key: cord-104269-9r7rqqqk
authors: nan
title: Sorting signals in the MHC class II invariant chain cytoplasmic tail and transmembrane region determine trafficking to an endocytic processing compartment
date: 1994-07-02
journal: J Cell Biol
DOI: nan
sha: 
doc_id: 104269
cord_uid: 9r7rqqqk

Targeting of MHC class II molecules to the endocytic compartment where they encounter processed antigen is determined by the invariant chain (Ii). By analysis of Ii-transferrin receptor (TR) chimera trafficking, we have identified sorting signals in the Ii cytoplasmic tail and transmembrane region that mediate this process. Two non-tyrosine-based sorting signals in the Ii cytoplasmic tail were identified that mediate localization to plasma membrane clathrin-coated pits and promote rapid endocytosis. Leu7 and Ile8 were required for the activity of the signal most distal to the cell membrane whereas Pro15 Met16 Leu17 were important for the membrane-proximal signal. The same or overlapping non- tyrosine-based sorting signals are essential for delivery of Ii-TR chimeras, either by an intracellular route or via the plasma membrane, to an endocytic compartment where they are rapidly degraded. The Ii transmembrane region is also required for efficient delivery to this endocytic processing compartment and contains a signal distinct from the Ii cytoplasmic tail. More than 80% of the Ii-TR chimera containing the Ii cytoplasmic tail and transmembrane region is delivered directly to the endocytic pathway by an intracellular route, implying that the Ii sorting signals are efficiently recognized by sorting machinery located in the trans-Golgi.

Ii-TR chimera containing the Ii cytoplasmic tail and transmembrane region is delivered directly to the endocytic pathway by an intracellular route, implying that the Ii sorting signals are efficiently recognized by sorting machinery located in the trans-Golgi. M AJOR histocompatibility complex (MHC) 1 class 1I molecules are polymorphic cell surface glycoproteins expressed primarily on specialized antigenpresenting cells such as macrophages, dendritic cells, and B lymphocytes. They bind peptides derived from exogenous proteins and present them to CD4 + helper T cells as part of the mechanism for recognizing foreign antigens and stimulating an immune response. MHC class II molecules displayed on the cell surface are ct/3 heterodimers, but during their intracellular transport to the cell surface, they are transiently associated with a third nonpolymorphic polypeptide, the invariant chain (Ii) . Ii is a type II membrane protein with different isoforms that arise from alternative splicing of one exon and the use of two translation initiation sites (Strubin et al., 1986; O'Sullivan et al., 1987; Koch et al., 1987) .

Studies of mutant mice lacking Ii have established that Ii plays a critical role in the surface expression of class II molecules and in the ability of class II molecules to present native protein antigens (Viville et al., 1993; Bikoff et al., 1993) .. Ii determines the membrane trafficking of MHC class II a/3 dimers (Bakke and Dobberstein, 1990; Lotteau et al., 1990) and also blocks their peptide-binding site during transit through the biosynthetic pathway (Roche and Cresswell, 1990; Teyton et al., 1990; Roche et al., 1992) . Ij is assembled with class II a and 13 polypepfides in the endoplasmic reticulum, and, after its transport through the Golgi region, the a/3Ii complex is sorted out of the constitutive biosynthetic pathway to an acidic compartment within the endocytic system (Cresswell, 1985; Lamb et al., 1991; Pieters et al., 1991) . Within this prelysosomal (Peters et al., 1991) , early lysosomal (Harding and Geuze, 1993) , or endosomal com-pa~i,uent (Guagliardi et al., 1990) , Ii dissociates from class II a/3 dimers, permitting binding ofpeptides from exogenous foreign antigens generated in the endocytic pathway. Class II a/3 dimers loaded with peptide are then transported to the cell surface, while the free Ii is rapidly degraded (Blum and Cresswell, 1988; Pieters et al., 1991) . Although a/3Ii complexes are thought to be delivered to an endocytic compartment predominantly by an intracellular mute, small amounts of Ii and aflli complexes have been detected on the cell surface, and rapid internalization of a/3Ii complexes has been reported (Wraight et al., 1990; Lotteau et al., 1990; Roche et al., 1993) . These data suggest that at least some newly synthesized li molecules are transported to an cndocytic compartment via the plasma membrane, a route analogous to that taken by some integral lysosomal membrane proteins, e.g., lysosomal acid phosphatase (Braun et al., 1989; Peters et al., 1990) .

li contains one or more sorting signals within its aminoterminal cytoplasmic domain required for localization to the cndocytic system (Bakke and Dobberstcin, 1990; Lotteau et al., 1990; Simonsen ct al., 1993) . Ifli with a truncated cytoplasmic domain is coexpressed with class II o~/~ dimers, li fails to dissociate, and a~Ii complexes are expressed on the cell surface (Roche et al., 1992) . However, the sorting signals within the cytoplasmic domain of li have not been wellcharacterized, and neither the trafficking pathway taken by MHC class II cd3Ii complexes nor the cndocytic membrane compartment where o~/~ dimcrs encounter antigen have been clearly defined.

In this study, we have constructed Ii-human transferrin receptor (TR) chimeras with the aim of characterizing li sorting signals. We show that the cytoplasmic domain of li contains two independent non-tyrosine-based sorting signals, both of which mediate internalization via clathrincoated pits of the plasma membrane. Both signals contain two adjacent large hydrophobic nonaromatic residues suggesting that this may be a common feature of non-tyrosinebased sorting signals. Trafficking of Ii-TR chimeras to an cndocytic compartment where they are degraded is dependent upon these signals but occurs predominantly by a direct intracclhlar route, implying that the non-tyrosine-based sorting signals arc efficiently recognized by sorting machinery located in the trans-Golgi. This cndocytic compartment is distinct from the sorting and recycling endosomal compartments traversed by the wild-type TR, and, therefore, is located within the prelysosomal/lysosomal branch of the endocytic pathway. In addition, wc show that efficient delivery of Ii-TR chimeras to this endocytic compartment requires an independent signal in the Ii transmembrane region.

