Introduction

In December 2019, an outbreak of mysterious pneumonia was reported in Wuhan, Hubei province, China [1, 2]. A novel coronavirus, called severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which is closely related to SARS-CoV-1, was detected in patients and subsequently identified as the causative agent of the new respiratory disease in January 2020 [2, 3]. SARS-CoV-2 causes coronavirus disease 2019 (COVID-19), which was declared a global pandemic on March 11, 2020. The rapid national and international spread of SARS-CoV-2 has threatened human health and led to a global economic recession [4].

SARS-CoV-2 is the seventh human coronavirus that has been identified, and it belongs to the subgenus Sarbecovirus of the genus Betacoronavirus in the family Coronaviridae [28]. The complete genomic nucleotide sequence of the virus from each passage was deposited in the GenBank database under the accession numbers MZ930250 to MZ930252, MZ930392 to MZ930396, and MZ995184 to MZ995188.

Statistical analysis

All values are expressed as the mean ± standard deviation of the mean difference (SDM). Statistical analysis was conducted using the GraphPad Prism 7 software package (GraphPad Software, San Diego, CA). P-values ≤ 0.05 were considered to be statistically significant.

Results

Effect of trypsin in SARS-CoV-2 infection

To investigate the effect of trypsin treatment on viral replication, Vero E6 cells were incubated for 1 h with trypsin at different concentrations (0, 1.25, 2.5, 5, and 10 µg/ml) and infected with SARS-CoV-2 (P3) at an MOI of 1 for 1 h or mock-infected. The infected cells were washed once with PBS and cultivated in cell growth medium without trypsin. According to the MTT cell viability assay, none of the trypsin dosages examined in this study resulted in a measurable level of cell death (Table 1). We initially quantified virus production by virus titration at 24 hpi. The virus yields remained unchanged, with a mean titer range of 106.53–106.73 TCID50/ml regardless of the presence or absence of trypsin before infection (Fig. 1A). SARS-CoV-2 replication was also measured by monitoring the intensity of the cytopathic effect (CPE) and was confirmed by IFA using an anti-N-protein MAb at 24 hpi. As shown in Fig. 1B, treating cells with trypsin before virus infection had no effect on SARS-CoV-2 infectivity at any of the concentrations tested compared to the untreated control. Furthermore, SARS-CoV-2 propagation in cell culture was unaffected by adding exogenous trypsin during the 1-h inoculation period (Fig. 1C and D). These results revealed that trypsin treatment before or during infection did not influence SARS-CoV-2 replication.

Table 1 Cell viability measured by MTT assay
Full size table
Fig. 1
figure 1

Trypsin-mediated enhancement of SARS-CoV-2 infection in cultured cells. (A and B) Vero E6 cells were preincubated with trypsin for 1 h before infection and then mock infected or infected with SARS-CoV-2 (P3) at an MOI of 1 for 1 h without trypsin addition. The virus-infected cells were then maintained in the absence of trypsin. (C and D) Vero E6 cells were mock infected or infected with SARS-CoV-2 (P3) at an MOI of 1 for 1 h in the presence of trypsin. The virus-infected cells were then maintained in the absence of trypsin. (E and F) Vero E6 cells were mock infected or infected with SARS-CoV-2 (P3) at an MOI of 1 for 1 h without trypsin addition. The virus-infected cells were then maintained in the presence of trypsin. The virus supernatants were collected at 24 hpi, and viral titers were determined (A, C, and E). SARS-CoV-2-specific CPE was monitored daily, and cells were photographed at 24 hpi using an inverted microscope at a magnification of 200× (top panels). For immunostaining, infected cells were fixed at 24 hpi and incubated with an MAb against the SARS-CoV-2 N protein, followed by incubation with Alexa green–conjugated goat anti-mouse secondary antibody (middle panels). The cells were then counterstained with DAPI (bottom panels) and examined under a fluorescence microscope at 200× magnification (B, D, and F). The values shown are the mean of three independent experiments, and error bars show the SDM. *, P = 0.001 to 0.05

