Improvement of influenza vaccine strain A/Vietnam/1194/2004 (H5N1) growth with the neuraminidase packaging sequence from A/Puerto Rico/8/34.
ABSTRACT H5N1 influenza candidate vaccine viruses were developed using the "6+2" approach. The hemagglutinin (HA) and neuraminidase (NA) genes were derived from the popular H5N1 virus and the remaining six internal segments were derived from the A/Puerto Rico/8/34 strain (H1N1, PR8). However, some of these candidate strains have been reported to produce relatively low yields in vaccine manufacture. In this study, we found that the NA vRNA of the A/Vietnam/1194/2004 strain (H5N1, VN1194) was poorly packaged into recombinant viruses with a backbone of PR8 genes, which resulted in the formation of defective virions that did not include the NA vRNA in the genome. Using recombinant DNA techniques, we constructed a chimeric NA gene with the coding region of VN1194 NA flanked by the packaging signal sequence of the PR8 NA gene (41 bp form the 3' end of the vRNA and 67 bp from the 5' end). The packaging of the NA vRNA was restored to normal levels in the recombinant viruses containing the chimeric NA gene. Recombinant viruses containing the chimeric NA replicated much better in chicken embryonated eggs than viruses with the wild-type NA from VN1194. These findings suggest a novel strategy to improve in ovo growth of vaccine strains and to increase the number of vaccine doses available to save people if a pandemic were to occur.
© 2012 Landes Bioscience.
Do not distribute.
2004 strain (H5N1, VN1194) was poorly packaged into recombinant viruses with a backbone of PR8 genes, which resulted
in the formation of defective virions that did not include the NA vRNA in the genome. Using recombinant DNA
techniques, we constructed a chimeric NA gene with the coding region of VN1194 NA flanked by the packaging signal
sequence of the PR8 NA gene (41 bp form the 3’ end of the vRNA and 67 bp from the 5’ end). The packaging of the NA
vRNA was restored to normal levels in the recombinant viruses containing the chimeric NA gene. Recombinant viruses
containing the chimeric NA replicated much better in chicken embryonated eggs than viruses with the wild-type NA
from VN1194. These findings suggest a novel strategy to improve in ovo growth of vaccine strains and to increase the
number of vaccine doses available to save people if a pandemic were to occur.
Improvement of influenza vaccine strain
A/Vietnam/1194/2004 (H5N1) growth
with the neuraminidase packaging sequence
from A/Puerto Rico/8/34
Weiqi Pan,1,2Zhenyuan Dong,1,2Weixu Meng,1,2Wei Zheng,1,2Ting Li,1Chufang Li,1Beiwu Zhang2and Ling Chen1,2
1State Key Laboratory of Respiratory Diseases, Institute of Respiratory Diseases, Guangzhou College of Medicine; Guangzhou, China;2State Key Laboratory of Respiratory Diseases,
Guangzhou Institute of Biomedicine and Health, Chinese Academy of Sciences; Guangzhou, China
Keywords: H5N1 influenza, vaccine, hemagglutinin, neuraminidase, packaging signal
Abbreviations: AEC, aminoethylcarbazole; BSA, bovine serum albumin; DMEM, Dulbecco’s modified Eagle’s medium;
EID50, 50% egg infectious dose; HA, hemagglutinin; HDV, hepatitis delta virus; MDCK, Madin-Darby canine kidney;
NA, neuraminidase; NIBSC, National Institute for Biological Standards and Control; NP, nucleoprotein; PA, polymerase acid;
PB1, polymerase basic 1; PB2, polymerase basic 2; PBS, phosphate balanced solution; PCR, polymerase chain reaction;
PFU, plaque-forming unit; qPCR, real-time quantative PCR; RNP, ribonucleoprotein; RT-PCR, reverse transcribe polymerase chain
reaction; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; 293T, human embryonic kidney cell culture;
UTR, un-translation region; WHO, World Health Organization
H5N1 influenza candidate vaccine viruses were developed using the “6 + 2” approach. The hemagglutinin (HA) and
neuraminidase (NA) genes were derived from the popular H5N1 virus and the remaining six internal segments were
derived from the A/Puerto Rico/8/34 strain (H1N1, PR8). However, some of these candidate strains have been reported to
produce relatively low yields in vaccine manufacture. In this study, we found that the NA vRNA of the A/Vietnam/1194/
Since the first human case of highly pathogenic avian influenza
(H5N1) infection was reported in Hong Kong in May 1997, the
H5N1 virus has been identified in poultry, migratory birds, and
mammals, including humans, in several countries in Southeast Asia,
Africa, and Europe (www.who.int/csr/disease/avian_influenza/ai_
timeline/en/index.html). These viruses possess a new H5 subtype
of hemagglutinin (HA), against which there is presently little
immunity in human populations. To date (up to June 10, 2011),
the World Health Organization (WHO) has received 556 reports
of humans infected with H5N1 avian influenza, of whom 325
have died (www.who.int/csr/disease/avian_influenza/country/en/).
