The SARS-CoV-2 Spike is a virulence determinant and plays a major role on the attenuated phenotype of Omicron virus in a feline model of infection

ABSTRACT

The role of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Omicron BA.1 Spike (S) on disease pathogenesis was investigated. For this, we generated recombinant viruses harboring the S D614G mutation (rWA1-D614G) and the Omicron BA.1 S gene (rWA1-Omi-S) in the backbone of the ancestral SARS-CoV-2 WA1 strain genome. The recombinant viruses were characterized in vitro and in vivo. Viral entry, cell-cell fusion, plaque size, and the replication kinetics of the rWA1-Omi-S virus were markedly impaired when compared to the rWA1-D614G virus, demonstrating a lower fusogenicity and ability to spread cell-to-cell of rWA1-Omi-S. To assess the contribution of the Omicron BA.1 S protein to SARS-CoV-2 pathogenesis, the pathogenicity of rWA1-D614G and rWA1-Omi-S viruses was compared in a feline model. While the rWA1-D614G-inoculated cats were lethargic and showed increased body temperatures on days 2 and 3 post-infection (pi), rWA1-Omi-S-inoculated cats remained subclinical and gained weight throughout the 14-day experimental period. Animals inoculated with rWA1-D614G presented higher infectious virus shedding in nasal secretions, when compared to rWA1-Omi-S-inoculated animals. In addition, tissue replication of the rWA1-Omi-S was markedly reduced compared to the rWA1-D614G, as evidenced by lower viral load in tissues on days 3 and 5 pi. Histologic examination of the nasal turbinate and lungs revealed intense inflammatory infiltration in rWA1-D614G-inoculated animals, whereas rWA1-Omi-S-inoculated cats presented only mild to modest inflammation. Together, these results demonstrate that the S protein is a major virulence determinant for SARS-CoV-2 playing a major role for the attenuated phenotype of the Omicron virus.

IMPORTANCE

We have demonstrated that the Omicron BA.1.1 variant presents lower pathogenicity when compared to D614G (B.1) lineage in a feline model of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. There are over 50 mutations across the Omicron genome, of which more than two-thirds are present in the Spike (S) protein. To assess the role of the Omicron BA.1 S on virus pathogenesis, recombinant viruses harboring the S D614G mutation (rWA1-D614G) and the Omicron BA.1 Spike gene (rWA1-Omi-S) in the backbone of the ancestral SARS-CoV-2 WA1 were generated. While the Omicron BA.1 S promoted early entry into cells, it led to impaired fusogenic activity and cell-cell spread. Infection studies with the recombinant viruses in a relevant naturally susceptible feline model of SARS-CoV-2 infection here revealed an attenuated phenotype of rWA1-Omi-S, demonstrating that the Omi-S is a major determinant of the attenuated disease phenotype of Omicron strains.

INTRODUCTION

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was first detected in a cluster of people presenting with severe pneumonia in Wuhan, Hubei province in China (1, 2–4). Since the emergence of SARS-CoV-2 in the human population, there have been 773 million confirmed coronavirus disease 2019 cases and over 7 million deaths, reported to the World Health Organization (WHO) as of 11 January 2024. While circulating in the human population, SARS-CoV-2 has continued to evolve, with new variants causing significant waves of infection worldwide. Importantly, several of the emerging variants present altered transmissibility, immune evasion capability, and pathogenicity (5). The first mutation established on the SARS-CoV-2 genome was the D614G substitution in the Spike (S) protein, detected for the first time in February 2020 and which is present in all SARS-CoV-2 lineages circulating worldwide since then (6). As genome mutations accumulated and viral properties changed over time, some SARS-CoV-2 lineages were classified as variants of concern (VOC) by the WHO (7). As of October 2023, five VOC have been defined: Alpha, Beta, Gamma, Delta, and Omicron. All these VOC rapidly became dominant upon their emergence, outcompeting previous variants regionally or globally. The SARS-CoV-2 Omicron VOC (BA.1) emerged in late 2021, and rapidly achieved global predominance by early 2022 (8, 9). A marked increase in the case numbers was observed worldwide when the Omicron BA.1 variant became predominant owing to high re-infection and vaccine breakthrough rates of this virus due to its ability to evade neutralizing antibody responses (10–15). Hospitalizations and deaths, however, were lower during the Omicron BA.1 surge, suggesting lower pathogenicity of the virus compared to previous SARS-CoV-2 lineages (16). It is possible, however, that pre-existing immunity in the human population generated by vaccination and/or previous infection(s) could also be a confounding factor in the perceived lower severity of infections with the Omicron variant (17).

The Omicron BA.1 sublineage differs from the ancestral Wuhan-Hu-1 virus by 59 amino acids, with 37 of these changes present in the S protein, a key viral protein involved in virus entry (receptor engagement and fusion) (18). Among these, 15 amino acid substitutions are in the receptor binding domain (RBD) with several additional mutations located near the furin cleavage site at the S1/S2 junction and near the S2ʹ site which could modulate host protease cleavage by furin and transmembrane serine proteases (TMPRSS2), respectively, potentially resulting in altered entry preference and fusogenicity, and likely affecting replication and pathogenicity (18–21). The Omicron BA.1 is less dependent on TMPRSS2, less efficient in S cleavage, less fusogenic, and presents an altered propensity to utilize the plasma membrane and endosomal pathways for virus entry. Specific mutations in the S protein have been linked to the attenuated phenotype of the Omicron BA.1 variant (19, 22). Additionally, amino acid mutations in the receptor binding motif contribute to escape from vaccine-induced humoral immune responses (18).

Experimental studies in animal models showed that the Omicron variant causes attenuated disease in humanized mice and hamsters (18, 23–25). Using a feline model, a naturally susceptible animal species, we have demonstrated that the Omicron BA.1.1 variant presents lower pathogenicity when compared to D614G (B.1) and the Delta lineages (26). To assess the contribution of the S protein in this viral phenotype, we generated recombinant viruses based on the ancestral SARS-CoV-2 WA1 strain backbone harboring the S D614G mutation (rWA1-D614G) and the Omicron BA.1 S gene (rWA1-Omi-S). The role of the S protein on virus entry/infectivity, fusion, and replication was characterized in vitro, and the infection dynamics, tissue tropism, and pathogenicity of the viruses were assessed and compared in vivo using the highly susceptible feline model of infection and pathogenesis.

RESULTS

The Omicron BA.1 S induces efficient viral entry, but leads to reduced replication of SARS-CoV-2 in vitro

The role and contribution of Omicron BA.1 S for virus infectivity, fusogenicity, cell-to-cell spread, and replication were investigated in cell culture in vitro. Two recombinant viruses based on the ancestral SARS-CoV-2 WA1 genome backbone harboring the S D614G mutation (rWA1-D614G) and the Omicron BA.1 S gene (rWA1-Omi-S) were generated and rescued (Fig. 1A). Initially, infectivity of rWA1-D614G and rWA1-Omi-S were investigated in cells stably expressing human (h) or cat (c) angiotensin converting enzyme 2 (hACE2_BHK21 and cACE2_BHK21, respectively), Vero E6, and Vero E6 TMPRSS2 cells and virus infectivity was determined by immunofluorescence assay (IFA) followed by quantification of infected cells. In both ACE2 expressing BHK21 cells and in Vero E6 TMPRSS2 cells, the rWA1-Omi-S virus showed higher infectivity at early time points of 4 h and 8 h post-infection (hpi) when compared to the rWA1-D614G (Fig. 1B and C). At 12 hpi, both viruses showed comparable infectivity in hACE2_BHK21 and cACE2_BHK21 cells as evidenced by similar percentage of infected cells; however, from 12 hpi onward, the rWA1-D614G overtook the rWA1-Omi-S virus, and at 24 hpi, a significantly higher percentage of infected cells were noted in rWA1-D614G-infected cells (Fig. 1B and C). On the other hand, in Vero E6 cells, both recombinant viruses showed comparable infectivity throughout the course of infection. We then performed multi-step growth curves with both rWA1-D614G and rWA1-Omi-S in hACE2 and cACE2 expressing BHK21 cells, and Vero E6 cells, to assess the effect of S on virus replication. In both ACE2 expressing BHK21 cells, the rWA1-Omi-S virus presented slightly lower viral yields, suggesting lower replication ability of this virus, in comparison to the ancestral rWA1-D614G virus (Fig. 1D). However, both recombinant viruses presented comparable replication kinetics in Vero E6 cells (Fig. 1D). These results are consistent with the infectivity assays performed in these cells (Fig. 1B and C).

