key: cord-0984843-anedg12x
authors: Bestle, Dorothea; Heindl, Miriam Ruth; Limburg, Hannah; van, Thuy Van Lam; Pilgram, Oliver; Moulton, Hong; Stein, David A.; Hardes, Kornelia; Eickmann, Markus; Dolnik, Olga; Rohde, Cornelius; Becker, Stephan; Klenk, Hans-Dieter; Garten, Wolfgang; Steinmetzer, Torsten; Böttcher-Friebertshäuser, Eva
title: TMPRSS2 and furin are both essential for proteolytic activation and spread of SARS-CoV-2 in human airway epithelial cells and provide promising drug targets
date: 2020-04-15
journal: bioRxiv
DOI: 10.1101/2020.04.15.042085
sha: c7a3ade5d2994efd9b2b29e94e874b822d560144
doc_id: 984843
cord_uid: anedg12x

In December 2019, a novel coronavirus named SARS-CoV-2 first reported in Wuhan, China, emerged and rapidly spread to numerous other countries globally, causing the current pandemic. SARS-CoV-2 causes acute infection of the respiratory tract (COVID-19) that can result in severe disease and lethality. Currently, there is no approved antiviral drug for treating COVID-19 patients and there is an urgent need for specific antiviral therapies and vaccines. In order for SARS-CoV-2 to enter cells, its surface glycoprotein spike (S) must be cleaved at two different sites by host cell proteases, which therefore represent potential drug targets. In the present study we investigated which host cell proteases activate the SARS-CoV-2 S protein in Calu-3 human airway epithelial cells. We show that S can be cleaved by both the proprotein convertase furin at the S1/S2 site and the transmembrane serine protease 2 (TMPRSS2) at the S2’ site. We demonstrate that TMPRSS2 is essential for activation of SARS-CoV-2 S in Calu-3 cells through antisense-mediated knockdown of TMPRSS2 expression. Further, we show that SARS-CoV-2 replication can be efficiently inhibited by two synthetic inhibitors of TMPRSS2 and also by the broad range serine protease inhibitor aprotinin. Additionally, SARS-CoV-2 replication was also strongly inhibited by the synthetic furin inhibitor MI-1851. Combining various TMPRSS2 inhibitors with MI-1851 produced more potent antiviral activity against SARS-CoV-2 than an equimolar amount of any single serine protease inhibitor. In contrast, inhibition of endosomal cathepsins by E64d did not affect virus replication. Our data demonstrate that both TMPRSS2 and furin are essential for SARS-CoV-2 activation in human airway cells and are promising drug targets for the treatment of COVID-19 either by targeting one of these proteases alone or by a combination of furin and TMPRSS2 inhibitors. Therefore, this approach has a high therapeutic potential for treatment of COVID-19.

In December 2019, a new coronavirus (CoV) emerged which rapidly spreads around the 56 world causing a pandemic never observed before with these viruses. The virus was identified 57 as a new member of the lineage b of the genus Betacoronavirus that also contains the 2002 58 severe acute respiratory syndrome (SARS)-CoV and was named SARS-CoV-2 by the WHO. 59

indicates that a trypsin-like serine protease is crucial for SARS-CoV-2 entry into these cells. 128

However, sequence analysis of the SARS-CoV-2 S protein suggests that furin may also be 129 involved in S processing ( Fig. 1B ; Coutard et al., 2020; Walls et al., 2020) . The S1/S2 site of 130 SARS-CoV-2 S protein contains an insertion of four amino acids providing a minimal furin 131 cleavage site (R-R-A-R 685 ↓) in contrast to the S protein of SARS-CoV. Instead, similar to 132 SARS-CoV the S2' cleavage site of SARS-CoV-2 S possesses a paired dibasic motif with a 133 single KR segment (KR 815 ↓) that is recognized by trypsin-like serine proteases. 134