A Cla I fragment containing the entire coding region of the wild-type human TR was cloned into the phagemid pBluescript SK (Stratagene, La JoLla, CA) . Oligonucleotide site-directed mutaganesis was used to generate the cytoplasmic taft and transmembrane domain of Ii (Iic-r and IirM) chimeric constructs from a pBluescript SK phagemid template of the wild-type TR by the method of Kunkel (1985) using the Mnta-gene mntagenesis kit (Bio-Rad Labs., Richmond, CA). The same method was used to generate the Iic'r+zM chimeric construct and mutant IiCT chimeric constructs from pBluescript SK phagemid templates of the cytoplasmic tail of Ii (IicT) . Oligonucleotides were prepared on a model 391 PCR-I~TE DNA synthesizer using phosphoranddite chemistry and purified by polyacrylamide gel electrophoresis (Appl. Biosystems, Inc., Foster City, CA). Mutants were selected by restriction mapping or differential hybridization, and Cla I fragments encoding the mutant receptors were excised and cloned into the retroviral expression vector BH-RCAS (Hughes et al., 1990) . The mutations were verified by dideoxynucleotide sequencing (Sanger et al., 1977; Tabor and Richardson, 1987) of the BH-RCAS constructs using the Sequenase kit (US Biochemical Corp., Cleveland, OH) according to the manufacturer's directions.

Chicken embryo fibroblasts (CEF) were prepared from fertilized eggs (SPAFAS, Norwich, CT) and grown in DME supplemented with 1% (vol/vol) chicken serum, 1% (vol/vol) defined bovine calf serum (Hyclone, Logan, LIT), and 2 % (vol/vol) tryptose phosphate broth (Difeo, Detroit, MI) . CEF were transfected with 30/~g retroviral construct DNA per 10-cm tissue culture plate of 40% confluent cells using the polybrene-dimethyl sulfoxide method (Kawai and Nishizawa, 1984) . One to two weeks after transfection with the BH-RCAS constructs, the CEF stably expressed the wild-type TR and chimeric Ii-TR constructs as a result of infection by recombinant virus. Surface expression levels of the wild-type TR and chimeric Ii-TR constructs was determined by measuring the binding of 125Ilabeled human transferrin (Tf) at 4°C. Diferric human "If (Miles Scientific, Naperville, IL) was labeled with 125I to a specific activity of 2--4 #Ci/~tg using Enzymobeads (Bio-Rad Labs.) according to the manufacturer's directions. CEF were plated in triplicate at a density of 7.5 × 104 cells/cm 2 in 24-well tissue culture plates 24 h before the binding assay (Costar Corp., Cambridge, MA). Cells were incubated in serum-free DME for 1 h at 37°C, and then washed once with ice-cold 0.15 M NaC1, 0.01 M sodium phosphate buffer (pH 7.4) containing 0.1% bovine serum albumin (BSA-PBS). t25Ilabeled Tf (4 #g/rnl) in 0.15 rnl of BSA-PBS was added to triplicate wells and incubated at 4°C for 60 rain. Cells were then washed three times with 0.5 ml of ice-cold BSA-PBS, removed from the wells with 0.5 ml of 1 M NaOH, and the radioactivity was counted in a gamma counter.

The apparent internalization efficiencies of the wild-type TR and chimeric Ii-TR constructs were estimated from measurements of the steady-state distribution of receptors at 37°C (Tanner and Leinhard, 1987) . CEF were plated in triplicate wells as described for the binding studies. The cells were first incubated in serum-free DME, and then incubated with 4/tg/ml t2Sllabeled "If in BSA-PBS for I h at 37°C. The labeling medium was removed, and the cells were washed three times with 1 ml of ice-cold BSA-PBS, and then incubated twice for 3 min with 0.5 m] 0.2 M acetic acid-0.5 M NaCI (pH 2.4) to remove surface-bound 1251-1abeled Tf (Hopkins and Trowbridge, 1983) , and removed from the wells with I M NaOH. Radioactivity in the acid wash and in the cell lysete was determined. More prolonged incubation with the acid wash did not change the amount of t251 released. At steady state, the rate of internalization, kint, of cell surface Tf.TR complexes, [TR] .... equals the rate of externalization, kext, of the internal pool of apoTf.TR complexes, [TR]~; i.e., kint'[TR]sur = kext'~I'R]int, assuming an insignificant rate of degradation of internalized receptors during the time required to achieve steady-state. The values of ['I'R]~ and [TR]mt can be obtained from steady-state binding of TR under saturating conditions at 37°C. As ke~t of apoTf.TR complexes is independent of signals in the TR cytoplasmic domain , k~xt values of mutant and wild-type receptors are identical so that their k~ values are proportional to their steady state distribution, [TR] 

The internalization efficiencies of wild-type TR and Ii-TR chimeras were also determined by measuring their ability to mediate iron uptake. Human apo-Tf was labeled with 59Fe (FeCI3; Amersham Corp., Arlington Heights, IL) to a specific activity of 5-10 #Ci/ttg using ultrilotriacetate (Bates and Schlabach, 1973) . Cells were plated in triplicate wells as described for the binding studies. The following day, cells were washed twice in prewarmed (37°C) serum-free DME, and then incubated in DME containing 0.1% BSA and 4 t~g/m159Fe-Tf at 37°C for 0, 1, 2, 3, and 4 h. At the indicated times, the medium was removed, and cells were washed three times with ice-cold BSA-PBS. Cells from triplicate wells for each time point were removed in 0.5 ml 1 M NaOH, and the radioactivity was counted in a gamma counter. The relative levels of wild-type TR and chimeric Ii-TR constructs expressed on the various CEF populations were determined in each experiment by measuring 125I-labeled Tf binding at 4°C for each weft. After preincubation for 1 h at 37°C in serum-free DME, triplicate wells of cells were incubated with 4/~g/rul t25I-labeled Tf on ice for 1 h, and then washed three times with 1 mi of ice-cold BSA-PBS, and the radioactivity bound to the cells was counted in a gamma counter.

CEF were plated in triplicate wells as described for the binding studies. Cells were preincobated in serum-free DME for 30 vain at 37°C, and then incubated with 125I-labeled "IY (4 #g/m1) in BSA-PBS for 1 h at 37°C. The medium was removed and the cells were washed three times with ice-cold BSA-PBS and incubated at 37°C with prewarmed (37°C) DME containing 0.1% BSA and 50 #g/ml unlabeled "If for 0, 5, 10, 15, 20, 40, or 60 rain. After incubation, the medium was collected, protein was precipitated in 10% TCA and removed by centrifugation, and then the acid-soluble and acid-insoluble radioactivity was counted in a gamma counter. The surfacebound and internalized "If in CEF was determined by the acid wash procedure described for the steady-state distribution assay.

