Specificity of two anti-class IHLA monoclonal antibodies that block class I recognition by the NKB1 killer cell inhibitory receptor

Copyright Q Munksgaard 1996 l h u e Antigens 1996: 48: 278-284 TISSUE ANTIGENS Printed in Denmark. All rights reserved ISSN oW1-281S Specificity of two anti-class I HLA monoclonal antibodies that block class I recognition by the NKB 1 killer cell inhibitory receptor J.E. Gumperz, J.C.M. Paterson, V. Litwin,N. Valiante,L.L. Lanier, P. Parham, A.-M. Little. Specificity of two anti-class I € L A monoclonal antibodies that block class I recognition by the NKB 1 killer cell inhibitory receptor. Tissue Antigens 1996: 48: 278-284.0 Munksgaard, 1996 Cytolysis by NK cells that possess the NKB 1 killer cell inhibitory receptor is inhibited by target cell expression of Bw4+ HLA-B molecules. The inhibitory effect can be prevented by addition of mAbs which block recognition of class I molecules by NKB 1. The epitopes recognized by two anti-class I mAbs, DX15 and DX16, which inhibit the interaction of NKB 1 with class I have been characterized. Binding of DX15 and DX16 to class I allotypes was inves. tigated by flow cytometric analysis of transfected cell lines which express just one HLA-A, B,or C allele, and by immunoprecipitationof class I molecules from HLA typed B-lymphoblastoid cell lines, followed by isoelectric focusing. The DX16 mAb recognizes class I allotypes which possess alanine at position 71 of the a,helix, and therefore has a specificity resembling that of the ME1 mAb but with broader specificity. Class I recognition by DX15 is affected by polymorphisms of the C-terminal part of the a,helix, and the N-terminal part of the cq! helix. DX15 thus appears to recognize a complex epitope near the end of the peptide binding groove which may be conformationally determined. Both antibodies are as effective as the anti-NKBl mAb (DX9) in preventing class I recognition by the NKBl receptor. DX16 also blocked recognition by a B*0702 allospecific CTL clone, whereas DX15 did not. Class I HLA molecules are best known for their interaction with the variable antigen receptors of cytolytic T lymphocytes (CTL).However, expression of class I molecules can also protect cells from lysis by natural killer (NK) cells (1,2). This is due to recognition of class I HLA molecules by a group of receptors, called killer cell inhibitory receptors (KIR) (3-5). Upon binding class I molecules of target cells, KIR generate an inhibitory signal which prevents cytolysis by NK cells. Target cells with impaired class I expression do not produce a strong inhibitory signal, and are consequently preferentially lysed by NK cells. KIR molecules belong to the immunoglobulin (Ig) superfamily (6-9), but are not closely related to the 278 J. E. Gumperzl, J. C. M., Paterson’,V. Litwin*, N. Valiantel, L. L. LanieP, P. Parham’ and A.-M. Little1i4 ‘Departments of Microbiology and Immunology and Structural Biology, Stanford University, Stanford, California, *Progenic Pharmaceuticals Inc., Tarrytown, New York, %epartment of Human Immunology, DNAX Research Institute of Molecular and Cellular Biology, Palo Alto, California, USA, ‘Anthony Nolan Research Institute, Royal Free Hospital, London, United Kingdom Key words: blocking antibody - class I receptor inhibition NK cell Received 12 July, accepted for publication 18 July 1996 - - antigen receptors of T cells or B cells. Two types of KIR molecules have been identified, the “p58” molecules, which have two extracellular Ig domains, and the “p70” molecules which have three Ig domains. The NKBl receptor is a p70 KIR molecule, which has been shown to have specificity for class I HLA-B allotypes that express the Bw4 serological epitope (9, 10). The class I HLA deficient 721.221 target cell line is efficiently lysed by NKBl’ NK clones, but transfectants of 72 1.221 which express Bw4’ HLA-B molecules show diminished lysis. Addition of certain monoclonal antibodies (mAbs) which bind the transfected class I allotype or the NIU3 I receptor, restores lysis to the level of the untransfected 7212 2 I cell line. Hence, binding of these