In the present study we demonstrate that the S protein of SARS-CoV-2 is activated by 135 TMPRSS2 and furin. We also show that inhibitors against both proteases strongly suppress 136 virus replication in human airway epithelial cells and that the combination of both types of 137 inhibitors has a synergistic effect on virus reduction. Our results show that this approach has 138 a high therapeutic potential for treatment of COVID-19. substrates possess a nonbasic residue in P2 position, such as Pseudomonas aeruginosa 148 exotoxin A or Shiga toxin (Rockwell et al., 2002; Garten, 2018) . To test, whether the S1/S2 149 sequence of SARS-CoV-2 S protein is efficiently cleaved by furin, a small series of 150 Fluorescence Resonance Energy Transfer (FRET) substrates was synthesized ( Fig. 2A ). All 151 compounds possess a 3-nitrotyrosine amide as P4ˈ residue and a 2-amino-benzoyl 152 fluorophore in P7 position. The analogous sequences of the S proteins from MERS-CoV, 153 SARS-CoV, and avian infectious bronchitis virus (IBV) strain Beaudette were prepared as 154 reference substrates. Moreover, two FRET substrates of the SARS-CoV-2 S1/S2 cleavage 155 site with P2 AK and AR mutations were synthesized, to evaluate whether they could 156 constitute even more efficient cleavage sites for furin than the wild-type. The FRET 157 substrates were tested in an enzyme kinetic assay with human furin, and their cleavage 158 efficiency is shown in Figure 2B . The FRET substrate of the SARS-CoV-2 S1/S2 cleavage 159 site was efficiently cleaved by recombinant furin. In contrast, the monobasic SARS-CoV 160

FRET substrate was not processed by furin. The MERS-CoV S1/S2 FRET substrate 161 possessing a dibasic R-X-X-R motif was cleaved by furin approximately 10-fold less 162 efficiently than the best substrates of this FRET series. The FRET substrate SARS-CoV-163 2_M1, which contains an optimized furin recognition site by virtue of an AK mutation in P2 164 position, was cleaved with similar efficiency compared to the wild-type sequence. However, 165 substitution of AR in the P2 position strongly enhanced cleavage by furin. As expected, the 166 analogous reference sequence of IBV was also processed by furin very efficiently. The data 167

show that the R-R-A-R motif at the S1/S2 cleavage site of SARS-CoV-2 S is efficiently 168 cleaved by furin in vitro. 169 170 SARS-Cov-2 spike protein is cleaved by both furin and TMPRSS2 171

We next examined whether the SARS-CoV-2 S protein is cleaved by endogenous furin in 172 HEK293 cells. Cells were transiently transfected with pCAGGS plasmid encoding the SARS-173

CoV-2 S protein with a C-terminal Myc-6xHis-tag, and incubated in the absence and 174 presence of the potent synthetic furin inhibitor MI-1851 (cf. Fig. S1 ; manuscript in 175 preparation). At 24 h post transfection, cell lysates were subjected to SDS-PAGE and 176

Western blot analysis using antibodies against the Myc epitope. As shown in Fig. 2C , the 177 uncleaved precursor S and the S2 subunit were detected in the absence of MI-1851, 178 indicating that S is cleaved by endogenous proteases at the S1/S2 site in HEK293 cells. S 179 cleavage was efficiently prevented by MI-1851. In contrast, S cleavage was not prevented by 180 the trypsin-like serine protease inhibitor aprotinin. Thus, the data indicate that SARS-CoV-2 181 S protein is cleaved by furin at the S1/S2 site in HEK293 cells. 182

We then investigated SARS-CoV-2 S cleavage by TMPRSS2. Since HEK293 cells do not 183 express endogenous TMPRSS2 (unpublished data; see also www.proteinatlas.org), we co-184 transfected the cells with pCAGGS-S-Myc-6xHis and pCAGGS-TMPRSS2. Then, the cells 185 were incubated in the absence or presence of MI-1851 to suppress S cleavage by 186 endogenous furin. Interestingly, two S cleavage products of approximately 95 and 80 kDa, 187 respectively, were detected upon co-expression of TMPRSS2 in the absence of MI-1851 188 ( Fig. 2C) , most likely S2 and S2', as they can both be detected by the Myc-specific antibody 189 (cf. Fig. 1A ). In the presence of MI-1851, only a minor S2 protein band was detected. 190

However, the amount of S2' protein present in transient TMPRSS2 expressing cells was 191 similar in MI-1851 treated and untreated cells, suggesting that S cleavage at the S2' site is 192 only caused by TMPRSS2 activity. The small amount of S2 protein detected in expressing cells in the presence of MI-1851 was likely due to residual furin activity rather 194 than cleavage of S at the S1/S2 site by TMPRSS2. Together, the data show that SARS-CoV-195 2 can be cleaved by furin and by TMPRSS2. The data further suggest that the proteases 196 cleave S at different sites with furin processing the S1/S2 site and TMPRSS2 cleaving at the 197 S2' site. 198 199

