A locally administered single cycle influenza vaccine expressing a non-fusogenic stabilised haemagglutinin stimulates strong T-cell and neutralising antibody immunity

word count: 228 13 Text word count: 4917 14 15 16 17 18 19 20 21 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint

22 Current influenza vaccination approaches protect against specific viral strains, but do not 23 consistently induce broad and long -lasting protection to the diversity of circulating influenza 24 viruses. Single cycle viruses delivered to the respiratory tract may offer a promising solution as 25 they safely express a diverse array of viral antigens by undergoing just one round of cell 26 infection in their host and stimulate broadly protective resident memo ry T-cell responses in the 27 lung. We have previously developed a vaccine candidate called S-FLU that is limited to a single 28 cycle of infection by inactivation of the hemagglutinin signal sequenc e and induces a broadly 29 cross-reactive T-cell response and antibodies to neuraminidase, but fails to induce neutralising 30 antibodies to hemagglutinin after intranasal administration. 31 This study describes the development of CLEARFLU, a derivative of S -FLU that is designed to 32 add a neutralising antibody response to hemagglutinin. In contrast to S -FLU, which does not 33 express a hemagglutinin molecule at the infected cell surface, CLEARFLU viruses express a 34 stabilised non-fusogenic hemagglutinin. They are equally limited to a single cycle of infection, 35 but induce a neutralising antibody response to the expressed hemagglutinin in addition to the 36 cytotoxic T lymphocyte ( CTL) responses to internal proteins and antibodies to neuraminidase 37 induced by S -FLU. This represents a notable advantage as CLEARFLU viruses may provide 38 sterile immunity against strain -matched challenge as well as non -sterile protection against a 39 broad range of influenza viruses. 40 Importance: 41 Influenza is a serious public health concern, causing seasonal epidemics as well as pandemics in 42 people. Influenza can also cause severe agricultural losses due to its circulation in farmed poultry 43 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint and swine. A major challenge in the control of influenza is the diversity of circulating viruses. 44 Developing vaccines which stimulate immunity to a wide array of influenza viruses is therefore 45 important for protecting human and animal populations from disease and death. In this study, we 46 describe a n approach for developing influenza vaccines which trigger immune mechanism s 47 shown to induce broad protection against a diversity of viruses, while also conserving the strong 48 protection against specific strains observed in existing vaccines. 49 50

51 Viruses that undergo a single replication cycle offer a promising platform for generating broadly 52 protective immune responses to influenza (1-3). They are designed to lack a functional copy of 53 one or more of the genes required for influenza replication and can only be propagated when the 54 related gene product is supplied in trans. While single cycle viruses can be propagated to high 55 titre in permissive cell culture, they are able to undergo only a single cycle of cell infection in 56 their natural hosts. 57 Single cycle influenza viruses (SCIV) with a functioning polymerase undergo genome 58 replication and expression in their host, leading to amplification of viral epitopes inside infected 59 cells and cell surface expression of encoded viral coat proteins. As SCIV typically contain all but 60 one intact viral segment, responses can be targeted to a wide array of RNA, peptide and protein 61 components. SCIV are therefore capable of generating strong antibody and T -cell responses, 62 including broadly protective cytotoxic T -cell responses to peptides from highly conserved 63 internal core proteins (4). SCIV hold an advantage over conventional live -attenuated vaccines in 64 that they infect very few cells and so are unlikely to cause pathology or mutate or reassort into 65 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint virulent forms. SCIV can be safely administered by aerosol to immunise the lung, which is 66 associated with local immunity through tissue-resident T-cells (5-7) and mucosal IgA (8). 67 Multiplasmid transfection is commonly used to produce influenza viruses de novo in cell culture 68 (9-11). To generate single cycle viruses, one or more producer plasmids can be mutated or 69 omitted. SCIV vaccine candidates with deficiencies in various genes have been developed, 70 including in matrix 2 (3), neuraminidase, the polymerase genes and the hemagglutinin . 71 Inactivating the hemagglutinin is particularly advantageous as it prevents functional 72 hemagglutinin reassortment. Hemagglutinin activity is also easy to compensate for in trans by 73 growing viruses in cells which express functional coating hemagglutinin on their surface. 74 Altering the coat protein in the producer cell lines (pseudotyping) can allow single cycle viruses 75 with the same core genes but varying coating hemagglutinins to be produced quickly from a 76 single seed virus. This has the advantage that the coating hemagglutinin can be selected to avoid 77 prior immunity and enhance the cell mediated immune response (12). 78 We have previously described the single cycle virus S-FLU, which is produced by inactivation of 79 the hemagglutinin signal sequence, generating a viral RNA called S -HA (signal minus 80 hemagglutinin) ( 13). Like for most other SCIVs lacking hemagglutinin activity, no 81 hemagglutinin epitopes are presented on the surface of cells infected with S -FLU. In other 82 designs, the majority of the hemagglutinin coding sequence can be replaced by other genes; for 83 example fluorescent proteins to follow cell infection ( 14), or NY -ESO-1 to induce tumour 84 immunity (12). S-Flu viruses have been helpful laboratory tools for evaluating the susceptibility 85 of different coat proteins to chemical inhibitors and antibodies. Recently, S -FLU viruses coated 86 in avian H7 hemagglutinins ( 15) and the glycoprotein from Ebola viruses (16,17) have been 87 evaluated as safe and accurate pseudotypes for assaying inhibition of cell entry. This enables the 88 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint characteristics of coat proteins from highly virulent viruses to be investigated in biosafety 89 containment 1/2 rather than in high containment facilities. 90 S-FLU viruses have shown promise as broadly protective vaccines in mice, ferrets and pigs. 91 They can be administered into the lung intranasally (mice and ferrets) or by aerosol (pigs) 92 without causing pathology. Heterologous protection is induced in the absence of a neutralising 93 antibody response to hemagglutinin at the low doses administered to the lung. Protection is 94 associated with strong T -cell responses, including CD8+ T -cell responses in the lung (7), and a 95 strain-specific antibody response to neuraminidase which reduces challenge virus titres in the 96 respiratory tract in mice (13,15) and ferrets (4,18) and may (19) or may not (18) reduce challenge 97 virus titres in pigs. No sterile immunity is evoked in any model species following administration 98 of S-FLU to the lung. While S -FLU viruses protect against severe disease, they therefore are not 99 able to prevent infection, even if they are perfectly antigenically matched to the challenge strain. 100 In this study, we explore whether expressing hemagglutinin epitopes from a non -functional 101 hemagglutinin molecule instead of S -HA or eGFP in the S -FLU expression cassette drives their 102 cell surface expression and stimulates neutralising antibody responses without negatively 103 impacting on the existing immunogenicity and single cycle nature of S-FLU viruses. 104 A multitude of single or double nucleotide mutations which significantly ablate or abolish 105 hemagglutinin function have been described previously. While alone these mutations generate 106 single cycle viruses with a high rate of reversion to infectivity, using them in combination could 107 allow for the generation of a full length, rationally designed, multi -mutated hemagglutinin 108 molecule which is irreversibly non-functional but antigenically preserved. 109 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Here, we describe the iterative design of a non -functional multi -mutated hemagglutinin 110 immunogen named CLEARFLU. We show that single cycle influenza viruses expressing 111 CLEARFLU hemagglutinin generate neutralising antibody responses as well as broadly reactive 112 T-cell responses in mice, when administered at low dose to the lung. 113

