The Journal of Immunology

Brief Residence at the Plasma Membrane of the MHC Class I-Related Chain B Is Due to Clathrin-Mediated Cholesterol-Dependent and Shedding1

Sonia Agu¨era-Gonza´lez,2 Philippe Boutet,2 Hugh T. Reyburn, and Mar Vale´s-Go´mez3

Recognition of MHC class I-related chain (MIC) molecules on the surface of target cells by the activating receptor NKG2D leads to their lysis by immune effector cells. Up-regulation of NKG2D ligands is broadly related to stress, although the detailed mo- lecular mechanisms that control the presence of these molecules at the plasma membrane are unclear. To investigate the post- translational mechanisms that control surface expression of the human NKG2D ligand MICB, we studied the subcellular local- ization and trafficking of this molecule. We found that in several cellular systems, the expression of MICB molecules on the cell surface is accompanied by an intracellular accumulation of the molecule in the trans-Golgi network and late endosome-related compartments. Surprisingly, MICB has a much shorter half-life at the plasma membrane than MHC molecules and this depends on both recycling to internal compartments and shedding to the extracellular medium. Internalization of MICB depends partially on clathrin, but importantly, the lipid environment of the membrane also plays a crucial role in this process. We suggest that the brief residence of MICB at the plasma membrane modulates, at least in part, the function of this molecule in the . The Journal of Immunology, 2009, 182: 4800–4808.

ajor histocompatibility complex class I-related instead, the expression of these is up-regulated in a chain (MIC)4 A and B are two MHC-like molecules number of pathological situations, mainly cancer and autoim- M encoded within the human MHC, but distinct from munity (for review, see Refs. 7 and 8). For example, MIC ex- ␤ classical HLA class I proteins in that they bind neither 2- pression has been found in tumors of various origins and the microglobulin nor antigenic peptides (1–4). MIC proteins bind presence of pathogens like Mycobacterium tuberculosis has the activating receptor NKG2D that is constitutively expressed been shown to up-regulate its expression (9). Recently, a num- ϩ on all NK cells and CD8 subpopulations and is up- ber of external stimuli such as DNA-damaging agents and pro- ϩ regulated after activation on CD4 T cells (5). Engagement of teasome inhibitors have been shown to induce expression of NKG2D by its ligands leads to the activation of lysis and cy- NKG2D ligands (10, 11). Thus, although the precise mecha- tokine by NK cells and T cells either directly or as a nisms that regulate MIC expression are still unknown, the pres- costimulatory factor (for review, see Ref. 6). Thus, the presence ence of MIC at the plasma membrane can be related broadly of MIC at the surface of putative target cells must be tightly with stress and this coincides with the presence of heat shock regulated so that the immune response is not triggered or hidden elements in the MIC promoters (3). The fact that MIC was inappropriately (which would lead to autoimmunity in the first detected by Western blot in a number of nontransformed cell instance and immune evasion in the latter). In fact, with the types, including keratinocytes, endothelial cells, and mono- exception of gastrointestinal epithelium, the vast majority of cytes, although they were not found at the cell surface of all of normal cells do not express MIC molecules at the cell surface; them (2, 12–14), suggests the existence of posttranscriptional and posttranslational mechanisms regulating MIC expression, in addition to transcriptional regulation. That NKG2D ligands Department of Pathology, University of Cambridge, United Kingdom can also be shed from the cells as soluble molecules represents Received for publication March 4, 2008. Accepted for publication February 4, 2009. an additional level of complexity in this system (15–17). The costs of publication of this article were defrayed in part by the payment of page MICB is a target for several proteins expressed by pathogens charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. that are involved in immune evasion. Interestingly, many of 1 This work was funded by a Leukemia Research Fund Project Grant (to H.T.R. and them are selective toward MICB and they do not affect the M.V.G.). M.V.G. is a recipient of a New Investigator Grant from the Medical Re- related MICA (18). There are still many questions about search Council and an International Joint Project from the Royal Society. S.A. was the large number of NKG2D ligands and whether this diversity supported by fellowships from the Fundacio´n Caja Madrid and Ibercaja. P.B. was supported in part by The Newton Trust. reflects somehow a difference in function (for review, see 2 S.A.-G. and P.B. contributed equally to the work presented. Ref. 19). With the goal of investigating the mechanisms that control 3 Address correspondence and reprint requests to Dr. Mar Vale´s-Go´mez, Department of Pathology, University of Cambridge, Tennis Court Road, Cambridge CB2 1QP, surface expression of MICB molecules at a posttranslational United Kingdom level, we studied the subcellular localization and trafficking of 4 Abbreviations used in this paper: MICA/B, MHC class I chain-related A/B; MICB. In the present study, we show, in several cell lines, that MESNA, 2,mercaptoethane sulfonic acid; DRM, detergent-resistant membrane; HCMV, human CMV; TGN, trans-Golgi network; CI-M6PR, cation-independent the expression of MICB molecules on the cell surface is ac- mannose-6-phosphate receptor; ER, ; EEA1, early endosome companied by an intracellular accumulation of the molecule in A1; LAMP-1, -activated membrane protein 1; CHX, cycloheximide; BFA, the trans-Golgi network (TGN) and late endosome-related com- brefeldin A; CHL, chloroquine; M␤CD, methyl-␤-cyclodextrin. partments. We further show that MICB has a short half-life at Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 the plasma membrane and that this depends on shedding and www.jimmunol.org/cgi/doi/10.4049/jimmunol.0800713 The Journal of Immunology 4801 clathrin-dependent endocytosis, although the integrity of cho- Table I. Colocalization Coefficients lesterol in the plasma membrane also influences the internal- ization of MICB protein. Pearson’s Coefficient Overlap

