Frozen Origami

Revealing the structures of megadalton-scale DNA complexes with nucleotide resolution - Massimo Kube, Fabian Kohler, Elija Feigl, Baki Nagel-Yueksel, Elena Marie Willner, Jonas Jörg Funke, Thomas Gerling, Pierre Marcel Stömmer, Maximilian Nicolas Honemann, Thomas Martin, Sjors Scheres, HD

Over the last decades, DNA nanoengineers have created ever more sophisticated molecular objects and devices. Surprisingly, much of the progress in the field has been achieved without ever seeing in detail the three-dimensional structures of the objects thus created. Could single-particle cryo EM become the lens of DNA nanoengineers? More than twelve years ago I started working with Andres Leschziner at Harvard University on solving cryo EM structures of 3D DNA Origami objects, which I had made as a PostDoc in William Shih’s lab. It was hopeless! The objects we made at the time were of low quality and far too dilute. Nonetheless, Andres kindly agreed to train me in the basics of cryo EM. The seed was thus planted to establish cryo EM as a tool for structural analysis of DNA origami.

In 2009, I started my own lab at Technische Universität München (TUM) located in Munich, Germany. I acquired a used 100kV Philips electron microscope about to be decommissioned – our first imaging system with which we could do at least negative-stain TEM imaging. We worked hard on making better DNA origami and more of them. Then I went looking for collaborators willing to perform cryo EM structural analysis. In general, it proved difficult to convince structural biologists to devote some of their precious beamtime to outlandish samples such as DNA origami.

Sjors Scheres at the Medical Research Council Laboratory of Molecular Biology (MRC LMB) in Cambridge, UK, saw some value in our line of work. He agreed to join forces on developing DNA-origami-based alignment supports to facilitate the analysis of guest proteins by cryo EM. Thomas Martin, one of the co-authors of the present work, had joined my lab as a PhD student in 2010 and began working on this project. In 2012, the first fruits of our joint work with Sjors appeared, in the form of the first cryo-EM 3D structure of a DNA origami, known as The Pointer, with data acquired at the cryo-EM facilities at the MRC LMB and with help from Xiaochen Bai, a PostDoc working with Sjors. For the first time we were able to see a DNA origami in 3D at a resolution where we could discern some helical detail such as grooves and all the crossovers. The pointer became a reference for computational prediction models like ENRG MD and OxDNA. Thomas went to do a PostDoc with Sjors and four years later we published our second paper together, on a prototype DNA origami support for structure determination of guest proteins.

By then, Sjors had become slightly disenchanted with DNA origami. We did not really succeed in improving resolution beyond what we achieved already in 2012. Given the frenzy for beamtime at the LMB Krios’ machines it was quite clear we needed our own cryo EM facilities and a proper team of microscopists to move forward. Needless to say, it was a major effort for me to build alliances and to convince funders and my host university to invest in cryo EM systems worth millions of Euros. In 2016, a Titan Krios system began its operation at TUM under the supervision of my team of cryo EM novices. Master’s students and young PhD students rolled up their sleeves and taught themselves how to do vitrification, data acquisition and processing, with patient advice here and there from Cambridge.

In the meantime, we also learned how to prepare high-quality DNA origami routinely at micro molar concentrations which finally allowed us to collect data on many different objects in free standing ice. Hence, we could also determine cryo EM maps of many different objects – which is reflected in the first part of the paper, led by Massimo Kube. It helped us appreciating phenomena such as twist deformations and how they relate to design details. Another advance that enabled us to finally move toward the sub 5Å resolution range was appreciating and dealing with conformational heterogeneity. This effort was led by Fabian Kohler. Sjors and his co-workers had developed multibody refinement methods to deal with exactly such phenomena in protein structures – we adapted it and came up with what we call “scanning focussed refinement”, where a region of interest is scanned over the structure to be determined. Such detailed analysis is the subject of the second part of the paper. The third part is about constructing atomic models to help interpreting the cryo EM maps, a tedious job had Elija Feigl not managed to tame it with a good degree of automation. The stage is set to use cryo EM as a lens for DNA nanoengineers: We can now look routinely at DNA origami with much improved detail and modify them until we are satisfied. Of course, the quest for ever higher resolution will go on.