My name is William Shih. I’m an associate professor of Biological Chemistry and Molecular Pharmacology at Harvard Medical School, Dana-Farber Cancer Institutes, and the Wyss Institute for Biologically Inspired Engineering. It’s my pleasure to share with you today some recent technical advances in the field of structural DNA nanotechnology, from my laboratory and those of my colleagues. We’re all familiar with the biological role of DNA as an information repository, principally for coding for protein sequence and for regulation of protein expression. And I’m not going to speak at all about that today. Instead, I’ll be talking about using DNA itself as a building material and harnessing that in order to us to construct nanoscale objects. For example, shown here is an electron micrograph of actin filaments that are about 7 nanometers (nm) in diameter. And below, we can see this peculiar Pac-Man-shaped object. This is a structure that was built entirely from DNA and we designed it for the purpose of taking bite-size chunks out of the actin filament. Well, this project hasn’t yet been successfully, however hopefully this image drives home the complementarity between the dimensions of our designed DNA nanostructures and biological macromolecular complexes. So I’ve been working in this field for over a decade now, and I’m continually surprised by advances in the field. I have these preconceived notions of the limitations of DNA, and they’re always shattered by the latest new discovery. And today I’m going to be sharing with you, in the first two sections, respectively, two recent developments that enable us to design and assemble DNA nanostructures of the size and complexity of this object shown here, about 30 nm in diameter. One of the most important goals for DNA nanotechnology is to self-assemble objects of ever-increasing complexity over time. For example, is it possible that, within the next decade or two, that we can self-assemble objects that are let’s say a thousand times as complicated, that have a thousand times as many unique components, as the object shown here? A second question is: what are these things good for? And in the third part of this lecture, I’ll be discussing some applications in my laboratory as tools for molecular biophysics and tools for future therapeutics where we think these objects might prove useful. We’re inspired from natural systems, we know that they can carry out many amazing behaviors: they can build, they can adapt, they can heal, they can reproduce. And these are capabilities that human technology struggles to reproduce on any kind of length scale, but what’s especially remarkable is the ability of life to carry these out on the molecular scale. So, for example, here on the left is a picture of a ribosome, and this is of course a machine that’s about 25 nm in diameter that takes information encoded in messenger RNAs and then translates that into a specific sequence of amino acids, to produce a polypeptide. Quite amazing machine. On the right, we have the T4 bacteriophage, that’s a bit larger in length scale. We can see that the viral capsid itself — head capsid — is about 100 nm in diameter, so it looks like a little nanoscale hypodermic syringe that docks onto the surface of its bacterial hosts and then injects its DNA cargo into the cell against an osmotic pressure gradient. Of course what makes this all possible is that living systems have invented molecular manufacturing and they’ve come up with a very robust and clever way of doing this. They, of course, synthesize these biopolymers — they can be polypeptide chains, polynucleotide chains — that then self-assemble into the desired structure. Now, if somebody asked you about ten years ago, “Would it be possible to generate an object of this complexity using some kind of human-based technology?” most people would have been skeptical. And yet I’m going to show you a new technology, DNA nanotechnology, developed in the last… especially with advances from the last ten years, that now make it possible for us to self-assemble, in a programmable way, structures of the same kind of complexity as you see here. Not yet of the functional complexity, but nevertheless we think this is an encouraging first step because along the pathway to this kind of functionality, we first need to master structural complexity. We’re going to be using DNA as our building material, and we know that DNA, very similar to proteins and other macromolecules from life, are very complicated molecules. There’s many different atoms, but it turns out the key point for DNA nanotechnology is that the robust base-pairing properties of DNA allow us to abstract away those chemical details, which is going to make the act of designing the nanostructures much simpler. And in fact there’s only three characteristics of the DNA that we need to remember for the purpose of DNA nanoconstruction. One, is that it’s a ladder with antiparallel strands. Secondly, there’s a right-handed twist for B-DNA, and we need to know that the twist is around 10.5 basepairs per turn. It turns out we can switch that around a little bit. And finally we need to know that A pairs with T, C pairs with G, and anytime you deviate from that pairing, you’re going to destabilize the structure. And it’s because the propensity of DNA to form this very regular structure, enforced very