M6PR 0.458 0.641 Materials and Methods KDEL 0.347 0.527 Cells, reagents, and Abs CD63 0.200 0.299 EEA1 0.383 0.520 HeLa, U373, and CV1 cells were maintained in DMEM supplemented with LAMP-1 0.099 0.190 10% FCS and antibiotics (DMEM/10% FCS). 721.221 and HCT116 were cultured in RPMI 1640 medium with 10% FCS and antibiotics. 721.221, U373, and CV1 transfected with MICB were previously de- scribed (20, 21). Anti-MICA/B monoclonal and polyclonal Abs were pur- chased from R&D Systems, anti-MICA (AMO1) and anti-MICB (BMO2) After blocking for 1 h with PBS containing 2% BSA, tissue culture super- were from Immatics, anti-MICA/B 6D4 was from Santa Cruz Biotechnol- natant was incubated for1hat37°C. The assay was developed using ogy, and isotype control mouse mAbs were purchased from Sigma-Al- biotinylated goat anti-MICB (0.1 ␮g/ml) followed by incubation with drich. For microscopy, sheep Ab directed against human TGN46 was from streptavidin-HRP (1/2000; Amersham Biosciences) and a peroxidase sub- Ј Serotec, mouse mAb specific for GM130 and early endosome A1 (EEA1) strate system (2,2 -azino-di-[3-ethylbenzthiazoline sulfonate] and H2O2 in were from BD Biosciences, mouse anti-human CD107a (lysosome-acti- a glycine/citric acid buffer; Roche). The absorbance was measured at 410 vated membrane (LAMP-1)) was from Southern Biotechnology Associ- nm with a reference wavelength of 490. Samples were analyzed in ates; mouse KDEL mAb was from Bioquote; rabbit anti-mouse tetramethyl duplicates. rhodamine isothiocyanate from DakoCytomation; Alexa Fluor 564 donkey anti-sheep, Alexa Fluor 488 donkey anti-sheep, Alexa Fluor 488 goat Internalization experiments anti-mouse, and Alexa Fluor 488 and 564 goat anti-rabbit Abs were Ab uptake. Ab uptake experiments were performed as previously de- from Molecular Probes; and anti-CI-M6PR polyclonal rabbit Ab was a scribed (23). Cells were incubated with Ab for 10 min at 4°C and then gift from P. Luzio (Cambridge Institute for Medical Research, Cam- transferred to 37°C for different lengths of time. To detect internalized bridge, U.K.). Anti-caveolin Ab was from BD Transduction Laborato- protein, cells were fixed, permeabilized, stained with secondary Ab, and ries. Unless otherwise indicated, chemicals were purchased from analyzed by confocal microscopy. Sigma-Aldrich. Cleavable biotinylation. For cleavable biotinylation experiments, the cells were washed three times with ice-cold PBS. Then incubated 30 min with Flow cytometry 0.25 mg/ml sulfosuccinimidyl-2-(biotinamido)ethyl-1,3-dithiopropionate For flow cytometry, 105 cells were incubated with mouse mAbs and bound (EZ-Link Sulfo-NHS-SS-Biotin; Pierce) in ice-cold PBS containing 1 mM Ј MgCl and 0.1 mM CaCl (PBSϩ). The reaction was quenched by washing Ab was visualized using either PE- or FITC-labeled F(ab )2 of goat anti- 2 2 mouse Ig (DakoCytomation). Samples were analyzed using a FACScan II the cells three times with ice-cold serum-free medium containing 0.1% flow cytometer (BD Biosciences). BSA and then once with ice-cold PBSϩ. Prewarmed (to 37°C) serum-free medium was added for the desired time period(s) to allow internalization of Confocal microscopy the biotinylated surface components, after which they were returned to 4°C. After one rinse with PBSϩ, biotin was stripped by three 20-min Cells were fixed with 4% p-formaldehyde at room temperature for 10 min, incubations with ice-cold 100 mM sodium 2-mercaptoethane sulfonic acid permeabilized by incubation with 0.1% saponin at room temperature for 10 (MESNA) in 50 mM Tris-HCl (pH 8.6), 100 mM NaCl, 1 mM EDTA, and min, stained with commercial anti-MICB Abs (R&D Systems) or costained 0.2% BSA. The cells were rinsed quickly twice again with PBSϩ, and with markers for intracellular compartments, and analyzed by confocal residual MESNA was quenched by a 10-min incubation with ice-cold 120 microscopy as previously described (22). mM iodoacetamide in PBSϩ. After two additional rinses with ice-cold All fluorescence images were obtained using a confocal microscope PBS, the cells were lysed and MICB was immunoprecipitated. The proteins (either a Leica DM IRE2 or IRBE confocal laser-scanning microscope). were resolved in SDS-PAGE and biotinylated proteins detected by Western Excitation wavelengths and filter sets were as supplied by the manufacturer blot with either streptavidin-HRP or MICB-specific Ab. (lines at 488 and 568 nm for Alexa Fluor 488 and Alexa Fluor 568, re- Estimation of MICB internalization by flow cytometry. Internalization of spectively). To reduce Fo¨rster resonance energy transfer, the emission MICB molecules was quantitated using a variant of a previously described spectrum for the first fluorochrome was made distinct from the excitation flow cytometric assay for estimating receptor internalization (24). Briefly, spectrum of the second, i.e., the setting of the spectra was narrowed down cells stably transfected with MICB were transiently transfected with plas- to 500–560 and 580–660 nm for Alexa Fluor 488 and Alexa Fluor 568, ϩ ϫ mids encoding the AP180 C terminus, epsin1 R63L H73L, and epsinR respectively. Images of fixed cells were taken using a 63 1.32 aperture (gift from Dr. H McMahon, LMB, Cambridge, U.K.) and EH29 using objective with the confocal pinhole set to 1Airy unit. JETPEI (PolyPlus Transfection SA). Twenty-four hours later, the cells Image acquisition. Images were obtained by scanning either a single focal were detached by incubation in PBS/0.2% BSA/10 mM EDTA, washed plane or a series of them across the cell using Leica TCS software. For twice (PBS/BSA), and resuspended in binding buffer (Ham’s F12 medium colocalization experiments, acquisition of images was conducted by suc- with 0.2% BSA, 10 ␮M GM6001, and 1 mM HEPES). Transfected cells cessive scanning of the 488 (green) and the 568 (red) channel, exciting one were divided into two aliquots and while one was held at 37oC, the other fluorochrome at a time by successive scanning of the green and the red was cooled to 4oC. MICA/B-specific mAb (2.5 ␮g/ml) was then added to channels independently. To explore the whole intracellular area, series each aliquot and the cells were incubated either at 37oC or on ice for 30 of sections (total interval z ϭ 2–4 ␮m) were acquired. Image acquisition min. After this time, the cells were washed into PBS/0.2% BSA/0.1% of the subsequent sections was conducted near the resolution limit of sodium azide and placed on ice followed by staining with PE-conjugated the optical system used, i.e., the optimal for a good analysis of colo- secondary goat anti-mouse Ig. Cells overexpressing the various constructs calization: 1024 ϫ 1024 pixels at z intervals of 0.1221 ␮m with a were gated on GFP fluorescence and the amount of MICB molecules re- magnification of 3–4 times. Colors and overlays shown in the figures maining in the plasma membrane after the surface labeling with specific were as collected by the microscope software and no further threshold mAb was determined for each condition. Duplicate samples (5 ϫ 104 cells/ adjustments were made. Adobe Photoshop software was used for com- sample) were analyzed in each experiment. The reduction in surface re- position of the figures. ceptor fluorescence after incubation at 37oC compared with 4oC represents Image analysis and quantitation. For visualization of colocalization, im- internalization of receptors. ages of composed overlays are presented in the manuscript; however, to analyze the degree of colocalization taking in account the three-dimen- Pulse-chase experiments sional nature of the sample, imaging processing was performed (data are Cells were washed and resuspended in either methionine-cysteine-free me- included in Table I). Huygens Deconvolution Software (Scientific Volume ␮ Imaging) was used for deconvolution and analysis. dium alone or with 100 M chloroquine or control. After a 30-min incu- bation at 37 °C, cells were spun and resuspended at 107 cells/ml with 35 ELISA PRO-MIX L-[ S] In Vitro Cell Labeling Mix (Amersham Biosciences) at 0.3 mCi/ml. Cells were incubated at 37°C for 15 min and then chased using Detection of soluble MICB was performed using a sandwich ELISA pro- complete medium with 10% FCS for the indicated times. Each time point cedure. The anti-MICB mAb from R&D Systems was used at 5 ␮g/ml. sample was washed in cold PBS and resuspended in lysis buffer (50 mM 4802 CELLULAR TRAFFICKING OF MICB

