Construction and identification of influenza plasmid pool imparting high yields to candidate vaccine viruses in Vero cell at low temperature

Abstract We generated plasmid pools for the rapid preparation of candidate vaccine strains, which could grow in the Vero cells at low temperature. Firstly, we cloned in the pHW2000 plasmid each of the eight gene segments (PB2, PB1, PA, hemagglutinin [HA], neuraminidase [NA], NS, NP, M) of two master donor strains (MDS), respectively, A/Yunnan/1/2005Vca(H3N2) and B/Yunnan/2/2005Vca(By), which had Vca phenotype (cold‐adapted phenotype in Vero cells). Secondly, the similar operation was implemented with each of the HA, NA and NP segments of circulating strains with epidemic potential (parental strains). The virus rescue techniques were employed in this study, according to the homology rate of HA segments between MDS and parental strains. Then, we harvested amount of new Vca virus strains. By transmission electron microscope, it could observe new viruses' diameter and length were from 100 to 120 nm. Importantly, these reassortant viruses could get high‐yield production in Vero cells at 25℃ from the beginning to the fourth generation, which was significantly differ from their original parental viruses. Additional, these production 16 new Vca strains could maintain enough antibody binding capacity and attenuation phenotype, which consisted with their MDS. So these plasmid pools constructed by mount of different influenza A and B virus gene fragments could present desired working performance and provide convenience and realization for more Vca reassortant virus as candidate vaccine strain if needing.


| INTRODUC TI ON
The influenza virus is an important respiratory pathogen that infects humans, including influenza A virus (IAV) and influenza B virus (IBV). 1 They are usually predispose to exacerbation of underlying disease or development of secondary bacterial infections, especially for pregnant women, older people and young children. 2,3 Every year, there are between three and five million cases of severe illness and around 500 000 deaths worldwide. [4][5][6] Additionally, the economic burden of influenza is considerable. In the United States, for instance annual costs of 0.4 billion dollars in healthcare utilization and 16.3 billion in work absenteeism are caused by influenza. [7][8][9][10] Vaccination is the most effective way to prevent infection and severe outcomes caused by influenza viruses. 11,12 However, serious problems with influenza vaccines are antigenic drift and antigen switching between candidate vaccine strains and seasonal strains or epidemics strains. Since 1971, World Health Organization (WHO) has provided formal recommendations for the composition of seasonal influenza vaccines. 13 Among all vaccines, the process of making influenza vaccines is considered uniquely complicated and difficult. 14,15 Viral gene reassortment is a phenomenon of generation of new viral strains in a cell infected with more than one strain of a split genome virus. It is usable as a tool for the purposeful production of novel viral strains. Influenza viruses, due to having a split genome, are subject to this production method. 16 Technically, two major approaches to generate new reassortant viruses, classical reassortment and reverse genetics exist at present. As for classical approach, its goal is to combine the desired hemagglutinin (HA) and neuraminidase (NA) viral segments from circulating strain with genes from master donor strain (MDS). After inoculating two influenza strains of the same type (A or B) in a 9-day-old embryonated chicken egg, there is a high probability of simultaneous infection of cells with both strains and intracellular reassortment of viral segments.
Theoretically, there might be as many as 256 possible gene combinations. Researchers search the many combinations for the influenza strain that contains the HA and NA genes from parent strain and genes from MDS that ensure that it is able to grow efficiently in eggs or cells.
Finally, this new reassortant strain will make up next year's vaccine.
If used reverse genetics approaches, scientists clone the HA and NA genes from parent strain into plasmid (eg pHW2000). Additional plasmids are created using the remaining six genes found in MDS. Then, scientists transfected the HA and NA plasmids from parent strain and the six plasmids carrying genes from MDS into animal cells growing in the laboratory.  17 Finally, the eight gene plasmids instruct the animal cells to make the desired new influenza strain.
However, the reality is usually that most wild-type seasonal influenza viruses encoding the recommended HA and NA antigens for immunization grow poorly in eggs even after sequential passages in the laboratory. [18][19][20] In the late 1960s, Kilbourne mitigated this problem by exploiting the exceptional replication efficiency of the laboratory adapted A/Puerto Rico/8/1934 (PR8) virus in eggs through genetic reassortment method. 20,21 Most of the inactivated seasonal influenza vaccines manufactured since the early 1980s utilize PR8-derived high-yield viruses that include only the HA and NA segments from the seasonal circulating influenza A strains (6:2 segment ratio approach). 22 That means six gene segment (PB1, PB2, PA, NP, NS and M) were from MDS virus strains (as PR8) and other two gene segments (HA and NA) from parental virus strains.
Recently, more research found that distinct nucleotide and amino acid sequence polymorphisms arising as a consequence of diverse passage histories may impact the antigen yields of new viruses in manufacturing. 23 The long vaccine production process (9 months for a cycle) 24 is largely caused by the use of chicken egg-based vaccine production technologies, and this duration creates a window of opportunity for new virus variants to emerge, often resulting in decreased vaccine efficacy. 25 The egg-based vaccines prone to mutations and changes in antigenicity during egg adaptation; long production time; dependent on egg supply, which might be limited in the case of pandemics. 26,27 Regarding the reverse genetic methods had rapid production and scale-up, production can be initiated from virus genetic sequence without live virus and no risk of mutations from laboratory passage.
In this study, we constructed a useful plasmid pool used for reverse genetic methods to prepare new influenza A and B viruses. We describe a new reassortant influenza strains production system, based on new sets of plasmids generated from circulating viruses and from MDS that are different from the egg-adapted PR8 system. This extensive influenza practical plasmid pool would produce high-yield candidate flu vaccine in Vero cell at low temperature. We also performed the quality control for each plasmid and determined the optimal segment ratio for each reassortant virus strain The 6:2 ratio was suitable for some high homology of HA gene, but 5:3 ratio was suitable for some particular strains, by using gene sequence comparison software (BLAST).