Full size image

Next, we investigated whether SARS-CoV-2 infectivity is enhanced by supplementing the medium with trypsin after SARS-CoV-2 infection. Vero E6 cells were infected with SARS-CoV-2 at an MOI of 1 for 1 h without trypsin and then propagated in FBS-free medium with trypsin at different concentrations (0, 1.25, 2.5, 5, and 10 µg/ml). Unless otherwise specified, trypsin was present throughout the infection period. When exogenous trypsin was added after inoculation, virus titers were considerably higher than without trypsin (Fig. 1E). The highest viral titer (107.52 TCID50/ml) was achieved in the presence of 5 µg of trypsin per ml, which was 1 log higher than the control mean titer (106.67 TCID50/ml). CPE observations and IFA analysis revealed that the virus infectivity increased when the virus-infected cells were cultivated in the presence of trypsin (Fig. 1F). In another experiment, Vero E6 cells were inoculated with tenfold serially diluted SARS-CoV-2 and maintained in the presence or absence of trypsin (5 µg/ml), and IFA was used to visualize infected cells (Fig. 2). The results revealed that when Vero E6 cells were propagated in the presence of trypsin, the number of infected cells was more than tenfold higher than in the cultures without trypsin. The virus growth medium (high-glucose DMEM supplemented with penicillin-streptomycin and 5 µg of trypsin per ml) was then used in all subsequent experiments unless otherwise indicated. We also tested the effect of another protease, elastase, on SARS-CoV-2 under the same experimental conditions. However, no apparent effect of elastase on SARS-CoV-2 replication was detected in cell culture when cells were treated with elastase at non-cytotoxic concentrations (5–10 µg/ml; Table 1) before, during, or after infection (Fig. 3).

Fig. 2
figure 2

Comparison of SARS-CoV-2 production in the presence and absence of trypsin. Vero E6 cells were mock infected or infected with tenfold serially diluted SARS-CoV-2 and further cultivated in the presence (A) or absence (B) of trypsin. The virus-infected cells were fixed at 24 hpi and incubated with an MAb against the SARS-CoV-2 N protein, followed by incubation with Alexa green–conjugated goat anti-mouse secondary antibody (top panels). The cells were then counterstained with DAPI (bottom panels) and examined under a fluorescence microscope at 200× magnification.

Full size image
Fig. 3
figure 3

Effect of elastase on SARS-CoV-2 infection. (A and B) Vero E6 cells were preincubated with elastase for 1 h before infection and then mock infected or infected with SARS-CoV-2 (P3) at an MOI of 1 for 1 h without elastase addition. The virus-infected cells were then maintained in the absence of elastase. (C and D) Vero E6 cells were mock infected or infected with SARS-CoV-2 (P3) at an MOI of 1 for 1 h in the presence of elastase. The virus-infected cells were then maintained in the absence of elastase. (E and F) Vero E6 cells were mock infected or infected with SARS-CoV-2 (P3) at an MOI of 1 for 1 h without elastase addition. The virus-infected cells were then maintained in the presence of elastase. The virus supernatants were collected at 24 hpi, and viral titers were determined (A, C, and E). For immunostaining, infected cells were fixed at 24 hpi and incubated with an MAb against the SARS-CoV-2 N protein, followed by incubation with Alexa green–conjugated goat anti-mouse secondary antibody (top panels). The cells were then counterstained with DAPI (bottom panels) and examined under a fluorescence microscope at 200× magnification (B, D, and F). The values shown are the mean of three independent experiments, and error bars show the SDM.

Full size image

To establish the point at which trypsin affects SARS-CoV-2 infection, we added trypsin to Vero E6 cells at different time points after infection. At 24 hpi, the extent of viral replication was assessed directly by virus titration. As shown in Fig. 4, the addition of trypsin at up to 2 hpi (i.e., 3 h after inoculation) resulted in a considerable enhancement of infectivity. However, when trypsin was introduced at or after 4 hpi, little or no increase in virus propagation was observed. These findings established that trypsin must be present during the early stages of viral infection in order to effectively enhance SARS-CoV-2 infection. Taken together, our results suggest that trypsin participates in the entry pathway of SARS-CoV-2.

