Table 1.
Viruses tested for stability during storage at 4°C and 27°C.
Table 2.
ELISA reagents used for capture ELISA.
Fig 1.
Development of capture ELISA to measure correctly folded HA stalk domain.
(A) and (B) describe the underlying concept of the assay. The assay was primarily developed to detect HA with conformationally intact stalk epitopes. The capture mAb in this assay recognizes the HA head linear epitope. On the other hand, the detection mAb recognizes the conformational epitope in the HA stalk. (A) If the stalk of the HA is properly folded, the detection mAb can bind and thus produces a signal. (B) If the stalk is misfolded, no signal is observed since the detection mAb cannot recognize the HA stalk. (C) HA concentration was calculated using a curve-fit model similar to parallel line analysis in which EC50 values are used to determine relative amounts. An EC50 analysis is shown using a recombinant protein standard with a known concentration for reference. The shift in EC50 between protein standard and test samples was then used to calculate the HA content of test samples. (D) The test samples were treated with 0.05% zwittergent for 1h to solubilize the influenza virus membrane and prevent transmembrane-driven rosette formation. EC50 values of virus preparations treated with or without zwittergent were calculated. Bars represent mean with error bars representing standard error of the mean.
Fig 2.
Selection and characterization of the detection mAb.
(A) Capture ELISAs were performed using three different detection mAbs including murine stalk mAbs KB2 (binds to H1, H5 and H6), 6F12 (pan-H1 binder), and human mAb CR9114 (pan-HA binder). The detection capability of these antibodies was measured using five different purified virus preparations: cH8/1Cal09N1, cH5/1 Cal09N1, DR13 (pandemic H1N1), NL09 (pandemic H1N1) and PR8 (H1N1). The amount measure with each mAb was normalized to the mean value of all three mAbs and expressed as percent of the mean. KB2 was selected for further assay development and characterized in more detail. (B) To understand if KB2 is sensitive to conformational changes of the HA a cH5/1Cal09ssN1 purified virus preparation was coated on ELISA plates and exposed to either PBS control, buffered solutions at pH4.4 (with and without reducing agent DTT) or at pH7 for one hour, and finally brought back to neutral pH prior to performing the ELISA. (C) Immunofluorescent staining of NL09 wildtype (WT) virus and escape mutant infected MDCK cells was done using KB2 and polyclonal serum against H1. NL09 was also passaged in the presence of irrelevant mAb as control. The right panel shows the structure of HA. Red denotes the globular head domain, while green denotes the stalk domain. Head antigenic sites are denoted in light red, and the highly conserved stalk epitopes are denoted in light green (FI6v3, F10, CR9114, CR6261) [23, 35, 36, 47]. The H45R position in the stalk of the KB2 escape mutant is denoted in blue (PDB 1RU7 [30]). (D) Competition ELISA indicates that KB2 and CR9114 compete for the same epitope in the stalk domain of the H1 HA.
Fig 3.
Effect of different stress treatments and group 2 stalk on assay performance.
Detergent treatment is necessary for the assay for optimal access antigenic sites on the HA trimers. In addition, conditions during the vaccine production process also require detergent and high salt concentrations. To test if these conditions influence the stalk conformation we treated recombinantly expressed HA protein (which is used as standard in the assay) with (A) 0.05% zwittergent and (B) 2% Triton X-100+340mM NaCl and compared the readout to the signal obtained with untreated HA. In addition, we tested the influence of different treatments on the concentrations of properly folded HA in purified viral preparations. We exposed the viral preparations to (C) low pH (pH 4 for 60 minutes) or (D) heat (100°C for 10 minutes) treatment. Post treatment HA concentrations were normalized to pre-treatment concentrations. (E) A capture ELISA was performed to measure the HA content of a group 2 cHA expressing virus, cH4/3N2, with a recombinant cH4/3 protein as standard.
Fig 4.
Stalk stability under different storage conditions.
HA content of crude cH5/1N1, cH8/1N1, cH11/1N1, cH12/1N1, wild type Cal09 and wild type DR13 preparations at 4°C and 27°C. The HA content of samples stored at 4°C was measured on days 0, 12, 30, 90 and 180; and HA content of samples stored at 27°C was measured on days 0, 12 and 30. In order to better understand the effects of the storage conditions on the stability of the two main HA domains (head and stalk), viruses were grouped based on either (A, B) head domain or (C, D) stalk domains. Data points in A and B represent an average of two replicates for the wild type strains and an average of three viruses (Cal09, Cal09ss, DR13) for cHA strains, while the data points in C and D represent an average of five viruses (cH5/1N1, cH8/1N1, cH11/1N1, cH12/1N1 and wild type) for Cal09 and DR13 stalk viruses and four viruses (cH5/1N1, cH8/1N1, cH11/1N1, cH12/1N1) for the Cal09ss stalk. The error bars represent the standard error of mean. All HA concentrations were normalized to concentrations measured on day 0.