| Construction of the gene plasmid pool
Total RNA was extracted from influenza virus culture solution (parental strain in Madin-Darby canine kidney (MDCK); MDS in Vero) using the QiAmp Viral RNA minikit (Qiagen, Valencia, CA) and reverse transcribed to cDNA and amplified using a one-step reaction system, and it was onestep RT-PCR kit, which was purchased from Qiagen (cat. no. 6180).
These gene segments have been obtained through RT-PCR method. Each pair of primers for each fragment is listed in Table 2.
Then, all gene fragments were sequenced by Shanghai Biological Engineering Technology Services Ltd (Shanghai). We analysed the HA, NA and NP genes by using gene sequence comparison software (BLAST). The genetic distance and phylogeny analysis were calculated.
Here, we employed the MEGA software online, which is molecular evolutionary genetics analysis software 31 and online gene sequence comparison software, called BLAST 32 for our research.
The eight gene fragments of the two donor strains were inserted into the bidirectional expression plasmid pHW2000, 17 whose schematic with markings is shown in Figure S1. The three gene fragments (HA, NA, NP) of

Size (bp)
A-PB2-F the parental strains were also inserted into pHW2000 by the same methods. The homologous recombination method was in accordance with our previous research 33 and that of Ljungberg et al. 34   Exclusion criteria were as follows: the PCR amplification of the plasmid encoded viral gene either (a) did not generate a full-length product; or (b) generated more than one product.

| Suitable ratio selection
For the production of reassortant Vca strains, we decided the use of

| HEK 293T transfection and propagation on embryonated chicken egg
To select the appropriate sets of 8 gene fragments and transfected them into HEK 293T cell, the given steps were as follows:

| Serial passages in Vero cells at 25°C
We seeded the 0.

