 Okay, so I'd like to begin by thanking Nawa and other organizers for inviting me You know, it's my great pleasure to come here and To speak in this really fascinating meeting and also take a break from my teaching in Colorado so Now this diagram shows you a membrane protein on the cell surface Now these protein are the windows and the doors of the cell Now this protein include receptors Channels and the transporters and they allow the cell to communicate with the outside environment Now as you can imagine the surface levels of these proteins must be precisely controlled now too much of a protein or too little of a protein Can often disrupt the physiological pathway associated with the protein and in humans this may often cause diseases So here I can give you one example many of you already know Now if you have two high levels activate EGR receptors on the cell surface This is going to cause uncontrolled proliferation of the cells and this may cause cancer Now on the other hand if you have a two low level or what we'll call transporter as you're going to see a moment This is going to disrupt the blue coat blood glucose homeostasis and in humans. This may cause type 2 diabetes Now how is the surface level protein is a determined? It's a determined by the balance of exocytosis and the endocytosis Now exocytosis is a Vesco fusion event which involves the fusion or exocytic Vesco with a plasma membrane and this delivers the cargo to the plaza membrane Indocytosis on the other hand removes the cargo from the plaza membrane and the returns it to the endocytic compartments Now remarkably both exocytosis and the endocytosis can be regulated by stimulus and And in this way the surface level of a membrane protein Can be adjusted acutely according to a physiological demands and here stimulus could be a second messenger. It could be a false relations Now a major model system we have been using to study a surface of protein homeostasis Is the insulin controlled for translocation? Now glu4 is quite an Euroglucal transporter Because under the basal condition glu4 is sequestered in intracellular compartments So it's not on the plaza membrane Now this is because your blood sugar level is normal and you don't need to uptake a glucose and At this basal stage the exocytosis and the endocytosis is at equilibrium So now imagine you have food So your blood sugar goes up and this triggers insulin secretion from the pancreas Which then travel in the bloodstream and reach the target tissues mostly adipocytes and the skeletal muscles Now when you insulin bind to the receptor on the plaza membrane this activates insulin signaling cascade Which eventually leads to the activation of exocytosis at the same time the endocytosis Is moderately reduced and the result is a translocation of glu4 protein from vesicle to the cell surface and This is a translocation allows the transporter to uptake a glucose into the cell either burning into ATP or storage Now this is your blood glucose level after you have food that's at time zero is the time you have meal Now you can see your blood glucose level quickly goes up But also it rapidly returns to the normal range and this Returning to the normal range is the driven by insulin stimulate glu4 translocation and the one your blood glucose returns to the normal range insulin signaling terminates and this also shuts down exocytosis at the same time Endocytosis resumes and this leads to the retrieval of glu4 from cell surface and the returning them to the vesicles Now you can see insulin control the surface level of glu4 Transporters by regulating exocytosis and the endocytosis now as much as we Love glu4 exocytosis the major goal we studied is to use as a paradigm to Establish the general principles of surface protein homeostasis For example, we expect our findings can be extended to understand the trafficking of neurotransmitter receptors in the neurons The transport of water channels in the kidney epithelial cells and the trafficking of immune receptors in the immune system Now today, I'm going to tell you two stories on the regulation of surface protein homeostasis I'm going to begin by a short story on the regulation of rap gtps by a protein called ribaf in the exocytosis Then I'm going to focus on a very recent finding on the regulation ap2 adapter formation by a gap in the endocytosis So I'll begin with exocytosis Exocytosis exocytosis as I mentioned is a vesicle fusion event Now this is after a rab and a tether bring the vesicle and the target of membrane together And the vesicle fusion is driven by two classes of molecules snares and the second one monkey team as I'm protein Now snares are membrane anchored proteins and you have a vesicle anchored snare kind of we snare and The target anchored snare called the t snares When the we snare and the t snare see each other they spontaneously assemble into a so-called trans snare complex And the trans snare assembly proceeds toward the membrane so you can imagine this is like a a zepering process Which forces the two membrane into a closed proximity to fuse Now SM protein is a soluble protein it can bind to the transnare complex and accelerate the transnare assembly