Transport of the lysosomal membrane glycoprotein lgp120 (lgp-A) to lysosomes does not require appearance on the plasma membrane

We have used stably transfected CHO cell lines to characterize the pathway of intracellular transport of the lgp120 (lgp-A) to lysosomes. Using several surface labeling and internalization assays, our results suggest that lgp120 can reach its final destination with or without prior appearance on the plasma membrane. The extent to which lgp120 was transported via the cell surface was determined by two factors: expression level and the presence of a conserved glycine-tyrosine motif in the cytoplasmic tail. In cells expressing low levels of wild-type lgp120, the majority of newly synthesized molecules reached lysosomes without becoming accessible to antibody or biotinylation reagents added extracellularly at 4 degrees C. With increased expression levels, however, an increased fraction of transfected lgp120, as well as some endogenous lgp-B, appeared on the plasma membrane. The fraction of newly synthesized lgp120 reaching the cell surface was also increased by mutations affecting the cytoplasmic domain tyrosine or glycine residues. A substantial fraction of both mutants reached the surface even at low expression levels. However, only the lgp120G----A7 mutant was rapidly internalized and delivered from the plasma membrane to lysosomes. Taken together, our results show that the majority of newly synthesized wild-type lgp120 does not appear to pass through the cell surface en route to lysosomes. Instead, it is likely that lysosomal targeting involves a saturable intracellular sorting site whose affinity for lgp's is dependent on a glycine-tyrosine motif in the lgp120 cytoplasmic tail.     

Sorting of mannose 6-phosphate receptors and lysosomal membrane proteins in endocytic vesicles

The intracellular distributions of the cation-independent mannose 6- phosphate receptor (MPR) and a 120-kD lysosomal membrane glycoprotein (lgp120) were studied in rat hepatoma cells. Using quantitative immunogold cytochemistry we found 10% of the cell's MPR located at the cell surface. In contrast, lgp120 was not detectable at the plasma membrane. Intracellularly, MPR mainly occurred in the trans-Golgi reticulum (TGR) and endosomes. lgp120, on the other hand, was confined to endosomes and lysosomes. MPR was present in both endosomal tubules and vacuoles, whereas lgp120 was confined to the endosomal vacuoles. In cells incubated for 5-60 min with the endocytic tracer cationized ferritin, four categories of endocytic vacuoles could be discerned, i.e., vacuoles designated MPR+/lgp120-, MPR+/lgp120+, MPR-/lgp120+, and vacuoles nonimmunolabeled for MPR and lgp120. Tracer first reached MPR+/lgp120-, then MPR+/lgp120+, and finally MPR-/lgp120+ vacuoles, which are assumed to represent lysosomes. To study the kinetics of appearance of endocytic tracers in MPR-and/or lgp120-containing pools in greater detail, cells were allowed to endocytose horse-radish peroxidase (HRP) for 5-90 min. The reduction in detectability of MPR and lgp120 antigenicity on Western blots, due to treatment of cell homogenates with 3'3-diaminobenzidine, was followed in time. We found that HRP reached the entire accessible pool of MPR almost immediately after internalization of the tracer, while prolonged periods of time were required for HRP to maximally access lgp120. The combined data suggest that MPR+/lgp120+ vacuoles are endocytic vacuoles, intermediate between MPR+/lgp120-endosomes and MPR-/lgp120+ lysosomes, and represent the site where MPR is sorted from lgp120 destined for lysosomes. We propose that MPR is sorted from lgp120 by selective lateral distribution of the receptor into the tubules of this compartment, resulting in the retention of lgp120 in the vacuoles and the net transport of lgp120 to lysosomes.     

Glycoproteins of the lysosomal membrane

Three glycoprotein antigens (120, 100, and 80 kD) were detected by mono- and/or polyclonal antibodies generated by immunization with highly purified rat liver lysosomal membranes. All of the antigens were judged to be integral membrane proteins based on the binding of Triton X-114. By immunofluorescence on normal rat kidney cells, a mouse monoclonal antibody to the 120-kD antigen co-stained with a polyclonal rabbit antibody that detected the 100- and 80-kD antigens as well as with antibodies to acid phosphatase, indicating that these antigens are preferentially localized in lysosomes. Few 120-kD-positive structures were found to be negative for acid phosphatase, suggesting that the antigen was not concentrated in organelles such as endosomes, which lack acid phosphatase. Immunoperoxidase cytochemistry also showed little reactivity in Golgi cisternae, coated vesicles, or on the plasma membrane. Digestion with endo-beta-N-acetylglucosaminidase H (Endo H) and endo-beta-N-acetylglucosaminidase F (Endo F) demonstrated that each of the antigens contained multiple N-linked oligosaccharide chains, most of which were of the complex (Endo H-resistant) type. The 120-kD protein was very heavily glycosylated, having at least 18 N-linked chains. It was also rich in sialic acid, since neuraminidase digestion increased the pI of the 120-kD protein from less than 4 to greater than 8. Taken together, these results strongly suggest that the glycoprotein components of the lysosomal membrane are synthesized in the rough endoplasmic reticulum and terminally glycosylated in the Golgi before delivery to lysosomes. We have provisionally designated these antigens lysosomal membrane glycoproteins lgp120, lgp100, lgp80.     