CEF expressing chimeric Ii-TR constructs were 1~I surface-labeled using lactoperoxidase for 10 min essentially as described , except that iodination was performed in 6-cm tissue culture dishes in a total volume of 0.5 rnl. CEF were then washed four times with ice-culd PBS and incubated at 37°C in complete medium for 0, 2, 3, 4, or 6 h. At each time point, cells were solub'dized on ice with 1% NP-40-PBS, and the chimeric Ii-TR constructs were immunoprecipitated using B3/25 monoclonal antibody specific for the extracellular domain of the human TR . Immunoprecipitates were analyzed on 7.5% SDS-polyacrylamide gels, which were then dried and exposed to preflashed XAR film (Eastman Kodak, Rochester, NY).

CEF were washed twice with methionine-free DME, and then pulse-labeled for 30 rain in 1 ml of DME containing 0.12 mCi/mi trans-asS-label (ICN Biomedicals, Irvine, CA) and 2% defined calf serum. Pulse-labeled cells were chased for 0, 2, 4, 8, or 24 h in complete medium. In one experiment, cells were preincubated for 4 h in medium containing 100 #g/ml leupeptin, and then pulse-labeled and chased for 0, 1, 2, 3, or 4 h in leupeptin-containing medium. At each time point, cells were soluffdized on ice with 1% NP-40-PBS. The wild-type TR and chimeric fi-TR constructs were immunoprecipitated from postnuclear supernatants using B3/25 monoclonal antibody and analyzed on 7.5% SDS-polyacrylamide gels . Dried gels were exposed to prettashed XAR film (Eastman Kodak, Rochester, NY). Quantitation of radioactivity was performed on a model 425 PhosphorImager (Molecular Dynamics, Sunnyvale, CA).

The surface-accessible pools of wild-type TR and Ii-TR chimeras were calculated from the nsI-labeled "If bound to intracellular and surface receptors at steady-state. The surface-accessible pool of chimera relative to wild4ype TR, [Ii-TR] s^, equals the cell surface chimem-Tf complexes, FIi-TR]s~, plus internalized chimera-Tf complexes [Ii-TR] int, multiplied by 100, divided by cell surface TR-Tf complexes [TR]sur, plus internalized TR-Tf complexes, [TR]int; i.e., [Ii-TR] 

receptors. This condition is met as steady-state conditions are achieved in ,~20 rain at 37°C (fling et al., 1990; see also above). In 20 rain, <10% of the total 125Llabeled "If bound to the most rapidly degraded chimera, Iicr+xu, is released into the medium as ~ soluble material (see Fig. 7 ). The fraction of Ii-TR chimeras trafficking via the cell surface is the surface accessible pool, [Ii-TR] sA, divided by the rate of biosynthesis of the Ii-TR chimera multiplied by the rate of degradation of [Ii-TR] sA; i.e., ([Ii-TRsA X degradation rate of synthesis.

As an example of the calculation, the surface expression of Iic'r+TM is 12% relative to wild-type TR, and at stendy-state 76% of the Iicr+~ chimera and 62% of the wild-type TR are internalized (see Results and Table  I ). Thus, [Iic'r+T~.dSA ffi (12 + [12 X 76/24]) X 100/(100 + [100 X 62/38]) = 19%. Quantitation of the data presented in Fig. 2 showed that eightfold more Iic~+~ chimera was synthesized than wild-type TR, whereas quantitation of the data presented in Fig. 8 showed that the [Iic'r+TM]SA is degraded eightfold faster than wild-type TR. Therefore, the fraction of Iic'r+TM chimeras trafficking via the cell surface relative to wildtype TR is 19 x 8/8% or 19%.

CEF and CEF expressing wild-type or Ii-TR chimeras were plated onto glass coverslips and cultured overnight. The cells were fixed with 2 % formaldehyde treatment for 15 rain at room temperature (RT), quenched in 0.27% NI-hC1, 0.37% glycine, pH 7.2, and incubated in P/H/S (PBS/10% horse serum/0.1% saponin) (for permeabilized samples) or P/H (for nonpermeab'diznd samples) for 30 min at RT. The samples were next incubated with monoclonal antibody B3/25 (undiluted hybridoma supernatants) followed by a second incubation with rat anti-mouse IgGi-FrIC (1:20 dilution; Zymed Labs.). Each incubation step was followed by three washes in P/H/S or P/H for 10 rain each. In all cases, antibodies were diluted in P/I-US or P/H, and primary and secondary antibodies were incubated for 1 h and 30 rain, respectively. Covershps were mounted in 1 mg/ml p-phenylenediamine in a 1:10 mixture of PBS:glycerol and sealed with nail polish. Slides were analyzed on a Nikon Optiphot microscope equipped with epifluorescence and a DAGE MTI series SIT video camera (Dage-MTI, Michigan City, IN) coupled to an Image 1-AT (Universal Ima~ng~ Media, PA) image analysis system. Micrographs were prepared with a Sony Color Video Printer (UP-5000).

For double-label indirect immunofluorescence, ceils were fixed, permeabilized, and incubated with the following combinations of antibodies: rabbit anti-human TR antiserum (1:500 dilution), goat anti-rabbit Ig-FITC (1:20 dilution; American Qualex), mouse monoclonal antibody anti-LEP100 (undiluted tissue culture supernatant, Developmental Studies Hybridoma Bank, University of Iowa), goat anti-mouse IgG-Texas red (1:100: Zymed Laboratories). Slides were analyzed on a Leitz fluorescence microscope equipped with a Vario Orthomat II camera system and a dual wavelength (FiTC/Texas red) epifluorescence filter module. Micrographs were prepared with Ektachrome Pl600 Professional film processed at ASA 800.

Colloidal gold (6 and 10 tun diameter) sols were made as described by Slot and ~ (1985) . Protein A gold and rabbit anti-mouse Ig gold antibody conjugates (1978) were made by standard methods (de Mey, 1986 ) and cryosections made and immunolabeled essentially as described by Tokuyasu (1978) . The sections were incubated 30 rain with a rabbit antiserum a vaingt human TR (Trowbridge, I.S., unpublished results) followed by 6-am gold protein A for 20 rain, and then rinsed for 20 rain with five changes of PBS and incubated with anti-LEP 100 mouse monoclooal antibody for 30 rain followed by 10-nm gold rabbit anti-mouse Ig antibody for 20 rain before an additional five rinses in PBS. Omitting either first or second specific antibodies indicated nonspecific labeling was less than 5%. Staining and embedding were carried out as described previously (Hopkim, 1983 ) and the sections were examined at 80 keV in a Philips 301 electron microscope. Quantitation of the Iicr+m chimera in clathrin-coated pits was determined on conventional plastic sections from cells incubated with B3/25-gold complexes at 4°C as described earlier (Miller et al., 1991) .