114 CLEARFLU design and expression 115 The design of CLEARFLU hemagglutinins was refined through three iterations using H1, H3 116 and H7 hemagglutinins (Table 1). Each CLEARFLU design included seven independent sets of 117 mutations known to block or heavily ablate hemagglutinin fusion activity, acting through up to 118 five different mechanisms: resistance to proteolytic cleavage ( 20,21), conformation locking 119 through disulphide bonding ( 22,23), receptor inactivation ( 24,25), inhibition of the fusion 120 peptide (26) and B-loop inactivation (27). 121 In the first design, the selected mutations disrupted binding by some antibodies (T1 -3B, T3-5D, 122 T2-6D (28) and MEDI8852 (29)) to the stem of H1 and H3 hemagglutinins expressed in cell 123 lines (Figure 1). This binding was restored for CLEARFLU version 2 by reverting mutations 124 W14A and W21A to wild -type and replacing these with I6G and G8A, which also inactivate the 125 fusion peptide but are closer to the N terminus and do not interfere with the binding site for these 126 antibodies. Antibody binding to the hemagglutinin globular head did not appear to be affected by 127 the introduction of any candidate mutations. 128 To verify that the mutations in the CLEARFLU version 2 design did indeed inactivate 129 hemagglutinin activity, we tested the replication of S -FLU viruses in cell lines expressing H7 130 A/Hong Kong/125/2017 hemagglutinins each with a single inactivating mutation (see Figure 2). 131 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint As S -FLU viruses express eGFP instead of hemagglutinin, green comma -shaped plaques of 132 infected cells are formed only in cell lines expressing functional hemagglutinin. We found no 133 evidence that the Y98F mutation previously described to inactivate the sialic acid binding site 134 reduced the growth of S -FLU in H7 transduced MDCK-SIAT1 cells. This supports previous 135 observations by others that Y98F leads to reduced receptor binding, but only limits infection in 136 cells with low sialic acid expression ( 30,25). As MDCK -SIAT1 cells over express α -2,6-sialic 137 acids, this effect was not seen here (31). The G8A mutation appeared to significantly reduce the 138 size of viral plaques, and we did not observe peaks of fluorescence in middle dilutions to suggest 139 efficient expansion of single viral clones. This suggests that the intermediate fluorescence 140 recorded for G8A is due to many viruses replicating poorly and that G8A reduces but does not 141 abolish fusion activity in this context. All the other mutations appeared to completely abolish 142 hemagglutinin-dependent replication. 143 To further refine the design of CLEARFLU, we removed the Y98F mutation to avoid disrupting 144 untested neutralising epitopes around the receptor binding site. To maintain the number of 145 inactivating mutations, we introduced two proline residues at positions 63 and 70 in HA2, which 146 together prevent conformational changes in the B loop required for fusion ( 27) and are not 147 permissive of S-Flu replication (Figure 2). CLEARFLU version 3 thus contains an intact receptor 148 binding site (Y98), seven fully inactivating mutation sets (R329Q, G1Q, L2G, I6G, 30/47C, 149 212/216C, 63/70P) and the partially attenuating mutation G8A. An antibody panel specific for 150 H7 haemagglutinin (32) confirmed that this combination of mutations did not disrupt protein 151 folding (Figure 3). 152 Genomic stability of CLEARFLU viruses 153 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Viruses expressing H7 A/Hong Kong/125/2017 CLEARFLU version 3 in the hemagglutinin 154 expression cassette were generated by multiplasmid transfection of HEK 293T cells ( 9). These 155 viruses were expanded in MDCK -SIAT1 cells stably expressing H1 A/Puerto Rico/8/1934 156 functional hemagglutinin. 157 CLEARFLU viruses grew to titres >10E6 TCID50/ml / >10E7CID50/ml (see methods) , but 158 while CLEARFLU was strongly displayed on the surface of some infected cells, many viral 159 plaques did not stain with H7 -specific antibody 4A14 (32), which targets the receptor biding site 160 (Figure 4). This suggested that CLEARFLU HA expression had been lost or that the receptor 161 binding site had been disrupted. The loss of full length CLEARFLU during expansion may not 162 be surprising given that CLEARFLU expression is not expected to confer any replicative 163 advantage in these conditions. 164 Previous work in our laboratory has shown that S-FLU viruses grow relatively poorly in MDCK-165 SIAT1 cells expressing Ebola glycoprotein (GP) ( 16). We therefore trialled expanding 166 CLEARFLU viruses in cells expressing GP to determine whether this would select for 167 expression of the intact hemagglutinin receptor binding site in CLEARFLU version 3 by 168 providing an efficient sialic acid binding function. 169 CLEARFLU viruses indeed grew to higher titres and formed denser plaques in MDCK-GP cells 170 than S -FLU viruses, and plaques produced by CLEARFLU viruses which did not stain with 171 hemagglutinin antibody were similar in morphology to small diffuse plaques produced by S-FLU 172 viruses (Figure 5). We tested this apparent selective pressure and genomic stability under these 173 conditions by passaging the virus at low multiplicity of infection (MOI) six times in MDCK-GP 174 cells, representing a 4x10 18 expansion by round six. After six rounds, CLEARFLU viruses still 175 grew to high titre and we were able to detect CLEARFLU expression, however the proportion of 176 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint clones which stained with 4A14 decreased from over 90% to around 10% (Figure 6). 177 Remarkably, sequencing revealed that no mutations were identified in the coding sequences of 178 any core viral genes in the round six culture. 179 In isolating viral clones from the round six culture, we were able to identify several CLEARFLU 180 viruses which replicated efficiently to form dense plaques despite not staining with 4A14. Four 181 of these clones were sequenced and all were found to carry the T148A mutation (N2 numbering; 182 T132A in N1 numbering) in their neuraminidase. The T148A substitution has been previously 183 reported to disrupt a conserved site for N -linked glycosylation in the B1L23 loop 184 [HSNGTVKDR] close to the active site of the enzyme (33). 185 To confirm that this substitution was responsible for the enhanced growth in these conditions, we 186 introduced the T148A mutation alone into S-FLU. W e showed that this mutation enables the 187 neuraminidase to agglutinate red blood cells and enhances cell entry in the GP transduced 188 MDCK-SIAT1 cells in an oseltamivir and neuraminidase antibody sensitive manner ( see S.12 189 and S13 in the s upplementary material). This suggested that T148A enhanced cell entry by 190 enhancing the binding to sialic acid, possibly reducing the selection for expression of 191 CLEARFLU. Sequencing of the CLEARFLU HA segment of two T148A carrying clones 192 revealed a large truncation of 1620 nucleotides, leaving a 438 nucleotide fragment containing 193 intact packaging sequences but only 63 nucleotides of CLEARFLU coding sequence (Figure 7). 