Multiplex measurements of the impacts of ACE2 sequence and cell surface abundance during SARS-like coronavirus entry

ABSTRACT

SARS-like coronaviruses, including SARS-CoV and SARS-CoV-2, encode spike proteins that bind human ACE2 protein on the cell surface to enter target cells and cause infection. The efficiency of virus entry depends on ACE2 sequence and expression levels in target cells. A small fraction of humans encodes variants of ACE2, thus altering the biochemical properties at the protein interaction interface. All humans possess cells with vastly differing amounts of ACE2 on the cell surface, ranging from cell types with high expression in the gut and lungs to lower expression in the liver and pancreas. Mastering our understanding of spike-ACE2 interaction and infection requires experiments precisely perturbing both variables. Thus, we developed a synthetic cell engineering approach compatible with high throughput assays for pseudo-typed virus infection. Our assay system is capable of assessing both variables individually and in combination. We adapted an engineered HEK293T DNA recombinase landing pad cell line capable of expressing transgenic ACE2 sequences at highly precise levels. Infection with lentiviruses pseudotyped with the spikes of SARS-like coronaviruses revealed that high ACE2 abundance could mask the effects of impaired binding thereby making it challenging to know the role of affinity altering mutations during infection. We limited the ACE2 abundance on the cell surface by expressing transgenic ACE2 behind a suboptimal Kozak sequence, thereby altering its protein translation rate. This allowed us to understand how ACE2 sequence could impact its interaction with coronavirus spike proteins as two human ACE2 variants at the binding interface, K31D and D355N, exhibited reduced infection. Our experiments suggested that we need to better understand how ACE2 expression determines the susceptibility of cells for SARS-like coronavirus binding and infection. We thus created an ACE2 Kozak library consisting of ~4,096 Kozak variants, each conferring a different ACE2 protein translation rate thus resulting in a range of ACE2 steady-state abundances. Combining fluorescence-activated cell sorting and high-throughput DNA sequencing (FACS-seq) revealed the library to span two orders of magnitude of ACE2 abundance. Challenging this library of cells with spike pseudotyped lentiviruses revealed how ACE2 abundance correlated with infection rate. The library-based experiments yielded a dynamic range wider than traditional single sample infection assay, likely more representative of infection dynamics in vivo. Now that we have characterized the impacts of ACE2 abundance on infectivity in engineered cells, our next goal is to expand the comparison to physiologically relevant cells with endogenously expressed proteins. Modulating protein abundance levels will be key to creating maximally informative functional assays for any protein in cell-based assays, and we have laid the groundwork for being able to simultaneously test the impacts of protein abundance and sequence in combination for proteins involved in diverse cellular processes. This research was supported by a National Institute of Health (NIH) grant GM142886 (KAM).Copyright © 2023 The American Society for Biochemistry and Molecular Biology, Inc.