Samuel A. McKie gained experience with structural biology and protein modelling through work on polypharmacological drug design, during both under- and post-graduate studies. Later publishing work on contrasting the ribosomal structures of bacteria, archaea, and mammals towards rational antibiotic drug candidates. His efforts during the pandemic hiatus have focused on a 3D printing start-up, making anatomic, molecular, and topographical 3D models. This led to the research and production of a series of physical SARS-CoV-2 models.
For more information, visit www.chapelprints.co.uk (Facebbok | Instagram) or contact chapelprints@gmail.com.
Those that use the Protein Data Bank appreciate the critical role biomolecular models have for both research and education. Protein, DNA, and organic compound models have long been understood to offer an engaging way to educate. However, even with modern digital renderings and a wealth of open access to information, awareness of key concepts of structural biology is not yet common knowledge. Many do not truly grasp that their bodies are biological machines, and that diseases and medical treatments have rational, mechanistic explanations underlying their function. I believe original and engaging models are an important means to dispel disinformation and irrational thinking. The need to improve public knowledge about the structural biology of diseases and treatments is of course highlighted by the COVID-19 pandemic.
At times, the current pandemic has shown how a poor understanding of a disease can have broad effects on public health, from compliance with lockdown measures, to vaccine refusal, and even outright disease denialism (1). In response to the limited number of educational models depicting SARS-CoV-2, we have created several up-to-date renderings of the COVID-19 virion. Here I discuss the methods used in their creation, how they fit alongside other learning resources, and the pressing need for better general understanding of structural biology, particularly for that of SARS-CoV-2.
When considering the Covid-19 virus, the public mainly see 2D and 3D models presented as illustrations or animations. Actual micrographs are rarely used, and perhaps considered too unrefined. In the first six months of the pandemic, most models were inaccurate and crude; naturally limited by the research available. After the release of the initial and iconic illustration created by Alissa Eckert, and Dan Higgins for the US Centers for Disease Control and Prevention (CDC), one of the first scientifically constructed illustrations (2) was created by Veronica Falconieri Hays (who outlined her excellent work in a previous PDB-101 Education Corner article). By autumn 2020, all structural proteins of SARS-CoV-2 had been modelled either directly via experimental techniques or through in silico methods. Yet even now, many media outlets still use limited or inaccurate graphics to illustrate and discuss COVID-19.
More creative and modern techniques of depicting the virus are available but have not been widely utilised. To my knowledge our physical SARS-CoV-2 virion model is one of only four so far produced. The US National Institutes of Health produced a 3D printed model early in the pandemic that was used by Dr Fauci's office in many presentations. The model now resides at the Smithsonian's National Museum of American History; an updated version is available to download for those with 3D printers (3). A collective known as the Coronavirus Structural Task Force created a version that includes features suggesting Envelope and Membrane proteins (4). This model is more user-friendly for use with commonly available fused deposition modelling (FDM) 3D printers. Finally, Biological Models based in Brooklyn, have released a commercially available model of the virion and a collection of SARS-CoV-2 protein models (5).
For all the advantages of a 3D printed model, access is not an option for many individuals. Origami models are recognized as a widely available method of visualizing biomolecular structures–especially for younger audiences. As with the availability of different 3D print models, only a handful of paper models for the SARS-CoV-2 virion can be found online at this time of writing. Victoria Scott has produced a freely downloadable 3D model and another pseudo 3D option is commercially available from Origami Organelles (6-7). However, both lack visualizations of certain viral details. Our origami model gives a third option that depicts some of the viral proteins in more detail.
Our Origami COVID-19 Model. Selected steps of the model’s creation are shown. A digital download of the origami sheet and instruction booklet is available, with a video demonstration on YouTube
By early 2021, given the formidable amount of literature and impressive number of protein models elucidating the inner workings of SARS-CoV-2, we undertook construction of a digital molecular model of the SARS-CoV-2 virion. We search the PDB for key SARS-CoV-2 proteins. The mature virion contains only four types of protein structure: The Spike, Membrane, Envelope and Nucleocapsid proteins (8). In addition, the Spike protein has at least four major conformations (9).
Reconstruction of the membrane bound tail of the Spike protein trimer was greatly assisted by the excellent models from Charmm-GUI (10). Some discrepancies over the orientation of the post-fusion Spike protein were resolved through interrogation of the structure and sequence. Importantly, an NMR structure from late 2020 formed the Envelope protein unit (11). Notably, the Membrane protein remains unresolved. For this structure, AlphaFold’s ab initio dimeric protein model was used (12). We limited our model to depict only the exterior of the virion, as the arrangement of the Nucleocapsid proteins and viral RNA was still uncertain in the literature during the research phase of the project.
