Recombinant rotavirus expressing the glycosylated S1 protein of SARS-CoV-2

Reverse genetic systems have been used to introduce heterologous sequences into the rotavirus segmented double-stranded (ds)RNA genome, enabling the generation of recombinant viruses that express foreign proteins and possibly serve as vaccine vectors. Notably, insertion of SARS-CoV-2 sequences into the segment seven (NSP3) RNA of simian SA11 rotavirus was previously shown to result in the production of recombinant viruses that efficiently expressed the N-terminal domain (NTD) and the receptor-binding domain (RBD) of the S1 region of the SARS-CoV-2 spike protein. However, efforts to generate a similar recombinant (r) SA11 virus that efficiently expressed full-length S1 were less successful. In this study, we describe modifications to the S1-coding cassette inserted in the segment seven RNA that allowed recovery of second-generation rSA11 viruses that efficiently expressed the ~120-kDa S1 protein. The ~120-kDa S1 products were shown to be glycosylated, based on treatment with endoglycosidase H, which reduced the protein to a size of ~80 kDa. Co-pulldown assays demonstrated that the ~120-kDa S1 proteins had affinity for the human ACE2 receptor. Although all the second-generation rSA11 viruses expressed glycosylated S1 with affinity for the ACE receptor, only the S1 product of one virus (rSA11/S1f) was appropriately recognized by anti-S1 antibodies, suggesting the rSA11/S1f virus expressed an authentic form of S1. Compared to the other second-generation rSA11 viruses, the design of the rSA11/S1f was unique, encoding an S1 product that did not include an N-terminal FLAG tag. Probably due to the impact of FLAG tags upstream of the S1 signal peptides, the S1 products of the other viruses (rSA11/3fS1 and rSA11/3fS1-His) may have undergone defective glycosylation, impeding antibody binding. In summary, these results indicate that recombinant rotaviruses can serve as expression vectors of foreign glycosylated proteins, raising the possibility of generating rotavirus-based vaccines that can induce protective immune responses against enteric and mucosal viruses with glycosylated capsid components, including SARS-CoV-2.


INTRODUCTION
Rotaviruses are a large genetically diverse group of segmented double-stranded (ds) RNA viruses that belong to the family Sedoreoviridae [1,2].The viruses are important human pathogens, representing the primary cause of acute potentially life-threatening gastroenteritis in children under 5 years of age [3,4].The extensive morbidity and mortality associated with rotavirus infections have stimulated the development of rotavirus vaccines and their introduction into the childhood immunization programmes of many countries, including the US [5,6].The most widely used rotavirus vaccines are formulated from live attenuated strains of the virus and are given orally to infants during the first 6 months of life [7].
Several studies have described the isolation and characterization of rotavirus variants that, due to sequence duplications, contain one or more unusually large genome segments [25][26][27][28][29][30].The most commonly described variants are those with sequence duplications present within segment five [30], which encodes the viral interferon antagonist NSP1 [31,32], or segment seven, which encodes the viral translation enhancer NSP3 [33][34][35].Variants have been identified with sequence duplications in segment five of almost 1.2 Kbp, attesting to the capacity of the rotavirus genome and virus particle to accommodate significant amounts of extra sequence [30].
Recombinant rotaviruses have been made by reverse genetics that, through the introduction of heterologous sequences into segments 5, 7, and 11 RNAs, are able to express foreign proteins [36,37].Because NSP1 is not essential for rotavirus replication, some recombinant rotaviruses expressing foreign proteins have been made by replacing portions of the NSP1 ORF with heterologous coding sequences [38][39][40].Other recombinant viruses have been made by placing the heterologous coding sequence at the end of the full-length NSP1 ORF (G.Yi, J. Patton, unpublished results).Recombinant rotaviruses that express foreign proteins from segment seven have been generated though placement of heterologous coding sequences at the end of the NSP3 ORF [41][42][43].Such segment seven modifications have been used to prepare recombinant viruses that express NSP3 fusion proteins, for example, NSP3 fused to the green epifluorescence protein, UnaG (NSP3-UnaG) [42].By insertion of a 2A stop-restart translation element in segment seven, between the coding sequence for NSP3 and a foreign protein, recombinant rotaviruses have been generated that express two separate protein products from segment seven [41,43,44].This approach has allowed the development of rotaviruses that express the complete complement of rotavirus proteins plus an additional non-rotaviral protein [45][46][47].Recombinant rotaviruses that express immunogenic regions of the capsid proteins of other viruses may allow the generation of combination vaccines that can induce protective immune responses to not only rotavirus, but also a second pathogenic virus, such as norovirus [46].
In an earlier study, we investigated the possibility of making recombinant rotaviruses that expressed portions of the spike (S) protein of SARS-CoV-2 [47].In this work, recombinant SA11 (rSA11) rotaviruses with segment seven modifications were recovered that expressed the N-terminal domain (NTD), the receptor binding domain (RBD), and the core domain (CR) of the SARS-CoV-2 S protein [47,48].A similar segment seven modification was used to make a recombinant virus (rSA11/fS1) containing the complete coding sequence of the SARS-CoV-2 S1 protein.The S1 protein is a cleavage fragment of the S protein that includes both the NTD and RBD [48] and is a primary target of neutralizing antibodies produced during SARS-CoV-2 infection [49][50][51][52][53][54].The open reading frame (ORF) in the modified segment seven RNA of the rSA11/fS1 virus included the coding cassette NSP3-2A-3xFLAG-S1 [47].Through the action of the 2A translation element, the segment seven RNA of the virus was expected to generate two products: NSP3 fused to a 2A peptide (NSP3-2A) and 3xFLAG-tagged S1 (fS1).In the NSP3-2A-3xFLAG-S1 cassette, a 3xFLAG tag was positioned immediately upstream of the S1 signal peptide, an element critical for synthesis of glycosylated S products [55].Immunoblot analysis of the products made by rSA11/fS1 indicated that although the virus efficiently made NSP3-2A, it was not efficient in generating the expected fS1 product, possibly due to instability or degradation of the S1 product, or impact of the FLAG tag on the function of the signal peptide [47].In the work described here, we have compared the S1 products made by the rSA11/fS1 virus to the products made by newly designed rSA11 viruses encoding S1 proteins differing in the nature of their terminal peptide tags.The results showed that one of newly designed rSA11 viruses (rSA11/S1f) efficiently expressed the S1 protein and that the S1 protein was glycosylated, localized to ER/Golgi vesicles, and had affinity for the extracellular domain of the ACE2 receptor [56].This is the first study demonstrating that recombinant rotaviruses can be used as expression vectors of glycosylated foreign proteins.

