Posttranslational Regulation of I-Ed by Affinity for CLIP Cornelia H. Rinderknecht, Michael P. Belmares, Tatiana L. W. Catanzarite, Alexander J. Bankovich, Tyson H. Holmes, This information is current as K. Christopher Garcia, Navreet K. Nanda, Robert Busch, of October 2, 2021. Susan Kovats and Elizabeth D. Mellins J Immunol 2007; 179:5907-5915; ; doi: 10.4049/jimmunol.179.9.5907 http://www.jimmunol.org/content/179/9/5907 Downloaded from

References This article cites 64 articles, 28 of which you can access for free at: http://www.jimmunol.org/content/179/9/5907.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on October 2, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Posttranslational Regulation of I-Ed by Affinity for CLIP1

Cornelia H. Rinderknecht,*† Michael P. Belmares,2‡ Tatiana L. W. Catanzarite,† Alexander J. Bankovich,3*§ Tyson H. Holmes,¶ K. Christopher Garcia,§ Navreet K. Nanda,ʈ Robert Busch,4† Susan Kovats,# and Elizabeth D. Mellins5†

Several MHC class II alleles linked with autoimmune diseases form unusually low stability complexes with CLIP, leading us to hypothesize that this is an important feature contributing to autoimmune pathogenesis. To investigate cellular consequences of altering class II/CLIP affinity, we evaluated invariant chain (Ii) mutants with varying CLIP affinity for a mouse class II allele, I-Ed, which has low affinity for wild-type CLIP and is associated with a mouse model of spontaneous, autoimmune joint inflammation. Increasing CLIP affinity for I-Ed resulted in increased cell surface and total cellular abundance and half-life of I-Ed. This reveals a post- chaperoning capacity of Ii via its CLIP peptides. Quantitative effects on I-Ed were less pronounced in DM-expressing cells, suggesting complementary chaperoning effects mediated by Ii and DM, and implying that the impact of allelic variation in CLIP affinity on immune responses will be highest in cells with limited DM activity. Differences in the ability Downloaded from of cell lines expressing wild-type or high-CLIP-affinity mutant Ii to present Ag to T cells suggest a model in which increased CLIP affinity for class II serves to restrict peptide loading to DM-containing compartments, ensuring proper editing of antigenic peptides. The Journal of Immunology, 2007, 179: 5907–5915.

ajor histocompatibility complex class II molecules are in these diseases and their animal models, it is postulated that class

polymorphic glycoproteins that are constitutively ex- II disease associations are related to the role of MHC class II http://www.jimmunol.org/ pressed on professional APCs, including dendritic molecules in Ag presentation. However, the specific mechanisms M6 cells (DC), B cells, and macrophages, and can be induced in other that link the polymorphisms to autoimmune pathogenesis have re- cell types by cytokines (1). Recognition of class II MHC/peptide mained elusive. A number of observations, including the associa- complexes by TCR is critical for appropriate thymic selection, tion of certain class II alleles with multiple autoimmune diseases immune responses to pathogens and tumors, and the maintenance with unrelated Ags (2, 3), argue that a general feature of class II of peripheral tolerance. Epidemiological studies have revealed alleles plays an important role in determining their association strong and statistically well-substantiated associations of MHC with susceptibility to . class II alleles with particular autoimmune diseases, including The major steps in the proper assembly, transport, and loading (RA), type 1 diabetes, multiple sclerosis, and of class II molecules have been well-described and are influenced by guest on October 2, 2021 ϩ celiac disease (2). As CD4 T cells are implicated in pathogenesis by three major accessory molecules: invariant chain (Ii), DM, and DO (reviewed in Refs. 3 and 4). Briefly, nascent class II molecules *Program in Immunology, †Department of Pediatrics, ‡Department of Chemistry, are assembled onto Ii trimers in the endoplasmic reticulum (ER). § ␣␤ Department of Molecular and Cellular Physiology, Structural Biology, and Howard Ii directs the trafficking of these nonameric ( )3Ii3 complexes out Hughes Medical Institute, ¶Division of Biostatistics, Department of Health Research and Policy, Stanford University, Stanford, CA 94305; ʈLaboratory of Cellular and Molecular of the ER, through the Golgi apparatus, and into the endosomal Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of class II-loading compartments. Different class II alleles vary in Health, Bethesda, MD 20892; and #Arthritis and Immunology Research Program, Okla- their dependence upon Ii for efficient assembly and egress from the homa Medical Research Foundation, Oklahoma City, OK 73104 ER (5). In endosomal compartments, Ii is degraded, leaving a Received for publication March 26, 2007. Accepted for publication August 17, 2007. nested set of peptides called CLIP as a surrogate ligand in the The costs of publication of this article were defrayed in part by the payment of page peptide-binding groove of the MHC class II molecule. Exchange charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. of CLIP for a diverse array of peptides is promoted by the peptide- 1 This work was supported by funds from the National Institutes of Health (to E.D.M., exchange catalyst DM (H-2M or H2-DM in mice, HLA-DM in C.H.R., K.C.G.), the National Science Foundation (to C.H.R.), the Arthritis Founda- humans), which is also thought to stabilize empty MHC class II tion (to E.D.M.), an Immunology Training Grant (to A.J.B.), the Howard Hughes Medical Institute (to T.L.W.C., K.C.G.), and the Lucille Packard Foundation for intermediates and edit the repertoire of bound endosomal peptides, Children’s Health (to T.H.H.). favoring stable peptide/MHC class II complexes. Another nonclas- 2 Current address: Arbor Vita Corporation, Sunnyvale, CA 94085. sical class II molecule, DO (H-2O or H2-DO in mice, HLA-DO in 3 Current address: Department of Microbiology and Immunology, University of Cal- humans), is a negative regulator of DM function expressed in a ifornia, San Francisco, CA 94143. subset of APCs, including B cells, thymic medullary epithelial 4 Current address: Sidney Sussex College and Department of Medicine, Cambridge cells, and specific subsets of DC (4, 6–9). Reduced or absent DM University, Cambridge, U.K. function (for example, due to low DM expression or to coexpres- 5 Address correspondence and reprint requests to Dr. Elizabeth D. Mellins, Depart- sion of DO) results in the accumulation of class II/CLIP ment of Pediatrics, Stanford University School of Medicine, 269 Campus Drive, CCSR 2115c, Stanford, CA 94305. E-mail address: [email protected] complexes. 6 Abbreviations used in this paper: DC, ; RA, rheumatoid arthritis; Ii, Of the molecules implicated in general mechanisms of Ag pre- invariant chain; ER, endoplasmic reticulum; HEL, hen egg white lysozyme; MFI, sentation, CLIP is well positioned to be involved in mechanisms mean fluorescence intensity; wt, wild type; GEE, generalized estimating equation; that link particular MHC class II polymorphisms to autoimmune co-IP, coimmunoprecipitation; CHX, cycloheximide. disease. CLIP occupies the class II-binding groove much like an- Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 tigenic peptides do, and different MHC class II alleles differ widely www.jimmunol.org 5908 CLIP CHAPERONING OF I-Ed

