this is not phage display but don't worry phage display is actually even cooler it's a way that we can get bacteria infecting viruses called phages to display little bits of protein and this allows us to do things like find antibodies that bind and from like combinatorial libraries and to find where antibodies bind and to do all this really really cool stuff a lot of stuff other than just antibodies um some really really cool work so face display is this really versatile technique and it was first invented in 1985 um well at least the first proof of concept was shown by gregory winter um he and george smith would run the nobel prize for it in 2018 um which they shared with francis arnold who is really awesome too um and this technique is just really really cool and so i want to tell you more about it how it works some of the uses um so yeah this thing is crazy useful like humera the um the drug humera it was one of the best-selling drugs it was made using um phage display technology um to go through all of these libraries of potential antibody and find this antibody humanize the mouth mouse antibody using this technique um and developing this drug that is super duper useful so as i mentioned there are a lot of different variations to page display and there are a lot of different um techniques that you can use to optimize various things i'll get into some of these but first here's just the general overview of how it works and then we'll get way into the details so some of these terms don't sound like familiar to you now don't worry we're gonna um dive into them and hopefully you'll be less confused by the end so basically a phage is a bacteriophage it's a virus that infects bacteria and what you do is you get this bacteriophage to display a fusion protein so basically it has its protein but then you stick a bit of a protein you want it to show into its protein into this protein that sticks off the surface of this phage so what happens is that you can create like a library of millions of different phage clones where you insert foreign dna into this code protein gene then it's gonna because dna has instructions for making protein this is gonna add a little bit of that protein it's gonna modify the protein so it's displaying your protein in its protein and then because you're making this library with millions of these different clones you're going to have lots and lots of different versions of this beige then what you can do is you can basically mix them you can get them to infect bacteria and so you can make a lot of a lot of them wait now you want to find the ones that are going to bind to something of interest so there are different ways that you can do this so basically sometimes this is called like bio painting too you can you need to like immobilize the target of interest so this can be like on a plate sometimes it's on beads and you basically mix your face library that's displaying all these different protein bits that you want to test and then you see which ones bind so you mix the phages with the thing that the immobilized target the ones that bind are going to stick the ones that don't bind aren't going to stick and so you can wash off those non-specific binders such as with increasing the salt concentration after you wash it you're going to be left with ones that stick but these might not be the best ones that's like super duper great stickers but they're they're at least um moderate stickers that they were able to survive the wash then go to a loot so you want to unstick these specific binders which you can do using slightly harsher conditions such as changing the ph and now what you're going to do is you're going to take those phage that you have just eluded so these should be the specific binders and you typically want to make a lot more of them because you're going to want to isolate its dna and so you want to have lots of copies of the dna and we can you you can use techniques to make more copies of that in vitro so like outside of these pages but you want to have more to start with so that you don't it doesn't just get lost in the noise so in the amplify step you let these phages infect more bacteria to make more so that's the great thing about these phages is that they have like the bacteria infecting viruses so they're already good at getting into bacteria and getting bacteria to make lots of copies of them and so we can get bacteria to make lots of copies of them which is how we can do these techniques where we're testing like millions of different phases and it's not like incredibly overwhelming because it's well really like make the bacteria do the hard work for us and so the bacteria are going to make lots of copies of this phase and then these phages now we can look and see what what the dna was that was connected to the binding protein so the protein that's on the surface is going to be the protein that was instructed by the dna that we inserted but remember this is typically like a random library so there can be tons and tons of different stuff in in here and we don't know what it is so we have these millions of different phases with these millions of different proteins corresponding to millions of different dna sequences and we need to find out what the dna sequence was that caused this protein to be made that was specifically bound and so you can sequence this um so you don't have to sequence the whole thing because you know what the region was around like of this fusion gene and stuff you know the region where you inserted the foreign dna so now you can amplify that region or you can sequence that region to find what the dna was that was inside of there and now what you can do is you can take that dna and you can clone it into some other expression vector um such as like a plasmid so a circular piece of dna we can put into bacterial cells and get the bacteria cells to make more of the protein and that we can like purify it off purify it out and then we can test it for various functions sometimes too because you do this you often do this in multiple multiple rounds though so there's a few reasons for this one is that you can keep making the wash conditions harsher so that you don't so that you find the best of the best you can also because there's going to be some random mutations introduced throughout this process because when you're making when the phage is making