Localization of TGN38 to the trans-Golgi network: involvement of a cytoplasmic tyrosine-containing sequence

Whereas the previous experiments suggested that TGN localization and endocytosis determinants were related but not identical, we were still intrigued by the possibility that the observed localization of the Tac-TGN38 chimeras could sim-ply correspond to the distribution of any molecule internalized by virtue of a tyrosine-containing motif. This prompted us to compare the distribution of T-G-G with that of the transferrin receptor, which also carries a cytoplasmic tyrosine-containing internalization motif (YTRF). The recycling pool of transferrin receptors in CV-I cells was marked by incubation of intact cells for 2 h at 37~ with rhodamineconjugated transferrin. Under these conditions, intracellular transferrin would be expected to be bound to recycling receptors (Klausner et al., 1983; Dautry-Varsat et al., 1983) , and to be localized mainly to early endosomes (Hopkins, 1983) , with smaller amounts present in coated vesicles and tubules in the area of TGN (Willingham and Pastan, 1985; Fishman and Fine, 1987; Stoorvogel et al., 1988) . Cells were then fixed, permeabilized, and stained for T-G-G using an antibody to Tac and a fluorescein-conjugated second antibody (Fig. 8) . No internalized transferrin could be detected in the central TGN structure containing T-G-G using this methodology; instead, transferrin was found within vesicles distributed throughout the cytoplasm (Fig. 8, a and b) . As mentioned earlier, in some cells T-G-G was also found in cytoplasmic vesicles; the presence of internalized transferrin in the same vesicles identified them as early endosomal structures (Fig. 8, a and b) . Similarly, the distribution of endogenous transferrin receptors detected with specific antibodies differed from that of T-G-G in the TGN of human RD4 cells (data not shown).