So let's just hear this. From reverse and that is that read history and there's nowhere in it you know in this western. Coast off road on and count that that uses an Emory and I you know five years running faster than history and he won and if you're nineteen I. Was. OK thanks this it's Thanks for the invitation to come here it's certainly a pleasure to come over at the tech. On occasion and I end up coming here you know pretty often actually I'm on the committee of one of the students I think in biology so. One of the graduate students in biology so. Maybe the research area that we work in is not so familiar to you. But. It's still we have the same challenges that anyone does in design how do we design things that have a particular set of properties of which you might be interested in exploiting for applications for example and I've been working in the area of peptide in protein nanomaterials basically for twenty plus years since I was a post-doc at the University of Massachusetts and started working in this area so one of the things is that I wanted to sort of focus on with this idea of the design challenge and and design challenge I think one can sort of apply broadly and a number of types of materials feel just like how do we control you know structure of the sort of molecular into the Superman like it or an extended levels which is going to have a profound effect obviously on the properties of something and coming from the way we do from biology slash chemistry perspective it's often the way that we view the problem is as a bottom up type of design rather than top down which one might think of maybe from electronics in some ways but. You know we think of things in different ways. So. How we got into this in this is a good sort of summary of what the what's behind the logic in which we approach the problems that I'm going to talk about. And so one of the fundamental fundamental concepts of biology is especially as we go progressed from information storage to transfer and action is that sequence dictates structure and function and this is true not only for folded proteins but also true for you know proteins that behave in terms of the materials properties and that's primarily our interest so we take inspiration from nature and example that's been around for quite a while but is really only beginning to be understood to the point where it can move towards applications is spider silk and so for example the sequence of this were were first determined and there's a number of types of silk proteins but the major one the dragline silk is I want to talk about it has the most interesting fiber properties sequences were known from from gene sequencing basically and using the genetic code to elucidate what the sequence was at the protein level back I think probably it's been more than twenty years ago now probably close to twenty five years and what they saw was something interesting is that the sequence of the socks themselves consist of an alternating. Sequence almost like a block or polymer if you prefer something like that is an analogy of Allen enrich domain shown here in yellow and glycine rich domains which are highlighted to various degrees here and these two types of them you know acids have different properties and it turns out that they sort of segregate in the fiber or the alanine rich domains form Crystal in beta sheet type regions and the glycine rich domains be more flexible that lacking sort of the substance you want on the Alpha carbon to have what they have greater cover more flexible conformational properties and they interconnect these things in sort of form an amorphous network in which the beta she crystallize for imbedded in that's gets the fiber It's a unique come. A nation of both tensile strength compressive strength which is which is unusual to see and type of synthetic material for example and so it's all sort of nano nano scale structural and a structure that's in coded within the poly peptide sequence and so when we think about the types of problems the work on wheat and the general approach that we employ in my lab we can summarise this in a number of questions can we make synthetic protein or peptide materials that emulate and potentially expand upon the properties of the native protein materials for the native protein materials are evolved to work under a specific set of conditions like in a spider web for example where we might want to work under conditions that are well beyond those. Not all pretty proteins are not going to be able to work under all conditions you know going to be working at really high temperature for example but you know can we work on to understand you know other given sets of conditions that would be beyond those of nature and do things beyond the native role of proteins and so that that is sort of a general question. And can we design these proteins rationally and synthesize I'm using conventional chemical and like it or genetic approaches. Can we synthesize them I think that's clearly yes I mean it's been done now in the genetic engineering goes back almost fifty years now chemical synthesis of peptides almost equally the same. But the rational design is really is really the critical consideration because we don't really understand how for example prepared types of proteins fold to sort of give the sort of manifold of different types of structures that we see in nature it's still an open question and so then there's always going to be a question and this gets to the design aspect of of whether how much confidence can we have to design a particular structure and if we design something will it will it have the types of properties that we envisioned when we ultimately do this obviously we like to be informed by some we would want to root around in the dark but we like to be informed by some sort of guiding principles we don't have full confidence in the fact a week and. See an example of this in the talk but we don't have full confidence that will always be able to get the structure that we desire and sometimes the sort of an unexpected results are more interesting than what we would have expected in the first place so can we determine the structures you know using Carrier characterization meant methods that enable correlation between a design and the observed super molecular architecture so if we're looking at assemblies of things we have to be able to you know characterize them across the length scales and and peptides and proteins often assemble in what's called a hierarchical fashion insight that you have information encoded at lower levels that gets propagated hierarchically and because of the sequence specificity that's encoded within peptide and protein sequences one has the potential for at least in principle. Creating higher order structures with with with a great degree of specificity and one can see this inside of a cell for example that that cells are able to control their internal environment and