Happy to share our new preprint characterizing a recent mutation in SARS-CoV-2 Spike protein (D614G) and its potential functional impact on the ability of the virus to transduce human cells.
First, this work was spearheaded by a talented postdoc in our group #ZharkoDaniloski and is a wonderful example of curiosity-based science.
In April, we learned that one particular mutation had become dominant among COVID-19 cases worldwide. Here’s a nice visualization using the Nextstrain website to see this mutation in the Spike protein appearing around February: 
Given its rapid spread and Spike’s key role in enabling viral entry into target cells, we wondered whether this mutation might enhance the Spike protein’s ability to transduce human cells. So, Zharko tested it.
Our group relies heavily on VSV-g pseudotyped lentiviruses for gene delivery in human cells and we were happy to see that there is a long history of Spike-pseudotyped lentiviruses, e.g. Deng & colleagues (2004) with SARS-CoV-1.
We packaged a EGFP-encoding lentivirus with no pseudotyping, a pseudotype with the Spike D614 variant (same as in the original sequence from Wuhan) or a pseudotype with the Spike G614 variant.
To our surprise, the Spike G614 variant increased transduction of human lung, liver and colon cells at every dose of virus that we tested with up to an 8-fold increase in viral transduction.
This was true both in cells that artifically overexpress the primary entry receptor (ACE2) and also non-ACE2 overexpressing cells.
The next question was, How does this single amino acid variant drive such a major change in transduction efficacy? For this, some details about our current understanding of how the Spike protein works are needed.
The SARS-CoV-2 Spike protein has two major regions (S1 & S2) with a protease cleavage site separating them. S1 binds to the cellular receptor (e.g. ACE2) while S2 enables fusion of viral and host membranes.
Here’s a 3D structure of the Spike protein with the S1/S2 cleavage sites labeled (taken from Hoffman et al., Molecular Cell, 2020):
When we expressed the Spike protein in human cells, we found that the G614 variant was more resistant to cleavage. On this gel, we can visualize the full-length Spike and also the S2 and S2’ cleavage fragments.
What is S2’? There are actually two cleavage events that need to occur for the virus to enter. After binding of ACE2, the S1/S2 cleavage exposes another cleavage site (S2’). S2' cleavage reveals a hydrophobic fusion polypetide which inserts into the host membrane.
Using a C-terminus epitope-tagged Spike, we found that the G614 variant seems less prone to cleavage. Even though cleavage is required during entry, cleavage before entry can result in a Spike that can no longer bind ACE2.
Previous work showed the host protease furin is likely responsible for the first (S1/S2) cleavage event. So, we immunoprecipitated each Spike variant and performed in vitro digestion with several furin concentrations. The G614 variant is less readily cleaved by furin in vitro:
Taken together, these experiments imply that the greater fraction of uncleaved Spike with the G614 variant may allow each newly-assembled virion to include more receptor binding-capable Spike protein.
In addition to the Spike G614 functional data in cell culture, we wondered what the impact of the mutation was on COVID-19 patients. There have been several studies looking into this but no clear consensus (e.g. Korber et al. 2020, Dorp et al. 2020, Bhattacharyya et al. 2020) .
Inspired by earlier work by Tim Cardozo @nyuniversity, we looked at the relationship between prevalence of the Spike G614 variant and the per-country case-fatality rate using ~22,000 viral genomes. We found a modest but significant positive correlation. Each dot is 1 country.
We also noticed that qPCR data from COVID-19 patients from two independent studies (from @bettekorber1 and from @trvrb) consistently detects 3-fold more viral RNA in patients with the Spike G614 variant. Perhaps there is a biological effect of the Spike variant in COVID patients?
In summary, we demonstrated that the SARS-CoV-2 Spike D614G variant increases transduction of the virus across a broad range of human cell types, including cells from lung, liver and colon. And G614 appears to be more resistant to proteolytic cleavage.
And perhaps most reassuringly, Hyeryun Choe, Michael Farzan and colleagues recently found a similar result of increased transduction with G614 using a pseudotyped retroviral (MLV) system.
Their preprint with this elegant work is here (and includes several other exciting experiments):
biorxiv.org/content/10.110…
biorxiv.org/content/10.110…
One important caveat of our work is that we use a pseudotyped lentivirus model, which has a different virion assembly pathway. Future work with isogenic SARS-CoV-2 strains (differing at D614G) are needed to bolster the functional differences seen in the pseudotyped virions.
Also, although there are interesting correlations with case-fatality rates and viral load in COVID-19 patients, it is still unclear whether the Spike variant has an impact on disease severity or real-world viral transmission.
It has been a so much fun to work with Zharko & @CathyxGuo on this project and there is much we still don’t know about D614G & other Spike mutations. We are also grateful for guidance from expert virologists @virusninja & #TristanJordan @MountSinaiNYC & @eghedin @nyuniversity 🙏