Table I. Effect of Ii mutations on persistence of I-Ed/CLIP I-Ed (23). 14-4-4S is a mAb to the common I-E␣ chain (24). 14-4-4S cells coimmunoprecipitation for production of supernatant and ascites were obtained from ATCC, clone HB-32. Purified, FITC-conjugated 14-4-4S, and an IgG2a-FITC isotype control were obtained from Southern Biotechnology Associates. The fol- Anchor Pocket P1 P4 P6 P9 lowing Abs were used together with appropriate HRP-conjugated second- ␤ wt CLIP anchor residue M90 A93 P95 M98 ary Abs for Western blotting: anti- -actin (Sigma-Aldrich), In-1 (rat anti- mouse Ii; BD Biosciences/BD Pharmingen), anti-Tetra-His (Qiagen), and Greatest effect RKrabbit-anti-I-Ed␣ cytoplasmic tail antiserum (25) (gift of R. N. Germain). Moderate effect F,aI,L K,I,F Little or no detectable effect G,I,L,V F,Lb Vectors and transfections a The single most effective mutant at each pocket is shown in bold. Mutations in Methods for site-directed mutagenesis, cloning, and retroviral transduction the “moderate” category are listed from most to least effective. have been previously described (26). Briefly, mutant p31 murine Ii cDNAs b M98L was included in this analysis as a control: it was expected to have no effect were generated by site-directed mutagenesis, using overlap extension PCR d on CLIP affinity for I-E . and high-fidelity Pfu polymerase, with a pGEM-mIi-p31 construct as the original template (gift of E. K. Bikoff, University of Oxford, Oxford, U.K.). Ii cDNAs were cloned into the retroviral vector pBMN-IRES-Neo and in their affinity for CLIP (10, 11). We previously compared RA- verified by sequencing. Vectors were transfected into Phoenix-A cells by associated alleles of DR4 (DR*0401, *0404, and *0405) with calcium phosphate precipitation. Phoenix-A cell supernatant containing closely related, non-RA-associated alleles (DR*0402 and *0403) retroviral particles was harvested and used to infect A20, 3A5, or L cells. Polyclonal populations expressing the Ii constructs were obtained by G418 and found that the disease-associated alleles form less stable com- selection. plexes with CLIP both in vitro and in living cells (12). Evidence from others supports our hypothesis, as a disproportionate number Pulse chase and immunoprecipitation Downloaded from of disease-associated alleles have been found to have low affinity Cells were washed in cysteine/methionine-free DMEM (Invitrogen Life for CLIP (reviewed in Ref. 3). The low affinity for CLIP leads Technologies) to remove unlabeled cysteine and methionine. Cells were these alleles to spontaneously release CLIP peptides, even in the starved for 1–2 h (at 37°C, 5% CO2) in starvation medium (Cys/Met-free absence of DM (12, 13). DMEM plus 10% dialyzed FBS plus 2 mM L-glutamine), then labeled for 1 h (at 37°C, 5% CO ) with 125 ␮Ci/ml ExpreSS [35S]labeling mix These observations led us to propose that low affinity for CLIP 2 (PerkinElmer). Excess free radiolabel was removed by washing with com- is a critical feature of disease-linked MHC class II alleles and plete medium (DMEM plus 10% FBS plus 2 mM L-glutamine), and cells http://www.jimmunol.org/ contributes to disease pathogenesis. We reasoned that, to influence were chased in complete medium for the designated times. Cells were lysed an in a whole organism, low CLIP affinity in buffer containing 1% Nonidet P-40, 50 mM Tris-HCl, 150 mM NaCl, must have molecular consequences detectable at the level of indi- and 5 mM EDTA (pH 8.0) and complete protease inhibitors (Roche Di- agnostics). Lysates were cleared of nuclear and cellular debris by centrif- vidual APCs. Comparisons of different class II alleles with varying ugation, and were precleared several times with combinations of normal CLIP affinity are confounded by additional variations in the char- mouse serum, rabbit-anti-mouse IgG (Zymed Laboratories), Pansorbin acteristics of the alleles, such as peptide-binding specificity and (heat-killed Streptococcus aureus cells; Calbiochem/EMD Biosciences), inherent dimer stability. Thus, we chose to introduce targeted mu- and protein A-Sepharose (Amersham Pharmacia Biotech). Samples were normalized either for starting cell number at time 0 or for counts following tations into the CLIP region of Ii constructs to vary CLIP affinity d d d I-E IP, using a 1450 Microbeta beta counter (PerkinElmer/Wallac). I-E by guest on October 2, 2021 for a single disease-linked class II allele. I-E is a murine class II was immunoprecipitated by incubating lysates with protein A-Sepharose allele with extremely low affinity for wt CLIP (10, 11), and is beads coated with 14-4-4S ascites or concentrated supernatant. Proteins associated with a spontaneous model of autoimmune joint inflam- were eluted from the beads by boiling in reducing SDS sample buffer and mation which may share some features with human RA (A. Caton, separated by SDS-PAGE. Gels were treated with Amplify (Amersham Bio- sciences), dried under vacuum, and exposed to radiography film (Kodak). unpublished observations). In this study, we identify mutations in d Ii that increase the affinity of CLIP for I-E , and use the resulting Flow cytometry panel of Ii mutants to determine the effect of varying CLIP affinity on I-Ed abundance, turnover, and Ag presentation capacity in Cells were stained on ice with FITC-conjugated 14-4-4S or isotype control. For cell surface FACS with L cells, propidium iodide was used to exclude model APC lines. dead/dying cells. For intracellular staining, fixation and permeabilization was performed using the Cytofix/Cytoperm kit (BD Biosciences/BD Materials and Methods Pharmingen). Data were collected using a FACScan or FACSCalibur flow Molecular modeling cytometer (BD Biosciences) and CellQuest Pro software (BD Biosciences), and were analyzed using FlowJo software (Tree Star). The mean fluores- The structure of I-Ed was predicted based on the solved crystal structure of cence intensity (MFI) of isotype controls was routinely under 10. MFI of I-Ek (14, 15) using the LOOK Software Suite (Molecular Applications staining on cells expressing mutant Ii was normalized to the appropriate Group). Structural and sequence comparison revealed homology within the (untagged or 6xHis-tagged) wild-type (wt) control within the same exper- d k ϫ ϭ P1 and P9 anchor pockets of I-E and I-E . Amino acids determined by iment: 100% MFImut/MFIwt MFI of mutant as a percentage of wt. experimental binding data to be highly favorable for binding I-Ek at these Data represent staining of polyclonal populations from multiple indepen- pockets (16) were predicted to be favorable for binding I-Ed, and were dent transfections and selections. modeled with I-Ed to confirm a qualitative fit of shape, size, hydrophobic- ity, and H-bond interactions. Modeling of all mutants reported in this study Immunoblotting and densitometry (Table I) was repeated using Modeler 8v2 (17). These coordinates were used to generate structure figures describing possible peptide-MHC inter- Cells were harvested and solubilized in lysis buffer containing 1% Nonidet I-Ed actions in this system with Pymol (18). P-40. For analysis of steady-state levels of in untreated cells, protein content of lysates was determined by Bradford assay and titrated amounts Cell lines and Abs of lysate normalized for protein amount were boiled in reducing SDS sam- ple buffer. For analysis of the half-life of I-Ed, cells were treated with 10 A20 is a B lymphoma cell line from a BALB/c mouse (H-2d haplotype; ␮g/ml cycloheximide (CHX). Samples were collected at the indicated time American Type Culture Collection (ATCC) clone TIB-208) (19). 3A5 is a points, solubilized in lysis buffer containing 1% Nonidet P-40, and nor- derivative of A20 which lacks H-2M␣ expression (20, 21). RT 10.3 C1 are malized for cell equivalents before boiling in reducing SDS sample buffer. L cells (a murine fibroblast cell line) transfected with I-Ed (22) (gift of Samples were separated by SDS-PAGE and transferred to Immobilon poly- R. N. Germain, National Institute of Allergy and Infectious Diseases, Be- vinylidene difluoride membrane (Millipore). Binding of primary Abs to the thesda, MD). Phoenix-A cells are a retroviral packaging cell line (gift of membrane was detected by HRP-conjugated secondary Abs followed by G. P. Nolan, Stanford University, Stanford, CA). D2-4-8 is a hy- Western Lightning ECL substrates (PerkinElmer Life Sciences) and expo- bridoma specific for hen egg white lysozyme (HEL) 106–116 bound to sure to Hyperfilm ECL (Amersham Biosciences). Due to the wide variation The Journal of Immunology 5909

model (30). Plate was modeled as a random factor to account for correla- tion among wells within an experiment. For A20, analysis used generalized estimating equations (GEE) (31), because data were too leptokurtic to meet normality assumptions of mixed models. Mixed model and GEE analyses were performed in SAS version 9.1 (SAS Institute).

