Identification of the critical residues of TMPRSS2 for entry and host range of human coronavirus HKU1

ABSTRACT

Human coronavirus (CoV) HKU1 infection typically causes common cold but can lead to pneumonia in children, older people, and immunosuppressed individuals. Recently, human transmembrane serine protease 2 (hTMPRSS2) was identified as the functional receptor for HKU1, but its region and residues critical for HKU1 S binding remain elusive. In this study, we find that HKU1 could utilize human and hamster, but not rat, mouse, or bat TMPRSS2 for virus entry, displaying a narrow host range. Using human–bat TMPRSS2 chimeras, we show that the serine peptidase (SP) domain of TMPRSS2 is essential for entry of HKU1. Further extensive mutagenesis analyses of the C-terminal regions of SP domains of human and bat TMPRSS2s identify residues 417 and 469 critical for entry of HKU1. Replacement of either D417 or Y469 with asparagine in hTMPRSS2 abolishes its abilities to mediate entry of HKU1 S pseudovirions and cell–cell fusion, whereas substitution of N417 with D or N469 with Y in bat TMPRSS2 (bTMPRSS2) renders it supporting HKU1 entry. Our findings contribute to a deeper understanding of coronavirus–receptor interactions and cross-species transmission.

IMPORTANCE

The interactions of coronavirus (CoV) S proteins with their cognate receptors determine the host range and cross-species transmission potential. Recently, human transmembrane serine protease 2 (hTMPRSS2) was found to be the receptor for HKU1. Here, we show that the TMPRSS2 of hamster, but not rat, mouse, or bat, can serve as a functional entry receptor for HKU1. Moreover, swapping the residues at the positions of 417 and 469 of bTMPRSS2 with the corresponding residues of hTMPRSS2 confers it supporting entry of HKU1 S pseudovirions, indicating the critical role of these residues in HKU1 entry. Our study identified the critical residues in hTMPRSS2 responsible for receptor interaction and host range of HKU1.

INTRODUCTION

Coronaviruses (CoVs) are enveloped, single-stranded, positive-sense RNA viruses causing respiratory, gastrointestinal, or neurotropic diseases in mammals and birds. To date, nine CoVs capable of infecting humans were discovered, including 229E (1), OC43 (2, 3), NL63 (4), HKU1 (5), severe acute respiratory syndrome coronavirus (SARS-CoV) (6–8), the Middle East respiratory syndrome coronavirus (MERS-CoV) (9), the recently emerged SARS-CoV-2 (10–12), porcine delta coronavirus (PDCoV) (13), and canine coronavirus-human pneumonia-2018 (CCoV-HuPn-2018) (14). The first four cause common cold-like symptoms but can lead to pneumonia in children, older people, and immunosuppressed individuals, whereas infection by SARS-CoV, MERS-CoV, and SARS-CoV-2 causes pneumonia and even death. PDCoV has been documented in at least two independent zoonoses (13), and CCoV-HuPn-2018 was isolated from a Malaysian patient with pneumonia, and viral RNA was detected in eight patients hospitalized with pneumonia in Malaysia (14).

The spike (S) proteins of coronaviruses mediate virus entry by binding to their cellular receptors and mediating fusion of viral and cell membranes (15). The interactions of the S proteins and their cognate receptors determine cell tropism and host range of coronaviruses. To date, there are four proteinaceous receptors for hCoVs identified, and all of them are cell-surface proteases. The aminopeptidase N (APN) is the receptor for 229E (16), while NL-63, SARS-CoV, and SARS-CoV-2 employ angiotensin-converting enzyme 2 (ACE2) for virus entry (11, 17–20). MERS-CoV utilizes human dipeptidyl peptidase 4 (DPP4) as its functional receptor (21), whereas human transmembrane serine protease 2 (TMPRSS2) was recently reported as an entry receptor for HKU1 (22). The sialic acids have also been found to be a critical attachment factor facilitating entry of OC43, HKU1, and several other CoVs (23–25).

HKU1 was first detected in Hong Kong in 2004 (5) and belongs to the Embecovirus subgenus of Betacoronavirus genus, together with OC43 and mouse hepatitis virus (MHV). Like the other CoVs, HKU1 S protein is also a trimer and contains two subunits, S1 and S2. Its receptor-binding domain is located at the C-terminal domain (CTD) of S1(26), of which W515 and R517 are critical for receptor binding (22, 26, 27). There are three genotypes of HKU1—designated genotype A, B, and C (28). The S proteins of genotype A and genotype B share about 80% of amino acid sequence identity, whereas the S proteins of genotype B and genotype C are almost identical. In sharp contrast to the S protein of SARS-CoV-2, the prefusion state of HKU1 S protein adopts an all-closed conformation (29–31). Upon binding of 9-O acetyl sialic acids to its N-terminal domain (NTD), the HKU1 S protein transits from the closed conformation to an open state that can bind to its protein receptor TMPRSS2 (32), and its fusion potential is then activated through protease cleavage by either TMPRSS2 at plasma membrane or cathepsins inside lysosome (33, 34). However, a recent study revealed that the S protein of HKU1 genotype B exhibited both closed and open conformations, and binding of hTMPRSS2 resulted in more open conformations (35).

