of February 2, 2019.This information is current as

Novel MHC Class I Structures on Exosomes

Antoniou and Simon J. PowisM. S. Nimmo, Catherine Botting, Alan Prescott, Antony N. Sarah Lynch, Susana G. Santos, Elaine C. Campbell, Ailish

http://www.jimmunol.org/content/183/3/1884doi: 10.4049/jimmunol.0900798July 2009;

2009; 183:1884-1891; Prepublished online 13J Immunol 

Referenceshttp://www.jimmunol.org/content/183/3/1884.full#ref-list-1

, 24 of which you can access for free at: cites 49 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2009 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on February 2, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on February 2, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

Novel MHC Class I Structures on Exosomes1

Sarah Lynch,* Susana G. Santos,* Elaine C. Campbell,* Ailish M. S. Nimmo,*Catherine Botting,† Alan Prescott,‡ Antony N. Antoniou,§ and Simon J. Powis2*

Exosomes are nanometer-sized vesicles released by a number of cell types including those of the immune system, and often containnumerous immune recognition molecules including MHC molecules. We demonstrate in this study that exosomes can display asignificant proportion of their MHC class I (MHC I) content in the form of disulfide-linked MHC I dimers. These MHC I dimerscan be detected after release from various cell lines, human monocyte-derived dendritic cells, and can also be found in humanplasma. Exosome-associated dimers exhibit novel characteristics which include 1) being composed of folded MHC I, as detectedby conformational-dependent Abs, and 2) dimers forming between two different MHC I alleles. We show that dimer formation ismediated through cysteine residues located in the cytoplasmic tail domains of many MHC I molecules, and is associated with a lowlevel of glutathione in exosomes when compared with whole cell lysates. We propose these exosomal MHC I dimers as novelstructures for recognition by immune receptors. The Journal of Immunology, 2009, 183: 1884–1891.

E xosomes are small vesicles in the size range 50–150 nmthat form by the inward budding of endosomes to gener-ate multivesicular bodies (MVBs)3 (1–4). A proportion of

these MVBs can fuse with the plasma membrane, releasing theirinternal vesicles to the extracellular environment. A wide varietyof cell types has been reported to secrete exosomes, including re-ticulocytes (5), neurons (6), epithelial cells (7), lymphocytes (8),mast cells (9), and dendritic cells (DC) (10).

Of particular interest, exosomes from immune cell types arereplete with many molecules involved in ligand-receptor interac-tions that can lead to the activation or modulation of various im-mune responses. For example, peptide-pulsed exosomes from DCcan induce tumor rejection in mice (10), while exosomes generatedfrom DC incubated with the immunoregulatory cytokine IL-10 candown-regulate disease activity in a collagen-induced murine modelof arthritis (1). This ability of exosomes to act as “cell-free” vac-cines has already been tested in phase I clinical trials against mel-anoma, small-cell lung cancer, and colorectal cancer (11–13).

MHC class I (MHC I) molecules function by binding shortendogenous peptides during their assembly in the endoplasmicreticulum (ER) and presenting them to CD8� T cells, thus ef-ficiently allowing the detection and elimination of virally in-

fected or potentially tumorigenic cells (14). MHC I moleculesare composed of an MHC-encoded H chain of �43 kDa, a 12kDa non-MHC encoded �2-microglobulin L chain (�2m), andan 8 –9 residue peptide. The recognition of MHC I molecules bythe TCR system involves inclusion of the MHC I moleculewithin immune synapses, with clustering increasing the sensi-tivity of detection (15, 16).

In addition to being recognized by CD8� T cells, which ishighly peptide- and MHC I-allele specific, it is now known thatmany MHC I and MHC I-like molecules are also recognized byreceptors often found expressed on cells of the NK lineage, in-cluding killer cell Ig-like receptors (KIRs) and leukocyte Ig-likereceptors (LIRs) (17).

While studying the ankylosing spondylitis (AS)-associated HLA-B27 class I allele in an EBV-transformed B cell line, we discoveredthat exosomes contained enhanced amounts of MHC I dimer struc-tures. HLA-B27 H chain homodimers have been implicated in thepathogenesis of AS (18–20). However, rather than the partially mis-folded HLA-B27 H chain homodimers implicated in AS, we foundthe exosome-associated HLA-B27 dimers to be fully folded. We dem-onstrate that cysteine residues located in the cytoplasmic tail domainparticipate in the formation of these structures. Taken together, thesedata demonstrate the existence of dimeric MHC I structures on exo-somes that may act as a novel ligand for immune receptors.

Materials and MethodsCell lines and Abs

The human EBV-transformed B cell line Jesthom was obtained from AdamBenham (University of Durham, Durham, U.K.) and Health ProtectionAgency Culture Collections (88052004), and cultured in RPMI 1640 (Invitro-gen) with 20% FBS (Invitrogen). KG-1 cells were obtained from Health Pro-tection Agency Culture Collections (86111306) and cultured in IMDM (In-vitrogen) with 20% FBS. KG-1 cells were matured with 10 ng/ml PMA and100 ng/ml ionomycin where required. Rat C58 thymoma cells were cultured inRPMI 1640 with 5% FBS. HLA class I negative .221 cells were a gift fromSalim Khakoo (Imperial College, London, U.K.), and were cultured in RPMI1640 with 5% FBS. Human monocyte-derived DCs were generated in DMEMplus 10% FBS from a 20 ml blood donation (obtained after ethical review bythe Bute Medical School ethics committee), expanding the 2 h plastic-adherentpopulation in GM-CSF and IL-4 for 4–5 days, followed by stimulation with100 ng/ml LPS (Sigma-Aldrich) for the indicated times. Abs used were: HC10,recognizing unfolded HLA-B and -C (21); HCA2, recognizing unfoldedHLA-A (22); ME1, recognizing folded HLA-B27 (23); BB7.2, recognizingfolded HLA-A2 (24); W6/32, recognizing folded HLA-A, -B, and -C (25); V5,

*Bute Medical School, University of St. Andrews, Fife, United Kingdom; †Biomo-lecular Sciences, University of St. Andrews, Fife, United Kingdom; ‡Division of CellBiology and Immunology, College of Life Sciences, University of Dundee, Dundee,United Kingdom; and §Division of Infection and Immunity, University College Lon-don, London, United Kingdom

Received for publication March 12, 2009. Accepted for publication May 30, 2009.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 S.L. was funded by the University of St. Andrews Maitland-Ramsey studentship.A.M.S.N. was supported by bursaries from the Association of Physicians of GreatBritain and Ireland and the Wolfson Trust. A.N.A. was supported by a U.K. ArthritisResearch Campaign Fellowship.2 Address correspondence and reprint requests to Dr. Simon Powis, Universityof St. Andrews, Westburn Lane, St. Andrews, U.K. E-mail address: sjp10@st-andrews.ac.uk3 Abbreviations used in this paper: MVB, multivesicular body; DC, dendritic cell;MHC I, MHC class I; ER, endoplasmic reticulum; �2m, �2-microglobulin; KIR, killercell Ig-like receptor; LIR, leukocyte Ig-like receptor; AS, ankylosing spondylitis; TfR,transferrin receptor; GSH, glutathione; NEM, N-ethylmaleimide; pI, isoelectric point.

Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00

The Journal of Immunology

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0900798

by guest on February 2, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

recognizing the V5 epitope tag (26); anti-transferrin receptor (TfR), recogniz-ing transferrin receptor; anti-ERp57 (gifted by N. Bulleid, University ofManchester, Manchester U.K.); anti-CD48; anti-transporter associated withAg processing (TAP1) (gifted by R. Tampe, Frankfurt, Germany); anti-ezrinand moesin (gifted by F. Gunn-Moore, University of St. Andrews, St An-drews, U.K.); anti-HLA-B27-FITC (One Lambda); anti-Tsg101 (Santa CruzBiotechnology); anti-ALIX (Santa Cruz Biotechnology).

Isolation of exosomes

Exosomes were isolated based on published procedures (10, 27). In brief,cells were incubated overnight in either serum-free medium or mediumcontaining 5% FBS or 2.5% exosome-free FBS (pre-spun at 100,000 � g).The exosomal MHC I dimer signal, determined by immunoblotting withHC10, was identical in all three cultures (data not shown). Culture super-natants were then depleted of cells (1000 � g, 10 min), alkylated with 10mM N-ethylmaleimide (NEM), and either filtered at 0.2 �m followed by a100,000 � g spin for 2 h, or spun at 10,000 � g for 10 min followed by100,000 � g for 2 h. Exosomes destined for incubation with live cells werenot subjected to NEM treatment. In all other cases, exosome pellets wereresuspended in a small volume of PBS with 10 mM NEM, and quantifiedby spectrophotometry or Bradford assay. Typical yields for Jesthom cellswere around 1 �g of exosomes per one million cells cultured overnight.Lower yields were obtained for all other lines used in this study.

Microscopy and mass spectrometry

For electron microscopy, exosomes were placed on a carbon and pioloform-coated grid, blotted dry, and stained with 3% uranyl acetate in water. Alter-natively exosomes were diluted in water and collected on a carbon film onmica and again stained with uranyl acetate; the film was collected on an un-coated grid. Imaging was performed on an FEI Tecnai 12 TEM (FEI). Forimmunofluorescence microscopy, cells were fixed in 2% formaldehyde in PBSfor 20 min, then blocked and permeabilized with 1% BSA, 0.2% saponin instaining buffer. Cells were stained with FITC-conjugated anti-B27. After ex-tensive washing, cells were resuspended in PBS and mounted on cover slideswith Vectashield containing Dapi (Vector Laboratories). Deconvolution mi-croscopy was done using a wide-field optical sectioning microscope, Delta-Vision Restoration Imaging System (Applied Precision). For each cell, a z-series of 15 to 45 images at 0.35 �m intervals was captured and processedusing constrained interactive deconvolution via SoftWoRx 3.0 software (Ap-plied Precision). Further image analysis was performed using NIH Image Jsoftware.

Mass spectrometric identification of proteins was performed from anSDS-PAGE gel stained with SimplyBlue SafeStain (Invitrogen). The re-gion of interest was excised, cut into 1-mm cubes, and subjected to in-geldigestion using a ProGest Investigator robot (Genomic Solutions), usingstandard protocols (28). In brief, gel cubes were destained in acetonitrile,reduced, alkylated, and digested with trypsin at 37°C. Peptides were ex-tracted with 10% formic acid, concentrated down, and separated using anUltiMate nanoLC (LC Packings) equipped with a PepMap C18 trap col-umn. The eluent was sprayed into a Q-Star Pulsar XL tandem mass spec-trometer (Applied Biosystems) and analyzed in Information DependentAcquisition mode. The tandem mass spectrometry data file generated wasanalyzed using the Mascot 2.1 search engine (Matrix Science) againstMSDB database. The data was searched with tolerances of 0.2 Da for theprecursor and fragment ions, trypsin as the cleavage enzyme, one missedcleavage, carbamidomethyl modification of cysteines as a fixed modifica-tion and methionine oxidation selected as a variable modification. TheMascot search results were accepted if a protein hit included at least onepeptide with a score above the homology threshold and the MS/MS inter-pretation accounted for the major peaks in the spectrum.

Immunoblotting and immunoprecipitation

Cell lysates were prepared by resuspending cells in lysis buffer (1% NP40,150 mM NaCl, 10 mM Tris (pH 7.3), 1 mM PMSF, 10 mM NEM) for 10min on ice, then spinning at 20,000 � g for 10 min. Lysates and exosomeswere resuspended in nonreducing or reducing sample buffer as required,heated at 90°C for 3 min, run on 8% SDS-PAGE gels, and transferred tonitrocellulose membranes (BA85, Whatman). Membranes were incubatedwith the indicated Abs and relevant HRP-coupled second stage reagentsand visualized with Femto Western (Perbio) chemiluminescence reagentsand a Fuji LAS3000 image analyzer. Immunoprecipitations were per-formed using cell lysates or exosome lysates (buffer as above), using theindicated Abs and protein G-Sepharose beads (Sigma-Aldrich). Beads werewashed in lysis buffer and resuspended in the relevant sample buffer beforeimmunoblotting as above. Two-dimensional gel analysis was performed aspreviously described (29), using nonreducing sample buffer, followed byimmunoblotting as above.

Flow cytometry and glutathione (GSH) assay

Flow cytometry was performed on cells stained with the indicated Abs andFITC-coupled second stage reagents, followed by analysis on a FACScan(Becton Dickinson).

Flow cytometry of exosomes was performed as follows: 30 �g of exo-somes in 100 �l PBS were incubated with 20 �l of latex beads (4 �maldehyde/sulfate, Invitrogen) that had been sonicated for 5 min. The bead/exosome mixture was incubated at room temperature for 15 min, resus-pended with 1 ml PBS, and incubated with mixing at 4°C for 1 h. Beadswere pelleted at 20,000 � g for 2 min, then resuspended in 100 mMglycine for 30 min. Beads were washed twice with PBS, resuspended in 0.5ml PBS, and 50 �l stained with the indicated Abs using standard protocols.Glutathione was measured using the Glutathione Assay Kit (Sigma-Aldrich; CS0260). Samples, quantified by the Bradford method, weredeproteinated and total GSH measured in a reaction where GSH causes thereduction of 5,5H-dithiobis(2-nitrobenzoic) acid to the yellow product5-thio-2-nitrobenzoic acid, which was measured at 380 nm using a FLU-Ostar Optima plate reader (BMG Labtech). Standard curves were preparedand the amount of GSH in cell lysates and exosome lysates determined.Samples were measured in duplicate.

