Mutagenesis Around Residue 176 on HLA-B*0702 Characterizes Multiple Distinct Epitopes for Anti-HLA Antibodies Jane A. McCutcheon and Charles T. Lutz

ABSTRACT: Preformed antibodies against HLAA,B,C molecules cause hyperacute rejection of transplanted allogeneic tissues. To understand better the molecular basis of hyperacute rejection, narrowly reactive anti-HLA-B*0702 monoclonal antibodies have been studied. Previous epitope mapping studies of these monoclonal antibodies by mutating B*0702 have conflicted with antibody-blocking studies. To resolve these discrepancies, we mutated B*0702 residues around the antigenically important residue 176, and measured antiB*0702 antibody binding. Antibody MB40.2 binding is abrogated by mutations at residues 169, 180, and 182, close to residue 176 in the primary structure. However, MB40.2 binding is not affected by 12 other B*0702 mutations close to residue 176 in the tertiary structure. This suggests that MB40.2 may recognize a sequential B*0702 epitope including residues between positions 169-182.

Antibody BB7.1 binding requires B*0702 c~2-domain residues 166 and 169. Competition for B*0702 residue 169 explains why MB40.2 and BB7.1 crossblock. Because BB7.1 binding also requires B*0702 ~l-domain residues, BB7.1 may contact both c~-helices, straddling the B*0702 peptide-binding groove. Previous results showed that both MB40.2 and MB40.3 binding require B*0702 residues 176/178. However, MB40.3 binding is not affected by any of 15 other mutations near residue 176. This suggests that MB40.3 does not contact residues 176/178; rather, residues 176/178 appear to affect MB40.3 binding by subtly influencing B*0702 conformation. Thus, monoclonal antibodies influenced by a defined B*0702 region around residue 176 appear to recognize three different types of epitopes. This suggests that human alloantibodies also recognize diverse types of HLA epitopes. Human Immunology 35, 125-131 (1992)

ABBREVIATIONS B7/27 intradomain recombinants between B*0702 and B'2705

mAb

monoclonal antibody

INTRODUCTION H L A class I transplantation antigens consist of a highly polymorphic heavy chain noncovalently coupled to the invariant fl2-microglobulin light chain [1]. The m e m brane proximal H L A heavy-chain a3 domain and /32microglobulin fold into immunoglobulinlike structures that support the polymorphic o~1 and a2 domains. The O~ 1 and o~2 domains fold into ol-helices and fl-strands that

From the Departments of Pathology and Microbiology, University of Iowa College of Medicine, Iowa City, Iowa, USA. Dr. J.A. McCutcheon's current address is the Department of Cell Biology, Fairchild Center, Stanford University Medical Center, Stanford, CA 94305-5400, USA. Address reprint requests to Dr. C.T. Lutz, Department of Pathology, University of lowa, Iowa City, IA 52242, USA. Received May 29, 1992; accepted August 26, 1992. Human Immunology 35, 125-131 (1992) © American Society for Histocompatibility and Immunogenetics. 1992

form the walls and floor of the peptide binding groove [1-7]. Class I H L A molecules are found on nearly all nucleated cells. Anti-HLA alloantibodies mediate hyperacute rejection of solid-organ transplants [8] and destruction of transfused platelets [9]. In addition, antibodies against class I molecules have been implicated in autoimmune diseases, such as the HLA-B27-associated spondyloarthropathies [10]. Thus a considerable effort has been directed toward understanding how antibodies recognize class I molecules [ 1 0 - 2 4 ] . Because human alloantibodies and many rodent monoclonal antibodies (mAbs) share H L A allele specificity and binding sites [18, 22, 23], rodent anti-HLA mAbs can be used to study clinically relevant H L A epitopes. 125 0198-8859/92/$5.00

