SARS-CoV-2 intra-host single-nucleotide variants associated with disease severity

Abstract

Abstract Variants of severe acute respiratory syndrome coronavirus 2 frequently arise within infected individuals. Here, we explored the level and pattern of intra-host viral diversity in association with disease severity. Then, we analyzed information underlying these nucleotide changes to infer the impetus including mutational signatures and immune selection from neutralizing antibody or T-cell recognition. From 23 January to 31 March 2020, a set of cross-sectional samples were collected from individuals with homogeneous founder virus regardless of disease severity. Intra-host single-nucleotide variants (iSNVs) were enumerated using deep sequencing. Human leukocyte antigen (HLA) alleles were genotyped by Sanger sequencing. Medical records were collected and reviewed by attending physicians. A total of 836 iSNVs (3–106 per sample) were identified and distributed in a highly individualized pattern. The number of iSNVs paced with infection duration peaked within days and declined thereafter. These iSNVs did not stochastically arise due to a strong bias toward C > U/G > A and U > C/A > G substitutions in reciprocal proportion with escalating disease severity. Eight nonsynonymous iSNVs in the receptor-binding domain could escape from neutralization, and eighteen iSNVs were significantly associated with specific HLA alleles. The level and pattern of iSNVs reflect the in vivo viral–host interaction and the disease pathogenesis.