Court Filings

A novel first primary anchor extends the MHC class II I-Ad binding motif to encompass nine amino acids

Description
A novel first primary anchor extends the MHC class II I-Ad binding motif to encompass nine amino acids
Categories
Published
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  International Immunology, Vol. 9, No. 8, pp. 1185–1193   ©   1997 Oxford University Press  A novel first primary anchor extends theMHC class II I-A d binding motif toencompass nine amino acids Kristian Bartnes, Francisco Leon, Jean Paul Briand 1 , Paul J. Travers 2 and Kristian Hannestad Department of Immunology, Institute of Medical Biology, University of Tromsø, School of Medicine, 9037Tromsø, Norway 1 Institut de Biologie Moleculaire et Cellulaire, UPR 9021 CNRS, Strasbourg, France 2 Imperial Cancer Research Fund Unit for Structural Molecular Biology, Department of Crystallography,Birkbeck College, London WC1E 7HX, UK Keywords  : autoimmunity, epitope, Ig, peptide, TCR, T lymphocyte AbstractThe MHC class II molecule I-A d has been reported to bind peptides containing a motif of sixconsecutive amino acids. We demonstrate that binding of the murine IgG2a b heavy chainallopeptide  γ  2a b 435–451 (Kabat numbering) to I-A d is strongly enhanced by a novel first primaryanchor (P1) three residues N-terminal to this hexamer. This is based on flow cytometricassessment of the I-A d binding capacity of  γ  2a b peptide analogues, their antigenicity for I-A d -restricted T cell clones and molecular modelling. The P1 pocket is broadly specific since aliphatic,aromatic, acidic, the basic histidine and small polar side chains all allowed good binding. Bycontrast, asparagine, arginine and glycine reduced the binding capacity 10-, 16- and  > 100-foldrespectively. Truncation or glycine substitution at P1 decreased antigenicity by a factor  > 1000.Nevertheless, I-A d -restricted T cells are not completely dependent on this anchor since highconcentrations of a peptide with glycine-substituted P1 elicited maximal responses. Additionalanchoring side chains are found at P4, P6 and P9. The autologous IgG2a a heavy chain sharesprominent epitopic residues with  γ  2a b 435–451 at P3, P5 and P8. However, the lysine of  γ  2a a at P9impairs binding to I-A d , which may explain why the  γ  2a b allopeptide-reactive T cells escapednegative selection. The data rationalize our observation (Bartnes, K. and Hannestad, K. 1997.  Eur.J. Immunol.  27:1124) that these T cells recognize a syngeneic B cell lymphoma, provided itspresentation of intrinsic  γ  2a a is enhanced by surface IgG2a a ligation.Introduction The MHC molecules are promiscuous peptide receptors. relative protection (8). Disease associations probably reflectthe role of MHC molecules in guiding T lymphocyte antigenThus, it has been estimated that an individual cell simultan-eously presents between 650 and 2000 different peptide recognition, since MHC polymorphism is most pronouncedat the peptide binding site, and susceptibility to rheumatoidsequences in association with a single class II variant, I-A d (1). Corresponding to the structural variability of the binding arthritis, type 1 diabetes mellitus and pemphigus vulgariscorrelates with allotypic class II sequences which influencesite, the spectrum of sequences that can be presented variesamong the different MHC allelic products. For both class I peptideselectivity(9–12).Based onthisstructuralinformation,peptides with high affinity for the respective disease-associ-and II molecules, allele-specific motifs have been definedwhich represent the features of primary sequence uniquely ated HLA class II variant have been found within candidateautoantigens(9,11,13).Thisindicatesthatelucidationofmotifsshared among peptides that bind well to a given MHC variant(2–6 and reviewed in 7). Alleles of the MHC influence, might aid identification of T cell epitopes involved in auto-immune, infectious and neoplastic diseases, and improvesometimes profoundly, the susceptibility to a range of auto-immune diseases, either by conferring an increased risk or our understanding of the molecular mechanisms involved in Correspondence to  : K. BartnesTransmitting editor: G. Ha¨mmerling  Received   6 December 1996,  accepted   30 April 1997  1186  A novel P1 anchor in the I-A d  motif  autoimmunity by comparing motifs of susceptibility associ- exposure to 1% paraformaldehyde (Sigma, St Louis, MO) for20 min atroom temperature asdescribed (27)priorto additionated, neutral and protective MHC variants.The murine class