strictly according to this Watson-Crick base pairing, that gives us its power in being able to generate these large structures with very little design work. The father of the field of DNA nanotechnology is Ned Seeman at NYU. He invented this field about thirty years ago. His training is as a crystallographer, and the way he came up with the idea is as follows: he was sitting in the campus pub one day, just drinking his beer, and suddenly what popped into his head was this woodcut from M.C. Escher. He had been collaborating with some of his friends on DNA Holliday junctions, and he had this eureka moment: why not replace the flying fish with DNA Holliday junctions? The notion was that if you could rationally design a porous crystal out of DNA, and then he could take the target protein he was interested in, and then dock that into each unit cell into a stereotyped orientation, then he would be able to impose that crystalline order on the target protein, and therefore make the X-ray crystallography problem easier for these large macromolecules that are otherwise difficult to crystallize. He’s been working on this problem for over thirty years, it’s an important goal, and he’s made some interesting progress, I’ll have more to say about that in the third segment. But in the meantime, he’s had some interesting landmark successes. So the first really noteworthy advance that he reported came out in 1992 or so in Nature, so this is a DNA cube where each edge of the wireframe cube is two turns of a double helix. Each face is a circular strand of DNA and the entire object has dimensions of about 10 nm. So the first time I think people saw this they thought, “Wow, this is really cool, but that doesn’t look like biology to me.” And what I hope convince you today is that this is in fact an extremely powerful technology. Yes, it’s fun, but it’s actually potentially very useful as well, for many different applications. Of course if we’re building from DNA strands and we’re just making double helices then that’s boring. The power of DNA nanotechnology is that we can build with branched junctions. With the previous example, the cube, each one of those vertices is a three-branch junction. But it turns out the most powerful motif so far in structural DNA nanotechnology has been a four-branch junction; a Holliday junction. So on the upper left here we have a schematic using simple letter notation of the strands, so you can see the cyan strand starts 5′-CCGG, goes to it’s 3′ and, and if you look closely at this you can see that there’s four different sequences and they have the proper sequence complementarity in order to generate a Holliday junction that actually is immobile. It can’t branch migrate due to its sequence. And we know from structural studies that this object likes to stack into two double helices that are connected at a joint, and it turns out this is really the building block that’s been the most fruitful for DNA nanotechnology. So the idea is as follows: if you only have one Holliday junction, now you have two helices that can wobble around. In order to fix those two helices to make a rigid building block, what we do is we simply introduce a second Holliday junction downstream, and now when we fix those two double helices with two Holliday junctions we have that rigid building block that we want, now with four sticky ends. The next step is to build two versions of this building block. In this example, we have a red one and we have a blue one. And we designed the sticky ends with the following complementarity in this example. So let’s say we make the sticky ends on the upper right-hand side of the red block, and we make that compatible with the lower left-hand side of the blue block. And so on and so forth, in order to create this kind of checkerboard fashion, hopefully you can see that we would be able to self-assemble these two bricks into an infinite two-dimensional lattice, as shown below. I don’t have the experimental images for this, but suffice it to say that this method actually worked. Quite amazing – you can design a two-dimensional crystal. The step after this would be to say, “Well, instead of two bricks, two tiles, what if I had ten tiles, or what if I had 100 tiles? Can I now make non-periodic structures that are highly complex, just with self-assembling tiles?” And unfortunately, nobody has really demonstrated this method, extending this particular method to 100’s of tiles, although what I’ll show you shortly is one method, DNA Origami, that can achieve this kind of complexity, and in the second segment, something called single-stranded bricks, that can do something very similar to what I just described. The method of DNA Origami is a particular flavor of structural DNA nanotechnology. It was developed by Paul Rothemund at Caltech, he published this in 2006, and the basic idea is as follows: so imagine you have a long single strand of DNA, the 7000 base genome of the M13 bacteriophage, that’s the gray strand in this animation. We know what the sequence is, and based on that known sequence, we chemically synthesize 100’s of short oligonucleotides that are 20-60 bases long that are programmed by Watson-Crick complementarity to pinch that long strand into a parallel array of helices, after heating everything up to about 65°C and then cooling down to room temperature over the course