International Immunology, Vol. 12, No. 3, pp. 305–312

© 2000 The Japanese Society for Immunology

Mapping the B cell superantigen binding site for HIV-1 gp120 on a VH3 Ig Mehran N. Neshat, Lee Goodglick, Kathleen Lim and Jonathan Braun1 Department of Pathology and Laboratory Medicine, and 1The Molecular Biology Institute, University of California–Los Angeles, 10833 Le Conte Avenue, Los Angeles, CA 90095-1732, USA Keywords: gp120, HIV-1, Ig, Protein A, superantigen, VH gene family

Abstract The emerging class of B cell superantigens includes HIV-1 gp120, which binds to many members of the VH3 Ig gene family. The present study addresses the structural features of VH3 antibodies conferring gp120 binding activity using a panel of recombinant full-length and Fab Ig proteins. Binding activity was fully conferred by the Fab portion of the Ig molecule. The VH region was the major determinant of binding; diverse light chains were permissive for gp120 binding. A series of recombinant VH3–VH1 chimeric molecules was created to analyze the contribution of different subregions of VH3 to gp120 binding. Hypervariable loop 1 (H1) substitution alone caused a 10-fold reduction in binding activity. The framework subregions (FR1, FR2 and FR3) and H2 also influenced binding, since substitutions of various combinations of these subregions conferred 10- to 100-fold binding reductions. We conclude that gp120 binding occurs through a nonconventional interaction involving multiple discontinuously arrayed residues spanning the VH, and including roles in gp120 contact and favorable conformation of the VH. Introduction The envelope glycoprotein of HIV-1, gp120, is one of an emerging family of microbial and somatically encoded superantigens for the Ig gene family (1–4). Ig superantigens bind to a high proportion of membrane-bound or soluble Ig through sequence-conserved subregions of certain Ig gene families and thus activate a large B cell subpopulation bearing the relevant superantigen-binding VH gene family members (5–7). In particular, Silverman and colleagues have demonstrated that Ig superantigens deliver a major impact on murine B cell clonal selection and activation (8). Studies of human HIV-1 infection have suggested an in vivo superantigenic role of HIV gp120, as judged by systematic changes in abundance and antibody production of VH3 B cells, and the depletion of common VH3 B cells in late disease and immunodeficiency lymphomas (6,9–12). A protective role for superantigenbinding Ig has been suggested by epidemiologic analysis (13). The selective interaction of Ig or T cell superantigens with entire families of Ig or TCR suggests that binding does not occur in the hypervariable (H) subregions, but instead involves conserved gene family peptides. In the case of human Ig genes, seven VH gene families have been defined based on

amino acid sequence homology, mainly localized in the framework (FR1 and FR3) subregions. Interfamily and interspecies comparison of Ig gene families suggests that certain family homologies are conserved phylogenetically. These homologies have given rise to the categorization of Ig families into three clans, which probably reflect the radiation of the primordial Ig gene family during vertebrate evolution. Among these, clan 3 includes the human VH3 family and VHIII subgroup of Kabat (14–18). The best characterized Ig superantigen is Staphylococcus aureus Protein A (SPA) (19–22). SPA binds the Fab portion of at least 15 members of the human VH3 family. VH3 is the largest family of variable heavy (VH) chain genes, comprised of ~22 members. Sequence comparisons between natural or mutagenized VH3 Ig that bind or do not bind SPA suggest that FR1, FR3 and an invariant portion of H2 are critical for SPA binding. FR1 and FR3 are highly conserved in VH3 family members. Structurally, these two β folds come together to form a surface-exposed region which has homology to the site of T cell superantigen binding on the TCR Vβ chain (17,23,24). For SPA, it is unknown whether FR1, FR3 and H2 form the actual Ig superantigen binding site or whether these regions structurally influence the binding site.