126

Using well-characterized mAbs that react with a limited number of HLA-B alleles, investigators have studied B*0702 as a model transplantation antigen. In examining narrowly reactive anti-B*0702 mAbs, two discrepancies have been noted. First, mAbs MB40.2 and BB 7.1 crossblock [ 18], implying that their epitopes overlap. However, intradomain recombinants between B*0702 and B'2705 (B7/B27) showed that MB40.2 binding requires B*0702 residues 177-180, at the Cterminal end of the ~2 domain, whereas BB7.1 binding requires B*0702 residues 63-70, in the middle of the oel domain oe-helix [21]. Thus, in contrast to mAbblocking studies, these data map the MB40.2 and BB7.1 epitopes to distinct sites. Second, MB40.2 and MB40.3 bind simultaneously to B*0702 [12, 18], indicating nonoverlapping epitopes. However, B7/B27 recombinants showed that both mAbs require B*0702 residues 177-180 [21]. Furthermore, the binding of both mAbs is abrogated by a double substitution at B*0702 residues 176/178 [ 19]. Thus, in contrast to mAb-blocking studies, these data map the MB40.2 and MB40.3 epitopes to the same site. Here we examine these conflicting data by mutating B*0702 near residue 176. We further characterize the MB40.2, BB7.1, and MB40.3 epitopes, and suggest a resolution for the conflicting data about MB40.2 and BB7.1 epitopes. MATERIALS A N D M E T H O D S

Cell lines. Hybridoma cells secreting isotype control mAbs, GAP A3 [13], CRll-351 [15], R17.217 [25], and MA2.1, and B*0702-binding mAbs, L368 [11]; MB40.5, W6/32, BB7.7, BBM1 (listed by Ways and Parham [26]); BB7.6, BB7.1, MB40.2, MB40.3 (listed by Parham [16]); and SFR8-B6 [14] were grown in a 50 : 50 mixture of DMEM-F12 (Gibco) 10% iron-supplemented bovine calf serum (Hyclone). mAbs were collected as supernatants. Ascites of B*0702-binding mAbs, 4E and F4/326 [17], were provided by Dr. S.Y. Yang. 721.221 human B-lymphoblasts were grown in RPMI 1640 (Gibco) with 10% iron-supplemented bovine calf serum. 721.221 cells contain no detectable HLA-A,B,C mRNA or cell surface protein [27], allowing transfected B*0702 variants to be analyzed in the absence of endogenous HLA class I molecules.

Mutagenesis. Residues were selected for mutagenesis based on A*0201 s-carbon coordinates (Brookhaven National Database), using the molecular modeling program Sybyl version 5.1 (Tripos) on a Silcon Graphics Iris Terminal. The B*0702 genomic sequence provided

J.A. McCutcheon and C.T. Lutz

by Dr. J. Barbosa (personal communication) was used to derive all mutagenic and sequencing oligonucleotides. Variants were generated as described [28], using singlestranded, uridine-containing DNA from Escherichia coli Bw313 [29] (0.04 mg), 2 pmol mutagenic primer, and 2 units of modified T7 DNA polymerase (Sequenase 2.0, United States Biochemical). Alternatively, variants were generated by a polymerase chain reaction protocol according to Ho et al. [30] with minor modifications. For both mutagenesis methods, we sequenced all nucleotides encoding mature HLA protein that were subjected to in vitro DNA synthesis except, in some cases, the last nucleotide of exon 1. The putative variants were sequenced with a kit (USB) using single-stranded or double-stranded DNA templates. After confirming the desired mutation and ruling out adventitious mutations, variants were subcloned into the pHEBO shuttle vector [27].

Transfection. 721.221 cells (5-10 × 106/ml) were incubated for 10 minutes at room temperature with 0.010.05 mg DNA. The cells were then electroporated by two 0.1-msec pulses of 1286 V/cm field strength in Hepes buffered saline [31] using a BTX T100 electroporator. The transfected cells were resuspended in 6 24 ml of a 50:50 mixture of 721.221 conditioned media and fresh RPMI-1640 supplemented with 10% iron-supplemented bovine calf serum, and plated at 1 ml/well in a 24-well plate (Costar). Transfectants were selected with hygromycin B (Calbiochem) at a final concentration of 0.15 mg/ml on day 3 and 0.3 mg/ml on day 5. Rapidly growing cells from multiple wells were pooled and maintained in 0.3 mg/ml hygromycin B.