Fig 1.

The role and contribution of Omicron BA.1 S to SARS-CoV-2 infectivity and replication. (A) Schematic presentation of the rSARS-CoV-2 harboring D614G (rWA1-D614G) and Omicron BA.1 S (rWA1-Omi-S). (B) IFA. BHK21 stably expressing human (hACE2_BHK21) and cat (cACE2_BHK21) ACE2, Vero E6, and Vero E6 TMPRSS2 cells were infected at 4°C [multiplicity of infection (MOI) = 0.1] with rWA1-D614G and rWA1-Omi-S viruses, adsorbed at 4°C for 1 h and then transferred and incubated at 37°C. At the indicated time points post-infection, cells were fixed and stained with a SARS-CoV-2 antibody against the viral nucleocapsid (N) protein (red) and 4′,6-diamidino-2-phenylindole (nucleus, blue). (C) Cells infected in (B) were counted and infectivity expressed as the percentage of infected cells based on the number of N-positive cells over the total number of cells. A total of four fields were counted per experiment. Three independent experiments were performed. (D) Multi-step growth curves. hACE2_BHK21, cACE2_BHK21, and Vero E6 cells were infected at 4°C (MOI = 0.1) with rWA1-D614G and rWA1-Omi-S, adsorbed at 4°C for 1 h and then transferred and incubated at 37°C. Cells were harvested at the indicated times post-infection, and virus titers were determined by limiting dilutions and expressed as 50% tissue culture infectious dose (TCID₅₀) per milliliter. The limit of detection for infectious virus titration was 10^1.05 TCID₅₀.mL⁻¹. Data are the means of three independent experiments (n = 3) ± SEM. Two-way analysis of variance followed by multiple comparisons test, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

The Omicron BA.1 S reduces fusogenicity and cell-to-cell spread of SARS-CoV-2 in vitro

The fusogenic activity of the Omicron BA.1 was evaluated and compared to the D614G S. For this, BHK21 cells stably expressing hACE2 or cACE2 were transfected with plasmids encoding the S protein from D614G and Omicron BA.1, and the fusogenic activity of the respective variant proteins was evaluated by determining the number of nuclei incorporated into the syncytia observed in transfected cells. Expression of Omicron BA.1 S in both hACE2- and cACE2-expressing BHK21 cells resulted in lower fusogenic activity as evidenced by significantly (P < 0.001) lower syncytia formation/nuclei incorporation when compared to expression of the ancestral D614G S protein (Fig. 2A and B). Notably, the lower fusogenicity of Omicron BA.1 S protein was confirmed in the context of virus infection. BHK21 cells expressing hACE2 or cACE2 infected with rWA1-Omi-S presented markedly lower fusogenicity as evidenced by lower number and smaller syncytia when compared to cells infected with the ancestral rWA1-D614G virus (Fig. 2C and D). Next, we assessed the effect of Omicron S protein on cell-to-cell spread of SARS-CoV-2 by using plaque assays. The rWA1-Omi-S formed significantly smaller plaques (P < 0.001) when compared to the rWA1-D614G virus in all three cell lines tested (hACE2_BHK21, cACE2_BHK21, and Vero E6) (Fig. 2E and F), indicating a reduced ability of the rWA1-Omi-S to spread from cell-to-cell and confirming the contribution of S protein for this phenotype. Collectively, these results demonstrate that while the Omicron BA.1 S enables efficient virus entry and infectivity, it reduces the overall fusogenicity and ability of the rWA-Omi-S virus to spread from cell-to-cell.

Fig 2.

The rSARS-CoV-2 WA1 harboring the Omicron BA.1 S (rWA1-Omi-S) presents reduced fusogenicity in vitro. (A) BHK21 cells stably expressing human (hACE2_BHK21) or cat ACE2 (cACE2_BHK21) were transfected with plasmids carrying either the D614G or the Omicron BA.1 S proteins and syncytia formation was visualized by IFA. (B) The fusogenic activity of the D614G and the Omicron BA.1 S was quantified by counting the incorporation of nuclei into syncytia observed in (A). The percentage of nuclei observed per field (n = 5) and incorporated into syncytia were calculated over the total number of nuclei per field. (C) Cells stably expressing human (hACE2_BHK21) and cat (cACE2_BHK21) ACE2 were inoculated (multiplicity of infection = 0.1) with rWA1-D614G and rWA1-Omi-S, adsorbed at 4°C for 1h and then transferred and incubated at 37°C. Syncytia formation was visualized by IFA using an N-specific monoclonal antibody (mAb). (D) The fusogenic activity of the rWA-D614G and the rWA1-Omi-S was quantified by counting the incorporation of nuclei into syncytia observed in (C). The percentage of nuclei observed per field (n = 4) and incorporated into syncytia was calculated over the total number of nuclei. (E) Plaque phenotype of the rWA1-D614G and rWA1-Omi-S viruses. hACE2_BHK21, cACE2_BHK21, and Vero E6 cells were infected with rWA1-D614G and rWA1-Omi-S and overlaid with 0.5% agarose medium. Plates were incubated at 37°C for 72 h, the agar overlay was removed, cells were fixed, and the monolayer was stained with 0.5% crystal violet. (F) Viral plaque size. The diameters of viral plaques were measured using a scale in millimeters. Data indicate means ± SEM. Two-way analysis of variance followed by multiple comparisons test, ***P < 0.001, ****P < 0.0001.

SARS-CoV-2 D614G and Omicron BA.1 S present similar ability to inhibit interferon-beta (IFN-β)-mediated luciferase activity

Given the reduced ability of the Omicron BA.1 strain to spread from cell-to-cell, we investigated the potential role of IFN pathway in inhibiting viral replication. The SARS-CoV-2 S protein has been shown to inhibit the IFN pathway by antagonizing viral RNA pattern recognition receptor RIG-I signaling (27). Using luciferase reporter assays, we assessed the effect of D614G and Omicron BA.1 S proteins on activation of the IFN-β and NF-κB signaling pathways. For this, HEK293T cells were co-transfected with plasmids encoding D614G or Omicron BA.1 S proteins, or an empty vector, and with IFN-β or NF-κB firefly luciferase reporter plasmids. At 24 h post-transfection, cells were stimulated with Sendai virus [SeV (Cantell strain), potent IFN inducer] or tumor necrosis factor alpha (TNF-α) (potent NF-κB pathway inducer) and incubated for 12 h. After incubation, firefly luciferase activity was measured and used to assess and compare the inhibitory activity of the D614G and Omicron BA.1 S proteins on these critical signaling pathways. Both D614G and Omicron BA.1 S proteins significantly (P < 0.01) inhibited SeV-induced IFN-β signaling, and no differences between the S variants were noted (Fig. 3A). No inhibitory activity in the NF-κB signaling was observed (Fig. 3B and C). Expression of D614G and Omicron BA.1 S in transfected cells was confirmed by Western blot (Fig. 3D). Importantly, results from the Western blot also demonstrated a lower cleavability of the precursor Omicron BA.1 S protein, as evidenced by markedly lower levels of the S2 subunit in Omicron BA.1 S-transfected cell when compared to D614G S (Fig. 3D). These findings suggest that the lower fusogenicity of the Omicron BA.1 S and the impaired ability of the rWA1-Omi-S virus to spread from cell-to-cell are likely due to a less efficient cleavage of S, and further rule out the involvement of the innate IFN response on this important biological property.

Fig 3.

Innate immune inhibition by SARS-CoV-2 S protein. HEK293T cells were transiently co-transfected with plasmids encoding D614G and Omicron BA.1 S proteins and with IFN-β (A) or NF-κB (B and C) firefly luciferase reporter plasmids. At 24 h post-transfection, cells were then stimulated with SeV (A and B) or TNF-α (C) for 12 h. Cell lysates were harvested, and the Firefly luciferase activity was measured using a luminometer. The luminescence ratio obtained from the target reporters (IFN-β-Luc or NF-κB-Luc) and from the control Renilla luciferase reporter (pRN-Luc) was calculated to normalize for transfection efficiency. The relative IFN-β- or NF-κB-driven luciferase activity was calculated as fold change over the unstimulated cells. (D) HEK293T cells were transfected with plasmids encoding D614G and Omicron BA.1 S proteins and protein expression was detected by Western blot using an antibody against the S protein. β-actin was included as a loading control. (A–C) Data indicate mean ± SEM from four independent experiments. One-way analysis of variance with multiple comparison test, **P < 0.01; ns, not significant.

