Periodic Reporting for period 1 - OPT-ASSEMBLY (Optimal design of frustrated self assembly building blocks)
Okres sprawozdawczy: 2022-09-01 do 2024-08-31
Many nanostructures in nature form via the headless self assembly of elements, sometimes orders of magnitudes smaller than the resulting structure. Examples include viral protein capsids, bacterial microcompartments and other shelled organelles. These structures are diverse in shape and functionality, yet, they all form from a few types of building blocks, spontaneously, governed by pure thermodynamics. Nature's remarkable success in economically and robustly assembling such structures of various sizes,
shapes and functionalities has long puzzled scientists and was envied by material engineers, and with
good reasons: the ability to assemble synthetic nanostructures gives rise to a series of technological applications including drug delivery, virus trapping and photonic applications. A relatively recent method to implement such self assembling building blocks is DNA origami, where DNA strands are carefully designed and folded into triangular blocks which, in turn, can attach to each other and self assemble into the pre-programmed shape. In many cases, however, the assembly process goes off-pathway resulting in off-target structures which are useless from the technological standpoint. Finding the optimal design principles in creating building blocks which assemble into the desired target with high fidelity was the main objective of the project. While collaborating closely with experimentalists, the project itself was purely theoretical and computational and we used a coarse grained Monte Carlo simulation scheme developed by us to explore various design principles and guide the experiments. Inspired by viral capsids, initially we focused on creating small, icosahedral-shaped shells from a single subunit species, then soon transitioned to build higher complexity shells with more technological interest.
shapes and functionalities has long puzzled scientists and was envied by material engineers, and with
good reasons: the ability to assemble synthetic nanostructures gives rise to a series of technological applications including drug delivery, virus trapping and photonic applications. A relatively recent method to implement such self assembling building blocks is DNA origami, where DNA strands are carefully designed and folded into triangular blocks which, in turn, can attach to each other and self assemble into the pre-programmed shape. In many cases, however, the assembly process goes off-pathway resulting in off-target structures which are useless from the technological standpoint. Finding the optimal design principles in creating building blocks which assemble into the desired target with high fidelity was the main objective of the project. While collaborating closely with experimentalists, the project itself was purely theoretical and computational and we used a coarse grained Monte Carlo simulation scheme developed by us to explore various design principles and guide the experiments. Inspired by viral capsids, initially we focused on creating small, icosahedral-shaped shells from a single subunit species, then soon transitioned to build higher complexity shells with more technological interest.
We developed, debugged and optimized the aforementioned Monte Carlo simulation method for the self assembly of triangular subunits. The simulation was key in achieving our objectives because it only follows the assembly of a single structure making it orders of magnitudes faster than other available methods. This increased performance allowed us to explore many complex design principles. Assembly yields can be improved by two major ways: i) by optimizing the assembly conditions and ii) by increased subunit complexity (i.e. using more subunit species).
Initially we attempted to follow i). Using a single subunit type, our aim was to optimize the assembly conditions and achieve larger and larger icosahedral shells. We used advanced Bayesian optimization techniques and achieved fairly high yields up to a T4 structure (an icosahedral shell of 80 subunits). The method, however, was not scalable as we found that yields degrade for larger structures. Recent developments in experiments, however, allow for the use of a fairly large number of subunit species, unfortunately, accompanied by poor control over the subunit compliance. Motivated by this advancement, we chose to follow a combination of route i) and ii), namely to identify the optimal design which provides high yields with minimal complexity and is the most robust to the assembly conditions.
After adjusting the simulation to support multiple species, we targeted the assembly of triply periodic minimal surfaces (TPMS) which may be used as photonic crystals. For the subunit design we used symmetry principles and developed the equivalent geometric construction to viral capsids where subunits in equivalent symmetry environments are of the same species. We tested this design against the assembly conditions for the three species of known TPMS and found that intermediate subunit flexibility is required for high fidelity and fast self assembly. Additionally, we identified the main defect mechanism leading to off-target assemblies, namely the formation of disclinations which may be eliminated either by larger subunit rigidity or the increase of subunit complexity. Such structures have since been assembled experimentally as well. Our results are published in PNAS, https://doi.org/10.1073/pnas.2315648121(odnośnik otworzy się w nowym oknie)
After the success in TPMS assembly, we moved back to icosahedral shell assembly. Our aim was to extend subunit complexity beyond the symmetry requirements and identify the economical designs which give high yields at a wider range of assembly conditions. We developed a construction which allows for the gradual increase in complexity and we tested the performance of each of those. We found that more complex designs are not necessarily better and that they only give a better yield if they help eliminating defects. A manuscript summarizing these results is currently in progress.