Results Identification of mutations in murine Ii that increase affinity of CLIP for I-Ed FIGURE 1. Identification of Ii mutants with increased CLIP affinity for I-Ed. 3A5 cells transfected with mutant Ii constructs were labeled with Mutations in murine Ii that were likely to increase the affinity of [35S]cysteine and methionine, and I-Ed was immunoprecipitated with the CLIP for I-Ed were chosen based on predictions using crystal anti-I-E␣ Ab 14-4-4S. Samples were normalized for starting cell number at structures of I-Ek (14, 15) to model I-Ed, and on published peptide- time 0 and analyzed by SDS-PAGE. Mutations in Ii that increase the binding motifs (32–34). None of the wt CLIP anchors (M90, A93, d amount and duration of CLIP co-IP with I-E were identified as high- P95, and M98) conformed to known I-Ed-binding motifs. The P1 affinity mutations. Shown is one experiment that is representative of re- and P9 methionine anchors in CLIP are thought to be moderate peated experiments with single, double, and triple mutants, in the presence anchors for most class II alleles, leaving room for allele-specific and absence of a 6xHis tag, at varying chase time points. improvement. Our random mutagenesis of DR (homolog of I-E) revealed that mutations which increased the affinity of DR3 for wt CLIP were centered around the P4 and P6 pockets (35). In addi- in the sensitivity of the primary Abs being used for detection, all experi- Downloaded from ments included two gels run in parallel on the same day from the same tion, molecular modeling suggested that the reported strong asso- lysates: one with high protein amounts loaded for detection of I-Ed (anti- ciation between wt CLIP and I-Ed␣/I-Eb␤ molecules (36) was due I-Ed␣ antiserum) and His-Ii (anti-Tetra-His), and one gel with low protein to differences between I-Ed␤ and I-Eb␤ centered around the P4 and ␤ amounts loaded for detection of -actin and total cellular Ii (In-1). Den- P6 pockets (data not shown). Based on these observations, we sitometry was performed using a Bio-Rad GS-710 densitometer and Quan- tityOne software (Bio-Rad). made selected mutations at P1, P4, P6, and P9 (Table I) and tested these empirically, as the relative contributions of specific anchors T cell stimulation assays are often context dependent (37). Single amino acid mutations http://www.jimmunol.org/ D2-4-8 T hybridoma cells (105) were cultured with A20 or 3A5 transfec- were incorporated into cDNA constructs for the p31 isoform of tants (5 ϫ 104) and different concentrations of whole HEL (Sigma-Aldrich) murine Ii. Most constructs (specified in figures) included a C-ter- or HEL 106–116 peptide, as indicated in the figure labels and legends. All minal 6xHis tag for subsequent differentiation between endoge- cultures were done in duplicate. Culture supernatants were collected at nous and transfected Ii in cell lines. Mutant Ii constructs and wt 20–24 h, and assayed for IL-2 with the BD OptEIA Mouse IL-2 ELISA kit and BD OptEIA TMB substrate (BD Biosciences). control constructs were expressed in A20 (a murine line with H-2d haplotype) and 3A5 (a DM-deficient derivative of A20), Statistical methods and in L cells (a murine fibroblast cell line that does not express d

For the analysis of FACS data in Figs. 2–5, normalized MFIs of wt and DM) transfected with I-E . In each of these cell lines, transfected by guest on October 2, 2021 mutant Ii transfectants were compared using paired Student’s t tests Ii must compete with endogenous Ii. Expression of endogenous Ii (GraphPad Prism; GraphPad Software). Within each figure, we corrected in L cells varies with the sublines; the L cells used in this study for multiple comparisons across t tests using sequential Bonferroni adjust- express endogenous Ii at levels comparable to those found in A20 ment (27). For titration analyses, MFI was modeled as a logistic response of log -transformed 14-4-4S concentration. Nonlinear regression used the and 3A5 (data not shown). 10 d quasi-Newton method (28). A t test was used to compare the positions of To screen the Ii mutants for increased affinity for I-E , we used the plateaus (at 5 ␮g) between groups. SEs for t tests were calculated using a pulse-chase assay in 3A5 cells. In the absence of DM, class a first-order ␦ approximation (29). For Fig. 7, in preparation for analysis, II/CLIP complexes are expected to accumulate. However, because data were smoothed by calculating the mean for each combination of pep- I-Ed has extremely low affinity for wt CLIP, very little CLIP is tide concentration and genotype within a plate (experiment). Analysis was d then performed on these means. For 3A5, we compared the mean for wt coimmunoprecipitated (co-IP) with I-E , and the CLIP band van- against the mean for each mutant, at each dose level, using a mixed-effect ishes rapidly as CLIP spontaneously dissociates from I-Ed (Fig. 1,

FIGURE 2. Increased cell surface I-Ed staining in the presence of high-affinity CLIP. 3A5 cells transfected with mutant Ii constructs with (A) or without (B) a 6xHis tag were stained with 14-4-4S-FITC. The MFI of isotype controls was routinely under 10 (data not shown). MFI of staining on cells expressing mutant Ii is normalized to the appropriate (untagged or 6xHis-tagged) wt control within the same experiment. Data represent staining of polyclonal p Ͻ ,ءء ;p Ͻ 0.05 ,ء :populations from multiple independent transfections and selections. Statistical significance was determined by paired Student’s t test p Ͻ 0.001. With sequential Bonferroni correction for multiple comparisons, differences between His-A93I and His-wt and between His-A93F ,ءءء ;0.01 and His-wt are not significant. 5910 CLIP CHAPERONING OF I-Ed left). The addition of a single high-affinity mutation at the P9 pocket (M98K) results in a clear increase both in the maximal amount of I-Ed-associated CLIP, and in the duration of I-Ed/CLIP association (Fig. 1, center). The effects of the 14 single amino acid mutations on persistence of I-Ed/CLIP co-IP are summarized in Table I. The effects of individual mutations on CLIP retention by I-Ed are cumulative, as Ii constructs containing two or three mu- tations incrementally increase accumulation and retention of CLIP (Fig. 1, right, and data not shown). Thus, our single, double, and triple Ii mutants constitute a panel of Ii mutants with a wide range of CLIP affinity for I-Ed. Because this assay uses radioactively labeled methionine for detection, and both P1 and P9 mutations remove a methionine label from CLIP, this assay substantially un- derestimates the amount of mutant CLIP on these gels. Parallel pulse-chase experiments in A20 cells (DMϩ) do not show signif- icant accumulation of mutant CLIP peptides (data not shown), sug- gesting that these mutants remain DM-susceptible and are within a physiologically relevant range of CLIP affinity.

