Expression of recombinant HA1 protein for specific detection of influenza A/H1N1/2009 antibodies in human serum



Yan Li, National Microbiology Laboratory, Public Health Agency of Canada, 1015 Arlington Street, Winnipeg, Manitoba, Canada R3E 3R2. Tel: +204 789 6045; fax: 204 789 2082; email:


The hemagglutinin genes (HA1 subunit) from human and animal 2009 pandemic H1N1 virus isolates were expressed with a baculovirus vector. Recombinant HA1 (rHA1) protein-based ELISA was evaluated for detection of specific influenza A(H1N1)pdm09 antibodies in serum samples from vaccinated humans. It was found that rHA1 ELISA consistently differentiated between antibodies recognizing the seasonal influenza H1N1 and pdm09 viruses, with a concordance of 94% as compared to the hemagglutination inhibition test. This study suggests the utility of rHA1 ELISA in serosurveillance.

List of Abbreviations:

confidence interval




hemagglutination inhibition






0.01 M PBS pH 7.4 containing 0.05% Tween-20

Influenza A viruses belong to the family Orthomyxoviridae [1]. Their genome is segmented, consisting of eight single-stranded, negative sense RNA molecules, which encode 11 viral proteins [2]. Influenza A viruses are classified into different subtypes according to their membrane glycoproteins, HA and NA, which are very important for induction of antibody responses in hosts. According to the distinct antigenicity of HA and NA molecules, 16 HA subtypes (H1–H16) and 9 NA subtypes (N1–N9) have been identified so far. Since the April 2009 global outbreak, the swine-origin 2009 A(H1N1) influenza virus has spread rapidly to numerous countries, resulted in a declaration of an influenza pandemic by the World Health Organization on 11 June 2009. The novel H1N1 influenza strain, designated as A(H1N1)pdm09, is a triple-reassortant swine influenza A (H1) virus containing genes from avian, swine, and human influenza viruses [3], [4]. This event highlights once again the unpredictability of influenza pandemics and the importance of global influenza surveillance among humans and animals. Although it is difficult to determine the exact number of A(H1N1)pdm09 cases worldwide, it is estimated that millions of people were infected between April and December 2009. By the end of 2009, the WHO reported that 12,000 people had died from A(H1N1)pdm09 infection, half of them in North America (Mexico, Canada and the USA) [5].

Serological methods have been widely used for influenza surveillance, vaccine development and evaluation, and sometimes for detection of novel influenza virus infections in humans and animals [6]. HA, the most antigenic surface protein, mediates attachment to host cell receptors and virus entry. Therefore, it is the most important antigen against which neutralizing antibodies are directed and is a crucial component of current vaccines. The HI test is considered the gold standard and has been widely used for detection of strain-specific serum antibodies to influenza. Recently, the MN method has been used more frequently due to its sensitivity in detection of antibodies to avian influenza A (H5N1) virus in human serum [7]. However, in sero-epidemiological studies, it is common to analyze hundreds to thousands of serum samples. Using standard HI or MN assays demands an enormous amount of the virus. When highly pathogenic or novel influenza strains are involved, specialized laboratories with stronger containment facilities and specially trained personnel are required for safe implementation of virus culture and testing [5]. In contrast, ELISA is a suitable alternative for seroepidemiological surveillance testing because it is highly sensitive and easy to perform [5], [8]. In this study, we describe the development and use of an indirect ELISA method that is based on the use of recombinant HA1 proteins of the 2009 H1N1 viruses isolated from humans in Mexico and from swine in Canada and that specifically recognizes A(H1N1)pdm09 virus antibodies in human serum samples. The Research Ethics Board at Mount Sinai Hospital approved this study before we enrolled any participants.

Genetic analysis has shown that the A(H1N1)pdm09 HA protein has a low (79.6%) amino acid identity with the HA of seasonal H1N1 virus, A/Brisbane/59/2007. Using virus-specific antiserum produced in ferrets, Rowe et al. performed an antigenic comparison of these viruses by HI and MN assays [7]. As shown in Table 1, antiserum raised against the A(H1N1)pdm09 strain does not cross-react with the seasonal human influenza A (H1N1) strain, A/Brisbane/59/07 and vice versa. These results clearly demonstrate that A(H1N1)pdm09 virus is antigenically distinct from seasonal human H1N1 [4]. Therefore, researchers PCR amplifed the HA1 genes from two North American pandemic H1N1 2009 isolates, A/Mexico/InDRE4487/2009 and A/swine/Alberta/OTH-33–21, cloned them into the pBlueBac4.5 vector (Invitrogen, Burlington, ON, Canada) and expressed them with a baculovirus expression system in Spodoptera frugiperda (SF21) insect cells [9], [10]. The sequence of each cloned HA1 gene was identical to its corresponding original isolate. The HA1 gene is 1032 bases long, encoding a protein of 344 amino acids with a predicted molecular mass of 39 kDa. This HA1 gene conserves all antigenic sites previously reported for HA [11] and 6 N-glycosylation sites, which are normally viewed as playing a key role in antigen-antibody recognition.