Fig. 4
figure 4

Effect of proteases on SARS-CoV-2 propagation at early time points postinfection. Vero E6 cells were mock infected or infected with SARS-CoV-2 (P3) at an MOI of 1. At the indicated time points postinfection, trypsin (A) or elastase (B) was added to achieve a final concentration of 5 or 10 µg/ml, respectively. At 12 hpi, the culture supernatant was collected, and virus production was quantified by virus titration. Results are presented as the mean of three independent experiments, and error bars show the SDM. *, P < 0.05; **, P < 0.001

Full size image

Effect of trypsin treatment on virus internalization

We next investigated whether trypsin facilitates the entry of SARS-CoV-2 in cultured cells. To establish the parameters for the viral internalization assay used to monitor SARS-CoV-2 entry, we used a proteinase K infection assay to determine the rate of virus attachment and penetration by quantifying bound and internalized virus particles, respectively. To determine the effectiveness of proteinase K treatment in removing bound viruses, Vero E6 cells were infected with SARS-CoV-2 for 1 h at 4°C and then treated with proteinase K at the specified doses for 45 min at 4°C. The quantity of virus bound to the cells was measured using qRT-PCR to determine the SARS-CoV-2 genome copy number. As shown in Fig. 5A, proteinase K treatment had a substantial effect on the number of virions bound at the cell surface, suggesting that virus attached to the cells is efficiently removed by proteinase K treatment.

Fig. 5
figure 5

Effect of trypsin on SARS-CoV-2 entry. (A) Vero E6 cells were incubated with SARS-CoV-2 at an MOI of 1 at 4°C for 1 h, after which the unbound virus was removed and the cells were treated with proteinase K (0.5–1 mg/ml) at 4°C for 45 min. The cells were collected in TRIzol for RNA isolation and determination of the SARS-CoV-2 RNA copy number. Results are shown as a percentage of SARS-CoV-2 RNA copy number compared with controls (ctrl) in which PBS was substituted for proteinase K. (B) Vero E6 cells were infected with SARS-CoV-2 at an MOI of 1 at 4°C for 1 h and washed with cold PBS. The infected cells were then incubated in the presence or absence of trypsin (5 µg/ml), either at 4°C (binding; blue bars) or 37°C (internalization; red bars), for an additional 1 h. The virus-infected cells that were incubated at 37°C were then treated with proteinase K (0.5 mg/ml) at 37°C for 45 min. The infected cells were then serially diluted and plated onto fresh Vero E6 cells. At 24 h post-incubation, bound or internalized viruses were titrated. Data are expressed as the mean of three independent experiments performed in triplicate, and error bars represent the SDM. *, P < 0.05; **, P < 0.001

Full size image

We then examined the two phases of viral entry – attachment and penetration – in the presence of trypsin using virus binding and internalization assays. Vero E6 cells were inoculated with SARS-CoV-2 at 4°C for 1 h to enable only virus attachment and then maintained at 4° or 37°C in the presence of trypsin to limit or permit virus internalization, respectively, after which proteinase K was used to remove any remaining virus particles from the cell surface. Serial dilutions of infected cells were performed on fresh Vero E6 cell monolayers, and the viral titers were determined (Fig. 5B). Regardless of the presence or absence of trypsin, the viral titers were equivalent among cells treated at 4°C to allow virus binding but inhibit penetration, However, virus production was considerably enhanced by trypsin treatment in cells maintained at 37°C to allow virus internalization to progress when compared to the vehicle control without trypsin under the same conditions. These findings suggest that trypsin treatment accelerates internalization of SARS-CoV-2.