Although person-to-person spread has been limited, H5N1 viruses
represent a potential source of a future influenza pandemic.1,2
Vaccination against pandemic influenza during the early wave of
infections is considered to be the most effective option to limit
the spreadofthe virus.The WHO recommends the developmentof
6:2 reassortant vaccine strains of H5N1 by plasmid-based reverse
genetics.3In each of the 6:2 reassortant viruses, the modified
avirulent HA and NA genes are derived from an H5N1 human
isolate, and the remaining six genes are derived from the A/Puerto
Rico/8/34 (H1N1, PR8) strain, which is known to grow well in
eggs. The vaccine candidate strain, which carries the HA and NA
genes from the A/Vietnam/1194/2004 (H5N1, VN1194) strain,
was determined to be a suitable vaccine candidate based upon its
safety and immunogenicity.4-6And ithas been distributed to vaccine
manufacturers worldwide for use in vaccine manufacture. However,
it has been reported that the yield of HA antigen from this
recombinant vaccine strain is lower than expected.7Therefore, to
Correspondence to: Ling Chen; Email: firstname.lastname@example.org
Submitted: 06/15/11; Revised: 10/14/11; Accepted: 10/19/11
Human Vaccines & Immunotherapeutics 8:2, 252–259; February 2012;G2012 Landes Bioscience
252Human Vaccines & Immunotherapeutics Volume 8 Issue 2
© 2012 Landes Bioscience.
Do not distribute.
quickly meet the demand of a future epidemic, new methods with
improve the HA production need to be developed.
The genome of the influenza A viruses contain eight negative-
sense RNA segments. The viral RNAs, together with three
polymerase proteins and the nucleoprotein (NP), form viral
ribonucleoprotein (vRNP) complexes.8In the final step of viral
replication cycle, eight vRNAs are packaged into progeny virions
in the form of vRNP complexes in an equimolar ratio.9Although
both a random incorporation model and a selective incorporation
model have been proposed for the packaging of the influenza
vRNAs, accumulating research suggests that the vRNAs are
selectively packaged into budding virions via packaging signals
residing at both ends of each vRNA.10-20Mismatched 3' and 5'
ends from different types of influenza viruses and different seg-
ments dramatically impair the packaging efficiency and virulence
of the virus both in vitro and in vivo.11,21,22Furthermore, silent
mutations in these nucleotides also led to defects in gene
packaging and viral replication.12-14,17,18,20In the present study, a
series of recombinant viruses with chimeric NA genes based on
the PR8 NA and VN1194 NA genes were generated by reverse
genetics to explore the role of the NA packaging signal in the
growth of vaccine virus. By comparing the original VN1194 NA
vaccine strain and the chimeric PR8-VN1194 NA virus, we found
that the addition of the packaging signal sequence of the PR8
NA gene to the target VN1194NA vRNA segment can improve
the viral growth titer and HA antigen content. Our finding
may significantly improve influenza vaccine yield, which could
result in the production of increased vaccine doses and lower
Sequences of the 3' and 5' ends of HA and NA segments from
NIBRG-14. Because there is no complete cDNA sequences that
included the un-translation regions (UTR) of segments HA and
NA segments from NIBRG-14 or A/Vietnam/1194 (H5N1)
strains reported in GeneBank, we first performed an analysis to
determine the 3' and 5' end sequences of these two segments. Full
length cDNA was circularized using CircLigase, which is an ATP-
dependent ligase that can circularize single-stranded DNA
templates with a 5'-phosphate and a 3'-hydroxyl group that are
longer than 30 nucleotides. To isolate the cDNA ends, the
circularized cDNA was used as template for inverse PCR with
nested pairs of gene specific primers that were directed away from
each other. PCR products were cloned into the pMD-18T vector
and then sequenced by plasmid DNA sequencing. The nucleotide
sequences of the 5' and 3' UTR sequences of each cDNA, from
the stop codon (TAG) to the first codon (ATG) of the ORF, are
shown in Figure1(panel A). Sequence alignment of the 3' and 5'
ends (1–41 bp; 1347–1413 bp) of the PR8 NA and VN1194 NA
vRNA revealed that there are six nucleotide differences in these
regions between the two strains (Fig.1, panel B). G4A, G13A,
and C1387U are located in the UTR. A38G is a silent mutation
in the coding region, and G1372C results in a serine to threonine
substitution. Additionally, VN1194 NA has seven adenines, while
PR8 NA has six adenines in the poly (A) site.