lots of copies then there's going to be some mutations introduced because nothing's perfect when it's doing the copying process so these mutations might make the binding worse but they might also make the binding better and so if they make the binding worse then those are just going to get washed off especially since you have harsher conditions but if the mutations these random mutations happen to make the binding stronger now those mutations are going to survive the harsher conditions and so you're going to end up enriching for even by their binders to make this even more likely you can actually use like mutator strains um where the bacteria where there's less error correction and there's more mutation prone um copying so that mutations are more likely and therefore you're more likely to you have like faster evolution of the binding partner so there are lots of different combinations that you can imagine using phage display so you can we'll talk a lot about a lot of the technologies are for antibodies so you can have antibodies um oh so you can find antibodies you can find what antibodies bind to so basically if you want to find antibodies you can stick antibody genes into here and have little bits of antibodies little antibodies displayed from the surface and then have something you want to test if the antibodies bind to you can also do the opposite so if you want to find what an antibody binds to so that is the antigen or or the epitope of a specific antigen so like the exact site it binds to you can stick um different things into here like different peptides you can stick like little short pieces of dna for various peptides and then these peptides will be displayed and then you can see uh find the find which antibodies bind to it um and so you can test various things like what might be some auto reacting auto antibody reactive things um viral reacting things various really cool things and then you can kind of like reverse engineer in all sorts of ways um as well as find drugs that bind specific peptides pharmacists peptides that bind specific drugs all sorts of really cool techniques for when you want to test a lot a lot of things um have the potential for some molecular evolution and various things so really really cool technique and now let's get into the details okay so let's start with this phage um so a phase phase is short for bacteriophage and there we use phages a lot in biochemistry so actually yesterday um in my post yesterday i was talking about the um hershey chase experiments and in that case they were using a t2 phage to help show that nucleic acids and not protein were the source of genetic information this is a different type of phage so it's this t2 phase you can see it has this kind of like shell like structurally shelly thing um and then it has this linear double-stranded dna genome the phages that we're using for phage display are different type of phage they're called filamentous phages it includes like m13 f1 and fd they have a singular they have a single strand circular dna genome um and then they have this long filamentous shape major bacteria after docking onto this bacterial uphillis and then like injecting their dna inside they have a simple genome which is good when you want to manipulate things and basically they have a few different multiple coke proteins so proteins are the ones that are like out on the surface and so some of them are like really important for getting the antibacteria and some of them there are parts of them that we can add things to so we want to add things to the part that's actually going to stick out um of the phage and so a lot of work was done for example in that key early paper but to determine where in this phase is best for display there are a few different there are a couple options that are often used um most typically this like p3 as well as sometimes p8 so p8 it has about like 2700 copies of it which you might think would be really really good but it's kind of like way too much so you kind of if you have a lot of copies of something things can stick even if like one copy alone wouldn't be strong enough to stick um so we'll get more into this idea of ability later but where you have a lot of copies of something then if they're kind of all weekly binding it can kind of build together and help make a stronger binding and so this p8 um is less than ideal but the rest of the proteins have about five copies and p3 is this one that is fairly really useful because it like sticks out and you have like five copies of it and so it'll and allows some flexibility for you to insert foreign protein but what you're not you're actually inserting the foreign dna to make telling it to make the protein and so we call this like a fusion protein where you have like its protein with a bit of your protein fused into it and so it's actually typically fused not into the actual like end because that needs to be used to attach to the the bacteria and stuff so but you can see it kind of like in the middle and there are different spots where people stick things but you can stick in different different bits of dna and get it to display different fusion proteins and then you can create a library of different ones test if they bind things and then look back and see what the foreign dna binders had so if we go back to our panning strategy we would basically get these to bind and then get amplify them and then look and sequence what was in here this is super duper useful because there are so many possible combinations that can be used to make a protein so there are 20 um common amino acids or protein letters and you can have these different combinations of them and so there are just like so proteins are made up of these combinations of them different proteins have different combinations of them and have then the different combinations are going to affect how they fold and function there are techniques where you can try to predict what how proteins fold there's a lot of progress on that front but there's still just so many possible combinations that we need a way to kind of test them all which is where phase display comes in really handy because we don't have to basically we don't have to waste our too much time with the ones that aren't helpful because those ones just won't get selected so we make that big library and the ones that aren't useful will just