Endomorphin-1 (EM-1) and endomorphin-2 (EM-2) represent two opioid active tetrapeptides with high affinity and selectivity for the μ-opioid (MOP) receptor. Both EM-1 and EM-2 exhibit strong inhibition of pain signals in the central nervous system (CNS). In contrast to these compounds, the undecapeptide substance P (SP) facilitates pain influx in the CNS. SP has been implicated in a number of functions in the central nervous system, including pain processing and reward. Its aminoterminal fragment SP_1–7 has been shown to modulate several actions of SP in the CNS, the nociceptive effect included. Although the actions of SP_1–7 have been known for long no specific receptor for the SP fragment has yet been cloned. In this study, we demonstrate the presence of specific binding sites for the heptapeptide in the rat spinal cord. The binding affinity for unlabeled SP_1–7 to the specific sites for the labeled heptapeptide highly exceeded those of SP and other C- or N-terminal fragments thereof. The NK-1, NK-2 and NK-3 receptor ligands [Sar⁹, Met(O₂)¹¹]SP, R396 and senktide, respectively, showed no or negligible binding. Moreover, both EM-1 and EM-2 were found to interact with SP_1–7 binding. However, a significant difference in binding affinity between the two opioid active tetrapeptides was observed. As recorded from replacement curves the affinity of EM-2 was 10 times weaker than that for SP_1–7 but about 100 times higher than that of EM-1. Among other Tyr-Pro-containing peptides Tyr-MIF-1 but not Tyr-W-MIF-1 exhibited affinity of similar potency as EM-2. These results strengthen the previously observed differences between EM-1 and EM-2 in various functional studies. Moreover, using a cell line (C6) expressing the MOP receptor it was shown that the labeled SP_1–7 did not interact with binding to this receptor and no functional response was seen for the SP heptapeptide on the MOP receptor by means of stimulation in the GTPγS assay. This suggests that the identified SP_1–7 binding sites, with high affinity also for EM-2, are not identical to the MOP receptor and apparently not to any of the known tachykinin receptors.