Recombinant viruses
Detailed procedures for generating and recovering recombinant SA11 rotaviruses have been published before [42,57].Briefly, BHK-T7 cells were transfected with SA11 pT7 plasmids and pCMV-NP868R using Mirus TransIT-LT1 transfection reagent.pT7/NSP2SA11 and pT7/NSP5SA11 were included in transfection mixtures at levels three-fold higher than the other plasmids.As necessary, the pT7/NSP3SA11 plasmid was replaced with pT7/NSP3-2A-3fS1, pT7/NSP3-2A-S1f or pT7/NSP3-2A-3fS1-His.The transfected BHK-T7 cells were overseeded with MA104 cells at 2 days post-infection, and the growth medium was adjusted to a final concentration of 0.5 µg ml −1 trypsin (porcine Type IX pancreatic trypsin, Sigma Aldrich).Three days after overseeding, cells in the media overlay were subject to three rounds of freeze-thaw and lysates clarified by low-speed centrifugation.Virus in lysates were amplified on MA104 cells, recovered by plaque isolation, and then amplified again by growth on MA104 cells [57,58].Viral dsRNAs were recovered by TRIzol (Thermo Fisher) extraction, resolved by polyacrylamide gel electrophoresis, and detected by staining with ethidium bromide [57,58].cDNAs were generated from dsRNAs using a Superscript III One-Step RT-PCR Platinum Taq kit (Thermo Fisher) and appropriate segment seven (NSP3) primers and sequenced by Eurofins Genomics.

Endoglycosidase H (Endo H) assay
MA104 cell monolayers in 6-well plates were mock-infected or infected with rSA11 viruses (5 PFU per cell).At 9 h p.i., cell monolayers were washed and scraped into phosphate-buffered saline (PBS), pelleted by low-speed centrifugation, and resuspended in 250 µl per well of IP lysis buffer.The presence of glycosylated proteins in the cell lysates was assessed using Promega Endoglycosidase H (Endo H) assay reagents (Promega V4871).Briefly, 27 µl samples of cell lysates were combined with 3 µl of 10X Denaturing Solution, heated to 95 °C for 5 min, and cooled to room temperature.The heat-treated lysates were mixed with 3 µl of nuclease-free water, 4 µl of 10X Endo H Reaction Buffer and 3 µl of Endo H enzyme, then incubated at 37 °C for 16 h.Proteins in the processed samples were detected by immunoblot assay, as described above.

