key: cord-0778509-affr2ya7
authors: Diebold, Ola; Gonzalez, Victoria; Venditti, Luca; Sharp, Colin; Blake, Rosemary A.; Stevens, Joanne; Caddy, Sarah; Digard, Paul; Borodavka, Alexander; Gaunt, Eleanor
title: Engineering a Vaccine Platform using Rotavirus A to Express SARS-CoV-2 Spike Epitopes
date: 2022-03-27
journal: bioRxiv
DOI: 10.1101/2022.03.23.485570
sha: ced442f285561e539958920c6bdafefd77b14b37
doc_id: 778509
cord_uid: affr2ya7

Human rotavirus (RV) vaccines used worldwide have been developed using live attenuated platforms. The recent development of a reverse genetics system for RVs has delivered the possibility of engineering chimeric viruses expressing heterologous peptides from other virus species to generate polyvalent vaccines. We tested the feasibility of this using two approaches. Firstly, we inserted short SARS-CoV-2 spike peptides into the hypervariable region of the simian SA11 RV strain viral protein (VP) 4. Secondly, we fused the receptor binding domain (RBD) of the SARS-CoV-2 spike protein, or the shorter receptor binding motif (RBM) nested within the RBD, to the C-terminus of non-structural protein (NSP) 3 of the bovine RF strain RV, with or without an intervening T2A peptide. Mutating the hypervariable region of SA11 VP4 impeded viral replication, and for these mutants no cross-reactivity with spike antibodies was detected. To rescue NSP3 mutants, we established a plasmid-based reverse genetics system for the bovine RF strain. Except for the RBD mutant, all NSP3 mutants delivered endpoint titres and replication kinetics comparable to that of the WT virus. In ELISAs, cell lysates of an NSP3 mutant expressing the RBD peptide showed cross reactivity with a SARS-CoV-2 RBD antibody. 3D bovine gut enteroids were susceptible to infection by all NSP3 mutants but only RBM mutant showed cross reactivity with SARS-CoV-2 RBD antibody. The tolerability of large peptide insertions in the NSP3 segment highlights the potential for this approach in the development of vaccine vectors targeting multiple enteric pathogens simultaneously. IMPORTANCE We explored the use of rotaviruses (RVs) to express heterologous peptides, using SARS-CoV-2 as an exemplar. Small SARS-CoV-2 peptide insertion (<34 amino acids) into the hypervariable region of the viral protein 4 (VP4) of RV SA11 strain resulted in reduced viral titre and replication, thus limiting its use as a potential vaccine expression platform. To test RF strain for its tolerance for peptide insertions, we constructed a reverse genetics system. NSP3 was C-terminally tagged with SARS-CoV-2 spike peptides of up to 193 amino acids. With a T2A-separated 193 amino acid tag on NSP3, there was little effect on the viral rescue efficiency, titre and replication. Tagged NSP3 elicited cross-reactivity with SARS-CoV-2 spike antibodies in ELISA. This is the first report describing epitope tagging of VP4, and of a reverse genetics system for the RF strain. We highlight the potential for development of RV vaccine vectors targeting multiple enteric pathogens simultaneously.

possibility of engineering chimeric viruses expressing heterologous peptides from other virus 23 species to generate polyvalent vaccines. We tested the feasibility of this using two 24 approaches. Firstly, we inserted short SARS-CoV-2 spike peptides into the hypervariable 25 region of the simian SA11 RV strain viral protein (VP) 4. Secondly, we fused the receptor 26 binding domain (RBD) of the SARS-CoV-2 spike protein, or the shorter receptor binding 27 motif (RBM) nested within the RBD, to the C-terminus of non-structural protein (NSP) 3 of 28 the bovine RF strain RV, with or without an intervening T2A peptide. Mutating the 29 hypervariable region of SA11 VP4 impeded viral replication, and for these mutants no cross-30 reactivity with spike antibodies was detected. To rescue NSP3 mutants, we established a 31 plasmid-based reverse genetics system for the bovine RF strain. Except for the RBD mutant, 32 all NSP3 mutants delivered endpoint titres and replication kinetics comparable to that of the 33 WT virus. In ELISAs, cell lysates of an NSP3 mutant expressing the RBD peptide showed 34 cross reactivity with a SARS-CoV-2 RBD antibody. 3D bovine gut enteroids were susceptible 35 to infection by all NSP3 mutants but only RBM mutant showed cross reactivity with SARS-36

CoV-2 RBD antibody. The tolerability of large peptide insertions in the NSP3 segment 37 highlights the potential for this approach in the development of vaccine vectors targeting 38 multiple enteric pathogens simultaneously. 39 40 IMPORTANCE 41 We explored the use of rotaviruses (RVs) to express heterologous peptides, using SARS-42

CoV-2 as an exemplar. Small SARS-CoV-2 peptide insertion (<34 amino acids) into the 43 hypervariable region of the viral protein 4 (VP4) of RV SA11 strain resulted in reduced viral 44 titre and replication, thus limiting its use as a potential vaccine expression platform. To test 45 RF strain for its tolerance for peptide insertions, we constructed a reverse genetics system. 46 NSP3 was C-terminally tagged with SARS-CoV-2 spike peptides of up to 193 amino acids. 47 INTRODUCTION reassortants and RV reporter expression systems, and conceptualise novel vaccine 90 platforms [43] [44] [45] . Using the plasmid-only reverse genetics system, recombinant RVs 91 harbouring chimeric VP4 proteins that showed efficient replication in cell culture and 92 neutralising activity in mice have also been engineered [46] [47] [48] . Protection from infection with 93 RV is primarily mediated by heterotypic neutralising antibodies that target VP4 and/ or VP7. 94 VP4 is therefore highly immunogenic and an important target for adaptive immunity [49, 50] . 95

Thus, the potential for VP4 to express heterologous epitopes from different RV strains may 96 provide a delivery platform for expression of different vaccine antigens, though peptide 97 insertions into VP4 has not previously been tested. Additionally, the plasmid-only reverse 98 genetics system has been utilised to generate a repertoire of recombinant RVs expressing 99 fluorescent reporter proteins [48, 51] . The C-terminus of the SA11 NSP3 open reading frame 100 (ORF) was fused to a porcine teschovirus translational 2A element followed by various 101 reporters including UnaG, mKate, mRuby or TagBFP to successfully yield two uncoupled 102 proteins without compromising virus replication [51] . A more recent study showed that the C-103 terminus of SA11 NSP3 can express different peptides of the severe acute respiratory 104 syndrome coronavirus 2 (SARS-CoV-2) spike protein with minimal impact on endpoint titres 105

