at these transport processes and so forth and phase separation processes. The kind of things that I know that you guys were interested in as well. But this is challenging work and we need to talk to more people about it.

Liam Holt (01:04:31):
Final slide, all the drugs that make you live forever–Metformin stops you from aging, so does red wine with it’s resveratrol, rapamycin, and starving yourself and making life horrible by putting apples in your cheeks. All of these things impact things like TORC, crowding, mechanical changes like inflammation and all of these things we think are playing a role and feeding into things like aberrant phase separation. We’re curious to keep on following down these leads. And I’ll finish up just by again acknowledging Greg, Tamas and Morgan, along with all of these other collaborators that we’ve been working with over the years. And stop again for any further questions.

Mark (01:05:23):
Thank you. Wonderful stuff. It’s wonderful how you weave together so many different themes in your research.

Liam Holt (01:05:30):
Yeah. We probably have too many themes in our research, always looking for collaborations.

Mark (01:05:36):
Absolutely. And it also makes me wonder about 2D cell culture, why it works at all.

Liam Holt (01:05:48):
Yeah. Yeah. 2D cell culture, it certainly has its limitations for the imaging we’re doing. Unfortunately we’re also using 2D cell culture because we need to have very high resolution imaging. But I think that we need to also move into things like organoid models and things like that.

Mark (01:06:11):
Yeah just the whole idea. I mean, all of your work shows how important these crowding effects are and the three dimensionality of the real system. And yet we are able sometimes to get quite useful data even from relatively simple 2D models, which it’s great that we can but it’s a good thing. It’s good fortunate that we can, given how much simpler those systems are than the real cellular environment. I think Alicia has a question.

Liam Holt (01:06:43):
Okay.

Alicia Zamudio (01:06:46):
Yes. Hi. I was wondering if you could expand on what you think is driving crowding in the nucleus, do you think its like really large molecules? Do you think its like really abundant molecules? Do you think it’s like RNA? What do you think it is?

Liam Holt (01:07:02):
Yeah, it’s a good question. Everything depends on length scale as well. We’ve been really focused on this 20 to 50 nanometer length scale because that’s the size of our nanoparticles. At that length scale, my guess is that chromatin and ribonuclear particles that, forming RNAs and so forth are probably dominant. I think there’s a lot of volume taken up by RNA. If you actually calculate the amount of volume that chromatin is supposedly taking up in the nucleuses, it’s not a tremendous fraction. You might imagine that it would be like half of it, but it’s not. It’s like I don’t remember the number but I’m the on the order of like 10%.

Liam Holt (01:07:51):
Nevertheless, I’m sure it’s important. Now if you get down to small things like individual proteins, I think it’s interesting that things like pioneer transcription factors are a very small typically, they don’t feel that kind of crowding very much at all. They can zip through these very, very crowded large complexes and make their way into little tight spaces. Not to say that they’re not interacting with structures or colliding with smaller things on the length scale of proteins. But then, it’s like the difference between trying to get through traffic on a bicycle versus in a truck. Right? I think that’s how biology manages to deal with these perturbations that grind everything to a halt at the 100 nanometers length scale. They still have pathways that can keep things going and adjust the system that are operating at the five nanometer length scale.

Emily Wu (01:08:49):
I have a question.

Liam Holt (01:08:58):
Hi Emily.

Emily (01:09:00):
You showed the KRAS mutation had a big impact changing this, the pressure. I was wondering if you look at the sequence of that, where the mutation is, does it make sense? Like is there a certain structure that’s involved that may drive, driving the difference in pressure?

Liam Holt (01:09:22):
Well, the KRAS molecule is just, it’s a small GTPase that is used to transmit information from growth back to signaling pathways and so forth and make decisions about whether to grow and divide, basically as an actuator of many signaling pathways. The mutation, usually the way that it works is that when it’s in the GDP-bound form, it’s off. And it’s not telling the cell to divide, for example. When it gets positive signals, it converts to the GTP-bound form and then tells cells to do things like divide. And then as the GTPase so it will spontaneously convert back to the GDP form.

Liam Holt (01:10:05):
The mutation that you find in cancer breaks the GTPase functionality. This G12B mutation means that the thing gets loaded with GTP, but then it’s GTPase doesn’t work anymore. And so it stays on all the time. And that means that it’s always telling the cell to divide and a bunch of other things, like it impacts the cytoskeleton, and I kind of alluded to that a little bit. It’s not that the structure of that particular mutation, is doing anything to crowding. It’s more than all of the pathways downstream, so the rate of biogenesis, the rate of cell division, the activity of the cytoskeleton, all of these things are what are ultimately defining the physical properties of the cell.