Increased CLIP affinity for I-Ed results in increased cell surface Downloaded from levels of I-Ed To begin studying the effects of varying CLIP affinity on I-Ed in a cellular context, we examined the level of cell surface I-Ed ex- pression in the presence of wt (low CLIP affinity) vs mutant (high CLIP affinity) Ii by flow cytometry. The Ii mutants that caused the greatest increase in CLIP association with I-Ed in pulse-chase ex- http://www.jimmunol.org/ periments (Table I) also resulted in increased cell surface staining with the monoclonal anti-I-E␣ Ab 14-4-4S in transfectants of the DM-deficient B cell line 3A5 (Fig. 2). The effects of individual mutations were again found to be cumulative, with an apparent plateau in the extent of the effect (Fig. 3). Furthermore, the in- crease in cell surface I-Ed in the presence of high-CLIP-affinity Ii is independent of the 6xHis tag (Fig. 3, A vs B) and of the partic- ular host cell line (Fig. 3, A vs C). by guest on October 2, 2021 The binding of some mAbs to cell surface class II molecules is FIGURE 3. Effects of individual Ii mutations on cell surface I-Ed stain- d influenced by the bound peptide repertoire and the conformation of ing are cumulative and reflect a change in cell surface abundance of I-E . the class II molecules. Thus, increased binding of 14-4-4S could Site-directed mutagenesis was used to generate Ii constructs with multiple CLIP mutations favorable for binding I-Ed. A–C, 3A5 cells transfected with reflect either an increase in abundance of I-Ed molecules at the cell untagged (A) or 6xHis-tagged (B) Ii constructs and L cells plus I-Ed trans- surface or change in conformation that increases affinity for 14-4- fected with untagged Ii constructs (C) were stained with 14-4-4S-FITC. 4S. To differentiate between these two scenarios, we measured MFI of cells expressing mutant Ii is normalized to the appropriate (un- d levels of I-E binding with titrated amounts of Ab. If differences in tagged or 6xHis-tagged) wt control within each experiment. Statistical sig- 14-4-4S binding reflect a change in Ab affinity for I-Ed without a nificance is determined by paired Student’s t test. In 3A5 (A and B), p Ͻ change in the number of binding sites, then these differences 0.001 for all wt vs mutant comparisons and for all M98K single-mutant vs should be overcome by saturating levels of Ab. If, however, the double- or triple-mutant comparisons. For double- vs triple-mutant com- number of I-Ed molecules is increased by the Ii mutation, we parisons, only His-FK vs His-FRK is statistically significant (p ϭ 0.0165). would expect to observe different plateau levels of staining with All of these differences are statistically significant even after correction for d Ͻ saturating amounts of Ab. Titrations showed that the observed dif- multiple comparisons. In L cells plus I-E (C), p 0.001 for wt vs FKK, wt vs FRK, and K vs FKK, and p ϭ 0.0056 for K vs FRK (all statistically ferences in 14-4-4S staining cannot be overcome by addition of ex- Ͻ d significant after correction for multiple comparisons). p 0.05 for wt vs cess Ab (Fig. 3D), indicating that increased affinity of CLIP for I-E FK and for K vs FK (not significant after correction for multiple compar- d results in increased abundance of I-E at the cell surface. Increased isons). D, 3A5 cells were stained with titrated amounts of 14-4-4S-FITC. d d cell surface I-E correlated with increased I-E /CLIP affinity and was Shown is one experiment representative of two experiments with the His- not simply a result of increased expression of the transfected Ii mu- M90F-M98K double mutant and several experiments with P1, P4, and P9 tant. Transfection with wt or 6xHis tagged wt Ii did not increase cell single mutants. Titration curves were fit by nonlinear regression. A t test surface I-Ed compared with untransfected cells or cells transfected was used to compare the positions of the plateaus (at 5 ␮g) between His-wt with control vectors lacking Ii (data not shown). and His-M90F-M98K titrations (p Ͻ 0.0001). Statistical analysis was not The same Ii mutations that increased I-Ed expression in DM- performed on titrations for single mutants. negative cells (3A5 and L cells, Figs. 2 and 3) also did so in

DM-expressing A20 cells (Fig. 4) across multiple transfections/ d selections and in both the presence and absence of a 6xHis tag. The The cell surface increase in I-E in the presence of high-affinity d CLIP is associated with an increase in total cellular abundance effects of varying CLIP affinity on cell surface I-E expression d levels were more modest in the presence of DM (Fig. 2 vs Fig. 4). of I-E For example, for M98K, MFI ϭ 141% of wt (mean MFI from all At least two distinct mechanisms could explain the increase in cell experiments) in 3A5 vs 113% of wt in A20. Cumulative effects of surface levels of I-Ed in the presence of high-affinity CLIP. Tightly multiple CLIP mutations were not detected in DMϩ cells. bound, high-affinity CLIP mutants could induce a more mature The Journal of Immunology 5911

FIGURE 4. The effect of CLIP affinity on cell surface I-Ed is less robust in the presence of DM. A20 cells transfected with mutant Ii constructs with (A) or without (B) a 6xHis tag were stained with 14-4-4S-FITC. MFI of stain- ing on cells expressing mutant Ii is normalized to the appropriate (untagged or 6xHis-tagged) wt control within the same experiment. Data represent staining of polyclonal populations from multiple independent transfections and selections. Statistical significance was deter- ;p Ͻ 0.05 ,ء :mined by paired the Student t test p Ͻ 0.001. With sequential ,ءءء ;p Ͻ 0.01 ,ءء Bonferroni correction for multiple compari- sons, the difference between triple mutant M90F-A93K-M98K and wt is not significant.

conformation of I-Ed that is preferentially exported to the cell surface, true, we treated 3A5 transfectants with CHX to block de novo d d causing an altered distribution of I-E molecules between intracellular protein synthesis and assessed the level of surviving I-E at various Downloaded from and cell surface compartments. Alternatively, increased CLIP affinity time points by Western blot. In the presence of wt Ii (low-affinity could result in an increase in total cellular (and surface) levels of I-Ed. CLIP), the amount of I-Ed in CHX-treated cells begins to decline Detection of total cellular levels of I-Ed in 3A5 transfectants by com- after 8 h (Fig. 6, left). However, in the presence of an Ii triple bined cell surface and intracellular FACS (Fig. 5A) and by Western mutant with high CLIP affinity, the amount of I-Ed remains blotting of whole cell lysates (Fig. 5, B and C) revealed an increase in constant over at least 22 h (Fig. 6, right), indicating that high- d d total cellular levels of I-E in the presence of high-affinity CLIP mu- affinity CLIP mutants protect against I-E turnover. Over the http://www.jimmunol.org/ tants. The increase was on the order of 10–40% by FACS for single same period of observation, ␤-actin levels remain constant, and and double mutants, and 1.6- to 4-fold by Western blot for triple total cellular Ii turns over at a similar rate in both cell lines, d mutants. It appears that the change in total cellular I-E is at least as confirming that the effect is specific to I-Ed. Steady-state levels d robust as the change in cell surface levels of I-E , if not more so. of 6xHis-tagged Ii (0 h CHX treatment) were lower in triple mutant transfectants than in wt transfectants (also shown in Fig. The half-life of d is increased in the presence of high-affinity I-E 5). Densitometry reveals a modest increase in the rate of turn- CLIP mutants over of triple mutant 6xHis-Ii compared with wt (Fig. 6). This The increase in steady-state abundance of I-Ed is likely to be the increase in rate of turnover may be sufficient to explain the by guest on October 2, 2021 result of an increase in the half-life of I-Ed. To test whether this is differences in steady-state 6xHis-Ii levels.

FIGURE 5. Increased affinity of CLIP for I-Ed is associated with an increase in total cellular abundance of I-Ed. A, Total cellular levels of I-Ed in 3A5 transfectants were assessed by combined cell surface and intracellular FACS with 14-4-4S-FITC. MFI of staining on cells expressing mutant Ii is normalized p Ͻ 0.001. All three ,ءءء ;p Ͻ 0.01 ,ءء :to the wt control within the same experiment. Statistical significance is determined by paired the Student t test comparisons show statistically significant differences even after correction for multiple comparisons. B, Western blotting was used to detect steady-state total cellular levels of I-Ed in titrated amounts of whole cell lysates from 3A5 transfectants. In-1 was used to detect total levels of endogenous and transfected Ii. Anti-Tetra-His was used to detect levels of transfected Ii. This experiment is one representative of several experiments. Results are consistent in the presence of both tagged and untagged Ii, comparing wt to M90F-A93R-M98K and M90F-A93K-M98K (data not shown). C, Densitometry of the bands shown in B. 5912 CLIP CHAPERONING OF I-Ed

FIGURE 6. Increased affinity of CLIP for I-Ed increases survival time of I-Ed. 3A5 cells expressing wt or mutant Ii were treated with CHX to block de novo protein synthesis. The level of surviving I-Ed was assessed by Western blot at the indicated time points. Samples are normalized to represent the same number of live cells. At the chosen dose of CHX (10 ␮g/ml), 3A5 cells ceased to expand, but retained Ͼ90% viability. Un- treated cells expanded exponentially and maintained constant levels of I-Ed over the period of observation. Shown is one representative experi- ment of four. Densitometry: band in- tensity was normalized to ␤-actin band intensity at the same time point to account for variation in loading,

and then expressed as the percent re- Downloaded from maining compared with the 0-h time point to illustrate rate of turnover.

