Homologous PB1 gene promotes the replication efficiency of avian influenza H7N4 candidate vaccine virus

Abstract Background The first and only case of human infection with the avian influenza A (H7N4) virus in China emerged in 2018. The H7N4 virus was distinct from previous H7N9 viruses and raised public concerns. Therefore, developing a suitable H7N4 candidate vaccine virus (CVV) remains crucial for potential pandemic preparedness. Methods We constructed a reassortant virus with a (6 + 2) genome composition, then introduced the polymerase basic protein 1 (PB1) from a wild‐type virus to develop a (5 + 3) reassortant virus through reverse genetics. We performed whole‐genome sequencing to confirm the genome stability, assessed the growth ability in MDCK cells, and analyzed virus antigenicity using hemagglutination inhibition assays. Subsequently, the effect of homologous PB1 on polymerase activity, viral protein yield, and pathogenicity was assessed. Results The (5 + 3) virus harbouring the homologous PB1 gene exhibited significantly improved growth characteristics, higher viral protein yield, and polymerase activity than the (6 + 2) virus. After successive passage in embryonated eggs, glutamic acid (E) substituted glycine(G) at position 218 (H3 numbering) in the hemagglutinin (HA) gene of both (5 + 3) and (6 + 2) viruses. The substitution improved the growth of the (6 + 2) virus but exhibited no significant effect or alteration on the antigenicity of the (5 + 3) virus. Moreover, the (5 + 3) virus exhibited low pathogenicity in chickens and ferrets. Conclusion Homologous PB1 of the H7N4 virus improves the growth ability while sustaining low pathogenicity. Collectively, the gene composition of the (5 + 3) reassortant virus is a suitable H7N4 CVV for potential pandemic preparedness.


| INTRODUCTION
Influenza viruses are highly contagious respiratory pathogens that threaten public health worldwide. Besides seasonal epidemics, influenza A viruses potentially influence the emergence of future pandemics whenever novel or avian influenza virus develops efficient human-to-human transmission and spreads globally. The first recognized human infection of highly pathogenic avian influenza A (H5N1) virus, previously confined to poultry, emerged in 1997 during a poultry outbreak in Hong Kong. 1 Since then, various avian influenza virus strains have infected humans, including H5N6, H7N9, H9N2, and H10N8, causing significant public health concerns. 2 In 2018, the first case of human infection with avian influenza A (H7N4) virus, causing severe pneumonia and acute respiratory distress, emerged in China. 3 Genome assembly showed 8 consensus gene segments of H7N4 from the throat swab (designated A/Jiangsu/1/2018, designated JS-Hu in this article), but the virus was not successfully isolated from the patient samples. 4 The JS-Hu genes are closely related to the influenza virus, A/Chicken/Jiangsu/1/2018 (JS-Ck, H7N4), which were eventually isolated from chickens fed by the patient. 4 However, the H7N4 viruses were genetically and antigenically distinct from previously reported H7N9 viruses, which caused persistent and severe human infections in previous years. 4 This prevalence of avian influenza viruses has raised great concerns regarding potential pandemics and warrants the developing corresponding vaccine strains for pandemic preparedness. As part of an influenza pandemic preparedness program for the novel H7N4 virus, the World Health Organization (WHO) recommended the JS-Ck-like virus as a candidate vaccine virus (CVV) based on the current antigenic, genetic, and epidemiologic data. 5 Vaccines are the most potent and effective tool for public health.
Owing to the lack of universal influenza vaccines, CVV development for specific influenza strains remains crucial. Thus, selecting high-yield CVVs is critical for large-scale pre-pandemic vaccine production. The reverse genetics (RG) method accurately controls the genetic composition of the reassortant viruses by an eight plasmid-based rescue system. This technique yields viral (6 + 2) genotype reassortant 'seeds' comprising the hemagglutinin (HA) and neuraminidase (NA) surface genes derived from wild-type (WT) epidemic viruses while other six "internal genes" are from the high yield A/Puerto Rico/ 8/1934(PR8) virus. However, previous studies showed that the frequency of high yield reassortant viruses derived from the WT homologous polymerase basic protein 1 (PB1) gene is high, and the PB1 segment from WT virus might enhance the replication efficiency of CVVs. [6][7][8] These reassortant viruses containing HA, NA, and PB1 from the WT combined with other genes from PR8, were designated as (5 + 3) reassortant viruses for selecting high yield CVVs. 9 In contrast, some studies reported that PB1 neither drives selecting the suitable virus nor selects the best yielding virus for vaccine production. 10,11 Moreover, PB1 affects the replication capacity and pathogenicity of the reassortant virus. 12,13 An effective CVV should have high replication capacity and antigenic consistency with the prototype strain, with low pathogenicity. 14 This study constructed two viral reassortants with (6 + 2) and (5 + 3) genomes generated from the JS-Ck virus using RG. The purpose was to develop a suitable H7N4 CVV that meets vaccine requirements.
The only difference between (6 + 2) and (5 + 3) viruses was the PB1 gene segment from PR8 or JS-Ck H7N4. We performed serial assessments to understand the effect of homologous PB1 on H7N4 CVV. This is the first known report on the development and evaluation of CVVs for H7N4 subtype influenza virus.