Human TMPRSS2 is a member of type II transmembrane serine proteases, consisting of 492 amino acids with an N-terminal intracellular domain, a transmembrane domain, a low-density lipoprotein receptor type-A (LDLR-A) domain, a class A scavenger receptor cysteine-rich (SRCR) domain, and a C-terminal trypsin-like serine peptidase (SP) domain (Fig. 1A) (36). TMPRSS2 is initially synthesized as an inactive zymogen and autocatalytically activated during the process of protein maturation (37). TMPRSS2 and other serine proteases are highly expressed in the respiratory tract, and their protease activities play important roles in triggering entry of influenza A virus (38–40), human metapneumovirus (41, 42), and many coronaviruses, including facilitating plasma membrane entry of SARS-CoV, MERS-CoV, and SARS-CoV-2 (19, 20, 42–46).

Fig 1.

The usage of TMPRSS2 orthologs by HKU1 to support virus entry. (A) Schematic diagram of human TMPRSS2. (B) Phylogenetic analysis of TMPRSS2 proteins from human, bat, mouse, rat, and hamster. (C) Western blot analysis for TMPRSS2 proteins from mammalian species in cell lysates. HEK293 cells were transfected with plasmids encoding TMPRSS2 orthologs from different mammalian species. At 40 h post-transfection, the TMPRSS2 proteins in cell lysates were detected by Western blot analysis using antibodies against FLAG tag. β-actin was used as a loading control. The experiments were performed at least three times and one representative is shown. (D) Transduction of 293 cells transiently expressing TMPRSS2 orthologs from different mammalian species by HKU1 S-pseudotyped lentiviruses. HEK293 cells transiently expressing TMPRSS2 orthologs were transduced by HKU1 S-pseudotyped lentiviruses. At 48 h post-transduction, the cells were lysed, and the luciferase activities in the cell lysates were measured to represent the pseudoviral entry efficiency. The experiments were carried out in triplicate and repeated at least three times. One representative is shown with error bars indicating SD. (E) The abilities of the TMPRSS2 proteins from different mammalian species to facilitate HKU1 S-mediated cell–cell fusion. HEK293T cells were co-transfected with plasmids encoding HKU1 S proteins and plasmids expressing EGFP using PEI. At 40 h post-transfection, the cells were lifted with trypsin and overlaid on 293 cells transiently expressing TMPRSS2 proteins from different mammalian species. After a 4-h coculture, the cell–cell fusion was observed using a fluorescent microscope. The experiments were performed at least three times and one representative is shown. (F) The quantitative analysis of cell–cell fusion. The area of cell–cell fusion was measured using ImageJ software and was normalized to the WT group. Data are shown as means ± SD.

Recently, human TMPRSS2 was identified as entry receptor for HKU1 virus (22). Strikingly, its protease activity is not required for entry of HKU1. The TMPRSS2 proteins are highly conserved among mammals; the TMPRSS2s of human, bat, hamster, mouse, and rat share over 76% the amino acid sequence identity with each other. Whether they could mediate HKU1 entry remains elusive. In this study, we determined the receptor activity of several animal TMPRSS2 orthologs for HKU1 and identified the critical domain and residues essential for HKU1 entry.

RESULTS

The hamster TMPRSS2 ortholog, but not those from bat, mouse, or rat, functions as a receptor for HKU1

The TMPRSS2s of bat, hamster, mouse, and rat share over 76% homology in amino acid sequence identity with hTMPRSS2 (Fig. 1B). To determine whether they might be able to be utilized for HKU1 entry, C-terminal Flag-tagged TMPRSS2 orthologs were ectopically expressed in HEK293 cells, and their susceptibilities to infection of HKU1 were evaluated. All four animal TMPRSS2 orthologs were expressed as well as hTMPRSS2 (Fig. 1C), and both full-length and proteolytically cleaved forms of TMPRSS2 were present in cell lysates, when detected using mouse monoclonal anti-FLAG tag M2 antibody. Various cleaved forms of TMPRSS2s were likely due to the self-cleavage activity of TMPRSS2 that occurs during its maturation process. Compared to mock control, HEK293 cells transiently expressing hTMPRSS2 exhibited about 78-fold increase of luciferase activities when transduced by lentiviral pseudovirions with HKU1 genotype B S proteins (Fig. 1D), consistent with the notion that hTMPRSS2 is the functional receptor for HKU1 (22, 47). Similarly, hTMPRSS2 also facilitated entry of HKU1 genotype A pseudovirions, albeit with reduced efficiency (Fig. 1D). In contrast, HEK293 cells expressing the TMPRSS2s of bat, hamster, mouse, or rat only displayed a background level of luciferase activities, indicating that they might be non-susceptible to HKU1 infection (Fig. 1D). The S-mediated cell–cell fusion assay was also performed to further evaluate the receptor activity of TMPRSS2 orthologs. HEK293T cells transiently co-expressing HKU1 S protein and EGFP were lifted with trypsin and overlaid on HEK293 cells expressing TMPRSS2 orthologs. After 4 h of incubation, the levels of syncytium formation were analyzed using a fluorescence microscope. Transient expression of hTMPRSS2 markedly enhanced HKU1 S-mediated syncytia (Fig. 1E and F), consistent with the previous report (22), whereas bat, hamster, mouse, and rat wild-type (WT) TMPRSS2s only exhibited a background level of syncytia, indicating that HKU1 might only have very limited host range.

