Antigenic characterization of recombinant hemagglutinin proteins derived from different avian influenza virus subtypes

Abstract

Since the advent of highly pathogenic variants of avian influenza virus (HPAIV), the main focus of avian influenza research has been the characterization and detection of HPAIV hemagglutinin (HA) from H5 and H7 subtypes. However, due to the high mutation and reassortation rate of influenza viruses, in theory any influenza strain may acquire increased pathogenicity irrespective of its subtype. A comprehensive antigenic characterization of influenza viruses encompassing all 16 HA and 9 neuraminidase subtypes will provide information useful for the design of differential diagnostic tools, and possibly, vaccines. We have expressed recombinant HA proteins from 3 different influenza virus HA subtypes in the baculovirus system. These proteins were used to generate polyclonal rabbit antisera, which were subsequently employed in epitope scanning analysis using peptide libraries spanning the entire HA. Here, we report the identification and characterization of linear, HA subtype-specific as well as inter subtype-conserved epitopes along the HA proteins. Selected subtype-specific epitopes were shown to be suitable for the differentiation of anti-HA antibodies in an ELISA.

Figure 1. Construction and purification of recombinant AIV HA.

Schematic drawing of the full-length HA protein and the relative location of the domains (black boxes) used for recombinant protein expression. The honeybee melittin (HBM) secretion signal (SS) is shown in grey. (B) Schematic illustration of recombinant HA proteins, shown as fusion of subtype-specific HA domains with the AIV ss or the HBM ss and a C-terminal 6xHis tag. (C-E) Western blot and SDS-PAGE analyses of cell culture supernatant (ccs) after Ni-NTA affinity chromatography. All proteins were secreted in the ccs in an uncleaved form and purified following an identical purification procedure. Differences in the number of positive elution fractions in the Western blot result from different quantities of recombinant protein bound to the Ni-NTA-column. SDS-PAGE results indicated that the most concentrated elution fractions contained almost exclusively the purified HA. Purification fractions are indicated with numbers, F =  flow through fraction, L =  load fraction.

Figure 2. Western blot analyses with pre-immune sera and antisera derived from immunized rabbits.

Unpurified recombinant H5 HA, H4 HA and H12 HA and a C-terminal 6x His tagged porcine IFN were blotted onto nitrocellulose membranes and analyzed with a monoclonal antibody against the His tag (A) or with serum from the 2nd bleeding (56 days post immunisation) from rabbits, immunized either with purified recombinant HA H5 (B), H4 (C) or H12 (D), respectively. Purified recombinant H5, H12 and H4 HA were blotted onto nitrocellulose membranes and analyzed with pre-immune rabbit sera (PI), a monoclonal antibody against the His tag and serum from the 3rd bleeding (114 days post infection) from rabbits, immunized with purified recombinant HA H5 (E), H4 (F) or H12 (G).

Figure 3. Visualization of data obtained from the peptide scanning analyses.

The signal intensity of each peptide on the dot blot membrane is shown as integrated intensity after subtraction of unspecific background and secondary antibody-related signals. Relative reactivity values are normalized by setting the highest value of each experiment at 100%. Sera were tested on membranes containing their homologous antigen. Examples of H5 HA dot blot and homologous reactivity patterns obtained by 3-fold probing and stripping of the membrane are shown in A and B, respectively. Cross-reactivity of sera was identified by testing each serum with each heterologous membrane; a representative plot obtained with sera against H5, H4 and H12 on the membrane containing H5-specific peptides is shown in C. Furthermore, cross-reactive epitopes were also identified by testing each of the sera on their respective heterologous membranes, as shown for serum against H5 on membranes representing H4 or H12 (D).

Figure 4. Principle of generation of semi-quantitative antigenic maps.

Integrated intensities from peptide scanning analyses were transferred to the aa sequence as shown here for the HA aa sequence from HPAIV A/tufted duck/Switzerland/V504/06(H5N1) within the sequence range from aa position 20 to 63, as indicated. (A) Semi quantitative display of the gross signal intensities in color-coded categories: 100%–40% (red), 40%–30% (yellow). (B) Semi-quantitative display of the net signal intensities in color code categories: 100%–70% (red), 70%–40% (green) and 40%–20% (light blue).

Figure 5. Antigenic maps of AIV HA domains reacting with their homologous sera.

Color code indicates (A) the gross signal intensities found (compare to Fig. 4A, 100%–40% (red) and 40%–30% (yellow)) and (B) the net signal intensities found (compare to Fig. 4B, 100%–70% (red), 70%–40% (yellow) and 40%–20% (light blue)).

Figure 6. Subtype-specific epitopes.

Integrated intensities from peptide scanning were transferred to the aa sequence of each AIV HA after subtraction of the heterologous from the homologous sera signals (H5, light blue; H4, green; H12, pink). All remaining epitopes are shown independent of their signal intensity in the homologous system.

Figure 7. Antibody differentiation with polyclonal animal sera against different AIV HA subtypes tested in ELISA.

(A) HA subtype-specific biotinylated peptide antigens (H5: biotin-Ttds-ANNSTEQVDTIMEKNVTVTHAQD-OH; H4: biotin-Ttds-DSEMNKLFERVRRQLRENAED KGNGCF-OH and for H12: biotin-Ttds-FTWAIHHPPTSDEQV-OH) were coated on ELISA plates at indicated amounts and tested with the 3 rabbit antisera (diluted 1∶500); H5 antigen (A), H4 (B), and H12 (C). Antigens H4, H5 and H12 HA were also tested with a chicken serum against H5 HA (diluted 1∶10) (D).