get washed off and the ones that are useful we get to keep so it's kind of like a brute force cracking approach except that there's no one right solution there can be lots of solutions and so this is one of the reasons why this is great um because we're doing it like you can do this like a random library and so you're not like biasing in terms of like getting stuck going down one path to find the one perfect thing by allowing there to be multiple really good things um so you can allow this diversity and um really this in your testing actual direct binding is instead of just like um prediction stuff so anyway it's really really cool um and that was not meant as a day get prediction stuff and it's like the the better you can actually mix the two and stuff so you can use this technique you can bias your library going into it if you have an idea of what might find well you can use a smaller library to stick in those things so you can make a library say from blood in a person that was infected with the various disease in order to try to isolate antibodies from that patient's blood you can also try to do you can also do things like you take the take those initial hits that you get and then use computational methods and molecular prediction and that sort of thing in order to try to improve the binding and then you can even use phages to test the various improvements and although if you don't have that many you would probably do some smaller scale thing so i've been talking a lot about antibodies so let's get into um how this all works um and so an antibody is just this little protein that binds to some specific thing and so we call the specific thing it binds to an antigen and the exact region on the antigen where it binds we call the epitope the the antibody is actually going to bind to the antigen in the special antigen binding sites and what's special about them is that these parts are unique so each antibody um has unique variable regions and these variable regions are where the antigen binding sites are located so there are different types of antibodies this is an igg antibody which is typically what is um this is like the most common one that you see in diagrams and that sort of thing and what's typically used in therapeutics these are y shapes they have this constant region and then the variable regions what happens in your body and in the body of animals is that different their genes for different variable regions and what happens in your antibody making cells your b cells they can kind of randomly choose which to make once they make a choice then they're stuck with it because they actually edit their dna so that they piece together different versions of those variable regions to make a unique antibody and since they're editing it out then they can only ever make that one but if they happen to get selected for if they happen to bind to something that is foreign and not um self-rep doesn't recognize it doesn't bind to something that is not foreign they'll get selected for more of that antibody will be made and then they can even undergo further like somatic hypermutation um where some changes can be made kind of like how we talked about after the selection like the different rounds you can introduce mutations that sort of thing can happen with the somatic hypermutation so you make better and better antibody binders but the complicated thing is that and what makes this so great is that you have so many different options that these b cells can choose from that there's so much potential diversity um and the phase just blaze away for us to kind of recreate that diversity and do this selection in a non-body way so basically those there's actually multiple of those generic of those unique parts that combine to make those antigen binding sites and then so this the constant region is i should mention is constant for a given organism so our constant chains are going to be different than like mouse constant chains different from goat constant strains and rabbit constant chains which is why we can use antibodies against these constant regions when we're using like secondary antibodies and western blots and stuff that doesn't make sense don't worry you need to you don't need to know about it for this post um just if you were curious i like to throw in those tidbits to other techniques then make connections and if i have posts on western blots if you're confused we can go into that but these have the antibodies have a heavy chain and a light change constant regions and unique regions so you have a constant heavy a constant light variable heavy and a variable light because that we're interested in these variable regions we're typically talking about this variable light in the variable heavy chains so each of these parts has a its own corresponding gene and then the different antibodies are formed by mixing and matching different versions of this gene so this is what's happening in those b cells when they're like choosing randomly what antibody to devote their life to making and so the antigen binding site is going to be formed from a combination of both these heavy and these light variable chains so what we can do is we can create a phase display library of antigen binding sites by mixing and matching the genes for different variable chains and then making fusion proteins where we stick them in the phageco protein and get them displayed and then use our infinity selection our like bio panning to see which ones bind and then isolate the bh and bl genes from those phages um so sounds great right but there can be issues although antibodies are small for proteins they're big for sticking onto phages so normally we only stick on parts of them so remember we're sticking this protein kind of like into this phase just normal protein in selene to make sure that we're not just messing up the face um in which case we wouldn't this technique would not be helpful but what is helpful to us is that the we don't need this the constant region is going to stay constant and what we would care most about are these variable regions this concentration is really important in our body and that sort of thing for helping it attach to various receptors and get splayed and all this stuff i'm not immunologist so i'm not going to try to get into that too much but i'm not saying the constant region doesn't matter it does but in terms