S1-ACE2 interaction assay
A Takara Capturem IP and Co-IP kit (Cat No: 635721) was used to assess the affinity of SARS-CoV-2 S1 expressed by rSA11 viruses for ACE2.Protein A spin columns and all necessary buffers were included in the Capturem IP kit.MA104 cell monolayers were mock-infected or infected with rSA11 viruses (5 PFU/cell).At 9 h p.i., the cells were washed and scraped into PBS, pelleted by low-speed centrifugation, and resuspended in Lysis/Equilibration Buffer containing protease inhibitor cocktail.After a 15 min incubation on ice, the lysate was clarified by centrifugation at 17 000 g for 10 min.Soluble hACE2-Fc (fchace2, InvivoGen), a recombinant protein consisting of the extracellular domain of human ACE2 fused to a human IgG1 Fc region, was added to the clarified lysates, to a final concentration of 20 µg per ml, and the mixture incubated overnight at 4 °C.To recover complexes formed between the hACE2-Fc and S1 proteins, lysate samples were loaded onto pre-equilibrated protein A spin columns, which were then centrifuged at 1000 g for 1 min at room temperature.After rinsing columns with Wash Buffer, proteins were eluted from columns by adding Elution Buffer and centrifugation at 1000 g for 1 min at room temperature.The eluted samples were immediately neutralized by adding Neutralization Buffer.Proteins in eluted samples were detected by immunoblot assay, as described above.

Immunofluorescence analysis
MA104 cells were grown on poly-l-lysine-coated coverslips in 12-well culture dishes and mock-infected or infected with 5 PFU of trypsin-activated rSA11 virus per cell.Prior to infection, virus inoculum was trypsin-activated by incubating with 10 µg porcine trypsin, type IX-S (Sigma-Aldrich), per millilitre for 30 min at 37 °C.At 9 h p.i., the cells were fixed by incubation with ice cold methanol for 3 min, then washed twice with PBS containing 1 % Triton X-100 (PBS-TX).The cells were incubated in PBS containing 5 % bovine serum albumin (BSA) for 30 min at room temperature and then incubated with rabbit S1/ABclonal antibody, and mouse monoclonal NSP2 (lot #171, 1 : 1500) or NSP4 antibody (lot #55/4, 1 : 800) in PBS containing 3 % BSA for 1 h at room temperature.Afterwards, the cells were washed thrice with PBS-TX, followed by incubation with Alexa 488 anti-rabbit IgG (green) and Alexa 594 anti-mouse IgG (red) (Molecular Probes) in PBS containing 3 % BSA for 30 min at room temperature.The cells were washed three times with PBS-TX, and the coverslips were mounted with ProLong antifade reagent containing 4,6-diamino-2-phenylindole (DAPI) (Invitrogen).Cells were analysed with a Nikon Eclipse NiE microscope (100× oil immersion objective) and images were captured with a Hamamatsu Orca-Flash 2.8 sCMOS high resolution camera using FITC and TRITC windows.

Genetic stability
The genetic stability of recombinant rotaviruses was assessed by serial passage on MA104-cell monolayers using 1 : 10 dilutions of infected cell lysates prepared in serum-free DMEM containing 0.5 µg ml −1 trypsin [45,58].Viral dsRNA was recovered by Trizol extraction from clarified cell lysates treated with RNase T1 to remove single-stranded RNA [42,57].Viral dsRNA was analysed by electrophoresis on 8 % polyacrylamide gels and detected by straining with ethidium bromide.