antibodies Anti-HLA class I that block NKBl recognition to the NK cell receptor or the class I ligand appears to block the inhibitory recognition. To investigate further class I recognition by the NKBl KIR, we have characterized the HLA class I epitopes recognized by two mAbs, DX15 and DX16, which block recognition. Material and methods Antibodies The DX15 and DX16 hybridomas (both IgG1) were derived by immunizing BALB/c mice with the VL186-1 NK cell clone and fusing their splenocytes with the Sp2/0 myeloma cell line. Hybridoma supernatants were then selected for the ability to permit augmented lysis of certain class I transfected target cells by NK cell clones (4). The immunizing NK clone VL186-1 was derived from a peripheral blood donor with serological HLA type: A2, 3 B7, 58 Cw3, 7. The other mAbs used were as follows: W6/32 (IgG2a) which recognizes a monomorphic determinant of HLA-A, B and C molecules (1l), 4E (IgG2a) which recognizes a determinant expressed by HLA-B and C molecules and members of the HLA-A19 family (12), MA2.1 (IgG1) which recognizes an epitope of HLA-A2 and HLA-B17 molecules (13); 116.5.28 (IgG2a) which recognizes the Bw4 serological epitope (14); ME1 (IgG1) (15) which recognizes an epitope expressed by HLA-B7 and related molecules (16). F(ab’), fragments of DX15 and DX16 were prepared by digesting the mAbs with immobilized pepsin (lOmg/ml in 10 ml of 0.2M sodium citrate, 0.15M NaCl buffer, pH 3.5, with 2.5 ml immobilized pepsin for 2hr at 37OC) (Pierce Chemicals, Rockford, IL). Fc containing species were then removed by protein A affinity chromatography. The F(ab’), fragment preparation was determined to be pure by SDS-PAGE analysis. F(ab’), fragments of MA2.1 were prepared as described (17). Flow cytornetric analyses of DX15 and DX16 reactivity , Transfectants of the class I negative B-lymphoblastoid cell line L721.221, expressing single class I HLA-A, B, or C alleles (18) were generated as described (10). For flow cytometrk analysis the DX15 and DX16 antibodies were labeled with fluorescein by conjugation to FITC, as described (19). The W6/32 and 4E IgG antibodies were conjugated to fluorescein (FL) by reaction with fluorescein-X-NHS (Molecular Probes, Eugene, OR). Aliquots of class I transfected 721.221 cells (approximately 5- 1Ox 1O’cells) were incubated with each fluorescein-mAb conjugate for 30 minutes on ice, then washed in PBS containing 1pg/rnl propidium iodide, and resuspended in 0.5 ml PBS. The stained cells were analyzed immediately using a FACscan (Becton Dickinson, San Jose, CA) (19). Cells which stained brightly with propidium iodide were considered to be dead, and were excluded from the data analysis. Results were reproduced in at least three independent experiments. For the mAb blocking experiment, cells were pre-incubated with 2 pg of the appropriate mAb for30minutes, then washed and . stained with DX15-FITC or DX 16-FITC as described above. lrnmunoprecipifation and IEF analyses of DX15 and DX16 reactivity Class I molecules were immunoprecipitated from EBV transformed B cell lines as previously described (20, 21). 5pg purified mAb (W6/32, DX15 and DX16) were used in each immunoprecipitation, except for 4E immunoprecipitates in which 5p1 ascites were used. lpl rabbit anti-mouse serum was added as a secondary antibody to allow precipitation of the DX15 and DX16 (IgG1) antigen complexes with protein A positive S.aureus cells. In most experiments sequential precipitations were performed. The first precipitation was with either DX15 or DX16. This was followed by precipitation with the 4E mAb and then the W6/32 mAb, to identify class I HLA molecules which did not react with the DX15 and DX16 mAbs. NK cell and CTL clones NK cell clones were generated as described (22). The HLA-B*0702 alloreactive CTL clone STC1.15 was obtained from a peripheral blood donor with the serological HLA type A24, 32:B14, 44.Mononuclear cells were isolated by density gradient centrifugation using Ficoll-Hypaque (Phannacia LKB Biotechnology Inc. NJ) and cultured in the presence of irradiated JY B-LCLs (HLA-A*0201, B*0702) at a responder to