Mutagenesis

Site-directed mutagenesis was performed using Stratagene Quickchangemethodology, using appropiate primers to mutate Cys67 to Ser67, Cys308 toAla308, and Cys325 to Ala325 in HLA-B*2705 cDNA, or to introduce a stopcodon in the cytoplasmic tail domain of HLA-A*0201 cDNA. Mutatgen-esis was confirmed by sequencing. Mutants, cloned into the pCR3.1 vector(Invitrogen) were transfected into KG-1, .221, or C58 cells by electropo-ration and selected with 1 mg/ml G418 (Melford). Expression was con-firmed by flow cytometry and immunoblotting.

ResultsExosome characterization from Jesthom and KG-1 cells

Exosomes were isolated by either differential centrifugation or byfiltration and then centrifugation (with essentially identical re-sults), from the HLA-A2- and -B27-expressing EBV-transformedB cell line Jesthom (30), and the HLA-A30-, -A31-, -B35-express-ing DC-like cell line KG-1 (31). Both exosome preparations, whenimaged by negative stain electron microscopy, displayed the char-acteristic exosomal cup-shaped morphology �100 nm in size (Fig.1A). Mass spectrometric identification of a number of protein spe-cies extracted from a nonreducing SDS-PAGE gel of Jesthom de-rived exosomes was also performed, and identified previously re-ported exosomal species (32) including myosin, TfR, annexins 4and 6, ezrin, moesin, MHC I and II, and CD20 (Fig. 1B). In ad-dition, we identified an MHC I signal in the size region of �80kDa, consistent with an MHC I H chain dimer. Furthermore, wealso identified the presence of CD48, previously unreported in exo-somes, which is a ligand for the NK receptor 2B4 (CD244) (33).

To further characterize Jesthom and KG-1-derived exosomes,we performed immunoblotting of reduced detergent cell lysatescompared with exosomes, and flow cytometry of exosomes ab-sorbed to latex beads. The exosome samples did not contain ex-amples of ER-resident species such as TAP1, or the oxidoreduc-tase ERp57, but did contain ezrin, moesin, MHC I, and also ALIXand tsg101, two species highly characteristic of exosomes (Fig.1C). Flow cytometry confirmed the presence of HLA-A and -Bmolecules on the extracellular side of Jesthom-derived exosomesusing mAbs BB7.2 and ME1, respectively. TfR and CD48 werealso observed (Fig. 1D). A representative negative control Ab (anti-V5 tag) stained the exosomes only very weakly.

Enhanced MHC I dimer structures on exosomes

As indicated above, the mass spectrometric data indicated the pres-ence of MHC I molecules in a size range consistent with the re-ported HLA-B27 H chain homodimers implicated in the pathogen-esis of ankylosing spondylitis. To further characterize these

1885The Journal of Immunology

by guest on February 2, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

putative HLA-B27 dimers on Jesthom-derived exosomes, we im-munoblotted nonreduced detergent cell lysates and exosomes withthe HLA-B recognizing Ab HC10. Unexpectedly, we observedthat the Jesthom-derived exosomes contained a much greater pro-portion of dimer molecules than the cell lysate, with �50% of theH chain migrating around 80 kDa (Fig. 2A). To confirm this spe-cies as an HLA-B27 H chain dimer we examined the sample underreducing conditions, wherein it migrated at the correct monomericH chain size of �43 kDa (Fig. 2B), indicating it to be a disulfide-linked species. It is also worthy of note that the nonreduced andreduced monomer heavy chains migrate differently in SDS-PAGE,due to the preservation of intrachain disulfide bonds in nonreduced

samples, which also affects the apparent migration of the dimerspecies, again due to preservation of intrachain disulfide bonds. Tofurther determine the composition of these dimeric structures, weexamined the sample using nonreducing two-dimensional electro-phoresis. H chain dimers should retain an identical isoelectric point(pI) to monomeric H chain, whereas an H chain complexed toanother protein would usually alter its overall pI. As shown in Fig.2C (top), the HLA-B27 signal resolved as monomers and dimerswith identical pI, thus indicating them to be authentic HLA-B27 Hchain dimers. Additionally, however, we also identified a dimer-sized orphan spot with no corresponding monomer spot (arrowed),which could represent HLA-B27 disulfide linked to another pro-tein. Previous use of this two-dimensional system indicated to usthat the other HLA class I allele expressed in Jesthom (HLA-A2)migrated to the left of HLA-B27 (29). We therefore hypothesizedthat this additional spot may be a disulfide-linked complex ofHLA-B27 and HLA-A2 heavy chains. Immunoblotting with theHLA-A recognizing Ab HCA2 (Fig. 2C, middle), which recog-nizes HLA-A2, also revealed the orphan spot (arrowed) andHLA-A2 dimers. Overlaying separately stained blots of HC10 andHCA2 permitted exact alignment of this spot, and costaining ofblots with both Abs also highlighted this same spot (Fig. 2C, bot-tom). Taken together, the above data strongly suggests that theJesthom exosomes contain HLA-B27 H chain monomers anddimers, HLA-A2 H chain monomers and dimers, and also a novelpopulation of dimers comprising HLA-A2 and HLA-B27.

Exosomal MHC I dimers are independent of residue cysteine 67and are fully folded

HLA-B27 H chain dimers are disulfide linked through a normallyunpaired cysteine residue at position 67 in the peptide groove (18).We therefore tested exosomes from the rat T cell line C58 and

FIGURE 1. Characterization of exosomes fromJesthom and KG-1 cells. A, Electron microscopy of exo-somes isolated from culture supernatants of Jesthom(top) and KG-1 (bottom). Size bars 200 nm. B, Massspectrometric identification of proteins from Jesthomderived exosomes after nonreducing SDS-PAGE. C,Immunoblotting of cell lysates and exosomes fromJesthom and KG-1 cells (nd, not determined). D, Flowcytometric analysis of Jesthom derived exosomes ab-sorbed to latex beads and stained with the indicatedAbs. Gray traces are unstained beads. 2° represents sec-ond stage FITC anti-mouse IgG alone.