and that by through the sequence of of the nucleic acids and proteins are able to to basically assemble into very highly organized and structures that are able to do things like operate far from equilibrium which is the conditions under which life occurs so we we would like to be able to understand our designs in terms of the super molecular architecture that's observed so this is almost like a feedback loop can we can we design a certain thing to have a particular architecture going to study the architecture and often is acquired requires application of a number of different types of physical methods because we're interrogating at different length scale structure at different length scales and so we need to be able to apply different methods to be able to do this and ultimately what we like to do is the same thing in nature doesn't code function and a function could be and you know anything we like to also potentially go beyond the sort of native function that you see in biological systems obviously which is what I sort of alluded to here. But we're still you know if we're controlling structure is still at it not primitive but still an early stage and coding function is even lagging further behind that and we take a lot of cues from already existing proteins in order to be able to encode to understand and potentially try to do this now. Something happened to this is no way over here I don't know why but it my computer just had a big X. and said there wasn't enough memory to do this even though I was using it in my office with many more windows open so I'm not exactly sure what's going on but I had a nice series of. Ribbon diagrams of structures of natural assemblies of flagellar filament and thirteen phage Pillai type three secretion system needle and the structures have been determined to varying degrees of certainty resolution basically being the application of the types of advanced characterization methods X. ray fiber diffraction cryo electron microscopy solid state enum are in some cases and we have a good idea I think within within the limits of resolution how these things work even though it's where we are you just have to use your imagination I guess my but. And but they have a number of really interesting roles controlled release directional transport locomotion and he's in that we'd like to be able to control and synthetic assemblies and what you would see if they were he if the if these images were indeed here was that you'd see that they're built up of a basically alpha helix all segments that interact with one another in a specific way and these are these the heel is the make these things actually have super healing symmetry and so what we like to be able to do is to be able to understand how to encode this this healable structure at a very fundamental level to build up things using simpler systems that recapture the structure of some of these organelles that have you know the properties that we like to emulate but do it in a way. That we can basically move beyond potentially nature and so the design principle comes into play right here and this is well you know how do we do this and so there is there's been proposals in the literature for a long time of how to build up organized assemblies of proteins and the simplest possible motif that you can have is an alpha helix I mean probably everyone's heard of that in biochemistry class but it's just a repeating structure as as are all heal Issy's based upon three point six I mean no acid repeat and. It's probably the most stable structure is the easiest one to understand and at least and we know that the most about helix all proteins because there are relatively simple to control the folding of compared to more complex proteins and the thing that you can think about this in the simplest possible way is to think about it is just a robin right but the interactions between rides can be programmed you know through the sequence of amino acids and so there was a proposal that appeared in the literature almost almost fifteen years ago now that if you took an alpha helix which is shown down the end these these amino acids are shown in schematically in terms of circles that wind around the helix this helix is shown to have two faces of red face in a blue face. And at these heels he's interacted the blue red face and blue face selectively with one another and this implies obviously you know local interactions between amino acid site chains that are stable but if they interacted selectively that once you form a closed cylinder of this type then and because of the geometric regularity of alpha helix he could predict how many would be in the structure so we thought that this would be a good place to begin. To design more complex assemblies and that we took a cue from a natural assembly which is of a protein known as Tulsi and not to get to you know bogged down in the structure this is this is a really interesting protein because it's responsible for drug resistance and gram negative bacteria it's basically a. Pump and if you look down through the center of it you would see that there's a hole in the middle through which it can open and close and pump drugs out of the system to to impart this behavior but part of the structure was a so-called Alpha cylinder alpha helix barrel that I described in the last in the last. Slide and so this is one of the few known examples but it turns out there are there that they also occur in. And bacteria phages In other words and bacterial viruses where they form the tube through which D.N.A. gets injected into the cell at least in some bacterial viruses and so this gets an indication of the potential for directional transport and these types of systems are very specific cargo or substrates so that would be something that we would like to try to understand and ultimately use so we started making a group of these these peptides based based upon the same sequence of a showed earlier very schematically an alpha helix in which we would put at different positions amino acids having different characteristics with the hope at least guided guided mostly through analysis of known structures that these things would interact in the way that was proposed earlier. By Walsh on Wilson and in the paper that I look at too and we thought well what types of amino acids should we put at these positions small hydrophobic PSI change the large hydrophobic side chains polar groups charged amino acids and so this would put the polar in charge would allow contact with a solution whereas a hydrophobic groups would mediate this interaction so on paper this looks simple enough and we based our designs on what was known about that protein that I alluded to earlier Telsey. And. We tried to promote the formation of all of larger diameter assemblies by putting a larger hydrophobic amino acids in them in the interior positions in the smaller ones in the exteriors. But we ran into a problem. And that. If you