CRISPR-Cas9 genetic screens reveal regulation of TMPRSS2 by the Elongin BC-VHL complex

Abstract

The TMPRSS2 cell surface protease is used by a broad range of respiratory viruses to facilitate entry into target cells. Together with ACE2, TMPRSS2 represents a key factor for SARS-CoV-2 infection, as TMPRSS2 mediates cleavage of viral spike protein, enabling direct fusion of the viral envelope with the host cell membrane. Since the start of the COVID-19 pandemic, TMPRSS2 has gained attention as a therapeutic target for protease inhibitors which would inhibit SARS-CoV-2 infection, but little is known about TMPRSS2 regulation, particularly in cell types physiologically relevant for SARS-CoV-2 infection. Here, we performed an unbiased genome-wide CRISPR-Cas9 library screen, together with a library targeted at epigenetic modifiers and transcriptional regulators, to identify cellular factors that modulate cell surface expression of TMPRSS2 in human colon epithelial cells. We find that endogenous TMPRSS2 is regulated by the Elongin BC-VHL complex and HIF transcription factors. Depletion of Elongin B or treatment of cells with PHD inhibitors resulted in downregulation of TMPRSS2 and inhibition of SARS-CoV-2 infection. We show that TMPRSS2 is still utilised by SARS-CoV-2 Omicron variants for entry into colonic epithelial cells. Our study enhances our understanding of the regulation of endogenous surface TMPRSS2 in cells physiologically relevant to SARS-CoV-2 infection.

Keywords: CRISPR-Cas9 screen; Colon epithelial cells; Coronavirus entry factors; Hypoxic regulation of surface proteins; Transmembrane serine proteases.

Fig. 1

Expression of TMPRSS2 and ACE2 at the surface of human cell lines. (A) Calu-3, RT4, LNCaP, Colo-205, CL-40 and Caco-2 cell lines were stained with the TMPRSS2-specific antibody or secondary antibody alone as a control and analysed by flow cytometry. (B) Cell lines from the panel 1A were stained with ACE2-specific antibody or secondary antibody alone as a control and analysed by flow cytometry. See also Figure S1.

Fig. 2

TMPRSS2 KO affects SARS-CoV-2 entry into colon epithelial Caco-2-ACE2 cells. (A) Caco-2-ACE2 Cas9 cells stably expressing sgRNAs specific for b2m, ACE2 or TMPRSS2 were stained with ACE2- and TMPRSS2-specific antibodies or secondary antibody alone and analysed by flow cytometry. (B) Caco-2-ACE2 Cas9 cells from the panel 2A were infected with rSARS-CoV-2 Venus or Omicron BA.2 at an MOI of 0.1 or 1, fixed 24 h later and analysed by automated microscopy. Y-axis indicates percentage of GFP (rSARS-CoV-2 Venus) or N-protein (BA.2)-positive cells. Data are presented as mean of n = 3 biological replicates ± s.d. The statistical significance was assessed by two-way ANOVA and Bonferroni’s multiple comparison correction. See also Figures S2, S3 and S4.

Fig. 3

Genome-wide and targeted CRISPR-Cas9 screens reveal Elongin B and VHL as regulators of TMPRSS2 expression. (A) Schematic workflow of the genome-wide and targeted CRISPR-Cas9 screens. Caco-2 cells were transduced with lentivirus encoding CRISPR library followed by selection with puromycin. The populations of cells with low TMPRSS2 expression were enriched by FACS with the TMPRSS2-specific antibody, subjected to genomic DNA isolation followed by Illumina sequencing of the integrated gRNAs. (B, C) MAGecK RRA (robust rank aggregation) scores of the enriched sgRNAs from the CRISPR-Cas9 screens performed with the genome-wide sgRNA library (B) and sgRNA library targeting Epigenetic Modifiers and Transcriptional Regulators (C).