Rescue of recombinant viruses. Based on the packaging signals
reported for NA,17the coding sequence of VN1194 NA was
flanked by 21 bases at the 3' end and 39 bases at the 5' end and
both the 3' and 5' UTR sequences from the PR8 NA vRNA
(Fig.1, Panel C). To assess the effect of the PR8 NA vRNA
packaging signal on the recombinant “6 + 2” H5N1 influenza
vaccine strain, a series of recombinant viruses with different
pairings of HA and NA genes were generated using a 12-plasmid
based system. Viral RNA from the rescued viruses was prepared
and reverse transcribed. Full-length of HA and NA were PCR
amplified with segment specific primers. Sequence analysis was
used to verified that the HA and NA sequences from the rescued
viruses after plaque purification were identical to that of the
constructed plasmids. The names of the recombinant viruses and
the compositions of the viral genomes are shown in Table 1.
Plaque immunostaining of recombinant viruses. Plaque assays
were performed to test the ability of recombinant viruses with
different combinations of HA and NA genes to form plaques in
MDCK cells. Three days post-infection, the cells were immuno-
stained with an anti-NP monoclonal antibody. Viruses containing
the wild-type VN1194 NA gene produced much smaller plaques
than the rescued wild-type PR8 virus, whereas the viruses
containing the PR8(21)VNNA(39) NA gene were able to form
plaques that were almost the same size as the plaques formed by
the wild-type PR8 (Fig.2). This observation suggests that the
insertion of packaging signals from the PR8 NA gene are able to
promote the replication of viruses containing the wild-type
VN1194 NA gene in MDCK cells.
NA vRNA levels in purified recombinant viruses. To assess
the incorporation of the chimeric and wild-type VN1194 NA
gene segments into rescued virions, 1mg of purified virion RNA
for each virus was analyzed using PAGE followed by SYBR II
staining. The bands corresponding to the NA vRNA from
PR8HA-VNNA and VNHA-VNNA viruses were hardly visible;
whereas the other seven vRNA segments in each virus were
present in approximately equimolar ratios. In contrast, after the
incorporation of the packaging signal sequence from PR8 NA, the
amount of NA vRNA in both PR8HA-PR8(21/39)VNNA and
VNHA-PR8(21/39)VNNA viruses were restored to similar level
as the other seven vRNA segments (Fig.3). This result indicates
that the wild-type VN1194 NA segment is not completely
“recognized” or packaged into virions with a PR8 backbone. The
inefficient incorporation of VN1194 NA RNA into the genome
of recombinant viruses resulted in defective virions without NA
vRNA and may have affected the growth of the virus. Our study
demonstrated that incorporation of the VN1194 NA into virions
can be improved by adding the 21- and 39-nucleotide coding
regions and the non-coding regions from the 3' and 5' termini of
the PR8 NA gene to the target NA.
qPCR analysis of packaged NA vRNA. Because RNA gel
electrophoresis only provides visible but not quantitative analysis
of RNA, we analyzed the packaging of NA vRNA using qPCR.
In the qPCR reactions, each gene segment was amplified using
specific primers, and the percentage of NA vRNA was determined
in comparison to the amount of NA vRNA in the wild-type PR8
virus. As shown in Table 3, the packaging efficiency of the
www.landesbioscience.com Human Vaccines & Immunotherapeutics253
© 2012 Landes Bioscience.
Do not distribute.
VN1194 viruses. The nucleotide numbering used for the PR8 NA was used. Sequence differences are shown in the box. The dash (-) is inserted for
maximum sequence homology. (C) Schematic diagram of the chimeric VN1194NA vRNA. The chimeric VN1194NA vRNA contains the 3’ UTR from the PR8
NA vRNA (20 nucleotides), 21 nucleotides from the PR8 NA coding region (with the start codon changed to ACG), the nucleotides for the complete
coding region of VN1194 NA in the negative sense, the 39 nucleotides corresponding to the PR8 NA C terminus, and the 5’ UTR from the PR8 NA vRNA.
This chimeric RNA is shown in the negative-sense orientation. The lengths of the regions are not to scale.
