So, just following up on this very interesting talk, what would they die of, then, if you

shown at least that the quality control systems were still working fine, what would they die of? Do they start costs? Yes. So, what we can find in cerevisiae, in all cases where we can make conclusions, for example,

if we prevent aggregate formation, the cells live longer. But what we observe is that actually this leads to a stress to the cell, because actually you have all these small seeds of misfollowed proteins that float around in the cytoplasm,

and that then uses stress response. And the first thing, I mentioned it before when I was asked, when you have this prototoxic stress in the cell, you actually downregulate the diffusion barrier, you open

the diffusion barrier, and now the circles start to leak out. So when you don't form an aggregate, you live longer, not because the aggregate is killing you, but you live longer because the misfolded protein are treated like a stress that opens a diffusion barrier and leads to the release of aging factors to the progeny.

Every single time where we have been able to prolongate the lifespan of a yeast by the cell, the DNA circles were leaking into the blood. And each time that we very surgically mutated by single amino acids that lead to detachment

of the circles from the pores, to let them leak in the blood, the cell is long-lived. So far it looks like what they always die off is accumulation of DNA circle. Now how do the DNA circle kill the cell?

Probably by the fact that they do change the nuclear pores. It's the model we are investigating now. So the circles, as they accumulate exponentially, pores accumulate with them exponentially as well. And so all mother cells are full of pores, but as I said, those pores are not classical

pores, they are specialized pores with some subunits gone and others that have joined the pore. And those pores are very good at imports, but very bad at exports. So they are specialized. So we think that this leads to an imbalance that actually kills the cell.

Do you see any change in the reactive oxygen species? We have not touched radioactive... No, not radioactive. We are not attached to ROS so far, because the ROS data is completely ambiguous.

If you prevent respiration, you are short-lived. You don't form ROS and you are short-lived. And you... Oh, in Cervis? You can inhibit certain complexes in the respiration, but still produce ROS.

Yes. But in the mutants where people were showing that they were preventing ROS formation, these mutants were short-lived. It looks like a certain level of ROS is necessary, and too much ROS is for sure deleterious.

You take a mammalian cell, primary cell from the mouse, and you culture it. It's a nest, right? After a while. Stop dividing. It's a nest. Age. Lifespan is short. The reason is because you grow them in atmospheric level of oxygen.

If you decrease the atmospheric level to 2%, then they will live forever. That's a known observation. Yeah, sure. There are many things, if you delete for every fifth gene in the yeast genome, if you delete it, you will live short.

So, do you live short? Why? Because you are sick? No. If you grow a cell outside of the body, in a petri dish, with high oxygen concentration to which it is not made to live with, and that you find them to live short, what

a surprise. Right? Because it's due to reactive oxygen species. Sure. And if you grow yeast cells in H2O2, they live short. Right. And also if you take away… If you grow them at 42 degrees, they live short.

Scavenger away. Scavenger across away, they live shorter. Sure. But many of these things, living short is an easy phenotype, right? Yes. Almost like that.

Do we have any questions for the small GTPase people? I'll ask one. I was curious for… Am I saying your name right? Mayo. Mayo. Mayo. Mayo. Mayo. Mayo. Mayo. Mayo. Mayo. I want to say a little bit more about Bragson. I was wondering, first of all, is it a natural product or a synthetic chemical?

And also, has there been any structure activity relationship? Can you perhaps put a fluorescent probe on it? Yeah, well, it's not a natural product. And we did indeed a SAR study with that. I couldn't show that, obviously, because it's not published and we have some valorization

issue as well. But we worked with analogs on one side and with mutant of the protein on the other side. So we're pretty confident in the binding mode and what is important or not in the molecule. Do you think you could put, say, a fluorescent probe on it somehow and retain its activity?

It might, let me see, it might be possible, but you have to try first, right? But we didn't try that yet. We have some other thing going on, but after the phenotype take place, like click chemistry

kind of thing. But so we are working on those kind of ideas, but we don't have a fluorescent compound, for example, right now. Is it rapidly reversible? I was just wondering because it's kind of an interesting binding mode. I didn't do kinetics on that, but yeah, you just you treat your cells for half an hour,

you just wash for half an hour, like meaning you just change the medium and half an hour later you fix the cell, do the microscopy, and it's almost completely fine. So you can see that it's not as compact at the beginning, but it's pretty good.