[52]. This has highlighted the potential to use RVs as expression vectors for development of 106 polyvalent vaccines for enteric viruses. encoding the SARS-CoV-2 spike protein in order to stimulate the production of neutralising 113 antibodies and T-cell mediated immune responses that target this protein [55] [56] [57] [58] . High 114 neutralising antibody titres are strongly associated with the receptor binding domain (RBD) 115 of spike, making it the most immunogenic antigen [59] [60] [61] [62] . To assess the potential for 116 generating chimeric vaccines using RV, we used SARS-CoV-2 as a timely exemplar to 117 introduce spike peptides into an RV backbone and determine whether chimeric viruses 118 showed cross-reactivity with spike antibodies. For this, the hypervariable region of SA11 119 VP4 (VP8* lectin domain) and the C-terminus of the bovine RF strain NSP3 were modified to 120 express SARS-CoV-2 spike epitopes. 121 122 This is the first report describing tagging of the surface protein VP4 of the RV with 123 heterologous peptides. We found that mutating the hypervariable region of SA11 VP4 124 reduced RV infectivity and mutants expressing spike peptides did not cross-react with 125 SARS-CoV-2 spike antibodies, suggesting that VP4 tagging is not a viable strategy for live 126 attenuated vaccine development. Using our established reverse genetics system for the 127 bovine RF strain RV, we rescued infectious viruses expressing either the RBD or the RBM of 128 the SARS-CoV-2 spike protein, with similar titres and replication kinetics to those of the wild 129 type (WT) virus. These viruses cross-reacted with RBD antibodies in ELISA, and were able 130 to infect bovine gut enteroids, inferring the potential of the system for use in live attenuated 131 vaccine development. 132 expressing SARS-CoV-2 spike peptides, we utilised two strategies involving both the simian 147 RV strain SA11 and the bovine RV strain RF. It was previously suggested that the surface 148 protein VP4 of the SA11 strain tolerates short immunogenic peptides inserted at specific 149 sites with minimal impact on viral replication or particle assembly [63] . Guided by the 150 available structure model of VP4 (PDB: 4V7Q), we have chosen an exposed loop outside 151 the sialic acid-binding domain located within the 'head' of the VP4 spike ( Fig. 1A) to 152 investigate the possibility of VP4 alteration. Moreover, several neutralisation escape mutants 153 were ascribed to the amino acid changes within this region [64], consistent with its 154 accessibility to antibodies. A number of B-cell linear epitopes derived from either the heptad 155 repeat 2 (HR2), N-terminal domain (NTD) or RBM regions of the SARS-CoV-2 spike protein 156

were selected for insertion [61, 65, 66] (Fig. 1A) . These epitopes were introduced into the Database available at VIPR (https://www.viprbrc.org/brc/). Only unstructured epitopes of up 162 to 15 residues were chosen ( Table 1 ). The SA11 strain was selected for this mutagenesis 163 due to the structural characterisation of VP4 of its close relative RRV, its user-friendly 164 reverse genetics system and its rapid growth kinetics. 165

As an alternative to VP4 tagging, we aimed to test whether tagging a RV strain more closely 166 related to the bovine virus backbone used in the pentavalent RotaTeq vaccine [69], could be 167 extrapolated to the bovine RF strain. Here we modified the C-terminus of the RF strain NSP3 168 ORF to express spike epitopes coding for the RBD or RBM of SARS-CoV-2, with or without 169 an intervening thosea asigna virus 2A (T2A) peptide (Fig. 1A) . The inclusion of the T2A was 170 employed in order to lower the risk of interfering with the function of the NSP3 gene. On the 171 other hand, increasing the segment size by incorporating the T2A peptide could further affect 172 the antigenic processing, hence both approaches were trialled. The panel of constructs was 173 assigned the notation RBM, T2A-RBM, RBD and T2A-RBD. 174 constructs were designed to encode each of the 11 RF gene segments, flanked at the 5' end 181 by a T7 promoter (T7P) and at the 3' end an antigenomic hepatitis delta virus (HDV) 182 ribozyme sequences, followed by the T7 terminator sequence (T7T) as in Kanai et al. [42] . 183

The constructs were synthesised by Invitrogen GeneArt on either pMK-RQ (kanamycin 184 resistance), pMA-RQ or pMA-T (ampicillin resistance) vectors. RF strain NSP3 constructs 185 RBM, T2A-RBM, RBD and T2A-RBD were ordered as gene blocks from Invitrogen GeneArt 186 and cloned into pT7-NSP2SA11 expression plasmid (Addgene #89169) after the NSP2 ORF 187 was removed using SmaI and SalI restriction enzymes. All plasmids were amplified by 188 transformation into chemically competent E. coli DH5α, except RF VP7 encoding plasmid for 189 which DH10β cells were used, and purified using QIAGEN ® Plasmid Midi Kit (QIAGEN) 190 according to the manufacturer's protocol. The presence and the size of the mutation in each 191 plasmid was verified by Sanger sequencing (GATC Biotech or Genewiz, Germany) using 192 primers listed in Table 2 . Sequence results were analysed in SSE v1.2 software [70] . 193 Bio-Rad) by heating up to 80°C for 2 hr under vacuum. Dried gels were placed in a sealed 211 cassette with an X-ray film (Fisher Scientific) overnight. X-ray films were developed using a 212

Konica SRX-101A X-ograph film processor following manufacturer's protocol. 213 214 Reverse genetics system. Viruses were recovered using the protocols described by Kanai 215 et al. [42] and Komoto et al. [43] , with slight modifications. At 70% confluency, monolayers of 216 BSR-T7 cells in 6-well plates were co-transfected with 11 plasmids corresponding to each 217 RV genome segment (2.5µg for plasmids encoding NSP2 and NSP5; 0.8µg for the 218 remaining plasmids) and plasmids encoding two vaccinia virus capping enzyme subunits 219 (pCAG-D1R and pCAG-D12L -0.8µg each) using 16µL Lipofectamine 2000 (Invitrogen) per 220 transfection reaction in a total volume of 200µL of Opti-MEM (Gibco). After 24 hr incubation 221 at 37°C 5% CO2, MA104 cells (1 x 10 5 cells/well) were added to transfected BSR-T7 cells 222 and co-cultured for 4 days in FBS-free DMEM supplemented with 0.5µg/mL porcine 223 pancreatic trypsin type IX (Sigma-Aldrich). Co-cultured cells were then lysed three times by 224 freeze/thaw and lysates were incubated with trypsin at a final concentration of 10µg/mL for 225 30 min at 37°C 5% CO2 to activate the virus. Lysates were then transferred to fresh MA104 226 cells in T25 flasks and incubated at 37°C 5% CO2 for 1 hr. After adsorption, MA104 cells 227