Now snares and SM protein they're the core engine for vesicle fusion In addition to this core engine good for exocytosis also requires specialized regular regulatory factors such as the Sinep, Thomason and the Dr. B Our previous work focused on biochemical dissection of these molecules since the most of the work has already been published Here I'm not going to our details And from from here I like to mention this protein I believe Only represent a minor fraction of the entire regulatory network for gluophore exocytosis Now the sequencing of a human genome predicted a large number of membrane proteins And for the majority of these membrane protein we know virtually nothing. So therefore I think the most important Direction in the field is to comprehensively identify new regulators for gluophore exocytosis Our request coincided with the one the CRISPR-Cas9 genome editing So we decided to perform unbiased genome-wide CRISPR screens So we took advantage of existing genome-wide CRISPR libraries, but also we make our own customer CRISPR libraries in this CRISPR libraries So these are all lentil virus based you have a single guide array and Also the Cas9 the single guide array recruits Cas9 protein To specific location in the genome to introduce lots of function mutations and the two perform screens we Use a HAJP gluophore reporter Now this reporter has a HA epitope inserted into actual cellular domain Now HA staining as you can see here Can stain the surface gluophore reporters On the other hand the GIP tells you the total protein in the cell now if we calculate HA staining and GIP staining This will give you the information about a relative amount of our gluophore reporter on the cell surface and we start a stable cell line expressing this gluophore reporter and we mutagenize using the CRISPR-Cas9 library and then we treat the cell with insulin And then we use the flow cytometry to sort for the cells with the loglophore on the cell surface And then we repeat the process for three times and then we took the sorted population and Recovered the single guide arrays and use the deep sequencing to analyze the abundance of the single guide arrays Then we use the bioinformatic tools to calculate the significant hits and finally we performed secondary screens and the individual validations So we perform a screen in multiple cell lines. We start proof-of-concept screen in the cancer cell lines, but eventually we focus on myoblast and pre-dipsense So very briefly our screen Identified a large number of single guide arrays that were significantly enriched compared to the control population So we then use this equation to calculate the abundance of the single guide arrays So we can derive significant hits from the screens While the screens identify a larger number of genes, the majority of the hits were not previously linked to gluophore exocytosis But we did recover known regulators such as Rep10, exocyst and almost the entire insulin signaling cascade So here I'm going to focus on one new factor identified in the screen called the Rabbi F Now Rabbi F, Rabbi interacting factor. It's a very small protein. It's a 14kd soluble protein Now when I did a sequence analysis The only prediction was it's a putative guanine nucleotide exchange factor for wraps or gap for wraps now now why you're probably gonna Well, I think I remember you told me you didn't believe Rabbi F was a gap from very beginning But I have to say it's kind of ahead of the field because if you search literature and Search the motifs This is the only information you got the entire protein is a ribo gap of course a putative gap Now what's a gap? RAP GTP is involved in the Weisgau tailoring and the RAP is a G protein and the RAP cycles between GTP and the GDP bond form and the GTP bond form is active and What the gap does is to promote GTP binding. So gap is a positive is a positive regulator for RAP Now the story of Rabbi F begin 25 years ago with the two papers published by Peter Novick and the Pietro decommedians groups So their papers were mostly based on in vitro biochemical acids and actually RAP F is also known as MS4 or DSS4 This is the first putative gap for RAP GTPs. Let's think about that. There are about 70 RAP GTPs in the human genome and this protein is the first putative gap But the surprisingly that case later is a biologic function was still unknown Next we use the CRISPR-Cas9 knockout ribo F in deep sets and In the wild type cell here, you can see insulin strongly promote gulfo reporter translocation to the cell surface So all these are flow cytometry based Analysis But in the RAP F knockout cells, you can see the translocation was largely abolished and this is confirmed the screen data Now these are exocytosis I say if you look at the slope of the curves This represent the speed of exocytosis and you can see here in the RAP F knockout cells Exocytosis was strongly reduced and we could fully rescue the exocytosis by introducing the wild type RAP F gene and these are confocal images images which also confirm the flow cytometry data in the wild type cells you can see an insulin Relocate gulfo reporter from intracellular compartment to the cell surface so I