SARS-CoV-2 entry at the cell surface facilitated by trypsin

To test whether SARS-CoV-2 enters cells by a pH-dependent endosomal pathway, we first evaluated the effect of BafA1, a lysosomotropic agent, on SARS-CoV-2 replication by virological analysis. Vero E6 cells were treated with BafA1 at concentrations of 0.1 and 0.5 µM or DMSO (0.5%) as a vehicle control before, during, or after infection. An MTT assay did not reveal any cellular cytotoxicity in the drug-treated cells at the concentrations used in the study (Table 1). Virus production was measured by virus titration and IFA using an anti-N protein MAb at 24 hpi (Fig. 6). Treating cells with BafA1 before or during infection significantly suppressed infection (Fig. 6A and C) and reduced the level of SARS-CoV-2 N expression (Fig. 6B and D) in a dose-dependent manner. By contrast, virus propagation remained unchanged when BafA1 was added after infection (Fig. 6E and F). These results suggest that SARS-CoV-2 entry occurs through an endosomal pathway.

Fig. 6
figure 6

Effect of BafA1 on SARS-CoV-2 infection. (A and B) Vero E6 cells were preincubated with BafA1 for 1 h before infection and then mock infected or infected with SARS-CoV-2 (P3) at an MOI of 1 for 1h without BafA1 addition. The virus-infected cells were then maintained in the absence of BafA1. (C and D) Vero E6 cells were mock infected or infected with SARS-CoV-2 (P3) at an MOI of 1 for 1 h in the presence of BafA1. The virus-infected cells were then maintained in the absence of BafA1. (E and F) Vero E6 cells were mock infected or infected with SARS-CoV-2 (P3) at an MOI of 1 for 1 h without BafA1 addition. The virus-infected cells were then maintained in the presence of BafA1. The virus supernatants were collected at 24 hpi, and viral titers were determined (A, C, and E). For immunostaining, infected cells were fixed at 24 hpi and incubated with an MAb against the SARS-CoV-2 N protein, followed by incubation with Alexa green–conjugated goat anti-mouse secondary antibody (top panels). The cells were then counterstained with DAPI (bottom panels) and examined under a fluorescence microscope at 200× magnification (B, D, and F). The values shown are the mean of three independent experiments, and error bars show the SDM. *, P = 0.001 to 0.05

Full size image

We then investigated the effectiveness of trypsin in facilitating SARS-CoV-2 entry directly at the cell surface. Vero E6 cells treated with BafA1 at a concentration of 0.5 µM were inoculated with SARS-CoV-2 at an MOI of 1 and kept at 4°C for 30 min to block the virus from entering cells. The cells were then treated with different proteases at room temperature for 5 min and maintained at 37°C for 6 h, and the amount of virus that was internalized was measured by qRT-PCR (Fig. 7). Trypsin greatly facilitated SARS-CoV-2 entry, whereas elastase did not influence viral entry (Fig. 7A). Trypsin treatment of cells prior to viral infection had no influence on SARS-CoV-2 internalization (Fig. 7B), showing that the effect of trypsin on the cells themselves is irrelevant for infection. These results imply that SARS-CoV-2 entry occurs via non-endosomal, direct fusion with the plasma membrane in the presence of trypsin, which cleaves the fusion-inducing S protein.

Fig. 7
figure 7

Entry of SARS-CoV-2 at the cell surface facilitated by trypsin. (A) Effect of proteases on the entry of SARS-CoV-2 into Vero E6 cells treated with BafA1. Vero E6 cells cultured in 6-well plates were treated with BafA1 at a concentration of 0.5 µM at 37℃ for 30 min, placed at 4℃ for 30 min, and infected with SARS-CoV-2 at an MOI of 1 for 30 min. Then, the cells were treated with various concentrations of trypsin or elastase at room temperature for 5 min and maintained in the presence of BafA1 for an additional 6 h. The amount of SARS-CoV-2 was measured quantitatively by real-time PCR. Cells not treated with BafA1 or those treated with BafA1 but not treated with trypsin or elastase were used as controls. (B) Effect of trypsin treatment before or after inoculation on SARS-CoV-2 infection in the presence of BafA1. Vero E6 cells were treated with BafA1 (0.5 µM) at 37℃ for 30 min and then treated with trypsin (10 and 30 µg/ml) at room temperature for 5 min before (pre) or after (post) virus inoculation. Viral infectivity was estimated quantitatively by real-time PCR. (C) SARS-CoV-2 kinetics after treatment with trypsin. Vero E6 cells were treated with BafA1, infected with SARS-CoV-2, and treated with 75 µg of trypsin per ml, as described in the legend to Fig. 7A. The production of SARS-CoV-2 was measured by real-time PCR at 3–6 h after virus inoculation. Vero E6 cells without any treatment were also infected as a control (untreated). The viral titers were expressed as genomic copies/ml. Data are expressed as the mean of three independent experiments, and error bars show the SDM. *, P < 0.05; **, P < 0.001