Our results have been presented and discussed at multiple scientific conferences and workshops including ones for the broader public: Middle European Cooperation in Statistical Physics (2023 and 2024), Non-equilibrium Physics of Self-Assembly: from Viruses to Nano-containers (2023, won the best poster award), Bad Honnef Physics School on Physics of Viruses (2023), Bolyai Society Interdisciplinary Dialogue Conference (2023). We held several seminars at UBB Cluj, CIFRA Bucharest, Brandeis Waltham and ESPCI Paris. Additionally, we took part in organizing the Saturday of Experiments at UBB (2023 and 2024) and presented our work to the broad public at other events like the Hungarian Cultural Days in Cluj (2023 and 2024).
Initially we attempted to follow i). Using a single subunit type, our aim was to optimize the assembly conditions and achieve larger and larger icosahedral shells. We used advanced Bayesian optimization techniques and achieved fairly high yields up to a T4 structure (an icosahedral shell of 80 subunits). The method, however, was not scalable as we found that yields degrade for larger structures. Recent developments in experiments, however, allow for the use of a fairly large number of subunit species, unfortunately, accompanied by poor control over the subunit compliance. Motivated by this advancement, we chose to follow a combination of route i) and ii), namely to identify the optimal design which provides high yields with minimal complexity and is the most robust to the assembly conditions.
After adjusting the simulation to support multiple species, we targeted the assembly of triply periodic minimal surfaces (TPMS) which may be used as photonic crystals. For the subunit design we used symmetry principles and developed the equivalent geometric construction to viral capsids where subunits in equivalent symmetry environments are of the same species. We tested this design against the assembly conditions for the three species of known TPMS and found that intermediate subunit flexibility is required for high fidelity and fast self assembly. Additionally, we identified the main defect mechanism leading to off-target assemblies, namely the formation of disclinations which may be eliminated either by larger subunit rigidity or the increase of subunit complexity. Such structures have since been assembled experimentally as well. Our results are published in PNAS, https://doi.org/10.1073/pnas.2315648121(odnośnik otworzy się w nowym oknie)
After the success in TPMS assembly, we moved back to icosahedral shell assembly. Our aim was to extend subunit complexity beyond the symmetry requirements and identify the economical designs which give high yields at a wider range of assembly conditions. We developed a construction which allows for the gradual increase in complexity and we tested the performance of each of those. We found that more complex designs are not necessarily better and that they only give a better yield if they help eliminating defects. A manuscript summarizing these results is currently in progress.
Our results have been presented and discussed at multiple scientific conferences and workshops including ones for the broader public: Middle European Cooperation in Statistical Physics (2023 and 2024), Non-equilibrium Physics of Self-Assembly: from Viruses to Nano-containers (2023, won the best poster award), Bad Honnef Physics School on Physics of Viruses (2023), Bolyai Society Interdisciplinary Dialogue Conference (2023). We held several seminars at UBB Cluj, CIFRA Bucharest, Brandeis Waltham and ESPCI Paris. Additionally, we took part in organizing the Saturday of Experiments at UBB (2023 and 2024) and presented our work to the broad public at other events like the Hungarian Cultural Days in Cluj (2023 and 2024).
The uniqueness of the project stems from the fact that, while the project itself was theoretical, collaboration with experimentalists allowed to identify and pursue the most relevant questions and test the answers right away. Combining symmetry considerations and efficient computer simulations, we developed design principles for triangular subunits which self assemble into TPMSs and icosahedral shells. Given the underlying lattice structure, we believe that these principles will be straightforward to adapt to other partially closed target structures (eg. helices, spherocylinders or ribbons). The power of constructing the desired nanostructures cannot be overestimated when it comes to potential applications. TPMSs could help building photonic crystals with carefully tailored optical properties. Icosahedral shells have already been proven to be successful in trapping viruses, but may also be used as nanocontainers or nanoreactors where the concentration of reagents is held at the required level.