194 In another clone which grew well, stained with 4A14 and did not contain the T148A mutation, 195 the CLEARFLU segment remained full-length and contained no mutations. 196 These results showed that expressi ng the CLEARFLU HA can provide a growth advantage 197 (presumably by enhancing sialic acid binding) to viruses propagated in cell lines transduced with 198 Ebola GP to provide a fusion function. However , substitution of NA T148A can also provide 199 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint enhanced sialic acid binding, and viruses with this mutation compete with those expressing 200 CLEARFLU HA. 201 Murine immune response to CLEARFLU viruses 202 As the immune response to S -FLU viruses is known to protect mice, ferrets and pigs against 203 heterologous challenge, we compared the immune response to CLEARFLU and S -FLU viruses 204 in mice. In our first experiment, we assessed the serum antibody response to viruses expressing 205 seasonal H1 (A/England/195/2009) and H3 (A/Hong Kong/5738/2014) CLEARFLU version 2 206 hemagglutinins. In the second, we compared local and systemic, hum oral and cellular immune 207 responses to a virus expressing H7 (A/Hong Kong/125/2017 ) CLEARFLU version 3 208 hemagglutinin to those generated by a strain-matched S-FLU. 209 Like S -FLU, CLEARFLU version 2 and CLEARFLU version 3 viruses can be safely 210 administered to mice both intranasally and intraperitoneally. We found that CLEARFLU viruses 211 offer an advantage over S -FLU viruses in that they produce a strong neutralising antibody 212 response to hemagglutinin even when administered to the airways (Figures 8 & 9). This response 213 was particularly strong for H1 and H7 hemagglutinins compared to H3. For CLEARFLU version 214 3, this neutralising response was strongest in the serum, but also detected in bronchoalveolar 215 lavage (BAL) and, while subtype -specific, showed some within -subtype cross-neutralisation in 216 the serum (Figure 9). The antibody response to the neuraminidase in the BAL and serum was 217 similar to that generated by S-FLU (15). 218 The CLEARFLU version 3 virus generated CD8+ T -cell responses to nucleoprotein in the lungs 219 and spleen comparable to a strain-matched S-FLU, which have previously been shown to protect 220 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint against heterologous - including heterosubtypic – challenge in mice and ferrets and partially 221 protect pigs (Figure 10). 222 Overall, these results indicate that CLEARFLU and S-FLU viruses have very similar 223 immunogenicities in mice, with the notable exception that the expression of CLEARFLU 224 hemagglutinins on the surface of cells is able to generate a strong neutralising antibody response 225 after intranasal delivery at low dose , and may therefore provide sterile immunity against strain -226 matched challenge. 227 228

229 We have described a methodology for designing haemagglutinins inactivated with multiple 230 mutations which can be safely expressed as part of single cycle viruses. We have shown that this 231 approach can be applied to both group 1 (H1) and group 2 (H3 & H7) hemagglutinins without 232 altering expression or antigenicity. In doing so, we have developed a promising vaccine platform 233 for stimulating a combination of broad local T-cell responses and strain-specific neutralising 234 antibodies after intranasal administration . The alternative way to induce neutralising antibody 235 with S-FLU is to combine intranasal administration with a larger dose in the periphery (13,34). A 236 single administration is clearly preferable. 237 CLEARFLU hemagglutinins expressed within the S -FLU expression cassette were displayed on 238 the cell surface and were immunogenic in mice. Although we identified that some CLEARFLU 239 expression is lost over time during expansion of viral stocks, both viral titre and the proportion of 240 viral clones still expressing CLEARFLU remained hi gh for five rounds of expansion, allowing 241 large quantities of virus to be produced without the need for recloning. While after expansion 242 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint some viral clones expressed a truncated CLEARFLU segment, we found no mutations in viral 243 clones with a full -length segment, indicating that the CLEARFLU segment was relatively 244 genetically stable. Truncation of the CLEARFLU segment appeared to be associated with a 245 T148A mutation in the neuraminidase which enabled neuraminidase -dependent sialic acid 246 binding and enhancement of cell entry. CLEARFLU viruses overall appeared to be genetically 247 stable, as no mutations were identified elsewhere in the viral genome. 248 Hemagglutinin-inactivated single cycle influenza viruses like CLEARFLU could be suitable as 249 pre-pandemic as well as seasonal vaccines as they do not encode a viable hemagglutinin that 250 could theoretically reassort with seasonal viruses. The presence of multiple independent 251 mutations scattered throughout the hemagglutinin renders negligible the probability of reversion 252 to pathogenicity and reduces the already low risk of homologous recombination with wildtype 253 hemagglutinins to yield functional new variants (35). As hemagglutinin is a key contributor to 254 antigenic shift, it is particularly beneficial to prevent this segment from acting as a donor 255 sequence for wildtype viruses. Thus an S-FLU expressing a CLEARFLU version 3 H7 256 haemagglutinin would be theoretically safer than a LAIV encoding a viable H7 sequence (36,37) 257 that could in principle reassort with a seasonal influenza. 258 Moreover, as the pseudotyping hemagglutinin is the determinant of cell entry but not the antigen 259 driving neutralising antibody responses after intranasal dosing, pseudotyping cell lines can be 260 selected to avoid pre -existing immunity while maximising sialic acid binding. Meanwhile, the 261 multi-mutated CLEARFLU hemagglutinin can be chosen entirely for its immunogenicity. This 262 could be a key advantage over other approaches like LAIV where pre-existing immunity reduces 263 vaccination efficiency against seasonal influenza. 264 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Unlike for many single cycle viruses which lack all or a large part of a viral gene, the full 265 complement of viral antigens should be expressed in RNA, protein and peptide form in the 266 appropriate cellular location after immunisation with CLEARFLU, maximising immunogenicity. 267 In particular, n eutralising antibody responses to both hemagglutinin and neuraminidase are 268 desirable as they act independently (38) to limit or abolish virus shedding and onward 269 transmission (39), reduce the risk of pathology in individuals with weak T -cell responses and 270 restrict the opportunity for wildtype virus evolution. CLEARFLU therefore represents a clear 271 advantage over S-FLU and other intranasal vaccine candidates which do not induce neutralising 272 antibody. 273 274