were washed and cultured in FBS-free DMEM supplemented with 0.5µg/mL trypsin type IX 228 for up to 7 days or until complete cytopathic effect was observed. Cells were then lysed 229 three times by freeze/thaw, pelleted and virus-containing supernatants (P1 stocks) were 230 aliquoted and stored at -80°C. To generate mutant SA11 or RF mutant viruses, plasmids 231 encoding either the SA11 VP4 or RF NSP3 gene segment were replaced with the 232 corresponding plasmid encoding SARS-CoV-2 spike epitopes. Mock preparations with the 233 mutated segment omitted were generated for use as negative controls throughout. All rescue 234 experiments were performed three times for each virus panel. The panels of viruses were 235 titred by plaque assays, and the presence of mutations in the target gene segments was 236 confirmed by RT-PCR and Sanger sequencing (GATC Biotech or Genewiz, Germany) as 237 described below. Properties of mutant viruses are summarised in Table 3 . 238 239 Cycler (Bio-Rad). SA11 VP4 PCR conditions were 1 cycle of 98°C for 30 sec, followed by 35 255 cycles of 20 sec at 98°C, 20 sec at 52°C and 2 min at 72°C, finishing with a 2 min incubation 256 at 72°C. RF NSP3 PCR conditions were 1 cycle of 5 min at 95°C, followed by 35 cycles of 257 30 sec at 95°C, 30 sec at 55°C and 2 min at 72°C, finishing with a 5 min incubation at 72°C. 258 PCR product length was confirmed by 0.8% agarose gel electrophoresis containing SYBR™ 259

Safe DNA Gel Stain and gels were imaged using the Odyssey ® XF imaging system (LI-260 COR). Analyses were performed with Image Studio™ Lite software (LI-COR). The presence 261 of sequence insertions was confirmed by sequencing at GATC Biotech or Genewiz, 262

Germany using primers listed in Table 2 . Sequence results were analysed in SSE v1.2 263 software [70] . 264 265 Plaque assay. Plaque assays for RVs were performed using adapted methods [71, 72] . 266

Confluent monolayers of MA104 cells in 6-well plates were washed with FBS-free DMEM 267 and infected with 800μL of ten-fold serially diluted virus for 1 hr at 37°C 5% CO2. Following 268 virus adsorption, 2mL/well overlay medium was added (1:1 ratio of 2.4% Avicel (FMC 269

Biopolymer) and FBS-free DMEM supplemented with 0.5μg/mL trypsin type IX) and 270 incubated for 4 days. Cells were then fixed for 1 hr with 1mL/well of 10% neutral buffered 271 formalin (NBF) (CellPath) and stained for 1 hr with 0.1% Toluidine blue (Sigma-Aldrich) 272 dissolved in H2O. 

RT-qPCR targeting segments VP1 (SA11 and RF), VP4 (SA11) and NSP3 (RF) were 287 designed using SSE v1.2 and OligoCalc software [70, 73] (Table 2) . To quantify total RNA levels of each segment relative to WT from timecourse experiments, 293 one-step RT-qPCR was performed using SensiFAST™ SYBR® Lo-ROX One-Step Kit 294 (Meridian Bioscience) with 2μL of RNA samples according to the manufacturer's protocol on 295 a Rotor-Gene Q apparatus (QIAGEN) using primers listed in Table 2 . RT-qPCR cycling 296 conditions were: 45°C for 10 min for reverse transcription, 2 min at 95°C, followed by 40 297 cycles of 10 sec at 95°C and 30 sec at 60°C. The conditions then increased from 50°C to 298 99°C at 1-degree increments to generate a melt curve to confirm specific amplification of 299 each gene. For relative expression of each gene, the results were analysed according to the 300 2 -ΔCt method. Average Ct values of each gene were normalised to WT at 8 hpi (when RNA 301 was first reliably detected for VP1, VP4 and NSP3) and the resulting ΔCt values were 302 adjusted for primer efficiency. Three independent experiments were performed in technical 303

To quantify the RNA copies of corresponding gene segments in mutant viruses for RNA:PFU 305 ratios, RNA from cDNA (approximately 100ng) of each gene segment was synthesised using 306 a MEGAscript™ T7 Transcription Kit (Invitrogen) followed by TURBO DNase treatment 307 (Invitrogen) and a clean-up with MinElute PCR Purification Kit (QIAGEN) according to 308 manufacturer's instructions. The extracted RNA was dissolved in 10μL nuclease-free water 309 (QIAGEN) and stored at -80°C. RNA was measured using Qubit™ RNA broad range assay 310 kit (Invitrogen) and serial dilutions of RNA were used as standards in one-step RT-qPCR. A 311 ratio of the transcript copy number of each gene segment in the mutant viruses to virus titre 312 was calculated and normalised to WT. Three independent experiments were performed in 313 technical triplicates with the standard done in duplicate. 314

The PCR efficiency (E) of each primer pair set was established by measuring serial dilutions 315 of cDNA of each segment in triplicate and calculated based on the slope of the standard 316 curve according to the formula E = (10 (-1/slope)-1 ) x 100. Threshold cycle (Ct) values equivalent 317 to mock samples and non-template control were considered to be negative. Table 4 ). After three 5 min washes with TBS-T, membranes were incubated for 1 hr at room 329 temperature with secondary antibodies diluted in TBS-T ( 330 Table 4 ). Following three 5 min washes with TBS-T, membranes were imaged using an 331

Odyssey ® XF imaging system (LI-COR). Analyses were performed with Image Studio™ 332

Lite software (LI-COR). 333 334 were dried onto filter paper and imaged using a Samsung Xpress C480FW scanner. 350

To visualise viral proteins of VP4 mutants, 25mL of clarified RV stock was used for 351 ultracentrifugation as described above. Viral pellets were resuspended in 1X TNC buffer 352 Laboratory Supplies), plates were blocked with 2% horse serum/PBS at room temperature 366 for 2 hr. After blocking, plates were incubated with 100μL/well primary antibody ( 367 Table 4 Infection of bovine enteroids with NSP3 mutants. 3D bovine enteroids were prepared as 381 described in Hamilton et al., 2018 [74] . Infection of organoids was carried out as described 382

by Derricott et al., 2019 [75] with modifications. 3D organoids were mechanically disrupted with 4% paraformaldehyde for 1.5 hr at 4°C with agitation. Following three washes with PBS, 399 enteroids were permeabilised with 0.5% Triton-X100 in PBS for 15 min and then blocked 400 with 2% horse serum in PBS for 1 hr, all at room temperature. Enteroids were incubated with 401 primary antibodies ( 402 Table 4 ) overnight at 4°C with agitation, then washed three times with PBS and incubated 403 with secondary antibodies ( 404 Table 4 ) and phalloidin (F-actin detection) (1:100) (Invitrogen) for 1 hr at room temperature. 405

All antibodies were diluted in 2% horse serum/PBS. Enteroids were then washed with PBS 406 three times with the addition of 4′,6-diamidino-2-phenylindole (DAPI) nuclear stain (1:5000) 407 (Invitrogen) for the final 10 min wash. Coverslips were rinsed in water and mounted onto 408 microscope slides (Thermo Scientific) with ProLong™ Gold Antifade Mountant (Invitrogen), 409 and imaged using a Zeiss LSM 710 confocal microscope at 630X magnification. Images 410 were analysed using the Zen Black software and processed using Photoshop v23. 