Full size image

Treatment with high concentrations (10–75 µg/ml) of trypsin resulted in an enhancement of virus entry and replication when compared with normal infection without trypsin (Fig. 7A; compare bars 6–8 with bar 1 from left). The replication kinetics of SARS-CoV-2 were then compared in cells treated with BafA1 (0.5 µM) and a high concentration (75 µg/ml) of trypsin to those in virus-infected cells maintained in the absence of BafA1 and trypsin. At each point during the early stage of infection, trypsin-treated cells produced substantially more virus (Fig. 7C).

Phenotypic and genotypic characterization of SARS-CoV-2 serially passaged in the presence or absence of trypsin

We serially propagated viruses in vitro for up to 50 passages in the presence or absence of trypsin and generated virus stocks every 10 passages, labeled P10, P20, P30, P40, and P50. Like the parental virus (P3), all SARS-CoV-2 strains cultured in the absence of trypsin produced CPE typical of viral infection, including cell rounding, clum** together in clusters, and detachment, in infected Vero E6 cells (Fig. 8, left panels). However, trypsin-adapted strains induced different patterns of CPE that included cell fusion and multinucleated cells, or syncytia (Fig. 8, right panels). Vacuoles and syncytia were larger and more predominant in cells infected with SARS-CoV-2 strains that were consecutively passaged in the presence of trypsin. As a consequence, the size of syncytia increased progressively with the serial passage number, and the high-passage P50 virus generated prevalent syncytia with many more nuclei than the low-passage P10 virus in the presence of trypsin (Fig. 8; compare panel n with panel j).

Fig. 8
figure 8

Cytopathic changes in virus-infected cells cultured in the presence or absence of trypsin. Fifty sequential passages were performed in Vero E6 cells in the presence or absence of trypsin. Vero E6 cells were mock infected or infected with each representative cell-adapted SARS-CoV-2 strain (P3, P10, P20, P30, P40, and P50) and maintained in the presence or absence of 5 µg of trypsin per ml. SARS-CoV-2-specific CPE was monitored daily, and cells were photographed at 24 hpi using an inverted microscope at a magnification of 200×.

Full size image

To examine the phenotypic characteristics of serially passaged SARS-CoV-2 strains in vitro, we evaluated the one-step growth rates of representative P3 and P10 strains cultured with or without trypsin (Fig. 9). Without trypsin treatment, the parental P3 virus had the fastest growth rate, showing a peak titer of 105.57 TCID50/ml at 24 hpi, after which its growth declined (Fig. 9A). By contrast, the growth rate of the P3 virus increased significantly in the presence of trypsin. At 12 hpi, the titer of the P3 virus increase rapidly in the presence of trypsin, continuing until 48 hpi, ranging from 106.33 to 107 TCID50/ml. In contrast, cell-adapted virus serially passaged under trypsin-free conditions produced growth curves and virus titers that were similar to those of the P3 virus, with a maximum at 24 hpi (Fig. 9B). Despite the comparable growth curves of the P3 and P10 strains, the trypsin-adapted virus grew faster and produced higher titers as the passage number increased. In particular, cells infected with the trypsin-adapted P10 virus reached virus titers of 107.93 TCID50/ml at 12 hpi, which continued to increase up to 48 hpi, reaching a maximum that was 100-fold higher than that of the parental or passaged strain without trypsin addition. Compared to the P10 virus cultured in the presence of trypsin, the growth patterns were comparable among further trypsin-adapted P20–P50 strains passaged in the presence of trypsin (data not shown). These results demonstrate that trypsin is able to promote SARS-CoV-2 infection by facilitating cell-to-cell fusion.

