Many melanoma epitopes are presented to cytotoxic T-lymphocytes (CTLs) by major histocompatibility complex (MHC) class I molecules, and it is reasonable to expect that the epitopes would be good substrates for the transporter associated with antigen processing (TAP), as TAP plays a major role in the transport of peptides into the endoplasmic reticulum (ER) for binding to MHC class I molecules. However, we have previously shown that several melanoma-associated epitopes, such as those derived from tyrosinase, gp100, MAGE-1 and MAGE-2 antigens, are in fact poor substrates for TAP. During the process of determining why these epitopes were capable of eliciting a strong CTL response, yet were poor substrates for TAP, it was observed that the epitopes possessed amino acids at their N-terminus that were deleterious for TAP binding as described for the peptide-binding motif for human TAP. We therefore postulated that the epitopes were transported by TAP as longer precursor molecules, and then trimmed in the ER to the appropriate size for presentation to T-cells. In an effort to test this hypothesis we synthesized a set of peptides, derived from the tyrosinase (YMNGTMSQV) and MAGE-1 (EADPTGHSY) epitopes, which possess N-terminal extensions of up to four amino acids. We show here that the longer peptides are indeed transported into the ER at a significantly higher level than the original epitopes. The data indicate that the longer melanoma-associated peptides are the preferred substrates for TAP, and further support the notion that peptides can be trimmed at the N-terminus in the ER during antigen processing.