We engineered a panel of SA11 strain VP4 plasmids with SARS-CoV-2 spike peptide 422 sequences inserted into the hypervariable region, and a panel of RF strain NSP3 plasmids 423 with 3' tags of SARS-CoV-2 RBM or RBD with or without a separating T2A peptide (Fig. 1A) . 424

To confirm the expression of the spike epitopes by the mutated VP4 and NSP3 segments, 425 coupled in vitro transcription and translation (IVT) reactions were carried out using rabbit 426 reticulocyte lysate system supplemented with [ 35 S] methionine ( Fig. 1B and 1C) . 427

Following SDS-PAGE and autoradiography, SA11 WT and mutated VP4 constructs 428 produced polypeptide products of the expected protein size throughout (Fig. 1B) . The WT 429 RF NSP3 construct expressed a protein of expected size (Fig. 1C, lane 2) . RF NSP3 430 segment incorporating the RBM motif produced a protein of a higher molecular weight 431 consistent with its predicted size (Fig. 1C, lane 3) . The T2A_RBM construct generated a 432 product containing the NSP3 protein that ran slightly higher than WT NSP3 due to the 433 residual C-terminal fusion of partial T2A sequence [76] (Fig. 1C, lane 4, red asterisk) . The 434 small size (~10kDa) of the unconjugated RBM peptide made it difficult to visualise due to co-435 migration with the dye front. Unseparated NSP3-T2A_RBM, seen as a minor product, was 436 probably produced as a result of unsuccessful ribosome skipping at the T2A site [76] (Fig.  437 1C, lane 4, black asterisk). As expected, translation of NSP3 fused to the RBD peptide also 438 produced a protein of a higher molecular weight than the WT (Fig. 1C, lane 5) . For the 439 T2A_RBD construct, untagged NSP3 was readily identifiable, and again a fainter band 440 corresponding to the predicted molecular weight of fused NSP3-T2A_RBD was also seen 441 (Fig. 1C , lane 6, red and black asterisks respectively). The RBD product (~23kDa) was 442

reproducibly not detectable, possibly due to discontinued translation as a result of ribosome 443 fall-off at the T2A site, or degradation in the rabbit reticulocyte lysate system. These results 444

show that the spike epitopes were successfully translated in a cell-free system and that the 445 T2A peptide was functional. 446

To generate the VP4 and NSP3 mutant viruses, we used an adapted version of the 448 previously published SA11 reverse genetics system or our established plasmid-only RF 449 reverse genetics system respectively [42, 43] ( Fig. 2A) . With the exception of the NTD 450 mutant, all VP4 mutants displayed significantly lower titres compared to the WT, with 451 approximately two-log10 decrease in the titres observed across the panel (Fig. 2B) . 452

Nonetheless, all VP4 mutants were successfully rescued on all attempts. The VP4 mutants 453 exhibited a similar but smaller speckled plaque phenotype to the WT (Fig. 2C) . Statistical 454 analyses of plaque sizes were not performed due to the ambiguities in determining the 455 peripheries of individual plaques and their non-uniform diameters within a well. 456

In contrast, compared to the WT virus, there were no statistically significant differences 457 observed in titres of any of the NSP3 mutants (Fig. 2D) . However, out of three attempts, 458 RBM was successfully rescued twice whilst T2A_RBM was rescued at all times, with one of 459 the rescues showing a two-log10 lower titre (Fig. 2D) . The RBD mutant was only rescued 460 once with a titre that was a log10 lower than the WT (Fig. 2D) . Although T2A_RBD rescued 461 on all attempts, it delivered a five-log10 variation in titres between the rescues (Fig. 2D) . The 462 viruses with NSP3 tags displayed smaller plaques than the WT, except T2A_RBM, whose 463 plaques had a similar morphology to that of WT (Fig. 2E) . Notably, in the presence of T2A 464 peptide, plaque sizes were larger than when RBM or RBD was fused directly to the C-465 terminus of NSP3 without T2A (Fig. 2E) . 466

Thus, small peptide insertions into the hypervariable region of SA11 VP4 significantly 467 reduced the virus titre, while the efficiency of viral rescue was affected by the size of the 468 peptides fused to the C-terminus of RF NSP3. Furthermore, these data demonstrated that 469 the RF strain of RV was successfully rescued from cloned cDNAs (confirmed by sequencing, 470 data not shown). 471

Following virus rescue, we first examined the effects of peptide insertion on RV replication in 473 cell culture. Multi-step growth curves were performed for the VP4 and NSP3 mutants after 474 infection of MA104 cells at a low MOI. From 24 hpi, all VP4 mutants had significantly lower 475 titres (≥ one-log10) than the WT virus for at least one timepoint during the infection (Fig. 3A) . 476

Next, we assessed the impact of the introduced mutations on the total RNA expression 477 levels for VP1 and VP4 viral transcripts. No marked differences in transcript levels of either 478 VP1 (Fig. 3B) or VP4 (Fig. 3C) were observed (p ≥ 0.05, paired t-test). 479

Conversely, the NSP3 mutants followed similar replication kinetics to RF WT with three-log10 480 increases in titres between 8 and 16 hpi, with titres plateauing thereafter (Fig. 3D) . 481

Throughout, no major differences in titres between any of the viruses in the panel were 482 observed. For the NSP3 virus panel, mutants produced higher levels of VP1 (Fig. 3E) and 483 NSP3 (Fig. 3F ) transcripts than the WT virus earlier in the time course. Higher RNA levels of 484 VP1 and NSP3 in the RBD mutant is likely a result of higher RNA input (Fig. 3E-F  485 respectively). Neither VP1 nor NSP3 transcript levels differed significantly throughout the 486 time course (p ≥ 0.05, paired t-test) (Fig. 3E-F) . 487