To investigate the signal-dependent trafficking of fi, we constructed Ii-TR chimeras consisting of the fi cytoplasmic tail, the TR extracellular domain, and the transmembrane region from either TR or from Ii (Fig. 1) . We reasoned that fi-TR chimeras would be expressed in sufficient amounts on the cell surface to identify Ii sorting signals using quantitative assays which measure Tf internalization and iron uptake . As there is increasing evidence that the same or closely related signals are recognized by the clathrin-based sorting machinery at the plasma membrane and trans-Golgi (Letourneur and Klausner, 1992 would also allow identification of Ii sorting signals operative along the intracelhlar trafficking pathway. Wild-type human TR and Ii-TR chimeras containing either the Iicr or Iicr÷ru were stably expressed in CEF using BH-RCAS, a replication-competent retroviral vector derived from Rous sarcoma virus Hughes et al., 1990) . Binding studies at 4°C using ~2SI-labeled human Tf indicated that both Ii-TR chimeras were expressed on the cell surface of infected CEF, but at lower levels than wild-type TR (51 + 13% Iicr and 12 ± 1% Iicr÷rM relative to wildtype TR).

Internalization of the Ii-TR chimeras was assayed by measuring the steady-state distribution of internalized receptors and their ability to mediate iron uptake Collawn et al., 1990) . The Iicr chimera was efficiently internalized, as judged by either iron uptake or steady-state assays (Table I) , indicating that the Ii cytoplasmic domain contains internalization signal(s) that can promote rapid endocytosis of the human TR. Electron microscopy (see Fig.  6 , inset) clearly demonstrates that Ii-TR chimeras internalize via clathrin-coated pits, 15% of the label on the plasma membrane being located within these domains. The apparent internalization efficiency of the Iicr+ru chimera was significantly higher than either wild-type TR or the Iicr chimera, implying that the transmembrane region of Ii also influences trafficking.

To locate the Ii cytoplasmic tail internalization signal(s), deletion mutants of the Iicr chimera were constructed (Fig.  1) . Deletion of residues 2-5 or residues 2-11 reduced the apparent internalization efficiency of the Iicr chimera to ~50%, whereas deletion of residues 2-17 completely abolished rapid endocytosis (Table II) . These results suggest that the Ii cytoplasmic tail contains two internalization signals, one located between residues 2-11, the other between residues 12-17.

Since the Ii cytoplasmic tail does not contain aromatic residues, the internalization signals identified by analysis of the Iicr deletion mutants cannot be tyrosine-based. However, the Ii cytoplasmic tail contains three leucine residues (residues 7, 14, and 17) within the first 17 amino-terminal residues (Fig. 1) , suggesting that internalization of the Ii-TR chimeras may be mediated by signals related to the dileucine-based sorting signals previously identified in the cytoplasmic tails of the T cell receptor CD3"y chain and the cation-independent and cation-dependent mannose-6-pbosphate receptors (Letourneur and Klausner, 1992; Johnson and Kornfeld, 1992a,b) . Although first identified as a lysosomal targeting signal, the CD37 chain di-leucine signal has been shown to function as an internalization signal (Letourneur and Klausner, 1992; White, S., J. F. Collawn, and I. S. Trowbridge, unpublished results). Di-leucine-based sorting signals have not been extensively characterized; however, isoleucine can substitute for the second leucine, but not the first, in the CD3~/di-leucine signal without loss of lysosomal targeting activity, whereas alteration of either leucine to alanine substantially reduces activity (Letourneur and Klausnet, 1992) . Because of the stricter requirement for leucine in the first position of the CD3~, chain signal, we initially thought that the two internalization signals identified by deletional analysis may involve I.,¢uTIle s and Leu~4pro is, respectively. To investigate this possibility, the two residues of either one or both of these di-peptides were altered to alanine. Strikingly, independently altering each di-peptide to two alanine residues reduced the apparent internalization efficiency of the Iicr chimera by ~,50%, whereas alteration of both dipeptides completely abrogated high efficiency endocytosis (Table ID . These results supported the idea that the Ii cytoplasmic tail contains two non-tyrosine-based signals with similar internalization activity that are related to the CD3"y chain di-leucine signal.

To further characterize the two sorting signals, each of the two residues that might be important for their activity were independently changed to alanine in mutant receptors in which the second signal had been inactivated by alanine substitutions. The results of this analysis clearly showed that both Leu 7 and He S were required for the signal most distal from the transmembrane region (Table HI) . However, whereas alteration of Pro ~s to alanine abolished the internalization activity of the membrane-proximal signal, alteration of Leu 14 to alanine increased its efficiency. This result implied that the membrane-proximal signal was novel and led us to examine whether the two hydrophobic residues, Met ~6 and Leu ~7, on the carboxy-terminal side of Pro ~5 were required. As shown in Table HI , alteration of either Met ~6 or Leu 17 to alanine completely abrogated the activity of the membrane-proximal sorting signal in the Ii cytoplasmic tall. We conclude, therefore, that one internalization signal in the Ii cytoplasmic taft involves Leu 7 and Ile s, whereas Pro 15, Met 16, or Leu ~7 are all important for the second signal.

Consistent with the distal signal being closely related to the CD33, chain di-leucine signal, Ile s could be modified to leucine without loss of activity. Interestingly, the nonconservative mutation of Pro ~5 to leucine did not reduce the internaliTation activity of the membrane-proximal signal (Table HI) .

Ii is thought to be transported as an ot/~Ii complex via the trans-Golgi directly to an acidic endocytic compathaent where it dissociates and is rapidly degraded (Cresswell, 1985; Lamb et al., 1991; Pieters et al., 1991) . To determine whether the cytoplasmic tail of Ii was sufficient to target TR to this endocytic compartment and whether the non-tyrosine-based sorting signals were involved, metabolic pulse- * The LT--*'A, P"*A, and P~L mutations were introduced into the mutant Iicr LP-*AA construct (internalization efficiency of 57% relative to the Iicr chimera; see Table II) , and the L~'-~A, PlS~A, P~s-"L, M16~A, and LIT-*A mutations were introduced into the mutant Iicr LI~AA construct (internalization efficiency of 50% relative to the Iicr chimera; see Table II ). Internalization efficiencies of mutant Iicr chimeras are expressed relative to the Iic-r chimeras as described in Table II Figure 2 . Rapid degradation of Ii-TR chimeras in a post-Golgi endocytic compartment. Equivalent cell numbers of CEF expressing either wild-type TR or the Ii-TR chimeras, Iicr, Iier+xM, or Iicr LI,LP--*AA,AA were pulselabeled for 30 rain with trans-35S-label and chased for various periods of time. TRs were then immunoprecipitated from post-nuclear supernatants and analyzed on SDS-polyacrylamide gels as described in Materials and Methods. Dried gels were exposed to pralashed XAR film for two days (Eastman Kodak, Rochester, NY).

