 Today I will be talking about something that is actually very new and dear to my heart because it's my PhD project. And I haven't talked about especially the production and the purification of KDPF-ABC in a long time. And I want to start by saying so while today we actually have several of those very pretty structures of the whole KDPF-ABC complex. When I started this protein in 2014 the idea of the structure we had looked like this. But before I want to tell you how I got from the left to the right side, I actually want to tell you a little bit about the protein itself. So KDPF-ABC is exclusively found in prokaryotic organisms and there it is expressed under potassium limiting conditions. So that means when the external potassium concentration falls below 100 micromolar, KDPF-ABC is expressed to actually secure the cell's survival. It does so by taking a potassium against a very steep gradient into the prokaryotic cell. This process is dependent on ATP hydrolysis and hence KDPF-ABC as a primary active transporter. KDPF-ABC as a whole consists of four different subunits. And the first one you already know now and actually I already stuck to the color scheme so you already saw that in Joff's presentation this is a P type ATPase. Has the four classical domains or the transmembrane domain and then three cytoplasmic and PNA domains. In KDPF-ABC the P type ATPase is paired with the potassium channel which we can see here in green. This is a member of the super family of potassium transporters, it's four-fold pseudo-symmetric and also has a central selectivity filter pool. Finally there are the two smallest subunits which are called KDPC and KDPF which probably only play minor roles in the transport. So KDPF-ABC was first discovered in the early 70s by Wolfgang Epstein and since then it has been vigorously studied. So amongst other things we know that it's highly specific for its substrate potassium. Also the hydrolysis of ATP can only happen in the presence of potassium. And the other way around transport of potassium can only happen if ATP is there for hydrolysis. Apart from that also singular minor assets that are crucial for ion selectivity but also transport function have been identified for this complex. But the molecular mechanism by which KDPF-ABC actually transports ions was highly speculated about in the recent years and only with the structures that we now have we are getting closer to having an idea of the mechanism. So yeah actually solving at that point still a crystal crystal structure of KDPF-ABC was the prime goal of my PhD thesis to answer this question. And yeah first I started with the production of the protein. So when I began this there was an expression system for KDPF-ABC that was based on the natural promoter which was the Bacteria II component system KDPD and KDPE. This system is actually induced by potassium limitation, a state that E. coli cells had to be adapted to over several days. And then in the final culture which was actually done in media that literally didn't contain any potassium. This state has to be closely monitored and then adapted by additions of small amounts of potassium that were just enough to allow the protein to function but not too much to actually stop protein production. As you can see this is very tedious, so it took a very long time to do that. You had to monitor it very closely and in the end the year that I got was not very convincing. So that was only 200 df1. So yeah basically the first thing that I did is was that I wanted to optimize a new expression system for the whole complex. And for this I turned to FX cloning, which is actually a very nice strategy if you want to generate several different constructs. This was first published in 2011 by Eric Gertzmann, Raymond Dutzler and actually contains a variety of different expression vectors that are different in tag position. You can get them with different promoters and they also fit for different expression hosts. So if you're looking to start a project might be worth looking into that. For me, I had to have several decisions that I had to take actually so I had to decide between a T7 and Arabian nose promoter for the expression in the coli. In this case, I actually favor the Arabian nose promoter because this is a tunable system so the amount of Arabian nose you add for the induction actually correlates to the amount of protein that is expressed. While the T7 promoter for me, rather work like an on and off switch so as soon as there's some IPTG around you will get high protein expression. Which actually in the case of KDP, FABC was not favorable because too much protein express meant that a very high amount of potassium was taken up into the cells and that actually sees cell growth and most of the cells died. So yeah, this was the decision for the Arabian nose promoter. Then I had to decide between N and C terminal tagging of the protein in both cases, whether his TEN tag. And what is very funny here is actually both of them worked and then the N terminal tag because all four subunits are on a single operon. So the N terminal tag would actually be here, which would mean tagged KDPF. That is only three kilodolton in size and it's still worked to purify the whole protein. But as you don't really see three kilodolton on any normal SDS page, I decided to go with the C terminal tag. That also worked and then also gave me the guarantee of the whole four subunits being expressed and also I could nicely see the 25 kilodolton of KDPC on any Western blotto SDS page. And then finally I did one more thing which was that I changed the induction of the expression for the KDPF gene to the late exponential growth phase. So only after the cells that actually reached ODs around 1.5 are induced with the Arabian nose and then only for one additional hour. Because again, although I lost, we used only 0.002% of the Arabian nose so close to nothing, the cell still sees growing after induction. But I had an expression system down. I got nice cell yields from that. And actually after membrane preparation, the next thing that I had a closer look at was solubilization of the KDPF ABC complex. And before I started, the common detergent that was used in like 95% of all publications was called Aminoxide WS35, which is an industrial scale detergent that came in this nice and handy five liter canister that is still there in the lab. Yeah, and was just very tedious to work with actually never gave any structural results. So it was clear that this mix of different detergents had to be exchanged for single defined detergent. So I actually performed a large scale detergent screen so I tested 70 different or around 70 detergent for their effectiveness and solubilizing KDPF ABC. And so what I did is that I prepared membranes at 10 mixed per mil, had small fractions of that added 1% of all the detergents overnight. Did a spin and then put the supernatant on a western plot always comparing to the membranes that had used for all of this. And as you can see, several of the detergents actually worked nicely for