Western blot analyses were then used to characterise the production of SARS-CoV-2 spike 488 polypeptides and RV VP6 using whole-cell lysates from the infection time course. As 489 different SARS-CoV-2 peptides were introduced in each mutant, anti-spike antibody affinity 490 may vary between mutants and so expression levels cannot be compared. Due to lack of 491 available antibodies, we were unable to measure RV VP4 and NSP3 protein levels directly. 492

For the detection of various SARS-CoV-2 spike peptides in cells infected with the VP4 493 mutants, a polyclonal SARS-CoV-2 spike antibody was used. As the spike peptides were 494 introduced into the hypervariable region of VP4, we considered the possibility of detecting 495 the full length VP4 (sizes indicated in Table 3) , and/or the VP8* cleaved product containing 496 the spike peptides. At 32 hpi, full length VP4 product was detected in HR2, NTD and RBM.2 497 mutants (Fig. 4A, black asterisks) . No SARS-CoV-2 spike peptide signal was detected for 498 the RBM.1 mutant, likely reflecting selective clonality of the antibody used (Fig. 4A) . 499

Interestingly, at 32 hpi, the upper band representing the uncleaved VP4 product was brighter 500 for RBM.2 (Fig. 4A, black asterisks) , whereas for NTD, the lower band possibly representing 501 the VP5* cleaved product was stronger than the upper band (Fig. 4A, green asterisks) . 502

These differences could be due to the distinct efficiencies of VP4 processing caused by the 503 inserted peptides. The hypervariable region where the spike peptides were introduced is 504 within the VP8* domain, and we would expect a product of around 30-31kDa in the event of 505 VP4 cleavage. Possibly, this is evident in NTD at 32 and 48 hpi (Fig. 4A, blue asterisks) , 506 although the NTD mutant showed several unexpected bands and so this may reflect non-507 specific antibody binding; a faint product of similar molecular weight was also detected in the 508 mock samples at 32 hpi (Fig. 4A) . 509

We also examined RV VP6 production which was detected in all VP4 viruses at 32 hpi (Fig.  510 4A). VP6 signal intensity did not correlate with spike signal intensity across viruses, again 511 likely due to variable affinities of the spike antibody for the different spike peptides 512 incorporated into VP4 (Fig. 4A) . 513

In cells infected with NSP3 mutants, detection of both RBD and RBM peptides was possible 514 using a polyclonal SARS-CoV-2 RBD antibody. Cross-reactivity for the RBM-expressing 515 virus was first detected at 8 hpi, with increased levels present at 16 and 32 hpi (Fig. 4B) . 516

Antibody cross-reactivity for the T2A_RBM virus was only visible between 16 and 32 hpi 517 (Fig. 4B) . The band detected is consistent in size with an unseparated NSP3-T2A_RBM 518 protein product (Table 3, Fig. 4B, black asterisks) . The RBM peptide of around 10kDa was 519 not detected, consistent with its in vitro translation efficiency (Fig. 4B, Fig. 1B) . For the RBD 520 virus, expression of the RBD peptide was first observed at 8 hpi but declined from 16 hpi 521 paralleling the disappearance of tubulin (likely due to cell death) (Fig. 4B ). In the T2A_RBD 522 mutant, both NSP3-conjugated and unconjugated RBD products were detectable from 16 hpi 523 (Fig. 4B , black and red asterisks respectively). Over time, the RBD peptide (~23kDa) 524 became more apparent, as the NSP3-RBD signal diminished, confirming the functionality of 525 the T2A element (Fig. 4B, red asterisks) . 526

It is unclear why the T2A-induced ribosomal skipping appeared to improve in efficiency over 527 the course of infection. It is possible that the stability of the fused peptides is lower than the 528 separated peptides. Similarly, over the course of infection, RBM protein levels increased 529 throughout, whereas RBD protein levels increased until 16 hpi after which they dropped 530 dramatically (Fig. 4B) . 531

The expression of VP6 was first observed at 8 hpi during infection with the RBD mutant, 532 correlating with the signal of the RBD peptide, but was only detected for the remaining 533 viruses from 16 hpi (Fig. 4B ). Since equal MOIs were used for time course infections, higher 534 input of genomic RNA copies could explain earlier V6 detection in the RBD mutant. This is 535 consistent with the higher transcript levels detected for the RBD mutant at 8 hpi ( Fig. 3E-F) . 536

In the event of a packaging defect of the RBD mutant more input genome copies would be 537 required to deliver equal numbers of infectious particles. 538

Infection with both mutants resulted in a similar drop in tubulin levels ( Fig. 4A-B) , suggesting 539 that the two mutants have different protein turnover rates. 540

In summary, introducing short SARS-CoV-2 peptides into the hypervariable region of VP4 541 impacted the virus yield, whereas fusing SARS-CoV-2 spike peptides to the C-terminus of 542 NSP3 with or without T2A did not affect the virus titre or replication kinetics. Nevertheless, 543 SARS-CoV-2 spike peptides were detectable in the majority of mutants using polyclonal 544 antibodies, encouraging further investigations into the potential of these strategies for 545 heterologous peptide presentation. 546

For viruses of the Reoviridae family, genome packaging is a tightly orchestrated and 548 controlled process, so increasing the RV genome size may affect packaging efficiency [30, 549 77-79] . To examine whether our mutations may have affected genome packaging, RNA was 550 extracted from equal volume of purified viruses and analysed by urea-PAGE and silver 551

The VP4 virus panel showed the expected constellation of genome segments, but as 553 samples were resolved for an extended period of time to visualise the small changes in 554 segment 4 (VP4) band sizes expected as a result of SARS-CoV-2 spike peptide insertion, 555 segments 10 and 11 ran off the gel (Fig. 5A) . When gels were run for a shorter time, no 556 differences in band densities of segments 10 and 11 were seen (data not shown). Mutated 557 VP4 segments migrated more slowly than the WT VP4 segment, reflecting the various sizes 558 of inserted spike sequences (Fig. 5A, highlighted in red) . Densitometry analysis showed no 559 substantial differences in band density between mutated VP4 and WT VP4 segments 560 (normalised to WT: HR2 = 0.96, NTD = 1.11, RBM.1 = 0.87, RBM.2 = 0.96), suggesting no 561 obvious packaging defects were introduced by small sequence insertions into the 562 hypervariable region of VP4 (Fig. 5A) . 563

The resolved segments for the NSP3 virus panel also showed the expected pattern of 11 564 RNA segments, as well as a prominent background band present in a mock infected sample 565 that migrated between segments 1 and 2 (Fig. 5B ). Segment 7 (encoding NSP3) from all 566 mutants migrated notably slower than from WT virus, corresponding with its increased gene 567 size (Fig. 5B, highlighted in red) . Co-migration of segments 7, 8 and 9 in the WT made it 568 difficult to reliably separate segment 7, precluding direct quantitative analyses (Fig. 5B) . 569