and Materials: 275 Experimental models 276 Madin-Darby Canine Kidney -sialyltransferase1 (MDCK-SIAT1) cells generated by transducing 277 MDCK cells with human 2,6 -sialtransferase1 (SIAT1) to express higher levels of sialyl -α2,6-278 galactose moieties ( 31) were obtained from the European Collection of Authenticated Cell 279 Cultures (ECACC 05071502). Human Embryonic Kidney (HEK) 293T cells (HEK cells stably 280 transduced to express SV40 Large T antigen) were obtained from the William Dunn School of 281 Pathology, Oxford (Ervin Fodor). All cells were grown in D10 (10% FCS + 2 mM L-glutamine 282 + 100 units/ml Penicillin + 100 μg/ml Streptomycin in DMEM) at 37°c with 5% C O2 and 283 passaged when confluent. MDCK -SIAT1 cells were harvested by incubating with Trypsin -284 EDTA for 10 minutes and centrifuging in 50ml of D10. 293T cells were harvested using 2mM 285 EDTA in phosphate buffered saline (PBS) and spun in 50ml PBS. Pellets were resuspended in 286 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint D10 for passaging, viral growth medium ( VGM) for viral assays or at about 10 million cells/ml 287 in freezing medium for storage. 288 C57BL/6 and BALB/c female mice were obtained fr om Envigo RMS Inc, Bicester and 289 maintained in individually ventilated cages at the Biomedical Services facility at the John 290 Radcliffe hospital, Oxford . They were 6 -8 weeks old at the start of experiments. Experiments 291 were conducted under Project License number PBA43A2E4 in accordance with the 3Rs. 292 Molecular studies 293 CLEARFLU hemagglutinins were designed by incorporating inactivating mutations into wild -294 type hemagglutinin sequences from A/England/195/2009 (H1), A/Hong Kong/5738/2014 (H3) 295 and A/Hong Kong/125/2017 (H7). DNAs were human codon -optimised. Restriction enzyme 296 cloning sites for NotI and EcoRI were added to either end of the sequence to facilitate cloning 297 into vector pHR-SIN or the S -FLU cassette for mammalian or viral expression respectively. All 298 sequences were ordered from GeneArt and re suspended in 50μl 10mM Tris/0.1mM EDTA. 299 Sequences are available in the supplementary material. 300 For subcloning into pHR-SIN (38) or the S -FLU expression cassette, 2μg of GeneArt plasmid 301 was digested using EcoRI and NotI in NEBuffer 3. 1. Digestion products were run on gels 302 containing 0.7% Agarose and 0.5μg/ml Ethidium Bromide in TAE buffer. Samples were loaded 303 using Orange G gel loading dye. Bands were image d and extracted under UV light and purified 304 using a QIAquick Gel Extraction Kit from Qiagen according to the manufacturer’s protocol . 305 Inserts were ligated into new vectors in a 30μl reaction containing 1.5μl T4 DNA ligase, 3μl T4 306 DNA ligase buffer, 3μl pre -digested and phosphatase treated vector and 15μl of insert. The 307 reactions were incubated at room temperature for at least 1 hour. 308 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint All plasmi ds were grown in DH5α E.coli . 2μl of plasmid was added to 5μl of DH5α 309 ultracompetent cells in Eppendorf tubes on ice. After a 40 -minute incubation, a heat shock was 310 performed for 30 seconds at 42°c , followed by incubation on ice for 1 minute . 1ml of LB was 311 added before incubating at 37°c for 1 hour and then spreading the broth on LB -Agar plates 312 containing ampicillin. After overnight incubation at 37°c, colonies were picked into 13ml LB 313 containing ampicillin for a further overnight incubation at 37°c on a 200rpm shaking platform. 314 0.9ml of the broth was stored at -80°c in Hogness solution for future use. The remaining cells 315 were then pelleted by centrifugation at 3000rpm for 45 minutes and the supernatant removed. 316 Plasmids were extracted from the pellets using a QIAprep Spin Miniprep Kit. To check pHR -317 SIN plasmids for correct insert length, 20μl digests with 2μl plasmid DNA and 0.5μl each of 318 NotI and EcoRI were set up. Phusion PCR (Thermo Scientific) was performed as recommended 319 by the manufacturer. PCR products were run on 0.7% agarose gels containing ethidium bromide 320 as described for digestion products. 321 Agilent QuickChange Lightning Site -Directed Mutagenesis kits were used to introduce specific 322 mutations by PCR into the hemagglutinin sequence of A/Hong Kong/125/2017 to generate 323 versions with a single inactivating mutation set. The sequence was subcloned into vector 324 pcDNA3.1- as described above for mutagenesis, and then subcloned into vector pHR -SIN 325 afterwards using the same restriction sites . The bacteria provided were grown up in the same 326 conditions as described for DH5α E.coli. 327 Transfection and transduction of cell lines 328 Multi-plasmid transfection of HEK 293T cells for generating recombinant viruses was performed 329 using Lipofectamine 2000 as described previously (13,9). 330 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint MDCK-SIAT1 cells were transduced to stably express hemagglutinin by lentivirus transduction. 331 First, HEK 293T cells were transfected with 1.33μg each of plasmids containing VSV -G, 332 Gag/Pol and the gene of interest in a pHR-SIN vector to produce lentivirus (40). After 48 hours, 333 the supernatant was incubated with MDCK -SIAT1 cells in the presence of 8μg/ml polybrene in 334 D10. After 24 hours, the lentivirus infection was repeated and cells were harvested once 335 confluent. 336 Staining cells for fluorescence-activated cell sorting (FACS) 337 Confluent T175 flasks of MDCK -SIAT1 cells were harvested and spun in D10 at 1400rpm. The 338 cell pellet was resuspended in 1ml of 20μg/ml primary antibody in D10 with 10mM HEPES and 339 incubated on ice for 1 hour. Cells were then washed with 50ml of D10 and resuspended in 1ml of 340 20μg/ml secondary antibody and incubated on ice for 1 hour. Cells were again washed with D10, 341 spun and resuspended in 7ml of D10 before being passed through a 0.22μm syringe filter and 342 sorted using a BD FACSAria III. Sorted cells were grown up and passaged in T175 flasks to 343 produce cell stocks for subsequent experiments. 344 Staining cells for analysis by flow cytometry 345 MDCK-SIAT1 cells were aliquoted into FACS tubes at approximately 1 million cells per tube. 346 The tubes were spun at 1400rpm and the cells were resuspended in 50μl of 20μg/ml primary 347 antibody diluted in cold FACS wash (1% FCS + 0.01% Azide in PBS) . After 1 hour at 4˚c, 2ml 348 of FACS wash was added and the cells were spun and resuspended in 50μl of 20μg/ml secondary 349 antibody. After a further hour at 4˚c, the cells were washed in FACS wash and resuspended in 350 300μl of FACS fixative (1% FCS + 1% Formalin in PBS) . Cells were analysed using a CyAn 351 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint ADP flow cytometer by Beckman or an Attune NxT flow cytometer by Invitrogen. Data was 352 analysed using FlowJo v10.6.1. Forward and side scatter were used to gate for single cells only. 353 Staining cells for plated viral assays 354 Confluent MDCK -SIAT1 cells in 96 well plates were stained with 40μl of 1-5μg/ml primary 355 antibody per well for 1 hour at 4˚c, washed thoroughly with PBS and then stained with 40μl of 356 5μg/ml secondary antibody in the same conditions. For staining nucleoprotein, cells were first 357 fixed by incubation with 100μl of 10% formalin for 30 minutes at 4˚c and then incubated with 358 100μl of permeabilization buffer (0.5% Triton X -100) at room temperate for 20 minutes. For 359 stains not involving nucleoprotein, cells were fixed after staining and were not permeabilised. 360 Double stains were performed by incubating with both primary antibodies - one mouse and one 361 human - simultaneously and then both secondary antibodies, as they do not cross -react 362 significantly. Antibodies were diluted in FACS wash. Fixed cells were stored in PBS . Images 363 were taken with a Zeiss fluorescence microscope using 10x magnification. Images were 364 managed using Fiji/Image J. Fluorescence was quantified using a CLARIOstar microplate reader 365 (BMG Labtech). 366 Virus titrations 367 Isolation of viral clones by limiting dilution and quantification by 50% tissue culture infectious 368 dose ( TCID50) was performed as described previously ( 13). Briefly, virus was added to 3E4 369 MDCK-SIAT1 cells in serial half-log dilutions in a 96 -well plate in 200μl volume of VGM with 370 TPCK-trypsin, with 8 replicates for each dilution. After 48 hours, cells were fixed, permeabilised 371 and stained as described above and the dilution at which 50% of wells were infected was 372 calculated using the Reed Muench method (41). 373 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint The cell infectious dose (CID50) was determined by adding 3E4 MDCK -SIAT1 cells to serial 374 two-fold dilutions of virus in duplicate rows in a 96 -well plate without the addition of trypsin . 