Tris (pH 7.6), 150 mM NaCl, 5 mM EDTA, 1% (v/v) Nonidet P-40, 1 ␮g/ml leupeptin, 1 ␮g/ml pepstatin, and 5 mM iodoacetamide). After a 30-min incubation on ice, samples were precleared by incubation, first with Pansorbin (Calbiochem), followed by protein A-Sepharose beads (Phar- macia). MICB was immunoprecipitated by incubation on ice with the ap- propriate Ab and recovered using protein A beads. Samples were washed three times and resuspended in SDS-PAGE-loading buffer. Proteins were resolved in SDS-PAGE. Gels were stained and exposed on Biomax MR film (Kodak). Immunoprecipitation Cells were lysed in 50 mM Tris (pH 7.6), 150 mM NaCl, 5 mM EDTA, 1% Nonidet P-40, 1 ␮g/ml leupeptin, and 1 ␮g/ml pepstatin. After a 30-min incubation on ice, samples were precleared by incubation with Pansorbin (Calbiochem). MICB was immunoprecipitated by incubation on ice with a mixture of the MICB-specific Abs BMO2 and 6D4 and recovered using protein G beads (Pharmacia). Samples were washed three times and SDS- PAGE-loading buffer was added. Proteins were resolved in SDS-PAGE. Western and dot blots Cell lysates were run on 12% SDS-PAGE gels and transferred to Immo- bilon-P (Millipore). The membrane was blocked using PBS containing 0.1% Tween 20 (PBS-T) and 5% nonfat dry milk. Detection of MICB was performed by incubating the membrane with biotinylated goat polyclonal anti-MICB Ab, followed by HRP-conjugated secondary reagent. Proteins were visualized using the ECL system (Amersham Pharmacia). Detection of ganglioside GM1 was performed by immobilizing 5 ␮l of the above fractions in nitrocellulose membrane, incubating with cholera toxin-HRP (Sigma-Aldrich), and detecting with ECL. Detergent-resistant membrane (DRM) fractionation Cells were lysed in TNE buffer (20 mM Tris (pH 7.5), 150 mM NaCl, and 5 mM EDTA) containing 1% Brij-58 and protease inhibitors and homog- enized using a Dounce homogenizer. Samples were diluted with an equal volume of 85% sucrose in TNE and then placed at the bottom of a step sucrose gradient and fractionated by centrifugation for 18–20 h at 200,000 ϫ g (25). One-milliliter fractions were collected from top to bot- tom and made to a final concentration of 0.2% deoxycholate. Proteins were separated on SDS-PAGE. Western blot was performed as described above FIGURE 1. A large proportion of MIC is intracellular. A, Flow cytom- using Abs against MICA/B and caveolin-1. etry analysis of surface MIC molecules. HCT116, HeLa, U373-MICA, and Efficiency of fractionation was assessed by detection of ganglioside U373-MICB were stained with anti-MICA (AMO1) and anti-MICB GM1 using cholera toxin-HRP. (BMO2) Abs as indicated or isotype control. B, Confocal microscopy of intracellular MICA and MICB in HCT116 cells. Cells were stained with Results anti-MICA (AMO1) and anti-MICB (BMO2) Abs and processed for con- MIC expression on the cell surface is accompanied by an focal microscopy. Images show a single focal plane across the cell (depth intracellular accumulation of the molecule in late of the plane 0.2849 ␮m) using the ϫ63 objective with a ϫ4 magnification. endosome/TGN-related compartments The series of images corresponding to the central part of the same cell are shown in supplemental Fig. 1. Scale bar, 4 ␮m. C, Confocal microscopy of In experiments exploring the mechanism of immune evasion used intracellular MICB in U373- and CV1-MICB transfectants. Cells were by human CMV (HCMV), we previously showed that HCMV- stained with anti-MICA/B Ab and processed for confocal microscopy. Im- infected cells lost surface expression of MICB and accumulated ages were obtained by scanning a single focal plane across the cell using MIC intracellularly. Interestingly, however, we observed that un- the ϫ63 objective without magnification. Scale bar, 20 ␮m. infected cells also expressed a large amount of intracellular protein (23). To study this observation further, we analyzed the distribu- tion of these molecules in a number of cell lines: HeLa and the alleles present in each cell line. For these reasons, we decided to colon carcinoma HCT116 naturally express MICA and MICB; the focus on the study of transfectants in the following experiments. glioma cell line U373 and the monkey cell line CV1 were stably To study the intracellular distribution of MICB molecules, cells transfected with MIC molecules (U373/CV1-MICA and -MICB). were costained with markers for intracellular compartments and The surface expression of MICA and MICB was studied by flow analyzed in confocal microscopy. Visualization of colocalization is cytometry (Fig. 1A). In general, MICB expression at the plasma presented as an overlay of the channels corresponding to MICB membrane was lower than that of MICA, but MICB was found and the markers for different compartments (Fig. 2). The perinu- inside HCT116 cells in a larger proportion than MICA (Fig. 1B; clear area where more MICB staining is found was magnified for 5 whole intracellular area supplemental Fig. 1 ). Examination of intracellular analysis and is also included in Fig. 2. To explore the both U373 and CV1 transfectants by confocal microscopy also whole intracellular area, series of sections were acquired through showed quite a large proportion of MICB intracellularly (Fig. 1C). the cell (supplemental Fig. 2) and qualitative evaluation and quan- Since the tumor cell lines HCT116 and HeLa express both MICA tification of data for colocalization purposes were performed (see and MICB, it is difficult to confidently discriminate between these Materials and Methods). The colocalization coefficients with the molecules, particularly since the Abs raised against either MICA different markers are shown in Table I. or MICB could have different cross-reactivity depending on the MICB colocalized partially with the TGN markers TGN46 (data not shown) and cation-independent mannose-6-phosphate receptor 5 The online version of this article contains supplemental material. (CI-M6PR), as well as partially with the endoplasmic reticulum The Journal of Immunology 4803

To confirm the TGN localization of MICB, cellular subfraction- ation was performed and the fractions were analyzed for the pres- ence of MICB and organelle markers (supplemental Fig. 3). MICB was detected in fractions enriched for markers of TGN, as well as in the fractions in which ER and early endosomes markers were found. The confocal microscopy and cellular subfractionation data would be consistent with MICB traveling between the TGN and the late endosomal system.