Fig. 4

Surface TMPRSS2 is regulated by Elongin B and VHL in Caco-2 cells. (A) Caco-2 Cas9 cells stably expressing sgRNAs targeting ELOB, VHL and b2m were stained with the TMPRSS2-specific antibody or secondary antibody alone as a control and analysed by flow cytometry. (B) Caco2 Cas9 cells from the panel 4A were lysed and analysed by immunoblot with the antibody specific for VHL and b-actin. (C) Caco-2, LNCAP and RT4 cells were treated with FG-4592 (roxadustat), stained with the TMPRSS2-specific antibody and analysed by flow cytometry. See also Figures S5 and S6.

Fig. 5

Surface TMPRSS2 is downregulated in a HIF-dependent manner. (A) Caco-2 Cas9 cells stably expressing sgRNAs targeting HIF1b (ARNT), HIF2a (EPAS1) and b2m were treated with FG-4592 (roxadustat) for 7 days, harvested and stained with the TMPRSS2-specific antibody or secondary antibody alone as a control and analysed by flow cytometry. (B, C) Immunoblot analysis of Caco2 Cas9 cells from the panel 5A. The cells were lysed and analysed by immunoblot with antibodies specific for HIF1b, HIF2a and b-actin. Asterisks denote non-specific bands. See also Figure S7.

Fig. 6

CRISPR Cas9-mediated depletion of Elongin B or PHD inhibitor treatment decrease SARS-CoV-2 infection of Calu-3 cells. (A) Calu-3 Cas9 cells stably expressing sgRNAs targeting b2m, Elongin B (TCEB2) or control sgRNAs were lysed and analysed by immunoblot with antibodies specific for HIF1⍺, HIF2⍺, Elongin B, VHL and b-actin. Asterisks denote non-specific bands. (B) Calu-3 Cas9 cells stably expressing two independent pairs of sgRNAs targeting Elongin B (TCEB2) or control sgRNAs were infected with rSARS-CoV-2 Venus at an MOI of 0.1, harvested 24 h later and subjected to RNA extraction followed by RT-qPCR analysis with the primers specific for SARS-CoV-2 nucleocapsid RNA and 18S. Data are presented as mean of n = 3 technical replicates ± s.d. The statistical significance was assessed by unpaired two-tailed t test. (C) Calu-3 cells were treated with 100 uM FG-4592 (roxadustat) or DMSO as a control for 72 h, infected with rSARS-CoV-2 Venus or Omicron BA.2 at an MOI of 0.1, harvested 24 h later and subjected to RNA extraction followed by RT-qPCR analysis with the primers specific for SARS-CoV-2 nucleocapsid RNA and 18S. Data are presented as mean of n = 3 technical replicates ± s.d. The statistical significance was assessed by unpaired two-tailed t test. (D) Depletion of Elongin B induces downregulation of ACE2 and TMPRSS2 in Calu-3 Cas9 cells. The cells from the panel 6A (right) were subjected to RNA isolation followed by RT-qPCR analysis with primers specific for TMPRSS2, ACE2, PGK1, CA9 and 18S. Data are presented as mean of n = 3 technical replicates ± s.d. The statistical significance was assessed by unpaired two-tailed t test. (E) PHD inhibitor treatment induces downregulation of ACE2 and TMPRSS2 in Calu-3 cells. The cells were treated with 100 uM FG-4592 (roxadustat) for 72 h and subjected to RT-qPCR analysis as described above (6D).

Fig. 7

Proposed mechanisms of TMPRSS2 regulation by the Elongin BC-VHL complex. Under normal physiological conditions, constitutively expressed HIF⍺ proteins are hydroxylated by prolyl hydroxylase domain (PHD) enzymes, recognised by the Elongin BC-VHL complex, ubiquitinated and targeted for proteasomal degradation. Under conditions of hypoxia, treatment with PHD inhibitors or depletion of the Elongin BC-VHL complex, the stabilised HIF⍺ protein heterodimerises with HIFβ and inhibits expression of TMPRSS2 either (A) through direct binding to the promoter or indirectly through induction of expression of a putative protein repressor (B) or microRNA (C). *It should be noted that since no hypoxia response elements were identified in the TMPRSS2 promoter, the regulation of TMPRSS2 expression will likely occur as described in scenarios (B) or (C).