| Stability of reassortant viruses
For the stability test, viruses were consecutively passaged six times through specific-pathogen-free (SPF) embryonated eggs following the WHO guidelines. 14 Each generation virus was stored at À80 C until further use. All viruses of each generation were sequenced by NGS and analyzed using the BioEdit (Version 7.1.3.0) software.

| Growth kinetics assay
MDCK cells were used to investigate growth characteristics of different viruses as described previously. 15

| Assay for polymerase activity
The polymerase activity was measured by minigenome transfection and virus infection, as previously described, with slight modifications. 13

| Plaque formation assay
MDCK cells were used to determine the plaque formation, and cells were infected with 0.01 MOI. The assay was performed according to our previous report. 17

| Hemagglutination assay and hemagglutination inhibition (HI) assay
Hemagglutination and HI assays were performed as described in the WHO Manual for Laboratory Diagnosis and Virological Surveillance of Influenza using 1% turkey erythrocytes. 18

| Purification and quantification of viral proteins
Viruses were purified from allantoic fluids through sucrose (Solarbio, Beijing, China) gradient centrifugation. 17 Total protein was determined by a BCA protein assay kit (Thermo Fisher Scientific, MA, USA) as previously described. 17 The protein was digested using PNGase F For antisera preparation, ferrets were infected with reassortant viruses through intranasal inoculation (500 μl per nostril). Blood samples were collected 14 days post-inoculation (dpi) and tested by HI assays to determine the antibody titer.
Pathogenicity assessment tests were conducted according to the WHO guidelines. 14 Briefly, 4-month-old ferrets (n = 4) were intranasally inoculated with 10 6 TCID 50 of the test virus in 1-ml sterile phosphate-buffered saline (PBS), and tissue samples were collected as previously described. 19

| Intravenous pathogenicity index in chickens
We assessed the intravenous pathogenicity index (IVPI) in SPF chickens to verify the virus pathogenicity. SPF chickens were purchased from Beijing Boehringer Ingelheim Vital Biotechnology Co., Ltd (Beijing, China). Meanwhile, the IVPI was conducted following the recommendation of the World Organization for Animal Health (OIE). 20 Viruses with IVPI values below 1.2 were considered to have low pathogenicity in chickens.

| Statistical analysis
Quantitative data are averages of three independent experiments and are expressed as means AE SD. Results were analyzed using the GraphPad Prism version 5.0 (GraphPad Software, Inc., CA, USA). A P-value of <0.05 was considered statistically significant.  Table 1.

| Growth characteristics of different reassortant viruses
We evaluated the virus replication ability in MDCK cells to compare the replication efficiency of viruses possessing different PB1 gene compositions. The results indicated that although JS-Ck62E replicated significantly better than JS-Ck62G (P < 0.001; Figure 1A), it grew significantly slower than JS-Ck53G and JS-Ck53E (P < 0.001; Figure 1A

| Effect of different PB1 on polymerase activity
The combination minigenome containing the PB1 gene from JS-Ck and PB2, PA, and NP from PR8 had higher polymerase activity than the combination with all four genes (PB2, PB1, PA, and NP) from PR8 (P < 0.05; Figure 2A). JS-Ck62E and JS-Ck53E viruses infected Vero cells after plasmid transfection. The results indicated that JS-Ck53E had better polymerase activity than JS-Ck62E (P < 0.05; Figure 2B).