Generation of recombinant viruses encoding S1 protein
In a previous study, a rSA11 virus (rSA11/3fS1) was generated with a modified segment seven (NSP3) RNA that encoded the SARS-CoV-2 S1 protein with a fused N-terminal 3xFLAG tag (NSP3-2A-3fS1) (Fig. 1).Notably, the 3xFLAG tag was positioned immediately upstream of the N-terminal signal peptide of the S1 protein.To gain a better understanding of factors affecting the poor expression of S1 by rSA11/3fS1, we generated two similar rSA11 viruses differing only in the nature of the peptide tags encoded upstream and downstream of the S1 ORF in the segment seven RNA.One of the viruses, rSA11/3fS1-His (Fig. 1), was identical to rSA11/NSP3-2A-3fS1, with the exception that the ORF in its segment seven RNA was engineered to introduce a 6xHis tag at the end of the S1 product.The rSA11/3fS1-His virus was generated to address the possibility that, due to cleavage of the signal peptide from the S1 product, the N-terminal 3xFLAG tag was lost, preventing accurate assessment of fS1 synthesis by the rSA11/NSP3-2A-3fS1 virus via immunoblot assay with anti-FLAG antibody [47].Instead, the production of S1 products could be assessed with anti-His antibody.The second recombinant virus that was made, rSA11/S1f (Fig. 1), contained a segment seven RNA designed to express S1 with a C-terminal 1xFLAG tag, but without any N-terminal tag (NSP3-2A-S1f).The usefulness of this virus was in examining the possibility that a tag positioned upstream of the S1 signal peptide might impede synthesis and glycosylation of the S1 protein on the endoplasmic reticulum.
The recombinant viruses, rSA11/S1f and rSA11/3fS1-His, were produced following the same reverse genetics procedure used previously to generate rSA11/3fS1 and the wild-type virus, rSA11/wt [42,47].The procedure included transfection of BHK-T7 cells with a set of T7 transcription vectors (pT7) expressing SA11 plus-sense (+)RNAs and a CMV expression plasmid (pCMV-NP868R) encoding the capping enzyme of African swine fever virus.In the transfection mixtures, T7 transcription vectors for NSP2 and NSP5 +RNAs (pT7/SA11NSP2 and pT7/SA11NSP5, respectively) were used at levels three-fold greater that the other pT7 vectors.The modified segment seven transcription vectors (pT7/3fS1, pT73fS1-His, and pT7/S1f, Fig. 1) used in generating rSA11 encoding S1 products was added to transfection mixtures in place of pT7/NSP3SA11.Recombinant viruses formed in transfected BHK-T7 cells were amplified by overseeding with MA104 cells and then isolated by plaque purification.Recombinant viruses were further amplified in MA104 cells.

Genomes and growth characteristics of rSA11s
The dsRNA genome segments of recombinant rotaviruses were resolved by gel electrophoresis to verify the presence of modified segment seven RNAs (Fig. 2a).The analysis showed that rSA11/3fS1, rSA11/S1f and rSA11/3fS1-His all lacked the 1.1-Kbp segment seven dsRNA typical of rSA11/wt.Instead, the S1-encoding rSA11 viruses contained segment seven dsRNAs that migrated on polyacrylamide gels near the position of the segment one dsRNA and had a size close to 3.3 Kbp.Thus, the segment seven RNAs of the mutant viruses were approximately three times the size of SA11/wt segment seven RNA.Sequencing verified that the segment seven sequence of the recombinant viruses matched those present in the pT7/NSP3-2A-3fS1, pT7/NSP3-2A-3fS1-His, and pT7/NSP3-2A-S1f plasmids (Fig. 1).The total size of genome segments of the recombinant viruses, rSA11/3fS1, rSA11/S1f and rSA11/3fS1-His, is 20.7-20.8Kbp, which is 2.1-2.2Kbp (or ~11 %) greater than rSA11/wt, pointing to the ability of the rotavirus genome and virion particle to accommodate large amounts of foreign sequence.
Analysis of the plaque phenotypes of rSA11/3fS1, rSA11/S1f and rSA11/3fS1-His showed that these viruses generated plaques on MA104 cell monolayers that were smaller than those produced by rSA11/wt (Fig. 2b).Similarly, recombinant viruses encoding the S1 protein grew to titres that were one-to two-logs lower than rSA11/wt (Fig. 2c).These findings are consistent with earlier reports which showed that the introduction of foreign sequences into segment seven generates recombinant viruses with plaque phenotypes that are smaller in diameter than wild-type virus and that grow to maximum titres that are less than wild-type virus [43,45,47].The poorer growth characteristics of recombinant viruses with larger segment seven RNAs may stem from alterations in the quantity of the NSP3 protein produced in infected cells or in the translational efficiency of the NSP3 (+)RNA.