want to point out these are Lepid droplets or a deep sets not the nucleus and here is the the boundary of the cells In contrast in the RAP F knockout cells you can see a insulin stimulated gulfo translocation was a largely abolished and Here I want to make an important the point is Discovery of the first the physiologic function of RAP F finally makes it possible to explore its medical mechanism Now since the RAP F is known as a RAP binding protein the first question we asked what's a RAP target? now if we Look at the screen again RAP F was identified in the same screen as the RAP 10, which is a non-regulator over Gulfo exocytosis so we speculated maybe RAP F controls Gulfo exocytosis by binding to RAP 10 and First we are prepared the recombinant protein and the test if these two protein bind to each other and This is the in vitro liposome called flotation assay and Here we anchored RAP 10 protein in the proteal liposomes Then we add a soluble RAP F to the liposome And after centrifugation the liposome migrate to the top over the gradient Together with the bonded protein Then we collect this fraction and analyze by STS page Okay, so here just want to you to focus on this link you can see a RAP 10 protein could bind to RAP F in the co-flotation assay and the binding appeared to be a stoichiometric By contrast protein-free liposome could not bind any RAP F protein So RAP F bind to RAP 10 directly And since RAP F is known as a putative gap we ask can RAP F function as a RAP 10 gap The structure of RAP F with a RAP 10 is still not known But we do have the structure or the complex with a RAP 8 Which is a kind of related RAP now based on this structure we We found that there are several residues at the binding interface So therefore we introduce the mutations either single island mutation or triple mutations to disrupt the interaction And this is the in vitro GDP release assay, which is commonly used to analyze RAP gaps Now without the RAP F you can see almost no change But when we added a wild type RAP F, you can see a rapid GDP release And when we introduced the mutant either a single mutant or the Triple island mutant you can see the GDP release rate was strongly reduced Now if you look at the initial rate of the triple mutation We estimate about a 24 the job in the In the GDP release So here came out the surprise when we introduced these mutants into a RAP F knockout cells, we found that these Gaff mutants Can rescue the group for access to the same level as the wild type gene Okay, and we repeated experiments in the multiple cell types several different conditions We always got the same results And these these data strongly suggest RAP F is actually not a gap you include for access atosis Now you put if a RAP F is not a gap. What's the biologic function? And then here came another surprise when we look at a RAP F Rep 10 expression in ribbon right back for knockout cells We found that the protein disappeared Okay, and we could rescue RAP 10 expression using either the wild type or the two mutant RAP F so The rescue of RAP 10 expression Correlate with the ability of this protein to rescue group for access atosis and Has nothing to do with their in vitro gap activities So therefore we propose RAP F functions by stabilizing RAP 10 protein in the group for access atosis Now to test this possibility we Also tried Lentil viral expression of RAP 10 in the knockout cells Now this work is important because we had to rule out right by if it doesn't control the transcription or epigenetic regulation RAP 10 expression so here we We found we could write it express RAP 10 in the wild type cells but very little protein was observed in the RAP F knockout cells and this data are consistent with a The previous data just showed you and the supported idea right by F stabilized RAP 10 protein and the one we treated the right by F knockout cells with MG 132 which is a proteasome inhibitor We found we could fully rescue RAP 10 expression and When we treat the MG the cells of is a both MG 132 and the secular hexamide Which is a protein translation inhibitor the rescue was abolished now this data suggest without a right by F RAP 10 protein could be Efficiently produced but they're quickly degraded without the function of a right by F Now here we proposed two models It's possible right by F stabilized Intrinsically unstable RAP 10 intermediate. So therefore it's more like a chaperone On the other hand, it's also possible right by F can protect the folded RAP 10 protein from proteasome degradation For example by masking a so-called a diagrams To distinguish between these two models we are reconstituted RAP 10 the right by F into E. Coli cells E. Coli as you know doesn't have any reps or right by F so we found RAP 10 protein could be produced in the E. Coli without a right by F But all the protein end up in the insoluble pellets. We couldn't get any protein from the supernatant we could Extract soluble RAP 10 protein only when right by F was co-expressed so therefore RAP F stabilize unstable RAP 10 intermediate and without RAP F RAP 10 is going to aggregate in E. Coli and then miss fold and get a degraded by the proteasome in the mammalian cells