Increased CLIP affinity impedes presentation of exogenous peptide, mutant M90F (Table I). At the P9 pocket, I-Ed and I-Ek are essentially k but not presentation of peptide processed from whole Ag identical (Ref. 34 and our modeling). The I-E P9 pocket is a long http://www.jimmunol.org/ ␤ Increased cell surface levels of class II molecules on APCs are tunnel culminating in 9E, which generally forms a salt bridge with generally thought to be associated with increased capacity to stim- the P9 residue (almost exclusively lysine) of bound peptides (15). The ulate T cells (38–40). However, despite considerable increases in dramatic effect of the P9 CLIP mutant M98K is likely attributable to d d cell surface I-Ed, presentation of an exogenously added, I-Ed-re- the formation of a similar salt bridge with I-E (Fig. 8C). At the I-E stricted HEL peptide by 3A5 (DMϪ) transfectants to a T cell hy- P4 pocket, aliphatic residues are favorable, but positively charged bridoma is reduced in the presence of all but one (M90F) of the anchors are preferred (32–34). Accordingly, we found that A93R and high-affinity CLIP mutants (Fig. 7, A–C). This effect is more pro- A93K were more effective at P4 than A93I and A93F. Our modeling nounced at low, more physiologically relevant peptide concentra- predicts that a positively charged P4 anchor is likely to form a salt ␤ by guest on October 2, 2021 tions, and can be overcome by the addition of excess peptide. In A20 bridge with a negatively charged residue ( 70D) at the side of the P4 (DMϩ) cells, effects of CLIP mutations on exogenous peptide pre- pocket (Fig. 8B). We did not detect improvement in CLIP binding to d sentation are diminished (Fig. 7D). The presentation of the same pep- I-E for the four tested P6 mutations. Although a motif can be defined d d tide from exogenously added whole HEL that must be processed by for the I-E P6 pocket (32–34), many naturally occurring I-E ligands the APC is not impeded by the presence of high-affinity CLIP mutants adhere to binding motifs at the P1, P4, and P9 pockets, but not at P6 (Fig. 7E). The whole Ag-presentation assays in A20 confirm that DM (34). It is likely that the P6 proline of wt CLIP is already acceptable, is able to mediate adequate exchange of the high-affinity mutant CLIP and it may also help maintain the polyproline type II helical confor- to allow presentation of antigenic peptides. As expected, 3A5 cells are mation of class II-associated peptides. The favorable mutations at P1, unable to present this HEL peptide from whole HEL, because it is a P4, and P9 all yield similar phenotypes, suggesting that increasing DM-dependent epitope (data not shown). CLIP affinity has common consequences, regardless of the pocket involved. Discussion Another goal of our study was to elucidate the effects of increased The first goal of this study was to identify mutations that increase CLIP affinity on class II in APCs. Our data show that high-affinity the strength of binding of CLIP to I-Ed within a physiological CLIP mutants protect against molecular turnover. This result reveals range, where CLIP remains DM susceptible. We found that sub- a biochemical basis for the chaperoning effect of high-affinity CLIP, stitutions at the P1, P4, and P9 residues of CLIP (in the context of consistent with the importance of peptide occupancy for class II sta- full-length Ii) had the most significant effects. Our results are con- bility and longevity (42–46). Ii mutants that caused the greatest sta- sistent with published peptide-binding motifs (32–34) and with bilization of CLIP/I-Ed complexes also resulted in increased cell sur- predictions from sequence and structural similarities and differ- face and total cellular abundance of I-Ed. The mechanism for this ences between I-Ed and I-Ek (our molecular modeling and Refs. 16 chaperoning and increase in abundance of class II by high-affinity and 34). I-Ed uses the common I-E ␣-chain and has a ␤-chain with CLIP mutants is most likely very simple: increased affinity of CLIP 92% sequence homology to that of I-Ek, a molecule whose struc- for I-Ed likely reduces spontaneous CLIP dissociation, protecting ture has been solved (14, 15, 41). The P1 pockets of I-E molecules against the subsequent inactivation and loss of empty or unstably are generally large, deep, and hydrophobic (37), preferring aro- loaded I-Ed molecules. The major advantage of CLIP over other po- matic and aliphatic anchor residues. The P1 pockets of I-Ed and I-Ek tentially sufficient affinity peptides, which are thought to be present in differ only at ␤87; ␤87F in I-Ek reduces the pocket volume, while limiting amounts in class II-loading compartments (44–48), is that it ␤87S in I-Ed leaves a pocket that accommodates large aromatic pep- is preloaded in the ER in the context of full-length Ii. In the absence tide anchors (Ref. 34, illustrated in Fig. 8A). Accordingly, we ob- of DM, the increase in total I-Ed was at least as large as the increase served some increase in CLIP affinity for I-Ed with aliphatic P1 mu- in cell surface I-Ed, arguing against redistribution of I-Ed from intra- tants M90I and M90L, and a greater increase with the aromatic P1 cellular compartments to the cell surface as a major effect. The Journal of Immunology 5913