You haven't tried to look in real time, though. No, no, no, no. I didn't. That will be interesting. There's lots of things to do, I think, in terms of visualization, real time, tough microscopy, this kind of approach. Yeah, definitely. We have a question here. So I have a question for Nawa and Bruno. So, well, you show the ER2-GALG and the plasma membrane trafficking, which is a classic

pathway, but there are also some report ER2-GALG membrane direct transfer and bypass the GALG. So I wonder how common this is and why you think certain cargo want to do this.

So what's the advantage? You want to bypass the GALG? So do you know about this? I mean, is this under certain conditions, in certain mutants, certain cargos? Yeah, more specialized conditions. So I am not sure that, I mean, what I know is this ER, for instance, plasma membrane

contact site, I think they are used mainly to transport or to exchange lipids, right? I think it's not really used for protein transport. Yeah, I think there are a few reports of proteins that can be also transported, but

I'm not sure how strong the evidence is. I don't know. I think the evidence is very strong for lipids because, I mean, this protein can, I mean, the protein that forms these contact sites, I mean, are transfer, lipid transfer protein, but I don't know.

You know, Katie? And so far there is not RAB-GTP as involved in this. I think so. Paul Snare. Not yet. So you were referring specifically to bypass mechanisms in terms of contact sites.

I don't think it's a lipids. Randy Scheckman had a review article a couple of years ago, so they called it unconventional secretion, which is basically protein secreted, by the way, not going through the Golgi. So those are separate mechanisms? I think it's very specialized.

Yeah, so that's a separate process and proteins are transferred. Yeah, so unconventional secretion mechanisms. And I guess are there RABs involved? I mean, I think this process is very well known in neurons, right? I mean, you have these two papers in eLife recently where you have a lot of

anglicosylated proteins that bypass the Golgi that are transferred to the plasma membrane. I think at least in the dendrite and axon, I think the paper thinks that it's due

to direct transport from the ER to the recycling endosomes and then to the, I think it's a recent paper in Cell or in Neurons about this. No, no, I can't just make a remark that YPT-11 in cerevisiae seems to be associated

with the ER plasma membrane contact sites at the butt cortex. But there is very little known about it's a RAB.

With some YPTs, it's not even clear if they belong to the YPTs or to the next next GTPAs family. So YPT-11 is I think one of them. But maybe, you know, now that if you say it, maybe they are involved in some membrane

process, it will be interesting. Tommy? I think I got confused in your talk. Yes, your talk. Me? Okay. Oh, you? Yeah, no. What a surprise. Tell me. So the constipation that these yeast cells are getting when they have the DNA attached

to the neutral pore that is leading. It's not really constipation, but yeah. So the entry goes, but the exit is difficult. Yeah. So the way I understood it is that if you have this piece of DNA, this plasmid hook

to the nuclear pore, that exacerbates the effects. The cells get sick because they are not getting a proper balance of egress and ingress into the nucleus. Is that correct? Yes. So that's the constipation effect. If you want. If you want. Okay.

Go ahead. So if you don't have that piece of DNA, then what you were saying is that you were connecting the fact that you were getting this trapping of the nuclear pore, depending on the biophysics, let's say, the ceramide, et cetera. Is that the way you were thinking it? Mm-hmm.

So, but under normal conditions, things do move correctly, right? So these are the consequences of your experimental setup of pushing the systems of the partition correctly in the nuclear pores. No. No, no. No, no. In a normal wild-type cells without pushing any things, you will see the pores accumulate

in the aging mother cell, and while she gives birth to daughter cells that have a normal number of pores. So how do you count the nuclear pores? So the way we have been counting nuclear pores so far is mainly by fluorescence intensity.

But that's, okay, the nuclear pore, sorry, the nuclear membrane that is forming in the budding cell, maybe it's just simply sampling by concentration surface, I mean, density of nuclear pores. I know you will get lower number of, mother nuclear pores simply because you have less

nuclear membrane, right? No, so we did, yeah, yeah, okay. I can give you more detail. I can give you more detail. If you let wild-type yeast cells age in which you have labeled specific nuclear poreings, some nuclear poreings will accumulate to, and we can show they are associated

with a pore, and these pores associate to very high levels such that at the end of her lifespan, a yeast mother cell can have more than 10 times more pores than a young cell in the wild-type.

