Biochemistry of SARS-CoV-2 Spike Protein and its Ocular Surface Membrane Receptor

National Eye Institute

Investigators

Linked publications & trials

Abstract

Novel coronavirus SARS-CoV-2, originating at the end of 2019 in Wuhan, China, causes the pandemic coronavirus disease COVID-19 that has had a major public health and economic impact in the USA and the rest of the world since then. The disease is quite heterogeneous and targets internal organs with many possible complications, morbidity, and significant world-wide mortality. Furthermore, the SARS-CoV-2 virus can target the eye causing viral conjunctivitis as was described in early cases in Wuhan. It has been estimated that 1/8 of COVID-19 cases have some form of ocular involvement, making it a subject of interest to vision research and to the NEI. In the course of its evolution, SARS-CoV-2 Spike glycoprotein (S protein) acquired a novel 4 amino acid insert -PRRA- at residues 681-684, absent in other lineage B -CoVs such as SARS-CoV, that is encoded by a novel 12-base RNA sequence which contains tandem rare codons. Our fundamental hypothesis was that this RNA sequence constitutes a ribosomal pausing site, with properties similar to premature stop codons. Alternatively, such sites may be involved in pausing, or parsing, of translation of large multi-domain proteins (such as S protein) to allow for proper folding of successive domains. This -PRRA- site is a furin protease cleavage site that also plays a major role in the virulence of SARS-CoV-2. Our mutagenesis experiments suggest that the insert may create a double-edged weapon (a combination of overlapping furin and translation pausing sites) that has allowed SARS-CoV-2 to infect its new host (human) more readily. This underlines the importance of ribosome pausing to allow efficient regulation of protein expression and, also, of co-translational subdomain folding. These results were published in the prior reporting period. Complicating the issue has been the appearance of SARS-CoV-2 variants, including mutations at the furin site (-HRRA- in alpha variant and -RRRA- in delta variant). The dominant current Omicron variants have a -HRRA- furin site. In addition, the Omicron variants contain the N679K mutation in S protein. We also wish to determine which receptors SARS-CoV-2 uses to enter ocular cells, as they appear to be different than those on other cells, such as lung cells. In the past year, we have been investigating the mechanism of entry into ocular cells by lentiviral particles pseudotyped with SARS-CoV-2 spike protein, finding that caveolae-mediated endocytosis using LDLR is the pathway for SARS-CoV-2 virus internalization in the ocular cell line ARPE-19. We have made the following progress: We found that, while Angiotensin-converting enzyme 2 (ACE2) is expressed in ARPE-19 cells, blocking ACE2 by antibody treatment did not prevent infection by SARS-CoV-2 spike pseudovirions, nor did antibody blockade of extracellular vimentin and other cholesterol-rich lipid raft proteins. Next, we implicated the role of cholesterol homeostasis in infection by showing that incubating cells with different cyclodextrins and oxysterol 25-hydroxycholesterol (25-HC) inhibits pseudovirion infection of ARPE-19. However, the effect of 25-HC is likely not via cholesterol biosynthesis, as incubation with lovastatin did not appreciably affect infection. Additionally, we determined that it was not likely to be an agonistic effect of 25-HC on LXR receptors, as the LXR agonist GW3965 had no significant effect on infection of ARPE-19 cells at up to 5 micromolar GW3965. We probed whether endocytic pathways were implicated but determined that clathrin-dependent and flotillin-dependent rafts were not involved. Furthermore, 20 micromolar chlorpromazine, an inhibitor of clathrin-mediated endocytosis (CME), also had little effect. In contrast, anti-dynamin I/II antibodies blocked the entry of SARS-CoV-2 spike pseudovirions, as did dynasore, a noncompetitive inhibitor of dynamin GTPase activity. Additionally, anti-caveolin-1 antibodies significantly blocked spike pseudotyped lentiviral infection of ARPE-19. However, nystatin, a classic inhibitor of caveolae-dependent endocytosis, did not affect infection while indomethacin inhibited only at 10 micromolar at the 48 h time point. Finally, we found that anti-LDLR antibodies block pseudovirion infection to a similar degree as anti-caveolin-1 and anti-dynamin I/II antibodies, while transfection with LDLR-specific siRNA led to a decrease in spike pseudotyped lentiviral infection, compared to scrambled control siRNAs. Thus, we concluded that SARS-CoV-2 spike pseudovirion infection in ARPE-19 cells is a dynamin-dependent process that is primarily mediated by LDLR. Additionally, in collaboration with Dr. A.V. Bocharov (NIH-CC) and Thomas Eggerman (NIH-CC), we tested our spike-pseudotyped lentiviruses in LDLR-stable transfected HeLa cell lines. This work was submitted for publication and was published in this reporting period. Our next step is to test our hypothesis on whole eye human retinal organoids generated from H9 embryonal stem cells (ESCs). We have found that organoids are susceptible to spike-pseudotyped lentivirus infection at 2 months but not after 3 months of differentiation. We will look at colocalization of virus and LDLR receptor on the surface of organoids by immunofluorescence microscopy and try to block infection of organoids with anti-LDLR, anti-dynamin, and anti-caveolin 1 antibodies. The dominant current Omicron variants have a -HRRA- furin site. In addition, the Omicron variants contain the N679K mutation in S protein. We have synthesized the Omicron version of the furin site in original S protein and will use this construct to generate lentiviral pseudovirions for infection of ARPE-19 cells and SEAM organoids. We wish to learn if the furin site is important for LDLR receptor recognition and internalization of S protein. These studies are being done in collaboration with Dr. Tim Blenkinsop, Icahn School of Medicine at Mount Sinai, NYC. This work is ongoing.