Multiple pathways of exocytosis, endocytosis, and membrane recycling: validation of a Golgi route

A number of pathways for intracellular membrane traffic have been detected in various cell types. The major established routes are: 1) the lysosomal pathway, which is the major route utilized in phagocytic and cultured cells; 2) the transcellular route, which represents the major type of traffic in nonfenestrated, capillary endothelial cells and which also appears to be the preferred route for the transport of immunoglobulins (intact) across cells; 3) the exocytosis pathway, utilized in secretory cells for discharge of secretory products, and which is also believed to be used for delivery of intrinsic membrane glycoproteins; 4) the plasmalemma to Golgi route, also highly developed in secretory cells, which is believed to be utilized for the recycling of secretory granule membranes; and 5) the biosynthetic pathways for transport of secretory products, lysosomal enzymes, and membrane proteins from the endoplasmic reticulum to the Golgi complex and for transport of lysosomal enzymes from the Golgi complex to lysosomes. It has become clear that cells repeatedly reutilize or recycle the membranes used in these various transport operations. Clathrin-coated vesicles have been found to be involved in transport along all these routes, which suggests that there are multiple populations of coated vesicles with different transport functions in every cell. It has become clear that the Golgi complex is the site where the membrane and product traffic converges and is sorted and directed to its correct destinations. The validation of a transport route from the cell surface to the Golgi complex raises the possibility that bound ligands and membrane constituents could be modified or repaired in transit during recycling through the Golgi complex, which is a biosynthetic compartment.     

Two lysosomal membrane proteins,LGP85 and LGP107, are delivered to late endosomes/lysosomes through different intracellular routes after exiting from the trans-Golgi network

Lysosomal membrane proteins are delivered from their synthesis site, the endoplasmic reticulum (ER) to late endosomes/lysosomes through the Golgi complex. It has been proposed that after leaving the Golgi they are transported either directly or indirectly (via the cell surface) to late endosomes/lysosomes. In the present study, we examined the transport routes taken by two structurally different lysosomal membrane proteins, LGP85 and LGP107, in rat 3Y1-B cells. Here we show that newly synthesized LGP85 and LGP107 are delivered to late endosomes/lysosomes via a direct route without passing through the cell surface. Interestingly, although LGP107 is delivered from the Golgi to early endosomes containing internalized horseradish peroxidase-conjugated transferrin (HRP-Tfn) en route to lysosomes, LGP85 does not pass through the HRP-Tfn-positive early endosomes. These results suggest, therefore, that LGP85 and LGP107 are sorted into distinct transport vesicles at the post-Golgi, presumably the trans-Golgi network (TGN), after which LGP85 is delivered directly to late endosomes/lysosomes, but significant fractions of LGP107 are targeted to early endosomes before transport to late endosomes/lysosomes. This study provides the first evidence that after exiting from the Golgi, LGP85 and LGP107 are targeted to late endosomes/lysosomes via a different pathway.     

Microtubule perturbation retards both the direct and the indirect apical pathway but does not affect sorting of plasma membrane proteins in intestinal epithelial cells (Caco-2).

Endogenous plasma membrane proteins are sorted from two sites in the human intestinal epithelial cell line Caco-2. Apical proteins are transported from the Golgi apparatus to the apical domain along a direct pathway and an indirect pathway via the basolateral membrane. In contrast, basolateral proteins never appear in the apical plasma membrane. Here we report on the effect of the microtubule-active drug nocodazole on the post-synthetic transport and sorting of plasma membrane proteins. Pulse-chase radiolabeling was combined with domain-specific cell surface assays to monitor the appearance of three apical and one basolateral protein in plasma membrane domains. Nocodazole was found to drastically retard both the direct transport of apical proteins from the Golgi apparatus and the indirect transport (transcytosis) from the basolateral membrane to the apical cell surface. In contrast, neither the transport rates of the basolateral membrane nor the sorting itself were significantly affected by the nocodazole treatment. We conclude that an intact microtubular network facilitates, but is not necessarily required for, the transport of apical membrane proteins along the two post-Golgi pathways to the brush border.     