Table 1. Serologic cross-reactivity between pandemic 2009 H1N1 and seasonal H1N1 viruses by using antisera from experimentally infected ferrets.
Seasonal H1N1 virus A/Brisbane/59/200712801280<10<10<10<10
Pandemic H1N1 virus A/Mexico/InDRE4487/2009<10<1025606405120640
Pandemic H1N1 virus A/California/07/2009<10<10512012802560640

We analyzed expression of H1N1-HA1 protein by Coomassie blue staining of 12% SDS-PAGE (1a). We observed a strongly stained protein band migrating at 42–55 kDa in the lysates of SF21 cells infected with recombinant AcNPV (1a, lanes 1 and 2). We confirmed the antigenicity and identity of the recombinant HA1 protein by western blotting (1b). The 42–55 kDa protein band was specifically recognized by A(H1N1)pdm09 HA-specific monoclonal antibody raised from recombinant HA1 protein of A/California/06/2009 strain, which is consistent with predication of potential N-linked glycosylation and likely reflects the different usage of these sites during post-translational modification in insect cells [10]. In contrast, the corresponding protein was not present in cells infected by wild-type AcNPV or in mock-infected cells (1b, lanes, Wt and Cell). We also evaluated the antigenicity of these proteins by western blot with ferret antiserum raised against the pandemic H1N1 strain (A/Mexico/InDRE4487/2009) and seasonal human H1N1 virus (A/Brisbane/59/2007). Both human and swine recombinant HA1 proteins were specifically recognized by ferret antiserum against the pandemic 2009 H1N1 virus and we observed no cross-reactivity with ferret antiserum specific for seasonal A/Brisbane/59/2007 (data not shown).

Figure 1.

Construction and expression of recombinant plasmid pBlueBac4.5–2009H1N1-HA1. A 1032 bp segment containing the HA1 genes of H1N1(2009) was amplified and cloned into baculovirus transfer vector pBluBac4.5 under the control of polyhedrin promoter. The primer set used for amplication is indicated by arrows. (a) The HA1 proteins were localized by Coomassie blue staining of the SDS-PAGE and (b) confirmed by western blot reactivity with a specific Mab that was generated from recombinant HA1 protein of A/California/06/2009 strain. M, marker protein; lane 1, recombinant virus Ac-Bac-2009H1N1Mexico-HA1 infected cells; lane 2, recombinant virus Ac-Bac-2009H1N1swine-HA1 infected cells; Wt, wild-type AcNPV-infected cells; Cell, uninfected SF21 cell control. The arrows indicate the locations of expressed HA1 protein. Molecular weight standards (M) are given in kDa.

Figure 2.

ELISA reactivity of antisera with recombinant HA1 proteins. The baculovirus expressed recombinant HA1 proteins from A/Mexico/InDRE4487/2009 (panel a) and A/swine/Alberta/OTH-33–21/2009 (panel b) were used as the coating antigens with antiserum (1:200 dilution) against A/Mexico/InDRE4487/2009 (black bar), A/California/7/09 (grey bar) and A/Brisbane/59/2007 (white bar). Results are representative of identical assays performed in duplicate. The prebleed nonspecific background control has been subtracted. OD, optical density.

Figure 3.

Specific evaluation of anti-A(H1N1)pdm09 antibodies in post-vaccination human serum samples by indirect ELISA. The recombinant HA proteins extracted from the cells infected by Ac-Bac-2009H1N1Mexico-HA1 were used as coat antigen. The post-vaccination serum samples were diluted 1:500 in 2.5% milk-PBST and allowed to react on a pre-blocked plate. Specific binding of HA1 proteins to anti-pdm09 antibodies is shown. The cutoff is indicated by a solid line. OD, optical density.

