Immunogenicity and safety of H1N1 influenza hemagglutinin protein expressed in a baculovirus system



Jae-Hwan Nam, Department of Biotechnology, Catholic University of Korea, 43-1 Yeokgok 2-dong, Wonmi-gu, Bucheon, Gyeonggi-do, 420-743, Korea. Tel: +82 2 164 4917; fax: 82 2 2164 4917; email:


Although most influenza vaccines are produced in eggs, new types of vaccines must be developed. In this study, the immunogenicity and safety of a baculovirus-expressed hemagglutinin (HA) of H1N1 influenza virus (Korea/01/2009; designated “HA-Bac-K”) was compared with those of a commercially available baculovirus-expressed HA (designated “HA-Bac-C”) and an Escherichia coli-expressed HA (designated “HA-E. Coli-K”). HA-Bac-K succeeded in inducing hemagglutination inhibition and neutralization antibodies in mouse and ferret models. The different immunogenicities observed may be attributable to the different expression systems and purification protocols used. Our work suggests that HA expressed in a baculovirus system is an effective and safe candidate influenza vaccine.

List of Abbreviations

Autographa californica nuclear polyhedrosis virus

E. coli

Escherichia coli




hematoxylin and eosin


hemagglutination inhibition





Influenza A and B viruses cause seasonal epidemics that result in significant morbidity and mortality. Among these pandemic strains, H1N1 (Spanish influenza virus) caused at least 50 million deaths in 1918 [1]. In 2009, the swine influenza A H1N1 virus spread rapidly and was classified as a pandemic strain by the World Health Organization in the months after its identification [1]. The influenza virus has two major glycoproteins on its cell surface, HA and NA. This virus has 16 HA and nine NA subtypes. HA regulates viral entry into the cell and, as the main antigenic driver, is important in the adaptive immune response [2]. The immune response to HA is directed against the head of the glycoprotein [3, 4]. Therefore, influenza virus HA is a significant target for vaccination against the virus.

Prevention of the spread of influenza virus is mainly achieved through vaccination. Three types of influenza vaccines are now available: whole virus, split and subunit vaccines [5]. However, because current vaccine production requires viral replication, vaccine production itself constitutes a risk in that virus from a production facility can escape and be disseminated into the human population [6]. Many of the vaccines used clinically are inactivated viruses that are propagated in embryonated chicken eggs [7]. However, production of egg-based influenza vaccines is dependent on the availability of embryonated eggs, which may be compromised in the event of outbreaks of avian diseases [8].

To circumvent these problems, the use of baculovirus expression systems to make new types of influenza vaccine is increasing. Baculovirus expression maintains the biological activity of the protein by ensuring the appropriate posttranslational modifications, such as glycosylation, disulfide bond formation and phosphorylation [9, 10]. Moreover, this expression system is not limited by the size of the DNA to be expressed; it can express DNAs from humans, microorganisms and animals [10, 11].

In this study, we constructed HA of H1N1 influenza virus (A/Korea/01/2009), with gp67 signal peptidein the AcNPV expression system, as a recombinant baculovirus-expressed HA (designated “HA-Bac-K”) in a secreted form. We compared its immunogenicity with that of a commercial HA protein (designated “HA-Bac-C”), which is also expressed in a baculovirus system. We also used Escherichia coli-expressed HA protein (designated “HA-E. Coli-K”), which constitutes a partial region of HA (HA1), to compare the expression systems and determine the extent to which immunogenicity is dependent on the expression system. To establish the protective activity of these proteins as vaccines, we confirmed the presence of specific HI and NT antibodies in mice and ferrets injected with these antigens. All procedures involving animals were approved by the institutional review board of the Catholic University of Korea.