Two-step flow cytometry. Cells were incubated, 2.5-5 × 105, on ice with 0.02 ml mAbs for 20 min, washed by centrifugation through a 0.5-ml bovine calf serum cushion, stained with 0.02 ml FITC-conjugated goat-antimouse IgG (Fisher) for 20 minutes on ice, and washed as above. Cells were resuspended in buffer containing propidium iodide (0.005 mg/ml) and immediately analyzed on a Becton-Dickinson 440 FACS. Live cell gates were set using propidium iodide content, forward scatter, and orthogonal scatter. Corrected median or mean fluorescence intensity was calculated by subtracting the goat-anti-mouse only control value (always similar to the isotype control value). Values were normalized to the signal produced by the pan-HLA mAb, W6/32 (corrected experimental value + corrected W6/32 value). For each variant, the corrected fluorescence intensity for each antibody was compared with a wild-type B*0702 transfectant in the same experiment; each transfectant was tested in at least three separate experiments.

HLA-B*0702 Epitopes Recognized by Antibodies

TABLE

1

127

Effect o f B * 0 7 0 2 m u t a t i o n s on antibody epitopes Antibody"

Transfected variant b

MB40.2

Broad Narrow mAb mAb BB7.1 MB40.3 average ~ average d

V165M E166Q R169L E180Q R181G A182P D183E

1.06 1.71 0.11 0.00 1.74 0.00 0.98

0.83 0.05 0.11 1.00 0.85 0.77 0.64

0.76 1.57 0.78 0.89 1.19 0.55 0.90

0.94 1.65 0.84 1.06 1.24 0.76 1.19

1.01 2.06 0.68 1.05 1.72 0.94 1.04

GIC SP2/V28M/Q54R ~ V28L Q32E E53G Q54R E55K R239G

0.93 1.04 1.25 1.54 1.25 1.88 1.45 1.28

1.79 0.88 0.80 1.27 1.15 0.95 0.97 0.98

0.76 0.88 0.93 1.24 1.23 0.85 1.26 0.98

0.91 1.05 0.76 1.22 0.84 0.83 0.72 0.70

0.98 0.90 0.79 1.07 1.11 0.95 1.22 1.07

" Change in median or mean fluorescence intensity for each antibody, normalized on pan-anti-HLA mAb, W6/32 (see Materials and Methods). Values markedly different from control mAb binding are shown in bold. b 721.221 cells transfected with variant B*0702 genes, as described in

cantly affect o r c o m p l e t e l y e l i m i n a t e m A b binding. F o r e x a m p l e , M B 4 0 . 2 b i n d s t h e B*0702 w i l d - t y p e transfectant at high levels (Fig. lc). In contrast, n e i t h e r M B 4 0 . 2 n o r the i s o t y p e c o n t r o l m A b , C R l l - 3 5 1 , b i n d (Fig. l b and d) a B * 0 7 0 2 variant with an arginine to leucine s u b s t i t u t i o n at p o s i t i o n 169 (R169L). M B 4 0 . 3 binds at high level to b o t h R 1 6 9 L and w i l d - t y p e B*0702 (Fig. lh). In t h r e e e x p e r i m e n t s , M B 4 0 . 3 b i n d i n g to the R 1 6 9 L variant was e i t h e r slightly g r e a t e r than o r slightly less than the p o s i t i v e c o n t r o l B*0702, b u t M B 4 0 . 2 b i n d i n g to the R 1 6 9 L variant was always u n d e t e c t a b l e . T h e loss o f a n t i b o d y e p i t o p e s is selective, as 12 o t h e r m A b s b i n d to the R 1 6 9 L variant at w i l d - t y p e B*0702

F I G U R E 1 B*0702 mutations have an all-or-nothing effect on anti-B*0702 mAb binding. HLA-A,B,C-negative 721.221 cells transfected with wild-type B*0702 (a, c, e, and g) or with B*0702 variant R169 (b, d, f, and h) were stained with isotype control mAb C R l l - 3 5 1 or with B*0702-binding mAbs MB40.2, BB7.1, or MB40.3. Fluoresence is shown on a four-decade log scale.

R169L

B7

Materials and Methods. Average staining produced by the broadly reactive antibodies L368, BBM1, MB40.5, BB7.7, 4E, and F4/326. d Average staining produced by the remaining narrowly reactive antibodies ME1, B27M1, BB7.6, and SFR8-B6. ' This variant contains three substitutions: S2P, V28M, and Q54R.