375 After overnight incubation, the cells were fixed, permeabilised and stained and the dilution of 376 virus at which 50% of cells in a well were infected was determined by linear interpolation. 377 In vivo work 378 Immunisation and sample collection : For intranasal immunisation, mice were anaesthetised 379 using 4.5% isofluorane and drops of virus were pipetted onto the nose to be inhaled one by one 380 to a total volume of 50μl. For intraperitoneal immunisation, 500μl of virus was injected into the 381 intraperitoneal space using a 0.5ml insulin syringe and a 29G needle. Mice were humanely killed 382 by exposure to slowly rising concentration of CO2 and death was confirmed by cervical 383 dislocation. Bronchoalveolar lavage (BAL) was obtained by nicking the trachea and using a 384 syringe to wash 1ml of sterile PBS through the lungs three times. Blood was harvested by 385 cardiac puncture using a 23 -25G needle and a 1 or 2ml syringe and stored in a BD microtainer 386 with SST gel for 30 minutes to allow clotting before spinning at 10000rpm for 5 minutes. Serum 387 was collected from above the interface gel and heat -inactivated at 56˚c for 30 minutes. Spleens 388 were cut into small pieces and passed through a 70μm cell strainer mesh to make single cell 389 suspensions in R10 ( 10% FCS + 2mM L -glutamine + 100 units/ml Penicillin + 100μg/ml 390 Streptomycin in RPMI-1640). Lungs were minced and treated with 4ml of enzyme solution 391 (2mg/ml Collagenase IV + 200 units in RPMI -1640) for 30 minutes at 37˚ before being passed 392 through a 70μm cell strainer mesh. The suspension was spun at 1200rpm for 5 minutes and the 393 cell pellets were t reated with RBC lysis buffer (Qiagen ) for 10 minutes and then washed with 394 R10. The cell pellets were resuspended in 5ml of 40% Percoll in a 15ml tube. A 5ml layer of 395 80% Percoll was laid down at the bottom of the tubes using a Pasteur pipette before spinning the 396 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint cells at 2000rpm for 25 minutes using a slow deceleration rate. Leukocytes were collected from 397 the Percoll interface and washed in R10 twice before being resuspended in R10. 398 Microneutralisation assays were performed as described previously ( 13) with slight 399 modifications. Sera were diluted 1 in 20 before use, BAL samples were not prediluted. For some 400 experiments, sera from animals in the same group were pooled. S -FLU viruses with known 401 neutralisation profiles were used as sources of hemagglutinin and were titrated before use to 402 ensure maximum signal. Briefly, sera or BAL were set up i n serial two -fold dilutions in 50μl 403 PBS in a 96 -well plate and incubated with 50μl S -FLU virus for 2 hours at 37˚c before adding 404 3E4 MDCK-SIAT cells in 100μl VGM. After overnight incubation at 37˚c, eGFP fluorescence 405 was read. 406 Enzyme-linked lectin assay s (ELLA) were performed to assay inhibition of neuraminidase 407 activity by immune sera or BAL (42, 43; modified by 15). S-FLU viruses were used as sources 408 of active neuraminidase and were selected to avoid interference by neutralising antibodies to 409 hemagglutinin. 96 -well ELISA plates were coated in fetuin by incubating in 50μl of fetuin 410 solution overnight at 4˚c. Doubling dilutions of sera or BAL were set up in 60μl, to which 60μl 411 of S-FLU virus was added for incub ation at 37˚c for 2 hours . Fetuin plates were washed four 412 times with PBS before adding 100μl of S-FLU and serum/BAL mix to each well and incubating 413 for 18 hours at 37˚c. After washing four times with PBS, 50μl PNA-HRP (peanut agglutinin 414 conjugated to horseradish peroxidase) solution was added for a 1.5 hour incubation. After 415 washing four times with PBS, 50μl of peroxidase substrate solution (OPD) was added and the 416 plates were developed for approximately 15 minutes before stopping the reaction with 50μl of 417 1M H2SO4. Absorbance at 492 nm was read immediately using a CLARIOstar microplate 418 reader. 419 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Detection of NP specific T cells : Approximately 1 million homogenised splenocytes or lung 420 cells were added to each well of a 96 well plate and washed with R10. MHC class I tetramers 421 (NP366-374 ASNENMETM labelled with APC) were diluted 1 in 200 in R10 and 200μl was added 422 to each well. After 30 minutes at 37˚c, the cells were washed with PBS. 25μl of Zombie Violet 423 was used at 1:1000 in PBS to stain dead cells by incubating in the dark at room temperature for 424 10 minutes. 25μl of an antibody mix was added to each well to simultaneously stain for surface 425 markers CD8 (APC-Cy7 53 -6.7), CD44 (Alexa Fluor 488 IM7) , Cd11b (Brilliant Violet 421 426 M1/70) and B220 (Brilliant Violet 421 RA3-6B2), with each antibody diluted 1 in 200 in FACS 427 buffer. After 30 minutes on ice in the dark, the cells were washed with PBS and fixed using 428 100μl of cold 5% formalin for 10 minutes at 4˚c, and then washed again and resuspended in PBS. 429 CD8+ T -cells positive for MHC tetramer were analysed using an Attune NxT flow cytometer 430 (Thermo Fisher Scientific) and FlowJo v10.6.1. Ultracompensation beads were used to account 431 for overlapping emission spectra. 432 Data availability: Data generated in this study are available from the corresponding author on 433 request. The products generated can be made available to other laboratories for research use. 434 435

436 This research was funded by the Biotechnology and Biosciences Research Council Doctoral 437 Training Partnership (BBSRC DTP) at the University of Oxford (Grant number: BB/M011224/1) 438 439 440 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint

441 1. Dudek, T., & Knipe, D. M. (2006). Replication-defective viruses as vaccines and vaccine 442 vectors. Virology , 344 , 230–239. https://doi.org/10.1016/j.virol.2005.09.020 443 2. Nogales, A., Baker, S. F., Domm, W., & Martínez-Sobrido, L. (2016). Development and 444 applications of single-cycle infectious influenza A virus (sciIAV). Virus Research , 216 , 26–445 40. https://doi.org/10.1016/j.virusres.2015.07.013 446 3. Eiden, J., Volckaert, B., Rudenko, O., Aitchison, R., Herber, R., Belshe, R., Greenberg, H., 447 Coelingh, K., Marshall, D., Kawaoka, Y., Neumann, G., Bilsel, P. (2022) M2-deficient single-448 replication influenza vaccine–induced immune responses associated with protection against 449 human challenge with highly drifted H3N2 influenza strain. The Journal of Infectious Diseases, 450 226 (1), 83–90. https://doi.org/10.1093/infdis/jiab374 451 4. Baz, M., Boonnak, K., Paskel, M., Santos, C., Powell, T., Townsend, A., & Subbarao, K. 452 (2015). Nonreplicating influenza A virus vaccines confer broad protection against lethal 453 challenge. MBio , 6 (5). https://doi.org/10.1128/mBio.01487-15 454 5. Mueller, S. N., & Mackay, L. K. (2016). Tissue-resident memory T cells: Local specialists in 455 immune defence. Nature Reviews Immunology , 16 (2), 79–456 89. https://doi.org/10.1038/nri.2015.3 457 6. Zheng, M. Z. M., Wakim, L.M. (2022) Tissue resident memory T cells in the respiratory 458 tract. Mucosal Immunology, 15, 379–388. https://doi.org/10.1038/s41385-021-00461-z 459 7. Zheng, M. Z. M., Tan T. K., Villalon-Letelier, F., Lau, H., Deng, Y.-M., Frtizlar, S., A 460 Valkenburg, S., Gu, H., Poon, L. L. M., Reading, P. C., Townsend, A. R. M.,Wakim, L.M. 461 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint (2023). Single-cycle influenza virus vaccine generates lung CD8+ Trm that cross-react against 462 viral variants and subvert virus escape mutants. Science Advances, 9(36):eadg3469. doi: 463 10.1126/sciadv.adg3469. 464 8. van Riet, E., Ainai, A., Suzuki, T., & Hasegawa, H. (2012). Mucosal IgA responses in 465 influenza virus infections; thoughts for vaccine design. Vaccine, 30 (40), 5893–466 5900. https://doi.org/10.1016/j.vaccine.2012.04.109 467 9. Fodor, E., Devenish, L., Engelhardt, O. G., Palese, P., Brownlee, G. G., & Garcίa, A. (1999). 