Nevertheless, no obvious defect in packaging was observed through visualisation of the 570 complete genome of RF mutants. 571

To further evaluate whether virion infectivity may have been affected by genome 572 mutagenesis, the genome copy number to PFU ratio was determined for the two panels of 573 viruses, measuring VP1 and either VP4 (SA11 panel) or NSP3 (RF panel) segments (Fig.  574 5C-F). No significant differences were observed in the levels of VP1 and VP4 segments 575 across the VP4 mutants relative to WT, although for all mutants a 1-2 log10 increase in RNA 576 copies required to make an infectious virion was observed for both VP1 and VP4 (p ≥ 0.05, 577 paired t-test) (Fig. 5C-D) . 578

In contrast, all NSP3 mutants, with the exception of RBD, had equivalent numbers of VP1 579 and NSP3 segments (Fig. 5E-F) . RBD had a significantly higher VP1 (p = 0.005, paired t-580 test) and NSP3 (p = 0.010, paired t-test) segment copy number:PFU ratio, with over 100-fold 581 more copies of RNA required to make a fully infectious particle (Fig. 5C-D) . 582 VP4 is important for viral attachment and entry as well as for the maturation of TLPs that constitute an 583 infectious virus [1]. Therefore, it was considered that mutation of VP4 may affect the assembly of the 584 viral structural proteins. To test this, purified VP4 viruses were further analysed by SDS-PAGE 585 followed by Coomassie staining (NSP3 viruses were not included because the spike epitopes were 586 fused to NSP3 which is not incorporated into virions) (Fig. 5G) . Apart from VP3 and VP7 proteins, 587 major structural proteins were detected throughout (Fig. 5G) 

A notable difference between the WT virus and its mutants was the cleavage pattern of VP4 594

in the presence of trypsin used for propagating the panel of mutant viruses. A band of the 595 predicted size for VP5* was visible in the WT (Fig. 5G, red arrow) . However, no 596 corresponding band was visible for any of the VP4 mutants expressing heterologous 597 peptides, further supporting our earlier hypothesis that peptide insertion into the VP4 598 hypervariable region impairs its processing into VP8* and VP5* (Fig. 4A) . 599

In contrast, VP8* was detected in all mutants, and it migrated slightly higher than WT VP8, 600 corresponding to the additional peptide sequences present in the hypervariable domain of 601 VP4 (Fig. 5G) . Nevertheless, VP8* band density was stronger for WT than for the mutants, 602 again consistent with inefficient tryptic cleavage of VP4. 603

Overall, no major effect on viral assembly was observed in either of the mutant viral panels, 604 with the exception of inserting the large RBD tag into NSP3 which caused a defect that could 605 be overcome by incorporation of a T2A element. 606

To further investigate the viability of using RV as a delivery vector for SARS-CoV-2 spike 608 antigens, we assessed whether the RV mutants expressing spike peptides would cross-react 609 with RBD antibodies in an indirect ELISA. For the VP4 mutants, no antibody cross-reactivity 610 was identified in this assay using the SARS-CoV-2 RBD antibody (Fig. 6A) . Additionally, no 611 cross-reactivity was observed using the SARS-CoV-2 spike antibodies to target the various 612 spike peptides in the VP4 mutants, or with the PR8-RBM positive control (data not shown). 613

Except the T2A_RBM mutant, all NSP3 mutants consistently showed a higher signal than 614 the background signal from the WT virus, although only the RBD mutant showed a 615 significantly higher signal than WT (p = 0.032, paired t-test) (Fig. 6B) . 616

Based on these observations, RF NSP3 may be a better target for heterologous peptide 617 conjugation than SA11 VP4 for expression of immunogenic antigens, which is at least partly 618 attributable to the larger insertion site tolerated in the corresponding genome region. 619

Cell culture-based assays identified the RF NSP3 gene as a strong candidate for expressing 621 immunogenic heterologous peptides. Orally administered live attenuated RV vaccines 622 replicate in the gastrointestinal tract, and so bovine intestinal organoids containing 623 enterocytes, goblet, Paneth, enteroendocrine and stem cells [75] were used to investigate if 624 a more physiologically representative system was susceptible to infection with the NSP3 625 mutants. A bovine organoid system was chosen as the RF strain was first identified in 626 diarrhoeic calves [80] . 627 3D bovine enteroids were infected with NSP3 mutants at an approximate MOI of 10, stained 628 with anti-VP6 and anti-RBD antibodies and imaged by confocal microscopy ( Fig. 7A and 7B  629 respectively). At 24 hpi, VP6 was predominantly detected in the epithelium comprised of 630 mature enterocytes lining the apical surface of the organoid lumen (Fig. 7A) , which is 631 consistent with previous findings showing their preferential infection by RVs [19, 20] . In 632 contrast, VP6 distribution in RBM and T2A_RBM mutants was detected around the nuclei of 633 cells located within the organoid lumen and in the punctate cytoplasmic inclusion bodies, 634 most likely viroplasms, sites of RV replication and assembly [1] (Fig. 7A, VP6 panel) . VP6 635 signal in the RBD and T2A_RBD mutants was not apparent in the lumen and was mainly 636 detected around the nuclei of cells in the organoid lining (Fig. 7A, VP6 panel) . 637

Following staining with the SARS-CoV-2 RBD antibody, RBM signal was detected around 638 the nuclei of cells in the organoid lining while other mutants did not produce any visible anti-639 RBD cross-reactivity (Fig. 7B ). This is comparable to western blotting data where RBM also 640 gave the strongest signal (Fig. 4B) . 641

We have used simian SA11 and bovine RF RV strains as viral vectors to express various 644 SARS-CoV-2 spike epitopes as a model system with which to test the potential for 645 expression of multivalent antigens. Tagging of the VP4 protein found on the surface of the 646 virion with smaller peptides consistently impaired viral growth and did not yield strong 647 antibody cross-reactivity. Conversely, relatively large foreign sequences could be tagged to 648 the C-terminus of NSP3 protein, mostly without impairing viral titres and replication kinetics, 649 and cross-reactivity with SARS-CoV-2 RBD antibodies was also demonstrable. 650

To investigate the feasibility of using RV as an expression vector, we analysed the effect of 651 introducing SARS-CoV-2 spike peptides into the 'head' region of the VP4 (i.e. VP8* domain), 652 outside the sialic acid binding domain. Since VP8* is used in several vaccine platforms such 653 as protein subunit or nanoparticle vaccines to induce RV-specific neutralising antibodies [81- TLPs (Fig. 5G) . In vitro assembly of TLPs has shown that DLPs require addition of VP4 672 before VP7, and that the tryptic cleavage occurs following addition of VP7 [90, 94] . 673

