N-glycation is a protein modification that occurs more often in, for example, antibodies [20]. Alternatively it could represent heterogeneity of VP1 due to N-terminal proteolysis. A 48-kDa VP1-VP2 dimer was observed in strain O1 Manisa but not in strains of other serotypes. This must represent a disulfide-bonded dimer since only O serotype strains contain a disulfide bond between cysteine 134 of VP1 and cysteine 130 of VP2 [14]. This is confirmed by analysis of tryptic digestion fragments. Trypsin treatment of FMDV strain

O1 Kaufbeuren results in cleavage of the VP1 C-terminus after residue 200 and cleavage in an exposed loop of Volasertib molecular weight VP1, known as the GH-loop, after residues 145 and 154 [17]. We observed cleavages at the same positions in SELDI-TOF-MS experiments of trypsin-treated FMDV O1 Manisa. We also observed a tryptic digestion fragment of 40.0 kDa corresponding to a VP1 degradation product linked to VP2. This confirms the presence of a VP1–VP2 dimer. The spectral peak corresponding to VP2 was predominantly identified based on its mass and because of its specific presence after immunocapture with FMDV specific VHHs. In trypsin digestion experiments we observed two peaks that suggested partial cleavage after VP2 residue 167 both in its single and its VP1 disulfide-bonded form. VP2 cleavage at this position is to our knowledge not observed before. The spectral learn more peak corresponding

to VP3 is more difficult to identify since it is predicted to have a mass intermediate between VP1 and VP2. Occasionally a peak of low height that could represent VP3 is detectable in SELDI-TOF-MS profiles (e.g. Fig. 2c). Furthermore, when the VP1 peaks until are absent due to trypsin treatment a peak at 24.0 kDa that could represent VP3 is visible. However, this peak has a lower height than the VP1 and VP2 peaks. This is unexpected since VP1–VP3 are present in equimolar amounts in FMDV particles [1]. VP3 of all FMDV serotypes is known to form disulfide bonds to other VP3 molecules [1]. Peaks that could

represent multimerized VP3 are readily visible in the spectra of all three FMDV strains, which could explain the low height of the putative VP3 monomer peak. Alternatively, the low height of the putative VP3 peak could be due to less efficient ionization of VP3. We used SELDI-TOF-MS analysis for the characterization of FMDV antigen during various stages of vaccine preparation. In FMDV antigen preparations we could readily detect PEG6000 and BSA as well as many other proteins that presumably originate from the BHK-21 cells used for viral propagation. Especially the ability to detect PEG6000 could be of use since this non-protein compound is more difficult to detect by other methods. We also observed some limited proteolytic degradation of VP1 when FMDV antigen was stored at the elevated temperature of 35 °C, but not when antigens were properly stored at 4 °C.