stably bind peptide due to disruption of the hydrogen-bonding net- work near the P1 and P2 pockets is rapidly degraded in the endosomes upon loss of Ii (49). However, when de novo syn- thesis was inhibited with CHX, we observed a drop in the remain- ing I-Ed in the presence of wt CLIP only after the 8-h time point (Fig. 6), which is beyond the expected time frame for CLIP gen- eration (Refs. 49 and 50 and Fig. 1) and endosomal residence (51). This may reflect a delay in degradation of I-Ed that is inactivated and/or aggregated shortly after CLIP generation. This scenario is supported by the fact that a pool of aggregated class II can be detected at steady-state in a B cell line (52). Alternatively, varying CLIP affinity may affect cell surface and recycling populations of class II, rather than nascent/endosomal class II. APC appear to discriminate between stable and unstable conformations of class II (45). The rescue of class II by high-affinity CLIP mutants and other high-affinity peptides could occur by providing sufficient stabili- zation to escape selective reinternalization and degradation of un- stable class II. Our ability to detect effects of mutant CLIP despite high levels of endogenous Ii suggests 1) that high-CLIP-affinity mutant Ii success- Downloaded from fully competes with endogenous wt Ii for assembly with I-Ed, and 2) that CLIP may contribute directly to early steps in the assembly of Ii with nascent class II molecules. Levels of total cellular Ii were indis- tinguishable in transfected and untransfected A20, 3A5, and L cells (data not shown), suggesting that the amount of transfected Ii is low compared with endogenous Ii. Interestingly, we detected reduced http://www.jimmunol.org/ steady-state levels and increased rate of turnover of 6xHis triple mu- tant Ii (Figs. 5 and 6, and data not shown). This may reflect the pref- erential assembly of high-CLIP-affinity Ii with nascent class II, and a subsequent increase in degradation as these Ii arrive in endosomes. Ii mutants M90L and M98L, which have ϳ24- and ϳ7-fold lower CLIP affinity for I-Ad than wt Ii, respectively (53), do not appear to compete effectively with endogenous wt Ii for assembly with I-Ad (data not shown). Within single APC, alleles with differing affinity for CLIP by guest on October 2, 2021 may compete for interaction with Ii in the ER, possibly contributing to allelic variation in abundance. Variation in CLIP affinity for I-Ed also affects the efficiency of presentation of an I-Ed-restricted HEL epitope. For presentation of exogenous HEL peptide by DM-null cells, the difficulty of dis- placing tightly bound CLIP outweighs the benefit of the increase in available I-Ed molecules. However, there may be a “window” of intermediate CLIP affinity (P1 single mutant M90F) in which CLIP affinity is high enough to stabilize and increase cell surface levels FIGURE 7. Presentation of an HEL epitope in the presence of high- of I-Ed, but still low enough to allow efficient peptide exchange, so CLIP-affinity Ii mutants. Presentation of the I-Ed-restricted HEL 106–116 that the outcome of the peptide presentation assay reflects the in- epitope was detected by a T cell hybridoma (D2-4-8). IL-2 production was d measured by ELISA. A, Presentation of exogenously added peptide by 3A5 creased I-E at the cell surface. The inhibition of peptide presen- (DMϪ) transfectants. One representative experiment of five. Error bars for tation by high-affinity mutant CLIP is almost completely masked ϩ d duplicates are omitted, as they are narrow enough to obscure the symbols. by the presence of DM. In DM APC, the majority of I-E mol- B–D, Summary of data from multiple experiments. For each Ag concen- ecules are loaded with a DM-edited repertoire of high-affinity pep- tration, IL-2 production is normalized to wt within the same experiment. B tides, and only the cohort of molecules that stochastically escapes and C, Presentation of exogenously added peptide by 3A5 transfectants (B, DM editing may be affected by the modulation in CLIP affinity. In untagged Ii constructs; C, 6xHis-tagged Ii constructs). D, Presentation of the presentation of peptide from whole HEL, there also may be a ϩ exogenously added peptide by A20 (DM ) transfectants. E, Presentation of “window” of intermediate CLIP affinity (M98K and M90F- HEL 106–116 from exogenously added whole Ag by A20 transfectants. M98K), in which CLIP affinity is sufficiently high to preserve high Means were compared between wt and each mutant using mixed effects I-Ed levels before interaction with DM, and sufficiently low to allow ,ءء ;p Ͻ 0.05 ,ء :modeling or GEE, as described in Materials and Methods p Ͻ 0.01. With sequential Bonferroni correction for multiple comparisons, for efficient CLIP removal by DM in favor of antigenic peptides. these differences are no longer statistically significant. Each bar represents Notably, the apparent windows for optimal presentation from peptide two to five experiments. and from whole Ag are offset, such that with increasing affinity within a limited range, cell surface peptide exchange is minimized just as presentation of a DM-selected epitope is optimized. This trend sug- A simple model would predict that a reduction in I-Ed abundance gests that increasing CLIP affinity for class II serves to restrict peptide in the presence of low-affinity CLIP would become apparent shortly loading to DM-containing compartments, ensuring proper editing of after Ii degradation and CLIP generation. This would be consistent antigenic peptides. This model is consistent with a previous study with a study showing that a mutant I-Ad molecule which is unable to which demonstrated a modest increase in the diversity of the I-Ab- 5914 CLIP CHAPERONING OF I-Ed

FIGURE 8. Modeling of CLIP M90F-A93R-M98K with I-Ed. I-Ed was modeled with wt (blue) or M90F-A93R-M98K (pink) CLIP, based on the solved structure of I-Ek.A, At the P1 pocket, the small I-Ed ␤87S residue leaves ample space to accommodate large P1 anchors such as CLIP M90F. B,Atthe P4 pocket, CLIP A93R (pictured) or A93K (data not shown) likely forms a salt bridge with I-Ed ␤70D. C, At the P9 pocket, CLIP M98K likely forms a salt bridge with I-Ed ␤9E, as is observed in crystal structures of I-Ek with P9K peptide anchors. restricted peptide repertoire on splenocytes from cells with mu- change (Refs. 59–62 and references therein). As the effects of CLIP tant low-CLIP-affinity Ii, as compared with wt high-CLIP-af- affinity on class II abundance are more robust in the absence of DM, Downloaded from finity Ii, even in the presence of DM (54). the impact of varied CLIP affinity on immune responses will be high- We expect that class II alleles will be differentially affected by est in cells with limited DM activity. In nonprofessional APC in variation in CLIP affinity. Gautam et al. (55) examined the fate of which class II expression is induced by IFN-␥, failure of class II and I-Ad and I-Au in the presence of mutant Ii with increased CLIP DM to colocalize in peptide-loading compartments can allow class II affinity for each class II allele. The high-CLIP-affinity Ii mutants to escape DM editing and chaperoning despite adequate expression moderately impaired Ag presentation from exogenous peptide and levels of DM (63). Coexpression of DO can also reduce DM function. http://www.jimmunol.org/ whole Ag. Honey et al. (54) used a reverse approach, choosing a DO is found in naive B cells, specific immature DC subsets, medul- high-CLIP-affinity allele, I-Ab, and creating a low-CLIP-affinity lary thymic epithelial cells, and in cells lining Hassall’s corpuscles, a mutant Ii. They noted an effect on cell surface I-Ab levels in the thymic structure involved in selection of regulatory T cells (6, 8, 9, 64, absence of DM, although they did not elucidate the mechanistic 65). These cells are involved in thymic selection and maintenance of basis of this phenotype, as we have done here. They also observed peripheral tolerance, suggesting that variation in CLIP affinity also changes in peptide and whole Ag presentation and overall peptide may be influential in establishment and maintenance of tolerance. repertoire diversity which, though modest, could be sufficient to Finally, multiple experimental systems suggest that prolonged or in- perturb carefully regulated thresholds for T cell stimulation. Dif- creased presence of CLIP may favor a Th2 response over a Th1 re- by guest on October 2, 2021 ferential allelic susceptibility to effects of CLIP affinity is probably sponse (66–68), a skewing expected to be protective against many determined by several factors, including inherent dimer stability autoimmune diseases. and efficiency of DM interaction. We would expect to see the greatest consequences of CLIP affinity in class II populations that Acknowledgments have low inherent dimer stability and/or reduced DM interaction, We thank Drs. Garry Nolan, Ronald Germain, and Elizabeth Bikoff for either due to allelic variation in affinity for DM or to altered DM generously sharing reagents. expression, localization, or regulation by DO in different APC. Disclosures These observations suggest numerous possible mechanisms for an The authors have no financial conflict of interest. influence of CLIP/class II affinity on susceptibility to autoimmunity. The Fathman and Nepom groups (56, 57) have proposed models in References which the instability of type 1 diabetes-associated class II alleles (I- 1. Glimcher, L. H., and C. J. Kara. 1992. Sequences and factors: a guide to MHC A97 in NOD mice and DQ8 or DQ3.2 in humans) leads to reduced class-II transcription. Annu. Rev. Immunol. 10: 13–49. 2. Jones, E. Y., L. Fugger, J. L. Strominger, and C. Siebold. 2006. MHC class II class II/peptide ligand density during T cell selection, which in turn proteins and disease: a structural perspective. Nat. Rev. Immunol. 6: 271–282. leads to compensatory selection of T cells with particularly high af- 3. Busch, R., C. H. Rinderknecht, S. Roh, A. W. Lee, J. J. Harding, T. Burster, finity for self peptides. The reduced abundance and half-life of class T. M. Hornell, and E. D. Mellins. 2005. Achieving stability through editing and chaperoning: regulation of MHC class II peptide binding and expression. Immu- II alleles in the presence of low-affinity CLIP could create a similar nol. Rev. 207: 242–260. situation. Reduced class II levels due to low CLIP affinity could also 4. Alfonso, C., and L. Karlsson. 2000. Nonclassical MHC class II molecules. Annu. impair thymic selection of regulatory T cells, which is thought to be Rev. Immunol. 18: 113–142. 5. Bikoff, E. K., R. N. Germain, and E. J. Robertson. 1995. Allelic differences dependent on strong recognition of self-peptide/MHC (58). Reduced affecting invariant chain dependency of MHC class II subunit assembly. Immu- CLIP affinity may also allow increased access of class II molecules to nity 2: 301–310. 6. Hornell, T. M., T. Burster, F. L. Jahnsen, A. Pashine, M. T. Ochoa, J. J. Harding, nonconventional, non-DM-regulated pathways of peptide loading, al- C. Macaubas, A. W. Lee, R. L. Modlin, and E. D. Mellins. 2006. Human dendritic tering the peptide repertoire and possibly lowering the overall stability cell expression of HLA-DO is subset specific and regulated by maturation. J. Im- of the class II/peptide complexes, thus increasing the susceptibility of munol. 176: 3536–3547. 7. Denzin, L. K., J. L. Fallas, M. Prendes, and W. Yi. 2005. Right place, right time, those complexes to further unsupervised peptide exchange at the cell right peptide: DO keeps DM focused. Immunol. Rev. 207: 279–292. surface. It is tempting to speculate that such cell surface exchange 8. Chen, X., L. M. Reed-Loisel, L. Karlsson, and P. E. Jensen. 2006. H2O expres- could contribute to presentation of cryptic self Ags, epitope spreading, sion in primary dendritic cells. J. Immunol. 176: 3548–3556. 9. Fallas, J. L., W. Yi, N. A. Draghi, H. M. O’Rourke, and L. K. Denzin. 2007. and/or initiation of an autoreactive immune response if an extracel- Expression patterns of H2-O in mouse B cells and dendritic cells correlate with lular source of peptides became available. Inflammation and tissue cell function. J. Immunol. 178: 1488–1497. 10. Liang, M. N., C. Beeson, K. Mason, and H. M. McConnell. 1995. Kinetics of the damage are likely to provide both the proteolytic environment to reactions between the invariant chain (85–99) peptide and proteins of the murine make peptides available and a reduced pH that assists peptide ex- class II MHC. Int. Immunol. 7: 1397–1404. The Journal of Immunology 5915