The first two authors contributed equally to this work Correspondence to: J. Braun Transmitting editor: K. L. Knight

Received 27 July 1999, accepted 18 November 1999

306 VH3 subregions for superantigen binding The present study characterizes the gp120 binding site of VH3 Ig. Using a panel of mAb and molecularly constructed chimeras, H1 appeared to be a major determinant for gp120 binding, with intermediate contributions by the other VH subregions. Materials Antibodies 5-2 and 5-3 (VH3-30 gene, VH3 clan) are monoclonal VH3 Fab fragments purified, cloned and characterized as previously described (25). Both 5-2 and 5-3, as well as the monoclonal anti-tetanus toxoid VH1 Fab, P313, were cloned into the bacteriophage expression vector. pComb3 and soluble Fab was produced and purified as described (25,26). P313 (VH1 family) and pComb3 were generous gifts from Dr Carlos Barbas III (UCSD). The anti-DNS murine monoclonals TAAO (J606 family, VH3 clan) and TSU (J558 family, VH1 clan) were generous gifts from Dr Sherie Morrison. These were constructed as IgG1 (both) and IgA (TAAO only) isotypes. Ig constructs The Ig superantigen binding site on VH3 Ig was analyzed by constructing Fab fragments containing exchanged regions between the P313 and 5-3 antibodies. The genes encoding both P313 and 5-3 were cloned into the bacteriophage expression vector, pComb3, including a C-terminal hexahistidine for purification purposes. DNA encoding the chimeric regions were constructed by splicing overlap extension (27,28) using Pfu polymerase and subcloned into the PCRscript cloning vector (Stratagene, La Jolla, CA). For technical reasons, 5-2 sequence was introduced into the N-terminal portion of the FR3 subregion during creation of the ∆H3 fusion proteins. The presence of the desired subfragments was confirmed by restriction endonuclease mapping. Subfragments were excised by restriction endonucleases and ligated into the pComb3 expression vector. The identity of heavy and light chain coding sequences were confirmed by DNA sequencing by dideoxy method. Recombinant Fab proteins were produced in XL-1 Blue cells as secreted protein from 0.5–1 L bacterial cultures after IPTG induction and purified using Ni2⫹ affinity chromatography as previously described (25). The proteins were analyzed by SDS–PAGE and Coomasie blue staining for purity and yield by titration and comparison with a human Fab⬘ IgG calibration standard. In addition, the products were assessed for yield of immunoreactive κ light chain by anti-Fab⬘ capture ELISA (25,29). The latter criteria permitted us to quantitate the amount of product forming dimeric VH–VL and retaining native conformation for isotype detection. Products were only accepted when the purity was ⬎80% by SDS–PAGE, when the Fab⬘ yield (by SDS–PAGE protein analysis) was between 0.3 and 1 mg/l of bacterial culture, and when the quantitation of Fab⬘ protein and immunoreactive VH–VL was concordant (within 20%). By these criteria, the quality of the different recombinant proteins was comparable. ELISA The direct binding ELISA was previously described (29). For all experiments, biotinylated goat F(ab⬘)2 anti-human or

-mouse secondary reagents were used, followed by 0.4 mg/ml poly-HRP40–streptavidin (Research Diagnostics, Flanders, NJ) and chromogenic substrate. Tetanus toxoid was a generous gift from Dr Andrew Saxon (UCLA). Peptides derived from the MN strain of HIV-1 gp120, were obtained from the AIDS Research and Reference Reagent Program, National Institutes of Health. Peptide inhibition studies were performed as described (29). The following gp120 derived peptides were used: peptide 1988 (C2 subregion, peptide 331–350), peptide 1959 (C3 subregion, peptide 231–250), peptide 1960 (C3 subregion, peptide 241–260) and peptide 1990 (C3 subregion, peptide 351–370). Peptides 1988, 1959 and 1960 are termed superantigen peptides based on their ability to inhibit the binding of Ig to gp120. Peptide 1990 had no inhibitory effect on superantigen binding and was therefore used as a control for an irrelevant peptide.