II variant I-A d is the restriction element for of antigen and T cells.T cells mediating herpes simplex virus-induced autoimmune Peptide binding assay  stromal keratitis (14) and confers protection against auto-immune diabetes mellitus (15), which is strongly associated RelativeI-A d bindingcapacitiesofvariouspeptidesweredeter-with a polymorphism at the class II  β  chain residue 57 in man mined essentially as described (28) by measuring their ability(16) as well as in the mouse (17). We report here an extended, to compete with biotinylated  γ  2a b 436–451 (B436 for short)nonameric motif for I-A d which encompasses a previously for I-A d binding. Various concentrations of competitor peptidereported hexameric motif (5) and includes a novel P1 anchor preincubated with 3  10 5 cells/well in V-bottom trays (Nunc,with broad specificity. Roskilde, Denmark) at 37°C. After 3 h, B436 was added (1.75 µ M) and incubation proceeded for 3 h in the presence of bothpeptides. After washing thrice the cells were resuspended in Methods R-phycoerythrin–streptavidin (PS) (Jackson Immunoresearch,WestGrove,PA)(5 µ g/ml)and furtherincubated for0.5hat4°C Synthetic peptides  followed bytwowashes.Theseintervalswereselected astheyThe single letter code for the amino acids is usedconferred the highest sensitivity of competition out of severalthroughout. Sequences are given in the figures and referredtested. All steps were performed in PBS with 0.1% BSA. Theto by numbering the peptide termini in accordance withfluorescence of 10 4 cells/sample was recorded on an Epicsthe corresponding IgG H chain residues [Kabat numberingProfile II flow cytometer equipped with a 488 nm argon laser(18)]. Most peptides were synthesized at Institut de Biologie(Coulter Electronics, Luton, UK). Background histograms,Moleculaire et Cellulaire, CNRS, Strasbourg, as describedobtained fromsamplesincubated withthecorresponding con-(19).  γ  2a b analogues 435–447(S446G), 435–447(K445A-centration of competitor, unbiotinylated  γ  2a b 436–451 insteadS446K), YFMYSGLG 4 ST and A 3 YA 2 LA 4 SA were purchasedofB436,and PS,weresubtractedusing theEpicsElitesoftwarefrom Chiron Mimotopes (Clayton, Australia). N-terminal(Coulter Electronics). Variation among duplicate samples wasbiotinylation of  γ  2a b 436–451 via a hexanoic acid linker  6%. The  γ  2a b 435–447 (L441A) analogue was included as awas achieved by reacting resin-bound free N-terminalreference competitor in all experiments.peptide with sulfosuccinimidyl-6-(biotinamido)hexanoate(Pierce, Rockford, IL) before cleavage from the resin, as  Molecular modelling  described (20). All peptides were purified by HPLC. PurityThe model of the I-A d molecule was built from that of thewas 95–99%.HLA-DR1(DRA/DRB1*0101)molecule(29,30)using theCOM-POSER suite of programs (31,32) contained within SYBYL Cells  (Tripos, St Louis, MO). In brief, the sequences of the I-A d We have previously described the T cell hybridomas A6 andchains (18) were built into the coordinates of the MHCB5 derived from CD4  , I-A d -restricted BALB/c clones (21)class II  α  and  β  chains of the HLA-DR1/influenza viruswhich recognize the IgG2a b heavy chain peptide  γ  2a b 435– haemagglutinin peptide structure (30). Insertions and dele-451 (19). The BALB/c B lymphoma A20 (22) (I-A d homozy-tions relative to the HLA-DR1 sequence were built initiallygous) and the MHC class II negative DBA mastocytoma P815using the loop searching algorithms within COMPOSER and(23) were purchased from ATCC (Rockville, MD).subjected to energy minimization. To model the  γ  2a b peptide,candidate residues for the P1 position were used to align the mAb  test sequence within the binding cleft. Alignments that gavemAb MK-D6 (anti-I-A d ) (24) and 14-4-4S (anti-I-E d,k,p,r ) (25)unacceptable contacts were eliminated and the remainingwere isolated from hybridoma culture supernatant on Proteinalignments were ranked by inspection of the peptide sideA–Sepharose columns (Pharmacia Fine Chemicals, Uppsala,chain interactions with the MHC molecule, particularly thoseSweden).Thecorresponding Bcellhybridomaswereobtainedwith the P1, P6 and P9 pockets, and by calculating interactionfrom ATCC. The I-A b/d -specific mAb 34-5-3S (26) was usedenergies for those pockets. Initially the main chain of thein the form of concentrated hybridoma culture supernatant.molecule and the C β  carbons were fixed. Following theFor the srcin and purification of mAb S43.10 specific for 4-convergence of the side refinement all of the atoms of thehydroxy-3-iodo-5-nitrophenyl