To assess whether the baculovirus-expressed HA1 protein could be used as an antigen for discriminating between antibodies recognizing seasonal influenza H1N1 and A(H1N1)pdm09, we extracted the recombinant protein from infected cells as described by Luo et al. [12] and evaluated it for its suitability as a coating antigen in an indirect ELISA. Briefly, we coated a 96-well plate (Nunc-Immuno, Mississauga, ON, Canada) with the pdm09 HA1 protein in 0.01 M PBS (pH 7.4), 500 ng/well, and incubated it overnight at 4°C. After washing with PBST, we blocked the plate with 5% dry milk. We diluted ferret sera against A/Mexico/InDRE4487/2009, A/California/07/09 and A/Brisbane/59/2007 in PBST containing 2.5% dry milk and placed 100 μL of each dilution in wells. After incubation for 1 hr at 37°C, we extensively washed the plate, added horseradish peroxidase-labelled anti-ferret IgG (Mandel, Guelph, ON, Canada) at a pre-determined dilution to each well and incubated them for 1 hr. After a final washing step, we visualized the reactions by adding substrate solution (Sure Blue Reserve TMB microwell peroxidase substrate; KPL, Gaithersburg MD, USA) which we stopped after 10 mins by adding TMB stop solution (KPL). We measured absorbance values at 450 nm with a multichannel spectrophotometer (SPECTRA Max Plus, Sunnyvale, CA, USA). Consistent with the findings of western blot analysis, baculovirus-expressed HA1 proteins were specifically recognized by pdm09 H1N1-specific antibodies (Fig. 2).

To further evaluate the utility of recombinant HA1-based ELISA for specific detection of A(H1N1)pdm09 antibodies, we used indirect ELISA to test 40 pairs of pre- and post-immunization sera from healthy subjects enrolled in a 2009/10 immunogenicity trial in which they were vaccinated with Arepanrix (GSK, Mississauga, ON, Canada). Sera had been collected at baseline and 21 days after immunization. We calculated the cutoff value for IgG anti-A(H1N1)pdm09 antibodies as the mean of five negative control sera plus three standard deviations. We determined the optimal antigen concentration and antibody dilution by checkerboard titration. We considered samples positive when the corrected optical density was above the cutoff (A450 > 0.78). Of the 40 pre-vaccination serum samples tested, ELISA scored 34/37 HI-negative samples (HI titers < 1:20) as negative (data not shown), giving 92% (95% CI 83, 100) specificity in comparison to the gold-standard HI test. Three pre-vaccination serum samples showed HI titers of 80, 320, and 640, respectively, indicating that the subjects had possibly been exposed to the A(H1N1)pdm09 virus before vaccination. As evident from the results (Fig. 3), the ELISA correlated well with the HI assay for the detection of virus-specific antibodies. Of the 33 HI-positive post-immunization sera with titers ≥ 1:40, ELISA scored 31 as positive, giving 94% (95% CI 86, 100) sensitivity and two as negative (Fig. 3, #10 and #35). In contrast, ELISA also scored all seven HI-negative samples as negative (Fig. 3, #7, #19, #21, #23, #24, #25, #26).

To further examine any potential cross-reactivity between pandemic and seasonal H1N1 sera, we used recombinant HA1-based ELISA to screen 36 pairs of human sera collected in late 2008 from people vaccinated with the 2008/09 trivalent inactivated influenza vaccine, which includes the A/Brisbane/59/2007 (H1N1)-like strain. All serum samples, including those with HI-positive seasonal H1N1 antibodies (titers of ≥ 1:40) failed to cross-react with the A(H1N1)pdm09 pandemic HA1 protein in both ELISA and HI tests (data not shown), suggesting that the HA1-based ELISA described here is both specific and sensitive.

In the present study, we present a virus-free ELISA method that allows specific recognition of influenza A(H1N1)pdm09 antibodies in human sera. In comparison with commonly used methods, this new method is rapid and does not require the use of viruses and poultry erythrocytes. Personnel with standard laboratory skills could easily reproduce and implement it in most laboratory settings. It would also be possible to extend the same protocol to various animal species to explore the results of exposure of various animals to the virus. Taken together, the rHA1 protein-based ELISA could serve as a useful tool for large epidemiological and clinical influenza studies.


We thank Drs. Yohannes Berhane and John Pasick for providing recombinant HA DNA plasmid of a swine isolate A/Swine/Alberta/OTH-33–21/2009. This study was support by the Canadian Food Inspection Agency and the Public Health Agency of Canada.


No authors have any conflict of interest.