Fig. 9
figure 9

One-step growth kinetics of SARS-CoV-2 strains passaged in the presence or absence of trypsin. Vero E6 cells were infected with each representative cell-adapted SARS-CoV-2 strain P3 (A) and P10 (B) and maintained in the presence or absence of 5 µg of trypsin per ml. At the indicated time points postinfection, culture supernatants were harvested, and virus titers were determined. Results are expressed as the mean of three independent experiments performed in triplicate, and error bars show the SDM. *, P < 0.05; **, P < 0.001

Full size image

To evaluate the genomic alterations that may have occurred during in vitro serial passage in Vero E6 cells in the presence or absence of trypsin, we determined the full-length nucleotide sequences of the parental P2 and its derived passages, P3–P50, using Sanger sequencing and RACE. The sequence data did not show any mutations in the P3 virus propagated in the presence (+) or absence (−) of trypsin compared with the original KCDC03 P2 strain. Although the 5′ and 3′ untranslated regions (UTRs) remained unchanged during in vitro serial passage in Vero E6 cells, mutations occurred in the protein-coding regions and the number of mutations increased gradually over time. At the nucleotide level, the genome sequences of P3(+) and P3(−) were nearly identical (99.94–99.99%) to that of the corresponding cell-adapted strain. In comparison to the parental P3(+) or P3(−) strain, the number of nucleotide/amino acid substitutions increased in direct proportion to the number of in vitro passages (Table 2).

Table 2 The number of nucleotide and amino acid differences between the parental KNU-SARS-CoV-2 P3 and cell-adapted viruses
Full size table

Interestingly, the number and location of the amino acid (aa) changes differed between the cell-culture-passaged strains in the presence or absence of trypsin (Fig. 10). The 50th-passage strain without trypsin addition contained nine aa mutations, including three aa deletions (DELs), whereas the P50(+) virus had 23 aa variations, including 13 aa DELs. The nine aa mutations in the P50(−) virus were distributed in open reading frames (ORFs) 1a, 2, 4, and 9, encoding nsp1, S, E, and N, respectively. Among these, five aa changes were in the S protein of P50(−). Notably, one aa mutation (R685S) was found in the S1/S2 furin cleavage site (FCS), which contains multiple basic amino acids (681PRRAR685) (Table 3), which occurred in P10(−). One- (M) and two- (QA) amino-acid DELs emerged independently in nsp1 and N, respectively, during the cell culture passages in the absence of trypsin, the former in P30(−) and the latter in P10(−) (Table 3). The M-DEL at position 85 in nsp1 resulted from a three-nucleotide (AUG) DEL covering the codon for methionine (M) at positions 253–255 in ORF1a (nt 490–492 at the genome level). At positions 418 and 419 in N, the QA-DEL arose from a C-to-T substitution (C1252T) at position 1252 in ORF9 (nt 29,497 at the genome level). This change resulted in a change from CAG coding for glutamine (Q) to a TAG terminator codon at positions 1252–1254 in ORF9 (nt 29,497–29,499 at the genome level), causing a premature termination resulting in a loss of two residues (Q and A) from the C-terminal end of the N protein.

Fig 10
figure 10

Schematic diagram of the amino acid differences between SARS-CoV-2 (P3) and its cell-adapted decedents (P10–P50). The organization of the SARS-CoV-2 genome, which is approximately 29.8 kb in length, is shown at the top. In the first diagram, blue arrows indicate the genes encoding nonstructural proteins (nsp1–16). In the second illustration, orange bars represent the identified ORFs. Light gray arrows represent the 5′ and 3′ untranslated regions. Lightly shaded areas are identical to those of SARS-CoV-2 (P3), and the vertical black bars represent individual amino acid positions where viruses from later passages differ from the P3 virus. Thin horizontal dashed lines indicate deletions. "S1/S2 FCS" represents the S1/S2 furin cleavage site (FCS), which contains multiple basic amino acids (681PRRAR685), and the vertical red bars indicate the R685S mutation in the S1/S2 FCS.