So we go from 100 pores to 1,000 pores, or even more. We have seen cells that are even more. And the asymmetry is maintained. The daughters are born with normal number of pores. If we prevent DNA circle accumulation, there are mutants that prevent DNA circle accumulation, then you don't accumulate pores.

If you detach the pore, the circles from the pores, then you don't accumulate pores. If you promote circle formation, then you accumulate pores. And each time, the lifespan of the cell correlates extremely well with the pore content.

The cells that have a lot of pores die early. The cells that accumulate pores much slower live much longer. What we know is that the circles, when it is at the pore, it recruits acetyltransferase, SAGA, and SAGA acetylates.

And this is something that happens in a normal cell cycle, in interface. And SAGA acetylates a number of nuclear poreings, such now we can mimic acetylation. If we mimic acetylation on these nuclear poreings, we accumulate pores without circles,

and the cells die earlier, as a short lifespan. Those nuclear poreings that we are playing with are involved in controlling the import versus export. In all cases, what we see is an increased import and a decreased export.

That's the deal. Does that clarify? I'm just kidding. That's a joke. That's a joke. David? So I'm wondering, forgive my ignorance on this, but for RAB proteins and ARF proteins,

what are the, in the most primitive eukaryotes, are they present? And what insights, if you look at the very, very simple eukaryotes, can you get about what's the most fundamental mechanism of RAB proteins? It seems like now there's maybe a diversity of function,

but was there, when this family formed, was there some insight from looking at the evolution? Maybe not repeat the question for posterity there. The question that we cannot answer. If we know about evolution of the RAB family,

so from what I remember, there will be... I'm trying to be an ARF, so I'm not... Oh, ARFs. ARF. Basically, there can be one. RABs, there is sort of like the minimal set, with and without duplication,

so there could be less than cerevisia. Does Giardia have RABs and ARFs? Who says... Just one thing. So the first eukaryotic common ancestor had ARFs and RABs.

It had very few ARFs, maybe just one, I'm not sure exactly how many, but very few and more RABs. So it seems to be a conserved feature to have few ARFs and more RABs. And this goes way back to the first eukaryote,

the last common ancestor of eukaryotes. Are they in archaea or in any prokaryotes? Prokaryotes, no. Well, so there's been a discovery, yeah, there's a discovery of a very recent, very recently of an archaeobacterium that's the closest to eukaryotic ancestor that has been identified.

And so there are ARF-related and RAB-related proteins in that archaeobacterium. So that's actually a very interesting path to follow. Yeah, that would be very interesting to follow up. They are different than the eukaryotic ARFs and RABs that we have,

but clearly there's signatures that make them, that put them into that family. So yeah, we don't know anything about this organism because it's only been identified at the genome level. So what would be very interesting is to see what its membrane compartments are like and that will await its identification. We only know about its genome through deep sequencing,

through its identification. Just to finish on this question, so this is highly, I am talking about RAB-CTP, so this highly conserved throughout evolution. And I think now we have genomic data on many, many species.

And I think there is, I mean, you need five or six, I mean, there is a minimal set of five or six RAB or YPT that are enough to sustain the secretary and probably endocytic pathway. So this RAB1 or YPT1, I think YPT3,

YPT6, YPT5, and maybe another one, and SEC4. And seven. I'm not sure about seven. Maybe YPT7.

I think about why not in bacteria probably because they don't have intracellular compartments. They don't need this machinery. But there will be some in bacteria that are parasites of human cells. They actually have sometimes RAB or RAB-interacting proteins

that they took, and you can see it transferred horizontally from hosts many, many years ago.

So not SNAREs as such, no. There are longan domain proteins, so the longan domain is present, and that is present in some SNAREs.

That does suggest exactly that, yes. Well, there's many questions, yes.

Stop this misconception of people saying RABs are involved only in fusion by saying they are involved in all aspects of cellular trafficking, not just fusion. Formation, too. And in the east, is there examples where

one transport step could have more than one YPT input? Yes. So one that I showed is, for example, going from trans-Golgi vesicles to the plasma membrane actually uses two RABs,

YPT3-1 and 3-2, or 3-2, and Sec4. So they share this step. But they share... The idea is that one goes halfway and then the other one takes off. Oh, you mean if they are parallel?