of an hour. At the end of the assembly, you end up with this parallel array of double helices where adjacent double helices are held together by these Holliday junction crossovers that I described to you a couple slides ago. So this is a half-crossover and then here we have a full DNA crossover. Importantly, what you just saw was an animation, not a simulation. In fact, we have a very poor understand of the order of events of folding these objects, we just know that if we program them in a way where all of the scaffold ends up basepaired to staple strands, then we have an extremely high probability of forming the desired structure. So it’s a very active area of research for us to try to understand better the mechanism of folding, and we hope that will actually help us to design more complex structures in the future. So Paul Rothemund used this method in 2006 to make structures such as this ‘disc with three holes’ that has dimensions of about 100 nm x 100 nm x 2 nm, this is an atomic force micrograph. The example in the upper left-hand corner represents, in size, just part of the upper lip of the object. So this is quite large by macromolecular standards. It’s like we have two ribosomes worth of molecular Silly Putty that we can mash into any desired two-dimensional cookie cutter shape. One of the very interesting things that he pioneered was that he developed a way to make this DNA origami where he made each one of the staple strands in two different flavors. So one flavor just made the structure as you saw. The second flavor had the identical sequence, but had a surface feature, a dumbbell that’s sticking out of one of its ends. And so what that means is any time he used the original flavor and he added it to the folding mix, then you’d get a plain vanilla DNA Origami surface at that location. But then if he replaced that sequence with the longer sequence, the one with the feature, now you get that same shape, but a bump over that feature. And in that way, he conceived that this rectangular DNA Origami could be treated as a molecular breadboard, where let’s say it has 200 different positions, we can decide at each position whether we want to create a bump or have no bump. In effect we have something that’s like a bitmap, and we can create new patterns simply by repipetting different patterns of the no-bump and plus-bump strands for each one of the locations. So for example, here we can see that he’s designing something that will say ‘DNA’ and have a little picture of DNA. These structures actually become very sticky at the ends because they have lots of blunt ends, and then they’ll make a continuous ribbon that says ‘DNA’. You can see that he made a map of the Americas. He’s a very humble guy, so he apologized to the rest of the world for stopping at the Americas, but DNA is a little bit expensive, so he stopped at the moment. Maybe by now he’s made the rest of the world. And you can program them to link up in specific ways, and in that way you can self-assemble two-dimensional crystalline objects. So what about getting to three dimensions, as I alluded? Well, we can get our initial inspiration from macroscale paper Origami, where we’re quite familiar that if we fold flat paper in many ways we can get quite intricate three-dimensional shapes. So this is the famous crane. And if you’re really diabolical, like Robert Lang, you might note that if you can fold these papers in especially intricate ways, then you can make incredibly complicated objects, that we can see some examples of here. Now, nothing I’m going to show you with DNA is as complicated as this, but again, as I mentioned, one of our goals is to scale to ever-increasing complexity, so we hope that someday we actually can self-assemble DNA into objects of this kind of complexity. So that group in Denmark that I just mentioned, of Andersen, Jørgen Kjems, Kurt Gothelf, they were able to design that M13 to fold into six different sheets, and then they programmed those six sheets to fold up into a three-dimensional box with a hollow inside. They designed a lid that can open in response to some kind of molecular key. So this was the first example of a three-dimensional hollow DNA Origami. So where my group wanted to contribute was to make solid three-dimensional Origami structures, and the idea is as follows: so first of all, we know that we can curl up DNA due to the helicity of the DNA helices, and I’m going to go through a little thought experiment just to give you a flavor of what this is about. So here we have, on the far left, three double helices that are arranged into a little DNA Origami. You can see, if you look closely, they’re connected by those Holliday junction crossovers to keep the helices parallel. And in this arrangement it’s making a flat sheet of three helices. So now imagine what would happen if we moved these crossovers on the top two base pairs to the left. Then that’s going to move that double helix behind the plane of the page. And likewise, if we move those two crossovers two base pairs to the right, that’s going to move that double helix in front of the plane of the page. And the take-home message here is that simply by shifting around the position of those crossovers with respect to each other, we can achieve curvature of these DNA Origami sheets along the axis of the double helices. So that’s the first