The severe acute respiratory syndrome-related coronavirus 2 (SARS-CoV-2) infects cells through interaction of its spike protein (SARS2-S) with Angiotensin-converting enzyme 2 (ACE2) and activation by proteases, in particular transmembrane protease serine 2 (TMPRSS2). Viruses can also spread through fusion of infected with uninfected cells. We compared the requirements of ACE2 expression, proteolytic activation, and the sensitivity to inhibitors for SARS2-S-mediated and SARS-CoV-S(SARS1-S)-mediated cell-cell fusion. SARS2-S-driven fusion was moderately increased by TMPRSS2 and strongly by ACE2, while SARS1-S-driven fusion was strongly increased by TMPRSS2 and less so by ACE2 expression. In contrast to SARS1-S, SARS2-S-mediated cell-cell fusion was efficiently activated by Batimastat-sensitive metalloproteases. Mutation of the S1/S2 proteolytic cleavage site reduced effector-target-cell fusion when ACE2 or TMPRSS2 were limiting and rendered SARS2-S-driven cell-cell fusion more dependent on TMPRSS2. When both ACE2 and TMPRSS2 were abundant, initial target-effector-cell fusion was unaltered compared to wt SARS2-S, but syncytia remained smaller. Mutation of the S2’ site specifically abrogated activation by TMPRSS2 for both cell-cell fusion and SARS2-S-driven pseudoparticle entry but still allowed for activation by metalloproteases for cell-cell fusion and by cathepsins for particle entry. Finally, we found that the TMPRSS2 inhibitor Bromhexine was unable to reduce TMPRSS2-activated cell-cell fusion by SARS1-S and SARS2-S as opposed to the inhibitor Camostat. Paradoxically, Bromhexine enhanced cell-cell fusion in the presence of TMPRSS2, while its metabolite Ambroxol exhibited inhibitory activity in some conditions. On Calu-3 lung cells, Ambroxol weakly inhibited SARS2-S-driven lentiviral pseudoparticle entry, and both substances exhibited a dose-dependent trend towards weak inhibition of authentic SARS-CoV-2. IMPORTANCE Cell-cell fusion allows the virus to infect neighboring cells without the need to produce free virus and contributes to tissue damage by creating virus-infected syncytia. Our results demonstrate that the S2’ cleavage site is essential for activation by TMPRSS2 and unravel important differences between SARS-CoV and SARS-CoV-2, among those greater dependence of SARS-CoV-2 on ACE2 expression and activation by metalloproteases for cell-cell fusion. Bromhexine, reportedly an inhibitor of TMPRSS2, is currently tested in clinical trials against coronavirus disease 2019. Our results indicate that Bromhexine enhances fusion in some conditions. We therefore caution against use of Bromhexine in higher dosage until its effects on SARS-CoV-2 spike activation are better understood. The related compound Ambroxol, which similarly to Bromhexine is clinically used as an expectorant, did not exhibit activating effects on cell-cell fusion. Both compounds exhibited weak inhibitory activity against SARS-CoV-2 infection at high concentrations, which might be clinically attainable for Ambroxol.

Kaposi’s sarcoma-associated herpesvirus (KSHV) is a human oncogenic virus associated with Kaposi’s sarcoma and two B-cell malignancies. The rhesus monkey rhadinovirus (RRV) is a virus of nonhuman primates that is closely related to KSHV. Eph family receptor tyrosine kinases (Ephs) are cellular receptors for the gH/gL glycoprotein complexes of both KSHV and RRV. Through sequence analysis and mutational screens, we identified conserved residues in the N-terminal domain of KSHV and RRV glycoprotein H that are critical for Eph-binding in vitro. Homology-based structural predictions of the KSHV and RRV gH/gL complexes based on the Epstein-Barr-Virus gH/gL crystal structure located these amino acids in a beta-hairpin on gH, which is likely stabilized by gL and is optimally positioned for protein-protein interactions. Guided by these predictions, we generated recombinant RRV and KSHV strains mutated in the conserved motif as well as an RRV gL null mutant. Inhibition experiments using these mutants confirmed that disruption of the identified Eph-interaction motif or of gL expression resulted in complete detargeting from Ephs. However, all mutants were infectious on all cell types tested, exhibiting normal attachment but a reduction in infectivity of up to one log order of magnitude. While Eph-binding-negative RRV mutants were replication-competent on fibroblasts, their infectivity was comparatively more reduced on endothelial cells with a substantial subpopulation of endothelial cells remaining resistant to infection. Together, this provides evidence for a cell type-specific use of Ephs by RRV. Furthermore, our results demonstrate that gL is dispensable for infection by RRV. Its deletion caused a reduction in infectivity similar to that observed after mutation of Eph-binding residues in gH. Our findings would be compatible with an ability of KSHV and RRV to use other, less efficient entry mediators in lieu of Ephs, although these host factors may not be uniformly expressed by all cells.

