Can a single amino acid mutation in the spike protein affect the infectivity and immunogenicity of SARS-CoV-2?
Recently, a great deal of attention has become focused on a specific SARS-CoV2 mutant in which amino acid residue 614 of the spike protein is changed from aspartic acid to the less bulky and more neutral glycine. This mutant, D614G, reported by Korber et al., has become increasingly predominant after first seemingly proliferating in Europe, then spreading rapidly elsewhere. Korber et al. suggested the mutant is more infectious, based on higher viral RNA titers from patients who were infected with it and its rapid prevalence, but it does not appear to be more pathogenic, based upon the clinical pictures of the patients. Residue 614 is not in the receptor binding domain. They suggested that the mechanism of enhanced infectivity could be due to glycine not forming a hydrogen bond with the neighboring spike protein subunit, allowing the subunits to dissociate more readily and thereby facilitating virion fusion with the cell membrane. There is a stunning virtual reality visualization of this on YouTube. Korber et al. also reported evidence for recombination between genomes carrying both this mutation as well as other mutations, indicating recombination and simultaneous infection of cells with more than one genotype. Although suggestive, no proof was presented for an actual increase in infectivity by the mutant.
Since that report, several other studies have addressed the issue of infectivity more directly. One caveat related to many of these reports is that they use lentiviral particles pseudo typed with coronavirus envelope proteins. Viral entry is measured by a co-transduced indicator gene, such as luciferase. Although this is thought to faithfully mimic coronaviral entry, it is an inherently artificial system. The general consensus of all these studies is that the G614 variant enters cells expressing the ACE2 receptor better than the D614 variant even though the variable residue is not in the receptor binding domain, that there is no difference in clinical outcome, that infection with the G614 variant results in higher viral RNA titers in nasal swabs, and that there is not a great deal of difference in antibody neutralization (which is good news for vaccine development).
Let’s consider the individual reports separately. Ozono et al. used the lentiviral pseudovirus method to sample five different naturally occurring mutations in the spike protein, including D614G, to characterize their behavior relative to the reference genotype. Their entry characteristics varied from having a lesser to a greater ability to enter cells expressing ACE2 and the protease TMPRSS2 cells, which greatly facilitates SARS-CoV2 entry. Significantly, the D614G mutant showed the most efficient entry. Interestingly, SARS-CoV1 was much more efficient at entry than was SARS-CoV2. They performed an in silico structural analysis that suggested that the SARS-CoV1 spike trimer has a more open configuration that would result in greater accessibility to the ACE2 receptor by the receptor binding domain. They also tested COVID 19 antisera from patients infected with the D614 variant, and showed no detectable differences in neutralization between the D614 and G614 variants.
Hu et al. also used a lentiviral pseudovirus system to analyze the D614G variants. As with Korber et al., they found the G614 variant to be globally distributed. Like Ozono et al., they found about a 2.5-fold increase in entry efficiency by the G614 variant, perhaps due to more efficient protease cleavage of its spike protein. Unlike Ozono et al., they found that a minority of COVID 18 antisera failed to neutralize the G614 variant to the same extent as the D614 variant. However, it is not clear with which variants the serum donors were infected.
Wagner et al. used a more natural but messier approach. They looked at viral loads, as measured by RT-PCR, and clinical status of patients in Washington state infected with either the G614 or D614 variants. They found that patients infected with the G614 variant had higher nasal viral RNA loads, but did not have a more severe clinical picture. The age of the patients infected with G614 skewed slightly younger (~3 years). They also found that G614 became increasingly more prevalent in Washington state over time.
Lorenzo-Redondo et al. (1) reported on patients in Chicago infected with one of what they called three clades of SARS-CoV2. Clade 1, which was introduced from Washington, contains the G614 phenotype, while clades 2 and 3 have the D614 phenotype. The origin of clade 2 was ascribed to Illinois, while clade 3 was introduced from New York. Clade 1 had higher viral loads than clade 2, in agreement with Wagner et al. Interestingly, when bronchial alveolar lavage samples were tested, there was little difference in viral RNA titers between the two clades, suggesting that the increased titers were specific to upper airway tissue. This could be a factor increased spread.
It should be pointed out that all the above results are in the form of preprints. In addition, the methods used to measure entry directly are somewhat artificial. However, taken together directly from patient data, it seems that G614 may in fact be more capable of spreading, perhaps because of more facile entry into cells, perhaps due to better proteolytic processing because of a more open quarternary structure. Fortunately, this mutation does not seem to worsen the clinical outcome of infection, nor does it seem to abrogate recognition by most neutralizing antibodies.
- R. Lorenzo-Redondo et al., A Unique Clade of SARS-CoV-2 Viruses is Associated with Lower Viral Loads in Patient Upper Airways. medRxiv, (2020).
- Tang, Leyan & Schulkins, Allison & Chen, Chun-Nan & Deshayes, Kurt & Kenney, John. (2020). The SARS-CoV-2 Spike Protein D614G Mutation Shows Increasing Dominance and May Confer a Structural Advantage to the Furin Cleavage Domain. 10.20944/preprints202005.0407.v1.
- Grubaugh, N.D., Hanage, W.P., Rasmussen, A.L., Making sense of mutation: what D614G means for the COVID-19 pandemic remains unclear, Cell (2020), doi: https:// doi.org/10.1016/j.cell.2020.06.040.