| Identification of the antigenicity of different reassortant viruses
An HI assay was conducted using ferret antisera from JS-Ck62G, JS-Ck62E, and JS-Ck53E to examine the influence of HA amino acid substitution on antigenicity. All viruses were immunogenic and induced HI antibody titers (320, 320, and 160, respectively) against homologous viruses. Each antiserum effectively inhibited all the tested RG H7N4 viruses. The difference in HI titer between the homologous and detected virus was below four-folds. The substitution in HA (G to E) did not change the viral antigenicity, and all tested viruses exhibited similar antigenicity (Table 2).  (Table 3). 19 The virus (2048 hemagglutinin unit) was diluted in sterile PBS

|
(1:10) and tested in the IVPI assay. All the chickens survived, and no clinical symptoms were observed during the 10-day observation period. The IVPI of the virus was zero, thus, below the 1.2 mark.
Therefore, the reassortant JS-Ck53E virus exhibited low pathogenicity and met the safety requirements as a CVV.

| DISCUSSION
Vaccines are currently the most effective strategies for preventing and controlling influenza. Antigenic drift or shift influences the ability of viruses to escape pre-existing immunity through mutations; thus, it is highly challenging to predict future epidemic influenza strains and develop a universal vaccine. Therefore, developing specific CVV with high replication ability and low pathogenicity remains crucial for vaccine production against the influenza virus, a potential cause of the next pandemic.
In this study, RG (5 + 3) and ( Relative polymerase activities are shown as the ratio of firefly luciferase to Renilla luciferase activity. Renilla luciferase was the internal control for normalizing transfection efficiency. The background control was three plasmids transfected without the PB1 subunit in the minigenome transfection assay. Virus-free culture media were the negative control in the virus infection assay. The relative luciferase activity associated with the minigenome of PR8 or JS-Ck62E virus was set at 100%. Each experiment was repeated three times. Statistical significance was assessed using an unpaired twotailed t test with GraphPad Prism 5.0. (*, P < 0.05) the risk of inducing antigenic changes. 17,25 We hope introducing the homologous PB1 gene will improve the virus replication capacity and compensate for the mutation as previously reported. 7,8,10 As anticipated, the results of replication ability indicated that (5 + 3) viruses were significantly better than (6 + 2) viruses. As a polymerase subunit, PB1 is critical for viral replication. 26 The results of polymerase activity in both minigenome transfection and viral infection verified that homologous PB1 provided better polymerase activity than PR8 PB1. The better polymerase activity influenced high growth capacity of (5 + 3) virus, consistent with previous reports that avian PB1 improves polymerase activity better than human PB1. 27,28 In addition, some PB1 mutations are related to polymerase activity or replication ability. 12,13,[29][30][31][32][33] Most PB1 mutations were the same in PR8 and JS-Ck, except for three motifs: 180G, 473 L, and 654N (in PR8 PB1) and 180E, 473V, and 654S (in JS-Ck PB1) (Table S1). G180E and N654S improve growth ability in the three mutations, and V473L decreases polymerase activity and replication efficiency. 30,33 The results revealed that these motifs might contribute, in part, to the difference in growth ability or polymerase activity of viruses.
Furthermore, PB1 enhanced the growth ability of the (5 + 3) virus but had no visible correlation with G218E substitution in HA, which was inconsistent with a previous study. 11 This substitution is prone to occur, 23 (Table S1), probably due to lack of mutation in N66S since S contributes to viral pathogenesis. 35 The results of this study suggest that homologous PB1 is critical for enhancing the growth ability of the (

CONFLICT OF INTEREST
The authors have no conflict of interest to declare.

ETHICS STATEMENT
All animal experiments were performed following the guidelines for animal experiments described and approved by the Animal Care Welfare Committee of the National Institute for Viral Disease Control and Prevention, China CDC.

DATA AVAILABILITY STATEMENT
Data are available on request from the corresponding author. Other organs include the brain, spleen, intestine and olfactory bulb of the brain. d /, denotes no virus titer.