To rule out the possibility that TMPRSS2 orthologs from bat, hamster, mouse, and rat fail to function as receptors for HKU1 might be due to lack of cell-surface expression, we conducted a cell-surface protein biotinylation assay to assess their expression. Similar to human TMPRSS2, both bat and mouse TMPRSS2s were detected on the cell surface, whereas hamster and rat TMPRSS2s were not (Fig. 2A), indicating that their low receptor activities might result from low cell-surface expression. Because protease activity of TMPRSS2 has significant effect on its expression on cell surface (37), camostat, a broad-spectrum serine protease inhibitor, was added during HEK293 cells expressing WT hamster and rat TMPRSS2s. As expected, camostat significantly increased the surface levels of TMPRSS2 orthologs, including those from rat and hamster (Fig. 2B). We then reassessed their receptor activity for HKU1 in the presence of camostat. Although WT bat, mouse, and rat TMPRSS2s still failed to mediate entry of HKU1 S pseudovirions, WT hamster TMPRSS2 exhibited a substantial increase in infectivity, even exceeding that of human TMPRSS2 without camostat (Fig. 2C), indicating its role as a functional receptor for HKU1. Camostat also enhanced HKU1 pseudoviral infectivity by over 30-fold in cells expression hTMPRSS2 (Fig. 2C). We also created catalytically inactive mutants of TMPRSS2 ortholog. Disabling TMPRSS2 protease activity led to a marked increase in the surface expression of rat and hamster TMPRSS2s (Fig. 2D). Although the catalytically inactive mutants of bat, mouse, and rat TMPRSS2s still failed to support HKU1 pseudoviral entry, the hamster TMPRSS2 S441A mutant gained the ability to mediate HKU1 entry, consistent with WT hamster TMPRSS2, facilitating HKU1 infection in the presence of camostat (Fig. 2C). Collectively, our results demonstrate that the low surface expression of WT hamster TMPRSS2 limits its ability to function as a receptor for HKU1. Inhibiting its catalytic activity increases surface expression, facilitating HKU1 infection.

Fig 2.

The hamster TMPRSS2 is a functional receptor for HKU1. (A, B, and D) Cell-surface expression of the WT TMPRSS2 orthologs in the absence of camostat (A) and in the presence of camostat (B), and the catalytically inactive TMPRSS2 orthologs (D). HEK293 cells were transfected with plasmids encoding the WT or catalytically inactive mutant TMPRSS2 orthologs. At 40 h post-transfection, the cells were incubated with Z-Link Sulfo-NHS-SS-Biotin, followed by lysis using RIPA. Biotinylated cell-surface proteins were enriched with NeutrAvidin beads, and the cell-surface expression of the WT or catalytically inactive mutant TMPRSS2 orthologs was detected by Western blot using mouse monoclonal anti-FLAG M2 antibody. β-actin and HLA-C were used as input control and biotinylation loading control, respectively. The experiments were performed at least three times and one representative is shown. (C) Transduction of 293 cells transiently expressing the WT or catalytically mutant TMPRSS2 orthologs from different mammalian species by HKU1 S-pseudotyped lentiviruses in the presence or absence of camostat. The experiments were carried out in triplicate and repeated at least three times. One representative is shown with error bars indicating SD.

The C-terminal region of SP domain of hTMPRSS2 is critical for HKU1 entry

Phylogenetic analysis reveals that, among four orthologs, bTMPRSS2 is closest to hTMPRSS2 with about 80.6% homology. To identify the specific region(s) of TMPRSS2 responsible for HKU1 S protein binding, we constructed two human–bat TMPRSS2 chimeras, hTMPRSS2-b(351-492) and bTMPRSS2-h(351-492), by swapping the last 142 amino acids at the C-terminal region between them (Fig. 3A). Both chimeric mutants were expressed well in HEK293 cells (Fig. 3B). Of note, the hTMPRSS2-b(351-492) migrated more slowly than hTMPRSS2 in SDS-PAGE gels, whereas the chimera bTMPRSS2-h(351-492) moved slightly faster than the WT bTMPRSS2 (Fig. 3B), likely resulting from the change of an additional N-linked glycosylation site at the C-terminal end of bTMPRSS2, which is absent in hTMPRSS2. The levels of human–bat TMPRSS2 chimeras on the cell surface were also evaluated by cell-surface protein biotinylation assay, and both chimeras were present on the cell surface at levels similar to their respective WT controls. Replacement of aa351-492 of hTMPRSS2 with that of bTMPRSS2 completely abolished the receptor activity of hTMPRSS2 (Fig. 3C), indicating that the aa351-492 of hTMPRSS2 might be critical for HKU1 entry. In contrast, the chimeric bTMPRSS2-h(351-492) with the aa351-492 of hTMPRSS2 substantially gained its ability to mediate HKU1 entry (Fig. 3C), further confirming the importance of the aa351-492 of hTMPRSS2 in HKU1 entry. We further examined the function of this region in HKU1 S binding by incubating 293T cells expressing chimeric TMPRSS2 with the soluble receptor-binding domain (RBD) of HKU1B. Although hTMPRSS2-b(351-492) chimera lost its ability to bind to HKU1B RBD, the bTMPRSS2-h(351-492) chimera gained the ability to bind to HKU1B RBD (Fig. 3D). These results demonstrated that the aa351-492 of hTMPRSS2 is involved in the binding of HKU1 S protein. We also performed the S-mediated cell–cell assay. Consistent with the results of pseudoviral transduction and receptor binding, the hTMPRSS2-b(351-492) chimera lost its ability of facilitating S-mediated syncytium formation, whereas bTMPRSS2-h(351-492) chimera acquired its capacity to promote S-mediated cell–cell fusion (Fig. 3E and F). Collectively, our data indicate that the C-terminal amino acids 351-492 within TMPRSS2 SP domain are critical for HKU1 entry.

