Design and analysis of wireframe DNA nanostructures

Abstract

In the last decades, the powerful self-assembly properties of DNA have been harnessed to produce complex structures at the nanoscale with high precision and yield. DNA origami is one of the most robust examples of this, where a 7000-nucleotide strand of biological origin is folded by hybridizing with hundreds of synthetic oligonucleotides, the programmed sequence of these “staple strands” determines the shape of the assembled object. The long “scaffold strand” permeates every helix of the assembled object acting as a backbone, finding the path for the scaffold strand is trivial in designs where the helices are packed on a parallel lattice but becomes challenging in other designs.

In this thesis we expand the design space of DNA origami to wireframe structures based on polyhedral meshes by the introduction of a software package consisting of: a routing algorithm for finding A-trail Eulerian circuits, a rapid physical simulation for converting the mesh to a DNA design with low strain, and vHelix, a graphical user interface for manual modification of the structure and processing of DNA sequences (Paper I). We find that this method can produce wireframe DNA origami structures with refined shapes and features, and we investigate these structures with negative stainedand cryo electron microscopy. The helices in these structures are not packed on a tight lattice and we find that they can assemble and remain stable at physiological salt concentrations unlike previously demonstrated 3D DNA origami.

We then expand this method to two-dimensional sheets (Paper II), first by generating three rectangular sheets with different vertex geometries and investigating them with atomic force microscopy to find that six-armed vertices are needed for non-distorted structures. The geometry with six-arm vertices is then used to generate four flat sheets with complex internal and external features, these structures fold with high yield to their programmed shape.

It is apparent from electron microscopy that these structures are not as rigid as structures based on the parallel packing of helices. In Paper III we study the effect of design choices on the rigidity of wireframe structures, specifically on rods where the flexibility can be estimated by measuring the persistence length. In addition to experiments we use coarse-grained molecular dynamics simulations to evaluate the rigidity in silico. We find that the rigidity of rods increases with increasing number of facets in the cross-section, and that the breakpoints between staples negatively affects rigidity and that his effect can be reduced by enzymatic ligation. The simulations reveal that the rigidity of the structures is greatly reduced by increasing the salt concentration.

In Paper IV we further explore the power of coarse-grained molecular dynamics simulation to predict the dynamics of DNA nanostructures. First, we track the end to end distance of the helices of a structure throughout a simulation and find that the behavior varies greatly between helices where some are practically rigid and others show large deformations. We then implement this concept in an iterative fashion where a structures rigidity is estimated by simulation and from this first generation a number of mutant structures are created by modifying one or more edges. These mutant structures are then simulated and the effect of the modifications are measured on their adjacent helices or on the entire structure, and modifications that are beneficial are inherited in the next generation of the structure. Using these methods, we are able to create a moderate evolution towards a lower flexibility in wireframe DNA origami structures.