Interaction of VP7 subunits with the VP5* foot domain stabilise and anchor the VP4 onto the 674 virion, allowing the protease to cleave only the linker sequence that bridges VP8* and VP5*, 675 which in our mutants is intact [22, 23, 67, 68] . It is possible that our peptide insertions may 676 have altered the VP4 conformation at this interaction site, disrupting VP4 cleavage and 677 rendering VP5* undetectable in our assays ( Fig. 4A and 5G) . 678 679 Finally, we were unable to detect any cross-reactivity of the SARS-CoV-2 spike peptides 680 with spike (not shown) or RBD antibodies in ELISA (Fig. 6A) Rescue of the RF mutant with the RBD peptide (193 amino acids) fused directly to NSP3 702 was achieved only once in three attempts (Fig. 2D) , possibly reflecting an impaired function 703 of NSP3. We also found that the RNA:PFU ratio was affected in the RBD mutant only (Fig.  704 5E-F), and that copy numbers of NSP3 and VP1 encoding transcripts were both affected. 705

This cannot be attributed to the longer length of this segment, as the NSP3-T2A_RBD 706 mutant did not demonstrate any packaging defect ( Fig. 5E-F) . Therefore, NSP3 exerts its 707 role in virus replication by regulating viral mRNA translation. This is consistent with the 708 observed increase in RNA:PFU ratios being nonspecific to a particular segment. 709

With the exception of T2A_RBM, NSP3 mutants expressing SARS-CoV-2 spike peptides 710 cross-reacted with the SARS-CoV-2 RBD antibody (Fig. 6B) , supporting the previously 711

proposed viability of using NSP3 as a tagging system [52] and demonstrating its application 712 to other RV strains. Follow up studies using sera from COVID-19 patients are needed to 713 confirm antibody responses to the antigens produced by the NSP3 mutants. 714

Overall, we conclude that the diminished infectivity of VP4 mutants accommodating only 715 small peptide insertions further limits the use of the hypervariable region for foreign 716 sequence expression and thus VP4 as a potential vaccine expression platform. On the other 717 hand, including the T2A element is beneficial for expressing foreign antigens as it allows co-718 expression of NSP3 and large peptides with little effect on the viral rescue efficiency, titre 719 and replication, all which are important traits for live attenuated vaccine development. In the 720 absence of the T2A element there may be a limit to the amount of additional foreign 721 sequences that the RF NSP3 can accommodate. Our results offer a possibility of utilising a 722 bovine RF RV as a backbone to facilitate the development of recombinant RV-based 723 vaccines; here we used SARS-CoV-2 as an example, but this can be extrapolated to other 724 gastrointestinal pathogens. 725

Acknowledgements. We want to thank lab members, central support unit (CSU), 726 bioimaging and technical staff for their support and assistance with this project. We thank the Kanai et al., 2017 and Komoto et al., 2018; created 988 with BioRender.com) [42, 43] . Full length cDNAs representing each of the 11 gene segments were 989 transfected into BSR-T7 cells with increasing amounts of two plasmids carrying NSP2 and NSP5 990 genes, along with two plasmids expressing vaccinia virus capping enzyme genes (D1R and D12L). 

Representative results from three independent rescues is shown (except for RBD, which only rescued 995 once). In (D) open circles show failed rescues plotted at the limit of detection.

Estimates of global, regional, and national morbidity, 745 mortality, and aetiologies of diarrhoeal diseases: a systematic analysis for the Global 746 Burden of Disease Study

Review of group A rotavirus strains reported in swine and cattle. Vet 753 Microbiol

Preweaned heifer management on US dairy operations: Part V

Factors associated with morbidity and mortality in preweaned dairy heifer calves

Factors associated with morbidity, mortality, and growth of 758 dairy heifer calves up to 3 months of age

Intestinal and extra-intestinal pathogenicity of a bovine reassortant 760 rotavirus in calves and piglets

Vaccines for preventing rotavirus diarrhoea: vaccines in use. The 762 Cochrane database of systematic reviews

Development of a rotavirus vaccine: clinical safety, 767 immunogenicity, and efficacy of the pentavalent rotavirus vaccine

Global Impact of Rotavirus Vaccination on Childhood 770 Hospitalizations and Mortality From Diarrhea. The Journal of Infectious Diseases

Implementation of a pre-calving vaccination programme 775 against rotavirus, coronavirus and enterotoxigenic Escherichia coli (F5) and 776 association with dairy calf survival

Group A Rotavirus Veterinary Vaccines. The Journal 778 of Infectious Diseases

Characterization of the NSP6 protein product of 782 rotavirus gene 11

Functional Studies on the Rotavirus Non-Structural Proteins NSP5 and 784 NSP6

Expression of rotavirus proteins encoded by alternative open 787 reading frames of genome segment 11

Rotavirus infection. Nat Rev Dis Primers

Rotavirus Interactions With Host Intestinal Epithelial Cells. 790 Frontiers in immunology

Structural rearrangements in the membrane penetration 792 protein of a non-enveloped virus

Trypsin cleavage stabilizes the rotavirus VP4 spike

Trypsin activation pathway of rotavirus infectivity

Cell attachment protein VP8* of a human rotavirus specifically interacts 798 with A-type histo-blood group antigen

The VP8 fragment of VP4 is the rhesus 802 rotavirus hemagglutinin

Rotavirus VP1 alone specifically binds to the 3' end of viral mRNA, but 804 the interaction is not sufficient to initiate minus-strand synthesis

Structural insights into the coupling of 807 virion assembly and rotavirus replication

Genome packaging in multi-811 segmented dsRNA viruses: distinct mechanisms with similar outcomes. Current 812 opinion in virology

A four-nucleotide translation enhancer in the 3'-816 terminal consensus sequence of the nonpolyadenylated mRNAs of rotavirus

Rotavirus RNA-binding protein NSP3 interacts with eIF4GI and 821 evicts the poly(A) binding protein from eIF4F. The EMBO journal

Nuclear localization of cytoplasmic poly(A)-binding protein upon 824 rotavirus infection involves the interaction of NSP3 with eIF4G and RoXaN

Rotavirus prevents the expression of host responses by blocking 827 the nucleocytoplasmic transport of polyadenylated mRNAs

Reverse genetics system for introduction of 832 site-specific mutations into the double-stranded RNA genome of infectious rotavirus

Rearranged genomic RNA segments offer a new approach to the 837 reverse genetics of rotaviruses