of binding then we care about the variable regions so we can just stick the variable regions on there so that this like sc fb this is like the tip of one of these arms of the y it only has the variable regions for each heavy and light chain connected with the linker and you can also stick just the individual variable heavy in variable light chains you can also stick in the fab which is like a single arm of the y so it also has part of the constant regions so this is a powerful technique for developing therapeutic antibodies so you probably have heard about therapeutic antibodies in the context of the coronavirus and how like neutralizing antibodies like monobot monoclonal antibodies or antibody cocktails can be used to block the to bind to the spike protein of the coronavirus and keep it from getting into cells but antibodies and are actually like a huge market for therapeutics and in like 2016 like six of the top 10 selling um drugs were monoclonal antibodies and so when you see and when you see drug names that end with a b that's for antibody and so antibodies are so useful because they combine just specific things and so they can say bind to and block various proteins involved in various diseases including the coronavirus but not just the coronavirus so scientists have a big interest in developing antibodies common ways for developing antibodies include traditional ways like injecting an animal with something that you want to act as an antigen letting the animal make the antibodies and then the antibodies will be secreted into the blood and you can isolate them if you typically when you want to get monoclonal antibodies so you want to isolate like a specific version so remember like this there's going to be lots of different b cells making lots of different antibodies in this mask that bind to the various the antigen that you want typically if you want you're interested in these like monoclonal antibodies we're talking about therapeutic purposes especially because we're going to want to make a lot of that so we want to isolate the dna from making this with monoclonal antibodies we need to isolate like the dna from making it or isolate the cell that it came from and then copy so the different techniques like hybridoma or sticking their copying their antibody dna sticking inspection expression cells and getting them to make more um with the hybridoma you fuse the a b cell so an antibody making cell with the tumor cell to make a hybridoma which can then grow and produce the antibodies but there are there's obviously um strong interest in trying to make techniques that avoid having to use an animal step not just for the protection of the animals and but also that these antibodies are going to be have the antibody the constant regions of the animal that they were made in unless you're using some sort of specialized animal that's been humanized so there are some like mice with humanized antibodies so they have like human and human constant regions but if you could just bypass all of that and directly engineer human antibodies with the phage library then that's even better um so there are huge like um huge libraries of the human antibody um a lot of this work has been done by greg winter where he actually like amplified the variable read the variable regions of human antibodies and then you can mix and match and make different faces that are displayed and selected for so you can actually do this with because different people are going to have different libraries because they have already gone through the selection process so you can kind of make it just super duper generic like every single possible combination but if you want you can start by like say if a person was had recovered from a various disease they would likely have antibodies that were good at um that recognized the proteins from from that virus from that bacteria and so then you could make libraries from their lymphocytes in particular and kind of start um get a head start in the race with your butt with your panning speaking of getting a head start um so the drug humera it is an antibody a monoclonal antibody against tnf alpha human necrosis factor alpha which is involved in some autoimmune diseases the antibody that was made was made during using phage display and it's a human antibody but it was actually made based off of a mouse humanizing a mouse antibody in a way so they had this mouse antibody that they knew was really really good at binding tnf alpha and so they what they did was they took the variable regions of this mouse antibody and they mixed and matched them with variable regions of human antibodies so they would take the mouse h and the mouse vh the variable eight heavy and the variable light mix them with the human heavy and the human light and find the human versions that um went well kind of with them and then mix those human versions um and do the selection this and so they're able to make a full human antibody so they don't have to worry about the human's body recognizing the antibody as foreign as i was mentioning before you can also kind of do the opposite so you can stick various potential epitopes so we can get these to display like peptides their little proteins various things that you want to see what the antibody is binding to and then you can take a single antibody and see what it's binding to so you're kind of doing this in reverse this kind of harkens back to what george smith originally thought was going to be useful for for like cloning genes when you didn't know what the gene um was so if you had like an antibody it begins to specific protein but you had no clue what that protein was or what like dna coded for it then you could stick random pieces of dna into this phase have them displayed immobilize the antibody against the protein and therefore select for the phages that had the dna that corresponded to the thing that's the antibody bound to and that way figure out its sequence if you can find that little piece um then you can use that piece to like fish out the whole gene for the product so i also get into a little more technical notes um so although we so the p3 it only only has like five copies um sometimes these can get cut off too so you might not even get five full copies but five is sometimes too much there's and sometimes you just want one so one would be like a monovalent display and