The rWA1-Omi-S leads to subclinical infection and limited viral shedding in cats

The dynamics of infection, virus replication, and virus shedding of rWA1-D614G and rWA1-Omi-S were assessed in domestic cats. Animals were inoculated intranasally with 5 × 10⁵ plaque-forming units (PFU) of rWA1-D614G or rWA1-Omi-S, and the clinical parameters of SARS-CoV-2 infection, including rectal temperature, body weight, and clinical signs of respiratory disease, were monitored daily (Fig. 4A). Animals inoculated with the rWA1-D614G became lethargic and presented increased body temperatures on days 2 and 3 post-infection (pi) (Fig. 4B), whereas rWA1-Omi-S-inoculated cats remained subclinical. Additionally, rWA1-D614G-inoculated cats lost weight between days 1 and 7 pi, while all rWA1-Omi-S-inoculated cats gained weight throughout the 14-day experimental period (1%–12% of their initial body weight) similar to the control mock-inoculated animals (Fig. 4C).

Fig 4.

rWA1-Omi-S leads to subclinical infection and limited viral shedding in cats. (A) Experimental study design. Body temperature (B) and body weight (C) following intranasal inoculation of rWA1-D614G and rWA1-Omi-S throughout the 14-day experimental period. SARS-CoV-2 RNA loads quantified by real-time reverse transcriptase PCR (rRT-PCR) in nasal (D) and oral (E) secretions, and feces (F) collected on days 0, 1, 3, 5, 7, 10, and 14 pi and in bronchoalveolar lavage fluid (BALF) (G) collected on days 3, 5, and 14 pi. Infectious SARS-CoV-2 loads in nasal (H) and oral (I) secretions, feces (J), and BALF (K) determined by virus titrations in rRT-PCR-positive samples. Virus titers were determined using endpoint dilutions and expressed as 50% tissue culture infectious dose (TCID₅₀) per mL. The limit of detection for infectious virus titration was 10^1.05 TCID₅₀.mL⁻¹. Data indicate mean ± SEM of three to nine animals per group per time point. Two-way analysis of variance with multiple comparison test, *P < 0.05; **P < 0.01; ***P < 0.005; ****P < 0.0001.

To assess virus the dynamics of replication and shedding of the recombinant SARS-CoV-2, nasal and oropharyngeal secretions and feces were collected using nasal (NS), oropharyngeal (OPS), and rectal swabs (RS). The samples were initially tested for the presence of SARS-CoV-2 RNA by real-time reverse transcriptase PCR (rRT-PCR). Viral RNA was detected between days 1 and 14 pi in NS and OPS secretions in the inoculated cats, with higher viral RNA loads detected between days 3 and 5 pi, which decreased thereafter through day 14 pi (Fig. 4D and E). Viral RNA load was higher in nasal secretions of rWA1-D614G-inoculated animals when compared to rWA1-Omi-S-inoculated cats on days 1 and 3 pi (P < 0.001). Similarly, viral RNA load was higher in OPS secretions of rWA1-D614G-inoculated animals when compared to rWA1-Omi-S-inoculated cats on days 1, 7, and 10 pi (P < 0.05) (Fig. 4E). Shedding of viral RNA in feces was markedly lower than in respiratory and oropharyngeal secretions and was characterized by intermittent detection of low amounts of viral RNA in feces, with rWA1-Omi-S-inoculated animals presenting lower RNA levels in feces on day 3 pi (P < 0.001) (Fig. 4F). All control cats (mock-inoculated) remained negative throughout the 14-day experimental period.

The dynamics of infectious virus shedding were also assessed in NS, OPS, and RS for the presence of infectious virus. High viral loads were detected in NS and OPS secretions of rWA1-D614G-inoculated cats (Fig. 4H and I). All rWA1-D614G-inoculated cats shed infectious SARS-CoV-2 between days 1 and 7 pi in nasal secretions, with viral titers ranging from 2.0 to 6.8 log 50% tissue culture infectious dose (TCID₅₀).mL⁻¹, whereas rWA1-Omi-S-inoculated animals shed significantly lower viral titers ranging from 1.8 to 5.5 log TCID₅₀.mL⁻¹ (Fig. 4H). Cats inoculated with rWA1-D614G shed infectious virus between days 1 and 7 pi in the oropharyngeal secretions, with viral titers ranging from 1.8 to 5.8 log TCID₅₀.mL⁻¹, while rWA1-Omi-S-inoculated cats shed lower viral titers ranging from 1.3 to 4.55 log TCID₅₀.mL⁻¹ (Fig. 4I). No infectious virus shedding was detected in feces in any of the groups (Fig. 4J).

We also assessed the viral RNA load and infectious virus titers in bronchoalveolar lavage fluid (BALF). Three cats from each group (control, rWA1-D614G, rWA1-Omi-S) were humanely euthanized on days 3, 5, and 14 pi (Fig. 4A). Viral RNA was detected in BALF of all inoculated animals on days 3 and 5 pi (Fig. 4G). All control cats (mock-inoculated) tested negative by rRT-PCR. On day 3 pi, infectious virus titers in BALF of cats infected with rWA1-D614G varied from 3.5 to 5.0 log TCID₅₀.mL⁻¹, whereas rWA1-Omi-S-inoculated animals presented lower infectious virus titers ranging from 1.3 and 2.5 log TCID₅₀.mL⁻¹ (P < 0.001) (Fig. 4K). On day 5 pi, viral titers in BALF from cats infected with rWA1-D614G varied from 3.0 to 5.0 log TCID₅₀.mL⁻¹, while rWA1-Omi-S-inoculated cats presented viral titers from 2.8 to 3.5 log TCID₅₀.mL⁻¹ (Fig. 4K). Viral RNA was detected in BALF from all inoculated cats on day 14 pi, (Fig. 4G); however, no infectious virus was detected (Fig. 4K). Together, these results demonstrate that the rWA1-Omi-S presents lower pathogenicity and replication ability when compared to the rWA1-D614G recombinant virus in cats.

rWA1-Omi-S presents reduced replication in tissues

The tissue tropism and replication sites of rWA1-D614G and rWA1-Omi-S were assessed following intranasal inoculation in cats. For this, nasal turbinate, palate/tonsil, retropharyngeal lymph node (LN), trachea, lung, mediastinal LN, heart, liver, spleen, kidney, small intestine, and mesenteric LN were collected from three cats per group on days 3, 5, and 14 pi following euthanasia and processed for rRT-PCR, infectious virus titration, and in situ hybridization and immunofluorescence staining. SARS-CoV-2 RNA was detected in several tissues sampled from each group, with higher viral RNA loads being detected on days 3 and 5 pi (Fig. 5A through C). The highest viral RNA loads were detected in the nasal turbinate on days 3 and 5 pi in rWA1-D614G-inoculated animals, with lower viral RNA loads detected in tissues from rWA1-Omi-S-inoculated cats (Fig. 5A and B). To assess virus replication in tissues, subgenomic viral RNA (sgRNA) was quantified by real-time quantitative PCR (RT-qPCR) targeting the envelop (E) gene. sgRNA was consistently detected in nasal turbinate, palate/tonsil, trachea, retropharyngeal LN, and lung from rWA1-D614G-inoculated cats on days 3 and 5 pi (Fig. 5D and E). The highest sgRNA loads were observed in nasal turbinate on days 3 and 5 pi (Fig. 5D and E). sgRNA was detected in nasal turbinate, soft palate/tonsil, retropharyngeal LN, trachea, and lung from rWA1-D614G-inoculated cats, whereas sgRNA was detected only in nasal turbinate and trachea from rWA1-Omi-S-inoculated cats on day 3 pi (Fig. 5D). On day 5 pi, sgRNA was detected in nasal turbinate, soft palate/tonsil, retropharyngeal LN, trachea, lung, and in mesenteric LN from one out of three cats regardless of the virus inoculated (Fig. 5E). On day 14 pi, viral RNA loads decreased when compared to early time points and sgRNA load was higher in retropharyngeal LN of rWA1-D614G-inoculated cats when compared to rWA1-Omi-S-inoculated animals (P < 0.001) (Fig. 5F). All tissues from the control animals (mock-inoculated) tested negative for viral RNA by RT-PCR (Fig. 5A through F).

