Cleavage and activation of the severe acute respiratory syndrome coronavirus spike protein by human airway trypsin-like protease

Abstract

The highly pathogenic severe acute respiratory syndrome coronavirus (SARS-CoV) poses a constant threat to human health. The viral spike protein (SARS-S) mediates host cell entry and is a potential target for antiviral intervention. Activation of SARS-S by host cell proteases is essential for SARS-CoV infectivity but remains incompletely understood. Here, we analyzed the role of the type II transmembrane serine proteases (TTSPs) human airway trypsin-like protease (HAT) and transmembrane protease, serine 2 (TMPRSS2), in SARS-S activation. We found that HAT activates SARS-S in the context of surrogate systems and authentic SARS-CoV infection and is coexpressed with the viral receptor angiotensin-converting enzyme 2 (ACE2) in bronchial epithelial cells and pneumocytes. HAT cleaved SARS-S at R667, as determined by mutagenesis and mass spectrometry, and activated SARS-S for cell-cell fusion in cis and trans, while the related pulmonary protease TMPRSS2 cleaved SARS-S at multiple sites and activated SARS-S only in trans. However, TMPRSS2 but not HAT expression rendered SARS-S-driven virus-cell fusion independent of cathepsin activity, indicating that HAT and TMPRSS2 activate SARS-S differentially. Collectively, our results show that HAT cleaves and activates SARS-S and might support viral spread in patients.

Fig. 1.

HAT cleaves SARS-S. (A) Cleavage of SARS-S by HAT in cotransfected 293T cells. Expression plasmids coding for SARS-S and the proteases indicated or empty vector (pcDNA) were transiently cotransfected into 293T cells, which were then treated with trypsin or PBS. Subsequently, S-protein cleavage was detected by Western blot analysis of cell lysates using a serum specific for the S1 subunit of SARS-S. Detection of β-actin served as a loading control. S_Fl, full-length SARS-S; S1, S1 subunit of SARS-S; asterisks, SARS-S cleavage fragments generated by TMPRSS2. (B) HAT cleaves SARS-S at arginine 667. Plasmids encoding wild-type SARS-S or the SARS-S R667A mutation were transfected into 293T cells jointly with TMPRSS2, HAT expression plasmids, or empty vector (pcDNA). Subsequently, the cells were treated with PBS or trypsin, and SARS-S cleavage was analyzed by Western blot analysis of cell lysates, using an S1-specific antiserum. Expression of β-actin in cell lysates was assessed as a loading control. (C) Cleavage of recombinant SARS-S (rSARS-S) by recombinant HAT. Recombinant SARS-S was incubated with the indicated concentrations of recombinant HAT, and cleavage products were analyzed by Western blot analysis employing a SARS-S1-specific serum. Recombinant HAT was also detected. (D) Separation of SARS-S cleavage products for mass spectrometric analysis. SARS-S cleavage products were separated by gel electrophoresis and visualized by colloidal Coomassie staining. Major differential bands appearing upon HAT treatment (see apparent molecular mass annotations on the right) were subjected to in-gel digest with trypsin or Asp-N, followed by mass spectrometric analysis. Thereby, we identified the 130-kDa band to be S2 with its N terminus starting at S668 and the 95-kDa band to be S1 with its C terminus ending at R667. Untreated SARS-S (170-kDa band in the left lane) was processed in parallel as a control.

Fig. 2.

Mass spectrometric analysis of Asp-N-digested SARS-S cleavage products. (A and B) Zoomed-in peptide survey spectra of the Asp-N digests. The mass signals at m/z 1,644.00 (A) and m/z 2,043.29 (B) represent the N-terminal Asp-N fragment of S2 (668-STSQKSIVAYTMSLGA-683, M + H⁺ calculated = 1,643.83 Da) and the C-terminal Asp-N fragment of S1 (649-DIPIGAGICASYHTVSLLR-667, M + H⁺ calculated = 2,043.069 Da with carboxamidomethylated Cys), respectively, as they were exclusively detected in the corresponding fractions. Note that the mass signal at m/z 1,659.99 in panel A represents the Met-oxidized variant of 668-STSQKSIVAYTMSLGA-683 (+16 mass units). (C and D) Mass spectrometric sequencing of the terminal Asp-N fragments. (C) The fragment ion mass spectrum of m/z 1,644 clearly confirmed the identity of the peptide to be 668-STSQKSIVAYTMSLGA-683 through a conclusive N-terminal b-ion series. The predominant occurrence of b-type ions is in agreement with the charge localization at Lys in position 5, and only this ion series is annotated for the sake of clarity (P, precursor). (D) The fragment ion mass spectrum of m/z 2,043 clearly confirmed the identity of the peptide to be 649-DIPIGAGICASYHTVSLLR-667 through a conclusive C-terminal y-ion series. The predominant occurrence of y-type ions is in agreement with the charge localization at the C-terminal Arg, and only this ion series is annotated for the sake of clarity. Mascot database searches of both fragment ion mass spectra against the Swiss-Prot database (with enzyme set to Arg-C plus Asp-N to mimic the cleavage scenario) identified spike glycoprotein (Swiss-Prot database accession number P59594) with significant MS/MS ion scores. a.u., arbitrary units.