this has to do with infinity and ability so affinity is how well each protein binds independently and the ability is the effects that come from multiple proteins working together it's easier for one to bind if there's another already tethered nearby so you can imagine if you have something with a bunch of binding sites and if one binds um this other one falls off but it's still stuck there because it's the one next to its bounds so then it comes back and it's like kind of like increasing your local concentration um increasing the probability of binding if what you care about is what you like what one kawa one copy binds so say if you were going to use some sort of antibody in a monomer form where you're just sticking like a single igg chain in there so you don't want to be like fooled into thinking that your antibody binds really well when it's really just this ability effect that's coming from this polyvalent display so polyvalent meaning that there's multiple copies available for like binding so this polyvalency is what's going to happen normally because you have those about five p3 copies but there are ways that you can get monovalent display there are some different ways but commonly do this using a phage mid and a helper phase so basically this a phage mid so you might have heard of a plasmid the circular piece of dna will often stick into um bacteria and the bacteria can kind of host it separately from their own genome and make copies of it and so a plasmid has an origin of replication initiation sequence for the bacteria so the bacteria can make copies of it and it can be propagated in bacteria separately from the bacterial chromosome a beige um is a virus that infects bacteria and it has its own origin of replication and a beige made as a sort of hybrid so it has some of the phase genes and the phage origin of replication but not enough to make not enough of the genes to make functional phage so basically it's going to act as a plasmid until you give it a helper phase that has the things that it needs that it's lacking so the phase mids um so they typically have like an antibiotic resistant selection gene too um so you can select for the bacteria that have them then they have the equal leg replication and then they have their own origin of replication and but in these patients they don't have everything that we need so they don't have say the things that they need to make the things that they need to like secrete themselves and coat themselves and all of this various things so they need help and this is where you can use a helper phage to add those other phase genes that they don't have and allow phage to get made so if you take a you take make a fusion protein where the p3 protein is in a phage mid so your phage mid is going to have the gene for the p3 protein but then the other phage genes are going to be provided by a helper phase you can let this bacteria grow with just your phage mist but they're not going to be they're going to be able to keep and copy this p3 plasmid but they're not going to be making actual phage until you add a helper protein when you add a helper phase or sorry when you add a helper beige when you add a helper phase you can add a helper phase that has a normal version of the p3 because this version of p3 the normal one doesn't have this bit of random protein stuck into it it's going to be more stable and it's going to be better incorporated so most of the phage is actually just going to be the wild type but some of it is going to be the one going to have a fusion protein and importantly you're not typically you're not likely to have ones with more than one fusion protein you're gonna have a lot without any but those are just gonna get washed off because they're not gonna get bind so it's not like fully efficient um but you shouldn't have polyvalency another thing is that the origin of replication for the phage mid is typically stronger than the one from the for the helper phase so this is going to be enriched you're going to enrich for keeping the stage mid around there's also some other technical reasons why you might want to use a phage man to mourn it my blog posts and stuff but you can do things like introduce amber suppression mutations so you can make soluble versions of it if you have this you introduce a stop codon that is can be recognized as not a stop codon if you grow it in certain bacterial strains it'll like keep going through and make the whole um and make the whole fusion protein on the on the phage but if you stick it in amber um if you stick it in non-suppressor um strength then a stop codon will be recognized and so you can have the soluble version of the protein made and they typically tend to transfect better in various things um so lots more on this and then the history in my blog post speaking of the history so as i mentioned um this one the 2018 nobel prize um george smith and sir gregory winter um they each got a quarter share in 2018 and the other half was shared by francis arnold so smith and winter got this for phase display of peptides and antibodies george smith um did this proof of concept in 1985 and then gregory winter really has done a lot of work with designing antibodies using it i'm not going to get too much into the historical stuff there's more on it in my blog post but i just i don't i want to make sure i get everything right and that sort of thing um but they should also i want to give a shout out to um so george smith and his nobel lecture which is really good he gives a shout out to some of the key people in this lab who did the work so steve parmly just developed the practical phase display vector and affinity selection as a grad student so heck yes and then the post doc jamie scott showed that you could do the affinity selection of peptides from random peptide libraries and robert davis chief manager and technician sequence the sequences of this hits and so i'm sure there were a lot a lot more people that were involved in this and there's been a lot of work done since then and continues to be a lot of work optimizing things and doing variations we'll talk about some of various variations um that we used but because science is always a team sport and that sort of thing but i also want to make sure that we got that i give credit where credit's due um when i know where credit is due um so yeah so really cool stuff