Proteomic analysis and immunodetection of the bovine milk osteopontin isoforms

The aim of this study was to characterize the osteopontin (OPN) secreted in bovine milk and to determine whether the different forms are the product of spliced variants. Spliced variants of the human gene and secreted osteopontin isoforms have been reported in human tumor tissue. In bovine milk, we identified 2 major forms: one corresponding to the full-length coding transcript and a truncated version of this form. No alternative spliced transcripts were detected in the lactating mammary gland tissue, in milk somatic cells, or in peripheral blood immune cells. The 60-kDa bovine osteopontin (bOPN) and a truncated 40-kDa protein isoform were confirmed by mass spectrometry and further characterized by immunoblotting using a panel of 6 antibodies targeting different domains of the protein. Of the 3 human anti-OPN antibodies targeting the N-terminal segment of the protein, only one detected all forms on sodium dodecyl sulfate-PAGE; one human anti-OPN antibody failed to detect bOPN, whereas the other detected only the 60-kDa protein, albeit barely in its phosphorylated form. Detection was generally more sensitive when the 60-kDa protein was dephosphorylated. Two polyclonal antibodies raised against bOPN were tested: one targeting the milk-purified bOPN (bOPN-121) and one targeting a bovine epitope (synthetic peptide) corresponding to a carboxy-terminal domain of the protein (bOPN-117). The bOPN-121 antibody detected all forms irrespective of the phosphorylation status of bOPN. The bOPN-117 and the mouse anti-human OPN (hOPN-4) antibodies, which recognized different domains of the carboxy-terminal segment of the protein, also preferentially detected the dephosphorylated 60-kDa protein. Whereas phosphorylation had a major effect on detection for several antibodies, deglycosylation slightly decreased immunodetection for the tested antibodies. In particular, phosphorylation is the major posttranslational modification that influenced the weak detection capacity of several antibodies. This fact needs to be taken into account for immunodetection of milk content. In conclusion, the OPN forms secreted in bovine milk are not the product of alternative splicing. The 40-kDa protein appears to be a truncated hypophosphorylated variant of the full-length 60-kDa form, which is highly phosphorylated. Together, the proteomic and immunoblotting analyses used to characterize bovine milk OPN revealed the complex nature of the bovine milk OPN forms.