Fig 3.

The C-terminal amino acids 351-492 within TMPRSS2 SP domain are critical for HKU1 S recognition. (A) Schematic representations of human TMPRSS2, bat TMPRSS2, and chimeric proteins hTMPRSS2-b(351-492) and bTMPRSS2-h(351-492). (B) Western blot analysis for the WT and chimeric TMPRSS2 proteins at cell lysates and on cell surfaces. HEK293 cells were transfected with plasmids encoding the WT or chimeric TMPRSS2 proteins. At 40 h post-transfection, the cells were incubated with Z-Link Sulfo-NHS-SS-Biotin, followed by lysis using RIPA. Biotinylated cell-surface proteins were enriched with NeutrAvidin beads, and the cell-surface expression of the WT and chimeric TMPRSS2 proteins was detected by Western blot using mouse monoclonal anti-FLAG M2 antibody. β-actin and HLA-C were used as input control and biotinylation loading control, respectively. The experiments were performed at least three times and one representative is shown. (C) Transduction of 293 cells transiently expressing the WT or chimeric TMPRSS2 proteins by HKU1 genotype B S-pseudotyped lentiviruses. HEK293 cells transiently expressing the WT or chimeric TMPRSS2 proteins were transduced by HKU1 genotype B S-pseudotyped lentiviruses. At 48 h post-transduction, the cells were lysed, and the luciferase activities in the cell lysates were measured to represent the pseudoviral entry efficiency. The experiments were carried out in triplicate and repeated at least three times. One representative is shown with error bars indicating SD. (D) Binding of HKU1B RBD to the chimeric TMPRSS2 proteins. HKE293 cells transiently expressing the WT or chimeric TMPRSS2 proteins were incubated with mFc-tagged HKU1B RBD proteins, followed by incubation with FITC-conjugated goat anti-mouse secondary antibody. Finally, the cells were analyzed by flow cytometry. The experiments were carried out in triplicate and repeated at least three times. The relative RBD binding was normalized to the WT hTMPRSS2. One representative is shown with error bars indicating SD. (E) The abilities of the WT and chimeric TMPRSS2 proteins to facilitate HKU1 genotype B S-mediated cell–cell fusion. HEK293T cells were co-transfected with plasmids encoding HKU1 genotype B S proteins and plasmids expressing EGFP using PEI. At 40 h post-transfection, the cells were lifted with trypsin and overlaid on 293 cells transiently expressing the WT or mutant TMPRSS2 proteins. After a 4-h coculture, the cell–cell fusion was observed using a fluorescent microscope. All cell–cell fusion experiments in Figure 3E were conducted simultaneously with the same negative (vector) and positive (human TMPRSS2) controls as in Figure 1E. The experiments were performed at least three times and one representative is shown. (F) The quantitative analysis of cell–cell fusion. The area of cell–cell fusion was measured using ImageJ software and was normalized to the WT group. Data are shown as means ± SD.

Residues D417 and Y469 of TMPRSS2 are crucial for HKU1 S binding

To further identify the residues in TMPRSS2 critical for HKU1 entry, we swapped residues in the aa351-492 region that differ between human and bat TMPRSS2s (Fig. 4A), either individually or in combination when adjacent residues varied, and their receptor activities for HKU1 entry were evaluated. As shown in Figure 4B, majority of hTMPRSS2 mutants were expressed at levels similar to or even higher than WT hTMPRSS2. Although most hTMPRSS2 mutants almost had no effect on entry efficiency by HKU1 S pseudovirions, two mutants with single mutation, D417N and Y469N, failed to mediate entry of HKU1 S pseudovirions (Fig. 4C), indicating that the residues D417 and Y469 of hTMPRSS2 might be critical for entry of HKU1 S pseudovirions. The levels of cell-surface expression of D417N and Y469N mutants were comparable to those of the WT hTMPRSS2, as determined by flow cytometry using antibodies against hTMPRSS2 (Fig. 4D). We further determined the binding capacities of D417N and Y469N mutants to HKU1B RBD. The affinities of both mutants to HKU1B RBD were decreased by more than 10-fold compared to that of the WT hTMPRSS2 (Fig. 4E). Cell–cell fusion assay also showed that single D417N or Y469N mutation in hTMPRSS2 abrogated its ability to support S-mediated syncytium formation (Fig. 4F and G), further confirming the key role of residues D417 and Y469 of hTMPRSS2 in HKU1 entry.

Fig 4.