stimulator ratio of 5: 1. Tissue culture medium was IMDM supplemented with 10% FCS, 2% human AB serum and lOOU/ml IL2. After two weeks of culture, cells were stained with Cy-chrome-conjugated anti-CD3 (Leu4) and PerCP-conjugated anti-CD8 (Becton Dickinson Immunocytometry Systems, Mountain View, CA). Cells expressing CD3 and CD8 were cloned at one cell per well using a single cell deposition system of the FACStar flow cytometer (Becton Dickinson, San Jose, CA). The cells were then cultured further with irradiated JY cells and allogeneic PBMCs, in culture medium containing IL-2. Clones reactive with HLA-B*0702 were identified by their ability to lyse the B*0702 expressing B-LCL targets: JY 279 Gumperz et al. Table 1 The class I HLA reactivity of the OX15 and OX16 rnAbs analyzed by flow cytornetry, or irnrnunoprecipitation followed by lsoelectric focusing. A 't' by irnrnunoprecipitationindicates that a band was clearly visible, '-' indicates no band was seen. By flow cytornetry a I-' Indicates staining of the transfectant wass20% of the signal obtained with the W6/32 antibody, a It' Indicates the signal was between 20% and 70% of the W6/32 staining, and 'tt'indicates it was270% of the W6/32 rnAb staining. OX16 rnAb: Class I Allele A1 A'0201 A'0211 A1 1 A24 A'2403 A.2501 A26 A29 A30 A31 A32 A34 A'3601 A'4301 A'6601 A'7401 8'0702 8'0703 8'0801 8'0802 8'1501 8'1502 8'1503 8'1508 8'1509 8'1513 8'1515 8'1516 8'1517 818 8'2702 8'2705 8'2708 835 8'3505 837 8'3801 8'3901 8'3905 842 8'4402 8'4601 8'4801 8'5101 8'5201 IEF OX15 rnAb: FACS overall IEF FACS Overall NO - - - NO - NO - - NO NO - NO NO- - NO NO NO NO NO NO - - - - - NO NO ND - tt NO - - - t t t NO - NO NO NO NO NO NO t t t - t t t tt - NO NO NO t t t - t - t t t t NO t - ND tt t NO NO t - NO - - NO NO NO - NO - - - NO NO t - t t t - t t NO t NO t t t NO t t tt t - NO NO t + t t ND ND t t + t t NO - - NO tt t t t - ND - 853 ND - t NO ND 6'5401 8'5501 8'5702 8'5801 8'6701 8'7301 8'7801 tt t ND t t tt t t t t NO t t NO t tt t ND tt t NO NO t - NO NO - + Nn 280 Nn - t tt t t+ t tt t tt t NO NO NO NO NO t t + t tt t t t tt t (HLA-A*0201, B*0702, Cw*O702), and 721.221 transfected with B*0702, and the lack of lysis of the untransfected L721.221 cell line and a HLA-B7 negative B-LCL LDB (HLA-A24,31 B44,61). Chromium release assays Cell killing by NK cell clones and the STC1.15 (anti-B*0702) CTL clone was measured using a standard 4-h 'lCr release assay. Target cells were labeled with 50pCi of N%''CrO, (Amersham Corp., Arlington Heights, IL) in 0.2 ml RPMI 1640 with 10% FCS at 37°C for 1.5 h. Assays were performed in duplicate or triplicate, and repeated at least twice for each effector and target cell combination. NK clone effector cells were used at an effector to target cell ratio of 6: 1, while that used for the CTL clone was 5: 1. Monoclonal antibodies were added to the wells where specified at a concentration of Ips! well. Specific lysis was calculated as percent lysis= lOOx(experimenta1 cpm - spontaneous cpm)/ total cpm -spontaneous cpm). The DX15 and DX16 mAbs emerged from an immunization and selection protocol designed to produce antibodies which block inhibitory class I recognition by NK cells (4). DX15 and DX16 were found to react with both NK cells and class I expressing target cells by flow cytometry. Immunoprecipitation from radiolabeled cell lysates with DX15 and DX16 revealed polypeptides of molecular weights 45 and 12 kDa upon SDS-PAGE analysis, the characteristic pattern for HLA class I molecules (data not shown). To confirm that DX15 and DX16 recognize class I HLA molecules, staining of the untransfected 721.221 cell line was compared with that of HLA class I transfectants by flow cytometry. Both DX15 and DX16 failed to bind untransfected 721.221 cells, but bound strongly to certain transfectants. including those expressing B*580 1 and B*0702. These data (not shown) are consistent with the serological HLA type of the cells used for immunization: A2, A3: B7, B58:Cw3, Cw7. t t tt Results t - NO NO NO NO NO overall ND t - Cw'0102 Cw'0304 Cw"0401 CW*O801 Cw'l503 FACS NO t ND IEF - - t - overall