Results Binding to gp120 requires the VH3 Fab but accommodates diverse VL sequences Previous studies indicate that the superantigen binding activity of gp120 has a strong preference for VH3 family Ig. Diverse heavy chain isotype-encoded sequences are permissive for this binding, reflecting the capacity of gp120 to bind endogenous serum VH3 Ig of α, γ or µ heavy chain origin (5,6,29). We directly verified this observation using recombinantly engineered IgA and IgG isotypes of TAAO, a murine VH3 clan protein of the J606 family (Fig. 1). The IgA and IgG forms of TAAO were each tested for gp120 binding (Fig. 1A). Both were active, whereas a recombinant murine VH1 clan protein of the J558 family, TSU, showed no superantigen binding activity. This represents the first example of a murine VH3 gene product tested for superantigenic gp120 binding activity. To validate this specificity of TAAO, we took advantage of the observation that superantigenic VH3–gp120 interaction is inhibited by gp120 peptides (1988, 1959 and 1960) bearing the gp120 superantigen decamer motif (29,30). As shown in Fig. 1(B), binding of TAAO to gp120 was inhibited by the superantigen peptide 1988, but not by an irrelevant gp120 peptide (1990). Moreover, recombinant VH3 antibodies engineered with only a Fab (hence bearing only CH1 of the constant region) were competent for gp120 binding (Fig. 2). These observations demonstrate that the heavy chain constant region was not required for gp120 binding. The contribution of light chain sequence to gp120 was tested in two ways. First, we compared the effect of light chains on gp120-binding activity of VH3 and VH1 recombinant human Fab proteins (Fig. 2A). 5-2 and 5-3 are two human VH3-30 molecular clones (anti-histone H1) composed of VA27 Vκ light chains with distinct junctional and FR4 (Jκ1 and Jκ3) sequences (see Fig. 4) (25). gp120 binding activities of 5-2 and 5-3 were similar. This finding suggested that the light chain sequence in this subregion were either non-critical or both permissive for gp120 binding. Second, we exchanged the light chain between a gp120binding VH3 Fab (5-3) and a gp120-non-binding VH1 Fab (P313, anti-tetanus toxoid) (26,29). In the chimeric 5-3 heavy chain ⫹ P313 light chain Fab, gp120 binding was maintained

VH3 subregions for superantigen binding 307 (Fig. 2A). In contrast, the converse chimeric Fab (P313 heavy chain ⫹ 5-3 light chain) failed to bind tetanus toxoid (Fig. 2B). These observations suggested that gp120 interaction

was dependent on the VH3 sequence. Neither 5-2 nor 5-3 exhibited SPA binding (data not shown), consistent with previous evidence that certain VH3 genes do not bear certain critical residues this binding interactions (2,4,21). FR1 and FR3 are not sufficient for gp120-binding We next analyzed the contribution of VH residues to gp120 binding. The general strategy was to engineer recombinant VH genes in which different residues of the VH domain were exchanged between VH3 and VH1 genes (Fig. 3; representative sequences shown in Fig. 4). VH chimeras were expressed with the VH1 cognate light chain, so that antitetanus toxoid activity in chimeras could be assessed. The results of the full data set are tabulated in Fig. 3. Representative binding curves are shown in Fig. 5(A) (gp120) and Fig. 6 (tetanus toxoid). For constructs with weak gp120 binding activity, specificity was validated using gp120 peptide competition (Fig. 5B). The contributions of FR1, FR3 and H2 were first evaluated, because these regions form a surface patch implicated in SPA binding (20,21). VH1 FR1 and FR3 (∆F13H3) fully abrogated gp120 binding (Figs 3 and 5A). VH1 FR1 (∆F1H3) substantially reduced gp120 binding (10–20% of wild-type Fab 5-3). These effects were not due to H3 alone, since this element modestly affected binding (∆H3, 30% of wild-type Fab 5-3). Conversely, using VH1 as the template, introduction of VH3 FR1 or FR3 (∆F123H23 or ∆F12H123) failed to confer detectable gp120binding activity. As a structural control, two constructs bearing all three VH1 hypervariable loops (∆F2H123 and ∆F12H123) were tested for tetanus toxoid binding. Both chimeric Fabs maintained tetanus toxoid binding activity, indicating that H1, H2 and H3 were properly aligned for conventional antibody–antigen interaction (Figs 3 and 6). Structural differences in binding of gp120 and conventional antigen H3 is critical for most conventional antibody–antigen interactions. The effect of this region was evaluated by introducing the VH1 H3 into the VH3 Fab (∆H3). As noted above, gp120 binding was only modestly reduced in this chimera (30% of Fab 5-3). When both H1 and H3 from VH1 were introduced into VH3 (∆H13), gp120 binding was profoundly impaired (3% of Fab 5-3) (Fig. 5A). The residual gp120 binding of these chimeras was gp120 specific, as the superantigen peptide

