Broadly distributed chemical reactivity of natural antibodies expressed in coordination with specific antigen binding activity.

Antibody (Ab) nucleophilic reactivity was studied using hapten and polypeptide antigens containing biotinylated phosphonate diester groups (covalently reactive antigen analogs, CRAs). Polyclonal IgG from healthy donors formed covalent adducts with a positively charged hapten CRA at levels superior to trypsin. Each of the 16 single chain Fv clones studied expressed a similar reactivity, indicating the V domain location of the nucleophiles and their broad distribution in diverse Abs. The formation of hapten CRA-Fv adducts was correlated with Fv proteolytic activity determined by cleavage of a model peptide substrate. Despite excellent nucleophilicity, proteolysis by IgG proceeded at lower rates than trypsin, suggesting that events occurring after nucleophilic attack on the substrate limit the rate of Ab proteolysis. The extracellular domain of the epidermal growth factor receptor with phosphonate diester groups at Lys side chains and a synthetic peptide corresponding to residues 421- 431 of human immunodeficiency virus glycoprotein (gp) 120 with the phosphonate diester at the C terminus formed covalent adducts with specific polyclonal and monoclonal Abs raised by immunization with epidermal growth factor receptor and synthetic gp120-(421- 436) devoid of phosphonate diester groups, respectively. Adduct formation was inhibited by extracellular domain of the epidermal growth factor receptor (exEGFB) and synthetic gp120-(421- 436) devoid of phosphonate groups, suggesting that the nucleophiles are located within the antigen binding sites. These results suggest the innate character of the Ab nucleophilic reactivity, its functional coordination with non-covalent adaptive binding interactions developing over the course of B cell maturation, and novel routes toward permanent inhibition of Abs.

Many enzymes exploit covalent interactions with substrates to catalyze chemical transformations. On the other hand, most studies on Ab 1 catalysis have focused on non-covalent binding forces as the mechanism by which the energy barrier between reactants and products is lowered, e.g. the electrostatic forces that stabilize the negatively charged oxyanionic transition state of ester hydrolysis (reviewed in Refs. 1 and 2). The underlying assumption has been that Abs interact with their ligands exclusively by non-covalent means. Initial indications that natural Abs express chemical reactivity indistinguishable from enzymes came from reports of proteolytic and nuclease activity of autoantibodies (3,4). Similar activities were later found in Ab light chains from multiple myeloma patients (5), alloantibodies from patients with transfusion-induced hemophilia (6), Abs raised by routine immunization with polypeptides (7,8), and anti-idiotypic Abs to anti-enzyme Abs (9). From mutagenesis and inhibitor studies, it appears that the proteolytic activity of natural Abs originates from nucleophilic mechanisms similar to those utilized by conventional serine proteases (10,11). The catalytic activity of natural Abs could be construed to violate the principles of B cell clonal selection theory. Antigen-specific Abs develop over the course of the immune response by sequence diversification of germ line genes encoding the V domains followed by selective antigen binding by B cell receptors with the greatest affinity, which stimulates clonal proliferation. Catalysis entails chemical transformation of the antigen and release of products (surface immunoglobulins associated with Ig␣ and Ig␤ subunits), which is predicted to result in the cessation of B cell selection. Therefore, adaptive selection of Ab catalytic activity may be a disfavored event. For this reason, catalysis by naturally occurring Abs is often assumed to be a molecular accident arising from sequence variability of the V domains as opposed to a general phenomenon with functional implications.