and This supports the idea right by F is a medical chaperone for RAP 10 not a gap So here I call it a HODIS chaperone because it's a different from the classical chaperones like HSV 70, HSV 90 both require ATP dependent cycles But RAP F is a small protein. It's not ATPS Instead it recognize the substrate and stabilize the protein which promote its folding and And then we perform the myspac analysis that we identified two additional RAP protein RAP 13 RAP 8 Also regulated by RAP F and this is our model RAP F as a chaperone they interact the substrate and stabilize These proteins which promote the folding of these RAP GTPs into the native confirmation So therefore they can promote exocytosis and without RAP F The RAPs cannot adopt native confirmation. They get degraded by the proteasome Brief summary we identified the first biological function of RAP F which enables us to Discovered unexpected function of RAP F in the GTP RAP GTPs regulation Because all the RAP GTPs function in the similar way We expect maybe this whole this chaperone model can also be applied to other RAP GTPs Okay, let's return to this diagram The surface level of membrane proteins Determined by the balance of exocytosis and the indocytosis now having talked about exocytosis I'm going to focus on indocytosis Yeah, go ahead. I mean Is this a chaperone for RAPs on the steady state conditions or Do cells have to go through something for this chaperone to be Not in group for exocytosis, but I think that's an interesting point Let's come back to yeast. This protein is conserved in yeast I was cloned first in the yeast, but when you knock out the protein, there is no phenotype So at a normal condition, it's not essential for yeast But what I imagine is a maybe a durian stress condition. This protein is recruited To a traffic cause a trafficking pathway and they become essential, but in group for exocytosis You always need it Does it bind to both GTP bounders? Yeah So we didn't address that but according to the non-structure with a RAP 8 It's actually bind to the nucleotide free form Which I think suggests it bind to the more upstream over the folding pathway And so far I don't think there is evidence showing it bind to the nucleotide bond form of RAP GTPs Yeah, is it a chaperone or a co-chaperone? So at least they doesn't seem to bind ATP No, no, it's a very small protein So what I can say is that this protein form direct complex and it's about a one-to-one ratio and If there is any additional protein bound to the complex, we don't know yet On this universe University of this I have a lot of questions about that First these are I presume these three are the only rams you've picked up Right, so that's right. At least in the cell type we studied these three rams are only are only proteins controlled by RABF So question is are there other isoforms of RABF in cells that could explain the other ones and just a comment rams are very nicely expressed in a coli in generally and And in my experience RABF is the only one that really is difficult I haven't tested right or or or or or are 13 But RABF of all the rams that I expressed in the coli was the only one that was very difficult to express so my my Expectation would be that this is a very specialized thing for these three rams that you're writing there Or and and the question then becomes what's special about the structure, right? So I I think I can I can tell you something So why I show this work to a colleague in the group for field and he said finally There is explanation why they couldn't produce RABF 10 in E. Coli They had to produce in the inside the cells and the comeback to your question is anything special about these three proteins and You know, I talked to people in the field and there is no reason to believe they're special And so you said that in other rams could be expressed readily in the E. Coli But that doesn't really rule out that there is a bacteria chaperones that perform similar functions promoted folding But again, it's all open question Can I just comment first of all I didn't say I don't believe because I never said I don't believe I don't think I don't believe Not in science I just I don't think What you only say that in writing I Don't think especially when when I have data Because it was only shown to that GDP can be released but not GDP can be Dequired and like other Jeff's and here I want to just make a suggestion that maybe what's Similar in these three is maybe just about simple biochemistry of Nucleotide affinity there affinity to nucleotides and Nucleotide free is known to be Not stable a state of all GTP as so maybe these three but that's also That's a true for all the RABGTPS, right? So I'm Nucleotide free RABGTPS unstable, right? I know but I'm saying maybe these three Have lower affinity than all the other RABGTPS to Look low types and maybe that's why they need this Agree, I think they're certainly many question need to be addressed. Yeah, yeah So related to that if you get rid of the gap AS 160 for So as 160 is a gap. Yeah, it's a gap You mean repeat experiments in the gap Okay, yeah, yeah, that's a good idea we haven't tested so I could be next step Yes Okay, so here we perform