Fig. 1. gp120-binding activity of a murine anti-DNS VH3 antibody, TAAO. Molecular clones of the murine anti-DNS hybridomas TAAO (VHJ606, VH3 clan) and TSU (VHJ558 family, VH1 clan) were recombinantly engineered with IgG1 and IgA (TAAO) or IgG1 (TSU) constant regions, and transfectoma proteins were purified and assayed in gp120-binding studies. (A) gp120 binding of murine TAAO. Binding of IgG and IgA TAAO and TSU proteins to gp120-coated wells. Values are direct ELISA absorbances. (B) Effect of gp120 superantigen peptide on gp120 binding. TAAO–IgA (50 ng/ml) was preincubated with 20mer peptide either including (1988) or lacking (1990) the gp120 superantigen sequence and tested for binding to gp120-coated plates. ELISA absorbances were recalculated as percent control for TAAO without peptide preincubation.

308 VH3 subregions for superantigen binding

Fig. 3. Diagram of chimeric VH3/VH1 gene constructs. Relative binding activity was calculated by determining the ratio of Fab concentrations for chimeric and wild-type proteins to yield equivalent ELISA absorbances. Ratios below detection (0.01) are tabulated as zero binding activity.

Fig. 2. Role of VL and VH in gp120 binding by recombinant Fab. Purified recombinant Fab proteins were produced with molecular clones of human VH3 (5-3 or 5-2, both anti-histone H1) and VH1 (P313, antitetanus toxoid), expressed either with their original or reciprocal light chains. Purified proteins were tested by ELISA for binding to gp120 (A) or tetanus toxoid (B). Relative absorbances were recalculated as maximum positive controls (gp120 binding of Fab 5-3 at 1 µg/ml; tetanus toxoid binding of Fab P313 at 0.1 µg/ml; range 0.7–1.2 OD492).

(1988) was inhibitory (Fig. 5B). The importance of H1 for gp120 binding was further emphasized by the fact that all constructs that contained VH1 H1 demonstrated little or no superantigen binding (Fig. 3). Nevertheless, introduction of VH3 H1 on the VH1 template (∆F123H23) was insufficient to confer gp120 binding (Fig. 5A). This construct also lost tetanus toxoid binding activity, consistent with the requirement of all three hypervariable loops for conventional antigen binding (Fig. 3). Discussion Ig superantigens are distinguished because their binding interaction with conserved VH family residues is distinct from conventional antigen-binding residues. HIV gp120 is the first example of a virally encoded Ig superantigen, based on its