RESULTS

Mutational strategy. B e c a u s e r e s i d u e s 1 7 6 / 1 7 8 w e r e i m p l i c a t e d in the M B 4 0 . 2 and M B 4 0 . 3 e p i t o p e s [19], w e s u b s t i t u t e d n e a r b y a m i n o acids r e s i d u e s in the B*0702 t e r t i a r y structure. A n t i b o d y typically contacts a large surface a r e a ( 6 0 0 - 9 0 0 42 ) on p r o t e i n antigen [ 3 2 - 3 5 ] . T h e r e f o r e , a 1 5 - 4 radius c e n t e r e d o n r e s i d u e 176 was c h o s e n to test a surface a r e a g r e a t e r than 700 42, w h i c h i n c l u d e s b o t h p o l y m o r p h i c and n o n p o l y m o r p h i c residues. B a s e d o n the A * 0 2 0 1 s t r u c t u r e ([2] and d a t a n o t shown), six sets o f ~ - c a r b o n a t o m s are close to r e s i d u e 176: r e s i d u e s 1 - 5 , 2 7 - 3 3 , 4 8 - 5 5 , 1 6 5 - 1 8 3 , 2 0 9 - 2 1 0 , and 239. W h e r e p o s s i b l e , at least two resid u e s w e r e m u t a t e d in each r e g i o n (Table 1), b u t resid u e s that d o n o t vary a m o n g m a m m a l i a n t r a n s p l a n t a t i o n antigens, such as H 3 and F210, w e r e n o t m u t a t e d . Each variant was t e s t e d in at least t h r e e s e p a r a t e e x p e r i m e n t s .

Mutations in B*0702 residues 1 6 9 - 1 8 2 abrogate MB40.2 binding. T h e m u t a t i o n s s t u d i e d e i t h e r d o n o t signifi-

b

a l!l c l!l d ,l/el l!l f •

I

I

I

I

g

CR11-351

BB7.1

I

h MB40.3

Log Fluorescence

128

levels (Table 1). Like R169L, the other mutations listed in Table 1 abrogate the binding of no more than one or two mAbs; thus the mutations tested probably do not cause large long-range B*0702 conformational changes. In addition to the R169L mutation, the E180Q and A182P mutations abrogate MB40.2 binding (Table 1). B*0702 residues 169, 180, and 182 are all close to residue 176, spanning an a-carbon atom distance of 17 (data not shown). Of the five B*0702 residues implicated in MB40.2 binding, four are charged amino acids that were replaced by neutral amino acids (Table 1). The MB40.2 epitope is further defined by B*0702 mutations, V165M, E166Q, R181G, and D183E, which do not affect antibody binding. Although mutations E 166Q and R 181G are nonconservative, substituting a charged amino acid with a neutral amino acid of equal or smaller size, mutations V165M and D183E have conservative substitutions. These data indicate that some, but not all, residues close to residue 176 affect MB40.2 binding. Antiprotein antibodies typically contact multiple residues that are close in the antigen tertiary structure, but distant in the primary structure [32-35]. To examine other regions close to B*0702 residue 176 in the tertiary structure, we mutated eight additional residues chosen from four distinct segments of the primary structure. Despite nonconservative amino acid substitutions, none of these mutations altered MB40.2 binding (Table 1). Mutations changed amino acid charge (Q32E, E53G, Q54R, E55K, and R239G) and increased amino acid size (GIC, S2P, Q54R, and E55K); one mutation caused multiple substitutions (S2P/V28M/Q54R). Thus, MB40.2 does not appear to contact these eight amino acid side chains that are close to residue 176 in the HLA tertiary structure, but distant in the primary structure. Combined with the ability of four mutations in residues 169-182 to abrogate MB40.2 binding, this suggests the possibility that MB40.2 recognizes a sequential B*0702 epitope.