key. So now let’s extend that and build an actual solid 3-dimensional structure. So here we have another representation of a DNA Origami where each one of these cylinders represents one of those double helices, so it’s similar to the example in the upper left, but now just rotated into this orientation. So this would represent the pattern of the scaffold running through those helices, but for the purpose of this explanation, I’m going to leave that invisible. It’s there, but I’m just not going to talk about it, that or the staple strands. And so what we’re going to do is we’re going to shift around the position of those crossovers so now these helices no longer prefer to be planar, but instead prefer to curl up into some kind of specific geometry. And in this example what we’re doing is we’re trying to curl up the structure into a corrugated S shape. Furthermore, anywhere where we have the orange that touches the white sheet that touches the blue sheet, we’re routing those staple strands through those interfaces. So for example, we might have a staple strand that starts 7 base pairs on this helix, and then goes 7 base pairs here, 7 base pairs, 7 base pairs, 7, 7 base pairs. And in that way, if the structure forms the way we intend it to, it should be highly crosslinked by these staple strands that are traversing the different helices. So it looks good on paper, okay, what happens in the test tube when we tried it? And perhaps we can say, “Of course, when we threw all the strands together and tried to fold the object, then it didn’t work.” We got a pile of molecular spaghetti that we could see under the electron micrograph. But we didn’t want to give up, and eventually Hendrik Dietz in the group came up with a key insight, which is it’s not that these 3-dimesional objects now are unstable thermodynamically, simply they’re more difficult to achieve kinetically. And so what we found is that we could only get appreciable yields of these objects if we folded them instead of for an hour, from 65°C to room temperature, if we folded them for more like a week, then we could start to get appreciable yields of the objects ranging from 10-50% yield. So we can see here one of the objects that was built by Shawn Douglas. Instead of 3 layers, it was with 10 layers. And then we have the electron micrographs below. We can see that we get a close resemblance between what we observe in the electron micrograph and the projection orientations of our designed structure. This is work that we published back in 2009, in the meantime, our group and others have been hard at work trying to improve the method. So the important thing here was that we could get something to fold at all, and now we’re trying to get better yields, improve the folding times. So there have been a couple of important discoveries since then. One has come from Hendrik Dietz’s lab in Munich, where they’ve discovered that these structures tend to have a favored temperature at which they fold faster than the other temperatures. So instead of spending the same amount of time at 65°C down to room temperature, for example this structure maybe folds faster at 50°C. So what they found is if they do most of their folding at 50°C then they can get it fold maybe an order of magnitude faster, which makes a lot of improvements for our lives as scientists designing them, they also suffer less thermal damage with a slower folding ramp. We’ve also learned some details about how to design the strands, the crossovers, the breakpoints, that I don’t have time to go into in this presentation, but I encourage you to look at some of our publications if you want to see the latest discoveries in how to make this process work better. So now I’m going to go through a panel, a gallery of different objects built using this method by our laboratory to give you a flavor of the generality of the method. S o the example on the top is what I just showed you, we call it the “Monolith”, it was built by Shawn Douglas. You might say that it looks a little bit like a nanoscale crystal, honeycomb array crystal, but it’s important to keep in mind that every element of the object is associated with a unique sequence and therefore is independently addressable. This is quite different from most nanoparticles that we see in synthetic nanotechnology today. The example on the bottom was built by Franziska Graf, we call it the “Genie Bottle”. We called it that because one version, not shown here, we only folded part of the M13 scaffold and the rest of it was coming out of the lip of the object kind of like wisps of smoke. These are all 20 nm scale bars. So here again, 20 nm scale bars, on the left is an object built by Shawn Douglas, we call it the “Square Nut”. It has a 7 nm hole in the middle, it has a front end and a back end, and if we make the sticky ends on the front end compatible with the sticky ends on the back end, then we can self-assemble filaments that are somewhat reminiscent of actin filaments and microtubules, although in this case they don’t yet demonstrate any dynamics. They’re just equilibrium formation of these long polymers. On the right is an object built by Tim Liedl, we call it the “Railed Bridge”. Again, every cylinder is one double helix, and we can see as we go through cross-sections of the object we have a different arrangement of double helices, and we can