rSA11 viruses expressing the S1 protein of SARS-CoV-2
To evaluate the nature of the S1 products expressed by recombinant viruses, lysates were prepared at 9 h p.i. from MA104 cells infected with rSA11/wt, rSA11/3fS1, rSA11/S1f and rSA11/3fS1-His.S1 products in the lysates were detected by immunoblot assay using anti-FLAG, anti-6xHis, and anti-S1 antibodies; rotavirus NSP3 and VP6 were detected using protein specific antisera.Analysis with anti-FLAG and anti-His antibodies (Fig. 2d) indicated that a ~120 kD protein was expressed by rSA11/3fS1, rSA11/ S1f and rSA11/3fS1-His but not by rSA11/wt.The 120-kD size corresponds to the glycosylated form of the S1 protein, suggesting that the S1 signal peptide in the recombinant viruses is functional (the size of nonglycosylated S1 is 80kD).Production of the 120-kD protein by the modified segment seven RNAs indicates that their 2A elements were functional, enabling the expression of S1 as a separate protein product.Indeed, recombinant viruses with the modified segment seven RNAs also produced an NSP3 product that was slightly larger in size than wild-type NSP3 (Fig. 2d).The larger NSP3 results from the presence of 2A residues left behind at the C-terminus of NSP3 following 2A activity, suggesting that the 2A element was active in infected cells producing S1.Notably, immunoblot analysis indicated lysates from rSA11/NSP3-2A-3fS1-infected cells contained not only a ~120 kD S1 product but also a small product of ~40 kD (Fig. 2d).The origin of the ~40 kD product is unknown but may result from an aberrant cleavage of the 120-kD S1 protein.Interestingly, the ~40 kD product was not detected in lysates from cells infected with viruses expressing S1 with either a C-terminal FLAG or 6xHis tag, suggesting that modification of the C-terminus may affect the protein's folding or modification such that the S1 protein becomes less prone to cleavage.Nonetheless, the instability of the S1 product may explain the poor detection of S1 expression reported earlier for the rSA11/3fS1 virus [47].Detection of the 120-kDa S1 proteins expressed by rSA11/S1f and rSA11/3fS1-His with antibodies recognizing tags present at the C-termini of the S1 proteins confirmed that the proteins represented full length products.
To further analyse the S1 products made by the rSA11/3fS1, rSA11/S1f and rSA11/3fS1-His viruses, we obtained a rabbit polyclonal anti-S1 antibody (ABclonal A20136) generated against an S1 fusion protein (amino acids 11-682) (Fig. 2d).Unexpectedly, immunoblot analysis showed that the S1/ABclonal antibody only recognized the 120 kDa product expressed by rSA11/S1f and not the 120 kDa products expressed by rSA11/3fS1 and rSA11/3fS1-His.The 120 kDa product of rSA11/S1f initiates with the S1 signal peptide, while the 120 kDa products of rSA11/3fS1 and rSA11/3fS1-His initiate with a 3xFLAG peptide, which is followed by the S1 signal peptide.It may be that the S1/ABclonal antibody recognizes an epitope that is altered by the presence of an N-terminal tag and, as a result, the antibody cannot recognize the 120 kDa products of rSA11/3fS1 and rSA11/3fS1-His.Perhaps, for example, the N-terminal FLAG-tags of rSA11/3fS1 and rSA11/3fS1-His may affect glycosylation or other modifications (e.g.palmatoylation, SUMOylation) of the S1 products in a manner that prevents recognition by the S1/ABclonal antibody.
To gain further insight into the immunoblot results obtained with the S1/ABclonal antibody, we probed blots prepared from rSA11/3fS1, rSA11/S1f and rSA11/3fS1-His infected cell lysates with a different S1 antibody.Unlike the rabbit polyclonal S1/ ABclonal antibody, this second antibody (S1/R and D) represents a mouse monoclonal antibody generated against S1 recombinant protein.The immunoblot analysis showed that the S1/R and D antibody strongly recognized the S1 product of the rSA11/S1f virus but was much less efficient in recognizing the S1 products of either the rSA11/3fS1 or rSA11/3fS1-His virus (Fig. 2e).These results are similar to those obtained with the S1/ABclone antibody, which recognized the S1 product of rSA11/S1f but failed to recognize the S1 products of the rSA11/3fS1 and rSA11/3fS1-His viruses.These results support the hypothesis that the presence of FLAG tags in front of the S1 signal may impact S1 glycosylation or other post-translation modification, impeding recognition by some anti-S1 antibodies.

Effect of endoglycosidase H (Endo H) treatment on expressed S1 protein
To further explore the possibility that the S1 products expressed by rSA11/3fS1, rSA11/S1f and rSA11/3fS1-His were glycosylated, we examined the effect that Endo H had on the electrophoretic migration patterns of the S1 products.To do this, we infected MA104 cells with the rSA11/3fS1, rSA11/S1f and rSA11/3fS1-His viruses, prepared lysates from the infected cells at 9 h p.i., and then treated portions of the lysates with Endo H using a Promega Endoglycosidase kit.The treated portions and untreated controls were resolved by gel electrophoresis, and immunoblot assay was used to determine the positions of the S1 products.Assays performed using the anti-FLAG and anti-6xHis antibodies showed that Endo H treatment significantly reduced the migration pattern of the S1 products of all three viruses from ~120 kDa to 80-90 kDa (Fig. 3).These results indicate that the S1 products are heavily glycosylated and contain Endo H-sensitive asparagine (N)-linked oligosaccharide chains.Immunoblot assays performed with the S1/ABclonal antibody likewise showed that Endo H caused a reduction in the size of the S1f product of the rSA11/S1f virus from ~120 kDa to 80-90 kDa.Notably, the S1/ABclonal antibody did not detect the Endo H-treated products of the rSA11/3fS1 and rSA11/3fS1-His viruses described above (Fig. 3).Endo H treatment truncates N-linked oligosaccharide chains on proteins, leaving behind N-acetylglucosamine attached to asparagine.It may be possible that the presence of the residual N-acetylglucosamine residues interfere with the ability of the S1 antibody to recognize epitopes located in the S1 products of the rSA11/3fS1 and rSA11/3fS1-His viruses, even though the S1 products have been treated with Endo H.

