Abstract
Translation of mRNA in alternate reading frames (ARF) is a naturally occurring process heretofore underappreciated as a generator of protein diversity. The MUC1 gene encodes MUC1-TM, a signal-transducing trans-membrane protein highly expressed in human malignancies. Here we show that an AUG codon downstream to the MUC1-TM initiation codon initiates an alternate reading frame thereby generating a novel protein, MUC1-ARF. MUC1-ARF, like its MUC1-TM 'parent' protein, contains a tandem repeat (VNTR) domain. However, the amino acid sequence of the MUC1-ARF tandem repeat as well as N- and C- sequences flanking it differ entirely from those of MUC1-TM. In vitro protein synthesis assays and extensive immunohistochemical as well as western blot analyses with MUC1-ARF specific monoclonal antibodies confirmed MUC1-ARF expression. Rather than being expressed at the cell membrane like MUC1-TM, immunostaining showed that MUC1-ARF protein localizes mainly in the nucleus: Immunohistochemical analyses of MUC1-expressing tissues demonstrated MUC1-ARF expression in the nuclei of secretory luminal epithelial cells. MUC1-ARF expression varies in different malignancies. While the malignant epithelial cells of pancreatic cancer show limited expression, in breast cancer tissue MUC1-ARF demonstrates strong nuclear expression. Proinflammatory cytokines upregulate expression of MUC1-ARF protein and co-immunoprecipitation analyses demonstrate association of MUC1-ARF with SH3 domain-containing proteins. Mass spectrometry performed on proteins coprecipitating with MUC1-ARF demonstrated Glucose-6-phosphate 1-dehydrogenase (G6PD) and Dynamin 2 (DNM2). These studies not only reveal that the MUC1 gene generates a previously unidentified MUC1-ARF protein, they also show that just like its 'parent' MUC1-TM protein, MUC1-ARF is apparently linked to signaling and malignancy, yet a definitive link to these processes and the roles it plays awaits a precise identification of its molecular functions. Comprising at least 524 amino acids, MUC1-ARF is, furthermore, the longest ARF protein heretofore described.