To confirm by detection of anti-influenza antibodies that the HA-Bac-K, HA-Bac-C, and HA-E. Coli-K proteins can be used as vaccines, each purified protein was used as an antigen in a western-blot-based analysis. Baculogold DNA (BD Bioscience, San Jose, CA, USA) and target DNA that included the full HA gene of the H1N1 influenza virus (A/Korea/01/2009) were used to generate HA-Bac-K. To express HA-Bac-K, baculovirus containing the HA gene was used to infect suspension cultures of Hi5 cells. After 3 days at 28°C, the culture medium was harvested and applied to a Ni-NTA column, which was equilibrated using 20 mM Tris–HCl (pH 8.0) and 200 mM NaCl. After washing with 50 mM imidazole, the precursor HA protein was eluted using an imidazole gradient before dialysis against 10 mM Tris–HCl (pH 8.0) and 50 mM NaCl, followed by hydrolysis with TPCK-treated trypsin for 3 hrs. The reaction was stopped by addition of 1 mM phenylmethylsulfonyl fluoride, after which the active form of HA was purified by Mono Q ion-exchange chromatography and Superdex 200HR size exclusion chromatography. HA-BAC-C was obtained from Sino Biological (Beijing, China; cat. # 11055-V08B). The DNA sequence of HA-Bac-C encodes the extracellular domain of influenza A virus H1N1 (A/California/04/2009). HA-Bac-C was 6 × histidine-tagged at the C-terminus for purification. First, 1.5 mL of Ni-NTA agarose was poured into a purification column and 6 mL of distilled water added. The resin was resuspended by inverting the column and 6 mL of native-binding buffer was added. Next, 8 mL of protein lysate was added to the column, which was incubated for 30–60 mins with gentle agitation to maintain the resin. The resin was washed with 8 mL of native wash buffer and allowed to settle by gravity; this was repeated four times. The protein was eluted with 8–12 mL of native elution buffer and 1 mL fractions were collected. HA-E. Coli-K was expressed as a truncated form (amino acids18–344) of the HA1 region of influenza virus H1N1 (A/Korea/01/2009) [12].

For detection of HA proteins, polyclonal antibodies from patient serum (A/Korea/01/2009; antibody concentration, 1:1,000) were used. The HA proteins produced by baculovirus were detected at sizes of about 60–70 kDa and the HA1 protein produced by E. coli was detected at 35 kDa (Fig. 1). Although HA-Bac-K and HA-Bac-C had the same amino acid sequences in their HA proteins, they had different molecular weights (Fig. 1). It was suspected that HA-Bac-C may contain an additional protein that is fused to HA protein for its purification and/or expression. However, further studies are required to obtain more detailed information. In any case, this result confirms that HA-Bac-K, HA-Bac-C and HA-E. Coli-K proteins were intact influenza virus HAs that react with patient serum.

Figure 1.

Antigenicity of the expressed HA proteins. HA-Bac-K and HA-Bac-C were produced from baculovirus expression systems and HA-E. Coli-K from an E. coli expression system. The vaccines were purified with rProteinA resin and an Ni–NTA column. The purified proteins were loaded onto an SDS–PAGE gel. The primary antibody used was serum from an infected patient. The size of HA-Bac-K was 65 kDa, of HA-Bac-C 70 kDa and of HA-E. Coli-K 35 kDa (•; index of size).

To determine the antibody titers in mice after immunization with the purified HA proteins, mice (6 weeks old, female, n = 10 per group) were injected twice i.m. with 30 μg of each purified HA protein and Imject Alum adjuvant (3:1 ratio, cat. # 77161; Thermo Scientific, Rockford, IL, USA) with a two week interval. Serum was collected from intraorbital vein blood to measure antibody production. Ten mice sera were pooled to measure the HI and NT titers. In the HI test, 100 μL of serum was added to 300 μL of receptor-destroying enzyme and incubated at 37°C overnight. PBS was incubated at 56°C for 30 mins and the serum mixture added to 600 μL of warm PBS. This serum mixture was added to packed red blood cells and incubated at 4°C for 1 hr. The mixture was then centrifuged and the supernatant collected. Twenty-five microliters of serially diluted serum and 25 μL of pandemic H1N1 influenza (A/Korea/01/2009) were placed in a well and incubated for 30 mins. Next, 50 μL of 0.5% turkey red blood cells were added to each well and incubated at room temperature for 1 hr. Ten days after the second immunization, the HI antibody titer of the mice was 1:160 in the HA-Bac-K group and 1:10 in the HA-Bac-C group (Table 1).

Table 1. Hemagglutination inhibition and neutralization antibody titers induced in mice immunized with the expressed HAs
 10 days pi60 days pi
HI titerNeutralization titerHI titerNeutralization titer
  1. ND, not detected; pi, post-inoculation.
HA-E. Coli-K1:801:201:801:10