NO NO NO NO tt NO FACS NO NO - tt - IEF t NO NO - Class IAllele t t - OX15 rnAb: tt t - OX16 mAb: t tt NO NO ~~ t tt tt Table 1 (continued) - t + Anti-HLA class I that block NKBl recognition Table 2 Blocking of DX15 and DX16 mAb binding to class I molecules by anti-class I mAbs. Class i HLA transfectants were incubated with a saturating concentration of the blocking mAb, then washed and incubated with a saturating concentration of the detection antibody: DX15, DX16, or goat anti-mouse IgG (GaM IgG) FlTC conjugate. Binding of the second step reagent was measured by flow cytometry. Results for DX15 and DX16 are expressed as a percentage of the staining in the absence of blocking antibodies. The GaM control staining shows the labeling of the transfectants by the primary (blocking) antibodies, and is expressed as a percentage of the staining obtalned with the monornorphic W6M2 anti-class I mAb. Detection mAb Class I HLA transfectant DX15-FiTC GaM IgG-FITC (control) 8'2705I721.221 B'5801I721.221 8'27051121,221 8'58011121.221 8'27051721.221 8'58011121.221 .. Blocking mAb 801#)1BO(wL MA21 ME11 100.7 38.3 134.8 104.7 4.5 97.2 26.0 83.6 133.5 103.4 88.0 0.0 ~~ DX16-FITC a 116.5.28 4E (Bwb) (Bw4) BlKnB1513 (Bw6) (Bw4) B21M)BMZB2705 (Bw6) (Bw4) (Bw4) BJ905BylOl (Bw6) (BW4) huuhnt ~ 77.7 55.7 138.5 103.4 101.9 109.6 70.1 60.5 15.0 12.9 117.4 90.2 Certain 721.221 class I transfectants were not stained by DX15 or DX16, indicating the mAbs recognize polymorphic epitopes of class I molecules. To investigate the class I determinants recognized by DX15 and DX16, their reactivity with a panel of class I allotypes was analyzed by flow cytometry of class I transfectants, and precipitation of class I molecules from HLA-typed B cell lines, followed by isoelectric focusing (IEF).(See table 1). To compensate for varying amounts of class I expression by the transfectants, the signal obtained using DX15 or DX16 in flow cytometric analysis was normalized by comparison with that of the monomorphic W6132 mAb. DX16 has a clear-cut specificity for class I allotypes possessing alanine at position 71 of the heavy chain. With the exception of B*4601, HLA-B and C allotypes that possess alanine 7 1 were immunoprecipitated with DX16, and/or had at least 70% of the W6/32 signal by flow cytometry (table 1). In contrast, class I allotypes lacking alanine 71 were not precipitated by DX16, andlor showed less than 20% of the W6132 signal by flow cytometry. Further substantiating the specificity, DX16 binds B*0702 which has alanine at position 7 1, but fails to precipitate B*0703, which only differs from B*0702 at residues 69-71. That B*4601, which has alanine 71, did not react with DX16 in flow cytometry or immunoprecipitation could be due to the unusual juxtaposition of residues characteristic of HLA-C molecules next to those common to HLA-B molecules, or the peculiarly restricted peptide specificity of B*4601. In contrast to DX16, the class I HLA specificity of DX15 is complex. Class I allotypes could not be divided easily into positive and negative groups Figure I. Results of a representative flow cytometric analysis of DX 15 staining of matched sets of HLA-B72 1.221 class I transfectants. The transfected HLA-B allotypes are identical except at the Bw4/Bw6 epitope. The specific signal for the DX15 mAb was divided by that for the WQ32 mAb for each transfectant, to normalize for differences in class I expression level. Similar results were obtained in three other independent experiments. based on DX15 reactivity. DX15 staining of transfectants ranged from background values to over 100% of the Wa32 staining. Some allotypes, including most of the HLA-A antigens tested, had no reactivity at all with DX15 (table 1). DX15 recognition of most HLA-B and HLA-C allotypes was usually moderate or good, but for some was poor or negative (table 1). Bw4+ HLA-B antigens and HLA-C allotypes which have asparagine or aspartate at position 77 of the heavy chain, generally showed the strongest binding by DX15. Moreover, DX15 bound 2-3 times better to five Bw4' HLA-B allotypes than to their otherwise identical