don't really know too much about protein protein chemistry I can tell you this briefly is that these these all the designs that we made were prone to rearrangement to a baby sheet and this is sort of an on control of the legal process and they are they interact in ways that we don't want you know we we didn't we wanted to and we wanted to generate Alpha healers is that would interact with one another and instead we got rearrangement the baby she now there's a way of understanding this and that's the fact that baby baby she to sort of an alternating sequence of and so if you put anything in like hydrophobic and hydrophilic amino acids they'll tend to favor beta for the formation because the polar repeat reinforces the structural repeat of the other other amino acid in and out of the plane so so we had to design against that and this is something that you see occurs often through the action of evolution is this God does this concept of negative design. And so evolution not only selects for a particular amino acid but against alternative structures and that's something that we don't often think about when we think about materials design we have but in complex systems such as this you need you need to not only design for a particular structure but against things that may be a comparable in energy by putting in you know repulsive interactions and we were able to do that the problem was it restricted the number of amino acids are the types of you know acids that we could put in the sequence pretty strongly. So the idea was and I'm not going to not going to go into this into much detail but the idea was that we could come up with a couple different peptides sequences and I won't go into the details of them but you can see that most of the amino acids and in each of these two peptides are identical you know and so at these positions the only difference is here this K. and here this are these are both positively charged amino acids K. is lysine in ours are Janine we thought this would be a relatively conservative or but nine substitution. That these things should behave very similarly. And so of course leading up to the point that they were not similar at all but and this gets again to the to the concept of. Design and how much influence we have upon it in biological systems at this point time so if we dissolve these peptides in solution and measure the spectroscopy using techniques that are there are diagnostic for different types of secondary structure we see that they behave very similarly again I won't go into the details of this but we see that the curves in each case which refer to different types of spectroscopic techniques are nearly identical and both consistent with the formation of Alpha healable assemblies which is what we which was at least encouraging because that's what we wanted. So. So do it does it do these things form higher order assemblies well that ultimately that was our goal is to form higher order assemblies that resemble like natural types of Helix all assemblies that one sees in say flagella or and you know other types of tubular assemblies and this was with the lysine containing peptide. And what we saw is if we just let it sit at room temperature for a while we saw these flexible fibers but slowly they rearranged in these into these assemblies with higher persistence lines long aspect ratio high persistence life assemblies and that's not practical to let it sit even at low temperature for four months with a protein but if you briefly an eel it at higher temperature you see and drive all of the peptide essentially into these assemblies so this was quite encouraging We didn't know what these things were at that point in time but at least it looked like we were forming something interesting. Which is always gratifying to see and when we did some higher resolution electron microscopy structural analysis at Brookhaven we saw on the even more interesting in that is that we think that this those initial flexible fibers that we saw actually of wind around each other to form these these tubes and so here is a typical bot filamentous virus tobacco mosaic via. And it's it's a prototypical virus it was the first virus it was ever crystallized and it's been studied in detail and used sort of a as a calibrate for these types of measurements and we see that all types of assemblies right here are a similar in aspect at least to tobacco mosaic virus slightly probably about two thirds the diameter of tobacco mosaic virus these are about twelve nanometers in diameter about tobacco mosaic viruses eighteen so but the way that these things assemble is quite interesting these fibers seem to be wrapping around each other and someway and in a way in and we did not at least anticipate this at the very beginning but the question is Well what was was going on and so we could do more detailed structural analysis. On it using using cryo electron microscopy cryo electron microscopy of a type and get some idea about the density of packing of these things in the assembly but we could not get any atomic level resolutions structural information using these methods. So I mentioned we had two peptides the two peptides are very similar in sequence in fact the only difference is we are very conservative from a biological point of view and for a chemical point of view from I guess from any type of structural point of view and if we substituted for positive the charge that you know acids of one type or for positively charged amino acids of another type if you talk to any protein structural biologists they would probably say that's a relatively benign substitution but if we take the assemblies of both then we can see both we can see by electron microscopy that they both formed assemblies and I'll come back to that in a minute but if we did some small angle scattering which gives an idea of the shape in solution and compared that to the red one is the the Argentine containing peptide in a balloon is a lysine containing peptide and we see that in general they look very similar we see this fine structure down here which is consistent with a hollow cylinder Basically you start to take my word for it if you're not. And sat if you're not familiar with sax measurement but you can see in other words the shapes of the. Curves are consistent with a one D. assembly hollow cylinder but they're shifted with respect to one another and in sort of this reciprocal distance parameter and that this is consistent with having a different diameter and which is something we could indeed see even only did the cryo lecture on microscopy of both under the same conditions and compare the two so this is we call this form one in form two this is the Argentine containing peptide and is the lacing containing half as are already seeing some difference here the diameter is quite different but what really does that imply what is significant of what significance at the structural at the most basic level structure