Studies on the mechanisms of autophagy: maturation of the autophagic vacuole

Data presented in the accompanying paper suggests nascent autophagic vacuoles are formed from RER (Dunn, W. A. 1990. J. Cell Biol. 110:1923-1933). In the present report, the maturation of newly formed or nascent autophagic vacuoles into degradative vacuoles was examined using morphological and biochemical methods combined with immunological probes. Within 15 min of formation, autophagic vacuoles acquired acid hydrolases and lysosomal membrane proteins, thus becoming degradative vacuoles. A previously undescribed type of autophagic vacuole was also identified having characteristics of both nascent and degradative vacuoles, but was different from lysosomes. This intermediate compartment contained only small amounts of cathepsin L in comparison to lysosomes and was bound by a double membrane, typical of nascent vacuoles. However, unlike nascent vacuoles vet comparable to degradative vacuoles, these vacuoles were acidic and contained the lysosomal membrane protein, lgp120, at the outer limiting membrane. The results were consistent with the stepwise acquisition of lysosomal membrane proteins and hydrolases. The presence of mannose-6-phosphate receptor in autophagic vacuoles suggested a possible role of this receptor in the delivery of newly synthesized hydrolases from the Golgi apparatus. However, tunicamycin had no significant effect on the amount of mature acid hydrolases present in a preparation of autophagic vacuoles isolated from a metrizamide gradient. Combined, the results suggested nascent autophagic vacuoles mature into degradative vacuoles in a stepwise fashion: (a) acquisition of lysosomal membrane proteins by fusing with a vesicle deficient in hydrolytic enzymes (e.g., prelysosome); (b) vacuole acidification; and (c) acquisition of hydrolases by fusing with preexisting lysosomes or Golgi apparatus-derived vesicles.     

Sorting of an internalized plasma membrane lipid between recycling and degradative pathways in normal and Niemann-Pick, type A fibroblasts

We examined the metabolism and intracellular transport of a fluorescent sphingomyelin analogue, N-(N-[6-[(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino]caproyl])- sphingosylphosphorylcholine (C6-NBD-SM), in both normal and Niemann-Pick, type A (NP-A) human skin fibroblast monolayers. C6-NBD-SM was integrated into the plasma membrane bilayer by transfer of C6-NBD-SM monomers from liposomes to cells at 7 degrees C. The cells were washed, and within 3 min of warming to 37 degrees C, both normal and NP-A fibroblasts had internalized C6-NBD-SM from the plasma membrane, resulting in a punctate pattern of intracellular fluorescence. Rates for C6-NBD-SM internalization and transport from intracellular compartments to the plasma membrane (recycling) were similar for normal and NP-A cells. With increasing time at 37 degrees C, internalized C6-NBD-SM accumulated in the lysosomes of NP-A fibroblasts, while normal fibroblasts showed increasing Golgi apparatus fluorescence with no observable lysosomal labeling. Since NP-A fibroblasts lack lysosomal (acid) sphingomyelinase (A-SMase), this result suggested that hydrolysis of C6-NBD-SM prevented its accumulation in the lysosomes of normal fibroblasts during its transport along the degradative pathway. We used the amount of C6-NBD-SM hydrolysis by A-SMase in normal cells as a measure of C6-NBD-SM transported from the cell surface to the lysosomes. After a lag period, C6-NBD-SM was delivered to the lysosomes at a rate of approximately 8%/h. This rate was approximately 18-19 fold slower than the rate of C6-NBD-SM recycling from intracellular compartments to the plasma membrane. Thus, small amounts of C6-NBD-SM were transported along the degradative pathway, while most endocytosed C6-NBD-SM was sorted for transport along the plasma membrane recycling pathway.     

Vitamin A excess alters membrane flow in rat liver

Pulse-chase methodology with [35S]methionine as label was employed to determine flow kinetics through the endoplasmic reticulum-Golgi apparatus-(lysosome-) secretory vesicle-plasma membrane export route in livers of animals receiving vitamin A excess by gavage. Overall fraction composition determined by morphometry and by analyses of marker enzymes was unchanged by vitamin administration. The vitamin modified the pattern of flow of proteins through the Golgi apparatus to the cell surface and to lysosomes. Altered flux was evidenced by a markedly reduced rate of labeling of lysosomes and a slightly increased rate of labeling of both total membrane proteins of the plasma membrane and of a specific membrane glycoprotein GP80. Also reduced was overall labeling of the Golgi apparatus. Differences in the rate or routes of trafficking of glycoproteins through the Golgi apparatus together with altered opportunities for processing might account for some of the alterations in glycoconjugate glycosylation associated with excess vitamin A administration.     