To construct the viral envelope, the online MemGen tool was used along with six key lipids: POPC, POPS, POPE, Cholesterol, POPI, and Cardiolipin (13-14). These lipids were chosen based on the membrane character of the ER–Golgi intermediate compartment, this where the SARS-CoV-2 virion buds (15). The MemGen tool produced a 70 by 70 angstrom lipid bilayer tile. We arranged 480 of these tiles manually using Chimera (protein editor and visualizer; 16) to produce an approximated sphere measuring roughly 90 nanometres in diameter. Unassigned gaps between tiles represented around 22% of the model’s surface area, these were filled appropriately with partial lipid tiles.
The scientific literature was reviewed to determine current estimates for the number of viral proteins per virion. Cryo-electron microscopy studies were particularly helpful in revealing SARS-CoV-2 has around 30 Spike proteins – mostly in the open or closed conformation (7-18). We believe many visualisations overpopulate the surface with Spike and Envelope proteins. Numerous studies suggest the virion includes only a handful of Envelope proteins, and between 300-1000 Membrane proteins (19-209). For our model, 25 Spike trimers and 5 Envelope protein units were randomly assigned to non-adjacent tiles, with all 450 remaining unoccupied tiles assigned Membrane protein units. Using Discovery Studio (3ds.com/biovia), the protein structures were inset manually into the lipid bilayer based on surface hydrophobicity profiles.
The resulting model contains over 7 million atoms, around 1000 protein macromolecules, and 480 PDB structures. This giant PDB file type had to be saved as multiple files. It was rendered into a printable STL file type via several programs - Discovery Studio (3ds.com/biovia), Rhino (rhino3d.com), Meshlab (meshlab.net), and MeshMixer (meshmixer.com). The resulting digital object was used to create a printable model of the SARS-CoV-2 virion as well as other models focusing on the viral proteins. Images taken from the virion model’s surface were used as the basis for the origami print.
Our digital macromolecule model can be used for 2D illustrations of the virion’s surface, depicting the relative size of protein types and differences in the Spike protein conformations. To our knowledge, this is the only attempt at combining the various PDB entries into one molecular model of the entire virion. With further work, this could be used as the basis for 3D visualisations and even more novel Virtual Reality models. The digital molecular model is available as a free download from our site at Chapel Prints. The printable model can be found on the popular 3D model sharing site Thingiverse.
Accurate physical models of the virus and its proteins offer unique tools for education and engagement. The fact the Spike proteins transition between closed, open, double-open, and post-fusion, highlights their mechanical nature. What the origami model lacks in detail and structural strength it makes up for in accessibility.
Available as a free digital download, anyone with a printer, paper, cocktail sticks and perhaps a little super glue can create a physical visualisation of the agent behind the pandemic. As a learning aid for a class, the focus required to fold the kit reframes the lesson from a passive theoretical exercise to a real practical problem.
Over the last 18 months a seemingly endless parade of articles, media segments, and discussions have rightfully focused on one of the largest global health crises in living memory. Although the public been made familiar with the statistics of the pandemic, concepts surrounding the biology of infection and how the numerous vaccines against it work must also be brought into the general knowledge of the public.
Inaccurate illustrations and oversimplified cartoons of the Covid-19 virus dominate in the collective imagination. Physical models of the virus have remained absent from printed or televised media, at least here in the United Kingdom. At a time when many local institutions possess 3D printers, why haven’t all students and curious members of the public had a chance to view and hold the agent behind the misery of the last couple of years? A tangible physical model has the power to make the topic more real, magnifying the weight of evidence and succinctly summarising what is known about this unseeable virus.
The problem does not stem from a lack of information. Already, more than 1400 scientifically determined models of SARS-CoV-2 proteins have been made available in the PDB. Many freely downloadable resources aimed at educating students and the public do exist, with RCSB PDB a leading example of these efforts. But there seems to be a lack of demand and difficulties in disseminating accurate scientific material aimed at the general reader; this despite a massive shift towards democratising scientific information. Perhaps our methods of engaging the public with current academic knowledge require updating? Maybe we need to further utilize new technology and acknowledge new insidious threats from misinformation? I believe for molecular biology, physical models are an important part of finding novel ways to utilize the PDB library and educate the public.