FIGURE 2. Detection of MHC I dimers on exosomes. A, Three micro-grams of Jesthom cell and exosome lysate was immunoblotted with HC10,revealing enhanced dimer-like structures on exosomes. B, Nonreduced andreduced samples of Jesthom derived exosomes were immunoblotted withHC10. C, Nonreduced Jesthom-derived exosome samples were analyzedby two-dimensional electrophoresis, followed by immunoblotting withHC10 (top), HCA2 (middle), or HC10 and HCA2 mixed together (bottom).Arrows indicate the putative mixed heterodimer of one HLA-A and oneHLA-B H chain.

1886 EXOSOME MHC CLASS I

by guest on February 2, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

human KG-1 DC-like cells which had been transfected with cDNAconstructs for wild-type HLA-B*2705, and a Cys67-Ser67 mutant(B27.C67S). In both instances, exosomes retained the HLA-B27 H

chain dimer signal (Fig. 3A). The previously proposed HLA-B27H chain dimer structure is thought to involve partial unfolding ofthe �1 helix, rendering it less well recognized by conformationdependent Abs such as W6/32 and ME1, but more recognisable byHC10 (34). Immunoprecipitation and immunoblotting of Jesthomexosomes lysed in detergent (Fig. 3B) revealed low HC10 signal,but strong signals for both monomers and dimers recognized bythe conformation dependent Abs W6/32 (HLA-A, -B, and -C),ME1 (HLA-B, including -B27), and BB7.2 (HLA-A, includingHLA-A2). Dimer structures (and monomers) could also be isolatedwith an anti-�2m specific Ab (data not shown), further confirmingthe folded nature of the H chain within these structures. We alsoused immunoprecipitation with the conformation-dependent AbBB7.2 to confirm the existence of the mixed A2-B27 dimer shownin Fig. 2C. Immunoprecipitation of HLA-A2 from a lysate ofJesthom exosomes was followed by immunoblotting for HLA-B27with HC10, which revealed a dimer band indicative of a mixedA2-B27 conjugate (Fig. 3D).

The HLA-A2 molecule does not contain the unpaired Cys67 res-idue of HLA-B27, therefore the above data strongly indicate thatthe MHC I dimers isolated from exosomes are not the same pop-ulation which form through the unpaired Cys67 residue present inHLA-B27.

Involvement of cytoplasmic domain cysteines and glutathione inexosomal MHC I dimers

If exosomal MHC I dimers are not Cys67 dependent, and are also fullyfolded, we hypothesized that a disulfide linkage may be forming be-tween unpaired cysteine residues located in the cytoplasmic tail ofmany MHC I molecules. All HLA-B alleles possess Cys308, which

FIGURE 3. Exosomal MHC class I dimers are folded MHC class I mol-ecules. A, Exosomes from rat C58 (left) and human KG-1 cells (right),expressing wild-type HLA-B27 or a mutated Cys67 to Ser67 version (C67S)were immunoblotted with HC10. B, Immunoprecipitation of MHC I mol-ecules from a lysate of Jesthom derived exosomes. Samples were immu-noblotted with either HC10 or HCA2 as indicated. C, Jesthom exosomeswere immunoprecipitated for HLA-A2 (BB7.2) or with an irrelevant con-trol mAb (V5), and immunoblotted for HLA-B27 with HC10. A sample ofthe input exosome lysate used in the immunoprecipitation is shown (right).

FIGURE 4. A role for cytoplasmictail domain cysteines and GSH inthe formation of exosomal MHC Idimers. A, Sequence of the cytoplas-mic tail domain regions of HLA-B27and HLA-A2, with cysteine residuesunderlined. B, Immunoblotting ofexosomes isolated from the superna-tants of Jesthom cells culture with in-creasing amounts of GSH. C, Similarto B, exosomes were isolated fromcells cultured in 50 �M 2-ME. D, Im-munoblotting of exosomes isolatedfrom rat C58 cells expressing wild-type HLA-A or a cytoplasmic tailtruncation mutant. E, Exosomes iso-lated from .221 cells expressingHLA-B27 with cysteine to alaninemutations at positions 308 (C308A)or 325 (C325A) were immunoblottedwith HC10. F, Total GSH contentwas measured from 7 �g of Jesthomcell and exosome lysate. Results arerepresentative of two independentexperiments.

1887The Journal of Immunology

by guest on February 2, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

lies on the border of the transmembrane-spanning domain and cyto-plasmic domain (Fig. 4A). In addition, some HLA-B alleles possessCys325 in the cytoplasmic domain. No HLA-A allele contains Cys308,but all have Cys339 close to the end of the cytoplasmic domain.

The main cellular thiol responsible for maintaining the reducingenvironment of the cytoplasm is GSH, the intracellular concentra-tion of which can be as high as 10 mM (35). To investigate the roleof GSH and the cytoplasmic tail cysteine residues in exosomalMHC I dimer formation, we first raised the external concentrationof GSH in the medium up to 10 mM and examined the MHC Idimer signal of exosomes isolated from the culture supernatant. Inthe presence of 10 mM GSH, MHC I dimers were present in sig-nificantly lower amounts (Fig. 4B). Many T cell assays include2-ME in the cell culture medium, however its presence at 50 �Mdid not prevent MHC class I dimer formation (Fig. 4C). We nexttested a mutant HLA-A2 molecule in which the tail domain wastruncated by 25 residues, thus removing Cys339. Exosomes iso-lated from rat C58 cells transfected with this “tail-less” HLA-A2mutant failed to form dimers (Fig. 4D). We also mutated the in-dividual cysteines residues located at positions 308 and 325 inHLA-B27 and examined exosomes from HLA class I negative.221 cells expressing these mutants (Fig. 4E). Wild-type HLA-B27and mutant C308A maintained dimer structures, whereas mutantC325A essentially failed to form dimers, directly implicating thecysteine residues in the cytoplasmic tail domains of HLA-A2 andHLA-B27 in dimer formation. Longer exposure of this image re-vealed a faint dimer signal in the C325A mutant, suggesting thatCys308 might contribute to a minor pool of dimer structures.

We also performed a direct measurement of the total GSH con-tent present in Jesthom cell lysate compared with exosomes, whichrevealed low levels of GSH in the exosomes, equivalent to �1.7nmol/�g lysate in whole cells compared with 0.4 nmol/�g lysatein exosomes (Fig. 4F).

In view of the low GSH level detected in exosomes, we next askedwhether extreme depletion of intracellular GSH in whole cells wouldmimic the situation in exosomes and promote MHC dimer formation.Intracellular GSH was rapidly removed from Jesthom cells using thethiol-oxidant diamide (Fig. 5). MHC I dimers were strongly inducedwith 0.25 mM diamide and above. Importantly, these MHC I dimerswere immunoprecipitated with W6/32, indicating them to be fullyfolded MHC I molecules. These data support a model wherein exo-somes do not contain enough reducing potential, in the form of GSH,to prevent disulfide bonds forming between available cysteine resi-dues in the cytoplasmic tails of MHC I molecules.