The foregoing restrictions do not apply to the initial step in the catalytic cycle of serine proteases. In analogy with conventional enzymes, a nucleophile belonging to a proteolytic Ab ( Fig  1A, Nu) is conceived to initiate nucleophilic attack on the antigen following the formation of the non-covalent ground state complex. Adaptive development of Ab nucleophilicity is fully compatible with B cell clonal selection if the outcome is the formation of a covalent acyl-Ab complex as occupancy of the B cell receptor will be maintained. Whether the catalytic cycle is completed depends on the efficiency of hydrolysis of the acyl-Ab complex and release of the product. Recently, hapten phosphonate esters have been developed as probes for covalent binding to the active site nucleophiles in Abs displaying serine protease and serine esterase activity (11,12) (designated CRAs). These compounds can be applied for direct study of Ab chemical activity independent of additional activities needed for accomplishment of catalysis. In addition, the phosphonates can be placed within peptides and proteins (Fig. 1, B and C) for studying the interplay between Ab nucleophilic reactivity and non-covalent forces permitting specific recognition of individual polypeptides.
We describe here observations suggesting the broad distribution of nucleophilic reactivity in IgG and recombinant Fv preparations at levels exceeding that of the conventional serine protease trypsin. Originally prepared as probes for catalytic Abs, CRA analogs of EGFR and an HIV gp120-synthetic peptide were observed to form covalent adducts with ordinary Abs raised by immunization with antigens devoid of the phosphonate groups, suggesting that adaptive maturation processes favor the expression of nucleophilic reactivity. These observations argue for Ab nucleophilicity as a force responsible for shaping the expressed Ab repertoire and suggest novel routes toward permanent inactivation of Abs.
Irreversible CRA Binding-Following incubation of biotinylated CRAs with Abs or trypsin (porcine, type IX, Sigma) in 50 mM Tris, HCl, 100 mM glycine, 0.1 mM CHAPS, pH 7.7, at 37°C, the reaction mixtures were boiled (5 min) in 2% SDS and subjected to SDS-PAGE (4 -20%, Bio-Rad, or 8 -25% Phast gels, Amersham Biosciences). Electroblotting and biotin detection procedures using streptavidin-horseradish peroxidase and a chemiluminescent substrate (Supersignal, Pierce) have been described previously (11). Imaging and quantification was on x-ray film (Eastman Kodak Co.) using Unscan-it software (Silk Scientific, Orem, UT) or Fluoro-STM MultiImager (Bio-Rad). Band intensities are expressed in arbitrary area units (AAU). Valid comparisons of band intensities from different experiments is not possible because exposure and development times were not held constant. Diisopropyl fluorophosphate (DFP) (Sigma) was kept at 4°C until used. In some experiments, biotinylated BSA (8 mol of biotin/mol protein, Pierce) was electrophoresed at several concentrations in parallel with the samples and the biotin content of the CRA adducts was determined. Pseudo-first order rate constants (k obs ) were computed from reaction progress curves by fitting to the equation B t ϭ B max (1 Ϫ exp(Ϫk obs t)), where B t represents adduct concentration at various times and B max represents the initial Ab concentration. Immunoblotting with goat anti-mouse IgG Abs was as described previously (7).
Proteolysis Assay-Catalytic activity was measured by fluorimetric determination ( ex 360 nm, em 470 nm; Varian Cary Eclipse) of the cleavage of amide bond linking aminomethylcoumarin to the C-termi-nal amino acid in short peptide-MCA substrates (Pro-Phe-Arg-MCA, Boc-Glu-Ala-Arg-MCA, and Boc-Ile-Glu-Ala-Arg-MCA (200 M), Peptide International) (10). Catalysts were incubated with peptide-MCA substrates in 50 mM Tris HCl, 0.1 M glycine, 0.025% Tween 20, pH 8.0, at 37°C in 96-well plates. In some assays, a comparison of IgG and trypsin proteolytic activity was done in 10 mM sodium phosphate, pH 7.4, 0.137 M NaCl, 2.7 mM KCl, 0.1 mM CHAPS. Authentic aminomethylcoumarin (Peptide International) was used to construct a standard curve from which product release was computed in molar values.

