The paper in Nature Microbiology can be found here:
After many years of looking at, and presenting, countless electron microscopy (EM) images of vaccinia virus (VACV) undergoing fusion with host cells I had a “sit up in bed moment” (see Figure 1).
The next day I scoured the literature looking for EM examples of poxvirus fusion. Turns out, there are only a couple dozen images, but over 95% of them showed virions fusing at their tips. The next day, Ricardo and I debated whether the fusion machinery would be located on one end of the virus, or on both ends.
I proceeded to stain VACV bound to a cover slip with an antibody directed against a component of the fusion machinery and took the sample to the structured illumination microscope (SIM). The SIM increases resolution by 2-fold compared to conventional microscopy allowing us to achieve visualization of sub-viral structures in VACV.
Figure 1. The eureka moment! PIs with Anxiety™
Seeing is believing
Looking at the samples, I realized that this was one of those rare amazing moments in science when you are the first person to see something completely novel. What I observed and discovered was that the poxvirus fusion machinery sits at both tips of individual virions. This clearly explained the fusion orientation bias observed in the EM images.
By the way, Ricardo won the debate. While it made biological sense that the fusion machinery would be on both poles - based on the bilateral symmetry of poxvirus particles - I was hoping the fusion machinery would localize to a single tip. This would have provided the missing “point of reference” for 3-D mapping of virus proteins.
SIM and STORM
Having developed VirusMapper, Robert Gray was the perfect person to run with this project. With a generous gift from Gary Cohen, who developed and passed along antibodies to the various EFC components, Rob generated localization models of the VACV entry fusion complexes (EFCs) by combining SIM and single particle averaging. Of note, all members of the EFC, which is comprised of 11 different proteins, were found at both tips of virus particles (Figure 2).
Figure 2. VirusMapper models of the average localization of various VACV EFC components.
To complement and extend this finding, David Albrecht applied single molecule localization microscopy. Looking at EFC components on individual virus particles, David confirmed that the EFCs reside at the tips of virions and showed that the EFCs were actually in clusters (Figure 3)
Figure 3. STORM imaging and reconstruction of the localization of EFC component L1 on a single virion.
A picture is worth a thousand words
Rob and David then began to investigate the localization and clustering of EFCs on VACV mutants that lack individual EFC components, so-called fusion mutants. They uncovered a perfect correlation between defective VACV fusion, EFC depolarization and disrupted EFC clustering. So, they drew a model (Figure 4).
Figure 4.Model of EFC localization and clustering on WT vs. EFC mutant viruses.
Long story short - Structure drives function
While we could correlate defective fusion with disrupted EFC polarization, up to this point we had only observed this when a component of the EFC was missing. We basically had a “chicken or egg causality dilemma”. We needed a way to disrupt the localization of the EFCs on VACV particles without breaking the fusion machinery itself.
Fortuitously, a VACV protein called A27 which was previously implicated in virus fusion, is not a component of the EFC. Our hope was that by removing A27 from the virus, we would disrupt EFC localization without affecting EFC function. This would allow us to determine the impact of EFC clustering and polarization on virus fusion. Lucky for us, this is exactly what happened. Deletion of A27 from the virus disrupted clustering and resulted in depolarization of EFCs from the tips of virus particles.
We analyzed the A27 mutant for hemi-fusion, the process by which the outer leaflets of virus and cell lipid bilayers mix, and full fusion, when virus and cell lipid bilayers completely merge to open a fusion pore. We found that A27-regulated clustering and polarization of VACV EFCs was critical for the process of full fusion.
Tip(s) of the iceberg
This study is the first direct demonstration that the position of virus fusion machinery on a virus particle is directly linked with its ability to drive efficient full fusion. This was of course only made possible by the use of Super-resolution microscopy which provided us with the opportunity to resolve the position of virus proteins, on single virions, within 10’s of nanometers, while maintaining molecular specificity.
Having now mapped 25 virion proteins with sub-viral resolution, the power of this technique is just coming to light. We can now begin to investigate virus assembly intermediates and define the changes in virus structure that drive maturation - at a molecular level - on single viruses and in infected cells. Hence, we may be sitting on the cusp of a new era in molecular virology where the ability to correlate changes in virus protein architecture with infectious virus assembly and disassembly can be used to inform anti-viral design targeted at disrupting virus meta-stability.
Written by Jason Mercer and edited by Ricardo Henriques on behalf of the other authors.
- Jason Mercer is an Associate Professor & MRC group leader at the MRC-LMCB, University College London, UK
- Ricardo Henriques is an Associate Professor at the the MRC-LMCB and Departmentof Cell and Developmental Biology, University College London, UK