An endogenous MDCK lysosomal membrane glycoprotein is targeted basolaterally before delivery to lysosomes

Using surface immunoprecipitation at 37 degrees C to "catch" the transient apical or basolateral appearance of an endogenous MDCK lysosomal membrane glycoprotein, the AC17 antigen, we demonstrate that the bulk of newly synthesized AC17 antigen is polarly targeted from the Golgi apparatus to the basolateral plasma membrane or early endosomes and is then transported to lysosomes via the endocytic pathway. The AC17 antigen exhibits very similar properties to members of the family of lysosomal-associated membrane glycoproteins (LAMPs). Parallel studies of an avian LAMP, LEP100, transfected into MDCK cells revealed colocalization of the two proteins to lysosomes, identical biosynthetic and degradation rates, and similar low levels of steady-state expression on both the apical (0.8%) and basolateral (2.1%) membranes. After treatment of the cells with chloroquine, newly synthesized AC17 antigen, while still initially targeted basolaterally, appears stably in both the apical and basolateral domains, consistent with the depletion of the AC17 antigen from lysosomes and its recycling in a nonpolar fashion to the cell surface.     

The pathway and targeting signal for delivery of the integral membrane glycoprotein LEP100 to lysosomes

A complete set of chimeras was made between the lysosomal membrane glycoprotein LEP100 and the plasma membrane-directed vesicular stomatitis virus G protein, combining a glycosylated lumenal or ectodomain, a single transmembrane domain, and a cytosolic carboxyl-terminal domain. These chimeras, the parent molecules, and a truncated form of LEP100 lacking the transmembrane and cytosolic domains were expressed in mouse L cells. Only LEP100 and chimeras that included the cytosolic 11 amino acid carboxyl terminus of LEP100 were targeted to lysosomes. The other chimeras accumulated in the plasma membrane, and truncated LEP100 was secreted. Chimeras that included the extracellular domain of vesicular stomatitis G protein and the carboxyl terminus of LEP100 were targeted to lysosomes and very rapidly degraded. Therefore, in chimera-expressing cells, virtually all the chimeric molecules were newly synthesized and still in the biosynthesis and lysosomal targeting pathways. The behavior of one of these chimeras was studied in detail. After its processing in the Golgi apparatus, the chimera entered the plasma membrane/endosome compartment and rapidly cycled between the plasma membrane and endosomes before going to lysosomes. In pulse-expression experiments, a large population of chimeric molecules was observed to appear transiently in the plasma membrane by immunofluorescence microscopy. Soon after protein synthesis was inhibited, this surface population disappeared. When lysosomal proteolysis was inhibited, chimeric molecules accumulated in lysosomes. These data suggest that the plasma membrane/early endosome compartment is on the pathway to the lysosomal membrane. This explains why mutations that block endocytosis result in the accumulation of lysosomal membrane proteins in the plasma membrane.     

The mannose 6-phosphate receptor and the biogenesis of lysosomes

Localization of the 215 kd mannose 6-phosphate receptor (MPR) was studied in normal rat kidney cells. Low levels of receptor were detected in the trans Golgi network, Golgi stack, plasma membrane, and peripheral endosomes. The bulk of the receptor was localized to an acidic, reticular-vesicular structure adjacent to the Golgi complex. The structure also labeled with antibodies to lysosomal enzymes and a lysosomal membrane glycoprotein (lgp120). While lysosome-like, this structure is not a typical lysosome that is devoid of MPRs. The endocytic marker alpha 2 macroglobulin-gold entered the structure at 37 degrees C, but not at 20 degrees C. With prolonged chase, most of the marker was transported from the structure into lysosomes. We propose that the MPR/lgp-enriched structure is a specialized endosome (prelysosome) that serves as an intermediate compartment into which endocytic vesicles discharge their contents, and where lysosomal enzymes are released from the MPR and packaged along with newly synthesized lysosomal glycoproteins into lysosomes.     