Ab Nucleophilicity Identified with Hapten CRAs-Phospho-
nate hapten CRAs I-III ( Fig. 1) are analogs of known active site-directed inhibitors of serine proteases (19). Similar to the serine protease trypsin, IgG from a healthy human subject formed adducts with CRA I that were resistant to boiling and the denaturant SDS (Fig. 2, IgG, 150-kDa adducts; trypsin, 21-kDa adducts). Pooled IgG from immunologically unmanipu-lated BALB/c mice formed similar CRA I adducts. The positively charged amidino group in CRA I was originally incorporated in this compound to allow selective recognition of trypsin, which displays preference for basic residues at the P1 site (the residue immediately adjacent to the cleavage site in peptide substrates) (for review see Ref. 20). CRA II lacks the positively charged amidino group adjacent to the covalently reactive phosphorus atom. IgG was 240-fold less reactive with CRA II than with CRA I, suggesting the trypsin-like P1 specificity of Abs. CRA III, which contains a weaker leaving group than CRA II did not form detectable adducts with IgG (the presence of methoxy-leaving groups reduces the electrophilicity of the phosphorus atom, and methoxy-containing phosphonate diesters are reported to bind weakly with certain serine proteases) (for review see Ref. 21). Increasing formation of covalent CRA I adducts with IgG and trypsin was evident as a FIG. 1. A and B, reaction of serine protease-like Abs with antigens and phosphonate diester-containing antigen analogs, respectively. C, CRA structures. In A, Nu denotes a nucleophile. AgЈ and AgЉ are components of the antigenic epitope at which non-covalent contact with the Ab occurs; AgЈ-Lys-OH is the N-terminal antigen fragment; and NH 2 -AgЉ is the C-terminal antigen fragment. The active site nucleophile attacks the carbonyl carbon of the scissile bond in the antigen (substrate) to form the tetrahedral transition-state complex. The C-terminal antigen fragment is released, and the acyl-Ab complex is formed. Hydrolysis of the acyl-Ab complex results in the release of the N-terminal antigen fragment and regeneration of the catalytic Ab. The catalytic rate constant k cat is the sum of k 3Јϩk3Љ . In B, the Ab nucleophile attacks the electrophilic phosphonate diester (instead of the carbonyl group) and the phosphonate-containing antigen (CRA) recapitulates the remaining interactions in the ground and transition state Ab-Ag complex (non-covalent binding at peptide epitopes), but unlike the acyl-Ab intermediate, the phosphonyl-Ab adduct is a stable product. In C, CRA I is an active site-directed inhibitor of trypsin-like enzymes. CRAs II and III are CRA I derivatives devoid of the side chain amidino function. CRA III contains the weaker methoxy leaving group. CRAs IV and V are intended, respectively, to permit detection of nucleophiles in specific Abs to exEGFR and residues 421-431 of gp120. The biotin and phosphonate diester groups were incorporated in CRA IV at Lys side chains. CRA Va contains biotin at the N terminus and the phosphonate diester at the C terminus. function of reaction time (Fig. 2B). The velocity of the reaction for IgG was 14.5-fold greater than for trypsin measured under identical conditions (172.7 Ϯ 14.2 and 11.9 Ϯ 0.6 AAU/min, respectively, from linear regression of Fig. 2B data). Assuming that hydrolysis of the phosphonylated-protein complex is equivalent (see reaction schemes in Fig. 1), it may be concluded that the nucleophilic efficiency of IgG is superior to that of trypsin.
IgG preparations from healthy humans and immunologically unmanipulated mice have been documented to cleave small model peptide substrates on the C-terminal side of basic residues. The cleavage activity was observed in each of several IgG preparations examined. The activity co-migrated with intact 150-kDa IgG in denaturing gel filtration studies, and it was expressed by Fab (fragment antigen binding) preparations made by papain digestion (22). In this study, we compared the proteolytic activity of trypsin and IgG from a healthy human subject (the same preparation as in nucleophilicity studies illustrated in Fig. 2). With Glu-Ala-Arg-MCA and Pro-Phe-Arg-MCA substrates, initial rates of proteolysis by IgG were, respectively, 1.8 ϫ 10 5 -and 6.8 ϫ 10 5 -fold smaller than by trypsin (Fig. 3, A and B, determined from the slopes of the progress curves). Glu-Ala-Arg-MCA is the preferred substrate for trypsin. Glu-Ala-Arg-MCA and Pro-Phe-Arg-MCA are the preferred substrates for human IgG determined from previous screening of a panel of peptide-MCA substrates (22). The magnitude of proteolysis by this IgG preparation falls within the range reported previously for other human IgG preparations. Despite its superior nucleophilic reactivity, the IgG is evidently a poor catalyst compared with trypsin.
CRA I and DFP (another active site-directed inhibitor of serine proteases) inhibited the catalytic activity of IgG-catalyzed peptide-MCA cleavage (Fig. 3C), and DFP inhibited the irreversible binding of CRA I by the IgG (by 95%). These results provide assurance that CRA I binds the catalytic sites of IgG. As DFP binds the active site of serine proteases, its inhibitory effect confirms the serine protease character of the CRA I binding sites of IgG. Electrophoresis of CRA I-IgG adducts under reducing conditions revealed labeling of both subunits by the hapten CRA, evident as biotin-containing bands at 50-kDa heavy chain bands and 25-kDa light chain bands (Fig.  3D). Irreversible CRA I binding activity of IgG was lost by preheating the protein at 60°C for 10 min, indicating the dependence of the nucleophilic reactivity on the native protein conformation.
Each of the five polyclonal IgG preparations from healthy humans displayed irreversible binding to CRA I (Table I) Fig. 4). Analyzed by electrophoresis under non-reducing conditions, some scFv reaction mixtures contained CRA I adducts at 55-90 kDa in addition to the monomer scFv adducts at 27 kDa. All of the CRA-adduct bands were also stainable with Ab to c-Myc, confirming the presence of scFv in the adducts (the recombinant proteins contain a 10-residue c-Myc peptide) (10). The tendency of scFv to form aggregates has been reported previously (23). Diminished levels of CRA I-adducts were detected when an scFv clone was treated with DFP prior to CRA I treatment (by 72%). The rate of covalent adduct formation by different Fv clones was variable over a 34-fold range (Table I), indicating distinct levels of nucleophilic reactivity of different Abs. The reactivity of the five polyclonal IgG samples, which represent mixtures of different Abs, was less variable (by 5.4). A comparison of the peptide-MCA cleaving activity (Glu-Ala-Arg-MCA substrate) and irreversible CRA I binding by the scFv clones indicated a strong correlation (p Ͻ 0.005, r 2 ϭ 0.77) (Fig. 4B), confirming the functional importance of superior nucleophilic reactivity.
Specific Covalent Binding of Peptidyl and Protein CRA-Protein CRA IV and peptide CRA Va were analyzed to assess whether antigen-specific Abs can express nucleophilic reactivity coordinated with non-covalent recognition of the antigen. CRA IV is the extracellular domain of a tumor-associated protein, exEGFR, presenting diverse antigenic epitopes derivitized at Lys side chains with the phosphonate diester (19 mol/mol) along with a small amount of biotin to allow the detection of adducts. SDS-electrophoresis of CRA IV revealed a major silver-stained and biotin-containing band with a nominal mass of 90 kDa (mass of exEGFR, 85 kDa; mass of hapten phosphonate group, 714 Da). CRA V corresponds to residues 421-431 of the HIV coat protein gp120 along with the amidino surrogate of Lys-432 and the covalently reactive phosphonate diester group located at the C terminus. The purity and chemical characterization of this peptidyl CRA has been reported previously (18). Abs raised by routine immunization with exEGFR and the synthetic peptide corresponding to residues 421-436 of gp120 were initially employed to assure the antigenic integrity of these CRAs. ELISA studies indicated that the binding of CRAs IV and Va (conjugated to BSA) by polyclonal Abs to exEGFR and synthetic gp120-(421-436), respectively, was only marginally lower than of the control antigens devoid of phosphonate diester groups, i.e. exEGFR and gp120-(421-436), respectively (Fig. 5). Evidently, the epitope structure of the two antigens is preserved despite the introduction of the phosphonate diester in Lys side chains (CRA IV) and at the C terminus (CRA Va). No binding of anti-exEGFR or anti-gp120-(421-436) Abs to immobilized calmodulin and albumin was detected (A 490 Ͻ 0.05 at antisera dilution 1:1000), confirming the absence of nonspecific protein binding effects. Immobilized CRA IV and CRA Va (conjugated to BSA) did not display unusual binding to nonimmune Abs used as controls for ELISA, indicating that the phosphonate diester group does not result in indiscriminate covalent binding effects.
Covalent binding by the Abs was studied using denaturing electrophoresis as described for the hapten CRAs. Saturable formation of biotin-containing IV adducts with Abs to exEGFR was evident (nominal mass of 250 kDa). CRA IV adducts of non-immune IgG were not detectable (Fig. 6). Because CRA IV concentration is small (0.2 M in Fig. 6), the formation of adducts similar to those observed using hapten CRA I is not predicted (Fig. 2, 100 M CRA I). Few or no adducts were formed in the presence of exEGFR (1 M), but adduct formation was not impeded by an equivalent concentration of calmodulin, indicating that the covalent binding reaction is at or near the antigen binding site of the Abs. The 250-kDa CRA IV adducts were stainable with anti-IgG (data not shown). Each of three commercially available monoclonal Abs to exEGFR formed covalent adducts with CRA IVa (according to the suppliers, Ab C225 binds residues 351-364 in the extracellular domain of EGFR; the linear peptide determinant recognized by Abs H11 and C111.6 is not known, but both Abs bind the extracellular domain of the protein), an irrelevant monoclonal Ab did not, and the formation of the adducts by the monoclonal Abs was inhibited by exEGFR devoid of phosphonate diester groups but not by the unrelated protein calmodulin. Essentially, similar results were obtained using CRA Va (Fig. 7). The formation of biotin-containing 152-kDa adducts was saturable as a function of time (mass of CRA Va, 2.2 kDa), adduct formation was inhibited by the gp120-(421-436)-BSA conjugate (3 M) but not an equivalent concentration of BSA, and the reaction with non-immune Abs proceeded slowly compared with the specific Abs.
The pseudo-first order rate constant k obs for accumulation of CRA IV adducts of polyclonal IgG to exEGFR was 1.0 Ϯ 0.1 h Ϫ1 . Because no reaction was detected with non-immune IgG, a precise estimate of k obs is not possible. Using the detection sensitivity of the imaging system as the upper limit for accumulation of adducts over the period of observation in Fig. 6 (133 AAU), the upper limit for k obs is 7.2 ϫ 10 Ϫ3 h Ϫ1 . Similarly, k obs for accumulation of anti-peptide IgG adducts of CRA Va was 496-fold greater than that of non-immune IgG adducts (17.8 Ϯ 3.3 h Ϫ1 and 0.4 ϫ 10 Ϫ1 Ϯ 0.1 ϫ 10 Ϫ1 h Ϫ1 , respectively) (see data in Fig. 7). DISCUSSION Activated nucleophilic residues in conventional serine proteases react covalently with phosphonate diester probes, e.g. the Ser residue activated by hydrogen bonding in the catalytic Ser-His-Asp triad of serine proteases. The presence of such nucleophiles in proteolytic and esterolytic Abs has been deduced from mutagenesis and covalent phosphonate binding studies (10 -12). Nucleophilic attack on the substrate is the rate-limiting step in catalysis by certain enzymes (24). As the reported catalytic rate constants (k cat ) of Abs are generally orders of magnitude lower than those of enzymes, it has generally been assumed that the deficiency resides in the nucleophilic reactivity of Abs. Studies reported here indicate otherwise. Despite their low proteolytic activity, IgG preparations displayed stronger nucleophilic reactivity than trypsin determined from rates of the formation of covalent adducts with hapten phosphonate diesters. The study of polyclonal IgG and individual scFv clones indicated an apparently universal nucleophilic reactivity. In control experiments, the reactivity was lost upon thermal denaturation, consistent with expectations that activation of the nucleophile is dependent on the native structure of the protein. Covalent Ab binding to the phosphonate diester was inhibited by the established serine proteasereactive reagent DFP. Moreover, Ab proteolytic activity was inhibited by the phosphonate as well as DFP, confirming the serine protease-like character of nucleophiles reactive with the phosphonate. These studies suggest nucleophilic reactivity as an intrinsic property of Abs expressed independent of noncovalent antigen binding forces developed over the course of the immune response. This conclusion is consistent with our previous report (25) that the catalytic triad of a proteolytic Ab light chain is encoded by a germ line V gene.
Both Ab subunits of IgG displayed covalent binding of hapten phosphonate diester CRA I, consistent with studies in which catalytic Ser nucleophiles have been identified in the light (8,10) and heavy chains (12,26). The study of recombinant scFv clones confirmed the presence of nucleophilic sites in the V domains. The nucleophiles are located within or in the immediate vicinity of the antigen binding site as suggested by observations of improved covalent binding of antigen-specific Abs to protein CRA IV and peptidyl CRA Va. We did not examine the presence of nucleophilic sites in the constant domains, as the present study was conducted in the context of catalytic activity attributed to the V domains. As the genes encoding the V and constant domains express certain sequence identities (27), the existence of constant domain nucleophiles can not be excluded. Notwithstanding their impressive nucleophilic reactivity, the rate of catalysis by Abs is limited. Presumably, this is because of energetic barriers associated with the deacylation and product release steps (Fig. 1). This statement does not conflict with observations of correlated proteolysis and nucleophilicity of the scFv clones, because increased accumulation of the acylated reaction intermediate will accelerate proteolysis according to the laws of mass action regardless of limitations at subsequent steps in the reaction cycle. In addition to proteases, diverse enzymes involved in chemical transformation of lipids, carbohydrates, and nucleic acids owe their catalytic power to covalent mechanisms (28 -30). Some of these enzymes are reported to react with phosphonate probes (31). An aldolase Ab has been raised by immunization with a phosphonate diester hapten (32), but its relationship with innate Ab nucleophilicity is unclear. In addition to protease and esterase activities, Abs express nuclease (4), peroxidase (33), and kinase (34) activities. Conceivably, nucleophilic Ab reactivity described here may play a role in these reactions.
Specific polyclonal and monoclonal Abs to EGFR and synthetic gp120-(421-436) peptide displayed covalent binding to the CRA analogs of these antigens (CRAs IV and Va, respectively) at levels substantially greater than non-immune IgG, indicating that the nucleophiles express their reactivity in coordination with non-covalent antigen binding interactions. Therefore, non-covalent Ab-antigen binding may be interpreted as a mechanism that permits more efficient delivery of the electrophiles (phosphonate groups) to the Ab nucleophiles. The cognate antigens devoid of phosphonate diester groups inhibited the covalent reaction, suggesting spatial proximity be-tween the nucleophile and residues at which non-covalent binding takes place. The following conditions must be met to explain the experimentally observed antigen-specific formation of the CRA adducts. (a) The germ line-encoded nucleophiles must be retained in the Ab combining sites, or novel nucleophiles must generated over the course of adaptive Ab specialization. (b) A mechanism must be available to allow an improved approach of the Ab nucleophile within covalent binding distance of the phosphonate probe. Precise spatial alignment of Ab nucleophiles in register with the phosphonate groups in CRAs IV and Va is unlikely because the Abs were raised by immunization with polypeptides that do not contain these groups. Conversely, the phosphonate electrophiles were placed at the side chain Lys residues of protein CRA IV and the C terminus of peptide CRA Va without foreknowledge of the spatial relationship between the non-covalent and nucleophilic binding sites in the Abs. These considerations suggest that the nucleophiles enjoy sufficient conformational freedom to make contact with imprecisely located phosphonate electrophiles in the antigenic epitope. The mobility of individual amino acids in Ab-combining sites following binding to antigen has been reported by other groups (35,36). Previous epitope mapping and mutagenesis studies (11,37) indicate that the catalytic residues of proteolytic Abs participate minimally in stabilizing the Ab-antigen ground state complex, suggesting that the mobility of the nucleophile is not restricted by non-covalent binding interactions. Further support for this model is available from observations that monoclonal Abs to vasoactive intestinal peptide (38) and gp41 (8) can cleave multiple peptide bonds in these antigens, presumably by the formation of alternate transition states in which the nucleophile is free to initiate attack on spatially neighboring peptide bonds.
Adaptive improvement in the rate of catalysis by Abs is limited by the mechanisms responsible for clonal selection of B cells. If product release exceeds the rate of transmembrane signaling by the B cell receptor, which is necessary to stimulate cell division, cellular proliferation will cease. On the other hand, there is no bar to adaptive improvement of Ab nucleophilicity as suggested by the results of this study. The improved nucleophilic reactivities of antigen-specific Abs described here results from routine immunization with polypeptides. It is difficult to ascribe the reactivity to a fortuitous immunological phenomenon, because it was observed in polyclonal Abs directed to two different antigens and three distinct monoclonal Abs. Nucleophilic attack on the natural counterparts of the phosphonate groups in CRAs IV and Va, i.e. the electrophilic carbonyl groups in the peptide backbone and side chain amides, is predicted to result in the formation of covalent acyl-Ab complexes (Fig. 1), allowing prolonged occupancy of the B cell receptor and favoring emergence of Abs with improved reactivity. Admittedly, the phosphonate diester group in CRAs is more electrophilic than the carbonyl group in proteins antigens, but Ab nucleophilicity is comparable or superior to that of trypsin, suggesting the feasibility of nucleophilic Ab attack on protein antigens. Two examples of Abs with the ability to form irreversible covalent complexes with hapten antigens have been reported previously (39,40), and certain Abs display SDS-resistant binding to albumin. 2 Ab-nucleophilic reactivity could conceivably contribute to Ab antigen binding without formation of stable covalent bonds. For instance, the nucleophilic reaction may lead to a structure with partial covalent character that does not progress to the acyl-Ab complex because no mechanism is available to donate a proton to the nitrogen atom of the leaving group (C-terminal peptide fragment in Fig. 1; ammonia is released if attack occurs on side chain amide groups). Important biological effects have been ascribed to the proteolytic activity of Abs found in autoimmune, alloimmune, and lymphoproliferative disease (41), e.g. interference with the immunoregulatory (42) and smooth muscle relaxant effects (43) of the neuropeptide vasoactive intestinal peptide. In view of enhanced covalent Ab binding of phosphonate diester groups facilitated by non-covalent binding interactions, peptidyl and proteinic CRAs may be hypothesized to permit permanent and selective blockade of the catalytic activity. Moreover, to the extent that expression of nucleophilicity coordinated with noncovalent antigen binding is a universal Ab characteristic, CRA inhibition may be a generally useful means to inhibit Ab biological effects regardless of catalytic activity. CRAs IV and Va, for instance, may be used to study the functional roles of Abs from patients with systemic sclerosis and lupus, which are reported to bind EGFR (44) and synthetic gp120-(421-436) (45), respectively.