Generation of an Avian-Mammalian Rotavirus Reassortant by Using 839 a Helper Virus-Dependent Reverse Genetics System

Entirely plasmid-based reverse genetics system for rotaviruses

Generation of Recombinant Rotaviruses Expressing Fluorescent 844 Proteins by Using an Optimized Reverse Genetics System

An Optimized Reverse Genetics System Suitable for 846 Efficient Recovery of Simian, Human, and Murine-Like Rotaviruses

Generation of Recombinant Rotavirus Expressing NSP3-UnaG 849 Fusion Protein by a Simplified Reverse Genetics System

Reverse Genetics Approach for Developing Rotavirus Vaccine 851 Candidates Carrying VP4 and VP7 Genes Cloned from Clinical Isolates of Human 852 Rotavirus

Rescue of Infectious Rotavirus Reassortants by a Reverse 854 Genetics System Is Restricted by the Receptor-Binding Region of VP4. Viruses

Recent advances in rotavirus reverse genetics and its 857 utilization in basic research and vaccine development. Archives of virology

VP4-and VP7-specific antibodies mediate heterotypic immunity to 860 rotavirus in humans. Science translational medicine

Human VP8* mAbs neutralize rotavirus selectively in human 862 intestinal epithelial cells

Expression of Separate Heterologous Proteins from the 864

Segment Using a Translational 2A Stop-Restart Element

Rotavirus as an Expression Platform of the SARS-CoV-2 867 Spike Protein. bioRxiv

SARS-CoV-2 vaccines in development

SARS-CoV-2 vaccines strategies: a comprehensive review of 871 phase 3 candidates. NPJ Vaccines

A Single-Dose Intranasal ChAd Vaccine Protects Upper and 873 Lower Respiratory Tracts against SARS-CoV-2. Cell

Ad26 vector-based COVID-19 vaccine encoding a prefusion-stabilized 875 SARS-CoV-2 Spike immunogen induces potent humoral and cellular immune 876 responses. NPJ vaccines

COVID-19 vaccine BNT162b1 elicits human antibody and TH1 T cell 878 responses

mRNA vaccine-elicited antibodies to SARS-CoV-2 and circulating 880 variants

A vaccine targeting the RBD of the S protein of SARS-CoV-2 induces 882 protective immunity

RBD-Fc-based COVID-19 vaccine candidate induces highly potent 884 SARS-CoV-2 neutralizing antibody response. Signal Transduction and Targeted 885 Therapy

CoV-2 Spike Receptor-Binding Domain by Structure-Guided High-Resolution

Structure of the SARS-CoV-2 spike receptor-binding domain bound to 890 the ACE2 receptor

Generation of Recombinant Rotavirus with an Antigenic Mosaic of 892 Cross-Reactive Neutralization Epitopes on VP4

The rhesus rotavirus VP4 sialic acid binding domain has a 894 galectin fold with a novel carbohydrate binding site. The EMBO journal

Functional mapping of B-cell linear epitopes of SARS-CoV-2 in COVID-897 19 convalescent population

Immunodominant linear B cell epitopes in the spike and 899 membrane proteins of SARS-CoV-2 identified by immunoinformatics prediction and 900 immunoassay

Functional refolding of the penetration protein on a non-902 enveloped virus

Safety and efficacy of a pentavalent human-bovine (WC3) 906 reassortant rotavirus vaccine

SSE: a nucleotide and amino acid sequence analysis platform. BMC 908 Res Notes

Culturing, storage, and quantification of 910 rotaviruses. Current protocols in microbiology

OligoCalc: an online oligonucleotide properties calculator

Development of in vitro enteroids derived from bovine small 916 intestinal crypts

Developing a 3D intestinal epithelium model for livestock species

Protein-mediated RNA folding governs sequence-specific 922 interactions between rotavirus genome segments

Assortment and packaging of the segmented 924 rotavirus genome

Structural basis of rotavirus RNA chaperone displacement and 926 RNA annealing

Rotavirus spike protein ΔVP8* as a novel carrier protein for 932 conjugate vaccine platform with demonstrated antigenic potential for use as bivalent 933 vaccine

A Nanoparticle-Based Trivalent Vaccine Targeting the Glycan Binding 935 VP8* Domains of Rotaviruses

Norovirus P particle, a novel platform for vaccine development and 937 antibody production

Tandem truncated rotavirus VP8* subunit protein with T cell epitope 939 as non-replicating parenteral vaccine is highly immunogenic. Hum Vaccin 940 Immunother

Expression and characterization of a novel truncated rotavirus VP4 for 942 the development of a recombinant rotavirus vaccine. Vaccine

Trypsin cleavage stabilizes the rotavirus VP4 spike

Viral membrane fusion

A rotavirus spike protein conformational intermediate binds lipid 949 bilayers

Proteolysis of monomeric 951 recombinant rotavirus VP4 yields an oligomeric VP5* core

Effect of mutations in VP5 hydrophobic loops on rotavirus cell entry

Generation of recombinant rotavirus with an antigenic mosaic of 956 cross-reactive neutralization epitopes on VP4

VP5* rearranges when rotavirus uncoats

Assembly of highly infectious rotavirus particles 960 recoated with recombinant outer capsid proteins

Rotavirus vaccines: recent 962 developments and future considerations

Challenging the Roles of NSP3 and Untranslated Regions in 964 Rotavirus mRNA Translation

Expression of VP6 and SARS-CoV-2 spike peptides in infected cells. Cells infected at 1003 low MOI were harvested at 1, 8, 16, 24, 32 and 48 hpi

In (A) black 1006 asterisks denote uncleaved VP4 product. Cleaved VP4 products VP5* and VP8* are marked by green 1007 and blue asterisks respectively. In (B) black asterisks show T2A read-through product and red 1008 asterisks identify separated products. Representative results from three independent experiments are 1009 shown

Extracted 1013 RNA from virus stocks was analysed by urea-PAGE and silver staining for SA11 mutants (A) and RF 1014 mutants (B). Individual RNA segments are labelled in black, and mutated RNA segment notations are 1015 in red. Lane "L" represents High Range RNA Ladder showing band size of RNA transcripts. RNA:PFU 1016 ratios of VP1 and either SA11 VP4 (C -D) or RF NSP3 (E -F) genes were determined by RT-qPCR 1017 and a ratio of copy number to viral titre was calculated

Electrophoretic profile of viral proteins 1020 from purified VP4 mutants visualised using Coomassie brilliant blue. Lane "L" represents protein 1021 ladder indicating the molecular weight markers (kDa)

Coloured symbols represent individual data points obtained from three 1027 independent experiments. OD (405nm) signal for all mutants was normalised to WT (background 1028 signal). PR8-RBM represents positive control