Residues D417 and Y469 of hTMPRSS2 are crucial for HKU1 S-mediated pseudovirus entry and cell–cell fusion. (A) Sequence alignment of amino acids 351-492 of TMPRSS2 proteins from human and bat. (B) Western blot analysis for the WT and mutant hTMPRSS2 proteins. HEK293 cells were transfected with plasmids encoding the WT or mutant hTMPRSS2 proteins. At 40 h post-transfection, the TMPRSS2 proteins in cell lysates were detected by Western blot analysis using antibodies against FLAG tag. β-actin was used as a loading control. The experiments were performed at least three times and one representative is shown. (C) Transduction of 293 cells transiently expressing the WT or mutant hTMPRSS2 proteins by HKU1 genotype B S-pseudotyped lentiviruses. HEK293 cells transiently expressing the WT or mutant hTMPRSS2 proteins were transduced by HKU1 genotype B S-pseudotyped lentiviruses. At 48 h post-transduction, the cells were lysed, and the luciferase activities in the cell lysates were measured to represent the pseudoviral entry efficiency. The experiments were carried out in triplicate and repeated at least three times. One representative is shown with error bars indicating SD. (D) Analysis of hTMPRSS2 mutants on the cell surface by flow cytometry assay. HEK293 cells were transfected with plasmids encoding the WT or mutant hTMPRSS2s. At 40 h post-transfection, the cells were incubated with rabbit anti-TMPRSS2 antibody, followed by FITC-conjugated goat anti-rabbit secondary antibody. After washing, the cells were analyzed by flow cytometry. The experiments were performed at least three times and one representative is shown with error bars indicating SD. (E) Binding of HKU1B RBD to hTMPRSS2 mutants. HKE293 cells transiently expressing the WT or mutant hTMPRSS2 proteins were incubated with mFc-tagged HKU1B RBD proteins, followed by incubation with FITC-conjugated goat anti-mouse secondary antibody. Finally, the cells were analyzed by flow cytometry. The experiments were carried out in triplicate and repeated at least three times. The relative RBD binding was normalized to the WT hTMPRSS2. One representative is shown with error bars indicating SD. (F) The abilities of the WT and mutant hTMPRSS2 proteins to facilitate HKU1 genotype B S-mediated cell–cell fusion. HEK293T cells were co-transfected with plasmids encoding HKU1 genotype B S proteins and plasmids expressing EGFP using PEI. At 40 h post-transfection, the cells were lifted with trypsin and overlaid on 293 cells transiently expressing the WT or mutant hTMPRSS2 proteins. After a 4-h coculture, the cell–cell fusion was observed using a fluorescent microscope. All cell–cell fusion experiments in Figure 4F were conducted simultaneously with the same negative (vector) and positive (human TMPRSS2) controls as in Figure 1E. The experiments were performed at least three times and one representative is shown. (G) The quantitative analysis of cell–cell fusion. The area of cell–cell fusion was measured using ImageJ software and was normalized to the WT group. Data are shown as means ± SD.

Most of bTMPRSS2 mutants were expressed at levels similar to or even higher than WT bTMPRSS2 (Fig. 5A) and remained non-susceptible for entry of HKU1 S pseudovirons. However, single individual N417D and N469Y mutation in bTMPRSS2 exhibited significant increase of HKU1 S pseudoviral entry by more than 10-fold (Fig. 5B), compared to WT bTMPRSS2, further supporting that residues 417 and 469 in bTMPRSS2 might be essential for HKU1 entry. To investigate whether these two substitutions have any synergistic effect, we further constructed a bTMPRSS2 double mutant, N417D/N469Y. The N417D/N469Y mutant was also expressed as well as WT bTMPRSS2 in cell lysates and on cell surfaces (Fig. 5E). The double mutant mediated a significantly higher level of syncytium formation with HKU1 S expressing cells and a significantly higher transduction efficiency with HKU1 S pseudovirions than those of both single mutants (Fig. 5C, D, and F), indicating the synergistic effect of these two mutations. The binding affinities of bTMPRSS2 mutants to HKU1B RBD were further evaluated by flow cytometry. Although the bTMPRSS2 mutants N417D and N469Y gained the ability of binding to HKU1B RBD compared to the WT bTMPRSS2, the double mutants N417D/N469Y showed an even greater increase in its binding affinity (Fig. 5G), further supporting the crucial role of residues 417 and 469 of TMPRSS2 in HKU1 S binding.

Fig 5.