MICB has a short half-life at the plasma membrane The presence of MICB molecules in late endosome/TGN-type or- ganelles suggests that the molecule might be traveling within var- ious compartments of the endocytic pathway and that some of the observed intracellular molecules could be the result of down-mod- ulation from the plasma membrane. To follow the fate of the pro- tein after it reaches the plasma membrane, Ab uptake experiments were performed and analyzed by confocal microscopy. After a 30-min incubation with Ab, a large proportion of MICB previously present at the plasma membrane was found intracellularly. The internalized MICB was then costained with markers for intracel- lular compartments (supplemental Fig. 4, A and B). At short incu- bation times, MICB was observed in vesicles that did not colocal- ize with CI-M6PR, although after longer incubation times, some colocalization between MICB and CI-M6PR was noted. The kinetics of MICB endocytosis was also studied in flow cy- tometry experiments. 721.221 cells transfected with MICB were stained with PE-labeled MICB-specific Ab and incubated for dif- ferent periods of time to allow internalization. After the indicated times, Ab bound at the plasma membrane was removed by acid wash (29) and the Ab that remained in the internal compartments was detected by flow cytometry. Surface-bound Ab was efficiently removed by acid wash without affecting viability of the cells (at time 0 min, 94% of cells lost all of the staining after the acid wash FIGURE 2. MIC colocalizes with markers for TGN and endosomes. and were still viable), and after incubation to allow internalization, CV1-MICB transfectants were fixed, permeabilized, and costained for endocytosed MICB staining was still observed (supplemental MICB (green panels) and markers of intracellular compartments (CI- M6PR (TGN), KDEL (ER), CD63 (late endosome, MVB), EEA1 (early Fig. 4c). endosome), LAMP1 (lysosome, late endosome), and GM130 medial To rule out that the observed internalization of the protein Golgi)] (red panels, as indicated). Images were obtained by successive was due to cross-linking with Ab, the phenomenon was studied scanning of the green and red fluorochromes (see Materials and Methods). in a biochemical assay independent of Ab. Cells were surface The first three columns show a single focal plane of a central region in the labeled with cleavable biotin and, after incubation at 37°C to z-axis of the series of focal planes acquired (full series shown in supple- allow internalization of the protein, the cells were treated with mental Fig. 2) using the ϫ63 objective. In parallel, images of the same cells MESNA (a cell-impermeant reducing agent) to remove the bi- were obtained with a ϫ4 magnification (shown in the last column) to otin moiety from the labeled proteins that remain at the plasma optimize the optical parameters for colocalization analysis. The experiment membrane, after which only intracellular labeled proteins are was performed five times and an average of four cells per marker were detected. As a control for the proper elimination of surface- analyzed per experiment. Similar results were obtained with the U373- MICB transfectants. Scale bars: merge, 20 ␮m; zoom, 5 ␮m. labeled proteins, cells were incubated at 4°C and treated with MESNA. No biotinylated protein was detected, indicating that no MIC protein had been internalized (Fig. 3A, lane 2). As a (ER) marker, KDEL, early endosomes (EEA1), and the late endo- positive control, cells incubated at 4°C but not treated with some marker CD63. Little colocalization was found with the other MESNA showed the total amount of cell surface biotinylated markers used: ER (PDI, ERp72; data not shown), medial Golgi MIC (Fig. 3A, lane 1). This experiment shows a band corre- (GM130), and LAMP-1. The CI-M6PR is involved in the sorting and sponding to internalized protein in cells treated with MESNA transport of newly synthesized lysosomal hydrolases that contain after incubation at 37°C. This result confirms the ability of mannose-6-phosphate moieties on their N-glycan chains, which serve MICB to internalize. The various size bands of MICB observed as sorting signals for TGN-to-endosome transport and in steady state in these experiments probably represent posttranslational mod- can be used as a marker for TGN (26). LAMPs are lysosomal integral ifications such as glycosylation. However it is also clear from membrane glycoproteins and LAMP-1 is often used as a lysosomal this experiment that a significant amount of MIC protein is lost marker, whereas CD63 (LIMP-I, LAMP-3) is a member of the tet- during this incubation at 37oC. This could include both degra- raspanin superfamily enriched in intraluminal vesicles in late endo- dation and shedding to the extracellular medium. somes/ that can travel from the TGN to lysosomes (27). It Given that MIC proteins are known to be shed from the cell has been shown recently that LAMP-1 and LAMP-2 are present in surface (15, 16), we next studied the dynamics of MIC expression separate TGN-derived vesicles from those containing mannose phos- and shedding in this system. To assess the stability of MICB phate receptor and the adaptor protein AP-1 (28). molecules at the cell surface, flow cytometry experiments were 4804 CELLULAR TRAFFICKING OF MICB

FIGURE 3. MICB has a short half-life at the plasma membrane. A, Cleavable biotinylation. U373-MICB cells were labeled with disulfide-biotin. Labeled proteins were allowed to internalize for2hat37°C and, after that period of time, surface biotin was removed by treatment with MESNA. Biotinylated MICB (left panel) and total MICB (right panel) were assessed by Western blot (WB). The total amount of surface biotinylation can be observed in the lane corresponding to the control at 4°C untreated with MESNA (lane 1); removal of all surface biotinylation labeling corresponds to the control at 4°C treated with MESNA (lane 2). A band corresponding to internalized MICB can be observed in the lane corresponding to cells treated with MESNA after incubation at 37°C. This experiment is representative of four experiments. B, Flow cytometry analysis of surface MICB molecules after treatment of U373-MICB cells with BFA, BFA and CHX (BFA ϩ CHX) and CHL as indicated. After incubation with the compound for the indicated time, U373-MICB cells were stained with anti-MICA/B Ab. Similar results were obtained with the cell line 721.221. Data are representative of three experiments. C, Soluble MICB in the tissue culture supernatant of U373-MICB cells was detected by ELISA. A sandwich ELISA was performed using anti-MICB mAb from R&D Systems as catching Ab and biotinylated goat anti-MICB for detection, followed by incubation with streptavidin-HRP. The absorbance was measured at 410 nm with a reference wavelength of 490. Samples were analyzed in duplicates. The dotted line indicates the background for the assay (medium from untransfected cells). The inset shows SDS-PAGE and Western blot analysis using anti-MICA/B Ab of tissue culture supernatant from untransfected cells (Med) and U373-MICB transfectants. The 50-kDa molecular mass marker is indicated. Similar results were obtained with the CV1-MICB transfectants. Experiments were repeated three times. D, The maturation of MICB does not change in the presence of CHL. Pulse-chase experiment followed by immunoprecipitation of MICB (see Materials and Methods) from cells either untreated or treated with 100 ␮M CHL (raises the pH of lysosomes). The proteins were resolved in SDS-PAGE. This experiment was performed twice.