Fig 5.

rWA1-Omi-S presents reduced replication in tissues. SARS-CoV-2 RNA loads assessed by rRT-PCR in tissues of cats collected on days 3 (A), 5 (B), and 14 (C) pi. Subgenomic SARS-CoV-2 RNA (sgRNA) loads (replicating RNA) assessed by RT-qPCR in tissues of cats collected on days 3 (D), 5 (E), and 14 (F) pi. Infectious SARS-CoV-2 in tissues assessed by virus titration in rRT-PCR-positive tissue samples collected on days 3 (G), 5 (H), and 14 (I) pi. Virus titers were determined using endpoint dilutions and expressed as TCID₅₀.mL⁻¹. The limit of detection for infectious virus titration was 10^1.05 TCID₅₀.mL⁻¹. Data indicate mean ± SEM of three animals per group per time point. Two-way analysis of variance with multiple comparison test, *P < 0.05; **P < 0.01; ***P < 0.005; ****P < 0.0001.

The presence and titers of infectious SARS-CoV-2 in tissues were next assessed by virus titrations in rRT-PCR-positive tissues. Detection of infectious virus and infectious viral loads was consistent with detection of sgRNA (Fig. 5D through I). Infectious SARS-CoV-2 was consistently detected in respiratory tract tissues on days 3 and 5 pi including in nasal turbinate, palate/tonsil, retropharyngeal LN, trachea, and lung from cats from both groups (Fig. 5G and H). Viral loads, however, were higher in tissues from rWA1-D614G-inoculated cats. The highest viral titers were observed in the nasal turbinate (titers ranging 5.8 to 7.3 log TCID₅₀.mL⁻¹) from rWA1-D614G-inoculated cats on days 3 and 5 pi, while markedly lower viral loads (4.0 to 5.0 log TCID₅₀.mL⁻¹) were detected in rWA1-Omi-S inoculated cats (Fig. 5G and H). Interestingly, although, on day 5 pi, infectious virus titers detected in rWA1-D614G-inoculated animals were lower than in day 3 pi, viral loads in tissues from rWA1-Omi-S-inoculated cats were slightly higher than viral titers detected on day 3 pi (Fig. 5G and H). Together, these results indicate reduced replication of the rWA1-Omi-S in the respiratory tract and associated lymphoid tissues when compared to the rWA1-D614G recombinant virus in cats.

rWA1-Omi-S presents reduced replication and induces milder pathological changes in the respiratory tract of cats

The tissue distribution of SARS-CoV-2 in the respiratory tract and associated lymphoid tissues was assessed by in situ hybridization (ISH) and IFA staining. For this, nasal turbinate and lung from all animals were examined by ISH using the RNAscope ZZ technology and by IFA using a SARS-CoV-2 N-specific antibody. While intense labeling for viral RNA and staining for the viral N protein were consistently observed in nasal turbinate from rWA1-D614G-inoculated cats, only modest and sporadic staining was detected in tissues from rWA1-Omi-S-inoculated animals on days 3 and 5 pi (Fig. 6A and B). When we compared the intensity and distribution of both viral RNA and N protein in nasal turbinate from the two inoculated groups, the lower infectivity of the rWA1-Omi-S virus was evident and characterized by markedly lower levels of viral RNA and N detection across all animals tested (Fig. 6A and B).

Fig 6.

rWA1-Omi-S presents reduced replication in the respiratory tract of cats. IFA, ISH, and hematoxylin and eosin (HE) staining in nasal turbinate (A–C) and lung (D–F) of cats inoculated with rWA1-D614G and rWA1-Omi-S and control cats. Nasal turbinate and lungs were collected from rWA1-D614G- (A and D) and rWA-1-Omi-S- (B and E) or mock-inoculated (C and F) cats on day 3 pi. The IFA was performed using a monoclonal antibody against N protein of SARS-CoV-2 (red labeling). In ISH, viral RNA (red labeling) was detected using a probe targeting the SARS-CoV-2 S RNA. Note that both IFA and ISH showed intense labeling on tissues from rWA1-D614G-inoculated cats and less abundant labeling on tissues from rWA1-Omi-S-inoculated cats. Representative images of nasal turbinate and lungs collected from three inoculated cats with rWA1-D614G (A and D) or rWA1-Omi-S (B and E) and one mock-inoculated cat are shown.

Epithelial cells of the nasal mucosa were the predominant cell type positive for SARS-CoV-2 in the rWA1-D614G-inoculated animals, with extensive virus labeling observed on day 3 pi. Additionally, middle and basilar areas of the epithelium were also sporadically stained (Fig. 6A). In the lung, sparse staining of bronchial cells was observed on days 3 and 5 pi (Fig. 6D). Staining in the lung was more frequent in the interstitial regions especially in cells of the bronchiolar glands. In rWA1-Omi-S-inoculated cats, viral RNA and N protein staining in both the nasal turbinate and lung were consistently lower (number of cells and intensity of staining) than that observed in tissues from rWA1-D614G-inoculated animals (Fig. 6A, B, D, and E). No viral RNA or N protein staining was observed in tissues from control animals (Fig. 6C and F). These results demonstrate that the rWA1-Omi-S virus presents limited tissue distribution, replication, and spread in the respiratory tract of cats when compared to rWA1-D614G .

Histological examination was performed on tissues collected on days 3, 5, and 14 pi. The rWA1-D614G-inoculated cats presented with aggregates of fibrin, cellular debris, and leukocytes partially filling the nasal passage on days 3 (Fig. 6A) and 5 pi. Marked sloughing of nasal epithelial cells and mild to moderate epithelial necrosis were also noticed. In the lungs, bronchiolar necrosis with accumulation of exudates in the lumen containing necrotic epithelial cells and mononuclear inflammatory cells was observed on days 3 and 5 pi. Mononuclear infiltration was also noticed in the submucosa of bronchi and bronchiole on days 3 (Fig. 6D) and 5 pi. Thickening of the alveolar septa was consistently noticed at all time points examined with abundant mononuclear cellular infiltration on day 3 pi (Fig. 6D).

In general, histological changes observed in rWA1-Omi-S-inoculated cats were milder than those observed in rWA1-D614G-infeted animals. Deciliation of the nasal epithelial cells and infiltration of mononuclear cells were observed in the submucosa of the nasal passage from rWA1-Omi-S inoculated cats (Fig. 6B). Small accumulation of fibrin and cellular aggregates in the nasal passage was also noticed on days 3 and 5 pi. In the lungs, mild exudation was noticed in bronchiolar lumen on days 3 and 5 pi. Thickening of alveolar septa with mononuclear infiltration was noticed on days 3 (Fig. 6D) and 5 pi. All control cats presented normal histology in the nasal passage (Fig. 6C) and lungs (Fig. 6F). Together, these results demonstrate that the rWA1-Omi-S presents a limited pathogenicity in the infected cats compared to rWA1-D614G.

Antibody responses following rWA1-D614G and rWA1-Omi-S infection in cats

The antibody responses to SARS-CoV-2 were assessed by luciferase immunoprecipitation system (LIPS) and virus neutralization (VN) assays. We used the LIPS assay to determine the kinetics of antibody responses against the N protein in the sera of the rWA1-D614G- and rWA1-Omi-S-inoculated cats. Serum samples from control and rWA1-D614G- and rWA1-Omi-S-inoculated cats were collected on days 0, 7, and 14 pi and tested by LIPS assay (n = 3 cats per group) (28). None of the control or inoculated cats reacted strongly to the NLuc-tagged N protein on days 0 and 7 pi (Fig. 7A). At 14 dpi, serum from both rWA1-D614G- and rWA1-Omi-S-inoculated animals showed strong binding to the NLuc-tagged N protein antigen confirming presence of N protein-specific antibodies, with rWA1-D614G-infected cats presenting higher antibody levels as evidenced by higher luminescence than the rWA1-Omi-S-inoculated cats (Fig. 7A).

Fig 7.