does that sort of imply and it's always been a challenge to structurally characterize assemblies of this type at atomic level resolution because you can't do in many cases the conventional types of structural characterization like an M R Or is there a solution M.R. or. X. ray crystallography which will allow you to get Tomic level resolution. Structural information and but there's been a sort of a revolution in the last few years and the resolution of one could obtain from say cryo electron microscopy which in combination with modeling unable to now to get to near atomic level resolution and so with with some collaboration. We were able to solve the structures of these two assemblies and to near atomic level resolution at least to the point where we could see that not only the trajectory of the backbone but also the packing of the site change within the assemblies and what we see if we look at these is that they're fundamentally very different structures. Certain aspects are similar but the but again getting to this idea of hierarchical assembly of very basic sequence or structural information being propagated over many lines scales we see that at least at some point in the assembly in the hierarchy of assembly that they begin to diverged. And I think that we can we can sort of as a get to analyze why this is the case so this is the lysine containing structure and this is the Argentine containing peptide in the lacing containing peptide this one is consists of by layers of heel Issy's In other words the asymmetric unit at least from that perspective is a pair of heel Issy's we can't resolve whether they're pointing in the same direction or in opposite directions at the level of resolution that we have on the structure but we can fundamentally see that unit. And the other peptide it's based upon a monolayer of Helix the asymmetric unit is a single helix and so at a fundamental level they're quite different in terms of a single helix type of Helix in a single layer making up the structure of the one on the right and here a by layer of Helix making making up the structure on the left. And you can even see a bigger difference if you look just down the end of the tube. And what you see here is that this one actually has in the a four fold symmetry in the structure and this one actually has three fold so McChrystal bit harder to see because of the way that the heel is he's packed together so three fold symmetry versus four fold symmetry intersect and are seems at least in this case to be part of the helix hanging out of the end of each strip of each. Of the layers of heel Issy's so. This tells us at least at some point they mention the structures are diverging and the question is how and why. And we went back and we can model the Sachs data by taking a structures and so we know that these structures experiment agrees with what the calculated profile for each of the each of the assemblies and so we know the persistence solution and that these is an independent check that they look good the structures look good so. So I mention that we see in each case Sheila sees sort of stacking on top of one another. And. What we saw which was interesting is that within one of these stacks of field season there's four stacks in this and three stacks in this one but within a stack of healer sees the healer sees interact in either structure quite similarly in fact you can if you just overlay them on top of one another you can see that they're almost nearly coincident and so within one of these layers of heel Issy's they are stacks of heel A C's They are very very similar and so it's somehow the way that these stacks come together is causing a difference in structure and so we've been able to analyze these using Come conventional bioinformatics approaches and show that they fall into the class of heel Issy's that we originally based our design on which was which was good so it would seem like the initial design was good but there's an element of unpredictability once you get beyond a certain level of structure and we don't really understand that So here are two structures again just shown and this is the one with four four fold symmetry and three fold symmetry based on a by layer and based on a model layers of Ulysses the only differences are four amino acids shown in a one letter abbreviations we mutate for our genes down over to four life scenes one type of positively charged amino acid for another. Why did you structure so different at the at the ultimate level of structure why do we see this difference and the corollary to asking this question is now that because of the high resolution structural information that's available from Cry electron microscopy that we can start to ask these types of questions which were can which conventional structural biologists and small molecule chemists have been able to ask for a while as to the influence of sequence and structure and that's something which is here to for has been a difficult to do using extended assemblies of this type and so why and so the experiment to do would be to start mutating or start you know substituting one of these amino acids for another with in a constant back. And seeing if we can sort of cause one structure to flip into another by one or fewer amino acid changes and so the structure which is a higher resolution structure is a so-called form one structure with the Argentines and if you take a look at the structure we Swedes saw that even at lower levels of resolution. That there are these Argentine side chains that are in the center of the peptide that become ordered even at lower resolution and they seem to be making a critical interaction hydrogen bonding and capping off the helix on the on an adjacent stack and so we thought this might be the mutation to make for example as we go back and look at the sequence again it's these two Argentines which are separated by three amino acids which means they're on the same side of the helix and so they could form a little sort of. Interacting pair or dyad of of amino acids that could sort of bind to or interact with the sea terminus of an adjacent helix. And so if we mutated one of these what would happen if we just took the left these three Argentines intact and mutated this one which is our thirteen over to a lysine. And what it does and so this is this is the this is the lysine only this is the Argentine only peptide this is a single mutant it looks more like this and we've been able to show this by other methods and so one single amino acid changes enough to flip the structure over to the other one and that's a pretty profound observation that we we wouldn't would have never guessed without having high resolution structural information but. The fact that a single amino acid change which is a very conservative substitution conflict between two different structures is something that we hadn't anticipated and gets to this idea of well how reliable are design rules right