binding requirement for conserved regions of VH3. The present study demonstrates that the interaction was dependent on the Fab but not Fc region and was permissive for diverse VL sequences, extending previous studies of this issue (5,29). Analysis of VH3–VH1 chimeric Fab proteins indicates that the interaction of VH3 Ig with gp120 directly or indirectly involves multiple residues spanning the VH domain. Three technical issues should be considered in interpreting this study. First, recombinant grafting may introduce detrimental and unpredictable effects on conformation (and hence binding activity), and this may be aggravated in non-germlineencoded sequences (16,31). Second, the quality of the recombinant Fabs is a significant issue. These were products of a bacterial expression system and non-functional protein could artifactually reduce the apparent gp120 binding activity of the recombinant product. Several criteria were used to assess and standardize the quality of the different recombinant Fabs used in this study (see Materials). However, these criteria are limited in defining conformational integrity and this remaining uncertainty needs to be considered in evaluating mutagenesis structure–function studies. Third, it should be noted that the Fab constructs included the IgG1 CH1; the possible contribution of this subregion to antigen binding has been structurally analyzed by Schroeder and colleagues (17), and has not been ruled out in this study. Binding to gp120 is a non-conventional interaction Antibodies derived from a typical immune response (such as a Fab P313) bind their cognate antigen typically using various

VH3 subregions for superantigen binding 309

Fig. 4. Sequences of representative VH genes and constructs. Hypervariable loops (H1–H3) use the definitions of Chothia (16,33) and numbering is in accordance with the Kabat nomenclature (34). Note that 5-2 sequence was introduced into the N-terminal portion of the FR3 subregion in creation of the ∆H3 fusion proteins. Asterisks denote VH3 residues previously implicated in gp120 binding (32). Dashes indicate sequence identity with 5-3.

hypervariable loops from the heavy and light chains. Accordingly, we observed that substitution of 5-3 hypervariable loops abolished binding of the chimeric Fabs to tetanus toxoid. In contrast, the interaction of VH3 Ig with gp120 displayed distinct requirements. First, H3 did not play a major role in superantigen binding. Accordingly, although 5-2, 5-3 and TAAO all bound gp120, there was no similarity in the hypervariable loops from any of the clones. Second, exchange of H3 from a non-reactive VH1 Fab (P313) to 5-3 did not result in a major loss in gp120 binding. Framework regions alone do not encode gp120 binding FR1 and FR3 are candidate subregions for gp120 binding, since they display strong sequence conservation among VH3 family members. These two sets of β strands come together to form a solvent-exposed surface which potentially could be a site for non-conventional binding (17). Consistent with this hypothesis, there is evidence that FR1 and FR3 plus H2 are necessary for SPA binding to VH3 Ig (20,21). In the present study, these subregions were insufficient for gp120 binding

(∆F2H123 and ∆F2H13). Unfortunately, it was not possible to confirm the integrity of these chimeras for SPA binding, since the wild-type VH3 Fab used in this study lacked SPA binding activity. VH3 antibodies with reduced or absent SPA binding have previously been analyzed for divergent residues impairing this interaction (4,20,21). Comparison with the Fab 5-2 and 5-3 reveals several potential deleterious residues: FR1 (15K, 35H), H2 (53D, 55-58/RKKK, 68F) and FR3 (82L). The roles of FR1, FR3 and H2 are nonetheless implicated in several Fab chimeras. VH1 replacement at each of these subregions reduced gp120 binding activity by a 3-fold increment and in combination reduced binding 10- to 100fold. While these findings may reflect the loss of gp120interacting residues, they may be due to indirect (e.g. conformational) effects. In this respect, replacement of FR2 similarly reduced gp120 binding, even though this subregion almost entirely lacks solvent exposure. Residues in this FR2 may indirectly affect packing and conformation of associated solvent-exposed subregions. The role of H3 substitution is unclear, because the binding reduction was modest and was

310 VH3 subregions for superantigen binding

Fig. 6. Binding of tetanus toxoid by Fab chimeras. Purified recombinant Fab proteins were reacted with TT-coated wells and absorbances (percent control) were as in Fig. 2.