Is that really true, or is that just true for certain types of granules, and then YPT3-1 could take all the way to the, could bypass Sec4 for certain cargo or something? No, overexpression of YPT3-1 does not suppress Sec4.

Okay, still, whatever Sec4 transports could be very important, and lead to lethality. But that doesn't necessarily mean that YPT3-1 could still take some all the way, no?

Well, it doesn't interact also with the specific effectors of Sec4, so I don't think so. Yes. So this in the endocytic cluster of the vesicles, we see after the vesicle buds from the membrane,

we get two waves of two different RABs coming in. One is RAB5, and the other one is RAB35. And they have different, so they're recognizing different lipids, they're having different effectors, and they come, even though they come in the same vesicle, when you trap them, right, they come in waves that are not synchronized,

and they're slightly shifted in time, right? So, I think the RAB5 is very important for homotypic fusion, RAB35, I'm not exactly sure, maybe it's more direct. Fast recycling. No, no, but this is the endocytic code of the vesicle.

So, it's not recycling yet. The first paper on RAB35 was made up by Aloy Schein. No, no, by the way, I don't mean to... Sorry. Go ahead. No, what I'm trying to say is, if you actually track on the vesicle the RAB, right,

they're coming, but that vesicle will actually go to a RAB5 still, right? So, it's... Are they mutually exclusive? No, they're not. Yeah, that's the point. No, but they're looking at different... They're not looking at the same effectors. Well, of course, no, but I think that...

No, I mean, the very first paper on RAB35, the current biology in 2006, I mean, so we measure the internalization of transferring, okay? And, I mean, the message was that RAB35 was involved in very fast recycling of transferring,

even between and before RAB4. So, we don't know if these vesicles that form, I mean, reach the endosomes and then recycle, is what I wanted to say. Well, sorry, we like to counter that, right? So, those experiments are based on, let's say, dominant mutants, right?

Right? No, in SCRNA. At that time, we used SCRNA. All right. So, either you do that or you do a depletion by RNAi, right? Something, right? So, these are slow effects. It takes a while until you build it. I'm trying to make the point that

if you just watch the arrival now of the proteins without perturbation in the system, right? As I said, in the endocytic code, the two proteins are coming, right? Into all the code vessels, right? And it's regardless whether they are transferring or not. So, there's no discrimination between transferring

an EGF receptor or whatever. They don't care, right? The sorting happens later, right? So, I think what I'm trying to also bring into discussion is that we have the experiment that we do where we do the slower perturbations and we get the read on the concepts of that as compared to now the ability to see what's going on

without the perturbation. And there are some differences, right? But this is not incompatible with the fact that RAP-certify could be involved in fast recycling, right? I mean, RAP-certify associates with the vesicle and then... No, I just said that they... I have no issues with that. That's fine.

I was just saying that the vesicle that pinches from the membrane and came in it yet has to decide what to do, right? I mean, is it going to recycle or is it going to re-enter? This vesicle, this very, very first vesicle is going to hit in one first compartment and then there's going to be a sorting, right?

And now I'm going to recycle, I'm going to go a bit deeper, right? But that very first vesicle doesn't know yet what to do. And in spite of that, it's collecting the two RAPs, right? So, right? We were discussing a little bit about... Just a follow-up on this, if you don't mind.

As we were discussing at lunch, I think the likelihood in mammalian cells, given the complexity of the trafficking, is that you're always going to have more than one RAP simultaneously in the same vesicles or most of those

because what they're creating is domains and those domains create different functionalities and you're probably going to need different functionalities sometimes simultaneously or with different kinetics that they're forming and removing and disappearing, etc.

So it's all this combinatorial scheme where the RAPs would be. I think that's the most likely scenario and that's entirely consistent with you seeing more than one RAP coming in even at the vesicle. I just want to confirm, this is not overexpression, yes?

So, I just want to add to the conversation

what could be the role of cargo because we saw the talk of Bruno and you had RAP6 and that could move to the plasma membrane, to the focal ignitions, to melanosomes. We saw in the talk of Nava that you had YPT-1

that could go to the Golgi or to the plus. So is there any connection between the cargo that is in the vesicle that could determine the recruitment of RAPs and that determines the recruitment of effector proteins.