Fig. 3.

Expression of HAT does not induce SARS-S shedding. VLPs were produced in 293T cells by coexpression of HIV p55-Gag and SARS-S in the absence and presence of coexpressed TMPRSS2, HAT, or cotransfected empty vector. Supernatants were collected and subjected to ultrafiltration followed by ultracentrifugation through a 20% sucrose cushion. Afterwards, pellets and supernatants of the ultracentrifuge reactions were treated with trypsin or PBS and analyzed for the presence of S protein, using a serum specific for the S2 subunit of SARS-S. In parallel, the presence of HIV p55-Gag was determined. UC pellet, VLP preparation subjected to ultrafiltration followed by ultracentrifugation and analysis of the pellets; UC-Sup, VLP preparation subjected to ultrafiltration followed by ultracentrifugation and analysis of the supernatants of ultracentrifuge reactions; S_Fl, full-length SARS-S; S2, S2 subunit of SARS-S; asterisks, SARS-S cleavage fragments generated by TMPRSS2.

Fig. 4.

HAT activates SARS-S for cell-cell fusion in cis and in trans. (A) cis activation of SARS-S by HAT. 293T effector cells were cotransfected with pGAL4-VP16 expression plasmid, SARS-S plasmid, and plasmids encoding the indicated proteases or empty plasmid (pcDNA) and mixed with target cells cotransfected with ACE2 expression plasmid or empty plasmid and a plasmid encoding luciferase under the control of a promoter with five GAL4 binding sites. The cell mixtures were then treated with PBS or trypsin, and the luciferase activities in cell lysates were quantified at 48 h after cell mixing. The results of a representative experiment performed in triplicate are shown; error bars indicate standard deviations (SDs). Similar results were observed in an independent experiment. (B) trans activation of SARS-S by HAT. The cell-cell fusion assay was performed as described in the legend for panel A, but proteases were expressed in target cells. The results of a representative experiment performed in triplicate are shown and were confirmed in two separate experiments. Error bars indicate SDs.

Fig. 5.

Activation of SARS-S by HAT does not bypass the requirement for cathepsin activity for virus-cell fusion. (A) cis cleavage of SARS-S by HAT does not rescue SARS-S-driven infectious entry from blockade by cathepsin inhibitors. Lentiviral pseudotypes bearing SARS-S were generated in 293T cells coexpressing the proteases indicated or cotransfected with empty plasmid (pcDNA). The pseudotypes were used to infect 293T cells engineered to express high levels of ACE2 and preincubated with the indicated concentrations of the cathepsin B/L inhibitor MDL28170 or the lysosomotropic agent NH₄Cl. Luciferase activities in cell lysates were determined at 72 h postinfection. The results of a representative experiment performed in triplicate are shown; error bars indicate SDs. Comparable results were obtained in two independent experiments. (B) trans cleavage of SARS-S by HAT does not rescue SARS-S-driven infectious entry from blockade by cathepsin inhibitors. The experiment was carried out as described in the legend for panel A, but proteases were expressed in viral target cells. The results of a representative experiment performed in triplicate are shown and were confirmed in two separate experiments. Error bars indicate SDs.

Fig. 6.

HAT and ACE2 are coexpressed in human lung, and HAT activates the SARS coronavirus for cell-cell fusion. (A, left) Section of lung immunostained for HAT, using the peroxidase technique (brown), demonstrating strong expression by intrapulmonary bronchial epithelial cells (B) and alveolar macrophages (M). Weaker positive immunostaining suggested expression at a lower level by type 1 pneumocytes (P1; thin flat cells) and type 2 pneumocytes (P2; plumper cells). (A, right) Serial section of lung immunostained for ACE2, using the peroxidase technique (brown), demonstrating expression by intrapulmonary bronchial epithelial cells (B), alveolar macrophages (M), and type 2 pneumocytes (P2), although no expression of ACE2 by type 1 pneumocytes is seen. Bar, 50 μm (the bar pertains to both panels). (B) ACE2-expressing 293T cells were transfected with HAT, TMPRSS2, or TMPRSS4 expression plasmids or transfected with empty plasmid (control) and infected with SARS-CoV (Frankfurt-1; multiplicity of infection, 0.1). At 24 h postinfection, the cells were fixed with paraformaldehyde (8%) and permeabilized, and SARS-CoV antigen was detected by immunostaining using a human antiserum and a goat anti-human Cy3-labeled secondary antibody (red). In parallel, proteases were detected with mouse anti-HAT, mouse anti-TMPRSS2, or mouse anti-Myc (for Myc-tagged TMPRSS4) and a Cy2-conjugated goat anti-mouse antibody (green). Nuclei were stained with DAPI (blue). Arrows, examples of syncytium formation that were partially magnified (white squares). Bars, 25 μm. Similar results were obtained in an independent experiment.