performed using cycloheximide (CHX, to block protein synthe- loss of MICB seen by3hofchase confirmed biochemically the sis), brefeldin A (BFA, inhibitor of vesicular trafficking), and a short half-life of MICB molecules suggested by the FACS combination of both (Fig. 3B). Interestingly, treatment with BFA experiments. and CHX led to a significant loss of MICB after 4 h, as did treat- All together, these data suggest that, at steady state, MICB re- ment with CHX alone (data not shown). A significant component mains only for a short time at the plasma membrane and that the of surface MICB must be involved in vesicular trafficking since the amount of MICB at the cell surface is most probably regulated by inclusion of BFA alone in the experiment also provoked a marked a combination of endocytosis and protein loss (by degradation and decrease in surface staining. Similar data were obtained in exper- shedding). iments with 721.221 cells transfected with MICB (data not shown). These data show that MICB is different than class I MHC MICB endocytosis and clathrin-dependent pathways proteins, which remain stably at the plasma membrane for 18–24 The mechanism of MICB endocytosis was next investigated. It h (30). An inhibitor of protein degradation, such as the weak base is generally accepted that there are two main endocytic path- chloroquine (CHL), that raises the pH of late endosomes and ly- ways based on the dependence or independence of clathrin (for sosomes, impairing their function, was also included in the exper- review, see Refs. 31–35). Both processes can be explored by the iments. The stability of MICB at the cell surface was not affected use of mutants or drugs that interfere with specific components by the presence of this compound. The shedding of MICB to the necessary for the correct function of the pathway (reviewed in tissue culture medium was analyzed by ELISA (Fig. 3C). Signif- Ref. 36). icant shedding of MICB was observed after1hofincubation and Because the cytoplasmic tail of MICA/B molecules contain a this soluble MIC accumulated over time (Fig. 3, inset). Treatment dileucine motif that could bind AP2, we first studied clathrin- with BFA, CHX, or CHL did not inhibit shedding. The fact that dependent endocytosis, a process that requires a variety of MICB is present in lysosome-related organelles, but that the levels adaptor proteins and the generation of coated pits that bud into of surface and soluble protein do not change when CHL is added vesicles. suggest that the protein must be trafficking through these compart- Eps15 is an adaptor protein found in clathrin-coated pits re- ments rather than merely being degraded in lysosomes. This was quired for the early steps of clathrin-dependent endocytosis. confirmed by the inability of CHL treatment to alter the loss of The EH29 dominant-negative mutant of Eps15 affects the as- MICB observed in pulse-chase experiments (Fig. 3D). The marked sembly of clathrin-coated pits and strongly inhibits transferrin The Journal of Immunology 4805

FIGURE 4. MICB endocytosis and clathrin-mediated pathways. A, U373-MICB cells were transfected with the indicated dominant-negative mutants. EH29 affects the generation of coated pits by inhibiting Eps15; the inactive variant of Eps15 (DIII⌬2) was used as a control. All of the dominant-negative mutants are expressed as fusions to GFP to allow visualization of transfected cells. Cells were incubated with anti-MIC Ab at 4°C, followed by uptake of the Ab at 37°C for 20 min. After that, cells were fixed, permeabilized, and stained with secondary Ab coupled to Alexa Fluor 568 (red) and analyzed by confocal microscopy. Similar results were obtained with the CV1-MICB transfectants. Scale bar, 20 ␮m. This experiment was performed four times. B, Flow cytometric estimation of MICB internalization. U373-MICB cells were transiently transfected with plasmids to disrupt endocytosis by overexpression of AP180 C terminus, epsin1 R63L ϩ H73L, epsinR, and EH29. The reduction of MICB at the plasma membrane produced by incubation of cells at 37°C for 30 min was determined by flow cytometry. Bars represent reduction in surface receptor p Ͻ 0.05 in a t test compared with mock. Cells were cultured in the presence ,ءء .(fluorescence (relative to control U373-MICB cells held at 4°C of GM6001 to inhibit metalloprotease-mediated shedding during the 30-min incubation with Ab. Data were obtained in duplicate and are repre- sentative of two experiments. C, U373-MICB cells were preincubated for 30 min with 8 ␮g/ml CHL. After the treatment, Ab uptake experiments were performed as in A. Scale bar, 40 ␮m. This experiment was repeated three times. receptor internalization (37, 38). The effect of this construct on epsinR (inhibits AP1 budding events), and EH29 (Fig. 4B). We MICB endocytosis was studied in Ab uptake experiments, in observed a marked reduction in internalization of cell surface which an inactive variant of Eps15 (DIII⌬2) was used as a MICB in cells transfected with AP180, epsin1 R63L ϩ H73L, control. To identify transfected cells, these variants of Eps15 and EH29. are tagged with GFP (Fig. 4A). After 20 min of Ab uptake, we In parallel, experiments using other methods to interfere with could observe an accumulation of MICB in mock-transfected clathrin-mediated endocytosis were also performed. Hypertonic cells and cells expressing the inactive mutant, whereas in cells medium promotes the disappearance of clathrin-coated vesicles transfected with EH29 a considerable proportion of MICB was and plasma membrane-coated pits by preventing the interaction of still at the plasma membrane. In supplemental Figs. 5 and 6, clathrin with adaptors required for endocytosis (39). Clathrin-de- parallel experiments performing series of sections through the pendent endocytosis can also be inhibited by the use of chlorprom- whole intracellular area are depicted. In these experiments, it is azine (40). Thus, Ab uptake experiments were also performed in still possible to observe some intracellular MICB after inhibi- the presence of chlorpromazine (Fig. 4C) and 0.32 M sucrose (sup- tion of clathrin in the EH29 mutant, although not as much as in plemental Fig. 7A). The presence of these drugs impaired the up- the inactive mutant DIII⌬2. Since clathrin blockade did not take of MICB, confirming the data obtained previously. completely abolish MIC endocytosis, these data suggest that another mechanism of internalization might be involved. MICB endocytosis and clathrin-independent pathways To estimate the amount of MICB that was internalized via Clathrin-independent endocytosis includes a group of pathways clathrin, flow cytometry experiments were performed. U373- still not well characterized, but often associated with plasma mem- MICB cells were transiently transfected with plasmids encoding brane domains known as lipid rafts/caveolae. These microdomains the AP180 C terminus (inhibits clathrin-dependent budding are enriched in cholesterol, sphingolipids, and caveolin (for a re- events), epsin1 R63L ϩ H73L (inhibits AP2 budding events), view, see Refs. 41 and 42). Endocytosis via lipid rafts/caveolae is 4806 CELLULAR TRAFFICKING OF MICB

FIGURE 5. MICB endocytosis and clathrin-independent pathways. A, U373-MICB cells were preincubated for 30 min with 10 mM M␤CD. For labeling, either fluorescent-cholera toxin or fluorescent transferrin were added during the preincubation period as indicated. Ab uptake experiments were performed as in Fig. 4A. B, Flow cytometric estimation of MICB inter- nalization. U373-MICB cells were treated with either M␤CD or sucrose. The reduction of MICB at the plasma membrane produced by incubation of cells at 37°C for 30 min was deter- mined by flow cytometry. Bars repre- sent reduction in surface receptor fluorescence (relative to control U373-MICB cells held at 4°C). Cells were cultured in the presence of GM6001 to inhibit metalloprotease- mediated shedding during the 30-min p Ͻ 0.05 and ,ءء .incubation with Ab p Ͻ 0.005 in a t test compared ,ءءء with mock. Data were obtained in du- plicate and are representative of two experiments. Scale bar, 40 ␮m.