The bTMPRSS2 mutants N417D and N469Y acquired the ability to support HKU1 S-mediated pseudovirus entry and cell–cell fusion. (A) Western blot analysis for the WT and mutant bTMPRSS2 proteins. HEK293 cells were transfected with plasmids encoding the WT or mutant bTMPRSS2 proteins. At 40 h post-transfection, the TMPRSS2 proteins in cell lysates were detected by Western blot analysis using antibodies against FLAG tag. β-actin was used as a loading control. The experiments were performed at least three times and one representative is shown. (B) Transduction of 293 cells transiently expressing the WT or mutant bTMPRSS2 proteins by HKU1 genotype B S-pseudotyped lentiviruses. 293 cells transiently expressing the WT or mutant bTMPRSS2 proteins were transduced by HKU1 genotype B S-pseudotyped lentiviruses. At 48 h post-transduction, the cells were lysed, and the luciferase activities in the cell lysates were measured to represent the pseudoviral entry efficiency. The experiments were carried out in triplicate and repeated at least three times. One representative is shown with error bars indicating SD. (C) The ability of the WT and mutant bTMPRSS2 proteins to facilitate HKU1 genotype B S-mediated cell–cell fusion. HEK293T cells were co-transfected with plasmids encoding HKU1 genotype B S proteins and plasmids expressing EGFP using PEI. At 40 h post-transfection, the cells were lifted with trypsin and overlaid on 293 cells transiently expressing the WT or mutant bTMPRSS2 proteins. After a 4-h coculture, the cell–cell fusion was observed using a fluorescent microscope. All cell–cell fusion experiments in Figure 5C were conducted simultaneously with the same negative (vector) and positive (human TMPRSS2) controls as in Figure 1E. The experiments were performed at least three times and one representative is shown. (D) The quantitative analysis of cell–cell fusion. The area of cell–cell fusion was measured using ImageJ software and was normalized to the WT group. Data are shown as means ± SD. (E) Western blot analysis for the WT and mutants N417D, N469Y, and N417D/N469Y bTMPRSS2 proteins. HEK293 cells were transfected with plasmids encoding the bTMPRSS2 WT and mutants N417D, N469Y, and N417D/N469Y. At 40 h post-transfection, the cells were incubated with Z-Link Sulfo-NHS-SS-Biotin, followed by lysis using RIPA. Biotinylated cell-surface proteins were enriched with NeutrAvidin beads, and the cell-surface expression of the WT or mutant bTMPRSS2 proteins was detected by Western blot using mouse monoclonal anti-FLAG M2 antibody. β-actin and HLA-C were used as input control and biotinylation loading control, respectively. (F) Transduction of 293 cells transiently expressing the bTMPRSS2 WT or mutants N417D, N469Y, and N417D/N469Y by HKU1 genotype B S-pseudotyped lentiviruses. HEK293 cells transiently expressing the bTMPRSS2 WT or mutants N417D, N469Y, and N417D/N469Y were transduced by HKU1 genotype B S-pseudotyped lentiviruses. At 48 h post-transduction, the cells were lysed, and the luciferase activities in the cell lysates were measured to represent the pseudoviral entry efficiency. The experiments were carried out in triplicate and repeated at least three times. One representative is shown with error bars indicating SD. (G) Binding of HKU1B RBD to bTMPRSS2 mutants. HKE293 cells transiently expressing the WT or mutant bTMPRSS2 proteins were incubated with mFc-tagged HKU1B RBD proteins, followed by incubation with FITC-conjugated goat anti-mouse secondary antibody. Finally, the cells were analyzed by flow cytometry. The experiments were carried out in triplicate and repeated at least three times. The relative RBD binding was normalized to the WT hTMPRSS2. One representative is shown with error bars indicating SD.

DISCUSSION

Coronavirus initiates viral infection by binding to their receptors via the S proteins. The interactions between the S proteins and their respective receptors largely determine cell and tissue tropism, host range, and potential cross-species transmission of coronaviruses (48, 49). Human TMPRSS2 was recently identified as the cellular receptor for HKU1 (22). In our previous study, we found that residues W515 and R517 within the C-terminal domain of S1 subunit are critical for HKU1 infection (27). The mutants W515A and R517A almost abolished the S protein binding to the receptor hTMPRSS2 (22). However, the molecular mechanisms underlying how hTMPRSS2 interacts with the HKU1 S protein remain largely elusive. Although hTMPRSS2 and bTMPRSS2 share a high similarity at the amino acid level, bTMPRSS2 lacks the ability to mediate HKU1 entry. By assessing the abilities of the chimeric human–bat TMPRSS2 proteins to support HKU1 entry, we determined that the C-terminal amino acids 351-492 within the SP domain of hTMPRSS2 are crucial for hTMPRSS2 interacting with the HKU1 S protein. These findings are in agreement with the previous reports that the anti-hTMPRSS2 nanobodies blocked the entry of HKU1 pseudoviruses and suppressed the enzymatic activities of hTMPRSS2 (22). However, the catalytic activity of hTMPRSS2 is dispensable for HKU1 S-mediated pseudovirus entry, as two catalytically inactive hTMPRSS2 mutants, R255Q and S441A, still supported the entry of HKU1 (22).

In this study, we identified residues D417 and Y469 of hTMPRSS2 critical for HKU1 entry. D417 and Y469 are located in two adjacent loops of TMPRSS2, loop 3 and loop 2, respectively (Fig. 6A), constituting the major binding sites for HKU1 S protein. Of note, residues 417 and 469 of TMPRSS2 are not conserved among human, bat, rat, mouse, and hamster with D417 and Y469 for hTMPRSS2 and N417 and N469/L469 for the others except for Y469 in rat TMPRSS2 (Fig. 6B). Although both hamster and mouse TMPRSS2s share the same amino acids at positions 417 and 469, only hamster TMPRSS2 serves as an entry receptor for HKU1, suggesting that other differing residues between the two might account for this discrepancy. Moreover, despite hamster TMPRSS2’s ability to efficiently mediate HKU1 entry, HEK293 cells expressing WT hamster TMPRSS2 alone remain non-susceptible to HKU1 entry due to its substantially low cell-surface expression. The addition of camostat or the introduction of an enzymatic activity mutation significantly enhances its surface expression, thereby enabling its receptor activity for HKU1.

Fig 6.

The key HKU1 S-binding sites in TMPRSS2 proteins. (A) Structure of hMPRSS2 (PDI:7meq). The SP domain is colored in pink, the LDLR domain is colored in cyan. The yellow shadow shows the potential HKU1 S-binding interface. (B) Comparison of residues 417 and 469 in TMPRSS2 proteins from a variety of mammalian species. GenBank accession numbers for rabbit (XP_051693971), cat (XP_023094478.2), dog (XP_038299491.1), monkey (XP_011724292.1), and camel (XP_064332182.1).