Bw6' HLA-B counterparts, suggesting an important role for residues 77-83 which determine the Bw4 and Bw6 specificities in DX 15 recognition (figure 1). Most HLA-A allotypes are not recognized by DX15, despite having asparagine or aspartate at position 77, indicating these residues are not sufficient to determine DX15 binding. The only HLA-A allotypes which were recognized by DX15 are members of the HLA-Al9 family, which differ from other HLA-A molecules in their substitution of arginine for histidine at position 151 of the a 2 domain, a feature shared by all HLA-B and C molecules. However, expression of arginine at position 151 cannot completely account for the specificity of DX15, since some HLA-B molecules are not recognized (table 1). Thus, the epitope recognized by DX15 is not easily correlated with a single amino acid or a linear sequence of residues in the class I structure. DX15 and DX16 specificity was further investigated by assessing the ability of other anti-HLA class I antibodies to block DX 15 and DX 16 binding to class I molecules. Antibodies tested for blocking 281 Gumperz et al. 60 T "T-- ' 35 .9 4 25.. m 15 .. 5 .. ac I -5 l n i m m*b b M nm/.ni n0mu.m smml.ui awml.ni owmini mmAb rDXllmAb tDXl6mAb +YEl& tKN.lrnAb Target Cell Figure 2. a) A representative cytotoxicity experiment using the NKB l+clone HY640.4 with untransfected and HLA-B*5801 transfected 721.22 1 target cells. Similar results were obtained using 5 other NKB 1+ NK clones and were repeated in an independent experiment. b) A cytotoxicity experiment showing anti-class I mAb blocking of HLA-B *0702 recognition by the STC1.15 CTL clone. Similar results were obtained in an independent experiment. were: ME1, 116.5.28 (anti-Bw4), 4E, and MA2.1 (12-17). The ME1 mAb recognizes an epitope determined by residues alanine 69 glutamine or lysine, 70, and alanine 71 of the heavy chain (15, 16); the anti-Bw4 antibody recognizes molecules with leucine 82 (Gumperz et al., submitted); 4E binds class I molecules with glutamine 143 and arginine at position 151 (12); and the epitope recognized by MA2.1 includes glycine 62 (24-26). DX16 binding was strongly inhibited by pre-binding of ME1 and MA2.1, and slightly affected by 4E and the anti-Bw4 mAb (table 2). Efficient blocking by the ME1 and MA2.1 mAbs agrees with DX16 recognizing alanine 7 1, since these mAbs recognize residues including and in close proximity to position 71 of the a,helix. Slight blocking by the anti-Bw4 and 4E antibodies could be due to steric hindrance of DX16 binding. Surprisingly, DX15 binding was only reduced by pre-incubation with the 4E mAb: pre-binding of antibodies recognizing epitopes within the a, helix had little or no effect (table 2). This suggests the DX15 antibody does not contact much of the a,helix, despite the effect of polymorphisms within the Bw4Bw6 epitope on DX15 reactivity. Analyses of 282 epitopes recognized by several other class I mAbs has similarly suggested they may be influenced by changes which are not directly part of the binding site (27-29). Thus the epitope recognized by DX15 is probably a complex determinant influenced by residues from both the a, and os, helices. The ability of the anti-class I mAbs, DX15, DX16, ME1, and MA2.1 to block class I-mediated inhibition of lysis by NKB1' NK clones was investigated in cytotoxicity assays. Compared to class I deficient 721.221 target cells, lysis of B*5801 transfected 721.221 target cells by NKB1' NK clones was reduced, but in the presence of the anti-NKB1 mAb DX9 lysis of both targets was equivalent (4). The DX9 anti-NU31 mAb was therefore included in the cytotoxicity assays as a control. In the presence of either DX15 or DX16, lysis of B*5801 transfected cells was restored to the level .of the untransfected cell line (Fig. 2a). The augmentation of target cell lysis in the presence of the DX15 and DX16 anti-class I mAbs was identical to that obtained with the DX9 anti-NKB 1 antibody, indicating all three antibodies are equally effective in blocking the receptor ligand interaction (Fig. 2a). The ME1 mAb does not bind the B*5801 target cell and