Immunoprecipitates were quantitated on a model 425 Phos-phorImager (Molecular Dynamics). The data shown are from one of two similar experiments.

chase experiments were performed. CEF expressing either TR, Iicr, Iicr+r~ or the Iicr LI,LP---AA,AA mutant were pulse-labeled with trans-3sS-label for 30 min and chased for various lengths of time; TR and Ii-TR chimeras were then isolated by immunoprecipitation and analyzed by SDS-PAGE (Fig. 2) . Notably, both the Iicr÷rM and Iier chimeras were rapidly degraded with half-lives of < 2 h and '~,4 h, respectively, similar to the rate of turnover of native Ii in antigen-presenting cells. In contrast, the wild-type TR had a half-life of '~24 h, consistent with previous estimates of wild-type human TR turnover rates in CEF . The Iicr÷a~ chimera was consistently degraded >2-fold more rapidly than the Iier chimera, implying that the Ii transmembrane region influences intracellular trafficking. After 2 h (Fig. 2) , the Mr of TR and Ii-TR chimeras increased to that of the mature glycoprotein and the molecules became partially resistant to endoglycosidase H digestion (data not shown). This indicates that the Ii-TR chimeras traverse the Golgi where glycosylation is completed and are degraded in a post-Golgi compartment. The Ii-TR chimera lacking the non-tyrosine-based signals was degraded more slowly, indicating that these signals are required for delivery to this compartment. Further, the rate of synthesis of the Iicx and Iicr+m chimeras was 4-and 8-fold higher, respectively, than that of wild-type TR even though the chimeric receptors were expressed in lower amounts on the cell surface suggesting most of the chimeric molecules were trafficking by a direct intracellular route to the endocytic compartment where they were degraded. Degradation of Ii is partially inhibited by lysosomotropic agents such as chloroquine, monensin, and leupeptin (Ilium and Cresswell, 1988; Nguyen and Humphreys 1989; Pieters et al., 1991; Zachgo et al., 1992) . Similarly, leupeptin significantly inhibited the degradation of the Iicr+rM chimera, as indicated by the slower rate of disappearance of the intact molecule and the appearance of partial proteolytic products derived from the TR external domain (Fig. 3) , which is consistent with degradation occurring in an endocytic compartment. Comparable results were obtained with chloroquine treatment (data not shown).

Ii-TR chimeras trafficking through the Golgi complex were identified using electron microscopy immunocytochemistry. Gold complexes bearing antibody specific for the external domain of TR were applied to cryosections and showed Ii-TR distributed throughout the flattened cisternae of the Golgi stack and within coated vesicles in the trans-Golgi area (see Fig. 6 ).

The intracellular and cell surface distribution of the Ii-TR chimeras was determined by immunofluorescence microscopy. The Iicr+rM chimera was found predominantly in intracellular vesicles, as indicated by the virtual lack of staining of non-permeabilized cells (Fig. 4 A) , compared to the strong punctate staining of permeabilized cells (Fig. 4 B) . CEF expressing the Iicr chimera exhibited more surface staining (Fig. 4 E) , but this chimera was also found predominantly in intracellular vesicles (Fig. 4 F) . The Iicr LI,LP ~ AA,AA chimera was expressed mainly on the cell surface ( Fig. 4 G) , but some staining of intracellular vesicles was still observed (Fig. 4 H) . Two-color immunofluorescence studies comparing the intracellular distribution of the Iicr+rM chimera with LEP100, an endogenous chicken lysosomal integral membrane protein Fambrough, 1986, 1987; Mathews et al., 1992) , showed significant colocalization, indicating that the hybrid molecules traffic along the prelysosomal segment of the endocytic pathway (Fig. 5 C) . Conversely, wild-type TR showed virtually no colocalization with LEP100 (Fig. 5 A) . The Iicv chimera also partially colocalized with LEP100 (Fig. 5 B) , whereas the Iicr LI,LP-*AA,AA (Fig. 5 D) gave a staining pattern similar to that of wild-type TR.

To quantitatively determine the altered trafficking of Iicr÷m chimera, the steady-state distribution of TR and Ii-TR chimeras was determined by gold label immunocytochemistry using antibody specific for the external domain of human TR. The extent to which the TR and Iicr+m chimera were colocalized with LEP100, an antigen located predominantly in the prelysosomal branch of the endocytic pathway Fambrough, 1986, 1987) , was quantified by double labeling. As shown in Fig. 6 , the distribution of TR, Iicr÷m, and LEP100 is clearly restricted to the perimeter membranes of endocytic elements. Quantitative analysis shows that while relatively low amounts (11.2 %) of TR can be found in LEP100-containing compartments, 64 % of the labeled compartments in cells transfected with Iicr+~M contain both Iic~÷~ and LEP100.

To determine whether fi-TR chimeras expressed on the cell surface traffic to the endocytic compartment where degradation occurs, cells were incubated with 1=I-labeled Tf at 37°C for 1 h to load the endocytic pathway with receptorligand complexes. The cells were then rapidly washed, and the reappearance of intact and degraded Tf in the medium Time (min) Figure 7 . Degradation of 1T bound to Ii-TR chimeras. Equivalent cell numbers of CEF expressing either wild-type TR, Iicr, IicT+m, or Iicr LI,LP-*AA,AA were preincubated in serum-free DME for 30 rain at 37°C, and then incubated with 125I-labeled Tf for 1 h at 37°C. The cells were then washed and reincubated at 37°C in DME containing 50 #g/ml unlabeled Tf for various times. Acid-soluble radioactivity (B) or acid-insoluble 125I-labeled Tf (A) released into the medium, as well as surface-bound 125I-labeled "IT (n) and internalized 12SI-labeled Tf (0) were determined as described in Materials and Methods and are expressed as a percent of total radioactivity recovered. Each point is the average of triplicate determinations from a representative experiment.

was monitored by measuring TCA insoluble and soluble radioactivity. As expected, the apo-Tf released into the medium from cells expressing the wild-type TR was undergraded ( Fig. 7) , as TR-apo-Tf complexes are efficiently recycled back to the cell surface through the sorting and recycling endosomal compartments Hopkins and Trowbridge, 1983; Dunn et al., 1989; Weissman et al., 1986; Trowbridge et al., 1993) . In stalking contrast, ,x,35% of the l~I-labeled Tf released from cells expressing the Iicr÷m chimera was degraded, implying that this fraction of the chimeric receptors traffic directly from the cell surface to an endocytic compartment where they are degraded. A significantly smaller fraction (~,20%) of Tf bound to the Iicr chimera was degraded (Fig. 7) , indicating that the Ii transmembrane region influences trafficking from the cell surface to this compartment. The Ii-TR chimera lacking nontyrosine-based sorting signals was degraded only to a slightly higher extent than wild-type TR.