KDPF ABC, which was pretty surprising for us at that time point because as said this Aminox seed was said to be the whole Holy Grail in the case of KDP. Yeah, then to to further evaluate this and actually check for the functionality of KDPF ABC and different detergents. I used an HPA essay so you just check how much ATP is hydrolyzed over a certain amount of time. And as you can see, I mean it's just an example of some of them and actually most of them also retained nice activity of course, for school and 12 kind of as a negative control here. But, yeah, as you can see some of them worked. And in the end I actually decided for DDM so a lot of screening to end up with solubilization and DDM in the case of KDPF ABC. And yeah, I went forward with that. And, yeah, actually then I spent another good amount of time on. Yeah, the purification of KDPF ABC after solubilization because also there. Not that much not that much published at that time. And I ended up with a three step purification strategy. So, what I did is I bound to solubilize protein to a nickel cephalrose resin. And then eluted by a three C protease cleavage actually have a three C cleavage site in between the his tech and the protein. And you can see the illusion here so that is already quite pure. And one nice thing is that because the three C protease itself is actually his tech it remains bound to the column. So I actually added to severe buffer and the protease directly to the beats just had that steering for a certain amount of time and eluted eluted on the protein and not the protease that stayed on the beast, as you can see here in this image is all illusion. The next step was an iron exchange chromatography to remove additional contaminants. Actually, so the isolated point of KDPF ABC is around 7.5 so I decided to do that at pH eight. With an anion exchanger so positively charged resin that's accused of a rose for example. I eluted with a sodium chloride gradient between 10 and then 500 millimolar. And as you can see a nicely separate the protein which is here from all contaminants that are high into 60 absorbance that otherwise would be very hard to get rid of. And then the final step was a gel filtration for buff exchange so to get rid of the high of the salt concentration at this point so which was also quite high. So around 300 millimolar and yeah to go for final buffers for crystallization. And yeah, this is the result. This is the SDS page that comes with that again KDPF is not visible to its small size, but it's there. And we can all agree that this is a very nice peak but unfortunately, I tried to crystallize this for, I would say, better part of my PhD, and I never got anything so that is the most frustrating stretch that I had. And actually, then a few years in a crystal structure of KDPF ABC was published by by Björn and David, and I would have never gotten the trick that they use like she does is the buffer composition. They use for Jeff attrition to 1.1% of better OG, and then 0.5 mixed per mil of team PCs actually adding a lot of lipid to that. Once I had that protocol, I was actually reproduced able to reproduce their results and also got some crystals that defected to between 3.5 and 4 angstrom at that point. So that actually worked for me too, which is quite nice. But luckily, in around 2018, I got another chance at this actually in cryo EM because we were approached by Christina Paulino and her people. And we actually were able to put this on a cryo EM grid and this didn't take a lot of tries, actually, because the sample was just perfect, I guess, for cryo EM analysis. In two to three trials, we had our first glimpses of what it looked like. Yeah, and if you actually want to know how we turned out and what what the mechanism was that that we proposed in this first or with these first cryo EM structures of KDPF ABC. You can have a look at this 2018 paper, which we published at that time. Yeah, then I would also like to mention the first crystal structure that I already had two slides before, which already revealed some of the mysteries of KDPF ABC. And then finally, I want to point to this pre print, which was just published or not published, but out there last week by Jacob was now leading the KDPF ABC project who together with Robin and Lisa had a very close look at how the iron transport mechanism actually works down to single and minor assets. And this is totally worth a read, I would say. And with this, I'm already at the acknowledgments. So, yeah, this is work for my PhD work largely done in the group of England at the good university in Frankfurt already mentioned Jacob, and then there's also a gun dollar to postdocs in the lab with also involved in the project. Yeah, Christina Lisa, who we did the cryo EM with at the University of Groningen, and then also fill in Robin who do some MD simulations on the project. And I also have to thank my new home. So this is here in the lab of Paulson, where I know actually switch to a mammalian expression system so something very new for me and also very exciting and they have given me great new home. And I want to thank them and the old gang and also all of you for your attention. And I'm open for questions now or also later down. Yeah. Thank you so much. Thank you, Charlotte for that very nice story, right. Super nice. That's a quick peek you had on that sample looked perfect for Chris, right. We have a few questions. Julie Tucker wants to say this well done for preserving Charlotte persevering Charlotte sorry. I would like to see that the second profile for the key the pf ABC that crystallize looks less monodisperse than the one did not any comments. I actually think it's the lipids. I mean, I think that like the specific mix of OG and DMPC that is already close to what you would use for like visceral crystallization so DMPC is used a lot on that. And this is also why the peak doesn't look that or less sharp I would say because of the lipids but I think it just might do the trick that it's like an estate between a bicell and something else that works for crystallization. Yeah, I mean David Drew was kind of into that as well so today is earlier today so yeah I get also the feeling. And one question I had was actually how much how much yield did you get. I'm actually not super lot for bacterial protein expressed in E. coli so 12 liters I got like half a milligram of protein. Yeah. Still you could work with it. Yeah, I mean it was enough for for I mean. I did a lot of very big expressions during my presentation of course but then once query and work that was very nice. So if you would start all over, what would you do, would you start with cryom from start or go both ways. I think again that the years are what is very determining for that right I mean cryom you can do that with such little nowadays and for crystallography. So if you want to go big and then even for it to be or something like that you need so much material that it all depends on what you get out, I would say. Okay, so with that I want to thank Charlotte and all the speakers for today.