now and can we design higher order structures with any degree of confidence and return to that question hoops so so if you want to. Do something else like if we wanted to go back in the opposite direction and say well could we convert this structure here over to this one by mutating one or so of these amino acids over to the opposite one of the Argentine if we did the exact opposite of what I talked about a minute ago mutated this lysine over to an Argentine it completely inhibits the ability of the thing to form either type of assembly. So that was unexpected but upon thinking about it a bit more we thought well maybe both of these Arjen these are needed so we did this double mutation and mutated both over and when we did that actually. It mutated this structure over to this one so back over to the to the four fold structure and we were able to confirm this again by the by the power spectra. Associated with the assemblies and we saw this flipping little symmetry OK so a single one or two amino acid changes can cause a profound changes structure so the implications of this are you know some I'll get to this in a minute what are the design implications of something like this I mean did does it have the does it give us any insight are we just like pushing the boulder continuously uphill only to have to roll back over us so can we design peptides with any type of confidence and so I guess I'll get back to that question so the other question was well maybe we discovered something interesting and that's that these are Janine separated by three M. you know acids occurring on the same side of the helix is a functional motif that can specify a particular type of quaternary structure higher order structure that we saw in the first structure here and so if we made a longer peptide move this sequence down that should mean the walls of the tube based would have based on the heel Issy's should be longer the diameter should be wider and so if we did this mean if we did this on the sequence what we saw was that we could we could cause a conversion of the original assembly which we called Form one over to this assembly which we call form three and so we're increasing as you can easily see to die. Amator of these assemblies now and we could we could study this by conventional methods we haven't solved the structure but if we studied buys my small angle scattering we see now that the so-called form three structure falls in between form to and form one we think is maintaining the structure of form one but just having it the peptide be longer you know having it may be thirty six amino acids instead of twenty nine extending it out by a repeat unit is causing the diameter of the assembly to become larger. And we're just this this motif. This this are it doesn't really matter what's in between them that is R X X X R motif is what's as a certain specific type of interaction that encodes that's in code within it which we just sort of discovered fortuitously And so again that's the consequence of being having the ability to obtain high resolution structural information atomic level resolute near atomic level resolution structural information so. So the mention the. Protein structure biologist talked about something called design ability and I mean in some ways it's tautological you know it's redundant you know the design of all proteins it's something that you can design but does it have any real meaning. And the way that they think about this design ability is having a secret amino acid sequence Well I should say haven't how to have a structure of a protein of a pet or peptide and seeing how stable it is in sequence space which means is a relatively robust structure and in other words if you can find a lot of have sequences that are compatible with a particular structure then it means it's design of one of the words you have some confidence and so it's sort of a functional definition but nonetheless it's. It's it's useful for understanding what types of proteins may be designed it doesn't necessarily rule out anything because evolution selects. For other things just that up and then a particular structure is structured you don't want the protein to be super stable because it won't be recycled You don't want approach you want the protein to be foldable under biological conditions so there are other things that are selected by evolution rather than just robustness in sequence space. But so that when I'm talking about that I'm talking about tertiary structure like a single protein folded up so the things that I'm talking about when I talk about assemblies is quaternary structure multiple proteins are peptides coming together to form an assembly and right now there's a lot of there's a big question whether quaternary structure is design a bill whether it's robust in sequence space in other words. And one of the arguments against that is human is hemoglobin all of you probably familiar with hemoglobin as the oxygen carrier in red blood cells and one thing that's well known about hemoglobin is that. You know under certain specific conditions you can make a mutation at the interface that occurs at the interface at the surface of one of the beta globe and units and that will cause fire realization to occur and so normally hemoglobin is it is a touch America structure consists of two different types of subunits shown in red and and blue but if you make a single mutation at at a particular Mino acid that's on the surface of beta globin sub sub units. Then it causes uncontrollable a humorous ation basically to form assemblies and if you think about the size of hemoglobin in the fact that a single amino acid sorry that a single Mino acid can do this very. I guess that's a sign that a single meal as a change can do this then it becomes really well how what hope do we have of controlling higher order assembly and proteins However there are things that one could argue the opposite as well and you know I guess protein structure biologists are contentious are opinionated sorts so so they can argue that well in certain cases there are. There are Mino acid sequences in certain types of proteins and here's an example one the form the super helix structure that in which one could argue that the sequences are very highly conserved then. And that they can form these types of his extended assemblies with some degree of predictability and so the question is then identifying those and so we started looking at this question and seeing if some of these assemblies that. Can be some of these types of protein units can be can be built into more extended assemblies and so here's an example of this. Maybe in this audience you'll appreciate the title Sol annoyed but so these types of proteins are known as solid white proteins because they're repeated structures over and over again and one can consider them sort of as wrapping around like a solid light I guess they form extended structures and so we've taken single