along with residues 26 and 27, are critical determinants of H1 canonical structure (16). Conclusion Fig. 5. Binding of gp120 by Fab chimeras. (A) Purified recombinant Fab proteins at the indicated concentrations were reacted with gp120coated wells and ELISA absorbances were expressed as percent control (see Fig. 2). (B) Fab proteins (1 µg/ml) were preincubated with gp120 20mer peptides either bearing (1959 and 1988) or lacking (1990) the gp120 superantigen decamer sequence. Mixtures were reacted with gp120-coated wells and analyzed for peptide inhibition as in Fig. 1.

associated with a substitution of C-terminal FR3. In this regard, the substitution FR3 residues are probably buried and may unfavorably affect packing of the adjacent FR3 strand with regard to its role in gp120 binding. H1 is important for gp120 binding Chimeric Fabs were also constructed to examine the role of H1 through H2 in gp120 binding. A large effect was produced by the replacement of VH3 H1 (∆H3 to ∆H13), which reduced binding activity 10-fold. It is notable that this major change resulted from a minor sequence change in H1 (three residues). M34 is present in most VH3 family members and is conserved in all VH3 gp120-binding Ig so far examined. Although this methionine is most likely buried within the antigen binding site, it is thought to interact with residues in the loop connecting the FR1 to H1 (residues 24 and 29). These amino acids,

Karray et al. have recently structurally analyzed VH3–gp120 binding using a sequence comparison strategy with a panel of binding and non-binding sequence-defined Ig proteins (32). The present study complements and confirms the work of Karray et al. using a mutagenesis experimental strategy. Karray et al. observed 16 VH positions in which favorable sequences correlated with gp120 binding. These were distributed among four of the subregions (4 FR1 residues, 2 H1, 0 FR2, 1 H2, 9 FR3, 0 H3 and 0 FR4). Modeling suggested that the side chains of 14 positions were solvent exposed in a dispersed array on the side of the antibody opposite the VH–VL interface. In the present study, mutagenesis experimentation has provided a direct test of the role of these subregions in gp120 binding. Our findings are consistent with the assessment of Karray et al. First, gp120 binding proteins contained most of the favorable sequences (16 out of 16, 16 out of 16 and 13 out of 16 for 5-2, 5-3 and TAAO respectively), whereas they occurred in only seven out of 16 positions for the gp120 nonbinding P313. Second, subregion substitutions revealed that favorable sequences are present in the region spanning between the C-terminal half of FR1 to the N-terminal half of FR3. Third, favorable sequences were present in each of the subregions, with the probable exception of FR2. We note that these subregions probably contribute both contact residues and residues required for favorable conformation. The influ-

VH3 subregions for superantigen binding 311 ence of FR residues on hypervariable canonical structure and, reciprocally, of hypervariable loops on the flanking FR structures has been well documented. Therefore, the information to date does not permit a detailed prediction of the structural interaction between VH3 and gp120. However, it permits the conclusion that, in contrast to conventional antigen binding, VH3 binding to gp120 depends on a discontinuous array of residues broadly distributed on the VH molecule opposite the VH–VL interface. Acknowledgements This study was supported by NIH CA12800, AI38545 and DK46763, the Universitywide AIDS Research Program, and the AIDS Research and Reference Reagent Program, NIH. We wish to thank and honor the memory of Dr Kathelyn Steimer (Chiron Corp.) for gifts of gp120SF2, Carlos Barbas III for pComb3 and the P313, P313, Dr Andy Saxon for tetanus toxoid, and Dr Sherie Morrison for TAAO and TSU.

Abbreviations FR H SPA TT

framework region hypervariable loop Staphylococcus aureus Protein A tetanus toxoid