So, thank you for asking this question. This is exactly the way that we are thinking. When I talked about the module, I think that the cargo is the one that will determine which JEF will come, which RAP will come, but also will affect which effector because otherwise that's our model and we are working on it.

And we have some evidence that it's true. So how do you make a carrier in the secretory pathway that only has cargo X and not cargo Y? Well, just to answer your question,

together with Judith, we are not doing it in the secretory pathway because it's easier to do it in autophagy and we have two different types of autophagy depending on the cargo. And these are all not stressed, not under stress, selective, which now require go to different organelles.

And yes, what? But I think the problem in the normal secretory pathway, let's say, you're dealing with smaller carriers. I don't understand how you will end up having a carrier that has only one type of cargo, so it can go here or there.

No, she's… OK, so that's the problem. I think, as you know, you have a few examples of a direct interaction between RAB and the cargo. I mean, for instance, between RAB-11 and Beta-Adenergic Receptor and this drive, probably the recycling of Beta-Adenergic Receptor

to the plasma membrane. So this is known for this. I think there is also the example of RAB-21, it's a work by Joanna Ivesca, and I think it involves RAB-21 for recycling of a pool of integrins.

It's not the bulk, but you're doing the sorting. No, we talk about… Yeah, especially… And also, actually, I suppose there was a report, I think, from Keith Mostoff a long time ago that RAB-3 was directly interacting with the poly-HGA.

For transcytosis. Yeah, because I think, when I was working on the hops, I thought, OK, it's on the bicycle and it's there,

but I realized that it's much too static a picture. It's probably a continuum of recruiting RABs or hops or tethers. And then, in the end, like Miguel is saying, it will form domains or, depending on the cargo,

one of these factors that is recruiting may kind of win. And then you get that this physical recruits these effector proteins and it makes a kind of decision that, OK, so I should go left or right.

But I can imagine that cargo could be very determining in that. And also sorting is not 100%. So you may take some cargo to the wrong side, but I think the majority of cargo may then influence what is happening.

It's just a form of food. OK. Oh, yes. I have a question for Ilbar. He's good. Have you found some other proteins, a part of RAB-3,

which I believe is linked to memory on aging? Yes. So proteins with this type of domains, PolyQ particularly, there's a full study that had been done from the lab of Sue Lindquist. And the yeast genome contains at least 150 by the threshold they used.

And actually, if you loosen a little, not much per bit, the threshold, it goes rapidly to 200, 250 proteins. And actually, we needed to loosen a little bit their criteria to get to E3. E3 was not in their 150 top. So there are way more.

And we know of at least five more proteins that are actually aggregating in response to pheromone and that are involved in the phenotype that you saw. So it's not just E3. E3 is a key one, but it's not alone. And we are now characterizing a few more that respond to heat shock.

For example, that are necessary for stress memory. Described a long time ago by a pastor, actually. It's called Acquired Thermotorents. And we have some hints about also some proteins

with this type of domains involved in toxin response. Are there some protein aggregation linked to memory and aging in those organisms?

So a protein aggregation has been a classical hallmark of aging cells. But generally, nobody knows what are in those aggregates. We know in terms of memory that the Kendall and Lindquist lab had worked on at least one other protein that aggregates in response to stimulation in the synapse,

aggregate in response to stimulation of the synapse, and is required for long-term potentiation of the synapse. So directly involving the memory. And there is the same protein in drosophila is involved in the memory of courtship.

Our males, flies, have to learn courtship. They do this by a protein that aggregate with a similar aggregation domain. And that protein happens to be the closest homolog to V3. We have a question here. So the question from a computer person to you. So you said the cell has to distinguish DNA cells

from non-cells, basically compared to very long strings. And there are a lot of computer algorithms for this. So are there any kind of them being used? For example, one each can be used to compare two signatures, cryptographic signatures, or two strings?

So we think so far from all what we have been able to do, it's completely sequence independent. So we don't think that they are recognizing any of the sequence. So what we think they recognize is in cerevisiae, it's whether they have a centromere or not.

We simply need to add these 150 nucleotides to any piece of DNA. And now it will condense during mitosis. And in Bombay, it's simply the history. The fact that you saw that it wasn't in the chromosome was sufficient.