It has been reported that the sequence Tyr20-X-Arg-Phe23 present within the cytoplasmic tail of the transferrin receptor may represent a tyrosine internalization signal (Collawn, J.F., Stangel, M., Kuhn, L.A., Esekogwu, V., Jing, S., Trowbridge, I.S., and Tainer, J. A. (1990) Cell 63, 1061-1072). However, as Tyr20 is not conserved between species (Alvarez, E., Girones, N., and Davis, R. J. (1990) Biochem. J. 267, 31-35), the functional role of the putative tyrosine internalization signal is not clear. To address this question, we constructed a series of 32 deletions and point mutations within the cytoplasmic tail of the human transferrin receptor. The effect of these mutations on the apparent first order rate constant for receptor endocytosis was examined. It was found that the region of the cytoplasmic tail that is proximal to the transmembrane domain (residues 28-58) is dispensable for rapid endocytosis. In contrast, the distal region of the cytoplasmic tail (residues 1-27) was found to be both necessary and sufficient for the rapid internalization of the transferrin receptor. The region identified includes Tyr20-X-Arg-Phe23, but is significantly larger than this tetrapeptide. It is therefore likely that structural information in addition to the proposed tyrosine internalization signal is required for endocytosis. To test this hypothesis, we investigated whether a heterologous tyrosine internalization signal (from the low density lipoprotein receptor) could function to cause the rapid endocytosis of the transferrin receptor. It was observed that this heterologous tyrosine internalization signal did not allow rapid endocytosis. We conclude that the putative tyrosine internalization signal (Tyr20-Thr-Arg-Phe23) is not sufficient to determine rapid endocytosis of the transferrin receptor. The data reported here indicate that the transferrin receptor internalization signal is formed by a larger cytoplasmic tail structure located at the amino terminus of the receptor.