Residues 176/178 may be distant from the MB40.3-binding site. Both MB40.2 and MB40.3 binding are selectively abrogated by a double substitution at residues 176/178; the binding of 10 other antibodies is not affected [19]. Interestingly, none of the 15 mostly nonconservative mutations close to residues 176/178 affect MB40.3 binding (Table 1), making it unlikely that MB40.3 contacts residues 176 or 178. Rather, our data suggest that the 176/178 mutation affects a distant MB40.3 contact site by subtly changing B*0702 conformation. BB7.1 binding requires B*0702 a2 residues 166 and •69. BB7.1 and MB40.2 crossblock each other's binding to

J.A. McCutcheon and C.T. Lutz

B*0702 [18], but the BB7.1 epitope has been mapped to al residues 63, 67, and/or 70 [21]. We find the R169L mutation abrogates both BB7.1 and MB40.2 binding (Fig. 1 and Table 1). BB7.1 binding also is abrogated by the E166Q mutation, but BB7.1 binding is not affected by any of the 13 other mutations near residue 176 (Table 1). These data suggest that BB7.1 could bind residues in both a-helices. Because the R169L mutation abrogates both MB40.2 and BB7.1 binding, competition for this side chain could explain the ability of these mAbs to crossblock. DISCUSSION MB40.2 and MB40.3 were raised against B*4001, but react with other HLA-B alleles and have been studied extensively using B*0702 [16, 18]. MB40.2 and MB40.3 simultaneously bind B*0702 [12, 18], indicating that these antibodies contact nonoverlapping B*0702 sites. Paradoxically, both mAb require B*0702 az-domain residues 177-180 [21]; both MB40.2 binding and MB40.3 binding are abrogated specifically by the K176N/K178T B*0702 mutation [19]. Thus previous mutagenesis studies implied that MB40.2 and MB40.3 contact the same a2 site. To investigate the apparent contradiction between crossblocking and mutagenesis studies, we mutated 15 B*0702 residues close to residue 176 (Table 1). B*0702 and B'4001 share all residues mutated in this study, except residue 32 [1], making it likely that the current findings apply to both antigens. MB40.2 binding requires residues 169, 180, and 182 (Table 1), all close to residue 176 in the primary structure. Based on the crystal structure of A*0201, the side chains of residues 169, 180, and 182 point "out" toward the solvent. In contrast, MB40.2 does not require side chains from residues 165, 166, 181, and 183. The residue 165 side chain points toward the peptide binding groove, the residue 181 side chain points "underneath" the final loop of the a-helix, and the residue 183 side chain points toward the a3 domain. Thus, the B*0702 residue 165,166, 181, and 183 side chains may not be accessible for MB40.2 binding. This is not surprising because, even in short peptide epitopes, not all antigen side chains interact with antibody [36]. Two other possibilities could explain the failure of residues 165, 166, and 183 to affect MB40.2 binding. First, these residues may lie outside the epitope. Second, the conservative amino acid changes at residues 165 and 183 may mask any potential effect. The current data do not distinguish between these possibilities. MB40.2 binding is not affected by eight largely nonconservative B*0702 mutations from four sites close to residue 176 in the tertiary structure (Fig. 2), but distant

HLA-B*0702 Epitopes Recognized by Antibodies

FIGURE 2 The B*0702 sequential epitope recognized by mAb MB40.2. A drawing of B*0702 is based on the A*0201 molecule [2], oriented with the cell membrane at the bottom of the figure. B*0702 mutations at positions 169, 176/178, 180, and 182 abrogate MB40.2 binding and are indicated by hatching. Mutations at 12 positions that do not affect MB40.2 binding (see Table 1 for listing) are indicated by stippling.

in the primary structure. Therefore, these side chains from nearby B*0702 sites are not involved in MB40.2 binding. Combined, these data suggest that MB40.2 may recognize a sequential B*0702 epitope formed by residues between 169 and 182 (Table 1 and Fig. 2). This result explains why MB40.2 binding depends upon B*0702 residues 177-180, or residues 176/178, in the earlier studies [19, 21]. Antibody binding depends on favorable energetics [37]. We speculate that only a few B*0702 mutations affect MB40.2 binding because each required residue provides considerable energy for binding. For example, B*0702 residues R169, E180, and K176 are charged, thereby having the potential to form high-energy ionic bonds with MB40.2. Novomy et al. [37] have calculated that two or three high-energy ionic bonds provide most of the energy for antibody-antigen complex for-