Calu-3 human airway epithelial cells 201

Next, we wished to investigate whether TMPRSS2 is involved in proteolytic activation and 202 multicycle replication of SARS-CoV-2 in Calu-3 human airway epithelial cells. To specifically 203 knockdown TMPRSS2 activity, we previously developed an antisense peptide-conjugated 204 phosphorodiamidate morpholino oligomer (PPMO) (Böttcher-Friebertshäuser et al., 2011) . at 48 h p.i. (Fig. 3B) . 233

To confirm knockdown of enzymatically active TMPRSS2 expression, Calu-3 cells were 234 treated with PPMO or remained untreated for 24 h, after which TMPRSS2-specific mRNA 235 was isolated and analysed by RT-PCR as described previously (Böttcher-Friebertshäuser et 236 al., 2011) . Total RNA was analysed with primers designed to amplify nucleotides 108 to 1336 237 of TMPRSS2-mRNA. A full-length PCR product of 1228 bp was amplified from untreated and 238 scramble PPMO treated Calu-3 cells, whereas a shorter PCR fragment of about 1100 bp was 239 amplified from T-ex5 PPMO-treated cells (Fig. 3C ). Sequencing revealed that the truncated 240 TMPRSS2-mRNA lacked the entire exon 5 (data not shown). To further confirm that T-ex5 241 PPMO single dose treatment prior to infection still interferes with TMPRSS2-mRNA splicing 242 at 72 h p.i., total RNA was isolated from infected cells at 72 h p.i. and amplified as described In sum, our data demonstrate that inhibition of either TMPRSS2 or furin strongly inhibits 301 SARS-CoV-2 in Calu-3 human airway cells, indicating that both proteases are critical for S 302 activation. In contrast, endosomal cathepsins are dispensable or not involved at all in SARS-303

CoV-2 activation in these cells. The data demonstrate that SARS-CoV-2 replication can be efficiently reduced by inhibiting 328 either TMPRSS2 or furin activity, demonstrating that both proteases are crucial for SARS-329

CoV-2 activation. In conclusion, our data demonstrate that both TMPRSS2 and furin cleave the SARS-CoV-2 S 364 protein and are essential for virus multicycle replication in Calu-3 human airway cells. The 365 results indicate that TMPRSS2 and furin cleave S at different sites -furin at the S1/S2 site 366 and TMPRSS2 at the S2' site -and that TMPRSS2 and furin cannot compensate for each 367 other in SARS-CoV-2 S activation. Hence, inhibition of either one of these critical proteases 368

can render the S protein of SARS-CoV-2 unable to efficiently mediate virus entry and 369 membrane fusion. Therefore, TMPRSS2 and furin provide promising drug targets for Proteolytic processing of CoV S is a complex process that requires cleavage at two different 378 sites and is yet not fully understood. The amino acid sequence at the S1/S2 and S2' 379 cleavage sites varies among CoVs (Fig. 1B) , suggesting that partially different proteases 380 may be involved in activation. Sequence analyses of the S protein of the novel emerged 381 SARS-CoV-2 suggested that the R-R-A-R motif at the S1/S2 site may by sensitive to 382 cleavage by furin, while the S2' site contains a single R residue that can be cleaved by Hoffmann et al., 2020). In the present study, we demonstrate that the SARS-CoV-2 S protein 385 is cleaved by furin and by TMPRSS2. Furthermore, we show that multicycle replication of 386 SARS-CoV-2 in Calu-3 human airway cells is strongly suppressed by inhibiting TMPRSS2 387 and furin activity, demonstrating that both proteases are crucial for S activation in these cells. 388