understand from this example that it is kind of analogous to sculpture. That you could imagine the sculptor begins with a solid block of marble, in our case these parallel arrays of double helices, and in design space we’re chipping away at that solid block to achieve whatever 3-dimensional structure we actually want in relief. Once we have our final design, then what we’re doing is we’re compiling that 3-dimensional structure into a series of DNA strands that are going to self-assemble with the M13 scaffold into that object. Here’s an object built by Björn Högberg when he was in the group, we called it the “Slotted Cross”, I’ll have more to say about this object in the next slide. This is another crossed object we called the “Stacked Cross”, built by Hendrik Dietz. Again, these are all 20 nm scale bars. This one looks a little bit like stacked molecular celery. We even designed a little molecular cavity on the top where we initially imagined we could host protein guests on the inside of that cavity. So let’s take a closer look at that “Slotted Cross” from Björn Högberg. So here what he’s done is he’s generated an animation where he’s stylized the routing of the scaffold strand through the structure. It’s designed as an “H-domain” and an “O-domain”, and the middle of the H-domain is designed to pass through the middle of the O-domain and it’s all folding from just one long M13 scaffold. I was quite amazed that this folded at all, but the yields are not so great at the moment, just a few percent. So now what I’m doing is I’m zooming in on what we call the strand diagram that describes the blueprint of the object. It’s like we take all the helices and then we splay them out onto a 2-dimensional surface. And in this case the blue represents that M13 scaffold strand and those colored strands represent the staple strands. And this part of the object is the upper left-hand corner of the H-domain. And so if you look closely you can see that the staple strands, what they’re doing is they’re binding to part of the scaffold strand and then they’re crossing over to a different part of the scaffold strand to pull those components together to make the 3-dimesional shape. We can zoom out, and then here you can get an appreciation that it is kind of like a blueprint. You can make out which part is the H-domain , which part is the O-domain, and if you look closely you can actually see where the H-domain and O-domain are being connected by that long scaffold strand. All of the examples that I’ve shown you so far have been built using this honeycomb lattice paradigm where we’re using these corrugated sheets. It turns out that it more naturally fits the preferred twist of DNA at 10.5 basepairs per turn, but it turns out we can also self-assemble these DNA sheets in a square lattice format. The only proviso is that now we have to underwind the DNA to 10.67 basepairs per turn, which is slightly disfavored. And, quite interesting, what happens is the structure will still form, but it then compensates by having a global supertwist in the right-handed direction. So it’s quite analogous to how plasmid DNA, for example, will have a right-handed supertwist when it’s underwound, as we find in most cells. One very important development in the field is software with a graphical user interface to make it accessible to people who are outside the field, but also just to make the process faster, more robust and convenient, for experienced practitioners. So for this really powerful software suite called “cadnano”, we owe our thanks to Shawn Douglas, he developed this software when he was a graduate student in my group, now he’s an assistant professor at UCSF, at the time of this filming. So I encourage you to check out the software, he’s continually improving it, at cadnano.org. And what we can do, now again with the graphical user interface, within an hour or so, we can design different shapes and then compile that into the sequence of DNA strands that can self-assemble into that object. What if you wanted to build larger structures? Well, the most obvious idea is to just get more parts. So you can remember as a kid, the first time you got a Lego set it was enthralling, but then about two hours later, you now were hungry for additional Lego pieces. So that’s the big drive for our field: can we get more Lego pieces into the structure? But in the meantime we can do other things that will allow us to get a little bit bigger. So one example is just to build with wireframes, that have high strength-to-weight. So in this example what we’ve done is we’ve added staple strands that fold that M13 scaffold into this wireframe structure. Each one of these struts in this example is a 6-helix DNA nanotube, and then we designed sticky ends such that they’re compatible, and we can get this structure, this Z-shaped structure, to fold into a double triangle, with now 10 of these 6-helix bundle termini, each with a unique set of sticky ends. In this example, what we did is we programmed 3 separate double triangles to form in 3 separate test tubes, and we programmed it to form this network on the bottom. This is a Schlegel diagram, and for those of you who might recognize this, you might see that this is actually a Schlegel diagram for a wireframe icosahedron. This object has an overall diameter of about 100 nm, each one of the struts has