Antibody responses following rWA1-D614G and rWA1-Omi-S infection in cats. The kinetics of antibody responses against N protein measured by LIPS assay (A). Neutralizing antibody (NAb) responses to SARS-CoV-2 and cross-reactivity between rWA1-D614G and rWA1-Omi-S recombinants were assessed by VN (cuttoff = 100% neutralization) (B and C). Data indicate mean ± SEM of three animals per group per time point. Two-way analysis of variance with multiple comparison test, *P < 0.05; **P < 0.01; ***P < 0.005; ****P < 0.0001.

Neutralizing antibody (NAb) responses to SARS-CoV-2 and cross-reactivity between rWA1-D614G and rWA1-Omi-S were evaluated. All rWA1-D614G-inoculated cats presented NAb titers as early as day 7 pi, while NAbs in rWA1-Omi-S-inoculated cats were only detected on day 14 pi. Neutralizing responses to the homologous virus were significantly higher than responses to the heterologous virus indicating low cross-neutralizing responses (P < 0.001) (Fig. 7B and C). NAb titers in rWA1-D614G serum ranged from 64 to 128 against rWA1-D614G virus, whereas no NAbs titers were detected against rWA1-Omi-S (titer lower than 8). rWA1-Omi-S infection induced lower and delayed NAbs, which were efficient in neutralizing the virus (Fig. 7C). All control animals remained seronegative throughout the experiment. These results confirm seroconversion of all rWA1-D614G- and rWA1-Omi-S-inoculated cats, demonstrating that all inoculated animals got infected. Moreover, results showed limited cross-neutralizing activity between rWA1-D614G and rWA1-Omi-S serum.

DISCUSSION

Here, we demonstrated that the S protein plays an important role in SARS-CoV-2 virulence and is a major determinant of the attenuated phenotype of SARS-CoV-2 Omicron BA.1 variant. In cell culture systems, the Omicron BA.1 S seems to mediate efficient entry into host cells; however, the rWA1-Omi-S presented reduced fusogenicity and ability to spread from cell-to-cell when compared to the B.1 D614G S bearing rWA1-D614G. Using a relevant naturally susceptible feline model of infection, we demonstrated that the rWA1-Omi-S induced subclinical infection, reduced replication in the upper respiratory tract, and decreased virus shedding and pathology, when compared to the ancestral rWA1-D614G virus. These findings obtained in a species that is naturally susceptible to SARS-CoV-2 and in the absence of pre-existing immunity, a confounding factor that precludes definitive conclusions about SARS-CoV-2 pathogenicity in humans, demonstrate that the S protein is a major determinant of SARS-CoV-2 virulence and pathogenesis.

The S protein of the Omicron BA.1 variant possesses 36 mutations compared to the ancestral D614G lineage. Several of the affected residues modulate receptor binding, protease cleavage efficiency, and cell-to-cell fusion, and contribute to virus replication and escape from NAbs. The S protein binds to the hACE2 receptor and mediates entry into susceptible host cells. Studies showed that the S protein of Omicron variant utilizes the hACE2 receptor with enhanced efficiency when compared to ancestral SARS-CoV-2 strains, a feature that has been speculated to contribute to rapid transmission of the virus (19, 29–31). Omicron BA.1 S protein presents an enhanced RBD-ACE2 binding interface through the interactions of amino acids N501Y, Q498R, and T478K with the host receptor in target cells. Notably, these mutations are reminiscent features of other variants, including Alpha (N501Y) and Delta (L452R and T478K) (31). Additional mutations such as Q493R and S477N also contribute to enhanced S-ACE2 interactions (19, 32). Notably, results of the infectivity assays in the present study using infectious recombinant viruses (rWA1-D614G and rWA1-Omi-S) demonstrated a more efficient entry of the Omicron BA.1 S harboring virus when compared to the ancestral D614G S virus. As the infection progressed, however, the rWA1-D614G virus, owing to its enhanced fusogenic activity and ability to spread from cell-to-cell, surpassed the rWA1-Omi-S-infection, resulting in higher spread and replication efficiency of the virus.

Recent studies demonstrated that the Omicron variant BA.1 is less dependent on transmembrane serine protease 2, less efficient in S cleavage, less fusogenic, and favors the endosomal pathway for virus entry (21, 30, 33). In line with this, results here with the Omicron S protein and the infectivity assays with the rWA1-Omi-S showed reduced fusogenicity and syncytia formation in hACE2_BHK21 and cACE2_BHK21 cells as compared to the D614G S and the rWA1-D614G virus, indicating that the S protein of Omicron contributes to this phenotype. The lower fusogenicity of the Omicron BA.1 S could also affect the ability of the virus to spread from cell-to-cell. This was confirmed by the smaller viral plaques observed for the rWA1-Omi-S compared to the rWA1-D614G. S mutations H655Y and T547K have been shown to contribute to the low fusogenicity of Omicron variants, while the H655Y mutation also dictates the enhanced endosome entry pathway utilization of Omicron (21, 34). SARS-CoV-2 S antagonizes viral RNA recognition by RIG-I blocking the induction of IFN and downstream IFN-stimulated genes (27). Subsequent studies showed that the S protein interacts with the signal transducer and activator of transcription 1 to block its association with Janus kinase 1 (35). Considering about 36 amino acid mutations in the S protein between Omicron BA.1 and B.1 D614G variants, we examined if these S proteins modulate innate immune inhibitory pathways differently. Differential modulation of innate responses and, more specifically, of IFN pathways could, for example, account for the reduced plaque size and lower ability of the rWA1-Omi-S virus to spread from cell-to-cell. Results from an IFN-β-driven luciferase reporter assay, however, revealed comparable inhibition of the IFN-β reporter activity by B.1 D614G and Omicron BA.1 S proteins. Together, these results rule out impaired IFN-β inhibition as a factor contributing to the lower replication ability and cell-to-cell spread observed for the rWA1-Omi-S virus. These observations further point toward the decreased cleavability and fusogenic activity of the Omicron BA.1 S as a major driving force contributing to the lower overall infectivity and ability of the virus to spread from cell-to-cell.

The Omicron BA.1 variant is more immune evasive and less virulent than the other major identified variants (10, 12–15, 19, 24, 25, 29, 30, 32, 33, 36–38). In Syrian hamsters, the SARS-CoV-2 B.1 D614G produced significant illness and higher viral loads in secretions and tissues, while the Omicron BA.1-inoculated animals were subclinical and had reduced virus replication and shedding (39). Other studies in hamsters and humanized K18 transgenic hACE2 mice also showed lower levels of lung infection and replication and decreased clinical disease and pathology in Omicron-inoculated animals when compared to historical isolates or other SARS-CoV-2 variants (24, 33, 36). Similarly, using domestic cats, a naturally susceptible animal species, we recently showed that a wild-type Omicron BA.1.1 causes lower pathogenicity than wild-type ancestral B.1 and Delta variants (26). Here, by using recombinant chimeric viruses (rWA1-D614G and rWA1-Omi-S), we sought to determine and dissect the role and contribution of the S protein for the attenuated disease phenotype of Omicron BA.1. While rWA1-D614G-inoculated cats were febrile, lethargic, and lost or maintained steady body weight, rWA1-Omi-S-inoculated cats remained subclinical and gained weight throughout the experiment. Similarly, the rWA1-Omi-S presented significantly lower replication in the respiratory tract and associated lymphoid tissues of inoculated animals, as evidenced by lower virus shedding and milder viral replication and pathological changes in target respiratory organs. These results confirm a major role for the Omicron BA.1 S for the attenuation of the virus in the feline model of infection. However, the role and contribution of Omicron-specific mutations that fall outside of S protein for virus infection and disease pathogenesis cannot be formally ruled out. Recent studies using K18 transgenic hACE2 mice showed that S contributes to the attenuated phenotype of Omicron BA.1, and mutations outside S protein are also likely to contribute to the virulence of the virus in this mouse model (18, 40). In these studies, intranasal inoculation of K18 hACE2 mice with Omicron BA.1 caused no significant body weight loss, while the rWA1 bearing Omicron S caused less severe disease than the ancestral virus but was not as attenuated as the natural Omicron (18). Survival assessments revealed non-fatal infection in Omicron-infected mice compared to the 100% and 80% mortality in ancestral D614G- and Omi-S-infected mice (18). Similarly, Barut and colleagues observed Delta S-mediated increased replication and pathogenicity with Omicron-BA.1 S-mediating reduced replication and pathogenicity in K18 hACE2 mice, suggesting that the S contributes to SARS-CoV-2 replication and pathogenicity (25). Another study in K18 transgenic hACE2 mice also showed that mutations outside of the S protein contribute to the attenuation of the Omicron variant (40). While the ancestral WA1 virus bearing the Omicron S protein was lethal, the Omicron virus bearing WA1 S protein was attenuated in K18 transgenic hACE2 mice (40). More recently, another study showed that while Omicron BA.1 and BA.2 cause minimal to negligible morbidity and mortality in K18 hACE2 mice, animals infected with Omicron BA.5 exhibited high weight loss and mortality that was dependent on the age of the mice (41).

Mutations in the S protein of the Omicron variants, particularly in the RBD, lead to widespread escape from NAb responses causing vaccine breakthrough infection and increased transmission (10, 11, 13–15, 30). As a result, Omicron showed lower neutralizing sensitivity to antibodies elicited by earlier variants or by immunization (10, 11, 15, 30). NAbs were detected in serum of rWA1-D614G-infected cats as early as day 7 pi, while NAbs were detected only on day 14 pi for the rWA1-Omi-S. Cross-neutralization assays showed no cross-neutralizing activity between rWA1-D614G and rWA1-Omi-S serum, confirming that the S protein contributes to the neutralizing activity and the Omicron S extensively escapes neutralization by antibodies elicited against ancestral WA1-D614G S. Our previous study also showed no or little cross-reactivity in serum neutralization between SARS-CoV-2 B.1- and Omicron BA.1.1-infected cat serum (26).

In summary, our study demonstrates the critical role and contribution of the Omicron BA.1 S protein for virus infectivity, cell-to-cell spread, and replication in vitro and in vivo. Our in vitro studies show that Omicron BA.1 S presents a lower fusogenicity ability when compared to the ancestral D614G S, which leads to decreased cell-to-cell spread of the virus. These results were confirmed in vivo in a feline model of infection. Importantly, while other non-S mutations may also contribute to the lower pathogenicity of Omicron BA.1 variant, our results demonstrate that the S is a major player contributing to virulence and pathogenesis of SARS-CoV-2.

MATERIALS AND METHODS

Cells and viruses

Vero E6 (ATCC CRL-1586) and Vero E6 expressing human TMPRSS2 (JCRB Cell Bank, JCRB1819) were cultured in Dulbecco’s modified eagle medium (DMEM), supplemented with 10% fetal bovine serum (FBS), L-glutamine (2 mM), penicillin (100 U.mL⁻¹), streptomycin (100 µg.mL⁻¹), and gentamicin (50 µg.mL⁻¹). BHK21 and HEK293T cells were maintained in complete growth media consisting of Minimum Essential Media (Corning, 10-010-CV), supplemented with 10% FBS, penicillin (100 U.mL⁻¹) and streptomycin (100 µg.mL⁻¹). The cell cultures were maintained at 37°C with 5% CO₂. SeV (Cantell strain) was propagated in 11-day-old embryonated chicken eggs and titrated using hemagglutination assay.

The bacterial artificial chromosome (BAC) harboring the whole genome of the ancestral SARS-CoV-2 WA1-D614G (rWA1-D614G) and S gene of Omicron BA.1 variant in the backbone of the ancestral rWA1-D614G (rWA1-Omi-S) was generated as previously described (42). The S gene from Omicron BA.1 strain was amplified from SARS-CoV-2 isolate hCoV19/USA/MD-HP25001/2022 (GISAID accession no. EPI_ISL_9245416). Recombinant viruses were rescued under biosafety level 3 (BSL-3) laboratory conditions following strict biosecurity procedures using Vero E6 TMPRSS2. Briefly, Vero E6 TMPRSS2 cells were seeded in six-well plates (3 × 10⁵ per well) and transfected with 2 µg/well of the respective SARS-CoV-2 BAC DNA [pBAC-WA1-D614G or pBAC-WA1-Omi-S (BA.1 S)] using Lipofectamine 3000. After 24 h, culture medium was replaced with complete growth media containing 5% FBS and incubated for an additional 48 h. The culture supernatant was collected and labeled as P0. The recombinant viruses rWA1-D614G and rWA1-Omi-S were passaged twice in Vero E6 TMPRSS2 cells. The whole genome sequences of the virus stocks were determined using a tiled amplicon approach and the Oxford Nanopore MK1C sequencing platform as previously described by us (43) to confirm that no mutations occurred during rescue and amplification in cell culture and the sequences were deposited in GenBank (accession no. OR626632-OR626633) and raw data files were deposited under Bioproject ID (PRJNA1066270). The titers of virus stocks were determined by plaque assays, calculated according to the Spearman and Karber method, and expressed as plaque-forming units per milliliter (PFU.mL⁻¹). Sequenced verified viral stocks were used in all experiments described in this study.

Generation of ACE2 stably expressing cell lines

Human and cat ACE2 expressing lentiviral plasmids (pscALPS-hACE2 and pscALPS-cACE2) were obtained from Addgene (158081 and 158082, respectively). Each gene was tagged with a c-terminal Myc-tag epitope to facilitate detection. Lentiviral particles encoding hACE2 and cACE2 were generated in HEK293T cells. For this, 7.2 × 10⁵ cells were seeded in each well of a six-well plate and allowed to adhere overnight. Each well was then transfected with the lentiviral packaging vectors psPAX2 (1 µg, Addgene, 12260) and pMD2.G (1 µg, Addgene, 12259), plus the pscALPS-hACE2 or -cACE2 (1 µg) or the empty pscALPS_puro (Addgene, 128504) vector as a negative control. Lentiviral particles were harvested from the supernatant of transfected cells at 72 h post-transfection, cleared by centrifugation, aliquoted, and stored at −80°C until use.

Target BHK21 cells in exponential growth phase were seeded 7.2 × 10⁵ cells in a six-well plate and allowed to adhere overnight. Prior to transduction, aliquots of the lentivirus were brought to room temperature (RT) and gently inverted to mix. Media was removed from the BHK21 cells and 1 mL of each lentivirus was added per well and adsorbed for 2 h at 37°C with plate rocking every 30 min. After adsorption, 2 mL of complete growth media was added, and cells were incubated for 72 h. At 72 h, media was replaced with complete growth media containing 5 µg.mL⁻¹ of Puromycin Dihydrochloride for selection (Gibco, A1113803). Following complete death of non-transduced cells, expression of hACE and cACE2 was validated via immunofluorescence and immunoblots through detection of the c-terminal Myc-tag with a 1:500 dilution of antibody Myc-Tag-9B11 Mouse mAb (Cell Signaling Technology, 2276).

SARS-CoV-2 infectivity assay

To assess infectivity of rWA1-D614G and rWA1-Omi-S viruses, hACE2_BHK21, cACE2_BHK21, Vero E6, and Vero E6 TMPRSS2 cells were seeded into eight-well glass chamber slides (Millipore, PEZGS0816). At 24 h, when cells were approximately 80% confluent, they were inoculated with rWA1-D614G and rWA1-Omi-S at a multiplicity of infection (MOI) of 1 and incubated at 4°C for 1 h. Following incubation, the inoculum was removed, 1 mL of complete growth media was added to each well, and plates were incubated at 37°C with 5% CO₂. At 4, 8, 12, and 24 h post-inoculation, cells were fixed with 3.7% formaldehyde solution for 30 min and washed with phosphate-buffered saline (PBS). To monitor SARS-CoV-2 entry, infectivity and spread cells were stained with a SARS-CoV-2 anti-N rabbit polyclonal antibody for 1 h followed by 594-conjugated secondary Goat anti-Rabbit IgG (VWR, GtxRb-003-D594NHSX) for 1 h. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (ThermoFisher Scientific, 62248). The percentage of infected cells at each time point was quantified using ImageJ software.