therefore does not prevent inhibition with this allotype. Cytotoxicity experiments testing inhibition mediated by B*2705, an allotype recognized by ME1, demonstrated that ME1 blocks comparably to the DX9 mAb (data not shown). MA2.1 blocked B*5801 mediated inhibition of cytotoxicity, but was consistently less effective than the DX9 mAb (Fig. 2a). The increased lysis in the presence of the anti-class I mAbs could not be attributed to antibody dependent cell-mediated cytotoxicity (ADCC), because the experiments with DX15, DX16, and MA2.1 were performed with F(ab'), fragments, and ME1 is of the IgGl isotype, which does not permit ADCC by human NK cells (30). Moreover the NK cell clones tested were derived from donors with different HLA types (data not shown), suggesting the increased lysis was due to binding of the antibodies to class I antigens of the target cell, rather than the effector cell. The ability of DX15 and DX16 to inhibit recognition of a class I HLA allotype by CTL was addressed using the STC1.15 clone, which is allospecific for target cells expressing HLA-B*0702, an allotype recognized well by both DX15 and DX16 (tablel). The ME1 mAb which binds B*0702, and the MA2.1 mAb which does not; were included as controls. The untransfected 72 1.22 1 cell line showed less than 5% specific lysis by the STC1.15 clone, and this was not affected by addition of the anti-class I mAbs (data not shown). In contrast, the B*0702/721.221 transfectant was lysed, and killing Anti-HLA class I that block NKBl recognition was reduced by addition of either DX16 or ME1 (Fig. 2b). Addition of the DX15 mAb did not reduce killing, and the MA2.1 mAb which does not bind B*0702 likewise had no effect. Thus, binding of DX16 to HLA-B*0702 blocked recognition by a T cell clone, whereas DX15 binding did not. Discussion From this analysis the DX16 mAb appears to recognize an epitope of the a,helix, determined by the presence of alanine at position 7 1. This specificity is related to that of the ME1 mAb, which binds class I molecules that possess an alanine, lysine, alanine (AKA), or alanine, glutamine, alanine (AQA), at residues 69, 70, and 71 (15, 16). However, DX16 differs from ME1 in that it recognizes B*5801 and all of the HLA-C allotypes tested. Consistent with a binding site for DX16 on the a;helix, is the ability of other class I mAbs which recognize epitopes of the a,helix to block its binding. DX15 binding is also affected by polymorphisms within the C-terminal part of the a,helix, since the allotypes best recognized are HLA-B and C molecules with asparagine or aspartate at position 77. Furthermore in analyzing the binding to pairs of HLA-B allotypes that only differ at residues 77-83 forming the Bw4 and Bw6 epitopes, DX15 binds more strongly to the Bw4’ allotypes. However, binding of DX15 to class I molecules is only blocked by the 4E mAb, which recognizes an epitope focused on residue 151 of the a, helix, and not by antibodies which recognize epitopes of the a,helix. Although influenced by residues of the a,helix, DX15 may not directly contact this part of the class I molecule. The crystal structures of B*3501 and B*5301 (two HLA-B allotypes differing only at the Bw4/Bw6 epitope) complexed with specific peptides may help interpretation of the class I reactivity of DX15 (31, 32). Unexpected from their amino acid sequences is that the conformation of the N-terminal part of the a, helix of the Bw6’ B*3501 differs from that of the Bw4’ B*5301 (31, 32). The difference appears to be due to contact of residues in the a, helix with the C-terminal residue of the bound peptide, which mainly affects the shape of the a,helix from positions 141-152 (a region that includes the 4E mAb epitope). Thus Bw4/Bw6 differences in the a,helix may indirectly affect the conformation of the helix, possibly‘ in part by altering the peptide binding properties of the class I molecule (Barber et al., in preparation). The epitope recognized by DX15 might therefore be a part of the N-terminal region of the a, helix affected by polymorphism in the C-terminal end of the a, helix. Alterna- tively, the