Binding of expressed S1 protein to the ACE2 receptor
Attachment of SARS-CoV-2 to target cells is mediated by the interaction of the RBD of the S protein with the human ACE2 (hACE2) transmembrane protein.To determine whether the S1 proteins expressed by rSA11 viruses were able to bind hACE2, we infected MA104 cells with rSA11/3fS1, rSA11/S1f and rSA11/3fS1-His.At 9 h p.i., clarified lysates prepared from the cells were incubated with soluble hACE2-Fc, a recombinant protein that contains the ACE2 extracellular domain fused to an IgG1 Fc tag.After incubation, protein A spin columns were used to recover the hACE2-Fc protein from the samples.After washing the spin Fig. 3. Endo H treatment of SARS CoV-2 S1 proteins expressed by rSA11 viruses.Whole cell lysates prepared at 9 h p.i. from MA104 cells infected with the indicated rSA11 viruses were mock-treated or treated with Endo H.The samples were examined by immunoblot assay using FLAG/His antibody to detect S1 proteins.The same blot was stripped and re-probed with antibody specific for SARS CoV-2 S1(αS1/ABclonal).Immunoblots were also probed with antibodies specific for rotavirus NSP3 and VP6, and for β-actin.columns, bound proteins were eluted and examined by immunoblot assay.As shown in Fig. 4a, analysis of the eluted proteins by immunoblot analysis with anti-FLAG and anti-6xHis antibodies indicated that the S1 proteins expressed by rSA11/3fS1, rSA11/ S1f and rSA11/3fS1-His viruses all bound hACE2-Fc.This result suggests that the RBD domains of the S1 proteins generated by the rSA11 viruses are functional.Similar immunoblot analysis performed with S1/ABclonal antibody also indicated that the S1 product of rSA11/S1f virus had affinity for hACE2-Fc and, thus, contained a functional RBD (Fig. 4a).Because the S1/ABclonal antibody does not recognize the S1 products of the rSA11/NSP3-2A-3fS1 and rSA11/NSP3-2A-3fS1-His viruses, it was not possible to use the antibody in immunoblot assays to assess interaction of their S1 products with hACE2-Fc.Analysis of the flow-through fractions collected from the protein A spin columns indicated that they contained similar amounts of cellular β-actin and that the infected cells expressed similar amounts of the VP6 inner capsid protein (Fig. 4b).Analysis of the flow through fractions also indicated that incubation of hACE2-Fc with the rSA11/S1f-infected cell lysate caused partial depletion of the S1f product (Fig. 4b), consistent with the idea that the S1f product had affinity for the RBD of the spike protein.

Intracellular localization of expressed S1 protein
The results above indicated that only the S1f product of the rSA11/S1f virus contained a functional epitope for the S1/ABclonal antibody.To gain insight into the distribution of the S1f product in infected cells, MA104 cells were infected with rSA11/wt and rSA11/S1f viruses and then, at 9 h p.i., examined by immunofluorescence assay using the S1/ABclonal antibody, an antibody recognizing a major component of rotavirus replication factories (viroplasms, anti-NSP2), and an antibody recognizing a rotavirus ER/Golgi resident transmembrane protein (anti-NSP4).As expected, immunofluorescence signal was not detected in mock infected cells analysed with the S1/ABclonal, NSP2, or NSP4 antibodies (Fig. 5).Likewise, signal was not detected with the S1/ ABclonal antibody in rSA11/wt-infected cells.Analysis of rSA11/S1f infected cells with the S1/ABclonal antibody indicated that a large portion of the S1f product accumulated in large cytoplasmic punctate structures (Fig. 5).The S1f fluorescence did not colocalize with viroplasms detected by the NSP2 antibody, indicating the S1f product does not accumulate in rotavirus replication factories.In contrast, the S1f fluorescence co-colocalized extensively with the viral NSP4 ER/Golgi transmembrane protein (Fig. 5).This finding is consistent with results provided above that the S1f product is a N-mannose linked glycoprotein.Although the S1f product localizes to the same site of the infected cell where the outer capsid of the rotavirus particle is assembled (ER/ Golgi) [2,8], this does prevent the formation of infectious rotavirus (Fig. 2c).However, we cannot rule out the possibility the co-accumulation of S1f in the ER/Golgi may be partially responsible for the reduction of lower virus titres formed in rSA11/ S1f-infected cells as compared to rSA11/wt-infected cells (Fig. 2c).