VN1194 NA vRNA in the PR8HA-VNNA and VNHA-VNNA
viruses was 8.2% and 33.7%, respectively. The packaing effici-
ency of VN1194 NA segment was restored to the level of wild-
type PR8 after the addition of the PR8 NA vRNA packaging
regions to both ends of VN1194 NA vRNA segment, as shown
with the PR8HA-PR8(21/39)VNNA and VNHA-PR8 (21/39)
Growth kinetics of recombinant viruses. To compare the
growth properties of the viruses containing the PR8(21)VNNA
(39) NA gene with those of the viruses containing the original
VN1194 NA gene, the kinetics of virus yield were assessed by the
inoculating 10 d old embryonated eggs with 100 EID50of each
virus stock. At 12 h intervals post-inoculation, viral titers were
determined using the plaque immunostaining assay and HA
assays. Two-way ANOVA with a Bonferroni’s multiple com-
parison test was used to analyze the data, and differences were
considered statistically significant at p , 0.05. On the PR8 HA
gene and PR8 internal genes background, from 12 h to 60 h,
the PR8HA-PR8(21/39)VNNA virus replicated significantly
better than the PR8HA-VNNA virus, producing 80-fold higher
pfu titers than the PR8HA-VNNA virus at 48 h (Fig.4A). The
HA titer of the PR8HA-PR8(21/39)VNNA virus was also 2- to
4-fold higher than that of the PR8HA-VNNA virus at 24 h to
60 h post-inoculation (Fig.4B). On the VN1194 HA gene
and PR8 internal genes background, VNHA-PR8(21/39)VNNA
produced 45-fold higher pfu titer (Fig.4A) and 2.5-fold higher
HA titer (Fig.4B) than the VNHA-VNNA (the vaccine strain
virus) at 48 h post inoculation. Taken together, these data
suggested that the growth of the vaccine virus, NIBRG-14,
could be significantly enhanced if the coding sequence of the
VN1194 NA gene was flanked by the packaging signal from PR8
Table1. Genome background of the rescued recombinant viruses
Virus HA gene NA gene Internal genes
wt PR8PR8 PR8 PR8
PR8HA-PR8(21/39)VNNAPR8 PR8(21)VNNA(39) PR8
Figure1. (A) The cDNA sequences of the 3’ and 5’ UTRs of the NIBRG-14 HA and NA segments. The predicted 5’ and 3’ UTR sequences of four different
clones were analyzed for each segment using the circularized cDNA and nested-reverse PCR strategy. The stop codon (TAG) and start codon (ATG)
of each open reading frame are shown in bold letters. The common sequences of 13 nucleotides at the 3’ end and the 12 nucleotides at the 5’ end of the
influenza A virus are enclosed in the box. (B) Alignment of the 5’ and 3’ packaging signal sequences of the NA and HA segments from the PR8 and
254 Human Vaccines & Immunotherapeutics Volume 8 Issue 2
© 2012 Landes Bioscience.
Do not distribute.
Assessment of HA protein in recombinant viruses. To evaluate
whether the enhanced growth in embryonated eggs correlated
with an increased yield of HA antigen, equal volumes of each of
concentrated allantonic fluids (200 times from collection) were
deglycosylatd and analyzed by SDS-PAGE. The HA content of
the recombinant viruses was compared between two groups
according to the background of the HA gene: the PR8 HA group
and the VN1194 HA group. In each group, the calculated HA
content of viruses containing the PR8(21)VNNA(39) gene was
shown relative to the virus containing the wild-type VN1194NA
gene. The results showed that there was a 2.7-fold (for VNHA)
and 2.8-fold (for PR8 HA) increase in HA concentrations in
viruses containing the PR8(21)VNNA(39) gene than in viruses
containing the wild type VN1194 NA gene (Fig.5), respectively.
This result indicates that more HA doses of inactivated vaccines
could be manufactured from a given number of eggs by the
insertion of the PR8 NA packaging signal at both ends of the
VN1194 NA coding sequence.
According to the recommendation of the WHO, H5N1 vaccine
seed viruses were developed by reverse genetics using a “6 + 2”
approach. In each of the 6:2 reassortants, the modified HA and
the original NA segments are derived from an H5N1 reference
virus and the remaining six segments are from strain A/PR/8/34.3
The vaccine seed virus should have the antigenic properties of
the targeted reference virus, and ideally, maintain the growth
properties of the original donor strain, A/PR/8/34, for more
efficient vaccine production. However, some of these vaccine
strain viruses do not grow as well as the original donor strain
A/PR/8/34 even though they possess the same six “internal”
genes.23During an influenza pandemic, manufacturers would
need to produce as many doses of vaccine as possible in the
shortest amount of time to protect the public from infection.