DX15 binding site could comprise features from both the C-terminal part of the a,helix and the N-terminal part of the a,helix. Our results suggest DX15 and DX16 interfere directly with class I binding by NKB 1, probably by binding a part of the class I molecule that overlaps with that targeted by the receptor. The MA2.1 mAb, which does not block NKB1 recognition as effectively as DX15 and DX16, may bind an epitope which only partially overlaps with that of the NKB 1 rece tor. As the combining site of an antibody is 20 by 30 A, binding of an antibody to the a,helix of the class I molecule could potentially obscure at least two thirds of the length of the helix. Since MA2.1 recognizes an epitope on the N-terminal portion of the a,helix, DX16 recognizes a determinant of the middle of the a,helix, and DX15 may bind near the C-terminal end of the peptide binding groove, differences in the degree of blocking between MA2.1 and DX15 and DX16 are consistent with observations that polymorphisms within the C-terminal part of the a,helix affect NKBl recognition of class I molecules (10). DX16 also blocked B*0702 recognition by an allospecific CTL clone, whereas DX15 did not. Thus the NKBl receptor might approach class I molecules from the C-terminal end of the peptide binding groove or from the “side”, whereas the receptor of the T cell clone probably approaches the class I molecule from the “top”. Bp - Acknowledgments We thank Dr. Keith Gelsthorpe for anti-Bw4 monoclonal antibody 116.5.28. This research was supported by NIH grants A117892 and AI22039 to Peter Parham. N. Valiante is supported by a postdoctoral fellowship from the Cancer Research Institute. DNAX Research Institute is supported by Schering-Plough Corporation. References Storkus W, Alexander J. Payne JA, CresswellP,Dawson JR. Reversal of natural killing susceptibility in target cells expressing hansfected class I genes. Pmc Nut1 Acad Sci U S A 1989: 86: 2361-2364. Shimizu Y,DeMars R. Demonstration by class I gene transfer that reduced susceptibility of human cells to natural killer cell-mediated lysis in inversely correlated with HLA class I antigen expression. Eur J Immunol 1989: 19: 447451. Moretta A, Vitale M, Bottino C et al. P58 molecules as putative receptors of major histocompatibility complex (MHC) class I molecules in human natural killer (NK) cells. Anti-pS8 antibodies reconstitute lysis of MHC class Iprotected cells in NK clones displaying different specificities. J E x p Med 1993: 178: 597-604. 283 Gumperz et al. 4. Litwin V, Gumperz J, Parham P, Phillips JH, Lanier LL. NKB 1: a natural killer cell receptor involved in the recognition of polymorphic HLA-B molecules. J Exp Med 1994: 180: 537-543. 5. Wagtmann N, Rajagopalan S , Winter C, Peruzzi M, Long EO. Killer cell inhibitory receptors specific for HLA-C and HLA-B identifiedby direct binding and by functional transfer. Immunity 1995: 3: 801-809. 6. Burshtyn DN, Scharenberg AM, Wagtmann N et al. Recruitment of tyrosine phosphatase HCP by the killer cell inhibitory receptor. Immunity 1996: 4: 77-85. 7. Wagtmann N, Biassoni R, Cantoni C et al. Molecular clones of the p58 NK cell receptor reveal immunoglobulinrelated molecules with diversity in both the extra- and intracellular domains. Immunity 1995: 2: 439-449. 8. Colonna M, Samaridis J. Cloning of immunoglobulin-superfamily members associated with HLA-C and HLA-B recognition by human natural killer cells. Science 1995: 268: 405-408. 9. D’Andre A, Chang C, Franz-Bacon K, McClanahan r, Phillips JH,Lanier L. Molecular cloning of NKB 1: a natural killer cell receptor for HLA-B allotypes. J Immunol 1995: 155: 23062310. 10. Gumperz JE, Litwin V, Phillips JH, Lanier LL, Parham P. The Bw4 public epitope of HLA-B molecules confers reactivity with natural killer cell clones that express NKB 1, a putative HLA receptor. J Exp Med 1995: 181: 1133-1 144. 11. Barnstable CJ, Bodmer WF, Brown G et al. Production of monoclonal antibodies to group A erythrocytes, HLA and other human cell surface antigens - new tools for genetic analysis. Cell 1978: 14: 9-20. 