HLA haplotype influences the ability to form exosomal MHC Idimers

The data presented in Fig. 3A posed a question to our proposedmodel: why were there no dimers in exosomes from control, non-B27 transfected KG-1 cells? The IMGT/HLA database (www.ebi.ac.uk) confirmed that no HLA-B35 allele (as expressed by KG-1cells) contains Cys325, and would therefore not form HC10-reac-tive dimers according to our model. This also further indicates thatCys308 does not contribute strongly to exosomal MHC class Idimer formation (Fig. 4E). Conversely, the HLA-A31 and -A35alleles (also expressed by KG-1) do contain Cys339, and wouldtherefore be predicted to form exosomal MHC I dimers, detectableby HCA2, which proved to be the case (Fig. 6A). We further testedthis model using the TAP-deficient cell line T2, which expressesonly HLA-A*0201 (with Cys339) and HLA-B*5101 (withoutCys325), and its rat TAP-restored derivative (36), denoted T3. T2has very low levels of HLA-B51, but significant levels of HLA-A2(Fig. 6B), because of the ability of the latter to bind leader peptidesin the ER (37). TAP1 and TAP2 restoration results in increasedexpression of both alleles in the T3 line. Comparisons of cell ly-sates and exosomes indicated that no HC10 dimers were detectedin either T2 or T3 exosomes, but HCA2 reactive dimers were seenweakly in T2 and more strongly in T3 cells (Fig. 6C). Thus, dimerformation follows both possession of cytoplasmic domain cysteineresidues, and is enhanced by higher levels of MHC I expression.

We next asked whether human monocyte-derived DC secreteexosomes displaying MHC I dimers. We compared a non-HLA-B27 individual with an HLA-B27-expressing individual. DC werestimulated for up to 72 h with LPS and exosomes isolated andimmunoblotted with HC10 (Fig. 6D). Exosomes from HLA-B27-expressing cells (DC exo 2) contained more MHC I dimers anddisplayed them at an earlier time after LPS activation comparedwith non-HLA-B27 exosomes. Exosomal MHC I dimers couldalso be detected in the plasma of a healthy non-HLA-B27 indi-vidual (Fig. 6E). Thus, exosomal MHC I dimers occur in vivo aswell as in vitro.

Finally, we asked whether exosomal MHC I dimers could betransferred from one cell to another. Jesthom-derived exosomes,which are strongly HC10-dimer positive (Fig. 6F, input) were in-cubated for 24 h with KG-1 cells, which have no HC10 reactivedimers (Figs. 3A and 6A). HC10 immunoblotting revealed thatdimers were detected in KG-1 cells incubated with exosomes (Fig.6F), and fluorescence microscopy with anti-HLA-B27-FITC re-vealed distinct regions where the HLA-B27 positive exosomes ac-cumulated in a representative KG-1 cell (Fig. 6G). Thus, exosomescan transfer novel MHC class I structures to cells which are notthemselves expressing such molecules.

DiscussionWe have demonstrated in this study that MHC I molecules canexist in exosomes in a dimeric form, which the evidence suggestsconsists of two fully folded MHC I molecules attached through adisulfide bond between cysteine residues located in the cytoplas-mic domain. We propose that these structures may confer novelimmunomodulatory roles for MHC I molecules on exosomes.

The detection of MHC I molecules in dimeric form has beenreported on the surface of cells (38). More recently, HLA-B27dimers have been studied in detail because of their proposed rolein the pathogenesis of AS (18, 34). A partially folded, potentiallypeptide loaded, �2m-free H chain dimer has been described whichcan be recognized by several KIR and LILR receptors, includingKIR3DL2 (39), and in the HLA-B27 transgenic rat model, folded

FIGURE 5. Induction of folded MHC I dimers by treatment with dia-mide. Jesthom cells were incubated with the indicated concentrations ofdiamide for 10 min. Cell lysates and W6/32 immunoprecipitations wereimmunoblotted with HC10. A nonspecific band in the 80-kDa region of theW6/32 immunoprecipitate is indicated with an asterisk. The MHC I dimerband migrates immediately beneath this band.

1888 EXOSOME MHC CLASS I

by guest on February 2, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

W6/32 reactive dimers were also seen at low levels (20). Our cur-rent study has several unique observations, namely 1) the previ-ously unreported detection of MHC I dimers in exosomes, 2) theprevalence of dimers, wherein they are highly enhanced in exo-somes compared with cell lysates, 3) their fully folded nature inpreference to unfolded/partially misfolded structures, and 4) theirformation by both HLA-A and HLA-B alleles, with the addedability, in at least one instance to form mixed heterodimers be-tween an HLA-A and an HLA-B molecule.

Taking each of the above points in turn, previous studies do notspecify the use of nonreduced vs reduced samples in SDS-PAGE.We presume, therefore, that the majority of studies on exosomes todate where MHC I has been detected have involved SDS-PAGE ofreduced samples, thus precluding their detection as disulfide-linked structures, whereas in our studies of HLA-B27 we analyzenonreduced samples to preserve such dimers. We were surprisedby the dramatic levels of dimers we detected in exosomes com-pared with cell lysates (Fig. 2A). Because our initial observationswere made in an HLA-B27-expressing cell line, we anticipatedthat we had identified on exosomes a rich source of the Cys67-Cys67 linked HLA-B27 H chain homodimers (named B272) thathave been characterized by Bowness and colleagues (34). How-ever, our subsequent experiments indicated we have identified adistinct set of MHC I dimers because the exosomal dimers appearfully folded, and can also involve HLA-A molecules.

The folded nature of the exosomal MHC I dimers raises a sig-nificant issue. The endosomal route to MVBs, where exosomes are

derived from, may alter the repertoire of peptides bound to MHCI molecules, either by removing those bound relatively weakly, orby loading peptides from exogenous sources, thus potentially con-tributing to cross-presentation. A detailed analysis of peptideseluted from exosomal MHC I molecules is required to address thisissue, although the current lack of an Ab to isolate MHC I dimersfrom monomers currently precludes any discrimination of the pep-tides from these two structures.