To confirm the inference from these studies that the Ii-TR chimeras traffic from the plasma membrane to an endocytic compartment where they are degraded, CEF expressing IicT, Iicr+m, or Iicr LI,LP-*AA,AA were surface-iodinated and the rate of degradation of receptors transiently located at the plasma membrane directly determined. As shown in Fig. 8 , surface-iodinated Iicr chimera (half-life •6 h) was degraded more rapidly than the chimera lacking non-tyrosine-based signals. Surface-iodlnated Iicr+~ was degraded at an even faster rate (half-life ~2-3 h).

The Iicr+~ chimera was degraded more rapidly than the Iicv chimera, whether the chimeras were delivered directly from the trans-Golgi to the endocytic pathway or via the cell surface, implying that the Ii transmembrane region was influencing trafficking along both pathways. To determine whether the Ii transmembrane region was sufficient to promote delivery of TR to the degradative endocytic compartment, a chimera consisting of the TR external domain and cytoplasmic tail and the Ii transmembrane region (Iim) was constructed. The Ii~ chimera was expressed on the cell surface at a level of 23 + 6 % relative to the wild-type TR and was internalized with a relative efficiency of 76.0 + Figure 8 . Degradation of 125I surface-labeled Ii-TR chimeras. Equivalent cell numbers of CEF expressing either Iicr, IicT+I"M or Iic~ LI,LP-*AA,AA chimeras were surface-iodinated, and then incubated for the times indicated. The Ii-TR chimeras were immunoprecipitated from post-nuclear snpernatants and analyzed on SDS-polyacrylamide gels as described in Materials and Methods. Dried gels were exposed to preflashed XAR film (Eastman Kodak, Rochester, NY) for one day (licr L1,LP-~A4,A4 and licr) or 2 d (licr+rM). Figure 9 . li transmembrane region targets the TR to a post-Golgi degradative compartment. Equivalent cell numbers of CEF expressing either wild-type TR or the IiTM chimeric receptor were pulse-labeled with trans-aSS-label and chased as described in the legend for Fig. 2 . TRs were then immunoprecipitated from postnuclear supernatants and analyzed on SDS-polyacrylamide gels as described in Materials and Methods. Dried gels were exposed to preflashed XAR film for two days (Eastman Kodak, Rochester, NY), and immtmoprecipitates were quantitated on a model 425 Phosphorlmager (Molecular Dynamics). 0.5%. Importantly, pulse-chase analysis showed that the fully glycosylated IiTM chimera was rapidly degraded (Fig.  9 ) with a half-life of ,,06 h, similar to that of the Iicr chimera.

By constructing 11-TR chimeras comprised of the cytoplasmic domain of Ii, the external domain of the human TR, and the transmembrane region of either Ii or TR, we have been able to take advantage of the quantitative assays available for measuring the rate of internalization of TRs from the cell surface. This experimental strategy has allowed us to demonstrate that the Ii cytoplasmic tall can promote rapid internalization of TR and to identify residues within the cytoplasmic tall of 1i that are required for this activity.

Our data indicate that the 11 cytoplasmic tail contains two independent non-tyrosine-based sorting signals that promote rapid internalization. The same or overlapping signals are recognized in the trans-Golgi and mediate direct sorting along an intracellular route to the endocytic pathway. One signal involves Leu 7 and Ile s and appears to be analogous to the di-leucine lysosomal targeting signal identified in the cytoplasmic domain of the CD3"y chain (Letourneur and Klausner, 1992) . Deletion of residues 2-5 from the Icr chimera cytoplasmic tall reduces the internalization activity to ,~50%, implying that, as for the di-leucine signal of the CD33, chain (Letourneur and Klausner, 1992) , residues to the amino-terminal side of Leuqle s are required for activity. The second sorting signal requires prols, Met16, and Leu ~7, and, therefore, differs significantly from the dileucine and leucine-isoleucine motifs identified previously (Letourner and Klausner, 1992; Johnson and Kornfeld, 1992a,b; Ogata and Fukuda, 1994) . Nevertheless, all these signals, including both 11 signals, contain two adjacent large nonaromatic hydrophobic residues and may, therefore, represent a family of related structural motifs specified by short linear arrays of amino acids that differ in specific sequence, analogous to tyrosine-based signals (Trowbridge et al., 1993) . Additional mutagenesis is required, however, to define these sorting signals more precisely, to identify any common features they may share, and to delineate any differences between signals recognized in the trans-Golgi or at the plasma membrane. At present, we believe it most appropriate to refer to them as non-tyrosine-based signals, a neutral term that serves to distinguish them from tyrosine-based signals.

Our results are consistent with those of Bakke and Dobberstein (1990), who identified amino acids 12-15 in the Ii cytoplasmic tail as important for targeting Ii to the endocytic pathway. Lotteau et al. (1990) also concluded that a sorting signal was located within residues 10-16 of the Ii cytoplasmic tail. The distal signal involving Leuqle s was not identified in either of these studies, however (see also Romagnoli et al., 1993) , illustrating the difficulty of detecting multiple signals without quantitative assays. Recently, in agreement with our data, Pieters et al. (1993) have independently concluded that the Ii cytoplasmic tail contains two sorting signals, one involving LeuTIle 8, the other localized to residues 12-29.