References 1 Goodglick, L., Townsley-Fuchs, J., Fairhurst, R. M., Neshat, M. N., Marcelo, D., Mosier, D. E. and Braun, J. 1996 The HIV-1 gp120 superantigen: structural features and its role in viral host–pathogen interaction. In Zouali, M., ed., B Cell Superantigens, p. 145. Landes, New York. 2 Zouali, M. 1995. B-cell superantigens: implications for selection of the human antibody repertoire. Immunol. Today 16:399. 3 Pospisil, R., Young-Cooper, G. O. and Mage, R. G. 1995. Preferential expansion and survival of B lymphocytes based on VH framework 1 and framework 3 expression: ‘positive’ selection in appendix of normal and VH-mutant rabbits. Proc. Natl Acad. Sci. USA 92:6961. 4 Silverman, G. J. 1997. B-cell superantigens. Immunol. Today 18:379. 5 Berberian, L., Goodglick, L., Kipps, T. J. and Braun, J. 1993. Immunoglobulin VH3 gene products: natural ligands for HIV gp120. Science 261:1588. 6 Berberian, L., Shukla, J., Jefferis, R. and Braun, J. 1994. Effects of HIV infection on VH3 (D12 idiotope) B cells in vivo. J. AIDS 7:641. 7 Kristiansen, S. V., Pascual, V. and Lipsky, P. E. 1994. Staphylococcal protein A induces biased production of Ig by VH3-expressing B lymphocytes. J. Immunol. 153:2974. 8 Silverman, G. J., Nayak, J. V., Warnatz, K., Hajjar, F. F., Cary, S., Tighe, H. and Curtiss, V. E. 1998. The dual phases of the response to neonatal exposure to a VH family-restricted staphylococcal B cell superantigen. J. Immunol. 161:5720. 9 Braun, J., Berberian, L., King, L., Sanz, I. and Govan, H. L. I. 1992. Restricted use of fetal VH3 immunoglobulin genes by unselected B cells in the adult. Predominance of 56p1-like VH genes in common variable immunodeficiency. J. Clin. Invest. 89:1395. 10 David, D., Goossens, D., Desgranges, C., The`ze, J. and Zouali, M. 1995. Molecular characterization of human monoclonal antibodies specific for several HIV proteins: analysis of the VH3 family expression. Immunol. Lett. 47:107. 11 Bessudo, A., Cherepakhin, V., Johnson, T. A., Rassenti, L. Z., Feigal, E. and Kipps, T. J. 1996. Favored use of immunoglobulin VH4 genes in AIDS-associated B-cell lymphoma. Blood 88:252. 12 Przybylski, G. K., Goldman, J., Ng, V. L., McGrath, M. S., Herndier, B. G., Schenkein, D. P., Monroe, J. G. and Silberstein, L. E. 1996.