It is also pertinent to ask where these MHC I dimer structuresare forming, and the related question as to the role played by GSH.Our data suggest that the low level of GSH in exosomes allowsdisulfide bonds to form between the cytoplasmic tail residues ofMHC I molecules, a situation not found in the strong reducingenvironment maintained in the cell cytoplasm. We suggest that thisnormally inhibits the formation of large populations of MHC Idimers of this type in normal cell membranes, but dimers can beinduced experimentally by the removal of GSH with diamide (Fig.5). The GSH assay system we have used here measures total GSHcontent, by also reducing any GSSG present, thus it is unlikely thatexosomes lack reducing capacity due to GSH being sequestered asGSSG. We are currently attempting to determine whether GSH islost at the stage when MVBs are formed, perhaps because not allof the necessary enzymes required are included when exosomesform, although they do contain at least one key enzyme, GST (40).Alternatively, GSH may be lost after release of the exosomes,when they may equilibrate their level of GSH to the low �M levelfound in serum and in culture medium (41).

FIGURE 6. MHC haplotype influ-ences dimer formation and dimertransfer between cells. A, NonreducedKG-1 cell and exosome lysates wereimmunoblotted with HC10 and HCA2.B, TAP-deficient T2 cells and ratTAP1 plus 2 restored T3 cells werestained with Abs ME1 (recognizingHLA-B51) and BB7.2 (recognizingHLA-A2). 2° represents second stageFITC anti-mouse IgG alone. C, T2and T3 cell and exosome lysateswere immunoblotted with HC10 andHCA2. D, Exosomes from monocyte-derived DC from a non-B27 express-ing (DCexo1) and a B27 expressing(DCexo2) individual were immuno-blotted with HC10. E, Immunoblotwith HC10 of exosomes isolated fromplasma from a non-HLA-B27 indi-vidual. The asterisk indicates a non-specific band, which varies betweendifferent serum samples, which wehave not yet identified. F and G,KG-1 cells were incubated overnightwith or without Jesthom derived exo-somes and then lysed and immuno-blotted with HC10 or stained withFITC anti-B27 (green) and imaged bymicroscopy. The nucleus was stainedwith DAPI (blue).

1889The Journal of Immunology

by guest on February 2, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

The biological role of the unpaired cysteine residues in the cy-toplasmic domains of MHC I molecules remains poorly defined. Inthis study, we propose they have a key role in the formation ofexosomal MHC I dimers. Previous studies have identified thatthese cysteines can be modified by palmitylation (42, 43). Of greatpotential interest, in the latter study by Gruda and coworkers (44),mutation of the cytoplasmic tail domain cysteines in HLA-B7 pre-vented recognition of HLA-B7 expressing cells by a solubleLIR1-Ig fusion protein. It could be speculated that preventing thedimerization of a small population of cell surface MHC I by mu-tation of the cytoplasmic domain cysteines results in a loss ofLIR1-Ig binding. Our current data does not suggest that Cys308 hasa major role in exosomal dimer formation, which does not entirelyfit with the data from Gruda and coworkers. Cys308 has, however,been identified as forming a transient disulfide linkage with tapasinduring MHC I assembly in the endoplasmic reticulum, so remainsaccessible during at least some part of the lifespan of an MHC Imolecule.

In this current study, we have not addressed what biological roleexosomal MHC I dimers may have, although we can speculate ontwo broad possibilities. Firstly, clustering of MHC I molecules isknown to enhance their recognition by T cells, and indeed chem-ically inducing the formation of MHC I dimers enhances theirability to present peptides at low concentrations (15). The reportedsupine orientation of MHC I molecules in the membrane couldalso facilitate dimer formation by allowing close proximity of thecytoplasmic domains (45). Secondly, it is also possible that exo-somal MHC I dimers may exert their influence by interacting withreceptors of the NK cell lineage. Indeed, we report in this study thepresence of the NK ligand CD48 on exosomes, indicating that theyhave the capacity to be recognized by NK cells. In this context wemay also be able to compare the exosomal MHC I dimers wedescribe in this study to those formed by the nonclassical HLA-Gclass I molecule, which is expressed almost exclusively on tro-phoblast cells during pregnancy, and which may function to gen-erate dominant immunosuppressive effects to protect a developingfetus. HLA-G has a shortened cytoplasmic domain, with no cys-teine residues, and thus cannot form dimers of the type we describein this study. However, HLA-G is unique among MHC I mole-cules, as it possesses an unpaired Cys42 residue, situated on anexternal loop of the �1 domain, which allows it to form dimers(46). Such HLA-G dimers have been shown to be up to 100 timesmore efficient at LILRB1 signaling (47). Thus, it is possible thatthe exosomal MHC class I dimers we describe in this study may bepotent ligands for recognition by NK lineage receptors biased to-ward HLA-A and -B molecules. Of great potential significance,recent observations in a phase I clinical trial of dendritic cell-derived exosomes has shown the significant restoration and acti-vation of NK cells in a cohort of patients with melanoma (48), thushighlighting the ability of exosomes to interact with NK cells.

We have also described in this study that a dimorphism exists inthe ability of exosomes to form MHC class I dimers. Although allHLA-A alleles contain Cys339, only a proportion of HLA-B allelescontain Cys325. For example, alleles with Cys325 include HLA-B07, �08, �27, �41, �42, and �44, whereas those without in-clude -B14, -B15, -B35, -B37, -B40, and -B51. Thus, each indi-vidual’s HLA haplotype will determine the presence of MHC classI exosomal dimers. In theory dimers may also form between dif-ferent alleles, for example a heterozygous individual expressingHLA-B07 and -B44 could form exosomal B07-B07 dimers, B44-B44 dimers, and B07-B44 dimers. Furthermore, as we have shownin this study with exosomes from the Jesthom cell line, it is alsopossible to form dimers between HLA-A and -B molecules (Fig.2C), though at lower levels, perhaps due to steric constraints

caused by the relative positions of the cysteines in the cytoplasmictails (49). Thus, our data describes for the first time a population ofdimeric MHC I molecules on exosomes, which may represent anew range of structures for recognition by the various MHC Ioriented receptors expressed on cells of the immune system.

DisclosuresThe authors have no financial conflict of interest.

References1. Kim, S. H., E. R. Lechman, N. Bianco, R. Menon, A. Keravala, J. Nash, Z. Mi,

S. C. Watkins, A. Gambotto, and P. D. Robbins. 2005. Exosomes derived fromIL-10-treated dendritic cells can suppress inflammation and collagen-induced ar-thritis. J. Immunol. 174: 6440–6448.

2. Lehmann, B. D., M. S. Paine, A. M. Brooks, J. A. McCubrey, R. H. Renegar,R. Wang, and D. M. Terrian. 2008. Senescence-associated exosome release fromhuman prostate cancer cells. Cancer Res. 68: 7864–7871.

3. Raposo, G., H. W. Nijman, W. Stoorvogel, R. Liejendekker, C. V. Harding,C. J. Melief, and H. J. Geuze. 1996. B lymphocytes secrete antigen-presentingvesicles. J. Exp. Med. 183: 1161–1172.