repeats that cut the can't they came from some of these proteins have been able to show that they form highly ordered extended assemblies we haven't determined a structure of these yet but least we show that they're promising and we can sort of go in and from the geometrical parameters associated with these native protein assemblies crystal structures that are that are known we can sort of calculate what a model structure should be and we want to test that hypothesis. But then the other thing is can we sort of design things using less biased from nature or in other words you know not taking directly a unit from a known protein or a consensus sequence from a known protein but do some sort of prediction and so the question is you know how do you do this and so there's ways of predicting protein structures computationally. But all of them have some drawbacks even though some of them have been highly successful and particularly when you're talking about highly extended assemblies with the Griese of symmetry we haven't gotten to that point yet where we can actually. Design needs from first principles and. So how do we know that we're not going to get something like what I described in the first part of the talk something it's rearranging to sue an undesirable. Structure something that we can't predict in other words we want some sort of predictability and in the things that we design and so I started working with a computational biologist who were green as a Dartmouth College and what we started to think about was well OK you can't do this starting from first principles yet is to computationally intensive and we don't really understand all of the interactions even to program into the design engines as so how can we do this not directly taking from nature you know not directly taking a specific motif from an unknown protein but at least taking it taking advantage of some of the information that we have from structural analysis of native proteins and so what he would guess that he's a computational biologist structural biologist and what he started it has been working on his idea of looking at the protein databases just a huge store of information structural information and if we look for protein if we look for specific structural motifs that are over represented in the protein database they should indicate that these sort of elements are designed for stable and so instead of trying to just sort of evaluate the stability of something in isolation or we look through the protein database to sort of mine it to sort of narrow down regions of conformational space to correspond it to particular structural elements that may be useful for building up assemblies and so he came up and then then we did the then we do what's mean we know the general. Protein motif to begin with then we can do computational design with less you know less and with less intensive expenditure of resources so in this case what he designed was a a peptide that has a helix and a strand So two different types of conventional protein secondary structure. The two most and most known that are connected by a loop and identified sequences using by mining the. Protein database that were consistent with this structure and when and we went in and synthesize these are structurally characterized I'm going to skip this but. So what so we were able to. Synthesize one of these peptides and we show that indeed it does form the types of assemblies that we think if we compare it to the model system so we could calculate you know some parameters like mass prolonged from the from the model system and compare it to that what we measured from electron microscopy and they seem to agree fairly well we wouldn't really know given a number of surprises we've seen in this field we won't really know until we solve the structure whether this is. Going to be useful or not so it probably won't have enough time to talk about all my slides but I'll talk a little bit about some other work that we're doing and so which again has some possible relevance at least to the way that you think about about the materials and so we've also been trying to extend these concepts from one dimensional assemblies to two dimensional assemblies which are quite a bit more challenging to. Fabricate and in comparison like one can always see one dimensional assemblies occur pretty frequently in nature but two dimensional assemblies are quite uncommon and so the analogy although it's really a poor analogy at least in terms of function to graphene Yes a graphene form is extended you know the localized assembly and you know it's a two dimensional assembly However if we're talking about proteins obviously not going to recapitulate recapitulate recapture the electronic properties of graphene but at least the formation of these extended assemblies might be something that we can do and might be on a lot different roles for for these types of assemblies as well if we can make them base from proteins Now some of this has been done for. D.N.A. using D.N.A. origami or other types of D.N.A. not a knot of structures but proteins we think have greater flexibility and I think has been shown the ability to incorporate catalytic activity and very selective recognition outside of the sort of. The sort of structures of D.N.A.. So so we could potentially do this they could have potentially make these types of assemblies from proteins or not of when we started this or not a lot of examples and again in many of the ways we get into these things by accident or fortuitously or whatever serendipitously And so we thought these might have potentially interesting properties but they're challenging to control the synthesis in general and this is true for a lot of today it simply is not just proteins However there is an example in nature and this is this so-called bacterial surface layers which protect the surface of some A certain types of bacteria and Archaea and they form very highly organized assembly that this is the form of things with different underlying symmetry this one has hexagonal symmetry and you can see it right here. But there's other structures that have different types of of symmetry as well it's all determined by you know the the nature of the interactions in the playing group. So we got into this by accident because we were looking at college and. College is an important protein because it's I don't know people say it's like thirty to thirty five percent of the protein in your body have seen figures vary there's a there's a lot of different types of college and twenty plus different types of college and and it is basically forms a scaffolding for tissue growth and differentiation and so it's an important protein at least from biomedical perspective and so we started thinking about well how can we sort of it's and a simple system recapture some of the properties of college and and in particular we wanted to take short peptides and allow them to grow up into