Figure3. Polyacrylamide gel electrophoresis of viral RNA. vRNAs were
extracted from the purified allantoic fluid of eggs infected with wild-type
PR8, PR8HA-VNNA, PR8HA-PR8(21/39)VNNA, VNHA-VNNA or VNHA-PR8
(21/39)VNNA viruses. One microgram of purified viral RNAs were loaded
onto a 2.8% polyacrylamide gel containing 7.7 M urea and visualized
following staining with SYBRII. The RNAs that encode the polymerase
proteins, hemagglutinin (HA), nucleoprotein (NP), neuraminidase (NA),
the matrix proteins (M), and the nonstructural proteins (NS) are
indicated. The RNA concentration of each band of RNA ladder is 120 ng.
Table2. Oligonucleotide primers for isolation of 5’and 3’ end sequences of
HA and NA gene of NIBRG-14
Figure2. Plaquephenotypeof thewild-type PR8 and recombinantviruses.
MDCK cells were infected with wild-type PR8 virus, PR8HA-VNNA virus,
PR8HA-PR8(21/39)VNNA virus, VNHA-VNNA virus, or VNHA-PR8 (21/39)
VNNA virus and overlaid with agarose. Three days post-infection, the cells
were subjected to immunostaining with an anti-NP monoclonal antibody.
Table3. Packaging efficiency of the NA vRNA into virions
PR8 PR8HA-VNNA PR8HA-PR8(21/39)VNNAVNHA-VNNA VNHA-PR8(21/39)VNNA
100.0 8.2 ± 0.999.3 ± 4.3 33.7 ± 2.5101.2 ± 1.1
The values are presented as the percentage of incorporated NA vRNA compared with the incorporation into wild-type PR8 virions. The results are presented
as the average and standard deviations from two independent experiments, with the assay performed in triplicate (n = 6).
www.landesbioscience.com Human Vaccines & Immunotherapeutics255
© 2012 Landes Bioscience.
Do not distribute.
Therefore, the low yield of HA antigen from a candidate influenza
vaccine virus would be a matter of concern. It would be
worthwhile to improve the viral growth and the yield of HA
antigen from candidate influenza vaccine viruses that have
lower than expected yields. Horimoto et al.24reported that 7:1
reassortant viruses containing PR8 NA and modified HA from
H5N1 viruses grew significantly better than 6:2 reassortant
viruses containing modified HA and NA from H5N1 virus.
Although HA is the major protective antigen in inactivated
vaccines, NA is the other surface antigen that plays an important
role in the influenza-specific immune response.25,26Therefore, NA
should be an indispensable component of the vaccine seed virus.
In this study, we found that wild-type VN1194 NA was not
being completely “recognized” by the packaging machinery on the
PR8 internal genes with the VN1194 HA gene background or
the PR8 internal genes with PR8 HA gene background. This
observation is consistent with the report by Fujii et al.,17which
demonstrated that both the noncoding and coding regions of the
3' and 5' ends of the NA vRNA are required for incorporation of
the NA segment into virions. The 21 nucleotides in the 3' coding
region appeared to be critical for efficient vRNA incorporation.
Sequence alignment of the 3' and 5' ends (1–41 bp; 1,347–1,413
bp) of the PR8 NA and VN1194 NA vRNA demonstrated that
there are six nucleotide differences in these regions (Fig.1, panel
B). To determine whether these differences are responsible for the
decreased packaging efficiency of the VN1194 NA into the PR8
gene background, without changing the amino acid sequence of
VN1194 NA, the coding sequence of VN1194 NA was flanked
with 21 bases at the 3' end and 39 bases at the 5' end and the
Figure4. Growth properties of recombinant viruses. Ten-day old
embryonated eggs were inoculated with 100 EID50of virus.
At the indicated times after inoculation, virus titer in the allantoic fluid
was determined using plaque immunostaining (A) and an HA assay (B).
The values are presented as the mean and standard deviations from
three independent experiments.
Figure5. Quantitation of HA content in allantoic fluid 48 h post-infection.
The HA content of PR8HA-PR8(21/39)VNNA and VNHA-PR8(21/39)VNNA
viruses are shown relative to the PR8HA-VNNA and VNHA-VNNA,
respectively. The data are presented as the mean and standard deviations
from three independent experiments, and error bars show standard
deviation of the results. Viruses with a significantly increased HA
concentration are indicated by an asterisk (*) (p , 0.05, Student’s t-test
with two-tailed analysis).
256 Human Vaccines & ImmunotherapeuticsVolume 8 Issue 2