4. Wang, G. J., Y. Liu, A. Qin, S. V. Shah, Z. B. Deng, X. Xiang, Z. Cheng, C. Liu,J. Wang, L. Zhang, W. E. Grizzle, and H. G. Zhang. 2008. Thymus exosomes-like particles induce regulatory T cells. J. Immunol. 181: 5242–5248.

5. Johnstone, R. M., M. Adam, J. R. Hammond, L. Orr, and C. Turbide. 1987.Vesicle formation during reticulocyte maturation: association of plasma mem-brane activities with released vesicles (exosomes). J. Biol. Chem. 262:9412–9420.

6. Faure, J., G. Lachenal, M. Court, J. Hirrlinger, C. Chatellard-Causse, B. Blot,J. Grange, G. Schoehn, Y. Goldberg, V. Boyer, et al. 2006. Exosomes are re-leased by cultured cortical neurones. Mol. Cell Neurosci. 31: 642–648.

7. van Niel, G., G. Raposo, C. Candalh, M. Boussac, R. Hershberg,N. Cerf-Bensussan, and M. Heyman. 2001. Intestinal epithelial cells secrete exo-some-like vesicles. Gastroenterology 121: 337–349.

8. Blanchard, N., D. Lankar, F. Faure, A. Regnault, C. Dumont, G. Raposo, andC. Hivroz. 2002. TCR activation of human T cells induces the production ofexosomes bearing the TCR/CD3/� complex. J. Immunol. 168: 3235–3241.

9. Raposo, G., D. Tenza, S. Mecheri, R. Peronet, C. Bonnerot, and C. Desaymard.1997. Accumulation of major histocompatibility complex class II molecules inmast cell secretory granules and their release upon degranulation. Mol. Biol Cell.8: 2631–2645.

10. Zitvogel, L., A. Regnault, A. Lozier, J. Wolfers, C. Flament, D. Tenza,P. Ricciardi-Castagnoli, G. Raposo, and S. Amigorena. 1998. Eradication of es-tablished murine tumors using a novel cell-free vaccine: dendritic cell-derivedexosomes. Nat. Med. 4: 594–600.

11. Escudier, B., T. Dorval, N. Chaput, F. Andre, M. P. Caby, S. Novault,C. Flament, C. Leboulaire, C. Borg, S. Amigorena, et al. 2005. Vaccination ofmetastatic melanoma patients with autologous dendritic cell (DC) derived-exo-somes: results of the first phase I clinical trial. J. Translational Med. (electronicresource) 3: 10.

12. Dai, S., D. Wei, Z. Wu, X. Zhou, X. Wei, H. Huang, and G. Li. 2008. Phase Iclinical trial of autologous ascites-derived exosomes combined with GM-CSF forcolorectal cancer. Mol. Ther. 16: 782–790.

13. Morse, M. A., J. Garst, T. Osada, S. Khan, A. Hobeika, T. M. Clay, N. Valente,R. Shreeniwas, M. A. Sutton, A. Delcayre, et al. 2005. A phase I study of dexo-some immunotherapy in patients with advanced non-small cell lung cancer.J. Translational Med. (electronic resource) 3: 9.

14. Antoniou, A. N., S. J. Powis, and T. Elliott. 2003. Assembly and export of MHCclass I peptide ligands. Curr. Opin. Immunol. 15: 75–81.

15. Fooksman, D. R., G. K. Gronvall, Q. Tang, and M. Edidin. 2006. Clustering classI MHC modulates sensitivity of T cell recognition. J. Immunol. 176: 6673–6680.

16. Segura, J. M., P. Guillaume, S. Mark, D. Dojcinovic, A. Johannsen, G. Bosshard,G. Angelov, D. F. Legler, H. Vogel, and I. F. Luescher. 2008. Increased mobilityof major histocompatibility complex I-peptide complexes decreases the sensitiv-ity of antigen recognition. J. Biol. Chem. 283: 24254–24263.

17. Bryceson, Y. T., and E. O. Long. 2008. Line of attack: NK cell specificity andintegration of signals. Curr. Opin. Immunol. 20: 344–352.

18. Allen, R. L., C. A. O’Callaghan, A. J. McMichael, and P. Bowness. 1999. Cuttingedge: HLA-B27 can form a novel � 2-microglobulin-free heavy chain ho-modimer structure. J. Immunol. 162: 5045–5048.

19. Kollnberger, S., L. A. Bird, M. Roddis, C. Hacquard-Bouder, H. Kubagawa,H. C. Bodmer, M. Breban, A. J. McMichael, and P. Bowness. 2004. HLA-B27heavy chain homodimers are expressed in HLA-B27 transgenic rodent models ofspondyloarthritis and are ligands for paired Ig-like receptors. J. Immunol. 173:1699–1710.

20. Turner, M. J., D. P. Sowders, M. L. DeLay, R. Mohapatra, S. Bai, J. A. Smith,J. R. Brandewie, J. D. Taurog, and R. A. Colbert. 2005. HLA-B27 misfolding intransgenic rats is associated with activation of the unfolded protein response.J. Immunol. 175: 2438–2448.

21. Stam, N. J., H. Spits, and H. L. Ploegh. 1986. Monoclonal antibodies raisedagainst denatured HLA-B locus heavy chains permit biochemical characterizationof certain HLA-C locus products. J. Immunol. 137: 2299–2306.

22. Stam, N. J., T. M. Vroom, P. J. Peters, E. B. Pastoors, and H. L. Ploegh. 1990.HLA-A- and HLA-B-specific monoclonal antibodies reactive with free heavy

1890 EXOSOME MHC CLASS I

by guest on February 2, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

chains in Western blots, in formalin-fixed, paraffin-embedded tissue sections andin cryo-immuno-electron microscopy. Int. Immunol. 2: 113–125.

23. Ellis, S. A., C. Taylor, and A. McMichael. 1982. Recognition of HLA-B27 andrelated antigen by a monoclonal antibody. Hum. Immunol. 5: 49–59.

24. Parham, P., and F. M. Brodsky. 1981. Partial purification and some properties ofBB7.2: a cytotoxic monoclonal antibody with specificity for HLA-A2 and a vari-ant of HLA-A28. Hum. Immunol. 3: 277–299.

25. Barnstable, C. J., W. F. Bodmer, G. Brown, G. Galfre, C. Milstein,A. F. Williams, and A. Ziegler. 1978. Production of monoclonal antibodies togroup A erythrocytes, HLA and other human cell surface antigens-new tools forgenetic analysis. Cell 14: 9–20.