468 Rescue of influenza A virus from recombinant DNA. Journal of Virology, 73 (11), 9679–9682. 469 10. Neumann, G., Watanabe T., Ito, H., Watanabe, S., Goto, H., Gao, P., Hughes, M., Perez, 470 D.R., Donis, R., Hoffmann, E., Hobom, G., Kawaoka, Y. (1999). Generation of influenza A 471 viruses entirely from cloned cDNAs. Proceedings of the National Academy of Sciences, 96 (16), 472 9345-9350. doi: 10.1073/pnas.96.16.9345 473 11. Schickli, J. H., Flandorfer, A., Nakaya, T., Martinez-Sobrido, L., Garcia-Sastre, A., Palese, P. 474 (2001). Plasmid-only rescue of influenza A virus vaccine candidates. Philosophical Transactions 475 of the Royal Society London B Biological Sciences, 356(1416): 1965–476 1973. doi: 10.1098/rstb.2001.0979 477 12. Kadasamy M., Gileadi U., Rijal, P., Tan, T.K., Lee, L. N., Chen, J., Prota, G., Klenerman P., 478 Townsend, A., Cerundolo, V. (2023) Recombinant single-cycle influenza virus with 479 exchangeable pseudotypes allows repeated immunization to augment anti-tumour immunity with 480 immune checkpoint inhibitors. eLife 12:e76414. https://doi.org/10.7554/eLife.76414 481 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint 13. Powell, T. J., Silk, J. D., Sharps, J., Fodor, E., & Townsend, A. R. M. (2012). Pseudotyped 482 influenza A virus as a vaccine for the induction of heterotypic immunity. Journal of Virology, 86 483 (24), 13397–13406. https://doi.org/10.1128/jvi.01820-12 484 14. MacLean, A., Richmond, N., Koneva, L., Attar, M., Medina, C. A. P., Thornton, E. E., 485 Gomes, A. C., El-Turabi, A., Bachmann, M. F., Rijal, P., Tan, T. K., Townsend, A., Sansom, S. 486 N., Bannard, O. & Arnon, T. I. (2022). Secondary influenza challenge triggers resident memory 487 B cell migration and rapid relocation to boost antibody secretion at infected sites. Immunity 55, 488 718–733. https://doi.org/10.1016/j.immuni.2022.03.003 489 15. Powell, Timothy J., Rijal, P., McEwen-Smith, R. M., Byun, H., Hardwick, M., Schimanski, 490 L. M., Huang, K. Y. A., Daniels, R. S., & Townsend, A. R. M. (2019). A single cycle influenza 491 virus coated in H7 haemagglutinin generates neutralizing antibody responses to haemagglutinin 492 and neuraminidase glycoproteins and protection from heterotypic challenge. Journal of General 493 Virology , 100 (3), 431–445. https://doi.org/10.1099/jgv.0.001228 494 16. Xiao, J., Rijal, P., Schimanski, L., Tharkeshwar, A. K., Wright, E., Annaert, W., & 495 Townsend, A. (2018). Characterization of an influenza virus pseudotyped with Ebolavirus 496 glycoprotein. Journal of Virology , JVI.00941-17. https://doi.org/10.1128/jvi.00941-17 497 17. Rijal, P., Elias, S. C., Machado, S. R., Xiao, J., Schimanski, L., O'Dowd, V., Baker, T., 498 Barry, E., Mendelsohn, S. C., Cherry, C. J., Jin, J., Labbé, G. M., Donnellan, F. R., Rampling, T., 499 Dowall, S., Rayner, E., Findlay-Wilson, S., Carroll, M., Guo, J., Xu, X.-N.; Huang, K.-Y. A., 500 Takada, A., Burgess, G., McMillan, D., Popplewell, A., Lightwood, D. J., Draper, S. J., 501 Townsend, A. R. (2019). Therapeutic monoclonal antibodies for Ebola virus infection derived 502 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint from vaccinated humans. Cell Reports, ISSN: 2211-1247, Vol: 27, Issue: 1, Page: 172-186.e7. 503 https://doi.org/10.1016/j.celrep.2019.03.020 504 18. Holzer, B., Morgan, S. B., Matsuoka, Y., Edmans, M., Salguero, F. J., Everett, H., Brookes, 505 S. M., Porter, E., MacLoughlin, R., Charleston, B., Subbarao, K., Townsend, A., & Tchilian, E. 506 (2018). Comparison of heterosubtypic protection in ferrets and pigs induced by a single-cycle 507 influenza vaccine. The Journal of Immunology, 200 (12), 4068–508 4077. https://doi.org/10.4049/jimmunol.1800142 509 19. Morgan, S. B., Hemmink, J. D., Porter, E., Harley, R., Shelton, H., Aramouni, M., Everett, 510 H. E., Brookes, S. M., Bailey, M., Townsend, A. M., Charleston, B., & Tchilian, E. (2016). 511 Aerosol delivery of a candidate universal influenza vaccine reduces viral load in pigs challenged 512 with pandemic H1N1 virus. The Journal of Immunology, 196 (12), 5014–513 5023. https://doi.org/10.4049/jimmunol.1502632 514 20. Orlich, M., Linder, D., & Rott, R. (1995). Trypsin-resistant protease activation mutants of an 515 influenza virus. Journal of General Virology, 76 , 625–633. 516 21. Chen, J., Lee, K. H., Steinhauer, D. A., Stevens, D. J., Skehel, J. J., & Wiley, D. C. (1998). 517 Structure of the hemagglutinin precursor cleavage site, a determinant of influenza pathogenicity 518 and the origin of the labile conformation. Cell, 95(3):409-17. doi: 10.1016/s0092-519 8674(00)81771-7. 520 22. Godley, L., Pfeifer, J., Steinhauer, D., Ely, B., Shaw, G., Kaufmann, R., Suchanek, E., Pabo, 521 C., Skehel, J. J., Wiley, D. C., & Wharton, S. (1992). Introduction of intersubunit disulfide bonds 522 in the membrane-distal region of the influenza hemagglutinin abolishes membrane fusion 523 activity. Cell, 68, 635–645. 524 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint 23. Lee, P. S., Zhu, X., Yu, W., & Wilson, I. A. (2015). Design and structure of an engineered 525 disulfide-stabilized influenza virus hemagglutinin trimer. Journal of Virology, 89 (14), 7417–526 7420. https://doi.org/10.1128/jvi.00808-15 527 24. Skehel, John J, & Wiley, D. C. (2000). Receptor binding and membrane fusion in virus entry: 528 The influenza hemagglutinin. Annual Reviews in Biochemistry, 69 , 531–529 569. www.annualreviews.org 530 25. Meisner, J., Szretter, K. J., Bradley, K. C., Langley, W. A., Li, Z.-N., Lee, B.-J., Thoennes, 531 S., Martin, J., Skehel, J. J., Russell, R. J., Katz, J. M., & Steinhauer, D. A. (2008). Infectivity 532 studies of influenza virus hemagglutinin receptor binding site mutants in mice. Journal of 533 Virology, 82 (10), 5079–5083. https://doi.org/10.1128/jvi.01958-07 534 26. Cross, K. J., Langley, W. A., Russell, R. J., Skehel, J. J., & Steinhauer, D. A. (2009). 535 Composition and functions of the influenza fusion peptide. Protein & Peptide Letters, 16 , 766–536 778. 537 27. Gruenke, J. A., Armstrong, R. T., Newcomb, W. W., Brown, J. C., & White, J. M. (2002). 538 New insights into the spring-loaded conformational change of influenza virus hemagglutinin. 539 Journal of Virology, 76 (9), 4456–4466. https://doi.org/10.1128/jvi.76.9.4456-4466.2002 540 28. Huang, K.-Y. A., Rijal, P., Schimanski, L., Powell, T. J., Lin, T.-Y., McCauley, J. W., 541 Daniels, R. S., Townsend, A. R. (2015). Focused antibody response to influenza linked to 542 antigenic drift. The Journal of Clinical Investigation, 125 (7), 2631-2645. 543 https://doi.org/10.1172/JCI81104. 544 29. Kallewaard, N.L., Corti, D., Collings, P. J., Neu, U., McAuliffe, J. M., Benjamin, E., 545 Wachter-Rosati, L., Palmer-Hill, F. J., Yuan, A. Q., Walker, P. A., Vorlaender, M. K., Bianchi, 546 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint S., Guarino, B., De Marco, A., Vanzetta, F., Agatic, G., Foglierini, M., Pinna, D., Fernandez-547 Rodriguez, B., Fruehwirth, A., Silacci, C., Ogrodowicz, R. W., Martin, S. R., Sallusto, F., 548 Suzich, J. A., Lanzavecchia, A., Zhu, Q., Gamblin, S. J., Skehel, J. J. (2016). Structure and 549 function analysis of an antibody recognizing all influenza A subtypes. Cell, 166 (3), 596-608. 550 https://doi.org/10.1016/j.cell.2016.05.073 551 30. Martin, J., Wharton, S. A., Lin, Y. P., Takemoto, D. K., Skehel, J. J., Wiley, D. C., & 552 Steinhauer, D. A. (1998). Studies of the binding properties of influenza hemagglutinin receptor-553 site mutants. Virology, 241 , 101–111. 554 31. Matrosovich, M., Matrosovich, T., Carr, J., Roberts, N. A., & Klenk, H.-D. (2003). 555 Overexpression of the α-2,6-sialyltransferase in MDCK cells increases influenza virus sensitivity 556 to neuraminidase inhibitors. Journal of Virology, 77 (15), 8418–557 8425. https://doi.org/10.1128/jvi.77.15.8418-8425.2003 558 32. Huang, KY.A., Rijal, P., Jiang, H., Wang, B., Schimanski, L., Dong, T., Liu., Y.