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

Evidence for early B-cell activation preceding the development of Epstein–Barr virus-negative acquired immunodeficiency syndrome-related lymphoma. Blood 88:4620. Townsley-Fuchs, J., Kam, L., Fairhurst, R. M., Gange, S. J., Goodglick, L., Giorgi, J. V., Sidell, N., Detels, R. and Braun, J. 1996. HIV-1 gp120 superantigen-binding serum antibodies. A host susceptibility factor in homosexual HIV-1 transmission. J. Clin. Invest. 98:1794. Cook, G. P., Tomlinson, I. M., Walter, G., Riethman, H., Carter, N. P., Buluwela, L., Winter, G. and Rabbitts, T. H. 1994. A map of the human immunoglobulin VH locus completed by analysis of the telomeric region of chromosome 14q. Nat. Genet. 7:162. van Dijk, K. W., Mortari, F., Kirkham, P. M., Schroeder, H. W., Jr and Milner, E. C. B. 1993. The human immunoglobulin VH7 gene family consists of a small, polymorphic group of six to eight gene segments dispersed throughout the VH locus. Eur. J. Immunol. 23:832. Chothia, C., Lesk, A. M., Gherardi, E., Tomlinson, I. M., Walter, G., Marks, J. D., Llewelyn, M. B. and Winter, G. 1992. Structural repertoire of the human VH segments. J. Mol. Biol. 227:799. Kirkham, P. M., Mortari, F., Newton, J. A. and Schroeder, H. W., Jr. 1992. Immunoglobulin VH clan and family identity predicts variable domain structure and may influence antigen binding. EMBO J. 11:603. Schroeder, H. W., Jr., Hillson, J. L. and Perlmutter, R. M. 1990. Structure and evolution of mammalian VH families. Int. Immunol. 2:41. Sasso, E. H., Silverman, G. J. and Mannik, M. 1991. Human IgA and IgG F(ab⬘)2 that bind to staphylococcal protein A belong to the VHIII subgroup. J. Immunol. 147:1877. Hillson, J. L., Karr, N. S., Oppliger, I. R., Mannik, M. and Sasso, E. H. 1993. The structural basis of germline-encoded VH3 immunoglobulin binding to staphylococcal protein A. J. Exp. Med. 178:331. Randen, I., Potter, K. N., Li, Y., Thompson, K. M., Pascual, V., Forre, O., Natvig, J. B. and Capra, J. D. 1993. Complementaritydetermining region 2 is implicated in the binding of staphylococcal protein A to human immunoglobulin VHIII variable regions. Eur. J. Immunol. 23:2682. Roben, P. W., Salem, A. N. and Silverman, G. J. 1995. VH3 family antibodies bind domain D of staphylococcal protein A. J. Immunol. 154:6437. Saul, F. A. and Poljak, R. J. 1993. Structural patterns at residue positions 9, 18, 67 and 82 in the VH framework regions of human and murine immunoglobulins. J. Mol. Biol. 230:15. Fields, B. A., Malchiodi, E. L., Li, H., Ysern, X., Stauffacher, C. V., Chlievert, P. M., Karjalainen, K. and Mariuzza, R. A. 1996. Crystal structure of a T-cell receptor β-chain complexed with a superantigen. Nature 384:188. Eggena, M., Targan, S. R., Iwanczyk, L., Vidrich, A., Gordon, L. K. and Braun, J. 1996. Phage display cloning and characterization of an immunogenetic marker (perinuclear anti-neutrophil cytoplasmic antibody) in ulcerative colitis. J. Immunol. 156:4005. Barbas, C. F., Kang, A. S., Lerner, R. and Benkovic, S. J. 1991. Assembly of combinatorial antibody libraries on phage surfaces: the gene III site. Proc. Natl Acad. Sci. USA 88:7978. Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K. and Pease, L. R. 1989. Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77:51. Horton, R. M., Hunt, H. D., Ho, S. N., Pullen, J. K. and Pease, L. R. 1989. Engineering hybrid genes without the use of restriction enzymes: gene splicing by overlap extension. Gene 77:61. Goodglick, L., Zevit, N., Neshat, M. S. and Braun, J. 1995. Mapping the Ig superantigen-binding site of HIV-1 gp120. J. Immunol. 155:5151. Karray, S. and Zouali, M. 1997. Identification of the B cell superantigen-binding site of HIV-1 gp120. Proc. Natl Acad. Sci. USA 94; 1356. Winter, G., Griffiths, A. D., Hawkins, R. E. and Hoogenboom, H. R. 1994. Making antibodies by phage display technology. Annu. Rev. Immunol. 12:433.

312 VH3 subregions for superantigen binding 32 Karray, S., Juompan, L., Maroun, R. C., Isenberg, D., Silverman, G. J. and Zouali, M. 1998. Structural basis of the gp120 superantigenbinding site on human immunoglobulins. J. Immunol. 161:6681. 33 Chothia, C., Lesk, A. M., Tramontano, A., Levitt, M., Smith-Gill, S. J., Air, G., Sheriff, S., Padlan, E. A., Davies, D., Tulip, W. R., Colman, P. M., Spinelli, S., Alzari, P. M. and Poljak, R. J. 1989.

Conformations of immunoglobulin hypervariable regions. Nature 342:877. 34 Kabat, E. A., Wu, T. T., Reid-Miller, M., Perry, H., Gottesman, K. S. and Foeller, C. 1991. Sequences of Proteins of Immunological Interest. USDHHS, US Government Printing Office, National Institutes of Health, Washington, DC.