129

marion, even in complexes with large contact areas. In contrast, the A 182P mutation probably affects MB40.2 binding by steric hindrance. Although we cannot rule out a contribution from residues distant in the B*0702 primary structure or from hydrophobic interactions, the current data are consistent with the idea that MB40.2 binds a sequential B*0702 epitope via a few high-energy bonds. Noncrossblocking mAbs can bind different faces of the same residues [24]. For example, the noncrossblocking mAbs HyHEL-10 and D1.3 contact different atoms of the same hen egg lysozyme residue, R21 [24, 32, 34]. Thus, one mechanism to account for the ability of the noncrossblocking mAbs MB40.2 and MB40.3 to be abrogated by mutations at 176/178 is that these mAbs bind to different faces of these residues. In contrast to the 176/178 mutation, however, none of 15 other mutations near residue 176 affect MB40.3 binding (Table i and Fig. 2). We conclude that MB40.3 does not contact any of the 15 side chains tested, making it unlikely that MB40.3 contacts residues 176-180. Rather, our data suggest that the 176/178 mutation [19] and the B7/B27 recombinants [21] affect a distant MB40.3 contact site by subtly changing B*0702 conformation. This hypothesis is consistent with earlier mAb binding studies [16]. MB40.3 binding causes both BB7.1 and an unrelated mAb, ME1, to dissociate from B*0702 [ 16]. The kinetics of these interactions [ 16] suggest that MB40.3 induces a specific B*0702 conformation required for binding. Thus the MB40.3-induced change in B*0702 conformation might be prohibited by B*0702 mutations distant from the MB40.3 contact site. Previous studies have shown that BB7.1 and MB40.2 compete for binding to B*0702 [18]. However, B7/ B27 recombinants showed that MB40.2 requires B*0702 ~2-domain residues 177-180, whereas BB7.1 requires B*0702 ~l-domain residues 63, 67, and 70 [21]. Our data resolve this apparent discrepancy. We show that both MB40.2 and BB7.1 binding are abrogated by the R169L mutation and that BB7.1 binding is also abrogated by the E166Q mutation. Because B*0702 and B'2705 do not differ at residues 166 and 169 [1] our data easily reconcile previous crossblocking data with the results from B7/B27 recombinants. Based on our data and previous results [21], we can envision at least two models of BB7.1 binding. First, BB7.1 may straddle the peptide binding groove, contacting both the ~z and o~2~-helices. Second, because OL1 residues 63, 67, and 70 implicated in BB7.1 binding [21] all point into the peptide binding groove [1-3], the BB7.1 epitope may depend directly or indirectly on bound peptide. Our results point out the difficulties inherent in HLA

130

epitope mapping based exclusively upon substituting residues that differ between two HLA alleles. Mutation of residue 182 helped define the MB40.2 epitope and mutation of residue 166 helped define the BB7.1 epitope, even though these residues are not polymorphic at the HLA-B locus [1]. More importantly, the R169L mutation was crucial for mapping the MB40.2 and BB7.1 epitopes, but residue 169 is invariant among known HLA-A,B,C alleles [1]. Other residues chosen for mutagenesis (Table 1) show little or no polymorphism [1]. A mutagenesis strategy that relied upon differences between two HLA alleles would not have allowed a detailed examination of epitopes mapped to the o~2-o~3-domain junction. The region around residue 176 appears clinically important. Based on the sequences of binding and nonbinding HLA alleles and blocking with mAbs, human alloantibodies have been mapped to B*0702 residues 166-182 [23]. Furthermore, residues 166 and 174 determine the binding o f mouse alloantibodies to mouse H-2 transplantation antigens [38]. Thus it is significant that the B*0702 1 6 6 - 1 8 2 region influences three types of epitopes. MB40.2 may recognize an apparently sequential B*0702 epitope. BB7.1 appears to recognize an assembled epitope that may straddle the B*0702 peptide-binding groove. Finally, MB40.3 recognizes an epitope that appears to be distant from residues 1 6 6 182, yet clearly is influenced by this antigenically important site. Because mAbs influenced by a defined B*0702 region appear to recognize three distinct types of epitopes, it is likely that human anti-HLA antibodies will also recognize diverse types of epitopes.

