Arrays were analysed on a PCS4000 ProteinChip Reader using the Protein Chip software version 3.0.6 (Ciphergen Biosystems, Inc., BEZ235 nmr Fremont, CA). The protocol averaged 10 laser shots per pixel with a focus mass of 24,000 Da, a matrix attenuation of 1000 Da and a range of 0–200,000 Da. The All-in-1 Protein Standard II (BioRad) was analysed on an NP20 array using the same analysis protocol. The following peaks were identified in the resulting spectrum and used to create

an internal calibration: hirudin BHVK (6964.0 Da), bovine cytochrome c (12230.92 Da), equine cardiac myoglobin (16951.51 Da) and bovine carbonic anhydrase (29023.66 Da). This internal calibration was applied to the spectra as an external calibration. The presented data are baseline subtracted and normalized by total ion current. Peaks with a signal-to-noise (S/N) ratio below 7 were not considered in subsequent analysis. FMDV antigen concentrated by PEG6000 precipitation is normally used for LBH589 clinical trial vaccine preparation. Such crude antigen preparations contain many proteins, most of

which are presumably derived from the BHK-21 cells used for virus propagation, as can be revealed by SDS-PAGE analysis (Fig. 1) of strains O1 Manisa (lane 2), Asia 1 Shamir (lane 4) and A24 Cruzeiro (lane 6). When the FMDV antigen of these strains is further purified by ultracentrifugation through a sucrose cushion it predominantly consists of three proteins migrating at about 23–25 kDa (Fig. 1, lanes 3, 5 and 7) which presumably represent VP1, VP2 and VP3. To facilitate the identification of the spectral peaks corresponding to the FMDV structural proteins

we used these purified antigens in SELDI-TOF-MS analysis employing NP20 arrays, which binds all proteins (Fig. 2a–c). The spectral peaks found were compared to the molecular masses predicted by translation of the RNA sequences (Table 1). For all three strains the peak at 9.0 kDa corresponds to myristoylated VP4, the peak at 23.2–23.3 kDa corresponds to VP1 and the peak at 24.5–24.9 kDa corresponds to VP2. Since these peaks are quite broad an accurate determination of their molecular mass is difficult. The molecular mass of VP3 is predicted to be intermediate between VP1 and VP2 (Table 1). A peak corresponding Astemizole to VP3 is more difficult to identify. Only in the profile of strain O1 Manisa a small peak can be seen at 24.1 kDa that could represent VP3 (Fig. 2c). The peak at 48 kDa that is observed with strain O1 Manisa but not with the two strains of other serotypes corresponds quite well to a VP1–VP2 dimer (Fig. 2c). For each serotype we also observe peaks of lower height at a normalized mass (m/z) of about 11.6 and 12.2 kDa, which is half the molecular mass of VP1 and VP2, and therefore represents double protonated forms of these proteins. For all three strains a repetitive pattern of peaks that differ by about 24 kDa is present in the molecular range above 50 kDa.