frequently defined using agents that disrupt or deplete cholesterol that treatment with M␤CD reduced the uptake of mAb specific at the plasma membrane. for MICA/B (Fig. 5B). To find out whether “caveolin-mediated” endocytosis is in- These results suggest that MICB is internalized through a clath- volved in MICB down-modulation, we studied the ability of the rin-mediated, cholesterol-dependent pathway. Such a mechanism protein to recycle from the plasma membrane in the presence of has been previously described for the entry of and bacterial methyl-␤-cyclodextrin (M␤CD). M␤CD binds to cholesterol toxins in the cell (45, 46). and selectively extracts it from the outer leaflet of the plasma membrane. We observed that cells preincubated with M␤CD MICB associates with DRMs accumulated MICB at the plasma membrane (Fig. 5A), suggest- The finding that the trafficking of MICB could be influenced by the ing a role for cholesterol in the trafficking of MICB. As con- lipid environment led us to investigate the association of this pro- trols, fluorescently labeled transferrin and cholera toxin were tein with DRM domains. DRMs are microdomains rich in choles- included in the experiment. As expected, when M␤CD is terol and sphingolipids, especially GM1 ganglioside (41, 42). present, a large proportion of cholera toxin remains at the DRMs can be separated from other membrane components by plasma membrane, indicating a lower efficiency in its uptake. ultracentrifugation in sucrose gradients after solubilizing using de- However transferrin, a protein internalized through clathrin- tergents such as Triton X-100 or Brij-58 (47). Caveolin-1, a com- coated pits, did get internalized, although with lower efficiency ponent of caveolae, is usually found in DRMs and can be used as than in control cells. It has been previously shown that acute a marker for the efficiency of fractionation. As shown in Fig. 6, a depletion of cholesterol also reduces the rate of internalization proportion of MICB was found in detergent-insoluble fractions of transferrin (43, 44). In supplementary Fig. 7B, a parallel rich in GM1 sphingolipid and caveolin-1. This finding confirms the experiment performing series of sections through the whole in- location of at least some MICB molecules in domains of the mem- tracellular area is depicted. Flow cytometry analysis confirmed brane rich in cholesterol. The Journal of Immunology 4807

context, the short half-life of MICB would mean a rapid loss of surface protein, even after induction of MIC transcription. There are still outstanding questions about the large number of ligands recognized by NKG2D and the way that several pathogens have evolved to affect particular ligands (18, 19). For example, although only one HCMV product has been found to down-mod- ulate MICA (52), several mechanisms target MICB, including UL16 (21, 53) and recently discovered virally encoded micro- RNAs (54). The features of MICB trafficking described here FIGURE 6. Association of MICB with DRMs. U373-MICB cells might favor encounter with the hijacker UL16 protein within were lysed in buffer containing 1% Brij-58 and homogenized using a the cytoplasmic compartments or, alternatively, the viral pro- Dounce homogenizer. Samples were fractionated in a discontinuous tein could simply enhance a process that already happens in the sucrose gradient. Fractions were collected and proteins were separated uninfected cell. on SDS-PAGE. Western blot was performed using Abs against MICA/B and Another finding from this work is that the mechanism of MICB caveolin-1 (Cav-1). Efficiency of fractionation was assessed by detection of endocytosis is dependent on clathrin but also affected by choles- ganglioside GM1 using cholera toxin-HRP on dot blots. This result is repre- terol. The involvement of lipids in clathrin-mediated pathways has sentative of at least five independent experiments and similar results were been previously reported, especially for the uptake of toxins and obtained with the CV1-MICB transfectants. viruses toward the (45, 46). Also, internalization of the apolipoprotein E receptor 2 internalizes through a clathrin-medi- ated pathway after association with caveolar and noncaveolar lipid Discussion rafts (55). Therefore, recent data support the existence of “hybrid” Given that recognition of NKG2D-L at the cell surface by their endocytosis and trafficking pathways in the cell whose mecha- receptor leads to killing of the target cell, a very tight regulation of nisms and functional significance remain unclear. The possibility the expression of these proteins must exist. There are examples in of association of MICA with membrane lipids has been suggested the literature of cells that express mRNA for one or more of the (56); however, the presence of MIC protein in DRMs was not NKG2D-L but do not show any expression of protein at the plasma demonstrated biochemically. In this study, we could observe a pro- membrane (48). A posttranslational mechanism that involves the portion of MICB in DRMs and future work will focus on exploring degradation of NKG2D-L RNA has been suggested as a manner of the extent and significance of this finding. Further studies are also regulating the amount of these proteins in the cell (14). In this needed to fully understand the mechanisms by which MICB asso- study, we propose that the amount of MICB present at the cell ciates with lipids and the implications for trafficking and organi- surface can also be modulated by the particular features of the zation in the plasma membrane. trafficking of the protein. Indeed, the main finding of the study is the short time of residence of MICB at the cell surface, confirmed Acknowledgments by its rapid loss from the plasma membrane upon incubation with We thank N. Miller for assistance with cell sorting; P. Luzio for CI-M6PR- inhibitors of protein synthesis. Several cellular processes seem to specific Ab; E. Veiga, R. Cassady-Cain, P. Roda-Navarro, and C. Chan for be involved in the loss of MICB from the cell surface, including helpful discussions; J. Trowsdale and A. Kelly for critical reading of this internalization toward intracellular compartments and shedding to manuscript; J. Skepper for assistance with quantitation analysis; and T. Fischer for helpful discussions. the extracellular medium. We show that MICB is endocytosed and trafficks through TGN Disclosures and the endosomal/lysosomal system but most probably recycles The authors have no financial conflict of interest. back to the plasma membrane, since inhibitors of vesicular traf- ficking affect the amount of MICB at the cell surface (Fig. 2B). It References is clear that MICB is not present in lysosomes (membrane-bound 1. Bahram, S., M. Bresnahan, D. E. Geraghty, and T. Spies. 1994. A second lineage acidic organelles that contain mature acid-dependent hydrolases of mammalian major histocompatibility complex class I . Proc. Natl. Acad. Sci. USA 91: 6259–6263. and LAMPs but lack mannose-6-phosphate receptors), since it co- 2. Zwirner, N. W., M. A. Fernandez-Vina, and P. Stastny. 1998. MICA, a new localizes with CI-M6PR and the increase of lysosomal pH does not polymorphic HLA-related antigen, is expressed mainly by keratinocytes, endo- thelial cells, and monocytes. Immunogenetics 47: 139–148. affect the expression and shedding of MICB. Thus, MICB could be 3. Groh, V., S. Bahram, S. Bauer, A. Herman, M. Beauchamp, and T. Spies. 1996. present in specialized compartments that share features of late en- Cell stress-regulated human major histocompatibility complex class I gene ex- dosomes/lysosomes but are functionally and morphologically dis- pressed in gastrointestinal epithelium. Proc. Natl. Acad. Sci. USA 93: 12445–12450. tinct, such as the MHC class II compartment or other secretory 4. Groh, V., A. Steinle, S. Bauer, and T. Spies. 1998. Recognition of stress-induced organelles (49, 50). However, multiple routes converge in the TGN MHC molecules by intestinal epithelial ␥␦ T cells. Science 279: 1737–1740. and it will be important to further characterize the pathway fol- 5. Bauer, S., V. Groh, J. Wu, A. Steinle, J. H. Phillips, L. L. Lanier, and T. Spies. 1999. Activation of NK cells and T cells by NKG2D, a receptor for stress- lowed by these molecules. inducible MICA. Science 285: 727–729. The short half-life of MICB at the plasma membrane of putative 6. Moretta, A., C. Bottino, M. Vitale, D. Pende, C. Cantoni, M. C. Mingari, R. Biassoni, and L. Moretta. 2001. Activating receptors and coreceptors involved target cells is likely to have a direct effect on their function in in human -mediated cytolysis. Annu. Rev. Immunol. 19: recognition by the immune system. Given that NKG2D-L are in- 197–223. dicators of cellular stress, their brief permanence at the cell surface 7. Collins, R. W. 2004. Human MHC class I chain related (MIC) genes: their bio- logical function and relevance to disease and transplantation. Eur. J. Immuno- suggests that continued expression requires continued stress. This genet. 31: 105–114. may be a mechanism to minimize immune elimination of cells 8. Mistry, A. R., and C. A. O’Callaghan. 2007. Regulation of ligands for the acti- during transient situations of cell damage. Moreover, a rapid loss vating receptor NKG2D. Immunology 121: 439–447. 9. Das, H., V. Groh, C. Kuijl, M. Sugita, C. T. Morita, T. Spies, and J. F. Bukowski. of MICB protein might be of crucial importance for the recogni- 2001. MICA engagement by human V␥2V␦2 T cells enhances their antigen- tion of pathogens that affect the cellular machinery of transcrip- dependent effector function. Immunity 15: 83–93. 10. Gasser, S., S. Orsulic, E. J. Brown, and D. H. Raulet. 2005. The DNA damage tion/translation. For example, it has long been known that herpes- pathway regulates innate immune system ligands of the NKG2D receptor. Nature viruses switch off the synthesis of cellular proteins (51). In this 436: 1186–1190. 4808 CELLULAR TRAFFICKING OF MICB