ACKNOWLEDGMENTS

We thank D.A. Jensen, T. Foy, and R. Nunez for technical assistance; J. Barbosa, R. DeMars, S.Y. Yang, P. Parham, and R. Karr for reagents; and J. Barbosa for B*0702 sequence information and unpublished results. We thank K. Smith, N. Goeken, G. Bishop, J. Kemp, R. Olson, and P. Parham for valuable discussions. This work was supported by American Cancer Society grant JFRA-256, March of Dimes Birth Defects Foundation Basil O'Connor grant 5-741, and NIH grant AI27879.

REFERENCES 1. Bjorkman PJ, Parham P: Structure, function and diversity of class I major histocompatibility complex molecules. Annu Rev Biochem 59:253, 1990. 2. Bjorkman PJ, Saper MA, Samraoui B, Bennett WS, Strominger JL, Wiley DC: Structure of the human class I histocompatibility antigen, HLA-A2. Nature 329:506, 1987. 3. Saper MA, Bjorkman PJ, Wiley DC: Refined structure of

J.A. McCutcheon and C.T. Lutz

4.

5.

6.

7.

8.

9.

10.

11.

the human histocompatibility antigen HLA-A2 at 2.6 resolution. J Mol Biol 219:227, 1991. Madden DR, Gorga JC, Strominger JL, Wiley DC: The structure of HLA-B27 reveals nonamer self-peptides bound in an extended conformation. Nature 353:321, 1991. Falk K, Rotzschke O, Rammensee H-G: Cellular peptide composition governed by major histocompatibility complex class I molecules. Nature 348:248, 1990. Falk K, Rotzschke O, Stevanovic S, Jung G, Rammensee H-G: Allele-specific motifs revealed by sequencing of self-peptides eluted from MHC molecules. Nature 351:290, 1991. Jardetzky TS, Lane WS, Robinson RA, Madden DR, Wiley DC: Identification of self peptides bound to purified HLA-B27. Nature 353:326, 1991. Patel R, Terasaki PI: Significance of the positive cross match test in kidney transplantation N Engl J Med 280:735, 1969. Yankee RA, Grumet FC, Rogentine GN: Platelet transfusion therapy: the selection of compatible platelet donors for refractory patients by lymphocyte HL-A typing. N EnglJ Med 281:1208, 1969. Schwimmbeck PL, Yu DTY, Oldstone MBA: Autoantibodies to HLA-B27 in the sera of HLA-B27 patients with ankylosing spondylitis and Reiter's syndrome: molecular mimicry with Klebsiella pneumoniae as potential mechanism of autoimmune disease. J Exp Med 166:173, 1987. Swansom RA, Tracy RP, Katzmann JA, Wilson DM, Young DS:/~2-Microglobulin determined by radioimmunoassay with monoclonal antibody. Clin Chem 28:2033, 1982.

12. Parham P: Monoclonal antibodies against two separate alloantigenic sites of HLA-B40*. Immunogenetics 13: 509, 1981. 13. Berger AE, Davis JE, Cresswell P: Monoclonal antibody to HLA-A3. Hybridoma 1:87, 1982. 14. Radka SF, Kostyu DD, Amos B: A monoclonal antibody directed against the HLA-Bw6 epitope. J Immunol 128:2804, 1982. 15. Russo CK, Ng A-K, Pellegrino MA, Ferrone S: The monoclonal antibody CRll-351 discriminates HLA-A2 variants indentified by T cells. Immunogenetics 18:23, 1983. 16. Parham P: Changes in conformation with loss of alloantigenic determinants of a histocompatibility antigen (HLAB7) induced by monoclonal antibodies. J Immunol 132:2975, 1984. 17. Yang SY, Morishima Y, Collins NH, Altar CT, Pollack MS, Yunis EJ, Dupont B: Comparison of 1-dimensional IEF patterns for serologically detectable HLA-A and B allotypes, lmmunogenetics 19:217, 1984. 18. Parham P, Antonelli P, Herzenberg LA, Kipps TJ, Fuller A, Ward FE: Further studies on the epitopes of HLA-B7

HLA-B*0702 Epitopes Recognized by Antibodies

131

defined by murine monoclonal antibodies. Hum Immunol 15:44, 1986.

site-specific mutagenesis without phenotypic selection. Methods Enzymol 154:367, 1987.