For the NT test, inactivated serum and 200 plaque-forming units of H1N1 influenza virus (Korea/01/2009) were mixed at a 1:1 volumetric ratio and incubated at 37°C for 1 hr. A549 cells were infected with the serum and virus mixture, then incubated at 37°C. After 2 hrs, autoclaved agarose and 2 × DMEM (with 4% FBS) were mixed at a 1:1 volumetric ratio. The cells were separated from the virus and mixed with 3 mL of the agarose mixture. At 5 days after infection, the cells were stained with crystal violet and the viral titers counted based on plaque numbers. The NT antibody titer was 1:80 in the HA-Bac-K group, but only <1:5 in the HA-Bac-C group (Table 1). In the HA-E. Coli-K group, the HI antibody titer was 1:80 and the NT antibody titer 1:20 (Table 1). Sixty days after the second immunization, HI and NT antibodies of the HA-Bac-C group were undetectable. However, the HI antibody titers of the HA-Bac-K and HA-E. Coli-K groups were maintained at 1:80 (Table 1). The NT antibody titer of HA-Bac-K was 1:20 and that of HA-E. Coli-K 1:10 (Table 1). Although the NT antibody titers of HA-Bac-K and HA-E. Coli-K did decrease, they were maintained for 60 days after immunization. Generally, HI antibody titers of 1:30–40 induced by immunization with an influenza vaccine prevent 50% of infections [13]. In our study, HA-Bac-K and HA-E. Coli-K displayed HI antibody titers of 1:80–1:160 at 10 days after the second immunization and HI antibody titers of 1:80 were maintained for 60 days after immunization (Table 1). Therefore, the immunogenicities of both recombinantly expressed HAs seem appropriate for candidate vaccines directed against the influenza virus. Based on these data, HA-Bac-K and HA-E. Coli-K were used in further tests in ferrets to analyze their immunogenicity.

Ferrets (n = 3 per group) were injected with HA-Bac-K or HA-E. Coli-K i.m. to confirm their safety and immunogenicity. The ferrets were injected twice with 40 μg of protein and Imject Alum adjuvant (3:1 ratio) per ferret with a two week interval. Sera were collected from the ferrets 10 days after the second immunization. HI and NT titers were determined separately in the three ferret sera. All the data for HI and NT antibody titers include standard deviations. HI antibody titers were determined against two influenza strains (A/California/07/2009 and A/Korea/01/2009). The HI antibody titers of both HA-Bac-K and HA-E. Coli-K were 1:160 against A/California/07/2009 and 1:80 against A/Korea/01/2009, whereas only HA-Bac-K showed an NT antibody titer (of 1:160; Table 2). Thus, interestingly, HA-Bac-K appeared to induce neutralizing antibodies more efficiently than did HA-E. Coli-K, perhaps because the protein expressed in the mammalian expression system was appropriately posttranslationally modified and/or because HA-E. Coli-K is a truncated form of HA (HA1).

Table 2. Hemagglutination inhibition and neutralization antibody titers induced in ferrets immunized with the expressed HAs
 HI titerNeutralization titer
  1. ND, not detected.
HA-Bac-K1:160 ± 01:80 ± 01:160 ± 0
HA-E. Coli-K1:160 ± 01:80 ± 0ND

The lung pathology of the animals was examined after immunization to confirm the safety of the vaccine. Mice and ferrets were injected twice with the expressed HAs at a 2 week interval. Ten days after the second immunization, the mice and ferrets were killed and their lungs collected for pathological analysis. No inflammation was observed in the lungs of either the mice or the ferrets (Fig. 2). The lungs of the mice treated with HA-Bac-K, HA-Bac-C, or HA-E. Coli-K did not differ from those of normal controls (Fig. 2a). As in the mice, no inflammation was observed in the lungs of the ferrets treated with HA-Bac-K or HA-E. Coli-K (Fig. 2b). All immunized mice and ferrets also had normal body temperatures and active movement (data not shown). Thus, none of the purified HAs induced inflammation in the animals tested.

Figure 2.

Histopathological findings in test animals' lungs after immunization with the expressed HA proteins. (a) Mice were immunized twice, at an interval of 2 weeks, with the expressed HAs. Ten days after the second immunization, the mice were killed and their lungs collected for immunohistochemical analysis. (b) Ferrets were immunized twice, at an interval of 2 weeks, with the expressed HAs. Ten days after the second immunization, their lungs were collected. All the lungs collected from the mice and ferrets were fixed in paraffin, stained with hematoxylin and eosin (HE) and their pathology assessed. No evidence of inflammation was found.

In our hands, although HA-Bac-C was HA expressed in a baculovirus system, it did not induce adequate immunogenicity against the influenza virus. However, HA-Bac-K induced sufficient HI and NT antibodies to protect against viral infection. This suggests that the purification protocol used for an expressed protein may be critical in maintaining the immunogenicity of a candidate subunit vaccine. In summary, the baculovirus expression system can be used for production of candidate vaccines directed against influenza virus.


This work was supported by grants from the Korea Centers for Disease Control and Prevention (2010-E43009-00), the Korean Healthcare Technology Research and Development Project of the Ministry of Health and Welfare (A103001), the Gyeonggi Regional Research Center of the Catholic University of Korea and the Mid-career Researcher Program through NRF funded by MEST (2010-0029242, K.H.K).


None of the authors has any conflict of interest associated with this study.