11. Sette, A., S. Southwood, J. Miller, and E. Appella. 1995. Binding of major his- tional consequences of altering a peptide MHC anchor residue. J. Immunol. 166: tocompatibility complex class II to the invariant chain-derived peptide, CLIP, is 3345–3354. regulated by allelic polymorphism in class II. J. Exp. Med. 181: 677–683. 42. Germain, R. N. 1995. Binding domain regulation of MHC class II molecule 12. Patil, N. S., A. Pashine, M. P. Belmares, W. Liu, B. Kaneshiro, J. Rabinowitz, assembly, trafficking, fate, and function. Semin. Immunol. 7: 361–372. H. McConnell, and E. D. Mellins. 2001. Rheumatoid arthritis (RA)-associated 43. Rabinowitz, J. D., M. Vrljic, P. M. Kasson, M. N. Liang, R. Busch, J. J. Boniface, HLA-DR alleles form less stable complexes with class II-associated invariant M. M. Davis, and H. M. McConnell. 1998. Formation of a highly peptide-recep- chain peptide than non-RA-associated HLA-DR alleles. J. Immunol. 167: tive state of class II MHC. Immunity 9: 699–709. 7157–7168. 44. Germain, R. N., and A. G. Rinker, Jr. 1993. Peptide binding inhibits protein 13. Hausmann, D. H., B. Yu, S. Hausmann, and K. W. Wucherpfennig. 1999. pH- aggregation of invariant-chain free class II dimers and promotes surface expres- dependent peptide binding properties of the type I diabetes-associated I-Ag7 mol- sion of occupied molecules. Nature 363: 725–728. ecule: rapid release of CLIP at an endosomal pH. J. Exp. Med. 189: 1723–1734. 45. Nelson, C. A., S. J. Petzold, and E. R. Unanue. 1994. Peptides determine the 14. Fremont, D. H., W. A. Hendrickson, P. Marrack, and J. Kappler. 1996. Structures lifespan of MHC class II molecules in the antigen-presenting cell. Nature 371: of an MHC class II molecule with covalently bound single peptides. Science 272: 250–252. 1001–1004. 46. Agrawal, B., E. Fraga, K. Kane, and B. Singh. 1994. Up-regulation of the MHC 15. Fremont, D. H., S. Dai, H. Chiang, F. Crawford, P. Marrack, and J. Kappler. class II molecules on B cells by peptide ligands. J. Immunol. 152: 965–975. 2002. Structural basis of cytochrome c presentation by IEk. J. Exp. Med. 195: 47. Zhong, G., F. Castellino, P. Romagnoli, and R. N. Germain. 1996. Evidence that 1043–1052. binding site occupancy is necessary and sufficient for effective major histocom- 16. Reay, P. A., R. M. Kantor, and M. M. Davis. 1994. Use of global amino acid patibility complex (MHC) class II transport through the secretory pathway rede- replacements to define the requirements for MHC binding and T cell recognition fines the primary function of class II-associated invariant chain peptides (CLIP). of moth cytochrome c (93–103). J. Immunol. 152: 3946–3957. J. Exp. Med. 184: 2061–2066. 17. Sali, A., L. Potterton, F. Yuan, H. van Vlijmen, and M. Karplus. 1995. Evaluation 48. Germain, R. N., and L. R. Hendrix. 1991. MHC class II structure, occupancy and of comparative protein modeling by MODELLER. Proteins 23: 318–326. surface expression determined by post-endoplasmic reticulum antigen binding. 18. DeLano, W. L. 2002. The PyMOL Molecular Graphics System. DeLano Scien- Nature 353: 134–139. tific, Palo Alto, CA. 49. Ceman, S., S. Wu, T. S. Jardetzky, and A. J. Sant. 1998. Alteration of a single 19. Kim, K. J., C. Kanellopoulos-Langevin, R. M. Merwin, D. H. Sachs, and hydrogen bond between class II molecules and peptide results in rapid degrada- R. Asofsky. 1979. Establishment and characterization of BALB/c lymphoma tion of class II molecules after invariant chain removal. J. Exp. Med. 188: lines with B cell properties. J. Immunol. 122: 549–554. 2139–2149. Downloaded from 20. Dang, L. H., L. L. Lien, B. Benacerraf, and K. L. Rock. 1993. A mutant antigen- 50. Guerra, C. B., R. Busch, R. C. Doebele, W. Liu, T. Sawada, W. W. Kwok, presenting cell defective in expresses class II MHC mole- M. D. Chang, and E. D. Mellins. 1998. Novel glycosylation of HLA-DR ␣ dis- cules with an altered conformation. J. Immunol. 150: 4206–4217. rupts antigen presentation without altering endosomal localization. J. Immunol. 21. Russell, H. I., I. A. York, K. L. Rock, and J. J. Monaco. 1999. Class II antigen 160: 4289–4297. processing defects in two H2d mouse cell lines are caused by point mutations in 51. Neefjes, J. J., V. Stollorz, P. J. Peters, H. J. Geuze, and H. L. Ploegh. 1990. The the H2-DMa gene. Eur. J. Immunol. 29: 905–911. biosynthetic pathway of MHC class II but not class I molecules intersects the 22. Germain, R. N., and H. Quill. 1986. Unexpected expression of a unique mixed- endocytic route. Cell 61: 171–183.