Fusogenicity assay

To assess and compare the fusogenicity of D614G and Omicron BA.1 S proteins, hACE2_BHK21, cACE2_BHK21, or control BHK21 cells were seeded into eight-well glass chamber slides (Millipore, PEZGS0816) to a final confluency of 80%. Initially, cells were transfected with 250 ng of plasmids encoding the SARS-CoV-2 S genes delivered by pTwist-SARS-CoV-2 Δ18 D614G (Addgene, 164437) or pTwist-SARS-CoV-2 Δ18 B.1.1.529 (Addgene, 179907), incubated overnight at 37°C with 5% CO₂, and fixed with 3.7% formaldehyde solution for 20 min. To assess the S fusogenicity during infection, hACE2_BHK21, cACE2_BHK21, or control BHK21 cells were seeded in eight-well glass chamber slides (Millipore, PEZGS0816) to a final confluency of 80% and inoculated with rWA1-D614G or rWA1-Omi-S viruses (MOI = 0.1). Viruses were adsorbed to cells for 1 h at 4°C, after which the inoculum was removed and 1 mL complete growth media was added to the cells. Cells were incubated for 24 h at 37°C with 5% CO₂ and fixed with 3.7% formaldehyde solution for 30 min. After fixation, transfected and infected cells were stained with an anti-N rabbit polyclonal antibody followed by incubation with secondary 594-conjugated Goat anti-Rabbit IgG (VWR, GtxRb-003-D594NHSX). Nuclei were stained with DAPI (ThermoFisher Scientific, 62248). For imaging, PBS was removed from all wells, the outer well walls were removed, and the glass slides were mounted with mounting media (Ibidi, 50–305-707) and coverslips. Representative images were taken across three replicates at 10×, and individual nuclei were counted using ImageJ. The fusogenicity of each S protein was quantitated for each of the three replicate images using the formula: percentage of nuclei incorporated into syncytia = (# of syncytia enclosed nuclei) / (total # of nuclei in field).

Plaque size determination

The ability of rWA1-D614G and rWA1-Omi-S to spread from cell-to-cell was evaluated using plaque assays. For this, Vero E6, hACE2_BHK21, and cACE2_BHK21 cells were seeded in six-well plates (3 × 10⁵ cells per well) and at 24 h were infected with rWA1-D614G and rWA1-Omi-S (10 PFU per well). After 1 h incubation at 37°C, the inoculum was removed and 2 mL of media containing 2× complete growth media and 1% SeaKem agarose (final conc. 1× media 0.5% agarose) was added to each well. After polymerization of agarose, the plate was transferred to the incubator 37°C 5% CO₂ for 72 h. The agarose overlay was removed, and cells were fixed with 3.7% formaldehyde solution for 30 min and stained with 0.5% crystal violet solution for 10 min at room temperature. The images and plaque size were determined using a Keyence BZ-X810 Microscope.

Viral growth kinetics

The replication kinetics of rWA1-D614G and rWA1-Omi-S were examined in Vero E6 (120,000 cells per well), hACE2_BHK21, and cACE2_BHK21 (360,000 cells per well) cells seeded in 12-well plates and incubated at 37°C for 24 h until they reached 80%–90% confluency. Cells were inoculated with rWA1-D614G and rWA1-Omi-S with an MOI of 0.1 (500 µL per well) and incubated at 4°C for 1 h. Following incubation, the inoculum was removed, and 1 mL of complete growth media was added per well and cells were incubated at 37°C with 5% CO₂. Cells and supernatant were harvested at 4, 8, 12, 24, and 48 h post-inoculation and stored at −80°C. Time point 0 h was an aliquot of virus inoculum stored −80°C as soon as inoculation was completed. Virus titers were determined in Vero E6 TMPRRSS2 cells on each time point using endpoint dilutions and the Spearman and Karber’s method and expressed as TCID₅₀.mL^–1.

Luciferase reporter assay

The ability of the D614G and Omicron BA.1 S to modulate innate immune pathway activation was investigated using luciferase assays. For this, HEK293T cells were seeded in 24-well plate at 2 × 10⁵ cells per well and transfected with pIFN-β-Luc or pNF-κB-Luc (200 ng/well) and pRN-Luc (50 ng) reporter plasmids in combination with S protein expressing plasmids pTwist-SARS-CoV-2 Δ18 D614G (Addgene, 164437) or pTwist-SARS-CoV-2 Δ18 B.1.1.529 (Addgene, 179907) or a pTwist empty vector (250 ng) using Lipofectamine 3000 (Invitrogen, L3000001). At 24 h post-transfection, cells were stimulated with SeV, Cantell strain (100 hemagglutination units per well), or TNF-α (25 ng/well) for 12 h. After stimulation, cells were lysed and luciferase activity was measured using the dual luciferase assay (Promega, E2940). Luminescence was measured using a luminometer plate reader (BioTek Synergy LX Multimode Reader). The ratio of luminescence obtained from the target reporters (IFN-β-Luc or NF-κB-Luc) to luminescence from the control Renilla reporter (pRN-Luc) was calculated to normalize the transfection efficiency. Then, the relative IFN-β- or NF-κB-driven luciferase activity was calculated as fold change over the unstimulated cells. Confirmation of the S expression in HEK293T-transfected cells was performed by Western blot using antibody against SARS-CoV-2 S2 protein (Sino Biological).

Animals housing and experimental design

A total of 27 20- to 65-month-old domestic cats (Felis catus) [five females and four males (n = 9) per group] were obtained from Clinvet (Waverly, NY, USA). Animals were donated to Cornell University to support the reduction of animal use in research. All animals were housed in the animal biosafety level 3 facility at the East Campus Research Facility at Cornell University. After acclimation, and prior to virus inoculation, nasal swabs and serum were collected from all animals and tested by rRT-PCR for SARS-CoV-2 and serum neutralization using SARS-CoV-2 B.1 and Omicron lineages which confirmed that the animals were negative for SARS-CoV-2 (26). On day 0, cats were anesthetized and inoculated intranasally with 1 mL (0.5 mL per nostril) of a virus suspension containing 5 × 10⁵ PFU of rWA1-D614G or rWA1-Omi-S. Control cats were mock-inoculated with Vero E6 cell culture medium supernatant. All animals were maintained individually in Horsfall high efficiency particulate air (HEPA)-filtered cages. Body temperatures and weight were measured daily until day 7, and then on days 10 and 14 pi. OPS, NS, and RS were collected under sedation (dexmedetomidine) on days 0, 1, 3, 5, 7, 10, and 14 post-inoculation. Upon collection, swabs were placed in sterile tubes containing 1 mL of viral transport medium (Corning, Glendale, AZ, USA) and stored at −80°C until processed for further analyses. Blood was collected through jugular venipuncture using a 3 mL sterile syringe and 21G × 1” needle and transferred into serum separator tubes on day 4 before inoculation, and on days 0, 7, and 14 pi. The blood tubes were centrifuged at 1,200 × g for 10 min and serum was aliquoted and stored at −20°C until further analysis. Three cats from each group were humanely euthanized on days 3, 5, and 14 pi. Following necropsy, BALF, and tissues, including nasal turbinate, palate/tonsil, retropharyngeal LN, trachea, lung, mediastinal LN, heart, liver, spleen, kidney, small intestine, and mesenteric LN, were collected and processed for rRT-PCR and virus titration. Additionally, nasal turbinate, trachea, and lung were collected and processed for standard microscopic examination and were also processed by ISH and in situ IFA. For this, tissue sections of approximately 0.5 cm in width were fixed by immersion in 10% neutral buffered formalin (20 volumes fixative to 1 vol tissue) for approximately 72 h, and then transferred to 80% ethanol, followed by standard paraffin embedding techniques. Slides for standard microscopic examination were stained with hematoxylin and eosin (HE).

Nucleic acid isolation and rRT-PCR

Nucleic acid was extracted from NS, OPS, RS, BALF, and tissue samples collected at necropsy. For NS, OPS, RS, and BALF samples, 200 µL of cleared swab supernatant was used for nucleic acid extraction. For tissues, 0.3 g of each tissue were minced with a sterile disposable scalpel, resuspended in 3 mL DMEM (10% wt/vol), and homogenized using a stomacher (one speed cycle of 60 s, Stomacher 80 Biomaster). Then, the tissue homogenate supernatant was centrifuged at 2,000 × g for 10 min and 200 µL of cleared supernatant was used for RNA extraction using the MagMax Core extraction kit (Thermo Fisher, Waltham, MA, USA) and the automated KingFisher Flex nucleic acid extractor (Thermo Fisher, Waltham, MA, USA) following the manufacturer’s recommendations. The rRT-PCR for total viral RNA detection was performed using the EZ-SARS-CoV-2 Real-Time RT-PCR assay (Tetracore Inc., Rockville, MD, USA), which detects both genomic and subgenomic viral RNA targeting the virus N gene. An internal inhibition control was included in all reactions. Positive and negative amplification controls were run side-by-side with test samples. For SARS-CoV-2 subgenomic RNA detection, a RT-qPCR reaction targeting the virus envelope (E) gene was used following the primers and protocols previously described (44). Both RT-PCR (for total viral RNA detection) and RT-qPCR assay (for specific subgenomic RNA detection) were verified using a standard curve by using 10-fold serial dilutions from 10⁰ to 10⁻⁸ of virus suspension containing 10⁶ TCID₅₀.mL⁻¹ for each of the SARS-CoV-2 variants used in the study. Relative viral genome copy numbers were calculated based on the standard curve and determined using GraphPad Prispm 9 (GraphPad, La Jolla, CA, USA). The amount of viral RNA detected in samples was expressed as log (genome copy number) per milliliter.