a similar screen and It's a largely the same but I want to point out the difference we so here we didn't treat of the cell with insulin we We sort for the cell the way the high group for reporter on the plasma membrane without a steam stimulation And the idea here is we're going to recover the mutant cells defective in the endocytosis So the screen recovered known regulators of endocytosis such as AP to Adapter subunits. I'm going to come back shortly and TPC 1d4 is this is actually a inhibitor for Exocytosis, so not only the screen recovered endocytic regulators. They also can recover inhibitors or exocytosis But again, most of the genes were not previously known to be involved in the group for endocytosis So here I'm going to focus on one gene called a gap So before that I want to mention the screen recovered subunits our AP to adapter complex Now AP to adapter is a tetramer two large subunits alpha and beta and The one medium subunit mu and the one small subunit of Sigma Now our screen identified MP AP to S1 which encode the Sigma subunit We also identify AP to M1 which encode the mule subunit So we didn't recover the alpha subunit, which is actually expected because there are two genes AP to A1 and AP AP to A2 genetic screens cannot recover redundant genes and The a the beta subunit the same story The AP to B1 encodes the beta subunit, but the gene can be fully rescued I would say compensated by AP one B1 So we couldn't recover alpha and the beta, but the idea is fairly clear. We recovered AP to adapter subunits Now what's a gap now if you ask someone in the field nobody have heard about this protein It's a 34 KD soluble protein. That's why some people you've called it P34 Alpha and the gamma adapting binding protein. I'm going to come back to the naming shortly Now this is known as the IRC 6p in East, but the no phenotype in East I like to mention a gap is frequently mutated in the human disease called a punctate pommel planter Kyrtodoma PPK now. This is the high-flow insufficiency mutations Which means just one allele mutation given the phenotype we see if both alleles are mutated Apparently the organism cannot survive Now this disease is characterized by the thickening of the palm and the soul skin and Of course the cause of the disease By a gap of mutation is still unclear and this is because we still don't know how a gap works Again, we knock out a gap in adipose using CRISPR-Cas9 and you can see if you focus on the orange bars It cause constitutive Translocation of the reporter to the self-surface. So therefore insulin regulation is disrupted The reporter go to the self-surface even without stimulation and We confirm the flow cytometry data by confocal imaging Again, these are well-tap cells well-tap cells before mostly in the intracellular compartment But if you look at a gap of knockout cells, it's already on the self-surface and the phenotype is very similar to AP2S1 knockout cells And AP2S1 as you remember encodes a subunit of the AP2 adapter And next we examined The indocytosis of a cargo. So here is a measurement of indocytosis or the fluorescent labeled antibody Recognizing group 4. Now, this is a very commonly used as say for indocytosis in the field Now when we knock out a gap, you can see a indocytosis of the cargo was abolished and the phenotype was at least as severe as AP2S1 knockouts and the transfer receptor it's a classic Indocytic cargo of a class or media the indocytosis when we knock out a gap There was a strong accumulation of the protein on the self-surface and the phenotype was a fairly similar to AP2S1 knockout Now this data clearly showed a gap is essential to class or media the indocytosis Now I said virtually nothing is known about this protein the only hint that came from East to hybrid screens I think probably two decades ago Now Scotty Robinson's group found a gap Combined to alpha subunit in East to hybrid However, we still don't know if this binding is a direct interaction or if this interaction is a biological significance and next we prepare the recombinant protein and the test the direct interaction and This protein from E. Coli we found a GST tag alpha subunit indeed Interact with a gap and the GST itself couldn't and the interaction here appeared to be a stoichiometric Now how does a gap regulate AP2 adapter formation? There's just some background light But hopefully you can see in the wild type cells on a self surface. There are abandoned AP to a thank you AP to a punk to on the surface, but the strikingly When we knock out a gap this punk to largely disappeared And so this is accompanied by a translocation of group for a reporter from interstellar the compartment to the south surface and we confirm this confocal data by a Turf microscopy Which monitor the events near the plaza membrane? But so here again in the wild type cells we observed a large number of AP to punk to But they all disappeared in a gap of knockout cells And here I can imagine two possibilities Maybe a gap is important for surface recruitment of AP to adapter or They could be essential for the stability of AP to adapter Well, it turned out the second possibility was correct When we knock out a gap, you