Rapid redistribution of Golgi proteins into the ER in cells treated with brefeldin A: evidence for membrane cycling from Golgi to ER

In cells treated with brefeldin A (BFA), movement of newly synthesized membrane proteins from the endoplasmic reticulum (ER) to the Golgi apparatus was blocked. Surprisingly, the glycoproteins retained in the ER were rapidly processed by cis/medial Golgi enzymes but not by trans Golgi enzymes. An explanation for these observations was provided from morphological studies at both the light and electron microscopic levels using markers for the cis/medial and trans Golgi. They revealed a rapid and dramatic redistribution to the ER of components of the cis/medial but not the trans Golgi in response to treatment with BFA. Upon removal of BFA, the morphology of the Golgi apparatus was rapidly reestablished and proteins normally transported out of the ER were efficiently and rapidly sorted to their final destinations. These results suggest that BFA disrupts a dynamic membrane-recycling pathway between the ER and cis/medial Golgi, effectively blocking membrane transport out of but not back to the ER.     

Intracellular trafficking of lysosomal membrane proteins

Lysosomes are the site of degradation of obsolete intracellular material during autophagy and of extracellular macromolecules following endocytosis and phagocytosis. The membrane of lysosomes and late endosomes is enriched in highly glycosylated transmembrane proteins of largely unknown function. Significant progress has been made in recent years towards elucidating the pathways by which these lysosomal membrane proteins are delivered to late endosomes and lysosomes. While some lysosomal membrane proteins follow the constitutive secretory pathway and reach lysosomes indirectly via the cell surface and endocytosis, others exit the trans-Golgi network in clathrin-coated vesicles for direct delivery to endosomes and lysosomes. Sorting from the Golgi or the plasma membrane into the endosomal system is mediated by signals encoded by the short cytosolic domain of these proteins. This review will discuss the role of lysosomal membrane proteins in the biogenesis of the late endosomal and lysosomal membranes, with particular emphasis on the structural features and molecular mechanisms underlying the intracellular trafficking of these proteins.     

Targeting of membrane proteins to endosomes and lysosomes

The pathways involved in targeting membrane proteins to lysosomes are extraordinarily complex. Newly synthesized proteins in the ER are transported to the Golgi complex, and upon arrival at the trans Golgi network (TGN) are targeted either directly to endosomes, or first to the cell surface from where they can be rapidly internalized into the endocytic pathway for delivery to lysosomes. The routes to endosomes are specified by sorting motifs in the cytoplasmic tails of the proteins that are recognized at the TGN or plasma membrane. The molecular details of these processes are just emerging.     

Lysosome associated membrane protein 1 (Lamp1) traffics directly from the TGN to early endosomes

The precise trafficking routes followed by newly synthesized lysosomal membrane proteins after exit from the Golgi are unclear. To study these events we created a novel chimera (YAL) having a lumenal domain comprising two tyrosine sulfation motifs fused to avidin, and the transmembrane and cytoplasmic domains of lysosome associated membrane protein 1 (Lamp1). The newly synthesized protein rapidly transited from the trans- Golgi Network (TGN) to lysosomes (t(1/2) approximately 30 min after a lag of 15-20 min). However, labeled chimera was captured by biotinylated probes endocytosed for only 5 min, indicating that the initial site of entry into the endocytic pathway was early endosomes. Capture required export of YAL from the TGN, and endocytosis of the biotinylated reagent, and was essentially quantitative within 2 h of chase, suggesting that all molecules were following an identical route. There was no evidence of YAL trafficking via the cell surface. Fusion of TGN-derived vesicles with 5 min endosomes could be recapitulated in vitro, but neither late endosomes nor lysosomes could serve as acceptor compartments. This suggests that contrary to previous conclusions, most if not all newly synthesized Lamp1 traffics from the TGN to early endosomes prior to delivery to late endosomes and lysosomes.     

Effect of acidotropic amines on the accumulation of newly synthesized membrane and luminal proteins in Chinese-hamster ovary (CHO) cell lysosomes.