The extent to which newly synthesized 11 molecules, alone or complexed to MHC class II c~/~ dimers, traffic to the endocytic pathway by an indirect route via the cell surface has been a matter of debate (Peters et al., 1991; Pieters et al., 1991; Roche et al., 1993) . Ii-TR chimeras trafficking to the plasma membrane can be quantitated because receptors appearing transiently at the cell surface can be labeled with exogenous Tf. Thus, the surface-accessible pools of the Ii-TR chimeras can be determined by measuring ~25I-labeled Tf bound to surface and internalized chimeric receptors under steady-state conditions. The surface-accessible pool of the Iicr+rM chimera was only 19% of the wild-type TR pool despite the fact that the ,,o 8-fold increase in the degradation rate of the Iicr+rM surface-accessible pool is offset by a similar increased rate of biosynthesis relative to wild-type TR (see Materials and Methods for details of the calculation). Assuming that essentially all wild-type TRs bind Tf under steady-state labeling conditions, it can be calculated that only 19 % of the Iicr+rM chimera traffics via the cell surface, and, therefore, that >80% of the chimera must be delivered by a direct intraceilular route to the endocytic compartment where it is degraded. Similarly, •50% of the Iicr chimera is degraded without ever being displayed on the cell surface. The lysosomal membrane glycoprotein, Lgp-A (LEP100, LAMP-1), also traffics along the endocytic pathway to the lysosome by a direct intracellular route or via the cell surface, although which route is the major pathway is less clear and may depend upon expression level (Williarns and Fukuda, 1990; Mathews et al., 1992; Halter and Mellman, 1992) .

The fate of Ii-TR chimeras which traffic via the plasma membrane can also be determined by loading them with exogenous Tf and monitoring the reappearance of intact and degraded Tf in the medium. The results of such studies indicated that ,~35% of 11CT+TM chimeras transiently displayed on the plasma membrane were degraded after internalization, whereas the remaining ,,065 % were recycled back to the cell surface. An even higher fraction (,,080%) of the 11cr chimera recycled back to the cell surface. These results indicate that the Ii-TR chimeras displayed on the plasma membrane recycle several times before being sorted to the endocytic compartment where they are degraded. In this respect, 11-TR chimeras expressed on the cell surface behave similarly to lysosomal acid phosphatase which is transported to the cell surface and undergoes multiple rounds of internalization and recycling before being transferred to the lysosome (Braun et al., 1989) .

Studies on the trafficking of the cation-independent mannose-6-phosphate receptor (M6PR) have clearly established that a direct route from the trans-Golgi to the endocytic pathway exists (Kornfield and Mellman 1989) . It has been shown that M6PR are concentrated within clathrin-coated buds in the trans-Golgi (Rijnboutt et al., 1992; Klumperman et al., 1993) and there is evidence suggesting that they are then delivered to an early part of the endocytic pathway, most probably the TR-containing endosome (Ludwig et al., 1991; M6resse and Hoflack, 1993) . Previous work on Ii (Guagliardi et al., 1990; Pieters et al., 1991; Zachgo et al., 1992; Romagnoli et al., 1993) also suggest that entry into the endocytic pathway is via the TR-containing endosome.

Our data show that the cytoplasmic domain of Ii contains two signals recognized in the trans-Golgi that are the same or overlap with the internalization signals recognized by clathrin-coated domains at the plasma membrane. This relationship implies that the signals active in the trans-Golgi are recognized by clathrin-based sorting machinery. It is also apparent, however, that for direct transfer of Ii from the trans-Golgi to the endocytic pathway to operate with full efficiency, additional signal information within the transmembrane region is required. The Ii transmembrane region also increases the efficiency of trafficking from the plasma membrane to the degradative endocytic compartment. Golgi retention signals have previously been localized to the transmembrane region (Machamer and Rose, 1987; Swift and Machamer, 1991; Munro, 1991; Nilsson et al., 1991; Wong et al., 1992) . However, there is no precedent for the transmembrane region of an integral membrane protein influencing sorting in the trans-Golgi or trafficking along the endocytic pathway.

Our evidence shows that Ii, like the M6PR (Kornfield and Mellman, 1989), enters the TR-containing endosome from both the trans-Golgi and the plasma membrane. However, the data also indicate that the Ii-TR chimeras which enter directly from the trans-Golgi are delivered to a degradafive compartment more efficiently than those which enter from the plasma membrane. Our recent observations on the TR-contalning endosome in CEF (Hopkins et al., 1994) show that it is a morphologically complex compatmient and suggest that it consists of at least two interconnected subcompartments in which internalized TRs are processed at different rates. Receptors entering the TR-containing endosome can either be recycled rapidly (tee in 10 rain) or remain for more prolonged periods (t~r2 in 30 min) before returning to the surface. The data on Ii-TR chimeras can thus be explained by proposing that most of the chimeras which enter from the cell surface travel via the rapidly recycling pathway and that on each cycle only •30% of them are sorted towards the lysosome. Direct transfer from the trans-Golgi, though more efficient, probably occurs via the subcompartment of the endosome which lies within its immediate vicinity in the pericentriolar area and through which TR are processed more slowly.

Identification of a sorting signal in the Ii transmembrane region that complements the non-tyrosine-based sorting signals in the Ii cytoplasmic tail provides a plausible explanation for two earlier observations. Pieters et al. (1993) showed that the Ii cytoplasmic tall was sufficient to target a heterologous protein, influenza neuraminidase, to the endocytic pathway. However, they noted significant differences between the localization of the Ii-neuraminidase chimera and Ii within the endocytic pathway and concluded that other sorting information may be contained within the Ii transmembrane region and/or extracellular domain. In a second study, Nilsson et al. (1991) reported that an Ii chimera containing the cytoplasmic tail of 8-1,4 galactosyltransferase, an enzyme normally localized to the trans-Golgi by a retention signal within its transmembrane region, paradoxically had a subcellular distribution similar to wild-type Ii (p31). The galactosyltransferase cytoplasmic tall contains a dileucine sequence which may function as a sorting signal and when complemented by the Ii transmembrane region signal promote similar trafficking to wild-type Ii.

Finally, Dintzis and Pfeffer (1990) observed that the M6PR cytoplasmic tall was not sufficient to target the epidermal growth factor receptor to prelysosomes and suggested that additional sorting information in the M6PR transmembrane region and/or extracellular domain may be required.

The identification of the Ii sorting signals required for targeting MHC class H molecules to the compartment where they encounter antigen may have practical implications since, in principle, it should be possible to target endogenously synthesized recombinant proteins containing Ii sorting signals to this compartment for degradation and selective presentation by MHC class II to T helper cells.