larger college in assemblies and the first way we did this was to. Take blocks of charge peptides a college in structure is relatively unique it has a very limited to it typically uses a very live. It pool of amino acids and some of these are post translation we modified In other words there and semantically modified after synthesis but we thought well even within is within the confines of this structure we could potentially you know get collagen to assemble into five rolls and we put a positively charged college in Domain A negatively charged college in the main in a neutral college and domain and hopefully a positive and negative would align up and allow for fibro formation to occur and that's what we saw to be the case here where we could for the first time I think show that we could form banded five rolls which resemble those of native collagen. But for a number of reasons if you came very difficult to control the assembly of these types of systems in fact and the reason why as I'll get to is of is is that they were forming other types of assemblies in the mixture as well so we started thinking well I mention these college and peptides could form five rules and they form fireballs by the positively charged regions aligning with the negatively charged regions and then them just kind of staggering down it forms a college and actually forms a triple helix three college and peptides come together and if they're off alignment they can form individual fibers which can pack together however we notice that it was possible for some of these that was also very readily theoretically or conceptually possible for the triple he'll seize the pack in layers like this in which the positively charged offsets the negatively charged or compensates whereas the neutral block is in the middle. And we made some of these peptides. And spectroscopic lead a look identical to college and they had the college and signature and circular Dyker ism which is consistent with triple helix. I'll skip this but what we saw instead of Fire a formation is the formation of a really highly organized two dimensional assemblies and this. First I would say it was disappointing why are we seeing these little flakes here and in E.-M. why are we seeing the studios selling about the more we thought about it I mean if you look at the structure it seems to have a extremely high degree of internal order and you can even see the curvature at the corners. So it's it's not like anything we've ever seen before it just spontaneously assembles in solution and at the first peptide we made actually. Forms. Again it forms college in these college into the assemblies. However it's polymorphic in both what I would call the X. Y. direction which is the sheet direction and in a Z. direction you can tell by the differential electron contrast in the M. and so they're forming probably multiple layers stacked but also the growth is uncontrolled other than the fact that it's ordered in the X. Y. direction. And if we look at the structure by atomic force microscopy what we can see is evidence for the formation of. Of layered structures basically can see the layer right here like a little terrace on the surface which the line is going through and if you look at the the thinnest sheets. If we take the height measurements by A.F.M. on the thinnest sheets we see is that it's about ten nanometers in diameter and if you go and the college in peptide that we made was thirty six amino acids in. Length and if you projected that college and have we projected that peptide sequence onto a college in triple helix and measured what the distance should we should be theoretically it would be ten point three Now to meters which is the height of a single layer and so this tends to say that the peptides are packing the triple heels these are packing perpendicular to the surface of the sheet so they have a very organized sense of packing and we could see this from a number of types of methods we could see this by small angle scattering. In which we can see not only the evidence for the two D. nature. The assembly which could easily see by E.M. but also by the fact that there's Bragg reflections even in solutions so this was done in a synchrotron at Argonne and what it shows is the strong drag like reflections here particularly in the X. Y. plane where where we have you know which is larger and so you have organized structure extending over on a longer length scale and that the relationship between the two Bragg diffraction peaks is consistent with the well mation of a square lattice. And you can see this by electron diffraction as well so you could see you know the the major If bright that the my major and minor bright light as a diffraction lattice basically were offset by forty five degrees as one would expect for packing of this type and we see the four fold symmetry so. It's. To me at least at this point time was unprecedented that we saw something like this and now there are a few more examples including for example from our group that we see order to dissemblers like this. We are trying to understand how to control the assembly of these types of materials we know that we can do it through that through controlling the chemistry of the peptide you know the words if we make this central the central block longer the central sequence the longer it was originally for repeats if we make it seven what that does is increases the thermal stability. Of the assembly now it melts using a protein terminology at sixty degrees about of in thirty degrees approximately so now we made more thermally robust assemblies. We still form sheets if you take if in measurements. The. The height now is around thirteen and a meters which because we have a longer peptide still agrees with the length of the college and peptide projecting projected into the length of the peptide projected onto the college a triple helix there are still stacked we think perpendicular to the surface of the sheet and because this type of pack. Would localize the ends of the peptide on the surface of the college and. She weak if we functionalize the ends of the college and we could functionalize the surface of the sheet so that binds the gold nanoparticles For example you know it through through charge charge charge interaction attraction between opposite charges because the ends of the peptide are on caps and charged but we can also tag it with a biotin group and end with a and strip that in tag or nanoparticles which selectively recognise with high affinity to buy it and it's localized on the surface anchored to particles that down on the surface and so we can functionalize the surface of these things so the nice thing about having robust structural platforms like this is that we can do chemistry with them but the surface is because we can control the surface by controlling a sequence of the peptide and we've been able to show this now in several different systems that we published over the