Open reading frame mining identifies a TLR4 binding domain in the primary sequence of ECRG4.

The embedding of small peptide ligands within large inactive pre-pro-precursor proteins encoded by orphan open reading frames (ORFs) makes them difficult to identify and study. To address this problem, we generated oligonucleotide (< 100-400 base pair) combinatorial libraries from either the epidermal growth factor (EGF) ORF that encodes the > 1200 amino acid EGF precursor protein or the orphan ECRG4 ORF, that encodes a 148 amino acid Esophageal Cancer Related Gene 4 (ECRG4), a putative cytokine precursor protein of up to eight ligands. After phage display and 3-4 rounds of biopanning for phage internalization into prostate cancer epithelial cells, sequencing identified the 53-amino acid EGF ligand encoded by the 5' region of the EGF ORF and three distinct domains within the primary sequence of ECRG4: its membrane targeting hydrophobic signal peptide, an unanticipated amino terminus domain at ECRG4 and a C-terminus ECRG4 domain. Using HEK-blue cells transfected with the innate immunity receptor complex, we show that both ECRG4 and ECRG4 enter cells by interaction with the TLR4 immune complex but neither stimulate NFkB. Taken together, the results help establish that phage display can be used to identify cryptic domains within ORFs of the human secretome and identify a novel TLR4-targeted internalization domain in the amino terminus of ECRG4 that may contribute to its effects on cell migration, immune cell activation and tumor suppression.