The signal peptide is a critical component in the secretory expression of protein in eukaryotic cells. It has been verified that the signal peptide of mouse nerve growth factor could mediate the secretory expression of beta-endorphin in cultured non-neuronal cells. Although there is a counterpart of nerve growth factor in human genome, no research about the signal sequence from human genome has been reported. The function of mediating secretory expression is affected by many factors. We assumed that the counterpart from human genome could function as the signal peptide from mouse nerve growth factor does and these two signal sequences had different efficiency in mediating secretory expression of beta-endorphin, but we could not figure out which one had a better function. To validate our hypothesis and give an answer to the question, we constructed two eukaryotic vectors, pcDNA3.1-hEP and pcDNA3.1-mEP, containing human and mouse signal sequences in fusion genes, respectively. RT-PCR showed that the constructed fusion genes were expressed in NIH3T3 cells. We also found that the detected beta-endorphin by the immunofluorescent technique was mainly in the cytoplasm of NIH3T3 cells. The concentration of beta-endorphin in the culture medium by RIA is 280.33 +/- 24.16 (pg/ml) and 191.04 +/- 7.96 (pg/ml) from pcDNA3.1-hEP and pcDNA3.1-mEP, respectively, and there was a significant statistical difference between them (P < 0.05). A difference existed between them and that from blank vector individually (P < 0.01). These findings suggest that our constructed fusion gene containing the signal sequence of human nerve growth factor can be secretorily expressed and the efficiency of the signal peptide from human nerve growth factor is higher than that of mouse signal peptide.