11. Vales-Go´mez, M., S. E. Chisholm, R. L. Cassady-Cain, P. Roda-Navarro, and 34. Mousavi, S. A., L. Malerod, T. Berg, and R. Kjeken. 2004. Clathrin-dependent H. T. Reyburn. 2008. Selective induction of NKG2D ligand expression by pro- endocytosis. Biochem. J. 377: 1–16. teasome inhibitors. Cancer Res. 68: 1–9. 35. Kirkham, M., and R. G. Parton. 2005. Clathrin-independent endocytosis: new 12. Zwirner, N. W., K. Dole, and P. Stastny. 1999. Differential surface expression of insights into caveolae and non-caveolar lipid raft carriers. Biochim. Biophys. Acta MICA by endothelial cells, fibroblasts, keratinocytes, and monocytes. Hum. Im- 1746: 349–363. munol. 60: 323–330. 36. Marks, D. L., R. D. Singh, A. Choudhury, C. L. Wheatley, and R. E. Pagano. 13. Jinushi, M., T. Takehara, T. Tatsumi, T. Kanto, V. Groh, T. Spies, T. Suzuki, 2005. Use of fluorescent sphingolipid analogs to study lipid transport along the T. Miyagi, and N. Hayashi. 2003. Autocrine/paracrine IL-15 that is required for endocytic pathway. Methods 36: 186–195. type I IFN-mediated dendritic cell expression of MHC class I-related chain A and 37. Benmerah, A., M. Bayrou, N. Cerf-Bensussan, and A. Dautry-Varsat. 1999. In- B is impaired in hepatitis C . J. Immunol. 171: 5423–5429. hibition of clathrin-coated pit assembly by an Eps15 mutant. J. Cell Sci. 112: 14. Boissel, N., D. Rea, V. Tieng, N. Dulphy, M. Brun, J. M. Cayuela, P. Rousselot, 1303–1311. R. Tamouza, P. Le Bouteiller, F. X. Mahon, et al. 2006. BCR/ABL oncogene 38. Benmerah, A., V. Poupon, N. Cerf-Bensussan, and A. Dautry-Varsat. 2000. Map- directly controls MHC class I chain-related molecule A expression in chronic ping of Eps15 domains involved in its targeting to clathrin-coated pits. J. Biol. myelogenous leukemia. J. Immunol. 176: 5108–5116. Chem. 275: 3288–3295. 15. Salih, H. R., H. G. Rammensee, and A. Steinle. 2002. Cutting edge: down-reg- 39. Hansen, S. H., K. Sandvig, and B. van Deurs. 1993. Clathrin and HA2 adaptors: ulation of MICA on human tumors by proteolytic shedding. J. Immunol. 169: effects of potassium depletion, hypertonic medium, and acidification. 4098–4102. J. Cell Biol. 121: 61–72. 16. Groh, V., J. Wu, C. Yee, and T. Spies. 2002. Tumour-derived soluble MIC li- 40. Singh, R. D., V. Puri, J. T. Valiyaveettil, D. L. Marks, R. Bittman, and gands impair expression of NKG2D and T-cell activation. Nature 419: 734–738. R. E. Pagano. 2003. Selective caveolin-1-dependent endocytosis of glycosphin- 17. Salih, H. R., D. Goehlsdorf, and A. Steinle. 2006. Release of MICB molecules by golipids. Mol. Biol. Cell 14: 3254–3265. tumor cells: mechanism and soluble MICB in sera of cancer patients. Hum. Im- 41. Nabi, I. R., and P. U. Le. 2003. Caveolae/raft-dependent endocytosis. J. Cell Biol. munol. 67: 188–195. 161: 673–677. 18. Lopez-Botet, M., A. Angulo, and M. Guma. 2004. Natural killer cell receptors for 42. van Deurs, B., K. Roepstorff, A. M. Hommelgaard, and K. Sandvig. 2003. Caveo- major histocompatibility complex class I and related molecules in cytomegalo- lae: anchored, multifunctional platforms in the lipid ocean. Trends Cell Biol. 13: virus infection. Tissue Antigens 63: 195–203. 92–100. 19. Eagle, R. A., and J. Trowsdale. 2007. Promiscuity and the single receptor: 43. Subtil, A., I. Gaidarov, K. Kobylarz, M. A. Lampson, J. H. Keen, and NKG2D. Nat. Rev. Immunol. 7: 737–744. T. E. McGraw. 1999. Acute cholesterol depletion inhibits clathrin-coated pit bud- 20. Roda-Navarro, P., M. Vales-Gomez, S. E. Chisholm, and H. T. Reyburn. 2006. ding. Proc. Natl. Acad. Sci. USA 96: 6775–6780. Transfer of NKG2D and MICB at the cytotoxic NK cell immune synapse cor- 44. Rodal, A. A., J. W. Tetreault, P. Lappalainen, D. G. Drubin, and D. C. Amberg. relates with a reduction in NK cell cytotoxic function. Proc. Natl. Acad. Sci. USA 1999. Aip1p interacts with cofilin to disassemble actin filaments. J. Cell Biol. 103: 11258–11263. 145: 1251–1264. 21. Vales-Gomez, M., H. Browne, and H. T. Reyburn. 2003. Expression of the UL16 45. Vela, E. M., L. Zhang, T. M. Colpitts, R. A. Davey, and J. F. Aronson. 2007. glycoprotein of human cytomegalovirus protects the virus-infected cell from at- Arenavirus entry occurs through a cholesterol-dependent, non-caveolar, clathrin- tack by natural killer cells. BMC Immunol. 4: 4. mediated endocytic mechanism. Virology 369: 1–11. 22. Vales-Gomez, M., and H. T. Reyburn. 