acetyl (NIP), see references inmolecule were released and the optimization was performed(21). These mAb are all IgG2a from mouse.on the entire model. Analysis of T cell responses  Results T cell hybridomas were challenged with peptides in thepresence of A20 or irradiated (2000 rad) BALB/c spleen cells Establishment of an assay for peptide binding to I-A d  as antigen-presenting cells (APC) and IL-2 activity in 24h supernatant was determined by assessing [ 3 H]thymidine Binding of biotin-conjugated peptides to MHC class II molec-ules on B cell surfaces can be assessed quantitatively by([ 3 H]dThd) uptake of HT-2 cells, as reported (21). Functionalcompetition experiments were performed as described (19) flow cytometry (20,28). To obtain a probe for measuring I-A d binding, we biotinylated the IgG2a b heavy chain peptide  γ  2a b using A20 as APC and with the competitor peptide presentthroughout the assay. When indicated, APC were fixed by 436–451 known to be effectively presented in association with  A novel P1 anchor in the I-A d  motif   1187 I-A d (19). This peptide, B436, was antigenic for the B5 T required an 80-fold excess of  γ  2a b 436–451. This could reflectan increased affinity for I-A d conferred by the N-terminal biotincell hybridoma (Fig. 1), demonstrating that it productivelyassociated with I-A d . Dose-dependent binding of B436 to the moiety, just like the improved binding to I-E d seen with N-terminally modified peptide ligands (33). The fact that B436A20 (I-A d ) cells was readily detectable by flow cytometrywhile the signal with MHC class II negative P815 cells was was 40-fold more antigenic than the unbiotinylated peptide(Fig. 1) supports this interpretation.negligible (Fig. 2A). At the concentration of B436 used incompetition experiments (1.75  µ M), the fluorescence was To further probe the specificity, selected mAb were testedfor ability to inhibit binding of B436 to A20 cells (Fig. 2C).on average 8.2 times higher than in control samples withunbiotinylated  γ  2a b 436–451 (not shown). Presence of unbio- The I-A d -reactive mAb MK-D6 and 34-5-3S both exhibiteddose-dependent inhibition reaching 95 and 83% respectively.tinylated  γ  2a b 436–451 in high concentrations abrogatedthe signal (Fig. 2B), demonstrating that the two analogues No inhibition was seen with the isotype-matched control mAb14-4-4S (anti-I-E d ) or S43.10 (anti-NIP). Neither binding ofcompeted for the same binding site. Notably, 50% inhibitionB436 nor competition with mAb or unbiotinylated peptidesrequired internalization, since live and fixed cells performedsimilarly in this assay (data not shown).Having established that B436 specifically bound I-A d , weexamined whetheraflowcytometric assaybased ontheabilityofunlabeled peptidestocompetewithB436would beconsist-ent with functional experiments employing the  γ  2a b 435–451-reactive T cell clones. The  γ  1 a 435–451 peptide completelyinhibited presentation of the  γ  2a b allopeptide to Tcells (Fig. 3)as well as the fluorescent signal of the B436/I-A d interaction(see Fig. 6). Furthermore, from independent assessment ofpeptide binding to purified I-A d molecules it was predictedthatreplacing  γ  2a b residueL441withglycinewould negativelyaffect binding (5). Indeed, the L441G substitution conferred a92-fold loss of relative binding capacity in the flow cytometricassay (Fig. 4),  γ  2a b 435–447 (L441G) was 700-fold less anti- Fig. 1.  Antigenicity of peptides  γ  2a b 436–451  genicthan γ  2a b 436–451(Fig.1)anditdidnotdetectablyinhibit (FMYSKLRVQKSTWERG), γ  2a b 435–447(L441G)(YFMYSKGRVQKST) T cell responses against the  γ  2a b epitope (Fig. 3). Binding and B436. B5 T cell hybridomas (5  10 4  /well) were co-cultured with analysisinthepresenceofproteaseinhibitorsand Tcellstimu- A20 cells (10 5  /well) and various concentrations of peptide. IL-2- lation experiments with fixed APC in serum-free medium indi- responses were determined as described (21). Similar results wereobtained withaldehyde-fixed APCand withhybridomaA6(notshown).  cated that estimates of relative I-A d binding capacity and Fig. 2.  (A) Binding of B436 to murine cell lines. A20 and P815 cells were incubated with B436 in different concentrations, stained with 5 µ g/ml PS and analyzed by flow cytometry. Between steps, samples were washed in PBS with 0.1% BSA. Shown is the mean fluorescenceafter subtraction of background histograms obtained with the corresponding concentration of unbiotinylated  γ  2a b 436–451 plus PS (mean ofduplicates). (B and C) Specificity of binding of B436 for I-A d on A20 cells. A20 cells were preincubated without competitor or with unbiotinylated γ  2a b 436–451 or  γ  2a b 435–447(L441G) for 3 h at 37°C (B) or with mAb 34–5–3S, MK-D6, 14-4-4S or S43.10 for 0.5 h at 4°C (C). After additionof B436 (1.75  µ M), incubation proceeded for another 3 h at 37°C followed by washing, staining and flow cytometric analysis. Shown is themean fluorescence after subtraction of background histograms obtained with the corresponding concentrations of competitor, unbiotinylated γ  2a b 436–451 plus PS (mean of duplicates).  1188  A novel P1 anchor in the I-A d  motif  adjacent 439. Truncation at position 438 impaired binding tothe same extent as the glycine substitution. These observa-tions demonstrate that Y438 is a major anchor and that thepeptide sequence recognized by I-A d spans nine aminoacids. Hence, the I-A d motif is contained within a nonamer inwhich Y438 likely represents the first primary anchor. Thespecificity of the pocket accommodating P1 is strikinglydegenerate. Notably, the Y438G substituted analogue elicitedmaximal T cell IL-2 responses at high concentrations (Fig.5). Thus, although the anchor at this position contributespowerfully to antigenicity it is not absolutely required for Tcell recognition.A set of multiply substituted analogues was also analysedfor I-A d binding (Fig. 6). When collectively introduced on apoly(A) backbone in their srcinal relative positions, residuesY438, L441 and S446 conferred good binding, supportingtheir assignment as primary anchors. No binding was detect-able with the multiply glycine substituted YFMYSGLG 4 STalthough the amino acids in the four anchor positions met therequirements of the motif. The native K440 and K445 in the Fig. 3.  Capacity of  γ  2a b 435–447(L441G) and  γ  1 a 435–451(YFVYSKLNVQKSNWEAG) to compete with  γ  2a b 435–447 for  tripleglycinesubstituted analogueYFMYSKLG 3 KSTconferred presentation to T cell hybridoma B5. Fixed A20 cells (10 5  /well) at least 16-fold better binding. Since the side chains of K440 incubated overnight at 37°C with competitor at various conc entrations and K445 were not implicated in I-A d binding (Fig. 4), the prior to addition of T cells (5  10 4  /well) and  γ  2a b 435–447 at a final poor binding of YFMYSGLG 4 STlikely reflected conformational concentration of 0.25  µ M. (The L441G analogue could not be tested rearrangements, consistent with the high level of structural forTcellinhibitionathighconcentrationssinceitwasapartialagonist.)Cultures without competitor and without any peptide were included as  flexibility typical for poly(G) stretches (34). controls. IL-2 responses were determined as described (21). Individual lysine substitutions for V443 and S446 stronglyreduced binding capacity (Fig. 4). These replacements bothresulted in two basic residues becoming neighbours in thepeptide. To examine the possibility that the reduced bindingantigenicity were not biased by differential susceptibility toproteolysis (not shown). Collectively the results indicate that resulted from unfavourable (basic–basic) intrapeptide sidechain interactions, we tested two analogues with the doubletheflowcytometric assayisspecific and reliablyreflectsdiffer-ences in I-A d binding capacity of various peptide ligands. The substitutions A442–R443 and T442–K443 [position 442accepts alanine (19 and this study) and threonine (5)]. Bothadvantages of the flow cytometry-based method are relativesimplicity and universality, i.e. peptides can be tested regard- bound as poorly as the singly substituted K443 analogue(Figs 4 and 6). Similarly, the negative effect of K446 (Fig.less of antigenicity or TCRantagonism.4) was maintained with the additional substitution of the Role of individual peptide side chains in I-A d  binding  neighbouring K445 with alanine (Fig. 6) [which singly allowsgood binding (Fig. 4)]. In conclusion, the negative effect onRecognition of IgG2a b by the two I-A d -restricted T cell clonesA6 and B5 depended on residues  γ  2a b 440–446 (19) which binding could be predicted from single residue replacementsand occurred independentlyofbasic–basic intrapeptideinter-encompass a reported I-A d binding motif (5). To extendprevious knowledge regarding the contribution of individual actions.peptide side chains in I-A d and TCR engagement, we TCR discrimination between non-self   γ   2a  b  and self   γ   1 a  and  assessed the I-A d binding capacity and antigenicity of a set γ   2a  a  of peptides with single residue alterations within the 438–446sequence (Fig. 4). Residue 438 was included since molecular The T cell hybridomas A6 and B5 were derived from T h 1-typeclonesthatsuppressed theIgG2a b butnottheIgG2a a allotypemodelling (see below) implicated residue 438 as the P1anchor. At positions 438, 441, 443 and 446 certain single  in vivo   when adoptively transferred to (BALB/c  B10.D2)F 1 (H-2 d , IgC Ha/b ) mice (21). Nevertheless, hybridoma B5 alsoresidue substitutions diminished binding   10-fold (and T cellresponses  100-fold), indicating that the corresponding side recognized syngeneic B cells presenting  γ  2a a peptidesderived from their intrinsic IgG2a a (35). The basis for thischains served as anchors. Of these, S446 was identifiedpreviously (19). cross-reaction with self, which was evident also with thesynthetic  γ  2a a 435–451 peptide (19), was examined by single-The most important finding relates to position 438. Out of12 substitutions selected to represent a broad variety of residue substitutions for their effect on antigenicity (Fig. 4).Residues K440, R442, Q444 and K445 were necessary fornatural amino acids, only arginine, asparagine and glycineconferred a   10-fold decreased I-A d binding capacity. The the epitope defined by both clones since individual substitu-tions at these positions abolished T cell recognition withoutglycine-substituted analogue was outstanding, having a relat-ive binding capacity   1% and antigenicity   0.1% of that of affecting the strength of I-A d binding (Fig. 4). [Since theK445Aanaloguewasarelativelyweakcompetitorinfunctionalthe wild-type peptide. This effect was position-specific sinceglycine was tolerated at several other positions including the assays, K445 was previously erroneously implicated as an  A novel P1 anchor in the I-A d  motif   1189 Fig. 4.  Antigenicity and I-A d binding capacity of  γ  2a b 439–451 and single amino acid substituted analogues of  γ  2a b 435–447. IL-2 responsesof T cell hybridomas A6 (top panel) and B5 (middle panel) (5  10 4  /well) were determined as described (21), using irradiated BALB/c spleencells (5  10 5  /well) as APC.  γ  2a b 439–451 is represented by an asterisk at position 438 to indicate truncation. Relative stimulatory capacitieswere calculated by dividing the molar concentration of the parent peptide ( γ  2a b 435–447 for substituted analogues,  γ  2a b 438–451 for theanalogue truncated at 438) required to obtain 50% of maximal antigen-induced hybridoma IL-2 response by the molar concentration ofanalogue required to obtain 50% stimulation. For each analogue, the median values of three to 10 experiments are shown. The antigenicity of10 of these analogues was reported previously (19). Minor discrepancies from (19) reflect adjustment of the median as data from morerepetitions were incorporated. The analogues were tested for capacity to compete with B436 for binding to A20 cells (lower panel) asdescribed in Methods. Relative I-A d binding capacities were calculated by dividing the molar concentration of the parent peptide required toreduce the mean fluorescence in the presence of 1.75  µ M B436 with 50% by the molar concentration of each analogue required to obtain50% inhibition (for the weakest binders, calculation was based on 30% inhibition). Median values of two to six experiments are shown. anchor residue (19).] Of these epitopic residues, only Q444 (Fig. 4) which within the critical 438–446 sequence representsthe only allelic difference besides position 444.represents an allelic difference between the a and b allotypesof  γ  2a. Since replacement of Q444 ( γ  2a b ) with E ( γ  2a a ) Peptide 435–451 of self  γ  1 a , which is highly homologousto the corresponding  γ  2a peptides, was non-antigenic (19).reduced the antigenicity for B5 only 43-fold, these datapredicted that residues K440, R442, E444 and K445 of the Its I-A d binding capacity was similar to that of  γ  2a b 435– 447 (Fig. 6), in keeping with the fact that it contained allautologous  γ  2a a sequence collectively mimicked the epitopedefined by the TCR of hybridoma B5. Nevertheless, the critical anchors defined by single residue substitutions. Inthe determinant core [ γ  2a b 440–446 (19)]  γ  1 a differs fromantigenicity of  γ  2a a 435–451 was 6000-fold weaker than theallopeptide (19). This could mainly be attributed to a lower I-  γ  2a b only at position 442 which was a TCR contactresidue of the allopeptide. Thus, discrimination between theA d binding capacity (Fig. 6) conferred by  γ  2a a K446 in P9
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks