Effect of amino acid substitutions in the heavy ... - Semantic Scholar

11 downloads 0 Views 219KB Size Report
Dec 2, 1997 - Abstract. In this study, we applied site-directed mutagenesis to the Fab fragment of a mouse IgM (IE12) that was previously shown to inhibit the ...
International Immunology, Vol. 10, No. 3, pp. 341–346

© 1998 Oxford University Press

Effect of amino acid substitutions in the heavy chain CDR3 of an autoantibody on its reactivity Minou Adib-Conquy, Miche`le Gilbert and Stratis Avrameas Immunocytochimie, CNRS URA 1961, Institut Pasteur, 25 rue du Dr Roux, 75724 Paris Cedex 15, France Keywords: mutagenesis, polyreactivity Abstract In this study, we applied site-directed mutagenesis to the Fab fragment of a mouse IgM (IE12) that was previously shown to inhibit the binding of IgG to autoantigens by interacting with their variable regions. Its native structure was very similar to that of a polyreactive natural IgM (ppc15–30). Indeed, they both use the same light chain and the same VH, D and JH segments. However, the N regions differ and the D is translated in two different reading frames, giving different amino acid compositions of the heavy chain CDR3 (HCDR3). Site-directed mutagenesis modified the HCDR3 of IE12 compared to that of the natural antibody and the resulting effects on its reactivity were analyzed. Because the HCDR3 of IE12 is very rich in aliphatic residues, which are hydrophobic, we replaced them with the more hydrophilic residues of the HCDR3 of the polyreactive IgM. In addition, we evaluated the impact of the proline residues in the HCDR3 of IE12 on its activity, because they are known to restrict backbone flexibility. We found that a more hydrophilic HCDR3 conferred to the IE12 Fab a polyreactive profile. Prolines seem to play an important role in this context, because when they were replaced by glycines, the resulting Fab fragments were highly polyreactive. Our results suggest that, for polyreactivity, hydrophilicity and a certain plasticity of the HCDR3 seem to be necessary. Greater flexibility of the CDR, particularly the HCDR3, might be an important characteristic for polyreactive antibodies. Introduction Among the studies dealing with natural autoantibodies, we previously characterized a population of IgM antibodies that, by interacting with IgG F(ab9)2 fragments, inhibited the latter’s binding to autoantigens such as actin, DNA, myosin and tubulin (1–4) We also generated and characterized monoclonal IgM antibodies having such reactivity (5). One of these antibodies, termed IE12, reacted with F(ab9)2 fragments of mouse IgG and the hapten trinitrophenol (TNP), was shown to inhibit the binding of autoreactive IgG to self-antigens and also bound to B lymphocytes from normal mice (6). We sequenced the variable regions of the heavy and light chains of IE12 and found that they were very similar to those of a highly polyreactive natural IgM (ppc15–30) (7). The light chains used by the two antibodies are λ1 and are 100% homologous in their amino acid composition. Furthermore, they use the same VH–D–JH segments with differences in amino acid composition concentrated in the CDR3 region. In order to study the relationship between the structure and function of IE12 by site-directed mutagenesis, we constructed

an expression system in Escherichia coli that permitted the secretion of the Fab fragment of this IgM antibody into the periplasmic space of the bacteria (8). Our aim was to modify the HCDR3 of IE12, as compared to that of ppc15–30, and to analyze the resulting modifications of its reactivity. In this report, we characterized the bacterially-secreted Fab fragments of this antibody after site-directed mutagenesis and compared them with the unmutated Fab fragment. Finally, we analyzed the effect of pH or ionic strength variations on the reactivity of the mutant Fab fragments, because in a previous study we observed that the reactivities of polyspecific antibodies, in contrast to those of monospecific antibodies, were particularly sensitive to these environmental parameters (9). Methods Bacterial strain and plasmid The plasmid pUC19 and the E. coli strain DH5a were used for the cloning and expression of the Fab fragment of IE12.

Correspondence to: M. Adib-Conquy Transmitting editor: H.-F. Bach

Received 14 September 1997, accepted 2 December 1997

342 Study of polyreactivity by mutagenesis Table 1. Mutagenesis primers Primers

Nucleotide sequence

VH 1 P96→G P104→G D1 D2

59-CTG TGA CAA AAG CCG AGG TGC AGC TGG TGG AGT CTG-39 59-GTG CAA GAC ATG GCC GTA TTA CTA C-39 59-CGG TAG TAG CTG GAT ATT ACT ATG C-39 59-GTA TTA CTG TGC AAG ATG GCA TTA CGG TAG TAG CTC-39 59-CAT TAC GGT AGT AGC TAC TAT GCT ATG GAC-39

Fig. 1. Diagram of the Fab expression vector. The dicistronic operon is under the transcriptional control of the malE promoter; RBS indicates the ribosome-binding sites; ss refers to signal sequences and phoA to alkaline phosphatase.

All DNA manipulations were carried out using standard methodology (10). Fab expression plasmid and mutagenesis The expression system, a dicistronic operon, was constructed in the pUC19 plasmid, with the VHCH1 sequences of IE12 fused to the alkaline phosphatase signal sequence (ssphoA) and VLCL sequences fused to the maltose MalE signal sequence (ssmalE). Both heavy and light chains were put under the control of the malE promoter (pmalE) (Fig. 1). This expression system was described in detail elsewhere (8). Site-directed mutagenesis was achieved using the Transformer site-directed mutagenesis kit (Clontech, Palo Alto, CA). The mutagenic primers are listed in Table 1. Single amino acid replacements were done in one round, while the replacement of the HCDR3 of IE12 by the one of ppc15–30 was done in two steps, first with the D1 primer and then with the D2 primer on a clone carrying the D1 mutation. The selection primers from Clontech were the AatII–EcoRV trans oligo for the first round of mutagenesis and the EcoRV–AatII switch oligo for a second round mutation on the same plasmid. After each mutagenesis, plasmids were extracted from several clones and sequenced using Sanger’s method (11) and Sequenase Version 2 kit (Amersham, Little Chalfont, UK). Those carrying the desired mutation were conserved and used for the expression of the Fab fragment. Expression and purification of the bacterially secreted Fab fragments Cultures of E. coli harboring the plasmids were grown in LB medium containing 100 µg/ml of ampicillin. The bacteria were first grown overnight at 30°C, the cells were then pelleted, and resuspended in LB medium containing 100 µg/ml of ampicillin and 1% maltose. The cells were grown at 25°C and the OD600, initially 1, was monitored until it reached 2. At the end of the culture, the bacteria were pelleted, the supernatant was tested for the presence of Ig fragments and the periplasmic extract was prepared according to Bre´ge´ge`re et al. (12). The Ig frag-

ments were purified by affinity chromatography on a sheep anti-mouse Ig immunoadsorbent (0.85 mg of antibodies/ml of acrylamide-agarose Aca 3.4 beads; column volume: 15 ml). Briefly, the periplasmic extract was concentrated using a Diaflo system (Amicon, Danvers, MA) and incubated for 1 h at room temperature on the immunoadsorbent. The column was then washed with PBS until the OD280 was below 0.05 and bound material was then eluted with 0.2 M glycine–HCl buffer, pH 2.8. The eluted fraction was neutralized with 1 M K2HPO4 and dialyzed extensively against PBS. The Fab fragments, eluted from the immunoadsorbent, were tested by ELISA on plates coated with goat anti-mouse Ig antibodies and trinitrophenol (TNP) coupled to ovalbumin (OVA) using biotin-conjugated goat anti-mouse λ chain antibodies (Southern Biotechnology Associates, Birmingham, AL) at 50 ng/ml in PBS containing 0.1% Tween 20 and 0.5% gelatin (PBST-G) and streptavidin–β-galactosidase (Southern Biotechnology Associates) (for more details, please see the ELISA section). If the quantity and the activity were satisfactory they were coupled to biotin (13). Briefly, the Fab fragments were dialyzed overnight against 0.1 M sodium bicarbonate in a final volume of 1 ml. d-Biotin-N-hydroxysuccinimide ester (Sigma, St Louis, MO) was diluted at 0.1 M in dimethylformamide and 10 µl/ml of Fab were added, mixed and incubated for 1 h at room temperature. The mixture was then dialyzed overnight against PBS and stored with 50% glycerol at 220°C. Antigens Mouse muscle actin and myosin, brain tubulin, and thyroglobulin were prepared according to described methods (14– 17). TNP, arsonate and phenyloxazolone were coupled to OVA (18). p-Aminophenyltrimethylammonium hapten was coupled to BSA (TMA-BSA) (19). IgG F(ab9)2 fragments were purchased from Jackson Laboratories (West Grove, PA); double-stranded DNA type I from calf thymus, sperm whale myoglobin and lipopolysaccharide (LPS) were obtained from Sigma. Laminin was prepared from mouse red blood cells as described (20). ELISA Polystyrene microtiter plates (Maxisorp; Nunc, Roskilde, Denmark) were coated with antigen or goat anti-mouse Ig antibodies at 5 µg/ml in 0.1 M carbonate-bicarbonate buffer, pH 9.5, for 1 h at 37°C. The wells were saturated with 0.5% gelatin in PBS for 1 h at 37°C, and washed thoroughly with PBS containing 0.1% Tween 20 (PBS-T). Biotin-coupled Fab fragments diluted in PBS-T-G were incubated with coated plates for 2 h at 37°C. After five washes with PBS-T, the plates were incubated for 1 h at 37°C with streptavidin–βgalactosidase diluted 1/1000 for 1 h at 37°C. After five washes,

Study of polyreactivity by mutagenesis 343

Fig. 2. Nucleic acid (A) and amino acid (B) compositions of the heavy chain CDR3 of IE12 compared to that of the natural polyreactive IgM ppc15–30.

the enzyme activity was revealed with a saturated solution of 4-methylumbelliferyl-β-D-galactopyranoside (Sigma) in PBS (21). After 1 h, the reaction was stopped with 2 M K2CO3 and the fluorescent signal was read using a Fluoroskan ELISAplate reader (Flow, McLean, VA). Inhibition experiments and Kd calculation Purified Fab fragments, at a concentration giving 50% of the maximal fluorescence in titration curves, were mixed with increasing amounts of antigen diluted in PBS-T-G. After 16 h at room temperature, the mixtures were transferred into microtiter plates coated with the same antigen and saturated with gelatin. The plates were incubated for 2 h at room temperature. After five washes, the antibodies bound to the solid-phase antigen were detected by ELISA. The percent inhibition was calculated in comparison to the binding in the absence of inhibitor and Kd were calculated from Klotz plots as described (22). Effect of pH and ionic strength The Fab fragments were titered in a classical ELISA test in PBS-T-G and a concentration of antibody in the linear part of the titration curve was chosen for each of them. The antibody activity was measured at different pH and NaCl concentrations for the Fab–antigen incubation step. For pH experiments, the 10 mM phosphate buffer contained 0.15 M NaCl adjusted to pH 5.4, 6.4, 7.4, 8.4 or 9.4. For the ionic strength experiments, the Fab were diluted in 10 mM phosphate buffer, pH 7.1, containing 0, 0.03, 0.15, 0.3 or 0.6 M NaCl. The Fab were incubated on the plates for 2.5 h at room temperature. All the subsequent steps were the same as those described above for ELISA experiments. Results Mutant reactivity As shown in Fig. 2, the HCDR3 of IE12 contains two prolines, one in each N region. The DH used by this antibody is the same as that of the natural polyreactive IgM ppc15–30 but it

is translated in two different reading frames and bordered by different N regions, thus leading to completely different amino acid compositions. Our first aim was to evaluate the impact of the proline residues in the IE12 HCDR3 on its activity, keeping in mind that this amino acid is known to restrict backbone flexibility (23,24) (mutants P96→G, P104→G and P96/104→G). Furthermore, because the IE12 HCDR3 is very rich in aliphatic residues that are hydrophobic, we replaced them by those of the ppc15–30 HCDR3 which are more hydrophilic (mutant Dmut). The four mutants obtained were characterized by ELISA on a panel of self antigen, haptens and LPS, and compared to the Fab fragment corresponding to the wild-type antibody. Figure 3 shows their reactivity profiles at one concentration (100 ng/ml) chosen because it is in the linear part of their titration curves. Absolute fluorescence values are shown in Fig. 3(A), while the percentages of the wild-type reactivity, which was considered to be 100% activity, are presented in Fig. 3(B). It can be seen that replacing one proline by a glycine (P96→G and P104→G) gave rise to Fab fragments showing polyreactive profiles and strong activities against myoglobin and myosin. The P96→G mutant had considerably higher anti-TMA activity (7.5 times) and showed moderate binding to actin, DNA and LPS. The P104→G mutant showed a very similar profile, however, with stronger reactivity with myoglobin and a lower binding to TMA. The replacement of both prolines by glycines (P96/104→G) did not increase the number of antigen recognized by this antibody, but gave a mutant having a different reactivity profile. Indeed, P96/104→G reacted strongly with myosin, and moderately with actin, myoglobin and TMA. Finally, Dmut was also polyreactive, reacting like the single proline mutant P96→G, strongly with myoglobin and more moderately with myosin, DNA and TMA. Inhibition experiments and dissociation constant (Kd) Inhibition experiments were performed on the two antigens strongly recognized by mutant Fab fragments, i.e. myoglobin and myosin. These experiments could not be performed on the wild-type Fab fragment, because its reactivity with these

344 Study of polyreactivity by mutagenesis

Fig. 3. Reactivity of the bacterially expressed Fab fragments of IE12 tested by ELISA. Results are presented as fluorescence values obtained on the various tested antigens (A) or as a fraction of the values of the wild-type (WT) Fab which was taken as 100% (B).

Fig. 4. Inhibition of the binding of the mutant Fab on myoglobin (A) and myosin (B) by the same soluble antigen tested by ELISA.

Table 2. Affinity of the mutant Fab fragments two antigens was too weak. The binding of the mutant Fab fragments to immobilized myosin and myoglobin could be inhibited by the corresponding soluble antigen in a dosedependent manner (Fig. 4). Their Kd were calculated and are given in Table 2. The values obtained on myosin were not very different for all the mutants. The Kd obtained for P104→G and Dmut were similar on both antigen, while the affinity of P96→G for myoglobin was 4 and 7 times lower respectively when compared to P104→G and Dmut. P96/104→G had the lowest affinities for both antigens, especially myoglobin. Effect of pH or ionic strength In a previous study comparing poly- and monospecific IgG mAb we observed that polyreactive mAb were more sensitive to environmental changes, such as pH and NaCl concentration variations, than monoreactive ones (9). We have conducted the same type of experiments on the mutant Fab fragments in order to compare their behavior with that of intact polyreactive antibodies. Figure 5(A and B) shows the reactivities of the mutants at various pH on myoglobin and myosin. One can see that reactivity decreased as a function of increasing pH for almost all of them, especially on myoglobin. Indeed, except for P104→G which exhibited its lowest activity at pH 5.4, mutants bound most strongly to the antigen at acidic pH. When tested at various NaCl concentrations (Fig. 5C and D), all the Fab fragments showed the same profile, with maximum

Mutant

P96→G P104→G P96/104→G Dmut

Kd (3 10–6 M) Myoglobin

Myosin

10.90 2.75 56.00 1.54

0.47 0.71 4.00 1.75

binding to myoglobin at 0.03 M and to myosin at 0.15 M NaCl, and decreasing reactivity with increasing salt concentrations. Discussion In this study, we applied site-directed mutagenesis to the HCDR3 region of an autoantibody (IE12), reacting with IgG F(ab9)2 fragments and TNP,. in an attempt to transform its activity towards polyreactivity. For this purpose, we used an expression system in E. coli and generated Fab fragments corresponding to each mutant. This expression system was previously described in detail (8). The reactivity of the native IE12 Fab fragment was similar to that of the whole IgM. However, the affinity of the Fab was lower than that of the intact antibody for IgG F(ab9)2 fragments, while it was similar

Study of polyreactivity by mutagenesis 345

Fig. 5. Reactivities of the mutant Fab fragments of IE12 tested on myoglobin and myosin at different pH (A and B) and NaCl concentrations (C and D).

for TNP. This difference can be explained by the fact that the global avidity of an IgM is higher than the affinity of its paratopes because of its pentameric structure. However, the multivalent structure does not seem to be important for the binding to a small molecule such as the hapten TNP, since the affinity of the Fab was comparable to that of the IgM. Our present data indicate that a more hydrophilic composition of the HCDR3 conferred to the Fab a polyreactive profile, whereas the reactivity of a Fab containing a hydrophobic HCDR3 rich in aliphatic residues was more restricted. Furthermore, prolines seem to play an important role in this context, because their presence at each end of the HCDR3 would restrict the antibody’s flexibility and, in turn, its reactivity. Indeed, when these prolines were replaced by glycines, the resulting Fab fragments exhibited highly polyreactive profiles. Several studies have previously shown the crucial role of the HCDR3 for antibody reactivity, particularly for polyreactivity (7,25). Experiments of CDR shuffling between a mono- and a polyreactive mAb have confirmed the importance of this region. Ichiyoshi and Casali showed that just by transferring the HCDR3 of an antibody they could obtain chimeric mAb having exactly the same reactivity as the donor HCDR3 (26). This was true for both a mono- and a polyreactive mAb. Crouzier et al., however, obtained a slightly different result. In their example only the HCDR3 of the monospecific mAb could transfer the activity; the polyreactive HCDR3 associated with the VH of the monospecific mAb and its own light chain did not show any activity (27).

In our experiments we wanted to elucidate the importance of the HCDR3 composition or its characteristics in polyreactivity. A comparative study on poly- and monospecific antibodies concluded that the CDR of polyreactive antibodies contained more arginine and lysine, and thus could bind to various charged antigen (28); however, similar conclusions were drawn from the study on lupus monospecific anti-DNA antibodies whose CDR are also mostly rich in cationic residues (29). In our example, the hydrophilicity of the HCDR3 seems to be more important than its content in charged residues. Indeed, the HCDR3 from ppc15–30 did not contain charged residues and nonetheless was able to confer polyreactivity to IE12. Another important point was the study of the prolines bordering the HCDR3. This amino acid is known to reduce the flexibility of the polypeptidic chain and could be important for the capacity of an antibody to be multireactive or not. We replaced the prolines by glycines that, on the contrary, confered a high flexibility because of their absence of a lateral chain (30). Indeed, the resulting mutants showed a polyreactive profile. However, when both prolines were replaced, the reactivity of the Fab was different from that of single mutants and the affinities calculated for two antigens were also lower, perhaps because with two glycines the plasticity of this region could be still further enhanced and result in a loss of affinity. In a previous study on polyreactive IgG, our data suggested that polyreactive IgG might have a more plastic structure than monoreactive IgG (9). In contrast

346 Study of polyreactivity by mutagenesis to monospecific IgG, their reactivities were considerably modified by environmental changes, such as pH or ionic strength. In the present study, we showed that this sensitivity is also true for polyreactive Fab fragments, and that they behave like intact polyreactive IgG molecules at various pH and NaCl concentrations. The polyreactivity of Fab fragments has also been shown to be sensitive to temperature modifications (31). Finally, our results also support the existance of a multireactive antigen binding site, because the modification of one residue resulted in the capacity to bind to several antigens. In conclusion, our results indicate that polyreactivity requires hydrophilicity and a certain plasticity of the HCDR3. Greater flexibility of the CDR, particularly the HCDR3, might be an important characteristic for polyreactive antibodies. Acknowledgements We thank J. Jacobson for rereading the manuscript. This work was supported by grant no. 6257 from the Association pour la Recherche sur le Cancer.

Abbreviations HCDR3 LPS OVA TMA TNP

complementarity-determining region 3 of the heavy chain lipopolysaccharide ovalbumin trimethylammonium trinitrophenol

References 1 Karsenti, E., Guilbert, B., Bornens, M. and Avrameas, S. 1977. Antibodies to tubulin in normal non immunized animals. Proc. Natl Acad. Sci. USA 74:3997. 2 Lutz, H. V. and Wipf, G. 1982. Naturally occurring autoantibodies to skeletal proteins from human red blood cells. J. Immunol. 128:1695. 3 Avrameas, S. 1991. Natural autoantibodies: from ‘horror autotoxicus’ to ‘gnothi seauton’. Immunol. Today 12:154. 4 Adib, M., Ragimbeau, J., Avrameas, S. and Ternynck, T. 1990. IgG autoantibody activity in normal mouse serum is controlled by IgM. J. Immunol. 145:3807. 5 Adib-Conquy, M., Avrameas, S. and Ternynck, T. 1993. Monoclonal IgG and IgM obtained after polyclonal activation, show reactivities similar to those of polyclonal natural autoantibodies. Mol. Immunol. 30:119. 6 Adib-Conquy, M., Gilbert, M., Christodoulou, C. and Avrameas, S. 1994. Reactivity and structure of a mouse anti-F(ab9)2 IgM. Comparison of its variable region sequences with those of a structurally close polyreactive natural IgM. Mol. Immunol. 31:555. 7 Chen, C., Stenzel-Poore, M. P. and Rittenberg, M. B. 1991. Natural auto- and polyreactive antibodies differing from antigen induced antibodies in the H chain CDR3. J. Immunol. 147:2359. 8 Adib-Conquy, M., Gilbert, M., Christodoulou, C. and Avrameas, S. 1995. Bacterial secretion of the Fab fragment of a mouse monoclonal IgM that reacts with IgG variable regions. Prot. Eng. 8:859. 9 Labrousse, H., Adib-Conquy, M. and Avrameas, S. 1994. Effects of pH or ionic strength on the reactivities of mono- and polyspecific IgG antibodies. Res. Immunol. 145:541. 10 Sambrook, J., Fritsch, E. F. and Maniatis, T. 1989. Molecular Cloning: A Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

11 Sanger, F., Nicklen, S. and Coulson, A. R. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl Acad. Sci. USA 74:5463. 12 Bre´ge´ge`re, F., Schwartz, J. and Bedouelle, H. 1994. Bifunctional hybrids between the variable domains of an immunoglobulin and the maltose-binding protein of Escherichia coli; production, purification and antigen binding. Prot. Eng. 7:271. 13 Guesdon, J.-L., Ternynck, T. and Avrameas, S. 1979. The use of avidin–biotin interaction in immunoenzymatic techniques. J. Histochem. Cytochem. 27:1131. 14 Spudich, J. A. and Watt, S. 1971. The regulation of rabbit skeletal muscle contraction. J. Biol. Chem. 246:4866. 15 Whalen, R. G., Butler-Browne, G. S. and Gros, F. 1978. Identification of a novel form of myosin light chain present in embryonic muscle tissue and cultured muscle cells. J. Mol. Biol. 126:415. 16 Shelanski, M. L., Gaskin, F. and Cantor, C. 1973. Microtubule assembly in the absence of added nucleotides. Proc. Natl Acad. Sci. USA 70:765. 17 Rolland, M. and Lissitsky, S. 1976. Endogenous proteolytic activity and constituent polypeptide chains of sheep and pig 19 S thyroglobulin. Biophys. Acta 427:696. 18 Little, J. R. and Eisen, H. N. 1976. Preparation and characterization of antibodies specific for 2,4,6-trinitrophenyl group. Biochemistry 5:3385. 19 Chesebro, B. and Metzger, H. 1972. Affinity labeling of a phosphorylcholine binding mouse myeloma protein. Biochemistry 11:766. 20 Timpl, R., Rohde, H., Robey, P. G., Rennard, S. I., Foidart, J. M. and Martin, G. R. 1979. Laminin. A glycoprotein from basement membranes. J. Biol. Chem. 254:9933. 21 Labrousse, H. and Avrameas, S. 1987. A method for the quantification of a colored or fluorescent signal in enzyme immunoassays by photodensitometry. J. Immunol. Methods 103:9. 22 Friguet, B., Chaffotte, A. F., Djavadi-Ohaniance, L. and Goldberg, M. E. 1985. Measurements of the true affinity constant in solution of antigen–antibody complexes by enzyme-linked immunoadsorbent assay. J. Immunol. Methods 77:305. 23 Vermersch, P. S., Tesmer, J. J. G., Lemon, D. D. and Quiocho, F. A. 1990. A Pro to Gly mutation in the hinge of the arabinose-binding protein enhances binding and alters specificity. J. Biol. Chem. 265:16592. 24 MacArthur, M. W. and Thornton, J. M. 1991. Influence of proline residues on protein conformation. J. Mol. Biol. 218:397. 25 Martin, T., Duffy, S. F., Carson, D. A. and Kipps, T. J. 1992. Evidence for somatic selection of natural autoantibodies. J. Exp. Med. 175:983. 26 Ichiyoshi, Y. and Casali, P. 1994. Analysis of the structural correlates for antibody polyreactivity by multiple reassortments of chimeric human immunoglobulin heavy and light chain V segments. J. Exp. Med. 180:885. 27 Crouzier, R., Martin, T. and Pasquali, J.-L. 1995. Heavy chain variable region, light chain variable region, and heavy chain CDR3 influences on the mono- and polyreactivity and on the affinity of human monoclonal rheumatoid factors. J. Immunol. 154:4526. 28 Gonzalez-Quintial, R., Baccala, R., Alzari, P. M., Nahmias, C., Mazza, G., Fougereau, M. and Avrameas, S. 1990. Poly (Glu60Ala30Tyr10) (GAT)-induced IgG monoclonal antibodies crossreact with various self and non-self antigens through the complementarity determining regions. Comparison with IgM monoclonal polyreactive natural antibodies. Eur. J. Immunol. 20:2383. 29 Datta, S. K., Patel, H. and Berry, D. 1987. Induction of a cationic shift in IgG anti-DNA autoantibodies. Role of T helper cells with classical and novel phenotypes in three murine models of lupus nephritis. J. Exp. Med. 165:1252. 30 Ramachandran, G. N. and Sasisekhara, V. 1969. Conformation of polypeptides and proteins. Adv. Prot. Chem. 23:283. 31 Ditzel, H. J., Itoh, K. and Burton, D. R. 1996. Determinants of polyreactivity in a large panel of recombinant human antibodies from HIV-1 infection. J. Immunol. 157:739.