| Mass production of reassortant viruses
The mass production of reassortant virus in Vero cells and paren-

| Serum preparation
We collected 2-3mL fasting venous blood, then separated the serum by centrifugation with 2301 g for 20 minutes and stored the patient sera at −20°C for further use. These sera were treated with receptor destroying enzyme (Denka Seiken, Tokyo, Japan) to remove nonspecific agglutination inhibitors, and it was heated to 56°C for 1 hour to inactivate its complement.
For the procedure of immunodiffusion testing, here, we standard-

| Attenuation phenotype of reassortant virus
There were three main test parameters to be calculated: median lethal dose (LD 50

| Statistical analysis of experimental data
All of the data are represented as the mean ± SD of three or more independent experiments. Since the data were homogenous, analysis of variance, the Student-Newman-Keuls test and Pearson's correlation were used. All of the analyses were performed using SPSS software, version 20.0 (SPSS Inc., Chicago, IL, USA). P values less than 0.05 were considered to be statistically significant, and less than 0.01 were considered to be very significant.

| Identification of the target gene fragments
We

| Observation of virus particles by transmission electron microscope
Using 2% phosphotungstic acid as the washing liquid for background, the reassortant viruses were observed using an electron microscope. The procedure of preparation of virus samples for EM and the procedure were described in our previous research. [41][42][43] The diameter and length of these viruses were approximately from 100 to 120 nm (Figure 2A). The morphology was complete, and the background was clear. The envelope structure of the virus was clearly visible and showed typical spherical shape influenza virus particles. Its size and shape were similar to that of the parental strain ( Figure 2B), indicating that reverse genetic to reassortant influenza virus had a normal and complete morphological structure.

| Growth characteristics of 16 reassortant viruses in Vero cell at 25°C
We cultured these 16 reassortant virus and the original parental virus in Vero cells at 25°C, respectively, from the first to fourth passages in a parallel testing. Then, we calculated their viral titter in Vero cells, illustrating these changes trend in Figure 3. These reassortant viruses could get high-yield production in Vero cells at 25°C from the beginning to the fourth generation. However, the original parental viruses could not be serially passaged in Vero cells at 25°C, which declined from the second generation. As shown, the reassortant virus showed a 100-fold increase in growth titters by the fourth serial passage. This results could be in facilitating industrial production.

| Identification of attenuation of reassortant virus in mice
The tissue culture infected dose 50% (TCID 50  We obtained the reassortant viruses by using reverse genetics methods. The values of LD 50 were 1000-fold lower in parental strains than in reassortant strains. It illustrated these two donor virus strains had provided a safety backbone for future using. As for LD 01 and LD 0 , it also found this trend. The minimal lethal dose (LD 01 ) could increase from 10 3 to 10 4 , and the maximal tolerance dose (LD 0 ) increased from 10 4 to 10 7 after reassortment experiment (Table 3).
Additionally, the weight loss is a critical indicator in vaccine safety evaluation. Recorded the weight of all experiment animals every day. And the results were consistent with PhD Long's research. 18 All parental strains could directly conduct weight loss exceeded 20%. But after reassortment method production, if

F I G U R E 1
The results of 1% agarose gel electrophoresis with eight target typical gene plasmids by PCR detecting for these plasmid pools infected with the new strains, the animals were in good condition.
Their weights grew slowly and steady, illustrating that reassortant virus was safe for live attenuated influenza vaccines (LAIVs) development.

| Vaccination
Influenza viruses cause annual seasonal epidemics and occasional pandemics of human respiratory disease, which represent a serious public health and economic problem. Fortunately, it could be most effectively prevented through vaccination. 13 But there is a challenge here, that influenza viruses suffer continual antigenic variation, which requires either the annual reformulation of seasonal influenza vaccines or the rapid generation of vaccines against potential pandemic virus strains.
There are two methods to process influenza vaccine virus selection and development, reverse genetics and classical reassortment approach. Their goal is to combine the desired HA and NA genes from circulating strain with the six other genes from MDS, which grows well in eggs or cell cultures.