a length of about 45 nm. And here on the lower left-hand corner we can see an animation, macroscale animation, reenactment, of the self-assembly of these double triangles into a wireframe icosahedron. What we find is that this process works in the test tube as well – no hands required. So again, what we do is we fold each of the double triangles in three separate test tubes, we then mix them together to form the desired wireframe object. So let’s take a look using electron microscopy. So here we see with a 1 micron (um) scale bar, we see a bunch of objects that seem to have about the right size, about 100 nm in diameter. There’s aggregates as well, so the self-assembly’s not perfect, but we’re glass-half-full kind of folk, we’re encouraged by something that works even partially. So now we’ve zoomed in, you have a 500 nm scale bar, and we can tell that there’s some kind of wireframe action going on. Zoom in some more, now we have a 200 nm scale bar, and it’s starting to look like the wireframe structure that we imagined. Of course you have some mis-assemblies as well. And then now if we go to the highest effective magnification for this negative stain method, 100 nm scale bar, we can see the objects, in fact, they look like they have lots of these triangular faces, they look like they have 5-fold vertices. And we’re able to make an object that now is something like five times the mass of a ribosome, it has overall dimensions the size of a medium size virus. And this is all just powered by Watson-Crick base pairing: A pairs with T, C pairs with G. It’s remarkable that we can push it this far, ~but we’re greedy and we dream about being able to extend this to objects that are a thousand times more complex or even more than that some day. Another kind of wireframe structure from macroscale engineering that inspired us are these floating compression sculptures from the artist Ken Snelson. And the idea here for these sculptures is that you have these beams, that are bearing compression, that aren’t touching each other directly, but instead they’re connected by cables that are bearing tension. And if you wire this up in the correct way, then it’s a balance between the tension of the cables and the compression of the beams, and you end up with an object that has high strength-to-weight and has other interesting features. For example, if those cables have some elasticity, then if you put a global force on the object, then it will deform, and every individual strut will rearrange in 3-dimensional space. When you now relieve that stress, then it’ll bounce back to the original shape. So we wanted to see if we could implement this using DNA Origami. This is work that was led by Tim Liedl and Björn Högberg when they were in the group, in collaboration with Don Ingber. So what they did was to design the staple strands to fold this M13 scaffold into 3 different struts, each of the struts in this case is 13 helices. It’s actually grabbing 3 separate segments of that scaffold in order to make each one of those 13-helix struts. So we again add everything together, heat it up, cool it down, and remarkably enough you can form structures like this in the test tube. In fact, we started to play games about looking at how much stress we could put the objects under and have them still fold. So what happens is that you have these single-stranded DNA elements that are acting like entropic coils – they’re exerting tension. And if we simply design those cables to be shorter, have fewer number of bases, then it’s going to exert a larger force over the same design distance between the two compressed elements. And what we found by continually shortening these cables is that we could self-assemble the structures up to about 14 piconewtons of force, that was the calculated force for the shortest cable that were able to self-assemble the objects. In other words, we’re able to self-assemble these DNA objects against twice the force that can be generated by powerful cytoskeletal motors such as kinesin or myosin. This is all powered by just DNA base pairing. We believe that these kind of structures may prove useful for applications in tissue engineering and regenerative medicine. So of course cell biologists have noted for a while now that cells, especially going through development, can communicate with their outside environment, with each other, using mechanics. So they might pull on the extracellular matrix and that extracellular matrix may pull back, and you might have, by introducing deformations into the extracellular matrix or within the cytoskeleton of the cell, you can trigger biochemical events. So we envision a day where we can use these kind of DNA nanostructures that can deform in response to some kind of mechanical stress and then translate into a biochemical event, it could be release of a growth factor, or maybe it could involve catalysis of some kind of chemical reaction. So we believe that this could be useful for regenerative medicine. So the last thing that I’d like to show you for this section is work from Hendrik Dietz, Shawn Douglas assisted on this work. Everything that I’ve shown you thus far has involved double helices that are straight. And Hendrik wanted to ask the question, “Could you build structures, curved structures, where the helices now are following an arc, instead of going straight?” And