Virus isolation and titrations

Samples that tested positive for SARS-CoV-2 by rRT-PCR were subjected to virus isolation under BSL-3 conditions at the Animal Health Diagnostic Center Research Suite at Cornell University. NS, OPS, RS, BALF, and tissues homogenate supernatant were subjected to endpoint titrations. For this, samples were subjected to limiting dilutions and inoculated into Vero E6 TMPRSS2 cells cultures prepared 24 h in advance in 96-well plates. At 48 h pi, cells were fixed and subjected to IFA as described in a previous study (26). The limit of detection for infectious virus titration is 10^1.05 TCID₅₀.mL^–1. Virus titers were determined on each time point using endpoint dilutions and the Spearman and Karber’s method and expressed as TCID₅₀.mL^–1.

In situ RNA detection

Paraffin-embedded tissues from days 3 and 5 pi were sectioned at 5 µm and subjected to ISH using the RNAscope ZZ probe technology (Advanced Cell Diagnostics, Newark, CA). Nasal turbinate, trachea, and lung from inoculated and controls were subjected to ISH using the RNAscope 2.5 HD Reagents–RED kit (Advanced Cell Diagnostics) following the manufacturer’s instructions and using a probe targeting SARS-CoV-2 S RNA (V-nCoV2019-S probe ref # 848561). A probe targeting feline host protein peptidylprolyl isomerase B was used as a positive control (Advanced Cell Diagnostics cat # 455011). A probe targeting DapB gene from Bacillus subtilis strain SMY was used as a negative control (Advanced Cell Diagnostics cat # 310043).

In situ IFA

Paraffin-embedded tissues from days 3 and 5 pi were sectioned at 5 µm and subjected to IFA. Tissues from inoculated and control cats including nasal turbinate, palate/tonsil, retropharyngeal LN, trachea, lung, and heart were subjected to IFA. Formalin-fixed paraffin-embedded tissues were deparaffinized with xylene and rehydrated through a series of graded alcohol solutions. Antigen unmasking was performed using Tris-based antigen unmasking solution pH 9.0 (Vector Laboratories ref # H-3301) by boiling the slides in the unmasking solution for 20 min. After 10 min at 0.2% Triton X-100 (in phosphate-buffered saline) at RT, and 30-min blocking using a goat normal serum (1% in PBS) at RT, tissues were subjected to IFA. A mouse monoclonal antibody (mAb) targeting SARS-CoV-2 N protein was used as a primary antibody (SARS-CoV-2 N mAb clone B61G11) (26) and incubated for 45 min at RT, followed by a 30-min incubation at RT with a goat anti-mouse IgG antibody (goat anti-mouse IgG, Alexa Fluor 594). Nuclear counterstain was performed with DAPI (10 min at RT).

Histology

For the histological examination, tissue sections of approximately 0.5 cm in width were fixed by immersion in 10% neutral buffered formalin (≥20 volumes fixative to 1 vol tissue) for approximately 72 h, and then transferred to 80% ethanol, followed by standard paraffin embedding techniques. Nasal turbinate and lung tissues collected from inoculated and control cats were processed and stained with hematoxylin and eosin (H&E). The stained tissues were subjected to histological examination by a trained veterinary pathologist.

LIPS

The N protein of SARS-CoV-2 B.1 D614G variant (NYI67-20 strain) was cloned into Nano-Luc vector pNLF1-N [CMV/Hygro] (N terminus) (Cat. N1351, Promega) as a fusion protein. Briefly, the N protein of SARS-CoV-2 was amplified by PCR and cloned in-frame to the 5´end of the NLuc gene between the restriction enzymes SacII and NotI. A FLAG tag was added to 5´ end of N for the confirmation of expression. The resulting plasmid was sequenced to verify the authenticity of the inserted gene and the expression of the N protein in HEK293T cells was confirmed by IFA and Western blot assays. The NLuc-tagged antigen was generated in HEK293T cells transfected with 10 µg of pNLF1-N plasmid containing SARS-CoV-2 N protein as described previously (28). At 48 h post-transfection, cells were lysed with radioimmunoprecipitation assay (RIPA) lysis and extraction buffer (ThermoFisher Scientific, Rockford, IL, USA), supplemented with 1× protease inhibitor (Pierce Protease Inhibitor Tablet, EDTA-Free, ThermoFisher Scientific, Rockford, IL, USA) and the light units (LU) per microliter of lysates were measured using Nano-Glo Luciferase Assay System (Promega) in a luminometer (BioTek Synergy LX Multimode Reader). The LIPS assay was performed following protocols described in a previous study (28) with some modifications. Briefly, 50 µL NLuc-antigen (10⁷ LU per well) and 50 µL serum (heat inactivated and diluted 1:50) were mixed in a 96-well plate and incubated for 1 h on a rotary shaker at room temperature. Then, the NLuc-antigen and serum mixture were incubated with 10 µL of a 30% suspension of Ultralink protein A/G beads (Pierce Biotechnology, Rockford, IL) in PBS in a 96-well plate. After 1-h incubation, the antigen-serum-bead mixture was transferred into a 96-well filter HTS plate (Millipore, Bedford, MA). The plate was washed eight times with buffer A followed by twice with PBS on a vacuum manifold as described previously (28). After that, 50 µL of Nano-Glo Luciferase reagent (Nano-Glo Luciferase Assay System, Promega) was added to each well, incubated for 3 min at RT in dark. The plate was then read in a luminometer (BioTek Synergy LX Multimode Reader) and the luminescence (LU) corresponded to the quantity of N protein-specific antibody present in the test serum was calculated. Pooled serum from non-inoculated control cats were used as negative control.

Neutralizing antibodies

Neutralizing antibody responses to SARS-CoV-2 were assessed by VN assay performed under BSL-3 laboratory conditions. Serum samples collected on days 0, 7, and 14 pi were tested against both the recombinant viruses rWA1-D614G and rWA1-Omi-S. For the VN assay, twofold serial dilutions (1:8 to 1:1,024) of serum samples were incubated with 100–200 TCID₅₀ of rWA1-D614G or rWA1-Omi-S for 1 h at 37°C. Following incubation of serum and virus, 50 µL of a cell suspension of Vero E6 cells was added to each well of a 96-well plate and incubated for 48 h at 37°C with 5% CO₂. At 48 h pi, cells were fixed and subjected to IFA as described in a previous study (26). Neutralizing antibody titers were expressed as the reciprocal of the highest dilution of serum that completely inhibited SARS-CoV-2 infection/replication. FBS and positive and negative serum samples from white-tailed deer (45) were used as controls.

Statistical analysis and data plotting

Statistical analysis was performed by two-way analysis of variance followed by multiple comparisons. Statistical analysis and data plotting were performed using the GraphPad Prism software (version 9.0.1). Figures 1A and 4A and the graphic abstract were created with BioRender.com.

ETHICS APPROVAL

The protocols and procedures for generation of the recombinant viruses were reviewed and approved by the Cornell University Institutional Biosafety Committee (MUA 16373-2). Handling of all recombinant SARS-CoV-2 was performed under BSL-3 conditions in the BSL3 and ABSL3 facilities at College of Veterinary Medicine, Cornell University, following strict biosecurity procedures and following standard guidelines. All animals were handled in accordance with the Animal Welfare Act. All animals were maintained individually in Horsfall HEPA-filtered cages. The study procedures were reviewed and approved by the Institutional Animal Care and Use Committee at Cornell University (IACUC approval number 2020-0064).