can see the alpha subunit largely disappeared At the same time the beta subunit was strongly reduced and the male subunit Was also abolished our antibody didn't work for Sigma so far But you can get the idea without a gap AP to adapters were degraded And of course, this is a highly reminiscent of a web app story. I just told you so therefore We are speculated maybe a gap also stabilized The AP to adapters since the bind to our sub subunit. Maybe it's a shaper for alpha subunit and Again, we recast it a gap and alpha subunit in E. Coli Which doesn't have rap which doesn't have a gap which doesn't have class remediated in those hypothesis Again, so here if you look at the pallet, this is after extraction which represent the insoluble fraction We could express the alpha subunit without a gap But all the protein end up in the pallet. We couldn't extract any protein in the supernet However, when we co-express a gap now we could extract the protein in the supernet and this again is similar to web app in the web 10 story, which suggests the alpha subunit itself is unstable in bacteria it's a Informal aggregates and in mammalian cells. It's recognized as a means for the protein and the degraded On the other hand a gap combined to offer subunit and the stabilized Therefore by coincidence, we think a gap is also hold this shaper on for alpha subunit And next we look at a subset of subsettler localization of the protein in the So this is the alpha subunit on the plaza membrane. You can see the punkta but a gap by contrast shows a diffusive pattern and Which which is a typical of cytoplasm Which leads to our hypothesis a gap Actually regulates upstream event in AP to adaptor formation And it does not follow AP to adaptor to the plaza membrane Now one prediction from the shaper model is a a gap function should be independent of its location in the cell To test this possibility we tether a gap to the ER surface by fusing it to a ER membrane protein cut ATF-6 And This is going to target the protein to the ER surface So this is a western blot you can see the fusion protein expressed at the similar level as a gap and there was a no degradation and I Think it's quite an interesting way. We found this fusion protein Could To a large degree restored transfer receptor indocytosis even it's not as close as a gap rescue But it clearly showed this a ER anchored protein can function. So therefore ER well a gap function is not dependent on its localization Which is consistent with a shaper model Now, this is the crucial test a Crucial prediction from the shaper model is forced expression of the alpha subunit should rescue a gap knockout phenotype and here we took the advantage of very strong promoter wires to derive the promoter and when we all expressed alpha subunit I Could fully rescue the transferrin receptor In internalization and as good as a gap rescue and The base subunit rescue our expression didn't work and I expected expression order for AP to subunits could rescue So I think we can do two key conclusions from this data The first is restoring AP to expression artificially Can bypass requirement for a gap and All these data indeed suggest a gap is a whole days of shaper on for the upper subunit But I think a gap is more than just a shaper on so I'm going to show you some data and Look forward to your comments So what happens after a gap bind to upper subunit? To monitor a gap interaction with AP to subunits We divide the HA tag system In which HA tag is added to all the four subunits of the AP to subunits and You can see this is sigma HA this is mu HA and alpha and the beta They migrate at the same location in sts page Okay, so here is the co IP data We found a gap it could have put on alpha subunit as expected, but it could not put on the sigma subunit by itself However, when we put in all the four subunits a gap could have put on alpha subunit and the sigma subunit We couldn't we never see the mu subunit And also we perform additional analysis different the co IP Design using an untagged version of alpha subunit. We could only see Afra and Sigma subunits. We never could see a beta or mu subunits And if we confirm this co IP data by using a recombinant protein, so here I just want to see this lane We can find so we saw the sigma subunit bind to Alpha subunit and a gap. It's very small protein. So the staining is a very faint, but I think it's about a stoichiometric interaction So they're therefore the a gap alpha subunit dimer Subsequent they recruit the sigma subunit. So I like to put what I talked about so far into this model We think a AP to adapter assembly is not a spontaneous process as previously assumed instead It's a highly organized process the initiator of the process is a dimmer or a gap and offer subunit and The dimmer here recruits the sigma subunit Since beta and the mu subunit does not bind to a gap We propose they can replace a gap from the alpha and the sigma subunit Which allows the formation of the tetramer. So what happens after is well known So the story I just told you it's a prequel of what the story that has been told again again and What happens of without a gap the entire AP to assemble assembly process