Genetic stability of rSA11 viruses expressing S1f
Successful development of rotavirus as vaccine vector system requires the recombinant virus to be sufficiently genetically stable to allow scale up and production of large amounts of vaccine virus expressing a foreign sequence.To evaluate the genetic stability of rSA11/S1f, the virus was subjected to five rounds of serial passage at low multiplicity of infection in MA104 cells.In this procedure, infected cell lysates were diluted 1 : 10 to generate the inoculum used subsequent rounds of infection.Gel electrophoresis showed no difference in the viral dsRNA recovered from passage 1-to 5-infected cell lysates (Fig. 6), suggesting that the rSA11/S1f virus was genetically stable, at least though five rounds of serial passage.Immunoblot analysis showed that rSA11/S1f virus serially passaged up to five times in MA104 cells continued to direct the expression of the S1f product, consistent with the idea that this virus strain is genetically stable despite containing 2.1 Kbp of extra sequence.Interestingly, similar recombinant viruses (rSA11/ fVP1 and rSA11/VP1-h) carrying 1.7 Kbp of norovirus capsid sequence was genetically unstable, with its modified segment seven fully lost from the virus pool upon five rounds of serial passage at 1 : 10 dilution [45].This finding suggests that it is not just the length of foreign sequence inserted into the rotavirus genome that influences genetic stability but also the nature of the foreign sequence (e.g.nucleotide composition, secondary structure) or the encoded protein.

DISCUSSION
In this study, we examined three rSA11 strains engineered to express the SARS-CoV-2 S1 through modification of the segment seven (NSP3) dsRNA.Although the segment seven RNAs of all three strains encoded S1 proteins with N-terminal signal peptides, two included sequence information that introduced a 3xFLAG tag immediately upstream of the signal peptide.The presence of the 3xFLAG was correlated with the inability of the S1 products of the rSA11/3fS1 and rSA11/3fS1-His viruses to be efficiently recognized by two widely used anti-S1 antibodies (Fig. 2d).On the other hand, the S1f product of the rSA11/S1f virus, which lacks a peptide tag at its N-terminus, was readily detected by the anti-S1 antibodies.The apparent size of the S1 product expressed by all three virus strains (rSA11/S1f, rSA11/3fS1, and rSA11/3fS1-His) was ~120 kDa, which is consistent with glycosylated S1 generated during SARS-CoV-2 infection.Endo H treatment of the 120-kDa S1 product of the rSA11/S1f, rSA11/3fS1, and rSA11/3fS1-His viruses resulted in a size reduction to 80-90 kDa, providing additional evidence that the S1 products had undergone N-linked glycosylation (Fig. 3).Although all three S1 products were glycosylated, there were differences between the products that extended beyond the fact that only the S1f form could be recognized by the anti-S1 antibodies.For example, a significant portion of the 3fS1 product of the rSA11/3fS1 virus was cleaved in infected cells, yielding a major 40 kDa glycosylated S1 fragment (Fig. 2d).This cleavage was not seen for the S1f and 3fS1-His products of the rSA11/S1f and rSA11/3fS1-His viruses, suggesting that the 3fS1 has an aberrant protein-fold that makes it prone to cleavage by an unidentified host protease.Collectively, our data indicates that the presence of a 3xFLAG tag upstream of the S1 signal peptide may impact the S1 product at multiple levels, including its glycosylation, fold structure and/or display of antibody epitopes.Our data also suggest that the S1f product of the rSA11/S1f virus has associated properties expected for authentic S1, suggesting that it may be possible to use rotaviruses as vaccine vectors, in developing a combined rotavirus-SARS-CoV-2 vaccine for immunization of infants and young children.To our knowledge, this is the first study demonstrating that rotaviruses can be used as vector systems for the expression of glycosylated proteins.In previous work, we showed that by modifying the segment seven dsRNA, it may be possible to develop rotaviruses as vaccine vectors for the expression of nonglycosylated capsid proteins of other enteric and mucosal viruses, such as norovirus [45], hepatitis A, C, and E viruses, polioviruses, adenoviruses.and enteroviruses.The results here show that it may be possible to extend the use of rotaviruses to include their development as vaccine vectors for the expression of the glycosylated capsid proteins of other enteric and mucosal viruses (HIV, coronavirus, Zika virus, and Ebola virus).The widespread use of rotavirus vaccines throughout the world makes rotavirus vectors prime candidates for the development of combination vaccines in infants and young children.
Our findings confirm previous observations that the rotavirus genome can accommodate a remarkable amount of foreign sequence [43,45,47].Based on work with the rSA11/S1 virus strains ( [47], and herein), the rotavirus genome can accommodate at least 2.1 Kbp of additional RNA sequence.A surprising finding from this study was that the genetic stability of rSA11/S1f viruses was remarkably greater than what has been observed for other rSA11 strains, in which smaller amounts of foreign sequence had been introduced into the rotavirus genome [45].Discovering what factors influence genetic stability will be important for fully developing rotaviruses as vaccine vectors systems.