last couple years and in some cases. We can make highly you for reasons we don't really understand we can make highly uniform populations of these not a sheets and again we don't really understand we think it's some sort of frustrated growth process we're can only grow to a certain certain length before the the earth certain with before the. A packing becomes a disfavored further addition becomes disfavored we don't really understand it but I think because we can control and this is this is one of the great aspects of design of peptides because you control the sequence you can potentially control aspects of the structure in and across the length scales if you understand how the peptides interact within these types of assemblies we don't always understand that we're trying to get to a bit of a greater degree of understanding in these systems so that we can have some forward predictability and design of these higher order structures. It's an exciting time to be able to do this because of the developments of new types of high resolution structural methods sick. Give us this insight but we're not quite at the point that we would like to be so I'll skip a lot of this. So. And thanks Well first of all to for the invitation and for the attention of everyone during my talk but thanks to a lot of collaborators who assisted us in various aspects of this work but then it Adelman a good war Gorean or a close collaborators of mine good work as a computational biologist and as an expert in cryo electron microscopy particularly healable reconstruction. Some other collaborators a different places Zeile being usual is that Argon Joe was a prick cave and Elizabeth writes at Emory She's an expert our local expert and the students who worked on this and funding was from and ISAF and DIA we thanks for your attention again I could answer. So which. OK. We think it's we think that if you look at the model structure there is very specific hydrogen bonding in charge of charge charge interactions with the amino acids in the sea terminus of the adjacent helix or helix on an adjacent layer and so it's forming we think these two are genes separated by approximately six angstroms are forming a little receptor for the amino acids at the sea terminus of the opposite helix and so there's very specific interactions that are occurring at least they appear to be predicted by modeling of the E.M.F. below. Yeah. Yeah. You can but I think this and this gets back to the to the element of design that this was a very imperial design that we came up with just based upon its very very it's a simple structural considerations and actually it only assembles under a very narrow ph range of between three point five to four point five once it's assembled it's stable but we're sort of perched on a thermodynamic precipice I would say on either side of falls off steeply and so it only assembles under a given set of conditions and so we really don't have the opportunity too much to play around with the conditions to see if it ever converts because we have limited sort of. Wiggle room you know to do that and I think part of the reason why the reason was again is a serendipitous thing is we got lucky that this thing worked and we checked it down in that very specific Ph. Range. But. When we're designing a sequence like this and when you does even even computation design of synthetic sequences. You can often fall into you know strange situations of which it only folds or assembles in a relatively narrow window and if you get too far away from that you might there might be can a lot of kinetic traps or whatever and I think so we we really can't go beyond that where I really couldn't say why that's the case you know in any sort of way that would sort of build confidence in the structure and the analysis of what's going on but what I can say is that you know it gets again to this importance of this design principle of having some confidence in the design and having some confidence and and be in the ability to predict things and so that's what we're trying to get to because I think this is because it was a completely unnatural sequence we just it could have formed basically nothing I say was more light more likely than not the form nothing as we've seen because we made a lot of peptides and for nothing. And so we were just lucky in this case that it form within a very specific range of PH it it formed an interesting structure and we could solve the structure by crikey we M. It opened up a lot of bit of avenues for for us to move down but hopefully we can do it with having a little bit more. Predictability in the design and because just for the fact what we get stuck in some kind of you know there's no guarantee that something's going to fall just because you draw it on a piece of paper I think. Yeah. Yeah in that case you don't want it you know but you can in certain cases see that. Yeah I think you know again what you have to do is look at the way that nature assembles something assembles things like the flagellum which is a mode at rate it's just an A.T.P. driven motor and it's got some really interesting properties like it switches between a left handed and right handed helix. So it's a mechanic sensitive switch basically and that one is switch is it from right handed a left hand it changes the way that the bacteria moves you know and in one in one in one direction when it rotates it forms when it forms a right handed heel if you can swim around towards an object basically having having a particular directional motion but if it switches it if it relates and you have a way in which is to a left handed heel it's not just tumbles in place basically trying to sense the presence of a nutrient so so if we could try to understand how these molecular sort of these native sort of macro molecular machines work then we have the hope of designing something I think it's pretty complicated still to do that. Individual atoms I'm not so sure about I mean if they've done some of that with you seen not an expert in the field but I've seen thing making gears by A.F.M.. Well in that case when you're working with hard materials it's quite a bit different there's that there's a chance that the A.F.M. tip could perturb the structure of these of these proteins I'm sure you could probably find conditions under which it would work but the question is like what what are the types of interactions that one could engineer between elements that are essential the soft materials. Yeah with the. I mean the phosphates make up the backbone is primarily the specificity is given by the specificity of hard and binding interactions between the base pairs so all those things you are. Right so I think I think you probably have to use self-assembly to get to a point where you could start manipulating things at larger Lansky are not necessarily large but a larger one you. Choose me as God I don't know if the time is going to finish your. Questions there.