Structural epitope profiling identifies antibodies associated with critical COVID-19 and long COVID

Even within a single protein, antibody binding can have beneficial, neutral, or harmful effects during the response to infection. Resolving a polyclonal antibody repertoire across a pathogen’s proteome to specific epitopes may therefore explain much of the heterogeneity in susceptibility to infectious disease. However, the three-dimensional nature of antibody-epitope interactions makes the discovery of non-obvious targets challenging. We implemented a novel computational method and synthetic biology pipeline for identifying epitopes that are functionally important in the SARS-CoV-2 proteome and identified an IgM-dominant response to an exposed Membrane protein epitope which to our knowledge is the strongest correlate of severe disease identified to date (adjusted OR 72.14, 95% CI: 9.71 – 1300.15), stronger even than the exponential association of severe disease with age. We also identify persistence (> 2 years) of this IgM response in individuals with longCOVID, and a correlation with fatigue and depression symptom burden. The repetitive arrangement of this epitope and the pattern of isotype class switching is consistent with this being a previously unrecognized T independent antigen. These findings point to a coronavirus host-pathogen interaction characteristic of severe virus driven immune pathology. This epitope is a promising vaccine and therapeutic target as it is highly conserved through SARS-CoV-2 variant evolution in humans to date and in related coronaviruses (e.g. SARS-CoV), showing far less evolutionary plasticity than targets on the Spike protein. This provides a promising biomarker for longCOVID and a target to complement Spike-directed vaccination which could broaden humoral protection from severe or persistent disease or novel coronavirus spillovers.