Just collapse all the subunits got degraded Well in summary We showed a gap is a master regulator of our AP to adapter assembly In addition as a shaperone for offer subunit we propose maybe a gap can prevent AP to binding to non-cognitive the cytosolic proteins in our many cytosolic protein contain the dilucine Signals, but they should not be recognized by AP to adapters Now if for the AP to adapter is a fully formed The beta subunit can block the binding site Unless until it's opened up on the plaza membrane But what happens before beta can mask the binding sites and we propose a Gap may be functioned by blocking the binding sites. So to prevent the binding to the non-cognitive proteins Now because other modern marriage a track trafficking factors like AP one AP three cop one Face the same challenges of assembly and the specificity issues We think it may be there are other a gap like molecules that control their assembly Now before I finish I like to return to the PPK disease Now here is a diagram of the skin cell contact the contact between skin cells are driven by a series of membrane proteins and the cell the solid proteins Which include a decimal glean decimal plaking dp here and the keratin interestingly Most of these proteins are also found to be mutated in the PPK disease Along with a gap now the mutations of these genes make a total sense because you can imagine these their mutations are going to Compromise in the integrity of the skin cell contact and therefore cause the skin disease But how does a gap fit into the picture? So here I just want to show you a preliminary observation we did a proteomic analysis of Surface of proteins in a gap knockout cells and we found a gap can Mutations downregulate a protein called a decimal plaking So without a gap this protein level on the surface was strongly reduced now This is a very preliminary data Additional work need to be done, but I think it has a potential to explain Why a gap of mutations can cause the skin disease? So before I rip up I like to convey the message There are still many mysteries in the mysteries in the Indocytosis and exocytosis None need to be solved Our screen identified many membrane proteins which show the very strong knockout phenotype But the virtually nothing is known about their biologic functions now if you do a sequence prediction you still got nothing In my opinion moving forward a very exciting direction is to find out how these a protein regulate cargo flow in the cell Okay, so with that I Like to thank the people Then Lauren great students hijab postdoc. So they did all the work. I show you today We return to this and thank you for your attention So this is very cool that you're regulating a P2 Assembly, but could you tell us why it's not legal? Yeah, so I Know class through a classroom is a list of because of the trafficking from the Golgi. It's actually Cause more lethality Actually, you can knock out the AP to and they're not a list of insert in cells It's lethal in you eliminate a P2 and it's lethal in the cell culture In animals I can understand Elimination is lethal in animals and if you eliminate Yeah animals for sure. Yeah, why this is so important for the Folding of the P2 The only phenotype you're getting as you said, well, that's a high-poly insufficiency, right? So just one a little mutation. So now remember I study a Westco fusion protein monkey team for many years high-poly insufficiency of mutation of monkey team one only cause Certain neuronal phenotype or all the patients are normal. Okay, so I think Why skin cells are more sensitive? Well, maybe some mathematic modeling I cannot tell so but overall I don't think other tissues are gonna be a sensitive and again, I Agree knock out to a AP to in animals gonna kill the animals, but in the cell culture levels I don't think there are essential genes Well, if you don't have a P2 and so I mean cells divide with no AP to they're very sick Anise's problems. I mean lots of issues. So I think you could be right. I think you are right For these clonal CRISPR knockout cells Eventually we select for the mutants that can grow right so it's it could be Compensation it could well, it's not gonna be phenotypic a compensation, right? Or could be a hypo morphic mutations Reduce the most of the activity, but there's some residual activity that keep the cells alive Yeah, well Well, when we knock out essential genes in the proteasome the cells can grow better sometimes No, but if you eliminated to So this actually can be This can be precessively check. There is a essential gene this We check a couple years ago with I don't think that these are essential genes Maybe surprisingly at least not in the cancer cell lines. Yeah Yeah, yeah, there are several screens Well, I cannot say for normal cells, but at least in cancer cells. They are not essential genes Yeah, it's a comment about the rabbeff story So if I understand you thought it's a sharp wrong if you remove it Rob tennis degree that rabbeff But I'm not so sure that this Excludes the fact that the real function could be something else No, it because it's common to find situations in which you have a diamond which both Proteins have Can