-M., Chang, 559 P., Iqbal, M., Wang M.-C., Chen, Z., Song, R., Chong, C.-C., Yang, J.-H., Qi, J., Li, T.-Y., Li, 560 A., Powell, T. J., Jan., J.-T., Ma, C., Gao., G., Shi, Y., Townsend, A. R. (2019). Structure–561 function analysis of neutralizing antibodies to H7N9 influenza from naturally infected 562 humans. Nature Microbiology 4, 306–315 https://doi.org/10.1038/s41564-018-0303-7 563 33. Varghese, J., Laver, W. & Colman, P. (1983). Structure of the influenza virus glycoprotein 564 antigen neuraminidase at 2.9 Å resolution. Nature 303, 35–40. https://doi.org/10.1038/303035a0 565 34. Martini, V., Paudyal B., Chrun, T., McNee, A., Edmans, M., Atangana Maza, E., Clark, B., 566 Nunez, A., Dolton, G., Sewell, A., Beverley, P., MacLoughlin, R., Townsend, A., Tchilian, E. 567 (2021) Simultaneous aerosol and intramuscular immunization with influenza vaccine induces 568 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint powerful protective local T cell and systemic antibody immune responses in pigs. Journal of 569 Immunology, 206 (3), 652-663. doi: 10.4049/jimmunol.2001086. 570 35. Boni, M. F., Zhou, Y., Taubenberger, J. K., & Holmes, E. C. (2008). Homologous 571 recombination is very rare or absent in human influenza A virus. Journal of Virology, 82 (10), 572 4807–4811. https://doi.org/10.1128/jvi.02683-07 573 36. Sobhanie, M., Matsuoka, Y., Jegaskanda, S., Fitzgerald, T., Mallory, R., Chen, Z., Luke, C., 574 Treanor, J., Subbarao, K. (2016). Evaluation of the safety and immunogenicity of a candidate 575 Pandemic Live Attenuated Influenza Vaccine (pLAIV) against influenza A(H7N9). The Journal 576 of Infectious Diseases, 213 (6), 922–929. https://doi.org/10.1093/infdis/jiv526 577 37. Chen, Z., Baz, M., Lu, J., Paskel, M., Santos, C., Subbarao, K., Jin, H., Matsuoka, Y. (2014). 578 Development of a high-yield live attenuated H7N9 influenza virus vaccine that provides 579 protection against homologous and heterologous H7 wild-type viruses in ferrets. Journal of 580 virology, 88(12):7016-23. doi: 10.1128/JVI.00100-14 581 38. Memoli, M. J., Shaw, P., A., Han, A., Czajowski, L., Reed, S., Athoha, R., Brsitol, T., Fargis, 582 S., Risos, K., Powers, J. H., Davey R. D. Jr, Taubenberger, J. K. (2016). Evaluation of 583 antihemagglutinin and antineuraminidase antibodies as correlates of protection in an influenza 584 A/H1N1 virus healthy human challenge model. mBio, 7(2): e00417-16. 585 doi: 10.1128/mBio.00417-16 586 39. Tan, J., O’Dell, G., Hernandez, M. M., Sordillo, E. M., Kahn, Z., Kriti, D., van Bakel, H., 587 Ellebedy, A. H., Wilson, P., C., Simon, V., Krammer, F., McMahon, M. (2022). Human Anti-588 neuraminidase Antibodies Reduce Airborne Transmission of Clinical Influenza Virus Isolates in 589 the Guinea Pig Model. Journal of Virology, 96(2): e01421-21. doi: 10.1128/JVI.01421-21 590 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint 40. Demaison, C., Parsley, K., Brouns, G., Scherr, M., Battmer, K., Kinnon, C., Grez, M., & 591 Thrasher, A. J. (2002). High-level transduction and gene expression in hematopoietic 592 repopulating cells using a human immunodeficiency virus type 1 based lentiviral vector 593 containing an internal spleen focus forming virus promoter. Human Gene Therapy, 13 , 803–813. 594 41. Reed, L. J., & Muench, H. (1938). A simple method for estimating fifty per cent endpoints. 595 The American Journal of Hygiene, 27 (3), 493–497. 596 42. Couzens L., Gao J., Westgeest K., Sandbulte M., Lugovtsev V., Fouchier R., Eichelberger M. 597 (2014). An optimized enzyme-linked lectin assay to measure influenza A virus neuraminidase 598 inhibition antibody titers in human sera. Journal of Virological Methods, 210, 7-14. doi: 599 10.1016/j.jviromet.2014.09.003. 600 43. Lambré C. R.,Terzidis H.,Greffard A., Webster R. G. (1991). An enzyme-linked lectin assay 601 for sialidase. Clinica Chimica Acta, 198 (3), 183-193. doi: 10.1016/0009-8981(91)90352-d. 602 44. Burke, D. F., & Smith, D. J. (2014). A recommended numbering scheme for influenza A HA 603 subtypes. PLoS ONE , 9 (11). https://doi.org/10.1371/journal.pone.0112302 604 605 Figures 606 Figure 1 – Antibody binding panel to CLEARFLU versions 1 & 2 expressed in stably 607 transduced MDCK-SIAT1 cells. 608 MDCK-SIAT1 cells were transduced to express H1 A/England/195/2009 and H3 A/Hong 609 Kong/5738/2014 CLEARFLU and wildtype hemagglutinins and sorted for high expression with 610 antibodies T3 -3C (28) (H1) or AM3C (H3) (produced in -house, manuscript in preparation). 611 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Hemagglutinins were bound with head or stem -targeted human primary antibodies, and binding 612 was detected with a FIT-C labelled goat anti-human IgG secondary antibody. 613 A. Bar graph showing median fluorescence. The average median fluorescence for three 614 populations of unstained cells which were not transduced to express hemagglutinin is shown as 615 the no-HA control. 616 B. Histogram plots showing the relative frequency of cells detected with a given fluorescence 617 intensity. Grey plots show unstained cells which were not transduced to express hemagglutinin. 618 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint 619 620 621 Figure 2 – Replication of S-FLU virus in cell lines expressing hemagglutinin with a single 622 CLEARFLU candidate mutation 623 MDCK-SIAT1 cells were transduced to express the selected mutant H7 A/Hong Kong/125/2017 624 hemagglutinin and sorted for high expression with H7 antibody 4A14 (Huang 2019). Cells were 625 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint infected with a n eGFP-expressing S -FLU virus pseudotyped with functional hemagglutinin in 626 half-log dilutions, starting at 1000 TCID50 per well. Eight replicates were used at each dilution. 627 Bars show the fluorescence of each well in a 96 -well plate after 48 hours. Fluorescence 628 microscopy images show cells in dilution 1 expressing eGFP after S-FLU infection, with clusters 629 of infected cells indicating virus replication. The experiment was split into three sets, each of 630 which was repeated with the same results. Virus used for infection : H7N1 S-FLU ([S-631 eGFP/N1(A/Puerto Rico/8/1934)] coated in H7(A/Netherlands/219/2003)) 632 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint 633 634 635 636 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Figure 3 – Antibody binding to CLEARFLU version 3 expressed in stably transduced 637 MDCK-SIAT1 cells 638 MDCK-SIAT1 cells were stably transduced to express H7 A/Hong Kong/125/2017 CLEARFLU 639 version 3 hemagglutinin and sorted for high expression with H7 antibody 4A14. Hemagglutinins 640 were bound with head or stem -targeted human primary antibodies, and binding was detected 641 with an Alexa Fluor 647 labelled secondary antibody. 642 A. Bar graph showing median fluorescence. 643 B. Histogram plots showing the relative frequency of cells detected with a given fluorescence 644 intensity. 645 646 647 648 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Figure 4 – Titration and hemagglutinin expression of CLEARFLU version 3 viruses 649 A, B & C. Two aliquots of CLEARFLU version 3 virus reference stock were titrated by CID50 650 and TCID50 using antibody 2 -8C (Powell et al., 2012) specific to nucleoprotein and an Alexa 651 Fluor 647 labelled secondary antibody. Error bars show the standard deviation around the mean 652 from duplicates. Plot B is representative of two titrations for each aliquot, the results of which 653 are presented in C. 654 D. Fluorescence microscopy images show the expansion of CLEARFLU version 3 virus clones 655 in MDCK -SIAT1 cells transduced to express PR8 hemagglutinin. Cells were infected as per 656 TCID50 protocol but were stained after 24 hours with primary antibodies 2 -8C (biotin -657 conjugated) and 4A14 and then with secondary layers Alexa Fluor 647 streptavidin and goat 658 anti-human Alexa Fluor 488. Clones were defined as clusters of nucleoprotein -stained cells, 659 often with comet-like trails. 