Emily (01:10:52):
Okay, great. Thank you. Thank you. I have another question. You said to show that the difference when these nanoparticle gets out of the nucleus to the cytoplasm, only happens during mitosis? That model is so unique and it’s very interesting. Have you looked at, so in addition to the chromatin change you observed, is there any other stuff that’s going on? I mean they find it because it looks like a very self restricted event that happening only pushing this out. I’m sure that the physical pressure is one of the reason, but I just feel like there must be something else happening contributing to it.

Liam Holt (01:11:43):
Wait, you’re going to have to repeat that question because I got kicked off Zoom right as you said it. Just go back to the beginning and just asking again for me.

Emily (01:11:50):
Yeah. Sorry, I didn’t notice you froze. When you observed the nanobody that’s coming out of the nucleus to the cytoplasm, and it only happens during mitosis. That’s a very interesting model that you can show a lot of differences. Just not limited to the chromatin change where it’s driven by additional pressure to pushing them out. Have you looked at other things that might be happening together with this chromatin change? Because I can’t imagine that the physical force to push it out is the only reason because… It might be, but there might be other changes.

Emily (01:12:39):
Another thing that comes to me is, because you have tested all these TORC inhibitors and all that. Have you just tried to put them in it together with our model? Because I know rapamycin also regulates proliferation in mitosis. If you put compound in that model, does that change? Because you already showed that it changes the nanoparticle… well, the first part of the presentation.

Liam Holt (01:13:15):
I guess to paraphrase what you were asking: can I relate the first part of the talk to how particles are being excluded from the nucleus? Like do these things somehow talk to one another? That’s a really interesting question. I think we’d have to do the experiment a little bit differently because if we just had the cells with rapamycin they won’t divide. But what we could do, I could imagine is arrest cells in mitosis and then maybe osmotically perturb them to change crowding and you might predict that that would have interesting impacts on chromatin structure.

Liam Holt (01:13:53):
They would then change the physical nature of the decondensing and the clustering of chromosomes. Yeah, that would be interesting to think about. Aside from the physical nature of how dense the chromatin is and how it might physically through the size of the mesh being too large or too small for nanobots to get in, we’re also curious about whether things like the chemical nature of the chromatin versus the nanoparticle could be important. We know that our nanoparticles have a negative surface charge. What if they had a positive surface charge, would they then go into the nucleus? The things that get you into the nucleus are things like very basic sequences, nuclear localization sequences.

Liam Holt (01:14:46):
I don’t know, that would be really interesting to try and so yeah. Well we actually set up to do some of those experiments but we haven’t got there yet. But that’s an interesting question.

Mark (01:14:57):
Great. Maybe one last quick Diana, maybe, one last question.

Diana (01:15:01):
Well you actually kind of touched on my questions. It was basically how to balance between the mesh size in the, how compatible the surface of the nanoparticles is. I was wondering if say you functionalize them with like a small,[crosstalk 01:15:22] HB1 alpha for instance, if you were able to partition them into the heterochromatin in addition to [chrosstalk 01:15:29]

Liam Holt (01:15:30):
That’s a really interesting question. Yeah. Even not thinking about the nuclear exclusion during mytosis, but if you look at the interface and whether particles go and do they go heterochromatin or not. We could certainly do that and we definitely want to do that as well. I think that would be super cool. We know that the behavior of molecules in general depends tremendously on that surface charge. But Poolman had really interesting work showing that negatively charged molecules, typically diffuse way more rapidly, like orders of magnitude, two orders of magnitude, more rapidly than positively charged molecules. That’s in the cytoplasm.

Liam Holt (01:16:11):
And the nucleoplasm, I don’t think we really know that much. And then if we could look through this whole physical diagram of size versus charge versus hydrophobicity and see how things start to segregate. I think that that’s really interesting. And then disorder and how all of that plays a role and how that disorder is modulated in terms of its chemical properties as you post-translational modifications, you can imagine that you’d have dramatic effects on how things partition.

Diana (01:16:50):
Yeah. That’s really interesting. Looking forward to see what comes out.

Mark (01:16:55):
Well, great. I think we’ll wrap up Liam. Wonderful presentation. Great to hear about you, your work. And I bet there will be a lot of other followup questions. What always happens is we go off and we talk about lectures and come back later with more. If that’s okay with you, we’ll come back with some additional thoughts.

Liam Holt (01:17:13):
Absolutely. Yeah. If you either want to email me, if you want to get me back on zoom, if you have questions in person. Either way, I’d love to talk more. Thank you so much for inviting me Mark, it’s been really fun and thank you all for your attention and your really great, your great questions.

Mark (01:17:28):
Thank you. Very good.

Emily (01:17:28):
Thank you so much.

Mark (01:17:30):
See you later

Liam Holt (01:17:33):
Bye. Bye.