26. Hanke, T., P. Szawlowski, and R. E. Randall. 1992. Construction of solid matrix-antibody-antigen complexes containing simian immunodeficiency virus p27 us-ing tag-specific monoclonal antibody and tag-linked antigen. J. Gen. Virol. 73:653–660.

27. Segura, E., C. Nicco, B. Lombard, P. Veron, G. Raposo, F. Batteux,S. Amigorena, and C. Thery. 2005. ICAM-1 on exosomes from mature dendriticcells is critical for efficient naive T-cell priming. Blood 106: 216–223.

28. Shevchenko, A., M. Wilm, O. Vorm, and M. Mann. 1996. Mass spectrometricsequencing of proteins silver-stained polyacrylamide gels. Anal. Chem. 68:850–858.

29. Antoniou, A. N., S. Ford, E. S. Pilley, N. Blake, and S. J. Powis. 2002. Interac-tions formed by individually expressed TAP1 and TAP2 polypeptide subunits.Immunology 106: 182–189.

30. Lemin, A. J., K. Saleki, M. van Lith, and A. M. Benham. 2007. Activation of theunfolded protein response and alternative splicing of ATF6� in HLA-B27 pos-itive lymphocytes. FEBS Lett. 581: 1819–1824.

31. Ackerman, A. L., and P. Cresswell. 2003. Regulation of MHC class I transport inhuman dendritic cells and the dendritic-like cell line KG-1. J. Immunol. 170:4178–4188.

32. Mignot, G., S. Roux, C. Thery, E. Segura, and L. Zitvogel. 2006. Prospects forexosomes in immunotherapy of cancer. J. Cell. Mol. Med. 10: 376–388.

33. Velikovsky, C. A., L. Deng, L. K. Chlewicki, M. M. Fernandez, V. Kumar, andR. A. Mariuzza. 2007. Structure of natural killer receptor 2B4 bound to CD48reveals basis for heterophilic recognition in signaling lymphocyte activation mol-ecule family. Immunity 27: 572–584.

34. Bowness, P. 2002. HLA B27 in health and disease: a double-edged sword? Rheu-matology 41: 857–868.

35. Jessop, C. E., and N. J. Bulleid. 2004. Glutathione directly reduces an oxi-doreductase in the endoplasmic reticulum of mammalian cells. J. Biol. Chem.279: 55341–55347.

36. Momburg, F., V. Ortiz-Navarrete, J. Neefjes, E. Goulmy, Y. van de Wal,H. Spits, S. J. Powis, G. W. Butcher, J. C. Howard, P. Walden, et al. 1992.Proteasome subunits encoded by the major histocompatibility complex are notessential for antigen presentation. Nature 360: 174–177.

37. Hunt, D. F., R. A. Henderson, J. Shabanowitz, K. Sakaguchi, H. Michel,N. Sevilir, A. L. Cox, E. Appella, and V. H. Engelhard. 1992. Characterization of

peptides bound to the class I MHC molecule HLA-A2.1 by mass spectrometry.Science 255: 1261–1263.

38. Capps, G. G., B. E. Robinson, K. D. Lewis, and M. C. Zuniga. 1993. In vivodimeric association of class I MHC heavy chains: possible relationship to class IMHC heavy chain-� 2-microglobulin dissociation. J. Immunol. 151: 159–169.

39. Kollnberger, S., A. Chan, M. Y. Sun, L. Y. Chen, C. Wright, K. di Gleria,A. McMichael, and P. Bowness. 2007. Interaction of HLA-B27 homodimers withKIR3DL1 and KIR3DL2, unlike HLA-B27 heterotrimers, is independent of thesequence of bound peptide. Eur. J. Immunol. 37: 1313–1322.

40. Pisitkun, T., R. F. Shen, and M. A. Knepper. 2004. Identification and proteomicprofiling of exosomes in human urine. Proc. Natl. Acad. Sci. USA 101:13368–13373.

41. Wring, S. A., J. P. Hart, and B. J. Birch. 1989. Determination of glutathione inhuman plasma using high-performance liquid chromatography with electrochem-ical detection with a carbon-epoxy resin composite electrode chemically modifiedwith cobalt phthalocyanine. Analyst 114: 1571–1573.

42. Gruda, R., H. Achdout, N. Stern-Ginossar, R. Gazit, G. Betser-Cohen,I. Manaster, G. Katz, T. Gonen-Gross, B. Tirosh, and O. Mandelboim. 2007.Intracellular cysteine residues in the tail of MHC class I proteins are crucial forextracellular recognition by leukocyte Ig-like receptor 1. J. Immunol. 179:3655–3661.

43. Kaufman, J. F., M. S. Krangel, and J. L. Strominger. 1984. Cysteines in thetransmembrane region of major histocompatibility complex antigens are fattyacylated via thioester bonds. J. Biol. Chem. 259: 7230–7238.

44. Chambers, J. E., C. E. Jessop, and N. J. Bulleid. 2008. Formation of a majorhistocompatibility complex class I tapasin disulfide indicates a change in spatialorganization of the peptide-loading complex during assembly. J. Biol. Chem. 283:1862–1869.

45. Mitra, A. K., H. Celia, G. Ren, J. G. Luz, I. A. Wilson, and L. Teyton. 2004.Supine orientation of a murine MHC class I molecule on the membrane bilayer.Curr. Biol. 14: 718–724.

46. Apps, R., L. Gardner, A. M. Sharkey, N. Holmes, and A. Moffett. 2007. Ahomodimeric complex of HLA-G on normal trophoblast cells modulates antigen-presenting cells via LILRB1. Eur. J. Immunol. 37: 1924–1937.

47. Shiroishi, M., K. Kuroki, T. Ose, L. Rasubala, I. Shiratori, H. Arase, K. Tsumoto,I. Kumagai, D. Kohda, and K. Maenaka. 2006. Efficient leukocyte Ig-like recep-tor signaling and crystal structure of disulfide-linked HLA-G dimer. J. Biol.Chem. 281: 10439–10447.

48. Viaud, S., M. Terme, C. Flament, J. Taieb, F. Andre, S. Novault, B. Escudier,C. Robert, S. Caillat-Zucman, T. Tursz, L. Zitvogel, and N. Chaput. 2009. Den-dritic cell-derived exosomes promote natural killer cell activation and prolifera-tion: a role for NKG2D ligands and IL-15R�. PLoS ONE 4: e4942.

49. Capps, G. G., and M. C. Zuniga. 1993. The cytoplasmic domain of the H-2Ldclass I major histocompatibility complex molecule is differentially accessible toimmunological and biochemical probes during transport to the cell surface.J. Biol. Chem. 268: 21263–21270.

1891The Journal of Immunology

by guest on February 2, 2019http://w

ww

.jimm

unol.org/D

ownloaded from