| Vero cells adaptation
First of all, the suitable cell line is very important to study the virus adaptability. We employed the Vero cell line in this study.
But it is true that the advantages of the use of the interferondeficient Vero cell line for the production of candidate vaccine strains, which is in order to decrease the fitness of the virus to in- But if interferon from another source is added to the culture, the process of virus-infected Vero E6 cells will be interfered, because they have the interferon-alpha/beta receptor. The defect is due to a ~9-Mb deletion in chromosome 12 of Vero cells, causing the loss of the type I interferon gene cluster. 51  The diameter of the immunoprecipitate circle (mm) produced non-infective influenza that remained predominantly cell associated. But the infectious virus particles could be prepared through HA titre testing and TCID 50 testing in our research.
Honestly, we study influenza A and B virus strain, but not C type, which was not used as vaccine. It was a limitation of our research.
PhD Kaverin NV 54 found that influenza virus replication in Vero cells was impaired by rapid inactivation of trypsin in the culture fluid by a factor secreted by Vero cells and that repeated addition of trypsin to the culture medium allows for multicycle growth of influenza A. In fact, not only influenza A but also B type virus needed trypsin during their replication in Vero cells.
In the same year, PhD Govorkova EA 55 suggested that by adding trypsin the efficiency of primary isolation for circulating H3N2 strains was similar in Vero and MDCK cells and that the amino acid sequence of HA1 was similar if passaging in Vero cells or MDCK.
This is a great try and founding. Our goal is in high-yield industrial production, including but not limited to isolation of influenza virus.

| Reverse genetics method using
As for the reverse genetics method, we employed the pHW2000 cloning system, which was carried out in the laboratory to generate reassortant viruses as vaccine candidate strains. This cloning system has the advantage of including promoters for pol I and pol II in opposite directions on the sides of the cloned viral segments in order to generate both mRNAs and copies of the viral genome. 56 The target viral segment is inserted between the sites for a pol II promoter element (truncated CMV promoter) and a pol I promoter element (truncated human pol I promoter) that are present in positive and negative orientation, respectively. Therefore, from this plasmid, both viral mRNA (pol II) and vRNA (pol I) can be synthesized without the need of additional helper viruses or further transfection with other polymerase genes.
In this study, we used homologous recombination between PCR products and the pHW2000 plasmid instead of digestion with BsmBI or BsaI and insertion into BsmBI sites in the linearized pHW2000. 17 This method was useful during our research. Previously, Professor  E coli) to carry the plasmid allows for easy amplification and retrieval of specific clones from the pool for analysis. It also provided a convenient transport to different laboratory. Of course, the quality control of the plasmid pool was very important by our research founding, which described previously. 34 In general for the production of reassortant viruses of the A/ This study with plasmid pool, influenza reverse genetics approaches and their implementation had supplied rapid, convenient method to rescue safe and efficacy candidate vaccine strains.

| Further using
Theoretically, the findings of this study will not only be used for the It could solve the keynote of LAIV developing with Vero cell cold-adapted strain (Vca). These properties were optimal for efficient growth and vaccine production. Additionally, we were not only studied the IAVs but also IBVs for the plasmid pool construction to generate the reassortant viruses.
The cold adaptation characteristic of LAIV is very useful and important. The human nasal or upper respiratory tract temperature is between 25°C and 28°C, which is suitable for LAIV replication. Then, the inoculation pathway could occur through nasal immunization, which provides a convenient alternative for subjects and a similar route to natural infection. This injection route can elicit mucosal cellular immunity in addition to humoral immunity, thus providing strong protective efficacy and long-lasting immunity.
Collectively, our study provides feasibility of rapid, quantities and high recombination efficiency for future IIV or LAIV vaccine by these plasmid pool.

| Limitations
However, more immunization experiment and challenge testing should be taken in the future for practical protective effect against seasonal influenza or avian Influenza. Back to the origin MDV, they were differ from PR8, and they and their reassortant virus require more human clinical trails to confirm their safety and immunogenicity.
On the next step, we should analyse more data after many more passages on the Vero cell line at 25°C, which included the evolution of the viral titter; the conservation of surface antigenicity and of HA sequence; and the conservation of attenuation in vivo.
The immunogenicity of the viral strains (H5N1, H7N9, H9N2 or H1N1-pdm09) had not been tested in immunodiffusion, that is a big stumbling stone during our research. We cannot collect the positive serum from patients infected with H5N1, H7N9 and H9N2. We have plan injecting the inactivated reassortant avian virus into the goat, to harvest the positive serum. Then, we could do AGID testing for avian virus. Maybe in the next flu season, we hope to carry out research on this project.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data, models and code generated or used during the study appear in the submitted article.