Given that D417N mutant failed to mediate HKU1 entry, and D and N mainly differ by a carboxyl group or negative charge, the salt bridge(s) formed by D417 with S protein might be critical for its receptor activity. The Y469N substitution replaced a bulky and hydrophobic-hydroxyphenyl group with a hydrophilic side chain and yielded a nonfunctional receptor, indicating that hydrophobic interactions from residue 469 of TMPRSS2 with S protein might also be critical. Interestingly, monkey TMPRSS2 also has N469 (Fig. 6B), suggesting that monkey also might not be susceptible for HKU1 virus infection. HKU1 viruses have thought to be originated from rodents (50). However, our data showed that neither rat nor mouse TMPRSS2 could act as the functional receptor for HKU1 entry. We also observed a difference in outcomes when using hTMPRSS2 or bTMPRSS2 as a backbone when investigating the role of residues 417 and 469 in HKU1 entry. Mutation of either D417 or Y469 in hTMPRSS2 to N resulted in complete loss of the receptor activity for HKU1, whereas substitution of either N417 or N469 in bTMPRSS2 with the corresponding amino acids of hTMPRSS2 only converted bTMPRSS2 into a partially functional receptor for HKU1. This suggests that other amino acid differences between hTMPRSS2 and bTMPRSS2 may account for the discrepancy. The bTMPRSS2 amino acid sequence used in this study is from Rhinolophus ferrumequinum bats. We analyzed the bat TMPRSS2 amino acid sequences available in the database and found that the residues at positions 417 and 469 are well conserved across all of them. Therefore, we speculate that HKU1 might be unable to infect any bat species. As an RNA virus, CoV constantly evolves and quickly adapts new hosts through mutation and recombination after cross-species transmission (46, 51), which might result in usage of different receptors for virus entry, although the exact mechanism remains elusive.

During the preparation of our manuscript, two preprint manuscripts and one published paper independently reported the complex structures of the HKU1-S/hTMPRSS2 (35, 47, 52). These studies also found residues 417 and 469 in hTMPRSS2 critical for its interaction with the S protein of HKU1, consistent with our findings that substitutions of residues 417 and 469 in hTMPRSS2 to their corresponding residues in bTMPRSS2 disable its receptor activity for HKU1. Although 293 cells transiently expressing hTMPRSS2 supported HKU1 S-mediated pseudovirus entry, the authentic HKU1 virus failed to replicate in these cells (data not shown), indicating the presence of post-entry block for HKU1 virus in 293 cells. Moreover, we also found that pseudovirions bearing the HKU1B S protein exhibited a higher transduction efficiency in 293 cells expressing hTMPRSS2 compared to pseudovirions bearing the HKU1A S protein, consistent with the observation that HKU1B CTD had a higher affinity constant with hTMPRSS2 protein than that of HKU1A (22). In the structure of the HKU1-S/hTMPRSS2, residue D417 of hTMPRSS2 interacts with H488 and W515 of HKU1A S protein and with K487 of HKU1B S protein, while residue Y469 of hTMPRSS2 interacts with R517, L521, and Y528 of HKU1A S protein and with L521 and Y528 of HKU1B S protein. W515, R517, L521, and Y528 are conserved between the S proteins of HKU1A and HKU1B, and only residues 487 and 488 vary between HKU1A and 1B. Individually or simultaneously swapping residues 487 and 488 between the S proteins of HKU1A and 1B showed that residue 487 in HKU1A S protein and residue 488 in HKU1B S protein are critical for the entry of pseudovirions on hTMPRSS2 (data not shown). Nevertheless, the discrepancy in transduction between HKU1A and 1B cannot be attributed to residues 487 and 488 of the S protein. Interestingly, all of HKU1A, 1B, 1C strains are extensively circulated among human population and propagate very well in human airway epithelial cells (5, 53–55), indicating that additional factors other than hTMPRSS2 might also be critical for HKU1 virus entry.

N-linked glycosylations of coronaviral receptors exhibit great impact on the virus–receptor interaction (56–59). Addition of an N-linked glycosylation to the S-binding region of human APN abolished its receptor activity for 229E (59). Similarly, variations in glycosylation patterns of DPP4 orthologs from different animal species play a critical role in determining the receptor utilization by MERS-CoV (56, 57). In the case of TMPRSS2, there is an extra N-linked glycosylation present at residue N476 of bTMPRSS2, which is absent in hTMPRSS2. However, removal or addition of this glycosylation almost had no effect on the receptor activity of TMPRSS2 for HKU1, indicating that this glycosylation might not play any role in the binding of the HKU1 S protein.

In conclusion, we identified the SP region and residues 417 and 469 in hTMPRSS2 essential for entry of HKU1. The sequence variations at residues 417 and 469 in TMPRSS2s have major effect on susceptibility and host range of HKU1. Our findings provide a better understanding of mechanism of entry and cross-species transmission of CoV.

MATERIALS AND METHODS

Cell lines

HEK293T cells and HEK293 cells were maintained in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin, streptomycin, and fungizone (PSF) at 37°C with 5% CO₂.

Plasmids

The human codon-optimized DNA sequences of HKU1 genotype A (GenBank: AY597011) and genotype B (GenBank: AGT17758.1) S proteins were synthesized by GenScript and cloned into pcDNA3.1+. The DNA fragment encoding the HKU1B S RBD (residues 310–676) with an N-terminal signal peptide and a C-terminal mouse Fc tag was subcloned into pcDNA3.1+ to generate the recombinant plasmid pcDNA3.1-HKU1B-RBD-mFc. The plasmid encoding human TMPRSS2 protein with C-terminal FLAG and His tags was purchased from WZ Biosciences Inc (Shandong, China). The coding sequences of TMPRSS2 proteins for mouse (GenBank: AF199362.1), rat (GenBank: AB073550.1), bat (GenBank: XM_033088803.1), and hamster (GenBank: XP_012971683.1) were synthesized by GenScript (Nanjing, China) and cloned into p3×FLAG-CMV14. The lentiviral packaging plasmid psPAX2 was obtained from Addgene. The lentiviral transfer plasmid pLenti-luc-GFP was a generous gift from Fang Li, Duke University. All mutagenesis experiments were performed using a Q5 mutagenesis kit (NEB, Ipswich, MA, USA).