The SARS-Coronavirus-2 (SARS-CoV-2) infects cells through interaction of its spike protein (SARS-CoV-2-S) with the ACE2 receptor and activation by proteases, in particular TMPRSS2. Viruses can also spread through fusion of infected with uninfected cells. We therefore analyzed cell-cell fusion activity of SARS-CoV-2-S with regard to the requirements for ACE2 expression, proteolytic activation, and sensitivity to inhibitors and compared it to SARS-CoV-S. We compared S-protein-driven fusion with target cells recombinantly overexpressing ACE2, TMPRSS2, or both. SARS-CoV-2-S-driven fusion was moderately increased by TMPRSS2 and strongly by ACE2, while the reverse observation was made for SARS-CoV-S. TMPRSS2-mediated effects were inhibited by the serine protease inhibitor Camostat. Effector-target-cell fusion by SARS-CoV-2-S was only affected by Camostat when receptor expression was limiting or when the S1/S2 cleavage site was mutated. Mutational ablation of the SARS-CoV-2-S S2’ cleavage site abrogated any effects of TMPRSS2 on fusion. Mutation of the SARS-CoV-2-S S1/S2 cleavage site reduced effector-target-cell fusion when ACE2 or TMPRSS2 were limiting. When both factors were abundant, initial target-effector-cell fusion was unaltered, but syncytia remained smaller over time. Overall, its polybasic cleavage site renders SARS-CoV-2-S-mediated cell-cell fusion less dependent on TMPRSS2 activity on target cells. Unexpectedly, we observed enhancement of SARS-CoV-2-S-mediated fusion by Bromhexine, another TMPRSS2 inhibitor. This effect required intact proteolytic cleavage sites, suggesting interference of Bromhexine with proteolytic priming, but not in a therapeutically desired way. Infection with SARS-CoV-2-S-pseudotyped particles clearly differed in the requirements for proteolytic activation from cell-cell fusion. TMPRSS2 strongly enhanced infection, which was reversed by Camostat but not by Bromhexine.IMPORTANCECell-cell fusion allows the virus to infect additional target cells without the need to produce free virus. Fusion likely also contributes to tissue damage by creating virus-infected syncytia. Our results demonstrate that the S2’ cleavage site is essential for activation by TMPRSS2 in trans and unravel important differences between SARS-CoV and SARS-CoV-2. Bromhexine, an inhibitor of the TMPRSS2 protease, is currently tested in clinical trials against COVID-19. Our results indicate that Bromhexine does not inhibit SARS-CoV-2-S-mediated particle entry and enhances fusion. We therefore caution against overly optimistic use of Bromhexine in higher dosage in clinical trials or as a therapy, at least until its effects on SARS-CoV-2 spike activation are better understood. The related compound Ambroxol, which similar to Bromhexine is clinically used as an expectorant, did not exhibit activating effects on SARS-CoV-2-S-mediated fusion and may therefore currently represent a better choice in therapeutic regimens for COVID-19.

Kaposi’s sarcoma-associated herpesvirus (KSHV) is the causative agent of Kaposi’s sarcoma and is associated with two B cell malignancies, primary effusion lymphoma (PEL) and the plasmablastic variant of multicentric Castleman’s disease. On several adherent cell types, EphA2 functions as a cellular receptor for the gH/gL glycoprotein complex of KSHV. KSHV gH/gL also has previously been found to interact weakly with other members of the Eph family of receptor tyrosine kinases (Ephs), and other A-type Ephs have been shown to be able to compensate for the absence of EphA2 using overexpression systems. However, whether these interactions are of functional consequence at endogenous protein levels has remained unclear so far. Here, we demonstrate for the first time that endogenously expressed EphA7 in BJAB B cells is critical for the cell-to-cell transmission of KSHV from producer iSLK cells to BJAB target cells. The BJAB lymphoblastoid cell line often serves as a model for B cell infection and expresses only low levels of all Eph family receptors other than EphA7. Endogenous EphA7 could be precipitated from the cellular lysate of BJAB cells using recombinant gH/gL, and knockout of EphA7 significantly reduced transmission of KSHV into BJAB target cells. Knockout of EphA5, the second most expressed A-type Eph in BJAB cells, had a similar, although less pronounced, effect on KSHV infection. Receptor function of EphA7 was conserved for cell-free infection by the related rhesus monkey rhadinovirus (RRV), which is relatively even more dependent on EphA7 for infection of BJAB cells.IMPORTANCEInfection of B cells is relevant for two KSHV-associated malignancies, the plasmablastic variant of multicentric Castleman’s disease and PEL. Therefore, elucidating the process of B cell infection is important for the understanding of KSHV pathogenesis. While the high-affinity receptor for the gH/gL glycoprotein complex, EphA2, has been shown to function as an entry receptor for various types of adherent cells, the gH/gL complex can also interact with other Eph receptor tyrosine kinases with lower avidity. We analyzed the Eph interactions required for infection of BJAB cells, a model for B cell infection by KSHV. We identified EphA7 as the principal Eph receptor for infection of BJAB cells by KSHV and the related rhesus monkey rhadinovirus. While two analyzed PEL cell lines exhibited high EphA2 and low EphA7 expression, a third PEL cell line, BCBL-1, showed high EphA7 and low EphA2 expression, indicating a possible relevance for KSHV pathology.

IFITM proteins are the first line of defense against infection by many pathogens and may also have therapeutic importance, as they, among other effectors, mediate the antiviral effect of interferons. Neither their function against herpesviruses nor their mechanism of action is well understood.