the basic strategy for implementing this is as follows: so here we have, again every cylinder represents a double helix, these planes that are slicing through the double helices represent the positions at which those crossovers are occurring. So it turns out in this example they’re only occurring every 7 basepairs. And he asked the question, well, what would happen if he replaced the double helical segments on the top, so the orange segments, with shorter double helices that only have 6 base pairs between planes. And what if he replaced the helices on the bottom, the blue ones, instead of 7 basepair segments, he had 8 basepair segments. So mechanically, now on the top, those elements are going to be under tension because you have less material in the same amount of space, they’re going to be stretched out. The helices are the bottom are going to be under compression, because we now just stuffed more material into the same amount of space. And the system is under stress and so it’s going to relax, of course, by bending. So this is the way to relieve that tension on the top and compression on the bottom. Does this actually work when we attempt this in the test tube? And the answer is yes. So Hendrik implemented this, using an 18 helix DNA bundle that’s illustrated on the upper left-hand side. And so what he did was he had a stereotyped straight region, these white regions, and then he had an experimental region that’s indicated here in red. So that’s where he’s going to be introducing those longer and shorter elements to induce the bending of the structure. You can see for the control you get this nice rigid straight object. So what happens when he introduces some small number of shorter strands in the double helices on top, and then longer ones on the bottom. He could get a reliably predicted 30 degree arc at that position. If he has roughly twice the number of perturbations, then you can get to a 60 degree angle. Kept on going, you get 90 degree, you can get a 120 degree angle, that’s quite remarkable. This is now getting down to a 10 nm radius of curvature. But then he kept on going, and he found he could go all the way to 180 degrees in this example. So this is something that has a 6 nm radius of curvature , it’s comparable to the tightness of wrapping of DNA double helices around histones in a nucleosome. So in that case that’s powered by protein-DNA interactions, in our case this is powered by DNA base pairing interactions. So here what we have is an animation prepared by Shawn that explains the bending principle. So again, what we’re doing is we’re introducing more basepairs, or long double helices on the left, and then shorter ones on the right. And you can see a little graph on the lower left-hand side that tells us how many basepairs per turn that we have for each of these different elements. And at the most extreme example we’re actually getting 15 basepairs per turn on the left, which is severely underwound, and only six basepairs per turn on the right, which is severely overwound. And I was quite flabbergasted that it should be possible for us to torture DNA to this extent. Now in fact once you get to those extremes, our folding yields do start to go down, so we can see that we’re at the edge of what we can do to DNA, but still it’s quite remarkable that DNA is so robust, that the 10.5 basepairs per turn is simply what it prefers to do, but if you put enough stress on it, you can make it do things that deviate from that ideal by quite a bit. So Hendrik and Shawn now used the method to make a variety of different structures. So on the upper left-hand corner we have a 6-helix DNA bundle that’s folded into a series of 180 degree arcs of increasing radii of curvature, so you make a spiral. On the lower left-hand corner we have an object that’s programmed to self-assemble into a beach ball, out of 6-helix bundles. You can see objects that are making concave triangles, this is designed by Shawn Douglas. And then here we have those 120 degree arcs that are repurposed, so we made sticky ends on the two ends of this little boomerang to be complementary, so that you can have three identical versions of them will come together to make a larger triangle. So in conclusion, hopefully I’ve persuaded you that DNA Origami is a highly versatile method for building both 2-dimensional and 3-dimensional structures of quite remarkable complexity, about twice the mass of a ribosome. Where we’re moving to next is to try to build structures that are more complicated. You might wonder what’s preventing us from building something 1000 times larger already, today. And the main problem is that we have errors in the self-assembly. For example, for one of these objects, we might have a yield, in the best cases, 75% of so, which might sound pretty good. But now if you wanted to build an object that’s 1000 times bigger, then you might argue that the probability, if you just mixed these things together, 1000 of them together, the probability that none of the 1000 would have any defect, would be 0.75^1000, which if you do the math, that’s basically zero. So there’s a lot of activity in the field trying to improve the fidelity of this self-assembly, other methods like hierarchical self-assembly, error correction, that’ll allow us to scale amount complexity and build really very complicated objects of the future.