660 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint 661 662 663 664 665 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Figure 5 – Plaque morphology of CLEARFLU version 3 viruses grown in cells expressing 666 Ebola glycoprotein (GP). 667 Cells were infected with S-FLU or CLEARFLU version 3 virus as per TCID50 protocol and 668 stained after 48 hours with anti-nucleoprotein mouse antibody AA5H and an Alexa Fluor 647 669 secondary antibody (red). Cells infected with CLEARFLU version 3 virus were also stained with 670 H7 hemagglutinin human antibody 4A14 and an Alexa Fluor 488 secondary antibody (green). 671 Fluorescence microscopy images were taken of plaques indicative of clonal expansion. 672 CLEARFLU virus plaques which did not stain with 4A14 are indicated with a white dash. 673 Viruses: H7N1 S-FLU ([S-eGFP/N1(A/Puerto Rico/8/1934)] coated in H7 A/Hong 674 Kong/125/2014) & H7 A/Hong Kong 125/2014 CLEARFLU version 3 reference stock. 675 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint 676 677 678 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Figure 6 – Proportion of clones expressing CLEARFLU version 3 hemagglutinin after 679 passage at low MOI in MDCK-SIAT1 cells expressing Ebola glycoprotein 680 CLEARFLU version 3 reference stock was used to infect a large flask of MDCK -SIAT1 cells 681 expressing Ebola GP at a multiplicity of infection of around 0.01 by TCID50, and then passaged 682 at a 1 in 1000 dilution every 48 hours six times. Aliquots from each round were used to infect 683 MDCK-SIAT1 cells expressing GP in limiting dilution, and viral clones were identified by 684 nucleoprotein staining with AA5H (Alexa Fluor 647 secondary). Clones were scored as positive 685 for CLEARFLU expression if they also stained with 4A14 (Alexa Fluor 488 secondary). Each 686 bar shows an independent count. Error bars show 95% confidence intervals calculated using the 687 Wilson/Brown method for proportions. 688 689 690 691 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Figure 7 – Design of the CLEARFLU version 3 viral expression cassette and truncation 692 after repeated passage in MDCK-SIAT1 cells expressing Ebola glycoprotein. 693 The CLEARFLU expression cassette contains intact packaging signals from the S-FLU 694 expression cassette, which in this design differs from that described in Powell et al. 2012 by the 695 alteration of two additional interfering ATG codons upstream of the NotI site. After six rounds of 696 passage at low multiplicity of infection in MDCK-SIAT1 cells expressing Ebola glycoprotein, 697 two viral clones expressed a truncated CLEARFLU construct containing only 63 base pairs of 698 coding sequence flanked by intact PR8 untranslated regions (UTRs). 699 700 701 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Figure 8 – Murine serum antibody response following immunisation with CLEARFLU 702 version 2 703 Serum antibody response in mice immunised with H1 or H3 CLEARFLU version 2 viruses 704 pseudotyped with H5 or H7 coats. Mice were immunised with 2E6 CID50 virus intranasally 705 (IN), or 1E6 CID50 intranasally and 1E7 CID50 intraperi toneally (IP) twice 28 days apart. Sera 706 were harvested 14 days after the second dose for characterisation by microneutralisation (A) and 707 enzyme linked lectin assay (B). Sera from 5 mice were pooled for each group, except the viral 708 growth medium (VGM) only control group, for which 4 mice were used. Error bars show the 709 standard deviation from within -assay duplicates, with results from two independent assays 710 shown. 711 A. Neutralising antibody response to H1, H3, H5 and H7 S-FLU by immune mouse sera or 712 control antibody MEDI8852 713 B. Inhibition of N1 neuraminidase activity by mouse immune sera or control antibody Z2B3. 714 715 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Figure 9 – Murine serum and bronchoalveolar lavage (BAL) antibody response following 716 immunisation with CLEARFLU version 3 717 Serum and BAL antibody response in mice immunised with H7 CLEARFLU version 3 718 pseudotyped with an H1 (PR8) coat or a strain -matched H7 S-FLU. Mice a-d in each group were 719 immunised twice 14 days apart with 1E5 TCID50 virus intranasally (IN), 1E6 TCID50 720 intraperitoneally (IP) or 1E5 TCID50 IN and 1E6 TCID50 IP. Sera and BAL were harvested 68 721 days after the second dose for characterisation by enzyme -linked lectin assay (A) and 722 microneutralisation (B). Error bars show the standard deviation from within -assay duplicates 723 when both duplicates fell above the detection threshold. 724 A. Inhibition of N1 neuraminidase activity by mouse immune sera, BAL or control antibody 725 Z2B3. Sera from the four mice were pooled for each group. BAL could not be collected from 726 mouse b in the CLEARFLU IN+IP group, mouse a in the S-FLU IN+IP group or mouse a in the 727 VGM group. 728 B. Neutralising antibody response to H7 S-FLU by immune mouse sera, BAL or control 729 antibody MEDI8852. The top two plots show neutralisation of strain -matched A/Hong 730 Kong/225/2017 S-FLU by individual mouse sera or BAL. BAL could not be collected from 731 mouse a in the S-FLU IN+IP group or mouse a in the VGM control group. The bottom plot 732 shows neutralisation of unmatched H7 A/Netherlands/219/2003 and A/Taiwan/1/2017 S-FLU by 733 pooled immune sera. 734 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint 735 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Figure 10 – Murine T-cell response to CLEARFLU version 3 736 Mice were immunised as described in Figure 9. Circles indicate the percentage of CD8+ T -cells 737 positive for MHC tetramer in the lungs and spleen of each mouse a-d in each group. Bars show 738 the mean and standard deviation for each group. For spleens, data could not be collected for 1 739 individual in the CLEARFLU IN, CEARFLU IP, CLEARFLU IN+IP and S-FLU IP groups and 740 for 2 individuals in the S -FLU IN+IP and VGM control groups. Multiple t tests were used to 741 compare CLEARFLU and S-FLU groups for each administration method. The Holm -Sidak 742

was used to correct for multiple comparisons. All differences were non -significant at the 743 p=0.05 level except for IP groups in the spleen where p=0.0078. 744 745 746 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint Tables 747 Table 1 Mutations introduced to inactivate wild -type hemagglutinin genes in each version of the 748 CLEARFLU design 749 Mutation type Mutation position Rationale CLEARFLU version (H3 numbering system (41)) 1 2 3 Cleavage resistance HA1-R329Q Cleavage of the precursor HA0 into HA1 and HA2 by trypsin is necessary for conformational changes required for cell entry. Viruses with mutations at the cleavage site cannot replicate due to changes in protease sensitivity (20,21). ✓ ✓ ✓ Locked by inter- monomer disulphide bonds Head: HA1-212C & HA1-216C The introduction of inter -chain disulphide bridges between head regions prevents membrane fusion for cell entry ( 22). Disulphide bridges between stem regions increase the stability of the hemagglutinin trimer and may prevent fusion (23). ✓ ✓ ✓ Stem: HA1-30C & HA2-47C ✓ ✓ ✓ Receptor binding abolished HA1-Y98F The change from tyrosine to phenylalanine prevents the formation of a hydrogen bond with sialic acid residues necessary for optimum binding ( 24) and attenuates influenza in mice, but has a high rate of back - mutation (25). ✓ ✓ Fusion peptide inhibited HA2-G1Q HA2-L2G Mutations in the fusion peptide abolish fusion. The residues which are most highly conserved are most likely to block function when mutated (26). ✓ ✓ ✓ HA2-W14A HA2-W21A ✓ HA2-I6G HA2-G8A ✓ ✓ B Loop inactivation by proline residues HA2-F63P & HA2-F70P Introducing two proline residues in the B loop prevents conformational changes needed for fusion (27). ✓ 750 .CC-BY-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted March 4, 2024. ; https://doi.org/10.1101/2024.03.01.582898doi: bioRxiv preprint