12. Trapani JA, Mizuno S , Kang SH, Yang SY, Dupont B. Molecular mapping of a new public HLA class I epitope shared by all HLA-B and HLA-C antigens and defined by a monoclonal antibody. Immunogenetics 1989: 29: 25-32. 13. McMichael A. Parham P, Rust N, Brodsky F. A monoclonal antibody that recognizes an antigenic determinant shared by HLA-A2 and B17. Hum ImmunolI980: 1: 121-129. 14. Arnett KL, Adam EJ,Gumperz JE, et al. Expression of an unusual Bw4 epitope by a subtype of HLA-B8 (B*0802). Tissue Antigens 1995: 46: 316-321. 15. Ellis SA, Taylor C, Hildreth JEK,McMichael AJ.An HLA class I specific monoclonal antibody that fails to bind to all HLA-A antigens. Hum Immunol 1985: 13: 13-19. 16. Parham P,Antonelli P, Herzenberg LA, Kipps TJ, Fuller A, Ward FE. Further studies on the epitopes of HLA-37 defined by murine monoclonal antibodies. Hum Immrinol 1986: 15: 44-67. 17. Parham P, Androlewicz MJ, Brodsky FM,Holmes NJ, Ways J. Monoclonal antibodies: purification, fragmentation and application to structural and functional studies of class I MHC antigens. J Immunol Methodr 1982: 53: 133-173. 18. Shimizu Y, Geraghty DE, Koller BH, Om HT, D t Mars R. Transfer and expression of three cloned human non-HLAA,B,C class I major histocompatibility complex genes in mutant lymphoblastoid cells. Proc Natl Acad Sci U S A 1988: 85: 227-231. 19. Lanier LL, Recktenwald DJ. Multicolor immunofluorescence and flow cytometry. Methods (Orlando) 1991: 2: 192-199. 20. Yang SY. A standardized method for detection of HLA-A 284 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. and HLA-B alleles by one-dimensional isoelectric focusing (IEF) gel electrophoresis. In: Dupont B. ed. fmmunobiology of HLA. Vol. I. New York: Springer-Verlag, 332-335. Little A-M, Parham P. HLA class I gene and protein sequence polymorphisms. In: Dyer P. Middleton D. ed. Histocompatibility testing: a practical approach. The Practical Approach Series. Oxford: IRL Press: 159-190. Yssel H, Devries JE, Koken M, van Blitterswijk W, Spits H. Serum-free medium for the generation and the propagation of functional human cytotoxic and helper T cell clones. J Immunol Methods 1984: 72: 219-227. Barber LD. Percival L, Valiante NM et al. The inter-locus recombinant HLA-B*4601 has high selectivity in peptide binding and functions characteristic of HLA-C. J Exp Med 1996: 184.735-740. Ways JP, Coppin HL, Parham P. The complete primary structure of HLA-Bw58. J Biol Chem 1985: 260: 1192411933. Ways JP, Rothbard JB, Parham P. Amino acid residues 56 to 69 of HLA-A2 specify an antigenic determinant shared by HLA-A2 and HLA-B17. J Immunol 1986: 137: 217222. Santos-Aguado J, Barbosa JA, Biro PA, Strorninger JL. Molecular characterization of serological recognition sites in the human HLA-A2 molecule. J Immunol 1988: 141: 28 11-28 18. Santos-Aguado J, Biro PA, Fuhrmann U, Strominger JL, Barbosa JA. Amino acid sequences in the a, domain and not glycosylation are important in HLA-A2/P2-microglobulin association and cell surface expression. Mol Cell Biol 1987: 7: 982-990. Ways JP, Lawlor DA, Wan AM, Parham P. A transposable epitope of HLA-B7, B40 molecules. Immunogenetics 1987: 25: 323-328. McCutcheon JA, Smith KD, Valenzuela A, Aalbers K, Lutz CT. HLA-B*0702 antibody epitopes are affected indirectly by distant antigen residues. Hum Immiinol 1993: 36: 69-75. Trinichieri G, Valiante N. Receptors for the Fc fragment of IgG on natural killer cells. Nut Immun 1993: 12: 218-234. Smith KJ, Reid SW, Harlos K et al. Bound water structure and polymorphic amino acids act together to allow the binding of different peptides to MHC class I HLA-B53. Immunity 1996: 4: 215-228. Smith KJ, Reid SW. Stuart DI, McMichael AJ,Jones EY, Bell JI. An altered position of the Q helix in revealed by the crystal structure of HLA-B*3501. Immunie 1996: 4: 203-2 13. Amit AG. Mariuzza RA, Phillips SEV,Poljak RJ, Three-dimensional structure of an antigen-antibody complex at 2.8A resolution. Science 1986: 233: 747-753. Address: Peter Parham Departments of Structural Biology and Microbiology and Immunology Stanford University School of Medicine Fairchild Building Stanford, CA 94305-5400 USA Fax +I 415-723-8464