I just check this question? Do you check it on indulgences or express proteins because the shopper all it would be different the reptile, right? Yeah, the reptile is all endogenous. Okay, and Also, we could For six press the reptile using a viral promoter. We restore reptile protein We totally rescue rabbeff knock out phenotype. So I think your question has a two parts first I think it's a fairly clear in our case. It's a chaperone, but there's a function In another way, you know, could be a gap In a different possible. I don't think I guess but in that case, I think it's a possible Yeah, it's good. Is it completely clear that it doesn't help Some I would call it we have function majority function on the activity around 10 But then of course it also stabilizes Because in diamond as this we did so in general this is the case for other diners I Think that's a totally possible for example a gap Even to chaperone, but I think as I showed it's more than that It's organized the whole assembly possibly maybe a rabbeff is doing that also Stabilize right, but also control downstream function. So right now we just don't have the evidence So quick follow-up on this and then I have another question Does it work on second form? Second second. Yeah, okay the east. Yeah, well the east I Remember if you knock out a rabbeff or homologue, you don't see anything, right? you see a little bit of effect only when you have a Rabbeff mutation that you overexpress is to rescue I think that only demonstrate those protein interacts, right? So we haven't tested in the east yet But I think you yeah, I don't see how you can test a function without a knock out phenotype It's kind of hard. So the other question was on the very last part of what you're saying about the phenotype I was confused how down regulation that there's no blacking would have anything to do with it right, so I Think that this is going to be an indirect effect We did a proteomics and knock out a gap cause a massive Dispersion Inbalances on the surface of proteins. So a large number of protein either you get too much or you get too less I think are going to be for example, maybe a cargo adapter on the plus membrane which is Reduced but in turn you get a moral another protein because of a lack of inhibition So I think it's going to be really hard to interpret at this moment. We don't know why It causes the disease. I just one possibility and I think I can likely can be a indirect effect For it all completely independent Yeah, so in that case, I would think we probably want to rescue a p2 Undercheck it, right? So Your phenotype which which has to do with critical site dysfunction Precise a patient Yeah, so that's going to be a crucial experiment that to study the disease we're thinking To introduce the precisely the same mutation into a PS out and the difference into into a keratinocytes I think That's going to tell us a lot about this protein function It could very well be that your protein Whatever the name is a gap He's doing what it's doing and it is involved in cell cycle specific cell junction disassembly And that is being affected and therefore what you end up with is a cell that does not Disassemble desmosomes when they have to divide and as a result what you end up with is creating a Pethenotype that is like that cell right to work, but you're still thinking the The more directed causes the endocytosis defect, right? Yeah, I agree So what happens after a p2 week? We cannot say for sure. It's just a speculation Okay, so we can choose to small comment question, but yeah, have you checked the mutant that caused the phenotype That's right, so we So many of the mutations actually are premature terminations Right around the middle and I see terminus. We didn't have with a precess mutation, but we have done a truncation which may make The disease in mutation. So that one was a totally inactive. So we saw in the cytosis of the fact So in our assay you need the entire protein How extensive it is this notion of how many chaperones to help So folding right because I mean the same is true with HSC Yeah, and so well Yeah, yes, so what we can we can talk more, but I think we're actually touching even Much a bigger problem for multiple subunits like AP2 or snares, you know, they are expressed at a different Chrome some location different promoters. How do you coordinate their assembly, right? Because they're expressed at different levels. I think maybe this is a one solution for that you have a initiator from one subunit and Just go from there and everything else If they're expressed more or less, it doesn't matter. They if too much they're gonna get degraded If we're too less, they're gonna wait for initiator to get ready. So I think this may be a general way to assemble Multimeric proteins in the mammalian cells How sensitive? But at least in a gap case is very sensitive because without a gap you don't get any AP2 adapters What's your second question is other substrates? Yeah, so I we have some preliminary data actually AP1 the gamma subunit also disappears with a gap knockout But the Delta Adapting was normal. So AP3 was fine, but AP2 and AP1 They were affected So All right, thank you