Our data indicate that furin cleaves at the S1/S2 site, whereas TMPRSS2 cleaves at the S2' 389 site. The effective processing of the S1/S2 site by furin was additionally confirmed by for furin cleavage at the S1/S2 site. Likewise, strong inhibition of SARS-CoV-2 replication by 397 knockdown of TMPRSS2 activity using T-ex5 PPMO or treatment of Calu-3 cells with 398 aprotinin, MI-432 and MI-1900, respectively, indicates that furin cannot compensate for the 399 lack of TMPRSS2 in S activation. This was further confirmed by using an analogous FRET 400 substrate derived from the S2' cleavage site of the SARS-CoV-2 S protein (Fig. S2) . Kinetic 401 measurements clearly revealed that this substrate cannot be cleaved by furin (Fig. S2) . Thus, 402

we could experimentally demonstrate for the first time that furin only activates the S1/S2 site, The S protein of IBV strain Beaudette contains multibasic motifs at the S1/S2 and S2' site 435 substrates of the S1/S2 and S2' site of the IBV Beaudette S protein were efficiently cleaved 439 by furin (Fig. 2B and Fig. S2B) . However, the advantage of furin-cleavable multibasic motifs 440 at the S1/S2 and/or S2' site for multicycle replication, cellular tropism and pathogenicity of 441 HCoVs remains to be determined. HCoV-OC43 and HCoV-HKU1 possess a furin cleavage 442 motif at the S1/S2 site. In contrast, the S proteins of the 2002 SARS-CoV, HCoV-229E and 443

HCoV-NL63 possess single arginine residues at both cleavage sites (see also Fig. 2B and 444

CoV S lacks the 4-mer insertion at the S1/S2 site ( Fig. 1B; Walls et al., 2020) . The S protein 446 of MERS-CoV contains a dibasic motif of the sequence R-X-X-R at both S1/S2 and S2' site. Thus, combination of protease inhibitors (e.g. aprotinin or camostat) and antivirals provides a 537 promising approach to block SARS-CoV-2 replication that should be tested in cell cultures 538 and animal models and should furthermore be considered as therapeutic strategy for the 539 treatment of COVID-19. 540

In summary, we demonstrate that TMPRSS2 and furin are essential for activation and 541 The cDNA encoding the SARS-CoV-2 spike protein of isolate Wuhan-Hu-1 (GenBank 572 accession number MN908947; codon-optimized, sequence available upon request) with a C-573

terminal Myc-6xHis-tag was synthesized at Eurofins and subcloned into in the pCAGGS 574 expression plasmid using XhoI and NheI restriction sites (pCAGGS-S-Myc-6xHis). 575

Expression plasmid pCAGGS-TMPRSS2 encoding the cDNA of human TMPRSS2 has been 576 described previously (Böttcher et al., 2006) . . The S1/S2 and S2′ cleavage sites and subunits S1, S2 and S2' are indicated by black and coloured arrows, respectively. For immunochemical detection recombinant S is expressed with a C-terminally fused Myc-6xHis-tag peptide in our study. B) Alignment of the amino acid sequences at the S1/S2 and S2' cleavage site of the S proteins of different human coronaviruses (HCoV) and avian infectious bronchitis virus (IBV) strain Beaudette. 1-4) . T-ex5 treated cells were inoculated with SARS-CoV-2 as described above and incubated in the absence of PPMO for 72 h (lane 4). Total RNA was isolated and analysed by RT-PCR using primers designed to amplify 1228 nt of full-length TMPRSS2-mRNA. Full-length and truncated PCR product lacking i) S must be cleaved at two sites, S1/S2 and S2', to trigger fusion of viral and cellular membranes during virus entry in order to release the virus genome into the host cell. CoV S cleavage is believed to occur sequentially, with cleavage at the S1/S2 site occurring first and subsequent cleavage at the S2′ site. Furin processes the S1/S2 site, whereas TMPRSS2 cleaves at the S2' site, and both proteases cannot compensate each other. Inhibition of either furin (ii) or TMPRSS2 (iii) or simultaneous inhibition of both proteases (iv) renders the S protein fusion-inactive and prevents virus entry. Inhibition of TMPRSS2 prevents exposure of the fusion peptide at the N-terminus of the S2' subunit (iii and iv). Inhibition of furin cleavage at the S1/S2 site may directly interfere with virus entry and membrane fusion by steric blockage of conformational changes (ii, upper scheme) or may prevent exposure of the S2' site to TMPRSS2 (ii, lower scheme). Fusion-competent S is indicated in blue, fusionincompetent S in grey. 