Production and transduction of S-pseudotyped lentiviruses

For S-pseudotyped lentivirus production, plasmids encoding HKU1 S proteins were co-transfected with psPAX2 and pLenti-Luc-GFP into HEK293T cells using polyetherimide (PEI). After a 40-h incubation, the supernatants containing pseudovirions were subjected to centrifugation at 1,000 × g for 10 min to remove cell debris. For pseudoviral transduction, HEK293 cells transiently overexpressing diverse TMPRSS2 proteins were seeded into 24-well plates. The following day, the cells were incubated with HKU1 S pseudovirions overnight. At 48 h post-transduction, the cells were lysed using the Steady-Glo Reagent from Promega (Madison, WI, USA). The transduction efficiency was assessed by quantifying the luciferase activity with a PerkinElmer EnVision Multilabel Plate Reader.

Detection of TMPRSS2 proteins by Western blot analysis

HEK293 cells were transfected with plasmids encoding TMPRSS2 proteins using PEI. At 40 h after transfection, the cells were lysed with a RIPA buffer solution composed of 50 mM Tris-HCl at pH 7.4, 150 mM NaCl, 1% NP-40, 0.1% SDS, 1 mM EDTA, and 1% sodium deoxycholate, in the presence of protease inhibitors from Selleck (Shanghai, China). The cell lysates were resolved by SDS-PAGE on a 10% gel and were transferred onto a nitrocellulose membrane. The TMPRSS2 proteins were detected using a mouse monoclonal anti-FLAG M2 antibody (1:2,000 dilution). β-actin, used as a loading control, was detected with a mouse monoclonal anti-β-actin antibody (1:5,000 dilution) from Sigma (St. Louis, MO, USA).

Cell–cell fusion assay

HEK293T cells were transfected with plasmids encoding HKU1 S proteins and plasmids encoding EGFP. At 40 h post-transfection, the cells were treated with 0.25% trypsin for 2 min to detach them. Subsequently, roughly one-third of these detached cells were placed onto a layer of HEK293 cells transiently expressing diverse TMPRSS2 proteins. After a 4-h incubation, formation of syncytium was visualized using a Nikon TE2000 epifluorescence microscope equipped with MetaMorph software (Molecular Devices).

Protein expression and purification

Expression and purification of HKU1B RBD proteins were carried out as previously described (60). Briefly, Expi293F cells were transiently transfected with the plasmids encoding mFc-tagged HKU1B RBD proteins using Expifectamine 293 Transfection kit (Life Technologies, USA) according to the manufacturer’s instructions. Cell culture supernatants were collected 72 h later and were purified using Strep-Tactin XT Superflow high-capacity resin (IBA Lifesciences). After washing with 10-bed volumes of wash buffer (20 mM Tris, 200 mM NaCl, pH 8.0), the RBD proteins were eluted by the elution buffer (wash buffer with 50 mM biotin). After elution, the elution buffer was replaced with a sodium citrate buffer (0.1 M sodium citrate, pH 8.0) using a centrifugal filter (Millipore).

Cell-surface protein biotinylation assay

The cell-surface biotinylation assay has been performed as previously described (61). HEK293 cells transiently expressing FLAG-tagged TMPRSS2 proteins were washed with 1× phosphate-buffered saline (PBS) and then detached by 1 mM EDTA. After washing twice with 1× PBS, the cells were incubated with 0.25 mg/mL Z-Link Sulfo-NHS-SS-Biotin (#21331, Thermo Fisher Scientific) for 2 h on ice. The reaction was quenched by adding lysine to a final concentration of 0.5 mg/mL. After washing, the cells were lysated with RIPA lysis buffer on ice for 30 min, and then centrifuged to remove the pellet. The supernatants were incubated with Streptavidin Agarose Resin (#20347, Thermo Fisher Scientific) for 2 h. After washing with RIPA lysis buffer, the cell-surface biotinylated proteins were eluted for 10 min at 98°C in 1× SDS-PAGE sample loading buffer with 100 mM dithiothreitol, and then the cell-surface expression of TMPRSS2 proteins was analyzed by Western blot analysis using antibodies against FLAG tag.

Flow cytometry analysis of HKU1 RBD binding

The receptor-binding assay has been performed as previously described (62). HEK293T cells were transfected with plasmids encoding the WT or mutant TMPRSS2 proteins. After 40 h of transfection, the cells were digested with 1 mM EDTA for 5 min. After washing three times with 3% FBS in 1× PBS, the cells were incubated with 5 µg/mL mFc-tagged HKU1B RBD proteins for 1 h on ice. Subsequently, the cells were washed three times with 3% FBS in 1× PBS, and then incubated with fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse secondary antibody (1:500) (Jackson ImmunoResearch, Pennsylvania, USA). After washing, the cells were then analyzed by flow cytometry.

DATA AVAILABILITY

All relevant data are within the article.