Full size image
Table 3 Summary of amino acid mutations during in vitro serial passages
Full size table

The 23 aa variations present in the viral genome after the 50th passage in the presence of trypsin were dispersed randomly in ORFs 1a, 1b, 2, 5, 6, 7b, and 8, encoding nsp6, nsp13, S, M, and accessory proteins, respectively (Table 2). Although the number (five) of aa mutations in the S protein of P50(+) was identical to that in P50(−), their positions in the P50(+) S protein were completely different from those in P50(−) (Fig. 10). Intriguingly, a 13-aa DEL occurred at positions 31–43 in ORF7b of P40(+) and was maintained until the 50th cell culture passage in the presence of trypsin (Table 3). The S–A-DEL resulted from a C-to-A substitution (C92A) at position 92 in ORF7b (nt 27,819 at the genome level). This mutation changed the sequence TCA, encoding serine (S), to a TAA termination codon at positions 91–93 in ORF7b (nt 27,818–27,820 at the genome level), leading to an early termination, eliminating 13 aa residues from the C-terminus of ORF7b.

Discussion

Receptor binding and subsequent proteolytic priming of the S protein are prerequisites for coronavirus entry into host cells. The priming event can be mediated by various host cell proteases, including trypsin, TMPRSS2, furin, and cathepsin [29]. Trypsin has been shown previously to promote the replication of porcine coronaviruses, including the emerging/re-emerging swine coronaviruses PEDV and PDCoV, and is therefore commonly used in their isolation and cultivation [15, 17, 19]. Furthermore, protease-mediated enhancement of infection is known for SARS-CoV-1, as well as viruses belonging to other families, such as influenza and parainfluenza viruses [13]. In this study, we demonstrated that trypsin treatment enhances SARS-CoV-2 infection in cultured cells when it is added at early time points in infection (at least until 2 hpi).

The entry process of SARS-CoV-2 begins with the attachment of viral particles to the cell surface through the interaction between the viral S protein and its cellular receptor ACE2 [36, 37]. The trypsin-adapted strains formed more extensive syncytia, leading to an increase in vacuolated areas, which were proportional to the number of cell culture passages. If the proteolytic cleavage of the S protein occurs within the functional S1/S2 FSC, mutations in S2 might be involved in the formation of syncytia by cells infected with the trypsin-adapted virus. It is therefore notable that N960I and V961A mutations were identified in S2 of the P20(+) through to P50(+) strains and that additional K849R and Q949R mutations emerged in the S2 fusion domain of P50(+), suggesting that these genetic changes may contribute to cytopathology. Interestingly, the trypsin-adapted P40(+) and P50(+) viruses had a large 13-aa DEL in a short ORF7b protein composed of 43 aa residues. Although the precise role of this accessory protein is unknown, we cannot exclude the involvement of ORF7b in SARS-CoV-2 propagation in vitro.

In conclusion, the present study showed that trypsin enhances the replication of SARS-CoV-2 in cultured cells and facilitates viral entry by promoting a direct fusion process at the cell surface. Our findings indicate that SARS-CoV-2 has the potential to use different pathways to enter cells, depending on the presence of trypsin. Remarkably, the trypsin-triggered SARS-CoV-2 non-endosomal entry at the cell surface facilitated more-efficient infection than the endosomal pathway in the absence of trypsin. Moreover, the Vero-E6-cell-adapted SARS-CoV-2 strains grown in the presence of trypsin exhibited clear syncytia formation and robust infection in cultured cells, whereas the strains without trypsin addition failed to form syncytia in infected cells and to enhance viral infection. These results suggest that trypsin or other proteases produced in the lungs or small intestine might enhance the replication of SARS-CoV-2 in these organs, leading to severe tissue damage. Conversely, these data may provide a practical methodology to improve the isolation and propagation of SARS-CoV-2 for the development of vaccines and other therapeutic agents. We also identified genetic mutations during serial in vitro passages in the presence or absence of trypsin that might be involved in the proteolytic priming or fusion activity of the S protein. Thus, our genetic data provide fundamental insights for future research involving reverse genetics to investigate the specific effect of mutations that occur during cell adaptation on SARS-CoV-2 replication and pathogenesis.