2006. Intracellular trafficking of the 46. Abrami, L., S. Liu, P. Cosson, S. H. Leppla, and F. G. van der Goot. 2003. HCMV immunoevasin UL16 depends on elements present in both its cytoplasmic Anthrax toxin triggers endocytosis of its receptor via a lipid raft-mediated clath- and transmembrane domains. J. Mol. Biol. 363: 908–917. rin-dependent process. J. Cell Biol. 160: 321–328. 23. Vales-Gomez, M., A. Winterhalter, P. Roda-Navarro, A. Zimmermann, L. Boyle, H. Hengel, A. Brooks, and H. T. Reyburn. 2006. The human cytomegalovirus 47. Horejsi, V. 2003. The roles of membrane microdomains (rafts) in T cell activa- glycoprotein UL16 traffics through the plasma membrane and the nuclear enve- tion. Immunol. Rev. 191: 148–164. lope. Cell Microbiol. 8: 581–590. 48. Borchers, M. T., N. L. Harris, S. C. Wesselkamper, M. Vitucci, and D. Cosman. 24. Tsao, P. I., and M. von Zastrow. 2000. Type-specific sorting of G protein-coupled 2006. NKG2D ligands are expressed on stressed human airway epithelial cells. receptors after endocytosis. J. Biol. Chem. 275: 11130–11140. Am. J. Physiol. 291: L222–L231. 25. Pozo, D., M. Vales-Gomez, N. Mavaddat, S. C. Williamson, S. E. Chisholm, and 49. Raposo, G., M. S. Marks, and D. F. Cutler. 2007. Lysosome-related organelles: H. Reyburn. 2006. CD161 (human NKR-P1A) signaling in NK cells involves the driving post-Golgi compartments into specialisation. Curr. Opin. Cell Biol. 19: activation of acid sphingomyelinase. J. Immunol. 176: 2397–2406. 394–401. 26. Kornfeld, S. 1992. Structure and function of the mannose 6-phosphate/insulinlike 50. Dell’Angelica, E. C., C. Mullins, S. Caplan, and J. S. Bonifacino. 2000. Lyso- growth factor II receptors. Annu. Rev. Biochem. 61: 307–330. some-related organelles. FASEB J. 14: 1265–1278. 27. Kobayashi, T., U. M. Vischer, C. Rosnoblet, C. Lebrand, M. Lindsay, 51. Everett, R. D. 1987. The regulation of transcription of viral and cellular genes by R. G. Parton, E. K. Kruithof, and J. Gruenberg. 2000. The tetraspanin CD63/ herpesvirus immediate-early gene products (Review). Anticancer Res. 7: lamp3 cycles between endocytic and secretory compartments in human endothe- 589–604. lial cells. Mol. Biol. Cell 11: 1829–1843. 52. Chalupny, N. J., A. Rein-Weston, S. Dosch, and D. Cosman. 2006. Down-reg- 28. Karlsson, K., and S. R. Carlsson. 1998. Sorting of lysosomal membrane glyco- ulation of the NKG2D ligand MICA by the human cytomegalovirus glycoprotein proteins lamp-1 and lamp-2 into vesicles distinct from mannose 6-phosphate UL142. Biochem. Biophys. Res. Commun. 346: 175–181. receptor/␥-adaptin vesicles at the trans-Golgi network. J. Biol. Chem. 273: 53. Wu, J., N. J. Chalupny, T. J. Manley, S. R. Riddell, D. Cosman, and T. Spies. 18966–18973. 2003. Intracellular retention of the MHC class I-related chain B ligand of 29. Cefai, D., H. Schneider, O. Matangkasombut, H. Kang, J. Brody, and C. E. Rudd. NKG2D by the human cytomegalovirus UL16 glycoprotein. J. Immunol. 170: 1998. CD28 receptor endocytosis is targeted by that disrupt phospha- 4196–4200. tidylinositol 3-kinase binding and costimulation. J. Immunol. 160: 2223–2230. 54. Stern-Ginossar, N., N. Elefant, A. Zimmermann, D. G. Wolf, N. Saleh, M. Biton, 30. Neefjes, J. J., V. Stollorz, P. J. Peters, H. J. Geuze, and H. L. Ploegh. 1990. The E. Horwitz, Z. Prokocimer, M. Prichard, G. Hahn, et al. 2007. Host immune biosynthetic pathway of MHC class II but not class I molecules intersects the system gene targeting by a viral miRNA. Science 317: 376–381. endocytic route. Cell 61: 171–183. 55. Cuitino, L., R. Matute, C. Retamal, G. Bu, N. C. Inestrosa, and M. P. Marzolo. 31. Gonzalez-Gaitan, M., and H. Stenmark. 2003. Endocytosis and signaling: a re- 2005. ApoER2 is endocytosed by a clathrin-mediated process involving the adap- lationship under development. Cell 115: 513–521. tor protein Dab2 independent of its rafts’ association. Traffic 6: 820–838. 32. Gruenberg, J., and H. Stenmark. 2004. The biogenesis of multivesicular endo- 56. Eleme, K., S. B. Taner, B. Onfelt, L. M. Collinson, F. E. McCann, somes. Nat. Rev. Mol. Cell Biol. 5: 317–323. N. J. Chalupny, D. Cosman, C. Hopkins, A. I. Magee, and D. M. Davis. 2004. 33. Mellman, I. 1996. Endocytosis and molecular sorting. Annu. Rev. Cell Dev. Biol. Cell surface organization of stress-inducible proteins ULBP and MICA that stim- 12: 575–625. ulate human NK cells and T cells via NKG2D. J. Exp. Med. 199: 1005–1010.