Fig. 1 .
Fig. 1.Modified segment seven (NSP3) plasmids used to generate rSA11s encoding the SARS-CoV-2 S1 protein.Schematic indicates the nucleotide positions of the coding sequences for NSP3, porcine teschovirus 2A element, 3X or 1X FLAG (FL), 6X His (His) and the complete S1 in pT7 plasmids.The red arrow notes the position of the 2A translational stop-restart site, and the asterisk notes the end of the ORF.Sizes (aa) of encoded NSP3 and S1 proteins are given.Virus (rSA11) recovered using the pT7 plasmid is indicated in parenthesis.T7 (T7 RNA polymerase promoter sequence), Rz (Hepatitis D virus ribozyme), UTR (untranslated region).

Fig. 2 .
Fig. 2. Properties of rSA11 viruses expressing SARS CoV-2 S1 proteins.(a) Viral dsRNA was recovered from MA104 cells infected with the indicated rSA11 isolates, resolved by gel electrophoresis, and detected by ethidium-bromide staining.RNA segments of rSA11/wt are labelled 1 to 11. Locations of segment seven dsRNAs are identified with black arrows.(b) Plaque assays were performed using MA104 cells and detected by crystal-violet staining.Titres were determined from two separate plaque assays and data was analysed by unpaired Student's t-test using Prism version 10. **, P<0.05.Bars reflect 95 % confidence levels.(c) Titres reached by rSA11 isolates upon complete CPE in MA104 cells were determined by plaque assay.(d) Lysates were prepared from cells infected with the indicated rSA11 viruses at 9 h p.i and examined by immunoblot assay using FLAG/His antibodies to detect S proteins.The same blot was stripped and re-probed with a rabbit polyclonal antibody specific for SARS CoV-2 S1 (αS1/ABclonal), rotavirus NSP3 and VP6, and β-actin.(e) Same lysates as in (d) were examined by immunoblot assay using mouse monoclonal anti-S1 antibody (αS1/R and D) and antibodies against rotavirus NSP3, VP6, and β-actin.

Fig. 4 .
Fig.4.Binding of the SARS-CoV-2 S1 product to the human ACE2 receptor.(a) Whole cell lysates (WCL) were prepared at 9 h p.i. from MA104 cells infected with the indicated rSA11 viruses.The lysates were incubated with hACE2-Fc, a recombinant protein consisting of the extracellular domain of human ACE2 fused to an IgG1 Fc tag.Protein-A spin columns were used to recover hACE2-Fc, and associated proteins, from the lysates.The recovered samples were examined by immunoblot assay using antibodies specific for S1 products (FLAG/His antibody, SARS CoV-2 S1 antibody (ABclonal A20136).The red arrow head identifies S1 protein recovered by the pulldown assay.MWM, molecular weight markers.(b) The residual S1 proteins in the flow through fractions of the protein-A spin columns was determined by immunoblot assay using FLAG/His antibody.The same blot was re-probed with antibodies specific for SARS CoV-2 S1 (αS1/ABclonal), rotavirus VP6, and cellular β-actin.

Fig. 5 .
Fig.5.Localization of SARS CoV-2 S1 protein in rSA11/S1f-infected cells.MA104 cells were mock infected or infected with rSA11/S1f and, at 9 h p.i., fixed with ice cold methanol.Afterwards, the cells were incubated with rabbit S1/ABclonal antibody, mouse NSP2 or NSP4 antibody, followed by Alexa 488 anti-rabbit IgG (green) and Alexa 594 anti-mouse IgG (red) to determine the locations of SARS-CoV-2 S1 and rotavirus NSP2 and NSP4 proteins.Nuclei were detected by staining with DAPI.Cells were analysed with a Nikon Eclipse NiE microscope (100× oil immersion objective) and images were captured with a Hamamatsu Orca-Flash 2.8 sCMOS high resolution camera.Images include 20 µm size bars.