Vectorization of morpholino oligomers by the (R-Ahx-R)4 peptide allows efficient splicing correction in 693 the absence of endosomolytic agents

Matriptase, HAT, and TMPRSS2 696 activate the hemagglutinin of H9N2 influenza A viruses

Activation of the SARS coronavirus spike protein via 698 sequential proteolytic cleavage at two distinct sites

Cathepsin L functionally cleaves the severe acute respiratory 700 syndrome coronavirus class I fusion protein upstream of rather than adjacent to the fusion peptide

Proteolytic 703 activation of influenza viruses by serine proteases TMPRSS2 and HAT from human airway epithelium

Hemagglutinin activating host cell proteases provide promising drug targets for the treatment of 707 influenza A and B virus infections

Inhibition of influenza virus infection 709 in human airway cell cultures by an antisense peptide-conjugated morpholino oligomer targeting the 710 hemagglutinin-activating protease TMPRSS2

Membrane-Anchored Serine Proteases: Host Cell Factors in 712

Proteolytic Activation of Viral Glycoproteins

Activation of Viruses by Host Proteases

Type II transmembrane serine proteases

Coronavirus cell entry occurs through the endo-/lysosomal 718 pathway in a proteolysis-dependent manner

The S2 Subunit 720 of QX-type Infectious Bronchitis Coronavirus Spike Protein Is an Essential Determinant of 721 Neurotropism

The spike glycoprotein 723 of the new coronavirus 2019-nCoV contains a furin-like cleavage site absent in CoV of the same 724 clade

Furin cleavage of the SARS coronavirus spike glycoprotein 726 enhances cell-cell fusion but does not affect virion entry

Morpholino Oligomers Targeting PB1 728 and NP Genes Enhance Survival of Mice Infected with Highly Pathogenic Influenza A H7N7 Virus

Entry, Replication, Immune Evasion, and 781 Neurotoxicity of Synthetically Engineered Bat-Borne Mumps Virus

Inhibition of Respiratory Syncytial Virus Infections with Morpholino Oligomers in Cell Cultures 784 and in Mice

Major Activating Protease of Influenza A Virus in Primary Human Airway Cells and Influenza B Virus in 789 Human Type II Pneumocytes

Peptidomimetic furin inhibitor MI-701 in combination with oseltamivir and ribavirin efficiently 792 blocks propagation of highly pathogenic avian influenza viruses and delays high level oseltamivir 793 resistance in MDCK cells

Inhibition of Influenza A 795 H3N8 Infections in Mice by Morpholino Oligomers

Characterization of a highly conserved domain 798 within the severe acute respiratory syndrome coronavirus spike protein S2 domain with characteristics 799 of a viral fusion peptide

Respiratory Syndrome Coronavirus Spike Protein Is Not Activated Directly by Cellular Furin during 802 Viral Entry into Target Cells

Identification of the first synthetic inhibitors of the type II transmembrane serine protease TMPRSS2 805 suitable for inhibition of influenza virus activation

Metabolic fate of 14C-camostat mesylate in man, rat and dog after intravenous administration

Host cell proteases: Critical determinants of coronavirus tropism and 810 pathogenesis

Morpholinos and their peptide conjugates: therapeutic promise and 812 challenge for Duchenne muscular dystrophy

Reproductive and Respiratory Syndrome Virus Infection in Piglets by a Peptide-Conjugated 815

Proteolytic processing 818 of Middle East respiratory syndrome coronavirus spikes expands virus tropism

Homology Modeling of TMPRSS2

Yields Candidate Drugs That May Inhibit Entry of SARS-CoV-2 into Human

Precursor processing by kex2/furin 824 proteases

Untreated cells (w/o) and DMSO treated cells were used as controls. Cell viability of untreated cells was set as 100 %. Results are mean values ± SD (n=3). C) Antiviral activity of combinations of TMPRSS2 and furin inhibitors against SARS-CoV-2 in human airway epithelial cells. Calu-3 cells were inoculated with SARS-CoV-2 at a MOI of 0.001 as described above and then incubated in the presence of single protease inhibitors or inhibitor combinations at the indicated concentrations. Virus titers in supernatants were determined by TCID 50 at 16, 24, 48 and 72 h p.i.. Data are mean values ± SD of two to three independent experiments. D) Calu-3 cells were treated with PPMO for 24 h, then infected with SARS-CoV-2 as described above and incubated in the absence of PPMO (w/o, scramble and T-ex5) and with or without 10 µM of furin inhibitor treatment (MI-1851) for 72 h