isotype class II MHC molecule by transfected L-cells. Nature 320: 72–75. 52. Kropshofer, H., S. O. Arndt, G. Moldenhauer, G. J. Hammerling, and A. B. Vogt. http://www.jimmunol.org/ 23. Nanda, N. K., and A. J. Sant. 2000. DM determines the cryptic and immuno- 1997. HLA-DM acts as a molecular chaperone and rescues empty HLA-DR mol- dominant fate of T cell epitopes. J. Exp. Med. 192: 781–788. ecules at lysosomal pH. Immunity 6: 293–302. 24. Ozato, K., N. M. Mayer, and D. H. Sachs. 1982. Monoclonal antibodies to mouse 53. Liang, M. N., C. Lee, Y. Xia, and H. M. McConnell. 1996. Molecular modeling major histocompatibility complex antigens. Transplantation 34: 113–120. and design of invariant chain peptides with altered dissociation kinetics from 25. Sant, A. J., L. R. Hendrix, J. E. Coligan, W. L. Maloy, and R. N. Germain. 1991. class II MHC. Biochemistry 35: 14734–14742. Defective intracellular transport as a common mechanism limiting expression of 54. Honey, K., K. Forbush, P. E. Jensen, and A. Y. Rudensky. 2004. Effect of de- inappropriately paired class II major histocompatibility complex ␣/␤ chains. creasing the affinity of the class II-associated invariant chain peptide on the MHC J. Exp. Med. 174: 799–808. class II peptide repertoire in the presence or absence of H-2M. J. Immunol. 172: 26. Doebele, R. C., R. Busch, H. M. Scott, A. Pashine, and E. D. Mellins. 2000. 4142–4150. Determination of the HLA-DM interaction site on HLA-DR molecules. Immunity 55. Gautam, A. M., M. Yang, P. J. Milburn, R. Baker, A. Bhatnagar, J. McCluskey, 13: 517–527. and T. Boston. 1997. Identification of residues in the class II-associated Ii peptide

27. Holm, S. 1979. A simple sequentially rejective multiple test procedure. Scand. (CLIP) region of invariant chain that affect efficiency of MHC class II-mediated by guest on October 2, 2021 J. Stat. 6: 65–70. antigen presentation in an allele-dependent manner. J. Immunol. 159: 2782–2788. 28. Asaithambi, N. S. 1995. Numerical Analysis: Theory and Practice. Saunders 56. Ridgway, W. M., M. Fasso, and C. G. Fathman. 1999. A new look at MHC and College Publishing, Fort Worth. autoimmune disease. Science 284: 749–751. 29. Rice, J. A. 1995. Mathematical Statistics and Data Analysis. Duxbury Press, 57. Nepom, G. T., and W. W. Kwok. 1998. Molecular basis for HLA-DQ associa- Belmont. tions with IDDM. Diabetes 47: 1177–1184. 30. McCulloch, C. E., and S. R. Searle. 2001. Generalized, Linear, and Mixed Mod- 58. Picca, C. C., J. Larkin, 3rd, A. Boesteanu, M. A. Lerman, A. L. Rankin, and els. John Wiley and Sons, New York. A. J. Caton. 2006. Role of TCR specificity in CD4ϩ CD25ϩ regulatory T-cell 31. Liang, K. Y., and S. L. Zeger. 1986. Longitudinal analysis using generalized selection. Immunol. Rev. 212: 74–85. linear models. Biometrika 73: 13–22. 59. Parak, W. J., S. Dannohl, M. George, M. K. Schuler, J. Schaumburger, 32. Rammensee, H. G., T. Friede, and S. Stevanovic. 1995. MHC ligands and peptide H. E. Gaub, O. Muller, and W. K. Aicher. 2000. Metabolic activation stimulates motifs: first listing. Immunogenetics 41: 178–228. acid production in synovial fibroblasts. J. Rheumatol. 27: 2312–2322. 33. Rammensee, H., J. Bachmann, N. P. Emmerich, O. A. Bachor, and S. Stevanovic. 60. Wilkins, R. J., and A. C. Hall. 1995. Control of matrix synthesis in isolated 1999. SYFPEITHI: database for MHC ligands and peptide motifs. Immunoge- bovine chondrocytes by extracellular and intracellular pH. J. Cell. Physiol. 164: netics 50: 213–219. 474–481. 34. Schild, H., U. Gru¨neberg, G. Pougialis, H. J. Wallny, W. Keilholz, S. Stevanovic, 61. Radu, C. G., A. Nijagal, J. McLaughlin, L. Wang, and O. N. Witte. 2005. Dif- and H. G. Rammensee. 1995. Natural ligand motifs of H-2E molecules are allele ferential proton sensitivity of related G protein-coupled receptors T cell death- specific and illustrate homology to HLA-DR molecules. Int. Immunol. 7: associated gene 8 and G2A expressed in immune cells. Proc. Natl. Acad. Sci. 1957–1965. USA 102: 1632–1637. 35. Doebele, R. C., A. Pashine, W. Liu, D. M. Zaller, M. Belmares, R. Busch, and 62. Ng, A. W., A. Bidani, and T. A. Heming. 2004. Innate host defense of the lung: E. D. Mellins. 2003. Point mutations in or near the antigen-binding groove of effects of lung-lining fluid pH. Lung 182: 297–317. HLA-DR3 implicate class II-associated invariant chain peptide affinity as a con- 63. Muczynski, K. A., S. K. Anderson, and D. Pious. 1998. Discoordinate surface straint on MHC class II polymorphism. J. Immunol. 170: 4683–4692. expression of IFN-␥-induced HLA class II proteins in nonprofessional antigen- 36. Alfonso, C., G. S. Williams, and L. Karlsson. 2003. H2-O influence on antigen presenting cells with absence of DM and class II colocalization. J. Immunol. 160: presentation in H2-E-expressing mice. Eur. J. Immunol. 33: 2014–2021. 3207–3216. 37. McFarland, B. J., and C. Beeson. 2002. Binding interactions between peptides 64. Douek, D. C., and D. M. Altmann. 1997. HLA-DO is an intracellular class II and proteins of the class II major histocompatibility complex. Med. Res. Rev. 22: molecule with distinctive thymic expression. Int. Immunol. 9: 355–364. 168–203. 65. Watanabe, N., Y. H. Wang, H. K. Lee, T. Ito, Y. H. Wang, W. Cao, and Y. J. Liu. 38. Dadaglio, G., C. A. Nelson, M. B. Deck, S. J. Petzold, and E. R. Unanue. 1997. 2005. Hassall’s corpuscles instruct dendritic cells to induce CD4ϩCD25ϩ regu- Characterization and quantitation of peptide-MHC complexes produced from hen latory T cells in human . Nature 436: 1181–1185. egg lysozyme using a monoclonal antibody. Immunity 6: 727–738. 66. Rohn, T. A., M. Boes, D. Wolters, S. Spindeldreher, B. Muller, H. Langen, 39. Mellins, E., B. Arp, D. Singh, B. Carreno, L. Smith, A. H. Johnson, and D. Pious. H. Ploegh, A. B. Vogt, and H. Kropshofer. 2004. Upregulation of the CLIP self 1990. Point mutations define positions in HLA-DR3 molecules that affect antigen peptide on mature dendritic cells antagonizes T helper type 1 polarization. Nat. presentation. Proc. Natl. Acad. Sci. USA 87: 4785–4789. Immunol. 5: 909–918. 40. Chang, S. T., J. J. Linderman, and D. E. Kirschner. 2005. Multiple mechanisms 67. Brown, D. R., K. Swier, N. H. Moskowitz, M. F. Naujokas, R. M. Locksley, and allow Mycobacterium tuberculosis to continuously inhibit MHC class II-medi- S. L. Reiner. 1997. T helper subset differentiation in the absence of invariant ated antigen presentation by macrophages. Proc. Natl. Acad. Sci. USA 102: chain. J. Exp. Med. 185: 31–41. 4530–4535. 68. Chaturvedi, P., R. Hengeveld, M. A. Zechel, E. Lee-Chan, and B. Singh. 2000. 41. Kersh, G. J., M. J. Miley, C. A. Nelson, A. Grakoui, S. Horvath, D. L. Doner- The functional role of class II-associated invariant chain peptide (CLIP) in its meyer, J. Kappler, P. M. Allen, and D. H. Fremont. 2001. Structural and func- ability to variably modulate immune responses. Int. Immunol. 12: 757–765.