Lysosomes were isolated by sequential gradient centrifugation [Madden, Wirt & Storrie (1987) Arch. Biochem. Biophys. 257, 27-38] from control or acidotropic-amine-treated Chinese-hamster ovary (CHO) cells. By marker-enzyme analysis, the preparation from chloroquine or NH4Cl-treated cells was about 25-fold enriched for lysosomes compared with the postnuclear supernatant and contained little or no marker activities for the plasma membrane, rough endoplasmic reticulum, Golgi apparatus, mitochondria, cytosol and peroxisomes. The yield of amine-treated lysosomes was about 60% relative to the postnuclear supernatant. Electron microscopy and cytochemistry demonstrated that the amine-treated preparation was highly purified. Cytochemical analyses after a short-term pulse of horseradish peroxidase revealed that endosomal contamination of the lysosomal preparation was less than 1%. Lysosomal polypeptides were biosynthetically labelled with [35S]methionine and identified by SDS/polyacrylamide-gel electrophoresis. As expected, the bulk accumulation of luminal proteins into lysosomes was decreased. The bulk accumulation of membrane proteins was increased by acidotropic amine treatment. There were also several qualitative differences in each lysosomal compartment, with new species observed and other species absent. These data suggest that a low pH is not necessary for the normal accumulation of the bulk of membrane proteins in lysosomes and that membrane trafficking from Golgi apparatus to lysosomes occurs at a high rate in acidotropic-amine-treated CHO cells.     

Accumulation of membrane glycoproteins in lysosomes requires a tyrosine residue at a particular position in the cytoplasmic tail

Human lysosome membrane glycoprotein h-lamp-1 is a highly N-glycosylated protein found predominantly in lysosomes, with low levels present at the cell surface. The signal required for delivery of h-lamp-1 to lysosomes was investigated by analyzing the intracellular distribution of h-lamp-1 with altered amino acid sequences expressed from mutated cDNA clones. A cytoplasmic tail tyrosine residue found conserved in chicken, rodent, and human deduced amino acid sequences was discovered to be necessary for efficient lysosomal transport of h-lamp-1 in COS-1 cells. In addition, the position of the tyrosine residue relative to the membrane and carboxyl terminus also determined lysosomal expression. Supplanting the wild-type h-lamp-1 cytoplasmic tail onto a cell surface reporter glycoprotein was sufficient to cause redistribution of the chimera to lysosomes. A similar chimeric protein replacing the cytoplasmic tyrosine residue with an alanine was not expressed in lysosomes. Altered proteins that were not transported to lysosomes were found to accumulate at the cell surface, and unlike wild-type lysosomal membrane glycoproteins, were unable to undergo endocytosis. These data indicate that lysosomal membrane glycoproteins are sorted to lysosomes by a cytoplasmic signal containing tyrosine in a specific position, and the sorting signal may be recognized both in the trans-Golgi network and at the cell surface.     

Protein sorting upon exit from the endoplasmic reticulum

It is currently thought that all secretory proteins travel together to the Golgi apparatus where they are sorted to different destinations. However, the specific requirements for transport of GPI-anchored proteins from the endoplasmic reticulum to the Golgi apparatus in yeast could be explained if protein sorting occurs earlier in the pathway. Using an in vitro assay that reconstitutes a single round of budding from the endoplasmic reticulum, we found that GPI-anchored proteins and other secretory proteins exit the endoplasmic reticulum in distinct vesicles. Therefore, GPI-anchored proteins are sorted from other proteins, in particular other plasma membrane proteins, at an early stage of the secretory pathway. These results have wide implications for the mechanism of protein exit from the endoplasmic reticulum.     

Differential use of two AP-3-mediated pathways by lysosomal membrane proteins

The adaptor protein complex AP-3 is involved in the sorting of lysosomal membrane proteins to late endosomes/lysosomes. It is unclear whether AP-3-containing vesicles form at the trans-Golgi network (TGN) or early endosomes. We have compared the trafficking routes of endolyn/CD164 and 'typical' lysosomal membrane glycoproteins (lgp120/lamp-1 and CD63/lamp-3) containing cytosolic YXXPhi-targeting motifs preceded by asparagine and glycine, respectively. Endolyn, which has a NYHTL-motif, is concentrated in lysosomes, but also occurs in endosomes and at the cell surface. We observed predominant interaction of the NYHTL-motif with the mu-subunits of AP-3 in the yeast two-hybrid system. Endolyn was mislocalized to the cell surface in AP-3-deficient pearl cells, confirming a major role of AP-3 in endolyn traffic. However, lysosomal delivery of endolyn (or a NYHTL-reporter), but not GYXXPhi-containing proteins, was practically abolished when AP-2-mediated endocytosis or traffic from early to late endosomes was inhibited in NRK and 3T3 cells. This indicates that endolyn is mostly transported along the indirect lysosomal pathway (via the cell surface), rather than directly from the TGN to late endosomes/lysosomes. Our results suggest that AP-3 mediates lysosomal sorting of some membrane proteins in early endosomes in addition to sorting of proteins with intrinsically strong AP-3-interacting lysosomal targeting motifs at the TGN.