19. Barbosa JA, Santos-Aguado J, Mentzer SJ, Strominger JL, Burakoff SJ, Biro PA: Site-directed mutagenesis of class I HLA genes: role ofglycosylation in surface expression and functional recognition. J Exp Med 166:1329, 1987. 20. Taurog JD, E1-Zaatari FAK: In vitro mutagenesis of HLA-B27: substitution of an unpaired cysteine residue in the 1 domain causes loss of antibody-defined epitopes. J Clin Invest 82:987, 1988.

29. Sambrook J, Fritsch EF, Maniatis T: Molecular Cloning: A Laboratory Manual. Cold Spring Harbor, NY, Cold Spring Harbor Laboratory, 1989. 30. Ho SN, Hunt HD, Horton RM, Pullen JK, Pease LR: Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77:51, 1989.

21. Toubert A, Raffoux C, Boretto J, Sire J, Sodoyer R, Thurau SR, Amor B, Colombane J, Lemonnier FA, Jordan BR: Epitope mapping of HLA-B27 and HLA-B7 antigens by using intradomain recombinants. J Immunol 141: 2503, 1988. 22. Fuller AA, Trevithick JE, Rodey GE, Parham P, Fuller TC: Topographic map of the HLA-A2 CREG epitopes using human alloantibody probes. Hum Immunol 28: 284, 1990. 23. Fuller AA, Rodey GE, Parham P, Fuller TC: Epitope map of the HLA-B7 CREG using affinity-purified human alloantibody probes. Hum Immunol 28:306, 1990.

32. Amit AG, Maruizza RA, Phillips SEV, Poljak RJ: Threedimensional structure of an antigen-antibody complex at 2.8 ~ resolution. Science 233:747, 1986.

24. Benjamin DC: Molecular approaches to the study of B cell epitopes. Int Rev Immunol 7:149, 1991. 25. Lesley JF, Schulte RJ: Section of cell lines resistant to anti-transferrin receptor antibody: evidence for a mutation in transferrin receptor. Mol Cell Biol 4:1675, 1984. 26. Ways JP, Parham P: The antigenic structure of HLA-A2: an analysis with competitive binding assays and monoclonal antibodies. J Immunol 131:856, 1983. 27. Shimizu Y, Geraghty DE, Koller BH, Orr HT, DeMars R: Transfer and expression of three cloned human nonHLA-A,B,C, class I major histocompatibility complex genes in mutant lymphoblastoid cells. Proc Natl Acad Sci USA 85:227, 1988. 28. Kunkel TA, Roberts JD, Zakour RA: Rapid and efficient

31. Chu G, Hayakawa H, Berg P: Electroporation for the efficient transfection of mammalian cells with DNA. Nucleic Acids Res 15:1311, 1987.

33. Sheriff S, Silverton EW, Padlan EA, Cohen GH, SmithGill SJ, Finzel BC, Davies DR: Three-dimensional structure of an antibody-antigen complex. Proc Natl Acad Sci USA 84:8075, 1987. 34. Padlan EA, Silverton EW, Sheriff S, Cohen GH, SmithGill SJ, Davies DR: Structure of an antibody-antigen complex: crystal structure of the HyHEL-10-Fab lysozyme complex. Proc Natl Acad Sci USA 86:5938, 1989. 35. Colman PM, Laver WG, Varghese JN, Baker AT, Tulloch PA, Air GM, Webster RG: Three-dimensional structure of a complex of antibody with influenza virus neuraminidase. Nature 326:358, 1987. 36. Getzoff ED, Geysen HM, Rodda SJ, Alexander H, Tainer JA, Lerner RA: Mechanisms of antibody binding to a protein. Science 235:1191, 1987. 37. NovomyJ, Bruccoleri RE, Saul FA: On the attribution of binding energy in antigen-antibody complexes McPC 603, D1.3 and HyHEL-5. Biochemistry 28:4735, 1989. 38. Ajitkumar P, Geier SS, Kesari KV, Borriello F, Nakagawa M, Bluestone JA, Saper MA, Wiley DC, Nathenson SG: Evidence that multiple residues on both the a-helices of the class I MHC molecule are simultaneously recognized by the T cell receptor. Cell 54:47, 1988.