Abbreviations used in this paper: CEF, chicken embryo fibroblasts; CT, cytoplasmic tail

Iicr, the cytoplasmic tail of Ii

Ii transmembrane region

M6PR, mannose-6-phosphate receptor

MCH class H-associated invariant chain contains a sorting signal for endosomal compartments

The reaction of ferric salts with transferrin

Defective major histocompatibility complex class II assembly, transport, peptide acquisition, and CD4 + T cell selection in mice lacking invariant chain expression

Role for intracellular proteases in the processing and transport of class II HLA antigens

Lysosomal acid phosphatase is transported to lysosomes via the cell surface

Transferrin receptor internalization sequence YXRF implicates a type I turn as the structural recognition motif for endocytosis

Intracellular class II HLA antigens are accessible to transferrin-neuraminidase conjugates internalized by receptor-mediated endocytosis

The preparation and use of gold probes

The mannose 6-phosphate receptor cytoplasmic domain is not sufficient to alter the cellular distribution of a chimeric EGF receptor

Iterative fractionation of recycling receptors from lysosomally destined ligands in an early sorting endosome

Co-localization of molecules involved in antigen processing and presentation in an early endocytic compartment

Immunogenic peptides bind to class II MHC molecules in an early lysosomal compartment

Transport of the lysosomal membrane glycoprotein lgpl20 (lgp-A) to lysosomes does not require appearance on the plasma membrane

Intracellular routing of transferrin and transferrin receptors in epidermoid carcinoma A431 cells

Internalization and processing of transferrin and the transferrin receptor in human carcinoma A431 cells

Trowbridge. 1994. In migrating fibroblasts recycling receptors are concentrated in narrow tubules in the pericentriolar area and then routed to the plasma membrane of the leading lameUa

Vectors and genes for improvement of animal strains

Nonacylated human transferrin receptors are rapidly internalized and mediate iron uptake

Role of the human transferrin receptor cytoplasmic domain in endocytosis: localization of a specific signal sequence for internalization

A His-Leu-Leu sequence near the carboxyl terminus of the cytoplasmic domain of the cation-dependent Marmose 6-pbosphate receptor is necessary for the lysosomal enzyme sorting function

The cytoplasmic taft of the Marmose 6-phosphate/insulin-like growth factor-H receptor has two signals for lysosomal enzyme sorting in the Golgi

New procedure for DNA transfection with polycation and dimethyl sulfoxide

Differences in the endosomal distributions of the two Mannose 6-phosphate receptors

Primary structure of the gene for the murine Ia antigen-associated in variant chains (Ii). An alternatively spliced exon encodes a cysteine rich domain highly homologous to a repetitive sequence in thyroglobulin

The biogenesis of lysosomes

Rapid and efficient site-specific mutagenesis without phenotypic selection

Invariant chain targets HLA class II molecules to acidic endosomes containing internaiized influenza virus

A novel di-leucine motif and a tyrosine-based motif independently mediate lysosomal targeting and endocytosis of CD3 chains

Lysosomal membrane dynamics: structure and interorganellar movement of a major lysosornal membrane glycoprotein

Cycling of the integral membrane glycoprotein, LEP100, between plasma membrane and lysosomes: kinetic and morphological analysis

Intracelhilar transport of class II MHC molecules directed by invariant chain

Distribution of newly synthesized lysosomal enzymes in the endocytic pathway of normal rat kidney cells

A specific transmembrane domain of a coronavirns El glycoprotein is required for its retention in the Golgi

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

Phosphorylation of the cation-independent Mannose 6-phosphate receptor is closely associated with its exit from the trans-Golgi network

Transferrin receptors promote the formation of clathrin lattices

Sequences within and adjacent to the transmembrane segment of a-2,6-sialyltransferase specify Golgi retention. EMBO (Eur

Time course of intracellular associations, processing, and cleavages of Ii forms and class II major histocompatibility complex molecules

The membrane spanning domain of 8-1,4-galactosyltransferas¢ specifies trans Golgi localization

Lysosomal targeting of Limp II membrane glycoprotein requires a novel Leu-lle motif at a particular position in its cytoplasmic domain

Biosynthesis of the human transferrin receptor in cultured cells

Four la invariant chain forms derive from a single gene by alternate splicing and alternate initiation of transcription/translation

Targeting ofa lysosomal membrane protein: a tyrosine-containing endocytosis signal in the cytoplasmic tail of lysosomal acid phoaphatase is necessary and sufficient for targeting to lysosomes

Segregation of MHC class II molecules from MHC class I molecules in the Golgi complex for transport to lysosomal compartments

Intracellular transport and localization of major histocompatibility complex class H molecules and associated invariant chain

The MHC class IIassociatext invariant chain contains two endosomal targeting signals within its cytoplasmic tail

Identification of subcellular compartments involved in biosynthetic processing of cathepsin D

Invariant chain association with HLA-DR molecules inhibits immunogenic peptide binding

Stable surface expression of invariant chain prevents peptide presentation by HLA-DR

Cell surface HLA-DR-invariant chain complexes are targeted to endosomes by rapid internalization

Relationship between invariant chain expression and major histocompatibility complex class II transport into early and late endocytic compartments

DNA sequencing with chainterminating inhibitors

Intracelhilar distribution of the MHC class II molecules and the as-social~l invariant chain 0i) in different cell lines

A new method of preparing gold probes for multiple-labelling cytochemistry

Two forms of the la antigenassociated invariant chain result from alternative initiations at two in-phase AUGs

A Golgi retention signal in a membrane-spanning domain of coronavirus E1 protein

DNA sequence analysis with a modified bacteriophage T7 DNA polymerase

Insulin elicits a redistribution of transferrin receptors in 3T3-L 1 edipocytes through an increase in the rate constant for receptor externalization

Invariant chain distinguishes between the exogenous and endogenous antigen presentation pathways

A study of positive staining of ultrathin frozen sections

Human cell surface glycoprotein related to cell proliferation is the receptor for transferrin

Signal-dependent membrane protein trafficking in the endocytic pathway

Mice lacking the MHC class H-associated invariant chain

Exposure of K562 cells to anti-receptor monoclonal antibody OKT9 results in rapid redistribution and enhanced degradation of the transferrin receptor

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

The 17-residue transmembrane domain of ~-galactoside o~2,6-sialyltransferase is sufficient for Golgi retention

Human major histecompatibility complex class II invariant chain is expressed on the cell surface

A block in degrad-tion of MHC class H-associated invariant chain correlates with a reduction in transport from endosome carrier vesicles to the prelysosome compartment

We thank Adele Gibson for expert technical assistance and Albert Tousson for helpful discussions and advice on the immunofluorescence experiments. The LEP100 monoclonal antibody used in these studies was obtained from